XENOTRANSPLANTATION BASIC RESEARCH AND CLINICAL APPLICATIONS EDITED BY
JEFFREY L. PLATT, MD
HUMANA PRESS
i
Xenotransplantation
ii
iii
Xenotransplantation Basic Research and Clinical Applications
Edited by
Jeffrey L. Platt, MD Transplantation Biology, Mayo Clinic Rochester, MN
HUMANA PRESS TOTOWA, NEW JERSEY
iv © 2002 Humana Press Inc. 999 Riverview Drive, Suite 208 Totowa, New Jersey 07512
www.humanapress.com For additional copies, pricing for bulk purchases, and/or information about other Humana titles, contact Humana at the above address or at any of the following numbers: Tel.: 973-256-1699; Fax: 973-256-8341, E-mail:
[email protected]; or visit our Website: http://humanapress.com All rights reserved. No part of this book may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, microfilming, recording, or otherwise without written permission from the Publisher. All articles, comments, opinions, conclusions, or recommendations are those of the author(s), and do not necessarily reflect the views of the publisher. Due diligence has been taken by the publishers, editors, and authors of this book to assure the accuracy of the information published and to describe generally accepted practices. The contributors herein have carefully checked to ensure that the drug selections and dosages set forth in this text are accurate and in accord with the standards accepted at the time of publication. Notwithstanding, as new research, changes in government regulations, and knowledge from clinical experience relating to drug therapy and drug reactions constantly occurs, the reader is advised to check the product information provided by the manufacturer of each drug for any change in dosages or for additional warnings and contraindications. This is of utmost importance when the recommended drug herein is a new or infrequently used drug. It is the responsibility of the treating physician to determine dosages and treatment strategies for individual patients. Further it is the responsibility of the health care provider to ascertain the Food and Drug Administration status of each drug or device used in their clinical practice. The publisher, editors, and authors are not responsible for errors or omissions or for any consequences from the application of the information presented in this book and make no warranty, express or implied, with respect to the contents in this publication. Cover design by Patricia F. Cleary. This publication is printed on acid-free paper. ' ANSI Z39.48-1984 (American National Standards Institute) Permanence of Paper for Printed Library Materials. Photocopy Authorization Policy: Authorization to photocopy items for internal or personal use, or the internal or personal use of specific clients, is granted by Humana Press Inc., provided that the base fee of US $10.00 per copy, plus US $00.25 per page, is paid directly to the Copyright Clearance Center at 222 Rosewood Drive, Danvers, MA 01923. For those organizations that have been granted a photocopy license from the CCC, a separate system of payment has been arranged and is acceptable to Humana Press Inc. The fee code for users of the Transactional Reporting Service is: [0-89603-674-X/02 $10.00 + $00.25]. Printed in the United States of America. 10 9 8 7 6 5 4 3 2 1 Library of Congress Cataloging-in-Publication Data Xenotransplantation: basic research and clinical applications / edited by Jeffrey L. Platt. p. cm. Includes bibliography and index. ISBN 0-89603-674-X 1. Xenografts. I. Platt, Jeffrey L. QR188.8.X453 2002 617.9’5–dc21 2001051652
v
Preface No field of medicine has engendered greater excitement or enjoyed greater success than the field of transplantation. Organ transplantation allows the “cure” of disease by replacing failing organs with physiologically normal organs. Tissue transplantation and tissue engineering allow not only the replacement of abnormal cells, such as bone marrow cells, but also the possibility of using a transplant to impart novel physiologic functions. The major limitation to applying transplantation for the treatment of disease is a shortage of human donors. This shortage limits transplant procedures to as few as five percent of those that would be carried out if the supply of organs and tissues were unlimited. Because of this shortage and because of recent advances in fundamental knowledge, there has been a crescendo of interest in xenotransplantation, the use of animals in lieu of humans as organ and tissue donors. For many years, xenotransplantation has seemed only a distant prospect because of the severe immune responses of the host against the graft. Recent studies, however, have revealed the molecular basis of these immune responses and have given rise to novel therapeutic approaches for circumventing them. For example, the generation of transgenic animals expressing human complement regulatory proteins or human glycosyltransferases raises the prospect that the severest type of rejection can be avoided without manipulating the xenograft recipient. Thus, xenotransplantation has quickly moved to center stage in the field of transplantation, engaging the interest of clinicians, basic scientists, and academicians. Xenotransplantation: Basic Research and Clinical Applications compiles and explains the fundamental molecular and cell biology that has been applied with such advantage in the emerging fields of transplant immunology and xenotransplantation. The contributors to this book are established authorities in transplant immunology and molecular and cell biology. This book provides a base of knowledge for the practitioner, fellow, and student, and those involved in biotechnology and related sciences. Jeffrey L. Platt, MD v
vi
vii
Contents Preface ................................................................................................ v Contributors ..................................................................................... ix 1
Molecular and Cellular Hurdles to Xenotransplantation Jeffrey L. Platt ......................................................................... 1
2
Pathological Responses to Xenotransplantation Matilde Bustos and Jeffrey L. Platt ...................................... 45
3
Natural Xenoreactive Antibodies Uri Galili ............................................................................... 57
4
Specificity of Xenoreactive Natural Antibodies William Parker, Paul B. Yu, and Yuko C. Nakamura ..........73
5
Biophysical Properties of Xenoreactive Natural Antibodies William Parker, Ryan C. Fields, and Yuko C. Nakamura..................................................... 87
6
The Origin of Xenoreactive Natural Antibodies Paul B. Yu............................................................................ 103
7
Synthesis of Carbohydrate Antigens Recognized by Xenoreactive Antibodies Mauro S. Sandrin and Ian F. C. McKenzie ........................119
8
The Complement Barrier to Xenotransplantation Agustin P. Dalmasso ...........................................................139
9
Defects and Amplification of Costimulation Across the Species Nicola Rogers and Robert Lechler .....................................173
vii
viii
Contents
10
Antibody-Dependent Effects on Cellular Immunity Antonello Pileggi, R. Damaris Molano, Thierry Berney, and Luca Inverardi .........................................................199
11
Disordered Regulation of Coagulation and Platelet Activation in Xenotransplantation Simon C. Robson .................................................................215
12
Current Applications of Cellular Xenografts Albert S. B. Edge .................................................................247
Index ...............................................................................................265
ix
Contributors THIERRY BERNEY • Diabetes Research Institute, University of Miami School of Medicine, Miami, FL MATILDE BUSTOS • Department of Surgery, Universidad de Navarra, Pamplona, Spain AGUSTIN P. DALMASSO • Departments of Surgery and Laboratory Medicine and Pathology, University of Minnesota Medical School, Minneapolis, MN ALBERT S. B. EDGE • Diacrin, Inc., Charlestown, MA RYAN C. FIELDS • Department of Surgery, Duke University Medical Center, Durham, NC URI GALILI • Departments of Cardiovascular-Thoracic Surgery and Immunology and Microbiology, Rush Medical College, Chicago, IL LUCA INVERARDI • Diabetes Research Institute, University of Miami School of Medicine, Miami, FL ROBERT LECHLER • Department of Immunology, Imperial College of Science, Technology, and Medicine, London, UK B RUCE L OVELAND • Molecular Immunogenetics Laboratory, Austin Research Institute, Austin and Repatriation Medical Centre, Heidelberg, Victoria, Australia IAN F. C. MCKENZIE • Molecular Immunogenetics Laboratory, Austin Research Institute, Austin and Repatriation Medical Centre, Heidelberg, Victoria, Australia R. DAMARIS MOLANO • Diabetes Research Institute, University of Miami School of Medicine, Miami, FL YUKO C. NAKAMURA • Department of Surgery, Duke University Medical Center, Durham, NC WILLIAM PARKER • Department of Surgery, Duke University Medical Center, Durham, NC ANTONELLO PILEGGI • Diabetes Research Institute, University of Miami School of Medicine, Miami, FL JEFFREY L. PLATT • Transplantation Biology and the Departments of Surgery, Immunology, and Pediatrics, Mayo Clinic, Rochester, MN ix
x
Contributors
SIMON C. ROBSON • Center for Immunobiology, Department of Medicine, Beth Israel-Deaconess Medical Center, Boston, MA NICOLA ROGERS • Department of Immunology, Imperial College of Science, Technology, and Medicine, London, UK MAURO S. SANDRIN • Molecular Immunogenetics Laboratory, Austin Research Institute, Austin and Repatriation Medical Centre, Heidelberg, Victoria, Australia PAUL B. YU • Division of Cardiology, Massachusetts General Hospital, Boston, MA
Hurdles to Xenotransplantation
1
1
Molecular and Cellular Hurdles to Xenotransplantation Jeffrey L. Platt, MD, PhD INTRODUCTION
The transplantation of organs, tissues, or cells between individuals of different species has been of increasing interest in recent years because the use of animals as organ and tissue donors as a source of transplants would overcome the severe and worsening shortage of human organs available for transplantation. This shortage restricts the application of organ transplantation to 5–15% of the patients who might benefit in the United States (1). Indeed, the shortage of donor organs is now widely acknowledged to be the major limitation of transplantation. Interest in xenotransplantation also arises because a xenotransplant might, in principle, be less susceptible to infection by viruses or other agents that caused the primary disease. Avoiding viral infection was the rationale for several baboon-to-human liver transplants (2) and for the transplantation of baboon bone marrow in a human patient with AIDS. Interest in xenotransplantation may further arise because the animal source can be subjected to genetic engineering, and such engineering might provide a vehicle for delivery of genes or enduring expression of genes (3). For example, genetic engineering might be used to express therapeutic genes in stem cells.
TISSUE SOURCE AND DONOR FACTORS IN XENOTRANSPLANTATION The type of organ or tissue transplanted and the phylogenetic distance between the donor and the recipient have profound importance for the immune response to xenotransplantation (Fig. 1). The type of organ
From: Xenotransplantation: Basic Research and Clinical Applications Edited by: Jeffrey L. Platt © Humana Press Inc., Totowa, NJ
1
2
Platt
Fig. 1. The immunological response to xenotransplantation. The immune response to xenotransplantation can be classified according to whether the graft consists of isolated cells or free tissues, such as islets of Langerhans or of a primarily vascularized organ such as the kidney or heart. (A) Vascularized organ grafts are subject to hyperacute and acute vascular rejection caused by the action of antidonor antibodies on donor endothelium. If hyperacute or acute vascular rejection are averted, the graft may undergo accommodation, a condition in which the graft appears to resist injury despite the return of anti-donor antibodies to the circulation and the presence of an intact complement system. A vascularized organ graft may also be subject to cellular rejection and chronic rejection more or less like the corresponding types of rejection observed in allografts. (B) Free tissue grafts are subject to failure caused by primary nonfunction, failure of neovascularization or failure of the microenvironment to support the survival and function of the foreign tissue. If the free tissue or isolated cells engraft, they are then subject to cellular or humoral rejection.
or tissue grafted determines the nature of the blood supply and, thus, the nature of the contact between the transplant and the immune system of the host (Fig. 2). The type of transplant also determines the relative importance of “growth” factors for engraftment and the survival and function of the graft. The phylogenetic distance determines the compatibility of growth factors and growth factor receptors between donor and recipient (4,5). The type of graft, i.e., cell, tissue, or organ, also determines how donor antigen comes to be presented to the immune system of the recipient. Xenografts consisting of isolated cells, such as hepatocytes, derive their vascular supply entirely by the in-growth of blood vessels of the recipient (6) (Table 1). Free tissue grafts, such as pancreatic islets or
Hurdles to Xenotransplantation
3
Fig. 2. Antigen presentation in a free tissue graft: implications for the immunogenicity of xenografts. Free tissue xenografts are generally vascularized by spontaneous anastomosis of donor and recipient blood vessels and by neovascularization. Spontaneous anastomosis allows the presentation of foreign antigen by foreign MHC as modeled in the upper left of the figure. Neovascularization results in presentation of foreign antigen by recipient MHC as in the upper right and lower parts of the figure. If neovascularization or spontaneous anastomosis were to be impaired, the mechanism of antigen presentation might differ from the mechanism that would predominate in a free tissue allograft.
Table 1 Classification of xenografts Type of xenograft (example) isolated cells (hepatocytes, bone marrow) free tissue (pancreatic islets, skin) organ (kidney, heart)
Type of vascular supply
Microenvironment
neovascularization
recipient
neovascularization + anastomosis of donor and recipient vessels primary anastomosis of donor and recipient vessels
donor and recipient donor
4
Platt
skin, derive their vascular supply in part from the host and in part from the spontaneous anastomosis of donor and recipient blood vessels. The blood vessels of the recipient in cell and tissue grafts pose a barrier between the graft and the immune system of the recipient. This barrier may be sufficient to allow survival of xenografts with no more immunosuppression than allografts (7). Blood vessels of recipient origin might, in principle, take up and present antigen of the donor. Antigen presented in this way to T lymphocytes is said to be presented through the “indirect” pathway (8) (Fig. 3). The type of rejection typically seen in cell and tissue xenografts is cellular rejection. In contrast to cellular xenografts, organ xenografts provide their own blood vessels. The interaction of the immune system of the recipient with donor blood vessels gives rise to distinct types of vascular or humoral rejection (Fig. 1). Organ xenografts are also subject to cellular rejection. The donor blood vessels of organ xenografts may present antigen to T lymphocytes through the “direct” pathway (Fig. 3). The nature of the blood supply and the microenvironment may also determine the biological viability of the graft (Table 1). Cell and tissue grafts may depend on growth factors of recipient origin. If these growth factors are not compatible with the xenogeneic cells, the graft may fail. An important example of such incompatibility can be found in the transplantation of xenogeneic bone marrow (5). In contrast, organ xenografts generally provide the factors needed for survival of the xenogeneic cells in the graft. Genetic differences between the donor and the recipient are another important factor in xenotransplantation (Table 2). Phylogenetic distance between the donor and the recipient, of course, determines the number of antigens that might serve as a target of the immune response. However, certain genetically controlled traits have a disproportionate impact on the outcome of a xenograft. For example, the expression of a functional _1,3-galactosyltransferase gene in lower mammals leads to the synthesis of Gal_1-3Gal as the terminus of certain oligosaccharide chains. Gal_1-3Gal is recognized by certain naturally occurring antibodies made by humans, apes, and Old World monkeys (9). Although this antigen–antibody system constitutes a severe immunological barrier to xenotransplantation, the distribution of that glycosyltransferase in phylogeny is not a function of genetic distance (Table 3). Another trait under genetic control is the regulation of the complement system. The complement system is regulated by plasma and cell surface proteins
Hurdles to Xenotransplantation
5
Fig. 3. Antigen presentation in xenotransplantation. Antigen may be presented by the direct or indirect pathways. When direct antigen presentation occurs, the T cells of the recipient recognize undenatured MHC antigen of the donor expressed by donor APC or other donor cells. The peptide associated with MHC is of donor origin and may either dictate specificity or may be nominal with regard to specificity of recognition. For indirect antigen presentation, the T cells of the recipient recognizes donor peptide expressed in association with MHC of the recipient.
Table 2 Phylogeny and the Susceptibility to Hyperacute Rejection
Donor
Organ
Recipient
hamster guinea pig pig dog pig NW monkey OW monkey
heart heart kidney kidney heart heart heart
rat rat dog pig baboon baboon baboon
a
Graft a Survival
Histology of Rejected Graft
Selected Reference
4 1/4 d 1/4 h 1/3 h 2 1/2 h 3h 1h 6d
AVRb HAR HAR HAR HAR HAR AVR
211 212 25 25 213 57 214
The graft survival and histology shown are typical for this type of transplant in a recipient receiving a “routine” immunosuppression regimen. b Abbreviations used: NW, New World; OW, Old World; AVR, acute vascular rejection; HAR, hyperacute rejection.
6
Platt
Table 3 Phylogeny of Gal_1-3Gal and Natural Antibodies Specific for that Saccharide
Species chicken mouse rat pig dog New world monkey Old world monkey Baboon human
_1,3-galactosyl transferase
Expression of Gal_1-3Gal
Natural antiGal_1-3Gal antibodies
– + + + + + – – –
– + + + + + – – –
+ – – – – – + + +
that function in a species-specific fashion (10) (Table 4). For example, the cells of the guinea pig allow the unrestrained activation of the rat alternative pathway of complement, and, as a result, organs of guinea pigs transplanted into rats are subject to very rapid and vigorous rejection (11,12). Fortunately, the alternative pathway of the human complement system is not primarily activated on porcine cell surfaces (13–15), and, thus, pigs may serve as a useful source of cells and organs for xenotransplantation into humans. Other genetic differences that may contribute to the outcome of xenografts include the species-specific functioning of cell-associated complement regulatory proteins such as decay accelerating factor (DAF), CD59 (16), and the potential speciesspecific functioning of the thrombomodulin vis-à-vis thrombin and protein C (17).
THE BIOLOGICAL RESPONSES TO ORGAN XENOTRANSPLANTATION Figure 1 shows a model for the biological responses to xenotransplantation, which manifest predominantly as rejection reactions. The sections that follow will summarize the molecular and cellular mechanism underlying these rejection reactions and how this information can be used to develop strategies for the prevention and treatment of rejection.
Serum
7
Cells
Cow Rabbit Horse Pigeon Pig Human Dog Rat Guinea Pig Sheep
Hurdles to Xenotransplantation
Table 4 Activation of Complement on Xenogeneic Cells Via the Alternative Pathway
Cow
Rabbit
Horse
Pigeon
Pig
Human
Dog
Rat
Guinea Pig
Sheep
_ 2.6 0.59 2.34 <0.28 >4.4 <0.28 0.36 >4.4 <0.28
<0.28 – <0.28 0.48 <0.28 <0.28 1.43 0.28 0.41 <0.28
<0.28
<0.28 1.14 0.67 – 1.23 0.35 1.85 0.37 0.55 <0.28
<0.28 0.58 0.41 1.14 – 0.40 0.56 0.39 2.3 <0.28
<0.28 1.08 0.49 0.37 <0.28 – 4.4 0.41 0.55 <0.28
<0.28 1.27 2.34 1.78 2.46 <0.28 – 2.2 2.46 0.70
<0.28 0.35 0.74 <0.28 0.29 <0.28 4.4 – 0.61 <0.28
<0.28 0.45 <0.28 <0.28 <0.28 <0.28 4.4 0.41 – <0.28
<0.28 0.41 <0.28 <0.28 <0.28 0.73 >4.4 0.38 >4.4 –
– <0.28 <0.28 <0.28 3.4 <0.28 <0.28 <0.28
8
The number of units of complement per milliliter of serum of 10 different species capable of lysing 50% of 5 × 10 RBC’s of other species by alternative pathway. Adapted from: Edwards, J. Transplantation 1981; 31:226, (101).
7
8
Platt
Whole organ xenografts are subject to various types of rejection listed in Fig. 1. These types of rejection consist predominantly of vascular disease caused by interaction of the immune system of the recipient with donor blood vessels. Figuring prominently in these types of rejection is intravascular coagulation caused by the conversion of blood vessels from anticoagulant to procoagulant and ischemia caused by the constriction or occlusion of blood vessels in the graft.
THE BIOLOGICAL RESPONSES TO CELL AND FREE TISSUE XENOGRAFTS Cell and free tissue xenografts are not subject to hyperacute rejection but do generally undergo rejection over a period of days, a tempo often more rapid than that observed for free tissue allografts. Rejecting free tissue xenografts may be infiltrated by host T cells and macrophages (18–20). Free tissue and cellular xenotransplants may also fail because of “primary nonfunction” (21), which may result from immune or nonimmune factors. The extent to which free tissue xenografts are subject to humoral rejection following vascularization has not been established; however, studies involving the transfer of immune serum demonstrate that anti-donor antibodies can mediate rejection of free tissue grafts and, thus, suggest that humoral responses could be an important determinant of graft outcome (22–24).
Hyperacute Rejection Organs transplanted between certain species are subject to hyperacute xenograft rejection (25,26) (Table 2). Combinations of donor and recipient species in which hyperacute rejection regularly occurs in unmodified recipients are called “discordant”; combinations of donor and recipient species in which hyperacute rejection rarely occurs are called “concordant” (27). The susceptibility to hyperacute rejection among various species of organ transplant donors and recipients is summarized in Table 2. Hyperacute xenograft rejection is characterized histologically by interstitial hemorrhage, edema, platelet thrombi, and severe injury to endothelial cells (28–33). In some cases, prominent infiltration of neutrophils is observed. Evidence of injury to graft endothelial cells—the presumed target of the hyperacute rejection reaction—includes swelling, vesiculation, alteration in cellular junctions, detachment, and, in some cases, lysis.
Hurdles to Xenotransplantation
9
The pathogenesis of hyperacute rejection depends absolutely on the activation of complement. Activation of complement may occur as a result of: 1. The binding to the graft of complement-fixing xenoreactive antibodies, 2. The direct activation of complement via the alternative complement pathway on the foreign cell surfaces, and/or 3. The failure of complement regulation in the foreign organ (34).
Regardless of how complement is activated, most evidence suggests that endothelial cells are the primary target of complement in hyperacute rejection (14,16). In addition to the structural damage to endothelium, the immunopathology of hyperacute rejection invariably reveals complement components and often immunoglobulin of the recipient along endothelial surfaces of graft blood vessels (12,14,28,30,35,36). The earliest structural change is the aggregation of platelets in small blood vessels (37). Platelet aggregation may be caused by various factors including the interaction of platelets with matrix exposed by the action of complement on endothelium and the direct effects of complement on platelets (38). Our concept of the pathophysiology of hyperacute rejection is that it involves the loss of barrier and anticoagulant functions of endothelial cells (39). The loss of these functions allows the escape of vascular contents from blood vessels and the attachment and aggregation of platelets leading to formation of platelet thrombi. The sections that follow will discuss the components of the immune system that contribute to the development of hyperacute rejection and the mechanism through which these components might bring about hyperacute rejection.
Xenoreactive Natural Antibodies All species of higher vertebrates have natural antibodies. Natural antibodies are synthesized without a known history of sensitization (40,41). Some natural antibodies are “polyreactive” as they bind to multiple target antigens. Other natural antibodies recognize blood groups (40,42,43), still others bind to the surface of xenogeneic cells (44). Xenoreactive antibodies, as such, appear to be related to some antiblood group antibodies and are distinct from polyreactive antibodies (9,45). Xenoreactive natural antibodies trigger hyperacute xenograft rejection of porcine-to-primate xenografts (25,46). The concept that natural antibodies initiate xenograft rejection has dominated the field of xenotransplantation for 30 yr (25,47). Perper and
10
Platt
Najarian suggested that cytotoxic natural antibodies specific for the cells of disparate species initiate hyperacute xenograft rejection (25). This concept was subsequently supported by several observations. First, naturally occurring antibodies directed against donor cells can be found in the serum of all mammalian species (40,41). Second, xenoreactive antibodies are rapidly deposited in xenografted organs (14,26). Third, depletion of xenoreactive antibodies prevents hyperacute rejection of xenografts (14,25,28). Fourth, hyperacute rejection does not occur if the recipient is newborn and lacks circulating xenoreactive antibodies (48). Fifth, hyperacute rejection can be induced by the administration of xenoreactive antibodies to a xenograft recipient (35). A major advance in the field of xenotransplantation came with the determination that xenoreactive natural antibodies in humans predominantly recognize one epitope, Gal_1-3Gal (49–52). Gal_1-3Gal was studied extensively by Galili (53,54) who showed that the sugar is expressed on the cells of New World monkeys and lower mammals but not on the cells of humans, apes or baboons (Table 3) (51). Humans, apes, and baboons, which do not express that sugar, have natural antibodies specific for it (55). Most individuals have 5–40 µg IgM per mL of plasma directed against Gal_1-3Gal (56). The amount of IgG specific for Gal_1-3Gal varies from 0 to 20 µg/mL of plasma (45). Parker has shown that the functional properties and concentration in serum of xenoreactive natural antibodies specific for Gal_1-3Gal is very similar to that of antibodies against blood group A and B antigens and, on this basis, has proposed that the xenoreactive natural antibodies and the antiblood group A and B antibodies may be members of a common family of natural antibodies (9,45,56). Evidence that Gal_1-3Gal might be important in xenotransplantation was first suggested by Good et al., who found that the binding of human antibodies to porcine cells could be blocked by purified Gal_1-3Gal but not by unrelated structures (49). Neethling et al. (52) found that antibodies directed against Gal_1-3Gal can be eluted from porcine organs perfused by human plasma. Sandrin et al. (50) demonstrated that transfection of COS cells with the murine _1,3-galactosyltransferase gene, which catalyzes the addition of _Gal residues to Gal`14GlcNAc-R to yield Gal_1-3Gal`1-4GlcNAc-R, induces binding of human natural antibodies to the transfected cells. Collins et al. (51) demonstrated that enzymatic removal of _-galactose residues from porcine cells abrogates binding of xenoreactive antibodies to the treated cells and that expression of Gal_1-3Gal provides a sufficient basis for
Hurdles to Xenotransplantation
11
the development of hyperacute xenograft rejection when the heart of a New World monkey is transplanted into a baboon that has antibodies specific for Gal_1-3Gal (57). Sachs et al. (58,59) showed that depletion of Gal_1-3Gal from the blood of baboons prevents the hyperacute rejection of porcine organ xenografts. Of natural anti-Gal_1-3Gal antibodies, it is IgM rather than IgG that appears to be of greatest significance in initiating the rejection of porcine organs by nonhuman primate recipients. Although both human natural IgM and IgG anti-Gal_1-3Gal bind to porcine cells (44), it is the binding of IgM and not of IgG that predominately initiates complement activation (15,56,60). In fact, xenoreactive natural IgG often consists largely of IgG2, which inhibits complement activation by blocking the binding of IgM (61). Consistent with this concept is the observation that human IgG given as gammaglobulin to a xenograft recipient does not promote hyperacute rejection but rather prevents it by diverting reactive complement proteins away from the graft endothelium (62). Recent work by Kearns-Jonker et al. has shown that antibodies may be restricted to the IGHV 3-11 and IGHV 3-74 germline progenitors in humans exposed to bioartificial livers containing porcine hepatocytes (63). The conditions that allow antiGal_1-3Gal antibodies to bind to that epitope are complex. Parker (56), Holzknecht (64), and Cotterell (65) with the author have shown that expression of the Gal_1-3Gal epitope by itself may not be sufficient to allow significant binding of complement-fixing xenoreactive antibodies. Rather, these antibodies seem to bind preferentially to certain glycoproteins bearing that epitope. This finding emerged from analysis of the mechanisms determining the avidity of natural IgM antibodies for the Gal_1-3Gal on cell surfaces (56,64). One important factor in antibody binding appears to be the clustering of epitopes allowing multivalent interactions between antibody molecules and the cell surface. Multivalent interactions are necessary because the KD for monomeric interactions is as low as 10–4 M whereas the effective avidity for intact IgM is characterized by a KD of approx 10–10 M. The optimal clustering of epitopes is not random, however, as the binding of xenoreactive antibodies to purified proteins containing similar numbers of Gal_1-3Gal substitutions varies over a range of 1000-fold (64). Further evidence that the manner in which Gal_1-3Gal is expressed rather than the total number of epitopes dictates the extent of antibody binding emerged from studies on variations in antigen expression by the population of potential donors (66). Analysis of the level of binding of xenoreactive antibodies to cells from populations of pigs suggests that
12
Platt
there is up to a 10-fold range of antigen expression (66,67). The variation appears to have a genetic basis, although it is not a simple Mendelian trait. The implications of this observation were tested by Cotterell and Alvarado (65). They found that porcine organs with low levels of antigen expression absorb significantly less xenoreactive natural antibodies during perfusion with baboon blood than organs with a high level of antigen expression. Cotterell and Alvarado also found that while binding of human IgM to porcine cells varies over a range of nearly 10-fold, that range is independent of the total expression of Gal_1-3Gal (65), suggesting that the way in which that sugar is expressed rather than its absolute level of expression determines antibody binding. The core structures bearing the Gal_1-3Gal modifications recognized by xenoreactive natural antibodies include proteins of the integrin family and von Willebrand factor (64,68). The attachment of antibodies to endothelial cell integrins could disrupt endothelial integrity by hindering the ability of the integrins to contribute to cell–cell interactions (69). Antibody binding to endothelial cell integrins may also deliver signals to the cells (70,71), potentially contributing to endothelial cell activation, which, as discussed in a later section of this chapter, is thought to be important in acute vascular xenograft rejection. The binding of antibodies to endothelial-cell-associated von Willebrand factor might potentially alter the interactions between endothelial cells and platelets (72–74). Based on current understanding of the specificity of xenoreactive antibodies, it is possible to devise specific strategies for depletion of xenoreactive antibodies from the circulation of xenograft recipients. Affinity columns have been used previously for depleting isohemagglutinins, allowing transplantation of organs across ABO barriers (75), and it is reasonable to think that this approach could also be used for depleting xenoreactive antibodies. Such columns bear Gal_1-3Gal and have been used effectively in experimental animals (58,76). Another way to prevent graft injury initiated by xenoreactive natural antibodies is to inhibit their binding using soluble ligands. This approach has been used to prevent rejection of ABO-incompatible kidney transplants (77,78). Unfortunately, because the binding of xenoreactive antibodies to cell surfaces is very avid, high concentrations of a monomeric inhibitor would be needed. Nevertheless, such approaches are being developed (79,80). Yet another approach to preventing the interaction of xenoreactive antibodies with a xenograft is to seek out or develop pigs that have low
Hurdles to Xenotransplantation
13
levels of antigen expression. With the identification and cloning of the gene for the glycosyltransferase responsible for the synthesis of Gal_13Gal (81,82), the possibility of genetically engineering donor animals with decreased expression has been proposed. The recent cloning of pigs by nuclear transfer (83) and the demonstration that these approaches might allow gene targeting (84), raise the possibility that _1,3galactosyltransferase might be “knocked out.” Indeed, _1,3galactosyltransferase knock-out pigs are anticipated to be tested within the next 12–18 mo. An alternative strategy proposed by Sandrin et al. (85) and by Sharma et al. (86) involves the introduction of another glycosyltransferase, _1,2-fucosyltransferase, which would compete with _1,3galactosyltransferase for adding the terminal saccharide onto oligosaccharide chains. Transgenic mice and pigs expressing the H transferase have been made and found to have decreased expression of Gal_1-3Gal. Yet another strategy might involve the expression of _galactosidase in transgenic animals (87). As another approach to obtaining donors with low levels of antigen, there is the possibility of exploiting the natural variation in antigen expression as previously mentioned. Geller (66) and Cotterell (65) found that expression of lower levels of antigen by some pigs has a genetic basis. Perfusion of organs from low antigen-expressing animals with the blood of baboons leads to the deposition of very little IgM and C4 in contrast to similar experiments in which organs from normal pigs are perfused.
Complement Activation of complement is an essential step in the development of hyperacute rejection. The importance of complement in xenograft rejection is suggested by three observations. First, serum complement levels decrease precipitously upon perfusion of a discordant xenograft with blood of the recipient (30,88). Second, complement accumulates rapidly in rejecting xenografts (14). Third, hyperacute rejection of xenografts is always prevented if complement is effectively inhibited by agents such as cobra venom factor (89–92), or if the graft is placed in a recipient that is congenitally deficient in complement component C6 (93,94). The role of complement in xenotransplantation has been reviewed recently (33,95–97). Given the importance of complement as a mediator of xenograft rejection, a critical question is how the complement system of the recipi-
14
Platt
ent becomes activated in the xenograft. One mechanism that may lead to complement activation is binding of complement-fixing antibodies to the graft. The importance of complement-fixing antibodies was originally proposed by Gewurz et al. (88). Perper and Najarian (25) suggested that there exists a relationship between the phylogenetic distance between the donor and recipient of a xenograft and the production of xenoreactive antibodies. Although the importance of xenoreactive antibodies in the activation of complement in xenografts has been questioned (27,98), there now exists compelling evidence supporting this mechanism in clinically relevant pig-to-primate xenograft models. First, rejecting xenografts can be shown to contain C4 co-localized with IgM (14). Second, depletion of IgM or C2, but not factor B of the alternative complement pathway, from a human serum prevents activation of complement when that serum is applied to porcine endothelial cells (13,15,56,60,64,68). Third, activation of complement on porcine cells during exposure to a human serum is inhibited by C1 inhibitor (99). Fourth, depletion of xenoreactive antibodies under conditions in which the complement system remains intact prevents hyperacute rejection of pig-to-primate xenograft (14). Fifth, pig hearts transplanted into newborn baboons who have an intact complement system but low levels of xenoreactive antibodies do not undergo hyperacute rejection (48). On the other hand, there are experimental models in which hyperacute rejection can occur in the apparent absence of anti-donor antibodies (100), suggesting that components of natural immunity other than xenoreactive natural antibodies may mediate immune recognition (Table 3). Such mechanisms might include activation of the alternative pathway or foreign cell surfaces and attachment to cells or proteins other than natural antibodies capable of activating complement. The alternative complement pathway is activated on some xenogeneic cell surfaces (101). Miyagawa showed that activation of the alternative pathway in a guinea pig-to-rat xenograft model could trigger hyperacute rejection (11). Activation of the alternative complement pathway probably also initiates the rejection of porcine organs transplanted into dogs, a point of historic interest, as much of the early work in xenotransplantation was carried out in pig-to-dog renal xenografts (25,26). Evidence implicating the alternative complement pathway in these models includes: 1. That the recipient was found to have low or undetectable levels of antidonor antibodies (11,98). 2. That further depletion of those antibodies did not prevent hyperacute rejection (12).
Hurdles to Xenotransplantation
15
3. That the alternative, but not classical, pathway activity in the recipient’s plasma decreased after the graft was reperfused (102).
Hyperacute rejection initiated by activation of the alternative complement pathway is particularly fulminant, perhaps because the formation of C3 convertase complexes proceeds rapidly and diffusely and does not depend on the kinetics of antibody-antigen interaction. The possibility also remains that activation of the alternative complement pathway by xenoreactive antibodies will be found to be important in some experimental models.
Complement Regulation Xenografts may be especially susceptible to complement-mediated injury because of the ineffective regulation of complement activation. Under physiologic conditions, the activation of complement on autologous cells is controlled in part by cell-associated glycoproteins such as decay accelerating factor (DAF;CD55) and membrane co-factor protein (MCP; CD46), which inhibit complement activation at the C3 convertase step, and CD59, which prevents formation of the membrane attack complex (103). These proteins have a limited ability to control activation of heterologous complement (10) and, as a result, a xenograft might be especially susceptible to complement-mediated injury (11,16,104). The potential impact of a defect in complement regulation on endothelial cell surfaces is suggested by the work of Matsuo et al. (105), who found that inhibition of decay accelerating factor by blocking antibodies in a rat model causes development of prominent vascular changes. It is possible that the species specificity of complement regulatory proteins depends on the system in which specificity is tested. For example, Miyagawa (106) found that complement regulation is conditioned by other components of the cell surface of the target cell. Van den Berg and Morgan (107) found that CD59 and perhaps other complement regulatory proteins may function normally between species in some model systems. In any case, because hearts from transgenic pigs expressing even low levels of human decay accelerating factor and human CD59 resist the development of hyperacute rejection in primates (108,109), control of complement would appear to be an important element of the definition of discordance and, thus, the immune barrier to xenotransplantation (96). In addition to the control of complement by cell membrane-associated proteins, there is the potential role of immunoglobulin as a regula-
16
Platt
tor of complement. Frank, Fries, and Basta showed that immunoglobulin may serve as an alternative acceptor for C3 and C4 (110), diverting the activated components away from target cell surfaces. Consistent with this concept, Magee et al. (90) showed that administration of gammaglobulin to primate recipients of porcine cardiac xenografts prevents the development of hyperacute rejection. The immunopathology of the xenografts in the treated recipients revealed host immunoglobulins and C1q along endothelial surfaces but little or no evidence of other complement components.
Components of Complement Involved in the Pathogenesis of Hyperacute Rejection Another important issue is which components of complement actually mediate tissue injury in rejecting xenografts (111). Two lines of evidence suggest that the occurrence of hyperacute rejection depend on assembly of terminal complement complexes. First, hyperacute xenograft rejection does not occur in recipients inherently deficient of C6 (93,94,112). Second, inhibition of complement using anti-C5 antibodies prevents some of the features of hyperacute rejection (113). Although terminal complement complexes are necessary for the development of hyperacute rejection, the formation of the membrane attack complex may not be essential. We recently found that organs from transgenic pigs expressing human CD59 significantly resist assembly of the membrane attack complex and some aspects of tissue injury; however, the development of hyperacute rejection is not averted (114). These apparently contradictory observations may be explained by the observation C5b67 complexes disrupt the integrity of endothelial monolayers (115), as discussed later in the text. If assembly of terminal complexes is absolutely necessary for the development of hyperacute rejection to occur, that does not exclude the possibility that other active components of complement may influence the well being and function of an organ during the hours following transplantation. Attachment of C3b to graft endothelium provides a ligand for neutrophils bearing complement receptors (116) and other cells that might amplify the impact of ischemia-reperfusion injury. Formation of C3a and C5a may contribute to this process by activating phagocytic cells. C5a may also mediate the release of heparan sulfate from endothelial cells (117), a process which may deprive endothelial cells of a number of physiological functions (13), as discussed later in the text.
Hurdles to Xenotransplantation
17
Natural Killer Cells Some recent studies have suggested the possibility that natural killer cells might accumulate in and damage a vascularized xenograft. The accumulation of natural killer cells in the graft and activation of those cells might be enhanced by xenoreactive antibodies deposited in the graft (118). Furthermore, because natural killer cells are ordinarily controlled by receptors specific for MHC class I molecules, the cells might fail to recognize the disparate MHC class I molecules expressed in a xenograft and thus be particularly active against a xenograft (119). Although natural killer cells have not been detected in large numbers in grafts undergoing hyperacute xenograft rejection (12,92,120), these cells have been seen in acute vascular rejection (121), and it would seem reasonable to think they may play a role in the evolution of this lesion, to be discussed presently.
Pathogenesis of Hyperacute Rejection Hyperacute rejection is perhaps best understood as a condition in which the critical functions of endothelial cells are lost (39). It is as though the blood vessels had lost endothelium. The loss of barrier functions allows the rapid formation of interstitial hemorrhage; the interaction between platelets and the extravascular compartment promotes thrombosis. Loss of the protective property of endothelium allows injury by oxidants, complement, and, perhaps, other inflammatory agents. For a more-detailed consideration of this subject, the reader is referred to a recent review (122).
Complement-Mediated Cytotoxicity One mechanism that might account for loss of endothelial function would involve complement-mediated lysis of endothelial cells. Lysis of endothelial cells would allow the egress of blood cells and fluid from blood vessels and would expose the underlying matrix that contains components capable of initiating platelet aggregation and thrombosis. Endothelial destruction is occasionally seen in emerging hyperacute rejection lesions, particularly in the more aggressive models such as the guinea pig-to-rat. However, endothelial cell death is not a major finding early in the course of hyperacute rejection in the pig-to-primate xenografts (14,32,33). Thus, complement-mediated changes probably involve noncytotoxic processes.
18
Platt
Release of Heparan Sulfate One noncytotoxic event that may contribute to the pathogenesis of hyperacute rejection is loss of heparan sulfate from endothelium. Heparan sulfate is an acidic polysaccharide that is involved in many of the physiologic functions of blood vessels. Heparan sulfate contributes to the endothelial barrier function, to anticoagulation, and to protection against complement and oxidants (123,124). Hence, changes in the metabolism of heparan sulfate might account for some of the events associated with xenograft rejection (13). Consistent with this possibility, exposure of endothelial cells to human serum was found to cause release of up to 50% of endothelial cell-associated heparan sulfate within 30–60 min (13). Release of heparan sulfate from endothelial cells appears to be caused by C5a (117) and involves the activation of endothelial cell-associated proteases (125). Stevens et al. showed that approx 50% of biosynthetically labeled heparan sulfate is lost within 5 min of the reperfusion of vascularized xenografts, and manipulations that inhibit loss of heparan sulfate also prolong the survival of xenografts (126). Other inflammatory events may cause the release of heparan sulfate from endothelial cells, perhaps accounting for other inflammatory or immune-mediated changes in blood vessels. For example, Key et al. (127) showed that elastase released from stimulated neutrophils causes release of endothelial cell heparan sulfate, and Magee et al. (128) showed that these process may be amplified in the presence of oxidants.
Complement-Mediated Changes in Endothelial Cell Shape and the Formation of Intercellular Gaps Another noncytotoxic mechanism that might bring about the loss of endothelial cell function is a change in endothelial cell shape with corresponding disruption of cell–cell attachments. Saadi et al. (115) showed that activation of complement on endothelial cells causes a dramatic change in endothelial cell shape leading to formation of intercellular “gaps.” The formation of gaps could be mediated by C5b67 (or C5b6), although the membrane attack complex accelerates the process. The formation of gaps is associated with increased production of second messenger metabolites. For example, like other changes in the barrier function and morphology of endothelial cells, complement-mediated alteration in cell shape is associated with changes in the level of cAMP (129,130). Adding dibutyryl-cAMP, forskolin, an activator of adenylate cyclase, or methylisobutyl xanthine, an inhibitor of cyclic nucleotide phosphodiesterase, to the medium bathing endothelial cells
Hurdles to Xenotransplantation
19
prevents gap formation in response to antibody binding and complement activation. Consistent with the potential importance of this mechanism, the ultrastructure of endothelial cells in xenograft tissues obtained 5 min after reperfusion with recipient blood reveals the interposition of platelets between adjacent endothelial cells.
Other Complement-Mediated Changes Activation of complement on endothelial cell surfaces causes a variety of other changes, which may contribute to the manifestations of hyperacute xenograft rejection. For example, formation of the membrane attack complex on endothelial cell surfaces triggers the secretion of von Willebrand factor (131,132), which might trigger platelet aggregation and the formation and release of cell membrane vesicles that might promote coagulation. In addition to recruiting platelets, the secretion of von Willebrand factor may contribute to the overall humoral reaction in xenografts as it can serve as a target of complement-fixing human natural antibodies (64). The membrane attack complex also changes alteration to the surface properties of endothelial cells so as to promote assembly of prothrombinase complexes (133). Stimulation of endothelial cells and platelets causes expression of P-selectin, which can serve as a ligand for neutrophils and platelets (134,135). These events and perhaps the generation of small amounts of thrombin may account for the observation that platelet aggregation is the earliest morphologic change observed in hyperacute rejection. In addition to causing the formation of thrombi and obstruction to blood flow, the aggregation and activation of platelets may have a direct impact on the functions of endothelial cells (37). Tissue damage in hyperacute xenograft rejection may be amplified by reperfusion injury. Reperfusion injury is thought to involve local activation complement and to lead to endothelial cell dysfunction through various mechanisms, including the generation of toxic oxygen species such as superoxide anion and the release of proteolytic enzymes. These mechanisms alter endothelial cell structure and function, potentially rendering the xenograft more susceptible to injury by natural antibodies and complement. Consistent with this idea, we have recently demonstrated that oxidant stress increases the susceptibility of porcine endothelial cells to the noncytotoxic effects of human natural antibody and complement (128) and that providing the second messenger, cAMP, which inhibits some manifestations of preservation injury (136), also prevents some of the effects of complement on endothelium (137). These
20
Platt
studies provide preliminary evidence supporting the idea that reperfusion injury, even at a level too low to yield dysfunction of an isograft or allograft, may contribute to tissue injury in a xenograft. It should be mentioned, however, that there are some published reports suggesting that prolonged preservation can improve early xenograft function (138). One potential explanation for this may be that under some conditions, preservation may function like ischemic preconditioning.
Acute Vascular Rejection If hyperacute rejection is prevented, or, in the case of concordant xenografts, hyperacute rejection does not occur, a discordant xenograft is subject to another type of rejection, which we have called “acute vascular xenograft rejection” (12,139–141). Acute vascular rejection is characterized pathologically by endothelial injury and swelling, ischemia, and thrombosis. An infiltrate consisting of mononuclear leukocytes and neutrophils is often observed. Acute vascular rejection is sometimes called “delayed” xenograft rejection; however, the later term may be misleading because it suggests incorrectly that this type of rejection is a delayed manifestation of hyperacute rejection. That acute vascular rejection is not, indeed, a delayed form of hyperacute rejection is suggested by several lines of evidence. First, acute vascular rejection typically occurs when the complement system of a xenograft recipient is inhibited, a condition that always prevents the occurrence of hyperacute rejection (12,33). Second, acute vascular rejection occurs in complement-deficient animals that are unable to mount an hyperacute rejection response (94). Third, treatments such as the administration of antileukocyte antibodies that suppress acute vascular rejection by inhibiting inflammatory cells have no impact on hyperacute rejection (142). On the other hand, administration of gammaglobulin, which diverts complement away from the graft but increases antibody binding to the graft (90), prevents hyperacute rejection but not acute rejection and may even make the latter worse. Fourth, the most dramatic pathological manifestations of acute vascular rejection, especially endothelial swelling, focal ischemic injury, and diffuse fibrin thrombi are not typically seen in hyperacute rejection (12,39).
Role of Xenoreactive Antibodies in the Pathogenesis of Acute Vascular Rejection Several lines of evidence suggest that acute vascular xenograft rejection is caused by the continuing interaction of xenoreactive antibodies
Hurdles to Xenotransplantation
21
with the graft (96). First, during the days following the extracorporeal circulation of blood of patients with fulminant hepatic failure through porcine livers, the levels of anti-swine antibodies in the blood of the patients increases, and this increase is observed at a time that corresponds to the period during which a xenograft is subject to acute vascular rejection (62). Second, the removal of antibodies from the circulation of xenograft recipients and/or inhibition of antibody synthesis by treatment with cytotoxic agents delays or averts acute vascular rejection (58,143–146). Third, a type of rejection similar or identical to acute vascular rejection of xenografts is observed in allografts (147,148) and xenografts (149,150) in association with de novo appearance of antidonor antibodies. Fourth, acute vascular rejection of allografts and xenografts can be induced by the infusion of anti-donor antibodies (35).
The Role of Complement in Acute Vascular Rejection Whether complement contributes to the pathogenesis of acute vascular xenograft rejection and how complement might do so is uncertain. Clearly, the development of tissue lesions characteristic of acute vascular rejection is easily envisioned as a reflection of acute complementmediated changes in endothelial cell structure and function as previously discussed. However, a role for complement in acute vascular rejection is difficult to reconcile with the fact that acute vascular xenograft rejection is typically observed in xenograft recipients that had been depleted of complement (12). This difficulty may be resolved by three considerations: 1. Even the most potent inhibitors of complement are incompletely effective. 2. Under conditions of inflammation, endothelial cells synthesize complement components. 3. Some endothelial changes, such as the alteration in heparan sulfate metabolism, can be induced by very low levels of complement.
As examples of the third mechanism, less than 1% of the complement activity in normal serum is sufficient to mediate the release of heparan sulfate from porcine endothelial cells (13,125), and less than 10% of normal complement activity is sufficient to stimulate procoagulant and proinflammatory changes in endothelial cells (151). If complement contributes to the pathogenesis of acute vascular rejection, the membrane attack complex may not be essential as acute vascular rejection develops in recipients deficient in C6 (93,94).
22
Platt
Endothelial Cell Activation Based on the histopathology of acute vascular rejection, which includes infiltration of the organ by neutrophils, the thickening of endothelial cells, and pronounced thrombotic changes (12), the author proposed that the pathogenesis of acute vascular rejection might involve the activation of endothelial cells and thus the acquisition of new endothelial functions (39). This idea was supported by the observation that endothelial cells in grafts undergoing acute vascular rejection display phenotypic changes characteristic of activation (121). One new function presumably acquired by endothelium in acute vascular rejection is the ability to promote coagulation. Unlike quiescent endothelium, which prevents coagulation by serving as a barrier between underlying matrix and plasma coagulation proteins, the endothelium in rejecting xenografts allows coagulation proteins to come into contact with underlying matrix. In addition, the formation of the membrane attack complex on endothelial cell surfaces stimulates the formation of IL1_, which, in turn, stimulates the de novo synthesis of tissue factor (151) and proinflammtory products (39). Other changes such as loss of thrombomodulin; increased expression of E selectin (152), chemokines (153), prostaglandins (154); synthesis of plasminogen activator inhibitor type 1; and changes as previously discussed, such as the alteration of cell surface promoting the formation of prothrombinase complexes, may also contribute to the procoagulant posture. In addition, the production of small amounts of thrombin, platelet activating factor, or thomboxane A2 may stimulate platelets in the vicinity of the endothelium.
The Role of Platelets in Acute Vascular Rejection Although the thrombi in acute vascular rejection consist mainly of fibrin, the fibrin is invariably admixed with platelets. Bustos recently tested the idea that platelets, activated by small amounts of thrombin, might activate endothelial cells (37). Activated human platelets were found to express IL-1_ on the surface, which together with secreted cytokines could activate endothelium. The relative importance of platelets and complement in triggering endothelial cell activation, and thus acute vascular rejection, remains uncertain.
The Role of Inflammatory Cells in Acute Vascular Rejection Blakely et al. (121) suggested that monocytes, which are known to express tissue factor, a co-factor for the formation of prothrombinase complexes, might cause the fibrin deposition and thus the tissue mani-
Hurdles to Xenotransplantation
23
festations of acute vascular rejection. This view was supported by the observation that acute vascular rejection in rodents is associated with the influx of macrophages expressing tissue factor. On the other hand, the lesions early in the course of acute vascular rejection do not reveal significant numbers of macrophages (90,155) and the earliest expression of tissue factor is on graft endothelium (156). Nor is the onset or the pathology of acute vascular rejection more than modestly inhibited by the administration of agents that inhibit the interaction of inflammatory cells with xenogeneic endothelium (142), in contrast to the results achieved by antibody depletion, which significantly prolongs the survival of the transplants (155). Nor, conversely, do allotransplants with cellular rejection, which are often found to have large numbers of invading macrophages, necessarily exhibit lesions typical of acute vascular rejection (157). Thus, whereas a role for macrophages cannot be excluded at this juncture, it seems more likely that these cells are a marker for tissue injury rather than the cause in acute vascular rejection of xenotransplants. Yet another mechanism that might contribute to the pathogenesis of acute vascular rejection is the action of natural killer cells on graft endothelium. The potential involvement of natural killer cells has been of special interest, because the cells might be expected to be activated in xenograft recipients through stimulation carbohydrate and Fc receptors and failure of stimulation of MHC class I receptors (118,158). Consistent with this concept, human natural killer cells have been found to exert cytotoxic and noncytotoxic effects, such as induction of procoagulant activity on porcine endothelial cells in vitro (159,160). Although natural killer cells and other lymphocytes have been found in some rodent xenografts (161), this finding is not always observed (12), as they are not major components of pig-to-primate xenografts (90,155). Therefore, although lymphoid cells may well contribute to tissue injury in acute vascular rejection, they may not be essential for the manifestation of tissue lesions.
Accommodation When anti-donor antibodies are depleted from a xenograft recipient, acute vascular xenograft rejection may not occur, even after the antibodies return to the circulation. This condition, in which the graft seems to resist acute vascular rejection despite presence in the circulation of all inciting factors, is called “accommodation” (16). We first observed accommodation in ABO-incompatible kidney allografts (77,162)
24
Platt
and postulated that a similar phenomenon might occur in vascularized xenografts (14,16,163). Accommodation may also be observed in graft recipients with circulating anti-donor HLA antibodies (164,165). We have postulated that accommodation may arise through one or more of three mechanisms (14,16). First, there is the possibility that the anti-donor antibodies may change in their functional properties or specificity or both. Supporting this possibility are studies in ABO-incompatible allografts in which anti-donor antibodies were detected in the circulation but were generally not observed in biopsies of the transplanted organ (77,162). Another potential change in the anti-donor antibodies is toward predominance of IgG2, which activates complement poorly and might compete with complement fixing antibodies for binding to target cells (61). Second, there might occur a change in the antigen. This concept is supported by the finding that carbohydrate synthesis in the kidney changes following transplantation (166) and the observation that some saccharide antigens are modulated by antibody binding (167). Third, accommodation might involve a change in the endothelium so that the graft becomes inured to antibody binding and complement activation. The possibility that an allograft or xenograft might acquire resistance to tissue injury and rejection is supported by recent studies showing that accommodation in a rodent model is associated with expression of “protective” genes by endothelial cells (168,169) and by observations that with continued stimulation of endothelial cells by antibodies and/or complement, the sensitivity of those cells to injury decreases (170–172). As an example of one potential scenario, continued stimulation of endothelial cells with endotoxin or with IL-1 causes the cells to develop resistance to restimulation (173). Not only may the sensitivity to restimulation decrease, but the sensitivity to injury may also decrease. For example, stimulation of endothelial cells may cause an increased synthesis of decay-accelerating factor (174), which inhibits complement activation at the level of C3 convertase. Regardless of the biological basis for accommodation, further understanding of how this process is induced could be very useful, if not critical, to the clinical application to xenotransplantation. Not only might the occurrence of accommodation eliminate the need for continuing depletion of anti-donor antibodies from a graft recipient, it also might provide clues to manipulations of the donor or recipient that might prevent more chronic forms of graft rejection. The study of accom-
Hurdles to Xenotransplantation
25
modation may also yield insights into how blood vessels might be manipulated to alter sensitivity to other injurious processes.
Therapeutic Approaches to Acute Vascular Rejection Three general strategies might be envisioned for preventing acute vascular rejection. The first strategy involves the induction of accommodation. As discussed above, accommodation might be induced by temporary depletion of xenoreactive antibodies. Studies in the author’s laboratory have indicated that accommodation of pig organs transplanted into nonhuman primates can be achieved by depletion of all Ig (155) or of anti-Gal_1,3Gal antibodies (59,175). The second approach to preventing acute vascular rejection might involve induction of immunological tolerance, leading to decreased production of anti-xenograft antibodies. Humoral tolerance has been achieved in _1,3galactosyl transferase knock-out mice (these mice make anti-Gal_1,3Gal antibodies like humans) (176) by the transplantation of normal murine bone marrow to yield mixed chimerism and autologous cells transduced with _1,3galactosl transferase (177). These approaches do seem capable of preventing rejection of Gal_1,3Gal+ murine hearts by Gal_1-3Gal recipients. Whether the strategy would be effective between disparate species is still unknown owing to the difficulties of transplanting bone marrow between species. It is hoped, however, that treatment with porcine cytokines might allow the enduring engraftment of porcine bone marrow (5). A third strategy for preventing acute vascular rejection involves decreasing or eliminating expression of Gal_1,3Gal in the transplant. To the extent that Gal_1,3Gal is the major target of the xenoreactive antibodies that cause acute vascular rejection, as the author’s work suggests (51), this strategy would have the theoretical advantage of limiting the immunosuppressive treatments that would have to be applied to the recipient. Expression of H transferase, which catalyzes production of H antigen, to which humans are tolerant, in lieu of Gal_1,3Gal, has been applied with some success in pigs (85). However, studies by Parker et al. suggest that to effectively prevent acute vascular rejection, Gal_1,3Gal expression would have to be decreased by greater than 95% (178). The recent cloning of pigs by transfer of nuclei from cultured somatic cells to enucleated zygotes (83) raises the possibility the _1,3galactosyl transferase could be “knocked out” and thus expression of Gal_1-3Gal
26
Platt
could be eliminated. Although knocking out _1,3galactosyltransferase would seem to be preferred, it is possible that the elimination of Gal_1,3Gal might subject the recipient to increased risk of infection by porcine viruses such as the porcine endogenous retrovirus (PERV), which would otherwise be neutralized by anti-Gal_1,3Gal antibodies (179). Hence, it is not as yet clear whether the optimal strategy would be to eliminate that saccharide. A more salient question may be whether other elicited antibodies will emerge that will induce acute vascular rejection in the absence of Gal_1-3Gal.
XENOGENEIC CELLULAR IMMUNE RESPONSES The initial type of rejection observed in cellular or free tissue xenografts is acute cellular rejection. If hyperacute and acute vascular rejection are averted, an organ xenograft would next be subject to rejection by cell mediated immunity of the host against the donor (18,58,180). The central questions regarding cellular rejection of xenografts are the extent to which the cellular immune response to a xenograft might differ from the cellular immune response to an allograft and whether this response would be more severe and less subject to immune modulation than the corresponding allogeneic response. Preliminary answers to these questions have not been forthcoming because, until recently, it has been difficult to bring about survival of organ xenografts for more than a few days without the administration of massive doses of immunosuppressive agents. The studies of Alexandre (143) involving the transplantation of porcine kidneys into baboons do provide some evidence for the importance of T cells in as much as decreases in graft function were reversed by administration of immunosuppressive agents. Also preliminary, but nonetheless intriguing, are studies by Leventhal and co-workers demonstrating that rejection of guinea pig-to-rat cardiac xenografts could be hastened by transfer of lymphocytes from presensitized animals do likewise (181). Studies of the cellular immune response to xenografts of free tissues such as the islets of Langerhans or the skin have provided some insight into the similarities and potential differences between the cellular immune responses to xenografts and allografts. Similar to allografts, islet and skin xenografts are subject to cellular rejection, because they rapidly fail if the recipient is a normal mouse but survive if the recipient is a nude or a SCID mouse. In some cases, however, the failure of a free tissue xenograft occurs notably earlier than the failure of a free
Hurdles to Xenotransplantation
27
tissue allograft. This difference may reflect a greater susceptibility of the xenograft to “primary nonfunction.” What is not yet clear is whether the early failure of the xenografts was caused by the immune response of the recipient as in primary nonfunction of allografts (21) or whether, at least to some extent, the failure was caused by incompatibility of the graft with the foreign microenvironment. As one example of the latter, neovascularization of a free tissue xenograft may be impaired in comparison to neovascularization of a free tissue allograft. Such impairment could result in the xenograft being deprived of nutrition and the ready removal of waste products. Impaired neovascularization could also influence the way in which foreign antigens are presented in a xenograft (Fig. 2). On the other hand, free tissue xenografts may survive for extended periods of time in recipients treated with immunosuppression that is not more intense than the immunosuppression used for allografts (182,183). Moreover, anti-CD4, which has limited effectiveness in allotransplantation, may be especially effective in preventing xenograft rejection (20). Some recent in vitro observations have shed light on the potential differences between xenogeneic and allogeneic cellular immune responses. Proliferative responses to xenogeneic cells in mixed leukocyte cultures are often less exuberant than proliferative responses to allogeneic cells (184). This decreased proliferative response to xenogeneic stimulation was thought to reflect the smaller repertoire of T cells that might recognize xenogeneic cells directly, a reflection in turn of positive selection leading to more cross reactivity with allogeneic than with xenogeneic MHC. Indeed, the dampened in vitro response has led to the suggestion that cellular immunity to a xenograft might be less intense than cellular immunity to an allograft (18,184). This idea was further supported by studies suggesting that recognition of xenogeneic cells occurs predominantly through the indirect pathway, the T cells of the host being specific for peptides of xenogeneic origin complexed with MHC on host antigen-presenting cells. Because indirect recognition would lead to activation of a smaller fraction of the T-cell repertoire, the response arising in this way would be less intense and less rapid than the response to allostimulation arising through the direct pathway. Xenogeneic T-cell responses on the indirect pathway also reflect defects in the ability of T cells to respond to xenogeneic antigen presenting cells brought about by “incompatibility” of CD8 or CD4 with xenogeneic MHC leading to decreased co-receptor function (185,186) and by incompatibility of co-stimulatory pathways and cytokines (187).
28
Platt
Recent studies have shown, however, that mouse and human T cells can recognize xenogeneic cells, such as porcine endothelial cells, directly and that under the proper conditions xenoimmune responses arising in vitro may be as vigorous as alloimmune responses (188–191), although there is some evidence of a limited frequency of T-cell precursors with the ability to recognize directly highly disparate MHC (186). There are reasons to think, however, that cellular responses to xenografts might be as strong or stronger than cellular responses to allografts. First, virtually all foreign proteins contain amino acid sequences that differentiate one species from another. Thus, a xenotransplant, in contrast to an allotransplant, might give rise to a vast array of foreign peptides that could stimulate a strong response even though the response is restricted to the indirect pathway. This concept is supported by the work of Dorling et al. (192), who demonstrated that human T cells respond to primary stimulation by porcine peptides, whereas primary responses to allogeneic peptides cannot be detected. There must exist, then, a much higher frequency of human T cells committed to respond to porcine peptides, than to human peptides. Second, to the extent that the direct recognition contributes to immunoregulatory responses, defects in the direct pathway, such as those previously described, might lead in vivo to impaired immunoregulation and thus to stronger cellular immune response. Third, the humoral response to xenotransplantation, as it occurs in vivo, may give rise to increases in expression of cell adhesion molecules or release of proinflammatory mediators that could in turn amplify cellular immune responses (151). As one example, heparan sulfate released from endothelial cells by antiendothelial cell antibodies and complement (13,124,125) may activate antigen-presenting cells leading to increased ability to stimulate proliferative (193–195) and cytolytic T-cell responses (196). In addition to depriving endothelium of the function of heparan sulfate, the release of that molecule may have other consequences. Wrenshall et al. found that heparan sulfate in soluble form activates antigen-presenting cells leading to amplified proliferative (194) and cytolytic (196) T-cell responses. These changes are caused by the direct action of heparan sulfate on antigen-presenting cells leading to stimulation of several immunomodulatory pathways (193,197). Another aspect of the cellular immune response to xenogeneic cells that remains to be elucidated is the extent to which the response will be directed against MHC versus other proteins. Human cellular immune
Hurdles to Xenotransplantation
29
responses generated against porcine cells in vitro can be directed against porcine MHC (191). Whether the porcine MHC would be the major target of in vivo responses remains to be determined. Also uncertain is whether the response will be subject to the kinds of regulation that impact on alloimmune responses. Another question of potential import is the nature of the humoral immune response that would be elicited by a xenograft and how such a response might impact elicited cellular immunity. Already discussed is the observation that when a xenograft recipient is treated with immunosuppressive therapy the antibodies in the serum of the recipient following xenotransplantation may still recognize Gal_1-3Gal as the predominant target (62). On the other hand, it seems likely that antibodies against other donor antigens, particularly polypeptides, are likely to arise and pose an additional hurdle, perhaps in the form of acute or chronic vascular rejection. This question is clearly an important one for future investigation. Whether cellular immune responses to xenotransplants can be controlled by the same therapeutic approaches used to control cellular immune responses to allotransplants is a question of obvious practical import. Specific advice on this matter cannot be offered at the present time except to consider the apparently increased efficacy of anti-CD4 antibodies in preventing rejection of cellular xenotransplants (vide supra). Some have suggested that the cellular immune response to xenotransplants would be so intense that induction of tolerance would be necessary to prevent rejection (58). If tolerance is required to prevent cellular rejection of xenotransplants, it is uncertain which approach to induction and tolerance would be most effective. Studies in rodents have shown that bone marrow transplantation lending to mixed chimerism (198,199) and thymus transplantation in conjunction with bone marrow transplantation (200,201) might be effective.
CHRONIC REJECTION There is the possibility that elicited anti-donor antibodies will contribute to more indolent or chronic types of rejection of organ xenografts, which is chronic rejection (202). Although the importance of these responses for xenografts of all types might seem intuitive, there has been much emphasis in recent years on the contribution of the donor organ to chronic rejection. If donor factor is of greatest importance in the pathogenesis of chronic rejection, one might postulate that the xenograft
30
Platt
would be less susceptible to chronic changes. In any case, xenografts might offer a useful model system for sorting out donor vs recipient factors contributing to chronic lesions (203).
MOLECULAR INCOMPATIBILITIES This chapter has focused on the molecular and cellular facets of the immune response that contribute to the barrier to xenotransplantation. However, some measure of the xenogeneic immune response may be owed to incompatibilities of immune and inflammatory cascades. Examples of such incompatibilities, complement control and T-cell co-stimulation, have been given above. The author has speculated that these incompatibilities may not simply be random variation but may reflect adaptations (203). Molecular incompatibilities between species may also pose a physiologic barrier to xenotransplantation. Gritsch et al. (4) describe incompatibilities of growth factors that hinder bone marrow engraftment between species. Hammer (204) describes molecular incompatibilities as a severe barrier to xenotransplantation of the liver. Bach et al. (161) and we (96,97) have speculated that molecular incompatibilities could underlie the diffuse coagulation seen in acute vascular rejection. Ierino et al. (205) suggest that there may exist a propensity for coagulation in some conditions. The importance of these molecular barriers is still uncertain. It is possible that accommodation will eventually be seen as a compensatory state that overcomes these barriers (172,206).
OTHER HURDLES TO XENOTRANSPLANTATION As if the intense and diverse responses to xenotransplantation were not enough, two other hurdles warrant brief comment. One other hurdle is the physiology of the graft—the possibility that a graft might fail or fail to function because of incompatibility with the recipient. At present, the question of physiology would appear to be more theoretical than a real hurdle, as recent studies have shown that the heart, kidney, and lungs can function in highly disparate recipients. In the end, the greatest physiologic hurdle may prove to be impairment in graft function caused by xenoimmune responses. One other hurdle worthy of comment is that of infection. There is a possibility of transferring infections from the graft to the xenogeneic recipient, that is zoonosis. The main agent of concern at present is the PERV. PERV is a type C retrovirus that can infect human lymphoid cells in culture (207,208). Thus far, infection of humans has not been
Hurdles to Xenotransplantation
31
observed (209,210). Of more immediate concern is the possibility that immunosuppression or tolerance regimens might render the recipient relatively immunoincompetent and thus severely subject to infection. Still another concern is that viral infections of xenotransplants might not be adequately controlled by cell-mediated immune responses because viral peptides would be presented by xenogeneic MHC and, thus, relatively unrecognized by the T-cell repertoire of the recipient. Despite all of the hurdles discussed above, there is increasing enthusiasm about the prospects for clinical xenotransplantation. In part, this enthusiasm springs from the urgent need for donor organs and a sense of willingness to consider xenotransplantation as a rational approach to that problem. In part, this enthusiasm reflects recent success in the genetic engineering and cloning of pigs as potential xenotransplant donor and from the long-term survival of experimental xenografts performed using organs from these pigs (145). Indeed, some of these results would seem to suggest that there are no fundamental incompatibilities that would stand in the way of enduring survival of a xenograft. Whether or not the enthusiasm for xenotransplantation proves to be well founded, progress in the field of transplantation provides dramatic evidence of how the approach to medical problems is changing. As molecular hurdles to xenotransplantation are discovered, those hurdles are being addressed through the rational design of drugs and the genetic engineering of donor animals. Thus, if it is not possible to predict that xenotransplantation will enter the clinical arena in a few years, it does seem certain that knowledge in this area will continue to advance rapidly and, in doing so, that knowledge and rationale strategies that evolve will increasingly benefit the broader fields of immunology and clinical medicine.
REFERENCES 1. Evans RW, Orians CE, Ascher NL. The potential supply of organ donors: an assessment of the efficiency of organ procurement efforts in the United States. JAMA 1992; 267:239–246. 2. Starzl TE, Fung J, Tzakis A, et al. Baboon-to-human liver transplantation. Lancet 1993; 341:65–71. 3. Platt JL. New directions for organ transplantation. Nature 1998; 392 (Suppl):11–17. 4. Gritsch HA, Glaser RM, Emery DW, et al. The importance of nonimmune factors in reconstitution by discordant xenogeneic hematopoietic cells. Transplantation 1994; 57:906–917. 5. Yang Y-G, Chen AM, Garrett LJ, et al. Development and analysis of transgenic mice expressing porcine hematopoietic cytokines: a model for achieving durable porcine hematopoietic chimerism across an extensive xenogeneic barrier. Xenotransplantation 2000; 7:58–64.
