Ocular Periphery and Disorders

OCULAR PERIPHERY AND DISORDERS This page intentionally left blank OCULAR PERIPHERY AND DISORDERS EDITORS DARLENE A...

149 downloads 2313 Views 23MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

OCULAR PERIPHERY AND DISORDERS

This page intentionally left blank

OCULAR PERIPHERY AND DISORDERS EDITORS

DARLENE A. DARTT Harold F. Johnson Research Scholar Schepens Eye Research Institute Harvard Medical School Boston, MA USA

REZA DANA Massachusetts Eye and Ear Infirmary Schepens Eye Research Institute Harvard Medical School Boston, MA USA

PETER BEX Schepens Eye Research Institute Harvard Medical School Boston, MA USA

LINDA K. MCLOON Department of Ophthalmology University of Minnesota Minneapolis, Minnesota USA

PATRICIA D’AMORE Schepens Eye Research Institute Harvard Medical School Boston, MA USA

JERRY Y. NIEDERKORN UT Southwestern Medical Center at Dallas Dallas, Texas USA

Amsterdam • Boston • Heidelberg • London • New York • Oxford Paris • San Diego • San Francisco • Singapore • Sydney • Tokyo Academic Press is an imprint of Elsevier

Academic Press is an imprint of Elsevier 525 B Street, Suite 1900, San Diego, CA 92101-4495, USA The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, UK Copyright # 2011 Elsevier Ltd. All rights reserved Material in the text originally appeared in the Encyclopedia of the Eye, edited by Darlene A. Dartt, Joseph C. Beshare and Reza Dana (Elsevier Ltd. 2010) No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means electronic, mechanical, photocopying, recording or otherwise without the prior written permission of the publisher Permissions may be sought directly from Elsevier’s Science & Technology Rights Department in Oxford, UK: phone (+44) (0) 1865 843830; fax (+44) (0) 1865 853333; email: [email protected]. Alternatively you can submit your request online by visiting the Elsevier web site at (http://elsevier.com/locate/permissions), and selecting Obtaining permission to use Elsevier material Notice No responsibility is assumed by the publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein. Because of rapid advances in the medical sciences, in particular, independent verification of diagnoses and drug dosages should be made British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloging-in-Publication Data A catalog record for this book is available from the Library of Congress ISBN: 978-0-12-382042-6 For information on all Academic Press publications visit our website at www.elsevierdirect.com Printed and bound in Spain 11 12 13

10 9 8 7 6 5 4 3 2 1

CONTENTS

Contributors Introduction

I.

ix–xii xiii–xiv

EXTRAOCULAR AND EYELID MUSCLES: STRUCTURE, FUNCTION, AND PATHOPHYSIOLOGY

Eyelid Anatomy and the Pathophysiology of Blinking

C Evinger

Differentiation and Morphogenesis of Extraocular Muscles Extraocular Muscles: Extraocular Muscle Anatomy

3

D M Noden

L K McLoon and S P Christiansen

Extraocular Muscles: Extraocular Muscle Metabolism Extraocular Muscles: Proprioception and Proprioceptors

9 17

F H Andrade

27

R Blumer

33

Abnormal Eye Movements due to Disease of the Extraocular Muscles and Their Innervation A Serra and R J Leigh

39

Extraocular Muscles: Functional Assessment in the Clinic

45

II.

S P Christiansen and L K McLoon

STRUCTURE AND FUNCTION OF THE TEAR FILM, OCULAR ADNEXA, CORNEA AND CONJUNCTIVA IN HEALTH AND PATHOGENESIS IN DISEASE

Tear Film

J P Craig, A Tomlinson, and L McCann

Meibomian Glands and Lipid Layer Lacrimal Gland Overview

51

T J Millar, P Mudgil, and S Khanal

60

M C Edman, R R Marchelletta, and S F Hamm-Alvarez

68

Lacrimal Gland Hormone Regulation Lacrimal Gland Signaling: Neural

A K Mircheff, D W Warren, and J E Schechter D Zoukhri

83

Lids: Anatomy, Pathophysiology, Mucocutaneous Junction

T Wojno

Overview of Electrolyte and Fluid Transport Across the Conjunctiva Conjunctival Goblet Cells

74

R R Hodges and D A Dartt

O A Candia and L J Alvarez

91 99 108

Ocular Mucins

M Berry

116

Tear Drainage

F P Paulsen and L Bra¨uer

126

P Asbell and D Brocks

133

Cornea Overview

Corneal Epithelium: Cell Biology and Basic Science

M A Stepp

143

v

vi

Contents

Corneal Nerves: Anatomy Corneal Nerves: Function

C F Marfurt C Belmonte

150 158

Corneal Epithelium: Wound Healing Junctions, Attachment to Stroma Receptors, Matrix Metalloproteinases, Intracellular Communications G M Gordon and M E Fini

164

Corneal Epithelium: Transport and Permeability

171

Stem Cells of the Ocular Surface

P S Reinach, F Zhang, and J E Capo´-Aponte

Y Du and J L Funderburgh

178

The Surgical Treatment for Corneal Epithelial Stem Cell Deficiency, Corneal Epithelial Defect, and Peripheral Corneal Ulcer N Koizumi and S Kinoshita

185

Refractive Surgery

193

S Marcos, L Llorente, C Dorronsoro, and J Merayo-Lloves

Refractive Surgery and Inlays Contact Lenses

R M M A Nuijts, M Doors, N G Tahzib, and L P J Cruysberg

N Carnt, Y Wu, and F Stapleton

Imaging of the Cornea

201 207

S C Kaufman, M Fung, D Raja, and N Kramarevsky

213

The Corneal Stroma

J L Funderburgh

219

Corneal Dystrophies

B H Feldman and N A Afshari

226

Corneal Imaging: Clinical Corneal Scars

S Garg and R F Steinert

D G Dawson

Corneal Endothelium: Overview

256 D R Whikehart

Regulation of Corneal Endothelial Function

272

J A Bonanno and S P Srinivas

Regulation of Corneal Endothelial Cell Proliferation Artificial Cornea

234

283

Q Lu, T A Fuchsluger, and U V Jurkunas

M A Rafat, J M Hackett, P Fagerholm, and M Griffith

290 296

Drug Delivery to Cornea and Conjunctiva: Esterase- and Protease-Directed Prodrug Design R S Talluri, S Hariharan, P K Karla, and A K Mitra

303

Knock-Out Mice Models: Cornea, Conjunctiva, Eyelids and Lacrimal Gland C-Y Liu, and H Liu

315

Gene Therapy for the Cornea, Conjunctiva, and Lacrimal Gland C Siddappa, and R R Mohan

III.

W W-Y Kao,

A Sharma, A Ghosh, 327

IMMUNE REGULATION OF THE CORNEA AND CONJUNCTIVA AND ITS DYSREGULATION IN DISEASE

Adaptive Immune System and the Eye: Mucosal Immunity

A K Mircheff

Adaptive Immune System and the Eye: T Cell-Mediated Immunity R D Vicetti Miguel Innate Immune System and the Eye

339

K C McKenna and 347

M S Gregory

354

Dynamic Immunoregulatory Processes that Sustain Immune Privilege in the Eye

J Y Niederkorn

361

Immunosuppressive and Anti-Inflammatory Molecules that Maintain Immune Privilege of the Eye A W Taylor

367

Antigen-Presenting Cells in the Eye and Ocular Surface

373

Dry Eye: An Immune-Based Inflammation Penetrating Keratoplasty

P Hamrah and R Dana

M E Stern and S C Pflugfelder

T H Flynn and D F P Larkin

Immunopathogenesis of HSV Keratitis

390

K Buela, G Frank, J Knicklebein, and R Hendricks

Immunopathogenesis of Onchocerciasis (River Blindness)

381

E Pearlman and K Gentil

396 401

Contents

Immunopathogenesis of Pseudomonas Keratitis L D Hazlett Immunobiology of Acanthamoeba Keratitis J Y Niederkorn Molecular and Cellular Mechanisms in Allergic Conjunctivitis

vii

406 413 V L Calder

419

Pathogenesis of Fungal Keratitis E Pearlman, S Leal, A Tarabishy, Y Sun, L Szczotka-Flynn, Y Imamura, P Mukherjee, J Chandra, M Momany, S Hastings-Cowden, and M Ghannoum

426

Conjunctiva Immune Surveillance

431

E Knop and N Knop

Defense Mechanisms of Tears and Ocular Surface Corneal Epithelium: Response to Infection

A M McDermott

Elizabeth A Szliter-Berger and L D Hazlett

444 452

Inflammation of the Conjunctiva

T Nishida

459

Concept of Angiogenic Privilege

B Regenfuss and C Cursiefen

465

Corneal Angiogenesis

M S Cortina and D T Azar

Avascularity of the Cornea

IV.

R J C Albuquerque and J Ambati

470 478

VISUAL ACUITY RELATED TO THE CORNEA AND ITS DISORDERS

Pupil

P D R Gamlin and D H McDougal

Acuity

M D Crossland

Contrast Sensitivity Astigmatism Myopia

P Bex

487 494 500

M J Cox

506

F A Vera-Diaz

517

Amblyopia

D M Levi

525

Hyperopia

E Harb

529

Index

535

This page intentionally left blank

CONTRIBUTORS

N A Afshari Duke University Medical Center, Durham, NC, USA

V L Calder UCL Institute of Ophthalmology, London, UK

R J C Albuquerque University of Kentucky, Lexington, KY, USA

O A Candia Mount Sinai School of Medicine, New York, NY, USA

L J Alvarez Mount Sinai School of Medicine, New York, NY, USA

J E Capo´-Aponte U.S. Army Aeromedical Research Laboratory (USAARL), Fort Rucker, AL, USA

J Ambati University of Kentucky, Lexington, KY, USA

N Carnt Institute for Eye Research, Sydney, NSW, Australia

F H Andrade University of Kentucky Medical Center, Lexington, KY, USA

J Chandra Case Western Reserve University, Cleveland, OH, USA

P Asbell Mount Sinai Hospital, Department of Ophthalmology, New York, NY, USA D T Azar University of Illinois at Chicago, Chicago, IL, USA C Belmonte Instituto de Neurociencias de Alicante, Universidad Miguel Herna´ndez-Consejo Superior de Investigaciones Cientı´ficas, San Juan de Alicante, Spain M Berry Bristol Eye Hospital, Bristol, UK P Bex Schepens Eye Research Institute, Boston, MA, USA R Blumer Medical University of Vienna, Vienna, Austria J A Bonanno Indiana University, Bloomington, IN, USA D Brocks Mount Sinai Hospital, Department of Ophthalmology, New York, NY, USA L Bra¨uer Martin Luther University Halle-Wittenberg, Halle, Germany K Buela University of Pittsburgh School of Medicine, Pittsburgh, PA, USA

S P Christiansen Boston University School of Medicine, Boston, MA, USA M S Cortina University of Illinois at Chicago, Chicago, IL, USA M J Cox University of Bradford, Bradford, UK J P Craig University of Auckland, Auckland, New Zealand M D Crossland UCL Institute of Ophthalmology/Moorfields Eye Hospital, London, UK L P J Cruysberg University Hospital Maastricht, Maastricht, The Netherlands C Cursiefen Friedrich-Alexander University Erlangen-Nuernberg, Erlangen, Germany R Dana Harvard Medical School, Boston, MA, USA D A Dartt Schepens Eye Research Institute, Boston, MA, USA D G Dawson Emory University School of Medicine, Atlanta, GA, USA M Doors University Hospital Maastricht, Maastricht, The Netherlands

ix

x

Contributors

C Dorronsoro Consejo Superior de Investigaciones Cientı´ficas, Madrid, Spain

S F Hamm-Alvarez University of Southern California School of Pharmacy, Los Angeles, CA, USA

Y Du University of Pittsburgh, Pittsburgh, PA, USA

P Hamrah Harvard Medical School, Boston, MA, USA

M C Edman University of Southern California School of Pharmacy, Los Angeles, CA, USA

E Harb New England College of Optometry, Boston, MA, USA

C Evinger SUNY Stony Brook, Stony Brook, NY, USA

S Hariharan University of Missouri–Kansas City, Kansas City, MO, USA

P Fagerholm Linko¨ping University Hospital, Linko¨ping, Sweden B H Feldman Philadelphia Eye Associates, Philadelphia, PA, USA M E Fini University of Southern California, Los Angeles, CA, USA

S Hastings-Cowden University of Athens, Athens, GA, USA L D Hazlett Wayne State University School of Medicine, Detroit, MI, USA

T H Flynn Moorfields Eye Hospital, London, UK

R Hendricks University of Pittsburgh School of Medicine, Pittsburgh, PA, USA

G Frank University of Pittsburgh School of Medicine, Pittsburgh, PA, USA

R R Hodges Schepens Eye Research Institute, Boston, MA, USA

T A Fuchsluger Schepens Eye Research Institute, Boston, MA, USA J L Funderburgh University of Pittsburgh, Pittsburgh, PA, USA

Y Imamura Case Western Reserve University, Cleveland, OH, USA U V Jurkunas Schepens Eye Research Institute, Boston, MA, USA

M Fung University of Minnesota, Minneapolis, MN, USA

W W-Y Kao University of Cincinnati, Cincinnati, OH, USA

P D R Gamlin University of Alabama at Birmingham, Birmingham, AL, USA

P K Karla University of Missouri–Kansas City, Kansas City, MO, USA

S Garg University of California, Irvine, Irvine, CA, USA

S C Kaufman University of Minnesota, Minneapolis, MN, USA

K Gentil University of Bonn, Bonn, Germany

S Khanal University of Western Sydney, NSW, Australia

M Ghannoum Case Western Reserve University, Cleveland, OH, USA

S Kinoshita Kyoto Prefectural University of Medicine, Kyoto, Japan

A Ghosh University of Missouri–Columbia, Columbia, MO, USA G M Gordon University of Southern California, Los Angeles, CA, USA M S Gregory Schepens Eye Research Institute, Harvard Medical School, Boston, MA, USA

J Knicklebein University of Pittsburgh School of Medicine, Pittsburgh, PA, USA E Knop Charite´ – Universita¨tsmedizin Berlin, Berlin, Germany N Knop Hannover Medical School, Hannover, Germany

M Griffith University of Ottawa Eye Institute, Ottawa, ON, Canada

N Koizumi Kyoto Prefectural University of Medicine, Kyoto, Japan

J M Hackett University of Ottawa, Ottawa, ON, Canada

N Kramarevsky University of Minnesota, Minneapolis, MN, USA

Contributors D F P Larkin Moorfields Eye Hospital, London, UK

R R Mohan University of Missouri–Columbia, Columbia, MO, USA

S Leal Case Western Reserve University, Cleveland, OH, USA

M Momany University of Athens, Athens, GA, USA

R J Leigh Case Western University, Cleveland, OH, USA D M Levi University of California, Berkeley, Berkeley, CA, USA C-Y Liu University of Cincinnati, Cincinnati, OH, USA H Liu University of Cincinnati, Cincinnati, OH, USA L Llorente Consejo Superior de Investigaciones Cientı´ficas, Madrid, Spain Q Lu Schepens Eye Research Institute, Boston, MA, USA R R Marchelletta University of Southern California School of Pharmacy, Los Angeles, CA, USA S Marcos Consejo Superior de Investigaciones Cientı´ficas, Madrid, Spain C F Marfurt Indiana University School of Medicine – Northwest, Gary, IN, USA L McCann Glasgow Caledonian University, Glasgow, UK A M McDermott University of Houston, Houston, TX, USA D H McDougal University of Alabama at Birmingham, Birmingham, AL, USA K C McKenna University of Pittsburgh, Pittsburgh, PA, USA L K McLoon University of Minnesota, Minneapolis, MN, USA J Merayo-Lloves Consejo Superior de Investigaciones Cientı´ficas, Madrid, Spain T J Millar University of Western Sydney, NSW, Australia A K Mircheff University of Southern California, Los Angeles, CA, USA A K Mitra University of Missouri–Kansas City, Kansas City, MO, USA

xi

P Mudgil University of Western Sydney, NSW, Australia P Mukherjee Case Western Reserve University, Cleveland, OH, USA J Y Niederkorn University of Texas Southwestern Medical Center, Dallas, TX, USA T Nishida Yamaguchi University Graduate School of Medicine, Yamaguchi, Japan D M Noden Cornell University, Ithaca, NY, USA R M M A Nuijts University Hospital Maastricht, Maastricht, The Netherlands F P Paulsen Martin Luther University Halle-Wittenberg, Halle, Germany E Pearlman Case Western Reserve University, Cleveland, OH, USA S C Pflugfelder Baylor College of Medicine, Houston, TX, USA M A Rafat University of Ottawa Eye Institute, Ottawa, ON, Canada D Raja University of Minnesota, Minneapolis, MN, USA B Regenfuss Friedrich-Alexander University Erlangen-Nuernberg, Erlangen, Germany P S Reinach The State University of New York, New York, NY, USA J E Schechter University of Southern California, Los Angeles, CA, USA A Serra University of Sassari, Sassari, Italy A Sharma University of Missouri–Columbia, Columbia, MO, USA C Siddappa University of Missouri–Columbia, Columbia, MO, USA S P Srinivas Indiana University, Bloomington, IN, USA F Stapleton University of New South Wales, Sydney, NSW, Australia

xii

Contributors

R F Steinert University of California, Irvine, Irvine, CA, USA

A W Taylor Schepens Eye Research Institute, Boston, MA, USA

M A Stepp The George Washington University Medical Center, Washington, DC, USA

A Tomlinson Glasgow Caledonian University, Glasgow, UK

M E Stern Allergan Inc, Irvine, CA, USA

F A Vera-Diaz Schepens Eye Research Institute, Harvard Medical School, Boston, MA, USA

Y Sun Case Western Reserve University, Cleveland, OH, USA

R D Vicetti Miguel University of Pittsburgh, Pittsburgh, PA, USA

L Szczotka-Flynn Case Western Reserve University, Cleveland, OH, USA

D W Warren University of Southern California, Los Angeles, CA, USA

Elizabeth A Szliter-Berger Wayne State University School of Medicine, Detroit, MI, USA

D R Whikehart The University of Alabama at Birmingham, Birmingham, AL, USA

N G Tahzib University Hospital Maastricht, Maastricht, The Netherlands

T Wojno The Emory Clinic, Atlanta, GA, USA

R S Talluri University of Missouri–Kansas City, Kansas City, MO, USA A Tarabishy Case Western Reserve University, Cleveland, OH, USA

Y Wu Institute for Eye Research, Sydney, NSW, Australia F Zhang The State University of New York, New York, NY, USA D Zoukhri Tufts University, Boston, MA, USA

INTRODUCTION

Protection of the entire eye from the external environment and maintenance of a clear optical pathway through the aqueous humor, lens, and vitreous to the retina are the functions of the ocular periphery. The outermost portion of the periphery is the eyelids that protect the eye through blinking and preserve visual acuity through the movement of the eye by the exceedingly specialized extraocular muscles. The next layer of protection is the tear film, secreted by the ocular adnexa, and the epithelia of the ocular surface, composed of the cornea and conjunctiva. The tears and ocular surface epithelia protect the eye through numerous coordinating layers of structural and functional mechanisms. The tears and cornea also must retain their transparency and maintain a smooth optical surface. Dysfunction and dysregulation of the ocular periphery in disease can compromise the entire visual system and lead to loss of visual acuity, inflammation, and infection, thus jeopardizing the function of the entire eye and, in severe cases, cause loss of vision. This book focuses on both the normal functioning of the tissues of the ocular periphery and their pathophysiology in disease. This volume provides a unique collection of chapters on the multiple, diverse tissues that comprise the ocular periphery and function to protect vision. The goal of this book is to provide a comprehensive and contemporary review of the structure and function of the ocular periphery in health and disease. The book is organized into four sections including I. Extraocular and Eyelid Muscles: Structure, Function, and Pathophysiology, II. Structure and Function of the Tear Film, Ocular Adnexa, Cornea, and Conjunctiva in Health and Pathogenesis in Disease, III. Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease, and IV. Visual Acuity Related to the Cornea and Its Disorders. Section I is devoted to the extraocular muscles and the eye lids, the muscles that move the eye and the lids. Chapters include several chapters on the anatomy, function, metabolism, and pathophysiology of the specialized extraocular muscles, along with a chapter on eyelid function and pathophysiology. Final chapters discuss diseases of the extraocular muscles and the clinical diagnosis of the dysfunction of these muscles. The authors have particularly highlighted the special features of the extraorbital muscles that allow them to function without fatiguing, unlike other skeletal muscles as well as specific diseases that preferentially spare or involve these muscles critical for binocular vision. Section II focuses on two major areas: first, the tear film and the tissues that secrete it and second, the epithelia (cornea and conjunctiva) that form the anterior surface of the eye. Initial chapters in this section focus on the multiple components of the complex tear film and the mechanisms by which they are secreted by the meibomian gland, lacrimal gland, and conjunctiva. The next chapters discuss the structure and function of the conjunctiva and the three layers of the cornea, each of these layers with its specialized functions to maintain the clarity of the cornea while protecting the eye from mechanical, thermal, chemical, and pathogenic challenges of the external environment. Additional chapters in this section focus on corneal disease as well as new modalities for understanding ocular surface dysfunction and repairing this dysfunction. Section III highlights the unique mechanisms that the cornea uses to respond to immune and infectious challenges. Multiple chapters focus on different aspects of the immune and angiogenic privilege of the cornea that is unique to this tissue, and compare it with the more ‘‘normal’’ response of the conjunctiva. The many facets of the complex immune response of the cornea are presented. In addition, the multiple mechanisms that are responsible for the avascularity of the cornea are discussed. Another theme of this section is inflammation and its involvement with dry eye disease and infectious diseases of the ocular surface. All together, sections I, II, and III present the multiple layers of structure and diverse overlapping mechanisms that are in place to prevent breach of the interior of the eye by mechanical, thermal, chemical, and pathogenic threats from the external environment and the diseases that result when these defenses are overwhelmed. Section IV is devoted to the cornea, but in this section, the visual acuity of the cornea is highlighted. Chapters discuss normal visual optics and the conditions that result from changes in corneal shape that disrupt the visual axis and lead to decreased vision.

xiii

xiv

Introduction

Each chapter contains text readable to a scientist outside the field of the article, multiple multimedia and color figures to illustrate the most important points of the chapter, and a list of references to provide more in-depth information. Chapters are primarily directed at scientists looking for an entry point into a field tangential to their specialty as well as at graduate students and postdoctoral fellows in eye research. The chapters will be especially useful to scientists designing introductory or generalized courses that cover diverse fields of eye research. Scientists writing review articles or chapters will also find the book’s chapters especially useful as a starting point. The many introductory chapters are written at the level to be understood by undergraduates at university and public libraries, but include enough information to satisfy the more advanced needs of graduate students and postdoctoral fellows. The in-depth chapters on more focused research areas are ideal for postdoctoral fellows and experienced scientists. The plentiful illustrations will be especially helpful in understanding the more complicated points as well as illustrating basic processes and anatomy. Finally, I thank the other editors of this book who chose the chapters to be included in this book as well as the chapter authors who devoted considerable time to proof-reading these articles. I also thank all the authors for their excellent contributions and Robin R. Hodges for her excellent managerial assistance. Darlene A. Dartt, Ph.D. Boston, MA USA October 24, 2010

I. EXTRAOCULAR AND EYELID MUSCLES: STRUCTURE, FUNCTION, AND PATHOPHYSIOLOGY

This page intentionally left blank

Eyelid Anatomy and the Pathophysiology of Blinking C Evinger, SUNY Stony Brook, Stony Brook, NY, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Blepharospasm – A dystonic muscle contraction disorder characterized by forceful, bilateral, and uncontrolled closure of the eyelids. Hemifacial spasm – A muscle contraction disorder characterized by forceful uncontrolled contraction of the facial muscles on one side of the face. Levator palpebrae superioris muscle – A skeletal muscle innervated by cranial motor nerve 3 whose primary function is elevation of the upper eyelid. Mu¨ller’s muscle – A smooth muscle that runs from the inferior surface of the tendon of the levator palpebrae superioris to insert into the tarsal plate. It receives sympathetic nervous system innervation. Omnipause neurons – Part of the brainstem neural circuit that controls saccades. These neurons fire during fixation and cease firing before and during saccadic movements. Stimulation of omnipause neurons interrupts saccades. Orbicularis oculi muscle – A circumferential muscle of facial expression innervated by the facial nerve that lies just deep to the skin within both eyelids as well as the surrounding bones of the orbital margin. Its main function is closure of the eyelid. Paresis – Partial paralysis of a skeletal muscle resulting in muscle weakness. Retractor bulbi muscle – Skeletal muscles innervated by cranial nerve 6 whose function is to retract the eyeball into the orbit causing movement of the nictitating membrane over the surface of the eye. Saccades – Rapid eye movements that redirect the line of sight so that the image of interest falls on the fovea of the retina. Vergence – The coordinated movements of both eyes in opposite directions in order to maintain binocular vision.

Organization of the Eyelid System To understand the neural control of eyelids, the basis of neurological diseases affecting eyelid control, and how the

eyelids protect the eye, think about the evolutionary origins of eyelids. For fish, a class of vertebrates without eyelids, eye protection primarily requires avoiding objects hitting and damaging the cornea. To avoid objects hitting the eyeball, fish retract their eyes into the orbit by co-contracting their extraocular muscles. Thus, cornea protection was initially linked to neural circuits whose primary goal was to move the eye. When vertebrates moved onto the land, the development of eyelids was a critical step in reducing the dehydrating effects of air on the cornea. Although goblet cells, lacrimal and meibomian glands produce the fluids to coat the corneal surface, it is blinking of the eyelid that spreads the tears to restore the tear film, which maintains corneal hydration. In addition, blinking removes small objects from the surface of the eye and provides some protection from objects getting into the eye. Although blinking is essential for maintaining the cornea, lid closure has the undesirable side effect of blocking vision. Thus, an eyelid control system must generate blinks that minimally disrupt vision while adequately protecting the cornea. The nervous system deals with this constraint by developing fast eyelid closure and opening without carefully controlling absolute eyelid position. The other restriction on the nervous system’s management of the eyelids is that they must move synchronously with vertical eye movements to avoid blocking vision. Overcoming this problem requires the nervous system to control eyelid position accurately. The eyelid control system accomplishes this feat by linking itself to the eye movement system. The melding of the eyelid system with the eye movement system reveals itself first in the anatomical organization of the eyelids. Only four forces act on the upper eyelid (Figures 1(a) and 2). (1) The phasically active orbicularis oculi (OO) muscle actively closes the eyelid. The ipsilateral facial (VII) nucleus innervates the OO muscle. (2) The tonically active levator palpebrae superioris (LP) muscle actively elevates the upper eyelid. LP innervation arises from the oculomotor (III) nucleus. (3) Raising the eyelid stretches the superior transverse (Whitnall’s) ligament and the lateral and medial canthal tendons to create a passive downward force. Thus, the lowest energy state for the eyelid is closed. (4) Mu¨ller’s smooth muscle (Figure 2), which bridges the belly and the tendon of the LP, raises the eyelid approximately 1.5 mm with sympathetic activation. Post-ganglionic nerves from the superior cervical ganglion innervate Mu¨ller’s muscle. The interaction between the first three forces (OO, LP, and passive downward forces) enables

3

4

Extraocular and Eyelid Muscles: Structure, Function, and Pathophysiology

Human Pos

Closed

Open

LP

WL 10 deg

APO CT

100 ms

OOemg

(a) (c) SO Guinea pig (Pos)

Pos

R1

2 deg

LP

100 ms R2 OOemg (b)

OOemg (d) SO

Figure 1 Forces acting on the eyelid and mammalian blinking. (a) Opening the eye stretches the levator palpebrae (LP) aponeurosis (APO), the superior transverse ligament (Whitnall’s ligament, WL), and the medial and lateral canthal tendons (CT) to create passive downward forces. The lowest energy state for the eyelid system is closed. (b) Lid (Pos) lowering during a blink results from a transient relaxation of the LP followed by a phasic activation of the orbicularis oculi (OO) muscle. Raising the eyelid occurs as the LP resumes its tonic activity following the completion of OO activity. Gray indicates LP activity and black indicates OO activity. (c, d) Individual examples of a reflex blink evoked by stimulation of the supraorbital branch of the trigeminal nerve (SO ▲) for a human (c) and a guinea pig (d). R1 is the short latency response and R2 is the long latency response seen after nerve stimulation. Abbreviations: OOemg, orbicularis oculi EMG; Pos, upper eyelid position.

the eyelids to blink rapidly yet accurately match the vertical position of the eyeball. The characteristic rapid eyelid closure of a blink followed by lid opening at approximately half the speed of lid closure follows directly from the anatomical organization of the eyelid (Figures 1(b)–1(d)). A blink begins with relaxation of the tonically active LP muscle. LP relaxation releases the passive downward forces to initiate lid closure. The phasically active OO muscle discharge combines with the passive downward forces to lower the eyelid rapidly. As the OO muscle relaxes, the LP muscle slowly resumes its tonic discharge. This LP contraction raises the upper eyelid. Eyelid elevation is slower than lid closure because the LP muscle must work against the passive downward forces. The point at which the tonically active LP force matches the passive downward force created by tendon and ligament stretching determines final lid position. This anatomical organization is conserved so that the pattern of blinking is similar among mammals (Figures 1(c) and 1(d)).

In contrast to the interactions between the OO and LP muscles and passive downward forces with blinking, the coordination of eyelid motion with vertical eye movement arises from the antagonistic interactions between the LP muscle and the passive downward forces. The linkage with eye movements occurs because the LP behaves like the superior rectus muscle, which rotates the eye upward. Embryologically, the LP muscle arises from the superior rectus muscle, and LP motor neurons are always adjacent to superior rectus motor neurons in the oculomotor nucleus. LP and superior rectus motor neurons exhibit similar patterns of activity except during a blink. The tonic firing frequency of superior rectus and LP motor neurons correlates with vertical eye position. With an upward saccadic eye movement, superior rectus and LP motor neurons generate a burst of action potentials followed by an increased tonic firing frequency that holds the eye in the new elevated position. A downward saccade results from a cessation of tonic activity followed by a lower frequency tonic discharge to hold the eye in the

Eyelid Anatomy and the Pathophysiology of Blinking

5

Epithelium

Fornix Facial nerve

Preseptal region of orbicularis oculi muscle

Müller’s muscle Pretarsal region of orbicularis oculi muscle

Tarsal glands Conjunctiva Figure 2 Montage of a sagittal section of the eyelid from an adult rabbit stained with hematoxylin and eosin. Courtesy of Dr. Linda K. McLoon.

depressed position. When the LP motor neurons transiently cease discharging during a downward saccadic eye movement, unopposed passive downward forces lower the eyelid. When the LP motor neuron resumes its activity at a lower tonic firing frequency, the new balance point between passive downward forces and active upward LP muscle force establishes the final eyelid position. With an upward eye movement, the increased LP motor neuron firing frequency pulls the eyelid upward until the LP muscle and passive downward forces match. Although it seems counter-intuitive that passive downward forces rather than the OO muscle act as the antagonist to the LP muscle with eye movements, it is clear that the OO does not participate in eyelid movement with vertical saccades. For example, individuals with OO denervation created by seventh nerve palsy exhibit nearly normal saccadic lid movements with vertical saccadic eye movements, but abnormally slow blinks. Further evidence of the evolutionary linkage of blinking with the oculomotor system is that blinks frequently occur with saccadic eye movements. These gaze-evoked blinks most commonly accompany large saccades to visual targets that do not have a strong behavioral significance. The advantage of combining blinks with saccades is that visual suppression occurs during both blinks and saccades. A gaze-evoked blink refreshes the tear film,

but does not produce more loss of vision than the saccadic eye movement. The evolutionary linkage also appears in the eye movements associated with blinking, blink-evoked eye movements. When looking straight ahead, there is an adducting and downward rotation of the eye with each blink. Nevertheless, the state of the eyelid system determines the trajectory of blink-evoked eye movements. These movements exhibit an upward trajectory in both eyes and are smaller than normal in individuals with a unilateral seventh nerve palsy. Eyeball retraction is also a component of these blink-evoked eye movements. For mammals, the eyeball retraction with blinking results from extraocular muscle co-contraction and contraction of the retractor bulbi muscle. This accessory extraocular muscle is innervated by motor neurons in the accessory abducens nucleus that send their axons to the orbit as part of the VIth nerve. Extraocular muscle co-contraction with blinking in mammals appears to reflect the evolutionary origins of the eyelid control system from eye retraction of fish. Despite blink-evoked eye movements accompanying all blinks, neither gaze-evoked nor reflex blinks occurring with a saccade prevent the eye from achieving its desired endpoint. With a simultaneous blink and saccade, the eyes follow a complex trajectory instead of the nearly straight path of a saccadic eye movement alone. This complex

6

Extraocular and Eyelid Muscles: Structure, Function, and Pathophysiology

trajectory results in a significantly slower saccade than when the saccade occurs alone. The linkage between blinking and saccadic eye movements indicates that blinking interacts with the brainstem circuits that generate saccadic eye movements. This interaction is apparent when reflex blinks increase the speed of saccadic eye movements in individuals with abnormally slow saccades, initiate saccade oscillations, and alter the speed of vergence eye movements where the eyes are moving in opposite directions. The neurophysiological mechanism that may underlie these effects is the activity of omnipause neurons, which control fixation, and are associated with reflex blinks. Tonically active omnipause neurons gate saccadic eye movements. Omnipause neurons cease discharging immediately before and during all saccadic eye movements and reflex blinks. Microstimulation of omnipause neurons blocks reflex blinks as well as saccadic eye movements. The evolutionary origins of eyelid control functionally intertwine blinking and saccadic eye movements so that blinks frequently occur with saccadic eye movements, blinking modifies the trajectory of eye movements, and blinking and saccadic eye movements utilize some of the same brainstem neuronal circuits.

Modifiability of the Blink System In order to respond to challenges to the eyelid’s protective function, the neural control of blinking exhibits significant adaptive plasticity. The most common challenge to the protective function of blinking is the development of dry eye. Cornea irritation created by tear film inadequacy rapidly initiates several changes. First, the trigeminal system becomes more excitable, so that a given blinkevoking stimulus elicits a bigger blink than before cornea irritation. In addition to increased blink excitability, the trigeminal reflex blink circuit begins to ‘oscillate’ in response to a blink-evoking stimulus (Figure 3). In this condition, a single stimulus evokes a reflex blink followed

by blink oscillations, one or more additional large blinks. Blink oscillations occur at a constant interblink interval. These modifications compensate for dry eye in at least two ways. The reduced threshold for evoking a blink produces more frequent blinking so that blinks can occur before significant disruptions of the tear film. The larger blinks increase the amount of meibum released over the tear film. Increasing the thickness of the oily meibum layer superficial to the aqueous layer reduces aqueous evaporation. Disruptions of these normally adaptive mechanisms may underlie neurological disorders involving the eyelids, benign essential blepharospasm (BEB) and hemifacial spasm (HFS). We have an outline of the neural basis for these adaptive modifications of the blink reflex initiated by eye irritation. The adaptive changes occur in trigeminal blink circuits, but not at OO motor neurons or reticular neurons receiving bilateral trigeminal inputs. If just one eye experiences irritation, stimulating the ophthalmic branch of the trigeminal nerve on the irritated side elicits adaptive blink modifications in both eyelids. Stimulating the contralateral trigeminal nerve associated with the unaffected eye, however, elicits normal, unadapted blinks in both eyes. Consistent with this result, spinal trigeminal neurons in the corneal-evoked blink circuit discharge before blink oscillations, and this discharge correlates with the pattern of OO activity producing the blink oscillation. Although the adaptive processes clearly involve the trigeminal system, the cerebellum is also essential for blink adaptation. Blink-related neurons in the cerebellum are active with blink adaptation, and lesioning the cerebellum blocks this form of motor learning. A hypothesis for the function of this brainstem–cerebellum circuit is that the cerebellum recognizes an error signal produced by eye irritation and then initiates modifications in spinal trigeminal blink circuits to compensate for the eye irritation. These adaptive modifications in the trigeminal blink circuits may involve long-term potentiation (LTP)- and long-term depression (LTD)-like processes.

Reflex blink

SO

15 deg 250 (ms)

Blink oscillations

Figure 3 Stimulating the supraorbital branch of the trigeminal nerve (SO ▲) evokes a reflex blink followed by additional blinks with a constant interblink interval, blink oscillations. Each upper eyelid position trace is a record from a single trial from an individual with dry eye.

Eyelid Anatomy and the Pathophysiology of Blinking

As discussed in the next section, the cerebellum appears to be an important element in creating the symptoms of dystonic movement disorders such as BEB.

Benign Essential Blepharospasm and Hemifacial Spasm BEB is a focal dystonia characterized by involuntary bilateral spasms of lid closure, trigeminal reflex blink hyperexcitablity, and photophobia (i.e., excessive sensitivity to light). BEB frequently begins with a complaint of ocular discomfort. The available evidence indicates that BEB arises from the confluence of a predisposing and a precipitating factor. Although the predisposing factor has not been identified, the genetic basis for other forms of dystonia suggests that the predisposing factor for BEB is genetic. The precipitating factor appears to be ocular irritation. Consistent with eye irritation as the precipitating factor is that BEB characteristics appear to be an exaggeration of the normally adaptive response to dry eye. For most individuals, the spasms of lid closure in BEB are rapid, repetitive contractions of the OO muscle. This spasm pattern is equivalent to shortening the interblink interval of the blink oscillations developed in response to dry eye (Figure 3). BEB patients exhibit trigeminal reflex blink hyperexcitability. One adaptive response to dry eye is to elevate trigeminal reflex blink excitability, although this increase accompanying dry eye is not as profound as occurs with BEB. The trigeminal hyperexcitability with BEB is sufficient to explain the photophobia. Consistent with the exaggeration of dry eye hypothesis, photophobia is present with dry eye although the light sensitivity is not as debilitating as with BEB. Thus, the exaggerated dry eye hypothesis proposes that BEB begins with the onset of dry eye or significant eye irritation. The nervous system initiates blink modifications to compensate for this irritation, but the genetic predisposition prevents the nervous system from recognizing that the adaptive changes corrected the ocular irritation. The nervous system responds by further increasing these adaptive modifications until the normally compensatory mechanisms becomes so maladaptive as to create the BEB syndrome. Although genetic modifications probably create the predisposing environment for BEB, the genetics underlying this focal dystonia are not yet clear. Most investigators argue that there is an autosomal dominant transmission with reduced penetrance in BEB. One challenge to linking genetics to specific types of dystonia, however, is that individuals with the same genetic mutation may exhibit very different forms of dystonia. For example, individuals with the DYT1 mutation responsible for the most common form of generalized dystonia may exhibit generalized dystonia, focal dystonia, or may be asymptomatic. Individuals with generalized dystonia, focal dystonia, and

7

asymptomatic DYT1 carriers all exhibit a similar pattern of abnormal brain activity. Positron emission tomography (PET) and functional magnetic resonance imaging (fMRI) scans reveal hyperactivity in the basal ganglia, primary motor cortex, supplementary motor area, putamen, thalamus, and cerebellum. In patients with BEB, these areas are also active, but the focus of abnormal activity is in the brainstem and cerebellum. Abnormal cerebellar activity with BEB supports the hypothesis that abnormal blink adaptation involving the cerebellum is critical in the development of BEB. Consistent with the hypothesis that a genetic predisposing factor affects adaptation or motor learning, individuals with BEB exhibit exaggerated LTPlike plasticity of blink circuits relative to normal subjects. There are two clear examples in which exaggeration of the adaptive processes initiated by eye irritation or dry eye leads to BEB. A small number of individuals with Bell’s palsy develop blepharospasm. The ocular irritation produced by incomplete eye closure, the precipitating factor, initiates unabated modifications that produce the BEB syndrome. As predicted by the exaggerated dry eye hypothesis, implanting gold weights in the paretic (weakened) eyelid, enabling nearly complete closure of the paretic eyelid, reduces eye irritation and allows a resolution of BEB signs. Combining predisposing and precipitating factors can also create a BEB-like syndrome in rats. In this model, a chemical lesion reducing approximately 30% of the dopamine neurons in the substantia nigra pars compacta elevates the excitability of the trigeminal blink reflex circuits. This increased trigeminal excitability acts as the predisposing factor. Transient eye irritation produced by crushing a facial nerve branch providing a portion of the OO innervation acts as the precipitating factor. The reduced eyelid mobility created by this procedure causes eye irritation that initiates adaptive blink modifications. Although the eye irritation resolves following nerve regrowth, rats continue to exhibit spasms of lid closure caused by repetitive bursts of OO activity and hyperexcitable trigeminal reflex blinks. Thus, the evidence indicates that BEB occurs in individuals genetically predisposed to the disorder who experience a precipitating condition, ocular irritation. The ocular irritation initiates a series of normally adaptive modifications. In the presence of the predisposing condition which creates an abnormal environment for motor learning, these modifications become exaggerated to create the signs of BEB. HFS begins as spontaneous, unilateral spasms of eyelid closure. Over a period of weeks to months, the spasms expand to include the rest of the facial muscles on that side of the face. Another characteristic of HFS is synkinesis in which there is an involuntary activation of multiple muscles that normally do not act together. An example of synkinesis is that stimulating the supraorbital branch of the trigeminal nerve would strongly activate the mentalis muscle, as well as the OO, in HFS patients. Unlike BEB,

8

Extraocular and Eyelid Muscles: Structure, Function, and Pathophysiology

HFS is unilateral and involves most of the ipsilateral facial muscles in addition to the OO. There is strong evidence, however, that HFS also arises from a combination of predisposing and precipitating factors that disrupt normal motor learning. The accepted precipitating factor for HFS is arterial compression at the root entry zone of the facial nerve. The most common blood vessels affecting the facial nerve in HFS are the anterior inferior cerebellar artery, the posterior inferior cerebellar artery, or the vertebral artery. Microsurgical decompression of the facial nerve typically reduces or eliminates the spasms and synkinesis of HFS so that spasms disappear in 64% and synkinesis in 53% of HFS patients within the first week following surgery. After 2–8 months, 90% of patients are spasm or synkinesis free. If arterial compression of the facial nerve alone is responsible for these signs of HFS, however, then decompression surgery should eliminate the spasms and synkinesis in less than 2–8 months. To understand the basis for this delay, it is important to consider what neural modifications might occur in response to pulsatile arterial compression of the facial nerve. A strong argument that facial nerve compression alone is insufficient to cause HFS is that 15–25% of the population exhibits arterial compression of the facial nerve, but do not develop HFS. This observation shows that HFS requires a predisposing factor as well as the precipitating factor, arterial compression of the facial nerve. Pulsatile arterial compression of the facial nerve can generate many modifications in brainstem neural circuits. Compression injury to motor neuron axons effectively weakens facial muscles, which initiates adaptive modifications. Repetitive antidromic activation of facial motor neurons by pulsatile compression of axons can alter facial motor neuron excitability or motor neuron excitability may increase because of facial motor neuron axotomy, severing the facial nerve fibers. Repetitive orthodromic activation of facial muscles by pulsatile compression of motor neuron axons can lead to a reorganization of sensory trigeminal circuits. Simultaneous orthodromic (electrical impulses traveling in the normal direction) activation of muscles that normally do not act together, for example, mentalis and OO, causes trigeminal primary afferent sensory signals from mentalis and OO muscle contraction to reach second-order trigeminal neurons synchronously. This abnormal pairing of sensory inputs can restructure trigeminal receptive fields so that the second-order trigeminal neurons respond strongly to inputs from both mentalis and OO activity instead of weakly to one and strongly to the other. Pulsatile activation of the facial nerve can also augment reflex circuit excitability because of synchronous activation of circuit elements. Second-order trigeminal neurons receiving synchronous afferent inputs innervate facial motor neurons that are already depolarized by antidromic activation.

This activity pattern can strengthen trigeminal inputs onto facial motor neurons in a spike timing-dependent plasticity-like manner. It is clear that pulsatile activation of the facial nerve can produce blink modifications, but these changes should not cause HFS by themselves. The eyelid system normally modifies itself so as to perform appropriately in the face of changes in the motor system or sensory inputs. For example, creating unexpectedly large blinks by adding weights to the upper eyelids initiates a rapid reduction in the trigeminal drive onto OO motor neurons. Similarly, chronic, repetitive facial nerve stimulation, such as occurs with pulsatile facial nerve compression, reduces blink amplitude. Thus, pulsatile facial nerve compression is insufficient to cause spasms of lid closure because the blink system will modify itself to prevent spasms of lid closure. The significant number of humans not experiencing HFS, but exhibiting arterial compression of the facial nerve, further supports this interpretation. These observations indicate that HFS, like BEB, requires a predisposing factor to develop spasms. Like BEB, an autosomal-dominant genetic mutation with low penetrance may provide the predisposing condition, which disrupts normally adaptive processes to create a pathological condition.

Further Reading Bour, L. J., Aramideh, M., and de Visser, B. W. (2000). Neurophysiological aspects of eye and eyelid movements during blinking in humans. Journal of Neurophysiology 83: 166–176. Evinger, C., Manning, K. A., and Sibony, P. A. (1991). Eyelid movements. Mechanisms and normal data. Investigative Ophthalmology and Visual Science 32: 387–400. Fukuda, H., Ishikawa, M., and Okumura, R. (2003). Demonstration of neurovascular compression in trigeminal neuralgia and hemifacial spasm with magnetic resonance imaging: Comparison with surgical findings in 60 consecutive cases. Surgical Neurology 59: 93–99; discussion 99–100. Manning, K. A. and Evinger, C. (1986). Different forms of blinks and their two-stage control. Experimental Brain Research 64: 579–588. Nielsen, V. K. and Jannetta, P. J. (1984). Pathophysiology of hemifacial spasm: III. Effects of facial nerve decompression. Neurology 34: 891–897. Rambold, H., Sprenger, A., and Helmchen, C. (2002). Effects of voluntary blinks on saccades, vergence eye movements, and saccade–vergence interactions in humans. Journal of Neurophysiology 88: 1220–1233. Sibony, P. A. and Evinger, C. (1998). Normal and abnormal eyelid function. In: Miller, N. R. and Newman, N. J. (eds.) Walsh and Hoyt’s Clinical Neuro-Ophthalmology, vol. 1, pp. 1509–1594. Baltimore, MD: Williams and Wilkins. Sparks, D. L. (2002). The brainstem control of saccadic eye movements. Nature Reviews Neuroscience 3: 952–964. VanderWerf, F., Brassinga, P., Reits, D., Aramideh, M., and Ongerboer de Visser, B. (2003). Eyelid movements: Behavioral studies of blinking in humans under different stimulus conditions. Journal of Neurophysiology 89: 2784–2796. VanderWerf, F., Reits, D., Smit, A. E., and Metselaar, M. (2007). Blink recovery in patients with Bell’s palsy: A neurophysiological and behavioral longitudinal study. Investigative Ophthalmology and Visual Science 48: 203–213.

Differentiation and Morphogenesis of Extraocular Muscles D M Noden, Cornell University, Ithaca, NY, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Mesenchyme – Embryonic cells with a fibroblast-like appearance, surrounded by extracellular matrix, lacking tight junctions with their neighbors, and often capable of undergoing extensive migratory movements. These can be of several different embryonic origins, and include cells that will contribute to many lineages. Morphogenesis – It includes those processes that establish the correct locations and three-dimensional organization of tissues and organs. This includes the proper positioning of extraocular muscles around the globe and their attachments to the sclera and orbital skeleton. Myoblasts – Mitotically active cells committed to the skeletal muscle lineage but not yet expressing muscle-specific proteins such as desmin and myosins, which generally are not evident until after myoblasts fuse to form multinucleated myofibers. Myotome – The several regions of each somite that contain progenitors of skeletal muscle progenitors. Neural crest – Mesenchymal cells that are derived from neural fold tissues and that move peripherally along well-delineated pathways to form neurons and glia of the peripheral nervous system and, in the head region, the connective tissues of the midface and branchial regions. Paraxial mesoderm – Early embryonic cells that are located beside and beneath the developing brain and spinal cord, and include the precursors of most skeletal muscles.

Introduction Muscles that move and stabilize the eye have been highly conserved during vertebrate evolution. While a few remarkable adaptations have occurred, such as co-opting of the dorsal (superior) oblique muscle to generate protective heating for the brain in some fishes, these muscles have retained an anatomical organization linking axes of the body and the eye that arose hundreds of millions of years ago. Considering their ancient status, it is logical to assume that the early development of extraocular muscles would similarly be well conserved among different species, and

therefore amenable to comparative analyses that supplant the absence of direct examination in mammals, including humans. However, compared to trunk and limb muscles, our understanding of the origins of ocular muscles and the mechanisms that initiate then maintain their development is at best fragmentary. Myogenesis of skeletal muscles is a lengthy process, with several parameters continuing to be function-dependent throughout the life of an animal. Primary myogenesis spans the period during which populations of myoblasts – committed, mitotically active muscle progenitors – arise, emigrate to their sites of differentiation, fuse to form multinucleated innervated myofibers, and establish intimate connections with connective tissues. This population forms a scaffolding, including the delineation of global and orbital domains, within which secondary myogenesis occurs. During secondary myogenesis stages, previously sequestered latent myoblasts are activated to proliferate and differentiate, forming more than 90% of the myofibers present in mature muscles and generating region-specific specialized fiber types that in most species are present before or soon after birth.

Origins of Extraocular Muscles Striated (skeletal) muscles throughout the body arise within paraxial mesoderm, which is located in close apposition to the embryonic brain and spinal cord. Exceptions to this are the avian striated ciliary muscle that is of neural crest origin, and possibly the striated muscles of the esophageal wall; however, both of these are involuntary. Among voluntary muscles, some of the more ventrally located branchial muscles arise from lateral mesoderm that is contiguous with paraxial mesoderm. Many early accounts of head myogenesis placed the embryonic origin of some eye muscles, especially the lateral rectus, in the same category as branchial (pharyngeal) arch muscles, and ascribed both to a lateral mesoderm origin. These claims were based on the sites at which muscle condensations are first grossly evident in the embryo. However, with the advent of mapping methods and assays for early muscle-specific gene expression patterns, separate and distinct sites of origin for all eye muscles within preotic (i.e., located rostral, in front of the developing inner ear) paraxial mesoderm was confirmed (Figure 1). The sites of origin of extraocular muscles parallel the sites of emergence of the three cranial motor nerves that

9

10

Extraocular and Eyelid Muscles: Structure, Function, and Pathophysiology

Prosencephalon

Extraocular Medial rectus Ventral rectus Dorsal rectus Ventral oblique

Paraxial mesoderm Lateral mesoderm 1st arch Mandibular adductors Intermandibular

3rd arch Pharyngeals Stylopharyngeus Hypaxial Laryngeal Glossal Rectus capitis

Mesencephalon

Dorsal oblique

n. V

n. IV

Retractor bulbi Lateral rectus

Metencephalon

Pterygoideus 2nd arch Stapedial Digastricus, facials

n. III

n. VI

n. VII DO n. IX

Myelencephalon n. XII

1st som.

n. X

Epaxial Biventer, splenius complexus (a)

2nd som. 3rd som. (b)

Figure 1 Sites of origin of craniofacial striated muscles in avian embryos. Panel (a) shows, in dorsal view, the locations of each muscle primordia within paraxial and lateral mesoderm. Panel (b) illustrates one mapping method used in avian embryos, wherein a small bolus of replication-incompetent retrovirus injected into early mesoderm and the embryo harvested 2 weeks later and stained for the appropriate reporter gene, in this case a bacterial galactosidase. In this embryo, in which the eye has been removed to reveal underlying tissues, only the dorsal (superior) oblique muscle was labeled.

innervate them. However, in contrast to axial and branchial muscles, sites of myogenesis are not congruent with locations of motor nerve emergence. Indeed, each of these cranial nerves needs to elongate considerably through peripheral tissues before initial contacts with target muscles are established. Some axons, such as those of the abducens nerve, must extend longitudinally beneath the brain to reach their target lateral rectus muscles, while others, such as the oculomotor nerve fibers, project perpendicular to the floor of the brain before decussating to approach their several target muscles. In extant vertebrates, head paraxial mesoderm constitutes a sparse population of mesenchymal cells (Figure 2). This contrasts with the situation in the neck and trunk regions, where paraxial mesoderm first forms somites, which are segmentally arranged, cuboidal aggregates of epithelial cells. As each somite matures, it becomes delineated into distinct myogenic (myotome) and connective

tissue-forming (sclerotome) regions. The most cranial somite is located beside the hindbrain, immediately caudal to the otic vesicle, and paraxial mesoderm rostral to this site fails to form epithelial tissues and lacks overt evidence of segmentation. Head paraxial mesoderm is present adjacent to the prospective eye-forming regions of the rostral neural plate, but is largely displaced caudally as the optic vesicles emerge and expand lateral to the diencephalon. In the midline just in front of the notochord, this population of loose paraxial mesoderm cells is contiguous with a sparse and species-variable population of prechordal mesenchymal cells. Mapping experiments in avian embryos have shown that prechordal mesoderm contributes to the genesis of extraocular muscles innervated by the oculomotor nerve (Figure 3), but it is not known whether this contribution is exclusive of or supplementary to that of paraxial mesoderm. It is not known if the same occurs in mammalian embryos.

Differentiation and Morphogenesis of Extraocular Muscles

11

Neural crest

Optic vesicle

Mesencephalon

Pharynx Lateral mesoderm

Heart tube Neural crest

(a)

Paraxial mesoderm

(b)

Figure 2 Colorized scanning electron micrographs showing the early relations of neural crest (blue) to mesodermal (red) populations in dorsal (a) and transverse (b) views. Small arrows indicate the direction of movement of the neural crest cells.

(a)

(b)

(c)

Figure 3 Contributions of prechordal mesoderm to developing extraocular muscles is shown by labeling their precursors at stage 4–5 (early gastrulation) with DiI, a fluorescent membrane-binding tag (site ‘o’ in (a)), and fixing the embryos over a day later((b), stage 12, ventral view). Labeled cells in (c) are within the eye muscle-forming region of paraxial mesoderm.

Determination of Eye Muscle Precursors Head paraxial mesoderm contains progenitors for many tissues in addition to skeletal muscle. These include cartilages and bones associated with the braincase, loose connective tissues such as meninges and adipocytes, and endothelial cells. In contrast to somites, wherein these progenitor populations are largely segregated, it appears based on mapping studies that these diverse precursors are either intermingled or contiguous in head mesoderm. The significance of this lies in the problem of generating diverse lineages. Somite cells are held in fixed positions relative to the dorsal and ventral parts of the adjacent neural tube (hindbrain and spinal cord) and overlying surface epithelium, all of which provide combinations of positive and negative regulators of early myogenesis and skeletogenesis.

A further complication – and one essential for the development of all craniofacial musculoskeletal systems – is the presence of a large, later-arising population of mesenchymal cells called the neural crest (Figure 2). Derived from neural folds either during (mammals) or shortly following (birds) closure of the cephalic neural tube (brain), these cells acquire a mesenchymal phenotype and quickly move peripherally, mostly atop underlying paraxial mesoderm. Neural crest cells from the rostral midbrain level move rostrally and caudally around the optic stalk and posterior part of the optic vesicle, then spread outwardly as the vesicle is transformed into the optic cup. After delineation of the lens from the lens placode, crest cells secondarily invade the space created anterior to the lens and establish the posterior epithelium (endothelium) and stromal populations of the cornea.

12

Extraocular and Eyelid Muscles: Structure, Function, and Pathophysiology

In the trunk, members of the wnt family of growth factors are secreted by surface ectoderm and provide essential positive stimuli for muscle differentiation. The same are released by head surface epithelium, but here their effects are to retard myogenesis of branchial muscles. Arriving neural crest cells separates branchial muscle progenitors from the source of these negative effects and, augmented by the release of additional myogenesispromoting factors, allows myogenesis to progress. The extent to which eye muscle progenitors follow a similar scenario is unclear. Some extraocular precursors, particularly the lateral rectus progenitors, are embedded deep within paraxial mesoderm and are therefore quite distant from both surface ectoderm and, at early stages, neural crest cells. This deep location places the lateral rectus precursors close to the neural epithelium at the level of the future metencephalon (pons). Several experiments have established that this location provides essential cues for lateral rectus formation. When newly formed trunk paraxial mesoderm cells were excised, before they had formed somites, then grafted into the head, in place of prospective lateral rectus mesoderm cells, the transplanted cells formed a muscle that expressed molecular markers unique to the lateral rectus and established proper anatomical connections with the braincase and sclera. Small changes in the location of the implants resulted in grafted cells contributing to the dorsal oblique or branchial arch musculature. Thus, the sites within head paraxial mesoderm at which each muscle primordium forms and is specified as to its identity are highly localized. Placing a barrier between the brain and paraxial mesoderm at this region does not prevent myogenesis, but the developing muscle cells lack molecular features that define their specific identity. Together, these experiments suggest that a rich tableau of local signals is necessary for early eye muscle differentiation, with both general myogenic and individual eye muscle-specific components.

Molecular Signatures and Muscle Differentiation All skeletal myoblasts use members of a closely related set of muscle-specific transcription factors to promote and coordinate their differentiation. Two of these, myf 5 and myoD, are cross-activating regulators that are among the earliest muscle-specific genes expressed. The upstream regulatory components of these genes are body regionand muscle group-specific, and serve to integrate the diverse micro-environments surrounding each muscle group with a highly shared set of outcomes, for example, activation of genes for desmin, myosins, muscle-specific actins, and junctional receptors.

Expression of myf 5 then myoD genes in eye muscle precursors generally is slightly later than expression in trunk axial muscles, but is simultaneous with that of branchial muscles (Figure 4). Expression of these regulatory genes coincides with the onset of aggregation of most muscle precursors (Figure 5), although it is not known whether these aggregates represent the totality of muscle precursors or only a subset of primary myoblasts. By these criteria, extraocular muscles appear similar to other head and also to trunk and appendicular muscles. However, as additional features of trunk and head myogenic regulatory networks have been identified, the number of differences has exceeded the similarities, and a heretofore underlying complexity has been revealed (Table 1). This area of investigation is rapidly expanding, and rather than detail each gene currently described, a few examples of categories of differences among muscle groups will be presented. Pax3 is a regulatory gene expressed in axial and appendicular muscle precursors, and null mutations of this transcription factor (e.g., Splotch mutation) result in severe depletion of trunk muscles. However, it is not expressed in head muscle precursors, and null mutations have no discernable effect on branchial or extraocular muscles. Another pronounced difference is in the hepatocyte growth factor (HGF) – cMet growth factor-receptor complex, which is functionally required for the correct dissemination and differentiation of appendicular and tongue muscle precursors. Again, this pathway has no known role in eye and branchial myogenesis, even though HGF is expressed in and around the precursors of branchial arch, lateral rectus, and both oblique muscles. The latter example further illustrates the considerable heterogeneity among extraocular muscles. The lateral rectus is particularly enigmatic, being the only head muscle that expresses the ubiquitous trunk paraxial mesoderm marker paraxis, and together with the dorsal oblique, the transcription factor lbx1, which is present in trunk hypaxial (thoracic and abdominal wall) muscle precursors. A further complexity arises from the often-changing patterns of gene expression during the early stages of head paraxial mesoderm development. The transcription factor pitx2, which is a key mid-level participant in the integrated formation of left-right asymmetry for the heart and mid-gut, is initially expressed symmetrically and ubiquitously throughout head paraxial mesoderm (Figure 6). However, a day later, during early myogenesis stages, its expression becomes more restricted but includes the first branchial arch, lateral rectus, and both oblique muscles in addition to periocular neural crest cells. Another regulatory gene, Tbx1, which is located in the region of chromosome 22 wherein deletions cause the DiGeorge syndrome, is similarly expressed over a wide domain of mesoderm (and pharyngeal endoderm) before becoming restricted to branchial arch and the lateral rectus muscles.

Differentiation and Morphogenesis of Extraocular Muscles

2 day

3 day

4 day

5 day

Move

13

Muscle group Dorsal (superior) rectus

Move

Inferior (ventral) oblique Inferior (ventral) rectus

Move

Medial rectus Move

Superior (dorsal) oblique

Move

Lateral rectus

Move

Branchial arch

Migr.

Tongue, laryngeal Epaxial (neck) Migr.

Wing Onset of myoD transcription

Aggregation of myoblasts move migr.

Key

Onset of myf5 transcription

Myoblast movements or migrations

Myosin proteins present

Figure 4 Timetable of gene activation in extraocular and other head and trunk muscles. The most consistent difference between trunk and head muscles is that the latter show a prolonged delay between the onset of myoD expression and the synthesis of muscle-specific proteins.

2.5 day

LR BA2 BA1 BA3

DO

VO LR DR

HGC

MyoD

6 day

4 day

BA2 BA3

LR DO

BA3 BA2

DO

EP

BA1 VO

VR MR

VR HYP

HGC

Myf5

DR

BA1

VO

Myosin

Figure 5 Early differentiation and morphogenesis of head muscles in chick embryos. BA1, 2, 3, branchial arches; DO, dorsal oblique; DR, dorsal rectus; Ep, epaxial muscle precursors; Hyp, hypaxial precursors, HGC, hypoglossal cord that forms tongue muscles; LR, MR, VR, lateral, medial and ventral rectus muscles; VO, ventral oblique.

At present the significance of these spatially and temporally dynamic expression patterns is unknown. It is possible that early expressions presage the later focal appearance of certain muscles, but it is equally plausible that each of these genes has multiple and distinct functions associated with each stage. As extraocular muscles mature, they exhibit a progression of fiber types, evidenced by changes in the myosin

isoforms and related contractile and energy metabolism proteins expressed. Emergence of these complex patterns requires a series of interactions among developing myofibers, surrounding connective tissues, and innervation. In the avian embryo, the primary myofibers of most extraocular muscles express embryonic slow myosin isoforms. However, one muscle, the quadratus nictitans, which is homologous to retractor bulbi muscles and is innervated

14

Extraocular and Eyelid Muscles: Structure, Function, and Pathophysiology

Table 1

Summary of myogenic regulatory networks of head and trunk

Expression sites

Genes

Trunk

Limb and tongue

Branchial

EOMs

All muscle Trunk only

myf5, myoD noggin, ptc1 pax3, c-met barx2 lbx1 paraxis pitx2 tbx1 en2 pod1 hgf myoR

þ þ þ þ þ þ þ

þ þ þ þ þ þ (þ)

þ þ þ BA1 þ BA1 þ

þ þ LR, DO LR þ LR þ DO, VO DO, VO, L1

Trunk and Head

Head only

LR BA1 BA2

DO

Allantois Left (a)

Right (b)

Figure 6 At stage 8 ((a), dorsal view) Pitx2 is expressed symmetrically throughout head paraxial (and lateral) mesoderm populations, but only on the left within trunk mesoderm. By stage 21 ((b), 4 days) it is restricted to a subset of eye muscles, the first and second branchial arch muscle masses, and periocular mesenchyme.

by the accessory abducens nerve, the myofibers are either completely fast myosin expressing, as in the quail embryo, or mixed, as in the chick embryo (Figure 7). To explore the basis for these distinct, species-specific patterns, periocular neural crest cells of the chick were replaced with comparable populations from a quail embryo. The quadratus muscle in these chimeric embryos exhibits the quail donor phenotype (Figure 7), indicating that initial differentiation of fiber types requires interactions among myofibers and encompassing connective tissues.

Muscle Morphogenesis Except for muscles that remain closely associated with the vertebral column and skull, all muscle progenitors leave their sites of origin in paraxial mesoderm and disperse into peripheral tissues. Body wall muscles do so together with sclerotome-derived connective tissue precursors, and maintaining these close spatial relations is essential

for the morphogenesis of thoracic and abdominal muscles. For appendicular myogenesis, lateral myotome-derived cells move in sequential waves of primary and secondary myoblasts into nearby limb buds, where they form longitudinal bands of future dorsal and ventral groups before segregating into individual muscles. In the hindlimbs, some of these undergo secondary dispersal to form muscles of the perineal region. Branchial muscles are comparable to body wall muscles in that they initially exhibit a constant juxtaposition with the precursors of their connective tissues, which in the head are all derived from neural crest cells, and also with the motor nerves that innervate them. This maintained registration permits continuous exchange of signals among all three components during all stages of muscle differentiation and morphogenesis. This constant contiguity has most dramatically been demonstrated for the trapezius muscle, whose precursors arise among caudal branchial arch mesodermal populations and secondarily shift caudally to attach to the scapula. Mapping studies in both bird and mouse embryos revealed that the neural crest-derived connective tissues move with, and perhaps somewhat in advance of, these myoblasts and indeed contribute to the scapula. This recapitulates an ancestral condition in which the forelimb girdle articulated with the back of the skull, as is still present in fish. Again, however, the extraocular muscles exhibit a developmental scenario unlike any other muscles. As illustrated in Figures 5 and 8, these muscle precursors move towards, then around, the equatorial region of the developing eye to assume their definitive locations. During this process, each muscle leaves the company of surrounding mesoderm cells and becomes fully encompassed by neural crest cells, which secondarily penetrate the muscle mass and form internal (e.g., endomysium) as well as surrounding (perimysium, fascia, and tendon) connective tissues. These periocular crest cells need not have originated at the same axial level as the muscles. For example, the lateral rectus muscle, the neural crest cells that will form

Differentiation and Morphogenesis of Extraocular Muscles

D.R

15

D.R

Qd. N. Qd. N.

(a)

(c)

D.R

Qd. N. Qd. N.

(b)

(d)

Figure 7 Fiber-type determination in the quadratus nictitans (Qd. N.) muscle. (a, b) Sections through this muscle in chick and quail embryos processed with antibodies to slow myosin isoforms. The quail Qd. N. lacks slow fibers, whereas in the chick both fast and slow fibers are present. (c) A control embryo in which chick neural crest cells were transplanted into a chick host, and the Qd. N. developed normally. However, when quail crest cells were grafted into a chick host embryo (d), the exclusive fast donor phenotype resulted.

Mesencephalon

Isthmus Me cep tenhalo n LR n VR Myelencephalo MR

DR

DO

VO 1.0 mm

Telencephalon

Figure 8 The movements of the dorsal (yellow arrow) and ventral (green arrow) oblique muscles from their sites of origin to their terminal locations along the equatorial zone of the globe.

its connective tissues, and the abducens nerve that innervates it originate at three separate axial levels. Indeed, these primordia do not become intimately associated until each has independently approached the periocular environment. This negates the possibility of prolonged

interactions among contiguous progenitor populations, as occurs for branchial musculoskeletal systems. The mechanisms by which aggregates of extraocular muscle primordia change both absolute and relative positions remain enigmatic. There is no precedence elsewhere in the embryo for condensations of cells moving actively through surrounding tissues. However, several lines of evidence support a model based on passive displacement of eye muscle primordia. As was shown in Figure 2, the interface between neural crest and myogenic paraxial mesoderm is initially a flat plane. Changes in the relative positions of the eye due to flexures and differential growth of the brain and expansion of the optic cup introduce distortions in this plane, but the extent to which this might affect individual eye muscles has been difficult to define. In screening for a wide range of gene expression patterns, several were found that coincided with the patterns of movements taken by some extraocular muscles (Figure 9). These reveal a set of localized distortions of the neural crest-mesoderm interface. Finger-like projections of paraxial mesoderm expand dorsally and caudally around the optic cup, becoming interdigitated with periocular neural crest populations and passively bringing the dorsal and ventral oblique muscle primordia to their definitive

16

Extraocular and Eyelid Muscles: Structure, Function, and Pathophysiology

globe, where they become surrounded then infused by connective-tissue forming neural crest cells, which largely direct both the gross and microscopic differentiation of these muscles. BA1

See also: Extraocular Muscles: Extraocular Muscle Anatomy.

Further Reading Figure 9 Lateral view of a 3.5-day chick embryo showing the sites of expression of the MyoR gene. Note the finger-like projections (arrows) extending dorsal and caudal to the optic vesicle, along the sites at which the dorsal and ventral oblique are expanding.

locations. Subsequently, crest cells close behind each of these muscle primordia, creating the appearance of an island of myoblasts/myofibers amid a sea of crest cells. How these focal distortions are established is unknown. The cell surface adhesion molecule semaphorin 3A is expressed by mesodermal cells in these projections, but its role in not known.

Summary The early stages of extraocular muscle formation are well described but poorly understood mechanistically. They arise at discrete sites within unsegmented head paraxial mesoderm then launch into developmental programs that share some features with trunk and branchial muscles but are largely and surprisingly unique. Passive distortions of the mesoderm-neural crest interface bring these muscle primordia to their definitive locations around the ocular

Borue, X. and Noden, D. M. (2004). Normal and aberrant craniofacial myogenesis by grafted trunk somitic and segmental plate mesoderm. Development 131: 3967–3980. Bryson-Richardson, R. J. and Currie, P. D. (2008). The genetics of vertebrate myogenesis. Nature Reviews Genetics 9: 632–646. Evans, D. J. R. and Noden, D. M. (2006). Spatial relations between avian craniofacial neural crest and paraxial mesoderm cells. Developmental Dynamics 235: 1310–1325. Grenier, J., Teillet, M. A., Grifone, R., Kelly, R. G., and Duprez, D. (2009). Relationship between neural crest cells and cranial mesoderm during head muscle development. PLoS ONE 4: e4381. Marcucio, R. M. and Noden, D. M. (1999). Myotube heterogeneity in developing chick craniofacial skeletal muscles. Developmental Dynamics 214: 178–194. Noden, D. M. and Francis-West, P. (2006). The differentiation and morphogenesis of craniofacial muscles. Developmental Dynamics 235: 1194–1218. Noden, D. M., Marcucio, R. M., Borycki, A-G., and Emerson, C. P., Jr. (1999). Differentiation of avian craniofacial muscles. I. Patterns of early regulatory gene expression and myosin heavy chain synthesis. Developmental Dynamics 216: 96–112. Noden, D. M. and Schneider, R. A. (2006). Neural crest cells and the community of plan for craniofacial development: Historical debates and current perspectives. In: Saint-Jeannet, J. P. (ed.) Neural Crest Induction and Differentiation, pp. 1–23. Boston, MA: Landes Bioscience and Springer Science Media. Tzahor, E. (2009). Heart and craniofacial muscle development: A new developmental theme of distinct myogenic fields. Developmental Biology 327: 273–276. Yoshida, T., Vivatbutsiri, P., Morriss-Kay, G., Saga, Y., and Iseki, S. (2008). Cell lineage in mammalian craniofacial mesenchyme. Mechanisms of Development 125: 797–808.

Extraocular Muscles: Extraocular Muscle Anatomy L K McLoon, University of Minnesota, Minneapolis, MN, USA S P Christiansen, Boston University, Boston, MA, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Abducens nerve (CNVI) – The cranial motor nerve which controls contraction of the lateral rectus muscle. Aponeurosis – A large flat and dense connective tissue layer which anchors a muscle to its origin or insertion. Excyclotorsion – An outward rotation of the upper pole of the vertical midpoint of each eye. Incyclotorsion – An inward rotation of the upper pole of the vertical midpoint of each eye. Myosin heavy chain (MyHC) isoforms – The major contractile protein in muscle is myosin, which in turn is composed of two heavy and four light chains. MyHC isoforms are responsible for the shortening velocity of muscle fibers during muscle contraction. Oculomotor nerve (CNIII) – The cranial motor nerve which controls contraction of the levator palpebrae superioris muscle, the inferior, medial, and superior rectus muscles, as well as the inferior oblique muscle. It also contains parasympathetic preganglionic axons that are destined for the ciliary body and iris sphincter muscles of the eye. Ora serrata – The junction between the neurosensory retina and the ciliary body. Satellite cells – Myogenic precursor cells that reside between the sarcolemma of the muscle fiber and its surrounding basal lamina. They are responsible for myofiber repair and regeneration after injury or in disease. Strabismus – A disorder characterized by altered tonus or restrictive disease of the extraocular muscles resulting in loss of conjugate binocular vision. Trochlear nerve (CNIV) – The cranial motor nerve which controls contraction of the superior oblique muscle.

The extraocular muscles (EOMs) have an extremely complex anatomy, both at the gross anatomical and histological levels. The main function of the EOM is to move the eyes in the orbit such that the eyes can be precisely positioned to allow focusing of the visual world on corresponding regions of each retina. Within each bony orbit the EOM includes four rectus muscles (superior,

lateral, inferior, and medial) and two oblique muscles (superior and inferior; Figure 1). A seventh skeletal muscle within the human orbit is the levator palpebrae superioris (LPS) muscle. It is the superior-most muscle in the orbit directly inferior to the frontal bone forming the orbital roof. The levator lies directly inferior to the periorbita and inserts via a large aponeurosis into the eyelid skin. The descriptions of the EOM will be primarily based on human muscles for ease of presentation. The general characteristics of size, shape, fiber type, and the like are similar for all EOM in principle, although they vary in detail for each specific animal that has been examined.

Gross Anatomy of the EOM within the Orbit The four rectus muscles originate from the apex of the bony orbit by a common tendinous annulus (of Zinn). The tendinous annulus attaches to the greater and lesser wings of the sphenoid bone as well as to the periosteum, the dense connective tissue lining the orbit. The annulus crosses over the inferior portion of the superior orbital fissure and runs superior and medial to the optic foramen (Figure 2). The superior (Figure 3) and medial rectus muscles arise from the superior part of this annulus, while the inferior and lateral rectus muscles arise from its inferior part. These muscles are all surrounded by a connective tissue capsule called Tenon’s capsule and are described as forming the muscle cone. The superior oblique (SO) muscle originates from the periosteum slightly superior and medial to the tendinous annulus. The inferior oblique (IO) muscle is the only EOM that does not arise from the orbital apex, but rather originates from the lateral border of the lacrimal fossa, which is anterior and nasal within the orbit. The rectus muscles run anteriorly to insert on the sclera on the anterior pole of the globe, at a location superficial to the ora serrata. The lateral and medial rectus muscles in human adults are approximately 41 mm in length. The superior rectus (SR) is the longest, averaging 42 mm, while the inferior rectus is the shortest averaging 40 mm. The insertions of the muscles onto the globe vary in their distance from the corneal limbus, with the SR furthest and the medial rectus closest. According to a study by Fuchs in 1884 on cadaver eyes, the distance from the limbus of rectus muscle insertions onto the globe

17

18

Extraocular and Eyelid Muscles: Structure, Function, and Pathophysiology

Superior

LPS/SR SO ON LR

MR IR

Figure 1 Magnetic resonance imaging view of the orbit in cross section. LPS/SR: levator palpebrae superioris and superior rectus muscles. SO, superior oblique muscle; MR, medial rectus muscle; IR, inferior rectus muscle; LR, lateral rectus muscle; ON, optic nerve. MRI generously provided by Dr. Michael S. Lee at the University of Minnesota.

Superior

Medial

Figure 2 Bony orbit with the tendinous annulus indicated in the orbital apex by the black oval. It encloses the optic foramen superiorly and crosses the inferior portion of the superior orbital fissure.

decreases as one progresses sequentially in the following order: SR at 7.7 mm, lateral rectus at 6.9 mm, inferior rectus at 6.5 mm, and medial rectus at 5.5 mm. However, when measured in vivo during strabismus surgery the distances were shorter for all muscles: SR at 6.7 mm, lateral rectus at 6.2 mm, inferior rectus at 5.89 mm, and medial rectus at 4.5 mm. The insertions typically are circumferential to the limbus. However, wide variations in obliquity and regularity of the insertions are common, both from patient to patient and from muscle to muscle, even in the same eye. These differences are important in the context of planning surgery on the EOM for treatment of eye motility disorders, such as strabismus. Recent studies convincingly show that the distance from muscle

Figure 3 Dissection of the human orbit from the superior view with the orbital bony roof and the periorbita removed. The most superficial structure is the frontal nerve. Directly inferior to the frontal nerve is the levator palpebrae superioris (LPS, elevated by scissor tips). Inferior to the LPS is the superior rectus muscle.

insertion to the limbus has a high inter-individual variability, and there appears to be no correlation between insertional distance and the amount of deviation in strabismus patients. The lateral rectus is innervated by the abducens nerve (CNVI) on its intraconal or deep surface (Figure 4). The SR muscle is innervated on its intraconal surface by the superior division of the oculomotor nerve (CNIII), while the inferior and medial rectus muscles and the IO muscles

Extraocular Muscles: Extraocular Muscle Anatomy

19

Figure 5 Deep orbital dissection from the superior view. The levator palpebrae superioris and superior rectus muscles are reflected medially to allow visualization of the superior and inferior divisions of the oculomotor nerve (CNIII). Figure 4 Dissection of the human orbit from the superior view. The bony roof, periorbita, levator palpebrae superioris, and superior rectus muscles are removed. The abducens nerve can be traced from where it pieces the dura in the posterior cranial fossa, through the superior orbital fissure, which has been opened in this dissection, to where it enters the lateral rectus muscle which it innervates.

are innervated by the inferior division of CNIII (Figure 5). All the cranial motor nerves except for the trochlear nerve (CNIV) enter the muscles intraconally, and all the motor nerves enter at approximately one-third of the muscle’s length from the orbital apex (Figure 6). The six EOMs control the position of the eye in the orbit while orbital fat and fascia constrain the paths of the muscles within the orbit. The two horizontal rectus muscles of each eye are agonist–antagonist pairs with relatively straightforward function; the medial rectus adducts the eye and the lateral rectus abducts it (Figure 7). By contrast, the SO and the IO muscles and the two vertical rectus muscles have far more complex functions. The primary direction of movement caused by the superior and inferior rectus muscles is elevation and depression, respectively. However, due to the shape of the bony orbit and their sites of origin and insertion, the vertical recti have secondary and tertiary actions that are torsional and horizontal, respectively. The SR, for example, is a secondary incyclorotator, moving

the superior pole of the eye toward the nose; its tertiary function is adduction. The inferior rectus muscle, by contrast, is also a secondary excyclorotator, but similar to the SR is a tertiary adductor (Figure 8). Thus, the SR, if acting unopposed, would elevate, adduct and incyclotort the eye such that the eye would be looking up and in. The inferior rectus, if acting unopposed, would depress, adduct, and excyclotort the eye such that the eye would be looking down and in. The superior and IO muscles have a similarly complex cyclovertical functions. The primary action of these two muscles is rotation or torsion, but due to the angle which they take in the orbit toward their insertion on the sclera, they will also elevate (IO) or depress (SO) the eye; both abduct the eye. The remarkable balance and integrity of the ocular motor plant becomes evident when one considers that to look straight superiorly without moving the head, both the IO (a primary excyclorotator, secondary elevator, and tertiary abductor) and the SR (a primary elevator, secondary incyclorotator, and tertiary adductor) coordinately contract. The same is true for the inferior rectus and SO muscles moving the eye straight inferiorly. The SO muscle is the thinnest, roundest, and longest of the EOM. The muscle runs 32 mm along the border of the medial wall and roof of the orbit, and 10–15 mm from the orbital margin it becomes tendinous and passes through

20

Extraocular and Eyelid Muscles: Structure, Function, and Pathophysiology

SR

IO

Figure 8 The primary role of the superior rectus (SR) muscles is to elevate the eye; however, they have a secondary role in adducting the eye. The primary role of the inferior oblique (IO) muscles is to extort the eye, rotating the eye upward (elevation) and outward (abduction).

Figure 6 Deep orbital dissection from the superior view. The lateral, inferior, and medial rectus muscles are clearly visible, as are the nerves entering each muscle. The globe has been retracted anteriorly (by scissors).

Medial rectus adducts

Lateral rectus abducts

Figure 7 The lateral rectus (LR) and medial rectus (MR) muscles move the eyes in the horizontal plane. The MR adducts the eye, moving the eye toward the midline. The LR abducts the eye, moving the eye away from the midline. These muscles are antagonists; they have opposite functions.

a fibro-fascial structure called the trochlea, located anteriomedially under the orbital roof. The trochlea changes the orientation of the muscle, becoming the functional origin of this muscle. The tendon of the SO courses deep to the SR muscle to insert into the sclera laterally on the posterior pole of the globe. The position of the SO tendon insertion, posterior and temporal to the center of rotation of the globe, explains the complex cyclovertical function

Figure 9 The superior oblique muscle can be seen along the medial wall of the orbit. The trochlear nerve (CNIV) is seen coursing along the muscle’s superior surface and enters into the posterior 1/3 of the muscle on its medial side.

of the muscle. The trochlear nerve (CNIV) innervates this muscle, coursing first on the muscle’s superior side, and finally entering the muscle superiomedially at about the proximal one-third of the muscle’s length (Figure 9). The IO muscle is approximately 37 mm in length and travels in a similar orientation to the reflected tendon of the SO distal to the trochlea. The muscle runs inferior to

Extraocular Muscles: Extraocular Muscle Anatomy

the inferior rectus muscle and inserts into the sclera of the posterior pole on the lateral side of globe. The insertion is relatively close to the macula of the retina. The IO is innervated by the inferior division of the oculomotor nerve (CNIII) on its superior surface with the nerves entering the muscle at approximately the posterior one-third.

Histological Anatomy of the EOMs Overall Organization The EOMs have a complex anatomical organization at the microscopic level. These muscles differ from those in the limbs and body in that the muscle fibers themselves have extremely small cross-sectional areas, with an average in human EOM of 340 200 mm2 (Figure 10). In addition, there is a great deal of variability in the myofiber crosssectional areas; this heterogeneity is quite striking. In cross-section, two distinct layers can be seen in all six EOMs, the orbital and the global layers. The orbital layer faces the bony orbit and is composed of myofibers with extremely small cross-sectional areas, with a mean of

Orbital

21

260 160 mm2. The global layer faces the globe, and the myofibers are markedly bigger than those in the orbital layer with a mean of 440 200 mm2. However, they are still very small compared to body and limb skeletal muscle myofibers, which typically range from 3500 to 4000 mm2 in the human soleus muscle, as an example. The total number of myofibers found within each of the six EOMs varies significantly. When measured in the mid-belly region of the muscles, the numbers of myofibers in the orbital layer in human EOM range from 7400 to 14 600 and in the global layer the numbers range from 8000 to 16 400 myofibers. This variation in myofiber number is seen in other species where the fiber number has been examined, although the range varies significantly from human. For example, monkey EOMs have approximately half the number of myofibers compared to human EOM. In addition, total myofiber number decreases along the length of each muscle as the insertions are approached. This variation in fiber number is due to the fact that the majority of myofibers within the EOM does not run tendon to tendon, as has been shown by a number of investigators. This can be demonstrated quite easily by serially sectioning muscles and following individual myofibers in consecutive sections (Figure 11). This is also supported by electrophysiological evidence demonstrating that the force produced by stimulating two separate motor unit groups individually is often more than the force produced by stimulating both motor groups simultaneously. This nonlinearity, or loss of force in summated motor units, is postulated to be caused by the lateral dissipation of force due to myofibers that do not reach the tendon ends.

Innervation

Global

Figure 10 A cross section of the medial rectus muscle from a Rhesus monkey immunostained for the presence of the fast myosin heavy chain (MyHC) isoform, which labels all forms of the fast MyHC. The orbital layer is composed of myofibers with extremely small cross-sectional areas, while myofibers in the global layer are somewhat larger. Bar is 50 mm.

Most skeletal muscles receive their motor nerve innervation in approximately the middle, and the neuromuscular junctions form a single endplate zone. Neuromuscular junctions are a pentomer composed of four distinct subunits: a (2), b, g, and d. During maturation, the g subunit is replaced with an e subunit, forming the adult acetylcholine receptor. In contrast to limb skeletal muscle, the EOM maintains a subpopulation of neuromuscular junctions with the immature subunit configuration. While the vast majority of EOM muscle fibers is fasttwitch fibers and receive a typical en plaque type of neuromuscular junction somewhere along their length, the EOM also contains two types of multiply innervated myofibers. These account for approximately 10% of the fibers in the global and orbital layers. In the global layer, these multiply innervated myofibers contain slow-tonic myosin and appear to be innervated by small en grappe endplates along their length. These en grappe neuromuscular junctions retain the g subunit of the acetylcholine receptor, rather

22

Extraocular and Eyelid Muscles: Structure, Function, and Pathophysiology

(a)

(b)

Figure 11 Two serial sections from a normal rabbit superior rectus muscle. Several sets of muscle fibers in cross section are numerically identified. (a) Note two small myofibers present in this section, one to the right of group 5, 6, and 7 (black arrow), and one between fibers 3 and 4 (black arrow). (b) Note that in a section 20 mm from the section in (a) that those two small fiber ends have ended and thus are no longer present in the cross-section. Muscle sections immunostained to visualize dystrophin on the sarcolemma. Bar is 40 mm.

than the e subunit of the mature neuromuscular junction. There are reports in the literature that the en plaque endings in mammalian EOM can also express the immature g subunit of the acetylcholine receptor. This type of multiply innervated myofiber develops a slow-graded or tonic tension when the nerves are stimulated. The orbital layer has a second type of multiply innervated myofiber. These fibers have a neuromuscular junction in the central one-third of their length with typical en plaque structure. In addition, however, there are multiple en grappe endings at the myofiber ends. Thus, the central region displays fast-twitch properties with nerve stimulation while the fiber ends display slow-tonic contractile properties. As a result of fibers that do not run the origin-toinsertion length of the muscles and the presence of neuromuscular junctions on fiber ends, neuromuscular junctions are found throughout the length of the EOM. This has important implications for pharmacologic and surgical manipulations of the muscles in the treatment of eye motility disorders such as strabismus and nystagmus.

Skeletal Muscle Fiber Types From a physiological perspective, the EOMs have extremely fast contractile properties, produce low levels of force, and are relatively fatigue resistant. These properties are conferred on muscle fibers by their expression of specific contractile proteins such as the myosin heavy chain isoforms (MyHC), as well as differences in cellular organelles such as mitochondria and other cellular metabolic pathways. One unusual aspect of EOM histology is the expression, within and between these two layers, of a complex pattern of MyHC expression.

Myosin Isoform Complexity in the EOM Skeletal muscle fibers in body and limb muscles have generally been described as fast, expressing MyHC type IIa, IIb (nonhuman), and/or IIx/d, or slow, expressing MyHC type 1. However, this fiber type organization breaks down in the EOM, as was described by Mayr in 1971. On average, only 16% of the myofibers within the orbital layer are positive for the slow MyHC (MYH7), while 14% of the myofibers in the global layer are positive for slow MyHC. Thus, the vast majority of myofibers expresses the fast MyHC2a isoform (MYH2) which is responsible for the extremely fast contractile properties of the EOM. However, the EOMs express, in total, at least eight distinct MyHC isoforms: 2a, 2x/d, 1 (b-cardiac), developmental (embryonic), neonatal (perinatal), acardiac, slow-tonic, and EOM-specific. If serial sections of individual myofibers are examined immunohistochemically, it quickly becomes evident that single myofibers express more than one isoform (Figure 12). In the Kjellgren et al. study of adult human EOMs, single fibers immunostained for the slow or MyHC type 1 isoform can co-express either or both slow-tonic MyHC and acardiac MyHC. If just the slow myofibers are considered, they can express: 1. 2. 3. 4. 5. 6. 7. 8. 9.

only MyHC type 1, only slow-tonic, only a-cardiac, only EOM-specific, MyHC type1 and slow-tonic, MyHC type 1 and a-cardiac, myosin type 1 and EOM-specific, slow-tonic and a-cardiac, slow-tonic and EOM-specific,

Extraocular Muscles: Extraocular Muscle Anatomy

Fast myosin

Developmental myosin

Neonatal myosin

10. 11. 12. 13. 14.

23

a-cardiac and EOM-specific, slow type 1, slow-tonic, and a-cardiac, slow type 1, slow-tonic, and EOM-specific, slow-tonic, a-cardiac, and EOM-specific, or all four isoforms (Figure 13).

These combinations may result in 14 types of fibers. It is also known that embryonic (developmental) and neonatal-specific MyHC isoforms are also expressed by some of these fibers as well. The same complexity is seen with the myofibers positive for the fast MYHC type 2A. Single fibers can also express one of one or more of these isoforms: developmental, neonatal, EOM-specific, and 2x/d. These types of myofibers are referred to as hybrid fibers, and they can be seen, albeit to a lesser extent, in other muscles such as the diaphragm. Even when only MyHC expression characteristics are considered, the high degree of individual myofiber polymorphism seen in the collective data from many laboratories strongly supports the view that there is a continuum of myofiber types within the EOM. In some ways, trying to fit the EOM into the classical fiber typing scheme is misleading, as it does not deal effectively with the hybrid and mismatched fibers. This complexity has significant ramifications for muscle function. MyHC isoforms control muscle-shortening velocity, and it has been proposed that this type of polymorphism allows for fine-tuned control over a wide range of forces, velocities, and fatigue properties. From a teleological perspective, these coexpression patterns would allow the EOM to contract at high velocities but with minimal fatigue, a characteristic that is important for muscles that are continually functioning in order to maintain fixation of gaze on the fovea in an infinite number of eye positions. Additionally, studies have shown that the EOMs show rapid alterations in MyHC isoform expression in response to stretch, alterations in hormones, botulinum toxin treatment, denervation, and the like. Nonuniform Expression of MyHC Isoforms along the Muscle Length

Figure 12 Three serially cut cross sections from a region of a rabbit lateral rectus muscle approaching the anterior 1/3 of the muscle. They have been immunostained for fast, developmental, and neonatal myosin heavy chain (MyHC) isoforms. One group of muscle fibers has been circled in yellow with two fibers identified by a light-green arrow and a fiber numbered 2. Note that the fiber indicated by the green arrow is positive for fast and neonatal MyHC but negative for developmental MyHC. Note that fiber 2 is negative for fast MyHC but positive for developmental and neonatal MyHC. A second group of muscle fibers has been circled in red with two fibers identified as fiber 3 and 4. Fiber 3 is positive for fast MyHC but negative for both developmental and neonatal MyHC, while fiber 4 is positive for all three of these isoforms. The fiber indicated by the large blue arrow is negative for fast MyHC, but positive for both developmental and neonatal MyHC.

Immunohistological examination of cross-sections taken from the tendon ends and the middle region of an EOM shows that the overall percentages of specific MyHC isoforms change dramatically depending on the location along the muscle length (Figure 14). For example, in a study of rat EOM, within the orbital layer the mid-belly region expresses EOM-specific MyHC but this isoform is completely excluded from the tendon ends where the embryonic (developmental) MyHC isoform is present. This is seen even at the level of single isolated myofibers, where the fiber ends express neonatal MyHC and the mid-region of the same myofiber expresses fast MyHC. This type of non-uniform expression of MyHC along the length of single fibers also is seen in the intrafusal muscle

24

Extraocular and Eyelid Muscles: Structure, Function, and Pathophysiology

Type 1

Type 1 α-cardiac

Type 1 slow tonic

Type 1 slow tonic a-cardiac

α-cardiac

Slow tonic

Type 1 eom-specific

Type 1 slow tonic eom-specific

Slow tonic α-cardiac

Type 1 α-cardiac eom-specific

eom-specific

Slow tonic eom-specific

Slow tonic α-cardiac eom-specific

α-cardiac eom-specific

Type 1 slow tonic a-cardiac eom-specific

Some possible hybrid fibers in extraocular muscle Figure 13 Diagrammatic representation of possible co-expression patterns when four myosin heavy chain isoforms are considered. Each red rectangle represents a single myofiber. Based on Kjellgren, D., Thornell, L. E., Andersen, J., and Pedrosa-Domellof, F. (2003). Myosin heavy chain isoforms in human extraocular muscles. Investigative Ophthalmology and Visual Sciences 44: 1419–1425.

(a)

(b)

(c)

(d)

Figure 14 Cross section from a single superior rectus muscle of an adult monkey taken from the orbital (a, b) and global (c, d) regions of the mid-region of the muscle (a, c) and the tendon end (b, d) immunostained for fast myosin heavy chain isoform (MyHC) (a, b) and neonatal MyHC (c, d). The orbital region is at the top of the photomicrographs in (a) and (b). Note that the orbital layer is thicker in (a), but that the individual myofibers have a smaller cross-sectional area in the mid-region compared to the tendon end. Additionally, there are more fast-negative fibers in the tendon region of the orbital layer. In the global layer, the mid-region has 20% of its myofibers positive for the neonatal MyHC, while the tendon end is almost devoid of this isoform. Bars is 50 mm.

Extraocular Muscles: Extraocular Muscle Anatomy

fibers found in muscle spindles, another specialized myofiber structure. Using single myofiber reconstructions to localize neonatal MyHC isoform expression in single fibers, myofibers are found that are neonatal MyHCpositive from fiber tip to fiber tip, but it is more common to find single myofibers with variable percentages of the total fiber length expressing this isoform, including fibers where the expression is discontinuous. It is emerging that these types of hybrid fibers are present in other craniofacial muscles, such as jaw and laryngeal muscles. In limb and body skeletal muscles, innervation, neuromuscular activity, exercise (use/disuse), mechanical loading or unloading, hormones, and aging all cause adaptive changes in contractile properties and metabolic profiles. This supports the view that expression of contractile proteins within muscle fibers is extremely dynamic and possesses an incredible adaptability to meet the physiological demands placed upon them. EOMs are constantly active and appear to represent the far end of a continuum of skeletal muscle types relative to their ability to react and adapt to changing functional needs. As the MyHC controls shortening velocity, this MyHC polymorphism would result in a gradation of function within populations of single myofibers. In other words, a continuum of force and movement would be possible, as the contractile properties of each myofiber would reflect its particular subset of contractile proteins. An example of this is shown in Figure 15, where force was determined using a single fiber preparation and the MyHC content was analyzed for each single fiber using polyacrylamide gel separation. As can be seen, fibers that produce the same amount of force can have significantly different MyHC isoform expression patterns. While it has not been demonstrated specifically within EOM myofibers, it is a well-known characteristic of multinucleated myofibers that each myonucleus controls

25

the expression of proteins in what is called its myonuclear domain. An elegant study by the Hardeman laboratory showed that transcription occurs in pulses within individual myonuclei, and the activities of single nuclei are not in sync with each other. Each nucleus, and thus each myonuclear domain, is individually controlled, and protein synthesis is a dynamic process that is altered at the local level to respond to the particular stress or strain. The advantage conferred by this complexity within populations of individual EOM myofibers can be hypothesized as an adaptation to the functional needs of eye movements in maintaining binocular vision and highly coordinated vergence movements, where a continuum of contraction forces and speeds would be required. Functionally, the complex MyHC co-expression patterns, and their presumed continuous modulation, would allow finely tuned control over these movements, as the kinetics of the EOM would cover a wide range of eye positions and velocities. Other Molecules Heterogeneously Expressed The heterogeneity of individual myofibers is made even more complex not only by differences in other contractile proteins, such as myosin light chains and troponin, but also by other metabolic differences. Of the molecules that have been specifically examined, the EOM often has patterns of expression not seen in limb muscle. For example, myosinbinding protein C has three isoforms. Despite the fact that the vast majority of myofibers within human EOM is positive for fast MyHC, the EOM does not express the fast form of myosin-binding protein C. Recent work showed that, in contrast to limb skeletal muscles, high levels of glycolytic and oxidative pathways coexist within single myofibers in the EOM. This molecular mismatch would provide these very active muscles with both fatigue resistance and fast contractile properties.

Continuous Remodeling in Normal Adult EOM

Fiber 2

Fiber 1

0

10

20 IIA

30 40 50 60 70 MyHC isoform (%) IIX

Neonatal

80

90 IIB

Figure 15 Relative percentages of four myosin heavy chain isoforms from two single-skinned myofibers with the same shortening velocity of 9.4 fiber lengths per second as determined using single-skinned fiber physiology in the EOM of rabbits. Note that fiber 1 has three isoforms expressed, and it contains type IIX as its main isoform, while fiber 2 has only two isoforms expressed, and IIB is the isoform with the greatest amount of expression.

Early studies by Moss and LeBlond demonstrated that the myonuclei within mature, multinucleated myofibers are postmitotic. However, muscle has regenerative capacity that resides in myogenic precursor cells called satellite cells, and these cells become activated, divide, and are responsible for muscle repair and/or regeneration of new fibers in disease and after injury. The EOM in normal adult mammals maintain an elevated number of satellite cells throughout life (Figure 16), which divide and integrate continuously into apparently normal muscle fibers. Concomitantly with nuclear addition, apoptosis of individual myonuclei is seen with apparent segmental cytoplasmic remodeling. The factors that control this process are unknown, but this process represents another dynamic

26

Extraocular and Eyelid Muscles: Structure, Function, and Pathophysiology

Figure 16 Several longitudinally sectioned myofibers from the inferior oblique muscle of a human cadaver donor. Note three satellite cells positive for Pax-7, a marker of all satellite cells. Two satellite cells in close proximity to each other is a common observation in the extraocular muscles.

physiological property of the EOM that would allow continuous adaptation of single fiber functional properties in response to physiological needs. The existence of ongoing remodeling in normal adult EOM suggests that there may be ways to modulate this process in vivo to alter muscle size, force, or response to injury or disease. In particular, it suggests new hypotheses to explain the preferential sparing or involvement of the EOM in skeletal muscle disease.

Acknowledgments This work was supported by EY15313 and EY11375 from the National Eye Institute, the Minnesota Medical Foundation, the Minnesota Lions and Lionesses, Research to Prevent Blindness (RPB) Lew Wasserman Mid-Career Development Award (LKM), and an unrestricted grant to the Department of Ophthalmology from RPB. See also: Extraocular Muscles: Extraocular Muscle Metabolism; Extraocular Muscles: Functional Assessment in the Clinic; Eyelid Anatomy and the Pathophysiology of Blinking.

Further Reading Asmussen, G., Punkt, K., Bartsch, B., and Soukup, T. (2008). Specific metabolic properties of rat oculorotatory extraocular muscles can be linked to their low force requirements. Investigative Ophthalmology and Visual Sciences 49: 4865–4871.

Caiozzo, V. J., Baker, M. J., Huang, K., et al. (2003). Single-fiber myosin heavy chain polymorphism: How many patterns and what proportions? American Journal of Physiology – Regulatory, Integrative and Comparative Physiology 285: R570–R580. Harrison, A. R., Anderson, B. C., Thompson, L. V., and McLoon, L. K. (2007). Myofiber length and three-dimensional localization of NMJs in normal and botulinum toxin-treated adult extraocular muscles. Investigative Ophthalmology and Visual Sciences 48: 3594–3601. Jacoby, J., Chiarandini, D. J., and Stefani, E. (1989). Electrical properties of multiply innervated fibers in the orbital layer of rat extraocular muscles. Journal of Neurophysiology 61: 116–125. Kallestad, K. M. and McLoon, L. K. (2008). Myogenic precursor cells in the extraocular muscles. In Low, W. C. and Verfaillie, C. M. (eds.) Stem Cells and Regenerative Medicine. Hackensack, NJ: World Scientific. Kaminski, H. J., Kusner, L. L., and Block, C. H. (1996). Expression of acetylcholine receptor isoforms at extraocular muscle endplates. Investigative Ophthalmology and Visual Sciences 37: 345–351. Kjellgren, D., Thornell, L. E., Andersen, J., and Pedrosa-Domellof, F. (2003). Myosin heavy chain isoforms in human extraocular muscles. Investigative Ophthalmology and Visual Sciences 44: 1419–1425. Li, Z. B., Rossmanith, G. H., and Hoh, J. F. Y. (2000). Cross-bridge kinetics of rabbit single extraocular and limb muscle fibers. Investigative Ophthalmology and Visual Sciences 41: 3770–3774. Mayr, R. (1971). Structure and distribution of fiber types in the external eye muscles of the rat. Tissue and Cell 3: 433–462. McLoon, L. K., Rowe, J., Wirtschafter, J. D., and McCormick, K. M. (2004). Continuous myofiber remodeling in uninjured extraocular myofibers: Myonuclear turnover and evidence for apoptosis. Muscle and Nerve 29: 707–715. Shall, M. S., Dimitrova, D. M., and Goldberg, S. J. (2003). Extraocular motor unit and whole-muscle contractile properties in the squirrel monkey. Summation of forces and fiber morphology. Experimental Brain Research 151: 338–345. Stephenson, G. M. M. (2001). Hybrid skeletal muscle fibers: A rare or common phenomenon? Clinical and Experimental Pharmacology and Physiology 28: 692–702.

Extraocular Muscles: Extraocular Muscle Metabolism F H Andrade, University of Kentucky Medical Center, Lexington, KY, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Adenine nucleotide translocator – An inner mitochondrial protein that exchanges adenosine diphosphate (ADP) and adenosine triphosphate (ATP) between the mitochondrial matrix and the cytoplasm. It is also known as the ADP/ATP translocator. Chronic progressive external ophthalmoplegia – A syndrome characterized by progressive inability or difficulty to move the eyes and elevate the eyelids. It is a common manifestation of some mitochondrial diseases. Electron transport chain – The set of mitochondrial protein complexes that couples the oxidation of electron donors (such as NADH) to the reduction of electron acceptors (such as oxygen) in order to produce ATP. Gene expression profiling – The measurement of the activity (or expression) of large numbers of genes simultaneously. Glycolysis – The metabolic pathway that converts glucose to pyruvate, with a net result of 2 ATP and 2 NADH. M line – A dark band or line seen in the center of the sarcomeres, using electron microscopy. NADH – Nicotinamide adenine dinucleotide (NAD+) and its reduced form NADH are coenzymes involved in oxidation/reduction reactions as electron acceptors and donors. Oxidative phosphorylation – The metabolic pathway that uses the energy released during the oxidation of nutrients to produce ATP.

The extraocular muscles exhibit the greatest diversity among mammalian skeletal muscles, a likely consequence of the varied functional requirements imposed by the ocular motor system. Extraocular muscle fibers differ from typical limb and respiratory skeletal muscles in mitochondrial content, innervation/contractile patterns, contractile protein isoforms, hormone receptors, cell surface markers, and a variety of other cell and molecular properties that may relate to their unique functions. The divergence from the skeletal muscle stereotype is further exemplified by the fact that extraocular muscles do not conform to traditional fiber type classifications, which are based primarily on myosin isoform expression. The most

accepted fiber-type classification scheme for extraocular muscles includes six fiber types based upon: (1) distribution into orbital and global layers, (2) innervation status, single versus multiple nerve contacts per fiber, and (3) mitochondrial/oxidative enzyme content. Figure 1 shows representative micrographs of extensor digitorum longus (EDL, which is a predominantly type IIB fiber – fast, fatigable – limb skeletal muscle), diaphragm (mixed fiber type, fatigue-resistant respiratory muscle), and the extraocular muscle from rats. EDL is recruited sporadically and diaphragm is constantly active as is the extraocular muscle. Despite the wide difference in activity, EDL and diaphragm sections are mostly indistinguishable. For sure, there are important biochemical differences between the two muscles that reflect their specific adaptations to their respective activation patterns. However, the divergent requirements of EDL and diaphragm motor systems are met with fairly stereotypical muscle fibers, as evident from the micrographs. In contrast, the extraocular muscle fibers are very different: small round fibers with prominent mitochondria, suggestive of atypical contractile and metabolic properties. This article outlines newly identified unique aspects of extraocular muscle metabolism and how they may correspond to contractile function.

Insights from Gene Expression Profiling The fast and constant contractions of the extraocular muscles necessitate well-developed energy supply systems. It might be expected a priori that these muscles would upregulate all the main energy-supply metabolic pathways, from glycolysis to mitochondrial metabolism. Surprisingly, this may only apply to mitochondrial content, which is the highest reported in mammalian skeletal muscles. Studies comparing the gene expression profile of extraocular and limb muscles found that genes coding for key enzymes of glycogen synthesis and breakdown were repressed in the extraocular muscles. Glycogen content in the extraocular muscles is correspondingly reduced. These findings indicate that the extraocular muscles are seemingly less dependent on stored glycogen as a metabolic fuel than other skeletal muscles. They also suggest that the extraocular muscles rely, instead, on constant transport of blood-borne glucose and fatty acids through their extensive microvascular network. Interestingly, the expression of the lactate dehydrogenase (LDH) isoform that preferentially oxidizes lactate to pyruvate is increased

27

28

Extraocular and Eyelid Muscles: Structure, Function, and Pathophysiology

Figure 1 The extraocular muscles are not typical skeletal muscles. The figure shows representative micrographs of limb (left, extensor digitorum longus), respiratory (middle, diaphragm), and extraocular (right) muscle sections stained with Gomori’s trichrome. This technique stains mitochondria (and sarcoplasmic reticulum) a darker reddish-blue. Despite the differences in functional profiles (occasionally active limb muscle vs. constantly active diaphragm), notice the similarity in fiber size and shape: large polygonal muscle fibers with peripheral myonuclei (left and middle panels). Some fibers have darker cytosolic staining indicating higher mitochondrial content. In contrast, the extraocular muscle shows significantly smaller fibers with clumpy cytosolic and subsarcolemmal staining due to abundant mitochondria (right panel). Scale bars ¼ 25 mm.

in the extraocular muscles, compared to typical skeletal muscles, and that may allow them to use lactate as a fuel for aerobic pathways. The importance of these alternative metabolic pathways is now being tested in extraocular muscles.

Lactate: An Oxidizable Substrate for the Extraocular Muscle In limb skeletal muscles, glycogen breakdown drives glycolysis only during brief bursts of intense activity (Figure 2). In most muscles, metabolic demand during moderate activity is met by aerobic (mitochondrial) pathways. During periods of sustained peak activity, when mitochondrial capacity is exceeded in skeletal muscle, lactate is the end product of glycolysis in a reaction that reduces pyruvate at the expense of NADH and is catalyzed by LDH. For this and other reasons, increased production and accumulation of lactic acid during exercise has been associated with muscle fatigue. However, cells can also use the LDH reaction in the reverse direction from lactate oxidation to pyruvate, and lactate then goes on to become a substrate for aerobic metabolism. As mentioned above, the expression of the LDH isoform that preferentially oxidizes lactate to pyruvate is higher in the extraocular muscles. Combined with their high aerobic capacity, this reaction would allow the extraocular muscles to use lactate as a metabolic substrate. Cinnamate, a blocker of lactate transport, alone or in combination with exogenous lactate can be used to evaluate the role of lactate on fatigue resistance. Cinnamate accelerates fatigue in the extraocular muscles significantly: treated muscles lose their ability to generate force at a faster rate than untreated extraocular muscles. Conversely, cinnamate treatment does not affect the endurance or residual force of limb muscles. Replacing glucose with exogenous

lactate increases limb-muscle fatigability but has no effect on the extraocular muscles. However, the extraocular muscles fatigue faster when exposed to exogenous lactate combined with cinnamate treatment. These results indicate that LDH oxidation of lactate to pyruvate seems to be an important source of metabolic substrate for aerobic metabolism in the extraocular muscles. This conclusion is a significant deviation from the traditional view of lactate as a final waste product of glycolysis; increased lactate production and accumulation during vigorous contractile activity is typically associated with fatigue. Muscle fatigue is a complex phenomenon: substrate depletion, metabolite accumulation, and ionic imbalances are some of the factors that combine to reversibly impair contractile function. In the particular case of the extraocular muscles, lactate can be used via the LDH reaction as an additional substrate source for aerobic metabolism, a concept developed recently for other aerobic muscles and one that also applies to the nervous system.

Creatine Kinase, the Missing ATP Buffer in the Extraocular Muscle Skeletal muscles and other tissues with fluctuating metabolic needs rely on the creatine–phosphocreatine system to buffer intracellular ATP concentration: creatine kinase (CK) catalyzes the reversible transfer of the phosphoryl group from phosphocreatine to ADP in order to maintain constant ATP levels. Cellular CK activity is due to a family of oligomeric enzymes: two cytosolic, ubiquitous brain-type CK-B and muscle-type CK-M, and two mitochondrial isoforms, ubiquitous mitochondrial CK (uCK) and sarcomeric mitochondrial CK (sCK). In differentiated skeletal muscle, CK-MM and sCK are the predominant isoforms. In fast-twitch muscles, most CK activity is due to the CK-MM isoform, some of which is found

Extraocular Muscles: Extraocular Muscle Metabolism

29

Glycolysis Glucose ATP ADP Glucose 6-phosphate

Fructose 6-phosphate ATP ADP Fructose 1,6-diphosphate

Glyceraldehyde 3-phosphate

Glyceraldehyde 3-phosphate

NAD

2 ADP

2 ADP

NAD

NADH

2 ATP

2 ATP

NADH

NADH

Pyruvate

Pyruvate

NAD

NAD Lactate

NADH

(Krebs cycle)

Lactate

Figure 2 Glycolysis, the anaerobic breakdown of glucose. The figure presents a diagram showing the sequence of steps in glycolysis, from glucose to pyruvate. The end product, pyruvate, may move on to the Krebs cycle (inside the mitochondria) to continue substrate oxidation, or it may be reduced to lactate in the reaction catalyzed by LDH that restores NAD. Dashed arrows represent omitted steps.

associated with the sarcomeric M line, the sarcoplasmic reticulum, and T-tubules. This arrangement couples the CK-dependent ATP-buffering system to the cellular sites with the highest ATPase activity, and is needed for normal contractile function. Given the predicted need for ATP buffering in the extraocular muscles, we proposed that (1) CK isoform expression and activity in rat extraocular muscles would be higher and (2) the resistance of these muscles to fatigue would depend on CK activity. Instead, we found that messenger ribonucleic acid (mRNA) and protein levels for all (cytosolic and mitochondrial) CK isoforms are lower in the extraocular muscles than in limb muscles. The muscle-enriched isoforms, CK-M and sCK, are less abundant in extraocular muscle, despite the fact that the extraocular muscles have a higher mitochondrial content than limb muscles. Total CK activity is also correspondingly decreased in the extraocular muscles. Moreover, cytoskeletal components of the sarcomeric M line, where a significant fraction of cytosolic CK activity is found, are downregulated in the extraocular muscles as was initially suggested by gene expression profiling. To explore the role of CK activity on muscle function, the CK inhibitor 2,4-dinitro-1-fluorobenzene (DNFB) was used during an in vitro fatigue protocol. Treatment with DFNB accelerates the development of fatigue in limb muscle, but has no detectable effect on the

extraocular muscles. These data support the conclusion that CK activity is not an important ATP buffer in the extraocular muscles. The myokinase reaction (2 ADP ! ATP þ AMP), catalyzed by adenylate kinase (AK), serves as an additional ATP-buffering system in skeletal muscle. While total AK activity is similar in extraocular and limb muscles, the mRNA content for two putative mitochondrial AK isoforms (AK3 and AK4) is over 13-fold more abundant in the extraocular muscles. This suggests that the relative lack of CK in the extraocular muscles may be compensated by upregulation of selected AK isoforms.

Mitochondrial Content in the Extraocular Muscles Aerobic capacity is typically measured by mitochondrial volume density (percentage of muscle fiber volume occupied by mitochondria). In general, mitochondrial volume density is well matched to the metabolic needs of skeletal muscle and it scales almost linearly with maximal oxygen uptake among muscles and across mammalian species. In other words, the consensus is that changes in the oxidative (aerobic) capacity of mammalian skeletal muscles are met by corresponding increases or decreases in mitochondrial volume density. Since the mitochondrial content and the

30

Extraocular and Eyelid Muscles: Structure, Function, and Pathophysiology

activity of respiratory complexes and enzymes of mitochondrial metabolic pathways change in parallel, enzymatic activities are used as indices of mitochondrial content and aerobic capacity. Highly aerobic muscle groups in mammals have abundant capillaries and elevated mitochondrial volume density. The extraocular muscles have, arguably, the highest mitochondrial content of all mammalian skeletal muscles. However, the mechanism responsible for maintaining mitochondrial abundance in the extraocular muscles remains unclear. We have identified a number of transcription factors that influence mitochondrial biogenesis and that are upregulated in the extraocular muscles. Surprisingly, these factors are different from the mitochondrial biogenesis program initiated in response to endurance training.

Mitochondria as Calcium Sinks in the Extraocular Muscle The fast-contracting extraocular muscles rely on tight regulation of free cytosolic calcium concentration ([Ca2+]i). In principle, the extraocular muscles have the profile of very efficient calcium handling capacity: extensive and welldeveloped sarcoplasmic reticulum and the expression of fast calcium ATPase isoforms. Moreover, the extraocular muscles contain parvalbumin, a low-weight calcium-binding protein that serves as a temporary buffer to accelerate the removal of calcium off its binding sites on the myofilaments and facilitates muscle relaxation. Other investigators have already shown that the kinetics of calcium flux into mitochondria are fast enough to influence very rapid events such as neurotransmitter release from motor nerve terminals. The specific inhibition of mitochondrial calcium transport slows the relaxation of mitochondria-rich skeletal muscles. We recently reported that the magnitude and speed of calcium uptake by mitochondria are sufficient to influence contractile function. This property of the extraocular muscles appears to serve at least two complementary functions. First, it couples metabolic supply to demand because higher mitochondrial calcium stimulates enzymes that control substrate oxidation: pyruvate dehydrogenase, 2-oxoglutarate dehydrogenase, isocitrate dehydrogenase, and glycerol 3-phosphate dehydrogenase. The combined activity of these enzymes sustains a high NADH/NAD+ level and maximizes oxidative phosphorylation and ATP production. Second, by limiting the [Ca2+]i increase during contractions in response to submaximal stimulation frequencies, mitochondria widen the dynamic range of the extraocular muscles. In other words, the capacity of the extraocular muscles to produce force is spread over a wider stimulation frequency range, increasing the fine control of the effector arm of the ocular motor system.

Are Extraocular Muscle Mitochondria Different? The adenine nucleotide transporter 1 (Ant1) gene encodes an inner mitochondrial membrane protein that transports ADP into mitochondria and ATP from mitochondria to the cytosol. Mutations within Ant1 have been shown to produce a syndrome of chronic progressive external ophthalmoplegia (CPEO) in humans. Ant1 knockout (Ant1–/–) mice develop cardiomyopathy and severe exercise intolerance. Despite this dramatic phenotype, the extraocular muscles are mostly unaffected. Histologically, the extraocular muscles from Ant1–/– mice present a relatively mild mitochondrial myopathy. There are no measurable ocular motor abnormalities in Ant1–/– mice, and their peak eye velocities overlap with those measured in control mice. Moreover, their extraocular muscles do not show evidence of increased fatigability. In addition, the extraocular muscles have higher levels of Ant2 mRNA compared to the limb muscles. Ant2 is a nonskeletal muscle isoform previously described in the heart. Its presence in the extraocular muscles may explain the lack of effects of Ant1 loss, and it was the first documented difference between extraocular muscle and limb muscle mitochondria. The ability of muscles to perform aerobic work depends on their mitochondrial volume density, with the assumption that the composition of these organelles is fairly constant across muscle types and mammalian species. One of these components is the electron transport chain, a series of multimeric complexes (complexes I–IV, plus the ATP synthase which is sometimes called complex V) in the inner mitochondrial membrane responsible for most of the aerobic ATP generation (Figure 3). Recently, we found that the extraocular muscle mitochondria have lower content or lower activity of some enzyme complexes of the electron transport system, causing them to respire at slower rates. This is puzzling given that the extraocular muscles are constantly active and aerobic capacity was predicted to be elevated, given their high mitochondrial content. These findings are not explained by differences in the ultrastructure of extraocular muscle mitochondria: the surface area of their inner membrane is comparable to values reported for other skeletal muscle. Furthermore, the differences are not generalized or systematic: complex II content and activity, and complex III content are similar in mitochondria from triceps surae (a limb skeletal muscle) and extraocular muscle. Complexes I and IV give a more puzzling result: their activities are lower, but their content is higher in the extraocular muscle mitochondria. These are multimeric protein complexes, and differential expression of isoforms of some subunits has been described in skeletal muscle and other tissues.

Extraocular Muscles: Extraocular Muscle Metabolism

31

Outer membrane +

H+

H

H+ Intermembrane space

lll

l

lV

ll ATP synthase

H ADP + Pi

+

ATP

H2O

Krebs cycle

Matrix Fumarate

H2O

O2

NADH

NAD+ + H+

Succinate

O2 lV

lll

H+

ll

Inner membrane l

H+

Figure 3 Mitochondrial electron transport chain and ATP synthase. The figure presents a drawing showing the oxidative phosphorylation steps that couple the final substrate oxidation to the reduction of oxygen to water, pumping hydrogen ions (protons, Hþ) from the mitochondrial matrix to the intermembrane space. The substrates for this chain are NADH or succinate, shown here as originating from the Krebs cycle. Complex I (NADH dehydrogenase) oxidizes NADH and transfers the electrons to complex III, which in turn transfers the electrons to complex IV (cytochrome c oxidase). The latter is the complex that reduces oxygen to water. Complex II (succinate dehydrogenase) is not an Hþ pump; it funnels electrons from succinate to complex III, and then complex IV. The ATP synthase (complex V) is driven by the trans-inner membrane electrochemical potential generated by the movement of Hþ to the intermembrane space.

Therefore, the content of some electron transport chain complexes (I, IV, and V) and the subunit composition of some others (I and IV) may not be the same in the extraocular muscles compared to limb muscles. This demonstrates that the metabolic divergence between extraocular and limb muscles includes major differences in the composition and basic function of their respective mitochondrial populations. Intrinsic differences in mitochondrial structure and function may explain the susceptibility of the extraocular muscles to some hereditary and acquired mitochondrial myopathies such as CPEO and related syndromes. For example, the extraocular muscles present the most severe age-dependent loss of mitochondrial respiratory complex activity among muscles. There is a significant increase in the number of fibers with cytochrome c oxidase defects in the extraocular muscles of humans and other primates, even when compared to other highly aerobic muscles such as the diaphragm and heart. This can be at least partially explained by mitochondrial DNA mutations, presumably due to reactive oxygen species generated during mitochondrial respiration or present as part of a more generalized cellular oxidative stress.

Matching Mitochondrial Capacity to Contractile Function The primary role of mitochondria is to generate ATP. Recent studies lead to an obvious question: How do extraocular muscles sustain their contractile function with mitochondria that respire half as fast as mitochondria from other muscles? The content of respiratory complexes is one parameter behind tissue variations in mitochondrial respiration, although some argue that it is not particularly relevant for metabolic control. Under experimental conditions, mitochondrial respiration in the skeletal muscle and heart is regulated at the level of the respiratory chain, while in the liver, kidney, and brain it is controlled mainly at the phosphorylation level by ATP synthase (complex V) and phosphate carrier. That may not be the case in vivo, where different parameters such as cellular steady state, the energy demand, and the energy supply of the tissue may also regulate mitochondrial respiration. In the case of the extraocular muscles, allosteric regulation of respiratory complexes may combine with changing metabolite concentrations to maintain mitochondrial respiration closer to its theoretical maximum.

32

Extraocular and Eyelid Muscles: Structure, Function, and Pathophysiology

For example, a mechanism to enhance energy production in the extraocular muscle is mitochondrial calcium influx during contractile activity in order to activate enzyme systems that exert strong control on substrate oxidation, as mentioned above.

Acknowledgements The author’s work in this field is supported by the National Eye Institute (grant R01 EY012998). See also: Extraocular Muscles: Extraocular Muscle Anatomy.

Matching Energy Supply to Demand Initially inspired by morphological characteristics and gene-expression-profiling results, a more global perspective of extraocular muscle metabolism is beginning to emerge. First, glycogen content is low and the glycogenolysis pathway seems to be downregulated in the extraocular muscles. Second, CK activity and content, including the mitochondrial isoform, are lower in the extraocular muscles, indicating that phosphocreatine may be a less important temporal and spatial ATP buffer in these muscles. In other words, mitochondrial ATP production may be sufficiently high and close to cellular sinks as to obviate the need for an energy buffer. Third, the extraocular muscles can use lactate as an oxidizable substrate due to the presence of a LDH isoform that catalyzes the conversion of lactate to pyruvate that then goes to the Krebs cycle. Finally, the mitochondrial population in extraocular muscles appears to respond to a different biogenesis program, and exhibits atypical functional characteristics that may influence the contractile activity of these muscles significantly.

Further Reading Andrade, F. H. and McMullen, C. A. (2006). Lactate is a metabolic substrate that sustains extraocular muscle function. Pflu¨gers Archiv-European Journal of Physiology 452: 102–108. Andrade, F. H., McMullen, C. A., and Rumbaut, R. E. (2005). Mitochondria are fast Ca2+ sinks in rat extraocular muscle: A novel regulatory influence on contractile function and metabolism. Investigative Ophthalmology and Visual Science 46: 4541–4547. McMullen, C. A., Hayeß, K., and Andrade, F. H. (2005). Fatigue resistance of rat extraocular muscles does not depend on creatine kinase activity. BMC Physiology 5: 12. Porter, J. D., Khanna, S., Kaminski, H. J., et al. (2001). Extraocular muscle is defined by a fundamentally distinct gene expression profile. Proceedings of the National Academy of Sciences of the Unites States of America 98: 12062–12067. Spencer, R. F. and Porter, J. D. (2006). Biological organization of the extraocular muscles. Progress in Brain Research 151: 43–80. Yin, H., Stahl, J. S., Andrade, F. H., et al. (2005). Eliminating the Ant1 isoform produces a mouse with CPEO pathology but normal ocular motility. Investigative Ophthalmology and Visual Science 46: 4555–4562.

Extraocular Muscles: Proprioception and Proprioceptors R Blumer, Medical University of Vienna, Vienna, Austria ã 2010 Elsevier Ltd. All rights reserved.

Glossary Choline acetyltransferase – The enzyme responsible for the synthesis of the neurotransmitter acetylcholine, causing the transfer of acetate to choline. Choline transporter – These recapture choline from the synaptic cleft after acetylcholine release and degradation. This process is critical for new acetylcholine synthesis at the synapse. Golgi tendon organ – A proprioceptive organ that provides information to the brain about changes in muscle tension. In contrast to muscle spindles, these are in series with muscle fibers, interwoven with the collagen in the muscle tendon. Tension on the tendon caused by muscle contraction activates these proprioceptors. Muscle spindles – Proprioceptive organs that provide the brain with information about changes in muscle length and are organized in parallel with the skeletal muscle fibers. They contain modified muscle fibers called intrafusal fibers, in contrast with the muscle fibers themselves, which are called extrafusal fibers. They have a complex structure; they are surrounded by a connective tissue capsule, and contain several types of modified myofibers within them. They are innervated by sensory afferents. Myotendinous cylinders or palisade endings – Proprioceptive organs found at the myotendinous junction consisting of dense axonal branching which invests the tips of single muscle fibers. These are unique to the extraocular muscles. The nerves establish synaptic contacts with both collagen fibrils and muscles fibers at the myotendinous junction. The function of these structures is unknown. Proprioception – The sense that provides information about the location of various parts of the body in relation to each other and in relation to the space. Proprioceptors – Sensory receptors which are found in muscles and tendons that bring sense of body position to the brain. Vesicular acetylcholine transporter – A membrane protein which is necessary for the uptake of acetylcholine into synaptic vesicles.

Proprioception Proprioception refers to a sense that provides information about the location of various parts of the body in relation to each other and in relation to space. It is of practical importance for activities in everyday life and allows a person to use the foot pedal of a car properly while driving or to learn to walk in darkness. Moreover, sportsmen use specific training devices to sharpen their proprioceptive sense. Proprioceptive signals come from specialized sensory nerve endings called proprioceptors that occur throughout skeletal muscle. Typical proprioceptors in skeletal muscle are muscle spindles and Golgi tendon organs which constantly transmit information to the brain. In this way the brain knows, at any given time, the spatial position of our body parts. The eyes are the most mobile organs of the body, and vision is useful only if the brain knows the position of the eyes in the orbit. By knowing where the eyes are pointing, the brain is aware of the position of objects in the surrounding space: if objects are leftwards, straight ahead, or rightwards. Several studies indicate that the brain has access to proprioceptive information from the extraocular muscles (EOMs). Specifically, neuronal tracing experiments have demonstrated projections from the EOM to various peripheral and central nervous system structures, including the trigeminal ganglion, the mesencephalic trigeminal nucleus, the superior colliculus, the vestibular nuclei, and the cerebellum. A recent physiology experiment showed that the primary somatosensory cortex, which receives proprioceptive input from all other skeletal muscles, also receives signals from EOM. This new finding completes the somatotopic representation of the body in the primary somatosensory cortex which, thus far, had lacked a map of the eye muscles. Indication that there is proprioceptive input from the EOM has also come from psychophysical investigations. Patients suffering from strabismus were tested after surgery, and it was detected that they had deficits in spatial perception. These results were interpreted to mean that the surgical intervention has damaged the proprioceptors at the myotendinous junction resulting in a loss of eye position signals. Despite this evidence for EOM proprioception, there are also counterarguments. Specifically, no stretch reflex has been observed in the EOM of monkey. By cutting the ophthalmic nerve, which is supposed to carry the afferent fibers from EOM, deficits in eye movements would be

33

34

Extraocular and Eyelid Muscles: Structure, Function, and Pathophysiology

expected. However, findings indicate that deafferentation does not affect ocular alignment or eye movements, including saccades and smooth pursuit. Such observations led scientists to doubt whether EOM proprioception really exists. Instead, it has been hypothesized that the motor command that is sent to the EOMs is copied, called efference copy, and this copy provides the necessary information for the brain to be aware of the eye’s position. If there is sensory feedback from EOM, the eye muscles should have proprioceptors. In the last century, the EOMs of several mammalian species and man have been screened for muscle spindles and Golgi tendon organs. Interestingly, the endowment with classical proprioceptors varies widely among the species, and there are even some species that do not have proprioceptors at all. In view of these interspecies variations, it is not clear where the source of EOM proprioception lies. By searching for alternative sensory organs in the EOMs, palisade endings (also called myotendinous cylinders) have been detected, and so far, palisade endings have been observed in each species investigated. In the following section, we discuss EOM proprioceptors, including muscle spindles, Golgi tendon organs, and palisade endings. We give an overview about the occurrence, distribution, number, and structure of these organs and speculate about their putative function. Recent studies have focused on the molecular characteristics of palisade endings, and we also refer to these findings.

Muscle Spindles Occurrence, Distribution, and Number of Muscle Spindles Muscle spindles are regularly observed in the EOMs of even-toed ungulates (sheep, cow, camel, goat, and pig) and in the EOMs of primates (monkey and man). In other animal species, including fellidae (cat), rodents (rat and guinea pig), odd-toed ungulates (horse), and lagomorphs (rabbit), muscle spindles have not been found (Table 1). In the EOMs of even-toed ungulates, muscle spindles are uniformly distributed throughout the entire muscle length. The number of muscle spindles is remarkably high, and counts per muscle yield between 146 and 333 muscle spindles in pig; between 100 and 181 muscle spindles in camel; and more than 200 muscle spindles in cow. In man the distribution of muscle spindles exhibits differences when compared with that in even-toed ungulates. Specifically, muscle spindles are located predominantly in the proximal and distal parts of the EOM, and each muscle has a spindle-free zone approximately in the middle. The number of human EOM spindles varies between 13 and 42. Only in the inferior oblique muscle has a lower number of muscle spindles been counted (3–7). The density of human EOM spindles is comparable

Table 1 Occurrence of muscle spindles, Golgi tendon organs, and palisade endings in the extraocular muscles of man and mammals

Species

Muscle spindles

Golgi tendon organs

Man Monkey Fellidae Even-toed ungulates Odd-toed ungulates Lagomorphs Rodents

þ þ þ

þ

a

Palisade endings þ þ þa

b

þ þ

So far, palisade endings have only been demonstrated in sheep. So far not analyzed for palisade endings.

b

to that of muscle spindles in finely controlled skeletal muscle such as the hand lumbrical and deep dorsal neck muscles. In monkey (rhesus monkey and cynomolgus monkey), very few muscle spindles (2–6) have been observed in some EOMs, and none in the others. Structure of Muscle Spindles Muscle spindles in the EOMs of even-toed ungulates conform in their structure with those in other skeletal muscles. The muscle spindles have a fusiform shape with a wide central region (equatorial region) and two narrow polar regions. Muscle spindles are ensheathed by a capsule consisting of several layers of perineural cells. The capsule space is filled with a viscous fluid containing acidic mucopolysaccharides. Inside the capsule two types of intrafusal muscle fibers (nuclear chain fibers and nuclear bag fibers) can be distinguished, which both exhibit modifications concerning their myonuclei in the spindle’s equatorial region. Nuclear chain fibers have a single row of centrally arranged nuclei, whereas nuclear bag fibers show an accumulation of nuclei. In the muscle spindle’s equatorial region, a large tissue-free space (periaxial space) separates the intrafusal muscle fibers from the capsule. Muscle spindles in the EOMs of even-toed ungulates receive a double innervation from sensory and motor nerve fibers. In the equatorial region, both types of intrafusal muscle fibers are endowed with sensory nerve endings (annulospiral sensory endings) which are wrapped spirally around the muscle fibers. Whether a second type of sensory nerve ending (flower-spray ending) that is common in other mammalian skeletal muscle spindles is also present in ungulate EOM spindles is unclear. Fine structural analyses have shown that sensory nerve terminals contain mitochondria and a few clear vesicles. The synaptic cleft separating the nerve terminal from the muscle fiber surface is free from a basal lamina. At the muscle spindle’s pole, intrafusal muscle fibers receive motor terminals. Motor terminals contain mitochondria

Extraocular Muscles: Proprioception and Proprioceptors

and dense aggregations of clear vesicles, and the synaptic cleft is filled with a basal lamina. Muscle spindles in EOM of primates exhibit structural differences when compared with those in even-toed ungulates. Specifically, in most muscle spindles of monkey and man the periaxial space exhibits little or no expansion. Only in human infants have some muscle spindles with a wide periaxial space been observed. Thorough analyses of the intrafusal fiber composition have been done in human EOM spindles. The findings indicate that human EOM spindles contain nuclear chain fibers but most of them lack nuclear bag fibers. Only 2% of the spindles contain nuclear bag fibers and, when present, the bag region is poorly developed with only two nuclei lying side by side. In addition to nuclear chain fibers, anomalous muscle fibers are also regularly observed in human EOM spindles. Anomalous muscle fibers exhibit no nuclear modification in the spindle’s equatorial region and are indistinguishable from muscle fibers outside the spindle. The unique morphology of human EOM spindles was initially described in aged persons (67–83 years old) and later was confirmed in infants (Figure 1(a)). The innervation pattern of primate EOM spindles has only been analyzed in humans. In human EOM spindles, sensory nerve endings have been observed on nuclear chain and, when present, on nuclear bag fibers, but only 7% of the anomalous fibers are endowed with sensory nerve terminals. In their fine structure, sensory nerve terminals in human EOM spindles do not differ from sensory nerve terminals in EOM spindles of even-toed ungulates. At the muscle spindle’s pole, intrafusal muscle fibers are equipped with motor terminals. Motor terminals in human EOM spindles are identical in their structure with those in EOM spindles of even-toed ungulates (Figure 1(b)).

35

Function of Muscle Spindles Muscle spindles in mammalian skeletal muscle are stretch receptors which register changes in muscle length. Indications that muscle spindles in the EOMs of even-toed ungulates are capable of monitoring muscle length have come from electrophysiological investigations. Specifically, in goat and sheep the EOMs were stretched and afferent signals were recorded in the sensory trigeminal ganglion. Recorded signals exhibited characteristics that are the same as muscle spindles in other skeletal muscles. There is controversy whether muscle spindles in human EOMs are functional. Due to their unusual morphology, some authors suppose that human EOM muscle spindles are not functional. On the other hand, muscle spindles in human EOMs are numerous, and their nerve terminals exhibit a normal morphology. This is why other authors suggest that human EOM spindles are functional, and their unusual morphology might indicate special functional properties. In particular, as most human EOM muscle spindles lack nuclear bag fibers, muscle spindles might have a predominantly static function and monitor the degree of muscle stretch rather than the contraction velocity of muscle fibers.

Golgi Tendon Organs Occurrence, Distribution, and Number of Golgi Tendon Organs Golgi tendon organs are exclusively found in the EOMs of even-toed ungulates (pig, sheep, camel, and cow). They have not been found in other mammals and man. In even-toed ungulates, Golgi tendon organs are distributed throughout the proximal and distal EOM tendons, their number always being higher in the distal tendons (Table 1). The number of Golgi tendon organs per muscle has been counted to be

C

BL N

AF ST

NC

(a)

(b)

Figure 1 (a) Semi-thin cross section through an extraocular muscle spindle of a 2-year-old human infant and (b) ultra-thin cross section through a nuclear chain fiber. (a) The muscle spindle contains six nuclear chain fibers (NF) and one anomalous fiber (AF). The anomalous fiber is indistinguishable from muscle fibers outside the spindle. N, nerve and C, capsule. Scale bar ¼ 100 mm. (b) Nuclear chain fiber (NF) with a sensory nerve terminal (ST). BL, basal lamina. Scale bar ¼ 1 mm.

36

Extraocular and Eyelid Muscles: Structure, Function, and Pathophysiology

46–128 and 30–90 in pig and camel, respectively. In both species, Golgi tendon organs are more frequent in the rectus EOMs than in the oblique EOMs. Structure of Golgi Tendon Organs Golgi tendon organs in EOMs of even-toed ungulates exhibit a fusiform shape and are enclosed by a capsule of perineural cells. The capsule space is filled with a viscous fluid containing acidic mucopolysaccharides. The main component of the Golgi tendon organs are collagen bundles that pass through the organ. At one end of the organ the collagen fascicles are attached to muscle fibers, and at the other end the fascicles merge with the tendon of the muscle. Many Golgi tendon organs contain only collagen bundles, but there are others which contain both collagen bundles and muscle fibers. Such intracapsular muscle fibers penetrate the Golgi tendon organ at one end and either terminate in collagen bundles or, more rarely, pass through the tendon organ. All Golgi tendon organs in the EOMs of even-toed ungulates exhibit a wide space in the central region that separates the collagen bundles and muscle fibers, if present, from the capsule (Figure 2(a)). Each Golgi tendon organ is innervated by a single sensory nerve fiber. The nerve fiber penetrates the capsule at various points. Inside the organ, the axon divides into several preterminal branches which finally establish nerve terminals that contact the surrounding collagen fibrils. Nerve terminals are only partly covered with Schwann cells, and at the area of contact only a basal lamina lies between the nerve terminal and the neighboring collagen. Nerve terminals contain mitochondria and a few clear vesicles (Figure 2(b)).

With the exception of intracapsular muscle fibers and a more pronounced capsule space in the central region, Golgi tendon organs in even-toed ungulates share the structural features of Golgi tendon organs found in other mammalian skeletal muscles. Function of Golgi Tendon Organs Golgi tendon organs in mammalian skeletal muscle are sensitive to muscle contraction. During muscle fiber contraction collagen bundles are stretched, and the nerve terminals within the collagen are deformed, thereby generating a receptor potential. Golgi tendon organs in EOMs of ungulates are supposed to function analogously and register muscle fiber contraction. Muscle fibers passing through Golgi tendon organs are supposed to regulate the sensitivity of the organ.

Palisade Endings Occurrence, Distribution, and Number of Palisade Endings Palisade endings (myotendinous cylinders) are sensory end organs that are unique to EOMs. So far, palisade endings have been found in the EOMs of almost all species investigated, including fellidae (cat), lagomorphs (rabbit), even-toed ungulates (sheep), rodentia (rat), and primates (monkey and man). These organs are located at the distal and proximal myotendinous junctions. Palisade endings are plentiful in the EOMs of monkey and cat (Table 1). In the distal myotendon of a monkey medial rectus 350 palisade endings have been counted, and in a

S

N

ST

MF COL C

COL (a)

(b)

Figure 2 (a) Semi-thin cross section through a Golgi tendon organ of a calf extraocular muscle and (b) ultrathin section through a sensory nerve terminal. (a) The Golgi tendon organ is ensheathed by a capsule (C) and contains collagen bundles (COL) and one muscle fiber (MF). Nerve fiber (N). Scale bar ¼ 100 mm. (b) A sensory nerve terminal (ST) which is partly ensheathed by Schwann cells (S) contacts the surrounding collagen bundles (COL). Scale bar ¼ 1 mm.

Extraocular Muscles: Proprioception and Proprioceptors

cat medial rectus 94. A smaller number of this EOMspecific organ have been found in the distal EOM myotendons of rat (27) and human (20–30). Structure of Palisade Endings Innervation for the palisade endings arises from nerve fibers that come from the muscle and extend into the tendon. Within the tendon, the nerve fibers make a 180 loop and return to the muscle. At the muscle–tendon junction, the returning axons divide into preterminal branches. Preterminal axons establish nerve terminals around the muscle fiber tips which have the appearance of a palisade fence (Figure 3), which is also the reason why this formation is called a palisade ending. The whole palisade complex is ensheathed by a capsule of fibroblasts. Palisade ending is exclusively associated with the multiply innervated muscle fibers of the global (inner) layer of the EOMs. Such muscle fibers have several motor contacts along their length, and with respect to contraction they exhibit nontwitch characteristics. The multiply innervated muscle fibers have a unique innervation from small motoneurons located outside the borders of the main EOM nuclei. The fine structure of palisade endings was initially analyzed in cat and monkey and later in sheep, rabbit, and man. It was observed that the majority of palisade nerve terminals contact the collagen fibrils of the tendon, and only a few of them contact the muscle fiber tip. Nerve terminals contacting the collagen fibrils are only partly enwrapped with Schwann cells, and at the area of contact with the collagen only a basal lamina covers the nerve terminals. Such neurotendinous contacts contain dense aggregations of clear vesicles and mitochondria. Palisade nerve terminals contacting the muscle fiber are free from a basal lamina in the synaptic cleft, thereby resembling sensory nerve terminals on intrafusal fibers of muscle spindles. Identical to neurotendinous contacts, neuromuscular contacts contain mitochondria and a large number of clear vesicles. Interestingly, in palisade endings of man and monkey, neuromuscular contacts have a basal lamina in the synaptic cleft which is a feature of motor terminals. Palisade

37

endings in rabbits and rats are an exception. In both species, the palisade endings lack neurotendinous contacts and neuromuscular contacts are present exclusively.

Molecular Characteristics of Palisade Endings In cat and monkey, it has been recently demonstrated that palisade endings have a cholinergic phenotype. Utilizing immunohistochemistry, palisade endings have been labeled with all commercially available cholinergic markers, including antibodies against choline transporter (ChT), choline acetyltransferase (ChAT), and vesicular acetyl choline transporter (VAChT), as well as a-bungarotoxin. In the nervous system, ChT is used for the uptake of choline, ChAT is the synthesizing enzyme of acetylcholine, and VAChT is used to transport acetylcholine into the synaptic vesicles. a-Bungarotoxin is a snake venom that binds to nicotinic acetylcholine receptors, and this neurotoxin is widely used to detect motor terminals in skeletal muscle. In cat and monkey, it has been shown that the nerve fibers supplying palisade endings are ChAT immunoreactive. The palisade complexes, including palisade nerve terminals, are ChAT positive as well. In monkey, it also has been demonstrated that palisade nerve terminals exhibit ChT/VAChT immunoreactivity, and neuromuscular contacts, when present, exhibit a-bungarotoxin binding. Finally, in some cases it has been detected that nerve fibers supplying palisade endings establish a-bungarotoxin-positive neuromuscular contacts outside the palisade complex (Figure 3(b)). Function of Palisade Endings So far, physiological studies on palisade ending are missing, and their function remains speculative. Indication that palisade endings are sensory organs comes from morphological studies and a single neuronal tracing experiment. Specifically, morphological studies show that palisade endings have nerve terminals contacting the tendon, and nerve terminals in apposition to collagen are arguably sensory. Palisade nerve terminals contacting

Figure 3 Palisade endings: (a) three-dimensional reconstruction of a palisade ending and (b) palisade ending labeled with anti-neurofilament (general marker for nerve fibers) and anti-ChAT (marker for cholinergic nerve fibers). Muscle fibers are counterstained with phalloidin. The tendon is not labeled and is continuous with the muscle fiber tip to the right. (a) A nerve fiber (green) coming from the muscle extends into the tendon and turns back to establish nerve terminals (red) around a muscle fiber tip. The muscle fiber is white. (b) This shows a nerve fiber forming a palisade ending. The nerve fiber and the palisade ending are positive for neurofilament (red) and ChAT (green). Muscle fiber (white). Scale bar ¼ 100 mm.

38

Extraocular and Eyelid Muscles: Structure, Function, and Pathophysiology

the muscle fibers lack a basal lamina in the synaptic cleft which is common with sensory nerve terminals in muscle spindles. Moreover, by injecting neuronal tracer into the sensory trigeminal ganglion, structures resembling palisade endings have been labeled. Palisade endings lie in series to the multiply innervated muscle fibers of the global (inner) EOM layer, and it is supposed that palisade endings register contraction of such muscle fibers. Although a consensus exists that palisade endings are sensory, there are other findings which favor a motor role for palisade endings. In particular, immunohistochemical studies have demonstrated that palisade endings are cholinergic, a feature common with motor nerve terminals. Nerve fibers forming palisade endings also establish motor neuromuscular contacts outside the palisade complex. In a nerve degeneration experiment a lesion of the oculomotor nucleus was performed and, in addition to the expected loss of motor terminals, the palisade endings degenerate as well. The functional significance of palisade endings with a motor nature is difficult to predict. In particular, it is unclear what effect cholinergic neurotendinous contacts would have on the surrounding collagen. At the moment the function of palisade endings is still a matter of discussion, and for clarification physiological studies are highly warranted.

Further Reading Billig, I., Buisseret, D. C., and Buisseret, P. (1997). Identification of nerve endings in cat extraocular muscles. The Anatomical Record 248: 566–575. Blumer, R., Konakci, K. Z., Brugger, P. C., et al. (2003). Muscle spindles and Golgi tendon organs in bovine calf extraocular muscle studied by

means of double-fluorescent labeling, electron microscopy, and three-dimensional reconstruction. Experimental Eye Research 77: 447–462. Blumer, R., Lukas, J. R., Aigner, M., et al. (1999). Fine structural analysis of extraocular muscle spindles of a two-year-old human infant. Investigative Ophthalmology and Visual Science 40: 55–64. Blumer, R., Wasicky, R., Brugger, P. C., et al. (2001). Number, distribution and morphological particularities of encapsulated proprioceptors in pig extraocular muscle. Investigative Ophthalmology and Visual Science 42: 3085–3094. Blumer, R., Wasicky, R., and Lukas, J. R. (2001). Presence and structure of innervated myotendinous cylinders in rabbit extraocular muscle. Experimental Eye Research 73: 787–796. Buisseret, P. (1995). Influence of extraocular muscle proprioception on vision. Physiological Reviews 75: 323–338. Buttner, E. J. A., Konakci, K. Z., and Blumer, R. (2005). Sensory control of extraocular muscles. Progress in Brain Research 15(1): 81–93. Donaldson, I. M. L. (2000). The functions of proprioceptors of the eye muscles. Philosophical Transactions of the Royal Society of London 355: 1685–1754. Konakci, K. Z., Streicher, J., Hoetzenecker, W., et al. (2005). Molecular characteristics suggest an effector function of palisade endings. Investigative Ophthalmology and Visual Science 46: 155–165. Konakci, K. Z., Streicher, J., Hoetzenecker, W., et al. (2005). Palisade endings in extraocular muscles of the monkey are immunoreactive for choline acetyltransferase and vesicular acetylcholine transporter. Investigative Ophthalmology and Visual Science 46: 4548–4554. Lukas, J. R., Aigner, M., Blumer, R., Heinzl, H., and Mayr, R. (1994). Number and distribution of neuromuscular spindles in human extraocular muscles. Investigative Ophthalmology and Visual Science 35: 4317–4327. Lukas, J. R., Blumer, R., Denk, M., et al. (2000). Innervated myotendinous cylinders in human extraocular muscle. Investigative Ophthalmology and Visual Science 41: 2422–2431. Ruskell, G. L. (1989). The fine structure of human extraocular muscle spindles and their potential proprioceptive capacity. Journal of Anatomy 167: 199–214. Ruskell, G. L. (1990). Golgi tendon organs in the proximal tendon of sheep extraocular muscle. The Anatomical Record 227: 25–31. Ruskell, G. L. (1999). Extraocular muscle proprioceptors and proprioception. Progress in Retinal and Eye Research 18: 269–291.

Abnormal Eye Movements due to Disease of the Extraocular Muscles and Their Innervation A Serra, University of Sassari, Sassari, Italy R J Leigh, Case Western University, Cleveland, OH, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Amblyopia – Lazy eye with poor vision because of misalignment of the eyes during development. Chronic progressive ophthalmoplegia (CPEO) – This term describes a number of disorders affecting the extraocular muscles that lead to progressive limitation of eye motion. Diplopia – Double vision. Esotropia – Misalignment of the eyes, one of which turns in (toward the nose)–cross-eyed. Exotropia – Misalignment of the eyes, one of which turns out (away from the nose)–wall-eyed. Kearns–Sayre syndrome – One cause of CPEO (see above) that is inherited from the mother and affects other tissues, such as the heart. Myasthenia gravis – A disorder causing muscle fatigue that is due to failure of the nerves to stimulate the muscles to contract. Nystagmus – An oscillation of the eyes (shimmering or jumping eyes). Optokinetic reflexes – Eye movements induced by moving a visual pattern in front of the eyes. Ptosis – Droopy lids. Saccades – Rapid eye movements that are used to move the point of visual fixation from one feature of interest to the next. Smooth pursuit – Eye movements that smoothly follow a moving object, such as a bird in the sky. Strabismus – Misalignment of the eyes; the eyes point in different directions. Vestibulo-ocular reflexes – Eye movements induced by head movements, which stimulate the balance mechanism in the inner ear.

Introduction In this article we apply current knowledge of the extraocular muscles (EOMs) and their brainstem innervation to develop working hypotheses to account for a range of abnormal eye movements. To be concise, we have mainly selected diseases with well-defined processes that affect

specific sites, from muscle to premotor neurons in the brainstem (Figures 1 and 2). This bottom-up approach is somewhat reductionist and simplified, but we hope that it will provide insights for readers with a broad range of interests. For more comprehensive reviews, readers can turn to sources listed at the end of this article. A prerequisite for understanding disordered ocular motility is that eye movements can be systemically examined (Table 1). Thus, restricting attention to the most evident disturbance (e.g., strabismus) will impoverish interpretation of the underlying disorder. Conversely, considering the properties of saccades, pursuit, vestibular, and vergence eye movements, as well as the presence of any visual deficits, will enrich the understanding of the pathogenesis of the disorder.

Effects of Disorders of the EOMs on Eye Movements The EOMs (Figure 1, site 1) possess unique properties that make them resistant to some diseases and susceptible to others. Thus, on the one hand, the EOMs are spared in Duchenne muscular dystrophy, even when the disease is well advanced, a finding that has prompted much research. On the other hand, EOMs are rich in mitochondria, which is appropriate for the sustained contraction required for precise gaze control, but which makes them susceptible to mitochondrial disorders. Although such disorders may arise in childhood along with involvement of other tissues, such as heart muscle (Kearns–Sayre syndrome), these may present throughout adulthood with the syndrome of chronic progressive ophthalmoplegia (CPEO). Such individuals have ptosis and a limited range of eye movements. The complaint of diplopia is rare in CPEO, and although this had been ascribed to equal involvement of each of the eye muscles, another explanation seems more likely. Thus, slow progression of CPEO allows time for the visual system to adapt, and suppress images from one eye. Another interesting finding is that such patients may also make relatively quick eye movements (saccades or vestibular eye movements), despite a limited range of movement. This anomaly may be due to sparing of fast type myosin heavy chain (MyHC) EOM-specific global EOM fibers, which have fewer mitochondria.

39

40

Extraocular and Eyelid Muscles: Structure, Function, and Pathophysiology

1 MR

LR

2

Ach CN III

Oculomotor nerve

Nodulus

Vergence

Dentate nucl.

Abducens nerve ATD

3

5 6

ICP

MLF

PPRF EBN

glu/asp

glu/asp

glu SVN

CN VI Abd. nucl.

4

glu

/as

IBN

p

CN VII

gly gly

gly

CTT MLF CN VI

PPRF MVN NPH MLF IVN

ICP

Ach

8

7

NPH

MVN

MRF Inf. olivary nucl.

Figure 1 Anatomic scheme for the synthesis of signals for horizontal eye movements. The abducens nucleus (CN VI) contains abducens motoneurons that innervate the ipsilateral lateral rectus muscle (LR), and abducens internuclear neurons that send an ascending projection in the contralateral MLF to contact medial rectus (MR) motoneurons in the contralateral third nerve nucleus (CN III). From the horizontal semicircular canal, primary afferents on the vestibular nerve project mainly to the MVN, where they synapse and then send an excitatory connection to the contralateral abducens nucleus and an inhibitory projection to the ipsilateral abducens nucleus. Saccadic inputs reach the abducens nucleus from ipsilateral excitatory burst neurons (EBNs) and contralateral inhibitory burst neurons (IBNs). Eye position information (the output of the neural integrator) reaches the abducens nucleus from neurons within the nucleus prepositus hypoglossi (NPH) and adjacent MVN. The medial rectus motoneurons in CN III also receive a command for vergence eye movements. Putative neurotransmitters for each pathway are shown: Ach, acetylcholine; asp, aspartate; glu, glutamate; gly, glycine. The anatomic sections on the right correspond to the level of the arrowheads on the schematic on the left. Abd. nucl., abducens nucleus; CN III, oculomotor nerve; CN IV, trochlear nerve; CN VI, abducens nerve; CN VII, facial nerve; CTT, central tegmental tract; ICP, inferior cerebellar peduncle; IVN, inferior vestibular nucleus; Inf. olivary nucl., inferior olivary nucleus; MVN, medial vestibular nucleus; MRF, medullary reticular formation; SVN, superior vestibular nucleus. Numbers indicate lesion sites that are discussed in the text. Adapted from Leigh, R. J. and Zee, D. S. (2006). The Neurology of Eye Movements, 4th edn. New York: Oxford University Press.

Besides mitochondrial disorders, the EOM may be affected by other genetic diseases such as hereditary disorders of myosin, or acquired disorders that present as a restrictive ophthalmopathy, such as thyroid disease. Thyroid ophthalmopathy, which has been attributed to accumulation of glycosaminoglycans in the orbit, often presents with vertical diplopia that is worse on wakening. Associated lid retraction and exophthalmia are common manifestations.

Effects of Disorders of the Neuromuscular Junction on Eye Movements Eye movements are especially susceptible to a disease affecting the neuromuscular junction (Figure 1, site 2), classical myasthenia gravis, which is due to an abnormal immune attack on the postsynaptic acetylcholine receptor. In half of all patients with myasthenia gravis, diplopia or ptosis is the presenting complaint and, in about 80%

Abnormal Eye Movements due to Disease of the Extraocular Muscles and Their Innervation

41

Superior colliculus Thalamus

PC

CG Cerebellum

TR ND

riMLF EBN

MRF IV

3

MLF

1 INC lBN

MB

CN IV 2

CN VII

III PPRF EBN

CN III

NPH

VI

NRTP CN VI Med RF IBN

Rostral

3 mm

Caudal

Figure 2 A sagittal section of the monkey brain stem showing the locations of premotor burst neurons: excitatory burst neurons for horizontal saccades lie in the paramedian pontine reticular formation (PPRF) and, for vertical and torsional saccades lie in the rostral interstitial nucleus of the medial longitudinal fasciculus (rostral iMLF). Burst neurons project to ocular motoneurons lying in the abducens nucleus (VI), the trochlear nucleus (IV) and the oculomotor nucleus (III). Omnipause neurons (indicated by an asterisk) lie in the midline raphe of the pons between the rootlets of the abducens nerve (CN VI) and gate the activity of burst neurons. CG, central gray; MB, mammillary body; MT, mammillothalamic tract; CN III, rootlets of the oculomotor nerve; CN IV, trochlear nerve; ND, nucleus of Darkschewitsch; NRTP, nucleus reticularis tegmenti pontis; PC, posterior commissure; NPH, nucleus prepositus hypoglossi; TR, tractus retroflexus. The arrow refers to the Horsley-Clarke plane of section. Numbers indicate lesion sites that are discussed in the text. Courtesy of Dr. Jean Bu¨ttner-Ennever).

of patients, movements of the eyes and lids are ultimately abnormal, with fluctuating weakness. Why are the EOMs so susceptible to diseases affecting the neuromuscular junction? One physiological reason arises from the demands made of eye movements to sustain the precise alignment of the eyes required for single binocular vision. It follows that fluctuating weakness due to myasthenia often causes ocular misalignment and diplopia. One morphological reason is that the postsynaptic junction of the EOM is poorly folded, thereby reducing the potential area for acetylcholine receptors. It follows that the EOM will be especially susceptible to loss of acetylcholine receptors. However, one subtype of EOM fibers, the MyHC EOMpositive myofibers (fast twitch/fatigable), which seems important for fast eye movements, does have substantial folding of its postjunctional membranes and, therefore, seems less susceptible to fatigue. Thus, it is interesting to note that patients with severe ocular myasthenia and little residual movement often retain the ability to make fast movements (quiver movements)-presumably due to preserved activity of their MyHC EOM-positive fast twitch myofibers.

Why is diplopia a common complaint in ocular myasthenia but rare in CPEO? The view that the eyes move conjugately in CPEO but not in myasthenia is not supported by measurements of eye movements. A more cogent reason is that the weakness in ocular myasthenia is highly variable (hence the characteristic symptom of fatigue), whereas in CPEO it evolves slowly and steadily. Thus, in CPEO the visual system has time to adapt to the loss of binocular correspondence, whereas in ocular myasthenia visual inputs are continually varying. This is not to state that adaptation of eye movements does not occur in myasthenia: the converse is the case, and is often evident by the occurrence of abnormally large eye movements immediately following pharmacological reversal of the neuromuscular failure by intravenous injection of the acetylcholine esterase inhibitor, edrophonium. Although ocular myasthenia is the most common disease to affect the EOM neuromuscular junction, other disorders that can impair eye movements include systemic botulism, neuromuscular blocking agents, and the Lambert-Eaton myasthenic syndrome (LEMS), which is due to the impaired release of acetylcholine secondary to autoimmune

42

Extraocular and Eyelid Muscles: Structure, Function, and Pathophysiology

Table 1

Functional classes of human eye movements

Class of eye movement Vestibular

Visual fixation Optokinetic Smooth pursuit

Nystagmus quick phases Saccades Vergence

Main function Holds images of the seen world steady on the retina during brief head rotations or translations Holds the image of a stationary object on the fovea by minimizing ocular drifts Holds images of the seen world steady on the retina during sustained head rotation Holds the image of a small moving target on the fovea; or holds the image of a small near target on the retina during linear self-motion; with optokinetic responses, aids gaze stabilization during sustained head rotation Reset the eyes during prolonged rotation and direct gaze toward the oncoming visual scene Bring images of objects of interest onto the fovea Moves the eyes in opposite directions so that images of a single object are placed or held simultaneously on the fovea of each eye

Adapted from Leigh, R. J. and Zee, D. S. (2006). The Neurology of Eye Movements, 4th edn. New York: Oxford University Press.

both receive increased innervation. One consequence of this adaptation is that movements of the weak eye may improve (unless paralysis is complete). A second consequence is that the strong eye, which also receives increased innervation, would, for example, make leftward saccades that overshoot the visual target. Such behavior is sustained for some time after adaptation even if the weak eye is covered and the strong eye views. This is a special example of plastic adaptation or motor learning, a property that depends heavily on the cerebellum, and which has been a subject of research interest for the past quartercentury. Because of the yoking mechanism by which eye movements are made conjugate in the brainstem circuitry (see the next section), there is a limitation to how much adaptive mechanisms can contribute to the recovery of the weakness due to nerve palsy. Nonetheless, such adaptive mechanisms undoubtedly contribute to the recovery from ocular motor nerve palsies. Paradoxically, in the case of trochlear nerve palsy, there is some recent evidence that such mechanisms may be maladaptive.

Effects of Disorders of the Brainstem Circuitry on Eye Movements Horizontal Movements

attack on presynaptic P/Q voltage-gated calcium channels. As opposed to myasthenia, in LEMS, repetitive saccades may change from hypometric (under-shooting) to hypermetric (over-shooting) as a consequence of the characteristic facilitation of muscle strength.

Effects of Disorders of the Oculomotor, Trochlear, and Abducens Nerves on Eye Movements Palsies of the nerves innervating EOMs are common in clinical practice and cause diplopia and selective patterns of weakness of the muscles they supply. Such paralytic strabismus (misalignment of the visual axes) is greatest when the affected patient attempts to look in the direction of the weak muscle. Thus, in the case of left abducens palsy (Figure 1, site 3), the patient cannot abduct (turn out) the affected eye to the left. However, such weakness and the attendant diplopia is a stimulus to adapt the neural signals that move the eyes. Such adaptive changes are evident if the strong eye is covered and the weak eye forced to view the world. A similar situation occurs naturally if the weak eye is also the visually dominant eye. In either case, the level of innervation is increased in motoneurons supplying muscles that induce corresponding movements in each eye. Thus, in our example, the left lateral rectus, which turns the left eye out, and the right medial rectus, which turns the right eye in, would

The brainstem machinery whereby the eyes are coordinated to move together (conjugately) in the horizontal plane is summarized in Figure 1. The abducens nucleus, which lies in the pons, may be regarded as the horizontal gaze center (Figure 1, site 4). Thus, the abducens nucleus receives inputs for each functional class of eye movements, including saccades, smooth pursuit, vestibular and optokinetic reflexes. It follows that each of these classes of conjugate eye movement (Table 1) may be independently affected by disease. The abducens nucleus contains two main groups of neurons: abducens motoneurons and abducens internuclear neurons. Axons of abducens motoneurons project in the sixth (abducens) cranial nerve to innervate the lateral rectus muscle. Axons of abducens internuclear neurons cross the midline and ascend in the contralateral medial longitudinal fasciculus (MLF, Figure 1, site 5) to contact medial rectus motoneurons in the oculomotor nucleus, which lies in the midbrain. Axons of medial rectus motoneurons project in the third (oculomotor) cranial nerve to innervate the medial rectus muscle. It follows that lesions affecting the abducens nucleus (Figure 1, site 4) will impair movements of both eyes to the side of the lesion (horizontal gaze palsy). It also follows that lesions of the MLF (Figure 1, site 5) will impair the ability of the ipsilateral eye to adduct; this is called internuclear ophthalmoplegia (INO), because the coordination of the abducens motoneurons and oculomotor medial rectus motoneurons is disrupted. Multiple sclerosis (MS) is the

Abnormal Eye Movements due to Disease of the Extraocular Muscles and Their Innervation

most common etiology for an INO in young persons, especially when it is present bilaterally. Patients with INO due to demyelination of the MLF in MS show slowing or absent movements of the adducting eye because the MLF can no longer conduct high-frequency signals between the abducens and the oculomotor nuclei. However, vergence movements may be preserved either in abducens nucleus lesions or in INO, since vergence commands project directly to the oculomotor nucleus (Figure 1). Horizontal saccades depend on premotor burst neurons, which lie in the paramedian pontine reticular formation (PPRF, Figure 1, site 6), and generate a high-frequency pulse of action potentials. Disorders affecting burst neurons of the PPRF selectively slow, or abolish, horizontal saccades. In contrast with abducens nucleus lesions, which cause complete horizontal gaze palsy, lesions of the PPRF usually spare ipsilateral smooth pursuit and vestibular eye movements. The vestibulo-ocular reflex (VOR) for horizontal head rotations depends on vestibular afferents from the lateral semicircular canals, which relay their signal to the contralateral abducens nucleus via the medial vestibular nucleus (MVN, Figure 1, site 7). Wernicke’s encephalopathy, a disorder due to thiamine deficiency that occurs in alcoholics, involves the vestibular nuclei and may impair the horizontal VOR. The nucleus prepositus hypoglossi (NPH , Figure 1, site 8), the adjacent MVN, and the cerebellum play an important role in holding the eyes in an eccentric position (e.g., far right gaze) against the elastic pull of the orbital tissues. This function depends on mathematical integration of premotor (visual, vestibular, saccadic) signals by the NPH/MVN–cerebellar network. Impaired function of this network (leaky integration) due, for example, to intoxication with alcohol, causes the eyes to drift back to the center, leading to gaze-evoked nystagmus. It has also been postulated that the ocular motor neural integrator network may also become unstable causing either increasing velocity drifts away from center position or quasisinusoidal eye oscillations (acquired pendular nystagmus). Vertical Movements The coordination of eye movements in the vertical plane depends heavily upon neural circuits in the midbrain. However, there is no single vertical gaze center similar to the abducens nucleus for horizontal gaze. The oculomotor and trochlear nuclei (Figure 2) house motoneurons that innervate EOMs that rotate the eyes mainly vertically (superior and inferior rectus muscles) or mainly torsionally (around the line of sight-the superior and inferior oblique muscles). These motoneurons receive their saccadic input from the rostral interstitial nucleus of the medial longitudinal fasciculus (riMLF), which lies in the prerubral fields of the rostral midbrain. Bilateral lesions involving the riMLF (Figure 2, site 1) cause slow

43

or absent vertical saccades, such as in progressive supranuclear palsy (PSP), a Parkinsonian disorder. Unilateral lesions of riMLF, such as in rostral brainstem strokes, cause loss of torsional rapid movements that rotate the top pole of the eye toward the side of the lesion. The signals for vertical vestibular and pursuit eye movements ascend from the medulla and pons to the midbrain in the MLF and other pathways. Thus, bilateral MLF lesions (Figure 2, site 2), which occur in MS, cause impaired vertical pursuit and vestibular responses (as well as bilateral adduction failure during horizontal saccades, as described in the previous section). The interstitial nucleus of Cajal plays an important role in holding vertical eccentric gaze steady (e.g., far upward gaze); lesions here (Figure 2, site 3) cause gaze-evoked nystagmus on upward or downward gaze, postulated to be due to a leaky ocular motor integrator. The superior colliculus is a midbrain tectal structure that receives inputs from the cortical eye fields, and is important for triggering both horizontal and vertical saccades. Functional imaging studies in humans have demonstrated activation of the superior colliculus during generation of short-latency (express) saccades. Neural circuits important for the generation of vergence eye movements are also located in the pretectum and midbrain, but pontine nuclei and their projections to the cerebellum also contribute. Thus, disturbances of vergence eye movements are encountered with lesions, such as strokes, throughout the brainstem. However, abducens nucleus lesions and INO (Figure 1, sites 4 and 5) usually spare vergence movements.

Congenital Misalignment of the Eyes (Infantile Strabismus) and Attendant Nystagmus Ocular misalignment from infancy may be due to disorders of the orbital tissues, the innervation of EOM, or as a consequence of failure to develop binocular vision. The failure to develop binocular vision usually presents as the fusional maldevelopment nystagmus syndrome (FMNS), which includes amblyopia of one eye, strabismus (commonly esotropia and dissociated vertical deviation, with upward deviation of the covered eye) and latent nystagmus. Latent nystagmus is a jerk nystagmus comprising slow drifts of the eyes off target and a rapid resetting component that is absent when both eyes are viewing but appears when one eye is covered. The quick components of latent nystagmus beat away from the covered eye, and the nystagmus reverses direction upon covering of either eye. In most patients, the nystagmus is present (but low amplitude) when both eyes are uncovered, and is termed manifest latent nystagmus. Thus, although binocular viewing is possible, affected individuals almost invariably choose to fix with one eye and suppress the image from the other. Latent nystagmus can be

44

Extraocular and Eyelid Muscles: Structure, Function, and Pathophysiology

induced experimentally in monkeys, either by depriving them of binocular vision early in life, or by surgically creating strabismus. In monkeys with latent nystagmus, the brainstem nucleus of the optic tract (NOT) shows abnormal electrophysiological properties. In normal monkeys, NOT neurons respond to visual stimuli presented to either eye. However, in monkeys with latent nystagmus, NOT neurons are driven mainly by the contralateral eye. Furthermore, inactivation of NOT with muscimol abolishes latent nystagmus in monkeys who have been deprived of binocular vision. Since the NOT projects to vestibular circuits concerned with gaze control during head rotations, one current view of the pathogenesis of latent nystagmus is that it represents the consequences of imbalance of visual inputs to the vestibular system, as if the subject was being rotated toward the side of the viewing eye.

Effect of Visual System Disorders on Eye Movements Patients with a broad range of retinal disorders causing blindness is often a familial disorder. Both show continuous jerk nystagmus, with components in all three planes, which changes in direction over the course of seconds or minutes. The drifting null point, the eye position at which nystagmus changes direction, probably reflects an inability to calibrate the ocular motor system. Animals raised in a strobe illuminated environment, which deprives them of retinal image motion while still providing position cues, also develop spontaneous ocular oscillations. Gene therapy used to restore vision to dogs blind due to an inherited retinal disease resulted in a decrease in their associated nystagmus. Nystagmus is also a feature of albinism, which is associated with abnormal development of visual pathways and optic nerve hypoplasia.

Infantile Forms of Nystagmus in Individual with Normal Visual Systems Infantile nystagmus syndrome (INS), or congenital nystagmus, may be present at birth but usually develops during infancy. The nystagmus is almost always conjugate and horizontal, even on up or down gaze, with a small torsional component. It is usually accentuated by the attempt to fix upon an object and by attention or anxiety. Up to 30% of patients with INS have strabismus but, even in individuals lacking strabismus, stereovision is usually degraded, partly due to retinal image motion. Head turns are common in INS and are used to bring the eye in the orbit close to the null point or zone, at which nystagmus is minimized. Some patients with INS also show head oscillations; such head movements could not act as an adaptive strategy to improve vision unless the VOR was negated.

It seems possible that the head tremor and ocular oscillations in INS represent the output of a common neural mechanism. Measurements of nystagmus in INS demonstrate typical waveforms with increasing slow-phase velocity and the superimposed presence, during each cycle of oscillations (usually after a quick phase), of a brief period when the eye is still and is pointed at the object of regard. Such foveation periods are probably one reason why many individuals with INS have near-normal vision and why most do not complain of oscillopsia (illusory motion of the seen world), in spite of otherwise nearly continuous movement of their eyes. INS, either with or without associated visual system abnormalities, is often a familial disorder. Both autosomal dominant and sex-linked recessive forms of inheritance have been reported. Although several hypotheses for the pathogenesis of INS have been offered, no animal models exist. At present, it seems possible that genetic studies will identify the underlying molecular mechanisms and point researchers to the neural disturbance causing INS.

Conclusions Recent progress in understanding disorders of the EOMs and their innervation from the viewpoint of molecular biology and genetics is approaching the point where it can be combined with behavioral and electrophysiological studies. For example, recent evidence indicates that each functional class of eye movements (Table 1) is served by a separate population of ocular motoneurons that receive specific premotor inputs. It follows that each functional class of eye movements may depend on distinct molecular mechanisms or morphological characteristics, from premotor neurons to EOM. Human diseases provide many opportunities to study behavioral effects of a disease when the disease process affects a specific site-such as the acetylcholine receptor in myasthenia gravis. In this way, insights from basic science have a growing impact on clinical ophthalmology and neurology, and vice versa. See also: Extraocular Muscles: Extraocular Muscle Anatomy; Extraocular Muscles: Extraocular Muscle Metabolism; Extraocular Muscles: Functional Assessment in the Clinic.

Further Reading Kennard, C. and Leigh, R. J. (2008). Using eye movements as an experimental probe of brain function. A symposium in honor of Jean Bu¨ttner-Ennever. Progress in Brain Research 171: 1–603. Leigh, R. J. and Zee, D. S. (2006). The Neurology of Eye Movements, 4th edn. New York: Oxford University Press. Leigh, R. J. and Devereaux, M. W. (2008). Advances in Understanding Mechanisms and Treatment of Infantile Forms of Nystagmus. New York: Oxford University Press.

Extraocular Muscles: Functional Assessment in the Clinic S P Christiansen, Boston University School of Medicine, Boston, MA, USA L K McLoon, University of Minnesota, Minneapolis, MN, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Binocular vision – The simultaneous perception by both eyes of two slightly disparate images of the same target on corresponding retinal elements resulting in a single three-dimensional image. Cover test – The use of an ocular occluder over one eye or alternately occluding the eyes, either alone or in conjunction with prisms, to detect the presence of an ocular deviation and to measure its magnitude. Diplopia – Double vision caused by a misalignment of the eyes resulting from the same image stimulating noncorresponding retinal elements in the two eyes. Esotropia – A form of strabismus where there is a nasal-ward deviation of the nonfixing eye. Exotropia – A form of strabismus where there is a temporal deviation of the nonfixing eye. Phoria – This is a latent misalignment of the eyes kept under control by fusional mechanisms and is contrasted with a tropia which is a manifest constant or intermittent deviation of the eyes. A phoria can be seen only when fixation is interrupted, as during a cover test. Recession surgery – In recession surgery, the overacting extraocular muscle is surgically removed from the sclera and resutured in a more posterior location on the globe. The goal is to decrease the rotational effect of muscle contraction. Resection surgery – In resection surgery, the underacting extraocular muscle is surgically removed from the sclera, a portion of the insertional end is removed, and the remaining, now shorter, muscle is resutured to its original insertional site on the globe. The goal is to increase the rotational effect of muscle contraction. Strabismus – A latent or manifest misalignment of the eyes. Tropia – A manifest misalignment of the visual axes of both eyes.

Normal Eye Movements There are six extraocular muscles responsible for eye movement within each orbit. These muscles are innervated by

cranial nerves III (superior rectus, medial rectus, inferior rectus, and inferior oblique), IV (superior oblique), and VI (lateral rectus). The medial rectus muscles are primarily responsible for adduction, pulling the eyes toward the nose. The lateral rectus muscles are responsible for abduction, pulling the eyes temporally. These horizontal movements are the most straightforward of the six extraocular muscles. The remaining four are cyclovertical muscles and have more complex function related to the fact that forwardoriented eyes are housed in laterally directed orbits. This means that the midline of these muscles does not consistently lie over the center of rotation of the globe in any position of gaze. If examined from the superior view, the bony medial orbital walls are parallel to each other and are in the sagittal plane. The lateral walls, however, are at a 45 angle from the plane of the medial walls. Since all but the inferior oblique muscles take their origin from the orbital apex, contraction of the superior and inferior rectus and superior oblique muscles will have a rotational or torsional component. The same is true for the inferior oblique, which originates from the anterior and inferior nasal orbital wall and courses posteriorly and laterally to insert onto the globe inferior to the belly of the inferior rectus muscle. The direction of contraction of the inferior oblique muscle thus also results in both a torsional and vertical movement of the eye. The function of the individual extraocular muscles has been more extensively covered elsewhere in this encyclopedia. To summarize, however, the superior rectus muscle and the inferior oblique muscles are the principal elevators of the eye while the inferior rectus muscle and superior oblique muscle are the principal depressors of the eye. The incyclotorters of the eye are the superior oblique and the superior rectus while the excyclotorters are the inferior oblique and the inferior rectus muscles. Each of the cyclovertical muscles also has minor horizontal function. It is important to recognize that the vertical or torsional component of each of the cyclovertical muscles changes depending on whether the eye is held in adduction or abduction. There are two basic kinds of eye movements: saccade and pursuit. Saccades are rapid and subserve fast changes in fixation. They are generated by a pulse-step pattern of innervation from the brainstem. An estimate of saccadic velocity can be gained by clinical observation alone, often with the use of an optokinetic nystagmus (OKN) drum or flag that the examiner uses to drive repeated changes in fixation, first in one direction and then in another.

45

46

Extraocular and Eyelid Muscles: Structure, Function, and Pathophysiology

RSR

LIO

RLR

LMR

RIR

LSO

RSR

RIR

LSR

LIR

RIO

LSR

RMR

LLR

RSO

LIR

(b) Figure 1 (a) Composite photographs showing a subject displaying the nine cardinal positions of gaze and (b) Chart showing the principal yoked muscles from the right (pink) and left (green) eyes responsible for movement of the eyes into the nine cardinal positions of gaze.

A decrease in saccadic velocity can be seen in patients with extraocular muscle palsy. Pursuit movements are slower, and are driven by smooth following movements in response to a target moved slowly before the patient. There should be no lag or saccadic interlude during pursuit movements. Gaze refers to movement of both eyes together. Gaze is conjugate if both eyes move the same amount, at the same speed, and in the same direction. Thus, in right gaze, both eyes look to the right and reach the intended target at the same time. Gaze is disconjugate if both eyes move in opposite directions or there is substantial failure of one eye to reach the target. Therefore, convergence and divergence movements are disconjugate. Ductions are movements of one eye examined under monocular viewing conditions. For example, we refer to adduction as a nasalward movement of an eye when the other eye is covered, eliminating any binocular adjustment in eye position. In contrast, versions are movements of both eyes examined under binocular viewing conditions. Dextroversion is movement of both eyes to the right; levoversion is movement of the eyes to the left. Version testing is helpful for assessing over- or under-action of a muscle compared to its yoked muscle in the other eye. As conjugate gaze position shifts, yoked muscles contract in response to gaze-evoked increases in innervational frequency (Figures 1(a) and 1(b)). However, innervation to the antagonist muscles is inhibited. Therefore, for gaze right, the right lateral and the left medial muscles

contract, but the right medial and left lateral rectus muscles are innervationally inactive. Because the eyes must move in an accurate, balanced, and coordinated fashion when gaze position is changed, complex central nervous system control mechanisms are required. Input from the frontal and parietal cortex and from the cerebellum is routed through a neural integrator in the brainstem whose function and control is still being investigated. Feedback from the afferent visual system and from proprioceptive input from the extraocular muscles modulates innervational tone to all the extraocular muscles and is important for long-term calibration of eye movements. Ocular Motility Assessment in the Clinic When an individual is looking straight ahead, in what is called primary gaze, a light falling on the eyes from a distance will be perceived by an examiner as a corneal reflex or reflection in approximately the same position in the pupil of each eye (Figure 1(a), central photograph). If the eyes are not aligned, there will be a nasal, temporal, or vertical offset of the corneal reflex in one eye compared to the other eye. The amount of offset of the reflex can be estimated subjectively by the examiner or measured with prisms. Care must be taken in the patient who has a very small degree of misalignment as these tests are not sensitive enough to detect it. To confirm normal alignment of the eyes, a cover test is performed. If there is no refixation movement of either eye

Extraocular Muscles: Functional Assessment in the Clinic

when the other eye is covered, the alignment of the eyes is considered orthotropic. This does not imply, however, that the eyes are normally aligned since latent misalignment of the eyes may be controlled by fusional mechanisms that can be remarkably robust. To determine if an individual’s eyes are normally aligned, an alternate cover test is performed. Here, the eyes are alternately occluded. If no refixation movement of the eye under cover occurs when the cover is moved to the other eye, then the individual is considered to have normal eye alignment and is deemed orthophoric. The alternate cover test can be used in all the cardinal positions of gaze to determine if changes in gaze position result in any misalignment. If refixation movement is detected with alternate cover testing, then prisms may be used with the cover test to measure the magnitude of the misalignment. A convergent misalignment of the eyes is called esotropia; divergent misalignment is called exotropia; and vertical misalignment is called hypertropia. In addition to alignment testing, an examiner performs certain sensory tests to determine the quality of binocular function. These tests detect the presence of normal or abnormal retinal correspondence, the presence of suppression or diplopia, even if not subjectively present, the quality of stereoacuity if present, and the presence of torsion. The nature of this text does not allow an indepth review of such testing, but it is useful to document, as binocular function is an important aspect of normal visual experience, and is critical to the assessment of outcomes after treatment for strabismus. Children with strabismus typically do not experience diplopia unless the onset of misalignment has been rapid as might be seen in an acute CN VI palsy due to increased intracranial pressure or due to a brain tumor. More often, children with strabismus suppress the nonfixing eye, an adaptive mechanism that obviates diplopia, but places the child at risk for the development of amblyopia, a nonorganic loss of vision that may be permanent if not treated during the child’s period of visual plasticity. Therefore, during the examination of children being evaluated for strabismus, careful assessment of visual acuity is essential. During a motility examination, all patients should be carefully observed for the presence of nystagmus (rhythmic to and-fro movements of the eyes), muscle weakness (paresis), muscle restriction, or binocular-gaze deficits along with abnormal head posturing, head nodding, or other adaptive mechanisms that may arise as a result of an abnormal ocular motor condition. Clinical Treatment for Primary Eye Motility Disorders Once strabismus has been diagnosed, then a decision must be made regarding how to treat it. In children, glasses are an important consideration. Children with esotropia who

47

are hyperopic (far-sighted) will often have an accommodative component to their strabismus. Control of accommodation by correcting the hyperopia may eliminate or reduce the angle of esotropia. Glasses are occasionally used in children with exotropia as well. Certainly, if a child has a significant refractive error, glasses will be required to optimize acuity. A child with strabismus and amblyopia will likely be given treatment for the amblyopia which may include occlusion therapy or optical penalization with atropine drops. Treatment of amblyopia is usually recommended before more definitive treatment of the strabismus is undertaken. Other nonsurgical treatments for strabismus include prisms, if the angle of misalignment is small or, in less common scenarios, exercises such as convergence training. When a decision has been made to proceed with surgery, there are several available options. Typical incisional surgery addresses the misalignment mechanically. If a muscle is overacting, then it is weakened. This is usually done by recession surgery in which the muscle insertion on the sclera is transected and then attached with sutures more posteriorly on the globe. This decreases the mechanical advantage of the muscle by reducing the arc of contact of the muscle on the globe. By contrast, if a muscle is underacting relative to its antagonist, then the muscle insertion is resected. In this surgery, a portion of the insertion is removed and the shortened muscle is reattached to the original muscle insertion on the sclera. This surgery works by shortening the tether length of the muscle and increasing its mechanical advantage relative to its antagonist. The amount of recession or resection is titrated to the angle of the strabismus with larger amounts of surgery for larger angles of misalignment. There are numerous variations on this theme, and the literature is full of unique means of weakening or strengthening muscles. In certain situations, such as when a muscle is significantly paralyzed, healthy muscle insertions may be transposed to approximate the insertion of the paralyzed muscle to assist its function. Sometimes, this is done in conjunction with botulinum toxin injection into the antagonist muscle to weaken it and to improve the rotation of the globe in the direction of the paralyzed muscle. Botulinum toxin A has heralded the beginning of a new era in strabismus surgery. First introduced into clinical use in the early 1980s by Dr. Alan Scott at SmithKettlewell in San Francisco, botulinum toxin weakens muscle by chemical denervation of the muscle. Release of the neurotransmitter, acetylcholine, into the synaptic cleft of neuromuscular junctions (NMJs) of treated skeletal muscle (including extraocular muscle) is blocked. This temporary paralysis of the synapses of the NMJs results in a spread of NMJ sites across the surface of the muscle. Treatment effect is maximal for approximately 6 weeks and then begins to diminish as the terminal nerves regrow and form new NMJs. Ultimately, there is a return of

48

Extraocular and Eyelid Muscles: Structure, Function, and Pathophysiology

function at the sites of the original NMJs and retraction of the sprouted nerves. Treatment effect is essentially void by 3 months postinjection. Botulinum injection has been used by some clinicians for the treatment of infantile and childhood forms of strabismus, especially esotropia, but its use has not been widely adopted because of the frequent need for reinjections, especially for larger angles of misalignment. However, pharmacologic treatment of strabismus is attractive because of decreased operative times required for injection compared with typical incisional surgery, decreased scarring, and preserved biomechanical relationships of the muscle, and orbital soft tissues. In the recent literature, there have been a number of reports of the use of new candidate drugs for both weakening and strengthening extraocular muscle in experimental animals. Although much research is still needed, the era of drug treatment for strabismus is dawning, and heralds the possibility of both reducing the short-term risks of strabismus surgery and improving the long-term outcomes of our interventions.

Acknowledgments This work was supported by EY15313 and EY11375 from the National Eye Institute, the Minnesota Medical

Foundation, the Minnesota Lions and Lionesses, Research to Prevent Blindness (RPB) Lew Wasserman Mid-Career Development Award (LKM), and an unrestricted grant to the Department of Ophthalmology from RPB. See also: Abnormal Eye Movements due to Disease of the Extraocular Muscles and Their Innervation; Extraocular Muscles: Extraocular Muscle Anatomy; Extraocular Muscles: Extraocular Muscle Metabolism.

Further Reading Anderson, B., Christiansen, S. P., and McLoon, L. K. (2008). Myogenic growth factors can decrease extraocular muscle force generation: A potential biological approach to the treatment of strabismus. Investigative Ophthalmology and Visual Science 49: 221–229. Leigh, R. J. and Zee, D. S. (1999). The Neurology of Eye Movements, 3rd edn. New York: Oxford University Press. McLoon, L. K., Anderson, B., and Christiansen, S. P. (2006). Sustained release of insulin growth factor-I results in stronger extraocular muscle. Journal of the American Association of Pediatric Ophthalmology and Strabismus 10: 424–429. McLoon, L. K. and Christiansen, S. P. (2005). Pharmacological approaches for the treatment of strabismus. Drugs of the Future 30: 319–327. Wong, A. (2008). Eye Movement Disorders. New York: Oxford University Press.

II. STRUCTURE AND FUNCTION OF THE TEAR FILM, OCULAR ADNEXA, CORNEA AND CONJUNCTIVA IN HEALTH AND PATHOGENESIS IN DISEASE

This page intentionally left blank

Tear Film J P Craig, University of Auckland, Auckland, New Zealand A Tomlinson and L McCann, Glasgow Caledonian University, Glasgow, UK ã 2010 Elsevier Ltd. All rights reserved.

Glossary Fluorescein sodium – A topical agent used extensively as a diagnostic tool in ophthalmology to enhance tear film visibility or to highlight epithelial cell loss. The molecule is highly fluorescent, with excitation and emission occurring at 494 and 521 nm, respectively. Interference bandpass filters are commonly combined with the observation systems used in ophthalmology to optimize visualization of the fluorescence alone. Lacrimal gland – The lacrimal gland is a compound tubuloalveolar gland, similar to the salivary gland, situated superotemporally in the orbit, which secretes aqueous tear fluid. Lacrimal sac – The lacrimal sac forms part of the tear drainage system, collecting tear fluid from the ocular surface via the puncta and canaliculi. Blinking controls the pumping action of the lacrimal sac into the nasolacrimal duct for drainage into the nasal cavity. Meibomian gland – Vertically oriented tubulo-acinar glands, embedded in the upper and lower tarsal plates, which release meibum (lipid). Videokeratoscopy – A computerized, dynamic technique, based on the principle of keratoscopy, used traditionally to assess the shape of the anterior surface of the cornea (corneal topography) from the reflection of a series of projected concentric rings.

The tear film is a thin film of fluid, which covers the exposed ocular surface. Essential for the health and normal function of the eye and visual system, any abnormality in quantity or quality of the tear film can lead to signs and symptoms of dry eye disease and ultimately to a loss of vision.

The Role of the Tear Film The tear film has a number of important functions, the first of which, as the most anterior element of the visual system, is maintenance of high-quality vision. Alterations in the stability of the tear film due to abnormal tear evaporation, production, and/or drainage can cause optical aberrations and adversely affect retinal image quality.

Secondly, the tear film plays an important role in ocular surface defence. Environmental challenges such as extremes of temperature or humidity, and exposure to irritants such as pollutants and allergens, can have a detrimental effect on the tear film. The tear film must be sufficiently robust to be able to withstand these challenges and be capable of responding rapidly with reflex tearing to help flush out irritants when required. External and adnexal infectious agents pose an additional risk to the exposed ocular surface. Antimicrobial components of the tear film, which include lysozyme, lactoferrin, and immunoglobulin A, help to protect the ocular surface from microbial infection. Lubrication is another important tear film function. The non-Newtonian rheological properties of the tear film mucins enable the tear film to lubricate the corneal and mucosal surfaces. The normal blinking mechanism draws the tear film across the ocular surface, enhancing comfort and cushioning the ocular surfaces from the shearing forces present during the blink, while the mucins that trap and coat foreign particles in the tear film for removal at the caruncle, confer further epithelial surface protection. Finally, the tear film plays a vital nutritive role in the transport of substances necessary for corneal metabolism and regeneration. Uniquely avascular for transparency, the cornea requires a nonvascular route for the supply of oxygen, electrolytes, growth factors, and nutrients to, and for the removal of metabolic by-products such as carbon dioxide from the ocular surface. While glucose diffuses primarily from, the aqueous humor, oxygen must be transported to the tissue through the tear film, either from the air in the open eye state or via the palpebral conjunctival vessels in the closed eye state.

Structure and Thickness of the Tear Film Initial reports described the tear film as trilaminar in structure, consisting of a thin superficial lipid layer, an intermediate aqueous layer, and an underlying mucous layer. Each of these layers has the potential to be affected by different conditions resulting in qualitative and quantitative changes. Almost half a century later, it was proposed that interfaces existed between the layers, giving rise to a six-layer model, with an oily layer, a polar lipid monolayer, an absorbed mucoid layer, an aqueous layer, and a mucoid layer on a glycocalyx base. The carbohydrate-rich

51

52

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Lipid component

Aqueous component

Mucous component Corneal epithelium

MUC5AC

SP-A

MUC5B

MUC7

MUC1

SP-D

MUC4

SP-B

MUC16

SP-C

Figure 1 Diagrammatic representation of our current understanding of tear film structure. The tear film comprises a thin superficial lipid layer, and an aqueous-mucin continuum increasing in mucin concentration toward the glycocalyx, adjacent to the ocular surface epithelium. Adapted from Bra¨uer, L. and Paulsen, F. P. (2008). Tear film and ocular surface surfactants. Journal of Epithelial Biology and Pharmacology 1: 62–67, with permission.

glycocalyx, produced by the surface cells of the corneal epithelium and subsurface vesicles of the conjunctival epithelium, is believed to attach the tear film to the surface of the epithelial cells. The most recent studies do not differentiate this number of distinct layers, but instead suggest the existence of an aqueous-mucin continuum that contains a decreasing concentration of dissolved mucus toward the superficial lipid layer, and is anchored to the epithelium by glycocalyx (Figure 1). The thickness of the precorneal tear film has proven to be a subject of great debate. Early estimates placed the thickness of the tear film in the region of between 4 and 8 mm. Later, on the basis of noninvasive techniques such as interferometry, it was proposed that due to a previously underestimated contribution from the mucous layer, the tear film thickness was closer to 40 mm in thickness. However, the most recent findings using techniques such as tomography and reflectance spectra propose values closer to the original measurements, suggesting that the tear film thickness is approximately 3 mm.

The Lipid Layer The superficial lipid layer of the tear film forms the initial barrier between the ocular surface and the environment. This thin, oily layer approximates 100 nm in thickness, although values ranging between 10 and 600 nm have

been reported. It is derived primarily from the meibomian glands, with additional lipid secreted by the eyelid glands of Moll and Zeiss. The lipids are excreted as meibum onto the ocular surface through the gland orifices located at the mucocutaneous junction of the lid margins. Between blinks, the lipid layer forms in two distinct phases. An inner, thin, polar layer spreads as a monolayer across the aqueous in the initial phase after the blink, then a thicker, outer, nonpolar layer follows, creating a final lipid structure with multiple layers. The lipid layer must be spread evenly by the blink to form a continuous layer without excessively thin or thick patches in order to inhibit evaporation and to prevent accelerated tear breakup from mucin contamination, respectively. Table 1 describes the proportions of the major lipid components of meibum. The polar layer consists of phospholipids, free fatty acids, and cerebrosides, while the less surface-active, nonpolar layer comprises mainly wax esters and sterol esters. The lipid layer confers a number of important protective functions including the formation of a hydrophobic barrier to prevent tear overflow onto the lids and to provide a water-tight seal during overnight lid closure, and the prevention of tear film contamination by skin lipids. However, arguably one of the most critical roles of the superficial lipid layer is to retard evaporation from the ocular surface. The polar lipids of the ocular tear film in the normal eye are capable of reducing its rate of evaporation by about 80–90%.

Tear Film

Tear Evaporation Numerous investigators have measured evaporation of fluid from the tear film, since it was established that the lipid layer retarded evaporation in a rabbit model, in 1961. Later work, also in a rabbit model, passed dry air over a cornea enclosed within a chamber. From the weight of water collected, the evaporative rate was measured as 10.1 10–7 g cm–2 s–1, and a fourfold increase in evaporation was found to occur with the removal of the rabbit tear film lipid layer. A similar increase in human tear film Table 1

Major lipid components of meibum

Component Synthesised lipids Wax esters Sterol esters Triglycerides Diglycerides Monoglycerides Fatty alchohol Hydrocarbons Membrane-derived lipids Cerebrosides Ceramides Phospholipids Degeneration products Free fatty acids

Percentage (%) 44 33 5 2 Trace Trace 2 4 Trace 8 2

Adapted from McCulley, J. P. and Shine, W. E. (2003). Meibomian gland function and the tear lipid layer. Ocular Surface 1(3): 97–106, with permission.

53

evaporation has since been confirmed in patients with incomplete or absent lipid layers (Figure 2). The use of different techniques for measurement of tear film evaporation makes comparison of evaporation rates in different studies difficult because the absolute values recorded are technique-dependent. However, a pattern to the observations reported in the literature does exist, making evaporation rate a useful measurement in the differential diagnosis of dry eye. In most cases, significant increases from normal tear film evaporation are seen in patients with aqueous deficient dry eye (ADDE), evaporative dry eye (EDE), and meibomian gland dysfunction (MGD). The evaporation in normal eyes averages 13.57 6.52 10–7 g cm–2 s–1, while in ADDE the values average 17.91 10.49 10–7 g cm–2 s–1, and in EDE, 25.34 13.8 10–7 g cm–2 s–1.

The Aqueous Layer The aqueous component of the tear film is a watery phase, bordering the lipid layer and comprising most of the tear film thickness. It is produced principally by the main lacrimal gland and accessory lacrimal glands of Krause and Wolfring although additional water and electrolytes are secreted by the epithelial cells of the ocular surface. The typical or basal level of tear flow present is believed to originate mainly from the accessory glands while the reflex tears, produced in response to mechanical, noxious, or emotional stimuli, arise from the main lacrimal gland.

Relative humidity sensor

Temperature sensor

Water vapour from ocular surface

Figure 2 Tear film evaporation rate measured by a modified ServoMed EP-3 Evaporimeter (Kinna, Sweden). This technique involves the measurement of the vapor pressure gradient from recordings of relative humidity and temperature at two points a known distance above the ocular surface. Reprinted from The Ocular Surface (www.theocularsurface.com), with permission.

54

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

During sleep, tear production is minimal but in the normal eye, in the open eye state, sensory stimulation of the exposed ocular surface induces tear production at a rate that varies according to the demands of the external environment. The secretion of electrolytes, protein, and water onto the ocular surface serves to nourish and protect the epithelia and convey messages between the structures bathed in aqueous. Corneal innervation is denser than that of any other part of the body, resulting in extreme pain if the corneal epithelium is damaged. Sensory nerve supply to the ocular surface arises from the trigeminal nerve. Stimulation of these nerve endings causes the release of neuropeptides such as substance P and calcitonin gene-related peptide (CGRP), which, through initiation of the inflammatory cascade, is believed to be an important step in the pathogenesis of many cases of dry eye. The lacrimal and meibomian glands are innervated by parasympathetic efferent nerve fibers (muscarinic and vaso-intestinal peptide (VIP)-ergic fibers) and to some extent they, and the blood vessels supplying them, are sympathetically innervated through tyrosine hydroxylase (TH) and neuropeptide Y fibers. Parasympathetic efferent nerve terminals surrounding the goblet cells suggest that conjunctival secretions are also under neurogenic control. The aqueous phase has a number of important responsibilities. These include creating a nurturing environment for the epithelial cells of the ocular surface, carrying essential nutrients and oxygen to the cornea, allowing cell movement over the ocular surface, and washing away epithelial debris, toxic elements, and foreign bodies. The major electrolytes present in the tear film are sodium, potassium, bicarbonate, and chloride, with magnesium, calcium, nitrate phosphate, and sulfate present in smaller quantities. The electrolytes dictate the osmolarity of tears, besides acting as a buffer to maintain pH and playing a role in maintaining epithelial integrity. An increase in the electrolyte concentration, described as hyperosmolarity, can cause damage to the ocular surface. The tear film protein concentration is approximately 10% that of plasma. The proportion of lacrimal gland versus serum-derived proteins and enzymes varies with tear flow rate, epithelial surface stimulation, blinking, and ocular surface disease. The tear proteins are involved in defense of the ocular surface and the maintenance of tear film stability. Electrophoresis has confirmed the presence of approximately 80 different components of human tear proteins. Around 30 proteins have been identified, half of which are enzymes. The principal tear proteins are lysozyme, lactoferrin, albumin, tear-specific pre-albumin, and globulins. Table 2 shows typical concentrations of the most significant tear proteins. The tear film also contains antioxidants such as vitamin C and tyrosine, which scavenge free radicals from within the tear film, while the

Table 2

Average concentration of the principal tear proteins

Protein component Total protein Lysozyme Albumin Tear specific prealbumin Lactoferrin Immunoglobulins (IgA, IgG, IgM, and IgE)

Average concentration (mg ml 1) 7.51 2.36 1.30 1.23 1.84 0.43

Adapted from Sariri R. and Ghafoori, H. (2008). Tear proteins in health, disease, and contact lens wear. Biochemistry (Moscow) 73(4): 381–392, with permission.

abundance of growth factors facilitates constant epithelial regeneration and promotes wound healing. Alterations in tear composition or inflammatory changes within the conjunctival vascular endothelia can act as the stimulus to ocular surface inflammation in which both cellular and soluble mediators play a significant role. The numbers of T lymphocytes and the relative proportions of activated T cells are increased in dry eye. The ocular surface epithelial cells are directly involved in such ocular surface inflammation with the release of a number of pro-inflammatory cytokines such as interleukin (IL)-1a, IL-1b, IL6, IL8, transforming growth factor beta 1 (TGF-b1) and tumor necrosis factor alpha (TNFa), and increased expression of immune activation molecules such as CD54 and HLA-DR. Increased proteolytic enzyme levels and activity have been observed in dry eye with, in particular, high levels of matrix metalloproteinase 9 (MMP9), which are not present on the normal ocular surface. The inflammatory markers described as precipitating dry eye are also recognized to perpetuate ocular surface inflammation, triggering an escalating cycle of ocular irritation, inflammation, apoptosis, and tear film dysfunction and instability, epithelial cell disease, and disruption of corneal epithelial barrier function.

Tear Production Traditional methods of measuring tear production rates are based on absorption of tears by Schirmer strips or cotton threads; however, both tests have been found to be poor quantifiers of tear production; the Schirmer test is marred by low specificity and sensitivity and the exact parameter measured with the cotton thread test has been questioned. As a result, a number of tests have been devised to measure the rate of disappearance of a dye marker placed in the tear film, as new tears are produced and the waste eliminated. In most studies in recent years, the rate of disappearance of instilled sodium fluorescein dye has been used to determine tear turnover (TTR) by the technique of fluorophotometry (Figure 3).

Tear Film

4.2

55

C0

FCE (ng ml−1)

4.0 3.8 3.6 3.4 3.2 3.0

0

2

4

6

8

10

12

14

16

Time (min) Figure 3 Commercial fluorophotometer (Fluorotron Master, Coherent Radiation Inc, CA, USA) shown with a typical trace of ocular surface fluorescence decay following instillation of fluorescein sodium into the eye. A biphasic curve of fluorescence is observed with initial rapid decay (due to reflex tearing) followed by a more gradual decay (due to basal tear turnover). Adapted from The Ocular Surface (www.theocularsurface.com), with permission.

The values reported for tear turnover (%min–1) and tear flow (ml min–1) in the major studies in the literature for normal and dry eye subjects of studies using the commercial fluorophotometer have recently been collated. The data reported for normals in the majority of studies ranges from 10% to 20% min–1, which equates to an average basal tear flow rate of 1.03 0.39 ml min–1 (16.19 5.10% min–1). For dry eye, in all its forms, it averages 0.58 0.28 ml min–1 (9.36 5.68% min–1) and, within the dry eye subtypes, averages 0.40 0.10 ml min–1 (7.71 1.02% min–1) and 0.71 0.25 ml min–1 (11.95 4.25% min–1) for ADDE and for EDE, respectively. These are the rates of tear production under nonstimulated conditions in normal and dry eyes. However, the eye is capable of producing copious reflex tears under provocative conditions, providing the lacrimal gland has the ability to function at the required capacity. Reflex rates have been quoted as approximately 100-fold those under basal conditions.

The Mucin Layer The innermost, mucin layer of the tear film lies adjacent to the hydrophobic epithelial cells of the ocular surface. The layer consists of soluble, gel-forming mucins, which are capable of retaining large quantities of water, and corneal and conjunctival epithelial mucins (principally MUC1, 2, 4, and 16), which form the glycocalyx. The glycocalyx functions, through the membrane-spanning domain of MUC1, to anchor the soluble mucin layer to the plasma membrane of the corneal and conjunctival epithelial cells, while the soluble mucins interact with these transmembrane mucins and with the overlying aqueous layer, to form a water-retaining gel. The most significant soluble mucin for the ocular surface is MUC5AC, secreted by the goblet cells of the conjunctiva. The high-molecular-weight glycoproteins, with additional proteins, electrolytes, and cellular material that

contribute to the mucous layer, enable fulfilment of several important functions in the maintenance of a healthy ocular surface. In addition to providing a hydrophilic surface upon which to support a stable aqueous layer, the mucous layer offers protection against the shear force of blinking and environmental insult, and facilitates maintenance of a smooth ocular surface for optical clarity. The constituents are also believed to protect the ocular surface by inhibiting inflammatory cell adhesion.

Tear Distribution and Stability The distribution of tear fluid on the ocular surface is highly dependent on the blink. Lid closure during a blink progresses from the temporal to the nasal side of the eye spreading tears across the ocular surface and facilitating tear drainage through the lacrimal puncta. The inter-blink period in normal individuals averages 4.0 2.0 s and is significantly decreased in patients with dry eye (to 1.5 0.9 s); a high blink rate in dry eye patients maximizes the tear supply to the ocular surface. In detailed reading tasks, requiring concentration, the blink rate drops to about a half (from 22.4 8.9 to 10.5 6.5 min–1). In the clinical setting, tear film stability has traditionally been measured following the instillation of fluorescein sodium solution into the tear film, to improve visualization of the film. Tear breakup time has been defined as the time taken for the tear film to form a dark spot or streak, following a blink. However, subsequent awareness of the disruptive effect of fluorescein instillation on the tear film has encouraged use of noninvasive techniques where tear film stability is determined by observing mires reflected from the tear film surface, for signs of disruption or distortion following a blink. A tear breakup time of greater than 10 s is considered normal while values less than 5 s are suggestive of dry eye. Values between 5 and 10 s are

56

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

generally considered to correspond to borderline dry eye, although it should be noted that this reported range was originally established for Caucasian eyes, and Asian eyes may exhibit significantly shorter tear film stabilities. In noninvasive techniques, without instillation of fluorescein, the reported cut-off values are longer with mean values around double those of the traditional fluorescein breakup test. The distribution of the tear film can further be observed in vivo using thin film interferometry. Interference fringes are produced by light reflected at the air-lipid and at the lipid-aqueous boundaries of the tear film due to the changes in refractive index. Specular reflection from the lipid layer precludes a clear view of the aqueous layer of the precorneal tear film although where the lipid layer is very thin or absent, aqueous fringes may be observed. Based on this optical principle, a number of clinical instruments, together with qualitative grading systems have been developed. These are useful for observing the structure of the tear film and offer some insight into its stability. Significant differences in appearance (and grade)

have been observed in dry eye conditions, with the partial or complete absence of the lipid layer being a feature. Recent work in this field has concentrated on developing quantitative analyses of interferometric images from the tear film of normal and dry eye patients (Figures 4 and 5, respectively). With the use of kinetic analysis of sequential interference images, it has been possible to quantify the lipid-spread time of tears in normal and dry eye patients. This spread time, defined as the time taken for the lipid film to reach a stable interference image, is significantly slower in ADDE, at 2.17 1.09 s, than it is in normal eyes (0.36 0.22 s). Because of this slower spread time, the resultant lipid film has been found to be thicker on the inferior cornea than the superior cornea, with the thickness being measured from a color reference chart created from the reflectance images of thin film interference generated by a white light source. Almost 90% of the patients with aqueous tear deficiency exhibit an interferometric pattern with vertical streaking, rather than the horizontal propagation typically observed in the superior corneal region.

Figure 4 Series of images obtained by dynamic thin film interferometry in a normal, asymptomatic subject. The images are obtained at 1 s intervals, following a blink. The lipid layer of the normal tear film reaches a relatively stable pattern within the first second after the blink. This pattern is then stable for about 6 s. Reprinted from The Ocular Surface (www.theocularsurface.com), with permission.

Tear Film

57

Figure 5 Series of thin film interferometry images obtained from a patient with severe dry eye. The patient had primary Sjo¨gren’s syndrome with a tear turnover rate of 4% min–1, evaporation of 25.5 g cm 2 s 1, volume of 3.9 ml and osmolarity of 337.6 mOsm ml–1. The images are obtained at 1 s intervals following a blink. The lipid layer of the tear film is incomplete and variable in thickness, exhibiting color fringe patterns. A stable pattern is reached in 2–3 s after the blink, but this pattern begins to be disrupted within the next 3 s. Reprinted from The Ocular Surface (www.theocularsurface.com), with permission.

Evaluation of tear film particle movement can also provide an indication of the time necessary to obtain stability of the tear film after the blink. The observed particles are thought to be accumulations of newly secreted lipid from the meibomian glands. Measuring the displacement of these tear film particles immediately after a blink has shown that the time necessary to reach zero velocity (tear stabilization time) is 1.05 0.3 s. A commercial thin film interferometer has been developed, which enables the specular reflection from the tear surface to be monitored digitally and the tear film interference patterns classified. Research with this apparatus has shown that thicker lipid layers are associated with greater tear film stability. A number of grading systems have been developed mostly assessing the uniformity of the interference fringe pattern. A change in color and loss of uniformity in distribution indicates tear film instability. Such patterns are found more commonly in dry eyes in association with thin lipid layers and reduced stability. Assessment of the reflected images from the cornea and tear film has been used to evaluate tear film quality and stabilization following the blink. High-speed videokeratoscopy assesses the regularity indices, such as surface regularity index (SRI) and surface asymmetry index (SAI), in the time interval following a blink. These indices

have been found to correlate significantly with the results of standard diagnostic tests for dry eye, such as symptoms, tear breakup time, Schirmer test, fluorescein staining score, and best corrected visual acuity.

Tear Film Osmolarity Adequate production, retention, distribution, and balanced elimination of tears are necessary for ocular surface health and normal function. Any imbalance of these components can lead to the condition of dry eye. A single biophysical measurement that captures the balance of inputs and outputs from the tear film dynamics is tear osmolarity, the end-product of variations in tear dynamics. Normal homeostasis requires regulated tear flow, the primary driver of which is osmolarity. Hyperosmolarity is thus an important biomarker for dry eye disease. Tear hyperosmolarity has been found to be the primary cause of discomfort, ocular surface damage, and inflammation in dry eye. In studies of rabbit eyes, tear osmolarity has been found to be a function of tear flow rate and evaporation. In rabbit conjunctival cell cultures, hyperosmolarity has been demonstrated to decrease the density of goblet cells and, in humans, a 17% decrease in

58

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

goblet cells density for subjects with dry eye has been reported. Granulocyte survival is significantly decreased with increases in solute concentration. Rabbit cells cultured in hyperosmolar states, above 330 mOs ml–1, show significant morphological changes, similar to those seen in subjects with dry eye. Hyperosmolarity-induced changes in surface cells in dry eye can be correlated with the degree and distribution of rose bengal staining. Measuring tear osmolarity is of benefit in the diagnosis of conditions such as dry eye. In a meta-analysis of human tear osmolarity values recorded in studies between 1978 and 2004 with freezing point depression (FPD) and vapor pressure (VP) osmolarity tests, normal values averaged 302.0 9.7 compared with 326.9 22.1 mOs ml–1 for patients with dry eye disease.

Drainage of Tears A principal means of elimination of tears from the eye is by drainage through the puncta of the eye. Tears then pass through the canaliculi, the lacrimal sac, and finally the nasolacrimal duct before reaching the nose. A technique for measuring tear turnover, which allows direct observation of tear drainage, involves instilling a radioactive dye into the tear film. In the technique of lacrimal scintigraphy a small quantity (0.013 mls) radioactive tracer such as technetium 99 (99M Tc), is introduced into the lower marginal tear strip. The distribution of the tracer is imaged serially by a gamma camera as it passes down the lacrimal drainage system (Figure 6 a–c). Images are typically taken at 10-s intervals for 1 min and then at less frequent intervals until all of the tracer has drained into the nasal cavity. The technique has been used to quantify tear turnover from the eye and drainage through

the lacrimal system. The drainage through this system is not linear, as a significant number of naso-lacrimal folds and ducts offer physiological obstruction to normal tear flow, and variable tear flow has been shown to be a typical feature of the drainage facility in asymptomatic individuals (Figure 6(d) and (e)). Therefore, most models of lacrimal drainage favor compartmental analysis to evaluate tear flow through the system, with separate components for the conjunctival sac, lacrimal sac, the nasolacrimal duct, and the nasal cavity. Although most quantitative lacrimal scintigraphy measurements describe the transit time of the radioactive tracer through the system, the compartmental model has be used to estimate tear flow rates. Depending on the number of compartments considered, basal flow rates have been estimated to fall between 0.45 and 8 ml min–1. Using a single compartment model for decay of the radioactive tracer on the conjunctival surface, mean values of reflex and basal turnover of 3.33 1.95 ml min–1 and 0.56 0.32 ml min–1, respectively, have been recorded by gamma scintigraphy. The mechanism of lacrimal drainage and the influence of blinking on the mechanics of the system have been observed by high-speed photography and by intracanalicular pressure measurements. Taking an anatomical approach and observing the lacrimal systems of human cadavers has shown that the surrounding vascular plexus of the lacrimal sac and the nasolacrimal duct is comparable to a cavernous body. While regulating the blood flow, the specialized blood vessels of this body permit opening and closing of the lumen of the lacrimal passage, which is effected by the bulging and subsiding of the cavernous body, thereby regulating tear outflow from the eye. Attempts have been made to quantify the regulation of tear outflow by measurement of the transit time of a fluorescein drop from the conjunctival sac into the inferior meatus

(a)

(b)

(c)

(d)

(e)

Figure 6 Gamma camera (a–c) used in the recording of intensity of a radioactive dye at various stages as it passes through the lacrimal system (d). In many cases of normal systems, the tracer does not proceed beyond the lacrimal sac (e). Reprinted from The Ocular Surface (www.theocularsurface.com), with permission.

Tear Film

of the nose. Application of a decongestant drug or placement of a foreign body on the ocular surface have both been found to significantly prolong the dye transit time, indicated restricted drainage through the lacrimal system in these conditions. It has therefore been concluded that the cavernous body of the lacrimal sac and naso-lacrimal duct plays an important role in the physiology of tear outflow regulation; it is subject to autonomic control and is integrated into a complex neural reflex feedback mechanism between the blood vessels, the cavernous body, and the ocular surface.

59

interdependent and have a close relationship with those of the adjacent ocular tissues such that failure of any one of aspect of the tear film or lacrimal system can cause imbalance and result in dry eye. See also: Conjunctival Goblet Cells; Contact Lenses; Defense Mechanisms of Tears and Ocular Surface; Dry Eye: An Immune-Based Inflammation; Eyelid Anatomy and the Pathophysiology of Blinking; Inflammation of the Conjunctiva; Lacrimal Gland Overview; Lids: Anatomy, Pathophysiology, Mucocutaneous Junction; Meibomian Glands and Lipid Layer; Tear Drainage.

Absorption of Tears by the Ocular Surface Another method by which tears can be eliminated from the eye is by absorption into the tissues of the ocular surface and the drainage system. The possibility has been suggested that the epithelial lining of the drainage system absorbs tear fluid before it reaches the nose. It has been shown in an animal model that lipophilic substances are absorbed from the tear fluid by the epithelium of the naso-lacrimal duct and that the cavernous body surrounding this duct may play a role in drainage of absorbed fluid. No quantification of fluid volume eliminated by this route has been reported. However, tears absorbed in the blood vessels of the cavernous body may, because these vessels connect to the blood vessels of the outer eye, have a role in a biofeedback mechanism for tear production. Observations of the absorption of tear film onto the anterior ocular surface have been made in studies of corneal permeability. The proportion absorbed, in the absence of compromised corneal function, appears to be small at 0.24 0.13% of the dye instilled in the eye. The lacrimal system of the human eye is, in the vast majority of individuals, a robust system, which allows the ocular surface to maintain its health and normal function throughout life, and under modest provocation. It is only in a relatively small proportion (15%) that the imbalance between evaporative loss and tear production results in dry eye. Recent research has confirmed that an increase in this ratio of approximately 2–3 times, as most often occurs in older individuals, appears to lead the condition of dry eye. The tears covering the anterior ocular surface, form a dynamic structure with a complex nature and a number of important functions. The tear film components are

Further Reading Bron, A. J., Yokoi, N., Gaffney, E., and Tiffany, J. M. (2009). Predicted phenotypes of dry eye: Proposed consequences of its natural history. Ocular Surface 7(2): 78–92. Craig, J. P. (2002). Structure and function of the preocular tear film. In: Korb, D. R. (ed.) The Tear Film: Structure, Function and Clinical Examination, pp. 18–50. London: Elsevier Health Sciences. Dartt, D. A. (2004). Dysfunctional neural regulation of lacrimal gland secretion and its role in the pathogenesis of dry eye syndromes. Ocular Surface 2(2): 76–91. Doane, M. G. (1994). Abnormalities of the structure of the superficial lipid layer on the in vivo dry-eye tear film. Advances in Experimental Medicine and Biology 350: 489–493. Gilbard, J. P. (1985). Tear film osmolarity and keratoconjunctivitis sicca. Contact Lens Association of Ophthalmologists Journal 11(3): 243–250. Gipson, I. K., Hori, Y., and Argu¨eso, P. (2004). Character of ocular surface mucins and their alteration in dry eye disease. Ocular Surface 2(2): 131–148. King-Smith, P. E., Fink, B. A., Fogt, N., et al. (2000). The thickness of the human precorneal tear film: Evidence from reflection spectra. Investigative Ophthalmology and Visual Science 41(11): 3348–3359. Mathers, W. D. and Choi, D. (2004). Cluster analysis of patients with ocular surface disease, blepharitis, and dry eye. Archives of Ophthalmology 122(11): 1700–1704. McCulley, J. P. and Shine, W. E. (2003). Meibomian gland function and the tear lipid layer. Ocular Surface 1(3): 97–106. Sariri, R. and Ghafoori, H. (2008). Tear proteins in health, disease, and contact lens wear. Biochemistry (Moscow) 73(4): 381–392. Stern, M. E., Beuerman, R. W., and Pflugfelder, S. (2004). Dry Eye and Ocular Surface Disorders; the Normal Tear Film and Ocular Surface. New York: Marcel Dekker. Tiffany, J. M. (2008). The normal tear film. Developments in Ophthalmology 41: 1–20. Tomlinson, A. and Khanal, S. (2005). Assessment of tear film dynamics: Quantification approach. Ocular Surface 3(2): 81–95. van Best, J. A., Benitez del Castillo, J. M., and Coulangeon, L. M. (1995). Measurement of basal tear turnover using a standardized protocol. European concerted action on ocular fluorometry. Graefes Archive for Clinical and Experimental Ophthalmology 233(1): 1–7.

Meibomian Glands and Lipid Layer T J Millar, P Mudgil, and S Khanal, University of Western Sydney, NSW, Australia ã 2010 Elsevier Ltd. All rights reserved.

Glossary Acinus – A gland that is shaped like a hollow sphere with the gland cells lining the sphere and secreting into the center of the sphere. The secretions are removed from the center of the sphere by a duct. HMG-CoA – The 3-hydroxy-3-methyl-glutarylcoenzyme A is a precursor molecule for lipid synthesis with the small precursor molecule attached to a carrier molecule, coenzyme A. Holocrine – A mechanism of secretion by a gland wherein the whole gland cell is secreted. Hydrophilic – A substance that dissolves readily in water (water loving). Hydrophobic – A substance that does not like water. Fats, lipids, and oils are common hydrophobic substances. mN m 1(millinewtons per meter) – A unit used for measuring surface pressure relative to that of water which is regarded as 0 mN m 1. Osmolarity – A measure of the number of individual molecules dissolved in water. It is important to cells because water can pass readily through a cell membrane, but the dissolved chemicals in the cytoplasm cannot. Hence, a cell will either take on water or release water depending upon whether its osmolarity is more (hyperosmolar) or less (hypoosmolar) than its environment, respectively. Refractive index – When light travels from one medium to another, for example, from air into water, it is bent. The refractive index is a measure of the extent to which the light is bent and is a constant for a particular substance. Tarsal plate – A sheet of fibrous cartilage in the eyelids of mammals that gives the eyelids their stiffness and shape.

Overview Meibomian glands are a series of fat (lipid)-producing glands found in the upper and lower eyelids of mammals, named after a German anatomist, Heinrich Meibom (1638–1700), who recorded their presence in De Vasis Palpebrarum Novis Epistola (1666). In humans, there are 30–40 evenly spread glands in the upper lid and 20–30 in the lower lid (Figure 1). Each gland is aligned vertically in

60

the eyelid and located within the tarsal plate (a sheet of fibrous cartilage that gives the eyelids their stiffness and shape) which lies closer to the ocular surface than the dermal surface. Structurally, meibomian glands have a central tubular duct surrounded by grape-like acini (glands). The duct is blind at one end and the other end opens onto the eyelid margin. By everting the lower eyelid, the openings can be readily seen as a row of small dots behind the eyelashes (Figure 2). Phylogenetically, meibomian glands are present in marsupials, but are absent in the two monotremes (echidna, Tachyglossus aculeatus; platypus, Ornithorhynchus anatinus) that we have studied. They are not present in reptiles or birds. It is believed that many of these animals use a different gland, the harderian gland, to secrete lipids onto the ocular surface. While the distribution and appearance of meibomian glands in other mammals are generally similar to those of humans, this is not always the case. Some species of voles and musk rats have few glands, for example, Microtus pinetorum, which has only two large glands in the upper eyelid – one at the medial canthus and the other at the lateral canthus. Whales have neither meibomian glands nor a tarsal plate in the eyelid. Dolphins and sea lions have a very oily secretion in their tears, but this is thought to originate from the harderian gland. Currently, the literature is not clear about the presence or absence of meibomian glands in sea-dwelling mammals. However, it is of interest to note that dolphins, sea lions, and sea otters have no eyelashes (or eyebrows), which means that they also lack the other major sebaceous glands in the eyelid, for example, the glands of Zeis which are the sebaceous glands of the eyelashes. The major function of the meibomian gland is to supply the main components of the outer layer of the tear film. The tear film is a thin (7–10 mm thick), watery fluid that covers the exposed surface of our eyes (Figure 3). Lipids from the meibomian glands are secreted onto the inner margin of the eyelids where they contact and then spread over the aqueous part of the tear film to form a covering layer (90 nm thick) in contact with the air. This layer is referred to as the lipid layer of the tear film. It is believed to decrease evaporation from the tear film and hence prevent dry eyes. However, this role for the meibomian lipids is by no means certain, and it is likely that it has other roles such as preventing tears from flowing onto the skin and skin lipids from flowing onto the ocular surface, assisting the spread of the tear film over the eye by lowering the surface tension, and forming a watertight seal when the lids are closed. The Meibomian lipids provide a smooth and

Meibomian Glands and Lipid Layer

Figure 1 Arrangement of the meibomian glands in the upper and lower eyelids.

Figure 2 Meibomian gland orifices (arrows) in the lower lid margin.

highly refractive (1.4766 at 589 nm and 35 C) surface. Clinically, a mechanism for measuring tear breakup time (TBUT) is to observe changes to interference colors of the surface layer of the tear film. This can be used as one measure of the tear film performance.

The Lipid Layer of the Tear Film The lipid layer of the tear film provides an optically smooth surface at the interface between the air and the aqueous part of the tear film. Although the structure of this layer has not been determined, the meibomian lipids form a major component (Figure 3). A useful model of the lipid layer, based on the idea that only lipids are present, was developed by McCulley and Shine. The model is presented as a crystalline array, and, while very useful

61

for developing understanding, the lipid layer is unlikely to be crystalline in practice. This model proposed that the wax esters, steryl esters, triglycerides, and hydrocarbons (hydrophobic lipids) from the meibomian glands reside in the outermost layers and are linked to the aqueous subphase by polar lipids (phospholipids, free cholesterol, and free fatty acids). Some of these polar lipids may arise from the aqueous layer of the tear film and others from meibomian lipids and hydrolysis (cleavage) of the wax esters, steryl esters, and triglycerides. These polar lipids have a hydrophobic part which interacts with the hydrophobic lipids and a hydrophilic part that is able to interact with water (the aqueous layer). Assuming that the outer layer of the tear film is entirely made from lipids implies that the interfacial molecules (surfactants) must be polar lipids. An alternate model proposing that proteins and mucins also contribute to this layer is more realistic. Strong evidence comes from surface tension measurements. It has been found that the surface tension of tears is 42–46 mN m 1 and this can only be achieved by a mixture of lipids and proteins and not with meibomian lipids alone. The interaction of meibomian lipids with tear proteins is often cautiously presented as lipids occupying the outer surface of the lipid layer, and the inner surface interacting with proteins from the aqueous layer. A more recent model includes proteins and mucins as integrated parts of this layer (Figure 3). Although there has been a focus on lipocalin, an abundant lipid-binding protein in tears, being the main tear protein interacting with the meibomian lipids, lysozyme and lactoferrin may be more involved. Lipocalin is thought to scavenge lipids that have adhered to the epithelial cells of the ocular surface and lipids that are in the aqueous layer. Although it has been claimed that these are then transported to the outer lipid layer, this may not be the case. Once a lipid is bound into the central pocket of lipocalin, lipocalin is in a low-energy state and unlikely to interact with the meibomian lipids at the outer surface of the tear film. The presence of proteins in the lipid layer has important conceptual implications because they are large molecules with complex mixtures of hydrophobic, hydrophilic, and distinctly charged components. These properties mean that they can unfold and form a range of shapes according to their local molecular environment. Due to this unfolding, it is possible for them to extend across the lipid layer and interact with the hydrophobic lipids and other proteins. As the model suggests, this means that the layer comprises a complex mixture of islands of proteins, islands of lipids, islands of mucins, and various mixtures of these (Figure 3). This model is more akin to models for cell membranes and for lung surfactant. Some advantages of this model are that: the outer layer would be a noncollapsible viscoelastic gel; it would allow for the lowest free energy states of the proteins in contact with lipids; and the changes in salt concentrations in the tears

62

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Key for molecules shown on the right side

L

Lipocalin

Cholesterol ester

Lysozyme

Wax ester

Lactoferrin

Phospholipid

IgA

Ganglioside

Mucins

Fatty acid

Lipid layer

Aqueous/mucin layer

Corneal epithelium Figure 3 Tear film comprising an outer lipid layer covering an aqueous/mucin layer. The exploded view of the lipid layer shows a mixture of different lipids and denatured proteins (see key).

(e.g., hyperosmolarity) would affect the ability of the proteins to unfold and interact with lipids in the outer layer. Taking these one at a time, the lipid-only model needs some mechanism to spread the lipids over the aqueous surface after a blink. Both phospholipids and mucins have been proposed as enabling this by spreading slightly ahead of the hydrophobic lipids, and this concept also means that a new lipid layer is formed at each blink. Lack of change over a number of blinks to the interference patterns formed by the outer layer of the tear film does not support this concept. However, a noncollapsible viscoelastic gel that would be formed with proteins and mucins in the layer obviates the need for spreading and would account for the consistent interference pattern over a number of blinks. Furthermore, in other fields of study, mixtures of molecules in a liquid environment autoassemble into states of lowest free energy, which means that when lipids and proteins are placed together, they will mix at the molecular level rather than remain separate as the lipid-only model suggests. Some of the proteins can act as the surfactants, meaning that while phospholipids may co-jointly serve as surfactants, they are not absolutely necessary. It has also been found that while lipids alone are not capable of lowering the surface tension to the levels found in tears, mixtures of tear proteins and tear lipids possess that ability. Hyperosmolarity of the tears has the strongest correlation with a dry eye. Salt concentrations have strong effects on how proteins fold; therefore, it is possible that high salt concentrations would affect the folding of proteins and hence their ability to interact with the lipid layer to form a stable outer layer.

In turn, this could affect the surface tension and spreading of the outer layer and, consequently, its protective function.

Meibomian Glands Anatomy and Histology There are no anatomical differences in the meibomian glands of human males and females. The ducts in humans are approximately 1.6 mm long with the central ducts being slightly longer than the nasal and temporal ducts, and are surrounded by a dense banding of elastic fibers. The ducts are lined by keratinized epithelium and lie nearly 780 mm from the dermal surface of the eyelid. A horny cell layer overlies one or two layers of intermediate cells that rest on cuboidal basal cells connected to a basement membrane, and there is no difference in this appearance between the proximal and distal portions of the duct (Figure 4). Acini are arranged circularly around the central duct and are connected to it by short ductules. The acinar cells are distinct from the ductal cells with no keratohyaline granules or lamellar bodies in the acinar cells, and no lipid vesicles in the ductal cells. The acinar cells continually differentiate into holocrine-secreting cells from basal acinar cells. They contain an abundance of smooth endoplasmic reticulum that surrounds the lipid vesicles. These cells also develop from the division of basal cells and move toward the center of the acinus (migration rate of 0.62 mm d 1 in rats), and slowly increase in neutral lipid content and in the size of the lipid-containing vesicles. The acinar cells die and gradually breakdown, leaving a

Meibomian Glands and Lipid Layer

63

central epithelial cells. At 15 weeks (100 mm), cuboidal secretory cells line the ducts. The upper and lower eyelids are fused during this whole process. At 7.5 months (250 mm), after the eyelids have separated, acini are present, epithelial cells plugging the meibomian gland orifices disintegrate, and secretion begins just prior to birth. A similar process takes place in the mouse; however, in this case, the eyelids are still fused at birth, which is also the time when the first signs of meibomian gland development show, again from the eyelid margin. In contrast to meibomian gland development, eyelashes in humans begin development earlier at 8 weeks (35 mm) as solid epithelial invaginations from the external face of the eyelid margin. Figure 4 An illustration of meibomian glands and duct. A: Ghosts of acini not filled in with detail. B: Acini showing nucleated basal cells around the periphery which gradually lose their nuclei as they mature and move to the center of the acinus. C: Secondary duct surrounded by ductal cells (stratified epithelium) containing meibomian secretion and cell remnants. D: Main duct surrounded by ductal cells containing meibomian secretion and cell remnants. The arrow shows a bundle of nerve endings close to the acinus but separated by collagen fibers.

lipid mass that is secreted via the duct. In rats, it takes close to 9 days for a cell to migrate to the center of the acinus, and it probably divides twice during this time. Surrounding the glands is a distinct extracellular matrix comprising collagen types I, II, and IV, aggrecan, dermatan sulfate, and chondroitin-6-sulfate. Close to the acini are numerous unmyelinated varicose nerve fibers with boutons in contact with collagen fibers within the basal lamina of the basal acinar cells. These are mainly parasympathetic fibers which contain the neurotransmitter, acetylcholine, and the neuropeptide vasoactive intestinal polypeptide. The cell bodies for these fibers lie in the pterygopalatine ganglion and their fibers reach the eyelid by the greater petrosal nerve. Preganglionic neurons are ipsilateral and cholinergic, and lie in the superior salivary nucleus located lateral, dorsal, and caudal to the superior olive and lateral, dorsal, and rostral to the facial nucleus. Sympathetic innervation is sparse and mainly associated with blood vessels. Sensory nerve fibers are also sparsely distributed close to the basal region of acini, and these are immunoreactive for the neuropeptides substance P and calcitonin gene-related peptide. Development Meibomian glands in humans develop as small, solid epithelial invaginations from the ocular face of the eyelid margin at approximately 9 weeks of development (crownrump length of 40 mm). By 12 weeks (60 mm), the epithelial growth has extended the depth of the tarsal plate and has a central tube that was formed from apoptosis of the

Composition of Meibomian Lipids The meibomian gland is often referred to as a modified sebaceous gland. In this case, the term sebaceous means lipid producing rather than sebum producing because the composition of the lipids differs from those produced by the sebaceous glands of hair follicles. The main lipid types produced by meibomian glands are wax and steryl esters (60–70%) which are very hydrophobic (dislike water). Wax esters are formed by linking a long-chain carboxylic acid (fatty acid) to a long-chain alcohol (fatty alcohol). Since different fatty acids and fatty alcohols are linked together, the wax esters are a complex family of lipids and their detailed structure varies between species. In humans, oleic acid is the most prevalent fatty acid found in these waxes. Similarly, the steryl esters are mainly cholesterol esters which are formed by linking cholesterol to a longchain fatty acid. These fatty acids are generally longer than those found in the wax esters. Small amounts of other lipids (mainly polar) – such as mono-, di-, and triglycerides; fatty acids; fatty alcohols; free cholesterol; and phospholipids – make up the remainder. While phospholipids are readily detected in meibomian lipids of rabbits, there is contention as to whether they are a component of human meibomian lipids. This is important because in models of the lipid layer of the tear film, phospholipids are crucial as a link between the hydrophobic molecules (wax and cholesterol esters) and the aqueous layer. If they are not present in the meibomian lipids, then they must be derived from elsewhere, such as the aqueous layer of the tear film, or alternative surfactants need to be present in the model. The nature and mixture of the lipids give them a melting range of 19–33 C, which means that they are fluid on the ocular surface. Meibomian Lipid Turnover and Synthesis Since the meibomian gland is a holocrine gland, lipid turnover is related to the cell turnover rate and the lipids are synthesized by the glands rather than being adsorbed from the bloodstream. For instance, the levels of

64

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

cholesterol found in meibomian lipids are independent of cholesterol levels in the blood. Synthesis of the straightchain fatty acids is typical of elsewhere in the body, occurring in the cytoplasm using acetyl-coenzyme A (CoA) and malonyl-CoA as the starting molecules and fatty acid synthase as the major enzyme. Some of the fatty acids have branching and, at least in rabbits, the branched carbon chains are derived mainly from the amino acids valine and isoleucine. The acyltransferases seem to be nonselective and will connect any fatty acid to a fatty alcohol. Similarly, cholesterol synthesis is typical of other tissues converting b-hydroxy-b-methylglutarylCoA to mevalonate, isoprene, squalene, and cholesterol. The acylcholesterol transferase, involved with cholesterol ester synthesis, appears to be selective for longer-chain fatty acids based on the predominance of long-chain fatty acids in cholesterol esters. The meibomian acinar cells have nuclear androgen receptors and the level of androgens or, more likely, the ratio between androgens and estrogens is critical for controlling lipid synthesis in the meibomian glands. Stimulating androgen receptors increases gene transcription in enzymes associated with fatty acid and cholesterol synthetic pathways (adenosine triphosphate (ATP)-citrate lyase, acetyl-CoA synthase, acetoacetyl-CoA synthase, 3-hydroxy-3-methylglutary (HMG)-CoA synthase 1, HMG-CoA reductase, acetyl-CoA carboxylase, glyceraldehyde-3-phosphate dehydrogenase, and sterol regulatory element-binding protein 1 and 2) and hence stimulates lipid production. Androgen deficiency has been associated with meibomian gland dysfunction (MGD) and dry eye. P2Y2 receptor gene expression has also been detected in meibomian gland acini. This suggests that extracellular ATP or UTP might influence lipid synthesis or composition through activation of G proteins. Specific changes to lipid composition through this pathway have not been investigated. The total amount of meibomian lipids on the lid margins has been estimated and children under 14 years of age showed the lowest levels (1.5 mg mm 2 lid margin). After puberty, there is a steady increment with age until the late 60s (3.26 mg mm 2) and throughout this period, males have nearly 10% greater levels than females. Given that one of the reasons for dry eye is insufficient meibomian lipid secretion and that there is an increase in the incidence of dry eye with age, it is surprising that the lowest levels of meibomian lipids have been found in children. Morning and afternoon basal levels are the same and there is no correlation between lid temperature (30–34 C) and basal levels of meibomian lipids. However, deliberately increasing the eyelid temperature from 33 to 37 C increases the lipid values on the lid margin by close to 25%. Despite the presence of parasympathetic nerves around the acini, lipid secretion is most likely due to the mechanical force of blinking which causes compression of the territorial fibrocartilaginous matrix that surrounds the

meibomian glands. This has been shown by measuring the reappearance of lipids on the eyelid margin after cleaning with an organic solvent. No lipids appear until a blink occurs (3 min was the longest). After approximately 10 blinks, the levels return to nearly a third of their basal levels. It is estimated that close to 10 mg of lipids are delivered per blink and that there is approximately 20–40 times excess basal amount of lipids available on the eyelid margin than what is required for forming a complete lipid layer on the tear film. How the lipids are removed from the eye is uncertain. It is believed that most of them flow over the eyelid margin onto the skin and eyelashes. This constant flow prevents the skin lipids from contaminating the tear film. It has been shown that skin lipids disrupt the tear film. Some lipids are likely to bind to proteins of the aqueous layer, particularly lipocalin, and are removed with the aqueous layer through the lacrimal ducts. The crusty buildup that collects in the corner of the eye during sleep is primarily a mixture of lipids and mucins and thus another mechanism for removing lipids from the ocular surface.

Pathology of the Meibomian Gland Disorders of the meibomian glands are manifest by the obstruction of the gland orifices, inflammation, or loss of the glands. This is often associated with one or more of the following: thickening of the lid margin, exaggerated vascularization around the gland orifices, and hyperkeratinization. Meibomian gland diseases are usually more uncomfortable rather than painful and, when chronic, are associated with dry eye which can be very painful. Absence or deficiency of meibomian glands is often congenital. Clinically, the state of meibomian glands is determined by examining their morphology and function. In eye clinics, the orifices of the glands lining both the upper and lower eyelids are observed through a slit lamp biomicroscope. If a gland is blocked, the orifice appears swollen on the lid margin. Some practitioners also squeeze the lower eyelid gently to expel meibomian lipids. A clear fluid is considered to be normal, whereas a thick, yellowy secretion is an indication of meibomian gland disorder. If excessive pressure is applied, a thick pasty expression can be obtained from people with normal meibomian gland function (Figure 5). This technique is used for obtaining meibomian lipids for experimental purposes or for analysis. In research settings, more specific tests are performed to assess meibomian gland function. Transillumination of the lower eyelids is widely used to evaluate the morphology of the glands. In particular, shorter-than-normal meibomian glands and meibomian gland dropout are strong indications of MGD (Figure 6). Changes to the shape and form of the glands do not occur with aging,

Meibomian Glands and Lipid Layer

although expression of secretion is commonly more difficult. A clear, noninvasive view of meibomian glands can be achieved by using infrared light, but the need for specialized equipment means that this technique is only used in a few research laboratories. Rather than examine the meibomian glands themselves, observation of the lipid layer of the tear film is regarded as an indirect measure of both the quantity and the quality of meibomian gland secretions. Using specialized interferometry, the lipid layer stability, distribution, dynamics, and thickness can be assessed. Generally, a thick amorphous layer is an indication of high-quality meibomian oil secretions and, hence, excellent meibomian gland function. Such a layer is also commonly associated with longer TBUTs, presumably because of the lowered surface tension. The

65

results obtained from tear breakup tests have been shown to be comparable to interferometry findings, although a cause–effect relationship between the two is yet to be established. Another method for evaluating meibomian gland function indirectly is to measure evaporation from the ocular surface. This is a difficult technique and is normally confined to research settings. Evaporimetry is based on the theory that evaporation from the ocular surface is normally minimal due to the well-spread lipid layer acting as a blanket, and that an inadequate lipid layer or disruption to this lipid layer causes tear evaporation to increase. It is thought to be the cause of evaporative dry eye. However, this idea has been difficult to substantiate and a wide variability in tear evaporation rates has been reported irrespective of the presence or absence of dry eye and, more importantly, the appearance of the lipid layer. Further, a large quantity of lipid on the tear surface does not necessarily correlate with an adequate barrier to evaporation. Evaporative dry eye can occur with an excessively thick lipid layer. Current areas of research center on whether the biochemical composition of the meibomian lipids can influence their surface activity and ability to diminish evaporation, but clear outcomes are still in the future.

Chronic Blepharitis

Figure 5 Hard squeezing of meibomian glands. The secretions are indicated by arrows.

The term blepharitis has different meanings depending upon the user. Acute blepharitis (normally just called blepharitis) is an infection of the anterior eyelid and Staphylococcus epidermidis or Staphylococcus aureus are the most likely cause. Chronic blepharitis is caused by dysfunction of the meibomian glands and is synonymous with

Region of meibomian gland dropout Meibomian gland in eyelid

Figure 6 The transillumination of eyelid showing the dark meibomian gland acini. There is meibomian gland dropout in the middle region of the lid. Courtesy of Jerry Paugh, Southern California College of Optometry.

66

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Chalazion

Stye

Figure 7 A chalazion appears as a swelling deep within the eyelid. A stye is associated with the eyelashes (infection of the eyelash follicle). Courtesy of Rob Terry, Institute of Eye Research, University of New South Wales.

MGD and posterior blepharitis. Many studies have looked at the association between blepharitis or chronic blepharitis and dry eye. However, the conclusions drawn about the association depend on the definitions being used in the study. Although MGD occurs in nearly three-quarters of patients with chronic blepharitis, it also occurs in approximately 20% of people with normal tear function. In MGD, the orifices of the meibomian glands are blocked, reducing the secretion of meibomian lipids onto the ocular surface. Regular warm compresses help to open the orifices and allow normal lipid secretion. Some patients with chronic blepharitis have similar symptoms to those of dry eye and are prescribed tear lubricants for palliative purposes, that is, they do not resolve the blepharitis. Chalazion The blockage of the meibomian glands can lead to formation of a chalazion (Figure 7). This is a cyst on the eyelid that is normally sterile and composed of a lipid granuloma. It looks similar to a stye, which is caused by an infected sebaceous gland of the eyelash, but can be easily distinguished clinically because a chalazion is painless and develops gradually, whereas a stye is always painful and forms over a few days. For both conditions, warm compresses are recommended. In extreme cases of chalazion, it is surgically incised and the granulomatous material is removed by curettage. Antibiotics are often prescribed to treat a stye. Surgical Damage Treatment of trachoma is a common cause for surgical damage to the meibomian gland. Trachoma, a leading cause of blindness, is an infectious disease of the palpebral conjunctiva that leads to the eyelids folding inward (entropion), causing the lashes to rub against the cornea. The lids, and hence the meibomian glands and ducts, are cut to relieve this condition. It is yet to be determined whether this compromises the functionality of the outer lipid layer of the tear film. Other surgical procedures such

as correction of lid malpositioning, particularly ptosis (drooping of the upper eyelid) and genetic entropion, can also sometimes require the cutting of the meibomian glands.

Contact Lenses and the Lipid Layer Anomalies of the lipid layer, in themselves, are not a deterrent for contact lens wear. Lipids or proteins or both are deposited on contact lenses during wear. These deposits can block the small pores of the contact lenses, which are essential for the passage of air to the cornea for its metabolism. It is impossible to ascertain beforehand how long a contact lens needs to be worn for it to be unsuitable for an individual as the amount and pattern of lipid deposition depends on the composition of the ocular lipids, which can vary between people, and the specific material the lens is made from. Contact lens cleaning agents are designed for the specific type of contact lens and normally contain a surface active agent that removes lipid deposits. See also: Dry Eye: An Immune-Based Inflammation; Ocular Mucins; Tear Film.

Further Reading Bron, A. J., Benjamin, L., and Snibson, G. R. (1991). Meibomian gland disease, classification and grading of lid changes. Eye 5: 395–411. Bron, A. J., Tiffany, J. M., Gouveia, S. M., Yokoi, N., and Voon, L. W. (2004). Functional aspects of the tear film lipid layer. Experimental Eye Research 78: 347–360. Butovich, I. A., Millar, T. J., and Ham, B. M. (2008). Understanding and analysing Meibomian lipids – a review. Current Eye Research 33: 405–420. Glasgow, B. J., Marshall, G., Gasymov, O. K., et al. (1999). Tear lipocalins: Potential scavengers for the corneal surface. Investigative Ophthalmology and Visual Science 40: 3100–3107. Goto, E. and Tseng, S. C. G. (2005). Kinetic analysis of tear interference images in aqueous tear deficiency dry eye before and after punctual occlusion. Investigative Ophthalmology and Visual Science 44: 1897–1905. Gouveia, S. M. and Tiffany, J. M. (2005). Human tear viscosity: An interactive role for proteins and lipids. Biochimica et Biophysica Acta 1753: 155–163.

Meibomian Glands and Lipid Layer Holly, F. J. (1973). Formation and rupture of the tear film. Experimental Eye Research 15: 515–525. Hykin, P. G. and Bron, A. J. (1992). Age related morphological changes in lid margin and Meibomian gland anatomy. Cornea 11: 334–342. Jester, J. V., Nicolaides, N., and Smith, R. E. (1981). Meibomian gland studies: Histologic and ultrastructural investigations. Investigative Ophthalmology and Visual Science 20: 537–547. Mathers, W. (2004). Evaporation from the ocular surface. Experimental Eye Research 78: 389–394. McCulley, J. P. and Shine, W. (1997). A compositional based model for the tear film lipid layer. Transactions of the American Ophthalmological Society 55: 79–93.

67

Millar, T. J., Tragoulias, S. T., Anderton, P. J., et al. (2006). The surface activity of purified ocular mucin at the air–liquid interface and interactions with meibomian lipids. Cornea 25: 91–100. Nagyova, B. and Tiffany, J. M. (1999). Components responsible for the surface tension of human tears. Current Eye Research 19: 4–11. Sullivan, D. A., Sullivan, B. D., Evans, J. E., et al. (2002). Androgen deficiency, Meibomian gland dysfunction, and evaporative dry eye. Annals of the New York Academy of Science 966: 211–222. Tiffany, J. M. (1995). Physiological functions of the Meibomian glands. Progress in Retinal and Eye Research 14: 47–74.

Lacrimal Gland Overview M C Edman, R R Marchelletta, and S F Hamm-Alvarez, University of Southern California School of Pharmacy, Los Angeles, CA, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Acinus – Originating from the Latin word grape, it refers to the sac-like ending of a secretory exocrine gland. Endocytosis – The process of internalization of plasma membrane as well as membrane-bound constituents and extracellular fluid by invagination of the plasma membrane, budding of the membrane vesicle, and its movement to the interior. Different types of endocytosis are known, including clathrin-mediated and caveolar endocytosis. Exocytosis – The process by which a cell releases the contents of secretory vesicles to the extracellular environment by fusion of secretory vesicles with a plasma membrane domain. Motor proteins – Mechanochemical proteins that utilize the energy of ATP hydrolysis to generate motive force along a polar surface, typically an actin filament or a microtubule. Rab proteins – Small GTP-binding proteins that utilize the GTP binding and hydrolysis cycle to trigger protein on and off states, and which serve as molecular zip codes to specify the accurate sorting and targeting of membranes. SNARE proteins – Proteins associated with donor and acceptor membranes which associate to form a fusion pore, allowing the contents of opposing membrane vesicles to intermingle, or allowing the extrusion of membrane-encapsulated contents to the cell exterior. Transcytosis – The process by which macromolecules are transported through a polarized cell. trans-Golgi network – A post-Golgi processing compartment responsible for the accurate segregation of contents into membrane vesicles destined for regulated exocytosis, constitutive exocytosis, or for targeting to intracellular membrane compartments.

Anatomy of the Main Lacrimal Gland The human main lacrimal gland, located laterally above the eye, measures approximately 20 12 5 mm and has an almond-like shape. The major part of the gland, designated as the orbital portion, or the intraorbital gland,

68

is located in the shallow lacrimal fossa of the frontal bone, while the smaller palpebral, or extraorbital portion, which is separated from the orbital portion by the lateral horn of the levator palpebrae muscle, is located above the temporal segment of the superior conjunctival fornix. In contrast, the mouse and rat have two pairs of lacrimal glands including a small orbital gland which is located laterally beneath the upper lid and a larger extraorbital gland which is located ventral and anterior to the eye. The rabbit lacrimal gland is unusually large and is comprised of a larger portion (4 cm) located below the eye and a smaller portion (0.5 cm) located above the eye. The lacrimal gland is constituted largely (80%) of acinar epithelial cells organized within the tubuloacinar units that are arranged into multiple globuli surrounded by fibrovascular septa. The remaining 20% of the mass of the lacrimal gland is composed of ducts, nerves, myoepithelial cells, leukocytes, and connective tissue. A schematic diagram showing the positioning of the gland relative to the ocular surface and the organization of several of the cell types within the gland is shown in Figure 1.

Cell Types within the Lacrimal Gland Acinar Cells The acinar epithelial cells within the lacrimal gland are triangular-shaped cuboidal cells organized in single cell layers in clusters with a narrow microvillus-covered apical domain oriented toward a central lumen and a more extensive basolateral domain which faces the tissue interstitium. Tight junctions near the apices segregate these two domains and result in polarization of the cells, which also are cytoplasmically coupled through gap junctions, thus making the acinus a single functional unit. The apical side of the cell is enriched in numerous large (1–2 mm) secretory granules or vesicles, containing proteins released upon cell stimulation, while the Golgi apparatus and endoplasmatic reticulum compartments are located more toward the basolateral side adjacent to the basolateral nucleus. Figure 2 shows the characteristic distribution of secretory vesicles and other organelles within an acinar cell from mouse lacrimal gland. Ductal Cells The lumena of several acini come together to form a duct; each duct merges with others into gradually larger ducts

Lacrimal Gland Overview

69

Myoepithelial Cells L D

CNS

N

BM AM

LG

SV

ME

OS

NE

Figure 1 Schematic diagram showing the positioning of the human main lacrimal gland relative to the ocular surface and the organization of several of the cell types within the gland. AM, apical membrane; BM, basolateral membrane; CNS, central nervous system; D, duct; L, lumen; LG, lacrimal gland; ME, myoepithelial cell; N, nucleus; NE, nerve ending; OS, ocular surface; SV, secretory vesicle.

N

The acini are surrounded by stellate-shaped myoepithelial cells with long slender processes. The myoepithelial cells not only exhibit characteristics of other epithelial cells, such as expression of cytokeratin, but also exhibit properties of smooth muscle cells such as expression of a-smooth muscle actin. The exact role of the myoepithelial cells in the regulation and maintenance of the lacrimal gland remains unclear, but it has been shown that they express receptors for neurotransmitters, suggesting that they play a role in facilitating the secretion from the lacrimal gland. It is also likely that an important role is to support and maintain the structure of the lacrimal gland. Bone-marrow-derived Cell Population The lacrimal gland is a part of the mucosal-associated lymphoid tissue (MALT). The bone-marrow-derived cells in the lacrimal gland are mainly immunoglobulin A (IgA)-producing plasma cells and T and B lymphocytes, but macrophages and mast cells are also present. The bone-marrow-derived cells are clustered into lymphoid follicles scattered in the stroma surrounding the acini.

Innervation of the Lacrimal Gland SV

L 2000 nm

Figure 2 Transmission electron micrograph of mouse lacrimal gland. N, nucleus; L, lumen; SV, secretory vesicle.

that finally, in humans, drain into 6–12 major ducts with openings in the upper lateral fornix. In rat and mice, the ducts from the extraorbital gland join to a single major duct that then joins the duct from the intraorbital gland before it empties onto the conjunctiva in the lateral canthus of the eye. In rabbit, a single duct each forms from the upper and lower portions of the gland, which empty onto the conjunctiva of the upper and lower lids, respectively, near the temporal angle. The ducts are formed by one to two layers of cuboidal epithelial cells. Similar to the acinar cell, the ductal epithelial cells of small ducts are polarized by tight junctions in the apical area. However, in these cells, the Golgi and endoplasmic reticulum are located more apically, and secretory vesicle content is lower. Interlobular ducts are embedded with and supported by perivasculoductal connective tissue containing associated structures such as nerve fibers, capillaries, and mast cells and a dense population of fibroblast-associated collagen fibrils.

The lacrimal gland is innervated by parasympathetic, sympathetic, and sensory nerves. Parasympathetic nerves originate in the lacrimal nucleus of the pons and travel along the nervus intermedius, the deep and superficial petrosal nerves, and the vividian nerve before they synapse in the pterygopalatine ganglion. The postganglionic parasympathetic fibers can take different routes to the lacrimal gland. They leave the ganglion through the pterygopalatine nerves but can also reach the lacrimal gland via the maxillary portion of the trigeminal, the zygomatic, or the lacrimal nerves. Parasympathetic fibers can also travel along a branch of the middle meningeal artery to join the ophthalmic or lacrimal artery en route to the lacrimal gland. Sympathetic nerves originate from the superior cervical ganglion and travel along with the parasympathetic nerves through the pterygopalatine ganglion, reaching the lacrimal gland through the lacrimal branch of the zygomatic branch of the maxillary trigeminal nerve that joins the ophthalmic branch of the trigeminal nerve. The sensory nerves innervating the lacrimal gland carry sensory information from the gland through the ophthalmic branch of the trigeminal nerve to the trigeminal ganglion. The parasympathetic nerves, being the most abundant, regulate the lacrimal gland mainly through the release of neurotransmitters, acetylcholine and vasoactive intestinal

70

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

peptide (VIP), with a possible co-secretion of nitric oxide (NO). Acetylcholine activates M3 muscarinic receptors located in the basolateral membrane of the lacrimal cells, while VIP binds to VIP receptors that are similarly located. The sympathetic nerves exert their effects on the lacrimal gland through release of norepinephrine that binds to a- and b-adrenergic receptors, and possibly through neuropeptide Y receptors also located at the acinar cell basolateral membranes. The sensory nerves release substance P and calcitonin gene-related peptide. Not every individual acinar cell is independently innervated; rather cells that are not directly innervated can respond to stimulation of neighboring cells due to the intercellular gap junctions that connect the cells. The density of synapses within each acinus varies according to the species: in the rat, orbital glands fewer than 15% of acinar cells have an adjoining nerve fiber in contrast to the mouse orbital glands where close to 100% of the cells have an adjoining nerve fiber.

Blood Supply The main blood supply to the lacrimal gland is not only through the lacrimal artery, a branch of the ophthalmic artery, but it also receives minor contributions from the infraorbital and the meningeal arteries. The veins mainly follow the same pathways as the arteries inside the orbit and drain into the superior ophthalmic vein.

Contents of Lacrimal Fluid The tear film consists of three layers: a mucus layer located directly above the ocular surface epithelium, an aqueous layer, and a thin external lipid layer. The lacrimal fluid produced by the main lacrimal gland constitutes the major part of the aqueous layer of the tear film, to which the accessory glands of Krause and Wolfring and the ocular surface epithelium also contribute, in a minor fashion. Although the major part of the aqueous layer is water, it also contains electrolytes and a high concentration of proteins. Human tear fluid, for instance, has a protein concentration of about 8 mg ml1. Although the three layers of the tear film largely originate from different sources, these sources can contribute in part to each layer; therefore, it is hard to determine the origin of a specific protein. Recently, researchers identified 419 different proteins in human tear fluid; however, many of these may not have an active function in the tear fluid but are simply debris shed from epithelial cells. Three proteins constitute 80% of the total protein within the aqueous tear film, that is, lipocalin, lysozyme C, and lactoferrin. The functions of the different proteins in the tear fluid are varied. For instance, many proteins contribute

to the antimicrobial properties of the tear fluid. Secretory IgA and cytokines are involved in immune responses, while others such as lysozyme C and lactoferrin provide a more direct defense against bacteria. The novel protein, lacritin, acts as a mitogen in corneal regeneration. Other proteins in the lacrimal fluid are involved in diverse activities in wound healing, blood coagulation, and oxidative stress reduction – all functions essential to maintain a healthy ocular surface. The protein pattern of both active and inactive proteins in the tear fluid can reflect disease states, including diabetes, dry eye, and cystic fibrosis.