32
Platt
6. Platt JL. Xenotransplanting hepatocytes: the triumph of a cup half full. Nat Med 1997; 3:26–27. 7. Gunsalus JR, Brady DA, Coulter SM, Gray BM, Edge A. Reduction of serum cholesterol in watanabe rabbits by xenogeneic hepatocellular transplantation. Nat Med 1997; 3:48–53. 8. Shoskes DA, Wood KJ. Indirect presentation of MHC antigens in transplantation. Immunol Today 1994; 15:32–38. 9. Parker W, Yu PB, Holzknecht ZE, Lundberg-Swanson K, Buckley RH, Platt JL. Specificity and function of “natural” antibodies in immunodeficient subjects: clues to B-cell lineage and development. J Clin Immunol 1997; 17:311–321. 10. Atkinson JP, Oglesby TJ, White D, Adams EA, Liszewski MK. Separation of self from non-self in the complement system: a role for membrane cofactor protein and decay accelerating factor. Clin Exp Immunol 1991; 86 (Supp. 1):27–30. 11. Miyagawa S, Hirose H, Shirakura R, et al. The mechanism of discordant xenograft rejection. Transplantation 1988; 46:825–830. 12. Leventhal JR, Matas AJ, Sun LH, et al. The immunopathology of cardiac xenograft rejection in the guinea pig-to-rat model. Transplantation 1993; 56:1–8. 13. Platt JL, Vercellotti GM, Lindman BJ, Oegema TR, Jr., Bach FH, Dalmasso AP. Release of heparan sulfate from endothelial cells: implications for the pathogenesis of hyperacute rejection. J Exp Med 1990; 171:1363–1368. 14. Platt JL, Fischel RJ, Matas AJ, Reif SA, Bolman RM, Bach FH. Immunopathology of hyperacute xenograft rejection in a swine-to-primate model. Transplantation 1991; 52:214–220. 15. Dalmasso AP, Vercellotti GM, Fischel RJ, Bolman RM, Bach FH, Platt JL. Mechanism of complement activation in the hyperacute rejection of porcine organs transplanted into primate recipients. Am J Pathol 1992; 140:1157–1166. 16. Platt JL, Vercellotti GM, Dalmasso AP, et al. Transplantation of discordant xenografts: a review of progress. Immunol Today 1990; 11:450–456. 17. Lawson JH, Daniels L, Platt JL. The evaluation of thrombomodulin activity in porcine to human xenotransplantation. Transplant Proc 1997; 29:884–885. 18. Auchincloss H Jr. Xenogeneic transplantation. Transplantation 1988; 46:1–20. 19. Egdahl RH, Good RA, Varco RL. Studies in homograft and heterograft survival. Surgery 1957; 42:228–239. 20. Pierson RN, Winn HJ, Russell PS, Auchincloss H. Xenogeneic skin graft rejection is especially dependent on CD4+ T cells. J Exp Med 1989; 170:991–996. 21. Kaufman DB, Platt JL, Rabe FL, Stock PG, Sutherland DER. The immunological basis of islet allograft primary non-function. J Exp Med 1990; 172:291–302. 22. Winn HJ, Baldamus CA, Jooste SV, Russell PS. Acute destruction by humoral antibody of rat skin grafted to mice: the role of complement and polymorphonuclear leukocytes. J Exp Med 1973; 137:893–910. 23. Baldamus CA, McKenzie IF, Winn HJ, Russell PS. Acute destruction by humoral antibody of rat skin grafted to mice. J Immunol 1973; 110:1532–1541. 24. Jooste SV, Colvin RB, Winn HJ. The vascular bed as the primary target in the destruction of skin grafts by antiserum. J Exp Med 1981; 154:1332–1341. 25. Perper RJ, Najarian JS. Experimental renal heterotransplantation. I. In widely divergent species. Transplantation 1966; 4:377–388. 26. Giles GR, Boehmig HJ, Lilly J, et al. Mechanism and modification of rejection of heterografts between divergent species. Transplant Proc 1970; 2:522–537.
Hurdles to Xenotransplantation
33
27. Calne RY. Organ transplantation between widely disparate species. Transplant Proc 1970; 2:550–556. 28. Rose AG, Cooper DKC, Human PA, Reichenspurner H, Reichart B. Histopathology of hyperacute rejection of the heart: experimental and clinical observations in allografts and xenografts. J Heart Lung Transplant 1991; 10:223–234. 29. Forbes RDC, Kuramochi T, Guttmann RD, Klassen J, Knaack J. A controlled sequential morphologic study of hyperacute cardiac allograft rejection in the rat. Lab Inv 1975; 33:280–288. 30. Busch GJ, Martins ACP, Hollenberg NK, Wilson RE, Colman RW. A primate model of hyperacute renal allograft rejection. Am J Pathol 1975; 79:31–56. 31. Colman RB, Habal M, Hollenberg NK, Birtch AG, Busch GJ. Hyperacute renal allograft rejection in the primate. Am J Pathol 1976; 82:25–42. 32. Platt JL. Xenotransplantation. In: Wood KJ, Morris P, Raddy K, eds. Handbook of Transplant Immunology. London, Medical & Scientific Publications, 1995, 259–299. 33. Platt JL. Hyperacute xenograft rejection. Medical Intelligence Unit. Austin: R.G. Landes, 1995. 34. Platt JL, Bach FH. The barrier to xenotransplantation. Transplantation 1991; 52: 937–947. 35. Perper RJ, Najarian JS. Experimental renal heterotransplantation. III. Passive transfer of transplantation immunity. Transplantation 1967; 5:514–533. 36. Larsen S, Starklint H. Histopathology of kidney xenograft rejection. In: Cooper DKC, Kemp E, Reemtsma K, White DJG, eds. Xenotransplantation. the transplantation of organs and tissues between species. New York: Springer-Verlag, 1991,181–206. 37. Bustos M, Platt JL. Platelet-endothelial cell interactions in a xenograft model. Transplant Proc 1997; 29:886. 38. Bustos M, Saadi S, Platt JL. Platelet-mediated activation of endothelial cells: implications for the pathogenesis of transplant rejection. Transplantation 2001; 72:509–515. 39. Saadi S, Ihrcke NS, Platt JL. Pathophysiology of xenograft rejection. In: Lieberman R, Morris R, eds. Principles of immunomodulatory drug development in transplantation and autoimmunity. New York: Raven Press, 1996,31–45. 40. Boyden SV. Natural antibodies and the immune response. Adv Immunol 1964; 5:1–28. 41. Hammer C. Preformed natural antibodies (PNAB) and possibilities of modulation of hyperacute xenogeneic rejection (HXAR). Transplant Proc 1989; 21: 522–523. 42. Parker W, Platt JL. Xenoreactive human natural antibodies. In: Peter JB, Schoenfeld Y, eds. Autoantibodies. Amsterdam: Elsevier Science B.V., 1996,846–851. 43. Parker W, Bruno D, Platt J. Xenoreactive natural antibodies in the world of natural antibodies: typical or unique? Transpl Immunol 1995; 3:181–191. 44. Platt JL, Turman MA, Noreen HJ, Fischel RJ, Bolman RM, Bach FH. An ELISA assay for xenoreactive natural antibodies. Transplantation 1990; 49:1000–1001. 45. Parker W, Lundberg-Swanson K, Holzknecht ZE, et al. Isohemagglutinins and xenoreactive antibodies are members of a distinct family of natural antibodies. Hum Immunol 1996; 45:94–104.
34
Platt
46. Gewurz H, Clark DS, Finstad J, et al. Role of the complement system in graft rejections in experimental animals and man. Ann N Y Acad Sci 1966; 129: 673–713. 47. Clark DS, Gewurz H, Good RA, Varco RL. Complement fixation during homograft rejection. Surg Forum 1964; 15:144–146. 48. Kaplon RJ, Michler RE, Xu H, Kwiatkowski PA, Edwards NM, Platt JL. Absence of hyperacute rejection in newborn pig-to-baboon cardiac xenografts. Transplantation 1994; 59:1–6. 49. Good AH, Cooper DKC, Malcolm AJ, et al. Identification of carbohydrate structures that bind human antiporcine antibodies: implications for discordant xenografting in humans. Transplant Proc 1992; 24:559–562. 50. Sandrin MS, Vaughan HA, Dabkowski PL, McKenzie IFC. Anti-pig IgM antibodies in human serum react predominantly with Gal_(1,3)Gal epitopes. Proc Natl Acad Sci 1993; 90:11,391–11,395. 51. Collins BH, Parker W, Platt JL. Characterization of porcine endothelial cell determinants recognized by human natural antibodies. Xenotransplantation 1994; 1:36–46. 52. Neethling FA, Koren E, Ye Y, et al. Protection of pig kidney (PK15) cells from the cytotoxic effect of anti-pig antibodies by _-galactosyl oligosaccharides. Transplantation 1994; 57:959–963. 53. Galili U, Rachmilewitz EA, Peleg A, Flechner I. A unique natural human IgG antibody with anti-_-galactosyl specificity. J Exp Med 1984; 160:1519–1531. 54. Galili U. Interaction of the natural anti-Gal antibody with _-galactosyl epitopes: a major obstacle for xenotransplantation in humans. Immunol Today 1993; 14: 480–482. 55. Galili U, Clark MR, Shohet SB, Buehler J, Macher BA. Evolutionary relationship between the natural anti-Gal antibody and the Gal _1-3Gal epitope in primates. Proc Natl Acad Sci USA 1987; 84:1369–1373. 56. Parker W, Bruno D, Holzknecht ZE, Platt JL. Characterization and affinity isolation of xenoreactive human natural antibodies. J Immunol 1994; 153:3791–3803. 57. Collins BH, Cotterell AH, McCurry KR, et al. Hyperacute rejection of cardiac xenografts between primate species: evidence to support the significance of the _-Galactosyl determinant. J Immunol 1995; 154:5500–5510. 58. Sachs DH, Sablinski T. Tolerance across discordant xenogeneic barriers. Xenotransplantation 1995; 2:234–239. 59. Kozlowski T, Shimizu A, Lambrigts D, et al. Porcine kidney and heart transplantation in baboons undergoing a tolerance induction regimen and antibody adsorption. Transplantation 1999; 67:18–30. 60. Platt JL, Lindman BJ, Geller RL, et al. The role of natural antibodies in the activation of xenogenic endothelial cells. Transplantation 1991; 52:1037–1043. 61. Yu PB, Holzknecht ZE, Bruno D, Parker W, Platt JL. Modulation of natural IgM binding and complement activation by natural IgG antibodies. J Immunol 1996; 157:5163–5168. 62. Cotterell AH, Collins BH, Parker W, Harland RC, Platt JL. The humoral immune response in humans following cross-perfusion of porcine organs. Transplantation 1995; 60:861–868. 63. Kearns-Jonker M, Swensson J, Ghiuzeli C, et al. The human antibody response to porcine xenoantigens is encoded by IGHV3-11 and IGHV3-74 IgVH germline progenitors. J Immunol 1999; 163:4399–4412.
Hurdles to Xenotransplantation
35
64. Platt JL, Holzknecht ZE. Porcine platelet antigens recognized by human xenoreactive natural antibodies. Transplantation 1994; 57:327–335. 65. Cotterell AH, Alvarado CG, Logan JS, Platt JL. Variation in expression of porcine antigens recognized by human xenoreactive natural antibodies. Transplant Proc 1995; 27:278–279. 66. Geller RL, Rubinstein P, Platt JL. Variation in expression of porcine xenogeneic antigens. Transplantation 1994; 58:272–277. 67. Alvarado CG, Cotterell AH, McCurry KR, et al. Variation in the level of xenoantigen expression in porcine organs. Transplantation 1995; 59:1589–1596. 68. Holzknecht ZE, Platt JL. Identification of porcine endothelial cell membrane antigens recognized by human xenoreactive antibodies. J Immunol 1995; 154:4565–4575. 69. Lampugnani MG, Resnati M, Dejana E, Marchisio PC. The role of integrins in the maintenance of endothelial monolayer integrity. J Cell Biol 1991; 112:479–490. 70. Kornberg LJ, Earp HS, Turner CE, Prockop C, Juliano RL. Signal transduction by integrins: increased protein tyrosine phosphorylation caused by clustering of b1 integrins. Proc Natl Acad Sci USA 1991; 88:8392–8396. 71. Hynes RO. Integrins: versatility, modulation, and signaling in cell adhesion. Cell 1992; 69:11–25. 72. Baumgartner HR, Tschopp TB, Meyer D. Shear rate dependent inhibition of platelet adhesion and aggregation on collagenous surfaces by antibodies to human factor VIII/von Willebrand factor. Br J Haematol 1980; 44:127–139. 73. Turitto VT, Weiss HJ, Zimmerman TS, Sussman II. Factor VIII/von Willebrand factor in subendothelium mediates platelet adhesion. Blood 1985; 65:823–831. 74. Roth GJ. Platelets and blood vessels: the adhesion event. Immunol Today 1992; 13:100–105. 75. Bannett AD, McAlack RF, Raja R, Baquero A, Morris M. Experiences with known ABO-mismatched renal transplants. Transplant Proc 1987; 19:4543–4546. 76. Lin SS, Kooyman DL, Daniels LJ, et al. The role of natural anti-Gal_1-3Gal antibodies in hyperacute rejection of pig-to-baboon cardiac xenotransplants. Transpl Immunol 1997; 5:212–218. 77. Chopek MW, Simmons RL, Platt JL. ABO-incompatible renal transplantation: initial immunopathologic evaluation. Transplant Proc 1987; 19:4553–4557. 78. Cooper DKC, Ye Y, Niekrasz M, et al. Specific intravenous carbohydrate therapy: a new concept in inhibiting antibody-mediated rejection—experience with ABO-incompatible cardiac allografting in the baboon. Transplantation 1993; 56:769–777. 79. Ye Y, Neethling FA, Niekrasz M, et al. Evidence that intravenously administered _-Galactosyl carbohydrates reduce baboon serum cytotoxicity to pig kidney cells (PK15) and transplanted pig hearts. Transplantation 1994; 58:330–337. 80. Rieben R, Bovin NV, Korchagina EY, et al. Xenotransplantation. In vitro analysis of synthetic alpha-galactosyl inhibitors of human anti-Galalpha1—>3Gal IgM and IgG antibodies. Glycobiology 2000; 10:141–148. 81. Dabkowski PL, Vaughan HA, McKenzie IF, Sandrin MS. Isolation of a cDNA clone encoding the pig alpha 1,3galactosyltransferase. Transplant Proc 1994; 26:1335. 82. Strahan K, Gu F, Andersson L, Gustafsson K. Pig _1,3 galactosyltransferase: sequence of a full length cDNA clone and chromosomal localisation of the corresponding gene. Transplant Proc 1995; 27:245–248.
36
Platt
83. Onishi A, Iwamoto M, Akita T, et al Pig cloning by microinjection of fetal fibroblast nuclei. Science 2000; 289:1188–1190. 84. McCreath KJ, Howcroft J, Campbell KHS, Colman A, Schnieke AE, Kind AJ. Production of gene-targeted sheep by nuclear transfer from cultured somatic cells. Nature 2000; 405:1066–1069. 85. Sandrin MS, Fodor WL, Mouhtouris E, et al. Enzymatic remodelling of the carbohydrate surface of a xenogeneic cell substantially reduces human antibody binding and complement-mediated cytolysis. Nat Med 1995; 1:1261–1267. 86. Sharma A, Okabe JF, Birch P, Platt JL, Logan JS. Reduction in the level of Gal (_1,3) Gal in transgenic mice and pigs by the expression of an _(1,2) fucosyltransferase. Proc Natl Acad Sci 1996; 93:7190–7195. 87. Osman N, McKenzie IF, Ostenried K, Ioannou YA, Desnick RJ, Sandrin MS. Combined transgenic expression of alpha-galactosidase and alpha1,2fucosyltransferase leads to optimal reduction in the major xenoepitope Galalpha(1,3)Gal. Proc Natl Acad Sci 1997; 94:14,677–14,682. 88. Gewurtz H, Clark DS, Finstad J, Kelly WD, Varco RL. Role of the complement system in graft rejections in experimental animals and man. Ann N Y Acad Sci 1966; 129:673–713. 89. Pruitt SK, Kirk AD, Bollinger RR, et al. The effect of soluble complement receptor type 1 on hyperacute rejection of porcine xenografts. Transplantation 1994; 57: 363–370. 90. Magee JC, Collins BH, Harland RC, et al. Immunoglobulin prevents complement mediated hyperacute rejection in swine-to-primate xenotransplantation. J Clin Invest 1995; 96:2404–2412. 91. Gewurz H, Clark DS, Cooper MD, Varco RL, Good RA. Effect of cobra venominduced inhibition of complement activity on allograft and xenograft rejection reactions. Transplantation 1967; 5:1296–1303. 92. Leventhal JR, Dalmasso AP, Cromwell JW, et al. Prolongation of cardiac xenograft survival by depletion of complement. Transplantation 1993; 55:857–866. 93. Chartrand C, O’Regan C, Robitaille P, Pinto-Blonde M. Delayed rejection of cardiac xenografts in C6-deficient rabbits. Immunology 1979; 38:245–248. 94. Brauer RB, Baldwin III WM, Daha MR, Pruitt SK, Sanfilippo F. Use of C6-deficient rats to evaluate the mechanism of hyperacute rejection of discordant cardiac xenografts. J Immunol 1993; 151:7240–7248. 95. Platt JL, Logan JS. The generation and use of transgenic animals for xenotransplantation. In Houdebine LM, ed. Transgenic animals: generation and use. Amsterdam, Harwood Academic Publishers, 1997,455–460. 96. Parker W, Saadi S, Lin SS, Holzknecht ZE, Bustos M, Platt JL. Transplantation of discordant xenografts: a challenge revisited. Immunol Today 1996; 17:373–378. 97. Lawson JH, Platt JL. Molecular barriers to xenotransplantation. Transplantation 1996; 62:303–310. 98. Johnston P, Lim S, Wright L, White D. Hyperacute xenograft rejection takes place in the complete absence of antibodies but requires the presence of functional complement. Transplant Proc 1991; 23:877–884. 99. Dalmasso AP, Platt JL. Prevention of complement-mediated activation of xenogeneic endothelial cells in an in vitro model of xenograft hyperacute rejection by C1 inhibitor. Transplantation 1993; 56:1171–1176. 100. Johnston PS, Wang M-W, Lim SML, Wright LJ, White DJG. Discordant xenograft rejection in an antibody-free model. Transplantation 1992; 54:573–576.
Hurdles to Xenotransplantation
37
101. Edwards J. Complement activation by xenogeneic red blood cells. Transplantation 1981; 31:226–227. 102. Muller-Eberhard HJ. Molecular organization and function of the complement system. Annu Rev Biochem 1988; 57:321–347. 103. Hourcade D, Holers VM, Atkinson JP. The regulators of complement activation (RCA) gene cluster. Adv Immunol 1989; 45:381–416. 104. Dalmasso AP, Vercellotti GM, Platt JL, Bach FH. Inhibition of complementmediated endothelial cell cytotoxicity by decay accelerating factor: potential for prevention of xenograft hyperacute rejection. Transplantation 1991; 52:530–533. 105. Matsuo S, Ichida S, Takizawa H, et al. In vivo effects of monoclonal antibodies that functionally inhibit complement regulatory proteins in rats. J Exp Med 1994; 180:1619–1627. 106. Miyagawa S, Shirakura R, Matsumiya G, et al. Test for ability of decay-accelerating factor (DAF, CD55) and CD59 to alleviate complement-mediated damage of xeno-erythrocytes. Scand J Immunol 1993; 38:37–44. 107. Van den Berg CW, Morgan BP. Complement-inhibiting activities of human CD59 and analogues from rat, sheep, and pig are not homologously restricted. J Immunol 1994; 152:4095–4101. 108. McCurry KR, Kooyman DL, Alvarado CG, et al. Human complement regulatory proteins protect swine-to-primate cardiac xenografts from humoral injury. Nat Med 1995; 1:423–427. 109. Byrne GW, McCurry KR, Martin MJ, McClellan SM, Platt JL, Logan JS. Transgenic pigs expressing human CD59 and decay-accelerating factor produce an intrinsic barrier to complement-mediated damage. Transplantation 1997; 63:149–155. 110. Basta M, Kirshbom P, Frank MM, Fries LF. Mechanism of therapeutic effect on high-dose intravenous immunoglobulin: attenuation of acute, complement-dependent immune damage in a guinea pig model. J Clin Invest 1989; 84:1974–1981. 111. Platt JL, Saadi S. The role of complement in transplantation. Mol Immunol 1999; 36:965–971. 112. Zhow XJ, Niesen N, Pawlowski I, et al. Prolongation of survival of discordant kidney xenografts by C6 deficiency. Transplantation 1990; 50:896–898. 113. Kroshus TJ, Rollins SA, Dalmasso AP, et al. Complement inhibition with an antiC5 monoclonal antibody prevents acute cardiac tissue injury in an ex vivo model of pig-to-human xenotransplantation. Transplantation 1995; 60:1194–1202. 114. Diamond LE, McCurry KR, Oldham ER, et al. Characterization of transgenic pigs expressing functionally active human CD59 on cardiac endothelium. Transplantation 1996; 61:1241–1249. 115. Saadi S, Platt JL. Transient perturbation of endothelial integrity induced by natural antibodies and complement. J Exp Med 1995; 181:21–31. 116. Vercellotti GM, Platt JL, Bach FH, Dalmasso AP. Neutrophil adhesion to xenogeneic endothelium via iC3b. J Immunol 1991; 146:730–734. 117. Platt JL, Dalmasso AP, Lindman BJ, Ihrcke NS, Bach FH. The role of C5a and antibody in the release of heparan sulfate from endothelial cells. Eur J Immunol 1991; 21:2887–2890. 118. Inverardi L, Samaja M, Motterlini R, Mangili F, Bender JR, Pardi R. Early recognition of a discordant xenogeneic organ by human circulating lymphocytes. J Immunol 1992; 149:1416–1423. 119. Lanier LL, Phillips JH. Inhibitory MHC class I receptors on NK cells and T cells. Immunol Today 1996; 17:86–91.
38
Platt
120. Collins BH, Chari RS, Magee JC, et al. Mechanisms of injury in porcine livers perfused with blood of humans with fulminant hepatic failure. Transplantation 1994; 58:1162–1171. 121. Blakely ML, Van Der Werf WJ, Berndt MC, Dalmasso AP, Bach FH, Hancock WW. Activation of intragraft endothelial and mononuclear cells during discordant xenograft rejection. Transplantation 1994; 58:1059–1066. 122. Saadi S, Platt JL. Endothelial cell responses to complement activation. In Volanakis JE, Frank MM, eds The human complement system in health and disease. New York, Marcel Dekker, Inc., 1998, 335–353. 123. Platt JL. Mechanisms of tissue injury in hyperacute xenograft rejection. ASAIO J 1992; 38:8–16. 124. Ihrcke NS, Wrenshall LE, Lindman BJ, Platt JL. Role of heparan sulfate in immune system-blood vessel interactions. Immunol Today 1993; 14:500–505. 125. Ihrcke NS, Platt JL. Shedding of heparan sulfate proteoglycan by stimulated endothelial cells: evidence for proteolysis of cell surface molecules. J Cell Physiol 1996; 168:625–637. 126. Stevens RB, Wang YL, Kaji H, et al. Administration of nonanticoagulant heparin inhibits the loss of glycosaminoglycans from xenogeneic cardiac grafts and prolongs graft survival. Transplant Proc 1993; 25:382. 127. Key NS, Platt JL, Vercellotti GM. Vascular endothelial cell proteoglycans are susceptible to cleavage by neutrophils. Arterio Thromb 1992; 12:836–842. 128. Magee JC, Platt JL, Oldham KT, Guice KS. Oxidant stress increases suceptibility of porcine endothelial cells to injury by xenoreactive antibodies and complement. Transplant Proc 1994; 26:1170. 129. Stelzner TJ, Weil JV, O’Brien RF. Role of cyclic adenosine monophosphate in the induction of endothelial barrier properties. J Cell Physiol 1989; 139:157–166. 130. Naito M, Hayashi T, Kuzuya M, Funaki C, Asai K, Kuzuya F. Vascular endothelial cell migration in vitro roles of cyclic nucleotides, calcium ion and cytoskeletal system. Artery 1989; 17:21–31. 131. Hattori R, Hamilton KK, Fugate RD, McEver RP, Sims PJ. Stimulated secretion of endothelial von Willebrand factor is accompanied by rapid redistribution to the cell surface of the intracellular granule membrane protein GMP-140. J Biol Chem 1989; 264:7768–7771. 132. Hattori R, Hamilton KK, McEver RP, Sims PJ. Complement proteins C5b-9 induce secretion of high molecular weight multimers of endothelial von Willebrand factor and translocation of granule membrane protein GMP-140 to the cell surface. J Biol Chem 1989; 264:9053–9060. 133. Hamilton KK, Hattori R, Esmon CT, Sims PJ. Complement proteins C5b-9 induce vesiculation of the endothelial plasma membrane and expose catalytic surface for assembly of the prothrombinase enzyme complex. J Biol Chem 1990; 265:3809–3814. 134. McEver RP, Beckstead JH, Moore KL, Marshall-Carlson L, Bainton DF. GMP140, a platelet alpha-granule membrane protein, is also synthesized by vascular endothelial cells and is localized in Weibel-Palade bodies. J Clin Invest 1989; 84:92–99. 135. Geng JG, Bevilacqua MP, Moore KL, et al. Rapid neutrophil adhesion to activated endothelium mediated by GMP-140. Nature 1990; 343:757–760.
Hurdles to Xenotransplantation
39
136. Pinsky D, Oz M, Liao H, et al. Restoration of the cAMP second messenger pathway enhances cardiac preservation for transplantation in a heterotopic rat model. J Clin Invest 1993; 92:2994–3002. 137. Saadi S, Ihrcke NS, Platt JL. Endothelial cell shape and hyperacute rejection. Transplant Proc 1994; 26:1149. 138. Loss M, Kunz R, Przemeck M, et al. Influence of cold ischemia time, pretransplant anti-porcine antibodies and donor/recipient size matching on hyperacute graft rejection following discordant porcine to cynomolgus kidney transplantation. Transplantation 2000; 69:1155–1159. 139. Matas AJ, Sibley R, Mauer M, Sutherland DER, Simmons RL, Najarian JS. The value of needle renal allograft biopsy. Ann Surg 1983; 197:226–237. 140. Perper RJ, Najarian JS. Experimental renal heterotransplantation. II. Closely related species. Transplantation 1966; 4:700–711. 141. Halloran PF, Schlaut J, Solez K, Srinivasa NS. The significance of the anti-class I response. II. Clinical and pathologic features of renal transplants with anti-class I-like antibody. Transplantation 1992; 53:550–555. 142. Zehr KJ, Herskowitz A, Lee PC, Kumar P, Gillinov AM, Baumgartner WA. Neutrophil adhesion and complement inhibition prolongs survival of cardiac xenografts in discordant species. Transplantation 1994; 57:900–906. 143. Alexandre GPJ, Gianello P, Latinne D, et al. Plasmapheresis and splenectomy in experimental renal xenotransplantation. In Hardy MA, ed. Xenograft 25. New York, Elsevier Science Publishers, 1989; 259–266. 144. Leventhal JR, John R, Fryer JP, et al. Removal of baboon and human antiporcine IgG and IgM natural antibodies by immunoadsorption: results of in vitro and in vivo studies. Transplantation 1995; 59:294–300. 145. Thompson C. Humanised pigs hearts boost xenotransplantation. Lancet 1995; 346:766. 146. Lin SS, Platt JL. The role of immunoabsorption in xenotransplantation. In: Brunkhorst R, Koch KM, Koll R, eds. Klinische immunadsorption. Stuttgart, Wissenschaftliche Verlagsgesellschaft mbH, 2000,133–136. 147. Paul LC, Claas FHJ, van Es LA, Kalff MW, de Graeff J. Accelerated rejection of a renal allograft associated with pretransplantation antibodies directed against donor antigens on endothelium and monocytes. New Eng J Med 1979; 300:1258–1260. 148. Abbas AK, Corson JM, Carpenter CB, Galvanek EG, Merrill JP, Dammin GJ. Immunologic enhancement of rat renal allografts. Am J Pathol 1974; 75:271–284. 149. Hasan R, van den Bogaerde JB, Wallwork J, White DJG. Evidence that long-term survival of concordant xenografts is achieved by inhibition of antispecies antibody production. Transplantation 1992; 54:408–413. 150. Fujino Y, Kawamura T, Hullett DA, Sollinger HW. Evaluation of cyclosporine, mycophenolate mofetil, and brequinar sodium combination therapy on hamsterto-rat cardiac xenotransplantation. Transplantation 1994; 57:41–46. 151. Saadi S, Holzknecht RA, Patte CP, Stern DM, Platt JL. Complement-mediated regulation of tissue factor activity in endothelium. J Exp Med 1995; 182:1807–1814. 152. Saadi S, Holzknecht RA, Patte CP, Platt JL. Endothelial cell activation by pore forming structures: pivotal role for IL-1_. Circulation 2000; 101:1867–1873. 153. Selvan RS, Kapadia HB, Platt JL. Complement-induced expression of chemokine genes in endothelium: regulation by IL-1-dependent and -independent mechanisms. J Immunol 1998; 161:4388–4395.
40
Platt
154. Bustos M, Coffman TM, Saadi S, Platt JL. Modulation of eicosanoid metabolism in endothelial cells in a xenograft model: role of cyclooxygenase-2. J Clin Invest 1997; 100:1150–1158. 155. Lin SS, Weidner BC, Byrne GW, et al. The role of antibodies in acute vascular rejection of pig-to-baboon cardiac transplants. J Clin Invest 1998; 101:1745–1756. 156. Nagayasu T, Saadi S, Holzknecht RA, Patte CA, Plummer TB, Platt JL. Expression of tissue factor mRNA in cardiac xenografts: clues to the pathogenesis of acute vascular rejection. Transplantation 2000; 69:475–482. 157. Platt JL, LeBien TW, Michael AF. Interstitial mononuclear cell populations in renal graft rejection: Identification by monoclonal antibodies in tissue sections. J Exp Med 1982; 155:17–30. 158. Inverardi L, Clissi B, Stolzer AL, Bender JR, Pardi R. Overlapping recognition of xenogeneic carbohydrate ligands by human natural killer lymphocytes and natural antibodies. Transplant Proc 1996; 28:552. 159. Malyguine AM, Saadi S, Platt JL, Dawson JR. Human natural killer cells induce morphologic changes in porcine endothelial cell monolayers. Transplantation 1996; 61:161–164. 160. Malyguine AM, Saadi S, Holzknecht RA, et al. Induction of procoagulant function in porcine endothelial cells by human NK cells. J Immunol 1997; 159:4659–4664. 161. Bach FH, Winkler H, Ferran C, Hancock WW, Robson SC. Delayed xenograft rejection. Immunol Today 1996; 17:379–384. 162. Alexandre GPJ, Squifflet JP, De Bruyere M, et al. Present experiences in a series of 26 ABO-incompatible living donor renal allografts. Transplant Proc 1987; 19:4538–4542. 163. Platt JL, Lindman BJ, Bach FH. Natural antibody targets on discordant endothelium: molecular characterization and consequences of antibody binding. Transplant Proc 1991; 23:815–816. 164. Palmer A, Welsh K, Gjorstrup P, Taube D, Bewick M, Thick M. Removal of antiHLA antibodies by extracorporeal immunoadsorption to enable renal transplantation. Lancet 1989; 1:10–12. 165. Ross CN, Gaskin G, Gregor-MacGregor S, et al. Renal transplantation following immunoadsorption in highly sensitized recipients. Transplantation 1993; 55:785–789. 166. Ulfvin A, Backer AE, Clausen H, et al. Expression of glycolipid blood group antigens in single human kidneys: change in antigen expression of rejected ABO incompatible kidney grafts. Kidney Int 1993; 44:1289–1297. 167. Andres G, Yamaguchi N, Brett J, Caldwell PRB, Godman G, Stern D. Cellular mechanisms of adaptation of grafts to antibody. Transpl Immunol 1996; 4:1–17. 168. Bach FH, Ferran C, Hechenleitner P, et al. Accommodation of vascularized xenografts: expression of “protective genes” by donor endothelial cells in a host Th2 cytokine environment. Nat Med 1997; 3:196–204. 169. Bach FH, Hancock WW, Ferran C. Protective genes expressed in endothelial cells: a regulatory response to injury. Immunol Today 1997; 18:483–6. 170. Dalmasso AP, He T, Benson BA. Human IgM xenoreactive antibodies can induce resistance of porcine endothelial cells to complement-mediated injury. Xenotransplantation 1996; 3:54–62. 171. Delikouras A, Hayes M, Malde P, Lechler RI, Dorling A. Nitric oxide-mediated expression of Bcl- 2 and Bcl-xl and protection from TNF_-mediated apoptosis in porcine endothelial cells after exposure to low concentrations of xenoreactive natural antibody. Transplantation 2001; 71:599–605.
Hurdles to Xenotransplantation
41
172. Holzknecht ZE, Platt JL. Accommodation and the reversibility of biological systems. Transplantation 2001; 71:594–595. 173. Busso N, Huet S, Nicodeme E, Hiernaux J, Hyafil F. Refractory period phenomenon in the induction of tissue factor expression on endothelial cells. Blood 1991; 78:2027–2035. 174. Shibata T, Cosio FG, Birmingham DJ. Complement activation induces the expression of decay-accelerating factor on human mesangial cells. J Immunol 1991; 147:3901–3908. 175. Lin SS, Hanaway MJ, Gonzalez-Stawinski GV, et al. The role of anti-Gal_1-3Gal antibodies in acute vascular rejection and accommodation of xenografts. Transplantation (Rapid Communication) 2000; 70:1667–1674. 176. Thall AD, Maly P, Lowe JB. Oocyte Gal_1,3Gal epitopes implicated in sperm adhesion to the zona pellucida glycoprotein ZP3 are not required for fertilization in the mouse. The Journal of Biological Chemistry 1995; 270:21,437–21,440. 177. Yang Y-G, deGoma E, Ohdan H, Bracy JL, Xu Y, Iacomini J, Thall AD, Sykes M. Tolerization of anti-Gal_1-3Gal natural antibody-forming B cells by induction of mixed chimerism. J Exp Med 1998; 187:1335–1342. 178. Parker W, Lin SS, Platt JL. Antigen expression in xenotransplantation: how low must it go? Transplantation 2001; 71:313–319. 179. Rother RP, Fodor WL, Springhorn JP, Birks CW, Setter E, Sandrin MS, Squinto SP, Rollins SA. A novel mechanism of retrovirus inactivation in human serum mediated by anti-_-galactosyl natural antibody. J Exp Med 1995; 182:1345–1355. 180. Geller RL, Turman MA, Dalmasso AP, Platt JL. The natural immune barrier to xenotransplantation. J Am Soc Nephrol 1993; 3:1189–1200. 181. Fryer JP, Leventhal JR, Dalmasso AP, Chen S, Simone PA, Goswitz JJ, Reinsmoen NJ, Matas AJ. Beyond hyperacute rejection. Transplantation 1995; 59:171–176. ´ 182. Ricordi C, Lacy PE, Sterbenz K, Davie JM. Low-temperature culture of human islets or in vivo treatment with L3T4 antibody produces a marked prolongation of islet human-to-mouse xenograft survival. Proc Natl Acad Sci USA 1987; 84:8080–8084. 183. Marchetti P, Scharp DW, Finke EH, Swanson CJ, Olack BJ, Gerasimidi-Vazeou D, Giannarelli R, Navalesi R, Lacy PE. Prolonged survival of discordant porcine islet xenografts. Transplantation 1996; 61:1100–1102. 184. Alter BJ, Bach FH. Cellular basis of the proliferative response of human T cells to mouse xenoantigens. J Exp Med 1990; 191:333–338. 185. Irwin MJ, Heath WR, Sherman LA. Species-restricted interactions between CD8 and the a3 domain of class I influence the magnitude of the xenogeneic response. J Exp Med 1989; 170:1091–1101. 186. Batten P, Heaton T, Fuller-Espie S, Lechler RI. Human anti-mouse xenorecognition. J Immunol 1995; 155:1057–1065. 187. Moses RD, Winn HJ, Auchincloss Jr H. Multiple defects in cell surface molecule interactions across species differences are responsible for diminished xenogeneic T cell responses. Transplantation 1992; 53:203–209. 188. Murray AG, Khodadoust MM, Pober JS, Bothwell ALM. Porcine aortic endothelial cells activate human T cells: direct presentation of MHC antigens and costimulation by ligands for human CD2 and CD28. Immunity 1994; 1:57–63. 189. Kirk AD, Li RA, Kinch MS, Abernethy KA, Doyle C, Bollinger RR. The human antiporcine cellular repertoire. In vitro studies of acquired and innate cellular responsiveness. Transplantation 1993; 55:924–931.
42
Platt
190. Rollins SA, Kennedy SP, Chodera AJ, Elliott EA, Zavoico GB, Matis LA. Evidence that activation of human T cells by porcine endothelium involves direct recognition of porcine SLA and costimulation by porcine ligands for LFA-1 and CD2. Transplantation 1994; 57:1709–1716. 191. Yamada K, Sachs DH, DerSimonian H. Human anti-porcine xenogeneic T cell response. evidence for allelic specificity of mixed leukocyte reaction and for both direct and indirect pathways of recognition. J Immunol 1995; 155:5249–5256. 192. Dorling A, Lombardi G, Binns R, Lechler RI. Detection of primary direct and indirect human anti-porcine T cell responses using a porcine dendritic cell population. Eur J Immunol 1996; 26:1378–1387. 193. Wrenshall LE, Cerra FB, Singh RK, Platt JL. Heparan sulfate initiates signals in murine macrophages leading to divergent biological outcomes. J Immunol 1995; 154:871–880. 194. Wrenshall LE, Cerra FB, Carlson A, Bach FH, Platt JL. Regulation of murine splenocyte responses by heparan sulfate. J Immunol 1991; 147:455–459. 195. Kodaira Y, Wrenshall LE, Nair S, Gilboa E, Platt JL. Heparan sulfate modulates allogeneic T cell responses stimulated by dendritic cells. FASEB J 1998; 12:A582 (3380). 196. Wrenshall LE, Carlson A, Cerra FB, Platt JL. Modulation of cytolytic T cell responses by heparan sulfate. Transplantation 1994; 57:1087–1094. 197. Kodaira Y, Nair SK, Wrenshall LE, Gilboa E, Platt JL. Phenotypic and functional maturation of dendritic cells modulated by heparan sulfate. J Immunol 2000; 165:1599–1604. 198. Li H, Ricordi C, Demetris AJ, Kaufman CL, Korbanic C, Hronakes ML, Ildstad ST. Mixed xenogeneic chimerism (mouse+rat —>mouse) to induce donor-specific tolerance to sequential or simultaneous islet xenografts. Transplantation 1994; 57:592–598. 199. Robinson LA, Tu L, Steeber DA, Preis O, Platt JL, Tedder TF. The role of adhesion molecules in human leukocyte attachment to porcine vascular endothelium: Implications for xenotransplantation. J Immunol 1998; 161:6931–6938. 200. Zhao Y, Swenson K, Sergio JJ, Arn JS, Sachs DH, Sykes M. Skin graft tolerance across a discordant xenogeneic barrier. Nat Med 1996; 2:1211–1216. 201. Zhao Y, Fishman JA, Sergio JJ, Oliveros JL, Pearson DA, Szot GL, Wilkinson RA, Arn JS, Sachs DH, Sykes M. Immune restoration by fetal pig thymus grafts in T cell-depleted, thymectomized mice. J Immunol 1997; 158:1641–1649. 202. Hancock WW, Buelow R, Sayegh MH, Turka LA. Antibody-induced transplant arteriosclerosis is prevented by graft expression of anti-oxidant and anti-apoptotic genes. Nat Med 1998; 4:1392–1396. 203. Cascalho M, Platt JL. The immunological barrier to xenotransplantation. Immunity 2001; 14:437–446. 204. Hammer C. Physiological obstacles after xenotransplantation. Ann N Y Acad Sci 1998; 862:19–27. 205. Ierino FL, Kozlowski T, Siegel JB, et al. Disseminated intravascular coagulation in association with the delayed rejection of pig-to-baboon renal xenografts. Transplantation 1998; 66:1439–1450. 206. Holzknecht ZE, Platt JL. The fine cytokine line between graft acceptance and rejection. Nat Med 2000; 6:497–498.
Hurdles to Xenotransplantation
43
207. Martin U, Kiessig V, Blusch JH, et al. Expression of pig endogenous retrovirus by primary porcine endothelial cells and infection of human cells. Lancet 1998; 352:692–694. 208. Patience C, Takeuchi Y, Weiss RA. Infection of human cells by an endogenous retrovirus of pigs. Nat Med 1997; 3:282–286. 209. Paradis K, Langford G, Long Z, et al. Search for cross-species transmission of porcine endogenous retrovirus in patients treated with living pig tissue. Science 1999; 285:1236–1241. 210. Patience C, Patton GS, Takeuchi Y, et al. No evidence of pig DNA or retroviral infection in patients with short-term extracorporeal connection to pig kidneys. Lancet 1998; 352:699–701. 211. Rosengard BR, Adachi H, Ueda K, et al. Differences in the pathogenesis of firstset allograft rejection and acute xenograft rejection as determined by sequential morphologic analysis. J Heart Transpl 1986; 5:263–266. 212. Jamieson SW. Xenograft hyperacute rejection: a new model. Transplantation 1974; 17:533–534. 213. Cooper DKC, Human PA, Lexer G, et al. Effects of cyclosporine and antibody adsorption on pig cardiac xenograft survival in the baboon. J Heart Transpl 1988; 7:238–246. 214. Roslin MS, Tranbaugh RE, Panza A, et al. One-year monkey heart xenograft survival in cyclosporine-treated baboons. Transplantation 1992; 54:949–955.
vi
Pathological Responses to Xenotransplantation
2
45
Pathological Responses to Xenotransplantation Matilde Bustos, MD, PhD and Jeffrey L. Platt, MD, PhD INTRODUCTION
The idea of transplanting animal organs into patients with organ failure is not new. When the development of vascular anastomosis made organ transplantation feasible from a surgical perspective, a few clinical renal xenografts were attempted. In 1906, Jaboulay (1) described the xenotransplantation of pig and goat grafts into humans. Neither pig nor goat grafts functioned, and the failure of the xenograft did not allow vascular thrombosis to be observed. At the same time, Unger performed xenotransplantation using organs from nonhuman primates with similar results (2). In 1923, Harol Neuhof affirmed that thrombosis or hemorrhage in the xenotransplant could be prevented (2). However, technical imperfection and the lack of understanding of immunological host reactivity led to waning interest in xenotransplantation. The first reports of successful clinical xenotransplantation appeared in the literature as recently as 1960. The initial attempts were performed with monkeys and baboons as donors. Reemtsma et al. (3) utilized chimpanzees and Starzl et al. (4) reported use of a series of baboons-tohuman renal xenografts. These transplants did not suffer immediate failure like those performed by Jaboulay, and, indeed, some of the transplants functioned for months. However, the outcome of the transplant was generally unsatisfactory, as the recipients suffered repeated episodes of rejection or transplantation infection and all eventually died. The pathological changes in cross-species xenotransplantation are described by Porter et al. (5) as interstitial cellular infiltrates with edema, patchy hemorrhage, and patchy infarction. Although these early attempts From: Xenotransplantation: Basic Research and Clinical Applications Edited by: Jeffrey L. Platt © Humana Press Inc., Totowa, NJ
45
46
Bustos and Platt
failed from a clinical perspective, today they are perhaps responsible for the rise of interest in xenotransplantation. In 1966, Kissmeyer-Nielson et al. (6) describe “hyperacute rejection” of clinical allotransplants as a cause of early graft failure. By this time, Perper and Najarian (7) find that xenotransplantation might have two different outcomes. Organs transplanted between closely related species such as sheep-to-goat or chimpanzee-to-human function for a period of days before rejection ensues, and the characteristics of rejection resemble those of allografts. However, species that are phylogenetically distant, such as guinea pig-to-rat and pig-to-human, exhibit a course dramatically different owing to a hyperacute rejection reaction much like that described by Kissmeyer-Nelson (6). In 1970, Calne (8) formalized this concept. Species combinations in which xenografts are not subject to hyperacute rejection are called “concordant” and species combinations in which xenografts are subject to hyperacute rejection are called “discordant.” It would seem logical that the best xenograft donor from a physiologic and immunologic perspective would be phylogenetically close to the recipient (concordant xenograft). However, although these procedures may help individual patients, they will not solve the overall problem of donor shortage because few nonhuman primates of appropriate size can be found. The transplantation of pig organs is preferred because organs of appropriate size might be available in large numbers at low cost and because the transplantation of porcine organs engenders less risk of zoonosis than the transplantation of primate organs (9). These advantages have prompted surgeons and scientists to focus on a pig-toprimate model as the final preclinical model. However, the use of porcine organs represents a discordant model owing to the presence of preformed natural antibodies to pig antigens. In the study of discordant xenografts, small animal models, such as guinea pig-to-rat models, constitute the most abundant source of information about histologic and immunologic changes. The different antibody–antigen systems and complement activation pathways involved in small animal models differ significantly from the processes in humans and nonhuman primates. Thus, the information from small animals is insufficient for clinical application. The following section describes the pathology of xenografts in a pigto-baboon model with insights into the causes and pathogenesis of vari-
Pathological Responses to Xenotransplantation
47
ous types of xenograft rejection and suggests new rational therapeutic strategies for future clinical application of xenotransplantation.
HYPERACUTE REJECTION An organ transplanted into unmanipulated, phylogenetically disparate recipients is subject to hyperacute rejection, which destroys the graft. Hyperacute rejection begins immediately on reperfusion of a xenogeneic organ graft, destroying the graft between minutes to hours. The first clinical and pathological description of hyperacute rejection is commonly credited to Kissmeyer-Nielsen in 1966 (6). The pathological features of the rejected organs contained extensive microvascular thrombosis and neutrophil infiltration similar to what is seen in a generalized Schwartzman reaction. The recipients had titers of antibodies directed against donor kidney extracts. Based on these findings, KissmeyerNelson et al. (6) concluded that the rejection reaction is caused by preexisting antibodies directed against foreign antigens in the graft. Thus, what Kissmeyer-Nelson et al. provided for the first time is not only the pathologic description of hyperacute rejection, but they proposed that antibodies against tissue antigens of the donor could mediate a form of rejection that is unique clinically and pathologically. Platt et al. (10) describe this kind of rejection in heart, kidney, and lung xenografts. Macroscopically, blood flow to the transplant organ begins to decline and changes in coloration of the external surface of the xenograft are evident. The tempo of hyperacute rejection varies from experiment to experiment and in the combination of donor and recipient. In species combinations such as pig-to-primate, in which hyperacute rejection is initiated by natural antibodies, the titer of these antibodies is probably the most important factor in determining the rate of rejection. In other species such as guinea pig-to-rat, where hyperacute rejection does not depend on natural antibodies but rather reflects direct activation of the recipients’ complement system on donor cells, hyperacute rejection is especially rapid and explosive. Microscopically, hyperacute rejection is characterized by platelet aggregates and erythrocyte sludge in the lumen of blood vessels. As rejection progresses, the pathological features are dominated by interstitial hemorrhage and thrombosis with posterior destruction of vessels. At ultrastructural levels, the damage to endothelial cells becomes more evident, showing alteration in cellular junctions, with platelet attachment to blood vessels
48
Bustos and Platt
and small vessels appearing to be collapsed. In the evolution of this picture of hyperacute rejection, electron microscopy shows a distortion of the endothelium with irregular surfaces and separation from the underlying matrix. At this point, many capillaries are often found to be occluded by platelets with some erythrocytes. Advanced lesions show rupturing of vessels with extravasation to the interstitium. The immunopathology of hyperacute rejection has been described in detail by Platt et al. (10). Platt and colleagues reveal classical pathway components C1q, C2, and C4 deposited along blood vessels. The alternative pathway components factor B or properdin are observed in some, but not in all, tissues. Also, the presence of immunoglobulin deposits of recipient origin is found along the endothelial cell surfaces of graft blood vessels. The immunopathological studies suggested by Platt et al. are as follows: 1. The endothelial cells constitute the primary target of the immune reaction. 2. In most cases, complement activation in pig-to-primate xenografts is initiated by activation of the classical complement pathway.
ACUTE VASCULAR REJECTION Experimental approaches to the prevention of discordant xenograft hyperacute rejection are explored in pig-to-primate experimental models (Table 1). All of these manipulations combined with heavy pharmacologic immunosuppressive therapy extend graft survival. Although hyperacute rejection can be prevented by those approaches, another kind of rejection can also occur, namely, acute vascular rejection (11). Acute vascular rejection has also been referred to as delayed hyperacute rejection by others (12). When hyperacute rejection is averted according to approaches that have been mentioned, the xenograft becomes subject to acute vascular rejection, which destroys the graft over a period of hours to days. This type of rejection is now viewed as a major immunologic barrier to the clinical application of xenotransplantation. Although acute vascular rejection might be considered to be a delayed form of hyperacute rejection, there is much evidence that suggests acute vascular rejection is distinct from hyperacute rejection because the pathogenesis and the pathology of acute vascular rejection are different from that of hyperacute rejection.
Pathological Responses to Xenotransplantation
49
Table 1 Therapeutic target Xenoreactive antibodies
Complement (C)
Donor modification
Therapy
Mechanism
Plasmaphersis
Depletion of Ab+C
Absorbent columns Anti-idiotype antibody
Depletion of Ab+C Inhibition of Ab production/binding
Anti-B-cells agents
Inhibition of Ab production
Soluble antigen
Inhibition of Ab binding
Cobra venom factor Depletion of complement SCR1 Inhibition of complement Gamma globulin Diversion of binding Transgenic for human complement regulatory proteins H-transferase transgenic pigs
References 14
11, 23–26
27,28
Acute vascular rejection may be related pathogenetically to the activation of graft endothelial cells, but the events that incite endothelial cell activation are subject to controversy. Bach and co-workers (12) propose that acute vascular rejection is caused by biological processes that occur independently of the immune reaction of the host against the graft. Based on four lines of evidence, Platt and co-workers (13) propose that acute vascular rejection is triggered by persistent interaction of xenoreactive antibodies with graft tissue as follows: 1. Primates from which xenografts are removed after rejection have a sudden increase in antidonor antibody levels, implying that the xenograft is continually exposed to xenoreactive antibodies and is actively absorbing them from circulation.
50
Bustos and Platt
2. Acute vascular rejection of allografts and concordant xenografts is associated with the presence of antidonor antibodies in the blood or can be induced by administration of antidonor antibodies. 3. Patients exposed to porcine antigens from extracorporeal circulation through porcine livers experience an increase in the titer of xenoreactive antibodies within a few days, coinciding with the time when a xenograft is subject to acute vascular rejection, and suggesting that immune stimulation has occurred. 4. Cytotoxic agents such as cyclophosphamide that inhibit the synthesis of antibodies appear to delay or avert acute vascular rejection.
Although the importance of antibodies in the development of acute vascular rejection seems evident, the exact nature of those antibodies is less certain (14). The histopathological changes observed on transgenic porcine organs after transplantation into baboons begin as soon as 1 h after transplantation. The most common change is prominent endothelial cell swelling in the capillaries with some red cells trapped within the lumen of the vessels. All the vessels are intact and myocytes are well preserved. After 24 h, endothelial cells exhibit marked swelling with increased nuclear size. Some capillaries appear to be occluded with a “rope-like” appearance that is described as typical of acute vascular rejection in allografts. Some vessels appear congested. In general, cardiomyocytes are preserved at 24 h, although in some areas myocytes show shrinkage with moderate nuclear pyknosis. Biopsies taken 72 h after transplantation and at later times, close to the time of rejection, have capillaries with the same features as at earlier times. Some capillaries remain open but some vessels show fibrin in the lumen and are occluded by swelling of endothelial cells and by various types of blood cells. Cardiomyocytes that lack striations and vacuolization of the cytoplasm are seen. Swelling of endothelial cells remain the main feature, whereas destruction of the vascular wall is not often observed. Cardiomyocytes appear to be damaged, having a wavy shape and pyknotic nucleus in areas associated with infiltrate of mononuclear cells, and in other areas, mild infiltrate of neutrophils. Infiltration of mononuclear cells appears around the blood vessels first and later in the interstitium, destroying the cardiac cells. Electron microscopy confirms, at the ultrastructural level, the findings by light microscopy and the events shown by immunofluorescence. Moreover, electron microscopy allows the study of vascular structure in detail. In biopsies taken in the first hour, endothelial cells show slight swelling without inflammatory cells in the interstitium. No fibrin is seen
Pathological Responses to Xenotransplantation
51
in the vessels or subendothelial domains. After 24 h, biopsies reveal histologic features of various grades of “damage” to endothelial cells. At the beginning, the normal flat cytoplasm of endothelial cells is disrupted by the appearance of multiple pinocytic vesicles. The vesicles appear along the luminal surface as well as central and peripheral aspects of the cell. In contrast to normal endothelial cells, which contain few organelles, the cytoplasm of endothelial cells in the organ transplants contain numerous ribosomes. At this stage, the basal membrane remains intact. The endothelial cells are thicker than in normal cells. Moreover, the flattened nucleus of the normal cell is changed by a protrudent nucleus into the lumen, giving the vessel a general undulant appearance. The interendothelial junctions are dense, long, and irregular. The most prominent characteristic in electron micrographs is an irregular lumen surface that contrasts with the smooth surface on normal endothelial cells. Cytoplasmic blebs or evaginations of the plasma membrane are a common feature of the lumen. This change appears more prominent at early time points. During this process of blebbing, cytoplasmic material appears to be lost. In these early lesions, the interstitium is increased in area, but inflammatory cells are not observed. Vessels are surrounded by edema (with fibroblasts and collagen). Myocytes show some damage in patches observed by lack of striations. Biopsies taken 3–7 d after transplantation, before the organ is rejected, show invariable changes present on the endothelial cells. The cytoplasmic volume of endothelial cells is increased. Endothelial cells protrude into the capillary lumen and, because of severe swelling, endothelial cells appear to be enfolding and occluding almost the total lumen of the capillaries. The cytoplasm reveals pallor, probably owing to excess water uptake diluting the cytoplasm matrix, and appears relatively structureless. Organelles and inclusions are separated by electron-lucent areas of cytoplasm. The endothelial cells in other larger vessels reveal moderate swelling of the cytoplasm with irregular and undulant surfaces. The endothelial cell surfaces develop long projections, called filopodia, important for binding blood cells. Platelets and white cells appear to be trapped or attached to the endothelium. The lumen of capillaries in advanced stages of rejection appears to be occupied by fibrin strands and sometimes the lumen is occluded by fibrin clots containing white cells and platelets. Platelets appear to be degranulated and in contact with fibrin and white cells. Subendothelial fibrin is observed and sometimes there is evidence of discontinuity between endothelial cells.
52
Bustos and Platt
It is important to mention that the process of rejection is dynamic, and it is common to see the juxtaposition of moderately damaged vessels next to severely damaged vessels. Ultrastructural changes in the stages of rejection show many necrotic endothelial cells and myocytes. Between the myocytes it is possible to observe fibrin strands that disrupt the cardiac cells. The immunopathological study of acute vascular rejection reveals the presence of IgM on the vessels within 1 h after transplantation. In some cases, IgM fluorescence decreases at 24 h and remains low for as long as 3 d. In other cases, IgM deposition remains at the same intensity as seen in the first hour. Deposition of IgG is not observed at 1 h after transplant; however, after 24 h, and especially after 3 d, IgG staining is apparent. The presence of IgG is seen even in the interstitium, suggesting that IgG is leaking from the vessels. Endothelial cells in normal porcine tissues are positive for MHC class I and MHC class II. MHC class I protein levels remain the same until 3 d after transplantation, at which time MHC class I expression on the surface of cardiomyocytes increases. The increase of MHC I is coincident with the presence of cellular infiltration, although the infiltration appears around vessels first. After 3 d, blood vessels are strongly MHC II positive, while a mild cellular infiltrate outside the vessels is also positive for MHC II. The presence of MHC II fluorescence around the vessels correlates with the presence of infiltrate. Biopsies taken early after transplantation do not show cellular infiltration. Infiltration by CD16+ cells is occasionally seen, and the presence of CD2+ cells appears around the vessels at d 7 and later in the interstitium. The influx of macrophages and PMN is present in biopsies associated with the presence of ischemia. Although platelet thrombi in capillaries are a typical feature of hyperacute and acute vascular rejection, the presence of platelets along the vessels is observed as a small component with the presence of fibrin. The analysis of vessels shows progressive deposition of fibrin, with small fibrin thrombi in the vessels by the third day. As the lesion progresses, the presence of fibrin is detected in the interstitium, reflecting barrier failure provided by endothelial cells.
DELAYED HYPERACUTE REJECTION Although acute vascular rejection might be considered to be a delayed form of hyperacute rejection (12), there is much evidence that acute vascular rejection is distinct from hyperacute rejection. First, acute vascular rejection is observed in allografts and concordant xenografts in
Pathological Responses to Xenotransplantation
53
which hyperacute rejection normally does not occur (15,16). Second, the pathology of acute vascular rejection differs from that of hyperacute rejection (16,17). Third, although the pathogenesis of acute vascular rejection is generally thought to reflect activation of endothelial cells in the transplant (18,19), the course of hyperacute rejection proceeds too rapidly to allow significant effects from endothelial cell activation. Fourth, acute vascular rejection develops when the complement system of the recipient is inactivated, a condition that invariably precludes the development of hyperacute rejection. Thus, we think that the term of delayed hyperacute rejection could be reserved to the pathological picture dominated by occlusion of erythrocytes, venular and capillary thrombi, interstitial hemorrhage, and influx of neutrophils in the same proportion to the extravasated erythrocytes and disruption of the capillaries. The pathological features of this condition are thus indistinguishable from hyperacute rejection.
CHRONIC REJECTION The possibility that pig-to-primate xenografts may be subject to chronic rejection, as allografts, remains to be explored. A limited number of studies in small animal models suggest that graft vascular disease may be an important impediment to long-term xenograft survival. Scheringa et al. (20), using a hamster-to-rat aorta transplantation model (concordant xenograft), show that features common to allograft chronic rejection, namely, intimal proliferation and infiltrating macrophages and T-cells, are the same with this xenograft model. Recently, Shen et al. (21) induced chronic rejection in hamster hearts transplanted into Lewis rats treated with leflunomide. Such lesions in xenografts involve arterial tree damage with histological similarities as well as differences with allografts. In summary, they describe differences in the injury pattern mainly involved with larger sized arteries in xenografts with morphologically more aggressive lesions in xenografts than allografts, such as fibrinoid necrosis, marked intimal edema with a large accumulation of extracellular matrix with or without mononuclear cell infiltration. Thus, xenografts represent a more intensive and aggressive process of arterial injury that is less favorable to long-term graft survival. At least two considerations have to be made in the interpretation of the above descriptions. First, the model used in the description of chronic rejection is a concordant xenograft, and, second, the use of small animal models cannot address many of the problems seen in the large animal discordant xenografts. For instance, the significance of anti-Gal anti-
54
Bustos and Platt
bodies (IgM or IgG) in chronic rejection could not be evaluated properly in these models. Understanding the role of these antibodies could provide important information in the search for new immunosuppressive drugs or an approach to tolerance induction. Galili (22) evaluates the role of antiGal IgG in chronic xenograft rejection and the association between _-Gal epitope expression and inflammatory infiltrates. Galili concludes that anti-Gal IgG would induce xenograft destruction by antibody dependent cell-mediated cytotoxicity (ADCC) by activation of endothelial cells and by increasing activation of T cells against xenograft antigens. Although the histopathology of chronic rejection in pig-toprimate transplants is unknown, one can easily imagine that this kind of rejection could be more intense and aggressive than seen in allografts and would justify a search for new immunologic approaches to overcome.
REFERENCES 1. U.S. Renal Data System, USRDS 1996 Annual Data Report. Bethesda, MD: National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases, 1996. 2. Reemtsma K. Xenotransplantation: a brief history of clinical experiences: 1900– 1965. In: Cooper DKC, Kemp E, Reemtsma K, White DJG, eds. Xenotransplantation. The transplantation of organs and tissues between species. New York: Springer-Verlag, 1991, 9–22. 3. Reemtsma K, McCracken BH, Schlegel JU, et al. Renal heterotransplantation in man. Ann Surg 1964; 160:384–410. 4. Starzl TE, Marchioro TL, Peters GN, et al. Renal heterotransplantation from baboon to man: experience with 6 cases. Transplantation 1964; 2:752–776. 5. Porter KA, Marchioro TL, Starzl TE. Pathological changes in six treated baboonto-man renal heterotransplants. Brit J Urol 1965; 37:274–284. 6. Kissmeyer-Nielsen F, Olsen S, Petersen VP, Fjeldborg O. Hyperacute rejection of kidney allografts, associated with pre-existing humoral antibodies against donor cells. Lancet 1966; 2:662–665. 7. Perper RJ, Najarian JS. Experimental renal heterotransplantation. I. In widely divergent species. Transplantation 1966; 4:377–388. 8. Calne RY. Organ transplantation between widely disparate species. Transplant Proc 1970; 2:550–556. 9. Cooper DKC, Ye Y, Rolf LL, Jr., Zuhdi N. The pig as potential organ donor for man. In: Cooper DKC, Kemp E, Reemtsma K, White DJG, eds. Xenotransplantation. The transplantation of organs and tissues between species. New York: Springer-Verlag, 1991,481–500. 10. Platt JL. Hyperacute xenograft rejection. Medical Intelligence Unit. Austin: R.G. Landes, 1995.
Pathological Responses to Xenotransplantation
55
11. Magee JC, Collins BH, Harland RC, et al. Immunoglobulin prevents complement mediated hyperacute rejection in swine-to-primate xenotransplantation. J Clin Invest 1995; 96:2404–2412. 12. Bach FH, Winkler H, Ferran C, Hancock WW, Robson SC. Delayed xenograft rejection. Immunol Today 1996; 17:379–384. 13. Platt JL, Lin SS, McGregor CGA. Acute vascular rejection. Xenotransplantation 1998; 5:169–175. 14. Lin SS, Weidner BC, Byrne GW, et al. The role of antibodies in acute vascular rejection of pig-to-baboon cardiac transplants. J Clin Invest 1998; 101:1745–1756. 15. McPaul JJ, Stastny P, Freeman RB. Specificities of antibodies eluted from human cadaveric renal allografts. Journal for Clinical Investigation 1981; 67:1405–1414. 16. Porter KA: Renal transplantation. In Heptinstall RH, ed. Pathology of the kidney. Vol. 3. Boston: Little Brown and Company, 1992,1799–1933. 17. Rose AG, Cooper DKC. Ultrastructure of hyperacute rejection in cardiac xenografts. In: Cooper DKC, Kemp E, Reemtsma K, White DJG, eds. Xenotransplantation. the transplantation of organs and tissues between species. New York: Springer-Verlag, 1991,243–252. 18. Platt JL, Vercellotti GM, Dalmasso AP, et al. Transplantation of discordant xenografts: a review of progress. Immunol Today 1990; 11:450–456. 19. Blakely ML, Van Der Werf WJ, Berndt MC, Dalmasso AP, Bach FH, Hancock WW. Activation of intragraft endothelial and mononuclear cells during discordant xenograft rejection. Transplantation 1994; 58:1059–1066. 20. Scheringa M, Buchner B, Geerling RA, et al. Chronic rejection after concordant xenografting. Transplant Proc 1994; 26:1346–1347. 21. Shen J, Chong AS, Xiao F, et al. Histological characterization and pharmacological control of chronic rejection in xenogeneic and allogeneic heart transplantation. Transplantation 1998; 66:692–698. 22. Galili U. Significance of anti-Gal IgG in chronic xenograft rejection. Transplant Proc 1999; 31:940–941. 23. Kobayashi T, Taniguchi S, Ye Y, et al. Delayed xenograft rejection in C3-depleted discordant (pig-to-baboon) cardiac xenografts treated with cobra venom factor. Transplant Proc 1996; 28:560. 24. Kobayashi T, Taniguchi S, Neethling FA, et al. Delayed xenograft rejection of pigto-baboon cardiac transplants after cobra venom factor therapy. Transplantation 1997; 64:1255–1261. 25. Leventhal JR, Dalmasso AP, Cromwell JW, et al.Prolongation of cardiac xenograft survival by depletion of complement. Transplantation 1993; 55:857–866. 26. Marsh HC Jr., Ryan US. Therapeutic effect of soluble complement receptor type I xenotransplantation. In: Cooper DKC, Kemp E, eds. Xenotransplantation. Heidelberg: Springer, 1997,437–447. 27. McCurry KR, Kooyman DL, Alvarado CG, et al. Human complement regulatory proteins protect swine-to-primate cardiac xenografts from humoral injury. Nat Med 1995; 1:423–427. 28. Sharma A, Okabe JF, Birch P, Platt JL, Logan JS. Reduction in the level of Gal (_1,3) Gal in transgenic mice and pigs by the expression of an _(1,2) fucosyltransferase. Proc Natl Acad Sci 1996; 93:7190–7195.
vi
Natural Xenoreactive Anitbodies
3
57
Natural Xenoreactive Antibodies Uri Galili INTRODUCTION
Many of the immune mechanisms that cause xenograft rejection in humans differ greatly from the mechanisms that mediate allograft rejection. Whereas allograft rejection is primarily the result of T-cell response to MHC antigens, much of the xenograft rejection process is the result of antibody activity. The amino acid sequence of almost every protein on pig cells is likely to differ by several percent from the homologous protein in humans. Thus, many pig proteins would serve as immunogenic entities that induce antibody formation in xenograft recipients. It is possible that the currently used immunosuppressive regimens can effectively prevent the production of such anti-xenopeptide antibodies. However, the major antibody that mediates xenograft rejection is an anti-carbohydrate antibody, termed anti-Gal. This antibody is present as a natural antibody in humans, and it interacts specifically with the carbohydrate structure Gal_1-3Gal_1-4GlcNAc-R, termed the _-gal epitope. (This epitope was also designated Gal_1-3Gal determinant and _-galactosyl epitope.) Moreover, in xenograft recipients this antibody activity increases by up to 300-fold, effectively contributing to chronic rejection of the xenograft. Unfortunately, immunosuppressive regimens currently used for prevention of allograft rejection are ineffective in preventing this massive increase in anti-Gal activity, which results from the immune response to _-gal epitopes on the xenograft. This chapter discusses anti-Gal, the _-gal epitope, the evolutionary and biosynthetic basis for their reciprocal distribution in mammals, the changes occurring in this antibody in xenograft recipients, and possible contribution of other natural antibodies to xenograft rejection.
From: Xenotransplantation: Basic Research and Clinical Applications Edited by: Jeffrey L. Platt © Humana Press Inc., Totowa, NJ
57
58
Galili
CHARACTERISTICS OF THE NATURAL ANTI-GAL ANTIBODY Anti-Gal is a natural antibody that constitutes approx 1% of the immunoglobulins in humans. It was first isolated by affinity chromatography on melibiose (Gal_1-6Gal) - sepharose columns, elution of the bound antibodies with free _<methyl-galactoside, and dialysis of the eluate for the removal of the free carbohydrate (1). The activity of this antibody was then demonstrated by its ability to agglutinate rabbit red blood cells and specific inhibition of this agglutination by carbohydrate molecules that contain galactose linked in _
Natural Xenoreactive Anitbodies
59
Class analysis of this antibody has indicated that anti-Gal is present in human serum in all three isotypes, IgG being the most prevalent (9). Subclass analysis of anti-Gal IgG molecules demonstrated their production in all subclasses. Two studies reported IgG1 to be the most prevalent subclass (14,15), whereas other studies reported that IgG2 subclass of anti-Gal is the subclass with the highest concentration in the serum (11,16). It is not clear whether these differences result from different reagents used in these assays, or there are large differences in production of subclasses in various individuals. Anti-Gal also comprises a large proportion of circulating IgM (9,11,16–18) and is also found as IgA molecules (9,18,19). In secretions such as saliva, milk, colostrum, and bile, anti-Gal is found primarily as IgA molecules (18). Anti-Gal IgA was found to be markedly elevated in the serum of patients with Henoch-Schönlein purpura and IgA nephropathy (19). The proportion of B lymphocytes capable of producing anti-Gal could be estimated by transformation of human blood lymphocytes with Epstein Barr virus (EBV) and determination of the proportion of lymphocytes producing this antibody. As many as 1% of the transformed lymphocytes were found to secrete anti-Gal in all individuals tested (20). In comparison, only 0.2–0.25% of the transformed lymphocytes were found to produce anti-blood group A or B antibodies in individuals that lacked the corresponding blood group antigens (20). Analysis of the heavy chain genes (VH genes) of the clones producing anti-Gal demonstrated a variety of genes most of which are clustered in the VH3 gene family (21). As indicated above, anti-Gal is produced in all humans throughout life. Upon birth, anti-Gal is found in the newborn only as the IgG isotype, which is of maternal origin (1,22,23). The titer of antibody in the cord blood is similar to that in the maternal blood (1). Anti-Gal decreases to approx 20% of its normal level by the age of 3–6 mo and thereafter it increases and reaches the adult level by the age of 2–4 yr (1). In a large proportion of individuals above the age of 75, anti-Gal activity was found to be significantly lower than in younger individuals (24). This may reflect the general age-associated decrease in the efficacy of the immune response in the elderly.
60
Galili
DISTRIBUTION OF ANTI-GAL AND _-GAL EPITOPES IN MAMMALS The abundant production of anti-Gal in humans implies that humans do not synthesize _
(N-acetyllactosamine)
(uridine diphosphate galactose)
A (_-gal epitope)
It is of interest to note that the _-gal epitope is synthesized only in mammals, whereas nonmammalian vertebrates such as birds, reptiles, amphibians, and fish lack this epitope (27). This implies that the _1,3GT appeared evolutionarily only in ancestral mammals and not in other vertebrates. Because marsupials also express _-gal epitopes, it is probable that _1,3GT appeared in mammals before marsupials and placentals diverged from each other. In view of the fact that Old World primates
Natural Xenoreactive Anitbodies
61
lack _-gal epitopes, it is likely that the activity of this enzyme was suppressed in ancestral Old World monkeys and apes after they diverged from New World monkeys, 30 to 40 million years ago. This suppression of _1,3GT activity in ancestral Old World primates could be the result of the appearance of a pathogen that was endemic to the Old World, was detrimental to primates, and expressed _-gal epitopes. Such a pathogen would have exerted a selective pressure on primates to suppress autologous _-gal epitope expression and lose immune tolerance toward it. This, in turn, could enable production of anti-Gal as a means of defense against such a putative pathogen. This hypothesis has been supported by several observations on the ability of anti-Gal to induce complementmediated lysis in a variety of microbial pathogens including viruses (29,30), bacteria (13), and protozoa (31). The cloning of _1,3GT gene in mouse (32) and cow (33) enabled the elucidation of the molecular mechanisms that results in the suppression of _1,3GT activity in ancestral Old World primates. Despite the lack of _1,3GT activity in humans and Old World monkeys, the gene coding for this enzyme is present as a pseudogene (33–35). However, no _1,3GT mRNA could be detected in monkey or human cells, implying that the enhancer/promoter regions of this gene are inactive and therefore the gene does not undergo transcription (33,36). In addition, the human and ape _1,3GT pseudogene was found to contain frame shift mutations which cause the production of a truncated and catalytically inactive enzyme in the event that the _1,3GT gene is aberrantly transcribed (34,35,37). Thus, synthesis of _
EFFECT OF ANTI-GAL ON XENOGRAFTS The evolutionary suppression of the _1,3GT gene and the subsequent production of anti-Gal has generated an immunologic barrier between humans producing large amounts of anti-Gal and mammals with cells that express the ligand for this antibody, i.e., the _
62
Galili
indicated that in vivo binding of anti-Gal to _
PRODUCTION OF ANTI-GAL IgG IN XENOGRAFT RECIPIENTS It could be argued that removal of anti-Gal by affinity columns prior to transplantation may prevent rejection of xenografts. However, it was found that following such treatment anti-Gal reappears in the circulation within 4–5 d because of the ongoing immune response to the gastrointestinal flora (50). Moreover, upon transplantation of xenografts into human or monkeys the immune system increases production of anti-Gal IgG so that the titer of this antibody may increase by 30- to 300fold (9,51). This immune response is highly detrimental to xenografts because, as described below, it effectively facilitates chronic xenograft rejection. The strong immune response to _
Natural Xenoreactive Anitbodies
63
activity in the serum of these patients has indicated that it increased by 20–100-fold within the period of 25–50 d post transplantation (9,53). This was the result of increase in titer of the three immunoglobulin classes; however, most of the increase was observed with the anti-Gal IgG isotype (9). It should be stressed that this increase in anti-Gal activity occurred despite the immunosuppressive treatment, which successfully prevented the rejection of the kidney allografts (51,52). Since total IgG concentration in the serum of the xenograft recipients did not change significantly posttransplantation, it is probable that the 100-fold increase in titer of anti-Gal in some of the patients was not only because of the increase in the concentration of the antibody but also the result of increased affinity of anti-Gal. Direct measurement of anti-Gal affinity in these sera, by equilibrium dialysis with radiolabeled free _
64
Galili
70% were T cells and 30% were macrophages (54). The infiltration of T cells seemed to be associated with _95% decrease in the infiltration of T cells into the xenograft (55). It should be stressed that these xenografts contained no live cells and most of the _
DETRIMENTAL EFFECTS OF HIGH AFFINITY ANTI-GAL The studies described above on anti-Gal response in xenograft recipients imply that the immune system responds to _
Natural Xenoreactive Anitbodies
65
The high affinity anti-Gal produced in xenograft recipients is very detrimental to xenografts. As indicated above, these antibodies can target xenoantigens to APCs. In addition, high affinity anti-Gal can very effectively mediate ADCC by binding to _
ARE THERE OTHER NATURAL ANTIBODIES THAT ARE INVOLVED IN XENOGRAFT REJECTION? The wide diversity of natural antibodies produced against a large variety of bacterial antigens and against other environmental antigens raises the question of whether there are additional natural antibodies that may contribute to xenograft rejection. One approach for studying this issue is to measure overall antibody binding to pig cells incubated in human sera, before and after specific removal of anti-Gal. Such experiments were performed with pig endothelial cells, and the binding of IgG molecules was assessed by subsequent binding of 125I-protein A. Most (>85%) of the immunoglobulin molecules in human or Old World monkey sera that could bind to pig endothelial cells were eliminated by specific removal of anti-Gal from the sera (59). It is not clear whether the remaining antibodies which bind to the endothelial cells are anti-Gal that was not successfully removed by adsorption, or do they have other
66
Galili
specificities. Similar studies with primate sera that lack anti-Gal, i.e., sera from the New World squirrel monkey, demonstrated the lack of natural antibodies to pig endothelial cells (59). This finding supports the assumption that anti-Gal is the only significant natural antibody that binds to pig cells. Other species may have large amounts of natural antibodies to pig cells that are not anti-Gal and that are absent in humans. Such antibodies are found in large amounts in dogs. Dog sera contain high titers of natural antibodies to pig endothelial cells that can induce complementmediated lysis of pig cells (59). Evidently, these antibodies are not antiGal because dogs synthesize _
Natural Xenoreactive Anitbodies
67
of such an immune response is the production of non-anti-Gal antibodies to pig cartilage in monkeys transplanted with pig cartilage devoid of _
CONCLUSIONS Anti-Gal is the major xenoreactive antibody in humans and monkeys. It comprises 1% of circulating immunoglobulins and it interacts specifically with _
REFERENCES 1. Galili U, Rachmilewitz EA, Peleg A, Flechner I. A unique natural human IgG antibody with anti-_
68
Galili
4. Galili U, Macher BA, Buehler J, Shohet SB. Human natural anti-_
Natural Xenoreactive Anitbodies
69
21. Wang L, Radic MZ, Galili U. Human anti-Gal heavy chain genes: Preferential use of VH3 and the presence of somatic mutations. J Immunol 1995; 155:1276–1285. 22. Minanov OP, Itescu S, Neethling FA, et al. Anti-Gal IgG antibodies in sera of newborn humans and baboons and its significance in pig xenotransplantation. Transplantation 1997;63:182–186. 23. Galili U, Minanov O, Michler RE, Stone KR. High affinity anti-Gal IgG in chronic rejection of xenografts. Xenotransplantation 1997; 4:127–131. 24. Wang L, Anaraki F, Henion TR, Galili U. Variations in activity of the human natural anti-Gal antibody in young and elderly populations. J Gerontol (Med Sci) 1995; 50A:M227–M233. 25. Spiro RG, Bhoyroo VD. Occurrence of _
70
Galili
38. Galili U. Interaction of the natural anti-Gal antibody with _
Natural Xenoreactive Anitbodies
71
54. Stone KR, Walgenbach AW, Abrams T, Nelson J, Gellett N, Galili U. Porcine and bovine cartilage transplants in cynomolgus monkey: I. A model for chronic xenograft rejection. Tranplantation 1997; 63:640–645. 55. Stone KR, Ayala G, Goldstein J, Hurst R, Walgenbach A, Galili U. Porcine cartilage transplants in cynomolgus monkey: III. Transplantation of _
vi
Xenoreactive Natural Antibodies
4
73
Specificity of Xenoreactive Natural Antibodies William Parker, PhD, Paul B. Yu, MD and Yuko C. Nakamura, BS
INTRODUCTION The specificity of primate antibodies that bind to porcine tissues and organs is of great interest given the importance of antibodies in the rejection of porcine-to-human xenografts, as discussed in Chapter 3. Multiple epitopes on almost any porcine protein, as well as many porcine carbohydrates (1), might be recognized as foreign if surveyed by the human immune system. Indeed, this fact may have a large impact on the direction scientists take as they attempt to achieve the long-term survival of xenografts. However, naturally occurring xenoreactive antibodies, unlike the vast number of antibodies that might be elicited against a pig organ, are generally thought to be limited in specificity. In this chapter we will focus on the specificity of natural xenoreactive antibodies, those xenoreactive antibodies that already exist in the serum of individuals, despite their lack of history of exposure to xenogeneic antigens. It is these antibodies that pose the initial barrier to xenotransplantation. Furthermore, even if natural antibodies are insufficient to cause the rejection of a particular organ, they may initiate pathologic processes that result in a decrease in function of a transplanted organ. For example, binding of antibodies and/or subsequent activation of complement by those antibodies may dramatically increase the activity of antigen presenting cells, resulting in a devastating elicited immune response greater than the response seen in allografts.
From: Xenotransplantation: Basic Research and Clinical Applications Edited by: Jeffrey L. Platt © Humana Press Inc., Totowa, NJ
73
74
Parker, Yu, and Nakamura
XENOREACTIVE ANTIBODIES SPECIFIC FOR GAL_1–3GAL: FINE SPECIFICITY It has been known for some time that most of the natural antibodies that bind to porcine cell surfaces are specific for the disaccharide Gal_1– 3Gal (2). First discovered and identified as xenoreactive in the mid1980s (3), anti-Gal_1–3Gal antibodies were identified as being potentially important to porcine-to-primate xenotransplantation in 1992 (4). Work in multiple laboratories has provided conclusive evidence that these xenoreactive antibodies, specific for Gal_1–3Gal, are a critical component of the immune process responsible for the hyperacute rejection of porcine organs by primates (2,5). Because of the large role that anti-Gal_1–3Gal antibodies play in the hyperacute rejection of xenogeneic organs, considerable work has been devoted toward determining the fine specificity of anti-Gal_1–3Gal antibodies. Early work by Wieslander indicated that some anti-Gal_1– 3Gal antibodies would bind to _-CH3 galactopyranoside, Gal_1–6Gal, Gal_1–4Gal, and galactose (6). These findings suggested that subterminal galactose was necessary for binding of some antibodies and unnecessary for the binding of others (Fig. 1). Later work by the author and colleagues (7) showed that the requirement for a subterminal saccharide was highly variable between different individuals. In some individuals, anti-Gal_1–3Gal antibodies required a subterminal galactose for binding, although anti-Gal_1–3Gal antibodies from other individuals did not require a subterminal galactose for binding, with most individuals having a mixture of the two distinct types of anti-Gal_1–3Gal antibodies (7). Teneberg showed that at least some anti-Gal_1–3Gal antibodies do not require a terminal nonreducing Gal, as some of the antibodies recognized the structure GalNAc_1–3Gal`1–4GlcNAc`1–3Gal`1–4Glc`1Cer (8). Binding of anti Gal_1<3Gal antibodies to this structure is not surprising, since the nonreducing terminus of the molecule is similar to Gal_1–3Gal, except that it has an N-acetyl substitution on the terminal galactose. Thus, like blood group antibodies that bind to both blood group A (terminal GalNAc) and blood group B (terminal Gal), at least some anti-Gal_1– 3Gal antibodies may recognize both terminal Gal and terminal GalNAc. The structure recognized by anti-Gal_1–3Gal on porcine cells is Gal_1–3Gal`1–4GlcNAc-R, where R is a glycoprotein or glycolipid. Wieslander (6) was the first to provide evidence that the second subterminal sugar from the nonreducing end (GlcNAc) was not required for binding, but that its presence did afford much tighter binding for all anti-
Xenoreactive Natural Antibodies
75
Fig. 1. The fine specificity of anti-Gal_1–3Gal antibodies. A drawing of the Gal_1–3Gal`1–4GlcNAc structure is shown and important parts of that structure for binding of anti-Gal_1–3Gal antibodies are indicated.
Gal_1–3Gal antibodies (Fig. 1). More recent work by others (7,9) has confirmed these results using antibodies from other individuals. Furthermore, Neethling (9) demonstrated that the anti-Gal_1–3Gal recognized Gal_1–3Gal`1–4GlcNAc more avidly than any other saccharide, but still bound more tightly to structures such as Gal_1–3Gal`1–4Gal than to Gal_1–3Gal. Thus, the fine specificity of anti-Gal_1–3Gal antibodies for the first residue from the nonreducing terminus is dramatically different than the fine specificity for the second residue from the non-reducing terminus. The requirement for the subterminal Gal seems to be all or none, and it varies from individual to individual, whereas the requirement for the subterminal GlcNAc seems to be universal and moderate (Fig. 1). Several other lines of work have also provided information regarding the fine specificity of anti-Gal_1–3Gal antibodies. Some work has been directed toward the binding of anti-Gal_1–3Gal antibodies to blood group B {Gal_1–3(Fuc_1–2)Gal} (Fig. 1). Although it is clear that anti-Gal_1–3Gal antibodies from an individual who is tolerant to blood group B will not bind to blood group B, some, but not all, anti-Gal_1– 3Gal antibodies from other individuals apparently will bind to blood group B (10). Thus, the presence of the branched Fuc residue, like the absence of the subterminal Gal residue, apparently prevents the binding of some anti-Gal_1–3Gal antibodies and does not affect the binding of others (Fig. 1). Binding of anti-Gal_1–3Gal antibodies to Gal_1– 3Gal`1-4(Fuc_1–3)GlcNAc`1–3Gal`1–4GlcNAc`1-Cer has also
76
Parker, Yu, and Nakamura
been observed (11). Therefore, fucose substitution on the GlcNAc of the Gal_1–3Gal`1–4GlcNAc-R apparently does not interfere with the binding of anti-Gal_1–3Gal antibodies. A few molecules that are not related to _-galactose are recognized by anti-Gal_1–3Gal antibodies. The recognition of these “mimeotopes” by anti-Gal_1–3Gal_1<3Gal antibodies may provide some insight into the fine specificity of the antibodies, although more-detailed studies of how these molecules fit into the binding pocket of the antibody will be required. Several peptides that are memeotopes for Gal_1–3Gal and that are recognized by anti-Galal-3Gal antibodies have been identified (12). Perhaps more interesting is the finding that the structures GlcNAc`1–3Gal`1–4GlcNAc`1–3Gal`1–4Glc`1-Cer and GalNAc`1– 3Gal`1– 4GlcN Ac`1–3Gal`1–4Glc`1-Cer are also recognized by anti-Gal_1–3Gal antibodies (8). These stuctures are recognized by other proteins specific for _-galactose, and, therefore, are apparently mimeotopes of Gal_1–3Gal. One issue of particular note is that fine specificity of only a very limited population of anti-_ galactosyl antibodies has been discussed. Anti-Gal_1–3Gal antibodies, those anti-_-galactosyl antibodies that are xenoreactive, comprise only a fraction of naturally occurring anti_-galactosyl antibodies. Furthermore, there is a wide range of naturally occurring antibodies specific for one or more of a wide range of _-galactosyl structures (13). In addition, although the experimental work discussed above is limited to anti-_-galactosyl antibodies found in humans and Old World primates, anti-_-galactosyl antibodies are phylogenetically disperse (14,15). Among mammals, it is thought that only apes, humans, and Old World monkeys produce anti-Gal_1–3Gal antibodies, since other mammals produce the Gal_1–3Gal antigen. However, even this notion may be overly simplistic. As an example, mice, which express Gal_1–3Gal, are known in some cases to produce antibodies that bind to Gal_1–3Gal expressed by porcine cells (16), indicating that mice apparently produce antibodies that distinguish between porcine (foreign) Gal_1–3Gal and murine (self) Gal_1–3Gal based on some, as yet unknown, feature of the naturally occurring carbohydrate structures, possibly those adjacent to Gal_1–3Gal. Although not the topic of this chapter, it is noteworthy that, although, under most conditions the vast majority of xenoreactive antibodies that bind to porcine cell surfaces are specific for Gal_1–3Gal, there is, perhaps surprisingly, considerable controversy surrounding almost every characteristic of the antibodies specific for Gal_1–3Gal. For example,
Xenoreactive Natural Antibodies
77
some laboratories have found anti–Gal_1–3Gal antibodies to be predominately IgG (3), whereas others have found them to be a mixture of IgM and IgG, with the amount of IgG depending on blood type and species (17,18). A second point of disagreement concerns the uniqueness of anti-Gal_1–3Gal antibodies. One view is that they are completely unique, being unlike any other previously identified natural antibody (3). Others, however, have found that they are very similar to isohemagglutinins (18) and, possibly, to other anti-_-galactosyl antibodies that do not bind to Gal_1–3Gal (6,14). Thus, although the specificity, and even the fine specificity, of most xenoreactive antibodies is understood, other characteristics of these antibodies remain in question.
XENOREACTIVE NATURAL ANTIBODIES NOT SPECIFIC FOR GAL_1–3GAL There is mounting evidence that xenoreactive antibodies with specificities for saccharides other than Gal_1–3Gal are present in normal human serum and in normal primate serum. Although it has not been evaluated, these antibodies might contribute to hyperacute rejection, even if they are not sufficient to initiate the process by themselves, and they might contribute to later forms of xenograft rejection, even if antiGal_1–3Gal antibodies or the Gal_1–3Gal antigen is eliminated. Furthermore, it may be difficult to achieve tolerance for antigens to which the host has already responded and produced natural antibodies. For these reasons, we will give consideration to natural antibodies other than anti-Gal_1–3Gal, even though the putative physiologic role of these antibodies in the xenograft is completely unknown. There is conclusive evidence that natural antibodies besides antiGal_1–3Gal are reactive to porcine antigens. Blood group A, expressed by some pigs and recognized by many humans, is, perhaps, the most obvious example. Blood group A is weakly expressed on the endothelium of pigs, perhaps alleviating, for now, any concern about the role of anti-blood group A antibodies in the pathology of xenotransplant rejection. However, this antigen is expressed more strongly in other porcine tissue types, such as epithelium, and pigs expressing blood group A may eventually prove unsuitable as donors for individuals with anti-blood group A antibodies. Some natural antibodies that have been identified as potentially xenoreactive include antibodies specific for sulfatide (SO4–3Gal-R) (19–21), anti-Pk antibodies, and anti-T antibodies. However, it is
78
Parker, Yu, and Nakamura
unknown whether any of these antibodies are able to recognize intact porcine tissue. Anti-Pk antibodies and anti-T antibodies are not found in all human sera. Furthermore, given the strong negative charge of sulfatide, it is not clear whether binding to that antigen is “specific” or not (see discussion of polyreactive xenoreactive antibodies below). Thus, although antibodies specific for these carbohydrates might potentially play a role in xenograft rejection in some individuals, there are a number of issues that must be addressed before this can be determined.
EVIDENCE OF THE PRESENCE OF XENOREACTIVE NATURAL ANTIBODIES SPECIFIC FOR ANTIGENS OTHER THAN GAL_1–3GAL One of the primary approaches used to survey for the presence of xenogeneic antigens is Western blotting. This method has a number of limitations, since it cannot detect many conformationally dependent epitopes or epitopes on glycolipids. Furthermore, studies using Western blotting may identify many antigens that are not exposed on the cell surface and that are, therefore, irrelevant to xenotransplantation. With these limitations in mind, Western blot analyses performed in a number of laboratories have detected binding of human natural antibodies to antigens other than Gal_1–3Gal (22–25). However, most of the antibodies not specific for Gal_1–3Gal, which bind to bands on Western blots, are not adsorbed by a porcine organ (22,23,25), indicating that either the antigens are not expressed in large quantity on the endothelial surface of the organ and/or that there are too many antibodies to be adsorbed. Another approach that has been used to detect xenogeneic antigens is an enzyme-linked immunosorbant assay (ELISA) using cultured endothelial cells as the target (26). Using this approach, the presence of natural antibodies not specific for Gal_1–3Gal is reproducibly observed. Some, but not all, of these antibodies are adsorbed by porcine organs, indicating that there are at least some xenoreactive natural antibodies that are not specific for Gal_1–3Gal but which are xenoreactive. More important, we have observed that, in baboons, the amount of antibodies binding to antigens other than Gal_1–3Gal increases dramatically following exposure to porcine organs (27). Analysis by both ELISA and Western blotting suggests that the natural antibodies not specific for Gal_1–3Gal bind to the same targets as the elicited antibodies not specific for Gal_1–3Gal (27). The finding that xenoreactive natural antibodies have the same specificity as elicited antibodies suggests that the
Xenoreactive Natural Antibodies
79
natural antibodies do, indeed, recognize antigens present in a porcine organ and that these antigens, like Gal_1–3Gal, are recognized by the primate immune system. However, the nature of the targets, and thus the specificity of the antibodies, remains unknown.
XENOREACTIVE ANTIBODIES AND POLYREACTIVITY Polyreactive antibodies are antibodies that have a broad reactivity, binding to a number of unrelated antigens, such as dinitrophenol, and a variety of autoantigens, including DNA, cytoskeletal proteins, such as actin and tubulin, and serum proteins, such as IgG and thyroglobulin. Thus, one test for polyreactivity is to determine whether or not an antibody has the ability to bind to a variety of antigens, such as those mentioned above. It has recently been reported that anti-Gal_1–3Gal antibodies are actually polyreactive (28). However, all of the reagents used to test for polyreactivity (DNA, actin, myosin, and tubulin) were apparently derived from cows or pigs, and, thus, all of the observed reactivity may have been due to Gal_1–3Gal modifications on xenogeneic glycoproteins or to contamination of preparations with Gal_1– 3Gal. Similar problems are evident in early reports, suggesting that most xenoreactive antibodies are polyreactive (29). By definition, reactivity of antibodies with Gal_1–3Gal does not indicate that an antibody is polyreactive, regardless of the ubiquitous nature of Gal_1–3Gal. Turman, Platt and colleagues provided the first evidence that polyreactive antibodies might be xenoreactive when they identified human monoclonal antibodies that were polyreactive and that bound to porcine antigens, as judged by Western blotting (29). These human antibodies were found to bind to a variety of autoantigens, in addition to xenogeneic antigens, confirming that the antibodies were, indeed, polyreactive. An analysis of the binding of the monoclonal polyreactive antibodies to porcine antigens immobilized on Western blots (29) reveals that the antibodies do not apparently bind to antigens recognized by antiGal_1–3Gal antibodies. This observation indicates that the monoclonal polyreactive antibodies evaluated by Turman were not specific for Gal_1–3Gal in their recognition of porcine antigens. In further work by Platt’s laboratory, Geller et al. found that an anti-idiotype that recognized a monoclonal polyreactive antibody bound in a similar pattern as IgM in porcine tissues obtained from pig-to-rhesus monkey and pig-to-baboon xenografts (30). These studies provided initial support for the idea that polyreactive antibodies may be deposited in xenogeneic organ grafts.
80
Parker, Yu, and Nakamura
Additional findings supporting the idea that some xenoreactive antibodies might be polyreactive have been reported by the author, in collaboration with Platt, using an ELISA with cultured porcine cells as a target. In the human serum tested, the binding of natural xenoreactive IgM not specific for Gal_1–3Gal was inhibited by human insulin (10%), human Fc (40%), human albumin (40%), and human thyroglobulin (60%) (31). These observations suggest that at least some xenoreactive antibodies are polyreactive. However, it was not determined whether or not the antibodies, whose binding could be inhibited by various autoantigens, were the same as those antibodies that are adsorbed by a porcine organ. Further indication or evidence that some xenoreactive antibodies might be polyreactive is provided by the finding that human antibodies bind to sulfatide from pig kidneys, because sulfatide is a highly negatively charged molecule that might be the target of polyreactive or “nonspecific” antibodies (19). However, the exact nature of the anti-sulfatide antibodies, and whether or not they will bind to an intact porcine organ, remains in question. The concentration of antibodies that are not specific for Gal_1–3Gal provides further indication that at least some of these antibodies might be polyreactive. Our studies have revealed that the xenoreactive antibodies not specific for Gal_1–3Gal are more abundant in serum than are anti-Gal_1–3Gal antibodies, even though the amount of antigen on the porcine cell surface is rather low. Evidence of this is seen in the binding of human xenoreactive antibodies not specific for Gal_1–3Gal as a function of serum concentration (Fig. 2). The binding of these antibodies is greater at very low serum concentrations than is the binding of anti-Gal_1–3Gal antibodies, suggesting that these antibodies are several times more prevalent in human serum than are anti-Gal_1–3Gal antibodies. This concentration, probably greater than 10% of total serum immunoglobulin, is consistent with polyreactive antibodies. Furthermore, the functional avidity of these antibodies is consistent with polyreactive antibodies, being lower than that of anti-Gal_1–3Gal antibodies (see Chapter 5). However, there may be a large number of specific targets on porcine cells that are recognized by a wide variety of specific, but low affinity, antibodies, accounting for the high concentration of these antibodies. Regardless of whether or not xenoreactive antibodies that are not specific for Gal_1–3Gal are polyreactive, they do, apparently, com-
Xenoreactive Natural Antibodies
81
Fig. 2. The binding of human xenoreactive natural antibodies to Gal_1–3Gal and to other determinants on cultured porcine aortic endothelial cells. The binding of human xenoreactive IgM to cultured porcine aortic endothelial cell (circles) and to cultured porcine aortic endothelial cells from which Gal_1–3Gal had been eliminated by cleavage with _-galactosidase (squares) was determined by ELISA. Most binding was eliminated by cleavage of the Gal_1–3Gal epitope, indicating that most of the bound antibodies were specific for this epitope. However, the level of binding to cells lacking Gal_1–3Gal was maximum even at low concentrations (< 10%) of serum, suggesting that the antibodies which bound to uncharacterized determinants (not Gal_1–3Gal) were relatively abundant in the serum. This experiment was previously reported by Collins et al. (36).
prise the majority of xenoreactive antibodies. Thus, there may be a very large number of natural antibodies that must be eliminated or blocked before long-term xenograft survival can be achieved. However, the concentration of antigen recognized by antibodies not specific for Gal_1– 3Gal, especially in an intact organ, is unknown. Furthermore, the biological role of these antibodies also remains unknown.
RARE OR UNCOMMON XENOREACTIVE NATURAL ANTIBODIES In the course of many years of using human blood products in medicine, a number of rare antibodies have been identified that may cause a pathologic reaction if the recipient is exposed to certain allogeneic blood
82
Parker, Yu, and Nakamura
products. Although evidence that similarly rare xenoreactive antibodies exist is sparse, we will examine some recent findings. Perhaps the most striking evidence that some individuals may have rare, preformed xenoreactive antibodies comes from independent observations published in 1998 by Taylor and colleagues (32) and by Naziruddin and colleagues (33). Taylor found that high levels of anti-HLA antibodies, in patients sensitized to HLA, were associated with high levels of nonanti-Gal_1–3Gal antibodies that were reactive with pig lymphocytes. These antibodies were not found in the eight normal patients that were tested, suggesting that some patients, who might be especially likely to receive a xenograft (patients sensitized to HLA), will have xenoreactive antibodies other than anti-Gal_1–3Gal antibodies. Naziruddin (33) tested purified anti-HLA antibodies and demonstrated that the antibodies bound to porcine lymphocytes. They concluded that anti-HLA antibodies present in the sera of sensitized individuals reacts with swine leukocyte antigens. It is not known whether or not a subset of the normal population, which also has high anti-HLA titers, might have these xenoreactive antibodies not specific for Gal_1–3Gal. In most normal individuals, the isotype and subtype of anti-Gal_1– 3Gal antibodies are IgM and IgG2 (34,35), as might be expected for natural antibodies against a carbohydrate antigen. However, high levels of anti-Gal_1–3Gal IgG1 have been identified in about 5% of normal individuals tested (35). Although the specificity of the antibody is not unusual, the subtype apparently is. These findings have been confirmed by several techniques, including characterization of antibodies directly bound to antigen and characterization of antibodies purified by immunoprecipitation and by affinity isolation (35). Furthermore, the concentrations of xenoreactive IgM and xenoreactive IgG1, determined by these methods, account for the anti-porcine complement fixing activity in human sera, whereas the concentrations of xenoreactive IgM and total xenoreactive IgG alone do not account for the anti-porcine complement fixing activity in human sera (35). The serum of individuals with this uncommon subtype profile is much more cytotoxic to porcine cells than is the serum from other individuals, as might be expected since xenoreactive IgG1 is much more effective at complement activation than is xenoreactive IgG2 or even xenoreactive IgM (see Chapter 5). These experiments provide direct evidence that the effector function of human natural antibodies may be significantly impacted by uncommon xenoreactive antibodies. However, it is not known to what extent rare
Xenoreactive Natural Antibodies
83
antibodies not specific for Gal_1–3Gal might play a role in the rejection of xenografts.
ACKNOWLEDGMENTS The authors thank Sarah C. Hensel for expert assistance in preparation of the manuscript.
REFERENCES 1. Cooper DK. Xenoantigens and xenoantibodies. Xenotransplantation 1998; 5:6–17. 2. Platt JL. The immunological barriers to xenotransplantation. Crit Rev Immunol 1996;16: 331–358. 3. Galili U, Macher BA, Buehler J, et al. Human natural anti-_-galactosyl IgG. II. the specific recognition of _(1–3)-linked glactose residues. J Exp Med 1985; 162: 573–582. 4. Good AH, Cooper DKC, Malcolm AJ, et al. Identification of carbohydrate structures that bind human antiporcine antibodies: implications for discordant xenografting in humans. Transplant Proc 1992; 24: 559–562. 5. Galili U, Wang L, LaTemple DC, et al. The natural anti-Gal antibody. Subcell Biochem 1999; 32: 79–106. 6. Wieslander J, Mansson O, Kallin E, et al. Specificity of human antibodies against Gal_1–3Gal carbohydrate epitope and distinction from natural antibodies reacting with Gal_1–2Gal or Gal_1-4Gal. Glycoconj J 1990; 7:85–100. 7. Parker W, Lateef J, Everett ML, et al. Specificity of xenoreactive anti-gal_ 1-3gal IgM for a-galactosyl ligands. Glycobiology 1996; 6:499–506. 8. Teneberg S, Lonnroth I, Torres Lopez JF, et al. Molecular mimicry in the recognition of glycosphingolipids by Gal alpha 3 Gal beta 4 GlcNAc beta-binding Clostridium difficile toxin A, human natural anti alpha-galactosyl IgG and the monoclonal antibody Gal-13: characterization of a binding-active human glycosphingolipid, non-identical with the animal receptor. Glycobiology 1996; 6:599–609. 9. Neethling FA, Joziasse D, Bovin N, et al. The reducing end of _Gal oligosaccharides contributes to their efficiency in blocking natural antibodies of human and baboon sera. Transpl Int 1996; 9:98–101. 10. Galili U, Buehler J, Shohet SB, et al. The human natural anti-Gal IgG: III. The subtlety of immune tolerance in man as demonstrated by crossreactivity between natural anti-Gal and anti-B antibodies. J Exp Med 1987; 163:693–704. 11. Bouhours D, Liaigre J, Naulet J, et al. A novel glycosphingolipid expressed in pig kidney: Gal alpha 1–3Lewis(x) hexaglycosylceramide. Glycoconj J 1997; 14:29–38. 12. Kooyman DL, McClellan SB, Parker W, et al. Identification and characterization of a galactosyl peptide mimetic. implications for use in removing xenoreactive anti-A Gal antibodies. Transplantation 1996; 61:851–855. 13. Parker W, Stitzenberg KB, Yu PB, et al. Biophysical characteristics of anti-Gal_13Gal IgM binding to cell surfaces: implications for xenotransplantation. Transplantation 2001; 71:440–446. 14. Parker W, Lin SS, Yu PB, et al. Naturally occurring anti-_-galactosyl antibodies: Relationship to xenoreactive anti-_-galactosyl antibodies. Glycobiology 1999; 9:865–873.
84
Parker, Yu, and Nakamura
15. Bouhours JF, Richard C, Ruvoen N, et al. Characterization of a polyclonal antiGalalpha1–3Gal antibody from chicken. Glycoconj J 1998; 15:93–99. 16. Takagaki M, Knibbs RN, Roth J, et al. Monoclonal antibodies that recognize the trisaccharide epitope Gal alpha 1–3Gal beta 1–4GlcNAc present on Ehrlich tumor cell membrane glycoproteins. Histochemistry 1993; 100:139–147. 17. McMorrow IM, Comrack CA, Nazarey PP, et al. Relationship between ABO blood group and levels of Gal alpha,3Galactose-reactive human immunoglobulin G. Transplantation 1997; 64:546–549. 18. Parker W, Lundberg-Swanson K, Holzknecht ZE, et al. Isohemagglutinins and xenoreactive antibodies are members of a distinct family of natural antibodies. Hum Immunol 1996; 45:94–104. 19. Rydberg L, Bjorck S, Hallberg E, et al. Extracorporeal (ex vivo) connection of pig kidneys to humans. II. the anti-pig antibody response. Xenotransplantation 1996; 3:340–353. 20. Samuelsson BE, Rydberg L, Breimer ME, et al. Natural antibodies and human xenotransplantation. Immunol Rev 1994; 141:151–168. 21. Holgersson J, Cairns TD, Karlsson EC, et al. Carbohydrate specificity of human immunoglobulin-M antibodies with pig lymphocytotoxic activity. Transplant Proc 1992; 24:605–608. 22. Parker W, Bruno D, Holzknecht ZE, et al. Characterization and affinity isolation of xenoreactive human natural antibodies. J Immunol 1994; 153:3791–3803. 23. Lin SS, Parker W, Holzknecht ZE, et al. Quantitative evaluation of porcine endothelial cell antigens recognized by human natural antibodies: an analysis by western blotting. Xenotransplantation 1996; 3:120–127. 24. Thibaudeau K, Anegon I, Lemauff B, et al. Human natural antibodies to porcine platelets. Transplantation 1994; 57:1110–1115. 25. Vaughan HA, McKenzie IFC, Sandrin MS. Biochemical studies of pig xenoantigens detected by naturally occuring human antibodies and the galactose_(1–3)galactose reactive lectin. Transplantation 1995; 59:102–109. 26. Platt JL, Turman MA, Noreen HJ, et al. An ELISA assay for xenoreactive natural antibodies. Transplantation 1990; 49:1000–1001. 27. McCurry KR, Parker W, Cotterell AH, et al. Humoral responses in pig-to-baboon cardiac transplantation: implications for the pathogenesis and treatment of acute vascular rejection and for accommodation. Hum Immunol 1997; 58:91–105. 28. Satapathy AK, Ravindran B. Naturally occurring alpha-galactosyl antibodies in human sera display polyreactivity. Immunol Lett 1999; 69:347–351. 29. Turman MA, Casali P, Notkins AL, et al. Polyreactivity and antigen specificity of human xenoreactive monoclonal and serum natural antibodies. Transplantation 1991; 52:710–717. 30. Geller RL, Bach FH, Turman MA, et al. Evidence that polyreactive antibodies are deposited in rejected discordant xenografts. Transplantation 1993; 55:168–172. 31. Parker W, Yu PB, Holzknecht ZE, et al. Specificity and function of “natural” antibodies in immunodeficient subjects: clues to B-cell lineage and development. J Clin Immunol 1997; 17:311–321. 32. Taylor CJ, Tang KG, Smith SI, et al. HLA-specific antibodies in highly sensitized patients can cause a positive crossmatch against pig lymphocytes. Transplantation 1998; 65:1634–1641.
Xenoreactive Natural Antibodies
85
33. Naziruddin B, Durriya S, Phelan D, et al. HLA antibodies present in the sera of sensitized patients awaiting renal transplant are also reactive to swine leukocyte antigens. Transplantation 1998; 66:1074–1080. 34. Ross JR, Kirk AD, Ibrahim SE, et al. Characterization of human anti-porcine “natural antibodies” recovered from ex vivo perfused hearts*predominance of IgM and IgG2. Transplantation 1993; 55:1144–1150. 35. Yu PB, Holzknecht ZE, Bruno D, et al. Modulation of natural IgM binding and complement activation by natural IgG antibodies. J Immunol 1996; 157:5163–5168. 36. Collins BH, Parker W, Platt JL. Characterization of porcine endothelial cell determinants recognized by human natural antibodies. Xenotransplantation 1994; 1:36–46.
vi
Xenoractive Natural Antibodies
5
87
Biophysical Properties of Xenoreactive Natural Antibodies W illiam P ar ker, PhD, Ryan C. Fields, and Yuko C. Nakamur a, BS
BS, BA
INTRODUCTION Elucidating the biophysical properties of a given biomolecule is a critically important step to understanding the biological function of that molecule. Thus, given the medical importance of xenoreactive natural antibodies and other natural antibodies, some attention has been given to the biophysical properties of these molecules. One approach to studying the biophysical properties of antibodies is to use a purified monoclonal antibody and measure the parameters associated with the monovalent interaction of that antibody with an antigen. This approach is attractive because the antibody - antigen interaction may be precisely characterized both kinetically and thermodynamically. Unfortunately, the results of such studies have an unclear relationship to processes that actually occur in vivo, where binding phenomena are considerably more complicated. Another approach to studying the biophysical properties of xenoreactive antibodies is to characterize the binding of these antibodies to cell surfaces. Clearly, this binding is much more physiologically relevant than binding of antibodies to monovalent ligands. However, to some, studying the biophysical properties of antibody binding to cell surfaces might be considered a futile endeavor. Concerns arise because the standard approaches used to characterize the biophysical properties of biomolecules cannot readily be used to characterize the biophysical properties of natural antibodies and their targets. Yet, as will be discussed, such biophysical studies have provided considerable insight into the structure/function relationships of antibody binding to cell surfaces. From: Xenotransplantation: Basic Research and Clinical Applications Edited by: Jeffrey L. Platt © Humana Press Inc., Totowa, NJ
87
88
Parker, Fields, and Nakamura
In this chapter, we will consider the biophysical properties of antibody binding to cell surfaces and discuss the implications of those studies for the development of a physical model of antibody binding to cell surfaces. Before considering the biophysical nature of xenoreactive antibodies, the limitations of standard biophysical approaches when applied to xenoreactive natural antibodies and the approaches used to overcome those limitations will be considered.
APPROACHES TO STUDYING THE BIOPHYSICAL PROPERTIES OF XENOREACTIVE NATURAL ANTIBODIES The biophysical parameters associated with the binding of molecules to cell surfaces are generally not easily assessed because the cell surface itself adds a high degree of complexity. As an example, consider the binding constant (Keq), a quantity frequently used to describe the affinity of an intermolecular interaction. A binding constant can, under “normal” conditions (in solution), be determined by calculating the concentrations of unbound ligand [L], ligand - receptor complex [RL], and unbound receptor [R], since Keq= [RL]/ [R] [L] for the interaction R + L RL. This approach will not work on a cell surface because the volume of a cell surface cannot be readily determined and, thus, [R] and [RL] on a cell surface cannot be expressed in a useful unit of concentration (number of molecules per unit volume). One commonly used approach to circumvent this problem is to measure [L] when half of the total receptor is bound, i.e., when [R] = [RL]. Under these conditions, Keq = 1/[L], and the actual units associated with [R] and [RL] become unimportant. Although 1/[L] can be readily determined, it corresponds to a “functional Keq,” which has a relative or functional meaning with little or no absolute meaning, because factors such as the shape of the reaction vessel will affect the functional Keq. Thus, the functional Keq determined for the ligand–cell surface receptor interaction in flask X with volume Y is comparable to other functional Keq values determined in flask X with volume Y, but not with functional Keq values determined using any other container or volume. Yet another problem associated with determining affinity (Keq) is that antibody binding is multivalent. Such interactions are associated with an on-reaction (nR + LA RnL; n is the number of cell surface receptors bound by a single ligand), which is a multistep process that may be rather variable and complex. Thus, the well-known term “avid-
Xenoractive Natural Antibodies
89
ity,” rather than affinity, is applied to multivalent interactions, including the interactions of antibodies with cell surfaces. Because we do not determine the avidity in solution, the avidity of antibodies for a cell surface will be relative, depending on conditions, and, therefore, shall be referred to as a “functional avidity,” rather than an absolute avidity. There are other complexities associated with functional avidity, which derive from the fact that the binding of many antibodies, including xenoreactive antibodies, to cell surfaces is not actually represented by nR + L RnL. Instead, the reaction on a cell surface is better approximated by two equations (Fig. 1). The first, nR + LA RnL, is an RnL2 + L1 irreversible binding reaction. The second, RnL 1 + L2 (L1 and L2 are any two ligands), is a multistep pseudoequilibrium reaction involving exchange of free and bound antibody (ligand) on the
Fig. 1. Binding of ligand (antibody) to cell surface receptors (membrane bound antigens). The binding interactions can be approximated as two distinct reactions, one (Reaction #1) involving polyvalent and irreversible binding of ligand 1 (L ) to cell surface receptors (R), and the second (Reaction #2) involving dis1 2 placement of one ligand (L ) by a second ligand (L )
90
Parker, Fields, and Nakamura
cell surface. Evidence of these two processes, one irreversible and one psuedoequilibrium, is conclusive (1,2). However, the two reactions are only approximations, because R, L, and RnL are all highly heterogeneous species, because n might be somewhat variable, and because one or more R within a given Rn “cluster” might exchange with unbound R in equilibrium. Furthermore, as discussed below, some RnL may be formed in a reversible fashion, although these unstable RnL complexes may be difficult to detect under many conditions. These and other factors that prevent the determination of a true affinity (Keq), as well as true on and off rate constants (k1 and k-1) and true thermodynamic parameters (6G, 6H, and 6S), are shown in Table 1. Despite the impossibility of describing the binding of xenoreactive natural antibodies to a cell surface by a single binding constant or any other precise physical parameter, it is still possible to measure biophysical parameters associated with the binding of xenoreactive natural antibodies. These measures are useful in evaluating the structure/function relationship of the antibodies and the surfaces they bind to, as long as the limitations of these numbers and their actual meaning are kept in mind. For example, the “quality” of binding of xenoreactive natural antibodies to porcine cells can be assessed and is referred to as “functional avidity.” However, xenoreactive natural antibodies are a polyclonal population, and for that reason the term “functional avidity” should probably be modified to obtain a more descriptive term such as “composite functional avidity” or “apparent functional avidity.” Still, the term functional avidity is generally used and is, perhaps, not too misleading, given that xenoreactive anti-Gal_1-3Gal IgM behave as a relatively homogeneous species. [The same cannot be said of antiGal_1-3Gal IgG, however (3).] Despite the limitations, evaluating the biophysical properties of xenoreactive antibodies has provided considerable insight into the relationship between the structure and function of xenoreactive antibodies and the cell surfaces to which they bind.
ENERGETICS The change in free energy (6G) associated with binding is, of course, an indicator of the favorability of a given binding interaction, and is directly related to the binding constant (Keq) by the well-known equation 6G = -RT lnKeq. The 6G of a given reaction has two contributing components [the change in enthalpy (6H) and the change in entropy (6S)], which are related to 6G by the well-known Gibbs free energy
Xenoractive Natural Antibodies
91
Table 1. Properties of Monovalent Interactions versus Interactions of Xenoreactive Natural Antibodies with Cell Surfaces Binding interactions that can be defined by equilibrium and rate constants
Binding of xenoreactive antibodies to cell surfaces
Unique receptor
Receptor (antigen) highly variable: 1. Same epitope may be expressed in essentially and infinite number of environments 2. Various epitopes may be present
Unique ligand
Ligand (antibody) is polyclonal
Unique receptor ligand interaction
A virtually infinite variety of antibodyantigen interactions are possible
Assays readily available
No standard assay available
Binding is reversible
Binding is irreversible in some cases
Monovalent interactions
Polyvalent interactions: the same antibody may bind in different ways to the same epitopes
Determined using purified components
Purified components not generally available for several reasons: 1. Purified Ig M not very stable Frequent lack of sufficient material 2. for purification from serum 3. Difficulty in growing human hybridomas in media free of antigen (serum free medium)
equation (6G = 6H - T 6S). We will discuss these two components in regard to general principles of protein carbohydrate interactions, and then evaluate how these principles are reflected in the binding of antibodies to a cell surface. The change in enthalpy (6H) upon antibody binding to a monovalent ligand is a reflection of the “goodness of fit” of the antigen binding site and the ligand. This fit involves both hydrogen bonds and van der Walls interactions in protein - carbohydrate interactions, and it dictates the “specificity” of the antibody. In addition, when antibodies bind
92
Parker, Fields, and Nakamura
multivalently to cell surfaces, 6H is also affected by the overall geometry of the binding site, including the orientation and relative distance separating two or more ligands that are bound by different Fab regions of the same antibody. Perhaps the most striking thermodynamic feature of protein - carbohydrate interactions is the marked negative entropy change (6S) associated with the binding. The reason the change in entropy is highly positive, and thus unfavorable, is because the carbohydrate is highly flexible, having a great amount of torsional freedom, which results from the very low torsional potential energy for glycosidic linkages (4). Upon binding to protein, the constantly moving, poorly restrained (high entropy) carbohydrate is locked into a much more constrained (low entropy) state. A good fit of a particular carbohydrate in the binding pocket of an antibody (high enthalpy) will restrict the movement of that carbohydrate (low entropy) more so than will interaction with a poor binding pocket (low enthalpy, high entropy). Thus, there is an “entropy - enthalpy compensation” in the interaction of carbohydrate antigens with antibodies, as illustrated in Fig. 2. On one hand, the increasing enthalpy serves to favor the formation of a tight carbohydrate - antibody interaction. On the other hand, the decreasing entropy of such a tight fit favors the dissociation of the complex. The high entropy inherent in carbohydrate structures provides direct information regarding the conformation of those molecules in solution. Although the lowest-energy form is most probable, it generally represents less than 1% of the total oligosaccharide ensemble, as determined by nuclear magnetic resonance (NMR) . Apparently, changes in temperature from 0 to 80ºC do not affect the population of various conformational states (5). Thus, the carbohydrate antigens seen on the cell surfaces of porcine cells by anti-Gal_1-3Gal antibodies are highly flexible, and are expected to populate a large variety of conformations, only some of which may participate in binding. The high 6S associated with protein - carbohydrate interactions has been reported for natural anti-carbohydrate antibodies, including antiGal_1-3Gal antibodies. In vitro, at 4ºC, virtually all anti-Gal_1-3Gal antibodies will bind to a porcine cell, whereas, at 37ºC, only about 10% of those antibodies bind under conditions of antigen saturation (6). Of the unbound fraction (37ºC), about 10% will bind when incubated with a new set of porcine cells. This result suggests that the low amount of
Xenoractive Natural Antibodies
93
Fig. 2. The balance between entropy (6S) and enthalpy (6H) in protein - carbohydrate interactions. The entropy and enthalpy contribution to the free energy (-T6S and 6H, respectively) of binding for a number of protein - carbohydrate interactions is shown, with each point corresponding to a different protein carbohydrate interaction. All of the interactions chosen for evaluation in this analysis had a KD of about 3 µM (6G = -7.5 kcal/mol). The values of -T6S and -6H increase for larger ligands so that the points corresponding to protein binding to monosaccharides are at the lower left and the point corresponding to protein binding to tetrasaccharide appears at the upper right. The results shown in the graph were originally compiled from other sources and reported by Carver and colleagues (5).
binding at higher temperature is apparently due to lower avidity of all antibodies, and not to the presence of just a few higher avidity antibodies. Although neither 6G, 6H, nor 6S have been determined for these interactions (due to the complexity described above), there is clearly a dramatic decrease in entropy associated with binding. These studies are consistent with the observation that xenoreactive antibodies bind to very flexible structures such as disaccharides and trisaccharides on the cell surface.
94
Parker, Fields, and Nakamura
EQUILIBRIUM In general, divalent binding of antibody is 100- to 1000-fold stronger than monovalent binding. Valencies greater than 3, as is expected for the binding of xenoreactive IgM to a cell surface, result in irreversible binding. The importance of a measurement of avidity, or even the concept of equilibrium, is not immediately apparent if binding is irreversible (2) . One might even wonder how changes in avidity can be observed, given that binding is irreversible. The key to this apparent contradiction lies in the heterogeneity of binding interactions. Although xenoreactive IgM specific for Gal_1-3Gal are relatively homogeneous in terms of binding avidity (6), the total number of interactions that result in irreversible binding may change, depending on conditions. That is, as functional avidity decreases (i.e., as with an increase in temperature, as described above), the number of sites able to be bound irreversibly by xenoreactive natural antibodies decreases and/or the number of conformations in which a xenoreactive natural antibody can bind to a given site decreases. In this way, the “stringency of binding” is apparently affected, and some interactions that are irreversible under one condition are no longer irreversible under more stringent conditions. On the other hand, very favorable (highly avid) interactions would be irreversible under a wide range of conditions. Thus, divalent binding of antibody may be 1000-fold stronger than monovalent binding and further increases in valency result in irreversible binding. Perhaps surprisingly, even tighter binding than irreversible binding has been observed. This binding, which will be referred to as “biplaner” irreversible binding, apparently involves the binding of polyvalent ligands (i.e., antibodies) to cell surface receptors (i.e., transmembrane glycoproteins containing Gal_1-3Gal) that are (or that become) noncovalently cross-linked on the cytoplasmic side of the plasma membrane (Fig. 3). The distinguishing feature of biplaner irreversible binding is that free soluble ligand or cell surface receptor does not compete off bound ligand, even though interactions are noncovalent. Biplaner irreversibility has previously been observed when polyvalent antigens (the ligand) bind to sIg (the cell surface receptor) (7). The finding that anti-Gal_1-3Gal IgM participate in biplaner irreversible binding on porcine cells (8) suggests that these antibodies are binding to molecules that are cross-linked on the cytoplasmic side of the cell membrane. Such cross-linking might happen by attachment of cell surface antigens to cytoskelatal proteins or other cytoplasmic structures (Fig. 3). This idea is consistent with experimental data showing that the
Xenoractive Natural Antibodies
95
Fig. 3. “Biplaner” irreversible binding. Cross-linking of antigen antibody complexes in two planes (one above and one below the plasma membrane) might result in a type of binding which cannot be competed off by soluble ligand or by soluble receptor. (A) As postulated by Goldstein and Wofsy , irreversible binding that cannot be competed off by soluble ligand or soluble receptor might be a result of multivalent antigens binding to sIg which are noncovalently crosslinked on the cytoplasmic side of the plasma membrane. (B) In a similar fashion, binding of IgM to cell surface glycoproteins (integrins containing Gal_1-3Gal) cross-linked on the cytoplasmic side of the cell surface might result in irreversible binding which cannot be competed off by soluble ligand or soluble receptor. Although a bivalent antibody is shown in the diagram for the purpose of illustration, it is expected that multivalent antibodies (IgM) would be required to effectively participate in such binding.
major sources of Gal_1-3Gal on porcine cell surfaces are integrins (9,10). Integrins are known to interact with the cytoskeleton and are found on the surface of endothelial cells, including the apical surface, which is exposed to the blood stream under normal conditions (11).
96
Parker, Fields, and Nakamura
KINETICS Kinetic data for the interaction between antibodies and their respective carbohydrate antigens have concentrated on the oligosaccharide components of the cell membranes of pathogenic bacteria and the carbohydrates associated with viral coats. Studies have characterized these antibody - carbohydrate interactions with respect to the effect of modifying both the antigen (oligosaccharide length) and the antibody (Ig valency). Taken together, these studies help to paint a picture of how antibodies and carbohydrates come together to interact on the surface of antigenic substances. The interaction of IgG with a group B streptococcal antigen shows high association binding kinetics with a dissociation constant (Kd) of 1.1 × 10-8 (4,12). In particular, Hiroshi et al. found on-rates to have a dramatic dependence on carbohydrate chain length in the binding of IgG with di-, tetra-, and hexa-saccharides derived from pneumococcas (12). A decrease in the stability of the antibody - oligosaccharide complex with decreasing chain length was observed. This decrease was accounted for by the difference in association rate constants, which were 1.7×104 M-1s-1 for the di-, 3.7×105 M-1s-1 for the tetra-, and 1.1×106 M-1s-1 for the hexa-saccharide. All three dissociation constants were of the same magnitude, suggesting a two-step binding process in which the dissociation of the intermediate product is much faster than the conversion of the intermediate product to the final antibody - antigen complex. The effect of valency on the binding kinetics of antibodies to carbohydrates has also been evaluated. The interaction of IgG with a Salmonella surface polysaccharide shows that the dimeric forms of IgG interact with the carbohydrate antigen more tightly than the monomeric form (13). This is demonstrated by the 20-fold slower off-rate of the dimeric form in contrast to the monomeric form. Furthermore, the dimeric form showed a five-fold increase in association rate constants, which also contributes to the better dimeric IgG - antigen interaction. These results suggest that the affinity gain obtained from going to a divalent antibody is not simply the product of the individual monomeric rate constants, as has been proposed (13). In this study, the increase in association rate was associated with the tendency of IgG to aggregate when interacting with the multivalent carbohydrate antigen. Taken together, the decrease in off-rate and increase in association rate translates into a 100-fold increase in functional affinity for the divalent form.
Xenoractive Natural Antibodies
97
Our own studies on the binding kinetics of anti-Gal_1-3Gal IgM to cell surfaces reveal that the on-rate of binding is first order with respect to antibody concentration, and fourth to fifth order with respect to antigen concentration (14). These studies indicate that the “typical” binding reactions occurring on the cell surface are 4R + LAR4L and 5R + LAR5L, as might be expected for the binding of IgM. Our studies on the effect of temperature on k1 show that the effect is complex, increasing the functional k1 from 0 to 15ºC and decreasing the functional k1 from 15 to 37ºC (15). Studies of other antibody - carbohydrate systems on the effect of temperature on kinetics are lacking. Further kinetic studies indicate that on-rate is dependent on viscosity, and is, therefore, apparently diffusion limited (15). This observation suggests that the kinetics of antibody binding may be very different in vivo and in vitro, as rapid mixing and a large surface area in vivo may facilitate much more rapid binding. Studies of the equilibrium and kinetics of xenoreactive natural antibodies binding to porcine cells has resulted in several observations, not described above, which have potential medical importance. For example, apparent loss of components of the glycocalyx, a condition that may occur when endothelial cells become activated, has been seen to increase the functional avidity and k1 of antibody binding to porcine cells. This observation may indicate that initial damage to a particular organ, by reperfusion or other injury, might increase the effectiveness of subsequent antibody-mediated immunity. This observation also has some technical implications, as fixation with various reagents is known to cause shedding of glycocalyx, and as antibodies are known to prevent shedding of glycocalyx by cross-linking the glycocalyx noncovalently (8) . As another example of the potential medical importance of biophysical studies of xenoreactive natural antibodies, anti-Gal_1-3Gal antibodies have been found to bind more avidly than other anti-porcine antibodies that have been characterized (16). This observation might suggest that anti-Gal_1-3Gal antibodies are more cytotoxic toward porcine cells than are other xenoreactive antibodies.
FUNCTION: INTERACTION WITH C1Q COMPONENT OF C1 COMPLEX Since the primary effector function of Ig in xenotransplantation is apparently the activation of complement (as discussed in Chapter 3), the interactions of antibody with complement will be considered in some detail. Activation of the complement pathway requires the binding of
98
Parker, Fields, and Nakamura
the C1 complex to antibody. The binding of C1 complex and antibody is the first “committed” step in complement activation. Thus, an understanding of the properties of this interaction is necessary to characterize the complement activation process. The C1 complex consists of a C1q molecule bound to one C1r and one C1s. The C1q molecule is itself made up of six subunits. Each subunit has a collagen-like tail, responsible for binding to C-reactive protein, and a globular head that binds the Fc portion of antibodies. Each head binds to one Fc domain. Binding of at least two heads of the subunits that make up the C1q molecule activates C1q, leading to activation of the classical complement cascade. Plasma IgM is a planar, pentameric molecule that doesn’t bind to C1q. However, the binding of IgM to a pathogen causes a conformational change in the IgM molecule from the planar form to a “staple” form, exposing the Fc binding sites to C1q. Contrastingly, C1q binds with low affinity to certain subclasses of IgG in solution. In order to activate C1q via IgG, at least two IgG molecules must bind C1q with the two IgG molecules bound to separate antigens on a membrane surface within 30 - 40 nm of each other. This requires many IgG molecules to bind to a given pathogen. The result is that IgM, in theory, can activate the classical complement pathway much better than IgG. However, several studies have shown that C1q is responsible for binding to IgM and that binding is poor between the C1 complex and IgM (17–19). For example, Weiner determined that the binding of C1q on a solid phase to IgM happens with a K on the order of 3×108 M-1 (17). Using the same approach, Weiner was unable to measure any binding of C1 to IgM due to a 107 M-1 cutoff in the sensitivity of the ELISA test. Another study showed that the binding of IgM to C1q occurred with a K of 4.0×108 M-1 , although no binding to C1 could be detected (17). Further work has demonstrated that some binding of IgM to C1 that has been identified can actually be accounted for by binding of the IgM to C1q, which has dissociated from the C1 complex (17,18). In contrast to IgM, it has been found that IgG immune complexes bind both C1q (K about 107 M-1) (20,21) and C1 (K about 3×107 M-1) (22) . Since C1, not C1q, is the biologically active molecule, these studies indicate that IgG is a better activator of complement than is IgM. Predictions of the relative complement fixing activity of IgG and IgM appear to hold true for xenoreactive antibodies. Yu and colleagues found that IgG1, the subtype of IgG elicited by exposure to xenogeneic antigen (23) , activated complement about five times better than did IgM (24).
Xenoractive Natural Antibodies
99
This finding has particular impact on xenotransplantation, as any elicited antibodies (predominantly IgG) that might be produced may be more damaging to a porcine organ than are natural antibodies [predominantly IgM and, in some individuals, IgG2 (24,25), a subtype of IgG that does not activate complement well compared to IgG1].
CONCLUSIONS There is still much work to be done regarding the biophysical properties of natural antibodies. One might argue that such an endeavor is not worthwhile, because, as has been described, the very nature of the antibodies and the antigens they recognize precludes precise characterization, a hallmark of modern biophysics and biochemistry. However, we would argue a different point of view, for although biophysical studies on nonideal and heterogeneous systems are certainly more challenging than studies on simpler systems, the information gained is more biologically relevant and cannot be adequately modeled by a simple system in many cases. Our own studies of xenoreactive natural antibodies binding to cell surfaces have provided insight into the nature of the antigens recognized, and have supported and augmented other biochemical studies. However, the work has been conducted using cultured cells, and, thus, the ideas and physical models derived from these experiments will need to be critically compared with results obtained in vivo. Regardless of the future role of xenotransplantation in medicine, the biophysical properties of xenoreactive natural antibodies are of great importance, since the biological relevance of xenoreactive natural antibodies and other natural antibodies extends beyond the field of xenotransplantation. For example, still poorly characterized, natural humoral immunity is of great medical importance, relating to a variety of conditions including inflammatory bowel disease and various conditions of sepsis. Thus, the careful characterization of the biophysical properties of xenoreactive natural antibodies and other natural antibodies is an important goal, despite the limitations inherent in the study of such a complex system.
ACKNOWLEDGMENTS The authors thank Sarah C. Hensel for expert assistance in preparing this manuscript.
100
Parker, Fields, and Nakamura
REFERENCES 1. Kyriakos RJ, Shih LB, Ong GL, et al.: The fate of antibodies bound to the surface of tumor cells in vitro. Cancer Res 1992; 52:835–842. 2. Ong GL, Mattes MJ. Re-evaluation of the concept of functional affinity as applied to bivalent antibody binding to cell surface antigens. Mol Immunol 1993; 30:1455–1462. 3. Parker W, Lin SS, Yu PB, et al. Naturally occurring anti-_-galactosyl antibodies: Relationship to xenoreactive anti-a-galactosyl antibodies.Glycobiology 1999;9:865–873. 4. Feldman RG, Breukels MA, David S, et al. Properties of human anti-group B streptococcal type III capsular IgG antibody. Clin Immunol Immunopathol 1998; 86:161–169. 5. Carver JP, Michnick SW, Imberty A, et al. Oligosaccharide-protein interactions: a three-dimensional view. CIBA Found Symp 1989; 145:6–18. 6. Parker W, Bruno D, Holzknecht ZE, et al. Characterization and affinity isolation of xenoreactive human natural antibodies. J Immunol 1994; 153:3791–3803. 7. Goldstein B, Wofsy C:. Why is it so hard to dissociate multivalent antigens from cell-surface antibodies? Immunol Today 1996; 17:77–80. 8. Parker W, Holzknecht ZE, Song A, et al. Fate of antigen in xenotransplantation: implications for acute vascular rejection and accommodation. Am J Pathol 1998; 152:829–839. 9. Platt JL, Lindman BJ, Chen H, et al. Endothelial cell antigens recognized by xenoreactive human natural antibodies. Transplantation 1990; 50:817–822. 10. Lin SS, Parker W, Holzknecht ZE, et al. Quantitative evaluation of porcine endothelial cell antigens recognized by human natural antibodies: an analysis by western blotting. Xenotransplantation 1996; 3:120–127. 11. Conforti G, Dominguez-Jimenez C, Zanetti A, et al. Human endothelial cells express integrin receptors on the luminal aspect of their membrane. Blood 1992; 80:437–446. 12. Maeda H, Schmidt-Kessen A, Engel J, et al. Kinetics of binding of oligosaccharides to a homogeneous pneumococcal antibody: dependence on antigen chain length suggests a labile intermediate complex. Biochemistry 1977; 16:4086–4089. 13. MacKenzie CR, Hirama T, Deng SJ, et al. Analysis by surface plasmon resonance of the influence of valence on the ligand binding affinity and kinetics of an anticarbohydrate antibody. J Biol Chem 1996; 271:1527–1533. 14. Lin SS, Parker W, Everett ML, et al. Differential recognition by proteins of a-galactosyl residues on endothelial cell surfaces. Glycobiology 1998; 8:433–443. 15. Parker W, Stitzenberg KB, Yu PB, et al. Biophysical characteristics of anti-Gal_13Gal IgM binding to cell surfaces: implications for xenotransplantation. Transplantation 2001; 71:440–446. 16. Parker W, Yu PB, Holzknecht ZE, et al. Specificity and function of “natural” antibodies in immunodeficient subjects: clues to B-cell lineage and development. J Clin Immunol 1997; 17:311–321. 17. Weiner EM. On the interaction of the first complement component C1 and its subunit C1q with solid-phase IgM immune complexes. Scand J Immunol 1988; 28:425–430.
Xenoractive Natural Antibodies
101
18. Taylor B, Wright JF, Arya S, et al. C1q binding properties of monomer and polymer forms of mouse IgM mu-chain variants. Pro544Gly and Pro434Ala. J Immunol 1994; 153:5303–5313. 19. Kimak E, Tomaszewski JJ. Interaction of C1q subcomponent with immunoglobulin M. Acta Biochim Pol 1986; 33:179–185. 20. Hughes-Jones NC:. Functional affinity constants of the reaction between 125Ilabelled C1q and C1q binders and their use in the measurement of plasma C1q concentrations. Immunology 1977; 32:191–198. 21. Hughes-Jones NC, Gorick BD, Howard JC, et al. Antibody density on rat red cells determines the rate of activation of the complement component C1. Eur J Immunol 1985; 15:976–980. 22. Hughes-Jones NC, Gorick BD. The binding and activation of the Clr-Cls subunit of the first component of human complement. Mol Immunol 1982; 19:1105–1112. 23. Yu PB, Parker W, Everett M, et al. Immunochemical properties of anti-Gala1-3Gal after sensitization with xenogeneic tissues. J Clin Immunol 1999; 19:116–126. 24. Yu PB, Holzknecht ZE, Bruno D, et al. Modulation of natural IgM binding and complement activation by natural IgG antibodies. J Immunol 1996; 157:5163–5168. 25. Ross JR, Kirk AD, Ibrahim SE, et al. Characterization of human anti-porcine “natural antibodies” recovered from ex vivo perfused hearts*predominance of IgM and IgG2. Transplantation 1993; 55:1144–1150.
vi
Origin of Xenoreactive Natural Antibodies
6
103
The Origin of Xenoreactive Natural Antibodies Paul B. Yu, MD, PhD INTRODUCTION
Xenoreactive natural antibodies, being a subset of natural antibodies, likely have a conserved role in primordial immunity. Natural antibodies were first described by Landsteiner, who showed that the serum of normal subjects contained substances that would agglutinate and lyse erythrocytes from other individuals or species without prior sensitization (1). Natural antibodies also react with various bacteria, leading to opsonization and complement-mediated destruction (2–4). The prevalence of natural antibodies and reactivity with various microbes led to the idea that natural antibodies may form an essential component of innate immunity against invasive bacteria (5–7). As the expression of natural antibodies is conserved in phylogeny (8–10), the function of natural antibodies in host defense may represent a form of primitive immunity. In addition to a direct role in host defense, natural antibodies are thought to facilitate secondary immune responses, to perform surveillance for neoplastic cells, to help clear senescent cells, and to regulate autoimmunity (6,11–14). Xenoreactive natural antibodies would also appear to present an immune barrier to viral transmission and organ transplantation across species (15,16). Natural antibodies appear to exist in several general types (Table 1). The first natural antibodies discovered were the isohemagglutinins, which recognize blood group carbohydrates on erythrocytes, causing transfusion reactions between incompatible individuals without previous sensitization (1). Related to the isohemagglutinins are natural antibodies that recognize microbial macromolecules such as phosphatidylcholine, phosphorylcholine, and lipoteichoic acid (7,17), From: Xenotransplantation: Basic Research and Clinical Applications Edited by: Jeffrey L. Platt © Humana Press Inc., Totowa, NJ
103
104
Yu Table 1. The Representation of Natural Antibodies in the Repertoire
Natural antibody
Percentage of total Ig
Reference
Human isohemagglutinins Anti-Gal_1–3Gal (human) Polyreactive Ab (human) Polyreactive Ab (mice)
IgM 0.5–1.0 1–4 33–50 ~50
(23) (23,31,32) (6,30) (33,34)
IgG 0.25–0.5 0–1.0
and natural antibodies that recognize bacterial polysaccharides or carbohydrates (18–20). Natural antibodies such as these may confer protective immunity against some bacteria and parasites or may limit their pathogenicity until the development of acquired immunity (7,17,21,22). A well-studied example of a natural anti-carbohydrate antibody are the human antibodies to Gal_1–3Gal, which appear to be closely related to the isohemagglutinins (23) and which mediate the rejection of xenografts from lower mammals (24–26). Yet another class of natural antibody is the “polyreactive” antibody, which binds to multiple structurally distinct ligands, such as heat shock proteins, nucleic acids, or cytoskeletal proteins, and may thus be autoreactive (27–28). Polyreactive natural antibodies are thought to have a fundamental role in host defense and immune regulation (6,29,30). Natural antibodies constitute a substantial fraction of the total Ig. Isohemagglutinins and natural anti-Gal_1–3Gal antibodies each account for approx 1% of the IgM, and as much as 0.5–1% of the IgG in humans (32,35) by some reports. Polyreactive natural antibodies may account for up to 50% of the IgM in normal humans, and perhaps an equal fraction of the IgM in mice (6,30,33,34). Polyreactive natural antibodies account for significant fractions of the circulating IgM in lower species such as the trout and shark (8–10), suggesting a function in primitive immune defense that has been retained in higher animals. Xenoreactive natural antibodies, defined as the naturally occurring human antibodies that are found to react with tissues of other species, are comprised of a continuum of antibody specificities. This continuum ranges from natural anti-carbohydrate antibodies with apparent monospecificity, to polyreactive natural antibodies that recognize diverse xenoantigens and autoantigens. The origin of polyreactive natural antibodies, some of which are xenoreactive and may thus have important roles in xenotransplantation, is a subject of extensive research by other
Origin of Xenoreactive Natural Antibodies
105
groups that is summarized in this chapter. The primary focus of this chapter, however, is the origin of antibodies directed against the terminal carbohydrate Gal_1–3Gal (Figs. 1 and 2), the primary antigenic target of human xenoreactive natural antibodies against the tissues of New World monkeys and lower mammals, including the pig (26).