Mechanisms of Protein Secretion in the Lacrimal Gland The acinar cells of the lacrimal gland are professional excretory cells that engage in several types of secretion that collectively contribute to the lacrimal fluid. Protein secretion at the apical membrane into lacrimal fluid can be subdivided into several types: regulated exocytosis, constitutive exocytosis, and transcytosis. Regulated exocytosis is a process in which proteins destined for a particular plasma membrane domain are sorted into secretory vesicles after their biosynthesis within a post-Golgi sorting compartment called the trans-Golgi network. These secretory vesicles mature and migrate toward the site of release, where they are stored until the appropriate signal triggers their movement and fusion with the acceptor membrane domain. Examples of content proteins thought to be released from the lacrimal acini in animal model systems through regulated exocytosis at the apical plasma membrane include peroxidase and b-hexosaminidase. Constitutive exocytosis occurs, similarly, as components for release to the exterior of the cell are sequestered into vesicles at the trans-Golgi network that are immediately targeted to the acceptor membrane. Unlike regulated exocytosis, constitutive exocytosis is not reliant on extracellular activation of receptors by a ligand (such as a hormone or a neurotransmitter) to elicit the event. Although both forms of exocytosis have been observed in lacrimal acini, most studies have focused on the role of regulated exocytosis in the release of proteins at the apical plasma membrane of the acinar cell into lacrimal fluid. The secretory vesicles in the lacrimal gland acinar cells are generally larger (1–2 mm) and considerably more heterogeneous in both size and content compared to vesicles in other exocrine glands such as exocrine pancreas and the salivary gland. The spectrum of proteins secreted from the lacrimal acinar secretory vesicles appear to span a greater functional range than the spectrum release from other exocrine tissues as well, including nutrient and protective factors as well as factors that protect the mucosal surface from pathogens. This is an area of very active research since there are an unusually large number of

Lacrimal Gland Overview

proteins of unknown function in the lacrimal gland secretory proteome. Transcytosis is a process in which material internalized at the basolateral membrane is recruited into vesicles by endocytosis, followed by the movement of these vesicles to apically located compartments, and ultimately their targeting to the apical plasma membrane for release. Major cargo known to be carried through this pathway includes dimeric IgA, through association with the polymeric IgA receptor. Although not specifically characterized in lacrimal acini, other abundant tear proteins, including albumin and transferrin, are known to be transported through transcytotic pathways in other epithelial cells, suggesting that these proteins may be comparably transcytosed into lacrimal fluid by lacrimal acini. Regulated exocytosis and transcytosis utilize a number of different processes, globally referred to as membrane trafficking, to achieve the unidirectional transport of cargo-laden vesicles over short and long distances followed by their targeted release. Several of these processes have been studied in acinar cells, including cytoskeleton and motor proteins, rab proteins, and soluble N-ethylmaleimide-sensitive factor attachment protein receptors or SNARE proteins. All mammalian cells contain filamentous structures collectively referred to as the ‘cytoskeleton’, which includes actin filaments, intermediate filaments, and microtubules. Each of these structures is formed from individual subunit proteins that exist in equilibrium with polymeric assemblies. Cytoskeletal filaments are critical in maintaining the integrity of cell shape as well as conferring cellular polarity or asymmetry, a function critical in aiding the movement of materials to different membrane domains in a polarized cell. Filamentous actin and microtubules, in particular, participate in several capacities in the movement of membrane vesicles to the apical plasma membrane, where the release of proteins into lacrimal fluid takes place. Microtubule- and actindependent membrane transport events can be facilitated either by the use of compressive force associated with cytoskeletal assembly to physically compress or direct membrane traffic, and/or by the use of these polymers as tracks which support the movement of motor proteins which carry membrane vesicles to specific destinations. In lacrimal acini, actin filaments are localized in a dense network below the apical membrane called the subapical actin network, and this network is also present to a lesser extent below the basolateral membrane. Beneath the subapical actin, the ends of microtubules are anchored. Microtubules extend from the subapical region to the basolateral membrane. Both actin filaments and microtubules sustain aspects of protein secretion in lacrimal acinar cells. When microtubules are disrupted in acinar cells using the agent, nocodazole, stimulated protein secretion is reduced because the microtubule scaffolding required for vesicle motility has been disrupted. Other studies have

71

suggested that a particular motor protein, cytoplasmic dynein, is required for the movement of membrane vesicles involved in secretory vesicle maturation and, possibly, transcytotic vesicle transport, to the subapical cytoplasm. The subapical actin cytoskeleton plays complex roles in lacrimal acinar secretion. With its location immediately beneath the apical plasma membrane in a dense network, it poses an intracellular barrier for vesicle fusion to the apical membrane. For fusion to occur, this actin barrier must be disassembled to allow access of large secretory vesicle to the apical plasma membrane. Recent work has shown in fact that regions of the actin cytoskeleton located immediately beneath the regions of apical plasma membrane do disassemble, but that actin filaments also reassemble and contract around the base of multiple fusing secretory vesicles. The force generated through compression and retraction of actin filaments toward the apical membrane aids in compound fusion and content extrusion from the fusing vesicles. Regulated exocytosis can therefore be further characterized in lacrimal acini into a type known as multivesicular exocytosis. Two specific actin-dependent motors have been implicated so far in this actin remodeling and compound fusion of activated secretory vesicles, a conventional myosin motor known as nonmuscle myosin 2 and an unconventional myosin motor known as myosin 5c, with the possibility that other members of the myosin motor superfamily may also participate in this complex process. Other major membrane trafficking effectors that have been implicated in acinar cell protein secretion include rab proteins. Rabs are major effectors of all intracellular steps of membrane trafficking and fusion in the eukaryotic cell, serving as the molecular address code on donor membrane vesicles which specify the acceptor compartment destination. GTP binding and hydrolysis serves as the on/off switch that activates these proteins. Specific rabs are localized to distinct compartments, conferring identity to these compartments. For instance, rab3D is enriched in secretory vesicles in lacrimal acinar cells and appears to regulate compound fusion of these vesicles. Other data suggest that rab27 isoforms also participate in the maturation and fusion of secretory vesicles during regulated exocytosis in lacrimal acini. By analogy with other systems, rab4 and rab5 isoforms are likely to participate in early events in acinar transcytosis, specifically basolateral endocytosis and sorting, while rab 11 isoforms are enriched in apical endosomes and may facilitate terminal transcytotic traffic of materials destined for the lacrimal fluid. Specific types of SNARE proteins are located on donor and acceptor membranes and interact to form fusion pores which allow membrane contents to mingle or secretory vesicle contents to be extruded to the cell exterior. Several types of SNARE proteins have been demonstrated in lacrimal acini. Regulated exocytosis of secretory vesicles in acini is thought to use both vesicle-associated

72

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Resting

AM

TGN

membrane protein 2 (VAMP 2) and VAMP 8 on donor membranes, and syntaxin 2 and SNAP23 on acceptor membranes. Figure 3 shows the organization of the two major protein secretory pathways that contribute proteins to lacrimal fluid: regulated exocytosis and transcytosis, as well as some of the effectors within each pathway. Morphological analysis of ductal epithelial cells has revealed the presence of large secretory vesicles, presumably containing additional constituents destined for release into the lacrimal fluid. However, due to limitations in the ability to isolate these individual cells and conduct cellular investigations into the membrane trafficking mechanisms, little is known about the precise mechanisms involved in ductal cell exocytosis and transcytosis.

BM

Mechanisms and Regulation of Electrolyte and Water Secretion by the Lacrimal Gland

(a) AM

Stimulated

Acinar Cells

TGN

(b)

BM Actin

Microtubules

Snares

Rabs

Unconventional myosin

Conventional myosin

Microtubule motor Figure 3 Protein secretion in lacrimal acinar cells at the apical membrane. (a) Depicts vesicles participating in the transcytotic pathway from the basolateral membrane (BM) to the apical membrane (AM) as shown in blue vesicles. Depicted as well is the maturation and movement to the AM of secretory vesicles after budding from the trans-Golgi network (TGN) as shown in red vesicles. Initially both transcytosis from the BM and movement from the TGN are reliant on microtubule-based motor proteins. As the vesicles move to the actin-rich subapical region, unconventional myosin motors become important in actin-based movement through the subapical actin network. (b) Depicts multivesicular exocytosis after stimulation with secretogogs such as carbachol. Unconventional myosins such as myosin 5c have been shown to have an important role in the reorganization of actin filaments around clusters of secretory vesicles primed for fusion. Actin and conventional myosins then work together to compress the fusing secretory vesicles and to promote content extrusion. Rab and SNARE

The lacrimal fluid is hypertonic due to a high Cl and Kþ 2þ content, whereas the levels of Naþ, HCO are 3 , and Ca similar to the plasma levels. The electrolyte concentration of the lacrimal fluid is however not static, but varies with the flow rate to become more isotonic with an accelerated flow rate. Fluid secretion is an osmotic process driven by ion movement through the membrane of the acinar cells. Parasympathetic stimulation of the acinar cells triggers an acute increase of cytosolic Ca2þ and adenosine 30 ,50 cyclic monophosphate (cAMP) which opens Cl channels in the apical membrane, resulting in a movement of Cl into the lumen. The increase in cytosolic Ca2þ also activates Kþ channels in the apical as well as the basolateral membrane, causing Kþ to move out of the cell. Naþ follows the flux of Cl and Kþ, moving from the basolateral side toward the lumen traveling through paracellular channels between the cells. To maintain an isotonic secretion, water exits the cell through water-channel proteins called aquaporins. The movement of Cl and Kþ out of the cell is dependent on their electrochemical gradient, that is, the intracellular concentration of these ions must be higher than in the extracellular fluid. This is made possible by ion pumps and co-transporters located in the basolateral membrane. Naþ/Kþ-ATPase transports Kþ into the cell and Naþ out of the cell; the Naþ/Kþ/2Cl co-transporter (NKCC1) moves all three ion types into þ þ the cell; and Cl/HCO 3 and Na /H anti-porters trans þ port Cl and Na into the cell and Hþ and HCO 3 out of the cell. Figure 4 illustrates the ion channels and transporters present in the lacrimal acinar cell. proteins participate in the targeting and fusion events in each pathway. It should be noted that the rabs and the SNAREs participating in transcytosis and exocytosis are different for each pathway.

Lacrimal Gland Overview K+

Cl– Lumen

Ca2+

cAMP

K+ Cl–

Na+ ATP ADP

BM Na K+ 2Cl–

K+ HCO3

Na+

H+

Figure 4 Schematic diagram showing the major ion transporters active in the electrolyte and water release from the lacrimal gland acinar cell. Increases in cytosolic Ca2þ and cAMP following neural stimulation opens Kþ and C channels, resulting in an outward flux of these ions. Naþ/Kþ-ATPase transports Kþ into the cell and Naþ out of the cell, the Naþ/Kþ/2Cl co-transporter moves all three ion types into the cell, and þ þ þ Cl/HCO 3 and Na /H anti-porters transport Cl and Na into þ the cell and H and HCO3 out of the cell.

Ductal Cells Due to difficulties in specifically isolating the ductal cells, the ion transport mechanisms have not been extensively studied. However, it has been hypothesized that the water and ion transport events continue as the lacrimal fluid travels through the ductal system in a pattern comparable to that in the acinar cells. Recently, several approaches to study the ductal cells have been developed, including microdissection and culturing of individual ducts and laser capture microdissection of individual ductal cells. In these studies, some of the most common acid/base transporters were characterized in the ductal cells. The same work also showed that the ion transport in the ductal cells can be regulated by parasympathetic neurotransmitters. Furthermore, studies showed that the lacrimal fluid released from the acinar cells is isotonic, leading to the hypothesis that the ductal cells are responsible for the high Kþ levels in tears. The finding that NaþKþ-ATPase and NKCC1 are expressed at higher levels in the ductal cells than in the acinar cells supports this hypothesis.

Conclusion Previous and ongoing studies have established many of the functions of the principal cells of the lacrimal gland. The major cell type, the acinar cell, is responsible for the regulated release of proteins, fluid, and electrolytes into

73

the lacrimal fluid, while ductal cells appear to further modify the electrolyte composition and likely also contribute additional proteins to the aqueous tear film. Some of the signaling pathways have been elucidated that stimulate the production of lacrimal fluid, while some of the molecular mechanisms involved in the fundamental exocytotic and transcytotic events have likewise been elucidated. However, the complexity of the signaling and membrane trafficking events even in lacrimal acinar cells, the best-studied cell type in this complex organ, means that considerable work remains to be done. Although some insights regarding changes in signaling and membrane trafficking pathways that result in altered production of lacrimal fluid have been obtained in dry eye disorders, considerable additional work is required in order to truly understand the etiology of these disorders. In some studies, changes in tear protein composition have been associated with dry eye disorders, so future challenges also include the identification of tear biomarkers that can be used to diagnose different types of dry eye disorders to aid in determination of the appropriate course of treatment. See also: Dry Eye: An Immune-Based Inflammation; Innate Immune System and the Eye; Lacrimal Gland Hormone Regulation; Lacrimal Gland Signaling: Neural; Meibomian Glands and Lipid Layer.

Further Reading Cohen, A. J., Mercandetti, M., and Brazzo, B. G. (eds.) (2006). The Lacrimal System, Diagnosis, Management and Surgery. New York: Springer. Hodges, R. R. and Dartt, D. A. (2003). Regulatory pathways in lacrimal gland epithelium. International Review of Cytology 231: 129–196. Jerdeva, G., Wu, K., Yarber, F. A., et al. (2005). Actin and non-muscle myosin II facilitate apical exocytosis of tear proteins in rabbit lacrimal acinar epithelial cells. Journal of Cell Science 118: 4797–4812. Marchelletta, R. R., Jacobs, D., Schechter, J. E., Cheney, R., and Hamm-Alvarez, S. F. (2008). Myosin Vc facilitates actin filament remodeling and compound fusion of mature secretory vesicles during exocytosis in lacrimal acini. American Journal of Physiology (Cell Physiology) 295: C13–C28. Pflugfelder, S. C., Beuerman, R. W., and Stern, M. E. (eds.) (2004). Dry Eye and Ocular Surface Disorders. New York: Marcel Dekker, Inc. Ubels, J. L., Hoffman, H. M., Srikanth, S., Resau, J. S., and Webb, C. P. (2006). Gene expression in rat lacrimal gland duct cells collected using laser capture microdissection: Evidence for Kþ secretion by duct cells. Investigative Ophthalmology and Visiual Science 47: 1876–1885. Walcott, B., Moore, L., Birzgalis, A., Claros, N., and Brink, P. R. (2002). A model of fluid secretion by the acinar cells of the mouse lacrimal gland. Advances in Experimental Medicine and Biology 506(Pt. A): 191–197. Wu, K., da Costa, S. R., Jerdeva, G., et al. (2006). Mechanisms of exocytosis in lacrimal gland. Experimental Eye Research 83: 84–96. Zierhut, M., Stern, M. E., and Sullivan, D. A. (eds.) (2005). Immunology of the Lacrimal Gland, Tear Film and Ocular Surface. New York: Taylor and Francis.

Lacrimal Gland Hormone Regulation A K Mircheff, D W Warren, and J E Schechter, University of Southern California, Los Angeles, CA, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary CD86 – A co-receptor expressed on the surfaces of antigen-presenting cells. When engaging either of its cognate receptors – CD28 and CTLA-4 – on the surface of T cells, it generates signals essential for T-cell activation and activates signaling cascades within the antigen-presenting cells. Chemokines – The proteins that promote recruitment of lymphocytes and leukocytes to inflamed tissues and to lymphoid tissues. Hypophysectomy – The surgical removal of the pituitary gland. Interferon gamma (IFN-g) – A cytokine released primarily by T cells and natural killer cells that induces T cells to express the TH1 phenotype, activates macrophage to express microbicidal functions, and induces B cells to switch from immunoglobulin M (IgM) to complement-fixing IgG isotypes. Interleukin 1 alpha and 1 beta (IL-1a, IL-1b) – The related cytokines released primarily from macrophages, endothelial cells, and epithelial cells that induce inflammatory responses. Interleukin 6 (IL-6) – A cytokine that promotes inflammatory responses and supports survival of T and B cells. Interleukin 12a (IL-12a) – A cytokine released in innate immune responses that induces expression of IFN-g and, thereby, promotes the evolution of adaptive responses mediated by TH1 cells. Lactation – The production and secretion of milk. Lactogenesis – The secretory differentiation of the mammary epithelium. Orchiectomy – The surgical removal of the testes. Sex hormone-binding globulin (SHBG) – A protein which binds estrogens and androgens. It is produced by the liver and secreted into the blood. Estrogens stimulate its production and androgens suppress its production. Sodium-potassium-dependent ATPase (Na,K-ATPase) – The sodium–potassium pump enzyme; it generates the chemiosmotic energy essential for lacrimal fluid production by pumping Na+ out of, and K+ into, the cytosol. Transforming growth factor-beta (TGF-b) – A cytokine released by T cells and macrophages, as well as by some epithelial cells and mesenchymal cells. Its actions on immune cells include: inhibiting T-cell proliferation and expression of effector

74

functions; inhibiting B cells from proliferating and inducing them to undergo IgM-to-IgA isotype classswitch recombination; and suppressing macrophage activation. It often exerts antiproliferative or proapoptotic influences on epithelial cells.

Gender-Related Dimorphisms The lacrimal glands produce most of the aqueous fluid that comprises the milieu exte´rieur sustaining the live, metabolically active cells of the superficial layers of the cornea and conjunctiva, and insufficient production of this fluid and alterations of its composition are associated with dry eye disease. Because dry eye disease is considerably more prevalent among women, it has seemed intriguing that morphological differences can be readily discerned between the acini – that is, the primary secretory structures of the lacrimal glands – of male and female rats. The structural dimorphisms were, some years ago, found to be accompanied by equally striking biochemical and functional dimorphisms. Classic work by Sullivan and colleagues showed that many of the evident dimorphisms are supported by the higher levels of androgens characteristic of males. However, it is taking much longer to learn how, in mechanistic terms, the gender-related dimorphisms might relate to females’ greater predilection for lacrimal dysfunction. Indeed, one of the first functional dimorphisms to be documented appeared paradoxical: basal precorneal tear volume is smaller in intact male rats than in females, and it increases in males after they are castrated – a surgical maneuver that causes the size of the acini to decrease, that is, to become more female like. One of the products that the lacrimal glands contribute to the ocular surface fluid is secretory immunoglobulin A (sIgA), which is the effector of the adaptive mucosal immune defense against microbial infection. The lacrimal glands of rats exhibit several readily quantified dimorphisms relating to the production and secretion of dimeric IgA (dIgA). The stromal spaces of the lacrimal glands of male rats are populated by larger numbers of dimeric IgA-secreting plasmacytes. Whole gland extracts contain greater masses of dIgA. Glandular epithelial cells express higher levels of the polymeric Ig receptor (pIgR), which mediates uptake of dimeric dIgA at the stromal-facing surface of the lacrimal epithelium, chaperones it through

Lacrimal Gland Hormone Regulation

the cells’ transcytotic apparatus, and provides the secretory component (SC) portion of secretory IgA (sIgA). Lacrimal gland fluid from sexually mature male rats contains both more sIgA and SC.

Sex Steroids Androgens As noted above, it was established early on that the androgens support the gender-related dimorphisms of acinar size and of precorneal tear volume, which is presumably related to basal rates of lacrimal fluid production. It was also found that the androgens also support epithelial cell expression of pIgR and secretion of SC. However, androgen effects on the numbers of IgAþ-plasmacytes populating the glands’ stromal spaces varied considerably among individual animals. Although different laboratories have reported discrepant findings, there is some evidence for the theoretical paradigm that the androgens exert general trophic influences on the glandular epithelium. Hypophysectomizing rats decreases circulating levels of gonadal and adrenal steroids, as well as pituitary hormones (see Box 1). This maneuver reduces the lacrimal glands’ gross weight and their weight as a fraction of total body weight. In some studies, administration of dihydrotestosterone (DHT) did not change lacrimal gland weight appreciably. In other studies, administration of DHT partially reversed hypophysectomy-induced decreases in the total amounts of protein and Na,K-ATPase catalytic activity measured in lacrimal gland lysates. Ovariectomizing female rabbits decreases serum sex steroid levels. This maneuver decreased the total protein and DNA contents of lacrimal gland lysates. Administration of DHT prevented the ovariectomy-induced decreases, and it increased the Na,K-ATPase catalytic activity measured in lacrimal gland lysates. Surprisingly, in view of the role Na,K-ATPase plays in lacrimal fluid production, DHT did not increase the basal rate of lacrimal fluid production. However, it significantly increased the volume of fluid intact glands produced when they were stimulated with cholinergic agonists. These findings make it clear that several independent layers of regulation determine lacrimal fluid production: long-term regulation of the mass of cells comprising the glandular epithelium and of the levels at which the epithelial cells transcribe the genes specifying Na,K-ATPase and other iontransport proteins, and acute regulation – presumably neurally mediated – of the transport proteins’ functional states. A more complex paradigm, however, is needed to account for why the magnitude of the lacrimal gland regression caused by ovariectomy is small compared to the extent of atrophy that nuclear magnetic resonance (NMR) imaging studies have documented in the lacrimal glands of aging females. One possible paradigm is that in mature, but not aged, female rabbits the lacrimal glands compensate for ovariectomy-induced loss of testosterone

75

by converting the weak androgen, dihydroepiandrosterone (DHEA) – which is produced by the adrenal cortex (see Box 1) – to testosterone and DHT. A second is that the loss of a small trophic influence must be compounded over time before its consequences become evident. A third is that androgens influence other parameters in addition to the cellular mass of the epithelium, and that it is the consequences of loss of those influences that are compounded over time. In ex vivo studies, DHT increased proliferation in models of acinar cells from rabbit lacrimal glands. Compared to the action of epidermal growth factor (EGF), however, the influence of DHT was relatively modest. When Azzarolo and colleagues tested the hypothesis that androgens support the number of cells in the epithelium by preventing apoptosis, as well as by promoting cell proliferation, they found that epithelial cell apoptosis is relatively rare, but plasmacytes in the glands’ stromal space began apoptosing within an hour following ovariectomy. Moreover, administration of DHT prevented ovariectomy-induced plasmacyte apoptosis. The finding that androgen withdrawal leads to a wave of apoptosis in the lacrimal gland plasmacyte population is one of several that accord with the paradigm that their higher levels of androgens protect men both from Sjo¨gren’s autoimmune dacryoadenitis and from the histopathophysiological syndrome commonly found in aging women by influencing immunophysiological processes within the gland. Indeed, it has been found that administration of DHT suppresses lacrimal gland disease in certain mouse models for Sjo¨gren’s syndrome, and preliminary clinical experiences suggest that androgen supplementation of hormone replacement therapy may improve symptoms and clinical signs in menopausal women with inflammatory autoimmune lacrimal diseases. In the models in which androgen administration is therapeutic, the responses are more pronounced in the lacrimal glands than in other affected organs. This finding led Sullivan and coworkers to propose that the androgens control the expression of critical immunoregulatory paracrine mediators by lacrimal gland epithelial cells. Recent microarray analyses indicate that androgens influence the expression of large numbers of gene transcripts in the lacrimal glands and corneas of mice. Some of the androgen actions might be expected to diminish inflammatory processes. For example, testosterone decreases expression of certain chemokines; interferon (IFN) response factors 4 and 7; caspase-1, which converts the inactive interleukin (IL)-1b precursor to active IL-1b; and the prolactin receptor, which – as discussed below – mediates mitogenic responses both in B and T cells and induces Tcells to express IFN-g. However, other testosterone actions might be expected to enhance inflammatory processes. For example, testosterone increases expression of IL-6; IL-12a; the chemokines CCL1, CCL8, CCL28, CCL5, and CXCL4; CD86; and interferon response factor 5.

76

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Box 1.

Regulation of reproductive hormone bioavailability

Figure 1 summarizes factors and feedback interactions that determine the bioavailabilities and local actions of the steroid reproductive hormones. Hypothalamic neurons release gonadotropin-releasing hormone (GNRH), which stimulates release of LH and FSH by the anterior pituitary. LH and FSH, in turn, stimulate synthesis of estradiol and progesterone in the ovaries, and synthesis of testosterone in the testes. Notably, testosterone is an obligatory precursor in the syntheses of estradiol and progesterone, and total serum testosterone levels increase as estradiol and progesterone levels increase. Progesterone can be converted to testosterone, but estradiol cannot. The anterior pituitary also produces adrenal corticotropic hormone (ACTH), which is induced by corticotrophin releasing hormone from hypothalamic neurons. ACTH, in turn, increases adrenal synthesis of the glucocorticoids and dihydroepiandrosterone (DHEA). DHEA can be converted to testosterone in peripheral tissues. Androgen target tissues frequently convert testosterone to dihydrotestosterone, a higher affinity ligand for androgen receptors. The liver plays an indirect but critical role in determining testosterone bioavailability, producing sex hormone binding globulin (SHBG) in response to increasing levels of estradiol and testosterone in the serum. In some cases SHBG sequesters the hormones, reducing their bioavailability. In other cases SHBG-sex steroid hormone complexes exert biological activities by interacting with nonclassical receptors at the surfaces of target cells, rather than with the classical receptors that traffic from the cytosol to the nucleus. After menopause, the ovarian stroma continues to produce some testosterone and androstenedione. During hormone replacement therapy, however, the exogenous estrogens suppress GnRH release, thus reducing ovarian testosterone production. Exogenous estrogens further decrease testosterone bioavailability by stimulating increased hepatic production of SHGB. Production of testosterone declines gradually as men age. Production of DHEA declines similarly in women and men. In women, reproductive steroid hormone and prolactin levels vary systematically during the menstrual cycle. They fluctuate, with more individual variation, during the perimenopause. They then remain consistently low postmenopausally, while prolactin levels are decreased only modestly. Estradiol, progesterone, and prolactin levels increase markedly over the course of a pregnancy. The steroid levels decrease abruptly at parturition; prolactin declines more gradually during a nonlactating puerperium, and it remains elevated during lactation. In contrast to positive regulation of LH, FSH, and ACTH by hypothalamic factors, the release of prolactin by the anterior pituitary is negatively regulated, that is, suppressed, by dopamine produced by hypothalamic neurons. Estradiol increases production of prolactin in the anterior pituitary, while prolactin suppresses estradiol production by acting upstream to suppress release of GNRH. As noted in the text, a number of peripheral tissues in addition to the pituitary produce prolactin.

Hypothalamus

Pituitary

CRH

Dopamine Gonads GNRH

ACTH LH, FSH Prolactin

Adrenal cortex

Testosterone

Cortisol

Gonads

DHEA Progesterone

Testosterone SHBG

Estradiol

DHT On Off

Liver

Figure 1 Production, regulation, and interactions of reproductive hormones. The pituitary produces the protein hormones, LH (luteinizing hormone), FSH (follicle stimulating hormone), ACTH (adrenalcorticotropic hormone), and prolactin. LH and FSH stimulate gonadal production of estradiol, progesterone, and testosterone. ACTH stimulates production of cortisol, as well as dihydroepiandrosterone (DHEA). Estradiol stimulates hepatic production of sex hormone binding globulins, which may either sequester steroids, particularly testosterone, reducing their bioavailability, or potentiate their actions by allowing them to bind to unconventional receptors at the surfaces of target cells. Estradiol also increases pituitary production of prolactin. Both prolactin and estrogen mediate negative feedback signals that decrease LH and FSH production.

Lacrimal Gland Hormone Regulation

Before considering tentative theoretical paradigms that might explain how the androgens might confer protection against dry eye disease, it is necessary to first review the influences other reproductive hormones exert on the lacrimal glands. Estradiol and Progesterone Whereas estradiol and testosterone often have opposing actions, Azzarolo and coworkers found that estradiol – like DHT – prevents ovariectomy-induced apoptosis of lacrimal gland plasmacytes. Estradiol does not appear to influence pIgR expression in rat lacrimal epithelial cells. However, tear lactoperoxidase levels vary during the estrus cycle in rats and the menstrual cycle in humans – correlating with the changes in estradiol and progesterone levels. Microarray analyses of mouse lacrimal gland extracts indicate that estradiol and progesterone influence the expression of numerous gene transcripts. They increase expression of the chemokines CCL2 and CXCL15 and decrease expression of FoxP3 – a central transcription factor in regulatory lymphocyte function; these influences might seem consistent with the greater risk for inflammatory lacrimal gland disease in females. However, estradiol and progesterone also decrease expression of CD86, IL-12, and the chemokines CCL6, CCL12, and CCL28 – influences which might be expected to diminish inflammatory responses.

Prolactin Prolactin is produced by the pituitary gland; as its name implies, its first discovered function was support of lactogenesis and lactation. However, prolactin has also been found to function as an autocrine/intracrine and paracrine mediator in a number of physiological systems. In the immune system, it acts as a mitogenic cytokine for Tcells and B cells, and as a differentiation factor for T cells – inducing them to express the prototypical TH1 cytokine, IFN-g. Administration of prolactin to hypophysectomized male rats increased Na, K-ATPase catalytic activity in lacrimal gland. However, a number of reports indicate that serum prolactin levels are elevated in women with Sjo¨gren’s syndrome and other autoimmune diseases. A study of reproductive hormone influences on lacrimal function revealed that increasing serum prolactin levels within the normal range of values for nonpregnant, nonlactating women were strongly correlated with decreasing lacrimal function – independently of menopausal status and use of estrogen replacement therapy. Exocrine Products and Autocrine/Intracrine and Paracrine Mediators While they respond to prolactin as a classic hormone, lacrimal gland epithelial cells also express prolactin, and

77

they secrete it both as an exocrine secretory product and as a paracrine mediator. In the lacrimal glands of normal, nonpregnant female rabbits, immunopositivities for prolactin, as well as for transforming growth factor-beta (TGF-b), EGF, fibroblast growth factor (FGF)-2, are localized preferentially – but not exclusively – in ductal epithelial cells. In both acinar and ductal cells, the cytokine and growth factor immunopositivities are concentrated in the apical cytoplasm. Prolactin is localized in the regulated exocrine secretory vesicles; the mechanisms by which epithelial cells of the rabbit lacrimal gland secrete the other cytokines and growth factors have not been elucidated. Like prolactin and other secretory vesiclecontent proteins, TGF-b is released to the fluid forming within the lumen of the acinus-duct system in response to stimulation with cholinergic agonists. As illustrated in Figure 1, lacrimal epithelial cells use their apical recycling endosome and early basolateral endosome as a transcytotic secretory apparatus. It is this apparatus which secretes sIgA and some free SC into the lumena of the acinus-duct system. They also use the early and recycling endosomes as a paracrine secretory apparatus that delivers products to the underlying stromal space. Both transcytotic secretion and paracrine secretion occur constitutively; although they can be accelerated by stimulation with cholinergic agonists, the steady-state pools of secreted products in the endosomes are quite small compared to the pools of products stored in regulated exocrine secretory vesicles. Experiments with ex vivo acinar cell models showed that increasing the concentration of prolactin in the ambient medium induces increased transcription of prolactin messenger RNA (mRNA). Increasing epithelial cell prolactin expression or increasing the prolactin concentration in the ambient medium decreased the amount of secretory proteins stored in apical secretory vesicles, and it induced the cells to express a novel population of regulated secretory vesicles that accumulated in the basal cytoplasm and released their contents at the basolateral plasma membrane in response to acute cholinergic stimulation. Immunogold localization studies demonstrated that when acinar cells endocytose prolactin from their ambient medium, they traffic it dually to the endosomes that comprise their constitutive transcytotic-paracrine apparatus and to the secretory vesicles of the novel, induced paracrine apparatus. Thus, when acinar epithelial cells internalize prolactin secreted from the pituitary or from ductal epithelial cells, they may recycle it as a paracrine mediator. Serum prolactin levels do not differ greatly between normal men and normal, nonpregnant women. However, serum prolactin levels increase markedly during pregnancy, and, by the time a pregnancy reaches term, mean serum prolactin levels are 10- to 20-fold greater than the levels in nonpregnant, nonlactating females. Thus, the physiological hyperprolactinemia of pregnancy has

78

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Early endosome

Recycling endosome Multivesicular body

Stroma

Golgi ER Lumen Regulated paracrine secretory vesicles Late endosome

TGN Exocrine secretory vesicles

Storage lysosome Pre-lysosome Autolysosome

Prolactin

Constitutive traffic

Prolactin receptor

Induced by prolactin

Maturation

Endocytosed prolactin

Cholinergically induced exocytosis

Locally expressed prolactin

Figure 1 As in other exocrine glands, lacrimal gland epithelial cells use secretory vesicles to secrete proteins into the fluid being produced in the lumena of the acinus-duct system. They also use their early basolateral endosomes and apical recycling endosomes as a transcytotic apparatus to secrete SC and sIgA. Furthermore, the constitutive traffic of transport vesicles from the endosomes to the basolateral plasma membranes secreted paracrine mediators to underlying stromal space. Elevated levels of prolactin induce the cells to express a novel population of regulated paracrine secretory vesicles and decrease their population of exocrine secretory vesicles. The induced paracrine secretory vesicles allow ductal epithelial cells to secrete more prolactin and TGF-b to the stroma, and they allow acinar cells that have endocytosed prolactin from the stroma to recycle it as a paracrine secretory product.

seemed to offer a natural model in which to study prolactin’s influences on the lacrimal glands.

Influences of Prolactin, Estradiol, and Progesterone during Pregnancy The lacrimal glands of nonpregnant, sexually mature female rabbits normally contain small aggregates of

lymphocytes and plasmacytes, localized in the stromal spaces surrounding and spanning between venules and interlobular ducts. It might be noted that these are the same sites where the ectopic lymphoid tissues characteristic of Sjo¨gren’s dacryoadenitis develop. The immunoarchitecture undergoes a remarkable change during pregnancy, and it remains in the altered state throughout lactation and for some weeks following weaning. By the time a pregnancy reaches term, the aggregates have largely dissipated, and

Lacrimal Gland Hormone Regulation

lymphocytes and plasmacytes are primarily located in the thin stromal spaces surrounding acini. The immunoarchitectural change is associated with several notable cytophysiological changes and functional changes. The basal rate of lacrimal gland fluid production decreases, while the rate of fluid production under cholinergic stimulation increases. Immunopositivities for TGF-b and prolactin increase substantially, and their localizations shift from the apical cytoplasm to the basal cytoplasm. It now appears that these redistributions occur because the novel paracrine secretory apparatus induced by the increased serum prolactin level captures prolactin and TGF-b away from the regulated exocrine secretory pathway. Accordingly, the level of prolactin excreted in lacrimal gland fluid decreases, and TGF-b becomes scarcely detectable. Experiments that have not yet been published indicate that when ovariectomized rabbits are implanted with sustained-release pellets establishing pregnancy-like serum levels of estradiol and progesterone, the patterns of lymphocyte organization and of TGF-b and prolactin expression and localization change to resemble the patterns characteristic of pregnancy. Thus, estradiol, progesterone, or the two steroid hormones in concert act on ductal epithelial cells to increase their expression TGF-b. They may increase ductal epithelial cell prolactin expression either directly, or indirectly, that is, by increasing pituitary prolactin secretion (see Box 1). The increased level of prolactin then directs both mediators away from the regulated exocrine protein-secretion apparatus and directs prolactin into the novel paracrine apparatus. As discussed below, the changes that occur in the lacrimal glands during pregnancy are analogous to those which occur over roughly the same time in the mammary glands. While the lacrimal glands are accessory organ of the visual system, the mammary glands are accessory organs of the reproductive system. Both glands also are effector organs of the mucosal immune system. Other organs of the male and female reproductive systems also play parallel roles as adaptive mucosal immune system effector organs, and their immunophysiological functions are, likewise, influenced by the reproductive hormones.

Reproductive Hormone Influences on Other Mucosal Immune System Tissues Neither the testes nor the ovaries are normally populated by IgA+-plasmacytes. However, in males, IgA+ cells are abundant in the urethral glands and prostate; they are also present in the seminal vesicles in some species, but not others. Orchiectomizing male rats has little effect on the amount of IgA in the prostate and seminal

79

vesicles; subsequent administration of DHT causes a slight increase of the IgA content of the prostate, but not the seminal vesicle. There is little pIgR expression in the testes, vas deferens, or epididymis, but a significant level of expression in the seminal vesicles and a 20-fold greater level in the prostate. Orchiectomy decreases pIgR expression, and subsequent administration of DHT increases it threefold in the seminal vesicles and fourfold in the prostate. Interestingly, estradiol has no effect on pIgR expression in the seminal vesicles but doubles it in the prostate. In females, IgA+ immune cells are abundant in the lamina propria of the fallopian tubes. They are sparse in the endometrium. Since endometrial gland epithelial cells contain IgM+ and IgA+, as well as J chain, it may be that the uterine lining secretes Igs derived primarily from the serum, rather than from local plasmacytes. In contrast, IgA+ cells are abundant within the epithelia and lamina propria of the endocervix, although somewhat less abundant in the ectocervix and vagina. Epithelial expression of pIgR roughly parallels the abundance of IgA+ cells. The level is significant in the fallopian tubes and endocervix. While there seems to be no clear evidence that pIgR is expressed in the ectocervix and vagina, the level of sIgA in cervical mucus increases just prior to ovulation and remains elevated throughout the luteal phase, and the uterine fluid contains a high level of sIgA throughout pregnancy. As the alveolar epithelium of the mammary glands develops during pregnancy, ductal epithelial cells are induced to express pIgR, and the glands’ stromal spaces become populated by dIgA+-plasmacytes. During lactation, the mammary epithelium secretes sIgA as well as lactoperoxidase and other innate mucosal immune effector molecules. Like the other changes occurring during lactogenesis, the induction of mucosal immune effector functions is controlled by interacting influences of estradiol, progesterone, and prolactin. Most studies of hormonal influences on the mammary glands have been motivated by interest in normal lactogenesis, lactation, and postweaning involution, and in mammary carcinoma – rather than focused on the mammary glands’ mucosal immune functions. This work has shown that the systemic hormones orchestrate lactogenesis and lactation, in part, by regulating the expression of autocrine/intracrine and paracrine mediators. The ductal network of the mammary gland develops during puberty, largely under the influence of estradiol. Prior to pregnancy, TGF-b – which is expressed both by ductal epithelial cells and periductal mesenchymal cells – exerts pro-apoptotic and antiproliferative influences that prevent development of the alveoli – and, during pregnancy, alveolar development depends on the concerted influences of prolactin and progesterone. Increasing progesterone levels increase the abundance of TGF-b, but

80

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

they also increase expression of EGF, FGF-2, and TGFalpha (TGF-a) – which abrogate TGF-b’s anti proliferative and pro-apoptotic influences. Notably, the expression of prolactin by mammary epithelial cells also increases at this time. Despite its evident synergy with prolactin in promoting lactogenesis, the elevated level of progesterone inhibits lactation. Recent evidence suggests it does so by increasing expression of Wnt-5b, which is thought to maintain the undifferentiated state by promoting nuclear translocation of b-catenin, and by increasing expression of insulin-like growth factor-binding protein (IGFBP-5) – which suppresses insulin-like growth factor signaling. These inhibitory influences are removed and lactation becomes possible at parturition, when production of progesterone is abruptly suppressed. Certain of the estradiol-, progesterone-, and prolactininduced mediators that determine development of the mammary epithelium also determine expression of the mammary glands’ mucosal immune functions. As reviewed elsewhere in the encyclopedia, TGF-b typically acts as a differentiation factor for immature dIgA+-expressing plasmablasts, inducing them to mature into dIgA-secreting plasmacytes as they arrive at mucosal immune effector sites. Normal plasmacytes – like other bone-marrowderived cells – express an intrinsic apoptotic program, and their ongoing survival requires that this program be abrogated by survival signals from the local milieu. As noted above, prolactin plays a role in inducing ductal epithelial cells to express pIgR. Its known mitogenic influences on T cells and B cells suggest that it may also act as one of the factors which support the mature plasmacytes’ survival. Thus, prolactin may contribute to a counterpoise against plasmacyte’s apoptotic program as well as against the pro-apoptotic and antiproliferative influences TGF-b would otherwise exert on both the plasmacytes and the alveolar epithelium. Evidently, this counterpoise is maintained as the levels of TGF-b, prolactin, and other estrogen- and progesterone-dependent factors increase during pregnancy, and it supports expansion of the population of plasmacytes that will produce dIgA for secretion in the milk. Similar interactions between the sex steroids and prolactin may account for the reproductive hormones’ influences on the other mucosal effector organs of the female and male reproductive systems, and the data available so far indicate that they do so for the lacrimal glands, as well.

Counterpoises between Contradictory Signals Figure 2 summarizes the spatial and temporal actions of the differentiation and survival factors, as they are

organized in the lacrimal glands. The notions that plasmacytes express an intrinsic apoptotic program which must constantly be abrogated and that the steady-state pools of paracrine secretory products in lacrimal epithelial cells are small may help explain why they begin undergoing apoptosis so abruptly after ovariectomy. The pools of paracrine secretory products would deplete rapidly after the signals that support their ongoing expression are removed. Unpublished findings that prolactin immunoreactivity is present in the nuclei of plasmacytes seem to confirm that prolactin is one of the lacrimal epithelial paracrine mediators the influence plasmacytes. This may only be part of the explanation; however, and it is possible that testosterone, estradiol, and, perhaps, progesterone as well also might interact with prolactin and other survival factors to generate synergistic signals that maintain plasmacyte survival. While prolactin may be one of several factors that provide mitogenic signals abrogating TGF-b’s pro-apoptotic and antiproliferative influences, it appears that TGF-b may provide a counterpoise to prolactin’s lymphoproliferative and proinflammatory influences. As reviewed elsewhere in the encyclopedia, there is evidence that the transcytotic apparatus mucosal epithelial cells use to internalize dIgA and release sIgA into the fluid they produce inevitably secretes a significant burden of autoantigens to the underlying stromal spaces. Thus, newly matured dendritic cells that emigrate from the lacrimal glands to the draining lymph nodes carry with them lacrimal epithelial autoantigens. They process the autoantigens to generate epitopes that their surface MHC class-II molecules will present to CD4+ T cells, and they also release the autoantigens for sampling by IgM+-B-cell antigen receptors. There is now evidence that TGF-b induces immature dendritic cells to mature into immunosuppressive antigen-presenting cells that prevent proliferation of autoreactive lymphocytes within the lacrimal glands and draining lymph node. Moreover, it is possible that dendritic cells that have matured within the lacrimal glands might also function as tolerogenic antigen-presenting cells – inducing the generation of TH3 or TR1 regulatory cells. Microarray studies have clarified that the reproductive hormones influence the expression of other cytokines and growth factors apart from TGF-b and prolactin in the lacrimal glands. Further work will be needed to determine the extents to which the various mediators are expressed by infiltrating immune cells, acinar and ductal epithelial cells, and mesenchymal cells. Other factors in addition to the reproductive hormones are likely to influence epithelial expression of TGF-b and of prolactin and the plasmacyte survival factors. Nevertheless, the concept that the reproductive hormones orchestrate counterpoises between contradictory signals, summarized in Figure 3, may lead to detailed paradigms that explain why Sjo¨gren’s dacryoadenitis and the common histopathological syndrome

Lacrimal Gland Hormone Regulation

81

Acinus

Plasmacyte

Interlobular duct

Mature dendritic cell

Plasmablast

Immature dendritic cell

Afferent lymph vessels Venule Dimeric IgA

Autoantigen

Survival factor

Secretory IgA

Maturation

Differentiation factor

Secretory component (SC)

Polymeric IgA receptor (pIgR)

Figure 2 Interlobular duct epithelial cells produce both differentiation factors, such as TGF-b and IL-10, which may also exert proapoptotic and immunoregulatory influences. They also produce survival factors, such as EGF, FGF, and prolactin, some of which also may exert mitogenic and proinflammatory influences. The counterpoises of contradictory influences maintain for the lacrimal glands’ exocrine functions as well as their mucosal immune functions. For example, TGF-b may support expression of the ductal epithelial phenotype, while FGF, EGF, and prolactin may support both survival of the ductal epithelium and development and survival of the acini. TGF-b induces plasmablasts to undergo terminal differentiation to dIgA-secreting plasmacytes, while prolactin may be one of several mediators that support the plasmacytes’ ongoing survival. TGF-b also induces immature dendritic cells that have taken up lacrimal autoantigens to differentiate as regulatory antigen presenting cells.

are both so much more prevalent among women, and why the onset of clinical dry eye disease is associated with events of the reproductive cycle and life cycle. There are several physiological states during which prolactin-mediated influences might become excessive with respect to the available counterpoises (see Box 1). Given the burden of epithelial autoantigens constitutively present in the stromal space of the lacrimal glands, one might predict that such states favor the accumulation of autoreactive T cells and B cells. In a direct experimental test of the hypothesis, an adenovirus vector was used to transiently increase prolactin expression in lacrimal glands of mature female rabbits. As has been reported in preliminary form, increased abundance of prolactin transcripts was accompanied by increased

abundances of mRNAs for IFN-g and TGF-a, as well as accumulation of large lymphocytic infiltrates and apparent formation of germinal centers. Moreover, the lymphocytic foci persisted for weeks after the prolactin mRNA levels returned to normal. Of interest also is a recent report that, after having a primary relative with an autoimmune disease, carrying a pregnancy to term is the second greatest risk factor for Sjo¨gren’s syndrome appears to accord with this prediction. Both findings suggest that autoimmune activation can be suppressed after systemic hormone levels have returned to normal, but that autoreactive memory cells may persist and become activated as the age-related loss of reproductive steroids changes the immunoregulatory signaling milieu within the lacrimal glands.

82

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Immunoregulation epithelial survival plasmacyte survival plasmablast differentiation

Atrophy

Inflammation

TGF-β, IL-10

Prolactin, EGF, FGF

Differentiation factors

Survival factors

Figure 3 The capacity of the lacrimal gland to perform its exocrine functions – that is, secretion of electrolytes, water, and proteins – and its mucosal immune functions – that is, maintenance of a population of mature, dIgA-secreting plasmablasts and transcytotic delivery of sIgA into the fluid being produced, while avoiding autoimmune inflammatory processes – depends on a system of counterpoises between contradictory influences. The reproductive hormones influence expression of many of the paracrine mediators that exert those influences.

See also: Adaptive Immune System and the Eye: Mucosal Immunity.

Further Reading Ariga, H., Edwards, J., and Sullivan, D. A. (1989). Androgen control of autoimmune expression in lacrimal glands of MRL/Mp-lpr/lpr mice. Clinical Immunology and Immunopathology 53: 499–508. Azzarolo, A. M., Eihausen, H., and Schechter, J. (2003). Estrogen prevention of lacrimal gland cell death and lymphocytic infiltration. Experimental Eye Research 77: 347–354.

Azzarolo, A. M., Mircheff, A. K., Kaswan, R. L., et al. (1997). Androgen support of lacrimal gland function. Endocrine 6: 39–45. Azzarolo, A. M., Wood, R. L., Mircheff, A. K., et al. (1999). Androgen influence on lacrimal gland apoptosis, necrosis and lymphocytic infiltration. Investigative Ophthalmology and Visual Science 40: 523–526. Bailey, J. P., Nieport, K. M., Herbst, M. P., et al. (2004). Prolactin and transforming growth factor-b signaling exert opposing effects on mammary gland morphogenesis, involution, and the Aky-forkhead pathway. Molecular Endocrinology 19: 1171–1184. Ding, C., Chang, N., Fong, Y. C., et al. (2006). Interacting influences of pregnancy and corneal injury on rabbit lacrimal gland immunoarchitecture and function. Investigative Ophthalmology and Visual Science 47: 1368–1375. Frey, W. H., Nelson, J. D., Frick, M. L., and Elde, R. P. (1986). Prolactin immunoreactivity in human tears and lacrimal gland: Possible implications for tear production. In: Holly, F. J. (ed.) The Preocular Tear Film in Health, Disease, and Contact Lens Wear, pp. 798–807. Lubbock, TX: Dry Eye Institute. Kolek, O., Gajowska, B., Godlewski, M. M., and Motyl, T. (2003). Antiproliferative and apoptotic effect of TGF-b1 in bovine mammary epithelial BME-UV1 cells. Comparative Biochemistry and Physiology C 134: 417–430. Mathers, W. D., Stovall, D., Lane, J. A., Zimmerman, M. B., and Johnson, S. (1998). Menopause and tear function: The influence of prolactin and sex hormones on human tear production. Cornea 17: 353–358. Mircheff, A. K., Wang, Y., de Saint Jean, M., et al. (2005). Mucosal immunity and self-tolerance in the ocular surface system. Ocular Surface 4: 182–193. Priori, R., Medda, E., Conti, F., et al. (2007). Risk factors for Sjo¨gren’s syndrome. Clinical and Experimental Rheumatology 25: 378–384. Richards, S. M., Liu, M., Jensen, R. V., et al. (2005). Androgen regulation of gene expression in the mouse lacrimal gland. Journal of Steroid Biochemistry and Molecular Biology 96: 401–413. Rosfjord, E. C. and Dickson, R. B. (1999). Growth factors, apoptosis, and survival of mammary epithelial cells. Journal of Mammary Gland Biology and Neoplasia 4: 229–237. Rudolph, M. C., McManaman, J. L., Hunter, L., Phang, T., and Neville, M. C. (2003). Functional development of the mammary gland: Use of expression profiling and trajectory clustering to reveal changes in gene expression during pregnancy, lactation, and involution. Journal of Mammary Gland Biology and Neoplasia 8: 287–307. Schechter, J., Carey, J., Wallace, M., and Wood, R. (2000). Distributions of growth factors and immune cells are altered in the lacrimal gland during pregnancy and lactation. Experimental Eye Research 71: 129–142. Sullivan, D. A., Kelleher, R. S., Vaerman, J. -P., and Hann, L. E. (1990). Androgen regulation of secretory component synthesis by lacrimal gland acinar cells in vitro. Journal of Immunology 145: 4238–4244. Suzuki, T., Schirra, F., Richards, S. M., et al. (2006). Estrogen’s and progesterone’s impact on gene expression in the mouse lacrimal gland. Investigative Ophthalmology and Visual Science 47: 158–168. Wang, Y., Chiu, C. T., Nakamura, T., et al. (2007). Traffic of endogenous, over-expressed, and endocytosed prolactin in rabbit lacrimal acinar cells. Experimental Eye Research 85: 749–761.

Lacrimal Gland Signaling: Neural D Zoukhri, Tufts University, Boston, MA, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Acinar cells – Highly polarized epithelial cells that form an acinus and whose primary function is to secrete proteins, electrolytes, and water. Exocytosis – The process in which molecules (such as secretory proteins) in a membrane-enclosed vesicle (secretory vesicle or granule) fuse with the plasma membrane and are then released outside the cell. Muscarinic receptors – A subtype of receptors for the neurotransmitter acetylcholine that is more responsive to muscarine than nicotine. Neurotransmitters – Chemicals released by neurons to modulate the function of a target cell. Preocular tear film – A complex and highly structured moist film which covers the bulbar and palpebral conjunctiva, and the cornea. It is composed of water, electrolytes, proteins, mucins, and lipids. Signal transduction – The biochemical events that conduct the signal of an external stimulus from the cell exterior, through the cell membrane, and into the cytoplasm.

Anatomy of the Lacrimal Gland The lacrimal gland is a compound tubuloalveolar serous gland composed primarily of acinar, ductal, and myoepithelial cells (Figure 1). Acinar cells account for over 80% of the cell type present in the lacrimal gland and form the site for synthesis, storage, and secretion of proteins. Several of these proteins have antibacterial or growth factor properties and are crucial to the health of the ocular surface. Acinar cells are highly polarized cells with tight junctions surrounding each acinar cell on the luminal side and thus separating the plasma membrane into apical (luminal) and basolateral (serosal) components. The basal portion of the cell contains a large nucleus, rough endoplasmic reticulum, mitochondria, and Golgi apparatus, while the apical portion is filled with secretory granules. Like the acinar cells, the duct cells are also polarized with the nuclei located basolaterally, whereas the rough endoplasmic reticulum and mitochondria are more apical. The primary function of the ductal cells is to modify the

primary fluid secreted by the acinar cells by absorbing or secreting water and electrolytes. The duct cells secrete a KCl-rich solution so that the final secreted lacrimal gland fluid is rich in K+. It has been estimated that as much as 30% of the volume of the final lacrimal gland fluid is secreted by the duct cells. The myoepithelial cells lie scattered between the acinar and ducts cells and the basement membrane and are interconnected by gap junctions and desmosomes. These cells are highly branched and contain multiple processes which surround the basal area of the acinar cells (Figure 1). The myoepithelial cells are thought to contract because they contain muscle contractile proteins (a-smooth muscle actin, myosin, and tropomyosin). The contraction of these cells would help expel the fluid out of the acini and the ducts onto the ocular surface. In support of a functional role of lacrimal gland myoepithelial cells, receptors and intracellular signaling molecules for parasympathetic neurotransmitters have been described. The lacrimal gland contains other cells: plasma cells, B and T cells, dendritic cells, macrophages, and mast cells. Immunoglobulin A (IgA)-positive plasma cells account for the majority of the mononuclear cells in the lacrimal gland. These cells synthesize and secrete IgA, which then is transported into acinar and ductal cells and secreted by these epithelial cells as secretory IgA.

Neural Control of Lacrimal Gland Secretion To ensure adequate production of the aqueous component of the preocular tear film, lacrimal gland secretion is under tight neural control. To this end, the lacrimal gland is densely innervated by the parasympathetic and sympathetic nervous system (Figure 2). Although scarce, sensory nerves are also present in the lacrimal gland (Figure 2). Nerves are located in close proximity to acinar, ductal, and myoepithelial cells, as well as blood vessels, and hence can control a wide variety of lacrimal gland functions. While each individual cell may not be innervated, gap junctions electrically and chemically connect cells within an acinus so that even noninnervated cells can respond to the neural stimulus. In the lacrimal gland, parasympathetic nerves contain the neurotransmitters acetylcholine and vasoactive intestinal peptide (VIP). Sympathetic nerves contain the neurotransmitters norepinephrine and neuropeptide Y (NPY).

83

84

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Sensory nerves contain the neurotransmitters substance P and calcitonin gene-related peptide (CGRP). Acetylcholine and VIP are potent stimuli of lacrimal gland protein and electrolyte/water secretion. Norepinephrine is also a potent stimulus of protein secretion, but a weak stimulus of electrolyte/water secretion. In contrast, NPY and CGRP are weak stimuli of protein secretion, while substance P does not appear to stimulate either protein or electrolyte/water secretion. The stimulation of lacrimal gland secretion occurs through a neural reflex arc originating from the ocular surface (Figure 3). Neural reflexes are initiated by stimulation of the afferent sensory nerves of the cornea and conjunctiva or by activation of the optic nerve in response to intense light. Efferent parasympathetic and sympathetic nerves of the lacrimal gland are then activated to release their neurotransmitters (Figure 3). The neurotransmitters interact with and activate specific receptors located on the basolateral membranes of acinar and duct cells, which then

initiates a cascade of intracellular events known as signal transduction. Activation of these signal transduction pathways induces fusion of the preformed secretory granules with the apical membrane to release secretory proteins into the lumen (Figure 3). To trigger electrolyte and water secretion, ion channels and pumps, located in the apical and basolateral membranes, are also activated.

Signal Transduction Pathways Activated in the Lacrimal Gland Signal transduction proceeds in three steps: (1) initiation of the signal by interaction of the ligand (neurotransmitter, neuropeptide, or hormone) with its receptor; (2) amplification of the signal through the interaction of the receptor/G protein/effector enzyme leading to the generation of second-messenger molecules; and (3) termination of the signal through the action of protein phosphatases and membrane pumps to bring the amount of phosphorylated proteins and ions, respectively, back to resting levels (Figure 4). Cholinergic Agonist-Activated Signal Transduction Pathways Acetylcholine, released from parasympathetic nerves, activates muscarinic receptors on the basolateral membrane of lacrimal gland cells. Of the five receptor subtypes (M1–5) identified, only the M3 or glandular subtype is present in the lacrimal gland. These receptors are coupled to the activation of phospholipases C and D (PLC and PLD, respectively) and the activation of the p42/p44 mitogen-activated protein kinase (p42/p44 MAPK, also known as extracellular signal-regulated kinase (ERK)) pathway (Figure 5).

Figure 1 Schematic of the lacrimal gland and photomicrographs showing the three major cell types that it is composed of. The acinar cells, which account for 80% of the cell type present in the lacrimal gland, and ductal cells were stained with hematoxylin and eosin. The myoepithelial cells were identified immunohistochemically (brown stain) using an antibody against a-smooth muscle actin.

PLC-coupled signaling pathway Lacrimal gland M3 receptors are coupled, via the Gprotein Gaq, to the effector enzyme PLC. Activated PLC hydrolyzes the plasma membrane lipid, phosphatidylinositol 4,5-bisphosphate (PIP2), to generate two

Figure 2 Photomicrographs depicting the innervation of murine lacrimal gland. Nerves were visualized using antibodies against the following neurotransmitters or enzymes: VIP for the parasympathetic nerves, dopamine b-hydroxylase for the sympathetic nerves, and CGRP for the sensory nerves.

Lacrimal Gland Signaling: Neural

85

Parasympathetic Sympathetic

Parasympath

+ +

+

S ym

+

Acinus

Brain etic

pat het ic Lacrimal gland

Secretion Neurotransmitters Secretory granules

Optic nerve

Tears

Secreted proteins

Cornea Sensory nerves Figure 3 Schematic of the neural reflex arc that controls lacrimal gland secretion. Neural reflexes are initiated by the stimulation of the afferent sensory nerves of the cornea and conjunctiva or by activation of the optic nerve. Efferent parasympathetic and sympathetic nerves of the lacrimal gland are then activated to release their neurotransmitters. The neurotransmitters activate specific receptors located on the basolateral membranes of acinar cells to stimulate secretion.

Nerve ending

Neurotransmitter (a)

Signal initiation Ion channel

Receptor G Signal amplification

Effector enzyme

(c) Second messengers

(b)

Inactive substrate

Active substrate

3

Signal termination

Protein phosphatases Figure 4 Schematic depicting the three steps involved in signal transduction in response to a neural stimulus: (a) signal initiation by interaction of the neurotransmitter with its receptor; (b) signal amplification through the interaction of the receptor with the G protein and effector enzyme to generate second-messenger molecules; and (c) signal termination through activation of protein phosphatases and membrane pumps.

second-messenger molecules, inositol 1,4,5-trisphosphate (IP3) and diacylglycerol (DAG; Figure 5). IP3, a water-soluble molecule, diffuses to the endoplasmic reticulum where Ca2+ is stored in an inactive, bound form. It also interacts with specific receptors located on the endoplasmic reticulum to release Ca2+ into the cytosol. Depletion of these Ca2+ stores leads to an increase in the influx of extracellular Ca2+ across the plasma membrane. The IP3 receptor is a homotetramer of 310 kDa each and constitutes one of the largest of all known ion channels. Binding sites for IP3 are located within the N-terminal

domain, whereas the C-terminal regions form the intrinsic Ca2+ channel. Multiple isoforms of IP3 receptor have been cloned. They share significant similarity to each other and are encoded by at least four genes. The activation of lacrimal gland cholinergic M3 receptors triggers a biphasic Ca2+ response: a rapid (usually referred to as peak) increase in [Ca2+]i due to IP3-induced release of Ca2+ from intracellular stores, followed by a slow and sustained (usually referred to as plateau) increase in [Ca2+]i due to influx of Ca2+ from the extracellular milieu. Both Ca2+ responses are necessary for cholinergic

86

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Parasympathetic nerves

Acetylcholine M3 Receptor

Ras Raf MEK

Gαq

Pyk2

DAG

PLC

PKC α, δ, ε

Src

Ca2+

IP3

PLD ER

ERK Ca2+ –

+

+ +

Exocytosis Figure 5 Schematic depicting the main signal transduction pathways activated by acetylcholine to stimulate protein secretion. The activation of lacrimal gland cholinergic receptors activates three main signaling pathways that either enhance (PKC, Ca2+) or attenuate (ERK) protein secretion. The net protein secretory output in response to acetylcholine stimulation will likely depend on the relative contribution of the stimulatory versus the inhibitory signal transduction pathways.

agonist-stimulated protein secretion since chelation of either the intracellular or the extracellular Ca2+ leads to complete inhibition of secretion. Ca2+ can stimulate secretion alone or do so by activating Ca2+ and calmodulindependent protein kinases to phosphorylate-specific substrates to cause secretion. The DAG formed from the hydrolysis of PIP2 activates protein kinase C (PKC, Figure 5). PKC is a family of closely related isozymes that has been divided into three categories based on structural and functional criteria. A first group, termed conventional PKCs (cPKCs), includes PKCa, -bI, -bII, and -g isoforms, which have a Ca2+- and DAG-dependent kinase activity. A second group, termed novel PKCs (nPKCs), includes PKCE, -d, -y, and -Z isoforms, which are Ca2+-independent and DAG-stimulated kinases. A third group, termed atypical PKCs (aPKCs), includes PKCz and -i/l isoforms, which are Ca2+- and DAG-independent kinases. Four isoforms of PKC are expressed in the rat lacrimal gland: one classical, PKCa; two novel, PKCd, and -E; and one atypical, PKCi/l. In an attempt to define the role that individual PKC isoforms might play in regulating lacrimal gland functions in response to cholinergic stimulation, isoform-specific peptide inhibitors of PKC were synthesized. These peptides were derived from the unique pseudosubstrate sequences of PKCa, -d, and -E, and were myristoylated at their N-terminus to make them cell permeant. Indeed, all PKC isoforms have a pseudosubstrate sequence in their N-terminal part, which interacts with the

catalytic domain to keep the enzyme inactive in resting cells. Using these peptides, it was shown that cholinergic agonists activate PKCa and -E to a larger extent and PKCd to a lesser extent, to induce protein secretion. It was also shown that PKCd and -E, but not -a, negatively modulate cholinergic-induced Ca2+ elevation in the lacrimal gland.

PLD-coupled signaling pathway PLD catalyzes the hydrolysis of membrane phospholipids (preferably phosphatidylcholine), producing phosphatidic acid and the free polar head group. Phosphatidic acid, by itself or after its conversion to DAG by a phosphohydrolase, is an important second-messenger molecule. Besides its hydrolytic activity, PLD possesses the unique ability to catalyze a transphosphatidylation reaction, in the presence of a primary alcohol, in which the phosphatidyl moiety of the phospholipid substrate is transferred to the primary alcohol producing the corresponding phosphatidylalcohol. Accumulation of such unique transphosphatidylation products has been used to detect PLD activity unambiguously in diverse cell types. Depending on the cell’s type, the receptor activation of PLD was shown to occur through mechanisms involving PKC activity, Ca2+, G proteins, or receptor-linked tyrosine kinases. Since PKC activation and Ca2+ mobilization are downstream to PLC stimulation, it has been suggested that PLD activation may be secondary to receptor activation of PLC.

Lacrimal Gland Signaling: Neural

Taking advantage of the transphosphatidyl reaction catalyzed by PLD, it was shown that the lacrimal gland contains a PLD activity. Cholinergic agonists, through the muscarinic receptor, stimulate both the hydrolytic activity of PLD to produce phopshatidic acid, as well as the transphosphatidyl reaction. However, if either Ca2+ is mobilized or PKC is activated, only the transphosphatidyl reaction is stimulated. This finding implied that cholinergic agonist activation of PLD in the lacrimal gland is not secondary to the activation of PLC by these agonists. MAPK-coupled signaling pathway

MAPK, also known as ERK, is a dual serine/threonine and tyrosine protein kinase. It is activated through phosphorylation by an MAPK kinase (known as MEK). MEK is also activated through phosphorylation by its upstream MAPK kinase kinase (known as Raf). Raf is activated when the small GTP-binding protein, Ras, is in its GTP-bound form. Depending on the cell’s type, Ras can be activated by several mechanisms, including the nonreceptor tyrosine kinases Pyk2 and Src, growth factors receptors, and PKC. In the lacrimal gland, the activation of MAPK attenuates protein secretion. The activation of MAPK by the M3 receptor was shown to involve the nonreceptor tyrosine kinases Pyk2 and Src, which in turn activate Ras and, ultimately, MAPK (Figure 5). Recent evidence showed that the activation of MAPK by cholinergic agonists is downstream of PLD activation. The mechanisms involved in PLD-mediated activation of MAPK in the lacrimal gland remain to be elucidated. The termination of the cholinergic signaling pathway involves receptor desensitization, activation of protein

phosphatases to dephosphorylate ERK and other phosphorylated substrates, and the activation of ion channels/ pumps to return the concentration of cytosolic Ca2+ to its resting levels. In summary, the activation of lacrimal gland cholinergic receptors in response to the parasympathetic neurotransmitter acetylcholine activates three main signaling pathways that either enhance (PKC, Ca2+) or attenuate (ERK) protein secretion (Figure 5). The net protein secretory output in response to acetylcholine stimulation will likely depend on the relative contribution of the stimulatory versus the inhibitory signal transduction pathways. VIP-Activated Signal Transduction Pathways VIP interacts with specific VIP receptors located on the basolateral membranes of lacrimal gland cells. Two types of VIP receptors have been identified, VIPRI and VIPRII, which are also known as VIPACR1 and VIPACR2, and both of them are expressed in the lacrimal gland, with VIPRI being the predominant type. Adenylate cyclase-coupled signaling pathway The VIP receptor uses the G protein Gas to activate the effector enzyme adenylyl cyclase (AC), which produces the second-messenger molecule, cyclic adenosine monophosphate (cAMP) (Figure 6). Molecular cloning has identified several isoforms of mammalian AC forming a family of at least 10 enzymes (ACI-X). There are at least three isoforms of AC (ACII, ACIII, and ACIV) present in the lacrimal gland, each having a unique localization. Although the regulation of AC enzymatic activity is complex and isoform specific, all AC isoforms are activated

Parasympathetic nerves

VIP

Ras Raf

87

VIP receptor I or II

Gαs

AC Inactive PKA

Ca2+ ATP

cAMP

R

C

R

C

MEK Active PKA

ERK –

+

C

R

C

R

Exocytosis Figure 6 Schematic depicting the main signal transduction pathways activated by the VIP to stimulate protein secretion. The activation of lacrimal gland VIP receptors activates two main signaling pathways that enhance protein secretion.

88

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

by Gas. Increases in the intracellular levels of cAMP lead to activation of protein kinase A (PKA), a ubiquitous serine and threonine protein kinase. In its inactive state, PKA consists of a complex of two catalytic (C) subunits and two regulatory (R) subunits (Figure 6). Binding of cAMP to the R subunit alleviates an autoinhibitory contact that releases the active C subunit (Figure 6). The active kinase is then free to phosphorylate specific protein substrates to stimulate lacrimal gland protein and fluid secretion.

signal-terminating mechanisms include desensitization of the VIPR and AC, sequestration of the PKA C subunits by the naturally occurring protein kinase inhibitor (PKI), and activation of protein phosphatases. a1-Adrenergic Agonist-Activated Signal Transduction Pathways Norepinephrine, released from the sympathetic nerves, binds to a1- and b-adrenergic receptors on lacrimal gland cells. b-Adrenergic receptors are coupled to activation of AC to activate a cAMP-dependent signal transduction pathway, as discussed for VIP. Of the three a1-adrenergic receptor subtypes (a1A, a1C, and a1D) identified, only the a1D subtype is expressed in the lacrimal gland.

MAPK-coupled signaling pathway

Recently, it has been shown that addition of VIP, exogenous cAMP, or analogs that increase cAMP levels inhibited both basal as well as cholinergic induced activation of MAPK in the lacrimal gland. One implication of these findings is that it could help explain the well-documented synergistic effect that addition of a cAMP along with a Ca2+-/ PKC-dependent agonist have on lacrimal gland protein secretion. Indeed, cholinergic agonists activate MAPK, which attenuates protein secretion. When the cAMP pathway is activated, MAPK activity is inhibited; this should alleviate the inhibitory effect that MAPK has on secretion and as a result, protein secretion will be potentiated if both the Ca2+/PKC and the cAMP pathways are activated simultaneously. The termination of VIP-activated signaling pathways likely includes activation of the cAMP-phosphodiesterase, which converts cAMP to the inactive 5’-AMP. Other

Ca2+- and PKC-coupled signaling pathways

In most exocrine tissues, a1-adrenergic agonists activate the same pathway as cholinergic agonists (i.e., activation of PLC and PLD). Surprisingly, in the lacrimal gland, a1-adrenergic agonists do not activate PLC or PLD, although their activation leads to a slight increase in cytosolic Ca2+ and to activation of PKC isoforms (Figure 7). To date, the effector enzyme(s) activated by lacrimal gland a1D-adrenergic receptors to mobilize Ca2+ and activate PKC is still unknown. Although Ca2+ mobilization in the lacrimal gland in response to adrenergic agonist stimulation is well

Sympathetic nerves

EGF

Norepinephrine α1D-AR

EGFR Pro-EGF MMP

Ras

Sc

Raf

SOS G rb 2

Gαq

? Ca2+

PKC

h

eNOS

ε

α, δ

MEK NO ERK GC UTP

cGMP

+ –

+ – Exocytosis Figure 7 Schematic depicting the main signal transduction pathways activated by the norepinephrine to stimulate protein secretion. The activation of lacrimal gland a1D-adrenergic receptors activates stimulatory pathways, including PKCE and cGMP, and inhibitory pathways, including PKCa, PKCd, and ERK. It is likely that the net lacrimal gland protein secretion in response to sympathetic stimulation is a balance between these stimulatory and inhibitory signal transduction pathways.

Lacrimal Gland Signaling: Neural

documented, the mechanisms involved are poorly understood. A role for IP3 has been ruled out since adrenergic agonists do not increase its production as they fail to activate PLC. It has been proposed that cyclic ADP ribose, which activates ryanodine receptors to release Ca2+ into the cytosol, might be involved. Other investigators proposed that nitric oxide (NO)-induced generation of cyclic guanosine monophosphate (cGMP) is involved in a1adrenergic agonist-induced mobilization of Ca2+. The role of PKC in a1-adrenergic agonist-induced lacrimal gland protein secretion has been studied using the myristoylated pseudosubstrate-derived peptides. It was shown that a1-adrenergic agonists activate three PKC isoforms – PKCa, -d, and -E. Activation of PKCE enhances protein secretion, whereas activation of PKCa and -d attenuates protein secretion. This is in contrast to the stimulatory effect that PKCa and -d isoforms have on protein secretion when activated by cholinergic agonists. These findings imply that the effect (inhibitory or stimulatory) of a given isoform of PKC in the lacrimal gland is stimulus dependent and might be dictated by the cellular location of PKC isoforms. MAPK-coupled signaling pathway

Similar to cholinergic agonists, a1-adrenergic agonists activate MAPK to attenuate lacrimal gland protein secretion. However, in contrast to the cholinergic pathway, activation of MAPK by the a1D-adrenergic receptor does not involve the nonreceptor tyrosine kinases Pyk2 and Src, but involves activation of the epidermal growth factor receptor (EGFR; Figure 7). The EGFR, also known as Erb1, is the prototypical member of the ErbB/EGFR family of receptors which consists of three additional members (ErbB2–4). EGFR is a type 1 transmembrane tyrosine kinase receptor consisting of an extracellular domain (ligand-binding site), a transmembrane domain, and the carboxy-terminal, an intracellular domain containing the tyrosine kinase motif. In addition, the carboxy-terminal domain contains tyrosine residues that become phosphorylated following ligand binding and receptor dimerization. Following receptor activation, several exogenous substrates that contain either Src homology 2 (SH2) or protein tyrosine binding (PTB) motifs are recruited to specific phosphorylated tyrosine residue. In the lacrimal gland, the activation of the EGFR by a1-adrenergic agonists occurs through a process termed as transactivation and involves the activation of a metalloproteinase and shedding of EGF (Figure 7). Following activation, Shc (an SH2 motif-containing protein) is recruited to the EGFR, followed by recruitment of Grb2 and the guanine nucleotide exchange factor protein, SOS (Figure 7). SOS stimulates the exchange of GDP for GTP on Ras and hence leads to its activation. Activated Ras triggers the activation of the MAPK cascade, leading to activation of ERK (Figure 7). Similar to the cholinergic pathway, a1-adrenergic-activated ERK has been shown to attenuate lacrimal gland protein secretion.

89

NO-coupled signaling pathway NO, along with L-citruline, is produced from L-arginine in the presence of O2- and NADPH-derived electrons. This reaction is catalyzed by the enzyme NO synthase (NOS). There are three well-characterized isoforms of NOS expressed by mammalian cells: neuronal NOS (nNOS also known as NOS1), inducible NOS (iNOS or NOS2), and endothelial NOS (eNOS or NOS3). Activation of nNOS and eNOS, but not iNOS, requires calmodulin and an increase in [Ca2+]. It has been shown recently that lacrimal gland a1-adrenergic receptors are coupled to the NO/cGMP pathway (Figure 7). Indeed, it was found that both nNOS and eNOS are expressed in the lacrimal gland. The addition of a1-adrenergic agonists led to generation of NO, presumably through activation of eNOS and not nNOS. NO, in turn, activates soluble guanylate cyclase to generate cGMP which enhances lacrimal gland protein secretion. The termination of the a1-adrenergic signaling pathway is likely to involve receptor desensitization, activation of protein phosphatases to dephosphorylate ERK and other phosphorylated substrates, and the activation cGMP-phosphodiesterase, which converts cGMP to the inactive 5’-GMP. In summary, the activation of lacrimal gland a1Dadrenergic receptors, in response to the sympathetic neurotransmitter norepinephrine, activates stimulatory pathways (including PKCE and cGMP) and inhibitory pathways (including PKCa, PKCd, and ERK; Figure 7). It is likely that the net lacrimal gland protein secretion in response to sympathetic stimulation is a balance between these stimulatory and inhibitory signal transduction pathways. See also: Adaptive Immune System and the Eye: Mucosal Immunity; Lacrimal Gland Hormone Regulation; Lacrimal Gland Overview; Tear Film.

Further Reading Botelho, S. Y., Hisada, M., and Fuenmayo, N. (1966). Functional innervation of the lacrimal gland in the cat. Archives of Ophthalmology 76: 581–588. Broad, L., Braun, F., Lievremont, J., et al. (2001). Role of the phospholipase C-inositol 1,4,5-trisphosphate pathway in calcium release-activated calcium current and capacitative calcium entry. Journal of Biological Chemistry 276: 15945–15952. Chen, L., Hodges, R. R., Funaki, C., et al. (2006). The effects of a1D-adrenergic receptors on shedding of biologically active EGF in freshly isolated lacrimal gland epithelial cells. American Journal of Physiology. Cell Physiology 291: C946–C956. Funaki, C., Hodges, R. R., and Dartt, D. A. (2007). Role of cAMP inhibition of p44/p42 mitogen-activated protein kinase in potentiation of protein secretion in rat lacrimal gland. American Journal of Physiology. Cell Physiology 293: C1551–C1560. Hodges, R. R. and Dartt, D. A. (2003). Regulatory pathways in lacrimal gland epithelium. International Review of Cytology 231: 129–196.

90

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Hodges, R., Rios, J., Vrouvlianis, J., et al. (2006). Role of protein kinase C, Ca2+, Pyk2 and c-Src in agonist activation of rat lacrimal gland p42/p44 MAPK. Investigative Ophthalmology and Visual Sciences 47: 3352–3359. Ota, I., Zoukhri, D., Hodges, R., et al. (2003). a1-Adrenergic and cholinergic agonists activate MAPK by separate mechanisms to

inhibit secretion in lacrimal gland. American Journal of Physiology. Cell Physiology 284: C168–C178. Wu, K., Jerdeva, G., da Costa, S., et al. (2006). Molecular mechanisms of lacrimal acinar secretory vesicle exocytosis. Experimental Eye Research 83: 84–96.

Lids: Anatomy, Pathophysiology, Mucocutaneous Junction T Wojno, The Emory Clinic, Atlanta, GA, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Actinic lesion – Dry, scaly, rough-textured patches that form after years of exposure to ultraviolet light, such as sunlight. Amblyopia – Disorder of the visual system that is characterized by poor or indistinct vision in an eye that is otherwise physically normal. Anisometropia – Condition in which the two eyes have unequal refractive power. Blepharitis – Chronic inflammation of the eyelids. Blepharoplasty – Surgical procedure intended to reshape the upper eyelid or lower eyelid by the removal or repositioning of excess tissue as well as by reinforcement of surrounding muscles and tendons. Dermatochalasis – Redundant, baggy eyelids. Ectopion – Outward malposition of the eyelid. Entropion – Inward malposition of the eyelid. Lagophthalmos – Inability to close the eye. Ptosis – Downward malposition of the upper eyelid. Strabismus – A condition in which the eyes are not properly aligned with each other. Trichiasis – Misdirected eyelashes.

Anatomy From a functional perspective, the upper lid can be divided into anterior, middle, and posterior lamellae (Figure 1). In the upper lid, the anterior lamella consists of the skin and orbicularis muscle while the posterior lamella consists of the conjunctiva, tarsus, levator, and Mu¨ller’s muscle. The middle lamella is the orbital septum and orbital fat. The thin eyelid skin covers the orbicularis muscle, which is functionally divided into the pretarsal, preseptal, and orbital parts. There is no discreet anatomic border to these components of the orbicularis. The levator muscle originates just superior to the annulus of Zinn at the orbital apex and changes from striated muscle to fibrous apponeurosis 15 mm above the superior border of tarsus. The levator inserts into the superior border and anterior surface of the tarsal plate. It is innervated by the third cranial nerve. Mu¨ller’s muscle is only 10–14 mm long and arises from the underbelly of the levator and inserts into the superior border of tarsus. It is composed of

smooth muscle fibers and is adrenergically innervated. The levator and Mu¨ller’s muscles function to open the upper lid while the orbicularis muscle closes it. The orbital septum is a thin, multilayered sheet of fibrous tissue separating the lid from the orbit. It arises from the superior orbital rim and inserts onto the levator aponeurosis 2–10 mm above the superior border of tarsus in Caucasians, 15 mm or more in blacks and at or below the superior border of tarsus in Asians. Small, fine, fibrous attachments extend from the levator to the subcutaneous tissue. These attachments and the insertion of the septum form the lid crease while the skin above the crease forms the lid fold. There are two fat pockets in the upper lid found nasally and centrally. The upper tarsus is a firm connective tissue usually 10–12 mm in height and 1 mm in thickness. The lower lid is likewise divided into three functional lamellae: an anterior layer of skin and orbicularis, a posterior layer of conjunctiva and lower lid retractors, and a middle layer of septum and orbital fat (Figure 2). The lower lid retractors are composed of the capsulopalpebral fascia (the equivalent of the levator in the upper lid) and the inferior tarsal muscle (the equivalent of the Mu¨ller’s muscle in the upper lid). The capsulopalpebral fascia is a fibrous band originating from the underbelly of the inferior rectus muscle that courses anteriorly, enveloping the inferior oblique, and inserts onto the inferior border of tarsus. It is functionally controlled by its origin from the inferior rectus muscle to retract the lower lid inferiorly when the eye looks downward, preserving unobstructed vision. The inferior tarsal muscle is often just scattered smooth muscle fibers intermixed with the capsulopalpebral fascia. As in the upper lid, the septum arises from the orbital rim and inserts on the inferior border of tarsus, often blending with the lower lid retractors. A lower lid crease occasionally exists but is usually less obvious. There are three fat pockets in the lower lid: nasal, central, and temporal (or lateral). The lower tarsus is usually 4–5 mm in height. The lid margin is the border between the anterior skin–muscle layer and the posterior tarsoconjunctival layer. There are two to three irregular rows of lashes, whose bulbs are embedded just below the skin surface within the orbicularis muscle. Posterior to the lash line are the meibomian gland orifices. These sebaceous glands are embedded within the tarsal plates and run the entire vertical length of the tarsus. There are about 25 of these glands in the upper lid and 20 in the lower lid. The mucocutaneous junction is just posterior to the meibomian

91

92

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Skin Subcutaneous tissue Preseptal orbicularis

Superior rectus muscle

Orbital septum Levator muscle

Preaponeurotic fat pad

Suspensory ligament of the upper fornix

Levator aponeurosis Fine attachment from levator aponeurosis to Müller’s muscle

Superior conjunctival fornix

Conjunctiva

Müller’s muscle Tarsus

Pretarsal orbicularis

* = Whitnall’s ligament Figure 1 Cross-section of the upper eyelid.

Conjunctiva Skin Inferior fornix

Pretarsal orbicularis Tarsus

Inferior rectus muscle

Suspensory ligament in the inferior fornix Inferior tarsal muscle Orbital septum Preseptal orbicularis Fat pad

Capsulo-palpebral fascia

Inferior oblique muscle Figure 2 Cross-section of the lower eyelid.

gland orifices. The gray line is a variably visible section of pretarsal muscle (muscle of Riolan) just anterior to the tarsus. Embedded within the lid margin are the apocrine glands of Moll and the sebaceous glands of Zeiss associated with the lash follicles. The upper and lower lids join laterally where the pretarsal heads of the orbicularis muscle form the lateral canthal tendon, which inserts into the orbital tubercle just posterior to the lateral orbital rim. Medially, the preseptal and pretarsal muscle form the medial canthal tendon,

whose anterior and posterior heads surround the lacrimal sac. The lacrimal puncta open on the lid margin 6 mm from the medial commissure of the lids.

Pathophysiology Dermatochalasis Dermatochalasis is the normal aging change in the upper and lower lids characterized by loose, redundant skin and

Lids: Anatomy, Pathophysiology, Mucocutaneous Junction

orbicularis muscle often with bulging of the orbital fat pockets (Figure 3). It may be associated with ptosis of the eyebrows and forehead relaxation. When severe in the upper lids, it can limit peripheral vision and obstruct the central visual axis. Lower lid dermatochalasis rarely affects the vision, but in rare cases, fat bulging can be so extreme so as to contact the patient’s glasses. Surgical treatment is aimed at removal of the excess skin, orbicularis, and fat.

Ptosis Lid ptosis is a lower-than-normal position of the upper lid margin (Figure 4). When the lid margin is 2 mm or less from the center of the pupil, the superior visual field is usually significantly obstructed. Ptosis is either congenital or acquired. Congenital ptosis is usually the result of a malformed levator muscle often with a family history. It is unilateral in 75% of cases and bilateral (although often

Figure 3 Dermatochalasis of all four eyelids.

93

asymmetric) in 25% of cases. It is associated with anisometropia, amblyopia, or strabismus in 30% of cases. In general, the more severe the ptosis, the more dystrophic appears the muscle histologically. In cases of severe congenital ptosis, striated muscle fibers are usually completely absent, totally replaced by fibro-fatty connective tissue. Treatment for mild-to-moderate congenital ptosis is to perform levator resection surgery, wherein 12–18 mm of the distal levator and the underlying Mu¨ller’s muscle is resected and the cut end resewn to the superior border of tarsus. This effectively shortens the muscle, resulting in a higher resting level of the lid on the globe but does not improve the overall movement of the lid. The most common problems post-operatively are undercorrection, overcorrection, or asymmetry of the lids often necessitating reoperation. Surgery induces lagophthalmos that increases with the amount of levator resected. Surprisingly, if done during childhood, lagophthalmos is usually well tolerated throughout the patient’s lifetime. In severe congenital ptosis, the levator is usually so dystrophic that resection is ineffective. In such cases a sling must be performed. In this surgery, autogenous or banked fascia or some alloplastic material is sewn into the tarsus and then threaded under the skin to the frontalis muscle in the forehead. The patient then elevates the lid by contracting the frontalis muscle, which pulls up the lid margin. Most patients do so automatically, resulting in effective lid opening. Acquired ptosis, usually seen with aging, results from thinning or dehiscence of the levator aponeurosis from the tarsal plate. Any condition, however, that causes lid swelling or stretching (cataract surgery, trauma, contact lens wear, etc.) can result in acquired ptosis. Much less frequently, acquired ptosis is due to actual deterioration of the levator muscle or true muscular dystrophy. The usual treatment of acquired ptosis is to shorten the levator aponeurosis but to a much smaller degree than done with congenital ptosis (usually 4–10 mm). In adults, ptosis repair, when performed bilaterally, is often combined with blepharoplasty surgery for optimal cosmesis. Like in congenital ptosis, undercorrection, overcorrection, and asymmetry are common problems. In adults, however, adjustments can often be accomplished in the office under local anesthesia. Lagophthalmos is generally to be avoided in adult surgery, since the cornea is usually very intolerant of any chronic exposure and can rapidly result in discomfort and even ulceration. Retraction

Figure 4 Ptosis of the right upper lid.

The opposite of ptosis, retraction is abnormal elevation of the upper lid or downward positioning of the lower lid (Figure 5). Most cases of upper lid retraction are due to thyroid eye disease resulting from contracture of the upper and lower lid retractors secondary to inflammation.

94

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Figure 5 Retraction of all four lids secondary to thyroid eye disease.

Figure 6 Involutional entropion of the right lower lid.

The lower lids may also retract due to age, associated with poor support from the cheekbones. Retraction due to thyroid eye disease is often associated with exopthalmos or an abnormal anterior displacement of the globe, which further increases the stare – so characteristic of this disease. Patients with retraction often have lagophthalmos, resulting in corneal exposure and irritation. Lower lid retraction is often seen as a normal physiologic variant in people with shallow orbits, especially common in black patients. Retraction is occasionally seen with overexcessive skin excision in blepharoplasty surgery. Treatment of upper lid retraction in thyroid eye disease consists of graded recession of the levator–Mu¨ller’s muscle complex superiorly so as to drop the upper lid margin down. Some surgeons insert spacers (autologous or banked tissues) between the recessed, cut edge of the levator–Mueller’s muscle complex and the superior border of the tarsus when performing this surgery. In the lower lid, retraction is treated with recession of the lower lid retractors very frequently combined with spacer grafts for additional support of the lower lid. In the upper lid, gravity works in favor of the correction while it works against it in the lower lid. If the retraction is due to skin shortage after blepharoplasty, then skin grafting may be necessary.

distinctive rolled-in appearance and can be reduced by pulling the lid against the lateral orbital rim, effectively tightening the lid. Surgical correction is aimed at correcting the horizontal laxity by resection of the redundant lid margin with resuspension at the lateral canthus. The vertical lid laxity may be corrected by plication of the lower lid retractors to the inferior border of tarsus. An effective repair is one which combines both of these procedures often performed through a lower lid blepharoplasty incision. Involutional entropion usually does not occur in the upper lid. Such aging changes usually result in ptosis, as discussed above, or in lash ptosis – a downward angulation of the lashes due to relaxation of the anterior lamella of the eyelid in which the lash follicles are embedded. Lash ptosis is often corrected as part of an upper lid blepharoplasty procedure. Cicatricial entropion may occur in both the upper and lower lids. It is due to vertical shortening of the posterior lamella of the eyelid, the tarsoconjunctival layer. It may be due to autoimmune disorders of the mucous membranes, inflammation, infection, surgery, trauma, or long-term use of glaucoma drops. Often, however, the cause is unclear. All surgeries for cicatricial entropion can be conceptualized as falling into three categories. The first are those that act by outward rotation of the lid margin (prototypical Weis procedure) usually involving a full-thickness, horizontal blepharotomy 4 mm from the lid margin. The second category involves expanding the shortened posterior lamella of the eyelid with grafts (usually buccal mucosa, amniotic membrane, or banked sclera). The third category involves procedures that split the lid margin at the gray line often with insertion of a spacer material, such as mucous membrane, to thicken the lid margin to the point that the lashes no longer rub against the globe. In some cases, the lash-bearing segment of the lid margin may simply be resected after splitting the lid margin.

Entropion Entropion is an inward turning of the upper or lower lid margin (Figure 6). This results in the lashes rubbing on the globe, causing irritation and even corneal ulceration. Senile or involutional lower lid entropion is common with age and is due to excess horizontal (canthal tendons) and vertical laxity (the lower lid retractors). This may be intermittent at first but usually progresses to a chronic condition. The lid margin appears to have a

Lids: Anatomy, Pathophysiology, Mucocutaneous Junction

Trichiasis Often confused with entropion, trichiasis is inward misdirection of lashes against the globe in the presence of normal lid margin position. Trichiasis may, however, coexist with cicatricial entropion. Trichiasis is caused by the same factors that cause cicatricial entropion but is often idiopathic. Focal trichiasis is often treated by simple epilation of the offending lashes. If recurrent, electrolysis, cryotherapy, or laser may be used. Repeat treatment is often necessary since these modalities will kill the visible offending lashes but not the lashes that are about to bud. For large areas of the lid margin, surgery as outlined above for ciciatricial entropion may be needed. Alternatively, if focal, a segmental resection of the involved lid margin can be performed. Distichiasis Distichiasis is an additional row of lashes that grow from the meibomian orifices on the posterior lid margin. Such lashes will rub against the globe causing corneal irritation. Distichiasis may be congenital, often with a family history or acquired due to lid inflammation causing metaplasia of the cells in the posterior layer of the lid margin. It is treated with electrolysis or cryotherapy often after splitting the lid margin to prevent damage to the normal lashes. It may also be treated by direct surgical excision of the offending lashes. Ectropion Ectropion is outward rotation of the lower lid margin away from the globe (Figure 7). The exposure of the globe and the palpebral conjunctiva results in irritation, corneal damage, and keratinization of the conjunctiva. Epiphora, increased tear production, often results when the lower punctum stands off the globe and cannot adequately drain tears from the eye. Involutional ectropion occurs as an

95

aging change secondary to horizontal lid laxity, mainly in the medial and lateral canthal tendons. Repair is accomplished by horizontal shortening of the redundant lid margin, usually at the lateral canthus, sometimes combined with vertical shortening of the lower lid retractors. When the problem is mainly medial, punctual ectropion, a spindle of conjunctiva and lower lid retractors is resected immediately below the lower punctum. Lower lid ectropion may coexist with retraction, as was discussed above. Involutional ectropion usually does not occur in the upper lid. Cicatricial ectropion is due to vertical shortening of the anterior, skin–muscle lamella of the upper or lower lid. It is usually secondary to trauma or skin disorders. In the upper lid, release of the scarred tissue is followed by skin grafting from the opposite upper lid, the retroauricular area, supraclavicular area, or forearm. In the lower lid, horizontal lid shortening is also often required since the chronically ectropic lid often has or develops a component of excess horizontal laxity. In the lower lid, a skin graft may be replaced by advancement of a skin–muscle flap from the surrounding area. Paralytic ectropion is due to paralysis of the seventh nerve. There is also lagophthalmos due to inability of the upper and lower lids to close. Exposure can be severe and can lead to corneal ulceration. Conservative treatment consists of ocular lubrication, moist chamber devices and lid taping. Tarsorrhaphy may be needed if the cornea dries out in spite of conservative measures. For those cases in which seventh nerve function will not return a gold weight or spring may be placed in the upper lid to counter the lagophthalmos. The lower lid usually needs to be tightened horizontally and often lifted vertically with a posterior lamellar graft or suspension sling of fascia or silicone. Although such treatments are helpful, patients never have a normal blink with any of these procedures. Floppy Lid Syndrome This uncommon syndrome is most frequently seen in obese, middle-aged males (Figure 8). The upper lid is extremely lax and spontaneously everts often while the patient sleeps and rubs against the pillow. Those affected have a severe papillary conjunctivitis, ropey mucoid discharge, and irritation. Floppy lid syndrome is accompanied by sleep apnea and hypertension. Symptoms can sometimes be controlled by having the patient wear a Fox shield over the eye while sleeping to reduce nocturnal eversion of the lid. Most often, horizontal upper lid shortening is needed along with tightening of the lower lid for control of symptoms. Common Malignant Eyelid Tumors

Figure 7 Involutional ectropion of the lower lids, worse on the right.

Basal cell carcinoma (BCC) is the most common eyelid malignancy accounting for 85–90% of such lesions.

96

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Figure 8 Demonstration of the floppy eyelid syndrome on the left.

BCC is an actinic lesion and is more common in fairskinned individuals on sun-exposed parts of the body. It is most common on the lower lid, followed by the medial canthus, upper lid, and lateral canthus. The tumor is classically a firm, nodular lesion with a central ulcer and pearly, vascular border (Figure 9). It may also present as an irritating, erythematous patch or rarely be entirely subcutaneous. Lesions of the medial canthus have a propensity for deep invasion into the eye socket. BCC only rarely metastasizes and generally grows very slowly. The preferred treatment is complete excision with frozen section control or by Mohs micrographic surgery usually performed by a specially trained dermatologist. It provides the best cure rate (98–99%) while saving as much normal tissue as possible. Excision is followed by reconstruction of the eyelid defect or occasionally by spontaneous healing if the defect is small or does not involve the lid margin. Radiation therapy is an alternative to surgery but not as effective due to the relatively high recurrence rate (5–20%). Recurrence is also difficult to detect after radiation due to tissue alteration from the treatment. Cryotherapy is sometimes used for small lesions but has similar problems with a relatively high recurrence rate (20–30%) and depigmentation and atrophy of treated tissue. Squamous cell carcinoma (SCC) accounts for about 5% of periocular malignancies. It too is an actinic lesion, being more common on sun-exposed areas of fair-skinned individuals. Again, the lower lid is the most common location. It most frequently presents as an erythematous, thickened patch with erosion of the involved tissue (Figure 10). A nodular subtype is occasionally observed. It may arise from preexisting actinic keratoses, which undergoes malignant transformation 20% of the time. SCC can metastasize through the regional lymphatics or spread along involved sensory nerves. Mohs removal and reconstruction are the preferred therapy. Radiation

Figure 9 Nodular basal cell carcinoma of the left lower lid.

Figure 10 Squamous cell carcinoma of the left medial canthus.

therapy is much less effective in SCC and thus is used only for palliation of surgically unresectable tumors. Topical therapy with 5-fluorouracil or imiquimod is a frequent treatment of the premalignant actinic keratosis. Sebaceous cell carcinoma accounts for about 5% of periocular malignancies but is increasing in frequency, which may be due in part to better pathological diagnosis. It arises from the meibomian glands, the glands of Zeiss, and sebaceous glands of the caruncle. It presents as a nodular lesion, which is often mistaken for a chalazion or a diffuse intraepithelial pattern that looks like a chronic conjunctivitis (pagetoid spread) (Figure 11). Because of this, the diagnosis is often delayed until biopsied. It is capable of both regional lymphatic and vascular spread. It is considered deadlier than SCC. Mohs microsurgery may be effective in tumor removal if the tumor is nodular, but pagetoid spread and skip lesions often necessitate wide sampling of the bulbar and palpebral conjunctiva (map biopsy). If diffusely spread, orbital exenteration is

Lids: Anatomy, Pathophysiology, Mucocutaneous Junction

Figure 11 Pagetoid spread of sebaceous cell carcinoma of the conjunctiva of the left lower lid.

97

Figure 12 Malignant melanoma of the right lower lid.

usually required. Sebaceous cell carcinoma is relatively radio resistant and responds poorly to chemotherapy. Malignant melanoma accounts for less than 5% of periocular malignancies, but it too is increasing in frequency (Figure 12). It too is an actinic lesion but other less well-defined causes play a role. It arises spontaneously usually as a nodule or in an existing nevus or area of lentigo maligna (intraepithelial tumor). It too is capable of spread through the regional lymphatics and the blood stream. Surgical excision is the treatment of choice, but requires permanent section histology to adequately assess tumor margins. Because of this, the resection of the tumor is often spread over several days (Slow Mohs) followed by reconstruction when margins are clear. Benign Eyelid Tumors A variety of benign lesions are found on the lids and periocular skin. The vast majority are of minimal functional significance, but patients frequently request removal for cosmetic reasons. The most common are nevi, inclusion and glandular cysts, seborrheic keratoses, verruca, skin tags, and benign glandular tumors. Simple excision can usually be carried out in the office under local anesthesia. Inflammatory and Infectious Disorders of the Lids Chalazion is a lipogranuloma of one of the meibomian glands of the tarsal plate. They arise relatively rapidly over a period of a few days, often with inflammation and discomfort. They can progress to form a chronic peasized, firm nodule in the lid (Figure 13). The initial treatment is to hot compress the involved lid frequently in the first few days in hopes of opening the obstructed gland. If not effective many will resolve spontaneously over the next several weeks to few months. Patients often

Figure 13 Chalazion of the right upper lid.

request removal for cosmetic reasons. This can usually be done in the office with local anesthesia. An alternative is to inject the chalazion with steroid, which is effective about 50% of the time. Hordeolum is a staphylococcal infection of one of the sebaceous glands of the lid. It presents as an acutely swollen, erythematous, painful nodule on the lid margin (external hordeolum) or on the palpebral conjunctiva (internal hordeolum). Hot compress, topical antibiotics, and surgical drainage are effective. Oral antibiotics may be needed if cellulitis occurs. Blepharitis is the most common inflammatory disorder of the eyelids characterized by redness, swelling, and irritation with visible crusting. It is a chronic condition characterized by intermittent exacerbations. There are three subtypes, but many patients show combinations of all three. In staphylococcal blepharitis, chronic colonization of the lid margin leads to inflammation from the bacterial toxins and antigens. Slit-lamp examination

98

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

reveals characteristic white collarettes around the base of the lashes. In seborrheic blepharitis, there are greasy scales (scurf) found on the lashes and often associated seborrheic dermatitis of the face and scalp. The third form, posterior blepharits (posterior lid margin disease, meibomian gland dysfunction, and meibomianits) is characterized by a change in the normal clear meibomian gland secretion to a thick, cloudy to yellow, oily discharge. Posterior blepharitis is often associated with acne rosacea and chalazia. All forms of blepharitis lead to chronic redness of the conjunctiva, sometimes with papillary hypertrophy. The tear film is often unstable as manifested by a rapid tear break-up time. Dry eye is a frequent association. In addition to the above findings, slit-lamp examination may reveal punctate epithelial erosions (PEEs) of the cornea and marginal corneal erosions due to staphylococcal hypersensitivity. There may be small, rounded domes on the meibomian orifices, manifestations of the thick, inspissated secretions. Digital pressure on the lid may cause the meibomian glands to express material that can, in severe cases, have a cheesy consistency. Treatment consists of warm compresses to thin the meibomian secretions and eyelid scrubs with baby

shampoo or commercially available products to clean the lid margin and express the meibomian glands. Topical antibiotic drops or ointment reduces the bacterial colonization of the lid margin and topical steroids help to control the inflammation. Oral, low-dose tetracycline (50 mg doxycycline per day) is helpful in reducing the meibomian gland discharge and normalizing the pH of the tear film. Oral erythromycin can be used if the patient is intolerant to tetracyclines, pregnant, nursing, or under the age of 12. See also: Eyelid Anatomy and the Pathophysiology of Blinking; Lacrimal Gland Overview; Tear Drainage.

Further Reading Jordan, D. R. and Anderson, R. A. (2000). Surgical Anatomy of the Ocular Adnexa: A Clinical Approach. San Francisco, CA: American Academy of Ophthalmology. Stewart, W. B. (2000). Surgery of the Eyelid, Orbit, and Lacrimal System. San Francisco, CA: American Academy of Ophthalmology.

Overview of Electrolyte and Fluid Transport Across the Conjunctiva O A Candia and L J Alvarez, Mount Sinai School of Medicine, New York, NY, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Short-circuit current – The short-circuit current (Isc) across tissues isolated within an Ussing chamber is defined as the charge flow per time, per the crosssectional area of the epithelium that is exposed to the bathing solutions, when the tissue is short-circuited by clamping the transepithelial voltage (PDt) to zero with an external circuit. The Isc is a current that circulates through the tissue and the external circuit thereby completing a closed loop. Because of this, the Isc enters the tissue across one surface and leaves across the other. Monitoring the Isc provides a continuous measure of the net charged flow of current across the transcellular pathways of the tissue. There is no net flow across the paracellular pathways in the short-circuited condition if the tissue is bathed bilaterally with solutions containing identical ionic concentrations. Adding drugs that affect the ionic channels or electrogenic elements within the membranes of the epithelium will affect the magnitude of the Isc. Transepithelial resistance – An epithelium can be considered to be comprised of an arrangement of resistance elements or resistors. This arrangement is most simply modeled in a multilayered epithelium as two resistors in series, namely Ra (the resistance of the apical membrane) and Rb (the resistance of the basolateral membrane). These resistances are shunted by a parallel resistor, Rshunt, which is the cumulative resistance of the paracellular pathways. As such, transepithelial resistance (Rt) is defined as follows: Rt ¼

ðRa þ Rb Þ Rshunt Ra þ Rb þ Rshunt

In the conjunctiva, Rshunt tends to be lower than the transcellular pathway. This leads to small measured changes in Rt when drugs that selectively affect either Ra or Rb are added. For example, potassium channel blockers should selectively increase Rb, but the effect on the measured Rt parameter is relatively smaller in the conjunctiva than in the corneal epithelium, which is a tight epithelium, that is, a tissue with a high Rshunt. Transepithelial voltage – Epithelial cells transport ions and thereby generate a transepithelial voltage (PDt). The generation of PDt requires (i) an asymmetric distribution of ion channels and

electrogenic transporters on the apical and basolateral sides of the tissue, and (ii) the presence of tight junction proteins between adjacent cells in the superficial epithelium to impede the flow of ions along the paracellular pathways. Unidirectional fluxes – Across epithelia isolated in Ussing-type chambers, a unidirectional flux of a substance (e.g., a radiolabeled electrolyte or water molecule) is its rate of translocation across the tissue from the bathing solution within one hemichamber to the contralateral bath, disregarding any counterbalancing flux in the opposite direction. In practice, unidirectional fluxes are measured in the apical-to-basolateral direction, and again in the basolateral-to-apical direction. A difference in the magnitude of these two unidirectional fluxes provides a measure of the net flux, for example, there is a net chloride flux across the conjunctiva in the basolateral-to-apical direction. Ussing chamber – A device designed by Hans Ussing in 1951 to originally study vectorial ion transport across the frog skin. It has since been modified (i.e., an Ussing-type arrangement) by many investigators to characterize electrolyte transport across various epithelial tissues including epithelia of the eye. This approach has significantly contributed to our understanding of how electrolytes are transported. The Ussing-type methodology entails two aspects. One is the chamber itself, which is constructed to hold the dimensions of a particular tissue, and enable it to be bathed bilaterally. The second aspect is the external electrical circuitry, which can be designed to measure transepithelial voltage, resistance, and current. The effects of pharmacological agents on the electrical parameters generated by the epithelium can be studied by applying test compounds unilaterally to either the apical-side or basolateral-side baths.

Introduction The conjunctiva and the corneal epithelium together form the ocular surface. The conjunctiva (from late Latin, feminine of conjunctivus, or conjoining) is in essence a connection (conjunction) between the eyelids, the sclera of the eyeball, and the cornea. It lines the posterior surface of the

99

100

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

eyelids (the palpebral conjunctiva) and the exposed, anterior aspect of the globe (the bulbar conjunctiva). The latter is loosely attached to the sclera of the eyeball, and translucent, thereby exposing the so-called ‘white of the eye’; it merges with the corneal epithelium at the limbus, which constitutes the edge of the cornea. The palpebral conjunctiva is tightly adherent to the eyelid. The space, lined by the conjunctiva, between the lids and the globe is known as the conjunctival sac. The bottom of the sac, which is unattached to the eyelids or to the eyeball, is known as the fornix, forniceal region, or conjunctival fold. The untethered nature of the fornix allows the eyeball to move freely. The conjunctival sac varies in size depending upon the degree to which the lids are open. The depths of the unextended sac in humans are about 14–16 mm superiorly and 9–11 mm inferiorly. The total surface area of the conjunctiva is about 9 and 17 times larger than that of the cornea in rabbits and humans, respectively. The lacrimal glands, which secrete tears, open into the superior fornix. The palpebral conjunctiva contains the openings of the lacrimal canaliculi, which allow tears within the conjunctival sac to drain into the nasal cavity. The vasculature of the palpebral conjunctiva is clearly visible within the tissue upon examining the posterior surface of the eyelid. In contrast, the bulbar conjunctiva is normally colorless, unless its vessels are dilated as a result of inflammation (conjunctivitis). The conjunctival vessels arise from a peripheral palpebral arcade and from the anterior ciliary arteries. Blood comes mostly from the orbit, but anastomoses with the facial system. Conjunctival innervation is mediated by the ophthalmic division of the trigeminal nerve. The conjunctiva is extensively innervated with adrenergic, cholinergic, and peptidergic fibers identified in various species. In general, the largest numbers of nerves present are sympathetic, with fewer parasympathetic and sensory nerves. The parasympathetic nerves contain the neurotransmitters acetylcholine and vasoactive intestinal peptide; the sympathetic nerves, norepinephrine and neuropeptide Y; and the sensory nerves, substance P and calcitonin-gene-related peptide. The major roles of the conjunctiva are: (1) to contribute to tear production by secreting electrolytes and fluid; (2) to modify the composition of the tear film by secreting mucins and lipids, and absorbing various organic compounds found in tears; and (3) to contribute to the resistance of the eye to infection by providing protection against microorganisms. The conjunctiva is comprised of an epithelium and an underlying stroma. The epithelium is embryologically related to, and anatomically continuous with, the epithelium of the upper airway. Within the conjunctival epithelium are goblet cells, which are specialized epithelial cells that function as unicellular mucus glands. The goblet cells secrete the

mucin component of the tear film, which consists of three layers, each of which is secreted by different cells. Secreted mucins constitute the inner layer of the tear film and serve as wetting agents that keep the apical, hydrophobic aspects of the corneal and conjunctival epithelia hydrated. The middle, aqueous layer of the tear film contains water, electrolytes, immunoglobulin A (IgA), glucose, and proteins (including antibacterial enzymes). This layer is primarily secreted by the main and accessory lacrimal glands; the latter glands of Krause and Wolfring flank the main lacrimal duct near the superior fornix. It is possible that the conjunctiva also contributes to this layer under basal conditions when the lacrimal glands are not stimulated. The outer, lipid layer of the tear film contains a fat mixture that is secreted by the meibomian glands that line the eyelids. This layer functions to reduce evaporation of the aqueous layer. Underlying the conjunctival epithelium, the connective tissue contains blood vessels, nerves, conjunctival glands, mast cells, and leukocytes including macrophages. The latter, defensive cells can be recruited in large numbers to an injury site on the ocular surface due to disruption of the barrier properties of the epithelium. They may also release paracrine-signaling agents that affect the transport properties of the epithelium, and certain leukocyte populations can also serve as antigen-presenting cells. Recent work has characterized the active transport properties of the conjunctival epithelium. The epithelium is capable of transporting fluid as a consequence of a sufficiently high water permeability bestowed by endogenous water channels (aquaporins) and transepithelial solute movement due to active transport mechanisms. This article includes a synopsis of the current understanding of the electrolyte and fluid transport across the conjunctiva.

Conjunctival Epithelium A primary role of all epithelial tissues, including those in the conjunctiva, is the absorption and/or secretion of fluid. In brief, two main elements are necessary for fluid movement across a membrane or a set of membranes: (1) the driving force represented by an osmotic gradient (or hydrostatic pressure), and (2) a water pathway represented by water channels (aquaporins) and the lipid bilayer. Thus, all fluid secretion or reabsorption is a consequence of the osmotic gradient created by active electrolyte transport, with the direction of fluid movement identical to that of the net transepithelial solute transport. To date, extensive, functional characterizations of the electrolyte transport properties of the conjunctival epithelium have been done only on the isolated rabbit conjunctiva. The epithelium of this species exhibits mechanisms that simultaneously mediate Na+ absorption and Cl secretion. The relative proportions of these oppositely directed functions varies considerably from one individual rabbit conjunctival

Overview of Electrolyte and Fluid Transport Across the Conjunctiva

specimen to another, for reasons that are unknown, but in general Cl transport is predominant. On average, the ratio of Cl secretion to Na+ absorption is about 1.5 to 1, suggesting that the rabbit epithelium can function primarily as a chloride-secreting epithelium potentially capable of moving fluid into the conjunctival sac. However, it must also be noted that a small proportion of conjunctival specimens exhibited a Na+ absorptive rate larger than the rate of Cl secretion. Morphologically, the epithelia of the rabbit bulbar, forniceal, and palpebral regions have distinct appearances (Figure 1). The bulbar epithelium appears columnar and thinner than the other sections with goblet cells present. It is as thick as two to three cell layers and packed irregularly. In the forniceal area, the number of cell layers increases to three or four with a greater abundance of goblet cells. From this region to the lid margin, a transition within the palpebral epithelium is evident with the number of goblet cells diminishing, and the epithelial

101

cells becoming more stratified. Species differences in morphology among mammalian conjunctivae have been described. Excluding variations in the number of goblet cells, the epithelial cells within each conjunctival region appear homogeneous, which suggests that both absorptive and secretory activities coexist within an individual epithelial cell. If so, the conjunctival epithelium exhibits a rare property among epithelia in that the transport functions for Na+ absorption and Cl secretion are not segregated in distinct cell types, and the transport rates for these oppositely directed functions are nearly equivalent in isolated conjunctivae under in vitro conditions. Regardless of the normal, physiologic direction of fluid movement across the human conjunctiva, it is clear that inhibiting reabsorption and/or stimulating secretion may have a beneficial effect by increasing the aqueous layer of the tear film in individuals with a tear-fluid deficit due to various lacrimal gland deficiencies.

G

(d)

(a)

G G

(e)

(b)

G

50 µm (c)

(f)

Figure 1 Histological sections of the rabbit limbal and conjunctival regions. (a) The limbal epithelium upon a highly vascularized stroma. (b) The bulbar conjunctival epithelium. There are goblet cells (G) present among the epithelial cells. (c) The forniceal epithelium nearest the bulbar region. (d) Another section of the forniceal epithelium, within which the goblet cells become more numerous relative to the bulbar region. The forniceal epithelium is thicker than the bulbar epithelium. (e) The palpebral epithelium near the forniceal region that is characterized by a decrease in the number of goblet cells and abundant lymphoid tissue. (f ) The palpebral epithelium closer to the eyelid margin. It is more stratified than the other conjunctival regions and devoid of goblet cells. Adapted from Wei, Z. G., Sun, T. T., and Lavker, R. M. (1996). Rabbit conjunctival and corneal epithelial cells belong to two separate lineages. Investigative Ophthalmology and Visual Science 37: 523–533.

102

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Bioelectric Studies on the Isolated Rabbit Conjunctiva The ionic transport systems potentially mediating the absorptive and/or secretory functions of an epithelium can be efficiently characterized by isolating the epithelium under an Ussing-type arrangement. With this method, a flat piece of an epithelium is dissected with some of its underlying connective tissue or stroma maintained in place to provide structural support. The thickness of the entire dissected preparation is usually about 1 mm, of which about 0.05 mm represents the epithelial cellular compartment. The isolated tissue is then positioned between two hemichambers, which when closed together result in the tissue serving as a partition separating the cavities of the two hemichambers. In this situation, the apical surface of the epithelium interfaces with the cavity of one hemichamber, while the basal surface attached to the underlying stroma interfaces with the contralateral chamber. Each hemichamber is then filled with a physiological solution to simultaneously bathe the two surfaces of the in vitro preparation. The rabbit conjunctiva serves as a fairly good specimen for this approach given its relatively large surface area, as well as the fact that the conjunctival sac can be removed nearly intact as a cylinder and then cut longitudinally to convert it to a flat epithelium that is easily mounted between Ussing-type hemichambers. Typically, about 0.5 cm2 of cross-sectional area of tissue is bilaterally exposed to the bathing solutions within the chambers. A negative aspect of this approach with the conjunctiva is that the epithelium seems to be relatively frail (when compared to the corneal epithelium) as its integrity deteriorates following several hours of isolation within the chambers. Nevertheless, useful electrophysiological experiments can be designed and informative data are obtainable. Upon isolation of an epithelial preparation, such as the conjunctiva, within the divided chambers, a potential

Apical aspect

Na Glucose

difference (PD) develops across the tissue. The PD is a consequence of the active transport of electrolytes by the epithelium, which spends metabolic energy to maintain ionic gradients between the cellular compartment and the extracellular bath. An epithelium will exhibit a negative intracellular voltage with respect to both the apical-side bathing solution and the stromal-side (basolateral) solution. Typically, the positive voltage of the stromal-side bath (PDs) with respect to the cellular compartment occurs because of the electrogenic Na+–K+ ATPase, which extrudes three Na+ ions for two imported K+ ions, and the fact that the cellular K+ ion concentration is above equilibrium, so that K+ will constantly efflux through K+ channels toward the stromal (basolateral) bath (Figure 2 shows an overview of the major transport elements found in the conjunctival epithelium). The Na+–K+ ATPase functions incessantly to maintain cellular K+ above equilibrium. The positive voltage of the apical-side bath (PDa) with respect to the cellular compartment is less positive than PDs. As such, a transepithelial PD (PDt) exists, which equals the difference between the PDs across the respective contralateral surfaces of the epithelium (PDt = PDa – PDs; and has a negative sign with PDs taken as reference). PDa is less positive with respect to the cellular compartment than is PDs because of the efflux of Cl via channels in the apical domain toward the apical-side bath, and in the case of the conjunctiva, an influx of Na+ also occurs via electrogenic transporters that are in the apical membrane (Figure 2). PDt can only exist if tight junctions are present in the epithelium, which is the case in the conjunctiva. These elements are located between the lateral membranes of the most superficial epithelial cells and form a resistance barrier that impedes the diffusion of ions from the contralateral bathing solutions through the paracellular pathways between the epithelial cells. Without the presence of tight junctions, ionic flows in the paracellular pathway would

Basolateral aspect

H

H2O H2O Na

Na

K

Cl H2O Na CI channels NSCC? include CFTR and CLCA AQP5 on apical surface

Na

K

Cl

Cl Cl

Na K H2O

HCO3

cAMP- and Ca-activated K channels are present

AQP3 on basolateral surface

Figure 2 Summary cartoon of the major transport elements found in the rabbit conjunctiva with water fluxes indicated with double arrows. NSCC, nonselective cation channels; CFTR, cystic fibrosis transmembrane conductance regulator (which has chloride channel activity); CLCA, calcium-activated chloride channel; AQP5, aquaporins homolog type 5; AQP3, aquaporins homolog type 3.

Overview of Electrolyte and Fluid Transport Across the Conjunctiva

short-circuit PDt, because PDa and PDs would still exist and result in a net movement of anions from the apical-side bath to the stromal-side bath through the paracellular pathways, and a net movement of cations in the opposite direction through the paracellular pathways. The Ussing-type chambers used to isolate the conjunctiva have ports for inserting electrodes into the bathing solutions to directly measure PDt. In addition, there are ports for current-sending electrodes. These are used to short-circuit PDt with an external circuit connected to an automatic voltage clamp that constantly maintains PDt ¼ 0 mV. The amount of current needed for this is continuously recorded and known as the short-circuit current (Isc). The Isc represents a real-time measure of the net transepithelial movement of electrolytes across the transcellular pathways of the tissue due to metabolically dependent active transport. As PDt is maintained at 0 mV, there is no net movement of electrolytes through the paracellular pathways, even if tight junctions were, in principle, not present. Transepithelial electrical resistance (Rt) can be determined by applying Ohm’s law to the measured PDt (under open-circuit conditions) and the measured Isc (under short-circuited conditions), or by measuring the amount of current necessary to offset the short-circuited condition by a few mV for a few seconds. Either approach gives identical measures of Rt with the conjunctiva. As alluded to above, the integrity of the conjunctival epithelium appears to degenerate following a prolonged period in the chamber. This is observed as a spontaneous, gradual decline in Rt. It appears that there is a loss in paracellular resistance (i.e., tight-junction structure may change with time in vitro) since the Rt decline occurs in the presence of a steady Isc. As noted, under the shortcircuited conditions, increases in paracellular ion movement do not result in a net flow across this pathway given the absence of a PD across the epithelium and identical electrolyte concentrations on each side of the preparation. As such, the Isc continues to measure net transcellular flow of electrolytes in experiments with symmetrical solutions. However, changes in conjunctival Rt in response to the additions of various drugs frequently underestimate changes in membrane resistance. This is because transcellular resistance is proportionally larger than paracellular resistance, which means that large changes in transcellular resistance are recorded as smaller changes when measuring conjunctival Rt. Nonetheless, the initial measurements of conjunctival Rt upon the isolation of fresh preparations within the divided chambers provide a good indication of the barrier properties of the epithelium. This is because the electrical resistances of both the cell membranes as well as the paracellular pathways contribute to the Rt measurement. The paracellular pathways of the in situ conjunctiva at the ocular surface allow for the passive diffusion of

103

hydrophilic solutes across the conjunctiva. Passive paracellular transport of electrolytes across the conjunctiva in vivo, which is analogous to the open-circuit situation in vitro, occurs because of gradients created by the transcellular mechanisms. In addition, cell-impermeable, hydrophilic solutes applied to the conjunctival sac may diffuse across the epithelium through the paracellular route. The transepithelial permeability of such solutes decreases with increased solute size. Tight junctions located at the apical-most aspect of an epithelium create the major barrier for the movement of solutes across all epithelia including the conjunctiva. However, the paracellular route varies considerably among epithelia in terms of permeability to solutes and electrical resistance. Rt measurements can range at least 1000-fold between highly resistant and so-called leaky epithelia. In some tissues, electron microscopy studies have correlated the ultrastructure of the tight junctions with the measured Rt values obtained in vitro. In general, the number of tight-junction strands along the apical-basal axis is proportional to the junctional resistance. Rt values from freshly isolated rabbit conjunctival epithelia are in the range of 1–2 kO cm2, with many studies reporting an average value of 1 kO cm2. This suggests that the conjunctival epithelium is a moderately tight epithelium. For comparison, Rt measurements of the rabbit corneal epithelium and rabbit corneal endothelium are 7–9 kO cm2 and 0.01–0.07 kO cm2, respectively. As such, the electrical resistances of the corneal epithelium and endothelium vary over a range of about 100-fold, and the conjunctiva exhibits an intermediary value. The fact that the measured Rt of the freshly isolated conjunctiva is so high also indicates that tight junctions must exist between the surface goblet cells and the most superficial stratified epithelial cells. An explanation as to why Rt declines with prolonged time in vitro remains to be evinced, but barrier resistance is physiologically regulated in other systems. The stable Isc recorded during the spontaneous Rt decline indicates that the epithelial cells have remained metabolically viable.

Electrolyte Transport Systems of the Rabbit Conjunctiva The bioelectrical approach discussed above has been used to determine the major electrolyte transport systems present in the rabbit conjunctival epithelium. In work done to date, the short-circuiting methodology was used to characterize transcellular transport. Such transport is energy dependent, and controlled by the tissue-specific profile of transporters and channels along the apical and basolateral membranes of the epithelium. The conjunctival apical membranes interface with the tears and the basolateral membranes interface with the paracellular pathways from the tight junctions to the basement membranes at the

104

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

stroma. Transport mechanisms in the tissue are generally uncovered by electrolyte substitution experiments, and the application of relatively specific drugs (i.e., channel blockers, channel openers, transporter inhibitors, etc.) to the bathing solutions of conjunctiva isolated in the divided chambers. In some cases, the identity of channels and transporters that were identified in electrophysiological experiments were corroborated with immunoblotting and immunohistochemcial observations. Identical to other Cl-secreting epithelia, the rabbit conjunctival epithelium has a basolateral bumetanidesensitive Cl uptake process (mediated by the Na+–K+–2Clcotransporter, NKCC1) positioned in series with apical Cl channels, including cystic fibrosis transmembrane conductance regulator (CFTR). In addition, Na+/H+ and Cl/HCO 3 exchangers exist in parallel in the basolateral membrane and can also mediate Cl uptake (Figure 2). Oppositely directed, electrogenic Na+ reabsorption is amiloride-insensitive, indicating the absence of the major epithelial Na+ channel (ENaC) at the apical surface, and occurs through a Na+-dependent cotransporter for glucose. Nonselective cation channels (NSCCs) were identified in whole-cell patch clamping of freshly isolated conjunctival epithelial cells and such channels may also reside at the apical surface. A transcellular Na+ movement occurs because the apical uptake mechanisms for Na+ exist in series with the basolaterally located Na+–K+ pump (Figure 2). Other electrogenic Na+-dependent uptake mechanisms at the apical surface were demonstrated by adding the transported substrates to the apical bath. Transport systems for amino acids, nucleosides, L-lactate, and di- and tri-peptides were evidenced in this manner because such compounds are normally not included in the physiological solutions used to bathe the in vitro preparations. The addition of such compounds to the apical bath results in a shortcircuit current stimulation. In the cases of the amino acids and nucleosides, the largest Isc stimulations occurred with L-arginine and uridine, respectively. The roles of these carriers are not firmly established. It is thought that these transport systems may clear the tear fluid of these compounds in either nonphysiologic or pathophysiological states of the ocular surface when excess amounts of such solutes might have leaked into tears. Results from protocols for immunoblotting and the immunofluorescent labeling of frozen sections from separately isolated bulbar and palpebral regions of the conjunctival epithelium indicated that the proteins for the Na+–glucose cotransporter, Na+–K+ ATPase, and Na+–K+–2Cl cotransporter are uniformly distributed throughout the conjunctiva. These observations suggest that despite stark differences in the regional morphology of the bulbar and palpebral regions, the entire conjunctival epithelium exhibits the elements for transepithelial Na+ and Cl transport.

Regulation of Epithelial Ion Transport in Rabbit Conjunctiva Currently, information on the regulation of electrolyte transport by the conjunctival epithelium is somewhat limited. This is because the characterization of the macroscopic electrolyte transport properties of this tissue, as measured in bicameral Ussing-type chambers, was begun relatively recently. Hence, many fundamental aspects of the tissue have not been elucidated. One underlying rationale for studying conjunctival transport is to define the secretory functions of the epithelium. This effort could prove to have utility in ameliorating complications from dry-eye diseases, and some progress has been made in this regard. Because of the large surface area of the conjunctival epithelium, active transport by conjunctiva with accompanying fluid secretion could, hypothetically, contribute a significant fraction of tear production, which is normally provided in healthy individuals by the lacrimal gland. Upon stimulation, the transepithelial conjunctival contribution could be greater. As commonly found in Cl-secreting epithelia, the exposure of the conjunctiva to secretogogues that increase either cell calcium or cyclic adenosine monophosphate (cAMP) stimulates transepithelial Cl fluxes and the Isc. The latter intracellular messenger can be increased in the conjunctival epithelium with forskolin (a direct stimulator of adenylyl cyclase), dibutyryl-cAMP (a cell permeable form of cAMP), 3-isobutyl-1-methyl-xanthine (IBMX, a nonselective phosphodiesterase inhibitor), rolipram (an inhibitor specific for cAMP-phosphodiesterase type IV), or epinephrine (a nonselective adrenergic agonist). In addition, these agents also increase the transconjunctival Isc under Cl-free conditions indicating that the increased cAMP levels also stimulate the Na+ absorptive activity of the epithelium. The increase in Na+ absorption has been attributed to a protein kinase A (PKA)-regulated, bariuminhibitable, basolateral K+ conductance in the rabbit conjunctival epithelial cells. One or more different types of K+ channels that have not yet been identified may mediate this K+ conductance. The stimulation of the PKA-gated K+ channels hyperpolarizes the negative cell potential relative to the bathing solutions and favors both the uptake of Na+ across the apical face and the efflux of Cl into the tears. There is evidence that apical Cl channels are also gated by PKA, particularly in the case of CFTR. In the shortcircuited conjunctiva, cAMP has a central role in coordinating simultaneous changes in apical Cl and basolateral K+ conductances to enable stimulations in the transcellular transport of Cl in the stromal-to-apical direction and of Na+ in the opposite direction. Should both absorptive and secretory mechanisms coexist within the same cell, such cAMP-evoked stimulations of the in vivo conjunctiva would, in principle, deplete the epithelium of KCl and reduce cellular volume, while Na+ and Cl move in opposite directions both trans- and

Overview of Electrolyte and Fluid Transport Across the Conjunctiva

paracellularly. Experimental measurements of net water fluxes across the isolated conjunctiva (under open-circuit conditions) indicate an increased fluid movement in the stromal (basolateral)-to-apical direction in response to cAMP, likely due to the higher rate of Cl secretion relative to Na+ absorption. Fluid flow occurs under open-circuit conditions, which is the situation in vivo. The dominant transport system, apparently Cl, will be transported mainly transcellularly, while Na+ will reverse from its net tear-to-stroma direction found under shortcircuit conditions to move paracellularly as a companion ion to neutralize the Cl charge and create a possible isotonic fluid at the apical surface. In open circuit, there would still be a transcellular movement of Na+ toward the stroma, and a paracellular movement of Cl toward the stroma, but the magnitude of these flows will be less than the net of Na+ and Cl secreted into tears (Figure 3). cAMP stimulates all flows and increases the net. Other effective secretogogues in the rabbit conjunctiva are: (1) 1-ethyl-2-benzimidazolinone (EBIO), a Cl and K+ channel opener that elicits electrophysiological effects similar to those of cAMP, although different subtypes of channels are likely involved; and (2) the nucleotide uridine 50 -triphosphate (UTP), which stimulates Cl secretion through P2Y2 receptors upon exogenous application to the apical-side bath. Of these, only the latter has been tested on net fluid movement across isolated conjunctivae and found to be a useful stimulant. Recently, synthetic P2Y2 agonists (e.g., diquafosol tetrasodium, which is also known as INS365) have been studied in clinical trials. Such agents are administered 4–5 times daily, and there is a time-dependent loss of efficacy that is observable in the data produced by such trials. This may be because P2Y2-receptor activation is often transitory due to the nature of the Ca2+ signal itself (through the phospholipase C-sensitive calcium signaling pathways) and the fact that purinergic agonists produce receptor desensitization from which recovery is slow. Yet currently, the use of purinergic agonists appears a suitable approach to palliate dry eye because of not only the stimulatory effects on conjunctival Cl secretion and fluid transport, but also the fact that purinergics serve as mucin secretogogues from conjunctival goblet cells. As such, purinergics appear to have utility in conserving the composition of the tear film. The established receptors that stimulate electrolyte and fluid secretion in the stromal-to-tear direction under open-circuit conditions are schematically presented in Figure 4. The specific channel subtypes activated by calcium and cAMP remain to be conclusively identified.

Fluid Transport Studies across Isolated Rabbit Conjunctiva Two commonly used methods to measure water fluxes across various epithelia have been applied to the excised,

Apical aspect

Basolateral aspect

H Cl3

Cl2

−

Na Glucose

+

Na1

Na Na

Cl

105

K K

Cl1

Cl

Na

Cl Cl

Na K

Na3

Na2 HCO3

Cl1 + Cl2 > Cl3 Na3 > Na1 + Na2 (Cl1 + Cl2) − Cl3 = Clnet Na3 − (Na1 + Na2) = Nanet Clnet = Nanet

Figure 3 A simplified model of the sodium and chloride flows across the rabbit conjunctiva under open-circuit conditions, which are analogous to the in vivo situation. Some transporters present in the epithelium have been omitted for clarity. Cl1 is the transcellular efflux of chloride via chloride channels in the apical domain. Cl2 is the paracellular movement of chloride in the stromal-to-tear direction. Cl3 is the paracellular movement of chloride in the tear-to-stromal direction. Na1 is the sodium efflux mediated by the sodium-potassium ATPase pump. Na2 is the paracellular movement of sodium in the tear-to-stromal direction, while Na3 is the paracellular movement of sodium in the opposite direction. In open-circuit, the tear-side (apical) bath will have a negative potential relative to that of the basolateral-side bath. Cations will thus flow in the paracellular pathways toward the tear side, while anions will flow in the opposite direction. There is also the possibility that some potassium will move along the paracellular pathways toward the tears. The sodium–potassium–chloride cotransporter in the basolateral membranes drives the transcellular movement of chloride. The net flux of sodium and chloride into the tears results in a net fluid transport across the conjunctiva. Flux relationships are indicated at the bottom of the figure by equations.

isolated rabbit conjunctiva: (1) unidirectional/diffusional flow with tritiated water (3H2O); and (2) net water flow by volumetric procedures. With method 1, the diffusion permeability coefficient, Pdw , is expressed in cm s1 and given by: Pdw ¼ Jdw =A Vw Cw

where, A is the area of the membrane (cm2), Vw is the partial volume of water (cm3 mol1), Cw is the concentration of water (mol cm3), and Jdw is the measured unidirectional H2O flux in cm3 s1. In this case, a two-compartment chamber is used. The tissue is mounted between compartments; 3H2O is added to one side and samples are taken periodically from both

106

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Apical aspect

UTP / INS365

CI

Basolateral aspect

Ca2+

P2Y2

CFTR

K Na

K

Ca2+ cAMP

CLCA2 CI channels?

SK4 channels activated via calmodulin?

CI

PKA

Na K cAMP

K

β2

cAMP-activatable K Epinephrine channels, possibly including KvLQT

Figure 4 Cartoon of the established receptors that stimulate electrolyte and fluid secretion in the stromal-to-tear direction under open-circuit conditions, which are analogous to the in vivo situation. Some transporters present in the epithelium have been omitted for clarity. The specific channel subtypes activated by calcium and cAMP remain unverified. Abbreviations: UTP (uridine 50 -triphosphate); INS365 (a synthetic activator of the P2Y2 receptor); P2Y2 (a receptor subtype for the pyrimidine, UTP, that triggers calcium signaling via the phosphoinositide pathway); CLCA2 (a calcium-activated chloride channel subtype thought to be present in the conjunctiva); CFTR (cystic fibrosis transmembrane conductance regulator, which mediates a cAMP-gated chloride conductance via PKA); SK4 (a potassium channel subtype thought to be present in the conjunctiva that is activated by calcium, possibly via calmodulin, a calciumregulatory protein); KvLQT (a potassium channel protein subunit linked to cAMP-activated potassium conductances). The presence of the above channel subtypes has been suggested from the expression levels of message in gene microarray assays of human conjunctiva, but the operational existence of the putative channel proteins has not yet been confirmed in functional experiments.

sides to determine the diffusion of 3H2O, which is proportional to Jdw. With method 2, the osmotic permeability coefficient, Pf , is expressed in cm s1 and can be calculated from the following expression: Pf ¼ Jv =A Vw DCs

where Jv (the measured net H2O flux) is expressed in cm3 s1 and DCs is the difference in solute concentration (mol cm3). For this approach, a two-compartment arrangement could also be used, along with a graduated capillary tube, or appropriate detection system, in order to directly measure Jv as a function of time. Unidirectional fluxes of water determined with 3 H2O (method 1) are usually large and similar in both directions. Thus, a small difference (the net volumetric flow, or Jv, which is detected directly with method 2) is difficult to detect by method 1 and is usually not calculated as a difference between two unidirectional fluxes. For example, in the case of the conjunctival epithelium, unidirectional water fluxes (Jdw) across the tissue are statistically identical in either direction, and have a magnitude 60-fold larger than the reported values for the net flux

(Jv) of fluid secreted to the tear side by the isolated conjunctiva (4–6 ml h1 cm2, using method 2 (a volumetric approach). Because of this discrepancy in magnitude, it is unfeasible (if not impossible) to calculate Jv as the difference between the two, relatively large, unidirectional fluxes in the opposite directions. However, method 1 (a diffusional approach) is useful for determining the effects of agents or various experimental conditions on water permeability (Pdw); because although labeled water will cross cell membranes via all available pathways – lipid bilayer, aquaporins, and other channels, the measurements of Jdw, which reflect Pdw, change equally in both directions when an experimental maneuver changes the water permeability of the epithelium. From diffusional water fluxes (Jdw) and mannitol fluxes it was determined that the conjunctival apical surface is highly permeable to water, and that the transepithelial water permeability (104 cm s1) exceeded the paracellular permeability (106 cm s1). A recently described element contributing to the water permeability of the apical surface is the water channel homolog known as aquaporin type 5 (AQP5). Generally, epithelia exhibit distinct AQPs in the apical and basolateral domains, and in the case of the conjunctiva, AQP3 is expressed in the lateral membranes. Together, AQP5 and AQP3 may be necessary in the conjunctiva for transepithelial fluid transport. AQP5 could serve as a potential target for pharmacological upregulation to enhance fluid secretion given that cAMP via PKA activity has been reported to increase the expression levels of this water channel at both transcriptional and posttranscriptional levels in other cell systems. From measurements of Jv, a spontaneous fluid transport across the isolated conjunctival epithelium in the basolateral-to-apical direction has been described, a property consistent with the more dominant Cl secretory activity of the tissue. As noted, the reported fluid secretion rates were 4–6 ml h1 cm2. This flow was dependent upon transepithelial electrolyte transport given its abolition by ouabain, sensitivity to K+ channel blockade, and Cl dependency. In addition, in experiments that increased the Na+ absorptive activity by raising the glucose concentration (to 25 mM) of the apical bath, fluid transport was inhibited by 77%; an inhibition that did not occur with a similar concentration of mannitol. Studies found that the fluid transport rate was increased (50–100 %) by Cl secretogogues that included purinergic agonists acting via P2Y2 receptors. As purinergic agonists stimulate mucin secretion by conjunctival goblet cells, it seems plausible that the roles of epithelial Cl transport include the hydration of mucins upon release. Overall, the conjunctival epithelium has sufficient water permeability and the transporters necessary to contribute significant fluid to the tear film (50 ml h1 based upon its total surface area). This level of fluid flow is sufficiently large that it may represent a baseline tear

Overview of Electrolyte and Fluid Transport Across the Conjunctiva

secretion beyond that contributed by the lacrimal gland, which mediates reflex tearing under neuronal control. It is not yet clear if the innervation of the conjunctiva directly regulates the rate of fluid transported across the conjunctiva in vivo. However, the transport systems of the conjunctiva can potentially be manipulated pharmacologically. See also: Antigen-Presenting Cells in the Eye and Ocular Surface; Cornea Overview; Corneal Angiogenesis; Imaging of the Cornea; Stem Cells of the Ocular Surface.

Further Reading Anderson, J. M. (2001). Molecular structure of tight junctions and their role in epithelial transport. News in Physiological Sciences 16: 126–130. Bron, A., Tripathi, R., and Tripathi, B. (eds.) (1997). Wolff’s Anatomy of the Eye and Orbit, 8th edn. London: Chapman and Hall. Candia, O. A. (2004). Electrolyte and fluid transport across corneal, conjunctival and lens epithelia. Experimental Eye Research 78: 527–535.

107

Dartt, D. A. (2004). Control of mucin production by ocular surface epithelial cells. Experimental Eye Research 78: 173–185. Gipson, I. K. and Argu¨eso, P. (2003). Role of mucins in the function of the corneal and conjunctival epithelia. International Review of Cytology 231: 1–49. Hosoya, K., Lee, V. H., and Kim, K. J. (2005). Roles of the conjunctiva in ocular drug delivery: A review of conjunctival transport mechanisms and their regulation. European Journal of Pharmaceutics and Biopharmaceutics 60: 227–240. Kaufman, P. L. and Alm, A. (eds.) (2003). Adler’s Physiology of the Eye, Clinical Application, 10th edn. St. Louis, MS: Mosby. Li, H., Sheppard, D. N., and Hug, M. J. (2004). Transepithelial electrical measurements with the Ussing chamber. Journal of Cystic Fibrosis 3(supplement 2): 123–126. Nichols, K. K., Yerxa, B., and Kellerman, D. J. (2004). Diquafosol tetrasodium: A novel dry eye therapy. Expert Opinion on Investigational Drugs 13: 47–54. Oen, H., Cheng, P., Turner, H. C., Alvarez, L. J., and Candia, O. A. (2006). Identification and localization of aquaporin 5 in the mammalian conjunctival epithelium. Experimental Eye Research 83: 995–998. Tiffany, J. M. (2008). The normal tear film. Developments in Ophthalmology 41: 1–20. Ussing, H. H. (1949). Transport of ions across cellular membranes. Physiological Reviews 29: 127–155. Wei, Z. G., Sun, T. T., and Lavker, R. M. (1996). Rabbit conjunctival and corneal epithelial cells belong to two separate lineages. Investigative Ophthalmology and Visual Science 37: 523–533.

Conjunctival Goblet Cells R R Hodges and D A Dartt, Schepens Eye Research Institute, Boston, MA, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Dry eye – A multifactorial disorder of the tear film characterized by either decreased tear production or increased tear evaporation. Glycosylation – The process by which a carbohydrate is added to protein. Goblet cells – Specialized epithelial cells that secrete mucins. Mucins – Large, highly glycosylated proteins. Signal transduction – The process by which a cell converts a signal or stimulus from outside the cell into a functional change.

Goblet cells are columnar epithelial cells that synthesize and secrete mucins, for example, the gel-forming mucin MUC5AC. These cells were originally termed goblet cells because of their distinctive goblet-like shape (Figure 1). The basal portion is narrow and shaped like the stem of a goblet containing the nucleus and organelles, while the apical portion of the cell is shaped like a cup due to the presence of numerous secretory granules. Goblet cells are found in all wet-surfaced epithelia such as the respiratory, gastrointestinal, and reproductive tracts, and the conjunctiva and are surrounded by stratified squamous epithelial cells. Goblet cells of the respiratory, gastrointestinal, and reproductive tracts have been extensively studied with regard to secretion and proliferation in healthy and diseased states. Conjunctival goblet cells have not been studied as extensively, but much information is available. The purpose of this article is to examine the current knowledge of conjunctival goblet cells.

Goblet Cell Development In humans, conjunctival goblet cells begin to appear in the eighth to ninth week of gestation in the region of the lid margin. By the 11th to 12th week, mature goblet cells can be seen containing secretory granules in the palpebral conjunctiva. Goblet cells appear in the bulbar conjunctiva around the 20th week. In the developing chick, goblet cells appear 2 days after hatching in the fornix and 3 days after hatching in the palpebral and bulbar conjunctivae. In rats, messenger ribonucleic acid (mRNA) expression for the conjunctival goblet cell-specific mucin, MUC5AC, first appears 1 day after birth, before eyelid

108

opening. This expression was detected in cells in the fornix region. By day 13 (before eyelid opening), a few round-shaped cells in the fornix expressed cytokeratin 7 (a cytokeratin expressed in goblet cells of adult rats) and were positive for alcian blue/periodic acid Schiff ’s (AB/PAS) reagent, a stain that that binds to sialomucins synthesized and secreted by the goblet cells. These cells also bind Ulex europaeus agglutinin I (UEA-I), which is a lectin that binds specific glycoprotein moieties present in adult goblet cells of the rat. In humans, the lectin Helix pomatia agglutinin (HPA) specifically binds to the goblet cell secretory products. On day 17, single and small clusters of these cells were seen distributed in the fornix and palpebral regions. Goblet cell clusters were seen by day 60. In addition to the appearance of the secretory product, by day 17, immunofluorescence microscopy showed nerves surrounding the basolateral portion of goblet cell clusters and the muscarinic receptor subtypes M2 and M3 and b1- and b2-adrenergic receptors were located on the goblet cells. The presence of nerves and neurotransmitter receptors on goblet cells around the time of eyelid opening (12–15 days after birth) implies that mucin secretion is regulated as the eyes open (Figure 2). There is considerable evidence that stem cells of the conjunctival epithelium, including goblet cells, are distinct from the stem cells of the cornea. This is despite the fact that conjunctival cells can, in the case of a corneal wound involving the limbus, rapidly migrate over the wound and eventually form an epithelium similar to the normal cornea that does not contain goblet cells. However, when bulbar, fornical, and palphebral conjunctivae were grown separately, each with 3T3 feeder cells, they did not express the cornea-specific cytokeratin pair of K3/K12. In addition, when cultured conjunctival cells were injected into athymic mice, the epithelial cyst formed contained epithelial and goblet cells. Given these data, it is thought that conjunctival stem cells are different from stem cells of the cornea. It is well established that corneal stem cells reside in the limbus, the area between the cornea and bulbar conjunctiva. The location of conjunctival stem cells is not as clear. In the mouse and rabbit, it is believed that the stem cells reside in the fornix based on the fact that slowcycling cells are clustered in the fornix, and these cells do not incorporate tritiated thymidine or bromodeoxyuradine that label dividing cells. In addition, cells in the fornix have the highest rate of proliferation in vitro. Cells grown from the bulbar conjunctiva and fornix had the same proliferative capacity as cells grown from

Conjunctival Goblet Cells

Figure 1 An electron micrograph of rat conjunctival goblet cells. Numerous secretory vesicles can be seen in the apical portion of the cells, while nuclei can be seen in the apical portion. Magnification 6000. Reprinted from Dartt, D. A. Regulation of mucin and fluid secretion by conjunctival epithelial cells. Progress in Retinal and Eye Research 21: 555–576. With kind permission of Elsevier.

Figure 2 Co-localization of the M3 muscarinic receptor with goblet cells from the developing rat conjunctiva. Sections from a 17-day-old rat conjunctiva were incubated with an antibody against M3 muscarinic receptor (shown in green) and the lectin UEA-I conjugated to rhodamine (shown in red) which is specific for rat goblet cells. Goblet cells were visualized with differential interference contrast microscopy. Arrows indicate the presence of M3 muscarinic receptor subjacent to the secretory vesicles (shown in red). Magnification 200.

the corneal limbus. Cells from all areas of the fornix and bulbar conjunctiva were capable of undergoing 80–100 cell divisions before reaching senescence, similar to limbal cells. Interestingly, goblet cells were present during the entire life span of the cultures of fornix and bulbar conjunctivae and the number of goblet cells increased during cultivation of the cell cultures. As goblet cells were present during the entire culture time, it seems likely that conjunctival stem cells are bipotent, that is, capable of differentiating into either goblet or stratified squamous cells. It is not known what causes the stem cell to differentiate

109

into either cell type. Possible explanations include genetic programming as it was demonstrated that conjunctival cells with high proliferative capacity differentiate into goblet cells at specific times in their cell cycle of duplication. The first goblet cells appear after a cell has achieved 45–50 doublings and, subsequently, another 10–20 doublings before senescence occurs. It is also possible that environmental effectors such as cytokines or growth factors play a role in the differentiation of a goblet cell. The number of goblet cells in the adult conjunctiva varies depending upon the location in the conjunctiva and species. For example, in the rabbit: (1) the tarsal conjunctiva contains the highest number of goblet cells while the bulbar conjunctiva contains the least, (2) the goblet cells of the tarsal conjunctiva were larger than those present in the bulbar conjunctiva, and (3) conjunctival goblet cells appeared as single cells interspersed throughout the stratified squamous epithelial cells. In contrast, in the rat: (1) the fornix contained the most goblet cells while very few were seen in the bulbar and limbal conjunctivae and (2) goblet cells appear as clusters with as many as 10 goblet cells present in some clusters in the fornix (Figure 3). Goblet cells in the human conjunctiva tend to appear as single cells (Figure 4), similar to the rabbit, though in certain areas, the density of goblet cells is high enough such that they appear to be clustered. The highest goblet cell density in human conjunctiva is in the inferior palpebral conjunctiva. It has been proposed that clusters of goblet cells, such as those that occur in rats, arise from the fact that goblet cells divide several times before senescence giving rise to the clusters. In contrast, in species such as rabbit and human, where goblet cells occur as single cells in the conjunctiva, differentiation into goblet cells occurs after the cell has undergone terminal differentiation.

Function of Conjunctival Goblet Cells The tear film is a thin layer of fluid that covers the ocular surface. Tears are secreted in response to decreased humidity, bright light, mechanical stimulation, bacterial and viral pathogens, and other environmental factors. The tear film is a stratified fluid layer consisting of three layers: (1) the outermost, which is a lipid layer secreted by meibomian glands of the upper and lower eyelids and is thought to be a barrier to evaporation; (2) the middle, which is an aqueous layer secreted essentially by the main and accessory lacrimal glands and contains water, electrolytes, and proteins such as growth factors and antibacterial proteins necessary for the health of the ocular surface; and (3) the innermost, which is a mucin layer, containing mucins that are not only secreted mainly by the conjunctival goblet cells, but also the stratified squamous cells of the cornea and conjunctiva. The mucin layer moves freely over the ocular surface toward the nasolacrimal

110

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Figure 3 Goblet cell clusters in rat conjunctiva. Rat conjunctival sections were incubated with the EGF receptor (shown in red) the lectin UEA-I conjugated to fluorescein isothiocyanate (FITC; shown in green) which is specific for rat goblet cells (left). Flat mounts of rat conjunctiva were prepared and stained with AB/PAS (right). Goblet cells appear as pink-to-purple cells. These micrographs indicate the clusters of goblet cells in rat conjunctiva. Magnification 200.

Figure 4 Goblet cell in human conjunctiva. Human conjunctival sections were incubated with an antibody specific to cytokeratin 7 (shown in red) which is specific for goblet cells (left). Impression cytology samples of human conjunctiva were prepared and stained with AB/PAS (right). Arrows indicate goblet cells. These micrographs indicate that goblet cells occur as single cells. Magnification 200.

duct with the blink providing the mechanism of movement. As it moves, it traps debris and pathogens. The function of conjunctival goblet cells is to synthesize and secrete mucins onto the ocular surface. Mucins are large, highly glycosylated proteins containing tandem repeats of amino acids that are rich in serine and threonine. Owing to the large amount of glycosylation, mucins are highly negatively charged molecules that are believed to be a barrier to pathogens. The glycosylation moieties are quickly hydrated upon exiting the cell, causing them to swell. Thus, mucins provide lubrication, water retention, and a barrier to infectious agents. Members of the mucin family can be subdivided based on whether they are secreted from the cell (secretory) or remain associated with the plasma membrane (membrane associated). While other members of the mucin family, namely MUC1, MUC2, MUC4, and MUC16, are present in the conjunctiva, only the gel-forming, secretory mucin MUC5AC has been identified in conjunctival goblet cells.

MUC5AC is a large gel-forming mucin that is closely related to the other gel-forming mucins MUC2, MUC5B, and MUC6. It contains four cysteine-rich domains (D domains), a tandom repeat region that can be duplicated 17–124 times, as well as a cysteine knot along with an additional cysteine region. The D domains, based on their large number of cysteines, form disulfide bridges with other MUC5AC molecules to form a large gel-like association of mucin molecules (Figure 5).

Control of Goblet Cell Proliferation and Mucin Secretion It is vital for the mucin in the tear film be of sufficient quantity and quality. The quantity of mucin depends on the: (1) number of goblet cells present (proliferation or differentation), (2) amount of mucin synthesized and

Conjunctival Goblet Cells

stored in secretory granules, (3) rate of mucin secretion, and (4) rate of mucin degradation. There are no studies on the rate of mucin synthesis or mucin degradation.

and HB-EGF increased goblet cell proliferation through activation of the EGF receptor. There are four types of EGF receptors, namely ErbB1 (the EGF receptor), ErbB2 (HER), ErbB3, and ErbB4. Each receptor binds to specific members of the EGF family. Activated EGF receptors form homo- or heterodimers and then recruit adaptor molecules including Shc/Grb2, phosphoinositide-3 kinase (PI-3K), phospholipase C (PLC)g, p38 mitogen-activated protein kinase (MAPK), and c-jun NH (2)-terminal kinase ( JNK). Each of these kinases initiates a cascade of kinases leading to a cellular response. The downstream kinase activated by Shc/Grb2 is the extracellular-related kinase 1/2 (ERK1/2); the downstream kinase activated by PI-3K is AKT; the downstream kinase activated by PLCg is protein kinase C (PKC); and the downstream kinase activated by JNK is c-jun (Figure 6). Further investigation of the role of ERK in EGFstimulated conjunctival goblet cell proliferation demonstrated that EGF increased the number of cells expressing ERK in their nucleus in a biphasic manner. Under basal conditions, ERK is present in the cytosol. Upon stimulation, ERK translocates to the nucleus where it phosphorylates proteins necessary for proliferation. The first, major peak occurs 1 min after the addition of EGF to cultured rat goblet cells. The number of goblet cells with ERK present in the nucleus returned to basal levels before increasing again approximately 18 h after the addition of EGF. The second peak corresponded with the appearance

Goblet Cell Proliferation With the development of a method to culture conjunctival goblet cells, studies of goblet cell proliferation are now underway. One important stimulus of goblet cell proliferation is the epidermal growth factor (EGF) family as measured by both a cell proliferation assay kit and by immunoflouresence microscopy using the an antibody against Ki-67, a protein known to be present in cells that have entered the cell cycle. EGF, transforming growth factor a (TGFa), heparin-binding EGF (HB-EGF), and heregulin are present in rat conjunctiva as well as cultured rat goblet cells as determined by reverse transcriptase polymerase chain reaction (RT-PCR), Western blot analysis, and immunofluorescence microscopy. EGF, TGFa,

Muc5AC

D domain Cysteine domain Repetitive domain Figure 5 A schematic of MUC5AC.

EGF

111

EGF

EGF receptor Ras

PIP2 PLCγ

MEKK-1

Raf

MKK-7

MEK

JNK

ERK 1/2

Sos

Grb2 Shc

IP3

DAG

PI-3K Ca2+ AKT

PKC 2+

Ca

?

ERK 1/2 C-jun

Transcription factors Transcription factors

Proliferation

Figure 6 A schematic representation of the EGF pathway leading to goblet cell proliferation. EGF receptor dimerizes upon EGF binding, recruiting the adaptor molecules leading to cell proliferation. PI-3K, phosphoinositide-3 kinase; PLCg, phospholipase Cg; JNK, c-Jun NH (2)-terminal kinase; ERK1/2, extracellular-related kinase 1/2; PKC, protein kinase C; PIP2, phosphatidylinositol bisphosphate; IP3, inositol trisphosphate; DAG, diacylglycerol; PKC, protein kinase C; MEK, mitogen-activated kinase kinase; MEKK-1, mitogenactivated kinase kinase-1; MKK-7, mitogen activated kinase kinase-7.

112

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

of Ki-67, indicating that goblet cells were starting to proliferate. The ERK inhibitor U0126, added 20 min prior to EGF, inhibited both peaks of the translocation of ERK as well as goblet cell proliferation. Interestingly, inhibition of the second peak of ERK translocation with U0126 also prevented EGF-stimulated proliferation. In addition to the activation of ERK, EGF also activates PKC isoforms -a and -e to, in turn, stimulate proliferation. Most proliferation studies to date have been performed on cultured goblet cells from rat. Several studies have examined the proliferation of human goblet cells in comparison to rat cells, and demonstrated that human goblet cells respond similarly to EGF. For example, EGF stimulates human goblet cell proliferation in a similar time- and concentration dependency. In both cell types, EGF activates ERK and PKC to stimulate proliferation. Due to the difficulty of obtaining a human conjunctiva and culturing human goblet cells, rat goblet cells are an excellent model for studying human goblet cell proliferation. Goblet Cell Mucin Secretion Goblet cell secretion occurs through an apocrine mechanism. In this mechanism, most or all the secretory vesicles fuse with one another upon stimulation and subsequently with the apical membrane releasing the mucin into the extracellular space. Therefore, the amount of mucin released is dependent upon the number of cells responding to a given stimulus. A strong stimulus would involve responses from more goblet cells than a weak stimulus. Taking advantage of this property, conjunctival samples can be treated with histochemical stains that recognize mucins in goblet cells, and the number of goblet cells can be counted. A decrease in the number of goblet cells would indicate an increase in secretion. Another method to measure conjunctival goblet cell secretion is based on the fact that the lectins UEA-I and HPA bind to specific carbohydrate residues found in MUC5AC, depending on the species. The amount of secreted mucin can be determined by an enzyme-linked lectin assay (ELLA) and Western blot or dot-blot analyses. Using the histochemical method of staining for mucins in a conjunctival button, it was demonstrated, for the first time, that goblet cell mucin secretion is neurally mediated, as a wound to the central cornea induced goblet cell mucin secretion. Sensory nerves in the cornea were activated by the wound causing a neural reflex arc in which the parasympathetic nerves that surround the goblet cells released their neurotransmitters to stimulate mucin secretion. These results were confirmed when parasympathetic nerves containing the neuropeptide vasoactive intestinal peptide (VIP) were demonstrated to be present in the conjunctiva surrounding the goblet cells subjacent to the secretory granules. Addition of exogenous VIP or

the cholinergic agonist carbachol (an analog of the parasympathetic neurotransmitter acetylcholine) stimulated goblet cell secretion. Other compounds which have been shown to stimulate conjunctival mucin secretion include activators of the purninergic receptor subtype P2Y2, such as ATP, UTP, and INS365, the neurotrophins nerve growth factor, and bone-derived neurotrophic factor, and the drug OPC-12759, an antigastric ulcer drug. Interestingly, no sympathetic neurotransmitter has been shown to stimulate goblet cell secretion despite the presence of the receptors for these neurotransmitters on goblet cells. The signal transduction pathways utilized by cholinergic agonists have been well studied. Cholinergic agonists bind to M2 and M3 muscarinic receptors which are present on conjunctival goblet cells. Classically, these receptors activate PLC, which hydrolyzes phosphatidylinositolbisphosphate into 1,4,5-inositol trisphosphate (IP3) and diacylglycerol (DAG) (Figure 7). IP3 induces the release of Ca2þ from the endoplasmic reticulum into the cytosol. It is not known if this occurs in goblet cells; however, what is known is that an increase in intracellular Ca2þ alone is sufficient to cause mucin secretion from conjunctival goblet cells. Calcium can also activate Ca2þ-/calmodulindependent protein kinase. However, inhibitors of Ca2þ/ calmodulin protein kinase do not have any effect on cholinergic agonist-stimulated mucin secretion. DAG is a phospholipid activator that along with Ca2þ activates PKC. Though a direct role for PKC involvement in conjunctival goblet cell mucin secretion has not been demonstrated, activators of PKC (phorbol esters) do stimulate secretion. In addition to the Ca2þ pathway, cholinergic agonists in goblet cells activate a second pathway leading to protein secretion. This pathway involves the transactivation of the EGF receptor through the stimulation of the focal adhesion kinase Pyk2. Pyk2, in turn, activates the nonreceptor tyrosine kinase p60Src. Pyk2 and p60Src are activated in goblet cells by Ca2þ and PKC as the addition of either a calcium ionophore or a phorbol ester increases phosphorylation (activation) of Pyk2 and p60Src and the use of PKC inhibitors inhibits cholinergic agonist-stimulated Pyk2 and p60Src phosphorylation (Figure 7). This again implies that PKC plays a vital role in conjunctival goblet cell secretion. The Pyk2/p60Src complex then transactivates the EGF receptor, recruiting the adaptor molecules Shc, Grb2, and the Ras guanine nucleotide exchange factor Sos. Sos binds to and activates the low molecular weight guanosine triphosphate (GTP)ase, Ras, causing the exchange of guanosine diphosphate (GDP) for GTP to activate Ras. Ras initiates another kinase cascade of Raf (also known as MAPK kinase kinase), mitogen-activated kinase kinase (MEK) (MAPK kinase), and ERK 1/2 (p44/p42 MAPK). Upon stimulation by cholinergic agonists, ERK activates proteins in the cytosol, leading to secretion. Inhibition of ERK inhibits cholinergic agonist-stimulated mucin secretion (Figure 7).

Conjunctival Goblet Cells

113

Parasympathetic nerves Ca2+

ACh Muscarinic receptors ? Gαq/11

EGFR EGFR RAS

PLC IP3

Ca2+

Ca2+

DAG

Pyk2

TK

TK

Src P Shc Grb2 Sos Raf

PKC

MEK Ca2+ ? ?

Secretion

ERK 1/2

Secretion

Secretion

Secretion

Figure 7 A schematic representation of the cholinergic pathway leading to goblet cell mucin secretion. Muscarinic receptors activate phospholipase C (PLC) to generate the production of inositol trisphosphate (IP3), which releases intracellular Ca2þ and diacylglycerol (DAG), which activates protein kinase C (PKC). The EGF receptor (EGFR) is transactivated leading to the activation of Ras, Raf, mitogen-activated kinase kinase (MEK), and ERK 1/2.

In contrast, when ERK is activated by exogenous EGF, ERK 1/2 translocates to the nucleus where it stimulates proliferation. The signaling pathways of the other compounds that stimulate mucin secretion are not well studied. Receptors for VIP (VPAC1 and VPAC2) are present on goblet cells both in vivo and in vitro. It is not known if exogenous VIP (which causes mucin secretion) acts through these receptors. As it is well established in other tissues that VIP increases the intracellular cyclic adenosine monophosphate (cAMP) concentration as well as the intracellular [Ca2þ], it is likely that a similar mechanism occurs in goblet cells. It is not know if the EGF family stimulates mucin secretion. As these growth factors activate ERK, which plays an important role in both secretion and proliferation, it is likely that they do. It is interesting that many of the same proteins and kinases are used by EGF to stimulate proliferation and by cholinergic agonists to stimulate mucin secretion. ERK1/2 is known to phosphorylate over 100 proteins. In order to confer specificity, ERK1/2, as is the case for many signaling molecules, is organized through the use of scaffold proteins. These proteins bind to two or more components of a signaling pathway in close proximity to assist with their interactions. Scaffolding proteins also target signaling molecules to particular areas of the cell in order to phosphorylate specific substrates. They can prevent cross talk between pathways. As the scaffolding proteins are also regulated, their stability can affect the duration of the signal. Several different scaffolding proteins have been identified in mammalian cells including kinase suppressor of Ras, MEK partner-1,

Morg1, IQGAP1, and b-arrestins 1 and 2. Not surprisingly, scaffolding proteins are required for the translocation of ERK1/2 to the nucleus and other proteins are responsible for maintaining ERK1/2 in the cytosol. In fact, ERK1/2 translocation to the nucleus requires phosphorylation by MEK, which causes conformational changes in ERK1/2, allowing for the formation of ERK dimers. This dimerization facilitates the translocation of phosphorylated ERK1/2 into the nucleus. In contrast, PEA-15, a small phosphoprotein, has been shown to anchor ERK1/2 in the cytosol.

Clinical Implications of Mucin Deficiency on the Ocular Surface The presence of goblet cells in the conjunctiva is of great importance to the health of the ocular surface. Dry eye, which is characterized by a deficient tear film or excessive evaporation, is a multifactorial disorder that causes damage to the ocular surface. One characteristic of dry eye is a decrease in the number of mucin-containing goblet cells. All in vivo studies to date identify goblet cells by their secretory product, either by staining with AB/PAS or by the presence of MUC5AC within the cell, and should be termed filled goblet cells. Thus, a decrease in the number of filled goblet cells in the conjunctival tissue indicates that goblet cells have secreted. Under chronic conditions, the decrease in filled goblet cells could indicate repeated stimulation such that mucin synthesis is unable to keep pace with secretion or indicate a loss of goblet cells through either a decrease in goblet cell proliferation,

114

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

dedifferentiation, or an increase in goblet cell death. It is difficult to know whether the loss of goblet cells is a cause of the disease or a result of it. Indeed, the mechanism by which goblet cells are lost from the conjunctiva has not been studied. In animal models of dry eye, proteins involved in apoptosis have been shown to be upregulated in conjunctival epithelial cells, though the effects on the goblet cells themselves were not clear. The loss of goblet cells in dry eye patients warrants further examination. The presence or absence of conjunctival goblet cells has been examined in different types of dry eye. These have been described below.

Sjo¨gren Syndrome Sjo¨gren syndrome is an autoimmune disease characterized by profound lymphocytic infiltration of the lacrimal and salivary glands, resulting in dry eye and dry mouth. It is not known whether the changes seen in the conjunctiva are a result of the autoimmune disease itself or as a result of the change in tears due to lacrimal gland destruction. Though the changes in the conjunctiva are relatively mild compared to the destruction of the lacrimal glands, there is a decrease in the number of mucin-containing goblet cells in patients with Sjo¨gren syndrome, as determined by histochemical staining methods. In addition, the amount of MUC5AC mRNA is decreased in patients with Sjo¨gren syndrome and the amount of MUC5AC detected in tears of these patients is also decreased.

Vitamin A Deficiency Vitamin A is vital for the health of the ocular surface as it is essential for the development of goblet cells in the conjunctiva. Vitamin A deficiency has been shown to cause keratinization and squamous metaplasia of the conjunctiva. The lack of goblet cells in the conjunctiva in vitamin-A-deficient patients has been documented through the use of impression cytology and histochemical staining methods. It has also been demonstrated that in vitamin-A-deficient rats, not only were there no mucincontaining goblet cells, but the mRNA for MUC5AC and MUC5AC protein also disappeared 20 weeks after the initiation of a vitamin-A-deficient diet. Reintroduction of vitamin A into the diet has been shown to increase the number of conjunctival mucin-containing goblet cells.

Topical Preservatives Preservatives, such as benzalkonium chloride, are often found in artificial tears and medications such as those used to treat glaucoma and anti-inflammatory medications. These patients often report the signs and symptoms of dry

eye. Benzalkonium chloride has been shown to decrease the number of mucin-containing goblet cells present in the conjunctiva. Ocular Cicatrical Pemphigoid Mucous membrane pemphigoid is an autoimmune disease that is characterized by blisters in the mucous membranes of the body including the mouth, nose, trachea, and conjunctiva. When the conjunctiva is involved, it is known as ocular cicatrical pemphigoid (OCP). This condition causes chronic conjunctivitis and, eventually, complete keratinization of the conjunctiva, resulting in severe dry eye. It has been observed that the mucin-containing goblet cell number is reduced in the later stages of the disease, likely as a result of the keratinization. Interestingly, it has been shown that the expression of a specific glycosyltransferase present exclusively in goblet cells in normal human conjunctiva is altered in late stages of OCP. This enzyme, which is responsible for the glycosylation of mucins, is expressed in stratified squamous cells in areas of the conjunctiva that were nonkeritinized in OCP patients, and, as the disease progressed, the expression disappeared. Laser-assisted In Situ Keratomileusis Patients undergoing laser-assisted in situ keratomileusis (LASIK) often experience dry eye symptoms. These symptoms are usually temporary, but can develop into chronic dry eye. It is known that the number of mucincontaining goblet cells decreases significantly within 1 week postsurgery, but returns to preoperative levels by 3 months after surgery. Diseases and disorders of goblet cells are a result not only of mucous underproduction, but also of mucous overproduction. Ocular Allergies The symptoms of allergic conjunctivitis include mucus production, ocular itching, foreign-body sensation, tearing, hyperemia, chemosis, and lid edema. Traditionally, allergic eye disease has been classified into: seasonal allergic conjunctivitis (SAC), perennial allergic conjunctivitis (PAC), vernal keratoconjunctivitis (VKC), atopic keratoconjunctivitis (AKC), and contact-lens-induced papillary conjunctivitis (CLPC). Each of the five categories of allergic conjunctivitis has a distinct pathology. In SAC and PAC, conjunctival inflammation is mild and of short duration. In VKC and AKC, conjunctival inflammation has an unclear history of exposure to allergen, is more severe, and lasts longer than SAC and PAC. A common symptom of allergic conjunctivitis in the human is alteration in mucus production. One hypothesis as to how mucous production is altered in allergic conjunctivitis is that activation of sensory nerves in the cornea

Conjunctival Goblet Cells

and conjunctiva, manifested by itchiness, foreign-body sensation, and increased tearing, could occur. This, in turn, could activate the efferent parasympathetic and sympathetic nerves that surround conjunctival goblet cells to release the neurotransmitters acetylcholine and VIP, which are known to stimulate conjunctival goblet cell secretion. When goblet cells secrete, all secretory granules are released at once. Under these chronic conditions, the decrease in filled goblet cells could indicate repeated stimulation with mucin synthesis unable to keep pace with secretion. In support of this, a decrease in filled conjunctival goblet cells in VKC and AKC patients, respectively, compared to control subjects, has been noted. In addition, the amount of MUC5AC RNA in the conjunctiva was decreased in AKC and VKC patients versus control. The excess mucus seen on the conjunctiva in individuals with VKC is a stringy mucus and does not necessarily represent goblet cell MUC5AC, but could be stratified squamous cell membrane-spanning mucins, poorly hydrated mucins, or mutated mucin variations. Additional study of the role of goblet cell mucin secretion in different types of ocular allergy is warranted. Results from animal models of allergic conjunctivitis, which most closely mimics SAC and PAC, found that the number of filled goblet cells decreased for 6 h after the final allergen challenge. Over 48 h, the number of filled goblet cells returned toward control values, indicating that goblet cell mucin secretion increased during delayed hypersensitivity, but that goblet cells refilled after allergen removal. MUC5AC RNA was also depleted by the final challenge and recovered at 6 h, indicating that the delayed hypersensitivity depletes MUC5AC RNA; however, it recovers quickly to begin synthesizing MUC5AC to refill the goblet cells.

Summary Much information has been gathered regarding the role of conjunctival goblet cells in the health of the ocular surface. Goblet cell proliferation, mucin synthesis, and secretion control the amount of mucin present on the ocular surface. These processes are regulated as growth factors increase proliferation and neural stimulation causes mucin secretion. However, many questions remain. Are the three processes that regulate mucin amount coordinated so that

115

one stimulus activates all three? While some of the intracellular pathways leading to proliferation or secretion are known, it is not known if these pathways are altered in diseases. The life span of a conjunctival goblet cell as well as the mechanisms by which goblet cells are lost in disorders of the tear film are unknown. It is also not known if goblet cells are able to secrete mucins multiple times, perhaps refilling secretory granules, in response to either a sustained signal or multiple stimuli. Further research into these questions may lead to the development of treatments for many of the ocular surface diseases. See also: Defense Mechanisms of Tears and Ocular Surface; Ocular Mucins; Overview of Electrolyte and Fluid Transport Across the Conjunctiva.

Further Reading Dartt, D. A. (2004). Control of mucin production by ocular surface epithelial cells. Experimental Eye Research 78: 173–185. Gipson, I. K. and Argueso, P. (2003). Role of mucins in the function of the corneal and conjunctival Epithelia. International Review of Cytology 231: 2–49. Gipson, I. K., Hori, Y., and Argueso, P. (2004). Character of ocular surface mucins and their alteration in dry eye disease. Ocular Surgery 2: 131–148. International Dry Eye Workshop (2007). The definition and classification of dry eye disease. Report of the definition and classification subcommittee of the International Dry Eye Workshop (2007). Ocular Surgery 5: 75–92. Lavker, R. and Sun, T-T. (2003). Epithelial stem cells: The eye provides a vision. Eye 17: 937–942. Pellegrini, G., Golisano, O., Paterna, P., et al. (1999). Location and clonal analysis of stem cells and their differentiated progeny in the human ocular surface. Journal of Cell Biology 145: 769–782. Ramos, J. W. (2008) The regulation of extracellular signal-regulated kinase (ERK) in mammalian cells. International Journal of Biochemistry and Cell Biology, 40: 2707–2719. Rios, J. D., Forde, K., Diebold, Y., et al. (2000). Development of conjunctival goblet cells and their neuroreceptor subtype expression. Investigative Ophthalmology and Visual Science 41: 2127–2137. Sellheyer, K. and Spitznas, M. (1988). Ultrastructural observation on the development of the human conjunctival epithelium. Graefe’s Archive of Clinical and Experimental Ophthalmology 226: 489–499. Shapiro, M. S., Friend, J., and Thoft, R. A. (1981). Corneal reepithelialization from the conjunctiva. Investigative Ophthalmology and Visual Science 21: 135–142. Shatos, M. A., Rios, J. D., Tepavcevic, V., Kanno, H., Hodges, R. R., and Dartt, D. A. (2001). Isolation, characterization, and propagation of rat conjunctival goblet cells in vitro. Investigative Ophthalmology and Visual Science 42: 1455–1464.

Ocular Mucins M Berry, Bristol Eye Hospital, Bristol, UK ã 2010 Elsevier Ltd. All rights reserved.

Glossary Alternative splicing – A variation mechanism in which linear combinations of exons are translated, resulting in a variety of mature products encoded by a single gene. Atomic force microscopy (AFM) – A technique of high-resolution imaging through the measurement of forces between atoms in the sample and those on the instrument tip. Glycan – The oligosaccharide portion of a glycoconjugate. Glycocalyx – An outer, carbohydrate-rich coating on the surface of cells. Glycoforms – Variations in the amount or compositions of oligosaccharides decorating the same peptide core. Meibomian glands – The special sebaceous glands at the rim of the eyelids that supply the lipid layer of the tear film. Mucins – A family of large, heavily glycosylated molecules, with most glycan chains O-linked to the peptide core. Persistence length – A measure of polymer stiffness; it is the length over which correlations in the direction of the tangent are lost. Reptation – Movement of a long polymer parallel to itself, similar to the movement of a snake. Tandem repeats – The adjacent repetition of a pattern of two or more nucleotides. Worm-like model – A model for the behavior of semiflexible polymers, considered continuously flexible. Young’s modulus – A measure of elasticity, defined as the ratio of stress to strain.

Introduction In addition to the well-defined anatomical blind sac formed by the cornea and conjunctiva, the meibomian and lacrimal glands, the ocular surface comprises a mucosal immune system, rich neural and endocrine loops, as well as the blink reflex. As with other mucosal systems the ocular surface is further integrated into the adaptive immunity of the organism, and into the microbial richness

116

of the outer environment. An additional and specific requirement of the ocular surface is the maintenance of transparency that applies to the cornea as well as to the preocular fluid. Bathing the exposed part of the outer eye is a complex fluid whose elements are secreted by the wet epithelia, and the lacrimal and meibomian glands. Mucins are the main component of a mucus gel and responsible for its viscoelastic properties. They form a dynamic matrix wetted by a plasma dialysate enriched by secretions from the lacrimal glands and topped by a layer of waxes and lipids originating in the meibomian glands. This fluid is periodically sheared and mixed by the movement of the lids during blinking. Underneath the mucous gel, the epithelial glycocalyx anchors the tear film to the ocular surface: mucins and glycoproteins are the major components of this layer.

Mucin Architecture A very rich glycosylation, with most sugar chains O-linked through N-acetylgalactosamine (GalNAc) to serine or threonine in the peptide core, is diagnostic of mucins. Sugar chains tend to be clustered in discrete regions resulting in concentrations of negative charges. In these high-charge regions the peptide core is rich in serine, threonine, and proline (PTS domains), and repeated sequences of aminoacids (variable number of tandem repeats, hence VNTR domains) are present, specific to the encoding gene. Other regions are less richly glycosylated and contain relatively more N-linked glycan chains, the latter necessary for mucin transit through intracellular microtubules during synthesis. In humans, ocular mucin regions of tens of nanometers seem to be almost naked: the molecular diameter measured in liquid with atomic force microscopy (AFM) is not significantly higher than that of aminoacids in a helix (Figure 1(a)), giving mucin polymers the appearance of strings of small beads, small beads of dense bottlebrushes. Toward the N- and C-termini of mucins, the arrangement of moduli is similar to those found in proteins involved in coagulation (von Willebrand factor, VWF domains), cysteine knots, or SEA (sea urchin sperm protein, enterokinase, and agrin) domains. These are useful in tracing the evolution of mucin genes. Genes encoding for PTS domains and multiple VWF domains (D1–D2–D3 PTS, as in secreted mucins) can be found early in the evolution of metazoa, preceding hemostasis or coagulation

Ocular Mucins

A

117

B

50 nm

50 nm

2

C 1 200

0

400

600

800

1000

2

D 1 200

0

400

600

800

1000

(a) MD

D1 D2

D⬘

Cys

MD

D4 B

D3

C CK

(b) Proteolytic cleavage site CT TD SEA

(c) von willebrand domains Cysteine Unique sequence Tandem repeats SEA domain Transmembrane domain TR Cytoplasmic tail CT

Figure 1 Architecture of ocular mucins deposited on mica (a) Portions of a hydrated human conjunctival mucin molecule imaged with AFM (A, B). The respective height profiles (C, D), highlight the short oligosaccharides of ocular mucins. All axes are in nanometres. Reproduced from McMaster, T. J. (1999). Atomic force microscopy of the submolecular architecture of hydrated ocular mucins. Biophysical Journal 77: 533–541. With permission from Biophysical Society. (b) Schematic of a mucin monomer, containing von Willebrand factor domains toward the C- and N-termini (D1, D2, D0 , D3, and D4, B, C, and CK, indicated by green labels) and central mucin domains (yellow). A more detailed schematic of PTS domains reveals unique sequences (brown-filled octagons) and tandem repeat regions (yellow octagons) interspersed with cysteins (blue). Most secreted mucins polymerize by disulfide-bonded linear concatenation of such monomers.(c) Schematic of a cell-surface-associated mucin, containing an SEA domain (sea urchin sperm protein, enterokinase, agrin) within which there is a proteolytic cleavage site, a transmembrane domain (TD), and a cytoplasmic tail (CT).

118

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

(Figure 1(b)). SEA domain mucins appear in vertebrates, while the auto-catalytically cleaved SEA domain is restricted to mammals (Figure 1(c)). Regions where the peptide core is extended by the insertion of the first sugar (GalNAc) and repulsion between negatively charged sugar chains alternate with more flexible polypeptide chain stretches, suggesting that mucins behave in solution like stiff random coils. Mucins occupy a large volume which indicates interpenetration of molecular domains at relatively low polymer concentrations. Using the worm-like model, calculations indicate that human ocular mucins are more flexible than DNA molecules of similar length; mucin persistence length is 35 nm and that of DNA is 50 nm. Polymer conformation and stiffness, for example, of human ocular MUC5AC glycoforms, are greatly influenced by the degree and nature of post-translational glycosylation. Mucin architecture is determinant of mucin role and function at the mucosal surface and in the gel. The organ, developmental state, and physiological status, in turn, affect mucin expression and details of glycosylation. Ocular mucins have short oligosaccharide chains: in humans they are mostly less than six sugars long, negatively charged and terminated in sialic acid, with fucosylation representing less than one-fifth of the sugars (Figures 2 and 3). In dogs and rabbits, glycans are mainly neutral and terminated in fucose and/or GalNAc. The short oligosaccharides might be related to transparency and (relatively fast) turnover of mucins, and terminal sugars to environmental microbiota. Mucin Families Mucin genes appeared through a combination of moduli existing in other proteins. The close connection between

mucin structure and function gives rise to a classification necessarily reflecting both. Surface-associated mucins Some mucins spend part of their life anchored into the apical cell membrane before they are shed into the luminal, that is, tear, fluid. Formerly known as membranebound mucins, they are now called cell-surface-associated mucins. These are heterodimers, with a large mucin subunit outside the cell, a (mostly hydrophobic) membranespanning region and an intracellular tail. A number of subfamilies are represented at the ocular surface: the mammalian-specific MUC1 with its SEA domain; the MUC16 that contains multiple SEA domains, not all of which are cleaved; and the MUC4 that has VWD but neither cysteine-rich domains nor SEA. Shedding of these mucins is thought to cause changes in the neighboring membrane domains, potentially transferring information to the cell interior. A further possibility is that information is conveyed through the cytoplasmic tail to the cytoskeleton. An important result of surface mucin release in the tear fluid is the renewal of the glycocalyx and the tear fluid itself. This group of mucins is heterogeneous and most genes also encode splice variants that are secreted: MUC4, a cell-surface mucin in normal cornea and conjunctiva, is a goblet cell mucin in some pterygia (Figure 4). MUC1/SEC, a splice variant of MUC1 that lacks the transmembrane domain and, therefore, results in a soluble, secreted form of MUC1, is present in human cornea and conjunctiva. Secreted mucins Milliseconds after the secretion of mucins stored in granules, often in specialized epithelial cells, their volume

GalNAc 45

NeuAc NeuAc2 Gal

%

30

GlcNAc Fuc 15 Man Glc 0 A

B

C

D

Increasing negative subunit charge Figure 2 Monosaccharide composition of human ocular mucins. The increase in sialic acids, mono- or di-acetylated (NeuAc, or Neu-Ac2) did not parallel the increase in negative subunit charge (A
Ocular Mucins

119

2AB 2AB

Core 1

Disialyl core 1

re2

Co

2AB Sialyl galactoseP

B

2A

2AB ore

yl c

ial -6 s

1

o2

Mon

2AB

Di 2-3 sialyl galactosyl core 2

Mono 2-3 sialyl core 1 Figure 3 Oligosaccharide composition of human ocular mucins. This pie chart indicates the abundance of the different glycans of purified human ocular mucins obtained by hydrazinolysis. Each sequence is named in the color of its proportion in the total glyco-repertoire, while its structure is indicated in the sugar notation presented below. Notable is the high proportion of sialylated oligosaccharide chains, almost half of which contain a2-3-linked sialic acids. A cautionary note: at present the method used for determining O-linked glycan composition affects the results. Glycochip methods often estimate a higher proportion of core 2 glycans than obtained by hydrazinolysis and HPLC. Notation of sugars in the figure is as follows: ^ N-acetylgalactosamine (GalNAc); ◊ galactose (Gal); N-acetylglucosamine (GlcNAc); $ N-acetylneuraminic acid (NeuAc). Solid lines denote b linkages, dashed lines mark a links.

▪

increases 100-fold in the extracellular fluid as a result of hydration. These secreted mucins have a well-defined VWF-D2-D3 PTS architecture, and Cys-rich domains or Cys knots. VWF domains and Cys are involved in the concatenation of mucin subunits (Figure 5). The secreted mucins – MUC2, MUC5AC, MUC5B, MUC7, MUC19, and MUC23 – have been found at the ocular surface, in cells, impressions, and tears. MUC20 messenger RNA (mRNA) has also been identified in the conjunctiva. With the notable exception of MUC7, secreted mucins form linear polymers of subunits linked by disulfide bonds and form gels, and are thus called secreted gel-forming mucins. Polymers extracted from cells (and protected from proteolysis) may be few microns long; in the secretion, submicron lengths dominate. MUC7 oligomers are like spokes of a wheel around a central, yet to be fully described, entity. Though MUC7 is not gel forming, it is found in gels, for example, saliva, from which it has been originally described.

Biosynthesis and Turnover

Figure 4 MUC4 alternative splicing in pterygium. In normal conjunctival tissue MUC4 is mainly a surface-associated mucin. In pterygia, triangular growths of the conjunctiva over the cornea, it can be present either as surface associated (red arrows) or secreted (yellow arrows) within prominent goblet cells. Where pterygial MUC4 is surface associated, goblet cells do not light up with anti-MUC4 antibodies (green arrows). MUC4 was visualized with antibody 4F12 (DSHB, Yowa, USA); counterstain: hemalum. Images: courtesy of Friedrich Paulsen, Halle University, Germany.

Synthetic pathways

Mucin genes account for nearly 4% of genes expressed in the normal conjunctiva, with a further 29% dedicated to glycosyltransferases. During synthesis, molecular species with varying properties are transported to the location corresponding to their stage in this complex process. MUC5AC might take over 2 h from initiation to storage

in the secretory granule. The making of secreted mucins starts with the synthesis of the peptide core in the endoplasmic reticulum, where N-linked sugars are also added. Following folding of the C- and N-termini, the peptides dimerize through S–S bonds between cysteine knot

120

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Ch

arg

ep

rofi

le -++---- S-S ---++

-++----++---++

vWF domains

-+ ----

Cys

Cys ---++

vWF domains

Mucin domains

Figure 5 Schematic of secreted mucin structure. Secreted mucins have a well-conserved architecture. Within mucin domains, repeated sequence(s) called tandem repeats are gene specific. The concatenation of mucin subunits gives rise to a polymer with concentrated regions of negative charge, interspersed with less charged, less glycosylated regions. vWF, Von Willebrand factor domains; Cys, cysteine.

domains at the C-termini. O-glycosylation is initiated and elaborated in the cis- and medial Golgi. Further polymerization, S–S bonds between VWD3 domains at the N-terminal, occurs in the trans-Golgi and the secretory granule. Cell-surface mucins too acquire their N-linked sugars in the endoplasmic reticulum, where the peptide core is cleaved into the two domains, and further glycosylated as these mucins progress through the Golgi. A signal peptide in the N domain localizes cell-surfaceassociated mucins at the apical membrane; the cytoplasmic domain starts with a signal sequence involved in retention at the plasma membrane and mucin recycling.

Glycosylation

glycan elongation, incorporating a measure of randomness in the glycan population. At the human ocular surface, ppGalNAcTs are distributed in an epithelial layer and in a cell-type-specific manner. Isoform –T3 was detected in all the epithelial layers, while –T2 in restricted to basal cells; –T4 is expressed in all apical cells. One member of the family, ppGalNAcT-6, was found in goblet cells only in normal subjects, but in apical cells of conjunctival epithelia in patients with ocular cicatricial pemphigoid, together with the –T2 isoform. How elongation of oligosaccharide chains is regulated is not yet understood: in humans, rabbit, and dog, ocular mucins have short sugar chains. The same mucin gene product in the respiratory or gastrointestinal mucosa is decorated by chains many tens of sugars long. The ensemble of human ocular surface enzymes is such that mucin oligosaccharides are negatively charged and mostly terminated in sialylated structures, including the histo-blood Lewis group antigens (sialyl Lex and sialyl Lea). The number of different structures is low, barely in double figures, compared with many hundreds in respiratory epithelia from one subject, for example. Histo-blood group antigens are involved in neutrophil adherence and activation. Because of the short chains, sugar epitopes that are cryptic in other mucosae are overt in human ocular mucins, for example, Tn or SialylTn that are richly present in the normal conjunctival epithelium. Which sugars or end groups are exposed in the glycan envelope of the tissue remains to be clarified for the ocular surface. Variation in this envelope is likely to modulate both immune effector cells and bacterial adhesion to the conjunctiva or cornea.

Glycosyltransferases

Mucin glycans are synthesized in the Golgi apparatus. The presence, activity, and localization of glycosyltransferases along the Golgi cisterns are the primary determinants of glycan chain density and sequence. The initial transfer of GalNAc from UDP-GalNAc to the hydroxyl group of Ser/Thr in the peptide backbone is catalyzed by uridine diphosphate GalNAc:polypeptide N-acetylgalactosaminyl transferases (ppGalNAcTs). Subsequent sequential sugar additions are catalyzed by glycosyltransferases (GalNAcTs). Nucleotide-sugar synthases and hydrolases, and genes encoding their transport proteins add further dimensions to the function and regulation of mucin molecules. In other mucosae, for example, respiratory epithelium, epidermal growth factor (EGF), Th2 cytokines, and alltrans retinoic acid alter the expression of GalNAc-Ts, some of which are also downregulated in cancer. Molecular details of transcriptional regulation of glycosylating enzymes are not known for the eye. The final density of glycosylation and chain structure strongly depends on the availability of sugar nucleotides, and competition between enzymes for acceptor intermediates during

Turnover Quantities, species, and glycoform composition of mucin populations on the ocular surface are a dynamic result of mucin production, secretion, and degradation. Not all the factors involved in this process are known: immune and infective agents modulate synthesis and secretion, as do neural stimuli; enzymes in the tears contribute to mucin degradation, while tearing and blinking are believed to remove spent mucus from the surface. There are no data on average half-life of mucins in the preocular fluid. Extraction of large and glycosylated mucin polymers from contact lenses suggests that some escape degradation during waking hours. Recycling

There is evidence that surface-associated mucins are reuptaken into the Golgi and re-glycosylated. In the cell lines where these experiments have been done, the final glycosylation of the mucin is different from its original, and, at least for MUC1, the uptake depends on mucin

Ocular Mucins

palmitoylation. It is not known whether this also occurs at the surface of the eye. Degradation

Bacteria are an essential player in mucin degradation: glycosidases and proteolytic enzymes contribute to mucus gel breakdown and renewal; adherent bacteria are probably wrapped in mucin epitopes (sacrificial epitopes) and removed from the surface. During sleep, neutrophil enzymes degrade the gel and cause or enhance the cleavage of surface-associated mucins, promoting recoating of epithelial surfaces with a mucus gel. Control of Secretion Mucins are a defensive secretion, augmented in response to pathogens and stimuli, including mechanical, thermal, and chemical insults (through nocioceptors in cornea and conjunctiva). Though the signaling pathway triggered by the external stimulus might be distinct, the involvement of protein tyrosine kinases, mitogen-activated protein kinases (MAPKs), and transcription factor NF-kappa B are relatively common. In the conjunctiva, the eicosanoid 15-(S)-hydroxy5,8,11,13-eicosatetraenoic acid (15(S)-HETE) stimulates MUC1, but not MUC2, MUC4 release, suggesting mucinspecific control of surface-associated mucin liberation in the tears. Further evidence for gene-product-specific modulation is derived from studies of a human corneal limbal epithelial cell line where matrix metalloproteinase-7 and neutrophil elastase induced the release of MUC16, but not of MUC1 or MUC4. Nucleotide agonists acting locally through P2Y2 purinoceptors (belonging to the major G-protein-coupled receptor (GPCR) family) on apical membranes of goblet cells provide a major regulatory system for mucin secretion. Cholinergic agonists are potent stimuli of mucin secretion. One of the pathways to secretion is through EGF receptor (EGFR) induction of MAPK. A fast Ca2+ sensor for the soluble NSF attachment protein receptor (SNARE) complex (the core machinery of membrane fusion) is essential for regulated secretion.

121

size-to-charge ratio of mucin populations is surprisingly well conserved: migration patterns of mucins from conjunctivas of men and women of different ages are similar, suggesting a tight control on the molecular size/charge ratio of intracellular mucins. Conservation of population parameters is also encountered in the largest mucins extracted from tears or contact lenses as shown in Figures 6 and 7. Ratios of the different mucin species sampled from the ocular surface showed little variation in asymptomatic contact lens wearers. The variation must therefore be more subtle and is yet to be understood.

Mucin Function Mucins are essential for the defense of the organ from external factors. Mucins are: (1) lubricants easing the movement of the lids over the surface of the eye; (2) a physical barrier; (3) a trap for microbes; and (4) a matrix where tear constituents among which enzymes, antimicrobial peptides, and other signal molecules fulfill their physiological role. The surface mucins that are richly represented on cornea and conjunctiva are expected to fulfill signaling functions, which are yet to be clearly understood. Adequate production of mucins is controlled by neural, hormonal, and other signaling networks. In addition,

250 KDa

Individual Variation It is not known whether the number of goblet cells varies in individuals. Their number, as assayed by impression cytology, represents those that are discharging in response to external stimuli, and depends on external factors, for example, wind, humidity, and spectacle wear. Goblet cell numbers decrease in severe, but not mild, dry eye disease. Allelic variation, for example, in the number of tandem repeats, is but one source of individual variation. Additionally, variation in glycosyltransferases and in the availability of donor sugars in different physiological states gives rise to potentially large individual variation. However, the

MUC5AC in individual tear samples Figure 6 Electrophoresis of MUC5AC from individual tear samples. Very little variation was seen between the mobilities of the largest MUC5AC polymers in tears from nine individuals. Samples (not equalized for protein content) were run on 1% agarose electrophoresis, vacuum blotted, and visualized after incubation with antibody CLH2 to MUC5AC peptide core. The graphic underneath shows details of mobility for the color-coded respective lane.

122

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

MUC1

MUC16

MUC4

MUC7

Figure 7 Electrophoretic profile of mucins in individual tear samples. Electrophoresis on NuPage Bis-Tris gels allows analysis of MUC1, MUC4, MUC7, and MUC16 and their fragmentation in individual samples. There are clear similarities in mobility, at least for species sufficiently represented to be detected with these methods.

Gel Formation Mucins adhere to surfaces through either sugars or peptide sequences. In a physiological buffer, sequential adhesions of a mucin lowered onto mica were most often spaced similar to the atoms on this surface, indicating that no region of the molecule is barred from adhering to the substrate. The early events in adhesion are governed by random diffusion. Later (i.e., after the first polymers have adhered), cooperative sequential adsorption occurs, creating regions of high mucin density. In dilute solutions, mucins reptate, that is, the polymer slides parallel to itself. A long mucin can thus extricate itself from entanglements, like a strand of spaghetti from a tangle. Mucin polymers associate through long-range hydrogen bonds; they interact with like-charged moieties (e.g., sugars on another polymer) through bridging divalent cations, and form disulfide bonds between unpaired cysteins. Lectinlike site adhesions and hydrophobic domain interactions are also expected to contribute to gel formation, additional to the entanglement of these flexible polymers (Figure 8). The ocular surface is covered by a stable and dynamic mucous gel. In vitro, the elastic properties of the gel are conserved in time and recover after addition of purified ocular mucin – paralleling the discharge of granules from a goblet cell in the conjunctiva (Figure 9). The dynamic nature of the gel can be gleaned in changes in the roughness of its surface, and the temporary oscillations in elastic qualities after intervention. Calcium chelation weakens

− S-S − − − S-S −

a massive discharge of mucins occurs as a response to environmental stimuli, from mechanical stimulation to bacterial lipopolysaccharide.

− −

−

−

− Cat+2 − − − − − − − − −

−

−

Cat+2 −

−

− −

−

− S-S

− − Cat+2 − − −

−

−

− Cat+2 − − −

Figure 8 Mucin adsorption and intermolecular bonds. Schematic of mucin adsorption and association (animation). Sugar or peptide-core moieties can adhere to a surface; intermolecular long-range bonds are established between charged epitopes, bridged by divalent cations in solutions, and through unpaired Cys moieties.

human preocular mucus gels: subsequent addition of Ca restores their initial elastic qualities. The gel collapses on mucin depolymerization. Shorter mucins (i.e., 500 nm long) are more mobile in a model gel than the longest polymers (1 mm long). It must be emphasized that the qualities of a gel depend not only on the mucins, but on all its constituents. Intracellular autocatalytic cleavage in the C-termini of MUC5AC is expected to generate reactive termini that form cross-links and enhance gel formation. The pattern of reactivities with antibodies to different parts of MUC5AC suggests that these cleavages occur in normal mucins, but not in those from dry eye patients. Significant for mucin function is the fact that small entities (compared to pore size) gain more local mobility in a mucin gel than in viscous polymer solution, as

Ocular Mucins

123

Epithelium

Figure 10 Schematic of the tear film anchoring to the ocular surface. Collage of confocal image of conjunctival epithelium and AFM images of cell surface associated (gray) and secreted mucins (blue). Glycocalyx is schematically shown in red.

Anchoring to the Ocular Surface (a)

Surface tension is responsible for maintaining the tear film onto the ocular surface. Apical epithelial cells of the cornea and conjunctiva express a rich glycocalyx that interfaces with the secreted phase of the fluid. Entanglement, hydrogen bonding, sugar–sugar interactions, and lectin–sugar interactions are all involved in the interpenetration of the sessile and the stirred layer of the preocular fluid (Figure 10). Surface-associated mucins are longer than most other glycocalyx components and are expected to penetrate and entangle with elements of the secreted fluid.

Cycle 1

E* (kPa)

2 20

3

Physical and Chemical Barriers

0 (b)

40 nm Mica

Figure 9 Purified mucin gels in vitro. (a) AFM topographical images during formation of mucin gels by repeated deposition of purified mucins on mica, in HEPES buffer (animation). (b) Characteristics of ocular mucin gels in vitro. Correlates of Young’s modulus (elastic modulus, E*) of the gels were calculated from indenting the gel surface with an AFM tip. These elastic characteristics of the gel were calculated at different depths (of the order of nanometres) within the gel, and differences reflect the effect of the stiff substrate. Elasticity being conserved, as illustrated for three cycles of indentation separated by more than 1 h, indicates that the gel is a stable structure. Throughout this period the gel was kept hydrated in HEPES buffer.

relatively large, typically 200–400-nm, fluid-filled pores open up. Thus, the transport of biologically significant molecules is enhanced in the preocular gel, while micronsized pathogens are still prevented from penetrating.

Particles (e.g., from mascara or cigarette smoke) can be seen trapped in the tear film for a period of time, after which they are eliminated from the ocular surface. This mechanism might involve wrapping in mucin or mucin aggregates, and it is also believed to be part of mucin turnover eliminating spent mucins. Sacrificial glycosylated epitopes, to which bacteria or viruses adhere before being eliminated, or which accept reactive molecular species, are also involved in the protective function of mucins. Mucins are resistant to proteolysis because the peptide core is shielded by sugars. Degradation occurs at specific sites giving rise to small mucin polymers that are very mobile in the mucin gel. The presence of cleavage sites outside the cell membrane to which cell-surface mucins are anchored underlies the turnover of the glycocalyx and tear film itself, for example, through the action of neutrophil enzymes. The same mechanism has a protective

124

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

value, when, in response to bacterial enzymes, the cleavage of mucins and portions of gel removes invading organisms from the surface of the eye. Lubrication Water lubricates sliding soft surfaces densely covered with bottlebrushes of oligosaccharide chains. Mucin polymers incorporate large quantities of aqueous tears that bridge between negatively charged oligosaccharides. This cushioning layer sustains the pressure of the moving lids, thus decreasing friction. Furthermore, tears lower their viscosity with shear, further easing the blink. In vitro, in a physiological buffer, human ocular mucin macromolecular aggregates slide past each other without forming adhesions. Tear Breakup Evaporation is the organizing event of tear breakup. The evaporation of water from the liquid trapped in the pores of the gel may cause an increase in salts, and importantly in divalent cations, which causes a local salting-out of mucins and other polyelectrolytes, in turn causing a local collapse of the scaffold, and thus a dry spot on the ocular surface. Antimicrobial Activity The ocular surface is not sterile; it hosts commensal bacteria that are part of the normal ocular surface physiology. Mucins are considered to have co-evolved with bacteria. Their mutual relationships result in continued and effective protection of the underlying tissues. Signaling between microflora and host mucosa impacts on local metabolism and immune function in ways which are a current focus of research. Carbohydrates, including mucin glycans, can be used as an energy source for bacteria. At the ocular surface, mucin oligosaccharides are protected from bacterial degradation by the acetylation of their terminal sugars: most sialic acids are mono-O-acetylated, and a minority diacetylated. Acetylation prevents or restricts bacterial glycolytic enzymes. A proportion of mucin terminal sugars is sulfated, which also protects from cleavage of the glycan by bacterial enzymes. For many bacteria, mucins are antiadhesive: denuding a cornea of mucins resulted in a large increase in the number of organisms attached to its surface. This effect can be achieved by mucin glycans interacting with bacterial adhesins and thus blocking them, and because bacteria coated in mucins do not bind to the mucosal surface. Mucins glide over those without any adhesions. In tears, Pseudomonas aeruginosa bind to specific oligosaccharide receptors containing sialic acids, and a substantial

proportion of glycans are potential receptors. Mucin glycans, however, contain 10 times fewer sialic acids in the a2–6 linkage – the receptor for this bacterial species– than a2–3-linked sialic acids that are not adhesive (Figure 3). This example illustrates the two-pronged defensive mechanism of binding and removing in the bulk of the preocular fluid by dissolved moieties, and increasing antiadhesion closer to the epithelial surface. To degrade complex substrates such as mucins, bacteria need an impressive array of glycosidases and peptidases acting in sequence. These are rarely in the arsenal of a single species, though they can be – and are – expressed by a bacterial community. Commensal bacteria degrade mucins: they cleave sugar chains and peptide backbones, changing the physicochemical characteristic of the mucin molecule. It is thought that in this way the commensal flora contributes to the renewal of the mucus component of the tear film.

Immune Protection In large epithelial tracts, a link between mucins innate mucosal immunity and mucosal inflammatory responses is provided by modulation of mucin expression by inflammatory cytokines such as interleukins (IL)-1b, IL-4, IL-6, IL-9, IL-13, interferons, tumor necrosis factor-a, or nitric oxide. Neutrophils – the main patrolling cells of the closed eyes – can also stimulate increases in production of both gel-forming and cell-surface mucins through neutrophil elastase. In vitro, their degranulation and activation were shown to be different on normal and dry eye mucins.

Clinical Relevance and Pathology Contact Lens Wear If all species present in the preocular fluid also adhere to the contact lens without prejudice to its optical properties and gas permeability, a mucin coating should add cushioning, and provide added antimicrobial protection to the ocular surface. Among mucins adhering to contact lenses there are very large and also (relatively) short mucin polymers, with both secreted and surface-associated mucins well represented. Mucins adherent to contact lenses might act as acceptors for reactive groups produced at the ocular surface. In these mucins, there exist subunits with much higher negative charge than observed in intracellular mucins. Naive wearers deposit more mucins on their lenses than experienced wearers, probably prior to habituation to the stimulus provided by the lens. Contact-lensinduced dry eye does not appear to alter the proportion of mucin species at the ocular surface.

Ocular Mucins

Dry Eye Syndromes Involvement of mucins in dry eye syndromes is suggested by a number of observations: (1) goblet cell numbers decrease in severe disease; (2) the quality of mucus can be manifestly affected with adherent plaques forming on the ocular surface; (3) ocular surface epithelia are (and feel) dry and even keratinized. The increase in squamous epithelial proliferation and relative (or absolute) decrease in goblet cells in dry eye syndromes highlight the different functions of surface-associated and secreted mucins in the preocular fluid. Mucin genes are not altered in the conjunctiva of patients with moderate nonimmune (non-Sjo¨gren) dry eye compared to normal conjunctiva. However, the expression of fucosyl- and sialyl-transferases is decreased, consistent with the decrease in sialylation observed in mucins of dry eye patients. Differences between neutrophil adherence and activation (dependent at least in part on fucosylated epitopes) on fields of normal mucins and mucins in Sjo¨gren patients suggest that these differences become more severe with the severity of disease. See also: Adaptive Immune System and the Eye: Mucosal Immunity; Conjunctival Goblet Cells; Defense Mechanisms of Tears and Ocular Surface; Imaging of the Cornea; Immunopathogenesis of Pseudomonas Keratitis; Innate Immune System and the Eye; Overview of Electrolyte and Fluid Transport Across the Conjunctiva.

Further Reading Argueso, P. (2003). The cell-layer- and cell-type-specific distribution of GalNAc-transferases in the ocular surface epithelia is altered during keratinization. Investigative Ophthalmology and Visual Science 44: 86–92. Basbaum, C. (1999). Control of mucin transcription by diverse injuryinduced signaling pathways. American Journal of Respiratory and Critical Care Medicine 160: S44–S48.

125

Berry, M. (2001). Exploring the molecular adhesion of ocular mucins. Biomacromolecules 2: 498–503. Carlstedt, I. (1985). Mucous glycoproteins: A gel of a problem. Essays in Biochemistry 20: 40–75. Carraway, K. L. (2000). Multiple facets of sialomucin complex/MUC4, a membrane mucin and erbb2 ligand, in tumors and tissues (Y2K update). Frontiers in Bioscience 5: D95–D107. Dartt, D. A. (2000). Regulation of conjunctival goblet cell secretion by Ca2+ and protein kinase C. Experimental Eye Research 71: 619–628. Fleiszig, S. M. J. (1994). Modulation of Pseudomonas aeruginosa adherence to the corneal surface by mucus. Infection and Immunity 62: 1799–1804. Gipson, I. K. (1997). Mucin genes expressed by the ocular surface epithelium. Progress in Retinal Eye and Research 16: 81–98. Hattrup, C. L. (2008). Structure and function of the cell surface (tethered) mucins. Annual Review of Physiology 70: 431–457. Imbert, Y. (2006). MUC1 splice variants in human ocular surface tissues: Possible differences between dry eye patients and normal controls. Experimental Eye Research 83: 493–501. Jumblatt, J. E. (1998). Regulation of ocular mucin secretion by P2Y2 nucleotide receptors in rabbit and human conjunctiva. Experimental Eye Research 67: 341–346. McMaster, T. J. (1999). Atomic force microscopy of the submolecular architecture of hydrated ocular mucins. Biophysical Journal 77: 533–541. Perez-Vilar, J. (2007). Mucin granule intraluminal organization. American Journal of Respiratory Cell and Molecular Biology 36: 183–190. Royle, L. (2008). Glycan structures of ocular surface mucins in man, rabbit and dog display species differences. Glycoconjugate Journal 25: 763–773. Sharon, N. (1989). Lectins as cell recognition molecules. Science 246: 227–234. Thornton, D. J. (2008). Structure and function of the polymeric mucins in airways mucus. Annual Review of Physiology 70: 459–486.

Relevant Websites http://www.functionalglycomics.org – Consortium for functional glycomics. http://www.hugo-international.org – HUGO Gene Nomenclature Committee. http://www.library.nhs.uk – National Library for Health Specialist Libraries. http://www.genenames.org/genefamily/muc.php – The HGNC database in 2008: a resource for the human genome.

Tear Drainage F P Paulsen and L Bra¨uer, Martin Luther University Halle-Wittenberg, Halle, Germany ã 2010 Elsevier Ltd. All rights reserved.

Glossary Dacryocystitis – An infection of the efferent tear ducts. Dacryolithiasis – The formation and presence of dacryoliths. Dacryostenosis – The obstruction or narrowing of one or both canaliculi or the nasolacrimal duct. It may be present at birth. Epiphora – An abnormal overflow of tears down the face. Horner’s muscle – A branch of the orbicularis oculi muscle passing behind the lacrimal sac; it contributes to the lacrimal pump. Natural killer (NK) cells – A type of cytotoxic lymphocytes that constitute a major component of the innate immune system. Pseudostratified epithelium – A type of epithelium that, though comprising only a single layer of cells, has its cell nuclei positioned in a manner suggestive of stratified epithelia. Rheology – The study of the flow of matter: mainly liquids but also soft solids or solids under conditions in which they flow rather than deform elastically. It applies to substances which have a complex structure, including muds, sludges, suspensions, polymers, many foods, bodily fluids, and other biological materials. The flows of these substances cannot be characterized by a single value of viscosity (at a fixed temperature) – instead the viscosity changes due to other factors.

The upper and lower canaliculi, lacrimal sac, and nasolacrimal duct are subsumed under the terms nasolacrimal ducts, efferent tear ducts, or lacrimal passages. After birth, the function of the nasolacrimal ducts is to drain tear fluid into the inferior meatus of the nose. The physiology of lacrimal drainage has been under study for over a century. Various mechanisms have been proposed to explain tear drainage, reflecting the unique anatomic configuration of the efferent tear ducts (Table 1). These include an active lacrimal pump mechanism that functions by contraction of the orbicularis eye muscle; a wringing-out mechanism governed by a system of helically arranged fibrillar structures; the bulging and subsiding of a cavernous body that surrounds the lacrimal sac and nasolacrimal duct; the action of epithelial secretion products; and physical factors such as capillarity, gravity, respiration,

126

evaporation, and absorption of tear fluid through the lining epithelium of the efferent tear ducts. Nevertheless, the complex mechanism by which the tear fluid is brought from the ocular surface to the inferior meatus of the nose is yet not completely understood. The median transit time of a single applied teardrop containing fluorescein dye has been shown to be 4.5 min or 8 min, depending on whether the fluorescein is applied without or with anaesthetizing the ocular surface, respectively. Without anaesthetizing, some reflex tearing of the lacrimal gland is initiated that increases lacrimal fluid volume, which in turn shortens the dye transit time. However, the passage time is subject to distinctive intraindividual variability with a standard deviation of 3.23 min and minimum and maximum values between 15 s and more than 18 min, respectively. There may be several factors that determine the high level of intraindividual variability in dye transit time: fluctuations within a single individual over time, family predisposition, emotional status, the fluid balance, basal tear film production, atmospheric conditions of testing, tear pump efficiency, hormonal status, and blink rate.

History The first exact description of the efferent tear duct system dates back to Giovanni Battista Carcano Leone (1574). Together with the work of Niels Stensen (1662) on tear secretions, Leone’s explanations led to a plausible concept for the entire lacrimal system.

Development During the third month of embryological development, the eyelid folds contact each other and fuse. The specialized structures of the eyelids develop during the period of fusion. Some epithelial cores from both margins of the lid folds get buried in them, at the inner sixth of the eyelids, to form the precursors of the puncta and canaliculi. The epithelial buds, which have grown inside the tarsus, become canalized, as do the epithelial cores that will form the puncta and canaliculi. In the sixth month, the nasolacrimal system becomes patent as a result of lysis in the central cells of both epithelial rods (one starting from the inner canthus and extending toward the nose and the other starting from the nasal mucosa and extending

Tear Drainage Table 1

Mechanisms of tear drainage

Active lacrimal pump mechanism aided by contraction of the lacrimal portion of the orbicularis muscle Distension of the lacrimal sac by the action of the lacrimal portion of the orbicularis muscle Epithelial secretion products (mucins, TFF peptides, and surfactant proteins) of the epithelia of the lacrimal sac and nasolacrimal duct Wringing-out mechanism governed by a system of helically arranged fibrillar structures Opening and closing of the lumen of the lacrimal passage effected by the bulging and subsiding of the cavernous body Capillarity Respiration Evaporation Absorption of tear fluid through the lining epithelia of the lacrimal sac and nasolacrimal duct

toward the inner canthus) at their site of junction. At birth, the nasolacrimal canal is patent from the puncta to the nasal mucosa, under the lower concha (also known as turbinate). The lower end of the lacrimal duct is separated from the inferior meatus of the nasal cavity by a membrane (Hasner’s membrane) consisting of the apposed mucosa lining the nasal fossa and the lower end of the duct. Many newborns suffer from congenital obstruction of the lacrimal pathways. The rate of congenital membranous stenosis of the lacrimal excretory systems in newborns has been reported to be as high as 50%. Fortunately, there is a high rate of spontaneous relief of the epiphora within the first 9 months of life. The repair of a lacrimal duct obstruction should therefore only rarely be performed prior to this age.

Anatomy and Dimensions The lacrimal passages consist of a bony passage and a membranous lacrimal passage. The bony passage is formed anteriorly by the frontal process of the maxilla and posteriorly by the lacrimal bone. The membranous part includes the lacrimal canaliculi, the lacrimal sac, and the nasolacrimal duct (Figure 1). Each canaliculus starts with a 0.25 mm (upper) to 0.3 mm (lower) large, round, oval, or slit-like lacrimal punctum with a nearly 2-mm-long vertical part. Consequently, the lacrimal canaliculus nearly runs at a right angle into the horizontal part, which measures approximately 8 mm. In most cases (c. 65–70%), the two canaliculi join to form a common canal that penetrates the wall of the lacrimal sac regularly 2–3 mm below the apex of the sac, termed fornix or fundus sacci lacrimalis. The vertical diameter of the sac is close to 12 mm, the saggital 5–6 mm, and the transversal 4–5 mm. The nasolacrimal duct normally measures 12.4 mm in adults. The bony coat is nearly 10 mm long and has a diameter of 4.6 mm.

127

The upper and lower canaliculi are lined by pseudostratified/stratified columnar epithelium and surrounded by a dense ring of connective tissue as well as muscle fibers of the lacrimal portion of Horner’s muscle (orbicularis oculi muscle and tensor tarsi muscle). The lacrimal sac and the nasolacrimal duct are lined with a double-layered epithelium, revealing a superficial columnar layer with microvilli and a deep flattened layer of basal cells. Both layers sometimes appear pseudostratified. Some cells of the nasolacrimal duct are lined by kinocilia (sincular kinocilium; a motile cilium on the apex of distinct cells, for example, cells covering the nasal cavity and nasal sinus). Besides epithelial cells, goblet cells are also integrated in the epithelium, sometimes forming intraepithelial mucous glands. Moreover, small seromucous glands are present in the lamina propria, especially in the fundus of the lacrimal sac.

Comparative Anatomy Unlike the human and ape nasolacrimal systems composed of upper and lower canaliculi, the lacrimal sac, and the nasolacrimal duct, the lacrimal systems in dogs, rabbits, cats, deer, pigs, and rats consist solely of the upper and lower canaliculi, leading directly into the nasolacrimal duct. Human, ape, dog, rabbit, cat, deer, and pig tissues reveal a pseudostratified, columnar epithelium with double layering in most areas, a basal cell layer and a superficial columnar layer. The rat shows a multilayered epithelium. The upper cell layers consist of larger squamous elements over several layers of essentially cuboidal cells. Goblet cells are integrated in the epithelia of humans, rats, and cats as solitary cells and in human and rat epithelia as intraepithelial mucous glands. By contrast, the epithelia of apes, dogs, rabbits, deer, and pigs contain no goblet cells. However, ape, dog, rabbit, and pig epithelia do contain many epithelial cells that show mildly positive staining with alcian blue (pH 1.0) in the upper cytoplasm; from investigations in dogs, it is known that this positive staining corresponds to mucins (authors’ observations). The cells with the mild staining are mostly arranged in cell groups. There are also epithelial areas without such cells or cell groups. Subepithelially, the lamina propria of the human lacrimal passage is composed of loose connective tissue containing elastic fibers and lymphatic cells and a rich venous plexus comparable to a cavernous body. A surrounding cavernous system of blood vessels is also found in apes, dogs, rabbits, deer, and pigs, but is absent in rats and cats. Small seromucous glands with excretory ducts opening into the lacrimal passage are integrated in the lamina propria of humans and pigs. None of the other animals possesses seromucous glands. Compared to humans, the efferent tear duct system of dogs and pigs is long. Therefore, the similarities between rabbit and human nasolacrimal ducts support the

128

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Upper lacrimal punctum Upper canaliculus

Lacrimal sac Cavernous body Nasolacrimal duct Lower canaliculus Lower lacrimal punctum Area of Hasner’s valve Inferior meatus

Figure 1 Nasolacrimal ducts. At the medial rim of the upper and lower lids, the lacrimal puncta open, leading into the lacrimal sac through the upper and lower canaliculi. The lacrimal sac is situated in the orbital lacrimal fossa and proceeds into the nasolacrimal duct. The nasolacrimal duct is surrounded by a bony canal created by the maxillary and lacrimal bones and opens into the inferior meatus of the nose. Both the lacrimal sac and nasolacrimal duct are surrounded by a vascular plexus comparable to a cavernous body that is connected to the cavernous system of the nose. From Putz / Pabst: Sobotta, Atlas der Anatomie des Menschen, 22. Auflage ã 2006 Elsevier GmbH, Urban & Fischer Verlag Mu¨nchen.

use of the rabbit for experimental studies of the efferent tear duct system.

Tear Transport through Canaliculi The drainage of tears involves a number of different mechanisms that are not completely understood. It has been suggested that physical factors such as gravity, respiration, and evaporation might play a role in the drainage of tears through the lacrimal passage. Brienen and Snell postulated that the main, and presumably the sole, force that impels lacrimal flow from the conjunctival sac is the pressure brought about by closing of the eyes; in all probability, their expansions and contractions are secondary consequences of pressure fluctuations in the conjunctival sac. Jones introduced the concept of the lacrimal pump system which functions with blinking and might be responsible for lacrimal drainage by analyzing the structure of the medial palpebral ligament and the palpebral part of the orbicularis oculi muscle. It has also been shown that, during blinking movements, the canaliculi and medial canthal tendon are compressed and a uniform volume of lacrimal fluid is squirted into the lacrimal sac. The expansion of the lacrimal sac then causes suction

during the opening phase of the blink, and, after the opening phase of the punctual areas, the canaliculi and lacrimal sac vacuum breaks to reload with tear fluid. The small canaliculi may also act as capillary tubes. Lacrimal fluid is attracted by capillarity into the lacrimal puncta, and, upon closing of the eyelids, the contraction of the preseptal muscle creates a negative pressure and sucks the tear fluid into the sac. The existence of negative pressure and the active transport of tears into the sac are, however, questioned by many. Nevertheless, the importance of Horner’s muscle becomes clear in cases of facial palsy. Tears are not pumped through the lacrimal system. Even with a Jones tube (a small tube allowing tears to drain into the nose) in place, there will be a decrease in tear flow if the orbicularis muscle function is insufficient. Support for the existence of a canalicular pump system on lid closure also came from experimental work carried out by others. Amrith et al. demonstrated that the puncta elevate and meet forcefully when the lids are half shut, and, on complete lid closure, the canaliculi and sac are compressed, forcing tears into the sac and nasolacrimal duct. The elastic expansion of the channels during lid opening and the drawing apart of the puncta break the vacuum, and the tear from the marginal strip is drawn into the puncta. Doane concluded that the force generated in the canaliculi

Tear Drainage

during lid closure alone is sufficient to transport the nonreflex secretion, as it is less than 1 ml min 1. If excess fluid is available, as in reflex tearing, it is possible that the sac may contribute to drawing the fluid. However, high-speed photographic and cinematographic techniques have not been useful in demonstrating what happens in the lacrimal sac during a blink. Even the role of gravity is not clear since Hurwitz concluded that gravity does play a significant role in the transport of tears.

Tear Transport through Lacrimal Sac and Nasolacrimal Duct It is a popular misinterpretation that the tear fluid is drained into the inferior meatus of the nose by the contraction of musculature surrounding the lacrimal sac and/or nasolacrimal duct. Underneath the epithelium, the lamina propriae of the lacrimal sac and nasolacrimal duct consist of loose connective tissue containing a thin layer of elastic fibers and a rich venous plexus situated under this tissue that is connected caudally to the cavernous body of the nasal inferior turbinate. Collagen bundles as well as elastic and reticular fibers between the blood vessels of a rich venous plexus are arranged in a helical pattern and run spirally from the fornix of the lacrimal sac to the outlet of the nasolacrimal duct, where they contribute biomechanically to tear outflow during blinking. Specialized types of blood vessels are distinguishable inside the vascular tissue and are comparable to a cavernous body. The blood vessels are specialized arteries (barrier arteries), venous lacunae (capacitance veins), veins (throttle veins), and arteriovenous anastomoses. They facilitate the opening and closure of the lumen of the lacrimal passage by swelling and shrinkage of the cavernous body. Swelling occurs when the barrier arteries (arteries with an additional muscular layer) are opened and the throttle veins (veins whose tunica media contains a muscle layer of helically arranged smooth muscle cells) are closed. Filling of the capacitance veins (widely convoluted venous

Canaliculus

Lacrimal sac

129

lacunae) occurs at the same time as closure of the lumen of the lacrimal passage. By contrast, closure of the barrier arteries and opening of the throttle veins reduce the blood flow to the capacitance veins, simultaneously allowing blood outflow from these veins with resultant shrinkage of the cavernous body and dilatation of the lumen of the lacrimal passage. Arteriovenous anastomoses enable direct blood flow between arteries and venous lacunae; thus, the subepithelially located capillary network can be avoided and rapid filling of capacitance veins is possible when the shunts of the arteriovenous anastomoses are open. While regulating the blood flow, the specialized blood vessels permit opening and closing of the lumen of the lacrimal passage, effected by the bulging and subsiding of the cavernous body, and simultaneously regulate tear outflow. The presence of the cavernous body is lacking in nearly all textbooks of anatomy and is therefore unknown to most nasolacrimal surgeons and radiologists. It is, however, densely innervated. Epiphora related to emotions such as sorrow or happiness occurs not only by increased tear secretion from the lacrimal gland and accessory lacrimal glands, but also by closure of the lacrimal passage. This mechanism acts, for example, to provide protection against foreign bodies that have entered the conjunctival sac: Not only is tear fluid production increased, but tear outflow is also interrupted by the swelling of the cavernous body to flush out the foreign body and protect the efferent tear ducts themselves. Moreover, it can be assumed that the valves in the lacrimal sac and nasolacrimal duct described in the past by Rosenmu¨ller, Hanske, Aubaret, Be´raud, Krause, and Taillefer could be caused by different swelling states of the cavernous body and must therefore be considered speculative. In fact, the cavernous body of the efferent tear ducts plays an important role in the physiology of tear outflow regulation and can be influenced pharmacologically. Interestingly, administration of a decongestant drug or insertion of a foreign body at the ocular surface prolong the tear transit time significantly, but by different mechanisms (Figure 2). The application of a

Foreign body

Decongestant + foreign body

Decongestant

Nasolacrimal duct (a)

(b)

(c)

(d)

Figure 2 A Schematic/anatomical model of the state of the cavernous body and lacrimal passage in the (a) resting state and (b–d) under different experimental conditions, indicating the specific swelling and compression of the cavernous body and how it permits or restricts tear drainage. Reproduced from Paulsen, F. P., Schaudig, U., and Thale, A. B. (2003). Drainage of tears: Impact on the ocular surface and lacrimal system. Ocular Surface 1: 180–191.

130

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

decongestant drug simultaneously with insertion of a foreign body shortens the tear transit time significantly compared to the effect of the decongestant drug alone, but there is no significant difference compared with application of a foreign body alone. The tear transit time is independent of side (right or left), gender, whether eyeglasses are worn, and whether the person is suffering from a common cold. Tear outflow is further supported by the regional distribution of epithelial secretion products in the lacrimal sac and nasolacrimal duct (see the section titled ‘Innate immune mechanisms’). These include membranebound and secretory mucins and trefoil factor family (TFF) peptides and, as recently shown, surfactant proteins. The epithelial secretion products might influence the rheology and flow of tears through the efferent tear passages. There is speculation that at the ocular surface, mucin composition, distribution, and function are influenced by shear forces generated during blinking. Such forces are absent in the nasolacrimal ducts, and other mechanisms are necessary to ease the flow of tears. From the data available so far, it can only be said that mucins, TFF peptides, and surfactant proteins are likely to interact and may thus affect tear outflow. However, concrete data addressing this feature are lacking.

Innate Immune Mechanisms As is the case with all mucosae, the surfaces of the lacrimal sac and the nasolacrimal duct are in constant interaction with environmental microorganisms and are therefore vulnerable to infection. Similar to conjunctiva and cornea, the mucosa of the nasolacrimal ducts has developed a number of different nonspecific defense systems that can protect against dacryocystitis (Table 2); the epithelial cells thus produce a spectrum of different antimicrobial substances, such as lysozyme, lactoferrin, and secretory phospholipase A2, as well as defensins, which protect against the physiological germ flora inside the lacrimal passage. When infectious and/or inflammatory Table 2 Functions of the epithelia of the lacrimal sac and nasolacrimal duct Secretion of antimicrobial substances (lysozyme, lactoferrin, secretory phospholipase A2, bactericidal-permeabilityincreasing protein, heparin-binding protein, human b-defensins, and surfactant proteins A and D) Secretion of mucins (MUC2, MUC5AC, MUC5B, MUC7, MUC8) and production of membrane-bound mucins (MUC1, MUC4, and MUC16) Secretion of trefoil factor (TF) peptides (TFF1 and TFF3) Secretion of surface active components (surfactant proteins B and C) Production of lipids Absorption of tear fluid components

dacryocystitis pose a threat, changes in the expression pattern occur, inducing the production of some of the antimicrobial substances, for example, antimicrobial peptides such as human inducible beta defensins 2 and 3, which are not produced under healthy conditions in the efferent tear ducts. Besides supporting tear outflow, the product of the mucus component formed by goblet cells and epithelial cells has been attributed largely to immunological response. It contains mucins MUC1, MUC2, MUC4, MUC5AC, MUC5B, MUC7, MUC8, and MUC16 and probably additional mucins. Moreover, the epithelium of the nasolacrimal ducts expresses and produces – as already mentioned – the TFF peptides TFF1 and TFF3. Disturbances in the balance of single mucins or TFF peptides are important in the development of dacryostenosis, daryolithiasis, and daryocystitis. Mucins have several functions. In addition to lubricating the mucosa and waterproofing to regulate epithelial cell hydration, mucins protect mucosal surfaces against potentially harmful substances; however, a variety of oral and intestinal bacteria have been shown to produce sialidase, an enzyme that can degrade mucins by removing sialic acid. Additionally, oral and intestinal bacteria synthesize an array of other glycosidases that can attack the oligosaccharide residues of mucins. Early results of current investigations reveal that such glycosidases are also present at the ocular surface. Finally, secretory immunoglobulin A (sIgA; the class of antibodies produced predominantly against ingested antigens, found in body secretions such as saliva, sweat, and tears, and functioning to prevent the attachment of viruses and bacteria to epithelial surfaces) is incorporated into the mucus layer of mucosal surfaces, supplementing the protective activity. It can interact with functionally diverse cells, including epithelial cells, B- and T-lymphocytes, natural killer (NK) cells, cells of the monocyte/macrophage lineage, and neutrophils. All of these latter cell types, as well as sIgA, are present on and in the nasolacrimal ducts and belong to the lacrimal mucosal immune system (see below). This defense is supported by the collectins (surfactant-associated proteins) SP-A and SP-D in the service of nonspecific natural immune defense and in the activation of the adaptive immune system. As a substance intrinsic to drained tear fluid, they protect the ocular surface in conjunction with IgA, defensins, and mucins against infection by Pseudomona aeruginosa, Staphylococcus aureus, and other pathogenic microbes in preventing the formation of dacryocystitis.

Adaptive Immune Mechanisms Subepithelially, lymphocytes and other defense cells are amply present inside the efferent tear ducts, sometimes

Tear Drainage

aggregated into follicles. Aggregated follicles are present in nearly a third of nasolacrimal ducts from unselected cadavers with no known history of disease involving the eye, efferent tear ducts, or the nose. These aggregations and the surrounding tissue fulfill the criteria for designation as mucosa-associated lymphoid tissue (MALT). They consist of organized mucosal lymphoid tissue characterized by the presence of reactive germinal centers and mantle zones. Around the mantle zone, there is an additional zone of somewhat larger cells corresponding to marginal zone cells. These larger cells extend into the overlying epithelium, forming a lymphoepithelium. In accordance with the terminology of MALT in other body regions, MALT of the human nasolacrimal ducts was termed TALT and, in conjunction with CALT of the conjunctiva, EALT for eye-associated lymphoid tissue. Current analysis of EALT, as well as different epithelial cells of the lacrimal passage, is interesting with regard to the induction of tolerance in the nasolacrimal system and at the ocular surface. Specific secretory immunity depends on a sophisticated cooperation between the mucosal B-cell system and an epithelial glycoprotein called the secretory component. The initial stimulation of Ig-producing B-cells is believed to occur mainly in organized MALT. It has become evident that considerable regionalization or compartmentalization exists in MALT, perhaps determined by different cellular expression profiles of adhesion molecules and/or the local antigenic repertoire. The antigenic stimulation of B-cells results in the generation of predominantly IgA-synthesizing blasts (an immature stage in cellular development before the appearance of the definitive characteristics of plasma cells) that leave the mucosae through efferent lymphatics, pass through the associated lymph nodes into the thoracic duct, and enter the circulation. The cells then return selectively to the lamina propria (nasolacrimal ducts) as plasma cells or memory B-cells by means of homing mechanisms and contribute to mucosal sIgA.

Absorption of Tear Fluid Components Recent animal experiments in rabbits have indicated that the components of tear fluid are absorbed in the nasolacrimal passage and transported into the surrounding cavernous body that is subject to autonomic control and regulates tear outflow. Under normal conditions, tear fluid components are constantly absorbed into the blood vessels of the surrounding cavernous body. These vessels are connected to the blood vessels of the outer eye and could act as a feedback signal for tear fluid production (Figure 3), which ceases if these tear components are not absorbed.

131

2

1

Figure 3 The normally constant absorption of tear fluid components into the blood vessels of the surrounding cavernous body of the nasolacrimal ducts and their transport to the lacrimal gland by blood vessel connections (1) could be a feedback signal for tear fluid production (2). Reproduced from Paulsen, F. P., Schaudig, U., and Thale, A. B. (2003). Drainage of tears: Impact on the ocular surface and lacrimal system. Ocular Surface 1: 180–191.

Conclusions The human efferent tear ducts are part of the lacrimal system. They consist of the upper and the lower lacrimal canaliculi, the lacrimal sac, and the nasolacrimal duct. As a draining and secretory system, the nasolacrimal ducts play a decisive role in tear transport and nonspecific immune defense. In this context, an active lacrimal pump mechanism that functions by contraction of the orbicularis eye muscle has the major impact on tear transport from the ocular surface into the lacrimal sac. From here, tears are transported by a wringing-out mechanism governed by the helical arrangement of fibrillar structures within the vascular system surrounding the lacrimal sac and nasolacrimal duct, the action of epithelial secretion products such as mucins, TFF peptides, surfactant proteins as well as probably others, and physical factors such as capillarity, gravity, respiration, and evaporation. Moreover, components of the tear fluid are absorbed by the epithelium of the nasolacrimal passage and transported into the surrounding vascular system of the lacrimal sac and nasolacrimal duct. This system is comparable to a cavernous body that is subject to autonomic control and also regulates tear outflow from the sac into the inferior meatus of the nose. TALT is present in the efferent tear ducts, displaying the cytomorphological and immunophenotypic features of mucosa-associated tissue MALT. See also: Adaptive Immune System and the Eye: Mucosal Immunity; Adaptive Immune System and the Eye: T Cell-Mediated Immunity; Conjunctiva Immune Surveillance; Conjunctival Goblet Cells; Defense Mechanisms of Tears and Ocular Surface; Inflammation of the Conjunctiva; Lacrimal Gland Overview; Ocular Mucins; Overview of Electrolyte and Fluid Transport Across the Conjunctiva.

132

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Further Reading Amrith, S., Goh, P. S., and Wang, S-C. (2007). Lacrimal sac volume measurement during eyelid closure and opening. Clinical and Experimental Ophthalmology 35: 135–139. Ayub, M., Thale, A., Hedderich, J., Tillmann, B., and Paulsen, F. (2003). The cavernous body of the human efferent tear ducts functions in regulation of tear outflow. Investigative Ophthalmology and Visual Science 44: 4900–4907. Barishak, Y. R. (2001). Embryology of the Eye and Its Adnexa. Basel: Karger. Bernal-Sprekelsen, M., Alobid, I., Ballesteros, F., et al. (2007). Dacryocystorhinostomy in children. In: Weber, R. K., Keerl, R., Schaefer, S. D., and Della Rocca, R. C. (eds.) Atlas of Lacrimal Surgery, pp. 69–71. Berlin: Springer. Bra¨uer, L. and Paulsen, F. P. (2008). Tear film and ocular surface surfactants. Journal of Epithelial Biology and Pharmacology 1: 62–67. Paulsen, F. (2003). The human nasolacrimal ducts. Advances in Anatomy, Embryology, and Cell Biology 170: 1–106.

Paulsen, F. (2006). Cell and molecular biology of human lacrimal gland and nasolacrimal duct mucins. International Review of Cytology 249: 229–279. Paulsen, F. (2007). Pathophysiological aspects of PANDO, dacryolithiasis, dry eye, and punctum plugs. In: Weber, R. K., Keerl, R., Schaefer, S. D., and Della Rocca, R. C. (eds.) Atlas of Lacrimal Surgery, pp. 15–27. Berlin: Springer. Paulsen, F. and Berry, M. (2006). Mucins and TFF peptides of the tear film and lacrimal apparatus. Progress in Histochemistry and Cytochemistry 41: 1–53. Paulsen, F., Fo¨ge, M., Thale, A., Tillmann, B., and Mentlein, R. (2002). Absorption of lipophilic substances from tear fluid by the epithelium of the nasolacrimal ducts. Investigative Ophthalmology and Visual Science 43: 3137–3143. Paulsen, F. P., Schaudig, U., and Thale, A. B. (2003). Drainage of tears: Impact on the ocular surface and lacrimal system. Ocular Surface 1: 180–191. Paulsen, F., Thale, A., Hallmann, U., Schaudig, U., and Tillmann, B. (2000). The cavernous body of the human efferent tear ducts – function in tear outflow mechanism. Investigative Ophthalmology and Visual Science 41: 965–970.

Cornea Overview P Asbell and D Brocks, Mount Sinai Hospital, Department of Ophthalmology, New York, NY, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Conductive keratoplasty (CK) – Refractive surgery that uses heat from radio waves to shrink the collagen in the cornea. It is useful for patients with hyperopia and presbyopia. Dry eye syndrome (DES) – Lack of quantity or quality of tears that lubricate the ocular surface. Disodium ethylenediaminetetraacetic acid (EDTA) – Chemical that chelates calcium. Descemet’s stripping automated endothelial keratoplasty (DSAEK) – Surgical procedure where the Descement’s basement membrane and endothelium of the host are removed and replaced with a thin posterior donor cornea lenticule in its place. Hyperopia – Disorder of vision where the eye focuses images behind the retina instead of on it so that distant objects can be seen better than near objects. Lamellar keratoplasty (LK) – Replacement of damaged or diseased anterior corneal stroma and Bowman’s membrane with donor material. Laser-assisted in situ keratomileusis (LASIK) – Refractive surgery that uses a laser to reshape the cornea. Laser-assisted subepithelial keratomileusis (LASEK) – Refractive surgery that uses a laser to reshape the anterior cornea. Laser thermokeratoplasty (LTK) – Refractive surgery using a mid-infrared laser shrinks the collagen of the corneal periphery. Penetrating keratoplasty (PK) – Corneal transplant using the entire cornea from a donor. Photorefractive keratectomy (PRK) – Refractive surgery where the corneal epithelium is removed and the stroma reshaped. Phototherapeutic keratectomy (PTK) – Procedure where the corneal epithelium is removed and an excimer laser is used to smooth the surface of the cornea. Radial keratotomy (RK) – Refractive surgery in which radial incisions are made in the cornea from the pupil to cornea. Scheimpflug imaging – It measures central corneal thickness and anterior chamber depth.

Anatomy of the Layers The cornea, often referred to as the window of the eye, is covered by the precorneal tear film, and provides a smooth, transparent medium for light rays to pass through. The average cornea measures approximately 12 mm horizontally and 11 mm vertically and seamlessly joins with the opacified sclera at its periphery. The normal cornea has five layers (epithelium, Bowman’s membrane, stroma, Descemet’s membrane, and endothelium), and has an approximate average central thickness of 540 mm.

Epithelium The corneal epithelium is composed of approximately 5–6 rows of stratified squamous epithelial cells. It is these cells in this configuration, along with the overlying tear film that help create the smooth, clear surface. The tight junctions between epithelial cells help to prevent the penetration of microbes and fluid into the corneal stroma. Epithelial cells are continuously being created by the basal limbal stem cells. New cells slowly migrate to the corneal surface where devitalized cells are lost and washed away in the tear film. The entire process takes approximately 2 weeks.

Epithelial Basement Membrane Posterior to the epithelium of the cornea is the proteinacous epithelial basement membrane. Although only an 50-nm thick membrane, the importance of this layer to maintaining a compact, clear, thin anterior corneal surface is obvious when dysfunction occurs, such as in map-dotfingerprint dystrophy (corneal epithelial basement membrane dystrophy). The thickened basement membrane present in this condition can be visualized on slit-lamp examination and it is the dysfunction of this layer that often leads to recurrent corneal erosions.

Bowman’s Membrane Beneath the epithelial basement membrane lies the acellular Bowman’s membrane. This membrane marks the transition from the epithelium to the cornea stroma.

133

134

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Stroma A majority of the cornea (approximately 90% of the full corneal thickness) is the stroma. The stroma is composed of a mixture of lamellae of collagen fibrils, proteoglycans, and keratocytes. The collagen and proteoglycans are produced by the keratocytes, which are dispersed throughout the stoma. The arrangement of collagen fibrils in the stroma plays an important role in maintaining corneal transparency.

exam, corneal thickness measurements, or through visualization with confocal microscopy, where the number, size, and shape of endothelial cells can be evaluated (Figure 2).

Function The cornea performs vital functions as a protective barrier and as an optically clear media.

Descemet’s Membrane

Protective Barrier

The endothelial basement membrane is termed Descement’s membrane. This basement membrane is continually laid down throughout life by the corneal endothelium and is approximately 4-mm thick at birth and approximately 12-mm thick in adulthood. Descemet’s membrane is firmly attached to the endothelium by hemidesmosomes (Figure 1).

The corneal tear film is of vital importance to protecting the eye from invading pathogens. However, in addition to the tear film, it is the tight junctions between the posterior epithelial cells that act as the final barrier in preventing the entrance of microbes into the corneal stroma. This barrier also acts to prevent the inflow of excessive fluid to the stoma from the outside environment.

Endothelium

Transparency

The corneal endothelium is only a single cell-layer thick. These hexagonal cells are of vital importance to maintaining the clarity of the cornea. The endothelial cells lie adjacent to the anterior chamber and work to help maintain the relative dehydrated state of the cornea. The electrolyte pump of the endothelial cells creates an osmotic gradient that draws fluid from the stromal tissue. Unlike the corneal epithelium, the endothelium is not replaced by limbal stem cells when damaged. The average person is born with approximately 3500 cells mm 2, which declines to approximately 2500 cells mm 2 in the eighth decade of life. When endothelial cells are lost or lose functionality, adjacent endothelial cells must enlarge or change shape to attempt to maintain deturgescence. Once a significant portion of endothelial cells are damaged or not functioning properly, corneal edema ensues. The loss of endothelial function can be evaluated by slit-lamp

There are three main attributes of the cornea that allow the tissue to maintain its transparency. First, the water content of the cornea tissue must be maintained at approximately 78%. This water content is dependent on both the epithelial and endothelial cells. The epithelium functions as the physical anterior barrier to the influx of excessive fluid into the corneal stroma. The endothelium is not only a posterior physical barrier to the influx of fluid into the corneal stroma but also uses its sodium–potassium pump to create an osmotic gradient that draws water out of the corneal stroma. Second, it is the compact arrangement and concentration of collagen, keratocytes, and extracellular matrix within the stroma that limits the scattering of light passing through the tissue. Third, the lack of blood vessels in the cornea allows the light rays that pass through the clear tissue to not be

25.4

25

Figure 1 Layers of the cornea as visualized by Scheimpflug imaging. Yellow arrow indicates epithelium, pink arrow indicates stroma, and white arrow indicates endothelium.

Cornea Overview

135

Cell density and area statistics (N = 36) Cell count: 1813 (cell/mm2) Normal: 1597–3120 (cell/mm2) Density [ROI]: 1813 (cell/mm2) Area [ROI]: 0.0199 (mm2) Density [POLY]: 0 (cell/mm2) Area [POLY]: 0.0000 (mm2) Distance 00 (µm)

Figure 2 Endothelial cells and cell count as measured by confocal microscopy.

obscured, refracted, or diffracted in any way by blood vessels. The lack of blood vessels also limits the effect that egress or ingress of fluid into or out of the blood vessels could have on corneal deturgescence (Figure 3).

Disease Processes The disease processes affecting the cornea are extensive. The following review is not meant as an exhaustive list, but rather as an introduction to many of the more commonly observed or discussed pathologies in clinical practice. The use of slit-lamp biomicroscopy to clearly identify which corneal layers are affected by the disease process is of vital importance to the corneal surgeon when choosing among the available medical and surgical treatment options.

Figure 3 Loss of corneal transparency from a mucopolysaccharidosis. The peripheral cornea remains opacified, while the central cornea is transparent following a penetrating keratoplasty.

Epithelial Disease Epithelial staining patterns

The use of vital dyes is required to discern the disease process affecting the corneal epithelium. These dyes include fluorescein, rose bengal, and lissamine green. Fluorescein will stain areas with an epithelial defect, while rose bengal and lissamine green will stain areas of devitalized epithelium.

By analyzing the staining pattern on the corneal surface, the possible etiology of the disease process (Figure 4) may become apparent (Table 1). Epithelial iron deposition Often, the deposition of iron can be visualized on slit-lamp examination of the corneal epithelium. The deposition of

136

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health Table 2

Figure 4 Diffuse damage to the corneal epithelium following the accumulation of environmental debris beneath the upper eyelid. Fluorescein was applied topically to show the epithelial defect (arrow).

Table 1 Examples of staining patterns associated with corneal pathology Diagnosis

Staining pattern

Herpes Simplex Virus keratitis Medication toxicity Exposure keratitis Superior limbic keratoconjunctivitis Superior palpebral conjunctiva foreign body Dry eye syndrome

Dendritic Punctate inferior/inferonasal Punctate inferior Punctate superior Vertical linear Diffuse punctate, with filaments when severe

iron is believed to be related to the uneven distribution of the tear film over the corneal surface in these areas (Table 2). Band keratopathy

Deposition of calcium in the epithelium and calcification of Bowman’s membrane is termed band keratopathy. Of unknown etiology, this process tends to occur in eyes with significant chronic disease. It has also been seen in patients with elevated serum calcium levels, such as occurs with hyperparathyroidism. Chemical/thermal burns

Significant chemical or thermal burns of the corneal epithelium can result in not only diffuse epithelial damage, but also damage to limbal stem cells, perhaps limiting the capacity of the epithelium to heal. These cases may require further surgical intervention, such as limbal stem cell transplant. Medication toxicity

Topical and systemic medications and preservatives used in topical medications are an important cause of induced epithelial changes.

Iron deposition in the corneal epithelium

Iron line name

Iron line location

Hudson-Stahli line Stocker’s line Ferry’s line Fleischer’s ring

Horizontal line in lower third of cornea At edge of pterygium At edge of glaucoma filtering bleb Around keratoconus cone

The commonly used eyedrop preservative benzalkonium (BAK), non-steroidal anti-inflammatory eyedrops, trifluridine, proparacaine, and tetracaine are a few examples of medications that can result in nonhealing epithelial defects. Usually, if identified early enough and if without infection, the induced punctate epithelial defects will rapidly heal. However, if left unchecked and continued use of such medications occurs, corneal melt and even perforation can occur. Deposition of material in the epithelium can occur with certain eyedrops and systemic medication. For example, the use of amiodarone can result in a whorl-like deposition of material. This is not an indication for stopping the medication, as it generally does not affect vision in any manner. Ciprofloxacin, when given topically, can result in a deposition of white crystals in epithelial defects. This can have visual symptoms, and, therefore, the medication should be stopped and replaced with an alternative antibiotic. Thygeson’s punctate keratitis The presence of round central punctate white opacities without corneal edema is the hallmark presentation of Thygeson’s punctate keratitis. The etiology of this disease is unknown. Patients generally present complaining of a significant foreign-body sensation and may have decreased vision if subepithelial haze is present. While usually controlled by steroid or cyclosporin topical treatment, this entity commonly recurs.

Subepithelial Disease Epithelial basement membrane dystrophy Also known as Cogan’s dystrophy and map-dotfingerprint dystrophy, this disease is characterized by areas of thickened epithelial basement membrane appearing as grey epithelial patches or lines along with microcystic epithelial changes appearing as epithelial dots. Although usually asymptomatic, the disease can result in recurrent corneal erosion which is frequently difficult to treat and requires medical or possibly even surgical intervention. Subepithelial infiltrates Following infection with adenovirus resulting in epidemic keratoconjunctivitis, patients may experience persistent haze in their vision. This may be related to persistent subepithelial infiltrates in the visual axis. These infiltrates

Cornea Overview

may improve with time. Although treatment with topical steroid may clear the infiltrates rapidly, they routinely reappear once steroid treatment is discontinued. A very slow taper is advocated in these situations. Overall, avoidance of topical steroids as primary treatment is recommended due to this issue of steroid dependence. Stromal Disease Infection

A variety of infections can result in significant damage to the corneal tissue. The pathogens include bacterial, viral, fungal, and parasitic microbes. Differentiating between bacterial corneal ulcers on examination can be quite difficult. A broad-spectrum antibiotic is generally initially prescribed until further information is obtained from corneal cultures. The more common bacterial pathogens include staphylococcus, streptococcus, and pseudomonas, but cultures must be used to rule out the more rare bacterial organisms such as Bacillus, Corynebacterium, Actinomyces, and Neisseria. While most bacteria require an epithelial defect of some degree to allow bacteria to enter the corneal tissue, it is important to note Neisseria can penetrate through intact epithelium (Figure 5). Herpes simplex virus is of significant importance to consider when evaluating a patient with infectious keratitis. While the characteristic dendritic appearance of the herpes virus moving along corneal nerves makes it fairly obvious to diagnose in these situations, the sometimes atypical appearing stromal infiltrates may require a culture to identify the causative organism. Fungal corneal ulcers, though more rare than bacterial ulcers, recently had a resurgence possibly related to contaminated contact lens solution. The corneal infiltrates tend to have a feathery appearance along their edges and do not respond to antibiotics. Sending fungal cultures at the same time as bacterial cultures is key to avoid late diagnosis of these infections.

Even more rare are parasitic corneal ulcers, such as acanthameoba. Acanthameoba ulcers are most often seen in patients with a history of contact lens wear who frequently have a recent history of swimming. Generally these ulcers are more painful than they appear they should be. Initial appearance on slit-lamp examination can mimic Herpes keratitis, and these ulcers are often initially misdiagnosed as such. Later in the course of the disease a ring infiltrate can present, sometimes mimicking the corneal pattern seen in topical anesthetic abuse. The acanthameoba parasites can be difficult to culture and may require corneal biopsy for diagnosis. Sometimes, it is also possible to visualize the parasitic cysts on confocal microscopy. Dystrophies Stromal corneal dystrophies are rare, yet important, inherited corneal diseases to identify. There are three main types of stromal corneal dystrophies, including granular, macular, and lattice dystrophy. Table 3 reviews the inheritance patterns, material deposited, and appropriate stain to visualize the material on pathologic specimen. Degenerations Corneal degenerations encompass a large category of corneal disease, which includes such more common processes as keratoconus, pellucid marginal degeneration, and arcus senilis. Keratoconus has an unclear etiology, but pathological specimens reveal degeneration of the corneal stroma, Descemet’s membrane breaks, and damage to Bowman’s membrane. Clinically, progressive steepening and thinning of the central cornea can be seen, which eventually can create a cornea too irregular for even hard contact lens wear. Other corneal changes that may be visualized include an iron line around the area of corneal steepening (Fleischer’s ring), stress lines on the posterior cornea (Vogt’s striae), a scissoring reflex on retinoscopy, or areas of corneal opacification and edema from ruptures of Descemet’s membrane (acute hydrops). These changes may be seen on slit-lamp examination; however, it may be necessary to employ computerized tomography or topography to diagnose early keratoconic (forme fruste) changes. Patients with advanced degeneration can consider either intrastromal ring segments or corneal transplantation (Figure 6). Table 3

Granular Macular Figure 5 Infectious keratitis with diffuse conjunctival injection (dilated blood vessels), corneal edema, and a dense corneal infiltrate (arrow).

137

Lattice

Stromal dystrophies Inheritance

Material deposited

Stain

Autosomal dominant Autosomal recessive Autosomal dominant

Hyaline Mucopolysaccharide

Masson trichrome Alcian blue

Amyloid

Congo red

138

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Figure 6 Corneal tomography revealing a central corneal elevation with corneal thinning in a pattern strongly suggestive of keratoconus. Color scale on right indicates thickness where purple is thickest and red is thinnest.

Pellucid marginal degeneration has a similar disease process as keratoconus; however, while the cornea in keratoconus is thinnest at the steepest location, the cornea in pellucid marginal degeneration is thinnest just below the steepest location, which is often near the inferior limbus. Similar pathologic changes can be seen and similar surgical interventions can be considered. Arcus senilis is a peripheral corneal degeneration that is quite common in those 50 years of age or older. When examined pathologically, the white ring seen peripherally consists of lipid deposits. This can be a sign of systemic hyperlipidemia in those younger than 50 years of age or can be a sign of carotid stenosis in those with asymmetric arcus senilis. Descemet’s copper deposition

The deposition of copper in Decesmet’s membrane can sometimes be seen in Wilson’s disease. This often difficult to diagnose disease of copper transport, which leads to copper deposition in the liver and brain, can sometimes be diagnosed quickly at the slit-lamp examination with visualization of the green–brown deposition of copper in the posterior cornea (Kayser–Fleischer ring).

be visualized on slit-lamp examination and on confocal microscopy. Confocal microscopy can also reveal the change in size (polymegathism) and change in shape (pleomorphism) of the dysfunctional endothelial cells. As the endothelium becomes further damaged, corneal edema worsens and may require medical treatment with hypertonic eyedrops and, eventually, even surgical intervention with cornea transplantation to improve visual acuity. Pseudophakic bullous keratopathy During phacoemulsification of the lens, significant stress can be placed on the corneal endothelium, either mechanically with intraocular instruments, or through transfer of energy in the form of ultrasound. These processes can result in loss of endothelial cells, with resultant increased pleomorphism and polymegathism in an attempt to maintain corneal deturgescence. When too much endothelial damage has occurred, persistent corneal edema may present postoperatively and may worsen with time. These patients may require medical intervention with hypertonic eyedrops or, eventually, may require cornea transplantation.

Surgery Endothelial Disease Fuchs’ dystrophy

Fuchs’ dystrophy is characterized by dysfunction of corneal endothelial cells. Wart-like excrescences are deposited by the endothelium into Descemet’s membrane, which can

Once a thorough corneal examination is completed, a proposed medical or surgical intervention can be planned. To understand which surgical options are available, slitlamp biomicroscopy must be used to accurately identify the layers of the cornea affected by the disease process.

Cornea Overview

139

The answer to this question will guide the physician to determine which of a vast array of surgical procedures could be undertaken. The grand scope of the surgical treatment of corneal disease cannot be covered, in full, in this article. A brief overview of some of the more frequently used interventions will be reviewed. Surgical Intervention of Epithelial Disease Amniotic membrane graft

Amniotic membranes have many applications in ocular surgery. For nonhealing or slowly healing epithelial defects, this tissue can be sutured in place over the anterior surface of the cornea, where it helps protect against further corneal degradation and promotes epithelial healing. The amniotic membrane provides a matrix on which new cells can grow, helps prevent excessive inflammation, and aids in preventing corneal scarring and neovascularization. Conjunctival flap

For poorly healing or nonhealing damage to the corneal surface, a conjunctival flap is an option. A partial or complete (Gunderson) flap can be dissected to cover the affected area, depending on the peripheral or central location of the corneal damage. Any patients with an active corneal process, such as infectious keratitis, are not good candidates for a conjunctival flap as it will not only interfere with appropriate treatment measures, but will also obscure the surgeon’s view when attempting to examine the affected tissue. Corneal glue

Cyanoacrylate tissue adhesive can be used to adhere small corneal perforations or severely thinned corneal tissue. Areas of perforated cornea from infection, trauma, postoperative wound leaks, or other etiologies can be treated with cynanoacrylate adhesive, which can at least temporarily renew the integrity of the globe. Disodium ethylenediaminetetraacetic acid chelation

Band keratopathy can present as a dense deposition of calcium hydroxyapetite in the visual axis. Following removal of the corneal epithelium, the anterior corneal calcium deposition can be removed by soaking the affected tissue in disodium EDTA, and scraping away the residual calcium with an ophthalmic blade. Patching or bandage contact lens must be placed until the epithelium heals adequately (Figure 7). Limbal stem cell transplant

Poorly healing epithelial defects may be a result of damaged epithelial limbal stem cells. When enough damage

Figure 7 Pre- (top) and postoperative (bottom) band keratopathy following disodium-EDTA chelation. Notice the white calcium hydroxypatite in the top photograph.

has occurred to these stem cells, vascularization of the corneal surface ensues. This can result from a variety of conditions such as infections, chemical burns, or post operatively. If the disease is unilateral, a stem cell autograft can be attempted from the fellow eye. If both eyes are affected, an allograft from donor tissue must be used, leading to a poorer prognosis and the necessity to systemically immunosuppress the patient postoperatively. Phototherapeutic keratectomy Phototherapeutic keratectomy (excimer laser) can be used to remove diseased tissue in the anterior cornea, such as deposits associated with anterior corneal dystrophies or anterior corneal scars. Although typically no significant refractive change is induced, with the removal of more tissue, the risk of corneal flattening inducing hyperopia increases. Postoperative scarring can occur, and many advocate the use of topical mitomycin C intraoperatively to decrease the risk of scarring. Pterygium excision Pterygium refers to a benign growth of the conjunctiva. A pterygium commonly grows from the nasal side of the sclera and is associated with ultraviolet-light exposure.

140

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

The most successful technique for pterygium excision has been under much debate for many years. The options of leaving behind a bare sclera, using a conjunctival autograft, rotational graft, or amniotic membrane graft are all under current use. The debate centers around the question of which technique results in the lowest frequency of recurrence. The gold standard seems to be pterygium excision with conjunctival autograft, despite several ongoing studies reviewing the newer option of using the amniotic membrane graft. Stromal puncture

The epithelial disease of anterior basement membrane dystrophy often leads to chronic recurrent corneal erosions. In an effort to create better adherence of the epithelium to its basement membrane and stroma, stromal puncture is attempted, usually with a 25-gauge needle inserted into the anterior stroma. Most commonly, this procedure is undertaken in affected areas outside of the visual axis and only in those patients that have failed medical therapy with lubrication and hypertonic solutions such as sodium chloride, which are also used to try to create a more firm adherence between the anterior cornea layers. Tarsorrhaphy

Tarsorrhaphy is a surgical procedure in which the eyelids are sutured together to protect the cornea. A temporary or permanent tarsorrhaphy can be one of the most important options in the corneal surgeons armamentarium. Particularly helpful in creating a more hospitable environment for epithelial healing, this process can be used to promote epithelial healing in instances such as post operatively following penetrating keratoplasty (PK), neurotrophic epithelial defects, Bells palsy, or chemical burns. A newer option – using botulinum toxin A to create a temporary tarsorrhaphy – is currently being studied as an alternative, which perhaps affords similar protection to a sutured tarsorrhapy while allowing for easier slit-lamp examination.

Penetrating keratoplasty Unlike lamellar keratoplasty, a PK involves the fullthickness removal of diseased corneal tissue from the host and replacement of all layers of corneal tissue with a donor cadaveric corneal button. The procedural technique has evolved over the past 100 years, resulting in the high success rate that is now enjoyed by the thousands of patients that undergo full-thickness corneal transplantations each year (Figure 8). Topical immunosuppression, usually with prednisolone acetate 1%, is generally a long-term necessity to prevent rejection. Irregular astigmatism generally requires suture manipulation or hard contact lens placement post operatively to attain the best-corrected visual acuity. Corneal biopsy For those patients with unusual corneal processes without clear diagnosis despite corneal culture or scrapings, corneal biopsy is the next option. Removing a small anterior portion of corneal tissue from the periphery affords a tissue sample for further pathological examination. Corneal laceration repair With the advent of the 10-0 nylon suture, the ability of the corneal surgeon to repair the integrity of the globe following penetrating corneal injury improved significantly. Ensuring the absence of vitreous or iris incarceration in the wound, that the wound edges are well apposed, and that the wound is watertight are of vital importance. Fulfilling these objectives as soon as possible after the penetrating injury occurs will hopefully help to promote healing and lower the risk of endophthalmitis (visionthreatening inflammation of the internal ocular tissues). Refractive surgery Since its introduction, the field of corneal refractive surgery has grown exponentially. Earlier procedures, such as

Surgical Intervention of Stromal Disease Anterior lamellar keratoplasty

Anterior corneal scars or deposits can be treated with an anterior lamellar keratoplasty. Rather than replacing a full-thickness corneal button as is accomplished in a PK, a partial-thickness trephination or dissection is accomplished to remove only the anterior diseased tissue without performing any manipulation of the posterior endothelial layer. This is an especially important option in young patients with excellent endothelial cell counts and morphology, but whose vision is obstructed by stromal opacities.

Figure 8 One-year status post penetrating keratoplasty. A single 10-0 prolene running suture remains.

Cornea Overview

Figure 9 Status post intrastromal ring-segment implantation.

Figure 10 Status post Descemet’s Stripping Automated Endothelial Keratoplasty. Notice the clear cornea.

radial keratotomy (RK) and laser thermokeratoplasty (LTK) – which have since fallen out of favor – provided important knowledge and insight into the manipulation of corneal tissue. The current options include procedures such as laser-assisted in situ keratomileusis (LASIK), photorefractive keratectomy (PRK), laser-assisted subepithelial keratomileusis (LASEK), conductive keratoplasty (CK), and intrastromal ring segments. The goal of these procedures is usually to remove the need for corrective lenses in all, or most, distance and/or near-daily situations. A thorough review of the importance of appropriate patient selection, preoperative evaluations, individual surgical options, the risks and benefits of each procedure, and the appropriate postoperative care is beyond the scope of this article (Figure 9).

Descemets Stripping Automated Endothelial Keratoplasty

Prosthetic keratoplasty

For those patients with extensive corneal neovascularization, severe damage to limbal stem cells, or who have failed prior PKs, the prosthetic keratoplasty is available (i.e., Boston type 1 keratoprosthesis). Although the technology has evolved substantially over the last 10 years, the keratoprosthesis generally still requires glaucoma surgical management concomitantly and is not appropriate for those patients requiring straightforward PK without a history of such issues as significant corneal vascularization or repeat PK failures. Surgical Intervention of Endothelial Disease The surgical manipulation and replacement of diseased corneal endothelium has taken substantial strides forward in recent years. After going through several generations of endothelial surgery with disappointing results, the ophthalmology community has arrived at the more promising current technique of Descement’s Stripping Automated Endothelial Keratoplasty (DASEK).

141

The most common diseases resulting in endothelial dysfunction are Fuchs’ dystrophy and pseudophakic bullous keratopathy. Although it has been well understood for quite some time that corneal edema in these situations was a result of dysfunctional corneal endothelium, until recently, a full-thickness (penetrating) keratoplasty was the sole viable option. The well-documented time-consuming process of helping these patients obtain their best corrected visual acuity following PK could not be avoided. Now, however, with the advent of the DSAEK procedure, the corneal surgeon can carefully dissect off the Descement’s basement membrane and endothelium of the host, and replace a thin posterior donor cornea lenticule in its place (Figure 10). Early studies suggest that, when successful, this process results in faster healing, less induced postoperative astigmatism, and patients reach their best-corrected visual acuity faster when compared to PK. See also: Adaptive Immune System and the Eye: Mucosal Immunity; Artificial Cornea; Contact Lenses; Corneal Dystrophies; Corneal Endothelium: Overview; Corneal Epithelium: Cell Biology and Basic Science; Corneal Epithelium: Response to Infection; Corneal Epithelium: Transport and Permeability; Corneal Imaging: Clinical; Corneal Nerves: Anatomy; Corneal Nerves: Function; Corneal Scars; The Corneal Stroma; Drug Delivery to Cornea and Conjunctiva: Esterase- and Protease-Directed Prodrug Design; Gene Therapy for the Cornea, Conjunctiva, and Lacrimal Gland; Imaging of the Cornea; Knock-Out Mice Models: Cornea, Conjunctiva, Eyelids and Lacrimal Gland; Lids: Anatomy, Pathophysiology, Mucocutaneous Junction; Refractive Surgery and Inlays; Regulation of Corneal Endothelial Cell Proliferation; Regulation of Corneal Endothelial Function; Stem Cells of the Ocular Surface; Tear Film; The Surgical

142

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Treatment for Corneal Epithelial Stem Cell Deficiency, Corneal Epithelial Defect, and Peripheral Corneal Ulcer.

Further Reading Ang, L., Chua, K., and Tan, D. (2007). Current concepts and techniques in pterygium treatment. Current Opinion in Ophthalmology 18: 308–313. Bahar, I., Kaiserman, I., McAllum, P., et al. (2008). Comparison of posterior lamellar keratoplasty techniques to penetrating keratoplasty. Ophthalmology 115: 1525–1533. Beebe, D. (2008). Maintaining transparency: A review of the developmental physiology and pathophysiology of two avascular tissues. Seminars in Cell and Developmental Biology 19: 125–133. Cauchi, P., Ang, G., Azuara-Blanco, A., et al. (2008). A systematic literature review of surgical interventions for limbal stem cell deficiency in humans. American Journal of Ophthalmology 146: 251–259. Dua, H. and Azuara-Blanco, A. (2000). Limbal stem cells of the corneal epithelium. Survey of Ophthalmology 44: 415–425. Gomes, J., Romano, A., Santos, M., et al. (2005). Amniotic membrane use in ophthalmology. Current Opinion in Ophthalmology 16: 233–240.

Hersh, P., Brint, S., Maloney, R., et al. (1998). Photorefractive keratectomy versus laser in situ keratomileusis for moderate to high myopia. Ophthalmology 105: 1512–1523. Ma, J., Graney, J., and Dohlman, C. (2005). Repeat penetrating keratoplasty versus the Boston keratoprosthesis in graft failure. International Ophthalmology Clinics 45: 49–59. Meek, K., Dennis, S., and Khan, S. (2003). Changes in the refractive index of the stroma and its extrafibrillar matrix when the cornea swells. Biophysical Journal 85: 2205–2212. Meek, K., Leonard, D. W., Connon, C. J., et al. (2003). Transparency, swelling and scarring in the corneal stroma. Eye 17: 927–936. McCarey, B., Edelhauser, H., and Lynn, M. (2008). Review of corneal endothelial specular microscopy for FDA clinical trials of refractive procedures, surgical devices and new intraocular drugs and solutions. Cornea 27: 1–16. Suh, L., Yoo, S., Deobhakta, B. S., et al. (2008). Complications of descemet’s stripping with automated endothelial keratoplasty. Ophthalmology 115: 1517–1524. Sutphin, J. (ed.) (2007). Basic and Clinical Science Course. External Disease and Cornea. San Francisco, CA: American Academy of Ophthalmology. Trattler, W. and Barnes, S. (2008). Current trends in advanced surface ablation. Current Opinion in Ophthalmology 19: 330–334. Yanoff, M. and Duker, J. (2004). Ophthalmology. St. Louis, MO: Mosby.

Corneal Epithelium: Cell Biology and Basic Science M A Stepp, The George Washington University Medical Center, Washington, DC, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Adhesion complex – The term used to refer to the different components of the cell: substrate junction that is present at the basal aspect of the stratified squamous epithelial tissues of the body, including the skin and the cornea. The adhesion complex includes not only the hemidesmosomes that contain a6b4 integrin and collagen XVII, but also the inner hemidesmosomal plaque that contains plectin and BPA230, the anchoring filaments that are made of laminin 332, the anchoring fibers and the adhesions plaques that both contain collagen type VII. This structure forms during development and must partially or completely disassemble during wound healing and tissue regeneration and then reassemble after healing is complete. If the adhesion complex fails to form due to mutations in one of its components, blistering of the epithelial sheets covering the skin and cornea occurs. Basement membrane zone – Sheets of epithelial cells that form the outer surfaces of the body are separated from their underlying mesechymal cells by basement membranes. Basement membranes are composed of the network-forming collagen type IV, the heparan sulfate proteoglycan perlecan, and several different types of laminins. The basement membrane zone consists of the basal plasma membrane surface of the basal epithelial cells, the hemidesmosomes, and the basement membrane itself. At the EM level, it consists of the inner and outer plaques of the hemidesmosomes, the lamina lucida, and the lamina densa. Defensins – These are small cysteine-rich cationic proteins first characterized in leukocytes in 1985. The name was chosen because these proteins have antibacterial, antifungi, and antiviral activity. They consist of 18–45 amino acids including six (in vertebrates) to eight conserved cysteine residues. They bind to proteins on the surfaces of pathogens and form much of the basis of what immunologists call the innate defense system. Human tears have at least 4 different defensins and their presence in the tears is regulated during the healing of corneal wounds. Glycocalyx – This term means sugar coat and in the cornea it refers to the thin film of sugar-containing material at the apical surface of the apical cells on the cornea. The primary molecules that make up the

glycocalyx in the cornea are called mucins. There are several different types of mucins, some secreted and some cell-surface bound, on the healthy, wet ocular surface. The mucins of the glycocalyx help the tear film to spread over the ocular surface. Integrins – A family of glycoproteins that function as heterodimers to mediate both cell:matrix interaction and cellular signaling. Integrins are integral membrane single pass proteins whose cytoplasmic domains bind to elements of the cytoskeleton including actin for most of the integrins and intermediate filaments for a6b4. Their extracellular domains bind to extracellular matrix proteins including collagens, fibronectin, and laminins. Integrins are considered mechanotransducers of forces from outside cells and tissues to the cytoskeleton to allow cells to change shape during development and wound healing. Niche – From the old French word nichier meaning to nest, this term is usually used to refer to the site where adult stem cells reside. In the cornea, the niche is believed to be located at the palisades of Vogt at the corneoscleral junction. Stem cell niches consist of both specialized extracellular matrix molecules that the stem cells adhere to as well as other cell types that play supportive roles by expressing cell:cell adhesion proteins and/or growth factors that act to maintain the stem cells within the niche and inhibit their differentiation. Refractive surgery – This term refers to surgical procedures developed to improve the refractive state of the eye and decrease dependence of the patient on glasses or contact lenses. The major forms of refractive surgery are laser-assisted in situ keratomileusis and photorefractive keratectomy. According to the American Society of Cataract and Refractive Surgery, more than 900 000 refractive procedures were performed in the US in 2005.

The Corneal Epithelium Has Vital Functions in Vision The corneal epithelium is a transparent covering that allows light entry into the eye. It is the outermost

143

144

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

GC

(a)

BM

MP

(b)

Figure 1 The surface epithelium of the cornea is specialized for light transmission and spreading of the tear film. (a) Transmission electron micrograph of the human corneal epithelium; a lower magnification image is in the inset. There are 3–5 cell layers and both apical and basal surfaces are flat and shear to minimize refraction of light. (b) The apical surface of the corneal epithelium has projections called microplicae. When shown at higher magnification as in the inset, the glycocalyx is seen covering the microplicae. The mucin proteins in the glycocalyx are hydrophilic and bind water which helps the tear film spread over the ocular surface. (a, b) Reproduced from Gipson, I. K. (1994). Anatomy of the conjunctiva, cornea, and limbus. In: Smolin and Thoft (eds.) The Cornea: Scientific Foundations and Clinical Practice, 3rd edn., Chapter 1, p. 8. New York: Little Brown and Company, with permission from I. K. Gipson.

layer of the cornea and is comprised of three to five anatomically distinct layers of stratified squamous nonkeratinizing epithelial cells (Figure 1(a)). It protects the inner neuronal tissues of the retina from microbial invasion and prevents water loss from the corneal stroma. When light waves hit the curved corneal surface and pass from the air to an aqueous medium, they are bent. This refraction of light waves focuses the light on the fovea of the retina. The crystalline lens is less powerful than the cornea at bending light waves because the light waves are already in an aqueous environment when they pass from the cornea to the lens. Although the lens does refract light, the differences in the refractive indices of the corneal stroma, the aqueous humor, and the lens are small compared to the difference in refractive index between air and the tear film at the corneal surface. Thus the cornea is the major refractive surface of the eye. Because of its significant impact on the refraction of light, changes in the curvature of the cornea alter the position where light focuses on the retina. This property of the cornea has lead to the development of procedures to alter corneal curvature to reduce or eliminate refractive errors such as myopia and astigmatism. One of the most popular of these procedures worldwide is refractive surgery using laserassisted in situ keratomileusis (LASIK) which involves cutting a flap of tissue at the front of the cornea through the epithelium and stroma, removing or ablating tissue from the stroma using an excimer laser, and repositioning the flap of tissue. LASIK patients will experience an immediate improvement in their sight; as a result, millions of these procedures are performed. It remains to be determined whether or not there will be long term consequences for people who have had LASIK or other corneal refractive procedures as their corneas age and

for this reason basic researchers in the field of corneal wound healing are concerned over the popularity of this procedure.

The Apical Squames Possess Specializations to Promote Tear Film Spreading The apical-most epithelial surface of the cornea has flattened cells called squames. The apical surfaces of these cells have protrusions called microplicae (Figure 1(b)). These are specialized structures that facilitate the spreading of the tear film. The microplicae have bound to them the glycocalyx that contains the membrane-associated mucins and other glycoproteins whose functions include facilitating tear film spread. The outermost layer of the tear film is a layer of lipid molecules that increases the stability of the tear film and prevents tear evaporation. The tear film itself has several functions. It prevents the eyelids from adhering to the corneal epithelium. It is the medium through which carbon dioxide, a product of corneal metabolism, is exchanged for oxygen from the air. In addition, it contains small molecules called defensins that inhibit the growth of microbial agents. When the corneal epithelium is injured, the concentration of these defensins in tears increases transiently and then, after the wound is closed, decreases back to the levels before wounding. These data suggest that defensins play important roles in protecting the cornea from infection especially after the epithelial barrier is disrupted during wound healing. Underneath the superficial apical layer are two to three layers of cells often referred to as wing cells by corneal biologists. The final layer that sits on the basement membrane is called the basal cell layer.

Corneal Epithelium: Cell Biology and Basic Science

The Epithelial Basal Cells Adhere to the Underlying Basement Membrane via tight Adhesive Junctions called Hemidesmosomes The corneal basal cells maintain the tight association of the epithelium with the basement membrane through specialized adhesion junctions called hemidesmosomes. A schematic representation of the adhesion complex, including the hemidesmosomes, is shown in Figure 2(a) and a transmission electron micrograph is shown in Figure 2(b). The hemidesomosomes form a tight rivet holding the epithelial cells on to the basement membrane and anterior stroma. Unlike adherens junctions and focal adhesions, hemidesmosomes utilize intermediate filaments rather than actin to stabilize cell adhesion and maintain cell:matrix adhesions. Hemidesmosomes must disassemble when the corneal epithelium is injured to permit the sheet of epithelial cells to move. The failure of the hemidesmosomes to disassemble during wound healing can lead to delayed wound closure, and failure to reassemble after migration is complete can lead to corneal epithelial erosions. The molecules that are present within hemidesmosomes that mediate these events are beginning to be characterized. a6b4 integrin acts as the primary mechanotransducer of forces from the basal cells to the extracellular matrix molecule laminin 332 in the basement membrane. In addition, a membrane-associated collagen, collagen XVII/BPA180, is also important for the stability of these structures. The function of these integral membrane molecules within the hemidesmosome is to associate with laminin 332, which makes up the anchoring filaments and type VII collagen that makes up the anchoring fibers. The anchoring filaments are found in the basement

145

membrane and the anchoring fibers in the anterior corneal stroma. Thus, the hemidesomosomes are an essential part of an adhesion complex that distributes sheering forces from the cell surface to deep within the stroma. This adhesion complex includes the intra- and extracellular aspects of the hemidesmosomal plaque at the plasma membrane as well as the anchoring filaments and anchoring fibers. The need for rapid mobilization of the corneal epithelial cells after a corneal wound combined with the sheerness of the corneal epithelial basement membrane forced a compromise during evolution: the basal surfaces of the corneal basal cells are more exposed to sheering forces from the environment compared to the basal surfaces of the basal cells in skin. The enfolding and ridges in the epidermis result in a basement membrane surface area that is increased in the basal keratinocytes relative to their lateral and apical surfaces. This provides increased protection from debridement of the epidermis. Because of the sheerness of the corneal epithelial basement membrane, it is more prone to recurrent epithelial erosions compared to the skin. However, this sheer surface also permits the corneal epithelial sheet to migrate rapidly and thereby minimizes dessication of the cornea and allows wounds to close quickly to minimize infections. While the cells that make up the corneal epithelial basal cell layer are the most proliferative of all the corneal epithelial layers, research has shown that the cornea epithelial basal cell layer does not contain the corneal epithelial stem cells (CESCs). If the CESCs were primarily located in the central cornea, they would be exposed to the DNAdamaging affects of ultraviolet light, and, lacking pigment to absorb light energy, would sustain irreversible mutations. To maintain the corneal epithelium throughout life,

HD PM LL LD BM AP LN: LN332 α6β4 integrin BPA 230

(a)

Interstitial collagens CN XVII/ BPA 180

AP IF: keratins CN VII Plectin

(b)

Figure 2 The adhesion complex maintains the tight adhesion of the corneal epithelium to the underlying basement membrane and Bowman’s layer and is shown schematically in (a) and by transmission electron micrograph in (b). It consists of the inner plaque with keratins, plectin, and BPA230, hemidesmosomes (HD) at the plasma membrane (PM) containing a6b4 integrin and type XVII collagen (CN)/BPA180, anchoring filaments at the basement membrane zone and Bowman’s layer containing laminin LN332, and the anchoring fibers and adhesion plaques (AP) containing type VII collagen. The lamina lucida (LL) and lamina densa (LD) contain type IV collagen (light green) and perlecan (dark green), respectively, in addition to numerous other proteins and proteoglycans. Type VII collagen indirectly links the intermediate filaments of the cytoskeleton to the interstitial collagens of the anterior lamellae of the corneal stroma. Adapted from Gipson, I. K. (1994). Anatomy of the conjunctiva, cornea, and limbus. In: Smolin and Thoft (eds.) The Cornea: Scientific Foundations and Clinical Practice, 3rd edn., Chapter 1, p. 10. New York: Little Brown and company, with permission from I. K. Gipson.

146

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

mammals maintain an epithelial stem cell population at the periphery of the cornea at an area called the corneoscleral junction or limbus. At this location, the CESCs are adjacent to a rich environment, often called a niche, which includes both melanocytes that produce melanin and absorb excess light to protect the stem cells from DNA damage, and blood vessels that release growth factors as well as immune cells. The idea that the limbus is a unique and distinct environment is supported by the fact that the composition of the basement membrane beneath the limbal basal cells is unique. It has different relative ratios of collagens and laminins compared to either the conjunctival or corneal basement membrane. Data characterizing the microanatomical relationships among the cells that make up the limbal niche are shown schematically in Figure 3(a). Different integrin heterodimers have been studied as potential markers for the CESCs. Figures 3(b) and 3(c) show the localization of integrins in the human and mouse limbus respectively. One integrin, a9b1, is only expressed in the unwounded adult cornea at the limbus; detailed studies in the mouse using bromodeoxyuridine to study label retaining cells have shown that a9b1 is not expressed on the stem cells themselves but is expressed on the cells that are the immediate progeny of the stem cells, which have been called transit amplifying cells. The CESCs have been shown to express high levels of a6b4- and b1-family integrins.

The Palisades of Vogt The palisades of Vogt were first described over 90 years ago but were named by Vogt in 1921. They can be seen at the corneal periphery at the limbus by a clinician using an ophthalmoscope as shown in Figures 4(a)–(c). A scanning electron micrograph of the palisades of Vogt is shown in Figure 4(d). The palisades of Vogt consist of ridges of stromal tissue covered by epithelial cells. After noting their loss during progression of certain corneal disease states, recent studies have shown that the palisades of Vogt become increasingly less prominent as the cornea ages and disappear entirely in patients with conditions that are associated with CESC deficiency. As discussed above and shown in Figure 3, the CESCs reside in the corneoscleral junction. These adult stem cells are small and relatively undifferentiated. Recently, in vivo confocal micrographs taken of human corneas as a function of age from 10- to 80-year olds show that the epithelial cells within the palisades of Vogt are smaller than those of the central cornea, that the palisades disappear with aging, and that the sizes of the corneal epithelial cells get larger as the palisades get smaller and disappear. These data are consistent with the hypothesis that the CESCs are located within the palisades of Vogt and that the aging cornea displays a progressive reduction in both

the palisades and the stem-like cells that reside there. While palisades of Vogt have not been reported in nonhuman corneas, it is likely that similar structures are present but are not as morphologically distinct as those in the young human cornea.

Bowman’s Layer Is an Acellular Zone Located Immediately under the Corneal Epithelial Basement Membrane Bowman’s layer is usually considered the second of the five layers of the cornea. It is a specialized acellular layer immediately beneath the epithelial basal cell basement membrane; in humans it is 12–18-mm thick and it begins at the corneoscleral junction. It is composed of an amorphous collection of collagen fibers intertwined with a rich collection of fine sensory nerve axons that mostly run parallel to the corneal surface. First observed in the chicken cornea, anatomically distinct Bowman’s layers have been widely reported to be present only in avian, primate, and cat corneas. It can be seen in Figure 1(a). Rabbits and mice were thought to lack a Bowman’s layer. More recent studies suggest that Bowman’s layers are found in mammalian, avian, and even fish corneas, but their thickness and composition vary significantly. The function of Bowman’s layer has remained controversial primarily because there are no true markers for Bowman’s layer and the lack of consensus regarding which species possess true Bowman’s layers; a better understanding of what constitutes a Bowman’s layer is necessary before its function can be determined. The most likely function for Bowman’s layer is to enhance corneal strength and/or integrity perhaps by providing a mesh-like environment through which the thicker more uniform anterior stromal collagen lamellae insert into the type VII collagen containing adhesion complex. This insertion would link and stabilize the corneal epithelial basal cell adhesion complex in the anterior corneal stroma. A secondary function for Bowman’s could be to facilitate nerve innervation of the ocular surface by providing a space through which axons can move during development.

The Collagen- and Proteoglycan-Rich Corneal Stroma with Its Stromal Cells, Descemet’s Membrane, and the Corneal Endothelial Cells are all Vital to the Health of the Cornea The third layer of the cornea is the stroma; it is transparent, compact, and fibrous and serves as a structural support. The clarity of the corneal stroma is due to the regularity of the diameter of the various collagen fibers that comprise it and the packing and specialized arrangement of these sheets, called lamellae, relative to one another. The collagen fiber spacing within and between lamellae is regulated by

Corneal Epithelium: Cell Biology and Basic Science

Conjunctiva

147

Cornea

Limbus

TDC PMC lTAC LC

Bo

eTAC SC

M N

M: LC: BM: N: BV:

(a)

MC

BM

BV

Melanocytes Langerhan’s cells Basement membrane Nerve Blood vessels

TDC: PMC: l TAC: eTAC: SC: MC: Bo:

Terminal differentiated cells Post-mitotic cells Late transient amplifying cells Early transient amplifying cells Stem cells Mesenchymal cells Bowman’s membrane

α9/α6

α9

α6

α9/β1

α9

β1

(b) α9/PI

β1/PI

α9/β1

α9/β1/PI

α9/PI

TN/PI

α9/TN

α9/TN/PI

(c)

Figure 3 The CESCs are present within the limbus at the palisades of Vogt where the epithelial cells show distinctive integrin expression profiles. The schematic representation in (a) summarizes our knowledge on the different cell types reported to be involved in maintaining the stem cell niche at the limbus. (b) Cross sections through the human limbus showing the palisades of Vogt and stained (1) for a9 integrin, which is a marker for cells arising directly from division of the CESCs, and for a6 integrin, a component of the hemidesmosomes that stains the basement membrane zone at the limbus; or (2) for a9 integrin and b1 integrin, an integrin shown to be expressed at high levels on the CESCs. a9 integrin is present in subsets of the basal cells present at both apical and basal sites along the ridges. (c) En face images taken from whole flat mounted mouse corneas stained with antibodies to reveal the localization of a9 and b1 integrin or stained to reveal the localization of a9 integrin and tenascin-C, an extracellular matrix molecule present only at the limbus of the unwounded mouse cornea; nuclei have been stained with the nuclear marker DAPI. Note that there are many more b1 positive cells than a9 positive cells and that tenascin-C and a9 integrin do not appear to co-localize but rather tenascin-C is found beneath the cells that are a9 integrin positive. The magnification bar in (b) equals 25 mm and the magnification bar in (c) equals 5 mm. (a) Adapted from Li, W., Hayashida, Y., Chen, Y. T., and Tseng, S. C. (2007). Niche regulation of corneal epithelial stem cells at the limbus. Cell Research 17: 26–36, with permission from Scheffer C. G. Tseng (c) Adapted from Stepp, M. A. and Zieske, J. D. (2005). The corneal epithelial stem cell niche. Ocular Surface 3: 15–26.

148

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Cornea

Cornea

Cornea

Limbus (a)

Conjunctiva

(b)

Limbus

(c)

Limbus

Palisades of vogt

(d) Figure 4 The palisades of Vogt are undulating pigmented ridges of tissue at the periphery of the cornea at a region called the corneoscleral junction or limbus. (a–c) The ridges of the palisades can be seen in vivo on the ocular surface (a). They can also be seen on sheets of human corneal epithelium removed with dispase from less pigmented (b) and more pigmented eyes (c). The fact that the corneal epithelial sheets contain pigment suggests that melanocytes are present intermixed with the epithelial cells at the limbus. (d) The palisades of Vogt are revealed in detail in this scanning electron micrograph of the de-epithelialized ocular surface. (a–c) Adapted from Li, W., Hayashida, Y., Chen, Y. T., and Tseng, S. C. (2007). Niche regulation of corneal epithelial stem cells at the limbus. Cell Research 17: 26–36, with permission from Scheffer C. G. Tseng. (d) Adapted from Gipson, I. K. (1989). The epithelial basement membrane zone of the limbus. Eye 3: 132–140, with permission from I. K. Gipson.

proteoglycans, including lumican, decorin, and bilycan. The proteoglycans bind and organize water molecules and constrain the spacing of the collagen fibers. The cells that produce and maintain the collagens and proteoglycans of the stromal matrix are mesenchymal cells of neural crest origin and are frequently referred to as corneal keratocytes. During wound healing, these cells become activated to become myofibroblasts and they function to facilitate wound closure. Between the stroma and the corneal endothelial cells there is a specialized homogenous basement membrane known as Descemet’s membrane. The single layers of flattened cells that make up the corneal endothelium are primarily responsible for secreting the extracellular matrix proteins that make up Descemet’s membrane. They are absolutely essential for a healthy cornea. The corneal endothelial cells have numerous channel proteins in their plasma membrane that pump ions and excess water in and out of the cornea to maintain its hydration state. The clarity of the corneal stroma is readily lost if the stroma begins to dry out or desiccate. Dessication disrupts the organization of the collagen molecules by removing the ordered water molecules associated with the proteoglycans. The endothelial

cells also pump nutrients from the aqueous humor into the corneal stroma. Nutrients diffuse across Descemet’s membrane and then passively diffuse outward to nourish the corneal stromal and epithelial cells. Because the cornea is avascular, it requires the corneal endothelial cells to provide its nourishment. It is likely that Descemet’s membrane acts as a sponge to trap nutrients delivered by the corneal endothelial cells. This would provide a sustained release of those nutrients to the rest of the cornea. In the healthy cornea, the corneal endothelial cells are more metabolically active compared to the corneal epithelium and corneal stromal cells. The generation of nutrient and ion gradients is energy dependent and these cells are constantly synthesizing channel and ion pump proteins. In contrast, in the unwounded cornea, the corneal stromal cells and the corneal epithelial cells are fairly quiescent and primarily utilize glycolysis to generate energy.

The Avascular Cornea The healthy central cornea is avascular to permit light to fall on the retina without distortion; blood and

Corneal Epithelium: Cell Biology and Basic Science

lymphatic vessels are present only at the corneal periphery at the limbus. Developing a better understanding of the mechanisms underlying the maintenance corneal avascularity is of extreme importance to cell biologists and many excellent studies have made progress in this area. We now know that the avascular status of the healthy cornea is maintained by the production by the corneal epithelial and stromal cells of several antiangiogenic factors, including angiostatin and endostatin. After corneal injury, corneal cells transiently produce angiogenic factors such as basic fibroblast growth factor and vascular endothelial growth factor (VEGF) disrupting the normal balance of pro- and anti-angiogenic factors and allowing for the ingrowth of blood vessels from the limbus. After healing is complete, these transient vessels undergo pruning and the avascular nature of the cornea is restored. However, prolonged inflammation or recurrent epithelial erosions can compromise corneal avascularity. The clarity of the cornea and its accessibility has given rise to several widely used assays to assess the mechanisms underlying angiogenesis. One is called the corneal pocket assay. This assay has been used by researchers to identify factors that inhibit new blood vessel growth primarily as a tool to learn how to inhibit the growth of tumor cells in various different types of cancer. A small pellet is inserted into an incision made on the cornea surface. If the pellet contains pro-angiogenic growth factors such as FGF or VEGF, blood vessels will sprout from the limbus and grow toward the pellet. The number of vessels that form and the speed with which they move toward the pellet can be quantified, and drugs that inhibit or enhance angiogenesis can be tested. A second procedure used for similar types of studies of angiogenesis involves placing a suture on the cornea. If the suture is placed at a regular distance from the limbus then the ingrowth of vessels from the limbal vasculature can be measured and the ability of various drugs to enhance or inhibit angiogenesis can also be assessed. See also: Artificial Cornea; Conjunctival Goblet Cells; Contact Lenses; Corneal Dystrophies; Corneal Endothelium: Overview; Corneal Epithelium: Response to Infection; Corneal Epithelium: Transport and Permeability;

149

Corneal Epithelium: Wound Healing Junctions, Attachment to Stroma Receptors, Matrix Metalloproteinases, Intracellular Communications; Corneal Imaging: Clinical; Corneal Nerves: Anatomy; Corneal Nerves: Function; Corneal Scars; The Corneal Stroma; Cornea Overview; Defense Mechanisms of Tears and Ocular Surface; Dry Eye: An Immune-Based Inflammation; Imaging of the Cornea; Ocular Mucins; Refractive Surgery and Inlays; Regulation of Corneal Endothelial Cell Proliferation; Regulation of Corneal Endothelial Function; Stem Cells of the Ocular Surface; The Surgical Treatment for Corneal Epithelial Stem Cell Deficiency, Corneal Epithelial Defect, and Peripheral Corneal Ulcer.

Further Reading Argu¨eso, P., Balaram, M., Spurr-Michaud, S., et al. (2002). Decreased levels of the goblet cell mucin MUC5AC in tears of patients with Sjo¨gren syndrome. Investigative Ophthalmology and Visual Science 43: 1004–1011. Dupps, W. J., Jr. and Wilson, S. E. (2006). Biomechanics and wound healing in the cornea. Experimental Eye Research 83: 709–720. Gipson, I. K. (2007). The ocular surface: The challenge to enable and protect vision: The Friedenwald lecture. Investigative Ophthalmology and Visual. Science 48: 4390–4398. Li, W., Hayashida, Y., Chen, Y. T., and Tseng, S. C. (2007). Niche regulation of corneal epithelial stem cells at the limbus. Cell Research 17: 26–36. Mu¨ller, L. J., Marfurt, C. F., Kruse, F., and Tervo, T. M. (2003). Corneal nerves: Structure, contents and function. Experimental Eye Research 76: 521–542. Pajoohesh-Ganji, A., Pal-Ghosh, S., Simmens, S. J., and Stepp, M. A. (2006). Integrins in slow-cycling corneal epithelial cells at the limbus in the mouse. Stem Cells 24: 1075–1086. Ruberti, J. W. and Zieske, J. D. (2008). Prelude to corneal tissue engineering – gaining control of collagen organization. Progress in Retinal and Eye Research 27: 549–577. Sakimoto, T., Rosenblatt, M. I., and Azar, D. T. (2006). Laser eye surgery for refractive errors. Lancet 367: 1432–1447. Stepp, M. A. (2006). Corneal integrins and their functions. Experimental Eye Research 83: 3–15. Stepp, M. A. and Zieske, J. D. (2005). The corneal epithelial stem cell niche. Ocular Surface 3: 15–26. Taneri, S., Zieske, J. D., and Azar, D. T. (2005). Evolution, techniques, clinical outcomes, and pathophysiology of LASEK: Review of the literature. Survey of Ophthalmology 49: 576–602. Wilson, S. E., Chaurasia, S. S., and Medeiros, F. W. (2007). Apoptosis in the initiation, modulation and termination of the corneal wound healing response. Experimental Eye Research 85: 305–311.

Corneal Nerves: Anatomy C F Marfurt, Indiana University School of Medicine – Northwest, Gary, IN, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Chemotropic guidance – A process whereby a chemical substance, usually secreted by some target cell, attracts growing axons. Cochet–Bonnet esthesiometer – A hand-held instrument that uses the pressure transmitted by a nylon monofilament of known diameter and variable length to measure the mechanical sensitivity of the cornea. Iridectomy – A surgical procedure that removes a small piece of the iris, most often for treatment of closed-angle glaucoma or melanoma of the iris. Laser-assisted in situ keratomileusis (LASIK) – A form of refractive laser eye surgery designed to change the shape of the cornea. Neurite – Any cytoplasmic extension from a neuron; the term is used frequently to describe developing or regenerating axons that have not yet attained their mature adult form. Neurotrophic epitheliopathy – The minor degenerative changes of the corneal epithelium caused by damage or functional impairment of the corneal innervation, especially as occurs after LASIK surgery. Neurotrophic keratitis – A serious degenerative condition of the cornea, affecting especially the corneal epithelium, caused by impairment of the corneal sensory innervation. It often manifests as corneal epithelial defects, ulceration, melting, and diminished wound healing. Phacoemulsification – A procedure in which a cataractous lens is broken up (emulsified) by ultrasound and aspirated from the eye; an intraocular lens is then inserted. Receptive field – The area of the cornea (or other body part) supplied by a single sensory nerve fiber. Trabeculectomy – A surgical procedure that removes part of the trabeculum of the eye in order to facilitate drainage of aqueous humor. It is performed for the relief of increased intraocular pressure associated with glaucoma. Trophic substances – The molecules that promote the growth, differentiation, and survival of specific cell populations.

150

Origins of Corneal Nerves The cornea is richly supplied by sensory neurons located in the ophthalmic region of the trigeminal ganglion (Figure 1). Although the entire sensory innervation of the mammalian cornea is derived from a relatively small number (50–450) of neurons, each neuron supports as many as 200–3000 individual corneal nerve endings. The sensory nerves reach the eye through the nasociliary branch of the ophthalmic nerve. In humans, the nasociliary nerve branches typically into two long ciliary nerves, one nasal and the other temporal, which course directly to the posterior pole of the eye, and a communicating branch carrying sensory fibers to the ciliary ganglion. Five to ten short ciliary nerves, carrying a mixture of sensory and autonomic fibers, emerge from the anterior pole of the ciliary ganglion and together with the long ciliary nerves pierce the posterior globe in the vicinity of the optic nerve. After penetrating the sclera, the nerves enter the suprachoroidal space and course anteriorly toward the cornea. While in transit to the anterior eye segment, the fibers branch repeatedly and eventually form a series of prominent, evenly spaced nerve bundles that approach the corneoscleral limbus uniformly from all directions. Some mammalian corneas also receive a modest sympathetic innervation from the superior cervical ganglion; however, the existence and magnitude of this contribution vary widely among species. In rabbits and cats, corneal sympathetic nerves may constitute as much as 10–15% of the total corneal innervation, while in humans and other primates corneal sympathetic fibers are exceedingly rare or absent. Avery sparse parasympathetic innervation has been reported in rat and cat corneas; however, the meager nature of this innervation and its absence from most mammalian corneas suggest that they are likely of little physiologic significance.

Architecture of the Corneal Innervation Limbal Plexus Prior to entering the cornea, the nerve bundles traverse the limbus and contribute fibers to the limbal, or pericorneal plexus, a dense, ring-like meshwork of nerve fibers that completely surrounds the peripheral cornea. The three-dimensional structure of the limbal plexus varies considerably in anatomical complexity according to species and contains complex mixtures of sensory, sympathetic, and parasympathetic nerves. Most limbal

Corneal Nerves: Anatomy

Posterior and anterior ethmoidal nerve Ciliary ganglion Frontal nerve Lacrimal nerve Nasociliary nerve Branch from carotid plexus Oculomotor nerve

151

Supraorbital nerve

Long ciliary nerves

Supratrochlear nerve

Infratrochlear nerve

V

Zygomatic Nerve to inferior nerve rectus Short ciliary nerves Nerve to inferior oblique

(a)

Infraorbital nerve Zygomaticotemporal and zygomaticofacial nerves Supratrochlear nerve Supraorbital nerve

Infratrochlear nerve Anterior ethmoidal nerve Zygomatic nerve

Posterior ethmoidal nerve Long ciliary nerves

Lacrimal nerve Frontal nerve

Short ciliary nerves

Ciliary ganglion

Ophthalmic branch Nasociliary nerve V

(b) Figure 1 The branches of the ophthalmic division of the trigeminal nerve as seen from the lateral side (a) and from above (b). Sensory nerves to the cornea travel mainly in the nasociliary nerve and its ocular branches, the long and short ciliary nerves. Reproduced from figure 25-5 in Liu, G. T. (2004). The trigeminal nerve and its central connections. In: Miller, N. R. and Newman, N. J. (eds.). Walsh and Hoyt’s Clinical Neuro-Ophthalmology, vol. 2, pp. 1233–1274. New York: Lippincott, Williams and Wilkins with permission from Lippincott, Williams and Wilkins.

nerves supply functional innervation to the limbal vasculature; however, occasional fibers travel through the limbal stroma unrelated to blood vessels and may provide sensory or autonomic innervation to resident cells. Stromal and Subepithelial Plexuses Most nerves enter the peripheral cornea in a series of radially directed major stromal nerve bundles, each containing as many as several dozen axons, arranged at regular intervals around the limbal circumference. On average, 60–80 major stromal nerves supply innervation

to the human cornea, while 12–20 bundles supply rabbit, cat, and dog corneas. Additional numbers of smaller nerve fascicles enter the peripheral cornea slightly superficial to the main stromal bundles and provide limited innervation to the perilimbal and peripheral cornea. At their point of entry in the peripheral cornea, most corneal nerves are unmyelinated C-fibers; however, perhaps as many as 30% are finely myelinated A-delta fibers that shed their myelin sheaths within a millimeter or so after entering the cornea. The axons then continue into the stroma as flattened, ribbon-like fascicles surrounded only by Schwann cell cytoplasm and basal lamina.

152

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Soon after entering the peripheral cornea, the major stromal bundles branch repeatedly to form elaborate, tree-like arbors whose distal branches anastomose extensively with neighboring branches to form a dense anterior stromal plexus. The plexus occupies approximately the anterior 25–50% of the corneal stroma, depending on the species, and consists of complex admixtures of seemingly randomly organized small and medium-sized nerve bundles and individual axons with straight, curvilinear, and tortuous trajectories. The plexus becomes increasingly denser and anatomically complex in the anterior direction. The most superficial layer of the anterior stromal plexus in humans and other large mammals is especially dense and is referred to as the subepithelial plexus (SEP). In contrast, the posterior half of the corneal stroma and the endothelium are, with rare exceptions, devoid of nerve fibers. A small but indeterminate number of nerve fibers in the anterior stromal and SEPs appear to terminate in the stroma, usually as slightly bulbous terminal swellings or free nerve endings. At the electron microscopic level, the stromal fibers and their terminals resemble morphologically free nerve endings found in the corneal epithelium, suggesting that stromal nerve endings may subserve undetermined sensory transduction functions. Some stromal fibers form intimate relationships with keratocytes and are occasionally enwrapped in keratocyte cytoplasmic extensions that may provide a morphological substrate for reciprocal functional interactions. Individual stromal axons entering at the corneoscleral limbus may travel as much as three-quarters of the way across the cornea before ending and, as a result of repetitive branching, possess receptive fields that range in size from less than 1 mm2 to as much as 50 or more square millimeters and may cover up to 20% or more of the corneal surface. The receptive fields are generally round, oval, or wedge-shaped in outline and often extend several millimeters beyond the cornea onto the adjacent limbus and bulbar conjunctiva. The large sizes and extensive overlap between adjacent receptive fields, coupled with convergent mechanisms in the central nervous system, explain why stimulation of the corneal surface is poorly localized.

Subbasal Nerve Plexus The subbasal nerve plexus comprises the densest part of the corneal innervation and is so-named because the nerves that comprise this plexus travel in the subnuclear part of the basal epithelium or between the basal epithelial cells and their basal laminae. Estimates of subbasal nerve fiber density in the central human cornea are expressed conventionally as the total length of all nerve fibers and branches per square millimeter, and range from

15–27 mm mm–2 (by in vivo confocal microscopy) to 40–55 mm mm–2 (by immunohistochemical staining of corneal whole mounts). Approximately 400–500 widely spaced stromal nerves penetrate Bowman’s membrane, mainly in the peripheral and intermediate cornea, to give origin to the human subbasal plexus. Additional subbasal nerves enter the peripheral cornea directly from the limbal plexus. Relatively few stromal nerves penetrate Bowman’s membrane in the central 2 mm of the human cornea; hence, the latter region receives most of its epithelial innervation from long subbasal nerves that invade the central cornea from more peripheral origins. Immediately after penetrating Bowman’s membrane, each stromal nerve bends at a 90 angle and branches simultaneously into 2–20 thinner nerve fascicles termed subbasal nerves. Subbasal nerves course in the horizontal plane, roughly parallel to one another and to the epithelial basal lamina, for 1–2 mm in rats and up to 6 mm or more in humans. The neuroanatomical structure thus formed by a stromal axon branching into multiple, parallel subbasal daughter fibers is termed an epithelial leash, a unique, two-dimensional sensory nerve specialization found only in the cornea (Figure 2). At the electron microscopic level, human subbasal nerves average 1.5 mm in diameter and may contain up to 40 individual unmyelinated axons (Figure 3). Immediately upon entering the basal epithelium, the axons shed their Schwann cell investments and continue as naked axon cylinders. Subbasal nerves in adjacent leashes interconnect frequently with one another. As a result, the boundaries between individual leashes are soon lost and a composite subbasal nerve plexus is formed. When viewed in its entirety, the subbasal nerve plexus forms a gentle spiral or whorl-like pattern on the curved corneal surface (Figure 4). The center of the spiral, often called the vortex, is located in human corneas approximately 2–3 mm inferior and nasal to the corneal apex. As a consequence of this arrangement, subbasal nerves in the superior and apical human cornea are oriented vertically, whereas subbasal nerves in other corneal regions may be oriented horizontally or obliquely, consistent with their geographical locations within the whorl-like plexus. The mechanisms that govern the formation and maintenance of this spiral-like pattern remain uncertain; however, it has been established that basal epithelial cells and subbasal nerves migrate centripetally in tandem. According to one theory, basal epithelial cells derived from stem cells in the corneoscleral limbus migrate centripetally in a whorl-like fashion toward the corneal apex in response to chemotrophic guidance, electromagnetic cues, and population pressures; the subbasal nerves, occupying cytoplasmic invaginations or narrow intercellular spaces between adjacent columns of migrating cells, are pulled along and undergo compensatory horizontal elongation. Alternatively, subbasal nerves may develop whorl-like,

Corneal Nerves: Anatomy

153

Intraepithelial terminals Epithelium

Basal lamina

Subbasal nerves

Stroma

Stromal nerve in subepithelial plexus

Figure 2 Innervation of the rabbit corneal epithelium. Stromal nerves penetrate the basal lamina and branch into a leash-like assemblage of horizontally oriented fibers called subbbasal nerves; the latter nerves give rise to a profusion of intraepithelial terminals. Modified from figure 6 in Rozsa, A. J. and Beuerman, R. W. (1982). Density and organization of free nerve endings in the corneal epithelium of the rabbit. Pain 14: 105–120. With kind permission from International Association for the Study of Pain.

Superior Nasal

Figure 3 Electron micrograph of a subbasal nerve from a human cornea. The nerve, cut in cross section, contains eight individual unmyelinated axons. M, mitochondria. Calibration bar is 1 mm. Reproduced from figure 5c in Muller, L. J., et al. (2003). Corneal nerves: Structure, contents and function. Experimental Eye Research 76: 521–542. With kind permission from Elsevier.

curvilinear orientations independent of epithelial cell dynamics and may, in turn, provide a structural scaffold that patterns and directs epithelial cell migration. Intraepithelial Nerve Terminals As the subbasal nerves course horizontally through the basal epithelium, they give rise to a profusion of thin,

Figure 4 Architecture of the subbasal nerve plexus in the human cornea. The area illustrated is 6.5 mm in diameter and centered on the corneal apex. Subbasal nerves converge in a gentle, whorl-like pattern on a region approximately 2–3 mm inferonasal to the corneal apex known as the vortex.

occasionally beaded terminal axons that ascend vertically or obliquely, often with a modest amount of additional branching, into the more superficial epithelial layers before ending (Figure 5). In some mammalian corneas,

154

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

additional epithelial nerves originate as single axons directly from the SEP. The end of each intraepithelial fiber is tipped by a conspicuous, bulbous terminal expansion. At the ultrastructural level, these expansions contain abundant small clear vesicles, varying numbers of large dense-cored vesicles, mitochondria, glycogen particles, neurofilaments, and neurotubules. Morphologically, they resemble nociceptor nerve endings described in other tissues. The neurochemical content of the small clear vesicles is uncertain but may include excitatory amino acids, while the large, dense-cored vesicles likely contain substance P (SP), calcitonin gene-related peptide (CGRP), and/or other neuropeptides. The terminals are located throughout all layers of the corneal epithelium but are especially numerous in the basal and wing cell layers. Some terminals may extend to within a few microns of the corneal surface. The epithelial cell membranes facing the nerve terminals often demonstrate numerous invaginations and, in many cases, the epithelial cell cytoplasm totally surrounds the nerve ending. The intimate morphological relationship thus formed between nerve terminal and epithelial cell does not constitute a true synapse, but may nevertheless provide a favorable environment for the exchange of diffusible substances and receptor-mediated interactions. The intimate contacts may also allow nerve endings to sense osmotic changes in epithelial cell shape and volume brought about by dessication of the ocular surface. When osmotic stimulation reaches threshold, the nerve fibers fire and activate brainstem circuits that promote reflex tear production and blinking in an effort to maintain the physiologic integrity of the ocular surface.

Figure 5 Intraepithelial nerve terminals (arrows) in the human corneal epithelium stained by neurotubulin immunohistochemistry. Arrowheads, subbasal nerve.

The innervation density of the corneal epithelium is probably the highest of any surface epithelium, and the central corneal epithelium of humans and rabbits contains approximately 5000–8000 nerve terminals per square millimeter. Both nerve terminal density and corneal sensitivity are greatest in the central cornea and decrease progressively in the peripheral direction. The richness of the corneal epithelial innervation provides a nociceptive detection system of unparalleled sensitivity and it is hypothesized that injuries to individual epithelial cells may be sufficient to trigger pain perception. Both corneal sensitivity and nerve density decrease progressively as a function of age and may contribute to the pathogenesis of dry eye disease in some elderly patients.

Corneal Nerve Neurochemistry Corneal sensory nerves help maintain a healthy ocular surface by activating brainstem circuits that stimulate reflex tear production and blinking, and by releasing trophic substances, including numerous neuropeptides, which promote corneal epithelial physiologic renewal and wound repair. Corneal nerves contain, in varying proportions, the same neuropeptides that are expressed in other ocular nerves. Each corneal fiber population (sensory, sympathetic, and parasympathetic) maintains a distinctive phenotypic signature; however, the chemical coding is complex and most fibers likely express combinations of neuropeptides rather than individual markers. To date, 12 different neuropeptides have been detected by radioimmunoassay or immunohistochemistry in the mammalian cornea. Corneal sensory nerves may contain one or more of six different neuropeptides. Two of these peptides, CGRP and SP, are found in especially high percentages of corneal nerves and remain the most studied and well characterized of the corneal peptidergic nerves (Figure 6).

Figure 6 Calcitonin gene-related peptide (CGRP)immunoreactive subbasal nerves in the rat corneal epithelium.

Corneal Nerves: Anatomy

CGRP and SP are expressed in approximately 30–60% and 10–20% of corneal sensory nerves, respectively, and have been found in all mammalian corneas investigated to date. The vast majority of nerves that contains CGRP also contain SP and the two peptides probably co-localize in the same vesicles. Other neuropeptides expressed in more limited numbers of corneal sensory nerves include: neurokinin A (a member of the tachykinin family) secretoneurin (a member of the chromogranin/secretogranin family), pituitary adenylate cyclase-activating peptide (PACAP, a member of the vasoactive intestinal polypeptide (VIP)-glucagon-secretin super family), and galanin. The extent to which these peptides coexist with CGRP and SP or represent distinct populations of corneal sensory nerves remains to be determined. Despite the richness of the corneal peptidergic innervation, it does not fully account for the known density of the corneal sensory innervation and it is therefore likely that many corneal sensory nerves do not express neuropeptides. Immunohistochemical investigations of rat corneal sensory nerves suggest that up to 40% of rodent corneal nerves do not express known neuropeptides. Most of the so-called nonpeptidergic nerves express a cell-surface glycoconjugate that binds the plant isolectin Bandaireae simplicifolia IB4 and contain the enzyme, fluoride-resistant acid phosphatase (FRAP). The neurotransmitters expressed by this prominent population of corneal IB4-positive nerves remain to be determined; however, excitatory amino acids such as aspartame and glutamate, or unknown neuropeptides that remain to be characterized, are likely candidates. Corneal autonomic nerves, which constitute in most species only a small percentage of corneal nerve fibers, are also neurochemically diverse. Corneal sympathetic nerves express (in addition to noradrenalin) serotonin and neuropeptide Y (NPY), while corneal parasympathetic nerves express VIP, met-enkephalin, NPY, and galanin. Functional roles of most neuropeptides in the mammalian cornea remain unclear and knowledge of their physiological roles has not kept pace with the results of immunohistochemical analyses. An important exception to this statement is SP, which exerts essential trophic functions on the corneal epithelium. SP receptors are present on corneal and limbal epithelial cells and it is thought that, under resting physiologic conditions, SP released from corneal sensory nerve terminals promotes corneal epithelial maintenance and physiological renewal by activating cellular pathways that stimulate epithelial cell proliferation, migration, and adhesion. In addition, topical application of SP has been reported to accelerate the rate of corneal epithelial wound healing in both experimental animal models and clinical patients with persistent corneal epithelial defects. Damage to corneal sensory nerves by surgery or trauma deprives the corneal epithelium of SP and other essential nerve-derived trophic substances and is associated with a variety of ocular

155

surface disorders, including dry eye disease, epitheliopathy, and neurotrophic keratitis. Even less is known of the functional roles of the cornea’s limited autonomic innervation. In animal models, corneal sympathetic nerves act through adrenergic mechanisms to modulate corneal epithelial cell ion transport processes, epithelial cell proliferation and mitosis, and cell migration during epithelial wound healing. In the corneoscleral limbus, NPY augments the vasoconstrictor effects of noradrenalin and exerts strong angiogenic effects.

Corneal Nerve Remodeling Corneal subbasal nerves and their intraepithelial terminals are dynamic structures that undergo morphological rearrangements continuously under normal physiologic conditions. Time-lapse, in vivo confocal microscopic examination of living human eyes reveals that subbasal nerves slide centripetally in tandem with their neighboring basal epithelial cells at rates of 10–20 mm day–1 and that this nerve elongation occurs through the addition of new nerve material near the site of nerve penetration at Bowman’s membrane. As subbasal nerves cannot elongate indefinitely, it is hypothesized that the distal nerve segments eventually degenerate or slough into the tear film. Intraepithelial nerve terminals also undergo continuous remodeling through combinations of long-term, nervedirected reconfigurations and passive, short-term reorganization in response to outward migrations of differentiating epithelial cells.

Corneal Nerve Regeneration after Ocular Surgery Corneal nerves are transected during a variety of corneal and anterior-segment surgical procedures, including refractive surgery, perlimbal incisions performed for cataract surgery, iridectomy and trabeculectomy, and penetrating keratoplasty (PK; corneal transplantation). Corneal nerves depend for their survival on axoplasmic transport of essential substances from their parent nerve cell bodies in the trigeminal ganglion; thus, surgical procedures that interrupt corneal nerve fibers cause rapid degeneration of the distal axons, decreased corneal sensitivity, and compromised functional integrity of the ocular surface. Corneal nerves are capable of regeneration; however, it is a slow, imperfect process and the regeneration that takes place after most corneal surgeries is characterized by reduced nerve density, alterations in nerve architecture, and diminished corneal sensitivity. The more proximally the nerves are cut, the more delayed and incomplete the regeneration process will be. Thus, surgical disruption of

156

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

the subbasal and subepithelial nerve plexuses produce, in general, less serious and more short-term damage to the corneal innervation than do deep or penetrating incisions that affect major stromal nerve bundles. Refractive Surgery In laser-assisted in situ keratomileusis (LASIK), a mechanical or femtosecond laser microkeratome is used to create a circular flap of corneal epithelium and superficial stroma with a hinge at one end; the flap is then folded back and an excimer laser is used to remove several tens of microns of stroma from beneath the flap. The microtome severs subbasal and subepithelial nerves all around the flap margin except at the hinge, and the excimer laser destroys additional nerves in the anterior stromal bed. Regenerating nerve fibers emerge first as neurites from the subbasal plexus at the hinge and, to a lesser extent, elsewhere around the flap margin, followed by a second generation of neurites originating from the transected stromal nerve trunks. Abnormal proteoglycans present at the LASIK scar interface between the stromal bed and the overlying flap are postulated to impede successful reinnervation from the anterior stromal plexus. Regeneration is slow and incomplete, with subbasal and stromal nerve densities returning to less than half of preoperative levels by 12 months after LASIK and failing to reach preoperative levels even 3–5 years later. In addition, the architecture and morphology of the regenerated subbasal nerve plexus are often abnormal. Long-term studies are needed to determine whether subbasal nerve density will eventually return to pre-LASIK levels. Although corneal reinnervation after LASIK surgery is incomplete, the corneal sensitivity, as determined by Cochet–Bonnet esthesiometry, in most cases returns to preoperative levels by 6–12 months and the short-term clinical outcome in most patients is excellent. Most patients experience a decrease in corneal sensitivity and mild-to-moderate dry eye after LASIK surgery that lasts for only a few days; however, approximately 5% of LASIK patients suffer long-term dry eye symptoms that may be related at least in part to impaired corneal reinnervation and interruption of the neural circuits that drive reflex tear production. Some patients with chronic dry eye after LASIK experience persistent and aberrant pain sensations that likely reflect sensitization and altered responsiveness of the immature, regenerating nerves. In photorefractive keratectomy (PRK), the corneal epithelium is removed and an excimer laser is used to ablate the most anterior portion of the corneal stroma. Nerve regeneration and recovery of corneal sensitivity under these conditions, in which a flap is not cut in the cornea, reportedly occurs more quickly than following LASIK surgery; nevertheless, subbasal nerve density, architecture, and corneal sensitivity remain depressed for up to 1–2 years postoperatively.

Cataract Surgery Penetrating perilimbal incisions, such as those performed for cataract surgery, are curvilinear incisions that are a few millimeters in length, which transect small numbers of major stromal nerve trunks near their sites of entry into the cornea. Neural regeneration proceeds slowly through a combination of nerve regrowth and collateral sprouting from adjacent, uninjured stromal nerves. The success of nerve regeneration varies from individual to individual; some nerve trunks fail to regenerate and, the ones that do regenerate, often contain diminished numbers of axons or form tangled masses of disorganized nerves in previously denervated areas of the stroma. Since the advent of phacoemulsification, the incision sizes used in cataract surgery have decreased steadily to as little as 1 mm and the risk of significant injury to the corneal innervation has been minimized. Penetrating Keratoplasty PK requires a full-thickness, 360 corneal incision that cuts all corneal nerves and results in complete denervation of the transplanted cornea. Nerve regeneration proceeds from the peripheral recipient cornea into the donor cornea very slowly and, even many years later, the innervation density of the grafted tissue remains far less than that of the host, peripheral cornea (Figure 7). Stromal nerves, in particular, regenerate very poorly after PK, perhaps because when the graft is introduced Schwann cell channels in the donor cornea are misaligned with the stromal nerve stumps in the host cornea. This contrasts to the perlimbal incisions used in cataract surgery, where nerve trunks and Schwann cell channels on opposing sides of the incision remain closely aligned. The limited nerve regeneration that does take place following PK derives mostly from small numbers of subbasal nerve fibers that elongate through epithelial bridges at the graft margin to enter directly into the donor basal epithelium. Many of the regenerated subbasal nerves are shorter than normal and exhibit atypical orientations and morphologies. Even several decades after surgery, median subbasal nerve density and corneal sensitivity in clear grafts remain significantly reduced compared to healthy corneas, and in about one-half of cases no subbasal nerves are visible by routine in vivo confocal microscopy. The return of sensation to the donor tissue is highly variable and, in many cases, hypoesthesia persists for many years after initial surgery. Fortunately, a return to normal innervation levels is not necessary for corneal clarity. Transplanted corneas can have severe nerve deficits and hypoesthesia for many years but remain clear; nevertheless, it is theorized that the impaired sensory innervation after PK may contribute to the relatively high frequency of epithelial complications observed after this procedure.

Corneal Nerves: Anatomy

157

Figure 7 Subbasal nerve fibers in human corneas as demonstrated by in vivo confocal microscopy. (a) Normal subbasal nerves in a control subject. (b) Decreased subbasal nerve fiber density after penetrating keratoplasty. Reproduced from figure 1 in Niederer, R. L., et al. (2007). Corneal innervation and cellular changes after corneal transplantation: An in vivo confocal microscopy study. Investigative Ophthalmology and Visual Science 48: 621–626. With kind permission from Investigative Ophthalmology and Visual Science (IOVS).

Mechanisms of Nerve Regrowth The mechanisms that stimulate and direct neurite outgrowth from injured and intact areas of the corneal innervation following local nerve injury are incompletely understood and, in most cases, these mechanisms are unable to restore the corneal innervation completely to its preoperative density, architecture, and sensory function. Growth factors expressed by corneal epithelial cells and released at the wound site following injury may play important roles. Nerve growth factor (NGF), a small protein that promotes the growth, survival, and maintenance of peripheral sensory nerve fibers, is constitutively expressed in corneal epithelial cells and is upregulated after epithelial wounding. Topical NGF administration stimulates corneal nerve regeneration and recovery of corneal sensitivity after PRK and LASIK in rabbits, and promotes healing of neurotrophic ulcers in clinical patients. Development of additional molecules that promote regeneration of injured nerves and restore the functional integrity of the ocular surface after corneal surgery is needed. See also: Cornea Overview; Corneal Nerves: Function; Refractive Surgery; Refractive Surgery and Inlays.

Further Reading Auran, J. D., Koester, C. J., Kleiman, N. J., et al. (1995). Scanning slit confocal microscopic observation of cell morphology and movement within the normal human anterior cornea. Ophthalmology 102: 33–41. Calvillo, M. P., McLaren, J. W., Hodge, D. O., and Bourne, W. M. (2004). Corneal reinnervation after LASIK: Prospective 3-year longitudinal

study. Investigative Ophthalmology and Visual Science 45: 3991–3996. deLeeuw, M. A. and Chan, K. Y. (1989). Corneal nerve regeneration. Correlation between morphology and restoration of sensitivity. Investigative Ophthalmology and Visual Science 30: 1980–1990. Jones, M. A. and Marfurt, C. F. (1998). Peptidergic innervation of the rat cornea. Experimental Eye Research 66: 421–435. Marfurt, C. F. (2000). Nervous control of the cornea. In: Burnstock, G. and Sillito, A. (eds.) Nervous Control of the Eye, pp. 41–92. Amsterdam: Harwood Academic Publishers. Marfurt, C. F. (2009). Peptidergic innervation of the cornea: Anatomical and functional considerations. In: Troger, J. and Kieselbach, G. (eds.) Neuropeptides in the Eye, pp. 22–37. Trivandrum, India: Research Signpost. Mu¨ller, L. J., Pels, L., and Vrensen, G. F. J. M. (1996). Ultrastructural organization of human corneal nerves. Investigative Ophthalmology and Visual Science 37: 476–488. Mu¨ller, L. J., Marfurt, C. F., Kruse, F., and Tervo, T. (2003). Corneal nerves: Structure, contents, and function. Experimental Eye Research 76: 521–542. Niederer, R. L, Perumal, D., Sherwin, T., and McGhee, C. (2007). Corneal innervation and cellular changes after corneal transplantation: An in vivo confocal microscopy study. Investigative Ophthalmology and Visual Science 48: 621–626. Oliveira-Soto, L. and Efron, N. (2001). Morphology of corneal nerves using confocal microscopy. Cornea 20: 374–384. Patel, D. V. and McGhee, C. (2005). Mapping of the normal human corneal sub-basal nerve plexus by in vivo laser scanning confocal microscopy. Investigative Ophthalmology and Visual Science 46: 4485–4488. Patel, D. V. and McGhee, N. J. (2009) In vivo confocal microscopy of human corneal nerves in health, in ocular and systemic disease, and following corneal surgery: A review. British Journal of Ophthalmology 93: 853–860. Ro´zsa, A. J. and Beuerman, R. W. (1982). Density and organization of free nerve endings in the corneal epithelium of the rabbit. Pain 14: 105–120. Tervo, T., Vannas, A., Tervo, K., and Holden, B. A. (1985). Histochemical evidence of limited reinnervation of human corneal grafts. Acta Ophthalmologica 63: 207–214. Zander, E. and Weddell, G. (1951). Observations of the innervation of the cornea. Journal of Anatomy 85: 68–99.

Corneal Nerves: Function C Belmonte, Instituto de Neurociencias de Alicante, Universidad Miguel Herna´ndez-Consejo Superior de Investigaciones Cientı´ficas, San Juan de Alicante, Spain ã 2010 Elsevier Ltd. All rights reserved.

Glossary Dysesthesias – Unpleasant abnormal sensations, spontaneous or evoked, such as burning, dryness, itching, electric shock, and pins and needles caused by the lesions of the peripheral or central nervous system. Ectopic activity – The abnormal generation of propagated nerve impulses in areas of the sensory neuron membrane that are normally not spontaneously excitable, often generated by peripheral sensory axons or cell bodies of injured sensory ganglion neurons. Ion channels – Pore-forming proteins located in the cell membrane of all living cells that selectively regulate the flow of ions (cations or anions) into and out of the cell. The ion channel switches between open and closed when the protein undergoes a conformational change, helping to establish and control a voltage gradient (membrane potential) across the plasma membrane. Nerve impulse – The electrical signal conducted along the axon of neurons by which information is conveyed within the nervous system. Neuropathic pain – Pain of pathological origin resulting of the abnormal functioning of the peripheral and/or central neural pathways involved in the detection of noxious stimuli. Sensory afferents – Peripheral branches of sensory neurons located in the dorsal root or trigeminal ganglia that innervate body tissues and carry sensory information to the brain. When covered with a myelin sheath (myelinated or A fibers), they conduct nerve impulses at rapid speeds (over 3 m s–1), while unmyelinated (or C) fibers lack the myelin sheath and have conduction velocities below 2 m s–1. Sensory receptors – The terminal portion of peripheral sensory axons specialized in the transduction of physical or chemical forces into a discharge of short-lasting (1 ms) membrane depolarizations, referred to as nerve impulses, which travel rapidly along the axon. Signal transduction pathways – The binding of extracellular signaling molecules (or ligands) to cell-surface receptors triggers ordered sequences of biochemical reactions carried out by enzymes activated by second messengers that finally result in a cellular event as, for instance, the opening or closing of specific ion channels.

The cornea is innervated by different functional types of sensory afferent fibers that are selectively activated by physical and chemical stimuli. They give rise to conscious sensations of variable quality that are referred to the eye and/or the periocular region. The cell bodies of corneal sensory afferents, most of which are of small or medium size, are located in the ipsilateral trigeminal ganglion and reach the cornea through the ciliary nerves. Sensory axons enter the cornea grouped into a variable number of radially oriented nerve bundles that branch extensively and form the sub-basal plexus below the basal epithelial cells. From this plexus, naked single fibers ascend vertically between the epithelial cells ending at variable depths. Immunocytochemical staining of the cell soma and axons of ocular sensory neurons reveals the presence of neuropeptides, principally substance P (SP) and calcitonin gene-related peptide (CGRP). Neuropeptides participate in neurogenic inflammation and promote corneal healing and/or maintenance of epithelial integrity either alone or in combination with growth factors.

Functional Properties of Corneal Sensory Receptors Functional studies of corneal sensory fibers have been performed recording electrophysiologically in single corneal nerve fibers nerve impulses evoked by physical and chemical stimuli applied to the corneal surface. The majority of corneal sensory nerve fibers, about 70%, are polymodal nociceptors (Figure 1). They are activated by near-noxious mechanical energy, heat, chemical irritants, and by a large variety of endogenous chemical mediators released by damaged corneal tissue, resident inflammatory cells, or those which leak from vessels in the periphery of the cornea (the limbus). Some of the polymodal nociceptor fibers belong to the group of thin myelinated (A-delta) nerve fibers, but most of them are unmylelinated C fibers. Polymodal nociceptors respond to their natural stimuli with a continuous, irregular discharge of nerve impulses that persist as long as the stimulus is maintained. They have a firing frequency roughly proportional to the intensity of the stimulating force. Therefore, the impulse discharge of polymodal nociceptors not only signals the presence of a noxious stimulus, but also encodes its intensity and duration in a certain degree. Corneal polymodal nociceptors have a mechanical threshold slightly lower than mechanonociceptors 158

Corneal Nerves: Function

Stimuli Thermal Cold Mechanical Heat

Ongoing activity 1

Chemical Receptor type Mechanoreceptor (low threshold)

2

Mechanonociceptor (high threshold)

3

Polymodal nociceptor

4

Cold receptor ⬚C

mN

159

⬚C

pH

2

3

4

1 4

4

Figure 1 Schematic representation of the types of sensory receptors found in the cornea. In the upper part of the figure, the presence of ongoing activity at rest and the discharge of nerve impulses to the different modalities of stimuli have been displayed for each functional type of corneal sensory receptor type. The shape and time course of the stimulus applied are represented in the lower line. In the lower part of the figure, a scheme of the anterior segment of the eye shows the location and size of the receptor areas of the different types of receptors presented in the upper part of the figure. Modified from Belmonte, C., Aracil, A., Acosta, M. C., Luna, C., and Gallar, J. (2004). Nerves and sensations from the eye surface. Ocular Surface 2: 29–34.

(Figure 2B(a)) and, when stimulated with heat, they begin to fire at temperatures over 39–40 C (Figure 2B(b)). A fraction of polymodal fibers (c. 50%) also increase their firing rate when the corneal temperature is reduced below 29 C. Many chemical agents known to excite polymodal nociceptors of other territories also activate ocular nociceptors. Acidic solutions (of pH 5.0–6.5), or gas jets containing increasing concentrations of CO2 (that in contact with the aqueous corneal surface forms carbonic acid and drops the local pH), evoke an impulse discharge in corneal polymodal nociceptors (Figure 2B(c)). Polymodal nociceptors are likely the origin of unpleasant sensations evoked by near-noxious and injurious chemical, thermal, and mechanical stimuli acting on the cornea. Approximately 15–20% of the axons innervating the cornea (all thin myelinated) respond only to mechanical forces in the order of magnitude close to that required to damage corneal epithelial cells. Accordingly, they belong to the mechanonociceptor type (Figure 2A). Fibers of this class of receptor fire only one or a few nerve impulses in response to brief or sustained indentations of the corneal surface and, often, also when the stimulus is removed. Thus, they are phasic sensory receptors that signal the

presence of the stimulus and, in a very limited degree, its intensity and duration. The threshold force required to activate mechanonociceptors is apparently low (c. 0.6 mN), far below the force that activates mechanonociceptor fibers in the skin. However, this intensity might be sufficient to damage unkeratinized corneal epithelium. Mechanonociceptors in the cornea are probably responsible for the acute, sharp sensation of pain produced by touching the corneal surface. The after-sensations of pain elicited by noxious stimuli are probably explained by the more sustained activity exhibited by polymodal nociceptors. Another category of corneal nerve fibers that represents 10–15% of the total population is cold-sensitive thermal receptors. These are A-delta and C fibers that discharge spontaneously at rest and increase their firing rate when the normal temperature of the corneal surface (c. 33 C) is reduced while they are transiently silenced upon warming. They also increase their firing rate as soon as the temperature of the cornea drops because of evaporation at the corneal surface, application of cold solutions, or blowing of cold air on the cornea. Cold receptor fibers are able to detect and encode as a change in impulse

160

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

A

(a)

(b) 47 ⬚C 35 ⬚C

(c)

B

Temp. (⬚C)

Discharge freq. (impulse/s)

15

0 30 20 10 0 50

35 50 Time (s)

0

C

Number of impulses/step

130

100

100

100

10

1

38

44

50

50

0 35

D

frequency and small temperature variations of 0.1 C or less, thus allowing for the perception of decreases in corneal temperature as a conscious sensation of innocuous cooling. Finally, it has been suggested that the cornea possesses mechanically insensitive, silent nociceptors, that is, nerve terminals that are not activated by mechanical or thermal stimuli when the tissue is intact, but, in the case of local inflammation, become responsive to these exogenous stimuli as well as to a variety of endogenous chemicals. Although the experimental evidence for their presence in the cornea is only indirect, they have been identified in virtually all other somatic tissues. Thus, it seems likely that such nociceptors also exist in the cornea. Figure 1 presents a schematic diagram of the firing response to mechanical, thermal, and chemical stimuli of the various functional types of sensory endings identified electrophysiologically in the cornea. The detection of stimuli by corneal receptor terminals, as occurs with sensory receptors of other tissues of the body, depends on membrane signaling proteins which convert the stimulus energy into a conformational change, leading to an alteration in ionic permeability and an electrical depolarization of the membrane of the nerve terminal, the generator potential. This potential change in the peripheral nerve endings gives rise to propagated nerve impulses in more proximal portions of the axon which travel centripetally to the brain. In nociceptors, most transduction molecules are ion channels that are

38

41

44

Temperature (⬚C)

47

50

Figure 2 A and B: Impulse response to different stimulus modalities of corneal sensory fibers. A: Transient discharge evoked in a mechanonociceptor fiber by a sustained mechanical indentation. B: Response of a polymodal nociceptor fiber to (a) mechanical indentations of increasing amplitude (80 and 150 mm); (b) stepwise heating of the corneal surface from 34 to 47 C; and (c) application of a drop of 10-mM acetic acid to the corneal surface. In all cases, upper traces depict the nerve impulse recordings and lower traces the stimulus waveform; time scales: A and B(a) = 1 s; B(b) = 15 s; and B(c) = 0.5 s. C and D: Sensitization to heat of corneal polymodal nociceptors. C: Peristimulus histograms of a corneal unit showing the first (upper) and second (lower) response to two identical stepwise heatings separated by 3 min. Each bar indicates the number of impulses evoked per second and the lower trace shows the stimulus waveform. E: Mean stimulus response relation of eight corneal units in response to the first (filled circles) and the second (open circles) stepwise heating. Each point represents the mean number of impulses evoked at the temperature indicated in the abscissa. The bars indicate the standard error of mean. A and B reproduced from Belmonte, C. Gallar, J. The primary nociceptive neuron: A nerve cell with many functions. In: Rowe M. J. (ed.) Somatosensory processing: From single neuron to brain imaging. Sydney, NSW: Harwood Academic Publishers. C and D reproduced from Belmonte, C. and Giraldez, F. (1981). Responses of cat corneal sensory receptors to mechanical and thermal stimulation. Journal of Physiology 321: 355–368.

Corneal Nerves: Function

directly gated by the stimulus or open by intracellular messenger systems that can, in turn, be activated by a variety of membrane proteins. Several classes of ion channels have been associated with the transduction of the various forms of energy and the production of the generator potential at nociceptor nerve terminals. A channel protein named transient receptor potential type vanilloid 1 (TRPV1) receptor, a nonselective cation channel with pronounced permeability for Ca2+ ions, is the main molecular substrate for the ability of polymodal nociceptors to respond to acid, heat, certain irritant chemicals, and inflammatory mediators. Another ion channel of the same family – transient receptor potential cation channel, subfamily A, member 1 (TRPA1) – also exhibits a prominent sensitivity to pungent chemicals. Thus, these channels serve as molecular integrators for multiple types of stimuli. The nature of the molecular entities involved in the transduction of mechanical forces at nociceptor endings is still uncertain. Stretch-activated ion channels have been identified in the membrane of mammalian sensory neurons. Extracellular matrix attachments have been proposed as the mechanism involved in the transmission of external mechanical forces to the neuronal surface and, subsequently, to ion channels activated or inactivated by stretch. Finally, TRPM8 and leak K+ channels have been suggested as cold sensors in cold thermoreceptor corneal fibers. The depolarization of sensory nerve terminals in the cornea secondary to ion channel opening has a variable duration. In the case of rapidly adapting corneal mechanosensory endings, only one or few nerve impulses are generated. Thus, they serve primarily to signal the presence of a stimulus. Polymodal nociceptors and cold receptors encode the intensity and the time course of the stimulus in their firing rate and duration. They therefore provide information to the brain on both the modality and the characteristics of the stimulus (Figure 1).

Response of Corneal Receptors to Local Inflammation One of the most prominent features of polymodal nociceptors is that sustained or repetitive stimulation tends to augment their response to new noxious stimuli. This phenomenon, referred to as sensitization, is also present in corneal nociceptors. Sensitization is characterized by a reduction of the impulse firing threshold such that impulse discharges are now evoked by intensities of the stimulus within the innocuous range. Moreover, noxious stimuli elicit a stronger and more sustained impulse discharge. Finally, sensitized corneal nociceptors often develop an irregular, low-frequency ongoing activity of nerve impulses in the absence of stimulation (Figures 2D and 2E). This spontaneous activity, coupled with the enhanced

161

responsiveness to stimulating forces, is the peripheral substrate of the augmented pain sensitivity and the presence of pain sensation in the absence of stimulus, exhibited by injured and/or inflamed corneas. The sensitization of corneal nociceptors occurs as a result of the local release by injury of a large variety of inflammatory mediators from damaged cells and from resident and invasive cells of the immune system. Endogenous chemicals include H+ and K+ ions, adenosine and adenosine triphosphate (ATP), serotonin, histamine, platelet-activating factor, bradykinin, prostaglandins, leukotrienes, tromboxanes, interleukins, tumor-necrosis factor, and nerve growth factor (NGF). These mediators are responsible for local inflammatory reactions in the cornea and conjunctiva (vasodilatation, plasma extravasation, and cell migration), and act on membrane channels and receptor proteins at corneal nociceptor nerve endings, provoking short-term changes in their opening probability. TRPV1 and TRPA1 channels at the membrane of polymodal nociceptors are direct or indirect targets for many of the inflammatory mediators released upon corneal injury. The enhanced ion flow and increased membrane excitability resulting of ion channel activation create the augmented impulse firing and reduced threshold that characterize the process of sensitization of injured corneal nerve terminals (Figures 2D and 2E). Inflammatory mediators use different signaling pathways to exert their effects on channel activity. For instance, prostaglandin E2, ATP, and adenosine or 5-hydroxytryptamine (5-HT) activate the protein kinase A (PKA) signaling pathway, whereas other mediators such as bradykinin act on protein kinase C to produce sensitization. The sensitization of corneal nociceptors develops within minutes of an acute corneal injury and normally fades when inflammation subsides. Long-lasting corneal inflammation, as occurs, for instance, in chronic dry eye, may produce more permanent changes in nociceptive terminals, including modified expression of the existing receptor molecules and expression of new ones possibly mediated by growth factors, in particular NGF. The nerve impulses evoked by noxious stimuli in corneal nociceptors not only travel centripetally to the brain, but also invade antidromically other nonstimulated peripheral branches of the parent axon, causing the release of neuropeptides contained into their peripheral endings. CGRP and SP released by depolarized nociceptor endings contribute to local inflammatory responses and amplify the proinflammatory effect of other endogenous mediators. This neurogenic inflammation mediated by the sensory nerves affects noninjured areas of the cornea and conjunctiva and explains the extension of inflammation to distant, intact tissues (conjunctiva, iris, and ciliary body) following a limited corneal lesion.

162

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Effects of Injury on Corneal Nerves After ocular surgery, accidental trauma, and nerve injury secondary to certain ocular or systemic diseases, corneal sensory nerves may become damaged at different points of their trajectory. It is well established that peripheral axotomy changes the morphology and functional properties of nerve fibers projecting to the cornea substantially. Denervated areas are invaded by outgrowths of adjacent noninjured nerve fibers and sprouts of the injured axons. Some of terminal stumps of severed axons are entrapped by scar tissue and form microneuromas. The expression of genes that encode ion channels (in particular sodium and potassium channels) and receptor proteins by primary sensory neurons is also modified by peripheral nerve injury. This alters nerve excitability and favors the development of ectopic discharges and abnormal responsiveness in damaged neurons, giving rise to peripheral neuropathic pain.

Sensitivity of the Intact Cornea Clinical exploration of the sensitivity of the cornea or conjunctiva to mechanical stimulation is normally performed by gently touching the ocular surface with a wisp of cotton, and observing the blink reflex or comparing the subjective sensation with that evoked in the fellow eye. A more quantitative approach is obtained using a calibrated hair of variable length (the Cochet–Bonnet esthesiometer). The Belmonte noncontact gas esthesiometer uses an air jet of adjustable flow and temperature that contains CO2 in a variable concentration to reduce local pH, which allows separate mechanical, thermal, or chemical stimulation applied to a restricted area of the cornea or the conjunctiva. Using these procedures, changes in normal corneal sensitivity in relation to age, sex, pregnancy, iris color, and use of contact lenses; pathological changes associated to ocular surgery; corneal pathologies such as herpes virus infections, keratitis, iritis, uveitis, and glaucoma; or systemic diseases such as fibromyalgia or diabetes have been detected in multiple clinical and experimental studies. Studies with the Belmonte esthesiometer have determined that the quality of the sensation evoked from the intact cornea and conjunctiva depends on the modality of the stimulus applied. Furthermore, electrophysiological recordings of nerve impulse activity in experimental animals applying mechanical, thermal, and acidic stimuli to the cornea with the Belmonte esthesiometer showed that each of them excited, in a variable degree, the various functional subpopulations of corneal nerve fibers. In humans, the sensations produced by suprathreshold mechanical, chemical, and heat stimulation of the cornea

possessed, in each case, a distinct quality that allowed the psychophysical identification of the stimulus modality, but always included a component of irritation. In contrast, the application of cold pulses to a corneal spot that moderately decreased (1–3 C) the local temperature evoked almost exclusively an innocuous cooling sensation, which acquired an additional component of irritation when more pronounced temperature reductions were applied. The psychophysical characteristics of sensations evoked by stimulation of the bulbar conjunctiva are quite similar to those of the cornea, except that sensitivity was comparatively lower and light mechanical stimuli were felt as nonirritating.

Sensitivity of the Injured Cornea Following surgical procedures for cataract removal, retinal detachment, and glaucoma, and particularly after surgery performed to correct refractive defects (radial keratotomy, photorefractive keratectomy, laser-assisted in situ keratomileusis, and keratoplasty), nerves directed to the cornea are usually damaged in a degree that, for photorefractive surgery, depends on the extent of the corneal lesion. Severed nerves start to regenerate soon after injury, but only a part of them succeed in penetrating the injured corneal tissue, and corneal innervation remains disorganized and reduced in number, months and even years after surgery. In parallel with the morphological changes, functional alterations have been described. Axotomized corneal neurons and nerve fibers innervating the injured cornea exhibit an altered threshold to their natural stimuli and abnormal responsiveness to mechanical, chemical, and thermal stimuli. As a result of the disturbances in peripheral innervation, corneal sensitivity to mechanical stimulation is impaired, with a remarkable increase in threshold of the denervated areas that takes months to recover, and may never return to normal values. Transplanted corneas or implanted lenticules in epikeratophakia remain totally anesthetic for years, or recover, at best, a very limited mechanical sensitivity usually restricted to the periphery of the transplant. After laser-assisted in situ keratomileusis (LASIK) surgery, corneal sensitivity to mechanical and chemical stimulation measured with the Belmonte esthesiometer remained significantly below control 3 and 6 months post-LASIK, becoming close to normal only 2 years postsurgery. However, at that time, a few patients with deep photo ablations presented severe sensory impairments. In parallel with the hypoesthesia just described, spontaneous pain sensations and dysesthesias may appear following refractive surgery. The lower sensitivity of the cornea to external corneal stimulation is explained by the reduced number and the altered threshold of corneal

Corneal Nerves: Function

sensory nerve fibers after surgery. The additional development of unpleasant dryness sensations and/or pain referred to the eye in some patients is attributable to the spontaneous firing and abnormal responsiveness of some of the injured corneal nerve fibers that were unable to fully recover their transducing capacities and exhibit abnormal excitability after injury. These disturbances seem to affect mainly polymodal nociceptor and cold corneal nerve fibers. Thus, these two functional subpopulations of corneal nerves appear to be the main peripheral source of abnormal sensations following corneal surgery. See also: Corneal Nerves: Anatomy.

163

Further Reading Belmonte, C. (2007). Eye dryness sensations after refractive surgery. Impaired tear secretion or ‘phantom’ cornea? Journal of Refractive Surgery 23: 598–602. Belmonte, C., Acosta, M. C., and Gallar, J. (2004). Neural basis of sensation in intact and injured corneas. Experimental Eye Research 78: 513–525. Belmonte, C. and Tervo, T. T. (2005). Pain in and around the eye. In: McMahon, S. and Koltzenburg, M. (eds.) Wall and Melzack’s Textbook of Pain, 5th edn., pp. 887–901. London: Elsevier. Belmonte, C. and Viana, F. (2007). Transduction and encoding of noxious stimuli. In: Schmidt, R. F. and Willis, W. (eds.) Encyclopedia of Pain, vol. 3, pp. 2515–2528. Berlin: Springer. Mu¨ller, L. J., Marfurt, C. F., Kruse, F., and Tervo, T. M. (2003). Corneal nerves: Structure, contents and function. Experimental Eye Research 76: 521–542.

Corneal Epithelium: Wound Healing Junctions, Attachment to Stroma Receptors, Matrix Metalloproteinases, Intracellular Communications G M Gordon and M E Fini, University of Southern California, Los Angeles, CA, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Apoptosis – The most common form of physiological (as opposed to pathological) cell death. Apoptosis is an active process requiring metabolic activity by the dying cell; often characterized by shrinkage of the cell, cleavage of the DNA into fragments that give a so-called laddering pattern on gels and by condensation and margination of chromatin. Cytoskeleton – General term for the internal components of animal cells which give them structural strength and motility: plant cells and bacteria use an extracellular cell wall instead. The major components of cytoskeleton are the microfilaments (of actin), microtubules (of tubulin), and intermediate filament systems in cells. Diabetic retinopathy – Damage to the retina caused by diabetes mellitus. Dystrophic epidermolysis bullosa – A rare disorder caused by a mutation in the keratin gene and is characterized by the presence of extremely fragile skin and recurrent blister formation. Filopodia – A thin protrusion from a cell, usually supported by microfilaments; may be functionally the linear equivalent of the leading lamella. Idiopathetic pulmonary fibrosis – A chronic, progressive interstitial lung disease with an unknown cause. Phagocytosis – Uptake of particulate material by a cell (endocytosis). Phenotype – The characteristics displayed by an organism under a particular set of environmental factors, regardless of the actual genotype of the organism. Recurrent corneal erosion – A disorder of the eye characterized by failure of the epithelial cells to attach to the basement membrane.

Corneal Epithelial Wound Healing: Introduction The main function of the cornea is to prevent infectious invasion and to retain a smooth, optically clear surface to

164

transmit light into the retina. The cornea has evolved a complex wound healing process for dealing with any insults to this tissue. The overall goal of corneal wound healing, as in most healing processes, is to repair the damaged tissue to resemble the unwounded tissue as closely as possible. In the cornea, this is especially important, as a failure to recapitulate an unwounded state can lead to visual impairments and a lower quality of life. The process of corneal wound healing is multifaceted involving many cell bound and secreted factors which work together to induce large-scale gene expression and phenotypic alterations in order to coordinate cell migration, proliferation, and survival rates for optimum regeneration and restoration of tissue function. While the process of corneal epithelial wound healing is well understood on the macro-molecular level (as will be discussed), there remain many micromolecular aspects to be elucidated if we are to completely understand this process.

Corneal Epithelial Wound Healing: Phases of Wound Healing Process The corneal epithelial wound healing process can be described as having three overlapping phases: the lag phase, the migration phase, and the proliferation/ restratification phase (Figure 1). The lag phase occurs immediately following any disruption of the corneal epithelium such as a scratch or puncture. During this phase, polymorphonuclear (PMN) cells from the tear fluid clear damaged and necrotic tissue from the wound area by phagocytosis, resulting in a microscopic enlargement of the wound area. These PMN cells generally disappear shortly after complete resurfacing of the wound by the epithelial cells though they can persist in cases of infection. Concurrently, a provisional matrix composed mainly of fibrin and fibronectin is secreted by the epithelial cells and the lacrimal gland and is laid down over the wound area. The provisional matrix acts as a substrate for migrating epithelial cells as well as an activator of integrin receptors on the cell surface to keep intact intracellular migration signals, mediated by the small GTPase proteins Rho and ROCK. Also, during this phase, the epithelial cells must prepare themselves for the migratory phase. For cells proximal to the wound edge, this includes a dramatic alteration

Corneal Epithelium

(a)

(b)

(c)

(d)

(e)

(f)

165

Figure 1 The stages of corneal epithelial wound healing. Histological examination of corneal resurfacing. (a) The unwounded cornea. (b) 1 h after removal of the corneal epithelium without penetrating the basement membrane cells at the wound margin have begun to alter their morphology and gene expression in preparation for migration. (c) 8 h and (d) 18 h after wounding, epithelial cells have begun to migrate over the provisional matrix. Note that there is a monolayer of migrating cells; the spacing between the cells and the basement membrane in (c) is an artifact of the sectioning process. (e) 24 h after wounding, cells have completely resurfaced the wound and begin the third phase: proliferation/restratification. By 7 days out (f), the corneal layers are repopulated and the ocular surface resembles the unwounded cornea. Arrowhead in (c) indicates leading edge; arrows in (e) indicate inflammatory cells. Curved arrows in (e) indicate fibroblasts.

in their gene expression profile including a secession of proliferation, and an alteration in cytokine and protease secretion, an alteration of the cytoskeleton including formation of lamellapodia and fillipodia, and an alteration of cell–cell and cell–matrix adhesions. Cells farther from the wound continue to proliferate to replace cells that enter into migratory phase. Additionally, during this phase, loss of superficial keratocytes due to apoptosis can be observed directly beneath the wound even when the basement membrane (BM) remains intact immediately following wounding. In a nonpenetrating wound, these stromal keratocytes are slowly replaced by proliferation of the surviving keratocytes without fibroblast activation or myofibroblast differentiation. In penetrating wounds, secreted factors induce fibroblast activation and proliferation and myofibroblast differentiation of the remaining keratocytes as will be discussed. The second phase of corneal wound healing is the migratory phase. It is important that the wound resurfaces quickly, as prolonged or chronic wound healing scenarios may lead to degradation of the BM due to an increased proteolytic profile. This may in turn lead to ulceration of the underlying stroma, activation of the keratocytes into myofibroblasts, scarring, and possible visual impairments. The migratory phase is characterized by the centripetal migration of a continuous monolayer of nondividing corneal epithelial cells (CECs) over the wound and usually lasts between 24 and 36 h depending on the size of the wound and animal species. Cells at the wound edge flatten out and migrate while those more distal to the wound are

induced to proliferate to replace the migratory cells. Cell migration rates vary by species, with rabbit CECs migrating at around 64 mm day–1 and mouse CECs at around 17 mm day–1. During this phase, epithelial cells must be anchored enough to the underlying provisional matrix so that they are not removed casually (blinking, rubbing, etc.), but if the attachments are too strong, the cells will be unable to migrate. Laminin-5 has been implicated as a major determinant in the adhesion of epithelial cells to the underlying matrix. In the unwounded cornea, processed laminin-5 is present and helps hemidesmosomes in epithelial cells adhere strongly to the collagen VII anchoring fibrils. However, following wounding, the presence of the processed form becomes secondary to that of the unprocessed form which is not bound by cells as strongly. Unprocessed laminin-5 is deposited in the provisional matrix and, along with other provisional matrix components such as fibronectin, promotes the migration of epithelial cells rather than stationary adherence. It has been proposed that epithelial cell migration is driven by an adhesion/deadhesion cycle that is regulated by two things: (1) a careful ratio of adherence molecules for attachment and proteolytic enzymes that degrade these attachments and (2) intracellular extension and contractile signals that alter the actin cytoskeleton morphology. Migratory cells first extend lamellapodia and fillopodia forward and form temporary focal adhesion complexes with the provisional matrix. Focal adhesions at the tail end of the cell are removed by a combination of proteolytic enzymes that cleave the complex and

166

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

intracellular contractile mechanisms that rip the trailing focal adhesions from the underlying stratum, sometimes even leaving part of the cell membrane behind. This cycle allows the cells to loosely adhere to the underlying matrix so that they can move due to contraction and reformation of the actin cytoskeleton. The third phase of corneal wound healing is the proliferation/restratification phase. During this phase, the basal cells which have just finished the migratory phase and were kept in a nondividing state begin to proliferate once more in order to repopulate then differentiate and restratify the corneal epithelial layers and smooth out any irregularities in the BM. Also during this phase, the epithelial basement membrane (EBM) is reformed (in penetrating or large debridement wounds), the provisional matrix is broken down, and permanent attachments are reformed between the basal epithelial cells and the EBM. In nonpenetrating wounds, this phase is usually accomplished within a few weeks. In penetrating wounds, remodeling of the ECM may persist for months and even years after the insult, and may never completely regenerate.

Corneal Nerves The cornea is one of the most densely innervated tissues of the body and corneal nerves have been shown to play a significant role in the maintenance of the cornea and in various corneal diseases. There are three networks of corneal nerves that enter the stroma and innervate the mid-stromal, bowman’s layer, and epithelial layers. Regeneration of epithelial nerves has been shown to depend on the type of insult sustained. For example, reinnervation of the epithelial layer is much faster following laser ablation of the epithelium versus manual debridement and reinnervation density correlates with renewed corneal sensitivity. Following penetrating wounds, as in a laser-assisted in situ keratomileusis (LASIK) procedure, nerve bundles in the Bowmans layer and stromal flap (those that were severed) disappear completely. Within the first 6 months, nerves first reinnervate the stromal flap, then the Bowman’s layer. A cycle of regeneration has been observed where the number of nerves is brought back to normal levels by 2 years post-LASIK, then reduced significantly by 3 years out though the reason and mechanism of this loss remains unknown. Nerve densities in the stroma below the wound remain unchanged following wounding. Although there is much greater complexity involved in penetrating wound healing scenarios, including the drastic alteration of the stromal keratocyte phenotype with the possibility of permanent scarring, the overall phases of the penetrating corneal wound healing (i.e., lag, migration, and proliferation/restratification) is similar to those of the nonpenetrating process. Important additional differences include cell invasion from limbal

blood vessels, the greater cytokine expression profile, and the much longer time frame for remodeling of the wound microenvironment due to the greater scale of remodeling that must take place in the stroma in order to maintain the optimal arrangement for light transmission.

Corneal Epithelial Wound Healing: Cell–Cell and Cell–Matrix Junctions Cell–cell and cell–matrix junctions are critical components of both unwounded and wounded cell environments and every tissue in the body has a unique profile. Along with secreted factors, these junctions are responsible for all the possible signaling triggers that can alter gene expression and affect cell phenotype, function, and survival. Additionally, these junctions in combination with cellular cytoskeletons give a tissue its structure and help define function. Following wounding, some of these junctions must be broken down while others must be formed to allow cells to migrate while still others must remain to retain as much of a barrier to the outside environment and ensure a cohesive, coordinated sheet of cells resurfaces the wound rather than cells migrating individually. There are four known types of cell–cell interactions: gap junctions, tight junctions, adherins junctions, and desmosomes. Cell–Cell Junctions: Gap Junctions Gap junctions are formed by connexins which are homoor heterohexameric proteins that form on the lateral side of cells and can form both homo- and heterotypic interactions with each other in order to connect the cytoplasm of two neighboring cells between an intercellular space. Gap junctions allow the passage of lowmolecular-weight proteins (water, ions, secondary messengers, electrical impulses, and low-molecular-weight (<1 kDa) metabolites and nutrients) to pass freely from one cell to another, allowing cells to communicate quickly with each other in order to coordinate such physiological processes as development and regeneration. Connexins(cxs) are named according to their molecular weight; thus, connxin43 is a 43-kDa protein. Multiple connexins have been observed in the cornea by studying various mammalian species (Table 1). Connexin 43 is the most well-documented connexin in the corneal epithelium and it is found primarily in the basal epithelial cells with expression decreasing progressively toward the superficial layers. Other connexins such as cxs26, 30, 31.1, 37, and 50 have been observed with spatially distinct patterns in the unwounded cornea. No connexins have been detected in migrating CECs following wounding though gap junctions in cells distal to the wound remain; thus, gap junctions in cells proximal to the wound are broken down during the lag phase and expression is downregulated throughout the

Corneal Epithelium Table 1 cornea

Cell–cell junctions in the unwounded and wounded

Unwounded Gap Jxns C26 C30 C30.3 C31 C31.1 C31.9 C32 C36 C37 c40 C40.1 C43 C45 C46 C47 C50 C58 C62 Tight Jxns Adherens Jxns Desmosomes

Wounded

retention of tight junctions presumably helps ensure the integrity of the corneal surface as much as possible and helps to keep cells migrating together as a cohesive sheet. Cell–Cell Junctions: Adherens Junctions

X X

X

X

X

X

X X X

167

X X

migration phase. Following resurfacing, gap junction reformation correlates with the reappearance of laminin-1 in the EBM. Expression of connexin43 has been attributed to Rho, but not ROCK signaling. The functional significance of the loss of gap junctions in migrating cells has yet to be determined. Cell–Cell Junctions: Tight Junctions Tight junctions are homeotypic interactions by transmembrane protein structures which link the actin cytoskeleton of neighboring cells and act like cellular gaskets. These structures are composed mainly of claudin occludin and junctional adhesion molecule A ( JAMA) and have several purposes: they link neighboring cells, separate proteins in the apical cell membrane from those in the basal–lateral, and they prevent their passage of ions and other small molecules from penetrating between neighboring cells, thus contributing to corneal deturgesence. In the cornea, tight junctions have been detected in the superficial and wing cells. Early work in the field had found that tight junctions were reduced at the wound margin following wounding but reformed quickly behind the migrating front of epithelia. However, more recently, it has been shown that these junctions persist following corneal epithelial ablation with no significant alteration in expression. This newer finding correlates well with an older finding that showed a component of the tight junction complex, ZO-1, is upregulated following removal of superficial CECs. While there is no direct evidence yet,

Adherens junctions are also found on the lateral side of cells and keep neighboring cells firmly attached to each other by anchoring to the actin cytoskeleton. Adherens junctions in the cornea are formed by homeotypic dimers of transmembrane E-cadherin molecules which bind each other in the intercellular space. E-cadherin is calcium dependent and is anchored intracellularly to the actin cytoskeleton by vinculin and a- and b-catenin molecules. E-cadherin-containing adherens junctions have been found throughout the cornea epithelium, and wounding does not seem to affect the expression or distribution of these junctions. Similar to tight junctions, adherens junctions may be retained in order to keep cells migrating together as a cohesive sheet, though again, direct proof is still lacking. Cell–Cell Junctions: Desmosomes The final type of cell–cell junction is the desmosome. Desmosomes are similar to adherens junctions in that they both function to keep neighboring cells attached to each other and they are both composed of molecules from the cadherin family. However, unlike adherens junctions which are composed of the cadherin family member E-cadherin, desmosomes are composed mainly of the caherin family members desmoglein and desmocollin. Also, while the adherens junction is anchored to the actin cytoskeleton, desmosomes are anchored by desmoplakin, plakoglobin, and plakophillin to the intermediate filament (IF), or cytokeratin, cytoskeleton. Interestingly, unlike adherens junctions, desmosomes do alter their expression pattern following wounding. It has been found that similar to connexin43, desmoglein is downregulated in migrating CECs, and this downregulation persists until resurfacing is complete and laminin-1 reappears below the epithelial cells. Like gap junctions, the functional significance of the loss of desmosomes in migrating cells has yet to be determined, though it may be necessary for the morphological changes needed for migration. Cell–Matrix Junctions: Hemidesmosomes and Adhesion Complexes In addition to cell–cell interactions, cell–matrix interactions are also critical components of tissue function and structure in normal, wounded, and diseased tissue, and a lack of adherence to an underlying matrix may result in apoptosis by ceratin cell types (such as epithelia) and a separation of the dermal and epidermal layers. Cells can be connected to an underlying extracellular matrix by

168

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

adhesions that are anchored intracellularly to either the actin or IF cytoskeleton, both of which utilize integrins to form these anchors. Integrins are not only structural proteins which affect cell attachment, but they are also potent signaling molecules that can affect cell proliferation, migration, and survival properties. Intergins form heterodimers of an a- and b-chain and the different combination of chains determines the extracellular binding partner; some have only one substrate while others have multiple substrates with differing binding kinetics. In mammals, there are 19 known a-chains and 9 known b-chains which can assemble into about 25 distinct integrins. Integrins a2b1, a3b1, a6b1, a9b1, avb1, avb5, avb6, and a6b4 have all been localized to the corneal epithelium with the strongest expression being found in the basal epithelial cells and getting progressively weaker as the cells get closer to being sloughed off into the tear fluid. This progressive loss of integrins ensures that cells that migrate away from the basal epithelium proliferate less and provide a less adhesive surface for microbial adherence and infection of the corneal surface. Hemidesmosomes are present at the basal membrane of unwounded basal epithelial cells and are composed of the a6b4 integrin as well as adaptor proteins for linking this protein to the IF cytoskeleton. In the unwounded cornea, hemidesmosomes are tightly bound to anchoring fibers composed of collagen VII which pass perpendicular through the BM then make a 90 turn and wrap around collagen III bundles. Following wounding, hemidesmosomes are broken down and are no longer detectable for 70–200 mm outside the wound area in order to allow cells to begin the migratory phase. Although the hemidesmosomes are absent, a6b4 integrin expression is actually increased during resurfacing, though these integrins are no longer associated with hemidesmosomes but instead form focal adhesion complexes (discussed earlier) which now connect the components of the underlying provisional matrix such as unprocessed laminin-5 with the actin cytoskeleton. In penetrating wounds, reformation of the basal lamina and hemidesmosome complexes occurs concurrently in about 6–8 weeks, though studies in monkeys have shown that reformation of the anchoring fibrils may take as long as 18 months to reform. All of these time frames are approximations and can be affected by variables such as the size of the wound, the age and species of the subject, and if there are any complicating factors such as infection or genetic abnormalities.

Secreted Factors Involved in Corneal Epithelial Wound Healing Along with cell–cell and cell–matrix junctions, secreted factors such as cytokines, growth factors, and proteases are also responsible for regulating normal tissue maintenance

and wound healing properties. Cytokines and growth factors are small molecules that can bind and activate their cognate cell surface receptors. These activated receptors then transduce the signal to the nucleus via various intracellular signaling pathways such as the mitogen-activated protein kinase (MAPK), Smad, and nuclear factor kappa-lightchain-enhancer of activated B cells (NF-kB) pathways. However, signaling is not necessarily linear and cross-talk between the pathways can and often does occur. In other epithelia of the body, these factors are usually secreted by platelets or other cell types that enter a wound from the blood stream. The cornea, however, is an avascular tissue and thus all signals must come from either the tear fluid or the epithelial and stromal cells themselves. Corneal epithelial and stromal cells have been shown to secrete a host of cytokines and growth factors. Transcriptional analysis of primary corneal epithelial cultures has found insulin-like growth factor (IGF), epidermal growth factor (EGF), transforming growth factor (TGF)-a, b-fibroblast growth factor (FGF), TGF-b1, TGF-b2; in addition, all of their cognate receptors are expressed by both corneal epithelial and stromal cells, though to varying degrees, and may act via both autocrine and paracrine mechanisms. Some other important factors have more specific localization patterns and generally act via paracrine signaling. For example, both keratinocyte growth factor (KGF) and hepatocyte growth factor (HGF) are only expressed by stromal keratocytes though their cognate receptors are much more preferentially expressed by epithelial cells, while plateletderived growth factor (PDGF)-bb is expressed by epithelial cells and its cognate receptor is only expressed in stromal cells. The inflammatory cytokines interleukin (IL)-1a and IL-1b are only expressed by epithelial cells in the unwounded state, but their cognate receptor is present on both epithelial and stromal cells. However, stimulation of keratocytes by IL-1 as occurs during penetrating wounds can induce an autocrine feedback loop; thus, keratocytes can express IL-1. Furthermore, this IL-1 feedback loop also induces other secreted factors such as HGF and KGF expression by keratocytes. Cytokines can induce gene expression changes that can alter the proliferation, migration, and survival properties of a cell. For example, it has been shown that KGF and EGF can induce CEC proliferation. Also, while neither EGF, PDGF-bb, IL-1, nor TNF-a alone can induce epithelial cell migration, when combined with fibronectin as a substrate, these factors can all significantly speed up migration rates as compared to fibronectin alone. This further emphasizes the complex and combinatorial signaling that controls corneal epithelia wound healing. The role of the TGF-b family in epithelial migration is controversial with some labs showing that it enhances epithelial migration rates, and others finding that it slows down resurfacing rates. Further studies must be done to reconcile these seemingly opposite results.

Corneal Epithelium

Matrix Metalloproteinases Matrix metalloproteinases (MMPs) are a family of structurally related zinc-dependent proteases with overlapping substrate specificities. MMPs are known regulators and effectors of many cellular processes, including tumor progression, development, and epithelial wound healing as well as normal tissue maintenance. The MMP family consists of over 20 members whose substrates have come to include all extracellular membrane proteins as well as many cell surface proteins (i.e., cadherins, heparin-binding EGF (HB-EGF), etc.) and secreted proteins (i.e., TGF-b, IL-1, TNF-a, etc.) which can be both activated and degraded by MMPs. Thus, MMPs can affect gene expression by alterating intracellular (nuclear) signaling by cleavage of cell–cell and cell–matrix interactions as well as by altering the active cytokine and growth factor microenvironment. Due to their broad substrate specificity and potent proteolytic activity, MMPs must be tightly controlled, and are regulated at the transcriptional and posttranslational levels. MMPs can be either secreted or membrane bound and contain a propeptide at their N-terminal that must be removed before they can become proteolytically active. MMPs are critical components of the wound healing process, as global MMP inhibition by pharmacological agents results in a failure to resurface a wound. Similarly, an overexpression of MMPs is also involved in many epithelial disorders such as idiopathic pulmonary fibrosis in the lung, dystrophic epidermolysis bullosa in the skin, and diabetic retinopathy and recurrent corneal erosion in the eye. Therefore, a precise level of MMP expression is necessary for normal tissue maintenance and optimal wound repair. In the unwounded cornea, most MMPs are expressed at basal or undetectable levels, except for MMP-7 and -14 which are expressed constitutively. Following removal of the epithelium, several MMPs are upregulated and various family members take on a distinct localization pattern, though the expression patterns of MMP-7 and -14 do not change significantly. MMP-9 is the most well-studied family member and is strongly upregulated at the very tip of the migrating epithelial cells. MMP-10 becomes upregulated by all migratory cells, and MMP-13 expression is upregulated by cells throughout the wound area and in the periphery of the wound. MMP-12 mRNA upregulation has also been detected in the peripheral cells, but this upregulation has not been detected for the actual protein. Penetration of the EBM induces expression of more MMPs such as MMP-2, -3, and possibly -8 which are used to remodel the stromal compartment. Following resurfacing of nonpenetrating wounds, MMP levels are generally decreased, though MMP-9 expression persists and spreads distally from the wound closure with the timing of its expression correlating with degradation of

169

the provisional matrix. This correlation has been confirmed using MMP-9 knockout (KO) mice which are unable to degrade the fibrin in the provisional matrix which remains as a corneal haze. However, a recent study in human eyes has found that people displaying uncomplicated LASIK surgeries (penetrating wound) may, up to 7 years later, still have EBM irregularities around the wound edges that lead to stromal–epithelial contact causing MMP expression and indicating a continuing wound healing process. While MMP-9 is critical for overall corneal wound regeneration, it is not critical for corneal re-epithelialization, and in fact acts to slow the rate of re-epithelialization. Other MMPs, however, are critical for corneal resurfacing as global inhibition of MMPs retards resurfacing. However, which specific MMPs are necessary, what their substrates are, and why these MMPs are necessary is still unknown. These questions are especially difficult to answer because loss of an MMP, as in a knockout or knockdown model, can be compensated for by other MMPs due to the overlapping substrate specificities as indicated by the low number of MMP KO mice displaying drastically different phenotypes.

Corneal Epithelial Wound Healing: Conclusion While the corneal epithelial wound healing process shares some similar properties with skin, lung, and gut epithelial wound healing, there are important differences. For example, whereas skin epithelia can heal by a scarring or vascularization process, the corneal wound healing process must minimize these parameters in order to retain optimal visual clarity. Formation of scar tissue on the molecular level means the accumulation of a-smooth muscle actin-containing fibroblasts also known as granular tissue myofibroblasts that form the fibrous tissue in the early scar before the secretion and formation of the older collagen scar tissue. As the central cornea is an avascular tissue, formation of a-smooth muscle actin expressing cells is initially derived solely from stromal keratocytes. Stromal keratocytes are induced to differentiate into a-smooth muscle actin expressing cells by secreted factors expressed by the CECs and present in the tear fluid such as TGF-b. In the unwounded eye, the EBM binds TGF-b and acts as a physical barrier to prevent the stromal keratocytes from being activated. Over a remodeling period of weeks, months, or even years, myofibroblasts can slowly disappear from a wound site via an unknown mechanism, possibly apoptosis. This disappearance is by no means guaranteed and even if it does occur, it may or may not improve visual acuity depending on whether or not a collagenous scar tissue has formed. The severity of the wound response is correlated with the severity of the wound; wounds that do not penetrate the EBM tend to regenerate completely, whereas wounds

170

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

that do penetrate the EBM tend to induce myofibroblast differentiation, require much longer to heal, and may never quite regain the original architecture. Thus, the most important factor involved in corneal wound healing is whether or not the EBM has been disrupted. While penetration of the BM is arguably the most important factor in corneal wound healing, it is by no means the only determining factor; the size of the wound, the age of the subject, and the causative agent are also important factors. Additionally, any defect in the tear fluid composition or integrity may further impair the wound healing process. All of these factors can affect the cellular microenvironment by alteration of cell adhesion molecules (CAM) and secreted factors which both coordinate and drive the wound healing process. See also: Corneal Epithelium: Transport and Permeability; Corneal Nerves: Anatomy; Refractive Surgery and Inlays.

Further Reading Fini, M. E., Cook, J. R., and Mohan, R. (1998). Proteolytic mechanisms in corneal ulceration and repair. Archives of Dermatological Research 290(supplement): S12–S23. Fini, M. E. and Stramer, B. M. (2005). How the cornea heals: Corneaspecific repair mechanisms affecting surgical outcomes. Cornea 24 (8 supplement): S2–S11. Gipson, I. K., Spurr-Michaud, S. J., and Tisdale, A. S. (1988). Hemidesmosomes and anchoring fibril collagen appear

synchronously during development and wound healing. Developmental Biology 126(2): 253–262. Imanishi, J., Kamiyama, K., Iguchi, I., et al. (2000). Growth factors: Importance in wound healing and maintenance of transparency of the cornea. Progress in Retinal and Eye Research 19(1): 113–129. Li, D. Q. and Tseng, C. G. (1995). Three patterns of cytokine expression potentially involved in epithelial–fibroblast interactions of human ocular surface. Journal of Cell Physiology 163(1): 61–79. Lu, L., Reinach, P. S., and Kao, W. W.-Y. (2001). Corneal epithelial wound healing. Experimental Biology and Medicine (Maywood) 226(7): 653–664. Mohan, R., Chintala, S. K., Jung, J. C., et al. (2002). Matrix metalloproteinase gelatinase B (MMP-9) coordinates and effects epithelial regeneration. Journal of Biological Chemistry 277(3): 2065–2072. Schultz, G. S., White, M., Mitchell, R., et al. (1987). Epithelial wound healing enhanced by transforming growth factor-alpha and vaccinia growth factor. Science 235(4786): 350–352. Sivak, J. M. and Fini, M. E. (2002). MMPs in the eye: Emerging roles for matrix metalloproteinases in ocular physiology. Progress in Retinal and Eye Research 21(1): 1–14. Steele, C. (1999). Corneal wound healing: A review. Part I. Optometry Today 24: 28–32. Stepp, M. A. (2006). Corneal integrins and their functions. Experimental Eye Research 83(1): 3–15. Stepp, M. A., Spurr-Michaud, S., and Gipson, I. K. (1993). Integrins in the wounded and unwounded stratified squamous epithelium of the cornea. Investigative Ophthalmology and Visual Science 34(5): 1829–1844. Stramer, B. M., Zieske, J. D., Jung, J. C., Austin, J. S., and Fini, M. E. (2003). Molecular mechanisms controlling the fibrotic repair phenotype in cornea: Implications for surgical outcomes. Investigative Ophthalmology and Visual Science 44(10): 4237–4246. Suzuki, K., Saito, J., Yanai, R., et al. (2003). Cell–matrix and cell–cell interactions during corneal epithelial wound healing. Progress in Retinal and Eye Research 22(2): 113–133. Zieske, J. D. (2001). Extracellular matrix and wound healing. Current Opinion in Ophthalmology 12(4): 237–241.

Corneal Epithelium: Transport and Permeability P S Reinach, The State University of New York, New York, NY, USA F Zhang, The State University of New York, New York, NY, USA J E Capo´-Aponte, U.S. Army Aeromedical Research Laboratory (USAARL), Fort Rucker, AL, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Barrier – Protein-containing formations between cells in the more superficial layers of the corneal epithelium to prevent infiltration of pathogens into the underlying stroma. Deturgescence – Physiological state of corneal hydration required for optimal transparency. Electrochemical equilibrium – Membrane voltage and intracellular ionic activity at which there is no net flux of a specific ion across a membrane. Net flux – Algebraic difference between unidirectional ionic fluxes across a membrane, which is indicative of active ion-transport activity. Transient receptor potential protein – Membrane proteins in a superfamily of 27 different genes that arrange themselves in tetrameric configurations to form plasma membrane channels, which can be activated by a large variety of stimuli.

Introduction Corneal epithelial ion transport and permeability underlie the ability of this tissue to maintain corneal transparency. In this article, we describe how different types of receptors modulate through second messenger-signaling control of ion transporter function and cell membrane permeability. Figure 1 depicts a schematic summarizing the current picture of ion transporters and channels mediating control of cornea epithelial renewal and tissue transparency. To provide insight into the outcome of dysregulation of transporter activity and permeability, we discuss strategies that may circumvent these pathophysiological consequences. The cornea provides about 75% of the total refractive power required for normal vision. This function depends on the ability of the cornea to remain transparent. The corneal epithelial layer undergoes continuous renewal every 14–28 days. This process preserves its integrity and assures that the cornea is protected from environmental pathogenic infiltration. Such protection is provided by tight junctional continuity forming a moderately high resistance barrier between neighboring cells. This resistance is essentially only selectively permeable to low-molecular-weight

ionic species. As the epithelial top layers continuously undergo terminal differentiation followed by being sloughed off, the maintenance of tight junctional integrity is critical for the corneal epithelial to provide its barrier function. Any disruption of upper cell layer apposition that is not rapidly repaired through wound healing renders the cornea vulnerable to infection, swelling, and opacification leading to losses in visual acuity. Losses in tissue transparency occur since swelling of the corneal ground substance leads to disruption of extracellular matrix organization and an irregular corneal surface. These changes cause impinging light to be reflected or scattered, instead of being refracted. Therefore, studies directed toward understanding how the corneal undergoes continuous renewal are relevant for delineating strategies to restore corneal transparency in a clinical setting.

Corneal Hydration Control and Transparency In humans, the cornea is approximately 500–550 mm thick and is composed of three layers: epithelium, stroma, and endothelium. These tissues have different embryonic origin. Corneal thickness is dependent on the hydration state of the stromal ground substance, whose physiochemical properties cause continuous fluid imbibition. Excessive fluid uptake by the stromal ground substance lying between collagen lamellae results in corneal swelling and opacity. The stromal ground substance continuously imbibes fluid from the tears and the aqueous humor in the anterior chamber facing the endothelium. This imbibing process is equivalent to a negative pressure of approximately 60 mmHg. In order for this suction effect not to lead to excessive swelling, it is essential that the corneal epithelium, in concert, with the underlying endothelial layer mediate osmotically coupled water flow outward from the stroma through net ion transport. The endothelial layer facing the anterior chamber provides most of the dehydrating function whereas the epithelial layer contribution is relatively minor. In other words, epithelial layermediated net fluid transport toward the tears from the stroma appears to play a fine-tuning role in maintaining a state of corneal hydration commensurate with normal vision. Nevertheless, under stimulated conditions, the rabbit epithelial layer can provide up to 25% of the total

171

172

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Ca2+ Basolateral membrane H+

Ca2+

K+ Na+

K+

2K+

Na+

3Na+

Na+ = 15 mM Tears Na+ H+

Ca2+ ~ 50 nM

Cl− Apical membrane Tight junction

Na+ K+ 2Cl−

K+ = 150 mM Cl− = 30 mM Cl−

K+ Cl−

Stroma

HCO−3 Cl−

Paracellular

Figure 1 Current model describing ion transporters and channels in corneal epithelium: Arrow heads point in directions of net fluxes. At the tear-side facing apical membrane, there is a Na+:H+ exchanger and a conductive pathway for net Cl and K+ efflux. There is also Na+ entry at this membrane. At the stromal and paracellular sides is the basolateral membrane. It has a K+ efflux pathway along with the Na+–K+ pump coupled to an electroneutral Na+:K+:2Cl cotransporter. A Cl :HCO3 electroneutral cotransporter and a Ca2+ channel reflective of different subtypes are shown. A K+:Cl cotransporter is also indicated.

dehydrating function. However, if the epithelial layer integrity is damaged, fluid leakage from the tears cannot be compensated by the dehydrating function of the innermost endothelial layer. In this case, the stroma will continuously swell. Therefore, epithelial wound healing through cell proliferation, migration, and tight junction renewal is crucial for maintaining the corneal barrier function.

Importance of Transport Mechanisms to Epithelial Function The corneal epithelial electrical resistance is dependent on the intactness of its tight junctions forming appositions in the upper layers of the epithelial functional syncytium. In order to provide an effective barrier function, the electrical resistance of the corneal epithelium has to be relatively high (i.e., at least 1 kO cm2). In addition, the ability of the epithelium to mediate net ion transport from the stroma into the tears is dependent on tight junctional resistance. If the resistance falls due to injury, there is a corresponding decrease in net ion transport, resulting in compromise of the barrier function followed by a decline in osmotically coupled fluid flow toward the tears.

In all species studied, the corneal epithelial layer elicits secondary active chloride (Cl ) transport from the stroma to the tears whereas net sodium (Na+) transport occurs in the opposite direction, from the tears to the stroma. There is variability in the magnitude of active Na+ transport due to differences in the Na+ permselectivity of the tear-sidefacing apical membrane. In rabbits and humans, the relative Na+ and Cl permeabilities are similar to one another whereas in the amphibian this membrane is essentially Cl permselective. There is large variability among different species with regard to the ratio between tear-directed Cl transport and stroma-directed Na+ transport. In the rabbit and human, this ratio under baseline conditions is about 50%, whereas in the amphibian (i.e., toad and bullfrog) it is at least 90%. Despite such differences, osmotically coupled fluid flow across this layer has been identified both in the isolated rabbit and amphibian cornea suggesting that net Cl fluxes toward the tears exceed stroma-directed Na+ fluxes. Corneal transparency is dependent on net influx and efflux of inorganic and organic osmolytes. For example, amino acid uptake mechanisms have been identified in this tissue. By definition, an active uptake mechanism depends on metabolic energy generated by aerobic and anaerobic metabolism. Its metabolic dependence is evident, since exposure to metabolic inhibitors results in

Corneal Epithelium: Transport and Permeability

corneal swelling and the development of opacification. These changes occur because corneal deturgescence (i.e., physiological hydration state) and maintenance of epithelial health are dependent on the ability of this tissue to: (1) elicit osmolyte and drug extrusion against opposing electrochemical gradients and (2) accumulate organic substrates to levels above those in the extracellular bathing solution. This array of very different transport functions are performed by transporters of distinct ionic and substrate selectivity.

Primary and Secondary Ionic Transport Mechanisms The corneal epithelial cells express both primary and secondary ionic transport mechanisms. As in all epithelial cells, Na+:K+ ATPase expression occurs along the basolateral membrane facing the stroma and the paracellular pathway between the neighboring epithelial cells. This transport mechanism is primary since its ATPase activity directly elicits electrogenic Na+ extrusion into the stromal and paracellular media. Such efflux is coupled to intracellular K+ accumulation above its predicted electrochemical value. On the other hand, secondary ionic transporter function is driven by the Na+ and K+ electrochemical gradients established by Na+:K+ ATPase activity. In all species, there is Na+:K+:2Cl co-transport activity in the basolateral membrane, which is reflective of secondary active ion transport function. This cotransporter functions to accumulate intracellular Cl to levels above its predicted electrochemical values. Net Cl transport into the tears occurs as a consequence of its efflux by electrodiffusion across the tear-side facing essentially Cl permselective apical membrane. In order to maintain charge neutrality, Na+ moves in parallel as a counter ion through the paracellular pathway. Another primary active ion transporter identified in the corneal epithelium is the H+ pump. Its activity in combination with two different types of cotransporters is required for mediating regulation of intracellular pH. These cotransporters are Cl :HCO3 and Na+:H+ antiporters whose transport directionality is determined by the electrochemical gradients of the involved ions. Specifically, the Cl :HCO3 antiport will drive HCO3 out of the cell, provided there is sufficient carbonic anhydrase activity to raise intracellular bicarbonate levels above those in the external medium. Under such conditions, Cl will be taken up from the stroma into the cell interior in exchange for bicarbonate efflux into the external medium. In this scenario, intracellular pH falls resulting in increases in the activity of two different alkalinizing mechanisms: Na+:H+ antiport activity in combination with H+ pump ATPase activity. The Na+:H+ antiport can provide an alkalinizing function if the Na+:K+ pump

173

activity is sufficient to establish a steep enough Na+ gradient between the external milieu and the intracellular compartment to exchange H+ from the cell interior for downhill Na+ movement into the cell interior. Cl efflux out of the epithelial cells across the apical membrane is dependent on membrane voltage electronegativity relative to the bathing solution. Numerous studies have shown that the intracellular membrane voltage is about 50 mV. This value is dictated to a large extent by the activity of the electrogenic Na+:K+ pump and the relative K+ permeability of the basolateral membrane. The electrogenic Na+:K+ exchange pump contributes since its pump Na+:K+ stoichiometry exceeds a value of 1. In other words, with each pump cycle, more Na+ ions are extruded than K+ ions are taken up into the cell interior from the stroma. In any case, due to the activity of the Na+:K+ pump, Cl is accumulated into the epithelial cells above its predicted electrochemical equilibrium value. In turn, K+ diffuses downward and outward, essentially across the basolateral rather than the apical membrane since the former barrier has a much higher K+ permselectivity. This downward movement of K+ establishes a negative membrane voltage which is dependent on the activity of the Na+:K+ pump and the relative basolateral membrane K+ permeability. Another factor affecting the level to which K+ is elevated above electrochemical equilibrium is the tightness of coupling between the Na+:K+:2Cl cotransporter and the Na+:K+ pump. Stimulation of the Na+:K+:2Cl cotransporter may occur as a consequence of increases in Na+: K+ pump activity leading to enhancement of intracellular K+ accumulation. A rise in K+ can then be compensated for by a rise in basolateral K+ membrane permeability, which will increase outward directed K+ efflux and the membrane voltage electronegativity. This chain of events will stimulate net Cl efflux from the stroma into the tears since the driving force for Cl efflux into the tears, across the highly Cl permselective apical membrane, is the magnitude of the membrane voltage electronegativity. Therefore, it is through the concerted activity of the aforementioned transport mechanisms, along with appropriate regulation of the cell-limiting membranes, that the corneal epithelium can elicit net Cl transport from the stroma toward the tears and regulate its intracellular pH within bounds that are required for the maintenance of anterior ocular surface health.

Coupling of Ionic Transport Mechanisms to Stromal Deturgescence It is commonly thought that the driving force for fluid transport out from the stroma across the epithelial cells into the tears is local osmosis. The fact that the corneal epithelium elicits net Cl transport into the tears suggests

174

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

that this ionic flux may be sufficient to account for the magnitude of outward-directed fluid transport. This expectation has been proven to be correct since, under open circuit conditions, net Cl transport is commensurate with measured isotonic fluid transport. The dependence of transepithelial fluid transport into the tears on net Cl transport was validated by showing that drugs which either stimulate or inhibit net Cl transport have corresponding effects on fluid transport. The fact that there is a coupling between fluid and net Cl transport prompted investigators to probe for receptor-mediated control of net Cl transport since corneal epithelial health is dependent on innervation of this tissue. This dependence suggests that neurotransmitter release from nerve endings mediate – through receptor stimulation – responses supportive of corneal epithelial functions. Such a notion was validated by showing that loss of neural input results in impaired corneal epithelial regeneration and recurrent erosion. Accordingly, it has also been shown that the corneal epithelium expresses receptor subtypes responsive to adrenergic and cholinergic agonists released from sympathetic and parasympathetic branches of the autonomic nervous system. Other studies indicated that stimulation of these receptors elicit regulation of corneal epithelial renewal, ion transport activity, transparency, and fluid transport. The regulation of cell volume by the aforementioned ionic transport mechanisms underlies the ability of the corneal epithelium to elicit net fluid transport from the stroma into the tears. It is possible that fluid transport is dependent on serial repetitive swelling and shrinking of cell volume. In this way, swelling occurs as a result of initial stimulation of net osmolyte uptake. Such rises occur in response to increases in Na+:K+ pump activity leading to rises in sodium chloride (NaCl) and potassium chloride (KCl) uptake through the Na+:K+:2Cl cotransporter. As a consequence of an increase in osmolyte uptake, fluid flow into the cells increases. Subsequently, the epithelial cell volume decreases as a result of stimulation of ionic efflux (i.e., KCl) into the tears. The swelling response is referred to as regulatory volume increase (RVI) due to increases in NaCl and KCl uptake. Subsequent shrinkage occurs as a consequence of RVI and is dependent on increases in KCl efflux leading to osmotically coupled fluid loss. This cell shrinkage is known as regulatory volume decrease (RVD). Therefore, fluid transport is tightly regulated by alternating stimulation and inhibition of osmolyte uptake and efflux into the corneal epithelium. How such tight and time-dependent control of these different transport mechanisms is mediated is unknown. It may be dependent on the ability of different receptor types to elicit – through a variety of second messenger pathways – rapid changes in the activities of these ion transporters mediating alternating RVI and RVD responses.

In addition to mediating fluid transport under isotonic conditions as a possible result of differential activation and inhibition of ion-transport activity and membrane permeability, RVD and RVI responses can be induced through variations in bathing solution osmolarity similar to those encountered in daily living and those identified in some types of dry eye disease. In order for cell volume regulatory responses to change fluid transport rates and thereby maintain corneal homeostasis, despite being exposed to an environmental hypertonic or hypotonic challenge, it is also necessary to have a coordinated regulation of ion transporters and parallel cell membrane permeability. Such regulation occurs through receptormediated events that can modulate ion transporter rates to either increase or decrease net ionic influx or efflux from corneal epithelial cells. These receptors elicit control of ion transporter function and cell permeability through second messengers that either concomitantly increase or inhibit ion transport rates through stimulation or inhibition of active ion transporter, cotransporters, and cell membrane permeability. Such coordinated control is required for the maintenance of corneal transparency since they underlie RVD and RVI responses driving fluid out of stroma into the tears. Depending on the type of osmotic challenge, these responses sustain corneal epithelial barrier function. However, in individuals afflicted with dry eye disease, one problem is compromise of barrier function leading to stromal infection. Even though, in these patients, there may be activation of a RVI response due to chronic exposure to hypertonic tears, it may not be adequate to restore isotonic cell volume resulting in disruption of barrier function. This suggestion has prompted a host of studies over the last 30 years that are focused on characterizing receptor-mediated regulation of corneal epithelial active ion transport underlying regulatory volume responses. Their results have pointed investigators in directions that may lead to the identification of novel strategies for improving the rates of corneal epithelial renewal as well as restoring its transparency and refractive properties following an injury.

Other Transport Mechanisms The limited corneal epithelial layer permeability presents a formidable barrier to solute and drug permeation into the ocular interior. This hindrance has prompted efforts to probe for the expression of transporters that could facilitate their uptake into the eye. A number of different transporters were identified that can elicit this effect. They include a carnitine/organic cation transporter and a sodium-dependent amino acid transporter. On the other hand, rather than identifying a drug-influx mechanism,

Corneal Epithelium: Transport and Permeability

a multi-drug-resistant (MDR)-1 efflux mechanism has been described. Such an efflux process does not promote drug uptake into the ocular interior if they are substrates for the MDR-1 pumps. Therefore, improving drug delivery into the ocular interior is in most cases focused on increasing their hydrophobicity so that they can more readily permeate by membrane diffusion. One example of such an improvement is exemplified by rendering epinephrine more lipid soluble through its derivatization with a lipid. An alternative is to inhibit MDR-1-elicited drug efflux. However, none of the currently used inhibitors have adequate selectivity for this purpose. Receptor-Mediated Control of Corneal Epithelial Ionic Transport Functions Adrenergic subtype (a1, a2, and b), serotonergic, and cholinergic receptor-linked functions have been identified in the corneal epithelium. Adrenergic receptors mediate their control of ionic transport and membrane permselectivity through second messenger pathways involving increases in intracellular Ca2+ and modulation of adenylate cyclase activity. Such variations change either the basolateral or apical membrane permeabilities resulting in modulation of active Cl transport through changes in the intracellular potential difference. For example, b-adrenergic receptor stimulation has a corresponding effect on net Cl transport by enhancing K+ basolateral membrane permselectivity as well as increasing Cl electrodiffusion across the apical membrane into the tears. Less is known about the signaling pathways eliciting cholinergic and serotonergic receptor control of net ion transport. The corneal epithelium and accessory anterior ocular tissues express a host of cytokines that are critical to the maintenance of corneal epithelial functions. In particular, cytokine expression is critical for inducing control of cell proliferation, migration, and terminal differentiation. Their importance has become self-evident from studies on the mechanisms of corneal epithelial wound healing induced by injury. Numerous cytokine expression levels are upregulated to hasten corneal epithelial wound healing through stimulation of cell migration and proliferation. Such a realization has heightened the interest in determining the mechanisms of regulation of cell-signaling pathways linking their cognate receptor activation to these responses. It is believed that these studies will identify potential drug targets to improve in a clinical setting the outcome of injury-induced corneal wound healing. Active ion transporters and membrane channels underlying receptor activation are components of a myriad of cell signaling pathways that mediate cytokine receptor control of epithelial renewal. One of the most potent and efficacious mitogens hastening corneal epithelial renewal is epidermal growth factor (EGF). This mitogen induces increases in cell proliferation and migration through

175

stimulation of different ion-transport mechanisms, ionic permeabilities, and receptor-linked channels. They include increases in the activity of the basolateral membrane localized Na+:K+:2Cl cotransporter resulting from stimulation of plasma membrane Ca2+ influx and activation of the mitogen-activated protein kinase (MAPK) superfamily. The increase in Ca2+ influx through a Ca2+ channel is dependent on EGF receptor (EGFR)-induced stimulation of membrane-associated phospholipase C (PLC) followed by hydrolysis of a phosphoinositide to induce Ca2+ release from an intracellular Ca2+ store. Another ionic influx affected by EGFR stimulation is K+ efflux through the basolateral membrane. In addition, mitogenic responses to EGF require that during cell-cycle progression cell volume is modulated to accommodate increases in the parent cell of genomic content prior to cell division. Such modulation is, in part, dependent on changes in K+:Cl cotransporter expression and activity. Another example of a cytokine whose induced effects are dependent on modulation of channel activity is the tumor necrosis factor-alpha (TNFa). Cell-Volume Control and Epithelial Renewal Epithelial renewal is a dynamic process that is dependent on tight regulation of cell volume. Prior to cell division, volume expansion of a parent cell is required to accommodate doubling of the nuclear and cytoplasmic components in preparation for their equal distribution into daughter cells. Similarly, changes in cell volume are requisite for cell migration, as this process involves repeated and coordinated leading-edge cytoplasmic volume extension, along with retraction at the opposite pole. As modulation of iontransport activity and membrane permeability underlie changes in cell volume, cytokine-induced control of renewal is, therefore, dependent upon modulation of ionic influx and efflux. Such control requires that there is synchronized release of specific cytokines that alter ion transport at appropriate times during the cell cycle. Numerous cytokines mediate the required regulation for error-free proliferation and migration. Therefore, cytokine receptor control of ion transport activity is of critical importance in the maintenance of corneal epithelial renewal, transparency, and its refractive properties. Ca2+ Channel and Pump Activity Receptor-induced Ca2+ signaling contributes to the regulation of net Cl transport and proliferation of corneal epithelial cells. In order for Ca2+ to link receptor activation to these responses, its intracellular concentration must be regulated at sub-micromolar levels. Such regulation occurs through different types of Ca2+/Mg2+-ATPase plasma membrane and endoplasmic reticulum transporters and channels. The plasma membrane pump counterpart is selectively stimulated by

176

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

calmodulin, whereas one type of channel in this domain is activated as a direct consequence of emptying of intracellular Ca2+ stores. This channel is formed by a tetrameric configuration of transient receptor potential (TRP) protein subunits. Each of the different types of TRP channels identified in the corneal epithelium is activated by different stimuli. Currently, three different types are expressed and studied in corneal epithelial cells. Even though their activation leads in all cases to increases in plasma membrane Ca2+ influx, the types of responses elicited are very different from one another. The four identified types are TRPC4, TRPV1, TRPV4, and TRPM8. In human cornea epithelial cells, TRPC4 is activated during exposure to EGF. TRPV1 is stimulated by an agonist contained in an extract of red peppers (capsaicin) leading to increases in proinflammatory cytokine expression. TRPV4 is activated by a decreased bathing-solution osmolarity and elicits a RVD response. On the other hand, TRPM8 is temperature sensitive. The Ca2+ signaling role attributable to activation of each of these three different TRP isoforms is dependent on the ability of endoplasmic reticulum Ca2+ transporters to rapidly take up into intracellular stores the Ca2+ that has flowed in through these different plasma membrane TRP channels. Otherwise, a sustained rise in intracellular Ca2+ concentration may be cytotoxic. The importance of Ca2+ signaling in mediating receptor control of corneal epithelial renewal and transparency suggests that its modulation by drugs is of potential value in a clinical setting. K+, Cl–, and Na+ Channels Modulation of K+ channel activity is essential for mediating different responses associated with corneal epithelial function. For example, changes on K+ channel activity mediate EGF-induced mitogenic responses, ultraviolet (UV) lightinduced apoptosis, and adrenoceptor-induced increases in net Cl transport. In these cells, there is expression of Ca2+-dependent and inwardly rectifying K+ channels. Even though activation of these different K+ channels types induces intracellular K+ losses, the responses are very different from one another. Additional studies are required to understand how their activation leads to such disparate responses. Chloride channel activity modulation occurs in response to adrenergic receptor-induced increases intracellular Ca2+ concentration and rises in intracellular cyclic adenosine monophosphate (cAMP) levels. Corneal epithelial cells express cystic fibrosis transmembrane regulator (CFTR) channels that are stimulated by rises in cAMP and underlie stimulation of net Cl transport. They also express a Cl channel designated as ClC-3 and their activation is suggested to underlie a regulatory volume response mediating shrinkage during exposure to a hypotonic challenge.

There is some evidence for the expression of tetrodotoxin-blockable Na+ channels in corneal epithelial cells. The functional importance of such activity is unclear since their activation only occurs at membrane voltages far less negative than those described in corneal epithelial cells.

Summary Ion transporter and membrane permeability regulation are essential to the maintenance of corneal epithelial health. Such control is essential for mediating responses that are required for corneal epithelial renewal and the maintenance of corneal transparency. The corneal epithelial cells express both primary and secondary ionic transport mechanisms whose regulation occurs through receptor activation. These ion transporters in concert with changes in membrane permeability underlie: (1) osmotically coupled fluid flow; (2) cell-volume regulation; (3) intracellular pH regulation; and (4) Ca2+ signaling. Such control assures that epithelial renewal maintains barrier function and the finetuning contribution capability of the corneal epithelium to elicit adequate fluid egress from the cornea for sustaining tissue transparency. Additional studies are warranted to identify novel drug targets in the signaling pathways mediating receptor control of ion transporter and channels. This endeavor will possibly identify improved techniques for restoring corneal epithelial function subsequent to injury. See also: Corneal Epithelium: Response to Infection; Dry Eye: An Immune-Based Inflammation; Stem Cells of the Ocular Surface; The Surgical Treatment for Corneal Epithelial Stem Cell Deficiency, Corneal Epithelial Defect, and Peripheral Corneal Ulcer.

Further Reading Al-Nakkash, L., Iserovich, P., Coca-Prados, M., Yang, H., and Reinach, P. S. (2004). Functional and molecular characterization of a volume-activated chloride channel in rabbit corneal epithelial cells. Journal of Membrane Biology 201: 41–49. Al-Nakkash, L. and Reinach, P. S. (2001). Activation of a CFTRmediated chloride current in a rabbit corneal epithelial cell line. Investigative Ophthalmology and Vision Science 42: 2364–2370. Bildin, V. N., Yang, H., Crook, R. B., Fischbarg, J., and Reinach, P. S. (2000). Adaptation by corneal epithelial cells to chronic hypertonic stress depends on upregulation of Na:K:2Cl cotransporter gene and protein expression and ion transport activity. Journal of Membrane Biology 177: 41–50. Candia, O. A. and Alvarez, L. J. (2008). Fluid transport phenomena in ocular epithelia. Progress in Retinal and Eye Research 27: 197–212. Capo´-Aponte, J. E., Wang, Z., Bildin, V. N., Pokorny, K. S., and Reinach, P. S. (2007). Fate of hypertonicity-stressed corneal epithelial cells depends on differential MAPK activation and p38MAPK/Na-K-2Cl cotransporter1 interaction. Experimental Eye Research 84: 361–372.

Corneal Epithelium: Transport and Permeability Lu, L. (2006). Stress-induced corneal epithelial apoptosis mediated by K+ channel activation. Progress in Retinal Eye Research 25: 515–538. Lu, L., Reinach, P. S., and Kao, W. W. (2001). Corneal epithelial wound healing. Experimental Biology and Medicine 226: 653–664. Pan, Z., Yang, H., Mergler, S., et al. (2008). Dependence of regulatory volume decrease on transient receptor potential vanilloid 4 (TRPV4) expression in human corneal epithelial cells. Cell Calcium 44: 374–385. Reinach, P. S., Capo´-Aponte, J. E., Mergler, S., and Pokorny, K. S. (2008). Roles of corneal epithelial ion transport mechanisms in mediating responses to cytokines and osmotic stress. In: Tombran-Tink, J. and Barnstable, C. J. (eds.) Ocular Transporters: In Ophthalmic Diseases and Drug Delivery, Series Ophthalmology Research, pp. 17–46. Totowa, NJ: Humana Press. Reinach, P. S., Holmberg, N., and Chiesa, R. (1991). Identification of calmodulin-sensitive Ca2+-transporting ATPase in the plasma membrane of bovine corneal epithelial cell. Biochimica et Biophysica Acta 1068: 1–8.

177

Wolosin, J. M. and Candia, O. A. (1987). Cl– secretagogues increase basolateral K+ conductance of frog corneal epithelium. American Journal of Physiology 253: C555–C560. Wu, X., Yang, H., Iserovich, P., Fischbarg, J., and Reinach, P. S. (1997). Regulatory volume decrease by SV40-transformed rabbit corneal epithelial cells requires ryanodine-sensitive Ca2+-induced Ca2+ release. Journal of Membrane Biology 158: 127–136. Yang, H., Mergler, S., Sun, X., et al. (2005). TRPC4 knockdown suppresses epidermal growth factor-induced store-operated channel activation and growth in human corneal epithelial cells. Journal of Biological Chemistry 280: 32230–32237. Yang, H., Sun, X., Wang, Z., et al. (2003). EGF stimulates growth by enhancing capacitative calcium entry in corneal epithelial cells. Journal of Membrane Biology 194: 47–58. Zhang, F., Yang, H., Wang, Z., et al. (2007). Transient receptor potential vanilloid 1 activation induces inflammatory cytokine release in corneal epithelium through MAPK signaling. Journal of Cellular Physiology 213: 730–739.

Stem Cells of the Ocular Surface Y Du and J L Funderburgh, University of Pittsburgh, Pittsburgh, PA, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Basement membrane – Every epithelial sheet elaborates a noncellular dual-layered membrane of collagen and proteoglycans to which the cells are anchored. In addition to anchoring the cells, basement membranes (also called basal laminae) serve to maintain epithelial differentiation and to provide a biological barrier preventing contact between the epithelial cells and the underlying mesenchymal tissue. Cell therapy – Stem cells are injected directly into pathological tissues to restore tissue function. Conjunctiva – The squamous epithelium overlaying the sclera and inner eyelid. Corneal stroma – This tough connective tissue provides 90% of the corneal thickness and consists mostly of collagen, proteoglycans, and water. Cells make up 4% of the stromal tissue. Crypts – These structures, originally named in the intestine, represent small dead-end tubes that protrude from the surface of an epithelial layer into the surrounding mesenchymal tissue. Cytokeratin – Hair and the epidermis consist primarily of the fibrous proteins, keratins. Moist epithelia such as intestine and cornea were originally thought not to be ‘keratinized’ but the outer layers cells in these tissues express intracellular keratin proteins as they differentiate. Due to the large number of keratin genes, keratin expression is highly tissue specific. Endothelium – Corneal endothelium is a single epithelial layer of cells on the posterior (inner) side of the cornea which provides a hydrodynamic pumping function to maintain corneal hydration. It is unrelated to vascular endothelium. Keratocyte – These are dendritic, quiescent cells of the corneal stroma. They produce the complex extracellular matrix of the corneal stroma but, during wound healing, also produce an opaque scar tissue. Limbus – In the eye, the border between the transparent cornea and the opaque sclera is known as the limbus. The limbal region of the overlying epithelia is continuous with no obvious boundary, but the conjunctival and corneal cells can be distinguished by the types of cytokeratins that are expressed. Neural crest – In early embryonic development a small population of cells between the ectoderm and

178

the neural tube migrates out into the embryo, differentiating into a wide variety of tissues including autonomic nerves, cartilage, smooth muscles, and melanocytes. A number of ocular tissues (corneal stroma and iris) are derived from neural crest. Niche – A stem cell niche is a phrase loosely used to describe the microenvironment in which stem cells are found, which interacts with stem cells to regulate stem cell fate. Squamous epithelium – Multilayered epithelia show differences between the basal cells attached to the basal lamina and the more superficial cells of the layer. The apical cells flatten and express tissuespecific keratins. These outer cells die by apoptosis and leave the layer (known as desquamation) to be replaced by underlying cells. Single-layer epithelia such as the corneal endothelium do not undergo the process of desquamation and renewal.

The Ocular Surface Anatomy Beneath the protective tear film, the human ocular surface is covered with a contiguous layer of moist squamous epithelium, consisting of the bulbar conjunctival epithelium over the sclera and the transparent corneal epithelium of the cornea (Figure 1). The bulbar conjunctival epithelium rests on a loose, vascularized connective tissue which allows the movement of the eyelid over the sclera and maintains the limbal vascular supply. This epithelium is contiguous with the corneal epithelium and with the conjunctiva of the fornix located in the folded region between the sclera and the eyelid. The conjunctival epithelium contains goblet cells – the source of soluble mucins in the tear film. These cells are an essential element in maintenance of the integrity of the ocular surface. The cornea is covered with a phenotypically unique epithelial sheet 5–11-cell layers deep which is transparent to light and devoid of goblet cells. The flattened cells of the corneal surface overlie wing-shaped cells – which, in turn, rest on a cuboidal basal epithelium. Unlike the conjunctival epithelium, the corneal epithelium is firmly connected to rigid tissue of the stroma via an anchoring complex involving keratin and collagen fibrils extending into the acellular anterior collagenous layer of the stroma. The interface between the corneal and conjunctival epithelium is known as the

Stem Cells of the Ocular Surface

corneoscleral limbus (Figure 1(a)). This region features a network of physical folds in the surface of stromal tissue known as the palisades of Vogt (Figure 1). As described below, the palisades harbor a small population of cells identified as limbal stem cells (LSC). The limbal region also contains pigmented cells and a population of immune cells related to the Langerhans’ cells of the dermis. Underlying the corneal epithelium is the corneal stroma, a tough, transparent connective tissue making up 90% of the thickness of the cornea (Figure 1(c)). The stroma contains layers (lamellae) of aligned collagen fibrils sandwiching quiescent keratocytes, flattened mesenchymal cells. The specialized arrangement of the collagenous ultrastructure in the stroma is responsible for the remarkable tensile strength of this tissue as well as its unique light transparency. On the posterior side of the stroma a single cell layer – the corneal endothelium – is separated from the stromal tissue by Descemet’s membrane. The three tissue layers of the cornea each make an important physiological contribution to maintenance of corneal transparency. In addition, the cornea provides 75% of the refractive power required to focus the light on the retina and acts as an effective biological barrier. Development Formation of the human cornea begins at approximately 5–6 weeks of gestation. After the lens vesicle pinches off from the surface ectoderm of the head, the overlying

ectoderm transforms into a layer of cuboidal epithelial cells, which continue to develop into the corneal epithelium. At 6–7 weeks, neural crest cells migrate between this epithelium and the lens, forming the corneal endothelium. Shortly afterward, a second wave of cell migration from neural crest forms the stroma – which then begins to generate collagenous matrix in the 8th week. Homeostasis and Repair The cellular layers of the cornea and conjunctiva differ markedly in several important characteristics of homeostasis and repair. In the corneal epithelium, the superficial cells are lost by exfoliation and replaced by the underlying wing cells. These, in turn, are continually replaced by mitotically active basal cells. In addition to the basal to superficial migration of corneal epithelial cells, there is a well-documented migration of epithelial cells from the periphery to the center of the cornea, measured at 2–3 mm per day. The dynamics of the corneal epithelium have been characterized as the X,Y,Z equation, in which the exfoliation rate (X ) equals the rate of production of new cells from the basal cells (Z) plus cells added by the centripetal migration of cells (Y ). When injured, the corneal epithelium closes wounds first by migration of the intact sheet over the denuded area, followed by a delay before increased cell division is observed. Bulbar conjunctival epithelium is contiguous with the corneal epithelium without any obvious physical barrier; however,

Limbus

Cornea

Cornea

Limbus

Palisades

179

Iris Palisades Limbal crypts

Bulbar conjunctiva

Bulbar conjunctiva

(b)

(a) Limbus Palisades Limbal crypts Conjunctiva

Corneal epithelium

Limbus

Keratocytes

Cornea Terminally differentiated cells

Stroma Conjunctiva

Corneal endothelium

Bowman’s layer Postmitotic cells Late transient amplifying cells

Langerhan’s cells

Iris

Early transient amplifying cells Stem cells

Lens

(c)

(d)

Mesenchymal cells

Epithelial basement membrane

Figure 1 The ocular surface. (a) The location of the corneal, conjunctival, and limbal surfaces are displayed along with the location of the palisades of Vogt. (b) Artists rendering illustrating a close-up view of the anatomy of the palisades of Vogt and the dead-end crypts that branch off from these folds in the tissue. (c) Cross-sectional illustration of the corneoscleral junction. (d) Close-up illustration of the limbal epithelium region and limbal stem cell niche. Original drawings by Kira Lathrop.

180

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

these cells express different cytokeratin markers and are found not to migrate in the unwounded eye. Cell division by basal cells and exfoliation is observed throughout the conjunctiva. The cells of the corneal stroma – the stromal keratocytes – show little cell division in normal adults. The keratocytes undergo rapid cell division after localization in the cornea in late embryogenesis, but, following birth, the keratocyte cell number stabilizes and little or no mitosis can be detected throughout life. In the case of inflammation or wounding, however, the stromal keratocytes become activated and mitotic. The activated keratocytes adopt a fibroblastic phenotype and, later, display characteristics of myofibroblasts. Connective tissue secreted by these cells during wound healing is not transparent but becomes opaque scar tissue. Following healing, the cells become quiescent, but human corneal scars are very slow to resolve; it is not clear if the resident cells ever return to a fully keratocytic phenotype. These properties suggest a limited process of tissue renewal in the corneal stroma. Renewal of corneal endothelial cells is even more limited than that of the keratocytes. Following childhood, human corneal endothelial cells do not divide. Compensation for endothelial damage is accomplished by flattening of the remaining cells to cover the posterior surface of the cornea. In vitro as well, human corneal endothelial cells show only limited ability to divide following infancy. These characteristics have led to a conventional view that, while the corneal epithelium is maintained by a stem cell population, the stroma and the endothelium – with limited ability for self-renewal – are not maintained by mitotically active, tissue-resident stem cells.

Properties of Stem Cells Stem cells, by definition, undergo asymmetric cell division; that is, they undergo self-renewal while giving rise to differentiated daughter cells. Embryonic stem cells derived from the inner cell mass of the blastocyst are pluripotent, giving rise to most cells of the body. In culture, embryonic stem cells can be propagated indefinitely in an undifferentiated state. Although they can be isolated from umbilical cord blood, fully pluripotent stem cells are thought to be scarce or absent in adult tissues. Stem cells in adult organisms have long been associated with self-renewing tissues, such as the hematopoietic system, dermis, and intestine. The resident stem cells in these tissues are generally capable of generating only one type of cell, making them unipotent. In recent years, understanding of a new class of stem cells has emerged: known as mesenchymal stem cells (MSC), these cells appear to be present in small numbers in many somatic tissues and can also be isolated from bone marrow. Studies

show these MSC to participate in injury repair, and – when expanded in culture – the MSC exhibit potential to differentiate into a number of lineages, and thus can be considered as being multipotent. Considerable effort has gone into identifying common characteristics of adult stem cells, but in spite of such efforts there appears to a dearth of global phenotypic markers for such cells. In vitro, clonal growth, extended life span, and the ability to express phenotypic markers of multiple differentiated cell types have been useful in stem cell identification, but cell surface markers for stem cells – as well as properties enabling isolation of these populations – often need to be defined for individual tissues. In self-renewing tissues, stem cells appear to be localized in an anatomical niche, a restricted microenvironment which interacts with stem cells to regulate stem cell fate. The niche provides physical and chemical factors that both control the replication and maintain the differentiation potential of the stem cells. Typically, the niche is near a vascular bed, usually containing a population of unrelated cells that serve as feeder cells.

Stem Cells in the Corneal Epithelium Characteristics of Limbal Stem Cells Abundant research supports the idea that a stem cell population is localized in the corneoscleral limbal region. These cells – termed limbal stem cells (LSC) – share a number of features with the stem cells of other selfrenewing tissues. They have small cell size and high nuclear-to-cytoplasmic ratio. They lack expression of differentiation markers expressed by corneal epithelial cells, specifically cytokeratins (CK) 3 and 12. Like stem cells in other self-renewing tissues, the LSC divide very infrequently (slow cycling), and, therefore, DNA-labeling agents – such as bromodeoxyuridine – are retained in these cells for months. Label retention has often been used in identifying stem cell candidates in intact tissues. Cells from the limbal region also grow clonally in large colonies known as holoclones, whereas clones from the central cornea are less abundant and grow through fewer population doublings. Furthermore, unlike central corneal cells, LSC proliferation is resistant to inhibition by phorbol esters. LSC express several genes linked to stem cell self-renewal, including CEBPD, Bmi1, and Notch1. Cornea-specific ablation of Notch1 resulted in differentiation of LSC into hyperplastic, keratinized, epidermal-like cells, indicating the probability that Notch1 expression is essential in maintenance of the LSC stem potential. LSC also express a transporter protein, ABCG2, responsible for efflux of fluorescent dye Hoechst 33342. Expression of ABCG2 enables isolation of cells using fluorescence-activated cell sorting (FACS). Cells isolated using this dye-efflux assay are known as a side population,

Stem Cells of the Ocular Surface

based on the distribution pattern of the fluorescent cells following FACS. Side-population cells from the corneal epithelium are present exclusively in the limbus; following sorting, they exhibit slow cycling and clonal growth properties consistent with their identification as stem cells. Ability to isolate LSC has allowed a detailed comparison of the differences between this population and the other cells of the corneal epithelium. A summary of these differences is given in Table 1. Limbal Stem Cell Niche Centripetal migration of pigmented cells in healing epithelial wounds was first observed in the 1940s, and, in 1971, Davanger and Evenson proposed that the source of the migrating cells was the palisades of Vogt. These are a series of radially oriented fibrovascular ridges concentrated along the upper and lower corneoscleral limbus (Figure 1(a) and (b)) described originally in the early twentieth century. The morphology of these features has been elucidated by Daniels and by Dua, showing that they are present throughout the limbus but more concentrated in the superior and inferior regions. The ridges and valleys of the palisades are maintained by a specialized vascular system and form an interwoven network, often terminating in tunnels or ‘crypts’ under the surface of the stroma (Figure 1(b)). The bases of these crypts contain cells strongly expressing the several marker genes used in identifying LSC (Table 1), providing further evidence that these crypts represent the LSC niche. Recently, it was observed that pigmented melanocytes in the limbal crypts express N-cadherin, as do the LSC. During expansion in Table 1 Expression markers for determining epithelial stem cell differentiation Gene

Limbal stem

Cornea: Basal

Positive (stem cell) markers DN-p63a þþ ABCG2 þþþ Vimentin þþþ Integrin a9 þþþ þ Keratin 19 þþ N-Cadherin þþ GDNF þþþ Integrin ß1 a-enolase þþþ þþ TrkA þþþ þþ Notch1 þþþ þþ C/EBPdelta þ Bmi1 þ Negative (differentiation) markers CK3/CK12 þþþ Involucrin þþþ Integrin a6 þþþ Cx 43 þþþ E-Cadherin þ/ þþþ Modified from Secker and Daniels (2008).

Cornea: Apical þ þ þþþ þþþ þþþ þþþ þþþ

181

culture the LSC lose N-cadherin expression as they differentiate. N-cadherin-dependent cell–cell interactions with melanocytes may consequently represent a feature of the niche environment involved in maintaining the slow cycling and stem cell potential of the LSC. The presence of label-retaining cells in the limbus and the centripetal migration of the more rapidly dividing cells in the corneal epithelium prompted Schermer and colleagues to suggest that the slow-cycling LSC are precursors to the mitotic basal cells throughout the epithelium. These mitotic daughter cells are termed transient amplifying (TA) cells. The TA cells are hypothesized to migrate from the limbus throughout the central cornea before ceasing cell division and differentiating to the superficial cells of the corneal surface. Labeling with both tritiated thymidine and bromodeoxyuridine followed by various chase periods provided kinetics showing that migrating basal cells are indeed mitotic, but are not label retaining – that is, cycle faster than the LSC. The LSC appear small and round compared to the TA cells, thus may be more primitive than the TA cells. Cloning studies also show the mitotic cells in the central stroma to be less able than LSC to generate the large holoclones indicative of high replicative potential. TA cells are not fully committed to corneal epithelial differentiation. TA cells in adult central corneal epithelium – when transplanted to embryonic dermis – have demonstrated potential to differentiate into hair follicles and other dermal tissue types. Thus, these basal cells of the central epithelium maintain a clear stem cell potential. Are these central corneal stem cells progeny of the LSC? DNA-labeling kinetics are consistent with an interpretation that slow-cycling cells in the limbus are the origin of mitotic cells of the central cornea. Lamellar keratoplasty in male rabbits with corneal tissue from females led to the gradual replacement of the sex chromatin in the central corneal graft, suggesting that LSC from the host were the source of the cells that mature and exfoliate from the superficial central cornea. In spite of these and a number of related studies, however, there is still only indirect evidence that LSC are the progenitors of the mitotic basal TA cells in undisturbed corneal epithelium. In fact (as discussed below), recent studies suggest the possibility that the LSC may contribute to the central corneal epithelial cell populations only in healing wounds. Limbal Stem Cell Deficiency It is well documented that, following wounding, the LSC become more mitotically active and participate in centripetal migration. Cells from the limbus can, in fact, rapidly repair wounds involving the entire central corneal epithelium. However, injuries or conditions resulting in loss or inactivation of the LSC present a markedly

182

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

different outcome. After loss of the limbal cells, cells from the conjunctiva – including goblet cells – migrate across the corneal surface. This conjunctivalization results in inflammation, neovascularization, and growth of a fibrovascular pannus which severely reduces visual acuity. This phenomenon, known as limbal stem cell deficiency, can result from either hereditary or acquired causes. Transplantation of central corneal tissue without the accompanying LSC provides only temporary amelioration of this condition. Successful treatment, however, can be affected by autologous transplantation of limbal tissue from a contralateral, unaffected eye to the conjunctivalized eye. Reestablishing the LSC population can restore a stable and transparent corneal epithelium both in experimental animal studies and in human patients. Allografts can serve the same purpose but require immune suppression to prevent rejections. Expansion of LSC in culture has also been used to prepare epithelial sheets that restore long-term human epithelial function as well. It is worth noting that autologous, cultured epithelial sheets from buccal epithelial cells – as well as from conjunctival cells cultured under conditions in which goblet cells do not form – can also restore epithelial function. The biological implications of this use of nonlimbal cells to restore corneal epithelial function have not been fully explored. It is not clear, for example, if stem cells in these cultured layers colonize the limbal crypts and become a kind of ectopic, but functional, LSC population. The conclusions of the studies on limbal stem cell deficiency and its therapy are that the LSC are effective in replacing the corneal epithelial layer in a wound healing situation, and that the LSC create a biological barrier preventing conjunctival migration onto the corneal surface. Understanding the Role for LSC in Corneal Homeostasis Several studies have suggested the possibility that during normal corneal homeostasis the LSC may not contribute cells to the corneal epithelium. Studies of the destrinknockout mouse showed that these mice lack the centripetal migration of corneal epithelium. Destrin is an actin-binding protein involved in cytoskeletal dynamics and thus may be involved directly in such motility. The corneal epithelium forms correctly in these mice, but, later in life, the tissue becomes vascularized and hyperplastic. The presence of mitotic basal cells in the absence of migration suggests that the basal cells are not TA progeny of the LSC but arise from a distributed stem cell population in these mice. Thus, the LSC may not be necessary for establishing a fully differentiated corneal epithelium. In a later study using mice, limbal tissue expressing a b-galactosidase marker gene was transplanted to the limbus of non-b-Galexpressing nude mice. Over a period of months, b-Gal was not found in the central cornea of unwounded mice, but

b-Gal cells rapidly migrated into the central cornea after wounding. b-Gal-expressing cells from central cornea were able to reconstitute a full epithelial sheet following transplantation, and furthermore, the reconstitution after transplantation could be carried out serially. These results indicate the existence of a nonlimbal population of corneal epithelial stem cells in mice that can give rise to a functional corneal epithelium. Such studies indicate LSC may not be necessary for formation or normal homeostasis of mouse cornea but support the principle that the LSC does participate in restoration of the tissue after wounding. There are no experimental data extending these results to humans, but until such data are presented, the role of the LSC in normal tissue homeostasis of the central corneal epithelium should be considered as unresolved.

Conjunctival Stem Cells The bulbar conjunctiva contains basal mitotic cells that migrate to the surface and become quiescent as they differentiate; however, the cell layer does not undergo a lateral migration similar to that of the cornea. Differentiated conjunctival epithelium also differs from corneal epithelium in the expression of cytokeratins. Whereas cornea is positive for CK3 and CK12, conjunctiva lacks these keratins while expressing CK19, an antigen absent in cornea. Label-retaining cells can be observed in the fornix region, and clonal culture of these gives rise to both epithelial cells and goblet cells. Thus, the fornix appears to contain bipotent stem cells. Similarly, cells in the fornix are more strongly stimulated to proliferate by phorbol ester than those in the bulbar epithelium. These studies support the idea of a stem cell population in a fornicial niche. Clonogenicity studies, however, support the idea that cells with extended replicative potential (e.g., stem cells) are present both in the fornix and the bulbar surface. The lack of lateral migration of the cells of the bulbar conjunctiva, however, strongly supports the idea that during normal tissue homeostasis the tissue is maintained by a distributed population of resident stem cells. Careful examination of stem-like cells in the bulbar conjunctiva has identified rare clusters of cells expressing the CK3/12 keratins characteristic of the cornea. Furthermore, transplantation of limbal b-galactosidase-expressing tissue into the limbal region of nude mice showed that, following wounding, some of these cells migrate into the conjunctiva and differentiate to both conjunctival epithelial and goblet cell phenotypes. These studies suggest that some of the stem cells distributed in the bulbar conjunctiva may derive from the LSC. Conversely, cells cultured from the bulbar region can be used to reconstitute the corneal epithelium in transplantation studies. Following transplantation to the cornea, these conjunctival-derived cells express the cornea-specific keratins CK3/12. The idea of

Stem Cells of the Ocular Surface

conjunctival/corneal transdifferentiation, once popular, is currently out of favor; however, these experiments support the possibility that stem cells in both the conjunctiva and cornea can – and do, in fact – produce differentiated cells in the neighboring tissue. The extent and the conditions under which this phenomenon occurs require further elucidation.

Corneal Stromal Stem Cells The corneal stroma is a mesenchymal connective tissue making up 90% of the corneal thickness, with physical properties that provide the cornea its essential character. The stroma is formed during late embryogenesis by a population of neural crest cells migrating from the periocular mesenchyme. Chicken keratocytes from late embryogenesis retain neural crest progenitor properties even after transplantation into a new environment along cranial neural crest migratory passageways. In adult mammals, however, numerous in vitro experiments show that keratocytes rapidly lose their characteristic phenotype following several population doublings. Such a loss of phenotype occurs in healing wounds in vivo as well as in vitro. Recently, the authors found that the stroma of bovine corneas contain a small population of cells exhibiting self-renewal ability for an extended number of population doublings in culture. These corneal stromal stem-like cells were clonogenic and proliferated in vitro for over 100 doublings. A similar population of stem cells was isolated from human corneas as a side population using FACS. These stromal stem cells demonstrated potential for differentiation into several noncorneal cell types – a characteristic similar to that found in adult stem cells from other mesenchymal tissues. These cells also expressed several genes clearly designating them as tissue-specific mesenchymal stem cells, including ABCG2, Notch1, BMI1, SIX2, and KIT. ABCG2-expressing cells were localized in the limbal stroma subjacent to the epithelial basement membrane in the regions near the LSC niche. Human corneal stromal cells remained viable for months after injection into mouse corneal stroma and were able to increase the transparency of lumican-knockout mouse corneas. In serum-free monolayer culture, keratocytes express stroma-specific gene products, but do not accumulate or organize extracellular matrix resembling that of the corneal stroma. However, when human stromal stem cells were cultured in attachment-free conditions they produced a stromal-like tissue including the stroma-specific keratan sulfate proteoglycan keratocan and generated parallel layers of collagen fibers resembling those in the stroma. These results suggest the potential use of stromal stem cells in bioengineering of corneal stroma or in direct stem cellbased therapy for corneal scars or corneal dystrophies.

183

Corneal Endothelial Stem Cells The corneal endothelium is a single layer of flat hexagonal cells forming a boundary between the corneal stroma and the anterior chamber. This layer of cells functions as a pump to regulate stromal hydration. Although – like the keratocytes – the corneal endothelium is derived from neural crest, the endothelial cell characteristics are different from those of keratocytes. The human corneal endothelial cells are arrested in G1 phase in vivo and do not normally replicate to replace dead or injured cells. This lack of cell division results in a physiological reduction of cell density of about 0.3–0.5% per year. Scattered evidence, however, suggests the potential for some mitotic events in human corneal endothelium. Mitotic figures were observed in vivo by specular microscopy study following a rejection reaction on a corneal graft. Clusters of cells smaller than surrounding cells suggested that, at least under some circumstances, mitosis occurs in the endothelium of the adult human cornea. Recently, Yokoo and coworkers identified cells in the human corneal endothelium able to form cell-spheres in attachmentindependent culture. Cells in these spheres, formed under conditions similar to those used for isolation of neural stem cells, can be expanded and generate daughter cells expressing neuronal and mesenchymal molecular markers. These properties suggest a stem cell origin for the cells forming the spheres. The sphere-forming cells also adapted the polygonal morphology characteristic of endothelial cells, suggesting the presence of endothelial progenitors. These precursors were effective in vivo in restoring endothelial function in an animal model of corneal endothelial deficiency. Both peripheral and central rabbit corneal endothelia contain a significant number of precursors, but the peripheral endothelium contains more precursors and has a stronger self-renewal capacity than the central region by sphereforming assay. As the long-term culture of human endothelial cells has not been carried out and because of a lack of both stem cell and endothelial markers, positive identification of the proposed endothelial stem cells in situ has yet to be accomplished.

Conclusions The ocular surface contains multiple populations of cells with stem cell-like properties. Some of these are localized to a distinctive niche in the corneoscleral limbus but other populations are more dispersed. Stem cells also are present in corneal stroma and endothelium. Stem cells support selfrenewal of the corneal and conjunctival epithelia and the limbal stem cells prevent conjunctivalization of the corneal surface; however, relationships among the different stem cell populations in normal tissue homeostasis – and in

184

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

response to wounding – are not yet fully characterized. Ocular-surface stem cells have a high potential for use in cell-based therapy for a variety of ocular pathologies, and in tissue engineering. Continuation of the characterization of ocular stem cells is, therefore, a high priority both for understanding the biology of the ocular surface and for development of sight-saving therapies.

Acknowledgments The authors wish to thank Kira Lathrop for the excellent illustration and Martha Funderburgh for proofreading the manuscript. This work was supported by NIH Grant EY016415 and Research to Prevent Blindness Inc. See also: Corneal Epithelium: Wound Healing Junctions, Attachment to Stroma Receptors, Matrix Metalloproteinases, Intracellular Communications; The Surgical Treatment for Corneal Epithelial Stem Cell Deficiency, Corneal Epithelial Defect, and Peripheral Corneal Ulcer.

Further Reading Daniels, J. T., Dart, J. K., Tuft, S. J., and Khaw, P. T. (2001). Corneal stem cells in review. Wound Repair Regen 9(6): 483–494.

Daniels, J. T., Notara, M., Shortt, A. J., et al. (2007). Limbal epithelial stem cell therapy. Expert Opinion on Biological Therapy 7(1): 1–3. Davanger, M. and Evensen, A. (1971). Role of the pericorneal papillary structure in renewal of corneal epithelium. Nature 229(5286): 560–561. Du, Y., Funderburgh, M. L., Mann, M. M., SundarRaj, N., and Funderburgh, J. L. (2005). Multipotent stem cells in human corneal stroma. Stem Cells (Dayton, Ohio) 23(9): 1266–1275. Lavker, R. M. and Sun, T. T. (2003). Epithelial stem cells: The eye provides a vision. Eye (London, England) 17(8): 937–942. Lavker, R. M., Tseng, S. C., and Sun, T. T. (2004). Corneal epithelial stem cells at the limbus: Looking at some old problems from a new angle. Experimental Eye Research 78(3): 433–446. Limb, G. A., Daniels, J. T., Cambrey, A. D., et al. (2006). Current prospects for adult stem cell-based therapies in ocular repair and regeneration. Current Eye Research 31(5): 381–390. Majo, F., Rochat, A., Nicolas, M., Jaoude, G. A., and Barrandon, Y. (2008). Oligopotent stem cells are distributed throughout the mammalian ocular surface. Nature 456(7219): 250–254. Mimura, T., Yokoo, S., Araie, M., Amano, S., and Yamagami, S. (2005). Treatment of rabbit bullous keratopathy with precursors derived from cultured human corneal endothelium. Investigative Ophthalmology and Visual Science 46(10): 3637–3644. Secker, G. A. and Daniels, J. T. (2008). Corneal epithelial stem cells: Deficiency and regulation. Stem Cell Reviews and Reports 4(3): 159–168. Shortt, A. J., Secker, G. A., Notara, M. D., et al. (2007). Transplantation of ex vivo cultured limbal epithelial stem cells: A review of techniques and clinical results. Survey of Ophthalmology 52(5): 483–502. Thoft, R. A. and Friend, J. (1983). The X, Y, Z hypothesis of corneal epithelial maintenance. Investigative Ophthalmology and Visual Science 24(10): 1442–1443.

The Surgical Treatment for Corneal Epithelial Stem Cell Deficiency, Corneal Epithelial Defect, and Peripheral Corneal Ulcer N Koizumi and S Kinoshita, Kyoto Prefectural University of Medicine, Kyoto, Japan ã 2010 Elsevier Ltd. All rights reserved.

Glossary Allogeneic – Taken from different individuals of the same species. Aniridic keratopathy – The corneal opacification in congenital absence of the iris. Atopic or vernal keratoconjunctivitis – The ocular conditions resulting from allergies. Type I and IV hypersensitivities have been demonstrated to play a role in the allergic response. Both disease processes manifest with classical symptoms of ocular allergy. Cicatrization – The formation of scar tissue. Corneal lenticules – Corneal disks, usually obtained from a donor. Keratoepithelioplasty (KEP) – The transplantation of peripheral corneal lenticules harvested from donor tissue for the treatment of severe ocular surface diseases. Keratolimbal allograft (KLAL) – A surgical procedure in which limbal tissue with peripheral cornea is obtained from donor eyes and transplanted to the recipient eyes. KLAL is performed to treat severe bilateral ocular surface disorders combined with limbal stem cell deficiencies. Lamellar keratoplasty – An operation in which diseased corneal tissue is removed and replaced by lamellar corneal tissue from a donor. The procedure is performed either to improve vision (optical keratoplasty) or to provide structural support for the cornea (tectonic keratoplasty). Limbal transplantation – The transplantation of limbal tissue including stem cells. Autografts and allografts of limbal transplantations were developed to improve the outcome of ocular surface reconstruction. Mooren’s ulcer – A rapidly progressive, painful, ulcerative keratitis, which initially affects the peripheral cornea and may spread circumferentially and then centrally. Ocular cicatricial pemphigoid (OPC) – A chronic disease that produces adhesions and progressive cicatrization and shrinkage of the conjunctival, oral, and vaginal mucous membranes.

Penetrating keratoplasty – The corneal transplant involving the replacement of all layers of the cornea, yet retaining the peripheral cornea. Stevens–Johnson syndrome – A condition affecting the skin in which cell death causes the epidermis to separate from the dermis. The syndrome is thought to be a hypersensitivity complex affecting the skin and the mucous membranes. Superficial keratectomy – The removal of corneal epithelium and anterior stroma. Symblepharon – Adhesion of the eyeball to one or both eyelids. The ex vivo expansion of corneal epithelial cells/ oral mucosal epithelial cells – A form of ocular surface reconstruction using cultivated corneal epithelial/oral mucosal epithelial cell sheets that are developed using tissue engineering techniques. Several types of cultivated epithelial sheets, with or without carrier materials, are used for the treatment of severe ocular surface diseases, such as Stevens–Johnson syndrome, ocular cicatricial pemphigoid, and severe chemical burns.

Introduction The concept of an ocular surface has been widely accepted in the field of ophthalmology and investigations in this area have greatly improved our understanding of the important role that the ocular surface plays in the maintenance of vision and ocular health. The healthy ocular surface is composed of corneal and conjunctival epithelia, each of which has a distinct cellular phenotype. These two types of epithelia, with the presence of an intact tear film, maintain the ocular surface integrity. The corneal epithelium, especially, plays a critical role in maintaining corneal transparency and avascularity. On the basis of numerous investigations, it is now believed that corneal epithelial stem cells exist in the basal layer of the limbal regions where palisades of Vogt are seen in normal human subjects. Severe damage to the limbal

185

186

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

region, for example, damage caused by limbal stem cell deficiencies, results in serious corneal surface problems such as persistent epithelial defects, conjunctivalization with superficial vascularization, keratinization, scarring, etc., with an associated severe loss of vision. In order to rescue such damaged ocular surfaces, surgical modalities have been developed over the past 20 years that are aimed at reconstructing the diseased ocular surface epithelium. In this article, we explain the surgical treatment for total stem cell deficiency.

Corneal Epithelial Transplantation for Total Stem Cell Deficiency In Vivo Expansion of Corneal Epithelial Cells (Keratolimbal Allografts)

stem cell deficiencies. This procedure is also applicable to patients with unilateral disease who do not wish to use limbal tissue obtained from their healthy eye. KLAL is a surgery most suitable for diseases with total stem cell deficiency with less inflammation and less conjunctival cicatrization, such as ocular surface tumors (conjunctival intraepithelial neoplasia or squamous cell carcinoma) or aniridic keratopathy. For cases involving stem cell deficiency with severe inflammation, such as those resulting from chemical injury, Stevens–Johnson syndrome (SJS), or ocular cicatricial pemphigoid (OCP), a KLAL is applicable if the inflammation can be well controlled prior to surgery by steroid and immunosuppressive treatment and if conjunctival involvement is not severe. When the conjunctival scarring is severe, amniotic membrane (AM) transplantation combined with a KLAL is performed to reconstruct the conjunctival fornix (Figure 1).

History and concept of ocular surface reconstruction

The concept of ocular surface reconstruction was first introduced via an autologous conjunctival transplantation for unilateral chemical injury reported in 1977 by Thoft. Thereafter, Thoft described a new technique known as keratoepithelioplasty (KEP), which involves the transplantation of peripheral corneal lenticules harvested from donor tissue for the treatment of severe ocular surface diseases. Following this development, autografts or allografts of limbal transplantations were developed to improve the outcome of ocular surface reconstruction. These surgical procedures, which involve the utilization of donor limbal stem cells in conjunction with the peripheral corneal lenticules, are now classified as keratolimbal allografts (KLALs), a kind of in vivo expansion of limbal stem cells. Indications

KLAL is a procedure in which limbal tissue with peripheral cornea is obtained from donor eyes and transplanted to the recipient eyes. KLAL is performed to treat severe bilateral ocular surface disorders combined with limbal

(a)

Surgical procedure For KLAL, fresh donor corneoscleral tissue preserved at 4 C is used. Ideally, the donor tissue should be as fresh as possible when used, with surgery being performed within 6 days after preparation. After the central cornea of the donor tissue is excised with a 7.0–7.5-mm trephine, the peripheral cornea with scleral rim is sectioned into four to five pieces of lamellar grafts (lenticules). The residual corneoscleral rim after conventional penetrating keratoplasty is also useful. Under a surgical microscope, the excess peripheral scleral tissue of each lenticule is removed by scissors. Then, the posterior two-thirds of the corneal stroma that is attached by Descemet’s membrane and the corneal endothelium are removed by lamellar dissection using spring scissors. After trimming the edge of each lenticule by spring scissors, the lenticules are placed onto the limbal area of the patient’s eye and secured with two to three interrupted sutures per lenticule using 10-0 nylon. Immediately after surgery, a therapeutic soft contact lens is placed on the ocular surface to prevent donor epithelial damage and promote smooth corneal epithelial healing.

(b)

Figure 1 Keratoepithelioplasty combined with lamellar keratoplasty and amniotic membrane transplantation for chemical injury. (a) Before surgery, the patient’s cornea was covered with conjunctival tissue. (b) One year after surgery, the corneal surface is covered with clear corneal epithelium and the patient recovered good vision.

The Surgical Treatment for Corneal Epithelial Stem Cell

Postoperative management

Compared to conventional penetrating keratoplasty, limbal allografts are at significantly higher risk for immunological rejection. Postoperatively, 0.1% dexamethasone and 0.05% cyclosporine A should be instilled topically, and dry-eye patients should receive preservative-free artificial tears. Systemic corticosteroids (betamethasone, 1 mg per day), cyclosporine A (3 mg per kg of body weight per day), and cyclophosphamide (1 mg per day), as well as mycophenolate mofetil (1 g per day), in some cases, should be used to prevent postoperative inflammation and immunological rejection. Systemic immunosuppression as described above should be used for at least 6 months postoperatively, after which it can be gradually reduced depending on clinical characteristics (Table 1). In many cases, it is necessary to administer a low dose of cyclosporine A (1 mg per kg of body weight per day) for up to 2–3 years. A therapeutic soft contact lens should be used for several years after surgery, changing the lens once every 2–4 weeks, as it has been shown that continuous coverage of the corneal surface including the limbal area by use of a soft contact lens is effective for preventing immunological rejection as well as mechanical damage of the corneal epithelium. Although the working mechanism is unknown, it is speculated that the continuous soft contact lens wear may prevent the exposure of the donor limbal tissue to host immunocompetent lymphocytes in tears. Other surgical procedures

Conjunctival limbal autograft (CLAU) is a procedure used for a unilateral stem cell deficiency, in which limbal tissue attached to a conjunctival carrier is transplanted from the healthy eye of the patient. The biggest advantage of this procedure is that no immunosuppression is required for an autograft. Living-related conjunctival limbal allograft (LR-CLAL) is a similar surgery to CLAU, yet in this procedure, a living relative of the patient is the source of the limbal Table 1 Immunosuppressive treatment after all-corneal epithelial transplantation Agent Topical corticosteroids Cyclosporine A Systemic corticosteroids Cyclosporine A Mycophenolate Cyclophosphamide (for Stevens–Johnson syndrome)

Dose and duration 4 times per day, several years 4 times per day, several years 2–4 mg per day, 2–4 weeks 100–200 mg per day, 6 months 1 g per day, 6 months 50–100 mg per day, 3 months

187

tissue used for transplantation. LR-CLAL is applicable for bilateral stem cell deficiency. A major concern associated with both CLAU and LR-CLAL is the risk of stem cell deficiency of the donor eye. Prior to the surgery, it is vital to exclude any possibility of limbal stem cell damage in the donor eye, and it is important to continue observing the condition of the donor eye through postsurgical follow-up. Compared to the KLAL, the amount of limbal tissue that can be taken from the donor eye is limited in CLAU and LRCLAL. Thus, partial stem cell deficiency is the appropriate condition for the use of those two procedures. Ex Vivo Expansion of Corneal Epithelial Cells History and concepts There is no doubt that corneal epithelial transplantations, including limbal autografts and KLALs, have helped to improve the outcome of ocular surface reconstruction in a number of situations. However, in severe ocular surface diseases, such as SJS or OCP, severe inflammation interferes with in vivo epithelial healing and results in a persistent epithelial defect. An alternative concept to the in vivo expansion of corneal epithelial cells is the ex vivo expansion of corneal epithelium using a tissue engineering technique. For this purpose, many investigators endeavored to reconstruct corneal epithelial sheets on carrier materials, such as collagen sheets, and corneal stromal carriers to create stratified corneal epithelial cell layers. Some groups tried to reconstruct not only the epithelial sheet, but also three layers of corneal tissue – a corneal equivalent – using cell-line cells supported by natural and synthetic polymers. This kind of corneal equivalent is now ready to be used for testing toxicity and drug efficacy, but it is not ready for clinical application. Despite the potential drawbacks of cultivated corneal epithelial transplantation, its first clinical application was demonstrated in 1997. A method was developed to reconstruct stratified corneal epithelial cell sheets on petrolatum gauze or a soft contact lens as a carrier. Two patients, who had unilateral chemical burns, were treated by transplanting cultivated corneal epithelial cells taken from the limbus of the healthy contralateral eye. The wellestablished keratinocyte-culturing method, which involves the use of 3T3 feeder layers to help maintain epithelial stem cells, was used. AM as a suitable carrier for corneal epithelial cell culture

Researchers soon realized the potential of AM as a carrier for corneal epithelial stem cell culture. AM is the innermost layer of the fetal membrane, and it is composed of a monolayer of amniotic epithelial cells, a thick basement membrane, and an avascular stroma. AM has been used for several years in a range of ocular surgeries, with or

188

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

without limbal transplantation, and has proven to be useful for the treatment of thermal and chemical injuries, severe pterygium, persistent or deep corneal ulcers, OCP, SJS, and other limbal stem cell deficiencies. AM is known to have many unique characteristics that are beneficial to ocular surface reconstruction. Notably, AM inhibits conjunctival fibroblasts by suppressing the transforming-growth-factor-beta signaling system, and it also prevents myofibroblastic differentiation of normal fibroblasts. Furthermore, the normal differentiation of conjunctival epithelial cells is encouraged after AM transplantation. The basement membrane of AM is reported to resemble that of the conjunctival epithelium as well as that of corneal epithelium. In addition, growth factors, such as epidermal growth factor (EGF), keratinocyte growth factor (KGF), and hepatocyte growth factor (HGF), detected in AM may also play a role in accelerated epithelialization after AM transplantation. A high and therapeutic level of nerve growth factor (NGF) is also reported to be present in AM. It has been reported that AM has antiinflammatory effects by inducing the suppression of interleukin 1a and interleukin 1b in limbal epithelial cells, as well as by trapping and preventing polymorphonuclear cells from infiltrating into corneal stroma. Based on these interesting clinical and laboratory findings, preserved AM is considered to be one of the most appropriate carrier materials for transplantation of cultivated corneal epithelial cells. Although there is still debate surrounding its use, including the merits and demerits of the denuding process for amniotic epithelial cells and controversy regarding methods, the clinical results of cultivated corneal epithelial transplantation using an AM carrier are encouraging.

Cell culture procedure

Denuded AM is useful to promote the prompt migration of corneal epithelial cells in vitro, and AM has the potential to make well-stratified and differentiated corneal epithelial

(a)

cell layers that express corneal-epithelium-specific keratins K3 and K12. In our clinical experiences, we have found that the ocular surface condition of candidates for cultivated corneal epithelial transplantation is very severe, often accompanied by complications such as severe aqueousdeficient dry eye and eyelid abnormality. For these patients, we consider it essential to transplant well-stratified epithelial cell layers that have developed barrier functions with well-controlled proliferative activity in basal cells and differentiated superficial cells. For this purpose, a culture system using an air-lifting method to promote epithelial cells via tight-junction formation was developed. By airlifting, we have obtained cultivated epithelial cell sheets with smaller intercellular spaces in the superficial cells and with an epithelial barrier function. We have also attempted to transplant cultivated corneal epithelial cells, including limbal stem cells, and have developed a cellsuspension culture system capable of supplying cultivated corneal epithelial sheets that are well developed, potentially allowing the transplantation of more corneal epithelial stem cells.

Indications In 1999, the Institutional Review Board of Kyoto Prefectural University of Medicine, Kyoto, Japan, approved the transplantation of cultivated corneal epithelial cell sheets. The use of cultivated corneal epithelial sheet transplantation was restricted to those patients who had poor visual prognosis with conventional corneal epithelial transplantation, such as a KLAL. Thus, cultivated corneal epithelial transplantation was performed on 39 eyes of 36 patients with total stem cell deficiencies such as severe chemical injury, SJS, and OCP (Figure 2). While the acute-phase eyes with persistent epithelial defects received cultivated corneal epithelial transplantation for the purpose of covering the corneal surface, alleviating intensive inflammation, and avoiding complications that accompany persistent epithelial defects, the chronic-phase eyes

(b)

Figure 2 Ocular surface reconstruction using ex vivo expanded corneal epithelial cells for the chronic phase of Stevens–Johnson syndrome. (a) Before surgery, the patient suffered from total stem cell deficiency. (b) One year after surgery, the corneal surface is covered with clear corneal epithelium.

The Surgical Treatment for Corneal Epithelial Stem Cell

received cultivated corneal epithelial transplantation to obtain better visual function. Surgical procedure

In our surgical procedure, scarred conjunctival tissue overlying the ocular surface from the cornea is removed up to approximately 3 mm outside of the limbus. After removing the subconjunctival tissue, the small tips of several microsponges containing 0.04% mitomycin C are placed in the subconjunctival space adjacent to the cornea for 5 min and vigorous saline washing is then performed to prevent the development of subconjunctival fibrosis after surgery. A cultivated corneal epithelial sheet on AM is then transplanted onto the corneal surface and sutured using 10-0 nylon. A therapeutic soft contact lens is then applied. For the chronic-phase eyes with corneal stromal scarring, lamellar keratoplasty is first performed with the use of preserved donor grafts to replace the scarred corneal stroma, followed by cultivated corneal epithelial transplantation. Postoperative management

Postoperatively, 0.1% dexamethasone and 0.05% cyclosporine A are instilled topically, and dry-eye patients receive preservative-free artificial tears. Systemic corticosteroids (betamethasone, 1 mg per day), cyclosporine A (3 mg per kg of body weight per day), and cyclophosphamide (1 mg per day), as well as mycophenolate mofetil (1 g per day), in some cases, should be used to prevent postoperative inflammation and immunological rejections. Systemic immunosuppression as described above should be used for at least 6 months postoperatively, after which it can be gradually reduced depending on clinical characteristics (Table 1). In many cases, it is necessary to administer a low dose of cyclosporine A (1 mg per kg of body weight per day) for up to 2–3 years. Clinical outcome of allogeneic cultivated corneal epithelial transplantation

The epithelial integrity was satisfactory in all cases, as evidenced by the fact that the transplanted corneal epithelium did not stain with sodium fluorescein just after being transferred onto the ocular surface during surgery. In addition, in every case there was no epithelial damage to the transplanted corneal epithelium 48 h after transplantation. The transplanted AM did not disturb the visual acuity, and clarity increased day by day. Surprisingly, the preoperative ocular surface inflammation, which had not been controlled by conventional treatment, decreased rapidly after surgery in all of the acute-phase patients. In the chronic-phase eyes, the long-term visual prognosis and epithelial stability were varied in the three kinds of diseases discussed below. In the case of severe chemical injury, the transplanted corneal epithelium was clear and stable up to 8 years after transplantation, and very little

189

conjunctival inflammation was present during the entire postoperative period. On the other hand, in patients with Stevens–Johnson syndrome, mild to moderate ocular surface inflammation occurred several months after surgery, and then decreased during the following 18 months. Whereas subconjunctival fibrosis had not progressed in the eyes with SJS, conjunctival scarring such as symblepharon and shortening of the fornix had progressed in the eyes with OCP. In most of the chronic-phase patients with SJS and OCP, the phenotypes of ocular surface cells on AM gradually changed from donor to host epithelial cells over a couple of years; however, subepithelial scarring and neovascularization did not progress. In other words, host conjunctival epithelium replacement on AM occurred without scarring. This phenomenon is considered to be partly due to a mild rejection of the transplanted corneal epithelial cells. Although graft survival was not very long in some eyes in these chronic cases, the ocular surface maintained its transparency and the patients obtained a better visual function than before surgery. It is possible to perform regrafting of cultivated corneal epithelium in which the severity of epithelial opacity progressed after an episode of rejection or persistent conjunctival inflammation. In unilateral cases, autologous cultivated corneal epithelial transplantation is applicable. There is less damage to the contralateral eye than has been the case with limbal autografts, and the cultivated corneal epithelial sheets formed well-stratified epithelial layers from the very small amount of limbal tissue. After a substantial followup period, the transplanted epithelium remained transparent and stable, and the patient achieved good visual acuity with no complications in the healthy contralateral eye. Ex Vivo Expansion of Oral Mucosal Epithelial Cells Concept Due to the fact that severe ocular surface diseases are usually bilateral, allogeneic corneal epithelial transplantation (either KLAL or cultivated corneal epithelial transplantation) is normally performed. However, these procedures not only require sufficient donor tissue, but they also are accompanied by the risk of rejection; therefore, prolonged immunosuppression is required that severely affects the clinical results. With these drawbacks in mind, we have established cultivated oral mucosal epithelial transplantation using autologous tissue. Cell culture procedure Small oral biopsies (approximately 2–3 mm in size) are obtained from the oral cavity under local anesthesia. The biopsy specimens are then incubated with enzymatic reagents, such as dispase and trypsin – ethylenediaminetetraacetic acid (EDTA), to separate the cells from the underlying connective tissue. The resultant single-cell suspension

190

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

of oral mucosal epithelial cells is then co-cultured for 2–3 weeks on a denuded AM carrier with inactivated 3T3 fibroblasts. Toward the end of the culture period, an airlifting technique is used to facilitate epithelial differentiation and stratification. The oral mucosal epithelial cells cultivated on AM will show five to six layers of stratification and appear very similar to in vivo normal corneal epithelium. The cultivated oral mucosal epithelial sheet will show nonkeratinized, mucosal-specific keratins 4 and 13, and corneaspecific keratin 3; however, keratinization-related keratin 1 or keratin 10 will not be detectable. Under appropriate culture conditions, oral mucosal epithelial cells cultivated on AM have the potential ability to differentiate into cornealike epithelial cells.

Indications

In our clinical trials, we applied cultivated oral mucosal epithelial transplantation in two different forms of surgery with a total of 50 eyes. One form of surgery was reconstruction of the corneal surface of a severe bilateral corneal stem cell deficiency, using a cultivated oral mucosal epithelial sheet instead of allogeneic corneal epithelium. The other was reconstruction of the conjunctival fornix in patients with OCP, SJS, and chemical and thermal burns (Figure 3).

Surgical procedure and postoperative medications

The surgical procedure is almost the same as that of cultivated corneal epithelial sheet transplantation. After complete removal of damaged tissue on the corneal surface and subconjunctival fibroblastic tissue, residual subconjunctival tissue is treated for 5 min with 0.04% mitomycin C, followed by vigorous repeated washing with saline in order to suppress the excessive preoperative inflammation and subconjunctival fibrosis. Then, the cultivated oral mucosal epithelial sheet on AM is transplanted onto the corneal surface and secured with 10-0 nylon sutures at the limbus.

(a)

The integrity of the cultivated oral mucosal epithelium is then confirmed via intraoperative fluorescein staining, and a therapeutic soft contact lens is applied. Postoperatively, topical antibiotics and corticosteroids are usually applied. Due to the fact that it is an autograft, immunosuppressives are not necessary, except for corticosteroids and cyclosporine to control the inflammation of the original disease.

Clinical outcome of cultivated autologous oral mucosal epithelial transplantation

Using slit-lamp examination with fluorescein staining, the survival of the transplanted epithelium can be confirmed 48 h after surgery. An epithelial phenotype of transplanted cultivated oral mucosal epithelium will be somewhat distinguishable from the conjunctival epithelium by fluorescein staining. Our preliminary data show the successful survival of autologous cultivated oral mucosal epithelium on the ocular surface without returning to an in vivo oral tissue phenotype, as was previously the case with oral mucosal transplantation. This major difference can be explained by the elimination of the subconjunctival fibrous tissue and vascular component in oral mucosa during the tissue culture system. It is possible that AM has some effect on this phenomenon as well. One adverse effect of this procedure is that the transplanted cultivated oral mucosal epithelium can sometimes show some neovascularization in the peripheral cornea with epithelial thickening. For cases with poor visual recovery due to the optical corneal opacity, the two-step surgical combination of cultivated autologous oral mucosal epithelial transplantation followed by penetrating keratoplasty is advised. Cultivated oral mucosal epithelial transplantation is also useful for reconstruction of the conjunctival fornix; this form of surgery is successful in cases of cicatricial pemphigoid, chemical injury, etc. However, it is important to be aware of abnormal postoperative fibrovascular proliferation caused by primary diseases, which is still critical to the long-term prognosis.

(b)

Figure 3 Ocular surface reconstruction using cultivated autologous oral mucosal epithelium for severe total stem cell deficiency in ocular cicatricial pemphigoid. (a) Before surgery. (b) Two months after surgery, the corneal surface is covered with cultivated oral mucosal epithelium and the fornix is well reconstructed by the surgery.

The Surgical Treatment for Corneal Epithelial Stem Cell

Phototherapeutic Keratectomy for Corneal Epithelial Disorders Phototherapeutic keratectomy (PTK) using an excimer laser is a good therapeutic tool for a variety of corneal surface disorders, including corneal degenerations and dystrophies, epithelial adherence problems, persistent epithelial defects, corneal irregularities, and superficial stromal scars. PTK, with or without manual superficial keratectomy, can make the patient’s corneal surface smooth and can effectively improve visual acuity or relieve symptoms such as pain, glare, and tearing. PTK for Corneal Epithelial Defect Indications

Recurrent epithelial erosions associated with posttraumatic or epithelial basement dystrophy resistant to the conventional therapy, including lubricative medications, bandage soft contact lens, or epithelial debridement, are good indications for PTK. PTK is also very effective for Schield ulcer seen in patients with either vernal or atopic keratoconjunctivitis. Surgical procedure

Prior to an excimer laser abrasion to the surface, any disadherent epithelium adjacent to the epithelial defect, but not degenerated epithelium, should be removed by gentle manual debridement. Then, a 10–20-0m abrasion should be performed either focally or diffusely depending on the area for treatment. The area to be abraded should encompass the area of epithelial defect and include 1 mm of adjacent cornea. Application of artificial-tear eye drops just prior to laser abrasion will enhance the smoothness of the cornea surface. Postoperative management

After the laser abrasion, a disposable bandage contact lens is applied to enhance reepithelialization and reduce the pain. For the best results, the patient should wear the contact leans for up to 3 months to establish permanent epithelial-basement-membrane adhesions. Topical antibiotics and corticosteroids are essential to prevent infection and decrease postoperative inflammation. We suggest the instillation of 0.1% betamethasone and 0.3% levofloxacin 4 times per day for 2 weeks, followed by a taperingoff of the dosage.

Tectonic Lamellar Keratoplasty for Peripheral Corneal Ulcers Concept Lamellar keratoplasty is an operation in which diseased corneal tissue is removed and replaced by lamellar

191

corneal tissue from a donor. The procedure is performed either to improve vision (optical keratoplasty) or to provide structural support for the cornea (tectonic keratoplasty). Lamellar keratoplasty can be performed to replace just a portion of the corneal thickness when the endothelium is healthy. Depending on the location of the corneal abnormality, it may be sufficient to replace just the anterior layers with lamellar keratoplasty, or the full thickness of the corneal stroma without endothelium, with deep lamellar keratoplasty. Postoperative care involving the use of appropriate immunosuppressive therapy often influences the results of optic keratoplasty as well as tectonic keratoplasty. Rheumatoid Arthritis Peripheral corneal ulceration is associated with scleral or episcleral inflammation in rheumatoid arthritis (RA) patients. When corneal thinning progresses or perforation occurs in reaction to conventional steroid therapy, KEP combined with peripheral lamellar keratoplasty is effective. In RA patients, paracentral corneal perforation is also often observed. Small-size lamellar keratoplasty is a good surgical treatment for paracentral corneal perforation. However, it is important to bear in mind that the prognosis for penetrating keratoplasty is not good for RA patients. Pre- and postoperative antirheumatic therapies improve surgical results. We recommend the application of therapeutic soft contact lenses at the conclusion of surgery; these should then be continuously used for several years to avoid the infiltration of immunoreactive cells from tears and prevent the recurrence of ulceration. Topical corticosteroids and antibiotic drops should be applied 4 times per day. The frequency should then be decreased as inflammatory signs subside to a level of two to three drops per day for several months. Careful removal of the sutures is performed during the first and second postoperative months to avoid epithelial damage. To avoid the recurrence of the original disease, corticosteroid drops should be continued once or twice per day for a number of years. Mooren’s Ulcer For the treatment of severe Mooren’s ulcer, which does not respond to steroids or immunosuppressive therapy using cyclosporine A, one surgical option which is available is KEP, with or without lamellar keratoplasty. By transplanting the Bowman’s layer with a thin corneal stroma onto the sclera adjacent to the ulcerated area, inflamed conjunctival tissue is unable to invade the corneal surface and cellular infiltration of the ulcerated peripheral cornea is prevented (Figure 4). Pre- and postoperative anti-inflammatory therapy, including systemic cyclosporine A, is important for the achievement of good surgical results. In our cornea service, we administer oral

192

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

(a)

(b)

Figure 4 Keratoepithelioplasty for Mooren’s ulcer. (a) Before surgery (visible from the 1 o’clock to 6 o’clock position). (b) Ten years after keratoepithelioplasty. There was no recurrence of Mooren’s ulcer.

cyclosporine A (100–200 mg per day) and oral betamethasone (2–4 mg), as well as topical 0.1% dexamethasone (4 times per day) and antibiotics (0.3% ofloxacin). As mentioned earlier, careful removal of the sutures is performed during the first and second postoperative months to avoid epithelial damage. To avoid the recurrence of the original disease, the extended use of therapeutic soft contact lenses and corticosteroid eye drops (1–2 times per day) should be continued for as many years as possible. See also: Contact Lenses; Corneal Epithelium: Cell Biology and Basic Science; Corneal Epithelium: Wound Healing Junctions, Attachment to Stroma Receptors, Matrix Metalloproteinases, Intracellular Communications; Defense Mechanisms of Tears and Ocular Surface; Lids: Anatomy, Pathophysiology, Mucocutaneous Junction; Stem Cells of the Ocular Surface.

Further Reading Das, S. and Seitz, B. (2008). Recurrent corneal erosion syndrome. Survey of Ophthalmolgy 53: 3–15. Dua, H. S. and Azuara-Blanco, A. (1999). Amniotic membrane transplantation. British Journal Ophtalmology 83: 748–752. Holland, E. J., Schwartz, G. S., and Nordlund, M. L. (2005). Surgical technique for ocular surface reconstruction. In: Krachmer, J. H., Mannis, M. J., and Holland, E. J. (eds.) Cornea, 2nd edn., pp. 1799–1812. Philadelphia, PA: Elsevier Mosby. Kenyon, K. R. (1989). Limbal autograft transplantation for chemical and thermal burns. Developments in Ophthalmology 18: 53–58. Kinoshita, S., Koizumi, N., and Nakamura, T. (2004). Transplantable cultivated mucosal epithelial sheet for ocular surface reconstruction. Experimental Eye Research 78: 483–491. Kinoshita, S., Koizumi, N., Sotozono, C., et al. (2004). Concept and clinical application of cultivated epithelial transplantation for ocular surface disorders. Ocularsurface 2: 21–33. Kinoshita, S., Ohashi, Y., Ohji, M., and Manabe, R. (1991). Long-term results of keratoepithelioplasty in Mooren’s ulcer. Ophthalmology 98: 438–445.

Refractive Surgery S Marcos, L Llorente, C Dorronsoro, and J Merayo-Lloves, Consejo Superior de Investigaciones Cientı´ficas, Madrid, Spain ã 2010 Elsevier Ltd. All rights reserved.

Glossary Aberrations – Phase deviations from the ideal wave front measured at the pupil plane. Aberrometers measure local ray deviations, which are fitted to the derivatives of the wave aberration, usually expressed as a Zernike polynomial expansion. A relevant aberration in the human eye (and particularly after standard refractive surgery) is spherical aberration (which results in peripheral rays converging on a different plane than central rays). A common optical quality metric derived from the wave aberration is the root mean square (RMS) wave-front error. Ablation profile – Corneal tissue that needs to be removed at each location to produce the desired change in corneal power. Generally, the ablation-profile equation is converted into the number of laser pulses to be applied at each location. Asphericity – Parameter used to describe the deviation of the anterior corneal surface from a sphere. The corneal surface can be fitted to a conic section using the apical radius of curvature and the eccentricity e (variation of this curve with distance from the apex). Asphericity Q is defined as –e2, with the surface represented by the following equation: (X 2+Y 2)+(1+Q)Z22ZR = 0. For a sphere Q = 0, a typical cornea shows an asphericity Q =0.26; a surface with zero spherical aberration should have an asphericity Q =0.52; an oblate surface (exhibiting positive spherical aberration) will have Q > 0; and a prolate surface (exhibiting low positive spherical aberration or negative) will have Q < 0. Beer–Lambert law – Law governing the photoablation of the corneal tissue by excimer laser. The depth of ablated material is proportional to the logarithm of the laser fluence (relative to the ablation threshold). Contrast-sensitivity function (CSF) – The contrast-sensitivity function represents the minimum subjectively discernible contrast as a function of spatial frequency. It typically has an inverted-U shape, peaking at around 4 cycles per degree (c/deg), with sensitivity decreasing on either side of the peak. The shape of the CSF is determined by the properties of the visual neurons and the optical aberrations of the eye. Other factors affecting the

CSF are pupil diameter and luminance. CSF is a more sensitive measure of changes in visual quality following a change in the optics (such as that produced by refractive surgery) than visual acuity. Excimer laser – Laser producing stimulated emission after electrical discharge forming dimers or complexes, emitting typically ultraviolet (UV) light. Lasers applied in refractive surgery use a combination of Argon (as inert compound) and fluorine (reactive gas) and emit at 193 nm. The excimer lasers are well suited to remove exceptionally fine layers of surface material (particularly, biological matter and organic compounds) by disrupting the molecular bonds of the tissue, through ablation rather than burning, leaving the remainder of the material almost intact. Lasers used in refractive surgery have fluences typically ranging between 120 and 400 mJ cm2. Laser-assisted in situ keratomileusis (LASIK) – Corneal refractive surgery technique which involves the creation of a thin flap on the cornea, folding it to enable remodeling of the tissue underneath with laser and repositioning the flap back after the corneal ablation has been performed. Modulation-transfer function (MTF) – Optical function representing the contrast degradation by an optical system as a function of spatial frequency. Factors affecting the MTF are diffraction (pupil size), optical aberrations, scattering, and wavelength. The ocular MTF is a low-pass function, with a cut-off frequency at around 70 c/deg. Munnerlyn formula – Equation on which standard ablation algorithms for corneal refractive surgery are based. The corneal tissue to be removed is a lenticule with an anterior radius of curvature equal to the preoperative corneal radius and the posterior radius of curvature equal to the postoperative corneal radius (easily related with the attempted correction). A parabolic approximation of the Munnerlyn formula states that the depth of the ablation (in microns) per diopter of refractive change is equal to the square of the optical ablation zone measured in millimeters, divided by 3. Optical zone – Area of the cornea where the ablation algorithm is applied. For the same attempted

193

194

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

correction, smaller optical zones require less ablation depth. However, small optical zones can severely compromise vision if they are smaller than pupil diameter. A transition zone around the optical zone is created to avoid an abrupt step between treated and untreated areas.

Short Historical Background of Corneal Refractive Surgery Corneal refractive surgery has become, in recent years, a popular surgical approach to correct ametropia. Table 1 depicts highlights of refractive surgery history in chronological order. The first surgical procedure that achieved a change in corneal power by relaxation through corneal incision was radial keratotomy (RK). First attempts of incisional corneal surgery date from 1885 onward.

Table 1 1898 1930 1948

1960s 1978 1975–79

1983 1988 1989

1991 1995 1999 1999

2000 2002 2002 2003

Chronology of corneal refractive surgery The basic principles of radial keratotomy are laid up by Dr. Lans (the Netherlands) Pioneering work on corneal incisions is performed by Dr. Tsutomu Sato (Japan) The first surgical techniques to reshape the cornea (freezing a corneal flap, reshaping it on a lathe and placing it back) are performed by Dr. Jose´ Barrarquer (Colombia) Radial keratotomy is developed by Dr. Fyodorov (Russia) Radial keratotomy is introduced in the US by Dr. Leo Bores Excimer Laser Technology is developed. Dr. Srinavasan (IBM Laboratories, USA) foresees the potential of the interactions of laser and biological tissue The use of excimer laser to remove corneal tissue is described by Dr. Stephen Trokel (USA) The first excimer treatment on a human eye is performed by Dr. Teo Seiler (Germany) The benefits of performing PRK after a flap was removed were theorized and experimented by Dr. Pallikaris (Greece) and Lucio Buratto (Italy) The first LASIK procedure is performed by Dr. Stephen Brint in the United States FDA approves excimer laser for refractive surgery (PRK) to correct myopia First excimer lasers approved to perform LASIK surgery LASEK (a surface ablation procedure in which the epithelium is removed with alcohol) is introduced by Dr. Massimo Camellin (Italy) FDA approves LASIK surgery for hyperopia Wavefront-guided LASIK approved for custom correction Femtosecond laser flap removal approved by FDA First Epi-LASIK (epithelium mechanically removed) procedures performed by Dr. Pallikaris (Greece)

The technique evolved through the 1930s and 1940s to the early 1980s when Fyodorov developed a systematic and more predictable RK procedure that he applied to thousands of patients. In the 1960s, Barraquer invented keratomileusis – the first lamellar surgical technique. This technique consisted of separating a thin layer of the superficial corneal tissue using a microkeratome, removing a small piece of cornea, which was frozen and then reshaped using a lathe, and suturing it back into place. However, it was only in the late 1980s, when excimer lasers were developed at IBM, that their excellent properties for micromachining of biological tissue and organic materials were identified by Srinivasan. In the 1980s, Stephen Trokel applied, for the first time, an argonfluoride excimer laser to remove tissue in bovine corneas, following previous mechanical removal of the outer layer of the cornea (corneal epithelium) to treat ametropia, thus giving birth to photorefractive keratectomy (PRK). However, PRK was limited by unpredictability in higher ranges of refractive error and higher risk of corneal haze after surgery. In the 1990s, Pallikaris combined these two techniques (keratomileusis and PRK), creating the laser-assisted in situ keratomileusis (LASIK), which has become the most popular refractive surgery technique (see the section titled ‘The LASIK procedure’). Today, faster lasers, larger spot areas, bladeless flap creation, intraoperative pachymetry, and wave-front-optimized and wave-front-guided techniques have significantly improved the reliability of the procedure compared to that of 1991. Nonetheless, the fundamental limitations of excimer lasers, the limited corneal thickness (particularly in the presence of the flap), and undesirable destruction of corneal nerves have spawned research into many alternatives to standard LASIK, including laser-assisted subepithelial keratomileusis (LASEK) or Epi-LASIK, which aim at combining the advantages of surface ablations such as PRK with those of LASIK surgery. Although safety and efficacy, and refraction predictability of PRK and LASIK are high, complaints of decreased vision and glare in mesopic and scotopic light levels, that is, night-vision problems, exist. Haze, halos, and increased optical aberrations are attributed to cause visual degradation, particularly in eyes that had undergone high refraction corrections. Several questions are still open today: proper transfer of the ablation profile to the cornea, wound healing, biological response, corneal biomechanics, microstructural stromal changes, and longterm healing. The implementation of aberrometry in refractive surgery has meant a turning point in the history of laser refractive surgery since – along with other technological advances including improvements in surgical lasers (such as flying spot lasers), ablation algorithms, and eyetracking – the measurement of ocular wave aberrations has opened the potential for improved refractive surgery, aiming not only at correcting refractive errors but also to minimize optical aberrations of the eye.

Refractive Surgery

The LASIK Procedure Figure 1 illustrates a LASIK procedure. In this technique, a hinged flap is created by means of a microkeratome, and folded back to leave the stroma exposed. To create the flap, a corneal suction ring is applied to the eye, holding the eye in place. Once the eye is immobilized, the flap is created. This process is achieved with a mechanical microkeratome using a metal blade, or more recently a femtosecond laser microkeratome. A hinge is left at one end of this flap. The flap is folded back, revealing the stroma, the middle section of the cornea. An excimer laser is then used to photoablate the stroma. For treatment of myopia, the central cornea is flattened and experiences a deeper ablation than the periphery. For a hyperopic treatment, the outer area of the optical zone experiences deeper ablation than the central area resulting in a conelike corneal profile. After the laser has reshaped the stromal layer, the LASIK flap is carefully repositioned over the treatment area by the surgeon and checked for the presence of air bubbles, debris, and proper fit on the eye.

Figure 1 Illustration of a LASIK procedure. A flap is lifted and the exposed stroma ablated with excimer laser pulses according to the programmed ablation profile following which the flap is repositioned.

(a)

195

The flap remains in position by natural adhesion until healing is completed.

Corneal Refractive Surgery, Wound Healing, and Haze The corneal response to laser refractive surgery induces keratocyte apoptosis immediately following the procedure. Proliferation and migration of keratocytes begins within 12–24 h, giving rise to activated keratocytes and myofibroblasts, which are critical components in the wound-healing cascade. Myofibroblasts and newly synthesized extracellular matrix play a major role in haze formation and regression due to stromal remodeling. The timing, intensity, and spatial distribution of wound healing vary significantly between LASIK and PRK. PRK involves injury on a broader area and removal of the epithelium, epithelial basement membrane, Bowman’s layer, and a portion of the anterior stroma, while LASIK leaves these structures relatively undisturbed, except at the flap margin. Figure 2 depicts immunohistology microscopic images of excised corneas in an avian model following PRK (5, 15, and 30 days following surgery), showing the distribution of myofibroblasts. Refractive regression is a major challenge following PRK for myopia, hyperopia, and astigmatism, especially for high levels of correction, and is both more common and more pronounced than the regression following LASIK. The source of regression is attributed to differential changes in the thickness of the cornea due to a combination of stromal remodeling and epithelial hyperplasia. The intensity of the corneal response is related to the magnitude of attempted treatment.

Corneal Refractive Surgery, Optical Aberrations, and Visual Quality One of the most important side effects of standard refractive surgery is the induction of higher-order aberrations. Early

(b)

Figure 2 Immunohistology of chicken corneas following PRK, stained for alpha smooth muscle myosin (a-SMA) as a marker for myofibroblasts ((a) 6 days and (b) 30 days). Magnification 200. From Merayo-Lloves, J., Yan˜ez, B., Mayo, A., Martı´n, R., and Pastor, J. C. (2001). Experimental model of corneal haze. Journal of Refractive Surgery 17: 696–699.

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

studies based on corneal topography showed that while defocus or astigmatism are generally successfully corrected, refractive surgery (RK, PRK, and LASIK) increased the amount of corneal aberrations. In addition, the distribution of aberrations changed from the third-order dominance found in normal subjects, to fourth-order dominance. This increase in corneal aberrations correlates well with the decrease found in contrast sensitivity. Seiler and colleagues, in standard myopic PRK (15 eyes, mean preoperative spherical error ¼ 4.8 D), and Moreno–Barriuso and colleagues, in standard myopic LASIK (22 eyes, mean preoperative spherical error ¼ 6.5 D), measured, for the first time, the changes in the total aberration pattern induced by either type of surgery. Both studies found a significant increase in third- and higher-order aberrations (by a factor of 4.2 and 1.9 in the root mean square (RMS), respectively). The larger increase occurred for spherical and third-order aberrations. The changes of total spherical aberrations are not fully accounted by changes in the anterior corneal surface. In all eyes, total spherical aberration increased slightly less than corneal aberrations, likely due to significant changes in the posterior corneal shape (shifting toward more negative values of spherical aberration). The increase in the total spherical aberration is highly correlated to the amount of spherical error corrected, and it is associated with an increase in corneal asphericity. Changes of corneal and total aberrations with LASIK surgery for hyperopia are even higher than those for LASIK surgery for myopia. While spherical aberration becomes more positive following myopic LASIK, it shifts toward negative values following hyperopic LASIK. For the same absolute amount of correction, the absolute increase of corneal spherical aberration is larger with hyperopic LASIK. Figure 3 shows wave pre- and postoperative high-order

aberration patterns in patients that had undergone myopic LASIK and hyperopic LASIK. Figure 4 compares the induced aberration (total and corneal) following myopic and hyperopic LASIK, respectively. Modulation-transfer functions (MTFs) can be computed from the measured wave aberrations, and the optical changes can be compared to the visual changes (measured in terms of contrast-sensitivity function (CSF)). Marcos and colleagues found that the decrease in the MTF

Total aberration

Hyperopic LASIK

Corneal aberration

1.5 1 0.5 0 −0.5 −1 −1.5 −2

0

2

4 6 8 10 Amount of pre-op spherical error (D)

14

Hyperopic LASIK Total

Corneal

Corneal µm

µm Pre

12

Figure 4 Induced spherical aberration vs. spherical correction (positive for hyperopia and negative for myopia). Corneal spherical aberration increases at a rate of 0.17 mm D1 in myopic LASIK and –0.23 mm D1 in hyperopic LASIK. Total spherical aberration increases at a rate of 0.09 mm D1 in myopic LASIK and –0.06 mm D1 in hyperopic LASIK. The inset depicts the fourth-order spherical-aberration Zernike term. From Llorente, L., Barbero, B., Merayo, J., and Marcos, S. (2004). Changes in corneal and total aberrations induced by LASIK surgery for hyperopia. Journal of Refractive Surgery 20: 203–216.

Myopic LASIK Total

Myopic LASIK

2 Spherical aberration (µm)

196

4

2

Pre

0 2 RMS = 0.54 µm

RMS = 0.61 µm

RMS = 0.34 µm

0

RMS = 0.63 µm

−4

−2 Post

−2

Post

RMS = 1.40 µm

RMS = 1.49 µm

RMS = 1.13 µm

RMS = 0.82 µm

Figure 3 Examples of corneal and total wave aberrations (third- and higher-order aberrations) before and after LASIK surgery for myopia (left panel) and hyperopia (right panel). The increase in aberrations is indicated by the increased RMS. Following surgery, the total aberration map is dominated by the corneal contribution, primarily by positive spherical aberration following myopic LASIK and by negative spherical aberration following hyperopic LASIK. From Marcos, S., Barbero, B., Llorente, L., and Merayo-Lloves, J. (2001). Optical response to LASIK for myopia from total and corneal aberrations. Investigative Ophthalmology and Visual Science 42: 3349–3356.

Refractive Surgery

(between 3 and 18 cycles per degree (c/deg)) was 1.38 – similar to the decrease in the CSF, by a factor of 1.51 (in the same spatial frequency range) – on average in a group of 22 eyes that had undergone LASIK surgery for myopia. This indicates that the increase in optical aberrations plays a major role in the decrease of visual quality following LASIK. Figure 5 shows pre- and postoperative MTF and CSF, for 3-mm pupils, and the pre/post contrast ratio as a function of spatial frequency.

197

the energy is not properly transferred onto the cornea, due to changes in laser efficiency across the corneal surface. Several authors have derived theoretical expressions (based on the Beer–Lambert law and Fresnel equations) to account for the laser-energy losses from the center to the periphery of the cornea. Figure 6 depicts average results on 13 patients of postoperative corneal asphericities following computer simulation of the Munnerlyn ablation pattern (and its parabolic approximation), directly or considering laser-efficiency changes across the cornea (based on the equations proposed by Jime´nez and colleagues), in comparison with average preoperative asphericities and real postoperative asphericities in the same eyes. An interesting approach to the understanding of the induction of spherical aberration is the ablation of plastic spherical surfaces, which are subject to geometrical effects but not biomechanical response. Ablation of poly(methyl methacrylate) (PMMA)-model eyes produces an increase in spherical aberration, similar to the increase found in real corneas. In addition, a comparison of the ablation profile of flat and spherical surfaces allows direct estimation of the laser-efficiency losses on PMMA, which can be extrapolated to corneal tissue, without relying on the exact knowledge of the ablation profile programmed into the laser system, and

Causes for Spherical Aberration Increase Following Corneal Refractive Surgery The causes for the increase of spherical aberration (and corneal asphericity) are still not well understood. Computer simulations of the postoperative corneal shape following subtraction of the standard ablation pattern (Munnerlyn equation) performed on real preoperative corneal elevation maps do not show the increased corneal asphericity found clinically. A parabolic approximation of this equation induces a slight increase of corneal asphericity, but much less than is found experimentally. It is likely that much of the discrepancy is due to the fact that

MTF

1.0

CSF

Modulation transfer

MTF 0.8 0.6 0.4 0.2 0.0

0

10

20

30

3 6 12 18 Spatial frequency (c / deg)

Spatial frequency (c / deg) Contrast ratio POST/PRE

Contrast sensitivity

1.0

CSF

120

80

40

0

0

4

8

12

16

Spatial frequency (c / deg)

20

0.8

MTF ratio 0.6

CSF ratio

0.4 0.2

0

10

20

30

Spatial frequency (c / deg)

Figure 5 Illustration of the estimation of the MTF from the measured wave aberration, and of the measured CSF, using a clinical gold standard (CVS1000, Vectorvision), in the upper-right panel. Average preoperative MTF (in green) and postoperative MTF (in red), horizontal sections, in the upper-left panel. Average preoperative CSF (in green) and postoperative MTF (in red) for vertical gratings, in the lower-left panel. The area under the curve (shaded in green and red, respectively) between 3 and 18 c/deg was used as a metric for optical quality. The ratio of the area under the MTF decreased by a factor of 1.38, and under the CSF by a factor of 1.51 – indicating that the optical changes have a similar visual impact. The lower-right panel shows the ratio of the MTF and the CSF post/pre as a function of spatial frequency. The drop in contrast follows the same trend for both the MTF and the CSF. Data are average of 22 eyes that had undergone refractive surgery for myopia (between –2.5 and –13 D spherical correction), for undilated pupils. Marcos, S. (2001). Aberrations and visual performance following standard laser vision correction. Journal of Refractive Surgery 17: 596–601.

198

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

5

Prolate 0 < Q < −1

Postoperative corneal asphericity

Y

4

Sphere Q=0 X

Clinical data

Z

Experimental ablation profile and efficiency

Oblate Q > 0

3

Experimental ablation profile and theoretical eff.

2

Experimental ablation profile

1

Parabollic profile 0

−1

Munnerlyn profile

0

2

4

6

8

10

12

14

Spherical correction (D) Figure 6 Postoperative corneal asphericity following LASIK surgery for myopia as a function of preoperative spherical error. The graph shows experimental postoperative data from patients – estimations from a virtual surgery using the experimental ablation profile measured on PMMA flat surfaces (and a conversion factor from PMMA to cornea) and an experimentally measured laser-efficiency factor (which takes into account the energy loss from the center to the periphery), estimations from a virtual surgery using the experimental ablation profile and a theoretical laser-efficiency factor, estimations from the experimental ablation profile, the Munnerlyn profile, and its parabolic approximation without considering laser-efficiency losses. From Cano, D., Barbero, B., and Marcos, S. (2004). Comparison of real and computer-simulated outcomes of LASIK refractive surgery. Journal of the Optical Society of America A 21: 926–936. and Dorronsoro, C., Cano, D., Merayo, J., and Marcos, S. (2006). Experiments on PMMA models to predict the impact of corneal refractive surgery on corneal shape. Optics Express 14: 6142–6156.

without the approximation of the theoretical approaches. Figure 6 shows estimated postoperative corneal asphericities following direct subtraction of the ablation profile obtained experimentally from profilometry of flat plastic surfaces and also considering the experimental efficiency factor.

(keratoconus) and reducing the risk of mechanical instability by leaving a minimal residual stromal thickness (of 300 mm or more) has proved to reduce the occurrence of corneal ectasia greatly. The change in the dynamical response of the cornea, inherent to its biomechanical properties, also affects the measurement of intraocular pressure with standard tonometers, which assume normalized values of corneal elasticity.

Corneal Biomechanical Effects in Refractive Surgery Although not fully understood, it seems clear that the biomechanical properties of the cornea change following refractive surgery. RK relied fully on the mechanical relaxation of the cornea following incisions. During PRK, LASIK, or any other procedure involving central ablation, corneal lamellae are severed. In a simple elastic shell model, if considered alone, this would result in corneal steepening. However, it has been suggested that a lamellar tension relaxation in the peripheral stroma occurs which produces a compression of the anterior cornea and central flattening. The elastic modulus of the residual stromal bed and the corneal shear strength likely change following surgery, as the cohesive forces among lamellae, stromal swelling pressure, are affected by the procedure. These effects have a potential relevance in the pathogenesis of ectasia, characterized by a progressive thinning and a progressive central and inferior steepening of the cornea, affecting 0.3% of the LASIK patients, although a careful identification of patients with preexisting corneal pathology

Other Side Effects and Complications of PRK and LASIK Apart from the indicated increased aberrations (resulting in halos and ghost images), haze and loss of best-corrected contrast sensitivity, other potential complications may occur during or following surgery. The most common side effect from refractive surgery is dry eye (occurring in 36% of patients), diffuse lamellar keratitis (2.3%), flap complications including slipped flap, debris or growth under flap and flap striae (0.244%), infection (0.4%), epithelial ingrowth (0.1%), glare, and light sensitivity.

Safety, Efficacy, and Satisfaction of LASIK An average of 700 000 patients in the US undergo LASIK annually. To date, more than 28.3 million LASIK

Refractive Surgery

199

Toward an Optimization of the Corneal Refractive Surgery Procedure

procedures have been performed worldwide. Collectively, 7830 patients (representing 16 502 eyes) participated in clinical trials from 1993 to 2005. In April 2008, the Food and Drug Administration (FDA) reaffirmed the safety and efficacy of the LASIK procedure. Although the number of patient complaints has increased in the last few years, the postoperative visual outcomes have improved in the most recent studies. In FDA studies recruiting patients that had undergone surgery prior to 2000, 1.4% of patients lost two lines or more of best spectacle-corrected visual acuity (BSCVA) versus 0.6% in studies after 2000. Prior to 2000, 1.68% of patients with a preoperative BSCVA 20/20 or higher had a postoperative BSCVA 20/25 or higher, compared with 0.16% after 2000. The surveys determining patient satisfaction with LASIK have found most patients satisfied, with satisfaction ranging between 92% and 98%. A metaanalysis – dated March 2008, performed by the American Society of Cataract and Refractive Surgery over 3000 peer-reviewed articles published over the past 10 years in clinical journals from around the world, including 19 studies comprising 2200 patients – that looked directly at satisfaction, revealed a 95.4% patient satisfaction rate among LASIK patients worldwide.

Recent technological advances in refractive surgery include high-frequency eye-tracking, improved laser-delivery systems, and flap creation by femtosecond lasers. The availability of clinical aberrometers, and flying spot technology, led to the development of wave-front-guided ablation profiles – aiming not only at correcting defocus and astigmatism, but also the eye’s high-order aberrations. However, major efforts must still focus on preventing the induction of high-order aberrations (particularly spherical aberration) by the procedure, particularly as these are not necessarily inherent to ablation algorithm (despite its simplicity, the standard Munnerlyn algorithm does not induce spherical aberration). Approaches range from theoretical correction of the ablation profile to empirical adjustment of the attempted corneal asphericity. Undoubtedly, a proper characterization of the ablation algorithm, and a calibration of the ablation pattern created by the laser is critical to achieve the desired postoperative corneal shape. Plastic models (in PMMA, and, more recently, a semi-rigid contact lens material – Filofocon A) have been proposed as calibration models for refractive surgery, and their ablation properties

0.005 0 −0.005

Ablation depth (␮m)

−0.01 −0.015 −0.02

−3D

−0.025 −0.03 −0.035 6 4

−6D

2 0 −2 −4 −6

−6

−4

−2

0

2

4

6

Surface location (mm) Figure 7 Ablation profiles –6 and –3 D spherical corrections measured experimentally by subtraction of pre- and postablation corneal topographies of PMMA corneal models. From Dorronsoro, C., Cano, D., Merayo, J., and Marcos, S. (2006). Experiments on PMMA models to predict the impact of corneal refractive surgery on corneal shape. Optics Express 14: 6142–6156.

200

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

have been thoroughly studied. Figure 7 shows an example of the ablation profile recorded on a plastic material and measured using noncontact profilometry. To date, empirical adjustments of the ablation normogram have – based on population average data – compensated for deviations from ametropia systematically found following surgery. However, customized corrections require a deeper knowledge of ablation, physical effects of the role of corneal hydration, and of the contribution of corneal biomechanical effects and their inter-individual variations on the refractive and highorder aberration outcomes. A fine control of the ablation profile and the biomechanical response will likely expand the range of applications of refractive surgery into presbyopic treatments. As an emerging technology, corneal collagen cross-linking is starting to be applied following refractive surgery as a way to control corneal ectasia by increasing corneal stiffness. Other technologies further in the track include solidestate ultraviolet (UV) lasers and femtosecond lasers. Diode-pumbed UV laser, using a solide-state laser crystal as the laser medium and nonlinear crystals for frequency conversion instead of the high-voltage gas discharge of excimer lasers, are a safer, more stable, compact, and less expensive alternative to gas-operated excimer lasers. On the other hand, femtosecond lasers use ultrashort pulses (as opposed to pulse durations in the range of nanosecond or picosecond in excimer lasers) which allow laser–tissue interactions characterized by significantly smaller and more deterministic photodisruptive energy thresholds, as well as reduced shock waves and smaller cavitation bubbles, which result in smoother surface quality. Lamellar procedures (keratomileusis) as well as lenticule removal are envisaged with this procedure. See also: Cornea Overview; Corneal Nerves: Anatomy; Hyperopia; Myopia; Refractive Surgery and Inlays.

Further Reading Applegate, R. A., Hilmantel, G., and Howland, H. C. (1996). Corneal aberrations increase with the magnitude of radial keratotomy refractive correction. Optometry and Vision Science 73(9): 585–589. Applegate, R. A. and Howland, H. C. (1997). Refractive surgery, optical aberrations, and visual perfomance. Journal of Refractive Surgery 13: 295–299. Buratto, L. and Brint, S. F. (eds.) (2003). Custom LASIK: Surgical Techniques and Commplications. Thorofare, NJ: Slack. Cano, D., Barbero, B., and Marcos, S. (2004). Comparison of real and computer-simulated outcomes of LASIK refractive surgery. Journal of the Optical Society of America A 21: 926–936.

Dorronsoro, C., Cano, D., Merayo, J., and Marcos, S. (2006). Experiments on PMMA models to predict the impact of corneal refractive surgery on corneal shape. Optics Express 14: 6142–6156. Dorronsoro, C., Siegel, J., Remon, L., and Marcos, S. (2008). Suitability of Filofocon A and PMMA for experimental models in excimer laser ablation refractive surgery. Optics Express 16: 20955–20967. Dupps, W. and Wilson, S. (2006). Biomechanics and wound healing in the cornea. Experimental Eye Research 83: 709–720. Jime´nez, J., Anera, R., Jime´nez del Barco, L., and Hita, E. (2002). Effect on laser-ablation algorithms of reflection losses and nonnormal incidence on the anterior cornea. Applied Physics Letters 81(8): 1521–1523. Krueger, R., Applegate, R. A., and MacRae, S. (eds.) (2004). Wavefront Customized Visual Correction: The Quest for Super Vision II. Thorofare, NJ: Slack. Llorente, L., Barbero, B., Merayo, J., and Marcos, S. (2004). Changes in corneal and total aberrations induced by LASIK surgery for hyperopia. Journal of Refractive Surgery 20: 203–216. Marcos, S. (2001). Aberrations and visual performance following standard laser vision correction. Journal of Refractive Surgery 17: 596–601. Marcos, S., Barbero, B., Llorente, L., and Merayo-Lloves, J. (2001). Optical response to LASIK for myopia from total and corneal aberrations. Investigative Ophthalmology and Visual Science 42: 3349–3356. Marcos, S., Cano, D., and Barbero, S. (2003). The increase of corneal asphericity after standard myopic LASIK surgery is not inherent to the Munnerlyn algorithm. Journal of Refractive Surgery 19: 592–596. Merayo-Lloves, J., Yan˜ez, B., Mayo, A., Martı´n, R., and Pastor, J. C. (2001). Experimental model of corneal haze. Journal of Refractive Surgery 17: 696–699. Moreno-Barriuso, E., Merayo-Lloves, J., Marcos, S., et al. (2001). Ocular aberrations before and after myopic corneal refractive surgery: LASIK-induced changes measured with laser ray tracing. Investigative Ophthalmology and Visual Science 42: 1396–1403. Mrochen, M., Krueger, R., Bueeler, M., and Seiler, T. (2002). Aberrationsensing and wavefront-guided laser in situ keratomileusis: Management of decentered ablation. Journal of Refractive Surgery 18: 418–429. Netto, M. V., Mohan, R. R., Ambro´sio, R., Jr., et al. (2005). Wound healing in the cornea: A review of refractive surgery complications and new prospects for therapy. Cornea 24: 509–522. Pallikaris, I. G., Agarwal, S., and Agarwal, A. (2003). Refractive Surgery. Thorofare, NJ: Slack. Seiler, T., Kaemmerer, M., Mierdel, P., and Krinke, H.-E. (2000). Ocular optical aberrations after photorefractive keratectomy for myopia and myopic astigmatism. Archive of Ophthalmology 118: 17–21.

Relevant Websites http://www.geteyesmart.org – American Academy of Ophthalmology and Its Partners. http://www.ascrs.org – American Society of Cataract and Refractive Surgery (ASCRS). http://www.escrs.org – European Society of Cataract and Refractive Surgery. http://www.aao.org/isrs – International Society of Refractive Surgery. http://www.fda.gov – US Food and Drug Administration. http://www.vision.csic.es – Visual Optics and Biophotonics Lab, Instituto de Optica, CSIC.

Refractive Surgery and Inlays R M M A Nuijts, M Doors, N G Tahzib, and L P J Cruysberg, University Hospital Maastricht, Maastricht, The Netherlands ã 2010 Elsevier Ltd. All rights reserved.

Glossary Astigmatism – A refractive defect in which vision is blurred due to the inability of the optics of the eye to focus a point object into a sharp, focused image on the retina due to an irregular or toric curvature of the cornea or lens. Cycloplegic refraction – A measurement of the refractive state of the eye without the effects of accommodation. A cycloplegic drop is used to temporarily paralyze the accommodation muscle. Corneal ectasia – A serious complication involving a cone-like bulging of the cornea following its weakening during laser-assisted in situ keratomileusis (LASIK). Diffuse lamellar keratitis – Noninfectious inflammatory complication of LASIK. Form fruste keratoconus – An abortive form of bulging of the cornea. Hyperopia – Refractive defect caused by an eye that is too short or a cornea that is too flat, so that images focus at a point behind the retina. It is also called farsightedness. Intracorneal ring (ICR) – Small ring inserted into the periphery of the cornea to change its shape and correct nearsightedness. Laser-assisted subepithelial keratectomy (LASEK) – A refractive surgery where only the corneal epithelia is cut to reshape the cornea. Laser-assisted in situ keratomileusis (LASIK) – A refractive surgery where the corneal epithelia and stroma is cut to reshape the cornea. Manifest – Easily seen. Myopia – Refractive defect of the eye where the light focuses in front of the retina rather than on the retina. It is also called nearsightedness. Photorefractive keratectomy (PRK) – A refractive surgery where the corneal epithelia is removed to reshape the cornea. Presbyopia – Reduced ability to see near objects caused by loss if the elasticity of the lens. Radial keratotomy (RK) – Surgical procedure to correct myopia where radial incisions are made into the cornea at precise depths allowing the sides of the cornea to bulge out and flatten the central cornea.

Introduction Corneal refractive surgery offers the patient the possibility of becoming independent of spectacles and/or contact lenses. The attainment of this treatment goal is of particular importance to individuals who are restricted in their professional and social life by their contact lens or spectacle intolerance. The developments and outcomes of various refractive surgery techniques have received increasing attention in the medical literature and public media. This phenomenon is mainly related to the numerous success stories and the dramatic changes achieved by correction of the refractive error and the resultant independence of spectacles and contact lenses. Numerous corneal refractive surgery techniques are available for the correction of refractive errors, with the majority of treatments consisting of myopic and myopic– astigmatic corrections. The field of refractive surgery has greatly evolved since the commencement of excimer laser treatments and surgical implantations of corneal inlays. The techniques have been refined and are continually evolving to be more specifically directed toward the individual optical design. The optical system can differ greatly between individuals and depends on various factors, such as the amount of the refractive error and the degree of optical aberrations. Therefore, laser-ablation techniques have changed from the standard correction of the refractive error to personalized and optimized laser treatments and from broad-beam to scanning-spot or flying-spot devices. This article presents an overview of the available corneal refractive procedures and their outcomes.

Radial Keratotomy Prior to the popularity of excimer photoablative refractive surgery, the technique of radial keratotomy (RK) was among the most widely used surgical techniques for the correction of myopia. RK involves making deep radial incisions in the paracentral and peripheral anterior cornea using a diamond blade knife. The technique results in the flattening of the central corneal curvature and steepening of the peripheral area, which reduces the degree of myopia. The number of RK incisions, diameter of the optic zone, and patient age determine the refractive outcome after RK. Incision direction was shown to be another predictor, with

201

202

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

the centripetal (vs. the centrifugal) incision decreasing myopia to a higher degree. Although the treatment by RK initially resulted in satisfactory refractive results, it appeared not to be as predictable as current refractive surgery techniques. The Prospective Evaluation of Radial Keratotomy (PERK) study was a nine-center clinical trial which analyzed the long-term (10-year) effects and stability of myopic RK (with a range of 2.00 to 8.75 diopters (D)). They showed that 53% of eyes achieved an uncorrected visual acuity (UCVA) 20/20, and 85% of eyes achieved UCVA of 20/40. They also showed that 38% of eyes had a refractive error within 0.50 D and 60% 1.0 D of the intended correction. A common and challenging side effect of RK was the development of secondary and progressive hyperopia. This hyperopic shift occurred in 43% of reported cases, with an additional incidence of 1–2% annually. A less common side effect following RK is the development of irregular astigmatism, which can be induced by the intersection of the incisions with the visual axis or by the eccentricity of the optical zone. Some other side effects are fluctuating vision and glare. Apart from treating myopia, RK has also been used for the correction of astigmatism (also known as arcuate keratotomy), although the predictability of this technique is known to be slightly less than that for the correction of myopia. The procedure has been shown to be an effective and safe method for correcting moderate to severe naturally occurring astigmatism. The popularity of RK has declined since the approval of the excimer laser in 1995, due to the superior outcomes of photorefractive keratectomy (PRK) and LASIK. However, keratotomy techniques (arcuate keratotomy and limbal relaxing incisions) are still used for the treatment of astigmatism in cataract surgery and in postsurgical patients.

Photorefractive Keratectomy In the early 1990s, PRK was the main treatment for low-to-moderate myopia. This technique went through various developments, varying from laser systems, to treatment algorithms, to the choice of transition and ablation zones. The treatment involves the use of a far-ultraviolet (193-nm) argon fluoride excimer laser, which permanently removes the most anterior portion of the corneal stromal tissue in a very precise manner. The ablation occurs with minimal damage to the adjacent corneal tissue. Prior to the performance of the laser-ablation procedure, the corneal epithelium is removed – either manually with a blade or a rotating brush or after alcohol administration. Afterward, a bandage contact lens is applied on the treated corneal surface.

Short-term problems following PRK include discomfort in the first 24 h; a delay in visual recovery lasting 3–5 days during epithelial healing; and a loss of corneal transparency – also called haze – lasting weeks to months following the procedure. PRK ablations often show an immediate postoperative hyperopic shift, due to a thinner epithelium. The hyperopic shift is often compensated by a period of regression that stabilizes between 1 and 6 months. Refractive stability after PRK is generally achieved after 6 months to 1 year and is maintained for up to a period of 5 years. Long-term studies on the outcome of PRK found no evidence of progressive time-dependent hyperopic shift or late regression, with trace haze in 4% after 12 years with no loss of best-corrected visual acuity (BCVA). In general, corneal haze was transient and decreases rapidly 1 year after treatment.

Laser-assisted in situ Keratomileusis The technique of LASIK was first described in 1991. The surgical technique includes the creation of an epithelialstromal flap using a microkeratome. The flap is attached to the periphery of the cornea by a hinge of uncut tissue and has a diameter of 8–10 mm. When using the mechanical microkeratome, the thickness of the flap ranges between 130 mm (with the newest microkeratomes) and 180 mm (with the older microkeratomes). Subsequently the flap is peeled back and ablation of the corneal stroma is performed using an excimer laser. Following the photoablation, the flap is repositioned on the treated corneal stroma (Figure 1). The introduction of this technique meant a major change in the field of refractive surgery. The side effects associated with PRK made LASIK treatment the leading procedure in refractive surgery. The popularity of LASIK is related to the relatively fast visual recovery time, minimal discomfort immediately following treatment, and the minimal incidence of haze. For low-to-moderate myopia (less than 6 D), LASIK has proven to be very effective, predictable, and safe – achieving an UCVA of 20/40 or more in 86–100% of eyes and an UCVA of 20/20 in 45–94% of eyes. The technique has shown to achieve a very accurate correction, with 71–96% of eyes achieving a refractive error within 0.50 D of the intended correction and 88–100% of eyes within 1.00 D of the intended correction. For moderate-to-high myopia (> 6.0 D), the results show more variation. Since PRK and LASIK candidates typically have healthy eyes, achieving and maintaining high levels of (subjective) satisfaction after surgery are very important. In 2005, a clinical study showed that the overall patient satisfaction

Refractive Surgery and Inlays

203

Figure 1 Schematic LASIK-procedure: first the creation of the flap after which the flap is peeled back, then eximer laser ablation of the corneal stroma, and third repositioning of the flap on the treated corneal stroma.

following LASIK treatment was 4.10 0.71 (a score of 5 meaning that the patient was totally satisfied). Patients are generally very satisfied with their uncorrected vision, visual recovery, and quality of life following LASIK treatment, with the majority of patients reporting that they would have the surgery again (92.3%), if required. Myopic regression is a condition that can occur after LASIK. The risk of myopic regression increases with the degree of preoperative myopia and patient age. Long-term studies on LASIK for treatment of moderate and extreme myopia showed a trend toward myopic regression, changing from 52–96% of eyes within 1.0 D of the attempted correction after 1 year to 46–91% after 5–6-years followup. In contrast, LASIK studies with lower degrees of preoperative myopia (< 6.0 D) show stable visual results during long-term follow-up. Although the risks associated with LASIK are considered to be low, intraoperative and postoperative flap-related complications are sight threatening and have resulted in a permanent loss of BCVA. The overall incidence of intraoperative LASIK-flap complications – such as incomplete flaps, buttonholes, free caps, and torn flaps – is approximately 4%. Postoperative flap-related complications include diffuse lamellar keratitis, infection, spontaneously or trauma-related flap displacement, and epithelial ingrowth. Furthermore, LASIK may cause (transient) dry eyes, which may be related to the neurotrophic effects of cutting the nerves during the creation of the flap. It tends to resolve 6–9 months following LASIK treatment, as the nerves grow back into the flap. Despite the aim of many surgeons to keep the residual corneal thickness of the stromal bed at least 250 mm, postoperative corneal ectasia (dilation) may occur following LASIK treatment. This rare, but important, complication seems to be related to biomechanical changes in the cornea after treatment and occurs at rates much lower than 1%. Risk factors that might contribute to the development of ectasia following LASIK have been suggested to be: high intraocular pressure, irregular topography, thin corneas, thin remaining corneal beds, forme fruste keratoconus, thick corneal flaps, large optical zones, and, possibly, high myopia.

Laser-Assisted Subepithelial Keratectomy LASEK aims to preserve the original anatomy of the cornea and to avoid potential risks posed by the creation of a LASIK flap. The treatment is, in fact, a blend of PRK and LASIK, aiming to decrease the potential complications of the two treatments. In LASEK treatment, diluted ethanol solution is applied to loosen the corneal epithelium, following which the epithelium is partially removed from Bowman’s layer, leaving it connected only at a hinge. Laser treatment is applied directly to Bowman’s layer, and afterwards the epithelial sheet is placed back over the treated stroma. The eye is covered by a bandage contact lens to prevent movement of the epithelial flap due to blinking and eye movements. LASEK does not have the risk of flap-related complications such as with LASIK, because LASEK can easily be converted to the PRK procedure if the epithelial flap tears or breaks. Furthermore, a larger residual bed is created – which retains the cornea’s biomechanical strength and reduces the risk of corneal ectasia associated with LASIK. One of the main therapeutic advantages of LASEK is that it can be performed in cases in which LASIK may be contraindicated. These include eyes with thin, steep, and flat corneas; epithelial basement-membrane dystrophy; large pupils (requiring wider and, therefore, deeper ablations); higher myopia; and deep-set eyes or tight orbits. Reports have shown that LASEK is a safe, effective, and predictable treatment, which can be seen as a good alternative to LASIK and PRK for the surgical correction of myopia. A major review of the literature showed that 95% of eyes achieved an UCVA 20/40 and 74% 20/20. Seventy-four percent of eyes achieved a refraction within 0.5 D of the desired refraction and 90% of eyes were within 1.0 D of the desired refraction. Loss of two or more lines of BCVA was demonstrated in 2% of eyes. Comparing LASEK with PRK and LASIK, it has been indicated that the recovery period following LASEK is shorter than that following PRK, but might be somewhat slower than that following LASIK. Discomfort following LASEK seems to be less than after PRK, which is probably

204

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

related to the fact that the epithelial flap acts as a biological therapeutic lens that protects the ablated stroma. However, other studies have demonstrated that the epithelial flap is probably not viable, is replaced by regenerated epithelial cells, and, as such, does not provide advantages in comparison to PRK. The biological properties of the epithelium might inhibit haze formation following the LASEK treatment. Furthermore, the procedure can be combined with the use of mitomycin-C (MMC) – a cytostatic drug known to inhibit proliferating cells and can be used to prevent postablation corneal haze in high-risk cases, such as in patients requiring retreatment.

Epithelial Laser in situ Keratomileusis (epi-LASIK) Epi-LASIK uses a modified microkeratome (epikeratome) to create a thin corneal epithelial flap before surface ablation is performed. The blade and the angle of cutting of the epikeratome are optimized for a subepithelial dissection, which does not disrupt the corneal stroma such as the LASIK microkeratome. The difference between epiLASIK and LASEK is that the separation of the epithelium is obtained mechanically without the use of alcohol. Epi-LASIK has been proposed as a safe alternative to LASIK and is especially suitable for patients with low-tomoderate myopia and myopic astigmatism, thin corneas, and in individuals with steeper or flatter corneas, where the cutting of a LASIK flap could potentially impose flaprelated complications. The healing period and visual recovery tends to be slower than traditional LASIK. Postoperative discomfort usually occurs within the first 48 h after surgery. A recent study presented the 1-year results of epiLASIK and stated that this technique is a safe and efficient method for the correction of low and moderate myopia, demonstrating that all of the eyes treated reached an UCVA of 20/40 or better and 86% an UCVA of 20/20 or better. More than 80% of eyes were within 0.5 D of the attempted correction and 97% were within 1 D of the attempted correction. In comparison with LASEK, it has been suggested that the incidence of haze following ablation of the cornea is lower with epi-LASIK. Longer-term clinical studies are needed to confirm the reported results on epi-LASIK.

Femtosecond Laser in situ Keratomileusis (FS-LASIK) At the beginning of this century the femtosecond laser, which creates a corneal flap by vaporizing small volumes of tissue using micro-photodisruption at a predetermined

depth, was introduced. Furthermore, when using the femtosecond laser, the flap thickness is much more accurate and thinner (between 90 and 110 mm) than with the excimer laser. The corneal flaps can also be customized with a variable flap thickness and diameter based on the requirements of the patient. The greatest benefit of these thinner femtosecondcreated flaps is that they result in a greater stability of the cornea when compared to mechanically created flaps. It has been described that the greatest strength of the cornea lies within the first 150 mm of the cornea. Thus, thinner flaps will help to protect the integrity of the cornea and will lead to thicker residual stromal beds, which results in a decreased risk of corneal ectasia. Studies have shown that flaps created with a femtosecond laser provide better visual results than flaps created with a microkeratome. Furthermore, FS-LASIK demonstrates better visual outcomes than PRK in the first 6 months of follow-up. Following this period, both treatments show similar visual results. As for epi-LASIK, long-term followup studies are required to demonstrate the efficacy and safety of FS-LASIK. Potential limitations of the femtosecond laser include increased costs of the procedure, increased surgical time, and a higher incidence of diffuse lamellar keratitis. The use of topical corticosteroids keeps the last complication at manageable levels. Furthermore, the newer highfrequency femtosecond lasers will probably diminish the incidence of diffuse lamellar keratitis.

Treatment of Hyperopia Photorefractive keratectomy, LASIK, and LASEK can also be applied for the treatment of low hyperopia (<2 D). In hyperopic treatments, most of the laser ablation is located at the periphery of the treatment zone. Mechanical weakening of the peripheral cornea might lead to a forward-bowing of the central cornea, which increases the intended laser effect. These biomechanical changes and a different wound healing cause increased levels of regression following the hyperopic treatment. Therefore, only low levels of hyperopia can be treated using laser refractive surgery. The Food and Drug Administration (FDA)-approved studies investigating the treatment of low hyperopia using LASIK showed that 90% of eyes achieved a UCVA 20/40 and only 63% a UCVA 20/20. Sixty-seven percent of eyes achieved a refraction within 0.5 D of the desired refraction and 90% of eyes were within 1.0 D of the desired refraction. Loss of two or more lines of BCVA was demonstrated in almost 2% of eyes after more than 3 months of follow-up. Hyperopic PRK and LASEK treatments show similar clinical results when compared to hyperopic LASIK treatments.

Refractive Surgery and Inlays

General Side Effects of Laser Refractive Surgery Although many developments in keratorefractive surgical techniques have improved the clinical outcome and have shown great success rates, several quality-of-vision problems have been reported. Qualitative visual disturbances can affect patients’ daily activities and include subjective complaints such as glare, halos, and difficulty with night driving. These complaints are more likely to occur with laser corrections of more than 7–8 D of myopia or more than 2–3 D of hyperopia and often diminish after the first six postoperative months. Glare, halos, and night-vision complaints may be attributed to a loss of contrast sensitivity or low-contrast visual acuity. These complaints have been described after all refractive surgery techniques, varying in degrees of incidence. Reports on patient satisfaction following LASIK treatment showed that predictors for night-vision complaints can include: . . . . . .

preoperative high levels of myopia (more than 5 D), advanced age, a flatter preoperative corneal curvature, surgical enhancements, optical zones <6 mm, postoperative residual refractive error >0.5 D from emmetropia (normal refraction), and . postoperative residual cylinder – a type of higher-order aberration (HOA) of the cornea. Remarkably, pupil size was not shown to be a significant predictor of night-vision complaints in any of these studies. There is variable evidence in the literature on excluding patients based on large pupil size. It has been suggested in the past that a large pupil, in combination with a small optical zone, is a dominant factor leading to increased night-vision complaints. However, other recent studies demonstrate that the correlation between pupil size and night-vision complaints or between night-vision complaints and the pupil– optical zone disparity is much less critical than previously thought. Pupil size seems to indeed be a significant predictor of glare and halos following LASIK, especially in the first postoperative month, yet it was demonstrated that pupil size is not a significant variable 6 or 12 months following treatment. Postoperative remodeling of the corneal shape by the epithelium may be responsible for these findings. Other common complications of laser refractive surgery include under- and overcorrection (30%), irregular astigmatism (30%), and dry eyes (4–30%, depending on the type of refractive treatment).

Wave front and Laser Refractive Surgery When applying conventional laser refractive surgery, the ablations are calculated using the data obtained during

205

manifest and cycloplegic refractions. However, HOAs can cause glare and halos and lead to decreased quality of vision. Wave front technology was developed to categorize and limit the amount of HOA induced by refractive surgery. The wave front sensor measures defocus, astigmatism, and total and individual HOA. Customized wave front-guided corneal ablation combines wave front sensing and wave front correction and, therefore, corrects refractive errors beyond spherical and cylindrical errors. Over the last decade, several clinical reports have studied HOAs in refractive surgery patients. Some of these studies have shown an increase in patient satisfaction, reduced night-vision complaints, and a lower increase of HOA following wave front-guided treatments, compared to conventional ablation; however, more and larger randomized studies are needed to further analyze and validate the results of these treatments. At the present time, it is not clear whether the excellent results are due to an improved postoperative asphericity profile or the consequence of treating the preexistent HOAs.

Corneal Inlays The use of corneal inlays as a refractive procedure involves the insertion of a synthetic or biological material into the cornea, which changes the refractive power of the eye by either altering the anterior corneal curvature, or the refractive index of the inlay material, or a combination of these two mechanisms. These inlays can be placed in the stroma or beneath the epithelium of the cornea. The main advantage of this technique over the above-mentioned laser refractive treatments is the reversibility of the procedure. The inlay material has to meet various physical and biological characteristics in order to minimize complications. Biological materials have shown to be biodegradable and insufficiently permeable to maintain a healthy cornea. Furthermore, hydrogel inlays have been demonstrated to be biocompatible but have insufficient porosity to maintain optimal nutrient flow. Many complications, including corneal haze, epithelial thinning, inlay encapsulation, epithelial opacification, corneal vascularisation, inlay decentration, and fibrosis, have been reported following the implantation of corneal inlays. At present, a phase 1 trial is being conducted using a synthetic corneal inlay made of a polymer of perfluoropolyether, which is placed into the stroma following the creation of a corneal flap with a microkeratome. The study involves implantation of the inlays in unsighted eyes. Despite the experimental phase of these synthetic inlays, it is believed that they might meet all the required physical and biological characteristics which will help minimize the above-mentioned complications. Corneal inlays have been developed not only to correct myopia and hyperopia, but also to correct presbyopia.

206

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

diminish the risks of adverse effects. LASIK has the advantage of rapid visual recovery and minimal discomfort and, therefore, has become the most widely performed refractive surgery technique, associated with high patient satisfaction. However, mechanical microkeratome-assisted LASIK displays more complications due to the creation of a flap and might cause more induced HOAs, which is why some surgeons have switched to PRK or LASEK. In the near future, FS-LASIK might become the recommended treatment, but a longer follow-up is needed to ensure longterm stability and safety.

Figure 2 Patient with two intracorneal ring segments (Intacs).

These presbyopic inlays are designed to be implanted in the nondominant eye and have a pinhole optic, which increases the depth of focus. They are currently under investigation in the United States and Europe.

Intracorneal Rings (ICR) Since 1999 the use of ICR segments, also known as Intacs, has been approved by the United States FDA for the correction of low myopia (less than 4 D) and low astigmatism (1 D). These ring segments are curved and made of polymethylmethacrylate. They can be inserted into the peripheral cornea through corneal channels, which are created by a microkeratome or by a femtosecond laser, at about 70% of the corneal depth (Figure 2). They aim to expand and flatten the corneal surface. ICR have some important advantages over laser refractive surgery: there is no abrasion of the cornea, less risk of corneal haze or scarring, the visual axis is spared from treatment, there is no flap creation and, therefore, no flaprelated complications, and the procedure is reversible. Complications following ICR implantation include induced astigmatism – which is the most common complication – extrusion of the ring, dry eye syndrome, and infectious keratitis. In the treatment of low myopia, the visual results of ICRs seem to be similar to the results following laser refractive surgery. However, ICRs are mainly used to improve visual outcomes in patients with keratoconus, post-LASIK ectasia, or other ectatic disorders.

Conclusion In conclusion, for the treatment of patients with low-tomoderate myopia, PRK, LASIK, and LASEK have all shown to produce stable and predictable results with an excellent safety profile. Patient selection is crucial to ensure high postoperative patient satisfaction and to

See also: Cornea Overview; Hyperopia; Myopia; Refractive Surgery; The Surgical Treatment for Corneal Epithelial Stem Cell Deficiency, Corneal Epithelial Defect, and Peripheral Corneal Ulcer.

Further Reading Alio, J. L., Muftuoglu, O., Ortiz, D., et al. (2008). Ten-year follow-up of laser in situ keratomileusis for myopia of up to –10 diopters. American Journal of Ophthalmology 145: 46–54. Choi, D. M., Thompson, R. W., and Price, F. W. (2002). Incisional refractive surgery. Current Opinion in Ophthalmology 13: 237–241. Fan-Paul, N. I., Li, J., Miller, J. S., and Florakis, G. J. (2002). Night vision disturbances after corneal refractive surgery. Survey of Ophthalmology 47: 533–546. Guell, J. L. (2005). Are intracorneal rings still useful in refractive surgery? Current Opinion in Ophthalmology 16: 260–265. Katsanevaki, V. J., Kalyvianaki, M. I., Kavroulaki, D. S., and Pallikaris, I. G. (2007). One-year clinical results after epi-lasik for myopia. Ophthalmology 114: 1111–1117. Nuijts, R. M., Nabar, V. A., Hament, W. J., and Eggink, F. A. (2002). Wavefront-guided versus standard laser in situ keratomileusis to correct low to moderate myopia. Journal of Cataract and Refractive Surgery 28: 1907–1913. Pop, M. and Payette, Y. (2004). Risk factors for night vision complaints after LASIK for myopia. Ophthalmology 111: 3–10. Rajan, M. S., Jaycock, P., O’Brart, D., Nystrom, H. H., and Marshall, J. (2004). A long-term study of photorefractive keratectomy; 12-year follow-up. Ophthalmology 111: 1813–1824. Sakimoto, T., Rosenblatt, M. I., and Azar, D. T. (2006). Laser eye surgery for refractive errors. Lancet 367: 1432–1447. Slade, S. G. (2008). Thin-flap laser-assisted in situ keratomileusis. Current Opinion in Ophthalmology 19: 325–329. Sugar, A., Rapuano, C. J., Culbertson, W. W., et al. (2002). Laser in situ keratomileusis for myopia and astigmatism: Safety and efficacy: A report by the American Academy of Ophthalmology. Ophthalmology 109: 175–187. Sweeney, D. F., Vannas, A., Hughes, T. C., et al. (2008). Synthetic corneal inlays. Clinical and Experimental Optometry 91: 56–66. Tahzib, N. G., Bootsma, S. J., Eggink, F. A., Nabar, V. A., and Nuijts, R. M. (2005). Functional outcomes and patient satisfaction after laser in situ keratomileusis for correction of myopia. Journal of Cataract and Refractive Surgery 31: 1943–1951. Taneri, S., Zieske, J. D., and Azar, D. T. (2004). Evolution, techniques, clinical outcomes, and pathophysiology of LASEK: Review of the literature. Survey of Ophthalmology 49: 576–602. Waring, G. O., Lynn, M. J., and McDonnell, P. J. (1994). Results of the prospective evaluation of radial keratotomy (PERK) study 10 years after surgery. Archives of Ophthalmology 112: 1298–1308.

Relevant Website http://www.intacsforkeratoconus.com – Intacs Corneal Implants.

Contact Lenses N Carnt and Y Wu, Institute for Eye Research, Sydney, NSW, Australia F Stapleton, University of New South Wales, Sydney, NSW, Australia ã 2010 Elsevier Ltd. All rights reserved.

Glossary Contact lens-induced papillary conjunctivitis (CLPC) – Enlarged papillae accompanied by redness of the upper palpebral conjunctiva. Symptoms include discomfort, itching, mucous discharge, and foreign body sensation. Corneal infiltrates – Corneal inflammatory response which appears as single or multiple, focal or diffuse lesions in the corneal stroma. Endothelial polymegethism – Variation in the size of the endothelial cells of the cornea as a result of disturbed metabolism. HEMA – Hydroxyethylmethacrylate monomer forming the basis of many soft contact lens material polymers. Hypermetropia – Also known as farsightedness, it is a defect of vision caused by an imperfection in the eye (often when the eyeball is too short or when the lens cannot become round enough), causing difficulty focusing on near objects. Lens-induced chronic hypoxia – Corneal oxygen deprivation that results from contact lenses that greatly impede oxygen diffusion to the cornea. Chronic hypoxia may cause changes in corneal structure and neovascularization. Multipurpose solutions (MPSs) – Solution for cleaning, rinsing, disinfecting, and storing your contact lenses. Myopia – The ability to see close objects more clearly than distant objects. Myopia can be caused by a longer-than-normal eyeball or by any condition that prevents light rays from focusing on the retina. Orthokeratology – The practice of reshaping the cornea by wearing specially designed rigid contact lenses. These lenses are usually worn only during sleep and they reshape the front surface of the eye, correcting the refractive error and allowing clear vision. PMMA – Polymethylmethacrylate material used for oxygen impermeable hard contact lenses. RGP – Rigid gas permeable materials used for oxygen permeable hard contact lenses. Superior epithelial arcuate lesions (SEALs) – Whitish, opalescent lesions, located adjacent or at the superior limbus, caused predominantly by a mechanical interaction between the (usually) silicone hydrogel or hydrogel contact lens, the eyelid, and the superior peripheral cornea.

Introduction Contact lenses provide a safe and reversible means of refractive error correction; lenses are currently worn by 150 million individuals worldwide. Contact lenses have many optical, sporting, cosmetic, therapeutic, and vocational advantages over conventional spectacles. The vast majority of wearers use contact lenses for the correction of simple low refractive errors but other applications include presbyopia, astigmatism, high refractive error, and other medical, cosmetic, or therapeutic indications. Given their close association with the ocular surface, contact lens wear may have an impact on ocular surface anatomy and physiology and they may be associated with adverse effects, including ocular surface infection or inflammation. The impact of contact lens wear varies with lens type, mode of wear, indication for lens wear, and with wearer factors including demographics, hygiene, compliance, and behavior factors.