CELLULAR ORIGIN OF NATURAL ANTIBODIES AND NATURAL ANTIBODIES TO GALI1–3GAL A central question has been the cellular origin of natural antibodies. An identified source of many natural antibodies are B cells of the B1 lineage, an apparently primordial subset of B cells originally identified by the CD5 surface marker (36). B1 cells include both the CD5+ B1a population and a “sister” CD5– B1b population with similar functional properties (37). B1 B cells have been called “the carriers of natural immunity” because they produce up to 80–100% of serum IgM and 50% of the serum IgA and because they are the source of polyreactive natural antibodies (38–40). B1a cells are the major type of B cells in peritoneal and mucosal lymphoid tissues, but are rare in the spleen and peripheral lymph nodes, including germinal centers, and absent in the bone marrow (41–42). The localization of B1 cells to the mucosa suggests that polyreactive natural antibodies, particularly secretory IgA, might function to exclude pathogens from the body or neutralize their products (6,43,44). B1 cells account for >50% of peripheral and >90% of splenic B cells in human neonates (45), suggesting a role of B1 cells in host defense before maturation of specific immunity. A limitation of studies aiming to study the B1 or primordial B-cell subset has been their phenotypic identification, and it appears that expression of the CD5 surface marker on B cells may in some instances reflect activation or inducible differentiation than a separate lineage (46,47). The expression of CD5 on B cells is neither necessary nor sufficient for identifying a primordial B-cell subset, but the weight of many studies on the B1 cell subset strongly supports a primordial B-cell origin of polyreactive natural antibodies. Following these studies, it was proposed that xenoreactive natural antibodies are also synthesized by B1 or primordial B cells (48). This appeared to be true to the extent that some xenoreactive natural antibodies are polyreactive, supported by the observations that 1. The binding of some xenoreactive natural antibodies was blocked by multiple antigens, including thyroglobulin, a common ligand of polyreactive antibodies.
106
Yu
Fig. 1. Structure of Gal_1–3Gal_1–GlcNAc.
Fig. 2. Structures of Gal_1–3Gal, blood group A and B antigens.
2. Some antibodies synthesized by transformed and cloned B1 B cells bind to xenogeneic endothelial cells. 3. At least one idiotype of such a polyreactive antibody is detectable in the immune deposits of some rejecting xenografts (49). On the other hand, we found that the xenogeneic carbohydrate epitope Gal_1–3Gal is recognized primarily by monospecific antibodies (50) , making a B1 origin in the case of this antibody less likely. Pitre et al.
Origin of Xenoreactive Natural Antibodies
107
reported that both B1 B cells and B2 B cells bound to porcine thyroglobulin, a glycoprotein bearing Gal_1–3Gal substitutions (51). A limitation of this report was that some but not all B cells that bound this glycoprotein produced immunoglobulin specific for Gal_1–3Gal, leaving unanswered the question of whether anti-Gal_1–3Gal antibodies are produced by B1 B cells, B2 B cells, or both. Current data indicate that natural antibodies recognizing carbohydrates are not exclusive products of the B1 B cell subset. Human natural antibodies against pneumococcal polysaccharides appear to be produced by CD5–B2 cells (52). Both CD5+ and CD5- peripheral B cells without bias may produce anti-carbohydrate natural antibodies specific for blood group A and B antigens based upon flow cytometry purification and an ELISAspot immunoassay (53), although such studies could not resolve potential differences in subclass or affinity of the detected antibodies. Our own studies support the idea that Gal_1–3Gal is potentially recognized by either CD5+ or CD5- peripheral B cells without bias (54). We analyzed peripheral CD19+ B cells that may represent precursor and memory populations of B cells that recognize natural antibodies to Gal_1–3Gal. Single Ag-specific B cells were sorted by FACS using a fluor-labeled neoglycoconjugate. The supernatants of sorted B cells cocultured with CD40L-bearing cells and IL-2 were assayed for antiGal_1–3Gal production to verify specificity of the isolation. To test the diversity of B cells specific for Gal_1–3Gal, we evaluated the fine specificity by competition of Ag binding with various oligosaccharides, identifying a population with narrow specificity for Gal_1–3Gal. To determine the compartment of Ag-specific cells, the coexpression of CD5 as a marker for the B1 lineage, of sIgG as a marker of memory B cells. Using stringent controls for specificity, we found the frequency of B cells in the periphery specific for Gal_1–3Gal were significantly less than the 1% previously reported (55), ranging from 0.1% to 0.6% of circulating B cells. Antigen-specific B cells lacked significant bias in expression of CD5 or VH3 heavy chains compared to unsorted B cells. However, CD5+ B cells specific for Gal_1–3Gal appeared to have less avidity for antigen compared to CD5- cells, suggesting functional differences between these populations. The heterogeneity of the Ag-specific B cells identified by these studies is consistent with multiple populations of B cells that may recognize Gal_1–3Gal, with no exclusive source of anti-Gal_1–3Gal antibodies. Corroborating these findings, some in-vivo insights into the cellular origin of anti-Gal_1–3Gal and other anti-carbohydrate antibodies were
108
Yu
gained in our studies on the expression of natural anti-Gal_1–3Gal antibodies in immunodeficient individuals (56). Individuals with the hyper-IgM phenotype, who are apparently intact in the expression of polyreactive natural antibodies, had in more than half of cases profound defects in the ability to synthesize anti-Gal_1–3Gal, indicating that the B cell origin of anti-Gal_1–3Gal is not equivalent to that of polyreactive natural antibodies. Anti-Gal_1–3Gal and other anti-carbohydrate natural antibodies thus do not derive exclusively from a primordial B cell subset, and may originate from more “conventional” B cells such as the B2 subset.
THE Ig GENES ENCODING XENOREACTIVE NATURAL ANTIBODIES The immunogenetics of the natural antibody response to xenogeneic antigens may be a relevant issue in attempts to modulate this response in cross-species transplantation. The immunogenetics of polyreactive natural antibodies and natural anti-Gal_1–3Gal antibodies appear to be quite distinct. Although conventional antibodies derive from diverse and mutated Ig genes, polyreactive natural antibodies are generally encoded by germline Ig genes or genes less mutated than those encoding monospecific antibodies (30). Consistent with this idea, human peripheral B1a cell heavy and light chains are significantly less mutated than B2 cells (57,58). The low frequency of mutation in natural polyreactive antibodies may reflect unique properties of B1 B cells, such as the fact that B1 cells typically respond to antigens independently of T cells (59,60). Polyreactive natural antibodies expressed by the B1 cells thus appear to express a mostly preimmune humoral repertoire. It should be noted that germline Ig usage is not equated with polyreactivity, however, as both germline monoreactive Ig and mutated polyreactive Ig exist (30) . The use of germline Ig genes by many B1 cells producing polyreactive natural antibodies led to the proposal that the structural determinants of polyreactivity are encoded by particular Ig variable (V) genes. However, several observations weigh against this proposal. Human hybridomas producing polyreactive antibodies utilize the same diverse V genes as hybridomas producing monospecific antibodies (30). It is also found that peripheral B1a cells utilize the same repertoire of heavy and light chain variable genes as peripheral B2 B cells (57,58). Polyreactivity may instead be determined by particular Ig rearrangements. The large binding grooves encoded by the third complementarity-determining
Origin of Xenoreactive Natural Antibodies
109
region (CDR3) in some polyreactive antibodies was proposed to account for polyreactivity (30), but this feature is not universal (57,58,61). Alternatively, polyreactivity might reflect the presence of positively charged residues in the CDR3 that have affinity for negatively charged molecules (30), but this is also not universal . Although the structural features determining polyreactivity are not known, it is likely that they are specified by CDR3, which is felt to be the major determinant of specificity in general (62). If the genetic basis for the polyreactivity of some natural antibodies is not clear, polyreactive antibodies do clearly share some cross-reactive idiotypes (63,64), indicating the presence of some shared structural motifs. The Ig gene usage of anti-Gal_1–3Gal antibodies is polyclonal and likely mutated rather than germline, resembling that of conventional antibodies. The Ig gene diversity of anti-Gal_1–3Gal was first analyzed by Galili using isoelectric focusing of antibody purified from an Ag-silica matrix (65). The isoelectric focusing pattern found by this technique was polyclonal. In a simliar approach, our laboratory purified anti-Gal_1–3Gal by specific elution from porcine vascular endothelial cells with Gal_1–3Gal oligosaccharide, and subjected the Ig µ and a chains to Edman degradation sequencing. The results showed that the majority of anti-Gal_1–3Gal IgM and IgG utilize a framework I region common to 19 of 23 functional VH3 family genes (Yu, unpublished data). This technique could not identify products of specific genes, nor could it resolve minor constituents of the antibodies. The heavy chain genes encoding anti-Gal_1–3Gal antibodies were first studied by analyzing Epstein-Barr virus (EBV)-transformed human B cells or recombinant Ig selected from a phage library (55,66) specific for Gal_1–3Gal, showing that most anti-Gal_1–3Gal antibodies are somatically mutated, and derive from the VH3 family without obvious preference for a particular gene. Studies using EBV-transformed B cells or recombinant anti-Gal_1–3Gal are limited in that they may potentially not represent the repertoire of serum antibodies, which in products of B cells in gutassociated lymphoid tissue might not be produced by B cells in the periphery nor B cells readily transformed by EBV. In another more direct approach we analzyed highly purified anti-Gal_1–3Gal obtained from immunoprecipitation by immunoblotting with a panel of VH family-specific anti-peptide antibodies (54,67–70). This method could resolve the VH family composition of circulating natural and elicited anti-Gal_1–3Gal antibodies, as well as differences in the VH usage of different Ig isotypes. The latter issue was of relevance given putative
110
Yu
differences in their mechanism of production (71). We found again that IgM are mostly products of the VH3 family, but could not ascertain bias for particular VH3 genes. On the other hand, IgG are equally expressed by members of VH1, VH3, and VH4, the families responsible for the majority of circulating serum antibodies. Our results would appear to be consistent with isoelectric focusing analysis by Rieben et al. of the closely related isohemagglutinins (72). These studies revealed an oligoclonal pattern for IgM and a polyclonal pattern for the IgG. Together these studies suggest that the natural IgM anti-carbohydrate antibodies are encoded by a limited number of genes, while IgG anticarbohydrate antibodies may result from diversification of such genes or recruitment of novel genes.
POTENTIAL STIMULI FOR XENOREACTIVE NATURAL ANTIBODY PRODUCTION Whether natural antibodies are constitutive products of the immune system or are produced in response to antigenic stimulation has long been a matter of controversy. At the turn of the century, natural antibodies directed against bacterial antigens and xenogeneic cells were originally thought to be produced spontaneously by the immune system (2,73). Later, Wiener found that many natural antibacterial antibodies did not occur in animals raised in a germ-free environment, and proposed that most natural antibodies resulted from occult sensitization with microbes (3). Springer then demonstrated that anti-blood group B antibodies and other antibacterial natural antibodies in chicks were produced only after colonization with normal microbial flora (4). Refinements of these prototype experiments in other germ-free and antigen-free animal models have confirmed that normal microbial flora are necessary for the production of many but not all natural antibodies (74,78). Some natural antibodies, such as the “heteroagglutinin” or xenoreactive antibodies, were expressed in germ-free models and did not appear to require microbial exposure. It is now evident that most of such spontaneously occurring antibodies are in fact polyreactive natural antibodies (30,77), in contrast to anti-carbohydrate antibodies such as isohemagglutinins and anti-Gal_1–3Gal, which appear to require environmental contact for development, such as gut bacteria. Self- or “endogenous” antigens may play a critical role in the expression of some natural antibodies. The repertoire of polyreactive natural antibodies that are expressed in antigen-free and germ-free animals
Origin of Xenoreactive Natural Antibodies
111
appears to vary with the genetic make-up of the animals, particularly the VH loci or MHC class I haplotypes (79–82). Self-antigens such as the Ig heavy chains and MHC molecules, as well as molecules acquired transplacentally or from colostrum, may thus have important roles in shaping the natural humoral repertoire in either a positive or negative fashion (83). Although microbial flora augment the production of natural antibodies, the flora might do so by specific antigen sensitization, as Wiener proposed, or by the nonspecific or polyclonal activation of B cells. The expression of polyreactive natural antibodies does not require microbial flora, yet is greatly enhanced by it (84,85), suggesting that microbes provide a nonspecific costimulation. For example, microbial stimulus is required for antibody expression in mice carrying a transgenic autoreactive anti-erythrocyte antibody, as no mice develop hemolytic anemia when raised in germ-free conditions, while 50% of mice develop hemolytic anemia under normal conditions (86). Microbes might stimulate B cells nonspecifically with molecules that are cross-reactive with B-cell antigen receptors, or with molecules that activate B cells in a polyclonal or antigen-independent fashion. Various microbial constituents could activate B cells in a polyclonal or antigen-independent manner. Bacterial molecules such as lipopolysaccharide and other lipoproteins, bacterial DNA sequences such as unmethylated CpG motifs, and bacterial porins cause B cells to proliferate and secrete antibodies without antigen receptor signaling (87). In support of a role of mitogens in stimulating the production of natural polyreactive antibodies, there is an increased frequency of autoreactive specificities such as anti-DNA and rheumatoid factor among the antibodies produced when B cells are stimulated in vitro by these mitogens (88,89). Other microbial products cross-link B-cell antigen receptors without regard to their antigen specificity, thus functioning as B-cell superantigens (90). For example, staphylococcal protein A activates B cells expressing VH3derived surface Ig, causing a profound expansion of these B cells in vivo and in vitro (91). Staphylococcal enterotoxin D, with the help of activated T cells, activates B cells expressing VH4-derived surface Ig in vitro (92). Peptostreptococcal protein L preferentially binds to certain light chain families and may thereby stimulate B cells (93). The gut lumenal protein Fv may function as an endogenous superantigen, as it binds to VH3 family Ig and B cells expressing VH3 family Ig (94). Superantigen stimulation may explain the observation in the mature B-cell repertoire of consistent skewing toward certain heavy and light chain V genes families compared to their germline frequency (57,58,90,95).
112
Yu
With regard to costimulation, polyreactive natural antibodies would appear to develop by nonconventional modes of B-cell stimulation, such as T-cell-independent modes used primarily in B1 cells, and perhaps by some of the nonconventional (i.e., TI–2, superantigen) modes described above as well. With regard to costimulation of B cells in the development of natural anti-carbohydrate antibodies, some limited insights were gained by studying the expression of xenoreactive natural antibodies and isohemagglutinins in immunodeficient subjects (56). Antibodies to Gal_1–3Gal in all individuals were essentially monospecific, reacting only with that sugar or very close derivatives, and thus distinct from polyreactive natural antibodies. Subjects with the hyper-IgM syndrome, having disrupted CD40/CD40L signaling, were able to synthesize abundant quantities of xenoreactive polyreactive antibodies [by presumably thymus–independent (TI) mechanisms] but were profoundly deficient in synthesizing anti-Gal_1–3Gal and isohemagglutinins, indicating a requirement for CD40/CD40L cosignaling for an intact response. These results suggest that the B cells giving rise to natural anti-Gal_1–3Gal and isohemagglutins are evolved by a thymus-dependent (TD) mechanism, supporting the notion that “natural” anti-Gal_1–3Gal and isohemagglutinins are not products of the primary immune repertoire but are elicited by an antigen. The cosegregation of isohemagglutinins and anti-Gal_1–3Gal in immunodeficiency provides further evidence for an integral relationship between natural anti-Gal_1–3Gal and isohemagglutinins, as had been suggested previously (23). These data corroborate other studies suggesting that the anti-Gal_1–3Gal response, like other natural anti-carbohydrate antibodies, are unlikely to originate exclusively from the B1 compartment. Other in-vivo work by our group suggests a TD basis for the generation of natural anti-Gal_1–3Gal antibodies. Previous work established that exposure of higher primates to xenogeneic tissues elicits strong IgG and IgM responses to Gal_1–3Gal (96,97). Although the natural IgG did not appear to be important in hyperacute rejection, the IgG produced after sensitization could cause hyperacute rejection in passive transfer experiments (98) and was proposed to play a significant role in the rejection of grafts when hyperacute rejection is evaded (99). It was not known whether antigen-stimulation led to a TI response or a TD response, presumably against proteins or hapten-protein complexes. The time course of the response (peaking between 7 and 8 d), the amplitude (2-log), and the presence of long-term anamnesis (97) were most consistent with a TD response.
Origin of Xenoreactive Natural Antibodies
113
REFERENCES 1. Landsteiner K, and Prasek E. Ueber die beziehung der antikorper zu der prazipitablen substanz des serums. Zeitschrift fur Immunitaetsforschung und Experimentelle Therapie 1911; 10:68. 2. Landsteiner K. Zur kenntnis der antifermentativen, lytischen und agglutinierenden wirkungen des blutserums und der lymphe. Centralblatt Fur Bakteriologie und Parasitenkunde 1900; 29:357. 3. Wiener A S. Origin of naturally occurring hemagglutinins and hemolysins: a review. Journal of Immunology 1951; 66:287. 4. Springer G F, Horton R E, and Forbes M. Origin of anti-human blood group B agglutinins in white leghorn chicks. Journal of Experimental Medicine 1959; 110:221. 5. Boyden S V. Natural antibodies and the immune response. Advances in Immunology 1966; 5:1. 6. Avrameas S. Natural autoantibodies: from ‘horror autotoxicus’ to ‘gnothi seauton’. Immunology Today 1991; 12:154. 7. Boes M, Prodeus A P, Schmidt T, Carroll M C, and Chen J. A critical role of natural immunoglobulin M in immediate defense against systemic bacterial infection. J Exp Med 1998; 188:2381.2 8. Gonzalez R, Charlemagne J, Mahana W, and Avrameas S. Specificity of natural serum antibodies present in phylogenetically distinct fish species. Immunology 1998; 63:31. 9. Marchalonis J J, Hohman V S, Thomas C, and Schluter S F. Antibody production in sharks and humans: a role for natural antibodies. Dev Comp Immunol 1993; 17:41. 10. Marchalonis J J, Schluter S F, Bernstein R M, and Hohman V S. Antibodies of sharks: revolution and evolution. Immunol Rev 1998; 166:103. 11. Dempsey P W, Allison M E, Akkaraju S, Goodnow C C, and Fearon D T. C3d of complement as a molecular adjuvant: bridging innate and acquired immunity. Science 1996; 271:348. 12. Fearon D T, and Locksley R M. The instructive role of innate immunity in the acquired immune response. Science 1996; 272:50. 13. Carroll M C. and Prodeus A P. Linkages of innate and adaptive immunity. Curr Opin Immunol 1998;10:36. 14. Wang H, and Chow D A. Protein kinase C expression links natural antibody binding with surveillance of activated and preneoplastic cells. Scand J Immunol 1999; 49:381. 15. Rother R P, Fodor W L, Springhorn J W, et al. A novel mechanism of retrovirus inactivation in human serum mediated by anti-alpha-galactosyl natural antibody. Journal of Experimental Medicine 1995; 182:1345. 16. Galili U. Interaction of the natural anti-Gal antibody with a-galactosyl epitopes: a major obstacle for xenotransplantation in humans. Immunology Today 1993; 14:480. 17. Challacombe S J, Bergmeier L A, Czerkinsky C, and Rees A S. Natural antibodies in man to Streptococcus mutans: specificity and quantification. Immunology 1984; 52:143. 18. Tomiyama T, Lake D, Masuho Y, and Hersh E M. Recognition of human immunodeficiency virus glycoproteins by natural anti-carbohydrate antibodies in human serum. Biochem Biophys Res Commun 1991; 177:279.
114
Yu
19. Slutzky G M, and Greenblatt C L. Identification of galactose as the immunodominant sugar of leishmanial excreted factor and subsequent labeling with galactose oxidase and sodium boro[3H]hydride. Infect Immun 1982; 37:10. 20. Kimura H, Umeda M, and Marcus D M. Expression of a cross-reactive idiotope on naturally occurring murine antibodies against 3-fucosyllactosamine, levan, and dextran. J Immunol 1989; 142:3477. 21. Avila J L, Rojas M, and Galili U. Immunogenic Gal alpha 1——3Gal carbohydrate epitopes are present on pathogenic American Trypanosoma and Leishmania. J Immunol 1989; 142:2828. 22. McConville M J, Homans S W, Thomas-Oates J E, Dell A, and Bacic A. Structures of the glycoinositolphospholipids from Leishmania major. A family of novel galactofuranose-containing glycolipids. J Biol Chem 1990; 265:7385. 23. Parker W, Lundberg-Swanson K L, Holzknecht Z E, et al. Isohemagglutinins and xenoreactive antibodies are members of a distinct family of natural antibodies. Human Immunology 1996; 45:94. 24. Perper R J., and Najarian J S. Experimental renal heterotransplantation. I. In widely divergent species. Transplantation 1966; 4:377. 25. Good A H, Cooper D K C, Malcolm A J, et al. 1992. Identification of carbohydrate structures that bind human antiporcine antibodies: implications for discordant xenografting in humans. Transplantation Proceedings 24:559. 26. Cooper D, Good A, Koren E, et al. Identification of a-galactosyl and other carbohydrate epitopes that are bound by human anti-pig antibodies: relevance to discordant xenografting in man. Transplant Immunology 1993; 1:198. 27. Schettino E W, Chai S K, Kasaian M T, Schroeder H W Jr., and Casali P. VHDJH gene sequences and antigen reactivity of monoclonal antibodies produced by human B-1 cells: evidence for somatic selection. J Immunol 1997; 158:2477. 28. Guilbert B, Dighiero G, and Avrameas S. Naturally occurring antibodies against nine common antigens in human sera. I. Detection, isolation, and characterization. Journal of Immunology 1982; 128:2779. 29. Grabar,P. Autoantibodies and the physiological role of immunoglobulins. Immunology Today 1983; 4:337. 30. Casali P, and Schettino E W. Structure and function of natural antibodies. Current Topics in Microbiology and Immunology 1996; 210:167. 31. Yu P B , Holzknecht Z E, Bruno D, Parker W, and Platt J L. Modulation of natural IgM binding and complement activation by natural IgG antibodies. Journal of Immunology 1996; 157:5163. 32. Galili U, Rachmilewitz E A, Peleg A, and Flechner I. A unique natural human IgG antibody with anti-a-galactosyl specificity. Journal of Experimental Medicine 1984; 160:1519. 33. Rousseau P G, Mallett C P, and Smith-Gill S J. A substantial proportion of the adult BALB/c available B cell repertoire consists of multireactive B cells. Mol Immunol 1989; 26:993. 34. Hartman A B, Mallett C P, Srinivasappa J, Prabhakar B S, Notkins, A L and SmithGill S J. Organ reactive autoantibodies from non-immunized adult BALB/c mice are polyreactive and express non-biased VH gene usage. Mol Immunol 1989; 26:359. 35. Parker W, Bruno D, Holzknecht Z E, and Platt J L. Xenoreactive natural antibodies: isolation and initial characterization. Journal of Immunology 1994; 153:3791.
Origin of Xenoreactive Natural Antibodies
115
36. Hayakawa K , Hardy R R, Parks D R, and Herzenberg L A. The “LY-1 B” cell subpopulation in normal, immunodefective, and autoimmune mice. Journal of Experimental Medicine 1983; 157:202. 37. Kantor A B, Stall A M, Adams S, and Herzenberg L A. Differential development of progenitor activity for three B-cell lineages. Proc Natl Acad Sci U S A 1992; 89:3320. 38. Hardy R R. Variable gene usage, physiology and development of Ly-1+ (CD5+) B cells. Curr Opin Immunol 1992; 4:181. 39. Forster I, and Rajewsky K. Expansion and functional activity of Ly-1+ B cells upon transfer of peritoneal cells into allotype-congenic, newborn mice. Eur J Immunol 1987; 17:521. 40. Baumgarth N, Herman O C, Jager G C, Brown L, and Herzenberg L A. Innate and acquired humoral immunities to influenza virus are mediated by distinct arms of the immune system. Proc Natl Acad Sci U S A 1999; 96:2250. 41. Hayakawa K, Hardy R R, and Herzenberg L A. Progenitors for Ly-1 B cells are distinct from progenitors for other B cells. J Exp Med 1985; 161:1554. 42. Stall A M, Wells S M , and Lam K P. B-1 cells: unique origins and functions. Semin Immunol 1996; 8:45. 43. Quan C, Berneman A, Pires R, Avrameas S, and Bouvet J P. Salivary natural antibodies as a basic immune barrier against group A streptococci. Adv Exp Med Biol 1997; 418:881. 44. Quan C P, Berneman A, Pires R, Avrameas S, and Bouvet J P. Natural polyreactive secretory immunoglobulin A autoantibodies as a possible barrier to infection in humans. Infect Immun 1997; 65:3997. 45. Bhat N M, Kantor A B, Bieber M M, Stall A M, Herzenberg L A, and Teng N N. The ontogeny and functional characteristics of human B-1 (CD5+ B) cells. Int Immunol 1992; 4:243. 46. Freedman A S, Freeman G, Whitman J, Segil J, Daley J, Levine H, and Nadler L M. Expression and regulation of CD5 on in vitro activated human B cells. Eur J Immunol 1989; 19:849. 47. Haughton G, Arnold L W, Whitmore A C, and Clarke S H. B-1 cells are made, not born. Immunology Today 1993; 14:84. 48. Turman M A, Casali P, Notkins A L, Bach F H, and Platt J L. Polyreactivity and antigen specificity of human xenoreactive monoclonal and serum natural antibodies. Transplantation 1991; 52:710. 49. Geller R L, Bach F H, Turman M A, Casali P, and Platt J L. Evidence that polyreactive antibodies are deposited in rejected discordant xenografts. Transplantation 1993; 55:168. 50. Parker W, Lateef J, Everett M L, and Platt J L. Specificity of xenoreactive antigala1-3gal IgM for a-galactosyl ligands. Glycobiology 1996; 6:499. 51. Pitre J, Moller E, and Satake M. Immunoselection of human lymphocytes producing xenoreactive natural antibodies against Gala1-3Gal terminal residues. Xenotransplantation 1995; 2:271. 52. Barrett D J, Sleasman J W, Schatz D A, and Steinitz M. Human anti-pneumococcal polysaccharide antibodies are secreted by the CD5- B cell lineage. Cell Immunol 1992; 143:66. 53. Wuttke N J, Macardle P J, and Zola H. Blood group antibodies are made by CD5+ and by CD5- B cells. Immunol Cell Biol 1997; 75:478.
116
Yu
54. Yu P B, Parker W, Silverman G J, Nayak J V, and Platt J L. Distinct VH gene utilization in natural and acquired anti-Gal(alpha)1-3Gal antibodies. In preparation. 55. Wang, L., M. Z. Radic, and U. Galili. 1995. Human anti-gal heavy chain genes. Journal of Immunology 1999; 155:1276. 56. Parker W, Yu P B, Holzknecht Z E, Lundberg K, Buckley R H, and Platt J L. Specificity and function of “natural” antibodies in immunodeficient subjects: clues to B cell lineage and development. J Clin Immunol 1997; 17:311. 57. Foster S. J , Brezinschek H P, Brezinschek R I, and Lipsky P E. Molecular mechanisms and selective influences that shape the kappa gene repertoire of IgM+ B cells. Journal of Clinical Investigation 1997; 99:1614. 58. Brezinschek H P, Foster S J, Brezinschek R I, Dorner T, Domiati-Saad R, and Lipsky P E. Analysis of the human VH gene repertoire. Differential effects of selection and somatic hypermutation on human peripheral CD5(+)/IgM+ and CD5()/IgM+ B cells. Journal of Clinical Investigation 1997; 99:2488. 59. Kantor A B, and Herzenberg L A. Origin of murine B cell lineages. Annu Rev Immunol 1993; m11:501. 60. Neron S, and Lemieux R. CD5+ B cell-dependent regulation of the murine T-cell independent immune response against the human blood group A antigen. Immunol Invest 1997; 26:631. 61. Diaw L, Magnac C, Pritsch O, Buckle M, Alzari P M, and Dighiero G. Structural and affinity studies of IgM polyreactive natural autoantibodies. J Immunol 1997; 158:968. 62. Seidl K . Wilshire J A, MacKenzie J D, Kantor A B, and Herzenberg L A. Predominant VH genes expressed in innate antibodies are associated with distinctive antigen-binding sites. Proc Natl Acad Sci U S A 1999; 96:2262. 63. Holmberg D , Andersson A, Carlsson L, and Forsgren S. Establishment and functional implications of B-cell connectivity. Immunol Rev 1989; 110:89. 64. Lymberi P, Dighiero G, Ternynck T, and Avrameas S. A high incidence of crossreactive idiotypes among murine natural autoantibodies. Eur J Immunol 1985; 15:702. 65. Galili U, Macher B A, Buehler J, and Shohet S B. Human natural anti-a-galactosyl IgG: the specific recognition of a(1-3)-linked glactose residues. Journal of Experimental Medicine 1985; 162:573. 66. Wang L, Radic M Z, Siegel D, Chang T, Bracy J, and Galili U. Cloning of antiGal Fabs from combinatorial phage display libraries: structural analysis and comparison of Fab expression in pComb3H and pComb8 phage. Molecular Immunology 1997; 34:609. 67. Silverman G, and Carson D. Structural characterization of human monoclonal cold agglutinins: evidence for a distinct primary sequence-defined Vh4 idiotype. European Journal of Immunology 1990; 20:351. 68. Silverman G J , Goni F, Fernandez J, Chen P P, Frangione B, and Carson D A. Distinct patterns of heavy chain variable region subgroup use by human monoclonal autoantibodies of different specificity. J Exp Med 1988; 168:2361. 69. Silverman G J, and Carson D A. Structural characterization of human monoclonal cold agglutinins: evidence for a distinct primary sequence-defined VH4 idiotype [published erratum appears in Eur J Immunol 1990 Apr;20(4):952]. European Journal of Immunology 1990; 20:351. 70. Silverman G J, and Lucas A H. Variable region diversity in human circulating antibodies specific for the capsular polysaccharide of Haemophilus influenzae type
Origin of Xenoreactive Natural Antibodies
71.
72.
73. 74. 75. 76.
77.
78. 79. 80.
81.
82. 83. 84.
85. 86.
87.
88.
117
b. Preferential usage of two types of VH3 heavy chains. Journal of Clinical Investigation 1991; 88:911. Kay L A. Cellular basis of immune response to antigens of ABO blood-group system. Capacity to provide help during response to T-cell-dependent ABO-system antigens is restricted to individuals of blood group O. Lancet 1984; 2:1369. Rieben R, Frauenfelder A, and Nydegger U E. Spectrotype analysis of human ABO antibodies: evidence for different clonal heterogeneity of IgM, IgG, and IgA antibody populations. Vox Sang 1996; 70:104. Ehrlich P. On immunity with special reference to cell life. Proceedings of the Royal Society of London 1900; 66:424. Springer G F, and Horton R E. Blood group isoantibody stimulation in man by feeding blood group-active bacteria. Journal of Clinical Investigation 1969; 48:1280. Thorbecke G J. Some histological and functional aspects of lymphoid tissue in germfree animals. Annual Review of New York Academy of Sciences 1959; 78:237. Tlaskalova-Hogenova H, Mandel L, Stepankova R, et al. Autoimmunity: from physiology to pathology. Natural antibodies, mucosal immunity and development of B cell repertoire. Folia Biol 1992; 38:202. Cukrowska B, Sinkora J, Rehakova Z, et al. 1996. Isotype and antibody specificity of spontaneously formed immunoglobulins in pig fetuses and germ-free piglets: production by CD5- B cells. Immunology 88:611. Michael J G. Natural antibodies. Curr Top Microbiol Immunol 1969; 48:43. Coutinho A. Lymphocyte survival and V-region repertoire selection. Immunol Today 1993; 14:38. HauryM, Sundblad A, Grandien S, Barreau C, Coutinho A, and Nobrega A. The repertoire of serum IgM in normal mice is largely independent of external antigenic contact. Eur J Immunol 1997; 27:1557. Vasconcellos R, Nobrega A, Haury M, Viale A C , and Coutinho A. Genetic control of natural antibody repertoires: I. IgH, MHC and TCR beta loci. Eur J Immunol 1998; 28:1104. Coutinho, A. Beyond clonal selection and network. Immunological Review 1989; 110:63. Malanchere E, Huetz F, and Coutinho A. Maternal IgG stimulates B lineage cell development in the progeny. Eur J Immunol 1997; 27:788. Hooijkaas H, Benner R, Pleasants J R, and Wostmann B S. Isotypes and specificities of immunoglobulins produced by germ-free mice fed chemically defined ultrafiltered “antigen-free” diet. Eur J Immunol 1984; 14:1127. Pereira P, Forni L, Larsson E L, Cooper M, Heusser C, and Coutinho A. Autonomous activation of B and T cells in antigen-free mice. Eur J Immunol 1986; 16:685. Murakami M, Nakajima K, YamazakiK, Muraguchi T, Serikawa T, and Honjo T. Effects of breeding environments on generation and activation of autoreactive B-1 cells in anti-red blood cell autoantibody transgenic mice. J Exp Med 1997; 185:791. Snapper C M, and Mond J J. A model for induction of T cell-independent humoral immunity in response to polysaccharide antigens. Journal of Immunology 1996; ; 157:2229. Pisetsky D S, Jelinek D F. McAnally L M, Reich C F, and Lipsky P E. In vitro autoantibody production by normal adult and cord blood B cells. Journal of Clinical Investigation 1990; 85:899.
118
Yu
89. Liang H, Nishioka Y, Reich C F, Pisetsky D S, and Lipsky P E. Activation of human B cells by phosphorothioate oligodeoxynucleotides. Journal of Clinical Investigation 1996; 98:1119. 90. Silverman G.J. B-cell superantigens. Immunology Today 1997; 18:379. 91. Silverman G J, Nayak J V, Wagenknecht R, and Warnatz K. In vivo consequences of B cell superantigen immunization. Ann N Y Acad Sci 1997; 815:105. 92. Domiati-Saad R, Attrep J F, Brezinschek H P, Cherrie A H, Karp D R, and Lipsky P E. Staphylococcal enterotoxin D functions as a human B cell superantigen by rescuing VH4-expressing B cells from apoptosis. Journal of Immunology 1996; 156:3608. 93. Akerstrom B, and Bjorck L. Protein L: an immunoglobulin light chain-binding bacterial protein. Characterization of binding and physicochemical properties. J Biol Chem 1989; 264:19740. 94. Silverman G J, Roben P, Bouvet J P, and Sasano M. Superantigen properties of a human sialoprotein involved in gut-associated immunity. Journal of Clinical Investigation 1995; 96:417. 95. Domiati-Saad R, and Lipsky P E. B cell superantigens: potential modifiers of the normal human B cell repertoire. Int Rev Immunol 1997; 14:309. 96. Satake M, Kawagishi N, Rydberg L, Samuelsson B, Tibell A, Groth C, and Moller E. Limited specificity of xenoantibodies in diabetic patients transplanted with fetal porcine islet cell clusters. Main antibody reactivity against a-linked galactose-containing epitopes. Xenotransplantation 1994; 1:89. 97. Cotterell A H , Collins B H, Parker W, Harland R C, and Platt J L. The humoral immune response in humans following cross-perfusion of porcine organs. Transplantation 1995; 60:861. 98. Platt J L, Fischel R J, Matas A J, Reif S A, Bolman R E, and Bach F H. Immunopathology of hyperacute xenograft rejection in a swine-to-primate model. Transplantation 1991; 52:214. 99. Lin S S , Weidner B C, Byrne G W et al. 1998. The role of antibodies in acute vascular rejection of pig-to-baboon cardiac transplants. Journal of Clinical Investigation 101:1745.
Xenoantigen Synthesis
7
119
Synthesis of Carbohydrate Antigens Recognized by Xenoreactive Antibodies Mauro S. Sandrin, PhD, DSc, and Ian F.C. McKenzie, MD, PhD
INTRODUCTION The problem in transplanting pig organs to humans is the occurrence of hyperacute rejection. Natural antibodies (particularly IgM, but also IgG) bind to antigens on the endothelium lining the blood vessels, fix complement, and lead to endothelial cell activation and intravascular thrombosis within minutes. Hyperacute rejection can be prevented by the depletion, blocking, or removal of any of the three components. Here we examine the importance of carbohydrate antigens in xenograft rejection, in particular the Gal_(1,3)Gal epitope; the _1,3galactosyltransferase enzyme responsible for generating Gal_(1,3)Gal and transgenic strategies designed to eliminate or reduce expression of Gal_(1,3)Gal, such that the epitope can no longer be detected by natural human antibodies.
NATURAL ANTI-GALA(1,3)GAL ANTIBODIES Humans have substantial amounts of natural anti-pig antibodies (i.e., antibodies not dependent on prior immunization with pig tissues) that cause hyperacute rejection of transplanted vascularized pig organs. Humans and pigs diverged from a common evolutionary ancestor some 64 million years ago, and as there are at present so many obvious differences between these two species, it would be logical to conclude that there would be many xenoantigenic differences (i.e., antigens important From: Xenotransplantation: Basic Research and Clinical Applications Edited by: Jeffrey L. Platt © Humana Press Inc., Totowa, NJ
119
120
Sandrin and McKenzie
in the human immune response to pigs) between the two species that would make the prevention or eradication of hyperacute rejection virtually impossible. This prediction is wrong, as it is clear from the results of a number of investigations that the major target of natural human antibodies is the carbohydrate epitope Gal_(1,3)Gal (1–6). The importance of this epitope was demonstrated by several key observations: 1. The identification of natural antibodies to Gal_(1,3)Gal. The first description of anti-Gal_(1,3)Gal antibodies can be traced back to Landsteiner’s original description of a “B-like” reaction of normal human serum with erythrocytes from several nonprimate species, an observation that arose from his studies of human ABO blood groups. More recently, Galili made the observation that approx 1% of total human IgG is to Gal_(1,3)Gal (7–9). Others have confirmed and extended these studies, and it is now clear that both IgG and IgM antiGal_(1,3)Gal antibodies are present in large amounts (1–3,5,10). The importance of demonstrating occurrence of anti-Gal_(1,3)Gal IgM antibodies (1) is that these are considered to be the most important antibody class involved in hyperacute rejection (11,12). 2. Inhibition of binding of anti-pig antibodies to key pig targets by carbohydrates in vitro and in vivo. Several monosaccharides, disaccharides, and synthetic oligosaccharides have been used to inhibit or neutralize the binding of anti-pig antibodies to Gal_(1,3)Gal+ cells in vitro assays including hemagglutination, direct binding, and cytotoxicity on cells such as lymphocytes and endothelial cells (1,6,13–15). These studies show that _
Xenoantigen Synthesis
121
complement] was able to lyse COS cells after they were transfected with the _1,3galactosyltransferase cDNA (13). Furthermore, absorption studies demonstrated that Gal_(1,3)Gal+ COS cells could remove all anti-pig antibodies from human serum (4). Based on the results of these different studies, it is now clear that Gal_(1,3)Gal is the major target of human xenoreactive antibodies.
GALA(1,3)GAL ANTIGENS The distribution of Gal_(1,3)Gal on different tissue has been examined histologically (22–24) and demonstrates that Gal(_1,3)Gal is widely distributed, being found on the surface of most pig cells, but with some notable exceptions. Virtually all endothelial cells lining arterioles, capillaries, and venules express large amounts of Gal_(1,3)Gal. All hepatic cells express Gal_(1,3)Gal and in the kidney the largest amount is found in the proximal convoluted tubules, less in the distal tubules, none in the collecting duct, and little in the glomerulus (23). By contrast, pancreatic islet cells and heart muscle fibers were virtually nonexpressing other than in small blood vessels and pancreatic ducts (23). The large amount in capillaries explain the hyperacute rejection at the molecular level, owing to natural IgM reacting with Gal_(1,3)Gal on the vessels. Gal_(1,3)Gal is expressed on many molecules as a normal component of glycosylation, including O-linked and N-linked oligosaccharide chains and glycolipids, where it is used as an alternative to sialic acid as an uncharged terminal carbohydrate residue. On the cell surface, Gal_(1,3)Gal is expressed on many different molecules (25), e.g., on platelets, Gal_(1,3)Gal is predominantly found on fibrinogen, _2 integrin, and `3 integrin (26), on endothelial cells, more than 20 glycoproteins carry Gal_(1,3)Gal (25), some of which have been identified as DM-GRASP, and the _1, _ v, _ 3/_ 5, `1, and `3 integrins (27). Furthermore, Gal_(1,3)Gal is expressed on many secreted proteins such as thyroglobulin, immunoglobulin, and laminin (28,29). As many different molecules are Gal_(1,3)Gal+ it is not possible to separately eradicate each—rather, the synthetic machinery producing Gal_(1,3)Gal has to be targeted.
SYNTHESIS OF GAL_(1,3)GAL Complex glycan moieties, including glycolipids and N-linked and O-linked glycans of glycoproteins, are synthesized by the sequential
122
Sandrin and McKenzie
and ordered addition of monosaccharide residues within the Golgi complex of the cell. The Golgi complex consists of a reticular network within which reside all the appropriate enzymes and transporters for the synthesis of complex carbohydrates. It is within the lumen of the Golgi complex that elongation of carbohydrate chains occurs, and it is therefore not surprising that the catalytic domains of glycosyltransferases (enzymes responsible for complex carbohydrate synthesis) reside. All glycosyltransferases isolated thus far are predicted to be type II integral membrane protein (amino terminal orientated toward the cytoplasm) comprising a short cytoplasmic tail, a transmembrane region of approx 16 amino acids, a stalk region, and a carboxyl-terminal catalytic domain that resides within the lumen of the Golgi (30). The glycosyltransferase responsible for Gal_(1,3)Gal synthesis is the _1,3galactosyltransferase (31), the enzyme that catalyzes the addition of a terminal _
Xenoantigen Synthesis
123
Fig. 1. Biosynthetic pathway for synthesis of Gal_(1,3)Gal. Pathway begins with N-acetyl-lactosamine (Gal_(1,4)GlcNAc), and the _1,3galactosyltransferase enzyme adds galactose to generate Gal_(1,3)Gal (A). Both gene inactivation by homologous recombination (B) and transgenic approaches using _galactosidase (C) would prevent production of _1,3galactosyltransferase, and eliminate Gal_(1,3)Gal. Also shown is an alternative transgenic approach to utilize the substrate of _1,3galactosyltransferase by _1,2fucosyltransferase (D) or _sialyltransferase (E).
strategies (see below). Studies in the mouse suggests four isoforms of this enzyme (33), and two isoforms have been identified in the pig (21,35). An interesting observation is that Old World monkeys and humans have anti-Gal_(1,3)Gal antibodies (7), as these species lack the Gal_(1,3)galactosyltransferase and thus do not synthesize Gal_(1,3)Gal. The reason for lack of a functional _1,3galactosyltransferase in these species is that the homologous genomic sequences are pseudogenes (37–39)—the human genome contains two pseudogenes, each containing several frame shift and nonsense mutations, one located on chromosome 9q33–34 and the other on chromosome 2q14–15 (40). Similar mutations have been shown in the Old World monkey
124
Sandrin and McKenzie
Fig. 2. Alignment of the amino-acid sequences comparison of _GTs. Single letter amino acid code used for the pig (p-GT), mouse (m-GT), bovine (b-GT), marmoset (marm-GT), human B transferase (UDP-Gal: Fuc(_1,2)Gal(_1,3)Gal transferase). Identical amino acid boxed, represents break introduced into amino acid sequence for maximal alignment.
pseudogene (39). It is interesting to note that the pseudogene on chromosome 9 is linked to the A and B blood group transferases (41), demonstrating an evolutionary relationship between all of these glycosyltransferase genes (3)—particularly as the predicted amino acid sequences of the A and B blood group transferases (which transfer _
Xenoantigen Synthesis
125
these glycosyltransferases in the different species: both the human and Old World monkey transferases have an absolute requirement for the subterminal galactose to be fucosylated, the opposite is true for the murine, bovine, and pig _1,3galactosyltransferases such that these transferases will only transfer a galactose to a nonfucosylated galactose. This observation forms the basis of alternative approaches to produce Gal_(1,3)Gal- pigs using a different glycosyltransferase to compete for substrate (see below).
STRATEGIES TO DECREASE THE AMOUNT OF GALA(1,3)GAL The discovery that a single enzyme produces the major xenoantigen for pig-to-human xenotransplantation has led to widespread interest in the possibility of genetically altering the pig to make a donor whose organs would not undergo hyperacute rejection. To eliminate or reduce expression of Gal_(1,3)Gal such that the epitope is no longer recognized by natural human antibodies, two approaches are possible (Fig. 1): either to inactivate the _1,3galactosyltransferase gene by homologous recombination, or alternatively, by using transgenes encoding (a) glycosyltransferases that can competitively and effectively inhibit the activity of the _1,3galactosyltransferase or (b) glycosidases that remove Gal_(1,3)Gal.
Gene Inactivation of the _1,3galactosyltransferase The most obvious way of producing Gal_(1,3)Gal- pigs is to disable the _1,3galactosyltransferase gene by homologous recombination. At present homologous recombination technology is generally limited to small species such as mouse and rat, and not feasible in pigs, although the technique may soon expand to other species (see below). However, what will be the physiological consequences of inactivating the _1,3galactosyltransferase gene in pigs and will this be lethal as is the inactivation of some other glycosyltransferases (42–45)? Insights into some of these questions can be gained from naturally occurring mutations in humans, and produced in Gal o/o knock-out mice. Naturally occurring mutations to inactivate the _1,3galactosyltransferase gene of humans and Old World monkeys has caused no deleterious effects—not unexpected as mutations inactivating other human glycosyltransferase have not led to deleterious effects: O blood group individuals have a nonfunctional B transferase, owing to several nucle-
126
Sandrin and McKenzie
otide substitutions that introduce in-phase termination codons into the gene (46); individuals with the “Bombay” phenotype lack a functional _1,2fucosyltransferase (47). Thus one would predict that other species with inactivated _1,3galactosyltransferase would also be viable and produce anti-Gal_(1,3)Gal antibodies. Such is the case with chickens [which have no Gal_(1,3)Gal] (7,48) and Gal o/o mice (49,50). Homozygous mouse strains with an inactivated _1,3galactosyltransferase gene have been developed, by targeting exon 9 (the exon encoding the catalytic domain) for disruption (49,50). Inactivation of the gene is not lethal, the mice completely lack Gal_(1,3)Gal in all tissues, and like humans produce natural anti-Gal_(1,3)Gal antibodies. The Gal o/o mice are a valuable resource as a small animal model to study Gal_(1,3)Gal-dependent hyperacute rejection of xenografts (51) or delayed xenograft rejection (52), or to rapidly evaluate the efficacy of combining different transgenic strategies to produce the optimal donor animal. In this model, using Gal o/o mice as recipients of heart grafts from Gal_(1,3)Gal+ animals are rejected within 30 min by antiGal_(1,3)Gal IgM antibodies and complement-dependent mechanisms (51). By contrast we have found that cultured pig islets are completely resistant to anti-Gal_(1,3)Gal antibodies in vivo and are not rejected by circumstances that reject Gal_(1,3)Gal+ mouse hearts (53). By contrast, grafts of Gal_(1,3)Gal+ bone marrow to Gal o/o mice showed 50% survival compared to syngeneic bone marrow, and if the Gal o/o mice were immunized to have high titered high affinity anti-Gal_(1,3)Gal antibodies, bone marrow was entirely rejected (in preparation). However, giving larger amounts of Gal_(1,3)Gal + bone marrow cells together with Gal o/o bone marrow cells leads to mixed chimerism in lethally irradiated Gal o/o mice, leads to B cell tolerance, i.e., antiGal_(1,3)Gal antibodies do not reappear (54). Retroviral gene transfer was used in another study to express the _1,3galactosyltransferase in Gal o/o bone marrow cells and induce B cell tolerance (55). Gal o/o hearts are not hyperacutely rejected by normal or hyperimmune (antibody titers >1/20,00) mice and point to the desirability of producing Gal o/o pigs. The technique of animal cloning by “nuclear transfer” may present the solution required to produce Gal_(1,3)Gal- pigs, and cloned sheep, obtained from the nucleus of a fully differentiated cell line, have been described (56), and, more recently, goats (57), cattle (58), and mice (59). Furthermore, transgenic sheep (60) and calves (58) have been produced from nuclei of fibro-
Xenoantigen Synthesis
127
blasts transfected in vitro to express a foreign protein. It is highly likely that gene inactivation will be coupled with the nuclear transfer technology and that this will be applied successfully to the pig.
Transgenic Approaches to Reduce Gal_(1,3)Gal We devised an alternate strategy to the knock-out for reducing expression of the Gal_(1,3)Gal epitope by the competitive removal of the precursor by transgenic expression of another glycosyltransferase, which “deviates” the glycosylation pathway away from Gal_(1,3)Gal (Fig. 1) and leads to the production of another carbohydrate, preferably one not recognized by natural human antibodies (3,61). The first glycosyltransferase to be used in this approach was the _1,2fucosyltransferase (61).
_1,2fucosyltransferase _1,2fucosyltransferase was chosen as an appropriate enzyme to decrease the expression of Gal_(1,3)Gal as, like _1,3galactosyltransferase, it uses N-acetyl lactosamine as the acceptor for an _90%, with a high level of H substance expression (61,63). Concurrently, there was a significant decrease in human antibody binding and decreased susceptibility of these cells to lysis by human serum and complement (61–63). The mechanism of exclusion of Gal_(1,3)Gal by _1,2fucosyltransferase was examined, and it could be demonstrated that cytoplasmic tails of the two glycosyltransferases (_1,2fucosyltransferase and _1,3galactosyl-transferase) determines the topology of these within the Golgi complex, and plays a central role in the temporal order of action of these enzymes (64).
128
Sandrin and McKenzie
Transgenic mice that ubiquitously express the _1,2fucosyltransferase show a decrease in Gal_(1,3)Gal expression in all tissues, including endothelial cells, and a decrease in natural human antibody binding (62,65). More recently transgenic pigs expressing the _1,2fucosyltransferase have been reported from three centers (66,67) with all showing similar results to transgenic mice, i.e., reduction in the cell surface expression of Gal_(1,3)Gal. To determine whether the small amount of Gal_(1,3)Gal remaining in an _1,2fucosyltransferase transgenic animal would be sufficient to react with human antibody to trigger hyperacute rejection, we examined heart grafts from _1,2fucosyltransferase transgenic mice in Gal o/o mice (51). In this model, _1,2fucosyltransferase transgenic hearts survive >120 min, with no evidence of hyperacute rejection (51), although the grafts are rejected within 18 h (91). This suggests that additional modifications are required for total avoidance of antibody-mediated rejection. One such modification could be the use of other fucosyltransferases such as the Secretor type _1,2fucosyltransferase which has a broader acceptor specificity than the H type _1,2fucosyltransferase (FUT1), and synergizes with FUT1 to give >99% reduction of Gal_(1,3)Gal in COS cells (68). OTHER GLYCOSYLTRANSFERASES Other glycosyltransferases that also utilize N-acetyl lactosamine as a substrate should have similar effects to the fucosyltransferase. Sialyltransferases have been examined as transgenes in vitro, and gave disparate results (Fig. 1). One study using _2,3sialyltransferase showed some reduction in the levels of Gal_(1,3)Gal, although not as dramatic as that seen with the _1,2fucosyltransferase (69,70). In contrast no significant reduction of Gal_(1,3)Gal was obtained by the expression of either _2,3sialyltransferase or _2,6sialyltransferase (67). Both the _2,3sialyltransferase and the _1,3galactosyltransferase are involved in transferring a terminal sugar and have previously been shown that these compete for the same substrate within a cell (71). As it is likely that both are located in the same Golgi compartment, the efficiency of _<sialyltransferase may be improved by the redirection of this enzyme to an earlier Golgi compartment using the cytoplasmic tail of _1,2fucosyltransferase as discussed above. Another approach uses the overexpression of N-acetlyglucosaminyl transferase III (GNTIII), which effectively reduces of Gal_(1,3)Gal expression both in vitro (70) and in vivo in transgenic mice (72). This is likely to be due to the inhibition of oligosaccharide processing by other glycosyltransferases by the branched GlcNAc residue, as the
Xenoantigen Synthesis
129
GNTIII is involved in the production of bisecting GlcNAc residues in branched N-linked oligosaccharides. A-GALACTOSIDASE
An entirely different approach to competitive removal of the precursor is to enzymatically cleave Gal_(1,3)Gal after glycosylation has occurred. This can be achieved by using an _
ROLE OF GALA(1,3)GAL IN CELL-MEDIATED RESPONSES Up to this point we have been discussing the role of Gal_(1,3)Gal in the hyperacute response. What is the role of Gal_(1,3)Gal in
130
Sandrin and McKenzie
noncomplement mediated aspects of the hyperacute reaction, delayed xenograft rejection, or cell-mediate xenograft rejection? Would the strategies outlined above have any influence in such response? Endothelial cell activation plays an important role in the hyperacute rejection response as well as the delayed xenograft rejection phenomenon (77). Gal_(1,3)Gal antibodies can cause both type I and type II activation of endothelial cells, independent of complement fixation, giving rise to an immediate inflammatory response (78,79). Furthermore, the IB4 lectin and high affinity anti-Gal_(1,3)Gal antibodies gave stronger responses compared with naturally occurring antibodies, suggesting that, in the absence of hyperacute rejection, elicited antiGal_(1,3)Gal antibodies alone may contribute to xenograft rejection. This view is supported by the accelerated rejection seen in pig-to-baboon kidney xenotransplants pretreated with high affinity chicken antiGal_(1,3)Gal IgY antibodies, which do not fix mammalian complement (48). It is likely that these antibodies binding irreversibly to activation molecules on the endothelial cells causing endothelial cell activation and immediate damage. Another aspect of the xenogeneic response to consider is the role of Natural Killer (NK) cells and macrophages. NK cells have been implicated in the delayed xenograft response (80–83), either through direct recognition of antigen on the xenogeneic cell or through Fc receptor recognition of cell surface bound IgG. Naturally occurring and induced anti-Gal_(1,3)Gal IgG antibodies, as shown in pig-to-human islet xenografts (84), may play a role in such a response and human NK cells have been shown to recognize Gal_(1,3)Gal (85,86), although the nature of the NK receptor which recognizes Gal_(1,3)Gal is not known. Of interest is the finding that the susceptibility of pig endothelial cells to human NK cell lysis is decreased by expression of _1,2fucosyltransferase (86). In addition human monocytes have an important role in the T-cell independent xenograft response (87), and can directly adhere to and be activated by pig endothelial cells and in turn activate endothelial cells (88). The binding of monocytes to endothelial cells is with sialic acid moieties, but it is not clear if monocytes can also recognize Gal_(1,3)Gal directly as do NK cells. The expression of the _1,2fucosyltransferase not only decrease cell surface expression of Gal_(1,3)Gal, but also results in the reduction of terminal sialylation (89,90). Pig endothelial cells expressing the _1,2fucosyltransferase show reduced adherence
Xenoantigen Synthesis
131
and activation of human monocytes (90). Thus, the _1,2fucosyltransferase strategy, originally devised to reduce the binding of antiGal_(1,3)Gal antibodies to prevent hyperacute rejection, also has a dramatic effect on recognition of xenogeneic cells by human NK cells and monocytes.
CONCLUSION It is now clear that the Gal_(1,3)Gal is the first barrier to pig-to-human xenotransplantation, as it is the target for antibody-mediated hyperacute rejection. To totally remove Gal_(1,3)Gal in pig organs will require either a new knockout technology such as nuclear transfer, or a combination of three or more transgenes, each of which has a different site and mode of action on Gal_(1,3)Gal synthesis and will reduce Gal_(1,3)Gal to sufficient levels to prevent hyperacute rejection. Human complement regulators may also be needed in such a composite transgenic animal. An important role of Gal_(1,3)Gal antibody-dependent- and independent NK cell and monocyte-mediated xenograft rejection is also becoming more evident. What still remains to be addressed is to what level Gal_(1,3)Gal has to be reduced to prevent recognition by these cells or to preclude the production of high affinity antibodies that may cause direct activation of endothelial cells or destruction by Fc receptor mediated mechanisms. These issues can only be studied and hopefully addressed once an animal is engineered that has no targets for antibody, no complement activation and therefore no hyperacute rejection.
REFERENCES 1. Sandrin MS, Vaughan HA, Dabkowski PL, McKenzie IFC. Anti-pig IgM antibodies in human serum reacts predominantly with Gal_(1,3)Gal epitopes. Proc Natl Acad Sci USA 1993; 90:11,391–11,395. 2. Sandrin MS, McKenzie IFC. Gal_(1,3)Gal, the major xenoantigen(s) recognized in pigs by human natural antibodies. Immunol Rev 1994; 141:169–190. 3. Sandrin MS, Vaughan HA, McKenzie IFC. Identification of Gal_(1,3)Gal as the major epitope for pig-to-human vascularised xenografts. Transplant Rev 1994; 8:134–149. 4. McKenzie IFC, Vaughan HA, Sandrin MS. How important are anti-Gal_(1-3)Gal antibodies in pig to human xenotransplants? Xeno 1994; 2:107–110. 5. Cooper DK, Koren E, Oriol R. Oligosaccharides and discordant xenotransplantation. Immunol Rev 1994; 141:31–58. 6. Good AH, Cooper DKC, Malcolm AJ, et al. Identification of carbohydrate structures that bind human antiporcine antibodies: implications for discordant xenografting in humans. Transplant Proc 1992; 24:559–562.
132
Sandrin and McKenzie
7. Galili U, Shohet SB, Korbin E, Stults CLM, Macher BA. Man, apes and Old World monkeys differ from other mammals in the expression of the _
Xenoantigen Synthesis
133
23. McKenzie IF, Xing PX, Vaughan HA, Prenzoska J, Dabkowski PL, Sandrin MS. Distribution of the major xenoantigen (gal (_1–3)gal) for pig to human xenografts. Transpl Immunol 1994; 2 (2):81–86. 24. Oriol R, Ye Y, Koren E, Cooper DK. Carbohydrate antigens of pig tissues reacting with human natural antibodies as potential targets for hyperacute vascular rejection in pig-to-man organ xenotransplantation. Transplantation 1993; 56 (6):1433–1442. 25. Vaughan HA, McKenzie IFC, Sandrin MS. Biochemical studies of pig xenoantigens detected by naturally occurring human antibodies and the galactose_(1–3)galactose reactive lectin. Transplantation 1995; 59 (1):102–109. 26. Platt JL, Holznecht ZE. Porcine platelet antigens recognised by human xenoreactive natural antibodies. Transplantation 1994; 57:327–335. 27. Holzknecht ZE, Platt JL. Identification of porcine endothelial cell membrane antigens recognized by human xenoreactive natural antibodies. J Immunol 1995; 154:4565–4575. 28. Thall A, Galili U. Distribution of Gal_1A3Gal`1A4GlcNAc residues on secreted mammalian glycoproteins (thyroglobulin, fibrinogen, and immunoglobulin G) as measured by a sensitive solid-phase radioimmunoassay. Biochemistry 1990; 29:3959–3965. 29. Arumugham RG, Hsieh TC, Tanzer ML, Laine RA. Structures of the asparaginelinked sugar chains of laminin. Biochim Biophys Acta 1986; 883:112–126. 30. Joziasse DH. Mammalian glycosyltransferases: genomic organization and protein structure. Glycobiology 1992; 2:271–277. 31. Blanken WM, Van den Eijnden DH. Biosynthesis of terminal Gal_1–3Gal_1– 4GlcNAc oligosaccharide sequences on glycoconjugates. Purification and acceptor specificity of a UDP-Gal:N-acetyllactosaminide _1–3-galactosyltransferase from calf thymus. J Biol Chem 1985; 260:12,927–12,934. 32. Larsen RD, Rajan VP, Ruff M, Kukowska-Latallo J, Cummings RD, Lowe JB. Isolation of a cDNA encoding a murine UDPgalactose:`
134
39.
40.
41.
42. 43.
44.
45.
46. 47.
48. 49.
50. 51.
52.
53.
54.
55.
Sandrin and McKenzie UDP-Gal:`
Xenoantigen Synthesis
135
56. Wilmut I, Schnieke AE, McWhir J, Kind AJ, Campbell KHS. Viable offspring derived from fetal and adult mammalian cells. Nature 1997; 385:810–813. 57. Yong Z, Yuqiang L. Nuclear-cytoplasmic interaction and development of goat embryos reconstructed by nuclear transplantation: production of goats by serially cloning embryos. Biol Reprod 1998; 58:266–269. 58. Cibelli JB, Stice SL, Golueke PJ, et al. Cloned transgenic calves produced from nonquiescent fetal fibroblasts. Science 1998; 280:1256–1258. 59. Wakayama T, Perry ACF, Zuccotti M, Johnson KR, Yanagimachi R. Full-term development of mice from enucleated oocytes injected with cumulus cell nuclei. Nature 1998; 394:369–374. 60. Schnieke AE, Kind AJ, Ritchie WA, et al. Human factor IX transgenic sheep produced by transfer of nuclei from transfected fetal fibroblasts. Science 1997; 278:2130–2133. 61. Sandrin MS, Fodor WL, Mouhtouris E, et al. Enzymatic remodeling of the carbohydrate surface of a xenogenic cell substantially reduces human antibody binding and complement-mediated cytolysis. Nature Medicine 1995; 1:1261–1267. 62. Cohney S, McKenzie IFC, Patton K, et al. Down regulation of Gal_(1,3)Gal expression in mice by the _1,2fucosyltransferase transgene. Transplantation 1997; 64:495–500. 63. Sandrin MS, Fodor WL, Cohney S, et al. Reduction of the major porcine xenoantigen Gal_(1,3)Gal by expression of _(1,2)fucosyltransferase. Xenotransplantation 1996; 3:134–140. 64. Osman N, McKenzie IFC, Mouhtouris E, Sandrin MS. Switching amino terminal cytoplasmic domains of _1,2fucosyltransferase and _1,3galactosyltransferase alters the expression of H substance and Gal_(1,3)Gal. J Biol Chem 1996; 271:33,105–33,109. 65. Chen C-G, Fisicaro N, Shinkel TA, et al. Reduction in Gal_(1,3)Gal epitope expression in transgenic mice expressing human H-transferase. Xenotransplantation 1996; 3:69–75. 66. Koike C, Kannagi R, Takamura Y, et al. Introduction of _(1,2)-fucosyltransferase and its effect on _
136
Sandrin and McKenzie
72. Takemura M, Miyagawa S, Ihara Y, et al. Suppression of the xenoantigen Gal_(1,3)Gal by N-acetylglucosaminyltransferase III (GnT-III) in transgenic mice. Transplant Proc 1997; 29:895–896. 73. Cairns T, Hammelmann D, Gray K, Welsh K, Larson G. Enzymatic removal from various tissues of galactose_1,3-galactose target antigens of human antispecies antibodies. Transplant Proc 1994; 26:1279–1280. 74. Satake M, Kawagishi N, Rydberg L, et al. Limited specificity of xenoantibodies in diabetic patients transplanted with fetal porcine islet cell clusters. Main antibody reactivity against a-linked galactose-containing epitopes. Xenotransplantation 1994; 1:89–101. 75. Osman N, McKenzie IFC, Ostenreid K, Ioannou YA, Desnick RJ, Sandrin MS. Combined transgenic expression of _
Xenoantigen Synthesis
137
88. Millan MT, Geczy C, Stuhlmeier KM, Goodman DJ, Ferran C, Bach FH. Human monocytes activate porcine endothelial cells, resulting in increased E-selectin, interleukin-8, monocyte chemotactic protein-1, and plasminogen activator inhibitor type-1 expression. Transplantation 1997; 63:421–429. 89. Sepp A, Skacel P, Lindstedt R, Lechler RI. Expression of _<1,3-galactose and other type 2 oligosaccharide structure in a porcine endothelial cell line transfected with human _<1,2-fucosyltransferase. J Biol Chem 1997; 272:23,104–23,110. 90. Kwiatkowski P, Artrip JH, Edwards NM, et al. High level porcine endothelial cell expression of _(1,2)-fucosyltransferase reduces human monocyte adhesion and activation. Transplantation 1999; 67:219–226. 91. McKenzie IFC, Li YQ, Patton K, Sandrin MS. Fucosyl transferase (H) transgenic heart transplants to Gal-/- mice. Transplantation 2000; 70:1205–1209.
vi
Complement in Xenotransplantation
8
139
The Complement Barrier to Xenotransplantation Agustin P. Dalmasso, MD INTRODUCTION
Early work on xenotransplantation indicated that complement activation is a major mediator of organ xenograft rejection (1,2). This knowledge provided the impetus for numerous studies addressing complement activation in a recipient bearing a xenograft and the mechanisms of complement-mediated xenogeneic injury (3,4). In phylogenetically distant, discordant combinations, such as pig-to-primate, complement activation occurs immediately after initiation of organ reperfusion, resulting in hyperacute rejection (HAR) of the xenograft. In concordant combinations, such as hamster-to-rat, complement may participate in delayed rejection once anti-graft antibodies have been generated. Therefore, several approaches have been explored to prevent complementmediated damage to xenografts. This chapter is comprised of a brief overview of the reaction mechanisms of complement, followed by a discussion of the role of complement in the pathogenesis of tissue injury in xenograft rejection, a review of approaches to inhibit complement activation as an important component of strategies to avoid xenograft rejection, and finally, a brief discussion of the significance of complement in accommodation.
OVERVIEW OF THE COMPLEMENT SYSTEM Complement has been initially recognized as a major component of natural immunity and the main effector system of humoral immunity. In recent years, while these fundamental roles have been reaffirmed, its biological significance has been expanded to encompass a role in regulation of the immune response and in other biological processes. The From: Xenotransplantation: Basic Research and Clinical Applications Edited by: Jeffrey L. Platt © Humana Press Inc., Totowa, NJ
139
140
Dalmasso Table 1 Proteins of the Complement System
Soluble proteins Proteins with recognition, activation and/or effector function a Classical pathway: C1q (0.2) , C1r (0.4), C1s(0.4), C4 (2.0), C2 (0.2), C3 (6.3) Lectin pathway: MBLb, MASPc-1, MASP-2 Alternative pathway: C3 (6.3), B (2.5), D (0.04) Membrane attack complex: C5 (0.4), C6 (0.5), C7 (0.5), C8 (0.4), C9 (0.8) Positive regulator: P (0.5) Negative regulators: C1 inhibitor (2.6), C4-binding protein (0.4), H (3.3), I (0.4) Vitronectin (6.2), Clusterin (6.2) Membrane-associated proteins d e f Negative regulators: CR 1, DAF , MCP , CD59 Receptors: C1qR, CR1, CR2, CR3, CR4, C3aR, C5aR a
b
Plasma concentration (µM) is given in parenthesis; MBL, mannan-binding lectin; MASP, MBL-associated serine protease; dcomplement receptor; eDAF, decayf accelerating factor; MCP, membrane cofactor protein.
c
complement system is comprised of about 35 plasma and membrane proteins, including control proteins and cell membrane receptors that recognize fragments derived from complement proteins (Table 1). Most soluble complement proteins are present in plasma and other body fluids in inactive form, which require activation and sequential interaction to generate biologically active protein fragments and complexes. Two recently published books are comprehensive sources of reviews on most aspects of the complement system of current interest (5,6).
Activation Mechanisms Complement activation may take place through one of three pathways, designated the classical, alternative, and lectin pathways (Fig. 1). During the process of activation a complement protein acquires either proteolytic activity or a specific-binding property to form polymeric complexes with other complement proteins. Enzymatic activity may result from the proteolytic fragmentation of the protein, usually from the complement protein in the activation sequence preceding the protein of interest. All pathways induce formation of enzyme complexes called C3
Complement in Xenotransplantation
141
Fig. 1. Diagram pathways of complement activation. Sites of action of complement inhibitors are indicated by interrupted lines.
convertases that activate C3, and C5 convertases that activate C5. Finally, these reaction steps result in assembly of the membrane attack complex (MAC). The proteins of the classical pathway and MAC are designated “components” and are abbreviated as C1, C2, and so on. Initiation of classical pathway activation is due to binding of C1q to an activator substance, usually antigen - antibody complexes containing IgM or IgG (7,8). C1q, together with C1r, C1s, and Ca2+, constitute the C1 complex. C1 can also be activated independently of immune complexes by a variety of substances, e.g., C-reactive protein, serum amyloid protein, uric acid crystals, endotoxin and other microbial substances, and even certain mammalian subcellular components. A conformational change in C1q follows its binding to the activator and results in activation of the proenzymes C1r and C1s. As activation of C1 proceeds, C1r undergoes autocatalysis, resulting in activated C1r, that cleaves C1s. The next step of the complement reaction is the formation of C4b2a, or classical pathway C3 convertase (Fig. 1). Activated C1s acts on C4, resulting in fragments C4a and C4b; the larger product, C4b, binds to the immune complex or the cell membrane and functions as a receptor for C2, then C2 is cleaved
142
Dalmasso
by C1s into fragments C2a and C2b. C2a remains associated with C4b and acts enzymatically on C3, generating a large number of C3a and C3b fragments. After dissociation of C3a from C3b, a thiolester bond in C3b undergoes cleavage, exposing reactive groups that can bind covalently to free amino or hydroxyl groups on acceptor molecules (9). From the many C3b molecules that are produced, some bind to C4b2a, forming C4b2a3b, or C5 convertase. This enzyme then cleaves C5 into C5a and C5b, and C5b interacts with terminal complement proteins, resulting in assembly of the MAC. The proteins of the alternative pathway are designated “factors” and are abbreviated with capital letters. One form of alternative pathway activation involves a spontaneous allosteric, nonenzymatic, modification of C3 that results in cleavage of its internal thiolester bond and incorporation of H2O, yielding C3H2O. This may be the first step in the generation of an alternative pathway C3 convertase (8,10). C3H2O is able to associate with B, which now becomes susceptible to cleavage by D, yielding fragments Ba and Bb; Ba dissociates to the fluid phase, and Bb remains weakly bound to the complex. C3H2OBb can function as an alternative pathway C3 convertase that may cleave C3 into C3a and C3b. By generating C3b this process may result in assembly of a more effective alternative pathway C3 convertase as C3b binds B more stably than C3H2O does. As more C3 is turned over, cleavage of its internal thiolester bond yields the highly reactive carbonyl that binds covalently to various acceptor macromolecules. Moreover, incorporation of another C3b molecule into C3bBb results in (C3b)2Bb, or alternative pathway C5 convertase, that acts on C5 to generate C5a and C5b. The most important activators of the alternative pathway are surface components of various microorganisms; other activators include polymers of IgA. A third pathway of complement activation, called the lectin pathway, has recently been characterized (11) (Fig. 1). This pathway is independent of antibody and C1 and begins with binding of the plasma protein mannan-binding lectin (MBL) to mannose or N-acetylglucosamine on the surface of microorganisms. A second lectin in plasma, ficolin/P35, is also capable of activating the lectin pathway (12). These proteins are structurally similar to C1q and, when bound to microorganisms, can activate two C1s-like zymogens called MASP-1 and MASP-2 (MASP for MBL-associated serine protease), which in turn may cleave C2 and C4, forming the classical pathway C3 convertase. MBL and ficolin are thought to play important roles in innate immunity.