Types of Contact Lenses Hard lenses, made from polymethylmethacrylate (PMMA), were first available in 1934 and soft lenses during the 1960s. Lens materials have evolved since with the advent of gas permeable rigid materials and more recently silicone hydrogel materials. An ideal material and lens design would maintain normal ocular surface physiology, have a low rate of complications, provide vision correction, and be comfortable during wear (Table 1 and Figures 1–6). Probably the most exciting recent innovation in the field has been the introduction of silicone hydrogel materials. The high oxygen permeability of these new siliconebased products has limited the effects of hypoxia and largely eliminated a range of physiological complications such as corneal edema, microcysts, corneal vascularization, endothelial polymegethism, overwear syndrome, and corneal exhaustion. Lenses may be worn on a daily wear basis, where they are removed, cleaned and disinfected overnight, and reinserted the following day. Daily lens use is the most common modality prescribed. Complication rates are generally lowest with this wear modality; however, conditions associated with toxic or allergic responses to lens care products occur more commonly with daily lens use. Daily disposable contact lenses (lenses worn once during the day and discarded on removal) have eliminated many

207

208 Table 1

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health Contact lens types and materials

CL type

Material

Comments

Illustration

Rigid corneal contact lenses

PMMA: polymethylmethacrylate

. Readily machined and polished . Impermeable to oxygen . Average lens diameter range 7–12 mm

Figure 1

RGP: rigid gas permeable (fluorosiloxane acrylates, perfluoroethers)

. Higher oxygen permeability compared with PMMA . Average corneal lens diameter range 9–12 mm

Figure 1

Orthokeratology lenses: generally highly oxygen permeable rigid gas permeable materials

. Overnight use of orthokeratology lenses reshapes the cornea to reduce the refractive error with overnight use . Larger diameter lenses with an oblate profile used to reduce myopic refractive errors

Figure 2

Hydrogel: hydroxyethylmethacrylate (HEMA), with other monomers, such as vinyl pyrollidone, vinyl chloride, methacrylic acid and others, added to alter ionicity and water content of the material

. Flexible polymers which conform to the shape of the cornea and limbus . Average lens diameter 13–15 mm . Water content of polymer determines oxygen permeability

Figure 3

Silicone hydrogel: silicone incorporated into a diverse group of hydrogel monomers. Wettability is enhanced by either surface treatment or by the incorporation of soluble polymers within the bulk material

. Higher oxygen permeability than hydrogel contact lenses, consequently less corneal hypoxia . Lens design similar to hydrogel contact lenses . Generally stiffer modulus than hydrogel contact lenses

Figure 3

Hybrid

. An RGP center with a soft hydrogel or silicone hydrogel periphery . Useful in fitting irregular corneas

Figure 4

. Scleral lenses use the sclera rather than the cornea as the bearing surface . Consequently a range of abnormal corneal topographies may be fitted . The overall diameter ranges from 16 mm (miniscleral) to 24 mm (full scleral) and lenses are made in RGP materials

Figures 5 and 6

Soft contact lenses

Other contact lenses

Hard/soft combination Scleral/semiscleral/mini-scleral lenses

Adapted from Fonn, D., Reyes, M., Terry R., and Williams L. (2000) The IACLE Contact Lens Course, 1st edn., vols. 2 and 8. Sydney: The International Association of Contact Lens Educators.

Figure 1 Rigid contact lens of 9.5 mm diameter.

Figure 2 Orthokeratology contact lens.

Contact Lenses

209

Figure 3 Hydrogel lens of 14 mm diameter. Figure 6 Full scleral contact lens on eye.

complications related to care solution use and those due to inadequate contact lens or storage case hygiene. A proportion of daily lens users (13–23%) report occasional unscheduled overnight use of lenses of up to one night per week. Certain lenses may be prescribed for overnight use, where they are worn continuously for up to 30 consecutive nights.

How Contact Lens Wear Affects the Ocular Surface

Figure 4 Hybrid lens on eye.

Hydrogel and silicone hydrogel contact lenses are on average 2–3 mm larger than the cornea and overlay the limbus and surrounding bulbar conjunctiva, consequently contact lenses interact directly with the corneal, limbal, bulbar, and tarsal conjunctival epithelia, and with the tear film. Their impact on the ocular surface can be characterized as follows. Effects of Contact Lens Wear on the Corneal and Limbal Epithelium

Figure 5 Rigid lenses from left to right: corneal RGP; orthokeratology RGP, mini-scleral, full scleral.

Lens-induced chronic hypoxia can lead to changes in the corneal epithelium such as thinning, lower oxygen uptake rates, and increased microcysts. In addition, limbal hyperemia is increased and encroachment of blood vessels into the corneal stroma can occur. The smaller diameter of rigid lenses results in less interaction between the lens and limbal or conjunctival epithelium, generally less corneal hypoxia and greater tear exchange compared with soft contact lenses. Epithelial microcysts are one of the more obvious clinical markers of hypoxia, appearing as small, irregular dots in retro illumination of the cornea with a slit lamp biomicroscope at high magnification. They are comprised of degenerated cellular material and are most often seen after 3 months of extended wear of low oxygen transmissible

210

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

lenses. They are transitory, exhibiting a spike in numbers approximately 1 month after hypoxic stress is removed and returning to normal within 3 months. All contact lenses slow the corneal epithelial cell renewal cycle by slowing epithelial cell proliferation and migration and lowering the exfoliation rate. At the limbus, stem cells continually provide new basal corneal epithelial cells, where the rate of production increases during ocular surface damage. Recurrent erosion, chronic inflammation, and blood vessel growth onto the cornea (vascularization) can occur as a result of compromise to the limbal stem cells. Basal epithelial cells produced from slow cycling stem cells replicate and migrate horizontally, differentiating upward, before undergoing either apoptosis (programmed cell death) or necrosis and exfoliation into the precorneal tear film. Soft contact lens wear, in particular, lens-related hypoxia, has been reported as a possible cause of limbal stem cell damage. To a lesser extent contact lenses may have a mechanical effect on the limbal stem cells, which are protected somewhat due to their location at the level of the basal epithelium. Alteration of the Lens: Cornea Resurfacing Mechanism The precorneal tear film maintains the normal optical, lubrication, and defence functions of the ocular surface. Wear of a contact lens may destabilize the tear film by compartmentalizing the tear film into prelens and postlens regions, by changing the tear film structure and components and by reducing tear exchange. Compared with the normal precorneal tear film, the prelens tear film is thinner, less stable, and exhibits higher evaporation. The increased evaporation rate is thought to be related to a thinner lipid layer which is more prone to contamination. A thinner lipid layer, lens dehydration, and an unstable tear film can lead to corneal epithelial dessication. All contact lenses inhibit normal tear exchange by trapping debris, toxins, antigens, and microorganisms beneath the lens and interfering with the normal resurfacing action of blinking. Increasing the retention time of metabolic debris, toxins, or antigens at the ocular surface has been implicated in the development of inflammatory conditions. The rate of tear exchange varies with lens type, material and wearer characteristics. For example, rigid gas permeable (RGP) contact lenses allow more rapid tear exchange than hydrogel or silicone hydrogel lenses and are associated with a lower rate of inflammatory complications compared to soft lens wear. Adverse Events Inflammation

Corneal inflammation occurs relatively often in contact lens wear, affecting 7% of wearers. The risk of inflammation

varies with different lens materials and wear modalities. The severity of inflammation varies and several systems have been used to characterize and classify corneal inflammation. Contact lens-related corneal inflammation is characterized by white blood cells, usually polymorphonuclear leukocytes, invading the cornea stroma (infiltrates). The infiltrates appear as small, usually round, opaque lesions (focal) or dispersed in a haze (diffuse). Usually limbal and bulbar redness accompanies the infiltrates. Epithelial damage can be present. Inflammation can be as a result of mechanical trauma but more often results from toxic reactions to bacteria. In the closed eye environment, bacterial toxin concentration behind the lens as well as compromise to the normal lid-tear resurfacing and alteration of tear defence proteins may lead to more frequent inflammation in extended compared to daily wear. Contact lens-induced papillary conjunctivitis (CLPC) is an inflammatory reaction of the upper palpebral conjunctiva, caused by mechanical trauma and/or allergic response to the lens or deposits on the surface of the lens. It is the most common contact lens-related adverse event that leads to discontinuation of lens wear. Clinically, it presents as raised papillae, which comprise lymphocytes and plasma surrounding central blood vessels and associated hyperemia. There are two forms of CLPC, one that involves the entire upper palpebral area (generalized) and the other a small cluster (localized). CLPC is more frequent in soft than rigid lenses, and more common in extended compared to daily wear. The generalized form is more common in softer, hydrogel lenses which attract more protein deposits; this form likely represents a hypersensitivity (both type I and type IV) response, whereas localized disease is more likely to be found with the stiffer silicone hydrogel lenses and is thought to be related to mechanical trauma. Infection Corneal infection is a rare but severe complication of contact lens wear, affecting 4 per 10, 000 wearers per year, with higher rates (20 per 10, 000) per year in overnight lens use. These estimates have remained remarkably consistent over a 20-year period, despite advances in contact lens materials, wear modalities, and care system technologies. Permanent vision loss occurs in 6 per 100, 000 wearers per year or in 10–15% of cases of infection. Lower rates of severe keratitis and vision loss have been reported with daily disposable contact lens use. A poorer disease outcome is strongly associated with recovery of an environmental pathogen and a delay in receiving appropriate treatment. Unlike corneal infections in noncontact lens wearers, Pseudomonas aeruginosa is the most common pathogen associated with contact lens-related disease, accounting for 50–60% of culture proven cases. More recently, environmental organisms such as Acanthamoeba and the fungi

Contact Lenses

Fusarium have been associated with outbreaks of disease, particularly disease linked with the use of specific multipurpose solutions (MPSs). In the case of Acanthamoeba keratitis, strong links with contaminated water have been identified. Mechanical effects

Contact lenses can cause mechanical effects on the cornea due to their physical presence. For soft contact lenses, superior epithelial arcuate corneal lesions (SEALs), white heaped lesions occurring adjacent to and concentric with the superior limbus are thought to be due to a combination of tight upper-lid pressure, stiffness of the lens material, poor surface, and tight fitting of soft lenses. These are usually associated with foreign body and discomfort symptoms and resolve with interruption to lens wear within 1 week. Erosions, a full thickness epithelial loss, appear as a plug of epithelium that has been lifted off the cornea. Resolution is rapid and often results in clumping of the epithelial cells as healing takes place. Care must be taken to observe good hygiene and not patch the lesion, as the break in the epithelium renders the cornea more prone to infection. Sometimes prophylactic antibiotic therapy is used. Symptoms can be quite painful and can be eased by use of ocular lubricants. For RGP lenses, the most common mechanical effect is abrasion due to debris trapped behind the contact lens. Corneal abrasions are more commonly seen with rigid lenses compared with soft. Often the patient is symptomatic despite quite spectacular appearances; resolution is rapid, usually within a day or two. Abrasions can also occur due to fingernail injuries on removal of lenses.

211

risk of low-grade inflammatory events. MPSs have been associated with decreased epithelial barrier function and increased susceptibility to bacterial binding; however, no evidence to date shows that solution-related staining is a risk factor for infection. Further exploration of the mechanisms and implications of short- and long-term corneal staining is underway. Corneal Infection Other than overnight wear of contact lenses, a range of hygiene, compliance, and demographic factors have been consistently associated with disease. In contemporary lens wearers these include increasing the degree of lens use in daily wear, poor-storage case hygiene, smoking, Internet purchase, less than 6 months wear experience, and socioeconomic class. Further study is directed toward exploration of new risk factors, particularly establishing the impact of wearer risk behavior such as internet purchase: evaluation of specific hygiene practice such as rub and rinse and establishing the impact of second generation silicone hydrogel lens materials on the risk of infection. Ongoing surveillance enables monitoring of disease outbreaks and new or unusual causative organisms. In corneal infection associated particularly with daily wear of contact lenses, the storage case has been identified as the source of organisms. This has resulted in the uptake of antimicrobial technologies, such as silver, into contact lens storage cases to limit microbial adhesion and colonization of the contact lens storage case. Early reports indicate some reduction in levels of certain microorganisms, but the long-term impact of these technologies in clinical use has not yet been established.

Current Hot Topics in Contact Lens Research

Orthokeratology

Contact Lens Care Products

Orthokeratology is a nonsurgical approach to temporarily correct or eliminate refractive error through rigid lens wear to reversibly change the curvature of the cornea. More recent approaches include using specially designed rigid contact lenses worn overnight to enable clear vision without lens wear during the daytime. In the case of myopia correction, a reverse-geometry lens with a base curve flatter than the peripheral curve of the back surface is worn. The reduction in myopia occurs through central corneal epithelial thinning. While these effects are reversible there is discussion whether the impact of orthokeratology lenses on the corneal epithelium increases the risk of corneal infection. Ongoing studies endeavor to optimize the refractive effects both in myopia and hypermetropia, to understand the mechanisms in epithelial thinning and to establish whether orthokeratology can be used to permanently reduce or eliminate refractive error or to prevent myopia progression.

MPSs dominate the CL care market and are used by 99% of soft lens wearers. In recent years, the industry has been focused on comfort and convenience for the wearer, formulating new preservatives and incorporating comfort additives into care products. Corneal staining (visualization of damaged epithelial cells by instillation of sodium fluorescein) is the most reliable clinical way to assess corneal integrity. Solutionrelated corneal staining has been reported from a 2-h provocative test and over 3 months of lens wear with a range of lens and solution combinations. Performance has not been predictable in terms of solution preservative or across different silicone hydrogel lens types, likely due to the more complicated lens chemistry compared to hydrogel lenses. Solution-related corneal staining has been associated with decreased comfort and increased

212

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Summary A range of materials and wear modalities are available in contemporary contact lens practice to correct a wide range of refractive errors, potentially to slow or prevent refractive error development and to meet other medical or cosmetic requirements. Contact lenses interact closely with the ocular surface and all lenses, to a lesser or greater degree, will impact on the conjunctiva, lid, cornea, and tear film physiology. While the majority of their effects and complications are reversible and nonsevere, corneal infection remains the most severe response which can potentially result in vision loss due to scarring and perforation. Safe contact lens wear requires awareness by wearers to ensure that their eyes ‘‘Feel good, see well and look good’’ at all times. In the rare event of corneal infection, morbidity can be limited by prompt treatment by the practitioner. Recent outbreaks of corneal infection caused by unusual environmental organisms have resulted in the industry demanding more appropriate care solution testing for licensing purposes, and improved guidelines and recommendations for contact lens wearers. Research into the impact of care solutions on the ocular surface, optimization of orthokeratology, eludication of new risk factors for corneal infection, and ongoing disease surveillance will better inform practitioners and wearers and result in improved outcomes for contact lens wearers. See also: Corneal Epithelium: Transport and Permeability; Corneal Epithelium: Response to Infection; Immunopathogenesis of Pseudomonas Keratitis; Inflammation of the Conjunctiva.

Further Reading Carnt, N., Jalbert, I., Stretton, S., Naduvilath, T., and Papas, E. (2007). Solution toxicity in soft contact lens daily wear is associated with corneal inflammation. Optometry and Vision Science 84: 309–315.

Dart, J. K. G., Radford, C. F., Minassian, D., Verma, S., and Stapleton, F. (2008). Risk factors for microbial keratitis with contemporary contact lenses: A case-control study. Ophthalmology 115: 1647–1654. e1643. Fon, D., Reyers, M., Terry, R., and Williams, L. (2000). The IACLE Contact Lens Course, 1st edn., vols. 2 and 8. Sydney: The International Association of Contact Lens Educators. Holden, B. A., Sweeney, D. F., Vannas, A., Nilsson, K. T., and Efron, N. (1985). Effects of long-term extended contact lens wear on the human cornea. Investigative Ophthalmology and Visual Science 26: 1489–1501. Keay, L., Edwards, K., Naduvilath, T., Forde, K., and Stapleton, F. (2006). Factors affecting the morbidity of contact lens related microbial keratitis: A population study. Investigative Ophthalmology and Visual Science 47: 4302–4308. Keay, L., Jalbert, I., Sweeney, D. F., and Holden, B. A. (2001). Microcysts: Clinical significance and differential diagnosis. Optometry (St Louis, MO) 72: 452–460. Li, S. L., Ladage, P. M., Yamamoto, T., et al. (2003). Effects of contact lens care solutions on surface exfoliation and bacterial binding to corneal epithelial cells. Eye Contact Lens 29: 27–30. O’Hare, N., Stapleton, F., Naduvilath, T., et al. (2003). Interaction between the contact lens and the ocular surface in the aetiology of superior epithelial arcuate lesions. Advances in Experimental Medicine and Biology 506: 973–980. Ren, D. H., Petroll, W. M., Jester, J. V., and Cavanagh, H. D. (1999). The effect of rigid gas permeable contact lens wear on proliferation of rabbit corneal and conjunctival epithelial cells. CLAO Journal 25: 136–141. Stapleton, F., Keay, L., Jalbert, I., and Cole, N. (2007). The epidemiology of contact lens related infiltrates. Optometry and Vision Science 84: 257–272. Stapleton, F., Keay, L., Edwards, K., et al. (2008). The incidence of contact lens related microbial keratitis. Ophthalmology 115: 1655–1662. Swarbrick, H. A. (2006). Orthokeratology review and update. Clinical and Experimental Optometry 89: 124–143. Sweeney, D. F., Jalbert, I., Covey, M., et al. (2003). Clinical characterisation of corneal infiltrative events observed with soft contact lens wear. Cornea 22: 435–442. Thai, L. C., Tomlinson, A., and Doane, M. G. (2004). Effect of contact lens materials on tear physiology. Optometry and Vision Science 81: 194–204. Truong, T., Wofford, L., and Lin, M. (2005). Effects of lens-care solutions on corneal epithelial barrier function. Optometry and Vision Science 82: E-Abstract 055009. Vermeltfoort, P. B. J., Hooymans, J. M. M., Busscher, H. J., and van der Mei, H. C. (2008). Bacterial transmission from lens storage cases to contact lenses – effects of lens care solutions and silver impregnation of cases. Journal of Biomedical Materials Research 87B(1): 237–243.

Imaging of the Cornea S C Kaufman, M Fung, D Raja, and N Kramarevsky, University of Minnesota, Minneapolis, MN, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Chemosis – Swelling of the iris due to swelling of the bulbar conjunctiva. Corneal verticillata – Congenital whorl-like opacities in the cornea. DSAEK – Descemet’s Stripping Automated Endothelial Keratoplasty, a form of keratoplasty which involves the removal of the host’s central Descemet’s membrane and corneal endothelium, which is subsequently replaced with a disc shaped graft of donor corneal stroma, Descemet’s membrane and endothelium. Follicles – Lymphoid tissue in the conjunctival stroma. Fuchs’ dystrophy – A common adult-onset form of corneal dystrophy with autosomal dominant inheritance. In this disease, the endothelial cells in the cornea gradually deteriorate. Guttae – Drop-like. Keratic precipitates – Fibrous deposits on the posterior surface of the cornea, usually associated with uveitis. Laser-assisted in situ keratomileusis (LASIK) – Refractive surgery to correct myopia, hyperopia, and astigmatism. Lattice dystrophy – Hereditary corneal dystrophy in which there is an accumulation of amyloid deposits throughout the middle and anterior stroma of the cornea. Lenticle – Relating to the lens. Pappillas – Small projection of tissue commonly triggered by constant irritation of the conjunctiva by contact lenses. Pachymetry – Measurement of the thickness of the cornea. Pigment dispersion syndrome (PDS) – Condition when the iris pigment epithelium and the lens come into contact. This leads to mechanical disruption of the iris resulting in release of pigment granules into the posterior chamber, which follows the flow of aqueous into the anterior chamber angle. The pigment can block the aqueous outflow resulting in elevated intraocular pressure with possible damage to the optic nerve. Posterior polymorphous dystrophy – Disease involving metaplasia and overgrowth of corneal endothelial cells.

Pseudoexfoliation syndrome – Characterized by flakes of granular material at the pupillary margin of the iris and throughout the inner surface of the anterior chamber. It is also associated with secondary open-angle glaucoma. Pseudophakic bullous keratopathy – Corneal edema occurring following cataract extraction. Sclerotic scatter – Biomicroscope illumination that scatters light throughout the cornea. Subluxation – Partial dislocation of an organ.

Slit-Lamp Biomicroscopy The slit-lamp biomicroscope is a versatile device that is the primary diagnostic instrument used during the clinical examination of the cornea and the external structures of the eye and adenexa. It has two primary components mounted on a common axis – the slit illuminator and the biomicroscope. The slit-beam-illumination unit is essentially a projector with a light beam that is adjustable in height, width, direction, and intensity. The examiner can create a sharply focused narrow slit beam which can illuminate and isolate fine detail in otherwise translucent tissue. The biomicroscope component is a binocular Galilean telescope that can produce excellent resolution at multiple magnifications. A headrest stabilizes the patient, and adjustable oculars allow the examiner to focus a stereoscopic image. The biomicroscope and the illuminator are arranged to be parfocal (focus on the same plane) and isocentric (the slit beam is centered in the field of view), which is necessary for its function. This essential setup allows for both direct illumination and – with shifting of the alignment – indirect illumination. During diffuse illumination, the light beam is broadened to its largest aperture and kept at both low magnification and reduced intensity. The illuminator is directed at the eye from an oblique angle and rotated through its arc of travel from side to side. When the light is applied tangentially, surface changes are enhanced and dimensionalized through the effect of creating highlights and shadows. Subtle alterations from normal topography become more clearly noticeable. Direct illumination can highlight conjunctival changes such as chemosis, follicles, and papillae and can further direct the examiner’s exam.

213

214

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Broad-beam illumination can help the examiner visualize opaque lesions that may reflect or absorb light. This technique can be invaluable in evaluating the anterior segment structures from the cornea to the iris. When no abnormalities are seen with this technique, other forms of illumination should be explored. Narrow-beam illumination allows the examiner to evaluate transparent tissues such as the cornea and lens through virtual cross-sections. When used under high magnification, the principal layers of the corneal can be distinguished clearly with great precision. In addition, a narrow beam can be used to identify the presence of anterior chamber cell and flare. Cell and flare is best seen against the background of the dark pupil with an intense but shortened slit light beam. Specular reflections represent the normal light reflexes that bounce off of an ocular surface. These reflections actually correspond to the mirrored reflections from the light source itself. The most important application of specular reflection is in the evaluation of the corneal endothelium. This involves arming the slit beam at an angle of 40–60 from the viewing arm and using a short slit. Then, the specular reflection of the light bulb’s filament is identified on the tear film. Adjacent to this area (on the side away from the light source), the less bright endothelial reflection can be viewed. Moving the biomicroscope slightly forward will bring the endothelial cellular detail into focus. A magnification of 25 to 40 is usually needed to obtain a clear view of the endothelial mosaic. Cell density and morphology can be evaluated; guttae and keratitis precipitates will appear as nonreflective dark areas. Indirect illumination can provide unique information. In this method of illumination, the light beam is directed to an area adjacent to the area needing to be examined. Under higher magnification powers, the light beam may need to be decentered from the normal isocentric position. The area being evaluated is seen via retro-illumination from the deeper layers. This method is highly effective for observing detail in areas that are deep to more anterior lesions which may prevent light from penetrating through. For example, an embedded foreign body that is obscured by tissue reaction can be better mapped out through proximal illumination. Sclerotic scatter takes advantage of the optical phenomenon total internal reflection seen in the cornea. The examiner should direct a de-centered – but intense – light beam at the corneoscleral junction. In a normal cornea the light is only seen where it intersects and reflects the sclera. A ring of light around the limbus is seen, but the cornea will remain dark against an essentially dark background. In an abnormal cornea, the light that is reflected at the sclera will illuminate an abnormality. Thus, abnormalities can be seen in totality and be better recognized as part of a broader disease pattern (e.g., corneal verticillata). Retro-illumination of the iris combines both direct illumination and indirect retro-illumination to reveal

different details about the area being examined. When the light beam is directed tangentially toward the cornea, direct retro-illumination from the iris can reveal opaque corneal lesions. In contrast, the areas adjacent to the light beam can reveal subtle refractile changes of lower density in the cornea, through indirect retro-illumination from the iris. The interface against both light and dark backgrounds is what shapes the light to reveal details through retro-illumination. Examples of corneal pathology which are best revealed through retro-illumination from the iris include folds in Descemet’s membrane, and the linear stromal corneal changes in Lattice dystrophy. Retroillumination from the fundus reflects light from the retinal pigment epithelium through the pupil to reveal changes in the cornea, lens, and anterior vitreous. A well-dilated pupil will achieve the most effect because light will not be scattered from the iris. In contrast to sclerotic scatter where abnormalities are best seen against a dark background, this technique uses brightfield illumination to reveal pathology. Lenticular changes, such as posterior subcapsular cataracts or subluxation, can be clearly visualized against the color reflected from the retinal pigmented epithelium. Retroillumination of the iris and lens can reveal iris defects seen in glaucomatous diseases such as pigment-dispersion syndrome and pseudoexfoliaiton syndrome.

Specular Microscopy Specular microscopy was developed by David Maurice in 1968. This device was used to examine the corneal endothelium, ex vivo, in the laboratory. This first microscope had an effective magnification of 500, but was not practical for clinical use. In 1975, Bourne, Kaufman, and Lange developed clinical specular microscopes. Specular reflections arise from light which is reflected from the interfaces of materials with different indices of refraction. This occurs when the angle of incidence is equal to the angle of reflection. Thus, the difference between the index of refraction between the corneal endothelium and the aqueous produces a specular reflection. Many types of specular microscopes exist, which can be divided into contact and non-contact, and clinical (horizontal) and upright (which is used by eyebanks). Other types of specular microscopes are capable of visualizing other layers within the cornea. The advent of the specular microscope allowed the ophthalmologist direct observation of the corneal endothelium in patients with Fuchs’ dystrophy, posterior polymorphous dystrophy, psuedophakic bullous keratopathy, and other disorders of the corneal endothelium (Figure 1). If the cornea is significantly edematous, the specular reflection can be masked – which prohibits the visualization of the corneal endothelium. Specular microscopy is also used by eye banks to examine the corneal endothelium of

Imaging of the Cornea

I.D. FW BW

RECORD

CENTER

PRINT

CHIN REST

AUTO UTIL. PCHY

R

University of Minnesota

Figure 1 The specular photomicrograph shows a typical image from a noncontact specular microscope. From the information in this image, a cell count can be obtained and the degree of polymegathism and pleomorphism can be assessed.

donor corneas. The endothelial cell counts – which are reported as cells per square millimeter – are used to determine the suitability and health of the transplant tissue. Corneal endothelial cells typically demonstrate a hexagonal shape. A deviation from the normal hexagonal endothelial mosiac is termed pleomorphism. Although the size of corneal endothelial cells may vary slightly, in a healthy cornea, the cells should be similar in size. When there is great variablity in cell size, this is termed polymegathism.

Confocal Microscopy Confocal microscopy employs a point source of light, which is focused on a thin section of the specimen. The confocal point detector is used to collect the resulting reflected signal. The use of a pinhole light source and its conjugate pinhole detector trades field of view for enhanced resolution by eliminating light which is reflected from structures above and below the focal plan under observation. A full field of view must be obtained by scanning many regions of the specimen. This can be achieved by rotating discs that contain multiple conjugate point detectors-pinhole light sources that allow for even scanning of the tissue. Other confocal systems use a slit beam which scans the specimen with a mechano-optical mirror system. White-light or a laser can be used at the light source. These systems are respectively termed white light confocal microscopes or laser scanning confocal microscopes. The confocal microscope examines the cornea in coronally oriented sections versus the typical sagittal sections that are common to most histological tissue preparations (Figure 2). The corneal epithelium appears as a large cellular mosaic with bright, hyper-reflective central nuclei (Figure 2(a)). The basal epithelium consists of

215

smaller cells and, generally, does not exhibit visible nuclei (Figure 2(b)). The corneal stroma appears as a collection of hyper-reflective, bean-shaped keratocyte nucei (Figure 2(d) and 2(e)). The dendritic-appearing cell bodies of the keratocytes are only visible when the keratocytes are active. Corneal nerves can be seen passing through the stroma (Figure 2(c)). The deeper corneal nerves are large compared to the subepithelial nerve plexus, which resembles a fine filamentous membrane (Figure 2(e)). The corneal endothelium appears next (Figure 2(f)). Because the confocal microscope is able to eliminate aberrant light, corneal edema rarely obscures the corneal endothelium – unlike in the specular microscope. Images that are obtained during confocal microscopy are native video; individual still-frames can be captured, as well as video. The images can be stored digitally or on video tape. Because of the ability to scan through the cornea and other tissues, video images can be particularly informative. Furthermore, the images are typically registered by z-axis location (depth within the cornea or other tissue). Thus, three-dimensional (3D) reconstructions can be produced which reveal ex vivo histopathology – like those depicting sections of the cornea. The advantages of confocal microscopy are as follows: (1) the resolution can be better than the resolution obtained with the conventional light microscopy, allowing for imaging of the epithelial surface, stroma, nerves, and endothelium (2) high-resolution images can be obtained in vivo without the need for staining or processing of the cornea. The clinical applications of confocal microscopy include the identification of corneal infectious agents, including bacteria, fungi, and Acanthamoeba. It has been used in clinical research to assess the effect of corneal wound-healing responses after refractive surgery and to characterize corneal dystrophies (Figure 3). Because the device can measure in the z-axis, it can be used to determine the depth of structures within the cornea, such as the thickness of the laser-assisted in situ keratomileusis (LASIK) flap or the residual bed, and the depth of corneal scars. These devices may also be useful in diagnosing neoplasia of the cornea and conjunctiva.

Ultrasound Biomicroscopy High-frequency ultrasound biomicroscopy (UBM) produces high-resolution images of the anterior segment. Cross-sectional images of the eye are produced using a frequency range of 25–100 MHz. Resolution ranges from 20 to 100 mm. Higher frequencies provide higher resolution but the signal is increasingly attenuated, thereby limiting penetration of the signal. Newer technology can amplify the returning signal based on the depth of the structure under examination. The transducer has also

216

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

evolved. An arc transducer follows the curvature of the cornea and permits the visualization of the entire cornea and anterior segment. The analog signal is digitized and can be used to construct a 3D-representation of the anterior segment. The required use of a water bath – as a fluid coupler – may limit the applicability of the UBM.

Cornea volume [65], 10/15/2008, OD University of Minnesota

# 2/40: 0 µm

UBM is used to examine many pathological conditions of the cornea, ocular adnexa, tumors of the anterior segment, intraocular lens, and diseases of the sclera. Because the UBM can image intraocular structures, it can be used to view intraocular cysts, narrow anterior chamber angle, and intraocular foreign bodies.

Cornea section [11], 12/6/2007, OD # 1/1: 27 mm

(a)

University of Minnesota

(c)

Cornea section [10], 12/6/2007, OD # 1/1: 13 µm (b) Figure 2 (Continued)

University of Minnesota

Cornea section [2], 12/6/2007, OS # 1/1: 75 mm (d)

University of Minnesota

Imaging of the Cornea

217

Cornea section [10], 2/28/2008, OS

Cornea section [14], 12/6/2007, OD University of Minnesota

# 1/1: 581 µm

# 1/1: 285 µm

(e)

University of Minnesota

Figure 3 A confocal micrograph of the corneal endothelium of a patient with Fuchs’ dystrophy. The scant number of corneal endothelial cells is evident.

Anterior Segment Optical Coherence Tomography Anterior segment optical coherence tomography (OCT) is an imaging technique that can provide detailed in vivo visualization of the anterior chamber. This noninvasive, noncontact device uses OCT to produce direct crosssectional images that can be measured and used for diagnostic purposes. Using low coherence tomography, the light is set along two different optical paths: a sample path into the eye and a reference path of the interferometer. The light source is a 1310-nm superluminescent light-emitting diode (SLD). The light returning from the sample and reference paths are then combined at the photo-detector. The strength of the return signal is a measure of the reflectance of a small volume of

Cornea section [7], 12/6/2007, OD # 1/1: 536 µm

University of Minnesota

(f) Figure 2 Confocal microscopy permits the examination of each layer of the cornea. (a) The superficial corneal epithelium appears as a cellular mosaic of nucleated cells. (b) The basal corneal epithelial cells are smaller than the more superficial layers

of epithelium. (c) Just below the corneal epithelium lies the sub-basal nerve layer. This fine meshwork of corneal nerves is responsible for the exquisite sensitivity of the cornea. (d) Confocal images of the anterior corneal stroma reveal the nuclei of the keratocytes. Contrasting the density of the anterior stromal keratocyte nuclei with the keratocyte density in (e), demonstrates that there is a greater density of keratocytes in the anterior stroma. (e) This confocal image of deep corneal stroma reveals a field of keratocyte nuclei and a corneal nerve (hyper-reflective linear structure). (f) The hexagonal mosaic of the corneal endothelium is seen with the confocal microscope.

218

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

University of Minnessota S/W version: 1.2.0.1 Patient ID: Gopher, Goldie High res. corneal

OD

OS 0⬚

180⬚

0.14 mm

Figure 4 An OCT of the cornea of a LASIK patient. The blue caliper demonstrates the measurement of the thickness of the LASIK flap.

tissue. Varying the optical lengths of the reference paths at each of the scanning points determines the axial depth of the tissue signal. By moving the scanning spot across the eye, multiple A-scans align to form a two-dimensional image. There are many versatile uses of this device in assessing the anterior segment. The anterior chamber dimensions (depth, diameter, etc.) can accurately be measured. The iris and pupil – as well as the crystalline, pseudophakic, and refractive lens implants – can also be evaluated. A highly detailed evaluation of the angle structures can be assessed as well. In vivo measurements of the cornea can be evaluated in aid of surgical and refractive procedures as well as in diagnosing pathological processes. The device can measure post-LASIK corneal flap and residual stromal bed thickness, as well as produce full-thickness pachymetry maps which can assist both in refractive and glaucoma surgical planning (Figure 4). With the growing number of Descemet’s Stripping Automated Endothelial Keratoplasty (DSAEK) surgeries being performed, this tool has found an additional role in post-surgical lenticle-placement evaluation. Despite its numerous uses, using light as its medium limits its penetration into the eye. For example, ciliary body tumors cannot always be adequately visualized. In addition, image quality can be greatly reduced when imaging dense corneal opacities.

Acknowledgments We would like to acknowledge the support of Research to Prevent Blindness and the Minnesota Lions Club.

See also: Corneal Dystrophies; Corneal Endothelium: Overview; Corneal Epithelium: Cell Biology and Basic Science; Corneal Nerves: Anatomy; Corneal Scars; The Corneal Stroma; Ocular Mucins; Overview of Electrolyte and Fluid Transport Across the Conjunctiva; Refractive Surgery and Inlays; The Surgical Treatment for Corneal Epithelial Stem Cell Deficiency, Corneal Epithelial Defect, and Peripheral Corneal Ulcer.

Further Reading Chiou, A. G., Beuerman, R. W., Kaufman, S. C., and Kaufman, H. E. (1999). Confocal microscopy in lattice corneal dystrophy. Graefe’s Archive for Clinical and Experimental Ophthalmology 237: 697–701. Chiou, A. G., Kaufman, S. C., Beuerman, R. W., Maitchouk, D., and Kaufman, H. E. (1999). Confocal microscopy in posterior polymorphous corneal dystrophy. International Journal of Ophthalmology 213: 211–213. Dhaliwal, J. S., Kaufman, S. C., and Chiou, A. G. (2007). Current applications of clinical confocal microscopy. Current Opinion in Ophthalmology 18: 300–307. Kaufman, S. C., Musch, D. C., Belin, M. W., Cohen, E. J., Meisler, D. M., Reinhart, W. J., Udell, I. J., and Van Meter, W. S. (2004). Confocal microscopy: A report by the American Academy of Ophthalmology. Ophthalmology 111(2): 396–406. (Review.) Ledford, J. and Sanders, V. (2006). The Slit Lamp Primer, 2nd edn. New York: Slack Inc. Ramos, J. L., Li, Y., and Huang, D. (2009). Clinical and research applications of anterior segment optical coherence tomography – a review. Clinical and Experimental Ophthalmology 37(1): 81–89. Wylegała, E., Teper, S., Nowin´ska, A. K., Milka, M., and Dobrowolski, D. (2009). Anterior segment imaging: Fourier-domain optical coherence tomography versus time-domain optical coherence tomography. Journal of Cataract and Refractive Surgery 35(8): 1410–1414.

The Corneal Stroma J L Funderburgh, University of Pittsburgh, Pittsburgh, PA, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Adherens junctions – Protein complexes that occur at cell–cell junctions, involving calcium-dependent homophilic interactions of a family of transmembrane proteins called cadherins. Connexin – Gap-junction proteins; family of structurally related transmembrane proteins that assemble to form vertebrate gap junctions. Each gap junction is composed of two hemichannels, or connexons, which are themselves each constructed out of six connexin molecules. Corneal dystrophies – Group of disorders characterized by a noninflammatory, inherited, bilateral opacity of the cornea. Crystallins – Water-soluble structural proteins found in the lens of the eye, accounting for the transparency of the structure. Ectomesenchyme – It has similar properties to mesenchyme. The major difference is that ectomesenchyme arises from neural crest cells, which are a critical group of cells that form in the cranial region during early vertebrate development. Glycan – It refers to a polysaccharide or oligosaccharide. Glycan may also be used to refer to the carbohydrate portion of a glycoconjugate, such as a glycoprotein, glycolipid, or a proteoglycan. Glycosaminoglycans – Long, linear carbohydrate polymers that are negatively charged under physiological conditions, due to the occurrence of sulfate and uronic acid groups. Hurler’s syndrome – Known as mucopolysaccharidosis type I (MPS I), Hurler’s disease, or gargoylism, this genetic disorder results in the buildup of mucopolysaccharides due to a deficiency of alpha-L iduronidase, an enzyme responsible for the degradation of mucopolysaccharides in lysosomes. Without this enzyme, a buildup of dermatan sulfate occurs in the body. Symptoms appear during childhood and early death can occur due to organ damage. Keratan sulfate – Also called keratosulfate, it is any of several sulfated glycosaminoglycans (structural carbohydrates) found especially in the cornea, cartilage, and bone. Keratan sulfates are large, highly hydrated molecules which, in joints, can act as a cushion to absorb mechanical shock. Keratocytes – The basic cell type found in the corneal stroma. The keratocytes are sparse in

distribution, occupying less than 5–10% of the stromal volume. Lamellae – A lamella is a thin plate-like structure, often one among many lamellae very close to one another, with open space between. Lumican – Also known as LUM, it is a human gene. This gene encodes a member of the small, leucinerich proteoglycan (SLRP) family that includes decorin, biglycan, fibromodulin, keratocan, epiphycan, and osteoglycin. In these molecules, the protein moiety binds collagen fibrils and the highly charged hydrophilic glycosaminoglycans regulate interfibrillar spacings. Not only is lumican the major keratan sulfate proteoglycan of the cornea, but it is also distributed in interstitial collagenous matrices throughout the body. Lumican may regulate collagen fibril organization and circumferential growth, corneal transparency, and epithelial cell migration and tissue repair. Macular corneal dystrophy – An autosomal recessive condition, which is the least common but the most severe of the three major stromal corneal dystrophies. It is characterized by multiple, graywhite opacities that are present in the corneal stroma and that extend out into the peripheral cornea. Mesenchyme – Loosely organized connective tissue present in the embryo regardless of origin. Viscous in consistency, mesenchyme contains collagen bundles and fibroblasts. Myofibroblast – Cell with a phenotype between a fibroblast and a smooth muscle cell in differentiation. It can contract by using smooth muscle-type actin–myosin complex, rich in a form of actin called alpha-smooth muscle actin. These cells are then capable of speeding wound repair by contracting the edges of the wound. Neural crest – Transient component of the ectoderm, located between the neural tube and the epidermis of an embryo during neural tube formation. Neural crest cells migrate during neurulation, an embryological event marked by neural tube closure. Proteoglycan – Special class of glycoproteins that are heavily glycosylated. They consist of a core protein with one or more covalently attached glycosaminoglycan (GAG) chain(s). Scheie’s syndrome – Mildest form of mucopolysaccharidosis type I (MPS I) (see Hurler’s syndrome).

219

220

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Sutural fibers – The elasmobranch (sharks, skates, and rays) cornea resists swelling because of unique structural adaptations in the stroma called sutural fibers. These are collagen fibers running anterior-toposterior, perpendicularly to the stromal lamellae, tying the anterior limiting lamella to Descemet’s membrane. Transforming growth factor-beta – Transforming growth factor-beta (TGF-b) is a small protein-growth factor with a broad array of functions. It controls proliferation, cellular differentiation, and other functions in most cells. It plays a role in immunity, cancer, and heart disease.

Stromal Anatomy In humans, the cornea has a diameter of about 11.5 mm and a thickness of 0.5–0.6 mm in the center and 0.6–0.8 mm at the periphery. Almost 90% of the human cornea is composed of stroma (Figure 1(a)). By weight, water in the extracellular matrix makes up 65% of the stroma and cellular water 11%. This hydration is roughly similar to that of cartilage, but surprisingly, it is higher than that of bone, muscle, or adipose tissue. As discussed below, the proteoglycans of the stroma are hydrophilic and, given free access to water, the tissue will imbibe up to 10-fold its normal content of water. Thus the tissue composition, particularly its hydration, is a dynamic property. The majority of the stroma consists of thin sheets (lamellae) of tightly packed collagen fibrils. In human corneas, there are up to 200–250 such lamellae, each about 2-mm thick. Within each lamella, the collagen fibrils run parallel to the corneal surface, parallel to one another, and regularly spaced. Ends of the fibrils are not abundant and thus fibrils extend essentially the entire width of the cornea. The lamellae are largely self-contained, but occasional bundles of fibrils extend from one lamella to another. Directional orientation of the fibril layers varies between neighboring lamellae. In the posterior stroma the fibril orientation is almost perpendicular from one layer to the next. Anteriorly, it is more oblique. This arrangement of layers of parallel rigid rods in a friable matrix is not unlike that of composite structural materials such as fiberglass or reinforced concrete. Such an arrangement gives the cornea its remarkable toughness and tensile strength. Numerous electron micrographic studies have documented the striking regularity of the collagen in the stroma. The collagen fibrils in central stroma have a diameter of about 31 nm and the distribution of diameters is narrow, giving a remarkably homogeneous distribution of collagen within each fibril bundle. Within the fibril bundles electron-dense material, identified as proteoglycans,

encase the individual fibrils and form bridges between neighboring fibrils (Figure 1(b)). In central stroma of normal human cornea, the proteoglycan bridges occur at highly regular intervals, and the bridging structures are of uniform lengths, about 1.8 nm. Such tight interaction between fibrils is considered to participate in generating the parallel alignment of the fibrils in each bundle and the highly regular spacing between neighboring fibrils. As discussed below, this lattice-like structure of the collagen fibrils in the central stroma is thought to be essential for corneal transparency. The anterior portion of most corneal stromas is limited by an acellular layer of dense, irregularly organized collagen immediately subjacent to the epithelial basement membrane. This anterior limiting lamella (ALL) is also known as Bowman’s layer or Bowman’s membrane. In human corneas, the ALL is about 10-mm thick in the central cornea and absent in the periphery. ALL is also prominent in chickens and some other terrestrial mammals, but it is very thin or not detected in other species such as felines. Collagen fibrils in the ALL are randomly interwoven to form a dense, felt-like sheet composed primarily of collagen types I, III, and V. Collagen VII, associated with anchoring fibrils of the overlying epithelium, is also present within the ALL. The posterior of the ALL merges with the lamellar stroma via oblique fibril bundles, recently revealed using two-photon microscopy. These connecting fibers appear to serve as a stabilizing feature, anchoring the anterior lamellae to the more rigid ALL and indirectly to the epithelial basement membrane. These anterior anchoring fibers appear analogous to the wellknown sutural fibers present in corneas of some species of sharks. The sutural fibers traverse the cornea perpendicular to the orientation of the lamellae and prevent the stroma from swelling when exposed to water. The oblique anterior anchoring fibrils recent identified in human corneas may provide a similar function in that studies show the swelling of human corneas in water occurs almost exclusively in central and posterior stroma. The extremely dense collagen in the ALL has prompted speculation that it may serve as a defense against bacterial or viral infection; however, little hard evidence supports such a role.

Stromal Development The embryonic origins of the corneal stroma were detailed in a elegant study by Hay and Revel in 1969. In developing chicks, a wave of migrating neural crest cells, destined to become corneal endothelium, moves between the overlying ectoderm and the lens during early embryogenesis. Shortly thereafter, the epithelium secretes an acellular layer of matrix, termed the primary stroma, which becomes hydrated and swells before a second wave of neural crest cells move into the stroma and begins active

The Corneal Stroma

221

(a) Epithelium (b) Anterior limiting lamella (c) ‘Anchoring’ collagen

(d) Lamellar collagen (e) Keratocytes (f) Descemet’s membrane (g) Endothelium

(A)

Glycosaminoglycan

(B)

Proteoglycan core protein

Fibril (C)

Figure 1 Components of the corneal stroma. (A) Cross section of the cornea with the expanded cutaway illustrating (a) epithelium, (b) anterior limiting lamella, (c) anterior ‘anchoring’ collagen fibrils, (d) aligned lamellar collagen in central and posterior stroma, (e) keratocytes sandwiched between lamellae, (f) Descemet’s membrane (g) endothelium. (B) Interaction between collagen fibrils and stromal proteoglycans. The core proteins bind fibrillar collagen at regular intervals and glycosaminoglycan chains protrude into the interfibrillar space. (C) Cultured primary bovine keratocyte with its extensive cellular processes (Green, actin; Red, vinculin).

secretion of the abundant extracellular matrix of the stroma. The orientation of collagen in the primary stroma is thought to direct formation of lamellae elaborated by the invading stromal cells. At day 14 (before hatching at day 21), the stroma undergoes dehydration in response to thyroxine, leading to thinning of the stroma and initiation of stromal transparency. Mammalian corneas show a somewhat different developmental pattern. No obvious primary stroma is present, and the endothelium and stromal cells are formed after a single influx of neural crest cells. After the endothelium is formed, neural crest cells in the stroma begin to elaborate new extracellular matrix. In mice and rabbits, cells in the stroma are mitotically

active until after birth. In mice, approximately at the time of eye opening (12–14 days postnatal) a decrease in the number of stromal cells occurs along with thinning and dehydration of the stroma. Simultaneously, the cornea-specific glycosaminoglycan – keratan sulfate – appears in the stroma, the stromal cells become quiescent, stromal proteins known as corneal crystallins accumulate in keratocytes, and transparency of the cornea increases significantly. Human stroma, by contrast, matures earlier in development. Keratan sulfate and corneal transparency are observed in embryos at 10–12 weeks of gestation, and infants are born with fully transparent corneas. Despite these species-related differences, the pattern of neural

222

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

crest population of the stromal space, secretion of cornealspecific matrix components following formation of the endothelial barrier, and corneal thinning due to dehydration present a common pattern of stromal development.

Stromal Cells Cells populating the adult stroma are known as keratocytes. Although the same term has been used for fish epidermal cells, corneal keratocytes are neural crestderived mesenchymal cells (ectomesenchyme). In adult tissue, the keratocytes are located between collagenous lamellae, exhibiting a flattened cell body with numerous extended cellular processes (Figure 1(c)). Keratocytes are the population of cells responsible for deposition and maintenance of the extracellular matrix which provides strength and transparency to the cornea. This population of cells is highly interconnected via their processes with numerous junctions between neighboring cells. At sites of contact, gap junctions containing connexin 43 and adherens junctions with cadherin 11 are present. A localized stimulus, such as an epithelial scratch, activates keratocytes throughout the stroma, as indicated by uptake of the dye neutral red. Such a global response to a localized stimulus demonstrates presence of a network of active cell–cell communications maintained by interconnected keratocytes. Following maturity, keratocytes remain quiescent, almost never exhibiting mitotic nuclei or uptake of DNA precursors. Studies of keratocyte turnover in mammals find half-lives too long to estimate accurately. DNA of keratocytes in adult humans shows a high incidence of acquired chromosomal abnormalities, but that of infants and children does not. This fact suggests that keratocytes do not regularly enter the cell cycle in which DNA repair is initiated, but accumulate damage from environmental exposure to light and oxidative stresses over long periods of time. Thus, in the nonwounded cornea, keratocytes may exist for years or decades without turnover. In spite of their quiescence, keratocytes maintain an active synthesis of the extracellular matrix. Proteoglycan secretion is maintained at a high rate throughout life. Because collagen becomes cross-linked in adult corneas, however, collagen turnover is reduced in adult keratocytes compared to that in embryonic tissue. Adult keratocytes also retain the ability to enter the cell cycle and divide. Nearly 100% of keratocytes isolated by collagenase digestion of bovine corneas enter the cell cycle when exposed to serum or mitogens. Collagenase-isolated primary keratocytes cultured in serum-free conditions maintain the keratocyte phenotype, as defined by quiescence – a dendritic morphology (Figure 1(c)) – and expression of high levels of cornea-specific products (Table 1); however, when exposed to serum or other mitogens, these cells dedifferentiate into a fibroblastic phenotype similar to

that of other mesenchymal cells cultured in serum. Following short-term culture in serum, some keratocyte properties return under quiescent conditions, but after multiple passages in culture transition to the fibroblast phenotype appears to be irreversible. When cultured corneal fibroblasts are exposed to transforming growth factor-beta (TGF-b), the cells express mRNA and protein for alpha-smooth muscle actin and assume a contractile phenotype known as the myofibroblast. TGF-b also induces a range of new gene products, many associated with corneal scarring or fibrosis in vivo. Genes and gene products which have been identified in corneal fibrotic (scar) tissue and also as products of myofibroblasts are shown in the top rows of Table 1. These include cellassociated markers such as Thy-1, matrix proteins (fibronectin, SPARC, tenascin c, etc.), collagens, proteoglycan proteins (biglycan), and glycosaminoglycans. The last four entries of Table 1 show components expressed at high levels in keratocytes which are reduced or disappear in corneal scars and are not expressed by myofibroblasts. These include corneal crystallins (such as ALDH3), the keratan sulfate proteoglycan keratocan, and the cellsurface marker CD34. As described below, the characteristic changes in collagens, crystallins, and proteoglycans occurring during keratocyte–myofibroblast transition directly impact stromal transparency. Thus, loss of transparency in extracellular matrix deposited in response to wound healing can be directly attributed to responses of the keratocytes to wounding. Smooth muscle actin-containing cells appear in healing corneas 1–2 weeks following wounding, and this appearance can be blocked with antibodies to TGF-b. This inhibition suggests that in vivo, as in vitro, TGF-b is an important inducer of fibrotic scar tissue. Several weeks following healing of corneal wounds, smooth muscle actin-containing cells are no longer seen in the wound area. It not clear, however, that the loss of smooth

Table 1

Markers of stromal fibrosis

Gene or product Dermatan sulfate Hyaluronan Biglycan Tenascin C SPARC Fibrillin-1 Collagen 1 Collagen 3 Thy-1 Aldehydehyde dehydrogenase 3A1 Keratocan Keratan sulfate Collagen a3(IV) CD34

Normal stroma

Fibrotic stroma and myofibroblast

þ 0 þ/ 0 0 0 þþ þ/ þ/ þþþþ

þþþ þþþþ þþþ þþþ þþþ þþþ þþþþ þþ þþ þ

þþþ þþþþ þþ þþþ

þ 0 0 0

The Corneal Stroma

muscle actin expression corresponds to a return to normal stromal extracellular matrix production. In fact, experimental animal studies show expression of fibrotic markers for months following wounding, and human scar tissue can persist for many decades. In addition to keratocytes, other cell types have been observed in the stroma. Most prominent are bone marrowderived cells. Mice in which bone marrow cells express green fluorescent proteins have about 10% of stromal cells demonstrating fluorescence. Most stromal leucocytes express the CD11b marker, but not other dendrite, granulocyte, T-cell, or NK markers, placing them in the monocyte/macrophage lineage. Following minor damage to the epithelium, the number of green inflammatory cells in the stroma increases dramatically within a few hours. These transient inflammatory cells are primarily neutrophils. A population of stem or progenitor cells has also been identified in the corneal stroma. Transplantation of avian corneal keratocytes into the neural crest migratory pathways of early embryos showed that some of the transplanted cells changed phenotype, migrating and differentiating into a variety of neural crest-derived tissues, thus demonstrating that progenitor cell potential is maintained following stromal differentiation. In mammals, similarly, stromal cells with stem cell properties can be isolated using a variety of techniques including spheroid culture, cloning, and fluorescence-activated cell sorting. These cells exhibit multipotent differentiation, clonal growth, and expression of a number of stem-cell-associated genes. These stromal stem cells do not express characteristic keratocyte markers, but can do so under selected culture conditions or when injected into corneal stroma in vivo. These stromal stem cells can restore transparency to lumican null mice with stromal haze, suggesting they may be appropriate for cell-based therapy of corneal diseases. The stromal stem cells appear to reside in the anterior region of the stroma near the limbus, but it is currently unknown what specific roles they play in corneal maintenance or repair.

223

functioned like a crystal, producing destructive interference of scattered light, allowing nonscattered light to pass. Mathematical models by McCally, Farrell, and others supported this concept, suggesting that, as long as the collagen fibrils were small, tightly packed, and aligned, transparency was possible. The crystal-like regularity of the stromal collagen allows for analysis using classical techniques such as X-ray diffraction. In a large number of such studies, Keith Meek and collaborators have confirmed the high regularity of spacing of nearest neighbor collagen fibrils and loss of that regularity in virtually every instance in which the cornea loses transparency. These studies have confirmed stromal hydration as one of the most critical parameters in maintenance of transparency. Excess water is adsorbed by the proteoglycans between collagen fibrils, disrupting the critical spacing and eliminating the order essential for transparency. Stromal transparency is consequently a highly dynamic property, dependent both on the removal of water from the tissue by the pumping action of the endothelial layer and by the water-binding properties of the stromal proteoglycans. Recently, a new aspect of corneal transparency has come to notice with the discovery of abundant soluble proteins in keratocytes and other cellular tissues of the cornea. These proteins, termed corneal crystallins, are typically enzymes with housekeeping functions present in corneal cells at concentrations vastly exceeding that in other tissues. This group of proteins includes isoforms of aldehyde dehydrogenase (ALDH), transketolase, and up to a dozen or more unrelated enzymes. The variety of proteins involved and their unusual abundance has led to the suggestion that – like lens crystallins – these proteins serve to alter the refractive index of cells, thus reducing light scatter by corneal cells. Studies by Jester and coworkers have shown light scatter by keratocytes correlates with abundance of corneal crystallins in these cells, supporting a role for crystallins in stromal transparency.

Stromal Extracellular Matrix Transparency The cornea is one of the few complex biological tissues with high transparency to light, and the biophysical properties which allow such transparency have been the subject of debate for decades. In most transparent tissues such as the lens, all structures have a similar index of refraction so that light is not scattered as is passes through the tissue. Birefringence studies of the cornea, however, demonstrate that the corneal collagen has an index of refraction different from the extrafibrillar matrix surrounding it, and thus should scatter light, producing an opalescent or white appearance. David Maurice proposed, in 1957, that the highly regular structure of the collagen fibrils in the stroma

Collagens are the most abundant proteins in the cornea, and most of the collagen in the stroma is involved in the collagen fibrils of the lamellae. Studies by Birk and coworkers have shown that fibrils in the stroma are heterotypic, containing both types I and V collagens in the same fibril. The triple-helical domain of the type V collagen molecules is buried within the fibril with its NH2-terminal domains exposed at the fibril surface. These exposed domains alter lateral association of collagen molecules during fibrillogenesis and, therefore, limit fibril diameter. The abundance of collagen type V is a likely factor in the small diameter of the stromal fibrils compared to fibrils in dermis and sclera, which contain mostly collagen types I and III. In corneal scarring and

224

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

haze, collagen type III is elevated in the stroma suggesting a direct correlation between the abundance of the small type I and V fibrils and corneal transparency. Type XII collagen is found associated at regular intervals with stromal fibrillar collagen. The long form of this fibril-associated (FACIT) collagen can be modified with chondroitin sulfate, making it also a fibril-associated proteoglycan. Mice expressing a truncated form of type XII show altered collagen fibril spacing in the stroma. In addition to fibrillar collagen, stroma is very rich in type VI – a nonfibrillar collagen. This protein interacts with cells and numerous components of the stroma forming a beaded network throughout the stroma. Type VI microfibrils can be visualized running perpendicularly to the fibrils in lamellae and may help stabilize the lamellar structure. Numerous other collagen types have also been identified in the stroma in small amounts. Type IV collagens, often associated with basement membrane, are seen in the stroma to be associated with keratocytes. One of these, alpha3(IV) is downregulated as keratocytes become activated by wound healing or mitogens, and thus serves as a marker of keratocyte phenotype. Proteoglycans are the second most abundant component of the corneal stroma. The corneal proteoglycans all belong to the small leucine-rich proteoglycan (SLRP) family, consisting of proteins of about 40 kDa, decorated with several N-linked oligosaccharides and one to two glycosaminoglycan chains. Normal adult stroma has one proteoglycan protein, decorin, which is modified with dermatan sulfate and three more SLRP proteins– lumican, keratocan, and osteoglycin (mimecan) – which have keratan sulfate chains. During healing, a fifth SLRP protein, biglycan, is detected in the stroma. Biglycan joins decorin as a second dermatan sulfate-containing proteoglycan in scar tissue. These SLRP proteoglycans contain multiple leucine repeat regions (LRR) – motifs involved in protein–protein binding. Each of the stromal proteoglycans binds collagen in a repeating pattern along the length of the fibrils with the keratan sulfate-containing SLRP’s binding sites different from that of decorin. X-ray studies and molecular modeling reveal that SLRP proteins fold into compact horseshoe-shapes with the glycosylation on the convex portion of the curve and the collagen-binding region on the inner face. Thus, proteoglycan association with collagen produces glycosaminoglycan chains protruding from the fibril with a ‘bottle brush’ like appearance, as documented by deep freeze etch electron microscopy (Figure 1(b)). The interactions between proteoglycan and collagen has an effect on the rate and size of the fibrils formed during fibrillogenesis, effecting the diameter and length of the collagen fibrils, and thus the physical properties of the tissue. Decorin-knockout mice thus have weakened skin, whereas lumican-knockout mice exhibit large and heterogeneous collagen fibrils in the posterior corneal stroma leading to corneal haze. These in vivo models confirm the importance

of the SLRP proteins in maintenance of the stromal ultrastructure required for vision. In addition to structural roles, the SLRP proteins interact directly with cells and growth factors. Decorin activates the EGF receptor on cell surfaces and also binds to and inactivates TGF-b, leading to reduced fibrosis in experimental models. Lumican stimulates attachment of macrophages and also has been shown to stimulate the healing of corneal epithelial wounds. Lumican-knockout mice show reduced responsivness to lipopolysaccharide-induced septic shock, and poor induction of proinflammatory cytokines. Lumican core protein also binds the CXC-Chemokine KC (CXCL1) and thus regulates neutrophilic infiltration. The SLRP proteins, therefore, clearly play important roles in mediating the response of the stroma to injury and inflammation. The glycosaminoglycan chains (glycans) – decorating SLRP proteins – constitute about half of the molecular weight of the proteoglycans. Because of their high level of sulfation, these glycans are hydrophilic, providing the impetus for influx of water into the stroma against which the endothelium provides an active pump. Dermatan sulfate is a widespread and abundant glycan which, in the cornea, is less highly sulfated than in skin or sclera. Keratan sulfate, while detectable in many tissues, is present in abundance only in cartilage and cornea. In addition, only in the cornea are lumican, keratocan, and mimecan glycanated with keratan sulfate, making the stromal keratan sulfate proteoglycans an abundant, structurally unique, tissue-specific class of matrix molecules. Keratan sulfate binds water differently than dermatan sulfate, and it is clear that the ratio of these glycosaminoglycans in the stroma is important for stromal transparency. In genetic diseases such as Scheie’s and Hurler’s syndromes, dermatan sulfate cannot be degraded and accumulates in the cornea. In such cases, intense corneal opacity occurs early in life. In macular corneal dystrophy, keratan sulfate is not properly sulfated. In this disease, the cornea loses transparency in the second decade of life. It is notable that, in scar tissue, keratan sulfate is absent or greatly reduced, whereas dermatan sulfate is more abundant and highly sulfated. It is thought that such long-term differences in the glycosaminoglycan content of scar tissue contributes to light scattering as a result of differential water binding by the two types of glycosaminoglycan. Biosynthesis of glycosaminoglycans is controlled differently from that of the proteins to which they are attached. For example, TGF-b treatment of keratocytes in vitro causes little change in lumican or decorin secretion; however, keratan sulfate modifying the lumican becomes dramatically shorter and virtually unsulfated during this treatment and dermatan sulfate modifying the decorin becomes longer and much more highly sulfated. These changes in glycanation can have dramatic effects on the properties of the proteoglycan. Lumican with short, unsulfated glycan chains serves as an

The Corneal Stroma

attachment substratum for macrophages and stimulates cell migration. Lumican modified with highly sulfated keratan sulfate, on the other hand, is antiadhesive and serves as a barrier for migrating cells. Thus the state of the glycan chains of the SLRP proteins serves not only to regulate water binding, but also other biological properties of the molecules. Modification of synthesis of the glycosaminoglycan is complex and not yet fully understood. Initiation, elongation, and sulfation of dermatan sulfate requires the participation of up to 16 different glycosyltransferase and sulfotransferase enzymes. For keratan sulfate, not all of participating enzymes are known as yet. One important gene has been clearly identified, however. CHST6 codes for a cornea-specific keratan sulfotransferase. This enzyme is mutated or nonfunctional in macular corneal dystrophy, leading to undersulfation of keratan sulfate. In healing wounds and fibrotic corneas, not only are keratan sulfate and dermatan sulfate altered but another glycosaminoglycan, hyaluronan (HA), is present. This glycan, not present in normal corneal tissue, appears rapidly in the stroma in most pathological conditions and remains for many months following healing. The hyaluronan biosynthetic enzyme HAS2 is upregulated in keratocytes in response to mitogens and TGF-b. Mice lacking HAS2 in the stroma do not express HA in response to induced inflammation, demonstrating that upregulation of HAS2 mRNA in the keratocytes is the source of HA during stromal pathology. Hyaluronan is a simple, unsulfated, acidic polysaccharide but is known to exhibit a large number of biological activities. It is associated with cell motility, inflammation, and with metastatic potential of cancer cells. In culture, knockdown of HAS2 mRNA reduces keratocyte ability to respond to TGF-b with fibrotic matrix components. These results suggest that HA might represent an extracellular signal that mediates scarring in the stroma. A final class of matrix components important to stromal function are the noncollagenous matrix proteins. Some of these are listed in Table 1. Another member of this group is protein BIGH3 (TGF-b-inducible gene H3). This secreted protein interacts with collagen and contains an RGD amino-acid sequence. It promotes cell attachment and in a number of systems the BIGH3 protein inhibits cell growth and motility. In the cornea, it has been shown that mutations in the gene coding for this protein (TGFBI) lead to a variety of corneal dystrophies presenting as opaque deposits in the stroma, usually in adults. Some of the specific syndromes caused by BIGH3 are granular dystrophy, Groenouw type I, Reis–Bucklers, lattice dystrophy type I, and Avellino dystrophy. The discovery that a minor component of the stromal extracellular matrix leads to marked disruption of vision attests to the complex biophysical equation involved in maintenance of stromal transparency. There are likely to be more important components of this system yet to be discovered.

225

Conclusion The corneal stroma is a physically tough tissue with the remarkable property of transparency to light. Stromal transparency is essential for vision, and corneal scarring obscures vision for millions of individuals worldwide. Although we have gained an understanding of how stromal fibrosis disrupts vision, we do not yet understand how this process might be reversed biologically. An important challenge for the future is to employ our understanding of stromal biology to design pharmaceutical or cell-based treatments to reverse the scarring process and restore the complex balance of cells, molecules, and water in order to provide vision for those affected by corneal blindness.

Acknowledgments The authors wish to thank Kira Lathrop and Martha Funderburgh for the excellent illustrations. This work was supported by NIH Grant EY016415 and Research to Prevent Blindness Inc. See also: Artificial Cornea; Corneal Dystrophies; Corneal Imaging: Clinical; Corneal Scars.

Further Reading Birk, D. E. (2001). Type V collagen: Heterotypic type I/V collagen interactions in the regulation of fibril assembly. Micron 32(3): 223–237. Farrell, R. A., McCally, R. L., and Tatham, P. E. (1973). Wave-length dependencies of light scattering in normal and cold swollen rabbit corneas and their structural implications. Journal of Physiology 233(3): 589–612. Fini, M. E. (1999). Keratocyte and fibroblast phenotypes in the repairing cornea. Progress in Retinal and Eye Research 18(4): 529–551. Guerriero, E., Chen, J., Sado, Y., et al. (2007). Loss of alpha3(IV) collagen expression associated with corneal keratocyte activation. Investigative Ophthalmology and Visual Science 48(2): 627–635. Hay, E. D. and Revel, J. P. (1969). Fine structure of the developing avian cornea. Monographs in Developmental Biology 1: 1–144. Jester, J. V. (2008). Corneal crystallins and the development of cellular transparency. Seminars in Cell and Developmental Biology 19(2): 82–93. Maurice, D. M. (1957). The structure and transparency of the cornea. Journal of Physiology 136(2): 263–286. McCally, R. L. and Farrell, R. A. (1982). Structural implications of small-angle light scattering from cornea. Experimental Eye Research 34(1): 99–113. Meek, K. M., Leonard, D. W., Connon, C. J., Dennis, S., and Khan, S. (2003). Transparency, swelling and scarring in the corneal stroma. Eye (London, England) 17(8): 927–936. Meek, K. M. and Quantock, A. J. (2001). The use of X-ray scattering techniques to determine corneal ultrastructure. Progress in Retinal and Eye Research 20(1): 95–137. Morishige, N., Petroll, W. M., Nishida, T., Kenney, M. C., and Jester, J. V. (2006). Noninvasive corneal stromal collagen imaging using twophoton-generated second-harmonic signals. Journal of Cataract and Refractive Surgery 32(11): 1784–1791.

Corneal Dystrophies B H Feldman, Philadelphia Eye Associates, Philadelphia, PA, USA N A Afshari, Duke University Medical Center, Durham, NC, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Amyloid – Insoluble, fibrous, and primarily extracellular protein aggregates exhibiting beta-sheet structures that are deposited in many local and systemic diseases. Anterior synechia – A pathological condition where the iris adheres to the cornea. Corectopia – An iris defect involving displacement of the pupil from its normal position. Corneal dystrophy – A bilateral inherited disorder of noninflammatory, progressive corneal disease. Corneal guttata – Wart-like excrescences of Descemet’s membrane that are associated with Fuchs’ corneal dystrophy. Lamellar keratoplasty – A partial thickness corneal transplant of either the anterior or posterior corneal layers. Penetrating keratoplasty – Full thickness corneal transplant. Phototherapeutic keratectomy – A surgical procedure in which the epithelium is removed and then an excimer laser is used to ablate abnormal anterior stromal tissue. Recurrent erosion – A syndrome of repeated epithelial defects due to abnormal adhesion of the epithelial basement membrane.

Introduction The corneal dystrophies are a group of noninflammatory, inherited, and bilateral disorders characterized by pathognomonic patterns of corneal deposition and morphological changes. These heterogeneous dystrophies are defined by their clinical characteristics, histological findings, and genetics. Traditionally, they have also been grouped anatomically into three categories of anterior, stromal, and posterior dystrophies. The anterior dystrophies affect the epithelium, epithelial basement membrane (EBM), or Bowman’s layer. The stromal dystrophies primarily affect the stroma, but may extend into the anterior corneal layers and may rarely affect Descemet’s membrane and the endothelium. The posterior dystrophies are primarily disorders of endothelial cells and the posterior portion of Descemet’s membrane. They may also alter the structure of the stroma and anterior cornea.

226

There is a wide variation in the type and degree of symptoms caused by the corneal dystrophies. Progressive accumulation of tissue deposits can lead to vision loss from corneal opacification and astigmatism. Disruption of normal cell function can lead to abnormal epithelial adhesion with resultant painful recurrent epithelial erosions or loss of endothelial cell activity with resultant corneal edema. Several of the dystrophies are associated with vision-threatening ocular or systemic manifestations. Together, corneal dystrophies are the primary indications for approximately 10–15% of the corneal transplantations performed in the United States each year.

Anterior Dystrophies Epithelial and EBM Dystrophies EBM Dystrophy Also known by the descriptive term, map-dot-fingerprint (MDF) dystrophy, or the eponym, Cogan’s dystrophy, EBM dystrophy (EBMD) is arguably the most common dystrophy, found in approximately 5% of the adult population, with a slight preponderance of women. While in rare instances an autosomal dominant inheritance pattern has been identified, in the vast majority of cases this disease is sporadic, making its designation as a corneal dystrophy controversial. On examination, several patterns of basement membrane and epithelial involvement can be observed, including dots or microcysts, map lines, and fingerprint lines. These abnormalities may fluctuate over time but are rarely progressive. The most common symptoms of EBMD are recurrent erosions and blurred vision. The recurrent erosions arise because of poor epithelial adhesion to the underlying basement membrane. These erosions may be triggered by the lid trauma caused by an innocuous blink and are most often noted upon eye opening in the morning – with pain sometimes – waking the patient from sleep. The pain, foreign body sensation, and blurring associated with the erosions typically last only several minutes, but for some may be prolonged and severe. In addition to these transient symptoms, the surface changes from EBMD in the visual axis may lead to irregular astigmatism and adversely affect a patient’s best-corrected visual acuity. Recurrent erosions are managed acutely with lubricants, bandage contact lenses, patching, and prophylactic antibiotics. Medical options for prophylaxis against these recurrent erosions begin with aggressive lubrication and an

Corneal Dystrophies

emphasis on the use of bedtime ointments to avoid epithelial dehydration which can potentiate poor epithelial adhesion. Additionally, it is beneficial to address co-existing ocular surface diseases such as blepharitis, and the use of oral tetracyclines may inhibit epithelial breakdown. Similarly, anti-inflammatory agents such as topical steroids may inhibit this breakdown, but the long-term use of these agents should be dealt with caution. If recurrent erosions persist despite medical interventions, epithelial adhesion can be improved through mechanical disruption with subsequent healing through debridement with diamond burr polishing, stromal needle puncture, or Nd:YAG laser puncture. For large or recurrent lesions in the visual axis, excimer laser phototherapeutic keratectomy may be preferred to avoid scarring due to other techniques. Microscopic examination of EBMD reveals a thickened epithelial layer with a thickened or redundant basement membrane that has extended into the epithelium in either linear propagations or sheets, representing the map and fingerprint lines, respectively. The dots are microcysts formed by these abnormal extensions of basement membrane that have entrapped degenerated epithelial cells. Meesman’s

This rare anterior dystrophy is an autosomal dominant disorder characterized by the development of intraepithelial vesicles in the central cornea that appear as early as in the first few years of life. Over time, there is progressive involvement of the mid-peripheral cornea and an increase in the number and density of vesicles. As the disease progresses, vision may slowly deteriorate and the eye may become irritated as the intraepithelial vesicles rupture. In contrast to EBMD, recurrent erosions are uncommon in this dystrophy. While most patients are asymptomatic, those who have ocular irritation from ruptured vesicles are treated with lubricants or soft contact lenses to manage these microerosions. The pathologic changes in the epithelium are thickening of the basement membrane and the formation of small epithelial cysts containing a material composed of degenerated keratin with cytoplasmic and nuclear debris, known as peculiar substance. Consistent with the epithelial nature of this dystrophy, Meesman’s lesions are known to recur in cornea transplants as the host epithelium repopulates the surface of the donor button. Genetic analysis has elicited responsible mutations in two genes (KRT3 and KRT 12) on chromosomes 12q13 and 17q12, respectively, which code for two corneal keratins (k3 and k12). Bowman’s Layer Dystrophies Reis-Buckler’s

Also known as corneal dystrophy of Bowman’s type 1 (CDB-1) and granular dystrophy type III (GD-3), this is

227

Figure 1 External photograph of Reis-Buckler’s dystrophy (corneal dystrophy of Bowman’s type 1 or granular dystrophy type 3) demonstrating geographic opacification of Bowman’s layer.

an autosomal dominant dystrophy that is marked by geographic opacification of Bowman’s layer, beginning as fine granular deposits that evolve into confluent opacities over time. Reis-Buckler’s arises during childhood or adolescence and often leads to frequent recurrent erosions and marked, progressive visual loss (Figure 1). The recurrent erosions can be managed in a manner similar to in EBMD, but the progression of opacification may necessitate extensive excimer laser phototherapuetic keratectomy. Despite initial clearing of the visual axis, these patients often have recurrence of opacities requiring additional treatment. Histological examination demonstrates that the normally acellular collagenous Bowman’s is disrupted, noncontiguous, or absent and is replaced with fibrocellular tissues. The opacities in Bowman’s layer are band-shaped, granular, and stain red with Masson’s trichrome. Similar to other granular dystrophies, these lesions appear as rodshaped bodies under electron microscopy. Multiple mutations of the human transforming growth factor inducible gene (TGFBI) – previously known by the misnomer, keratoepithelin – on chromosome 5q31 may lead to the Reis-Buckler’s phenotype. The most common mutations are Arg124 or Arg555 which are both thought to affect the solubility and stability of the TGFBI protein. Thiel-Behnke Initially classified as a variant of Reis-Buckler’s, ThielBehnke, or corneal dystrophy of Bowman’s type 2 (CDB-2), at times is similar in its clinical appearance but is always distinct in its histological and electron microscopic appearance. Also autosomal dominantly inherited, ThielBehnke has a slightly later onset of recurrent erosions, typically in the second decade of life, and vision loss is more slowly progressive. At the slit lamp, this dystrophy typically appears as honeycombed-shaped subepithelial

228

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

opacities. It can be managed similarly to Reis-Buckler’s and recurrence after keratectomy is infrequent. Examination of pathologic specimens reveals only weakly positive Masson’s trichome staining, in contrast to the strongly positive staining of Reis-Buckler’s. The pathognomonic features are the 8–10 nm curly fibers identified by electron microscopy. Similar to Reis-Buckler’s, this dystrophy is sometimes linked to the TGFBI gene. However, there is genotypic heterogenicity for Thiel-Behnke as evidenced by the discovery that mutations on chromosome 10q23-q24 can also lead to this phenotype. Stromal Dystrophies Lattice

There are three major types of lattice dystrophy and all are unified by the appearance of lattice lines on slit-lamp examination and amyloid deposition on histological examination. Lattice type 1. In type 1 lattice, or Biber–Haab–Dimmer corneal dystrophy, amyloid is deposited in the cornea but is not found elsewhere in the body. It can vary in its appearance, and there is often progression from round, ovoid and white, or small, filamentous, and refractile anterior stromal lesions to more nodular, threadlike, and thicker linear lesions that extend into deep stroma (Figure 2(a)). Initially, the stroma between lesions remains clear, but over time these spaces opacify and assume a ground-glass appearance. The limbus is typically spared. Signs of lattice dystrophy most often appear in early childhood, but symptoms of surface erosions, irregular astigmatism, and vision loss usually begin in the second or third decades of life. Recurrent erosions can be frequent and can be managed as described previously. Some authors believe that phototherapeutic keratectomy should be avoided (as well as laser-assisted in situ keratomileusis (LASIK) and photorefractive keratectomy (PRK)) because excimer laser, within the UV light spectrum, may trigger activation of TGFB and exacerbate this condition. For severe vision loss due to opacification, penetrating and lamellar keratoplasty may be warranted. Recurrence may occur in these grafts but presents differently than the primary lesions. Specimens are positive for amyloid and stain with Congo red, periodic acid-Schiff ’s reagent (PAS), Masson’s trichrome, and thioflavine-T fluorochrome, and are metachromatic with crystal violet (Figure 2(b)). As with all amyloid, they demonstrate apple-green birefringence under polarized light. Lattice type 2. Type 2 lattice, also known as Meretoja’s, familial, or Finnish amyloidosis syndrome, is characterized by both systemic and corneal amyloid deposition. Typically seen in families of Finnish, European, or Japanese origin, it is usually asymptomatic until early adulthood. Corneal slit-lamp examination shows more peripheral

Figure 2 (a) Slit-lamp photograph of lattice type 1 dystrophy. Arrow indicates amyloid deposition. (b) Congo red stain of cornea in lattice type 1 demonstrating amyloid deposition. Afshari, N. A., Mullally, J. E., Afshari, M. A. et al. (2001). Survey of patients with granular, lattice, avellino, and Reis-Bu¨cklers corneal dystrophies for mutations in the BIGH3 and gelsolin genes. Archives of Ophthalmology 119: 16–22. ã 2001 American Medical Association. All rights reserved.

deposits, with fewer and finer lattice lines, and a primarily sub-Bowman’s location of deposition. Patients begin to experience corneal changes in the third decade of life, but symptoms of reduced corneal sensation and frequent recurrent erosions are uncommon until the fifth decade. Overall, the visual prognosis is better than in type 1 with many patients not developing visual loss until late in the course of disease. While there is a decreased severity of corneal disease, the systemic manifestations can be serious and include dry, itchy skin, intermittent proteinuria, and cardiac abnormalities. Patients may develop severe mask-like facial paresis (loss of facial muscle motor function), blepharochalasis (inflammation

Corneal Dystrophies

of eyelid), protruding lips, and pendulous ears from amyloid deposition and secondary muscular dysfunction. Unique among the lattice dystrophies, type 2 does not arise from a mutation in the TGFBI gene. Instead, it has been linked to a dominant mutation of the gelsolin gene on 9q32–34 that encodes an amyloid precursor. Lattice types 3 and 3a. Type 3 lattice is the least severe and has an autosomal recessive pattern of inheritance. The lattice lines in type 3 are thicker and ropier in appearance, and later in onset. Vision is often not affected until the sixth or seventh decade of life, and recurrent erosions are rare. Like type 1, type 3 is associated with a defect in the TGFBI gene. Type 3a is similar in presentation to type 3 but follows an autosomal dominant inheritance pattern.

229

(a)

Granular Granular type 1

The most common of the stromal dystrophies, granular dystrophy is named for its typical appearance as crumblike, discrete, grainy opacities in the anterior stroma. Initially, these opacities are fine dots or lines and with time they assume their more characteristic appearance of anterior stromal drops, rings, or crumbs with intervening clear spaces. As the dystrophy progresses, the lesions coalesce and extend into deeper stroma, and the intervening spaces take on a ground-glass appearance. The lesions rarely extend to the limbus (Figure 3). Symptoms of vision loss do not usually occur until the fifth decade. Despite the anterior location of the dystrophy, recurrent erosions are rare. There are exceptions to this mild course, and some variants of the disease cause earlier and more severe vision loss. For patients with severe vision loss, management options include superficial keratectomy, phototherapeutic keratectomy, and lamellar or full thickness keratoplasty depending on the depth of involvement. Recurrence may occur even in full thickness grafts necessitating repeat procedures. With recurrence, the lesions may appear similar to primary disease or may consist of subepithelial lesions emanating from the graft periphery. The pathology of granular type 1 shows eosinophilic, rod-shaped, trapezoidal hyaline deposits in the stroma that are bright red on Masson’s trichome and weakly PAS positive. On electron microscopy, the amorphous material appears rod or trapezoidal in shape. Granular dystrophy type 1 is also due to a defect in the TGFBI gene and at least two causative mutations have been identified. Granular type 2

Previously known as Avellino dystrophy (because the initial reports were of families from Avellino, Italy) this variant of granular dystrophy is characterized both by granular deposits and lattice lines or stellate deposits. Early lesions are often ring-like and anterior and can appear similar to type 1 granular lesions. As the disease

(b)

(c) Figure 3 Slit-lamp photographs of mild (a) and moderate (b) granular type 1 corneal dystrophy. (c) Masson’s trichrome staining of hyaline deposits in granular dystrophy. Afshari, N. A., Mullally, J. E., Afshari, M. A. et al. (2001). Survey of patients with granular, lattice, avellino, and Reis-Bu¨cklers corneal dystrophies for mutations in the BIGH3 and gelsolin genes. Archives of Ophthalmology 119: 16–22. ã 2001 American Medical Association. All rights reserved.

progresses, deeper lesions appear and the lattice-like lesions are observed. With progression, the intervening clear spaces between lesions assume a hazy appearance.

230

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Symptoms may be more severe and the onset earlier than in type 1 granular dystrophy. Many patients experience vision loss by adolescence, but vision rarely drops below the 20/70 level. Recurrent erosions and related ocular discomfort occur in a minority of patients, but more frequently than in type 1 granular patients. Not every patient displays clinical evidence of both granular and lattice lesions, but on histologic examination these patients uniformly demonstrate both hyaline and amyloid deposits, the distinguishing pathologic findings of the two respective diseases. As both granular and lattice dystrophies can arise from dominant TGFBI gene defects, it is not surprising that TGFBI mutations have been found to be responsible for this dystrophy which displays characteristics of both diseases. The autosomal dominant inheritance of type 2 granular dystrophy has a high penetrance, but phenotypic expressivity is quite variable between family members. Macular The least common of three major stromal dystrophies, the autosomal recessive Macular dystrophy is also the most severe. Its clinical appearance is unique, seen at the slitlamp as diffuse, cloudy, gray–white lesions that begin superficially and centrally and then spread downward and outward. By late adolescence, these lesions may progress to involve the entire cornea from limbus to limbus and from anterior stroma to the endothelium. Anteriorly protruding nodules may cause highly irregular astigmatism, and Descemet’s membrane involvement may produce corneal guttata. Vision loss can begin in the first decade of life and may be severe by the third decade. Recurrent erosions are more frequent and may also begin at an early age. Due to photophobia, many patients seek treatment with tinted lenses, phototherapeutic keratectomy, or lamellar and penetrating keratoplasty. Unfortunately, recurrence is common and may lead to significant vision loss even in penetrating keratoplasty recipients. This recurrence most often begins in the periphery near the graft–host junction, and has been found to be more common in smaller-sized grafts. The key histopathologic finding of Macular dystrophy is glycosaminoglycan accumulation. This accumulation may occur both intracellularly and extracellularly and is found between stromal lamellae, subepithelially, and within keratocytes and endothelial cells. The deposits stain with Alcian blue, colloidal iron, metachromatic dyes, and PAS. The cornea may have decreased overall thickness with a thinned or absent Descemet’s membrane and an epithelium that is stretched thin over anterior stromal deposits. Unique in its autosomal recessive inheritance, Macular dystrophy is linked to various mutations on chromosome

16q2 and the CHST6 gene that encodes enzymes of keratan sulfate synthesis. This keratan sulfate defect is systemic and the varied expression of keratan sulfate in serum and cornea has led to the classification of several types of Macular dystrophy. In type 1, keratan sulfate is completely undetectable in the cornea and serum. In type 1a, the only detectable keratan sulfate is found within keratocytes, and in type 2 corneal keratan sulfate is present but levels are diminished. Rare Stromal Dystrophies Gelatinous drop-like dystrophy Presenting in the first decade of life with photophobia, foreign body sensation and decreased vision, this autosomal recessive dystrophy, also known as familial subepithelial amyloidosis, resembles band keratopathy. The opacities are subepithelial and anterior stromal, and may, over time, assume a mulberry-like gelatinous appearance. Keratectomy and keratoplasty may be performed with variable success due to recurrence. This dystrophy is linked to mutations in the M1S1 gene on chromosome 1p31. These defects lead to the accumulation of amyloid from a truncated surface glycoprotein.

Schnyder’s crystalline dystrophy Also known as Scandinavian dystrophy due to a reported cohort of patients in central Massachussets of Scandinavian origin, this autosomal dominant dystrophy occurs within the first year of life as central, crystalline, anterior stromal lesions that are either discoid, ring-like, or geographic in distribution. A significant minority may have a noncrystalline form which occurs later in life. As the lesions progress, they involve more posterior stroma. By the third decade a prominent arcus lipoides is seen, and by age 40 most patients have full thickness corneal opacities. While there is no correlation between this disease and elevated serum cholesterol, patients’ lipid and cholesterol levels should be checked as early arcus, before age 50, and can be an indicator of hyperlipidemia. Important diseases in the differential diagnosis of central crystalline corneal deposits should also be ruled out and these include Bietti’s corneoretinal dystrophy, cystinosis, and dysproteinemias such as multiple myeloma. Pain is infrequent because many of these patients develop decreased corneal sensation. Vision loss is the most common presentation and may be treated with lamellar or penetrating keratoplasty. Histologic examination of the corneal buttons reveal that the crystals are composed of cholesterol and lipid and the pathogenesis is believed to involve a defect in local lipid metabolism or transport. The genetic defect has been linked to chromosome 1p.

Corneal Dystrophies

Central cloudy dystrophy

First described by Francois, this autosomal dominant or, rarely, sporadic dystrophy appears as bilaterally symmetric central and deep stromal cloudy, gray–white, illdefined, snowflake-like lesions. The appearance is similar to posterior crocodile shagreen, and, similarly, is rarely symptomatic. Congenital hereditary stromal dystrophy

This dystrophy, which is often static or slowly progressive, has been described in several families and occurs at birth as a diffusely hazy cornea of normal thickness. The lesions are diffuse, bilateral, small, and located primarily in the anterior stroma. This generalized opacification may lead to profound vision loss and early penetrating keratoplasty may be warranted to prevent amblyopia. The pathogenesis involves disorganization of corneal lamellae with randomly arranged collagen fibrils and loss of corneal transparency. The inheritance pattern is autosomal dominant and a defect has been identified in the decorin gene that encodes a dermatan sulfate proteoglycan. Posterior amorphous dystrophy

Remarkable for central stromal thinning without ectasia or astigmatism, this autosomal dominant dystrophy begins in childhood and slowly progresses. It is characterized by bilateral gray sheets in the deep stroma extending to the limbus. The stroma may thin to about 300mm, but vision is only mildly affected. In addition to stromal thinning, histologic examination demonstrates a thick collagenous layer posterior to Descemet’s membrane as well as relative absence of stromal keratocytes. The endothelium is unaffected. Fleck dystrophy

Fleck dystrophy is usually discovered as an incidental finding, because it is asymptomatic in the majority of patients (recurrent erosions can rarely occur). It is autosomal dominant, may be present at birth or arise in infancy, and is rarely progressive. On slit-lamp examination, small white flecks can be seen in all stromal layers and represent swollen keratocytes with cytoplasmic vesicles due to membrane-bound vacuoles of lipid and mucopolysaccharide. On electron microscopy, flecks stain with oil-red O. A causative mutation on chromosme 2q35 in the PIP5K3 gene has been identified. This gene codes for an enzyme involved in post-Golgi vesicle processing of protein and lipids.

dystrophy that directly involves Descemet’s membrane and the endothelium, and may indirectly impact all layers of the cornea. The classic lesion of Fuchs’ is the central guttae, but a controversial nonguttate form has been described. Over time, the guttata spread peripherally and coalesce. Descemet’s may assume a thickened, grayish, irregular appearance, and this may eventually mask the guttata. If corneal edema develops, it begins posteriorly as evidenced by Descemet’s and deep-stromal wrinkles. As the corneal edema progresses, the stroma thickens and microcystic epithelial edema can usually be appreciated by the time the cornea thickness has increased by 100 mm. Later in the course of the disease, the cornea takes on a ground-glass appearance (Figure 4). Fuchs’ guttata are 2.5 times more likely to develop in women than men, and women are nearly 6 times more likely to develop Fuchs’ corneal edema. Symptoms of blurring are often first seen in during the morning hours because the cornea swells overnight due to decreased evaporation and the increased hypotonicity of the tear film. Some patients find relief from hypertonic solutions, decreasing intraocular pressure, or by facilitating tear evaporation with external devices such as blow dryers. Recurrent erosions often occur in advanced cases due to edema-induced anterior basement membrane-like lesions and ruptured epithelial bullae. Repeat epithelial defects may lead to fibrous scarring and neovascularization. These epithelial lesions may be managed with bandage contact lenses. As mentioned previously, many patients require management with penetrating or endothelial keratoplasty, especially after other intraocular surgeries which may lead to further loss of endothelial cell viability. The pathology of Fuchs’ is well characterized and primarily consists of changes in endothelial cell architecture and number, and abnormalities of posterior Descemet’s membrane. The number of endothelial cells decreases while the remaining cells show an increase in size, a less hexagonal and more irregular shape, and an

Posterior Dystrophies Fuchs’ Endothelial By far the most common corneal dystrophy to lead to corneal transplantation, Fuchs’ is an autosomal dominant

231

Figure 4 Slit-lamp photograph of Fuch’s dystrophy.

232

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

eccentricity of cellular nuclei. This can be seen in vivo with specular microscopy as a decreased endothelial cell density, polymegathism, and cellular polymorphism. On electron microscopy the cells look more fibroblastic with an increase in rough endoplasmic reticulum, lysosomes, vacuoles, and cytoplasmic filaments. The posterior nonbanded zone of Descemet’s membrane is thinner than normal, on an average less than 2 mm compared to 8 mm in a normal adult population. An abnormal PAS-positive posterior collagenous layer significantly increases the overall thickness of Descemet’s to 2–3 times that of normal corneas, and this layer is contiguous with guttae that jut out posteriorly, thinning the overlying endothelial cells and pushing their nuclei aside into the intervening valleys. With corneal edema, the corneal lamellae thicken, and histologically this is evidenced by a decrease in artifactual clefting. The genetic defects leading to Fuchs’ are numerous and the identified mutations have been linked to chromosome 1p32-1p34, as well as chromosomes 7, 15, 17, and X.

Posterior Polymorphous Membrane Typically beginning in the second or third decade of life, this autosomal dominant dystrophy of Descemet’s and the endothelium is named for the diversity of clinical findings seen on slit-lamp examination. The hallmark lesions are posterior vesicles, and these may be accompanied by bands or by diffuse opacities in 50% and 10% of the cases, respectively. The vesicles, misleadingly, look like transparent cysts of Descemet’s membrane. The bands, if present, are usually horizontal, have scalloped edges, and most commonly occur along the inferior paracentral cornea. The bands can be distinguished from tears or folds of Descemet’s because they lack tapered edges. Diffuse opacities range in size from 500 to 2000 mM, have a peau d’orange texture and are associated with adjacent posterior stromal haziness. As in Fuchs’, guttata may be seen and corneal edema can develop. Rarely, this dystrophy is associated with corneal steepening or ectasia. For the majority of patients the dystrophy causes no symptoms, is nonprogressive and may be picked up incidentally in the second or third decade of life. However, there is a wide spectrum of disease severity and some cases are vision threatening. In fact, severe corneal edema and clouding may be present at birth or early childhood, and this dystrophy needs to be considered in the differential diagnosis of congenital corneal clouding. While most cases are bilateral, there may be a marked degree of asymmetry in its presentation. In addition to corneal symptoms, patients may develop high intraocular pressure as a result of the peripheral anterior synechiae. For patients with corneal decompensation, penetrating keratoplasty has been the treatment of choice. The success

rate of keratoplasty is much higher for patients without significant broad preoperative synechiae or high pressure, and may be extremely low for patients with these problems. Recurrence after keratoplasty may occur in the form of a retrocorneal membrane. The pressure elevations associated with this dystrophy are difficult to manage medically or surgically. The pathologic findings demonstrate layered-endothelial cells that have assumed epithelial characteristics such as microvilli and rapid growth in culture, and stain with epithelial cell markers such as cytokeratin (CK), pancytokeratin, and CK7 (a glandular epithelial marker). Descemet’s membane is irregular with a typically normal anterior-banded zone but an absent or markedly abnormal posterior nonbanded zone. Much of the posterior zone is replaced by heterogenous collagenous components that comprise a 15–25-mM-thick posterior collagenous layer. Posterior synechiae may be found in up to a quarter of patients, and may be accompanied by iris defects including atrophy and corectopia. On specular microscopy, the various posterior polymorphous lesions can be further examined. The vesicles appear as well-demarcated dark round areas with lighter ridges and dots. The bands have shallow hills and valleys composed of confluent vesicles, and the diffuse lesions are well-demarcated reflective areas with enlarged, pleomorphic, indistinct endothelial cells that are surrounded by more normal appearing endothelial cells. At least two different loci are associated with posterior polymorphous dystrophy and they are found on chromosomes 20q11 and 1p34.3–p32. The resultant defects may affect the production of type VIII collagen, the predominant component of the anterior-banded zone. Congenital Hereditary Endothelial Dystrophy Present at birth or in the early postnatal period, this dystrophy is bilateral, symmetric, and diffuse, with corneal haze spanning from limbus to limbus. The cornea is very thick – 2–3 times normal – and there are no other associated anterior segment defects. Rarely, this corneal edema begins later in infancy or early childhood. Treatment consists of early keratoplasty. Examination of the excised corneal buttons reveals reduction, absence, or degeneration of endothelial cells and diffuse corneal edema. As in all posterior dystrophies, there is an abnormal posterior nonbanded zone which merges into a posterior collagenous layer. Inheritance is most often autosomal recessive and linked to chromsome 20p13, but rarely may be inherited as a dominant trait on chromosome 20p11.2–q11.2. See also: Corneal Endothelium: Overview; Corneal Epithelium: Cell Biology and Basic Science; Corneal

Corneal Dystrophies Epithelium: Transport and Permeability; Corneal Epithelium: Wound Healing Junctions, Attachment to Stroma Receptors, Matrix Metalloproteinases, Intracellular Communications; Corneal Imaging: Clinical; Corneal Scars; The Corneal Stroma; Cornea Overview; Imaging of the Cornea; Regulation of Corneal Endothelial Cell Proliferation; Regulation of Corneal Endothelial Function.

Further Reading Afshari, N. A., Mullally, J. E., Afshari, M. A., et al. (2001). Survey of patients with granular, lattice, avellino, and Reis-Bu¨cklers corneal dystrophies for mutations in the BIGH3 and gelsolin genes. Archives of Ophthalmology 119: 16–22. Afshari, N. A., Li, Y. J., Pericak-Vance, M. A., et al. (2009). Genome wide linkage scan in Fuchs endothelial corneal dystrophy. Investigative Ophthalmology and Visual Science 50: 1093–1097. Bron, A. J. (2000). Genetics of the corneal dystrophies: What we have learned in the past twenty-five years. Cornea 19: 699–711. Dinh, R., Rapuano, C. J., Cohen, E. J., et al. (1999). Recurrence of corneal dystrophy after excimer laser phototherapeutic keratectomy. Ophthalmology 106: 1490–1497.

233

Holland, E. J., Daya, S. M., Stone, E. M., et al. (1992). Avellino corneal dystrophy: Clinical manifestations and natural history. Ophthalmology 99: 1564–1568. Kang, P. C., Klintworth, G. K., Kim, T., et al. (2005). Trends in the indications for penetrating keratoplasty, 1980–2001. Cornea 24: 801–803. Krachmer, J. H., Mannis, M. J., and Holland, E. J. (2005). Cornea, 2nd edn. Philadelphia, PA: Elselvier Mosby. Kanski, J. J. (2003). Clinical Ophthalmology: A Systematic Approach, 5th edn. Edinburgh: Butterworth Heinemannn. Stone, E. M., Mathers, W. D., Rosenwasser, G. O., et al. (1994). Three autosomal dominant corneal dystrophies map to chromosome 5q. Nature Genetics 6: 46–51. Vasilliki, P. and Colby, K. (2008). Genetics of anterior and stromal corneal dystrophies. Seminars in Ophthalmomlogy 23: 9–17. Yanofff, M. and Duker, J. S. (2004). Ophthalmology, 2nd edn. St. Louis, MO: Mosby.

Relevant Websites http://www.emedicine.com – eMedicine: Ophthalmology Article. http://www.nei.nih.gov – Facts about the Cornea and Corneal Disease (NEI Health Information). http://www.cornealdystrophyfoundation.org – The Corneal Dystrophy Foundation.

Corneal Imaging: Clinical S Garg and R F Steinert, University of California, Irvine, Irvine, CA, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Confocal – Noncontact imaging modality based on the principle that light passed through an aperture and focused by an objective lens onto an area of interest. Optical coherence tomography (OCT) – Noncontact imaging modality in which back-reflected light or backscattered infrared light from internal tissue microstructure is analyzed to achieve two or three-dimensional cross-sectional tomographic images of optical reflectivity. Phototherapeutic keratectomy (PTK) – Surgery using a laser to treat various ocular disorders by removing tissue from the cornea.

Introduction Assessment of anterior segment structures is an integral part of the ophthalmic evaluation. Clinical techniques for examining the human cornea in vivo have greatly expanded over the last several decades. The clinician’s armamentarium includes slit lamp biomicroscopy, specular microscopy of the endothelium, computed corneal topography, high-frequency ultrasound, anterior segment optical coherence topography (OCT), and confocal microscopy. Advanced anterior segment imaging is a routine part of the anterior segment physicians’ practice. Confocal microscopic evaluation of the cornea in vivo began in the late 1980s. The invention of OCT in the early 1990s initially centered on retinal imaging and was subsequently modified for anterior segment applications. In the clinical setting, confocal microscopy and anterior segment OCT are noninvasive devices that allows for examination of normal and diseased corneas and anterior segment structures such as the angle, iris, and lens, aiding in both routine patient care and in managing complex pathology.

Confocal Microscopy Historical Overview Marvin Minsky invented the confocal microscope in 1955. His novel invention imaged tissue parallel to its surface.

234

Traditionally, images of the tissue were oriented perpendicular to their surface. Minsky exploited the pinhole effect to accomplish his goal. He proposed that both the illumination (condenser) and observation (objective) systems could be focused on a single point (both have common focal points, and thus the name confocal). Using a pinhole eliminates unwanted optical artifacts from light reflected above and below the tissue of interest, improving image quality. However, this increased resolution comes at a cost of a small field of view. A full field of view is accomplished by scanning. In 1968, the first tandem scanning confocal microscope (TSCM) was developed. This improvement used a rotating Nipkow disk to simultaneously scan multiple points on a stationary specimen. In 1985, the confocal was first used to describe imaging of human corneas ex vivo and rabbit corneas in vivo. Around the same time of development of the TSCM, Svishchev introduced the scanning two-sided mirror confocal microscope. This was later modified by Thaer to enable real-time scanning, the precursor to the modern slit scanning confocal microscope (SSCM). Current commercially available confocal microscopes include the Confoscan 4 (Nidek Technologies, Gamagori, Japan), Confoscan P4 (Tomey Corporation, Cambridge, MA, USA), Microphthal (Helmut Hund, Wetzlar, Germany), and the Heidelberg Retina Tomograpy II Rostock Cornea Module (Heidelberg Engineering, GmBH, Germany). How it Works Confocal microscopy is based on the principle that light passed through an aperture and focused by an objective lens onto an area of interest. The reflected light is then focused by a second objective lens through a second aperture to eliminate out of focus light. The ability of the system to discriminate light that is outside the focal plane results in images of higher X-, Y-, and Z-axis resolutions. The drawback is a small field of view (Figure 1). Moving the confocal system (scanning) over the stationary specimen allows for larger fields of view. The Z-axis resolution of the confocal microscope permits the dynamic scanning capability of the instrument, allowing in vivo corneal imaging without the need for stains or dyes. With computerized three-dimensional reconstruction, this technology has improved lateral and axial resolution to 1–6 and 4 –15 mm, respectively, and increased magnification up to 600. Image quality is affected by image contrast, the light source, the scanning method, the path of light in the cornea, and the optics of the objective.

Corneal Imaging: Clinical

235

Focal plane

First pinhole

Out of focus

Second pinhole

Cornea

Observer or camera Figure 1 Schematic of the optical principles of confocal microscopy. White light that passes through the first pinhole is focused on the focal plane in the cornea by the condensing lens. Returning light is diverted through the objective lens and a conjugate exit pinhole and reaches the observer or camera. Scattered out of focus light from below or above the focal plane (broken lines) is greatly limited by the pinholes and does not reach the observation system. Reprinted with permission from BJO.

Figure 2 Superficial epithelium.

Image separation (depth) is recorded by the movement of the objective between images. Water immersion objective lenses of high numerical apertures are typically used, as they eliminate surface reflections and provide good depth resolution. This requires a short working distance. In clinical practice, subject preparation for the scan is of great importance. Topical anesthesia, patient counseling of the short working distance, bright illumination, and use of coupling agents all aid in capturing useful images. Maintaining a perpendicular orientation to the corneal surface is necessary to avoid oblique sectioning. Figure 3 Basal epithelium.

Clinical Applications The normal cornea

The normal human cornea consists of five layers: epithelium, Bowman’s layer, stroma, Descemet’s membrane, and endothelium. All of these layers, with the exception of Descemet’s membrane, can be imaged by confocal microscopy. It is important to note that the more a cellular component reflects light, the brighter the image will appear on a confocal scan. From front to back, the corneal epithelium is composed of superficial, wing, and basal cells, with a normal thickness of approximately 50 mm. In the superficial layers, cells appear flat and polygonal with hyperreflective nuclei. Wing cells appear uniform in shape and size with dark nuclei. These cells are generally larger than basal cells, but smaller than superficial cells. The basal cells are smaller, more uniform in size, and have bright borders and highly reflective cell nuclei (Figures 2 and 3). Between the basal epithelium and Bowman’s layer are corneal nerves that appear as beaded, well-defined linear

Figure 4 Basal nerve plexus.

branching structures with homogeneous reflectivity (Figure 4). Bowman’s layer appears as an amorphous homogenous layer in the normal cornea. It is acellular, with randomly

236

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

dispersed collagen fibrils. On confocal microscopy, Bowman’s layer is imaged poorly without the aid of anatomical landmarks such as the highly reflective subbasal nerve plexus due to an average thickness of only 15 mm. In the corneal stroma, which typically is about 500-mm thick, keratocyte nuclei appear hyperreflective against a dark background. Confocal images exhibit poor reflectivity of stromal keratocyte cytoplasm, cell boundaries, and collagen substance. In the anterior stroma keratocyte nuclei appear as distinct, bright, and oval-round in random orientation. In the mid-stroma, keratocytes exhibit a more regular oval shape transitioning to elongated spindleshaped as the scan approaches the posterior stroma. Additionally, hyperreflective nerve fibers are sporadically seen coursing within the anterior and mid-stroma, but they are absent in the posterior stroma (Figures 5 and 6). Descemet’s membrane thickens throughout life. It is the basement membrane of the inner layer of endothelial cells. Because of the lack of cellularity and thinness, it is poorly captured by confocal microscopy.

Corneal endothelial cells are single layered, normally characterized by a regular hexagonal hyperreflective cell body, void of nuclei, and surrounded by hyporeflective borders (Figure 7). Pachymetry Confocal microscopy can be used to determine corneal thickness through a function known as confocal microscopy through focusing (CMTF) on the TSCM. This is accomplished by focusing in the Z-axis and determining the amount of light backscattering which in turn is plotted as an intensity profile curve. The differences in scattering of the various corneal layers allows for determination of each layer’s location. Overall, this method for determination of corneal thickness offers good repeatability, especially for determination of thin layers such as the epithelium and Bowman’s membrane. Nontandem models such as the Nidek Confoscan 4 utilize a contact ring at the limbus.

Applications in Pathology

Figure 5 Anterior stroma.

The treatment of infectious keratitis can be challenging. The golden standard for diagnosis of infectious keratitis is light microscopic examination and culture of corneal scrapings. However, confocal microscopy is a very useful tool in helping with diagnostic quandaries. Confocal microscopy is particularly helpful when Acanthamoeba or fungal elements are suspects as etiologies. Bacteria (2 mm) can theoretically be visualized; however, given their small size (near the typical resolution for confocal microscopy), clinical distinction is not possible. Fungal infections generally image as hyperreflective, elongated, filaments, or budding yeasts (Figures 8 and 9). In its cystic form, Acanthamoeba appears as a highly reflective, round, or ovoid double-wall structure with a diameter of 10–25 mm (Figure 10). Radial keratoneuritis may appear as irregularly enlarged nerve fibers.

Figure 6 Deep stroma.

Figure 7 Endothelium.

Corneal Imaging: Clinical

Figure 8 Fusarium keratitis.

237

Figure 11 Iridocorneal endothelial syndrome (ICE). Note the epithelial-like endothelial cells with hyperreflective nuclei, which is characteristic of ICE syndrome in which the endothelium grows across the angle and iris. Notice irregular rows and shape of the endothelial cells.

Figure 9 Fusarium keratitis.

Figure 12 Posterior polymorphous membranous dystrophy (PPMD). Note the hyporeflective vesicular changes which are characteristic of PPMD in which the endothelium has areas of large and irregularly shaped cells.

dystrophies, endothelial pathology, as well as evaluation of corneal ectatic disorders (Figures 11 and 12).

Anterior Segment OCT Historical Overview

Figure 10 Acanthaeomba.

Other applications

Confocal microscopy has also been used to research and clinically evaluate the effects of refractive surgery on the cornea, evaluation of corneal deposits and stromal

David Huang and colleagues developed OCT in the early 1990s. Joseph Izatt and colleagues first demonstrated corneal and anterior segment OCT in 1994. Over the next decade, the majority of developments in OCT technology focused on retinal imaging. The use of longer wavelengths, telecentric transverse scanning, and very high-speed axial scanning with a grating-based rapid scanning optical delay (RSOD) mechanism in the

238

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

reference arm allowed for improved imaging of the anterior segment. In October 2005, the Food and Drug Administration in the United States approved the Zeiss Visante anterior segment OCT, the first commercially available OCT device designed for the anterior segment.

SLD Sample

50/50

How it Works Anterior segment OST (AS-OCT) is a noncontact imaging technique based on Michelson low coherence inferometry. In conventional interferometry with long coherence length (laser interferometry), interference of light occurs over a distance of meters. In OCT, this interference is shortened to a distance of micrometers. Light in an OCT system is broken into two arms – a sample arm (containing the item of interest) and a reference arm (usually a mirror). The combination of reflected light from the sample arm and reference light from the reference arm gives rise to an interference pattern, but only if light from both arms have traveled the same optical distance (same meaning a difference of less than a coherence length). By scanning the mirror in the reference arm, a reflectivity profile of the sample can be obtained (time domain OCT). Areas of the sample that reflect with greater intensity will create greater interference than areas that do not. Any light that is outside the short coherence length will not interfere. This reflectivity profile, called an A-scan, contains information about the spatial dimensions and location of structures within the item of interest (Figure 13). This is analogous to B-scan ultrasonography; however, OCT uses light as compared to sound waves. OCT was initially used for retinal evaluation, where images are optimized with an 820-nm light. AS-OCT evolved from retinal OCT. For the Visante AS-OCT, Zeiss uses a longer wavelength (1310 nm) that allows for greater penetration through tissues that scatter light intensely such as the sclera and limbus, which in turn permits visualization of anterior segment structures such as the angle, ciliary body, and ciliary sulcus. Approximately 90% of the 1310-nm light is absorbed prior to reaching the retina allowing for the AS OCT to be used as a higher power than retinal OCT. This results in realtime imaging and decreased motion artifacts. Currently, there are three commercially available anterior segment OCT devices: the Visante OCT (Visante OCT, Carl Zeiss Meditech Inc, Dublin, CA, USA), the Slit Lamp OCT (SL-OCT, Heidelberg Engineering GmbH, Heidelberg, Germany), and the Optovue (RTVue with Cornea-Anterior Module, Freemont, CA, USA). The Visante OCT provides high-resolution corneal scans, anterior segment scans (anterior-chamber depth, anterior-chamber angle, angleto-angle distance), and pachymetry maps. It is reported to have an axial resolution of 18 mm and a transverse resolution of 60 mm. The SL-OCT is essentially a slit lamp biomicroscope-mounted OCT device allowing similar

Detector Reference Demodulator

AD

Computer

Figure 13 Schematics of the basic fiber-optic OCT system. Light from a superluminescent diode (SLD) is launched into a single mode-optical fiber. The light is equally split at the coupler into the sample and reference arms. Sample and reference reflections are recombined at the coupler and the interference pattern is converted to an electrical signal by the detector. The signal is demodulated and converted from analog to digital (AD) form for computer signal and image processing. To scan the reflections from various depths in the sample, the reference mirror is scanned over the equivalent range of delay. This produces a scan of sample reflectivity versus depth, also called an axial scan. From Steinert, R. E. and Huang, D. (2008). Anterior Segment Optical Coherence Tomography. Thorofare, NJ: SLACK Incorporated. Reprinted with permission from SLACK Incorporated.

measurements as the Visante OCT. The potential advantage of the SL-OCT is its attachment to the biomicroscope. The SL-OCT has a reported axial resolution of 25 mm and a transverse resolution of 20–100 mm. The speed of acquisition is 4 –8 frame s 1 for the Visante OCT and 1 frame s 1 for the SL-OCT. The Optovue OCT employs an interchangeable lens system for anterior-segment images. Its principal advantage is the use of spectral domain OCT technology, with a substantially higher resolution. However, scans are limited in width, so only a portion of the cornea can be imaged at one time, and the retina-optimized wavelength does not allow useful imaging of the angle. Clinical Applications Anterior segment OCT is a powerful tool for the clinician. It can be used to measure corneal thickness, pachymetry maps, total corneal power, corneal backscatter, angle configuration, anterior-segment tumors, anterior-segment depth, lens vault, corneal opacities, and corneal refractive implants. The following discussion pertains to applications of the Visante OCT. Keratoconus is a bilateral corneal ectasia characterized by progressive thinning and inferior protrusion. Diagnosis of this and other ectasias (e.g., pellucid marginal degeneration) is critical when evaluating patients for possible refractive surgery. Form fruste keratoconus is subclinical and can often be difficult to diagnose. Anterior segment OCT is a valuable tool in evaluating these patients

Corneal Imaging: Clinical

to aid in proper diagnosis. As both the Orbscan and Pentacam are based on slit-scanning principles, they tend to underestimate corneal thickness in keratoconic eyes. OCT has a higher resolution as compared to these other imaging modalities and therefore accurately maps the corneal thickness of normal, postoperative, and opacified corneas. The pachymetry map allows for accurate pachymetry readings over broad areas of the cornea

239

allowing for detection of subclinical thinning suggestive of corneal ectatic disorders (Figures 14 and 15). The Visante OCT can obtain high-resolution images of the cornea. The flap tool measures laser-assisted in situ keratomileusis (LASIK) flap thickness and residual stromal bed (RSB) in up to seven locations. This application allows for evaluation of LASIK flaps in uncomplicated and ectatic cases. RSB measurements are invaluable in providing

Figure 14 Normal visante OCT pachymetry map. The visante OCT allows for a global pachymetry map represented by numerical values and a corresponding color gradient. The cooler colors represent thicker pachymetry values and the warmer colors represent thinner pachymetry values.

Figure 15 Visante pachymetry map in a patient with keratoconus. Note the infero-temporal thinning represented by the warmer orange-red.

240

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

(a)

(b) Figure 16 Continued

Corneal Imaging: Clinical

S/W version: 1.0.12.1896 Patient ID: 1964500 Gender: male Age: 47 High res. corneal OD 180°

241

OS 0°

−1.49 mm

0.06 mm

1.50 mm

−2.50 mm

2.50 mm

206 µm 279 µm

240 µm 201 µm

231 µm 180 µm

207 µm 251 µm

194 µm 306 µm

(c)

S/W version: 1.0.12.1896 Patient ID: 1964500 Gender: male Age: 47 High res. corneal

OD

180°

OS 0°

−1.00 mm

0.02 mm

1.01 mm

−2.50 mm

296 µm 220 µm

2.49 mm

253 µm 144 µm

248 µm 101 µm

238 µm 119 µm

241 µm 207 µm

(d) Figure 16 Forty-seven-year-old male with a history of bilateral conventional LASIK, with subsequent wave front-guided enhancements. (a) Corneal topography of the right eye with mild irregularity noted on the anterior float. (b) Corneal topography of the left eye with marked steepening on the keratometric map, anterior irregularity, and abnormal posterior curvature consistent with postrefractive ectasia. (c) OCT of the right cornea with thick flap and thin residual stromal bed. The number anterior to the epithelium represents the location from the apex in millimeters, the first number on the endothelial aspect represents the thickness from the horizontal line anteriorly (i.e., the flap thickness), the second number on the endothelial aspect represents thickness from the horizontal line posteriorly (i.e., the residual stromal bed). (d) OCT of the left cornea with thick flap and very thin residual stromal bed. Note that centrally the flap measures 248 mm and the residual stromal bed only 101 mm. From Steinert, R. E. and Huang, D. (2008). Anterior Segment Optical Coherence Tomography. Thorofare, NJ: SLACK Incorporated. Reprinted with permission from SLACK Incorporated.

242

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

(a)

S/W version: 1.0.12.1896 Patient ID: 1950593-2 Gender: female Age: 52 High res. corneal 180°

OD

OS 0°

(b) Figure 17 Continued

information regarding the potential safety of performing an enhancement when lifting the flap. Additionally, the flap tool can be useful when evaluating a patient with postLASIK keraecastia (Figure 16(a)–16(d)). Likewise, OCT is useful in evaluations of refractive corneal inlays and intacs intracorneal ring segments (Figures 17(a)–17(c) and 18). Assessment of the depth of corneal ulcers is another valuable application of AS-OCT. Involved tissue appears hyperintense on OCT images. The clinician is able to follow the progression of keratitis by imaging the depth of involvement, the density of the infiltrate, and evaluating

for possible corneal thinning (Figures 19(a)–19(e) and 20(a)–20(e)). Penetrating (full thickness) and lamellar (partial thickness) corneal transplantation has been the focus of several technologic and surgical advancements over the past several years. The development of posterior lamellar keratoplasty (Descemet stripping endothelial keratoplasty, or DSEK), and femtosecond laser-enabled keratoplasty have dramatically changed corneal transplantation techniques. AS-OCT evaluation of these patients pre- and postoperatively is helpful in surgical planning and clinical follow-up (Figures 21–23).

Corneal Imaging: Clinical

243

S/W version: 1.0.12.1896 Patient ID: 1950627−8 Gender: female Age: 36 High res. corneal OD 180°

OS 0°

−0.24 mm

112 µm 520 µm

0.06 mm

186 µm 443 µm 4.59 mm

(c)

Figure 17 Fifty-two-year-old female with revision optics 5-mm corneal inlay. (a) Slit lamp photo. Note edge of inlay highlighted by black arrow. (b) Visualization of implant which appears dark on OCT highlighted by white arrow. (c) Flap tool used to measure inlay thickness and implant size. The yellow bars represent the flap tool indicating the depth of the inlay, and the blue bars represent the width of the inlay. From Steinert, R. E. and Huang, D. (2008). Anterior Segment Optical Coherence Tomography. Thorofare, NJ: SLACK Incorporated. Reprinted with permission from SLACK Incorporated.

S/W version: 1.1.0.Research.1991 Patient ID: Gender: male Age: 51 High res. corneal quad

OD

180°

OS 0°

−3.76 mm

474 µm 231 µm

3.37 mm

410 µm 257 µm

Figure 18 Use of flap tool to measure intacs depth. The intacs is highlighted by the white arrow and the yellow bars represent the flap tool. From Steinert, R. E. and Huang, D. (2008). Anterior Segment Optical Coherence Tomography. Thorofare, NJ: SLACK Incorporated. Reprinted with permission from SLACK Incorporated.

244

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Accurate measurement of corneal opacities is another clinical application of AS-OCT. Compared to ultrasonic pachymetry, ultrasound imaging, optical pachymetry, confocal microscopy, and optical low-coherence reflectometry, OCT has been shown to accurately map pachymetry in both normal and opacified corneas. This accuracy aids in clinical decision making with respect to ablative treatments such as phototherapeutic keratectomy, mechanical scraping and peeling, or their combination. Peripheral corneal pathologies, such as peripheral corneal and scleral melts, are also amenable to OCT scans. The longer wavelength (1310 nm) allows for full thickness imaging of the cornea and sclera through overlying opacities such as infiltrate, pannus (flap of tissue), and calcium plaques. Measurement of the depth of involvement and thickness of remaining tissue is useful for clinical and surgical planning (Figure 24(a)–24(c)). Beyond corneal applications, AS-OCT can also be utilized for glaucoma evaluations. Determination of angle configuration, visualization of normal angle structures (Figure 25(a)–25(c)), iris configuration (pupillary block configuration, pre- and postiridotomy (procedure

(a)

Figure 19 Continued

to create a hole in the iris to enhance the drainage passages blocked by a portion of the iris), plateau iris configuration, pre- and postiridoplasty, a surgical procedure where the position of the peripheral iris is changed), and evaluation of peripheral anterior synechiae closure (condition where the peripheral iris adheres to the cornea) (Figure 26(a)–26(c)) are among the many applications of AS-OCT pertaining to glaucoma.

Conclusion Confocal microscopy and optical coherence tomography imaging allow clinicians and researchers to evaluate the structure of the cornea and anterior segment at levels beyond slit lamp biomicroscopy. Each technology has its own advantages and drawbacks, with their applications generally complementing each other. Quantitative and qualitative measures of normal and pathologic states greatly improve the clinician’s ability to follow and treat.

Corneal Imaging: Clinical

S/W version: 1.1.2.1987 Patient ID: 20048609 Gender: male Age: 58 Anterior segment single OD 180°

OS

0°

(c)

S/W version: 1.1.2.1987 Patient ID: 20048609 Gender: male Age: 58 High res. corneal OD 270°

(d)

Figure 19 Continued

OS 90°

245

246

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

S/W version: 1.1.2.1987 Patient ID: 20048609 Gender: male Age: 58 High res. corneal OD

180°

OS

0°

(e) Figure 19 Fifty-eight-year-old male with misdiagnosed Fusarium keratitis. (a) Moderately dense full-thickness central corneal infiltrate (black arrow) and small hypopyon (white arrow). (b) Confocal microscopy shows multiple linear structures with branching at 45 , consistent with Fusarium species (later culture proven). (c–e) OCT showed a remarkable funnel of fibrin (white arrow) from the base of the ulcer to the pupil. From Steinert, R. E. and Huang, D. (2008). Anterior Segment Optical Coherence Tomography. Thorofare, NJ: SLACK Incorporated. Reprinted with permission from SLACK Incorporated.

Figure 20 Continued

Corneal Imaging: Clinical

S/W version: 1.0.12.1896 Patient ID: 19663061 Gender: unknown Age: 54 High res corneal

OD

271°

OS

91°

(b)

S/W version: 1.0.12.1896 Patient ID: 19663061 Gender: unknown Age: 54 High res corneal 271°

Figure 20 Continued

OS

91°

0.27 mm

(c)

OD

247

248

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

S/W version: 1.0.12.1896 Patient ID: 19663061 Gender: unknown Age: 54 High res. corneal quad

OD

270°

OS 90°

0.32 mm

(d)

S/W version: 1.0.12.1896 Patient ID: 19663061 Gender: unknown Age: 54 High res. corneal quad

OD

180°

OD

270°

0° 225°

90° 315°

OD

OD

45°

135°

0.50 mm

(e) Figure 20 Fifty-four-year-old female with severe infectious keratitis. Gram stain of the corneal scrapings showed gram-negative rods, and cultures were positive for Pseudomonas. (a) Large infiltrate with large hypopyon. (b) Vertical orientation of OCT shows 60–70% anterior chamber hypopyon. Note white arrow representing interface between cornea (anterior) and hypopyon (posterior). (c) Significant stromal thinning with 270 mm shown by caliper tool (blue bar). Note that normal corneal thickness is about 540 mm. (d) The patient was admitted and treated aggressively; 1 week later, the thinning was nonprogressive. (e) Over a 2-month period, the area of corneal thinning gradually improved to nearly normal thickness and the anterior chamber hypopyon regressed as shown in this high-resolution corneal quadrant image. Each image represents a corneal section through a particular axis which is shown above the image. From Steinert, R. E. and Huang, D. (2008). Anterior Segment Optical Coherence Tomography. Thorofare, NJ: SLACK Incorporated. Reprinted with permission from SLACK Incorporated.

Corneal Imaging: Clinical

S/W version: 2.0.1.88 Patient ID: Gender: unknown Age: 60 Enhanced high res> corneal

OD

180°

249

OS

0°

−0.56 mm

0.95 mm

451 µm 203 µm

486 µm 221 µm

−2.73 mm

483 µm 1 74 µm

Figure 21 DSEK (note attached graft with flap tool showing thickness of DSEK lenticule and compact overlying stroma).

S/W version: 2.0.1.88 Patient ID: Gender: unknown Age: 59 Enhanced high res. corneal OD 180°

OS 0°

Figure 22 Dislocated DSEK (note fluid cleft in interface (white arrow) and relative thickness of overlying stroma and DSEK lenticule).

250

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

S/W version: 1.1.2.1987 Patient ID: 1827371 Gender: male Age: 65 High res. corneal OD

180°

OS

0°

(c)

S/W version: 1.1.2.1987 Patient ID: 1827371 Gender: male Age: 65 High res. corneal 180°

(d) Figure 23 Continued

OD

OS 0°

Corneal Imaging: Clinical

S/W version: 1.1.2.1987 Patient ID: 1827371 Gender: male Age: 65 High res. corneal

OD

251

OS 0°

180°

(e)

S/W version: 1.1.2.1987 Patient ID: 1827371 Gender: male Age: 65 High res. corneal

OD

OS

0°

180°

Suture at 50% depth

(f)

Figure 23 Zig-zag shaped PKP femtosecond laser incisions. (a) Slit lamp photo showing well-aligned incision with smooth transition from donor to host. (b) Angled anterior incision is clearly visible (black arrow). (c) OCT at 1 month postoperatively shows excellent alignment of donor and host, both at the anterior and posterior surfaces (white arrows). (d) OCT at 3 months shows higher signal return at incision indicating stronger wound healing as compared to postop month one. (e) The rainbow color image highlights the zig-zag incision. (f) Suture depth is noted at 50% depth with perfect posterior tissue alignment and apposition. From Steinert, R. E. and Huang, D. (2008). Anterior Segment Optical Coherence Tomography. Thorofare, NJ: SLACK Incorporated. Reprinted with permission from SLACK Incorporated.

252

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

S/W version: 1.1.2.1987 Patient ID: 1707051 Gender: female Age: 41 High res. corneal quad OD

OS

270°

90°

0.44 mm

(a)

S/W version: 1.1.2.1987 Patient ID:1707051 Gender: female Age:41 High res. corneal quad

OD

OS

225°

45°

0.40 mm

(b) Figure 24 Continued

Corneal Imaging: Clinical

253

S/W version: 1.1.2.1987 Patient ID: 1707051 Gender: female Age: 41 High res. corneal quad

180°

270°

OS

OS

0° 225°

90° 315°

OS

45°

OS

135°

(c) Figure 24 Forty-one-year-old patient with a corneal melt. (a) Left eye showing superior corneal thinning (vertical orientation) as shown by the caliper tool (blue bar). Normal corneal thickness is about 540 mm. (b) Left eye showing supero-temporal thinning (oblique orientation). (c) Corneal quad scan of left eye showing a generalized superior corneal thinning (white arrows).

S/W version: 2.0.1.88 Patient ID: Gender: unknown Age: 68 Enhanced anterior segment single 180°

(a)

Figure 25 Continued

OD

OS 0°

254

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

S/W version: 2.0.1.88 Patient ID: Gender: unknown Age: 68 Enhanced anterior segment single OD

180°

OS

0°

38.9° 38.4°

(b)

S/W version: 2.0.1.88 Patient ID: Gender: unknown Age: 68 Enhanced anterior segment single 180°

OD

OS 0°

(c)

Figure 25 Sixty-eight-year-old male with normal angle configuration. (a) Horizontal meridian scan showing open angles. (b) Horizontal meridian scan with angle degree markers. (c) High resolution of angle at 180 .

Corneal Imaging: Clinical

255

(a)

(b)

Synechiae

(c) Figure 26 (a–c) Peripheral anterior synechiae (white arrows). From Steinert, R. E. and Huang, D. (2008). Anterior Segment Optical Coherence Tomography. Thorofare, NJ: SLACK Incorporated. Reprinted with permission from SLACK Incorporated.

See also: Penetrating Keratoplasty; Refractive Surgery.

Further Reading Chiou, A. G., Kaufman, S. C., Kaufman, H. E., et al. (2006). Clinical confocal microscopy. Survey of Ophthalmology 51: 482–500. Dhaliwal, J. S., Kaufman, S. C., and Chiou, A. G. (2007). Current applications of clinical confocal microscopy. Current Opinion in Ophthalmology 18: 300–307. Jalbert, I., Stapleton, F., Papas, E., et al. (2003). In vivo confocal microscopy of the human cornea. British Journal Ophthalmology 87: 225–236. Kaufman, S. C. and Kaufman, H. E. (2006). How has confocal helped us in refractive surgery? Current Opinion in Ophthalmology 17: 380–388.

Konstantopoulous, A., Hossain, P., and Anderson, D. F. (2007). Recent advances in ophthalmic anterior segment imaging: A new era for ophthalmic surgery? British Journal of Ophthalmology 91: 551–557. Lim, L. S., Aung, H. T., and Tan, D. T. (2008). Corneal imaging with anterior segment optical coherence tomography for lamellar keratoplasty procedures. American Journal of Ophthalmology 145: 81–90. Patel, D. V. and McGhee, C. N. (2007). Contemporary in vivo confocal microscopy of the living human cornea using white light and laser scanning techniques: A major review. Clinical Experiment Ophthalmology 35: 71–88. Schallhorn, J. M., Tang, M., Song, J. C., et al. (2008). Optical coherence tomography of clear corneal incisions for cataract surgery. Journal of Cataract and Refractive Surgery 34: 1561–1565. Steinert, R. F. and Huang, D. (2008). Anterior Segment Optical Coherence Tomography. Thorofare, NJ: SLACK.

Corneal Scars D G Dawson, Emory University School of Medicine, Atlanta, GA, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Corticosteroid – A class of steroid hormones that are produced in the adrenal cortex and are involved in a wide range of physiologic systems, such as the stress response, the immune response, regulation of inflammation, carbohydrate metabolism, protein catabolism, blood electrolyte levels, and behavior. In wound healing, its primary role is in suppressing the immune response and reducing the release of inflammatory cell mediators. Epithelial Laser in situ Keratomileusis (epi-LASIK) – An anterior surface ablative refractive surgical procedure that uses a unique epikeratome (a microkeratome with a blunt, oscillating blade) to mechanically separate the epithelium from the Bowman’s layer and stroma, while suction is applied, to make a hinged epithelial flap, similar to a traditional LASIK flap. Unlike LASIK, no sharp blades or intrastromal cuts are made. It is most similar to LASEK, but no alcohol is required. Extracellular matrix (ECM) – Any material produced by cells and secreted into the surrounding medium, but it usually is applied to the noncellular portion of living tissues. It typically is classified as basement membrane or interstitial ECM. Fibrosis – Formation of excessive fibrotic scar tissue in an organ or specific tissue of the body. Hypertrophic scar – An abnormal, excessively large fibrotic scar due to a pathologic reparative woundhealing response, where the scar is raised above the surrounding adjacent normal tissue, but it does not grow beyond the boundaries of the original wound. It often regresses in size and improves in appearance over a few years after. Keloid – An abnormal, excessively large fibrotic scar due to a pathologic reparative fibrotic wound-repair response, where the scar extends beyond the boundaries of the original wound in addition to being raised above the surrounding adjacent normal tissue. It is a more severe degree of pathologic scarring than a hypertrophic scar because it can carry on growing indefinitely into a large, benign growth that can present a significant cosmetic issue to the individual. Laser-assisted subepithelial keratomileusis (LASEK) – An anterior surface ablative refractive surgical procedure that reshapes the cornea to correct refractive errors of the eye and is basically a modified form of photorefractive keratectomy (PRK).

256

With LASEK, instead of removing or excising the epithelium completely as in PRK, the surgeon cuts a round semi-circular area of epithelium with a finebladed instrument called a trephine. A dilute alcohol solution is then applied to loosen the cut epithelium inside the semi-circle where it attaches to Bowman’s layer and then the surgeon gently lifts and folds back the hinged loosened epithelial flap, exposing the underlying Bowman’s layer and corneal stroma for the excimer laser treatment. Mitomycin C (MMC) – Mitomycins are a family of aziridine-containing natural products isolated from the soil fungus, Streptomyces caespitosus. One of these compounds, MMC, finds use as a chemotherapeutic agent or as a scar-reducing agent by virtue of its potent DNA cross-linking. Myofibroblast – An epithelial-, endothelial-, or mesenchymal-derived cell (e.g., keratocyte) that acquires morphological and biochemical features of smooth muscle cells, including the expression of alpha-smooth muscle actin (a-SMA). It can contract by using the actin smooth muscle filament complex, which can help speed up wound repair by contracting the edges of the wound and enhancing ECM deposition compared to that of simply an activated fibroblast. It has been suggested that in several fibrotic diseases (e.g., liver cirrhosis, kidney fibrosis, retroperitoneal fibrosis, corneal fibrotic scarring) that persistence of myofibroblasts and chronic wound contracture leads to a long-term fibrotic scar. Platelet-derived growth factor (PDGF) – A family of growth factors (GFs) involved in stimulating fibrotic wound repair and angiogenesis. Three PDGF isomers are found in humans: PDGF-AA, PDGF-AB, and PDGF-BB. In wound healing, its primary cellular source is platelets, which degranulate and release PDGF-AB. This initiates a fibrotic repair response by chemotactic and proliferative effects on fibroblasts and inflammatory cells. Regeneration – Regrowth of a damaged organ or tissue so that the original structure and function are restored back to normal. Transforming growth factor-beta (TGF-b) – A family of cytokine proteins that are expressed by almost all mammalian cells in three isoforms (TGF-b1, TGF-b2, and TGF-b3), which have different functions in wound healing. All three TGF-b isoforms play major roles in wound healing including chemotaxis

Corneal Scars

(signaling migration of macrophages, monocytes, fibroblasts, and neutrophils), fibroblast proliferation, collagen production, and angiogenesis. They act on target cells to generate a specific molecular signaling response by binding to cell surface TGF-b receptors, types RI, RII, and RIII, which then initiates activation of several intracellular signaling pathways. All three TGF-b isoforms are secreted from cells as a large (125–160 kDa) latent TGF-b complex (LL-TGF-b complex), composed of an active TGF-b dimer, a latency-associated peptide (LAP), and one of four possible isoforms of the latent TGF-b-binding protein (LTBP). Overall, the LL-TGF-b complex covalently binds with high affinity to ECM, such as fibronectin, through its specific LTBP domain to create storage pools of latent TGF-b, whereas the free active TGF-b dimer binds to TGF-b cell receptors or creates more storage pools by binding to the core protein portion of dermatan sulfate and heparin sulfate protoglycans (PGs) in the ECM.

Introduction Tissue or organs of mammalian embryos and fetuses up to the first-half of gestation heal through scar-free regenerative wound healing, whereas tissues in late gestation fetuses up to elderly adults heal more slowly (inversely related to postnatal age) and less efficiently through scarforming fibrotic wound healing. At the microscopic level, a fibrotic scar is basically defined as disorganized extracellular matrix (ECM) containing repair mesenchymal cell phenotypes as opposed to regenerative or uninjured normal tissue, which have normal ECM architecture containing quiescent mesenchymal cell phenotypes, or tissue fibrosis, which has disorganized ECM containing quiescent mesenchymal cell phenotypes. At the macroscopic level, a fibrotic scar leaves a mark or cicatrix on the skin or tissue after injury, often leading to an alternation in function or cosmesis, whereas tissue fibrosis leaves an inconspicuous mark and regenerative tissue leaves flawless tissue restoration. Thus, fibrotic scarring represents reparation failure to complete tissue fibrosis and both represent failures to achieve true tissue regeneration. Research studies on prenatal embryos and fetuses suggest that the mechanisms involved in the regenerative response are intrinsic to the tissue itself and are not caused by the sterile prenatal environment in the womb. The evolutionary reason for the transition from the scar-free regenerative response to the scar-forming fibrotic repair response in higher mammals, such as humans, is that such a variation helped an individual live longer to reproduce and

257

take care of their young (i.e., natural selection). The fibrotic repair response results in rapid inflammatory cell recruitment and subsequent wound contracture and filling of space, which can be organ or life saving. The main disadvantages of the fibrotic repair response are that it commonly results in the reconstitution of semifunctioning tissue or can fill the tissue with a cosmetically displeasing scar. In modern times, the result of this evolutionary mismatch is that a fibrotic scar can occur from even minor injury (e.g., sterile, simple surgical incision with suturing). In humans, tissue fibrosis or fibrotic scarring occurs after almost any type of injury to any tissue in the adult human body, including the eye. One exception is injury in which ectodermal cells alone are damaged (e.g., firstdegree burns, corneal abrasion, skin abrasion) and heal through scar-free cellular regeneration without any longterm fibrotic effects to the underlying interstitial mesenchymal cells. In fact, in certain cases where the context of the injury and the molecular signaling pathways invoked are shifted toward embryonic mechanisms, scar-free tissue regeneration may still occur as opposed to fibrotic scarring. Thus, tissue fibrosis or fibrotic scarring is not inevitable. For example, if oral mucosa (e.g., gums, palate, and inner cheek) is superficially damaged, complete regeneration occurs. Other examples include complete regeneration of the liver even if up to two-thirds of the liver is removed or complete regeneration of the skin in specific injury contexts, such as a pin-prick, a hypodermic needle stick, or multiple needle insertions (tattooing). In contrast, fibrotic scarring results from penetrating stab incision into the liver or various more severe injuries to the skin, such as: (1) a larger incision than a needle stick, (2) an excision, (3) suturing a wound closed with pro-inflammatory material (cat gut suture), and (4) leaving sutures in place too long or with incorrect tension. These examples illustrate the plasticity of the two wound-healing responses, fibrosis and regeneration. Differences between the desired responses (regeneration or secondarily minimal to mild tissue fibrosis) and the less desired responses (nonhealing wounds or excessive fibrotic scarring) are subtle, but fundamentally come down to three variables: (1) the molecular signaling pathways invoked, (2) the targeted responding cells, and (3) the context and degree in which the targeted responding cells are stimulated. Central to our understanding of the pathogenesis of tissue regeneration versus fibrotic scarring (and tissue fibrosis) has been the identification of several cytokines and growth factors (GFs) involved in initiating and/or regulating the ongoing wound-healing responses. The relationship between these cytokines and GFs and their predicted cellular responses is now only being partially understood. As such, it is currently known that cytokines and GFs, their associated target cell surface receptors, and their molecular signaling pathways play a significant role

258

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

in directing cellular behavior and/or phenotype in normal, diseased, and wound-healing states. Currently, the cytokine best known to play the most crucial role in mediating either tissue regeneration or fibrotic scarring is transforming growth factor-beta (TGF-b). Three isoforms of TGF-b have been identified in mammals, TGFb1, b2, and b3, all having major roles in wound healing. Studies comparing embryonic tissue regeneration versus adult fibrotic scarring have identified a variety of differences in their intrinsic cellular levels of expression and release. TGF-b3 elicits a regenerative response, whereas both TGF-b1 and TGF-b2 elicit a fibrotic-repair response. There are other notable disparities between the TGF b isoforms (Table 1). TGF-b is secreted extracellularly in a latent or inactive pro-peptide form, consisting of a 25 kDa disulfide-linked homodimeric TGF-b peptide, also known as the active form, that is bound to a latency-associated peptide (LAP), which itself is bound to a specific latent TGF-b-binding protein (LTBP). This 125–160 kDa TGF-b complex is known as a large latent Table 1 Comparison of embryonic and adult wound healing characteristics Wound healing characteristics Fluid environment Sterile environment Oxygen tension Temperature Speed of epithelialization/ wound closure Fibroblast migration/ proliferation Fibroblast cellularity Myofibroblastic differentiation Extracellular matrix deposition Extracellular matrix architecture Tensile wound strength Inflammation Immune system Inflammatory response Neutrophil transmigration Platelets/degranulation Cytokine profile the fibroblasts TGF-b1 and b2 TGF-b3 FGF PDGF Angiogenesis Wound healing response

Embryo/fetus

Adult

Present Present Lesser Warmer Faster

Absent Absent Greater Cooler Slower

Faster

Slower

Greater Absent

Lesser Present

Faster

Slower

Normal

Disorganized

Normal

Reduced

Immature Sparse Absent Absent

Mature Great Present Present

Low High High Absent Lesser Tissue regeneration

High Low Low High Greater Fibrotic scarring

Data from Ferguson, M. W. J. and O’Kane, S. (2004). Scar-free healing: From embryonic mechanisms to adult therapeutic intervention. Philosophical Transactions of the Royal Society, London B 359: 839–850.

TGF-b (LL-TGF-b) complex. It binds covalently to ECM (e.g., fibronectin) through its LTBP domain to create storage reservoirs of latent TGF-b. During wound healing or inflammation, proteolytic enzymatic cleavage (e.g., tissue plasminogen activator (TPA), plasmin, matrix metalloproteinases (MMPs), or cathepsin) or nonproteolytic dissociation (e.g., low pH, reactive oxygen species) of the active TGF-b dimer occurs from the LL-TGF-b complex, leading to TGF-b activation. Active TGF-b isoform dimer binding to its targeted cell surface receptor(s) leads to a cascade of intracellular signaling pathways, including a Smad pathway as well as other pathways, each resulting in a cell-specific set of responses. In general, TGF-b/receptor-mediated intracellular signaling pathways stimulate the production of various ECM proteins and inhibit ECM degrading enzymes (increases inhibitors of MMPs and decreases expression of MMPs), although exceptions to these principles abound. TGF-b also modulates cellular functions such as apoptosis, motility, mitosis, and differentiation or transdifferentiation into new cellular phenotypes. With wound healing, the various cytokine or GF cascades sometimes result in overlapping specific cellular responses that are not completely well understood. For mesenchymal fibrocytes (e.g., keratocytes), this set of responses includes fibrocyte migration and proliferation and subsequent differentiation into a metabolically activated repair cell type known as a fibroblast. Fibroblast cellular behavior results in the synthesis and deposition of a provisional fibronectin matrix comprised of ECM adhesion molecules (fibronectin), ECM space-filling molecules (hyaluronic acid), and ECM signaling molecules (tenascin). The next step is the conversion to early mature ECM, comprised of collagen fibrils, sulfated protoglycans (PGs), and possibly epithelial- or endothelialderived surface basement membranes (BMs). Under appropriate conditions, such as prolonged or excessive TGF-b exposure, some fibroblasts differentiate into another repair phenotype known as a myofibroblast, which is the central cellular mediators of the fibrotic repair response (Figure 1). Myofibroblasts originate by differentiation from fibrocytes or activated fibroblasts or by transdifferentiation, also known as metaplasia, from epithelial, endothelial, or local smooth muscle cells that are associated with blood vessels (Figure 1). Morphologically, myofibroblasts are larger than activated fibroblasts and uniquely express alpha-smooth muscle actin (a-SMA), an important morphologic marker identifying this cell type using immunohistochemistry techniques. Myofibroblasts are highly contractile repair cells that promote wound closure and contraction. Functionally, myofibroblasts also express and secrete greater amounts of cytokines and GFs, synthesize and deposit greater amounts of ECM, and have stronger and more focal ECM cell surface adhesion receptor sites compared to activated fibroblasts. Myofibroblasts first appear in the stroma directly under the epithelium and

Corneal Scars

259

TGF-β

−/+

Fibroblast (activated keratocyte)

Epithelial cell

Fibrocyte (keratocyte)

−/+

+

+

Epithelial BM

TGF-β, etc.

TGF-β

+

−/+

−/+

Fibronectinsulfated PGs n

io

sit

Ch

Myofibroblast

D

Immature or mature ECM

n, wound oattractio ion tion, chem e ECM deposit Proilfera matur on, and contracti

o ep

em

TGF-β, etc.

oa

ttra

cti

on

Macrophage

Tissue fibrosis or fibrotic scarring Figure 1 TGF-b is the main cytokine player and the myofibroblast is the main cellular player in the fibrotic repair response. A myofibroblast is derived from a fibroblast, an epithelial cell, a fibrocyte (bone-marrow derived cell?), or many other cell types, and it exerts a central role in cellular reconstitution, inflammation, and extracellular matrix (ECM) deposition in the pathogenesis of tissue fibrosis. Specific cell–cell interactions and ECM–cell interactions further modulate myofibroblast cell behavior. + = amplifying factor. ¼ suppressing factor.

then their domain advances to progressively deeper levels. Myofibroblast-mediated ECM deposition occurs in an orderly sequence from immature (primitive) to mature (fibrotic) stages: first fibronectin and other noncollagenous provisional ECM proteins (hyaluronic acid, tenascin, etc.), then collagen type III, and lastly collagen type I and sulfated PGs (e.g., decorin, lumican, biglycan). As the scar matures, most of the provisional ECM components are replaced by collagen fibrils (type I >> type III) and sulfated PGs. As judged by collagen fibril diameter and density, ECM usually first reconstitutes to the most mature and most fibrotic stage in the portion of the wound just under the epithelium and then it advances progressively to deeper levels. Additionally, during the evolution of maturation, the hypercellularity and hypervascularity decreases due to apoptosis and tensile strength increases due to increased maturity acquired collagen fibril cross-linking. The mechanism of late disappearance of myofibroblasts is uncertain, but probably is due to a combination of apoptosis and dedifferentiation of cells back to quiescent fibrocytes. In humans, fibrotic scarring commonly causes major medical problems in a variety of tissues. In the eye, fibrotic scarring causes visual impairment or even blindness. In the peripheral and central nervous systems, glial scarring prevents neuronal reconnections and hence blocks

reparative attempts at restoring normal neuronal function. In gastrointestinal and reproductive organs, strictures and adhesions can give rise to serious or life-threatening conditions, such as infertility, bowel obstruction, or chronic pain. In ligaments and tendons, fibrotic scarring restricts motility and decreases strength. Additionally, if the fibrotic repair response is incomplete, becomes excessive, or fails to appropriately terminate, pathologic wound healing occurs resulting in unhealed wounds, hypertrophic scars, or keloid scars, respectively. Overexpression and release of pro-fibrotic TGF-b1 and b2 cytokines, upregulation of TGF-b receptors RI and RII, and/or the formation of a positive, autoinducible feedback loop have been suggested as the primary reasons for excessive fibrotic scarring in hypertrophic or keloid scars. The potential for TGF-b autoinduction is usually self-limited in nonpathologic fibrotic repair responses through a negative feedback loop that terminates further TGF-b expression and release from that cell. In adult humans, the degree of fibrotic scarring versus tissue regeneration or tissue fibrosis depends on the following five factors: tissue site, sex, race, age, and magnitude and contamination of the wound. For tissue site, the gums, liver, and skin can either regenerate or scar, depending on the context of the injury. Most other tissues or organs in the body just repair themselves through tissue fibrosis or

260

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

fibrotic scarring, which is body site dependent. The deltoid, sternum, and conjunctiva scar more than the face, abdomen, legs, and cornea. Concerning sex, fertile females scar more than postmenopausal females and all males as estrogen has a major stimulatory effect on the fibrotic repair response. For race, darker pigmented individuals scar more than less pigmented individuals. Concerning age, young people in their teens and 20s scar more than infants and children, who have an immature immune response, and both heal better than older individuals, whose inflammatory response becomes sluggish and less effective with age and whose mesenchymal cells become senescent. Finally, the magnitude of injury and wound contamination alter the wound-healing response as larger, more inflamed, more contaminated wounds scar more than smaller, uncontaminated wounds.

Ocular Wound Healing The ocular fibrotic wound-repair response is quite similar to that seen in other tissues in the human body. In the eye, fibrotic scarring most commonly occurs in the conjunctiva, cornea, iris, lens, retina, and choroid (Figure 2). In vascularized ocular tissues, like the conjunctiva, iris, choroid, or even the cornea or retina after having chronic ocular disease, fibrovascular scars are basically granulation tissue. Scars normally develop in response to vascular stromal injury whereby serum plasma or blood leaks into

the wound resulting in a clot. This initiates a fibrovascular repair response predominantly through a platelet-derived growth factor (PDGF)-mediated platelet degranulation signaling pathway resulting in massive release of PDGF into the wound along with some TGF-b release. However, the eye also has several avascular transparent tissues (e.g., cornea, lens, vitreous, and portions of the retina) that also can repair themselves through fibrosis, albeit through four less intense TGF-b-mediated signaling pathways. TGF-b-mediated wound healing occurs because TGF-b and its associated receptors are constitutively expressed in the normal state by almost all cells residing in or on ocular tissues. The external ocular surface has other factors to additionally consider in that it is bathed in a tear film composed of lacrimal gland secretions that are rich in latent TGF-b1. Similarly, internal intraocular structures are bathed in aqueous humor composed of ciliary and lens epithelial secretions that are rich in latent TGF-b2. Latent TGF-b2 is thought to be one of the principal factors promoting the normal naive immune status of intraocular tissues, including the cornea, due to its ability to maintain resident tissue immune cells in an immature state. Secondarily, latent TGF-b2 has a direct immunosuppressive effect whereby it inhibits immune cell proliferation and inflammatory cell cytokine and GF release. At the concentration levels found in the normal aqueous humor, latent TGF-b2 is also thought to participate in the inhibition of angiogenesis into transparent parts

Vascular fibrotic scarring of the conjunctiva

Vascular fibrotic scarring of choroid

Vitreous humor

Lens Iris

Avascular fibrotic scarring of the lens

Cornea

Avascular or vascular fibrotic scarring of the cornea

Avascular or vascular fibrotic scarring of the retina

a

tin

Vascular fibrotic scarring of iris

Re

id

oro

Ch

Figure 2 After injury, tissue fibrosis or fibrotic scarring occurs in the cornea, conjunctiva, iris, lens, retina, or choroid. The origin of the mesenchymal repair cell types (activated fibroblasts, myofibroblasts) in each case is either from local fibrocytes (keratocytes, subconjunctival fibrocytes), local epithelial cells (lens epithelium, retinal pigment epithelium), local endothelial cells (corneal endothelium), or local smooth muscle cells that are associated with blood vessels (iris, choroid). The lens undergoes only avascular fibrotic scarring, while conjunctiva, iris, and choroid only undergo fibrovascular granulation tissue scarring; the cornea and retina can undergo both, depending on whether preexisting corneal or retinal disease is present or on the context and severity of the initial injury.

Corneal Scars

of the eye, although a primary role for latent TGF-b2 in preventing vascularization in the cornea remains controversial. In various ocular diseases or after injury, TGF-b release and subsequent extracellular activation usually initiates a fibrotic repair response. The TGF-b-mediated fibrotic repair response can lead to significant visual impairment. TGF-b can cause overcorrection, undercorrection, or regression of refractive effect leading to refractive instability. TGF-b can also cause loss of refractive function by inducing irregular astigmatism or it can cause loss of transparency by inducing corneal haze, cataract, capsular fibrosis, or vitreous opacification. Tractional distortion of ocular structures, such as proliferative vitreoretinopathy, or other ocular complications, such as wound dehiscence, epithelial ingrowth, infection, hypotony (low intraocular pressure), epiretinal membranes, or retinal detachment, can also be caused by TGF-b . Thus, this molecule is a subject of intense research. Avascular transparent tissue fibrosis or fibrotic scarring in the cornea, lens, and retina is perhaps even more intriguing to researchers since it provides a unique opportunity to directly examine the cell biology during wound healing caused by the four TGF-b-mediated signaling pathways in isolation from the PDGF-mediated platelet degranulation signaling pathway.

Corneal Wound Healing Although corneal neovascularization resulting from chronic ocular disease, such as corneal infection or chemical injury, may lead to a fibrovascular repair response (Figure 3, far-right), this type of repair response will not be focused on in this article. Instead, this article will strictly focus on the avascular corneal fibrotic repair response (Figure 3, middle). In the latter repair response, all cases have in common early inflammation that is minimally present up to 2 weeks after injury and maximally

261

present up to 1 month after injury, at least some transient myofibroblast differentiation that is minimally present up to 4–6 months after injury and maximally present longterm, and the long-term presence of disorganized collagenous ECM (Figure 4).

Cell–Cell Communication Cytokines and GFs Corneal wound healing is an exceedingly complex process mediated by autocrine (affecting the same cell), juxtacrine (affecting adjacent similar cell types), or, most commonly, paracrine (affecting other adjacent cell types) cell–cell interactions involving cytokines and GFs produced by epithelial, stromal keratocyte, endothelial, immune, lacrimal gland, and corneal nerve cells (Figure 4). The paracrine interactions usually affect cells at most 50–75 mm from one another, while the other modes of interactions operate over shorter distances. The earliest event involves cell-mediated release of several cytokines and GFs by directly injured resident tissue cells or secretion by resident tissue cells just adjacent to the injury site and their subsequent extracellular activation. In the normal, uninjured state, the participating cytokines and GFs are expressed constitutively and stored intracellularly by the corneal epithelium, corneal nerves, keratoctyes, and corneal endothelium. These cytokines play a vital role in the maintenance of normal corneal health, structure, and function as some are secreted in a limited controlled fashion, mostly in the latent or inactive form. The adjacent tear film produced by the lacrimal gland and other ocular surface epithelia and the aqueous humor produced by ciliary and lens epithelial cells’ secretions are another rich source of cytokines and GFs in the normal, uninjured state since these fluids supply various nutrients to the avascular cornea that otherwise would not be available. After wounding activates this direct resident tissue injury-induced

Neutrophil Macrophage

Myofibroblast

Myofibroblast Activated keratocyte Macrophage

Activated keratocyte Keratocyte

Neovascularization Keratocyte

Platelets Keratocyte

Figure 3 Schematic diagrams demonstrating the histology of the cornea during the active wound-healing phase, approximately 1 week after injury when re-epithelization has just been completed and epithelial thickness restored, in normal uninjured cornea (far-left), avascular fibrotic scarring (middle), and fibrovascular granulation tissue scarring (far-right). If these wounds eventually reach the conclusion of the remodeling phase (not shown), the middle illustration would end up with corneal tissue fibrosis or fibrotic scarring, whereas the far-right illustration would end up with fibrovascular corneal scarring.

262

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Corneal tissue injury Epithelial

Stromal

Endothelial

Direct injury-induced cytokine and GF release or secretion: TGF-β>>PDGF

Epithelial cell migration until re-epithelization

Endothelial cell migration Keratocyte apoptosis Persistent corneal edema or stromal in-growth

Persistent epithelial defects or epithelial cell ingrowth

Proliferation resumes and epithelial thickness returns to normal

Compensatory basal epithelial cell size changes and/or changes in the number of cell layers in the epithelium

Endothelial cell junctions re-established and pump function resumes Inflammation

Attachment complexes Keratocyte proliferation (i.e., hemidesmosomes, and migration anchoring fibrils, and anchoring plaques) re-synthesized

Recurrent corneal erosions

Transient differentiation into activated keratocytes Inadequate cell-cell TGF-β signaling ± ECM suppressive factors

Hypercellularity of wound with transient myofibroblastic differentiation Provisional matrix synthesis and deposition Prolonged or excessive cell-cell TGF-β signaling Adequate cell-cell TGF-β signaling ± ECM amplifying factors

Early provisional matrix synthesis and deposition and later replacement by prolonged or excessive collagen and sulfated PG synthesis and deposition as well as chronoic myofibroblastic differentiation

Chronic mechanical irritation to keratocytes

Persistent keratoyte activation and extracellular lipid collection = stromal crystalline-like deposits

± ECM suppressive/amplifying factors

Early provisional matrix synthesis and deposition and later replacement by collagen and sulfated PG synthesis and deposition

Hypocellular primitive matrix

Wound remodeling Wound remodeling

Hypercellular fibrotic scar

Persistent corneal edema

Injury-induced inflammatory cell-stromal interactions: macrophages (TGF-β>PDGF)

Epithelial- or endothelial−stromal interactions Tear film- or aqueous humor−stromal interactions

Normocellular or hypercellular tissue fibrosis

Steroids MMC

Figure 4 Diagram of normal corneal wound-healing pathways in black and abnormal or pathologic healing pathways in blue. Depending on the context of the injury and the molecular signaling pathways invoked, three types of long-term corneal stromal scars are possible: hypocellular primitive matrix, normocellular-to-hypercellular tissue fibrosis, and hypercellular fibrotic scars. The most functionally optimal type is that of normocellular-to-hypercellular tissue fibrosis as it is macroscopically transparent, yet still somewhat strong in tensile strength (30%). Imbalances in the fibrotic repair response result in the other two stromal scar types. An incomplete fibrotic repair response results in a hypocellular primitive matrix scar, which is transparent, but very weak in tensile strength (2–3%). An excessive, nonpathologic fibrotic repair response results in a hypercellular fibrotic scar, which is hazy, yet the strongest in tensile strength (40%). Note that complete tissue regeneration is not possible in the in vivo adult human cornea, but if it could be manipulated to occur it would be the best option since it would be both transparent and back to normal tensile strength (100%). Topical corticosteroid drops suppress (i.e., decreases inflammation and blunts the intensity of some remaining wound-healing steps) the wound-healing pathways in both the epithelium and stroma (stroma > epithelium), while single intraoperatively applied mitomycin C (MMC) appears to block only the stromal wound-healing pathway by preventing keratocyte proliferation. Adapted from Dawson, D. G., Edelhauser, H. F., and Grossniklaus, H. E. (2005). Long-term histopathologic findings in human corneal wounds after refractive surgical procedures. American Journal of Ophthalmology 139: 168–178.

cytokine and GF cascade, supplemental exposure to additional cytokines and GFs may occur by four dynamic signaling pathways (Figure 4): epithelial– or endothelial–stromal interactions, inflammatory cell–stromal interactions, tear film– or aqueous humor–stromal interactions, and keratocyte mechanotransduction. The major pertinent cytokines and GFs studied to date in regard to corneal wound healing include connective tissue growth factor (CTGF), epithelial growth factor (EGF), fibroblast growth factor (FGF), interleukin-1

(IL-1), nerve growth factor (NGF), PDGF, TGF-a, TGF-b, tumor necrosis factor-alpha (TNF-a), and vascular endothelial growth factor (VEGF). Currently, TGF-b is thought to be the most important one of this group in regard to stimulating a fibrotic repair response. The major resident cellular source of TGF-bs in the normal, uninjured cornea is the epithelium, which primarily expresses TGF-b2, while keratocytes and endothelial cells express only a little TGF-b1. No TGF-b3 expression has been found in the adult human cornea.

Corneal Scars

Resident cell surface TGF-b receptor sites are most numerous on keratocytes, while epithelial and endothelial cells have very little. After injury, increased expression of both TGF-b1 and -b2 and their associated receptors occurs, with TGF-b2 predominating in epithelial–stromal interactions (stroma within 75 mm from the epithelium) and TGF-b1 predominating in endothelial–stromal interactions (stroma within 75 mm from the endothelium). In contrast, stromal injury more than 75 mm from either surface invokes only slightly increased TGF-b1 expression and, most importantly, receives no paracrine supplementation from epithelial– or endothelial–stromal interactions. Although the normal tear film contains high concentrations of latent TGF-b1, immediately following corneal epithelial injury, stimulation of the corneal nerve-lacrimal gland reflex arc causes hypersecretion of tears and increased secretion of several pro-fibrotic cytokines and GFs, such as PDGF-BB, NGF, and additional TGF-b1, all of which are normally stored intracellularly in lacrimal gland cells. The higher amounts of latent TGF-b1 as well as the newly secreted PDGF-BB and NGF are then activated by plasmin or MMPs, which are both released or secreted by injured or healing corneal epithelial cells into the wound. This corneal nerve-lacrimal gland reflex arc secretion and protease release usually returns to baseline levels within 7 days of epithelial defect closure; hence it is a transient event. Other cytokines or GFs from the group listed above have intriguing complementary or antagonizing effects to TGF-b’s pro-fibrotic effects. As discussed in the ocular fibrovascular repair response section, PDGF’s pro-fibrotic effects are almost identical to TGF-b’s pro-fibrotic effects. In the normal, uninjured cornea, small amounts of PDGFBB are produced by the corneal epithelium that reacts with its receptors, which are most highly concentrated in the keratocytes of the stroma and the corneal endothelium. Also, a limited amount of PDGF-BB is secreted by macrophages and lacrimal gland cells. After epithelial injury, the expression, release, and activation of PDGF-BB increases in parallel to that of TGF-b. NGF, a trophic neuropeptide important in maintaining corneal epithelial health, is another GF that is complementary to TGF-b’s profibrotic effects as NGF is known to directly stimulate myofibroblastic transformation, but without the proliferative or ECM deposition effects of TGF-b. CTGF is also complementary to TGF-b as it too stimulates tissue fibrosis. However, CTGF is only potently expressed when stimulated by TGF-b, suggesting CTCF mediates several downstream actions of TGF-b. Thus, CTGF may be more important in chronic stages of fibrotic scarring as opposed to initiation or early regulation of the fibrotic repair response. Alternatively, CTGF may just synergize with TGF-b’s pro-fibrotic effects since CTGF binds to TGF-b and potentiates TGF-b binding to TGF-b type II receptors. In contrast, TNF-a directly antagonizes TGF-b’s profibrotic effects by inhibiting the TGF-b/Smad pathway.

263

This direct antifibrotic effect may seem counterintuitive since TNF-a is one of the major pro-inflammatory cytokines of the cornea. As can be appreciated, a correct combination and temporal sequencing of the various cytokines and GFs is required for proper healing to occur. Imbalances can result in incomplete wound repair, such as (1) persistent epithelial defects and corneal ulceration or melting, or (2) excessive pathological tissue fibrosis, such as stromal outgrowth, fibrodegenerative pannus formation, stromal ingrowth, or retrocorneal membrane formation. TGF-b signaling pathways Both historical clinical anecdotal observations and recent experimental investigations have suggested that epithelial–stromal interactions are the primary mediator of the avascular fibrotic repair response in the cornea. Thus, epithelial cell-derived substances substitute for those produced by platelets in the skin and other vascular tissues. The cornea consists of three cellular layers consisting of epithelium, interstitial stroma with interspersed keratocytes, and endothelium and two acellular layers consisting of an epithelial BM/Bowman’s layer complex and an endothelial Descemet’s membrane. Although Bowman’s layer is not re-formed if damaged, both BM layers are re-formed if damaged because they are produced by epithelial or endothelial cells. Both acellular layers seem to function clinically as physiological barriers to separate the three cellular layers from direct contact or communication with one another to maintain normal corneal homeostasis. Disruption of the epithelial BM/Bowman layer complex induces maximal epithelial–stromal interactions that are intense enough even in a minimal injury state to cause a fibrotic repair response. This is seen clinically with bullous keratopathy – a pathological condition in which small blisters (vesicles or bullae) are formed in the corneal epithelium due to endothelial dysfunction, at the host–graft interface after penetrating keratoplasty (aka full-thickness corneal transplantation; a surgical procedure where a damaged or diseased cornea is replaced by full-thickness donated corneal tissue), or with Bowman’s layer breaks (a degenerative disorder of the cornea in which structural changes cause it to thin and change in shape resulting an outward bulging of the tissue) as a subepithelial fibrotic scar, stromal outgrowth, or fibrodegenerative pannus formation (fibrocollagen connective tissue that proliferates in the anterior layers of the cornea in degenerative corneal disease). In some cases of bullous keratopathy, there may even be an no obvious microscopic full-thickness Bowman’s layer break, but rather just separation of the of epithelium and its associated BM from Bowman’s layer due to severe epithelial edema causing subepithelial bullae or blisters. In this extreme example of epithelial–stromal interactions, the keratocytes apparently migrate through trigeminal nerve fiber perforation sites found in Bowman’s layer to the

264

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

subepithelial space created by the edema, where they undergo a fibrotic repair response. Separation or breaks in Descemet’s membrane provide a similar trigger for causing a fibrotic repair response in the posterior cornea through endothelial–stromal interactions. For example, with penetrating corneal injury or breaks in Descemet’s membrane, whether caused by trauma, disease (keratoconus), or surgery, endothelial–stromal interactions stimulate fibrotic scarring resulting in a supraendothelial fibrotic scar, stromal ingrowth, or retrocorneal membrane formation, depending on the degree of apposition of the broken Descemet’s membrane wound margins. Therefore, current evidence suggests that the main TGF-b pro-fibrotic signaling pathway to block or suppress to prevent or reduce the chance for fibrotic scarring is that of direct epithelial–stromal or endothelial–stromal interactions. Three other TGF-b pro-fibrotic signaling pathways are also present in the cornea and can initiate an avascular fibrotic repair response, but more often they just sustain the response already initiated by epithelial- or endothelial–stromal interactions. As already discussed earlier, a robust inflammatory cell response is part of a normal or nonpathologic fibrotic repair response. However, along with the advantages of stimulating an immediate innate immune response and thus preventing or at the very least lowering the risk for infection, leukocytes also secrete many pro-fibrotic cytokines and GFs, such as TGF-b and, to a lesser extent, PDGF. These pro-fibrotic factors are disadvantageous because they cause keratocytes in the corneal wound to undergo a higher degree of tissue fibrosis, which can lead to chronic nonpathologic scarring or even pathologic scarring. Usually after sterile clean injury, the inflammatory response in the avascular cornea is quite mild compared to that of the skin or other vascular tissues. However, dirty injuries or wounds complicated by infection or with preexisting immune-related disease may make the inflammatory response so massive that it becomes the primary pro-fibrotic signaling pathway resulting in varying degrees of severe corneal scarring. Another TGF-b pro-fibrotic signaling pathway was already discussed in the the section entitled ‘Cytokines and GFs’ when discussing the epithelial cell injury-induced corneal nerve-lacrimal gland reflex arc. These tear film–stromal interactions are usually transient (<7 days) and minor in comparison to the two already discussed. However, with persistent epithelial defects or chronic corneal ulceration, tear film–stromal interactions may become of paramount importance in causing fibrotic scarring in addition to the chronic increased release of epithelial cellderived cytokines at the peripheral edge of the epithelial defect. The aqueous humor may have a similar role with posterior stromal injury. Finally, the fourth TGF-b profibrotic signaling pathway involves keratocyte mechanotransduction interactions. Historically, wound tension has been known for years to promote a higher degree of fibrotic

scarring. This mechanotransduction pathway was confirmed experimentally using cell culture experiments where mechanical strain on fibroblasts increased pro-fibrotic TGF-b1 expression, upregulated TGF-b receptors RI and RII, and markedly increased collagen expression. Findings by other researchers indicate that fibrocyte (e.g., keratocyte) cell surface integrins, which form cell–ECM adhesions, can act as strain gauges to external mechanical stress, triggering both direct intracellular signaling pathways or indirect release of paracrine cytokine and GFs factors, which stimulates the fibroblastic phenotype and augments ECM synthesis and deposition. Cell–ECM Communication Suppressing or amplifying factors The ECM, through intrinsic binding and storage properties or cellular receptor-mediated adhesive properties, collaborates with cytokines and GFs in regulating cell behavior and phenotype through bidirectional cytokine and GF–ECM and cell–ECM interactions. In the latter, ECM adhesion and surface characteristics influence cellular behavior through interaction with cell surface receptors. Clinical anecdotal observation supports the contention that having an intact corneal BM usually prevents fibrotic repair phenotypes from forming in the stroma and return of the BM usually correlates with loss of the fibrotic repair phenotype. Apparently, an intact epithelial BM or Descemet’s membrane is the key factor in suppressing direct maximal epithelial- or endothelial–stromal interactions from taking place in the normal, uninjured state by binding and storing the soluble LL-TGF-b complex. Thus, epitheial BM or Descemet’s membrane act as passive diffusion barriers to epithelial- or endothelial-derived latent TGF-b, respectively, and they ultimately regulate the bioavailability of TGF-b to the underlying or overlying corneal stroma. BMs are composed mainly of collagen type IV, entactin/nidogen, different laminins, and heparan and chondroitin sulfate PGs. The LL-TGF-b complex binding and storage property of BMs is likely related to heparin binding mechanisms intrinsic to the heparin sulfate PGs found in BMs, which bind to the soluble LL-TGF-b complex through either LTBP-1 or -3 domains. Fibronectin, which is found in the provisional matrix, also has LL-TGF-b complex binding and storage properties similar to that of BMs. Some sulfated PGs in the corneal stroma, like decorin, can even bind to the active dimer form of TGF-b through their core proteins; thus, additionally antagonizing the biological affects of TGF-b. BM-bound heparin sulfate PGs may have a similar function through their core proteins too. With injury, corneal BM barriers are breached so that normal maximal epithelial- or endothelial–stroma interactions take place in addition to the acute wound-healing cascade already started by direct resident tissue injury itself. For example, with combined superficial epithelial

Corneal Scars

and stromal injury (photorefractive keratectomy (PRK), sterile corneal ulceration), this stromal wounding response is clearly seen clinically as it frequently stimulates excessive subepithelial fibrotic scarring resulting in prolonged or even permanent subepithelial haze. Apparently, removal of the epithelial BM directly exposes anterior stromal keratocytes to maximal pro-fibrotic effects of TGF-b released by the injured epithelial cells or found in the neurally stimulated tear film. Moreover, after injury, return of an intact epithelial BM suppresses normal epithelial cell-derived TGF-b2 intracellular expression and, more importantly, binds the TGF-b2 extracellularly secreted from epithelial cells. Thus, epithelial and presumably endothelial BMs represent an efficient ECM antifibrotic mechanism that goes slightly beyond just simple barrier function. However, this ECM antifibrotic mechanism is not foolproof since chronic subepithelial haze still persists long-term or episodic subepithelial haze still exacerbates after ultraviolet light overexposure, despite having an intact epithelial BM. This persistent haze seems to be best explained by TGF-b signaling pathways that work only in the stroma, such as mechanotransduction or inflammatory cell TGF-b signaling pathways, respectively. These two TGF-b signaling pathways can act directly on activated keratocytes and/or myofibroblasts in the stroma without having to send cytokines or GFs through a BM. Regarding the latter UV light example, quiescent or regressed myofibroblasts in a subepithelial fibrotic scar may have long-term upregulated TGF-b receptor sites densities, making them hypersensitive to any TGF-b, whether caused by inflammation or late resident cell injury. This may then lead to an autocrine positive feedback loop of more TGF-b production from a single myofibroblast or from a paracrine positive feedback loop due to macrophage infiltration and further TGF-b release. Corneal cell–ECM interactions are an area that needs further study. Other disciplines have already shown that cells adhere to the ECM usually through their cell surface receptors (e.g., integrins, cell surface proteoglycans). For example, cell–ECM adhesion by integrins receptors regulates cellular homeostasis in multiple ways, including regulation through direct and indirect connections to the actin cytoskeleton, growth factor receptors, and intracellular signal transduction cascades. Integrins also sense mechanical strain arising from the ECM, thereby converting these stimuli to downstream signals modulating cell behavior. Disruption of this connection to the ECM has deleterious effects on cell survival, sometimes leading to a specific type of programmed cell death called anoikis. Anoikis is defined as apoptosis that is induced by inadequate or inappropriate cell–ECM interactions. Another direct extension of this cell–ECM interaction theme is clinical anecdotal reports and experimental animal research studies in which injuryinduced stromal surface irregularity, also known as irregular surface nanotopography, results in increased TGF-b

265

expression and a higher degree of corneal fibrotic scarring. Although irregular surface nanotopography is known to alter cell adhesion and positional information to the surface cells, such as epithelium and endothelium, it remains to be proven whether this pathway directly leads to myofibroblastic transformation since incomplete return of an intact epithelial BM has been reported as another possible cause. Bone marrow-derived stem cells Bone marrow-derived immune cells have been recently been detected in the corneal stroma and they may also participate in TGF-b molecular signaling pathways, especially the histiocytes (tissue macrophages). The relative contribution and importance of bone marrow-derived stem cells to regenerative or fibrotic repair responses needs to be explored further. Because of their extraordinary plasticity and their ability to secrete embryonic cytokines and GFs that promote tissue regeneration over fibrotic repair, bone marrow-derived stem cells are being explored as restorative cellular therapy in preclinical trials for myocardial infarction, neurodegenerative disorders, and osteogenesis imperfecta. However, some bone marrow-derived cells, or their progeny, may reside for prolonged periods in the stroma and could make an as yet unknown contribution to the wound-healing process. Cellular responses After injury-mediated induction of the TGF-b/receptor molecular signaling pathway, keratocytes undergo a series of repair phenotypic and behavioral changes that are temporally and spatially regulated during the corneal wound-repair process. Keratocytes are neural crestderived mesenchymal cells that are sandwiched between collagenous lamellae forming a closed, highly organized syncytium. Keratocytes communicate with one another through gap junctions present on their long dendritic processes. The adult human corneal stroma has approximately 2.4 million keratocytes that occupy approximately 10% of the stromal volume. In postnatal life, they remain in the corneal stroma as modified fibrocytes, where they inconspicuously maintain the ECM of the corneal stroma. Depending on the type of injury and the molecular signaling invoked, keratocytes can be stimulated to undergo cell death (apoptosis) or transform to become a migratory keratocyte, an activated keratocyte, or a myofibroblast. Within minutes after injury, exposure to cytokines and GFs, such as IL-1 and TNF-a, found in the neurally stimulated tear film or released from injured epithelial cells is thought to lead to keratocyte apoptosis. The phenotypic transition (actin-cytoskeleton rearrangement) from a keratocyte to a migratory keratocyte appears to be initiated by loss of gap junction contact and possibly loss of keratocyte ion channels. Behaviorally, migratory keratocytes serve to reconstitute the cellularity of the tissue by once again re-forming a large intercommunicating network of keratocytes. The phenotypic transition

266

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

from a keratocyte to an activated keratocyte and then possibly to a myofibroblast depends on the TGF-b ligand/receptor occupancy rate and the duration of this occupancy rate. The myofibroblast phenotype may also require synergistic PDGF co-stimulation. Behaviorally, activated keratocytes promote provisional fibronectin ECM synthesis and deposition, while myofibroblasts promote mature collagenous fibrotic ECM synthesis and deposition. It has been postulated that at least partial ECM regeneration may still be possible in vivo since an adult keratocyte stem cell subpopulation has been discovered near the limbus. These progenitor cells express the ocular development gene Pax6, which is an embryonic marker not expressed by resident stromal keratocytes. Overall, it has been suggested that variations in these two repair phenotypes (activated keratocyte and myofibroblast), especially the degree and duration of myofibroblast differentiation, ultimately determine the final clinical outcome of the wound-healing process, resulting in either primitive, incomplete wound repair, acceptable tissue fibrosis, or excessive fibrotic scarring. Three Phases of Corneal Wound Healing

TGF-β ligand/receptor signaling/# of myofibroblasts

Classically, corneal wound healing is typically divided into three overlapping phases (Figure 5): (1) inflammatory phase – cytokine release and amplified expression, keratocyte apoptosis, and inflammation; (2) active wound-healing Injury

phase – epithelial cell migration, proliferation, and regeneration, keratocyte migration and proliferation, keratocyte metabolic activation and eventual quiescence, myofibroblast development, ECM synthesis and deposition (immature primitive scar formation with or without transition to early mature fibrotic scar formation); and (3) remodeling phase – decrease in cellular density, persistence or disappearance of myofibroblasts, persistence or disappearance of primitive scar, and collagen fibril re-organization/remodeling (mature fibrotic scar formation). Inflammatory phase (lasts up to 2–4 weeks after injury)

As described earlier, after injury, damaged resident cells in the cornea release or adjacent surviving cells secrete a variety of cytokines and GFs into the wound. As the local concentration of these cytokines and GFs increases and these compounds are converted to their active form, one of the first morphologically observable changes in the corneal stroma following injury is a 50–75 mm zone of keratocyte apoptosis surrounding the actual direct injury site. This is followed by a subsequent transient influx of mixed acute and chronic inflammatory cells. Keratocyte apoptosis usually peaks at approximately 4 h after injury, but may still occur minimally up until approximately 1 week after the initial insult. After the initial peak of keratocyte apoptosis, an increased proportion of cells die

Inflammatory phase Active wound healing phase (injury to 2−4 weeks) (injury to 4−6 months)

Remodeling phase (4−6 months to 3−4 years)

Pathological fibrotic scar repair response

Chronic fibrotic scar repair response Prolonged or Incomplete excessive fibrotic repair Normal fibrotic repair response response repair response

Time Figure 5 Representative line graph of a normal fibrotic repair response (dashed black line) in the cornea that results in tissue fibrosis (long-term fibrotic ECM, transient (<6 months) myofibroblast differentiation) at the end of the active wound-healing phase. If the transient myofibroblastic diffentiation extends into the remodeling phase, but eventually goes away, it still is considered a normal fibrotic repair response, albeit prolonged or excessive. If myofibroblastic differentiation is chronic (persistent or long-term), it is considered a fibrotic scar (solid black line). In comparison, representation plots for incomplete (dotted black line) or pathologic (thickest solid black line) fibrotic repair responses are shown. In normal corneal wound healing, myofibroblastic differentiation and ECM synthesis and deposition are both TGF-b dependent processes, which are appropriately terminated through negative regulatory mechanisms during the active wound-healing phase. However, prolonged or excessive TGF-b production occurs through four possible signaling pathways that may result in excessive tissue fibrosis, fibrotic scarring, or pathologic fibrosis. In comparison, incomplete TGF-b production results in primitive or provisional fibronectin matrix deposition long-term wound healing.

Corneal Scars

through necrosis. Within 8 h of injury, neutrophils are attracted to the wound site by chemokines and these inflammatory cells function by killing microbes and even host resident cells with their secretion of free radicals. Neutrophils serve no direct role in amplifying or suppressing the fibrotic repair response through its secreted cytokines or GFs, but may indirectly amplify it by damaging more resident tissue cells. Within 24 h of injury, monocytes are attracted to the wound site where they differentiate into macrophages, which serve to degrade and remove dead or damaged cells and ECM. Keratocytes and epithelial-derived proteases and collagenases help macrophages in this degradation and removal process. Macrophages also directly amplify the fibrotic repair response through secretion of pro-fibrotic TGF-b 1 and some PDGF, which can potentially result in a paracrine positive feedback loop, if intense enough. Finally, within 3 days of injury, lymphocytes enter the wound, where they secrete antifibrotic cytokines and GFs. Overall, this direct injury-mediated innate inflammatory response typically adds more pro-fibrotic cytokines and GFs to the original direct resident tissue injury-mediated cytokine and GF cascade. The innate inflammatory response amplifies the cascade of events already taking place, commonly resulting in additional keratocyte proliferation and migration, further keratocyte differentiation to one of the various fibrotic repair phenotypes including the myofibroblastic stage. Thus, inflammation has both positive and negative consequences of preventing infection and amplifying the fibrotic repair response, respectively.

267

Remodeling phase (occurs from 4–6 months to 3–4 years after injury)

During the remodeling phase, corneal cellularity in the repair tissue decreases closer to normal because of the disappearance or persistence of residual keratocytes or myofibroblasts. The stromal cells also revert back to a more normal nonfibrotic reparative phenotype as actin stress fibril bundles disappear and TGF-b receptor site density gets closer to baseline. The ECM reorganizes closer to normal as collagen fibrils become larger in diameter, more regular, more orderly in arrangement, richer in collagens type I molecules, less rich in type III molecules, and more cross-linked. In addition, any residual hyaluronic acid in the provisional ECM is replaced by more mature, sulfated PGs. The remodeling phase is the dominant process guiding the early mature scar to complete the maturation process. As such, this phase may be considered more or less a functional improvement phase. After remodeling, corneal scars gain increased tensile strength, have less haze and less wound contracture, and are more cosmetically pleasing with a reduction in size and improvement in structure.. During the remodeling phase, the collagen fibril turnover rate is still higher than normal, most likely due to ongoing MMP activity. Unfortunately, if only an immature provisional fibronectin ECM is present after completion of the active wound-healing phase, then the remodeling phase is incapable of making the transition from the immature provisional matrix to the mature collagenous ECM and a provisional fibronectin matrix remains permanently in the corneal stroma. Termination of the fibrotic repair response (about 3–4 years after injury)

Active wound-healing phase (lasts up to 4–6 months after injury)

Proliferation and migration of residual surviving keratocytes begins 12–24 h after injury to reconstitute the cellularity of the injured area and typically continues for only several days after injury. The migrating keratoctyes are spindle-shaped in appearance and highly motile. Subsequently, migratory keratocytes differentiate into a metabolically activated repair phenotype called an activated keratocyte. Within 1–2 weeks after injury, myofibroblasts first begin to appear in the stromal wound under the epithelium and then develop in the deeper stroma down to a depth of approximately 50–75 mm. Myofibroblasts are a fibrotic repair phenotype characterized by significant deposition of a disorganized collagenous ECM, significant hypercellularity, and extensive wound contraction. An important physical characteristic of myofibroblasts in the cornea is their reduced transparency relative to other cell types in the corneal stroma, which may be due to a loss of soluble cytoplasmic corneal crystallins in combination with assembly of insoluble actin stress fiber bundles.

In humans, the long-term result of corneal wounding is the production of a hypercellular fibrotic scar (i.e., chronically hazy fibrotic scar) or a normocellular-to-hypercellular tissue fibrosis (i.e., transparent fibrotic tissue) in wound regions where epithelial– and endothelial–stromal interactions took place and a hypocellular primitive matrix (i.e., transparent immature provisional fibronectin ECM) in wound regions where keratocyte injury pathways took place in the absence of other supplemental healing responses (Figure 4). These three long-term histological tissue repair types have functional differences as the hypercellular fibrotic scar is strong (tensile strength approximately 40% of normal), but clinically hazy because of persistent myofibroblasts present in the repair tissue longterm. In contrast, the hypocellular primitive matrix is transparent, but it is very weak in tensile strength since it is composed of very little collagen fibrils (tensile strength approximately 2–3% of normal). Normocellular-to-hypercellular tissue fibrosis more or less is a hybrid of the two just described as it is macroscopically transparent and somewhat strong (tensile strength approximately 30% of normal). An additional variable to consider in this scheme

268

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

} Hypercellular fibrotic scar } Hypercellular fibrotic scar } Hypocellular primitive scar } Hypocellular primitive scar Suture

Hypercellular fibrotic scar { (b)

(a)

} Hypercellular fibrotic scar } Hypocellular primitive scar

Normocellular-to-hypercellular tissue fibrosis or hypercellular fibrotic scarring

(c)

(d)

Figure 6 Long-term histologic findings of corneas that had undergone sutured, temporal, clear-corneal cataract extraction surgery (a), astigmatic keratotomy (b), photorefractive keratectomy (c), and laser-assisted in situ keratomileusis (d). All corneas shown are greater than 4 years after surgery and, technically, they have completed the remodeling phase and have terminated the fibrotic repair response (toluidine blue 15X for a, 25X for b, and 100X for c and d). Adapted from Dawson, D. G., Edelhauser, H. F., and Grossniklaus, H. E. (2005). Long-term histopathologic findings in human corneal wounds after refractive surgical procedures. American Journal of Ophthalmology 139: 168–178.

is the fact that more precisely realigned wounds, such as sutured and unsutured wounds with minimal gaping and no epithelial cell plugging, heal better than poorly aligned wounds, such as wounds with wide wound gaping; epithelial plugging; or incarceration of Bowman’s layer, Descemet’s membrane, or uvea (Figure 6).

Table 2

Modulation of Scarring

Amniotic membrane

Mechanism of action of scar-reducing agents

Agent

Mechanism

Corticosteroids

Inhibition of inflammatory response Possible inhibition of keratocyte proliferation and collagen synthesis Inhibition of keratocyte proliferation Promotes keratocyte and myofibroblastic apoptosis Suppression of epithelial-stromal interactions Potent anti-fibrotic, anti-inflammatory, and anti-angiogenic properties Partially shifts the fibrotic repair response over to the scar-free regenerative response Decreased myofibroblast differentiation and collagen synthesis Gene transfer of TGF-b antagonists (e.g., fibromodulin) or blockers of TGF-b signaling (e.g., truncated TGF-b receptor II)

Chemotherapeutic agent (MMC)

Currently Available Scar-Reducing Therapies Corneal fibrotic scars remain difficult to cure. Thus, the best approach is prevention. At present, there is no pharmacologic agent or surgical procedure that is universally effective in ameliorating fibrotic scarring. The best strategy currently entails a polytherapeutic empirical approach involving short-term (approximately 1–3 months in duration) topical corticosteroids, a single intraoperative local application of mitomycin C (MMC), and possibly several alternative or emerging options (Table 2). The

Exogenous TGF-b3 supplementation Anti-TGF-b1 and b2 agents Gene therapy

Corneal Scars

easiest way to understand the place that a drug is located in the range of possible treatment options is to consider the TGF-b signaling pathway or the wound-healing phase that the compounds act on. Corticosteroids reduce scar formation primarily by suppressing the TGF-b-mediated inflammatory cell–stromal interactions and, secondarily, by possibly depressing steps in the active wound-healing phase, such as inhibition of fibroblast proliferation and diminished collagen synthesis. Regarding the fibroblasts, a previous study demonstrated that corticosteroids have a direct antiproliferative effect on ocular fibroblasts by altering the intracellular activity and expression of multiple genes that participate in fibrotic scar formation, including inhibition of TGF-b1, TGF-b2, and collagen expression. Thus, it should not be surprising that corticosteroids affect stromal wound healing to a greater degree than epithelial wound healing. Unfortunately, in randomized, placebo controlled, prospective clinical trials, corticosteroids produced no clinically significant longterm effects on either haze or refractive regression due to fibrotic scarring compared to placebo, but were definitely associated with a few unwanted side effects (high intraocular pressure and posterior subcapsular cataract formation). Thus, the role of steroids in the prevention of fibrotic scarring remains limited. Moreover, in view of their side-effect profile, steroids clearly would be unacceptable for longterm use in scar prevention and should be strictly used short-term (about 1–3 months from injury). Overall, corticosteroids will likely be used most effectively in future scar-reducing protocols as part of a polytherapeutic strategy since inflammation is just one of the four TGF-b signaling pathways involved in the avascular fibrotic repair response. When used, however, the timing of corticosteroid administration is crucial as the best results occur if used within hours of injury before neutrophils and profibrotic macrophages enter the wound. Once the injury-induced inflammatory response is initiated in the wound, it is somewhat self-amplifying, making it quite difficult to immediately suppress the response with corticosteroids. The mechanism by which MMC prevents or reduces fibrotic scar formation has not been fully elucidated. It appears to work by inhibiting cell proliferation due to its potent alkylating chemotherapeutic effect by crosslinking DNA after metabolic activation by way of reduction. This effect is downstream from the initiating pro-fibrotic events in the four TGF-b-mediated signaling pathways of the cornea. Thus, MMC may affect all the four pathways, at least acutely while it remains in the tissue at therapeutic concentrations. MMC is potent because it affects cells in all phases of the cell cycle. In high concentration, MMC also is known to promote direct keratoctye and myofibroblast apoptosis in the inflammatory phase via free radical injury, which may be another mechanism by which it works. MMC combined with low-dose corticosteroids is the

269

conventional treatment approach used today for corneal fibrotic scar prevention in high-risk cases. This combination is also used in the treatment of corneal fibrotic scars after additional maneuvers are performed to remove as much of the scar tissue as possible by surgical debridement (e.g., phototherapeutic keratectomy (PTK) before direct MMC local application). An alternative or supplementary scar-reducing treatment is that of using fresh or cryopreserved amniotic membrane (AM) allografts to cover treated areas of exposed injured corneal stroma to prevent fibrotic scarring. Human fetal AM is composed of three layers: a cuboidal to columnar epithelial monolayer, a 200–300-nm thick BM, and an avascular stroma. This membrane suppresses the inflammatory and active wound-healing phases of the fibrotic repair response. Intact AM–BM suppresses epithelial–stromal interactions since the stroma has intrinsic antifibrotic, anti-inflammatory, and antiangiogenic properties, presumably due to cytokine and GF binding to sulfated PGs in the AM stroma. AM also aids in re-epithelization due to its epithelial surface and its BM, which contains cell adhesion molecules, such as fibronectin. The primary disadvantages of AM are its expensive cost and increased surgical time. Thus, AM is used more often for severe corneal and/or limbal stem cell disease rather than repair of simple corneal wounds. Newer advanced surface ablation techniques, such as laser-assisted subepithelial keratectomy (LASEK) or epikeratome laser-assisted in situ keratomileusis (EpiLASIK), modify the PRK surgical technique using natural means to suppress direct epithelial–stromal interactions. For example, in LASEK surgery, the replaced epithelial flap retains as much intact epithelial BM as possible, which serves to bind and store direct injury-induced epithelialderived TGF-b. Thus, LASEK ingeniously takes advantage of natural antifibrotic cytokine and GF–ECM interactions to reduce the degree of fibrosis. EpiLASIK goes one step further in that the epithelial flap retains much healthier, living corneal epithelium as well as an intact epithelial BM. Thus, EpiLASIK additionally blunts profibrotic tear film–stromal interactions by natural means since a healthy, viable superficial squamous epithelial layer containing zonula occludens tight junctions serves as a physical barrier to direct stromal exposure to cytokines and GFs factors in the tear film. Emerging Scar-Reducing Therapies Extensive research has been performed to elucidate the fibrotic repair response, particularly in the skin, with the long-term goal of iatrogenically manipulating this response toward obtaining a clinical advantage. Recent molecular therapeutic investigations have concentrated on inhibiting myofibroblast differentiation by targeting TGF-b. A promising strategy currently being evaluated

270

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

in phase-I and-II clinical trials in the skin entails manipulating the ratio of the three TGF-b isoforms in the wound in favor of TGF-b3. Three independent, randomized, placebo controlled, prospective clinical trials have shown that exogenous TGF-b3 supplementary treatment resulted in clinically significant scar prevention or reduction with primary and multiple secondary endpoints in the skin. Inhibition of TGF-b1 and b2 binding to their associated receptors with topical anti-TGF-b antibodies has also been shown to reduce myofibroblastic differentiation, ECM deposition, and cell haze induced by PRK in animal models. By subtly altering the ratio of cytokines or GFs present in the wound during corneal wound healing, it is ultimately hoped that one day postnatal wounds could be induced to heal like embryonic wounds through tissue regeneration. An important caveat to these molecular therapies is that the timing of the application of these agents is critical, usually with the best results occurring if used immediately after injury or, at the very least, within 48 h of injury. There have been a few published studies, limited to animal models, using gene therapy for scar reduction. To date, fibroblasts have been used as the primary targets. Further investigation into the role of gene therapy for scar reduction is still needed since the practical use of such an approach is not presently feasible. Other possible emerging fibrotic scar-reducing therapies include systemically administered minocycline antibiotics, which significantly reduce the severity of hypertrophic scarring in a rabbit ear scar model. The mechanism by which minocycline reduces scar formation in this model remains unanswered, but the most plausible mechanism involves inhibition of MMPs resulting in the inhibition of fibroblast migration. Anti-intracellular Smad signaling pathway agents are another promising possibility.

Conclusion Significant advances have been made in understanding the mechanisms controlling the adult fibrotic wound-repair response in comparison to the embryonic scar-free regenerative response. Discovery of four TGF-b signaling pathways and the central role of myofibroblasts in corneal wound healing have been crucial to this end. Currently, the most effective corneal scar-reducing therapies involve a rather crude polytherapeutic empirical strategy of management involving short-term topical steroids and single intraoperative MMC application. Hopefully, in the near future, the recent advances in understanding the molecular and cell biology of fibrotic wound repair can translate into the development and clinical use of more promising agents, such as molecular, gene, or stem cell therapies. The capability to successfully manipulate the fibrotic repair process in the cornea in vivo offers many tantalizing

prospects from preventing blindness from corneal scarring to obtaining 20/8 perfect vision with optimal corneal wound healing after keratorefractive surgery to perfect tissue regeneration due to converting the adult fibrotic repair response back to an embryonic scar-free regenerative response to the possibility of restoring vision through tissue engineering of a human healthy cornea for replacement purposes.

Acknowledgments This work was supported in part by NIH Grants P30 EY06360 (Departmental Core Grant), T32EY07092 (DGD), R01EY00933 (HFE), and Research to Prevent Blindness, Inc., New York, New York. See also: Corneal Endothelium: Overview; Corneal Epithelium: Wound Healing Junctions, Attachment to Stroma Receptors, Matrix Metalloproteinases, Intracellular Communications; Cornea Overview; The Corneal Stroma; Refractive Surgery and Inlays; Regulation of Corneal Endothelial Cell Proliferation; Regulation of Corneal Endothelial Function; The Surgical Treatment for Corneal Epithelial Stem Cell Deficiency, Corneal Epithelial Defect, and Peripheral Corneal Ulcer.

Further Reading Dahlmann, A. H., Mireskandari, K., Cambrey, A. D., et al. (2005). Current and future prospects for the prevention of ocular fibrosis. Ophthalmology Clinics of North America 539–559. Ferguson, M. W. J. and O’Kane, S. (2004). Scar-free healing: From embryonic mechanisms to adult therapeutic intervention. Philosophical Transactions of the Royal Society B: Biological Sciences 359: 839–850. Fini, M. E., Stramer, B. M., Fini, M. E., and Stramer, B. M. (2005). How the cornea heals: Cornea-specific repair mechanisms affecting surgical outcomes. Cornea 24(8 supplement): S2–S11. Gabison, E. E., Huet, E., Baudouin, C., and Mensashi, S. (2009). Direct epithelial–stromal interaction in corneal wound healing: Role of EMMPRIN/CD147 in MMPs induction and beyond. Progress in Retinal Eye Research 28(1): 19–33. Klenkler, B., Sheardown, H., Jones, L., et al. (2007). Growth factors in the tear film: Role in tissue maintenance, wound healing, and ocular pathology. Ocular Surface 5(3): 228–239. Klenkler, B., Sheardown, H., Klenkler, B., and Sheardown, H. (2004). Growth factors in the anterior segment: Role in tissue maintenance, wound healing and ocular pathology. Experimental Eye Research 79(5): 677–688. LaGier, A. J., Yoo, S. H., Alfonso, E. C., et al. (2007). Inhibition of human corneal epithelial production of fibrotic mediator TGF-beta2 by basement membrane-like extracellular matrix. Investigative Ophthalmology and Visual Science 48(3): 1061–1071. Netto, M. V., Mohan, R. R., Ambrosio, R., Jr, et al. (2005). Wound healing in the cornea: A review of refractive surgery complications and new prospects for therapy. Cornea 24(5): 509–522. Obata, H., Tsuru, T., Obata, H., and Tsuru, T. (2007). Corneal wound healing from the perspective of keratoplasty specimens with special

Corneal Scars reference to the function of the Bowman layer and Descemet membrane. Cornea 26(9 supplement 1): S82–S89. Peled, Z., Liu, W., Levinson, H., et al. (2000). Cellular strain upregulated profibrotic growth factors and collagen gene expression. Surgical Forum 51: 591–593. Saika, S., Yamanaka, O., Sumioka, T., et al. (2008). Fibrotic disorders in the eye: Targets of gene therapy. Progress in Retinal and Eye Research 27: 177–196. Stramer, B. M., Mori, R., and Martin, P. (2007). The inflammation–fibrosis link? A Jekyll and Hyde role for blood cells

271

during wound repair. Journal of Investigative Dermatology 127: 1009–1017. Stramer, B. M., Zieske, J. D., Jung, J. C., et al. (2003). Molecular mechanisms controlling the fibrotic repair phenotype in cornea: Implications for surgical outcomes. Investigative Ophthalmology and Visual Science 44(10): 4237–4246. Wilson, S. E., Liu, J. J., and Mohan, R. R. (1999). Stromal–epithelial interactions in the cornea. Progress in Retinal and Eye Research 18(3): 293–309.

Corneal Endothelium: Overview D R Whikehart, The University of Alabama at Birmingham, Birmingham, AL, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Autosomal dominant – The property of inheritance of a disease or trait from a single parent. Chamber angle – This is the natural angle formed at the junction of the posterior limbus (at the trabecular meshwork), the ciliary body, and the iris. Glycosaminoglycan – A general term for any sugar polymer of alternating sugars and aminosugars (aka a GAG). Homeobox protein – A protein that is concerned with the embryological development of a multicellular organism. It is synthesized by a homeobox gene. Keratocyte – A cell type found in the corneal stroma that makes and maintains collagens and glycosaminoglycans (aka a stromal cell). Knock-out mouse – A genetically engineered mouse in which one or more specific genes have been made nonfunctional. Schlemm’s canal – A vessel behind the trabecular meshwork through which the aqueous fluid of the eye can drain into the venous system. Stem cell – A primitive cell that has not differentiated into a functional cell of a multicellular organism. Trabecular beam – A portion of the trabecular meshwork of the posterior limbus that consists of a rod of largely collagen material. When assembled in its typical complex pattern, trabecular beams offer resistance to the outflow of the aqueous fluid. Transient amplifying cell – A cell in the process of differentiation from a stem cell as a functioning cell.

Anatomy General Description The corneal endothelium consists of a monolayer of polygonal cells, primarily hexagonal in shape, such that each has dimensions of 18–20 mm (width) by 5 mm (thickness). These cells produce a predominately collagenous product known as Descemet’s membrane that is sandwiched between the anterior basal side of the endothelial cells and the posterior layer of the stroma. Descemet’s membrane is a basement membrane that may act as a cushion for endothelial cell trauma and, in general, such basement membranes have been assigned roles associated

272

with cell adhesion, migration, differentiation, and signal transduction. However, the exact roles of this membrane are presently unknown. Endothelial cells continually synthesize Descemet’s membrane throughout life. On the posterior apical side, endothelial cells make direct contact with the anterior aqueous fluid in the anterior chamber. This pH neutral fluid (pH 7.4) is virtually absent in serum proteins and lipids, but contains nourishing glucose, antioxidants, and a variety of cytokines. At the boundary of the endothelium, endothelial cells are joined to a limbus whose transitional area is composed of cells that may contain precursor stem cells for the endothelium. The circular diameter of the corneal endothelium is approximately 11.7 mm in the adult. As with the remainder of the cornea, there are normally no blood vessels in this tissue. Although the anterior cornea does possess nerve endings, none are found in the endothelium (see Figures 1–3).

Cell-to-Cell Junctions Corneal endothelial cells number about 3000 mm–2 in healthy, young adults and slowly decrease in number with age. The cells have a well-defined nucleus and their cytoplasm is packed with mitochondria necessitated by a high-energy requirement for the cells to act as a fluid pump. The cells also have a well-developed Golgi apparatus associated with the production of extracellular proteins needed for the assembly of Descemet’s membrane. The cell-to-cell junctions are somewhat tortuous and interdigitated, a form that helps to keep the cells together. The cellular membranes at these junctions are held together with well-described joining proteins, of which one assembled complex has been called a zonula occludens (tight junction). However, there is still a controversy about whether or not this is a true zonula occludens. The reason for this is that the channel joining adjacent endothelial cells must allow both sodium and water to flow into the anterior chamber. This junctional form has been described to be more focal or point-like than that of the true riveted tight junctions of a zonula occludens. Gap junctions have also been shown to exist between the cells for the purpose of transporting small molecules between each cell. Additionally, adhesion junctions to strengthen cell-to-cell fastening can be found there. The membranes facing the anterior chamber have a number of microvilli, typical of many endothelial cell types. However, no roles for the microvilli in these cells have been described (see Figures 3 and 4).

Corneal Endothelium: Overview

Iris

0.70

273

Cornea

a b Lens

11.7

0.50 d

0.70

Endothelium Figure 1 Overview of the anterior segment of the eye showing the cornea with the endothelium labeled in red. Numbers within the cornea are the cross-sectional dimensions in millimeters. The vertical line with the number represents the approximate diameter of an adult cornea in millimeters. Modified from Hogan, M. J., Alvarado, J. A., and Weddell, J. E. (1971). Histology of the Human Eye, p. 61. Philadelphia, PA: Saunders.

c

e

Figure 3 Three-dimensional, representational sketch of the corneal endothelium (c) with attached Descemet’s membrane (b) and posterior stroma (a). Endothelial cell microvilli are shown at (d). Intercellular channels are indicated at (e) while quasi-tight junctions occur at (f). Adapted from Hogan, M. J., Alvarado, J. A., and Weddell, J. E. (1971). Histology of the Human Eye, p. 101. Philadelphia, PA: Saunders.

A A a b

B C

C D

c

d e Figure 2 Cross-section of the human cornea. a, epithelium; b, Bowman’s membrane; c, stroma; d, Descemet’s membrane; e, endothelium. Adapted from Hogan, M. J., Alvarado, J. A., and Weddell, J. E. (1971). Histology of the Human Eye, p. 65. Philadelphia, PA: Saunders.

Embryology Initial Development It has been known for a long time that the corneal endothelium originates from neural crest, stem cells. These

E

Figure 4 Cross-section of human corneal endothelial cells. The large population of mitochondria are evident at (A) while an intercellular channel may be seen at (B). Two Golgi apparatus are pointed out at (C). The intercellular channel, labeled at (B) is shown to empty out at the aqueous chamber at (D). At (E) can be seen a microvillus. Modified from Hogan, M. J., Alvarado, J. A., and Weddell, J. E. (1971). Histology of the Human Eye, p. 103. Philadelphia, PA: Saunders.

cells migrate from folds of the neural ectoderm at weeks 4–5 of embryonic development. This is illustrated in Figures 5–7. Since the developing embryo is relatively small and neural crest cells are localized along each section of the neural folds, from which they detach, migration is not very distant for each group of cells. In eye development, optic sulci (the primordial eyes) appear as shallow pits along the neural plate at week 4. These sulci begin to protrude outward as hollow optic vesicles from the proencephalon or forebrain of the neural plate. At just over 30 days, a wave of mesenchymal cells (in this case, neural crest cells) migrates over the optic cup into the space between the anterior surface of the lens and the surface ectoderm (corneal epithelium). These cells will become

274

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Neural folds Neural crest cells

Surface ectoderm

Neural tube

(a)

Migrating neural crest cells

(b) Figure 5 The developing neural folds and neural tube. As the neural folds move away from the neural tube (a and b), neural crest cells break away from the neural fold tissue (a) and begin their migration (b) to specific developing tissues. Modified from Forrester, J. V., Dick, A. D., McMenamin, P. G., and Roberts, F. (2008). The Eye. Basic Science in Practice, 3rd edn., p. 111. Edinburgh: Saunders.

process, a third wave of neural crest cells migrates to the angle between the posterior cornea (endothelium) and the anterior edge of the optic cup (Figure 10). These cells eventually develop into the ciliary body and the iris. During this development, the tissues anterior to the chamber angle between the anterior eye cup and the endothelium becomes occupied by a mass of mesenchymal (neural crest) cells that remain, at first, undifferentiated. These cells develop into flat endothelial-like cells that bring about the trabecular beams (trabecular meshwork) and, separately, Schlemm’s canal cells. Some of the stellate cells, between the trabecular beams and the endothelial lining of Schlemm’s canal, appear to remain undifferentiated. Evidence to support this is seen in a number of stem cell markers in this area which change with wounding (Figure 11). This is an important observation as it suggests the retention of stem cells in a niche for the replacement of cells in the posterior limbus and for the corneal endothelium. The point is made again that evidence strongly suggests that nearly all the mesenchymal cells that invade these areas of the anterior segment are neural crest in origin. Role of Transcription Factors

Rhombencephalon Mesencephalon Migrating mesencephalic neural crest cells

Vll V Prosencephalon

lX

1 2 3

4

Figure 6 Migration pathways of neural crest cells. Neural crest cells are shown in green. The neural crest cells migrating to the eye originate primarily from the proencephalon (developing forebrain and brainstem) and the mesencephalon (developing midbrain). Migration is shown by dark blue arrows. Adapted from Forrester, J. V., Dick, A. D., McMenamin, P. G., and Roberts, F. (2008). The Eye. Basic Science in Practice, 3rd edn., p. 119. Edinburgh: Saunders.

the corneal endothelium. A second wave of mesenchymal cells (also neural crest cells) migrates around day 49 and places itself between the endothelium and epithelium. These cells are destined to become the keratocytes of the corneal stroma. Based on mouse model studies, it appears that those neural crest cells that will define the corneal endothelium at first remain in contact with the anterior lens and flatten out into a monolayer of cells. Then, the lens detaches from the endothelial monolayer to provide space for an anterior chamber where the aqueous fluid may enter (Figures 8 and 9). During this

Molecular biological mechanisms that control the overall development of the anterior segment remain incompletely described. This is also true for the corneal endothelium. What is known can explain some developmental effects. It is understood, for example, that inductive signals from the lens are partly responsible for endothelial cell differentiation. Defects in three lens genes that produce the homeobox proteins – MAF, FOXE3, and PITX3 – result in the inability of the lens to separate from the cornea (Figure 12).) Homeobox proteins assemble in a specific binding area of DNA to signal a specific RNA synthesis. They are required to bind to DNA in a set order for synthesis to begin. The three genes mentioned are concerned with producing genetic transcription factors that cause the synthesis of proteins necessary for lens–cornea separation. Other genetic abnormalities may also occur in the mesenchymal (stem) cells themselves that cause corneal development. In the endothelium itself, mutations in the genes PITX2, FOXC1, and PAX6, for example, are known to prevent satisfactory, functional corneal development. Normally, these genes produce the transcription factors for protein synthesis related to such development. PAX6, in particular, is required for making signaling molecules that cause transport of neural crest cells into the eye as well as the phenotypical development of the corneal endothelium itself. In addition, the amounts and origins of PAX6 proteins appear to be critical for the sequence signaling of corneal development. One signaling molecule that may be produced as a result of DNA binding of such transcription factors is transforming

Corneal Endothelium: Overview

Neural crest cells streaming over optic cup and stalk

End of week 4

Neural ectoderm of prosencephalon

Day 27 (4–5 mm) Otic placode Pharyngeal arches and clefts

Optic vesicle

275

Heart bulge

Lens placode Optic vesicle

Retinal disk Cavity of optic vesicle

Umbilical cord

Cavity of the optic stalk

Optic stalk

Figure 7 Cell migration pattern to the eye cup at the end of the 4th week. Neural crest cells are shown in green. Mesodermal cells are shown in red. Adapted from Forrester, J. V., Dick, A. D., McMenamin, P. G., and Roberts, F. (2008). The Eye. Basic Science in Practice, 3rd edn., p. 112. Edinburgh: Saunders.

Start of week 7 Day 44 (13–17 mm) Tunica vasculosa lentis

Lids forming

Hyaloid artery

Developing sclera

Mesenchyme between lens and surface ectoderm – future corneal endothelium and stroma

Developing choroid Figure 8 Established neural crest cells in the eye at the beginning of the 7th week. Neural crest cells are shown in green and include the future corneal endothelium and stroma. Adapted from Forrester, J. V., Dick, A. D., McMenamin, P. G., and Roberts, F. (2008). The Eye. Basic Science in Practice, 3rd edn., p. 113. Edinburgh: Saunders.

growth factor, beta 2 (TGF-b2). In TGF-b2 knock-out mice, that cannot produce this protein due to a lack of PAX6, the corneal endothelium is completely absent. It is also known that the overexpression of TGF-b1 (a related molecule using similar receptors) results in an absent endothelium. So, the case is made that these controlling proteins must be present in specific amounts and at the right time sequence to allow proper corneal endothelial cell development to occur.

Biochemistry and Metabolism Glucose and Energy Metabolism The corneal endothelium is a cell type that is required to have large outputs of energy to maintain the process of

deturgescence (discussed in the section entitled ‘Proteins synthesized for external transport’). In this regard, corneal endothelial cells maintain a high amount of adenosine triphosphate (ATP)-producing mitochondria as well as a defined Golgi apparatus for the complete synthesis, retention, and export of proteins. In studies of comparative carbohydrate metabolism, it has been shown that the amount of aerobic glycolysis (glucose breakdown to produce ATP) is 3 times higher than it is in the cells of the corneal epithelium and stromal keratocytes. Compared to the cells of the lens, corneal endothelial cells use aerobic glycolysis at better than 6 times the amount used by the lens. On a per cell basis, it is even estimated that corneal endothelial cells produce more ATP energy than individual cells of the ciliary body by 40%. Only cells of the retina exceed the cells of the corneal endothelium in

276

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Trabecular meshwork Nestin Alkaline phosphatase Telomerase

1

Transition zone Nestin Alkaline phosphatase Telomerase

2

3

Endothelium Telomerase

Schlemm’s canal Telomerase

Stromal collagen

Unwounded cornea

(a)

4

Trabecular meshwork Nestin, alkaline phosphatase Telomerase, +Oct3/4, Pax6, Wnt1, Sox2 Transition zone Nestin, alkaline phosphatase Telomerase, +Pax6, Wnt1 Sox2

Figure 9 A developing eye of a human fetus past 6 weeks. 1. the outer cornea (epithelium). 2. coalescing stem cells forming the endothelium and keratocytes. 3. the developing lens. 4. the inner layer of the retina. Modified from Forrester, J. V., Dick, A. D., McMenamin, P. G., and Roberts, F. (2008). The Eye. Basic Science in Practice, 3rd edn., p. 115. Edinburgh: Saunders.

2 1

3

OCM

PM

Figure 10 Development of tissues at the angle of the posterior limbus at 12 weeks. The figure shows the pupillary membrane (PM) beginning to form the optic cup margin (OCM) and indentation caused by vascular mesenchyme (double arrows). At (1) are the corneal endothelium (1st wave of neural crest cells); (2) the forming keratocytes (2nd wave of neural crest cells) and collagen; and (3) possible 3rd wave of invading neural crest cells to form the trabecular meshwork. Modified from Forrester, J. V., Dick, A. D., McMenamin, P. G., and Roberts, F. (2008). The Eye. Basic Science in Practice, 3rd edn., p. 131. Edinburgh: Saunders.

metabolic energy produced. Although direct evidence is scarce, it seems that corneal endothelial cells are not insulin dependent and, therefore, are not starved of glucose in diabetes. On the contrary, it is perplexing that glycation of sodium, potassium-ATPase (Na, K-ATPase) does not cause an inhibition of the enzyme in the diabetic. Cell Division and Replenishment The lack of any ready, apparent replication of endothelial cells has been an ongoing problem when cell replacement

Endothelium Telomerase, +Pax6, Sox2

Schlemm’s canal Telomerase, +Pax6, Wnt1

Stromal collagen Wounded cornea

(b) Figure 11 Stem cell and transient amplifying cell staining patterns at the peripheral corneal endothelium, transition zone, and trabecular meshwork of the posterior limbus in adult humans. Unwounded tissue is shown at (a) while wounded tissue is shown at (b). Tissues were stained with the following stem cell marker proteins: nestin, alkaline phosphatase, Wnt1, and Oct3/4. Tissues were stained with the following marker for both stem cells and transient amplifying cells: telomerase. Wounded tissues were stained with the following cell differentiation markers: Pax-6 and Sox2. Adapted from McGowan S. L., Edelhauser, H. F., Pfister, R. R., and Whikehart, D. R. (2007). Stem cell markers in the human posterior limbus and corneal endothelium of unwounded and wounded corneas. Molecular Vision 13: 1984–2000. http://www.molvis.org/molvis/v13/a224/.

might be required. In general, the population of these cells decline with age, but usually there remain sufficient numbers of cells at an advanced age to maintain corneal clarity. This has not seemed to be the case with wounding, trauma, and other dramatic forms of endothelial cell loss. Studies of the cell cycle in endothelial cell have indicated that the cells tend to remain in the G1 phase (perhaps even the Go temporary exit of cell division). This could be due to such factors as simple cell-to-cell contact inhibition or the presence of the TGF-b2 protein found in the aqueous humor. Departure from the G1 stage toward replication can be artificially induced with the E2F2 transcription factor since the E2F2 protein is known to bring cells into the S-phase of the cell cycle. Cells that have been transformed either with the SV40 large T-antigen or with the E6/E7 human papilloma virus can also act to initiate replication. The SV-40 and E6/E7 proteins inactivate the activity of retinoblastoma (Rb) and p53 cell cycle suppressor proteins by interacting with

Corneal Endothelium: Overview

A

B

Trabecular meshwork

200 μm

Figure 12 Cross-section of Foxe3 / (deficient) mouse eye showing small lens (A) and persistent attachment of lens to cornea (B). Adapted with permission from Medina-Martinez, O. and Jamrich M. (2007). Foxe view of lens development and disease. Development 134: 1455–1463.

them and allowing the E2F2 transcription factor to initiate the cell cycle. These are not the usual cell processes, however, and the lack of normal cell division remains unresolved. There is evidence, however, that endothelial cells are replaced in vivo. Studies have pointed to the existence of stem cells in the posterior limbus just beyond the boundary of the corneal endothelium (Figure 11). Initially, some investigators noted a higher than usual density of corneal endothelial cells that exist at the endothelial periphery. This suggested that replacement cells or germinating cells were present in this area. Later, stem cells were identified in the posterior limbus. This has been shown by labeling with stem cell markers such as Nestin and Oct3/4 in the limbus and the subsequent appearance of repair or developmental proteins such as PAX6 and Wnt1 in the limbus and the peripheral endothelium following wounding. The appearance of the transient amplifying cell marker telomerase in these areas strongly points to the possibility that stem cells, resident in the posterior limbus, give rise to new corneal endothelial cells. In fact, BrdU studies have confirmed the generation of new cells from the posterior limbus into the peripheral endothelium after wounding (Figure 13). This is analogous to what occurs in the corneal epithelium, but at a significantly lower rate of reproduction.

Cytokines and Immune Privilege The existence of cytokines, generated by the lens and other cells adjacent to the anterior chamber in the aqueous fluid, suggests that the endothelium may be influenced by its neighboring tissues. Cytokines are a broad mixture of polypeptides or proteins that are able to communicate a signal to a cell to initiate some change or response from the cell. In fact, the distinction between a

277

Corneal endothelium

Figure 13 BrdU fluorescence shown in the peripheral corneal endothelium and trabecular meshwork 48 h after a mechanical wound to the endothelium. Adapted from Whikehart, D. R., Parikh, C. H., Vaughn, A. V., Mishler, K., and Edelhauser, H. F. (2005). Evidence suggesting the existence of stem cells for the human corneal endothelium. Molecular Vision 11: 816–824. http://www.molvis.org/molvis/v11/a97/.

cytokine and a local hormone is considered moot by many. Cytokines were originally associated only with immune functions. An example of a cytokine that has both immune and nonimmune effects is the TGF-b, pointed out earlier. In the eye, TGF-b is known to modulate cell migration, proliferation, death, development, tissue repair, and many pathological processes as well. Generally, however, this cytokine is responsible for extracellular matrix production and the suppression of cell proliferation. There are three isoforms of TGF-b. Of the three, TGF-b2 is the predominant cytokine that is found in the aqueous fluid bathing the endothelium. The presence of TGF-b2 is essential to the cornea since its absence results in a failure of the endothelium to develop (see Figure 14). Immune privilege in the eye is a process in which the eye protects itself from undesirable immune characteristics, such as inflammation, as a device to preserve vision. For the anterior segment, that means the ability of light to pass through its tissues unimpeded. Immune privilege is the summation of many complex molecular and cellular mechanisms whose operation, as a whole, remains incompletely understood. It is also the process by which corneal transplants are more likely to be free of immune reactions than transplants in other parts of the body. In an immuneprivileged site, such as the corneal endothelium, active processes are set in motion to suppress immune reactions when an antigen enters the site. The aqueous humor is responsible for this action since it contains the necessary suppression cytokines that are released by cells bordering the anterior chamber. TGF-b2, in addition to its roles mentioned previously, inhibits T-cell activation and differentiation that are necessary to initiate inflammation.

278

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

6000

TGF-β1 (pg ml−1) Mean ± SD

5000 4000 3000 2000 1000 0 0

2

4

6

8

10

12

14

16

18

20

22

24

Days incubated (human, 4 ⬚C,

)

(human, 37 ⬚C, ) (rabbit, 37 ⬚C,

)

Figure 14 Levels of TGF-b1 in fluids bathing an excised cornea following mechanical wounding. At 37 C, it is seen that the levels of the cytokine, produced by the cornea, are increased as a mechanism to limit cell division about 48 h after wounding. Adapted from Whikehart, D. R., Parikh, C. H., Vaughn, A. V., Mishler, K., and Edelhauser, H. F. (2005). Evidence suggesting the existence of stem cells for the human corneal endothelium. Molecular Vision 11: 816–824. http://www.molvis.org/molvis/v11/a97/.

Other examples of cytokines that contribute to immune privilege are: a-melanocyte stimulating hormone, vasoactive stimulating peptide, calcitonin gene-related peptide, and somatostatin. A more recent finding has been the discovery of another cytokine: cluster of differentiation 95 ligand (CD95L) at immune-privileged sites. CD95L acts as a death signal that triggers apoptosis (programmed cell death) in CD95 sensitive T cells. The membrane protein form of CD95L, expressed on corneal endothelial cell membranes, has been shown to be important for the preservation of orthoptically placed corneal grafts and it does this by interacting with these T cells. Some contradictions in the understanding of immune suppression also exist. For example, it is not understood how the soluble versus membrane forms of CD95 function and why the CD95L form is important for graft acceptance. In addition, many of the cells that surround the anterior chamber have receptors for tumor necrosis factor-a. This cytokine, unfortunately, plays an important role in intraocular inflammation (endotoxin-induced uveitis) when it occurs. Proteins Synthesized for External Transport In human corneas, there are known anatomical subdivisions found in Descemet’s membrane: an anterior, banded zone and a nonbanded, amorphous zone (Figure 15). The banded zone is formed before birth while the amorphous zone is continuously synthesized during life. Four principal proteins have been found to constitute Descemet’s membrane in both zones: collagens type IV and VIII as well as laminin and fibronectin. All are considered to be synthesized by corneal endothelial cells throughout the lifetime of an individual. However, recent suggestions

BR

NBR

E

Figure 15 The banded (BR) and nonbanded (NBR) regions of Descemet’s membrane. The BR is formed before birth while the NBR is made continuously throughout life. E, the endothelium. 7000. Adapted from Forrester, J. V., Dick, A. D., McMenamin, P. G., and Roberts, F. (2008). The Eye. Basic Science in Practice, 3rd edn., p. 21. Edinburgh: Saunders.

contend that some evidence points to the posterior keratocytes as sources for some of this synthetic work. Collagen type IV was originally considered to be the principal collagen of Descemet’s membrane. It is described as the main structural protein of all basement membranes. The a-1–a-6 chains are found in Descemet’s membrane facing the stromal and endothelial sides of the membrane in both

Corneal Endothelium: Overview

The Role of the Corneal Endothelium in Deturgescence Introduction The average human cornea in the adult is maintained at a thickness of 500 mm in the center to about 700 mm at the periphery (Figure 1). Keeping these parameters stable is important for corneal clarity. Deturgescence is the physiological process of maintaining such a clear cornea. The process is active and continuous. Two forces are at work during deturgescence. In one process, the cornea swells and tends to become cloudy. Due to the large volume of negatively charged glycosaminoglycans (GAGs) bound to the proteoglycans in the cornea, there is a constant influx of cations into the stroma from fluids outside of this tissue. These cations are accompanied by water as an osmotic compensation and bring about corneal swelling. The stroma is composed of layers of collagenous lamellae with each collagen strand separated from its neighbor by an aquous space. Disruption of the geometric regularity in the aqueous spaces that contain the GAGs brings about a loss in the mutual interference of light that is refracted through the cornea. This results in the beginning of a loss of clarity which can continue through progressive degrees of opacity (cloudiness). If unchecked, this process would result in functional blindness. In the second process of deturgescence, excess water is actively transferred out of the stroma by the corneal endothelium into the aqueous fluid of the anterior chamber. The corneal epithelium takes on a rather passive role in this process by acting more as a barrier to water flow. Early experiments demonstrated the existence of this active process to be predominately in the endothelium. This was accomplished

by the selective removal of corneal outer layers to determine which side (epithelium or endothelium) was involved in pumping out water. A metabolic component was shown to exist when the deturgescent process was compared in experimental corneas at cold versus physiological temperatures (Figure 16). The Biochemistry of Active Deturgescence Active deturgescence is considered to occur as the result of the catalytic activity of two enzymes: sodium, potassiumstimulated adenosine triphosphatase (Na, K-ATPase) and bicarbonate-stimultated adenosine triphosphatase (–HCO3ATPase ). Na, K-ATPase resides in the basolateral membranes of corneal endothelial cells in sufficient quantities to ensure significant pumping activity. As an enzyme, Na, K-ATPase exists with a minimal structure of four polypeptide chains of which two (the achains) are catalytic and two (the b-chains) are structural stabilizers in the cell membrane. In carrying out pumping activity, the enzyme is energized by the hydrolysis of ATP to adenosine diphosphate (ADP) in which energy, contained in the released inorganic phosphate group, is transferred to an a-subunit. This energy provides for the kinetic transfer of two potassium ions into an endothelial cell while three sodium ions are virtually, simultaneously moved to the cell exterior. The result is the net movement of one cation outside the cell (Figure 17). The osmotic consequence of this is the simultaneous flow of water into the anterior chamber either directly or through the channels that lie between endothelial cells. Water therefore flows osmotically, due to ion transfer, through endothelial cells by means of proteins known as aquaporins.

+0.02 Δ corneal thickness (mm)

infants and adults. Type VIII collagen is a short nonfibrillar protein that may determine cell phenotype. In the infant, it is found on the endothelial face of the membrane, whereas it occurs on the stromal face of the membrane in the adult. Laminin is a noncollagenous glycoprotein that binds to other proteins to form sheets. Its association with type IV collagen in basement membranes is well known. Laminin occurs on the stromal and endothelial faces of infants, but it is not found on the stromal face of adult Descemet’s membrane. Fibronectin is a high-molecular-weight protein that can bind to membrane receptor proteins as well as collagen. It is involved in cell adhesion. Fibronectin only occurs on the stromal face of Descemet’s membrane in both infants and adults. Mucin-1 (MUC-1), a cell surface protein that has largely been considered to be present on the surface of the corneal epithelium, has also been found to be synthesized by the endothelium and transported to its apical surface. Its function there would be as an interfacing protein to the aqueous fluid.

279

0.00

−0.02

−0.04

−0.06 0

1

2 h

Figure 16 A temperature-reversal experiment in which a cornea is previously allowed to swell in the cold and then demonstrated to decrease in thickness when placed at room temperature for 2 h. From Dr. Henry Edelhauser.

280

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Transporter binds 3 Na+ from the inside of the cell EnzI

3 Na+

Phosphorylation favors P−EnzII ATP ADP

P−EnzII Transporter releases 3 Na+ to the outside and binds 2 K+ from the outside of the cell

P

P−EnzII

3 Na+ 2 K+

P

Pi +

2K

EnzI

Dephosphorylation favors EnzI Transporter releases 2K+ to the inside

Inside

Figure 18 Aquaporin-1 consists of four polypeptide chains (shown here in red and blue). The channel for water that they form is shown as a green arrow. Modified from Wikipedia/Aquaporin.

Outside

Figure 17 A postulated mechanism for the two a-subunits of Na, K-ATPase. The mechanism (beginning at the top of the figure) is initiated with the cytosolic binding of three sodium cations, followed by ATP hydrolysis, sodium external transport, two potassium cation external binding, potassium internal transport, and release. The cycle is then repeated. Adapted from Nelson, D. L. and Cox, M. M. (2005). Lehninger. Principles of Biochemistry, 4th edn., p. 399. New York: Freeman.

The water transport protein in the cornea is known as aquaporin-1 (AQP-1; Figure 18). Bicarbonate-ATPase has a role that is more difficult to define, but which has been considered by some to be supportive of the role of Na, K-ATPase. Evidence indicates that the enzyme has a molecular weight similar to the a-subunit of Na, K-ATPase. Part of the difficulty in understanding how or why this enzyme functions lies in the fact that it is resident in mitochondria rather than in plasma membranes. The enzyme functions, just as Na, K-ATPase, by obtaining energy from ATP hydrolysis and is, therefore, also a P-type ATPase. The energy incorporated into the enzyme causes it to transfer anions across membranes that include chloride as well as bicarbonate. In the literature, there has been considerable confusion about whether this enzyme is primarily a chloride or a bicarbonate transporting enzyme in situ in the corneal endothelium. Experimentally, it seems that either anion may be substituted for the other or that bicarbonate may be required to stimulate the transport of chloride. The importance of bicarbonate cannot be denied as,

experimentally, a decrease in bicarbonate, and the use of carbonic anhydrase inhibitors clearly shows a loss of deturgescence. Carbonic anhydrase converts water and carbon dioxide into bicarbonate and a proton. At this point, whether chloride or bicarbonate is of prime importance is unclear. The Physiological Control of Active Deturgescence It is evident that moving a net amount of cations from the corneal endothelium (via Na, K-ATPase) to the anterior chamber will result in transferring water from the endothelium to the aqueous to maintain an osmotic balance. This activity serves to remove excess water from the stroma (via aquaporin transport through the corneal endothelium) while water leaks back into the stroma simultaneously. Normally, the concentration of Na+ ions is higher in the anterior chamber than in the stroma, so the activity of Na, K-ATPase is required to push Na+ ions against the gradient of the anterior chamber. The two processes (pump and leak) operate in a fashion analogous to water leaking into a basement while it is being removed by a sump pump at the same time. Since Na, K-ATPase requires a substantial supply of ATP as an energy source from mitochondria, it can be speculated that the membrane potential and pH of that organelle requires that the potential and pH stability be well maintained. To what degree the operation of a –HCO3-ATPase in the mitochondria contributes to these phenomena is unknown.

Corneal Endothelium: Overview

There remain other physiological observations that challenge further investigation of the pump-leak hypothesis of deturgescence. For example, do the aquaporins (AQP-1 in the endothelium and AQP-5 in the epithelium) play more than a passive role in deturgescence? What function might TRPV4 (an osmolar-sensing protein in the epithelium) play in regulating stromal volume?

Genetic Diseases of the Corneal Endothelium Fuchs’ Endothelial Dystrophy This disease is considered to be the result of an autosomal dominant disorder. It is a true disease of the corneal endothelium rather than one that originates from another area of the cornea (Figure 19). It was first described in

281

1910. There are two forms of the disease: early onset (mutations associated with chromosomes 13 and 18) and late onset (perhaps found on chromosome 1 or at other locations). The early onset form is associated with a mutation of the COL8A2 gene which synthesizes collagen VIII, a major collagen of Descemet’s membrane. The exact identity of the gene defect in the late onset form has been proposed to be SLC4A11, a gene that causes the synthesis of a transport protein in endothelial cells. Both forms of the disease manifest themselves initially in the formation of guttae (gutattae) or buttons on the endothelium. Later development of the condition occurs with the formation of edema or swelling in the posterior stroma. The edema spreads anteriorly and leads to the development of corneal cloudiness. The stroma may take on the appearance of ground glass as the disease progresses. This disease progresses slowly and is more common in females than in males. Effective treatment is a penetrating keratoplasty (corneal transplant).

Related Posterior Membrane Dystrophies

(a)

Two dystrophies that have some similarities to Fuch’s dystrophy are posterior polymorphous dystrophy (PPMD) and congenital hereditary endothelial dystrophy (CHED). PPMD has been associated with four gene mutations on chromosomes 10 and 20 that affect the synthesis of collagens IV and VIII. In PPMD, there are differences in the effects on Descemet’s membrane and the corneal endothelium versus the effects that occur in Fuchs’ dystrophy. The nonbanded zone of Descemet’s membrane becomes absent or minimal while the endothelial cells can take on the appearance of epithelial cells. PPMD patients are often asymptomatic, but those receiving penetrating keratoplasties sometimes develop increased intraocular pressure associated with iris attachments (synechiae). By contrast in CHED, the genetic abnormalities have been associated with one or more gene mutations only on chromosome 20. There is a decided loss of endothelial cells rather than a thinning of cells as occurs with Fuchs’ dystrophy. The outcome of penetrating keratoplasty for CHED patients is mixed. See also: Regulation of Corneal Endothelial Cell Proliferation; Regulation of Corneal Endothelial Function.

(b) Figure 19 Endothelial cell layers as viewed from the posterior side of the cornea: (a) is a normal endothelium showing a regular pattern of polygonal cells and (b) is an example of moderately advanced Fuch’s dystrophy. It shows a marked decrease in cell numbers, abnormal nuclei, and cell displacement by guttatae from Descemet’s membrane. Adapted from Kratchmer, J. H., et al. (eds.) (2005). The Cornea, 2nd edn., vol. 1, p. 942. Philadelphia, PA: Elsevier.

Further Reading Cveki, A. and Tamm, E. R. (2004). Anterior eye development and ocular mesenchyme: New insights form mouse models and human diseases. BioEssays 26: 374–386. Forrester, J. V., Dick, A. D., McMenamin, P. G., and Roberts, F. (2008). The Eye. Basic Sciences in Practice, 3rd edn. Edinburgh: Saunders/ Elsevier.

282

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Gerencher, G. A. and Zhang, J. (2003). Chloride ATPase pumps in nature: Do they exist? Biological Reviews 78: 197–218. Hogan, M. J., Alvarado, J. A., and Weddell, J. E. (1971). Histology of the Human Eye. Philadelphia, PA: Saunders. Krachmer, J. H., Mannis, M. J., and Holland, E. J. (eds.) (2005). Cornea, 2nd edn., vol. 1, Philadelphia, PA: Mosby/Elsevier. Kratchmer, J. H., Mannis, M. J., and Holland, E. J. (eds.) (2005). The Cornea, 2nd edn., vol. 1, p. 942. Philadelphia, PA: Elsevier. McCarey, B. E., Edelhauser, H. F., and Lynn, M. J. (2008). Review of corneal endothelial specular microscopy for FDA trials of refractive procedures, surgical devices, and new intraocular drugs and solutions. Cornea 27: 1–16. McGowan, S. L., Edelhauser, H. F., Pfister, R. R., and Whikehart, D. R. (2007). Stem cell markers in the human posterior limbus and corneal endothelium of unwounded and wounded corneas. Molecular Vision 13: 1984–2000. Medina-Martinez, O. and Jamrich, M. (2007). Foxe view of lens development and disease. Development 134: 1455–1463. Mimura, T. and Joyce, N. C. (2006). Replication competence and senescence in central and peripheral human corneal endothelium. Investigative Ophthalmology and Visual Science 47: 1387–1396. Streilein, J. W. and Stein-Streilein, J. (2000). Does innate immune privilege exist? Journal of Leukocyte Biology 67: 479–486. Verkman, A. S. (2005). Aquaporins in endothelia. Kidney International 69: 1120–1123. Vithana, E. N., Morgan, P. E., Ramprasad, V., et al. (2008). SLC4A11 mutations in Fuchs endothelial corneal dystrophy. Human Molecular Genetics 17: 656–666.

Whikehart, D. R. (2003). Biochemistry of the Eye, 2nd edn. Philadelphia, PA: Butterworth-Heinemann/Elsevier. Whikehart, D. R., Parikh, C. H., Vaughn, A. V., Mishler, K., and Edelhauser, H. F. (2005). Evidence suggesting the existence of stem cells for the human corneal endothelium. Molecular Vision 11: 816–824. Wilson, S. E., Weng, J., Blair, S., He, Y. G., and Lloyd, S. (1995). Expression of E6/E7 or SV40 large T antigen-coding oncogenes in human corneal endothelial cells indicates regulated highproliferative capacity. Investigative Ophthalmology and Visual Science 36: 32–40. Zhu, C., Rawe, I., and Joyce, N. C. (2008). Differential protein expression in human corneal endothelial cells cultured from young and older donors. Molecular Vision 14: 1805–1814.

Relevant Websites http://www.eyesite.org – Diagrams and a video on corneal endothelial transplantation, The Eyesite. http://dev.biologists.org – The Company of Biologists Ltd: Development. (This on-line paper reports about the deleterious effects of either TGFa or EGF on corneal endothelial development in a transgenic mouse.).

Regulation of Corneal Endothelial Function J A Bonanno and S P Srinivas, Indiana University, Bloomington, IN, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Aquaporins – The water channels that facilitate water movement across the plasma membrane. Deturgescence – The removal of water from the cornea stroma to counteract edema. Guttata – Excresances of Descemet’s membrane (basement membrane of the corneal endothelium) produced by abnormal endothelial cells. Peri-junctional actomyosin ring (PAMR) – A Dense band of actin cytoskeleton found proximal to the apical junction complex. Pump-leak mechanism – The maintenance of corneal deturgescence in which endothelial active fluid transport (the pump) exactly counters the passive leak directed into the stroma. Tight junctions – The intercellular junctions at the apical domain of endothelial cells which occlude the paracellular space. Transendothelial electrical resistance (TER) – Dependent on the integrity of the tight junctions.

hydrophilic and exert a swelling pressure leading to imbibition of water across the relatively leaky endothelium at the posterior surface. Thus, if a bare piece of stroma is placed in saline, it will swell to many times the normal thickness. This stromal edema produces large variations in collagen-fiber spacing and consequentially increased light scatter, corneal haze, and in vivo produces diminished visual acuity. Loss of endothelial cells from surgical trauma or disease, for example, Fuchs’ endothelial dystrophy, results in corneal edema. These observations together with in vitro physiological studies indicate that the maintenance of stromal hydration is primarily dependent on the corneal endothelium. The pioneering work of David Maurice first showed that the corneal endothelium actively pumps water from stroma to anterior chamber. This pump exactly counterbalances the leak into the stroma, which is driven by the GAG-dependent swelling pressure. Therefore, the stromal hydration is maintained relatively constant and stromal transparency is preserved. This is often called the Pump–Leak hypothesis for maintenance of corneal hydration and is illustrated in Figure 1. Endothelial Barrier Function

Corneal Endothelial Function Stromal Swelling Pressure and Maintenance of Transparency The transparency of the cornea stroma is dependent on the tissue hydration. The stroma is composed of 200 or more lamellae, 2-mm thick with widths varying from 10 to 250 mm that span from limbus to limbus and overlap each other at varying orientations. The total thickness of the human stroma varies from 450 to 550 mm. Each lamella is composed of parallel strands of collagen (type I) and associated glycosaminoglycans (GAGs). Between the layers of overlapping lamellae are the keratocytes – a very flat, stellate-shaped cell that is responsible for producing collagen and GAGs and maintaining the stromal structure. The GAGs act as spacers between the collagen fibers. At normal stromal hydration, this space is 30 nm and is very uniform. Because the refractive index of collagen and the GAGs ground substance are significantly different, a random orientation of collagen fibers and varying fiber diameter (which is characteristic of the sclera) would produce an opaque tissue. However, because there is almost no variation in fiber diameter and the spacing between fibers is uniform, light scattered offaxis through the stroma is <10%. GAGs, however, are very

The tight junctions (TJs) of the corneal endothelium, although leaky (trans-endothelial resistance (TER) 25 O cm2), restrain fluid leak into the stroma. This constitutes the barrier function of the endothelium, and complements fluid-pump function in the regulation of stromal hydration. Despite the leakiness of the endothelium, breakdown of its TJs in the absence of an increase in fluid-pump activity results in corneal edema. In addition to this direct effect, edema could be enhanced further by the fact that TJs also influence the fluid pump function through two indirect mechanisms. First, intact TJs prevent dissipation of local osmotic gradients across the endothelium set up by ion-transport mechanisms by restraining solute back-flux through the paracellular space (i.e., gate function of TJs). Secondly, intact TJs are indispensable for the maintenance of apical-basal polarity of the ion-transport proteins. This is achieved by limiting their lateral diffusion of the membrane proteins (i.e., fence function of TJs). When the polarity of the transport mechanisms is compromised, a vectorial ionic movement and hence fluid transport cannot occur. Thus, the TJs of the endothelium not only restrain fluid leak into the stroma, but also form a principal determinant of the endothelial fluid-pump activity (Figure 2).

283

284

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Nutrition/Waste Removal One requirement for corneal transparency is the absence of blood and blood vessels. As such, corneal nutrition must come from the tears, the limbus, and/or the anterior chamber. The stratified squamous epithelium is a very tight barrier, and there are no transporters for nutrients such as glucose. However, the epithelium is permeable to hydrophobic nutrients such as oxygen and wastes, for example, carbon dioxide (CO2). Because of the long diffusion distances from the limbus, its contribution to overall corneal nutrition is minimal. Therefore, most of the nutrition and removal of wastes – for example, lactic acid – occurs across

the endothelium. The corneal endothelium expresses glucose transporters to facilitate uptake into the stroma to nourish keratocytes and the corneal epithelium. Similarly, the endothelium expresses lactic acid transporters to remove this end product of anaerobic glycolysis. Approximately 85% of the glucose consumed by the cornea goes through anaerobic glycolysis because of the relative paucity of mitochondria in the epithelial cells and keratocytes. The relative lack of mitochondria is presumably another strategy to minimize light scatter and enhance transparency.

Corneal Endothelial Transport Active Transport Epithelium

Keratocytes Collagen fibrils w/GAGs coating

Stroma

Aqueous

Endothelium Leak

Active fluidpump

Normal endothelium

Figure 1 Pump–Leak hypothesis for maintenance of corneal hydration. Stromal glycosaminoglycans (GAGs) are negatively charged hydrophilic molecules that exert a swelling pressure that draws water across the limiting layers (epithelium and endothelium). This is the tissue leak. The endothelial pump must exactly counterbalance the leak so that stromal hydration and transparency are maintained.

Following discovery that the endothelium was responsible for maintaining stromal hydration, it was shown that the endothelial pump was dependent on active transport. In contrast to corneal epithelial cells and keratocytes, endothelial cells have a very high mitochondrial density. Poisoning the mitochondria reduces the pump activity indicating that it is dependent on the availability of adenosine triphosphate (ATP). Furthermore, exposure of endothelium to the cardiac glycoside ouabain – which blocks the membrane Naþ,Kþ-ATPase – also inhibits the pump activity. The Naþ,Kþ-ATPase provides the ion gradients across cell membranes that drive secondary transporters, for example, Naþ:Kþ:2Cl cotransport, Naþ/Hþ exchange, and Naþ : 2HCO 3 cotransport, and anion channels that may participate in ion-coupled fluid secretion. A classic fluid-transport mechanism might include basolateral cotransporters and apical anion channels providing a pathway for vectorial ion fluxes that could be osmotically coupled to water transport.

Endothelial dysfunction

Tears

Phaco emulsification SP < 50 mmHg

SP = 50 mmHg Corneal graft rejection

Inflammatory mediators (e.g., TNF-α) Tight junctions

Leak

Pump

Tight junctions

Leak

Pump ???

Figure 2 Barrier integrity is essential for stromal hydration control: Fluid leak into the stroma through the paracellular space – determined by the barrier integrity – is driven by a hydraulic gradient equivalent to swelling pressure (SP) of 50 mmHg at normal stromal hydration. Despite the leaky nature of the endothelium (TER < 25 O cm2), stromal thickness is held constant by the fluid-pump mechanism, which counterbalances the fluid leak. When the barrier integrity breaks down, the pump mechanism cannot cope with the leak and hence stromal edema becomes inevitable. Inflammatory stress – which can reduce the barrier integrity – is not known to stimulate the pump function concomitantly. In fact, when the tight junctions are compromised, the fluid-pump mechanism cannot be sustained since the local osmotic gradients generated by ion transport are dissipated by futile solute back-flux.

Regulation of Corneal Endothelial Function

transports one Naþ with two HCO 3 ions into the cells from the stroma. Application of the anion transport inhibitor 4,40 -diisothiocyanatostilbene-2,20 -disulphonic acid (DIDS) or genetic knockdown of NBCe1 expression using siRNA significantly reduces bicarbonate uptake and transendothelial bicarbonate flux. DIDS also results in corneal swelling in vitro, suggesting that this anion-transport process has an important role in the endothelial pump activity. NBCe1 is responsible for the large Naþ-dependent (Cl independent) bicarbonate permeability of the basolateral membrane. The bicarbonate permeability of the apical membrane of the corneal endothelium is about one-third of basolateral membrane and it is independent of Naþ or Cl, suggesting that the smaller apical permeability is conferred by a channel. To date, only two apical anion channels have been described in the corneal endothelium – the cystic fibrosis transmembrane regulator (CFTR) and a calciumactivated chloride channel (CaCC). These channels are permeable to both Cl and HCO 3 in about a 5:1 ratio. Physiological experiments with rabbit corneas have shown, however, that CFTR-channel inhibitors have no effect on the endothelial pump rate. Furthermore, siRNA knockdown of CFTR in cultured cells – while inhibiting cyclic adenosine monophosphate (cAMP)-activated anion flux, did not change the basal bicarbonate flux. This is consistent with clinical studies that indicate normal corneal thickness and function in cystic fibrosis patients. Similarly, siRNA knockdown of CaCC channels had no effect on basal bicarbonate permeability, but reduced calcium-activated flux. To date, the nature of the apical HCO 3 transport remains unknown. Recent studies have also shown that removal of Cl will also produce significant corneal swelling in vitro,

Bicarbonate/Carbonic Anhydrase Early studies have shown that the endothelial pump is significantly inhibited in the absence of bicarbonate suggesting that bicarbonate transporters and anion channels may be components of the endothelial pump. A role for bicarbonate was strengthened when it was found that carbonic anhydrase inhibitors (CAIs) – applied directly to the endothelium – also slowed the pump. With the advent of topical carbonic anhydrase inhibitors to lower intraocular pressure (IOP), there was some concern that this could cause corneal edema. However, numerous clinical studies have shown that in humans with normal endothelial cell counts, topical CAIs do not cause corneal edema. Only when cell counts are low and/or in the presence of significant endothelial guttata have topical CAIs been shown to cause corneal edema. This suggests that whatever role bicarbonate/CA activity has in pump activity, there must be a large functional reserve. Interestingly, endothelial pump activity can be maintained in the absence of bicarbonate, but only when bicarbonate is substituted with a high concentration of another buffer. These observations suggest that part of the role of bicarbonate/CA in the endothelial pump may be its buffering capacity, possibly for lactic acid. Anion Transporters and Channels Several anion transporters and channels that could participate in a bicarbonate secretory pump mechanism are expressed in corneal endothelium. At the basolateral (stromal side) membrane, the sodium bicarbonate cotransporter (NBCe1) is highly expressed (Figure 3). This protein Anterior chamber

HCO3− + H+

TEP = −0.5 mV

CO2

CA IV

Tight junction

CaCC

CFTR

285

Tight junction

?

Cl−, HCO3− Cl−, HCO3− Cl−, HCO3−

Na+ Na pump

Na+

Cl−

K+

AE2

CA ll

K+

HCO3− + H+

HCO3− CO2

NBCe1

Stroma

Na+

K+

2Cl−

Na+

2HCO3−

Figure 3 Model for transendothelial bicarbonate transport. NBCe1 uses the inward [Na+] gradient to transport bicarbonate into the cell at the basolateral surface. This is facilitated by carbonic anhydrase II (CAII). Anion channels at the apical surface that are permeable to bicarbonate can provide an efflux pathway. CAIV is present on the apical surface and may facilitate net HCO 3 flux. A chloride transport pathway is also in place, which may only be activated during endothelial stress.

286

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

suggesting the possibility that both anions – Cl and HCO 3 – participate directly in the pump mechanism. In fact, the basolateral membrane expresses the Naþ: Kþ:2Cl cotransporter and the cytoplasmic [Cl] is 35 mM, which is above electrochemical equilibrium (Figure 3). Thus the potential for apical anion channel mediated Cl flux is in place. However, the highly specific inhibitor of the Naþ:Kþ:2Cl cotransporter – bumetanide – has no effect on corneal thickness in vitro. In addition, studies examining transendothelial Cl fluxes have been equivocal. This indicates that a basal Cl flux across the endothelium is unlikely to have a role in the pump mechanism. The presence of Cl may have two other roles that could support HCO 3 fluxes. First, because NBCe1 is electrogenic (1 Naþ : 2HCO 3 ), membrane potential hyperpolarization (e.g., from 50 to 65 mV) will stop efflux through NBCe1 dependent HCO 3 influx. Cl CFTR or other unidentified anion channels could dissipate the hyperpolarizing effects of NBCe1 activity. Second, the anion exchanger AE2 is expressed on the basolateral membrane as shown in the rabbit corneal endothelium. Thus, the transmembrane Cl gradient can help regulate intracellular [HCO 3 ]. Further studies are needed to elucidate the role of Cl or Cl transporters in the corneal endothelial pump mechanism. Figure 3 illustrates a possible model for bicarbonate transport.

differences in the lateral spaces between the cells and/or on the apical surfaces of cells within an unstirred layer. These osmotic gradients are the driving forces for water movement across cellular membranes. This standing gradient osmotic theory, first developed by Diamond and Bossert, has come under fire as a general mechanism for fluid transport. For example, in many epithelial cells – including the corneal endothelium – there is no evidence that these gradients exist. This has led to the consideration of other mechanisms, such as electro-osmosis, which has been championed by Jorge Fischbarg for the corneal endothelium. In this theory, cells generate a transepithelial potential (0.5 mV apical-side negative in corneal endothelium) that draws counter-ions – for example, Naþ – through a paracellular pathway that is ion specific. This produces electro-osmotic coupling across the TJ. Another mechanism for fluid transport across epithelial cells that has been recently developed is the cotransporter model. In this model, water is directly transported or coupled with the movement of the associated ions and metabolites within the cotransporter protein. Since the endothelium expresses several cotransporters – for examþ þ cotransporters – ple, 1Naþ:2HCO 3 and Na :K :2Cl and at least two monocarboxylic acid (lactic acid) cotransporters – MCT1 and MCT2 – there is potential for downhill fluid transport by this mechanism.

Aquaporin-1 (AQP1)

Barrier Integrity

Aquaporin-1 (AQP1) water channels are highly expressed on both apical and basolateral membranes of corneal endothelium. No other AQP channel is expressed in the corneal endothelium. AQP1 confers a very high osmotic permeability and allows for rapid cell-volume regulation in response to anisosmotic solutions. In AQP1-knockout mice, corneal endothelial cell osmotic permeability is reduced, as expected. Corneal de-swelling rates are also significantly reduced. This indicates that a significant amount of water flux across the endothelium – at least under non-steady-state conditions – is transcellular. However, steady-state corneal thickness of these mice is slightly thinner than wild-type mice. This result may indicate that the pathway for water fluxes driven by the pump and the leak (GAGs-dependent stromal swelling pressure) is the same, which would result in no net effect. Thus AQP1 (and AQP5 in the corneal epithelium) may be present to increase the rate of water fluxes in response to osmotic changes that can occur, for example, following eye closure or exposure to hypotonicity that occurs during swimming.

Barrier integrity – which implies resistance to diffusion of solutes or fluid leak through the paracellular pathway – is dependent on the TJs. TJs are supramolecular assemblies localized at the apical domain of epithelial/endothelial monolayers. The transmembrane molecules (occludins, claudins, and junctional adhesion molecule ( JAM)) – which are all expressed in corneal endothelium associated with the TJs – these transmembrane molecules of one cell interact with their homotypic counterparts in the neighboring cells. This interaction – facilitated by intercellular tethering forces through Ca2þ-dependent adherence junctions (AJs) – brings about occlusion of the paracellular space. The cytoplasmic domains of the transmembrane molecules of the TJs are structurally linked to a thick band of cortical actin cytoskeleton (called peri-junctional actomyosin ring (PAMR)) via adapter molecules such as zona occludens-1 (ZO-1). For AJs, the association with PAMR is mediated through catenins. The adapter molecules of both AJs and TJs also form a scaffold for a number of signaling molecules which are now implicated in the control of the stability of AJs and TJs and hence in the regulation of barrier integrity. Recent studies have demonstrated that increased contractility of the PAMR induces a breakdown of the barrier integrity through a reduction of the intercellular tethering forces at the TJs and AJs and also possibly through redistribution

Pump Mechanism The conventional view of epithelial cell secretion and absorption of water is that these cells create local osmotic

Regulation of Corneal Endothelial Function

of the AJ and TJ molecules at the apical junctional complex. An increase in contractility of the actin cytoskeleton is induced by an increase in the phosphorylation of the regulatory light chain of myosin II (also called myosin light chain or MLC; 20 kDa). Recent studies have shed light on the importance of actomyosin contraction in the regulation of corneal endothelial barrier integrity. MLC Phosphorylation and Actomyosin Contraction Phosphorylation of myosin light chain (MLC) – which is bound to the motor protein myosin II – induces actomyosin interaction resulting in increased contractility of the actin cytoskeleton (Figure 4). The extent of MLC phosphorylation is regulated by two opposing pathways: myosin light-chain kinase (MLCK)-driven phosphorylation, and myosin light-chain phosphatase (MLCP)-driven dephosphorylation. MLCK is activated after binding to the Ca2þ-calmodulin complex and its activity is dedicated to MLC phosphorylation. MLCK activity is also modulated by other protein kinases by direct phosphorylation, especially the large-size isoform of MLCK called vascular

287

endothelial cell MLCK (EC-MLCK; 220 kDa). MLCP is a hetero-trimeric complex consisting of PP1Cd (the catalytic subunit), the myosin-binding subunit (MYPT1; 130 kDa), and a small subunit of unknown function (M20). Phosphorylation of MYPT1 by Rho kinase (at Thr-696 and Thr-850) – a downstream effector of the small GTPase RhoA – inhibits the phosphatase activity of PP1Cd. Protein kinase C (PKC) isoforms also inhibit the activity of MLCP through phosphorylation of CPI-17 (PKC-activated 17-kDa inhibitor protein of type 1 phosphatase) and consequent inactivation of PP1Cd. Thus, activation of Rho kinase and/or PKC results in contraction of the actin cytoskeleton and a breakdown in barrier integrity. Effect of MLC Phosphorylation on Corneal Endothelium Barrier Integrity The mechanisms associated with MLC phosphorylation – which are well characterized in smooth muscle cells – are also present in corneal endothelial cells as demonstrated in several recent studies. Thus, it is now known that corneal endothelial cells – similar to vascular endothelial

Inactive

(a) pMLCK

MLCK ↑[Ca2+]i

PKA

[Ca2+]4–CaM

Active CaM MLCK–[Ca2+]4–CaM

(b) MLC

Relaxation

(c)

pMLC

Active MLCP

Inactive MLCP M20

PP1Cδ

M20

MYPT1

(d)

Contraction

PP1Cδ

Inactive MLCP M20

MYPT1

PP1Cd MYPT1

Plasma membrane P RhoA-GDP

RhoA-GTP

Rho kinase

GEF

PKA RhoA-GDP GDI

GAP

GTP

pCPI-17 PKC

GDP

GDI: GDP dissociation inhibitors GAP: GTPase activating proteins GEF: Guanine nucleotide exchange factor

CPI-17

Figure 4 Regulation of myosin light-chain (MLC) phosphorylation in the corneal endothelium. (a) MLC phosphorylation is catalyzed by Ca2+-Calmodulin-dependent myosin light-chain kinase (MLCK). Expression of both endothelial and smooth muscle isoforms is known in corneal endothelium. (b) MLC phosphorylation promotes actomyosin contraction. (c) MLCK activity is opposed by myosin light-chain phosphatase (MLCP), which catalyzes dephosphorylation of pMLC. MLCP is a heterotrimeric complex consisting of MYPT1 (a regulatory subunit), PP1Cd (the catalytic subunit), and M20 (function unknown). (d) Rho kinase – effector of RhoA – phosphorylates MYPT1. This inhibits PP1Cd. When RhoA is phosphorylated by protein kinase A (PKA) at its Ser-188, dissociation of RhoA-GDI from RhoA-GDP is opposed. Guanine exchange factors (GEFs) promote the release of GDP and subsequent binding to GTP; G-protein accelerating proteins (GAPs) stimulate the GTPase activity of RhoA; guanine nucleotide-dissociation inhibitor (GDIs) stabilize the inactive state of RhoA. Yet unidentified isoforms of PKC phosphorylate CPI-17 leading to inhibition of PP1Cd.

288

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

cells – possess both EC-MLCK and a smaller smooth muscle isoform of MLCK (SM-MLCK; 120 kDa). In addition, myosin II isoforms (myosin IIA and IIB) are also expressed in the endothelium. To show the importance of MLC phosphorylation and its effects on the corneal endothelial barrier integrity, several G-protein-coupled receptor (GPCR) agonists have been employed. Thus, thrombin – which activates the RhoA–Rho-kinase axis through Ga12/13-coupled PAR-1 receptors in corneal endothelial cells – led to an increase in MLC phosphorylation. This was found to disrupt PAMR with a concomitant breakdown of the barrier integrity. Pretreatment with ML-7 – a selective inhibitor of MLCK – could not significantly block the thrombin-induced MLC phosphorylation. However, the Rho-kinase inhibitor Y-27632 was effective in blocking the thrombin effect. These findings indicate the dominance of the RhoA-mediated Ca2þindependent pathway downstream of PAR-1 activation. The importance of the Ca2þ- and PKC-dependent mechanisms in MLC phosphorylation, however, have also been demonstrated independently through activation of histamine H1 receptors, which are coupled to Gaq/11 G-protein and are not known to activate RhoA. Thus, histamine-induced MLC phosphorylation and the resultant loss of endothelial barrier integrity could be suppressed by ML-7 and chelerythrine (a nonspecific PKC inhibitor). Taken together, these data have emphasized a strong role for actin cytoskeleton in the regulation of the endothelial barrier integrity and also suggest that bioactive factors and proinflammatory mediators found in the aqueous humor may influence the barrier function through their respective cell surface receptors. Transplantation Failure and Tumor Necrosis Factor Tumor necrosis factor (TNF-a) is a 17-KDa proinflammatory cytokine implicated in corneal endothelial failure during graft rejection (Figure 2). TNF-a mRNA has been detected in corneal allografts undergoing rejection, and TNF-a protein levels are significantly elevated in the aqueous humor and the serum of hosts that reject corneal allografts. A study with rabbit corneas has demonstrated that TNF-a breaks down the barrier integrity of the endothelial cells concomitant with disruption of actin cytoskeleton. The cytokine is well known to breakdown barrier integrity in vascular endothelial cells. Some of the important molecular mechanisms involved include activation of RhoA, MLC phosphorylation, significant loss of PAMR, microtubule disassembly secondary to activation of p38 MAP kinase, mobilization of oxidative stress, and formation of stress fibers. In recent studies with bovine corneal endothelium, it is becoming evident that TNF-a also induces disassembly of microtubules concomitant

with a gradual loss of barrier integrity and disappearance of PAMR. These effects could be blocked by pretreatment with paclitaxel, SB-203580 (a selective inhibitor of p38 MAP kinase), and inhibitors of matrix metalloproteinases. These results suggest that a number of mechanisms are involved in the breakdown of the barrier integrity in response to TNF-a which converges directly and indirectly on actin and microtubule cytoskeleton.

Regulation of Transport Activity and Barrier Integrity Adenosine, Soluble Adenylate Cyclase, and cAMP The corneal endothelial fluid pump is thought to proceed at one rate and when this rate matches the leak rate driven by the stromal GAGs, corneal hydration and thickness reaches a steady state. There are other forces that can influence the leak rate including evaporative loss across the epithelial surface and contact lens-induced hypoxic stimulation of epithelial lactate production, which can produce substantial additions to stromal osmotic pressure. There is no evidence that the pump speeds up when the cornea is edematous or slows down if the cornea thins. Shortly after it was clear that the endothelium was actively transporting water, it was discovered that the rate of fluid transport could be increased by the addition of adenosine. Later, it was shown that adenosine increases intracellular [cAMP] in endothelial cells through activation of adenosine A2b receptors. Other approaches that increase cAMP within the cells – for example, stimulating adenylate cyclase (AC) directly or inhibiting phosphodiesterase – also increased corneal endothelial fluid-transport rates. More recently, the expression of a new type of AC, called soluble AC (sAC), was shown in corneal endothelium. Unlike the transmembrane-linked adenylate cyclases, sAC is distributed throughout the 2þ cytoplasm and it is activated by HCO 3 and Ca . Because þ of the robust 1Na :2HCO3 cotransporter in endothelial cells, the sAC is active and raises the basal [cAMP] by 50%, suggesting that it may have a small role in maintaining basal fluid transport rates. Raising cAMP activates protein kinase A (PKA) and this demonstrably phosphorylates the apical CFTR channel, increases apical Cl and HCO 3 permeability, and increases transendothelial HCO 3 flux. Together, this could contribute to the increased fluid transport observed by increasing cAMP. Corneal endothelial cells can produce adenosine from ATP at the apical surface. When stressed, corneal endothelial cells release more ATP – which, when converted to adenosine, will enhance fluid transport and could help counter the negative effects of the stress.

Regulation of Corneal Endothelial Function

Role of cAMP–PKA axis in the Regulation of Barrier Integrity An in vitro study with rabbit cornea first showed that adenosine could also promote corneal deturgescence through enhanced barrier integrity. Consistent with this finding, several recent studies have shown that adenosine induces MLC dephosphorylation through mobilization of cAMP–PKA axis via A2b receptors as noted above. More importantly, consistent with MLC dephosphorylation, exposure to adenosine led to an increase in the barrier integrity as measured by the trans-endothelial electrical resistance. Similar findings were noted with extracellular ATP – which was found to undergo extracellular hydrolysis resulting in formation of adenosine and subsequent activation of A2b receptors. In addition to these findings, forskolin (direct activator of adenylate cyclase), adenosine, and ATP have been found to overcome thrombinand histamine-induced MLC phosphorylation as well as loss of barrier integrity. At the molecular level, the locus of action of elevated cAMP is also becoming evident. PKA is known to induce MLC dephosphorylation through modulation of the RhoA–Rho-kinase axis. One potential mechanism involves direct phosphorylation of RhoA by PKA and consequent increase in the affinity of the small guanosine triphosphatase (GTPase) to its guanosine diphosphate (GDP) dissociation inhibitor (GDI). An alternative mechanism involves direct phosphorylation of MYPT1 by PKA. The latter is known to prevent Rho kinase from phosphorylating MYPT1 leading to inactivation of MLCP. cAMP may also influence cell–cell adhesion through activation of a small GTPase, namely Rap1. This involves activation of Epac – a guanine nucleotideexchange factor (GEF) for Rap1. Activated Rap1 promotes formation of AJs, presumably through enhanced cadherin ligation.

Summary and Perspective The corneal endothelium is responsible for maintaining corneal hydration and transparency. Active transport processes – through mechanisms that are not fully elucidated – provide a Pump that exactly counterbalances the stromal glycosaminoglycan induced Leak. The Pump

289

is regulated through cAMP-dependent signaling that acts on components of ion transport and the barrier integrity of the endothelial monolayer. Further understanding of the Pump and Leak mechanisms are needed to provide medical therapies that could maintain stromal deturgescence in diseased or traumatized corneal endothelium. See also: Corneal Endothelium: Overview; The Corneal Stroma; Regulation of Corneal Endothelial Cell Proliferation.

Further Reading Bonanno, J. A. (2003). Identity and regulation of ion transport mechanisms in the corneal endothelium. Progress in Retinal and Eye Research 22(1): 69–94. Dikstein, S. and Maurice, D. M. (1972). The metabolic basis to the fluid pump in the cornea. Journal of Physiology 221(1): 29–41. Doughty, M. J. and Maurice, D. M. (1988). Bicarbonate sensitivity of rabbit corneal endothelium fluid pump in vitro. Investigative Ophthalmology and Visual Science 29(2): 216–223. Fischbarg, J., Diecke, F. P., Iserovich, P., and Rubashkin, A. (2006). The role of the tight junction in paracellular fluid transport across corneal endothelium. Electro-osmosis as a driving force. Journal of Membrane Biology 210(2): 117–130. Fischbarg, J. and Lim, J. (1974). Role of cations, anions, and carbonic anhydrase in fluid transport across rabbit corneal endothelium. Journal of Physiology 241: 647–675. Hodson, S. and Miller, F. (1976). The bicarbonate ion pump in the endothelium which regulates the hydration of rabbit cornea. Journal of Physiology 263: 563–577. Li, J., Sun, X. C., and Bonanno, J. A. (2005). Role of NBC1 in apical and basolateral HCO 3 permeabilities and transendothelial HCO3 fluxes in bovine corneal endothelium. American Journal of Physiology. Cell Physiology 288(3): C739–C746. Maurice, D. (1972). The location of the fluid pump in the cornea. Journal of Physiology 221: 43–54. Riley, M., Winkler, B., Starnes, C. A., and Peters, M. I. (1996). Adenosine promotes gulation of corneal hydration through cyclic adenosine monophosphate. Investigative Ophthalmology and Visual Science 37: 1–10. Satpathy, M., Gallagher, P., Lizotte-Waniewski, M., and Srinivas, S. P. (2004). Thrombin-induced phosphorylation of the regulatory light chain of myosin II in cultured bovine corneal endothelial cells. Experimental Eye Research 79: 477–486. Srinivas, S. P., Satpathy, M., Gallagher, P., Larivie`re, E., and Van Driessche, W. (2004). Adenosine induces dephosphorylation of myosin II regulatory light chain in cultured bovine corneal endothelial cells. Experimental Eye Research 79: 543–551. Srinivas, S. P., Satpathy, M., Guo, Y., and Anandan, V. (2006). Histamine-induced phosphorylation of the regulatory light chain of myosin II disrupts the barrier integrity of corneal endothelial cells. Investigative Ophthalmology Visual Science 47: 4011–4018.

Regulation of Corneal Endothelial Cell Proliferation Q Lu, T A Fuchsluger, and U V Jurkunas, Schepens Eye Research Institute, Boston, MA, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Cell cycle – Series of events that occur in order for a cell to divide. This cycle is divided into phases where the G0 (gap 0) is resting or quiescent phase; G1 (gap 1) is the synthesis of enzymes necessary for DNA replication; S phase is when DNA synthesis occurs; G2 (gap 2) phase involves the production of microtubules; M phase is the division of the cell into daughter cells. Corneal dystrophies – Conditions in which the cornea is altered without the presence of any inflammation, infection, or other eye disease. Cyclins and cyclin-dependent kinases (CDKs) – Cyclins act as the regulatory subunits while CDKs act as the catalytic subunits of an activated heterodimer. Neither cyclins nor CDKs are active in the absence of one another. CDKs are constitutively expressed in cells, whereas cyclins are synthesized at specific stages of the cell cycle, in response to various molecular signals. Descemet’s membrane – A specialized form of extracellular matrix separating corneal endothelial cells from corneal stroma. E2F – A group of genes that encodes a family of transcription factors. Epidermal growth factor (EGF) – Compound that promotes cell growth and differentiation. Mitogen – A chemical substance that induces cell division. Retinoblastoma gene – A gene whose protein is dysfunctional in many types of cancer. Tight junction – A junction between two cells composed of the junctional membrane. This acts as a selective barrier to small molecules and as a total barrier to large molecules. Transforming growth factor (TGF) – Used as a polypeptide growth factors, TGF is produced in many cell types and is involved in cellular development.

Background Mammalian corneal endothelial cells (CECs) are derived from neural crest cells of mesenchyme and form a monolayer in the inner portion of the cornea (Figure 1). The corneal

290

endothelium is attached to the Descemet’s membrane (DM) and is in direct contact with the aqueous humor. Its main function is to maintain corneal transparency by regulating corneal hydration. CECs contain numerous Na+–K+-adenosine triphosphatases (ATPases) that pump fluid out of the stroma to counteract the corneal tendency to swell. In addition, tight junctions between the CECs provide a barrier that prevents the influx of fluid from aqueous humor into the stroma. CEC number gradually declines with age (Figure 2). Since CECs do not proliferate in vivo and have a limited ability to regenerate, the loss of endothelial cells is permanent. Certain corneal pathologies, such as dystrophies, infections, and trauma, can lead to an accelerated loss of CECs and a compromise in the endothelial cell layer integrity. As a result, the cornea is unable to maintain its water balance and corneal edema ensues, clinically resulting loss of clarity and a decline in visual acuity.

Cell Cycle Progression, G1/S Transition and its Cell Cycle Regulators Mammalian cell cycles can be divided into four phases, including G1 phase (gap phase 1), S phase (DNA synthesis), G2 phase (gap phase 2), and M phase (mitosis). The entire process of cell cycle progression is controlled by a variety of regulatory proteins (Figure 3). Among them, cyclindependent kinases (CDKs) are the engine cores that promote cell cycle progression. CDKs generally remain at a constant level throughout the cell cycle, while their binding partners, cyclins, and post-translational modifiers, kinases and phosphatases, undergo periodic fluctuations throughout the cell cycle. During this process, negative regulatory protein cyclin-dependent kinase inhibitors (CDKIs) also play an important role in cell cycle progression. During the G1/S transition (Figure 3), CDKs 2, 4, 6 and their regulators control the cell transition to the S-phase. Association of CDK4 or CDK6 with D-type cyclins (Cyc D) is critical for G1 phase progression, whereas the association of CDK2 and cyclin E (Cyc E) fosters the initiation of the S phase. Both complexes, Cyc D–CDK4/6 and Cyc E–CDK2, are involved in phosphorylation of retinoblastoma gene (Rb) and a subsequent G1/S transition. During this process, two main families of CDKIs play an important role in regulating G1/S transition: inhibitor of CDK (INK) and CDK-interacting protein/cyclin-dependent kinase inhibitory protein (CIP/KIP). The INK protein, p16, is a

Regulation of Corneal Endothelial Cell Proliferation

a

b c

Figure 1 Ultrastructure of a human cornea: (a) epithelium, (b) stroma, and (c) endothelium.

291

competitive inhibitor of CDK4/6-Cyc D complex, while the CIP/KIP proteins, p21 and p27, serve as competitive inhibitors of CDK2–Cyc E complex. The G1/S phase transition features two consecutive steps: (1) Early G1 progression involves mitogen-dependent accumulation of cyclin D, sequestration of p21, and cyclin D-dependent phosphorylation of retinoblastoma gene (Rb) (2) Late G1 progression depends on phosphorylation and degradation of p27 and additional phosphorylation of Rb by Cyc E/CDK2. Full phosphorylation of Rb leads to its dissociation from E2F and ultimately activation of E2F, which is a main transcription factor involved in the G1/S transition (Figure 3).

G1 Phase Cell Cycle Arrest in CECs in vivo Human corneal endothelial cells (HCECs) do not divide in vivo sufficiently to replenish lost cells due to aging, trauma, or disease. HCECs exhibit no positive staining with Ki67, a marker of actively cycling cells. However, HCECs exhibit in situ staining with the key cycle cell regulators, such as cyclins D, E, and A. Such staining pattern which is similar to the one seen in limbal epithelium known to contain slow-cycling stem cells, points to the fact that CECs are arrested in the G1 phase rather than having permanently exited from the cell cycle. Possible mechanisms accounting for such G1 phase cell cycle arrest are contact inhibition, interaction with the extracellular matrix (ECM), and presence of transforming growth factor (TGF)-b2 in the aqueous humor (Figure 4).

(a)

(b) Figure 2 Physiological loss of human corneas endothelial cells: (a) in a newborn (cell density (CD) = 4347 cells mm–2) and (b) in a 60-year-old normal eye (CD = 2392 cells mm–2).

Cell–Cell Contact Inhibition Contact-dependent inhibition of cell proliferation is a well described phenomenon. In corneal endothelium, majority of studies on cell–cell contact inhibition have been performed in neonatal rats, since rats have an immature corneal endothelium at birth, unlike the other species. In neonatal rats, the number of CECs staining positively for bromo-deoxyuridine (BrdU), an S-phase marker, gradually decreased between postnatal days 1 and 13. After postnatal day 13, positive BrdU staining was no longer detectable. Stable cell–cell and cell–substrate contacts gradually formed, and monolayer maturation was complete between postnatal days 14 and 21. These studies showed a correlation between decreased proliferation and increased monolayer formation, underlying the importance of cell–cell contact inhibition in the maturation of endothelial monolayer. Consistent with these findings, treatment of corneal endothelial monolayer with ethylenediaminetetraacetate (EDTA), a calcium chelator which releases cells from

292

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Mitogen

Cyc D Cyc E

p27 p21

CDK2 Cyc D p16

CDK 4/6

Genes involved in G1/S transition

Rb E2F

P P P Rb

Genes involved in G1/S transition

E2F

Figure 3 Cell cycle regulators involved in G1/S transition.

G1 Phase Arrest of Corneal Endothelial Cells

Cell−cell contact-dependent inhibition

Cell adhesion molecules in CEC = cadherin, connexin 43, ZO-1

Presence of transforming growth factor-β2 (TGF- β2) in aqueous humor TGF-β1

Interaction of CEC with matrix (descemet’s membrane)

TGF-β2

TGF-β mediated transition

• Releasing cell−cell contact by EDTA • Knocking down connexin 43

p27

p27

CyclinD/CDK 4 complex

Cyclin E/CDK 2 complex

Inhibition of CDK2

Fibroblast growth factor-2 (FGF-2)

Collagen type VIII

?

PROLIFERATION Figure 4 Possible mechanisms involved in G1 phase cell cycle arrest of corneal endothelial cells.

cell–cell contact, has been shown to promote cell proliferation. CEC proliferation can be regulated by calcium levels due to the presence of several cell adhesion proteins that are calcium sensitive and maintain cells in the state of contact inhibition. The main cell adhesion molecules involved in CEC physiology are cadherins, zonula occludens, and connexins (Cx). Exposure of CECs to calciumfree medium resulted in loosening of apical junctions, loss of barrier function, and corneal edema. This effect was reversible by exposing the endothelial cells to calcium. The attempts to break these intercellular interactions have been successful in promoting endothelial cell proliferation, and providing promise to regenerating CEC in

traumatic and degenerative situations. These findings indicate that cell–cell contact inhibition may play an important role in the growth arrest status of CECs. It was demonstrated that knocking down one of the major connexins, Cx43 resulted in a significant increase in the number of actively proliferating CECs. Using the rat model, corneal endothelial scrape injuries were simultaneously applied with Cx43 antisense oligodeoxynucleotide, small interfering RNA, or adenovirus (CMV– Cx43–mRFP1) into the anterior chamber. Changes in Cx43 expression were analyzed by immunolabeling (ZO-1, alpha-smooth muscle actin (SMA), Cx43). While, the endothelial–mesenchymal transition/transformation

Regulation of Corneal Endothelial Cell Proliferation

after injury was inhibited, Cx43 knock-down induced proliferation of the corneal endothelium. DNA-binding transcription factors, including the members of the BTB/POZ-zinc finger protein family, might be involved in the signaling pathway of cell–cell contactinduced growth inhibition as well. Among the family members, the promyelocytic leukemia zinc finger protein (PLZF) gene specifically inhibits the transcription of genes involved in G1/S transition, such as cyclin A2 and c-myc. The expression of PLZF is closely related with cell–cell contact. Its expression is high in confluent cells, and treatment with EDTA to disrupt cell–cell contact decreases PLZF messenger RNA (mRNA) levels. Similarly, overexpression of PLZF has been found to inhibit cell proliferation. N-Cadherin is thought to be a potential upstream signaling molecule that regulates PLZF mRNA levels. Studies showed that p27, a CDKI-type protein, is also involved in the mediation of cell cycle arrest induced by cell–cell contact. Increased expression of p27 in the developing corneal endothelium is well correlated with the cessation of the proliferation in wild-type mice. On the other hand, p27 knock-out mice show prolonged postnatal period of CEC proliferation when compared with their wild-type counterparts. These data indicate that p27 plays an important role in contact-inhibition-mediated growth arrested in CEC. The Presence of Antiproliferative TGF-b2 in the Aqueous Humor In many cell types, TGF-b2 inhibits proliferation by inducing G1-phase arrest. CECs are in direct contact with aqueous humor containing TGF-b2. Both exogenous TGF-b2 and active TGF-b2 in rat aqueous humor inhibit S-phase entry in rat CECs. The effect of TGF-b2 may be mediated by CDK4 and p27. In CECs treated with TGF-b2, CDK4 synthesis is inhibited and p27 is mobilized from the cyclin D-CDK 4 complex into the cyclin E-CDK 2 complex to inhibit CDK 2 activity. Furthermore, TGF-b2 prevents phosphorylation of p27 and maintains p27 in an active form.

The Interaction Between CECs and Extracellular Matrix ECM provides an important microenvironment for cell adhesion, migration, growth, differentiation, and signal transduction. DM is a specialized form of ECM separating CECs from corneal stroma. One difference between infant and adult human corneal DM is that collagen VIII, the major component of DM, shifted from the endothelial face of DM in infant to the stromal side of the DM in the adult. It is not fully understood whether compositional differences of collagen VIII are responsible for the differences in CECs

293

growth potential from young and adult humans. In a mouse model, the complete lack of type VIII collagen leads to dysgenesis (abnormal development) of the anterior ocular segment and enlarged CECs with reduced cell density. Hence, collagen VIII may serve as an important component of matrix that stimulates CEC growth during embryonic development. It is possible that adult DM has structural differences that foster endothelial growth arrest, associated with adult corneas. Proteomic analysis revealed that there is an age-related increase in TGFb-induced protein content and proteolytic processing of the human corneal endothelium and DM complex.

What Do We Know From Primary and Subcultured CECs Human and other mammalian CECs can be isolated and subcultured in vitro, indicating that endothelium still retains some replicative capacity. Such studies may provide better understanding of the regulatory mechanisms involved in CECs proliferation. Age-related decrease in sensitivity to mitogen or growth factors has been observed in cultured HCECs. It is still controversial whether the corneal periphery has more replicative capacity compared to central region. E2F is a key transcription factor involved in the G1/S transition of CECs, and overexpression of one of its isoforms E2F2 has been shown to promote proliferation. Various upstream signals including protein kinase C (PKC) and fibroblast growth factor (FGF)-2 might also play an important role in the proliferation of cultured CECs. Age-Related Decrease in Sensitivity to Mitogen or Growth Factors in Cultured HCECs Donor age negatively affects the proliferative capacity of cultured HCEC. In cultured HCEC a decrease in proliferative capacity is accompanied by changes in morphology and cell density. CECs cultured from older donors show an increase in cell size and a decrease in overall cell density. Different expression levels of cell cycle regulators were studied in cultured HCEC from younger and older donors. Increased expression of CDKIs such as p21 and p16 were seen in HCEC from older donors when compared with those from younger donors. On the other hand, transfection of p27 siRNA was sufficient to promote proliferation in confluent cultures of HCECs from younger (<30 years old), but not from older donors (>60 years old). This suggests that inhibition of proliferation in older donors is regulated by other mechanisms in addition to p27. It is known that an age-dependent increase in expression of negative cell cycle regulators p21 and p16 might be among those mechanisms.

294

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Protein tyrosine phosphatase (PTP) plays a negative role in regulating epidermal growth factor (EGF) signaling pathway. PTP1B is a widely expressed nonreceptor PTP originally identified in placentas. It downregulates the EGF signaling pathway by dephosphorylating epidermal growth factor receptor (EGFR) and inhibition of PTB1B promotes S phase entry. Increased PTP1B activity was detected in HCECs from older donors and reduced proliferative activity in response to EGF in those cells is partly due to increased PTP1B activity. Comparison of Proliferative Capacity of HCECs from Central and Peripheral Regions It is still controversial whether the peripheral endothelium has more potential for cell division than the central endothelium. p53 is a negative cell cycle regulator; it inhibits cell division primarily through a p21 pathway. It was found that p53 and its family member TAp63 are highly expressed in central rather than in peripheral endothelium, supporting that there is a greater potential for cell proliferation in the peripheral. However, studies by other group have shown that HCECs cultured from both the central and peripheral areas are capable of cell division in response to serum and their proliferation rates are same. Overexpression of E2F2 Promotes Proliferation of CEC E2F family proteins are key transcription factors for genes involved in the cell cycle progression. Three isoforms of this family (E2F1, E2F2, and E2F3) play an important role in G1/S transition. Overexpression of E2F2 promotes cell proliferation in rabbit CECs by increasing the proliferation marker ki67 and cylin B (G2 phase cell cycle regulator). PKC Signaling Pathways in Regulating Proliferation of CECs PKC comprises a family of serine/threonine protein kinases, which play an important role in regulating proliferation in many cell types. Several PKC isoforms, including PKC-alpha, -betaII, -delta, -epsilon, -iota, -eta, -gamma, and –theta, were detected in CECs. PKC activity, in particular PKC-alpha and -epsilon activity, is important in promoting CEC proliferation. Inhibition of PKC activity prohibits G1/S-phase progression and reduces cyclin E protein levels in cultured rat CECs. FGF-2 Signaling Pathway FGF-2 is a component of DM. As a member of the FGF family, it is a multifunctional regulator of cell development,

differentiation, regeneration, senescence, proliferation, and migration. The biological actions of FGF-2 are mediated through transmembrane cell surface receptors that possess tyrosine kinase activity. There are four isoforms of FGF-2. Only the 24-kDa nuclear FGF-2 isoform induced by corneal endothelium modulation factor (CEMF) may be involved in cell proliferation. Phospholipase C gamma (PLC-gamma) or phosphoinositide 3-kinases (PI3 kinase) serve as a downstream signaling pathway in FGF-2mediated proliferation of rabbit CEC proliferation. Both PLC-gamma and PI3 kinase may utilize Cdk4 and p27 while exerting the mitogenic signal.

Summary Understanding of the regulatory mechanisms involved in CEC proliferation is important in the context of regenerative medicine. Since corneal endothelium does not divide in vivo, manipulation of the factors involved in this cell cycle can be used to expand endothelium for regenerative purposes. Such developments would bring a great promise to the development of the treatment strategies for both exogenous and endogenous corneal endotheliopathies, hopefully bypassing the need for allogeneic corneal transplantation. See also: Corneal Endothelium: Overview; Regulation of Corneal Endothelial Function.

Further Reading Bednarz, J., Teifel, M., Friedl, P., and Engelmann, K. (2000). Immortalization of human corneal endothelial cells using electroporation protocol optimized for human corneal endothelial and human retinal pigment epithelial cells. Acta Ohthalmologica Scandinavica 78: 130–136. Coqueret, O. (2002). Linking cyclins to transcriptional control. Gene 299: 35–55. Enomoto, K., Mimura, T., Harris, D. L., and Joyce, N. C. (2006). Age differences in cyclin-dependent kinase inhibitor expression and Rb hyperphosphorylation in human corneal endothelial cells. Investigative Ophthalmology and Visual Science 47: 4330–4340. Harris, D. L. and Joyce, N. C. (2007). Protein tyrosine phosphatase, PTP1B, expression and activity in rat corneal endothelial cells. Molecular Vision 13: 785–796. He, Y., Weng, J., Li, Q., Knauf, H. P., and Wilson, S. E. (1997). Fuchs’ corneal endothelial cells transduced with the human papilloma virus E6/E7 oncogenes. Experimental Eye Research 65: 135–142. Hopfer, U., Fukai, N., Hopfer, H., et al. (2005). Targeted disruption of Col8a1 and Col8a2 genes in mice leads to anterior segment abnormalities in the eye. FASEB Journal 19: 1232–1244. Joko, T., Nanba, D., Shiba, F., et al. (2007). Effects of promyelocytic leukemia zinc finger protein on the proliferation of cultured human corneal endothelial cells. Molecular Vision 13: 649–658. Joyce, N. C. (2003). Proliferative capacity of the corneal endothelium. Progress in Retinal and Eye Research 22: 359–389. Joyce, N. C., Harris, D. L., and Zieske, J. D. (1998). Mitotic inhibition of corneal endothelium in neonatal rats. Investigative Ophthalmology and Visual Science 39: 2572–2583.

Regulation of Corneal Endothelial Cell Proliferation Jurkunas, U. V., Bitar, M. S., and Rawe, I. M. (2009). Co-localization of increased transforming growth factor beta induced protein (TGFBIp) and clusterin expression in guttae of Fuchs endothelial corneal dystrophy patients. Investigative Ophthalmology and Visual Science 50(3): 1129–1136. Kabosova, A., Azar, D. T., Bannikov, G. A., et al. (2006). p27kip1 siRNA induces proliferation in corneal endothelial cells from young but not older donors. Investigative Ophthalmology and Visual Science 47: 4803–4809. Konomi, K., Zhu, C., Harris, D., and Joyce, N. C. (2005). Comparison of the proliferative capacity of human corneal endothelial cells from the central and peripheral areas. Investigative Ophthalmology and Visual Science 46: 4086–4091. McAlister, J. C., Joyce, N. C., Harris, D. L., Ali, R. R., and Larkin, D. F. (2005). Induction of replication in human corneal endothelial cells by E2F2 transcription factor cDNA transfer. Investigative Ophthalmology and Visual Science 46: 3597–3603.

295

Nakano, Y., Oyamada, M., Dai, P., et al. (2008). Connexin43 knockdown accelerates wound healing but inhibits mesenchymal transition after corneal endothelial injury in vivo. Investigative Ophthalmology and Visual Science 49: 93–104. Paull, A. C. and Whikehart, D. R. (2005). Expression of the p53 family of proteins in central and peripheral human corneal endothelial cells. Molecular Vision 11: 328–334. Sasaki, T., Sorokin, L. M., Steiner-Champliaud, M. F., et al. (2007). Compositional differences between infant and adult human corneal basement membranes. Investigative Ophthalmology and Visual Science 48: 4989–4999. Yoshida, K., Kase, S., Nakayama, K., et al. (2004). Involvement of p27KIP1 in the proliferation of the developing corneal endothelium. Investigative Ophthalmology and Visual Science 45: 2163–2167. Zhu, C. and Joyce, N. C. (2004). Proliferative response of corneal endothelial cells from young and older donors. Investigative Ophthalmology and Visual Science 45: 1743–1751.

Artificial Cornea M A Rafat, University of Ottawa Eye Institute, Ottawa, ON, Canada J M Hackett, University of Ottawa, Ottawa, ON, Canada P Fagerholm, Linko¨ping University Hospital, Linko¨ping, Sweden M Griffith, University of Ottawa Eye Institute, Ottawa, ON, Canada ã 2010 Elsevier Ltd. All rights reserved.

Glossary Bioengineered corneas – Are natural-based substitutes for human donor tissue that are designed to replace part or the full thickness of damaged or diseased corneas. Collagen – The most common naturally occurring structural protein found in all multi-cellular animals and accounts for approximately 30% of all body proteins. Interpenetrating polymeric networks (IPNs) – Polymeric structures comprising two or more networks that are interconnected on a molecular scale through chemical (covalent) and physical bonds. Keratoprosthesis – A type of artificial cornea that is designed to be implanted in a patient who has severe bilateral corneal disease for which a corneal transplant is not an option. Lenticles – A tiny disk slipped into the pocket of the patient’s own cornea between corneal epithelium and Bowman’s membrane for vision correction. Penetrating keratoplasty – A surgical procedure where a damaged or diseased cornea is replaced by donated corneal tissue which has been removed from a recently deceased individual having no known diseases which might affect the viability of the donated tissue. Photorefractive keratectomy (PRK), laser-assisted sub-epithelial keratectomy (LASEK), or laser-assisted in situ keratomileusis (LASIK) – Laser eye-surgery procedures for correcting a person’s vision can reduce the need for glasses or contact lenses.

The Need for Artificial Corneas The terms artificial corneas (ACs) or bioengineered corneas are widely used to describe corneal scaffolds that are designed to restore vision. These scaffolds are used as substitutes to human donor corneas, and can replace part or the full thickness of damaged or diseased corneas.

296

Corneal scaffolds can range from solely synthetic ocular prostheses through to tissue-engineered hydrogels that allow some regeneration of the host tissues. In addition, bioengineered lenticles may be implanted into the cornea to improve vision. This is achieved by altering the refractive properties of the eye, which is an alternative procedure to laser-assisted in-situ keratomileusis (LASIK), laser-assisted sub-epithelial keratectomy (LASEK), and photorefractive keratectomy (PRK). Compromising the transparency of the cornea interferes with its function. Once optical clarity is compromised, due to disease or damage, vision loss occurs and can result in corneal blindness (Figure 1). According to the World Health Organization (WHO), there are approximately 37 million people worldwide who possess bilateral blindness, as well as at least 124 million people who have impaired vision in both eyes. Corneal ulceration and ocular trauma are reported to be the major causes of corneal blindness, accounting for 1.5–2 million new cases annually. At present, transplantation of matched human donor tissue is the only widely acceptable treatment. Corneas are the most successful organ transplants – with an 86% graft-survival rate at the 1-year postoperative follow-up, 73% after a 5-year period, 62% after 10 years, and 55% at 15 years. The success of the transplantation is dependent upon the availability of goodquality donor tissue, as well as the patient’s condition. Inactive central scars or keratoconus are amenable to transplantation; however, alkali burns or neurotrophic scars that are secondary to Herpes zoster ophthalmicus have a poor prognosis. The cornea donor pool is in limited supply due to a longer life expectancy, combined with the aging population within North America. Demand for corneas is expected to increase, but a shortage in supply will most likely be experienced. The shortage is expected to be compounded by the increasing incidence of infectious diseases (HIV, hepatitis, Creutzfeldt–Jakob disease, etc.), as well as the growing popularity of refractive surgery. Surgically treated corneas are thinned, rendering them unacceptable as donor tissue. As alternatives to donor tissues, bioengineered corneas are designed to replace some or all of a damaged or diseased cornea. They range from prosthetic devices that solely address replacement of the cornea’s function, to tissue-engineered hydrogels that permit regeneration of host tissues. This article focuses on the efforts employed to build an in vitro model of the visual system.

Artificial Cornea

Childhood blindness 3.9%

Trachoma 3.6%

297

Corneal opacity 5.1%

Oncho cerciasis 0.8% Diabetic retinopathy 4.8% Cataract 47.8%

AMD 9%

Others 13.0%

Glaucoma 12.3%

Figure 1 Global causes of blindness as a percentage of total blindness in the year 2002. Trachoma and corneal opacity are cornea-related diseases comprising about 10% of the total causes of blindness. Adapted with permission from Fig. 1 in Resnikoff, S., Pascolini, D., Etya’ale, D., et al. (2004) Global data on visual impairment in the year 2002. Bulletin of the World Health Organization 82: 844–851, with permission from WHO Press.

In particular, we examine the development of novel biomaterials that serve as the building blocks for the fabrication of scaffolds in engineered tissues.

Desired Characteristics for an Implantable AC In order to be clinically applicable, fabricated corneal substitutes need to replicate the functions of the human cornea. The human cornea forms a transparent window through which light is transmitted to the retina, enabling one to see. A unique property of the cornea is its optical clarity, which accounts for over 70% of the light that is transmitted to the retina for vision. Corneal clarity is now believed to result from a combination of refractive-index matching, and the presence of structural components that are well below the wavelength of visible light. As alternatives to donor tissues, ACs are designed to replace some or all of a damaged or diseased cornea. They range from prosthetic devices that solely address replacement of the cornea’s function, to tissue-engineered hydrogels that permit regeneration of host tissues. In instances where corneal stem cells have been depleted by injury or disease, tissue-engineered lamellar implants reconstructed with stem cells have been transplanted. In situ methods using ultraviolet A (UVA) cross-linking have also been developed to strengthen weakened corneas. In addition to the clinical need, bioengineered corneas are also rapidly gaining importance in the area of in vitro toxicology. In Europe, there is

currently a ban on consumer product testing in animals (European Union Directive 76/768/EEC) that is expected to expand worldwide. Complex, fully innervated, physiologically active, three-dimensional (3D) organotypic corneal models are currently being developed and tested. ACs must meet certain requirements, without exception, to be successful. To properly integrate into host tissue, AC must be biocompatible and noncytotoxic. A watertight junction with the host tissue is essential for preventing infection and epithelial down-growth. Epithelial cell growth must be supported over the anterior surface, allowing a wettable, self-renewing layer that promotes a healthy tear-film formation. Nerve innervations must be supported for high touch sensitivity. Penetration and proliferation of host fibroblast cells must be promoted for tissue regeneration. Optical transparency >80% and light scatter of <5% should be exhibited, as well as a suitable morphology and curvature to obtain the appropriate refractive index. In addition, flexibility and sufficient tensile strength is required to allow surgical manipulation and fixation, as well as to protect the eye. The AC should exhibit sufficient swelling in aqueous solutions similar to that of native cornea, but at the same time be permeable to oxygen, nutrients such as glucose, and serum albumin – which is a major water-soluble protein in the human cornea. Lastly, the AC must be inexpensive and easy to fabricate. Artificial or bioengineered corneas developed to date range from prostheses – known as keratoprosthesis (KPro) – to naturally fabricated cell-based tissue equivalents, to bioengineered scaffolds that serve as templates for regeneration of host tissues.

298

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Synthetic Artificial Corneas or Keratoprosthesis Research into the development of an AC has existed for more than 200 years, with the original glass and quartz optics being put forward, in 1789, by Guillaume Pellier de Quengsy. Since then, there have been numerous attempts at developing ACs. Currently, four generations of KPros have been defined, according to the KPro Study group. KPros are synthetic implants designed to replace the central portion of an opaque cornea. First-generation KPros are comprised of monoblocks, or one-piece prostheses that are made from plastics such as poly methyl methacrylate (PMMA). An example of a second-generation KPros is the osteo-odonto keratoprosthesis (OOKP). This prosthesis, developed by Strampelli in 1964, consists of autologous tissue derived from tooth and bone, which surrounds a central PMMA optic. An osteodental skirt is preimplanted into the buccal mucosa, allowing colonization of fibroblasts to support its integration as an ocular implant. Third-generation KPros include a range of devices with plastic optics and metal parts that aid in anchoring the device to host tissues. Donor-tissue attachment can also be achieved through skirts, allowing for host integration. Fourth-generation KPros utilize the optic-skirt model, in which a solid optical core is surrounded by a porous skirt. This encourages biointegration with the adjacent host tissues to circumvent implant extrusion. Rigid synthetic polymers, such as PMMA, were used in the early attempts to develop artificial corneal transplants. While PMMA still remains a popular material, poly(2-hydroxyethyl methacrylate) (PHEMA) has been used more frequently in various types of KPros. In this article, we provide a synopsis of several examples of KPros that have either been tested clinically, or are currently in clinical use. KPros currently in clinical use include the following: the OCULAIDW KPro, Dohlman KPro, AlphaCor™ Kpro, OOKP KPro, BioKPro III, Seoul-type KPro, and Pintucci KPro. Figure 2(a) and 2(b) represents the OCULAIDW KPro, composed of an anti-conical shaped shaft that can be fixed into the host cornea or sclera. It creates a valve on the cornea to ensure a watertight environment. The pressure in the eye pushes the corneal rim around the 3-mm top of the KPro, while the steel-suture fixation on the sclera is designed to prevent extrusion. The Dohlman AC is a collarbutton-design KPro, composed of PMMA. It consists of a central optical stem that penetrates the full thickness of the cornea. This stem is sandwiched between two plates and sutured into place similar to a penetrating keratoplasty (PKP) graft (Figure 2(c)). The AlphaCor™ Kpro is one of the best-known keratoprosthetic devices fabricated as a one-piece device that comprises a transparent core and an opaque porous skirt. The implant is a 7-mm-diameter, one-piece, nonrigid synthetic cornea (Figure 2(d)). Composed of a transparent

central optic core of PHEMA gel, the KPro is designed to allow the passage of light into the posterior of the eye. The outer porous skirt – an opaque, high-water PHEMA – is designed to allow cell infiltration from the host, which anchors the prosthesis into place. This device gained the Food and Drug Administration (FDA)-approval in 2002 for use in patients with scarred, vascularized, or diseased corneal tissues who are either ineligible for conventional donor-tissue transplants or have had multiple previous graft failures. Early results suggest that the AlphaCor™ Kpro, previously known as the Chirila KPro, is clinically safe. Associated complications include the formation of retroprosthetic membranes, corneal melt, retained lenticular material, optic depositions, and rare cases of device extrusion. Contraindications include abnormal tear film, as well as uncontrollable high intraocular pressure. Topical administration of medroxyprogesterone has been shown to limit corneal melting of the device. However, this KPro has been effectively used to restore a degree of vision in patients considered untreatable by conventional corneal transplantation. The OOKP, described above, consists of autologous tissue derived from tooth, which surrounds a central PMMA optic. This KPro is one of the most successful, as it has a low extrusion rate due to the excellent integration of the skirt material with the host tissue. Associated complications with the OOKP include retroprosthetic membrane formation, glaucoma and decentration of the central optic, due to absorption of the osteodental skirt. The BioKPro III, designed by the Legeais group, was recently clinically evaluated at the Moorfields Eye Hospital, London, UK. The BioKpro III consists of a central 5-mm-diameter, 500-mm-thick silicone optic, and a surrounding opaque skirt made of porous fluorocarbon. The device was implanted into seven patients with severe corneal scarring, and monitoring occurred between 18 and 48 months. The results showed that the KPro failed in six patients, due to extrusion occurring between 2 and 28 months postoperatively. One patient, who had a thermal burn, retained the KPro and reported an improvement in vision. However, this patient also reported mucous accumulation on the optic. The Seoul-type KPro (S-KPro) consists of three sections: a long cylindrical optic surrounded by a mushroomshaped anterior flange, a skirt for corneal fixation, and haptics for scleral fixation. The 4-mm diameter optic is made of PMMA; the anterior flange is composed of fluorinated silicone approximately 0.2-mm thick and 6 mm in diameter; while the skirt is fabricated from expanded polytetrafluoroethylene (e-PTFE). Preliminary results from the first human trial indicate that complications including retinal detachment, formation of a retroprosthetic membrane, and extrusion have been identified. The Pintucci KPro consists of a 3-mm thick and 5-mm long optical cylinder made of PMMA. A woven, 0.7-mm

Artificial Cornea

299

6 mm 3 mm

Steel suture fixation

(a)

(b)

(c)

(d)

Figure 2 A glance at some of the keratoprostheses used in human clinical trials: (a) and (b) OCULAIDW KPro, (c) Dohlman or Boston KPro, and (d) AlphaCor™ Kpro. (a) and (b) Adapted with permission from Dr. Jan Worst Research Group, the Netherlands. (c) Adapted with permission from Dr. Esen Akpek of the Wilmer Eye Institute at Johns Hopkins.

thick and 10-mm circular Dacron membrane is added to the KPro for tissue integration. Like the OOKP, this device is preimplanted into the patient for colonization of the skirt. Thirty-one patients have received implants from 1997 to 2004, by means of clinical trials. Results indicated that no infections or retroprosthetic membranes were reported. Seventy-seven percent of implanted eyes improved enough to enable the patients to function independently. However, approximately 40% of implanted eyes had complications, although only a few cases were vision threatening. Recent developments in the design of KPros include a KPro developed by Storsberg and colleagues in Germany; the Stanford KPro; and recent work by Sheardown and colleagues on the coverage of KPros by extracellular matrix (ECM) proteins for enhanced epithelialization. The German KPro adheres to the eye without sutures, reducing inflammation and infection. Sheardown and colleagues have covalently attached cell-adhesion peptides to poly(dimethyl siloxane) (PDMS) surfaces. This surface modification has been effective, leading to a synergistic effect on corneal epithelial cell attachment when compared to single peptides only. The Stanford KPro has also been designed using a polymer network hydrogel, comprised of poly(ethylene glycol) and poly(acrylic acid) (PEG/PAA). When implanted in rabbit corneas, this hydrogel was retained and tolerated well in nine out of 10 cases for a 2-week period.

To date, however, no KPro meets the previously defined standards for a successful corneal implant. As such, no particular KPro is in widespread use to date, although recent versions appear to be promising.

Self-Assembled Corneal Equivalents There have been various attempts at developing a naturalbased self-assembled corneal equivalent. These range from the use of purely biological materials synthesized by cells in culture, to the use of noncorneal tissues as substitutes. Most bioengineering approaches to the restoration and repair of damaged tissue require scaffold materials upon which cells can attach, proliferate, and differentiate. Such scaffolds can be made using a self-assembly approach in which chemicals are used to stimulate the secretion of collagen, and other ECM molecules by fibroblast cells. Resulting sheets of scaffolds are stacked together to form a stroma, allowed to further integrate in vitro, and then epithelial cells are seeded on top of the stack. These constructs mimic corneal morphology, and the cells express appropriate tissue-specific markers. The main drawback is the time needed to produce enough self-assembled scaffolding for transplantation. A cornea equivalent, composed of a 3D, bovine, dermal collagen matrix has been developed by Minami and

300

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

colleagues for in vitro studies. Zieske and colleagues also developed an in vitro cornea – fabricated using primary rabbit stromal cells. Funderburgh and colleagues have used keratocytes of the corneal stroma to produce a transparent ECM that may be useful in cell-based corneal therapy or for the development of bioengineered corneas. Funderburgh and colleagues also determined there is a population of cells present in adult mammalian corneal stroma having the ability to divide extensively, generating differentiated keratocytes. Previously, the authors reconstructed a human cornea using immortalized human corneal cell lines. Each cell line was subjected to electrophysiological, biochemical, and morphological tests. This was carried out to determine the phenotype, which was compared to postmortem human corneal cells, before being used in the 3D reconstruction. Collagen–chondroitin sulfate was the base scaffold in which keratocytes were integrated, before epithelial and endothelial cells were layered above or below. Two weeks following construction, the resulting corneal equivalent was found to behave similarly to a normal cornea, with respect to morphology, transparency, ion and fluid transport, and gene expression following injury. Although this human corneal equivalent shared functional properties with the natural cornea, it was not designed to meet the mechanical characteristics needed for transplantation. However, these studies represent an important future directive toward the development of bioengineered corneal implants. Reconstructed corneal equivalents presently have use in the biomedical world, as they are used as replacements for animals in toxicology testing and pharmacological studies.

Bioengineered ACs that Address Regeneration To overcome the challenges of biocompatibility, inflammatory responses, and rejection, there have been attempts to promote varying degrees of corneal tissue regeneration through implants of bioengineered corneal ECM substitutes. In general, these matrix-mimetic materials range from simple cross-linked ECM macromolecules, such as collagen, to hybrids of ECM macromolecules and synthetic polymeric components. Prior to the assembly of these ECM mimetics for implantation, a bioactive scaffold material with accurate chemical/physical properties must be designed. It must be able to form robust scaffolds, promote cell differentiation/integration, and promote tissue formation in a uniform manner that is repeatable and reliable. Polymeric blends of collagen have been previously used to emulate the collagen–glycosaminoglycans scaffolding of the ECM. Various tissue-engineering applications have utilized this technology, for example, as a scaffold for artificial liver, skin scaffolds with nerves and dermal models, membranes for controlled drug release, and as an in vitro model to test antineoplastic agents. Many

efforts have been made to stabilize collagen, and its blends, by chemical cross-linking methods. These methods can be divided into two categories: bifunctional and amide-type. Several bi-functional reagents such as glutaraldehyde (GTA), polyethylene glycol diacrylate (PEGDA), and hexamethylene diisocyanate (HDC) have been used to bridge amine groups of lysine or hydroxylysine residues of collagen polypeptide chains. A major handicap of these cross-linking agents is the potential toxic effect of residual molecules and/or compounds released when the biomaterial is exposed to biological environments (i.e., during in vivo degradation). Amide-type cross-linkers such as carbodiimide, especially 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC), and N-hydroxysuccinimide (NHS) offer the main advantage of lower toxicity and better compatibility over GTA and HDC. However, collagen scaffolds stabilized by carbodiimide are not strong, yet are elastic enough for PKP transplantation. This is due to limited zero-length crosslinks; in addition, there are reaction sites on collagen molecules that are not linkable by carbodiimide. It is reported that EDC and NHS (EDC/NHS) can link carboxylic acid and amino groups located within 1 nm from each other. Therefore, functional groups that are located on adjacent collagen microfibrils are too far apart to be bridged by carbodiimide. With systems such as EDC/NHS, the increase in tensile strength, especially when induced by the increasing of a cross-linking agent, is associated with the decrease in elasticity and toughness. This compromise in integrity may be due to restraints placed on the mobility of the polymer network, a decrease in scaffold porosity, and/or a diffusion of reactive residues and byproducts out of the scaffolds. Despite the drawbacks, cross-linking agents have been used effectively to fabricate collagen based matrices. The application of these matrices has been mainly diagnostic; where the tissue is made and used in vitro for testing drug metabolism, uptake, and toxicity. The authors have previously reported collagen-based materials, ranging from corneal scaffolds based on the copolymer poly(N-isopropylacrylamide-co-acrylicacid-coacryloxylsuccinimide), to a simple EDC/NHS cross-linked collagen scaffold. These scaffolds allow regeneration of corneal cells and nerves, when implanted as lamellar grafts. However, these materials still lack the optimal toughness and elasticity required to withstand PKP surgical procedures, as well as normal day-to-day mechanical stresses. In our most recent ongoing clinical study, we reported corneal regeneration following the implantation of a bioengineered corneal substitute, based on recombinant collagen. Visual acuity, ocular surface quality, and corneal sensitivity are continuously improving in the first recipients of the implants. Figure 3 shows slit-lamp and optical coherence tomography (OCT) photographs of the operated cornea immediately following lamellar keratoplasty and at 9 months post operation. Such substitutes

Artificial Cornea

(a)

301

(b)

Bioengineered corneal implent

Host cornea

(c)

Figure 3 Operated cornea of a 37-year-old man following implantation of a bioengineered artificial cornea, which is transparent with a smooth corneal surface and is well accepted by the patient’s eye: (a) slit-lamp photograph of the implanted cornea right after lamellar keratoplasty (LKP) anchored with three overlying 10-0 nylon sutures. (b) slit-lamp photograph of the cornea 9 months following the LKP, (c) OCT (optical coherence tomography) image (ASOCT, Visante, Carl Zeiss Meditec, Jena, Germany) 9 months following the LKP.

may find use as temporary or emergency corneal replacements, where human tissue is unavailable. In several attempts to enhance mechanical properties of corneal substitutes, hybrid interpenetrating polymeric networks (IPNs) have been developed as ACs. Synthetic-based IPNs have been widely explored and used; however, bioengineered IPNs have not been investigated and examined for corneal applications until recently. Efforts were made by the authors to develop a bioengineered IPN as an AC. The goal was to develop an IPN scaffold that combines the bioactive features of biopolymers with the physical characteristics of synthetic polymers. Composite IPN structures – comprised of two or more interconnected networks on a molecular scale through chemical (covalent) and physical bonds – were developed. The scaffold material used bio-functional polymers that naturally occur in the native tissue (e.g., collagen) as the core material. As an alternative, tissue-mimetic polymers such as chitosan, 2-methacryloyloxy ethyl phosphorylocholine (MPC) or chondroitin sulfate were used as bio-interactive components; with poly (ethylene glycol) dibutyraldehyde (PEG-DBA) or poly(ethylene glycol) diacrylate (PEG-DA) being used as a synthetic long-range cross-linker. Bio-inert short-range cross-linkers, including EDC and NHS, were also used in conjunction with long-range cross-linkers. This composition allowed the formation of hybrid IPNs that are mimetic of the natural cornea. The IPN scaffolds demonstrated significantly enhanced mechanical strength and elasticity compared to

(a)

(b)

12 mm 0.5 mm ~3 mm

(c) Figure 4 Corneal implants: (a) bioengineered artificial cornea (IPN), (b) eye-bank human donor cornea, and (c) dimensions of a typical bioengineered artificial cornea.

their non-IPN counterpart. In addition, they demonstrated excellent optical properties, optimum mechanical properties and suturability, and good diffusivity to glucose and albumin. The IPNs had excellent biocompatibility, and were further tested by being implanted into pig corneas. Over the 12-month monitoring period, it was demonstrated that there was seamless host–graft integration with successful regeneration of host corneal epithelium, stroma, and nerves. Figure 4 depicts an IPN-bioengineered AC compared to an eye-bank human donor cornea.

302

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Future Directions In the cornea, regeneration of the host cornea could overcome the rejection problems and other postoperative complications from donor-tissue transplantation and KPros. In addition, corneal implants that allow nerve- and host-tissue regeneration could also circumvent problems after surgery, found both in human donor tissue and in synthetic KPros. At the current pace of product development and testing, viable alternatives to donor corneas for transplantation are not far off. Processing procedures can affect both the native material properties and the subsequent clinical utility of scaffolds intended for certain tissue-engineering applications. Irrespective of the method, the AC must be engineered to mimic morphological, physiological, and biochemical properties of the natural tissue as closely as possible. See also: Corneal Epithelium: Cell Biology and Basic Science; The Corneal Stroma; Cornea Overview; Penetrating Keratoplasty; Refractive Surgery and Inlays.

Further Reading Chirila, T. V., Hicks, C. R., Dalton, P. D., et al. (1998). Artificial cornea. Progress in Polymer Science 23: 447–473. Crawford, G. J., Hicks, C. R., Lou, X., et al. (2002). The Chirila keratoprosthesis: Phase I human clinical trial. Ophthalmology 109: 883–889. Dohlman, C. H., Harissi-Dagher, M., Khan, B. F., et al. (2006). Introduction to the use of the Boston keratoprosthesis. Expert Review of Ophthalmology 1(1): 41–48. Duan, D., Klenkler, B. J., and Sheardown, H. (2006). Progress in the development of a corneal replacement: Keratoprostheses and tissue-engineered corneas. Expert Review of Medical Devices 3: 59–72. Funderburgh, M. L., Du, Y., Mann, M. M., Raj, N. S., and Funderburgh, J. L. (2005). PAX6 expression identifies progenitor cells for corneal keratocytes. FASEB Journal 19(10): 1371–1373.

Germain, L., Carrier, P., Auger, F. A., Salesse, C., and Gue´rin, S. L. (2000). Can we produce a human corneal equivalent by tissue engineering? Progress in Retinal and Eye Research 19: 497–527. Griffith, M., Hakim, M., Shimmura, S., et al. (2002). Artificial human corneas: Scaffolds for transplantation and host regeneration. Cornea 21(7): S54–S61. Griffith, M., Osborne, R., Munger, R., et al. (1999). Functional human corneal equivalents constructed from cell lines. Science 286: 2169–2172. Hicks, C., Crawford, G., Chirila, T., et al. (2000). Development and clinical assessment of an artificial cornea. Progress in Retinal and Eye Research 19: 149–170. Kalayoglu, M. V. (2006). In search of the artificial cornea: Recent developments in keratoprostheses. Ophthalmology Technology Spotlight Medcompare. Leibowitz, H. M., Trinkhaus-Randall, V., Tsuk, A. G., and Franzbau, C. (1994). Progress in the development of a synthetic cornea. Progress in Retinal and Eye Research 13: 605–621. Minami, Y., Sugihara, H., and Oono, S. (1993). Reconstruction of cornea in three-dimensional collagen gel matrix. Investigative Ophthalmology and Visual Science 34: 2316–2324. Rafat, M., Li, F., Fagerholm, P., et al. (2008). PEG-stabilized carbodiimide crosslinked collagen–chitosan hydrogels for corneal tissue engineering. Biomaterials 29: 3960–3972. Resnikoff, S., Pascolini, D., Etya’ale, D., et al. (2004). Global data on visual impairment in the year 2002. Bulletin of the World Health Organization 82: 844–851. The National Coalition for Vision Health (2005). There’s still a critical shortage of corneas for transplantation 2005. http://www. visionhealth.ca/news/insert/shortage.htm (accessed July 2009). WHO (1997). Global initiative for the elimination of avoidable blindness. Geneva: World Health Organization. (unpublished document WHO/PBL/97.61/Rev 1). http://whqlibdoc.who.int/hq/1997/ WHO_PBL_97.61_Rev.1.pdf. WHO (2003). Human organ and tissue transplantation. World Health Organization, May 2003 (EB112/5 112th Session). http://www.who. int/ethics/topics/human_transplant/en (accessed July 2009). Worst, J. (2007). The ‘‘Champagne Cork’’ keratoprosthesis (KP). http:// www.janworst.com/projects/kp/frames/framekp1.htm. Yaghouti, F., Nouri, M., Abad, J. C., and Power, W. J. (2001). Keratoprosthesis: Preoperative prognostic categories. Cornea 20: 19–23.

Relevant Website http://www.asmr.org.au – The Australian Society for Medical Research, Following in the footsteps of Fred Hollow, Key Statistics 2006.

Drug Delivery to Cornea and Conjunctiva: Esterase- and Protease-Directed Prodrug Design R S Talluri, S Hariharan, P K Karla, and A K Mitra, University of Missouri–Kansas City, Kansas City, MO, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Amidases – The enzymes that catalyze the cleavage of carbon–nitrogen bonds in amides. Bioreversion – The conversion of a prodrug to an active form. Esterases – The enzymes that catalyze the hydrolysis of an ester into its alcohol and acid. Peptidases – The enzymes that catalyze the hydrolysis of peptides into amino acids. Prodrug – The drugs designed to be inactive until in vivo activation generates the active form of the drug. Transporter – The protein that translocates materials in biological systems resulting from expenditure of metabolic energy.

Topical Ocular Drug Delivery Topical drug delivery is the most acceptable form of treatment for the diseases affecting the anterior segment such as corneal epithelial and stromal keratitis, glaucoma, conjunctivitis, dry eye syndrome, iritis, uveitis, and blepharitis. The structure of the eye is shown in Figure 1. Topically applied drugs can reach the intraocular tissues either by corneal and/or noncorneal (conjunctival–scleral) pathway(s). Drugs traversing the corneal pathway should permeate through the corneal epithelium and stroma which are rate-limiting barriers for hydrophilic and lipophilic molecules, respectively. Conjunctival absorption could result in higher drug concentrations in the anterior as well posterior chambers depending on the mechanism of absorption. Conjunctival–scleral pathway is favored for the treatment of diseases in the posterior segment, as it can bypass the anterior chamber and thus permit direct access to intraocular tissues like sclera retina and vitreous humor. However, nasolacrimal drainage, tear dilution, as well as the outer conjunctiva and cornea, act as barriers to drug absorption. As a result, therapeutic concentrations in intraocular tissues following topical administration are difficult to achieve and about 1–5% of topically instilled dose reaches the anterior segment of the eye. Moreover, a larger fraction of the applied drug is eliminated from the precorneal area within 5 min, through drainage across

nasolacrimal duct into the systemic circulation. Fifty percent of the normal human tear film is replaced every 2–20 min. Such a high tear-turnover rate also reduces the drug-residence time in precorneal and conjunctival areas. Therefore, rapid tear-film drainage can also impede the drug absorption following topical administration. Topically administered agents have a low probability of reaching the posterior segment in significant amounts, as passage through the corneal and conjunctival epithelia, aqueous humor, and lens is required to reach the retina. Various strategies have been investigated in order to improve the corneal and conjunctival absorption of drugs instilled topically. Prodrug approach is one of the most promising and effective strategies currently being investigated for ophthalmic drug delivery. Exploring the inherent drug metabolism capability of ocular tissues is one of the important aspects of prodrug design. In this strategy, the drug molecule is modified chemically by attaching it to a promoiety to improve the physicochemical characteristics, such that higher drug absorption into the tissues can be achieved. Prodrugs are designed to be therapeutically inactive until in vivo activation to generate the parent drug. The compounds are synthesized by linking an appropriate chemical moiety to the parent drug, usually linking by an ester or an amide bond. Upon absorption into the tissue, the prodrug will be subjected to enzymatic hydrolysis (bioreversion by esterases/peptidases) to release the active parent drug (Figure 2). The rate of bioreversion depends upon various factors, including affinity of the prodrug linkage toward hydrolyzing enzyme(s), the capacity and turnover rate of the enzyme, etc. The enzymes responsible for hydrolysis of prodrugs are present ubiquitously in all biological fluids and tissues. For example, esterases are expressed throughout the body and can be utilized in the hydrolysis of an ester functional group. In ocular tissues, the esterase activity has been found to be the highest in iris-ciliary body followed by the cornea and the aqueous humor. Among ocular esterases, butyrylcholinesterase (BuChE) constitutes the major proportion compared to acetylcholinesterase (AChE) – except in the corneal epithelium of albino rabbits. Drugs and prodrugs containing ester linkages can undergo varying extents of esterase-mediated hydrolysis while permeating the cornea/conjunctiva and upon entering into the aqueous humor, iris, and ciliary body. Proteases or peptidases are primarily responsible for hydrolysis of amide linkage in

303

304

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Palpebral conjunctiva

Lacrimal gland

Cor

Bulbar conjunctiva Aqueous humor (anterior chamber)

nea

Iris Fonix (forniceal conjuctiva)

s

Len

Pupil Posterior chamber

Vitreous humor

Choroid Hyaloid canal

Sclera

Fovea Retina

Retinal blood vessels Optic nerve

Figure 1 Cross-sectional view of the eye. From Hosoya, K., Lee, V. H., and Kim, K. J. (2005). Roles of the conjunctiva in ocular drug delivery: A review of conjunctival transport mechanisms and their regulation. European Journal of Pharmaceutics and Biopharmaceutics 60: 227–240.

Drug

Lipophilic prodrug

Transporter targeted prodrug

Drug

Drug

Ester/amide linkage

Alkyl chain (promoiety)

Poor permeability

Promoiety Nutrient transporter

Cell membrane

Increased permeability

Cell cytosol Drug

Drug

Esterase/ protease Hydrolysis

Drug

Prodrug

Esterase/ protease

Drug

Hydrolysis

Figure 2 Schematic: Lipophilic and transporter targeted prodrug design.

peptides or peptide-based prodrugs. These enzymes are classified as either ‘endo-’ or ‘exo-’ depending on whether they cleave internal or external peptide bonds. The aminopeptidase activity is the highest in the corneal epithelium

and iris-ciliary body followed by conjunctiva and corneal stroma. Kashi and colleagues have shown that aminopeptidases, dipeptidyl peptidase, and dipeptidyl carboxylpeptidase are

Drug Delivery to Cornea and Conjunctiva: Esterase- and Protease-Directed Prodrug Design

involved in hydrolysis of enkhaphilins in rabbit oculartissue homogenates including conjunctiva, corneal stroma, iris ciliary body, lens, and tears. Prodrugs targeted toward membrane transporters expressed on the epithelial cells are perhaps the most exciting of all the current drug-delivery strategies. Epithelial cells express various nutrient transporters and receptors on their membrane surface. Analogs or prodrugs targeted toward these transporters can significantly enhance the absorption of poorly permeating therapeutic agents. Such prodrugs are recognized by membrane transporters as natural substrates and are translocated across the epithelial membranes. Once inside the cell, the conjugate will release the parent drug by enzymatic hydrolysis (Figure 2). Various nutrient transporters are expressed on the cornea and conjunctiva. Their utility in drug delivery will be discussed in subsequent sections.

Role of Cornea in Topical Drug Delivery Cornea is the outermost avascular and transparent domeshaped structure of the eye. It consists of five layers (in the direction from anterior to posterior): epithelium, Bowman’s layer, stroma, Descemet’s membrane, and corneal endothelium (Figure 3). The lipoidal corneal epithelium is comprised of five to six layers of tightly adherent columnar cells with tight-junction proteins called zonulae occludens – acting as a major barrier to hydrophilic drugs. On the other hand, the stroma – which is comprised of 90% water – lies directly beneath the corneal epithelium and acts as a rate-limiting barrier to lipophilic drugs. Thus,

even if a molecule is sufficiently lipophilic to rapidly cross the epithelium, penetration through the stroma is still rate limiting. In addition, the physicochemical properties of the drug itself limit its permeability across the cornea. More recently, the expression of multidrug resistance proteins such as the P-glycoprotein (P-gp) and multidrugresistance-associated proteins (MRPs) has been reported on rabbit and human corneal epithelium. They have gained attention lately, since majority of the drug molecules applied topically have been categorized as substrates to one or more of these efflux pumps. In fact, P-gp and MRP-2 localized on the rabbit corneal epithelium can act as a barrier to in vivo drug absorption through cornea. Lipophilic prodrug derivatization has been considered as a viable strategy to enhance transcorneal permeation of ocular therapeutic agents. High octanol/water coefficient of these prodrugs can facilitate permeation across the corneal epithelium. The water-laden stroma, in turn, does not act as a barrier to the regenerated hydrophilic parent drug – thereby enhancing the overall permeability of the prodrug across cornea. Esterase activity has been reported to be the highest in iris-ciliary body followed by cornea and the aqueous humor. Even though high levels of esterases are reported in the iris-ciliary body, the activity in the cornea is highly relevant since cornea acts as a major permeation pathway to these lipophilic ester prodrugs. Bulk of esterase-mediated hydrolysis takes place in the corneal epithelium where the esterase activity is about 2 times to that of stroma and endothelium. Lipophilic ester prodrug design has been employed for a variety of ocular therapeutic agents, which suffer from poor ocular absorption. Esterification of prostaglandin F2a

Corneal epithelium (six layers)

Bowman’s layer

Stroma

Descemet’s membrane Figure 3 Structure of cornea.

305

Endothelium

306

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

(PGF2a) analogs which exhibit higher potency than PGF2a have resulted in blockbuster drugs such as bimatoprost, travoprost, latanoprost, and isopropyl unoprostone. Travoprost, latanoprost, and isopropyl unoprostone are isopropyl esters of PGF2a analogs, whereas bimatoprost is an ethyl amide prodrug (Figure 4). The free acid form of all the above prodrugs has shown poor permeability across cornea, thus obviating the need for lipophilic ester prodrug design. Such strategy is particularly applicable to b-blockers such as timolol – widely used in the treatment of glaucoma – which suffers from high incidence of cardiovascular and respiratory side effects due to systemic absorption of the topically administered dose. Lipophilic prodrug derivatization of timolol to acetyl, propionyl, and butyryl ester prodrugs resulted in corneal permeabilities 2–3 times higher than timolol. Moreover, the enhanced corneal permeation leads to a four- to sixfold increase in the aqueous humor concentrations. Better corneal permeation and high aqueous humor concentration of lipophilic ester prodrugs of timolol consequently resulted in a twofold reduction in the topical dose. Reduced dose can, in turn, decrease the concentration in the systemic circulation, thereby reducing the incidence of cardiovascular and respiratory side effects. The lipophilic ester prodrug design has been extended to antiviral agents such as acyclovir (ACV) and ganciclovir (GCV). These are highly potent against Herpes simplex virus (HSV). Currently available therapy for HSV keratitis involves the use of a 1%-trifluorothymidine (TFT) solution. However, long-term treatment with TFT raises potential concerns due to its high cytotoxicity. ACV is a potent candidate effective against HSV which is available as a 3% ophthalmic ointment. However, due to various problems associated with the use of ointments in the eye, it has not been approved in the United States. Ocular bioavailability of these compounds is extremely limited because of poor corneal permeability. Corneal permeability coefficients increase with increasing lipophilicity for monoester prodrugs of GCV, with the optimal prodrug form having three or four carbon atoms in the side chain. The apparent permeability (Papp) of the valerate-ester prodrug of GCV was sixfold higher than the parent drug GCV, across the cornea. Acylation of ACV also led to improved corneal permeation of the parent drug. There is a linear relationship between the corneal permeability coefficient and the octanol/water partition coefficient with a positive slope, except for ACV isobutyrate – which displayed an anomalously low corneal permeability despite improved lipophilicity. The anomaly with branched alkyl side chain could be explained due to enhanced stability of these prodrugs toward esterases present in the corneal epithelium. The branched side chain offers steric hindrance for enzyme accessibility. The concept has been proven where the stability of homologous series of oxprenolol esters increases with

increasing carbon chain length of the side chain. It has been shown that hydrolysis rates are moderate for O-propionyl, O-butyryl, and O-valeryl prodrugs of oxprenolol, with O-acetyl-oxprenolol being highly unstable and O-pivaloyl-oxprenolol being highly stable (Table 1). Again, the steric hindrance offered by the bulky tertiary butyl group in the pivaloyl derivative has been cited as the primary reason for superior enzymatic stability. However, all of these prodrugs described above are highly lipophilic and possesses very low aqueous solubility to be formulated into aqueous drops for topical administration. Reports indicate that the aqueous solubility of the lipophilic ester prodrugs decrease with increasing carbon side-chain length. Low aqueous solubility has been considered as a major drawback for formulating these lipophilic ester prodrugs into eyedrops. Thus, for a compound to be effective topically and to be formulated into eyedrops, it must possess sufficient hydrophilicity and at the same time exhibit sufficient permeability across the cornea to reach therapeutic levels. Recently, transporter-targeted prodrug approach has received significant attention and a number of membrane transporters have been discovered in various ocular tissues such as the cornea, conjunctiva, and retina. These transporters are involved in the translocation of essential nutrients and xenobiotics across biological membranes. Ocular transporters include carriers for peptides, amino acids, glucose, lactate, and nucleosides/nucleobases and are primarily localized on the corneal epithelium, corneal endothelium, retinal pigmented epithelium (RPE), and retinal capillary endothelium. Prodrugs or analogs designed to target these transporters can significantly enhance the absorption of poorly permeating parent drug. Both solubility and the desired membrane permeability can be achieved by proper selection of the promoiety. Such prodrugs are recognized by the membrane transporters as substrates and are translocated across the epithelia. Subsequently, the prodrugs are enzymatically cleaved to release the parent drug and free ligand which in most cases is a nutrient and nontoxic. Of late, such transporter-targeted prodrug design has been applied to the antiviral drugs ACV and GCV. The existence of an oligopeptide transport system (PEPT1) on the rabbit corneal epithelium has been demonstrated which can transport peptidomimetic prodrugs of acyclovir. Permeation of valine-ester prodrug of acyclovir (L-Val-ACV) across the cornea has been found to be much higher than that of the parent drug ACV (Table 2). The transport of L-Val-ACV is saturable at higher concentrations, pH dependent, and competitively inhibited by other known PEPT1 substrates indicating its translocation by PEPT1 present on the cornea. This result prompted the investigators to synthesize a series of water-soluble dipeptide prodrugs of acyclovir targeting the peptide transport system on the cornea to improve the ocular bioavailability. The structures of ACV, valine-based amino acid ester of

Drug Delivery to Cornea and Conjunctiva: Esterase- and Protease-Directed Prodrug Design

O

O

HO

HO

OH

OH

HO

HO

(a)

(b)

HO

HO

OH OH

O

HO

O

HO

O

O

HO

O

F

O F (c)

(d)

F

OH

OH

O N H O

HO

HO

(e)

O

HO

HO

(f)

Figure 4 Chemical structures: (a) PGF2a, (b) 17-phenyl-PGF2a, (c) isopropyl unoprostone, (d) travoprost, (e) bimatoprost, and (f) latanoprost.

307

308

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Table 1 Half-life values for a homologous series of oxprenolol (O) esters in 0.05 M phosphate buffer (pH 7.4), 30% human plasma, aqueous humor, and corneal extract at 37 C (data taken from Geraldine, C. and Jordan M. 1998) Half-life (t1/2, min) Ester

Buffer (pH 7.4)

Human plasma 30%

Aqueous humor

Corneal extract

O-acetyl O-propionyl O-butyryl O-valeryl O-pivaloyl

9.1 10.4 19.1 21.1 2035.5

4.3 6.4 12.8 16.2 263.2

6.1 14.6 18.3 19.6 687.8

9.6 11.3 13.0 16.4 375.5

Table 2 Permeability of dipeptide prodrugs of ACV across freshly excised rabbit cornea Drug

Papp ( 106 cm s1)

ACV VACV VVACV GVACV YVACV VYACV

4.24 1.41a 12.1 0.44* 9.91 2.40* 12.4 1.42* 7.19 1.38* 8.34 1.12*

a

Control All values are expressed as mean SD. Statistically significant difference between control and test is represented by * (p < 0.05). ACV, acyclovir; VACV, val-acyclovir; VVACV, val-val-acyclovir; GVACV, gly-val-acyclovir; YVACV, tyr-val-acyclovir; VYACV, val-tyr-acyclovir. Data from Anand, B. S., Nashed, Y. E., and Mitra, A. K. (2003). Novel dipeptide prodrugs of acyclovir for ocular herpes infections: Bioreversion, antiviral activity and transport across rabbit cornea. Current Eye Research 26: 151-163, with permission from Taylor & Francis.

ACV (VACV), and valine- and glycine-based dipeptide conjugate of acyclovir – valine–valine–acyclovir (VVACV) and glycine–valine–acyclovir (GVACV) are shown in Figure 5. VACV is found to be hydrolyzed primarily by esterases to release ACV, whereas VVACV is hydrolyzed initially by aminopeptidases to release VACV, which is further acted upon by esterases to release ACV (Figure 6). Direct conversion of VVACV to ACV appears to be minimal. These prodrugs exhibited excellent solution stability and solubility allowing formulation into suitable eyedrops. All dipeptide prodrugs of ACV exhibited enhanced transcorneal permeability resulting in higher ocular bioavailability in rabbits, with GVACV being the highly permeable prodrug (Table 2). The prodrugs also exhibit higher antiviral efficacy against HSV epithelial keratitis and stromal keratitis and are less cytotoxic and more effective than trifluorothymidine (TFT) – a current drug of choice. In addition to peptide transporters, even amino acid transporters such as ASCT1 (Naþ-dependent neutral amino acid transporter) and B0,þ (Naþ-dependent neutral and cationic amino acid transporter) have been explored for ocular delivery of ACV. A series of amino acid ester prodrugs including alanine-ACV, serine-ACV, isoleucine-ACV,

g-glutamate-ACV, and valine-ACV were synthesized and evaluated by Katragadda and colleagues for in vivo corneal absorption against the parent drug, ACV. Results showed that the amino acid-ester prodrug, serine-ACV – owing to its enhanced stability – exhibited higher area under the curve, Cmax and Clast values, in comparison to ACV and seemed to be a promising candidate for the treatment of ocular HSV infections. Gunda and colleagues have applied similar chemical derivatization to GCV – an acyclic guanosine analog. GCV has shown to exhibit excellent antiviral activity against the herpes viruses but suffers from poor corneal permeability due to its hydrophilicity. Hence, in order to enhance the corneal permeability and ocular bio availability, dipeptide monoester prodrugs of GCV targeting the peptide transporter on the corneal epithelium were synthesized. Among the synthesized prodrugs, tyrosine– valine–GCVand valine–valine–GCV, exhibited significantly higher transcorneal permeability ex vivo and in vivo leading to higher ocular bioavailability when compared to GCV.

Role of Conjunctiva in Ocular Drug Delivery Conjunctiva is a transparent, highly vascularized mucous membrane that covers the sclera and lines the inner surfaces of eyelids. It covers almost 80% of exposed ocular surface and is comprised of many small blood vessels and tiny secretory glands. These glands produce tear film that lubricates and protects the eye during its movement in the socket. Three different types of conjunctival membranes have been identified based on its location. Palpebral or tarsal conjunctiva is the one lining the eyelids. Bulbar or ocular conjunctiva is a semipermeable and colorless membrane covering the eyeball, over the sclera. Fornix conjunctiva is where the inner part of the eyelids and the eyeball meet. It is loose and flexible, allowing free movement of the lids and eyeball. The tissue consists of pseudostratified columnar epithelium rich in goblet cells and it contains ductules of main lachrymal gland, accessory lachrymal glands, and lymphoid follicles. The space between the palpebral and bulbar conjunctiva is called the conjunctival cul-de-sac. Conjunctiva is typically composed of two layers, an outer epithelium and underlying stroma. The epithelium consists of 5–15 layers of

Drug Delivery to Cornea and Conjunctiva: Esterase- and Protease-Directed Prodrug Design

O

O

N

N

N

N

HN

HN

H2N

309

O

CH3

H2N

N

N

O O

H3C

O HO (a)

NH2

(b)

O

N HN

O

H2N

N

N

H N

O O

H2N O

(c)

O

N

HN

O

H2N

H N

N

N

O O

H2N O (d)

Figure 5 Chemical structures: (a) acyclovir, (b) valine–acyclovir, (c) valine–valine–acyclovir, and (d) glycine–valine–acyclovir.

stratified epithelial cells and is covered with microvilli. The stroma loosely attaches to the underlying sclera. It contains all the lymphatics and blood vessels. The larger surface area and pore density coupled with expression of various nutrient transporters listed in the subsequent sections makes the conjunctiva more amenable for topical drug delivery. The surface area of the conjunctiva is almost 9 and 17 times larger than that of cornea in rabbits and humans, respectively. Compared to cornea, the paracellular pore size as well as the pore density in conjunctiva is larger. Due to such increased pore size and density small peptides and oligonucleotides can permeate across conjunctival pores. However, for

hydrophilic drugs instilled topically, the lipophilic conjunctival epithelium acts as rate-limiting barrier for drug absorption. In addition, an external enzymatic barrier in the conjunctival epithelium – specifically the proteases – restricts the penetration of peptide drugs like enkephalins, substance P, and insulin. One of the major disadvantages associated with conjunctival absorption following topical administration is drainage of drug molecules by conjunctival blood vessels into the systemic circulation. Moreover, it has been shown that the conjunctiva expresses efflux transporters including P-gp on the apical side of conjunctival epithelium. These can limit the absorption of several fluoroquinolones such as levofloxacin, gatifloxacin, and

310

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Aminopeptidases Val–Val–ACV

Val–ACV

Esterases

Esterases

ACV Figure 6 Metabolic pathway of valine–valine–acyclovir (VVACV) to valine-based amino acid ester of acyclovir (VACV) and acyclovir (ACV). From Anand, B. S., Nashed, Y. E., and Mitra, A. K. (2003). Novel dipeptide prodrugs of acyclovir for ocular herpes infections: Bioreversion, antiviral activity and transport across rabbit cornea. Current Eye Research 26: 151–163, with permission from Taylor & Francis.

grepafloxacin instilled topically. Thus, improving the drug absorption across conjunctiva is one of the major challenges needed to be overcome. Based on the target tissue, conjunctival drug delivery can be divided into subconjunctival and transconjunctival delivery. Transconjunctival Pathway In this route, agents can be targeted toward the conjunctiva for treatment of local conjunctival infection as well as other anterior chamber diseases such as dry eye syndrome and glaucoma. Transconjunctival absorption could also result in higher concentrations in posterior ocular tissues through conjunctival–scleral pathway. However, subconjunctival pathway is widely exploited to deliver drugs to posterior ocular segment. Hydrophilic molecules can permeate through paracellular pathway between epithelial cells through the tight junctions. However, their penetration will be extremely low due to the small surface area of paracellular pathway compared to transcellular route. Transconjunctival drug penetration can be enhanced by increasing the lipophilicity of the drug molecule through chemical modification, mostly as prodrugs or analogs. Propranolol – with a log partition coefficient between octanol and water (log P) of 3.21 – is absorbed through cornea and conjunctiva up to tenfold greater than a hydrophilic drug of similar size, for example, sotalol with a log P of 0.62. Prodrug strategy has not been exploited in conjunctiva as extensively as in cornea. As discussed earlier, in vivo activation of the prodrug is very essential for the success of this approach. As ester and amide linkages are widely utilized in prodrug design, it is important to know the activities of esterases and amidases in conjunctiva. As mentioned before, esterase and aminopeptidase activity has been reported in conjunctiva along with the cornea and iris-ciliary body. These play a role in hydrolysis of ester and amide bases prodrugs in the conjunctiva.

The peptidases can also hydrolyze short-chain, biologically active peptides such as methionine and leucine enkephalins. There is good number of studies demonstrating esterases/amidases targeted prodrug approach to enhance drug absorption across conjunctiva. A few of them are discussed below. PGs can lower intraocular pressure (IOP) in openangle glaucoma. However, due to the low permeability of PGF2a across the cornea, a relatively high concentration of PGF2a is necessary for effective IOP reduction. This can result in conjunctival hyperemia, ocular discomfort, headaches, and other side effects. Chen and colleagues have evaluated lipophilic PGF2a ester prodrugs for improving the permeability of PGF2a following topical administration. A series of lipophilic esters such as PGF2a 1-isopropyl, 1, 11-lactone, 15-acetyl, 15-pivaloyl, 15-valeryl, and 11, 15-dipivaloyl esters have been evaluated for transport and bioreversion in rabbit cornea, conjunctiva, and iris-ciliary body (Figure 7). All of the prodrugs penetrated the rabbit cornea and conjunctiva faster than PGF2a except the 15-acetyl ester prodrug, which is equally permeable as PGF2a across conjunctiva. However, a direct relationship is not observed between the degree of apparent permeability and prodrug lipophilicity. The two most lipophilic prodrugs – the 15-valeryl and 11, 15-dipivaloyl esters – were less permeable in the cornea and conjunctiva than other prodrugs with comparatively lower lipophilicity. The high permeabilities of 1-isopropyl ester and 1, 11-lactone is attributed to their state of ionization at the site of absorption where they exist more in unionized form. The prodrug 1, 11-lactone hydrolyzed at a slower rate in the ocular tissues probably due to the intracyclic ester linkage. The bulky pivaloyl group in 15-pivaloyl prodrug rendered it enzymatically more stable than the 1-isopropyl and other 15-monoester prodrugs. Bulky moiety (steric hindrance) might have impeded access of esterases to hydrolyze the pivaloyl ester linkage. Similarly, the prodrug with two pivaloyl groups – 11, 15-dipivaloyl prodrug – is much more stable as expected. Thus, the size and structural branching of the promoiety can influence enzymatic hydrolysis of prodrugs which, in turn, can further determine the permeability across tissues. Corticosteroids find applications in the treatment of uveitis and postsurgical inflammation. However, these drugs can cause significant adverse effects such as increase in IOP, cataract, and herpetic reactivation. Dexamethasone (DX) is the one drug mostly indicated for ocular corticosteroid treatment. Several lipophilic esters of dexamethasone prodrugs have been evaluated for their ocular permeability and bioreversion in bovine conjunctival epithelial cell line (BCEC) and isolated rabbit cornea (Figure 8). This study is mainly aimed at selecting an optimum prodrug that improves the delivery of dexamethasone to the target tissue with minimum permeation across other ocular tissues to reduce adverse effects. The permeability

Drug Delivery to Cornea and Conjunctiva: Esterase- and Protease-Directed Prodrug Design

MWa

compound

log Pb R1(C-15)

R2(C-1)

311

R3(C-11)

OH CO2R2

R3O PGF2α 15-Acetyl ester 1-Isopropyl ester 1,11-Lactone

355 396 396 337 OH

OR1 H CH3CO H

1.26 2.16 2.50 2.61

H H H H CH2(CH3)2 H O C

O 439 15-Pivaloyl ester 15-Valeryl ester 439 11,15-Dipivaloyl ester 523

3.50 3.75 5.0

OH C(CH3)3CO H H C4H9CO C(CH3)3CO H

H H C(CH3)3CO

a Molecular weight. b Log P was determined from the Pomona Med. Chem. Software system or measured by HPLC.

Figure 7 Chemical structures of PGF 2a and its prodrugs. From Chien, D. S., Tang-Liu, D. D., and Woodward, D. F. (1997). Ocular penetration and bioconversion of prostaglandin F2alpha prodrugs in rabbit cornea and conjunctiva. Journal of Pharmaceutical Sciences 86: 1180–1186, with permission from John Wiley & Sons, Inc.

Dexamethasone (DX) ester

Abbreviation

21-sodium phosphate

DSP

21- metasulfobenzoate

DSB

21-acetate

DAC

17-propionate

DPR 17

21-propionate

DPR

21-butyrate

DBU

21-valerate

DVA

21-palmitate

DPALM

O C CH3

HO

17 CH3

F

21 CH2OH OH CH3

H

H

O

Figure 8 Dexamethasone esters and their abbreviations. From Civiale, C., Bucaria F., Piazza S., et al. (2004). Ocular permeability screening of dexamethasone esters through combined cellular and tissue systems. Journal of Ocular Pharmacology and Therapeutics 20: 75–84, with permission from JOPT.

of these dexamethasone prodrugs correlated well with their lipophilicity until a maximum value is reached, which corresponded to dexamethasone butyrate (log P ¼ 3.95). DSP and DSB prodrugs are hydrophilic and thus do not permeate across BCEC as expected. Other prodrugs exhibited higher permeability with increase in lipophilicity till DBU, after which it started to plateau as observed with DVA (Figure 9). A similar trend has been observed across the cornea with exceptions at the extreme ends of log P values (DSP and DVA) which has been attributed to the rate of hydrolysis of prodrugs in corneal tissue and BCEC an epithelial layer. DVA being highly lipophilic gets absorbed into corneal

epithelium and its further permeation is limited by hydrophilic corneal stroma unless it is hydrolyzed to relatively more hydrophilic dexamethasone. However, the hydrolysis of DVA to dexamethasone is very poor in cornea, confining the prodrug in the corneal epithelium (Table 3). This may be optimum for delivery of dexamethasone to cornea, where the prodrug in the corneal epithelium can hydrolyze slowly and sustain the release. Thus, it reduces the drug concentrations in aqueous humor and other intraocular tissues minimizing its side effects. Unlike in BCEC, DSP has been found to hydrolyze rapidly in the cornea creating a concentration gradient for higher absorption into cornea. DBU is the

312

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

prodrug that is completely hydrolyzed both in BCEC and cornea and also has tremendously improved the permeability of dexamethasone. This may be more suitable prodrug to deliver dexamethasone to other intraocular tissues. These studies clearly show that both the promoiety as well as the ability of the prodrug to get hydrolyzed are important in designing an optimum prodrug in ocular drug delivery. A series of alkyl, cycloalkyl, and aryl ester prodrugs of timolol has been evaluated to examine the effect of enzymatic lability of prodrugs on corneal and conjunctival penetration of timolol. Straight-chain alkyl and the unsubstituted cycloalkyl esters hydrolyzed more rapidly than their corresponding branched-chain and substituted analogs as well as the aryl esters. This might be due

Papp (10–6 cm s–1)

60

40

20

0 0

1

2

3

4

5

Log P Figure 9 Permeability rates of dexamethasone esters through BCEC (filled squares) and excised cornea (empty circles) versus log P. Left to right, 21-sodium phosphate ester (DSP), 21-metasulfobenzoate ester (DSB), dexamethasone (DX), 17-propionate ester (DPR17), 21-acetate ester ( DAC), 21-propionate ester (DPR), 21-butyrae ester (DBU), and 21-valerate (DVA). From Civiale, C., Bucaria F., Piazza S., et al. (2004). Ocular permeability screening of dexamethasone esters through combined cellular and tissue systems. Journal of Ocular Pharmacology and Therapeutics 20: 75–84, with permission from JOPT.

Table 3

to free accessibility of the ester linkage to esterases in straight-chain alkyl esters. The slower hydrolysis of branched chain alkyl esters might be due to steric hindrance to hydrolyzing enzymes. As expected, the esters with enzymatically labile straight-chain alkyl chains penetrated the cornea and conjunctiva at faster rates than the esters with branched alkyl chains that are less labile toward enzymatic hydrolysis. Thus, this study has shown that the rate of enzymatic hydrolysis can highly influence the corneal/ conjunctival absorption of prodrugs. Carrier-mediated transport represents an area of growing interest to pharmaceutical scientists. These transport systems play an important role in absorbing nutrients as well as their mimetic drugs. Transporter-targeted prodrug delivery can be an effective strategy to improve the permeability of poorly absorbed drugs, as conjunctiva has been shown to express various transporters for nutrients such as amino acid, peptides, L-lactate, and nucleosides. These transporters can be utilized for improved drug absorption across the conjunctiva. However, such a strategy has not been explored to a large extent. Functionally active sodium-dependent, carriermediated, monocarboxylate transport system has been reported on the mucosal or tear-side of the rabbit conjunctival epithelium. It has been shown to recognize and translocate nonsteroidal anti-inflammatory drugs (NSAIDs) and fluoroquinolone antibiotics administered topically. The conjunctiva expresses Naþ-glucose transporter (SGLT1) on the mucosal side and a sodium-dependent nucleoside transporter has been characterized that can be targeted by nucleoside mimetic antivirals on conjunctiva. Various amino acid-transporter systems including L-lysine and B0,þ are present on the apical side of rabbit conjunctiva. Among various transporters exploited for drug delivery, peptide transporters (PEPT1 and 2) have become popular due to their high capacity and wide substrate specificity. These play an important role in the translocation of di- and tripeptides and peptidomimetic drugs

Transport of dexamethasone and its esters across BCEC and rabbit cornea In vitro

Ex vivo

Steroid

Log P

Papp ( 10 cm s )

Conv. to DX (%)

Papp ( 106 cm s1)

Conv. to DX (%)

DX DSP DSB DAC DPR17 DPR DBU DVA

2.12 0.54 1.65 2.92 2.69 3.19 3.95 4.70

1.08 0.17 < 0.02 < 0.02 18.2 1.11 6.29 0.22 16.8 2.17 27.8 1.64 29.4 5.4

-

5.06 1.01 3.87 0.62 0.51 0.23 21.1 1.17 18.6 4.94 29.1 3.70 53.8 13.8 14.8 6.65

100 42 100 100 100 100 21

6

1

0 0 90 100 61 99 99

DX, Dexamethasone; DSP, 21-sodium phosphate ester; DSB, 21-metasulfobenzoate ester; DAC, 21-acetate ester; DPR17, 17-propionate ester; DPR, 21-propionate ester; DBU, 21-butyrae ester; DVA, 21-valerate. Data from Civiale, C., Bucaria, F., Piazza, S., et al. (2004). Ocular permeability screening of dexamethasone esters through combined cellular and tissue systems. Journal of Ocular Pharmacology and Therapeutics 20: 75-84, with permission from JOPT.

Drug Delivery to Cornea and Conjunctiva: Esterase- and Protease-Directed Prodrug Design

across various tissues. A proton-coupled dipeptide transporter process has been reported to be present on pigmented rabbit conjunctiva. A proton-coupled, saturable, and temperature-dependent uptake of a dipeptide on rabbit conjunctival epithelial cells (RCEC) was reported. However, the expression level of the dipeptide transporters seems to be rather low in conjunctiva. Overall, the presence of a variety of influx transporters on the conjunctiva offers an immense potential for targeted delivery after topical prodrug administration. Subconjunctival Delivery This approach of drug delivery is more popular and safer and less invasive than intravitreal injection. Moreover, it offers the potential advantage of localized, sustained delivery in the treatment of ocular diseases affecting the posterior segment such as age-related macular degeneration (AMD) and diabetic retinopathy. As this route can circumvent both the corneal and conjunctival barriers, it has immense potential for delivery of both small molecules and macromolecular drugs to posterior ocular tissues including choroid and retina. Moreover, sclera is much more permeable than conjunctiva. Drug delivery by this route exploits the large surface area, easy accessibility, and relatively high permeability of sclera. By this route, sustained intraocular therapeutic drug concentrations can be achieved without surgical implantation of controlled-release drug-delivery implant in the vitreous humor or by repeated intravitreal/periocular injections. In this regard, a subconjunctival injection of drug-loaded nanoparticulate system offers potential alternative. These systems can release the drug in a sustained manner as well as achieve higher drug concentrations in the target tissues. Drug-loaded nanoparticulate systems prepared with biodegradable polymers have been widely used in sustaining the drug release. Nanoparticles are usually taken up into the cell by endocytosis which will significantly enhance the uptake of nanoparticles into the targeted cells. Moreover, receptor-mediated endocytosis have attracted attention due to high capacity and targeting feasibility. Expression of various receptors by RPE will enable this strategy to be successfully applied for subconjunctival administration of receptor-targeted nanoparticles. In this strategy, a promoiety is conjugated to the polymer, which is used to prepare these nanoparticles. It is also ensured that these targeting promoieties are present on the surface of nanoparticles, such that it is recognized by a specific receptor present on the RPE and undergo receptor-mediated endocytosis. Thus, drug loaded nanoparticles improve drug absorption into RPE, enable targeting, and sustain the release for prolonged periods of time. In addition to this, a sustained drug release can also be achieved by dispersing the drug/prodrug in a thermosensitive gelling polymer. These are the polymers that are

313

liquids at room temperature and gels at body temperature. Drug/prodrug can be mixed homogenously in the polymer solution and injected subconjunctivally. Upon injection, the polymer solution gels and sustains the release of drug. The drug loading and release can be optimized to maintain therapeutic concentrations for prolonged periods of time.

Conclusion Prodrugs have proven to be an effective strategy for drug delivery to the anterior segment. Several examples of successfully marketed ophthalmic prodrugs including dipivalyl ester of epinephrine (dipivephrine) are available. In addition, ester prodrugs of PGF2a analogs such as bimatoprost, travoprost, and latanoprost are highly effective. However, the application of this strategy to a particular drug molecule depends upon its chemical structure. The drug molecule needs to have a specific functional group such as carboxyl/hydroxyl/amine group to facilitate the conversion to an ester or amide-based prodrug. Despite this, a significant amount of research needs to be performed to examine the feasibility of extending the prodrug strategy to deliver macromolecular drugs like proteins and antibodies which have been highly effective in treatment of various ocular diseases, especially the AMD. See also: Cornea Overview; Corneal Angiogenesis; Corneal Endothelium: Overview; Corneal Epithelium: Transport and Permeability; The Corneal Stroma; Overview of Electrolyte and Fluid Transport Across the Conjunctiva.

Further Reading Anand, B. S., Nashed, Y. E., and Mitra, A. K. (2003). Novel dipeptide prodrugs of acyclovir for ocular herpes infections: Bioreversion, antiviral activity and transport across rabbit cornea. Current Eye Research 26: 151–163. Chang, S. C., Bundgaard, H., Buur, A., and Lee, V. H. (1987). Improved corneal penetration of timolol by prodrugs as a means to reduce systemic drug load. Investigative Ophthalmology and Visual Science 28: 487–491. Chien, D. S., Sasaki, H., Bundgaard, H., Buur, A., and Lee, V. H. (1991). Role of enzymatic lability in the corneal and conjunctival penetration of timolol ester prodrugs in the pigmented rabbit. Pharmaceutical Research 8: 728–733. Chien, D. S., Tang-Lio, D. D., and Woodward, D. F (1997). Ocular penetration and bioconversion of prostaglandin F2alpha prodrugs in rabbit cornea and conjunctiva. Journal of Pharmaceutical Sciences 86: 1108–1186. Civiale, C., Bucaria, F., Piazza, S., et al. (2004). Ocular permeability screening of dexamethasone esters through combined cellular and tissue systems. Journal of Ocular Pharmacology and Therapeutics 20: 75–84. Geraldine, C. and Jordan, M. (1998). How an increase in the carbon chain length of the ester moiety affects the stability of a homologous series of oxprenolol esters in the presence of biological enzymes. Journal of Pharmaceutical Sciences 87: 880–885.

314

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Gunda, S., Hariharan, S., and Mitra, A. K. (2006). Corneal absorption and anterior chamber pharmacokinetics of dipeptide monoester prodrugs of ganciclovir (GCV): In vivo comparative evaluation of these prodrugs with Val-GCV and GCV in rabbits. Journal of Ocular Pharmacology and Therapeutics 22: 465–476. Horibe, Y., Hosoya, K., Kim, K. J., and Lee, V. H. (1998). Carriermediated transport of monocarboxylate drugs in the pigmented rabbit conjunctiva. Investigative Ophthalmology and Visual Science 39: 1436–1443. Hosoya, K., Lee, V. H., and Kim, K. J. (2005). Roles of the conjunctiva in ocular drug delivery: A review of conjunctival transport mechanisms and their regulation. European Journal of Pharmaceutics and Biopharmaceutics 60: 227–240. Hughes, P. M. and Mitra, A. K. (1993). Effect of acylation on the ocular disposition of acyclovir. II: Corneal permeability and anti-HSV 1 activity of 20 -esters in rabbit epithelial keratitis. Journal of Ocular Pharmacology 9: 299–309.

Kashi, S. D. and Lee, V. H. (1986). Hydrolysis of enkaphilins in homogenates of anterior segment tissues of the albino rabbit eye. Investigative Ophthalmology and Visual Science 27: 1300–1303. Katragadda, S., Gunda, S., Hariharan, S., and Mitra, A. K. (2008). Ocular pharmacokinetics of acyclovir amino acid ester prodrugs in the anterior chamber: Evaluation of their utility in treating ocular HSV infections. International Journal of Pharmaceutics 359: 15–24. Lee, V. H. (1983). Esterase activities in adult rabbit eyes. Journal of Pharmaceutical Sciences 72: 239–244. Stratford, R. E., Jr. and Lee, V. H. (1985). Ocular aminopeptidase activity and distribution in the albino rabbit. Current Eye Research 4: 995–999. Tirucherai, G. S., Dias, C., and Mitra, A. K. (2002). Corneal permeation of ganciclovir: Mechanism of ganciclovir permeation enhancement by acyl ester prodrug design. Journal of Ocular Pharmacology and Therapeutics 18: 535–548.

Knock-Out Mice Models: Cornea, Conjunctiva, Eyelids and Lacrimal Gland W W-Y Kao, C-Y Liu, and H Liu, University of Cincinnati, Cincinnati, OH, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Cre – The phage recombinase that catalyzes cyclization recombination of LoxP elements. EGF – The epidermal growth factor which binds to its receptor, EGFR. FGF – The family of growth factors involved in angiogenesis, wound healing, and development. Keratins – The family of structural proteins that are tough and insoluble. Keratin 12 is specific to the cornea. Keratocan – A member of keratan sulfate proteoglycans that is important for transparency of the cornea. LoxP – The locus of crossover within P1 phage. Reverse tetracycline transcriptional activator (rtTA) – Binds to Tet operator sequence 7 and activates transcription of the target gene in the presence of tetracycline (Tet-on system). Tetracycline operator element – The tetracyclineregulated promoter that contains tetracycline operator sequences, which regulate expression of a downstream gene.

Transgenesis Transgenesis through microinjection of cloned DNA into fertilized mouse eggs was first accomplished in the laboratories of Brinster, Costantini, Ruddle, Mintz, and Wagner. The availability of a cell-type-specific promoter is a prerequisite for the success of creating transgenic mouse lines that exhibit altered phenotypes caused by the presence of such reporter gene product in tissues of interest. Since then, the technology of transgenesis has advanced to create inducible transgenic mouse lines in which a reporter gene has a unique spatial and temporal expression pattern by administering antibiotics, hormones, and pheromones to experimental animals. The system usually consists of two transgenic mouse lines, one of which employs a tissue-specific promoter for the expression of a transgene that encodes a fusion protein of a transcription factor and hormone receptor and antibiotic suppressor. The other is a transgenic mouse carrying a reporter gene following the responsive elements of a transcription factors and suppressor. In the bitransgenic offspring from the mating of the two transgenic mice, the reporter genes can be turned on and off when ligands bind to the fusion proteins of transcription factor/receptor and transcription factor/suppressor fusion proteins. For example, tet-ON and tet-OFF system in overexpression of tet-OFGF7 by epidermal epithelium under the control of reverse tetracycline transcription activator (rtTA) driven by keratin 5 (K5) and K14 promoters of bitransgenic K5-rtTA/tet-OFGF7 and/or K14-tTA/tet-O-FGF7 mice.

Introduction Transgenesis, that is, the insertion of an exogenesis gene into an organism such that the gene is transmitted to the offspring, and gene targeting are among the most important biological techniques developed in the twentieth century. The studies of genetically modified mutant mice by transgenesis and gene targeting are of great value to elucidate the pathophysiology of altered gene functions and have greatly increased our knowledge of normal physiology and diseases in humans. They not only provide the means for the generation of animal models that are used to examine the pathogenesis of human diseases caused by altered genetic functions, but also allow the development of gene and cell therapy strategies to treat diseases. For example, the application of transgenesis and gene-targeting techniques opens the door of targeted introduction of genes to cells of diseased tissues, that is, gene therapy. Thus, lost cellular and tissue functions can be restored and diseases are cured.

Gene Targeting In late 1980, Doetschman et al. and Thomas et al. demonstrated that the mouse genome could be modified in vitro in embryonic stem cells by homologous recombination. It was subsequently demonstrated that the modified mouse genome could be transmitted to offspring through injection of such genetically modified embryonic stem cells into blastocysts. Since then, many gene-targeting strategies have been developed not only to ablate functional genes (i.e., knockout), but also to replace target genes with another functional gene and/or a mutant gene of interest (knock-in). These strategies allow us to examine gene function in experimental mouse lines that mimic pathogenesis of human diseases and yield useful information, leading to a better understanding of normal physiology and pathophysiology in humans. However, many of the altered genes often lead to embryonic lethality and detrimental effects on animal

315

316

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

development, limiting their use in studying biology of the ocular surface tissues. To circumvent the pitfalls, one of the strategies is conditional gene knock-out that employs Cre–LoxP system to ablate gene in a cell-type-specific manner. Tissue-specific gene ablation using Cre–LoxP system

The Cre–LoxP system was developed to avoid embryonic lethality as well as to confine the inactivation of the target gene in a cell- or tissue-specific manner. Cre is a phage recombinase that specifically deletes any DNA sequence flanked by two LoxP elements. The Cre–LoxP system consists of two mouse lines, one of which uses a tissue-/ cell-type-specific promoter for the expression of Cre. The other mouse line contains a modified genome in which two LoxP elements are inserted in two introns flanking functionally important exon(s) of the targeted gene by gene-targeting techniques. The LoxP-modified (floxed) gene remains fully functional, except in cells expressing Cre that is under the control of a tissue-specific promoter. Thus, the offspring are characterized by the ablation of the gene in a tissue-/cell-type-specific manner. The system allows the inactivation of the target gene in a single cell type and/or a limited number of cell types, depending on the specificity of the promoter, and reduces the probability of embryonic lethality of the experimental mice. Thus, it permits the analysis of physiological and pathophysiological consequences of the genetic alteration in mature animals.

Table 1

Pitfalls of Cre–LoxP system: Cryptic Cre expression by germ cells during gametogenesis

We have recently discovered that recombination of floxed genes happens during gametogenesis in several Cre/LoxP bitransgenic mouse lines that express tissue-specific Cre outside the gonads, for example, Kera-Cre, Krt12-Cre, and BF1-Cre mice. The frequency of the promiscuous LoxP/ Cre recombination varied in different lines of Cre driver mice and sex of the same driver mice with higher penetrance in male than in female double transgenic mice as well (Table 1). Polymerase chain reaction (PCR) and recombination analysis demonstrate recombination of floxed allele occurs during transition from spermatogonia (diploid) to primary spermatocyte (tetraploid) in testis. Thus, target floxed allele(s) are ubiquitously ablated in Cre/LoxP mice intended for tissue-specific gene deletion. Increasing evidence indicates that many cell-typespecific genes in adults are promiscuously expressed in gonad during gametogenesis. Therefore, the promiscuous expression of Cre recombinase driven by cell-typespecific promoters should always be examined to avoid misinterpretation resulting from haploid deficiency. Inducible Cre–LoxP system To overcome the pitfalls of promiscuous Cre expression during gametogenesis, we have developed a doxycyclineinducible tet-ON system to spatially and temporally ablate gene of interest in ocular surface tissues. This tet-ON system employs triple transgenic mice consisting of an ocularsurface tissue-specific promoter rtTA (Kera-rtTA (KR)),

Excision of floxed alleles during gametogenesis of double transgenic Cre/floxed mice

Father

Mother

Number of Cretg/ floxed mice Total D

KC4.3/ZEG KC4.1/ROSAR KC1/ZEG C57BL/6 K12Cre/w/ZEG C57BL/6 BF1Cre/R26R C57BL/6 WntlCre/R26R Tbr2f/f Smad4f/f

C57BL/6 C57BL/6 C57BL/6 KC4.3/ROSAR C57BL/6 K12Cre/w/ZEG CD-1 Wnt1-Cre/R26R C57BL/6 KC4/Tbr2f/w KC4.3/Smad4f/w

10 4 15 4 NAa NAa 7 1 1 18b 6b

a

10 4 15 2 NAa NAa 5 0 0 5 2

Number of Cre0/ floxed mice Total D

Percentage of penetrancec

10 6 16 1 13 27 9 5 7 26b 7b

100 100 100 60 46 7 69 0 0 23 38

10 6 16 1 6 2 6 0 0 5 3

ZEG transgene is on chromosome 11 same as Krt12. Thus, the ZEG allele is always co-segregated with the Krt12Cre allele Homozygous floxed Tbr2 and Smad4 alleles c Percentage of penetrance is calculated by number of mice with excised floxed allele divided by the total number of mice carrying floxed allele in each experimental group. Double transgenic mice were crossbred with wild-type mice, and homozygous Tbr2f/f and Smad4f/f mice. The excision event was determined by the expression of respective reporter gene activities or by PCR of the excised floxed alleles as described in Methods. CreTg: hemizygous Cre transgenic mice; Cre0: non-Cre transgenic mice. Reproduced from Weng, D. Y., Zhang, Y., Hayashi, Y. et al., (2008). Promiscuous recombination of LoxP alleles during gametogenesis in cornea Cre driver mice. Molecular Vision 14: 562–571. b

Knock-Out Mice Models: Cornea, Conjunctiva, Eyelids and Lacrimal Gland Table 2

317

Mouse lines for cornea-specific genetic modification

Mouse lines

Tissue/cell specificity

Induction

Function

Kera-rtTA/tet-O-reporter

Doxycycline

Overexpression

Not inducible Doxycycline

Gene ablation

Krt12-rtTA/tet-O-reporter Krt12-Cre

Stromal keratocytes in adult and neural crest cells in embryo Stromal keratocytes in adult and neural crest cells in embryo Stromal keratocytes in adult and neural crest cells in embryo Corneal epithelium Corneal epithelium

Krt12-rtTA/tet-O-Cre/Xf/f

Corneal epithelium

Doxycycline Not inducible Doxycycline

Pax6OS-rtTA/tet-Oreporter Pax6OS-rtTA/tet-O-Cre/Xf/f

Ocular surface epithelium and lacrimal gland (?)

Doxycycline

Ocular surface epithelium and lacrimal gland (?)

Doxycycline

Kera-Cre Kera-rtTA/tet-O-Cre/ Xf/f

Krt12-rtTA (K12R), and Pax6OS-rtTA (P6R) (Table 2), and tet-O-Cre (TC) transgenes, as well as an Xf/f (a LoxPmodified gene). Administration of doxycycline will activate rtTA that subsequently turns on the tet-O-Cre transgenes for the synthesis of Cre recombinase, which then excises the Xf/f gene in an ocular-surface tissue-specific manner (as described below).

Strategies of Cornea-Specific Genetic Modification: Transgenic and Knock-Out/ Knock-In Mice Identification of Ocular-Surface Tissue-Specific Promoter The availability of ocular-surface tissue-specific promoter is a prerequisite for the preparation of experimental animal models in which genetic modifications are made to study the consequence of the loss and/or gain of functions of genes in ocular surface tissues, that is, cornea, conjunctiva, lacrimal glands, and eyelids. To date, our laboratory has identified and characterized two corneaspecific genes: keratin 12 (Krt12) and keratocan (Kera) genes of corneal epithelium and stroma, respectively, and used transgenesis and gene-targeting techniques to create mouse lines in which the gene functions are altered in a corneaspecific manner. However, it should be noted that many systemic genetically modified mouse lines that were created by conventional systemic gene-targeting techniques and/or transgenesis with non-ocular-surface tissue-specific promoters also manifested pathogenesis in ocular surface tissues. Naturally, these mouse lines often exhibited pathogenesis in other tissues than ocular surface tissues as well. To date, only a few promoters that exhibit ocular surface tissue specificity have been identified, for example, keratocan (Kera), keratin 12 (Krt12), and modified Pax6 promoters.

Spatial and temporal gene ablation Overexpression Gene ablation Spatial and temporal gene ablation Overexpression Spatial and temporal gene ablation

It should be noted that there is no tissue-specific promoter available that can be used to create mouse lines in which genetic modifications are limited to conjunctiva, cornea endothelium, and lacrimal gland. Stromal keratocyte-specific promoter Keratan sulfate proteoglycans (KSPGs) play a pivotal role in the development and maintenance of corneal transparency. Keratocan, lumican, and mimecan (osteoglycin) are the major KSPGs in vertebrate corneas. We have cloned both the mouse keratocan gene and its complementary DNA (cDNA). The mouse keratocan gene spans approximately 6.5 kb of the mouse genome and contains three exons and two introns. Northern blotting and in situ hybridization were employed to examine keratocan gene expression during mouse development. Unlike lumican gene, which is expressed by many tissues other than cornea, keratocan messenger RNA (mRNA) is more selectively expressed in the corneal tissue of the adult mouse. During embryonic development, keratocan mRNA was first detected in periocular mesenchymal cells migrating toward developing corneas on embryonic day 13.5 (E13.5). Its expression was then gradually restricted to corneal stromal cells in between E14.5 and E18.5. Interestingly, keratocan mRNA can be detected in scleral cells of E15.5 embryos, but not in E18.5 embryos. In adult eyes, keratocan mRNA can be detected in corneal keratocytes, but not in scleral cells. To identify and characterize a keratocyte-specific promoter, we have cloned a 3.2-kb genomic DNA fragment 50 of the mouse Kera gene, which has promoter activities in driving the expression of reporter genes, for example, b-galactosidase (b-Gal), reverse tetracycline transcription activator (rtTA), Cre recombinase (Cre), by migrating periocular mesenchymal cells of neural crest origin during embryonic development and by cornea stromal keratocytes in adults. For example, in adult Kera-bGal transgenic mice, b-galactosidase activity was detected

318

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

only in cornea, not in other tissues (e.g., lens, retina, sclera, lung, heart, liver, diaphragm, kidney, and brain). In contrast, during ocular development, the spatial–temporal expression patterns of bGal reporter gene recapitulated that of endogenous Kera expression in mice. Using X-Gal staining, strong b-galactosidase activity was first detected in periocular tissues of E13.5 embryos, and restricted to corneal keratocytes at E14.5 and thereafter. Interestingly, in addition to cornea, b-galactosidase activity was transiently found in some nonocular tissues, that is, ears, snout, and limbs of embryos of E13.5 and E14.5, but was no longer detected in those tissues of E16.5 embryos. The transient expression of endogenous keratocan in nonocular tissues during embryonic development was confirmed by in situ hybridization. Taken together, the observations suggest that the 3.2-kb Kera promoter contains sufficient cis-regulatory elements to drive heterologous minigene expression in cells expressing keratocan.

Corneal-epithelium-specific promoter

Keratins are a group of water-insoluble proteins that form 10-nm intermediate filaments in all epithelial cells. Approximately 30 different keratin molecules have been identified, which can be divided into acidic and basic neutral subfamilies. In vivo, a basic keratin is usually coexpressed and paired with a particular acidic keratin. The expression of keratin pairs is tissue specific, differentiation dependent, and developmentally regulated. Expression of keratin 3/keratin 12 pair has been found in human, bovine, guinea pig, rabbit, and chicken corneas and is regarded as a marker for corneal-type epithelial differentiation. The expression of keratin 12 is restricted to the corneal epithelium. We have identified and cloned the corneal-epithelium-specific K12 keratin (Krt12) gene. Using gene gun to deliver Krt12-LacZ reporter gene constructs to rabbit corneal epithelium, we have identified the cis-regulatory element that is sufficient and necessary for corneal-epithelium-specific expression of the reporter genes. However, the use of conventional transgenesis techniques failed to identify a functional Krt12 promoter that was capable of driving the expression of reporter genes, for example, b-Gal, chloramphenicol acetyl transferase (CAT), and green fluorescent proteins (GFPs) in corneal epithelium. Transgenesis using lentiviral Krt12lacZ vectors had successfully generated mouse lines that expressed the reporter LacZ gene by corneal epithelium, but multiple insertions by the use of lentivirus vector compromised the efficiency of obtaining stable transgenic mice. To overcome the pitfalls, a gene-targeting construct containing an internal ribosomal entry site–reverse tetracycline transcription activator (IRES-rtTA) cassette was inserted into the Krt12 allele to produce knock-in Krt12-rtTA and Krt12-Cre driver mouse lines through gene-targeting techniques (see below).

Ocular-surface epithelium-specific Pax6 promoter (Pax6OS)

Pax6 is a regulatory gene with restricted expression and essential functions in the developing eye and pancreas and distinct domains of the central nervous system (CNS). Three conserved transcription start sites (P0, P1, and a) have been identified in the murine Pax6 locus. Furthermore, the use of transgenic mouse technology has identified the cis-regulatory elements controlling the tissue-specific expression of Pax6. Specifically, a 107-bp enhancer and a 1.1-kb sequence within the 4.6-kb untranslated region upstream of exon 0 are required to mediate Pax6 expression in the lens, cornea, lacrimal gland, conjunctiva, or pancreas. Another 530-bp enhancer fragment located downstream of the Pax6 translational start site is required for expression in the neural retina, the pigment layer of the retina, and the iris. Finally, a 5-kb fragment located between the promoters P0 and P1 can mediate expression into the dorsal telencephalon, the hindbrain, and the spinal cord. The identified Pax6/cisessential elements are highly conserved in pufferfish, mouse, and human DNA and contain binding sites for several transcription factors indicative of the cascade of control events. Corresponding regulatory elements from pufferfish are able to mimic the reporter expression in transgenic mice. Thus, the results indicate a structural and functional conservation of the Pax6 regulatory elements in the vertebrate genome. This segment of the mouse Pax-6 gene 50 flanking region is necessary and sufficient for reporter construct expression in components of the eye derived from non-neural ectoderm, for example, lens epithelium and ocular surface epithelium. This transcriptional control element has been used to prepare Pax6OSrtTA (P6R) driver mouse lines for transgene expression.

Ocular Surface Tissue-Specific Driver Mouse Lines Corneal stroma-specific mouse lines The mouse keratocan gene (Kera) expression tracks the corneal morphogenesis during eye development and becomes restricted to keratocytes of the adult, implicating a cornea-specific gene regulation of the mouse Kera. We have identified and cloned a 3.2-kb genomic DNA fragment 50 of the mouse Kera gene, which is capable of driving the expression of LacZ reporter gene that recapitulates the expression patterns of Kera. The keratocan promoter has been used to create transgenic mice, for example, Kera-biglycan, and Kera-Cre and Kera-rtTA driver mouse lines. The Kera-rtTA (KR) mice can be used to create bitransgenic Kera-rtTA/tet-O-reporter construct and tritransgenic KR/TC (tet-O-Cre)/Xf/f (X, gene of interest) mice for keratocyte-specific doxycycline-inducible transgene expression and gene ablation (Table 2).

Knock-Out Mice Models: Cornea, Conjunctiva, Eyelids and Lacrimal Gland

Corneal-epithelium-specific mouse lines

3

45 6

7

XmnI EcoRI

StuI

XbaI

XbaI

BamHI

EcoRI

XbaI

2

StuI XmnI

c-myc) are bicistronic in that they have IRESs in the mRNA that allows a second initiation of translation after the stop codon of the first reading frame for the synthesis of a second protein from the mRNA. Inclusion of such IRES elements in the reporter gene constructs has allowed the generation of transgenic mouse lines that express two proteins encoded by a single transgene. To create corneal-epithelium-specific gene ablation in mice, a targeting construct containing intron 2 to exon 8 of Krt12 gene was prepared in which an IRES-Cre and the phosphoglycerate kinase-neomyocin resistance gene (PGK-Neo) reporter genes were inserted right after the stop codon within exon 8 as shown in Figure 1. Germ line chimera mice were obtained through conventional gene-targeting techniques using embryonic stem cells. The K12-Cre mice were crossed with reporter

A similar strategy employing Cre–LoxP and tet-ON systems with keratin 12 promoter (Krt12) would allow us to prepare experimental mouse lines that express transgenes and ablate genes in a corneal-epithelium-specific manner. Thus, for many years we tried to identify and isolate a functional keratin 12 promoter for the preparation of corneal-epithelium-specific transgenic and Cre–LoxP mouse lines without success. To circumvent these difficulties, we have prepared mouse lines carrying IRES-Cre and IRES-rtTA reporter genes into the Krt12 locus through a targeted knock-in strategy. Unlike prokaryotes, most mammalian mRNAs are monocistronic (i.e., one message encodes one protein). However, some viral mRNAs and/or translationally regulated mRNAs (e.g., fibroblast growth factor 2 (FGF-2) and

1

319

8

EcoRI

(a)

IRES-Cre or Pgk-neo/pA rtTA

1

2

3

45 6

7

XmnI EcoRI

StuI

Pgk-DTA

XmnI

EcoRI

XbaI

XbaI

BamHI

XbaI

EcoRI

(b)

Δ8 IRES-Cre or Pgk-neo/pA rtTA

(c)

5 probe

3 probe

IRES

(d)

K12

rtTA or Cre

Figure 1 Generation of Krt12rtTA/+ and Krt12Cre/+ knock-in mice through gene targeting. (a) Krt12 allele. (b) Targeting vector: the IRESrtTA (and IRES-Cre) cassette, containing IRES, rtTA (Cre), and SV40-polyA, was cloned in-frame into the corresponding EcoRI/EcoRV site of pKrt12-4.8 30 to the stop codon in exon 8 of Krt12, creating a modified exon 8 that contains the entire 30 coding region of Krt12, IRES-rtTA, and SV4-polyA signals. This was followed by pgk-Neo minigene, untranslated exon 8, and polyadenylation (pA) of Krt12. Finally, a negative selection marker gene, diphtheria toxin A fragment (pgkpr-DTA) cassette, was placed on the 50 end of the targeting vector. Knock-in’ shows the predicted structure of a targeted knock-in allele after homologous recombination. (c) Knock-in Krt12 allele through homologous recombination. (d) Bi-cistronic mRNA derived from the Knock-in Krt12 allele. Panels (a), (b), and (c) of this figure are reprinted from Chikama, T. -L., Hayashi, Y., Liu, C. -Y. et al. (2005). Characterization of tetracycline inducible Krt12rtTA/þ/ tet-O-LacZ mice. Investigative Ophthalmology and Visual Science 46: 1966–1972.

320

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Tg(CAG-Bgeo/ALPP)1Lbe (ZAP) mice that harbor a transgene containing a chicken b-actin promoter, a floxed LacZ gene (two LoxP elements flanked at 50 and 30 ends of LacZ), and followed by alkaline phosphatase (AP) gene in tandem. The Cre activity was assessed by the detection of LacZ and expression of AP in corneas of the offspring as shown in Figure 2. Thus, the K12-Cre knock-in mice can be used to create mouse lines in which any floxed genes are ablated in corneal epithelium. Thus, it will allow us to investigate the role of a gene in corneal morphogenesis and homeostasis through the loss of function without threatening the life of the experimental animals. Using a similar strategy, a gene-targeting construct containing an IRES-rtTA cassette was inserted into the Krt12 allele to

produce a knock-in Krt12rtTA/w mouse line through genetargeting techniques (Figure 1). The Krt12rtTA/w knock-in mice were bred with tet-O-LacZ reporter mice to obtain Krt12rtTA/w/tet-O-LacZ bitransgenic mice. The cornealepithelium-specific expression of the LacZ gene was induced in bitransgenic mice by administration of doxycycline in the drinking water and chow as shown in Figure 3. To further expand the usefulness of Krt12-rtTA knock-in mice, a tritransgenic Krt12rtTA/w/tet-O-Cre/ZEG mouse line was prepared in which the expression of enhanced green fluorescence protein (EGFP) was activated by feeding mice doxycycline that subsequently excised the LacZ gene from the ZEG allele and allowed the expression of EGFP. Figure 4 shows the doxycycline-induced

(a)

(b)

(c)

(d)

(e)

(f) Cre/Cre

Figure 2 Histograms of X-gal and AP staining of Krt12 /ZAP bitransgenic mice. Corneas from bitransgenic mice at different ages were subjected to histochemistry staining for X-gal and alkaline phosphatase activities. (a) P15 (postnatal day 15); (b) P30; (c) P60; (d) P90; (e), P180; (f) P300. Cells express K12 and AP (alkaline phosphatase) positive were stained red, whereas the K12 negative cells were LacZ positive and stained blue. At P15, the expression of LacZ (blue) and AP (red) shows a mosaic pattern. At P90, almost all central cornea epithelia express K12 and stained red with sporadic blue cells. At P180 and P300, central corneas were stained red with blue cells located at limbus. Reproduced from Tanifuji-Terai, N., Terai, K., Hayashi, Y., et al. (2006). Expression of keratin 12 and maturation of corneal epithelium during development and postnatal growth. Investigative Ophthalmology and Visual Science 47: 545–551.

Knock-Out Mice Models: Cornea, Conjunctiva, Eyelids and Lacrimal Gland

BTg: No Dox

BTg: Dox 1D

BTg: Dox 2D

BTg: Dox 7D

BTg: Dox 14D

BTg: Dox 14D on, 14D off

BTg: Dox 14D on, 28D off

LacZ: Dox 14D

321

100 μm

Figure 3 In situ analysis of b-galactosidase enzyme activity induction by doxycycline in corneas of Krt12rtTA/+/tet-O-LacZ bitransgenic mice. Stereomicroscopy showed a side view of each eye after whole-mount b-galactosidase staining. Histological examination (lower image of each panel) of the same samples revealed that the b-galactosidase expression was restricted to corneal epithelium. Corneal epithelial cells began to express b-galactosidase in 24 h after doxycycline administration. The number of b-galactosidase expressing cells in BTg mouse with doxycycline increased in the course of time. Not all corneal epithelial cells expressed b-galactosidase was observed even in the maximum level. BTg, bitransgenic; LacZ, tet-O-LacZ single transgenic. Reproduced from Chikama, T. -I., Hayashi, Y., Liu, C. -Y. et al. (2005). Characterization of tetracycline Inducible Krt12rtTA/þ/tet-O-LacZ mice. Investigative Ophthalmology and Visual Science 46: 1966–1972.

expression of EGFP by corneal epithelial cells of Krt12rtTA/w/TC/ZEG mice. ZEG is a transgenic mouse line that has dual reporter genes of a LacZ flanked by LoxP, followed by EGFP driven by a chicken actin promoter. Chicken actin promoter drives the expression of LacZ in all cells except those of which express Cre and show green fluorescence due to the excision of LacZ and expression of EGFP. Using the strategies, we have prepared several mouse lines that can be used to overexpress transgenes and ablate genes of interest in a cornea-specific manner as summarized in Table 2. Ocular surface tissue-specific Pax6OS-rtTA mouse line

We have recently prepared an ocular surface epitheliumrtTA (Pax6OS-rtTA) driver mouse line with cis-regulatory elements of the 4.6-kb un-translated region upstream of exon 0 of Pax6, which mediate Pax6 expression in the lens, cornea, lacrimal gland, conjunctiva, or pancreas. The

bitransgenic Pax6OS-rtTA/Tet-O-EGFP mice express EGFP upon doxycycline induction (data not shown). This Pax6OS-rtTA mouse line is useful for generating experimental mice of overexpression of tet-O-transgene and ablation of floxed gene of tritransgenic (Pax6OS-rtTA/ tet-O-Cre/X f/f) mice.

Roles of Growth Factors on Ocular Surface Tissue Morphogenesis during Development and Wound Healing Elucidated from Transgenic Mice Corneal morphogenesis during eye development of vertebrates involves the differentiation of cells from surface ectoderm and the migration of periocular mesenchymal cells of neural crest origin. The differentiation of surface ectoderm gives rise to corneal and conjunctival epithelia of the ocular surface as well as to glandular epithelium,

322

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

ZEG

K12 rtTA/w /tet-O-Cre/ZEG

tet-O-Cre/ZEG

+Dox

Day 0

1⬘⬘

1⬘⬘

1⬘⬘

5.5⬘⬘

5.5⬘⬘

5.5⬘⬘

1⬘⬘

1⬘⬘

1⬘⬘

0.3⬘⬘

0.2⬘⬘

0.

1⬘⬘

1⬘⬘

1⬘⬘

Day 2

Day 4

Day 7

Day 14

−Dox

Figure 4 Expression of floxed reporter gene by doxycycline induction Tritransgenic Krt12/tet-O-Cre/ZEG mice were fed doxycycline chow for various periods of time. The experimental animals were examined under a ZEISS stereomicrscope with epi-fluorescence attachment. Strong green fluorescence was observed in tritransgenic Krt12/tet-O-Cre/ZEG mice after 2 days induction and then declined in 14 days after removal of doxycycline in the diet.

for example, lacrimal and meibomian glands. The mesenchymal cells of neural crest origin become corneal endothelial cells and keratocytes, and the stromal cells of other ocular surface tissues, that is, eyelids, iris, ciliary body, and trabecular meshworks. The corneal epithelium synthesizes components of extracellular matrix (ECM) for the formation of primary stroma when the lens detaches from the ectoderm during development. In vertebrate corneal development, the first wave of mesenchymal cells that migrate underneath the primary stroma forms the endothelium. The mesenchymal cells of the second wave invade the primary stroma and become keratocytes, which are responsible for

the formation of secondary stroma of adult vertebrates. The third waves of mesenchymal cells contribute to the stromas of eyelids, iris, ciliary body, and trabecular meshworks. This orderly cellular migration and differentiation are controlled by cues from various cytokines and the components of ECM, which are under constant remodeling during embryonic development. For example, members of the transforming growth factor beta (TGFb) superfamily play pivotal roles in embryonic development. However, the precise cytokines and their functions that modulate corneal morphogenesis remain unknown. It is very likely that cytokine signaling may modulate the expression of specific transcription factors that contribute

Knock-Out Mice Models: Cornea, Conjunctiva, Eyelids and Lacrimal Gland

to this orderly corneal morphogenesis during development or vice versa. TGFb Receptor Signaling Pathways during Development and Corneal Wound Healing Role of TGFb2 on development

TGFb has a pivotal role in embryonic development. In mammals, three isoforms of TGFb (b1, -2, and -3) are known. Members of TGFb family are multifunctional cytokines involved in development, tissue repair, and other physiological or pathologic processes. Among the knock-out mice of the three Tgfb isoforms, only Tgfb2–/– mice exhibit ocular pathology of thin corneal stroma, absence of corneal endothelium, fusion of cornea to lens, a phenotype resembling Peter’s and Axenfeld anomaly in humans, and accumulation of hyaline cells in vitreous. Delayed appearance of macrophages in ocular tissues was observed in Tgfb2–/–mice. Malfunctioning macrophages may account for accumulation of cell mass in vitreous of Tgfb2-null mice. InTgfb2–/–mice, fewer keratocytes were found in stroma that have a decreased accumulation of ECM; for example, lumican, keratocan, and collagen I were greatly diminished. The thinner stroma resulting from decreased ECM synthesis may account for the decreased cell number in the stroma of Tgfb2-null mice. The absence of TGFb2 did not compromise corneal epithelial cell proliferation, nor enhance apoptosis. Keratin 12 expression was not altered in Tgfb2–/–mice, implicating that TGFb signaling is not essential for cornea-type epithelium differentiation. This suggestion is further supported by the observation that ablation of TGFb type II receptor in Krt12Cre/Cre/Tbr2 f/f mice does not cause corneal epithelium anomaly (our unpublished observation). Role of TGF-b signaling in wound healing of corneal epithelium

Corneal epithelial defects must be rapidly resurfaced to avoid microbial infection and further damage to the underlying stroma. Epithelial healing is achieved by migration of the epithelial sheet to cover the denuded surface and enhanced cell proliferation to reestablish the epithelial stratification quickly after resurfacing. It is of interest to note that in the early phase of healing only one of the two cellular responses, cell migration, takes place, whereas cell proliferation is suppressed. Although cell migration promotes rapid re-epithelialization, the cessation of cell proliferation may impede healing if such cessation is prolonged. Various growth factors, including TGFb, orchestrate the behavior of healing corneal epithelium: for example, cell migration and/or proliferation, cell death, and protein synthesis. It has been demonstrated that the TGFb isoforms and their receptors are present in corneal and limbal epithelia and other supporting tissues (e.g., conjunctiva and tear fluid). Therefore, it has long been

323

speculated that the TGFb isoforms play pivotal roles in maintaining corneal homeostasis in a paracrine and autocrine fashion as TGFb inhibits cell proliferation of cultured keratinocytes and corneal epithelial cells in vitro, thus it may suppress corneal epithelial cell proliferation in vivo. This notion is further supported by the observation in which the administration of anti-TGFb-neutralizing antibodies reduces scar tissue formation in injured corneas. Recently, it was shown that epithelial debridement causes an upregulation of TGFb receptor expression on migrating corneal epithelial cells, suggesting that this ligand may have a pivotal role in modulation of functions of migrating corneal epithelial cells during wound healing. We recently examined the roles of TGF-b signaling pathways in regulating cell migration and proliferation of the healing of corneal epithelium debridement. TGF-b type II receptor (Tbr2) floxed mice were bred with Krt12-Cre mice to generate bitransgenic mice in which the Tbr2 gene was disrupted selectively in the corneal epithelial cells. Corneal epithelial debridement (2 mm in diameter) was created in 2-month-old bitransgenic Krt12Cre/Cre/Tbr2 f/f mice and their littermates as controls Krt12Cre/Cre/Tbr2f/w and Krt12Cre/Cre/Tbr2w/w. Our results indicated that corneal epithelium of Krt12Cre/ Cre /Tbr2 f/f mice exhibited delayed healing of debridement in comparison to that of control littermates that were heterozygous floxed and wild-type Tbr2. The naive uninjured corneal epithelium of Krt12Cre/Cre/Tbr2 f/f mice exhibited higher cell proliferative activities than controls as determined by BrdU incorporation. It is of interest to note that corneal epithelium debridement caused cessation of epithelial cell proliferation of all experimental mice in 6–12 h, irrespective of whether the Tbr2 was ablated or not. Immunohistochemistry using anti-phospho-p38 mitogen-activated protein kinase (MAPK) revealed, following epithelium debridement, that the activation of p38MAPK was seen in 6 h of injury in control mice. In contrast, phosphorylation and nuclear translocation of p38MAPK were markedly delayed in mice lacking Tbr2 in corneal epithelium compared to control mice. The observation is consistent with results of our previous studies, in which we demonstrated that addition of p38MAPK inhibitors blocked cell migration more markedly than neutralizing anti-TGFb antibody and enhanced cell proliferation in the injured corneal epithelium. The observation suggested that p38MAPK, but not the mothers against decapentaplegic (Smad, another signaling cascade active by TGFb) cascade, plays a major role in promoting cell migration and in suppressing cell proliferation in migrating epithelium. Role of FGF7 in Maintenance of Corneal Homeostasis During mammalian embryogenesis, epithelial– mesenchymal interactions play a determining role in

324

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

normal tissue patterning and development. FGF7 (also known as keratinocyte growth factor, KGF), a member of the FGF family, is a mesenchymally derived mitogen for epithelial cells in regulating epithelial cell behavior, as the FGF7 receptor is expressed by epithelial cells. Overexpression of human FGF7 by a crystalline promoter in the eye caused hyperproliferation of embryonic corneal epithelial cells and their subsequent differentiation into functional lacrimal gland-like tissues. This indicates that stimulation of the FGF7 receptor early in development, in surface ectoderm normally destined to form corneal epithelium, is sufficient to alter the fate of these cells. This further suggests that the correct spatial and temporal expression of FGFs plays a critical role in normal lacrimal gland induction. It is of interest to note that overexpression of FGF7 in Krt12rtTA/wt/tetO-FGF7 bitransgenic mice by doxycycline induction during embryonic development resulted in the formation of vascularized cornea with epithelium hyperplasia, resembling human ocular surface squamous neoplasia (OSSN) as shown in Figure 5. The phenotype variations of the two mouse models can be explained by the fact that a-crystalline expression by lens commences at E10–E11.5, whereas the expression of keratin by corneal epithelium begins at E14.5. EGFR/EGF, TGFa Signaling Pathways on Eyelids Morphogenesis Transgenic Kera-Bgn mice overexpressing biglycan, driven by keratocan promoter under the keratocan promoter, exhibit exposure keratitis and premature eye opening from noninfectious eyelid ulceration due to perturbation

of eyelid muscle formation and the failure of meibomiangland formation. In addition, in vitro analysis revealed that biglycan binds to TGFa, thus interrupting epidermal growth factor receptor (EGFR) signaling pathways essential for mesenchymal cell migration induced by eyelid epithelium. The defects of TGFa signaling by excess biglycan were further augmented by the interruption of the autocrine or paracrine loop of the EGFR signaling pathway of heparin-binding (HB)-EGF expression elicited by TGFa.10 These results are consistent with the notion that under physiological conditions, biglycan secreted by mesenchymal cells serves as a regulatory molecule for the formation of a TGFa gradient serving as a morphogen of eyelid morphogenesis (Figure 6). MEK kinase 1 (MEKK1) is an MAPK originally identified as an upstream activator for several MAPK pathways. During mouse embryogenesis, MEKK1 controls cell shape changes and formation of actin stress fibers that are required for sealing epidermis in the embryos in a process known as eyelid closure. MEKK1-null mice display eyeopen at birth (EOB), a phenotype found also in mice impaired in activin, a subgroup of the TGFb family, or in EGFR or its ligand TGFa, or in transcription factor c-Jun. Molecular analyses have revealed at least two signaling mechanisms in the control of eyelid closure. One is originated from the activins and is transduced through MEKK1, leading to transcription-independent actin stress fiber formation and transcription-dependent keratinocyte migration. Another is the TGFa/EGFR signal that is transduced through an MEKK1-independent pathway to the activation of the extracellular signal-regulated kinase (ERK) MAPK, which also leads to keratinocyte migration.

100 μm

Epithelial neoplasia with mesenchymal invasion Figure 5 Excess expression of FGF7 by corneal epithelium causing ocular surface squamous neoplasia (OSSN). Krt12-rtTA/tet-OFGF7 bitransgenic mice were obtained by crossbreeding single Krt12-rtTA and tet-O-FGF7 mice. The progeny were induced by feeding mothers doxycycline chow at the beginning of mating (E0.5) and until weaning of the pups at postnatal day 18 (P18). The pups were continuously fed doxycycline chow after weaning until sacrifice at P21. Enucleated eyes were subjected to histological examination. The bitransgenic Krt12-rtTA/tet-O-FGF7 mice showed bilateral OSSN.

Knock-Out Mice Models: Cornea, Conjunctiva, Eyelids and Lacrimal Gland

Wt epithelium

Tg epithelium

EGFR

TGFα

P

EGFR

TGFα

P

P

Cell migration P

325

HB-EGF

Cell migration

Biglycan P

EGFR

EGFR cJun P

P

HB-EGF

cJun P

cJun AP-1 target gene Mesenchymal cell

cJun AP-1 target gene Mesenchymal cell

Figure 6 The EGFR signaling mediated by TGF-a during eyelid morphogenesis. The presence of excess biglycan in the Kera-Bgn transgenic mice sequesters TGF-a and consequently perturbs the autocrine and/or paracrine loop of EGFR signaling pathways through HB-EGF and impairs mesenchymal cell migration. Reproduced from Hayashi, Y., Liu, C. -Y., Jester, J. V. et al. (2005). Excess biglycan interferes TGF-a signaling required for eyelid morphogenesis. Developmental Biology 277: 222–234.

c-Jun might serve as a connection between the two pathways. As embryonic eyelid closure is a specific morphogenetic process that is easily detectable, genetic mutant mice with EOB will be ideal models to understand the signaling mechanisms in the control of epithelial cell migration and the morphogenetic process of epithelial sheet movement. Thus, the observation supports the hypothesis that tissue morphogenesis during development is regulated by growth factors and cytokines, and is characterized by constant remodeling of ECM in response to signaling molecules, for example, growth factors, cytokines, and so forth. Proteoglycans that bind growth factors are potential regulators of tissue morphogenesis during embryonic development.

Conclusion: The Clinical Relevance of Tet-ON Mouse Models in Elucidating Pathophysiology of Ocular Surfaces Diseases Many transgenic and knock-out mice exhibit pathogenesis resembling human ocular surface diseases. Thus, the clinical manifestations of mouse lines can be used as clues for identifying inherited human disease of unknown etiology. However, embryonic lethality and congenital defects of the mouse lines do not allow further examination of the effects of altered genetic functions on pathophysiology of acquired diseases in adults. The difficulties can be overcome by preparing mouse lines of inducible transgene

expression, tissue-specific gene ablation, and inducible tissue-specific gene ablation. The conditional transgenic mouse lines will live normally until the administration of doxycycline, which induces expression of the transgene and/or ablation of gene of interest. Use of these genetically modified mouse lines can simulate the pathophysiology of ocular surface diseases, for example, wound healing, tumorigenesis, and irregular hormone and cytokine signaling that offsets homeostasis in adults.

Acknowledgments This work was supported by NIH grants EY 10556, EY 11845, and EY 13755, Challenge Grant for Research to Prevent Blindness, Inc., and Unrestricted grant from Ohio Lion Eye Research Foundation. See also: Conjunctival Goblet Cells; Cornea Overview; Corneal Epithelium: Cell Biology and Basic Science; Corneal Epithelium: Wound Healing Junctions, Attachment to Stroma Receptors, Matrix Metalloproteinases, Intracellular Communications; Gene Therapy for the Cornea, Conjunctiva, and Lacrimal Gland; Lacrimal Gland Hormone Regulation; Lacrimal Gland Overview; Lacrimal Gland Signaling: Neural; Lids: Anatomy, Pathophysiology, Mucocutaneous Junction; Overview of Electrolyte and Fluid Transport Across the Conjunctiva; The Surgical Treatment for Corneal Epithelial Stem Cell Deficiency, Corneal Epithelial Defect, and Peripheral Corneal Ulcer.

326

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Further Reading Camper, S. A., Saunders, T. L., Kendall, S. K., et al. (1995). Implementing transgenic and embryonic stem cell technology to study gene expression, cell–cell interactions and gene function. Biology of Reproduction 52: 246–257. Chikama, T., Hayashi, Y., Liu, C. Y., et al. (2005). Characterization of tetracycline-inducible bitransgenic Krt12rtTA/+/tet-O-LacZ Mice. Investigative Ophthalmology and Visual Science 46: 1966–1972. Funderburgh, J. L., Corpuz, L. M., Roth, M. R., et al. (1997). Mimecan, the 25-kDa corneal keratan sulfate proteoglycan, is a product of the gene producing osteoglycin. Journal of Biological Chemistry 272: 28089–28095. Hanks, M., Wurst, W., Anson-Cartwright, L., Auerbach, A. B., and Joyner, A. L. (1995). Rescue of the En-1 mutant phenotype by replacement of En-1 with En-2. Science 269: 679–682. Hayashi, Y., Liu, C. Y., Jester, J. J., et al. (2005). Excess biglycan causes eyelid malformation by perturbing muscle development and TGF-alpha signaling. Developmental Biology 277: 222–234. Kao, W. W. (2006). Ocular surface tissue morphogenesis in normal and disease states revealed by genetically modified mice. Cornea 25 (supplement 1): S7–S19. Kao, W. W. and Liu, C.-Y. (2003). The use of transgenic and knock-out mice in the investigation of ocular surface cell biology. The Ocular Surface 1: 5–19.

Kao, W. W., Xia, Y., Liu, C. Y., and Saika, S. (2008). Signaling pathways in morphogenesis of cornea and eyelid. The Ocular Surface 6: 9–23. Liu, C. Y., Shiraishi, A., Kao, C. W., et al. (1998). The cloning of mouse keratocan cDNA and genomic DNA and the characterization of its expression during eye development. Journal of Biological Chemistry 273: 22584–22588. Muller, U. (1999). Ten years of gene targeting: Targeted mouse mutants, from vector design to phenotype analysis. Mechanisms of Development 82: 3–21. Saika, S., Okada, Y., Miyamoto, T., et al. (2004). Role of p38 MAP kinase in regulation of cell migration and proliferation in healing corneal epithelium. Investigative Ophthalmology and Visual Science 45: 100–109. Weng, D. Y., Zhang, Y., Hayashi, Y., et al. (2008). Promiscuous recombination of LoxP alleles during gametogenesis in cornea Cre driver mice. Molecular Vision 14: 562–571. Xia, Y. and Kao, W. W. (2004). The signaling pathways in tissue morphogenesis: A lesson from mice with eye-open at birth phenotype. Biochemical Pharmacology 68: 997–1001. Zhang, L., Wang, W., Hayashi, Y., et al. (2003). A role for MEK kinase 1 in TGF-beta/activin-induced epithelium movement and embryonic eyelid closure. EMBO Journal 22: 4443–4454.

Gene Therapy for the Cornea, Conjunctiva, and Lacrimal Gland A Sharma, A Ghosh, and C Siddappa, University of Missouri–Columbia, Columbia, MO, USA R R Mohan, University of Missouri–Columbia, Columbia, MO, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Alloantigen – An antigen that is the part of an animal’s self-recognition system. The common alloantigens are the major histocompatibility complex and red blood cells. Astigmatism – The inability of the eye to focus sharp image on the retina due to the irregular shape of the cornea or lens. Capsid – The protein shell that surrounds genetic material of the virus. Corneal graft or penetrating keratoplasty – The replacement of the central diseased cornea with a healthy donor cornea. This process is also called corneal transplantation. Corneal neovascularization – The ingrowth of new blood vessels from limbal plexuses toward clear cornea. Corneal scarring or haze – The loss of corneal transparency and appearance of scar because of injury or abnormal wound healing. Corneal stroma – The tissue between the epithelium and endothelium of the cornea. It constitutes approximately 90% of corneal thickness and comprises collagens, keratocytes, and extracellular matrix. CRE-adenovirus vector – An adenovirus vector generated with the helper vector that has a packaging sequence flanked by Cre/lox-P system using human embryonic kidney 293 cells stably transfected with Cre recombinase. This results in selective deletion of packaging sequence from helper virus. Dominant-negative mutant construct – A plasmid encoding for a mutant protein that competes with wild-type (normal) protein within the same cell to inhibit its function. Enhanced green fluorescent protein (EGFP) – A 27-kDa protein isolated from jelly fish. It fluoresces green when exposed to blue light and is commonly used as a marker for gene transfer studies. Hyperopia – A term used to define eye defect in which near objects appear blurred because images are focused on the back of the retina instead of on the retina. This medical condition is also known as far- or long-sightedness. Keratitis – The inflammation in the cornea causing pain and discomfort.

Myopia – A term used to define eye defect in which distant objects appear blurred because images are focused in front of the retina instead of on the retina. It is also called near- or short-sightedness. Orthotopic graft/transplant – The transplantation or grafting of tissue at its normal position in the body. Reporter/marker gene – The gene(s) used to track and/or quantify levels of gene delivery in the cell. The common reporter genes are green fluorescent protein, beta galactosidase, etc. Serotype – A characteristic set of antigens used to distinguish closely related virus strains. Sonoporation – A technique used to produce small pores temporarily in the cell membrane using ultrasonic waves for introducing RNA, DNA, or small drugs into the cells. Vector – Virus or nonviral materials used as a vehicle for carrying genes into the cells.

Introduction Gene therapy is an attractive approach to treat ocular surface diseases and disorders (Table 1). Recent reports of improved visual function in adult patients with Leber’s congenital amaurosis with gene therapy attest the potential of this form of molecular medicine to cure eye disease and prevent blindness. The front of the eye, commonly referred as ocular surface, primarily consists of the tear film, cornea, and conjunctiva. The lids, tears, and lacrimal glands are also considered integral parts of the ocular surface, as they spread tears and protect the eye. Nearly all ocular surface diseases entail an abnormality in the cornea. Consequently, more research has been performed to develop gene therapy approaches for the cornea compared to other ocular surface constituents. This article provides an overview of ocular surface gene transfer studies performed by us and many other investigators.

Corneal Gene Therapy Methods Cornea is an attractive target for gene therapy because of its accessibility, immune-privileged status, and ability to be monitored visually. Gene transfer research in the

327

328

Genes used for gene therapy of the ocular surface

Cornea Graft rejection

Wound healing

Alkali burns

Scarring or haze

Neovascularization

Gene

Vector

Mode of transfection

Effect on resolution of disease process

Cytotoxic T-lymphocyte-associated antigen 4 (CTLA-4) CTLA-4 CTLA-4 CTLA-4

Adenovirus

Topical application

þ/

Adenovirus Adenovirus Minimalistic immunolistic immunologically defined gene expression vector Adenovirus Adenovirus Adenovirus Adenovirus Adenovirus

IV injection IP injection Gene gun

þ þ þ/

IP injection Organ culture Organ culture Organ culture IM injection

þ in ovine in rat þ in ovine in rat þ

Adeno-associated virus Plasmid

þ þ

Adenovirus Adenovirus Adenovirus

Organ culture Electroporation in the anterior chamber Topical application Topical application Topical application

Retrovirus

Topical application

þ

Retrovirus Plasmid

Topical application Intrastromal injection

þ þ

Adenovirus

Injection into anterior chamber

þ

Inducible T-cell co-stimulator (ICOS) Interleukin (IL) 10 antagonist Interleukin 12 antagonist Interleukin 4 Soluble type II transforming growth factor (TGF)b receptor Decorin Tissue plasminogen activator SMAD 7 Bone morphogenic protein 7 Peroxisome proliferator-activated receptor(PPAR)-g Herpes simplex virus (HSV) thymidine kinase Dominant-negative cyclin G1 Vascular endothelial growth factor receptor FLT-1 Vascular endothelial growth factor receptor FLT-1

þ þ þ

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Table 1

Dystrophies

Lacrimal gland Sjo¨gren syndrome Conjunctiva Wound healing

Subconjunctival injection

þ

Plasmid Plasmid Lentivirus

Intrastromal injection Intrastromal injection In vitro

þ þ þ

Electroporation Adeno-associated virus Plasmid

Organ culture Organ culture Topical application or intrastromal injection Topical application or intrastromal injection Injected into anterior chamber or intrastromal injection

þ þ þ

Plasmid

Topically, subconjunctival, IP, or IM injection

þ

IL2, 4, 10, interferon, tumor necrosis factor

Plasmid

Topical application

IL10 TNF

Adenovirus Adenovirus

Injected into lacrimal gland Injected into lacrimal gland

þ þ

Dominant-negative p38MAPK SMAD 7 PPARg

Adenovirus Adenovirus Adenovirus

Topical application Topical application Topical application

þ þ þ

IL 10

Plasmid

b-Glucuronidase

Adenovirus

HSV-1 glycoprotein D, B1

þ þ

Gene Therapy for the Cornea, Conjunctiva, and Lacrimal Gland

Disorders Increase immunity against HSV-induced ocular keratitis Herpes keratitis

Adeno-associated virus

Vascular endothelial growth factor receptor FLT-1 FLT23K FLT24K Endostatin/kringle-5 domain of plasminogen fusion protein Kringle 5 plasminogen Angiostation IL12

329

330

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

cornea primarily focuses on developing therapeutic modalities for common clinical problems such as corneal neovascularization, graft rejection, corneal scarring, and wound healing. The scope of developing novel interventional gene therapy strategies has been markedly enhanced because of an increased understanding of the molecular mechanisms and pathogenesis of corneal diseases. Multiple vectors, techniques, and strategies have been utilized to deliver foreign genes in the cornea. Employing in vitro, ex vivo, and in vivo models, the efficiency of numerous viral and plasmid vectors to deliver genes in the ocular surface tissues has been evaluated. Among viral vectors, adenovirus, adeno-associated virus (AAV), retrovirus, and lentivirus vectors were found to efficiently introduce genes in the cornea. However, none of these vectors is ideal, and each had its own pros and cons. Viral Vectors Adenovirus and retrovirus vectors provided short-term transgene expression in the mouse and rabbit cornea with moderate-to-severe inflammatory response. These vectors also transduced corneal and conjunctival epithelium efficiently in human cornea but failed to deliver genes in rat, rabbit, and sheep corneal epithelium. Interestingly, adenovirus vector showed the best transduction efficiency for corneal endothelial cells when compared to the vectors developed from equine immunodeficiency virus, lentivirus, or bacculovirus. Nonetheless, both adenovirus and retrovirus vectors are of limited use for corneal gene therapy because of their inability to transduce nondividing cells, low transduction efficiency for corneal cells, and propensity to induce immune reactions. AAV and disabled lentivirus vectors can transduce nondividing cells and provide long-term transgene expression, thus offering good alternatives for delivering genes into keratocytes and endothelium of the cornea. However, limited studies have been performed to evaluate the utility of these vectors for corneal gene therapy. We, for the first time, showed selective transgene delivery into keratocytes of the rabbit cornea in vivo with AAV2 vector using a lamellar flap technique. Figure 1 shows the levels

and locations of b-galactosidase (b-gal) reporter gene expression in the rabbit cornea. This study also revealed that direct contact of vector to stroma is critical for introducing therapeutic genes into rabbit keratocytes in vivo. Our subsequent studies with AAV serotypes 2 and 5 showed that AAV vectors provide long-term transgene expression in the mouse and rabbit stroma in vivo. The mouse eyes continued to express enhanced green fluorescent protein (EGFP) in the corneal stroma for 10 months, without showing any significant side effects (Figure 2). In recent years, several hybrid AAV vectors have been engineered using the genome of AAV serotype 2 and capsid protein of AAV serotypes 1–9 for gene therapy. The newly produced hybrid AAV vectors have shown improved trangene delivery in many tissues including the eye. However, besides the AAV2/5 vector, transduction efficiency of other newly developed hybrid AAV vectors for the cornea has not been investigated. Recently, our laboratory tested the transduction efficiency of AAV2/6, AAV2/8, and AAV2/9 vectors for the human corneal fibroblasts and epithelial cells in vitro and for mouse cornea in vivo. The results showed that the tested hybrid AAV vectors are more efficient when compared to nonhybrid AAV2 serotypes for each corneal cell type. Figure 3 demonstrates the levels of transgene delivery in human corneal epithelium (a) and human corneal fibroblast (b) delivered with AAV2/6 vector. The analysis of AAV2/6, AAV2/8, and AAV2/9 gene transfer data revealed that AAV2/6 vector is the most efficient among the three tested vectors for human corneal fibroblasts and epithelial cells in vitro. Although AAV vectors are efficient and safe for delivering genes in the cornea, unfortunately they are incapable of transporting large genes (>1.8 kb). Disabled lentivirus vectors have been used to deliver large genes to all three major cells of the human cornea in vitro and to mouse corneal endothelium in vivo. In addition, transgene delivered with lentivirus vectors showed long-term expression. Our research team noted efficient EGFP gene delivery into keratocytes of the mouse cornea in vivo with lentivirus vector that was applied to the mouse stroma for 2 min after removal of

Figure 1 AAV-mediated targeted transgene delivery into keratocytes of the rabbit cornea in vivo using lamellar flap technique. The AAV2 vector was topically applied on the stromal bed after making corneal flap with microkeratome. Significant levels of transgene delivery in the rabbit stroma at day 3 and day 10, respectively, are shown in (a) and (b).

Gene Therapy for the Cornea, Conjunctiva, and Lacrimal Gland

331

Figure 2 AAV-mediated transgene transduction into keratocytes of the mouse cornea in vivo observed at 10 months. Representative stereomicrograph of the cornea surface (a) and tissue section (b) images demonstrate detection of high levels of fluorescent GFP reporter gene expression in the mouse cornea in vivo delivered into mouse keratocytes with AAV5 vector applied topically.

Figure 3 Efficient transduction of cultured human corneal epithelial (a) and fibroblast (b) cells by the AAV2/6 hybrid vector detected 40 h after vector treatment. The transduction efficiency of AAV2/6, AAV2/8, and AAV2/9 vectors to transduce human corneal epithelial and fibroblast cells was compared. AAV2/6 vector showed the highest transduction efficiency for tested corneal cultures. Transduced cells expressing human placental alkaline phosphatase reporter gene are stained blue.

Figure 4 Lentivirus-mediated transgene delivery in the mouse cornea in vivo noted after 6 weeks of vector application. Two microliters of lentivirus titer was either topically applied for 2 min or microinjected in the cornea using a glass needle. Transgene delivery in the rabbit stroma noted at 2 weeks in the tissue sections of the cornea, shown in (a) and (b), which received vector through topical application (a) or microinjection (b) techniques.

epithelium. Figure 4 shows the quantity and distribution of transgene in the mouse cornea with lentivirus vector administered in the stroma topically or through microinjection technique. Detection of distinctly different levels and locations of transgene in the cornea supports our hypothesis that gene delivery in the cornea is regulated by both vector and vector-delivery techniques and not by the vector alone. No serious cytotoxicity or side effects have been reported with the AAV- or lentivirus-mediated gene delivery, suggesting that these vectors may be suitable

for gene therapy treatments for patients. However, safety remains a major concern for using lentivirus vectors because of their origin from human immunodeficiency virus. Nonviral Vectors Contrary to viral vectors, nonviral vectors are nontoxic, nonpathogenic, and nonimmunogenic, and can deliver large genes. Additionally, their production is simple and

332

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

cost effective. Various surgical, mechanical, electrical, or chemical approaches were used to administer plasmid DNA in the cornea in vitro, ex vivo, and in vivo. Intrastromal injection of plasmid DNA encoding reporter, vascular endothelial growth factor, or interleukin-1 receptor antagonist successfully delivered transgene into the mouse cornea. A major advantage of this method over viral technology was that it provided rapid transgene expression in the mouse cornea in vivo. Transgene expression was detected as early as 1h after administration, with a peak occurring in 12–24 h. Furthermore, the plasmid-injected corneas remained clear and free of inflammation. Plasmid applied during stromal hydration and filtration bleb surgeries delivered genes in rabbit corneas, in vivo. Techniques using electric current plasmid DNA delivery in corneal epithelium, keratocytes, and endothelium have been reported. Electrical current up to 200 V cm1 did not cause trauma, edema, or inflammation in the cornea but introduced only low levels of transgene. Higher electrical current increased transgene delivery but was associated with substantial corneal damage. Firing of DNA- or RNA-coated gold microparticles on the cornea with a gene gun introduced transgene in the corneal epithelium. However, gene-gun-mediated gene delivery was limited to epithelium and was associated with corneal damage. Lipids and polymers have been used frequently to deliver genes in corneal cells in vitro but provide inadequate levels of transgene in the cornea in vivo. Low transfection efficiency and short-term transgene expression are among the major challenges of nonviral gene therapy that need to be addressed. Use of nanoparticles as vectors has great potential for improving nonviral gene transfer.

Use of Gene Therapy to Treat Corneal Diseases Corneal Graft Rejection Many gene therapy studies for prevention of corneal graft rejection have focused on the use of cytotoxic Tlymphocyte-associated antigen-4 conjugated to human immunoglobulin G (IgG) heavy chain (CTLA-4 Ig). The rationale for this therapy is based on the fact that helper T-cell activation primarily involves the interaction between T-cell receptor/cluster of differentiation 3 (CD3) and the alloantigen/major histocompatibility complex II (MHCII) of the antigen-presenting cells (APCs). An additional, essential stimulus for T-cell activation is provided by the interaction between co-stimulatory molecules, such as APC B7 antigens, and the CD28 molecule of T cells. T-cell activation can be prevented by blocking this interaction with administration of chimeric protein, CTLA-4 Ig. Multiple studies have evaluated the efficiency of CTLA-4 Ig gene therapy to control corneal graft rejection.

The CTLA-4 Ig gene therapy was delivered to the donor rat corneas with adenovirus prior to transplantation topically or after transplantation by a single intravenous or intraperitoneal injection. Topical application of adenovirus expressing CTLA-4 Ig had a marginal effect on preventing graft rejection, whereas systemic administration markedly prolonged graft survival. In another study, minimalistic immunologically defined gene expression (MIDGE) vector encoding for CTLA-4 was delivered in the corneal epithelial cells with a gene gun 10 days after orthotopic corneal transplantation. A marginal beneficial effect of CTLA-4 expression was noted for graft survival in the tested model. However, a subsequent study of MIDGEmediated CTLA-4 gene transfer in mice demonstrated significant prolongation of graft survival when gene therapy was delivered 1 day before corneal transplantation. Contrary to the encouraging CTLA-4 gene transfer studies, the results have been disappointing when other costimulatory pathways of T-cell activation are inhibited. Adenovirus-mediated ICOS, which is an inducible costimulatory receptor expressed by activated T cells, ex vivo or systemic gene therapy did not result in a significant prolongation of corneal graft survival. Cytokines play an important role in corneal graft survival and rejection. Consistently elevated levels of proinflammatory cytokines, such as Th (T-helper cell) type 1, have been detected in the corneal tissues and aqueous humor of eyes with graft rejection. Th type 2 cytokines, such as interleukin (IL)-10 and IL-4, exert inhibitory effects on Th1 cytokines. The effects of elevation of anti-inflammatory Th type 2 cytokines or inhibition of proinflammatory Th type 1 cytokines on graft survival have been the focus of many gene therapy studies. Significantly increased corneal graft survival, from 18–20 days to 45–55 days, was observed in ovine donor corneas transduced with adenovirus-expressing IL-10 or IL-12 antagonist. On the other hand, adenovirus-mediated IL-10 or IL-12 antagonist gene therapy did not prevent corneal transplant rejection in a rat model. This disparity in results could be due to differences in the kinetics of graft rejection in the two animal species. Interestingly, IL-4 gene therapy was found to be ineffective in both the species. Corneal Wound Healing Wound healing plays important role in maintaining corneal transparency and normal visual function. Injury to the cornea is known to trigger unregulated woundhealing response that can lead to scar formation and loss of vision. Multiple growth factors and cytokines have been shown to regulate wound healing in the cornea. Out of many cytokines influencing corneal wound healing, transforming growth factor beta (TGFb) has been shown to play a central role in corneal wound healing, myofibroblast generation, and haze development. Soluble

Gene Therapy for the Cornea, Conjunctiva, and Lacrimal Gland

type II TGFb receptor binds TGFb and blocks its biological activity. Soluble type II TGFb receptor gene therapy has been shown to reduce corneal opacification in corneal injury model. A single intramuscular injection of recombinant adenovirus encoding for soluble type II TGFb receptor in the mouse eye prevented corneal edema, angiogenesis, inflammatory cell infiltration, and deposition of extracellular matrix in the cornea. High levels of soluble TGFb receptor were detected in the serum and corneal fluid of treated animals for 10 days but several side effects were also observed. This study demonstrated the potential application of gene therapy to treat corneal disorders but several issues such as safety, immunological reaction, etc., associated with adenovirusbased gene therapy still need to be resolved. AAV-mediated gene therapy may address some of these issues. Our laboratory, for the first time, demonstrated efficient transgene delivery into keratocytes of the normal and diseased (hazy and neovascularized) rabbit corneas in vivo with AAV2 and AAV5 vectors. The tested AAV vectors showed high levels and long-term transgene expression (over 4 months) in the rabbit corneas. Our ongoing gene transfer experiments include evaluation of AAVmediated decorin gene therapy to control corneal scarring and angiogenesis. Decorin is a small leucine-rich proteoglycan and blocks biological activity of TGFb. Our in vitro studies revealed that decorin gene delivery into rabbit and human keratocytes significantly reduces TGFb-driven transdifferentiation of keratocyte to myofibroblasts. This transformation is believed to cause corneal haze in vivo. Employing nonviral gene transfer approach, the role of fibrin deposition on corneal transparency was also studied in laser-induced fibrin clot formation model in vivo. Administration of plasmid-DNA-expressing tissue plasminogen activator gene through electroporation in the anterior chamber of the eye markedly reduced extracellular matrix deposition in treated eyes. The beneficial effects lasted for 4 days.

Corneal Alkali Burn Alkali burn injury to the cornea is a serious clinical problem that often leads to permanent visual loss due to ulceration, scarring, and/or neovascularization. A gene therapy approach has been used to treat corneal alkali burn by targeting modulation of TGFb super family genes using adenovirus vectors. Mothers against decapentaplegic homolog 7 (Smad7) is well documented to inhibit TGFb signaling. Topical application of Cre-adenovirus encoding for Smad7 effectively prevented alkali-induced corneal scaring. Bone morphogenic protein-7 (BMP7) is another member of TGFb super family and has been shown to antagonize the effects of TGFb. Adenovirusmediated BMP7gene transfer in the cornea has been

333

shown to accelerate re-epithelilization of corneal surface and suppress myofibroblast generation, monocytes/macrophages infiltration and macrophage chemoattractant protein-1 (MCP-1), and TGFb and collagen expression in the corneal stroma in alkali-induced corneal injury mouse model. However, BMP7 gene therapy was ineffective to suppress stromal neovascularization. Another study reported that topical application of Cre-adenoviral vector encoding for peroxisome proliferator-activated receptor-gamma (PPARg) reduced inflammatory and fibrogenic responses in the alkali burn mouse cornea. Adenovirus-mediated PPARg overexpression inhibited growth factors and upregulation of matrix metalloproteinases, monocytes/macrophages infiltration, and myofibroblasts production in healing cornea in vivo. Further, accelerated re-epithelization and basement membrane reconstruction was observed in mouse cornea overexpressing PPARg gene.

Corneal Scarring or Haze Corneal scarring is primarily caused by injury or infections to the eye. It is also a common side effect of the photorefractive keratectomy (PRK) surgery, particularly in patients undergoing high myopia correction. The PRK surgery is frequently used worldwide to treat myopia, hyperopia, and astigmatism. Accumulating evidences suggest that corneal scarring occur due to abnormal wound healing, deposition of disorganized extracellular matrix, keratocyte activation, and transdifferentiation of keratocytes to fibroblasts and myofibroblasts. Gene therapy offers a unique approach for intercepting the molecular events involved in the development of corneal haze. Blockade of keratocyte proliferation is an attractive approach for controlling corneal haze in vivo. Rabbit corneas transduced with retroviral vector encoding for herpes simplex virus (HSV) thymidine kinase gene after keratectomy, followed by topical application of ganciclovir, showed significant inhibition of laser-induced corneal haze. Another approach that has been tested to control corneal haze is the blockade of cyclins and cyclin-dependent kinases that play an active role in controlling cell division. Again using a retroviral vector, a dominant-negative mutant construct for cyclin G1 was delivered in the rabbit cornea after transepithelial phototherapeutic keratectomy surgery. Analysis of corneal tissues showed that dominant-negative cyclin G1 delivery decreased extracellular matrix production and markedly reduced the development of corneal haze in tested rabbit model. Apoptosis of activated keratocytes by the delivered dominant-negative cyclin is believed to be a mechanism for the decrease in corneal haze appearance. As noted above, we have also shown the potential of decorin gene therapy to inhibit corneal haze using an in vitro model.

334

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

Corneal Neovascularization Corneal insults, such as infection, degenerative disease, chemical damage, mechanical/surgical injury, and immunologic disease, can cause corneal neovascularization (CNV) that affects about 1.4 million Americans each year. Multiple lines of evidence suggest that cytokine vascular endothelial growth factor (VEGF) plays a key role in CNV development. Various gene therapy strategies have been developed and examined to suppress VEGF-induced corneal angiogenesis using experimental models. Intrastromal injection of plasmid encoding a soluble form of the VEGF receptors Flt-1, which neutralizes VEGF, significantly inhibited CNV in the mouse eye. Similar inhibition of CNV was demonstrated by sFlt1 gene transfer with adenovirus and AAV. In a novel approach for controlling CNV with gene therapy, plasmids encoding for Flt23K or Flt24K peptide (VEGF-binding domains of sFlt-1) coupled with endoplasmic reticulum-retaining peptide have been developed. These peptides are not secreted outside the cell and therefore are capable of neutralizing intracellular VEGF. Intrastromal injection of these plasmids in mouse eye inhibited CNV by as much as 50% in the experimental models of corneal angiogenesis. Recently, these scientists reported CNV reduction up to 40% in the mouse cornea after delivering Flt23K gene with albumin nanoparticles. Gene therapy has also been evaluated in a corneal transplant rabbit model to study the effect of gene transfer of a fusion protein. These rabbit corneas were transduced ex vivo with lentiviral vector encoding for endostatin and kringle-5 domain of plasminogen fusion protein. Endostatin is an anti-angiogenic peptide derived from type XVIII collagen, and Kringle 5 domain of plasminogen is a specific inhibitor for endothelial cell proliferation derived from the proteolytic fragment of human plasminogen. The endostatin-kringle-5 fusion protein effectively inhibited allogenic transplantation-induced neovascularization. Delivering kringle-5-plasminogen gene through electroporation was also demonstrated to be effective in reducing CNV in a rat model. Angiostatin is another potent, recently identified anti-angiogenic agent. AAV vectors encoding angiostatin gene effectively reduced alkali-induced CNV in rat corneas. In another study, mouse corneas transfected with plasmid encoding for IL12 or IL10, using topical application or intrastromal techniques, showed marked suppression of CNV. Interestingly, a careful review of the literature reveals that few gene transfer studies of CNV in experimental disease models have reported, discussed, or investigated the side effects or downsides of the gene transfer methods that were investigated. Corneal Dystrophies In the past decade, several genes and gene mutations, such as BIGH3, TGFb1, gelsolin, and CHTS6 have been identified to cause granular, lattice, Avellino, and/or

Reis–Bu¨cklers corneal dystrophies. However, a paucity of experimental models for evaluating the usefulness of gene therapy to cure genetic corneal dystrophies remains a major obstacle to the development of therapeutic interventions for these vision-threatening disorders. Very few gene transfer studies using corneal buttons collected from patients or rodent models have been performed to evaluate the efficacy of gene therapy in corneal dystrophies. Gene therapy approaches can be used as a tool for creating animal models of corneal dystrophies by overexpressing aberrant proteins in the cornea. Mucopolysaccharidosis is a group of metabolic disorders characterized by deficiency of lysosomal enzymes needed to degrade glycosaminoglycans. This disorder causes corneal abnormalities, including the loss of corneal transparency. Delivering human beta-glucuronidase gene in the cornea induced significant clearance of corneal clouding in mouse model of type VII mucopolysaccharidosis. The adenovirus vector encoding human betaglucuronidase gene was injected in the anterior chamber or intrastromal region to deliver transgene in the eye. A rapid clearance of lysososmal-storage vesicles in keratocytes and clearing of the cornea were observed. Other Corneal Disorders HSV keratitis is a leading cause of infectious blindness. Developing vaccination using gene transfer approaches have been the focus of several studies. Most of these studies used plasmid DNA encoding for HSV-1 glycoproteins, such as glycoprotein (g) D, gB1, or a cocktail of glycoproteins in an effort to increase immunity against HSV-induced ocular keratitis. The plasmids were administered in test animals through either topical application, or subconjunctival, intramuscular, or intraperitoneal injection. The intramuscular route of administration has been found to confer total protection against HSV keratitis, whereas topical application and subconjunctival administration prevented stromal keratitis but not the epithelial keratitis. In another study, enhancement of gD glycoprotein was observed when it was administered in combination with IL-2. Conversely, many investigators have reported limited success in controlling herpes keratitis with gene therapy, as a topical application of naked plasmid vector encoding for cytokines, such as IL2, IL4, IL10, interferon (IFN), or tumor necrosis factor (TNF) alpha, demonstrated very little or no benefit. Nonetheless, cytokine gene therapy has been shown to inhibit corneal lesion severity if used 3 days prior to HSV infection. Clinically, such treatment may be beneficial for preventing primary HSV keratitis. Lacrimal Gland Gene Therapy Lacrimal glands produce tears and tear proteins which lubricate, and supply nutrition and protection to ocular

Gene Therapy for the Cornea, Conjunctiva, and Lacrimal Gland

surface tissues. Several autoimmue disorders such as Sjo¨gren syndrome, rheumatoid arthritis, systemic lupus erythematosus, and uveitis affect lacrimal gland function and result in dry eye syndrome. This syndrome affects about 4% people in the United States. Unlike conventional therapy that demands repeated lubricant application, gene therapy could potentially require just one treatment. Genes delivered to lacrimal glands can modulate tear composition and flow rate. Initial studies tested the efficacy of vaccinia, herpes, and adenovirus vectors for delivering genes in lacrimal gland using ex vivo rat lacrimal gland. Vaccinia viral vector showed highest transgene delivery followed by adenovirus and herpes viral vector. Histological examination of tissues revealed that vaccinia vector delivered transgene in the lacrimal duct cells and acini, whereas adenovirus vector mainly transduced myoepithelial cells surrounding the lacrimal acini. Cellular degradation response was also noted with adenovirus vector, possibly due to vector’s toxicity. Subsequent studies showed IL-10 and TNF inhibitor gene delivery in cultured rabbit lacrimal gland epithelial cells by the adenovirus. In vivo studies examined the efficacy of these cytokines-based gene therapy for Sjo¨gren syndrome, using a rabbit model of dacryoadenitis. The prophylactic adenovirus-mediated IL-10 gene delivery resulted in protection against lacrimal gland immunopathology, ocular surface disease, and decrease in tear production. On the other hand, adenovirusmediated TNF inhibitor gene delivery increased basal tear production and its stability, and reduced corneal surface defects and intensity of immune cell infiltration in lacrimal gland. Adenovirus-mediated gene delivery has also been utilized to investigate cellular and physiological functions of the lacrimal gland. Examples include investigation of androgens in the pathophysiology of Sjo¨gren syndrome and the role of dyneins, protein kinases, and cytoskeletal proteins in the secretory functions of the lacrimal gland. Besides adenovirus, not many other viral vectors have been evaluated for lacrimal gland gene delivery in vivo. Retroviral vectors expressing the E6 and E7 genes of the human papillomavirus have been employed to generate immortalized lacrimal gland epithelial cell lines for laboratory studies. Conjunctiva Gene Therapy A handful of gene transfer studies have been performed for the conjunctiva. Recently, transfection efficiency of hyaluronic acid–chitosan nanoparticles to deliver genes in human conjuctival cells has been examined. These nanoparticles were found to be nontoxic to the conjuctival cells and exhibited moderate transfection efficiency (15%) for the conjunctiva. Topical application of these nanoparticles to the rabbit eye demonstrated a successful

335

delivery of reporter gene into conjunctival epithelium. Sonoporation (application of sound waves), following subconjuctival injection in the eye, delivered considerable levels of EGFP in the rat conjunctiva. Among viral vectors, adenovirus has been most widely investigated for transgene delivery in the conjunctiva. Modulation of wound healing and prevention of scarring in the conjunctiva has been the primary focus of conjunctival gene therapy studies. As in corneal injury, TGFb has been shown to play a critical role in conjunctiva wound healing and scarring. Activation of p38 mitogen-activated protein (MAP) kinase is one of the key signaling pathways activated by TGFb. Adenovirus-mediated delivery of p38 MAPkinase suppressed myofibroblast generation and decreased messenger RNA (mRNA) expression of connective tissue growth factor (CTGF) and monocyte/ macrophage chemoattractant protein-1 (MCP-1) in the mouse model of conjunctival injury. The protective effects of Smad7 gene transfer on conjunctival fibrosis have also been reported using a mouse model. Cultured subconjunctival fibroblasts transduced with adenoviral vector expressing Smad7 inhibited type-I collagen, a smooth muscle actin, and CTGF, whereas topical application of adenoviral vector expressing Smad7 gene prevented macrophage invasion and decreased VEGF and a smooth muscle actin levels in conjunctival fibroblasts in vivo. More recently, adenovirus-mediated PPARg gene transfer was found to be protective against conjunctival fibrosis. PPARg overexpression suppressed expression of type I collagen, fibronectin, and CTGF in cultured human conjunctival fibroblasts both at mRNA or protein level. In vivo experiments showed that PPARg gene delivery significantly decreased monocyte/macrophage invasion, myofibroblast generation, and blocked upregulation of cytokines/growth factors, collagen I, and a2 mRNA in the healing conjunctiva. In summary, many studies have demonstrated the ability of gene therapy to treat ocular surface diseases and improve visual function. Numerous vectors, techniques, and strategies have been identified for establishing novel gene therapy modalities to treat ocular surface disorders. Although the proof-of-principle experiments validated the potential and promise of gene therapy for ocular surface disorders, several obstacles remain to be cleared before gene therapy clinical trials for these diseases are instituted.

Acknowledgments The work is supported by the RO1EY17294 (RRM) grant from the National Eye Institute, National Institutes of Health, Bethesda, Maryland, USA and a grant from the Research to Prevent Blindness, New York, USA.

336

Structure and Function of the Tear Film, Ocular Adnexa, Cornea and Conjunctiva in Health

See also: Conjunctival Goblet Cells; Corneal Dystrophies; Corneal Epithelium: Cell Biology and Basic Science; Corneal Epithelium: Response to Infection; The Corneal Stroma; Cornea Overview; Defense Mechanisms of Tears and Ocular Surface; Drug Delivery to Cornea and Conjunctiva: Esterase- and Protease-Directed Prodrug Design; Imaging of the Cornea; Lacrimal Gland Hormone Regulation; Lacrimal Gland Overview; Lacrimal Gland Signaling: Neural; Overview of Electrolyte and Fluid Transport Across the Conjunctiva.

Further Reading Bainbridge, J. W., Tan, M. H., and Ali, R. R. (2006). Gene therapy progress and prospects: the eye. Gene Therapy 13: 1191–1197. Buch, P. K., Bainbridge, J. W., and Ali, R. R. (2008). AAV-mediated gene therapy for retinal disorders: From mouse to man. Gene Therapy 15: 849–857.

Jun, A. S. and Larkin, D. F. (2003). Prospects for gene therapy in corneal disease. Eye 17: 906–911. Klausner, E. A., Peer, D., Chapman, R. L., Multack, R. F., and Andurkar, S. V. (2007). Corneal gene therapy. Journal of Controlled Release 124: 107–133. Mohan, R. R., Schultz, G. S., Hong, J. W., Mohan, R. R., and Wilson, S. E. (2003). Gene transfer into rabbit keratocytes using AAV and lipid-mediated plasmid DNA vectors with a lamellar flap for stromal access. Experimental Eye Research 76: 373–383. Mohan, R. R., Sharma, A., Netto, M. V., Sinha, S., and Wilson, S. E. (2005). Gene therapy in the cornea. Progress in Retinal and Eye Research 24: 537–559. Saika, S., Yamanaka, O., Sumioka, T., et al. (2008). Fibrotic disorders in the eye: Targets of gene therapy. Progress in Retinal and Eye Research 27: 177–196. Selvam, S., Thomas, P. B., Hamm-Alvarez, S. F., et al. (2006). Current status of gene delivery and gene therapy in lacrimal gland using viral vectors. Advanced Drug Delivery Reviews 58: 1243–1257. Williams, K. A., Jessup, C. F., and Coster, D. J. (2004). Gene therapy approaches to prolonging corneal allograft survival. Expert Opinion on Biological Therapy 4: 1059–1071.

III. IMMUNE REGULATION OF THE CORNEA AND CONJUNCTIVA AND ITS DYSREGULATION IN DISEASE

This page intentionally left blank

Adaptive Immune System and the Eye: Mucosal Immunity A K Mircheff, University of Southern California, Los Angeles, CA, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary CD80 and CD86 – B7 costimulatory ligands expressed by antigen presenting cells. They interact with coreceptors CD28 and CTLA4 on T cells and are necessary, but not sufficient for T-cell activation. MHC class II (major histocompatibility complex class II molecules) – These molecules typically acquire autoantigen epitopes in endosomes and antigen processing compartments of antigen presenting cells and present them to antigen receptors of CD4+ T cells. TGF-b – Pleiotropic cytokine. One critical function in the mucosal immune system is to favor IgM-to-IgA isotype class switching; another is to promote differentiation of immature dendritic cells as regulatory antigen presenting cells. It may be mitogenic, antiproliferative, or pro-apoptotic, depending on target cell or synergistic interactions with other cytokines and growth factors.

The visual system interfaces with the external world at the epithelium of the corneal surface. The corneal epithelium is part of the convoluted yet topologically continuous surface that separates the body’s milieu inte´rieur from the external world. Beyond the limbal and conjunctival epithelia, it ranges in one direction through the lacrimal excretory ducts and network of ducts to the acini of the lacrimal gland. In the other direction it ranges through the lacrimal drainage system to the oral and pharyngeal mucosae; through ducts to the acini of the salivary glands, through the airways to the alveoli of the lungs; through the mucosae of the esophagus, stomach, intestine, and colon; through ducts to the acini of the pancreas; and through ducts to the canaliculi of the liver. These moist tissues are linked topologically, via the epidermis, with the mucosal and glandular epithelia of the urinary tract, the reproductive tracts, and the mammary glands. They also are linked functionally, via the traffic of immune cells through the lymph vessels, secondary lymphoid organs, and vasculature, to comprise a physiological system, the mucosal immune system. The metabolically active epithelial cells that comprise these surfaces perform diverse functions related to their roles in the visual, respiratory, gastrointestinal, liver, renal, and reproductive systems. However, they express several common functions. They either produce immense volumes

of fluid, for example, saliva, gastric juice, bile, pancreatic juice, and occasionally diarrhea, or produce thin, largely aqueous, but physically and chemically complex, films as homeostatic milieus exte´rieurs for themselves. They also execute innate immune functions, and, at specialized inductive and effector sites, adaptive immune functions, which protect them, their underlying stromas, and the rest of the body from particulate irritants, noxious chemicals, and infectious microbes. The central principle of the adaptive immune strategy is to use noncomplement-fixing immunoglobulins (i.e., dimeric IgA and IgG1) to prevent infection while avoiding inflammatory processes that would damage host tissues and compromise their functions. As improved sanitary systems, antibiotics, and vaccines decrease morbidity and mortality due to infection, dysfunctions of mucosal immune system tissues that result in chronic inflammatory processes join other chronic inflammatory diseases among the main categories of afflictions that burden aging populations. The chronic inflammatory mucosal immune disorders of the visual system are subsumed under the prosaic rubric, dry eye disease. Other names have been suggested, including dysfunctional tear syndrome and lacrimal keratoconjunctivitis, or, as this author would prefer, dacryokeratoconjunctivitis. Much of the current understanding of normal mucosal immune system physiology has come from studies of the gastrointestinal and respiratory systems. The gastrointestinal system must not only prevent infection by pathogens but it must also tolerate, and to a large extent, actively host, a rich flora of commensal microbes, which benefit the organism by processing or producing nutrients that would otherwise remain inaccessible or unavailable. When the mucosal immune barrier against infection fails, conventional innate and adaptive inflammatory mechanisms mount robust responses to rescue the host organism, but these responses typically induce diarrheal fluid loss, ulcerative damage to the mucosal and stromal tissues, and malabsorption of nutrients and electrolytes. The complex relationship the mucosal immune system maintains with its microbiota is paralleled by a nuanced relationship with the foods that the gastrointestinal system processes. Nutrient carbohydrates and proteins typically are digested to monosaccharides and amino acids before being absorbed across the intestinal epithelia. Nevertheless, antigen presenting cells continuously surveille the luminal contents, and both micropinocytosis and receptor-mediated transcytosis may transfer intact nutrient macromolecules across the epithelium. Clearly, it is in the organism’s interest to avoid inflammatory responses against them.

339

340

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

While a smaller mass and less rich diversity of commensal and infectious microbial flora challenge the ocular surface tissues, the imperative to avoid inflammatory responses while preventing infection is no less urgent than it is in the gastrointestinal system. The executive decisions largely are made by antigen presenting cells, but these decisions are based on information conveyed by paracrine mediators that are produced by parenchymal cells and mesenchymal cells, as well as by neurotransmitters and neuropeptides.

Organized Inductive Sites New generations of mucosal antigen presenting cells, T cells, and B cells are introduced to microbes and soluble molecules in organized signaling milieus that amplify B cells with high-affinity IgM and induce them to undergo somatic hypermutation and immunoglobulin isotype class switching while continuing to retain expression of the J chain. Some authors refer to the organized inductive sites generically as mucosa-associated lymphoid tissues (MALTs) and specifically as conjunctiva-associated lymphoid tissue (CALT), eye-associated lymphoid tissue (EALT), tonsils, adenoids, bronchus-associated lymphoid tissue (BALT), and Peyer’s patches and gut-associated lymphoid tissue (GALT). Other authors include both the organized inductive sites and the less organized effector sites together as MALT. Figure 1 illustrates stereotypical features of mucosal inductive site organization. The epithelial sheet that overlays an inductive site contains specialized cells with distinctive basal convolutions enfolding superficial dendritic cells. These so-called M cells phagocytose microbes and

Subepithelial zone

T-cell zone

engulf soluble molecules at their apical surfaces, transcytose them to their basal surfaces, and release them to the underlying space. Dendritic cells in the immediate subepithelial areas again engulf the transcytosed microbes and soluble molecules, which then traffic to acidic endosomal compartments containing major histocompatibility complex class II (MHC class II) molecules and hydrolytic enzymes. Peptides that are exposed in superficial domains of the proteins being degraded have the greatest likelihood of entering the binding grooves of MHC class II molecules. Peptides that remain outside the binding grooves are exposed to further degradation. Peptide sequences that are protected from proteolysis become the dominant epitopes that the MHC class II molecules will present to CD4+ T cells upon trafficking to the dendritic cells’ surface membranes. Whether or not a dendritic cell redistributes MHC class II molecules to its surface, upregulates expression of B7 costimulatory molecules, and migrates the short distance to the CD4+ T-cell-rich zone may depend on signals received early in the encounter with the material it has internalized. The best understood of these signals are conveyed by lipopolysaccharides and double-stranded DNAs, which activate toll-like receptors (TLRs). The TLR activation induces a dendritic cell to downregulate chemokine receptors and homing receptors that favor retention in the subepithelial zone and to upregulate receptors that favor migration to the T-cell-rich zone. Typically, TLR activation also induces dendritic cells to upregulate surface expression of the costimulatory B7 ligands, CD80 and CD86, as they redistribute MHC class II molecule–epitope complexes to their surface membranes.

B-cell zone

Germinal center

Figure 1 Stereotypical organization of mucosal immune inductive sites. After having taken up microbes or antigens transcytosed by M cells, dendritic cells may traffic directly to the B-cell zone, or they may migrate to the T-cell zone, where they activate antigen-specific CD4+ cells to express TH2 cytokines. B cells with antigen receptors of sufficient avidity internalize antigen, then use MHC class II molecules to present epitopes to activated TH2 cells. Ongoing TH2 activation promotes B-cell division and Ig hypermutation. In this signaling milieu, TGF-b induces IgM-to-IgA isotype class switching. In combination with B-cell growth factors, TGF-b in germinal centers promotes plasmablast proliferation and emigration via afferent lymph vessels.

Adaptive Immune System and the Eye: Mucosal Immunity

While signals from engulfed microbes are decisive in determining whether the dendritic cell will undergo complete functional maturation and activation, the activated phenotype that the dendritic cell will assume is determined largely by paracrine mediators that are released by the overlaying epithelium and surrounding mesenchymal cells. The signaling mediators that are known to be predominant are transforming growth factor-beta (TGF-b) and interleukin (IL)-10, and they induce maturing dendritic cells to also express TGF-b and IL-10. An additional population of dendritic cells are distinguished by the absence of MHC class II molecule expression and expression of chemokine and homing signal receptors that lead them to bypass the T-cell-rich regions and enter B-cell-rich zones, where there they will release the material they had internalized. Engagement of a naive CD4+ T cell’s antigen receptors by MHC class II molecule–epitope complexes generates the primary signal necessary, but not sufficient, for activation. Simultaneous engagement of CTLA-4 or CD28 at the T-cell surface by CD80 or CD86 at the dendritic cell surface provides the second signal essential for promoting T-cell activation and differentiation. However, the functional phenotype the activated CD4+ T-cell expresses is determined by paracrine mediators in the immediate signaling milieu. These are secreted by dendritic cells, epithelial cells, and mesenchymal cells. They induce CD4+ T cells to differentiate as TH2 cells and to change their panel of chemokine receptors and homing receptors to favor migration from the T-cell-rich zones and into the B-cell-rich zone. When molecules that dendritic cells have released encounter B cells with surface IgM antigen receptors that bind them with high enough affinity, they deliver a primary activation signal, which induces the B cell to endocytose the IgM–antigen complex to antigen-processing compartments and to begin expressing MHC class II molecules. Thus, CD4+ T cells arriving from the interfollicular zone will be presented with dominant epitopes that maintain them in their activated state. They continue secreting the TH2 cytokines needed to induce somatic hypermutation of B-cell immunoglobulin genes and promote selection of B-cell clones with increasing affinity for the determinant. The signaling milieus within some of the organized inductive sites, notably Peyer’s patches, selectively induce IgMto-IgA class switch recombination. The milieus in the inductive sites in the tonsils, adenoids, and airways may induce either IgM-to-IgA or IgM-to-IgG1 class switch recombination. In cases of IgA deficiency, the numbers of IgG1+ cells and IgM cells may increase in compensation. Whether induced to express IgA or IgM, the activated B cells continue to express J chain. The IgA+ B cells will express dimeric IgA (dIgA), while IgG1 B cells will express apparently irrelevant J chain. The activated B-cell blasts amplify in germinal centers, and they alter their panel of chemokine receptors and

341

homing receptors, so that they are induced to leave the inductive sites via lymph vessels, enter the circulation via the thoracic duct, and enter the mucosal immune effector sites. The IgA+ B cells expressing CCR10, the receptor for CCL28, traffic to the lamina propria of the oral cavity, lower airways, and colon. The IgA+ B cells expressing CCR9, the receptor for CCL25, traffic to the trachea, stomach, intestine, salivary glands, and mammary gland. The IgG1+ B cells expressing CXCR4, the receptor for CXCL12, traffic to the bone marrow. In each setting, the plasmablasts will take residence and mature to become long-lived memory cells or plasmacytes.

Effector Sites Specialized Niches Stereotypical features of the mucosal immune effector site are illustrated in the upper panel of Figure 2. Like IgG+ plasmablasts entering the bone marrow, the dIgA+ plasmablasts that arrive in the mucosal effector sites receive signals that induce them to mature into plasmacytes. One of the critical mediators of maturation signaling at the mucosal effector sites is TGF-b. A mature plasmacyte’s ongoing survival depends on the constant presence of additional signals to suppress the intrinsic apoptotic program. The candidate mediators of the survival signals include IL-6, B-cell activating factor (BAFF), a proliferation-inducing ligand (APRIL), and, perhaps, prolactin. In the rabbit lacrimal gland, which has been studied in some detail, epithelial expression of paracrine mediators is compartmentalized according to the epithelial histoarchitectural plan and vascular organization. Venules, the vascular elements that plasmablasts cross to enter the stromal space, tend to follow courses that parallel interlobular ducts. Epithelial cells of the interlobular ducts express TGF-b and prolactin at much higher levels than epithelial cells in the acini, and it appears they secrete these mediators both apically (i.e., into the fluid forming within the duct lumen) by way of their regulated exocrine secretory apparatus for proteins, and also basally, through either a constitutive paracrine secretory apparatus or a parallel paracrine apparatus that can be induced under certain circumstances. While TGF-b, as noted, induces dIgA+ plasmablast-to-plasmacyte differentiation, it may not be conducive to the plasmacytes’ ongoing survival. It appears that TGF-b signaling may be concentrated in the stromal spaces surrounding the venules and interlobular ducts. Preliminary findings in the author’s laboratory suggest that transcripts for the extracellular matrix proteoglycan, decorin, are two orders of magnitude more abundant than TGF-b transcripts in lacrimal gland epithelial cells. Because decorin binds to the extracellular matrix and also binds the latent TGF-b, it may create a diffusion barrier that keeps latent TGF-b

342

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

TGF-β IL-10 IL-6

Enk NPY SP α-MSH VIP ACh 5-HT nEpi

Figure 2 Normal (upper panel) and infected (lower panel) mucosal immune effector sites. IgA+ J chain+ plasmablasts, immature dendritic cells, and T cells enter the lamina propria of the conjunctiva and drainage system, stromal space of the lacrimal gland, and other effector sites from venules. Paracrine mediators induce plasmablasts to undergo terminal differentiation, then support their survival as plasmacytes. Of these mediators, TGF-b, IL-10, and IL-6 are provided, at least in part, by epithelial and stromal cells. Sensory, parasympathetic, and sympathetic nerve endings provide neurotransmitters and neuropeptides that may also influence plasmablasts, plasmacytes, dendritic cells, and lymphocytes. The normal signaling milieu induces dendritic cells that have taken up soluble proteins and cellular debris to mature as regulatory antigen presenting cells, which exit via afferent lymph vessels and induce regulatory T cells in the lymph nodes. The TLR activation by microbes that have evaded the immunospecific sIgA barrier induces dendritic cells to mature as immunoactivating antigen presenting cells that will elicit IgG responses in the lymph nodes.

concentrated in the periductal stroma. In contrast, prolactin may be free to equilibrate through the interstitial fluid, diffusing to plasmablasts throughout the gland’s stromal spaces. Preliminary studies suggest that acinar epithelial cells may be able to express IL-6. In contrast, mRNAs for APRIL and BAFF appear substantially more abundant in whole gland extracts than in ex vivo acinar

cells. If substantiated, this finding would imply that APRIL and BAFF are expressed by infiltrating lymphocytes, but it cannot exclude significant expression by ductal epithelial cells. The information so far available about the cytophysiological mechanisms that secrete paracrine mediators to the stroma has come from immunohistochemical studies

Adaptive Immune System and the Eye: Mucosal Immunity

of TGF-b, epidermal growth factor (EGF), fibroblast growth factor (FGF), and prolactin localizations in rabbit lacrimal gland and studies of the intracellular traffic of prolactin in ex vivo acinar cells from rabbit lacrimal gland. Under normal circumstances, the paracrine mediators are localized primarily in the apical cytoplasm, consistent with secretion into the duct lumen via the regulated exocrine secretory apparatus. Consistent with this inference, pilocarpine-induced lacrimal gland fluid contains high concentrations of TGF-b and prolactin. Two-color confocal immunofluorescence microscopy of the ex vivo model indicates that prolactin colocalizes with rab4, rab5, and rab11, which are effectors of traffic to and from endosomes, and electron microscopic-immunogold studies indicate that prolactin also is localized within mature secretory vesicles. These findings suggest that lacrimal epithelial cells may maintain two pools of the paracrine mediators they secrete: a large, often static, pool in the regulated exocrine secretory apparatus, and a small, but high throughput, pool in the paracrine apparatus.

Transfer of Immunoglobulins to the Milieu Exte´rieur pIgR and dIgA Several prominent aspects of lacrimal cytophysiology appear to reflect adaptations to the roles the epithelium plays in the mucosal immune system: maintaining the underlying stromal space as a niche for dIgA-secreting plasmacytes and then transferring dIgA from the stromal space to the fluid within the lumen of the acinus–duct system. The fundamental principles of the transcytotic mechanism for dIgA secretion were elucidated in studies of Madin-Darby canine kidney (MDCK) cells, derived from a canine kidney, which form continuous epithelial monolayers when cultured on microporous substrata. Renal tubular epithelial cells do not normally secrete IgA, but when MDCK cells are stably transduced to express the polymeric immunoglobulin receptor (pIgR), they gain the ability to internalize dIgA from their basal medium, which corresponds to the interstitial fluid, and secrete sIgA into their apical medium, which corresponds to the lumenal fluid of an exocrine gland or the surface fluid of a mucosal sheet. Figure 3 illustrates the transcytotic apparatus as it is thought to function in conjunctival epithelial cells. The pIgR makes its way from the biosynthetic apparatus, that is, the endoplasmic reticulum and Golgi complex, through the trans-Golgi network (TGN) and endosomes, to the basal–lateral plasma membranes, hitchhiking in transport vesicles at each step. The pIgR may or may not bind the J chain and associated immunoglobulin while its ligand-binding domain is exposed at the basal–lateral membrane. In either case, it is internalized in an endocytotic transport vesicle that will traffic to the

343

early endosome. Another transport vesicle will traffic pIgR and pIgR–dIgA complex to the recycling endosome, and yet another will traffic pIgR and pIgR–dIgA complex to the apical plasma membrane. Either during its transit through the recycling endosome, or upon arriving at the apical plasma membrane, the pIgR will encounter a protease that cleaves the extracytoplasmic, dIgA-binding domain, referred to as secretory component (SC), from the membrane-spanning domain. Thus, the cell releases both free SC and SC–dIgA complexes, that is, secretory IgA (sIgA), into the milieu exte´rieur. Consistent with the topological constraints that preclude traffic of transport vesicles between adjoining cells, expression of pIgR is concentrated in the apical-most layer of epithelial cells in the conjunctiva. Acute Regulation of pIgR Traffic The traffic of pIgR through lacrimal gland epithelial cells appears to be somewhat more complex than described in the MDCK cell paradigm. First, it appears that pIgR might cycle repeatedly to the basal–lateral plasma membrane from any compartment of the transcytotic apparatus, while free SC is preferentially trafficked from the recycling endosome to the apical plasma membrane. Second, some of the pIgR traffic from the TGN to the immature secretory vesicle, rather than to the transcytotic apparatus; SC derived from these pIgR remains in mature secretory vesicles until released into the lumen along with the classical secretory proteins. Interestingly, many of the classic lacrimal secretory proteins, such as b-hexosaminidase, lysozyme, and lactoperoxidase, are innate immune effector molecules. Free SC appears to perform important functions in the ocular surface fluid film, and the ability to sequester a portion of pIgR away from the transcytotic apparatus ensures that some free SCs always will be secreted, despite the large mass of dIgA that normally is present in the stromal space. A considerable volume of membrane traffic flows through the lacrimal acinar cell’s transcytotic apparatus. The micropinocytosis associated with endocytotic internalization of transport vesicles from the basal–lateral and apical surface membranes carries fluid phase markers into the cells, and such markers rapidly equilibrate throughout the aqueous phase within the lumena of the early endosome, recycling endosome, and TGN, as well as the late endosome, prelysosome, and lysosome. If ex vivo acinar cell models are acutely stimulated with the muscarinic acetylcholine receptor agonist, carbamylcholine (carbachol, CCh), they increase the rates at which they endocytose and exocytose fluid phase markers. If the CCh concentration is increased from 10 mM, which maximally accelerates fluid phase endocytosis but half-maximally accelerates protein secretion, to 100 mM, a concentration that maximally accelerates protein secretion, fluid phase endocytosis continues at the increased rate, but

344

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

Recycling endosome

Early endosome

er Golgi tgn

Late endosome Prelysosome

sIgA

SC Membrane mucin

pIgR

dIgA

Lysosomal protease

Figure 3 Transcytotic traffic of sIgA to the milieu exte´rieur. The apical-most layer of epithelial cells in stratified tissues, like the epithelial cells of lacrimal acini and ducts, express pIgR. At the basal–lateral surface membrane pIgR mediates endocytosis of dIgA, which it then chaperones through the early endosome and apical recycling endosome. Either within the apical recycling endosome or at the apical plasma membrane, the pIgR is cleaved by protease to release either free SC or dIgA–SC complex, that is, sIgA. In lacrimal gland epithelial cells, pIgR may cycle repeatedly through the TGN and transcytotic apparatus; it may enter immature secretory vesicles (not shown), where free SC is released and stored for subsequent secretion in response to secretagogue stimulation, or it may traffic through the late endosome, which is a gateway to the prelysosome and autophagic–lysosomal apparatus. The author proposes that the traffic of transport vesicles returning from the prelysosome and late endosome to the early endosome accounts for the constitutive release of an unusually heavy burden of autoantigens to the stromal space.

the volume of space that the fluid can reach decreases. If a marker is allowed to equilibrate through the fluid phases of the network of endomembrane compartments under resting conditions, subsequent acute stimulation with 100 mM CCh accelerates its release, but release then stops with a substantial portion of the marker remaining entrapped within the cell. These findings suggest that when the cell is maximally stimulated with CCh, it imposes a blockade of bidirectional traffic between proximal and distal compartments of the endocytotic pathway. Other studies indicate that the blockade affects traffic to the late endosome from both the early endosome and the TGN. A possible cytophysiological rationale for this phenomenon is suggested by preliminary studies of the kinetics of CChinduced SC secretion. The CCh acutely induces a burst of secretory vesicle exocytosis, which releases SC along with the other secretory proteins. Secretory vesicle exocytosis is complete within 10 min. However, the cells continue to

release SC at an increased rate for at least 60 min, while their total content of pIgR plus SC appears to remain unaltered. These phenomena suggest that the cell blockades traffic to the late endosome in order to divert an increased proportion of its newly synthesized pIgR from the degradative apparatus to the transcytotic apparatus. A Mechanism That Constitutively Secretes Autoantigens to the Stroma A consequence of the strategy of directing a large volume of transport vesicle traffic from the early endosome to the late endosome and prelysosome under nonstimulated conditions is that the reciprocal traffic of transport vesicles constitutively secretes lysosomal hydrolases and potential autoantigens that, in other cells, would remain in the autophagic–lysosomal apparatus and be degraded. Moreover, when traffic to the late endosome is blocked,

Adaptive Immune System and the Eye: Mucosal Immunity

the lysosomal hydrolases reflux into the compartments which remain accessible (i.e., the TGN), immature secretory vesicle, early endosome, and recycling endosome. Of particular interest, at least some of the lysosomal proteases are catalytically active and may process other proteins that have accumulated with them in the same compartments. This phenomenon might lead to the degradation of dominant autoantigen epitopes and lead to the eventual presentation of epitopes that under normal circumstances are degraded. FcgRn and IgG1 While the plasmacytes that produce dIgA reside in the stromal spaces of exocrine glands and the lamina propria of mucosal sheets, the plasmacytes that produce IgG1 typically reside in the bone marrow, and IgG1 must travel through the circulation to reach the epithelia that will secrete it. The same subcellular apparatus that transports SC and pIgR–dIgA complexes through epithelial cells likely also mediates the secretion of IgG1. However, IgG1 does not associate with J chain, and its traffic is mediated not by pIgR but by the neonatal Fcg receptor, FcgRn. The association between IgG1 and FcgRn is regulated by pH, such that binding affinity is highest in relatively acidic endosomal compartments and lower at the pH of the interstitial fluid. In the tissues where it has been studied, the IgG1–FcgRn complex tends to bypass the late endosome in favor of the early endosome and recycling endosome. Because dissociation of IgG1 from FcgRn does not require proteolytic processing, FcgRn can mediate multiple cycles of IgG1 absorption as well as secretion. The FcgRn can also mediate absorption and secretion of IgG1-immune complexes, in some cases secreting opsonized microbes and immune complexes and in others absorbing them in a physical state that minimizes the risk of infection but perpetuates T-cell and B-cell responses. The extent to which FcgRn might be expressed in the lacrimal glands and drainage system has not been investigated; recently, however, it has been found to be present in epithelial cells of rat conjunctiva.

Mucosal Tolerance and Response to Infection Commensal microbes in the gut are coated with sIgA, which normally prevents them from entering the epithelial lining. Notably, dendritic cells as well as dIgA-secreting plasmacytes populate the mucosal effector sites. They are present within the lamina propria and the epithelia and may extend processes that interdigitate between adjoining cells to surveille the milieu exte´rieur. As in the organized inductive sites, immature dendritic cells are endocytotically active, and under normal circumstances the material they

345

internalize consists of soluble molecules and cellular debris. The robustness of the phenomenon of oral tolerance suggests that the TGF-b- and IL-10-rich milieu directs them to mature as regulatory antigen presenting cells, which, when they arrive in lymph nodes, induce food antigenand, presumably, autoantigen-specific CD4+ T cells to mature as TH3 and TR1 regulatory cells. The regulatory cells might function in both the effector sites and the lymph nodes to prevent inflammatory mediators or activated effect T cells from redirecting dendritic cells toward expression of immunoactivating phenotypes. As illustrated in the lower panel of Figure 2, TLR on immature dendritic cells in the epithelium or lamina propria recognizes microbes that have evaded the immunospecific sIgA barrier, and TLR activation induces phagocytosis, redistribution of MHC class II molecules to the dendritic cell surface, and upregulation of surface CD80 and CD86. Certain microbes induce maturing dendritic cells to remain in the lamina propria and release antigen directly to B1 cells, stimulating activation independently of T-cell help. Typically, however, dendritic cells are activated to switch chemokine receptor and homing receptor expression to favor emigration via afferent lymph vessels. In the lymph nodes, they will induce activation of naive, antigenspecific CD4+ T cells as TH2 cells and generation of IgGproducing B cells.

Implications for Ocular Surface Pathophysiology Contemporary paradigms hold that dry eye disease results from insufficiency of the ocular surface fluid film. Such might be the consequence of impaired lacrimal gland fluid production, excessive evaporation of water from the ocular surface fluid due to deficient production of lipids in the Meibomian glands, and excessive breakup of the fluid film due to deficiencies of the normally hydrophilic substratum provided by the apical surfaces of corneal and conjunctival epithelial cells. There is a general consensus that, whatever the etiology, the principal feature of pathophysiological state is inflammation of the ocular surface tissues. Inflammatory processes plausibly account for much of the cytopathology that occurs within the lacrimal gland and leads to quiescence of impairment of its exocrine functions. This seems to be the case, not only in Sjo¨gren’s syndrome, but also in graft-versus-host disease, sarcoidosis, Wegener’s granulomatosis, and diffuse infiltrative lymphocytosis syndrome associated with HIV infection. A histopathological syndrome distinct from the familiar diagnoses and described primarily through postmortem studies is characterized by lymphocytic infiltration, acinar atrophy, ductal dilatation, and periductal fibrosis. Its underlying pathophysiological mechanisms have not been studied, although they appear not to involve production of autoantibodies.

346

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

The author suggests that these inflammatory processes normally are prevented by the same strategy that prevents immune responses to food antigens, soluble autoantigens, and cellular debris. There may be particular challenges to maintaining immunoregulation in the lacrimal glands, where vigorous traffic through the transcytotic apparatus releases an exceptionally heavy burden of potential autoantigens into the stromal space, and in the ocular surface tissues, where environmental stresses may induce release of inflammatory mediators that abrogate the local signaling milieu’s ability to generate regulatory antigen presenting cells. Additional challenges may result from the epithelia’s dependence on systemic hormones to maintain expression of the local signaling milieu. See also: Adaptive Immune System and the Eye: T CellMediated Immunity; Conjunctiva Immune Surveillance; Defense Mechanisms of Tears and Ocular Surface; Lacrimal Gland Overview; Tear Film.

Further Reading Brandtzaeg, P. and Johansen, F.-E. (2005). Mucosal B cells: Phenotypic characteristics, transcriptional regulation, and homing properties. Immunological Reviews 206: 32–63. Cain, C. and Phillips, T. E. (2008). Developmental changes in conjunctiva-associated lymphoid tissue of the rabbit. Investigative Ophthalmology and Visual Science 49: 644–649. Evans, E., Zhang, W., Jerdeva, G., et al. (2008). Direct interaction between Rab3D and the polymeric immunoglobulin receptor and trafficking through regulated secretory vesicles in lacrimal gland acinar cells. American Journal of Physiology. Cell Physiology 294: C662–C674. Franklin, R. M, Kenyon, K. R., and Tomasi, T. B. (1973). Immunohistologic studies of human lacrimal gland: Localization of

immunoglobulins, secretory component and lactoferrin. Journal of Immunology 110: 984–992. Kaetzel, C. S. (2005). The polymeric immunoglobulin receptor: Bridging innate and adaptive immune responses at mucosal surfaces. Immunological Reviews 206: 83–99. Kim, H., Fariss, R. N., Zhang, C., et al. (2008). Mapping of the neonatal Fc receptor in the rodent eye. Investigative Ophthalmology and Visual Science 49: 2025–2029. Knop, E., Knop, N., and Claus, P. (2008). Local production of secretory IgA in the eye-associated lymphoid tissue (EALT) of the normal human ocular surface. Investigative Ophthalmology and Visual Science 49: 2322–2329. Macpherson, A. J., McCoy, K. D., Johansen, F.-E., and Brandtzaeg, P. (2008). The immune geography of IgA induction and function. Mucosal Immunology 1: 11–22. Meagher, C. K., Liu, H., Moore, C. P., and Phillips, T. E. (2005). Conjunctival M cells selectively bind and translocate Maackia amurensis leukoagglutinin. Experimental Eye Research 80: 545–553. Mircheff, A. K., Wang, Y., de Saint Jean, M., Ding, C., and Schechter, J. E. (2007). Lacrimal epithelium mediates hormonal influences on APC and lymphocyte cycles in the ocular surface system. In: Zierhut, M., Rammensee, H. G., and Streilein, J. W. (eds.) Antigen Presenting Cells and the Eye, pp. 93–119. New York: Informa. Rose, C. M., Qian, L., Hakim, L., et al. (2005). Accumulation of catalytically active proteases in lacrimal gland acinar cell endosomes during chronic ex vivo muscarinic receptor stimulation. Scandinavian Journal of Immunology 61: 36–50. Seder, R. A., Marth, T., Sieve, M. C., et al. (1998). Factors involved in the differentiation of TGF-beta-producing cells from naive CD4+ T cells: IL-4 and IFN-g have opposing effects, while TGF-b positively regulates its own production. Journal of Immunology 160: 5719–5728. Spiekermann, G. M., Finn, P. W., Ward, E. S., et al. (2002). Receptor-mediated immunoglobulin G transport across mucosal barriers in adult life: Functional expression of FcRn in the mammalian lung. Journal of Experimental Medicine 196: 303–310. Wang, Y., Chiu, C. T., Nakamura, T., et al. (2007). Traffic of endogenous, over-expressed, and endocytosed prolactin in rabbit lacrimal acinar cells. Experimental Eye Research 85: 749–761. Weiner, H. L. (2001). Induction and mechanism of action of transforming growth factor-b-secreting Th3 regulatory cells. Immunological Reviews 182: 207–214.

Adaptive Immune System and the Eye: T Cell-Mediated Immunity K C McKenna and R D Vicetti Miguel, University of Pittsburgh, Pittsburgh, PA, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Antigen – A molecule recognized by receptors expressed by T cells (TCR) and B cells (immunoglobulin). Complement – A system of serum proteins and cell surface proteins that combine to form the membrane attack complex that perforates cell membranes. Component proteins of the complement cascade have enzymatic activity and generate inflammatory mediators. Cytokines – Soluble molecules that modulate immune responses. Inflammation – The hallmark of an immune response characterized by edema, redness, and pain. Inflammatory mediators induce vasodilation of local vessels which are leaky and promote edema and facilitate immune cell infiltration. While inflammation controls infection, it can also cause tissue damage, as the effector function of the innate immune response is nonspecific. Major histocompatibility complex (MHC) – Highly polymorphic molecule expressed on the cell surface that presents self and foreign peptides for T-cell recognition through the TCR. Opsonization – The coating of cell surfaces with molecules that facilitate phagocytosis. Phagocytosis – The process of pathogen engulfment through recognition of surface receptors. Polymorphonuclear leukocytes (PMN) – Cells of the innate immune response characterized by a multilobed nucleus including neutrophils and eosinophils. Signal transduction – The process by which signals received by the cell are converted into gene expression within the nucleus.

The human body and the microbes (bacteria, viruses, and parasites) that reside within it share the same simple mandate: to survive and to reproduce. The ground rules of the host are clear. The host–microbe relationship cannot cause damage that compromises the health of the host. When health is compromised, the host immune system eliminates or controls the growth of these unruly microbes, now termed pathogens. Immune responses in the eye are critical for protection from pathogens. For example, immunesuppressed HIVþ patients are more susceptible to ocular

infection. However, ocular immune responses also pose a threat as inflammation can damage the delicate microanatomy of the eye and compromise vision. Examples of blinding inflammation include viral, bacterial, and parasitic keratitis, retinitis, bacterial endophthalmitis, and autoimmune uveitis, which are reviewed elsewhere in this encyclopedia. As an evolutionary adaptation to preserve vision, immune responses in the eye are normally tightly regulated to control pathogens while minimizing inflammation.

Innate and Adaptive Immunity The two major arms of the immune system, innate and adaptive immunity, can be distinguished by the unique molecular structures they employ to recognize molecules expressed by pathogens or expressed by a host cell in response to a pathogen. The innate immune response utilizes a limited set of pathogen recognition receptors (PRRs) to detect pathogen-associated molecular patterns (PAMPs), which are conserved pathogen products. There are PRRs that recognize PAMPs expressed by all major classes of microbes for example: toll-like receptor (TLR) 4 recognizes lippopolysaccharide (LPS) of Gram-negative bacteria; TLR3 recognizes double-stranded RNA expressed by viruses; and TLR11 recognizes proteins expressed by parasites. PRRs can also recognize molecules associated with pathogen-induced cell death. For example, RAGE (receptor of advance glycation end products), TLR 2, 4, and 9 recognize nonoxidized high mobility group box 1 (HMGB1), which is released by normal cells upon necrotic but not apoptotic cell death. PRRs are expressed on epithelial cells at sites of microbial entry and on cells of the innate immune response which include: macrophages, dendritic cells (DC), and polymorphonuclear leukocytes (PMN). Pathogen recognition by the adaptive immune response is not predetermined by a limited set of receptors. Rather, a random process of gene segment rearrangement by recombinase activation genes (Rag-1 and Rag-2) generates over a billion different receptors tailored to recognize essentially any molecule. Cells of the adaptive immune response include B cells, which express immunoglobulin (Ig) molecules and T cells that express T-cell receptors (TCRs). Individual B or T cells express receptors with a single specificity creating a repertoire of many B and T cells each with different specificities. B cells mediate humoral immunity through production of secreted Ig molecules (antibodies) which directly recognize

347

348

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

unique shapes, and conformations of pathogen-associated molecules including proteins, carbohydrates, lipids, nucleic acids, and simple chemical groups. Cellular immunity is mediated by T cells through TCR recognition of pathogen-associated proteins processed into peptides and presented on the cell surface by the major histocompatibility complex (MHC). These differences in antigen recognition specialize humoral immunity for defense against extracellular pathogens and cell-mediated immunity for defense against intracellular pathogens. The focus of this article is a general overview of T-cellmediated immunity introducing how T-cell responses are regulated in the eye and conditions where ocular T-cell responses cause immunopathology. Antigen Trafficking, Processing, and Presentation to T cells The lymphatic system drains extracellular fluids through a network of collection channels and vessels that ultimately empty into the bloodstream through the thoracic duct. Interspersed between lymphatic vessels are lymph nodes that are filled with immune cells. This strategic placement of an immune organ within a drainage pathway allows the immune system to survey for foreign antigens contained within extracellular fluids from different regions of the body. For example, the eye is drained to cervical lymph nodes within the neck whereas the arms drain to axillary and brachial lymph nodes. Internal organs also have lymphatic drainage: lungs drain to mediastinal lymph nodes and intestine drains to mesenteric lymph nodes. T cells, which have not encountered their antigen, are found in the lymph nodes, spleen, and blood but not in nonlymphoid tissues, such as the eye. Therefore, these naive T cells are not the first responders to an ocular pathogen. Rather, innate immunity is the first line of defense. Recognition of PAMPs through PRRs induces rapid expression of inflammatory cytokines (IL (interleukin)-1, IL-6, and tumor necrosis factor (TNF)) that increase the production of chemokines, such as IL-8 and acute phase proteins (e.g., C-reactive protein). Chemokines promote the migration of immune cells to the eye while acute phase proteins opsonize microbial surfaces where they activate complement and promote phagocytosis by macrophages and PMN. To control the pathogen, infiltrating PMN and macrophages produce reactive oxygen species, including superoxide which can be transformed into H2O2 by superoxide dismutase. H2O2 and chloride are substrates for myeloperoxidase which generates antimicrobial hypochlorous acid, the active ingredient of bleach. PMN and macrophages also release granules containing antimicrobial molecules and generate nitric oxide (NO) that reacts with superoxide to form the potent oxidant peroxynitrite. In addition to controlling the pathogen, DCs and macrophages which have internalized pathogens

and/or pathogen products migrate to the lymph nodes through lymphatic vessels to activate pathogen-specific T cells. Adaptive immune responses depend on innate immune responses, as T cells only recognize cell-associated antigens. Hence, antigens that flow freely into the lymph node or that are carried by DCs and macrophages from the pathogen site have to be processed and presented by antigen-presenting cells (APCs) for recognition by T cells. As a general rule, pathogen-expressed proteins from the extracellular milieu are digested by phagocytic APC and presented through MHC class II molecules for recognition by CD4þ T cells, whereas intracellular pathogen-expressed proteins are processed and presented through MHC class I molecules for recognition by CD8þ T cells (Figure 1). However, certain types of APC, for example CD8aþ CD11cþ DCs from mice, cross present extracellular proteins through MHC class I molecules. T-cell activation Very few T cells within the entire T-cell repertoire are specific for any one particular antigen. Therefore, to most effectively respond to a pathogen, these antigen-specific T cells undergo cell division to expand their numbers. This process of clonal expansion takes time, which explains why adaptive immune responses lag behind innate immune responses. The activation of T cells to undergo proliferation results from several signals delivered from the APC. These signals initiate and sustain signal transduction cascades culminating in the transcription of genes necessary for T-cell proliferation, most importantly IL-2 (Figure 2). Transducing signals from the APC to T cells occurs by phosphorylation and dephosphorylation of intracellular substrates by kinases and phosphatases, respectively. Signal one is delivered by TCR engagement of MHC þ peptide complexes, which results in a conformational change in the TCR that makes immunoreceptor tyrosine-based activation motifs (ITAMs) accessible on chains of the CD3 complex which is comprised of five distinct molecules (g, d, e, z, and Z). The CD3 complex is associated with the TCR and is necessary for signal transduction upon TCR engagement, as the TCR has a short cytoplasmic domain incapable of signaling. Upon TCR antigen recognition, CD3 ITAMS are phosphoylated by p56lck, and p59fyn kinases, then bound by ZAP-70 ((CD3) zeta chain-associated protein of 70 kDa) kinase, which activates phospholipase Cg1 (PLCg1). PLCg1 is an enzyme which catabolizes membrane-bound phosphatidlyinositol-1, 4-bisphosphate (PIP2) into inositol-1,4,5-triphosphate (IP3) and diacylglcerol (DAG). IP3 induces an intracellular calcium increase which then activates the serine phosphatase calcineurin to dephosphorylate nuclear factor of activated T cells (NFAT). Dephosphorlated NFAT leaves the cytosol to enter the nucleus. DAG activates protein kinase C (PKC) initiating another cascade leading to degradation of the

Adaptive Immune System and the Eye: T Cell-Mediated Immunity

APC

349

APC Nucleus

Nucleus

ER

CLIP

ER

TAP Pr Cytosolic proteins

Extracellular proteins MHC class I

CD8

MHC class II CD4

TCR

TCR

CTL

Th

(a)

(b)

Figure 1 Antigen processing and presentation to T cells. (a) Proteins expressed within the cytosol are degraded into peptides by the proteasome (Pr), transported into the endoplasmic reticulum (ER) by the transporter associated with antigen processing (TAP) and then associate with newly formed MHC class I molecules which are transported to the cell surface for recognition by CD8þ cytolytic T lymphocytes (CTL). (b) Extracellular proteins are degraded into peptides in endosomal vesicles that fuse with exocytic vesicles containing MHC class II molecules associated with the class II invariant chain peptide (CLIP). CLIP is then removed by the protein DM and replaced by peptides derived from extracellular proteins. MHC class II + peptide complexes are then transported to the cell surface for recognition by CD4þ T helper (Th) cells.

Signal one

Signal one + singnal two

APC

APC

MHC

MHC

TCR

TCR

CD3

PLCγ PIP2 IP3

Calcineurin

iκB/NFκb

DGKα DAG

IP3

RasGRP PKC

NFAT NFκB AP-1

Ca2+

Ras MAPK

Calcineurin

IL-2 DGKα

Anergy

RasGRP PKC iκB/NFκb

Ras MAPK

NFAT NFκB AP-1

Nucleus

(a)

CD28

PLCγ PIP2

DGKα DAG

Ca2+

CD80/CD86

IL-2 DGKα

Nucleus

(b)

Activation

Figure 2 Signal transduction pathways of T-cell activation and anergy. (a) Engagement of TCR alone favors NFAT activation leading to inhibited IL-2 production and the induction of anergy-inducing genes. (b) Engagement of TCR along with co-stimulation provides balanced NFAT and NFkB activation and AP-1 formation leading to IL-2 production and repression of anergy-inducing genes. APC, antigen-presenting cell; MHC, major histocompatability complex; TCR, T-cell receptor; PLCg, phospholipase Cg; PIP2, phosphatidylinositol-1,4 bisphosphate; IP3, inositol 1,4,5-triphosphate; DGKa, diacylglycerol kinase alpha; DAG, diacylglycerol; PKC, protein kinase C; RasGRP, Ras guanyl releasing protein; MAPK, mitogen activated protein kinase; ikB, inhibitor of NFkB; NFkB, nuclear factor kB; NFAT, nuclear factor of activated T cells; AP-1, activator protein-1; IL-2, interleukin 2.

inhibitor of NFkB (ikB) and subsequent translocation of NFkB from the cytosol to the nucleus. DAG also activates the Ras pathway, which through mitogen-activated protein kinase (MAPK) leads to formation and activation of the activator protein-1 (AP-1) molecule. NFAT, NFkB, and AP-1 are all transcription factors that cooperate to induce transcription of IL-2 and other T-cell activation genes.

Signal one alone, however, is not sufficient to induce complete T-cell activation. Signal two is delivered by invariant molecules expressed on the cell surface of APC that are upregulated after PRR engagement of PAMPs. The best characterized of these co-stimulatory molecules are CD80 (B7-1) and CD86 (B7-2), which bind to CD28 on T cells. CD28

350

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

engagement further activates the Ras/MAPK pathway to promote IL-2 production by making the IL-2 promoter accessible to transcription factors by chromatin remodeling, and by stabilizing IL-2 mRNA. Immediately, T-cell proliferation ensues. The benefit of generating T cells with TCR capable of recognizing a tremendously diverse array of antigens carries the consequence that the same process generates TCR that recognize self molecules. While the most dangerous self-reactive T cells are deleted during T-cell development in the thymus by a process referred to as negative selection, many T cells with cross reactivity to both foreign and selfantigens escape negative selection. The requirement for two signals to induce T-cell activation may have developed to ensure that these cross-reactive T cells are activated only under appropriate circumstances; for example, when foreign antigen is encountered. During normal physiological conditions, APC that present self-antigens will not be activated by PAMPs to upregulate co-stimulatory molecules. Hence, signal one but not signal two will be delivered. Signaling through the TCR alone leads to long-lived T-cell unresponsiveness referred to as anergy, which is characterized by inhibited expression of IL-2. Therefore, anergy induction protects the body from the generation of autoimmunity while preserving the most extensive repertoire of TCR. The current paradigm for anergy induction suggests that signal one induces NFAT activation with only partial AP-1 activation causing a signaling imbalance that leads to the expression of anergy-inducing genes, such as diacylglycerol kinase alpha (DGKa) (Figure 2(a)). DGKa phosphorylates DAG promoting its degradation and further decreasing activation of the Ras/MAPK and PKC pathways. Ultimately, the IL-2 promoter becomes inaccessible to transcription factors due to chromatin remodeling, and as a result, the cell becomes anergic. Anergy induction also involves T-cell expression of inhibitory co-stimulatory molecules, such as cytotoxic T-lymphocyte antigen 4 (CTLA-4), which antagonize activating co-stimulatory molecules. T cell differentiation and effector function

Upon activation, T cells receive further instructions in the form of cytokines that promote differentiation that is best suited to respond to the offending pathogen. For example, CD4þ T helper (Th) responses are directed toward at least four effector lineages: Th1, Th2, Th17, and adaptive T regulatory cells (Treg) by exposure to particular cytokines IL-12/ interferong (IFNg), IL-4, transforming growth factor-b (TGF-b)/IL-6, or TGF-b, respectively (Figure 3). CD8þ T cells differentiate into cytolytic T lymphocytes (CTLs) by exposure to IL-12 and express granzymes and perforin. CD8þ T cells also produce cytokines and similar to Th cells are directed toward four effector lineages: Tc1, Tc2, Tc17, and Treg by exposure

to IL-12/IFNg, IL-4, TGF-b/IL-6, or TGF-b. PAMPs directly or indirectly induce the expression of these polarizing cytokines. For example, bacterial LPS and viral RNA induce DC expression of IL-12. The cellular source of IL-4 and TGF-b have not been defined. IL-6 is commonly expressed in inflammatory environments. These critical cytokines induce or repress the expression of key transcription factors. Th1 cells and CTL express T-bet while suppressing Gata-3 which is critical for Th2 development. Th17 cells do not express T-bet or Gata-3, but express retinoic acid-related orphan receptorgt (RORgt) which is fundamental for their development. Forkhead box P3 (FoxP3) is a signature transcription factor for Tregs. Each transcription factor promotes specific expression of subset-specific genes: Th1/Tc1 cells express IFNg, Th2/Tc2 cells produce IL-4, 5, and 13, Th17/Tc17 cells produce IL-17, and Treg cells produce TGF-b. The primary effector function of Th cells is to provide T cell help through co-stimulatory molecule expression and cytokine production, which orchestrates the immune response by inducing: T and B cell proliferation, Ig heavy chain class switching in B cells, and activation of innate cells. T follicular helper cells (Tfh) are another recently defined Th effector lineage characterized by expression of the chemokine receptor (CXCR5) and localization to B cell follicles within lymphoid tissues. Tfh cells promote B cell proliferation by production of IL-21. Activated B cells initially express IgM heavy chain molecules but upon engagement of Th cells through TCR and CD40L interactions, B cells change their Ig heavy chain isotype in response to Th expressed cytokines. For example, Th1-expressed IFNg or Th2 expressed IL-4 promotes switching to IgG or IgE isotypes, respectively. Antibodies opsonize pathogen surfaces to: neutralize toxins, inhibit pathogen invasion by blocking cell surface receptors critical for binding host cells, and activate complement. In addition, antibody heavy chain isotypes are bound by Fc receptors expressed by innate cells to mediate antibody-dependent cellular cytotoxicity (ADCC). Antibody engagement of Fc receptors expressed by neutrophils (FcgRI), macrophages (FcgRI), NK cells (FcgRIII), and eosinophils (FceR1), induces the release of lytic granules leading to targeted cell lysis. IgG molecules mediate ADCC through macrophages, neutrophils, and NK cells, whereas IgE molecules mediate ADCC through eosinophils. Each Th lineage displays very different effector functions with clear protective and pathological consequences. Th1 cells mediate type IV delayed-type hypersensitivity responses (DTH) through production of IFNg, which activates macrophages to increase their phagocytic activity and induce expression of TNF, NO, and reactive oxygen species. Hence, Th1 infiltration of the pathogen site amplifies inflammation initiated by PAMPs by further activation of macrophages. Th2 cells produce IL-4 and

Adaptive Immune System and the Eye: T Cell-Mediated Immunity

IL-21

CD4+ Tfh

?

CD4+ Th

Fβ

TG

TGFβ

et T-b IL-4

3

CD4+ Treg

IFNγ

CD4+ Th2

Gata-3 IL6/T GF RO β Rγ t

xP

Fo

Nγ

/IF

12

IL-

?

CD4+ Th1

351

CD4+ Th17

IL-4, IL-5, IL-13

IL-17

(a)

CD8+ Tc1 CTL

Granzymes Perforin lFNγ

γ

FN

/I 12

IL-

Fβ

CD8+ CTLp

TG

TGFβ

t be TIL-4

CD8+ Treg

CD8+ Tc2

IL-

6/T

IL-4

GF

β

CD8+ Tc17

IL-17

(b) Figure 3 T-cell effector differentiation. CD4+ Th cells (a) and CD8+ T cells (b) differentiate into distinct subsets based on exposure to particular cytokines which induce unique transcription factor expression and cytokine expression. Th, T helper cell; Tfh, T follicular helper cells; Treg, regulatory T cells; TGFb, transforming growth factor b; FoxP3, forkhead box P3; IL, interleukin; IFNg, interferon gamma; RORgt, retinoic acid-related orphan receptor gt; CTLp, cytolytic T lymphocyte precursors.

IL-5, which control extracellular helminthic infections by activating eosinophils. The recently described Th17 subset produces IL-17 and has been implicated in the response against a wide variety of pathogens requiring a strong inflammatory response predominated by neutrophils. The nonspecific effector mechanisms elicited by macrophages, neutrophils, and eosinophils are associated with significant tissue damage. In contrast, Treg cells inhibit immune responses by expression of CTLA-4 and/or by the production of immunosuppressive cytokines such as TGF-b and IL-10, which inhibits T-cell proliferation and decreases co-stimulatory molecule and cytokine expression of APC. Tregs also suppress cytokine production and lytic activity of T cells, at sites of inflammation. CD4þ T helper cells are also critical for the generation of long-lived memory CD8þ T cells that respond quickly and efficiently upon secondary exposure to antigen. CD8þ T cells that expand in numbers to a pathogen in

the absence of CD4þ T cells rapidly undergo activationinduced cell death (apoptosis) through TRAIL/TRAIL receptor interactions upon secondary exposure to antigen. In addition, memory CD8þ T cells generated in the presence of CD4þ Th cells are not maintained if transferred to mice deficient in CD4þ T cells, indicating that Th cells are somehow necessary to maintain long-lived memory CD8þ T cells in vivo. The primary effector function of CD8þ T cells is control of intracellular pathogens. CD8þ T cells differentiate into lytic effectors (CTL) which are characterized by expression of perforin and granzymes. CTL target infected cells by their expression of pathogen peptides presented on the cell surface by MHC class I molecules. Perforin molecules form pores in cell membranes which facilitate transfer of granzymes into the cytoplasm that activate caspases to induce apoptosis. Interestingly, granzyme B can also inhibit herpes simplex virus-1 (HSV-1) replication within latently infected neurons without

352

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

inducing apoptosis by caspase-mediated cleavage of viral proteins. Whether the neuron or the HSV-1 infection prevents apoptosis under this circumstance remains to be determined. CD8þ CTL also produce IFNg, which inactivates viral proteins and activates innate cells as previously mentioned. Differentiation of T-cell effectors is coincident with the differentiation of long-lived memory T cells. Memory T cells are divided into central (Tc) and effector (Te) memory subpopulations based on their location and unique expression of cell surface molecules. Tc memory cells localize to lymph nodes and spleen through expression of the chemokine receptor7 (CCR7) and CD62L, whereas Te memory cells are present in nonlymphoid tissues and do not express these molecules. Several cell surface molecules have been useful in distinguishing naı¨ve T cells from memory and effector T-cell populations. For example, naı¨ve T cells express low levels of CD44 and killer cell lectin-like receptor G1 (KLRG1) and high levels of the IL-7 receptor (CD127). In contrast, both effector and memory T-cell populations express high levels of CD44. The retention or reexpression of CD127 along with low expression of KLRG1 distinguishes memory T cells from CD127 low KLRG1 hi T cell effectors. The differentiation of effector and memory T cells from a single naı¨ve T-cell precursor is an area of active research and several models are being tested. Lineage determination could involve dedifferentiation of effector cells. Another model suggests that the strength of signal delivered by APC to T cells may favor a particular lineage. For example, T effector generation may be favored when inflammation is high, whereas T memory generation may be favored when inflammation is low. Inverse regulation of T-bet and eomesodermin transcription factors by IL-12 expressed during inflammation may be critical for lineage determination. Alternatively, lineage commitment may be determined upon the first cell division by asymmetric distribution of critical lineage determining molecules to daughter cells.

chamber escaped from the eye and induced systemic T-cell nonresponsiveness. Upon secondary exposure to the same antigen under immunogenic conditions, such as antigen administration with adjuvant, the magnitude of Th1 or Th2 type cell-mediated immune responses was markedly reduced in comparison to mice in which the first exposure to antigen was through a nonprivileged route (skin or conjunctiva of the eye). This phenomenon, termed anterior chamber-associated immune deviation (ACAID), is mediated by CD4þ Tregs, which inhibit the induction of T-cell effector responses and CD8þ Tregs, which inhibit the expression of T-cell effector responses specific for ocular antigens. ACAID may contribute to the tremendous success of corneal transplantation that does not routinely require donor/recipient MHC matching or systemic immunosuppression because mice that accept corneal allografts demonstrate reduced DTH responses to donor antigens. Ocular immune privilege is also maintained by soluble molecules contained within the aqueous humor and by cellassociated molecules expressed on tissues lining the anterior chamber, which are immunosuppressive (Table 1). These molecules directly inhibit T-cell effector function, convert T effectors into Treg, or inhibit the activation of

T-cell immune responses in the eye

VIP (vasoactive intestinal protein) CGRP (calcitonin generelated peptide) CD95 (FasL), TRAIL, and PD-L1 Soluble FasL CD86 (B7-2)

The eye was characterized as a site of immune privilege by the Nobel laureate Peter Medawar in 1948 based on the observation that foreign tissue transplanted in the anterior chamber persisted indefinitely whereas the same tissue transplanted in the skin was rapidly rejected by the host immune response. Based on the absence of demonstrable lymphatic drainage of the interior of the eye, Medawar concluded that ocular antigens were sequestered and the immune system was ignorant of their presence. However, subsequent investigations found that Medawar was incorrect. A reevaluation of ocular immune privilege almost 30 years later by Kaplan and Streilein indicated that soluble and cell-associated antigens injected into the anterior

Table 1 Immune suppressive factors that maintain ocular immune privilege Factor

Effect

TGF-b (Transforming growth factor-beta)

Decreases expression of CD40 and production of IL-12 and increases production of TGF-b by antigen-presenting cells (APC) leading to generation of T-cell hyporesponsiveness and/ or T regulatory cells (Treg) generation. Inhibits T-cell proliferation and T-cell effector function Conditions APC to generate CD4þ Treg Converts CD4þ Teffector into Treg Inhibits neutrophil function Inhibits T-cell proliferation

aMSH (alpha-melanocyte stimulating hormone)

MIF (macrophage migration inhibitory factor)

Inhibits production of nitric oxide by macrophages Induces apoptosis of effectors infiltrating the eye Inhibits neutrophil function Expressed on pigmented cells of the eye Induces functional nonresponsiveness of T cells upon engagement of cytotoxic T-lymphocyte antigen 4 (CTLA-4) Inhibits natural killer (NK) cell activity

Adaptive Immune System and the Eye: T Cell-Mediated Immunity

innate cells. These barriers to T-cell responses are thought to represent a necessary compromise to limit tissue damage during pathogen control. The inhibition of the adaptive immune response would suggest that the eye would be more susceptible to infection. However, this is not the case. Rather, the eye is more prone to immunopathology when T-cell responses break immune privilege to control an ocular pathogen. Several ocular infections induce T-cell responses that promote immunopathology. Viral (HSV-1) and bacterial (Pseudomonas aeruginosa) infections of the cornea induce CD4þ Th1 cells that infiltrate the cornea and orchestrate neutrophil infiltration leading to keratitis. Parasitic Onchocerca volvulus infections induce Th2 responses that contribute but are not alone sufficient to induce keratitis. Interestingly, endosymbiotic Wolbachia bacteria activate TLRs to induce infiltration of the cornea by neutrophils, which are the primary mediators of inflammation. Th17 and Th1 cells have been shown to contribute to immunopathology in experimental rodent models of uveitis. CD4þ Th cells are also critical in the rejection of corneal allografts through induction of a classic DTH response in the cornea and mediate lacrimal keratoconjunctivitis in mice exposed to dessicating stress. The factors that break immune privilege are not completely understood. However, changes in immunoregulatory elements due to inflammation must be involved. Interestingly, in animal studies, the type but not intensity of ocular inflammation has been shown to be critical in breaking immune privilege. For example, the induction of

353

ACAID is preserved in mice with LPS-induced uveitis but not in mice with inflammatory pigmentary glaucoma, although neutrophil inflammation is more pronounced in uveitic mice. Differences in the duration of inflammation may distinguish whether privilege is broken, as LPSinduced uveitis resolves within days whereas glaucoma persists for months. Identification of the regulatory factors affected by pathogens that break immune privilege will be critical for our understanding of how T-cell responses to ocular antigens are modulated to promote the elimination of ocular pathogens while producing minimal tissue damage. See also: Immunobiology of Acanthamoeba Keratitis; Immunopathogenesis of Onchocerciasis (River Blindness); Immunopathogenesis of Pseudomonas Keratitis; Pathogenesis of Fungal Keratitis.

Further Reading Abbas, A. K. and Lichtman, A. H. (2009). Basic Immunology: Functions and Disorders of the Immune System, 3rd edn. Philadelphia, PA: Saunders/Elsevier. Forrester, J. V., Dick, A. D., McMenamin, P. G., and Lee, W. R. (2008). The Eye: Basic Sciences in Practice, 3rd edn. Philadelphia, PA: Saunders/Elsevier. Niederkorn, J. Y. and Kaplan, H. J. (2007). Immune Response and the Eye, Chemical Immunology and Allergy, vol. 92, 2nd, rev. edn. Basel: Karger. Paul, W. E. (2008). Fundamental Immunology, 6th edn. Philadelphia, PA: Wolters Kulwer/Lippincot Williams and Wilkins.

Innate Immune System and the Eye M S Gregory, Schepens Eye Research Institute, Harvard Medical School, Boston, MA, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Endocytosis – The process by which cells absorb molecules, such as proteins, from outside the cell by engulfing them with their cell membrane to form an endosome. Endophthalmitis – An infection of the posterior of the eye. Opsonization – The process by which a pathogen or infected cell is marked for destruction by a phagocyte. Phagocytosis – The engulfment of solid particles, such as bacteria, by the cell membrane to form an internal phagosome.

Introduction Innate immunity comprises a large number of molecules and cells that recognize and respond rapidly to pathogens, providing immediate defense against infection. However, innate immunity also carries with it the potential of highly destructive inflammation that presents an important dilemma for the eye. Inflammation is necessary for successfully eradicating pathogens. An ideal response would eliminate the microorganisms before they are able to directly damage any ocular tissues. The innate immunity would be limited and produce little or no damage to the surrounding normal tissues. However, some types of ocular infections trigger inflammation that is either (1) insufficient to clear the microorganisms, resulting in direct destruction of ocular tissue by the pathogen, or (2) excessive inflammation that clears the microorganisms, but destroys a significant amount of normal tissue. Either of these two scenarios is undesirable and can lead to significant loss of vision. Therefore, a delicate balance must be achieved between the amount of inflammation required for pathogen clearance and the amount of nonspecific tissue damage. The innate immune system of the eye is similar to other mucosal surfaces. The first tier is passive consisting of several anatomic, physical, and chemical barriers that work together to prevent infection without inducing inflammation. The second tier is active consisting of cellular and secretory components that together cause acute inflammation aimed at eradicating the pathogen. The delicate

354

tissues of the eye that make up the visual axis (cornea, lens, and retina) have a very low tolerance for inflammation, as a very small amount of damage can produce a significant loss of vision. The two-tiered system helps to prevent unnecessary inflammation and the active mechanisms of innate immunity are only turned on once the passive barriers have been breached. Both the passive and active arms of ocular innate immunity are the focus of this article.

Passive Innate Defense System Anatomic and Physical Barriers Several anatomic and physical barriers protect the anterior and posterior of the eye from invading pathogens (Figure 1). The active arm of innate immunity is only triggered when pathogens breach these barriers. The cornea is exposed to the external environment, making the anterior segment highly vulnerable to potential pathogen invasion. Therefore, the anterior segment possesses a multilayer barrier system that includes: eyelids and eyelashes, tear film, and the corneal epithelium. By contrast, the posterior segment is not exposed to the external environment, and is therefore less vulnerable to infection. The critical barriers of the posterior segment include (1) the retinal pigment epithelium (RPE), which lies between the blood-rich choroid and the neural retina, and (2) the posterior lens capsule that forms the barrier between the anterior and posterior segments. Each component of the passive innate defense system is described briefly below. Eyelids and eyelashes The outermost barrier of the ocular surface consists of the eyelids and eyelashes. The eyelashes protect the ocular surface from dust and foreign debris. The regular blinking action of the eyelids moves the tears across the ocular surface, washing away potentially colonizing or infecting organisms. Tear film Tears form the second barrier, lubricating and protecting the ocular surface. Tears also posses a potent defense system that limits the growth, colonization, and survival of microorganisms. The tear film consists of three layers: the outermost lipid layer, an aqueous layer, and the inner mucus layer (Figure 2). The lipid layer lubricates the eyelid and slows evaporation of the aqueous tear film layer. The aqueous layer forms the major component of the tear film and contains numerous antimicrobial

Innate Immune System and the Eye

Bowman’s membrane Stroma

a critical component of the passive defense system, while at the same time, upregulation of mucin production and secretion can also be a product of the active arm of innate immunity in the eye.

Descemet’s membrane Eye lid and eye lashes

Corneal epithelium The final barrier of the ocular surface consists of nonkeratinized stratified epithelial cells bound together by tight junctions. The corneal epithelium acts as a physical barrier to invasion of microorganisms due to the presence of epithelial intercellular tight junctions and the rapid renewal of epithelial cells with frequent shedding of the superficial layers of potentially infected epithelium. As mentioned in the previous section, the epithelium also expresses membrane-bound mucins that inhibit bacterial binding to the epithelial surface and produce several of the antimicrobial factors that are present in the ocular tear film.

n

Posterior lens capsule The posterior lens capsule forms a physical barrier between the anterior and posterior segments of the eye after extracapsular cataract surgery and prevents the spread of microorganisms from the anterior chamber into the posterior chamber in the postsurgical eye. The best example of this is the fact that an intact posterior lens capsule is critical in preventing endophthalmitis following cataract surgery. Contamination of the aqueous humor can occur during cataract surgery. However, the pathogens are quickly cleared and endophthalmitis does not develop. By contrast, when the posterior capsule is breached, the rate of endophthalmitis increases significantly. This supports the finding that the anterior segment is much more efficient at clearing bacteria as compared to the posterior segment. Studies suggest the difference in ability to clear pathogens in the anterior versus posterior of the eye may be linked to expression of antimicrobial peptides. One major difference is that the AH is continuously secreted and drained, whereas the vitreous humor is not. The vitreous also offers greater opportunity for microbes to bind its fibrils. However, the exact molecular mechanisms involved remain unclear.

Endothelium

Corneal epithelium

Posterior lens capsule

alle

P.M

Tear film

355

Retinal pigmented epithelium

Figure 1 Anatomic and physical barriers of the eye. The eyelid, eye lashes, tear film, and corneal epithelium serve as barriers of the anterior segment of the eye. The posterior lens capsule and RPE serve as barriers of the posterior segment of the eye.

proteins including: lysozyme, lactoferrin, defensins, secretory IgA (sIgA), and complement. Many of these antimicrobial proteins are constitutively expressed and provide early, broad-spectrum protection against invading pathogens and also prevent the overgrowth of commensal bacteria. The innermost mucus layer of the tear film is made up of secreted and membrane-bound mucins that protect the epithelium from debris, pathogens, and desiccation. Mucins are high-molecular weight glycoproteins characterized by extensive O-glycosylation. Membrane-bound mucins expressed by the ocular surface epithelia include MUC1, MUC4, and MUC16. Secreted mucins are also found in the mucus layer and include MUC2, MUC5AC, and MUC19. The membrane-bound mucins anchor the ocular tear film to the corneal epithelium and are thought to act as a physical barrier against pathogen penetrance. Secreted mucins bind to pathogens in the tear film, facilitating their clearance from the ocular surface. Under normal conditions, mucin production and secretion by goblet cells and corneal epithelial cells are constitutive. However, mucin production can also be induced via Tolllike receptors (TLRs) expressed on the surface of corneal epithelial cells. Moreover, inflammatory cytokines, such as IL-1b, IL-6, and TNFa, have also been shown to induce mucin production and secretion. Together, these data reveal that constitutively expressed mucins make up

Retinal pigment epithelium The RPE consists of a single layer of cells joined by tight junctions that lie between the photoreceptors of the neural retina and the blood-rich choroid. The RPE serves multiple functions aimed at protecting and maintaining the health of the neural retina. RPE cells (1) phagocytose shed disks from the photoreceptor outer segments and recycle their components; (2) transport nutrients from the choroid to the retina; (3) absorb light; (4) provide adhesive properties for the retina; and (5) serve as a rich source of cytokines, chemokines, and growth factors. More recently, RPE have also been linked to immunity and have been

356

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

Lipid layer

Peter Mallen

Aqueous layer

Mucin layer

Epithelium

Lysozyme

LL-37

Defensins

sIgA

C3

C4

Membrane Secreted bound mucins mucins

Protein

Figure 2 Ocular tear film. The ocular tear film is composed of three layers: the outermost lipid layer, the aqueous layer, and the innermost mucus or mucin layer. The lipid layer lubricates the eyelid and slows evaporation of the aqueous tear film layer. The aqueous layer forms the major component of the tear film and contains numerous antimicrobial proteins including: lysozyme, cathelicidin (LL-37), defensins, secretary IgA (sIgA), and complement components (C3 and C4). The mucin layer acts as a physical barrier against pathogen invasion and consists of both membrane bound and secreted mucins.

shown to behave as antigen-presenting cells that phagocytose pathogens, produce cytokines, and present pathogenderived peptides to sensitized T cells. Therefore, while RPE acts as a physical barrier to the posterior segment, it is also important in shaping and regulating the adaptive immune response once this barrier has been breached.

Secretory phospholipase A2 Secretory phospholipase A2 exhibits potent antibacterial activity against Gram-positive bacteria. Similar to lysozyme, secretory phospholipase A2 dissolves bacterial membranes by hydrolyzing the principal phospholipid, phosphatidylglycerol.

Chemical Barriers

Cathelicidin (LL-37) LL-37 is a small cationic peptide with potent antimicrobial activity against Gram-positive and Gram-negative bacteria as well some viruses. The precise mechanism of action is incompletely understood, but it is widely believed that the antimicrobial activity is due to disruption of the microbial membrane or viral envelope.

In addition to the anatomic and physical barriers designed to block invasive pathogens, there are also a number of soluble factors that inhibit bacterial growth, adherence, and survival. Several of these factors are constitutively expressed at low levels, providing a baseline of protection from foreign pathogens. However, many of these factors can also be upregulated in response to pathogens and inflammation; thus, these become an important product of the active arm of innate immunity. Some of the more important factors are discussed in more detail below. Lysozyme

Lysozyme is bacteriacidal and makes up 20–40% of the total tear protein. Lysozyme kills bacteria by (1) binding to and creating pores in the bacterial cell wall, or (2) dissolving bacterial membranes by enzymatic digestion.

Defensins Beta defensins are expressed in epithelial cells that line mucosal surfaces such as the cornea. Similar to LL-37, defensins have a broad spectrum of antimicrobial activity and are effective against: Gram-positive and Gramnegative bacteria, fungi, and enveloped viruses. In addition to their antimicrobial activities, defensins modulate a variety of cellular activities including immune cell chemotaxis, epithelial proliferation, cytokine secretion, and stimulation of histamine release from mast cells. Human

Innate Immune System and the Eye

corneal epithelial cells constitutively express human beta defensin-1 (hBD-1) and hBD-3. By contrast, hBD-2 is induced in response to corneal injury, infection, or inflammation, and is approximately 10-fold more potent than hBD-1 with an even wider antibacterial spectrum. Therefore, while hBD1 and hBD-3 provide a baseline defense to protect the cornea from infection, upon injury or microbial invasion, hBD-2 is upregulated and displays increased antimicrobial activity. Lactoferrin

Lactoferrin is bacteriostatic and binds to and depletes iron from the tear film, which is required for microbial metabolism and growth. Lactoferrin is also bactericidal and permeabilizes membranes of Gram-positive and Gramnegative bacteria. Additional functions of lactoferrin have also been described: (1) inhibits biofilm development, (2) inhibits bacterial adhesion to host cells, (3) inhibits intracellular invasion, (4) amplifies apoptotic signals in infected cells, and (5) enhances bactericidal activity of neutrophils. Lipocalin-A

Lipocalin A prevents bacteria from obtaining iron, an essential nutrient for microorganism survival. However, unlike lactoferrin, lipocalin-A does not bind iron directly. Lipocalin-A inhibits the iron acquisition system of microbes by binding to and blocking microbial sidephores used to transport iron into bacteria. Secretory IgA

sIgA protects the ocular surface against colonization and possible invasion by pathogenic microorganisms by binding to bacteria and facilitating clearance. In addition, sIgA can opsonize bacteria for phagocytosis. Complement

Complement components (such as C3 and C4) are constitutively expressed in the tear film and are involved in phagocytic chemotaxis, opsonization, and lysis of bacteria. The eye is unusual in that there is a constitutive low level of activated complement that is present even in uninfected normal eyes. It is believed that this low level of activated complement provides innate immune surveillance of microbes and allows for a rapid activation of the full complement cascade upon pathogen invasion.

Active Innate Defense System Pattern Recognition Receptors Innate immunity develops rapidly and is described historically as nonspecific, while adaptive immunity develops slowly and is antigen-specific. However, the discovery of pattern recognition receptors (PRRs) that detect unique

357

pathogen-associated molecular patterns revealed a level of specificity in innate immunity that not only allows discrimination between self and non-self, but also allows the development of innate immunity tailored to specific pathogens, such as bacteria, viruses, and fungi. There are two classes of PRRs: (1) TLRs and (2) nucleotide-binding oligomerization domain (NOD)-like receptors (NLRs) (Table 1). TLRs are expressed on the cell surface and detect pathogens at the cell membrane. TLRs are also expressed within endosomes and detect pathogens that have been endocytosed. By contrast, NLRs are expressed within in the cytoplasm and detect the presence of microbial molecules inside the host cell. These two classes of PRRs are discussed in detail below. Toll-like receptors TLRs are type 1 transmembrane proteins with an extracellular domain for ligand binding composed of leucine rich repeats and a cytoplasmic domain for intracellular signaling which is known as the Toll/IL-1 receptor (TIR) domain. TLRs recognize bacteria, viruses, fungi, protozoa, and endogenous ligands, such as heat shock proteins and fibrinogen. Triggering of the TLR leads to activation of the transcription factor, nuclear factor-kappa B (NF-kB), and the expression of pro-inflammatory molecules, such as TNF-a, IL-1, and IL-2. To date, 10 human TLRs have been identified and each TLR has a unique ligand specificity (Table 1). TLRs were first identified on innate immune cells: neutrophils, macrophages, monocytes, and dendritic cells. More recently, TLRs were also identified on epithelial cells that lie at the host/environment interface: skin, gastrointestinal tract, respiratory tract, and urogenital tract. Furthermore, several TLRs have been identified on ocular tissue in both the anterior and posterior segment of the eye, including cornea, iris, ciliary body, choroid, and RPE. Similar to other barrier epithelium, several TLRs have been identified on the RPE and corneal epithelial cells and are vital for sensing microbes and triggering a rapid response to eliminate the pathogen. The study of TLRs in ocular immunity is still relatively new and debate remains over which TLRs are expressed in the eye and whether TLRs are expressed on the cell surface or within endosomes. However, it is clear that TLRs are required for initiation of innate immunity in the eye. This is supported by studies using TLR knockout mice that demonstrate a significant decrease in inflammation characterized by decreased neutrophil infiltration and increased susceptibility to infection in the absence of TLRs. NOD-like receptors The NLRs comprise a large family of cytoplasmic PRRs that recognize bacteria and endogenous danger signals such as uric acid (Table 1). All members of the NLR family share a conserved NOD. However, the NLR family

358 Table 1

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease TLR and NLR expression within the eye Ligand

Location in the eye Cornea, conjunctiva, RPEa Cornea, onjunctiva, RPE Cornea, conjunctiva, RPE Cornea, conjunctiva, iris, ciliary body, choroid, whole retina, RPE Cornea, RPE Conjunctiva, RPE Cornea, conjunctiva, RPE None detected Cornea, conjunctiva, RPE

TLR10 TLR11 NOD1

Triacyl lipopeptides Glycolipds, lipopeptides, lipoproteins, PGN, LTA, HSP70, zymosan dsRNA (viruses) Lipid A (Gram-negative bateria), LPS, bacterial HSP60, RSV coat protein Flagellin Diacyl lipopeptides ssRNA (viruses) ssRNA (virsues) Unmethylated CpG motifs of bacterial DNA, dsDNA (viruses and bacteria) N/d profillin iE-DAP

NOD2 NALP1 NALP2 NALP3 NALP10

MDP Cell rupture N/d Bacterial RNA, toxins and ATP, uric acid N/d

TLR1 TLR2 TLR3 TLR4 TLR5 TLR6 TLR7 TLR8 TLR9

Cornea, conjunctiva, RPE N/d Anterior and posterior portions of the eye,b HCE-T and HCEc Anterior portion, HCE-T and HCE HCE-T and HCE HCE-T and HCE HCE-T and HCE HCE-T

a

RPE: primary cultures of human retinal pigment epithelial cells. Anterior portion contains corneal tissues; posterior portion contains all other ocular tissues. c HCE-T, SV-40 immortalized human corneal epithelial cell line; HCE, human primary corneal epithelial cells. TLR, Toll-like receptor; NLR, NOD-like receptors (caspase recruitment domain-containing NODs and pyrin domain-containing NALPs); PGN, peptidoglycan; LTA, lipoteichoic acid; iE-DAP, g-D-glutamyl-meso-diaminopimelic acid; RSV, respiratory syncytial virus; MDP, muramyl-dipeptide; N/d, not determined. b

can be subdivided into three groups based upon their N-terminal domains: caspase recruitment domain (NODs), pyrin domain (NALPs), or baculovirus inhibitor repeat (neuronal apoptosis inhibitor proteins, NAIPs). At present, the human NLR family comprises of 23 proteins, while at least 34 NLR genes have been identified in mice. NALPs are unique in that upon binding to their ligand, NALPs form a complex termed the inflammasome resulting in caspase-1 activation and the release of active IL-1b. Recently, NALP1, 3, and 10 were identified in corneal epithelial cells, but their potential function in regulating ocular innate immunity has not been determined. NAIPs inhibit caspase effectors and are mainly expressed in neurons where their primary role is to protect against apoptosis. NAIPs have not yet been found in ocular tissues. NODs are the most extensively studied members of the NLR family and are expressed in ocular tissue. Both NOD1 and NOD2 are constitutively expressed within the eye; NOD1 in the anterior and posterior segments of the eye and NOD2 only in the anterior segment. However, which specific tissues express NODs is unknown. NOD1 and NOD2 recognize specific subcomponents of bacterial peptidoglycan. NOD1 recognizes D-g-glutamyl-meso-diaminopimelic acid, while NOD2 recognizes muramyl dipeptide. Recently, a mutation in NOD2 was linked to the development of uveitis in patients with Blau syndrome, suggesting that cytoplasmic

NODs may be important in ocular inflammation. While the importance of NOD receptors in innate immunity and the pathogenesis of inflammatory disease are recognized outside the eye, the study of NODs within the eye is at an early stage. A complete understanding of ocular host defense will require a better understanding of where NODs and other NLRs are expressed and how they regulate innate immunity within the eye.

Complement Similar to PRRs, the complement system acts as an innate immune surveillance system, detecting the first signs of pathogen invasion. The complement system was first identified as a biochemical cascade of serum proteins that help or complement antibodies to clear pathogens and mark them for opsinization by phagocytes. It is now known, however, that the complement system can also be activated directly by microbial products via the alternative pathway. Several studies demonstrate that complement is constitutively active at low levels in the eye. This is thought to be a primary defense mechanism of the eye against pathogenic infection. Upon pathogen invasion the complement system is further activated to clear the infection through (1) generation of inflammatory factors (C3 and C5a), (2) chemotaxis of phagocytes (C3 and C5a), (3) opsonization of Ab-coated cells (C3b), and (4) lysis of bacteria and virus-infected cells (C8 and C9).