Complement in Xenotransplantation
143
If complement activation by any pathway results in generation of C5b, assembly of the MAC may take place. C5b associates with C6 and C7 to form C5b-7, which has high affinity for membrane lipid bilayers. Membrane-associated C5b-7 binds one molecule of C8, forming C5b8, which then may bind multiple molecules of C9. The strong affinity for phospholipids of C5b-8 causes perturbations of the membrane bilayer that result in increased permeability. The barrier function of the membrane is further reduced by incorporation of C9; polymerized C9 may form a highly organized tubular structure that functions as a large pore (13).
Control Mechanisms Several mechanisms participate in regulation of complement activation at every stage of the complement reaction (Fig. 1). These mechanisms have evolved to protect the host from excessive complement activation that may occur during inflammation or defense against microorganisms. This process is carried out by several plasma and membrane proteins, listed in Table 1. With a few exceptions, including C1 inhibitor (C1 inh), soluble control proteins are designated “factors” and are abbreviated with capital letters. Activation of the classical pathway is controlled by C1 inh, a protein of the serpin family that binds irreversibly to activated C1r and C1s, and blocks their enzymatic activity. C1 inh also blocks activated MASP, the contact system of kinin generation, and the intrinsic coagulation pathway. Critical plasma levels of C1 inh are required to prevent formation of pathogenic fragments derived from complement or the contact system. If these levels are not met, as in heterozygous genetic deficiencies or in acquired deficiencies, episodes of potentially life-threatening angioneurotic edema may develop (14). Another stage of the complement reaction that is under tight negative control is at the level of the C3 and C5 convertases; this is achieved through two general mechanisms (see Fig. 1). The first consists of the dissociation from the convertases of the enzymatically active proteins C2a and Bb, which to a significant extent occurs spontaneously. However, enhanced dissociation, as well as inhibition of formation of new complexes, takes place when the convertases interact with the plasma proteins C4b-binding protein (C4bp) and factor H, or with the membrane proteins decay-accelerating factor (DAF) and complement receptor 1 (CR1) (15). On the other hand, enhanced stability of the alternative pathway C3 convertase may be brought about by incorporation of the positive regulator properdin into the complex. A second mechanism of
144
Dalmasso
negative control at the C3-, C5-convertase level consists of the proteolytic degradation of C3b and C4b by factor I. Factor I can cleave its substrates only after C3b or C4b binds a cofactor protein from plasma (factor H or C4bp) or from the cell membrane (CR1 or membrane cofactor protein [MCP]). Bound C3b that is not part of a convertase can also be inactivated in this manner by factor I, yielding C3bi. The process of MAC assembly, which is the last stage of the complement reaction, is controlled by the membrane-associated protein CD59 (16). CD59 prevents MAC formation by inhibiting both the binding of C9 to C5b-8 and the incorporation of additional C9 into C5b-9. Finally the plasma proteins vitronectin and clusterin may interact with lateacting components as they assemble into complexes in plasma, preventing binding to membranes. The complement inhibitors DAF, MCP, and CD59 act intrinsically, inhibiting the convertases near the membrane area where they are located. DAF and CD59 are linked to the membrane by means of glycosyl phosphatidylinositol (GPI). The physiologic significance of these inhibitors to protect host cells against activation of autologous complement is demonstrated in patients with paroxysmal nocturnal hemoglobinuria. In this condition a population of blood cells derived from a precursor cell with an acquired mutation in the gene for an enzyme involved in GPI synthesis lacks the GPI-linked membrane proteins. Therefore, these cells lack DAF and CD59. These patients may have hemolytic anemia because their red cells are highly susceptible to complement-mediated lysis (17). The deficiency of CD59 is thought to be the important factor in the pathogenesis of this condition. MCP and CR1 are linked to the membrane by means of a hydrophobic peptide. CR1 is an elongated molecule, with binding sites for C3b and C4b that are relatively far from the cell membrane; for this reason CR1 is more effective in protecting against complement activation triggered beyond the membrane surface itself (15). Cells that are in contact with blood or other biological fluids containing complement generally express membrane-associate complement inhibitors. These proteins are inhibitory to homologous complement and in certain cases are less inhibitory to complement from a different species. In other cases they may inhibit heterologous and homologous complement with similar efficacy; for example, pig CD59 is equally effective against human and pig complement (18,19).
Complement in Xenotransplantation
145
Biological Activities Complement has a broad range of biological activities, which reside in various protein fragments and complexes generated during the course of the complement reaction. Often the complement proteins induce their activities after binding to specific membrane receptors (Table 1). The main activities of complement comprise enhancement of antibody production, solubilization and removal of immune complexes, opsonization and killing of microorganisms, and inflammation. The role of complement in enhancement of antibody production is important in the response to low doses of antigen. This effect is due to the interaction of fragments C3d,g/C3d, derived from proteolysis of C3bi, with complement receptor 2 (CR2) on B cells and follicular dendritic cells. In germinal centers the presentation of antigen to B lymphocytes is optimized by the interaction of immune complex-bound C3d,g with CR2 on the surface of follicular dendritic cells. This mechanism provides a large enhancement in antibody formation by facilitating the interaction of antigen with the B lymphocyte signaling complex formed by membrane-associated IgM, CR2, CD19, and other molecules (20). Complement plays a crucial role in prevention of immune complex formation as well as in their solubilization. These activities are mediated by binding of C1q, C3b, and C4b to the immunoglobulin molecules in the complexes (21). In addition, immune complexes are removed from the circulation when C3b that is deposited on the immune complexes binds to CR1 on red blood cells; the complexes are then transported by the red cells and cleared from the circulation by the liver and spleen. The significance of this mechanism is demonstrated by the high prevalence of immune complex disease in patients with certain complement deficiencies. In addition to preventing tissue damage by controlling formation of immune complexes, the classical pathway has been implicated in elimination of cells after they undergo apoptosis (22,23). Without an intact classical pathway apoptotic cells may persist in tissues such as the kidney, where cellular antigens may become immunogenic and contribute to the development of autoimmune diseases. Complement has evolved to protect the host against invading microorganisms. These defense mechanisms consist of opsonization and killing of pathogens, and promotion of inflammation (21). Microorganisms coated with C3b and C3bi demonstrate increased adherence to neutro-
146
Dalmasso
phils and macrophages through binding to CR1 and complement receptor 3 (CR3), respectively; then the organisms undergo phagocytosis. The importance of this complement function is demonstrated by the frequent occurrence of infections with Gram-positive bacteria in C3-deficient patients. On the other hand, C3b and C4b may neutralize virus infectivity and the MAC can mediate lysis of certain viruses and killing of Gram-negative bacteria. Individuals who are deficient in a component of the MAC have a higher incidence of neisserial septicemia than normocomplementemic individuals. Complement promotes inflammation through several mechanisms, including increase in vascular permeability, attraction of phagocytic cells, promotion of phagocytosis, and cytotoxicity. C3a and C5a induce vascular permeability, contraction of smooth muscle, release of mediators from mast cells and polymorphonuclear leukocytes, and generation of oxygen radicals. In addition C5a acts on leukocytes as a chemotactic agent, promotes cell adhesion, and stimulates the cyclooxygenase pathway. The Bb fragment acts on macrophages facilitating cell spreading and killing efficiency. The terminal components activate the metabolism of arachidonic acid and the production of oxygen radicals. The MAC may also induce a procoagulant state in platelets and endothelial cells. Although complement has potent proinflammatory effects, individuals with severe complement deficiencies do not have defective inflammation, possibly because of the overlapping functions and redundancy of various mediators of inflammation. Killing of nucleated cells by the MAC is associated with important metabolic changes in the cell undergoing complement attack, including increased intracellular Ca2+ and lipid metabolism, and mitochondrial swelling. In contrast to red cells, nucleated cells undergo irreversible damage only after deposition of a large number of MAC complexes on their membranes, owing to the ability to resist complement-mediated killing by endocytosis and exocytosis of bound MAC (24). Although the proinflammatory effects of complement are usually protective to the host, they may also participate as a major pathophysiologic mechanism in various immunologic diseases.
ROLE OF COMPLEMENT IN XENOGRAFT REJECTION When the blood circulation is reestablished in a discordant organ xenograft complement activation takes place immediately, resulting in the destruction of the organ in minutes to a few hours. Unequivocal evidence for a crucial role of complement in HAR was published 36 yr
Complement in Xenotransplantation
147
ago by Nelson (1) and Gewurz et al. (2), who demonstrated that complement inactivation with cobra venom factor (CVF) markedly extended xenograft survival. Since then numerous studies have established that organ transplants in discordant combinations undergo complementdependent HAR. In contrast, transplants in related combinations (concordant) do not experience HAR, as complement activation is not triggered until antibodies against the graft are produced. The following observations, reviewed in (3), demonstrate that complement activation mediates HAR of discordant xenografts. First, serum complement levels fall abruptly in recipients of xenogeneic organs, in conjunction with rapid elevation in plasma levels of complement activation products, such as C3a and SC5b-9. Second, complement proteins are rapidly deposited on the vascular endothelium of the xenograft. Third, inhibition of complement activation in xenograft recipients consistently results in prolongation of graft survival. Fourth, recipients genetically deficient in a complement component uniformly demonstrate prolongation of graft survival.
Pathways of Complement Activation in Hyperacute Rejection An organ xenograft can activate complement via the classical pathway or the alternative pathway, depending on the donor - recipient combination (Table 2). Various strategies have been employed to establish the pathway of complement activation in xenotransplantation, including the use of recipient animals genetically deficient in a complement component that is part of one or the other activation pathway, the use of inhibitors that selectively interfere with one pathway, and the description of the pattern of deposition of complement proteins on the vascular endothelium early during rejection. If the recipient has preformed antibodies against the graft, such as in the pig-to-primate combination, complement is activated via the classical pathway (25). A pig heart transplanted into a primate exhibited deposits of classical pathway proteins that co-localized with IgM along the vasculature of the graft, but only trace amounts of alternative pathway proteins. In combinations that result in HAR in the absence of preformed antibodies, as in rats of certain strains transplanted with a guinea pig heart, complement is activated via the alternative pathway (26). In this case complement activation is triggered directly by the vascular endothelium of the xenograft. In the pig-to-primate combination, complement is not activated directly by the xenograft. Natural antibodies bind
148
Dalmasso Table 2
Complement Activation Pathway Initially Involved in Hyperacute Rejection in Models of Organ Xenotransplantation Donor-recipient combination Pig-to-primate Rat-to-cynomolgous monkey Mouse-to-rabbit Guinea pig-to-rat Rabbit-to-newborn pig Rat-to-fetal sheep
Complement pathway
Reference
Classical Classical Classical Alternative Alternative Alternative
25 131 35 26 132 133
rapidly to the vascular endothelium of the graft and IgM, but not IgG, effectively triggers complement activation (25); IgA natural antibodies do not appear to play a role in HAR (27). In several combinations other than pig-to-primate, HAR is also mediated by activation of the classical pathway (see Table 2); some of these models may be valuable to investigate various factors that may influence graft survival by taking advantage of genetically modified rodent donors already available or that may be produced at relatively low cost.
Role of Complement in Delayed Xenograft Rejection/Acute Vascular Rejection HAR can be prevented by removal of anti-porcine antibodies or inhibition of complement activation. After several days, however, the xenograft is rejected by a process called delayed xenograft rejection or acute vascular rejection (DXR/AVR) (28,29). Although the pathogenic mechanisms of DXR/AVR are incompletely understood, antibodies are thought to play a major role. Antibodies may be able to directly activate the endothelium of the xenograft and initiate a proinflammatory response; however, it is likely that antibodies cause injury to the xenograft indirectly, by mediating the binding of Natural Killer (NK) cells and monocytes (30,31). Finally, antibodies may also initiate complement activation, with generation of biologically active complement fragments that may damage the endothelium of the xenograft, as described below. This mechanism was suggested by the observation that, when complement was inactivated with CVF in conjunction with antibody depletion, graft survival was further extended to a few weeks, in comparison to either procedure alone; in this case rejection was ac-
Complement in Xenotransplantation
149
companied by intravascular coagulation and fibrin deposition (32,33). Although in the complete absence of complement in the recipient DXR/ AVR would occur, the contribution of the local synthesis of complement proteins by the endothelium of the donor organ has not been evaluated. In most models of xenotransplantation in which complement activation is incomplete, complement contributes significantly to tissue injury in DXR/AVR. The participation of the MAC is not essential for the development of DXR/AVR because it occurs in C6-deficient animals (34,35); however, the participation of earlier stages of complement activation may be significant. Numerous in vitro studies, reviewed in the following section, indicate that products of complement activation generated at various stages of the complement reaction are able to activate the graft endothelium and the recipient’s leukocytes and platelets. These studies support the view that complement, though essential for HAR, may also participate in DXR/AVR.
COMPLEMENT IN THE PATHOPHYSIOLOGY OF XENOGENEIC TISSUE INJURY In vivo Studies and Ex vivo Perfusion Experiments In unmodified recipients of a discordant xenograft complement activation triggers the events that result in HAR through mechanisms that include loss of the endothelial cell barrier function, vasoconstriction, and promotion of coagulation. Experiments in recipient animals with genetic complement deficiencies have identified the products of complement activation that are important for HAR. Early studies showed that, in contrast to normal rabbits, C6-deficient rabbits did not immediately reject a xenogeneic organ transplant, suggesting that the MAC is the major mediator of HAR (36,37). More recently, it was found that a guinea pig heart survived for 1–2 d in C6-deficient rats, in comparison to less than 20 min in controls, and that rejection was associated with granulocyte and monocyte infiltration (34). On the other hand, the survival of a cardiac xenograft in C6-deficient rats or normal rats treated with CVF was 3–4 d (38). The longer survival in CVF-treated rats was thought to be due to the destruction of C3 and C5 that follows complement activation with CVF, precluding generation of C3a, C3bi, and C5a at the site of the graft; in this case the tissues of the graft showed less cell infiltration. The significance of C3bi in rejection was demonstrated in experiments with C6-deficient rats that received recombinant neutro-
150
Dalmasso
phil-inhibiting factor, a hookworm glycoprotein that inhibits CR3 (CD11b/CD18) and blocks binding of neutrophils and macrophages to endothelial cell-bound C3bi (39). Administration of neutrophil-inhibiting factor increased xenograft survival beyond that in control C6D rats. Experiments with ex vivo perfusion models offer the advantage that they can be used to analyze mechanisms of hyperacute injury as they may occur in the pig-to-human combination. Studies with such models have suggested that also in human recipients C5a and the MAC may play the predominant role in HAR of a porcine heart (40). It was found that, whereas a porcine heart that was perfused with unmodified human blood survived only 25 min, addition of a monoclonal antibody against human C5 to the human blood extended heart survival for at least 6 h. Moreover, the inhibition of C5 prevented the development of histologic evidence of HAR and tissue deposition of MAC. The identification of the biological effects of the MAC that are most relevant for the induction of HAR has been difficult because HAR develops very rapidly. It has been anticipated that HAR results from a few of the effects of complement that in vitro are most injurious to the integrity of endothelial cell monolayers, including endothelial cell retraction and intercellular gap formation, and endothelial cell detachment with exposure of the subendothelial matrix.
In Vitro Studies An in vitro model of xenotransplantation consisting of cultured porcine endothelial cells and human anti-porcine antibodies and complement has been extensively used to delineate the effects of activated complement on the endothelium. With this model it has been shown that normal human serum is cytotoxic to porcine endothelial cells; critical for this effect of human serum are IgM natural antibodies and the classical pathway of complement activation (25). Although IgG and IgA natural antibodies or the alternative pathway do not appear to be primarily involved, dimeric IgA natural antibodies are able to activate the alternative pathway and cause endothelial cell killing (41). Although HAR is mediated by the MAC, several products of complement activation may cause injury to a xenograft when HAR is avoided, and may contribute to vascular injury, focal ischemia, and thrombosis in DXR/AVR. The potential for complement-mediated damage to a xenograft, however, is not limited to the time period when
Complement in Xenotransplantation
151
HAR and DXR/AVR usually develop, as complement activation may take place during the lifetime of the graft. Whenever complement activation is triggered by antibodies or other complement activators originating in damaged tissues of the graft, activation fragments of complement may cause significant tissue damage. Numerous in vitro studies indicate that complement products can stimulate the endothelial cells to become activated, proinflammatory, and procoagulant (3,42). In addition, when complement is activated on the vascular endothelium of a xenograft, complement fragments are released into the plasma and may bind to blood cells of the recipient as blood circulates through the graft, resulting in activation and recruitment of leukocytes and platelets. The main effects of the active fragments of complement are discussed below (see also Table 3); however, not all the studies described have been performed with pig endothelial cells and human complement. EFFECTS OF C1q C1q, when associated with antibody, may bind to one of several receptors on endothelial cells that are specific for different segments of the C1q molecule. Binding of aggregated C1q to endothelial cells may result in expression of adhesion molecules (43,44), followed by recruitment and activation of leukocytes. Circulating platelets may also be recruited to a xenograft after modified C1q binds to receptors on the platelet surface and activates the expression of adhesion molecules (45). EFFECTS OF C3bi Following complement activation, endothelial cell-bound C3b is cleaved by factor I, resulting in C3bi. C3bi mediates binding of neutrophils to the endothelium by interacting with CR3 (CD11b/CD18) on the neutrophil surface (46). This interaction likely requires first that CR3 undergoes a conformational modification, which may be induced by leukocyte stimulation with C5a or other substances (47,48). Studies under flow conditions have shown that, following exposure of porcine endothelial cells to human serum for 0.5 or 1.5 h, C3-dependent leukocyte adhesion and transmigration takes place (49). EFFECTS OF C3a AND C5a During the course of the complement reaction the anaphylatoxins C3a and C5a are released by enzymatic fragmentation of C3 and C5, respectively. C5a has stronger biologic activity than C3a. Although the
152
Dalmasso Table 3 Summary of Biological Effects of Complement Activation Products on Endothelial Cells of the Xenograft and Blood Cells of the Recipient
Complement activation product
Effects on endothelial cells of the xenograft
Effects on blood cells of the recipient
C1q
Expression of adhesion molecules
C3bi
Binding of PMN
Platelets: Activation of gp IIb-IIIa Expression of P-selectin Macrophages: Activation
C3a, C5a
Increased permeability Loss of heparan sulfate Expression of P-selectin
Neutrophils: Chemotaxis Generation of oxygen radicals
Increase adhesiveness for leukocytes Synthesis of oxygen radicals and enzymes
Expression of elastase Aggregation Release of granular contents
a
Platelets: Aggregation Release of granular contents C5b-9
c
Cell retraction; intercellular gaps Binding of PMN IL-8 and MCP-1 Prostaglandin synthesis Expression of selectins Release of basic fibroblast growth factor, PAFc d and PDGF Membrane vesiculation Expression of von Willebrand factor and tissue factor Cell death
Platelets: Assembly of pro thrombinase complex Production of IL-1_, Membrane vesiculation
a PMN, polymorphonuclear leukocytes; b MCP-1, Monocyte chemotactic protein-1; PAF, platelet activating factor; dPDGF, platelet-derived growth factor.
Complement in Xenotransplantation
153
anaphylatoxins share certain effects, such as vasoconstriction, other effects are specific for individual anaphylatoxins. Both endothelial cells and leukocytes display the specific receptor for C5a (50,51). C5a can induce release of endothelial cell-associated heparan sulfate from cells sensitized with human anti-pig natural antibodies. This loss was considered important to the pathogenesis of the altered barrier function, procoagulant changes, and injury by oxygen and free radicals in HAR (52). This effect of C5a is thought to be due to activation of an enzyme that cleaves heparan sulfate from endothelial cells (53,54). Other effects of C5a that may be important in xenograft rejection include stimulation of endothelial cell expression of adhesion molecules and synthesis of oxygen products (51,55). C5a can also induce tissue factor in human endothelial cells (56). C3a and C5a are chemotactic for neutrophils and monocytes and activate these cells to generate oxygen radicals and express adhesion molecules (57), promoting adhesion of these cells to the endothelium (55). C5a stimulates the synthesis of elastase in neutrophils and the release of their granular contents (58). C5a also causes aggregation and the release reaction in platelets. These effects suggest that C3a and C5a may be important mediators of injury to a xenogeneic endothelium. EFFECTS OF C5b-9 MAC-mediated endothelial cell killing is not considered essential for the occurrence of HAR, and endothelial cell killing is not involved in the mechanism of DXR/AVR. However, in vitro studies have suggested that sublytic amounts of terminal components may be important in both types of rejection. Sublytic terminal components have multiple effects on endothelial cells as well as on the blood cells of the recipient; most of these effects are proinflammatory or procoagulant. A general effect of the MAC in sublytic amounts is a rise in intracellular Ca2+, which is due to both enhanced influx and mobilization from intracellular storage, resulting in cell activation (24). Binding of C5b-7, C5b-8, and C5b-9 to porcine endothelial cell monolayers causes endothelial cell retraction, with development of intercellular gaps and exposure of the subendothelium; this may lead to loss of macromolecules and blood cells, and adhesion of platelets to collagen (59). These changes are reversible upon complete assembly of the MAC, but recovery of the monolayer integrity also requires participation of substances secreted by the endothelial cells. Induction of endothelial cell retraction and intercellular pores by terminal comple-
154
Dalmasso
ment complexes may be a major mechanism for development of intercellular edema and hemorrhage during HAR. Of interest are recent results of experiments under flow conditions in which porcine endothelial cells were perfused with human serum (49). Althought perfusion for 1.5 h induced leukocyte adhesion and transmigration that was mediated by C3, perfusion during 5 h resulted in C3-independent leukocyte adhesion and transmigration. The late effect was mediated by expression of VCAM-1 and ICAM-1 following complement-mediated activation of endothelial cell NF-gB, which was possibly triggered by the action on endothelial cells of the MAC or products derived from C5a-activated leukocytes. Other effects of the MAC on endothelial cells that promote inflammation include stimulation of MCP-1 and IL-8 secretion, which comprises an early phase that requires no de novo protein synthesis, and a late phase that in part requires IL-1_ as an intermediate step (60,61). IL-1_, which acts as an autocrine factor, also induces cyclooxygenase2 and thromboxane synthase, with release of PGE2 and TXA2, that may result in vasoconstriction, disruption of cytoskeletal actin microfilaments, widening of intercellular junctions, and increased permeability (62). Another possible mechanism of injury induced by the MAC is a reduction in vascular endothelium-dependent relaxation, as has been shown for coronary arteries (63). Additional manifestations of endothelial cell activation elicited by the MAC include expression of adhesion molecules (64) and production of basic fibroblast growth factor and platelet-activating factor (65). Several mechanisms account for the prothrombotic activity of the MAC, including the induction of endothelial cell and platelet membrane vesiculation, expression of von Willebrand factor, and assembly of the prothrombinase complex (64,66). Induction of endothelial cell detachment with exposure of the subendothelium may also be important because it may result in platelet deposition and activation. The MAC may also stimulate endothelial cells to synthesize tissue factor, which occurs over a period of 16–42 h and is secondary to complement-induced release of IL-1_ (67). The biological effects of complement discussed above are thought to be caused by insertion of the terminal complement complexes into the cell membrane. However, during complement activation, the terminal components also assemble in the fluid phase; in this case the complexes are unable to insert in a cell membrane. It has been reported that these
Complement in Xenotransplantation
155
complexes possess significant biological activity upon endothelial cells, being able to stimulate human umbilical vein endothelial cells to express tissue factor and adhesion molecules of the Ig superfamily (68). This observation raises the possibility that, in a xenogeneic organ transplant, C5b-9 that is formed in the fluid phase may also contribute to tissue injury.
INHIBITION OF COMPLEMENT ACTIVATION IN XENOTRANSPLANTATION Because of the prominent role of complement in xenograft rejection, there is great interest in developing effective methods to prevent complement activation in xenotransplantation. Residual complement activation may still take place after removing anti-graft natural antibodies and inhibiting the production of elicited antibodies, after blocking antigen sites to abrogate antibody binding, or reducing the number of antigen sites on the endothelium of the donor organ. Therefore, complement inhibition is necessary, primarily through expression of membrane-associated complement regulators in the donor organ, and possibly supplemented with administration of fluid phase complement inhibitors.
Complement Inhibition with Membrane-Associated Complement Regulators Because membrane-associated complement regulators such as DAF, MCP, and CD59 are important to protect autologous tissues from activated complement, it was thought that proper expression of these proteins in the graft would prevent complement activation and HAR. In addition, this approach could also protect the graft from complement activation that might occur throughout the life of the graft, triggered by antibodies or other mechanisms. SPECIES RESTRICTION OF COMPLEMENT REGULATORS AND XENOTRANSPLANTATION The membrane complement regulators are generally considered to be efficient to protect tissues from complement of the same species but less efficient against complement of other species. Based on this species restriction it was thought that a complement regulator such as pig DAF on pig cells might not be able to efficiently protect those cells from human complement. When human DAF was incorporated into pig endothelial cell membranes, it protected the cells from the cytotoxic
156
Dalmasso
effects of human complement (69). It was also shown that non-human endothelial cells transfected with cDNA for a human complement regulator resulted in expression of the corresponding protein on the cell membrane and were protected from cytotoxicity by human complement (70,71). In transgenic pigs that express human DAF, MCP, or CD59, the inhibitors generally had broad tissue distribution, and in some cases high level of expression was achieved (72–74). Solid organs from transgenic pigs expressing human DAF, MCP, or CD59 on the vascular endothelium that were perfused with human blood showed extended survival, reduced tissue damage, and reduced deposition of complement proteins distal to the site of the inhibition of complement activation (75–77). In vivo studies with pig organs transplanted into non-human primates demonstrated that organs expressing the inhibitors did not undergo HAR in unmodified recipients (78,79). This protection may be incomplete, as the organ in hours or a few days undergoes injury that may at least in part be complement-dependent. These studies indicate that expression of complement inhibitors of the recipient species in the xenograft affords protection against HAR; this may be similar to the relative resistance of ABO-incompatible human organ allografts against HAR in the presence of low levels of anti-A or anti-B antibodies. The experiments also confirm previous results with soluble complement inhibitors showing that, in addition to inhibiting complement, other interventions are needed to control DXR/AVR and cell-mediated rejection. Studies using organs from transgenic pigs transplanted into non-human primates that were unmodified or underwent various forms of antibody depletion and immunosuppression have recently been reviewed (80). Recipients of a life-supporting transgenic pig heart or kidney transplant survived a few weeks (81,82). Kidneys from transgenic pigs that express human DAF were able to function adequately when transplanted into a primate; administration of human erythropoietin was needed to maintain hemoglobin levels (83). One way to enhance protection against complement is to use organs from transgenic animals that express two complement inhibitors; these organs were better protected than organs expressing only one inhibitor (78,84,85). Recent studies in mice evaluated the effect of expressing a human complement inhibitor together with human H-transferase to reduce expression of _gal antigens (86). It was found that lymphocytes from mice that are transgenic for both CD59 and H-transferase are better protected from lysis by human serum than cells from mice expressing only one of those proteins. Of interest, these cells were equally protected
Complement in Xenotransplantation
157
as cells that expressed CD59 and no _gal (from _gal knock-out mice). Ex vivo perfusion studies showed that reduction in _gal expression prolonged the survival of a mouse heart perfused with human plasma beyond that of hearts expressing the complement inhibitors only (87,88). COMPLEMENT INHIBITORS OF THE DONOR SPECIES MAY PROTECT THE XENOGRAFT AGAINST FOREIGN COMPLEMENT Because of its implications to xenotransplantation the ability of pig complement regulators to effectively inhibit human complement have recently been investigated. It was found that pig MCP and CD59 can inhibit lysis of red cells by human complement (89,90). Moreover, with the recent isolation and cloning of pig CD59, it was shown that pig CD59, when expressed in certain cell lines, is as effective in protecting against human complement as it is against porcine complement (18,19). However, it has not been established whether pig CD59 inhibits killing of pig endothelial cells by human complement. It will also be of interest to ascertain whether porcine DAF is effective against human complement. The finding that porcine MCP and CD59 are functionally equivalent to their human counterparts is of potential importance to xenotransplantation. Pigs express high levels of MCP and CD59, in a pattern similar to humans. This suggests that the increased protection against human complement in pig organs expressing human inhibitors may be due to the increased availability of inhibitors. Therefore it has been hypothesized that increased expression of porcine inhibitors will obtain a similar result (91). Application of this development would require overexpression of the pig complement inhibitors in a donor pig, which may require the use of transgenesis. An advantage of using porcine complement regulators is that these molecules likely do not interact with human pathogens that may use the complement regulators as receptors for cell invasion (92,93). For example, human MCP serves as receptor for the measles virus, and human DAF serves as receptor for Echo viruses, Coxsackie B viruses, and enterovirus 70. The use of pig complement regulators would avoid the possibility that transgenic animals that express human inhibitors may become susceptible to certain human pathogens.
Complement Inhibition with Soluble Molecules Even with organ xenografts from donors that express appropriate levels of membrane complement inhibitors, soluble complement inhibitors may potentially have applications to clinical xenotransplantation,
158
Dalmasso
especially in the early period following engraftment. Soluble inhibitors may contribute to protect the organ from reperfusion injury and also from the initial complement activation due to residual antibody binding to an organ with reduced xenoantigen expression or a recipient that underwent antibody removal. A problem with the use of soluble inhibitors for prolonged periods of time is that it may compromise the protective role of complement against infectious agents and persistence of immune complexes. Various groups of soluble complement inhibitors of interest to xenotransplantation are listed in Table 4. In addition to using inhibitors that are administered systemically, in the future it might be possible to engineer donor pigs that synthesize and secrete a soluble complement inhibitor in the local environment of the graft. These donors would also overexpress a membrane complement inhibitor, but the locally produced soluble inhibitor would represent another protection against complement injury. Transgenic mice have already been developed that overexpress a complement inhibitor as a soluble protein and are protected from antibody-induced glomerular injury (94). SOLUBLE FORMS OF MEMBRANE INHIBITORS Because membrane-associated complement regulators are strong inhibitors of complement activation, it was thought that soluble counterparts of those proteins might be useful. Several such proteins have been developed by recombinant technology and tested in models of xenotransplantation. Thus, soluble human CR1 (sCR1) was found to prolong the survival of a guinea pig heart graft transplanted into rats (95,96). Administration of a 15 mg/kg single dose of sCR1 to cynomolgus monkeys prolonged the survival of a porcine heart transplant to 48–90 h, from less than 1 hr in untreated controls (97). These studies were extended with continuous infusion of sCR1 at 40 mg/kg/d, which maintained low levels of complement activity and prolonged the survival of the heart transplant up to 7 d; sCR1 in combination with immunosuppression extended survival for several weeks (98). A recombinant soluble chimeric protein derived from human DAF and MCP, called CAB-2, has also been shown to inhibit complement activation (99). In vitro CAB-2 inhibited cytotoxicity of porcine endothelial cells by human serum and, ex vivo, prolonged survival of a pig heart perfused with human blood, and inhibited the generation of C3a and SC5b-9 (100). In an in vivo pig-to-rhesus monkey model a heart transplant survived up to 4 d without immunosuppression, with marked reduction in generation of C3a and SC5b-9 and tissue deposition of C3
Complement in Xenotransplantation
159
Table 4 Soluble Inhibitors of Complement Activation of Potential Value for Xenotransplantation Type of inhibitor Soluble proteins derived from membrane-associated complement inhibitors
Inhibitor sCR1
CAB-2 Cell targeted molecules
sCR1 modified by sialyl glycosylation
Other complementrelated proteins
C1 inhibitor
Mechanism of inhibition Enhances dissociation and I- mediated proteolysis of C3- and C5-convertases Similar to sCR1
Localizes inhibitor to site of proinflammatory activation Hybrid of sCD59 Localizes inhibitor to and antibody against foreign vascular surface membrane antigen
IVIG
CVF
Irreversibly binds and neutralizes activated C1r and C1s Competes with foreign endothelium for binding of C3b and C4b Activates the AP and destroys C3, C5 and other components
Antibodies against complement proteins
Anti-C5 and anti-C8 Fv anti-C5
Inhibit C5 and C8, respectively Inhibits C5
Other inhibitors of potential interest
Compstatin
Binds C3, preventing cleavage of C3 by C3 convertases Inhibits convertases, enhances C1 inh function, etc Block C1q binding sites for Igs Block Ig binding sites for C1q Inhibit C5 cleavage by C5 convertases
Heparin
Ig peptides C1q-binding peptides C5 aptamers
160
Dalmasso
and C9 (101). It has been suggested that an advantage of CAB-2 may be that it preserves the antimicrobial function of complement because CAB-2 does not promote the inactivation of C3bi by factor I (99). Recently, hybrid or modified molecules have been developed to localize the complement inhibitor to a site of interest. Modified sCR1 containing covalently attached sialyl Lewis x was more efficient than unmodified sCR1 in reducing tissue injury in a mouse model of cerebral ischemia (102). Another fusion protein that may be of interest in xenotransplantation would target the inhibitor to the endothelial cell with an antibody fragment that recognizes an antigen of importance in triggering xenogeneic injury. The antibody fragment would block antibody binding but would not activate complement or bind to Fc receptors on recipient cells, while the complement inhibitor would function as in its membrane-associated form. A model to test this concept has been developed, employing soluble CD59 fused to an antibody-combining site from IgG with specificity for a hapten (103). The fusion protein specifically bound to Chinese hamster ovary cells that carried the hapten and provided protection against complement-mediated lysis. OTHER COMPLEMENT-RELATED PROTEINS The use of human proteins to inhibit complement clinically has the advantage that they are usually non-antigenic and, when necessary, can be used repeatedly. C1 inh is of interest for pig-to-primate transplantation because it selectively inhibits the classical pathway. Large amounts of C1 inh were needed in vitro for complete inhibition of complementmediated cytotoxicity (104); the requirement of a large amount of C1 inh reflects that C1 inh is only moderately effective with strong C1 activators such as antigen - antibody complexes. However, lower amounts of C1 inh were sufficient when used together with heparin; these substances inhibited complement activation synergistically (105). In an ex vivo perfusion model C1 inh reduced tissue injury of a pig kidney perfused with human blood (106). Intravenous IgG (IVIG) in large doses inhibited complement effectively, especially the classical pathway. It has been shown that administration of IVIG prolongs graft survival in guinea pig-to-rat (107,108) and in pig-to-primate combinations (109). CVF is a cobra glycoprotein analogous to C3b that has been used extensively in xenotransplantation. However, CVF is not suitable for clinical xenotransplantation because its administration results in generation of C3a and C5a that may cause
Complement in Xenotransplantation
161
pulmonary injury and other complications, and also elicits antibodies that make it ineffective after several days of administration (110). OTHER INHIBITORS Antibodies against certain complement components, especially recombinant humanized monoclonal antibodies, may be useful for protecting a xenograft (111). An anti-human C5 monoclonal antibody suppressed HAR in an ex vivo model of pig-to-human xenografting (40) and a monoclonal antibody against C8 protected a rat heart from damage caused by perfusion with human blood (112). A humanized, recombinant, single-chain Fv anti-human C5 was tested in patients undergoing cardiopulmonary bypass and found to effectively inhibit formation of SC5b-9, expression of leukocyte CD11b, and postoperative myocardial injury (113). These findings suggest that inhibition of C5 activation with a blocking antibody against C5 may represent a useful strategy to prevent MAC-mediated tissue injury. Compstatin is a 13-amino acid cyclic peptide that was found to prolong organ survival of a pig kidney perfused ex vivo with human blood; the compound inhibited generation of complement activation fragments and prevented tissue deposition of C3 and MAC (114). Heparin was tested as complement inhibitor in a guinea pig-to-rat model, where it prolonged the survival of the transplant and reduced the amount of C3b bound to the heart; it also inhibited the release of heparan sulfate from the transplanted heart (115). It was recently reported that a hexadecemeric multiple peptide of the C1q-binding site from human IgG1 inhibits lysis of pig red cells by human complement, with an I50 value of 1 µM (116). On the other hand, a C1q fragment that blocks the binding of IgM and IgG to C1q may also be of interest (117). An advantage of the latter two approaches for pig-to-primate transplantation is that they attempt to prevent classical pathway complement activation at its earliest stage. A new approach to inhibit complement activation with RNA aptamers consists of the derivation of RNA aptamers that bind C5 (118). One inhibitor aptamer was produced with a Kd of 2–5 nM.
COMPLEMENT IN ACCOMMODATION The term accommodation is used to describe the survival of a grafted organ in a recipient with anti-graft antibodies and normal complement, as may be observed in human renal allotransplantation across the ABO
162
Dalmasso
barrier or in patients with anti-HLA antibodies, and in certain instances of concordant xenotransplantation (119,120). This section discusses the significance of complement to accommodation. It is known that a proportion of cases of ABO-incompatible renal allografts in patients receiving conventional immunosuppression does not undergo HAR, especially if antibody levels are low. Moreover, HAR could be prevented in additional patients by removal of antibodies before and during a short time following engraftment (121,122); it was hypothesized that in these cases accommodation develops, as the transplant is not rejected when antibody levels returned after discontinuation of antibody removal (119,123). It is likely that complement regulators of the graft may play a role in protection against complement-mediated rejection while antibody levels are low, permitting the development of adaptive changes that may be essential for accommodation. Accommodation has been demonstrated in rodent models of concordant cardiac xenotransplantation. In these models, if complementdependent graft rejection is avoided by complement inactivation with a single dose of CVF and T-cell rejection is prevented, the graft survives in most recipients (124). A critical factor in accommodation is thought to be the overexpression in the tissues of the accommodated graft of genes for protective antioxidant and antiapoptotic proteins, especially heme oxygenase-1 (120,125). Induction of accommodation required inactivation of complement with CVF to avoid rejection by complement-fixing antibodies. However, it was found that accommodation could be induced in C6-deficient rats and in normocomplementemic rats given anti-C6 antibodies without administration of CVF (126). Therefore, the only role of CVF as part of the protocol to induce accommodation was the inactivation of MAC proteins; inactivation of C3, other components of the alternative pathway, or C5 is not required. The current understanding of accommodation in rodents suggests some of the modifications that would be required to achieve accommodation in pig-to-primate combinations. These modifications include inhibition of complement activation and induction of protective mechanisms against injury in the tissues of the graft, especially in the vascular endothelium. In vitro studies on induction of protection against MACmediated injury in pig endothelial cells suggest requirements for accommodation in primates that may be similar to those in rodents. Stimulation of pig endothelial cells with _gal-binding agonists elicited a late response that protects the cells against the effects of the MAC, and possibly also from cell-mediated damage (127–130). Resistance to
Complement in Xenotransplantation
163
complement was found to be a long-lasting manifestation of a slow activation process that requires protein synthesis and is associated with increased expression of mRNA for the MAC inhibitor CD59 and the corresponding membrane-associated protein (127,128). Protected cells also manifested overexpression of heme oxygenase-1 (120). These changes induced in vitro in porcine endothelial cells represent activation processes that may be of fundamental importance for the possibility of inducing accommodation in pig-to-primate models.
CONCLUSION Recent work on the role of complement in xenotransplantation has established that the MAC alone is responsible for HAR. However, when HAR is averted, various products of complement activation may play a role in the pathophysiology of xenograft tissue injury. The use of organs from transgenic donor pigs that express membrane-associated human complement regulators represents a major advance in xenotransplantation research, as these organs do not undergo HAR. It has been reported, however, that porcine complement regulators may be as effective as the corresponding human proteins to inhibit activation of human complement. This finding suggests that enhanced expression of the donor pig’s own complement regulators may be equally effective as donors that express the human proteins. Therefore, an area of major interest is to develop pigs engineered for overexpression of their own complement regulators. Then it will be possible to compare organs from these donors and from donors expressing the human inhibitors in pig-toprimate transplantation models. This approach, together with the engineering of donors to express reduced amounts of xenoantigens, and induction of immunological tolerance or accommodation, should provide continuing progress in the field. Regarding soluble complement inhibitors, the success that has been achieved to abrogate HAR in experimental animals suggests that they may be useful to reduce reperfusion injury and contribute to prevention of HAR when given before and soon after a transplant. Much work is needed to develop small nontoxic molecules that inhibit complement efficiently. Therefore, it appears likely that, as in the past, xenotransplantation research will continue to expand with regards to delineating the pathophysiologic role of complement in graft rejection and accommodation, as well as new approaches to more effectively inhibit complement activation.
164
Dalmasso
ACKNOWLEDGMENTS Work performed in the author’s laboratory has been supported by the National Institutes of Health, the Department of Veterans Affairs Medical Research, and the Minnesota Medical Foundation.
REFERENCES 1. Nelson RA. A new concept of immunosuppression in the hypersensitivity reactions and in transplantation immunity. Surv Ophthalmol 1966; 11:498–505. 2. Gewurz H, Clark DS, Cooper MD, Varco RL, Good RA. Effect of cobra venominduced inhibition of complement activity on allograft and xenograft rejection reactions. Transplantation 1967; 5:1296–1303. 3. Dalmasso AP. The role of complement in xenograft rejection. In: Cooper DKC, Kemp E, Platt JL, White DJG, eds. Xenotransplantation. The transplantation of organs and tissues between species, 2nd ed. Berlin: Springer-Verlag, 1997, 38–60. 4. Auchincloss H, Jr. Xenografting: a review. Transplant Rev 1990; 4:14–27. 5. Volanakis JE, Frank MM. The human complement system in health and disease. New York: Marcel Dekker, 1998. 6. Rother K, Till GO, Hansch GM. The complement system. Berlin: Springer, 1998. 7. Cooper NR. The classical complement pathway: activation and regulation of the first complement component. Adv Immunol 1985; 37:151–216. 8. Müller-Eberhard HJ. Molecular organization and function of the complement system. Ann Rev Biochem 1988; 57:321–347. 9. Levine RP, Dodds AW. The thioester bond of C3. Curr Topics Microbiol Immunol 1990; 153:73–82. 10. Pangburn MK, Müller-Eberhard HJ. The alternative pathway of complement. Springer Semin Immunopathol 1984; 7:163–203. 11. Matsushita M. The lectin pathway of the complement system. Microbiol Immunol 1996; 40:887–893. 12. Matsushita M, Endo Y, Fujita T. Complement-activating complex of ficolin and mannose-binding lectin-associated serine protease. J Immunol 2000; 164:2281–2284. 13. Dalmasso AP, Benson BA. Pore size of lesions induced by complement on red cell membranes and its relation to C5b-8, C5b-9 and poly C9. In: Podack ER, ed. Cytolytic lymphocytes and complement: Effectors of the immune system Vol. 1. Boca Raton: CRC Press, 1988: 207–219. 14. Davis AE. C1 inhibitor and hereditary angioneurotic edema. Ann Rev Immunol 1988; 6:595–628. 15. Hourcade D, Holers VM, Atkinson JP. The regulators of complement activation (RCA) gene cluster. Adv Immunol 1989; 45:381–416. 16. Meri S. Protectin (CD59). Complement lysis inhibitor and prototype domain in a new protein superfamily. The Immunologist 1994; 2:149–155. 17. Rosse WF, Ware RE. The molecular basis of paroxysmal nocturnal hemoglobinuria. Blood 1995; 86:3277–3286. 18. Hinchliffe SJ, Rushmere NK, Hanna SM, Morgan BP. Molecular cloning and functional characterization of the pig analogue of CD59: relevance to xenotransplantation. J Immunol 1998; 160:3924–3932.
Complement in Xenotransplantation
165
19. Maher SE, Pflugh DL, Larsen NJ, Rothschild MF, Bothwell AL. Structure/function characterization of porcine CD59: expression, chromosomal mapping, complement–inhibition, and costimulatory activity. Transplantation 1998; 66:1094–1100. 20. Fearon DT, Carter RH. The CD19/CR2/TAPA-1 complex of B lymphocytes: linking natural to acquired immunity. Ann Rev Immunol 1995; 13:127–149. 21. Morgan BP. Physiology and pathophysiology of complement: progress and trends. Crit Rev Clin Lab Sci 1995; 32:265–298. 22. Botto M, Dell’Agnola C, Bygrave AE, et al. Homozygous C1q deficiency causes glomerulonephritis associated with multiple apoptotic bodies. Nature Genetics 1988; 19:56–59. 23. Korb LC, Ahearn JM. C1q binds directly and specifically to surface blebs of apoptotic human keratinocytes: complement deficiency and systemic lupus erythematosus revisited. J Immunol 1997; 158:4525–4528. 24. Morgan BP. (1993) Cellular responses to the membrane attack complex. In: Whaley K, Loos M, Weiler JM, eds. Complement in health and disease Vol. 20. Boston: Kluwer Academic 1993: 325–351. 25. Dalmasso AP, Vercellotti GM, Fischel RJ, Bolman RM, Bach FH, Platt JL. Mechanism of complement activation in the hyperacute rejection of porcine organs transplanted into primate recipients. Am J Pathol 1992; 140:1157–1166. 26. Miyagawa S, Hirose H, Shirakura R, et al. The mechanism of discordant xenograft rejection. Transplantation 1998; 46:825–830. 27. Kroshus TJ, Bolman RM, Dalmasso AP. Selective IgM depletion prolongs organ survival in an ex vivo model of pig-to-human xenotransplantation. Transplantation 1996; 62:5–12. 28. Leventhal JR, Dalmasso AP, Cromwell J, et al. Prolongation of cardiac xenograft survival by depletion of complement. Transplantation 1993; 55:857–865. 29. Leventhal JR, Matas AJ, Sun, L.H., et al. The immunopathology of cardiac xenograft rejection in the guinea pig-to-rat model. Transplantation 1993; 56:1–8. 30. Fryer J, Leventhal JR, Dalmasso AP, et al. Cellular rejection in a discordant xenograft when hyperacute rejection is prevented: analysis using adoptive and passive transfer. Transplant Immunol 1994; 2:87–93. 31. Fryer JP, Leventhal JR, Dalmasso AP, et al. Beyond hyperacute rejection. Accelerated rejection in a discordant xenograft model by adoptive transfer of specific cell subsets. Transplantation 1995; 59:171–176. 32. Leventhal JR, Sakiyalak P, Witson J, et al.The synergistic effect of combined antibody and complement depletion on discordant cardiac xenograft survival in nonhuman primates. Transplantation 1994; 57:974–978. 33. Ierino FL, Kozlowski T, Siegel JB, et al. Disseminated intravascular coagulation in association with the delayed rejection of pig-to-baboon renal xenografts. Transplantation 1998; 66:1439–1450. 34. Brauer RB, Baldwin WMI, Daha MR, Pruitt SK, Sanfilippo F. Use of C6-deficient rats to evaluate the mechanism of hyperacute rejection of discordant cardiac xenografts. J Immunol 1993; 151:7240–7248. 35. Kerr SR, Dalmasso AP, Apasova EV, Chen SS, Kirschfink M, Matas AJ. Mouseto-rabbit xenotransplantation: a new small animal model of hyperacute rejection mediated by the classical complement pathway. Transplantation 1999;67:360–365. 36. Chartrand C, O’Regan S, Robitaille P, Pinto-Blonde M. Delayed rejection of cardiac xenografts in C6-deficient rabbits. Immunology 1979; 38:245–248.
166
Dalmasso
37. Zhow XJ, Niesen N, Pawlowski I, et al. Prolongation of survival of discordant kidney xenografts by C6 deficiency. Transplantation 1990; 50:896–898. 38. Johnson EM, Leventhal J, Dalmasso AP, et al. Inactivation of C3 and C5 prolongs cardiac xenograft survival and decreases leukocyte infiltration in a model of delayed xenograft rejection. Transplant Proc 1996; 28:603. 39. Johnson EM, Leventhal J, Dalmasso AP et al.Use of a novel CD11b/CD18 inhibitory agent in a C6 deficient rat to evaluate delayed xenograft rejection. Transplant Proc 1996; 28:728. 40. Kroshus TJ, Rollins SA, Dalmasso AP, et al. Complement inhibition with an antiC5 monoclonal antibody prevents acute cardiac tissue injury in an ex vivo model of pig-to-human xenotransplantation. Transplantation 1995; 60:1194–1202. 41. Schaapherder AFM, Gooszen HG, Te Bulte MTJW, Daha MR. Human complement activation via the alternative pathway on porcine endothelium initiated by IgA antibodies. Transplantation 1995; 60;287–291. 42. Saadi S, Platt JL. Endothelial cell responses to complement activation. In: Volanakis JE, Frank MM, eds. The human complement system in health and disease. New York: Marcel Dekker, 1998: 335–353. 43. Guo WX, Ghebrehiwet B, Weksler B, Schweitzer K, Peerschke EI. Up-regulation of endothelial cell binding proteins/receptors for complement component C1q by inflammatory cytokines. J Lab Clin Med 1999; 133:541–550. 44. Lozada C, Levin RI, Huie M, et al. Identification of C1q as the heat-labile serum cofactor required for immune complexes to stimulate endothelial expression of the adhesion molecules E-selectin and intercellular and vascular cell adhesion molecules 1. Proc Natl Acad Sci USA 1995; 92:8378–8382. 45. Peerschke EIB, Reid KBM, Ghebrehiwet B. Platelet activation by C1q results in the induction of _IIb/`3 integrins (GPIIb-IIIa) and the expression of P-selectin and procoagulant activity. J Exp Med 1993; 178:579–587. 46. Vercellotti GM, Platt JL, Bach FH, Dalmasso AP. Neutrophil adhesion to xenogeneic endothelium via iC3b. J Immunol 1991; 146:730–734. 47. Oxvig C, Lu C, Springer, TA. Conformational changes in tertiary structure near the ligand binding site of an integrin I domain. Proc Natl Acad Sci USA 1999; 96:2215–2220. 48. Vetvicka V, Thornton BP, Ross GD. Soluble beta-glucan polysaccharide binding to the lectin site of neutrophil or natural killer cell complement receptor type 3 (CD11b/ CD18) generates a primed state of the receptor capable of mediating cytotoxicity of iC3b-opsonized target cells. J Clin Invest 1996; 98:50–61. 49. Morigi M, Zoja C, Colleoni S, et al. Xenogeneic serum promotes leukocyte-endothelium interaction under flow through two temporally distinct pathways: role of complement and nuclear factor-gB. J Am Soc Nephrol 1999; 10:2197–2207. 50. Haviland DL, McCoy RL, Whitehead WT, et al. Cellular expression of the C5a anaphylatoxin receptor (C5aR): demonstration of C5aR on nonmyeloid cells of the liver and lung. J Immunol 1995; 154:1861–1869. 51. Murphy HS, Shayman JA, Till GO, et al. Superoxide responses of endothelial cells to C5a and TNF-alpha: divergent signal transduction pathways. Am J Physiol 1992; 263:L51–59. 52. Platt JL, Dalmasso AP, Lindman BJ, Ihrcke NS, Bach FH. The role of C5a and antibody in the release of heparan sulfate from endothelial cells. Eur J Immunol 1991; 21:2887–2890.
Complement in Xenotransplantation
167
53. Ihrcke NS, Platt JL. Shedding of heparan sulfate proteoglycan by stimulated endothelial cells: evidence for proteolysis of cell-surface molecules. J Cell Physiol 1996; 168:625–637. 54. Platt JL, Vercellotti GM, Lindman B, Oegema TR, Jr, Bach FH, Dalmasso AP. Release of heparan sulfate from endothelial cells: implications for pathogenesis of hyperacute rejection. J Exp Med 1990; 171:1363–1368. 55. Foreman KE, Glovsky MM, Warner RL, Horvath SJ, Ward PA. Comparative effect of C3a and C5a on adhesion molecule expression on neutrophils and endothelial cells. Inflammation 1996; 20:1–9. 56. Ikeda K, Nagasawa K, Horiuchi T, Tsuru T, Nishizaka H, Niho Y. C5a induces tissue factor activity on endothelial cells. Thromb Haemost 1997; 77:394–398. 57. Varani J, Ginsburg I, Schuger L, et al. Endothelial cell killing by neutrophils. Synergistic interaction of oxygen products and proteases. Am J Pathol 1989; 135:435–438. 58. Owen CA, Campbell MA, Sannes PL, Boukedes SS, Campbell EJ. Cell surfacebound elastase and cathepsin G on human neutrophils: a novel, non-oxidative mechanism by which neutrophils focus and preserve catalytic activity of serine proteinases. J Cell Biol 1995; 131:775–789. 59. Saadi S, Platt JL. Transient perturbation of endothelial integrity induced by natural antibodies and complement. J Exp Med 1995; 181:21–31. 60. Kilgore KS, Flory CM, Miller BF, Evans VM, Warren JS. The membrane attack complex of complement induces interleukin-8 and monocyte chemoattractant protein-1 secretion from human umbilical vein endothelial cells. Am J Pathol 1996; 149:953–961. 61. Selvan RS, Kapadia HB, Platt JL. Complement-induced expression of chemokine genes in endothelium: regulation by IL-1-dependent and -independent mechanisms. J Immunol 1998; 161:4388–4395. 62. Bustos M, Coffman TM, Saadi S, Platt JL. Modulation of eicosanoid metabolism in endothelial cells in a xenograft model. Role of cyclooxygenase-2. J Clin Invest 1997; 100:1150–1158. 63. Stahl GL, Reenstra WR, Frendl G. Complement-mediated loss of endotheliumdependent relaxation of porcine coronary arteries. Role of the terminal membrane attack complex. Circ Res 1995; 76:575–583. 64. Hattori R, Hamilton KK, McEver RP, Sims PJ. Complement proteins C5b-9 induce secretion of high molecular weight multimers of endothelial von Willebrand factor and translocation of granule membrane protein GMP-140 to the cell surface. J Biol Chem 1989; 264:9053–9060. 65. Benzaquen LR, Nicholson-Weller A, Halperin JA. Terminal complement proteins C5b-9 release basic fibroblast growth factor and platelet-derived growth factor from endothelial cells. J Exp Med 1994; 179:985–992. 66. Hamilton KK, Hattori R, Esmon CT, Sims PJ. Complement proteins C5b-9 induce vesiculation of the endothelial plasma membrane and expose catalytic surface for assembly of the prothrombinase enzyme complex. J Biol Chem 1990; 265:3809–3814. 67. Saadi S, Holzknecht RA, Patte C, Stern DM, Platt JL. Complement-mediated regulation of tissue factor activity in endothelium. J Exp Med 1995; 182: 1807–1814. 68. Tedesco F, Pausa M, Nardon E, Introna M, Mantovani A, Dobrina A. The cytolytically inactive terminal complement complex activates endothelial cells to
168
69.
70. 71.
72.
73.
74.
75.
76.
77.
78.
79.
80. 81. 82.
83. 84.
Dalmasso express adhesion molecules and tissue factor procoagulant activity. J Exp Med 1997; 185:1619–1627. Dalmasso AP, Vercellotti GM, Platt JL, Bach FH. Inhibition of complement-mediated endothelial cell cytotoxicity by decay accelerating factor. Potential for prevention of xenograft hyperacute rejection. Transplantation 1991; 52:530–533. Akami T, Sawada R, Minato N, et al. Cytoprotective effect of CD59 antigen on xenotransplantation immunity. Transplant Proc 1992; 24:485–487. Charreau B, Cassard A, Tesson L, et al.Protection of rat endothelial cells from primate complement-mediated lysis by expression of human CD59 and/or decayaccelerating factor. Transplantation 1994; 58:1222–1229. Fodor WL, Williams BL, Matis LA, et al. Expression of a functional human complement inhibitor in a transgenic pig as a model for the prevention of xenogeneic hyperacute organ rejection. Proc Natl Acad Sci USA 1994; 91:11,153–11,157. McCurry KR, Kooyman DL, Alvarado CG, et al. Human complement regulatory proteins protect swine-to-primate cardiac xenografts from humoral injury. Nature Med 1995; 1:423–427. Rosengard AM, Cary NRB, Langford GA, Tucker AW, Wallwork J, White DJG. Tissue expression of human complement inhibitor, decay-accelerating factor, in transgenic pigs: a potential approach for preventing xenograft rejection. Transplantation 1995; 59:1325–1333. Daggett CW, Yeatman M, Lodge AJ, et al. Swine lungs expressing human complement-regulatory proteins are protected against acute pulmonary dysfunction in a human plasma perfusion model. J Thoracic Cardiovasc Surg 1997; 113:390–398. Kroshus TJ, Bolman RM, Dalmasso AP, et al. Expression of human CD59 in transgenic pig organs enhances organ survival in an ex vivo xenogeneic perfusion model. Transplantation 1996; 61:1513–1521. Schmoeckel M, Nollert G, Shahmohammadi M, et al. Transgenic human decay accelerating factor makes normal pigs function as a concordant species. J Heart Lung Transplant 1997; 16:758–764. Byrne GW, McCurry KR, Martin MJ, McClellan SM, Platt JL, Logan JS. Transgenic pigs expressing human CD59 and decay-accelerating factor produce an intrinsic barrier to complement-mediated damage. Transplantation 1997; 63:149–155. Cozzi E, Yannoutsos Y, Langford GA, Pino-Chavez G, Wallwork J, White DJG. Effect of transgenic expression of human decay-accelerating factor on the inhibition of hyperacute rejection of pig organs. In: Cooper DKC, Kemp E, Platt JL, White DJG, eds. Xenotransplantation. The transplantation of organs and tissues between species 2nd ed. Berlin: Springer, 1997: 665–682. Lambrigts D, Sachs DH, Cooper DK. Discordant organ xenotransplantation in primates: world experience and current status. Transplantation 1998; 66:547–561. Schmoeckel M, Bhatti FN, Zaidi A, et al. Orthotopic heart transplantation in a transgenic pig-to-primate model. Transplantation 1998; 65:1570–1577. Zaidi A, Schmoeckel M, Bhatti F, et al. Life-supporting pig-to-primate renal xenotransplantation using genetically modified donors. Transplantation 1998; 65:1584–1590. Zaidi A, Bhatti F, Schmoeckel M, et al. Kidneys from HDAF transgenic pigs are physiologically compatible with primates. Transplant Proc 1998; 30:2465–2466. Cowan PJ, Somerville CA, Shinkel TA, et al. High-level endothelial expression of human CD59 prolongs heart function in an ex vivo model of xenograft rejection. Transplantation 1998; 65:826–831.
Complement in Xenotransplantation 85.
169
Mulder LCF, Mora M, Lazzeri M, et al. Human MCP and DAF double transgenic mice are protected from human complement attack in an in vivo model. Transplant Proc 1996; 28:589. 86. Costa C, Zhao L, Decesare S, Fodor WL. Comparative analysis of three genetic modifications designed to inhibit human serum-mediated cytolysis. Xenotransplantation 1999; 6:6–16. 87. Cowan PJ, Chen CG, Shinkel TA, et al. Knock out of alpha1,3galactosyltransferase or expression of alpha1,2-fucosyltransferase further protects CD55- and CD59-expressing mouse hearts in an ex vivo model of xenograft rejection. Transplantation 1998; 65:1599–1604. 88. van Denderen BJ, Salvaris E, Romanella M, et al. Combination of decay-accelerating factor expression and alpha1,3-galactosyltransferase knockout affords added protection from human complement-mediated injury. Transplantation 1997; 64:882–888. 89. van den Berg, C.W. and Morgan, B.P. (1994) Complement-inhibiting activities of human CD59 and analogues from rat, sheep and pig are not homologously restricted. J. Immunol., 152, 4095–4101. 90. van den Berg CW, Perez de la Lastra JM, Llanes D, Morgan BP. Purification and characterization of the pig analogue of human membrane cofactor protein (CD46/ MCP). J Immunol 1997; 158:1703–1709. 91. Perez de la Lastra JM, Hanna SM, Morgan BP. Distribution of membrane cofactor protein (MCP/CD46) on pig tissues. Relevance to xenotransplantation. Immunology 1999; 98:144–151. 92. Pascual M, French LE. Complement in human diseases: looking towards the 21st century. Immunol. Today 1995; 16:58–61. 93. Thorley BR, Milland J, Christiansen D, et al. Transgenic expression of a CD46 (membrane cofactor protein) minigene: studies of xenotransplantation and measles virus infection. Eur J Immunol 1997; 27:726–734. 94. Quigg RJ, He C, Lim A, et al. Transgenic mice overexpressing the complement inhibitor crry as a soluble protein are protected from antibody-induced glomerular injury. J Exp Med 1998; 188:1321–1331. 95. Pruitt SK, Baldwin WM, Marsh HC, Jr, Lin SS, Yeh CG, Bollinger RR. The effect of soluble complement receptor type 1 on hyperacute xenograft rejection. Transplantation 1991; 52:868–873. 96. Xia W, Fearon DT, Moore FD, Jr, Schoen FJ, Ortiz F, Kirkman RL. Prolongation of guinea pig cardiac xenograft survival in rats by soluble human complement receptor type 1. Transplant Proc 1992; 24:479–480. 97. Pruitt SK, Kirk AD, Bollinger RR, et al. The effect of soluble complement receptor type 1 on hyperacute rejection of porcine xenografts. Transplantation 1994; 57:363–370. 98. Pruitt SK, Bollinger RR, Collins BH, et al. Effect of continuous complement inhibition using soluble complement receptor type 1 on survival of pig-to-primate cardiac xenografts. Transplantation 1997; 63:900–902. 99. Higgins PJ, Ko JL, Lobell R, Sardonini C, Alessi MK, Yeh CG. A soluble chimeric complement inhibitory protein that possesses both decay-accelerating and factor I cofactor activities. J Immunol 1997; 158:2872–2881. 100. Kroshus TJ, Salerno CT, Yeh CG, Higgins PJ, Bolman RM, Dalmasso AP. A recombinant soluble chimeric complement inhibitor composed of human CD46
170
101.
102. 103. 104.
105.
106.
107.
108.
109.
110.
111. 112.
113.
114.
115.
116.
Dalmasso and CD55 reduces acute cardiac tissue injury in models of pig-to-primate heart transplantation. Transplantation 2000; 69:2282–2289. Salerno CT, Dalmasso AP, Kroshus TJ, et al. A recombinant soluble chimeric complement inhibitor CAB-2.0 prolongs xenograft survival in a heterotopic model of pig-to-primate cardiac transplantation. Surg Forum 1997; 48:265–267. Huang J, Kim LJ, Mealey R, et al. Neuronal protection in stroke by an sLexglycosylated complement inhibitory protein. Science 1999; 285:595–599. Zhang H, Yu J, Bajawa E, Morrison S.L, Tomlinson S. Targeting of functional antibody-CD59 fusion proteins to a cell surface. J Clin Invest 1999; 103:55–61. Dalmasso AP, Platt JL. Prevention of complement-mediated activation of xenogeneic endothelial cells in an in vitro model of xenograft hyperacute rejection by C1 inhibitor. Transplantation 1993; 56:1171–1176. Dalmasso AP, Platt JL. Potentiation of C1 inhibitor plus heparin in prevention of complement-mediated activation of endothelial cells in a model of xenograft hyperacute rejection. Transplant Proc 1994; 26:1246–1247. Fiane AE, Videm V, Johansen HT, Mellbye OJ, Nielsen EW, Mollnes TE. C1inhibitor attenuates hyperacute rejection and inhibits complement, leukocyte and platelet activation in an ex vivo pig-to-human perfusion model. Immunopharmacology 1999; 42:231–243. Gautreau C, Kojima T, Woimant G, Cardoso J, Devillier P, Houssin D. Use of intravenous immunoglobulin to delay xenogeneic hyperacute rejection. An in vivo and in vitro evaluation. Transplantation 1995; 60:903–907. Latremouille C, Haeffner-Cavaillon N, Goussef N, et al. Normal human polyclonal immunoglobulins for intravenous use significantly delay hyperacute xenograft rejection. Transplant Proc 1994; 26:1285. Magee JC, Collins BH, Harland RC, et al. Immunoglobulin prevents complement-mediated hyperacute rejection in swine-to-primate xenotransplantation. J Clin Invest 1995; 96:2404–2412. Taniguchi S, Kobayashi T, Neethling FA, et al. Cobra venom factor stimulates anti-alpha-galactose antibody production in baboons. Implications for pig-tohuman xenotransplantation. Transplantation 1996; 62:678–681. Matis LA, Rollins SA. Complement-specific antibodies: Designing novel antiinflammatories. Nature Med 1995; 1:839–841. Rollins SA, Matis LA, Springhorn JP, Setter E, Wolff DW. Monoclonal antibodies directed against human C5 and C8 block complement-mediated damage of xenogeneic cells and organs. Transplantation 1995; 60:1284–1292. Fitch JC, Rollins S, Matis L, et al. Pharmacology and biological efficacy of a recombinant, humanized, single-chain antibody C5 complement inhibitor in patients undergoing coronary artery bypass graft surgery with cardiopulmonary bypass. Circulation 1999; 100:2499–2506. Fiane AE, Mollnes TE, Videm V, et al. Compstatin, a peptide inhibitor of C3, prolongs survival of ex vivo perfused pig xenografts. Xenotransplantation 1999; 6:52–65. Stevens RB, Wang YL, Kaji H, et al. Administration of nonanticoagulant heparin inhibits the loss of glycosaminoglycans from xenogeneic cardiac grafts and prolongs graft survival. Transplant Proc 1993; 25:382. Kojima T, Del Carpio CA, Tajiri H, Yoshikawa K, Saga S, Yokoyama I. Inhibition of complement-mediated immune hemolysis by peptides derived from the constant domain of immunoglobulin. Transplantation 1999; 67:637–638.