Innate Immune System and the Eye

Cytokines, Chemokines, and Effector Cells Once pathogens breach the passive barriers of the eye and trigger the PRRs or the complement system, the primary function of innate immunity is to eliminate the invading pathogen as quickly as possible. To achieve this innate immunity must: (1) trigger an immediate immune response, (2) amplify the response, (3) clear the pathogens, and (4) activate the adaptive system if the pathogen cannot be cleared quickly. Cytokines, chemokines, and adhesion molecules play critical roles in regulating each of these stages. Initiation and amplification

PRRs (TLRs and NLRs) recognize microbial products in the earliest phase of host defense and activate many immune and inflammatory genes, the products of which are important in initiating and amplifying antimicrobial immunity. Triggering TLRs on either the corneal epithelium or the RPE leads to activation of NF-kB via the (1) MyD88-dependent or (2) MyD88-independent pathway. The MyD88-dependent pathway is utilized via most TLRs (TLR1, 2, 4, 5, 6, 7, 8, and 9) and leads to the production of proinflammatory cytokines and chemokines (IL-6, IL-8, IL-18, MIP-1, and TNF-a). By contrast, only TLR3 and TLR4 utilize the MyD88-independent pathway and leads to the production of IFN-a and IFN-b. Recruitment of neutrophils into inflamed tissues is controlled predominantly by two chemokines (MIP-2 (= IL-8 in humans) and KC). In Pseudomonas aeruginosa-induced corneal keratitis, elevated MIP-2 and KC correspond with increased infiltration of neutrophils. In the posterior segment, elevated TNF-a, IL-1b, and CINC (rat homolog of IL-8) contribute to the breakdown of the blood–retinal barrier and the recruitment of neutrophils in response to Staphylococcus aureus. The adhesion molecules ICAM-1 and E-selectin are also upregulated early in iris, ciliary body, and retinal vessels, serving to enhance the infiltration of neutrophils to the site of infection. Clearing the pathogen

In response to inflammatory cytokines and chemokines, neutrophils and macrophages are recruited to the site of infection. Bacterial clearance by neutrophils is accomplished by (1) phagocytosis, (2) generation of reactive oxygen species, and (3) the release of granule-associated enzymes: cathepsin G, myeloperoxidase, lactoferrin, and elastase. While the recruitment of neutrophils to the site of infection is essential for clearance of the pathogen, the persistence of neutrophils and the prolonged release of inflammatory mediators is also associated with nonspecific host tissue damage. Similar to neutrophils, macrophages also phagocytose and directly kill microbes as part of innate immunity. However, macrophages are also antigenpresenting cells, and as such participate in the development of the adaptive immune response. The natural killer (NK) cell is another important innate effector cell in host defense

359

against viral infections of the cornea such as herpes simplex virus type-1. NK cells respond in an antigen-independent manner and kill virus-infected host cells through the release of perforin and granzymes or through binding of the death receptors Fas and TRAIL-R on the target cell. If innate effector cells fail to clear the infection, adaptive immunity will take over and finish eradicating the pathogen. However, if the infection is successfully cleared, the final and most important step of innate immunity is preventing nonspecific host tissue damage. This is accomplished in the eye through several mechanisms that together make up innate immune privilege. Innate Immune Privilege The primary role of innate immunity is to rapidly eradicate invading pathogens through the induction of inflammation. As a general rule, the level of inflammation is proportional to the size and virulence of the infection. Small, nonvirulent infections are cleared by mild inflammation, while larger, more virulent infections induce intense inflammation. The potential danger of inflammation occurs when intense and/or prolonged inflammation threatens the surrounding normal tissue, resulting in nonspecific tissue damage and scarring. The potential danger of inflammation in the eye is magnified by the presence of irreplaceable and highly sensitive ocular tissues. Therefore, it is not surprising that immune privilege in the eye has multiple mechanisms to control innate immunity and limit nonspecific tissue damage. The aqueous humor contains multiple factors that directly inhibit innate immunity including: (1) TGF-b, soluble Fas ligand, and alpha-melanocyte stimulating hormone (inhibit neutrophil activation); (2) macrophage migration inhibitory factor (inhibits NK cell-dependent lysis of target cells); (3) calcitonin gene-related peptide (inhibits nitric oxide release from activated macrophages); and (4) complement regulatory factors: CD46, CD55, CD59, and Crry (inhibit complement activation). These factors work together to limit the damaging consequences of inflammation and to preserve the visual axis. Unfortunately, while these mechanisms evolved to limit local tissue destruction and preserve the visual axis, they may leave the eye more vulnerable to organisms whose virulence often requires a robust inflammatory response for eradication. Therefore, a delicate balance must be made between the amount of inflammation needed for eradicating the pathogen and the amount of nonspecific tissue damage. Link between Innate and Adaptive Immunity Innate and adaptive immunity have long been discussed as separate arms of the immune system. However, it is increasingly clear that they are indeed not separate but highly integrated. Several studies, both outside and inside

360

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

the eye, identified the dendritic cell as a central link between innate and adaptive immunity. Immature dendritic cells reside in peripheral tissues and through triggering of TLRs, participate in the primary immune response against microbial infections. Immature dendritic cells encounter invading pathogens, capture bacterial antigens, and migrate to the draining lymph node. Once in the lymph node, only mature dendritic cells can efficiently prime naı¨ve T cells and initiate adaptive immunity. Studies using TLR and MYD88 deficient mice, reveal that TLR signaling is required for bacteria-induced maturation of dendritic cells and induction of adaptive immunity. Therefore, TLRs on dendritic cells are essential for (1) sensing the microbe and initiating the immediate innate immune response, as well as (2) inducing the development of an adaptive immune response. While dendritic cells have been identified in the cornea and retina, additional studies are needed to completely understand their function in linking innate and adaptive immunity within the eye.

Conclusion Innate immunity is a critical first line of defense against ocular infections. However, the regulation of innate immunity within the eye is just beginning to be unraveled. The identification of TLRs and NLRs has provided new insights into the mechanisms of host defense and the pathogenesis of inflammatory diseases. A better understanding of how microbial agents and endogenous host factors interact with TLRs and NLRs in the eye will be critical in advancing our knowledge of the pathogenesis of infectious and noninfectious eye diseases. Moreover, a better understanding of these mechanisms will lead to the identification of new therapeutic targets for treating and preventing sight-threatening infections. See also: Corneal Epithelium: Wound Healing Junctions, Attachment to Stroma Receptors, Matrix Metalloproteinases, Intracellular Communications; Defense Mechanisms

of Tears and Ocular Surface; Immunosuppressive and AntiInflammatory Molecules that Maintain Immune Privilege of the Eye; Ocular Mucins.

Further Reading Banchereau, J. and Steinman, R. M. (1998). Dendritic cells and the control of immunity. Nature 392: 245–252. Creagh, E. M. and O’Neill, A. J. (2006). TLRs, NLRs and RLRs: A trinity of pathogen sensors that co-operate in innate immunity. Trends in Immunology 27: 352–357. Gregory, M., Callegan, M. C., and Gilmore, M. S. (2007). Role of bacterial and host factors in infectious endophthalmitis. Chemical Immunology and Allergy 92: 266–275. Haynes, R. J., McElveen, J. E., Dua, H. S., Tighe, P. J., and Liversidge, J. (2000). Expression of human beta-defensins in intraocular tissues. Journal of Investigative Ophthalmology and Visual Science 41: 3026–3031. Holtkamp, G. M., Kijlstra, A., Peek, R., and de Vos, A. F. (2001). Retinal pigment epithelium–immune system interactions: Cytokine production and cytokine-induced changes. Progress in Retinal and Eye Research 20: 29–48. Kaplan, H. J. and Niederkorn, J. Y. (2007). Regional immunity and immune privilege. Chemical Immunology and Allergy 92: 11–26. Kawai, T. and Akira, S. (2007). TLR signaling. Seminars in Immunology 19: 24–32. Kolls, J. K., McCray, P. B., and Chan, Y. R. (2008). Cytokine-mediated regulation of antimicrobial proteins. Nature Reviews Immunology 8: 829–835. Pearlman, E., Johnson, A., Adhikary, G., et al. (2008). Toll-like receptors at the ocular surface. Ocular Surface 6: 108–116. Rodriguez-Martinez, S., Cancion-Diaz, M. E., Jimenez-Zamudio, L., et al. (2005). TLRs and NODs mRNA expression pattern in healthy mouse eye. British Journal of Ophthalmology 89: 904–910. Sack, R. A., Nunes, I., Beaton, A., and Morris, C. (2001). Host-defense mechanism of the ocular surfaces. Bioscience Reports 21: 463–480. Sohn, J. H., Bora, P. S., Jha, P., et al. (2007). Complement, innate immunity and ocular disease. Chemical Immunology and Allergy 92: 105–114. Streilein, W. J. and Stein-Streilein, J. (2000). Does innate immune privilege exist? Journal of Leukocyte Biology 67: 479–487. Tosi, M. F. (2005). Innate immune responses to infection. Journal Allergy and Clinical Immunology 116: 241–249. Van Vilet, S. J., den Dunnen, J., Gringhuis, S. I., Geijtenbeek, T. B., and Van Kooyk, Y. (2007). Innate signaling and regulation of dendritic cell immunity. Current Opinion in Immunology 19: 435–440.

Dynamic Immunoregulatory Processes that Sustain Immune Privilege in the Eye J Y Niederkorn, University of Texas Southwestern Medical Center, Dallas, TX, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary ACAID (anterior chamber-associated immune deviation) – A dynamic antigen-specific downregulation of T-cell-mediated immunity that is elicited when antigens are introduced into the anterior chamber of the eye. ACAID is believed to be a mechanism to maintain homeostasis in the eye and reduce the likelihood of T-cell-dependent inflammation in response to noninfectious, nominal antigens that enter the eye. ACAID is also believed to be induced by orthotopic corneal allografts and promotes their survival. Alloantigens – Histocompatibility antigens that provoke immune rejection of organ transplants. These include both major histocompatibility complex antigens and antigens encoded by a wide diversity of minor histocompatibility genes. Delayed-type hypersensitivity – The T-cell- and immune-mediated inflammation that contributes to resistance to intracellular pathogens but also carries a heavy burden of collateral damage to innocent bystander tissues. TGF-b (transforming growth factor-b) – A cytokine that is produced by many cells and has pleiotropic effects on the immune system. TGF-b is present in the anterior chamber of the eye and is crucial for altering the behavior of ocular antigenpresenting cells within the eye such that they induce ACAID. Tregs (T regulatory cells) – These are T lymphocytes that can be induced in multiple ways and suppress T-cell-dependent immune processes. Introducing antigens into the anterior chamber or the vitreous cavity of the eye elicits the generation of T regulatory cells. VCAID (vitreous cavity-associated immune deviation) – Antigens introduced into the vitreous cavity or into the subretinal space elicit an immune deviation that is indistinguishable from ACAID and is characterized by antigen-specific suppression of T-cell-dependent inflammation.

Introduction The immune privilege of the eye was recognized over a century ago by the Dutch ophthalmologist van Dooremaal, who introduced a variety of foreign bodies and tissues into the eyes of animals as a means of studying cataractogenesis, and, in the process, unwittingly discovered the prolonged survival of mouse skin grafts placed in the anterior chamber (AC) of the dog eye. Evidence suggesting that tissue grafts might also enjoy prolonged survival in the eyes of humans surfaced when the first successful corneal transplant in a human subject was reported in 1905, a landmark event that occurred over 60 years before antirejection drugs were used in the first human heart transplant. In the 1940s, experimental pathologists transplanted human tumor cells into the AC of the rabbit eye as a bioassay for determining the malignancy of biopsy specimens. We now know that the survival of such tumor implants in the AC of the rabbit eye was not a property of malignant cells per se, but like the survival of corneal transplants, was a manifestation of ocular immune privilege. It was not until the early 1950s that the preeminent immunologist and Nobel Laureate, Sir Peter Medawar, recognized the significance of the prolonged survival of skin grafts placed into the eye and the brain and coined the term ‘‘immune privilege.’’ Although the concept of ocular immune privilege is widely recognized, it is frequently misunderstood or over-simplified. There are two common misconceptions of ocular immune privilege. The first misconception is that corneal transplants are universally exempt from immune rejection. While it is true that corneal allografts benefit from immune privilege and enjoy a high success rate compared to other categories of transplants, they can undergo immune rejection and, in some cases, the rejection cannot be prevented, even with potent immunosuppressive drugs. The second misconception is that ocular immune privilege is due to the absence of lymph vessels draining the interior of the eye, which would sequester antigens in the eye and deny them access to regional lymph nodes where they would elicit an immune response. Although there are no anatomically detectable patent lymphatics draining the interior of the eye, antigens and cells placed into the AC do in fact reach

361

362

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

the cervical lymph nodes in rodents. Thus, other explanations are needed to account for ocular immune privilege. Important clues as to the possible underlying mechanisms for ocular immune privilege surfaced in the mid-1970s in seminal studies from J. Wayne Streilein and colleagues, who demonstrated that introducing antigens into the AC elicited a deviation in the systemic immune responses resulting in the suppression of T-cell-mediated immunity. This anterior chamber-associated immune deviation (ACAID) is characterized by the antigen-specific suppression of Th1 immune responses, such as delayed-type hypersensitivity (DTH), and Th2-based inflammation, such as experimental allergic asthma in mice. Although antigens introduced into the AC suppress Th1- and Th2based immune inflammation, other immune elements are preserved. Noncomplement fixing antibodies are generated, while complement-fixing antibody production is silenced. Initial studies demonstrated that the induction of ACAID was associated with the production of IL-10, which at the time was considered a Th2 cytokine, and the suppression of the Th1 cytokine, interferon-g (IFN-g). This led some to conclude that ACAID was simply the preferential cross regulation of Th1 immune response by Th2 cytokines. However, further analysis revealed that ACAID also suppressed Th2-based inflammatory responses and that ACAID did not require the participation of a key Th2 cytokine, IL-4. Not only does ACAID suppress classical Th1 immune responses, but it also mitigates Th2-mediated allergic inflammatory lung disease. ACAID also deviates antibody responses by preserving the production of noncomplement-fixing antibodies while blocking the generation of complement-fixing antibodies. This reduces the likelihood of immune inflammation, which is precipitated by activation of the complement cascade. Once activated by complement-fixing antibodies, the complement cascade spins off pro-inflammatory molecules that stimulate granulocytic inflammation. Thus, ACAID suppresses both Th1- and Th2-based inflammation and the generation of complement-fixing antibodies and thereby reduces the likelihood that antigens entering the eye will elicit immune inflammation that could injure tissues in the eye that possess little or no regenerative properties.

Induction of ACAID Studies conducted over the past 30 years have shed light on the molecular and cellular basis of ACAID and have revealed the remarkable complexity of this systemic immunoregulatory phenomenon. The eye, thymus, spleen, and sympathetic nervous system are required for the induction of ACAID. Chemical sympathectomy prior to AC injection of antigen or removal of any of these organs within 3 days of AC injection prevents the induction of ACAID, and in

many cases, allows the development of robust Th1 immune responses, such as DTH. Ocular Phase of ACAID The induction of ACAID begins when antigens enter the AC. The eye is an active participant in the induction of ACAID, as enucleation of the eye within 3 days of AC injection not only prevents the induction of ACAID, but also results in active immunization and development of DTH to the antigens that were introduced into the AC. It is widely believed that within the eye, antigen is captured by F4/80+ macrophages, which under the influence of TGF-b (which is present in the aqueous humor), are imprinted to produce IL-10 while simultaneously extinguishing IL-12 production. The preferential production of IL-10 by ocular macrophages is crucial for the induction of ACAID, as macrophages from IL-10 knockout mice are incapable of inducing ACAID. TGF-b also stimulates ocular macrophages to produce macrophage inflammatory protein-2 (MIP-2), which is a potent chemokine that is pivotal in the splenic phase of ACAID (see below). Thymic Phase of ACAID Within 72 h of antigen entering the eye, F4/80+ ocular macrophages capture antigen and migrate to the thymus. Within the thymus, they evoke the generation of CD4 , CD8 , and NK1.1+ T cells (NKT cells), which then emerge from the thymus and enter the bloodstream where they migrate to the spleen. Other F4/80+ ocular macrophages are believed to migrate directly from the eye to the spleen. Both populations of F4/80+ ocular macrophages contribute to the generation of CD4+ and CD8+ T regulatory cells (Tregs) within the spleen. Splenic Phase of ACAID After entering the spleen, F4/80+ ocular macrophages secrete MIP-2, which attracts CD4+ NKT cells. The NKT cells in turn interact with the ocular macrophages and secrete the chemokine, RANTES, which recruits other cells into the marginal zone of the spleen. Within the marginal zone, F4/80+ ocular macrophages, NKT cells, B cells, and CD4+ T cells, in the presence of the third component of complement, collaborate to generate CD8+ Tregs. The CD8+ Tregs are the end-stage effector cells of ACAID that inhibit Th1- and Th2-based immune-mediated inflammation and promote corneal allograft survival. Sympathetic Nervous System and ACAID All three organs involved in the induction of ACAID – eye, spleen, and thymus – possess dense sympathetic innervations. Many immune responses are influenced by

Dynamic Immunoregulatory Processes that Sustain Immune Privilege in the Eye

the sympathetic nervous system, including ACAID. Animals subjected to chemical sympathectomy prior to AC injection of antigen fail to develop ACAID. It is not clear at what level the sympathetic nervous system exerts its effect, but it appears that it is not at the ocular phase, as chemical sympathectomy does not affect the generation of F4/80+ ocular macrophages. Thus, the induction and expression of ACAID are remarkably complex and require the active participation of multiple organs and organ systems including the circulatory system, sympathetic nervous system, thymus, spleen, and eye (Figure 1). ACAID T Regulatory Cells Two phenotypically and functionally distinct populations of Tregs are induced in ACAID. CD4+ Tregs inhibit the induction of T-cell-mediated immune responses, but do not suppress T-cell effector responses, such as DTH. Thus, CD4+ ACAID Tregs act at the afferent arm of the immune response and prevent the initiation of immune

363

responses, but have no effect if an immune response has already been initiated. CD4+ afferent Tregs are also needed for the generation of CD8+ Tregs that suppress immune responses at the effector stage. CD8+ Tregs are the end-stage regulatory cells of ACAID that block effector immune responses, such as DTH, even if the host has been previously immunized and is capable of mounting a robust DTH response. Both the CD4+ afferent Tregs and the CD8+ efferent Tregs are antigen-specific. It is not clear how the CD8+ ACAID Tregs suppress DTH and other T-cell immune effector responses, but evidence to date indicates that these Tregs do not lyse immune cells via a perforin-dependent mechanism or induce apoptosis by Fas/FasL interactions. Interestingly, CD8+ efferent Tregs express IFN-g receptors and require IFN-g to exert their suppressive properties. However, the precise molecular mechanisms whereby CD8+ efferent Tregs produce their effects remain to be elucidated. Although two distinct Treg populations have been detected in ACAID (i.e., CD4+ and CD8+), it is possible that other Tregs might also participate

Sympathetic innervation?

Thymus Trabecular meshwork d

oo

Bl

Spleen Bl

oo

IL-10 IL-12

F4/80+ APC

Antigen regurgitation

d

Antigen recognition CD25+ CD4+ T cell

? =

CD4+ T cell

NK T cell

IL-10

B cell ? BCR

?

γδ Qa-1 T cell

CD8+ T cell

?

Figure 1 Organ systems and immune cells involved in the induction of ACAID. Removal of the eye, thymus, or spleen within 72 h of AC injection prevents the induction of ACAID. Chemical sympathectomy prior to AC injection of antigen also prevents the induction of ACAID. BCR, B-cell receptor. Reproduced from Niederkorn, J. Y. (2006). See no evil, hear no evil, do no evil: The lessons of immune privilege. Nature Immunology 7: 354–359, with permission from Nature Publishing Group.

364

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

in the induction and expression of ACAID. Although, most, if not all, of the studies performed to date have examined Tregs in the spleens of mice with ACAID, it is entirely possible that Tregs may also be present in lymph nodes, such as the cervical and submandibular lymph nodes. What is the Relevance of ACAID? One might argue that ACAID and immune privilege create an immunological blind spot that renders the eye potentially vulnerable to infectious agents. However, the eye and the brain are part of the central nervous system and possess very limited regenerative properties. Thus, limiting the immunological options to viral or bacterial infections is desirable for the preservation of vision. DTH is crucial for controlling intracellular pathogens, while activation of the complement cascade is an important mechanism for clearing bacterial infections. However, DTH and complement elicit granulocytic inflammation. Granulocytes produce a potpourri of reactive oxygen species and proteases that are notorious for producing extensive collateral damage to innocent bystander tissues. If unrestrained, such inflammation would produce extensive necrosis of ocular tissues leading to blindness. Thus, suppressing DTH and the complement cascade has a clear benefit for preserving the functional integrity of the visual apparatus. The eye is also at risk from other T-cell-mediated immune responses, such as cytotoxic T lymphocyte (CTL)-mediated immunity. During virus infections, viral antigens are displayed on major histocompatibility complex (MHC) class I molecules of the virus-infected cells. The MHC class I/viral antigen complex serves as a docking station for CTL, which kill the virus-infected cells, and thus, eliminate the viral invaders. Although this is a highly efficient mechanism for controlling virus infections in most parts of the body, it would be devastating in the eye. However, CTL responses are suppressed as a consequence of ACAID. Interestingly, corneal endothelial cells and cells of the retina do not express the critical MHC class I molecules that are necessary for recognition and elimination of virus-infected cells by CTL. Thus, the active downregulation of CTL responses to ocular antigens combined with the absence of MHC class I molecules shield ocular cells from CTL-mediated injury. The preferential production of noncomplement fixing antiviral antibodies provides a level of protection by neutralizing viral particles without activating the complement cascade and provoking granulocytic inflammation. Although ACAID seems to be an adaptation to prevent unwitting immune-mediated injury to ocular tissues, it may also benefit the corneal transplant recipient. Animal studies have shown that hosts bearing successful long-term orthotopic corneal allografts display an immune deviation

that is reminiscent of ACAID. Moreover, AC injection of donor cells prior to transplantation results in a dramatic enhancement of corneal allograft survival in rodent models of penetrating keratoplasty.

Vitreous Cavity-Associated Immune Deviation Immune privilege is not restricted to the AC, but appears to be equally expressed in the vitreous cavity and in the subretinal space. Although less is known about the immune privilege in the poster segment of the eye, it is clear that antigens introduced into the vitreous cavity elicit an immune deviation that appears to be identical to ACAID and has been termed vitreous cavity-associated immune deviation (VCAID).

Ocular Regulatory Cells Induced In Situ The aqueous humor contains a myriad of anti-inflammatory and immunosuppressive molecules. Iris and ciliary body (I/CB) cells line a major portion of the AC and secrete constituents of the aqueous humor. I/CB cells not only secrete immunosuppressive and anti-inflammatory molecules such as TGF-b, but they also have the capacity to suppress T-cell proliferation and the secretion of pro-inflammatory cytokines, such as IFN-g, by cell contact-dependent mechanisms, which are independent of soluble anti-inflammatory molecules such as TGF-b, IL-10, and tumor necrosis factor-a (TNF-a ). I/CB cells are strategically located at sites where T cells can enter the eye and, thus, are positioned to exert their influence by inhibiting T-cell proliferation and production of IFN-g shortly after T cells enter the eye. Alpha-melanocyte-stimulating hormone (a-MSH) is one of the numerous immunosuppressive constituents of the aqueous humor. In addition to suppressing the production of pro-inflammatory cytokines, a-MSH stimulates T cells to produce anti-inflammatory cytokines such as TGF-b. Moreover, a-MSH converts Th1 cells into CD4+, CD25+ Tregs, which suppress DTH and mitigate experimental autoimmune uveitis (EAU). The induction of ACAID requires penetration of the eye (i.e., antigen injection via a needle) in order to introduce antigens into the AC. AC injections, even if performed in the least possible traumatic manner, elicit the local production of pro-inflammatory cytokines including TNF-a. Thus, ACAID can be envisioned as a form of immune tolerance induced by antigens entering the AC as a consequence of perforating injuries to the eye. However, the eye may need to regulate immune responses to endogenous ocular antigens. During ontogeny, some structures in the eye are isolated from the immune system and

Dynamic Immunoregulatory Processes that Sustain Immune Privilege in the Eye Table 1

365

Ocular-induced regulatory cells

Induction ACAID VCAID I/CB Endogenous retinal neoantigens EAUa a-MSHb

Antigen pecificity

Organs involved in induction

Regulatory cell phenotype

Yes Yes No Yes

Eye, spleen, thymus, sympathetic nervous system Eye Iris and ciliary body Retina

CD4+; CD8+; CD4+CD25+ ND I/CB cells ND

Yes Yes

Retina/uveal tract AC and retina

CD4+ CD4+

a

T regulatory cells that develop following resolution of EAU. Alpha-melanocyte stimulating hormone-induced regulatory cells induced in situ in the AC or in the retina. ND, not determined.

b

express tissue-specific antigens that, under some circumstances, can initiate an immune response. This has led some to suspect that ACAID and VCAID might be failsafe mechanisms that generate Tregs, which prevent immune responses to tissue-specific ocular antigens (e.g., lens crystallins and retinal antigens). Studies using transgenic mice have demonstrated that novel neoantigens engineered to be exclusively expressed in the retina induce the development of Tregs that suppress DTH responses, yet are significantly different from the Tregs induced by AC injection of antigens. Thus, endogenous ocular antigens arising in situ in an intact eye can elicit the generation of Tregs, which presumably maintain immune homeostasis in the eye and serve as buffers against autoimmune inflammation, although through mechanisms that are likely distinct from ACAID and VCAID.

Conclusions Ocular immune privilege is the product of multiple anatomical and physiological properties of the eye. The blood–eye barrier restricts entry of inflammatory and immune cells into the eye. Moreover, once in the eye, cells of the immune system encounter a milieu that is rich with soluble immunosuppressive and anti-inflammatory molecules. Cells lining the interior of the eye are decorated with membrane-bound molecules such as FasL, programmed death ligand-1 (PD-L1), and tumor necrosis factor-related apoptosis-inducing ligand, each of which can induce apoptosis of activated T cells. In addition, the limited expression of MHC complex molecules on the corneal endothelium and the retina render these cells invisible to CTL. Immune privilege is also maintained by dynamic immunoregulatory processes that are initiated when antigens are introduced into the eye by injection, corneal transplantation, or endogenous ocular antigens. Each of these dynamic immunoregulatory processes relies on

Tregs that act to either prevent the induction or expression of immune processes that inflict injury on tissues with little or no regenerative properties, while preserving immune effector mechanisms that provide protection against pathogens without damaging innocent bystander cells in the eye (Table 1). As stated over two decades ago by the preeminent ocular immunologist J. Wayne Streilein, the eye and the immune system are engaged in ‘‘a dangerous compromise.’’ This compromise protects the eye from unwitting immune-mediated injury at the risk of being vulnerable to ocular infections and perhaps blindness. The reader’s capacity to read through the articles in this encyclopedia is a testament to the success of this compromise. See also: Adaptive Immune System and the Eye: T CellMediated Immunity; Antigen-Presenting Cells in the Eye and Ocular Surface; Immunosuppressive and AntiInflammatory Molecules that Maintain Immune Privilege of the Eye.

Further Reading Ashour, H. M. and Niederkorn, J. Y. (2006). Peripheral tolerance via the anterior chamber of the eye: Role of B cells in MHC class I and II antigen presentation. Journal of Immunology 176: 5950–5956. Caspi, R. R. (2006). Ocular autoimmunity: The price of privilege? Immunological Reviews 213: 23–35. Cone, R. E., Li, X., Sharafieh, R., O’Rourke, J., and Vella, A. T. (2006). The suppression of delayed-type hypersensitivity by CD8+ regulatory T cells requires IFN-g. Immunology 120: 112–119. Faunce, D. E., Sonoda, K. H., and Stein-Streilein, J. (2001). MIP-2 recruits NKT cells to the spleen during tolerance induction. Journal of Immunology 166: 313–321. Faunce, D. E. and Stein-Streilein, J. (2002). NKT cell-derived RANTES recruits APCs and CD8+ T cells to the spleen during the generation of regulatory T cells in tolerance. Journal of Immunology 169: 31–38. Li, X., Taylor, S., Zegarelli, B., et al. (2004). The induction of splenic suppressor T cells through an immune-privileged site requires an intact sympathetic nervous system. Journal of Neuroimmunology 153: 40–49. McKenna, K. C. and Kapp, J. A. (2004). Ocular immune privilege and CTL tolerance. Immunologic Research 29: 103–112.

366

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

Niederkorn, J. Y. (2006). Anterior chamber-associated immune deviation and its impact on corneal allograft survival. Current Opinion in Organ Transplantation 11: 360–365. Niederkorn, J. Y. (2006). See no evil, hear no evil, do no evil: The lessons of immune privilege. Nature Immunology 7: 354–359. Niederkorn, J. Y. and Wang, S. (2005). Immune privilege of the eye and fetus: Parallel universes? Transplantation 80: 1139–1144.

Skelsey, M. E., Mellon, J., and Niederkorn, J. Y. (2001). Gamma delta T cells are needed for ocular immune privilege and corneal graft survival. Journal of Immunology 166: 4327–4333. Stein-Streilein, J. (2003). Invariant NKT cells as initiators, licensors, and facilitators of the adaptive immune response. Journal of Experimental Medicine 198: 1779–1783.

Immunosuppressive and Anti-Inflammatory Molecules that Maintain Immune Privilege of the Eye A W Taylor, Schepens Eye Research Institute, Boston, MA, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Adaptive Immunity – The part of the immune system utilizing T cells and B cells that adapt to specifically target, eliminate, and prevent pathogenic infections. Through adaptive immunity, immunological memory is established to further adapt immunity to mount stronger responses each time the pathogen is encountered. Immune homeostasis – The state of immunity in a stable, unperturbed tissue environment. Immune privilege – The immune status of the ocular microenvironment that has evolutionally adapted itself to prevent the induction of excess inflammation, thereby protecting its delicate structures from the damages of inflammation. It is also defined as any tissue site, such as the brain and eye, which affords survival of incompatible grafts without immunosuppressive therapy. Innate immunity – A more primitive defensive mechanism against infecting pathogens. The activation of innate immunity is through pathogenassociated molecules that bind specific recognition receptors on innate immune cells. Innate immunity is associated with phagocytosis, complement activation, and infiltration of neutrophils and macrophages. Neuropeptides – The low-molecular-weight proteins that are found in, and released from, centrally derived neurons; however, their production is not limited to neurons. They are produced by cells of endocrine glands, immune cells, and cells that make up immune-privileged tissue microenvironments.

Immunosuppression and Anti-Inflammatory Activity in Aqueous Humor The immune-privileged microenvironment of the eye suppresses the induction of inflammation. One of the most dominant mechanisms of this suppression, by which the immune-privileged eye prevents induction of inflammatory immunity, is the manipulation of the functionality of immune cells that enter the ocular microenvironment.

A well-characterized set of neuropeptides that targets specific immune cells and their activities mediates this manipulation. The result of this immunosuppression and immunoregulation is the induction of immune cells that are not only prevented from expressing proinflammatory functionalities, but also regulate other immune cells. The regulatory activity itself is not unique; it occurs at the resolution of immune responses and is part of the mechanisms that prevent the induction of autoimmunity. These are the mechanisms of immune homeostasis that tailor the immune response, and prevent uncontrolled immunity. What is unique is that within the ocular microenvironment these mechanisms are constantly active, and are mediated by constitutively present soluble factors that provide the eye with its unique form of immune homeostasis. Over the past two decades, we have come to understand that the ocular microenvironment is rich with immunosuppressive molecules that influence the activity of immune cells. Many of these molecules are found in the aqueous humor. The first indication that soluble mediators in the eye manipulate the function of immune cells was the finding that aqueous-humor-treated macrophages process and present antigen in a manner that promotes activation of suppressor T cells. This activity is associated with a phenomenon of systemic antigen-specific immunosuppression induced by placing an antigen into the anterior chamber of the eye. This phenomenon is called anteriorchamber-associated immune deviation (ACAID), which may be responsible for the highly successful acceptance of corneal grafts, and may be important in the survival of ocular stem cell and retinal transplants. Macrophages and dendritic cells stimulated with bacterial products that activate innate immune-mediated inflammation fail to mediate inflammation when they are treated with aqueous humor. The expected promotion of macrophages and dendritic-cell-antigen-presenting cell functionality to promote proinflammatory T-cell activation are also suppressed by aqueous humor treatment. Moreover, aqueous-humor-treated macrophages and dendritic cells produce anti-inflammatory cytokines, and present antigens in a manner that promotes immune regulation. Therefore, resident ocular macrophages and dendritic cells, while still able to respond to infectious agents and unhealthy cells, are inhibited from mediating an inflammatory response. It is not completely understood how aqueous humor manipulates the functionality of macrophages. It is possible that ocular resident macrophages function perfectly well in

367

368

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

clearing and defending the ocular microenvironment from a pathogen, but cannot recruit other immune cells to help control the infection through inflammation, or to mediate an effective wound response. T cells treated with aqueous humor are inhibited in proliferation, cytotoxic activity, and production of proinflammatory cytokines. The CD4+ T cells treated with aqueous humor are not only suppressed in their production of the proinflammatory cytokine interferon-gamma (IFN-g), but they also produce the regulatory cytokine, transforming growth factor-beta (TGF-b). This change in cytokine production is associated with a change in T-cell functionality from inflammatory to regulatory. The resistance of the ocular microenvironment to the activation of inflammation mediated by T cells is seen when T cells that mediate delayed-type hypersensitivity are placed into the anterior chamber of the eye along with antigen and antigen-presenting cells. These T cells normally mediate inflammation if injected into the skin, a conventional immune site, with their antigen and antigenpresenting cells; however, in the anterior chamber, the T cells do not mediate inflammation. In addition, treating the T cells in vitro with aqueous humor and then transferring them into the skin fails to produce inflammation. Moreover, aqueous-humor-treated T cells function as regulatory T cells that can suppress the activation of other hypersensitivity-mediating T cells. Therefore, the constitutive immunoregulation and immunosuppressive soluble factors of aqueous humor prevent the activation of inflammatory immunity while turning the immune response onto itself to further regulate the immune response within the immune-privileged ocular microenvironment. There is an ever-growing list of identified factors in aqueous humor that have the potential to suppresses and regulate immunity (Table 1). Neuropeptides form a major group of these factors and most of these factors are found throughout the eye, suggesting that they regulate immunity in all tissue sites of the ocular microenvironment. The factors important for the regulation of T-cell activation and innate immunity in aqueous humor are TGF-b, alphamelanocyte-stimulating hormone (a-MSH), calcitoningene-related peptide (CGRP), somatostatin (SOM), and vasoactive intestinal peptide (VIP) (Figure 1). Other factors, such as Fas ligand (FasL), maybe important in eliminating activated T cells within the ocular microenvironment, and factors such as macrophage migration inhibitory factor (MIF), and complement inhibitors are important in the regulation of innate immunity. In addition to neuropeptides, the retina produces thrombospondin-1 (TSP-1), and pigment epithelium-derived factor (PEDF), which influence the activation and regulation of immunity. It is the collective activity of these molecules within an intact blood–ocular barrier that maintain the unique immune homeostasis of the ocular microenvironment, a process called immune privilege.

Table 1 Immunoregulating and immunosuppressive factors of ocular immune privilege Immune response

Regulation

Factors

Innate immunemediated inflammationa

Suppression

a-MSH, CGRP, PEDF, MIF, CRPe

Adaptive immunemediated inflammation APCb

Suppression

T cellc

Suppression

a-MSH, CGRP, VIP, TGF-b2 TGF-b2, a-MSH, VIP, SOM, FasL

Regulatory immunity ACAIDogenic APCd Treg cells

Induce Induce

TGF-b2, TSP-1 a-MSH, TGF-b2, SOM

a

Production of proinflammatory cytokines and antimicrobial molecules induced by bacterial products or by interleukin-1 and tumor necrosis factor. b Assayed for APC activation of hypersensitivity-mediating T cells. c Assayed for antigen-stimulated T-cell production of proinflammatory cytokines, proliferation, cytotoxic activity, and survival. d Assayed in an adoptive transfer of treated APC for induction of immune deviation. e Complement regulatory proteins.

The Immunoregulatory and Immunosuppressive Factors of The Immune-Privileged Eye Transforming Growth Factor-Beta TGF-b was the first identified immunosuppressive factor in aqueous humor. While it is common to discuss TGF-b in a generic manner, the most interesting aspect of the TGF-b produced in the eye is that a single isoform, TGF-b2, is the dominant form found in aqueous humor, with very little of the other forms of TGF-b being present in the aqueous humor. The dominant expression of TGFb2 over the other isoforms of TGF-b is a common feature of neurological tissues. The significance of expressing TGF-b2 over the other isoforms of TGF-b is not known. In addition, TGF-b is normally expressed in a latent form, requiring activation by other enzymes or binding proteins. It is through TGF-b2 that aqueous humor induces ACAID-mediating antigen-presenting cells (APCs). Treating macrophages in vitro with TGF-b2 induces the characteristics of the aqueous-humor-induced and the in-vivo-induced ACAID APCs. When exposed to TGF-b2, the macrophages increase their expression of the surface marker F4/80, and have reduced expression of coreceptors needed in T-cell activation. The macrophages express anti-inflammatory cytokines of interleukin-10 (IL-10) and activated TGF-b. The activated TGF-b production by the macrophages is probably associated with an

Immunosuppressive and Anti-Inflammatory Molecules that Maintain Immune Privilege of the Eye

suppression of antigen presentation that promotes inflammatory T-cell activation. If there is activation, it results in the generation of immune cells that produce additional TGF-b that further promotes immune privilege.

Cornea Aqueous humor

Alpha-Melanocyte-Stimulating Hormone Pe ter m

alle n

Lens

Iris Ciliary body

TGFβ2

369

Retina

α-MSH

VIP

CGRP

SOM

NPY

Figure 1 Distribution of soluble immunomodulating proteins in the eye. TGF-b, transforming growth factor-beta; a-MSH, alphamelanocyte-stimulating hormone; VIP, vasoactive intestinal peptide; CGRP, calcitonin-gene-related peptide; SOM, somatostatin; NPY, neuropeptide Y. Adapted from Taylor, A. W. (2009). Neuropeptides, aqueous humor, and ocular immune privilege. In: Troger, J., Kieselbach, G., and Bechrakis, N. (eds.) Neuropeptides in the Eye, pp. 79–91. Research Signpost, Kerala, India.

autocrine pathway involving TSP-1. This induction of ACAID-inducing APCs is not limited to the anterior chamber and aqueous humor. Injecting antigen into the retina induces a similar immune deviation involving TGF-b and TSP-1 as well. While the phenomenon of ACAID is an antigen-specific systemic immunosuppression initiated by ACAIDogenic APCs that migrate to the spleen, it is not clear how these APCs may function within the ocular microenvironment. Macrophages or dendritic cells that take up residence with the ocular microenvironment will likely be exposed to some level of activated TGF-b2, and as a result, they express low levels of molecules that promote T-cell action, and simultaneously produce anti-inflammatory cytokines. There is evidence to support this hypothesis. Expression of antigen-presenting molecules within the ocular microenvironment is rarely detected. However, the expression of anti-inflammatory cytokines, such as IL-10, has not been found in abundance in normal aqueous humor. This suggests that while the presentation of antigen by resident APCs in the eye is impaired by TGF-b, other functionalities induced by TGF-b in vitro may not occur within the ocular microenvironment. Some of the suppression of T-cell activation by aqueous humor may involve TGF-b2; however, neutralization of TGF-b2 in aqueous humor does not eliminate all of the T-cell suppression by aqueous humor. Aqueoushumor-treated T cells function as regulatory T cells, which produce TGF-b and suppress the action of other T cells. Therefore, it appears that one of the effects of TGF-b2 in the ocular microenvironment is the

The neuropeptide a-MSH is a 13-amino-acid-long neuropeptide that is derived from sequential endoproteolytic cleavage and post-translational modifications of the protein proopiomelanocortin (POMC) hormone. Originally described for its melanin-inducing activity in frogs, a-MSH has a fundamental role in modulating inflammatory responses. Injections of a-MSH suppress systemic inflammatory responses to endotoxin, and proinflammatory cytokines such as IL-1 and tumor necrosis factor (TNF). Its anti-inflammatory activity is greatest on macrophages, dendritic cells, and neutrophils where it can suppress the induction of reactive oxygen intermediates, nitric oxide, proinflammatory cytokine production, and immune cell migration. It also enhances its own receptor expression and production in macrophages, which in turn, promotes an anti-inflammatory autocrine loop. a-MSH also induces IL-10 production by macrophages and dendritic cells. The anti-inflammatory activity of aqueous humor resembles the anti-inflammatory activity of a-MSH. a-MSH is constitutively present in the aqueous humor in pg ml–1 amounts that are highly antiinflammatory. In aqueous humor, a-MSH has two roles: it suppresses proinflammatory cytokine production by endotoxin-stimulated macrophages and it induces regulatory activity in T cells. While a-MSH suppresses proinflammatory cytokine production, it does not affect antigen presentation other than causing the APCs to present antigen in a manner that does not promote inflammatory T-cell activation. a-MSH-treated macrophages that are stimulated with endotoxin have their intracellular signaling pathway from endotoxin-bound receptors blocked by a-MSH treatment. Thus, instead of a classically activated macrophage producing proinflammatory cytokines and anti-microbial molecules, it produces anti-inflammatory cytokines. This suggests that the ocular microenvironment has a pathway to clear microbial molecules or pathogens without inducing an inflammatory response. Whether this is an effective process is yet to be confirmed. The effect of aqueous humor a-MSH on T cells is profound. It is possible to change an antigen-specific proinflammatory T-cell response to an antigen-specific regulatory T-cell response by treating T cells with aqueous humor or with a-MSH. In this process, TGF-b2 helps in promoting a-MSH induction of regulatory activity in T cells. Flow cytometric analysis of the a-MSH-treated T cells has shown that they express the regulatory T-cell marker CD25 and that this activity is limited to CD4+

370

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

T cells. Unlike other types of regulatory T cells, TGF-b is the only cytokine that appears to be produced by these a-MSH-induced regulatory T cells. Such regulatory T cells can be generated in vitro and used in adoptive transfer experiments to show that they require antigen specificity to activate their suppressive functionality. When their antigen specificity is for retinal autoantigens, they can be used to suppress autoimmune uveitis, and promote retinal allograft survival. Recently, it has been reported that mice that naturally recovered from autoimmune uveitis produce regulatory CD4+ T cells that are specific for retinal autoantigens. Importantly, such CD4+ T-regulatory cells are not detected in mice that do not require the receptor for a-MSH on their T cells. There are four a-MSH receptors that are differentially expressed on immune cells. For T cells, it is the melanocortin 5 receptor (MC5r) that is required for a-MSH to induce regulatory T-cell activity. Interestingly, the retinas of postexperimental autoimmune uveitis (EAU) mice, which do not express MC5r, are severely damaged with losses in photoreceptors. This severe damage is not seen in normal post-EAU mouse retinas. The retina itself also expresses MC5r, and there is evidence that the melanocortin family of proteins is needed for normal ocular development. This suggests that a-MSH, which is produced by several layers of cells in the retina, is important for healthy retinal development, survival, and immune privilege. It has been proposed that immune privilege is an evolutionary adaptation to protect the eye from inflammation, and a-MSH may be one these adaptations of a molecule originally used for other purposes, but now has the added role of being immunosuppressive and immunoregulatory. Other Neuropeptides Three neuropeptides were examined for their presence, and a possible role in aqueous humor immunosuppression because they are found in the neurons that innervate the anterior chamber and some are produced by neural cells of the retina. SOM, CGRP, and VIPs are constitutively present in aqueous humor in ng ml–1 amounts, are expressed in the retina, and target different cells of the immune response. While SOM is found in aqueous humor, its presence was not required for aqueous-humor-mediated immunosuppression. We discovered that SOM induced a-MSH production in T cells that in turn caused the T cells to become regulatory T cells. Therefore, SOM contributes to immune privilege by further promoting production of antiinflammatory and immunoregulating factors by immune cells. There are some contradictory findings regarding the role of SOM in the retina. A protein produced by a retinal pigment epithelial (RPE) cell called PEDF was found to suppress the induction of inflammatory activity in macrophages. However, this suppression was countered by SOM. Activating macrophages by treatment with RPE-derived

factors results in enhanced nitric oxide production, yet the macrophages continue to produce anti-inflammatory cytokines. The meaning for this contradictory finding is not clear. It is possible that macrophages are alternatively activated within the retina and possibly in other tissue sites of the ocular microenvironment. The role of VIP in immune privilege still remains a bit of a mystery. This neuropeptide is in aqueous humor, and it suppresses the activation and proliferation of T cells; however, it does not induce regulatory T cells. The suppression of T-cell proliferation by VIP is only 50%, and there is some speculation that selected populations of T cells are responsive to VIP. Therefore, it is possible that whole aqueous humor selectively expands a subpopulation of T cells that is not responsive to VIP. Maybe these are the regulatory T cells. VIP receptors are present on macrophages, but there is nothing known about the role of aqueous humor VIP on antigen presentation, and on inflammatory activity of macrophages. One problem is that it is not clear whether VIP exists in the ocular microenvironment as a whole molecule or as immunoreactive functional peptide. It is possible that in aqueous humor, or in other regions of the ocular microenvironment, different types of VIP fragments are present and affect different target cells and have different effects on immunity. Whether it is a whole polypeptide or a fragmented peptide, VIP is a contributing factor to the immunosuppression seen within the immune-privileged eye. Unlike the other neuropeptides in aqueous humor, GCRP does not target T cells, but instead, influences macrophage activity. Most mature T cells are unresponsive to CGRP. The CGRP in aqueous humor suppresses nitric oxide generation by macrophages that have been stimulated with endotoxin and IFN-g. Neutralization of CGRP activity in aqueous humor also eliminates aqueous-humor-mediated suppression of nitric oxide production by inflammatory macrophages. The concentration of CGRP in healthy aqueous humor is 20-fold less than its concentration in uveitic aqueous humor. At the higher concentration found in uveitic eyes, CGRP has no effect on inflammatory macrophage function. This suggests that in normal conditions, there is a specifically maintained concentration of CGRP for immunosuppression. Studies on CGRP in aqueous humor have brought to light several issues about the homeostatic environment of the immune-privileged ocular microenvironment. The immune-privileged ocular microenvironment needs to not only maintain a constitutive level of a specific set of factors, but also must maintain them at physiological and functional concentrations. Other Molecules There are several publications that individually describe the immunosuppressive activity of proteins that are not

Immunosuppressive and Anti-Inflammatory Molecules that Maintain Immune Privilege of the Eye

always considered immunosuppressive. Two that have already been discussed above, PEDF acting as an antiinflammatory cytokine, and TSP-1 being important in TGF-b activation and in the process of ACAID, are usually considered anti-angiogenic factors. Their regulation of angiogenesis may be their main function within the eye, but may have evolved like a-MSH, also to function as an immunoregulatory factor. The inflammatory cytokine, macrophage MIF, was considered an important molecule produced by activated T cells to keep macrophages migrating from sites of inflammation. In the eye it has been found to be constitutively expressed in aqueous humor and has a role in preventing natural killer (NK) cells from killing cells not expressing major histocompatibility complex (MHC) antigens. Transformed and injured cells express altered or reduced MHC class I antigens on their surface, and NK cells see this as signal to kill the cell. Since MHC class I molecules are expressed at low levels within the ocular microenvironment, the presence of MIF protects these cells from NK cell attack. The regulation of NK cells and inflammatory macrophages is part of the ocular microenvironment that controls innate immunity. Another component of innate immunity is the complement cascade pathways that release protein fragments that induce migration and activation of immune cells, vascular leakage, and cellular lysis. There are constitutively expressed inhibitors of complement within the eye. In aqueous humor, there are two inhibitors that prevent activation of the alternative complement pathway. Although complement inhibitors are present in the eye, there is evidence that complement activation maybe occurring at a low level that might be necessary to maintain normal aqueous flow. Finally, there is the expression of FasL, a membranebound immunosuppressive molecule that is expressed by cells throughout the ocular microenvironment. When FasL binds to Fas, a molecule on activated immune cells, it triggers programmed cell death. It has been suggested that it is a major contributor to immune privilege and is necessary for graft survival within the ocular microenvironment. However, its expression has not explained the finding that FasL expression can induce neutrophil activation and destruction of corneal grafts. Recently, there is some evidence that FasL exists as a soluble molecule in aqueous humor, and that soluble FasL may block neutrophil activation, and act as an immunosuppressive molecule. This could mean that the balance between soluble FasL and membrane-bound FasL is what is important for maintaining immune privilege in the eye. In addition, surviving allografts of other tissues when engineered to express FasL promote induction of graft-specific regulatory T cells. While expression of FasL may keep activated immune cells at bay, it could be a selective element within ocular microenvironment promoting activation of regulatory immune cells.

371

Application of the Lessons of Immune Privilege The rapidly expanding discoveries of the mechanisms of immune privilege demonstrate that within the ocular microenvironment are active processes for suppressing inflammatory immunity, promulgating alternative activation of immune cells, and mediating regulatory immunity are present. Many of these mechanisms are normally found at various phases of a conventional immune response, especially in the resolution phase, and it appears that the responsiveness of the immune cells to the ocular immunosuppressive and immunoregulating mechanisms is the same whether the cells are taken from the eye or from other tissues. Therefore, it is becoming clear that such mechanisms of ocular immune privilege can be imposed onto immunity to prevent, cure, and establish or reestablish immune tolerance in various hypersensitivity responses, autoimmune diseases, and prevent allograft rejection in tissues other than just the eye. The mechanisms of ACAID and a-MSH-mediated suppression of immunity have been used to demonstrate the potential of applying the lessons of ocular immune privilege as a therapy. The therapeutic approaches for both mechanisms are direct applications that in some ways are personalized therapies tailored to the disease and the patient. The ACAID therapy involves the use of patient’s monocytes, which are treated ex vivo with TGF-b and antigen and reinfused into the same patient. The a-MSH therapy involves either injecting the neuropeptide into the tissue site or collecting the patients’ own immune cells and treating them ex vivo with a-MSH while they are restimulated with autoantigen. The feasibility of these therapeutic approaches has been demonstrated using rodent models of autoimmune diseases, allografts, and hypersensitivity. The therapeutic utilization of the mechanisms of ocular immune privilege is in its infancy, and has a strong potential in being a new direction in immunotherapy.

Conclusion Starting with the first experimental description of ocular immune privilege by Medawar in the 1940s, the understanding of the mechanisms of immune privilege has grown. Along with this understanding is a change in the concept of immune privilege. At first, immune privilege was viewed as an interesting experimental phenomenon that was explained by the unique anatomical features of the ocular microenvironment, which include the anterior chamber’s lack of direct lymphatic drainage and the presence of an ocular–blood barrier. These passive mechanisms suggested that the immune system was ignorant of the presence of antigen within the eye. However, today we

372

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

know that the immune system perceives antigen placed into the eye, but it is the active engagement with the ocular microenvironment that regulates and controls immunity within the eye and systemically to the intraocular antigen. This active engagement is mediated by soluble immunoregulatory and immunosuppressive factors and neuropeptides. Our discovery and understanding of the mechanisms of ocular immune privilege will not only lead to potentially new immunotherapeutic approaches, but may also reveal the mechanisms of how the immune system regulates itself. See also: Adaptive Immune System and the Eye: T CellMediated Immunity; Antigen-Presenting Cells in the Eye and Ocular Surface; Innate Immune System and the Eye.

Further Reading Apte, R. S., Sinha, D., Mayhew, E., Wistow, G. J., and Niederkorn, J. Y. (1998). Cutting edge: Role of macrophage migration inhibitory factor in inhibiting NK cell activity and preserving immune privilege. Journal of Immunology 160: 5693–5696. Cousins, S. W., McCabe, M. M., Danielpour, D., and Streilein, J. W. (1991). Identification of transforming growth factor-beta as an immunosuppressive factor in aqueous humor. Investigative Ophthalmology and Visual Science 32: 33–43. Granstein, R., Staszewski, R., Knisely, T., et al. (1990). Aqueous humor contains transforming growth factor-b and a small (<3500 daltons) inhibitor of thymocyte proliferation. Journal of Immunology 144: 3021–3027. Griffith, T. S., Brunner, T., Fletcher, S. M., Green, D. R., and Ferguson, T. A. (1995). Fas ligand-induced apoptosis as a mechanism of immune privilege. Science 270: 1189–1192. Han, D., Tian, Y., Zhang, M., Zhou, Z., and Lu, J. (2007). Prevention and treatment of experimental autoimmune encephalomyelitis

with recombinant adeno-associated virus-mediated alpha-melanocyte-stimulating hormone-transduced PLP139-151specific T cells. Gene Therapy 14: 383–395. Ng, T. F., Kitaichi, N., and Taylor, A. W. (2007). In vitro generated autoimmune regulatory T cells enhance intravitreous allogeneic retinal graft survival. Investigative Ophthalmology and Visual Science 48: 5112–5117. Nishida, T., Miyata, S., Itoh, Y., et al. (2004). Anti-inflammatory effects of alpha-melanocyte-stimulating hormone against rat endotoxin-induced uveitis and the time course of inflammatory agents in aqueous humor. International Immunopharmacology 4: 1059–1066. Taylor, A. W. (2003). Modulation of regulatory T cell immunity by the neuropeptide alpha-melanocyte stimulating hormone. Cellular and Molecular Biology (Noisy-le-grand) 49: 143–149. Taylor, A. W. (2007). Ocular immunosuppressive microenvironment. Chemical Immunology and Allergy 92: 71–85. Taylor, A. W., Streilein, J. W., and Cousins, S. W. (1992). Identification of alpha-melanocyte stimulating hormone as a potential immunosuppressive factor in aqueous humor. Current Eye Research 11: 1199–1206. Taylor, A. W., Streilein, J. W., and Cousins, S. W. (1994). Immunoreactive vasoactive intestinal peptide contributes to the immunosuppressive activity of normal aqueous humor. Journal of Immunology 153: 1080–1086. Taylor, A. W. and Yee, D. G. (2003). Somatostatin is an immunosuppressive factor in aqueous humor. Investigative Ophthalmology and Visual Science 44: 2644–2649. Taylor, A. W., Yee, D. G., and Streilein, J. W. (1998). Suppression of nitric oxide generated by inflammatory macrophages by calcitonin gene-related peptide in aqueous humor. Investigative Ophthalmology and Visual Science 39: 1372–1378. Wilbanks, G. A. and Streilein, J. W. (1992). Fluids from immune privileged sites endow macrophages with the capacity to induce antigen-specific immune deviation via a mechanism involving transforming growth factor-beta. European Journal of Immunology 22: 1031–1036. Zhang-Hoover, J. and Stein-Streilein, J. (2007). Therapies based on principles of ocular immune privilege. Chemical Immunology and Allergy 92: 317–327.

Antigen-Presenting Cells in the Eye and Ocular Surface P Hamrah and R Dana, Harvard Medical School, Boston, MA, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary ACAID (anterior chamber-associated immune deviation) – Systemic inhibition of delayed-type hypersensitivity reactions to antigens which have previously been placed into the anterior chamber of the eye. APC (antigen-presenting cell) – A cell that displays foreign antigen complex with major histocompatibility complex on its surface. CX3CR1 – CX3CR1 (Fractalkine receptor) is important for homing of Langerhans-like dendritic cells to the corneal epithelium. DC (dendritic cell) – Professional antigenpresenting cells. DC-LAMP/CD208 – A member of the lysosomeassociated membrane glycoprotein (LAMP) family, specifically expressed by mature dendritic cells. DC-SIGN/CD209 – A type 2 transmembrane protein that also contains a mannose-binding (C-type lectin) domain, expressed on dendritic cells. EAU (experimental autoimmune uveitis) – A disease of the neural retina induced by immunization with retinal antigens. F4/80 – Antibody that recognizes both dendritic cells and macrophages. GFP (green fluorescent protein) – Originally isolated from the jellyfish Aequorea victoria that fluoresces green when exposed to blue light. The GFP gene is frequently used as a reporter of expression. Animals have been created that express GFP as a proof-of-concept that a gene can be expressed throughout a given organism. LCs (Langerhans cells) – Dendritic cells that typically reside in the epithelium or epidermis. MHC class II (major histocompatibility complex class II) – These are necessary to present antigen to T cells.

Introduction Antigen-presenting cells (APCs) serve as the immune sentinels to the foreign world and can be subdivided into professional and nonprofessional APCs. In the eye, professional APCs, such as dendritic cells (DCs), epithelial

Langerhans cells (LCs), macrophages, and B cells, are derived from hematopoietic stem and progenitor cells in the bone marrow (BM), forming an integral part of the immune system. Nonprofessional APCs are found among nonlymphoid cells (e.g., vascular endothelial cells, corneal endothelial cells, and keratocytes) and have a low T-cell stimulatory capacity. However, they can gain requisite signals for T-cell priming under certain circumstances (e.g., inflammation). DCs are specialized APCs that play a dual role in inducing adaptive immune responses to foreign antigens and in maintaining T-cell tolerance to self. DCs can also play an important role in innate immunity due to their capacity to respond acutely to inflammatory insults or danger signals in peripheral tissues. DCs consist of several distinct populations that can be differentiated by surface and intracellular phenotypic markers, immunological function, and anatomic location. In mice, DCs variously express the CD11c integrin and MHC class-II (MHC-II) molecules, and are further phenotypically distinguished by their differential expression of CD8a, CD4, and CD11b, as well as a growing list of other new markers. Irrespective of their phenotype and immunological role, DCs exert their activity in the eye remote from their place of origin, where they utilize their advanced migratory skills for navigation. DC progenitors are not restricted to the BM and can be found in multiple locations. These progenitors can differentiate into DCs upon challenge in peripheral tissues. Fully differentiated DCs are found in healthy tissues as immunologically immature cells, being able to sample foreign antigens, but not able to prime naive T cells. Immature DCs express negligible amounts of MHC-II on their surface, and lack the requisite accessory (costimulatory) signals for T-cell activation, such as CD40, CD80 (B7-1), and CD86 (B7-2). In their immature state, they remain alert until signals in the extracellular milieu through inflammatory mediators (derived from microbes or distressed bystander cells) induce a rapid change in function, also known as activation or maturation. Maturation induces redistribution of MHC molecules from the intracellular endocytic compartments of DCs to the cell surface, allowing for T-cell stimulation. Macrophages reside in virtually every tissue, are an integral part of the innate immune response, and synthesize and secrete a variety of powerful biological molecules. They develop from myeloid progenitor cells, enter the bloodstream as monocytes, and migrate into tissues as macrophages. Monocytes are circulating precursors for

373

374

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

tissue DCs and macrophages, being able to maintain or replenish populations in the peripheral tissues during homeostasis. Macrophages express low levels of MHC-II and costimulatory molecules that enable them to act as APCs, even though much less efficient than DCs. Macrophages are generally poorly responsive to activation signals, and also play a role in other processes, including immune regulation and suppression, tissue reorganization, angiogenesis, and lymphangiogenesis. APCs, including macrophages and DCs, are found in a variety of ocular tissues, including the cornea, conjunctiva, iris, ciliary body, sclera, retina, and choroid.

Antigen-Presenting Cells of the Ocular Surface The immune-mediated responses of the ocular surface are influenced by its unique anatomy and physiology. The ocular surface consists of three distinct anatomical regions: the cornea, the limbus, and the conjunctiva that function both independently and in concert as specific barriers against microbial, immunogenic, and traumatic insults. Although the conjunctiva and cornea are anatomically proximate and are bathed in the same tear film, their immune responses are distinctly different from each other. Two populations of BM-derived cells, (1) macrophages or

monocytes and (2) DCs/ LCs, form the main APC arm of the ocular surface immune response (Figure 1).

Corneal APCs Epithelial Langerhans cells During homeostasis, peripheral resident LCs, a subset of DCs, are the only cells that constitutively express MHC-II in the corneal epithelium (Table 1). While a large number of LCs are MHC-IIþ in the periphery, a large population of MHC-II-negative immature LCs are present both in the periphery and the center of the epithelium, with the center being exclusively negative for MHC-II and costimulatory markers. These immature LCs are capable of expressing MHC-II and costimulatory markers during inflammation and migrate to draining lymph nodes (LNs) to present antigen. Phenotypically, both the peripheral and central murine LCs are CD11cþCD11b, with the density of these cells decreasing from the limbus toward the center. These LCs have a classic dendritic morphology with long processes interdigitating among the corneal epithelial cells. In mice lacking the chemokine receptor CX3CR1, homing of immature LCs to the epithelium is markedly impaired. APCs can be observed in living healthy corneas by modern in vivo confocal microscopy. Similar to rodents, the density of APCs declines from the limbus to the

Sclera CD45+, CD11c+, CD11b−, CD8a−, GR-1−, MHC class II−, CD80−, CD86−

CD45+, CD11c+, CD11b−, CD8a−, GR-1−, MHC class II+, CD80+, CD86+

Cell type

Langerhans cell

Markers

Location

CD45+, CD11c+

Limbal, equatorial, peripapillary

CD45+, CD11b+

Limbal, equatorial, peripapillary

Immature

Mature

Normal cornea Iris/ciliary body Cell type

Epithelium

Markers

Location

MHC class II+

Epithelium

F4/80+, MHC class II− Stroma

Stroma

MHC class II+

Stroma

Center Periphery

Periphery

Retina CD45+, CD11c−, CD11b+, F4/80+ CD8−, GR-1−, MHC class II−, CD80−, CD86−

Macrophage

+

CD45+, CD11c+, CD11b+, CD8a−, GR-1−, MHC class II+, CD80+, CD86+ CD40+

Dendritic cell Mature

Cell type

+

CD45 , CD11c , CD11b+, CD8a−, GR-1−, MHC class II−, CD80−, CD86−

Dendritic cell Immature

llen

a P. M

Langerhans cell

Markers

Nerve fiber layer ganglion cell layer, inner plexiform layer, outer plexiform layer

MHC class II+, 33D1+

Juxtapapillary regions (strain variability)

F4/80+, MHC class II−

Perivascular

Macrophage Immature

Key

Mature

Dendritic cell

Microglia

Figure 1 Schematic of antigen-presenting cells of various tissues in the eye.

Location

CD11b+, F4/80+, CD8a+, CD80+, MHC class II+

Antigen-Presenting Cells in the Eye and Ocular Surface Table 1

375

Antigen-presenting cell markers in normal corneal tissue

Tissue/cell type Corneal epithelium (Mouse) Langerhans cells (mature/ immature) Langerhans cells (immature) Corneal epithelium (Human) Langerhans cells Langerhans cells Corneal stroma (Mouse) Dendritic cells (mature/ immature) Dendritic cells (Immature) Macrophages Corneal stroma (Human) Dendritic cells Dendritic cells Macrophages

Markers

Location

CD45þ, CD11cþ, CD11b, CD8a GR-1, MHC class IIþ/, CD80þ/ CD86þ/

Periphery

CD45þ, CD11cþ, CD11b, CD8a GR-1, MHC class II, CD80, CD86

Center/periphery

CD45þ, CD11cþ, CD207þ, CD1aþ, HLA-DRþ, CD11b, DC-SIGN, DC-LAMP CD45þ, CD11cþ, CD207, CD1a, HLA-DR, CD11b, DC-SIGN, DC-LAMP

Periphery Center/periphery

CD45þ, CD11cþ, CD11bþ, CD8a GR-1, MHC class IIþ/, CD80þ/, CD86þ/, CD40þ/ CD45þ, CD11cþ, CD11bþ, CD8a GR-1, MHC class II, CD80, CD86 CD45þ, CD11c, CD11bþ, F4/80þ CD8, GR-1, MHC class II, CD80, CD86

Periphery

CD45þ, mostly CD11cþ, CD11bþ, HLA-DRþ, CD207, CD1a, DC-SIGN, few DC-LAMPþ CD45þ, DC-SIGN, DC-LAMP mostly CD11cþ, CD11bþ, HLA-DR, CD207, CD1a CD45þ, CD11bþ, CD11c, HLA-DR, CD207, CD1a, DC-SIGN, DC-LAMP

Periphery/few central Central/periphery

center in the healthy human cornea. In the corneal limbal epithelium, DCs are present in almost every healthy subject, while in the central cornea only some 20–30% of healthy controls show APCs. LCs are located at a depth of 35–60 mm, mostly at the level of basal epithelial cells and the subbasal nerve plexus. Phenotypically, peripheral LCs in freshly cultured human corneas are Langerin (CD207)þ/CD1aþ/CD11cþ/HLA-DRþ, with no Langerin expression on central LCs. The expression of high levels of CD1a and Langerin on peripheral LCs suggests a unique role of these cells in initiating immune responses to microbial pathogens. Stromal APCs Resident DCs reside in the periphery and center of the anterior corneal stroma. Phenotypically, these DCs are CD11cþCD11bþCD8a demonstrating their monocytic lineage, although a small number of plasmocytoid DCs have been described. Peripheral stromal DCs are MHC-IIþ and positive for the costimulatory markers CD80, CD86, and CD40. The stromal center, however, contains exclusively MHC-IICD80CD86 DCs, similar to those of the highly immature LCs in the epithelium. The density of murine stromal DCs decreases from the limbus toward the center of the cornea. A population of undifferentiated monocytic precursor cells distinct from DC and macrophage populations also resides in the corneal stroma. Thus, in contrast to other organs, where terminally differentiated populations of resident DCs and/or macrophages outnumber colonizing precursors, large numbers of DCs within the cornea remain in a relatively undifferentiated state. The absence of MHC-II

Center/periphery Center/periphery

Periphery/few central

molecules in the normal cornea might actively maintain tolerance to foreign antigens, as antigen presentation to T cells by immature DCs can lead to anergy of T cells and subsequent tolerance, protecting the cornea from immune-mediated damage, when the insults are minor. Resident CD11cCD11bþ corneal macrophages are present in the posterior stroma of the normal mouse and human cornea, and are distinct from the DCs described in the anterior stroma. They are located in the peripheral, paracentral, and central regions. These resident stromal macrophages likely provide a critical first line of defense against pathogens that breach the epithelial barrier of the cornea by producing antimicrobial substances, as well as other inflammatory cytokines and chemokines to attract and activate additional macrophages, neutrophils, and DCs. In freshly cultured human corneas, DCs express DCSIGN and are detected mainly peripherally and in the anterior stroma, having only variable CD11c expression. Most of these cells are HLA-DR, with few mature DCs expressing DC-LAMP/HLA-DR or costimulatory markers. These DCs can be found in the cornea even after long-term culture. DC-LAMPþ mature DCs are only partially DC-SIGNþ, implying that the peripheral stroma harbors two sets of rare mature DCs, those that coexpress DC-SIGN (mostly CD11c) and those that do not coexpress DC-SIGN (CD11cþ). Studies in BM chimera mice have demonstrated a turnover rate for BM-derived cells in the stroma at around 24% at 2 weeks. Replenishment occurs initially in the peripheral cornea and the anterior stroma. By 8 weeks, turnover reaches 75%, reaching a plateau between 2 and 6 months. Close to one-third of migrating cells into the central and peripheral cornea are DCs.

376

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

Corneal APCs in Inflammation Microbial products stimulate the immune system by interacting with toll-like receptors (TLRs) on APCs and other cells. The interaction between TLRs and their ligands activates APCs toward maturity. In addition, the release of proinflammatory cytokines, including interleukin (IL)-1b, GM-CSF, tumor necrosis factor (TNF)-a, and lipopolysaccharides (LPS), or heat-shock proteins from dying cells, creates a microenvironment that activates immature DCs. DCs themselves are also important producers of these proinflammatory cytokines, which act in an autocrine fashion to promote DC activation and maturation. Resident immature epithelial LCs and stromal DCs in the central cornea can significantly upregulate maturation markers, including MHC-II and costimulatory markers, within 24 h after induction of inflammation (Table 2). Table 2

In addition to the resident APC population, APCs are also recruited into the cornea from the limbal areas through upregulation of IL-1 and TNF-a during inflammation (Figure 2). In general, the migration of APCs to peripheral tissues requires the concerted activity of cell adhesion molecules and chemotactic factors. Cell adhesion molecules regulate both cell–cell and cell–matrix interactions, while chemokines provide directionality to local and infiltrating APCs. Suppression of these cytokines leads to downmodulation of APC migration into the cornea. IL-1 and TNF-a can act in concert to recruit APCs from the limbus into the cornea by mediating the expression of cell adhesion molecules and chemokines. Recruitment of macrophages into the cornea plays a crucial role in inducing inflammatory neovascularization by supplying or amplifying signals essential for pathological hemangiogenesis. Macrophages, but not DCs, physically

Antigen-presenting cell markers in inflamed corneal tissue

Tissue/cell type Corneal epithelium (mouse) Langerhans cells (mature/immature) Corneal stroma (mouse) Dendritic cells (Mature/immature) Macrophages

Markers

Location

CD45þ, CD11cþ, CD11b, CD8a GR-1, MHC class IIþ/, CD80þ/, CD86þ/

Periphery/center

CD45þ, CD11cþ, CD11bþ, CD8a GR-1, MHC class IIþ/, CD80þ/, CD86þ/, CD40þ/ CD45þ, CD11c, CD11bþ, F4/80þ CD8, GR-1, MHC class II, CD80, CD86

Periphery/center

Normal cornea CD45+, CD11c+, CD11b−, CD8a−, GR-1−, MHC class II+, CD80+, CD86+

Mature CD45+, CD11c+, CD11b−, CD8a−, GR-1−, MHC class II−, CD80−, CD86−

Immature Langerhans cell

CD45+, CD11c−, CD11b+, F4/80+ CD8−, GR-1−, MHC class II−, CD80−, CD86−

Epithelium

Stroma

Macrophage

CD45+, CD11c+, CD11b+, CD8a− GR-1−, MHC class II+, CD80+, CD86+ CD40+

Mature

n alle P.M

CD45+, CD11c+, CD11b+, CD8a−, GR-1−, MHC class II−, CD80−, CD86− Immature

Dendritic cell

Inflamed cornea

Figure 2 Schematic of the effect of inflammation on corneal antigen-presenting cells.

Periphery/center

Antigen-Presenting Cells in the Eye and Ocular Surface

contribute to lymphangiogenesis under pathological conditions and express lymphatic endothelial markers such as LYVE-1 and Prox-1 under inflamed conditions. Macrophages are capable of forming tube-like structures that express LYVE-1 and podoplanin, and are actively involved in lymphangiogenesis. The Function of APCs in Corneal Transplantation The process of corneal transplant rejection includes an induction phase, called the afferent arm, and an expression phase, called the efferent arm. In the afferent arm the host becomes sensitized to the donor antigens by means of APCs , presenting antigens to T cells. This process can take place through two different pathways. The direct pathway, involving donor APCs that sensitize the host directly and the indirect pathway, involving host APCs that move toward the graft, take up donor antigens, and then present these antigens to T cells in draining LNs. While both direct and indirect alloreactive T cells can mediate graft rejection, host sensitization to donor antigens of corneal grafts occurs through both pathways of sensitization, especially in high-risk corneal grafting, where transplantation occurs in an inflamed bed. CD40 is a critical costimulatory molecule expressed by many APCs (including corneal DCs and LCs), whose ligation by CD154 leads to overexpression of other costimulatory molecules, and IL-12 – critical factors in priming a T-cell response. Blocking the interaction of CD40 with CD40 ligand/CD154 can block both the direct and indirect pathways of allosensitization, by preventing T-cell priming, but without promoting active tolerance. Disruption of the eye–LN axis in the setting of corneal transplantation has been shown to lead to both complete prevention of host allosensitization and the indefinite survival of corneal grafts, demonstrating the functional relevance of corneal APC trafficking to draining LNs. Both donor and host-derived corneal APCs are capable of migrating efficiently to host LNs within 24 h after corneal transplantation. Since the cornea is alymphatic, this is achieved through sprouting of new lymphatic vessels into the cornea upon inflammation, and through migration of APCs toward the lymphatic-rich limbus and conjunctiva. Signaling through vascular endothelial growth factor receptor-3 (VEGFR-3) is critical for DC access to lymphatics, and selective blockade of this pathway can impair DC flow to LNs and the induction of alloimmunity, leading to reduction in the rate of graft rejection. On the one hand, migration of DC to LNs is facilitated by the interaction of the chemokine receptor CCR7 on their surface, with CCL21 secreted by the lymphatic vessels. On the other, CCR1 and CCR5 have been shown to be responsible for the recruitment of immature DCs in inflamed tissues, through their ligands

377

CCL4, CCL5, and CCL7, with only stromal DCs expressing CCR1. Blockade of CCR1 leads to significant reduction in the rate of graft rejection, indicating the important role of stromal DCs in the alloimmune response. Interestingly, depletion of donor APCs before transplantation, however, does not have a significant effect on promoting graft survival, even in the high-risk setting, suggesting that these cells, other than promoting immunization, may also be relevant in the induction of maintenance of tolerance. The Function of APCs in Microbial Keratitis Herpetic stromal keratitis (HSK) is an inflammatory disorder induced by herpes simplex virus (HSV)-1 infection and is characterized by T-cell-dependent destruction of corneal tissues. The number of MHC-IIþ LCs present in the central areas of the cornea has been shown to correlate with the degree of corneal damage. Virally induced migration or maturation of LCs in the cornea precedes the development of HSK. Induction of LC migration into the central cornea before HSV-1 infection results in an accelerated and enhanced delayed-type hypersensitivity (DTH) response to HSV-1 antigens, and in an increased severity of HSK. Contrary, depletion of DCs reduces the incidence and severity of HSK, suggesting a role for DCs in the induction of a T-cell response. These findings have led to the conclusion that HSV-1 infection results in de novo migration of LCs from the limbus, which in turn might play a role in the immunopathology of HSK though presentation of antigens to T cells in the infected cornea. Pseudomonas aeruginosa is an opportunistic pathogen associated with sight-threatening keratitis, whose outcome is largely determined by the host inflammatory response. Specific susceptible mouse strains challenged with P. aeruginosa undergo corneal perforation, while other strains are resistant. While induction of LCs into the central cornea of already susceptible strains before infection does not alter the outcome of disease, induction of LCs into the central cornea of resistant strains converts these to a susceptible phenotype. LCs in these mice express the costimulatory molecule B7-1, enhancing their capacity to present antigen to T cells. Further, macrophages control resistance to P. aeruginosa corneal infection through regulation of neutrophil number and apoptosis, bacterial killing and balancing pro- and anti-inflammatory cytokine levels.

Conjunctival APCs In the naive conjunctiva, predominantly MHC-IIþ DCs are consistently detected from birth in the subepithelial layer and substantia propria. There are species and strain-specific differences in the numbers of these cells, with 200–400 LCs mm2 in humans as compared to 100–150 LCs mm2 in mice, with rat and guinea pig numbers intermediate between these two. Further, the

378

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

number of LCs is not static and increases with age. Moreover, a significant variability in LC density is also found in different regions of the conjunctiva. The largest number of epithelial LCs is found in the palpebral and inferior fornical region, followed by the medial and inferior epibulbar conjunctiva. DCs of the substantia propria are distributed most densely in the superior and medial epibulbar conjunctiva. This variability within the conjunctiva is interesting, and may be related to exogenous antigenic challenge secondary to the direction of normal tear drainage or to microenvironmental differences within the conjunctiva. Finally, macrophages have a density of 6.5 cells/mm2 in the tarsal epithelium and 32.2 cells/mm2 in the tarsal substantia propria, with similar numbers in the bulbar conjunctiva. Role of APCs in Allergic Eye Disease Allergic eye disease is a spectrum of diseases that share a common initiating mechanism and pattern of inflammation. While B cells and T cells are the mediators of the allergic immune response, DCs are the actual initiators and modulators of this response. Mast cells mediate the effector phase of the allergic response, whereas DCs are critical in determining the nature of the allergic response. During the sensitization phase, allergens encounter DCs on the ocular surface. Following allergen challenge, there is a marked influx of conventional DCs and plasmocytoid DCs into the subepithelial layer and throughout the substantia propria. DCs then process and present the allergen to T cells in association with the MHC-II. These T cells are then polarized in favor of the development of allergen sensitivity.

APCs of the Uvea The uveal tract, the vascularized middle layer of the eye, consists of the iris, ciliary body, and choroid. It contains rich networks of F4/80þ APCs that reside and traffic through the eye (Figure 1 and Table 3). These populations include large numbers of macrophages and to a lesser extent, immature DCs, maintaining local immunological homeostasis, and play a role in inflammatory processes and immune-mediated diseases. Dendritiform and pleiomorphic macrophages are distributed in a regular array within the rat iris and ciliary body stroma (600–700 cells/ mm2). The iris contains a network of MHC-IIþ DC Table 3

Antigen-presenting cells of the normal uvea

Tissue/cell type Iris/ciliary body stroma Macrophages Dendritic cells Ciliary body epithelium Dendritic cells

Markers F4/80þ, MHC class II MHC class IIþ MHC class IIþ

(400–600 cells/mm2) within the iris stroma and ciliary epithelium, with few DCs in the uveal tract expressing costimulatory molecules. BM chimera studies demonstrate replenishment of BM-derived cells starting at 2 weeks, with almost complete turnover by 8 weeks in the iris stroma and posterior iris surface. Replenishment rates in the uveal tract are similar in the choroid, iris, and ciliary body stroma, although DCs in the ciliary epithelium replenish at a slightly slower rate starting at 4 weeks. Intraocular DCs, after contact with aqueous humor of the anterior chamber, migrate through the trabecular meshwork to the spleen. Additionally, they begin to secret IL-10 and transforming growth factor (TGF)-b in an autocrine fashion, thereby creating a microenvironment that is rich in tolerance-inducing mediators. These DCs promote the effective suppression of T-cell-dependent inflammatory reactions in the lymphoid organs, inducing sufficient levels of tolerance. DCs in the tissues lining the anterior chamber represent a rich network of APCs and are the most likely candidates for transmitting antigenspecific signals from the anterior chamber in vivo and in experimental models such as anterior chamber-associated immune deviation (ACAID). Anterior Chamber-Associated Immune Deviation The eye receives immune protection against pathogens, while avoiding inflammatory and immunological damage. The selective inability to develop delayed-hypersensitivity responses following antigen invasion into the anterior segment of the eye is highly dependent on DCs, which form the basis of an extraordinary phenomenon called ACAID. The immune response begins with intraocular capture of antigen by specialized ocular F4/80þ APCs in the iris/ ciliary body. ACAID-inducing APCs create a microenvironment rich in TGF-b and IL-10, but deficient in IL-12, thus failing to upregulate CD40. These APCs then migrate through the trabecular meshwork and the venous circulation, preferentially to the marginal zone of the spleen, where they become part of an intricate and highly specific cluster of immune cells. The end result is the emergence of a population of antigen-specific T-regulatory lymphocytes that return to the eye and suppress DTH response. A similar process has been described in the vitreous and other posterior compartments of the eye. Role of APCs in Age-Related Macular Degeneration AMD is the most common cause of legal blindness in elderly individuals of industrialized countries. The presence of complement factor proteins in drusen in AMD eyes and single nucleotide polymorphisms (SNPs) for complement factor regulatory genes in individuals with AMD implicate inflammation as an important component

Antigen-Presenting Cells in the Eye and Ocular Surface

in this disease. Choroidal macrophages are proposed as key players in the removal of age-related accumulation of extracellular debris at the choroidal–retinal interface. Observation of aging mice deficient in CCR2 or CCL-2 indicates that defective clearance or scavenging mechanism by resident choroidal macrophages may, in part, be responsible for the presence of drusen deposits at the choroidal–retinal interface. In addition to the potential role of choroidal macrophages, the discovery of agedependent accumulation of subretinal microglia has recently implicated this population of cells as potential initiators of neovascularization and photoreceptor damage. Role of APCs in EAU Resident APCs of the normal human uvea are endowed with the complete LPS receptor complex and are strategically positioned in perivascular and subepithelial locations for surveying blood-borne or intraocular LPS. LPS may act as an adjuvant by activating APC maturation in the presence of the putative uveitogenic self-antigens and thus mediate the breakdown of peripheral tolerance resulting in the induction of an autoimmune response. APCs that capture self-antigens, present them to autoreactive T cells and induce T-cell tolerance by deletion or anergy, as these APCs are relatively immature. TLRs, however, can convert tolerogenic signals to activating signals by promoting APC maturation. DCs and macrophages act as local APCs in the induction of uveoretinitis. Specifically, MHC-IIþ DCs appear at the time of disease onset and continue to be recruited during the inflammatory process, indicating their role in initiation if EAU. MHC-II macrophages expressing antigens, however, are prominent during the peak phase of tissue damage in the retina and choroid. Depletion of these cells causes a delay in the onset and a reduction in the severity of EAU.

APCs of the Retina The presence of the blood–retinal barrier and a predominantly immunosuppressive intraocular environment contribute to the suppression of local immune responses to retinal antigens. Nevertheless, retinal inflammation is not uncommon. Retinal antigen-specific T cells must encounter cognate antigen on APCs within the retina to initiate retinal inflammation. Several distinct populations of myeloid-derived cells reside in the retina, namely, the more prevalent retinal microglia (CD11bþF4/80þCD8aþ CD80þMHC-IIþ), as well as perivascular macrophages (Figure 1 and Table 4). Perivascular macrophages have poor antigen-presenting capability and are not thought to be absolutely essential for disease induction. Further, a population of BM-derived MHC-IIþ 33D1þ DCs has been identified in mice, of which small numbers reside in

Table 4

379

Antigen-presenting cells of the normal retina

Cell type

Markers þ

Location þ

Retinal microglia

CD11b , F4/80 , CD8aþ, CD80þ, MHC class IIþ

Dendritic cells

MHC class IIþ, 33D1þ

Macrophages

F4/80þ, MHC class II

Nerve fiber layer ganglion cell layer, inner plexiform layer, outer plexiform layer Peripheral margin and juxtapapillary regions (strain variability) Perivascular

the peripheral margin and juxtapapillary regions. Of note, the distribution and phenotype of these DCs within the retinas differs between mouse strains exhibiting different disease susceptibility. In EAU-resistant mice, DCs are MHC-II (low/–). Conversely, DCs are MHC-IIþ in EAU-susceptible mice. Microglia reside in the nerve fiber/ganglion cell layer, inner plexiform layer, and outer plexiform layer of the retina. Resting microglia play various roles in host defense, immunoregulation, and tissue repair and rapidly increase in numbers in response to various insults in the retina. Retinal microglia respond to photoreceptor light-induced injury or degeneration by migration from the inner retinal layers toward the photoreceptor layer and subretinal space, where they phagocytose photoreceptor debris and remain for prolonged periods. Resident host microglia residing in the inner retina are the principal source of the phagocytic microglia that accumulate in the photoreceptor layer and subretinal space during aging or retinal degeneration. BM chimera studies demonstrate reconstitution of myeloid cells in the retina beginning at 4 weeks. Migrating cells are evident at the juxtapapillary margin and migrating deeper into the retinal layers, and almost completely replenish between 2 and 6 months, depending on the strain. Turnover of microglia within the retinal microenvironment occurs at a much slower rate than other peripheral tissue macrophages. When photoreceptor degeneration is induced, large numbers of microglia/macrophages are observed in the injured retina, starting at 12 h after injury, and peaking at 24 h. In addition, the number of MHC-IIþ cells in the retina increases greatly after retinal injury. In response to retinal damage, numerous BM-derived cells migrate to the retina from the ciliary body, optic nerve, and retinal vessels and differentiate into microglia. The higher rate of immunologic activation and the increased specificity to the damaged site appear to be the characteristic features of BM-derived microglia.

APCs of the Sclera BM-derived cells have been described in the sclera. In BM chimeras, BM-derived cells replenish the sclera

380 Table 5

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease Antigen-presenting cells of the normal sclera

Cell type

Markers þ

Location þ

Dendritic cells

CD45 , CD11c

Macrophages

CD45þ, CD11bþ

Limbal, equatorial, peripapillary Limbal, equatorial, peripapillary

through limbal vessels and optic nerve vessels, migrating into the equatorial zone. These cells are CD11cþ or CD11bþ DCs and macrophages are found among the scleral fibroblasts (Figure 1 and Table 5). During EAU, significant infiltration of these BM-derived cells takes place into the sclera, contributing to the ocular immune response.

Conclusions The integrity of the visual system in the face of everchanging immune challenges is vital. The unique immune homeostasis and immunological status of the eye and ocular surface is fascinating and continuously evolving. It is compelling that most of the APCs described herein were only discovered less than 10 years ago, emphasizing how much still has to be learned about APCs and their function in the eye. The constitutive presence of APCs in the ocular tissues has significant implications for a variety of infectious, autoimmune, and inflammatory responses in the eye. Since the presence of resident ocular APCs was largely unknown until very recently, many paradigms have already been shifted and many more will need to be rethought in the future. Understanding the mechanisms

that lead to APC maturation, activation, and trafficking may well lead to novel approaches in the induction of tolerance, autoimmunity, and vaccine therapy. See also: Adaptive Immune System and the Eye: Mucosal Immunity; Adaptive Immune System and the Eye: T Cell-Mediated Immunity; Dry Eye: An ImmuneBased Inflammation; Dynamic Immunoregulatory Processes that Sustain Immune Privilege in the Eye; Immunosuppressive and Anti-Inflammatory Molecules that Maintain Immune Privilege of the Eye; Innate Immune System and the Eye; Penetrating Keratoplasty.

Further Reading Dana, R. (2004). Corneal antigen-presenting cells: Diversity, plasticity, and disguise: The Cogan lecture. Investigative Ophthalmology and Visual Science 45: 722–727. Hamrah, P. and Dana, R. (2007). Corneal antigen-presenting cells. Chemical Immunology and Allergy 92: 58–70. Hamrah, P., Huq, S. O., Liu, Y., Zhang, Q., and Dana, M. R. (2003). Corneal immunity is mediated by heterogeneous population of antigen-presenting cells. Journal of Leukocyte Biology 74: 172–178. Kezic, J. and McMenamin, P. G. (2008). Differential turnover rates of monocyte-derived cells in varied ocular tissue microenvironments. Journal of Leukocyte Biology 84: 721–729. McMenamin, P. G. (1999). Dendritic cells and macrophages in the uveal tract of the normal mouse eye. British Journal of Ophthalmology 83: 598–604. Novak, N., Siepmann, K., Zierhut, M., and Bieber, T. (2003). The good, the bad and the ugly – APCs of the eye. Trends in Immunology 24: 570–574. Streilein, J. W. (2003). Ocular immune privilege: Therapeutic opportunities from an experiment of nature. Nature Reviews Immunology 3: 879–889. Xu, H., Dawson, R., Forrester, J. V., and Liversidge, J. (2007). Identification of novel dendritic cell populations in normal mouse retina. Investigative Ophthalmology and Visual Science 48: 1701–1710.

Dry Eye: An Immune-Based Inflammation M E Stern, Allergan Inc, Irvine, CA, USA S C Pflugfelder, Baylor College of Medicine, Houston, TX, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Autoimmunity – An immune response of an organism against any of its own tissues, cells, or cellular components. Diseases, such as rheumatoid arthritis, systemic lupus erythematosus, multiple sclerosis, and dry eye are considered autoimmune based. CD4þ T cells – T helper cells (Th), a subgroup of lymphocytes that play an important role in establishing and maximizing the capacity of the immune response against invading extracellular pathogens, for example, bacteria and parasites. The CD4þ T cells bearing T-cell receptors that recognize self-antigen, that is, autoantigen, contribute to the immunopathogenesis of several autoimmune diseases, for example, rheumatoid arthritis, multiple sclerosis, and dry eye. Dry eye – An ocular surface immune-based inflammatory disease resulting from an unstable tear film composition mediated by dysfunction of a complex Lacrimal Function Unit (LFU: cornea, conjunctiva, lacrimal glands, and meibomian glands), which causes damage to the interpalpebral ocular surface and is associated with symptoms of ocular discomfort. Dry eye is also known as lacrimal keratoconjunctivtis (LKC) and most recently, dysfunctional tear syndrome (DTS). Desiccating stress (DS) – Following exposure to DS in low humidity (<40%) mice display similar clinical and histopathological features to human patients with dry eye, including rapid and coordinated upregulation of proinflammatory cytokines, decreased tear production and goblet cell number, surface epithelial apoptosis, and increased cellular infiltration, for example, CD4þ T cells into the LFU. Goblet cells – Glandular simple columnar conjunctival epithelial cells that function to secrete mucus. Lacrimal glands – Glands located in the upper, distal portion of the orbit of each eye that secrete the aqueous layer of the tear film. Lacrimal function unit (LFU) – The lacrimal functional unit is composed of the lacrimal glands (both main and accessory), the ocular surface (cornea, conjunctiva, goblet cells, and meibomian glands), and the interconnecting innervation that coordinates afferent (ocular surface to the brain) and efferent (brain to the ocular surface tissues and associated glands) signals. The LFU is responsible

for maintaining the quantity and quality of the tear fluid. Meibomian glands – Sebaceous glands at the rim of the eyelids responsible for the supply of sebum, an oily substance that prevents evaporation of the eye’s tear film. MHC class II (major histocompatability complex class II) – This is responsible for presenting antigen fragments to T helper cells by binding exclusively to the T-cell receptor present of the surface of CD4þ T cells. The MHC class II is involved in presentation of antigen derived from extracellular pathogens, thus providing specificity for the generation of adaptive immunity. The MHC class II molecules bearing selfantigen (autoantigen) may trigger activation of autoreactive CD4þ T cells and autoimmunity.

Defining the Problem Epidemiology of Dry Eye In 1993 the National Eye Institute (NEI)/Industry workshop formally defined dry eye as a ‘‘disorder of the tear film due to tear deficiency or excessive evaporation, which causes damage to the interpalpebral ocular surface and is associated with symptoms of discomfort.’’ Recently, we and others proposed a more comprehensive definition of dry eye based on the increasing evidence demonstrating that ocular surface immune-based inflammation and ocular surface epithelial diseases result from dysfunction of a complex Lacrimal Function Unit (LFU) and the resultant unstable tear film. To provide guidelines for selection of treatment, the Delphi panel of experts coined the term dysfunctional tear syndrome (DTS) based on symptoms and signs (not tests) of dry eye disease. Dry eye is a highly prevalent condition and one of the leading causes of visits to ophthalmologists and optometrists in the United States. Epidemiologic studies have reported that dry eye affects up to 11% of people 30–60 years of age and 15% of those 65 years of age or older. As many as 12 million Americans have moderate to severe dry eye and this number is likely to increase as the population ages. Dry eye affects 0.1–33% of the worldwide population; the wide range of variability is dictated by the study and diagnostic criteria used.

381

382

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

Dry Eye Syndrome in Peri- and Post-menopausal Women Dry eye is more common in women than men (2:1) and the incidence increases with age. The role of sex hormones in dry eye has been reported in several studies. For example, data from 36 995 female health professionals ranging from age 49 to 89 years old (estimated at least 3.2 million (or 7.8%) women aged 50 years and older) suffer from dry eye disease in the US. The incidence appears to be higher among older women (9.8%, 75 years) than women under 50 years of age (5.7%). The incidence of clinically significant rosacea is also higher among aging females and occurs in approximately 30% of menopausal women. In fact, it is predicted that up to 75% of peri-menopausal women with facial rosacea will develop ocular involvement. Estrogen may have detrimental effects on the tear film and could influence the development of dry eye; although the ratio between estrogens and androgens may be a better indicator. It was reported that women who receive hormone replacement therapy, especially with unopposed estrogen therapy, have an increased risk of developing dry eye disease. On the other hand, androgen deficiency and/or imbalance in estrogen–androgen levels are also associated with dry eye. Along these lines, androgen deficiency as seen in Sjo¨gren syndrome and Sjo¨gren’s syndrome keratoconjunctivitis sicca (KCS) occurs almost exclusively in women. Furthermore, women who suffer from premature ovarian failure lack both estrogens and androgens and exhibit more ocular surface damage and dry eye-related symptoms than age-matched controls. Patients on Anti-androgen Therapy Androgenic hormones play an important role in supporting the secretory immune function of the lacrimal glands and meibomian glands. The meibomian glands are the main androgen target organs on the ocular surface. Androgen deficiency that may occur during menopause, aging in both sexes, autoimmune disorders (e.g., Sjo¨gren’s syndrome, Systemic Lupus Erythematosus, rheumatoid arthritis, RA), complete androgen insensitivity syndrome (i.e., women with dysfunctional androgen receptors, congenital androgen insensitivity syndrome), and the use of anti-androgen medications (e.g., for prostatic cancer or hypertrophy) is associated with meibomian gland dysfunction, tear film instability, and a significant increase in dry eye signs and symptoms. Studies showed that anti-androgen treatment is paralleled by significant changes in the fatty acid profiles of neutral lipid fractions in meibomian gland secretions. Conversely, treatment with androgens has been reported to alleviate dry eye conditions and stimulate tear flow in Sjo¨gren’s syndrome patients.

Clinical Features of Dry Eye Chronic Pain Ocular surface neuropathy in dry eye Ocular surface pain and discomfort in severe sicca disease may partially result from the well-documented neuropathy associated with Sjo¨gren’s syndrome, which is categorized with the neuropathies associated with connective tissue disease. Indeed, clinical evidence has shown that peripheral sensory neuropathy may be an important presenting sign for Sjo¨gren’s patients. In accordance, ocular surface discomfort is often the initial motivation for dry eye patients to visit the ophthalmologist. In affected individuals, ganglioside-specific antibodies are found in peripheral nerves, dorsal root ganglia, and dorsal roots, and inflammatory cells are localized within the ganglia. Although the trigeminal system has not been studied in as much detail, the ocular surface discomfort of dry eye may be a form of sensory neuropathy; however, this theory requires confirmation. Small diameter myelinated and unmyelinated axons in the cornea are potential targets for peripheral nerve disorders, and inflammatory cells infiltrating the ocular surface are well documented in dry eye. These cells, in combination with ganglioside-specific antibodies and other neural proteins, could cause local degeneration of small diameter axons and axon terminals. Cranial neuropathies may be more common in Sjo¨gren’s syndrome than is currently recognized, and the dysthesias associated with the cornea may indicate an inflammatory neuropathy within the trigeminal system.

Comorbidities Patients with lacrimal keratoconjunctivitis (LKC) typically experience ocular discomfort. The most common symptoms include scratchiness, grittiness, foreign body sensation, burning, and itching; these symptoms are exacerbated by prolonged visual activity (e.g., viewing a video display terminal) and environmental stresses, such as low humidity and air drafts. The LKC patients often complain of blurred and fluctuating vision that stimulates increased blink frequency, an unconscious response to clear the visual field. Together, these symptoms contribute to severe ocular fatigue and many patients report that they are unable to read or concentrate for more than a few minutes at a time. Also, LKC can cause considerable ocular morbidity. The thinned and unstable precorneal tear layer and the altered corneal epithelial barrier function that accompany LKC are major risk factors for sterile keratolysis (loss of uppermost layer of cells in cornea) and microbial keratitis (infection of the cornea). Severe and recurrent corneal ulceration mediated by LKC can ultimately lead to reduced vision, blindness, and in severe cases, loss of the eye. Pre-existing LKC is an important cause of complications following corneal surgery. Complications include

Dry Eye: An Immune-Based Inflammation

How We Secrete Normal Tears

penetrating keratoplasty and LASIK, and may lead to decreased vision, pain, epithelial and stromal wound healing problems, haze, ulceration, and predisposition to microbial infections. Surgical amputation of the corneal sensory nerves that drive glandular secretion, a direct consequence of LASIK and other refractive procedures, negatively impacts the integrated ocular surface secretory gland functional unit. This exacerbates pre-existing LKC and most likely results in new cases of dry eye.

The Lacrimal Functional Unit The ophthalmic pathology seen in Sjo¨gren’s syndrome and chronic dry eye surrounds an immune-based inflammatory disruption of the LFU (Figure 1). The LFU is composed of the ocular surface (cornea, conjunctiva, conjunctival blood vessels), the lacrimal glands (main and accessory (Wolfring and Krauss)), and the interconnecting innervations (V, VII). This tear secreting reflex is also modulated by hormonal and immune factors. The role of the LFU is to secrete a precise tear film composition that maintains a homeostatic environment around the epithelial cells of the ocular surface.

Quality of Life Impact The LKC symptoms significantly impact quality of life documented by utility scores. Utility scores quantify how many years a subject would give up from the end of his/her life in exchange for avoiding a particular malady. Utility scores for dry eye were found to be similar to those from patients with angina. The chronic and unremitting nature of dry eye syndrome can lead to despair, depression, decreased productivity, and in some cases permanent job disability. The physical and psychological impact of LKC symptoms is similar to that experienced by patients with other chronic regional pain syndromes, such as those affecting the lower back.

Carotid artery

383

The general role of the LFU in homeostasis and disease

The purpose of the tightly controlled ocular surface environment is to preserve corneal clarity and vision. The main and accessory lacrimal glands, the corneal limbus, and the meibomian glands provide the vital supportive function to protect the sensitive epithelial surfaces of the conjunctival and corneal tissues from environmental

Long ciliary nerve Nasociliary nerve Short ciliary nerves Lacrimal nerve

Frontal nerve

V Nucleus V1

Sup.salivatory nucleus

Ciliary ganglion

V2 V3

VII motor nucleus

Geniculate ganglion

Sphenopalatine ganglion

Infraorbital nerve

Conjunctival afferents

Afferent sensory fibers Efferent parasympathetic fibers Efferent sympathetic fibers

Figure 1 The lacrimal functional unit. Subconscious stimulation of the free nerve endings within the cornea generates afferent nerve impulses through the ophthalmic branch of the Trigeminal Nerve (V) to the midbrain (pons). The afferent signals are integrated in the midbrain and then travel via the efferent branch through the pterygopalatine ganglion, terminating in the main and accessory (Wolfring and Krause) lacrimal glands. Evidence suggests that this pathway also controls secretion from meibomian glands and conjunctival goblet cells. Proper function of the LFU supports homeostasis on the ocular surface by controlling secretion of the three major tear film components (mucin, aqueous, and lipid) to maintain the optimal quantity and quality of tear fluid; however, dysfunction of the LFU may lead to altered tear film composition and dry eye disease.

384

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

injury that results in pain and decreased visual acuity. The function of the LFU is to control secretion of tear constituents that help sustain a stable, anti-infective, and epithelial supportive tear layer essential for optimal optical performance. Signals emanating from ocular surface sensory nerves supply continuous input into the CNS that tells the brain what changes are occurring within the ocular surface milieu. The brain then sends signals to the specialized support tissues, for example, lacrimal and meibomian glands that are programmed to secrete the optimal tear quantity and composition. The process by which normal tears are secreted is initiated following corneal nerve stimulation. The process occurs unconsciously and in response to many stimuli; however, environmentally induced dry spot formation is thought to be among the most common. Out of necessity for survival, evolution has shaped the cornea to become the most densely sensory nerve innervated epithelial surface in the body. Conduction of pain originates from myelinated and unmyelinated nerves that terminate in the cornea, limbus, and conjunctival epithelium. Neural receptors in the cornea are free nerve endings that terminate in all of the corneal epithelial layers and are protected from direct irritation by zonula occludens and the tear mucin gel. Afferent (ocular surface to the brain) nerve traffic through the ophthalmic branch of the Trigeminal Nerve (V) enters the central nervous system in the area of the pons (midbrain) and the para spinal sympathetic tract. These signals are integrated with cortical and other inputs and are then transmitted to the efferent (brain to ocular surface tissues and associated glands) secretomotor impulses resulting in secretion of the homeostatic tear-film components. Tear secretion by the lacrimal gland also occurs in response to neural stimulation. The acini, ducts, and blood vessels of the lacrimal gland are innervated by parasympathetic, sympathetic, and sensory nerves. The initial signal originates from the parasympathetic cholinergic nerves via acetylcholine release, which then binds to M3 muscarinic acetycholine receptors on the basolateral cell membrane of secretory epithelia. At the same time, vasoactive intestinal peptide (VIP) binds to VIPergic receptors, and norepinephrine, a sympathetic neurotransmitter binds to a1- and b-adrenergic receptors. Neural innervation of the accessory lacrimal glands has also been reported and fibers positive for CGRP and substance P are associated with secretory tubules, interlobular and excretory ducts, and blood vessels. However, the degree of neural influence over accessory lacrimal glands is still being elucidated. Sensory, sympathetic, and parasympathetic neuropeptides are present in the ocular surface tissues and associated glands. Conjunctival goblet cells have a secretory response to the parasympathetic cholinergic muscarinic output from the pterygopalatine ganglion. Goblet cells express M3-muscarinic receptors on their membranes. The M1 and M2 receptors are located throughout the

conjunctiva. The presence of a1A- and b3-adrenergic receptors on conjunctival goblet cells suggests the presence of sympathetic innervation. In addition, transmission electron microscopy of meibomian glands revealed the presence of unmyelinated axons with granular and agranular vesicles. Substance P- and CGRP-positive axons have also been identified, but their function is uncertain, as these neurological peptides would be expected to conduct information into, rather than away from, the CNS. It is predicted that parasympathetic fibers innervating the meibomian glands are indeed present at higher levels. Parasympathetic neurotransmitters neuropeptide Y and VIP have been found around the meibomian glands, as well as tyrosine hydroxylase in sympathetic axons, implicating that both types of autonomic nerves may play an important role in stimulating lipid secretion onto the ocular surface. Patients with LKC commonly complain of constant corneal sensations normally described as a gritty, sandy, or itchy. These complaints are usually accompanied with a pathophysiological state that indicates a chronic state of inflammation and a disadvantageous change in tear film composition. Infiltrating inflammatory cells within the ocular surface tissues have been reported in dry eye. These inflammatory cells, in addition to ganglioside-specific antibodies and other neural proteins, may result in regional degeneration of small diameter axons and their terminals. Chronic dysfunction of the LFU results in a shift toward inflammation and persistent psychological distress.

Events on the Ocular Surface Environmental Impact on the Ocular Surface The immune response is designed to defend against stress and/or microbial assaults on the ocular surface and paradoxically may also contribute to autoimmunity. Along these lines, regulatory mechanisms have evolved to modulate the afferent and efferent arms of the immune response to preserve tissue and limit activation of autoreactive lymphocytes following acute inflammation. The afferent events leading to cellular immunity include antigen processing and presentation by ocular surface antigen presenting cells (APCs) and migration of these cells to the draining lymph nodes. Afferent immune processes are modulated by a wide variety of anti-inflammatory factors that include cellular, for example, T regulatory cells (Tregs), and diffusible factors, for example, transforming growth factor beta (TGF-b) and interleukin (IL)-1 receptor antagonist, that favor protective immunity without breaking self-tolerance. The afferent arm of the immune response dictates the efferent response, the phase involved with antigen driven homing of primed and targeted lymphocytes to tissue-specific inflammatory sites. The efferent response is initiated in the secondary

Dry Eye: An Immune-Based Inflammation

lymphoid organs and amplified on the ocular surface via cell-to-cell interactions between lymphocytes and APCs; activation and differentiation of lymphocytes and migration to the ocular surface is tightly regulated to control the efferent immune response. Indeed, immunoregulation on the ocular surface is the result of the coordinated effort between a wide variety of immune players. However, when these mechanisms are compromised the ocular surface may become susceptible to chronic and/or autoimmunemediated ocular disease, such is the case with dry eye. Afferent arm of the immune response: immunoregulation

Redundant mechanisms regulate the afferent immune response to guard against activation and infiltration of autoreactive lymphocytes to the ocular surface tissues. For instance, there is a predominance of intraepithelial lymphocytes, for example, CD4þ/CD8þTregs and gamma delta Tcells in the normal conjunctival epithelium; these cells are thought to harbor immunoregulatory functions similar to those found in other mucosal tissues, such as the intestine. The conjunctiva and cornea are also covered by mucin, which forms a barrier, guarding against unwarranted infiltration of inflammatory cells into the epithelium. Furthermore, the cornea lacks lymphatic and blood vessels, mature APCs, and resident Tcells, thereby reducing the incidence of chronic inflammation and bystander cell damage on ocular surface following an acute inflammatory insult. The tear fluid also contains high concentrations of soluble immunoregulatory factors that help maintain homeostasis before, during, and after environmental challenge. For example, androgenic hormones provide an immunosuppressive umbrella to help protect the secretory function of the lacrimal and the meibomian glands. In addition, the corneal epithelium expresses vascular endothelial growth factor (VEGF) receptors that function in part to sequester soluble VEGF, which ultimately reduces the stimulus for neovascularization after ocular surface challenge. Neurotrophic factors produced by the limbal corneal epithelia, such as glial cell line-derived neurotrophic factor (GDNF) also appear to have immunoregulatory activity. Transforming growth factor beta (TGF-b), a cytokine that can inhibit the function of APCs and effector T-cell proliferation, is secreted by goblet cells and is found in high levels within the tear fluid. Indeed, the presence of TGF-b may bias conjunctival APCs to activate Tregs instead of effector T cells, thereby preventing the activation/infiltration of autoreactive T cells. In addition, interleukin 1 receptor antagonist (IL-1RA) mutes the effects of the potent proinflammatory cytokine IL-1. The action of tissue inhibitor of matrix metalloproteinase (TIMP-1) inhibits matrix metalloproteinases (MMPs), which play a dominant role in promoting immune cell infiltration into ocular surface tissues during inflammation.

385

Efferent arm of the immune response: immunoregulation

The efferent arm of the immune response includes several mechanisms to prevent spurious activation and infiltration of autoreactive lymphocytes to the ocular surface during acute inflammation. It is critically important that the immune response is sufficient to eliminate the current threat, and then tempered to avoid chronic inflammation and tissue destruction. Similar to the afferent arm, recent studies have shown that the efferent arm of the immune response is also regulated by the concerted effort of Tregs, anti-inflammatory cytokines, and other factors vital for preventing autoimmunity. Activation and differentiation of CD4þ and CD8þ Tregs in the secondary lymphoid organs are critical for limiting bystander tissue damage and maintaining selftolerance and are emerging as important efferent immune modulators in the eye. Indeed, CD4þ Tregs present in C57BL/6 mice decrease clinical and histopathological disease in a mouse model of dry eye, which are exacerbated when mice are depleted of CD4þCD25þFoxP3þ Tregs. In addition, in vitro expanded CD4þCD25þFoxp3þ Tregs mute ocular surface inflammation in a Th1-mediated adoptive transfer model of dry eye disease. These data suggest that CD4þCD25þ Tregs present in the secondary lymphoid organs and ocular surface tissues inhibit the pathogenic effect of autoreactive effector T cells in an effort to maintain homeostasis. Restricted homing of effector T cells from the regional lymphoid organs to the ocular surface also facilitates immunoregulation following acute insult. For example, the programmed death ligand-1 (PD-L1) has been implicated in protecting the ocular surface from unwanted T cell infiltration and tissue injury. The PDL-1 is a negative regulator of T-cell activation; the interaction between PDL-1 and the PD-1 receptor (expressed on activated T cells) inhibits lymphocyte proliferation and cytokine secretion. Furthermore, PD-1-deficient mice develop spontaneous autoimmunity. The PDL-1 is expressed constitutively on the ocular surface and is upregulated in both human cell lines stimulated with proinflammatory cytokines and in patients with ocular inflammation. In a mouse model of corneal allograft transplantation, PDL-1 was shown to promote apoptosis of the infiltrating CD4þ and CD8þ T cells that was associated with sustained corneal allograft survival. By contrast, PDL-1 blockade resulted in increased lymphocyte infiltration within the ocular surface tissues and enhanced allograft rejection. Exactly where and how CD4þ and CD8þ Tregs exert their regulatory affects is a current area of intense research. Activation and differentiation of antigen-specific Tregs is mediated by interaction with APCs within the lymphoid organs and is influenced by the local cytokine milieu. It is clear that CD4þ Tregs are involved in suppressing inflammation on the ocular surface; however, the underlying

386

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

Afferent arm of the immune response during autoimmune-based inflammation

mechanisms driving activation and differentiation of these cells are currently unknown. Following ocular surface insults, it is suspected that Tregs function to suppress inflammation by dampening T-cell priming in the lymphoid organs, and inhibit T-cell effector function within inflamed ocular surface tissues. The immunosuppressive properties of Tregs are likely to occur by cell contact-dependent, for example, Treg:APC and/or Treg:T cell, and cell contactindependent mechanisms, that is, anti-inflammatory cytokine production. One possibility is that CD4þ Tregs temper the inflammatory response on the ocular surface by secreting TGF-b and IL-10, which may bias APC-mediated activation and differentiation of other regulatory lymphocytes in the lymphoid tissue and/or decrease Th1-mediated inflammation locally within ocular surface tissues.

IL-6

During dry eye the afferent arm of the immune response is initiated following environmental stress-mediated desiccation and/or increased tear osmolarity (Figure 2). The initial response disrupts the protective barrier by promoting proteolysis of tight junction proteins and alters the pattern of epithelial differentiation towards squamous metaplasia and decreased mucus production. Osmotic stress activates signaling pathways in a variety of cell types, including the ocular surface epithelia. Exposure to increased osmolarity in vivo or in vitro activates mitogenactivated protein kinase (MAPK) pathways, including p38 and c-Jun N-terminal kinases, and nuclear factor (NF)-kB in the ocular surface epithelia; desiccating and osmotic

A

IL-12

IL-1α

TNF

-α

va

nju

IL-12

G CCL5

CXCL10

IL-17 IL-17

IL-6

nc

tiv a

IL-17

IFNg

IFNg

nju

IL-17

IFNy

TNF-α

MMP’s

Co

IL-17

ti nc

Co

CCL5

Cornea

IL-17

IL-17

IL-1β

MMP’s

IL-6

B

C

IFNg IL-1β

CXCL10 TGF

IFNg

D

-β

M-1

ICA

Environmental stress Antigen Immature DC Activated DC

Efferent circul

Naive CD4+ T cell

t ly mph atics

Activated CD4+ T cell Tissue destruction

Af

fer en

n atio

F CXCR3

E

Cervical lymph node

LFA-1 CCR5 Figure 2 Autoimmune cycle of chronic inflammation during the immunopathogenesis of dry eye disease. Afferent arm of the immune response: (a) An environmental stimulus initiates acute inflammation on the ocular surface stimulating upregulation of proinflammatory cytokines (e.g., TNF-a, IL-1a and IL-1b), matrix metalloproteinases (MMPs), adhesion molecules (e.g., ICAM-1), and chemokines (e.g., CCL5 and CXCL10) within the conjunctival and corneal epithelium that act in concert to perpetuate the immune response; (b) Antigen presenting cells (e.g., dendritic cells) process autoantigen and (c) following activation, (d) traffic to the draining cervical lymph nodes (CLNs) via the afferent lymphatics, where they (e) present antigen to autoreactive CD4þ T cells. Efferent arm of the immune response: (f) Activated CD4þ T cells bearing specific adhesion molecules and chemokine receptors (e.g., CCR5 and CXCR3) migrate specifically to the ocular surface tissues (conjunctiva and cornea are shown), including the meibomian and lacrimal glands where they infiltrate the tissue and (g) release proinflammatory cytokines (IFN-g and IL-17) that promote chronic inflammation and tissue destruction.

Dry Eye: An Immune-Based Inflammation

stress-mediated MAPK activation stimulates corneal epithelial cells to produce of a variety of proinflammatory mediators. In addition, the altered barrier facilitates diffusion of soluble inflammatory factors into the epithelium and stroma and inflammatory cell infiltration into the ocular surface tissues. Exposure to desiccating stress (DS) initiates ocular surface epithelial cells to release proinflammatory cytokines and chemokines (e.g., IL-1b, TNF-a, IFN-g, IL-8, CXCL10, MMP-1, -3, -9, -10, and -13). For instance, IFN-g upregulates the adhesion molecule ICAM-1 expression within the epithelium, stromal fibroblasts and vascular endothelium, and increases vascular permeability in animal models and patients with non-Sjo¨gren’s or Sjo¨gren’s syndrome-mediated dry eye. Indeed, effector CD4þ T cells express high levels of lymphocyte function-associated antigen 1 (LFA-1), the cognate binding partner to ICAM-1. The upregulation of adhesion molecules on T cells and endothelial cells coupled with increased vascular permeability results in peripheral immune cell infiltration. The IL-1a and IL-1b levels are also elevated in the tears and conjunctiva from Sjo¨gren’s syndrome-associated tear deficiency patients and in a mouse model of dry eye. The IL-1 also induces upregulation of adhesion molecules localized on endothelial cells and stimulates expression of chemokines, which act in concert to facilitate leukocyte infiltration. The TGF-bdependent suppression of DC activation is also compromised as TGF-b2-secreting conjunctival goblet cells undergo apoptosis during the immunopathogenesis of dry eye. Along these lines, IL-1 and TNF-a also activate immature APCs in the cornea and conjunctiva, which results in increased expression of MHC class II antigens, co-stimulatory molecules, VEGFr3, and CCR7. These molecules act together to coordinate APC migration to the draining lymph nodes and activation of effector T cells that drive the efferent arm of the immune response. Efferent arm of the immune response during autoimmune-based inflammation

The efferent arm of the immune response is mediated by autoreactive T cells that (1) are activated within secondary lymphoid tissue via cell-to-cell contact with ocular surface-derived APCs, (2) are targeted to ocular surface tissues by acquisition of trafficking molecules and chemokine receptors, and (3) compromise the integrity of the ocular surface and contribute to prolonged inflammation (Figure 2). During the immunopathogenesis of dry eye, activation of autoreactive CD4þ T cells is thought to be driven within secondary lymphoid tissue by ocular surface-derived dendritic cells bearing self-antigen presented in the context of MHC class II. This theory is well supported by animal studies demonstrating that CD4þ T cells from the cervical lymph nodes of dry eye mice specifically home to the ocular surface tissue and mediate

387

dry eye disease when adoptively transferred to athymic nude recipient mice. Importantly, DS-specific CD4þ T cells are not detected in any other tissues indicating that these cells are targeted to the ocular surface during activation within the secondary lymphoid organs. Tissue-specific targeting of autoreactive T cells occurs within secondary lymphoid organs and is dictated by interactions between adhesion molecules and chemokine receptors that bind and respond to ligands expressed locally within inflamed tissues. Expression of LFA-1 on the T cells from patients with non-Sjo¨gren’s or Sjo¨gren’s syndrome-mediated dry eye and high-level ICAM-1 expression within ocular surface tissues of these patients suggest that LFA-1:ICAM-1 binding contributes to T-cell infiltration. Chemokine receptor signaling also appears to contribute to efferent homing of autoreactive T cells during dry eye disease. Elevated expression of the chemokine receptors CCR5 and CXCR3 and the chemokine ligands, CCL3, CCL4, CCL5, CXCL9, and CXCL10 has been detected within the cornea and conjunctiva of dry eye mice. In addition, CCR5 is also expressed on cells within conjunctival epithelium of patients with dry eye and CCL5 and CXCL10 are upregulated in human conjunctival epithelium cells in response to cytokine stimulation. These results suggest that the CCL5:CCR5 and CXCL9/ CXCL10:CXCR3 signaling axes play a role in T-cell trafficking during dry eye disease. The T-cell infiltration compromises the integrity of the ocular surface and drives chronic autoimmunemediated inflammation. In the mouse model of dry eye, accumulation of CD4þ T cells within ocular surface tissues correlates with increased cytokine production (e.g., IFN-g, IL-1b, TNF-a, and MMP-9), epithelial cell apoptosis, and decreased goblet cell density, tear production, and turnover. The presence of T cells in human dry eye patients also correlates with similar pathology. Indeed, CD4þ T cells are a prominent source of IFN-g, which can induce expression of a variety of proinflammatory factors, including trafficking molecules, such as ICAM-1, CCL5, and CXCL10, and the pro-apoptotic proteins, Fas-FasL. These factors may contribute to chronic inflammation on the ocular surface by attracting infiltrating T cells and macrophages and perpetuate bystander tissue damage, including epithelial cell apoptosis and nerve damage. The IFN-g was also shown to activate a cornified envelope precursor and genes involved in conjunctival epithelial differentiation, implicating CD4þ T cells as mediators of epithelial cell abnormalities and keratinization of the ocular surface. The CD4þ T cell accumulation is also associated with increased levels of IFN-g in the tears of dry eye mice and is inversely related to goblet cell density and conjunctival squamous metaplasia. Furthermore, exogenous administration of IFN-g to IFN-gdeficient mice results in goblet cell loss. These findings suggest that CD4þ T-cell-derived IFN-g is a major

388

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

component of chronic inflammation, epithelial cell metaplasia, and tissue destruction observed during dry eye. Emerging evidence demonstrating (1) the presence of Th17-specific CD4þ T cells within the draining cervical lymph nodes of mice with experimental dry eye and (2) high-level expression of IL-6 and IL-17 in experimental dry eye and human patients suggests that in addition to IFN-g-producing Th1 cells, IL-17-producing Th17 cells also play a pivotal role in the underlying immunopathogenesis of dry eye.

Past, Future, and Current Therapies Artificial Tears Individuals with aqueous tear deficiency have decreased tear film stability and diminished tear volume. There are several therapeutic options for these patients. Artificial tears, applied topically to the ocular surface, are polymerbased. The type of polymer used determines the artificial tear viscosity, retention time, and adhesion to the ocular surface. For example, artificial tears containing hyaluronic acid exhibit non-Newtonian properties and relatively long retention times. Other types of polymers used in different artificial tears include cellulose esters (increases tear viscosity), polyvinyl alcohol (provides optimal wetting), povidone (superior wetting), and carbomers (longer retention times). There are tear gels made with polyacrylic acid that provide greater retention times than artificial tears. Some formulations add a lipid component, such as castor oil, that helps to prevent evaporation. Many tears contain electrolytes and buffers to help normalize the tear pH. Artificial tears offer provisional relief of eye irritation but do not reverse conjunctival squamous metaplasia. Preservatives in artificial tears such as benzalkonium chloride can induce ocular surface epithelial toxicity if frequently applied on patients with low tear turnover or individuals that have punctual occlusion. Preservativefree artificial tears may be a consideration for patients using artificial tears more than four times per day. Artificial tears provide a palliative therapy to dry eye patients, but do not prevent the underlying cause of disease. Inflammation! Corticosteroids Dry eye is an immune-based inflammatory disease. Chronic dry eye requires topical treatment with therapies designed to manage inflammation. Corticosteroids effectively block multiple inflammatory pathways including proinflammatory cytokine and chemokine secretion, synthesis of matrix metalloproteinases and prostaglandins, and cell adhesion molecule expression. Activated steroid receptors bind to DNA and control gene expression and impede transcriptional regulators (AP-1 and NFkB) of proinflammatory genes. Topical corticosteroid use, while effective, is

normally prescribed for short-term use (up to 4 weeks) due to the plethora of potential side effects including glaucoma, cataracts, and ocular infection. Topical nonpreserved methylprednisolone (1%) treatment of 15 Sjo¨gren’s syndrome patients three times daily for 2 weeks followed by punctual occlusion resulted in moderate to complete relief of disease symptoms. Cyclosporine Restasis (topical CsA, 0.5%) is the only FDA-approved therapeutic for dry eye syndrome. A fungal-derived peptide, CsA, inhibits nuclear translocation of cytoplasmic transcription factors that are necessary for T-cell activation and the production of pro-inflammatory cytokines. The CsA was first identified as therapy for dry eye in dogs with spontaneous keratoconjunctivitis sicca. In human dry eye patients, treatment with topical CsA reduced conjunctival cellular infiltration, IL-6 levels, and increased conjunctival goblet cell numbers. In two 6-month independent FDA Phase III clinical trials CsA treatment resulted in a significant (p 0.05) improvement in corneal fluorescein staining and anesthetized Schirmer test values in patients treated with CsA (0.05 or 0.1%) compared to patients treated with vehicle alone. There was no indication of a dose-dependent effect with a safety profile of 0.05% and 0.1%. Moreover, patients did not show any serious adverse effects other than occasional burning and stinging. In addition, CsA also increased tear production in patients. There were no detectable levels of CsA in the blood of patients treated with CsA for 12 months, suggesting that topical CsA does not reach high enough levels to impact the systemic immune response. In addition to the clinical improvement in CsA-treated patients, there was also a marked decrease in expression of HLA-DR and IL-6 by conjunctival epithelial cells. Infiltration of CD3þ, CD4þ, and CD8þ T cells was also decreased in the conjunctiva of patients treated with CsA, but increased in those treated with vehicle alone. Mucin Secretagogues Secretagogues induce tear production by the lacrimal glands and the ocular surface epithelia. Pilocarpine and cevimeline are cholinergic agonists approved for oral administration. Patients taking Pilocarpine 5 mg four times daily reported a significantly greater overall improvement in ocular problems. Patients taking cevimeline had improvement in ocular irritation symptoms and aqueous tear production. A potential future secretagogue, Diquafosol tetrasodium, is an agonist for the P2Y2 receptor. Small molecule tyrosine kinase receptor agonists that induce mucin MUC5AC secretion from conjunctival goblet cells have also been developed.

Dry Eye: An Immune-Based Inflammation

Tetracyclines Tetracyclines, traditionally used as antibiotics, have been reported to have a number of anti-inflammatory properties including inhibition of proinflammatory cytokines, MMP production, and nitric oxide production. On the ocular surface, tetracyclines reduce human corneal epithelial production of IL-1 and MMPs, preserving the ocular surface epithelial barrier, and blocking the activation of destructive cytokines. See also: Adaptive Immune System and the Eye: Mucosal Immunity; Conjunctiva Immune Surveillance; Conjunctival Goblet Cells; Defense Mechanisms of Tears and Ocular Surface; Lacrimal Gland Hormone Regulation; Lacrimal Gland Overview; Lacrimal Gland Signaling: Neural; Meibomian Glands and Lipid Layer; Ocular

389

Mucins; Overview of Electrolyte and Fluid Transport Across the Conjunctiva; Tear Film.

Further Reading Lam, H., Bleiden, L., de Paiva, C. S., et al. (2009). Tear cytokine profiles in dysfunctional tear syndrome. American Journal of Ophthalmology 147(2): 198–205. Niederkorn, J. Y., Stern, M. E., Pflugfelder, S. C., et al. (2006). Desiccating stress induces T cell-mediated Sjo¨gren’s syndrome-like lacrimal keratoconjunctivitis. Journal of Immunology 176(7): 3950–3957. Siemasko, K. F., Gao, J., Calder, V. L., et al. (2008). In vitro expanded CD4þCD25þFoxp3þ regulatory T cells maintain a normal phenotype and suppress immune-mediated ocular surface inflammation. Investigative Ophthalmology and Visual Science 49(12): 5434–5440. Stern, M. E., Beuerman, R., and Pflugfelder, S. C. (2009). Dry Eye and the Ocular Surface. New York: Marcel Dekker.

Penetrating Keratoplasty T H Flynn and D F P Larkin, Moorfields Eye Hospital, London, UK ã 2010 Elsevier Ltd. All rights reserved.

Glossary Afferent and efferent components of allogeneic response – Afferent arm of the immune response is the inductive stage in which antigens are presented to T lymphocytes in lymph nodes. Efferent arm of the allogeneic response is the stage in which T cells and antibodies are generated and are available to mediate graft rejection or protection from infectious agents. Alloantigen – An antigen present in some, but not all, individuals of the same species. Alloreactive – The reaction of lymphocytes or antibodies with alloantigens. Anterior chamber-associated immune deviation – Systemic downregulation of antigen-specific cell-mediated immunity that is induced when antigens are introduced into the anterior chamber of the eye. Delayed-type hypersensitivity – Immune response consisting mostly of T cells that develops 24–72 h after exposure to antigen. Immune privilege – Condition in which immune responses are suppressed or downregulated. Indirect and direct allorecognition – Two pathways whereby immune responses to histocompatibility antigens on a corneal allograft elicit an immune response and ultimately allograft rejection. Indirect pathway occurs when host-derived antigen presenting cells reprocess antigens from the allograft. Direct pathway involves direct stimulation of the host’s T lymphocytes by major histocompatibility antigens expressed on the corneal allograft.

clinical circumstance in which immune homeostasis of the eye might seem undermined. However, in those patients in whom irreversible transplant rejection occurs, the selective targeting of immune-mediated injury to donor cells is a striking example of protection of host ocular tissue from bystander injury – afferent and efferent components of the allogeneic response in tandem maintaining immune homeostasis of the eye. Large cohort outcome studies have identified a number of factors on multivariate analysis which, if present, have a statistically significant detrimental effect on corneal graft survival. These are a previous ipsilateral failed graft, ipsilateral ocular inflammation, vascularization of the recipient cornea, and the primary corneal diagnosis. In most patients, these factors increase risk of graft failure by immune rejection.

Corneal Immune Privilege Apart from the specific diagnoses of keratoconus and Fuchs endothelial disease, few of the common indications for corneal transplantation can truly be considered low rejection risk. Nevertheless, it is clear that corneal transplants do enjoy comparative immune privilege. Early investigators attributed the immune privilege of the cornea entirely to its lack of vascularity, that is, sequestration of alloantigen from the immune response. There is no doubt that this is an important factor. Animal models and large human cohort studies have identified recipient corneal vascularization as the most significant factor conferring high rejection risk in multivariate analyses. In fact, the cornea is an immune-privileged tissue situated in and transplanted to an immune-privileged site. Immune-Privileged Tissue

Penetrating Keratoplasty: Indications and Survival Penetrating keratoplasty, or full thickness corneal transplantation, is the surgical procedure most commonly used in management of blinding corneal disease. Overall graft survival rates in corneal transplantation are similar to those in cadaveric renal transplantation: this indicates that the impact of corneal graft failure, due in most patients to allogeneic rejection, is significant, even if very few of the cornea patients receive oral immunosuppression as prophylaxis. In this way, corneal transplantation is one

390

A number of factors are known to contribute to the relative immune privilege of the cornea as a tissue. . On a gross level, the normal cornea is isolated from circulating immune cells due to (1) its avascular nature and (2) the blood–aqueous barrier. . Until relatively recently, the cornea was thought to contain no passenger antigen-presenting cells (APCs). Recent work has established that it does, in fact, contain APCs but that they are immature and do not express major histocompatibility complex (MHC) class II in the normal setting. Subsequently, the normal cornea is

Penetrating Keratoplasty

devoid of lymphatics to transport APCs. Experimental studies which increase the numbers of passenger APCs in the donor cornea have been shown to erode immune privilege and increase the direct component of allorecognition. Because in clinical transplantation normal avascular donor cornea is grafted, high rejection risk status is conferred by vascularization or inflammation of the residual host cornea. . The stroma and endothelium have low immunogenicity. . The endothelium, which is the most important target in rejection, expresses (i) low levels of MHC class I and II and (ii) high levels of Fas ligand which induces apoptosis in alloreactive T-lymphocyte cells and protects the graft.

391

and signs of stromal rejection have two patterns. It more frequently manifests as subepithelial opacities. These infiltrates have a similar appearance to those seen in adenovirus viral keratitis but are seen only in the donor cornea. This type of stromal rejection is often asymptomatic, but like epithelial rejection it may precede the onset of a more severe and visually significant process. Rejection of the deeper stroma results in spreading graft opacification and decreased visual acuity. Endothelial rejection usually presents with discomfort and decreased visual acuity. A line of leukocytes may, over a period of days, be seen migrating across the donor endothelium leaving edematous stroma in its wake (Figure 2). This line often spreads out like a wave from an area of deep vascularization to the graft host junction. Alternatively, a

Immune-Privileged Site The cornea, or specifically its endothelium, constitutes the anterior boundary of the anterior chamber which has been shown to be an immune-privileged site by Medawar in one of the earliest studies of this concept. The aqueous humor in contact with the endothelial cells contains high levels of immunoregulatory proteins such as transforming growth factor beta (TGF-b). In addition, antigen placed in the anterior chamber of the eye alters the immune response (anterior chamber-associated immune deviation or ACAID) to subsequent exposure to the antigen even at a different site. Antigen from the anterior chamber leaves the eye via several pathways but at least some leaves via the conventional aqueous outflow pathway and travels to the spleen. There the interaction of antigen, natural killer (NK), T cells, B cells, and gd T cells induces a type of operational tolerance.

Figure 1 Transparent corneal transplant 8 months following surgery. The donor cornea is fully transparent, indicating normal endothelial function. Both recipient and donor cornea are avascular.

Clinical Features of Corneal Graft Rejection Selective rejection signs can be observed in the epithelium, stroma, and endothelium of donor cornea in the human. The corneal endothelial cell monolayer controls stromal hydration, which has an essential role in transparency and transmission of light (Figure 1). As (1) these cells do not have mitotic capability and (2) following uncomplicated corneal transplantation the endothelial cell monolayer density declines at an even faster rate than in health, rejection of the endothelial layer is a terminal threat to graft transparency unless reversed sufficiently early. Even if reversed by treatment, a proportion of endothelial cells is lost. Patients with isolated epithelial rejection are often asymptomatic. In epithelial rejection, an elevated curvilinear opaque line is seen on the epithelium. However, epithelial rejection is direct evidence that the recipient has been sensitized to the graft and may progress to rejection of the deeper corneal layers. The clinical symptoms

Figure 2 Endothelial rejection is indicated by the demarcated region of donor corneal opacification bordered by a horizontal linear opacity. Vascularization to the superior graft periphery can also be seen.

392

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

Table 1

Summary of cellular steps to corneal transplant destruction

Direct pathway of antigen presentation

Indirect pathway of antigen presentation

Low risk

High risk

1

Host APCs infiltrate graft

Few recipient cornea APCs

Egress of host APCs bearing antigen Egress of host APC from anterior uvea bearing antigen shed from endothelium

#MHC II expression

More recipient cornea APCs "MHC II expression

Few lymphatics

Lymphatics in situ

Indirect antigen presentation ‘‘Quiet’’ eye #MHC II

Direct antigen presentation Inflamed eye "MHC II

#Costimulatory molecules ACAID Avascular cornea

"Costimulatory molecules Erosion of ACAID Vascular cornea

‘‘Quiet’’ eye #MHC expression

Inflamed eye "MHC expression

2

Egress of donor APCs

Direct priming of T lymphocytes (afferent allorecognition)

‘‘Indirect’’ priming of T lymphocytes (afferent allorecognition)

3

Exposure of circulating primed lymphocytes and other leukocytes to graft

Exposure of circulating primed lymphocytes and other leukocytes to graft

4

Recognition of alloantigen (efferent allorecognition) Recruitment of other effector cells

Recognition of alloantigen (efferent allorecognition) Recruitment of other effector cells

5

APC – antigen presenting cell; MHC – major histocompatibility complex; ACAID – anterior chamber-associated immune deviation.

more diffuse corneal edema may be seen with diffuse keratitic precipitates of variable density. In all cases, there are visible cells in the anterior chamber and the signs of inflammation and of endothelial dysfunction are limited to the donor tissue, demonstrating specificity of the allogeneic response.

Pathogenesis of Rejection Current understanding of the pathway to rejection of donor cornea following penetrating keratoplasty will be described in terms of the afferent component or sensitization, in which donor transplant antigens are recognized and processed, and the effector component. Components of the allogeneic response in low and high rejection risk corneal transplantation are summarized in Table 1. Afferent Mechanisms and Components Histocompatibility antigens

Otherwise known as transplantation antigens, these are proteins and peptides derived from donor cells; in most forms of transplantation, the most potent are class I and class II molecules of the MHC. Additional transplantation antigens termed minor histocompatibility antigens also induce allogeneic tissue rejection. The designation of

major or minor refers to the relative importance of these antigens in vascularized organ transplants and may be quite inaccurate in corneal transplantation, in which, for example, minor H antigen appears to be relatively more important than MHC antigens. Published evidence from several laboratories indicates that following transplantation recipient APCs enter the graft and endocytose exogenous alloantigen. APCs then migrate to the local lymph node where the alloantigen is presented on MHC class II molecules to naive CD4 cells and on MHC class I to naive CD8 cells. APC-associated antigen may be detected in the draining lymph node within hours of corneal transplantation. The exogenous antigen in question could be either a donor minor histocompatibility antigen or part of a donor major histocompatibility antigen. One self APC activates both CD4 and CD8 cells in what is known as the three-cell model of alloantigen presentation. Direct and indirect alloantigen recognition The type of antigen presentation by recipient APCs outlined above is termed indirect antigen presentation. Another form of antigen presentation, unique to the transplantation, is direct antigen presentation mediated by donor APCs transplanted within the graft (passenger leukocytes). According to the concept of self-restriction, alloantigen presented by these cells should not be recognized

Penetrating Keratoplasty

by host T cells if the major histocompatibility antigens are not matched. In reality, the alloantigens are recognized by a significant number of host T cells. Some of these lymphocytes may recognize, and be primed, by the alloantigenic MHC molecule itself regardless of the peptide it bears. This method of antigen presentation is consistent with the three-cell model and in vascularized organ grafts directly primed T cells constitute 90% of the alloreactive cells during acute graft rejection. With its lack of mature resident APCs and lymphatics, the cornea would appear to be poorly equipped to facilitate antigen presentation via the direct route. This is confirmed by studies which show that the indirect route of antigen presentation plays a more prominent role in low-risk corneal transplantation. This lack of influence of the direct route may explain the following findings which are on first consideration counterintuitive: . In human studies, MHC class I and particularly class II matching of corneal grafts has shown no convincing survival benefit. . In animal studies, mismatches in minor rather than major histocompatiblity antigens have been shown to be equally or more important in influencing graft survival. Further experimental evidence suggests that some of the factors which confer immune privilege do so by minimizing or preventing direct antigen presentation and that in erosion of immune privilege (i.e., in high-risk grafts) the direct route of antigen presentation becomes relatively more prominent. Nevertheless, the indirect route of antigen presentation is sufficient to induce sensitization.

T-lymphocyte activation

In the lymph node the MHC–peptide complex interacts with the T cell receptor (TCR), a complex cell surface receptor. The vigor of the T cell response following presentation of antigen is variable and dependent on the nature of the dendritic cell itself, the affinity of the clonotypic TCR for the MHC–peptide complex in question, the state of the T cell ( naive, memory), soluble factors in the immediate microenvironment (cytokines and chemokines), and the interactions between accessory (adhesion and costimulatory) molecules. Once a T cell is activated in the lymph node, there is rapid clonal expansion of alloantigen-specific T cells which enter the circulation. The lifespan of these cells is limited and it follows that there is a limited window of opportunity for these cells to bring about graft destruction in the absence of continuous antigenic stimulation. The rapid expansion of T cells is followed by contraction as many effector T cells undergo apoptosis. Memory (central and effector) cells make up part of the T cell repertoire thereafter.

393

Effector Mechanisms Once primed in the regional lymph nodes, activated lymphocytes enter the peripheral circulation. The avascular nature of the cornea and the blood–aqueous barrier provide barriers to immune cell infiltration and endothelial cell destruction. In the case of vascularized corneas, immune cells have easier access to graft antigens/cells. The nature of graft-infiltrating cells in corneal allograft rejection has been studied in human and animal pathological specimens. The cell types which appear in the highest numbers and with the greatest consistency are CD4+ and CD8+ lymphocytes of the adaptive immune system, and innate immunity component macrophages and NK cells. The presence of a cell in a tissue during rejection does not prove that the cell has an effector functional role: the important questions as to which cells cause endothelial cell destruction and by what mechanism(s) remain poorly understood. For instance, the mechanisms of allorecognition in the effector stage of graft rejection are unclear. A particular conundrum has been the question of how indirectly primed T cells (host MHC(+mH) molecules) can recognize antigen on donor cells (donor MHC(+mH) molecules). The discovery of a semidirect pathway of antigen presentation, whereby recipient APCs present whole donor MHC molecules as well as their own MHC molecules, provides a possible explanation for this but has not been demonstrated experimentally in corneal transplantation. T lymphocytes Because both CD4 and CD8 have been found in pathological specimens of rejected corneal grafts, much interest has fallen on the roles of these cells in corneal graft rejection. Several studies have demonstrated the presence of two distinct lymphocyte populations in response to a corneal allograft. One group appear to be CD4+ and are activated by indirect presentation of alloantigen. The other group are CD8+ cells with direct specificity for alloantigen. CD8+ cells act directly on target cells and are cytotoxic but it appears that CD8+ cells are less important in corneal graft rejection than in other organs. While CD4+ cells have the capacity using FasL to be directly cytotoxic, their primary modus operandi in corneal graft rejection appears to be via delayed-type hypersensitivity (DTH), by secreting cytokines and recruiting other cells such as macrophages. The evidence for role of CD4+ cells is supported by the finding that DTH responses rather than cytotoxic responses are found in rejectors of corneal grafts. We may conclude from findings in several laboratories that CD4+ cells play a more important role than CD8+ cells in graft rejection under most conditions but that either cell type may mediate rejection and that neither is essential for the process. However, there is considerable redundancy within the allogeneic response, with several lines of investigation supporting alternative cellular pathways for graft destruction.

394

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

Macrophages

The heavy mononuclear cell infiltrate in rejected grafts is in keeping with a DTH reaction. Depletion of corneal macrophages with clodronate liposomes prolongs corneal graft survival in rats, and macrophages have been shown to be necessary as APCs rather than as effector cells. However, cells of monocyte/macrophage lineage have been shown to be the dominant cell type in human aqueous humor during acute endothelial rejection. The function(s) of these cells in the effector arm of the rejection process remains unknown. NK cells

NK cells of the innate immune system have been found in rejected corneal grafts and in the aqueous humor of experimental animals with corneal allograft rejection. These cells usually function in the elimination of virally infected cells. The default function of an NK cell is to kill any cell with which it comes in contact. Only the presence of self MHC class I on the cell inhibits this process. In vitro studies have demonstrated the capacity of NK cells to kill allogeneic corneal endothelial cells, so these cells are likely to be functionally active in the efferent arm of rejection. Breakdown of immune privilege?

In a high rejection risk transplant there is a preexisting erosion of immune privilege at one or more of these steps mentioned in section ‘‘Immune-privileged Tissue’’. Low rejection risk grafts that reject later may be thought of as grafts that have acquired high rejection risk characteristics due to breakdown of immune privilege. In nonvascularized corneas, immune rejection occurring months or years after transplantation is often seen to be preceded by a local episode of alloantigen-independent inflammation (e.g., loosened transplant suture, bacterial suture-associated infection, and recurrent herpetic infection) which may lead to recruitment of immune-competent cells, angiogenesis, lymphangiogenesis, and upregulation of MHC molecules on the graft cells. Each step and the factors within it contributing to immune privilege are reasonably well understood but the extent to which one step inevitably follows the preceding one is less clear. We may ask of grafts which are not rejected, whether the recipient has not been sensitized due to the immune system not seeing the antigen (ignorance of the alloantigen), whether the immune system has seen the antigen but does not or cannot mount a response (tolerance of the antigen), or whether the immune system has seen the antigen and been sensitized but its effector cells cannot see the target antigen due to sequestration of the graft in its avascular bed. While it is tempting to speculate that a single step exists, manipulation of which would induce tolerance or absolute immune privilege in all cases, it is far more likely that the relative contributions to immune privilege at each step is different for each person and for each graft and

there is no factor contributing to immune privilege that cannot be overcome by one of the many redundant cellular pathways and mechanisms known to bring about rejection.

Treatment of Rejection The mainstay of treatment for established rejection is intensive topical corticosteroid treatment. The most commonly used regimen is prednisolone acetate 1% or dexamethasone 0.1% hourly. This treatment effectively suppresses graft inflammation but once inflammation has been suppressed, the question of whether graft clarity will return depends on the extent to which the endothelium has been damaged. Topical corticosteroids influence effector cells such as T cells and macrophages in the cornea chiefly by inducing the expression of anti-inflammatory genes (Annexin-1, SLPI) and repressing the expression of proinflammatory genes (cytokines, chemokines, adhesion molecules, and MHC molecules). Inhibition of IL-2 receptor production inhibits T cell proliferation but this may not be an important effect of topical treatment as T cell proliferation occurs quite distal to the site of application in the regional lymph nodes. Corticosteroids also affect dendritic cell (DC) function and have been shown to alter cytokine production, to induce apoptosis in DCs and to delay DC maturation with resultant impairment of antigen presentation. Corticosteroids inhibit angiogenesis but this is unlikely to be relevant in setting of acute rejection. While some clinicians treat endothelial rejection with systemic as well as topical corticosteroid, a trial of intravenous methylprednisolone in addition to intensive topical treatment did not show an improvement in outcome compared with topical treatment alone.

Prevention of Rejection The key to minimizing immune-mediated graft failure in patients is a dual strategy of (1) identifying pre- and posttransplant those at high risk of rejection, tailoring their management appropriately and (2) educating graft recipients as to the signs and symptoms of rejection. Preoperative risk factors for rejection include unmodifiable factors such as a previously rejected ipsilateral graft or previous herpetic keratitis and factors which are modifiable to a greater or lesser degree such as corneal vascularization or active external eye inflammation. All ocular inflammation should be brought under control where possible before elective corneal transplantation. A degree of regression of corneal vessels may be induced by topical steroid treatment particularly in an inflamed cornea. More established vessels may be difficult to treat. Of most concern are deep vessels and vessels close to the (projected) graft–host interface. Inhibition of formation

Penetrating Keratoplasty

of lymphatic vessels from preexisting lymphatic vessels (lymphangiogenesis) in the host cornea has been shown to prevent or delay graft rejection by inhibiting APC egress and sensitization to alloantigen. Antivascular endothelial growth factor (anti-VEGF) treatment is a promising area for future investigation as it may help to mitigate both the afferent (lymphangiogensis) and efferent (angiogenesis) arms of the immune response to allogeneic cornea. One rational approach to management of transplants at high rejection risk is the use of systemic immunosuppression with calcineurin inhibitors to prevent alloreactive T cell clonal expansion. Unfortunately, there is no robust evidence favoring any such regimes and complete absence of randomized trials. In grafts not at high risk of rejection, very long-term local immunosuppression with topical corticosteroid may be useful in preventing rejection but the benefit must be weighed against such risks as glaucoma, susceptibility to infection, and impaired corneal wound healing. In the postoperative period, patients presenting with alloantigen-independent ocular surface inflammation, such as suture loosening, should be treated promptly and aggressively. See also: Adaptive Immune System and the Eye: T CellMediated Immunity; Antigen-Presenting Cells in the Eye and Ocular Surface; Dynamic Immunoregulatory Processes that Sustain Immune Privilege in the Eye;

395

Immunosuppressive and Anti-Inflammatory Molecules that Maintain Immune Privilege of the Eye.

Further Reading Barnes, P. J. (2006). How corticosteroids control inflammation: Quintiles Prize Lecture 2005. British Journal of Pharmacology 148: 245–254. Ferguson, T. A. and Griffith, T. S. (2006). A vision of cell death: Fas ligand and immune privilege 10 years later. Immunological Reviews 213: 228–238. Fu, H., Larkin, D. F. P., and George, A. J. T. (2008). Immune modulation in corneal transplantation. Transplantation Reviews 22: 105–115. Hamrah, P., Liu, Y., Zhang, Q., and Dana, M. R. (2003). The corneal stroma is endowed with a significant number of resident dendritic cells. Investigative Ophthalmology and Visual Science 44: 581–589. Herrera, O. B., Golshayan, D., Tibbott, R., et al. (2004). A novel pathway of alloantigen presentation by dendritic cells. Journal of Immunology 173: 4828–4837. Katami, M. (1991). Corneal transplantation – immunologically privileged status. Eye 5: 528–548. Niederkorn, J. Y. (2006). See no evil, hear no evil, do no evil: The lessons of immune privilege. Nature Immunology 7: 354–359. Rogers, N. J. and Lechler, R. I. (2001). Allorecognition. American Journal of Transplantation 1: 97–102. Streilein, J. W., Wilbanks, G. A., Taylor, A., and Cousins, S. (1992). Eye-derived cytokines and the immunosuppressive intraocular microenvironment: A review. Current Eye Research 11(supplement): 41–47. Williams, K. A. and Coster, D. J. (2007). The immunobiology of corneal transplantation. Transplantation 84: 806–813. Williams, K. A., Esterman, A. J., Bartlett, C., et al. (2006). How effective is penetrating corneal transplantation? Factors influencing long-term outcome in multivariate analysis. Transplantation 81: 896–901.

Immunopathogenesis of HSV Keratitis K Buela, G Frank, J Knicklebein, and R Hendricks, University of Pittsburgh School of Medicine, Pittsburgh, PA, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Anterograde axonal transport – Transport through the cytoplasm of axons from the neuron cell body or soma towards the synapse. Chemokine – A combined form of the words chemotactic (able to attract cells) and cytokines. Chemokines are a family of low-molecular-weight chemotactic cytokines that attract and activate leukocytes at sites of infection and inflammation. Cytokine – A large and diverse family of intercellular molecular messengers that, like hormones, transmit information from one cell to another. The term is often used to denote polypeptides with immunomodulatory activity. HSV-1 latency – The retention of a complete viral genome for extended periods without production of infectious virions. Immunosurveillance – The continuous monitoring function of the immune system whereby it recognizes and reacts against aberrant cells arising within the body. Regulatory T cells – Specialized subpopulation of T cells that act to suppress activation of the immune system and thereby maintain immune system homeostasis and tolerance to self-antigens. Retrograde axonal transport – The transport of vesicles from the synaptic region of an axon toward the cell body or soma. T cells – A population of lymphocytes that derive from bone marrow precursors, mature in the thymus, and play a central role in cell-mediated immunity. Trigeminal ganglia – The ganglia that house the cell bodies of primary afferent neurons innervating the head and neck. Virion – A mature infectious virus particle existing outside a cell.

Natural History of Herpes Stromal Keratitis Although herpes stromal keratitis (HSK) has been studied in a variety of animal models, much of our current knowledge regarding the immunopathogenesis of HSK has derived from murine models due to the plethora of

396

immunologic reagents available to dissect the immune response in mice. Strain differences in susceptibility to HSK have been noted with A/J>Balb/c>C57BL/6. With most strains of HSV-1, corneal infection requires surface abrasion to facilitate entry of the virus into corneal epithelial cells. An epithelial lesion then forms as a result of virus replication in and destruction of epithelial cells. The lesions are transient, resolving by 4 days post infection (dpi), with eradication of replicating HSV-1 by 7–8 dpi. Healing of the lesions is aided by the rapid turnover of corneal epithelial cells.

Eradication of Replicating HSV-1 from the Corneal Epithelium Viral clearance appears to be largely a function of the innate immune system, consistent with the rapid kinetics of virus elimination. Evidence suggests a role for neutrophils, natural killer (NK) cells, and gd T cells in the eradication of replicating virus from the cornea. These cells are attracted to and activated within the cornea as a result of the release of cytokines, including cyclooxygenase 2, prostaglandin E2, type 1 interferon, interleukin (IL)-1, IL-6, IL-17, granulocyte macrophage colony stimulating factor (GM-CSF), and tumor necrosis factor (TNF); and chemokines including CXCL10 (IP-10), CXCL9 (MIG), CXCL2 (MIP-2), and CCL3 (MIP-1a). The production of these cytokines and chemokines appears to be initiated by ligation of tolllike receptors (TLRs), especially TLR2 and TLR9, and in response to cytokines including IL-17, IL-6, and interferongamma (IFN-g). The initial neutrophilic infiltrate peaks in HSV-1 infected mouse corneas around 4 dpi, and declines between 6 and 8 dpi, coincident with the peak and decline of viral titers in the mouse cornea. The adaptive immune system appears to play a less significant role in the initial clearance of HSV-1 from the infected cornea following primary infection. However, it is noteworthy that severe combined immune-deficient (SCID) mice that lack an adaptive immune system do not fully eradicate replicating virus from the cornea. Accordingly, activated CD4 T cells have been detected within the infected cornea by 3 dpi, although their role in inhibiting HSV-1 replication within the cornea has not been established. It is possible, and indeed likely that the failure of SCID mice to eradicate replicating HSV-1 from the cornea during primary infection results from their inability to fully establish and maintain viral latency in the trigeminal

Immunopathogenesis of HSV Keratitis

ganglion. Replicating virus in the trigeminal ganglion could then be transported back to and shed into the cornea, perpetuating the corneal infection. The role of T cells in clearing virus from the corneal epithelium following reactivation has not been studied. Antibodies also appear to play an important role in the clearance of HSV-1 from the cornea during both primary and recurrent disease.

HSV-1 Colonization of Sensory Ganglia During the course of virus replication in epithelia of the cornea, oral, or nasal passages of mice, the virus gains access to the termini of sensory neurons and is transported to neuronal nuclei in trigeminal ganglia. There it replicates briefly and then establishes a latent infection that is characterized by retention of a functional viral genome without production of infectious virions. Evidence suggests that control of acute HSV-1 replication and establishment of latency in trigeminal ganglia is mediated largely by an innate response of macrophages and NK cells mediated by TNFa, IFN-g, and nitric oxide. However, complete eradication of replicating virus requires reinforcement by CD8+ T cells and gd T cells. Latently infected neurons within the ophthalmic branch of trigeminal ganglia then serve as a source of virus for recurrent corneal disease. HSV-specific CD8+ T cells then remain in the trigeminal ganglion in close apposition to infected neurons for the life of the mouse. These CD8+ T cells express a persistent activation phenotype and have been shown to inhibit HSV-1 reactivation from latency both in vivo and in ex vivo ganglionic cultures. A similar association of activated CD8+ T cells with latently infected neurons has also been established in human ganglia. Moreover, psychological stress, which is associated with recurrent herpetic disease in humans, causes a transient compromise of the function of ganglionic CD8+ T cells and HSV-1 reactivation from latency in trigeminal ganglia of mice. These findings suggest that maintenance of HSV-1 latency and prevention of recurrent shedding of virus at mucosal surfaces such as the cornea requires constant immunosurveillance by CD8+ T cells.

Herpes Stromal Keratitis Following resolution of the epithelial lesion, corneal clarity is reestablished in mice. However, by 7–10 dpi corneal opacity develops accompanied by ingrowth of blood vessels into the normally avascular cornea, and massive leukocytic infiltration. These are manifestations of HSK, which in mice progresses from a non-necrotizing form with little epithelial involvement to severe necrotizing keratitis in which stromal inflammation gives rise to epithelial necrosis. The necrotizing form represents about 7% of human HSK, with 88% exhibiting the non-necrotizing form, and

397

5% exhibiting a mixture of both forms. It is not clear if the reduced incidence of necrotizing HSK in humans relative to mice is due to medical management of inflammation in human corneas or a reflection of fundamentally different pathogenic mechanisms in most human cases of HSK. T Cells in HSK Early studies revealed that HSK failed to develop in athymic nude mice that lacked T cells, but could be reconstituted in these mice by T-cell adoptive transfer. The CD4+ T cell population appears to be the predominant mediator of HSK in most mouse models, with CD8+ T cells tending to reduce disease severity. However, CD8+ T cells can also mediate a milder and more transient form of HSK when mice are deficient in CD4+ T cells or infected with certain strains of HSV-1. Human corneas with HSK contain both CD4+ and CD8+ T cells that are specific for HSV antigens, though the relative contribution of these cells in the progression or resolution of HSK cannot be evaluated. Recently, a population of CD4+ T cells referred to as natural T regulatory cells (nTregs) that co express CD25 and the forkhead/winged-helix transcriptional regulator, FoxP3 were shown to moderate the severity of HSK. The fact that these nTregs maintain the capacity to regulate HSK when expanded in vitro and adoptively transferred into infected mice offers a potentially useful therapeutic modality. Cytokines and Their Role in HSK Cytokines provide a means of communication among leukocytes and between leukocytes and parenchymal cells at sites of infection and/or antigen deposition. Under optimal conditions cytokines orchestrate an immunoinflammatory response that eliminates a pathogen with minimal damage to host tissue. In HSK resolution of the immunoinflammatory lesions in the cornea is accompanied by scar tissue formation. This trade-off of strength for clarity is unacceptable in a tissue within the visual axis. Moreover, the fact that in both mice and humans HSK can develop in the absence of replicating virus in the cornea renders HSK a purely immunopathological process. Thus, regulating the production and/or function of the cytokines that orchestrate HSK would reduce tissue damage without compromising protection of the host. Most studies in the murine models have implicated the Th1 cytokines as playing a primary role in the pathogenesis of HSK. The prototypic Th1 cytokine, IFN-g, aids neutrophil extravasation into the cornea at least in part through upregulation of platelet/endothelial cell adhesion molecule 1 (PECAM-1, CD31) on local vasculature. Th1 cells also produce IL-2 that has been shown to directly or indirectly regulate the migration of neutrophils into the central cornea during HSK. The Th1-associated cytokine

398

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

IL-12 also contributes to HSK development, presumably through its capacity to regulate IFN-g production. The cytokine IL-6 also plays a central role in orchestrating HSK in part through its ability to induce the neutrophil chemokine MIP-2 (CXCL2). Recently, the functional repertoire of CD4+ T cells was expanded to include Th17 cells. The IL-17 cytokine family represents the signature cytokines of Th17 cells, and is now implicated in a variety of inflammatory processes and autoimmune diseases. IL-17 is produced in both human and mouse corneas during HSK. Both human and mouse keratocytes express receptors for IL-17, and engagement of this receptor induces their production of neutrophil chemotactic factors as well as GM-CSF. The chemotactic factors attract neutrophils into the cornea while GM-CSF activates them and maintains their viability. Recent studies support a role for IL-17 in neutrophilic infiltration during the early stages of HSK. The requirement for IL-17 is transient and presumably superseded by Th1 cytokines as HSK progresses. Regulating the production or blocking the function of the Th1 and Th17 cytokines offer potential therapeutic benefit in preventing the progression of HSK. In addition, the Th1/Th2/Treg cytokine IL-10 has a proven capacity to ameliorate HSK. The general anti-inflammatory properties include inhibition of the production and function of IFN-g. Antigen-Presenting Cells in HSK Antigen-presenting cell (APC) refers to the ability of certain cells, including macrophages, dendritic cells, and B lymphocytes to take up, process, and present antigenic peptides to naive CD4+ T cells in the context of major histocompatibility complex class II (MHC class II) and co-stimulatory molecules. The once prevalent view that the normal cornea is devoid of these specialized cells has now been dispelled by several recent histological studies of mouse and human corneas. Dendritic cells (DCs) are present in the basal layer of the corneal epithelium and macrophages are present throughout the corneal stroma, with the density of both cell types gradually diminishing from the peripheral to the central cornea. During HSV-1 infection of the corneal epithelium, the resident DCs are reinforced by immigrants from the limbus. Studies employing ultraviolet irradiation to diminish DCs or cautery to increase their presence in the cornea prior to HSV-1 infection suggested a role for these cells in directly or indirectly inducing a delayed-type hypersensitivity response in lymphoid organs. Since DCs exhibit transient survival following HSV-1 infection, it is likely that either noninfected corneal DCs acquire viral antigens from corneal epithelial cells for direct presentation to T cells in the lymph nodes, or infected corneal DCs travel to the lymph nodes where they are phagocytosed by

resident DCs and viral antigens are cross-presented to T cells. The DC depletion studies also suggested an important role for corneal DC in presenting HSV-1 antigens to CD4+ T cells upon infiltration of the infected cornea. This is consistent with the observation that CD4+ T cells induce a second and more profound infiltration of DC into the cornea at the time of onset of HSK. It should be noted, however, that the method of DC depletion employed in these early studies was not specific, and the role of DC and DC subpopulations in the efferent and afferent limbs of the CD4+ T cell response in HSK awaits confirmation in studies employing specific depletion strategies now available. Although many cell types can be induced to express MHC class II antigens, professional APCs are uniquely able to co-express the co-stimulatory molecules that are required for the activation of naive and, in many cases, effector CD4+ T cells. In a mouse model of HSK, blocking the interaction of the ligands B7-1 (CD86) and B7-2 (CD80) on APCs with the co-stimulatory molecule CD28 on CD4+ T cells locally within the cornea inhibited the HSK progression. In contrast, ligation of the inhibitory receptor program death 1 (PD-1) by its ligand B7-H1 (CD274) appears to inhibit HSK progression, though it is not clear if this effect is exerted in the cornea or lymphoid organs. HSK progression was not influenced by blocking the interaction of the APC activating receptor OX40 by its T-cell ligand OX40L (CD154). These studies suggest that local inhibition of co-stimulatory molecule or augmentation of inhibitory molecule signaling in CD4+ T cells might hold therapeutic potential in the management of HSK. Chemokine Involvement in HSK Chemokines are chemotactic cytokines that activate and direct the migration of leukocytes into sites of inflammation. Chemokines not only orchestrate the innate immune response that eradicates replicating HSV-1 from the cornea (see above), but also have a central role in regulating HSK. Defining the role of chemokines in inflammatory processes is complicated by the fact that they are pleiotropic, have overlapping functions, share receptors, and can bind to multiple receptors. Not surprisingly, studies of their involvement in HSK, have in some cases, produced conflicting results. While IP-10 (CXCL10) and MCP-1 (CCL2) appear to be important for directing CD4+ T cells into the cornea, IP-10 might also inhibit neovascularization (see below). The chemokine CCL20, expressed by HSV-1-infected corneal epithelial cells and corneal keratocytes when stimulated by IL-1b and TNFa, appears to be involved in DC infiltration during both epithelial lesions and the secondary infiltration during HSK development. The chemoattractants MIP-2 (CXCL2), MIP-1a (CCL3), and RANTES (CCL5) have all been implicated in guiding neutrophils into the cornea

Immunopathogenesis of HSV Keratitis

during HSK. A variety of neutralizing antibodies, and peptide and nonpeptide inhibitors of these chemokines have been developed and might have therapeutic potential in managing HSK, though their successful use will likely be complicated by the redundant nature of the chemokines. Angiogenesis in HSV-1 Stromal Keratitis Neovascularization of the normally avascular cornea appears to be a requisite step in the development of severe HSK in mice. Both vascular endothelial growth factor (VEGF) and matrix metalloproteinase (MMP)-9 have an important role in the neovascularization of corneas with HSK. Production of these factors is induced by IL-1, IL-6, and MIP-2. Corneal cell production of soluble VEGF receptors prevents vascularization of the cornea during steady state. Clearly, the capacity of the soluble VEGF receptor to neutralize VEGF and prevent its binding to VEGF receptors on the limbal vasculariture is overwhelmed by the amount of VEGF produced in HSV-1 infected corneas. Neovascularization provides enhanced corneal access of blood-borne leukocytes that mediate HSK, and blocking neovascularization effectively prevents HSK progression. Based on the established roles for vascularization and VEGF in mouse models of HSK, and the prominent presence of corneal vessels in necrotizing keratitis in humans, therapeutic approaches geared toward blocking vascularization hold great promise in the management of HSK.

Models of HSK Pathogenesis Despite extensive advances in our understanding of the pathogenesis of HSK in mice, one of the most basic questions (viz., what stimulates the CD4+ T cells that mediate HSK) remains unanswered. Three models of CD4+ T-cell activation in HSV-1 infected corneas have been advanced and supported by experimental data. These include: (1) bystander activation by the cytokine milieu present in the cornea, (2) autoimmunity to corneal tissue resulting from molecular mimicry by viral proteins, and (3) virus-specific activation. The concept that bystander activation of CD4+ T cells mediated by the cytokine milieu within the infected cornea is sufficient to induce HSK was proposed based on findings in mice whose T cells all express transgenic Tcell receptors (TCRs) specific for ovalbumin (ova). When the corneas of these mice were infected with HSV-1, they developed HSK that was mediated by ova-specific CD4+ T cells. However, these mice failed to control HSV-1 replication in the cornea, and succumbed to infection early during HSK development. Indeed, when HSV-1 replication was controlled in the corneas of these with kinetics similar to that of immunologically normal mice,

399

the mice failed to develop HSK. Thus, in immunologically normal mice (and presumably people) bystander activation of CD4+ T cells might contribute to, but is not itself sufficient for, the development of HSK. Other studies incorporating BALB/c mice that are congenic for the IgH locus concluded that HSK develops as a result of an autoaggressive attack of CD4+ T cells on corneal tissue. This autoimmune response was induced by a peptide contained in the viral UL6 protein that mimicked a sequence in a corneal protein and in the Ig heavy chain of one of the congenic strains. The BALB/c strain that expressed the mimicked peptide in its Ig heavy chains did not respond to the peptide due to clonal deletion, and were highly resistant to HSK. In contrast, the congenic strain that failed to express the UL6-mimicked peptide in its Ig heavy chain generated a response to UL6 that crossreacted with a corneal protein. The authors concluded from these findings that HSK is an autoimmune disease arising from antigenic mimicry by the HSV-1 UL6 protein. Unfortunately, these intriguing findings were not reproduced in subsequent studies by another group, and no UL-6-specific or cornea-specific CD4+ T cells have been isolated from mouse or human corneas with HSK. Thus, the involvement of autoimmunity in HSK remains uncertain. A third possibility is that reactivity of HSV-specific CD4+ T cells to HSV antigens within the cornea triggers HSK. This hypothesis is consistent with the fact that HSV-1-specific CD4+ T cells have been isolated from both human and murine corneas at varying stages of keratitis. Further support for this concept came from a study in which partial tolerization of CD4+ T cells to HSV-1 antigens was associated with reduced severity of HSK. However, no studies to date have established a clear role for HSV-1-specific CD4+ T cells in directly mediating HSK. Based on the available data, it is likely that the initiation of HSK results from stimulation of HSVspecific CD4+ T cells in the cornea, while bystander activation might contribute to the progression and chronicity of the inflammation.

Epidemiology and Pathogenesis of Human HSK Clinically, ocular HSV-1 infection remains a significant cause of visual impairment worldwide. In developed nations, there are approximately 8.4–13.2 new ocular HSV-1 infections per 100 000 person-years with an overall incidence, including recurrences, of 20.7–31.5 episodes per 100 000 person-years. Perhaps the most serious manifestation of ocular HSV-1 infection is the potentially blinding HSK, which accounts for only 2% of initial HSV-1 ocular presentations but represents 20–61% of recurrent disease.

400

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

Multiple forms of HSK exist and no classification system is universally accepted (here we use a classification system described by Liesegang). HSK commonly presents as a non-necrotizing or immune form, but can also present as the rarer but more serious necrotizing form. Based on limited data, it appears that both the necrotizing and immune presentations of human HSK share the features of neovascularization and leukocytic infiltration seen in the mouse model. Since HSK in the mouse progresses from disease resembling immune HSK to disease resembling necrotizing HSK, it is reasonable to propose that the two forms represent a continuum rather than distinct pathologies. However, there are important differences between the human and mouse HSK, including the fact that HSK rarely leads to corneal perforation in humans, while perforation is a frequent occurrence in the mouse model. Moreover, human HSK is often associated with replicating HSV-1 in the cornea and overlying epithelial lesions, whereas replicating HSV-1 is typically eradicated from the mouse cornea prior to the onset of HSK.

of CD4+ T-cell pathways necessary for production of inflammatory cytokines, such as IL-2 and IFN-g. Furthermore, CsA inhibits corneal neovascularization in murine models of HSK. Successful medical treatment of human HSK has important implications concerning the theories of HSK pathogenesis. CsA functions to inhibit TCR-induced calcineurin signaling that leads to production of inflammatory cytokines, such as IL-2. Therefore, effective treatment of HSK with CsA implicates TCR signaling, favoring CD4+ T-cell activation by viral or autoantigens, but not by cytokines. Studies in mice emphasize an important role for neovascularization in HSK progression. Mouse studies also implicate VEGF as a critical regulator of vascularization associated with HSK. The current availability of therapeutic agents such as Avastin and Leucentis that inhibit vascularization by blocking VEGF offers the potential for exciting new treatment alternatives.

Management of Human HSV Stromal Keratitis and Implications on Pathogenesis

Acknowledgments

Management of HSK is complicated by the fact that the cells that are responsible for eliminating the virus also contribute to immunopathology in the cornea. Treating the inflammation alone can significantly exacerbate viral replication. In contrast, treating with antivirals alone can lead to uncontrolled inflammatory damage to the cornea. The current standard of care for both non-necrotizing and necrotizing HSV stromal keratitis includes topical corticosteroids and topical antivirals. Corticosteroids are used to diminish the immunopathological component of stromal disease, while the antivirals prevent further viral replication. However, corticosteroids applied to the ocular surface can lead to adverse side effects including development of cataracts or glaucoma. A less well-studied treatment option for patients suffering from HSV stromal keratitis is topical cyclosporin A (CsA) in addition to topical antivirals. Several small noncontrolled trials in humans with HSK suggested that treatment with CsA may be beneficial in cases resistant to corticosteroids. Additionally, studies using murine models of HSK revealed that CsA effectively reduced stromal inflammation and haze in a dose-dependent manner. Similar to corticosteroids, CsA also functions to dampen the immune system. CsA blocks transcriptional activation

The authors would like to thank Kira Lathrop, MAMS, for technical assistance within this article. See also: Adaptive Immune System and the Eye: T CellMediated Immunity; Antigen-Presenting Cells in the Eye and Ocular Surface; Avascularity of the Cornea; Corneal Epithelium: Response to Infection; Immunopathogenesis of Pseudomonas Keratitis.

Further Reading Biswas, P. S. and Rouse, B. T. (2005). Early events in HSV keratitis – setting the stage for a blinding disease. Microbes and Infection 7: 799–810. Lepisto, A. J., Frank, G. M., and Hendricks, R. L. (2007). How herpes simplex virus type 1 rescinds corneal privilege. In: Niederkorn, J. Y. and Kaplan, H. J. (eds.) Immune Response and the Eye, vol. 92, pp. 203–212. Karger: Basel. Liesegang, T. J. (1999). Classification of herpes simplex virus keratitis and anterior uveitis. Cornea 18: 127–146. Metcalf, J. F., Hamilton, D. S., and Reichert, R. W. (1979). Herpetic keratitis in athymic (nude) mice. Infection and Immunity 26: 1164–1171. Sheridan, B. S., Knickelbein, J. E., and Hendricks, R. L. (2007). CD8 T cells and latent herpes simplex virus type 1: Keeping the peace in sensory ganglia. Expert Opinion on Biological Therapy 7: 1323–1331.

Immunopathogenesis of Onchocerciasis (River Blindness) E Pearlman, Case Western Reserve University, Cleveland, OH, USA K Gentil, University of Bonn, Bonn, Germany ã 2010 Elsevier Ltd. All rights reserved.

Glossary Microfilariae – First stage (L1) larvae of filarial nematodes including Onchocera volvulus, the causative agent of river blindness. TLR (toll-like receptor) – A family of surface and endosomal receptors that are expressed on mammalian cells, including the cornea. These receptors recognize microbial products and transmit cell signals that culminate in the elaboration of chemokines and cytokines. Wolbachia – Obligate intracellular bacteria that exist as endosymbionts in filarial nematodes including Onchocerca volvulus.

Onchocerciasis (river blindness) remains endemic in a number of sub-Saharan African countries and has foci in Yemen and in Latin America. Most recent (2006) estimates indicate that there are 37 million individuals infected with Onchocerca volvulus. The life cycle of all filarial nematodes includes transmission through insect vectors, with Simulium blackflies transmitting O. volvulus. First stage larvae (microfilariae, L1) are ingested during a blood meal and migrate through the insect gut, thorax and into the salivary gland having undergone two molts to the third-stage larvae (L3). Infective L3 enter the mammalian host during a second insect blood meal, where they develop to L4 stage and then adult males and females. Adult male and female worms live for over 10 years in subcutaneous tissues, producing millions of microfilariae over their lifespan. Microfilariae can survive for over 1 year in the skin, and can enter the anterior and porterior segments of the eye. While alive, microfilariae appear to cause minimal damage, and individuals can be very heavily infected; however, when the larvae die either by natural attrition or following chemotherapy, the host immune response causes acute and chronic tissue damage, with severe onchodermatitis in the skin, visual impairment, and blindness. Figure 1 shows examples of sclerosing keratitis and of cellular infiltration and vascularization in the corneal stroma of an infected individual in west Africa.

The Role of Endosymbiotic Wolbachia Bacteria in Onchocerciasis Intracytoplasmic Rickettsia-like bacteria were first described in filarial nematodes in 1977, and later identified

as Wolbachia pipientis. Although 75% of insect species and a number of crustaceans harbor Wolbachia as endosymbionts, filarial nematodes are the only group of parasitic worms that are infected, most likely because they are the only nematode group with an obligate insect host. In filarial nematodes, Wolbachia are detected with in cells in the hypodermis and uterus, in immature microfilariae in the uterus, and in mature microfilariae in the skin and cornea. The bacteria are most numerous in the mammalian host compared with the insect stage, and appear to have an essential, though poorly understood, role in nematode embryogenesis. Antibiotic (Doxycycline) treatment of filaria-infected individuals effectively sterilizes the adult females, reducing overall microfilaria numbers in the skin and blocking disease transmission.

Role of Wolbachia in Pathogenesis – Evidence from Infected Individuals The role of Wolbachia in the pathogenesis of filarial disease has been implicated from observations made after antifilarial therapy. Elevated Wolbachia DNA and even intact Wolbachia are detected in the blood, and are associated with the proinflammatory cytokines seen in patients with post-treatment side effects such as fever, edema, and headache. Using quantitative PCR for Wolbachia surface protein (WSP) gene, which is present as a single copy per organism, the number of bacteria per worm was found to be similar in all insect stages of B. malayi. However, within 7 days in the mammalian host, bacteria numbers increased 600-fold and showed a high ratio of Wolbachia/nematode DNA in L4 larvae, indicating rapid bacterial replication within the worms. This Wolbachia load is maintained in adult males, but increases in females during embryogenesis. Wolbachia are therefore a target for antibiotic treatment, and patients given a course of doxycyline in addition to annual ivermectin treatment have reduced systemic microfilariae, and significantly fewer adult worm number in subcutaneous nodules. Wolbachia appear to mediate recruitment of neutrophils, as the number of neutrophils in nodules from doxycycline-treated individuals is greatly reduced compared with untreated individuals. An additional line of evidence for a role for Wolbachia in the pathogenesis of onchocerciasis relates to earlier studies showing that two strains of O. volvulus that differ in virulence exist in West Africa based on DNA probes using a noncoding repeat sequence, and the strain

401

402

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

(b)

V

(a)

(c)

Figure 1 Ocular onchocerciasis. (a) Blind individual in endemic region of Cote D’Ivoire, West Africa, 1999, with both corneas opaque as a result of infection with Onchocerca volvulus (photograph by Eric Pearlman); (b) sclerosing keratitis, showing central region of corneal opacification and neovascularization (with permission from Dr Hugh Taylor); (c) corneal section showing intact microfilariae (arrow) in the corneal stroma. Blood vessels (V) and cellular infiltrate are also visible (from Armed Forces Institute of Pathology).

shown to cause more severe ocular disease has significantly higher Wolbachia loads compared with the second, less virulent strain, indicating a correlation between virulence and Wolbachia in ocular onchocerciasis. Taken together, these findings strongly support the notion of an important role for Wolbachia in the proinflammatory response and pathogenesis of onchocerciasis.

The Role of Wolbachia in the Pathogenesis of ocular onchocerciasis – Lessons from the Murine Model of O. volvulus/Wolbachia Keratitis Microfilariae invade both the anterior and the posterior segments of the eye. In the latter case, they cause uveitis and chorioretinitis, resulting in loss of vision; therefore, although Onchocerca keratitis is more frequent and more readily detectable, corneal transplants are not conducted as the patients generally also have posterior segment disease. Because eyes from human cases of onchocerciasis are not available, the host response to Onchocerca has been examined in the skin of infected individuals. As such, Onchocerca infected individuals show microfilariae in the dermis surrounded by neutrophils, eosinophils, or macrophages. The likely explanation is that neutrophils surround recently dead and degenerating worms in the skin, whereas macrophages and eosinophils migrate to the site at later

time points. Our findings in a murine model of Onchocerca keratitis shows that neutrophils surround microfilariae in the cornea within 24 h and immunogold labeling of the major Wolbachia surface protein shows neutrophils in close proximity to Wolbachia (Figure 2). Using this mouse model of O. volvulus keratitis in which filaria/Wolbachia extracts are injected into the corneal stroma, St. Andre and colleagues demonstrated that endosymbiotic Wolbachia bacteria are essential for the pathogenesis of O. volvulus keratitis as: (1) O. volvulus from individuals depleted of Wolbachia by antibiotic treatment do not induce corneal inflammation; (2) related filarial species containing Wolbachia induce keratitis in contrast to filarial species lacking Wolbachia; and (3) isolated Wolbachia induce neutrophil recruitment to the corneal stroma.

Wolbachia and TLRs in the Cornea TLRs are surface and endosomal receptors that are expressed in the cornea and respond to microbial products such as lipopolysaccharide (TLR4) and lipoproteins (TLR2). TLR2 forms heterodimers with TLR1 or TLR6 to initiate signaling through adaptor molecules and induce nuclear factor kappa B (NFkB) translocation to the nucleus, and results in production of pro-inflammatory and chemotactic cytokines. Our findings using gene knock-out mice clearly demonstrate that either O. volvulus extracts

Immunopathogenesis of Onchocerciasis (River Blindness)

Stroma ml

B (a)

(b)

PMN

PM

N

PMN

mf

mf (c)

D (d)

(e) Figure 2 Presence of endosymbiotic Wolbachia bacteria in microfilaria and adult female worms. C57BL /6 mice were injected into the corneal stroma with microfilariae, corneas were removed after 4 h or 18 h and thin sections were immunostained with anti-Wolbachia Surface Protein (WSP) and visualized with IgG conjugated to 15 nm gold particles. Sections were counterstained with uranyl acetate and lead citrate, and examined by electron microscopy. (a, b) 4 h after injection. WSP was clearly detected inside microfilariae in the corneal stroma (arrows). Mf: microfilariae. (c, d) 18 h after injection microfilariae containing Wolbachia were surrounded by neutrophils (PMN). WSP labeled with gold particles (arrows) are present in the microfilariae adjacent to the neutrophils. Magnifications: (a) 4800; (b) 8400; (c) 5300; (d) 16 000. Reprinted with permission from Gillette-Ferguson, I., Hise, A. G., McGarry, H. F., et al. (2004). Wolbachia-induced neutrophil activation in a mouse model of ocular onchocerciasis (river blindness). Infection and Immunity 72: 5687. (e) Longitudinal section of adult female Brugia malayi immunostained with anti-WSP showing WSP positive microfilariae in the uterus (magnification 400). Reprinted from Science; photomicrograph by Amy Hise.

403

containing Wolbachia, or isolated Wolbachia bacteria selectively activate TLR2 and TLR6, and adaptor molecules MyD88 and Mal. Figure 3 shows that corneal inflammation (neutrophil infiltration and increased corneal haze) is entirely dependent on activating TLR2. Moreover, studies using chimeric mice have shown that TLR2 expressed on bone-marrow-derived cells have an important role in provoking corneal inflammation. Taken together, findings from our group and others indicate that Wolbachia induces TLR2 activation in resident macrophages in the corneal stroma, and produce proinflammatory cytokines and CXC chemokines, which mediate neutrophil recruitment from peripheral, limbal vessels into the corneal stroma. Neutrophil responses to Wolbachia are also dependent on TLR2/MyD88, which mediate cytokine production by these cells, and may contribute to degranulation and secretion of reactive oxygen species and matrix metalloproteinases, resulting in cell death and loss of corneal clarity. In chronically infected, untreated individuals, there is also an ongoing adaptive immune response, due to repeated invasion of microfilariae into the corneal stroma, and prolonged worm degeneration and release of Wolbachia. Infiltrating eosinophils and macrophages also contribute to tissue damage, manifesting as corneal opacification, loss of vision, and blindness. We found that a third role for the TLR2 is filaria/Wolbachia activation of dendritic cells and T-cell production of IFN-g but not IL-4 or IL-5. IFN-g also has an indirect role in enhancing pro-inflammatory and chemotactic cytokine production, thereby increasing neutrophil recruitment to the corneal stroma (Figure 4). Together, these findings demonstrate that TLR2 governs the host response to Wolbachia at several levels, including systemic and corneal responses, and may be a target for blocking corneal inflammation.

Identification of a Wolbachia TLR2/TLR6 Ligand The TLR2/TLR6 heterodimer is activated by diacylated lipoproteins. To identify possible Wolbachia lipoproteins that activate TLR2/TLR6, Taylor and colleagues searched the lipoprotein databases, and identified Brugia malayi Wolbachia peptidoglycan-associated lipoprotein PAL (wBmPAL). Synthetic diacylated wBmPAL was shown to selectively induce IFN-g production, to induce systemic TNF-a in a murine model of lymphatic filariasis, and to induce corneal inflammation in a TLR2/TLR6–dependent manner. These data indicate that the interaction between these (and likely other) lipoproteins and TLR2/6 in the cornea is essential for the development of O. volvulus keratitis.

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

HBSS

404

6000 4000 OvAg

NIMP-R14+cells

8000

2000 0

(a)

C57Bl/6

TLR2−/−

TLR4−/−

TLR2/4−/−

(b)

350

# cells/section

300 250 200 150 100 50 0 Recipient

B6

TLR2

B6

TLR2

HBSS

Donor

B6

TLR2

TLR2

B6

All

(c)

OvAg

Figure 3 O. volvulus /Wolbachia keratitis is dependent on TLR2. O. volvulus extract containing Wolbachia (OvAg) was injected into the corneal stroma of C57BL/6, TLR2/, TLR4/ and TLR2/4/ mice. After 18 h, mice were sacrificed and corneas were examined as described. (a) Single cell suspension was prepared from the corneal stroma, and total neutrophils were detected by flow cytometry using MAb NIMP-R14. (b, c) Bone marrow cells from donor C57BL/6 / eGFP+ and TLR2/ mice were used to reconstitute sublethally irradiated recipient C57BL/6 or TLR2/ mice. After 2 weeks, chimeric mice were injected intrastromally with O. volvulus extract (OvAg) or saline (HBSS), and 18 h later, 5 mm corneal sections were examined by fluorescence microscopy. Corneal sections were examined for neutrophils by immunohistochemistry. (b) Representative corneas from irradiated C57BL/6 mice reconstituted with C57BL/6 / eGFPþ bone marrow cells. After 2 weeks, mice were either untreated (naive), or examined 24 h after injection with either saline (HBSS) or O. volvulus extract (OvAg). (c) Total neutrophils per corneal section. Data points represent individual corneas; these data points are combined from three repeat experiments. Reprinted from Gillette-Ferguson, I., Daehnel, K., Hise, A. G., et al. (2007). Toll-like receptor 2 regulates CXC chemokine production and neutrophil recruitment to the cornea in Onchocerca volvulus/Wolbachia-induced keratitis. Infection and Immunity 75: 5908–5915.

Conclusion River blindness appears to be mostly under control due to mass distribution of ivermectin (Mectizan), which kills microfilariae and has led to reduced prevalence of disease. Recent studies targeting Wolbachia for antibiotic treatment now demonstrate that doxycycline can also be used to treat infected individuals. This is not only a paradigm shift in treating filarial-infected individuals, but given the risk for ivermectin resistance, also provides a critical second approach to treatment, although this is still at the clinical trial stage. The Bill and Melinda Gates Foundation recently provided funds to develop novel antibiotics as an adjunct treatment for river blindness and lymphatic filariasis, and are described in the antiWolbachia website. Overall, the increased understanding

of the pathogenesis of this disease and targeting Wolbachia in particular bring these devastating diseases closer to being controlled.

Acknowledgments We gratefully acknowledge the contribution of our colleagues in these studies, including Illona Gillette-Ferguson, Amy Hise, Achim Hoerauf and Mark Taylor. This work was supported by NIH grants EY10320 and EY11373, by the Research to Prevent Blindness Foundation and the Ohio Lions Eye Research Foundation. See also: Innate Immune System and the Eye; Pathogenesis of Fungal Keratitis.

Immunopathogenesis of Onchocerciasis (River Blindness)

C57BL/6 TLR2−/− TLR4−/−

Medium OvAg

P = 0.0110

Anti-CD3 0

500 Cells/5 μm section

Medium OvAg Anti-CD3

(a)

100

200 7000 14 000 IFN-γ (pg ml−1)

Neutrophils

0 (b) 500

P < 0.0001 400

400

300

300

200

200

100

100

0 (c)

405

100 200 300 400 500 600 700 IL-5 (pg ml−1) Eosinophils

p = 0.24

0 C57BL/6

IFNγ –/–

(d)

C57BL/6

IFNγ −/−

Figure 4 IFN-g production is dependent on TLR2 and regulates neutrophil recruitment to the corneal stroma. (a, b) Mice were immunized three times with soluble O. volvulus extract containing Wolbachia (OvAg). After the final immunization, mice were sacrificed and spleens removed for in vitro recall response with soluble OvAg. As a positive control, splenocytes were incubated with stimulatory anti-CD3. Cytokine production by splenocytes after 72 h of culture was measured by ELISA. Note TLR2-dependent IFN-g, but not IL-5 production. (c, d) C57BL/6 and IFNg/ mice were immunized 3 times subcutaneously with OvAg, and injected into the corneal stroma with OvAg. After 24 h, 5 mm corneal sections were immunostained for neutrophils or eosinophils, and the number of positive cells in the corneal stroma was assessed by fluorescence microscopy. Each data point represents an individual cornea, and the experiment was repeated 3 times. Note IFN-g-dependent recruitment of neutrophils, but not eosinophils to the cornea. (a, b) Daehnel, K., GilletteFerguson, I., Hise, A. G., et al. (2007). Filaria/Wolbachia activation of dendritic cells and development of Th1-associated responses is dependent on Toll-like receptor 2 in a mouse model of ocular onchocerciasis (river blindness). Parasite Immunology 29: 455–465. (c, d) Gentil, K. and Pearlman, E. (2009). IFN-g and IL-1R1 regulate neutrophil recruitment to the corneal stroma in a murine model of Onchocerca volvulus keratitis (river blindness). Infection and Immunity 77(4): 1606–1612.

Further Reading Boatin, B. A. and Richards, F. O., Jr. (2006). Control of onchocerciasis. Advances in Parasitology 61: 349–394. Brattig, N. W. (2004). Pathogenesis and host responses in human onchocerciasis: Impact of Onchocerca filariae and Wolbachia endobacteria. Microbes and Infection/Institut Pasteur 6: 113–128. Daehnel, K., Gillette-Ferguson, I., Hise, A. G., et al. (2007). Filaria/ Wolbachia activation of dendritic cells and development of Th1associated responses is dependent on Toll-like receptor 2 in a mouse model of ocular onchocerciasis (river blindness). Parasite Immunology 29: 455–465. Gentil, K. and Pearlman, E. (2009). IFN-g and IL-1R1 regulate neutrophil recruitment to the corneal stroma in a murine model of Onchocerca volvulus keratitis (river blindness). Infection and Immunity 77(4): 1606–1612. Gillette-Ferguson, I., Daehnel, K., Hise, A. G., et al. (2007). Toll-like receptor 2 regulates CXC chemokine production and neutrophil recruitment to the cornea in Onchocerca volvulus/Wolbachiainduced keratitis. Infection and Immunity 75: 5908–5915. Gillette-Ferguson, I., Hise, A. G., McGarry, H. F., et al. (2004). Wolbachia-induced neutrophil activation in a mouse model of ocular onchocerciasis (river blindness). Infection and Immunity 72: 5687–5692. Higazi, T. B., Filiano, A., Katholi, C. R., et al. (2005). Wolbachia endosymbiont levels in severe and mild strains of Onchocerca volvulus. Molecular and Biochemical Parasitology 141: 109–112. Hise, A. G., Daehnel, K., Gillette-Ferguson, I., et al. (2007). Innate immune responses to endosymbiotic Wolbachia bacteria in Brugia

malayi and Onchocerca volvulus are dependent on TLR2, TLR6, MyD88, and Mal, but not TLR4, TRIF, or TRAM. Journal of Immunology 178: 1068–1076. Hoerauf, A. (2008). Filariasis: New drugs and new opportunities for lymphatic filariasis and onchocerciasis. Current Opinion in Infectious Diseases 21: 673–681. Hoerauf, A., Mand, S., Adjei, O., Fleischer, B., and Buttner, D. W. (2001). Depletion of Wolbachia endobacteria in Onchocerca volvulus by doxycycline and microfilaridermia after ivermectin treatment. Lancet 357: 1415–1416. Pearlman, E., Johnson, A., Adhikary, G., et al. (2008). Toll-like receptors at the ocular surface. Ocular Surface 6: 108–116. Saint Andre, A., Blackwell, N. M., Hall, L. R., et al. (2002). The role of endosymbiotic Wolbachia bacteria in the pathogenesis of river blindness. Science 295: 1892–1895. Taylor, M. J., Bandi, C., and Hoerauf, A. (2005). Wolbachia bacterial endosymbionts of filarial nematodes. Advances in Parasitology 60: 245–284. Turner, J., Langley, R. S., Johnston, K. L., et al. (2009). Filarial Wolbachia lipoprotein stimulates innate and adaptive inflammatory responses through TLR2 and TLR6 and induce disease manifestations of lymphatic filariasis and river blindness. Journal of Biological Chemistry 284: 22364–22378.

Relevant Websites http://www.a-wol.net – A-WOL – Anti-Wolbachia Consortium.

Immunopathogenesis of Pseudomonas Keratitis L D Hazlett, Wayne State University School of Medicine, Detroit, MI, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Akt1 – Also known as Akt or protein kinase B, it is an important molecule in mammalian cellular signaling and inhibits apoptotic processes. Eicosanoids – Signaling molecules made by oxygenation of 20-carbon essential fatty acids. There are four families of eicosanoids: the prostaglandins, prostacyclins, thromboxanes, and the leukotrienes. IRF1 (interferon regulatory factor 1) – A member of the interferon regulatory transcription factor family. IRF1 serves as an activator of interferons alpha and beta transcription, and in mouse it has been shown to be required for double-stranded RNA induction of these genes. IRF1 also functions as a transcription activator of genes induced by interferons alpha, beta, and gamma. Further, IRF1 has been shown to play roles in regulating apoptosis and tumor-suppression. Keratitis – A condition in which the cornea becomes inflamed. The condition is often marked by moderate to intense pain and usually involves impaired eyesight. LPS (lipopolysaccharide, endotoxin) – Major outer membrane component of Gram-negative bacteria comprising a lipid A core (endotoxin) and polysaccharide of varying length and composition. Toll-like receptor 4 (TLR4) and MD2 (myeloid differentiation 2) bind to the lipid A moiety. MAPK (mitogen-activated protein kinase) – Serine/threonine-specific protein kinases that respond to extracellular stimuli (mitogens) and regulate various cellular activities, such as gene expression, mitosis, differentiation, and cell survival/ apoptosis. Matrix metalloproteinase-9 (MMP-9) – Gelatinase B, 92 kDa gelatinase, 92 kDa type IV collagenase is a biological enzyme. Proteins of the MMP family are involved in the breakdown of extracellular matrix in normal physiological processes, as well as in disease processes. Most matrix metalloproteinases (MMPs) are secreted as inactive pro-proteins which are activated when cleaved by extracellular proteinases. The enzyme encoded by this gene degrades type IV and V collagens. NF-kB (nuclear factor-kappa B) – A protein complex that is a transcription factor. NF-kB is found in almost all animal cell types and is involved in cellular responses to stimuli such as stress,

406

cytokines, free radicals, ultraviolet irradiation, and bacterial or viral antigens. Peroxynitrite – An anion, with the formula ONOO , it is an unstable valence isomer of nitrate, NO3 , which has the same formula but a different structure. Although peroxynitrous acid is highly reactive, its conjugate base, peroxynitrite is stable in basic solution. SIGIRR (single immunoglobulin interleukin 1 receptor (IL-1R)-related protein) – An inhibitory member of this receptor superfamily. SIGIRR seems to temper cellular activation by IL-1, LPS, and probably other activators of receptors in the TLR–IL1R superfamily, such that the biological outcome will be the result of a balance between activation by a receptor and dampening by SIGIRR. SIGIRR therefore acts as a brake on the TLR system, which may be essential for regulating the detrimental effects of innate immunity, as occurs in sepsis and chronic inflammation. siRNA (short interfering RNA or silencing RNA) – A class of 20–25 nucleotide-long double-stranded RNA molecules that play a variety of roles in biology. Most notably, siRNA is involved in the RNA interference pathway, where it interferes with the expression of a specific gene. T helper (Th) 1 (e.g., predominant IL-12 and interferon gamma production) and Th2 (predominant IL-4, 5, 10, 13) response – Response involving effector T cells that play an important role in establishing and maximizing the capabilities of the immune system.

Introduction In the United States, microbial keratitis is frequently associated with complications resulting from use of extended wear contact lenses with an incidence of about 30 000 cases annually. The cost of treatment is estimated at between $15 and $30 million, representing a considerable medical and economic impact. Of the bacterial organisms able to induce keratitis, Pseudomonas aeruginosa remains an important Gram-negative pathogen. The bacteria is often referred to as opportunistic, as it is capable of inducing keratitis not only in extended wear contact lens users, but

Immunopathogenesis of Pseudomonas Keratitis

also in more tropical climates, and in patients that are either debilitated or hospitalized. Most of the complications of bacterial keratitis are structural alterations of the cornea, but other sight-threatening problems include development of secondary glaucoma and cataract. These consequences are largely caused by the host’s inflammatory response, but the influence of bacterial toxins, exoproducts, and toxicity from antibiotic treatment cannot be overlooked. There is also a recent increase in the incidence of reported antibiotic resistant strains and failure of an initially promising treatment using antimicrobial peptides to manage keratitis, suggesting that better understanding of the pathogenic mechanisms of disease induction by this pathogen will be critical to development of improved therapeutic strategies. P. aeruginosa, like most other microorganisms, typically requires surface injury to permit corneal invasion. Because it has few nutritional requirements, it can adapt to a variety of ecological conditions and niches, such as preserved ophthalmic solutions and the hospital environment. Pseudomonal and other Gram-negative bacterial infections often present as a rapidly progressing, suppurative stromal infiltrate with a marked mucopurulent exudate. Yellowish coagulative necrosis surrounded by inflammatory epithelial edema is distinctive and stromal ulceration can lead to stromal tissue destruction and vision loss. A ring infiltrate may appear in the surrounding paracentral cornea and hypopyon (a dense inflammatory coagulum) is usually present; in addition, descemetocele (ulcer penetrated through cornea) formation or corneal perforation are not uncommon. Animal models of bacterial keratitis continue to be of value in the study of this disease and are produced by topical bacterial application after abrading the epithelium, by intrastromal inoculation, or by placing a contaminated suture or contact lens on the cornea. These approaches and models have led to increased understanding of the mechanisms of corneal inflammation and innate immunity that are operative in bacterial keratitis. Bacterial eradication by neutrophils (PMN) involves phagocytosis, lysosomal degranulation and bacterial killing within the acidic lysosomal compartment of the cell. Phagocytosis and intracellular degranulation by PMN also involve oxidative attack, production of toxic oxygen metabolites, triggering of the respiratory burst, biosynthesis of superoxide anions, and other oxidizing agents such as hydrogen peroxide and formation of peroxynitrite. Phagocytic secretion and lysis result in release of extracellular lysosomal enzymes, including, but not limited to, elastase, collagenase and myeloperoxidase (MPO). These enzymes and the oxygen-derived free radicals cause stromal destruction by breaking down collagen, digesting glycosaminoglycans, and disrupting stromal keratocytes. Nitric oxide (NO) also mediates vasodilation and can be important in bacterial killing as well as in bystander tissue

407

damage. These and other substances released from activated PMN and other inflammatory cells (e.g., macrophages, Mf) contribute to stromal necrosis and corneal edema during bacterial disease. Bacterial endotoxin and exotoxins also stimulate Mf to release biologically active substances including, but not limited to, interleukin (IL)-1, IL-6, and tumor necrosis factor (TNF)-a, cytokines that synergize to elicit inflammatory events. Selected cytokines, certain eicosanoids, and other molecular mediators are also involved in ulceration and angiogenesis during bacterial keratitis. Maintenance of leukocyte recruitment during inflammation requires intercellular communication between infiltrating leukocytes, the epithelium, neuropeptides such as substance P (SP) and vasoactive intestinal peptide (VIP), vascular endothelium, and resident stromal cells. These events are mediated by generation of early response cytokines (e.g., IL-1), the expression of adhesion molecules such as intercellular adhesion molecule-1, and the production of chemotactic cytokines and chemokines.

PMN, Cytokines, and Chemokines If leukocytes such as PMN persist within the cornea, destructive pathology ensues, including stromal scarring and perforation, potentially requiring corneal transplantation. PMN infiltration into inflamed tissue is controlled largely by local production of inflammatory mediators. In the mouse, two members of the CXC family of chemokines, MIP-2 (functional homolog of human IL-8) and KC, are potent chemoattractants and activators of PMN. In corneal infections, MIP-2 has been shown to be the major chemokine that attracts PMN into the P. aeruginosa infected cornea and persistence of PMN in the cornea of susceptible (cornea perforates) C57BL/6 versus resistant BALB/c (no corneal perforation) mice (Table 1) was found to correlate with higher MIP-2 chemokine levels (both mRNA and protein). IL-1, produced by Mf, monocytes, and resident corneal cells also influences PMN infiltration into tissues. When tested, IL-1a and b (mRNA and protein) were elevated in the infected cornea of C57BL/6 over BALB/c mice. Furthermore, after infection, MMP-9 was shown to upregulate chemotactic cytokines/chemokines (IL-1b and MIP-2), contribute to degradation of collagen IV, and overall, enhanced P. aeruginosa keratitis. In contrast, neutralization of IL-1b in infected C57BL/6 mice reduced Table 1

Response to P. aeruginosa corneal infection in mice

C57BL/6 (B6)

BALB/c

Th1 responder Corneal perforation IFN-g, IL-12

Th2 responder Corneal healing IFN-g, IL-10

408

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

disease severity, evidenced by reduction of PMN in cornea (MPO assay), decreased bacterial load, and decreased levels of MIP-2 at both the mRNA and protein levels. The use of caspase 1 (enzyme that processes IL-1b to generate the mature cytokine) inhibitor treatment in C57BL/6 mice confirmed these data, even when inhibitor treatment was initiated 18 h after disease onset. In addition, improvement was augmented when the caspase 1 inhibitor was given after infection together with the antibiotic ciprofloxacin. A live attenuated P. aeruginosa vaccine also has been tested and found to elicit outer membrane protein-specific active and passive protection against corneal infection.

T Cells and IL-12 The role of other cells, such as T cells in P. aeruginosa corneal infection, was first studied in inbred C57BL/6 wild type and cytotoxic, CD8(+) T deficient, b2 microglobulin knockout mice (on the C57BL/6 background). Corneas of both groups of mice perforated by 7 days post-infection (p.i.) and histopathology was similar, with infiltration of PMN within 24 h p.i. In contrast, corneas of wild-type mice antibody depleted of helper, CD4(+) T cells and infected with P. aeruginosa did not perforate at 7 days p.i., versus mice depleted of CD8(+) T cells or treated with an irrelevant antibody. Antibody neutralization of IFN-g before infecting C57BL/6 mice also prevented corneal perforation and was associated with a lower delayed type hypersensitivity response when compared with C57BL/6 mice similarly treated with an irrelevant antibody. These data support that a CD4(+) T cell T helper type 1 (Th1) dominant response following P. aeruginosa infection is associated with genetic susceptibility and corneal perforation in C57BL/6 mice and provided the first evidence that CD4(+) T cells are important in development of severe keratitis and eventual corneal perforation. In addition, use of gene array studies confirmed a Th1 versus Th2 bias of C57BL/6 versus BALB/c mice to infection with Pseudomonas. Other studies investigated whether IL-12 (IL-12 p40) was associated with IFN-g production and the susceptibility response of C57BL/6 mice after P. aeruginosa challenge. IL-12 p40 knockout mice (C57BL/6 background) versus wild-type mice were tested to examine disease progression in endogenous absence of the cytokine. When tested, both groups of mice were susceptible to corneal challenge with P. aeruginosa and corneal perforation was observed at 5–7 days p.i. Semi-quantitative reverse transcription polymerase chain reaction (RT-PCR) and enzyme-linked immunosorbent assay analyses confirmed that IL-12 p40 message and protein levels were elevated after infection in the wild type over the expected absence of IL-12 p40 in the knockout mouse cornea. Immunostaining for IL-12 in wildtype C57BL/6 mice revealed that stromal PMN were at least one source of the cytokine.

IL-18 and IFN-g The role of IL-18 and IFN-g production in the resistance response of the predominantly Th2 responding BALB/c mouse was also tested. Semi-quantitative RT-PCR detected IFN-g mRNA expression levels in the cornea of infected mice at 1–7 days p.i. Cytokine levels were significantly upregulated when compared with control, uninfected normal mouse corneas. Using RT-PCR, IL-18 mRNA expression was detected constitutively in the normal, uninfected cornea, but levels were significantly elevated 1–7 days p.i. To test whether IL-18 regulated IFN-g production, BALB/c mice were injected with an anti-IL-18 monoclonal antibody. Treatment decreased corneal IFN-g mRNA levels and both bacterial load and disease severity increased when compared to immunoglobulin G injected control mice. Data provided evidence that IL-18 is critical to the resistance response of BALB/c mice by induction of IFN-g and that IFN-g is required for bacterial killing/ stasis in the cornea. Another separate study showed that the killing effect of IFN-g was indirect, through regulation of NO levels in the infected cornea. IFN-g and SP Further study of the resistance response in BALB/c mice examined the role of the pro-inflammatory neuropeptide, SP in IFN-g production. This study provided evidence that natural killer (NK) cells were required to produce IFN-g; that the cells expressed the neurokinin-1 receptor (NK1R, the major SP receptor); and that they directly and tightly regulated IFN-g production through SP interaction with this receptor, suggesting a unique link between the nervous system and development of innate immunity in the cornea. On the other hand, a disparate distribution of SP in the infected cornea of susceptible C57BL/6 (higher levels) versus resistant BALB/c (lower levels) mice also has been documented and blocking the interaction of SP with its major receptor (NK1R) in C57BL/6 mice improved disease outcome, supporting a role for SP in development of the susceptible phenotype after P. aeruginosa corneal infection. Thus, the amount of SP and its interaction with available NK1R sites after infection contributes either to resistance or susceptibility, and appears an important component of keratitis outcome in these murine models.

Neuropeptides SP In this regard, SP involvement in the pathophysiology of inflammatory disease generally has been evidenced by aberrant levels of SP and SP containing nerve fibers,

Immunopathogenesis of Pseudomonas Keratitis

as well as NK1R, in diseased tissues. SP has been shown to elicit cytokine secretion (IL-2, IL-4, IL-10, and IFN-g) from mouse T cells. In addition, it was demonstrated that human bronchial epithelial cells produce IL-6, IL-8, and TNF-a after SP treatment. SP-induced cytokine production and secretion by leukocytes, including T cells, Mf, and dendritic cells leads to the release of a number of inflammatory mediators such as additional cytokines, oxygen radicals, arachidonic acid derivatives, and histamine, all of which further amplify the inflammatory response. VIP Recent studies have also provided ample evidence for another neuropeptide, VIP, functioning as a potent endogenous anti-inflammatory molecule affecting the immune response antithetically when compared to SP. VIP regulates inflammatory mediators via several transduction pathways and transcription factors essential for gene activation, such as nuclear factor-kappa B, interferon regulatory factor-1 (IRF-1), mitogen-activated protein kinase (MAPK), and cAMP response element. VIP downregulates the production of several pro-inflammatory cytokines, including: TNF-a, IL-1, IL-6, IL-12, and IFN-g, while stimulating production of anti-inflammatory cytokines IL-10, IL-1 receptor (R) antagonist, and TGF-b. Investigation of the effect of VIP in a murine endotoxin challenge model showed that after treatment with VIP, levels of TNF-a and IL-6 in serum and peritoneal fluid were reduced by almost 50%. Regarding the eye, VIP treatment converted the susceptible phenotype (corneal perforation) to resistant (no perforation) in a mouse model of P. aeruginosa-induced infection via downregulation of pro-inflammatory mediators, upregulation of anti-inflammatory molecules, and modulation of host inflammatory cell activation. Thus VIP, a 28-amino-acid peptide, delivered by several types of neurons to immune organs and lymphoid tissues in the heart, gastrointestinal tract, lungs, kidney, cornea and skin, is anti-inflammatory. In fact, evidence indicated a differential response to VIP between infected BALB/c (more) and C57BL/6 (less) mice, due to disparate VIPR1 expression by Mf (which can be induced in a dose-dependent manner by VIP itself). Mf are known to play a key role in regulating/ balancing pro- and anti-inflammatory activity in the resistant (BALB/c) and susceptible, C57BL/6 murine models; therefore, evidence that VIP influences the functional behavior of these cells further supports a more salient role for this neuropeptide in regulating the inflammatory response.

TLR Mouse eye infection models have also been used to study the role of Toll receptors in disease. The Toll family of

409

receptors, conserved throughout evolution from flies to humans, is central in initiating innate immune responses. This family of receptors, composed of trans-membrane molecules, links the extracellular compartment where contact and recognition of microbial pathogens occurs and the intracellular compartment, where signaling cascades leading to cellular responses are initiated. Gene array data showed that the expression of TLRs and related molecules including CD14, soluble IL-1 receptor antagonist, TLR-6, and IL-18R-accessory-protein were significantly elevated in susceptible (C57BL/6) versus resistant (BALB/c) mice following challenge with P. aeruginosa. In another model system involving induction of sterile keratitis, when C3H/HeJ (Toll-like receptor 4, TLR4, point mutation) versus control mice were treated with lipopolysaccharide (LPS) from P. aeruginosa, a significant increase in stromal thickness and haze was seen in the cornea of TLR4 sufficient control, but not in TLR4 deficient mutant mice. The severity of disease coincided with PMN stromal infiltration, indicating that TLR4 signaling enhances corneal disease. In contrast, in a bacterial keratitis model, TLR4 mRNA expression was markedly increased in the cornea of resistant BALB/c mice after bacterial infection. These data led to testing corneas from TLR4-deficient and wild-type control (BALB/c) mice after challenge with live P. aeruginosa to determine the role of TLR4 in bacterial keratitis. Given that TLR4 deficiency was suggested to be protective in sterile keratitis, we might predict that corneas of TLR4-deficient mice would be less susceptible and exhibit a decreased inflammatory response to bacterial infection. In marked contrast, TLR4-deficient versus control mice exhibited significantly increased inflammation and corneal perforation instead of healing after infection. Furthermore, data from clinical score, slit lamp, and histopathology confirmed that TLR4-deficient versus wild-type control mice exhibited significantly increased corneal disease with more opacity and more severe stromal swelling and destruction. In addition, bacterial load (more than 10-fold higher) and PMN recruitment (MPO activity) were markedly upregulated in the infected cornea of TLR4-deficient versus TLR4-sufficient control mice. The data provide strong evidence that TLR4 is required for the resistance response of BALB/c mice to P. aeruginosa challenge and, unlike the sterile keratitis model, TLR4 is required for disease resolution in bacterial keratitis. Overall, it appears conflicting that TLR4 is critical in the pathology of corneal disease in sterile keratitis, while it is protective in bacterial keratitis and required for host resistance. In fact, these data illustrate the characteristic double-edge sword activity of TLR4 activation. On the one hand, in the keratitis model, TLR4 recognized LPS, a component of P. aeruginosa, and initiated an innate immune response that was important for bacterial clearance. TLR4 deficiency impaired bacterial clearance, led

410

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

to overgrowth of bacteria, an overwhelming PMN infiltrate and excessive pro-inflammatory cytokine production. This in turn contributed to corneal destruction and perforation. In the sterile keratitis model, activation of LPS-TLR4 signaling in TLR4 sufficient mice, lead to pro-inflammatory cytokine production and PMN infiltration, increased stromal thickness and contributed to haze production which increased, albeit transiently, corneal perturbation when compared with TLR4 mutant mice. Negative regulators of TLR are also of importance and recent evidence showed that one of them, single immunoglobulin IL-1R related molecule is differentially expressed in BALB/c (resistant) versus C57BL/6 (susceptible) mice. This Toll receptor is critical in resistance to P. aeruginosa infection in BALB/c mice, functioning to downregulate type 1 immunity and negatively regulating sustained IL-1 and TLR4 signaling. siRNA treatment to knockdown TLR9 was also found to influence the outcome of bacterial keratitis, and lead to reduced inflammation, but with the unwanted effect of decreased bacterial killing.

Apoptosis Delay in apoptosis in bacterial keratitis, as evidenced by terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) staining, also contributes to corneal perforation in C57BL/6 mice. Consistent with this finding, Bcl-2, an anti-apoptotic gene, was significantly upregulated in C57BL/6 mouse cornea at 18 h p.i., suggesting that the delayed onset of apoptosis in C57BL/6 mouse cornea may be, in part, due to upregulation and signaling of this gene. These data are also consistent with previous studies showing that over-expression of Bcl-2 reduces lymphocyte apoptosis in P. aeruginosa-induced pneumonia. In the process of apoptosis, execution of cells largely depends on proteolytic cleavage and activation of caspase 3. In this regard, BALB/c versus C57BL/6 mouse cornea expressed more intense staining for activated caspase 3 at 1 day p.i. compared to the delayed peak intensity in C57BL/6 mice. The two mouse groups were also disparate for expression of caspase 9, with significantly more mRNA expression in BALB/c over C57BL/6 mice. Although only hypothetical, these data suggest that different pathways of apoptosis may be operative in the infected cornea of the two groups of mice. Our studies, using a combination of TUNEL and immunostaining, as well as PMN depletion, also provided evidence that the corneal apoptotic cells identified in both groups of mice were predominantly PMN and confirmed that apoptosis of these cells is delayed in C57BL/6 mice. We hypothesize that earlier apoptosis of PMN in resistant BALB/c mice is consistent with effective elimination of invading bacteria, while inducing minimal tissue damage due to unresolved and persistent inflammation.

Neuropeptides The balance between apoptosis and cell survival, as well as the tissue milieu and timing of apoptosis, is critical in immune defense. In this regard, the neuropeptide SP, mentioned above, is a potent anti-apoptotic regulator and can exacerbate inflammation. SP has been shown to stimulate phosphorylation of the anti-apoptotic molecule Akt in colonic mucosa both in vivo and in vitro, preventing apoptosis in humans with colitis. In another in vitro study, SP induced p53, Bcl-2, and NO expression in peritoneal Mf, blocking apoptotic signals. SP delays spontaneous apoptosis of PMN in a dose-dependent fashion by its interaction with the NK1R in vitro, and this effect could be inhibited by application of the NK1R antagonist GR82334. In this regard, C57BL/6 mice treated with another NK1R antagonist, Spantide I, showed significantly improved disease outcome and an earlier onset of apoptosis, similar to the pattern observed in naturally resistant BALB/c mice. Consistent with earlier onset of apoptosis, mRNA expression of caspase 3 was also significantly upregulated earlier in the cornea of Spantide I, versus control treated animals. The data suggest that the protective mechanism of Spantide I treatment in C57BL/6 mice involves induction of earlier PMN apoptosis in the infected cornea, with less bystander tissue damage. Whether in C57BL/6 mice, the effects of Spantide I are directly mediated via the PMN or indirectly through Mf regulation of PMN also required resolution. Clodronate depletion of Mf with or without Spantide I treatment revealed that in the absence of the Mf, apoptosis was reduced/absent in the cornea. To further explore the role of this cell in susceptibility and resistance, Mf from both groups of mice (C57BL/6 and BALB/c) were incubated in the presence of SP together with LPS or with LPS alone. Significantly fewer apoptotic cells were detected in cells from C57BL/6 mice in the presence of SP and LPS versus LPS alone, while the same combined treatment (SP and LPS) did not decrease the number of LPS induced apoptotic Mf from BALB/c mice. To determine the mechanism for the disparate response to SP treatment between Mf from the two mouse groups, expression levels of the NK1R were comparatively tested. Although cells from both groups expressed the receptor, with a slightly weaker signal in BALB/c cells, after LPS stimulation, mRNA expression for the NK1R was only detected in cells from C57BL/6 mice. These data suggest that the possible mechanism for the absence of the anti-apoptotic effects observed in BALB/c Mf after SP treatment may involve a low level of NK1R expression on the cells and possible rapid depletion of the receptor upon LPS stimulation. In this regard, VIPR1, the major receptor for the neuropeptide VIP, was also reported to be expressed disparately in Mf from C57BL/6 (less) versus BALB/c (more) mice.

Immunopathogenesis of Pseudomonas Keratitis

411

initiated and progresses similar to the human pattern. A photograph of the new animal model is provided in Figure 1. This model will allow us to explore treatment paradigms that have been efficacious in the mouse and other rodent species and may more quickly result in clinical trial of successful candidates. See also: Corneal Epithelium: Response to Infection.

Figure 1 P. aeruginosa infection in Marmoset monkey model.

Mf also may provide distinct activation signals for Th1/Th2 differentiation. In this regard, others have reported that Leishmania major infected Mf enhanced the proliferation and IL-4 secretion of Th2 T cells, but inhibited the response of Th1 T cells. When testing for this possibility, we detected that Mf from C57BL/6 mice expressed significantly more IL-12, while BALB/c Mf expressed more IL-10 after LPS stimulation. IL-10 appears protective in the BALB/c infected cornea, as after subconjunctival injection of clodronate-containing liposomes to deplete these cells, higher levels of IFN-g and lower levels of IL-10 were detected and resistant mice were converted to the susceptible phenotype. It was also reported that VIP treated C57BL/6 mice showed improved disease outcome and increased IL-10 expression after P. aeruginosa corneal infection. Furthermore, the data suggest that SP, which acts in an anti-apoptotic manner toward activated C57BL/6 mouse Mf, may also enhance an IL-12 driven, Th1 type immune response and thus further contribute to the susceptibility of this mouse strain to P. aeruginosa infection.

New Animal Model Without doubt, experimental animal models of bacterial infection with P. aeruginosa have provided us with important insights into mechanisms underlying ocular infection and inflammation and our understanding of the effector and regulatory mechanisms involved in disease continues to grow. However, our understanding and knowledge of the precise mechanisms operative in human cases of keratitis (sterile and infectious) remains much more limited. In this regard, development of a new primate model of keratitis would be a timely and important translational approach to facilitate eventual human application of what has been learned in rodent and other species. The Marmoset monkey, with 98% homology to the human genome, provides such a model. Initial studies have been undertaken in the monkey and show that with scarification of the cornea (similar to that done in the mouse), infection can be

Further Reading Delgado, M., Pozo, D., and Ganea, D. (2004). The significance of vasoactive intestinal peptide in immunomodulation. Pharmacological Reviews 56(2): 249–290. DeVane, L. The fibromyalgia community, substance P: A new era, a new role. http://fmscommunity.org/subp.htm (accessed June 2009). Dinarello, C. A. (2005). Blocking IL-1 in systemic inflammation. Journal of Experimental Medicine 201(9): 1355–1359. Ferguson, T. A. and Griffith, T. S. (2007). The role of Fas ligand and TNF-related apoptosis-inducing ligand (TRAIL) in the ocular immune response. Chemical Immunology and Allergy 92: 140–154. Harrison, S. and Geppetti, P. (2001). Substance P. International Journal of Biochemistry and Cell Biology 33(6): 555–576. Hazlett, L. D. (2004). Corneal response to Pseudomonas aeruginosa infection. Progress in Retinal and Eye Research 23(1): 1–30. Janeway, C. A., Jr. and Medzhitov, R. (2002). Innate immune recognition. Annual Review of Immunology 20: 197–216. Kernacki, K. A., Barrett, R. P., Hobden, J. A., and Hazlett, L. D. (2000). Macrophage inflammatory protein-2 is a mediator of polymorphonuclear neutrophil influx in ocular bacterial infection. Journal of Immunology 164(2): 1037–1045. Lighvani, S., Huang, X., Trivedi, P. P., Swanborg, R. H., and Hazlett, L. D. (2005). Substance P regulates natural killer cell interferon-gamma production and resistance to Pseudomonas aeruginosa infection. European Journal of Immunology 35(5): 1567–1575. McClellan, S. A., Huang, X., Barrett, R. P., van Rooijen, N., and Hazlett, L. D. (2003). Macrophages restrict Pseudomonas aeruginosa growth, regulate polymorphonuclear neutrophil influx, and balance pro- and anti-inflammatory cytokines in BALB/c mice. Journal of Immunology 170(10): 5219–5227. Meek, B., Speijer, D., de Jong, P. T., de Smet, M. D., and Peek, R. (2003). The ocular humoral immune response in health and disease. Progress in Retinal and Eye Research 22(3): 391–415. Rudner, X. L., Kernacki, K. A., Barrett, R. P., and Hazlett, L. D. (2000). Prolonged elevation of IL-1 in Pseudomonas aeruginosa ocular infection regulates macrophage-inflammatory protein-2 production, polymorphonuclear neutrophil persistence, and corneal perforation. Journal of Immunology 164(12): 6576–6582. Strand, F. L. (1999). Neuropeptides: Regulators of Physiological Processes. Cambridge, MA: MIT Press. Substance P: A modulator of inflammation. (1998). http://www. woongbee.com/Cytokine/Cytokine%20bulletin/Spring%201998/ spring1998-3.htm (accessed June 2009). Szliter, E. A., Lighvani, S., Barrett, R. P., and Hazlett, L. D. (2007). Vasoactive intestinal peptide balances pro- and anti-inflammatory cytokines in the Pseudomonas aeruginosa-infected cornea and protects against corneal perforation. Journal of Immunology 178(2): 1105–1114. Taylor, P. R., Martinez-Pomares, L., Stacey, M., et al. (2005). Macrophage receptors and immune recognition. Annual Review of Immunology 23: 901–944. Todar, K. (2008). Todar’s Online Textbook of Bacteriology: Immune Defense against Bacterial Pathogens: Innate Immunity. http:// textbookofbacteriology.net/innate.html (accessed June 2009). Winkler, J. D. (ed.) (1999). Apoptosis and Inflammation. Basel: Birkhauser Verlag.

412

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

Relevant Websites http://www.diseasesdatabase.com – Diseases Database Ver 1.8; Medical lists and links Diseases Database, Vasoactive Intestinal Peptide.

http://www.macrophages.com – Macrophages.com. http://users.rcn.com – RCN Corporation: Apoptosis. http://users.rcn.com – RCN Corporation: Innate Immunity. http://www.ResearchApoptosis.com – Research Apoptosis.

Immunobiology of Acanthamoeba Keratitis J Y Niederkorn, University of Texas Southwestern Medical Center, Dallas, TX, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Clodronate – Drug that preferentially kills macrophages, but has no known deleterious effects on other cells of the innate or adaptive immune responses. Interferon-g (IFN-g) – Cytokine produced by T cells that activates macrophages and enhances their capacity to kill Acanthamoeba trophozoites. Mannose-binding protein (MBP) – Lectin receptor that is expressed on Acanthamoeba trophozoites and facilitates their adherence to mannosylated proteins that are expressed on corneal epithelial cells. Major histocompatibility complex (MHC) class-I antigens – Antigens that allow cytotoxic T lymphocytes to recognize and kill cells infected with viruses and some protozoal parasites. Mannose-induced protease 133 (MIP-133) – The 133-kDa protease that is induced when Acanthamoeba trophozoites engage mannose in the cell walls of bacteria or upon mannosylated proteins on corneal epithelial cells. This protease facilitates invasion of Acanthamoeba trophozoites by degrading basement membranes and it also induces apoptosis of corneal cells. Mucosal immunity – Immune responses that are primarily in the form of secretory immunoglobulin A (IgA) antibodies, which preferentially accumulate in the milk, tears, and in mucosal secretions. This form of T-cell-dependent immunity acts primarily to prevent pathogens from entering the body via mucosal surfaces and rarely directly kills microorganisms. Trophozoite – Amoebic phase of Acanthamoeba spp. Acanthamoeba spp. can exist as either dormant cysts or as the active trophozoite (amoebic) phase. Unlike cysts, trophozoites are invasive, produce pathogenic proteases, and directly kill host cells by apoptosis and direct cytolysis.

Introduction Acanthamoeba spp. are the causative agents for Acanthamoeba keratitis (AK) and can be isolated from virtually any terrestrial, aquatic, and marine environment. Viable Acanthamoeba

spp. have even been isolated from eyewash stations, bottled water, and contact lens cases of asymptomatic contact lens wearers. Acanthamoeba can exist either as a dormant cyst or as the active amoebic stage called the trophozoite. Trophozoites are the active vegetative stage that normally exist as free-living amoebae and feed on bacteria and fungi. Trophozoites are approximately the size of a leukocyte (10–25 mm) and are readily identified by their spiny pseudopodia that give them a sea urchin-like appearance. Acanthamoeba cysts are the dormant stage and are approximately half the size of the trophozoite. The cyst wall is comprised primarily of protein and cellulose. Interestingly, the latter molecule is not normally found in animals, but is restricted to members of the plant kingdom including bacteria and fungi. The cyst is remarkably resistant to environmental agents and can remain viable even after 20 years of storage at room temperature or following treatment with over 250 000 rads of gamma irradiation or doses of ultraviolet B (UVB) irradiation that are known to kill every category of mammalian cells. Although AK is believed to be caused by corneal infections produced by trophozoites adhering to contact lenses, cysts can also adhere to contact lenses, and, under certain circumstances, can produce corneal infections in experimental animals. Cysts can persist in corneal tissue for up to 31 months following antiamoebic treatment and may be the underlying cause for recrudescence in patients, especially those who receive corneal transplants to restore the vision lost as a consequence of AK. Corticosteroids are often used to extinguish the inflammation that is provoked in AK. However, corticosteroids have been shown to induce cysts to excyst and transform into infectious trophozoites. Moreover, corticosteroids activate trophozoites and render them more pathogenic and invasive. Thus, corticosteroid treatment may unwittingly exacerbate AK and contribute to the recrudescence that has been reported in AK patients who are treated with topical corticosteroids to prevent immune rejection of their corneal transplants. In spite of the ubiquitous distribution of Acanthamoeba spp. in the environment and widespread contact lens wear, AK is remarkably rare. Moreover, environmental exposure to Acanthamoeba spp. is commonplace; up to 100% of the normal individuals with no history of AK possess serum antibodies to Acanthamoeba antigens, suggesting previous environmental exposure to Acanthamoeba spp. Viable Acanthamoebae can be isolated from the contact lens cases of individuals with no symptoms or history of AK. Collectively, these findings suggest that a large portion of the population is exposed to Acanthamoeba spp. and

413

414

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

may express some degree of immunity against corneal Acanthamoeba infection. This proposition begs the obvious question as to which immune component provides protection and how does it do it.

infectious agents, thereby providing much-needed time for the development of the adaptive immune response.

Anatomical and Physiological Barriers to Corneal Infections with Acanthamoeba

In vitro studies have demonstrated that macrophages are capable of killing Acanthamoeba trophozoites. Moreover, depletion of periocular macrophages by subconjunctival injection of liposomes containing the macrophagicidal drug clodronate results in a dramatic exacerbation of AK in experimental animals. Likewise, in vitro or in vivo exposure to liposomes containing interferon-g (IFN-g) activates macrophages, increases their capacity to kill Acanthamoeba trophozoites, and mitigates AK in experimental animals. In animal models, AK is a self-limiting disease that resolves in 4–5 weeks. However, depletion of conjunctival macrophages results in progressive AK that does not resolve and mimics the human counterpart. The extraordinarily low incidence of AK is not commensurate with the enormous number of contact lens wearers and the ubiquitous distribution of Acanthamoeba spp. This suggests that the presence of an additional risk factor is involved in the development of AK. It is tempting to speculate that patients who develop AK represent a small population of individuals who have underlying deficiencies in their conjunctival macrophage population or altered macrophage function.

Clinicians have long suspected that contact lenses served as vectors for transmitting Acanthamoeba trophozoites to the corneal surface, and that antecedent injury to the corneal surface created epithelial defects that permitted trophozoites to gain a foothold in the cornea. Less than one-third of the cases of AK involve both eyes, even though contact lens wearers typically store their lenses in the same contact lens case, use the same contact lens solutions, and use the same finger for inserting their contact lenses. If a preexisting corneal epithelial barrier defect were not necessary, one would expect that the overwhelming majority of AK cases would be bilateral. Studies in animal models of AK confirm the importance of preexisting corneal epithelial defects in the development of AK that is produced by applying Acanthamoeba-laden contact lenses to the abraded corneal surface. However, other animal studies have shown that breaching the intact corneal epithelium by direct injection of trophozoites into the corneal stroma produces AK. Thus, the intact corneal epithelium provides a barrier for preventing the establishment of ocular infections with Acanthamoeba trophozoites. In addition to the physical barrier provided by an intact corneal epithelium, the ocular surface is bathed in tears, which contain multiple factors that inhibit trophozoite binding and cytopathic effects. Although never formally proven, it is suspected that the shear forces produced by the blinking eyelid interfere with trophozoite adherence to the corneal epithelium and reduce the likelihood of the trophozoites gaining a foothold in the cornea.

Innate Immune System and Resistance to Acanthamoeba Infections The immune system is divided into two functionally distinct components: (1) the innate immune apparatus – which is characterized by its nimble response, but lack of antigen specificity – and (2) the adaptive immune apparatus – which, although slower in its response, provides long-lasting immunity and memory. Elements of the innate immune apparatus are the first responders via their detection of pathogen-associated molecular patterns (PAMPS) that are widely and promiscuously expressed by many microorganisms. By utilizing PAMPS, macrophages and neutrophils are able to rapidly identify invading microorganisms and mount an initial response that restrains the

Role of Macrophages in the Resistance to Acanthamoeba Infections

Role of Neutrophils in the Resistance to Acanthamoeba Infections The neutrophil is another constituent of the innate immune system that plays an important role in the resistance and resolution of AK. Neutrophils are consistently found in AK lesions – both in patients and in experimental animals. Neutrophils are highly effective in detecting the presence of both trophozoites and cysts. They kill both cysts and trophozoites in a myeloperoxidase-dependent manner. Blocking chemotactic responses of conjunctival neutrophils or depleting neutrophils with anti-neutrophilic antiserum results in progressive AK. Likewise, intracorneal injection of macrophage inflammatory protein-2 (MIP-2) – a potent chemoattractant for neutrophils – results in a swift infiltration of neutrophils into the central cornea and in an accelerated resolution of AK in experimental animals. Ocular Acanthamoeba infections rarely progress beyond the cornea and are not known to produce endophthalmitis. Only three reports in the literature suggest that Acanthamoeba infections of the cornea progress to the posterior segment of the eye and involve the choroid or retina. Moreover, only one publication provides histopathological documentation of Acanthamoeba cysts in the posterior segment of the eye in a single AK patient. Moreover, the patient in this study had received four

Immunobiology of Acanthamoeba Keratitis

separate corneal transplants in the affected eye. Likewise, more than 16 years of experience with both the pig and Chinese hamster models of AK has failed to produce any evidence of Acanthamoeba infections progressing posterior from the cornea or involving the uveal tract or retina. In vitro studies have shown that trophozoites can penetrate Descemet’s membrane and are, theoretically, capable of entering the anterior chamber. Further studies showed that intraocular injection of 1 million trophozoites into the anterior chamber of the eye in Chinese hamsters did not produce intraocular infection. A swift neutrophilic infiltrate eliminated the injected trophozoites without inflicting collateral damage to the intraocular tissues. Humoral Factors of the Innate Immune System that Affect Resistance to Acanthamoeba Infections Tears contain a potpourri of antimicrobial factors that protect the ocular surface from pathogenic insults. Among these are lysozyme, lactoferrin, and complement components. Lysozyme is active against Gram-negative bacteria and some fungi, but is ineffective against Gram-positive bacteria. By contrast, lactoferrin and transferrin are effective in controlling Gram-positive bacteria. Complement is present in tears and can be activated by the alternate pathway via bacterial products or by the classical pathway by antibodies. Thus, the complement system straddles the innate and adaptive immune systems. Complement appears to have little or no effect in controlling AK, as pathogenic Acanthamoeba spp. express complement-regulatory proteins – including decay-accelerating factor – which disable the complement system. Tears and milk also contain a factor that inhibits trophozoite adherence and cytolysis of corneal cells. Interestingly, the factor in both milk and tears is not an Ig. The milk-borne factor is >100 kDa and is inactivated by proteinase K, indicating that it is a protein. The two major proteins in milk – a1-antitrypsin and a1antichemytrypsin – are not the milk-borne factors, as neither of these proteins blocks trophozoite adherence or cytolytic activity. Thus, both humoral and cellular elements of the innate immune system can contribute to the resistance to AK (Table 1). Table 1

415

Adaptive Immune System and Resistance to Acanthamoeba Infections Elements of the innate immune system serve as the first responders to pathogens and are characterized by capacity to act immediately in response to microbial infections. However, the innate immune system acting alone cannot clear all microbial infections. Unlike the innate immune system, the adaptive immune apparatus needs a jump start to generate its effector elements, but – once engaged – is crucial for the recovery from microbial infections and the establishment of immunity to future infections. The innate and adaptive immune systems do not function in isolation, but communicate with each other in a coordinated immune response. Macrophages and dendritic cells present antigens to T and B cells, which process is facilitated by another innate immune system component – complement. T and B cells have the capacity to generate an endless array of antigen receptors that, when confronted with antigens expressed on antigen-presenting cells, culminates in the generation of antibodies and T cells that possess exquisite specificity and are used by antibodies to identify and kill bacteria and neutralize viruses. T cells utilize their antigen receptors and CD8 surface molecules to identify and kill virus-infected cells. The adaptive immune system is characterized by its exquisite specificity and memory. The efficacy of preventive immunization relies entirely on the capacity of the vaccine to activate crucial elements of the adaptive immune system. One has to look no further than the biology of acquired immune deficiency syndrome (AIDS) to recognize the importance of the adaptive immune system. With few exceptions, recovery and survival from microbial infections is contingent upon the effective activation of the adaptive immune system. However, AK is a notable exception to this rule. There is no evidence to date that patients with AIDS have an increased incidence of AK, suggesting that a disabled adaptive immune system does not increase the susceptibility to corneal infections with Acanthamoeba even though Acanthamoeba spp. express antigens that are capable of activating the adaptive immune system. Fifty to one hundred percent of the individuals with no past history of AK possess serum and tear antibodies specific for Acanthamoeba antigens – indicating that Acanthamoeba antigens can

Elements of the innate immune system that control ocular Acanthamoeba infections

Component

Function

Evidence for role in protection

Neutrophils

Kill trophozoites

Macrophages

Kill trophozoites

Nonimmunoglobulin tear-borne factor

Prevents trophozoite adherence to cornea

Treatment with anti-neutrophil serum exacerbates AK; stimulating neutrophil infiltration into the cornea mitigates AK; neutrophils kill trophozoites in vitro. Depletion of conjunctival macrophages exacerbates AK; activating conjunctival macrophages mitigates AK; macrophages kill trophozoites in vitro. In vitro assays demonstrate that tears from normal animals block adherence of trophozoites to corneal cells and prevent trophozoite-mediated cytolysis of corneal cells.

416

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

activate the adaptive immune system and that exposure to these antigens is remarkably widespread. The T-cell arm of the adaptive immune system is also activated by Acanthamoeba antigens, as 50% of the normal population with no known history of AK demonstrate T-cell lymphoproliferative responses to Acanthamoeba antigens. In spite of this, patients whose initial AK has been brought under control with antimicrobial therapy can experience recrudescence. This suggests that even if the adaptive immune system has been activated, it is ineffectual in controlling AK. This impression has been confirmed in animal models of AK. Both in Chinese hamsters and pigs, subcutaneous immunization with Acanthamoeba antigens elicits robust IgG antibody and T-cell responses, yet fails to protect the animals against ocular challenge with Acanthamoeba trophozoites. Moreover, resolution of AK does not render animals resistant to second ocular infections with trophozoites; repeat ocular Acanthamoeba infections are as severe as the primary infections.

Role of the Mucosal Immune System in Resistance to AK On first blush, it would appear that the adaptive immune system is utterly incapable of preventing or resolving AK. However, animal studies and serological studies on human AK patients suggest that secretory IgA antibodies in the tears may prevent initial corneal Acanthamoeba infections. It is widely accepted that corneal infection with Acanthamoeba spp. begins when trophozoites adhere to mannosylated proteins on the corneal epithelium. Mannosylated proteins appear to be key ligands for trophozoite attachment – which is facilitated by the expression of a 136-kDa mannose-binding protein (MBP) that is expressed on the trophozoite cell membrane. Mannose expression on the corneal epithelium is dramatically upregulated by contact lens wear or by mild trauma to the Table 2

ocular surface. In addition to promoting the adhesion of trophozoites, mannose stimulates the trophozoites to secrete a 133-kDa pathogenic protease (mannose-induced protease-133 (MIP-133)), which facilitates trophozoite invasion and cytolysis of corneal cells. Upregulation of mannosylated protein expression on the corneal epithelium is crucial for the establishment of AK. However, blocking the interaction between mannosylated glycoproteins on the corneal epithelium and the MBP on the trophozoite cell membrane is an effective strategy for preventing AK. Immunization through mucosal surfaces such as the gastrointestinal tract preferentially stimulates the generation of secretory IgA antibodies that appear in the tears and in mucosal secretions. Animals immunized orally with the Acanthamoeba MBP develop secretory IgA antibodies in the tears, which prevent the development of AK. Secretory IgA antibodies block the adhesion of Acanthamoeba trophozoites to corneal epithelial cells, and thereby, disrupt the first key step in the pathogenic cascade of AK. Anti-MBP secretory IgA antibodies are not toxic to trophozoites, even in the presence of complement and are ineffectual if they are generated after a corneal infection had been established. Thus, IgA antibodies in the tears are the only known component of the adaptive immune system that has an effect on the development of AK (Table 2).

Evading the Adaptive Immune Response Exposure to Acanthamoeba antigens is commonplace as evidenced by the high incidence of serum antibodies and T-cell responses to Acanthamoeba antigens in individuals with no previous history of Acanthamoeba infections. In experimental animals, subcutaneous and intramuscular immunization with Acanthamoeba antigens elicits high titers of IgG antibodies and T-cell activation, yet fails to protect against AK. Although the adaptive immune

Role of adaptive immunity in preventing ocular acanthamoeba infections

Immune element

Role in AK

Evidence for effect

Serum IgG

None

Delayed-type hypersensitivity Complement

None

Cytotoxic T lymphocytes

None

Secretory IgA

Prevents initial corneal infection

Animals and humans with anti-Acanthamoeba IgG antibodies can develop AC. AntiAcanthamoeba antiserum fails to kill trophozoites in vitro, even in the presence of complement. Subcutaneous immunization with Acanthamoeba antigens induces robust delayed-type hypersensitivity to Acanthamoeba but fails to prevent or mitigate AK. Trophozoites resist complement-mediated lysis due to their expression of complement decay accelerating factor. Trophozoites are extracellular pathogens that do not express the crucial major histocompatibility complex class I restricting elements that are necessary for cytotoxic T lymphocyte function. Mucosal immunization with surface antigens expressed on trophozoites induces production of secretory IgA in the tears; anti-Acanthamoeba IgA prevents trophozoite adherence to corneal cells and prevents trophozoite-mediated cytolysis of corneal cells; passive transfer of IgA monoclonal antibody against Acanthamoeba surface eptiotpes protects animals against ocular Acanthamoeba infections.

None

Immunobiology of Acanthamoeba Keratitis

system is capable of eliminating a wide range of microbial agents, it is ineffectual in preventing corneal infections with Acanthamoeba or recrudescence of previous corneal infections. Acanthamoeba spp. employ at least three strategies to evade immune elimination: (1) inactivating immune effector elements; (2) escaping detection by cytotoxic T lymphocytes; and (3) forming dormant cysts that escape immune detection (Table 3). Acanthamoeba trophozoites can activate the complement cascade via the alternate pathway or by the direct pathway by complement-fixing IgG antibodies that recognize surface determinants on the trophozoite cell membrane. Complement components are present in the tears and at the ocular surface, yet trophozoites escape complement-mediated cytolysis – both in vitro and in vivo. The ability of trophozoites to elude the ravages of complement is ostensibly due to their expression of complement-regulatory proteins (i.e., decay-accelerating factor) – which disable the complement cascade. Complement-regulatory proteins are also present on the corneal epithelium and may further interfere with complementmediated cytolysis of trophozoites at the ocular surface. Trophozoites secrete a variety of proteases that are known to degrade Igs, including secretory IgA antibodies in the tears. The importance of this evasive strategy is limited, as IgA antibodies in the tears can prevent the initial establishment of Acanthamoeba infections of the cornea. Acanthamoeba trophozoites are extracellular pathogens and thus, escape detection and elimination by cytotoxic T lymphocytes – which are only able to kill intracellular pathogens whose antigens are presented on host cells that are expressing major histocompatibility complex (MHC) class-I molecules. Table 3

Acanthamoeba immune escape mechanisms

Factor or mechanism Secretion of serine proteases Complement decay accelerating factor Encystment

Immune privilege of cornea

Extracellular infection

Potential effect in Acanthamoeba keratitis Degrade immunoglobulins; induce apoptosis of leukocytes Inactivate complement and disable complement-fixing antibodies Escape detection by inflammatory and immune cells; resistant to phagocytosis by non-activated macrophages; weakly immunogenic; do not produce chemoattractants for leukocytes Corneal infections fail to elicit IgG antibody or delayed-type hypersensitivity responses to Acanthamoeba antigens Trophozoites are extracellular pathogens and do not reside in host cells that express MHC class I antigens. Thus, they escape recognition and attack by cytotoxic T lymphocytes

417

Protozoal parasites have the capacity to encyst and, thereby, escape immune detection and elimination. Acanthamoeba spp. are one of the few protozoal parasites of humans that form cysts in tissues. Although most intestinal protozoal parasites can encyst, the cysts reside transiently within the lumen of the large intestine and are typically eliminated in the feces. Toxoplasma gondii can encyst within mammalian cells and not in extracellular sites. Acanthamoeba cysts occur in tissue and have been reported to persist in corneal tissue for up to 31 months. Acanthamoeba cysts are remarkably resistant to a variety of agents and can remain viable for decades. Moreover, Acanthamoeba cysts are poor immunogens, have only weak chemoattractive properties, and often reside in an immune-privileged site (i.e., the cornea), thereby enhancing their ability to escape immune detection.

Anti-Disease Vaccine for AK Effector elements of the adaptive immune system appear to be incapable of killing or inactivating Acanthamoeba trophozoites. Thus, immunotherapeutic strategies are limited to two potential targets: (1) preventing the initial adherence of trophozoites to the corneal surface, or (2) inactivating the trophozoite-borne pathogenic molecules that damage the cornea. As stated earlier, secretory IgA antibodies directed against the cell-membrane molecules on the trophozoite can prevent the establishment of corneal infections by blocking the adherence of trophozoites to the corneal epithelium, but only if the antibodies are present in the tears when trophozoites are first introduced to the corneal surface. IgA antibodies are unable to alter the course of disease once the trophozoites have gained a foothold in the cornea. Trophozoites elaborate a variety of proteases that produce extensive cytolysis and apoptosis of corneal cells, degradation of basement membranes, and melting of the stroma. One of the most important trophozoite-borne proteases is a 133-kDa serine protease – MIP-133 – which is secreted in response to trophozoites binding to mannosylated glycoproteins on the corneal epithelium. Mucosal immunization with MIP-133 elicits the generation of secretory IgA antibodies that inhibit the cytolysis of corneal cells and neutralize the enzymatic degradation of stromal proteins. Hosts mucosally immunized with MIP-133 have much milder disease compared to controls. An attractive feature of eliciting mucosal immunity against a pathogenic molecule is that the ocular surface is bathed in tears containing neutralizing antibodies which are continuously replenished. An anti-disease vaccine could be implemented at the time AK is diagnosed and used in combination with conventional anti-Acanthamoeba chemotherapy.

418

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

Conclusions The immunobiology of AK is puzzling. Acanthamoeba spp. are free-living amoebae that can be found in virtually any environmental niche, yet rarely cause disease. Environmental exposure to Acanthamoeba spp. is commonplace and results in the activation of the adaptive immune response as evidenced by the high percentage of individuals who display antibody and T-cell responses to Acanthamoeba antigens. The low incidence of AK combined with the high frequencies of anti-Acanthamoeba antibody and T-cell immunity suggest that the adaptive immune response protects against corneal infection with Acanthamoeba spp. However, animal studies suggest otherwise and indicate that – with the exception of secretory IgA antibodies – the adaptive immune system is incapable of preventing or controlling Acanthamoeba infections. By contrast, the innate immune apparatus appears to control ocular Acanthamoeba infections. This article offers the following theory to explain the immunobiology of AK. Environmental exposure to Acanthamoeba spp. is common and occurs via mucosal surfaces by breathing airborne Acanthamoebae that can readily be isolated from heating and cooling vents or by ingestion of Acanthamoebae that are found in a variety of vegetables and fruits. Most environmental isolates of Acanthamoeba spp. are nonpathogenic, but are immunogenic. The presence of anti-Acanthamoeba IgA antibodies in the tears protects against ocular infection, especially in individuals without corneal epithelial defects. Occasionally, Acanthamoeba trophozoites will reach the ocular surface and escape detection by IgA antibodies in the tears, but are eliminated by neutrophils and macrophages. The small number of individuals who develop AK might have underlying deficiencies in their tear antibody titers or in their periocular neutrophil and macrophage repertoire.

Such immunological blind spots might account for corneal Acanthamoeba infections. This theory remains to be confirmed, but is consistent with results from animal studies and with clinical findings from AK patients. See also: Adaptive Immune System and the Eye: Mucosal Immunity; Adaptive Immune System and the Eye: T Cell-Mediated Immunity; Innate Immune System and the Eye.

Further Reading Clarke, D. W. and Niederkorn, J. Y. (2006). The immunobiology of Acanthamoeba keratitis. Microbes and Infection 8: 1400–1405. Clarke, D. W. and Niederkorn, J. Y. (2006). The pathophysiology of Acanthamoeba keratitis. Trends in Parasitology 22: 175–180. Khan, N. A. (2003). Pathogenesis of Acanthamoeba infections. Microbial Pathogenesis 34: 277–285. Khan, N. A. (2006). Acanthamoeba: Biology and increasing importance in human health. FEMS Microbiology Reviews 30: 564–595. Kumar, R. and Lloyd, D. (2002). Recent advances in the treatment of Acanthamoeba keratitis. Clinical Infectious Diseases 35: 434–441. Leher, H., Silvany, R., Alizadeh, H., Huang, J., and Niederkorn, J. Y. (1998). Mannose induces the release of cytopathic factors from Acanthamoeba castellanii. Infection and Immunity 66: 5–10. Leher, H., Zaragoza, F., Taherzadeh, S., Alizadeh, H., and Niederkorn, J. Y. (1999). Monoclonal IgA antibodies protect against Acanthamoeba keratitis. Experimental Eye Research 69: 75–84. Li, L. and Sun, X. (2008). Impaired innate immunity of ocular surface is the key bridge between extended contact lens wearing and occurrence of Acanthamoeba keratitis. Medical Hypotheses 70: 260–264. Marciano-Cabral, F. and Cabral, G. (2003). Acanthamoeba spp. as agents of disease in humans. Clinical Microbiology Reviews 16: 273–307. McClellan, K., Howard, K., Niederkorn, J., and Alizadeh, H. (2001). The effect of corticosteroid on Acanthamoeba cysts and trophozoites. Invesigative Ophthalmology and Visual Science 42: 2885–2893. Niederkorn, J. Y., Alizadeh, H., Leher, H., and McCulley, J. P. (1999). The pathogenesis of Acanthamoeba keratitis. Microbes and Infection 1: 437–443.

Molecular and Cellular Mechanisms in Allergic Conjunctivitis V L Calder, UCL Institute of Ophthalmology, London, UK ã 2010 Elsevier Ltd. All rights reserved.

Glossary Allergen – A protein to which the immune system responds. Amblyopia – Vision disorder characterized by poor or indistinct vision in an eye that is otherwise physically normal. Atopy – Allergic response in an area not in contact with allergen. Conjunctiva – Tissue lining of the ocular surface comprised of epithelium and stromal layers. Bulbar conjunctiva covers the outer surface of the eye; tarsal (palpebral) conjunctiva lines the eyelids. HLA-DR – A transmembrane human major histocompatibility complex 2 family member. Hyperemia – Increase of blood flow to a portion of the body. Photophobia – An abnormal sensitivity to light. Rhinitis – Inflammation of some internal areas of the nose. The primary symptom of rhinitis is nasal dripping.

The symptoms of SAC include itching, watering, redness, and swelling of the eyes and lids, and there may be increased discharge. There is often an associated rhinitis and a history of atopy. During the pollen season, the ocular signs can be dramatic, usually affecting both eyes to a similar degree. Edema of the lids and conjunctiva can be mild and often outweighs the degree of hyperemia, giving a milky or pink appearance to the eye. In severe cases the swelling can be gross, with the inability to open the lids and a ballooning out of the conjunctiva termed chemosis, particularly after exposure to high aeroallergen concentrations or after rubbing of the eye. Eversion of the lids to reveal the tarsal conjunctiva demonstrates some hyperemia and mild infiltration of the conjunctiva, leading to a loss of transparency and thickening, with diffuse small inflammatory excrescences known as papillae. Since there is no serious limbal disease or conjunctival scarring, and the cornea is not involved, the visual acuity remains normal. Outside the pollen season the eye examination is normal.

Perennial Allergic Conjunctivitis The term ocular allergy describes a spectrum of clinical conditions, ranging from the common, milder conditions of seasonal and perennial allergic conjunctivitis (SAC, PAC), to the rare but more severe diseases, vernal keratoconjunctivitis (VKC) and atopic keratoconjunctivitis (AKC). This article describes the clinically different subtypes of ocular allergy (classification summarized in Table 1) and our current understanding of the cellular and molecular pathways involved in the different forms of ocular allergy.

Ocular Allergies Seasonal Allergic Conjunctivitis SAC, or hay fever, is the most common form of ocular allergy and, in fact, the most common of all ocular disorders. SAC occurs only during the pollen season, the timing being dependent on which pollen is allergenic for that individual (e.g., tree, grass, weed). In countries where pollen seasons are extended, SAC can affect individuals for up to 10 months per year. It can occur at any age but is more frequently seen in children and young adults and the severity tends to lessen with age.

PAC is another common ocular allergy with many similarities to SAC but with a very different time course. Since the allergens in PAC are present for most or all of the year, PAC is not seasonal. The disorder is again most frequently and severely seen in children and young adults. House dust mite (Dermatophagoides pteronyssinus) is the most common sensitizing allergen but animal hair and dander, molds and other allergens can also induce PAC. The symptoms are perennial and include ocular itch, discomfort, watering, redness, and some discharge. Patients may be able to correlate symptoms with exposure to an allergen. House dust mite allergy sufferers have a history of symptoms worse in the morning. Approximately one-third have an associated rhinitis, and a family and/or personal atopic history is very common. The clinical appearance is of a mild conjunctival inflammation and clinical signs may be very slight. The bulbar conjunctiva may be slightly red and edematous and the tarsal conjunctiva shows mild to moderate hyperemia, infiltration, and fine papillae. Lid edema is usually mild. Similar to SAC, in PAC there is no conjunctival scarring, or corneal involvement, such that vision is not affected. Due to the continuing presence of the allergens, the resultant inflammation in PAC is more chronic and hence the immunopathology of PAC differs from that of SAC.

419

420 Table 1

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease Classification of ocular allergic diseases

Disease

Timing

Age group

Prevalance

Keratopathy

Sight threatening

SAC

Seasonal

No

Perennial Seasonal – Perennial if severe

Very common Common Uncommon

No

PAC VKC

Majority children and young adults Adult Children

Common Yes

No Yes

AKC

Perennial

Adult

Rare

Yes

Yes

Course Mild, nonprogressive, often resolves Not serious, nonprogressive Serious, but usually resolves in 2–10 years with good outcome if well managed. Can change into AKC Serious and progressive, vision often reduced

SAC, seasonal allergic conjunctivitis; PAC, perennial allergic conjunctivitis; VKC, vernal keratoconjunctivitis; AKC, atopic keratoconjunctivitis.

Vernal Keratoconjunctivitis VKC is a more serious ocular allergy of childhood. It makes up 0.1–0.5% of ocular disease in the developed world but is more common and much more severe in hot dry countries, especially the Middle East, West Africa, and the Mediterranean. It is a chronic form of conjunctivitis, usually with seasonal exacerbations. In the United Kingdom, VKC is a rare, self-limiting, often seasonal ocular allergy that affects children and young adults, most of which are male (85%) and many have a history of atopy. The symptoms are worse in the spring and summer but, in severe disease, will last all year. Patients complain of severe itching, discomfort or pain, photophobia, stringy discharge, blurred vision, and difficulty opening the eyes in the morning. The ocular signs can be very asymmetrical. Conjunctival signs are maximal in the superior tarsal conjunctiva and limbus and the heavily inflamed lid may droop (ptosis). The conjunctival surfaces are hyperemic, edematous, and infiltrated, and a stringy mucoid discharge is present. The tarsal conjunctival tissues are densely infiltrated, with papillae that are often giant (>1 mm in diameter, known as cobblestone papillae). The limbus can have discrete swellings or, less often, diffuse hyperemia and inflammation, and the presence of small white chalky deposits (Trantas’ dots) is typical of vernal limbitis. In the later stages, fine reticular white scarring may be seen, but this does not lead to significant shrinkage and distortion of the ocular surface as in some cicatrising conjunctival diseases such as AKC (see below). Visual acuity can be affected by involvement of the cornea (keratopathy), which is most marked in the upper third of the cornea as a result of greater exposure to toxic inflammatory mediators, not mechanical rubbing by the papillae. At its mildest, there is a punctate disturbance of the epithelium, which may coalesce to form a discrete epithelial defect (macroerosion). Deposition of mucus, fibrin, and inflammatory debris can then result in the formation of a shallow oval plaque (or shield) ulcer, which repels the hydrophilic tears and the epithelial-healing

response. Herpetic and bacterial corneal infection may occur. In the later stages, scarring of the cornea may lead to permanent visual reduction. Steroid treatment-related complications, and (because of the young age group) sensorydeprivation amblyopia also contribute to the potential for long-term visual loss. Atopic Keratoconjunctivitis AKC is the least common but most serious of the ocular allergies. It is a life-long condition that affects adults who have systemic atopic disease, either atopic dermatitis or chronic asthma. The usual onset is in the late teens but, unlike VKC, the disease is persistent and may be relentlessly progressive; occasionally the disease can begin in childhood. AKC is a highly symptomatic disorder with severe itching, pain, watering, stickiness, and redness. There is usually facial atopic dermatitis involving the eyelids. The lid margins show severe blepharitis (chronic inflammation of the lash follicles and meibomian glands) and are thickened and hyperemic, posteriorly rounded, sometimes keratinized and the lid anatomy may be distorted with ectropion (outwardly turning eyelid), entropion (inwardly turning eyelid), trichiasis (inturning lashes), loss of lashes, and notching. The whole conjunctiva is affected and shows intense infiltration, papillae (which may be giant) and sometimes scarring with linear and reticular white scar tissue, lid to conjunctiva adhesions and shrinkage or loss of the conjunctival sac and secondary lid distortions. Marked limbal inflammation can develop and Trantas’ dots may occur. The disease may never affect the cornea, in which case it is sometimes referred to as atopic blepharo-conjunctivitis (ABC); in this situation, the overall inflammation is generally less severe. The cornea can be affected as a direct effect of the inflammatory process or may be damaged secondarily following extensive changes to the usually protective ocular surface by processes such as continual mechanical trauma, reduced lid protection, or severe loss of conjunctival tear

Molecular and Cellular Mechanisms in Allergic Conjunctivitis

production. Significant visual acuity reduction due to corneal involvement occurs in 40–70% of the cases. Keratopathy may consist of punctate and macroscopic epithelial defects, filamentary keratitis, plaque ulcer, progressive scarring, neovascularization (with or without lipid deposition), thinning, and secondary corneal infections (herpetic, bacterial, and fungal). Associations between AKC and eye rubbing, keratoconus, atopic cataract, and retinal detachment have been reported.

Cellular Mechanisms in Ocular Allergy The predominating immune mechanism occurring in SAC is an immediate (type 1) hypersensitivity response whereby conjunctival mast cells (MCs) (Table 2) and their secreted products primarily orchestrate the inflammatory response. In contrast, the cellular responses in PAC involve MC to a certain extent, although neutrophils and some T cells have also been detected in the conjunctival tissues, probably recruited as a result of the release of chemokines which attract these cells to the site of inflammation during the persistent allergen-driven inflammatory response. During VKC, studies have identified cells of both innate and adaptive immune responses becoming activated with T lymphocytes and eosinophils predominating, as well as MCs, neutrophils, and other cells infiltrating the conjunctival epithelium and stroma. In AKC, the predominant cell types infiltrating the conjunctival tissues are T cells, eosinophils, and neutrophils. In both VKC and AKC there are alterations to the epithelium and evidence of tissue remodeling and collagen deposition. Conjunctival MCs MCs are important effector cells at all mucosal sites including the conjunctiva, where rapid responses are necessary. During SAC, conjunctival MC become activated as a direct result of allergen cross-linking of surface IgE receptors (FceR1), resulting in degranulation and release Table 2 Summary of cell types in conjunctival tissues in ocular allergy

Disease SAC PAC VKC AKC

Predominant inflammatory cells

Involvement of tissueresident cells

MCCT MCCT and MCCT, neutrophils, T cells Eosinophils, CD4+T cells, neutrophils, MC CD4+T cells, eosinophils, neutrophils, MC

None None

421

of histamine, leukotrienes, proteases, prostaglandins, cytokines, and chemokines (Figure 1). This rapid MC histamine response causes the itching characteristic of SAC. Upon binding to its receptors (H1 and H2), histamine induces vascular leakage, resulting in further cellular infiltration of eosinophils and neutrophils from the blood, leading to chemosis. Relative increases in mucosal MC (MCT) were identified in tarsal conjunctival tissue specimens in SAC whereas increased numbers of both MCT and connective tissue type (MCCT) MC phenotypes have been detected in both tarsal conjunctival epithelial and substantia propria layers in PAC. The different pattern of MC subset activation occurring in PAC probably reflects the more persistent response to allergen. Therapeutic intervention in SAC initially focused on the use of topical antihistamines in the form of eye drops, to neutralize the localized effects of the histamine secreted by the degranulating MC. Another therapeutic approach has used MC stabilizing drugs (e.g., sodium chromoglycate) to inhibit release of histamine and other secretagogues from the cells. Combinations of antihistamines with chromolyns have been used to treat the signs and symptoms of SAC for many years. However, more recent topical antiallergic drugs for treating SAC (e.g., azelastine, epinastine, ketotifen, olopatadine) combine antihistamine action with MC stabilization. The additional benefits of these drugs are in their ability to selectively prevent MC secretion of various inflammatory mediators including histamine, as well as cytokines and chemokines. Due to the presence of other cell types in the more chronic forms of ocular allergy, these antiallergic drugs are most effective in SAC and are not effective for the more severe forms of ocular allergy when used alone, but can be of some benefit when given in combination with other anti-inflammatory drugs such as steroids or cyclosporine A (see below). While specific allergen immunotherapy has been successfully used for treating other forms of allergy where the specific allergens are known, it has not been widely used for treating SAC since there is a wide variation among individuals affected with SAC in terms of their allergen responsiveness, with many

Fibroblasts, epithelium Epithelium, fibroblasts

SAC, seasonal allergic conjunctivitis; PAC, perennial allergic conjunctivitis; VKC, vernal keratoconjunctivitis; AKC, atopic keratoconjunctivitis; MC, mast cells; MCCT, mucosal mast cells; MCCT, connective tissue mast cells.

(a)

(b)

Figure 1 Conjunctival mast cells in vitro : (a) Unstimulated mast cells (small arrows); (b) mast cells stimulated via FecR crosslinking (large arrows). Stimulated mast cells form clusters prior to degranulating.