Complement in Xenotransplantation
171
117. Kishore U, Perdikoulis MV, Strong P, Reid KBM. A homotrimer of the B-chain globular head region of Clq is an inhibitor of Clq-mediated complement activation. Mol Immunol 1998; 35:375. 118. Biesecker G, Dihel L, Enney K, Bendele RA. Derivation of RNA aptamer inhibitors of human complement C5. Immunopharmacology 1999; 42:219–230. 119. Bach FH, Turman MA, Vercellotti GM, Platt JL, Dalmasso AP. Accommodation: a working paradigm for progressing toward clinical discordant xenografting. Transplant Proc 1991; 23:205–207. 120. Soares MP, Lin Y, Sato K, Stuhlmeier KM, Bach FH. Accommodation. Immunol Today 1999; 20:434–437. 121. Alexandre GPJ, Squifflet JP, De Bruyère M, et al. Present experiences in a series of 26 ABO-incompatible living donor renal allografts. Transplant Proc 1987; 19:4538–4542. 122. Slapak M, Digard N, Ahmed M, Shell T, Thompson F. Renal transplantation across the ABO barrier: a 9-year experience. Transplant Proc 1990; 22:1425–1428. 123. Bannett AD, McAlack RF, Morris M, Chopek MW, Platt JL. ABO incompatible renal transplantation: a qualitative analysis of native endothelial tissue ABO antigens after transplantation. Transplant Proc 1989; 21:783–785. 124. Hasan R, Van den Bogaerde J, Forty J, Wright L, Wallwork J, White DJG. Xenograft adaptation is dependent on the presence of antispecies antibody, not prolonged residence in the recipient. Transplant Proc 1992; 24:531–532. 125. Soares, M.P., Lin, Y., Anrather, J., et al. Expression of heme oxygenase-1 can determine cardiac xenograft survival. Nature Med 1998; 4:1073–1077. 126. Suhr BD, Guzman-Paz M, Apasova EP, Matas AJ, Dalmasso AP. Induction of accommodation in the hamster-to-rat model requires inhibition of the membrane attack complex of complement. Transplant Proc 2000;32:976. 127. Dalmasso AP, He T, Benson BA. Human IgM xenoreactive natural antibodies can induce resistance of porcine endothelial cells to complement-mediated injury. Xenotransplantation 1996; 3:54–62. 128. Dalmasso AP, Benson BA, Johnson JS, Lancto C, Abrahamsen MS. Resistance against the membrane attack complex of complement induced in porcine endothelial cells with a gal_(1–3)gal binding lectin: up-regulation of CD59 expression. J Immunol 2000; 164:3764–3773. 129. Dorling A, Stocker C, Tsao T, Haskard DO, Lechler RI. In vitro accommodation of immortalized porcine endothelial cells: resistance to complement mediated lysis and down-regulation of VCAM expression induced by low concentrations of polyclonal human IgG antipig antibodies. Transplantation 1996; 62:1127–1136. 130. Dorling A, Delikouras A, Nohadani M, Polak J, Lechler RI. In vitro accommodation of porcine endothelial cells by low dose human anti-pig antibody: reduced binding of human lymphocytes by accommodated cells associated with increased nitric oxide production. Xenotransplantation 1998; 5:84 - 92. 131. Azimzadeh A, Wolf P, Dalmasso AP, et al. Assessment of hyperacute rejection in a rat-to-primate cardiac xenograft model. Transplantation 1996; 61:1305–1313. 132. Johnston PS, Lim SML, Wang MW, Wright L, White DJG. Hyperacute rejection of xenografts in the complete absence of antibody. Transplant Proc 1991; 23:877–879. 133. Rajasinghe HA, Reddy VM, Hancock WW, Sayegh MH, Hanley FL. Key role of the alternate complement pathway in hyperacute rejection of rat hearts transplanted into fetal sheep. Transplantation 1996; 62:407–411.
vi
Defects and Amplification of Costimulation
9
173
Defects and Amplification of Costimulation Across the Species Nicola Rogers, PhD and Robert Lechler, PhD, FRCP INTRODUCTION
The initiation of a T-cell immune response requires the integration, by the T cell, of signals transduced by numerous cell surface molecular interactions. Central to these is the recognition of an antigenic major histocompatibility complex (MHC):peptide complex by the T cell’s antigen receptor (TCR). Ligation of the TCR leads to signal transduction through the CD3 complex, the signaling component of the receptor. Additional molecular interactions contribute to the elaboration and amplification of the TCR/CD3-transduced signal, which can be referred to as “signal 1.” These include the coreceptors CD8 for MHC class Irestricted T cells and CD4 for MHC class II-restricted cells. Another category of molecules, described as accessory molecules, contributes to the generation of signal 1. These include ICAM-1, LFA-3, and VCAM1, expressed by the antigen-presenting cell (APC), which interact with LFA-1, CD2, and VLA-4, respectively, expressed by the T cell. These interactions have two functions; first, they increase the avidity of T cell:APC conjugates, thereby increasing the chances of TCR occupancy by specific MHC molecule:peptide complexes. Second, the signals that they transduce amplify signal 1 (Fig. 1). However, it has been long appreciated that the activation of naive and resting memory T cells requires the receipt of a so-called “second,” or “costimulatory,” signal. The two-signal model of T-cell activation was first proposed by Bretscher and Cohn (1), and was substantiated by the work of Jenkins and Schwartz in the mid 1980s (2). The significance of costimulation lies, not only in the finding that signal 1 without signal 2 fails to activate From: Xenotransplantation: Basic Research and Clinical Applications Edited by: Jeffrey L. Platt © Humana Press Inc., Totowa, NJ
173
174
Rogers and Lechler
Fig. 1. Molecular ineractions between porcine stimulatory cells and human T-cells mediating Direct T-cell reactivity.
T cells (2–6), but in the many observations that a lack of costimulation can lead to abortive activation (5,7,8). Such abortive activation often induces T-cell nonresponsiveness either as a result of programmed cell death or the induction of T-cell anergy. The T-cell response to xenogeneic cells and tissues involves two distinct pathways, known as direct and indirect (9). The indirect pathway refers to xenoantigens being handled in exactly the same way as other protein antigens, in other words being taken up, processed, and presented by recipient APC in the form of peptides bound to recipient
Defects and Amplification of Costimulation
175
MHC molecules. This type of T-cell response does not involve any cross-species interactions in the generation of signal 1 and signal 2. The other pathway of T-cell xenorecognition, the direct pathway, results from CD4 and CD8 T cells interacting with the xenogeneic MHC molecules in intact form on the surface of xenogeneic APC. This will only culminate in recipient T-cell activation if the relevant interspecies molecular interactions occur with adequate efficiency to transduce the appropriate signals. Studies from our group, and from others, have demonstrated that both the indirect and the direct pathways of human antipig T-cell xenorecognition lead to vigorous T-cell responses in unprimed individuals (10–17). This chapter will be devoted to species compatibility of the molecular interactions involved in direct pathway responses. These cross-species contacts are not only important for T cell:APC interactions, but also play a key role in the recruitment of recipient leukocytes across xenogeneic endothelium (Fig. 2). This aspect of accessory and costimulatory molecule function will also be discussed.
The Role of Costimulation in T-cell Activation A large body of evidence now clearly supports the two-signal model of T-lymphocyte activation first proposed by Bretscher and Cohn (1). The in vitro induction of T-cell anergy in the absence of a second signal was first demonstrated by Jenkins and Schwartz in 1986 (2) using chemically fixed APC to present peptide antigen to CD4+ T-helper clones. A multitude of data has since been produced supporting the hypothesis that signal 1 in isolation fails to activate T cells and that costimulatory signaling results from contact with other cells rather than via soluble factors (2,3,7,9). Fibroblasts (18) or myoblasts (19) transfected with human class II MHC molecules, but not expressing costimulatory molecules (lacking signal 2) can efficiently display antigen MHC to class II-restricted T-cell clones, but fail to induce T-cell proliferation, rendering the responder cells anergic, i.e., unable to undergo autocrine proliferation upon subsequent stimulation. Alloreactive T-cell clones behave in a similar manner when exposed to costimulation-deficient fibroblast transfectants expressing allogeneic class II MHC (20). The context in which T cells first encounter antigen therefore has an important bearing on subsequent immune responsiveness. The nature of signal 2 has been well characterized. The crucial molecular interaction is between members of the B7 family of molecules on the APC and their primary ligand, CD28, on T cells because
176
Rogers and Lechler
Fig. 2. Leukocyte-porcine endothelial cell adhesion molecule interactions.
ligation of CD28 by either CD80 or CD86 is both necessary and sufficient to prevent the induction of anergy (21–30). Costimulatory signaling via B7/CD28, in concert with TCR/CD3-transduced signals, has been demonstrated to induce IL-2 production in both memory and naive T cells (31). The increase in IL-2 transcription and the stabilization of IL-2 mRNA following ligation of CD28 is due to the induction of the
Defects and Amplification of Costimulation
177
transcription factors AP-1 and NF-kappa B, which play a central role in IL-2 gene transcription (31–34). Studies using Chinese hamster ovary (CHO) cells transfected with cDNAs encoding human MHC II with or without CD80 or CD86 clearly illustrated the role of these molecules in amplifying IL-2 production (32). The induction of IL-2 following CD28 ligation results in the generation of long-lasting cell proliferation. These observations are supported by a growing body of in vitro data showing that although cells deficient in B7 fail to stimulate a primary mixed lymphocyte reaction (MLR), transfectants expressing high levels of B7 gained the capacity to stimulate the production of IL-2 by alloreactive T cells and to costimulate a polyclonal population of purified T cells cultured with immobilized anti-CD3 Mab (27). CD28 ligation not only enhances IL-2 production, which drives clonal expansion, but also augments the expression of the intrinsic survival gene, bcl-xl (35–37). Increased cell survival was found to correlate with upregulation of bcl-xl (37). The bcl-xl protein, however, is not induced following increased T-cell receptor cross-linking in the absence of CD28 ligation (37). Anti-costimulation molecule strategies aimed at either the receptors or their ligands are being used as therapeutic strategies for altering immune responses. Such approaches have been tested in model transplant systems to alter cell mediated responses thereby preventing graft rejection (38–41).
The Molecules Involved in Costimulation The major costimulatory molecules are CD80 and CD86 (B7.1 and B7.2; members of the B7 family) interacting with the T-cell counterreceptor CD28. The interaction between CD40 and CD154 (CD40 ligand) also contributes to the costimulation of T cells (42), although this may be due to the activating effects of CD40 ligation on the APC, leading to increased B7 expression, rather than to a direct effect on T cells. CD80 AND CD86 In vitro and in vivo data clearly demonstrate the crucial roles played by CD80 and CD86 in providing costimulation (21–28,30). In vivo, targeting the B7–CD28 interaction has been shown to prevent T-cell sensitization to graft antigens, thereby prolonging graft survival (41,43,44).
178
Rogers and Lechler
CD80 and CD86 belong to the immunoglobulin superfamily and are heavily glycosylated transmembrane proteins (29). CD80 was first identified as a B cell activation molecule in 1981 (45), followed by CD86 in 1993 (45). Both human CD80 and CD86, and the murine homologs, have now been cloned and functionally characterized (29). Both B7 isoforms are expressed on activated APC but exhibit distinctive patterns of expression and regulation (46). CD86 is constitutively expressed on APC and is rapidly up-regulated following activation. In contrast, CD80 is induced after activation of the APC and the kinetics of upregulation are much slower than that of CD86 (29,47). Porcine CD86, unlike its human and murine counterparts, is also constitutively expressed on endothelium (48), thereby increasing the immunogenicity of the transplanted tissue with respect to the generation of an anti-graft T-cell mediated response. CD80 and CD86 are highly homologous and are the natural ligands for the T-cell antigen CD28 (49). CD28 is constitutively expressed on resting T cells, but its expression is not static as levels increase following T-cell activation (49). Cytotoxic T-lymphocyte antigen-4 (CTLA-4), another cell surface glycoprotein and an Immunoglobulin superfamily member has been identified as a second receptor for the B7 family of molecules (50) and is homologous to CD28 with 31% sequence identity. Both B7 isoforms bind to CTLA-4 with 20-fold higher affinity than to CD28 (49,51). Unlike CD28, CTLA4 is not expressed on resting T cells but is induced following T-cell activation and expression is transient. In both human and murine systems, cell surface expression of CTLA-4 peaks 48 h after activation, returning to background levels by 96 h (51). Ligation of CD28 leads to the upregulation of CTLA4 mRNA (29,52). Although CD28-B7 receptor engagement results in an APC-derived costimulatory signal involved in antigen-specific IL-2 production (26,27), CTLA-4 appears to function as a negative regulator of T-cell activation (53–55). Cross-linking of CTLA4 receptors by anti-CTLA-4 antibodies has been demonstrated to antagonize CD28 ligation and, in addition, CTLA4 knock-out mice die due to uncontrolled lymphocyte proliferation within the first few weeks of life (56). Thus, CTLA4 ligation is thought to be crucial for the limitation and regulation of immune responses. The intracellular signaling events triggered by CTLA4 ligation have not yet, however, been clearly defined. CD40 CD40 is a 50kDa surface glycoprotein belonging to the tumor necrosis factor (TNF)-receptor superfamily. CD40 is expressed on all spe-
Defects and Amplification of Costimulation
179
cialized APC and many other cell types such as endothelial cells, thyroid epithelial cells, keratinocytes, and Langerhans cells (30). Its counterreceptor CD154 is a 33 kDa type II integral membrane protein (42) transiently expressed on activated but not resting T cells. The CD40–CD154 interaction has been demonstrated to play an important role in both the humoral and cellular arms of the immune response with a dominant role in B cell activation. Cross-linking of CD40 on B cells is essential for B-cell growth and isotype switching. Data from CD40 knock-out mice demonstrated that T cells primed in the absence of CD40 are unable to help B cells to form germinal centers or to class switch{\}immunoglobulin class switching to T-dependent antigens is defective{\} although responses to T-independent antigens are normal (57). In vitro studies have demonstrated that Mab specific for CD154 or soluble CD40Ig block the ability of activated T cells to activate B cells (42,58). Thus, stimulation of T cells via CD154 is important in their subsequent differentiation into cells that can help B cells. Recent studies provide evidence that CD40#-#CD154 interaction is also important in T-cell activation (59). Mice deficient in CD154 expression show defects in antigen-specific T-cell responses and reduced levels of IFN-a and IL-4 production. Peng et al. studied the effect of CD40-mediated costimulation for polyclonal T cell activation using P815 mouse mastocytoma cells transfected with CD40. CD40-mediated costimulation resulted in the induction of IL-2 production (60). Cross-linking of CD40 on B cells also results in the up-regulation of CD80 and CD86 expression (49). Levels of B7 expression (and thus APC capacity) are clearly up-regulated following CD40 signaling (30). Resting B cells do not normally express CD80/CD86 at high levels (49). Activation of B cells following simultaneous engagement of MHCpeptide/TCR and CD40–CD154 leads to the up-regulation of B7 family members on B cells, thereby enhancing the stimulation and subsequent activation of T cells (42,58,61). Thus, the CD40–CD154 interaction influences T-cell costimulatory activity by inducing expression of the B7 family of molecules and perhaps other accessory molecules. The clear synergistic effects of CD40 and CD80/CD86 indicate the importance of both costimulatory pathways for the initiation and amplification of T-cell dependent immune responses (43). CD40–CD154 interactions have also been shown to play a crucial role in the generation of cytotoxic T lymphocyte (CTL) responses by cross-priming (62–64). Modulation of CD40 on the surface of dendritic cells following engagement with antigen-specific T helper cells has
180
Rogers and Lechler
recently been demonstrated to condition the dendritic cell enabling it to subsequently stimulate CTL in the absence of T helper cells (63). Surface expression of the B7 isoforms on dendritic cells did not change with CD40 conditioning (63). Thus, CD40 modulation on the cell surface can determine the functional status of a dendritic cell in some undefined manner (62–64). Extensive studies have demonstrated the importance of blocking B7–CD28 and/or CD40–CD154 interactions in the context of both allotransplantation and xenotransplantation. Data strongly supporting this include the use of CTLA4–Ig to block signaling via CD28–B7 resulting in enhanced graft survival and the prevention of chronic rejection in a rat cardiac allograft model and a murine aortic allograft model (38–40). In these studies, administration of CTLA4–Ig caused partial or complete tolerance to graft antigen by inducing T-cell anergy. Treatment of allo pancreatic islet transplants with anti-CD80 (65) and antiCD86 antibody has also been demonstrated to inhibit transplant rejection (66). Similar results were obtained in models inhibiting CD40 signaling in a mouse cardiac allotransplant and a rhesus monkey renal allotransplant model (67–69). Two studies detailing the simultaneous blockade of signaling via CD28–B7 and CD40–CD154 in an allotransplant context prevented the onset of allograft rejection in both mice (43,69) and primates (70). In vitro and in vivo data have therefore clearly demonstrated that targeting the interactions mediated by the CD28–CD80/CD86 or CD40– CD154 interactions, separately or together, can prevent the sensitization of T cells to alloantigen and xenoantigen from engrafted tissue thereby prolonging graft survival.
The Molecules Involved in T Cell Adhesion Adhesion molecules are paramount to many cell–cell interactions that underlie and play a central role in the efficient functioning of the immune system. Adhesion molecules are critically involved in leukocyte traffic, recruitment into tissues, interaction with APC, and effector functions (71–73). There is a continuum, from adhesion receptors that serve to stabilize pairing between less avid receptors and ligands, to molecules involved largely in signal transduction and effector function. Following transplantation and reperfusion of a vascular graft, circulating leukocytes will first encounter the graft endothelium. To infiltrate
Defects and Amplification of Costimulation
181
the graft and to initiate the rejection response, leukocytes are required to traverse the endothelial barrier. As mentioned above selectins mediate the initial transient rolling steps, whereas stable adhesion and extravasation are determined by integrin mediated arrest and subsequent transmigration. The selectin family is comprised of three members (P-, E-, and L-selectin) (74,75). The three family members were identified according to the cell type on which they were first identified, E- (endothelium), P- (platelets), and L-(lymphocytes) selectin. Both E- and P-selectin are expressed on activated endothelium, mediating the initial transient rolling of leukocytes (75). Firm adhesion and tissue attachment are mediated by the increased avidity of leukocyte surface integrins from the exposure of rolling leukocytes to activating signals. Chemokines are the soluble agents that are thought to be the prime candidates triggering these activating signals (76). The dominant integrin-mediated interaction, which results in the arrest of selectin-mediated rolling of leukocytes and subsequent tissue attachment, has been demonstrated to be VCAM-1 on the endothelium with VLA-4 on the leukocytes, with a more minor role played by both ICAM-1 with LFA-1 and CD2 with LFA-3 (72,77–81). ICAM-1 and VCAM-1 are both members of the immunoglobulin superfamily and are variably glycosylated transmembrane proteins. Both the human and murine homologs have been cloned and functionally characterised (82–85). The expression of both ICAM-1 and VCAM-1 on a variety of cell types (lymphocytes, adherent macrophages and vascular endothelium) can be rapidly modulated in response to cytokine activation, e.g., IL-1, TNF-_, and IFN-a. Although constitutive expression of adhesion molecules on unactivated endothelium may not be sufficient to mediate leukocyte adhesion, activation of vascular endothelium is well established during organ ischemia resulting in the up-regulated expression of adhesion molecules. Studies in an allogeneic transplant context have demonstrated that inhibiting the interaction of these adhesion molecules can significantly prolong survival of the allograft (86–94), findings that can be extrapolated to the xenotransplantation context.
ACCESSORY AND COSTIMULATORY MOLECULE FUNCTION ACROSS SPECIES Evidence from Functional Data. Elucidating the role of T cells in xenoresponses has been facilitated by both in vitro experiments and in vivo studies of the transplantation of
182
Rogers and Lechler Table 1. Comparison of the CD86 Amino Acid Sequence Homology Across Species
Species comparison
Percentage amino acid homology
1
2
Human versus porcine Human versus murine3 Porcine versus murine
61.7% 40.7% 36.0%
Table 2. Comparison of the CD80 Amino Acid Sequence Homology Across Species Species comparison 4
Percentage Amino acid homology 5
Human versus murine 6 Human versus rat Rat versus murine
41.2% 40.1% 60.5%
Table 3. Comparison of the CD40 Amino Acid Sequence Homology Across Species Species comparison 7
Percentage Amino acid homology 8
Human versus murine Human versus bovine9 Bovine versus murine
61.2% 55.5% 43.2%
Table 4. Comparison of VCAM-1 Amino Acid Sequence Homology Across Species Species comparison 10
Percentage Amino acid homology 11
Human versus murine Human versus dog12 Dog versus murine
76.2% 81.5% 73.9%
Table 5. Comparison of ICAM-1 Amino Acid Sequence Homology Across Species Species comparison 13
Percentage amino acid homology 14
Human versus murine Human versus rat15 Human versus dog16 Murine versus rat Murine versus dog Dog versus rat
51.3% 45.8% 51.5% 76.4% 44.2% 46.4%
The percentage amino acid homology for the different species comparisons were calculated by aligning the translated cDNA sequences of the individual sequences. All of the sequences were obtained from the Genbank database. The accession numbers are
Defects and Amplification of Costimulation
183
nonvascularized grafts (skin and pancreatic islets) between species. This has enabled the definition of the role of T cells in the absence of both hyperacute rejection and the associated use of immunosuppressive drugs. Initial studies to investigate the nature of CD4+ T-cell-mediated xenoreactions used human T cells and murine stimulator cells as the in vitro model (95,96) revealing low level IL-2 production and reduced proliferative responses in comparison with allogeneic responses. Impaired interaction of the T-cell coreceptor CD4 with the MHC class II receptor and T-cell receptor bias were found to contribute to the weak human anti-murine xenoresponse. In addition, the responses predominantly involved the indirect pathway of recognition, although direct xenoresponses were also detected. In the context of CD8 restricted T-cell xenoresponses, the frequency of murine CTL precursors recognizing human HLA molecules was found to be substantially lower in comparison to the frequency of CTL precursors recognizing allogeneic murine H-2 antigen. That this diminished response was due to species specificity in the interaction of murine CD8 with human MHC class I and not due to the expression of other cell surface molecules was demonstrated using cell transfectants (96,97). Studies to investigate the provision of noncognate signals between the species demonstrated efficient costimulation of human T-cell proliferative responses by murine CD80 in both mitogenic and allogeneic systems (98). The binding of the murine adhesion molecule LFA-1 to human ICAM-1 is much reduced in comparison to intraspecies interactions. The complementary interaction between human LFA-1 and murine ICAM-1 is, however, significantly more efficient (99). A similar interaction of reduced efficiency involves murine LFA-3 and human CD2 (100,101). Sequence comparisons between costimulatory and adhesion molecules reveal varying degrees of homology between different species combinations (Tables 1–5). Thus, in context of certain species combinations, direct T-cell recognition appears to be weak due to a number of cross-species diffferences in the provision of both cognate and noncognate signals and the frequency of xenogeneic MHC-specific stimulator T cells. With the decision that porcine organs are now the most suitable for transplantation in a clinical context, the functional interactions in the human–pig species combination have been carefully investigated both in the context of T-cell activa1
2
3
4
5
6
7
as follows: U04343, L76099, L25606, M27533, X60958, AF010465, X60592, 8 MM83312, 9U57745, 10X53051, 11X67783, 12U32086, 13X06990, 14X52264, 15 16 D00913, L31625
184
Rogers and Lechler
tion and in leukocyte migration. Human anti-porcine T-cell responses have been extensively studied, revealing much stronger xenoresponses than had been observed in the human#-#mouse combination.
Human Anti-porcine Responses As mentioned above, stimulation of human T cells by porcine stimulators is very efficient. Studies within this laboratory, and others, have highlighted significant primary proliferative responses against porcine stimulator cells by human peripheral blood mononuclear cells (PBMC) (10–13,16,102–105). Evidence for the direct nature of this response is provided by the findings that depletion of human APC from the PBMC population did not affect the observed proliferation or IL-3 production and responses could be inhibited by the addition of anti-SLA class II monoclonal antibodies (102). Comparison between allogeneic and xenogeneic model systems in vitro have determined that precursor frequencies of human T cells with direct anti-porcine specificity are similar to the frequency of T cells with direct alloreactivity (106). Many studies have been performed to analyze the noncognate (costimulation and adhesion molecule) signals provided by porcine stimulator cells to human T cells, which underlie the dominant direct T-cell response, and between porcine endothelial cells and human leukocytes, which would support graft infiltration. In the context of T cell:APC interactions, it has been demonstrated that porcine stimulator cells can be as efficient as human APC at providing the noncognate interaction required for mitogenic T-cell proliferation and subsequent IL-2 production (107). The functional interaction of CD28 expressed on human T cells, and the counterreceptor CD86 expressed on porcine endothelial cells and bone-marrow-derived APC has also been formally demonstrated in vitro (48). Porcine CD86 has been cloned from aortic endothelial cells. Following transient transfection of porcine CD86, human umbilical vein endothelial cells strongly costimulated IL-2 production by human T cells. This costimulation of human T cells by porcine CD86 was shown to be as effective as costimulatory signals provided by human CD80 or CD86 and could be specifically blocked by the fusion protein CTLA4-Ig (48). The constitutive expression of CD86 by porcine endothelium is an additional factor contributing to the antigenicity of the porcine graft. This contrasts with allotransplantation, whereby once the specialized APC have been destroyed, allogeneic tissues are poor immunogens, owing to their lack of B7 expression. Expression of porcine CD80 and
Defects and Amplification of Costimulation
185
CD40, and their possible role in human T-cell stimulation have not, as yet, been reported. The functional interaction between porcine VCAM-1, ICAM-1, and CD2 with their ligands on human leukocytes have been demonstrated in their role mediating human leukcocyte adhesion to porcine endothelium (108,109). In addition it has recently been demonstrated that it is possible to detect functional interactions between porcine E- and L-selectin and the appropriate carbohydrate-bearing counterreceptors on human leukocytes (109). Antibody blockade of porcine E-selectin and the `2 integrins significantly reduced arrest and tissue attachment of human leukocytes to TNF-_ activated porcine endothelium (109). As mentioned above, functional interactions between the three major adhesion molecule pairs have been demonstrated in the human#-#pig species combination. Thus, although, as previously described, mouse LFA-1 does not interact with human ICAM-1 (99), human LFA-1 interacts efficiently with porcine ICAM. Anti-porcine ICAM-1antibody (122) and anti-human LFA-1 monoclonal antibodies (109) can efficiently inhibit adhesion of leukocytes to porcine endothelium. In similar studies, Mab specific for porcine VCAM-1 clearly blocked the human#-#porcine cellular interaction . Increased levels of adhesion were demonstrated when mitogen-activated, rather than resting, lymphocytes were incubated with porcine renal epithelial cells. The involvement of VCAM-1 and ICAM-1 in mediating adhesion was demonstrated by antibody blocking experiments (107). The functional interaction between human VLA-4 and porcine VCAM-1 has also been clearly demonstrated in human T-cell proliferation assays using immortalized porcine endothelial cells to efficiently costimulate the response of human T cells to a suboptimal dose of phytohemagglutinin (PHA). Addition of anti-porcine VCAM-1 MAb significantly inhibited the proliferative response (108,111). The CD2/LFA-3 pathway also significantly contributes to increased T-cell avidity to the APC. In vitro studies using porcine endothelium and human T cells have directly demonstrated interactions between human CD2 and porcine LFA-3 (108,111). Prior incubation of the human T cells with a blocking MAb directed against CD2 completely inhibited adhesion of human T cells to porcine aortic endothelium and to porcine erythrocytes which also express CD2. It has therefore been extensively demonstrated that porcine endothelial cells express the ligands for the human counterreceptors VLA-4, LFA-1, CD2, and CD28 and that cross-species interactions clearly con-
186
Rogers and Lechler
tribute to the activation of human T cells. Thus, although it is true that T-cell receptor#-#ligand interactions are generally more efficient within a species, intercellular interactions between the porcine-human species combination are adequate to initiate and maintain a T-cell response. The crucial role of these receptor#-#ligand interactions in T-cell activation and leukocyte extravasation, makes each molecule a potential target for tolerance inducing strategies.
STRATEGIES FOR INHIBITING COSTIMULATION ACROSS THE SPECIES Impetus for Novel Approaches to Immunosuppression in Xenotransplantation At present the major therapies to prevent cell-mediated rejection of organ transplants rely on systemic immunosuppressive drugs or Mab therapy directed against targets such as CD3, CD4, and CD25. There are several compelling reasons to think that novel approaches to the inhibition of cell-mediated responses should be pursued in the context of xenotransplantation. First, the direct human anti-pig T-cell response is surprisingly vigorous, and is quantitatively comparable to the direct alloresponse between HLA-mismatched pairs. Second, the direct human anti-pig T-cell response is unlikely to be self-limiting, as it appears to be in allotransplantation. As discussed above, once the bone-marrowderived specialized APC (passenger cells) are destroyed, allogeneic tissues are poor immunogens due to their lack of B7 expression. Indeed, the prolonged residence of an allograft tends to induce hyporesponsiveness in recipient T cells with direct anti-donor allospecificity (Mason, Lechler, and Hornick, unpublished observations). In contrast, it is probable that porcine vascularized tissues will retain their immunogenicity, owing to the constitutive expression of B7 molecules by vascular endothelium (11,112). As a consequence, the graft itself will continue to stimulate direct pathway T cells indefinitely. The third reason for devising novel strategies for immunosuppressing the T-cell xenoresponse is that xenotransplantation offers opportunities for graft-specific immunosuppression that are impossible in allotransplantation. Despite all the problems that species differences create, these very differences can be exploited. It should be possible to design reagents that will bind to donor (porcine) and not to recipient (human) antigens. This should avoid the undesirable complications of nonspecific immunosuppression.
Defects and Amplification of Costimulation
187
There are other opportunities offered by the use of animal organs for transplantation; access to both donor and recipient prior to transplantation will allow recipient pretreatment protocols involving donor cells. The transplant can be programmed once the desired effect has been achieved in terms of tolerance induction. The other obvious advantage offered by xenotransplantation is that the donor can be genetically manipulated, in order to render its tissues less susceptible to human immune attack. Thus far, genetic manipulation of the pig has been orientated toward alleviating hyperacute rejection (HAR). Now that it appears that HAR can be inhibited by complement regulation and/or antigen disguise, it is timely to contemplate other genetic manipulations that may inhibit the cellular response to porcine xenografts. Some possibilities are discussed below.
Injection of Reagents Designed to Inhibit Costimulatory or Accessory Molecule Function The molecules that play key roles in costimulating T cells, and in the adhesion of T cells to both APC and endothelial cells, are well defined, as discussed above. Considerable effort is currently being devoted to the inhibition of these interactions in allotransplantation. Much of this work is based on demonstrations of efficacy in experimental transplant models. For example, administration of a fusion protein, CTLA4-Ig, in combination with anti-CD40L Mab led to permanent cardiac allograft survival in mice (43). This combined approach has also been shown to be effective in a primate allotransplant model (113). Mabs specific for integrins including LFA-1 and VCAM have also led to prolonged mouse allograft survival (87,94). The most desirable outcome of any form of immunotherapy in transplantation is donor-specific T-cell tolerance. Whether or not costimulatory blockade, as achieved by coadministration of CTLA4-Ig and anti-CD40L, leads to tolerance is questionable. Although the mouse cardiac allografts enjoyed prolonged survival, the animals did not appear to be tolerant, as judged by their ability to reject a skin graft from the same donor strain. Indeed, rejection of the challenge skin graft led to slow rejection of the original cardiac allograft (43). These data raise some interesting questions, in that they suggest that complete inhibition of costimulation may lead to a state that is best described as immunological “ignorance,” whereby the graft is ignored, but the recipient is not actually tolerant. This may be a rather fragile state, in that any perturbation of the recipient’s response to the graft may precipitate a rejection response. These in vivo data are consistent with in vitro data that indi-
188
Rogers and Lechler
cate that the induction of tolerance requires partial activation. It is not yet clear whether costimulatory blockade can be tailored or timed such that abortive activation by the donor tissues is achieved, and tolerance is the end result. Application of these approaches in xenotransplantation is one attractive strategy. As emphasized above, the refinement that can be achieved when species-mismatched donors are used is that the immunotherapeutic reagent can be made donor-specific. This will avoid compromising the recipient’s immune system. Some experimental data have been described that confirm the potential utility of this strategy. Prolonged survival of human pancreatic islet xenografts was achieved in mice by administration of human CTLA4-Ig (41). It remains to be determined whether human and porcine B7 family molecules are sufficiently different to allow the generation of either fusion proteins or Mabs that will bind with clear preference for the pig molecules. As previously discussed porcine CD86 has been cloned and sequenced; although considerable homology exists with the human homolog, and porcine CD86 does costimulate human T cells, there is significant sequence divergence to think that it may be possible to generate pig-specific reagents. As outlined in Section 2.1, other porcine molecules that are involved in leukocyte:endothelial cell interactions and in T-cell:APC interactions do bind productively to their ligands expressed by human cells. These include VCAM, ICAM-1, and LFA-3. Reagents that bind specifically to these porcine molecules would also have potential in inhibiting the human anti-pig cellular response. As for the inhibition of costimulation, Mabs and fusion proteins may prove effective forms of therapy. There is also considerable interest in designing peptide inhibitors of integrin interactions for the treatment of human inflammatory disease. Again, it may be possible to generate peptides with specificity for porcine integrins; if so, smaller doses would be required in that the target tissue would be significantly smaller than when using peptides that react with recipient endothelium, for example.
Induction of Endogenous Antibodies Specific for Donor Costimulatory Molecules The administration of exogenous fusion proteins or antibodies has two disadvantages. First, these reagents currently need to be administered intravenously. This is a major inconvenience if the therapy has to be maintained beyond the first weeks after transplantation. Second,
Defects and Amplification of Costimulation
189
these forms of therapy commonly induce an immune response against the administered protein, which is of murine origin, in the case of Mabs, and may be of porcine origin in the case of a fusion protein such as CTLA4-Ig. This is a familiar problem with Mab therapy, and limits the half-life of the antibody until, eventually, the Mab becomes ineffective. The practice of humanizing monoclonal antibodies has reduced this problem; however, even humanized antibodies often induce an antiidiotypic response. With both of these limitations in mind, an alternative approach is under investigation, namely, to induce an endogenous antibody response of the desired specificity in the recipient, before xenotransplantation. The xenograft is predicted then to maintain the antibody response for as long as the offending antigen is displayed. The precedent experiment was conducted in mice that were immunized with peptides corresponding to part of the sequence of autologous gonadotrophin releasing hormone (GnRH). The peptides induced an autoantibody response, that was then maintained by the endogenous GnRH. As a consequence the immunized animals became sterile (114–116). This approach has been shown to inhibit the delivery of costimulation by porcine CD86 in a direct porcine islet to mouse transplant model, demonstrating significant prolongation of the islets.
Genetic Modification of the Donor Another of the advantages of xenotransplantation, as alluded to above, is the possibility of genetically modifying the donor. Although effort in this area has currently been devoted to strategies to inhibit HAR, it is timely to consider genetic means of inhibiting the cellular response to porcine xenografts. One set of approaches is based on attempts to inhibit the expression of key molecules, such as costimulatory molecules, or adhesion molecules such as ICAM-1, VCAM, and LFA-3. There are several candidate strategies that can be envisaged, which include the use of antisense constructs, ribozymes, and dominant negative transcription factors. Another generic approach may be best described as “intracellular knock out.” Conventional gene knock out by homologous recombination in embryonic stem cell lines is currently impossible in pigs, owing to the lack of such cell lines. Intracellular knock out refers to the intracellular expression of a genetically engineered antibody with specificity for the target molecule. Binding of the target molecule inside the cell by the
190
Rogers and Lechler
“intrabody” leads to intracellular retention and degradation. The form of antibody that is most appropriate to this approach is an sFv, which is a genetically engineered contiguous Fab fragment, so that the heavy and light chain variable segments are physically linked. Libraries of these sFv have been synthesized incorporating enormous diversity using randomly generated oligonucleotides encoding the hypervariable CDR3 region. When expressed as a coat protein on the surface of bacteriophage, sFv of the required specificty can be selected from the library provided that purified antigen is available. This approach has been successfully employed by two groups to inhibit the activity of the galactosyl transferase that generates the Gal_1-3Gal epitope recognized by the majority of xenoreactive natural antibodies (Sepp and Lechler, unpublished observations). Clearly, this strategy could be adapted for the purpose of inhibiting the expression of costimulatory and adhesion molecules. The other type of genetic modification that can be considered is one that leads to the expression of a molecule by the donor tissue that directly inhibits the function of recipient leukocytes. A classical example of this approach is the expression of Fas ligand. Two so-called “immunologically privileged” sites are the testis and the anterior chamber of the eye. Antigens introduced into these sites do not lead to a proinflammatory immune response. One of the mechanisms responsible for this is the constitutive expression of Fas ligand (117–120). Once this had been noted, several groups have genetically modified tissues to induce Fas ligand expression before transplantation. Cotransplantation of myoblasts expressing Fas ligand with allogeneic pancreatic islets led to prolonged islet graft survival (121). Others have failed to achieve the same success; nonetheless, this is an avenue that may be worth exploring in the context of xenotransplantation. Given the potent negative signaling role of CTLA4, its role in shutting down T-cell responses, and its possible role in the induction of T-cell anergy, strategies designed to preferentially signal recipient T cells through CTLA4, rather than CD28, may also hold promise.
CONCLUSIONS Recent advances in our understanding of the mechanisms, and, more specifically, the molecular interactions involved in transplant rejection, have enabled us to design graft-specific therapeutic strategies. The multitude of data arising from in vitro and in vivo allotransplantation
Defects and Amplification of Costimulation
191
systems have clearly provided a precedent for tolerance induction by the inhibition of these interactions. Thus, despite the significant T-cell response generated by human T cells against porcine stimulatory cells, owing to the functional interaction between costimulatory and adhesion molecules across the species divide, enhanced by the expression of porcine B7-2 on porcine endothelium, prospects for tolerizing the human anti-porcine T-cell response are encouraging.
REFERENCES 1. Bretscher P, Cohn M. A theory of self-nonself discrimination. Science 1970; 169 (950):1042. 2. Jenkins MK, Schwartz RH. Antigen presentation by chemically modified splenocytes induces antigen-specific T cell unresponsiveness in vitro and in vivo. J Exp Med 1987; 165 (2):302. 3. Mueller DL, Jenkins MK, Schwartz RH. An accessory cell-derived costimulatory signal acts independently of protein kinase C activation to allow T cell proliferation and prevent the induction of unresponsiveness. J Immunol 1989; 142 (8):2617. 4. Mueller DL, Jenkins MK, Schwartz RH. Clonal expansion versus functional clonal inactivation: a costimulatory signalling pathway determines the outcome of T cell antigen receptor occupancy. Annu Rev Immunol 1989; 7:445. 5. Bretscher P. The two-signal model of lymphocyte activation twenty-one years later [see comments]. Immunol Today 1992; 13 (2):74. 6. Schwartz RH. A cell culture model for T lymphocyte clonal anergy. Science 1990; 248:1349. 7. Baird MA. Evidence that heat-treated antigen-presenting cells induce hyporesponsiveness in allogeneic T cells. Transplantation 1994; 57 (5):763. 8. Jenkins MK, Pardoll DM, Mizuguchi J, Chused TM, Schwartz RH. Molecular events in the induction of a nonresponsive state in interleukin 2-producing helper T-lymphocyte clones. Proc Natl Acad Sci U S A 1987; 84 (15):5409. 9. Dorling A, Lechler RI. The passenger leucocyte, dendritic call and antigen-presenting cells (APC). Philadelphia: Lippincott-Raven, 1996. 10. Kirk AD, Hall BL, Finn OJ, Bollinger RR. In vitro analysis of the human antiporcine T-cell repertoire. Transplant Proc 1992; 24 (2):602. 11. Dorling A, Lombardi G, Binns R, Lechler RI. Detection of primary direct and indirect human anti-porcine T cell responses using a porcine dendritic cell population. Eur J Immunol 1996; 26 (6):1378. 12. Yamada K, Sachs DH, DerSimonian H. Human anti-porcine xenogeneic T cell response. Evidence for allelic specificity of mixed leukocyte reaction and for both direct and indirect pathways of recognition. J Immunol 1995; 155 (11):5249. 13. Kirk AD, Li RA, Kinch MS, Abernethy KA, Doyle C, Bollinger RR. The human antiporcine cellular repertoire. In vitro studies of acquired and innate cellular responsiveness. Transplantation 1993; 55 (4):924. 14. Alter BJ, Bach FH. Cellular basis of the proliferative response of human T cells to mouse xenoantigens. J Exp Med 1990; 171 (1):333. 15. Dorling A, Binns R, Lechler RI. Cellular Xenoresponses: Observation of significant primary indirect human T cell anti-pig xenoresponses using co-stimulator-deficient or SLA class II-negative porcine stimulators. Xenotransplantation 1996; 3:112.
192
Rogers and Lechler
16. Kumagai-Braesch M, Johansson Borg A, Satake M, Moller E. Characteristics of direct and indirect activation of human T cells against allogeneic and porcine xenogeneic cells/peptides. Xenotransplantation 1997; 4:85. 17. Satake M, Kawagishii N, Moller E. Direct activation of human responder T cells by porcine stimulator cells leads to T cell proliferation and cytotoxic T cell development. Xenotransplantation 1996; 3:198. 18. Gimmi CD, Freeman GJ, Gribben JG, Gray G, Nadler LM. Human T-cell clonal anergy is induced by antigen presentation in the absence of B7 costimulation. Proc Natl Acad Sci USA 1993; 90 (14):6586. 19. Warrens AN, Zhang JY, Sidhu S, et al. Myoblasts fail to stimulate T cells but induce tolerance. Int Immunol 1994; 6 (6):847. 20. Boussiotis VA, Freeman GJ, Gray G, Gribben J, Nadler LM. B7 but not intercellular adhesion molecule-1 costimulation prevents the induction of human alloantigenspecific tolerance. J Exp Med 1993; 178 (5):1753. 21. Boussiotis VA, Freeman GJ, Gribben JG, Daley J, Gray G, Nadler LM. Activated human B lymphocytes express three CTLA-4 counterreceptors that costimulate T-cell activation. Proc Natl Acad Sci U S A 1993; 90 (23):11,059. 22. June CH, Bluestone JA, Nadler LM, Thompson CB. The B7 and CD28 receptor families. Immunol Today 1994; 15 (7):321. 23. Freeman GJ, Freedman AS, Segil JM, Lee G, Whitman JF, Nadler LM. B7, a new member of the Ig superfamily with unique expression on activated and neoplastic B cells. J Immunol 1989; 143 (8):2714. 24. Freeman GJ, Gribben JG, Boussiotis VA, et al. Cloning of B7-2: a CTLA-4 counterreceptor that costimulates human T cell proliferation [see comments]. Science 1993; 262 (5135):909. 25. Azuma M, Ito D, Yagita H, et al. B70 antigen is a second ligand for CTLA-4 and CD28. Nature 1993; 366 (6450):76. 26. Linsley PS, Clark EA, Ledbetter JA. T-cell antigen CD28 mediates adhesion with B cells by interacting with activation antigen B7/BB-1. Proc Natl Acad Sci USA 1990; 87 (13):5031. 27. Norton SD, Zuckerman L, Urdahl KB, Shefner R, Miller J, Jenkins MK. The CD28 ligand, B7, enhances IL-2 production by providing a costimulatory signal to T cells. J Immunol 1992; 149 (5):1556. 28. Galvin F, Freeman GJ, Razi Wolf Z, et al. Murine B7 antigen provides a sufficient costimulatory signal for antigen-specific and MHC-restricted T cell activation. J Immunol 1992; 149 (12):3802. 29. Schultze J, Nadler LM, Gribben JG. B7-mediated costimulation and the immune response. Blood Rev 1996; 10 (2):111. 30. van Gool SW, Vandenberghe P, de Boer M, Ceuppens JL. CD80, CD86 and CD40 provide accessory signals in a multiple-step T-cell activation model. Immunological Reviews 1996; 153:47. 31. Wingren AG, Parra E, Varga M, et al. T cell activation pathways: B7, LFA-3, and ICAM-1 shape unique T cell profiles. Crit Rev Immunol 1995; 15 (3–4):235. 32. Olsson C, Michaelsson E, Parra E, Pettersson U, Lando PA, Dohlsten M. Biased dependency of CD80 versus CD86 in the induction of transcription factors regulating the human IL-2 promoter. Int Immunol 1998; 10 (4):499. 33. Parra E, Wingren AG, Hedlund G, et al. Costimulation of human CD4+ T lymphocytes with B7 and lymphocyte function-associated antigen-3 results in distinct cell activation profiles. J Immunol 1994; 153 (6):2479.
Defects and Amplification of Costimulation
193
34. Parra E, Wingren AG, Hedlund G, Kalland T, Dohlsten M. The role of B7-1 and LFA-3 in costimulation of CD8+ T cells. J Immunol 1997; 158 (2):637. 35. Collette Y, Benziane A, Razanajaona D, Olive D. Distinct regulation of T-cell death by CD28 depending on both its aggregation and T-cell receptor triggering: a role for Fas-FasL. Blood 1998; 92 (4):1350. 36. Boise LH, Noel PJ, Thompson CB. CD28 and apoptosis. Curr Opin Immunol 1995; 7 (5):620. 37. Sperling AI, Auger JA, Ehst BD, Rulifson IC, Thompson CB, Bluestone JA. CD28/ B7 interactions deliver a unique signal to naive T cells that regulates cell survival but not early proliferation. J Immunol 1996; 157 (9):3909. 38. Baliga P, Chavin KD, Qin L, et al. CTLA4Ig prolongs allograft survival while suppressing cell-mediated immunity. Transplantation 1994; 58 (10):1082. 39. Turka LA, Linsley PS, Lin H, et al. T-cell activation by the CD28 ligand B7 is required for cardiac allograft rejection in vivo. Proc Natl Acad Sci USA 1992; 89 (22):11,102. 40. Lin H, Bolling SF, Linsley PS, et al. Long-term acceptance of major histocompatibility complex mismatched cardiac allografts induced by CTLA4Ig plus donorspecific transfusion. J Exp Med 1993; 178 (5):1801. 41. Lenschow DJ, Zeng Y, Thistlethwaite JR, et al. Long-term survival of xenogeneic pancreatic islet grafts induced by CTLA4lg [see comments]. Science 1992; 257 (5071):789. 42. Grewal IS, Flavell RA. The role of CD40 ligand in costimulation and T-cell activation. Immunol Rev 1996; 153:85. 43. Larsen CP, Elwood ET, Alexander DZ, et al. Long-term acceptance of skin and cardiac allografts after blocking CD40 and CD28 pathways. Nature 1996; 381 (6581):434. 44. Wecker H, Auchincloss H, Jr. Cellular mechanisms of rejection. Curr Opin Immunol 1992; 4 (5):561. 45. Freeman GJ, Borriello F, Hodes RJ, et al. Murine B7-2, an alternative CTLA4 counter-receptor that costimulates T cell proliferation and interleukin 2 production. J Exp Med 1993; 178 (6):2185. 46. Linsley PS, Greene JL, Brady W, Bajorath J, Ledbetter JA, Peach R. Human B7-1 (CD80) and B7-2 (CD86) bind with similar avidities but distinct kinetics to CD28 and CTLA-4 receptors [published erratum appears in Immunity 1995 Feb;2(2):following 203]. Immunity 1994; 1 (9):793. 47. Ellis JH, Burden MN, Vinogradov DV, Linge C, Crowe JS. Interactions of CD80 and CD86 with CD28 and CTLA4. The Journal of Immunology 1996; 56:2700. 48. Maher SE, Karmann K, Min W, Hughes CC, Pober JS, Bothwell AL. Porcine endothelial CD86 is a major costimulator of xenogeneic human T cells: cloning, sequencing, and functional expression in human endothelial cells. J Immunol 1996; 157 (9):3838. 49. Jenkins MK, Johnson JG. Molecules involved in T-cell costimulation. Curr Opin Immunol 1993; 5 (3):361. 50. Brunet JF, Denizot F, Luciani MF, et al. A new member of the immunoglobulin superfamily—CTLA-4. Nature 1987; 328 (6127):267. 51. Lenschow DJ, Walunas TL, Bluestone JA. CD28/B7 system of T cell costimulation. Annu Rev Immunol 1996; 14:233. 52. Lin H, Rathmell JC, Gray GS, Thompson CB, Leiden JM, Alegre M-L. Cytotoxic T lymphocyte antigen (CTLA4) blockade accelerates the acute rejection of cardiac
194
53. 54. 55. 56.
57. 58. 59. 60.
61. 62.
63.
64.
65.
66.
67.
68.
69.
70.
Rogers and Lechler allografts in Cd28-deficient mice: CTLA4 can function independently of CD28. Journal of Experimental Medicine 1998; 188 (1):199. Krummel MF, Allison JP. CD28 and CTLA-4 have opposing effects on the response of T cells to stimulation [see comments]. J Exp Med 1995; 182 (2): 459. Krummel MF, Allison JP. CTLA-4 engagement inhibits IL-2 accumulation and cell cycle progression upon activation of resting T cells. J Exp Med 1996; 183 (6):2533. Walunas TL, Lenschow DJ, Bakker CY, et al. CTLA-4 can function as a negative regulator of T cell activation. Immunity 1994; 1 (5):405. Waterhouse P, Penninger JM, Timms E, et al. Lymphoproliferative disorders with early lethality in mice deficient in Ctla-4 [see comments]. Science 1995; 270 (5238):985. van Essen D, Kikutani H, Gray D. CD40 ligand-transduced co-stimulation of T cells in the development of helper function. Nature 1995; 378 (6557):620. Grewal IS, Flavell RA. A central role of CD40 ligand in the regulation of CD4+ T-cell responses. Immunol Today 1996; 17 (9):410. Tang A, Judge TA, Turka LA. Blockade of CD40-CD40 ligand pathway induces tolerance in murine contact hypersensitivity. Eur J Immunol 1997; 27 (12):3143. Peng X, Kasran A, Warmerdam PA, de Boer M, Ceuppens JL. Accessory signaling by CD40 for T cell activation: induction of Th1 and Th2 cytokines and synergy with interleukin-12 for interferon-gamma production. Eur J Immunol 1996; 26 (7):1621. Grewal IS, Flavell RA. CD40 and CD154 in cell-mediated immunity. Annu Rev Immunol 1998; 16:111. Schoenberger SP, Toes RE, van der Voort EI, Offringa R, Melief CJ. T-cell help for cytotoxic T lymphocytes is mediated by CD40-CD40L interactions [see comments]. Nature 1998; 393 (6684):480. Ridge JP, Di Rosa F, Matzinger P. A conditioned dendritic cell can be a temporal bridge between a CD4+ T- helper and a T-killer cell [see comments]. Nature 1998; 393 (6684):474. Bennett SR, Carbone FR, Karamalis F, Flavell RA, Miller JF, Heath WR. Help for cytotoxic-T-cell responses is mediated by CD40 signalling [see comments]. Nature 1998; 393 (6684):478. Tran HM, Nickerson PW, Patel A, Strom TB, Allen RDM, O’Connell PJ. Shortterm suppression of the xeno-immune response with mCTLA4-Fc treatment. Xenotransplantation 1997; 4:222. Lenschow DJ, Zeng Y, Hathcock KS, et al. Inhibition of transplant rejection following treatment with anti-B7-2 and anti-B7-1 antibodies. Transplantation 1995; 60 (10):1171. Lu L, Li W, Fu F, et al. Blockade of the CD40-CD40 ligand pathway potentiates the capacity of donor-derived dendritic cell progenitors to induce long-term cardiac allograft survival. Transplantation 1997; 64 (12):1808. Larsen CP, Alexander DZ, Hollenbaugh D, et al. CD40-gp39 interactions play a critical role during allograft rejection. Suppression of allograft rejection by blockade of the CD40-gp39 pathway. Transplantation 1996; 61 (1):4. Sun H, Subbotin V, Chen C, et al. Prevention of chronic rejection in mouse aortic allografts by combined treatment with CTLA4-Ig and anti-CD40 ligand monoclonal antibody. Transplantation 1997; 64 (12):1838. Kirk AD, Harlan DM, Armstrong NN, et al. CTLA4-Ig and anti-CD40 ligand prevent renal allograft rejection in primates. Proc Natl Acad Sci U S A 1997; 94 (16):8789.
Defects and Amplification of Costimulation
195
71. Osborn L. Leukocyte adhesion to endothelium in inflammation. Cell 1990; 62 (1):3. 72. Springer TA. Adhesion receptors of the immune system. Nature 1990; 346 (6283):425. 73. Hynes RO. Integrins: versatility, modulation, and signaling in cell adhesion. Cell 1992; 69 (1):11. 74. Bevilacqua MP, Nelson RM. Selectins. Journal of Clinical Investigation 1993; 91:379. 75. Tedder TF, Steeber DA, Chen A, Engel P. The selectins: vascular adhesion molecules. Faseb J 1995; 9 (10):866. 76. Tanaka Y, Adams DH, Shaw S. Proteoglycans on endothelial cells present adhesion-inducing cytokines to leukocytes. Immunol Today 1993; 14 (3):111. 77. Springer TA. Traffic signals for lymphocyte recirculation and leukocyte emigration: the multistep paradigm. Cell 1994; 76 (2):301. 78. Butcher EC. Leukocyte-endothelial cell recognition: three (or more) steps to specificity and diversity. Cell 1991; 67 (6):1033. 79. Alon R, Kassner PD, Carr MW, Finger EB, Hemler ME, Springer TA. The integrin VLA4 supports tethering and rolling in flow on VCAM-1. J Cell Biol 1995; 128 (6):1243. 80. Carlos TM, Harlan JM. Leukocyte-endothelial adhesion molecules. Blood 1994; 84 (7):2068. 81. Jones DA, McIntire LV, Smith CW, Picker LJ. A two-step adhesion cascade for T cell/endothelial cell interactions under flow conditions. J Clin Invest 1994; 94 (6):2443. 82. Tessitore A, Pastore L, Rispoli A, et al. Two gamma-interferon sites activation sites on the promotor of the human ICAM-1 gene are required for induction of transcription of Interferon gamma. 1995. 83. Ballantyne CM. Nucleotide sequence of the cDNA for murine intercellular adhesion molecule-1 (ICAM-1). Nucleic Acids Research 1989; 17:5853. 84. Polte TR. Full length vascular cell adhesion molecule. Nucleic Acid Research 1990; 18:5901. 85. Araki M, Araki K, Vassalli P. Cloning and sequencing of mouse VCAM-1 cDNA. Gene 1993; 126:261. 86. van Dijken PJ, Ghayur T, Mauch P, Down J, Burakoff SJ, Ferrara JL. Evidence that anti-LFA-1 in vivo improves engraftment and survival after allogeneic bone marrow transplantation. Transplantation 1990; 49 (5):882. 87. Isobe M, Yagita H, Okumura K, Ihara A. Specific acceptance of cardiac allograft after treatment with antibodies to ICAM-1 and LFA-1. Science 1992; 255 (5048):1125. 88. Berlin PJ, Bacher JD, Sharrow SO, Gonzalez C, Gress RE. Monoclonal antibodies against human T cell adhesion molecules— modulation of immune function in nonhuman primates. Transplantation 1992; 53 (4):840. 89. Fischer A, Griscelli C, Blanche S, et al. Prevention of graft failure by an anti-HLFA1 monoclonal antibody in HLA-mismatched bone-marrow transplantation. Lancet 1986; 2 (8515):1058. 90. Le Mauff B, Hourmant M, Rougier JP, et al. Effect of anti-LFA1 (CD11a) monoclonal antibodies in acute rejection in human kidney transplantation. Transplantation 1991; 52 (2):291. 91. Kameoka H, Ishibashi M, Tamatani T, et al. The immunosuppressive action of antiCD18 monoclonal antibody in rat heterotopic heart allotransplantation. Transplantation 1993; 55 (3):665.
196
Rogers and Lechler
92. Byrne JG, Smith WJ, Murphy MP, Couper GS, Appleyard RF, Cohn LH. Complete prevention of myocardial stunning, contracture, low-reflow, and edema after heart transplantation by blocking neutrophil adhesion molecules during reperfusion. J Thorac Cardiovasc Surg 1992; 104 (6):1589. 93. Cosimi AB, Conti D, Delmonico FL, et al. In vivo effects of monoclonal antibody to ICAM-1 (CD54) in nonhuman primates with renal allografts. J Immunol 1990; 144 (12):4604. 94. Orosz CG, Ohye RG, Pelletier RP, et al. Treatment with anti-vascular cell adhesion molecule 1 monoclonal antibody induces long-term murine cardiac allograft acceptance. Transplantation 1993; 56 (2):453. 95. Batten P, Heaton T, Fuller Espie S, Lechler RI. Human anti-mouse xenorecognition. Provision of noncognate interactions reveals plasticity of T cell repertoire. J Immunol 1995; 155 (3):1057. 96. Engelhard VH, Le AX, Holterman MJ. Species-specific structural differences in the alpha 1 + alpha 2 domains determine the frequency of murine cytotoxic T cell precursors stimulated by human and murine class I molecules. J Immunol 1988; 141 (6):1835. 97. Newberg MH, Ridge JP, Vining DR, Salter RD, Engelhard VH. Species specificity in the interaction of CD8 with the alpha 3 domain of MHC class I molecules. J Immunol 1992; 149 (1):136. 98. Hargreaves R, Logiou V, Lechler R. The primary alloresponse of human CD4+ T cells is dependent on B7 (CD80), augmented by CD58, but relatively uninfluenced by CD54 expression. Int Immunol 1995; 7 (9):1505. 99. Claiborne Johnston S, Dustin ML, Hibbs ML, Springer TA. On the species specificity of the interaction of LFA-1 with Intercellular adhesion molecules. The Journal of Immunology 1990; 145 (4):1181. 100. Moses RD, Winn HJ, Auchincloss H, Jr. Evidence that multiple defects in cellsurface molecule interactions across species differences are responsible for diminished xenogeneic T cell responses. Transplantation 1992; 53 (1):203. 101. Moses RD, Pierson RNd, Winn HJ, Auchincloss H, Jr. Xenogeneic proliferation and lymphokine production are dependent on CD4+ helper T cells and self antigenpresenting cells in the mouse. J Exp Med 1990; 172 (2):567. 102. Rollins SA, Kennedy SP, Chodera AJ, Elliott EA, Zavoico GB, Matis LA. Evidence that activation of human T cells by porcine endothelium involves direct recognition of porcine SLA and costimulation by porcine ligands for LFA-1 and CD2. Transplantation 1994; 57 (12):1709. 103. Murray AG, Khodadoust MM, Pober JS, Bothwell AL. Porcine aortic endothelial cells activate human T cells: direct presentation of MHC antigens and costimulation by ligands for human CD2 and CD28. Immunity 1994; 1 (1):57. 104. Dorling A, Binns R, Lechler RI. Direct human T-cell anti-pig xenoresponses are vigorous but significantly weaker than direct alloresponses. Transplant Proc 1996; 28 (2):653. 105. Dorling A, Binns R, Lechler RI. Significant primary indirect human T-cell anti-pig xenoresponses observed using immature porcine dendritic cells and SLA-class IInegative endothelial cells. Transplant Proc 1996; 28 (2):654. 106. Dorling A, Lechler RI. The T cell response to xenografts: molecular interactions and graft-specific immunosuppression. Xenotransplantation 1996; 4:68.
Defects and Amplification of Costimulation
197
107. Pleass HC, Forsythe JL, Proud G, Taylor RM, Kirby JA. Xenotransplantation: an examination of the adhesive interactions between human lymphocytes and porcine renal epithelial cells. Transpl Immunol 1994; 2 (3):225. 108. Mueller JP, Evans MJ, Cofiell R, Rother RP, Matis LA, Elliott EA. Porcine vascular cell adhesion molecule (VCAM) mediates endothelial cell adhesion to human cells. Transplantation 1995; 60:1299. 109. Robinson LA, Tu L, Steeber DA, Preis O, Platt JL, Tedder TF. The role of adhesion molecules in human leukocyte attachment to porcine vascular endothelium: implications for xenotransplantation. J Immunol 1998; 161 (12):6931. 110. Dorling A, Stocker C, Tsao T, Haskard DO, Lechler RI. In vitro accommodation of immortalized porcine endothelial cells: resistance to complement mediated lysis and down-regulation of VCAM expression induced by low concentrations of polyclonal human IgG antipig antibodies. Transplantation 1996; 62 (8):1127. 111. Dorling A, Stocker C, Tsao T, Haskard DO, Lechler RI. In vitro accomodation of immortalised porcine endothelial cells: resistance to complement mediated lysis and down-regulation of VCAM expression induced by low concentrations of polyclonal human IgG antipig antibodies. Transplantation 1996; 62:1127. 112. Pober JS, Cotran RS. Immunologic interactions of T lymphocytes with vascular endothelium. Adv Immunol 1991; 50:261. 113. Larsen CP, Pearson TC. The CD40 pathway in allograft rejection, acceptance, and tolerance. Curr Opin Immunol 1997; 9 (5):641. 114. Sad S, Gupta HM, Talwar GP, Raghupathy R. Carrier-induced suppression of the antibody response to a ‘self’ hapten. Immunology 1991; 74 (2):223. 115. Sad S, Rao K, Arora R, Talwar GP, Raghupathy R. Bypass of carrier-induced epitope-specific suppression using a T-helper epitope. Immunology 1992; 76 (4):599. 116. Sad S, Chauhan VS, Arunan K, Raghupathy R. Synthetic gonadotrophin-releasing hormone (GnRH) vaccines incorporating GnRH and synthetic T-helper epitopes. Vaccine 1993; 11 (11):1145. 117. Yagita H, Hanabuchi S, Asano Y, Tamura T, Nariuchi H, Okumura K. Fas-mediated cytotoxicity—a new immunoregulatory and pathogenic function of Th1 CD4+ T cells. Immunol Rev 1995; 146:223. 118. Golstein P. Fas-based T cell-mediated cytotoxicity. Curr Top Microbiol Immunol 1995; 198:25. 119. Lynch DH, Ramsdell F, Alderson MR. Fas and FasL in the homeostatic regulation of immune responses [see comments]. Immunol Today 1995; 16 (12):569. 120. Nagata S, Golstein P. The Fas death factor. Science 1995; 267 (5203):1449. 121. Lau HT, Yu M, Fontana A, Stoeckert CJ. Prevention of islet aloograft rejection with engineered myoblasts expressing FasL in mice. Science 1996; 273:109. 122. Stocker CJ, Sugars KL, Yarwood H, et al., Transplantation 2000 Aug 27; 70(4):579–586. 123. Rogers, N. J., Mirenda, V., Jackson, I., Dorling, A., Lechler, R. I. Costimulatory blockade by the induction of an endogenous xenospecific antibody response. Nat Immunol. 2000 Aug; 1(2):163–8.
vi
Antibody-Dependent Effects
199
Effects on 10 Antibody-Dependent Cellular Immunity Antonello Pileggi, R. Damaris Molano, Thierry Berney, and Luca Inverardi, MD INTRODUCTION Transplantation of tissues or organs between different species is referred to as Xenotransplantation. Combinations of donor and recipient species in which a xenograft is rapidly lost due to hyperacute rejection (HAR) are operationally defined as “discordant,” whereas combinations in which HAR does not occur are called “concordant.” HAR is initiated by binding of xenogeneic natural antibodies (XNA) to the vascular endothelial cells (VEC) of the implanted discordant xenogeneic organ. The presence of XNA in the recipient therefore defines a discordant species combination. XNA play a pleiomorphic role on the rejection of discordant xenografts. First they represent the primum movens of HAR, where IgM binding to the implanted organ efficiently activates complement (C) via the classic pathway progressing to activation of coagulation, thrombosis, and massive hemorrhage. VEC are the primary target of XNA binding: swelling, vesiculation, alterations in cellular junctions, detachment, and lysis are observed on the graft endothelium during HAR (12,31,72,73). Antigen specificity of XNA IgM binding has been defined, in humans, and appears to be directed at a carbohydrate moiety composed of two Gal molecules in alpha linkage (_Gal) (25). Second, XNA of G class with similar antigen specificity may play a prominent role in inducing/amplifying events characteristic of an additional mechanism of xenograft rejection that has been called delayed xenograft rejection (DXR). DXR is observed when HAR is prevented by inhibiting complement activation, with administration of cobra From: Xenotransplantation: Basic Research and Clinical Applications Edited by: Jeffrey L. Platt © Humana Press Inc., Totowa, NJ
199
200
Pileggi et al.
venom factor (CVF) or soluble complement receptor-1 (sCR1) (12,74). It is characterized by cell infiltration of the graft, mostly by natural killer (NK) cells, granulocytes, monocytes/macrophages, and by VEC activation, fibrin deposition, and platelet aggregation (5). IgG XNA binding to VEC during DXR mediates cell recruitment mainly via an antibodydependent cell cytotoxicity (ADCC)-like phenomenon. ADCC requires the presence of Fc receptors on the surface of cells recruited at the site of IgG binding. One of the cell subsets potentially highly relevant to this process is the NK cells, which exert their cytotoxic action on target cells through binding to IgG via the FcaRIII (CD16). Additionally, IgG XNA binding to VEC can induce VEC activation, an important hallmark of DXR (64,65). XNA, present in the vast majority of human subjects, are dynamically modulated in their titers upon exposure to xenoantigens (3,9,26,27,77,78,83). IgM concentration increase is likely to be T-cellindependent, while switch from IgM to IgG involves T-cell help (T-dependent) (53). CD4+ T cells are required for the generation of helper-dependent XNA that contribute to rejection both by complement-dependent mechanisms and by ADCC. Controversy exists as to the role of preformed natural antibodies in the recognition and rejection of nonvascularized xenogeneic tissues. Although XNA binding can be demonstrated in vitro on tissues such as islets of Langerhans, its role in inducing C activation and/or ADCC in vivo is unclear, although it has been suggested that preformed natural antibodies may be involved in the recognition of cell transplants, when large animals or primates are used as recipients (33,40). * It is widely accepted that pig tissues might represent the most suitable source of xenografts for humans. Nevertheless, the vigorous immune response to porcine organs and tissues remains the major obstacle to the successful use of those xenografts in clinical settings. Preformed natural antibodies represent the primum movens of hyperacute rejection of xenografts between discordant species. There is evidence that 85–95% of the human XNA are directed against a xenoantigen abundantly expressed on pig cells, Gal_1–3Gal`1–4GlcNAc-R, referred to as _-galactosyl or _Gal epitope (7,21,22). This epitope is synthesized by the enzyme _-1,3-galactosyltransferase, which is present in all mammals except humans, apes, and Old World monkeys (24). As a result, humans do not express the _Gal epitope, and instead normal human
Antibody-Dependent Effects
201
serum contains high xenoantibodies titers against _Gal. Other less-defined xenoantigens might exist, but their nature and role in the pathogenesis of xenograft recognition and rejection are not fully characterized. Anti-_Gal is a natural polyclonal antibody present in humans, as IgG (20), IgM, and IgA isotypes (47,66,76), that represent about 1% of the total circulating immunoglobulins. Anti-_Gal antibodies develop soon after birth, possibly in response to exposure to gastrointestinal microorganisms that express _Gal molecules (8,23,30). Most individuals have 5–40 µg of IgM per mL of plasma directed against _Gal (66), whereas the amount of IgG specific for _Gal varies from 0 to 20 µg per mL of plasma (67). It has been proposed that XNA and anti-blood group A and B antibodies may be members of a common family of natural antibodies. This is based on the observed similarity between the functional properties and concentration in serum of xenoreactive antibodies specific for _Gal and that of antibodies against blood group A and B antigens (66–68). It has been suggested that the generation of the xenospecific humoral repertoire might occur via one or both of two described pathways of B lymphocyte stimulation and immunoglobulin production. The first is characterized by T-cell-dependent proliferation of B cells and maturation in specificity and avidity of antibody binding following somatic mutations of the Ig Variable region genes. The second pathway is characterized by production of polyreactive natural antibodies in response to repetitive, polyvalent antigen stimulation (such as that associated with bacteria and infectious agents), encoded by Variable genes in germline configuration (10). The importance of this T-independent humoral mechanism has been stressed after observation of the increasing XNA levels in patients that received immunosuppression exposed to xenogeneic bioartificial liver treatment (3). Substantial IgM and IgG anti-_Gal titer increases have been identified in human serum after extracorporeal circulation through porcine livers (3,9,27,84), and in nonhuman primate serum in similar conditions (2,52). High immunoglobulin levels have also been detected in xenografted organs after transplantation and have been associated with graft rejection (45,48). XNA production in rodents has been attributed to a B cell subset characterized by the surface expression of the CD5 antigen. SCID mice reconstituted with purified rat CD5+ B cells were able to produce XNA, of the IgM isotype, after inoculation of hamster irradiated splenocytes,
202
Pileggi et al.
in a T-independent fashion. In animals reconstituted with CD5- B cells no XNA production was detected (53). In humans a CD5- B lymphocyte subset has been described, that is also capable of producing antibodies against _Gal pig residues (67). Attempts to avoid HAR by removal of anti-_Gal antibodies from serum of primate recipients prior to transplantation (using extracorporeal immunoabsorption of plasma through immunoaffinity columns) have been successful in delaying rejection and temporarily depleting baboons of anti-_Gal antibodies. Nevertheless, once the treatment was discontinued, IgG and IgM rose significantly despite the use of immunosuppressive therapy (19,44,48). The evidence available demonstrates not only an important role of xenoantibodies as complement activators and inducers of HAR, but also as mediators of cell-dependent cytotoxicity and possibly exerting direct effect on target cells by activating or damaging graft endothelial cells or inducing procoagulative changes (54). Humoral immune response to xenogeneic antigens seems to be one of the key factors, not only during HAR but also during DXR. * IgM binding to xenogeneic VEC occurs very rapidly after reperfusion of vascularized organs. Deposition of this immunoglobulin was detected on the endothelium of porcine organs as early as 1 min after transplantation into primates (72). Analysis of biopsies revealed IgM along most VEC surfaces, whereas IgG was found in interstitial spaces and in fibrin clots (72). Subsequently, IgM recognition of xenoantigens results in activation of the complement cascade that induces damage of cell membranes and consequent cell lysis. Complement activation also leads to procoagulative changes on the endothelial surface, initiating the coagulation cascade, fibrin deposition, and thrombosis of the vascularized graft (39). Depletion of IgM from human serum prevented the lysis of target cells and the activation of endothelium in vitro and prolonged organ survival in an ex vivo pig to human xenotransplantation model (46,66). Transfer of serum containing high levels of IgM obtained from rats rejecting xenografts induced HAR of hamster hearts in naive rats. When IgM depleted serum was used, HAR did not occur (53). * IgG specific for _Gal have been found in variable amounts in human serum, and some competition with IgM for binding to the _Gal epitope may exist. IgG in humans can be further subdivided according to structural differences in four subclasses—IgG1, IgG2, IgG3, and IgG4.