422

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

responding to more than one allergen. In the chronic forms of allergic eye disease (VKC, AKC), the conjunctivitis is due less to a direct allergen-specific response, but more an immune-mediated response involving local tissue resident cells, and other nonallergen-specific inflammatory cells. In those severe forms of ocular allergy, specific allergen immunotherapy would not be appropriate. Conjunctival Eosinophils and Neutrophils MCs are predominately responsible for SAC, but in more chronic forms of conjunctivitis other cell types have been identified as playing an important role. Some neutrophils and eosinophils have been found in less than half of symptomatic patients with SAC but little or no T cell infiltration was observed. In SAC and PAC, neutrophils and eosinophils have been observed at the site of inflammation, and both of these cell types are able to secrete a wide range of proinflammatory cytokines (interleukin (IL)-3, IL-4, IL-5, IL-6, transforming growth factor alpha (TGF-a), tumor necrosis factor alpha (TNF-a)), chemokines (IL-8, RANTES), and multiple mediators (granule proteins: eosinophil cationic protein, major basic protein, and eosinophil-derived cationic protein) to amplify the inflammatory response. In PAC, due to the presence of eosinophils, neutrophils, and some T cells, several cellmediated processes are likely to be involved. MC-targeted therapy alone is not an effective treatment for PAC, supporting the hypothesis that there is a complex network of cells contributing to the chronic inflammation in this condition. Recently, intranasal corticosteroids have demonstrated a promising therapeutic effect in relieving the ocular symptoms associated with perennial allergic rhinitis. In the more severe forms of ocular allergy (VKC and AKC), increased numbers of eosinophils have been detected in the conjunctival tissues, although it has also been found that it is the extent of eosinophil activation expressing intracellular adhesion molecule (ICAM)-1 or a transmembrane human major histocompatibility complex 2 family member (HLA-DR) that correlates more with disease severity than the overall numbers of eosinophils. There are differences between VKC and AKC in the patterns of cytokines which colocalize to conjunctival eosinophils, with those from VKC mainly expressing IL-3, IL-5, IL-6, and granulocyte/macrophage colony stimulating factor (GM-CSF) whereas, in AKC, eosinophils express mainly IL-4, IL-8, and GM-CSF. Although the specific cellular interactions are as yet unclear, these different cytokine profiles point to different eosinophilmediated pathways being involved in each severe form of ocular allergy. Conjunctival Lymphocytes There are very few T cells detected in normal and in SAC conjunctival tissue specimens, which are mainly situated

in the epithelial layer. There are a few T cells detectable in PAC but their phenotypes remain unknown. In contrast, immunostaining of tarsal conjunctival tissue specimens from VKC patients have found significantly increased numbers of lymphocytes which are mainly activated CD4+ T cells, localized to the subepithelial layers of the affected tissue. There is also an increased HLA-DR expression within the epithelium and stromal layers of the conjunctival tissues as compared with normal subjects and increased numbers of Langerhans’ cells and activated macrophages (CD68+) were also observed. T cell clones, derived from VKC conjunctival tissues, were functionally characterized as Th2-type, since in situ hybridization staining demonstrated an increased Th2 cytokine (IL-3, IL-4, and IL-5) mRNA expression in VKC in areas of maximum T cell infiltration (Figure 2). In support of these studies, VKC tear samples were found to have increased intracellular T cell expression of IL-4 in more than 60% of the specimens. Further analysis of tear specimens using multiplex bead cytokine arrays, found that IL-4, interferon-gamma (IFN-g), and IL-10 were all elevated in SAC and VKC in comparison with nonatopic controls. Although such studies do not identify the cellular source of the cytokines, nevertheless they illustrate the differentially activated cytokine pathways in each form of ocular allergy, perhaps due to the different cell types involved in each form of ocular allergy. Similar to VKC, conjunctival biopsy specimens in AKC were found to have increased numbers of activated CD4+T cells, HLA-DR expression, and cells of the monocyte/macrophage lineage as well as mRNA expression of the Th2 cytokines (IL-3, IL-4, and IL-5) in the stromal tissues. However, in contrast to VKC, there was also a significant increase in the expression of IL-2 mRNA, and in numbers of IFN-g expressing T cells, suggesting a more Th1-type T cell response in the most severe of the ocular allergic diseases. In support of this, conjunctival biopsy specimen-derived T cell lines from AKC were found to secrete significantly increased levels of IFN-g, indicative of Th1-T cells. It has thus been proposed that AKC is an

(a)

(b)

Figure 2 Light microscopy immunostaining of conjunctival biopsy tissue sections for IL-13 expression (brown): (a) weak staining for IL-13 localizing to goblet cells within the epithelial layer in SAC biopsy specimen; (b) intense staining for IL-13 (brown) localizing to mononuclear cells within the subepithelial and epithelial layers in VKC biopsy specimen (magnification 200).

Molecular and Cellular Mechanisms in Allergic Conjunctivitis

423

Mast cell Neutrophil

Immunopathogenesis in ocular allergy Clinical severity

Eosinophil

Dendritic cells

SAC

PAC

VKC

AKC B cells

Th1 T cell

Th2 T cell

Fibroblast activation

Collagen deposition

IL-4 IgE

Fibroblast

IL-4

Allergen Th2 cytokines

IL-8, RANTES

Th2/Th1 cytokines

Blood vessel

IFNγ IgE Histamine Blood vessel vasodilation

IgE

IL-5 IL-5 ECP, EDN, EPO

IgE

ECP, EDN, EPO

Figure 3 Schematic comparing the different types of immunopathogenis in ocular allergy. SAC, seasonal allergic conjunctivitis; PAC, perennial allergic conjunctivitis; VKC, vernal keratoconjunctivitis; AKC, atopic keratoconjunctivitis.

immune-mediated response involving Th1 T cells, whereas VKC involves a predominant Th2-type T cell response (Figure 3). Due to the severity of the inflammation in VKC and AKC, immunosuppressive drugs (steroids) are used to dampen the immune response. However, long-term steroid treatment can have serious side effects in the eye, causing raised intraocular pressure which can lead to glaucoma, and cataract formation. Following the identification of T cells and their cytokines within the conjunctival tissues in VKC and AKC, cyclosporine A (CsA; 2% in maize oil) was tested and found to be an effective steroidsparing treatment for VKC and AKC if administered locally as eye drops. Conjunctival Epithelial Cells As a consequence of chronic inflammation at the ocular surface, in particular in AKC, the epithelium can become

thickened. Immunostaining of conjunctival epithelial cells from conjunctival biopsies have demonstrated an increased expression of ICAM-1 and HLA-DR, but only in the most severe forms of allergic eye disease (VKC and AKC) and almost no expression of these costimulatory molecules in noninflamed control conjunctival tissues. The ability of conjunctival epithelial cells to express ICAM-1 might allow greater adhesion and recruitment of leukocytes, while the expression of HLA-DR molecules could simply reflect the activation status of the cells, although the possibility of conjunctival epithelial cells presenting antigen to T cells is still to be confirmed. Several in vitro studies have used conjunctival epithelial cells, either as primary cultures of cells isolated from biopsy specimens, or as immortalized epithelial cell lines. Upon activation of conjunctival epithelial cells in vitro, there is an upregulation of costimulatory molecules, including ICAM-1 and HLA-DR and secretion of various cytokines such as IL-6, CCL8 (IL-8), a potent

424 Table 3

Control VKC AKC

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease Summary of cytokines and molecules present in conjunctival tissues IL-2

IL-4

IL-5

MMP-1, -3, -9

IFN g

HLA-DR + epithelium

ICAM-1 + eosinophils

+

+ +++ ++

/+ +++ +++

+ ++ nd

+ ++ +++

++ ++

++ ++

/+ +++

VKC, Vernal keratoconjunctivitis; AKC, atopic keratoconjunctivitis; IL, interleukin; MMP, matrix metalloproteinases; IFN, interferon; HLA-DR, a transmembrane human major histocompatibility complex 2 family member; ICAM, intracellular adhesion molecule.

chemokine for neutrophils. The ability of conjunctival epithelial cells to secrete proinflammatory cytokines and chemokines suggests an important proinflammatory role for these cells (Table 3).

Conjunctival Fibroblasts In the severe forms of chronic ocular allergy, there is significant tissue remodeling involving collagen deposition and subepithelial fibrosis. Conjunctival biopsyderived fibroblasts in vitro have been found to secrete cytokines (e.g., IL-6), chemokines (e.g., MCP-1), as well as matrix metalloproteinase (MMP)-1 and -9 and tissue inhibitor of matrix metalloproteinase (TIMP)-1. These cells were also found to respond to the Th2 cytokines IL-4 and IL-13, by secreting increased levels of eotaxin-1, IL-6, and RANTES, whereas there was a significant reduction of MMPs. Immunostaining of conjunctival biopsies from noninflamed controls and VKC identified a significant increase in MMP-1, -3, -9, and -13 expression in VKC tissues. Tear levels of MMP-1 and MMP-9 have also been found to be increased in VKC in comparison with controls, which is probably a reflection of the ongoing fibrotic tissue response.

Molecular Mechanisms The ocular surface of the eye is a mucosal site and plays an important role in protecting the eye from infection through a mucosal barrier as well as supporting both innate and adaptive immune responses as described elsewhere in this encyclopedia. The ocular surface is protected by the presence and continuous production of tears and mucins to prevent binding of antigens, and within the tear fluid, antibodies can be found which will bind to, and activate opsonization of antigens by phagocytes. In addition, the expression of toll-like receptors (TLRs) provides another mechanism at the ocular surface, whereby binding of evolutionarily conserved microbial proteins (pathogen-associated molecular patterns, PAMPs) to these receptors induces innate and adaptive immune responses to break down and remove the invading pathogen. Immunostaining has been used to demonstrate expression of

TLR-2, -4, and -9 in healthy conjunctival tissues, with expression mainly in the stromal layers. In comparison, in VKC, there is an increased expression of TLR-4 on both epithelium and stroma, together with a decrease in TLR-9. In AKC, TLR-2 expression has been found to be increased on human primary cultures of conjunctival epithelial cells following exposure to Staphylococcus aureus, as well as increases in ICAM-1 and HLA-DR expression and secretion of TNF-a and IL-8. This has been proposed as a mechanism whereby S. aureus infection at the ocular surface in AKC could activate a host epithelial cell response. In AKC, S. aureus colonization can occur at the ocular surface, probably due to a compromised mucosal barrier during this severe form of ocular allergy.

Costimulatory Molecules in Ocular Allergy Immunohistochemical studies of tarsal and bulbar conjunctival biopsy specimens demonstrated expression of adhesion molecules, ICAM-1 and e-selectin, to be increased in SAC in comparison with controls. However, this increased expression was only detected during the pollen season and outside the pollen season, the levels returned to those of controls. This pattern of expression correlated with the degree of neutrophil or eosinophil infiltration in the bulbar tissue, suggesting an MC-mediated cell recruitment process. Expression of HLA-DR and ICAM-1 molecules has been used as markers of cell activation. In AKC, there is an upregulation of HLA-DR expression and ICAM-1, localized to the epithelial cells, suggesting epithelial cell activation, probably as a result of exposure to proinflammatory cytokines as well as TLR activation. In both VKC and AKC there is an upregulation of HLA-DR expression and ICAM-1, localized to the eosinophils which correlated with an enhanced activation of these cells. IgE in ocular allergy Total serum IgE levels are significantly increased in VKC than in controls. However, IgE levels are variable among those with ocular allergy and cannot be used as a reliable indicator of disease activity or severity. Studies investigating allergen-specific serum IgE levels have detected a range of allergen specificities. A greater percentage of

Molecular and Cellular Mechanisms in Allergic Conjunctivitis

VKC patients have specific serum IgE against D. pteronyssinus and Dermatophagoides farinae, whereas in SAC the specific serum IgE is against grass pollens. Allergenspecific IgE is also increased in tear specimens and there is a highly significant correlation with ocular allergy symptoms, supporting a diagnostic value for specific tear IgE, although limited tear volume restricts its use in routine immunoassays.

425

side effects and there is an urgent need for improved treatments which can be given topically to reduce the impact of potential side effects. See also: Adaptive Immune System and the Eye: Mucosal Immunity; Adaptive Immune System and the Eye: T CellMediated Immunity; Conjunctiva Immune Surveillance; Defense Mechanisms of Tears and Ocular Surface; Innate Immune System and the Eye; Overview of Electrolyte and Fluid Transport Across the Conjunctiva.

Cytokines and Chemokines in Ocular Allergy Throughout this article, cytokines and chemokines have been discussed in relation to particular cell types. However, cell-free tear specimens have also been studied for the presence of cytokines and chemokines in the various forms of ocular allergy. Studies of VKC tear specimens have detected increased levels of IL-4, IL-10, IFN-g, eotaxin, and TNF-a in comparison with noninflamed control tears. This correlates well with the enhanced expression of Th2-cytokines detected in isolated T cell clones from conjunctival biopsy specimens from VKC patients, as well as the increased percentages of intracellular IL-4-expressing T cells from tear specimens in VKC. In severe forms of AKC, increased tear levels of eotaxin-1 were found to correlate with increased numbers of eosinophils in tears, although the cellular source of the eotaxin was not identified. The production of proinflammatory cytokines and chemokines by infiltrating conjunctival T cells could provide a mechanism whereby local tissue resident cells such as conjunctival fibroblasts become involved, since collagen deposition and conjunctival tissue remodeling is considerable in chronic allergic eye disease. Comparing conjunctival biopsy specimens from VKC patients with controls, increased expression of RANTES, eotaxin, monocyte chemotactic protein (MCP)-1, and MCP-3 was detected, reflecting the range of inflammatory cells present. VKC conjunctival tissue expression of the chemokine receptor CXCR3 was found to be specifically localized to T cells, and the CXC chemokine Mig was highly expressed, suggesting an important role for this ligand in recruitment of activated T cells. In conclusion, tissue-based studies, combined with in vitro and in vivo models (not discussed in this article), have identified discrete cellular and molecular pathways in each form of ocular allergy, and this knowledge has allowed a more selective therapeutic approach. Nevertheless, currently available therapies for the more chronic forms of ocular allergic disease are limited by their

Further Reading Abelson, M. B. and Granet, D. (2006). Ocular allergy in pediatric practice. Current Allergy and Asthma Reports 6(4): 306–311. Blaiss, M. S. (2008). Evolving paradigm in the management of allergic rhinitis-associated ocular symptoms: Role of intranasal corticosteroids. Current Medical Research and Opinion 24(3): 821–836. Bonini, S., Gramiccioni, C., Bonini, M., and Bresciani, M. (2007). Practical approach to diagnosis and treatment of ocular allergy: A 1-year systematic review. Current Opinion in Allergy and Clinical Immunology 7(5): 446–449. Bonini, S., Sacchetti, M., Mantelli, F., and Lambiase, A. (2007). Clinical grading of vernal keratoconjunctivitis. Current Opinion in Allergy and Clinical Immunology 7(5): 436–441. Calonge, M. and Enrı´quez-de-Salamanca, A. (2005). The role of the conjunctival epithelium in ocular allergy. Current Opinion in Allergy and Clinical Immunology 5(5): 441–445. Calonge, M. and Herreras, J. M. (2007). Clinical grading of atopic keratoconjunctivitis. Current Opinion in Allergy and Clinical Immunology 7(5): 442–445. Dogru, M., Okada, N., Asano-Kato, N., et al. (2005). Atopic ocular surface disease: Implications on tear function and ocular surface mucins. Cornea 24(8 supplement): S18–S23. Fukuda, K., Kumagai, N., Fujitsu, Y., and Nishida, T. (2006). Fibroblasts as local immune modulators in ocular allergic disease. Allergology International 55(2): 121–129. Kumagai, N., Fukuda, K., Fujitsu, Y., Yamamoto, K., and Nishida, T. (2006). Role of structural cells of the cornea and conjunctiva in the pathogenesis of vernal keratoconjunctivitis. Progress in Retinal and Eye Research 25(2): 165–187. Leonardi, A., De Dominicis, C., and Motterle, L. (2007). Immunopathogenesis of ocular allergy: A schematic approach to different clinical entities. Current Opinion in Allergy and Clinical Immunology 7(5): 429–435. Leonardi, A., Motterle, L., and Bortolotti, M. (2008). Allergy and the eye. Clinical and Experimental Immunology 153(supplement 1): 17–21. Mantelli, F. and Argu¨eso, P. (2008). Functions of ocular surface mucins in health and disease. Current Opinion in Allergy and Clinical Immunology 8(5): 477–483. Micera, A., Stampachiacchiere, B., Aronni, S., dos Santos, M. S., and Lambiase, A. (2005). Toll-like receptors and the eye. Current Opinion in Allergy and Clinical Immunology 5(5): 451–458. Schultz, B. L. (2006). Pharmacology of ocular allergy. Current Opinion in Allergy and Clinical Immunology 6(5): 383–389. Stern, M. E., Siemasko, K. F., and Niederkorn, J. Y. (2005). The Th1/Th2 paradigm in ocular allergy. Current Opinion in Allergy and Clinical Immunology 5(5): 446–450.

Pathogenesis of Fungal Keratitis E Pearlman, S Leal, A Tarabishy, Y Sun, L Szczotka-Flynn, Y Imamura, P Mukherjee, and J Chandra, Case Western Reserve University, Cleveland, OH, USA M Momany and S Hastings-Cowden, University of Athens, Athens, GA, USA M Ghannoum, Case Western Reserve University, Cleveland, OH, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Biofilm – The microbial secretion of an extracellular matrix surrounding the organisms. Conidia – These are fungal spores. Matrix metalloproteinases (MMPs) – The proteases that participate in tissue remodeling, wound healing, and inflammation. Multipurpose solution (MPS) – The contact lens care products that are used to disinfect daily-wear contact lenses. Toll-like receptor (TLR) – A family of surface and endosomal receptors that recognize microbial products. TLR signaling induces production of proinflammatory and chemotactic cytokines and antimicrobial peptides.

Contact-Lens-Associated Fungal Keratitis In June 2006, the Centers for Disease Control and Prevention (CDC) Fusarium investigation team (Chang and colleagues) reported 318 cases of Fusarium keratitis, with 164 confirmed cases in 33 states and one US territory, although smaller outbreaks were reported in Singapore, Hong Kong, and France. The age group was between 12 and 83, with a median age of 41; 94% wore soft contact lenses; and keratoplasty was needed for 34%. (Examples of contact-lens-associated Fusarium keratitis are shown by Alfonso and colleagues.) The CDC study demonstrated a clear relation to the use of Bausch and Lomb Renu with MoistureLock multipurpose solution (MPS), and the number of cases of Fusarium keratitis dropped shortly after withdrawal of this product. As unopened bottles were sterile, and Fusarium can be readily isolated from sink and shower drains, the CDC report concluded that the source of infection was in the patients’ homes. However, although the report implies that poor lens care habits were involved, it became clear that Fusarium clinical isolates were more resistant to disinfectants in the lens care solution than CDC strains that were used for comparison. Moreover, resistance was related to the ability of the microorganism’s capacity to form biofilm (see below). Reports from several regions of

426

the USA, including Florida and San Francisco, described cases of contact-lens-associated Fusarium keratitis demonstrating severe corneal opacification and descemetocele formation (hernia of Descemet’s membrane), most of which required keratoplasty. The CDC also reported that the number of cases of Fusarium keratitis dropped after Renu with Moisture Lock was voluntarily withdrawn from the market. However, several cases have been reported subsequently that were not due to this lens care product and were most likely due to increased awareness of Fusarium, although the causes were not always apparent.

Biofilm Formation in Contact-LensAssociated Keratitis Biofilm is defined as microbial secretion of an extracellular matrix surrounding the organisms. Biofilm formation allows the organisms to resist antibiotics (20–1000 times more resistant than planktonic forms), and to host immune responses. The CDC report on the contact-lens-associated outbreak of Fusarium keratitis also suggested that biofilm formation contributes to the resistance phenotype, as bacterial biofilm can form on contact lenses and lens cases. Bacterial biofilms can be generated rapidly on contact lenses and may therefore contribute to the pathogenesis of keratitis and endophthalmitis. Imamura and co-workers showed that Fusarium forms a biofilm on silicone hydrogel contact lenses; furthermore, the architecture, thickness, and composition of the biofilm differ according to the contact lens type. It is likely that the conidia germinate on the contact lens surface, and favorable conditions allow biofilm development. Once the biofilm is formed, the organisms are more resistant to antifungal agents, including those in multipurpose lens care solutions. Consistent with this notion, the Fusarium strain used to test lens care solutions did not form a biofilm and was more sensitive to lens care solutions (Figure 1).

Keratitis Caused by Candida Candida species are the most common pathogenic yeast associated with keratitis. Candida albicans is part of the normal commensal flora; however, these organisms can cause opportunistic corneal infections in immunosuppressed

Pathogenesis of Fungal Keratitis

Etafilcon A

Galyfilcon A

Lotrafilcon A

A

B

C

A

B

C

(a)

(b) Figure 1 Fusarium biofilm formed on different soft contact lenses. (a) Fusarium conidia were incubated with each contact lens for 2 h, after which time the lenses were removed and incubated a further 48 h. Biofilms were formed by FSSC 1-b isolate MRL8609 on soft contact lenses, and their gross morphologies were imaged using a digital camera. All lenses tested supported biofilm formation by strain MRL8609. (b) The Fusarium FSSC 1-b strain MRL8609 was allowed to form mature biofilms on Etafilcon A silicon hydrogel contact lenses and then was stained with ConA and FUN1 dyes to show extracellular matrix (red) and live organisms (green). Stained lens-containing biofilms were analyzed by confocal scanning laser microscopy. Etafilcon A (A), galyfilcon A (B), lotrafilcon A (C), balafilcon A. Arrows indicate extracellular matrix in the biofilms. Similar results were found for C. albicans (not shown). Reprinted from Imamura, Y., Chandra, J., Mukherjee, P. K., et al. (2008). Fusarium and Candida albicans biofilms on soft contact lenses: Model development, influence of lens type, and susceptibility to lens care solutions. Antimicrobial Agents and Chemotherapy 52: 171–182.

individuals or following trauma or surgery. In contrast to filamentous fungi in which trauma is the major predisposing condition, Candida is primarily associated with therapeutic contact lenses, steroid use or immunosuppressive disease, and corneal surgery. In these dimorphic organisms, the yeast stage initially infects the cornea, and then germinates to form pseudohyphae in the corneal stroma. Candida produces a number of proteases and phospholipases (particularly phospholipase B) that facilitate their penetration through the cornea and contribute to tissue destruction. Using C. albicans mutants in a murine model of keratitis, Jackson and colleagues showed that C. albicans virulence depends on expression of genes encoding or regulating hyphal formation, but not genes regulating adherence.

Fungal Keratitis Associated with Trauma Although relatively rare in North America and Europe, filamentous fungi are among the most common causes of microbial keratitis and corneal ulcers in India, China, and Ghana. In the southern USA, Fusarium solani and Fusarium

427

oxysporum are the most common causes of mycotic keratitis, with Aspergillus species being the second most common cause, including A. fumigatus, A. niger, and A. nidulans. Corneal trauma is the main predisposing factor, and the incidence of fungal keratitis increases during harvest season, which is consistent with the majority of cases associated with agricultural work, where it affects more males than females of working age. Trauma can be caused by numerous factors, such as airborne soil and plant material. Filamentous fungi are ubiquitous in the environment, especially on plants, and the fungal spores (conidia) can penetrate the corneal epithelium, germinate in the stroma, and if unchecked by the host response or effective antifungal therapy, fungal hyphae will grow in the corneal stroma, penetrate Descemet’s membrane, and invade the anterior chamber. Role of Matrix Metalloproteinases in Fungal Keratitis Matrix metalloproteinases (MMPs) have an important role in tissue remodeling, wound healing, and inflammation. Rohini and co-workers examined human tears from fungal keratitis patients and corneal sections after keratoplasty and detected elevated collagenases MMP-2 and MMP-8, and the MMP-9 gelatinase, which is consistent with the activation and degranulation of infiltrating neutrophils. In addition to microbial killing, which is primarily mediated by oxygen radicals, neutrophils also prevent microbial dissemination by releasing MMPs and causing local tissue damage. Dong and co-workers as well as Mitchell and co-workers showed that MMP-2 and MMP-9 were also elevated in rabbit and mouse models of Fusarium and Candida keratitis, and Lin and colleagues demonstrated a role for MMP-8 in corneal inflammation by mediating breakdown of collagen and release of chemotactic Pro–Gly–Pro peptides, which then facilitate neutrophil migration through the cornea. Although differences between Aspergillus and Fusarium growth in the stroma have been reported by Xie and colleagues, it is not clear at present how this relates to MMP activity, or if there is a difference in protease production by these organisms. Role of Innate Immunity in Fungal Keratitis Since the 1960s, it has been apparent that the host immune response regulates fungal growth and the outcome of the infection. In rabbit and murine models of Fusarium and Candida keratitis in which either conidia or yeast is applied topically to the abraded epithelium, or is injected intrastromally, a neutrophil-rich cellular infiltrate into the corneal stroma ultimately clears the infection. However, subverting the host response by systemic

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

treatment with cyclophosphamide leads to increased hyphal penetration of the corneal stroma, decreased cellular infiltration, (especially neutrophils), and the unchecked growth of hyphae, causing corneal perforation. These results indicate that the host response plays a critical role in restricting fungal growth in the cornea. To characterize the host response to fungal challenge, Candida-infected corneas were processed for microarray analysis, which demonstrated that the proinflammatory cytokines, interleukin-1 (IL-1) and tumor necrosis factoralpha (TNF-a), were upregulated. To characterize the innate immune response, Tarabishy and colleagues injected Fusarium conidia into the corneal stroma of immunocompetent C57BL/6 and mice on the same genetic background in which genes related to the Toll-Like Receptor (TLR) family of pathogen recognition molecules were knocked out. Figure 2 shows that 6 h after intrastromal injection of 10 000 conidia, hyphae were detected in the corneal stroma of C57BL/6 mice and in mice in which the gene for the MyD88 adaptor molecule common to most TLRs and IL1R1 is mutated. Although a cellular infiltrate was detected in the peripheral cornea of C57BL/6 mice, this infiltrate was absent in MyD88–/– mice. Figure 3 shows that whereas C57BL/6 mice rapidly develop corneal opacification associated with a pronounced infiltrate and clear the organisms, MyD88–/– mice had delayed cellular infiltration, and even though neutrophils were recruited to the corneal stroma,

Control DEB

Murine Model of Aspergillus Keratitis Aspergillus species, the second most common cause of fungal keratitis after Fusarium, are also ubiquitous in the environment, and most people inhale hundreds of conidia daily. Although pulmonary aspergillosis occurs primarily in immunosuppressed individuals, this is not the case in keratitis, where the risk factors are similar to those of Fusarium, that is, the highest incidence is associated with agriculture and trauma. In addition, Aspergillus conidia are smaller than Fusarium conidia, can therefore penetrate deeper into the lungs, and likely also penetrate deeper into the corneal stroma. We generated a strain of Aspergillus fumigatus expressing a red fluorescent protein, and injected conidia into the corneal stroma of C57BL/6 mice. Figure 6 shows that after 24 h, the cornea is opaque. However, Figure 6(b) also shows that the presence of corneal opacities coincides with the presence of Aspergillus. Figure 6(c) shows higher magnification of hyphae in the corneal stroma. Ongoing studies are examining the role of the host response and Aspergillus virulence factors in the pathogenesis of this disease.

0.20

Central

Peripheral

C57BI/6

Metabolic activity (XTT OD 492 nm, mean ± SD)

0.25

they were unable to clear the organisms (Figures 4 and 5). Fusarium hyphae are detected throughout the cornea (Figure 5), which perforated within 4 days. Subsequent experiments showed that IL-1R1 is required for neutrophil recruitment to the cornea, whereas TLR4 is important for fungal killing.

0.15 0.10 0.05

0.00 MRL8609

MRL8996 ATCC36031 Fusarium isolates

SC5314

MyD88–/–

428

C. albicans

Fungal isolate Figure 2 Fusarium strain differences in biofilm formation on lotrafilcon A lenses. Biofilms were formed in the absence or presence of DEB (which inhibits biofilm formation) on a lotrafilcon A lens using the FSSC 2-c ATCC 36031 reference isolate or clinical isolate FSSC 1-b (MRL8609), FOSC 3-a (MRL8996). Biofilms were quantified using the XTT metabolic activity assay. Data represent means (+/– SDs) calculated from three separate experiments. Reprinted from Imamura, Y., Chandra, J., Mukherjee, P. K., et al. (2008). Fusarium and Candida albicans biofilms on soft contact lenses: Model development, influence of lens type, and susceptibility to lens care solutions. Antimicrobial Agents and Chemotherapy 52: 171–182.

Figure 3 In vivo confocal microscopy of Fusarium keratitis in C57BL/6 and MyD88–/– corneas. C57BL/6 and MyD88–/– mice were injected intrastromally with 1 104 conidia from a clinical isolate of F. oxysporum. After 6 h, mice were examined by in vivo confocal microscopy (Confoscan). Representative images from the central and peripheral corneal stroma are shown. Note the presence of hyphae in the central corneal stroma of C57BL/6 and MyD88–/– mice (a, c); however, a cellular infiltrate is present in the peripheral cornea of C57BL/6, but not MyD88–/– mice (b, d). Reprinted from Tarabishy, A. B., Aldabagh, B., Sun, Y., et al. (2008). MyD88 regulation of Fusarium keratitis is dependent on TLR4 and IL-1R1 but not TLR2. Journal of Immunology 181: 593–600.

Pathogenesis of Fungal Keratitis

d1

d2

C57BL/6

6h

429

MyD88–/–

CFU/eye

10000

1000

(a) 100 Clinical score

4

p = 0.0006

3

B6 (c)

MyD88

B6

6h

MyD88 24 h

B6

MyD88 48 h

2 1 0

(b)

MyD88–/–

C57BL/6

Figure 4 Fungal keratitis in C57BL/6 and MyD88–/– mice. Mice were injected intrastromally with 1 104 conidia from a clinical isolate of F. oxysporum as described above. Corneal opacification, CFU, and histology were examined by standard methods. (a, b) Corneal opacification in MyD88–/– mice was impaired at 24 h, but increased until 72 h after which time MyD88–/– corneas perforated, whereas C57BL/6 corneas eventually resolved. (c) CFU decreased in C57BL/6 mice over time, whereas Fusarium replicated in the corneas of MyD88–/– mice. Reprinted from Tarabishy, A. B., Aldabagh, B., Sun, Y., et al. (2008). MyD88 regulation of Fusarium keratitis is dependent on TLR4 and IL-1R1 but not TLR2. Journal of Immunology 181: 593–600.

Epi Str a/c (a)

(c)

(b)

(d)

Figure 5 Impaired cellular infiltration in MyD88–/– mice (a, b) Histological analysis in MyD88–/– corneas after PAS staining shows Fusarium hyphae penetrating Descemet’s membrane after 24 h, and growing in the stroma and anterior chamber after 48 h despite the presence of a cellular infiltrate. In contrast, there was an early and pronounced cellular infiltrate in C57BL/6 mice (c, d) comprised mostly of neutrophils. Reprinted from Tarabishy, A. B., Aldabagh, B., Sun, Y., et al. (2008). MyD88 regulation of Fusarium keratitis is dependent on TLR4 and IL-1R1 but not TLR2. Journal of Immunology 181: 593–600.

(a)

(b)

(c)

Figure 6 Aspergillus fumigatus in murine cornea. Aspergillus fumigatus was transfected with a plasmid expressing m-cherry. Conidia were injected into the corneal stroma of C57BL/6 mice, and after 24 h, corneas were examined by light (a) and fluorescence (b) microscopy. (c) Whole mount cornea examined by confocal microscopy.

require a combination of human and animal studies to determine the course of events leading to fungal killing and resolution of infection. Human disease correlates are particularly difficult to study; however, Bochud and co-workers showed that polymorphisms in TLR4 are associated with susceptibility to systemic aspergillosis; therefore, it is possible that TLR4 also mediates susceptibility to Aspergillus keratitis.

Acknowledgments Conclusions The pathogenesis of fungal keratitis depends on the balance between the host response and expression of fungal virulence factors. Although some mediators of innate immunity and fungal virulence factors have been identified, it will

Studies presented in this article were supported by NIH grant EY18362 (EP) and EY11373 (EP), by DE017486-01A1 (MAG) and R01DE 13932 (MAG), and by the Research to Prevent Blindness Foundation and the Ohio Lions Eye Research Foundation.

430

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

See also: Contact Lenses; Corneal Epithelium: Response to Infection; Innate Immune System and the Eye.

Further Reading Alfonso, E. C., Cantu-Dibildox, J., Munir, W. M., et al. (2006). Insurgence of Fusarium keratitis associated with contact lens wear. Archives of Ophthalmology 124: 941–947. Bharathi, M. J., Ramakrishnan, R., Meenakshi, R., et al. (2007). Microbial keratitis in South India: Influence of risk factors, climate, and geographical variation. Ophthalmic Epidemiology 14: 61–69. Bochud, P. Y., Chien, J. W., Marr, K. A., et al. (2008). Toll-like receptor 4 polymorphisms and aspergillosis in stem-cell transplantation. New England Journal of Medicine 359: 1766–1777. Chang, D. C., Grant, G. B., O’Donnell, K., et al. (2006). Multistate outbreak of Fusarium keratitis associated with use of a contact lens solution. Journal of the American Medical Association 296: 953–963. Dong, X., Shi, W., Zeng, Q., and Xie, L. (2005). Roles of adherence and matrix metalloproteinases in growth patterns of fungal pathogens in cornea. Current Eye Research 30: 613–620. Grant, G. B., Fridkin, S., Chang, D. C., and Park, B. J. (2007). Postrecall surveillance following a multistate Fusarium keratitis outbreak, 2004 through 2006. Journal of the American Medical Association 298: 2867–2868.

Imamura, Y., Chandra, J., Mukherjee, P. K., et al. (2008). Fusarium and Candida albicans biofilms on soft contact lenses: Model development, influence of lens type, and susceptibility to lens care solutions. Antimicrobial Agents and Chemotherapy 52: 171–182. Jackson, B. E., Wilhelmus, K. R., and Mitchell, B. M. (2007). Genetically regulated filamentation contributes to Candida albicans virulence during corneal infection. Microbial Pathogenesis 42: 88–93. Mitchell, B. M., Wu, T. G., Chong, E. M., Pate, J. C., and Wilhelmus, K. R. (2007). Expression of matrix metalloproteinases 2 and 9 in experimental corneal injury and fungal keratitis. Cornea 26: 589–593. Pearlman, E., Johnson, A., Adhikary, G., et al. (2008). Toll-like receptors at the ocular surface. Ocular Surface 6: 108–116. Rohini, G., Murugeswari, P., Prajna, N. V., Lalitha, P., and Muthukkaruppan, V. (2007). Matrix metalloproteinases (MMP-8, MMP-9) and the tissue inhibitors of metalloproteinases (TIMP-1, TIMP-2) in patients with fungal keratitis. Cornea 26: 207–211. Tarabishy, A. B., Aldabagh, B., Sun, Y., et al. (2008). MyD88 regulation of Fusarium keratitis is dependent on TLR4 and IL-1R1 but not TLR2. Journal of Immunology 181: 593–600. Wu, T. G., Keasler, V. V., Mitchell, B. M., and Wilhelmus, K. R. (2004). Immunosuppression affects the severity of experimental Fusarium solani keratitis. Journal of Infectious Diseases 190: 192–198. Xie, L., Zhai, H., Shi, W., et al. (2008). Hyphal growth patterns and recurrence of fungal keratitis after lamellar keratoplasty. Ophthalmology 115: 983–987. Yuan, X., Mitchell, B. M., and Wilhelmus, K. R. (2008). Gene profiling and signaling pathways of Candida albicans keratitis. Molecular Vision 14: 1792–1798.

Conjunctiva Immune Surveillance E Knop, Charite´ – Universita¨tsmedizin Berlin, Berlin, Germany N Knop, Hannover Medical School, Hannover, Germany ã 2010 Elsevier Ltd. All rights reserved.

Glossary Conjunctiva-associated lymphoid tissue (CALT) – It is the physiological protective mucosal immune tissue of the conjunctiva. It consists of lymphoid cells and accessory cells inside the mucosal tissue and can be divided into the epithelial and underlying connective tissue (lamina propria) compartments. It is arranged as a diffuse lymphoid effector tissue along the whole extension of the conjunctiva and has interspersed organized lymphoid follicles for afferent antigen uptake and effector cell generation. Dendric cells (DCs) – They are a special class of professional antigen-presenting cells (APC, together with macrophages and B-cells). They take up external antigens, degrade them into small fragments (epitopes), present them on MHC-class-II to T-helper cells, and hence, induce immune reactions. Depending on their maturation status which is influenced by the presence of inflammatory signals, they modulate between the inductions of tolerance versus inflammation. They also link the unspecific innate to the induced specific immune system and are hence key modulators of the immune reaction. Eye-associated lymphoid tissue (EALT) – This tissue summarizes all the lymphoid tissues of the extended mucosal ocular surface, that is, of ocular surface proper (conjunctiva and cornea) along with its mucosal adnexa (the lacrimal-drainageassociated lymphoid tissue, LDALT, and the lymphoid cells inside the lacrimal gland). EALT is in line with the mucosal immune system in other parts of the body (e.g., gut-associated lymphoid tissue (GALT) in the gut and bronchus-associated lymphoid tissue (BALT) in the airways). High endothelial venules (HEVs) – Specialized postcapillary venules that have an endothelium of bright roundish cells compared to the ordinary flat dense ones. They are located in lymphoid tissues and have tissue-specific adhesion molecules (vascular addressins) on the cell surface that specifically interact with homing receptors on circulating lymphocytes in order to maintain a regulated immigration of lymphocytes into the tissue. Human leukocyte antigen (HLA) – A system of the major histocompatibility complex () MHC)

in humans. It contains of a number of genes and their respective encoded proteins (that can act as antigens). The term HLA is frequently used to describe immunological self and nonself in the context of transplant rejection. Intercellular adhesion molecule 1 (ICAM-1) – An adhesion molecule (CD54 according to the immunological cluster of differentiation, CD, nomenclature) mainly on vascular endothelial cells which is upregulated in inflammation and promotes the increased immigration of leukocytes, that carry corresponding integrin receptors, into the tissue. Membraneous cells (M-cells) – Also called microfolded cells, they are a special type of cells in the modified epithelium overlying organized lymphoid follicles, the so-called follicle-associated epithelium (FAE). Their name refers to the fact that they have a different, usually smooth, surface ultrastructure compared to the ordinary epithelial cells. They form cellular pockets populated by groups of leukocytes which are separated from the lumen by a thin luminal cytoplasmic sheet. M-cells actively transcytose luminal antigens for uptake by the leukocytes and their subsequent presentation to and activation of T- and B-cells in order to generate antigen-specific effector cells. Major histocompatibility complex (MHC) – It is differentiated mainly into class-I and class-II. Their encoded proteins on the surface of cells perform the presentation of protein antigen fragments (epitopes) to immune cells. MHC-class-I is found on all nucleated cells and presents antigens produced inside the cell (either own or viral proteins after infection) to cytotoxic CD8 lymphocytes and natural killer cells. MHC-class-II, in contrast, occurs physiologically only on specialized antigenpresenting cells and presents foreign antigens to the CD4 receptor of T-helper cells. In inflammation, it can be upregulated by other cells. Lipopolysaccharide (LPS) – A component of the outer cell membrane of the wall of Gram-negative bacteria that acts as an endotoxine. The presence of LPS, that is detected for example, by toll-like receptors, signals the pathogenic nature of antigens to the immune system and elicits a strong inflammatory reaction.

431

432

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

Tolerance – Immune tolerance is a status in which the immune system is in a state of nonreactivity to an antigen in order to prevent inflammatory tissue-destructive reactions. Tolerance is actively generated and directed not only against the bodies’ own cellular self-antigens but also against nonpathogenic external antigens. If tolerance fails, autoimmunological disease or allergy may occur. Tolerance is the default mode of the mucosal immune system, including CALT and EALT, in order to preserve tissue integrity.

lymphocytes (IELs) that reside mainly in the basal layers (Figure 2(b) and Figure 3(a)) as well as dendritic cells (DCs), that have long narrow extensions, for uptake of antigens from the surface. The conjunctival epithelial surface is covered by small cytoplasmic protrusions (microvilli and microplicae) with a well-developed surface coat of filamentous projections (glycocalyx) that form a meshwork (Figure 1(c) and Figure 2(a)). Epithelial Immune Surveillance Takes Care of Environmental Antigens Physical and physicochemical barriers keep antigens outside

Conjunctival Morphology and Function Are Closely Interacting for Immune Surveillance Epithelial Defense Mechanisms Epithelial morphology and function

The conjunctiva is a moist mucous organ that consists of a surface epithelium and an underlying loose connective tissue (lamina propria), separated by the epithelial basement membrane. The epithelium of the human conjunctiva has, in contrast to small rodents (e.g., rat and mouse), a stratified nonsquamous morphology and consists of two to three cell layers of cubical cells in most parts. It becomes multilayered and assumes prismatic morphology at the fornix whereas it tends to become squamous toward the limbus (Figure 1(a) and 1(b)). Interspersed mucus-secreting goblet cells occur inside the epithelium as single cells or in small groups as well as intraepithelial

The structure of the conjunctival epithelium already contributes to basic protective mechanisms which can be considered as part of the innate defense. Epithelial cells are mechanically connected by desmosomes and have an apical belt of intercellular junctions including tight junctions that seal the intercellular space and limit the passive para-cellular leakage of antigens in and out of the tissue (Figure 2(a)). This physical cellular barrier is supplemented by the physicochemical barrier of the epithelial mucins, that consist of cell membrane-spanning mucins (glycocalyx) produced by the ordinary epithelial cells and of soluble mucins secreted by the goblet cells which mix with the aqueous phase. Together they form a layer in the range of few micrometers thickness, that is, a sticky gel to which microbes adhere and can hence be cleared by the constant renewal of the preocular tear film. Soluble protective factors, including secretory immunoglobin A (SIgA), are fixed to the mucin layer in order to make it an almost impenetrable and lethal barrier to antigens and in particular to microbes.

M

MV MP BZ

(a)

(b)

(c)

1 μm

Figure 1 Structure of the human conjunctival epithelium. (a) The epithelium of the human conjunctiva is stratified cuboidal in most regions and assumes more layers with prismatic surface cells toward the fornix. Interspersed goblet cells (BZ) release mucus (M) tufts onto the surface. (b) Goblet cells contain densely packed mucin granules and a flat or triangular nucleus. They may be slightly inclined if located in the relatively flat bulbar epithelium close to the limbus. The surface of the conjunctival epithelial cells shows numerous microprotrusions that result in a rough surface in low-magnification transmission electron microscopy. (c) In higher enlargement, microvilli (MV) and microplicae (MP) are seen which have a dense glycocalyx of fine molecular antennae (arrows) that project into the lumen and form a meshwork, as better seen in cross section (inset, 2). Reproduced from Knop, E. and Brewitt, H. (1992). Morphology of the Conjunctival Epithelium in Spectacle and Contact Lens Wearers – A Light and Electron Microscopic Study. Contactologia, Stuttgart: Enke Verlag.

Conjunctiva Immune Surveillance

FAE SlgA

AMP

433

Microbe

m-cell

Microbe SlgA

gc 8

gc 8

TLR

dc

8

tj

4

4

4

ENK pc

4

Lymph vessel

HEV

HEV

m∅ n Lymph vessel

(a)

bm

(b)

fi

mc

Diffuse lymphoid effector tissue

Blood circulation

(c)

Follicular organized inductive lymphoid tissue

Figure 2 Defense systems of the human conjunctiva. (a) The conjunctival epithelium has an array of defense systems consisting of the integrity of the surface epithelial cells (provided with pattern recognition receptors, TLR) that are sealed by apical tight junctions (tj), of the attached mucin layer that is enforced by adhering antimicrobial proteins and peptides (AMPs) including specific secretory IgA (SIgA) and of the overlying tear film (shaded blue) that contains similar protective molecules and provides a washing effect. (b) A diffuse effector tissue is formed by lymphoid cells of the specific adaptive immune system and by innate cells such as macrophages (mø), mast cells (mc), neutrophilic granulocytes (n), and dendritic cells (dc). They functionally interact with stromal fibrocytes (fi). Lymphoid cells consist of CD4þ and CD8þ T-cells (black circles) that constitute intraepithelial and lamina propria lymphocytes. Differentiated B-cells (plasma cells (pc), large blue) produce dimeric IgA, which is transported through the epithelium as SIgA. (c) Interspersed solitary lymphoid follicles consist of B-cells (small blue circles), frequently have a bright germinal center due to cell proliferation, have an apical follicle-associated epithelium (FAE) with M-cells for antigen transport but without goblet cells (gc) and have para-follicular T-cell (small black circles) zones with lymph vessels (yellow) and high endothelial venules (HEVs); small arrows indicate the direction of cell migration. The mechanisms for conjunctival immune surveillance are explained topographically in this figure from the epithelium (a) over the diffusely interspersed effector cells (b) toward the organized lymphoid follicles that generate the effector cells (c). Functionally, it is reverse because the effector cells generated in lymphoid follicles after antigen uptake and presentation recirculate via the blood circulation (symbol of heart and blood flow between (c) and (b)) to and migrate into the diffuse effector sites to exert their protective function by cell contact or by soluble mediators. The drawing is not to scale.

The mechanical washing effect of the tear film wipes away antigens and detritus

The tear film is an important functional component of the ocular surface mucosal protection system. Apart from providing the necessary moisture, the constant flow of tears over the ocular surface and in particular, over the cornea, together with the wiping effect of the lid margin with every blink, provides a constant mechanical washing. This discharges antigens and removes dust and cell detritus. Other parts of the ocular surface along the retropalpebral tear film are not so rapidly cleared so that antigens can stay in longer contact with the epithelium. Therefore, the tear film contains a large number of antimicrobial factors that contribute more specifically to the innate immune defense. Epithelial innate immune defense factors

The innate immune system uses pattern-related receptors (PRRs) that mainly detect conserved pathogen-associated molecular patterns (PAMPs) but also host antigens from destroyed cells. It reacts via effectors, which consist of soluble antimicrobial proteins and peptides (AMPs)

which bind to the microbial cell wall in order to destroy it or which interfere with the microbial metabolism. The innate immune system also employs production of soluble mediators, such as inflammatory cytokines and chemotactic cytokines (chemokines) that functionally couple the innate and adaptive immune answer. PRRs on epithelial cells provide an external alarm system

As soon as microbial antigens have breached the physicochemical barrier, they get in touch with epithelial PRRs (Figure 2(a), the most prominent of which is presently the diverse family of toll-like receptors (TLRs). Binding of their ligands causes activation of the host cell via a MyD88-dependent signaling pathway that activates a nuclear transcription factor, nuclear factor kappa B (NFkB), and results in production of inflammatory cytokines such as interleukin 6 (IL-6), interferon gamma (IFN-g), or tumor necrosis factor alpha (TNF-a). Subsequently, these induce the production of chemokines, adhesion molecules, and inducible AMP. Altogether this represents an inflammatory cascade with activation, first

434

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

I

I

P

I

P E

E

(a)

(b)

HEV

I I I

HEV E

PE

(c)

(d)

(e)

N

(f)

RER

(h)

(g)

C

M

HE

LU E LP

IgA

10 µm

LU E

SC

(i)

(j)

LP

(k)

10 µm

Figure 3 Characteristics of diffuse CALT. (a) Tarso-orbital conjunctiva. Plasma cells (P) and lymphocytes (l) form a diffuse lymphoid cell layer in the lamina propria covered by an epithelium with intraepithelial lymphocytes (arrowhead). A high endothelial venule (HEV) underneath has typical roundish endothelial cells (E) and contains lymphocytes within and around the wall (arrows). (b) TEM shows an intraepithelial lymphocyte (l) between epithelial cells (E) on the basement membrane (arrowheads). (c) Immunostaining indicates T-cells inside the epithelium (arrowhead), in the lymphoid layer (l) and around or in the wall (arrows) of a HEV (asterisk). (d) Ultrastructurally, a HEV shows large bright endothelial cells (E), a contractile pericyte layer (PE), and adjacent (l) or intramural (arrow) lymphocytes. (e) The vast majority of plasma cells in the lymphoid layer are IgA positive as also deposits in the epithelium (arrowheads), while IgM (f) is rare. (g) The epithelium is positive for the transporter SC. (h) A plasma cell lying in the loose collageneous (C) tissue has extended rough endoplasmic reticulum (RER), mitochondria (M), large nucleolus (N), and radial heterochromatin; (b, d, h: bar ¼ 1 mm; a, c, e–g: bar ¼ 10 mm). (i (1–3)) Lacrimal gland, LG, with lymphocytes (arrow), and plasma cells (arrowead) between the roundish acini. (2) IgA is found strongly in plasma cells and as weaker patchy staining in acinar epithelial cells, which more strongly express SC (3). (j (1–3)) Excretory lacrimal ducts that connect the LG to the conjunctiva have similar characteristics but in the epithelium IgA (2) and SC (3) are mainly expressed in the luminal layer; the duct has two cell layers but appears wider to the left due to oblique plane of section. (k (1,2)) Multiple-fluorescent staining for IgA (green), SC (red), and cell nuclei (blue) shows that the components of the secretory immune system are similarly arranged in the LG (1) and the conjunctiva (2, here orbital zone); bm level indicated by fine lines. IgA-positive plasma cells are diffusely interspersed in the LP of both tissues; in the LG frequently in groups. Epithelium (E) shows strong staining for SC;

Conjunctiva Immune Surveillance

of the epithelial cells and later also of lamina propria leukocytes and vascular endothelial cells. It induces leukocyte recruitment into the epithelium and their immigration from the blood stream into the tissue. The normal conjunctival epithelium expresses a number of different TLRs, similar to the cornea. TLR1, 2, 3, 5, and 6 were found in all conjunctival and limbal epithelial cell samples, TLR4 and 9 only inconstantly, but not TLR7, 8, and 10. TLR2 may only occur upon stimulation by IFN-g and bacterial cell wall extract, for example, in patients with ocular allergy. This results in upregulation of inflammatory markers, such as the intercellular adhesion molecule 1 (ICAM-1), human leukocyte antigen (HLA), TNF-a, and IL-8, in a dose-dependent manner. Bacterial-specific TLRs are of interest in ocular allergy because colonization by bacteria is common there. The activation of TLRs represents an important co-factor in ocular allergy and their blockade can significantly inhibit release of inflammatory mediators which may turn out as a promising new therapy option for ocular allergy. The conjunctival epithelium secretes diverse AMPs

The spectrum of epithelial derived AMPs is distinct for cornea and conjunctiva but overlapping. Conjunctival epithelium produces not only the human b-defensins (hBD)-1,2,3 and further AMPs such as liver-expressed antimicrobial peptides (LEAPs) 1 and 2 and cathelicidin (LL-37) but also macrophage inflammatory protein 3alpha (MIP-3a) and thymosin beta 4 (Tb-4). Some of the AMPs are constitutively produced, whereas others are inducible. hBD-2 is induced by inflammatory cytokines in ocular surface inflammation and by presence of bacterial LPS, while hBD3 is induced by infection and LL-37 by epithelial wounding. Conjunctival AMPs such as LL-37 are active against bacterial (Pseudomonas aeruginosa, Staphylococcus aureus, Staphylococcus epidermidis) and viral (Herpes simplex virus 1, adenovirus) infection. They can act as multifunctional factors in wound healing and signaling pathways. Interestingly, the antimicrobial activity of some of these AMPs (e.g., hBD-1, hBD-2, and Tbeta-4) is almost completely inhibited in the presence of tear fluid. This may indicate that not all epithelial AMPs are produced in order to act as tear film factors but rather play a major role for local protection inside the conjunctival epithelium itself. Apart from AMPs, there is a plethora of other protective proteins. AMPs continue downstream in the lacrimal drainage system.

435

Conventional antibacterial factors are surprisingly versatile defense tools

Even the ‘old fashioned’ established antimicrobial proteins in the tear film, such as lysozyme and lactoferrin, have surprising newly detected functions. Apart from being bactericidal, either by lysis of components in the Gram-positive bacterial cell wall (lysozyme) or by interfering with their iron metabolism (lactoferrin), they are also antifungal and antiviral. Through the absorption of the strongly inflammatory bacterial endotoxin lipopolysaccharide (LPS), which is a surface molecule on Gram-negative bacteria, they also act anti-inflammatory. They have further anti-inflammatory functions by their influence on antigenpresenting cells (APCs) and hence appear as key elements in host defense that link innate and adaptive immunity. Conjunctival Lamina Propria: Morphology and Function of the Diffuse Mucosal Immune System Diffusely arranged lymphoid and innate cells contribute to conjunctival immune surveillance

The lamina propria contains bone-marrow-derived cells and vessels of different types. Apart from capillaries and lymph vessels, specialized high endothelial venules (HEVs) occur. Vessels serve for the supply with nutrition and discharge of metabolites, for hormonal regulation of the tissue, and also for the migration of immune cells. Lymphocytes, together with accessory leukocyte populations (macrophages, granulocytes, mast cells, and DCs), form a diffuse lymphoid tissue (Figure 3(a)) which is regarded mainly as an effector site of CALT although antigen uptake via DCs can also occur here to a certain extent. The diffuse lymphoid tissue is located in the vast majority of the surface, except for the solitary lymphoid follicles. The thickness of this cell layer depends on the location along the conjunctiva, shows a certain topographical variation, and is frequently only one to two cells wide, which may be a reason why these cells have often been overlooked in the past. It also shows a certain interindividual variation that may depend on the immune status of the person. IEL also functionally belong to the diffuse effector cells (Figure 3(b)). Different subtypes of lymphocytes occur in the conjunctiva

Diffuse conjunctival lymphocytes are mainly CD3þ T-cells (Figure 3(c)) (whereas CD20þ B-cells are largely restricted to the solitary lymphoid follicles). They are activated (CD45Roþ, CD25þ) and express the human

goblet cells (asterisks) are negative for SC. Mixed color indicating both proteins (¼SIgA) is seen in the tubuloacinar lumina (LU) of the LG and frequently delineates the luminal cell surface. (a–h) Adapted from Knop, N. and Knop, E. (2000). Conjunctiva-associated lymphoid tissue in the human eye. Investigative Ophthalmology and Visual Science 41: 1270–1279. (i–k) Knop, E., Knop, N., and Claus, P. (2008). Local production of secretory IgA in the eye-associated lymphoid tissue (EALT) of the normal human ocular surface. Investigative Ophthalmology and Visual Science 49: 2322–2329.

436

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

mucosa lymphocyte antigen (HML-1, integrin aEb7) which substantiates the integration of CALT into the mucosal immune system. In the epithelium CD8þ, cytotoxic/suppressor cells prevail that have been proposed to act mainly in the suppressor mode and may hence provide a component of the immune tolerance at the ocular surface. Lamina propria lymphocytes, in contrast, consist of equal or prevailing amounts of CD4þ T-helper (Th) cells compared to CD8þ T-cells. All known types of T-cells exist in CALT and their immune regulation is considered below (see the section titled ‘Mechanisms of conjunctival immune regulation’). Plasma cells (see the section titled ‘The conjunctiva contributes actively to the secretory immune system’) account for about 20% of the conjunctival leukocytes in histology but their absolute number is in the range (2/3) of those in the lacrimal gland that was long regarded as the sole source of tear film SIgA. Conjunctival lymphocytes are even more abundant. This supports the concept that the conjunctiva considerably contributes to its own specific defense and it very well supports the recognition of a diffuse CALT along the whole extension of the human conjunctiva in tissue whole mounts. The conjunctiva contributes actively to the secretory immune system

Specific soluble antigen-receptors (immunoglobulins, Ig) produced by local mucosal differentiated B-lymphocytes (plasma cells) and transported through the overlying epithelium constitute the secretory immune system. This is a major mucosal defense mechanism and is also present in the conjunctiva. Mucosal Ig mainly consist of polymeric (p)IgA that also forms the predominant Ig in the tear film, besides small amounts of the other polymeric Ig (IgM) and trace amounts of IgG. During eye closure, overnight IgA is the predominant protein in the closed eye tear film, but only recently the components of the conjunctival secretory immune system could be consistently verified by immunohistochemistry and polymerase-chain reaction. Local conjunctival plasma cells produce mainly the antiinflammatory IgA

The vast majority of conjunctival plasma cells produce IgA and hence stain positive for it in immunohistochemistry (Figure 3(e)). IgM, which performs the initial acute secretory immune answer, is rarely observed (Figure 3(f)) and hence indicates that the physiological conjunctival diffuse lymphoid cells do not reflect any kind of reaction to an acute insult. The epithelial transporter molecule for both of them (pIgR, represented by its extracellular domain secretory component, SC) is strongly expressed throughout the human conjunctival epithelium (Figure 3(g)). After transport, SC remains linked to pIgA which together constitute secretory IgA (SIgA). Conjunctival plasma cells show a typical ultrastructure in transmission electron microscopy (Figure 3(h)).

IgA-positive plasma cells in the lamina propria and SC in the overlying epithelium are continuously expressed from the lacrimal gland (Figure 3(i) 1–3) along its excretory ducts (Figure 3(j) 1–3) into the conjunctiva and further within the lacrimal drainage system. In multifluorescent immune staining, the secretory immune system of the lacrimal gland and the conjunctiva show the same characteristics (Figure 3(k) 1,2). SIgA performs diverse protective and antiinflammatory functions at the ocular surface

SIgA is deposited onto the epithelial surface and into the tear film (Figure 2(a)). It contributes to the binding of specific antigens and to their immobilization and discharge. It binds to the surface of microbes and viruses and thereby limits their binding to and entrance into the tissue. It binds and thereby neutralizes bacterial toxins such as LPS. SIgA antibodies occur naturally to the physiological commensal ocular flora and are induced by the presence of pathological microbes, such as Acanthamoeba and Pseudomonas. IgA does not only exert immune functions at the luminal ocular surface but also locally inside the tissue. IgA has a low complement-binding activity and hence acts in an antiinflammatory fashion. Bound antigens are opsonized to phagocytes which facilitates microbe uptake and destruction. IgA can bind to pathogens that have already penetrated into the tissue including intracellular viral particles. During the vectorial transport of pIgA toward the lumen, the bound pathogens are cleared from the tissue. IgA-bound antigens have an antiinflammatory effect on signaling networks and immune regulation inside the tissue by induction of the tolerogenic cytokines TGF-b and IL-10 and by limiting the activation of DC. Lamina Propria Leukocytes Provide Immediate Innate Response against Invading Pathogens and can Orchestrate an Inflammatory Reaction Apart from the lymphocytes, various other types of bonemarrow-derived leukocytes exist in the diffuse conjunctival effector tissue that are all not purely pathogenic, but exert important protective immune functions. Macrophages are reportedly frequent in immunohistochemistry but less obvious in conventional histology (Figure 3(a)). Phagocytes, such as macrophages, granulocytes, and DCs, can engulf and devour the invader but in addition, they act in different ways. Macrophages mainly destroy pathogens by internal digestion but are also capable to present epitopes of the pathogen on the MHC-class-II molecule (MHC-II) to CD4þ T helper-cells for their subsequent activation. DCs digest pathogens mainly for this purpose. Granulocytes, also known as polymorphonuclear (PMN) cells, are usually the first cells that arrive at the site of acute defense reactions.

Conjunctiva Immune Surveillance

Protective functions of conjunctival lamina propria leukocytes

Neutrophile granulocytes are regularly observed in the normal human conjunctiva. They account for about 5% of the leukocytes and occur massively in the closed-eye tear film. In addition to their phagocytic capacity, they secrete several soluble factors such as antimicrobial proteins (e.g., lactoferrin, alpha-defensins, and cathelicidin) for immediate destruction of microbes. Furthermore, they secrete proteases (e.g., cathepsins, gelatinase, and neutrophile elastase) that lead to digestion of the extracellular matrix in order to provide space for accommodating the plethora of cells necessary to mount an effective inflammatory response. Eosinophile granulocytes are reportedly not observed in a normal conjunctiva but immigrate in parasitic infection. They produce chemokines and cytokines (e.g., RANTES [CCL5], TGF-b, TNF-a) for the activation and recruitment of other leukocytes including T-cells, as found in ocular allergy. Mast cells, apart from their potential physiological function, are mainly known for their inflammatory activity during IgE-mediated allergic disease where they release a variety of vasoactive mediators (e.g., histamine and heparin) and Th1 and Th2 cytokines (IL-4, IL-5, IL-6, and TNF-a) that can orchestrate an inflammatory response. HEVs Provide the Regulated Immigration of Bone-Marrow-Derived Cells into the Tissue The bone-marrow-derived cells that populate the conjunctiva all arrive here via the blood stream. Most of them stay here but lymphocytes, after being primed, and DCs, after antigen uptake, can also leave the tissue again via lymphatics. Although lymphocytes can emigrate through ordinary capillaries and venules, they do so with higher efficiency through conjunctival HEV via their tissuespecific adhesion molecules (lymphocyte homing molecules) that interact with vascular endothelial addressins. HEVs occur particularly in the para-follicular T-cell areas of lymphoid follicles (Figure 2(c)) but they are also found in the diffuse lymphoid effector tissue of the conjunctiva. Emigrated T-cells are frequently observed around HEV (Figures 3(c) and 3(d)). Conjunctival HEVs are a normal component of the lymphoid tissue and they have a characteristic ultrastructure similar to that in other mucosal organs.

Conjunctival Lymphoid Follicles Have a Typical Morphology and Function Solitary organized lymphoid follicles are interspersed into the diffuse effector tissue along the conjunctiva. They are relatively flat due to the limited space in the narrow conjunctival lamina propria but still show typical follicular characteristics. They consist of accumulations of

437

B-cells (Figure 4(a)). The overlying epithelium changes its morphology toward the apex into a follicle-associated epithelium (FAE) by losing the goblet cells, by assuming a flat cell shape (Figures 4(a)–4(c)), and by a rarefied expression of SC (Figure 4(c)). Groups of lymphocytes, including CD20þ B-cells (Figure 4(a)) and CD3þ T-cells, occur inside the epithelium and are separated from the lumen just by a very narrow epithelial lining. Altogether, this is a conspicuous sign for the presence of specialized M-cells that form intraepithelial pockets populated by lymphoid cells. M-cells serve for the uptake and transport of luminal antigens toward the lymphoid cells in the pocket that can detect the antigen and present it to naive lymphocytes. The typical morphology of conjunctival M-cells and their antigen transport have been verified in a number of animal species, including guinea pig, turkey, chicken, rabbit, dog, and monkey. It has been shown that CALT is able, for example, to induce a tolerance against retinal antigens upon their topical conjunctival instillation. The number of lymphoid follicles in elderly humans is relatively low (about 10 follicles per eye) with an average diameter of about 0.25 mm. Their small size again offers an explanation for the fact that CALT has frequently been overlooked in the past. They are more frequent in the upper conjunctiva than in the lower one (Figure 5(a)) and show a bilateral symmetry. In younger individuals, however, lymphoid follicles are more frequent and before puberty they occur in every person. Therefore, CALT follicles show a similar involution with age as observed for other locations of the mucosal immune system in general. Mucosal lymphoid B-cell follicles and their associated para-follicular T-cell zones (Figure 2(c)) serve for the generation of B- and T-effector lymphocytes, respectively (Figure 7).

The Topographical Distribution of CALT is in the Right Place to Assist Corneal Immune Surveillance If the distribution of CALT is used to draw a topographical map (Figure 5(b)), it corresponds to the position of the cornea during eye closure. CALT in the tarso-orbital regions of the palpebral conjunctiva is then in the right position to support the immune protection of the cornea, which itself is largely devoid of lymphoid cells and other leukocytes. CALT may act during blinking as an immunological windscreen wiper and during sleep as an immunological cushion that covers the cornea (Figure 5(c)). CALT can provide the cornea with innate and specific antibacterial peptides and proteins including SIgA that are not produced there. This concerns the usual daytime setting when the conjunctiva regularly glides over the cornea and wipes it clean. Even more so, CALT may be

438

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

(a)

(b)

(c)

Figure 4 Characteristics of follicular CALT. (a) Even smaller lenticular lymphocyte accumulations are primarily composed of B-cells in immunostaining (brown dots). Over the apex, the overlying epithelium becomes flatter ((b), this follicle is not exactly met at the apex) and changes into a follicle-associated epithelium (FAE, between arrowheads in (c)), where goblet cells are absent, immunostaining for secretory component is rarefied (c) and numerous intraepithelial lymphocytes (IEL) are present in groups, suspicious for M-cells. The IEL are arranged in groups, including B-cells (arrowhead in (a)). Immunostaining for CD20þ B-cells (a) and the IgA-transporter SC (c) of a small, almost flat, follicular accumulation that appears homogeneous, apart from disintegration at the location of the former germinal center, in HE staining (b); size bar in all figures ¼ 100 mm. From Knop, N. and Knop, E. (2000). Conjunctiva-associated lymphoid tissue in the human eye. Investigative Ophthalmology and Visual Science 41: 1270–1279.

relevant during nighttime when the eye is closed. Then, an upregulated level of proinflammatory factors from PMN cells is obtained as a temporary approach in order to dampen the growth of the entrapped microorganisms that enjoy a comfortable environment without disturbance. Due to the intimate contact, CALT can also detect corneal antigens and generate respective effectors.

recirculation of lymphocytes (Figure 2(b) and (c)) within the mucosal immune system in order to repopulate the ocular surface mucosal tissues and other mucosa-associated lymphoid tissue (MALT) locations and in return, EALTcan also share effector cells from other organs.

Mechanisms of Conjunctival Immune regulation CALT is a Part of the Complete Eye-Associated Lymphoid Tissue

CALT Is Physiologically Biased to Tolerogenic, Anti-Inflammatory Responses

The conjunctival mucosa is, at the temporal and nasal side (Figure 6), anatomically continuous, through the lacrimal excretory ducts, with the lacrimal gland and through the lacrimal canaliculi with the lacrimal drainage system, respectively. The histology clearly shows that a continuous mucosal immune system is also present from the periacinar tissue of the lacrimal gland throughout the conjunctiva into the lacrimal drainage system, that is, along the extended ocular surface. Together, this constitutes an eye-associated lymphoid tissue (EALT) and CALT is the regional part of it at the ocular surface proper (Figure 6). EALT is an undividable anatomical and functional unit and its different parts support each other in function. EALT is in line with the other parts of the mucosal immune system of the body, such as gut-associated lymphoid tissue (GALT) in the gut or bronchus-associated lymphoid tissue (BALT) in the bronchi. Therefore, primed effector cells from EALT can be distributed by the regulated

The mucosal immune regulation including CALT is maintained via the mode of antigen presentation by APC, on their MHC-class-II, to the T-cell-receptor (TCR) of naı¨ve CD4þ T-cells (To) and influenced by additional signals such as co-stimulatory molecules and the prevailing cytokine milieu within the tissue. This leads to the generation of different types of CD4þ Th cells which produce characteristic cytokine patterns and have different functions (Figure 7). Due to the prevalence of nonpathogenic antigens and the delicate tissue construction, CALT is biased toward anti-inflammatory immune answers. Tolerance is also necessary in order to avoid autoimmune reactions against own tissue constituents by self-reactive T-cells. Normally, CALT favors Th2 cells under the influence of cytokines such as IL-4. These interact with B-cells and produce cytokines (e.g., IL-4, IL-5, and IL-13) that promote B-cell Ig iso-type class switch to IgA and their differentiation into IgA-secreting plasma cell precursors

Conjunctiva Immune Surveillance

439

Average number of follices in the conjunctiva

Number

5 4 3 2 1

Localization

l

l

ita

sa

r Ta

O

rb

l

ica

rn Fo

r

ba

l Bu

Upper conjunctiva

Temporal

Medial

l

l

r

ba

l Bu

ita

ica

rn Fo

rb

O

l

sa

r Ta

Lower conjunctiva

Nasal

Nasal

Medial

Temporal

T

T

O

O

F F ENK

B

B Diffuse High Medium lymphpoid Few tissue

Follicles

Ocular surface with closed lids (en face view)

Crypt-associated CALT

Sagittal section

Figure 5 Topography of CALT – in the right position to assist corneal immune protection. (a) Morphometrical analysis of lymphoid follicles in the different zones (tarsal, orbital, fornical, bulbar) of upper (left) and lower (right) human conjunctiva, in a flat preparation of a tissue whole mount, shows a main expression in the tarso-orbital zones. More follicles are present in the upper lid and the average total number of CALT follicles in an elderly population is about 10 follicles per eye. (b) The topographical distribution of lymphoid follicles is the same as that of the diffuse lymphoid effector tissue in which they are interspersed, as seen in a topographical map of CALT in a complete flat whole mount of a human conjunctiva with the lid margin to the top and the nasal zone in the middle. The map differentiates the previously mentioned zones (T, O, F, B) as well as temporal, medial, and nasal locations. Increasing density of diffuse lymphoid cells is indicated as increased shades of gray. Hatched lines indicate the location of conjunctival crypts of Stieda that are associated with CALT. (c) If the topographical distribution of CALT is projected onto the bulbar surface in a closed lid situation, it is obvious that CALT covers the cornea as seen in frontal en face view (middle) and sagittal section (right). The central portion is covered by the tarsal crypts of Henle (open circles in middle figure) that are associated with CALT but not indicated in the topographical map in (b). From Knop, E. and Knop, N. (2003) Eye-Associated Lymphoid Tissue (EALT) is Continuously Spread Throughout the Ocular Surface from the Lacrimal Gland to the Lacrimal Drainage System. Der Ophthalmologe, Heidelberg: Springer.

440

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

T T

Lacrimal glandassociated lymphoid T tissue

T

T

T

Conjunctivaassociated lymphoid tissue (CALT)

T

T

T

T

Lacrimal drainageassociated lymphoid tissue (LDALT)

T T T

T T T T T B B B B B B B B B B T

T

T

T

T

T T T B B T B B B T B B B B B

TT

T

T

T

T

T

T

T

T

T T

T

T T T T T T

T

T T

T T

d

o Blo

od

T

T

Blo

T

m Ly

ph

Lym

ph

Lymphocyte recirculation

T

T

Lymph

Blood

T T T T T B BB B T B B B T B B BB

T

T

T

ENK

Figure 6 Eye-associated lymphoid tissue (EALT). Eye-associated lymphoid tissue integrates the continuous mucosal immune system of the conjunctiva (CALT) and its mucosal adnexa, composed of the lacrimal gland and lacrimal drainage system, which together form the extended functional ocular surface. The ocular tissues belong together embryologically and functionally. They are connected by the flow of tears (yellow arrows) which lets them share protective immune factors as well as antigens and they are furthermore connected by the regulated recirculation of lymphoid cells in the body via efferent lymph vessels and blood vessels. Lymphoid cells enter the tissues via blood vessels, including high endothelial venules (represented by roundish endothelial cells), and leave them via lymphatics. EALT is continuous throughout these organs in the form of a diffuse lymphoid effector tissue composed of T-lymphocytes (represented by black cells in the drawing) and plasma cells (represented by large blue cells) together with accessory leukocytes (not indicated here, compare Figure 2(b)). Inductive sites, in the form of lymphoid follicles, composed of B-cells (small blue cells) with adjacent para-follicular T-cell areas (compare Figure 2(c)), are present in CALT and LDALT. They serve for the uptake and presentation of antigens and for the subsequent generation of respective effector cells specific for the ocular surface relevant antigens. Not only the effector cells generated in EALT, but also those from other mucosal sites, can, after recirculation in the lymph and blood system, populate the diffuse lymphoid effector tissue that is present along the whole extended ocular surface including the large mucosa-associated gland (lacrimal gland). Adapted from Knop, E. and Knop, N. (2007). Anatomy and immunology of the ocular surface. In: Niederkorn, J. Y. and Kaplan, H. J. (eds.) Immune Response and the Eye. Chemical Immunology and Allergy. Basel: Karger Verlag.

and mature plasma cells as observed in the conjunctiva. Under the influence of mainly IL-6, the well-known inflammatory Th1 cells are formed that produce inflammatory cytokines such as IFN-g and TNF-a which have the physiological function to activate cells, in particular phagocytes, to destroy intracellular pathogens. If inflammatory cytokines and other danger signals, such as bacterial LPS or components of dead cells occur in the tissue, they can bind to TLRs and mediate the secretion of excess inflammatory cytokines that skew CALT toward inflammatory immune answers. The antagonistic action of Th1 and Th2 cells led to the construction of the Th1–Th2 paradigm for explanation of immune regulation and phenomena at the ocular surface. In recent years, however,

other anti-inflammatory (regulatory T-cells, Treg) and inflammatory (Th17 cells) were also observed which indicated that immune regulation is more complex (Figure 7) and needs further investigation. Deregulation of EALT Is a Central Component of Inflammatory Ocular Surface Disease Various stress mechanisms, for example, mechanical alteration, hyperosmolar tears or exposure to inflammatory cytokines can pathologically activate the ocular surface epithelium that responds by secretion of (further) inflammatory cytokines and proteases (such as matrixmetalloproteinase, MMP) and upregulates surface

Conjunctiva Immune Surveillance

TLR Costimulation

441

Microbe

MHC-II Antigen TCR

ENK

APC

IL-4 IL-4, IL-5 IL-13 B

To Th1

Th2

IFN-gamma TNF-alpha

IL-6 TGFbeta Treg

Th17

Macrophage IgA

TGFbeta

ANTI-INFLAMMATORY

IL-17

INFLAMMATORY

Figure 7 Generation of T- and B-effector cells in CALT. Immune regulation in CALT is governed via the presentation of antigens by antigen-presenting cells (APC) on their MHC-class-II (MHC-II) to the T-cell-receptor (TCR) of naive CD4þ T-cells (To), and assisted and modulated by co-stimulatory molecules. This is influenced (indicated by uninterrupted arrows) by cytokines (interleukins, IL) and by microbial antigens that are recognized, for example, by TLR. This leads to the generation (interrupted arrows) of different types of CD4þ T-helper cells (Th) that produce characteristic cytokine patterns and perform different functions. CALT is naturally biased toward anti-inflammatory immune answers. It favors the generation of Th2 that promote B-cells to differentiate into IgA secreting plasma cells. It also favors Treg that produce anti-inflammatory TGF-beta. Binding of microbial antigens to TLR and presence of IL-6 represents danger signals to the immune system and induces inflammatory immune answers via Th17 and Th1, the latter of which normally assist phagocytic cells to clear intracellular pathogens.

markers, such as ICAM-1 and MHC-class-II. This promotes an inflammatory process and through MCH-classII the epithelial cells acquire the potential for presentation of antigens, including self-antigens, to resident conjunctival T-cells (Figure 8(a)) that can induce a loss of natural conjunctival immune tolerance. Similar events are also shown for the acinar epithelial cells in inflammatory disease of the lacrimal gland where diverse perturbations result in altered intracellular protein traffic, alter the lacrimal acinar cell autoantigenic spectra, and upregulate MHC-class-II. This results in a loss of tolerance to own cell constituents, such as the M3 receptor with a subsequent autoimmune process. This again indicates that the natural tolerogenic bias can be lost in inflammatory disease and may be the underlying reason for a self-perpetuating inflammatory process at the ocular surface and its associated gland. In fact, in inflammatory ocular surface diseases such as dry eye disease, autoreactive T-cells are generated that are specific to ocular surface tissue. They can be

transferred and lead to destruction of the same ocular tissues in a naive recipient that has never experienced the pathological condition. Respective Tregs can prevent the tissue destruction and offer therapeutic potentials. In addition, wounding can allow the entry of nonpathogenic antigens into the tissue and their presentation to Tand B-cells, as observed in ocular allergy. Downstream effects are the activation of conjunctival vascular endothelial cells that upregulate adhesion molecules (such as ICAM-1, VCAM-1, or E-selectin) with subsequent recruitment of further leukocytes from the vascular compartment and the activation of bystander cells including stromal fibroblasts. They contribute to the accumulation of MMPs that lead to tissue degradation. Altogether, this constitutes an immune-mediated conjunctival inflammatory process (Figure 8(a)), that can be compared with events in other mucosal organs, for example, in inflammatory bowel disease, and is based on a deregulation of the physiologically protective CALT and perpetuated by several vicious circles.

442

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

Mechanisms for loss of physiological mucosal immune tolerance Epithelial alteration Irritation

Lacrimal gland Neural reflex-arc

Defect

AG

Tear deficiency

ENK

Ocular surface epithelium

ENK MHC-II+ Co-Stim.

Th17

Th17 To

DC

To Th1

Wetting defect

Conjunctiva

Th1

Epithelial defects

Th1

Th1

IFN-γ IL-6, IL-17 TNF-α Macrophage

Fibroblast

Impaired secretion

T

ICAM-1 VCAM-1 E-selectin

Activation and increased permeability of ocular surface epithelium

Vessel

T-cell activation Inflammatory cytokines

Matrixmetalloproteinase (a)

Immunemodulated inflammation

(b)

(Subclinical inflammation)

Disrupted equilibrium of cell-differentiation and tissue degradation

Figure 8 Deregulation of CALT determines immune-mediated inflammatory ocular surface disease. (a) Compromised integrity leads to loss of mucosal immunological tolerance and to immune-mediated inflammation. Irritation of the surface epithelium, for example, by tear film (light blue) defects, its infection or its wounding results in activation of the epithelial cells. These may respond by secretion of inflammatory cytokines and proteases (such as matrix-metalloproteinase, MMP) and by expression of the antigen-presenting molecule MHC-class-II (MHC-II) resulting in potential presentation of self-antigens to resident conjunctival T-cells. Wounding with physical defects can, in addition, allow the entry of nonpathogenic antigens into the tissue and their presentation in the context of inflammatory cytokines as observed in ocular allergy. As described for immune-regulation (Figure 7) this represents danger signals that contribute to a further accumulation of inflammatory cytokines in the tissue and to the generation of inflammatory, potentially autoreactive, types of T-cells. All of which is shown in inflammatory ocular surface disease. Downstream effects are the activation of vascular endothelial cells that upregulate adhesion molecules with subsequent recruitment of further leukocytes from the vascular compartment and the activation of bystander cells, including stromal cells (fibroblasts). They contribute to the accumulation of MMPs that lead to tissue degradation which all together constitutes an immune-mediated conjunctival mucosal inflammatory process that is based on a deregulation of the physiologically protective CALT. (b) Immune-mediated inflammation is a core mechanism that results in several vicious circles in the pathogenesis and propagation of ocular surface disease. An immune-mediated inflammation represents an important common factor in the vicious circles of ocular surface disease, including the dry eye syndrome and ocular allergy, which is first subclinical but tends to amplify if it is not limited. Tear film deficiency results in epithelial defects and these in turn are an important primary factor for onset of an immune-mediated inflammatory conjunctival process that tends to selfpropagation via several vicious circles. These include disturbance of afferent innervation resulting in impaired secretion with further tear film deficiency and increase of epithelial damage and in impairment of mature ocular surface differentiation (leading to squamous metaplasia) that results in wetting defects and amplification of epithelial destruction. (a) From Knop E. and Knop N. (2005). Influence of the Eye-associated Lymphoid Tissue (EALT) on Inflammatory Ocular Surface Disease. The Ocular Surface, Ethis Communications. (b) Adapted from Knop E. et al. (2003). Dry Eye Disease as a Complex Dysregulation of the Functional Anatomy of the Ocular Surface. New Impulses to Understanding Dry Eye Disease. Der Ophthalmologe, Heidelberg: Springer.

See also: Adaptive Immune System and the Eye: Mucosal Immunity; Corneal Epithelium: Response to Infection; Defense Mechanisms of Tears and Ocular Surface; Dry Eye: An Immune-Based Inflammation; Immunopathogenesis of Onchocerciasis (River Blindness); Molecular and Cellular Mechanisms in Allergic Conjunctivitis; Pathogenesis of Fungal Keratitis; Tear Drainage.

Further Reading Argueso, P. and Gipson, I. K. (2001). Epithelial mucins of the ocular surface: Structure, biosynthesis and function. Experimental Eye Research 73: 281–289. Brandtzaeg, P. and Pabst, R. (2004). Let’s go mucosal: Communication on slippery ground. Trends in Immunology 25: 570–577. Chodosh, J., Nordquist, R. E., and Kennedy, R. C. (1998). Comparative anatomy of mammalian conjunctival lymphoid tissue: A putative

mucosal immune site. Developmental and Comparative Immunology 22: 621–630. Dua, H. S., Gomes, J. A., Jindal, V. K., et al. (1994). Mucosa specific lymphocytes in the human conjunctiva, corneoscleral limbus and lacrimal gland. Current Eye Research 13: 87–93. Hingorani, M., Metz, D., and Lightman, S. L. (1997). Characterisation of the normal conjunctival leukocyte population. Experimental Eye Research 64: 905–912. Knop, E. and Brewitt, H. (1992). Morphology of the conjunctival epithelium in spectacle and contact lens wearers – a light and electron microscopic study. Contactologia 14: 108–120. Knop, N. and Knop, E. (2000). Conjunctiva-associated lymphoid tissue in the human eye. Investigative Ophthalmology and Visual Science 41: 1270–1279. Knop, E. and Knop, N. (2003). Eye-associated lymphoid tissue (EALT) is continuously spread throughout the ocular surface from the lacrimal gland to the lacrimal drainage system. Ophthalmologe 100(11): 929–942. Knop, E. and Knop, N. (2005). Influence of the eye-associated lymphoid tissue (EALT) on inflammatory ocular surface disease. Ocular Surface 3(4): S180–S186. Knop, E. and Knop, N. (2005). The role of eye-associated lymphoid tissue in corneal immune protection. Journal Anatomy 206: 271–285.

Conjunctiva Immune Surveillance Knop, N. and Knop, E. (2005). Ultrastructural anatomy of CALT follicles in the rabbit reveals characteristics of M-cells, germinal centers and high endothelial venules. Journal of Anatomy 207: 409–426. Knop, E. and Knop, N. (2007). Anatomy and immunology of the ocular surface. In: Niederkorn, J. Y. and Kaplan, H. J. (eds.) Immune Response and the Eye. Chemical Immunology and Allergy, vol. 92, pp. 36–49. Basel: Karger Verlag. Knop, E., Knop, N., and Brewitt, H. (2003). Dry eye disease as a complex dysregulation of the functional anatomy of the ocular surface. New impulses to understanding dry eye disease. Ophthalmologe 100: 917–928. Knop, E., Knop, N., and Claus, P. (2008). Local production of secretory IgA in the eye-associated lymphoid tissue (EALT) of the normal human ocular surface. Investigative Ophthalmology and Visual Science 49: 2322–2329. Liu, H., Meagher, C. K., Moore, C. P., and Phillips, T. E. (2005). M cells in the follicle-associated epithelium of the rabbit conjunctiva

443

preferentially bind and translocate latex beads. Investigative Ophthalmology and Visual Science 46: 4217–4223. McDermott, A. M., Perez, V., Huang, A. J., et al. (2005). Pathways of corneal and ocular surface inflammation: A perspective from the cullen symposium. Ocular Surface 3: S131–S138. Mircheff, A. K., Wang, Y., Jean, Mde S., et al. (2005). Mucosal immunity and self-tolerance in the ocular surface system. Ocular Surface 3: 182–192. Pflugfelder, S. C. and Stern, M. E. (2007). Future directions in therapeutic interventions for conjunctival inflammatory disorders. Current Opinion in Allergy and Clinical Immunology 7: 450–453. Sack, R. A., Beaton, A., Sathe, S., et al. (2000). Towards a closed eye model of the pre-ocular tear layer. Progress in Retinal and Eye Research 19: 649–668. Sullivan, D. A. (1999). Ocular mucosal immunity. In: Ogra, P. L., Mestecky, J., Lamm, M. E., et al. (eds.) Handbook of Mucosal Immunology, 2nd edn., pp. 1241–1281.

Defense Mechanisms of Tears and Ocular Surface A M McDermott, University of Houston, Houston, TX, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Antigen-presenting cell – Bone-marrow-derived cell that ingests antigen and presents it to T lymphocytes to trigger adaptive immune response. Antimicrobial peptide – Small protein, typically cationic, with broad spectrum antimicrobial activity. Pattern recognition receptor – Cell surface or intracellular protein receptor that binds conserved motifs (pathogen-associated molecular patterns) on microorganisms. Siderophore – Iron chelating compound secreted by bacteria and fungi to facilitate uptake of this essential nutrient.

the cornea is involved as this structure provides the majority of the refracting power of the eye and loss of visual acuity or even blindness may be the consequence. Given appropriate circumstances, a range of organisms can infect the ocular surface. The most common bacterial species are Pseudomonas aeruginosa, Serratia marcescens, Staphylococcus aureus, Staphylococcus epidermidis, and Streptococcus pneumoniae. Herpes simplex virus is the most common viral culprit. Fungal infection is less common than bacterial or viral infection, but is on the rise, with the most frequently isolated species in the USA being Aspergillus, Candida, and Fusarium. Some of the ocular surface and tear defenses that help prevent these infections are depicted in Figure 1 using the corneal epithelium as a model.

Defense Mechanisms of Tears Introduction The ocular surface is composed of the epithelia of the cornea, limbus, and conjunctiva, the anatomy and physiology of which are discussed in detail elsewhere in this encyclopedia. The tear film coats the epithelia and is a complex structure composed of an outer anterior-most lipid component that prevents evaporation, an aqueous component that has ions, soluble mucins, enzymes, and a range of other proteins and closest to the epithelial surface is a thick mucus, primarily composed of the gelforming mucin MUC5AC. Together, the ocular surface epithelia and tears: (1) create a formidable barrier that helps prevent microbial attachment in the first place, (2) bombard organisms with a plethora of chemicals to stop them dead in their tracks (or at least stop them proliferating), and (3) provide a detection system such that when an organism actually manages to circumvent the primary innate defenses, adaptive immunity can be activated to provide further help to eliminate the offending organism. The defense mechanisms are remarkably effective, as despite constant exposure to the external environment and frequent interaction with unwashed fingertips, the ocular surface rarely succumbs to infection. Generally only when there is physical disruption of the epithelial barrier such as occurs with contact lens wear or injury do the defenses encounter serious challenge and infections become a significant cause of morbidity. Infection can affect both the cornea (infectious keratitis) and conjunctiva (infectious conjunctivitis), but is most serious when

Tears provide both physical/mechanical and chemical defense to the ocular surface. The act of blinking moves tears toward the puncta and into the lacrimal sac, thus helping to wash away any potential pathogens before they have had time to interact with and invade the ocular surface epithelial cells. Furthermore, reflex tearing increases tear volume which helps to rapidly dilute harmful substances released by invading pathogens. Chemical entities with antimicrobial properties present in tears primarily originate from lacrimal and accessory gland secretions and the ocular surface epithelial cells. Other components include serum exudates and secreted products of neutrophils and other infiltrating cells. Several tear film components have been identified that have direct antimicrobial activity or which can otherwise limit pathogen entry and growth. The first identified was lysozyme which was shown to kill Gram-positive bacteria by Alexander Fleming (of penicillin fame) in 1922. This enzyme accounts for 20–40% of total tear protein and has the ability to catalyze the hydrolysis of 1,4-beta-linkages between N-acetylmuramic acid and N-acetyl-D-glucosamine in the peptidoglycan backbone of bacterial cell walls. It is also able to cleave chitodextrins in fungal cell walls. The compromised cell wall is then no longer able to maintain a stable osmotic environment and lysis of the organism ensues. Secretory phospholipase A2 (sPLA2) has been identified as the major tear protein active against Gram-positive bacteria, although it has no activity against Gram-negatives in the normal ionic environment of tears and notably is some 50-fold less abundant in tears than lysozyme. Secretory PLA2 hydrolyzes the sn-2-fatty acyl

444

Defense Mechanisms of Tears and Ocular Surface

445

Figure 1 Summary of defense mechanisms of tears and ocular surface. Schematic diagram of some of the various defenses used by tears and ocular surface epithelia to prevent infection (additional ones are discussed in the text). The corneal epithelium is depicted (note not all cell layers are shown) but as described in the text similar mechanisms are used by conjunctival epithelial cells. PRR, pattern recognition receptor (e.g., TLRs); SP-A, surfactant protein-A; SP-D, surfactant protein-D; sIgA, secretory IgA; sPLA2, secretory phospholipase A2.

moiety from phospholipids, in particular, phosphatidylglycerol, which is abundant on bacterial cell membranes. Other tear fluid components showing direct microbial killing include the cationic antimicrobial peptides (AMP; see the section entitled ‘Defense mechanisms of the ocular surface epithelia’) and the a-defensins human neutrophil peptide (HNP) 1, 2, and 3. As implicated by their names these peptides are produced by neutrophils and while present in only low concentrations under normal circumstances their level in tears rises after ocular surface injury. As depicted in Figure 2, these positively charged peptides exert their activity by interacting electrostatically with the negatively charged microbial cell membrane and form pores or otherwise disrupt the membrane leading to disturbances of respiration and metabolism, leakage of cell contents, and eventual death of the organism. A lytic effect of these peptides is also possible.

Beta-lysin and secretory leukocyte protease inhibitor (SLPI) are other examples of cationic proteins found in tears, which also interact with bacterial cell membranes and cause cell lysis. SLPI, which comes from lacrimal gland and ocular surface epithelial cells, has a defensinlike domain that confers antimicrobial activity and is also a potent inhibitor of neutrophil elastase. Thus, SLPI can both help prevent infection and protect host cells from the damaging effects of neutrophil enzymes. Elafin is another protein, which exerts similar dual functionality. Histatins, which are small histidine-rich AMPs with antifungal activity, have also been detected in the tears. A number of tear components limit bacterial growth rather than actually killing the invading organism. Lactoferrin represents some 21% of the total reflex tear protein and has a high capacity to bind divalent cations including iron, thus depriving many bacteria of this essential

446

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

Pertubation of membrane and loss of intracellular contents = microbial death

+

+

+

+

+ +

Figure 2 Mechanism of action of antimicrobial peptides. Positively charged peptides (shown in yellow) such as defensins and cathelicidin (LL-37) interact with negatively charged microbial membranes leading to disruption of the membrane, and possibly transient or stable pore formation, which results in leakage of intracellular contents, disturbance of metabolism, and death of the organism.

nutrient for growth. Furthermore, a highly basic sequence at the N-terminus (referred to as lactoferricin) allows lactoferrin to act as a cationic detergent and disrupt the cell membrane of some organisms. Tear lipocalin represents approximately 25% of the reflex tear proteins and was recently shown to be capable of binding siderophores produced by a range of bacteria and fungi. Siderophores are chelating compounds that transport iron into microorganisms. Thus, like lactoferrin, lipocalin exerts a bacteriostatic effect by interfering with the ability of pathogens to take up iron. The tear film also contains members of the collectin family of C-type lectins, surfactant proteins (SP)-A and -D. Collectins bind to carbohydrates on the surface of various microorganisms and to receptors on phagocytic cells, which in turn promotes their phagocytic activity. SP-A and SP-D are produced by lacrimal gland cells and also corneal and conjunctival epithelial cells. SP-D is known to exert growth inhibitory effects on some Gram-negative bacteria, to promote pathogen phagocytosis by mononuclear cells and through an interaction with lipopolysaccharide (LPS), which is present on the outer membrane of Gram-negative bacteria, can inhibit bacterial adhesion to target cells. SP-D inhibits corneal epithelial cell invasion by P. aeruginosa (a Gram-negative pathogen that commonly causes ulcers in contact lens wearers), possibly via an LPS-dependent mechanism. Secretory IgA (sIgA) is the predominant immunoglobulin in the tear film. This antibody is of great importance as it facilitates removal of pathogens right at the point of entry at the ocular surface. sIgA is not a particularly efficient activator of complement (important for preventing unwanted inflammation) or a good opsonin (although

neutrophils do have receptors for sIgA which when engaged could trigger phagocytosis). The major effector mechanism of sIgA is neutralization, which prevents attachment to host cells. sIgA can also bind to lectin-like adhesin molecules on pathogens causing them to aggregate and trapping them within the tear film. sIgA is produced by plasma cells (terminally differentiated B lymphocytes) residing in the lacrimal gland and in specialized areas of the conjunctiva referred to as conjunctival-associated lymphoid tissue. sIgA binds to specific receptors on lacrimal gland acinar cells, conjunctival epithelial cells, and is taken up by endocytosis, and then traverses the cell by transcytosis. This antibody is then released into the tears attached to a protein called secretory component, a fragment of the receptor to which the antibody was bound during its passage through the cells. Secretory component stabilizes the antibody and masks proteolytic sites so conferring resistance to host and pathogen proteases. Low levels of functionally active complement and complement regulatory proteins have also been detected in tears. An overview of the complement pathway is presented in Figure 3. The relative amounts of different components, namely abundant C3 and factor B, but less C1q, suggest that activation via the alternative pathway (i.e., spontaneous hydrolysis of C3) is the predominant mechanism. Possible sources of the various complement components are leakage of plasma through the conjunctival vessels during sleep, infiltrating neutrophils, and local synthesis by corneal and conjunctival epithelial cells. Activation of the complement pathway generates fragments involved in acute inflammatory responses, fragments that act as opsonins which facilitate target recognition by neutrophils and results in the formation of membrane attack complexes that can lyse pathogens (and host cells). The complement pathway is believed to be most active when the eyes are closed (see comments below on closed-eye tears). To prevent unnecessary activation and hence tissue damage, the complement pathway is regulated by a number of factors. This pathway is inhibited by molecules such as lactoferrin and vitronectin both of which are present in the tears and CD55 (decay accelerating factor) as well as CD59 which are membranebound molecules expressed by corneal and conjunctival epithelial cells. In immediate apposition to the superficial epithelial cells is a blanket of mucus, composed primarily of the gelforming mucin MUC5AC which is secreted by goblet cells in the conjunctiva in response to parasympathetic stimulation. This blanket interacts with the glycocalyx coating the superficial cells. Membrane spanning mucins MUC 1, 4, and 16 produced by the epithelial cells are important components of the glycocalyx and can be cleaved from the cell surface and released into the tear film. Mucins are known to help prevent bacteria from

Defense Mechanisms of Tears and Ocular Surface

447

Alternative pathway Microbial surfaces

C3

Factors B&D

C3b

C3 convertase

C5 convertase

Factor P MAC C6, C7, C8, C9 C3 Mannan-binding lectin

C4

+

C3b

C3a

C2

C3 convertase

C5

C5b

C5a

Target lysis

C5 convertase

Activated C1 Classical pathway

Antibody-antigen complex

Figure 3 Overview of the complement pathway. The major components of the three known pathways (classical (blue), alternative (yellow) and mannan-binding lectin (light blue)) that lead to activation of complement are shown. In tears the primary pathway is believed to be the alternative pathway but, as can been seen, regardless of the mechanism of activation the outcome is the same (green), i.e., production of intermediaries with enzymatic activity (the convertases), generation of inflammatory mediators such as C3a and C5a and of opsonins such as C5b which bind to pathogen surfaces to facilitate their recognition by phagocytes. The final product of the pathway is the membrane attack complex (MAC) in which several complement components come together to form pores on a pathogen surface leading to lysis and death of the pathogen.

reaching epithelial surfaces. This function has been attributed to a number of mechanisms, for example, mucin can bind and trap the bacteria, which are then effectively removed from the ocular surface by blinking. It should be noted, however, that the ability to interact with mucins varies widely among different organisms. There is also evidence that sIgA and positively charged proteins such as lysozyme and SLPI accumulate in the mucous blanket, thus providing a reservoir of antimicrobial agents. Therefore, mucins may trap microbes, which are then killed by accumulated antimicrobials or aggregated by sIgA and then cleared by blinking. Thus, tears are equipped with a plethora of chemical entities capable of neutralizing invading pathogens. While many tear components have independent antimicrobial effects, several are thought to cooperate in a synergistic fashion to yield maximal effect. For example, sequestering of cations by lactoferrin destabilizes the cell wall of Gramnegative bacteria making the peptidoglycan layer more accessible to cleavage by lysozyme. Also, it should be brought to the reader’s attention that the composition of the tears changes during sleep. Open-eye and reflex tears have primarily lysozyme, lactoferrin, lipocalin, and sIgA, whereas closed-eye tears have increased amounts of sIgA (up to 80% of total tear protein), complement components, and of serum-derived proteins. There is also a large influx of neutrophils within 2–3 h of eye closure, which

provides additional defense factors in the guise of AMPs and reactive oxygen species, for example. Overall, these changes appear to represent a shift to a subclinical state of inflammation, which is believed to be necessary to protect the ocular surface from invasion by entrapped pathogens while the lids are closed. There is also an increase in proteins such as SLPI and elafin, which have potent antiprotease activity, and vitronectin, which inhibits complement, which serve to protect the ocular surface cells in this proinflammatory environment.

Defense Mechanisms of the Ocular Surface Epithelia Mechanical/Physical Defenses The outermost superficial epithelial cells are bound by tight junctions, which effectively seal two cells together forming a barrier against free diffusion of fluids, electrolytes, and macromolecules as well as microorganisms and their secreted products. Tight junctions are also important in establishing and maintaining cell polarity. Polarized cells are characterized by differences in the composition and distribution of proteins and other surface molecules between apical and basolateral surfaces. This arrangement is maintained by the aforementioned tight junctions that segregate the domains and targeted delivery

448

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

that sends the molecules to their correct location. Disruption of polarity, such as occurs with corneal epithelial cells migrating to recover an injured area, has been shown to increase susceptibility to infection. This may be the result of a number of factors, for example, host cell receptors with which pathogens interact may be more abundant on basolateral surfaces. Also, loss of polarity disrupts the apical mucin-containing glycocalyx, which normally helps restrict bacterial attachment. Another important feature contributing to defense is the constant turnover of the ocular surface epithelial cells. Both corneal and conjunctival epithelia have a population of strategically located stem cells that provide new cells to replace those being shed into the tear film. In the absence of a penetrating injury, infection begins in the most superficial epithelial layers and so through the constant renewal of cells, outer infected cells may be sloughed off before there is time for the infection to spread to the lower epithelial layers. Pathogen Recognition While the tear film acts to prevent pathogens from reaching the ocular surface epithelial cells in the first place, it is very important that the cells have a system to recognize the presence of invading organisms if they do happen to conquer the outer defenses. Host cell proteins that mediate pathogen recognition are often referred to as pathogen recognition receptors (PRRs) and they recognize pathogen-associated molecular patterns present on bacteria, viruses, and fungi. The primary PRRs on epithelial cells are toll-like receptors (TLRs). These are type I transmembrane glycoproteins, which have an extracellular leucine-rich domain and a cytoplasmic domain homologous to the signaling domain of the interleukin (IL)-1 receptor. Of the 10 functional human TLRs that have been identified all have been reported to be expressed by corneal and conjunctival epithelial cells. Figure 4 shows a simplified diagram of the distribution of TLRs, their ligands, and signaling pathways. TLR1, 2, 4, 5, 6, and 10 are typically located at the cell surface. TLR2 forms heterodimers with TLR1 and with TLR6 and so can recognize a large variety of microbial products. For example, TLR2/6 heterodimers recognize lipoteichoic acid from Gram-positive bacteria and TLR2/1 heterodimers recognize triacyl lipoprotein/peptides of bacterial cell walls. TLR4 forms a complex with MD2 (also known as lymphocyte antigen 96) and cluster of differentiation 14 (CD14) protein and recognizes LPS from Gram-negative bacteria, while TLR5 recognizes flagellin, a component of bacterial flagella. TLR10, the ligand for which is unknown, is able to dimerize with TLR1 and TLR2. TLR3, 7, 8, and 9 are (typically) all located intracellularly, on endosomal membranes and recognize nucleic acids. TLR3 recognizes double-stranded RNA, a by-product of the replication of some viruses, whereas TLR7 and 8

recognize viral single-stranded RNA. TLR9 responds to unmethylated cytosine–phosphate–guanosine dinucleotide motifs found in both bacterial and viral DNA. Thus, by interacting with specific pathogen-derived molecules TLRs can detect the presence of a wide range of organisms, including those that replicate intracellularly. The engagement of TLRs with their specific microbial ligand results in activation of intracellular signaling pathways, leading to a variety of functional changes in the ocular surface epithelial cells. The latter include production of inflammatory cytokines such as IL-6 and chemokines such as IL-8 that will attract neutrophils for phagocytosis and AMPs such as human b-defensin-2 that can directly kill invading pathogens (see the section entitled ‘Antimicrobial peptides’). It is important that members of the normal ocular flora do not trigger TLR activation and hence cause unwanted inflammatory reactions at the ocular surface. To this end, it has been observed that flagellin from pathogenic, but not from nonpathogenic bacteria, can activate TLR5 in corneal epithelial cells. Expression of TLR5 (and possibly TLR4) appears to be restricted to basal and wing cells suggesting that TLR5 will only be activated when there is a breach in the corneal epithelium. Also, rather than being surface bound, TLR4 may be expressed intracellularly and so would not be available. Evidence for the expression of another class of PRRs, the cytoplasmic nucleotide-binding and oligomerization domain (NOD) proteins, has yet to be investigated for the human ocular surface. However, mouse anterior eye tissue expressed both NOD1 (which recognizes meso-DAP, a component of peptidoglycan in Gram-negative organisms) and NOD2 (which recognizes muramyl dipeptide found in both Gram-positive and -negative bacteria). Antimicrobial Peptides AMPs are small peptides, most less than 50 amino acids, that are amphipathic and typically carry an overall positive charge (+2 or greater) due to a relative excess of amino acids such as arginine and lysine. These peptides show a broad spectrum of antimicrobial activity and many have additional effects on mammalian cell behavior. The two major categories of mammalian AMPs are the defensins and cathelicidins. Human defensins are characterized by the presence of six cysteine residues that interact to form three disulfide bonds (the specific pattern of connectivity gives rise to two classes referred to as a and b). Both corneal and conjunctival epithelial cells express at least three b-defensins (hBDs). hBD-1 and hBD-3 are constitutively expressed, whereas the expression of hBD-2 is variable, being expressed by normal tissue only occasionally. Ocular surface hBD-2 expression is known to be inducible by exposure to both Gram-negative and -positive bacteria and

Defense Mechanisms of Tears and Ocular Surface

ria

LPS

cte

Ba

Unme

llin

e lag

MD2

F

op

Virus

ssRNA dsRNA

6 R T L LR 2 T

Lip

thylate d CpG

CD14

4 TLR 4 TLR

R5 TL

ins

te ro

449

R

8 7/

R

9

3

TL

R

TL

TL

T TL LR R 2 1

R3

TL

MyD88

AP-1 NF-κB

IRF 3/7

DNA

Cytokines, chemokines, cell surface markers, antimicrobial peptides Figure 4 Toll-like receptor activation. The major TLRs expressed by ocular surface epithelial cells and their known bacterial and viral ligands are shown. For clarity the details of the several signaling pathways that lead to transcription factor activation and gene transcription have been omitted. LPS, lipopolysaccharide; CpG, cytosine–phosphate–guanosine dinucleotide; ssRNA, single stranded RNA, dsRNA, double stranded RNA, MyD88, myeloid differentiation protein 88; IRF 3/7, interferon regulatory factor 3/7; AP-1, activating protein-1; NF-kB, nuclear factor kB.

bacterial products such as LPS, peptidoglycan, and lipoproteins. This upregulation is chiefly mediated via the activation of TLRs such as TLR2. In a recent study, expression of a novel beta defensin gene DEFB 109 was detected in the ocular surface epithelia, and interestingly its expression was decreased in inflammation and infection. As noted earlier, a-defensins HNP-1 through -3 are produced by neutrophils and are present in the normal tear film. They can also be detected in the cornea and conjunctiva when neutrophils infiltrate in response to a specific stimulus. The cathelicidins have a highly conserved N-terminal cathelin domain and a variable antimicrobial domain. Only one, LL-37, is expressed in humans. LL-37 is expressed by both corneal and conjunctival epithelial cells and its expression is increased in response to corneal epithelial injury and bacterial challenge with P. aeruginosa and S. aureus. LL-37 is also a major component of neutrophil granules;

thus, its ocular surface levels are expected to rise in situations leading to infiltration of these and other inflammatory cells. While defensins and LL-37 represent the main AMPs present at the ocular surface, others have been reported including liver expressed AMP-1 and -2, statherin, CCL28 and CXCL-1 (two of many antimicrobial cytokines), MIP3a, and thymosin b-4. However, as most of these molecules have other recognized functions, it is unlikely that antimicrobial effects are the major facet of their action at the ocular surface. The primary site of AMP action is the microbial cell membrane, electrostatic disruption of which leads to permeabilization, loss of essential intracellular components, and death (see Figure 2). However, intracellular targets may also be utilized leading to inhibition of protein, peptidoglycan, and nucleic acid synthesis and interference with the activity of bacterial heat-shock proteins. Epithelial b-defensins and LL-37 are effective against

450

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

common ocular surface bacterial pathogens in vitro with hBD-3 and LL-37 showing the broadest spectrum and most potent activity. Animal studies have revealed that experimental infection with P. aeruginosa in genetically modified mice unable to express cathelicidin causes much more severe disease and corneal damage than in normal wild-type mice, thus showing that AMPs are important for ocular surface defense in vivo. In addition to exerting direct antimicrobial effects, AMPs have other properties that help protect against pathogens and their destructive actions. For example, LL37 is known to bind and neutralize LPS. The latter is a product from Gram-negative bacteria and through activating TLR4 induces an inflammatory response and likely mediates much of the ocular surface damage that results from infections with pathogens such as P. aeruginosa. Thus, LL-37 may have a role in dampening LPS-mediated ocular surface inflammation and damage. Also, both defensins and LL-37 have been shown to be chemotactic for a variety of immune and inflammatory cells, including lymphocytes, monocytes, and immature dendritic cells thus may help draw these cells to a site of infection. AMPs also stimulate inflammatory and immune cell cytokine production, which in turn can modulate cellular functions. Additionally, hBD-3 has been shown to activate dendritic cells, raising the possibility that corneal/conjunctival epithelial hBD-3 may be able to activate epithelial Langerhans cells and stromal dendritic cells which in turn may initiate an adaptive immune response.

The Langerhans cells are typically found between basal epithelial cells and in the cornea their density is lowest in the central region and gradually increases toward the periphery. Also, most cells in central cornea appear to be small immature cells. In the periphery, larger mature major histocompatibility complex class II (MHC II) expressing cells with prominent dendritic processes are observed. Overall Langerhans cells perform a surveillance function, screening their environment for pathogens that have breached the defenses of the tear film and the epithelial barrier. If successful in their search they then activate adaptive immunity to help eliminate the invader. Sensory nerves, which are particularly abundant in the cornea, also provide an important contribution to ocular surface defense. When triggered nerves induce the production of reflex tears, which, by virtue of their increased volume, help wash pathogens from the ocular surface and dilute out their toxic products. Release of neuropeptides such as substance P from the nerve termini may also affect epithelial cell cytokine production that, as noted above, may modulate other aspects of host defense. The presence of a complement of nonpathogenic organisms also assists in preventing infection. Such commensals deplete the tears of nutrients, occupy attachment sites so preventing binding of pathogens, and produce bacteriocins that kill members of nonrelated species.

Other Contributions to Ocular Surface Epithelial Defense

In summary, the ocular surface and tears possess a wide range of chemicals and physical attributes that help prevent infection. Such redundancy is commonplace in biological systems, but is particularly important at the ocular surface where the inability to control and eliminate an infection can have dire consequences for visual function. Having multiple protective mechanisms is also necessary as pathogens are very adept at developing strategies to circumvent host defenses. While countering one specific mechanism is relatively easily achieved, developing multiple strategies is rather more challenging.

As noted earlier, the outermost superficial cells of both the cornea and conjunctiva are coated in a matrix of carbohydrate referred to as the glycocalyx. By projecting from the epithelial cell surface, membrane spanning mucins MUC 1, 4, and 16 of the glycocalyx physically prevent pathogens from reaching the cell membrane. Also, some organisms are repulsed by negatively charged glycosaminoglycans present on the mucin. The ocular surface epithelial cells also produce a variety of cytokines and chemokines that are important in protection from microbial invasion. IL-1 is an important cytokine released in response to trauma and injury and among a plethora of activities serves to regulate production of other molecules such as IL-6, growth regulated oncogene (GRO)-a, -b, -g, TNF-a, and IL-8, which in turn modulate inflammatory and immune cell infiltration and activation. Dispersed between the epithelial cells of the cornea and conjunctiva are bone-marrow-derived dendritic cells called Langerhans cells. These cells are highly potent antigen-presenting cells which capture antigen and when mature present it to T lymphocytes in nearby secondary lymphoid tissues so activating adaptive immunity.

Concluding Remarks

Acknowledgments The author acknowledges grant support from NIH, NSF, and the State of Texas for her work on ocular surface antimicrobial peptides and thanks Kimberly Thompson of the University of Houston College of Optometry audiovisual department for drawing the figures. See also: Antigen-Presenting Cells in the Eye and Ocular Surface; Conjunctiva Immune Surveillance; Corneal Epithelium: Cell Biology and Basic Science; Corneal Nerves: Anatomy; Corneal Nerves: Function; Immunopathogenesis of HSV Keratitis; Immunopathogenesis of

Defense Mechanisms of Tears and Ocular Surface Pseudomonas Keratitis; Overview of Electrolyte and Fluid Transport Across the Conjunctiva; Pathogenesis of Fungal Keratitis.

Further Reading Chen, G., Shaw, M. H., Kim, Y. G., and Nunez, G. (2009). Nod-like receptors: Role in innate immunity and inflammatory disease. Annual Review of Pathology 4: 365–398. Evans, D. J., McNamara, N. A., and Fleiszig, S. M. J. (2007). Life at the front: Dissecting bacterial–host interactions at the ocular surface. Ocular Surface 5: 213–227. Flanagan, J. L. and Willcox, M. D. P. (2009). Role of lactoferrin in the tear film. Biochimie 91: 35–43. Fleming, A. (1922). On a remarkable bacteriolytic element found in tissues and secretions. Proceedings of the Royal Society Series B 93: 306–317.

451

Gupta, G. and Surolia, A. (2007). Collectins: Sentinels of innate immunity. BioEssays 29: 452–464. Hamrah, P. and Dana, M. R. (2007). Corneal antigen-presenting cells. Chemical Immunology and Allergy 92: 58–70. Hazlett, L. D. (2007). Bacterial infections of the cornea (Pseudomonas aeruginosa). Chemical Immunology and Allergy 92: 185–194. McDermott, A. M. (2009). The role of antimicrobial peptides at the ocular surface. Ophthalmic Research 41: 60–75. Paulsen, F. P. and Berry, M. S. (2006). Mucins and TFF peptides of the tear film and lacrimal apparatus. Progress in Histochemistry and Cytochemistry 41: 1–53. Sack, R. A., Nunes, I., Beaton, A., and Morris, C. (2001). Host-defense mechanisms of the ocular surfaces. Bioscience Reports 21: 463–480. Shafer, W. M. (ed.) (2006). Antimicrobial Peptides and Human Disease. Berlin: Springer. Yu, F.-S. X. and Hazlett, L. D. (2006). Toll-like receptors and the eye. Investigative Ophthalmology and Visual Science 47: 1255–1263.

Corneal Epithelium: Response to Infection Elizabeth A Szliter-Berger and L D Hazlett, Wayne State University School of Medicine, Detroit, MI, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Caspases – Humans express 11 different cysteineaspartic acid proteases classified as either initiator or effector caspases, which play essential roles in apoptosis, necrosis, and inflammation. CD44 – A type I transmembrane glycoprotein that regulates conformational changes of integrin heterodimers and their ability to microcluster and anchor to the actin cytoskeleton. Desmosomes – A junctional complex of adhesion molecules and linking proteins in the plasma membrane for cell-to-cell adhesion and contributes to the structural integrity through linkage of keratin cytoskeletal filaments of adjoining cells. Epidermal growth factor (EGF) – Expressed by corneal epithelial and stromal cells, it promotes cell migration and the attachment of corneal epithelial cells to fibronectin. Hemidesmosomes – Integral membrane protein complexes of integrin hetrodimers in the basal cell plasma membrane anchoring cells to the extracellular matrix. Langerhans cells – Unique subset of dendritic cells located in mucosal stratified squamous epithelium and skin epidermis; as professional antigenpresenting cells, they express toll-like receptors (TLRs) as well as C-type lectin receptors. Melanocytes – Pigment-producing cells located within the uvea of the eye. Nuclear factor-kappa B (NF-kB) – A protein complex that functions as a transcription factor; NF-kB is found in almost all animal cell types and is involved in cellular responses to stimuli such as stress, cytokines, free radicals, ultraviolet irradiation, and bacterial or viral antigens. Pathogen-associated molecular patterns (PAMPs) – Small, highly conserved molecular motifs present in bacteria, yeast, or viruses but not in mammalian cells and induce inflammatory signaling via interaction with receptors families, such as TLRs. Tritiated thymidine-labeling – Technique used to label cells actively undergoing DNA synthesis; it estimates the proportion of S-phase cells in a cell population.

452

Trophic – It refers to a nutrition-derived function. Zonula occludens – Also known as tight junctions, these structures are primarily composed of occludins and claudins to form a virtual impermeable barrier between adjacent cell membranes.

Introduction Basic Structure and Function The corneal epithelium is a remarkably proficient tissue that combines structure and function to serve as the major refractive component of the eye and maintain ocular surface integrity; yet provide a critical barrier protecting the visual axis from the external environment. The corneal epithelium is the outermost layer of the cornea and is comprised of stratified, nonkeratinized, nonsecretory, squamous epithelial cells, intermixed with Langerhans’ cells and occasional dendritic melanocytes. This tissue layer is a highly organized structure that is avascular and almost perfectly transparent in order to preserve the optical properties of the cornea. It is 5–7 cell layers thick and contains three cell types (posterior to anterior): basal cells, wing cells, and superficial cells. The deepest layer is comprised of cuboidal basal cells. This single cell layer of progenitor cells undergoes mitosis at a rate of 10–15% per day, followed by intermediate differentiation of daughter cells into one to three layers of wing cells as they migrate toward the surface. Superficial to wing cells is a three- to four-cell thick layer of terminally differentiated squamous cells. These cells constantly degenerate and desquamate from the corneal surface in a continuous cycle of shedding of superficial cells and proliferation of cells in the basal layer resulting in complete renewal of the epithelium every 7 days, as has been demonstrated by mitotic rate measurements in basal cells and by tracking the migration of tritiated thymidine-labeled cells to the ocular surface. This regenerative process is further maintained by a constant renewal of basal cells from the limbal epithelium via stem cells located in the limbus that differentiate into basal cells, followed by centripetal migration into the cornea. Although the corneal epithelium is constructed as a highly effective, semipermeable membrane on the ocular surface,

Corneal Epithelium: Response to Infection

it is also well equipped to participate in the host response to invading pathogens and infection as described in greater detail below.

Corneal Infection The corneal epithelium, similar to other mucosal epithelial linings in the body, constitutes the eye’s first line of defense against microbial pathogens. The cornea has an immuneprivileged status, which includes features such as avascularity of the cornea and dearth of antigen-presenting cells (APCs) in the central region of the cornea in order to protect the visual axis. Although resident dendritic cells or Langerhans cells, known for their global antigen-presenting properties, are present in the corneal periphery, they can readily be recruited into the central cornea when necessary. In addition, the corneal epithelium plays an active role in host defense against invading pathogens. Cells of the corneal epithelium: (1) recognize pathogens and their byproducts and, as a result, (2) respond through expression and secretion of a network of proinflammatory cytokines and chemokines that recruit inflammatory cells into the cornea, (3) secrete antimicrobial products, and (4) promote wound healing and restoration of tissue homeostasis. Keratitis is a condition of corneal inflammation, which can be caused by a number of bacterial, viral, and fungal pathogens. Clinically, it is associated with symptoms such as redness, tearing, reduced visual clarity, corneal discharge, and severe pain. Bacterial keratitis is the leading cause of corneal infection. Staphylococci are the most commonly occurring organisms in bacterial conjunctivitis and keratitis, including Staphylococcus (S.) aureus, S. pneumoniae, S. intermedius, and a-hemolytic streptococcus. Pseudomonas aeruginosa (P. aeruginosa) is most frequently encountered in keratitis cases associated with extended contact lens wear and constitute 19–42% of bacterial keratitis cases. Other bacteria known to cause keratitis include Escherichia coli and Morganella morganii. Common pathogens associated with viral keratitis include, yet are not limited to, herpes simplex virus (HSV) and adenovirus. HSV infection is the most common cause of corneal blindness in the United States at present time. Approximately 400, 000 people in the United States have been infected with ocular herpes and 50, 000 initial and recurring cases of HSV keratitis are diagnosed annually. Fusarium, Aspergillus (both filamentary fungi), and Candida albicans (a yeast) constitute those fungi associated with the majority of fungal keratitis cases. Sterile keratitis also incites an immune response from the corneal epithelium; however, instead of bacteria adhering to and invading the

453

ocular surface (requisite for bacterial keratitis), the epithelium responds to the presence of bacterial endotoxin (lipopolysaccharide or LPS). Sterile keratitis can also occur when bacteria colonize contact lenses and subsequently, release endotoxin onto the corneal surface, a condition known as CLARE or contact lens-induced acute red eye. The effects of both the invading pathogen on corneal tissue and the host immune response to the pathogen lead to many structural alterations that are otherwise essential to maintaining transparency of the cornea and, as a result, directly compromise visual acuity. As such, components of the corneal epithelium have evolved to respond to infection and these are described in greater detail below.

TLRs and TLR-Related Molecules Toll-like receptors (TLRs) are evolutionarily conserved type I transmembrane protein receptors that are expressed by corneal epithelial cells (as well as inflammatory cells) and function to respond to pathogens as the corneal epithelium’s first line of defense. These receptors initiate innate immunity and are essential for host defense against infection. TLRs recognize a broad spectrum of pathogen-associated molecular patterns (PAMPs), ranging from LPS (predominately recognized by TLR4), flagellin (TLR5), peptidoglycan (TLR2), single- or double-stranded RNA (TLR3,-7,-8), and unmethylated CpG DNA (TLR9). TLRs signal through several adaptor molecules, including the common adaptor protein myeloid differentiation factor (MyD)88 and MD-2; however, the reader is referred to Further Reading section for more information. It has been proposed that corneal epithelial cells play a central role in regulation of inflammatory responses by expression of functional TLRs and adaptor molecules. Epithelial cells are thought to intrinsically respond to the presence of pathogens through TLR recognition of PAMPs. Of the 13 TLRs identified to date, human corneal epithelial cells have been shown to express TLR-1,-2,-4,-5,-6, and -9 either intercellularly or at the cell surface (and has yet to be fully elucidated). TLRs primarily associate as homodimers, with exceptions for TLR-1,-2, and -6, which form heterodimers. Upon recognition of PAMPs expressed by invading pathogens on the ocular surface, TLRs produce downstream signaling events which induce translocation of nuclear factor kappa B (NF-kB), a major transcription factor of numerous genes important in both innate and acquired immune responses. These genes include proinflammatory cytokines and chemokines, leading to activation of adhesion molecules, and subsequently resulting in macrophage and polymorphonuclear neutrophilic leukocyte (PMN) recruitment into the

454

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

cornea. Antimicrobial peptide expression by corneal epithelial cells is induced by TLR activation, as well. Although there is a common pathway for TLR activation (the myeloid differentiation protein–IL-1 receptor-associated kinase–tumor necrosis factor receptor-associated kinase (MyD88– IRAK–TRAF) pathway), individual TLRs most likely activate different, alternative, signaling pathways as well, and these remain under investigation (IL, interleukin).

TGF-b1, in particular, is a potent anti-inflammatory cytokine that modulates lymphocyte activation and promotes wound healing; regarding the cornea, it is further known to promote proliferation and lamellar differentiation of corneal epithelial cells via keratocyte-mediated stimulation.

Antimicrobial Molecules Cytokines/Chemokines The immune response of the corneal epithelium utilizes a variety of different substances to mount an attack upon organisms invading the ocular surface. Cytokines and chemokines are produced and endogenously released by corneal epithelial cells to directly and indirectly recruit and activate immune cells of both innate and adaptive immunity. Gram-negative bacterial endotoxin such as LPS, for example, elicit production of endogenous proinflammatory cytokines such as tumor necrosis factor (TNF)-a, interleukin (IL)-6, and IL-1. Protein levels for both IL-1a and IL-1b have been shown to be constitutively expressed in the normal human cornea. It has further been demonstrated that upon rupture of the epithelial cell membrane by infectious agents or trauma, IL-1a is passively released, contributing to increased vascular permeability, macrophage and lymphocyte infiltration and activation, angiogenesis, as well as regulatory effects on corneal fibroblasts. If left unchecked, these events would destroy the cornea; however, the epithelium also secretes both soluble and membrane-bound forms of IL-1RII, a receptor and natural antagonist of IL-1 in order to modulate the effects of this potent proinflammatory cytokine and consequently preserve visual function. In addition to IL-1, in vitro studies using primary human corneal epithelial cells (HCECs) and telomeraseimmortalized HCECs have revealed expression and secretion of IL-6, IL-8, and TNF-a following either P. aeruginosa or HSV-1 challenge. TNF-a is a major proinflammatory mediator that is known to promote apoptosis, inflammation, and regulate immune cells. IL-6 is both a pro- and anti-inflammatory cytokine that can potentiate the inflammatory response, yet also modulate inflammation through its inhibitory effects on TNF-a and IL-1, while activating IL-1 receptor antagonist and IL-10. IL-8 is a chemokine that is secreted by the epithelium under stress and functions as a strong chemoattractant for neutrophils (PMN); MIP-3a, also released by corneal epithelial cells after infection, promotes directed migration of leukocytes, such as immature dendritic cells and effector T cells. TGF-a,-b1, and -b2 are expressed also by corneal epithelial cells and contribute to re-epithelization, an essential step in resuming normal corneal function after infection.

The corneal epithelium also employs the action of small (100 amino acids or less), positively charged (arginineand lysine-rich) molecules known as antimicrobial peptides to further assist in combating invading pathogens. Over 500 naturally occurring antimicrobial peptides have been identified in mammals. Typically these molecules are cationic polypeptides that disrupt bacterial membranes through charge interactions and hydrophobic amino acids. Many antimicrobial peptides and their microbicidal effects are induced locally by inflammatory stimuli at the site of infection and act synergistically with other anti-inflammatory mechanisms (cytokines, inflammatory cells) in defending against microbial pathogenesis. Of the four distinct structural classes of antimicrobial molecules, recent studies have shown that corneal epithelial cells secrete peptides from the defensin and cathelicidin families, which are thought to help protect the eye through broad spectrum activity against microorganisms including Gram-positive and -negative bacteria, fungi, and certain enveloped viruses. These molecules and their roles in preventing microbial invasion and managing infection are discussed in greater detail below. b-Defensins The defensins include a- and b-defensin subfamilies, all of which are characterized by a b-sheet-rich fold and three disulfide bridges. A third class of y-defensins has been recently identified in rhesus macaque leukocytes. Although y-defensin mRNA is detectable in humans, these transcripts contain a premature stop codon preventing translation of functional protein. Leukocytes (PMN) and various types of epithelial cells have been shown to express both a- and b-defensins; however, the corneal epithelium has been demonstrated to produce and secrete only members of the b-defensin subfamily. Human b-defensins (hBDs) include 28 members, of which three have been associated with the corneal epithelium. Human b-defensin-1 (hBD-1) is constitutively expressed by corneal epithelial cells; while studies have shown that hBD-2 is induced by bacterial infection and bacterial products such as lipoprotein and lipotechoic acid, and is mediated by TLR2. In addition, cytokines TNF-a and IL-1 upregulate hBD-2 expression by corneal epithelial cells. Expression of hBD-3 is more variable,

Corneal Epithelium: Response to Infection

whereby some studies have indicated solely constitutive expression by corneal epithelial cells; and others have demonstrated inducible expression by TNF-a and interferon-gamma (IFN-g). Regardless, hBD-3 was shown by McDermott and colleagues to exert most potent antibacterial activity against S. epidermidis, S. aureus, and P. aeruginosa using in vitro antimicrobial assays, followed by hBD-2; while hBD-1 showed moderate activity against Pseudomonas, but no activity against Staphylococcus strains. In vivo, it was recently demonstrated that murine (m) BD2, but not mBD1, was protective in the P. aeruginosa-infected mouse cornea. The importance of these antimicrobial peptides has been demonstrated by knocking out the mouse b-defensin-1 gene, which led to less effective clearance of Haemophilus influenzae from the lung and increased colonization of Staphylococcus in the bladder. Cathelicidins Of the numerous members of the cathelicidin family, only LL-37 has been described in humans. Cathelicidins express a highly conserved cathelin domain and a less conserved, more variable antimicrobial region. LL-37, as its name suggests, is a 37-amino acid linear peptide expressed by inflammatory/immune cells and epithelial tissue. It is derived by cleavage of the precursor, human cationic antimicrobial protein 18 (hCAP18) and appears to function as both an antibacterial peptide and immunomodulator. Low expression levels of LL-37 are detected constitutively and subsequently upregulated following injury, infection, and exposure to IL-1b, as well as heatkilled P. aeruginosa. Regarding antimicrobial properties, LL-37 is thought to function in a similar manner to that described for defensins and work synergistically with corneal epithelial proteins such as defensins, lactoferrin, and lysozyme (latter two present in tear film). LL-37 is able to bind and neutralize LPS and lipotechoic acid, thus reducing the inflammatory response associated with these molecules. In addition to eradicating ocular pathogens, studies have demonstrated that LL-37 enhances the innate and adaptive immune response in the corneal epithelium through modulation of cytokine/chemokine expression. Using an in vitro stimulation assay, LL-37 was shown to induce production of IL-1b, IL-6, IL-8, and TNF-a by human corneal epithelial cells. Furthermore, this molecule promotes wound repair through enhanced cell migration, including fibronectin-induced migration by stimulating corneal epithelial cells. In addition to microbicidal activities of b-defensins and LL-37, these molecules also wield effects over immune cells and influence wound healing. They have been demonstrated to recruit and activate immune cells through the induction of cytokine and chemokine production by epithelial cells, which further actuate the cellular components of the immune response to corneal infection. Regarding

455

postinfection, both hBD-2 and LL-37 have been shown to be upregulated in an in vitro organ culture model of corneal epithelial wound healing suggesting roles for corneal epithelial cell migration and proliferation.

Complement System The complement system also contributes to the first line of innate immune defense against corneal infection. This critical system is composed of a series of effector and regulatory proteins that sequentially activate one another to generate biologically active molecules, such as opsonins and chemotaxins. The complement system is continuously activated at low levels in the eye under normal conditions, as supported by detection of soluble and membraneassociated complement regulatory proteins (e.g., rsCD59), which are also strongly expressed in the corneal epithelium for tight regulation of aberrant activation. Components of complement are also more heavily distributed in the peripheral versus the central cornea, potentially due to the diffusion of complement molecules from corneal limbal vessels. In response to infection activation of complement can occur via both the classical and alternate pathways.

Secretory IgA Although this molecule is not essential in ocular defense, secretory IgA (sIgA) does play a major role in the prevention of some corneal infections, including Pseudomonas and Acanthamoeba. In fact, over 75% of the general population contain anti-P. aeruginosa-specific IgAs in their tear film. sIgA protects the corneal epithelium by accumulating in the ocular mucin layer and displays an antigen–antibody clearance function. Aggregated sIgA opsonizes bacteria for PMN phagocytosis and processing via recognition by sIgA receptors expressed on the immune cell surface. IL-8, which is secreted by corneal epithelial cells during infection, further enhances the ability of sIgA to induce release of reactive oxygen species by PMN.

Adhesion Molecules Under normal conditions, integrity of the corneal epithelium depends upon a number of factors, including adhesion molecules. Corneal epithelial cells are interdigitated, particularly in the middle layers, and largely interconnected by desmosomes. Basal cells are firmly attached to the basement membrane, neighboring basal cells and overlying wing cells via hemidesmosomes. Tight junctional complexes, or zonula occludens, found only between superficial cells are

456

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

of great importance regarding the barrier function of the corneal epithelium. As such, superficial cells form a highly effective, semipermeable membrane on the corneal surface. Not only does the tight junctional barrier prevent decreases in net fluid transport out from the stroma, but it also prevents corneal penetration by pathogens. In addition, gap junctions connect the cells of all layers of the corneal epithelium and function as communication conduits between cells. These tight anatomical arrangements with virtual absence of intercellular spaces contribute to the remarkable transparency of the epithelium, but have deleterious effects when breached after infection (or wounding) and must be restored forthwith in order to restore/preserve visual function. Upon infection of the corneal epithelium, breakdown of tight junctional integrity occurs due to the loss or disruption of the outermost layers of the epithelium by invading pathogens. This results in a collapse of cell membrane permeability and selectivity, in addition to creating an unrestricted portal for further pathogen intrusion into the cornea until the corneal epithelium and its adhesion molecules are regenerated. Additional adhesion molecules such as selectins, intercellular adhesion molecule-1 (ICAM-1) and vascular cell adhesion molecule-1 (VCAM-1), are modified after infection due to the secretion of epithelial-derived cytokines, such as TNF-a. As a result, these barriers are further breached to allow the transmigration of inflammatory cells into the injured or infected tissue site. Recruitment of activated leukocytes to sites of infection is essential to the function of both inflammation and innate immunity. However, the extent of leukocyte recruitment contributes largely to the intensity of the inflammatory response, and if this is not well balanced it can result in considerable tissue damage. However, the corneal epithelium does play an active role in regenerating and repairing itself after infection. CD44 is among the molecules that corneal epithelial cells express to mediate wound healing after infection. This molecule is thought to be involved in cell–cell interactions that provide adhesive strength for the epithelial sheet and in cell–matrix interactions that occur during cell migration and the re-epithelialization process. Basal cells of the corneal epithelium have been demonstrated to secrete APLP2, an amyloid precursor-like protein that is suggested to influence reorganization of the extracellular matrix and dynamic cell–matrix adhesion during re-epithelialization. It has been demonstrated that the corneal epithelium (as well as stromal and endothelial cells) produces and secretes epithelial growth factor (EGF), which is thought to promote cell migration and mitosis of epithelial cells. Corneal epithelial cells then resume expression of molecules such as connexins 45/43 and a6b4 integrin, which participate in the formation of gap junctions and hemidesmosomes/desmosomes, respectively, thus restoring homeostasis of an intact corneal epithelium.

Neuropeptides Innervation of the Corneal Epithelium The cornea is among the most densely innervated tissues in the body. In regards to the corneal epithelium, the subbasal epithelial nerve plexus originates from the ophthalmic division of the trigeminal nerve via the anterior ciliary nerves and provides innervation to the basal epithelial cell layer and terminates within the superficial epithelial layers. These nerve fibers are predominately sensory (most classified as nociceptors) and serve a protective role, typically responding to mechanical, thermal, and chemical stimuli. They are stimulated during corneal abrasions and ulcers, and are extremely painful. It is estimated that single corneal sensory neurons support approximately 200 and 3000 individual nerve endings in the mouse and rabbit, respectively, demonstrating the high density of innervation in the corneal epithelium. Corneal nerves in the epithelium have a trophic function, as well. In addition, neuropeptides have been associated with corneal nerves and include: substance P (SP), calcitonin gene-related peptide (CGRP) and vasoactive intestinal peptide (VIP), pituitary adenylate cyclase activating peptide (PACAP), and neuropeptide Y (NPY) among others. Nerve damage can lead to transient or chronic neurotrophic deficits following infection of the corneal epithelium. Corneal denervation also significantly impairs the ability of the corneal epithelium to heal itself and predisposes newly healed corneas to periodic, spontaneous epithelial erosions. Although the nerves coursing through the human corneal epithelium are known to produce a variety of neuropeptides, the following sections focus on the importance of two of these molecules: SP and VIP.

Substance P SP is an 11-amino acid neuropeptide that is largely associated with proinflammatory events during corneal infection, including upregulation of TLR4 and TLR9 mRNA and of cytokines/chemokines IL-1b, TNF-a, IFN-g, and MIP-2 as demonstrated in a murine model of P. aeruginosa-induced keratitis. Physiologically relevant concentrations of SP are present in the normal human (and mouse) cornea and corneal epithelial cells express the NK1 receptor, which is the major physiological receptor for SP. The neuropeptide also has been associated with wound healing properties and together with insulin-like growth factor-1 (IGF-1) has been demonstrated to have a stimulatory effect on corneal epithelial cell migration, adhesion, and wound closure.

Vasoactive intestinal peptide VIP is a 28-amino acid neuropeptide that exerts immunomodulatory properties in the cornea following infection.

Corneal Epithelium: Response to Infection

It has been demonstrated in a murine model of P. aeruginosainduced ocular infection that VIP downregulates the production of several proinflammatory cytokines, including: TNF-a, IL-1, IL-6, IL-12, and IFN-g, while stimulating the production of anti-inflammatory cytokines IL-1 receptor antagonist, IL-10, and TGF-b. Macrophage and PMN activation was also shown to be influenced by the presence of VIP. Studies in the mouse further support a role for VIP in wound healing and restoration of immune homeostasis in the cornea; however, it has yet to be determined the extent of which is carried out by the epithelium versus stroma.

Thymosin-b4 As previously stated, after eradication of the invading pathogen, the corneal epithelium must heal and promptly resume normal activity. Thymosin (T)b-4 is a 43-amino acid protein produced by corneal epithelial cells that contributes to the resurfacing of the epithelium and regeneration of cell adhesion molecules to reconstitute the epithelial barrier. Studies have shown that Tb-4 levels are increased in murine corneas during re-epithelialization and also significantly enhance the migration of human corneal epithelial cells through upregulation of molecules associated with cell migration, including laminin-5 and matrix metalloproteinase 1 (MMP-1). Furthermore, Tb-4 modulates corneal cytokine production in an anti-inflammatory manner by downregulating levels of MIP-2 and TNF-a, as demonstrated in the murine cornea. Tb-4 has been demonstrated to inhibit apoptosis of human corneal epithelial cells through inhibition of caspases and suppression of Bcl-2

457

(an antiapoptotic factor) release from mitochondria. Lastly, this molecule has been demonstrated to modulate matrix metalloproteinase expression and prevent PMN infiltration in an alkali injured mouse cornea model, further supporting a key role for Tb-4 in the repair and remodeling of the injured cornea.

Conclusion The cornea generates approximately 80% of the eye’s refractive power; therefore, it is imperative that the corneal epithelium possess the ability to effectively and appropriately activate an innate immune response when pathogens are encountered on the ocular surface. The protective mechanisms employed by the epithelium have evolved to balance recognition and elimination of pathogens while limiting corneal damage and preserving the visual axis. In order to do so, the corneal epithelium is able to recognize invading organisms through TLR signaling, generate a network of cytokines and chemokines to recruit and activate host inflammatory cells, produce antimicrobial molecules, and provide activation signals to the complement system, all of which coalesce into an effective and efficient immune response as depicted schematically in Figure 1. The corneal epithelium also possesses the ability to regenerate itself and promote wound healing through expression and secretion of adhesion molecules and Tb-4; this process is enhanced by the presence of corneal nerve fibers that release neuropeptides such as SP and VIP, molecules that further contribute to the healing process and restoration of tissue homeostasis.

LPS Bacterium

Cytokines/ chemokines

slgA TLR c9/rsCD59 c9/GPI anchor Antimicrobial peptides

Tear film

Epithelium

Stroma

Nerve

Figure 1 The corneal epithelium also possesses the ability to regenerate itself and promote wound healing through expression and secretion of adhesion molecules and Tb-4; this process is enhanced by the presence of corneal nerve fibers that release neuropeptides such as SP and VIP, molecules that further contribute to the healing process and restoration of tissue homeostasis.

458

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

See also: Immunopathogenesis of Pseudomonas Keratitis.

Further Reading Akpek, E. K. and Gottsch, J. D. (2003). Immune defense at the ocular surface. Eye 17: 949–956. Bora, N. S., Jha, P., and Bora, P. S. (2008). The role of complement in ocular pathology. Seminars in Immunopathology 30: 85–90. Dohlman, C. H. (1971). The function of the corneal epithelium in health and disease. Investigative Ophthalmology 10(6): 383–407. Edelhauser, H. F. and Ubels, J. L. (2003). Adler’s Physiology of the Eye: The Cornea and the Sclera. St. Louis, MO: Mosby. Ganz, T. (2003). Defensins: Antimicrobial peptides of innate immunity. Nature Reviews Immunology 3: 710–720. Kumar, A., Zhang, J., and Yu, F. X. (2006). Toll-like receptor 2-mediated expression of b-defensins-2 in human corneal epithelial cells. Microbes and Infection 8: 380–389. Lu, L., Reinach, P. S., and Kao, W. W. (2001). Corneal epithelial wound healing. Experimental Biology and Medicine 226(7): 653–664. McDermott, A. M. (2004). Defensins and other antimicrobial peptides at the ocular surface. Ocular Surface 2(4): 229–247. McDermott, A. M. (2007). Ocular surface expression and in vitro activity of antimicrobial peptides. Current Eye Research 32(7–8): 595–609. Muller, L. J., Marfurt, C. F., Kruse, F., and Tervo, T. M. (2003). Corneal nerves: Structure, contents and function. Experimental Eye Research 76(5): 521–542.

Sack, R. A., Nunes, I., Beaton, A., and Morris, C. (2001). Host-defense mechanism of the ocular surfaces. Bioscience Reports 21(4): 463–480. Sosne, G., Qiu, P., and Kurpakus-Wheater, M. (2007). Thymosin b-4 and the eye: I can see clearly now the pain is gone. Annals of the New York Academy of Science 1112: 114–122. Szliter, E. A., Lighvani, S., Barrett, R. P., and Hazlett, L. D. (2007). Vasoactive intestinal peptide balances pro- and anti-inflammatory cytokines in the Pseudomonas aeruginosa-infected cornea and protects against corneal perforation. Journal of Immunology 178(2): 1105–1114. Uematsu, S. and Akira, S. (2008). Handbook of Experimental Pharmacology: Toll-like Receptors (TLRs) and Innate Immunity: Toll-Like Receptors (TLRs) and Their Ligands. Berlin: Springer. Wilson, S. E., Liu, J. J., and Mohan, R. R. (1999). Stromal–epithelial interactions in the cornea. Progress in Retinal and Eye Research 18 (3): 293–309. Zhang, J., Wu, X., and Yu, F. X. (2005). Inflammatory responses of corneal epithelial cells to Pseudomonas aeruginosa infection. Current Eye Research 30: 527–534.

Relevant Websites http://www.nei.nih.gov – Facts about the Cornea and Corneal Disease, National Eye Institute. http://www.revoptom.com – Handbook of Ocular Disease Management: Cornea.

Inflammation of the Conjunctiva T Nishida, Yamaguchi University Graduate School of Medicine, Yamaguchi, Japan ã 2010 Elsevier Ltd. All rights reserved.

Glossary Cell-adhesion molecule – Cell-adhesion molecules are cell surface proteins, usually glycoproteins, that mediate cell–cell adhesion. They play important roles in the assembly and maintenance of tissues, wound healing, morphogenic cellular movements, cell migration, and metastasis. Intercellular adhesion molecule-1 (ICAM-1) functions in leukocyte adhesion and inflammation. Its expression is induced in various cell types by interferon-g (IFN-g), and it mediates interactions with neutrophils in inflamed tissue. Vascular cell-adhesion molecule-1 (VCAM-1) is presented on the surface of various cell types, including endothelial cells, tissue macrophages, fibroblasts, and dendritic cells. Its expression is induced by cytokines, and it plays a key role in the recruitment of eosinophils to sites of inflammation. Collagenase (microbial) – Microbial collagenase is a metalloproteinase produced by bacteria that degrades helical regions of native collagen to yield small protein fragments. The preferred cleavage site is immediately upstream of the glycine residue in the sequence – proline–X–glycine–proline – where X is any amino acid. Six forms (or two classes) of microbial collagenase have been isolated from Clostridium histolyticum; these proteins are immunologically cross-reactive but possess different amino acid sequences and different specificities. Other variants have been isolated from Bacillus cereus, Empedobacter collagenolyticum, Pseudomonas marinoglutinosa, and species of Vibrio and Streptomyces. Helper T cell – Helper T cells constitute a specific subpopulation of CD4+ T cells that provides help to other cells of the immune system in mounting an immune response either through direct cell–cell interaction or the secretion of cytokines. They are also referred to as effector T cells. Several distinct subtypes of helper T cells, designated Th1, Th2, Th3, and Th17, have been identified. Matrix metalloproteinase – Matrix metalloproteinases (MMPs) constitute an important family of enzymes that regulate composition of the extracellular matrix. They are synthesized as inactive precursor proteins that consist of propeptide, catalytic, and hemopexin domains; proteolytic removal of the propeptide domain results in MMP activation. MMPs are zinc-dependent

endopeptidases that cleave one or several constituents of the extracellular matrix as well as nonmatrix proteins, and they play an important role in cleaving fibrillar collagen types I, II, and III into characteristic three-fourths and one-fourth fragments. Some MMPs are associated with the cell membrane, either through a transmembrane domain or through glycosylphosphatidylinositol anchor; such MMPs may act within the pericellular environment to influence cell migration. MMP-1, MMP-8, and MMP13 are also known as collagenase 1, collagenase 2 (neutrophil collagenase), and collagenase 3, respectively. Th1 cytokine – Th1 cytokines include interleukin (IL)-2, IFN-g, IL-12, and tumor necrosis factor-b. They are secreted by Th1 cells and play an important role in cell-mediated immunity and chronic inflammation. In general, Th1 responses are stimulated by intracellular pathogens, including viruses as well as certain mycobacteria, yeasts, and parasitic protozoans. Th2 cytokine – Th2 cytokines include IL-4, IL-5, IL-6, IL-10, and IL-13. They are secreted by Th2 cells and play a key role in the initiation of allergic responses. Th2 responses are also elicited by free-living bacteria and other parasites.

Inflammation Inflammation is a biological response of the living body to injury or other harmful insults including microbial pathogens, allergens, and physical or chemical agents. It serves to protect the body and is the precursor to wound healing. Classical signs and symptoms of inflammation include redness, swelling, heat, pain, and loss of tissue function. Thus, although inflammatory reactions are well regulated to maintain homeostasis of the body and to promote wound repair, they may result in bodily discomfort. In some instances, however, excessive inflammation may result in tissue damage. Classically, inflammation has been considered to begin with a reaction of vascular tissue that renders vessels permeable to blood cells at the site of injury, resulting in the extravasation of such cells. Recent advances in immunology and molecular cell biology have revealed the mechanisms of inflammation at the level of cellular

459

460

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

interactions and molecular networks. Allergens and infection by pathogens are the major pathological triggers for inflammation at the ocular surface.

The Conjunctiva and Cornea The ocular surface is composed of the cornea, conjunctiva, lacrimal glands, and associated lid structures. Both the conjunctiva and the cornea are derived from the embryonic epidermis, and they are separated from each other by tear fluid. External insults to the ocular surface evoke different types of inflammatory reactions in the conjunctiva and the cornea that are related to the anatomic differences and physiological roles of these two structures as well as to their connection via tear fluid (Figure 1). The conjunctiva is a semitransparent membrane that covers the surface of the eye from the back surface of the eyelids to the edge of the transparent cornea. It serves as a barrier at the surface of the eyeball and helps to protect against the invasion of biological, chemical, or physical agents without interrupting the free movement of the eyeball. The surface of the conjunctiva is covered by multiple layers of epithelial cells. The conjunctival epithelium is a relatively inefficient barrier, however, with the result that pathogens, allergens, and biologically active substances readily penetrate into the stroma of the conjunctiva. The conjunctival stroma consists of conjunctival fibroblasts, loosely packed collagen fibers, a vascular system, and abundant immune cells. The triggering of an inflammatory reaction by pathogens or allergens results in enlargement of the blood vessels of the conjunctiva and consequent increased blood flow (hyperemia). The associated

increase in the permeability of the conjunctival vascular system leads to leakage of liquid components and to the development of conjunctival edema. It also allows the infiltration of blood cells into the conjunctival stroma and the consequent activation of conjunctival fibroblasts. Like the conjunctiva, the cornea faces the external environment. However, unlike the conjunctiva, the cornea is transparent, and its surface must be maintained smooth for the proper transmission of light into the eye. The anatomic structure of the cornea is relatively simple compared with that of the conjunctiva. Its surface is also covered by multiple layers of epithelial cells, but the corneal epithelium provides a much tighter barrier than does the conjunctival epithelium. In the absence of any loss or dysfunction of its component cells, the corneal epithelium prevents the entry of pathogens and allergens into the corneal stroma. The cornea does not contain a vascular system. Although a small number of immune cells such as Langerhans cells, stromal dendritic cells, and macrophages are present in the cornea, its major cellular components are epithelial cells, stromal keratocytes (corneal fibroblasts), and endothelial cells. Both the conjunctiva and the cornea are innervated by the ophthalmic branch of the trigeminal nerve, but the cornea is the most sensitive tissue at the ocular surface, and indeed maybe in the entire body, as a result of the high density of sensory nerve endings in the corneal epithelium.

Tear Fluid: A Reservoir of Inflammatory Cells and Modulators Tear fluid functions as a lubricant between the tarsal conjunctiva and the surface of the cornea and serves to

Tear fluid

Conjunctiva

Simple structure Tight epithelial barrier Avascular tissue Few immunologic cells Faces the environment Wet tissue

Biological defense system

Extensive vascular system Loose epithelial barrier Abundant immune cells Faces the environment Wet tissue

Diseases

Neutrophils Eosinophils Lymphocytes Cytokines Chemokines MMPs, etc.

Cornea Anatomic characteristics

Infections (bacterial, viral, fungal, protosoal) Physical or chemical injuries Clinical problems Loss of transparency Irregular surface-scarring

Anatomic characteristics

Lubricant Pathway for inflammatory cells and bioactive substances

Diseases Allergic conjunctivitis Infections (bacterial, viral) Chemical injuries Clinical problems Swelling, itching Fibrosis due to scarring

Figure 1 Cornea–tear fluid–conjunctiva axis in ocular surface inflammation. MMP, matrix metalloproteinase.

Inflammation of the Conjunctiva

maintain the ocular surface wet. It is also important for ensuring the generation of a clear image on the retina. Moreover, it contributes to the biological defense system of the ocular surface, containing immunoglobulin, lactoferrin, lysozyme, and other protective proteins. With regard to inflammation at the ocular surface, tear fluid provides a pathway for the movement of inflammatory cells – such as neutrophils, eosinophils, and lymphocytes – between the conjunctiva and the cornea. It also serves as a reservoir of various inflammatory cytokines, chemokines, and growth factors as well as of nutrients and oxygen. Collagen-metabolizing enzymes such as matrix metalloproteinases (MMPs) are present in the tear fluid of individuals with certain ocular inflammatory conditions.

Allergic Reactions in the Conjunctiva The conjunctiva is a common site for allergic reactions (Figure 2). Clinical characteristics of conjunctival allergic disease include hyperemia, edema, the formation of papillary discharge, the development of corneal epithelial disorders, and, in some patients, corneal ulcer. Hyperemia and edema result from dilation and an increase in the

Allergy

permeability of the vascular system in the conjunctiva. Conjunctival fibroblasts are responsible for the formation of papillae. Mechanical injury caused by papillae as well as the effects of inflammatory cytokines, such as interleukin (IL)-4, IL-13, and tumor necrosis factor-alpha (TNF-a), are responsible for discharge and damage to the corneal epithelium. Disruption of corneal epithelial barrier-function results in the spread of inflammation to the cornea and the development of various types of corneal epithelial disorders. Corneal fibroblasts contribute to the pathology of corneal ulceration. The primary cells that mediate allergic reactions at the ocular surface include mast cells, vascular endothelial cells, eosinophils, T helper 2 (Th2) cells, and conjunctival fibroblasts, with corneal epithelial cells and corneal fibroblasts also contributing in some cases. Certain allergens that enter tear fluid from the environment are solubilized by the fluid and penetrate through the loose barrier provided by the conjunctival epithelium into the conjunctival stroma. In the stroma, the allergens trigger the secretion of histamine and inflammatory cytokines, such as IL-4, IL-13, TNF-a from mast cells, and IL-3 and IL-5 from Th2 cells. Histamine acts on the vascular endothelium to increase vessel

Histamine

Vascular endothelium

Mast cells Conjunctival stroma

Th2 cells Allergens IL-4 IL-13 TNF-α

461

Permeability

Hyperemia edema

IL-3 IL-5

Eosinophils

Conjunctival fibroblasts

Eotaxin

TARC

Collagen synthesis cell proliferation Papillae

Conjunctival epithelium (loose barrier) Tear fluid

Allergens

Mechanical injury

IL-4 IL-13 TNF-α

Corneal epithelium (tight barrier)

Eosinophils

TARC

Discharge

Epithelial disorders

Disruption of barrier function

IL-4 IL-13 TNF-α Corneal stroma

TARC

Eotaxin Eosinophils Ulcer

Corneal VCAM-1 fibroblasts

MMPs

Figure 2 Clinical characteristics of allergic reactions in the conjunctiva and the cornea. IL, interleukin; MMP, matrix metalloproteinase; Th2, T helper 2 cell; TARC, thymus and activation-regulated chemokine; TNF-a, tumor necrosis factor alpha; VCAM, vascular cell-adhesion molecule.

462

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

permeability and induce vessel enlargement, resulting in conjunctival hyperemia and edema. IL-4, IL-13, and TNF-a – released by mast cells – activate conjunctival fibroblasts and trigger their secretion of the chemokines eotaxin and thymus and activation-regulated chemokine (TARC). Eotaxin attracts eosinophils to the interstitial space, and the extravasated eosinophils are then activated by IL-3 and IL-5 released by Th2 cells. TARC attracts Th2 cells into the interstitial space, and these cells then serve as an additional source of IL-4, IL-13, and TNF-a. Exposure of conjunctival fibroblasts to IL-4, IL-13, and TNF-a also stimulates the synthesis of collagen and cell proliferation – effects that give rise to the formation of papillae. The protrusive shape of the papillae results in mechanical injury of both conjunctival and corneal epithelia; such injury together with the effects of IL-4, IL-13, and TNF-a that enter tear fluid from the conjunctiva lead to disruption of the barrier function of the corneal epithelium and to discharge. Eosinophils that enter tear fluid from the conjunctiva are then able to penetrate into the corneal stroma. IL-4, IL-13, and TNF-a also enter the corneal stroma from tear fluid and activate corneal fibroblasts to express TARC, eotaxin, and vascular cell-adhesion molecule-1 (VCAM-1) – a cell-adhesion molecule for eosinophils. The activated corneal fibroblasts also produce MMPs, which degrade collagen of the extracellular matrix in the corneal stroma, resulting in corneal ulceration. TARC released from corneal fibroblasts passes through tear fluid into the conjunctival stroma, where it further promotes the secretion of IL-4, IL-13, and TNF-a by Th2 cells in a vicious cycle. This scenario thus reveals that, although immune cells such as mast cells, Th2 cells, and eosinophils play a prominent role in allergic disorders at the ocular surface, resident fibroblasts in both the conjunctiva and cornea also contribute to the inflammatory process.

Infection of the Conjunctiva or Cornea The clinical characteristics of infection at the ocular surface include swelling, hyperemia at the conjunctiva, discharge, and epithelial defects and ulceration in the cornea. As with allergic reactions, the reactions of the conjunctiva and cornea to infection differ (Figure 3). The vascular system of the conjunctiva ensures a robust immune response to infection in this tissue, with conjunctivitis being a relatively mild clinical condition. However, the cornea is avascular and possesses few immune cells, with the result that corneal infection is more serious and may become sight threatening. If a pathogen survives the biological defense system in tear fluid, it readily penetrates the conjunctival epithelium and triggers the dilation and permeabilization of conjunctival blood vessels, resulting in swelling and

hyperemia. Neutrophils and Th1 cells enter the conjunctival stroma from the bloodstream and serve as the second line of defense against pathogens. Both neutrophils and Th1 cells secrete IL-1, with Th1 cells also secreting interferon-g (IFN-g). These cells and cytokines may be sufficient to inactivate the pathogen and limit the inflammatory response to the conjunctiva. However, pathogens also act on conjunctival epithelial cells to trigger the secretion of IL-8, IL-6, and TNF-a. These cytokines together with IL-1 and IFN-g can enter tear fluid and, in the presence of damage to the corneal epithelium, may penetrate into the corneal stroma and activate corneal fibroblasts. The tight barrier provided by the corneal epithelium normally prevents the entry of pathogens into the cornea. However, corneal epithelial injury can result in pathogen penetration into the corneal stroma. Pathogen-associated various factors such as lipopolysaccharide (LPS) of Gram-negative bacteria and peptidoglycan (PGN) of Gram-positive bacteria are recognized by toll-like receptors (TLRs) on the surface of corneal fibroblasts and trigger the production of IL-8 and the expression of intercellular adhesion molecule-1 (ICAM-1) by these cells. IL-8, IL-6, TNF-a, IFN-g, and IL-1 that enter the corneal stroma via tear fluid also induce IL-8 production by corneal fibroblasts. IL-8 then attracts neutrophils exuded (extravasated) from conjunctival blood vessels into the corneal stroma, and these cells interact with corneal fibroblasts via ICAM-1. IL-1 released from neutrophils further stimulates corneal fibroblasts. Corneal infection is associated with the production of two types of collagen-degrading enzymes: collagenase released from the pathogen and MMPs released from corneal fibroblasts. These enzymes destroy stromal collagen, eventually resulting in the development of corneal ulcer. Collagen destruction by MMPs released from activated corneal fibroblasts may continue even if the pathogen has been killed by antimicrobial treatment. Neutrophils were originally thought to destroy stromal collagen, but these cells were subsequently found to promote the production of MMPs by corneal fibroblasts rather than to degrade the collagen themselves. As with ocular allergy, corneal fibroblasts thus play a key role in the progression of the inflammatory response to corneal infection.

Tear Fluid as a Diagnostic Indicator of Inflammation The measurement of inflammatory cytokines or chemokines and the cellular components of tear fluid provides clinically important information on inflammation at the ocular surface. The presence of eosinophils in tear fluid thus confirms a diagnosis of allergic inflammation,

Inflammation of the Conjunctiva

463

Infection Vascular endothelium

Swelling

Permeability

Neutrophils

Conjunctival Stroma Pathogen

Hyperemia

Th1 cells IFN-γ IL-1

Conjunctival epithelium (loose barrier) Tear fluid

Injury

Epithelial cells

Pathogen

IL-8 IL-6 TNF-α

Corneal epithelium (tight barrier)

IFN-γ IL-1

Neutrophils

Epithelial defects

Disruption of barrier function

IL-8 IL-6 TNF-α

Discharge

IL-8

IFN-γ IL-1

Neutrophils

Pathogen Corneal Stroma

LPS PGN

Corneal fibroblasts

TLR

Ulcer

ICAM-1 IL-1

MMPs

Collagenases

Figure 3 Clinical characteristics of infection in the cornea and conjunctiva. ICAM, intracellular adhesion molecule; IFN-g, interferon-gamma; IL, interleukin; LPS, lipopolysaccharide; MMP, matrix metalloproteinase; PGN, peptidoglycan; Th2, T helper 2 cell; TARC, thymus and activation-regulated chemokine; TLR, toll-like receptor; TNF-a, tumor necrosis factor alpha.

whereas the presence of neutrophils is indicative of infectious inflammation. In addition to being of diagnostic value, the condition of the tear fluid can affect the progression of ocular surface inflammation. In individuals with dry eye, for example, the decrease in tear secretion and small volume of tear fluid may result in concentration of inflammatory cells and proteins. The condition of tear fluid should thus be taken into account in the treatment of patients with ocular surface inflammation.

Connection of the Conjunctiva and Cornea via Tear Fluid The surfaces of both the conjunctiva and the cornea are covered by epithelial cells. However, the biological responses of these two tissues to allergens or to pathogens differ markedly. The conjunctiva has a prominent vascular system and contains abundant immune cells, whereas the cornea is transparent and avascular and contains few immune cells. These anatomic differences between the

conjunctiva and cornea are reflected in the types of inflammatory condition that affect them. The conjunctiva is the principal target tissue for allergic reactions at the ocular surface, whereas the cornea is the main target for microbial infection or injury. The vascular system of the conjunctiva serves as a key source of immune cells in each of these conditions. The cornea is also affected by inflammatory reactions that occur in the conjunctiva, with the tear fluid that covers the surface of both the conjunctiva and the cornea serving as a conduit for the exchange of immune cells, cytokines, chemokines, and growth factors. The concept of inflammation was first described more than 2000 years ago as redness and swelling with heat and pain by Celsus. In the nineteenth century, the concept of loss of tissue function associated with inflammation was recognized. Recent advances in cell and molecular biology have revealed the cytokine and chemokine network that underlies inflammation. However, the availability of effective anti-inflammatory drugs other than steroids remains limited. Nonsteroidal antiallergic drugs have been developed and are effective for the treatment of allergic conjunctivitis. Nonsteroidal anti-inflammatory drugs (NSAIDs) are also effective in ameliorating inflammatory reactions.

464

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

However, anti-inflammatory agents that halt tissue destruction are needed. Further characterization of the bidirectional regulation of conjunctival and corneal resident cells via cytokines and chemokines, as well as immune cells, released into tear fluid may provide a basis for the development of new drugs effective for the treatment of inflammation at the ocular surface. See also: Adaptive Immune System and the Eye: Mucosal Immunity; Adaptive Immune System and the Eye: T Cell-Mediated Immunity; Antigen-Presenting Cells in the Eye and Ocular Surface; Conjunctiva Immune Surveillance; Conjunctival Goblet Cells; Defense Mechanisms of Tears and Ocular Surface; Dry Eye: An Immune-Based Inflammation; Immunopathogenesis of Pseudomonas Keratitis; Molecular and Cellular Mechanisms in Allergic Conjunctivitis; Ocular Mucins; Overview of Electrolyte and Fluid Transport Across the Conjunctiva.

Further Reading Hazlett, L. D. (2005). Role of innate and adaptive immunity in the pathogenesis of keratitis. Ocular Immunology and Inflammation 13: 133–138. Kolaczkowska, E., Chadzinska, M., and Plytyez, B. (2008). Basic concepts of inflammation – from pioneer studies until now. In: Romano, G. T. (ed.) Inflammation Research Perspectives, pp. 113–168. New York: Nova Science Publishers. Kumagai, N., Fukuda, K., Fujitsu, Y., Yamamoto, K., and Nishida, T. (2006). Role of structural cells of the cornea and conjunctiva in the pathogenesis of vernal keratoconjunctivitis. Progress in Retinal and Eye Research 25: 165–187. Kumar, V., Abbas, A. K., Fausto, N., and Mitchell, R. N. (2007). Robbins Basic Pathology, 8th edn., pp. 31–58. Philadelphia, PA: Saunders-Elsevier. Ley, K. (2001). History of inflammation research. In: Ley, K. (ed.) Physiology of Inflammation, pp. 1–10. New York: Oxford University Press. Pearlman, E., Johnson, A., Adhikary, G., et al. (2008). Toll-like receptors at the ocular surface. Ocular Surface 6: 108–116. Tuli, S. S., Schultz, G. S., and Downer, D. M. (2007). Science and strategy for preventing and managing corneal ulceration. Ocular Surface 5: 23–39.

Concept of Angiogenic Privilege B Regenfuss and C Cursiefen, Friedrich-Alexander University Erlangen-Nuernberg, Erlangen, Germany ã 2010 Elsevier Ltd. All rights reserved.

Glossary Angioblast – Mesenchymal tissue that differentiates into blood cells and vascular endothelium. Angiogenesis – Formation of new blood vessels by outgrowth from preexisting vessels. Intussusception – Formation of new blood vessels by splitting of existing vasculature. Keratoplasty – Corneal transplantation. Vasculogenesis – De novo blood-vessel formation from endothelial progenitor cells.

Introduction: Angiogenesis and Lymphangiogenesis (Hem)angiogenesis describes the process of new bloodvessel formation by outgrowth from preexisting vessels. Accordingly, lymphangiogenesis means the formation of lymphatic vessels from preexisting ones. Both processes are precisely regulated and play an essential role in physiological and pathophysiological events in the organism. In the context of the eye, pathological new blood and lymphatic vessels are associated with numerous disorders reducing visual acuity. New blood-vessel formation in the organism is achieved either by angiogenesis or by vasculogenesis. Both processes can be distinguished from each other and strongly differ in the way the vessels arise. Vasculogenesis occurs mainly during embryogenesis and implies de novo blood-vessel formation by endothelial progenitor cells. During embryonic development, angioblasts, a subset of mesodermal cells, differentiate into endothelial cells and form the early vascular plexus. After establishing the primary vascular plexus, new blood vessels can be generated through angiogenesis that means by sprouting from preexisting blood vessels or by intussusception (nonsprouting angiogenesis). Angiogenesis and vasculogenesis normally occur during embryonic development. For the vascularization of the central nervous system (CNS) and the kidneys, angiogenesis seems to be the more important process. Following birth, most blood vessels remain in a quiescent state except for the once in the hair cycle, in the female reproductive system and during wound healing. In the case of unregulated angiogenesis, however, neovascularization can occur in the adult organism and usually is associated with

diseases such as arthritis, tumors, or corneal and retinal disorders. During early development of the retina, which is embryologically derived from the diencephalon, vasculogenesis and angiogenesis take place. In 1970, Ashton first described the mechanism of vasculogenesis for blood vessel formation in the retina of the human embryo. He proposed that primitive mesenchymal cells, after invading the retina, differentiate into endothelial cells, thereby, forming a capillary network. Nowadays there is evidence that vasculogenesis and angiogenesis both are responsible for the vascular development of the human fetal retina. Hughes and colleagues suggest a mechanism where vasculogenesis pioneers the establishment of a rudimentary vascular plexus, whereas angiogenesis provides further expansion of the vascular network and cares for increasing vessel density. Considering the fact that the retina is a highly metabolic active tissue both mechanisms complement one another and contribute to meet the metabolic requirements of the developing retina. The developed retina is a highly vascularized tissue that shows a dual blood supply. The inner layer of the retina is supported by the centralis retinae artery, originating from the arteria ophthalmica. The outer layer – especially the receptors – receive blood from the arteriae chorioideae. In general, the eye is an efficiently vascularized organ and shows a significantly higher blood circulation compared to other organs with the same volume; however, there are exceptions at the anterior pole of the eye being completely devoid of blood and lymph vessels. Whereas posterior structures like the retina, as mentioned earlier, show a strongly branched blood-vessel network, the sclera is relatively low vascularized and the adjacent cornea and the vitreous even are devoid of blood and lymphatic vessels. Keeping up corneal avascularity comprises an active process and needs the balance between angiogenic and anti-angiogenic factors. In this process, the cornea maintains the transparency even under inflammatory or other pro-angiogenic conditions by different molecular mechanisms which are not completely elucidated to date. This ability is called the corneal angiogenic privilege.

Corneal Angiogenic Privilege Corneal Avascularity The corneal angiogenic privilege normally prevents the ingrowths of new vessels in the cornea even under

465

466

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

inflammatory or angiogenic conditions, thus maintaining corneal transparency and preventing loss of vision. A study with corneas from stillborn patients showed that this privilege is already present during intrauterine development at least from gestational age of 17 weeks. Contrary to the conjunctiva, in which blood and lymphatic vessels are detectable, no vessels are found in the cornea. This leads to the presumption that due to an early expression of anti-angiogenic and anti-lymphangiogenic factors, the cornea is primarily devoid of blood and lymphatic vessels and not as a result of regression of already existing vessels. The occurrence of neovascularization by angiogenic activity is initiated by a disbalance of angiogenic and antiangiogenic factors caused by upregulating angiogenic molecules as well as by downregulating angiogenesisinhibiting molecules. Interestingly, homozygote TSP-1 or TSP-2-knockout mice and even TSP-1/2-knockout mice showed no spontaneous corneal angiogenesis. The deficiency of an important anti-angiogenic factor like TSP-1 or TSP-2 does not result in a breakdown of the angiogenic privilege. This leads to the conclusion that at least during embryonic development the angiogenic privilege seems to be redundantly regulated by several anti-angiogenic factors. However, secondary to severe inflammation and several other diseases the angiogenic privilege can be overcome and an initial parallel outgrowth of blood and lymphatic vessels occurs. Several factors comprising angiogenic and anti-angiogenic molecules, the cornea itself, and adjacent structures like the limbus are known to be involved in affecting the angiogenic privilege (see Figure 1). Deprivation of the angiogenic privilege can lead to corneal neovascularization and, in consequence, loss of vision.

Tear film

Angiogenic and Anti-Angiogenic Molecules Involved in Corneal Neovascularization Numerous angiogenic and anti-angiogenic molecules have been identified in the cornea over the last years. Angiogenic molecules include vascular endothelial growth factors (VEGFs), basic fibroblast growth factors (bFGFs), and matrix metalloproteinases (MMPs). Angiostatin, endostatin, thrombospondins, and pigment epithelium-derived factor (PEDF) are some of the anti-angiogenic molecules detected in the cornea. The regulation of angiogenesis is due to the interaction of pro-angiogenic molecules and angiogenesis inhibitors, where tilting the balance toward pro-angiogenic factors can lead to neovascularization. Lymphangiogenesis seems to proceed in a similar way as angiogenesis, and can be activated in the adult during inflammation, immune responses, or malignant processes. Stimuli like hypoxia, for example, in the context of wound healing, can also trigger the induction of hemangiogenesis via hypoxia-inducible transcription factor (HIF), a key transcriptional regulator for VEGF-A. In contrast, lymphangiogenic VEGF-C cannot be upregulated by hypoxia but only by proinflammatory cytokines. Vascular endothelial growth factors The VEGF growth factor family currently consists of five members, VEGF/VEGF-A, PIGF, VEGF-B, VEGF-C, and VEGF-D. The growth factors are recognized by different VEGF receptors, namely VEGFR-1, VEGFR-2, and VEGFR-3. VEGF-A originally isolated from a human histiocytic lymphoma cell line U937 is secreted

Decoy receptors

Limbal barrier

Anterior border IPAS Endogenous angiogenic inhibitors

Limbal stem cells Internal border Anterior chamber

Figure 1 Several strategies are used by the normal cornea to maintain corneal avascularity (corneal angiogenic privilege): The cornea possesses several defense lines against invading vessels: an anterior border at the epithelial basement membrane, an internal border at Descemet’s membrane, and the limbal barrier beneath the limbal epithelial stem cell niche. Several mechanisms contribute to maintain the angiogenic privilege of the cornea: (a) endogenous inhibitors of angiogenesis, (b) decoy receptors neutralizing angiogenic growth factors, (c) anti-angiogenic stem cells, and (d) anti-hypoxia-driven-angiogenesis agents. ã M. Vogler.

Concept of Angiogenic Privilege

in five different isoforms, generated by alternative splicing: VEGF115, VEGF121, VEGF165, VEGF189, and VEGF206. Vascular endothelial growth factor

VEGF-A shows numerous activities such as inducing endothelial cell proliferation and migration, proteolytic activity, and stimulating microvascular leakage – all of them promoting angiogenesis. VEGF-A mediates its function through receptors VEGFR-1 and VEGFR-2. Additionally, VEGF-A was reported to promote angiogenesis via an indirect pathway by upregulating NRP1, a neuronal receptor that has recently been shown to act as an isoform-specific receptor for VEGF165. VEGF-A can be released during hypoxia, in inflammatory situations, and during glucose deficiency. It was shown that the expression of VEGF-A is upregulated in inflamed and vascularized human corneas. In conclusion, VEGF-A seems to play an important role in inducing wound and inflammation-related corneal neovascularization. This was confirmed by the fact that corneal neovascularization could be suppressed after implantation of VEGF-A neutralizing antibodies in the corneal stroma of rats and rabbits. Whereas early data proposed that VEGF stimulates selectively hemangiogenesis but not lymhangiogenesis, recent data also suggest an (indirect) lymphangiogenic role: endogenous VEGF can promote lymphangiogenesisvia the recruitment of bone marrow-derived macrophages, releasing lymphangiogenic growth factors such as VEGF-C and -D. This broadens the impact of VEGF-A, not only for pathological hemangiogenesis, but also for lymphangiogenesis – at least in context of inflammation-induced neovascularization. A specific role for VEGF-A in the regulation of lymphangiogenesis was also described for primary tumors. The tumors were shown to overexpress VEGF-A, thereby inducing sentinel lymph node lymphangiogenesis. In a mouse model of delayed-type hypersensitivity (DTH), lymphangiogenesis was promoted by VEGF-A that was produced in the inflamed tissue. In addition, genetic variety, that is, single-nucleotide polymorphisms in the gene coding for VEGF-A, is associated with eye diseases like neovascular age-related macular degeneration and diabetic retinopathy. VEGF-C and VEGF-D

VEGF-C and VEGF-D are the main growth factors for lymphangiogenesis and both mediate their function by binding to receptors VEGFR-2 and VEGFR-3, present on endothelial cells. VEGF-C stimulates migration of cultured endothelial cells in vitro and increases – in its fully processed form – vascular permeability, migration, and proliferation of endothelial cells. Recently, the decisive role for VEGF-C during lymphangiogenic development was demonstrated in a study where homozygote as well as heterozygote VEGF-C-lacking mice were shown

467

to have severe defects in the formation of lymphatic vessels. A study undertaken to analyze VEGF-C and its role in corneal neovascularization suggests VEGF-C to induce corneal lymphangiogenesis by binding to its cognate receptor VEGFR3 on lymphatic vessels in the conjunctiva. Inflammatory cells invading the cornea were identified as the main source of VEGF-C that was strongly upregulated 3 days following the injury. Besides its role as classic lymphangiogenic growth factor, VEGFC was reported to induce angiogenesis in vivo. VEGF-D also can act as a potent angiogenic factor, controlled by the nuclear oncogene c-fos and thus playing an important role in tumor invasion and tumor cell growth. Basic fibroblast growth factor Another important pro-angiogenic molecule – the basic fibroblast growth factor (bFGF/ FGF-2) – belongs to the FGF family. bFGF and acidic FGF (aFGF) expression has been demonstrated immunohistochemically in the outer retina of rat and mouse. bFGF was analyzed in several corneal neovascularization models and was recently demonstrated to induce angiogenesis as well as lymphangiogenesis in vivo in a mouse model corneal micropocket assay. Lymphangiogenesis was mediated by bFGF in an indirect way via VEGFR3 and was suppressed after inhibition of VEGFR3 signaling with anti-VEGFR3 antibodies. Both factors, bFGF and aFGF, were detectable in retinal pigment epithelial cells from choroidal neovascular membranes from human subjects with age-related macular degeneration (AMD), whereas there was only little immunoreactivity for the growth factors in retinal pigmented epithelial (RPE) cells from healthy eyes. This suggests an important role for aFGF and bFGF in the development of choroidal neovascularization. bFGF might play an indirect role in initiation of neovascularization and interacts with the VEGF signal-transduction pathways. This is supported by the fact that bFGF was found to be colocalized with VEGF in cells of epiretinal and choroidal neovascular membranes, suggesting that more than one growth factor may contribute to pathological angiogenesis. In favor for that theory is that mice with a disruption of the bFGF-coding gene can still develop choroidal neovascularization. Recently, bFGF was thought to take a role in progression and survival of retinoblastoma, a tumor producing significant amounts of bFGF. The differential production and response to isoforms of bFGF reveal bFGF as a growth factor influencing pathogenesis and chemoresistence of retinoblastoma. Inhibitory PAS (Per/Arnt/Sim) domain protein As mentioned earlier, the upregulation of angiogenic molecules like VEGF-A and angiopoietin-4 (Ang-4)

468

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

during hypoxia is mediated by hypoxia-inducible transcription factor-1 (HIF-1) and can induce an angiogenic response. Interestingly, hypoxic conditions in the cornea appearing, for example, during overnight closure of the eye, do not induce corneal neovascularization, indicating the presence of factors suppressing hypoxia-induced angiogenesis. However, prolonged contact lens use has been associated with corneal angiogenesis and hypoxia has been implicated in this process. IPAS, a basic helix–loop–helix (bHLH)/PAS protein, expressed in mouse corneal epithelium, was suggested as a negative regulator of HIF-mediated control of gene expression: only low levels of IPAS mRNA were detectable in primary cultures of mouse corneal cells under normoxic conditions, whereas under hypoxic conditions IPAS mRNA expression was upregulated. Following transfection of primary corneal cells with an IPAS antisense vector, VEGF mRNA expression – under normoxic and hypoxic conditions – was upregulated. Furthermore, in vivo experiments with mouse corneas containing pellets with IPAS antisense oligonucleotides showed a significantly induced neovascularization compared to the eyes treated with the IPAS sense oligonucleotide.

Limbal Barrier Function The border between sclera/conjunctiva and the transparent cornea, the limbus, is of great importance for angiogenesis: the loops and arcades of conjunctival capillaries as well as the lymphatic capillaries end in the limbal region. Corneal neovascularization, however, arises in the limbal area from preexisting pericorneal vessels (hemangiogenesis as well as lymphangiogenesis). Stem cells, required for corneal epithelial cell proliferation and differentiation, are located in the basal epithelium at the corneoscleral limbus. They were described to act as a barrier between conjunctival and corneal epithelium and as important for corneal wound healing. In normal situations, the limbal stem cells prevent conjunctival epithelial cells from migrating to the ocular surface thereby inhibiting corneal neovascularization. This limbalbarrier concept may contribute to the maintenance of the angiogenic privilege. The theory is supported by the observation of conjunctivalization of the corneal surface with subsequent vascularization in situations of loss or malfunction of the stem cells. Angiogenic Privilege and Immune Privilege

Cornea and Corneal Epithelium Consistent with the assumption of a redundantly organized angiogenic privilege, numerous endogenous anti-angiogenic factors in the cornea are described to be implicated in the regulation of angiogenesis. Years ago, the corneal epithelium itself was found to have angiogenic activity. In 1978, Eliason reported data from an in vivo system suggesting corneal epithelium as a source for an unknown vasostimulating substance. In vitro experiments from Eliason and Elliott also showed a stimulating effect of corneal epithelial homogenate and epithelial-conditioned medium on the proliferation of cultured rabbit vascular endothelial cells. Recent research attributes corneal epithelium an anti-angiogenic function: an intact corneal epithelium can suppress inflammation and corneal neovascularization in the graft following orthotopic transplantation in mice. Secondly, mice with de-epithelialized corneas have significantly increased recruitment of CD45þ inflammatory cells and an increased neovascular response compared to mice with an intact epithelium. One potent mechanism for an antiangiogenic function of corneal epithelium is the ectopic constitutive expression of VEGFR-3 on normal human corneal epithelial cells. VEGFR-3 can act as a decoy receptor to bind VEGF-C, thus functioning as a sink for the angiogenic molecules and inhibiting inflammation induced corneal hemangiogenesis and lymphangiogenesis. A similar task fulfils the soluble form of VEGFR-1 – expressed in the cornea – where it can neutralize VEGF-A.

Ingrowth of blood and lymphatic vessels into the cornea is incompatible with good vision. Visual acuity is impaired not only by vascularization itself, but also by secondary changes such as lipid keratopathy, corneal edema, or bleeding into the cornea, thereby reducing corneal clarity and transparency. As mentioned earlier, actively maintaining the avascularity even under inflammatory or other angiogenic conditions is ensured by the angiogenic privilege. It contributes, at least partly, to the occurrence of the prolonged graft survival of corneal allografts, called the immune privilege of the eye. The phenomenon of the immune-privileged site was first proposed by Medawar, in 1948, and has built a foundation for numerous research. Nowadays, ocular immune privilege is commonly seen as the fact that vulnerable organs or tissues are protected from pathogens without an immunogenic inflammation that would permanently damage those tissues and/or would lead to a loss of specialized functions. An immune response after transplantation in so-called low-risk eyes can only be noticed in around 10%, although under normal circumstances there is no HLA matching and only a topical, but no systemic, immunosuppression. In contrast, immune reactions in high-risk eyes with preceding corneal inflammation or neovascularization occur in over 50%. The pathologically vascularized recipient bed prior to corneal transplantation (i.e., penetrating keratoplasty), therefore, lowers the outcome of corneal transplantation and is an important risk factor for subsequent immune reactions.

Concept of Angiogenic Privilege

Immune responses are primarily mediated by corneal lymphatic vessels which form the afferent arc of the immune response. Via the lymphatic vessels invading antigen-presenting cells (APCs; dendritic cells from the graft or host) and antigenic material from the graft can be transported via conjunctival lymph vessels to the regional lymph nodes. The draining cervical lymph nodes were shown to be critically involved in promoting alloimmunity and allograft rejection. Following surgical removal of the cervical lymph nodes and following orthotopic corneal transplantation of fully mismatched high-risk allografts, over 90% of the hosts accepted the allograft. The importance of the lymphatic vessels, being the afferent arc of the immune response, offers new therapeutic opportunities for improving graft survival. Interfering with this pathway might restore the immune privileged status of the eye and ensures prolonged graft survival in low- and high-risk eyes. Early studies have shown that induction of donor-specific anterior chamber-associated immune deviation (ACAID) – manifestation of the ocular immune-privilege induced prolonged graft survival in high-risk eyes of C57BL/6 mice. Recently, it was shown that Integrin a5-blockade could significantly block the outgrowth of lymphatic vessels in the cornea. The angiostatic drug bevacizumab, a recombinant humanized monoclonal antibody against VEGF-A, inhibits corneal hemangiogenesis and lymphangiogenesis in vitro and in vivo. Furthermore, inhibition of corneal hemangiogenesis and lymphangiogenesis by a molecular VEGF-A trap leads to improved long-term graft survival. In addition to the inhibition of inflammatory lymphangiogenesis, alternative strategies like induction of regression of established lymphatic vessels in prevascularized corneas and influencing the recruitment of APCs could be possible methods for corneal anti-lymphangiogenic treatment. Recently, it was shown that even hemangiogenesis and lymphangiogenesis occurring following transplantation increase the risk for graft rejection after high-risk corneal transplantation.

469

See also: Avascularity of the Cornea; Corneal Angiogenesis.

Further Reading Azar, D. T. (2006). Corneal angiogenic privilege: Angiogenic and antiangiogenic factors in corneal avascularity, vasculogenesis, and wound healing (an American Ophthalmological Society thesis). Transactions of the American Ophthalmological Society 104: 264–302. Cebulla, C., Jockovich, M. E., Pina, Y., et al. (2008). Basic fibroblast growth factor impact on retinoblastoma progression and survival. Investigative Ophthalmology and Visual Science 49(12): 5215–5221. Chang, J. H., Gabison, E. E., Kato, T., and Azar, D. T. (2001). Corneal neovascularization. Current Opinion in Ophthalmology 12: 242–249. Churchill, A. J., Carter, J. G., Ramsden, C., et al. (2008). VEGF polymorphisms are associated with severity of diabetic retinopathy. Investigative Ophthalmology and Visual Science 49: 3611–3616. Cursiefen, C. (2007). Immune privilege and angiogenic privilege of the cornea. Chemical Immunology and Allergy 92: 50–57. Cursiefen, C., Chen, L., Dana, M. R., and Streilein, J. W. (2003a). Corneal lymphangiogenesis: Evidence, mechanisms, and implications for corneal transplant immunology. Cornea 22: 273–281. Cursiefen, C., Seitz, B., Dana, M. R., and Streilein, J. W. (2003b). Angiogenesis and lymphangiogenesis in the cornea. Pathogenesis, clinical implications and treatment options. Ophthalmologe 100: 292–299. Folkman, J. and Shing, Y. (1992). Angiogenesis. Journal of Biological Chemistry 267: 10931–10934. Hori, J. and Niederkorn, J. Y. (2007). Immunogenicity and immune privilege of corneal allografts. Chemical Immunology and Allergy 92: 290–299. Makino, Y., Cao, R., Svensson, K., et al. (2001). Inhibitory PAS domain protein is a negative regulator of hypoxia-inducible gene expression. Nature 414: 550–554. Niederkorn, J. Y. (1999). The immune privilege of corneal allografts. Transplantation 67: 1503–1508. Niederkorn, J. Y. (2007). Immune mechanisms of corneal allograft rejection. Current Eye Research 32: 1005–1016. Shweiki, D., Itin, A., Soffer, D., and Keshet, E. (1992). Vascular endothelial growth factor induced by hypoxia may mediate hypoxiainitiated angiogenesis. Nature 359: 843–845. Streilein, J. W. (2003a). Ocular immune privilege: The eye takes a dim but practical view of immunity and inflammation. Journal of Leukocyte Biology 74: 179–185. Streilein, J. W. (2003b). Ocular immune privilege: Therapeutic opportunities from an experiment of nature. Nature Reviews Immunology 3: 879–889.

Corneal Angiogenesis M S Cortina and D T Azar, University of Illinois at Chicago, Chicago, IL, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Angiogenesis – The formation of new vessels from preexisting vascular structures. Allograft – The tissue taken from one person for transplantation into another. Autograft – The tissue transplanted from one part of the body to another part of the same individual. Chemotaxis – The movement of neutrophils toward bacteria or an area of tissue damage. Conjunctival metaplasia – An abnormal epithelial differentiation represented by a spectrum of skin-like changes of conjunctival epithelium. Corneal extracellular matrix – The tissue that provides structural support to the cells in the cornea. Corneal pannus – The fibrovascular connective tissue that proliferates in the anterior layers of the peripheral cornea in inflammatory corneal disease. Growth factor – A naturally occurring protein capable of stimulating cellular growth, proliferation, and differentiation. Hypercapnia – High levels of carbon dioxide. Hypoxia – Oxygen deficiency. Limbal stem cell deficiency – The loss of stem cells in the limbus (ring around the base of the cornea which supports health of the corneal epithelium). Matrix metalloproteinases (MMPs) – The zincdependent endopeptidases capable of degrading all kinds of extracellular matrix proteins; they can also process bioactive molecules. Penetrating keratoplasty (PK) – The procedure in which a full-thickness button of cornea is removed from the recipient and replaced with a similar-sized or larger button of tissue from a donor. Vasculogenesis – The formation of new blood vessels from bone-marrow-derived angioblasts that occurs mainly during embryogenesis.

Introduction Under homeostatic conditions, the cornea is avascular, which is critical for corneal light transmission and proper optical performance. Corneal avascularity is maintained by tightly controlled biological anti-angiogenic events that counterbalance the effects of pro-angiogenic factors in the

470

cornea. Under pathological conditions, the balance may be shifted toward angiogenesis, leading to the formation of new blood vessels and lymphatic channels. New blood vessel formation or corneal neovascularization (NV) is a sight-threatening condition usually associated with inflammatory or infectious disorders. It is a major contributor to the loss of corneal transparency. The presence of corneal NV, in turn, elevates the risk of graft rejection and decreases the success of penetrating keratoplasty (PK).

Angiogenesis Angiogenesis is the main process of blood vessel formation in nonembryonic tissue. It involves the formation of new vessels from preexisting vascular structures and is the primary mechanism underlying corneal NV. New blood vessels also form during vasculogenesis, which is the formation of new blood vessels from bone-marrow-derived angioblasts and occurs mainly during embryogenesis. Angiogenesis is a complex process that starts with vasodilation of existing vessels and an increase in vascular permeability. This leads to extravasation of plasma proteins (such as fibrin), growth factors, and inflammatory mediators. The accumulation of plasma proteins in the surrounding tissue provides a supporting structure for subsequent endothelial cell (EC) migration. The combined presence of growth factors and inflammatory mediators stimulates the degradation of the extracellular matrix (ECM), making room for EC migration as well as releasing angiogenic factors anchored in the matrix. The newly released angiogenic factors then continue to activate ECs, which migrate from preexisting vessels and form sprouting tubes. The avascularity of the cornea dictates that nutrients for this tissue be obtained from adjacent tissues. The three major sources of nutrients are tear film, aqueous humor, and the pericorneal capillary plexus at the limbus. This plexus nourishes the peripheral cornea and is derived from ciliary arteries, which are branches of the ophthalmic artery. These vessels do not normally enter the cornea. However, new blood vessels may sprout from capillaries and venules of the pericorneal plexus under pathological conditions, leading to corneal NV.

Etiology and Epidemiology of Corneal NV Although the exact incidence of corneal NV is not known, it was estimated that this condition affects 1.4 million

Corneal Angiogenesis

patients in the United States annually and that 20% of corneal buttons obtained during PK show evidence of NV. The causes of corneal NV include immune, inflammatory, infectious, degenerative, and traumatic disorders. Corneal infections are the most common worldwide causes of corneal NV leading to vision loss. A classic example is trachoma, an infectious disease characterized by the formation of a superior pannus, which can extend to the central cornea and is often associated with corneal scarring, opacification, and loss of visual acuity. The incidence of trachoma in the US is low; however, this condition remains a major cause of blindness in other parts of the world (Figure 1). Corneal NV is also commonly associated with other severe bacterial and viral infections. The herpes virus family (primarily herpes simplex and herpes zoster viruses) is the primary cause of keratitis-induced NV in transplant buttons. In herpes-simplex-induced stromal keratitis, NV is essential for the pathogenesis of keratitis, and inhibition of angiogenesis can reduce the formation of corneal lesions. Although infections account for many US cases of corneal NV, the most common cause of corneal NV in the US is the use of contact lenses. In this case, hypoxia and hypercapnia are thought to be associated with the induction of NV (Table 1). The incidence of corneal NV after PK can be as high as 40% at 6–9 months after surgery. The prognosis of transplanting grafts into heavily vascularized corneas is poor. Graft failure has been reported to contribute to >30% of the histopathological diagnoses obtained from vascularized corneal buttons. Risk factors for corneal NV after PK in patients without active inflammation, previous corneal NV, or persistent epithelial defects include suture knots buried in the host stroma, active blepharitis, and a large recipient bed.

471

Immune disorders also contribute to corneal NV (Figure 2). These disorders can result in significant vision loss and include ocular pemphigoid, rosacea, atopic keratoconjunctivitis, and Stevens–Johnson syndrome. The incidence of corneal NV among patients with these disorders can be considerable. Long-term follow-up of patients with atopic keratoconjunctivitis, for example, revealed that the rate of corneal NV was as high as 60% during the disease course. Limbal stem cell deficiency, which may occur following trauma and chemical burns, is another cause of corneal NV. It may be also associated with aniridia and autoimmune disorders. Limbal stem cell deficiency produces not only corneal NV, but also corneal inflammation and conjunctivalization of the corneal epithelium. The restoration of corneal avascularity after successful limbal stem cell transplantation underscores the importance of the antiangiogenic and anti-inflammatory activity of normal corneal epithelial cells.

Clinical Manifestations Corneal NV can be classified as pannus or stromal NV. In the former, fibrovascular tissue is visible between the epithelium and Bowman layer. Inflammatory pannus is associated with prominent leukocyte infiltration and disruption of Bowman’s layer. In contrast, degenerative pannus is characterized by fewer inflammatory cells, an intact Bowman’s layer, and regression of the vascular component that leaves a layer of fibrous tissue. Stromal NV, located posterior to Bowman’s layer, is more commonly seen in the anterior two-third of the stroma. In herpetic and syphilitic interstitial keratitis, deep stromal vessels (and ghost vessels in the late quiescent stages) are seen just anterior to Descemet’s membrane (Figure 3). Visual acuity is reduced by corneal NV. Reduced visual acuity may be secondary to multiple factors. For example, opacity caused by circulating blood cells may interfere with visual acuity. Acuity may also be reduced by irregular architecture of vascular walls, a feature that induces higher-order aberrations. Other effects contributing to diminished visual acuity may include alteration in the spacing of stroma collagen fibers, fluid leakage, edema and lipid deposition in the surrounding tissue, and corneal surface irregularities (Figure 4).

Mechanisms Underlying the Maintenance of Corneal Avascularity Figure 1 Salzman’s nodular degeneration with associated superficial corneal neovascularization.

Several mechanisms have been proposed to contribute to corneal avascularity. These include:

472 Table 1

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease Causes of corneal neovascularization

Inflammatory disorders

Infectious keratitis

Degenerative/congenital

Traumatic/iatrogenic

Ocular pemphigoid Atopic conjunctivitis Rosacea Graft rejection Lyell’s syndrome Stevens–Johnson syndrome Graft vs. host disease

Herpes simplex Herpes zoster Pseudomonas Chlamydia Syphilis Candida Fusarium Aspergillus Onchocerciasis

Pterygium Terrien marginal degeneration Aniridia Degenerative pannus

Contact lens wear Alkali burns Ulceration Iatrogenic Stem cell deficiency

Figure 2 Corneal neovascularization and infectious keratitis in a patient with underlying severe dry eye syndrome secondary to rheumatoid arthritis.

Figure 3 Superficial and deep stromal neovascularization in a patient with neurotrophic cornea.

1. Corneal dehydration resulting in tightly packed collagen lamellae. The relationship between corneal NV and corneal edema was first reported in 1949 by Cogan, who postulated that the distention and bursting of the vessels preceding formation of the capillary sprouts were due to a decrease in external pressure that reduced

Figure 4 Lipid deposition secondary to corneal neovascularization. Note relative small vessel crossing the host-graft junction responsible for significant lipid deposit.

vessel wall support. However, subsequent studies showed that corneal edema alone is not sufficient to trigger corneal NV. 2. The angiostatic nature of corneal epithelial cells. Blood vessels are known to be capable of growing into corneas in the absence of epithelium. Early research suggested that corneal epithelial cells are a source of angiogenic factors. More recent studies suggest that the corneal epithelium has an anti-angiogenic effect. The presence of soluble vascular endothelial growth factor (VEGF) receptors 1 and 3, and other naturally occurring antiangiogenic factors, in the corneal epithelium contributes to its angiostatic nature. 3. The immune privilege of the cornea. The mechanisms underlying corneal immune privilege include low expression of major histocompatibility complex (MHC) antigens on corneal cells, expression of Fas ligand by the cornea, a relative paucity of mature antigen-presenting cells, and the presence of immunomodulatory molecules in the anterior chamber. Importantly, this state can be reversed by inflammation, and such reversal may contribute to vascularization. For example, polymorphonuclear leukocytes have the potential to initiate corneal NV through

Corneal Angiogenesis

4.

5.

6. 7.

the release of chemical mediators. Accordingly, there is a clear association between corneal inflammation and NV. Nevertheless, corneal NV may occur in the absence of inflammation. Low corneal temperature, extensive innervation, and movement of the aqueous humor across the cornea. The roles of these factors in the maintenance of corneal avascularity are currently unclear. The barrier function of limbal cells. The limbus is thought to prevent corneal NV by acting as a barrier to conjunctival growth over the cornea. This barrier function may be attributable to the ability of limbal stem cells to replenish the corneal epithelium, thus preventing invasion of conjunctival epithelium and avoiding NV. This hypothesis has been used to account for corneal NV following experimental limbal damage and stem cell dysfunction. It is also one of the explanations for corneal pannus observed in aniridia. This hypothesized barrier function of limbal cells supports the use of limbal stem cell transplantation as a definitive treatment for ocular-surface disorders. However, a physical barrier may not completely explain corneal avascularity. Low levels of angiogenic factors and active production of potent anti-angiogenic factors in the cornea during homeostasis. Active production of potent anti-angiogenic factors. Although the five previously mentioned factors likely contribute to maintenance of corneal avascularity, available evidence supports this as the main mechanism responsible for maintaining corneal avascularity.

Corneal Angiogenic Privilege: The Balance between Angiogenesis and Anti-Angiogenesis Multiple local and systemic signals are responsible for regulating growth and regression of new blood vessels. These signals include cyclic adenosine monophosphate (cAMP), steroid hormones, protein kinase C (PKC) agonists, polypeptide growth factors, oxygen, free radicals, glucose, cobalt, and iron. In the cornea, the tight equilibrium between these pro- and anti-angiogenic signals may be disrupted under pathological conditions. Ultimately, this may tip the balance toward an upregulation of proangiogenic factors or a downregulation of anti-angiogenic factors, in either case leading to corneal NV (Table 2). When the cornea is injured, wound healing often occurs in the absence of NV. This is the case for most adequately treated corneal infections. Healing after corneal surgery is also usually avascular. Corneal wound healing involves four phases. In the first phase, keratocytes near the area of epithelial debridement undergo apoptosis. In the second phase, adjacent keratocytes proliferate to repopulate the wound within 24–48 h after wounding. These keratocytes

Table 2

473

Factors involved in the regulation of angiogenesis

Angiogenic factors

Anti-angiogenic factors

Fibroblast growth factor (FGF) Vascular endothelial growth factor (VEGF) Placenta growth factor (PGF) Transforming growth factor-a (TGF-a) Transforming growth factor-b (TGF-b) Insulin-like growth factor (IGF) Leptin Integrins (anb3) Platelet-derived growth factors (PDGF) Matrix metalloproteinases (MMPs) Angiogenin

Endostatin Angiostatin

Hepatocyte growth factorscatter factor (HGF-SF) Tumor necrosis factor-a (TNF-a) Connective tissue growth factor (CTGF) Interleukin-8 (IL-8) Monocyte chemoattractant protein-1 (MCP-1) Platelet-activating factor (PAF)

Prolactin Matrix metalloproteinases (MMPs) Tissue inhibitor of MMPs (TIMPs) Thrombospondin Arresten Canstatin Tumstatin Pigment-epitheliumderived factor (PEDF) Tumor necrosis factor a (TNF-a) Interleukin-4 (IL-4) Interleukin-13 (IL-13) Fibulin Endoperellin Antithrombin Plasminogen activator inhibitor (PAI) Vasostatin Neostatin-7

transform into fibroblasts and migrate into the wound area. This process may take up to a week and is not accompanied by corneal NV. In the third phase, fibroblasts may transform into myofibroblasts. This occurs in laser-inflicted wounds lacking Bowman’s layer and in incisional wounds. Myofibroblasts may take up to a month to become apparent. Corneal NV is also absent in this phase of wound healing. The fourth and final phase involves stromal remodeling and is dependent on the original wound. When wound healing is accompanied by ECM turnover, angiogenesis in granulation tissue is usually observed. Some of the molecules that regulate angiogenesis are discussed below.

Angiogenic Molecules Vascular endothelial growth factor VEGF is a dimeric 46-kDa glycoprotein. This growth factor stimulates angiogenesis by increasing EC proliferation, migration, proteolytic activity, and capillary tube formation. It also significantly increases vascular permeability. The VEGF family includes VEGF-A, -B, -C, -D, placenta growth factor (PlGF), and the viral VEGF

474

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

homolog VEGF-E. VEGF-B promotes nonangiogenic tumor progression, while VEGF-C and -D participate in angiogenesis and lymphangiogenesis. VEGF-A also participates in angiogenesis and increases vascular permeability. Five isoforms of VEGF-A (VEGF115, 121, 165, 189, and 206) can be generated by alternative splicing of the same gene. The longer isoforms (VEGF189 and 206) are matrixbound, whereas the shorter isoforms (VEGF121 and 165) are freely diffusible. These VEGF isoforms produce different actions when secreted. For example, all isoforms increase in vascular permeability, but only VEGF121 and VEGF165 have mitogenic activity. VEGF121 has greater angiogenic activity than VEGF165 or VEGF189. On the other hand, VEGF165 is more potent than VEGF121 in induction of inflammation, intercellular adhesion molecule-1 (ICAM-1) expression in ECs, and chemotaxis of monocytes. This suggests that alternate splicing of VEGF messenger RNA (mRNA) can be regulated to achieve a range of physiologic actions. The VEGF family members act through binding to high-affinity receptor tyrosine kinases. Two high-affinity receptor tyrosine kinases have been identified for VEGFA: VEGFR-1 (fms-like tyrosine kinase-1 or Flt-1) and VEGFR-2 (kinase insert domain-containing receptor or KDR). VEGFR-3 (fms-like tyrosine kinase-4 or Flt-4) serves as a high-affinity receptor for VEGF-C and -D. Both VEGFR-1 and -2 are expressed primarily in vascular ECs, while VEGFR-3 is predominantly expressed in lymphatic ECs. VEGF-B binds to VEGFR-1 and has mild mitogenic activity. In contrast, binding of VEGF-D and -C to VEGFR-3 regulates the growth and differentiation of blood vessels and lymphatic endothelium. VEGF is produced by macrophages, T cells, astrocytes, pericytes, fibroblasts, retinal pigment epithelial cells, and smooth muscle cells. In addition, VEGF is expressed in all three cellular layers of the cornea. It is highly expressed in vascular ECs of limbal vessels and in new stromal vessels. Under inflammatory conditions, VEGF expression is increased in epithelial and vascular ECs, particularly near macrophage infiltrates and fibroblasts in corneal scars. Following corneal cautery, VEGF165 and 189 mRNA is increased at 48 h and returns to baseline by day 7. Immunohistochemistry has revealed that VEGF is initially expressed in neutrophils and later expressed in macrophages, demonstrating that VEGF production by leukocytes is associated with corneal NV. In addition, VEGF concentration is significantly increased in vascularized corneas as compared to normal corneas. In limbal-deficiency-induced corneal NV, VEGF mRNA and protein are induced after injury and are both temporally and spatially correlated with inflammation and NV. VEGF is not only induced during NV, but is also required for corneal angiogenesis. The indispensable role of VEGF in angiogenesis is shown by the finding that

stromal implantation of anti-VEGF antibodies inhibits NV in a rat model. Conversely, implantation of a Hydron pellet containing VEGF into the stroma induces severe corneal NV without significant inflammation. The effects of VEGF in the cornea are not limited to NV, as this growth factor has also been shown to regulate goblet cell migration. Studies analyzing the correlation between cornea NV and conjunctivalization showed that VEGFR-1 is present in the conjunctiva-like epithelium covering the cornea as well as in goblet cells, invading leukocytes, and the corneal vasculature. Inhibition of VEGF activity inhibited not only corneal NV, but also goblet cell density, suggesting that VEGF may promote goblet cell migration. Evidence suggests that VEGF also participates in corneal lymphangiogenesis. Corneal lymphangiogenesis may contribute to graft sensitization and rejection, following high-risk keratoplasty of vascularized corneas. VEGF-C binds to VEFGR-3 and induces lymphatic growth in the cornea. Interestingly, inhibition of lymphatic growth is observed after administration of a VEGF trap that neutralizes VEGF-A, but not VEGF-C or -D. This could be explained by the chemotactic effect on macrophages that release VEGF-C in inflamed corneas observed with VEGF-A. Thus, VEGF-A amplifies signals essential for lymphatic growth. In general, corneal lymphangiogenesis seems to correlate well with the degree of corneal hemangiogenesis. Recent studies have shown that VEGF, although present in the cornea, does not promote angiogenesis under normal conditions. VEGF-A found in corneal tissue is mostly bound to an alternative spliced secreted isoform of VEGFR-1 (sflt-1), which acts as a trap for secreted VEGF-A and in this way contributes to maintenance of corneal avascularity. In addition, VEGFR-3 is expressed in endothelial as well as epithelial cells in the cornea. When VEGF-C and -D bind to endothelial VEGFR-3, they stimulate proangiogenic signaling. In contrast, VEGFR-3 expressed by corneal epithelium acts as a decoy receptor sequestering VEGF but yet rendering it available when an angiogenic response is needed to enhance the immune defense. This VEGFR-3 sink system is a potent mechanism that inhibits inflammatory-induced angiogenesis. Basic fibroblast growth factor Basic fibroblast growth factor (bFGF) is another potent angiogenic factor. It is a member of the fibroblast growth factor (FGF) family, which includes 23 heparin-binding peptides widely expressed during cell differentiation, angiogenesis, mitogenesis, and wound healing. bFGF functions are mediated by the receptors FGFR-1, -2, -3, and-4. FGF recptor-1 (FGFR-1) is expressed in normal corneal epithelium, while bFGF is upregulated following injury. It is also upregulated following co-culture of corneal epithelial cells with vascular EC and keratocytes. The affinity of bFGF for its receptor differs according to the extent of

Corneal Angiogenesis

maturation of new vessels. This may be due to varying expression of heparan sulfate proteoglycans and highlights the role of ECM proteins in the regulation of corneal angiogenesis. Matrix metalloproteinases

The matrix metalloproteinases (MMPs) constitute a multigene family of zinc-binding proteolytic enzymes that participate in ECM remodeling. Many of the growth factors that modulate angiogenesis also influence MMP expression. These growth factors include VEGF, FGF-2, and tumor necrosis factor-alpha (TNF-a). Vascular ECs respond by secreting proteolytic enzymes that degrade the ECM to facilitate migration and differentiation of ECs. The MMPs that have identified in the cornea are collagenases I and II (MMP-1 and -13), stromelysin (MMP-3), matrilysin (MMP-7), membrane-type MMP (MT-MMP-14), and gelatinases A and B (MMP-2 and -9). Both MMP-2 and MMP-9 are proteolytically activated primarily by MT1MMP during capillary formation. Several reports suggest that these MMPs participate in vascular invasion by directly degrading the matrix or releasing matrix-bound cytokines and growth factors. Accordingly, inhibition of MMP-9 activity in the cornea decreases angiogenesis. However, given their ability to degrade ECM, MMPs exhibit a dual action in angiogenesis. For example, MMP-2 activation may release anti-angiogenic fragments, allow the production of potent angiostatic factors, or facilitate angiogenesis. Lipid mediators

One of the initial events that occurs after corneal injury is the release of arachidonic acid. In the corneal epithelium, arachidonic acid is then metabolized by cyclooxygenase (COX) to generate eicosanoids (such as 12- and 15-HETE), lipoxin A4 (LXA4), and prostaglandins. 12(S)-HETE is a powerful angiogenic factor, and COX inhibitors have been shown to reduce corneal angiogenesis in animal models. Plateletactivating factor is another potent lipid mediator released from the cell membrane after corneal injury. It contributes to corneal NV by increasing expression of VEGF, MMP-9, and urokinase plasminogen activator (uPA), all of which subsequently stimulate vascular EC migration.

Anti-Angiogenic Molecules Angiostatin

Angiostatin results from the cleavage of plasminogen. Several MMPs can cleave plasminogen to generate angiostatinlike molecules. The inhibitory effect of angiostatin on vascular ECs may be due to inhibition of adenosine triphosphate (ATP) synthesis in these cells, an effect that decreases EC migration and proliferation. Angiostatin binds to integrin alpha-v beta-3 (avb3) and affects angiogenesis as well as

475

developmental NV. It also induces vascular EC apoptosis mainly in areas of NV. All three layers of the cornea are able to synthesize plasminogen and angiostatin. Tears collected after overnight eye closure contain a significant amount of angiostatinrelated molecules known to have anti-angiogenic properties. This has also been shown in tears of contact lens-bearing patients, suggesting that these molecules play a role in preventing NV under hypoxic conditions. Corneal NV occurs following injection of anti-angiostatin antibodies into corneas having undergone post-excimer laser keratectomy. This supports the idea that plasminogen and angiostatin are important for the maintenance of corneal avascularity. Endostatin and neostatins Endostatin is a 20-kDa proteolytic fragment of collagen XVIII that exhibits anti-angiogenic activity. It was originally discovered as an angiogenic inhibitor purified from conditioned media of murine hemangioendothelioma cells. Endostatin inhibits bFGF-induced corneal NV as well as VEGF-induced vascular EC migration and proliferation. Collagen XVIII is localized mainly in the corneal vascular and epithelial basement membrane. Smaller fragments of collagen XVIII, known as Neostatins -7 and -14, are generated by the enzymatic activity of MMPs -7 and -14, respectively. They have potent antiangiogenic and anti-lymphangiogenic properties. Local production of endostatin and Neostatins -7 and -14 may occur during wound healing. Endostatin is Food and Drug Administration (FDA)-approved for the treatment of cancer-related NV. Pigment-epithelial-derived factor Pigment-epithelial-derived factor (PEDF) is a potent anti-angiogenic and neurotrophic factor that is found in multiple eye tissues including the cornea. In contrast to VEGF, which is induced under low oxygen conditions, PEDF expression is suppressed during hypoxia. PEDF induces EC apoptosis. It also has antipermeability and anti-inflammatory activity that counterbalances VEGF actions. Studies have shown that PEDF-blocking antibodies induce corneal NV when implanted into the stroma and that recombinant PEDF inhibits bFGF-induced corneal NV. These findings are consistent with an essential role for PEDF in maintaining the avascularity of ocular tissues. Given its effectiveness at countering VEGF activity, PEDF may be a good pharmacological inhibitor of angiogenesis. Arresten, canstatin, and tumstatin Arresten is a 26-kDa protein derived from the noncollagenous (NCl) domain of the type IV collagen a1 chain. This molecule has been shown to inhibit bFGFstimulated proliferation, migration, and tube formation of

476

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

cultured ECs. It also inhibits NV in vivo. Canstatin is a 24-kDa fragment of the type IV collagen a-chain. It also inhibits EC proliferation and tube formation. The mechanism of action appears to involve phosphoinositide 3-kinase/ protein kinase B (PI3K/Akt) inhibition and depends on signaling events transduced through membrane-death receptors. Tumstatin, a 28-kDa fragment of the type IV collagen a3 chain, also has anti-angiogenic activity.

Therapy Identifying and adequately treating the underlying cause of corneal NV is critical. Therapies for corneal NV may range from antimicrobial therapy for infectious keratitis to systemic immunosuppression for autoimmune diseases such as ocular cicatricial pemphigoid. Some of the established and investigational medical and surgical treatments for corneal NV are discussed below. Medical Treatments Anti-inflammatory compounds, such as steroids, have a long history of use for the suppression of inflammation and associated angiogenesis. The anti-angiogenic effects of steroid treatment are likely secondary to their antiinflammatory actions and include inhibition of chemotaxis and cytokine synthesis. Steroids have also been shown to inhibit vascular EC proliferation and migration. Unfortunately, the side effects of these compounds make long-term administration difficult in some patients. Moreover, their role in corneal NV that is not associated with inflammation is limited. Advances in our understanding of the mechanisms underlying ocular NV has led to the identification of new pharmacologic targets. Given the key role of VEGF in NV of the eye, attention has been directed to developing drugs that will counteract the activity of this factor. Bevacizumab is an anti-VEGF antibody that binds to all VEGF isoforms. This molecule inhibits VEGF-receptor interactions and in this way, blocks all VEGF actions. It is currently approved by the FDA to treat metastatic colorectal cancer. It has also been tested for the treatment of wet (neovascular) age-related macular degeneration (AMD). Ranibizumab is another anti-VEGF antibody that has been approved for use in the eye to treat wet AMD. Bevacizumab treatment of corneal NV has gained popularity since the successful use of this molecule to treat choroidal NV. Subconjunctival injection as well as topical application of this molecule has also been used with promising results to treat herpes simplex virus (HSV) keratitis, recurrent pterygia, rejection of corneal grafts, and Stevens–Johnson syndrome. However, data on these treatments are limited, and adverse effects such as loss of epithelial integrity and progression of thinning have been reported in a small number of patients.

Further investigation is required to establish efficacy, adequate dosing, and safety in the different clinical scenarios that present with corneal NV. Other forms of anti-VEGF therapy are currently undergoing clinical trials. One example is VEGF TRAP, a highaffinity VEGF antagonist designed to bind and neutralize VEGF in the circulation and within tissues. It binds to all isoforms of VEGF and to placental growth factor, which is a related pro-angiogenic factor. SIRNA-027, another anti-VEGF therapy, is a short interferon RNA designed to downregulate VEGFR-1 expression. PKC412 is an orally administered tyrosine kinase inhibitor that binds to the intracellular, enzymatically active domain of the VEGF receptor and prevents phosphorylation and activation of the VEGF signaling cascade. Some of these compounds may be available for use in the near future. Surgical Treatment One surgical approach for the treatment of corneal NV is laser therapy. The use of laser photocoagulation with a 577-nm yellow dye for the treatment of established corneal NV has been investigated. The effectiveness of this technique has been tested in clinically significant corneal NV resistant to medical therapy both before and after PK. Some reduction of corneal NV can be achieved; however, the benefit of laser photocoagulation prior to high-risk keratoplasty is unclear, and this technique does not appear to be useful for treating extensive corneal NV. An alternative to laser occlusion is fine-needle diathermy. This procedure is easy to perform, requiring only a 10-0 nylon suture and a unipolar diathermy unit. It produces occlusion of 50–100% of corneal NV and has been show to moderately benefit visual acuity in a series of 17 patients. Photodynamic therapy is currently used to treat choroidal NV. In this technique, a photo-sensitizer selectively accumulates in new vessels and is subsequently activated by a laser beam. This technique is currently under investigation in animal models of corneal NV. Finally, in some cases, conjunctival, limbal, or amniotic membrane transplantation may be required to restore the ocular surface. Conjunctival autograft and allograft transplantation have been shown to decrease corneal NV. Amniotic membrane has anti-angiogenic properties as well. Limbal autograft transplantation has been successful in cases of stem cell deficiency and conjunctival metaplasia. This technique not only treats the stem cell deficiency and decreases the angiogenic stimulus from chronic ulceration, but also directly inhibits vascular ECs. No ideal treatment is currently available for corneal NV. However, significant progress in the understanding of corneal angiogenesis has opened a new field of investigation that may lead to the development of novel therapeutic agents for the treatment of this condition.

Corneal Angiogenesis See also: Concept of Angiogenic Privilege.

Further Reading Ambati, B. K., Nozaki, M., Singh, N., et al. (2006). Corneal avascularity is due to soluble VEGF receptor-1. Nature 443: 993–997. Azar, D. (2006). Corneal angiogenic privilege: Angiogenic and antiangiogenic factors in corneal avascularity, vasculogenesis, and wound healing. Transactions of the American Ophthalmological Society 104: 264–302. Chang, J. H., Gabison, E. E., Kato, T., and Azar, D. T. (2001). Corneal neovascularization. Current Opinion in Ophthalmology 12: 242–249.

477

Cursiefen, C., Chen, L., Saint-Geniez, M., et al. (2006). Nonvascular VEGF receptor 3 expression by corneal epithelium maintains avascularity and vision. Proceedings of the National Academy of Sciences of the United States of America 103: 11405–11410. Dorrel, M., Uusitalo-Jarvinen, H., Aguilar, E., and Friedlander, M. (2006). Ocular neovascularization: Basic mechanisms and therapeutic advances. Survey of Ophthalmology 52: 3–19. Ma, D. H., Chen, J. K., Zhang, F., et al. (2006). Regulation of corneal angiogenesis in limbal stem cell deficiency. Progress in Retinal and Eye Research 25: 563–590. Zhang, S. X. and Ma, J. X. (2007). Ocular neovascularization: implication of endogenous angiogenic inhibitors and potential therapy. Progress in Retinal and Eye Research 26: 1–37.

Avascularity of the Cornea R J C Albuquerque and J Ambati, University of Kentucky, Lexington, KY, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Alloimmunity – A condition in which the body gains immunity, from another individual of the same species, against its own cells. Aniridia – Congenital disorder characterized by the abnormal deficient development of the iris and associated with corneal angiogenesis and poor vision. Atopic keratoconjunctivitis (AKC) – Allergic conjunctivitis where the conjunctiva is red and swollen. Untreated, AKC can progress to ulceration, scarring, cataracts, keratoconus, and corneal vascularization. Hemangiogenesis – It pertains to the specific growth of blood vessels. Limbus – The edge of the cornea where it joins the conjunctiva and the sclera. Lymphangiogenesis – It pertains to the specific growth of lymphatic vessels. Neovascularization – Formation of new blood and lymphatic vessels. Perforating keratoplasty – Corneal transplant with replacement of all layers of the cornea, but retaining the peripheral cornea.

Light, the substrate of vision, is required to transverse the full diameter of the eye globe and reach retinal photoreceptors giving rise to the intricate biophysical phenomenon of sight. The cornea, interfacing the outer world and the intra-ocular tissues, serves as an entry window through which light comes into the eye. Avascularity and optical transparency are directly related and a requirement for optimal vision. The absence of vessels in the cornea is also one of the pillars of corneal immune privilege, an important physiological phenomenon that is associated with the maintenance of corneal clarity and responsible for the high success rate of corneal transplants. The growth of blood and lymphatic vessels into the normally avascular cornea (neovascularization) is considered pathological as it impairs the passage of light resulting in severely deteriorated vision or complete corneal blindness, which together afflict over 200 million people worldwide. The absence of vascular structures in the cornea has been known for over 1000 years. But only recently, advances in molecular biology have led to improved

478

understanding of the homeostatic mechanisms underlying such phenomenon. The absence of vasculature (blood and lymphatic vessels) in the cornea is remarkably intriguing given the highly vascularized nature of the neighboring tissues such as the ocular conjunctiva. The abrupt and precise delineation of the limbal vasculature (Figure 1) suggests that corneal avascularity is an active process in which endogenous pro- and anti-angiogenic mechanisms are in harmony and that these molecular modulators of angiogenesis are differently expressed in these interfaces between the cornea and its neighboring tissues preventing the blood and lymphatic vessels from invading the avascular cornea.

Corneal Histology Histologically, the cornea is comprised of five layers (Figure 2). The epithelial layer coats its outermost surface and is composed of a thin nonkeratinized squamous stratified epithelium (only a few cells thick). Unlike the stratified squamous epithelium of the epidermis, that contains indented dermal papillae, the corneal epithelium lays flat on a thick basement membrane called Bowman’s membrane. The subjacent layer of the cornea, its stroma or substantia propria, is formed by tightly packed collagen fibers that are uniquely organized in a parallel fashion affording the cornea its crystal clear disposition. The corneal stroma is devoid of blood and lymphatic vessels and is populated by fibroblasts. In addition to fibroblasts, the substantia propria is also endowed with a heterogeneous population of cells including bone-marrow-derived cells, and antigen presenting dendritic cells, most of which, under normal physiological conditions, are still immature and remain quiescent. Descement’s membrane, a thick lamina propria, separates the corneal stroma from its innermost cellular layer: the corneal endothelium, which consists of a single layer of low cuboidal cells. The corneal endothelium is critical for water homeostasis as it actively transports excess fluid from the corneal stroma into the anterior chamber. This peculiar histological organization of the cornea with a thin epithelium layer and an active endothelium allows oxygen from the room-air and nutrients from the aqueous humor to diffuse through its full thickness. Because the loss of avascularity in the cornea results in impaired light transmission and poor vision, these unique histological features evolved over time and bestowed the cornea with the ability to remain viable and clear in the absence of a direct blood supply.

Avascularity of the Cornea

479

EPI BM

STR

DM END Figure 1 Photograph of the human limbus showing the abrupt termination of the conjunctival vasculature in the interface between the cornea (C) and the conjunctiva (CJ). Dotted line delineates the limbus.

Corneal Avascularity and Optical Clarity The visual impairment associated with the loss of corneal avascularity is not only related to the physical obliteration caused by the opaque vessels within the visual field. Corneal neovascularization also reduces visual acuity because the infiltrating vasculature disrupts the tightly packed collagen bundles eliciting opacities, especially in areas surrounding the newly formed vessels. Vascular leakage and edema, usually associated with inflammatory neovascularization, overwhelms the endothelium drainage capacity, creates fluid accumulation, disrupts corneal clarity, and perturbs light transmission. Additionally, the neovascular corneas may also have diminished transparency due to lipid deposits. This is observed in vascularized corneas following corneal herpetic infections. Altogether, these observations speak of the tight correlation that exists between corneal avascularity, corneal transparency, and optimal optical performance.

Endogenous Anti-Angiogenic Mechanisms The molecular homeostatic mechanisms supporting the lack of vessels in the cornea were unknown until recently. Long ago, it was postulated that corneal avascularity was a

Figure 2 Photomicrograph of the human cornea stained with H&E showing its histological layers. The epithelial layer (EPI) overlaying Bowman’s membrane (dotted line). The corneal stroma (STR) displaying its tightly packed parallel collagen fibers. Descement’s membrane (DM) interfacing the substantia propria and the corneal endothelium (END).

passive process. It was thought that the cornea was avascular simply because pro-angiogenic forces were not present. It is quite intriguing, however, that the avascular cornea is surrounded by extremely vascularized tissues, such as the ocular conjunctiva and iris. Because pro-angiogenic factors are a requirement for endothelial cell survival one may postulate that the cornea must be armed with angiostatic capabilities in order to counteract the constant angiogenic stimuli that derives from the adjacent vascular beds. This alternate hypothesis has challenged the previous beliefs, became accepted as a working hypothesis, and still stands as a paradigm of modern vascular biology. Currently, it is well known that corneal avascularity is an extremely active phenomenon, requiring an exact balance between pro- and anti-angiogenic forces, and not the mere absence of pro-angiogenic stimulation. Extrinsic and intrinsic mechanisms have been proposed to underlie the absence of blood and lymphatic vessels in the cornea. The aqueous humor, fluid that circulates through the anterior chamber of the eye, has been regarded as a major extrinsic inhibitor of corneal angiogenesis. It contains several soluble angiostatic molecules including heparan sulfate proteoglycans. Because vascular endothelial growth factor (VEGF) and basic fibroblast growth factor (bFGF), the foremost studied angiogenic factors, bind to

480

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

these glycoproteins with relative high affinity, it has been proposed that the aqueous humor sequesters VEGF and bFGF from the cornea into the anterior chamber. However, it is important to bear in mind that not all isoforms of VEGF, like VEGF121, shown to be expressed in the cornea, bind heparan sulphate. This suggests that additional mechanisms are in place to secure corneal avascularity. In reality, it has long been suggested that corneal avascularity is a phenomenon supported by a redundant system. Additional factors contributing to corneal avascularity include the intrinsic synthesis of angiostatic molecules. It has been shown that the cornea expresses several of these anti-angiogenic factors, like angiostatin, endostatin, thrombospondins (TSPs), interleukin-1 receptor antagonist, pigmented epithelium-derived factor (PEDF), non vascular VEGF receptor-3 and soluble VEGF receptor-1. The current paradigm of intrinsic angiostatic mechanisms in the cornea derived from the clinical observation that corneal epithelial dysfunction was often associated with neovascularization. This suggested that the antiangiogenic powers of the cornea resided primarily in its epithelial layer. In fact, most of the anti-antiangiogenic factors identified in the cornea have been localized primarily to the corneal epithelium. Angiostatin, a by-product of plasminogen, was first isolated in 1994. Its angiostatic effect was initially demonstrated in a cancerous tumor growth and metastasis model. Mechanistically, angiostatin is thought to inhibit ATP synthase activity, block endothelial cell migration and proliferation and also cause vascular endothelial cell apoptosis. Angiostatin has been detected in healthy human corneal extracts and it has been shown that human corneal epithelium in culture is capable of converting exogenous plasminogen into its angiostatic by-product. Angiostatin has also been implicated in the maintenance of angiogenic privilege in the normal cornea and after wound healing. Endostatin was first described in 1997. It is a proteolytic fragment of the caroboxyterminus of collagen XVIII. It has been portrayed as a potent inhibitor of angiogenesis and tumor growth. Collagen XVIII is a component of the basement membrane ubiquitously expressed and it also has been shown to exist in ocular tissues. Immunohistochemical studies have localized collagen XVIII and endostatin to the corneal epithelium, principally in the basal epithelium. Endostatin has been shown to block VEGFinduced phosphorylation of VEGFR-2 and inhibit endothelial cell proliferation and migration. The synthesis of endostatin in the cornea is upregulated during injury and has been shown to inhibit injury-induced corneal neovascularization. Although endostatin has been shown in the uninflammed cornea, mice deficient in collagen XVIII have normal avascular corneas, suggesting that endostatin is only one of many factors contributing to the maintenance of corneal avascularity.

TSP-1 and -2 are potent anti-angiogenesis protein. TSP-1 directly inhibits the migration and survival of endothelial cells by activation transforming growth factor-beta (TGF-b). TSP-2 inhibits vascular endothelial cell proliferation independently of TGF-b activation. Both molecules have been detected in the mouse and human corneas under normal physiological conditions. Exogenous administration of TSP-1 and/or -2 have been associated with diminished suture-induced corneal angiogenesis. However, the systemic ablation of TSP-1 and -2 was not associated with spontaneous angiogenesis in the cornea, suggesting that alternate redundant angiostatic mechanisms are operative in the cornea. Interleukin (IL)-1 receptor antagonist, a key modulator of IL-1 activity, was shown to be expressed in the normal human cornea. It was localized to the corneal epithelium and some stromal fibroblasts. In a mouse model of sutureinduced corneal neovascularization IL-1 receptor antagonist was shown to have anti-angiogenic properties and its exogenous administration was also associated with diminished infiltration of inflammatory cells into the cornea. IL-1 receptor antagonist deficiency in mice had no bearing in corneal avascularity, once again implying that corneal avascularity is secured by a multifactorial and redundant system. VEGFR-3, normally expressed in lymphatic endothelial cells (LECs), corneal dendritic cells, and macrophages, was shown to be ectopically expressed in corneal epithelial cells of human and mice. These ectopic receptors have been shown to work as an inhibitor of injuryinduced corneal angiogenesis in mice. Since its discovery in 1993, soluble VEGFR-1 (sVEGFR-1) has been extensively studied as a powerful inhibitor of VEGF-induced angiogenesis. It has been implicated in several pathological states, including preeclampsia, sepsis, arthritis, and cancer. Interestingly, VEGF, a powerful driver of angiogenesis, is expressed in the normal avascular cornea. Recently, it has been shown that sVEGFR-1 is co-expressed by the corneal epithelium serving as a VEGF manacle. Because the systemic ablation of VEGFR-1 gene is not compatible with survival, corneal specific deletion of VEGFR-1 was employed in mice and led to spontaneous corneal neovascularization. sVEGFR-1 is considered to be singularly essential for maintaining corneal avascularity of the uninjured cornea. The expression of sVEGFR-1 in the cornea was also shown to be conserved among mammals, including humans. One exception is the Manatee, whose cornea is spontaneously vascularized and lacks sVEGFR-1 expression. A similar splice variant of VEGFR-2, sVEGFR-2, was recently identified and described as the first specific endogenous inhibitor of lymphangiogenesis. In the cornea, sVEGFR-2 was shown to be singularly essential to maintaining the cornea devoid of lymphatic vessels, as its genetic deletion cause spontaneous invasion of lymphatic, but not blood vessels into the cornea.

Avascularity of the Cornea

481

Loss of Corneal Avascularity

Corneal Transplant and Avascularity

Corneal neovascularization occurs when the precise equilibrium between pro- and anti-angiogenic forces is disrupted. The loss of corneal avascularity is pathological. Several ocular disorders are hallmarked by corneal neovascularization. These neovascular disorders of the cornea range from benign contact lens-associated neovascularization to congenital and hard to manage ocular anomalies such as aniridia. Extended contact lens wear may induce neovascularization because the associated hypoxia triggers a steep rise in VEGF expression, overwhelming the natural antiangiogenic barriers. Other corneal disorders are coupled with pathological angiogenesis because of direct damage to the corneal epithelium. Corneal trauma, such as corneal abrasions or chemical (alkali) burn of the ocular surface, infections (herpetic keratitis), immune diseases such as atopic keratoconjuctivitis or rheumatoid arthritis, limbal cell deficiency, or congenital anomalies such as aniridia are all associated with the loss of avascularity. These clinical entities commonly require surgical management, like refractive surgery or perforating keratoplasty, both of which are negatively impacted by the preexistence of vessels in the cornea. This represents a significant clinical predicament and speaks of the critical nature of unveiling molecular targets that may be manipulated to treat the aberrant and disordered growth of vessels into the cornea.

In 1905, ophthalmologist Edward Zim performed the first corneal transplant in a human subject. Since then, corneal transplants have become the most common type of solid tissue transplantation in the world. Nearly 46 000 corneal transplants are preformed yearly in the US. In addition to being the most prevalent, corneal allograft transplantation, it is also the most successful intervention among other commonly transplanted organs. However, the long-term outcome of this intervention is greatly influenced by pre-operative risk factors, with corneal neovascularization (high-risk group) being an important negative predictor of corneal allograft survival. While graft survival is approximately 90% in the low-risk group (no pre-operative inflammation or neovascularization), these numbers are drastically reduced to roughly 35% in the high-risk group. Recent studies targeting corneal angiogenesis with VEGF-A binding molecules (VEGF-trapW) demonstrated that allograft survival is inversely related to the amount of neovascularization in the murine corneal transplantation model corroborating the aforementioned epidemiological observations. The loss of corneal avascularity is therefore a significant clinical quandary. The surgical procedures used in corneal allograft transplantation require delicate techniques to prevent adverse inflammatory reactions which may compromise outcome. The corneal graft is attached to the recipient’s ocular surface with the placement of small sutures. Paradoxically, in a vastly employed injury model of corneal angiogenesis, similar intrastromal sutures are used as a method of eliciting blood and lymphatic vessel growth. Because suture placement is a requirement for corneal transplantation as well as a pro-angiogenic stimulus, it is critical to understand the molecular underpinnings related to the growth of blood and lymphatic vessels into the cornea, so potential molecular targets could be identified and manipulated to promote corneal avascularity, optimal optical performance, and prevent corneal blindness.

Immune Privilege of the Avascular Cornea The absence of blood and lymphatic vessels in the cornea is known to play a critical role in maintaining its immune privilege, but other immune-protective mechanisms have been described. One such mechanism is referred to as anterior chamber-associated immune deviation (ACAID). ACAID is regarded as the ability of antigen-presenting cells (APCs) and antigens from anterior chamber-associated tissues (i.e., cornea) to directly enter the blood circulation through the trabecular meshwork homing to the spleen where immune tolerance is induced. Additionally, tissues from the anterior segment of the eye have been reported to express Fas-ligand which induces apoptosis in activated immune cells (Fas-receptor positive), thus protecting the cornea from damage by stimulated lymphocytes. These mechanisms are thought to collectively downregulate inflammation in the cornea, thereby preserving corneal clarity which is essential for optimal vision. Corneal avascularity is therefore one factor of several redundant active mechanisms aimed at preserving corneal transparency and optical light transmission.

Corneal Alymphaticity and Allograft Rejection Because the growth of blood and lymphatic vessels into the cornea are intimately intertwined, the individual contribution of each of these vasculatures to the fate of corneal allografts is not clearly understood. However, recent evidence suggests that the growth of lymphatic vessels into the cornea may be more tightly associated with loss of corneal immune privilege and critical for corneal allograft rejection than corneal hemangiogenesis. Substantial progress in the study of corneal lymphangiogenesis has taken place since the discovery of VEGFR-3

482

Immune Regulation of the Cornea and Conjunctiva and Its Dysregulation in Disease

and its ligands VEGF-C and –D. The identification of specific cellular makers preferentially expressed by LECs, such as LYVE-1, Prox1, and Podoplanin, has also propelled great advances to the field of lymphangiogenesis. Corneal lymphangiogenesis generally occurs after corneal injury and inflammation, which in turn is associated with increased levels of VEGF-C. The newly formed corneal lymphatic vessels give rise to an afferent route through which corneal transudate and APCs are carried from the interstitial space into the lymphatic system and later back into the blood circulation. This drainage pathway becomes extremely deleterious in the context of corneal transplantation. Under these circumstances, the alternative route bypassing the standard outflow pathway (i.e., trabecular meshwork in the anterior chamber) allows for antigens from the donor cornea to escape through the lymphatic system and into the draining lymph node where a graft rejection reaction is initiated. The significance of this alternate drainage pathway to corneal alloimmunity and graft rejection has been portrayed in studies demonstrating that removal of cervical lymph nodes significantly increases the transplant survival rates in the low-(noninflamed, nonvascularized) and high-risk (neovascularized) groups. Together, these observations suggest that improved molecular understanding of corneal lymphangiogenesis, as well as the identification of endogenous compounds with the ability to uncouple lymphangiogenesis from hemangiogenesis would shed light into our current understanding of allograft rejection and potentially unveil therapeutic targets to enhance the survival of corneal allograft.

The Avascular Cornea as an Angiogenesis Study Platform The avascular disposition and its ready accessibility have made the cornea an important platform for the study of angiogenesis allowing scientists to test the pro-and/or anti-angiogenic effects of several compounds in vivo. Numerous assays have been developed to study angiogenesis modulation utilizing the cornea. Direct intra-stromal injection of angiogenesis compounds have been performed in the mouse, rat and rabbit cornea. Models in which a transient chemical (alkali), physical (scraping), or thermal (cautherization) injury are incurred to the cornea to provoke an angiogenic response have been widely used. Prolonged injury of the cornea has been achieved with intra-stromal suture placement. The insertion of a small pellet containing pro-angiogenic molecules has also been described and termed corneal micropocket assay. The cornea stroma has even been utilized for the implantation of tumor cells. The reliability of these models and the easy visualization of corneal vessels have placed the cornea in

the forefront of discovery and in vivo testing of drugs for the treatment of disorders hallmarked by aberrant angiogenesis, particularly cancer. The ability of analyzing these molecules in vivo provides valuable insight regarding the angio-modulatory effects of such compounds.

Conclusions Avascularity of the cornea is intimately related to optical transparency and optimal vision. Hence, the loss of corneal avascularity is pathological and often results in impaired vision or corneal blindness. A precise balance between pro- and anti-angiogenic factors is essential to maintain the avascular disposition of the cornea. Even though it is well known that corneal immune privilege is a function of a constellation of factors, the absence of blood and lymphatic vessels in the cornea has proven to be one of its most important underlying mechanisms. A more precise understanding regarding the individual contribution of blood and lymphatic vessel growth to corneal alloimmunity is needed. The cornea, given its avascular nature and accessibility, is an ideal the platform for the in vivo testing of angiogenesis modulators. See also: Corneal Angiogenesis; Tear Drainage.

Further Reading Albuquerque, R. J. C., Hayashi, T., Cho, W. G., et al. (2009). Alternatively spliced vascular endothelial growth factor receptor-2 is an essential endogenous inhibitor of lymphatic vessel growth. Nature Medicine 15: 1023–1030. Ambati, B. K., Nozaki, M., Signh, N., et al. (2006). Corneal avascularity is due to soluble VEGF receptor-1. Nature 443: 993–997. Azar, D. T. (2006). Corneal angiogenic privilege: Angiogenic and antiangiogenic factors in corneal avascularity, vasculogenesis, and wound healing (an American Ophthalmological Society thesis). American Ophthalmological Society 104: 264–302. Cursiefen, C. (2007). Immune privilege and angiogenic privilege of the cornea. Chemical Immunology and Allergy 92: 50–57. Cursiefen, C., Chen, L., Saint-Geniez, M., et al. (2006). Nonvascular VEGF receptor 3 expression by corneal epithelium maintains avascularity and vision. Proceedings of the National Academy of Sciences of the United States of America 103: 11405–11410. Folkman, J. (2007). Angiogenesis: An organizing principle for drug discovery? Nature Reviews Drug Discovery 6: 273–286. Hirsch, E., Irikura, V. M., Paul, S. M., et al. (1996). Functions of interleukin 1 receptor antagonist in gene knockout and overproducing mice. Proceedings of the National Academy of Sciences of the United States of America 93: 11008–11013. Krachmer, J., Mannis, M., and Holland, E. (2004). Cornea. Amsterdam: Mosby. Lawler, J. (2000). The functions of thrombospondin-1 and -2. Current Opinion in Chemical Biology 12: 634–640. O’Reilly, M. S., Boehm, T., Shing, Y., et al. (1997). Endostatin: An endogenous inhibitor of angiogenesis and tumor growth. Cell 88: 277–285.

Avascularity of the Cornea O’Reilly, M. S., Holmgren, L., Shing, Y., et al. (1994). Angiostatin: A novel angiogenesis inhibitor that mediates the suppression of metastases by a Lewis lung carcinoma. Cell 79: 315–328. Spencer, W. H. (1996). Ophthalmic Pathology: An Atlas and Textbook. Philadelphia, PA: W.B. Saunders.

483

Whitcher, J. P., Srinivasan, M., and Upadhyay, M. P. (2001). Corneal blindness: A global perspective. Bulletin of the World Health Organization 79: 214–221. Yamagami, S., Dana, M. R., and Tsuru, T. (2002). Draining lymph nodes play an essential role in alloimmunity generated in response to high-risk corneal transplantation. Cornea 21: 405–409.

This page intentionally left blank

IV. VISUAL ACUITY RELATED TO THE CORNEA AND ITS DISORDERS

This page intentionally left blank

Pupil P D R Gamlin and D H McDougal, University of Alabama at Birmingham, Birmingham, AL, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Accommodation – A change in the refractive power of the crystalline lens of the eye. Intrinsically photosensitive retinal ganglion cells (ipRGCs) – The ganglion cells expressing a photopigment, melanopsin, that is intrinsically light sensitive. Miosis – Pupillary constriction. Mydriasis – Pupillary dilation. Pupillary light reflex (PLR) – The constriction of the pupil that is elicited by an increase in illumination of the retina.

Advantages of a Mobile Pupil The normal human pupil can change diameter from 8 to 1.5 mm, which corresponds to approximately a 30-fold change in area and almost a 1.5-log unit change in retinal irradiance. Although the visual system can operate over a 10-log unit range of lighting levels through the process of adaptation, it can take several minutes for optimum sensitivity to return after an abrupt increase or decrease in retinal illumination. The rapid control of retinal irradiance by the iris allows the visual system to more quickly regain optimal sensitivity by dampening fast changes in ambient lighting levels and by requiring less retinal adaptation for a given change in environmental lighting levels. However, changes in pupil size affect not only retinal illumination, but also diffraction, optical aberrations, and depth of focus of the eye. These factors differentially affect visual performance and, given changing environmental lighting conditions and visual tasks, the nervous system continuously modulates pupil diameter for optimal visual performance. The diffraction of light rays by an aperture is a major limiting factor in the resolution of an image in any optical system. The amount of disruption in image quality caused by diffraction at a circular aperture decreases as the size of the opening increases. Therefore, as pupil diameter increases, there is decreased degradation in retinal image quality caused by diffraction. In contrast to diffraction, the image-degrading effects of optical aberrations increase as aperture diameter increases. Therefore, as pupil diameter increases, the degradative effects of optical aberrations

also increase, and offset the benefits gained by reduced diffraction at larger pupil diameters. Over the normal range of pupillary diameter, diffraction impacts image quality less than optical aberration, and the optimal pupil diameter is therefore approximately between 2 and 4 mm. Along with diffraction and optical aberrations, defocus is an important determinate of retinal image quality. Although the pupil does not refract or focus light, it influences the depth of field of the eye. Depth of field is the range of distance in depth in which objects appear to be in focus. For example, when one reads a book, the power of the crystalline lens of the eyes changes in order to bring the text on the page into focus through a process called accommodation. With the eyes accommodated on the book, all objects within a range in front of and behind the book will also appear in focus. This range is called the depth of field and it is primarily dependent both on viewing distance and pupil diameter. When the viewing distance is held constant, the depth of field increases with decreases in pupil diameter, and therefore the pupil diameter can affect the focus of the retinal image. Clearly, a mobile pupil allows the nervous system to optimize retinal irradiance, diffraction, ocular aberrations, and depth of focus despite differing conditions and visual tasks. For example, across a range of daylight (photopic) luminances, pupil size corresponds to that required for the highest visual acuity, and the maximal information capacity of the retinal image. On the other hand, under low light (scotopic) conditions in which poorer retinal image quality can be tolerated due to the lower resolution of rod photoreceptors, the pupil dilates sufficiently to maximize the retinal illumination. Further evidence for the optimization of pupil diameter for differing visual tasks is evident in the pupillary near response (PNR). When the viewing distance changes from far to near, the pupils constrict to increase the field of view and reduce the retinal image defocus. This compensates for the decrease in the effective field of view that naturally occurs when viewing distance decreases (see the section titled ‘Pupillary near response’ for more details).

Overview of the Pathways Controlling Pupil Diameter A summary diagram of the afferent, central, and efferent pathways controlling pupil diameter is shown in Figure 1. This figure shows the iris musculature innervated by autonomic efferents from both the parasympathetic and sympathetic components of the autonomic nervous system

487

488

Visual Acuity Related to the Cornea and Its Disorders

Edinger – Westphal nucleus

Ciliary ganglion Retinal ganglion cell

Oculomotor nerve

Optic chiasm

Pretectal olivary nucleus

Sphincter pupillae

Dilator pupillae Superior cervical ganglion

Ciliospinal center Figure 1 Anatomical drawing showing the direct and consensual pupillary light reflex (PLR) pathways and the parasympathetic and sympathetic innervation of the iris in primates. The bilateral projection from the retina to the pretectum is also shown. The pretectal olivary nucleus receives input from the temporal retina of the ipsilateral eye and the nasal retina of the contralateral eye. The pretectal olivary nucleus projects bilaterally to the Edinger–Westphal (EW) nucleus, which contains parasympathetic, preganglionic, and pupilloconstriction neurons. The axons of these preganglionic neurons travel in the third cranial nerve to synapse upon postganglionic pupilloconstriction neurons in the ciliary ganglion. The axons of these postganglionic neurons leave the ciliary ganglion and enter the eye through the short ciliary nerves, and travel through the choroid to innervate the sphincter muscle of the iris. The sympathetic preganglionic pupillodilation neurons are found at the C8-T1 segmental levels of the spinal cord. The axons of these neurons project from the spinal cord through the dorsal roots and enter the sympathetic trunk, and then project rostrally to the superior cervical ganglion where they synapse with the postganglionic neurons. These postganglionic neurons project from the superior cervical ganglion through the neck and carotid plexus, and into the orbit of the eye. These fibers enter the eye either by passing through the ciliary ganglion and entering the short ciliary nerves, or bypassing the ciliary ganglion and entering via the long ciliary nerves (for clarity, only one of these alternative pathways is shown). Upon entering the eye, these axons travel through the choroid and innervate the dilator muscle of the iris. From McDougal, D. H. and Gamlin, P. D. R. (2008). Pupillary control pathways. In: Basbaum, A. I., Kaneko, A, Shepherd, G. M., et al. (eds) The Senses: A Comprehensive Reference, Vol 1: Vision 1, pp. 521–536. San Diego, CA: Academic Press.

(ANS). The parasympathetic component of the ANS innervates the sphincter pupillae muscle of the iris. The preganglionic parasympathetic fibers controlling the sphincter pupillae originate from neurons in the Edinger–Westphal (EW) nucleus, the autonomic subdivision of the third cranial nerve nucleus, and travel through the third cranial nerve to the ciliary ganglion, which is located within the orbit of the eye (see Figure 1). Within the ciliary ganglion, the preganglionic pupilloconstriction neurons form cholinergic, nicotinic synapses with the postganglionic neurons.

The axons of these postganglionic neurons leave the ciliary ganglion to enter the eye through the short ciliary nerves and travel to the iris. Here, they release acetylcholine, which acts on the muscarinic receptors of the sphincter pupillae (see Figure 2). The sympathetic component of the ANS innervates the dilator pupillae muscle. The preganglionic sympathetic neurons, which control pupillary dilation, are located in the C8-T1 segments of the spinal cord, a region termed the ciliospinal center of Budge (and Waller). The

Pupil

Anterior border Stroma

489

Pupillary ruff

Iris root Dilator pupillae

Sphincter pupillae

Pigmented epithelium

Ciliary process

Figure 2 Low-power photomicrograph of a cross section of the macaque iris. Scale ¼ 200 mm. From McDougal, D. H. and Gamlin, P. D. R. (2008). Pupillary control pathways. In: Basbaum, A. I., Kaneko, A, Shepherd, G. M., et al. (eds.) The Senses: A Comprehensive Reference, Vol 1: Vision 1, pp. 521–536. San Diego, CA: Academic Press.

axons of these preganglionic neurons project to the sympathetic chain and travel in the sympathetic trunk to the superior cervical ganglion. Within the superior cervical ganglion, the preganglionic axons form nicotinic, cholinergic synapses with postganglionic pupillodilation neurons. The axons of these postganglionic neurons project from the superior cervical ganglion to the orbit, where they enter the eye through the short and long ciliary nerves and travel to the iris (see Figure 1). Here, they release norepinephrine, which acts on the adrenoreceptors of the dilator muscle (see Figure 2).

Iris Musculature In a cross-section of the iris, the sphincter pupillae can be seen as an annular band of smooth muscle (100–170 mm thick; 0.7–1.0 mm wide) encircling the pupil (Figure 2). The sphincter, which is located in the posterior iris immediately anterior to the pigmented epithelium, interdigitates with the surrounding stroma and connects to the dilator muscle fibers. The smooth muscle cells of the sphincter are clustered in small bundles and connected by gap junctions. These gap junctions ensure synchronized contraction of the sphincter muscle. The sphincter receives muscarinic, cholinergic innervation from the short ciliary nerves – parasympathetic, postganglionic fibers arising from the ciliary ganglion. The dilator pupillae is composed of radially oriented smooth muscle fibers that are myoepithelial in origin. Individual fibers are approximately 50 mm long and 5–7 mm wide. In the pupillary zone, dilator muscle processes fuse with the sphincter pupillae, while peripherally, their processes attach to the ciliary body. Contraction of the dilator muscle pulls the pupillary margin toward the ciliary body.

Pupillary Light Reflex Description The pupillary light reflex (PLR) is the constriction of the pupil that is elicited by an increase in illumination of the retina. The direct PLR, present in virtually all vertebrates, is the constriction of the pupil in the same eye as that stimulated with light. The consensual PLR is the constriction of the pupil in the eye opposite to the eye stimulated with light. In mammals with laterally placed eyes, such as the rat and rabbit, the direct PLR is more pronounced than the consensual PLR. However, in those mammalian species with frontally placed eyes, such as humans and monkeys, the direct and consensual PLRs are essentially equal. An example of a human consensual PLR produced by two different wavelengths of light is shown in Figure 3. The PLR has traditionally been divided into two separate pathways based on the clinical manifestations of the defects in this reflex. The afferent pathway is composed of both the retinal cells that project to the pretectum as well as their recipient neurons, which project bilaterally to the EW nucleus (Figure 1). The efferent pathway is composed of the preganglionic pupilloconstriction fibers of the EW nucleus and their postganglionic recipient neurons in the ciliary ganglion, which project to the sphincter muscle of the iris (Figure 1).

Afferent Pathway The first neurons in the afferent pathway of the PLR are retinal ganglion cells. It has recently been recognized that this reflex in rodents and primates is driven predominantly by a unique subset of intrinsically photosensitive retinal ganglion cells (ipRGCs) which project to the pretectal olivary nucleus (PON), a small nucleus in the

490

Visual Acuity Related to the Cornea and Its Disorders

8 Light

7

Pupil diameter (mm)

6 5 4 3 2 1 0 0

5

10

15

20 25 Time (s)

30

35

40

45

Figure 3 Pupilloconstriction elicited by a 10-s light stimulus of 493-nm wavelength light at 14.0 log quanta cm–2 per second irradiance (blue trace), and 613-nm wavelength light at 14.1 log quanta cm–2 per second irradiance (red trace). Note that a 473-nm stimulus, which effectively activates the intrinsic photoresponse of intrinsically photosensitive retinal ganglion cells (ipRGCs), drives a larger pupillary response than the 613-nm stimulus (red trace), which does not effectively activate the intrinsic photoresponse of ipRGCs at this irradiance level. Note that the pupilloconstriction induced by the 473-nm light is maintained following stimulus offset. From McDougal, D. H. and Gamlin, P. D. R. (2008). Pupillary control pathways. In: Basbaum, A. I., Kaneko, A, Shepherd, G. M., et al. (eds.) The Senses: A Comprehensive Reference, Vol 1: Vision 1, pp. 521–536. San Diego, CA: Academic Press.

pretectum; the pretectum is located in the dorsal lateral aspect of the midbrain at the level of the superior colliculus (see Figure 1). Intrinsically photosensitive retinal ganglion cells

Prior to 2000, it was assumed that the PLR was driven by retinal ganglion cells which received light signals exclusively from rod and cone photoreceptors, which up to that time were the only known photoreceptive cells in the retina. However, recent studies have demonstrated that the PLR is driven predominantly by retinal ganglion cells which, unlike any other retinal ganglion cell class, are intrinsically photosensitive. The intrinsic photoresponse of these neurons, which is mediated by the photopigment melanopsin, presumably compensates for the adaptation of rod and cone photoreceptors, and serves to maintain pupilloconstriction during steady-state exposure at all photopic (daylight) illuminance levels. In addition to their intrinsic light-driven signal, it is clear that ipRGCs receive rod and cone inputs. In response to a pulse of light, intracellular recordings from these cells show a characteristic transient burst of neural activity at stimulus onset, which rapidly decays to a plateau of sustained activity that

often extends well past stimulus offset. The initial burst of neural activity is mediated by a rapidly adapting conemediated photoresponse, while the sustained activity is driven predominantly by the intrinsic response of these cells, although there is growing evidence for a rod contribution to this sustained activity under steady-state lighting conditions. IpRGCs project to the PON of rodents and primates, and they play a major role in pupillary responses. Monkeys and rodents with nonfunctional rod and cone photoreceptors but functional melanopsin-containing ipRGCs display a PLR; however, the reflex has a higher irradiance threshold than normal. Mice with ipRGCs lacking melanopsin also display PLR, but their pupils fail to constrict maximally in bright lights. Taken together, these results show that both the intrinsic photoresponse of ipRGCs and their classical photoreceptor inputs provide signals of retinal irradiance that drive the PLR. Additional studies further suggest that the influence of rod and cone photoreceptors on the pupillary light reflex is mediated exclusively through their inputs to ipRGCs. The intrinsic photoresponse of ipRGCs can also affect pupillary behavior in the absence of ongoing light stimulation. As noted above, ipRGCs encode stimulus irradiance through an elevation of firing rate that continues well beyond stimulus offset. Indeed, bright light stimuli can produce a prolonged pupillary constriction in humans that can persist up to 20 min after the light has been extinguished (see Figure 3). Experiments in primates, including humans, demonstrate that this prolonged pupillary constriction in darkness is mediated almost entirely by the intrinsic photoresponse of ipRGCs. Pretectal olivary nucleus The first relay in the afferent pathway of the PLR consists of luminance neurons within the PON, which receives direct retinal input. PON luminance neurons are characterized by tonic firing rates that increase with increases in retinal illuminance. In primates, these neurons exhibit a transient burst of activity followed by sustained tonic activity in response to increases in retinal illuminance. In addition, the tonic firing rate of these cells is proportional to retinal illuminance over at least a 3 log unit range of stimulus intensities in primates and in rats. Electrical microstimulation of the PON in rats and monkeys elicits pupilloconstriction at short latencies, and lesions of the PON in rats produce deficits in pupillomotor function. These results strongly suggest that luminance neurons within the PON mediate the PLR. In addition to retinal afferents, the PON also receives significant cortical, ventral thalamic, and midbrain inputs which may also have an influence on the PLR or other pupillary movements. Owing to its importance for the PLR, the best-described efferent projection of the PON is to the EW nucleus. However, the PON has been shown to project to a number

Pupil

of other targets, such as the hypothalamus, pons, and medulla that may also influence pupillary behavior. Efferent Pathway The efferent leg of the PLR begins with preganglionic pupilloconstriction neurons of the EW nucleus that project through the third cranial nerve to the ciliary ganglion (see Figure 1). The EW nucleus is a distinct nucleus of the midbrain, lying immediately dorsal to the oculomotor complex. It is located just ventral and lateral to the cerebral aqueduct at the level of the superior colliculus (see Figure 1). Evidence for the course of the efferent parasympathetic pupillary pathway and the importance of the EW nucleus in pupilloconstriction comes from electrical stimulation studies in the vicinity of EW nucleus that elicit pupilloconstriction in a variety of animal models. Within the ciliary ganglion, which is approximately 3 mm in size, and located 2–3 mm posterior to the globe and lateral to the optic nerve, the axons of the preganglionic neurons synapse with the postganglionic pupilloconstriction neurons. The axons of these postganglionic neurons leave the ciliary ganglion to enter the eye via the short ciliary nerves to innervate the sphincter muscle of the iris. Sympathetic Influences on the PLR It is generally agreed that the parasympathetic pathway discussed above is the primary route of pupillary constriction associated with the PLR. However, there is some evidence that increases in retinal illumination may cause a reduction in the tone of the dilator muscle of the iris through the sympathetic pathway outlined in Figure 1, and thus enhance the PLR. Studies in cats have shown a light-induced inhibition of postganglionic pupillodilation fibers at the level of the long ciliary nerves, and preganglionic, pupillodilation fibers at the level of the cervical sympathetic nerve. These studies found that the pupillodilation fibers were inhibited by light in an intensitydependent manner, that is, a more intense light brought about a greater inhibition in firing rate. However, these findings have not been replicated in primates, in which the evidence suggests that the sympathetic system does not contribute to the dynamics of the PLR and only contributes to tonic modulation of pupil diameter.

The Pupillary Near Response Description The PNR is the pupillary constriction associated with a change in viewing distance from far to near that occurs in primates including humans. When the eyes move from viewing a far object to viewing a near object, three

491

oculomotor responses occur. The eyes converge to bring the image of the object onto the fovea of each retina, the refractive power of the crystalline lens is adjusted to bring the image of the object into focus on the retina, and the pupil constricts. These collective processes are classically referred to as the near response or the near triad. Efferent Pathway of the PNR The PNR is thought to be driven solely by an increased drive to the sphincter muscle of the iris through the parasympathetic efferent pathway. Therefore, the neural control pathway of the PNR shares a common efferent pathway with the PLR, although the afferent inputs responsible for the PNR are more complex. The neural signals driving these two reflexes most likely converge at the EW nucleus, since the activity of PON luminance neurons is not correlated with pupil constriction during near viewing. Further, certain clinical neurological conditions are characterized by an intact PNR despite the absence of the PLR (light-near dissociation). It is generally accepted that preganglionic neurons in the EW nucleus drive the PNR as well as the PLR. However, it has not been determined if separate subpopulations of neurons exist in EW nucleus devoted exclusively to either the PLR or the PNR, or whether the same population of neurons drives pupillary constriction in both reflexes, although the latter seems most likely. Afferent Influences on the PNR Early investigations attempted to determine whether the PNR was driven primarily by ocular convergence or accommodation, the other two components of the near triad. Some studies found that the PNR was more closely associated with accommodation than with convergence. Other studies found a greater association with convergence, and even reported that the PNR was totally absent during some blur-driven accommodative responses. These conflicting results are likely a product of an incomplete disassociation between the convergence and accommodation systems during these experiments, as these two systems have been shown to be highly interdependent. A more modern view of the afferent influences controlling the PNR has recently emerged. In this view, the PNR is not seen as resulting from either accommodation or convergence alone, but as a separate output of the neural pathways that drive both accommodation and convergence. A number of brain areas play a role in controlling the near triad. These include cortical areas, such as extrastriate cortex, parietal cortex, frontal eye fields, as well as the cerebellum and the midbrain. Of particular interest to the PNR, is the supraoculomotor area of the midbrain, which lies just dorsal and lateral to the oculomotor nucleus. The supraoculomotor area contains near response cells which

492

Visual Acuity Related to the Cornea and Its Disorders

are modulated by both vergence and accommodation. These cells project to medial rectus motoneurons, and thus contribute to vergence eye movements. It seems likely that these cells also project to EW nucleus and are responsible for carrying the signal from the accommodation and convergence controller to the preganglionic, pupilloconstrictor neurons.