Antibody-Dependent Effects
203
Human IgG XNA are composed of all four IgG subclasses, with higher levels of IgG1 and IgG2, and lower ones of IgG3 and IgG4. In general, it is believed that a certain subclass specificity for discrete antigen categories exists. IgG1, IgG3, and IgG4, for example are thought to preferentially recognize protein antigens, whereas IgG2 bind to carbohydrate antigens (3,34,75). IgG subclasses are also characterized by different functional properties such as their C activation efficacy, where IgG1 and IgG3 activate efficiently, whereas IgG2 do not (84,85). IgG4 are also deficient in activating the classic pathway (58). It has been observed that IgG1 and IgG3 levels are higher after exposure to xenoantigens in vivo, and suggested that complement activation might need higher levels of those subclasses to result in any effect comparable to those mediated by IgM (3). In vitro studies that utilized human serum and cultured porcine aortic endothelial cells showed that binding of IgM to the VEC and deposition of C1q decreased as concentrations of anti-_mGal IgG increased in the serum. Also, it was demonstrated that xenoreactive IgG (mostly IgG2) were unable to induce complement fixation (83). This indicates a possible modulating role of IgG on the C fixation initiated by xenogeneic IgM on cellular targets and is consistent with experiments in which administration of large amounts of purified IgG to non human primates prevented the hyperacute rejection of porcine cardiac xenografts (55). When human serum enriched in IgG was added to porcine VEC culture, a dose-dependent decrease in deposition of iC3b, cytotoxicity, and heparan sulfate release were observed. The decrease was not due to alteration in antibody binding or consumption of C, but presumably reflected diminished formation of C3 convertase on VEC (55). * It is believed that the main role of xenogeneic IgG, rather than complement activation, might be of providing a ligand for Fc receptors present on neutrophils, eosinophils, monocytes, and NK cells, after binding to _Gal epitopes on the xenogeneic target (ADCC). These effector cells release their granule contents, leading to cell lysis in the graft (29,38,42,53). An additional role attributed to IgG is their capability of inducing VEC activation. VEC activation has been observed after stimulation with sera obtained from monkeys that received porcine cartilage xenografts in vitro (64,65). Interestingly, sera from the same monkeys prior to xenotransplantation were unable to activate VEC suggesting
204
Pileggi et al.
that either a high titer was necessary or that acquired antigen specificity/ affinity were needed for efficient VEC activation. Increased expression of E-selectin as a marker of VEC activation was detected only when anti-_Gal antibodies were present in the serum. These data led the authors to propose that interactions between high affinity anti_Gal antibodies, induced in vivo by the _Gal epitope expressed on the xenograft, and the VEC could represent a prominent mechanism of vascular endothelium activation in DXR. These two mechanisms (the availability of Fc fragments of IgG bound to VEC and VEC activation) are likely to contribute, in a synergistic fashion, to the recruitment of cells such as macrophages, polymorphonuclear (PMN), and NK cells in the rejecting graft (29). * It is generally accepted that DXR represents the major immunological hurdle to the clinical application of xenotransplantation. DXR is revealed when HAR is prevented and is characterized histologically by ischemia, swelling of endothelium, deposition of fibrin thrombi, and inflammatory alterations. It may result from graft endothelial cell activation and subsequent infiltration of the organ with activated host monocytes and NK cells. Disruption of endothelial continuity and exposure of subendothelial matrix leads to activation of coagulation cascade and platelet aggregation with the consequent thrombosis. It has been speculated that the interaction of XNA with the graft might represent the initiating event of DXR, since removal of XNA by preabsorption and inhibition of antibody synthesis with cytotoxic agents has been shown to be capable of delaying or even preventing DXR (52), whereas the infusion of anti donor antibodies could induce it (16). Evidence of the important role of _Gal specific XNA IgG has been obtained through transplantation of porcine meniscus cartilage, which is avascular and therefore does not undergo HAR. Thirty to three-hundred-fold increases in recipients’ anti-_Gal activity, mostly IgG, were observed within 4 wk after transplantation into cynomolgus monkeys, with extensive macrophage and T-cell infiltration within the implants. This inflammatory response was markedly reduced when cartilage tissue was treated with _Galactosidase to remove _Gal epitopes before implantation in primates, and only few infiltrating mononuclear cells, mostly macrophages, were observed two months after implantation. Absence of the _Gal epitopes might not only lead to inefficient recruitment of macrophages, but also to reduced opsonization and processing
Antibody-Dependent Effects
205
of xenoantigens by antigen-presenting cells and therefore T-cell activation (28,29). * Several ex vivo and in vitro observations have contributed to elucidate the mechanisms and subsets of cells implicated in the rejection of vascularized discordant xenografts. An ex vivo model in which rat hearts were perfused with human peripheral blood leukocytes (PBL) permitted the recognition of the interaction between circulating NK cells and the VEC (37,38). The study was designed to investigate the mechanisms of cellular immune recognition of discordant vascularized organs in the early stages, in an experimental system that mimics the in vivo situation. Human PBL were circulated in the rat hearts in a buffer containing heatinactivated human serum. In some of the experiments the human serum was pretreated to eliminate IgG; in other experiments IgG-depleted serum was reconstituted by the addition of purified human IgG. At the end of the experimental time (60 min), 30 – 40% of the originally infused cells could not be recovered from the recirculating solution, when unmanipulated whole human serum was present. When IgG-depleted serum was utilized, the rate of human cell sequestration dropped to 20%. The phenotypic analysis of the recovered cells showed a significant reduction of the CD16+CD56+ cells, corresponding to NK cells subset, and use of purified NK cell preparations confirmed their preferential sequestration and the amplifying role of IgG. Histological analysis of the organs confirmed that IgG and NK cells were binding to the inner surface of the vascular VEC. The NK cells had infiltrated the parenchyma, and vascular permeability was grossly altered (37). Sequestration was at least partly dependent on the functional integrity of leukocyte integrins, since blocking of LFA-1 and Mac-1 with specific monoclonal antibodies largely prevented NK retention and impairment of heart performance. Pretreatment of NK cells with blocking antibodies specific for FcaRIII significantly decreased IgG-dependent adhesion of NK cells to the xenogeneic VEC. In vitro experiments confirmed NK cell adhesion to VEC and enhancement of adhesion by the IgG (41). When the cytotoxic potential of human NK cells was evaluated in vitro on cultured VEC monolayers of different species, human NK cells mediated efficient lysis, and the efficiency also increased in the presence of human IgG in the assay (41). Therefore these data provided in vitro evidence of ADCC occurring following binding of IgG to xenoepitopes expressed on VEC, and
206
Pileggi et al.
recruitment of NK cells via engagement of FcaRIII. It additionally showed that a novel mechanism of direct recognition of VEC by NK cells might concurrently occur (14,15,36). To investigate whether target recognition by XNA and NK cells overlaps, human IgG F(ab’)2 fragments, specific for _Gal epitope, were used to mask the epitope before addition of NK cells in cocultures of porcine VEC. IgG binding to VEC was inhibited by F(ab’)2 that were also capable of interfering with the direct NK cells recognition and lysis of target cells. This observation confirmed that both XNA and NK cells recognize common target structures on xenogeneic VEC, suggesting that a comparable evolutionary pathway might exist in the selection of the repertoires of XNA and NK cells (41,42,80,83). It has also been reported that NK cells can activate VEC by direct recognition of endothelial antigens in vitro. This leads to the up-regulation of E-selectin and expression of the chemotactic cytokine IL-8. The addition of IgG to the assay enhanced E-selectin expression and cellular cytotoxicity, and induced tumor necrosis factor (TNF)-_ and interferon (IFN)-a secretion (19). NK cells have been reported to express E-selectin ligand (70), suggesting that activated VEC could enhance NK cell-mediated cytotoxicity. * Macrophages also participate in the response to xenografts in the DXR phenomenon and may cause graft injury. They have antigen-presenting capacity, can perform phagocytosis, and express Fc-receptors for IgG, being capable of mediating ADCC. Therefore, their role in DXR might comprise i) direct damage to VEC or other xenogeneic targets, ii) antigen presentation to T lymphocytes, and iii) ADCC. Antigen presentation could be rendered more efficient by the presence of IgG, leading to opsonization of relevant antigens. IgG also are required for the ADCC phenomenon. Evidence of this mechanism has been observed in cardiac xenografts (guinea pig into rat), where the presence of both B cells and macrophages accelerated rejection, whereas the tempo to rejection was significantly prolonged when one of the two cellular subsets was depleted (17). It could be speculated, therefore, that XNA synthesized posttransplant by B cell might bind to xenograft antigens and promote adherence and migration of activated macrophages via FcaR-dependent interactions. Alternatively, the activation of VEC following XNA binding might lead to expression of surface integrins (such as E- and P-selectin) and
Antibody-Dependent Effects
207
cytokines production (such as TNF-_, IFN-a), promoting cellular adhesion and infiltration of the xenograft, amplifying the response to xenogeneic tissues. * Furthermore, PMN are also involved in the rejection of vascularized xenografts, and some evidence of the role of XNA in mediating their recruitment and adhesion to VEC has been obtained in vitro. It has been observed that human sera containing IgG XNA were able to mediate increased adhesion of PMN to porcine VEC in vitro, and also to initiate damage to VEC via ADCC mechanisms along with peripheral blood mononuclear cells bearing FcaRIII; when the mononuclear cells were treated with anti-FcaRIII monoclonal antibody (3G8), lysis of VEC was inhibited substantially (6). In dynamic conditions in vitro adhesion and transmigration of human mononuclear cells on porcine VEC was augmented in the presence of human serum (60). This phenomenon has been attributed to two different mechanisms: the first depending directly on complement (C3) deposited on VEC in the early phase (30–90 min) of cell activation induced by xenogeneic serum; the second, delayed (5 h), possibly regulated by transcription of nuclear factor-kappa B (NFgB)-dependent genes and up-regulation of integrins expression (VCAM-1 and ICAM1) on activated VEC (61). * Receptors for each Ig class have been detected on several species’ leukocytes. Only the low affinity receptor for IgG, referred to as FcaRIII (CD16), has been identified and characterized biochemically and genetically on most of NK cells, macrophages, and a minor T cells subset (11,69). The transmembrane anchored CD16 isoform is a 50–70 kDa glycoprotein of the Ig superfamily, noncovalently associated with the a subunit of the high-affinity IgE receptor (Fc¡RI-a) in mouse NK cells and with Fc¡RI-a or the c subunit of T cell antigen receptor (TcR) complex in human NK cells (49). ADCC mediated by NK cells is initiated by the engagement of lowaffinity FcaRIII by the Fc portion of the antibodies. One of the earliest detectable signaling events following Fc_RIII receptor ligation is increased activation of Src-family tyrosine kinases. Following FcaRIIImediated activation, NK cells secrete cytokines, mediate ADCC, and may undergo apoptosis as a consequence of Fas ligand (FasL)-induced
208
Pileggi et al.
cell death. Similarity between the signal transduction and the effector functions induced by engagement of FcaRIII on NK cells and those due to the TcR on T lymphocytes have been observed (51). The cytotoxic effector mechanisms that NK cells utilize include the release of perforin/granzymes (35,43,62). Human NK cells incubated with porcine kidney cell line (PK15) are efficiently inhibited in their lytic potential when perforin/granzyme was chemically blocked, whereas blockade of the Fas/FasL pathway had no measurable effect (19). However, the Fas/Fas ligand-dependent cell mediated cytotoxicity has also been reported to be active in the killing of target cells in several experimental settings (1,18,56,63). Direct adhesion of human NK cells to porcine endothelium induces activation of VEC, with increased expression of E-selectin and IL-8 by VEC, and this phenomenon is accentuated by the addition of IgG (31). The activation of the VEC is initiated by ligation of NK membrane-bound lymphotoxin with the TNFR1 receptor expressed on the VEC (82). * The role of XNA in rejection of nonvascularized xenografts (such as cells and tissues) is quite controversial. Although there is consensus on the lack of HAR in rejection of cell transplants such as islets of Langerhans, somehow conflicting data exist on the mechanisms that mediate graft destruction. Xenogeneic islets of Langerhans, for example, have been shown to be bound by human and primate natural antibodies in vitro, but the specificity of antigen recognition has not been elucidated. _Gal epitopes appear to be expressed mainly, if not exclusively, on intra-islet VEC, but binding of human serum can be also detected on _Gal negative endocrine cells (57,59, and our unpublished data). Intracytoplasmatic staining of _Gal, on the other hand, has been reported in one recent publication in endocrine ` cells (78). In vivo analysis of Ig binding to implanted xenogeneic islets of Langerhans in primates has also yielded data supporting a role for XNA, and data denying it. Work from the group of Morris suggested, both in vitro and in vivo, an important role of Ig binding and consequent PMN opsonization and chemotaxis in the rejection of rabbit islets in cynomolgus monkeys (33), while no detectable Ig binding or complement deposition have been observed by the group of Korsgren, when fetal pig islet cell clusters (ICC) were implanted in cynomolgus (79). Recent data obtained in the same laboratory, in a model of neural cell implants, has demonstrated prolonged survival in Ig-deficient recipient rodents (50). Interestingly, the same recipients rejected neonatal ICC as efficiently as the control rodents, as
Antibody-Dependent Effects
209
well as the FcaR-deficient ones (4), suggesting a certain tissue specificity governing the mechanisms involved in xenogeneic cell recognition. It must be noted that substantial differences in the models utilized might explain the observed controversial data. In brief, our knowledge on the cellular and humoral mechanisms participating in xenograft rejection of tissues and cells needs to be expanded to obtain a working hypothesis that could be extrapolated to multiple experimental models. * In conclusion, antibody-dependent cellular events of discordant xenograft recognition appear to play a measurable role in DXR, where IgG binding to epitopes expressed mainly on VECs promotes recruitment of mononuclear and PMN cells via engagement of the membranebound FcaR. These events contribute significantly to VEC activation and/or destruction. Monocytes, NK cells, and PMNs are capable of transmigration and parenchymal infiltration, exerting direct effects on the graft, but they also contribute to the amplification of a concurrent or subsequent T cell response. At variance, the role of a cooperative interaction of Ig and cells in nonvascularized xenograft rejection still needs to be confirmed or excluded. It is reasonable to assume that successful xenotransplantation will require specific targeting of these mechanisms.
REFERENCES 1. Arase H, Arase N, Saito T. Fas-mediated cytotoxicity by freshly isolated natural killer cells. J Exp Med 1995; 181:1235–1238. 2. Azimzadeh A, Meyer C, Watier H, et al. Removal of primate xenoreactive natural antibodies by extracorporeal perfusion of pig kidneys and livers. Transpl Immunol 1998; 6:13–22. 3. Baquerizo A, Mhoyan A, Kearns-Jonker M, et al. Characterization of human xenoreactive antibodies in liver failure patients exposed to pig hepatocytes after bioartificial liver treatment. An ex vivo model of pig to human transplantation. Transplantation 1999; 67:5–18. 4. Benda B, Karlsson-Parra A, Ridderstad A, Korsgren O. Xenograft rejection of porcine islet-like cell clusters in immunoglobulin- or Fc-receptor a-deficient mice. Transplantation 1996; 62:1207–1211. 5. Blakely ML, Van der Werf WJ, Berndt MC, Dalmasso AP, Bach FH, Hancock WW. Activation of intragraft endothelial and mononuclear cells during discordant xenograft rejection. Transplantation 1994; 58:1059–1066. 6. Cooke SP, Pearson JD, Savage CO. Human IgG xenoreactive antibodies mediate damage to porcine endothelial cells in vitro by both humoral and cellular mechanisms. Transpl Immunol 1997; 5:39–44. 7. Cooper DKC, Koren E, Oriol R. Oligosaccharides and discordant xenotransplantation. Immunol Rev 1994; 141:31–58. 8. Cooper DKC, Thall AD. Xenoantigens and xenoantibodies: their modification. World J Surg 1997; 21:901–906.
210
Pileggi et al.
9. Cotterell AH, Collins BH, Parker W, Harland RC, Platt JL. The humoral immune response in humans following cross-perfusion of porcine organs. Transplantation 1995; 60:861–868. 10. Cramer DV, Wu GD. The humoral response to xenografts. Xeno 1996; 4:30. 11. Daeron M. Fc receptor biology. Ann Rev Immunol 1997; 15:203–234. 12. Dalmasso AP, Vercellotti GM, Platt JL, Bach FH. Inhibition of complement-mediated endothelial cell cytotoxicity by decay accelerating factor potential prevention of xenograft hyperacute rejection. Transplantation 1991; 52:530–533. 13. Dalmasso AP. The complement system in xenotransplantation. Immunopharmacology. 1992; 24:149–160. 14. Daniels BF, Nakamura MC, Rosen S, Yokoama WM, Seaman WE. Ly-49A, a receptor for H-2Dd, has a functional carbohydrate recognition domain. Immunity 1994; 1:785–792. 15. Drickamer C, Taylor ME. Biology of animal lectins. Annu Rev Cell Biol 1993; 9:237–264. 16. Freyer JP, Leventhal JR, Dalmasso AP, Chen S, Simone PA, Jessurun J, Sun LH, Reinsmoen NL, Matas AJ. Cellular rejection in discordant xenografts when hyperacute rejection is prevented: analysis using adoptive and passive transfer. Transpl Immunol 1994; 2:87–93. 17. Fryer JP, Leventhal JR, Dalmasso AP, et al. Beyond hyperacute rejection. Accelerated rejection in discordant xenograft model by adoptive transfer of specific cell subsets. Transplantation 1995; 59:171–176. 18. Fujiwara I, Nakajima H, Matsuda T, Mizuta N, Yamagishi H, Oka T. Fas and Fas Ligand in discordant xenogeneic Antibody-Dependent Cell-Mediated Cytotoxicity. Transplant Proc 1998; 30:2488–2490. 19. Fujiwara I, Nakajima H, Matsuda T, Mizuta N, Oka T. The molecular mechanism of apoptosis induced by xenogeneic cytotoxicity. Xenotransplantation 1998; 5:50–56. 20. Galili U, Rachmilewitz EA, Peleg A, Flechner I. A unique natural human IgG antibody with anti-_galactosyl specificity. J Exp Med 1984; 160:1519–1531. 21. Galili U, Macher BA, Buehler L, Shohet SB. Human natural anti-_-galactosyl IgG. II. The specific recognition of _(1,3)-linked galactose residues. J Exp Med 1985; 162:573–582. 22. Galili U, Buehler L, Sohet SB, Macher BA. The human natural anti-Gal IgG. III. The subtlety of immune tolerance in man as demonstrated by cross-reactivity between natural anti-Gal and anti-B antibodies. J Exp Med 1987; 165:693–704. 23. Galili U, Mandrell RE, Hamadeh RM, Sochet SB, Griffiss JM. Interaction between human natural anti-alpha-galactosyl immunoglobulin G and bacteria of the human flora. Infect Immun 1988; 57:1730–1737. 24. Galili U, Swanson K. Gene sequences suggest inactivation of _1,3galactosyltransferase in catarrhines after the divergence of apes from monkeys. Proc Natl Acad Sci USA 1991; 88:7401–7404. 25. Galili U. Interaction of natural anti-Gal antibody with _-galactosyl epitopes: a major obstacle for xenotransplantation in humans. Immunol Today 1993; 14:480–482. 26. Galili U, Tibell A, Samuelsson B, Rydberg L, Groth CG. Increased anti-Gal activity in diabetic patients transplanted with fetal porcine islet cell cluster. Transplantation 1995; 59:1549–1556. 27. Galili U, Minonov OP, Michler RE, Stone KR. High-affinity anti-Gal immunoglobulin G in chronic rejection of xenografts. Xenotransplantation 1997; 4:127–131.
Antibody-Dependent Effects
211
28. Galili U, La Temple DC, Walgenbach AW, Stone KR. Porcine and bovine cartilage transplants in cynomolgus monkey: II. Changes in anti-Gal response during chronic rejection. Transplantation 1997; 63:646–651. 29. Galili U. Significance of Anti-Gal in Chronic Xenograft Rejection. Transpl Proc 1999; 31:940–941. 30. Good AH, Cooper DKC, Malcom AJ, et al. Identification of carbohydrate structure that bind human anti-porcine antibodies: implications for discordant xenografting in humans. Transpl Proc 1992; 24:559–562. 31. Goodman DJ, von Albertini M, Willson A, Millan MT, Bach FH. Direct activation of porcine endothelial cells by human natural killer cells. Transplantation 1996 15;61:763–771. 32. Goodman DJ, Millan MT, Ferran C, Bach FH. Mechanism of delayed xenograft rejection. In: Kemp E, Reemtsa K, White D, Platt JL eds. Xenotransplantation, Berlin Heidelberg: Springer-Verlag, 1997: 77–94. 33. Hamelmann W, Gray DWR, Cairns TDJ, et al. Immediate destruction of xenogeneic islets in a primate model. Transplantation 1994; 58:1109–1114. 34. Hammarstrom L, Smith CIE. IgG subclasses in bacterial infections. Monogr Allergy 1986; 19:122–133. 35. Henkart PA, Williams MS, Zacharchuk CM, Sarin A. Do CTL kill target cells by inducing apoptosis? Semin Immunol 1997; 9:135–144. 36. Houchins JP, Yabe T, McSherry C, Bach FH. DNA sequence analysis of NKG2, a family of related cDNA clones encoding type II integral membrane proteins on human natural killer cells. J Exp Med 1991; 173:1017–1020. 37. Inverardi L, Samaja M, Motterlini R, Mangili F, Bender JR, Pardi R. Early recognition of discordant xenogeneic organ by human circulating lymphocytes. J Immunol 1992; 149:1416–1423. 38. Inverardi L, Samaja M, Marelli F, Bemder JR, Pardi R. Cellular early immune recognition of xenogeneic vascular endothelium. Transplant Proc 1992; 24:459–461. 39. Inverardi L, Pardi R. Early events in cell-mediated recognition of vascularized xenografts: cooperative interactions between selected lymphocytes subsets and natural antibodies. Immunol Rev 1994; 141:71–93. 40. Inverardi L, Ricordi C. Transplantation of pancreas and islets of Langerhans. A review of progress. Immunol Today 1996; 17:7–9. 41. Inverardi L, Clissi B, Stolzer AL, Bender JR, Sandrin MS, Pardi R Human natural killer lymphocytes directly recognize evolutionarily conserved oligosaccharide ligands expressed by xenogeneic tissues. Transplantation 1997; 63(9):1318–1330. 42. Inverardi L, Clissi B, Pardi R. Xenorecognition by NK cells and natural antibodies: a common evolutionary pathway? Xeno 1997; 5:9–11. 43. Kagi D, Ledermann B, Burki K, et al. Cytotoxicity mediated by T cells and natural killer cells is greatly impaired in perforin-deficient mice. Nature 1994; 369:31–37. 44. Kozlowski T, Ierino FL, Lambrigts D, et al. Depletion of anti-Gal(alpha)1-3Gal antibody in baboon by specific alpha-Gal immunoaffinity columns. Xenotransplantation 1998; 5:122–131. 45. Kozlowsky T, Shimizu A, Lambrigts D, et al. Porcine kidney and heart transplantation in baboons undergoing a tolerance induction regimen and antibody absorption. Transplantation 1999; 67:18–30. 46. Kroshus TJ, Bolman RM 3rd, Dalmasso AP. Selective IgM depletion prolongs organ survival in an ex vivo model of pig to human xenotransplantation. Transplantation 1996; 62:5–12.
212
Pileggi et al.
47. Kujundzic M, Koren E, Neethlig FA, et al. Variability of anti _-Gal antibodies in human serum and their relation to serum cytotoxicity against pig cells. Xenotransplantation 1994; 1:58. 48. Lambrigts D, Van Caster P, Xu Y, et al. Pharmacologic immunosuppressive therapy and extracorporeal immunoadsorption in the suppression of anti-alphaGal antibody in the baboon. Xenotransplantation 1998; 5:274–283. 49. Lanier LL. NK cell receptors. Ann Rev Immunol 1998; 16:359–393. 50. Larsson LC, Czech KA, Winder H, Korsgren O. Discordant neural tissue xenografts survive longer in immunoglobulin deficient mice. Transplantation 1999; 68:1153–1160. 51. Leibson PJ. Signal transduction during natural killer cell activation: inside the mind of a killer. Immunity 1997; 6:655–661. 52. Leventhal JR, John R, Fryer JP, et al. Removal of baboon and human antiporcine IgG and IgM natural antibodies by immunoadsorption. Transplantation 1995; 59:294–300. 53. Lin Y, Vandeputte M, Waer M. Suppression of T-independent IgM xenoantibody formation by leflunomide during xenografting of hamster hearts in rats. Transplantation 1998; 65:332–339. 54. Lin SS, Weidner BC, Byrne GW, et al. The role of antibodies in acute vascular rejection of pig to baboon cardiac transplants. J Clin Invest 1998; 101:1745–1756. 55. Magee JC, Collins BH, Harland RC, et al. Immunoglobulin prevents complementmediated hyperacute rejection in Swine-to- Primate Xenotransplantation. J Clin Invest 1995; 96:2404–2412. 56. Manilay JO, Sykes M. Natural killer cells and their role in graft rejection. Curr Opin Immunol 1998; 10:532–538. 57. McKenzie IFC, Koulmanda M, Mandel TE, Sandrin MS. Cutting edge: Pig islets are susceptible to “anti-pig” but not Gala(1,3)Gal antibodies plus complement in Gal o/ o mice. J Immunol 1998; 161:5116–5119. 58. Miletic VD, Frank MM. Complement-immunoglobulin interactions. Curr Opin Immunol 1995; 7:41–47. 59. Mirenda V, Le Mauff B, Boeffard F, et al. Intact pancreatic islet function despite humoral xenorecognition in the pig-to-monkey combination. Transplantation 1998; 66:1485–1495. 60. Morigi M, Zoja C, Colleoni S, et al. Xenogeneic serum promotes leukocyte adhesion to porcine endothelium under flow conditions, possibly through the activation of the transcription factor NF-kappa B. Xenotransplantation 1998; 5:57–60. 61. Morigi M, Zoja C, Colleoni S, et al. Xenogeneic serum promotes leukocyte-endothelium interaction under flow through two temporally distinct pathways: role of complement and nuclear factor-kappa B. J Am Soc Nephrol 1999; 10:2197–2207. 62. Nakajima H, Henkart PA. Cytotoxic lymphocyte granzymes trigger a target cell internal disintegration pathway leading to cytolysis and DNA breakdown. J Immunol 1994; 152:1057–1063. 63. Oshimi Y, Oda S, Honda Y, Nagata S, Miyazaki S. Involvement of Fas ligand and Fas-mediated pathways in the cytotoxicity of human natural killer cells. J Immunol 1996; 157:2909–2015. 64. Palmetshofer A, Galili U, Dalmasso AP, Robson S, Bach FH. A-Galactosyl epitope mediated activation of porcine aortic endothelial cells. Type I activation. Transplantation 1998; 65:844–853.
Antibody-Dependent Effects
213
65. Palmetshofer A, Galili U, Dalmasso AP, Robson S, Bach FH. A-Galactosyl epitope mediated activation of porcine aortic endothelial cells. Type II activation. Transplantation 1998; 65:971–978. 66. Parker W, Bruno O, Holzknecht ZE, Platt JL. Characterization and affinity isolation of xenoreactive human natural antibodies. J Immunol 1994; 153:3791–3803. 67. Parker W, Lundberg-Swanson KL, Holzknecht ZE, et al. Isohemagglutinins and xenoreactive antibodies: members of a distinct family of natural antibodies. Hum Immunol 1996; 45:94–104. 68. Parker W, Yu PB, Holzknecht ZE, Lundberg-Swanson KL, Buckley RH, Platt JL. Specificity and function of “natural” antibodies in immunodeficient subjects: clues to B cell lineage and development. J Clin Immunol 1997; 17:311–321. 69. Perussia B. Fc receptors on Natural Killer cells. In: Kärre K, and Colonna M eds. Specificity, Function, and Development of NK Cells. Berlin-Heidelberg :Springer - Verlag, 1988: 63–88. 70. Pinola M, Renkonin R, Majuri ML, Tiisala S, Saksela E. Characterization of the E-selectin ligand on NK cells. J Immunol 1994; 152:3586–3594. 71. Pitre J, Moeller E, Satake M. Human xenoreactive natural antibodies against Gal(1-3) pig terminal residues are not produced by CD5+ B-lymphocytes. Transplant Proc 1996; 28:545. 72. Platt JL, Fichel RJ, Matas AJ, Sun LH, Bolman RM, Bach FH. Immunopathology of hyperacute xenograft rejection in a swine-to-primate model. Transplantation 1991; 52:214–220. 73. Platt JL, Vercellotti GM, Lindman BJ, Oegama TR, Bach FH, Dalmasso AP. Relase of heparan sulfate from endothelial cells. Implications for pathogenesis of hyperacute rejection. J Exp Med 1990; 171:1363–1368. 74. Pruitt SK, Baldwin WM 3d, Marsh HC Jr, Lin SS, Yeh CG, Bollinger RR. The effect of soluble complement receptor type 1 on hyperacute xenograft rejection. Transplantation 1991; 52:868–873. 75. Ravindran B, Satapathy AK, Das MK. Naturally-occurring anti-_-galactosyl antibodies in human plasmodium falciparum infections: a possible role for autoantibodies in malaria. Immunol Lett 1988; 19:137–141. 76. Sandrin M, Vaughan HA, Dabkowsky PL, McKenzie IFC. Anti pig IgM antibodies in human serum react predominantly with Gal _1-3Gal epitopes. Proc Natl Acad Sci USA 1993; 90:11,391–11,395. 77. Satake M, Kumagai-Braesch M, Korsgren O, Andersson A, Möller E. Characterization of humoral human anti-porcine xenoreactivity. Clin Transplantation 1993; 7:281–288. 78. Satake M, Kawagishi N, Rydberg L, et al. Limited specificity of xenoantibodies in diabetic patients transplanted with fetal porcine islet cell clusters. Main antibody reactivity against _-linked galactose-containing epitopes. Xenotransplantation 1994; 1:89–101. 79. Shaaphender AFM, Wolvekamp MCJ, Te Bulte MTJW, Bouwman E, Gooszen HG, Daha MR. Porcine islet cells of Langerhans are destroyed by human complement and not by antibody-dependent cell-mediated mechanisms. Transplantation 1996; 62:29–33. 80. Söderlund J, Wennberg L, Castanos-Velez E, et al. Fetal porcine islet-like cell cluster transplanted to cynomolgus monkeys. Transplantation 1999; 67:784–791.
214
Pileggi et al.
81. Trinchieri G. Biology of natural killer cells. Adv Immunol 1989; 47187–374. 82. von Albertini M, Ferran C, Brostjan C, Bach FH, Goodman DJ. Membrane-associated lymphotoxin on natural killer cells activates endothelial cells via an NF-kappaBdependent pathway. Transplantation 1998; 66:1211–1219. 83. Yokoyama WM. Recognition structures on natural killer cells. Curr Opin Immunol 1993; 5:67–73. 84. Yu PB, Parker W, Everett ML, Fox IJ, Platt JL. Immunochemical properties of antigal_1-3Gal antibodies after sensitization with xenogeneic tissues. J Clin Immunol 1999; 19:116–126. 85. Yu P, Holzknecht ZE, Bruno D, Parker W, Platt JH. Modulation of natural IgM binding and complement activation by natural IgG antibodies. A role for IgG antiGal_1-3Gal antibodies. J Immunol 1996; 157:5163–5168.
Disordered Coagulation
215
11 Disordered Regulation
of Coagulation and Platelet Activation in Xenotransplantation Simon C. Robson, FRCP (UK), PhD INTRODUCTION
Over the past decade, substantial increases in transplant organ and recipient survival have been accompanied by a significant increase in the quality of life for patients with end-stage organ failure. However, the increasing access to organ transplant lists, coupled with static or even falling organ donation rates, have resulted in a doubling of the waiting time for patients receiving a cadaveric kidney at many major centers in the United States (http://www.unos.org). In addition, many patients waiting for suitable heart or liver donors die because of the lack of effective life-support systems. Living donor transplantation has the potential to alleviate renal allograft shortages but comparable procedures have been performed for lung and liver in only a few specialized centers to date. The proposed use of a unlimited supply of animal organs in clinical practice, viz., xenotransplantation, could provide a bridge to a successful allograft, or, more optimistically, may even substitute for allografts and provide for long-term graft survival (1–3). Unfortunately, all prior clinical applications of xenotransplantation have been failures when measured against the routine and effective use of allografts. Future trials in clinical xenotransplantation may be feasible once the immunological mechanisms of xenograft loss have been better determined. Such advances should result in novel therapies that can be tested in relevant animal models and shown to be effective with respect to graft survival and function at acceptable toxicity levels (2,4–6). UnfortuFrom: Xenotransplantation: Basic Research and Clinical Applications Edited by: Jeffrey L. Platt © Humana Press Inc., Totowa, NJ
215
216
Robson
nately, this scenario may not be immediately at hand as other nonimmunological barriers appear to exist. Under certain circumstances, profound disturbances in coagulation, platelet activation, and vascular injury are associated with the transplantation of vascularized xenografts or xenogeneic cells into primates. Indeed, intravascular coagulation is a feature of all forms of xenograft rejection and may represent an intrinsic barrier to this procedure. In this chapter, we address patterns of xenograft rejection, describe the factors that mediate the vascular injury and inflammation seen in xenograft rejection, and discuss the evidence for molecular incompatibilities between discordant xenografts or xenogeneic cells and primate blood constituents. These factors will contribute to the disordered regulation of clotting and platelet activation seen during xenotransplantation.
MECHANISMS OF XENOGRAFT REJECTION Recent developments in the fields of immunology and vascular biology have greatly expanded our understanding of the mechanisms by which xenografts are rejected (7–10). It has become apparent that the rejection response directed at a discordant xenograft is likely comprised of many separate elements that appear to have different kinetics and result in various manifestations of xenograft rejection (3,5,7). Despite this consideration, endothelial cell (EC) activation processes, with the accompanying vascular prothrombotic and inflammatory changes, are important manifestations of experimental xenograft rejection, irrespective of levels of complement (C) activation (7,10).
Hyperacute Rejection The pathogenesis of hyperacute rejection (HAR) relates to the binding of xenoreactive natural antibodies (XNA) to the graft endothelium and consequent activation of the C cascade by the classical pathways (11–13). These events result in perturbation of the involved graft endothelium causing interstitial edema, and hemorrhage with associated vascular thrombosis resulting in ischemic necrosis of the graft. The rapidity of this process (minutes to hours) precludes any absolute requirements for transcriptional up-regulation and synthesis of proinflammatory factors by vascular cells; this form of EC stimulation has been termed typeI activation and is associated with vascular disruption, parenchymal hemorrhage, and thrombosis (7,8).
Disordered Coagulation
217
Activated C components and thrombin among other mediators induce EC stimulation and platelet sequestration manifest by cellular retraction, exposure of the subendothelial matrix, von Willebrand factor (vWF) translocation to the EC surface, and loss of the antithrombotic phenotype provided by heparan sulfate (HS), thrombomodulin (TM), and the vascular ATP diphosphohydrolase (ATPDase/CD39) (14). The basis of this discordance relates to fundamental molecular incompatibilities. These result in the expression of xenoantigens, consequent binding of XNA, and the associated dysregulation of C activation by membrane-associated regulators of C (RCA; an incompatibility resulting in unopposed generation of activated C components) (15). XNA are directed at galactose-_1,3-galactose (gal) residues of xenogeneic glycoproteins (16) and appear to be the major immediate mediators of HAR injury in the discordant swine to primate combination (17,18). Primate recipients may be treated prophylactically by C inhibition to ameliorate these initiating events (19); however, EC activation with graft sequestration of platelets, mononuclear phagocytes, and natural killer (NK) cells are still observed under these circumstances (20). It is possible that C inhibition may be sufficient to preclude the rapid graft loss but does not prevent all of the earlier pathogenetic events of HAR that may evolve further to delayed forms of rejection (10). Novel molecular biological techniques have allowed the production of potential donor animals (pigs) with human transgenes directed toward amelioration of the C activation (21) and antibody interactions (22,23) shown to be of immediate importance in immediate xenograft rejection. Transplantation of transgenic pig organs that express human complement regulatory proteins (e.g., human decay accelerating factor; DAF, CD55) and CD59 (protectin) into primates provide an elegant approach to overcoming HAR. Complement inhibition following the grafting of these transgenic organs appears to be very effective in blocking HAR and the immediate C-mediated activation of platelets and coagulation; this topic has been reviewed recently (12). Unfortunately, the duration of this beneficial effect is not yet clear, and the prolongation of experimental porcine xenograft survival in primate models, chiefly using baboons, can be still measured only in days to weeks (24,25). Rejection events are associated with the deposition of XNA and elicited xenoreactive antibodies, local generation of procoagulants, vascular thrombosis, and ultimate graft loss in a process termed either acute vascular rejection (AVR) (11,26) or delayed xenograft rejection (DXR) (7,8).
218
Robson
Acute Vascular Rejection/Delayed Xenograft Rejection (AVR/DXR) Small animal models have been used to characterize certain aspects of AVR/DXR, e.g., the guinea pig heart to rat xenotransplantation model in which HAR is prevented by C-depletion in the recipient. Because the reactivity of XNA is unchanged, the initial C-component C1q-binding is unopposed and consequent activation of C to C3 unimpaired; hence, these models are not that informative in dissecting out defined nonlytic components of HAR from elements that occur during AVR/DXR (7). The studies of the fate of grafts in the pig-to-primate models, in which HAR has been prevented, have been complicated by three issues: the use of cynomolgus monkey recipients in which HAR is of variable occurrence, the use of neonatal donors or recipients, and highly variable treatment regimens that may not be clinically acceptable. In addition, the extensive pre- and posttransplantation treatments with consequential sepsis and immunosuppression have made detailed interpretation of the results difficult (12). In general terms, the pathological entity of AVR/DXR develops following the management of HAR by C suppression or transient depletion of XNA and has a time frame measured in days to weeks. Hence, as alluded to above, activation events initiated during graft reperfusion or HAR could evolve more fully in the still-viable but injured xenograft (20). Hypothetically, there are three prominent pathogenetic events that could result in AVR/DXR; this topic has been also reviewed extensively (7,8): 1. Abnormal Thromboregulation at the EC Surface. This may be considered both intrinsic (pre-existing) and/or acquired. The consequences are platelet sequestration and fibrin deposition in the vasculature; 2. Infiltration by Mononuclear Cells, viz. host NK cells and macrophages, with activation and production of cytokines and: 3. Autocrine and Paracrine EC Activation. This process of EC activation in AVR/DXR has been termed Type II, as the mechanisms are protein synthesis dependent. This latter development compounds any thrombotic elements and heightens mononuclear cell infiltration. Further perturbation of the quiescent vascular antithrombotic surface is linked to the production of procoagulants and release of prominent natural anticoagulants. Other features of the activated endothelium are dependent upon the new expression of adhesion molecules such as E-selectin, vascular cell
Disordered Coagulation
219
adhesion molecule (VCAM-1), and intercellular adhesion molecule (ICAM-1) with secretion of chemoattractant chemokines IL-8 and monocyte chemoattractant protein (MCP-1) (20). There have been many debates over both the exact sequences of events that lead to AVR/DXR and the relative importance of xenoantibodies (5,7). Currently, the evidence suggests that both XNA and elicited xenoreactive antibodies are associated with (and bind to) the vasculature of xenografts undergoing AVR/DXR. It seems likely that processes of rejection would be influenced by the levels of xenoreactive antibody and their potential effects on vascular integrins and adhesion proteins; several experimental data support this hypothesis (26). As alluded to above, discordant xenograft rejection may be considered as a continuum of events that occur at different time frames; these slower events can only be manifest if earlier onset and major injurious events are curtailed. Decreased and low levels of C-mediated injury may result in delayed up-regulation of procoagulants such as tissue factor (TF) that are important in the manifestations of thrombosis seen during the pathogenesis of AVR/DXR (27,28).
SELECTED MEDIATORS OF XENOGRAFT REJECTION Complement The central role of C activation in HAR can be clearly demonstrated by the prevention of this process by pretreatment of recipients with cobra venom factor (CVF), which depletes C by continuous activation to exhaustion (9,29). Soluble complement receptor type-1 (sCR-1) inhibits C activation and has comparable effects (19,30). In pig-toprimate xenografts, the C cascade is activated primarily via the classical pathway, while in rodents this also occurs by the alternate pathway (31). Under normal circumstances, basal levels of C activation are modulated by fluid phase and cell surface modulatory RCA. These [viz. CD46 (membrane cofactor protein; MCP), CD55, and CD59] preclude “innocent bystander” injury during periods of C activation e.g., during sepsis. The high levels of activation of C by IgM xenoantibody binding to the xenogeneic vasculature (32), coupled to documented interspecies molecular incompatibilities severely limits the ability of normal levels of porcine RCA to regulate human C (2,9). Overexpression of RCA in both mice and pigs markedly inhibits human C activation (33) and prevents HAR of pig to primate xenografts (34,35).
220
Robson
Xenoantibodies Almost all human or primate anti-pig XNA are directed against a single epitope, gal, widely expressed on glycolipids and glycoproteins (16,18). The _
Coagulation Factors and Platelet Mediators Whether these innate protective systems are recruited in a synergistic context by inflammatory reactions or could contribute to xenograft injury in a primary manner remains unclear. Certainly, inflammation plays an important role in the coagulation balance, viz., the local generation of procoagulants with the associated loss of heparan, TM, tissue factor pathway inhibitor (TFPI) (36), and vascular ATP diphosphohydrolase/CD39 activity following EC activation responses in vivo (37,38). These changes could potentiate any putative intrinsic thrombophilia within the rejecting xenograft (7). Such developments could exacerbate vascular damage and potentiate the activation of platelets and coagulation pathways resulting in graft infarction (10) and also lead to systemic complications with disseminated intravascular coagulation (DIC) (39,40). Clear evidence for the importance of coagulation factors or thrombotic mediators in xenograft rejection can be inferred by the beneficial effects of their inhibition that have been noted in several experimental models; beneficial results are usually comparable to those seen with C inhibition (41–43). Inhibition of platelet aggregation by treatment of xenograft recipients with antagonists to the platelet fibrinogen receptor, GPIIbIIIa (44,45), by the use of P-selectin or PAF antagonists (46–48), or by administration of a soluble ATPDase (49) has generally been shown to prolong graft survival in several discordant xenotransplantation models. In a contrary manner, deletion of cd39 in mutant mice hastens xenograft rejection and promotes vascular injury thereby indicating an important modulatory influence on graft survival (50–52).
Disordered Coagulation
221
DISORDERED THROMBOREGULATION IN VASCULARIZED XENOGRAFTS There are several historical precedents for the development of coagulation abnormalities and thrombocytopenia in association with solid organ xenograft rejection (53,54). More recently, hemoperfusion of porcine renal explants by human volunteers has resulted in significant thrombocytopenia with rapid onset of vascular injury (55); similar events have been described with ex vivo porcine liver hemoperfusions (56). Published reports detailing pig-to-primate xenotransplantation have documented substantively improved graft survival over the past 5 yr. In many of these series, bleeding does not appear to present overtly and hence coagulation parameters have not been examined in detail (24,34,57,58). Studies by White and colleagues examining transplantation of transgenic pig organs suggest that early vascular thrombosis, graft nonfunction and acute vascular rejection may present difficulties in the pig-to-baboon experiments; in their alternative models studying cynomolgus monkeys, an added complication was present in that hyperacute rejection of nontransgenic kidneys did not occur (59–64).
Disseminated Intravascular Coagulation In collaboration with Sachs and colleagues, we have determined that activation of coagulation factors by the xenograft vasculature has the potential to generate serious systemic hemostatic abnormalities with localized xenograft vascular injury progressing to a form of consumptive coagulopathy and then DIC (39). Plasmapheresis, column adsorption, and the administration of soluble complement receptor type-1 inhibitor (sCR-1) for the first eight postoperative days, were employed for the removal of XNA and inactivation of complement in the recipient baboons tested in this study. Both high- and low-dose sCR-1 efficiently blocked total hemolytic C levels as determined by CH50 assays during the time of administration. However, activation was not fully inhibited systemically during this time, as there was clear evidence for C3a generation and conversion of C3b to iC3b and to C3d (39). Typically, a bleeding diathesis was clinically evident from d 5 to 12 following the transplantation (39). Profound thrombocytopenia (with hematological abnormalities in keeping with DIC) was observed prior to and in association with the bleeding problem. The observed prolongation of prothrombin and partial thromboplastin times was accompanied by evidence for TF-mediated coagulation pathways, high levels of
222
Robson
thrombin generation (prothrombin fragment F1+2 production and thrombin–antithrombin complex formation; Fig. 1A,B), fibrinogen depletion, and production of high levels of the fibrin degradation product d-dimer. It is important to note that these disturbances resolved rapidly following the excision of the rejected xenografts in surviving animals. Histopathological examination of the rejected xenografts confirmed vascular injury and fibrin and platelet deposition with localized C activation. In these studies, baboons underwent bone marrow engraftment and were subject to conditioning regimens several months prior to the xenotransplantation procedure. The experimental animals had a considerable time period to recover from any vascular injury associated with these interventions. It is, however, possible that these procedures may have facilitated the development of DIC in the experimental animals. Certainly, irradiation and allogeneic or even autologous bone marrow and stem cell transplantation may be associated with thrombotic microangiopathy. In addition, depletion of plasma anticoagulants, such as anti-thrombin and other significant plasma antiproteases, could follow column perfusions and repeated aphereses; these procedures are associated with the activation of c (C3a and Bb generation) and coagulation factors (surrogate markers of thrombin generation on initial plasmapheresis; unpublished data). Conditioning events alone were not sufficient alone to perturb the hemostatic mechanism. However, these could have predisposed experimental animals to the development of disordered coagulation and hemostasis, that was then triggered by the onset of xenograft rejection. These severe alterations in hemostasis suggested the presence of a consumptive-type coagulopathy probably linked to the process of AVR/ DXR. Activation of coagulation factors by the xenograft vasculature, modulated in this instance by low levels of C activation via the classical pathway, may have the potential to generate serious systemic hemostatic abnormalities and DIC in the setting of xenograft rejection. We have been able to control these manifestations to some extent by combinations of agents that target thrombin generation and inhibit platelet aggregation (unpublished). The data presented by D’Apice and colleagues at the Nagoya Xenotransplantation meeting have detailed the transplantation of triple transgenic (CD55/CD59 and _1,2fucosyltransferase) kidneys into adult baboons without additional immunosuppression and reinforce the
Disordered Coagulation
223
Fig. 1. Markers of thrombin generation in DIC associated with xenotransplantation. (A) Substantive increases in F1+2 followed the xenotransplantation procedure. Levels of F1+2 then decreased until around postoperative day 5 where dramatic secondary increases in levels were observed. These peaked prior to rejection in all three animals. Levels had fallen by the time of graft nephrectomy in B75-34 and B75-18 but were still elevated in B75-13 at the earlier time of surgery and prior to demise. [Three experimental animals: B75-34 (*), B75-18 (䊏), and B75-13 (䊊)(39)] (B) Thrombin-antithrombin complexes (TAT) were determined by ELISA in the plasma samples and show a comparable pattern to that seen with other surrogate markers of prothrombin activation. However, the F1+2 levels do seem to precede TAT increases in keeping with the generation of thrombin preceding inactivation by complex formation with AT. (B75-34 (*), B75-18 (䊏) and B75-13 (䊊))
224
Robson
hypothesis that xenografts are associated with hemostatic abnormalities and coagulation disturbances. Control (nontransgenic) pig kidneys were rejected hyperacutely within minutes. Although the triple transgenic kidneys functioned for up to 5 d, recipients developed thrombocytopenia within hours of revascularization of the grafts; this was followed by DIC and worsening thrombocytopenia. Treatment with low-molecularweight heparin prevented this delayed progression and attenuated other aspects of the coagulopathy. These studies have confirmed the occurrence of DIC previously shown by us (39), and confirmed that this phenomenon was a consequence of the rejection process and not an artifact induced by the extensive conditioning regimen used to induce immunological tolerance in these studies. Our impression is that coagulopathies do not develop in the conditioned primates then exposed to allografts in comparable experiments. Our data suggest that the intensity of xenograft rejection, the process of C activation, XNA deposition, and putative molecular barriers, individually or collectively, are critical for the development of DIC (39). DIC has not been reported by other groups using cardiac or renal CD55 or CD55/CD59 grafts. One apparent common discriminating feature is the use of cyclophosphamide in groups that have not observed coagulopathy; however, it is not clear whether this may affect the DIC process directly or via attenuation of xenoantibody-mediated responses. In addition, the effects and documented benefits of other high-dose immunosuppressants on the vasculature (65,66) remain largely unexplored in discordant xenograft rejection. Our group is currently examining these possibilities.
Humoral Mediators and the Induction of Thrombotic Diathesis The development of DIC and the observation of the Shwartzman phenomenon in humoral-mediated rejection of homografts (67,68) support the hypothesis that antibody-mediated vascular injury may be implicated in the perturbation of coagulation seen with xenograft injury. Platt and colleagues have demonstrated that xenoreactive antibodies play a significant role in the pathogenesis of AVR/DXR and infer that this form of rejection might be treated by immunosuppressive and other therapies aimed at the humoral immune response to porcine antigens (26). However, currently available techniques to address the issue of xenoantibodies will also invariably influence coagulation factor levels
Disordered Coagulation
225
(affinity column perfusions) and EC or platelet activation (immunosuppressants). Saadi et al. have established that humoral injury to the vasculature, with the development of a procoagulant phenotype, is also related to C activation in vitro (27,69). Our more recent observations with specific gal epitope-mediated stimulation of EC have reinforced the postulate that xenoreactive antibodies may directly induce activation responses in the total absence of c (70–72). In these studies, porcine EC were incubated in the presence of the gal binding, Bandeiraea simplicifolia lectin (BS-I), and underwent type I activation with cellular shape changes associated with the formation of intercellular gaps (71). Porcine EC exposure to BS-I was also associated with the tyrosine phosphorylation of a protein (approx 130 kDa), not observed following LPS, TNF, or XNA stimulation (72). This lectin-induced tyrosine phosphorylation was not affected by cytochalasin D (inhibitor of actin filament polymerization), by genistein (inhibitor of tyrosine kinases), or by staurosporine (inhibitor of tyrosine phosphorylation and protein kinase C). Interestingly, porcine EC activation could be observed upon binding of various other lectins to the glycosylated moieties of membrane proteins. These other carbohydrate epitopes against which XNA may exist in certain models might represent minor xenoantigens from porcine to primates or may comprise the major xenoepitopes in other discordant xenograft models. In addition, incubation of EC with BS-I and monoclonal anti-gal IgM induced p42/44 map kinase and activated the transcription factor NF-gB The tetravalent gal-binding lectin BS-I, the wholly gal-specific isolectin BS-IB4, and elicited primate anti-pig xenoreactive antibodies (decomplemented cynomolgus monkey anti-porcine serum) all induced E-selectin protein expression in porcine EC; BS-I strongly induced E-selectin, P-selectin, ICAM-1, and IL-8 mRNA in an NF-gB dependent manner (70). It is important to note that human and primate XNA lacked this activity when tested at relevant concentrations. This type II activation response was gal specific, as preincubation with synthetic gal carbohydrate or adsorption of lectin or serum to rabbit, but not human, red blood cells removed the activating component. The activation response induced by BS-I was inhibited in the presence of genistein, a tyrosine kinase inhibitor, and by mepacrine, an inhibitor of phospholipase A2. Thus, several agonists that specifically bind to gal evoke both type I and type II EC activation. Binding of elicited xenoreactive anti-
226
Robson
bodies to gal (and potentially other) carbohydrate epitopes may contribute to AVR/DXR (potentially in the absence of c activation) (70–72).
VASCULAR INJURY FOLLOWING XENOGENEIC CELL TRANSPLANTATION In order to achieve specific immunological tolerance to allografts and concordant xenografts, Sachs, Sykes, and colleagues have developed models in which mixed hematopoietic chimerism have been established in rodents and nonhuman primates (73–75). In an attempt to extend this approach to the discordant pig-to-baboon combination, high doses of porcine peripheral blood leukocytes and mobilized progenitor cells (2–4×1010 cells/kg) have been infused into baboons undergoing non-myeloablative conditioning regimens (131–133). All recipients of porcine cells have developed a microangiopathic hemolytic anemia with thrombotic injury predominantly involving the microvasculature of the lung, kidneys, and brain. Thrombotic microangiopathy is a serious complication of bone marrow transplantation (BMT) that resembles thrombotic thrombocytopenic purpura (TTP); these pathological entities are characterized by the selective consumption of platelets, usually without prominent coagulation changes and the presence of platelet microthrombi in the microvasculature (76,77). Excessive intravascular platelet aggregation has been associated with appearance in plasma of unusually large vWF multimers in the classic forms of TTP (78), but not in thrombotic microangiopathy (79,80). We have recently performed investigations to analyze the pathobiology of this microangiopathic process and to test therapeutic interventions to ameliorate it. Both conditioned and naive baboons that received porcine cells developed intravascular platelet clumping progressing to severe thrombocytopenia (<10,000/mm3), intravascular hemolysis with schistocytosis (>10/hpf), increases in plasma lactate dehydrogenase (LDH) (2500–9000 U/L), transient neurological changes, renal insufficiency, and purpura. Autopsies on two baboons confirmed extensive platelet thrombi in the microcirculation. The conditioning regimen alone had self-limiting effects on platelet numbers and did not induce overt vascular injury. Baboons receiving the standard conditioning regimen and porcine cell infusions with the prophylactic administration of heparin, prostacycline, and high-dose steroids developed substantive albeit less profound thrombocytopenia (20,000/mm3) rarely requiring platelet
Disordered Coagulation
227
transfusions (Fig. 2), minimal schistocytosis (< 3 hpf), minor increase in LDH levels (< 1,000 U/L), and no clinical sequelae. No unusually large vWF multimers or changes in vWF protease activity were seen in plasma of baboons following the infusion of porcine cells.
Fig. 2. Porcine leukocyte/precursor cell infusions induce thrombocytopenia as a component of the microangiopathic injury in baboons. Changes in platelet counts in representative baboons following conditioning with whole body and thymic irradiation (WBI and TI) and anti-thymocyte globulin (ATG); prophylactic heparin, prostacycline and methylprednisone were also administered to the experimental animals. Note the initiation of thrombocytopenia occurring postirradiation and ATG, that is exacerbated after the cells were infused with platelet recovery within 15 d.
228
Robson
Although cyclosporin has been implicated in causing thrombotic microangiopathy and may induce vWF secretion from endothelial cells, this disorder occurred independently of cyclosporin administration and the conditioning regimen. In all baboons that survived with restoration of their platelet counts, resolution was accompanied by normalization of serum LDH levels and by a decrease in peripheral blood schistocytosis. The thrombotic microangiopathy occurring in our studies has some similarities to the clinical presentation of BMT-associated thrombotic microangiopathy in humans. Notably, this occurs in baboons without significant changes in the vWF multimer patterns or in vWF-cleaving protease activity. Also, plasma from affected baboons did not induce thrombocytopenia when injected into healthy baboons. However, the timing of onset of thrombotic microangiopathy observed in baboons following xenogeneic leukocytes and precursor cells was considered quite early when compared to clinical BMT-associated thrombotic microangiopathy (which typically occurs > 4 wk posttransplant) (81). The platelet consumption observed in baboons following porcine cell infusion could present in an accelerated manner secondary to interactions with the xenogeneic cells and consequent immune-mediated destruction. To counter this, all animals were splenectomized prior to infusion of cells but this appeared to make little difference. We are currently investigating the mechanisms whereby the porcine leukocyte and precursor cell preparations activate primate endothelial cells and platelets in vitro. Anticoagulant proteins expressed by the porcine cells are comparable to those on EC and would be largely ineffective across discordant xenogeneic species barriers. These molecular incompatibilities could predispose to the DIC seen with rejection of vascularized xenografts and the thrombotic microangiopathy seen in our studies with xenogeneic cellular transplants and are addressed in the next section.
XENOGENEIC MOLECULAR INCOMPATIBILITIES Theoretically, a major difference between allografts and xenografts is the degree of molecular compatibility between the graft and the recipient. This would be important not only for immunological recognition but also for the regulation of inflammatory responses and appropriate physiological functioning across species barriers. Specifically, allografts would have a low degree of molecular incompatibility or difference that would exist chiefly within the polymorphic
Disordered Coagulation
229
regions of the major histocompatibility complex (MHC); innate humoral responses would not be apparent in naive recipients. HAR of allografts could only occur with prior sensitization leading to anti-MHC antibodies or isoagglutinins. This modality of allograft rejection could be managed by prevention of allosensitization, appropriate screening, and avoidance or if the transplantation is absolutely required by preoperative plasmapheresis and high-dose immunosuppression to suppress B-cell function; such interventions may be required with acquired humoral-type rejection responses posttransplantation. Acute allograft rejection is a T-cell-mediated phenomenon and preventable or treatable by an array of potent and specific immunosuppressive agents. The vascular injury that occurs with humoral (and cytotoxic) reactions is associated with platelet sequestration and fibrin deposits. The molecular incompatibilities of xenografts extend much further than those of allografts. Those of the clinically relevant pig-to-human combination are still incompletely studied. Concordant xenografts do not induce HAR but both humoral and cellular mechanisms are intense, causing accelerated-type rejection. Discordant xenografts between widely disparate species immediately induce innate inflammatory responses and are subject to HAR because of preformed XNA. At an immunological level, the differences between pigs and humans are so great that grafts are destroyed before the adaptive immune system can be activated. It is interesting to speculate that the xeno-response may have alternative functions as an innate protective mechanism, e.g., against interspecies retroviral infections to protect the germ-line and genome of somatic cells. Certainly, aspects of these reactions are comparable to those targeting parasites, bacteria, and viruses. Success in overcoming HAR, the first barrier of molecular incompatibility relevant for the anti-gal XNA binding and c dysregulation, has uncovered further barriers relevant to AVR/DXR. These are also characterized by novel incompatibilities, such as cytokine/receptor, adhesion molecules/ligand, and, especially relevant here, in the regulation of the coagulation cascade, an important inflammatory component. Cellular and additional immunological reactions remain to be more fully explored. It is likely that the uncovered forms of xenograft rejection will be novel, likely to be initially poorly understood, and not fully amenable to current therapeutic approaches. The nature of several discordant interspecies incompatibilities is such that many interfere with (patho-) physiological interactions rather than
230
Robson
simply presenting as immunologically recognizable differences. More difficulties could be predicted for pig-to-primate liver xenotransplantation because of the complex functions of the liver, e.g., synthesis of the plasma proteins responsible for multiple molecular interactions with peripheral sites. The sophisticated and highly interactive, regulated systems comprised of coagulation and c proteins may not function appropriately with xenogeneic regulators expressed on the recipient vasculature and leukocytes. Similarly, lipid metabolism and bile secretion may be compromised by differing metabolic requirements between pig and primate. Xenotransplantation is likely to be more feasible for functionally less complex organs, such as the heart, which make few molecules critical to the rest of the body and can be adequately autoregulated or required to respond to simple phylogenetic preserved molecules, such as catecholamines or adenosine.
Natural Anticoagulants and Regulators Quiescent EC express effective anti-complement, anticoagulant, and platelet antiaggregatory mechanisms that maintain circulatory homeostasis and vascular integrity under physiological conditions (14). In addition to their modulation during cell activation (38,82), it is possible that certain of these factors may not be completely effective across species barriers because of molecular incompatibilities between activated coagulation components and their inhibitors (83,84). This scenario could be considered analogous to the highly specific porcine endothelial RCA that are inoperative against activated human C components (9,15). Many of these antiproteases and natural anticoagulants are also expressed by monocytes (85,86), so that infiltration of the xenograft by host cells may paradoxically improve the disordered thromboregulatory state (85). In a contrary manner, xenogeneic infused monocytes may initiate more severe inflammatory events in the recipient than would be anticipated by allogeneic cells. The functions of certain relevant porcine natural anticoagulants have been characterized to some extent. The first to receive attention was HS. These proteoglycans normally function as cofactors for antithrombin and, in addition, serve as a major anchor for other anticoagulants on the EC surface. HS appear to work well across species but tend to be lost from the surface of activated EC, as seen during xenograft rejection (82).
Disordered Coagulation
231
We have shown that porcine TFPI does not effectively neutralize human factor Xa (FXa) (36,85), and there is aberrant activation of both human prothrombin and factor X by porcine EC in vitro (87) and ex vivo (41). Hence, TF-dependent and independent activation of coagulation factor X may be markedly increased on porcine EC (relative to human aortic EC) because human factor Xa activity can be only adequately inhibited by human EC-associated TFPI. These TF interactions with factor VII (a) seem to be predominant in generating the procoagulant process in xenogeneic systems in vitro (8). The expression and very prominent presence of TF and fibrin in histological sections of cardiac xenografts rejected at 3–5 d is in keeping with these observations (36). The expression of TFPI has also been analyzed during HAR and AVR/DXR by immunohistochemistry. Sections from normal pig hearts show widespread vascular labeling for TFPI while the cardiac xenografts subject to HAR in untreated baboon recipients showed fibrin deposition within the microvasculature, and persistence of TFPI labeling on graft EC. Xenografts surviving until d 6, in recipients treated with CVF, show dense deposition of fibrin, lack of TFPI labeling, and marked expression of TF by intragraft macrophages. To compound the preexisting molecular incompatibility, TFPI appears to be lost from the vasculature during DXR (36). The described loss of HS is probably also responsible for TFPI functional downregulation in AVR/DXR. TM, a key anticoagulant expressed by EC, is an important regulator of thrombin activity and generates the important anticoagulant-activated protein C. Inflammatory mediators such as TNF are very effective in suppressing transcription of TM and its internalization from the cell surface and degradation, resulting in its rapid disappearance from the vasculature. In addition, porcine TM appears incompatible with human thrombin and protein C (87,88). Porcine EC generate significantly less activated human PC activity than the matched human EC cultures under identical experimental circumstances; the data suggest a specific failure of the porcine EC expressed TM to adequately bind human thrombin and then catalyze human PC, generating insufficient levels of activated PC (87). These factors contribute to substantial generation of human thrombin from prothrombin by porcine EC (87,89) that is also observed following ex vivo perfusion of pig hearts by human blood (41). When combined with the HS modulation in xenograft rejection, this TM incompatibility suggests that thrombin could be an important inflam-
232
Robson
matory mediator in the porcine vasculature exposed to human blood. Another factor contributing to the abnormally high levels of thrombin generation could be the failure of the protein C system to inhibit TF-induced thrombin generation (90,91); specifically, the synergistic effects of TM and TFPI (92) may be compromised. Finally, at the Xenotransplantation conference in Nagoya, Levy and colleagues also proposed induction of xenogeneic prothrombinases (fgl2/fibroleukin) on porcine EC that could directly generate thrombin from prothrombin. Thrombin both initiates clotting and is an important mediator of EC activation (93) or xenogeneic platelet aggregation (94). Thrombin may also stabilize fibrin clots and prevent their dissolution by activating the zymogen termed thrombin activable fibrinolysis inhibitor (TAFI) (95–97). Hence, therapies to address the disordered coagulation within xenografts will likely initially focus on anti-thrombin modalities to determine proof-of-principle and later consider both TF- and thrombin-mediated events. These interventions would also have to include anti-platelet modalities of treatment, while maintaining adequate hemostasis required for the surgical procedure.
Fibrinolytic Abnormalities Early reports also suggest abnormalities in the regulation of human plasminogen activators by porcine plasminogen activator inhibitor type I (PAI-1; associated with vascular EC) (98); these may compromise fibrinolytic pathways and further promote vascular occlusion in xenografts. However, expression and cloning of PAI-1 has demonstrated that there are no obvious functional differences between the human and porcine serpins (99,100). Platt and colleagues have recently demonstrated that porcine vascular EC convert to an antifibrinolytic state following stimulation with human XNA and C. The shift is at least partly explained by an increased level of PAI-1. This increased antifibrinolytic activity may contribute to the thrombotic diathesis seen in AVR/DXR (101).
PLATELETS AND REGULATION OF CLOTTING The assembly of coagulation factors on the platelet surface is facilitated by several events. First, there is expression of specific receptors for coagulation factor V that are up-regulated as a consequence of platelet activation. Second, the platelet membrane aminophospholipids are exposed as a result of microparticle formation, which enhance the binding and assembly of the prothrombinase complex. Third, platelets contain
Disordered Coagulation
233
in their alpha granules high concentrations of fibrinogen and other coagulation proteins that, when expressed on the cell surface or secreted into the extracellular environment, promote local fibrin deposition. Platelet aggregation involves an initiation trigger (thrombin, exposure to collagen, and so on) amplified by the release of ADP and serotonin. The associated generation of thrombin serves as a potent positive feedback step. The relationship between xenograft rejection and platelet reactivity is highly complex (102). For instance, platelet binding to the activated c component C1q could result in the activation of the platelet fibrinogen receptor (GPIIbIIIa) with consequent expression of P-selectin, and development of procoagulant activity on platelets (103). Models of the acute interaction of human/primate platelets and porcine EC in vitro suggest that the platelet aggregation may be initiated by C-activation and thrombin generation. Expression of the platelet adhesion molecule P-selectin might promote platelet–leukocyte interactions resulting in expression of a procoagulative environment stimulated by monocyte TF (104). These changes may be highly relevant for the thrombotic microangiopathic changes seen in baboons receiving infusions of porcine leukocytes. Endothelial–platelet interactions and development of platelet aggregates appear to be prominent factors in xenograft rejection (94,102). Platelet sequestration within the xenograft may be pathogenetically linked to the expression of porcine vWF that interacts with human platelet receptors with high affinity and causes activation responses (105–107). We have shown that porcine vWF-A1-domains can be expressed as glycosyl phosphatidylinositol linked FLAG fusion proteins on COS7-cells. These transfected cells are able to spontaneously induce GPIbdependent aggregation and intracellular Ca2+ uptake of platelets. Comparable responses were seen for human, baboon, and cynamolgus platelet preparations (108). The enhanced potential of porcine vWF to associate with human platelet glycoprotein GPIb suggests that this will represent an important barrier to xenograft acceptance, as this receptor triggers platelet activation ab initio (108). Hence, exposure and expression of vWF in the xenogeneic subendothelium following EC retraction or injury could result in massive activation of circulating platelets with formation of aggregates even prior to activation of coagulation (7). Vascular injury could be further exacerbated by the associated loss of vascular ATPDase/CD39, a putative thromboregulatory factor that would work well across discordant species (38,109). Adenine nucleotides and other factors released by platelets may contribute directly to the elicitation of vascular inflammatory responses (110) in xenograft rejection.
234
Robson
Combinations of select anti-platelet modalities with parallel approaches to control thrombin generation would be an optimal first step to regulating the thrombotic components of AVR/DXR. The challenge remains to either specifically target these interventions to the xenograft or to develop well-tolerated systemic pharmacological approaches.
VIRAL MEDIATED PROCOAGULANT PATHWAYS Even in the absence of xenoreactive antibodies and C activation, xenogeneic EC injury by, for example, certain viruses with induction of procoagulant activity (111–113) and cell adhesion receptors (114,115) could have serious consequences for long-term graft survival. These may be of greater import than in the equivalent allograft situation because of the greater potential for altered thromboregulation, altered immune defenses (116–118), and targeting of certain viruses to highly expressed human c regulatory proteins that bypass species-specific cellular resistance to infection (119,120). Recently, porcine retroviruses have been characterized (121) and have been shown to have the potential to infect human and porcine cells (122,123). Concerns regarding the possible xenogeneic transmission of retroviruses are being addressed and fall beyond the scope of this article. However, the observation that human retroviral infection has been associated with clinical events of recurrent vascular thrombosis (124) may prove to be relevant in xenotransplantation outcomes. There is also the potential for other enveloped viruses (such as the Herpesviridae) to directly activate prothrombin and initiate clotting on their surfaces in the absence of cells (125,126). Attempts to control thrombotic or inflammatory responses within the xenograft could theoretically promote viral replication in a manner similar to that predicted for c inhibition and removal of gal epitopes in transgenic animals (117,118).
PROPOSED ANTITHROMBOTIC INTERVENTIONS Pharmacological Regulation of Disordered Thrombin Generation within the Xenograft Vasculature Ex vivo studies have been conducted using a working model of HAR of porcine hearts perfused by fresh, heparinized human blood with or without a peptidomimetic thrombin inhibitor. Thrombin inhibition, in this xenoperfusion model, prolonged survival, enhanced function of the explanted organ, and improved histological features at the time of rejection (41).