Additional Cortical Influences on Pupillary Responses In addition to cortical afferents mediating the PNR, the pupil is also influenced by both visual and nonvisual cortical regions. These afferents manifest themselves as small changes in pupil diameter during presentation of visual stimuli such as colored stimuli and gratings, as well as nonvisual stimuli such as auditory tones, and even during higher-order cortical functions such as problem solving. These observations provide clear evidence that cortex exerts an influence on pupillary behavior, which therefore cannot be thought of as entirely reflexive in nature. Visually Mediated Cortical Influences on Pupillary Behavior Small pupillary constrictions have been shown to occur in both human and monkeys with the presentation of complex visual stimuli, even when the stimuli do not involve a change in viewing distance or retinal illuminance. Changes in stimulus attributes such as color, spatial frequency, or apparent motion produce such cortically mediated pupillary responses. Deficits in these pupillary responses are observed in humans with lesions to cortical areas involved in processing one or more of these stimulus characteristics. In addition, lesions of rostral inferior temporal cortex but not V4 in macaques abolish pupillary responses to chromatically modulated gratings. Task-Evoked Pupillary Responses In the early 1960s, Hess and colleagues published a series of papers which reported modulations in human pupillary diameter associated with complex cognitive processes such as subjective attitudes or mental activity. Later studies failed to replicate the findings relating pupillary dynamics to subjective attitudes, although the findings related to mental activity have been replicated and extensively studied. The small pupillary dilations associated with increased mental activity, or task-evoked pupillary responses (TEPRs), have now become a well-established tool of cognitive psychology. These pupillary responses are generally reported to vary in magnitude from 0.2 to 0.7 mm, and have been shown to correlate with cognitive

load across diverse functions, such as sensory perception, memory, language, and attention. TEPRs have been repeatedly shown to monotonically vary with the degree of mental activity required by a task as measured by other objective criterion such as reaction time and the extent of cortical activation indicated by positron emission tomography (PET) scan, and this has allowed TEPRs to be utilized successfully to empirically test theories of language processing and intelligence. Although the behavioral phenomenon of TEPRs has been extensively studied and quantified, little is known of the underlying neurophysiology that drives these responses. It has been suggested that they may be driven by noradrenergic projections from the locus ceruleus since the activity of neurons in this nucleus has been show to correlate with both pupil diameter and task-related events.

Influence of Alertness on Pupillary Behavior Since the muscles of the iris are controlled by the ANS, environmental or physiological conditions which cause changes in overall autonomic function can have a significant effect on pupillary behavior. Even though the environment or physiological conditions which produce the change in autonomic tone may not have a direct influence on the visual system, they may still manifest themselves through an affect on pupil diameter. Arousal Situations or stimuli which produce an emotional or startle response often produce a profound pupillary dilation. This effect is mediated through the hypothalamus, the brain area responsible for the integration of autonomic function. This integration allows for the coordination of the various functions of the ANS and often leads to global changes in the balance between the sympathetic and parasympathetic branches of the ANS. For example, an unexpected loud noise may produce a startle response which is characterized by increases in heart rate, respiratory rate, and pupil diameter; it is caused by a systemic increase in sympathetic tone mediated through the hypothalamus. This global increase in sympathetic tone can affect pupil diameter via activation of the pupillodilation centers of the spinal cord and inhibition of the pupilloconstriction neurons of EW nucleus. Neurons within the hypothalamus project to the sympathetic preganglionic pupillodilation neurons of the thoracic spinal cord. This direct effect of hypothalamic activation on pupil diameter can be shown through microstimulation of the posterior hypothalamus, which often causes rapid pupil dilation. Increase in sympathetic tone can also produce inhibition

Pupil

of pupilloconstriction neurons of EW nucleus via the influence of ascending neuromodulatory pathways. The hypothalamus is also the site at which autonomic function is regulated by the central nervous system through connections with the limbic system and cortical structures. The limbic system of the brain, which is responsible for emotions and short-term memory, has a direct connection to the hypothalamus and therefore can have significant effects on autonomic balance. Situations or stimuli which produce an intense emotional response are often accompanied by pupillary dilation, which is certainly mediated through limbic connections to the hypothalamus. In addition, cortical influences on the hypothalamus allow a wide variety of stimuli to effect autonomic tone and thus pupil diameter.

493

affect pupillary behavior in a wide range of animal models and human studies. These differential effects are most likely due to the both interspecies variability in the projections of these neuromodulatory fibers, as well as the differential activation of multiple brain areas implicated in pupillary behavior due to the extensive projections of these neuromodulators.

Acknowledgments This work was supported by NIH grant EY09380 and the EyeSight Foundation of Alabama. See also: Acuity.

Sleep Sleep has a pronounced effect on the ANS, specifically a reduction in sympathetic outflow and an increase in parasympathetic outflow. Given this overall trend, it is not surprising that pupillary behavior during sleep is characterized by prolonged constriction of the pupil. It has been shown that sleep-induced pupillary constriction persists in animals with lesions of the preganglionic sympathetic pupillodilation fibers. This suggests that the sleepinduced pupillary changes are mediated by an activation of the preganglionic parasympathetic pupilloconstriction fibers of the EW nucleus. Ascending Neuromodulatory Systems The ascending neuromodulatory systems of the midbrain and brainstem can have a variety of effects on pupillary behavior. These nuclei are the origin of neuromodulatory fibers which release dopamine, norepinephrine, histamine, and serotonin at a number of brain areas implicated in pupillary control. These neuromodulatory systems appear to be critical in the regulation of sleep and arousal, as well as autonomic regulation and cortical plasticity. In addition to these global neuromodulatory effects, all or some of which could have a profound influence on pupillary behavior, there is some evidence for a direct inhibition of pupilloconstriction neurons in the EW nucleus by adrenergic neurons originating from the locus ceruleus in a number of animal models. However, other studies in humans and rabbits have failed to find this direct noradrenergic inhibition of EW nucleus pupilloconstriction neurons and it has been suggested that this effect might be mediated by dopaminergic neurons in these species. Drugs which agonize or antagonize these neuromodulatory neurotransmitters have been found to differentially

Further Reading Barbur, J. L. (2003). Learning from the pupil – studies of basic mechanisms and clinical applications. In: Chalupa, L. M. and Werner, J. S. (eds.) The Visual Neurosciences, pp. 641–656. Cambridge, MA: MIT Press. Beatty, J. and Lucero-Wagoner, B. (2000). The pupillary system. In: Cacioppo, J. T., Berntson, G., and Tassinary, L. G. (eds.) Handbook of Psychophysiology, 2nd edn., pp. 142–162. Cambridge: Cambridge University Press. Berson, D. M. (2003). Strange vision: Ganglion cells as circadian photoreceptors. Trends in Neuroscience 26: 314–320. Bron, A. J., Tripathi, R. C., Tripathi, B. J., and Wolff, E. (1997). Wolff’s Anatomy of the Eye and Orbit. London: Chapman and Hall Medical. Busettini, C., Davison, R. C., and Gamlin, P. D. R. (2009). Vergence eye movements. In: Squire, L. (ed.) Encyclopedia of Neuroscience, vol. 10, pp. 75–84. Oxford: Elsevier. Charman, W. N. (1995). Optics of the eye. In: Bass, M. (ed.) Handbook of Optics, pp. 24.3–24.54. New York: McGraw-Hill. Gamlin, P. D. (2000). Functions of the Edinger–Westphal nucleus. In: Burnstock, G. and Sillito, A. M. (eds.) Nervous Control of the Eye, pp. 117–154. Binghamton, NY: Harwood Academic. Gamlin, P. D. (2005). The pretectum: Connections and oculomotorrelated roles. Progress in Brain Research 151: 379–405. Gamlin, P. D., McDougal, D. H., Pokomy, J., et al. (2007). Human and macaque pupil responses driven by melanopsin-containing retinal ganglion cells. Vision Research 47(7): 946–954. Kardon, R. H. (2005). Anatomy and physiology of the autonomic nervous system. In: Miller, N. R., Walsh, F. B., Biousse, V., and Hoyt, W. F. (eds.) Walsh and Hoyt’s Clinical Neuro-Ophthalmology, vol. 3, pp. 649–714. Baltimore, MD: Lippincott Williams and Wilkins. Kawasaki, A. (2005). Disorders of pupillary function, accommodation, and lacrimation. In: Miller, N. R., Walsh, F. B., Biousse, V., and Hoyt, W. F. (eds.) Walsh and Hoyt’s Clinical Neuro-Ophthalmology, vol. 3, pp. 739–804. Baltimore, MD: Lippincott Williams and Wilkins. Loewenfeld, I. E. and Lowenstein, O. (1993). The Pupil: Anatomy, Physiology, and Clinical Applications. Ames, IA: Iowa State University Press. McDougal, D. H. and Gamlin, P. D. R. (2008). Pupillary control pathways. In: Basbaum, A. I., Kaneko, A., Shepherd, G. M., et al. (eds.) The Senses: A Comprehensive Reference, Vol 1: Vision 1, pp. 521–536. San Diego, CA: Academic Press. Oyster, C. W. (1999). The Human Eye: Structure and Function, pp. 411–446. Sunderland, MA: Sinauer Associates.

Acuity M D Crossland, UCL Institute of Ophthalmology/Moorfields Eye Hospital, London, UK ã 2010 Elsevier Ltd. All rights reserved.

Glossary Cycles per degree – The number of complete phases of a grating (e.g., the distance between the center of a white bar and the center of the next bright bar in a square-wave grating; or the distance between two adjacent areas of maximum brightness on a sine-wave grating) contained in 1 of visual angle. Minimum angle of resolution – The size of the angle subtended at the eye of the smallest feature which can be reliably identified on an optotype. Minute of arc – One-sixtieth of a degree. Optotype – A letter, symbol, or other figure presented at a controlled size to measure vision. Visual angle – The angle, which a viewed object subtends at the eye.

Detection and Resolution Acuity Visual acuity can be defined in two broad ways. Detection acuity is measured by determining the size of the smallest object which can be reliably seen (is there a circle on the first or second screen?). Detection can be elicited reliably with targets, which subtend an angle at the eye as small as 1 s of arc (1/3600 ). Even a small point of light will stimulate several photoreceptors due to the point-spread function of the eye: that is, the way in which light is diffracted through the eye’s optics (Figure 1(a)). Tests that require the identification of a target are a measurement of resolution acuity. These tests frequently involve identifying a letter or reporting an object’s orientation (what direction is this letter C facing?). Acuity for these tests depends on the separation of the target features: if they are too close, the point-spread function from each element will overlap and they will not be identified (Figure 1(b)). The smallest separation of the elements required for identification of the target (Figure 1(c)) is known as the minimum angle of resolution (MAR). For an adult observer with good vision, a typical MAR for a centrally presented, high-contrast target can be as good as 30 s of arc (1/120 ). Figure 2 shows the feature critical for the MAR for some commonly used tests of visual acuity.

494

Measurement of Visual Acuity Visual acuity tests have been used for millennia: the ancient Egyptians are reported to have used discrimination of the twin stars of Mizar and Alcor as a measurement of vision. The most familiar clinical test of visual acuity, the Snellen chart, was introduced in 1862, and is still widely used today. Detection acuity is often measured psychophysically by means of a temporal two-alternative forced-choice experiment (did the light appear in the first or the second interval?). Detection acuity is rarely measured clinically. In psychophysical experiments of the visual system, resolution acuity is commonly measured by asking observers to report the orientation of a grating with variable separation between each dark and light bar (Figure 2(b)). In clinical practice, gratings are rarely used, with the exception of forced-choice preferential looking tests in preverbal children. These tests consist of a uniform gray field with an isoluminant grating toward one side of the chart (Figure 3(a)). In a featureless room, the test is presented to the child and the clinician observes whether the child looks toward the grating. The finest grating toward which the child repeatedly looks is recorded as the visual acuity. For cooperative patients, optotypes are more often used to measure clinical resolution acuity. The Landolt C (Figure 2(c)) is the standard to which letter visual acuity tests are compared. This target consists of a ring of fixed width with a gap, of height equal to the stroke width, at the top, left, right, or bottom of the circle. The observer is asked to report the position of this gap. The smallest gap whose position can be reliably reported is equivalent to the MAR. The National Academy of Sciences standard for visual acuity measurement advocates the presentation of 10 optotypes, of equivalent difficulty to the Landolt C, at each acuity size. The horizontal spacing between each optotype should be at least one character width, and vertical spacing between lines should be 1–2 times the height of the larger optotypes. It suggests that the number of characters on each line should be equal, and that the size difference between consecutive lines is 0.1 log units: in other words, for each target size, the next line should be approximately 1.26 times smaller. The Snellen chart (Figure 3(b)) does not meet these recommendations: the number of letters per line and step size between the lines are variable, as is the horizontal and

Acuity

495

Target

(a)

(b)

(c) (a)

2-D PSF

1-D PSF

(b)

Figure 1 Schematic illustration of the point-spread function of three visual targets: (a) a point target; (b) two adjacent lines, too close to be resolved; and (c) two adjacent lines, with sufficient separation to be resolved. Middle row: two-dimensional representation of the target point-spread function; bottom row: one-dimensional representation of the point-spread function; and red line indicates the sum of energy incident on the retina. PSF, point-spread function.

θ

(a) θ

(b)

θ

(c)

θ

(d)

Figure 2 Examples of the limiting feature for four commonly used resolution tasks: (a) two-point discrimination task; (b) grating; (c) Landolt C; and (d) Sloan letter E (note that white gap size is equal in width to black bar elements).

vertical spacing on the chart. There is also a marked difference in the legibility of different letters on the Snellen chart: a W, for example, has far less separation between the elements of the letter and is more difficult

(c)

Figure 3 (a) A forced-choice preferential looking test consisting of a grating against an isoluminant background. Note the peephole in the center for the clinician to observe the child’s visual behavior; (b) the Snellen chart; and (c) The ETDRS chart. ETDRS, Early treatment of diabetic retinopathy study.

to identify than a letter L. In the 1950s, Sloan suggested the use of 10 letters with a selection of vertical, horizontal, oblique, and round strokes which are each about as legible as a Landolt C. These Sloan letters are C, D, H, K, N, O, R, S, V, and Z. Each of the Sloan letters has a stroke width of the MAR and has a total height and width of five times the MAR. The Bailey–Lovie chart, introduced before the recommendations of the National Academy of Sciences, conforms to most of these requirements, although it only has five letters per line. Further, the letters on the Bailey–Lovie chart are taller than they are wide: their height-to-width ratio is 5:4 and they are selected from the British Standards set of letters (D, E, F, H, N, P, U, V, R, and Z). The ETDRS chart (Figure 3(c)), developed for the early treatment of diabetic retinopathy study (ETDRS), is similar in design but does use the recommended 5 5 Sloan letters. A criterion of 7/10 letters being read correctly for a line to be marked as seen was suggested by the National Academy of Sciences. This threshold reduces the chance of the line being scored correctly by chance (by a blind observer) to around 1 in 9 000 000. On a chart with five letters per line, recording a visual acuity where four of the five letters are read correctly equates to a chance success rate of 1 in 46 000. There is a theoretical advantage if the observer knows there are only 10 letters which can be presented on the chart: if an observer guesses from all 26 letters rather than the ten Sloan letters, the probability of the observer getting four out of five letters correct reduces to about 1 in 100 000.

496

Visual Acuity Related to the Cornea and Its Disorders

Test–retest variability of the Snellen chart is around 0.3 logMAR, while the ETDRS chart has far better repeatability (test–retest variability 0.1–0.2 logMAR). Despite the many limitations of the Snellen chart, it is still widely used in clinical practice. While this is likely to be largely due to clinicians’ familiarity with the Snellen chart, there is also a perception that Snellen acuity measurement is quicker than that on the Bailey–Lovie or ETDRS charts. Various modified versions of the ETDRS chart exist: for example, a version with an altered letter set (A, B, E, H, N, O, P, T, X, and Y) has been developed for use by readers of most European languages, including those based on Cyrillic or Hellenic alphabets. For observers unable to report letters on a sight chart, other frequently used optotypes include the tumbling E chart (formerly and less politically correctly known as the illiterate E chart), where a letter E is shown in each of four rotations; the HOTV chart, where only these four letters are used; symbols such as the Lea or Kay pictures; and simple shapes, such as the Cardiff card.

In much of Europe, the Snellen fraction is reduced into a decimal fraction. A further confusion with the Snellen system is that in countries not using the metric system, distances are expressed in feet rather than meters, with 20/20 being exactly equivalent to 6/6 but with a test distance of 20 ft rather than 6 m. Although Snellen recommended adoption of the metric system in 1875 and, in 1980, the US National Academy of Sciences favored adoption of a standard defined in meters, given the imminent adoption of the metric system, the feet system is still widely used in the USA, and among lay people in the UK. The accepted standard for expressing visual acuity in clinical research, and increasingly in clinical practice, is to use the base 10 logarithm of the MAR (logMAR), such that 0.0 logMAR is equivalent to 6/6 or 20/20, and 1.0 logMAR is the same as 6/60 or 20/200. Table 1 gives approximately equivalent values in MAR, cycles per degree, Snellen fractions in meters and feet, decimal acuity, and logMAR for a range of visual acuities.

Reporting Visual Acuity

Optical and Neural Limits on Visual Acuity

Clinicians have traditionally used Snellen fractions to record visual acuity, where the numerator is the test distance and the denominator the target size. The target size is expressed, counterintuitively, as the distance from which the target has an MAR of 1 min of arc. Therefore, a visual acuity of 6/6 indicates that from 6 m, letters with MAR 1-min arc are correctly identified, while a visual acuity of 3/36 indicates that from 3 m, the targets identified have a MAR of 1 min of arc when viewed from 36 m. The reciprocal of the Snellen fraction gives the visual acuity in MAR: so a visual acuity of 3/36 indicates a MAR of 12 min of arc.

Visual acuity is limited by many factors: the optics and refraction of the eye; the clarity of the optical media; the spacing and function of the retinal photoreceptors; the ratio of retinal ganglion cells to photoreceptors; and the resolution of the primary visual cortex and higher areas of visual processing. Each diopter of myopia reduces visual acuity: a –1.00DS myope will typically have uncorrected visual acuity of around 0.5 logMAR (6/18; 20/60) and a two-diopter myope will have vision of around 0.8 logMAR on a distance test. Hypermetropia can often be relieved by accommodation in young people, but each diopter of hypermetropia

Table 1

Visual acuity conversion table

a

MAR (min)

Cycles/ degree

Snellen (metric)

Snellen (feet)

Decimal

Log MAR

60 20 10 6.3 4 3.2 2 1.6 1.3 1 0.83 0.67 0.5 0.33

0.5 1.5 3 4.7 7.5 9.4 15 18.8 23 30 36 44 60 91

1/60 3/60 6/60 6/36 6/24 6/18 6/12 6/9 6/7.5 6/6 6/5 6/4 6/3 6/2

20/1200 20/400 20/200 20/120 20/80 20/60 20/40 20/30 20/25 20/20 b 20/17 b 20/13 20/10 20/7

0.017 0.05 0.1 0.17 0.25 0.33 0.5 0.67 0.8 1 1.2 1.5 2 3

1.8 1.3 1 0.8 0.6 0.5 0.3 0.2 0.1 0 0.1 0.2 0.3 0.4

a

Each row contains approximately equivalent values of visual acuity. Log MAR values have been rounded to 1 decimal place. On US Snellen charts, these lines are 20/16 and 20/12 respectively.

b

Acuity

Visual Acuity across the Retina Nonfoveal vision is limited by many elements. First, the eye’s optics are not optimized for viewing off the visual axis, and peripheral vision is subject to greater aberration than central vision. Second, the size of photoreceptors increases and their density falls with increasing eccentricity. The number of photoreceptors per retinal ganglion cell also increases, from less than one photoreceptor per ganglion cell in the fovea to more than 20 photoreceptors per ganglion cell in the far periphery. The volume of visual cortex devoted to noncentral retina is also proportionally lower. It is unsurprising, therefore, that visual acuity falls quickly with increasing distance from the fovea (Figure 4). This is one reason for the severely reduced visual acuity of people with central vision loss from diseases such as age-related macular disease.

Visual Acuity over Life Over the first year of life, visual acuity assessed by a preferential looking test appears to be reasonably stable

250 Letter visual acuity (min arc)

beyond the accommodative ability of the eye will reduce visual acuity by a similar amount to an equivalent degree of myopia. Astigmatism, particularly where the meridia of astigmatism are oblique, will also reduce uncorrected vision significantly. Other aberrations of the eye beyond defocus and astigmatism further limit visual acuity. Retinal image quality can be improved by viewing monochromatic stimuli (to reduce chromatic aberration) and by using a deformable mirror to correct coma, trefoil, and other higher-order aberrations of the eye. Under these ideal conditions, Williams and colleagues have shown that subjects are able to resolve gratings of up to 55 cycles per degree, equivalent to a visual acuity of approximately –0.30 logMAR (6/3; 20/10). Assuming that an image is perfectly focused on the retina, the next limit on visual resolution is the spacing of the retinal photoreceptors. In order to detect a grating, alternate black and white bars must fall on adjacent photoreceptors. This theoretical limit of vision, known as the Nyquist limit, is equivalent to a grating with light to dark separation of 1/√D, where D is the center-to-center separation of two photoreceptors. In the fovea, D is approximately 3 mm, equivalent to a visual angle of approximately 55 cycles per degree – almost identical to the value found by Williams. This confirms that in people with good vision, all of the limits on visual acuity are precortical. Amblyopia, where vision is reduced despite the absence of any eye disease, is dealt with elsewhere in the encyclopedia.

497

200

150

100

50

0

0

10

20

30 40 Eccentricity (⬚)

50

60

Figure 4 Letter visual acuity measured in peripheral vision as a function of degrees of eccentricity. Data from Anstis, S. M. (1974). Letter: A chart demonstrating variations in acuity with retinal position. Vision Research 14(7): 589–592.

at around 6 min of arc. Between a child’s first and third birthday, visual acuity improves exponentially to reach 1 min of arc. A further small improvement in resolution ability to approximately 0.75 min of arc is achieved by age 5 years. In the absence of eye disease, this value remains relatively constant until the sixth decade. In a populationbased study of nearly 5000 older adults, Klein found a decrease in visual acuity to a mean value of approximately 2 min of arc in those aged over 75 years. Of course, this reflects the age-related nature of many diseases which affect visual acuity, such as cataract, glaucoma, diabetic retinopathy, and age-related macular degeneration. Figure 5 plots data from the studies of Mayer and Klein. Visual Standards In most countries, there is a visual-acuity requirement for car drivers. While the level and measurement technique varies between countries, the acuity limit is usually approximately 0.3 logMAR. Commercial airline pilots are required to have a binocular visual acuity of 0.0 logMAR. Best corrected binocular visual acuity of 1.0 logMAR or poorer is used as a definition of low vision or partial sight in many countries, with acuity of worse than 1.3 logMAR being described as severe sight impairment.

Hyperacuity Some visual tasks can be performed with a far greater degree of precision than would be suggested by the MAR. Alignment tasks such as Vernier discrimination (where the offset of one line with respect to another is detected, Figure 6(a)) can be performed with misalignment of less

498

Visual Acuity Related to the Cornea and Its Disorders

6

Visual acuity (min arc)

5 4 3 2 1 0

1

10 Age (years)

100

Figure 5 Variation in visual acuity over life. From Mayer, D. L. and Dobson, V. (1982). Visual acuity development in infants and young children, as assessed by operant preferential looking. Data from Vision Research 22(9): 1141–1151 and Klein, R., Klein, B. E., Linton, K. L., and De Mets, D. L. (1991). The beaver dam eye study: Visual acuity. Ophthalmology 98(8): 1310–1315.

must be for it to be seen. If a target moves with velocity of 40 s1, the MAR is increased to about 2 min of arc, while at 80 s1, acuity is about 3 min of arc. In peripheral vision, slow image motion (less than 10 s1) slightly improves visual acuity for peripherally presented targets, perhaps because it breaks the phenomenon of Troxler fading. Target motion at the retina can be induced by target movement, by eye motion, or by head motion. Many eye diseases, particularly those of the macula, are associated with poor fixation stability of the eye. This poor eye stability increases retinal image motion, and is significantly associated with poorer visual function. Small degrees of head motion do not significantly decrease visual acuity under normal conditions, but have a marked deleterious effect for subjects viewing through telescopic spectacles. Therefore, subjects with macular disease who have poor fixation stability and who view through telescopic low-vision aids have a marked impairment in their dynamic visual acuity. See also: Amblyopia; Contrast Sensitivity; Pupil.

Further Reading

(b)

(a) Figure 6 Examples of hyperacuity tasks. Misalignment of the lower element is easily visible. (a) Vernier alignment; (b) dot alignment: the offset of the middle dot with respect to the upper and lower dot is easily discerned.

than 5 s of arc – considerably less than the center-tocenter spacing of a foveal photoreceptor. This is thought to be due to interpolation of the inputs of two or more adjacent neural elements.

Dynamic Visual Acuity Throughout this article, visual acuity has been discussed for static targets. If the target is moved, central visual acuity decreases: the faster the target moves, the larger it

Anstis, S. M. (1974). Letter: A chart demonstrating variations in acuity with retinal position. Vision Research 14(7): 589–592. Bailey, I. L. and Lovie, J. E. (1976). New design principles for visual acuity letter charts. American Journal of Optometry and Physiological Optics 53: 740–745. Bennett, A. G. and Rabbetts, R. B. (eds.) (1989). Visual acuity and contrast sensitivity. In: Clinical Visual Optics, pp. 23–72. Oxford: Butterworth-Heinemann. Brown, B. (1972). Resolution thresholds for moving targets at the fovea and in the peripheral retina. Vision Research 12(2): 293–304. Committee on vision. (1980). Recommended standard procedures for the clinical measurement and specification of visual acuity. Report of working group 39. Advances in Ophthalmology ¼ Fortschritte der Augenheilkunde ¼ Progres en Ophtalmologie 41: 103–148. Assembly of Behavioral and Social Sciences, National Research Council, National Academy of Sciences, Washington, DC Crossland, M. D., Culham, L. E., and Rubin, G. S. (2004). Fixation stability and reading speed in patients with newly developed macular disease. Ophthalmic and Physiological Optics 24: 327–333. Demer, J. L. and Amjadi, F. (1993). Dynamic visual acuity of normal subjects during vertical optotype and head motion. Investigative Ophthalmology and Visual Science 34(6): 1894–1906. Klein, R., Klein, B. E., Linton, K. L., and De Mets, D. L. (1991). The beaver dam eye study: Visual acuity. Ophthalmology 98(8): 1310–1315. Liang, J., Williams, D. R., and Miller, D. T. (1997). Supernormal vision and high-resolution retinal imaging through adaptive optics. Journal of the Optical Society of America. A, Optics, Image Science, and Vision 14: 2884–2892. Mayer, D. L. and Dobson, V. (1982). Visual acuity development in infants and young children, as assessed by operant preferential looking. Vision Research 22(9): 1141–1151. Plainis, S., Tzatzala, P., Orphanos, Y., and Tsilimbaris, M. K. (2007). A modified ETDRS visual acuity chart for European-wide use. Optometry and Vision Science 84(7): 647–653.

Acuity Rosser, D. A., Cousens, S. N., Murdoch, I. E., Fitzke, F. W., and Laidlaw, D. A. (2003). How sensitive to clinical change are ETDRS logMAR visual acuity measurements? Investigative Ophthalmology and Visual Science 44: 3278–3281. Thibos, L. N., Cheney, F. E., and Walsh, D. J. (1987). Retinal limits to the detection and resolution of gratings. Journal of the Optical

499

Society of America. A, Optics, Image Science, and Vision 4: 1524–1529. Westheimer, G. (1987). Visual acuity. In: Moses, R. A. and Hart, W. M. (eds.) Adler’s Physiology of the Eye: Clinical Application, pp. 415–428. St Louis, MO: Mosby.

Contrast Sensitivity P Bex, Schepens Eye Research Institute, Boston, MA, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Channels – The groups of visual sensors that are selective for a narrow range of image spatial or temporal structure. Contrast constancy – At high contrasts, apparent contrast is relatively independent of the parameters that strongly influence contrast-detection threshold. Contrast-detection threshold – The statistical contrast boundary below which contrast is too low for an image to be detected reliably and above which contrast is high enough for frequent image detection. Often defined as the contrast that produces 75% correct target identifications in forced-choice paradigms. Contrast sensitivity – The reciprocal of contrastdetection threshold that also represents the transition between visible and invisible images. Critical flicker frequency – The highest flicker rate of a full contrast image that can be detected reliably. Forced-choice paradigms – Robust behavioral method used to measure detection or discrimination thresholds. Observers are forced to select between two or more intervals, of which only one contains a target. Fourier analysis – Analytical method that calculates the simple sine-wave components whose linear sum forms a given complex image. Resolution limit – The highest spatial frequency of a full contrast image that can be detected reliably. Spatial frequency – The number of image cycles that fall within a given spatial distance, typically 1 of visual angle. Temporal frequency – The number of image cycles that fall within 1 s. Wavelets/gabors – A local filter that is the point-wise product of a two-dimensional (2D) spatial sine wave and a 2D Gaussian envelope.

Most people are familiar with image brightness and contrast from their controls on computer and television displays. The brightness control adjusts the mean luminance of the display uniformly, in order that the intensity of every point in the image increases when brightness is increased or decreases when brightness is reduced. The contrast control adjusts the difference between the lightest and darkest areas of the image. Increasing contrast

500

makes areas that are below mean luminance darker and areas that are above mean luminance lighter, without changing the mean value. Decreasing contrast draws all values toward the mean, thus making the whole image fainter, similar to viewing the image through fog. Figure 1 illustrates the effect of changing the contrast of a sine-wave striped pattern (the reasons for using a sine-wave pattern are described below). The top panel shows images of gratings whose contrast increases from 12.5% on the left to 100% on the right. The mean luminance of each image is the same. The traces in the bottom row plot luminance versus position for a horizontal slice through each image.

Contrast-Detection Threshold A powerful measure of visual sensitivity can be obtained by finding the minimum contrast that is necessary for an image to be detected. This minimum contrast is referred to as contrast-detection threshold (Cthresh) and it is important because it defines the transition at which an image moves from invisible to visible. One method to estimate Cthresh might be to allow a subject to adjust the contrast until an image is just visible. However, this method is highly subjective and large differences in individual criteria for just visible make this measure unreliable.

Psychophysical Assessment of Vision To overcome these problems, most researchers employ forced-choice procedures that require an observer to identify which of two or more intervals (the more the better) contain the target. An example of a four-alternative forcedchoice (4AFC) detection task is shown in Figure 2(a). In this case, a computer presents a target in one of four positions at random around a central fixation point. The observer’s task is to fixate the central dot and to indicate the location of the target, usually by pressing a computer button. Targets that are below Cthresh (sub-threshold) are rarely detected, whereas targets that are above Cthresh (supra-threshold) are usually detected. Contrast-detection thresholds are therefore probabilistic and are defined as the contrast at which they are correctly detected midway between chance and perfect performance. It is difficult to cheat on forced-choice methods or to change criteria – the target is either seen, in which case its position is correctly identified, or it is not seen, in which

Contrast Sensitivity

12.5%

25%

(a)

50%

(b)

(c)

501

100%

(d)

Figure 1 Image contrast. The top row shows the appearance of two-dimensional (2D) sine-grating patterns that are routinely used in vision research. The contrast of the sine grating increases from left (12.5%) to right (100%) as shown by the caption. The bottom row plots a horizontal section through each image and shows that contrast changes the luminance range separately from mean luminance.

Proportion correct

1 0.75 0.5 0.25 0 0.001

0.01

0.1

1

75% correct at approximately 2.5% contrast. The slope (s) can be used to infer how easily nearby contrasts can be discriminated from one another – a shallow slope means that a large contrast difference is required to achieve a given change in performance, whereas a steep slope means that a small change in the stimulus produces a large change in performance.

Contrast

(a)

(b)

Figure 2 Contrast detection. (a) Example of a four-alternative forced-choice (4 AFC) task. The observer is required to fixate the central dot and to indicate whether the target appeared top left, top right, bottom left, or bottom right. The target contrast is adjusted by computer to a level that produces 75% correct detection. (b) A typical psychometric function. Circles show the proportion of trials the target was detected (ordinate) as a function of the target contrast (abscissa). Error bars show 1 standard deviation. The curve shows the best-fitting cumulative normal function, from which the interpolated 75% correct point is taken as contrast-detection threshold.

case the subject is forced to guess. Notice that when guessing, the subject is still correct sometimes (25% if there are four alternatives, 33% if there are three, or 50% if there are two, etc.), as shown in the frequency of seeing curve in Figure 2(b), where, at low contrasts, performance is 25% correct. The data have been fit with a curve known as a psychometric function, in this case a cumulative Gaussian: Y ¼ g þ ð1 g Þ erf ðz=sqrtð2ÞÞ=2

where z ¼ (X – m)/s; g is the guess rate (0.25 in a 4AFC experiment). The mid-point (m) of the psychometric function is often taken as Cthresh – for a 2AFC task, this is 75% correct. In the example shown, the observer achieved

Spatial Frequency Channels Based on behavioral observations in humans and single unit recordings in mammalian visual systems, researchers discovered around half a century ago that the visual system analyses images at a series of relatively narrow spatial scales and orientations known as channels. Thus, fine and coarse image details are encoded separately and Fourier analysis can be used to study the image structure that is encoded by different visual processing channels. Fourier analysis computes the sum of basic sine waves whose linear sum produces the image. To illustrate the representations of an image that are available at different spatial scales, Figure 3(a) shows a typical image, together with its coarse (Figure 3(b)) and fine (Figure 3(c)) spatial structure. Visually responsive neurons in primary visual cortex, the first cortical projection from the retina through the lateral geniculate nucleus of the thalamus, respond to images only within a limited area of the visual field, known as the classical receptive field, and are selective for a limited range of spatial frequencies and orientations. These receptive fields are now routinely modeled as Gabor or wavelet functions, defined as:

502

Visual Acuity Related to the Cornea and Its Disorders

(a)

(b)

(c)

Figure 3 Spatial frequency in real images. (a) An image of Albert Einstein’s face is encoded at a range of spatial scales, from (b) coarse – low spatial frequency to (c) fine – high spatial frequency.

02 x þ g2 y 02 x0 þ ’ G ðx; y; l; ’; s; gÞ ¼ exp sin 2p 2s2 l

where x 0 ¼ x cos y þ y sin y and y 0 ¼ x sin y þ y cos y, l represents the wavelength, y the orientation, and c the phase of the sine-wave component. For the Gaussian window, s is the standard deviation and g is the spatial aspect ratio. Examples of Gabors are illustrated in Figure 4. On the top row, spatial frequency increases from left to right and all Gabors are of the same orientation 0 and contrast. On the bottom row, spatial frequency is fixed, but orientation is 45 , 90 , or 135 (from left to right). The visual system encodes image structure with a bank of such wavelet filters that represent the retinal image through patchwise local analysis. Figure 5 provides compelling demonstrations that our visual system employs a set of spatial frequency and orientation-selective channels. These demonstrations show that after prolonged viewing of a particular pattern (termed adaptation) the appearance of other patterns can be altered (termed an aftereffect). In these demonstrations, adapting to a pattern of one spatial frequency or orientation produces a loss in sensitivity in the channel that responds most to that pattern, but little change in channels tuned to other spatial frequencies or orientations. This localized loss in sensitivity produces a relative shift in the responses of our visual channels that cause us to experience changes in the appearance of the image. These observations have led to the widespread use of sine-wave grating patterns in basic and clinical vision research. In order to derive a measure of vision that reflects the sensitivity across our set of visual channels and to reflect the fact that functional vision requires us to detect and interact with objects of various sizes, contrastdetection thresholds are measured for gratings of a range of bar widths, expressed as spatial frequency or the number of grating cycles per unit distance. Figure 4 illustrates Gabors of differing spatial frequency; however, the size of one grating cycle on the retina depends on the distance

Figure 4 Gabor (wavelets) of differing spatial frequency and orientation. Top row: spatial frequency increases from left to right, orientation is fixed at 0 . Bottom row: Orientation increases from left to right: 45 , 90 , and 135 , spatial frequency is fixed.

from which it is viewed. Therefore, image sizes are usually calculated in terms of visual angle, which specifies the retinal image size. Figure 6 shows how visual angle is calculated and its relationship to image size and viewing distance. A convenient rule is that 1 cm viewed from 57 cm subtends a visual angle of 1 and roughly corresponds to a finger nail viewed at arm’s length.

Contrast Sensitivity Function Many researchers have shown that for sine-grating patterns, Cthresh strongly depends on spatial frequency. This fundamental observation is demonstrated in the classic image shown in Figure 7. Spatial frequency increases from left to right and contrast increases from top to bottom, so that contrast is constant across any horizontal line. Contrast-detection thresholds can be visualized on this figure as the imaginary curve along

Contrast Sensitivity

Figure 5 Demonstration of spatial frequency- and orientationselective aftereffects. First note that when you fixate the centre gray dot, the gratings in the middle row are of the same spatial frequency and orientation. Next, look back and forth between the black dots in the top row for around 10 s. Now, when you look at the center gray dot, the grating on the left appears to be of higher spatial frequency than the grating on the right. Next, look back and forth between the white dots in the bottom row for around 10 s. Now, when you look at the center gray dot, the grating on the left appears tilted counterclockwise, while the grating on the right appears tilted clockwise. These aftereffects are robust even though you know that the gratings in the middle row are the same. These demonstrations provide compelling evidence that visual processing involves channels that are narrowly tuned for spatial frequency and orientation.

which the grating changes from invisible (toward the top of the figure) to visible (toward the bottom of the figure). Most people report that the function peaks somewhere near the middle of the figure. Notice that the peak shifts as you move the figure closer or further away. This demonstrates the importance of visual angle rather than physical image size. Note that the highest spatial frequency that can be detected at maximum contrast is given by the rightmost point on a contrast sensitivity function (CSF). This is referred to as the resolution limit and is a quick and convenient method of assessing visual sensitivity than measuring the entire CSF. When measured with forced-choice procedures, (Figure 2) contrast-detection thresholds are lowest for

503

gratings around 2–5 cycles per degree of visual angle (c deg–1). By convention, the inverse of Cthresh (1/Cthresh) is usually reported and is termed contrast sensitivity. The rationale for the use of contrast sensitivity over contrastdetection threshold is most likely because the shape of the CSF is the same as that of the underlying modulation transfer function of the system. The circles in Figure 8 show the author’s contrast sensitivity as a function of spatial frequency measured with a forced-choice procedure. Error bars show 95% confidence intervals. The data have been fit (green curve) with the outputs of a set of spatial frequency channels shown by the colored curves. The channels are log spaced in spatial frequency (with peaks at 0.5, 1, 2, 4, 8, 16, or 32 c deg–1) and have the same bandwidth (1.4 octaves). The summed outputs of the set of filters provide a good fit to the data and this channel-based system is now a widely accepted model of early visual processing. The spatial frequency aftereffect shown in Figure 5 is easily explained with this channel-based model. Adapting to one spatial frequency reduces the responses of the channel that is most sensitive to that spatial frequency, but has little effect on the responses of other channels. When a different spatial frequency is subsequently viewed, the overall activity across the channels is shifted away from the adapted channel. This shift in the population response produces a shift in apparent spatial frequency away from the adapting frequency. An analogous model explains the shifts in orientation in the lower row of Figure 5, except that orientation-selective channels are adapted rather than spatial-frequency-selective channels. The CSF is highly dependent on the mean luminance of the display on which it is measured. This can easily be experienced by viewing Figure 7 with a pair of dark sunglasses (possibly two pairs), which moves the curve down (reducing sensitivity) and shifts the peak to lower spatial frequencies. The data in Figure 8 were collected on a standard computer monitor that has a mean luminance of 50 cd m–2 (candelas per square meter). Photopic, mesopic, and scotopic vision and changes in visual performance show that sensitivity to high spatial frequencies increases with mean luminance. This property is important because CSFs are routinely measured on relatively dim displays (e.g., 50–100 cd m–2) in the laboratory and in the clinic; however, the luminance of the real world is typically much greater. For example, the luminance of a cloudy sky is around 35 000 cd m–2, suggesting that standard experimental conditions may underestimate sensitivity to fine spatial structure.

Temporal Contrast Sensitivity In addition to a dependence on spatial frequency, contrast sensitivity also depends strongly on temporal frequency. Figure 1 illustrates spatial variation in luminance, but

504

Visual Acuity Related to the Cornea and Its Disorders

a = 0.5° 2094 miles

240 000 miles

Figure 6 Visual angle and viewing distance. The angular size of an object is calculated as 2*tan((0.5*h)/d), where h is the height of the object and d is the distance from which it is viewed. The example, which is not to scale, shows the angular size subtended by the moon is 0.5 . For comparison, the nail of the average index finger viewed at arm’s length subtends 1 .

200

Sensitivity

150

100

50

0 0.5

5

50

Spatial frequency (c deg–1) Figure 8 Spatial contrast sensitivity. Circles show contrast sensitivity (the reciprocal of contrast-detection threshold) for sine gratings of a range of spatial frequencies. Sensitivity peaks at around 2 c deg–1 under the conditions employed here and decreases at lower or higher spatial frequencies. The black curve is the summed sensitivity of the set of log-scaled channels shown by the colored curves and provides a good fit to the data. Figure 7 Illustration of the contrast sensitivity function (CSF). Spatial frequency increases from left to right, contrast increases from top to bottom. The contrast along any horizontal line is fixed. Different spatial frequencies become visible at different contrasts and define an imaginary curve that separates seen from unseen structure. Notice that if you move the image closer to your eye, the peak moves to the right and if you move it further away, the peak moves to the left. This demonstrates that contrast sensitivity depends on retinal not physical image size. If you wear one or two pairs of dark sunglasses, the curve shifts down and the peaks moves left, which demonstrates the dependence of the CSF on mean luminance.

imagine instead that the x-axis represents time, rather than space. Now the figure illustrates flicker. Flicker frequency can be varied in the same way as spatial frequency is varied in Figures 3 and 7. The circles in Figure 9 show how the author’s contrast sensitivity varies as a function of temporal frequency for a 2 c deg1 grating pattern. Sensitivity peaks around 5 Hz, at the mean luminance

used here (50 cd m2) and decreases at lower or higher temporal frequencies. These data are well fit (black curve) by a model with only two temporal channels, compared with the multiple channels that support spatial contrast sensitivity. One channel (red curve) is low-pass or sustained and is most sensitive to structure that is stationary or slowly changing over time. The second channel (blue curve) is band-pass or transient and is most sensitive to structure that changes at around 5 Hz. The spatial resolution limit falls steadily with distance from the fovea, an effect that can be experienced by viewing Figure 7 while fixating away from the center of the image. As you fixate further away, the threshold curve moves further down the figure and its peak shifts further to the left. Unlike spatial resolution, temporal resolution (the highest flicker rate that can be detected at any contrast, often called critical flicker fusion frequency)

Contrast Sensitivity

a phenomenon termed contrast constancy.Contrast constancy can be experienced in Figure 7 – while the transition between visible and invisible gratings has a curved shape, toward the bottom of the figure, the gratings appear to have similar contrast regardless of spatial frequency. This has important implications for image enhancement, which should therefore target only image components that are below their Cthresh.

300

Sensitivity

250 200 150 100 50 0 0.5

505

See also: Acuity. 5 Temporal frequency (Hz)

50

Figure 9 Temporal contrast sensitivity. Circles show contrast sensitivity (the reciprocal of contrast-detection threshold) for sine gratings of a range of temporal frequencies. The black curve is the summed sensitivity of the two log-scaled channels shown by the red and blue curves. The red curve has peak sensitivity at low temporal frequencies – that is, static images – and is termed a sustained channel. The blue curve has peak sensitivity around 6 Hz and is termed a transient channel.

increases moderately with distance from the fovea. This explains why older, 60-Hz computer displays can sometimes be seen to flicker when seen in the peripheral visual field, but not when viewed directly. Just as spatial contrast sensitivity depends on luminance, so does temporal contrast sensitivity. A 35-mm film is generally recorded at 24 frames per second, a refresh rate that could be easily detected at moderate light levels, as can be seen from Figure 9. For this reason, movie theaters are generally dark because sensitivity to high flicker rates is poor under those conditions. In addition, the visible 24-Hz image update rate is masked by flashing the illuminant at 48 Hz, so each frame is flashed twice. At supra-threshold contrasts, apparent contrast is relatively independent of spatial or temporal frequency,

Further Reading Bracewell, R. (1999). The Fourier Transform and Its Applications, 3rd edn. London: McGraw-Hill. Campbell, F. W. and Robson, J. G. (1968). Application of Fourier analysis to the visibility of gratings. Journal of Physiology 197: 551–566. Field, D. J. and Tolhurst, D. J. (1986). The structure and symmetry of simple-cell receptive-field profiles in the cat’s visual cortex. Proceedings of the Royal Society of London. Series B. Biological Sciences 228(1253): 379–400. Georgeson, M. A. (1990). Over the limit: Encoding contrast above threshold in human vision. In: Kulikowski, J. J. (ed.) Limits of Vision, pp. 106–119. London: Erlbaum. Hubel, D. H. and Wiesel, T. N. (1959). Receptive fields of single neurones in the cat’s striate cortex. Journal of Physiology 148: 574–591. Kelly, D. H. (1961). Visual responses to time-dependent stimuli. 1. Amplitude sensitivity measurements. Journal of the Optical Society of America. A, Optics, Image Science, and Vision 51: 422–429. Kulikowski, J. J. and Tolhurst, D. J. (1973). Psychophysical evidence for sustained and transient detectors in human vision. Journal of Physiology 232(1): 149–162. Landis, C. (1954). Determinants of the critical flicker-fusion threshold. Physiological Reviews 34(2): 259–286. O’Shea, R. P. (1991). Thumb’s rule tested: Visual angle of thumb’s width is about 2 deg. Perception 20(3): 415–418. Rovamo, J., Virsu, V., Laurinen, P., and Hyvarinen, L. (1982). Resolution of gratings oriented along and across meridians in peripheral vision. Investigative Ophthalmology and Visual Science 23: 666–670.

Astigmatism M J Cox, University of Bradford, Bradford, UK ã 2010 Elsevier Ltd. All rights reserved.

Glossary Against-the-rule – Ocular astigmatism in which the meridian with greater optical power in the eye is horizontal. Astigmat – An individual with ocular astigmatism. Axis meridian – The meridian of a cylindrical surface that is flat and consequently has no optical power. Emmetropization – The active process of reduction in refractive error toward an ideally focused system that occurs in the human eye during the first 2 years of development. High-order wave front aberrations – Wave front aberrations that are expressed using cubic, or higher, powers of the light ray’s distance from the pupil center to predict the amount of aberration. Meridional amblyopia – A lack of contrast sensitivity to high and medium spatial frequency contours oriented along a particular meridian without any refractive error or ocular pathological process affecting the visual function. Paraxial optics – Image forming through an optical system where only rays traveling close to the optical axis of the system and/or at small angles to this axis are considered. Penetrating keratoplasty – A surgical procedure to remove corneal material and replace it with material from a donor cornea. Stokes lens – A lens constructed from two symmetrically counter-rotating cylindrical lenses of equal absolute power but opposite sign. These combine to make a continuously variable power crossed-cylinder lens. Wave front aberrations – Deviations of a wave front of light propagating through an optical system from a perfect spherical wave front. With-the-rule – Ocular astigmatism in which the meridian with greater optical power in the eye is vertical.

The Definition and Etymology of Astigmatism The earliest forms of correction for visual loss caused by refractive errors in the eye used spherical spectacle lenses.

506

These are rotationally symmetrical about the optical axis and, considering paraxial optics, produce a point image from a point object on the axis of the lens. This type of image forming is known as stigmatic, from the Greek stigma, meaning a branding mark. Astigmatism describes an optical system where any nonpoint image is formed from a point object. In practice, astigmatism commonly refers to the simplest extension of stigmatic image formation, namely where the optical system forms two perpendicular line images from an axial point object, each at a different distance along the optical axis. In an astigmatic optical system the power varies as a function of the meridian, with a maximum and minimum power in meridians that are perpendicular. This variation in power approximates very well to a sinusoidal function, as seen in the formula Fy ¼ FSph þ

FCyl FCyl þ cos½2ðy aÞ 2 2

½1

where Fy is the power in the meridian at y , measuring angles anticlockwise from the horizontal, FSph is the power in the meridian of minimum power, FCyl is the difference in power between the meridians of maximum and minimum power, and a is the angle of the meridian of maximum power. The double angle ½2ðy aÞ in eqn [1] demonstrates that the power varies through a complete cycle as we rotate the meridian 180 around the optical axis. Newton is said to be the first to describe the variation in optical power with meridian and the consequent formation of line foci, but did so in the rather specialized form of oblique astigmatism. It was left to Thomas Young to first describe and measure ocular astigmatism, a finding which was a by-product of his attempts to measure his own refractive error as a starting point for investigating accommodative mechanisms in the human eye. He found his own astigmatism to be around 1.75 D and even had additional evidence to suggest that its source was a tilt in his crystalline lens. Airy was the first individual to measure and correct his ocular astigmatism. Wollaston, Ostwalt, and Tscherning discovered means by which spectacle lenses could be manufactured to minimize lens-induced oblique astigmatism. The above discussion concerns regular astigmatism, where two perpendicular axes of symmetry exist within the optical system and the relationship between power and the angle of the meridian is known. In irregular astigmatism this symmetry does not exist over the aperture through which light travels to form an image. Locally, over much smaller apertures, such symmetry may be present, but for

Astigmatism

the purposes of image formation by the optical system or eye it is absent. This results in objects that do not form line images but rather two elongated spreads of light, even at best focus. Furthermore, these two elongated images are not oriented perpendicularly. All eyes contain some degree of irregular astigmatism but this is rarely visually limiting except in pathological processes where the irregular astigmatism is large. Examples of such pathology include keratoconus; lenticonus; corneal scarring, inflammation, dystrophy, and degeneration; pterygium; mechanical effects on the cornea from neighboring structures such as the lids or sclera or following the use of rigid contact lenses; lens dislocation; localized lens index changes; and polycoria and ectopic pupils. In ocularly healthy individuals with unusually large pupils and statistically higher levels of irregular astigmatism, visual function can be affected in low and medium light levels.

Ocular Astigmatism: Prevalence and Age-Related Changes Ocular astigmatism is important for two main reasons. First, it prevents optimal retinal image formation and leads to a loss of contrast in the retinal image. It can be argued that it is more debilitating than either myopia or hypermetropia as, unlike myopia, there is no object distance at which a clear retinal image can be formed and unlike hypermetropia, it is not possible to overcome the refractive defect by using one’s accommodation. Second, if the high levels of astigmatism that are naturally present during early infancy do not reduce during the process of emmetropization, then permanent meridional amblyopia can occur leading to lack of visual sensitivity to small oriented details in later years, even when the astigmatism has been refractively corrected. In addition, some suggest that visual blur during early life may help to drive the development of myopia in later years. Uncorrected ocular astigmatism may be one such cause of visual blur, and an association between ocular astigmatism and myopia development has been found. Almost all neonates have significant amounts of astigmatism caused by an unusually steep cornea in one meridian, although the angle of this meridian with higher power is not consistent across the population. Early development reduces this cornea-generated astigmatism such that by 4–6 years of age only around one-twentieth of the population has ocular astigmatism in excess of 1 D and in the great majority of young astigmats, the meridian with the greatest power is vertical (or within 15 of vertical, the so called with-the-rule astigmatism). In later childhood (5–17 years of age) the proportion of individuals with at least 1 D of astigmatism increases up to around a quarter, with a higher risk for Asians and Hispanics (around a third), and a lower risk for African-Americans (around a fifth).

507

In adults, about two-thirds of the population have measureable astigmatism (> 0.25 D), but the majority of this is at low levels (<1 D). Estimates suggest that the prevalence of higher amounts of ocular astigmatism (>1 D) in adults ranges from about 10–20%, dropping to only 0.5% for astigmatism in excess of 4 D. A general trend can also be found in the adult population concerning the angle of the meridian of greatest power. In younger adults (<40 years of age) this is predominantly with-the-rule. In older adults it is predominantly against-the-rule. This change is believed to be due to the changes in the effects of the upper and lower lids on the cornea as the lids age and the tension generated by the lids on the cornea reduces. The lids squeeze on the upper and lower cornea in youth, flattening the cornea directly underneath them but steepening the corneal cap in the vertical meridian. The mechanical effects of the collagen structure within the cornea cause a consequent flattening of the cornea in the horizontal meridian. With aging, the cornea steepens overall, but the steepening in the vertical meridian is offset by the reduction in lid tension, leaving a steeper corneal curvature in the horizontal meridian and a consequent change from with-the-rule to against-the-rule ocular astigmatism.

The Origin of Ocular Astigmatism The cornea is the major refracting surface in the eye and the ocular astigmatism is best correlated with the astigmatism generated by the cornea. Both the anterior and posterior surfaces of the cornea generate astigmatism and show similar changes in curvature as a function of meridian. Due to the aqueous humor immediately behind the cornea, the refractive power of the posterior surface is only about one-tenth of the anterior surface power and of opposite sign. Keratometers, when used to predict the refractive power of the cornea from anterior corneal curvature measurements in a meridian, account for this by adjusting the value used for the refractive index of the cornea. The physiological value of the corneal refractive index is believed to be close to 1.38, but the majority of keratometers (and their keratographic successors) utilize a value of 1.3375. With the exception of corneal astigmatism associated with syndromic conditions, there is a complex pattern of heredity associated with corneal astigmatism in the general population, with some evidence of autosomal dominant inheritance patterns, but it seems likely that polygenic inheritance with variable penetrance exists. Unidentified environmental factors are also believed to influence the corneal astigmatism. There is good evidence to suggest that the corneal shape, and hence the corneally generated ocular astigmatism, is influenced by lid position, tension, and shape.

508

Visual Acuity Related to the Cornea and Its Disorders

Performing a task that alters the usual lid position, for instance prolonged reading, produces measureable but temporary changes in the corneal shape and astigmatism. Extra-ocular muscles may also influence the corneal shape, certainly showing effects following changes as dramatic as those induced by strabismus surgery where the extraocular muscles are recessed or resected. Corneal astigmatism is also associated with some welldescribed syndromes, such as Down’s syndrome and Treacher Collins syndrome (also known as Treacher Collins–Franceschetti syndrome or mandibulofacial dysostosis). Changes in the orientation of the palpebral aperture correlate with the axis of the astigmatism, the most powerful corneal meridian lying perpendicular to the axis connecting the inner and outer canthus. In addition to ocular astigmatism being generated by the cornea, it is also influenced by the shape and position of the anterior and posterior surfaces of the crystalline lens, as well as the refractive index distribution of the crystalline lens. Accurate in vivo measurements of the shape of the anterior and posterior surfaces of the crystalline lens are difficult to make. Consequently, it is not known how great a contribution crystalline lens surface shape makes to ocular astigmatism, as opposed to the effect of surface position and refractive index changes. What is known is that the internal astigmatism, that is, that generated by the crystalline lens, partially compensates for the corneal astigmatism and it is believed that some active feedback process is responsible for this compensation. Decentration and tilt of the optical axis of the crystalline lens with respect to that of the cornea will also induce astigmatism. The average internal astigmatism is 0.5 D against-the-rule and this does not appear to alter significantly with age. This is too great to be explained by crystalline lens tilt alone, as the required tilt is 4–5 times the average tilt found for the crystalline lens.

Refractive index changes in the cortex of the crystalline lens, commonly associated with age-related cortical cataract, are also able to induce internal astigmatism, with the axis of the astigmatism believed to be associated with the long axis of the cortical opacity. Other pathological conditions that lead to abnormally large displacement of the crystalline lens also induce astigmatism, including sectoral weakness in the zonular supports of the crystalline lens in conditions such as Marfan’s syndrome and homocystinuria, following blunt trauma, and associated with hypermature cataract. Lenticonus is a condition that results in conically shaped lens surfaces inducing large amounts of both regular and irregular astigmatism.

Image Formation and Refractive Specification in Astigmatism The basic image-forming process that occurs in astigmatic imaging can be seen in Figure 1. Standard Notation for Specifying Astigmatic Refraction Here, the front surface of the lens is a cylindrical surface. Note the straight line section of the lens front surface in the XX0 meridian compared to the curved section in the YY0 meridian. This cylindrical surface is created by rotating a line about an axis. The orientation of this axis defines the axis meridian of the cylindrical surface. The distance of the line from the axis defines the radius of curvature, and hence power of the surface. The back surface of the lens is a spherical surface, and the lens is described as a spherocylindrical lens. Although used at one time in the production of spectacle lenses, such lens designs are now very uncommon and nearly all spectacle lenses used to correct

Y⬘ CLC Anterior focal line

X

Posterior focal line

X⬘ Y Figure 1 Image formation from a distant axial point object through a sphero-cylindrical lens. Blue rays are traveling in the horizontal meridian. Green rays are traveling in the vertical meridian. The front surface of the lens is cylindrical with a horizontal axis meridian, XX0 . The bold lines show, from front to back, the anterior focal line, the circle of least confusion (CLC), and the posterior focal line.

Astigmatism

astigmatism use a toric back surface, which is defined using two perpendicular radii of curvature. The front surface power in the axis meridian is 0, while in the perpendicular meridian (the power meridian) it is FCyl. The angle of the axis meridian is measured from the perspective of looking at a person with astigmatism. An angle of 0 is represented by 180 but the degree sign is omitted. If the power of the spherical back surface in this lens was FSph, then we would denote the lens power as

509

At other image planes in the light pencil, the image is a horizontal ellipse if measured anterior to the CLC and a vertical ellipse if measured posterior to the CLC. Paraxial optics can be used to determine the size and position of these astigmatic focal images. For a distant object (vergence of 0 D) and a pupil diameter d, the length of the anterior focal line is dFCyl Fmax

½5

The length of the posterior focal line is

FSph DS=FCyl DC 180

where DS is dioptres of spherical power. DC is dioptres of cylindrical power. FCyl is the difference in power between the meridians of maximum (Fmax) and minimum (Fmin) absolute power and is a signed value. Conversion between positive cylinder and negative cylinder notation is a straightforward process of FSph0 ¼ FSph þ FCyl

½2

FCyl0 ¼ FCyl

½3

Axis0 ¼ ðAxis þ 90Þ mod 180

½4

A third much less commonly used notation is crossedcylinder notation where the Fmax and Fmin meridians are considered as two separate cylindrical lenses. If Fmax is þ2.00 DC 180 and Fmin is þ1.00 DC 90 then the crossed-cylinder specification for the lens is þ1:00 DC 90=þ 2:00 DC 180

Ocular Image Formation in Astigmatism The eye that suffers from astigmatism can usually be successfully studied by considering it to have a single toric corneal refracting surface with perpendicular meridians of maximum and minimum power. The image forming is very similar to that shown in Figure 1, where the more powerful meridian is vertical, denoting an eye having with-the-rule astigmatism. This astigmatic image pencil is known as the conoid of Stu¨rm. Rays from a distant axial point that fan out in the vertical meridian (shown in green) are brought to a focus closer to the lens than the rays that fan out in the horizontal meridian (shown in blue). Hence, in the focal plane of the vertical meridian there is horizontal defocus and a horizontal line image is formed. In the focal plane of the horizontal meridian there is a vertical line image. Dioptrically (but not geometrically) equidistant between these two line images, the defocus in the horizontal and vertical (and all other) meridians is equal and the defocused image has its smallest spread, the circle of least confusion (CLC).

dFCyl Fmin

½6

The diameter of the CLC is dFCyl Fmax þ Fmin

½7

Given that FCyl found in human eyes is <1 D in around 80% of the cases, and the average value for Fmax is around 60 D it is evident that eqns [5] and [6] will produce very similar results that are about twice the value of the result from eqn [7]. Thus, the focal lines give around twice as large an extent of blur as the CLC. This is helpful in explaining some of the effects of ocular astigmatism on human visual performance. Classification of Ocular Astigmatism Ocular astigmatism is divided into five categories depending upon where the retina lies with respect to the conoid of Stu¨rm: . The anterior focal line lies behind the retina – compound hypermetropic astigmatism; . The anterior focal line lies on the retina – simple hypermetropic astigmatism; . The anterior and posterior focal lines lie on either side of the retina – mixed astigmatism; . The posterior focal line lies on the retina – simple myopic astigmatism; . The posterior focal line lies in front of the retina – compound myopic astigmatism. Most individuals with astigmatism have one or other form of compound astigmatism. Astigmatic Blurring and Visual Perception Unless the CLC is coincident with the retina, as is only possible in mixed astigmatism or hypermetropic astigmatism when accommodation is used, the retinal blur patches in an uncorrected astigmat will be oriented. In myopic astigmatism the long axis of the blur patch will typically be oriented along the more powerful corneal

510

Visual Acuity Related to the Cornea and Its Disorders

meridian, and in hypermetropic astigmatism along the less powerful corneal meridian. Contours and edges in a scene that are parallel to the long axis of the blur patch will appear relatively clear, while those perpendicular to this axis will appear to be most blurred. As most scenes contain more contrast energy for horizontal and vertical orientations than for oblique orientations, there is often useful nonblurred information available to the viewer who has with-the-rule or against-the-rule astigmatism compared to a viewer with obliquely oriented astigmatism. This can be measured with visual acuity results, where with-the-rule and against-the-rule astigmatism produce a smaller loss of visual acuity than astigmatism along an oblique axis. In addition, as the line spacing in printed text is larger than the letter spacing, horizontal blurring is worse for reading than vertical blurring. An astigmat will benefit when reading printed text if

he/she accommodates to a position where the vertical focal line is on the retina. For a with-the-rule astigmat, this would typically mean more accommodation is required to bring the vertical focal line (which is more posterior) onto the retina, compared with the accommodation required by an against-the-rule astigmat, for whom the vertical focal line is more anterior. If moderate to large amounts on myopia are present in addition, then the text would be best held at a distance that places the vertical focal line on the retina. Figures 2 and 3 demonstrate the effects of astigmatic blur, compared to spherical blur, for Snellen letters and oriented contours (Figure 2) and an interior scene (Figure 3) when there is simple myopic astigmatism (or equivalent spherical error) present in with-the-rule (axis 180 ), against-the-rule (axis 90 ) and oblique (axis 135 ) meridians. The images are taken from the viewer’s

Cylinder axis 90⬚ Cylinder axis 180⬚

Cylinder axis 135⬚ Spherical blur Figure 2 The effects of astigmatic blur, compared to spherical blur, for Snellen letters and oriented contours. Simple myopic astigmatism (or equivalent spherical error) is present in the following meridians: with-the-rule (axis 180 ), upper left panel; against-the-rule (axis 90 ), upper right panel; and oblique (axis 135 ), lower left panel. The level of blur is that of a narrow failure to pass the European vision standard for driving.

Astigmatism

Cylinder axis 180⬚

Cylinder axis 90⬚

Cylinder axis 135⬚

Spherical blur

Figure 3 The effects of astigmatic blur, compared to spherical blur, for an interior scene. Simple myopic astigmatism (or equivalent spherical error) is present in the following meridians: with-the-rule (axis 180 ), upper left panel; against-the-rule (axis 90 ), upper right panel; and oblique (axis 135 ), lower left panel. The level of blur is that of a narrow failure to pass the European vision standard for driving.

perspective and the level of blur chosen is that of a narrow failure to pass the European vision standard for driving. The induced spherical error is half that of the induced cylindrical error such that the diameters of the blur circle for the spherical lens and the CLC for the cylindrical lens would be matched.

Oblique Astigmatism Even in an optical system that is rotationally symmetrical, the formation of images from off-axis object points will result in astigmatism, known as oblique astigmatism. This should not be confused with conventional astigmatism along an oblique axis meridian, which is due to different radii of curvature of the refracting surface in different meridians. In the eye, oblique astigmatism is important because the fovea is offset from the best approximation to an optical axis that the eye has, so that the objects that form an image on the fovea are nearly always slightly offset from this optical axis. It is also the dominant form of aberration in image formation for the peripheral retina. Figure 4 shows the image-forming properties in oblique astigmatism. Rays in the tangential plane, the one containing the optical axis and the chief ray, form a line image oriented in the sagittal plane. Rays in the sagittal plane (the

511

plane containing the chief ray that is perpendicular to the tangential plane) form a line image oriented in the tangential plane. In this biconvex lens, the tangential images form nearer the lens than the sagittal images, but for any given lens power the relative locations of the tangential and sagittal images depend upon the relative powers of the two surfaces of the lens, assuming that there is an aperture stop behind the refracting surface as is the case in the eye or when a spectacle lens is used in front of an eye. The distance from the tangential image to the image position found in a system free from oblique astigmatism is always three times the equivalent distance for the sagittal image. Circumferential image contours are formed clearly in the tangential image plane. In the sagittal image plane, radial image contours are formed clearly. Any ocular astigmatism caused by oblique astigmatism at the fovea can be corrected with the usual refractive techniques and the oblique astigmatism that is present in the periphery of the visual field is unimportant in normal visual function as the retinal ganglion cell mosaic limits the visual resolving power of the eye to a greater extent than the optics. It is argued that ocular oblique astigmatism acts as a useful high spatial frequency filter to prevent aliasing and the consequent perception of reverse motion in the peripheral visual field. Oblique astigmatism in refractive correction using spectacles is important. As the eye rotates to view different objects in the field of interest it will make different angles with the optical axis of any spectacle lens, and this will then cause different amounts of oblique astigmatism to be generated by the lens. Ocular Astigmatism During Near Work In addition to the nature of the visual task typically being different for near work compared to distance viewing, occasionally the ocular astigmatism itself, or its correction, can change. The power and/or axis of the ocular astigmatism at near compared to distance viewing may be affected by the following factors: . when the eyes converge and depress for near work they tend to excyclotort; . the suspensory ligaments of the crystalline lens may not be uniform, causing induced astigmatic changes in the lens during accommodation; . the ciliary muscle fibers themselves may not exert uniform force on the crystalline lens; . the crystalline lens may not be homogenously plastic, inducing irregular shape changes during accommodation; . the pupil or lens may displace with respect to the corneal axis during accommodation inducing some change in the oblique astigmatism. The power of the astigmatic correction using spectacles can change due to near vision effectivity when the astigmatic or spherical refractive error is large.

512

Visual Acuity Related to the Cornea and Its Disorders

Tangential focus

xis

al a

tic Op

ief

ray

Ch

tial gen Tan

e

plan

Sagittal focus

is

l ax

a ptic

O

ief Ch

ray

Sagittal focus Tangential focus

Sa

git

ta

lp

lan

e

xis

la

ca pti

O

y f ra

ie Ch

ial ent

ne

pla

g Tan Sa

git

ta

lp

lan

e

Figure 4 The image-forming properties for a biconvex spherical lens demonstrating oblique astigmatism for imaging a distant off-axis object point. Rays (green) in the tangential plane, upper panel, the one containing the optical axis and the chief ray, form a tangential focus, which is a line image oriented in the sagittal plane. Rays (blue) in the sagittal plane, middle panel, a plane perpendicular to the tangential plane that also contains the chief ray, form a sagittal focus, which is a line image oriented in the tangential plane. The lower panel demonstrates the relative position of these light rays and the induced astigmatism.

Specialized Notations for Specifying Astigmatic Refraction In addition to the standard sphero-cylindrical notation used to specify astigmatic refractive errors, more

specialized notations have been developed. The chief limitation of the standard notation happens when more than one refractive error needs to be considered, for instance when trying to determine population descriptive

Astigmatism

statistics for refractive errors, assessing change in refractive error, or when trying to determine the effects of a refractive error when used in conjunction with other optical equipment or other induced refractive errors. Two main alternative forms of notation have been suggested: power vector notation and matrix notation. Power vector notation resolves the optical power of any sphero-cylindrical lens into three components, a spherical component M, and two crossed-cylinder components, J0 and J45, with their axes at either 90 /180 or 45 /135 , respectively. Equal values of the components in this notation will lead to equal retinal blur circle sizes. Equations for computing the M, J0, J45 values from the conventional FSph/FCyl y are given below: M ¼ FSph þ

Jy ¼

FCyl 2

FCyl cosð2yÞ 2

FCyl sinð2yÞ J45 ¼ 2

f11 f21

f12 f22

½8

½9

Keratometry

½10

is defined in thin lenses such that f11 ¼ FSph þ FCyl sin2 ðyÞ ¼ M þ Jy

½11

pﬃﬃﬃ FCyl sinð2yÞ ¼ J45 2 2

½12

f22 ¼ FSph þ FCyl cos2 ðyÞ ¼ M Jy

½13

f12 ¼ f21 ¼

The Measurement of Ocular Astigmatism Various techniques are employed to measure the extent of ocular astigmatism. The majority are based upon locating the meridians of maximum and minimum power and then minimizing the size of the retinal blur ellipses along those meridians. This may be done objectively or subjectively. Alternatively, different components of the astigmatic power may be measured and the final astigmatic refraction inferred from them. Subjective methods include the Stenopaeic slit, Jackson’s crossed-cylinder, fan and block, and Stokes lens techniques. Objective methods include retinoscopy, automated refraction, keratometry, and wave front analysis. It is the latter two that have benefited from the greatest development effort in the last few years.

While this notation is useful for adding, subtracting, and multiplying by a scalar when dealing with thin sphero-cylindrical lenses, it is much less convenient to use when multiplying lens powers (as required when computing the powers of multiple lens systems or dealing with thick lens systems) or computing the prismatic effects of lens systems. The alternative matrix representation of lens power is more general and can be used in the above-mentioned instances. A 2 2 matrix, F¼

513

It is interesting to note the simple transformations between these two systems in thin lenses and their relationship to the Zernike basis function representation of wave front aberrations now commonly used when considering irregular astigmatic refractive error. The Zernike astigma2 pﬃﬃ times the tism coefficients C22 and C2þ2 are exactly 2r 6 value of J45 and Jy, respectively, given a pupil radius r, and using normalized variance terms. This has implications when measuring and correcting ocular astigmatism.

In this method, the image magnification produced by using the anterior corneal surface as a convex mirror is measured along the meridians of maximum and minimum curvature. The sagittal radius of curvature is inferred from this magnification and the corneal astigmatism is inferred from the sagittal curvature the corneal refractive index and the presumed corneal back surface curvature. The ocular astigmatism is then inferred by assuming a fixed value for the internal astigmatism. An extension to the principle of the keratometer is the corneal topographer that uses either reflected light from the cornea to measure the local curvature, or light (or even sound) scattered from the cornea – tear-film or cornea – aqueous interfaces to measure the local surface heights. It is possible to computer the local curvature from the local surface heights and vice versa. Unlike keratometers, such instruments clearly demonstrate the conicoidal nature of the cornea over most of its surface excluding the limbal region. Estimates of ocular astigmatism can be made by finding the best fitting overall match to the power map provided by the instrument. They also detect and measure irregular astigmatism much more accurately than keratometers. Figure 5 shows a corneal keratometric power map and surface height map from the right eye of an individual with approximately 4.50 D of regular with-the-rule ocular astigmatism. The symmetric bow tie pattern in the curvature map is a typical finding in these types of representations of corneal power in astigmatism and the corneal axis of astigmatism can be clearly identified from both maps. Wave Front Analysis Most of the methods of refractive measurement estimate the refractive power of the eye along one meridian at a

514

Visual Acuity Related to the Cornea and Its Disorders

105

58.00 56.00 54.00 52.00 50.00 48.00 46.00 44.00 42.00 40.00 38.00 36.00 34.00 32.00 30.00 28.00 26.00

90

60

135

45

150

30

165

15

180

0

195

345

T

N 210

330

225 240

300 255

105

270

90

285

OD Elevation BFS anterior

75

120

60

135

45

150

Sim K’s: Astig: 4.9 D @ 121 ⬚ Max: 6.6 mm @ 121 ⬚ Min: 7.3 mm @ 31 ⬚

30

165

3.0 MM zone: Mean power Astig power Steep axis Flat axis 5.0 MM zone: Mean power Astig power Steep axis Flat axis

15

0

180

345

195

T

N 330

210

0.005 mm color steps

Sim K’s : Astig: 4.9 D @ 121⬚ Max: 6.6 mm @ 121⬚ Min: 7.3 mm @ 31⬚ 3.0 MM Zone: Irreg: ± 3.0 D ± 1.9 D Mean power 48.4 Astig power 4.8 ± 2.4 D ± 24⬚ Steep axis 114 Flat axis ± 23⬚ 29 5.0MM Zone: Irreg: ± 4.3 D ± 3.0 D Mean power 46.9 ± 3.1 D Astig power 2.7 ± 32 ⬚ 106 Steep axis 21 ± 31 ⬚ Flat axis White-to-white [mm]: 11.3 Pupil diameter [mm]: 3.9 Thinnest: 489 um @ (-0.6,-0.4) ACD (Ep): 3.49 mm Kappa: 5.65⬚@ 216.66⬚ Kappa intercept: −0.15, 0.06

315

1.0 D color steps

0.080 0.070 0.060 0.050 0.040 0.030 0.020 0.010 0.000 −0.010 −0.020 −0.030 −0.040 −0.050 −0.060 −0.070 −0.080

Optical power keratometric

75

120

225

Irreg: 48.4 4.8 114 29 Irreg: 46.9 2.7 106 21

± 3.0 D ± 1.9 D ± 2.4 D ± 24 ⬚ ± 23 ⬚ ± 4.3 D ± 3.0 D ± 3.1 D ± 32 ⬚ ± 31⬚

White-to-white [mm]: 11.3 Pupil diameter [mm]: 3.9 Thinnest: 489 um @ (−0.6,−0.4) ACD (Ep): 3.49 mm Kappa: 5.65⬚@ 216.66⬚ Kappa intercept: −0.15, 0.06

315 240

300 255

270

285

Float OD

Figure 5 A corneal keratometric power map (upper image) and surface height map (lower image) from the right eye of an individual with approximately 4.50 D of regular with-the-rule ocular astigmatism. Color coding is for D in the upper map and millimeters in the lower one.

time. In these cases the refractive power in that meridian is a best fit to the combination of low-order refractive errors, high-order wave front aberration, and light scatter in the eye to produce the best perceived retinal image by viewing or varying light rays in only that meridian. If scatter or high-order aberrations are large (as in irregular astigmatism) then maximum and minimum power meridians may not be perpendicular giving a problem (albeit soluble) when attempting to arrive at a sphero-cylindrical refractive error. In addition, because change is restricted to one meridian at a time, and the imaging effects from the two meridians tested may interact, it is far from guaranteed that an optimum solution will be found. Furthermore,

intermediate meridians will also influence image quality and these are not assessed. The Stokes lens subjective method circumvents these problems to some extent by estimating along two perpendicular meridians at the same time. Corneal topography assessments can theoretically reduce the problem significantly by estimating ocular astigmatism from combining many individual samples of corneal power across the whole pupil area. Wave front analysis, where the refractive power of the eye is also sampled across multiple small apertures across the pupil, also estimates ocular astigmatism from this array of samples and is typically represented using Zernike polynomials. The Zernike terms are

Astigmatism

orthogonal to each other and therefore theoretically do not influence one another over a circular pupil. By considering the measurement or expression of ocular astigmatism using this basis, as opposed to along meridians of maximum and minimum power, it is possible to arrive at an astigmatic correction that is optimized for reducing the wave front aberration given a limitation of being able to use only sphero-cylindrical refractive correction such as can be worked onto spectacle lenses, contact lenses, and intraocular lenses. The spherical element of the correction is not well predicted using an equivalent method, however, due to the extra importance of central versus peripheral pupil rays in the wave front. Techniques that can use asymmetrical surfaces, such as laser refractive corneal ablation, directly make use of the additional higherorder terms of the Zernike expression of the optical defect to provide further optimization of the refractive correction. However, neither corneal topography nor wave front analysis can account for the effects of ocular scatter.

515

common today to use minimum tangential error, where the tangential image plane is designed to be of a specified power that is independent of the angle of viewing. While this leaves a small residual astigmatic error, it reduces the spherical power error found in punctal lens designs where power changes as a function of viewing angle. Spectacle lenses with an aspheric surface can also be used to reduce or eliminate oblique astigmatism with the additional benefit of a flatter front surface that makes the lens look less bulbous and also allows the lens to be thinner and hence lighter. The disadvantage is that aspheric lenses must be centered on the visual axis correctly or they induce, rather than reduce oblique astigmatism and other nonrotationally symmetric aberrations. Progressive addition lenses (PALs, also known as varifocals), with their aspheric surface induce astigmatism in a similar way as they can be centered for only two zones in the lens. Contact Lenses

The Correction of Ocular Astigmatism Spectacles Spectacles are the most common form of correction for ocular astigmatism. Except in unusual circumstances these lenses are successful in providing optimal visual function. In a limited number of individuals with unusually high irregular astigmatism and/or under conditions where the eye’s pupil is very large, the other nonrotationally symmetric optical aberrations such as coma, trefoil, tetrafoil, and secondary astigmatism can reduce the visual performance significantly despite the correction of regular astigmatism. Spectacle correction of high levels of regular astigmatism also produces spatial distortions due to the different levels of relative spectacle magnification in the different meridians. These make it difficult to binocularly fuse objects that are viewed through the more peripheral parts of the lenses. Patients corrected in this way will often learn to limit this effect by substituting head movement for eye movements when viewing outside a limited area from the optical center of the lens. Misperceived orientations of features and contours in the field of view also occur in high levels of regular astigmatism. Perceptual adaptation takes place over about a 2-week period to ameliorate this effect. Changing the power of the lenses significantly can, however, lead to a return of the perceptual problems. Tscherning offered a solution to the problem of spectacle lens-induced oblique astigmatism by deriving the necessary lens forms (distribution of lens power between the front and back surfaces) to eliminate oblique astigmatism from spectacle lenses with a given refractive index and for use viewing objects at a given distance. These lens forms are known as punctal (or point focal) lenses. It is more

Contact lens correction of astigmatism can be with rigid or soft lenses. Spherical rigid contact lenses can eliminate the majority of regular or irregular corneal astigmatism. Ensuring the correct fitting of the rigid contact lens limits the amount of correction that can be provided with this method and bitoric lenses, with front and back toric surfaces, are needed for high degrees of astigmatism. Toric front surface rigid contact lenses can be used in cases of internal astigmatism that require correction. Most soft contact lenses flex sufficiently to align to the shape of the anterior corneal surface and so do not correct corneal astigmatism through a tear lens in the same way that rigid lenses do. Front surface toric soft contact lenses are required for the proper correction of ocular astigmatism. In either rigid or soft contact lens correction of ocular astigmatism, many of the optical problems associated with spectacle correction are relieved. Meridional image magnification, near vision effectivity, and differential prismatic effects are all greatly reduced in contact lens correction. One of the principal difficulties in using contact lenses to correct astigmatism is in maintaining a stable rotational alignment between the contact lens and the cornea. Surgery Surgical control or correction of ocular astigmatism usually relies upon altering the shape of the cornea. This is most commonly done by making circumferential incisions in the corneal periphery. This reduces the tensile forces across the corneal stroma in that location, steepening the cornea locally, but flattening it across the pupil in the meridian which intersects with the incision. It is common because cataract surgery is common and most cataract surgery uses a corneal incision, the incision usually being

516

Visual Acuity Related to the Cornea and Its Disorders

made in the steepest corneal meridian to reduce postoperative astigmatism. Penetrating keratoplasty is sometimes needed in severe cases of irregular corneal astigmatism. In this case, the resulting corneal shape can be altered by increasing or decreasing the tensile forces created by sutures. It is even possible to remove corneal material in one meridian, a wedge resection, to stretch the cornea, and change its shape. For healthy eyes, laser refractive surgery and conductive keratoplasty are the two techniques commonly used to treat ocular astigmatism. Toric intraocular lens implants also exist that can be used to correct ocular astigmatism and these may be used following cataract surgery or clear lens extraction. Not Correcting Ocular Astigmatism It is not always desirable or necessary to correct ocular astigmatism in adults. The most useful property of astigmatism is that, if one is accepting limited amounts of blur, it creates depth of focus in the eye. As mentioned earlier, for near tasks blur along a vertical meridian can usually be tolerated more readily than horizontal blur. An individual with against-the-rule simple myopic astigmatism will have vertical blur conjugate with a near distance equivalent to the amount of astigmatism and horizontal blur conjugate with distance vision. As long as this patient can tolerate the distance blur, they will have functional vision at both distance and near without accommodative effort. Such a situation would be useful for later presbyopes and pseudophakes and free the individual from needing any form of optical correction. In addition, in some instances where individuals have very different ocular astigmatism in each eye the fusion difficulties that occur with spectacle correction make binocular vision impossible. If contact lenses cannot be tolerated and surgery is not an option, then there may be no alternative than to significantly undercorrect the astigmatism in the nondominant eye and allow visual suppression of the resulting blurred image to occur. Finally, it is sometimes the case that an individual with high levels of ocular astigmatism has a measureable change in that astigmatism but is not complaining of any significant visual difficulties. One would be very cautious about changing any refractive correction in this individual

as the perceptual adaptation to the meridional magnification differences would be affected by doing so and the person may end up feeling that he/she see worse rather than better. The same is true when large changes to the ocular astigmatism are found that do cause symptoms. Adapting to large changes in the correction can be intolerable for some people.

Summary Although ocular astigmatism is very common, affecting two out of every three adults, visually disabling ocular astigmatism is much less common, affecting only one in five. Correcting this astigmatism is straightforward in the great majority of cases, although the measurement and correction of both irregular astigmatism and high degrees of regular astigmatism remain challenging. Astigmatism induced when looking obliquely through spectacles is well controlled by using the appropriate lens design and contact lens and surgical correction of astigmatism is successful in many instances, even though the uptake of these latter forms of refractive correction is still low. Measuring and correcting irregular astigmatism has improved with the use of more accurate assessment of the corneal surface and with wave front aberrometry to sample the local refractive powers across a fine array of pupil locations. Optimally correcting irregular astigmatism with custom contact lenses and surgical refractive correction is proving to be challenging but successful trials have been performed. See also: Hyperopia; Myopia; Refractive Surgery.

Further Reading Freeman, M. H. and Hull, C. C. (2003). Optics, 11th edn. London: Butterworth-Heinemann. Harris, W. F., Raasch, T. W., and Thibos, L. N. (1997). Optometry and Vision Science: Feature Issue on Visual Optics 6: 339–463. Rabbetts, R. B. (2007). Bennett and Rabbetts’ Clinical Visual Optics, 4th edn. London: Butterworth-Heinemann. Read, S. A., Collins, M. J., and Carney, L. G. (2007). A review of astigmatism and its possible genesis. Clinical and Experimental Optometry 90: 5–19.

Myopia F A Vera-Diaz, Schepens Eye Research Institute, Harvard Medical School, Boston, MA, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Accommodation – The changes in optical power by the eye in order to maintain a clear image (focus) as objects are moved closer. This occurs through a process of ciliary muscle contraction and zonular relaxation that causes the elastic-like lens to round up and increase its optical power. Natural loss of accommodation with increasing age is called presbyopia. Astigmatism – An optical defect in which refractive power is not uniform in all directions (meridians). Light rays entering the eye are bent unequally by different meridians that prevent formation of a sharp image focus on the retina. Choroidal neovascularization – The creation of new, weak and leaky, blood vessels in the choroid layer of the eye. It is a common symptom of wet age-related macular degeneration and of pathological myopia. Glaucoma – The neuropathy that affects the optic nerve of the eye and involves loss of retinal ganglion cells in a characteristic pattern causing peripheral visual-field defects. Retinal detachment – A disorder of the eye in which some layers of the retina peel away from its underlying layers of support tissue. It is a medical emergency and if untreated it may cause partial or total vision loss.

Definition of Myopia: Health and Economic Implications Myopia, also called near- or shortsightedness, refers to the refractive state of the eye whereby the images of distant objects are focused in front of the retina when the accommodation system is relaxed. It can also be described as the refractive error in which the point conjugate with the retina, the far point of the eye, is located at some finite point in front of the eye. Therefore, light entering the eye has to originate from near objects, within the eye’s focal point, or diverged by concave lenses (minus power) in order to be focused on the retina of the myopic eye (Figure 1). Distant objects are otherwise perceived as blurred, hence the term nearsightedness. Squinting of the eyes is another symptom of myopia which can, rarely, produce headaches.

Myopia, in particular high myopia, is directly or indirectly associated with a number of ocular health complications that are potentially blinding. Moderate and high levels of myopia – greater than 5.00D – are a predisposing risk factor of reghmatogenous retinal detachment (lifetime risk is greater than 9%). High myopia is also a predisposing factor for open-angle glaucoma, myopic retinopathy, and myopic maculopathy. The increased elongation of the globe may be associated with degenerative changes in the sclera, choroid, Bruch’s membrane, retinal pigment epithelium (RPE), and neurosensory retina. There is increased incidence of fundus (posterior part of the eye) lesions (Figure 2) such as posterior staphyloma, atrophy of RPE and choroid, lacquer cracks in Bruch’s membrane, subretinal hemorrhages, lattice degeneration, pavingstone degeneration, pigmentary degeneration, white with or without pressure, retinal holes or tears, posterior vitreous detachment, macular holes, and choroidal neovascularization (CNV). Of these, CNV is the most common vision-threatening complication. Keratoconus, lens opacities, and increased complications following cataract surgery and pigmentary glaucoma are other associated complications. Although various methods of optical correction of myopia are possible, none changes the abnormally large size and shape of the myopic eye, with consequent thinning of the various layers, and therefore the risk of complications remains. Therefore, frequent complete eye health examinations, with pupillary dilation for comprehensive examination of the fundus, are necessary. Myopia is a significant public health problem and its rapid increase in prevalence in recent decades is associated with a significant financial burden. In the United States, the annual direct cost of refractive correction with eye glasses alone is estimated to be at least $3.8 billion. Further adding to the cost of myopia are eye examinations, time off work, other optical corrections (e.g., contact lenses), refractive surgery, and other treatments. In the United States, correction of refractive errors, including the cost of glasses, contact lenses and refractive surgery, consumed over 12 billion dollars in 1990. Although blurred vision resulting from myopia can often be corrected with visual aids, such as glasses, contact lenses, or refractive surgery, uncorrected refractive error is the major cause of visual impairment worldwide, accounting for at least 33% of visual impairments. In the United States, it is estimated that 5.3% of the visual impairments are due to uncorrected refractive errors. In addition, myopia – even if corrected – can be an impediment in certain professions; for example, military pilots and police officers.

517

518

Visual Acuity Related to the Cornea and Its Disorders

Natural History of Myopia Emmetropization is the process whereby the refractive components and the axial length of the eye come into balance during postnatal development in order to induce emmetropia (no refractive error). Most infants are hyperopic, and in those born myopic, the myopia typically decreases to reach emmetropia by toddler age. However, in a small population, for instance, premature infants with retinopathy of prematurity, myopia is present at birth and does not regress. By 12 months of age, the frequency distribution of the spherical equivalent becomes leptokurtic (as it is in adults) with the peak at low hyperopia. Results of a large longitudinal study by Gwiazda and colleagues, in 2000, showed that infantile astigmatism is associated with increased astigmatism and myopia during the school years. Myopia typically develops during the school years, progressing until adulthood, but it may also develop in adults. Progression typically ceases in the teenage years, although it may continue into the 30s. Generally, the annual progression is close to 0.50D in juvenile (8–12year-old) Caucasians and double that for juvenile Asians. There is a correlation with the age of onset and the final refractive status in adulthood, that is, children who become myopic at an earlier age (6 vs. 11 years) have a higher risk for myopia progression and higher degree of myopia later on. Refractive error in the adult population follows a leptokurtic distribution with the peak around emmetropia. Later in life, a myopic refractive shift may result due to crystalline lens changes.

Figure 1 (From top to bottom) Schematic drawing (color) of an emmetropic eye (top) with parallel rays of light focusing at the retina; a longer myopic eye (middle) with parallel rays of light focusing in front of the retina; and a longer myopic eye corrected with a concave lens, the light rays are diverged and now focus at the retina (bottom).

Structural Correlates, Molecular and Anatomical Changes in Myopia In the Aristotelian writings (c 330 BC), the condition of shortsightedness was already documented. The optics of myopia, however, were first elucidated by Johannes

Figure 2 Fundus photograph of a normal emmetropic eye (left) and an eye with pathological myopia (right). Modified photographs from original images from images.google.com

Myopia

Kepler in his initial Clarification of Ophthalmic Dioptrics (1604) when he correctly assumed that the incident light was brought to a focus in front of the retina. The reason for the displacement of the focus in myopic eyes elicited much attention. Some authors believed it was due to abnormal convexity of the lens; others attributed the defect either to an increased convexity of the cornea or to undue length of the globe, while some others described anatomically the unusual distance between the lens and the retina. Biometric investigations of myopic eyes have confirmed that increased axial length of the eye, particularly vitreous chamber elongation, is the main structural correlate in human myopia. Steeper corneal radius, deeper anterior chambers, and thinner crystalline lenses have also been found; however, these are less consistent findings. The relationship between the eye’s axial length and corneal radii (normally 3/1) is generally increased in myopic eyes. Retinal thickness is increased in the fovea, but decreases toward the periphery and it is significantly thinner in myopic eyes beyond the macula. In the 1990s, some studies showed that intraocular pressure (IOP) was related to myopia in children and therefore associated to the pathogenesis of myopia. However, a number of recent studies have refuted this theory with findings of no association between IOP and refractive error prior to or after the onset of myopia in children. Differences in the results may have resulted from ethnic differences. The protein composition in aqueous humor is different in myopic eyes, suggesting that those proteins could indicate a potential biomarker for myopia development. However, most myopic changes are found in the posterior segment of the eye. The sclera of myopic eyes, even with low or moderate amounts of myopia, exhibits a number of structural changes due to the stretching of the eye, including increased extensibility, narrowing and dissociation of collagen fiber bundles, increased prevalence of stellate fibrils, severe thinning, and reduced collagen content, among others. These differences in scleral structure, and therefore its functionality, translate into a relatively thinned and weakened sclera. As myopic eyes expand during myopia development, the sclera must increase its surface area. Either new tissue must be added or existing tissue remodeled. Results of biochemical and histological studies are generally consistent with the hypothesis of active remodeling of the sclera during myopic growth. Regulatory changes in scleral metabolism could be rapidly evoked by a change in visual conditions which can also regulate the direction of change in eye size (toward hyperopia or myopia). The sclera’s role on regulating eye growth and emmetropization and its potential role for preventing myopia warrant further investigation. Signaling cascades link retinal image processing to scleral growth in myopia. Chemical messengers are released from the retina toward the RPE where secondary

519

messengers are released through Bruch’s membrane and transmitted through the choroid to the sclera. The role of substances and transmitters in the retina and choroid, such as glucagons, insulin, early growth response factor-1 (EGR1 or ZENK), dopamine, nitric oxide synthase (NOS) inhibitors, gamma aminobutyric acid (GABA), muscarinic receptor subtypes (M1 and M4), and acetylcholine, is being investigated. Jody Rada and Lisa Palmer have suggested that increased choroidal permeability may represent a mechanism for controlling the rate of delivery of bioactive factors to the sclera to regulate the rate of glycosaminoglycan synthesis in the posterior sclera. Changes in collagen subtype expression and turnover of the normal scleral matrix (matrix metalloproteinase-2) are ultimately responsible for the anatomical changes in collagen fibril morphology and tissue thinning found in myopia. A role of growth factors, such as transforming growth factor beta and specific matrix–cell receptors (integrins), and altered scleral cell phenotype (myofibroblasts), in the scleral regulation of myopia development has been suggested. In addition, several research centers are beginning to map the myopia-associated locus and to identify the gene(s) responsible for myopia (see the section titled ‘Etiology of myopia’). Anthropometric measures such as body stature and weight have been associated to myopia. The Genes in Myopia (GEM) Twin Study group from Australia has recently (2008) reported that individuals in the heaviest quartile of weight of their study population had an increased incidence of myopia compared to those in the lightest weight quartile, but the relationship was significant only for females. Previous investigations have also associated myopia with taller and heavier individuals, but not all studies agree.

Classifications of Myopia The pattern of myopia development is complex and variable; therefore, it makes more sense to refer to ‘‘myopias’’ rather than a single condition of myopia. This complex pattern makes a classification of myopia difficult and has resulted in numerous different classifications being postulated, including: . Classification according to the degree of myopia. (1) Low, (2) moderate, and (3) high. The limits are still arbitrary, a consensus among experts is necessary if studies of prevalence are to be compared. Typically, low myopia refers to amounts between 0.50D and less than 3.00D; moderate refers to amounts between 3.00D and 6.00D; and high would be greater than 6.00D. . Ophthalmologic classification based on the fundus changes. (1) Simple or physiological (no fundus changes) and (2) degenerative or pathological myopia (fundus anomalies).

520

Visual Acuity Related to the Cornea and Its Disorders

. Classification according to progression of myopia. In 1984, Donders subdivided myopia progression into (1) stationary, (2) temporarily progressive, and (3) chronically progressive (also called malignant or deleterious) myopia. Nowadays, researchers classify myopia based on the progression of the refractive power: (1) stable myopia refers to the refractive error that has not increased more than -0.25D in a period greater than 2years, and (2) progressing myopia refers to greater increases over that period. . Classification according to the age of onset. Typically classified as (1) congenital, (2) infantile, (3) juvenile, and (4) adult myopia. It may also be classified as (1) congenital versus (2) acquired. Research studies classify myopia based on the age of onset: (1) late-onset (15 years or older), and (2) early-onset myopia (14 years or younger). . Classification according to the combination of components of the eye. (1) Refractive, correlation or combination myopia, and (2) component myopia (e.g., due to corneal curvature myopia, lens myopia, and axial myopia). . Classification according to presumed etiology. (1) Environmental versus (2) genetic. Also: (1) physiological myopia, (2) school myopia (due to close work), and (3) excessive myopia (i.e., caused by diseases). . Genetic classification. Dominant type, recessive type, a sex-linked recessive type, etc. . Biological classification of myopia. (1) Physiological or simple myopia as a biological variation of the normal distribution of the eye components, and (2) pathological (progressive or magna) myopia as falling outside the normal distribution. Clinical forms of myopia include: nocturnal myopia, due to drift in the accommodation state that increases the power of the eye under scotopic conditions, and pseudomyopia, false myopia due to physiological or pathological increased accommodation state.

at the ages of 6–8 years with the average annual progression quoted as ranging from 0.10D to 0.60D. Chinese children are predominantly myopic; in 1990, the prevalence was 37% and 50% in the 6–12-year-old group and 13–18-year-old group, respectively. Recent studies in China and Sweden show that the prevalence of myopia is now higher for teenagers. Prevalence is even higher among Taiwanese children: 56% of 12-year-old children and 76% of 15-year-old children are myopes. In the later 1990s in Japan, 43.5% of 12-year-old children and 66% of 17-year-olds were myopes. It can be appreciated that there are racial differences within groups of same-age populations, for example, Chinese and Taiwanese populations show higher prevalence rates than other populations. Myopia prevalence declines to some extent in the population over 45 years of age, which is likely due to an increased prevalence rate in younger populations than an age difference. The prevalence of myopia varies considerably between races, although it is difficult to ascertain the influence of environmental factors on these differences. In an attempt to account for environmental differences, a number of studies have investigated differences in racial groups living in the same location. In 1999, Lam and Edwards reviewed previous studies on myopia prevalence and concluded that the greatest prevalence of myopia occurs among Japanese, Chinese, and some Native American tribes. These studies take into account that the prevalence of myopia is lower in rural populations and it is affected by educational attainment and socioeconomic status. Some studies have found a slightly higher prevalence of myopia in females than in males; however, not all agree on this. It seems that the differences in gender are found if the prevalence is measured in younger adults when males and females are not at the same growth level.

Etiology of Myopia Epidemiology of Myopia In the early 1980s, more than 25% of the adult population in the United States and up to 75% in other developed countries, such as Taiwan, was myopic. Twenty years later, approximately 50% of general population in USA and up to 84% of the 16- to 18-year-old group in Taiwan was myopic. Based on a serial cross-sectional study from the Singapore Armed Forces, myopia prevalence has increased in military conscripts from 26% in the late 1970s, to 43% in the 1980s, 66% in the mid-1990s, and to 83% in the late 1990s. The prevalence of myopia varies with age; it is low in young children, increases in school-age children and young adult cohorts, and decreases in older age groups. Within populations, myopia tends to be initially observed

It is currently accepted that the etiology of myopia is complex with genetic and environmental factors playing a role. Understanding the relative contributions of genetic and environmental components is necessary to establish the mechanisms of myopia development and further attempt to arrest its progression. Separating these components is not trivial; for example, educational attainment is strongly influenced by genes, and therefore this risk factor should not solely be considered as an environmental risk factor; twins typically have the same environmental factors and hence should not be considered merely a genetic component. A century ago, in 1913, Steiger showed that myopia is influenced by genetic factors and is an expression of a range within the normal distribution. A number of early

Myopia

and recent studies with twins have shown that the concordance of refractive error is greater in monozygotic twins, indicating the existence of a genetic factor in myopia. Further, twin studies have provided evidence of correlation and heritability of a number of traits implicated in myopia. However, there are limitations associated to these studies. The use of twin studies is also important in the identification of myopia’s gene loci. A number of genetic loci have been identified as being linked with myopia. These include loci for high myopia occurring at Xq28 (MYP1), 18p11.31 (MYP 2), 12q23-24 (MYP3), 7q36 (MYP4), 17q21-22 (MYP 5), 4q22-27 (MYP11), and 2q37.1 (MYP12). Other candidate loci have been linked with low and moderate (common) myopia: 22q12 (MYP6), 11p13 (MYP7), 3q26 (MYP8), and 4q12 (MYP9). These linkages provide evidence that myopia is a polygenic disease with multiple and different alleles likely to contribute to different disease subtypes. Multiple familial studies also support a genetic component in myopia, suggesting a definite genetic basis for high myopia, and likely for low myopia. However, to date no candidate genes have been shown to account for even a modest fraction of the familial risk of myopia, and data are conflicting about whether a true association exists. It is likely that there is substantial genetic heterogeneity. In addition, further studies need to assess the relative roles of environmental factors and genetic influences, such as interactions of early-age nearwork and genotype, and the identification of phenotypes for etiologically different subgroups of myopia, for example, age of onset, presence of retinal degenerative changes, or response to treatments. See Schaeffel et al., Young, and the Genes of Myopia (GEM) Twin Group studies in the further reading section for reviews on the molecular basis of myopia.

521

Environmental factors, particularly nearwork, have been associated to myopia for centuries. Even though associated, nearwork has not yet been proven to be a causative factor. A number of aspects related to nearwork have been associated to myopia development and progression, including reading, cognitive effort during nearwork (with associated educational level and intelligence), lighting, and working distance. Accommodation and inaccuracies of the accommodation response are associated with myopia. Children who are to become myopic show greater inaccuracies of accommodation (greater lags) during near tasks. Accommodation effort, inaccuracies of the accommodation system during distance viewing, and accommodation flexibility have also been associated to myopia (nearwork-induced transient myopia (NITM) – a shift in refractive error due to inability to relax the accommodation). Evidence for active and visually guided emmetropization (and its failures, such as myopia) is beyond refute. A number of animal models of myopia show that emmetropization can be manipulated and myopia can be induced. Manipulation of the environment can be achieved by imposing a close-work environment, or most commonly by lens-induced myopia (using high-power negative – concave – lenses) or form-deprivation myopia (using diffuser lenses, sutured lids, etc.; Figure 3). The most common animal models for myopia are chicks, mice, pigs, and tree shrews. Defocus blur is thought to be the primary cue for emmetropization and myopia. Emmetropization uses blur as visual feedback to regulate eye growth; the system requires detection of blur probably at the level of outer retina (perhaps amacrine and bipolar cells) – via diffusion of signals across RPE and choroid – to then alter scleral matrix (likely through a modulation of proteoglycan synthesis). The amount of defocus blur may increase in those

Figure 3 Schematic of an animal model of eye compensation for lens-induced defocus. Blue rays: a positive convex lens placed in front of a normal eye causes the image to form in front of the retina, the eye compensates by slowing down its growth rate and it hence becomes hyperopic (shorter); the eye layers become thickened. Red rays: a negative concave lens placed in front of a normal eye causes the image to form behind the retina, the eye compensates by accelerating its growth rate and it hence becomes myopic (longer); the eye layers become thinner. Black rays: the normal eye. Modified from Wallman, J. and Winawer, J. (2004). Homeostasis of eye growth and the question of myopia. Neuron 43, 447–468.

522

Visual Acuity Related to the Cornea and Its Disorders

individuals who adapt more to blur, and hence blur adaptation may be a risk factor for myopia. The balance between central and peripheral defocus has been related to myopia, and animal models show that peripheral defocus alone can influence the rate of axial eye growth. Optical aberrations have been suggested to play a role in myopization, as some studies have found higher levels of aberrations in myopic eyes, but their association with myopia is not clear. Among environmental factors, increasing educational demand seems to be a risk factor for development of myopia. When comparing university student populations with general populations, a much higher rate of myopia progression is found in the university populations. For example, significant differences have been found in the prevalence of myopia among Norwegian university students when compared to the general population of Norway and other Nordic countries. The prevalence of myopia has also been found to be greater in Greek university students and Scandinavian medical students than in the general population of the respective countries. In an interesting study in 1993, Zylbermann and coworkers showed that myopia was more prevalent in Orthodox Jewish boys in Jerusalem compared to Jewish girls, since boys spend up to 16 h day reading religious text, whereas girls receive a similar education to that in Western countries. Educational demands have also been cited as a risk factor in the prevalence of myopia in Young’s 1969 work with Eskimo families in Alaska; while only two of the 130 Eskimo parents (all illiterate) were myopic, 60% of the children (required to attend school) became myopic. Studies show that other environmental factors, such as light levels (related to latitude and the absence of ultraviolet (UV) radiation), contamination, diet, and parental smoking, have been associated to the development of myopia but there is no proven causation. Children who spend more daytime outdoors are less likely to become myopic. This was a key discussion during the 12th International Myopia Conference in Australia. Following Jones and colleagues’ first report in 2002 on the beneficial effects of outdoor exposure, others have found that it is the amount of time spent outdoors, rather than any particular physical activity as it was previously suggested, which may help retard myopia. There is consensus among myopia research groups (Orinda Longitudinal Study of Myopia, Sydney Myopia Study, and Singapore Sharable Content Object Reference Model (SCORM) study) that outdoor time is protective against myopia. The mechanism for this protective effect is unknown. Mutti and coworkers (in 2002), the Correction of Myopia Evaluation Trial (COMET) group (in 2005), and others have found that parental myopia is a high-risk factor for myopia. Parental myopia may affect myopia with both genetic and environmental components. A theory of genetic predisposition to myopia with environmental

triggers, such as nearwork, urban life, and reduced outdoor exposure, is commonly accepted at present. For an update on molecular, structural, and functional studies in humans and animals, including twin studies; prevalence, progression, and risk factors in myopia; the influence of nearwork and outdoor activity in myopia; and therapies for myopia see the 2009 work by McBrien and coworkers.

Correction and Prevention: Clinical Management of Myopia Myopia may be corrected using visual optical aids such as spectacles, contact lenses, or, increasingly, refractive surgery. Correction of myopia is achieved by placing minor power (concave) lenses in front of the eye (Figure 1). The power of the lenses is measured in diopters (D) and corresponds to the inverse of the focal power of the lens. The eye care provider measures the refractive correction using objective (i.e., retinoscopy and autorefractometer) and subjective techniques to determine the lowest-power diverging lens that achieves best visual acuity. A binocular and eye health examination is required to establish the appropriate prescription. As per the American Optometric Association guidelines, the goals for management of the patient with myopia are clear, comfortable, efficient binocular vision and good ocular health. Low levels of myopia (less than 3.00D) are not corrected in infants and toddlers as it may disappear within 2 years of age and their visual world is close anyway, that is, they can see as far as they need to. Older children or young children with higher amounts of myopia need to be corrected to allow the visual system appropriate development with clear visual input. For adolescents and adults, in general, any degree of myopia should be corrected any time the patient is adversely affected by the lack of clear distance vision; therefore, patients’ needs are taken into account when prescribing a myopic correction. Astigmatism may occur in conjunction with myopia. When the degree of myopia is different (2.00D) between the two eyes, the condition is called anisometropic myopia (anisomyopia). The material of choice for the eye glasses will also depend on the patient’s characteristics (e.g., polycarbonate lenses are given to children) and the degree of the myopia (e.g., high-index, thinner lenses are recommended for higher prescriptions). Contact lenses (soft or gas permeable) are typically well accepted by myopic patients as the size of the retinal image is larger than with glasses, and they avoid visual-field restrictions. Whether spectacles or contact lenses are preferable in a given case depends upon numerous factors, including patient age, motivation, compliance, corneal physiology, and financial considerations. Orthokeratology is a technique of contact lens fitting which flattens the corneal surface over time to transiently reduce myopia.

Myopia

A number of options to correct myopia with refractive surgery are available. Since the approval of the use of the excimer laser in the 1990s to reshape the cornea, there has been significant development in the correction of myopia. Laser refractive surgery has surpassed other conventional surgical techniques in safety and efficacy. The suitability of refractive surgery needs to be determined on a caseto-case basis and after a thoughtful discussion between the surgeon and the patient. Current refractive surgery options include: 1. Excimer laser photorefractive keratectomy (PRK). A procedure in which the corneal power is decreased by laser ablation of the central cornea, the corneal epithelium is scraped away to allow the reshaping of the corneal stroma. 2. Laser-assisted in situ keratomileusis (LASIK). This was introduced in the mid-1990s and largely replaced PRK. Unlike PRK, a flap is made in the epithelium to permit access to the stroma. LASIK avoids most of the problems of corneal haze, postoperative pain, and slow rehabilitation seen in PRK, but complications are sometimes associated with the flap. Some studies show that LASIK surgery has predictable and stable results in refractive and visual outcomes in correcting moderate to high myopia on long-term follow-up. Refractive stability is maintained over 7 years, with no evidence of progressive late-onset complications. Recently, an alternative mechanical, epi-LASIK technique has shown comparable effective results to LASEK. 3. Laser epithelial keratomileusis (LASEK). In this procedure, the epithelium is treated with alcohol and then peeled back to permit reshaping of the underlying layer. It avoids all flap-related complications associated with LASIK and has less postoperative pain and faster recovery than PRK. 4. Wavefront-guided (WFG), or custom LASIK. This technique is used to avoid the induced positive spherical aberration typically found after conventional LASIK procedures. It is more advantageous with large pupils (most studies of conventional LASIK have shown a relationship between the diameter of the low-light pupil and complains of visual symptoms after surgery). However, WFG LASIK has all the same risks of conventional surgery. 5. A new procedure of PRK with intraoperative use of topical mitomicym C seems to be more effective than LASIK surgery for moderate myopia. 6. A number of surgical techniques use phakic intraocular lenses implanted in the posterior chamber. These are used for the correction of high myopia. Despite significant advances, certain limitations and complications exist in these refractive surgery procedures and patient education is essential before a decision is taken. In this rapidly changing field, co-management of patients and consultation of most recent literature are necessary.

523

None of these correction techniques prevents or treats the condition. A number of approaches and techniques, which are discussed below, have been advocated over the years to prevent, inhibit, and attempt to reverse myopia development. Positive Additions for Nearwork The underlying principle of using positive additions is to reduce the accommodative demand at near, as increased accommodative effort is thought to play a role in the development of myopia. Positive additions may be achieved with the undercorrection of spectacles which also leads to a distance undercorrection and therefore defocused images; hence, the use of bifocals and progressive lenses has been suggested. The technique is supported by animal research that has shown that wearing positive lenses induces hyperopia. However, good results on the use of positive additions may not be only dependent on reducing the magnitude of accommodation, but on other oculomotor factors as well. Results of different studies are contradictory, although some reduction in the rate of myopia progression has been found to occur. In particular, the COMET group has found that the progression of myopia slows down in myopic children who have esophoria at near and are treated with positive additions in the form of progressive addition lenses (PALs). Ongoing longitudinal studies are being undertaken in this area. Contact Lenses Although some studies have shown that contact lenses (especially hard contact lenses) reduce myopia progression, the mechanism of action is not well understood. However, the success of these techniques is thought to result from corneal curvature changes produced by the contact lenses, which suggests that the changes may be transient. Vision Therapy and Biofeedback Training Various forms of accommodative vision therapy have been advocated in the treatment of myopia. The aim of a training program is to reinforce and establish control over the accommodative response. Although there are reports that these techniques can induce a reduction in myopia, there are no masked studies with objective data supporting the usefulness of vision therapy for correcting or preventing the progression of myopia. In most published reports the quantified test was solely unaided visual acuity which improvement may be explained by an improved ability to interpret a blurred retinal image. Myopes of low degree commonly report that their vision seems poorer upon removal of their spectacles compared to that after a period without spectacle wear. This phenomenon is called blur adaptation and accounts for very small to no additive improvement in visual performance.

524

Visual Acuity Related to the Cornea and Its Disorders

Rigorous research studies are needed if accommodation biofeedback is to qualify as a method of clinical treatment of myopia. Pharmacological Treatment The first drugs to treat the progression of myopia were chosen based on the belief that myopia could be controlled either by the relaxation of accommodation or the reduction of IOP. Hence, cycloplegic agents such as tropicamide and atropine, and adrenergic antagonistic agents have been used in the past with no success. The difficulties resulting from the regular instillation of drugs, such as inconvenience, reading problems, discomfort, pupillary mydriasis, and the possibility of other adverse reactions to the drug, in addition to the lack of clear evidence of the effectiveness of these agents, provoked abandonment of these techniques. Although none of these pharmacological agents used has demonstrated an ability to control myopic progression successfully, for the past few years new approaches to myopia control are considering the importance of muscarinic receptors in myopia development. M1 antagonist pirenzepine and atropine are the newest drugs used in the attempt to slow down the progression of myopia. Atropine is the only pharmacological agent currently studied in clinical trials. Although the site of action of these drugs is unknown, studies in chicks suggest that these drugs act on the cartilaginous sclera to transiently inhibit glycosaminoglycan synthesis and slow down myopia development. Growth factors, such as insulin-like growth factor 1 (IGF-1), basic fibroblast growth factor (bFGF), and retinoic acid have been shown to be effective at controlling ocular growth in animal models and are promising therapeutic agents for high human myopia. Future therapies to slow down the progression of myopia in humans will be directed at altering the retinal–scleral signaling cascade involved in emmetropization. Correction and Prevention: Clinical Management of Myopia? In conclusion, the limited success achieved by the various methods of myopia control should not necessarily be interpreted as providing the absence of any relationship between myopia and nearwork. It is possible that once axial elongation has begun, continued visual feedback would produce additional axial elongation irrespective of any external intervention. See also: Refractive Surgery.

Further Reading American Optometric Association (1997). Care of the Patient with Myopia, Optometric Clinical Practice Guideline. Association of Optometrists. http://www.aoa.org/myopia.xml (accessed Jun. 2009). Charman, W. N. (2005). Aberrations and myopia. Ophthalmic and Physiological Optics 25: 285–301. Chen, C. Y.-C. (2007). Heritability and shared environment estimates for myopia and associated ocular biometric traits: The genes in myopia (GEM) family study. Human Genetics 121: 511–520. Curtin, B. J. (1985). The Myopias: Basic Science and Clinical Management. Philadelphia, PA: Harper and Row. Dirani, M., Shekar, S. N., and Baird, P. N. (2008). Adult-onset myopia: The genes in myopia (GEM) twin study. Investigative Ophthalmology and Vision Science 49: 3324–3327. Gilmartin, B. (2004). Myopia: Precedents for research in the twenty-first century. Clinical and Experimental Ophthalmology 32(3): 305–324. Grosvenor, T. P. (1998). Clinical Management of Myopia. Boston, MA: Butterworth-Heinemann. Gwiazda, J., Hyman, L., Dong, L. M., et al. (2007). Factors associated with high myopia after 7 years of follow-up in the Correction of Myopia Evaluation Trial (COMET) cohort. Ophthalmic Epidemiology 14(4): 230–237. Jones, L., Sinnott, L. T., Mutti, D. O., et al. (2007). Parental history of myopia, sports and outdoor activities, and future myopia. Investigative Ophthalmology and Vision Science 48: 3524–3532. Kurzt, D., Hyman, L., Gwiazda, J. E., et al. (2007). Role of parental myopia in the progression of myopia and its interaction with treatment in COMET children. Investigative Ophthalmology and Vision Science 48(2): 562–570. McBrien, N. A., Young, T. L., Pang, C. P., et al. (2009). Myopia: Recent advances in molecular studies; prevalence, progression and risk factors; emmetropization; therapies; optical links; peripheral refraction; sclera and ocular growth; signalling cascades; and animal models. Optometry and Vision Science 86(1): 45–66. Rada, J. A. and Palmer, L. (2007). Choroidal regulation of scleral glycosaminoglycan synthesis during recovery from induced myopia. Investigative Ophthalmology and Vision Science 48: 2957–2966. Rosenfield, M. and Gilmartin, B. (1998). Myopia and Nearwork. Oxford: Butterworth-Heinemann. Saw, S.-M., Gazzard, G., Au Eong, K. G., and Tan, D. T. (2002). Myopia: Attempts to arrest progression. British Journal of Ophthalmology 86: 1306–1311. Saw, S.-M., Tong, L., Chua, W. H., et al. (2005). Incidence and progression of myopia in singaporean school children. Investigative Ophthalmology and Vision Science 46: 51–57. Schaeffel, F., Simon, P., Feldkaemper, M., Ohngemach, S., and Williams, R. W. (2003). Molecular biology of myopia. An invited review. Clinical and Experimental Optometry 86(5): 295–307. Wade, N. J. (1999). A Natural History Of Vision. Cambridge, MA: MIT Press. Wallman, J. and Winawer, J. (2004). Homeostasis of eye growth and the question of myopia. Neuron 43: 447–468. Young, T. L. (2009). Molecular genetics of human myopia: An update. Optometry and Vision Science 86(1): E8–E22.

Relevant Website http://www.revoptom.com – Review of Optometry: Handbook of Ocular Disease Management: Pathological Myopia and Posterior Staphyl.

Amblyopia D M Levi, University of California, Berkeley, Berkeley, CA, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Anisometropia – A condition in which the two eyes have unequal refractive power so that the two eyes are in different states of myopia. Astigmatism – An optical defect causing blurred images due to failure to focus a point object into a sharp image on the retina. Sensitive period – An early developmental period that is particularly sensitive to development of amblyopia. Snellen acuity – Clarity of vision as measured by eye care professionals using a chart called the Snellen chart. Strabismus – Misregistration or misalignment of the images from the two eyes preventing the development of binocular vision.

Clinically, crowding may be a useful sign to aid in the diagnosis of amblyopia.

Amblyopia Is a Significant Public Health Problem Amblyopia can easily be reversed or eliminated when diagnosed and treated early in life. Thus, there is a premium on early detection of amblyopia and its risk factors. It has been estimated that perhaps as many as three-quarters of a million preschoolers are at risk for amblyopia in the United States, and roughly half of those may not be detected before school age. Moreover, detection is likely to be more delayed in low socioeconomic areas. Improved vision screening and access to treatment could, in principle, eliminate amblyopia as a public health issue.

Types of Amblyopia

What Is Amblyopia? Amblyopia (from the Greek, amblyos – blunt; opia – vision) is a developmental abnormality that results from physiological alterations in the visual cortex and impairs form vision. Amblyopia is clinically important because, aside from refractive error, it is the most frequent cause of vision loss in infants and young children, occurring naturally in about 2–4% of the population; and it is of basic interest because it reflects the neural impairment which can occur when normal visual development is disrupted. The damage produced by amblyopia is generally expressed in the clinical setting as a loss of visual acuity in an apparently healthy eye, despite appropriate optical correction; however, there is a great deal of evidence showing that amblyopia results in a broad range of neural, perceptual, and clinical abnormalities. Currently, there is no positive diagnostic test for amblyopia. Instead, amblyopia is diagnosed by exclusion: in patients with conditions such as strabismus and anisometropia, a diagnosis of amblyopia is made through the exclusion of uncorrected refractive error and underlying ocular pathology. Amblyopic patients (especially those with strabismic amblyopia) often exhibit crowding problems, meaning they have better visual acuity when letters are presented in isolation than when they are presented in a line or a full chart.

Amblyopia comes in different sizes (degree of loss) and flavors (types). The presence of amblyopia is almost always associated with an early history of abnormal visual experience: binocular misregistration (i.e., strabismus – a turned eye), image degradation (high refractive error and astigmatism, anisometropia), or, less commonly, form deprivation (congenital cataract, ptosis). The severity of the amblyopia appears to be associated with the degree of imbalance between the two eyes (e.g., dense unilateral cataract results in severe loss), and to the age at which the amblyogenic factor occurred. Precisely how these factors interact is as yet unknown, but it is evident that different early visual experiences result in different functional losses in amblyopia, and a significant factor that distinguishes performance among amblyopes is the presence or absence of binocular function. Binocular function is much more likely to be damaged when amblyopia results from binocular misregistration (strabismus) than from image blur (anisometropia).

The Site(s) of Amblyopia A longstanding question is the site of damage in amblyopia. Current opinion places the earliest functional physiological abnormalities in cortical area V1. Exhaustive anatomical and physiological experiments failed to find retinal alterations in monkeys reared with experimentally induced amblyopia. These same animals had marked

525

526

Visual Acuity Related to the Cornea and Its Disorders

abnormalities in V1. Moreover, although human electroretinogram (ERG) studies are equivocal, after optimizing optical focus, fixation alignment, and fixation stability, Robert Hess and colleagues found no pattern ERG deficit in deep amblyopes, in a spatial frequency range where there were obvious psychophysical deficits for the same stimuli. Although it is possible that retrograde degeneration may affect the lateral geniculate nucleus (where there is some shrinkage of the cells in the parvocellular layers) and retina, it seems unlikely that these effects contribute significantly to the behavioral losses. In contrast, amblyopia results in profound alterations in V1 both in cats and monkeys. In monkeys, visual deprivation (via lid suture) leads to a massive loss of neurons in V1 that can be driven by the deprived eye. Experimentally induced blur during development leads to a selective loss of V1 neurons tuned to high spatial frequencies and the spatial tuning of neurons may be markedly different when tested through the two eyes. Experimentally induced strabismus disrupts the binocular connections of cortical neurons. It is difficult to draw distinctions based on the type of rearing from the physiology, because the effects of abnormal visual experience are complicated by the onset, duration, and depth of deprivation; however, there is evidence that the physiological deficits in amblyopia do not fully explain the behavioral losses (in the same monkeys), suggesting that there may be deficits downstream from V1. The most dramatic changes in V1 involve alterations in binocularity. Specifically, neurons that appeared to be monocular often demonstrate clear binocular interactions during dichoptic stimulation. In strabismic (prism-reared) monkeys, there is marked binocular suppression during dichoptic stimulation suggesting that inhibitory connections are less susceptible to the effects of strabismus than excitatory connections. Interestingly, even very brief periods (just 3 days) of prism-induced strabismus at the height of the critical period (4 weeks in monkeys, which translates to about 4 months in humans) increased the prevalence of V1 neurons that exhibited binocular suppression without altering their sensitivity to interocular spatial phase disparity. This result suggests that the earliest change in V1 is increased binocular suppression and, importantly, that the suppression originates at a site downstream from where information from the two eyes is first combined. Very much less is known about the physiological effects of amblyopia on visual areas downstream from V1. Brainimaging studies using positron emission tomography and functional magnetic resonance imaging show a clear deficit in V1, and several studies have also found deficits in other areas (e.g., V2). However, it is difficult to discern whether these downstream losses are simply a pass-through effect from V1 or whether the V1 losses are amplified downstream. However, several imaging and psychophysical

studies are consistent with the idea that the abnormalities in V1 are amplified in V2 and possibly beyond. These studies show losses in second-order detection global form and motion integration, symmetry detection, and counting.

Sensitive Periods for the Development of Amblyopia Clinicians are well aware that amblyopia does not develop after 6–8 years of age, suggesting that there is a sensitive period for the development of amblyopia; however, in humans with naturally occurring amblyopia, the age of onset of the amblyogenic condition(s) is difficult to ascertain, and the effects of intervention combine to make it difficult to obtain a clear picture of the natural history of amblyopia development. Thus, much of our current understanding of the development of amblyopia accrues from animal studies, and from retrospective studies of clinical records. Technological improvements in infant testing have also provided more direct data on the development of naturally occurring amblyopia in humans and monkeys. All of these studies provide strong evidence for amblyopia induced by early deprivation. While the upper limit for susceptibility of binocular interactions (binocular summation and stereopsis) is not yet certain, it appears to be later than that for acuity or contrast sensitivity in monkeys, and may extend to at least 7 or 8 years (and possibly more) in humans. Psychophysical studies of interocular transfer in humans with a history of strabismus provide an indirect estimate of the period of susceptibility of binocular connections. The results of both studies suggest that binocular connections are highly vulnerable during the first 18 months of life, and remain susceptible to the effects of strabismus until at least age 7 years.

Traditional Treatment of Amblyopia For centuries, the primary treatment for amblyopia has consisted of patching or penalizing the fellow preferred eye, thus forcing the brain to use the weaker amblyopic eye. Typically, patients with mild to moderate amblyopia are prescribed complete occlusion for 2–6 waking hours per day, over several months to more than a year. Patients with moderate to severe amblyopia are often prescribed 6–10 h or more a day, and some clinicians recommend more aggressive full-time occlusion for severe amblyopia. As reported in a recent large-scale clinical study of children (3–8 years of age), the dose–response rate for occlusion is approximately 0.1 log unit (1 chart line) per 120 h of occlusion, and the treatment efficacy is 3–4 logMAR lines. The dose–response curve appears to plateau only after 100–400 h. The treatment outcome is dependent on

527

occlusion dose, the depth of amblyopia, binocular status, fixation pattern, the age at presentation, and patient compliance. Recent clinical studies suggest that atropine penalization may be just as effective as patching. The notion that there is a sensitive period (or periods) for the development of amblyopia has often been taken to indicate that there is also a critical period for the treatment of amblyopia. This concept grew out of the work of Claude Worth in 1903. Worth suggested that the presence of a sensory obstacle (e.g., unilateral strabismus) arrested the development of visual acuity (amblyopia of arrest), so that the patient’s acuity remained at the level achieved at the time of onset of strabismus. In this view, the depth of amblyopia is a direct function of the age of onset of the sensory obstacle. Worth further suggested that, if amblyopia of arrest were allowed to persist, amblyopia of extinction could occur as a result of binocular inhibition. In Worth’s view, only this extra loss of sensory function (i.e., the amblyopia of extinction) could be recovered by treatment. Although this latter notion is open to question in the light of present knowledge, the ideas of Worth have had a powerful influence upon both clinicians and basic scientists. Many of our currently held concepts of amblyopia, such as plasticity, sensitive periods, and abnormal binocular interaction, were already described more than a century ago, and gained currency with the work of Hubel and Wiesel in 1970 and the many anatomical and physiological studies that followed. Consequently, while amblyopia can often be reversed when treated early, treatment is generally not undertaken in older children and adults. Below we consider both experimental and clinical evidence for plasticity in the adult visual system that calls into question the notion of a sensitive period for treatment.

Experimental Treatment of Amblyopia Beyond the Sensitive Period

Clinical Studies

Anatomisch

Amblyopia

A B C D E

F

G

H

Visus 1.5 1.25 1.0 Funktionell

It is often stated that humans with amblyopia cannot be treated beyond a certain age; however, a review of the literature suggests otherwise. Recent clinical trials suggest that in children, 2 h of patching per day may be just as effective as 6 h per day. Moreover, treatment may be just as effective in older (13–17 years) patients who have not been previously treated as in younger (7–12 years) children. Plasticity in adults with amblyopia is also dramatically evident in the report of amblyopic patients whose visual acuity spontaneously improved in the wake of visual loss due to macular degeneration in the fellow eye. There are also reports suggesting that some adult amblyopes recover vision in their amblyopic eye following loss of vision in their fellow (nonamblyopic) eye. These studies are consistent with the notion that the connections from the amblyopic eye may be suppressed rather than destroyed. Loss of the fellow eye would allow these existing connections to be unmasked, as occurs in adult cats with retinal lesions (Figure 1).

Adults are capable of improving performance on sensory tasks through repeated practice or perceptual learning (‘yes, you can teach old dogs new tricks!’), and this learning is considered to be a form of neural plasticity that also has consequences in the cortex. Specifically, in adults with normal vision, practice can improve performance on a variety of visual tasks, and this learning can be quite specific (to the trained task, orientation, eye, etc.). Interestingly, similar neural plasticity exists in the visual system of adults with naturally occurring amblyopia due to anisometropia and/or strabismus, suggesting that perceptual learning may be a useful approach for amblyopia treatment. Perceptual learning can improve visual functions in amblyopia on a wide range of tasks, including: Vernier acuity, positional acuity, contrast sensitivity, and letter identification. Practicing each of these tasks results in improved performance on the practiced task. The specificity of perceptual learning noted above poses some interesting difficulties. If the improvement following practice was solely limited to the trained stimulus, condition and task, then the type of plasticity documented here would have very limited (if any) therapeutic value for amblyopia, since amblyopia is defined primarily on the basis of reduced Snellen acuity. Importantly, perceptual learning of many tasks (e.g., Vernier acuity, position discrimination, contrast sensitivity) appears to transfer, at least in part, to improvements in Snellen acuity, as does practicing contrast detection. In addition to visual acuity improvement, other degraded visual functions such as stereoacuity and visual counting improve as well.

0.8 0.6

Sehfunktion

0.4 0.2 0 Entwicklungspotenz

0

2

4

6

8

10 12 Jahre

14

16

18

20

Figure 1 The postnatal development of visual function. Cartoon illustrating visual functions (sehfunktion) developing at somewhat different rates, while the developmental potential (entwicklungspotenz, in the lower panel) dissipates over the years (Jahre). Reproduced from Teller, D. Y. and Movshon, J. A. (1986) Visual Development. Vision Research 26: 1483–1506.

528

Visual Acuity Related to the Cornea and Its Disorders

Acuity improvement (%)

See also: Astigmatism.

Further Reading

10

VPHDE Positional acuity Visual acuity OT: age 3−8 years OT: age 6−8 years (n = 2)

PVEHR EHVDF NUZFE

1 0

20

40 Hours

60

80

Figure 2 Improvement in positional acuity (solid circles) and Snellen acuity (open circles) of a severe juvenile amblyope (observer AL, 8.8 years old, with unilateral strabismus). The triangle shows the improvement based on occlusion alone. The gray line shows the improvement based on occlusion alone (OT: occlusion therapy) in 2 amblyopes (aged 6–8) with acuities similar to that of AL. Replotted from Li et al., 2007; Stewart et al., 2004, 2005, 2007.

Perceptual Learning as a Clinical Tool for Treating Amblyopia Occlusion therapy is the gold standard method for treating amblyopia. In all previous perceptual learning studies, the subjects are occluded while performing the visual task, so it is reasonable to ask whether active perceptual learning actually provides an added benefit over occlusion alone. Recent work suggests that occlusion plus perceptual learning may be more effective than occlusion alone (Figure 2). Combining occlusion with perceptual learning may be a useful method for obtaining the optimal treatment outcome in the shortest possible time. Eliminating or reducing the need to wear an eye patch in public would eliminate, or at the very least reduce, the emotional stress that often accompanies occlusion therapy. Over the centuries, there have been numerous attempts to increase the effectiveness of treatment. These attempts have a long and chequered history, ranging from the sublime to the ridiculous, and include: subcutaneous injection of strychnine, electrical stimulation of the retina and optic nerve, flashing lights, red filters and rotating gratings, administration of Levodopa/Carbidopa and shocks to the brain via transcranial magnetic stimulation. Few were subjected to rigorous scrutiny, and those that were often failed to stand up to it. Thus, any promising new method should be examined critically and there is a clear need for careful controlled studies.

Acknowledgment This work was supported by National Eye Institute grant R01EY01728 from the National Eye Institute.

Ciuffreda, K. J., Levi, D. M., and Selenow, A. (1991). Amblyopia: Basic and Clinical Aspects. Stoneham, MA: Butterworth-Heinemann. Fahle, M. (2004). Perceptual learning: A case for early selection. Journal of Vision 4: 879–890. Harwerth, R. S., Smith, E. L., III, Duncan, G. C., Crawford, M. L. J., and von Noorden, G. K. (1987). Multiple sensitive periods in the development of the primate visual system. Science 232: 235–238. Hubel, D. H. and Wiesel, T. N. (1970). The period of susceptibility to the physiological effects of unilateral eye closure in kittens. Journal of Physiology (London) 206: 419–436. Kiorpes, L. (2006). Visual processing in amblyopia: Animal Studies. Strabismus 14: 3–10. Levi, D. M. (2006). Visual processing in amblyopia: Human studies. Strabismus 14: 11–19. Levi, D. M. and Carkeet, A. (1993). Amblyopia: A consequence of abnormal visual development. In: Simons, K. (ed.) Early Visual Development, Normal and Abnormal, pp. 391–408. Oxford: Oxford University Press. Levi, D. M. and Polat, U. (1996). Neural plasticity in adults with amblyopia. Proceedings of the National Academy of Sciences of the United States of America 93: 6830–6834. Li, R. W., Provost, A., and Levi, D. M. (2007). Extended perceptual learning results in substantial recovery of positional acuity and visual acuity in juvenile amblyopia. Investigative Ophthalmology and Vision Science 48: 5046–5051. Mckee, S. P., Levi, D. M., and Movshon, J. A. (2003). The pattern of visual deficits in amblyopia. Journal of Vision 3: 380–405. Polat, U., Ma-Naim, T., Belkin, M., and Sagi, D. (2004). Improving vision in adult amblyopia by perceptual learning. Proceedings of the National Academy of Sciences of the United States of America 101: 6692–6697. Revell, M. J. (1971). Strabismus: A History Orthoptic Techniques. London: Barrie and Jenkins. Stewart, C. E., Fielder, A. R., Stephens, D. A., and Moseley, M. J. (2005). Treatment of unilateral amblyopia: factors influencing visual outcome. Investigative Ophthalmology and Vision Science 46: 3152–3160. Stewart, C. E., Moseley, M. J., Stephens, D. A., and Fielder, A. R. (2004). Treatment dose-response in amblyopia therapy: The Monitored Occlusion Treatment of Amblyopia Study (MOTAS). Investigative Ophthalmology and Vision Science 45: 3048–3054. Stewart, C. E., Stephens, D. A., Fielder, A. R., and Moseley, M. J. (2007). Modeling dose-response in amblyopia: toward a child-specific treatment plan. Investigative Ophthalmology and Vision Science 48: 2589–2594. Teller, D. Y. and Movshon, J. A. (1986). Visual Development. Vision Research 26: 1483–1506. Vereecken, E. P. and Brabant, P. (1984). Prognosis for vision in amblyopia after the loss of the good eye. Archives of Ophthalmology 102: 220–224. Wiesel, T. N. (1982). Postnatal development of the visual cortex and the influence of environment. Nature 299: 583–591. Worth, C. A. (1903). Squint: Its Causes, Pathology and Treatment. Philadelphia: The Blakiston. Wu, C. and Hunter, D. G. (2006). Amblyopia: Diagnostic and therapeutic options. American Journal of Ophthalmology 141: 175–184.

Hyperopia E Harb, New England College of Optometry, Boston, MA, USA ã 2010 Elsevier Ltd. All rights reserved.

Glossary Accommodation – Increase in optical power by the eye in order to maintain a clear image (focus) as objects are moved closer. It occurs through a process of ciliary muscle contraction and zonular relaxation that causes the elastic-like lens to round up and increase its optical power. Natural loss of accommodation with increasing age is called presbyopia. Amblyopia – Decreased vision in one or both eyes due to an amblyogenic factor (e.g., refraction, strabismus, and cataract) without detectable anatomic damage in the eye or visual pathways. Asthenopia – Vague eye discomfort arising from use of the eyes; it may consist of eyestrain, headache, and/or brow ache. It may be related to uncorrected refractive error or poor fusional amplitudes. Astigmatism – Optical defect in which refractive power is not uniform in all directions (meridians). Light rays entering the eye are bent unequally by different meridians that prevent formation of a sharp image focus on the retina. Convergence – Inward movement of both eyes toward each other, usually in an effort to maintain single binocular vision as an object approaches. Cycloplegic – An agent (typically eye drops) that paralyzes the accommodation system of the crystalline lens to eliminate variability in retinoscopy or refraction measures. Emmetropization – An active growth process in which the eye grows toward a refractive error of plano. Esotropia – Eye misalignment in which an eye deviates inward (toward nose) while the other fixates normally. Strabismus – Eye misalignment caused by extraocular muscle imbalance: one fovea is not directed at the same object as the other.

Definition and Classifications of Hyperopia Hyperopia, also termed hypermetropia or farsightedness, is a common refractive error in children and adults.

In a hyperopic eye, parallel rays of light entering the eye reach a focal point behind the plane of the retina, while accommodation is in a relaxed state. Hyperopia can produce variable symptoms, depending on the magnitude of hyperopia, the age of the individual, the status of the accommodative and convergence systems, and the demands placed on the visual system (distance vs. near). Individuals with uncorrected hyperopia may experience symptoms such as blurred vision, asthenopia, accommodative/binocular dysfunction, amblyopia, and/ or strabismus. The vast majority of cases of hyperopia are of a physiological nature. In terms of physiological optics, hyperopia occurs when the axial length of the eye is shorter than the refracting components the eye requires for light to focus precisely on the photoreceptor layer of the retina. Hyperopia may result in combination with or isolation from a relatively flat corneal curvature, insufficient crystalline lens power, increased lens thickness, short axial length, or variance of the normal separation of the optical components of the eye relative to each other. Astigmatism, the most common refractive error, is often present in conjunction with hyperopia. High hyperopia is associated with high levels of astigmatism, suggesting a breakdown in the process of emmetropization that results in a component-type refractive error. Hereditary factors are probably responsible for most cases of refractive error, including physiologic hyperopia, with environment playing some role in influencing the development and degree of the error. The American Optometric Association Clinical Practice Guideline has outlined the classifications for hyperopia in the following ways: 1. Hyperopia can be classified on the basis of structure and function. These three classifications are based solely on the structure of the eye: . simple hyperopia, due to normal biological variation, can be of axial or refractive etiology; . pathological hyperopia is caused by abnormal ocular anatomy due to mal-development, ocular disease, or trauma; and . functional hyperopia results from paralysis of accommodation. 2. Clinically, hyperopia may also be categorized based on degree of the refractive error. The magnitude of hyperopia can be classified in the following three ways: . low hyperopia consists of an error of þ2.00 diopters (D) or less;

529

Visual Acuity Related to the Cornea and Its Disorders

The sum of latent and manifest hyperopia is equal to the magnitude of hyperopia. Active accommodation mitigates some or all of hyperopia’s adverse effects on vision. The impact of accommodation is highly dependent on age, the amount of hyperopia and astigmatism, the status of the accommodative and vergence systems, and the demands placed upon the visual system. Latent hyperopia is typically overcome in the young patient by the action of accommodation, but may not be sustainable for long periods of time under conditions of visual stress. Signs and symptoms such as blurred vision, asthenopia, accommodative and binocular dysfunction, and strabismus may develop. These signs and symptoms occur more readily and to a greater degree in manifest hyperopia. In general, younger individuals with lower degrees of hyperopia and moderate visual demands are less adversely affected than older individuals, who have higher degrees of hyperopia and more demanding visual needs. Pathological hyperopia is a rarer type of hyperopia, is typically greater in magnitude, and may be due to several etiologies: mal-development of the eye during the prenatal or early postnatal period, a variety of corneal or lenticular changes, aphakia, chorioretinal or orbital inflammation or neoplasms, neurological- or pharmacologicalbased etiologies, micro- or nan-ophthalmia, anterior segment malformations, and acquired disorders. A number of congenital and/or genetic developmental disabilities and syndromes are also associated with high hyperopia.

Prevalence Although it is difficult to discuss the prevalence of hyperopia due to variations in its definition by researchers (e.g., with or without cycloplegia, spherical equivalent, and least hyperopic meridian), the prevalence of hyperopia is age related. Most full-term infants are mildly hyperopic (approximately þ2.00 D), while premature infants and those of low birth weight tend to be either less hyperopic or myopic (approximately þ0.24 D). The prevalence of

45 40

Infants Childern

35 30 25 20 15 10 5 0

−6.50 −5.50 −4.50 −3.50 −2.50 −1.50 −0.50 0.50 1.50 2.50 3.50 4.50 5.50 6.50 7.50 8.50 9.50 10.50

. moderate hyperopia includes a range of error from þ2.25 to þ5.00 D; . high hyperopia consists of an error over þ5.00 D. 3. An additional classification may be based upon the outcome of noncycloplegic and cycloplegic (1.0% cycloplentolate) refractions and takes the patient’s accommodative system in consideration: . manifest hyperopia, the amount of refractive error that is detected by noncycloplegic refraction; . latent hyperopia, the amount of refractive error that is detected only by cycloplegia, can be overcome by accommodation under noncycloplegic conditions.

Percent of children

530

Refractive error (D) Figure 1 Comparison of refractive error distributions between newborns and children illustrates a shift away from hyperopia during the emmetropization process. Reprinted from Cotter, S. A. (2007). Management of childhood hyperopia: A pediatric optometrist’s perspective. Optometry and Vision Science 84(2): 103–109.

refractive error among full-term infants has a normal bell-shaped distribution (see Figure 1). Up to 9% of 6–9-months-old infants have hyperopia greater than þ3.25 D, but this prevalence decreases to 3.6% in the 1-year-old population. Over the next 10–15 years of life, there is a further decrease in the prevalence of hyperopia and an increase in the frequency of myopia. With the development of presbyopia, latent hyperopia becomes manifest, contributing to an apparent increase in the prevalence of hyperopia.

Normal Time Course of Hyperopia Emmetropization results in a gradual decrease in the level of hyperopia in most patients (see Figure 2), but in patients who have high degrees of hyperopia the change occurs more rapidly. However, infants with high hyperopia are more likely to remain significantly hyperopic throughout childhood. These infants may also have an increased incidence of against-the-rule astigmatism, which appears to further decrease the reduction in hyperopia during emmetropization compared to hyperopic infants without significant astigmatism. During the school years, there is a slow but continued decreasing trend in the incidence and the magnitude of hyperopia, except in patients with high hyperopia, whose refractive error is more likely to remain relatively unchanged. During the years of presbyopia development, latent hyperopia may become manifest requiring the implementation of both distance and near correction, yet in older individuals (>75 years of age) a myopic refractive shift may ensue likely due to crystalline lens changes.

Hyperopia

Spherical equivalent, D

3

Present study Larsen8 Zadnlk et al18 Lue et al9 Muttl et al7

2

1

0 1

2

4

6 8 Age, y

10

12

14

Figure 2 Mean spherical equivalent refractive errors, as reported by several studies, plotted by age. A smooth curve (single exponential function) was plotted to illustrate the change in refractive error with age. Reprinted from Mayer, L., Hansen, R., Moore, B., Kim, S., and Fulton, A. (2001) Cycloplegic refractions in healthy children aged 1 to 48 months. Archives of Ophthalmology 119: 1625–1628.

Clinical Presentations of Hyperopia Young persons with hyperopia generally have sufficient levels of accommodation to maintain clear vision without producing asthenopia symptoms. However, when the level of hyperopia is too great or the accommodative reserves are insufficient due to age or fatigue, blurred vision and asthenopia may develop. Presbyopia brings an increase in absolute hyperopia, causing blur, especially at near. The influence of accommodation on the vergence system also plays a role in the presence or absence of symptoms in patients with hyperopia. Individuals with esophoria and inadequate negative fusional vergence ability are frequently symptomatic and may become esotropic as a result of the uncorrected hyperopia. Among the signs and symptoms of hyperopia are red or tearing eyes, squinting and facial contortions while reading, ocular fatigue or asthenopia, frequent blinking, constant or intermittent blurred vision, focusing problems, decreased binocularity and eye–hand coordination, and difficulty with or aversion to reading. The presence and severity of these symptoms varies widely. Some young patients with hyperopia, including those with moderate and high hyperopia, may be relatively free of signs and symptoms. Early detection of hyperopia may help prevent the complications of strabismus and amblyopia in young children. In older children, uncorrected hyperopia may affect learning abilities and in individuals of any age, it can contribute to ocular discomfort and visual inefficiency.

Risks of Uncorrected Hyperopia The major complications of moderate and high physiologic hyperopia in children are amblyopia and strabismus.

531

Infants with moderate to high hyperopia (>+3.50 D) are up to 13 times more likely to develop strabismus by 4 years of age if left uncorrected, and they are 6 times more likely to have reduced visual acuity than infants with low hyperopia or emmetropia. Children who had significant hyperopia during infancy are much more likely to develop amblyopia and strabismus by 4 years of age. The presence of anisometropic hyperopia further increases the risk of strabismus and amblyopia, especially if found beyond 3 years of age. The American Optometric Association has published guidelines stating that levels greater than 1.00 D of hyperopic anisometropia and 5.00 D of isometropic hyperopia are amblyogenic. Early detection and treatment of hyperopia may reduce the incidence and severity of consequential amblyopia and strabismus and is a major justification for universal vision evaluation of young children.

Importance of Early Detection of Significant Hyperopia Atkinson and colleagues showed that uncorrected hyperopia (>3.5 D in one meridian) might contribute to poor motor and cognitive development in younger children (9 months to 5.5 years) and/or learning problems in some older children. The precise mechanism of this relationship is unclear, but optical blur, accommodative and binocular dysfunction, and fatigue all appear to play roles. In fact, uncorrected infant hyperopia has been associated with mild delays in visuo-cognitive and visuo-motor development, but has appeared to reach the level of their emmetropic counterparts after 6 weeks of full time hyperopic spectacle wear in 3–5-year-olds. The substantial number of school-age children and young adults who have uncorrected significant hyperopia is evidence of the potential impact of this learning-related vision problem and the need for early detection screening programs.

Examination Techniques of Hyperopia Optical correction should be based on both static (normal accommodation) and cycloplegic (e.g., 1% cyclopentolate) retinoscopy, accommodative and binocular assessment, and AC/A (accommodative convergence/ accommodation) ratio. The correction should then be modified as needed to facilitate binocularity and compliance. Plus-power spherical or sphero-cylindrical lenses are prescribed to shift the focus of light from behind the eye to a point on the retina. Accommodation plays an important role in determining the prescription. Some older patients with hyperopia do not initially tolerate the full correction indicated by the manifest refraction, and many patients with latent hyperopia do not tolerate the full correction of hyperopia indicated under cycloplegia.

532

Visual Acuity Related to the Cornea and Its Disorders

However, young children with accommodative esotropia and hyperopia generally require only a short period of adaptation to tolerate full optical correction. Patients with latent hyperopia who prove intolerant to the use of full or partial hyperopic correction may benefit from initially wearing the correction only for near viewing; or alternatively, trial use of a short-acting cycloplegic agent may enhance acceptance of the optical correction. Patients with absolute hyperopia are more likely to accept nearly the full correction, because they typically experience immediate improvement in visual acuity.

Management of Hyperopia The specific elements of treatment (e.g., final spectacle prescription) should be tailored to individual patient needs. Among the factors to consider when planning treatment and management strategies are: the magnitude of the hyperopia (under dry and cycloplegic conditions), the presence of astigmatism and/or anisometropia, the patient’s age, the presence of an associated esotropia and/or amblyopia, the status of the accommodative and convergence systems, the demands placed on the visual system, and any symptoms. Among several available treatments for hyperopiarelated symptoms, optical correction of the refractive error with spectacles and contact lenses is the most commonly used modality. Newer high-index lens materials and aspheric lens designs have reduced the thickness and weight of high plus-power lenses, increasing wear ability and patient acceptance. Spectacles, especially those with lenses of polycarbonate material, provide protection against trauma to the eye and orbital area and are imperative in children. Soft or rigid contact lenses are an excellent alternative for some patients. Contact lenses not only provide better cosmesis and compliance, but they reduce aniseikonia in persons with anisometropia, improving binocularity. Multifocal or monovision contact lenses may be considered for patients who require additional near correction, but resist the use of multifocal spectacles because of their appearance. Patients who wear contact lenses are at increased risk for ocular complications due to corneal hypoxia, mechanical irritation, or infection; but nevertheless, improved vision makes contact lens wear a valuable treatment option for compliant patients. Vision therapy and modification of the patient’s habits and environment can be important in achieving definitive long-term remediation of symptoms. Such modifications include, but are not limited to, improved lighting, longer near working distances, using better-quality-printed material, taking frequent breaks when reading or working on a computer for sustained periods of time. These modifications all work to reduce the accommodative

demand placed on the patient and can allow, irrespective of spectacle wear, the patient to perform their daily near activities with less symptoms. Several refractive surgery techniques to correct hyperopia are under development. The major procedures currently being studied as possible therapies for hyperopia are – Holmium: YAG laser thermal keratoplasty, automated lamellar keratoplasty, spiral hexagonal keratotomy, excimer laser, and clear lens extraction with intraocular lens implantation. The American Academy of Ophthalmology has recently reviewed 36 research articles studying the efficacy and safety of refractive surgery for hyperopia and found that the surgery provides an effective and safe correction for lower ranges of hyperopia (<3.00 D). Although still considered to be investigative on a long-term treatment basis, laser-assisted in situ keratomileusis (LASIK) is Food and Drug Administration (FDA) approved for hyperopia up to þ6.00 D. There is no universal approach to the treatment of hyperopia. The goals of treatment are to reduce accommodative demand and to provide clear, comfortable vision and normal binocularity at all distances. It is not simply determination of the lens power required to focus light onto the retina, but a complex approach encompassing the patient’s visual needs, magnitude of accommodation, coexisting amblyopia and/or strabismus, and sensitivity. The following are specific management strategies appropriate for different age groups and conditions as outlined by the American Optometric Association’s Clinical Guidelines for Hyperopia. Young children (birth–10 years of age) with low to moderate hyperopia, but without strabismus, amblyopia, or other significant vision problems, usually require no treatment. However, even occasional evidence of decreased visual acuity, binocular anomalies, or functional vision problems may signal the need for treatment. Whereas the effects of uncorrected hyperopia may manifest as visual perceptual dysfunction, reading difficulties, or failure in school, any child with hyperopia who is experiencing learning or other school difficulties needs careful assessment and may require treatment. In most young hyperopic children, the process of emmetropization leads to a gradual reduction in the degree of hyperopia by 5–10 years of age. Some children do not go through this process, however. They remain significantly hyperopic and at increased risk for developing strabismus and amblyopia. Patients under age 5 who have hyperopia over 3.25 D appear to benefit from early optical correction to reduce the risk for strabismus and amblyopia. Clinical pediatric studies suggest that partial hyperopic prescriptions do not impede infants’ emmetropization by 36 months. Optical correction of hyperopia should generally be prescribed for young children who have moderate to high hyperopia. However, there are many differences in

Hyperopia

prescribing patterns within and between pediatric optometrists and ophthalmologists. Lyons and colleagues surveyed pediatric optometrists and ophthalmologists prescribing thresholds for hyperopia in 2-year-olds and found that 65% of pediatric optometrists used þ3.0 D of bilateral asymptomatic hyperopia as their threshold for prescribing and 25% used þ5.0 D, while pediatric ophthalmologists had less conservative thresholds with 66% using þ5.0 D and 25% using þ3.0 D. It should also be prescribed, along with other interventions (e.g., occlusion or active vision therapy), for all young patients with actual or suspected amblyopia or strabismus. Optical correction may be deferred for some patients with moderate hyperopia, but such patients should be considered at risk and re-examined periodically. Careful follow-up is essential, and frequent lens changes may be needed. It is not unusual for a significant increase in hyperopia to occur after optical correction has been worn for even a short time, due to the manifestation of latent hyperopia. It is also important to continually monitor for the presence of an esotropia during all follow-up visits. When compliance proves difficult, the clinician may encourage acceptance of the prescribed treatment by using cycloplegic agents to blur uncorrected vision. Contact lenses may be a good alternative for patients who do not comply with prescriptions for spectacle wear, especially those with anisometropia, high hyperopia with or without nystagmus, and hyperopia with accommodative esotropia. Special consideration should be given to several specific categories of problems in young children who have significant hyperopia. Prior to the onset of accommodative esotropia, which usually becomes evident at about 2–3 years of age, few children exhibit obvious signs of ocular problems, with the exception of intermittent esotropia in children who are ill or very tired. Early screening for refractive error usually detects hyperopia, but due to the relative infrequency of refractive screening, many children with underlying moderate to high hyperopia go undetected until the appearance of frank strabismus. Appropriate treatment includes the use of either single vision or multifocal spectacles depending on the patient’s binocular and accommodative status. Concurrent amblyopia, when present, may be treated by patching and active vision therapy. Less commonly, young children with bilateral high hyperopia develop isoametropic amblyopia due to the resulting constant state of severely blurred vision. Such patients may have an associated esotropia, or conversely, may not manifest esotropia because they make no attempt to accommodate. Optical correction of this condition is indicated in order to prevent or treat the associated amblyopia and/or strabismus; however, careful monitoring is warranted a previously nonexistent esotropia may present itself following correction. Partial correction may

533

inadvertently stimulate accommodative esotropia, because the patient now has good reason to attempt to overcome the remaining uncorrected hyperopia. Treatment to improve vision in the child with amblyopia may take a few years, but improvement is usually possible with full time spectacle wear and/or patching or pharmacological penalization. Many persons between the ages of 10 and 40 years who have low hyperopia require no correction because they have no symptoms. Ample accommodative reserves shelter them from visual problems related to their hyperopia. Under increased visual stress, such persons may develop symptoms that require correction. Wearing prescribed lenses with low amounts of plus power usually alleviates the problem. Patients with moderate degrees of hyperopia are more likely to require at least part-time correction, especially those who have significant near demands or have accommodative or binocular anomalies. Accommodative or binocular dysfunction associated with uncorrected low to moderate hyperopia may be treated by optical correction or vision therapy. Vision therapy may be instituted initially or after optical correction in patients who have significant binocular vision problems. The effects of visual habits and environment play an increasing role in determining the need for and characteristics of treatment. By the age of 30–35 years, most previously asymptomatic, uncorrected hyperopic patients begin to experience blur at near and visual discomfort under strenuous visual demand. Facultative hyperopia can no longer be sustained comfortably due to decreasing accommodative amplitudes. Latent hyperopia should be suspected when symptoms occur in conjunction with lower amplitude of accommodation than expected for the patient’s age. Cycloplegic retinoscopy can help identify this latent component. By the mid-1930s, accommodation takes noticeably longer while facility decreases, causing associated visual problems in many hyperopic persons previously free of symptoms. A prescription for the distance manifest (noncycloplegic) refraction for the patient to wear as needed (i.e., part time at near) often suffices. The patient may require additional correction with increasing age and visual demands at near. Before prescribing a permanent pair of spectacles, the optometrist may lend the patient a pair of spectacles (i.e., over-the-counter reading glasses) to demonstrate the potential benefits of optical correction of latent hyperopia. With the onset of presbyopia (loss of accommodation), maintaining focus at near becomes progressively more difficult, especially in poor illumination. Prescribing an optical correction for most or all of the distance manifest refraction, along with a near addition, can greatly improve vision and comfort. Hyperopia equal to or greater than þ1.00 D to þ1.50 D generally requires full-time distance correction, with a near addition for patients over about

534

Visual Acuity Related to the Cornea and Its Disorders

age 45. As facultative hyperopia becomes absolute, more plus power at distance is required. Progressive multifocal lenses enable clear focusing at a range of finite distances. A monovision, bifocal, or multifocal contact lens prescription is an option in some patients. Physiological hyperopia is not progressive. Therefore, the prognosis, which can be given at diagnosis, is generally excellent, except for those patients with both hyperopia and amblyopia or strabismus, for whom it is less certain. Appropriate optical correction almost always leads to clear and comfortable single binocular vision. Younger children who have significant hyperopia associated with amblyopia, strabismus, or anisometropia require intensive follow-up and treatment for their more complex problems, starting as early as 3–6 months of age. The timing of periodic preventive optometric care for uncomplicated hyperopia depends on the patient’s age and circumstances. For children with hyperopia, follow-up may be required as often as every 3–6 months depending on the concern for strabismus and/ or amblyopia, while for asymptomatic adults, every 1 or 2 years is generally sufficient. The underlying cause, rather than hyperopia itself, is the chief concern in patients with pathologic hyperopia. Because the causes of pathologic hyperopia are both uncommon and diverse, general statements concerning treatment must be limited to the need to correct the hyperopia in the best manner possible, depending on the underlying etiology. Conditions of a developmental or anatomic nature are rarely progressive. When useful vision is thought to be obtainable, the treatment of hyperopia resulting from nonprogressive conditions is similar to that for physiologic hyperopia. Patients with pathological hyperopia require treatment of their underlying conditions and, when indicated, referral to another eye care provider for special services. All patients treated for hyperopia with persistent symptoms require additional follow-up care to remediate their problem.

Conclusion Hyperopia is a common refractive disorder that has a close association with other consequential disorders, namely amblyopia and strabismus, especially in children. This makes hyperopia a greater risk factor for more permanent vision loss than myopia. In addition, it appears that uncorrected hyperopia has a detrimental effect on a child’s learning and development. Therefore, the early diagnosis

and treatment of significant hyperopia and its consequences can prevent a significant amount of visual disability in the general population. Because hyperopia is usually not readily apparent, preventive examination of all young children is essential. Although there is no clear consensus on the prescribing practices for hyperopia, clinicians must examine the visual needs and binocular system of each patient to best determine their final treatment plan. Periodic eye examinations are needed thereafter to help ensure the provision of treatment appropriate to the changing visual needs of the hyperopic patient. See also: Abnormal Eye Movements due to Disease of the Extraocular Muscles and Their Innervation; Amblyopia.

Further Reading AOA Consensus Panel on Care of the Patient with Hyperopia (1997). American Optometric Association Clinical Practice Guidelines: Hyperopia. http://www.aoa.org/documents/CPG-16.pdf (accessed Jun. 2009). Atkinson, J., Anker, S., Nardini, M., et al. (2002). Infant vision screening predicts failures on motor and cognitive tests up to school age. Strabismus 10(3): 187–198. Atkinson, J., Braddick, O., Nardini, M., and Anker, S. (2007). Infant hyperopia: Detection, distribution, changes and correlates – outcomes from the Cambridge infant screening programs. Optometry and Vision Science 84(2): 84–96. Atkinson, J., Nardini, M., Anker, S., Braddick, O., Hughes, C., and Rae, S. (2005). Refractive errors in infancy predict reduced performance on the movement assessment battery for children at 3½ and 5½ years. Developmental Medicine and Child Neurology 47: 243–251. Cotter, S. A. (2007). Management of childhood hyperopia: A pediatric optometrist’s perspective. Optometry and Vision Science 84(2): 103–109. Donahue, S. P. (2007). Prescribing spectacles in children: A pediatric ophthalmologist’s approach. Optometry and Vision Science 84(2): 110–114. Guzowski, M., Wang, J. J., Rochtchina, E., Rose, K. A., and Mitchell, P. (2003). Five-year refractive changes in an older population: The blue mountains eye study. Ophthalmology 110(7): 1364–1370. Lee, K. E., Klein, B. E. K., Klein, R., and Wong, T. Y. (2002). Changes in refraction over 10 years in an adult population: The beaver dam eye study. Investigative Ophthalmology and Visual Science 43(8): 2566–2571. Lyons, S. A., Jones, L. A., Walline, J. J., et al. (2004). A survey of clinical prescribing philosophies for hyperopia. Optometry and Vision Science 81(4): 233–237. Mayer, L., Hansen, R., Moore, B., Kim, S., and Fulton, A. (2001). Cycloplegic refractions in healthy children aged 1 to 48 months. Archives of Ophthalmology 119: 1625–1628. Varley, G. A., Huang, D., Rapuano, C. J., et al. (2004). LASIK for hyperopia, hyperopic astigmatism, and mixed astigmatism: A report by the American Academy of Ophthalmology. Ophthalmology 111(8): 1604–1617.

Subject Index Page numbers suffixed by t and f refer to tables and figures respectively. A AAV see Adeno-associated virus Abducens nerve (CNVI) 17 abnormal eye movements from disorders of 42 LR innervated by 18–19, 19f Aberrations 193 Ablation profile 193 Abnormal eye movements see also Eye movements EOM and 39, 40f, 41f brainstem disorders and horizontal movements 40f, 42 brainstem disorders and vertical movements 41f, 43 CNVI, CNIII, CNIV disorders effecting 42 congenital misalignment 43 disorders effecting 39 NMJ disorders and 40, 40f visual system disorders effecting 44 AC see Adenylyl cyclase; Artificial cornea ACAID see Anterior chamber-associated immune deviation Acanthamoeba keratitis 413 adaptive immune system’s resistance to 415 evading 416, 417t barriers to 414 innate immune system’s resistance to 414, 415, 415t macrophages and 414 mucosal immune system and 416, 416t neutrophils and 414 trophozoites for 417 Accommodation 487, 517, 529 hyperopia and 530 Acinar cells 83 in lacrimal gland 68, 69f, 83, 84f electrolyte/water secretion mechanisms by 72, 73f in meibomian glands 62–63 Acinus 60, 68 Actinic lesion 91 Active innate immune system see Innate immune system, active Acuity see Visual acuity ACV see Acyclovir Acyclovir (ACV) 306–308, 308t Adaptive immune system 347, 367 Acanthamoeba keratitis and 415 evading 416, 417t antigen trafficking, processing and T cell presentation in 348, 349f innate immune system’s link to 359 PRRS in 347–348 ADDE see Aqueous deficient dry eye Adenine nucleotide translocator 27 transporter 1 (Ant1) 30 transporter 2 (Ant2) 30 Adeno-associated virus (AAV) 330, 331f Adenosine diphosphate (ADP) 27 Adenosine triphosphate (ATP) 27, 284, 288 EOM and buffering system of 28–29 EOM mitochondria generating 31–32 synthase 30–31, 31f Adenylate cyclase soluble 288 VIP signal transduction and 87, 87f Adenylate kinase (AK) 28–29 Adenylyl cyclase (AC) 87 Adherens junctions (AJs) 219, 286–287 corneal epithelial wound healing and 167, 167t Adhesion complex 143 cell-matrix junctions and 167 hemidesmosomes of 145, 145f Adhesion molecules 359 corneal epithelium infection response of 455 ADP see Adenosine diphosphate Adrenergic agonists, signal transduction and 88 Ca2+ and PKC 88, 88f MAPK 88f, 89 NO 88f, 89

Afferent immune response 390 in autoimmune-based inflammation 386, 386f corneal graft rejection, mechanisms/components of 392 direct/indirect allorecognition 392 histocompatibility antigens 392 T-lymphocyte activation 393 immunoregulation and 385 AFM see Atomic force microscopy Against-the-rule 506, 507 Age-related macular degeneration (AMD) 378 AJs see Adherens junctions AK see Adenylate kinase AKC see Atopic keratoconjunctivitis Akt1 406 Aldehyde dehydrogenase (ALDH) 223 ALDH see Aldehyde dehydrogenase Alignment testing 46 Alkali burn 333 ALL see Anterior limiting lamella Allergen 419 Allergic conjunctivitis 419 Allergic eye disease 114 APCs and 378 Allergies see Ocular allergies Alloantigen 327, 361, 390 Allogeneic 185 Allograft 470 see also Keratolimbal allograft rejection 481 Alloimmunity 478 Alloreactive 390 Alpha-melanocyte-stimulating hormone 369 Alternative splicing 116 AM see Amniotic membrane transplantation Amblyopia 39, 91, 419, 525, 529 as public health problem 525 sensitive periods for development of 526 sites of 525 treatment for 47 clinical studies on 527, 527f perceptual learning as tool for 527, 528f traditional 526 types of 525 AMD see Age-related macular degeneration Amidases 303 Amniotic membrane transplantation (AM) 139 for CESC culture 187 clinical outcome of 189 indications for 188, 188f postoperative management for 187t, 189 procedure for 188 surgical procedure for 189 for fibrotic scarring 269 history/concepts of 187 ocular surface reconstruction and 188 AMPs see Antimicrobial peptides Amyloid 226 Androgens 75 Angioblast 465 Angiogenesis 465, 470 corneal avascularity as study platform for 482 process of 470 Angiogenic molecules bFGF 467 in CNV 466 in corneal angiogenic privilege 473, 473t lipid mediators 475 MMPs 475 VEGF 473 Angiogenic privilege see Corneal angiogenic privilege Angiostatin 475, 480 Anion transporters 285, 285f Aniridia 478 Aniridic keratopathy 185

535

536

Subject Index

Anisometropia 91, 525 Ant1 see Adenine nucleotide translocator Ant2 see Adenine nucleotide translocator Anterior chamber-associated immune deviation (ACAID) 352, 361, 362, 364, 365t, 368–369, 373, 390, 481 APCs and 378 induction of 362 ocular phase of 362 relevance of 364 splenic phase of 362 sympathetic nervous system and 362, 363f thymic phase of 362 Tregs in 363 Anterior lamellar keratoplasty 140 Anterior limiting lamella (ALL) 220 Anterior segment OCT (AS-OCT) 217, 218f see also Visante OCT clinical applications for 238, 239f, 241f, 243f, 246f, 248f, 249f, 251f, 253f, 254f, 255f devices for 238 principles of 238, 238f Anterior synechia 226 Anterograde axonal transport 396 Anti-angiogenic molecules 475 angiostatin 475 arresten, canstatin, tumstatin 475 in CNV 466 in corneal angiogenic privilege 473t, 475 endostatin and neostatin 475 PEDF 475 Antigen-presenting cells (APCs) 348–349, 368–369, 373, 444 ACAID-inducing 378 allergic eye disease and role of 378 AMD and 378 conjunctival 377 corneal 374 in inflammation 376, 376f, 376t LCs 374, 375t transplantation 377 of corneal stroma 375 EAU and role of 379 in HSK 398 HSK and function of 377 of ocular surface 374, 374f of retina 374f, 379, 379t of sclera 374f, 379, 380t types of 373 of uvea 374f, 378, 378t Antigens 347 adaptive immune system and, trafficking, processing and T cell presentation of 348, 349f histocompatibility 392 Antimicrobial peptides (AMPs) 433, 444 categories of 448–449 conjunctival epithelium secreting 435 corneal epithelium infection and response of 454 ocular surface defense mechanisms and 446f, 448 APCs see Antigen-presenting cells Apical squames 144, 144f Aponeurosis 17 Apoptosis 164 Pseudomonas aeruginosa keratitis and 410 AQP-1 see Aquaporin-1 Aquaporin-1 (AQP-1) 279–280, 280f, 283 in corneal endothelium 286 Aquaporins 283 Aqueous deficient dry eye (ADDE) 53 Aqueous humor immunosuppression in 367, 368t, 369f a-MSH and 369 neuropeptides and 370 SOM, CGRP, VIP in 370 T cells and 367–368 Aqueous layer of tear film 53 Arresten 475 Artificial cornea (AC) desired characteristics for 298 Kpro, types of 298, 299f need for 296, 297f regeneration with 300, 301f requirements for 297 self-assembled corneal equivalents 299

Artificial tears 388 Ascending neuromodulatory systems 493 AS-OCT see Anterior segment OCT Aspergillus keratitis 428, 429f Asphericity 193 Asthenopia 529 Astigmat 506 refraction 508 Astigmatism 201, 327, 517, 525, 529 blurring and visual perception in 509, 510f, 511f definition/etymology of 506 image formation in 508, 508f astigmat refraction 508, 512 ocular 508f, 509 oblique 511, 512f ocular classification of 509 contact lenses for 515 corneal keratometric power map for 513, 514f image formation in 508f, 509 measurement of 513 near work and 511 not correcting 516 origin of 507 prevalence and age-related changes in 507 spectacles for 515 surgery for 515 wave front technology for 205 with-the-rule/against-the-rule 507 Atomic force microscopy (AFM) 116 Atopic keratoconjunctivitis (AKC) 114, 185, 420, 420t, 478 see also Ocular allergies Atopy 419 ATP see Adenosine triphosphate Attendant nystagmus 43 Autoantigen secretion 344 Autograft 470 Autoimmunity 381 afferent immune response in inflammation of 386, 386f efferent immune response in inflammation of 386f, 387 Autosomal dominant 272 Avascular cornea 148 Axis meridian 506

B Bailey-Lovie chart 495, 495f BALT see Bronchus-associated lymphoid tissue Barriers 171 to Acanthamoeba keratitis 414 corneal endothelium and integrity of 286 cAMP-PKA axis in 289 MLC phosphorylation’s effects on 287 TJs and integrity of 286–287 Basal cell carcinoma (BCC) 95, 96f Basement membrane (BM) 143, 178 see also Epithelial basement membrane corneal epithelial basal cells adhering to 145, 145f degradation of 165 regeneration of Bowman’s membrane and 263–264 Basic fibroblast growth factor (bFGF) 467 BCC see Basal cell carcinoma BCEC see Bovine conjunctival epithelial cell BEB see Benign essential blepharospasm Beer-Lambert law 193, 197 Benign essential blepharospasm (BEB) 6, 7 causes of 7 genetics underlying 7 Best spectacle-corrected visual acuity (BSCVA) 198–199 bFGF see Basic fibroblast growth factor Bicarbonate-stimulated adenosine triphosphatase (HCO3-ATPase) 279, 280–281, 280f Binocular vision 45 Bioengineered corneas see Artificial cornea Biofilm 426 contact-lens-associated fungal keratitis and formation of 426, 427f Bioreversion 303–304 Blepharitis 91 chronic 65 in eyelids 97–98 posterior 97–98

Subject Index

seborrheic 97–98 staphylococcal 97–98 Blepharoplasty 91 Blepharospasm 3 benign essential 6, 7 causes of 7 genetics underlying 7 Blink oscillations 6, 6f Blink-evoked eye movements 5–6 Blinking anatomical organization of 4, 4f cornea irritation and 6, 6f modifiability of 6 protective function of 6 reflex, omnipause neurons blocking 6 saccades combined with 5–6 tear film distribution and 55 Blood vessels 129 BM see Basement membrane Bone-marrow-derived cells in corneal epithelial wound healing and ECM communication 265 in corneal stroma 223 in lacrimal gland 69 Bony orbit EOM in, anatomy of 17 muscles in 17, 18f Bony passage 127, 128f Botulinum toxin A 47–48 Bovine conjunctival epithelial cell (BCEC) 310–312, 312f, 312t Bowman’s membrane 133, 146 dystrophies of 227 Reis-Buckler’s dystrophy 227, 227f Thiel-Behnke 227 regeneration of BM and 263–264 Brainstem, abnormal eye movements and disorders of horizontal movements 40f, 42 vertical movements 41f, 43 BrdU see Bromo-deoxyuridine Bromo-deoxyuridine (BrdU) 291 Bronchus-associated lymphoid tissue (BALT) 340, 438 BSCVA see Best spectacle-corrected visual acuity

C Ca2+ see Cytosolic calcium concentration CaCC see Calcium-activated chloride channel Calcitonin gene-related peptide (CGRP) 54, 83–84, 153–154, 154f in aqueous humor 370 Calcium-activated chloride channel (CaCC) 285, 285f CALT see Conjunctiva-associated lymphoid tissue cAMP see Cyclic monophosphate cAMP-PKA axis 289 Canaliculi, tear transport through 128 Candida keratitis 426 see also Fungal keratitis Canstatin 475 Canthal tendons (CT) 4f Capacitance veins, blood flow in 129 Capsid 327 Caspases 452 Cataract surgery, corneal nerve regeneration after 156 Cathelicidin (LL-37) 356 corneal epithelium infection response of 455 CD4+ T cells 381, 385 CD44 452, 456 CD80 339, 379 CD86 74, 339 CDKs see Cyclin-dependent kinases CEC see Corneal endothelial cells Cell cycle 290 Cell therapy 178 Cell-adhesion molecule 459 Cell-cell junctions in corneal endothelium anatomy 272, 273f corneal epithelial wound healing and 166, 261 AJs 167, 167t cytokines and GFs in 261, 262f desmosomes 167, 167t gap junctions 166, 167t TGF-b signaling pathways in 263 tight junctions 167, 167t

Cell-matrix junctions, corneal epithelial wound healing and 166 hemidesmosomes and adhesion complexes 167 Cell-volume control 175 Central cloudy dystrophy 231 CESCs see Corneal epithelial stem cells CFTR see Cystic fibrosis transmembrane conductance regulator CGRP see Calcitonin gene-related peptide Chalazion 66, 66f in eyelids 97, 97f Chamber angle 272 Channels 500 ChAT see Choline acetyltransferase CHED see Congenital hereditary endothelial dystrophy Chemokines 74, 396 corneal epithelium infection and response of 454 in HSK 398 innate immune system regulation of 359 in ocular allergies 425 Pseudomonas aeruginosa keratitis and 407 Chemosis 213 Chemotaxis 470 Chemotropic guidance 150 Choline acetyltransferase (ChAT) 33, 37 Choline transporter (ChT) 33, 37 Cholinergic agonists 121 signal transduction and 112, 113f protein secretion in 112–113 Chronic blepharitis 65, 91 Chronic progressive external ophthalmoplegia (CPEO) 27, 39 diplopia in myasthenia gravis compared to 41 occurrence of 30 slow progression of 39 ChT see Choline transporter Cicatrical entropion 94 Cicatricial ectropion 95 Cicatrization 185 CK see Conductive keratoplasty; Creatine kinase Cl channels 176 CLAU see Conjunctival limbal autograft Clodronate 413 CLPC see Contact-lens-induced papillary conjunctivitis CNIII see Oculomotor nerve CNIV see Trochlear nerve CNV see Corneal neovascularization CNVI see Abducens nerve Cochet-Bonnet esthesiometer 150, 162 Cold-sensitive thermal receptors 159–160 Collagen 296 bundles 220, 221f in corneal stroma 223–224 in Golgi tendon organs 36, 36f Collagenase (microbial) 459 Complement 347 active innate immune system role of 358 corneal epithelium infection and response of 455 passive innate immune system role of 357 Conductive keratoplasty (CK) 133, 140–141 Confocal microscopy 215, 217f, 234 clinical applications of 235 normal cornea 235, 235f, 236f history of 234 pathological applications of 236, 237, 237f principles of 234, 235f Confocal microscopy through focusing (CMTF) 236 Congenital hereditary dystrophy endothelial 232 corneal endothelium and 281 stromal 231 Congenital hereditary endothelial dystrophy (CHED) 281 Congenital misalignment (infantile strabismus) 43 Conidia 426 Conjugate gaze 46, 46f Conjunctiva 99–100, 178, 419 allergic reactions in 461, 461f APCs of 377 corneal infection compared to 462, 463f epithelium AMP secretion of 435 fluid transfer across 100 immune surveillance of 432, 432f, 433f, 434f innate immune system defense factors of 433 PRRs and 433, 433f

537

538

Subject Index

Conjunctiva (continued) gene therapy 335 goblet cells in 100, 108, 109, 109f, 111f mucin synthesis/secretion by 110, 111f, 112, 113f proliferation of 110, 111, 111f IgA produced by 434f, 436 immunoregulation of 438 membrane types in 308–309 PD of 102–103 rabbit bioelectric studies on 102, 102f electrolyte transport systems of 103 epithelia of 101, 101f epithelial ion transport regulation in 104, 105f, 106f fluid transport studies across isolated 105 roles of 100 stem cells 182 tear fluid connecting cornea and 460, 461f inflammation and 463 topical drug delivery and role of 308 prodrug strategy in 310 transconjunctival pathway in 310, 311f, 312f, 312t transepithelial resistance of, determining 103 ussing chamber isolating 103 Conjunctiva-associated lymphoid tissue (CALT) 340, 431, 434f, 435 corneal immune surveillance with 437, 439f deregulation of 441f EALT and 438, 440f immunoregulation in 438, 441f topography of 439f Conjunctival epithelial cells 423, 424t Conjunctival fibroblasts 424 Conjunctival flap 139 Conjunctival lamina propria immune surveillance of 434f, 435 macrophages and 434f, 436 neutrophils protecting 437 T-lymphocytes in 434f, 435 Conjunctival limbal autograft (CLAU) 187 Conjunctival lymphoid follicles 437, 438f, 439f, 441f Conjunctival metaplasia 470 Connexin 219 Constitutive exocytosis 70–71 Contact lenses advantages of 207 corneal infection and 211 fungal keratitis and 426 biofilm formation in 426, 427f lipid layer of tear film and 66 MPSs and care products for 211 mucins and 124 for myopia 523 for ocular astigmatism 515 ocular surface and effects of 209 alterations for 210 corneal epithelium/limbal epithelium 209 infection and 210 inflammation and 210 mechanical effects 211 orthokeratology for 211 types of 207, 208f, 208t, 209f Contact-lens-induced papillary conjunctivitis (CLPC) 114, 207, 210 Continuous jerk nystagmus 44 Contrast constancy 500 Contrast-detection threshold 500 forced-choice paradigms for 500, 501f Contrast-sensitivity function (CSF) 193, 196–197, 500, 502, 504f pre-/post LASIK 197f temporal 501f, 503, 504f, 505f Convergence 529 Corectopia 226 Cornea see also Artificial cornea; specific parts anatomy of layers of 133, 305f APCs of 374 in inflammation 376, 376f, 376t LCs 374, 375t transplantation and 377 avascular 148 Bowman’s membrane 133, 146 dystrophies of 227

confocal microscopy for normal 235, 235f, 236f conjunctiva infection compared to 462, 463f corneal angiogenic privilege and 468 Descemet’s membrane and 134, 134f, 148 disease processes of 135 endothelial disease 138 epithelial disease 135 Fuch’s dystrophy 138 keratoconus 137, 138f pseudophakic bullous keratopathy 138 stromal disease 137 subepithelial disease 136 Thygeson’s punctate keratitis 136 epithelial basement membrane 133 epithelium 133 function of 134 as protective barrier 134 transparency and 134, 135f histology of 478, 479f homeostasis of FGF7 and maintenance of 323, 324f LSC in 182 hydration control of 171 immune privilege of 390 factors contributing to 390 site 391 immune surveillance of, CALT assisting 437, 439f infection of 211 infiltrates 207 inlays 205 irritation of, blinking and 6, 6f periphery, palisades of Vogt at 146, 148f refractive surgery and biomechanical effects of 198 sensitivity of injured 162 intact 162 LASIK and 162 sensory receptors of functional properties of 158, 159f, 160f local inflammation and 160f, 161 stem cells of 108 structure of 305f surgical intervention for 138 tear fluid connecting conjunctiva and 460, 461f inflammation and 463 topical drug delivery and role of 305, 305f, 307f, 308t, 309f transparency of 171, 172–173 transplants for 296–297 APCs function in 377 Corneal alymphaticity 481 Corneal angiogenic privilege 465, 466f angiogenic molecules in 473, 473t anti-angiogenic molecules in 473t, 475 cornea/corneal epithelium and 468 immune privilege and 467 regulatory factors of 473, 473t Corneal avascularity as angiogenesis study platform 482 immune privilege of 481 loss of 481 maintenance of 470, 471 optical clarity and 479 pro-/anti-angiogenic forces of 479 transplants and 481 Corneal biopsy 140 Corneal dystrophies 219, 226, 290 anterior 226 EBMD 226 Meesman’s lesions 227 of Bowman’s membrane 227 Reis-Buckler’s dystrophy 227, 227f Thiel-Behnke 227 gene therapy treating 332, 334 alkali burn 333 CNV 334 corneal epithelial wound healing 332 corneal graft rejection 332 fibrotic scarring 333 granular 229 type 1 229, 229f type 2 229

Subject Index

macular 219, 230 posterior 231 congenital hereditary endothelial dystrophy 232 Fuch’s dystrophy 231, 231f polymorphous disease 226 stromal 228 central cloudy dystrophy 231 congenital hereditary 231 fleck dystrophy 231 gelatinous drop-like 230 lattice types of 228, 228f posterior amorphous 231 rare 230 Schnyder’s crystalline dystrophy 230 Corneal ectasia 201 Corneal endothelial cells (CEC) background for 290, 291f cell-cell contact inhibition of 291 E2F promoting proliferation of 294 ECM interactions with 293 FGF-2 signaling pathway and proliferation of 294 G1 phase of 290, 292f cell cycle arrest in 291, 292f G1/S transition of 290–291, 292f human 291, 292f age-related decrease in sensitivity to mitogen/GF in 293 proliferative capacity from central v. peripheral regions of 294 PKC signaling pathways regulating proliferation of 294 progression of 290, 291f TGF-b in aqueous humor of 293 Corneal endothelium 134, 135f, 178 active transport and 284 regulation of 288 anatomy of 272, 273f cell-cell junctions in 272, 273f anion transporters and channels in 285, 285f AQP-1 in 286 barrier integrity in 286 cAMP-PKA axis in 289 MLC phosphorylation’s effects on 287 regulation of 288 bicarbonate/carbonic anhydrase and 285 biochemistry and metabolism in 275 cell division and replenishment in 276, 276f, 277f cytokines and immune privilege in 277, 278f deturgescence and role of 273f, 279, 279f biochemistry of active 279, 280f physiological control of active 280 development of 273, 274f, 275f, 276f function of 283 stromal swelling pressure and maintenance of transparency for 283, 284f TJs of 283, 284f genetic diseases of 281 Fuch’s dystrophy 281 PPMD/CHED 281 glucose and energy metabolism 275 nutrition/waste removal of 284 proteins synthesized for external transport in 278, 278f pump-leak mechanism in 286 stem cells 183 transcription factor’s role in 274, 277f Corneal epithelial stem cells (CESCs) 180 AM as carrier for 187 clinical outcome of 189 indications for 188, 188f postoperative management for 187t, 189 procedure for 188 surgical procedure for 189 ex vivo expansion of 185, 187 location of 145–146 LSC characteristics of 180, 181t in corneal homeostasis 182 deficiency of 181 niche of 179f, 181 at palisades of Vogt 145–146, 147f in vivo expansion of 186 Corneal epithelial wound healing 164, 261, 261f, 262f cell-cell junctions and 166, 261 AJs 167, 167t cytokines and GFs in 261, 262f

desmosomes 167, 167t gap junctions 166, 167t tight junctions 167, 167t cell-matrix junctions and 166 hemidesmosomes and adhesion complexes 167 corneal nerves and 166 ECM communication in 264 bone-marrow-derived cells in 265 cellular responses in 265 suppressing/amplifying factors of 264 gene therapy treating 332 keratocyte changes during 265–266 MMPs and 169 phases of 164, 165f, 266, 266f active wound-healing phase 267 inflammatory phase 266 remodeling phase 267 termination of fibrotic repair response 262f, 267, 268f secreted factors involved in 168 TGF-b signaling pathways in 263, 323 Corneal epithelium 99–100, 133 apical squames of 144, 144f basal cells adhering to BM of 145, 145f Ca2+ transport and 175 contact lens effects on 209 corneal angiogenic privilege and 468 functions of 143, 144f HSV-1 eradication from 396 infection of 452 adhesion molecule response to 455 AMP response to 454 complement system response to 455 cytokine/chemokine response to 454 hBD response to 454 LL-37 response to 455 neuropeptide response to 457 sIgA response to 455 SP response to 456 thymosin-b4 for 457 TLR response to 453 VIP response to 456 ion transport in 171, 172f importance of 172 primary/secondary 173 receptor-mediated control of 175 stromal deturgescence coupled with 173 K+, Cl , Na+ channels and 176 KLAL and transplantation of indications for 186, 186f postoperative management for 187, 187t surgical procedure for 186 mouse lines 319, 319f, 320f, 321f, 322f in passive innate immune system 355 PTK for defects of 191 structure/function of 452 transgenic and knock-out mice and specific promoter of 318 Corneal equivalents, self-assembled 299 Corneal erosion, recurrent 164, 226 Corneal gene therapy 327 methods for 327 AAV 330, 331f lentivirus vectors 330–331, 331f nonviral vectors 331 viral vectors 330, 330f, 331f Corneal glue 139 Corneal graft 327 Corneal graft rejection 332 afferent immune response mechanisms/components of 392 direct/indirect allorecognition 392 histocompatibility antigens 392 T-lymphocyte activation 393 clinical features of 391, 391f efferent immune response mechanisms/components of 393 immune privilege breakdown 394 macrophages 394 NK cells 394 T-lymphocyte activation 393 pathogenesis of 392, 392t prevention of 394 treatment of 394 Corneal guttata 226

539

540

Subject Index

Corneal keratometric power map for 513, 514f Corneal laceration repair 140 Corneal lenticules 185 Corneal neovascularization (CNV) 327, 478, 517 angiogenic/anti-angiogenic molecules in 466 clinical manifestations of 471, 472f etiology/epidemiology of 470, 471f, 472f, 472t gene therapy for 334 medical treatments for 476 surgical treatment for 476 therapy for 476 VEGF in 466 Corneal nerves architecture of 150 intraepithelial nerve terminals 153, 154f limbal plexus 150 SEP 151 stromal plexus 151 subbasal nerve plexus 152, 153f corneal epithelial wound healing and 166 injuries and 162 neurochemistry of 154, 154f origins of 150, 151f regeneration of cataract surgery 156 ocular surgery 155 PK 156, 157f refractive surgery 156 regrowth mechanisms of 157 remodeling 155 Corneal pannus 470 Corneal refractive surgery see Refractive surgery Corneal scars see Fibrotic scarring Corneal staining 211 Corneal stroma 134, 146, 178, 327 anatomy of 220, 221f APCs of 375 autoantigen secretion to 344 bone-marrow-derived cells in 223 collagen bundles in 223–224 development of 220 dystrophies of 228 central cloudy dystrophy 231 congenital hereditary 231 fleck dystrophy 231 gelatinous drop-like 230 lattice types of 228, 228f posterior amorphous 231 rare 230 Schnyder’s crystalline dystrophy 230 ECM 223 keratocytes of 221f, 222 mouse lines 317t, 318 PGs in 224 stem cells 183, 223 swelling pressure and maintenance of transparency of 283, 284f transparency of 223 Corneal verticillata 213 Corticosteroid 256, 268t, 269, 310–312 for dry eye 388 Cover test 45 CPEO see Chronic progressive external ophthalmoplegia Cre 315 CRE-adenovirus vector 327 Creatine kinase (CK) in EOM metabolism 28 sarcomeric mitochondrial 28–29 ubiquitous mitochondrial 28–29 Cre-LoxP system 319 inducible 316, 317t pitfalls of 316, 316t tissue-specific gene ablation using 316 Critical flicker frequency 500 Crypts 178 Crystallins 219, 223 CSF see Contrast-sensitivity function CT see Canthal tendons CTL see Cytotoxic T lymphocyte CTMF see Confocal microscopy through focusing CX3CR1 373 Cycles per degree 494

Cyclic monophosphate (cAMP) 72, 288 cAMP-PKA axis, in corneal endothelium barrier integrity 289 VIP and production of 87, 87f Cyclin-dependent kinases (CDKs) 290 Cyclins 290 Cyclopetic refraction 201 Cycloplegic 529, 531–532 Cyclosporine 388 Cystic fibrosis transmembrane conductance regulator (CFTR) 104 Cytokeratin 178 Cytokines 347, 396 corneal endothelium, immune privilege and 277, 278f corneal epithelial wound healing and 261, 262f corneal epithelium infection and response of 454 in HSK 397 innate immune system regulation of 359 in ocular allergies 425 Pseudomonas aeruginosa keratitis and 407 Th1 459 Th2 459 Cytoskeleton 164 subapical actin 71 Cytosolic calcium concentration (Ca2+) adrenergic agonist signal transduction and 88, 88f corneal epithelium and transport of 175 Cytotoxic T lymphocyte (CTL) 364

D Dacryocystitis 126 Dacryolithiasis 126 Dacryostenosis 126 DAG see Diacylglyerol DC see Dendritic cell DC-LAMP/CD208 373 DC-SIGN/CD209 373 De Vasis Palpebrarum Novis Epistola (Meibom) 60 Defensins 143, 356 Delayed-type hypersensitivity 361, 377, 390 Dendritic cell (DC) 373, 431 intraocular 378 Dermatochalasis 91, 92, 93f Descemet’s membrane 134, 134f, 148, 290 banded/nonbanded regions of 278–279, 278f copper deposition in 138 Descemet’s stripping automated endothelial keratoplasty (DSAEK) 133, 213, 218 endothelial disease and 141, 141f Desiccating stress (DS) 381, 387 Desmosomes 167, 167t, 452 Detection acuity 494, 495f Deturgescence 171, 275–276, 283 corneal endothelium’s role in 273f, 279, 279f biochemistry of active 279, 280f physiological control of active 280 Dexamethasone (DX) 310–312, 311f, 312t Dextroversion 46 DHEA see Dihydroepiandrosterone DHT see Dihydrotestosterone Diabetic retinopathy 164 Diacylglyerol (DAG) 112 Diffuse lamellar keratitis 201 dIgA see Dimeric IgA DiGeorge syndrome 12 Dihydroepiandrosterone (DHEA) 75 Dihydrotestosterone (DHT) hypophysectomy and 75 lacrimal gland and 75 Dimeric IgA (dIgA) 74–75, 80, 343 2,4-dinitro-1-fluorobenzene (DNFB) 28–29 Diplopia 39, 45 in CPEO compared to myasthenia gravis 41 Direct allorecognition 390 Corneal graft rejection afferent immune response and 392 Distichiasis 95 DNFB see 2,4-dinitro-1-fluorobenzene Dominant-negative mutant construct 327 Drug delivery see Topical drug delivery Dry eye 7, 108, 133, 381 anti-androgen therapy patients and 382

Subject Index

chronic pain with 382 LKC 382 in ocular surface 382 clinical features of 382 epidemiology of 381 gender and 74 goblet cells and types of LASIK 114 OCP 114 Sjo¨gren syndrome 114 topical preservatives 114 vitamin A deficiency 114 mucins and 125 in peri-/post-menopausal women 382 quality of life impact of 382–383 therapies for 388 artificial tears 388 corticosteroids 388 cyclosporine 388 mucin secretagogues 388 tetracyclines 389 DS see Desiccating stress DSAEK see Descemet’s stripping automated endothelial keratoplasty DTS see Dysfunctional tear syndrome Duchenne muscular dystrophy 39 Ductal cells, in lacrimal gland 68, 83, 84f electrolyte/water secretion mechanisms by 73 DX see Dexamethasone Dynamic visual acuity 498 Dysesthesias 158 Dysfunctional tear syndrome (DTS) 381 Dystrophic epidermolysis bullosa 164

E E2F 290 CEC proliferation promoted by 294 EALT see Eye-associated lymphoid tissue EAU see Experimental autoimmune uveitis EBM see Epithelial basement membrane EBMD see Epithelial basement membrane dystrophy ECM see Extracellular matrix Ectomesenchyme 219 Ectopic activity 158 Ectropion 91 cicatricial 95 eyelids and 95, 95f paralytic 95 EDE see Evaporative dry eye EDL see Extensor digitorum longus EDTA see Ethylenediaminetetraacetate Effector cells, innate immune system regulation of 359 Efferent immune response 390 in autoimmune-based inflammation 386f, 387 corneal graft rejection, mechanisms/components of 393 immune privilege breakdown 394 macrophages 394 NK cells 394 T-lymphocyte activation 393 immunoregulation and 385 Efferent tear ducts see also Nasolacrimal ducts cavernous body of 129–130, 129f development of 126 history of 126 mechanisms of 126, 127t EGF see Epidermal growth factor EGFP see Enhanced green fluorescent protein EGFR see Epidermal growth factor receptor Eicosanoids 406 Electrochemical equilibrium 171 Electron transport chain 27 EOM mitochondria content and 30–31 Emmetropization 506, 507, 518, 521–522, 529 Endocytosis 68, 354 Endophthalmitis 354 Endostatin 475, 480 Endothelial disease 138 Fuch’s dystrophy 138 pseudophakic bullous keratopathy 138

surgical intervention for 141 DSAEK 141, 141f Endothelial polymegathism 207 Endothelium of cornea see Corneal endothelium Enhanced green fluorescent protein (EGFP) 327 see also Green fluorescent protein Entropion 91 cicatrical 94 of eyelids 94, 94f involutional 94 EOM see Extraocular muscles Eosinophils 422 Epidermal growth factor (EGF) 290, 315, 452 ERK and 111–112 eyelid morphogenesis and 324, 325f mucin secretion and 113 receptors of 111 Epidermal growth factor receptor (EGFR) 324, 325f epi-LASIK see Epithelial laser in situ keratomileusis Epiphora 126 Epithelial basement membrane (EBM) 133, 226 dystrophy of 136, 226 reformation of 166 Epithelial basement membrane dystrophy (EBMD) 136, 226 Epithelial cell migration 165–166 Epithelial disease 135 band keratopathy and 136 chemical/thermal burns and 136 iron deposition in 135, 136t medication toxicity for 136 staining patterns for 135, 136f, 136t surgical intervention for 139 amniotic membrane graft 139 conjunctival flap 139 corneal glue 139 EDTA 139, 139f limbal stem cell transplant 139 pterygium excision 139 PTK 139 stromal puncture 140 tarsorrhaphy 140 Epithelial laser in situ keratomileusis (epi-LASIK) 204, 256 Epithelial renewal 175 ERK see Extracellular signal-regulated kinase Esotropia 39, 45 Esterases 303 Estradiol 77 pregnancy and 78 Estropia 529, 533 Ethylenediaminetetraacetate (EDTA) 133, 291–292 epithelial disease surgical intervention and 139, 139f Evaporation meibomian glands evaluation with 65 tear breakup from 124 from tear film 53, 53f Evaporative dry eye (EDE) 53 Excimer laser 193 Excyclotorsion 17 Exocytosis 68, 83 constitutive 70–71 process of 71 Exotropia 39, 45 Experimental autoimmune uveitis (EAU) 373 APCs in 379 Extensor digitorum longus (EDL) 27, 28f Extracellular matrix (ECM) 256, 470 CEC interactions with 293 corneal epithelial wound healing and communication with 264 bone-marrow-derived cells in 265 cellular responses in 265 suppressing/amplifying factors of 264 corneal stroma 223 fibrotic scarring and 257 Extracellular signal-regulated kinase (ERK) 84, 86f EGF and 111–112 Extraocular muscles (EOM) 17 abnormal eye movements and 39, 40f, 41f brainstem disorders and horizontal movements 40f, 42 brainstem disorders and vertical movements 41f, 43 CNVI, CNIII, CNIV disorders effecting 42 congenital misalignment 43

541

542

Subject Index

Extraocular muscles (EOM) (continued) disorders effecting 39 NMJ disorders and 40, 40f visual system disorders effecting 44 ATP buffering system and 28–29 in bony orbit, anatomy of 17 clinical assessment of 46 eye movement function of 45 fibers of, classification for 27, 28f Golgi tendon organs in 36 occurrence, distribution, and number of 34t, 35 structure of 36, 36f heterogeneity of 12 histological anatomy of 21 innervation 21 organization in 21, 21f, 22f metabolism of CK in 28 gene expression profiling and 27 glycogen in 27–28 LDH in 27–28 mitochondrial content in 29 ATP generation of 31–32 as calcium sinks 30 capacity matched to contractile function of 31 differences of 30 electron transport chain and 30–31 energy supply matched to demand in 32 morphogenesis development of 13f, 14–15, 15f gene expression patterns in 15–16, 16f muscle spindles in 34, 35 occurrence, distribution, and number of 34, 34t structure of 34, 35f MyHC isoforms in 22, 23f, 24f nonuniform expression along muscle length of 23, 24f, 25f myofibers in 21, 22f heterogeneity of 25 types of 21–22 origins of 9, 10f, 11f prechordal mesoderm and 10, 11f palisade endings in 37 molecular characteristics of 37, 37f occurrence, distribution, and number of 34t, 36 structure of 37, 37f proprioception and 33 arguments against 33–34 remodeling in 25 satellite cells within 25–26, 26f skeletal muscle fiber types in 22 Eye mobility 33 Eye movements see also Abnormal eye movements; Blinking; Pursuit movements; Saccades blink-evoked 5–6 EOM functions for 45 functional classes of 39, 42t normal 45, 46f Eye muscle progenitors determination of 11 LR formation and 12 myf5/myoD genes in 12, 13f pax3 gene 12 Eye-associated lymphoid tissue (EALT) 340, 431 CALT as part of 438, 440f deregulation of 440, 441f Eyelashes, in passive innate immune system 354 Eyelids benign tumors of 97 blepharitis in 97–98 chalazion in 97, 97f dermatochalasis of 92, 93f distichiasis and 95 ectropion and 95, 95f entropion of 94, 94f evolutionary origin of 3 floppy lid syndrome and 95, 96f forces acting on 3–4, 4f hordeolum in 97 inflammatory an infectious disorders of 97 lower, anatomy of 91, 92f malignant tumors of

BCC in 95, 96f malignant melanoma 97, 97f SCC 96, 96f sebaceous cell carcinoma 96–97, 97f margin of, anatomy of 91–92 meibomian glands in 60, 61f morphogenesis of, EGFR/EGF TGFa on 324, 325f organization of 3 in passive innate immune system 354 ptosis of 93, 93f retraction of 93, 94f transillumination of, meibomian glands and 64–65, 65f trichiasis and 95 upper, anatomy of 91, 92f

F F4/80 373 Fc Rn 345 Femtosecond laser in situ keratomileusis (FS-LASIK) 204 FGF see Fibroblast growth factor FGF7, in corneal homeostasis maintenance 323, 324f Fibroblast growth factor (FGF) 315 basic 467 FGF-2 signaling pathway 294 FGF7 in corneal homeostasis maintenance 323, 324f Fibrosis 256 Fibrotic scarring 257 AM for 269 ECM and 257 gene therapy for 333 injuries causing 257 medical problems caused by 259–260 modulation of 268 PRK causing 333 scar-reducing therapies for currently available 268, 268t emerging 269 TGF-b mediating 257–258, 258–259, 258t Filarial disease 401 Filopodia 164 Fleck dystrophy 231 Fleming, Alexander 444–445 Floppy lid syndrome 95, 96f Fluorescein sodium 51 tear film stability measured with 55–56, 56f, 57f FMNS see Fusional maldevelopment nystagmus syndrome Follicles 213 Forced-choice paradigms 500 for contrast-detection threshold 500, 501f Form fruste keratoconus 201 Fourier analysis 500 FS-LASIK see Femtosecond laser in situ keratomileusis Fuch’s dystrophy 138, 213, 214–215, 231, 231f corneal endothelium and 281 Fungal corneal ulcers 137 Fungal keratitis contact-lens-associated 426 biofilm formation in 426, 427f innate immune system and 427, 428f, 429f MMPs in 427 trauma associated with 427 Fusarium keratitis 426 see also Fungal keratitis Fusional maldevelopment nystagmus syndrome (FMNS) 43–44

G G1/S transition, of CEC 290–291, 292f Gabors 500, 501–502, 502f GAGs see Glycosaminoglycans GALT see Gut-associated lymphoid tissue Ganciclovir (GCV) 306–308, 308t Gap junctions 166, 167t Gaze conjugate 46, 46f disconjugate 46 primary 46, 46f GCV see Ganciclovir

Subject Index

Gelatinous drop-like dystrophy 230 Gene expression profiling 27 Gene therapy 327, 328t see also Corneal gene therapy conjunctiva 335 corneal dystrophy treatment with 332, 334 alkali burn 333 CNV 334 corneal epithelial wound healing 332 corneal graft rejection 332 fibrotic scarring 333 lacrimal gland 334 GFP see Green fluorescent protein GFs see Growth factors Glaucoma 517 Glycan 116, 219 Glycocalyx 116, 143 Glycoforms 116 Glycogen in EOM metabolism 27–28 glycolysis and breakdown of 28 Glycolysis 27 anaerobic breakdown of 29f glycogen breakdown in 28 Glycosaminoglycans (GAGs) 219, 272, 279 biosynthesis of 224–225 in PGs 224 swelling pressure and 283 Glycosylation 108 mucins and 120 Glycosyltransferases 120 Goblet cells 108, 381, 432 in conjunctiva 100, 108, 109, 109f, 111f mucin synthesis/secretion by 110, 111f, 112, 113f proliferation of 110, 111, 111f development of 108, 109f, 110f dry eye types and LASIK 114 OCP 114 Sjo¨gren syndrome 114 topical preservatives 114 vitamin A deficiency 114 VIP receptors in 113 Golgi apparatus 120 Golgi tendon organs 33 collagen bundles in 36, 36f in EOM 36 occurrence, distribution, and number of 34t, 35 structure of 36, 36f Granular dystrophies 229 type 1 229, 229f type 2 229 Green fluorescent protein (GFP) 318, 373 see also Enhanced green fluorescent protein Growth factors (GFs) 470 See also specific growth factors corneal epithelial wound healing and 261, 262f HCEC and age-related decrease in sensitivity to 293 in ocular surface tissue morphogenesis in transgenic mice 321 regeneration regulated by 257–258 wnt family of 12 Gut-associated lymphoid tissue (GALT) 340 Guttae 213 Guttata 283

H HA see Hyaluronan Haze 327, 333 hBDs see Human b-defensins HCEC see Human corneal endothelial cells HCO3-ATPase see Bicarbonate-stimulated adenosine triphosphatase Head myogenesis 9 regulatory networks of 14t paraxial mesoderm 10, 11f Helix pomatia agglutinin (HPA) 108 Helper T cell 459 HEMA see Hydroxethylmethacrylate Hemangiogenesis 478

Hemidesmosomes 452 adhesion complexes and 145, 145f cell-matrix junctions and 167 Hemifacial spasm (HFS) 3 course of 7–8 pulsatile arterial compression and 8 Hepatocyte growth factor (HGF) 12 Herpes simplex virus 137, 453 Herpes stromal keratitis (HSK) APCs in 398 causes of 397 chemokines in 398 cytokines in 397 epidemiology and pathogenesis of human 399 history of 396 HSV-1 angiogenesis in 399 management of 400 pathogenesis models for 399 T cells in 397 Herpetic stromal keratitis (HSK) 377 HEVs see High endothelial venules HFS see Hemifacial spasm HGF see Hepatocyte growth factor High endothelial venules (HEVs) 431, 433f, 434f, 437 Higher-order aberration (HOA) 205 wave front technology for 205 High-order wave front aberrations 506 Histocompatibility antigens 392 HLA see Human leukocyte antigen HLA-DR 419 HMG-CoA 60 HOA see Higher-order aberration Holorine 60 Homeobox protein 272, 274–275 Homeostasis corneal FGF7 and maintenance of 323, 324f LSC in 182 immune 367 LFU’s role in 383 ocular surface repair and 179 Hordeolum 97 Horizontal gaze palsy 40f, 42–43 Horner’s muscle 126, 128–129 HPA see Helix pomatia agglutinin HSK see Herpes stromal keratitis; Herpetic stromal keratitis HSV-1 corneal epithelium, eradication of 396 HSK angiogenesis in 399 management of 400 latency 396 in trigeminal ganglia 397 Huang, David 237–238 Human b-defensins (hBDs) 454 Human corneal endothelial cells (HCEC) 291, 292f age-related decrease in sensitivity to mitogen/GF in 293 proliferative capacity from central v. peripheral regions of 294 Human leukocyte antigen (HLA) 431 Hurler’s syndrome 219 Hyaluronan (HA) 225 Hydrophilic 60 Hydrophobic 60 Hydroxethylmethacrylate (HEMA) 207 Hyperacuity 497, 498f Hypercapnia 470 Hyperemia 419 Hypermetropia 207, 507 Hyperopia 133, 201, 327 accommodation and 530 classifications of 529 clinical presentations of 531 early detection of 531 examination techniques of 531 LASIK/LASEK treatment of 204 management of 532 prevalence of 530, 530f time course of 530, 531f uncorrected 531 Hyperosmolarity, of tear film 57–58

543

544

Subject Index

Hypertrophic scar 256 Hypophysectomy 74 DHT and 75 Hypoxia 470

I ICAM-1 see Intercellular adhesion molecule 1 ICR see Intracorneal ring Idiopathic pulmonary fibrosis 164 IFN- see Interferon gamma IgA see Immunoglobulin A IgE see Immunoglobulin E IgG1 345 IL-1a, IL-b see Interleukin 1 alpha and 1 beta IL-6 see Interleukin 6 IL-12 see Interleukin 12 IL-18 see Interleukin 18 Image contrast 500, 501f Immune homeostasis 367 Immune privilege 367, 390 see also Anterior chamber-associated immune deviation corneal 390 factors contributing to 390 site 391 corneal angiogenic privilege and 467 corneal endothelium, cytokines and 277, 278f corneal graft rejection efferent immune response and breakdown of 394 factors breaking 353 history of 361 immunosuppression/immunoregulation factors of 368 innate 359 lessons of 371 ocular-induced regulatory cells of 364, 365t suppressive factors maintaining 352–353, 352t VCAID and 364 Immune surveillance 396 conjunctival epithelium and 432 epithelial morphology and function in 432, 432f, 433f, 434f of conjunctival lamina propria 434f, 435 of cornea, CALT assisting 437, 439f Immunoglobulin A (IgA) 434f, 436 see also Secretory immunoglobulin A Immunoglobulin E (IgE) 424 Immunoglobulins IgG1 345 transfer to milieu exte´rior dIgA 343 pIgR 343 Immunoreceptor tyrosine-based activation motifs (ITAMs) 348–349 Immunoregulation afferent immune response of 385 in CALT 438, 441f of conjunctiva 438 efferent immune response of 385 immune privilege and 368 Immunosuppression in aqueous humor 367, 368t, 369f in aqueous humor, neuropeptides and 370 a-MSH and 369 immune privilege and 368 TGF-b and 368 Incyclotorsion 17 Indirect allorecognition 390 Corneal graft rejection afferent immune response and 392 Infantile nystagmus syndrome (INS) 44 Infantile strabismus 43 Inferior oblique (IO) 17, 19, 20f anatomy of 20–21 Inferior rectus muscle, CNIII innervating 18–19, 19f Inflammation 347 afferent immune response in autoimmunity and 386, 386f corneal APCs in 376, 376f, 376t corneal sensory receptors and local 160f, 161 efferent immune response in autoimmunity and 386f, 387 of ocular surface from contact lenses 210 signs/symptoms of 459 tear fluid and 460 as diagnostic indicator 462 tear fluid connecting cornea/conjunctiva and 463

Inhibitory PAS protein (IPAS) 467 Innate immune privilege 359 Innate immune system 347, 354, 367 Acanthamoeba keratitis and 414, 415, 415t active 357 complement system in 358 PRRs in 357, 358t adaptive immune system’s link to 359 of conjunctiva epithelium, defense factors of 433 cytokines/chemokines/effector cells regulating 359 fungal keratitis and 427, 428f, 429f passive 354 anatomic/physical barriers of 354, 355f chemical barriers of 356 complement system in 357 corneal epithelium 355 eyelids and eyelashes 354 posterior lens capsule 355 RPE 355 tear film in 354, 356f tiers of 354 INO see Internuclear ophthalmoplegia INS see Infantile nystagmus syndrome Integrins 143 Intercellular adhesion molecule 1 (ICAM-1) 431 Interferometer, thin film 57 Interferon gamma (IFN- ) 74, 413 in Pseudomonas aeruginosa keratitis IL-18 and 408 SP in 408 Interferon regulatory factor 1 (IRF1) 406 Interleukin 1 alpha and 1 beta (IL-1a, IL-b) 74, 480 Interleukin 6 (IL-6) 74 Interleukin 12 (IL-12) 74 in Pseudomonas aeruginosa keratitis 408 Interleukin 18 (IL-18) 408 Interlobular duct epithelial cells 80, 81f Internuclear ophthalmoplegia (INO) 42–43 Interpenetrating polymeric networks (IPNs) 296 Intracorneal ring (ICR) 201 for myopia 206, 206f Intraepithelial nerve terminals 153, 154f Intraocular pressure (IOP) 310 Intrinsically photosensitive retinal ganglion cells (IpRGCs) 487 in PLR afferent pathway 490, 490f Intussusception 465 Involutional entropion 94 IO see Inferior oblique Ion channels 158 sensory receptors and 159f, 161 Ion transport, in corneal epithelium 171, 172f importance of 172 primary/secondary 173 receptor-mediated control of 175 stromal deturgescence coupled with 173 IOP see Intraocular pressure IPAS see Inhibitory PAS protein IPNs see Interpenetrating polymeric networks IpRGCs see Intrinsically photosensitive retinal ganglion cells Ipsilateral facial nucleus 3–4, 4f IRF1 see Interferon regulatory factor 1 Iridectomy 150 Iris musculature 489, 489f ITAMs see Immunoreceptor tyrosine-based activation motifs

J Jones tube 128–129

K K+ channels 176 Kearns-Sayre syndrome 39 Keloid 256 KEP see Keratoepithelioplasty Kepler, Johannes 518–519 Kera see Mouse keratocan gene Keratan sulfate 219, 220 lumican modified by 224–225

Subject Index

Keratan sulfate proteoglycans (KSPGs) 317 Keratic precipitates 213 Keratins 315 Keratitis 327, 406 see also Fungal keratitis Acanthamoeba 413 adaptive immune system’s resistance to 415 adaptive immune system’s resistance to, evading 416, 417t barriers to 414 innate immune system’s resistance to 414, 415, 415t macrophages and 414 mucosal immune system and 416, 416t neutrophils and 414 trophozoites for 417 Aspergillus 428, 429f Candida 426 Fusarium 426 Pseudomonas aeruginosa 406 animal models for 411, 411f apoptosis 410 neuropeptides and 408, 410 PMN, cytokines, chemokines in 407 SP and 408 T cells and IL-12 in 408 TLRs 409 VIP and 409 Keratocan 315 mouse 318 Keratoconus 137, 138f diagnosing 238–239 Keratocytes 178, 219, 272 corneal epithelial wound healing and changes of 265–266 of corneal stroma 221f, 222 for transgenic and knock-out mice 317 Keratoepithelioplasty (KEP) 185 Keratolimbal allograft (KLAL) 185 corneal epithelial transplantation by indications for 186, 186f postoperative management for 187, 187t surgical procedure for 186 history of 186 Keratoplasty 465 Keratoprosthesis (KPro) 296, 298, 299f KLAL see Keratolimbal allograft Knock-out mice 272 see also Cre-LoxP system gene targeting for 315 models for 315–326 ocular-surface tissue-specific promoter identification in 317 strategies for 317 corneal-epithelium-specific promoter 318 keratocyte-specific promoter 317 ocular-surface tissue-specific drive lines in 318 Pax6 gene for 318 KPro see Keratoprosthesis KSPGs see Keratan sulfate proteoglycans

L Lacrimal fluid contents of 70 regulation layers of 75 Lacrimal function unit (LFU) 381 anatomy/function of 383, 383f disease and role of 383 homeostasis and role of 383 Lacrimal gland 51, 381 acinar cells in 68, 69f, 83, 84f electrolyte/water secretion mechanisms by 72, 73f anatomy of 68, 69f, 83, 84f blood supply to 70 bone-marrow-derived cells in 69 counterpoises between contradictory signals in 80 DHT and 75 differentiation and survival factors of 80, 81f drainage of 58, 58f ductal cells in 68, 83, 84f electrolyte/water secretion mechanisms by 73 estradiol and 77 gene therapy 334

545

hormone regulation of 74–82 innervation of 69 myoepithelial cells in 69, 83, 84f parasympathetic nerves in 83–84 prolactin and 77 protein secretion in 70, 72f MAPK activation in 86f, 87 reproductive hormones influencing exocrine functions of 80–81, 82f secretion neural control of 83, 84f neural reflex arc and stimulation of 84, 85f sIgA from 74–75 signal transduction activation in 84, 85f cholinergic agonist-activated 84, 86f sympathetic nerves in 83–84 Lacrimal keratocunjunctivitis (LKC) 384 dry eye chronic pain from 382 quality of life impact of 382–383 Lacrimal passages 126, 129f see also Nasolacrimal ducts Lacrimal sac 51 immune mechanisms of adaptive 130 innate 130, 130t tear transport through 129, 129f Lactate dehydrogenase (LDH) 27–28 Lactation 74 sIgA secretion during 79–80 Lactoferrin 357 Lactogenesis 74, 79–80 Lagophthalmos 91 Lambert-Eaton myasthenic syndrome (LEMS) 41–42 Lamellae 219, 220 Lamellar keratoplasty (LK) 133, 185, 226 anterior 140 for peripheral corneal ulcers 191 Mooren’s ulcer 191, 192f RA and 191 Langerhans cells (LCs) 373, 450, 452 as corneal APCs 374, 375t LASEK see Laser-assisted subepithelial keratomileusis Laser thermokeratoplasty (LTK) 133, 140–141 Laser-assisted in situ keratomileusis (LASIK) 114, 133, 140–141, 150, 193, 201, 213, 296 corneal sensitivity after 162 CSF/MTF pre-/post- 197f history of 194 history/evolution of 202, 203f hyperopia treatment with 204 LASEK compared to 203–204 myopic regression after 203 optical aberrations/visual quality after 195, 196f process of 144, 195, 195f safety/efficacy/satisfaction with 198 side effects of 198, 205 spherical aberration following 197, 198f studies on 198–199 wave front technology and 205 Laser-assisted subepithelial keratomileusis (LASEK) 133, 140–141, 201, 203, 256, 296 aims of 203 history of 194 hyperopia treatment with 204 LASIK compared to 203–204 PRK compared to 203–204 side effects of 205 wave front technology and 205 LASIK see Laser-assisted in situ keratomileusis Lateral rectus muscle (LR) 19, 20f abduction with 45 CNVI innervating 18–19, 19f eye muscle progenitors and formation of 12 Lattice dystrophy 213, 214, 228, 228f Lbx1transcription factor 12 LCs see Langerhans cells LDH see Lactate dehydrogenase Left abducens palsy 40f, 42 LEMS see Lambert-Eaton myasthenic syndrome Lens-induced chronic hypoxia 207, 209 Lenticle 213, 296 Lentivirus vectors 330–331, 331f

546

Subject Index

Leone, Giovanni Battista Carcano 126 Levator palpebrae superioris muscle (LPS) 3, 17 action of 3–4, 4f OO interacting with 4, 4f passive downward forces interacting with 4–5 LFU see Lacrimal function unit Lids see Eyelids Limbal barrier function 468 Limbal epithelium, contact lens effects on 209 Limbal plexus 150 Limbal stem cells (LSC) characteristics of 180, 181t in corneal homeostasis 182 deficiency of 181, 470 niche of 179f, 181 transplant of 139 Limbal transplantation 185 Limbus 478 vasculature of 478, 479f Lipid layer of tear film 52, 61 contact lenses and 66 meibomian glands and 52, 53t function of 60–61, 62f models for 61, 62f proteins in 61–62, 62f Lipids mediators 475 meibomian glands producing amount of 64 composition of 63 turnover and synthesis of 63 Lipocalin 61 lipocalin-A 357 Lipophilic ester prodrug design 305–306, 307f, 308t Lipopolysaccharide (LPS) 406, 431 LK see Lamellar keratoplasty LKC see Lacrimal keratocunjunctivitis LL-37 see Cathelicidin Local osmosis 173–174 Long-term depression (LTD) 6–7 Long-term potentiation (LTP) 6–7 LoxP 315 see also Cre-LoxP system LPS see Levator palpebrae superioris muscle; Lipopolysaccharide LR see Lateral rectus muscle LSC see Limbal stem cells LTD see Long-term depression LTK see Laser thermokeratoplasty LTP see Long-term potentiation Lubrication of mucins 124 tear film and 51 Lumican 219 keratan sulfate modifying 224–225 Lymphangiogenesis 465, 478 Lymphocytes see T-lymphocytes Lysozyme 356

M M line 27 Macrophages 266–267, 376 Acanthamoeba keratitis and 414 conjunctival lamina propria and 434f, 436 in corneal graft rejection efferent immune response 394 role of 373 Macular corneal dystrophy 219, 230 Magnetic resonance imaging (MRI) 7 Major histocompatibility complex (MHC) 347, 370–371, 413, 431 CTL and 364 Major histocompatibility complex class II molecules (MHC class II) 339, 340, 373, 379, 381 Malignant melanoma 97, 97f MALT see Mucosa-associated lymphoid tissue; Mucosal-associated lymphoid tissue Manifest 201 Mannose-binding protein (MBP) 413 Mannose-induced protease 133 (MIP-133) 413 MAPK see Mitogen-activated protein kinase MAR see Minimum angle of resolution Marker gene 327

Mast cells (MCs) 421, 421t ocular allergies and 421, 421f Matrix metalloproteinase-9 (MMP-9) 406 Matrix metalloproteinases (MMPs) 426, 459, 470, 475 corneal epithelial wound healing and 169 in fungal keratitis 427 MMP-9 406 Maurice, David 214, 223 pump-leak hypothesis and 283, 284f MBP see Mannose-binding protein M-cells see Membranous cells MCs see Mast cells Mechanonociceptors 159, 160f Medawar, Peter 352, 361–362 Medial longitudinal fasciculus (MLF) 40f, 42 bilateral 41f, 43 rostral interstitial nucleus of 41f, 43 Medial rectus muscle (MR) 17, 19, 20f adduction with 45 CNIII innervating 18–19, 19f Medial vestibular nucleus (MVN) 40f, 43 Meek, Keith 223 Meesman’s lesions 227 Meibom, Heinrich 60 Meibomian gland dysfunction (MGD) 53 chronic blepharitis from 65 Meibomian glands 51, 60–61, 116, 381 acinar cells in 62–63 anatomy and histology of 62, 63f development of 63 disorders/pathology of 64 chalazion 66, 66f chronic blepharitis 65 trachoma 66 evaporation for evaluating 65 in eyelids 60, 61f lipids produced by amount of 64 composition of 63 turnover and synthesis of 63 as modified sebaceous gland 63 orifices of 61f squeezing of 64–65, 65f tear film lipid layer and 52, 53t function of 60–61, 62f transillumination of eyelids showing 64–65, 65f Melanocytes 452 Membranous cells (M-cells) 431 Membranous lacrimal passage 127, 128f Menopausal women, dry in peri-/post- 382 Meridional amblyopia 506 Mesenchyme 9, 219 MGD see Meibomian gland dysfunction MHC see Major histocompatibility complex MHC class II see Major histocompatibility complex class II molecules Mice see Knock-out mice; Mouse lines Microfilariae 401 Milieu exte´rior, immunoglobulin transfer to 343 dIgA 343 pIgR 343 Millinewtons per meter (mN m 1) 60 Minimum angle of resolution (MAR) 494, 495f Minute of arc 494 Miosis 487 MIP-133 see Mannose-induced protease 133 Mitochondria, EOM content of 29 ATP generation of 31–32 as calcium sinks 30 capacity matched to contractile function of 31 differences of 30 electron transport chain and 30–31 energy supply matched to demand in 32 Mitogen 290 HCEC and age-related decrease in sensitivity to 293 Mitogen-activated protein kinase (MAPK) 84, 406 adrenergic agonists signal transduction and 88f, 89 p38MAPK 323 protein secretion in lacrimal gland and activation of 86f, 87 signal transduction and 86f, 87 VIP 88 Mitomycin C (MMC) 256, 268–269 MLC see Myosin light chain

Subject Index

MLF see Medial longitudinal fasciculus MMC see Mitomycin C MMP-9 see Matrix metalloproteinase-9 MMPs see Matrix metalloproteinases mN m 1 see Millinewtons per meter Modulation-transfer function (MTF) 193 computing/measuring 196–197 pre-/post LASIK 197f Mooren’s ulcer 185 LK for peripheral corneal ulcers and 191, 192f Morphogenesis 9 EOM development of 13f, 14–15, 15f gene expression patterns in 15–16, 16f muscle 14 trapezius muscle 14 Motility assessment see Ocular motility assessment Motor proteins 68, 71 Mouse keratocan gene (Kera) 318 Mouse lines see also Knock-out mice corneal epithelium-specific 319, 319f, 320f, 321f, 322f corneal stroma-specific 317t, 318 Pax6OS-rtTA 321 MPSs see Multipurpose solutions MR see Medial rectus muscle MRI see Magnetic resonance imaging MS see Multiple sclerosis MTF see Modulation-transfer function Mucin deficiency, on ocular surface 113 allergies and 114 OCP 114 Sjo¨gren syndrome 114 vitamin A deficiency 114 Mucin layer of tear film 55 Mucin secretagogues 388 Mucins 108, 116 antimicrobial activity of 124 architecture of 116, 117f, 118f, 119f biosynthesis of 119 classifications of 118 conjunctival goblet cells synthesizing/secreting, control of 110, 112, 113f contact lens wear and 124 dry eye and 125 EGF and secretion of 113 function of 121 gel formation of 122, 122f, 123f glycosylation and 120 immune protection of 124 individual variation of 121, 121f, 122f lubrication of 124 mucus gel and 116 physical/chemical barriers of 123 secreted 118, 120f control of 121 surface-associated 118, 119f synthetic pathways of 119 turnover of 119, 120 degradation for 121 recycling for 120 Mucopolysaccharidosis 334 Mucosa-associated lymphoid tissue (MALT) 340, 438 of nasolacrimal ducts 130–131 Mucosal immune system 413 Acanthamoeba keratitis and 416, 416t autoantigen secretion to corneal stroma 344 effector sites of 341 niches of 341, 342f Fc Rn and 345 IgG1 and 345 immunoglobulin transfer to milieu exte´rior in 343 inductive site organization in 340, 340f infection tolerance/response of 342f, 345 physiology of 339 reproductive hormones influencing tissues of 79 Mucosal-associated lymphoid tissue (MALT) 69 Mucus gel 116 Mu¨ller’s muscle 3 action of 3–4, 4f anatomy of 91, 92f Multiple sclerosis (MS) 42–43

Multipurpose solutions (MPSs) 207, 210–211, 426 contact lens care products and 211 Munnerlyn formula 193 Muscarinic receptors 83 Muscle spindles 33 in EOM 34, 35 occurrence, distribution, and number of 34, 34t structure of 34, 35f MVN see Medial vestibular nucleus Myasthenia gravis 39 diplopia in CPEO compared to 41 in NMJ 40–41 Mydriasis 487 Myf5 genes 12, 13f MyHC isoforms see Myosin heavy chain isoforms Myoblasts 9 myf5/myoD genes 12, 13f MyoD genes 12, 13f Myoepithelial cells, in lacrimal gland 69, 83, 84f Myofibers, in EOM 21, 22f heterogeneity of 25 types of 21–22 Myofibroblast 219, 256 reduced transparency of 267 in regeneration 258–259, 260f TGF-b and 222–223, 222t Myogenesis head 9 regulatory networks of 14t of paraxial mesoderm, neural crest and 11f, 15 of skeletal muscles 9 trunk, regulatory networks of 14t Myopia 201, 207, 327, 518f classifications of 519 clinical management of 518f, 522 contact lenses for 523 correction and prevention in 522 pharmacological treatment in 524 positive additions for nearwork in 523 refractive surgery options for 523 vision therapy and biofeedback training in 523 epidemiology of 520 etiology of 520, 521f health and economic impacts of 517 ICR for 206, 206f natural history of 518 structural correlates, molecular/anatomical changes in 518 Myopic regression, LASIK and 203 Myosin heavy chain isoforms (MyHC isoforms) 17 in EOM 22, 23f, 24f nonuniform expression along muscle length of 23, 24f, 25f Myosin light chain (MLC), phosphorylation of 287, 287f corneal endothelium barrier integrity and effect of 287 Myotendinous cylinders see Palisade endings Myotome 9

N Na, K-ATPase see Potassium-stimulated adenosine triphosphatase; Sodium-potassium-dependent ATPase Na+ channels 176 NAD+ see Nicotinamide adenine dinucleotide NADH 27, 28 Nasolacrimal ducts 126 anatomy/dimensions of 127, 128f comparative 127 immune mechanisms of adaptive 130 innate 130, 130t MALT of 130–131 sIgA in 130 tear absorption in 131, 131f tear transport through 129, 129f Natural killer cells (NK) 126, 130 in corneal graft rejection efferent immune response 394 regulation of 370–371 NBCel see Sodium bicarbonate cotransporter Neostatin 475 Neovascularization see Corneal neovascularization Nerve impulse 158 Net flux 171

547

548

Subject Index

Neural crest 9, 11, 11f, 178, 219, 220–222 action of 11 development of 274f, 275f myogenic paraxial mesoderm and 11f, 15 Neural reflex arc 84, 85f Neurite 150 Neuromuscular contacts 37 Neuromuscular junction (NMJ) 21 abnormal eye movement and disorders of 40, 40f en plaque/en grappe 21–22 myasthenia gravis at 40–41 Neuropathic pain 158 Neuropeptides 367, 368 corneal epithelium infection response of 457 Pseudomonas aeruginosa keratitis and 408, 410 apoptosis 410 SP and 408 TLRs 409 VIP and 409 Neurotendinous contacts 37 Neurotransmitters 83–84 Neurotrophic epitheliopathy 150 Neurotrophic keratitis 150 Neutrophils 266–267 Acanthamoeba keratitis and 414 conjunctival lamina propria and protective functions of 437 ocular allergies and 422 NF-kB see Nuclear factor-kappa B Niches 143, 178 LSC 179f, 181 of mucosal immune system 341, 342f Nicotinamide adenine dinucleotide (NAD+) 27 see also NADH Nitric oxide (NO) 88f, 89 NK see Natural killer cells NLRs see NOD-like receptors NMJ see Neuromuscular junction NMR see Nuclear magnetic resonance NO see Nitric oxide NO synthase (NOS) 89 NOD-like receptors (NLRs) 357, 358t Nonviral vectors 331 NOS see NO synthase NPH see Nucleus prepositus hypoglossi Nuclear factor-kappa B (NF-kB) 406, 452 Nuclear magnetic resonance (NMR) 75 Nucleotide agonists 121 Nucleus prepositus hypoglossi (NPH) 40f, 43 Nystagmus 39 attendant 43 continuous jerk 44 function of 42t infantile 44

O Oblique astigmatism 511, 512f OCP see Ocular cicatrical pemphigoid OCT see Optical coherence tomography Ocular allergies AKC 420, 420t cellular mechanisms in 421, 421t conjunctival epithelial cells and 423, 424t conjunctival fibroblasts and 424 cytokines, chemokines in 425 eosinophils and neutrophils in 422 MCs and 421, 421f molecular mechanisms of 424 costimulatory 424 IgE 424 PAC 419, 420t SAC 419, 420t T-lymphocytes and 422, 422f, 423f types of 420t, 423 VKC 420, 420t Ocular astigmatism classification of 509 contact lenses for 515 image formation in 508f, 509 measurement of 513 corneal keratometric power map for 513, 514f

near work and 511 not correcting 516 origin of 507 prevalence and age-related changes in 507 spectacles for 515 surgery for 515 wave front technology for 205 Ocular cicatrical pemphigoid (OCP) 114, 185, 186 Ocular irritation 7 Ocular motility assessment 46, 47 Ocular surface 178 anatomy of 178, 179f APCs of 374, 374f carbohydrates at 124 contact lens wear’s effects on 209 alterations for 210 corneal epithelium/limbal epithelium 209 infection and 210 inflammation and 210 mechanical effects 211 damage to 185 defense mechanisms of 445f, 447 AMPs and 446f, 448 mechanical/physical 447 PRRs in 448 development of 179 dry eye chronic pain in 382 environmental impact on 384 homeostasis and repair of 179 mucin deficiency on 113 allergies and 114 OCP 114 Sjo¨gren syndrome 114 vitamin A deficiency 114 pathophysiology of 345 reconstruction of AM and 188 history/concept of 186 sIgA at 433f, 436 tear absorption and 59 tear film and absorption of 59 anchoring of 123, 123f transgenic mice, GFs of morphogenesis in tissues of 321 Ocular wound healing 260, 260f TGF-b and 260–261 Ocular-induced regulatory cells 364, 365t Oculomotor nerve (CNIII) 17 abnormal eye movements from disorders of 42 inferior rectus muscle innervated by 18–19, 19f MR innervated by 18–19, 19f SR innervated by 18–19 Omnipause neurons 3 reflex blinking blocked by 6 Onchocerciasis (river blindness) 401, 402f Wolbachia in pathogenesis of 402, 403f OO see Orbicularis oculi muscle Opsonization 347, 354 Optical coherence tomography (OCT) 217, 218f, 234, 237 see also Anterior segment OCT; Visante OCT Optical zone 193 Optokinetic reflexes 39, 42t Optotype 494, 495f Ora serrata 17 Oral mucosal epithelial cells 185 ex vivo expansion of 189 cell culture procedure for 189 clinical outcome for 190 concept of 189 indications for 190, 190f surgical procedure and postoperative management for 190 Orbicularis oculi muscle (OO) 3 action of 3–4, 4f LP interacting with 4, 4f Orchiectomy 74 Orthokeratology 207 contact lenses and 211 Orthotopic graft/transplant 327 Osmolarity 60 of tear film 57 Oxidative phosphorylation 27

Subject Index

P PAC see Perennial allergic conjunctivitis Pachymetry 213 Palisade endings 33 in EOM function of 37 molecular characteristics of 37, 37f occurrence, distribution, and number of 34t, 36 structure of 37, 37f Palisades of Vogt CESCs at 145–146, 147f at corneal periphery 146, 148f Palmer, Lisa 519 PAMPs see Pathogen-associated molecular patterns PAMR see Peri-junctional actomyosin ring Papillas 213 Paralytic ectropion 95 Paramedian pontine reticular formation (PPRF) 40f, 43 Parasitic corneal ulcers 137 Parasympathetic nerves 69–70 in lacrimal gland 83–84 Paraxial mesoderm 9 head 10, 11f neural crest and myogenic 11f, 15 skeletal muscles in 9 Paraxial optics 506 Paraxis 12 Paresis 3 Passive downward forces 3–4, 4f LP interacting with 4–5 Passive innate immune system see Innate immune system, passive Pathogen recognition receptors (PRRs) 347 in active innate immune system 357, 358t in adaptive immune system 347–348 conjunctival epithelium and 433, 433f ocular surface defense mechanisms and 448 Pathogen-associated molecular patterns (PAMPs) 347, 452 Pattern recognition receptor 444 Pax3 gene 12 Pax6 gene 318 Pax6OS-rtTA 321 PD see Potential difference PDGF see Platelet-derived growth factor PDS see Pigment dispersion syndrome PEDF see Pigment-epithelial-derived factor PEEs see Punctate epithelial erosions Penetrating keratoplasty (PK) 133, 185, 226, 296, 327, 470, 506 corneal nerve regeneration after 156, 157f indications for 390 stromal disease surgical intervention with 140, 140f survival with 390 Peptidases 303–304 Perennial allergic conjunctivitis (PAC) 114, 419, 420t see also Ocular allergies Perforating keratoplasty 478 Peri-junctional actomyosin ring (PAMR) 283, 286–287 Peripheral corneal ulcers, LK for 191 Mooren’s ulcer 191, 192f RA and 191 Peroxynitrite 406 Persistence length 116 PET see Positron emission tomography PGs see Proteoglycans Phacoemulsification 150 Phagocytosis 164, 347, 354 Phenotype 164 Phoria 45 Phospholipase C (PLC) 84 Phospholipase D (PLD) 86 Phosphorylation, of MLC 287, 287f corneal endothelium barrier integrity and effect of 287 Photophobia 419 Photorefractive keratectomy (PRK) 133, 140–141, 201, 234, 296 fibrotic scarring from 333 history of 194 history/evolution of 202 LASEK compared to 203–204 side effects of 198 Phototherapeutic keratectomy (PTK) 133, 226 for corneal epithelial defect 191 epithelial disease surgical intervention and 139

Pigment dispersion syndrome (PDS) 213 Pigment-epithelial-derived factor (PEDF) 475 pIgR see Polymeric immunoglobulin receptor Pitx2 transcription factor 12, 14f PK see Penetrating keratoplasty PKA see Protein kinase A PKC see Protein kinase C Platelet-derived growth factor (PDGF) 256, 260 PLC see Phospholipase C PLD see Phospholipase D PLR see Pupillary light reflex PLZF see Promyelocytic leukemia zinc finger protein PMMA see Polymethylmethacrylate PMN cells see Polymorphonuclear cells PNR see Pupillary near response Polymeric immunoglobulin receptor (pIgR) 343, 344f acute regulation of traffic of 343 Polymethylmethacrylate (PMMA) 207 ablation of 197–198 Polymodal nociceptors 158–159, 160f Polymorphonuclear cells (PMN cells) 164, 347 Pseudomonas aeruginosa keratitis and 407 PON see Pretectal olivary nucleus Positron emission tomography (PET) 7 Posterior amorphous dystrophy 231 Posterior blepharitis 97–98 Posterior lens capsule, in passive innate immune system 355 Posterior polymorphous dystrophy (PPMD) 213, 226 corneal endothelium and 281 Potassium-stimulated adenosine triphosphatase (Na, K-ATPase) 279, 280–281, 280f Potential difference (PD) 102–103 PPMD see Posterior polymorphous dystrophy PPRF see Paramedian pontine reticular formation Prechordal mesoderm, EOM origins and 10, 11f Pregnancy estradiol and 78 progesterone and 78 prolactin and 77–78 TGF-b and 79–80 Preocular tear film 83 Presbyopia 201 Pretectal olivary nucleus (PON) 490 Preterminal axons 37 Primary eye motility disorders treatment 47 Primary gaze 46, 46f PRK see Photorefractive keratectomy Prodrugs 303 conjunctiva’s role in topical drug delivery and 310 design of 303–304, 304f effectiveness of 303, 313 lipophilic ester design for 305–306, 307f, 308t transporter-targeted 312 Progenitors see Eye muscle progenitors Progesterone 77 pregnancy and 78 Progressive supranuclear palsy (PSP) 43 Prolactin 77 lacrimal gland and 77 pregnancy and 77–78 Promyelocytic leukemia zinc finger protein (PLZF) 293 Proprioception 33 EOM and 33 arguments against 33–34 Proprioceptors 33 Prosthetic keratoplasty 141 Proteases 303–304 Protein cholinergic agonists signal transduction and secretion of 112–113 in corneal endothelium and synthesis for external transport 278, 278f EGFP 327 GFP 318, 373 homeobox 272, 274–275 lacrimal gland and secretion of 70, 72f MAPK activation in 86f, 87 in lipid layer of tear film 61–62, 62f MBP 413 motor 68, 71 rab 68, 71 membrane trafficking and 71 SNARE 68, 71 membrane trafficking and 71–72

549

550

Subject Index

Protein (continued) in tear film 54, 54t, 70 transient receptor potential 171 transporter 303 Protein kinase A (PKA) 104 cAMP-PKA axis, in corneal endothelium barrier integrity 289 Protein kinase C (PKC) 86, 86f adrenergic agonist signal transduction and 88, 88f CEC proliferation regulation and signaling pathways of 294 Proteoglycans (PGs) 219 in corneal stroma 224 GAGs in 224 IOP lowered with 310 secretion of 222–223 small leucine-rich 224 PRRs see Pathogen recognition receptors Pseudoexfoliation syndrome 213 Pseudomonas aeruginosa keratitis 406 animal models for 411, 411f IL-18 and IFN- in 408 neuropeptides and 408, 410 apoptosis 410 SP 408 TLRs 409 VIP and 409 PMN, cytokines, chemokines in 407 T cells and IL-12 in 408 Pseudophakic bullous keratopathy 138, 213 Pseudostratified epithelium 126 PSP see Progressive supranuclear palsy Pterygium excision 139 PTK see Phototherapeutic keratectomy Ptosis 39, 91 acquired 93 congenital 93 of eyelids 93, 93f Pulsatile arterial compression, HFS and 8 Pump-leak mechanism 284f in corneal endothelium 286 Maurice’s hypothesis of 283, 284f Punctate epithelial erosions (PEEs) 97–98 Pupil advantages of mobil 487 alertness of 492 arousal of 492 cortical influences on 492 TEPRs 492 visually mediated 492 diameter of, pathways controlling 487, 488f, 489f sleep and 493 Pupillary light reflex (PLR) 487, 488f, 489, 490f afferent pathway of 488f, 489 IpRGCs 490, 490f PON 490 anatomy of 488f efferent pathway of 488f, 491 sympathetic influences on 488f, 491 Pupillary near response (PNR) 491 afferent influences on 491 efferent pathway of 491 Pursuit movements smooth 39 function of 42t speed of 45–46

Q Quadratus nictitans muscle 13–14 fiber-type determination in 15f

R RA see Rheumatoid arthritis Rab protein 68, 71 membrane trafficking and 71 Rabbit conjunctiva bioelectric studies on 102, 102f electrolyte transport systems of 103 epithelia of 101, 101f ion transport regulation in 104, 105f, 106f

fluid transport studies across isolated 105 Rada, Jody 519 Radial keratotomy (RK) 133, 140–141, 201 history/evolution of 194, 201 Receptive field 150 Recession surgery 45 Rectus muscles inferior, CNIII innervating 18–19, 19f lateral 19, 20f abduction with 45 CNVI innervating 18–19, 19f eye muscle progenitors and formation of 12 medial 17, 19, 20f adduction with 45 CNIII innervating 18–19, 19f superior 17–18, 18f, 19, 20f CNIII innervating 18–19 Recurrent corneal erosion 164, 226 Reflex blinking 6 Refractive index 60 Refractive surgery 140, 141f, 143 see also Epithelial laser in situ keratomileusis; Femtosecond laser in situ keratomileusis; Laser-assisted in situ keratomileusis; Laser-assisted subepithelial keratomileusis; Photorefractive keratectomy; Radial keratotomy benefits of 201 corneal biomechanical effects in 198 corneal inlays and 205 corneal nerve regeneration after 156 evolution of 201 history of 194, 194t myopia and options for 523 optical aberrations/visual quality after 195, 196f optimization of 199, 199f spherical aberration following 197, 198f wave front technology and laser 205 wound healing/haze after 195, 195f Regeneration 256 with AC 300, 301f of BM/Bowman’s membrane 263–264 of corneal nerves cataract surgery 156 ocular surgery 155 PK 156, 157f refractive surgery 156 GFs regulating 257–258 TGF-b mediating 257–258, 258–259, 258t Regulatory T cells 396 Regulatory volume decrease (RVD) 174 Regulatory volume increase (RVI) 174 Reis-Buckler’s dystrophy 227, 227f Reporter gene 327 Reproductive hormones bioavailability of, regulation of 78f, 76 counterpoises between contradictory signals of 80 lacrimal gland exocrine functions and influence of 80–81, 82f mucosal immune system tissues and influence of 79 prolactin 77 lacrimal gland and 77 pregnancy and 77–78 Reptation 116 Resection surgery 45 Resolution acuity 494, 495f Resolution limit 500 Retina APCs of 374f, 379, 379t visual acuity across 497, 497f Retinal detachment 517 Retinal pigment epithelium (RPE) 355 Retinoblastoma gene 290 Retraction of eyelids 93, 94f Retractor bulbi muscle 3 Retrograde axonal transport 396 Reverse tetracycline transcriptional activator (rtTA) 315 Pax6OS-rtTA 321 RGP see Rigid gas permeable materials Rheology 126 Rheumatoid arthritis (RA) 191 Rhinitis 419 Rigid gas permeable materials (RGP) 207, 210 riMLF see Rostral interstitial nucleus of medial longitudinal fasciculus

Subject Index

River blindness see Onchocerciasis RK see Radial keratotomy Rostral interstitial nucleus of medial longitudinal fasciculus (riMLF) 41f, 43 RPE see Retinal pigment epithelium rtTA see Reverse tetracycline transcriptional activator RVD see Regulatory volume decrease RVI see Regulatory volume increase

S SAC see Seasonal allergic conjunctivitis Saccades 3, 39, 42t blinking combined with 5–6 downward 4–5 generation of 45–46 horizontal 40f, 43 SAI see Surface asymmetry index Sarcomeric mitochondrial CK (sCK) 28–29 Satellite cells 17 in EOM 25–26, 26f Scar-reducing therapies currently available 268, 268t emerging 269 Scars see Fibrotic scarring SCC see Sebaceous cell carcinoma; Squamous cell carcinoma Scheimpflug imaging 133 Schele’s syndrome 219 Schlemm’s canal 272 Schnyder’s crystalline dystrophy 230 sCK see Sarcomeric mitochondrial CK Sclera, APCs of 374f, 379, 380t Sclerotic scatter 213, 214 Scott, Alan 47–48 SEALs see Superior epithelial arcuate lesions Seasonal allergic conjunctivitis (SAC) 114, 419, 420t see also Ocular allergies Sebaceous cell carcinoma (SCC) 96–97, 97f Seborrheic blepharitis 97–98 Secreted mucins 118, 120f Secretory immunoglobulin A (sIgA) corneal epithelium infection and role of 455 lacrimal glands contributing 74–75 lactation and 79–80 in nasolacrimal ducts 130 at ocular surface 433f, 436 role of 357 in tear film 446 Secretory phospholipase A2 356 Self-assembled corneal equivalents 299 Sensitive period 525 for amblyopia development 526 Sensory afferents 158 Sensory receptors 158 cold-sensitive thermal 159–160 of cornea functional properties of 158, 159f, 160f local inflammation and 160f, 161 ion channels and 159f, 161 mechanonociceptors 159, 160f polymodal nociceptors 158–159, 160f silent nociceptors 160 SEP see Subepithelial plexus Serotype 327 Sex hormone-binding globulin (SHBG) 74 Sex hormones see Reproductive hormones Sex steroids 75 androgens 75 estradiol 77 pregnancy and 78 progesterone 77 prolactin 77 lacrimal gland and 77 pregnancy and 77–78 SHBG see Sex hormone-binding globulin Short interfering RNA (siRNA) 406 Short-circuit current 99 Siderophore 444 sIgA see Secretory immunoglobulin A SIGIRR see Single immunoglobulin interleukin 1 receptor-related protein Signal transduction 83, 108, 158, 347

551

adrenergic agonists and 88 Ca2+ and PKC 88, 88f MAPK 88f, 89 NO 88f, 89 cholinergic agonists and 112, 113f protein secretion in 112–113 lacrimal gland activation of 84, 85f cholinergic agonist-activated 84, 86f MAPK-coupled 86f, 87 PLC-coupled 84 PLD-coupled 86 VIP and 87 adenylate cyclase 87, 87f MAPK 88 Silent nociceptors 160 Single immunoglobulin interleukin 1 receptor-related protein (SIGIRR) 406 siRNA see Short interfering RNA Sjo¨gren syndrome 114 Skeletal muscles fiber types of, in EOM 22 myogenesis of 9 in paraxial mesoderm 9 Sleep 493 Slit-lamp biomicroscopy 213 SLRP see Small leucine-rich proteoglycan Small leucine-rich proteoglycan (SLRP) 224 Smooth pursuit 39, 42t SNARE protein 68, 71 membrane trafficking and 71–72 Snellen chart 494–495, 495f, 496, 510f, 525 SO see Superior oblique muscle Sodium bicarbonate cotransporter (NBCel) 285, 285f Sodium-potassium-dependent ATPase (Na, K-ATPase) 74 SOM see Somatostatin Somatostatin (SOM) 368 in aqueous humor 370 Somite cells 11 Sonoporation 327 SP see Substance P Spatial frequency 500, 501, 502f, 503f, 504f Specular microscopy 214, 215f Spherical aberration 197, 198f Split scanning confocal microscope (SSCM) 234 Squamous cell carcinoma (SCC) 96, 96f Squamous epithelium 178 SR see Superior rectus muscle SRI see Surface regularity index SSCM see Split scanning confocal microscope Staphylococcal blepharitis 97–98 Stem cells 272 see also Corneal epithelial stem cells; Limbal stem cells conjunctiva 182 cornea endothelium 183 corneal stroma 183, 223 properties of 180 Stensen, Niels 126 Stevens-Johnson syndrome 185, 186 Stokes lens 506 Strabismus 17, 39, 45, 91, 525, 529 botulinum toxin A in surgery for 47–48 treatment for 47 Streilein, J. Wayne 362 Striated muscles see Skeletal muscles Stroma see Corneal stroma Stromal deturgescence 173 Stromal disease 137 degenerations of 137 dystrophies of 137, 137t infection of 137, 137f surgical intervention for 140 anterior lamellar keratoplasty 140 corneal biopsy 140 corneal laceration repair 140 PK 140, 140f prosthetic keratoplasty 141 refractive surgery 140, 141f Stromal fibrosis 222–223, 222t Stromal plexus 151 Stromal puncture 140 Subapical actin cytoskeleton 71 Subbasal nerve plexus 152, 153f

552

Subject Index

Subepithelial disease 136 EBM dystrophy 136 subepithelial infiltrates 136 Subepithelial infiltrates 136 Subepithelial plexus (SEP) 151 Subluxation 213 Substance P (SP) 407 corneal epithelium infection response of 456 in Pseudomonas aeruginosa keratitis 408 IFN- production and 408 Superficial keratectomy 185 Superior epithelial arcuate lesions (SEALs) 207, 211 Superior oblique muscle (SO) 17, 19 anatomy of 19–20, 20f Superior rectus muscle (SR) 17–18, 18f, 19, 20f CNIII innervating 18–19 Surface asymmetry index (SAI) 57 Surface regularity index (SRI) 57 Surface-associated mucins 118, 119f Sutural fibers 220 Symblepharon 185 Sympathetic nerves 69 in lacrimal gland 83–84 Sympathetic nervous system, ACAID and 362, 363f

T T cell receptors (TCRs) 347–348, 348–349 T cells 396 activation of 348, 349f adaptive immune system and antigen presentation to 348, 349f aqueous humor and 367–368 CD4+ 381, 385 differentiation and effector function of 350, 351f eye immune responses of 352, 352t helper 459 in HSK 397 in Pseudomonas aeruginosa keratitis 408 regulatory 396 T helper response 406 T regulatory cells (Tregs) 361 in ACAID 363 Tandem repeats 116 Tandem scanning confocal microscope (TSCM) 234 CTMF on 236 Tarsal plate 60 Tarsorrhaphy 140 Task-evoked pupillary responses (TEPRs) 492 TBUT see Tear breakup time TCRs see T cell receptors Tear breakup time (TBUT) 60–61 Tear drainage see Efferent tear ducts; Lacrimal passages; Nasolacrimal ducts Tear ducts see Efferent tear ducts; Lacrimal passages; Nasolacrimal ducts Tear film 51 aqueous layer of 53 distribution of 55 blinking and 55 drainage of 58, 58f evaporation from 53, 53f hyperosmolarity 57–58 layers of 109–110 lipid layer of 52, 61 contact lenses and 66 meibomian glands and 52, 53t meibomian glands function in 60–61, 62f models for 61, 62f protein in 61–62, 62f lubrication and 51 mechanical washing effect of 433 mucin layer of 55 ocular surface and absorption of 59 anchoring of 123, 123f osmolarity 57 particle movement of 57 in passive innate immune system 354, 356f preocular 83 production of 54, 55f protein concentration in 54, 54t, 70 sIgA in 446

stability of 55 fluorescein sodium for measuring 55–56, 56f, 57f structure and thickness of 51, 52f videokeratoscopy for assessing 57 Tear fluid cornea and conjunctiva connected through 460, 461f inflammation and 463 inflammation and 460 as diagnostic indicator 462 Tear transport in canaliculi 128 in nasolacrimal ducts and lacrimal sac 129, 129f Tear turnover (TTR) 54, 55, 55f Tears defense mechanisms of 444, 445f, 446f, 447f drainage of 58, 58f, 127t evaporation and breakup of 124 nasolacrimal duct absorption of 131, 131f ocular surface absorption of 59 secretion of 109–110, 383, 383f, 384 Temporal contrast-sensitivity function 501f, 503, 504f, 505f Temporal frequency 500 Tendinous annulus 17, 18f Tenon’s capsule 17 TEPRs see Task-evoked pupillary responses TER see Transendothelial electrical resistance Testosterone, estradiol and 77 Tet-ON system 315, 325 see also Cre-LoxP system Tetracyclines for dry eye 389 operator element 315 TFF see Trefoil factor family TGFa see Transforming growth factor-alpha TGF-b see Transforming growth factor-beta TH see Tyrosine hydroxylase Th1 cytokine 459 Th2 cytokine 459 Thiel-Behnke 227 Thin film interferometer 57 Thygeson’s punctate keratitis 136 Thymosin-b4 457 Thyroid ophthalmopathy 40 Tight junctions (TJs) 167, 167t, 283, 290 barrier integrity and 286–287 breakdown of 456 of corneal endothelium 283, 284f TJs see Tight junctions TLRs see Toll-like receptors T-lymphocytes in conjunctival lamina propria 434f, 435 corneal graft rejection and afferent immune response 393 efferent immune response 393 ocular allergies and 422, 422f, 423f TNF-a see Tumor necrosis factor Tolerance 432 Toll-like receptors (TLRs) 357, 376, 401, 426 activation of 449f corneal epithelium infection and response of 453 distribution of 448, 449f Pseudomonas aeruginosa keratitis and 409 Wolbachia and 402, 404f, 405f TLR2/TLR6 403 Topical drug delivery 303 see also Prodrugs conjunctiva’s role in 308 prodrug strategy in 310 transconjunctival pathway in 310, 311f, 312f, 312t cornea’s role in 305, 305f, 307f, 308t, 309f subconjunctival 313 Trabecular beam 272 Trabeculectomy 150 Trachoma 66 Transcription factor corneal endothelium and role of 274, 277f Lbx1 12 pitx2 12, 14f Transcytosis 68, 71 Transendothelial electrical resistance (TER) 283

Subject Index

Transepithelial resistance 99 of conjunctiva, determining 103 Transepithelial voltage 99 Transforming growth factor-alpha (TGF-a) 324, 325f Transforming growth factor-beta (TGF-b) 74, 77, 220, 256, 290, 339, 341, 361 autoinduction of 259–260 in CEC aqueous humor 293 corneal epithelial wound healing and signaling pathways of 263, 323 in embryonic development 323 fibrotic scarring and regeneration mediated by 257–258, 258–259, 258t immunosuppression and 368 myofibroblasts and 222–223, 222t ocular wound healing and 260–261 pregnancy and 79–80 Transgenesis development of 315 of mice corneal-epithelium-specific promoter for 318 GFs in ocular surface tissue morphogenesis in 321 keratocyte-specific promoter for 317 ocular-surface tissue-specific drive lines of 318 ocular-surface tissue-specific promoter identification in 317 Pax6 gene for 318 strategies for 317 process of 315 Trans-Golgi network 68, 70–71 Transient amplifying cell 272 Transient receptor potential protein 171 Transillumination of eyelids 64–65, 65f Transporter protein 303, 312 Trapezius muscle morphogenesis 14 Trefoil factor family (TFF) 130 Tregs see T regulatory cells Trichiasis 91 eyelids and 95 Trigeminal ganglia 396 HSV-1 in 397 Trigeminal nerve 6–7 Tritiated thymidine-labeling 452 Trochlear nerve (CNIV) 17, 18–19 abnormal eye movements from disorders of 42 Trophic substances 150, 452 Trophozoites 413 for Acanthamoeba keratitis 417 Tropia 45 Trunk myogenesis 14t TSCM see Tandem scanning confocal microscope TTR see Tear turnover Tumor necrosis factor (TNF-a) 284f, 288 Tumors of eyelids BCC in 95, 96f benign 97 malignant melanoma 97, 97f SCC 96, 96f sebaceous cell carcinoma 96–97, 97f Tumstatin 475 Tyrosine hydroxylase (TH) 54

U Ubiquitous mitochondrial CK (uCK) 28–29 UBM see Ultrasound biomicroscopy uCK see Ubiquitous mitochondrial CK Ultrasound biomicroscopy (UBM) 215 Unidirectional fluxes 99 Ussing, Hans 99 Ussing chamber 99, 102 conjunctiva isolation with 103 Uvea, APCs of 374f, 378, 378t

V VAChT see Vesicular acetylcholine transporter Vascular endothelial growth factor (VEGF) 148–149, 385, 473 in CNV 466 Vasculogenesis 465, 470 Vasoactive intestinal peptide (VIP) 69–70, 83–84 in aqueous humor 370

cAMP production and 87, 87f corneal epithelium infection response of 456 goblet cells and receptors for 113 Pseudomonas aeruginosa keratitis and 409 signal transduction and 87 adenylate cyclase 87, 87f MAPK 88 VCAID see Vitreous cavity-associated immune deviation Vectors 327 lentivirus 330–331, 331f nonviral 331 viral 330, 330f, 331f VEGF see Vascular endothelial growth factor Vergence 3, 42t Vernal keratoconjunctivitis (VKC) 114, 185, 420, 420t see also Ocular allergies Vesicular acetylcholine transporter (VAChT) 33, 37 Vestibulo-ocular reflexes (VOR) 39, 42t Wernicke’s encephalopathy and 43 Videokeratoscopy 51 tear film assessments using 57 VIP see Vasoactive intestinal peptide Viral vectors 330, 330f, 331f Virion 396 Visante OCT 238 applications of 238 pachymetry map 239f Visual acuity detection and resolution 494, 495f dynamic 498 hyperacuity 497, 498f measurement of 494, 495f optical/neural limits on 496 over life 497, 498f reporting 496, 496t retina and 497, 497f standards for 497 Visual angle 494 Visual fixation 42t Vitamin A deficiency 114 Vitreous cavity-associated immune deviation (VCAID) 361, 364, 365t immune privilege and 364 VKC see Vernal keratoconjunctivitis VOR see Vestibulo-ocular reflexes

W Wave front aberrations 506 Wave front technology for HOAs 205 for ocular astigmatism 205 Wavelets 500, 501–502, 502f Wernicke’s encephalopathy 43 Whitnall’s ligament 3–4, 4f With-the-rule 506, 507 Wnt family of growth factors 12 Wolbachia 401 filarial disease role of 401 in onchocerciasis pathogenesis 402, 403f TLRs and 402, 404f, 405f TLR2/TLR6 403 Worm-like model 116 Wound healing see also Corneal epithelial wound healing ocular 260, 260f TGF-b and 260–261

Y Young’s modulus 116

Z Zim, Edward 481 Zonula occludens 452

553