Disordered Coagulation
235
The effect of thrombin inhibition was also tested in both HAR and AVR/DXR following decomplementation with CVF in normal Lewis (Lew) rats and intrinsically C6 deficient in PVG (C6-/PVG) recipient rats. In this system, thrombin inhibition significantly improved graft survival in HAR, but the agent tested failed to prolong survival in CVF treated Lew rats or the PVG (C6-/PVG) recipient rats (42). This discrepancy appeared related to non-specificity of the serine protease inhibitor studied with the potential activation of alternative pathway of c via the inhibition of factor I. Systemic toxicity at high doses have now precluded the use of this agent in further large animal studies; comparable pharmacological agents are undergoing further evaluation.
Overexpression of Human TM Retroviral transduction of pig EC with the gene encoding for human thrombomodulin (hTM) has resulted in expression of high levels of specific TM-cofactor activity on porcine EC in vitro. In these experiments, polyadenylation sites and regulatory elements in the 3' untranslated region of full-length hTM cDNA were replaced by a synthetic polylinker containing Sal I, EcoR I and BamH I restriction sites. The resulting hTM cDNA (1893 bp) was cloned as a Sal I fragment into the Xho I site of the retroviral vector pNTK-2 in front of an internal thymidine kinase promotor flanked by the two long terminal repeats (Fig. 3). Expression of hTM resulted in a 620-fold higher activation of human protein C in the presence of human thrombin when compared to mock transduced porcine EC (127). These data show that the functional deficiency of the anticoagulant protein C pathway may be corrected by somatic recombination. Expression of hTM by adenoviral vectors or other approaches (including transgenesis) may have therapeutic utility
Fig. 3. Retroviral vector for thrombomodulin expression. Truncated human Tm cDNA (1893 bp) was cloned as a Sal I fragment into the Xho I site of the retroviral vector pNTK-2, adjacent to an internal thymidine kinase promotor flanked by the two long terminal repeats (see text for details).
236
Robson
in the genetic modification of both porcine xenografts and bone marrow derived cells (128).
Platelet Inhibitors and Thromboregulatory Factors Inhibition of platelet aggregation by treatment of xenograft recipients with antagonists to the platelet fibrinogen receptor, GPIIbIIIa (44,45), by the use of P-selectin or PAF antagonists (46–48), or by administration of a soluble ATPDase (49) has been generally shown to prolong graft survival in several discordant xenotransplantation models. INHIBITORS OF VWF–GPIB Primate platelet aggregation mediated via GPIb–porcine vWF interactions could be controlled by the thromboregulatory mechanisms described above. Recent reports and our own unpublished data have also shown that prostacyclin and nitroprusside, vasodilator compounds that enhance cellular and platelet cAMP and cGMP concentrations, respectively, can cause a drastic inhibition of vWF-induced platelet responses (comparable to the xenogeneic situation) (129). More specific vWF-inhibitors [viz. aurin-tricarboxylic acid, soluble A1 domains (RG 12986) and VCL, antibodies, and so on] have been tested and shown to be effective in vitro (108). VWF-deficient pigs have been used as renal donors to xenoantibodydepleted baboons without effect on graft survival or on the microvascular thrombosis associated with AVR/DXR (130). However, the effect on platelet counts was not studied and circulating C activity was not inhibited. INHIBITORS OF FIBRINOGEN-GPIIbIIIa The integrin GPIIbIIIa is known to be crucial to the formation of platelet aggregates and potentiates adhesion to subendothelial matrices via fibrin(ogen), vWF, and vitronectin. We have already tested the effects of a specific GPIIbIIIa antagonist during xenograft rejection in small animals and have unpublished data in primates. Lew rats have received heterotopic guinea pig cardiac xenografts and were then treated with a peptidomimetic GPIIbIIIa antagonist alone (HAR model) or in combination with CVF (AVR/DXR model). Plasma from animals in the high-dose group completely inhibited platelet aggregation in vitro. Similarly, the agent induced a statistically significant increase in graft survival in both HAR and AVR/DXR groups. The combination of the GPIIbIIIa antagonist and CVF resulted in a significant decrease in
Disordered Coagulation
237
intragraft platelet aggregation, P-selectin expression, and leukocyte infiltration (44). THROMBOREGULATORY VASCULAR ATPDASES CD39 is an integral membrane protein of EC that degrades ATP and ADP to AMP with ultimate generation of adenosine. These nucleotides are released from damaged EC and promote platelet and EC activation. Adenosine has potent anti-platelet aggregatory and antiinflammatory actions. The loss of CD39 activity associated with EC activation may be a consequence of oxidative stress and contributes to the generation of a procoagulant phenotype by EC following exposure to oxidants, injury or activation. The cloning, characterization, and study of this and related agents in vascular inflammatory conditions, including xenograft rejection have been described (37,38,49,109). It is relevant to comment here that cd39deficient mice have markedly heightened thrombotic responses to vascular injury and cd39-deficient hearts are subject to more rapid and severe AVR/DXR when transplanted to CVF-treated sensitized rats. Extrapolation suggests that supplementation of ATPDases or overexpression of CD39 in xenografts may assist in overcoming the thrombotic components of AVR/DXR. Such experiments have been done and have largely validated this approach. This use of CD39 as an anti-platelet agent acting at an earlier phase of platelet activation would have certain benefits over the GPIIbIIIa antagonists that operate at a final common pathway and do not preclude release of platelet mediators. Additional antiinflammatory effects and the maintenance of vascular integrity by CD39, a natural vascular-associated thromboregulatory factor, are other putative theoretical benefits. Finally, targeted local expression of CD39 may have advantages to systemic administration of soluble NTPDase derivative. However, both options remain to be explored. Current pharmacological interventions to block platelet activation have been associated with limited therapeutic success and adverse clinical events. These concerns are particularly relevant in the setting of a surgical procedure or multiple requirements for vascular access. Available experimental data suggest that therapeutic modalities targeting nucleotide-mediated platelet activation and the evolution of rational drug delivery systems may facilitate control of the thrombotic component of xenograft rejection while preserving adequate hemostasis. ATPDases could be also safely combined with direct anti-thrombin modalities (including heparin) or certain fibrinolytic interventions.
238
Robson
CONCLUSIONS There are several proposed pharmacological therapies for dealing with the problem of C activation, XNA, and elicited xenoreactive antibodies in addition to the genetic engineering approach. Adequate immunosuppression, other interventions such as plasmapheresis, anti-C therapies, and the control of any resulting EC activation process following binding of xenoreactive antibodies will be of crucial import to the continued progress in this interesting area. As evident from consideration of putative pathogenetic modes, it can be argued that, although thrombosis may be a consequence of the immunological response to the xenograft and validly viewed as a component of rejection, it may also arise as a consequence of molecular incompatibilities of nonimmunological origin. These molecular barriers may persist even if the xenogeneic immune response is completely inhibited and xenoantibodies fully controlled. The presence of an incompatible thrombophilic vasculature within the xenograft may compromise longterm survival and function. Hence, the need for additional genetic engineering of donor animals with expression of human anticoagulants and platelet regulatory factors. An alternative approach may be to introduce long-term anticoagulation in recipients of xenografts. The concept that disordered thromboregulation is a component of AVR/DXR has led to novel ideas that may translate into approaches to therapy. As these problems are overcome, it is assumed that effective therapy for subsequent, graft-specific cellular responses will be necessary. However, whether all of this will lead to clinical application of discordant xenografting of immediately vascularized organs must await the appropriate experimentation and acceptance of the proposed risk/ benefit analysis by the transplant community.
ACKNOWLEDGMENTS I am grateful to Dr. D’Apice for recent important discussions in this area; also to Drs. Bach, Cooper, Fishman, Sachs, and White for sharing their expertise and advising me over the past five years. The important direct contributions of Drs. Alwayn, Buhler, Kopp, Young, Imai, Ierino, Lesnikoski, Candinas, Schulte am Esch, Siegel, and Goepfert to the experimentation reported on here have facilitated the generation and testing of the hypotheses proposed in this review. I apologize if I have
Disordered Coagulation
239
not adequately referenced the work of others in this area but are constrained by space requirements.
REFERENCES 1. Taniguchi S, Cooper D. Clinical xenotransplantation - past, present and future. Ann R Coll Surg Engl 1997; 79:13–19. 2. Platt JL, Bach FH. Discordant xenografting: challenges and controversies. Curr Opin Immunol 1991; 3:735–739. 3. Platt, J.L. Xenotransplantation - recent progress and current perspectives. Curr Opin Immunol 1996; 8:721–728. 4. Platt JL. The immunological barriers to xenotransplantation. Crit Rev Immunol 1996; 16:331–358. 5. Parker W, Saadi S, Lin SS, Holzknecht ZE, Bustos M, Platt JL. Transplantation of discordant xenografts - a challenge revisited. Immunol Today 1996; 17:373–378. 6. Platt JL, Bach FH. The barrier to xenotransplantation. Transplantation 1991; 52:937–947. 7. Bach FH, Winkler H, Ferran C, Hancock WW, Robson SC. Delayed xenograft rejection. Immunol Today 1996; 17:379–384. 8. Bach FH, Robson SC, Ferran C, et al. Endothelial cell activation and thromboregulation during xenograft rejection. Immunol Reviews 1994; 141:5–30. 9. Dalmasso AP. The complement system in xenotransplantation. Immunopharmacology 1992; 24:149–160. 10. Bach FH, Ferran C, Soares M, et al. Modification of vascular responses in xenotransplantation - inflammation and apoptosis. Nature Medicine 1997; 3:944–948. 11. Platt JL. New directions for organ transplantation. Nature 1998; 392:11 – 17. 12. Goodman DJ, Pearse MJ, Dapice AJF. Overcoming hyperacute xenograft rejection with transgenic animals. Biodrugs 1998; 9:219–234. 13. Dorling A, Lechler RI. Prospects for xenografting. Curr Opin Immunol 1994; 6:765–769. 14. Robson SC, Candinas D, Hancock WW, Wrighton C, Winkler H, Bach,FH. Role of endothelial cells in transplantation. Int Arch Allergy Immunol 1995; 106:305–322. 15. Dalmasso AP, Vercellotti GM, Fischel RJ, Bolman RM, Bach FH, Platt JL. Mechanism of complement activation in the hyperacute rejection of porcine organs transplanted into primate recipients. Amer J Pathol 1992; 140:1157–1166. 16. Galili U. Interaction of the natural anti-gal antibody with alpha-galactosyl epitopes - a major obstacle for xenotransplantation in humans. Immunol Today 1993; 14:480–482. 17. Cooper D, Ye Y, Niekrasz M, et al. Specific intravenous carbohydrate therapy - a new concept in inhibiting antibody-mediated rejection experience with ABO-incompatible cardiac allografting in the baboon. Transplantation 1993; 56:769–777. 18. Cooper DKC, Koren E, Oriol R. _-galactosyl oligosaccharides and discordant xenografting. Xeno 1994; 2. 19. Pruitt SK, Bollinger RR, Collins BH, et al. Effect of continuous complement inhibition using soluble complement receptor type 1 on survival of pig-to-primate cardiac xenografts. Transplantation 1997; 63:900–902. 20. Hancock WW. Delayed Xenograft Rejection [Review]. World J Surg 1997; 21:917–923.
240
Robson
21. Rosengard AM, Cary NR, Langford GA, Tucker AW, Wallwork J, White DJ. Tissue expression of human complement inhibitor, decay-accelerating factor, in transgenic pigs. A potential approach for preventing xenograft rejection. Transplantation 1995; 59:1325–1333. 22. Sandrin MS, Fodor WL, Mouhtouris E, et al. Enzymatic remodeling of the carbohydrate surface of a xenogeneic cell substantially reduces human antibody binding and complement-mediated cytolysis. Nat Med 1995; 1:1261–1267. 23. Sandrin MS, Fodor WL, Cohney S, et al. Reduction of the major porcine xenoantigen ga1-alpha(1,3)gal by expression of alpha(1,2)fucosyltransferase. Xeno-transplantation 1996; 3:134–140. 24. Byrne G, Mccurry KR, Martin MJ, Mcclellan SM, Platt JL, Logan JS. Transgenic pigs expressing human CD59 and decay-accelerating factor produce an intrinsic barrier to complement-mediated damage. Transplantation 1997; 63:149–155. 25. Diamond LE, Mccurry KR, Martin MJ, McClellan SB, Oldham ER, Platt JL, Logan JS. Characterization of transgenic pigs expressing functionally active human CD59 on cardiac endothelium. Transplantation 1996; 61:1241–1249. 26. Lin SS, Weidner BC, Byrne GW, et al. The role of antibodies in acute vascular rejection of pig-to-baboon cardiac transplants. J Clin Invest 1998; 101:1745–1756. 27. Saadi S, Holzknecht RA, Patte CP, Stern DM, Platt JL. Complement-mediated regulation of tissue factor activity in endothelium. J Exper Med 1995; 182:1807–1814. 28. Platt JL, Lin SS, McGregor CGA. Acute vascular rejection. Xenotransplantation 1998; 5:169–175. 29. Dalmasso AP, Platt JL, Bach FH. Reaction of complement with endothelial cells in a model of xenotransplantation. Clin Exper Immunol 1991; 1:31–35. 30. Candinas D, Lesnikoski BA, Robson SC, et al. Effect of repetitive high-dose treatment with soluble complement receptor type 1 and cobra venom factor on discordant xenograft survival. Transplantation 1996; 62:336–342. 31. Bach FH. Revisiting a challenge in transplantation: discordant xenografting. Hum Immunol 1991; 30:262–269. 32. Oriol R, Ye Y, Koren E, Cooper D. In vivo depletion of xenoreactive natural antibodies with an anti-mu monoclonal antibody. Transplantation 1993; 56:1427–1433. 33. McCurry KR, Kooyman DL, Diamond LE, et al. Human complement regulatory proteins in transgenic animals regulate complement activation in xenoperfused organs. Transplant Proc 1995; 27:317–318. 34. McCurry KR, Kooyman DL, Alvarado CG, et al. Human complement regulatory proteins protect swine-to-primate cardiac xenografts from humoral injury. Nat Med 1995; 1:423–427. 35. Koike C, Isobe K, Nakashima I, et al. Establishment of a human DAF/HRF20 double transgenic mouse line is not sufficient to suppress hyperacute rejection. Surgery Today1996; 26:993–998. 36. Kopp CW, Siegel JB, Hancock WW, et al. Effect of porcine endothelial tissue factor pathway inhibitor on human coagulation factors. Transplantation 1997; 63:749–758. 37. Candinas D, Koyamada N, Miyatake T, et al. Loss of rat glomerular ATP diphosphohydrolase activity during reperfusion injury is associated with oxidative stress reactions. Thromb Haemost 1996; 76:807–812. 38. Robson SC, Kaczmarek E, Siegel JB, et al. Loss of ATP diphosphohydrolase activity with endothelial cell activation. J Exper Med 1997; 185:153–163.
Disordered Coagulation
241
39. Ierino FL, Kozlowski T, Siegel JB, et al. (1998) Disseminated intravascular coagulation in association with the delayed rejection of pig-to-baboon renal xenografts. Transplantation 1998; 66:1439–1450. 40. Kozlowski T, Shimizu A, Lambrigts D, et al. Porcine kidney and heat transplantation in baboons undergoing a tolerance induction regimen and antibody adsorption. Transplantation 1999; 67:18–30. 41. Robson SC, Young VK. Cook NS, et al. Thrombin inhibition in an ex vivo model of hyperacute xenograft rejection. Transplantation 1996; 61:862–868. 42. Lesnikoski BA, Candinas D, Otsu I, Metternich R, Bach FH, Robson SC. Thrombin inhibition in discordant xenograft rejection. Xenotransplantation 1997; 4:140–146. 43. Jakobs FM, Davis EA, White T, Sanfilippo F, Baldwin WM. Prolonged discordant xenograft survival by inhibition of the intrinsic coagulation pathway in complement C6-deficient recipients. J Heart Lung Transplant 1998; 17:306–311. 44. Candinas D, Lesnikoski BA, Hancock WW, et al. Inhibition of platelet integrin GPIIbIIIa prolongs survival of discordant cardiac xenografts. Transplantation 1996; 62:1–5. 45. Robson SC, Young VK, Cook NS, et al. Inhibition of platelet gpIIbIIIa in an ex vivo model of hyperacute xenograft rejection does not prolong cardiac survival time. Xenotransplantation 1996; 3:43–52. 46. Makowka L, Chapman FA, Cramer DV, Qian SG, Sun H, Starzl TE. Plateletactivating factor and hyperacute rejection. The effect of a platelet-activating factor antagonist, SRI 63-441, on rejection of xenografts and allografts in sensitized hosts. Transplantation 1990; 50:359–365. 47. Ohair DP, Roza AM, Komorowski R, et al. Tulopafant, a PAF receptor antagonist, increases capillary patency and prolongs survival in discordant cardiac xenotransplants. Journal of Lipid Mediators. 7:79–84. 48. Coughlan AF, Berndt MC, Dunlop LC, Hancock WW. In vivo studies of P-selectin and platelet activating factor during endotoxemia, accelerated allograft rejection, and discordant xenograft rejection. Transplant Proc 1993; 25:2930–2931. 49. Koyamada N, Miyatake T, Candinas D, et al. (1996) Apyrase administration prolongs discordant xenograft survival. Transplantation 1996; 62:1739–1743. 50. Enjyoji K, Sevigny J, Lin Y, et al. Targeted disruption of cd39/ATP diphosphohydrolase results in disordered hemostasis and thromboregulation. Nat Med 1999; 5:1010–1017. 51. Imai M, Takigama K, Guckelberger O. Modulation of nucleoside triphosphate diphosphohydrolase-1/cd39 in xenograft rejection. Molec Med 1999; 5:743–752. 52. Imai M, Takigami K, Guckelberger O. Recombinant adenoviral mediated CD39 gene transfer prolongs cardiac xenograft survival. Transplantation 2000; In press. 53. Rosenberg JC, Broersma RJ, Bullemer G, Mammen EF, Lenaghan R, Rosenberg B.F. Relationship of platelets, blood coagulation, and fibrinolysis to hyperacute rejection of renal xenografts. Transplantation 1969; 8:152–161. 54. Broersma RJ, Bullemer GD, Rosenberg JC, Lenaghan R, Rosenberg BF, Mammen EF. Coagulation changes in hyperacute rejection of renal xenografts. Thrombosis et Diathesis Haemorrhagica - Supplementum 1969; 36:333–340. 55. Breimer ME, Bjorck S, Svalander CT, et al. Extracorporeal (ex vivo) connection of pig kidneys to humans. 1. Clinical data and studies of platelet destruction. Xenotransplantation 1996; 3:328–339. 56. Collins BH, Chari RS, Magee JC, et al. Immunopathology of porcine livers perfused with blood of humans with fulminant hepatic failure. Transplant Proc 1995; 27:280–281.
242
Robson
57. Davis EA, Jakobs F, Pruitt SK, et al. Overcoming rejection in pig-to-primate cardiac xenotransplantation. Transplant Proc 1997; 29:938–939. 58. Davis EA, Pruitt SK, Greene PS, et al. (1996) Inhibition of complement, evoked antibody, and cellular response prevents rejection of pig-to-primate cardiac xenografts. Transplantation 1996; 62:1018–1023. 59. Schmoeckel M, Nollert G, Shahmohammadi M, et al. Transgenic human decay accelerating factor makes normal pigs function as a concordant species. J Heart Lung Transplant 1997; 16:758–764. 60. Tucker, A.W., Carrington, C.A., Richards, A.C., Robson, S.C., and White, D. (1997) Endothelial cells from human decay acceleration factor transgenic pigs are protected against complement mediated tissue factor expression in vitro. Transplant Proc 29. 61. White DJG, Schmoeckel M, Cozzi E, Chavez G, Langford G. Genetic engineering of pigs for xenogeneic heart transplantation [French]. Biodrugs1997; 8:33–36. 62. Cozzi E, Tucker AW, Langford GA, et al. Characterization of pigs transgenic for human decay-accelerating factor. Transplantation 1997; 64:1383–1392. 63. Zaidi A, Schmoeckel M, Bhatti F, et al. (1998) Life-supporting pig-to-primate renal xenotransplantation using genetically modified donors. Transplantation 1997; 65:1584–1590. 64. Schmoeckel M, Bhatti FNK, Zaidi A, et al. Orthotopic heart transplantation in a transgenic pig-to-primate model. Transplantation 1998; 65:1570–1577. 65. Gibbons GH, Dzau VJ. The emerging concept of vascular remodeling N Engl J Med 1994;330:1431–1438. 66. Colvin M, Hilton J. Pharmacology of cyclophosphamide and metabolites. Cancer Treatment Reports 1981; 65:89–95. 67. Starzl TE, Boehmig HJ, Amemiya H, et al. Clotting changes, including disseminated intravascular coagulation, during rapid renal-homograft rejection. N Engl J Med 1970; 283:383–390. 68. Starzl TE, Lerner RA, Dixon FJ, Groth CG, Brettschneider L, Terasaki PI. Shwartzman reaction after human renal homotransplantation. N Engl J Med 1968; 278:642–648. 69. Saadi S, Platt JL. Transient perturbation of endothelial integrity induced by natural antibodies and complement. J Exper Med 1995; 181:21–31. 70. Palmetshofer A, Galili U, Dalmasso AP, Robson SC, Bach FH. Alpha-galactosyl epitope-mediated activation of porcine aortic endothelial cells - type II activation. Transplantation 1998; 65:971–978. 71. Palmetshofer A, Galili U, Dalmasso AP, Robson SC, Bach FH. Alpha-galactosyl epitope-mediated activation of porcine aortic endothelial cells - type I activation. Transplantation 1998; 65:844–853. 72. Palmetshofer A, Robson SC, Bach FH. Tyrosine phosphorylation following lectin mediated endothelial cell stimulation. Xenotransplantation1998; 5:61–66. 73. Gritsch HA, Glaser RM, Emery DW, et al. The importance of nonimmune factors in reconstitution by discordant xenogeneic hematopoietic cells. Transplantation 1994; 57:906–917. 74. Kawai T, Cosimi AB, Colvin RB, et al. Mixed allogeneic chimerism and renal allograft tolerance in cynomolgus monkeys. Transplantation 1995; 59:256–262. 75. Latinne D, Vitiello DM, Sachs DH, Sykes M. Tolerance to discordant xenografts. 1. Sharing of human natural antibody determinants on miniature swine bone marrow cells and endothelial cells. Transplantation 1994; 57:238–245.
Disordered Coagulation
243
76. Wada H, Kaneko T, Ohiwa M, et al. Increased levels of vascular endothelial cell markers in thrombotic thrombocytopenic purpura. Am J Hematol 1993; 44:101–105. 77. Takahashi H, Tatewaki W, Nakamura T, Hanano M, Wada K, Shibata A. Coagulation studies in thrombotic thrombocytopenic purpura, with special reference to von Willebrand factor and protein S. Am J Hematol 1989;30:14#–321. 78. Furlan M, Robles R, Solenthaler M, Wassmer M, Sandoz P, Lammle B. Deficient activity of von Willebrand factor-cleaving protease in chronic relapsing thrombotic thrombocytopenic purpura. Blood 1997; 89:3097–3103. 79. Sarode R, Mcfarland JG, Flomenberg N, et al. (1995) Therapeutic plasma exchange does not appear to be effective in the management of thrombotic thrombocytopenic purpura hemolytic uremic syndrome following bone marrow transplantation. Bone Marrow Transplant 1995; 16:271–275. 80. Wassmann B, Martin H, Elsner S, Bruecher J, Thaiss F, Stahl R, Hoelzer D. (1994) Microangiopathic hemolytic anemia and renal impairment following autologous bone marrow transplantation - a case of hemolytic uremic syndrome. Bone Marrow Transplant 1994;14:849–851. 81. Pucci G, Martino M, Morabito F, et al. Thrombotic thrombocytopenic purpura - a rare late complication of allogeneic bone marrow transplantation. Haematologica 1994; 79:371–373. 82. Platt JL, Vercellotti GM, Lindman BJ, Oegema TJ, Bach FH, Dalmasso AP. (1990) Release of heparan sulfate from endothelial cells. Implications for pathogenesis of hyperacute rejection. J Exp Med 1990; 171:1363–1368. 83. Robson SC, Kopp C. Disordered thromboregulation in discordant xenograft rejection. Life Science 1996; 6: 34–38. 84. Lawson JH, Platt JL. (1996) Molecular barriers to xenotransplantation. Transplantation 1996; 62:303–310. 85. Kopp CW, Robson SC, Siegel JB, et al. Regulation of monocyte tissue factor activity by allogeneic and xenogeneic endothelial cells. ThrombHaemost 1998; 79:529–538. 86. Grey ST, Hancock WW. A physiologic anti-inflammatory pathway based on thrombomodulin expression and generation of activated protein C by human mononuclear phagocytes. J Immunol 1996; 156:2256–2263. 87. Siegel JB, Grey ST, Lesnikoski BA, et al. Xenogeneic endothelial cells activate human prothrombin. Transplantation 1997; 64:888–896. 88. Lawson JH, Daniels LJ, Platt JL. The evaluation of thrombomodulin activity in porcine to human xenotransplantation. Transplant Proc 1997; 29:884–885. 89. Jurd KM, Gibbs RV, Hunt BJ. Activation of human prothrombin by porcine aortic endothelial cells - a potential barrier to pig to human xenotransplantation. Blood Coagul Fibrinolysis 1996; 7:336–343. 90. Van’t Veer C, Golden NJ, Kalafatis M, Mann KG. (1997) Inhibitory mechanism of the protein C pathway on tissue factor-induced thrombin generation - synergistic effect in combination with tissue factor pathway inhibitor. J Biol Chem 1997; 272:7983–7994. 91. Van’t Veer C, Mann KG. Regulation of tissue factor initiated thrombin generation by the stoichiometric inhibitors tissue factor pathway inhibitor, antithrombin–III, and heparin cofactor-II. J Biol Chem 1997; 272:4367–4377. 92. Van’t Veer C, Hackeng TM, Delahaye C, Sixma JJ, Bouma BN. Activated factor X and thrombin formation triggered by tissue factor on endothelial cell matrix in a flow model: effect of the tissue factor pathway inhibitor. Blood 1994; 84:1132–1142.
244
Robson
93. Anrather D, Millan MT, Palmetshofer A, et al. Thrombin activates nuclear factorkappa-B and potentiates endothelial cell activation by TNF. J Immunol 1997; 159:5620–5628. 94. Robson SC, Siegel JB, Lesnikoski BA, et al. Aggregation of human platelets induced by porcine endothelial cells is dependent upon both activation of complement and thrombin generation. Xenotransplantation 1996; 3:24–34. 95. Bajzar L, Nesheim M, Morser J, Tracy PB. Both cellular and soluble forms of thrombomodulin inhibit fibrinolysis by potentiating the activation of thrombinactivable fibrinolysis inhibitor. J Biol Chem 1998; 273:2792–2798. 96. Kokame K, Zheng XL, Sadler JE. Activation of thrombin-activable fibrinolysis inhibitor requires epidermal growth factor-like domain 3 of thrombomodulin and is inhibited competitively by protein C. J Biol Chem 1998; 273:12,135–12,139. 97. Bajzar L, Manuel R, Nesheim ME. Purification and characterization of TAFI, a thrombin-activable fibrinolysis inhibitor. J Biol Chem 1995; 270:14,477–14,484. 98. Fay WP, Murphy JG, Owen WG. High concentrations of active plasminogen activator inhibitor-1 in porcine coronary artery thrombi. Arterioscler Thromb Vasc Biol 1996; 16:1277–1284. 99. Debrock S, Declerck PJ. Identification of a functional epitope in plasminogen activator inhibitor-1, not localized in the reactive center loop. Thromb Haemost 1998; 79:597–601. 100. Bijnens AP, Knockaert I, Cousin E, Kruithof EK, Declerck PJ. Expression and characterization of recombinant porcine plasminogen activator inhibitor-1 [published erratum appears in Thromb Haemost 1997 May;77(5):1046]. Thromb Haemost 1997; 77:350–356. 101. Kalady MF, Lawson JH, Sorrell RD, Platt JL. Decreased fibrinolytic activity in porcine-to-primate cardiac xenotransplantation. Mol Med 1998; 4:629–637. 102. Robson SC, Kopp C, Lesnikoski E. Platelets in xenograft rejection. Xeno 1994; 2: 38–46. 103. Peerschke EIB, Reid KBM, Ghebrehiwet B. Platelet activation by C1q results in the induction of _II/`III integrins and the expression of P-selectin and procoagulant activity. J Exp Med 1993; 178:579–587. 104. Amirkhosravi, A., Alexander, M., May, K., et al. The importance of platelets in the expression of monocyte tissue factor antigen measured by a new whole blood flow cytometric assay. Thromb Haemost 1996; 75:87–95. 105. Robson SC, Schulte am Esch II J, Bach FH. Factors in xenograft rejection. Ann N Y Acad Sci Bioartificial Organs II. 875:261–276. 106. Pareti FI, Mazzucato M, Bottini E, Mannucci PM. Interaction of porcine von Willebrand factor with the platelet glycoproteins Ib and IIb/IIIa complex. Brit J Haematol 1992; 82:81–86. 107. Mazzucato M, Demarco L, Pradella P, Masotti A, Pareti FI. Porcine von Willebrand factor binding to human platelet GPIb induces transmembrane calcium influx. Thromb Haemost 1996; 75:655–660. 108. Schulte am Esch II J, Siegel JB, Cruz M, Anrather J, Robson SC. The A1 domain of von Willebrand factor expressed on cell membranes directly activates platelets. Blood 1997; 90:4425–4437. 109. Kaczmarek E, Koziak K, Sevigny J, et al. Identification and characterization of CD39 vascular ATP diphosphohydrolase. J Biol Chem 1996; 271:33,116–33,122.
Disordered Coagulation
245
110. Von Albertini M, Palmetshofer A, Kaczmarek E, et al. Extracellular ATP and ADP activate transcription factor NF-kappa-B and induce endothelial cell apoptosis. Biochem Biophys Res Commun 1998; 248:822–829. 111. Vandammieras M, Muller AD, Vanhinsbergh V, Mullers W, Bomans P, Bruggeman CA. The procoagulant response of cytomegalovirus infected endothelial cells. Thromb Haemost 1992; 68:364–370. 112. Li C, Fung LS, Chung S, et al. Monoclonal antiprothrombinase (3D4.3) prevents mortality from murine hepatitis virus (MHV-3) infection. J Exp Med 1992; 176:689–697. 113. Altieri DC, Etingin OR, Fair DS, et al. Structurally homologous ligand binding of integrin Mac-1 and viral glycoprotein-C receptors. Science 1991; 254:1200–1202. 114. Etingin OR, Silverstein RL, Hajjar DP. (1993) Von Willebrand factor mediates platelet adhesion to virally infected endothelial cells. Proc Nat Acad Sci USA 1993; 90:5153–5156. 115. Etingin OR, Silverstein RL, Hajjar DP. Identification of a monocyte receptor on herpesvirus- infected endothelial cells. Proc Nat Acad Sci USA 1991; 88:7200–7203. 116. Rother RP, Fodor WL, Springhorn JP, et al. (1995) A novel mechanism of retrovirus inactivation in human serum mediated by anti-alpha-galactosyl natural antibody. J Exp Med 1995; 182:1345–1355. 117. Michaels MG. Infectious concerns of cross-species transplantation - xenozoonoses. World J Surg 1997; 21:968–974. 118. Chapman LE, Folks TM, Salomon DR, Patterson AP, Eggerman TE, Noguchi PD. (Xenotransplantation and xenogeneic infections. N Engl J Med 1995; 333:1498–1501. 119. Thorley BR, Milland J, Christiansen D., et al. Transgenic expression of a CD46 (membrane cofactor protein) minigene - studies of xenotransplantation and measles virus infection. Eur J Immunol 1997; 27:726–734. 120. Mazure G, Grundy JE, Nygard G, et al. Measles virus induction of human endothelial cell tissue factor procoagulant activity in vitro. J Gen Virol 1994; 75:2863–2871. 121. Akiyoshi DE, Denaro M, Zhu HH, Greenstein JL, Banerjee P, Fishman JA. Identification of a full-length cDNA for an endogenous retrovirus of miniature swine. J Virol 1998; 72:4503–4507. 122. Wilson CA, Wong S, Muller J, Davidson CE, Rose TM, Burd P. (1998) Type C retrovirus released from porcine primary peripheral blood mononuclear cells infects human cells. J Virol 1998; 72:3082–3087. 123. Patience C, Takeuchi Y, Weiss RA. Infection of human cells by an endogenous retrovirus of pigs. Nat Med 1997; 3:282–286. 124. Toulon P, Lamine M, Ledjev I, Guez T, Holleman ME, Sereni D, Sicard D. Heparin cofactor-II deficiency in patients infected with the human immunodeficiency virus. Thromb Haemost 1993; 70:730–735. 125. Sutherland MR, Raynor CM, Leenknegt H, Wright JF, Pryzdial EL. (1997) Coagulation initiated on herpesviruses. Proc Nat Acad Sci USA 1997; 94:13,510–13,514. 126. Pryzdial EL, Wright JF. (1994) Prothrombinase assembly on an enveloped virus: evidence that the cytomegalovirus surface contains procoagulant phospholipid. Blood 1994; 84:3749–3757.
246
Robson
127. Kopp CW, Grey ST, Siegel JB, et al. Expression of human thrombomodulin cofactor activity in porcine endothelial cells. Transplantation 1998; 66:244–251. 128. Hayashi H, Lee RS, Germana S, et al. Retroviral vectors for long-term expression of allogeneic major histocompatibility complex transduced into syngeneic bone marrow cells. Transplan Proc 1995; 27:178–179. 129. Francesconi M, Casonato A, Pagan S, et al. Inhibitory effect of prostacyclin and nitroprusside on type IIb von willebrand factor–promoted platelet activation. Thromb Haemost 1996; 76:469–474. 130. Meyer C, Wolf P, Romain N, et al. Use of von Willebrand diseased kidney as donor in a pig-to-primate model of xenotransplantation. Transplantation 1999; 67:38–45. 131. Buhler L, Basker M, Always, IPJ, et al. Coagulation and thrombotic disorders associated with pig organ and hematopoietic cell transplantation in nonhuman primates. Transplantation 2000; 70:1323–1331. 132. Buhler L, Goepfert C, Kitamura H, et al. Porcine hematopoietic cell xenotransplantation in nonhuman primates is complicated by thrombotic microangiopathy. Bone Marrow Transplant 2001; 27:1227–1236. 133. Alwayn IPJ, Buhler L, Appel JZ, et al. Mechanisms of thrombotic microangiopathy following xenogeneic hematopoietic progenitor cell transplantation. Transplantation 2000; 71:1601–1609.
Current Applications of Cellular Xenografts
247
12 Current Applications of Cellular Xenografts Albert S.B. Edge INTRODUCTION Cellular xenografting is a new type of transplantation therapy that is finding increasing application in the treatment of a variety of diseases. The hundreds of distinct cell types present in mammals offer the possibility of treating diseases, disorders, and injuries characterized by cell death or dysfunction. This chapter will focus on the unique aspects of xenogeneic cell therapy and will discuss the specific applications in current use in the clinic and in animal models. Cell transplantation therapy can be viewed as serving two types of clinical need. In the first the cell is used primarily as a means of delivering a gene or gene product to a tissue that has a mutation causing a disease state. In the second, the cells are delivered to replace degenerating or apoptotic tissue and rebuild a functional organ. In the first application, the cell without genetic engineering is acting as a type of vehicle for gene therapy, whereas in the second, the cell is being used to repair a failing tissue in a mode more similar to tissue engineering. Although these approaches are distinct in theory, they overlap considerably in clinical practice. As an illustration of the types of diseases that can be treated by cell therapy for the correction of single gene defects or, alternatively, by introduction of a normal gene, metabolic defects in the liver are being treated by both cell therapy and gene therapy approaches. Familial hypercholesterolemia which results from a defect in the LDL receptor gene has been treated by gene replacement therapy and by cellular transplantation (1,2). A patient with Crigler–Najjar syndrome has been treated with normal human hepatocytes to replace the defective gene in From: Xenotransplantation: Basic Research and Clinical Applications Edited by: Jeffrey L. Platt © Humana Press Inc., Totowa, NJ
247
248
Edge
bilirubin metabolism, UDP-glucuronyltransferase (3), and the same defect has been treated by delivery of the gene by viral vectors in animal models of the disease (4). Hepatocytes could also be used for replacing clotting factors such as factor IX, which has been treated by gene therapy approaches. The cell in this paradigm is the ultimate vehicle for gene delivery as it has the advantage of metabolic integration that results from the expression of multiple genes and regulation by feedback mechanisms that coordinate cell metabolism. In the second type of application, cell transplantation therapy has been employed to replace degenerating or apoptotic tissue, a situation that would be difficult to envision being treated by gene therapy. Cartilage tissue is being repaired with articular chondrocytes (5,6). Neural cell transplantation for neurodegenerative diseases has seen clinical application (7). The lost neurons and connections in Parkinson’s disease and Huntington’s disease may be replenished by the transplantation of fetal neurons from human or animal sources (8–10). Remyelination of neurons in multiple sclerosis is another target for cellular therapy and animal models have shown that this may be feasible in some cases (11). Failing myocardial tissue is being treated with cardiomyocyte and myoblast transplantation in animal models (12). Diabetes is being treated by cell replacement therapy (13). In diabetes, cell therapy treatments overlap with genetic approaches insofar as diabetes has been treated with complex gene therapy protocols to make insulin secretion responsive to glucose concentrations without actually rebuilding the lost tissue. The limited application of cell therapy can be attributed to the lack of available human cells suitable for transplantation. Derivation of primary cells from humans is hindered by the lack of suitable organs, and cells that are kept in continuous culture have associated risks that have prevented their use in the clinic. The limitations to cell therapy are thus primarily due to cell availability, as illustrated by the fact that cell based therapies in which the cells are readily available, such as bone marrow replacement, are a standard part of current medical practice.
ADVANTAGES OF CELL TRANSPLANTATION The potential for broad therapeutic use of cell transplantation derives from the ability of cells to restore damaged tissue after integration into the host. Cells can be transplanted without many of the immunological barriers encountered with whole organ transplantation.
Current Applications of Cellular Xenografts
249
Cell transplantation confers an array of options in clinical practice: the ability to localize sites of cell dysfunction and tissue necrosis and ischemia using modern imaging techniques coupled with progress in radiology and stereotactic surgery is leading to the targeted reconstruction of damaged tissue. Indeed, the number of experimental and clinical applications of xenogeneic cell based therapy is rapidly increasing. Cell transplantation as opposed to whole organ transplantation carries a number of significant advantages: besides a lower immune barrier, cells are easily manipulated either by genetics or by incubation with enzymes, substrates, or media, cells can integrate and be regulated by the host, and the infusion of cells is far less invasive than the surgery involved in organ removal and replacement.
ADVANTAGES OF XENOTRANSPLANTATION Xenotransplantation has been considered as an alternative to allotransplantation beginning with attempted corneal transplants from animals to humans (14). Cellular xenotransplantation began with studies of porcine islets transplanted into diabetic animals (15,16). Many species have been considered for cellular xenotransplantation. Although primate cells might represent an obvious choice due to phylogenetic similarities, risks associated with transfer of infectious agents have limited research in this area. Porcine cells offer a number of important advantages. The supply of cells from porcine donors is virtually unlimited, and animals can be raised under quality controlled conditions to provide consistent cell populations. Pigs have been used extensively in research on xenotransplantation and have similar physiological responses in a number of systems (17,18). The advantages of using porcine cells becomes particularly apparent in clinical protocols that involve the use of human fetal tissue with the attendant ethical issues, lack of control over the timing of the harvest, and lack of control over tissue quality. Porcine fetal cells can be obtained under controlled conditions after timed pregnancies, allowing for reproducible cell preparations of high viability. Cells can be derived from whole organs of human or animal origin or may be derived from stem cells that have been expanded in culture and differentiated into the cell lineage needed. Human cells would obviously be desirable due to the lower immune barrier and the physiological compatibility, but the issue of obtaining sufficient human cells from
250
Edge
donor or cadaveric sources has precluded their regular use. Stem cells offer great promise but are far from being ready for human use. Issues such as the determination of fate of the cells remain to be worked out, and, in addition, safety issues such as the propensity to proliferate in vivo need to be addressed. However, with the advent of both pluripotential human stem cells (19,20) and stem cells derived from various organs (liver, bone marrow, brain), the potential for use of these cells is promising. Recent studies indicate that stem cells derived from adult tissue have a greater degree of plasticity than previously thought. For example, bone marrow stromal cells can be induced to form cardiomyocytes (21) that may be useful for cell transplantation for heart disease. Mesenchymal stem cells from bone marrow have also been differentiated into adipocytes, chondrocytes, and osteocytes (22), and recent data indicate that similar cells may give rise to neural cells (23). Neural stem cells can undergo a fate change to give rise to hematopoietic tissue (24). Cell types that are immortalized such as those in which the telomerase gene is expressed are under investigation as possible donor cells for transplantation. Animal cells have the advantage of availability in unlimited quantities from any organ. These cells are in use now in human trials as pointed out above. The major issues confronting the expanded therapeutic use of xenogeneic cells are inhibition of the immune response, physiological compatibility, and risk of infection, each of which will be discussed in this chapter. Given the practical and technical constraints of human tissue, porcine cellular transplantation has advanced experimentally and is currently being applied in clinical trials.
TRANSPLANTED CELLS INTEGRATE INTO THE HOST A particularly exciting aspect of this venture in cell-based therapies is the finding that transplanted cells from allo- and xenogeneic sources integrate into the tissue of interest where they carry out all of the physiological functions of normal cells (1,16,25). This physical and physiologic integration is apparent in the regulated responsiveness of the cells to host signaling. This provides an advantage over conventional gene therapy in which a gene or cDNA is introduced into a cell that may lack the cohort of associated molecules that participate in physiological homeostasis. In certain cases of genetically engineered cells, a single gene product is
Current Applications of Cellular Xenografts
251
delivered without the associated array of regulatory molecules provided by an intact cell of the appropriate tissue. For example, production of neurotransmitters is responsive to the synaptic environment and regulated in a complex manner requiring reuptake receptors in addition to degradation and synthesis, and is not readily replaced by a peripheral cell transfected with a gene for production of a given neurotransmitter. Regulation of proteins secreted by the liver can involve multiple transcription factors responsive to metabolic signals such as oncotic pressure in addition to a host of cytokines and feedback regulation on synthesis and degradation. This integrated function is unlikely to be achieved by engineering of a nonliver cell. In treatment of diabetes, the requirement for physiologic regulation is apparent for insulin, which must be regulated to control the wide variations in glucose intake from the diet and glucose utilization. Improper regulation in each of these areas should be overcome by the use of intact physiologically relevant cell types that have the innate regulatory mechanisms of the tissue into which they are being transplanted. One of the most remarkable findings in cell transplantation studies has been that these regulatory mechanisms are conserved across species such that cells transplanted from one species to another are capable of reading and responding to signals in the host environment. Porcine islets are regulated similarly to human islets, in that insulin is secreted in response to glucose. Porcine islets have been used in a clinical trial that showed limited survival and function of the cells (13). Porcine fetal neural cells transplanted across species are regulated in the appropriate manner, and fetal cells recognize signals in the adult brain from dissimilar species that lead them to differentiate and in the case of fetal neural cells form appropriate synapses (26–28).
IMMUNOLOGICAL ISSUES IN ORGAN AND CELL TRANSPLANTATION Immunosuppression continues to be a major problem for transplantation. Whole organ transplantation has evolved to the point that first year survival rates in excess of 80% are obtained for kidney, liver, and heart transplants. However, the standard immunosuppressive regimens employed have significant side effects and are less successful in preventing chronic rejection. Five year survival rates remain below 70% for these same organs. To hone the regimens for preventing graft rejec-
252
Edge
tion, the major current effort is to develop graft-specific immunosuppression that will allow survival of the graft without associated compromise of the immune system and its ability to fight infection and malignancy. Some of the most successful of these regimens prevent the expansion of graft-specific T cells by blocking the costimulatory signals required for T-cell activation. Both CTLA4Ig and anti-CD40 ligand block at this level of the immune response and have proven effective in animal models of transplantation (29). In a promising recent study, the combination of these reagents has permitted survival of renal and heart allografts for up to 100 d (30,31). The discovery of apoptotic pathways mediated by Fas ligand and tumor necrosis factor (TNF) has revealed another way of eliminating T cells. Transplantation of Fas-ligandexpressing tissues, thought initially to be a way of physiologically eliminating T cells (32), has yielded mixed results, as Fas ligand appears to attract neutrophils to the graft and has been reported to decrease graft survival (33–35). However, combining TGF-( with Fas-ligand-expressing tissue may solve this problem by preventing the inflammatory response (36). These and other approaches are thought to induce apoptosis or anergy of graft-specific T cells. Other approaches that are thought to inactivate T cells include the use of altered peptide ligands to induce anergy (37), the use of soluble peptides to interfere with MHC– T-cell-receptor interaction, and the use of antibodies against MHC molecules to inhibit T-cell recognition (38,39). In recent studies, an altered cytokine profile was observed when antibodies to MHC class I were used to inhibit a human anti-porcine response (40). Altered cytokine production by immune cells from a Th1 to a Th2 profile may inhibit graft reactive T cells. The mechanisms of T-cell-mediated rejection of xenografts appear to be similar to those invoked in allogeneic graft rejection. However, organ xenotransplantation is complicated by a number of naturally occurring effector mechanisms that can result in early rejection. Consequently, survival of xenotransplant organs continues to be extremely difficult to attain in animal studies with conventional immunosuppression, and, accordingly, clinical xenotransplantation of organs remains a possibility that has yet to be realized. In addition to T-cell-mediated responses, which can again be direct (if the donor MHC and costimulatory molecules are recognized by host T cells) or indirect (if the recipient processes and presents the xenogeneic antigens), a number of additional pathways are integral to the process of xenograft organ rejection. These
Current Applications of Cellular Xenografts
253
include the reactivity of naturally occurring antibodies with epitopes present on vascular endothelial cells leading to thrombosis and “hyperacute” rejection that takes place within minutes to hours (41–43). If hyperacute rejection is avoided, acute vascular rejection ensues in days to weeks. Human NK cells also recognize porcine cells owing to the lack of negative regulatory sequences in the porcine MHC molecules (44–46). It has been proposed that introduction of immunosuppressive genes in animals might aid in the future application of xenotransplantation. Immunological barriers to xenogeneic cellular transplants, as outlined below, are less of a problem than those confronting xenogeneic organs and we have therefore adopted the use of xenogeneic cells. Cell transplantation decreases the problems with rejection of foreign tissue. For xenotransplantation using porcine cells, a number of factors contribute to the increased resistance to rejection of the cells as compared to organs. These include (i) the ability to transplant cells into immunologically privileged sites (47,48) and (ii) the lack of vasculature in cellular transplants: With vascularized organs, the focus of the initial immune response is the vascular wall where hyperacute rejection occurs in a xenograft (49). Antibodies directed primarily against alpha-linked galactose on endothelial cells bind and initiate this process (42,43,50). In cell transplantation the principal site for xenorejection, the vascular endothelial cell, is not present, and as a result the problems preventing the application of organ xenografting are not an issue. Although natural antibodies and complement can kill cells directly in vitro, a number of studies have shown that xenogeneic cells resist hyperacute rejection (51), and cells may only face hyperacute rejection if endothelial cells are present in the preparation. This is the case, for example, with porcine islets that are subject to killing immediately after isolation when the vasculature is intact but not after several days in culture when endothelial cells are lost and only the endocrine cells survive (52). Acute vascular rejection is also not likely to be a problem for isolated porcine cells. However, more problematic is the delayed T-cell mediated rejection that can occur, and this needs to be addressed in transplantation protocols. If the cell preparation lacks antigen presenting cells and the cells are low in MHC class I and MHC class II as well as adhesion and costimulatory molecules, the cells may stimulate this arm of the immune response less than an organ containing multiple cell types. Thus, the immunosuppressive protocols currently used may be sufficient for xenogeneic cellular transplantation.
254
Edge
LIMITATIONS AND RISKS OF CELL THERAPY Limitations of cell therapy include the need to isolate the cell type of choice. Identification of the optimal site for cell infusion and the technical limitations of transplantation to a given site are other prerequisites that need to be addressed. Issues of optimal dose and compatibility across species also need to be considered. Risks associated with organ transplantation include the transmission of infections, complications of surgery, immunological rejection, and primary nonfunction of a graft. Cell transplantation carries similar risks, and recent attention has focused on the risks associated with the use of animal tissues for transplantation into humans. While the risks of human organ transplantation are well documented, the risks of xenotransplantation of cells are hypothetical. In particular, the risk of transmitting a porcine retrovirus or of generating a new form of a retrovirus by recombination between a porcine and a human endogenous sequence has received recent attention (53–55), but these risks are entirely speculative and the FDA has approved continued clinical trials of xenotransplantation with careful monitoring. Porcine tissues have been used previously in humans without infectious problems: pig skin grafts have been placed on immunosuppressed patients, and porcine insulin produced from porcine pancreas was used for decades without incident. Patients that have received porcine grafts for Parkinson’s disease (26) or for diabetes (56) as well as patients whose blood has been circulated through porcine kidneys for the treatment of kidney failure (57) have recently been tested for the transmission of porcine retroviruses using PCR methods developed for this purpose, and no evidence of transmission has arisen. In addition, pigs have not been reported to harbor prions and would therefore not transmit spongiform encephalopathies.
RECENT ADVANCES IN CELL THERAPY Xenogeneic Cell Based Therapies in the Clinic Currently, we are exploring the potential of porcine fetal neural cell transplantation not only in Parkinson’s disease but also in Huntington’s disease and focal epilepsy. Preliminary data have suggested that the transplanted porcine neurons are effective in alleviating the symptoms of Parkinson’s disease (9). Improvements in the United Parkinson’s Disease Rating Scale were found after transplantation in most patients.
Current Applications of Cellular Xenografts
255
Evidence of porcine neural cell survival was obtained from one of the 12 patients who died of a pulmonary embolus nearly 8 mo after transplantation, affording us the opportunity to examine autopsy tissue (26). Immunohistochemistry and in situ hybridization using porcine specific probes (58) demonstrated that the porcine neural cells had survived in this patient and that dopaminergic neurons had sent out processes into host tissue. In addition, the extent of lymphocyte infiltration and MHC class II expression in this cyclosporin treated patient was minimal. Diabetes has been treated with porcine islets in a series of 10 patients. Porcine C-peptide production was detected in several of these patients and some evidence for porcine islet survival was obtained through biopsies (13).
Experimental Xenogeneic Cell Based Therapies In a study in Watanabe rabbits, which lack functional LDL receptors, we showed that xenotransplanted hepatocytes could lower serum cholesterol by taking up LDL from the serum (1). In addition, xenogeneic hepatocyte transplantation has been used successfully in animal models of liver failure (59–61). Based on our results in animal studies we intend to treat patients with acute and chronic liver failure with porcine hepatocytes. Patients with fulminant hepatic failure have been reported to recover from coma after human hepatocyte transplantation (62), and recent results indicate that these cells may be beneficial for patients with chronic liver disease (63). Porcine hepatocytes are an attractive resource for this therapy because they can be obtained in the quantities required. Xenotransplantation of retinal pigment epithelial (RPE) cells shows promise in the treatment of macular degeneration in which photoreceptor loss is thought to be attributable to dysfunction of this cell layer. In animal models of macular degeneration xenotransplanted cells have been shown to survive in the subretinal space and functional improvement has been observed (64,65). Human clinical trials are underway using human fetal RPE cells for the treatment of this disease (66), and we have transplanted porcine RPE cells into animal models in preparation for clinical application of these cells. Another cell type that will be useful in the treatment of retinal degeneration involving the neural retina will be porcine fetal photoreceptor cells. The advantages of using porcine fetal cells for this clinical application are that the fetal cells of appropriate gestational age are available and that the subretinal space like the central nervous system (CNS) is an immunoprivileged site.
256
Edge
Additional applications for porcine neural cell transplantation are currently being tested in animal models. Fetal porcine spinal cord neurons are being tested in animal models of spinal cord injury. We anticipate that these porcine cells will integrate and restore sensory motor neuron function in these debilitating conditions. Fetal porcine cells from the fetal precursor of the striatum are currently being tested in an animal model of stroke induced by middle cerebral artery occlusion. An important application for fetal porcine cardiomyocytes of skeletal myoblasts is the treatment of ischemic heart disease. In this condition, damage to the myocardium is essentially irreversible owing to the inability of cardiomyocytes to regenerate. Fetal porcine cardiomyocytes have the potential to be transplanted into the areas of ischemic damage and the zone adjacent to the infarct that remains at risk for future damage. This concept has been tested in xenogeneic systems as well as autologous systems that have demonstrated both cell survival and improvement of myocardial function (12,67). Xenogeneic cartilage would be valuable for the repair of articular joint damage. Recent evidence indicates that autologous human chondrocytes can repair cartilage defects (5,6), but this procedure requires a biopsy of articular cartilage from the patient and a second procedure to implant the cells. The application of porcine chondrocytes in humans would allow replacement from a ready supply of cells without two invasive procedures. Treatment of multiple sclerosis with glial cells from xenogeneic sources would provide a practical way to obtain sufficient cells for treatment of the focal lesions found in the myelin sheaths of central and peripheral neurons in this disease. Initial success in animal studies suggest that this approach will provide benefit to patients with this disorder (11). Diffuse alveolar damage leading to adult respiratory distress syndrome is characterized by loss of the alveolar cells lining the bronchi of the lung. Cell replacement therapy in this condition could benefit patients by restoring the lung’s lining and thereby increasing the efficiency of oxygen uptake. Some of the loss of alveolar function in emphysema and bronchitis might also be treated by transplantation of these cells. Loss of inner ear cells is the most commmon cause of deafness (68). Hair cells in the cochlea can be damaged by noise exposure, aminoglycoside antibiotics, and viral infection (69). Cochlear ganglion cells are lost in auditory neuropathy resulting in severe hearing deficits (70). Cell therapy holds great potential for the treatment of these conditions.
Current Applications of Cellular Xenografts
257
In the kidney, glomerular cells might be useful for the treatment of primary glomerular insufficiency. This would include the use of mesangial cells for the treatment of diabetic kidney disease and the use of glomerular epithelial cells for the treatment of glomerular alterations that are characterized by loss of epithelial cells, glomerular basement membrane thickening, and proteinuria. Renal tubular cells might be valuable for repair of the tubules in acute tubular necrosis. Porcine renal tubule cells have recently been shown to provide metabolic function in an extracorporeal renal-assist device (71). Endothelial cells are susceptible to damage in a number of diseases and are not readily able to regenerate by entering mitosis. In saphenous vein grafting, lost endothelial cells might be replaced by transplantation of porcine vascular endothelial cells. Hepatic endothelial cells are the major source of factor VIII and should prove useful in the treatment of hemophilia A. Kidney endothelial cells form a key part of the glomerular filtration apparatus and might be expected to benefit patients with compromised glomerular function due to damaged endothelium from immune complex deposition or autoimmune glomerulonephritis. Porcine bladder and urinary tract cells might be harvested and used for the treatment of incontinence and urinary tract reconstruction. Transplantation of urothelial cells for reconstruction of the bladder wall has recently been shown to have promise for bladder disease (72). This list is certain to expand as this mode of therapy becomes accepted as an alternative to conventional treatments. The availability of porcine cells and their apparent compatibility with human tissues should allow the development of a variety of xenotransplantation strategies.
FUTURE CHALLENGES IN CELL TRANSPLANTATION THERAPY Cell Delivery The mode of delivery for cell transplantation is varied, and it is clear that improvements could be made that would result in better targeting of cells and improved graft survival. The improvements in imaging methods and in radiologic delivery will allow improved delivery of cells to sites of tissue damage. In the CNS improved stereotactic equipment and imaging methods are allowing more precise delivery of cells, and recent studies have shown that minimal disruption of tissue with finer canula tips results in less inflammation and better graft survival. Deliv-
258
Edge
ery of cells via catheters using percutaneous methods is being tested for hepatocyte transplantation and will be tested in delivery of cells to the heart. Vitreoretinal surgery has advanced to a point that delivery of retinal cells to precise layers of the retina is possible. Methods will need to be developed for delivery of cells to the inner ear and to vascular sites for repair of endothelial damage. If cells can be shown to home to sites of origin, cell delivery may be done intravenously. This improvement should lead to ways of enhancing cell survival in the tissue under consideration.
Improved Cell Survival In studies in which engraftment and survival of cells has been carefully measured, cell numbers in a graft have been found to be less than the number of cells infused (73,74). Inefficient cell engraftment can result from cell death or from loss of the cognate substratum for cell attachment resulting in inefficient repopulation of a target tissue. Manipulation of cells to inhibit apoptosis, particularly in the vulnerable phase when the cells have been removed from their normal substrate and have yet to establish themselves on the matrix of the host tissue, may be a means of improving cell survival. A recent study in which mouse hepatocytes transgenic for the anti-apoptotic gene, Bcl-2, were transplanted into anti-Fas-treated recipients showed that these cells had an advantage for liver repopulation (75). Addition of homing molecules by transfection or introduction into the germ line of the porcine donor may aid in this process. Selection of cells that have a better capacity to survive may also increase cell survival rates (74,76). Methods of permitting egress of cells from the vasculature may improve the efficiency of delivery of cells by nonsurgical methods employing percutaneous catheterization to the large vessels.
CONCLUSIONS Cellular transplantation is potentially applicable to a broad range of diseases, which include important public health problems that lack other means of treatment. The use of xenogeneic cells alleviates the problem of human organ availability and provides a solution for a reliable source of cells. The immunological issues facing the transplant scientist in organ transplantation are more severe than those encountered in cell transplantation, and several promising methods of avoiding transplant rejection are in hand. The testing of this technology through the ongoing
Current Applications of Cellular Xenografts
259
clinical trials and the reported success in treating several diseases with cellular therapy is generating increased interest in the efficacy of the proposed approaches. Porcine cellular xenotransplantation is currently being applied to the treatment of human disease.
REFERENCES 1. Gunsalus JR, Brady DA, Coulter SM, Gray BM, Edge ASB. Reduction of serum cholesterol in Watanabe rabbits by xenogeneic hepatocellular transplantation. Nat Med 1997; 3:48–53. 2. Grossman M, Raper SE, Kozarsky K, et al. Successful ex vivo gene therapy directed to liver in a patient with familial hypercholesterolemia. Nat Genet 1994; 6:335–341. 3. Fox IJ, Chowdhury R, Kaufman SS, et al. Treatment of the Crigler-Najjar syndrome Type I with hepatocyte transplantation. N Engl J Med 1998; 338:1422–1426. 4. Li Q, Murphree SS, Willer SS, Bolli R, French BA. Gene therapy with bilirubinUDP-glucuronyltransferase in the Gunn rat model of Crigler-Najjar syndrome type 1. Hum Gen Ther 1998; 9:497–505. 5. Brittberg M, Lindahl A, Nilsson A, Ohlsson C, Isaksson O, Peterson L. Treatment of deep articular cartilage defects in the knee with autologous chondrocyte transplantation. N Engl J Med 1994; 331:889–895. 6. Binette F, McQuaid DP, Haudenschild DR, Yaeger PC, McPherson JM, Tubo R. Expression of a stable articular cartilage phenotype without evidence of hypertrophy by adult human articular chondrocytes in vitro. J Ortop Res 1998; 28:241–251. 7. Lindvall O, Sawle G, Widner H, et al. Evidence for long-term survival and function of dopaminergic grafts in progressive Parkinson’s disease. Ann Neurol 1994; 35:172–180. 8. Lindvall O. Neural transplantation: a hope for patients with Parkinson’s disease. Neuroreport 1997; 8:iii-x. 9. Ellias SA, Palmer EP, Kott HS, et al. Fetal porcine ventral mesencephalic transplantation for Parkinson’s Disease: preliminary results. Mov Disord 1997; 12:839–840. 10. Hillaire MS, Shannon K, Schumacher J, Starr et al. Transplantation of fetal porcine striatal cells in Huntington’s disease: preliminary safety and efficacy resuts. Neurology 1998;50:S10.008. 11. Archer DR, Cudden PA, Lipshitz D, Duncan LD. Myelination of the canine central nervous system by glial cell transplantation: a model for repair of human myelin disease. Nat Med 1997; 3:54–59. 12. Kessler PD, Byrne BJ. Myoblast cell grafting into heart muscle: cellular biology and potential applications. Ann Rev Physiol 1999; 61:219–242. 13. Groth CG, Korsgren O, Tibell A, et al. Transplantation of porcine fetal pancreas to diabetic patients. Lancet 1994; 344:1402–1404. 14. Gibson T. Zoografting - A curious chapter in the history of plastic surgery. Br J Plast Surg 1955; 8:234. 15. Liu X, Federlin KF, Bretzel RG, Hering BJ, Brendel MD. Persistant reversal of diabetes by transplantation of fetal pig proislets into nude mice. Diabetes 1991; 40:858–866. 16. Li H, Ricordi C, Demetris AJ, et al. Mixed xenogeneic chimerism (mouse + rat to mouse) to indice donor-specific tolerance to sequential or simultaneous islet xenografts. Transplantation 1994; 57:592–598.
260
Edge
17. Sachs DH. The pig as a potential xenograft donor. Pathol Biol 1994; 42:217. 18. Fishman J. Miniature swine as organ donors for man: strategies for prevention of xenotransplant-associated infections. Xenotransplantation 1994; 1:47–57. 19. Shamblott MJ, Axelman J, Wang M, et al. Derivation of pluripotential stem cells from cultured human primordial germ cells. Proc Natl Acad Sci USA 1998; 95:13,726–13,731. 20. Thomson JA, Itskovutz-Eldor J, Shapiro SS, et al. Embryonic stem cell lines derived from human blastocysts. Science 1998; 282:1145–1147. 21. Makimo S, Fukuda K, Miyoshi S, et al. Cardiomyocytes can be generated from marrow stromal cells in vitro. J Clin Invest 1999; 103:697–705. 22. Pittenger MF, Mackay AM, Beck SC, et al. Multilineage potential of adult mesenchymal stem cells. Science 1999; 284:143–147. 23. Azizi SA, Stokes D, Augelli BJ, DiGirolamo C, Prockop DJ. Engraftment and migration of human bone marrow stromal cells implanted in the brains of albino rats - similarities to astrocyte grafts. Proc Natl Acad Sci USA 1998; 95:3908–3913. 24. Bjornson CRR, Rietze RL, Reynolds BA, Magli MC, Vescovi AL. Turning brain into blood: a hematopoietic fate adopted by adult neural stem cells in vivo. Science 1999; 283:534–537. 25. Isacson O, Deacon T. Neural transplantation studies reveal the brain’s capacity for continuous reconstruction. TINS 1997; 20:477–482. 26. Deacon T, Schumacher J, Dinsmore J, et al. Histological evidence of fetal pig neural cell survival after transplantation into a patient with Parkinson’s disease. Nat Med 1997; 3:350–353. 27. Galpern WR, Burns LH, Deacon TW, Dinsmore J, Isacson O. Xenotransplantation of porcine fetal ventral mesencephalon in a rat model of Parkinson’s disease: functional recovery and graft morphology. Exp Neurol 1996; 140:1–13. 28. Isacson O, Deacon TW, Pakzaban P, Galpern WR, Dinsmore J, Burns LH. Transplanted xenogeneic neural cells in neurodegenerative disease models exhibit remarkable axonal target specificity and distinct growth patterns of glial axonal fibers. Nat Med 1995; 1:1189–1194. 29. Bluestone JA. New perspectives of CD28-B7-mediated T cell costimulation. Immunity 1995; 2:555–559. 30. Kirk AD, Harlan DM, Armstrong NN, et al. CTLA4-Ig and anti-CD40 ligand prevent renal allograft rejection in primates. Proc Natl Acad Sci USA 1997; 94:8789–8794. 31. Larsen CP, Elwood ET, Alexander DZ, et al. Long-Term Acceptance Of Skin and Cardiac Allografts After Blocking Cd40 and Cd28 Pathways. Nature 1996; 381:434–438. 32. Lau HT, Yu M, Fontana A, Stoeckert CJ. Prevention of islet allograft rejection with engineered myoblasts expressing FasL in mice. Science 1996; 273:109–112. 33. Allison J, Georgiou HM, Strasser A, Vaux DL. Transgenic expression of CD95 ligand on islet beta cells induces a granulocytic infiltration but does not confer immune privilege upon islet allografts. Proc Natl Acad Sci USA 1997; 94:3943–3947. 34. Kang SM, Schneider DB, Lin ZH, et al. FAS ligand expression in Islets of Langerhans does not confer immune privilege and instead targets them for rapid destruction. Nat Med 1997; 3:738–743.
Current Applications of Cellular Xenografts
261
35. Muruve DA, Nicolson AG, Manfro RC, Strom TB, Sukhatme VP, Libermann TA. Adenovirus-mediated expression of FAS ligand induces hepatic apoptosis after systemic administration and apoptosis of ex-vivo infected pancreatic islet allografts and isografts. Hum Gene Ther 1997; 8:955–963. 36. Chen JJ, Sun Y, Nabel GJ. Regulation of the proinflammatory effects of Fas Ligand (CD95L). Science 1998; 282:1714–1717. 37. Sloan-Lancaster J, Evavold BD, Allen PM. Induction of T-cell anergy by altered T-cell-receptor lignad on live antigen-presenting cells. Nature 1993; 363:156–159. 38. Smith DM, Bluestone JA, Jeyarajah DR, et al. Inhibition of T cell activation by a monoclonal antibody reactive against the a3 domain of human MHC class I molecules. J Immunol 1994; 153:1054–1063. 39. Stock PG, Ascher NL, Bumgardner GL, Field MJ, Sutherland DER. Modulation of MHC class I antigen decreases pancreatic islet immunogenicity. J Surg Res 1989; 46:317–321. 40. DerSimonian H, Yatko C, Kohnson E, Rodrigue-Way A, Edge ASB. Human antiporcine T cell response: the role of anti-class I antibody masking in graft tolerance to cell transplantation. J Immunol 1999; 162:6993–7001. 41. Bach FH, Winkler H, Ferran C, Hancock WW, Robson SC. Delayed xenograft rejection. Immunol Today 1996; 17:379–384. 42. Galili U. Interaction of the natural anti-Gal antibody with alpha-galactosyl epitopes: a major obstacle for xenotransplantation in humans. Immunol Today 1993; 14:480–482. 43. LaVecchio JL, XDunne SC, Edge ASB. Enzymatic removal of alpha-galactosyl epitopes from porcine endothelial cells diminishes the cytotoxic effect of natural antibodies. Transplantation 1995; 60:841–847. 44. Seebach JD, Comrack C, Germana S, LeGuern C, Sachs DH, DerSimonian H. HLACw3 expression on porcine endothelial cells protects against xenogeneic cytotoxicity mediated by a subset of human NK cells. J Immunol 1997; 159:3655–3661. 45. Donnelly CE, Yatko C, Johnson EW, Edge ASB. Human natural killer cells account for non-MHC class I restricted cytolysis of porcine cells. Cell Immunol 1997; 175:171–178. 46. Sullivan JA, Oettinger HF, Sachs DH, Edge ASB. Analysis of ploymorphism in porcine MHC class I genes. Alterations in signals recognized by human cytotoxic lymphocytes. J Immunol 1997; 159:2318–2326. 47. Bellgrau D, Gold D, Selawry H, Moore J, Franzusoff A, Duke RC. A role for CD95 ligand in preventing graft rejection. Nature 1995; 377:630–632. 48. Griffith TS, Brunner T, Fletcher SM, Green DR, Ferguson TA. Fas ligand-induced apoptosis as a mechanism of immune privilege [see comments]. Science 1995; 270:1189–1192. 49. Bach FH, Robson SC, Winkler H, et al. Barriers to xenotransplantation. Nat Med 1995; 1:869–873. 50. Neethling FA, Koren E, Ye Y, et al. Protection of pig kidney (PK15) cells from the cytotoxic effect of anti-pig antibodies by a-galactosyl oligosaccharides. Transplantation 1994; 57:959–963. 51. Mirenda V, Le Mauff B, Cassard A, et al. Intact pig pancreatic islet function in the presence of human xenoreactive natural antibody binding and complement activation. Transplantation 1997; 63:1452–1462.
262
Edge
52. Markmann JF, Tomaszewski J, Posselt AM, et al. The effect of islet cell culture in vitro at 24°C on graft survival and MHC antigen expression. Transplantation 1990; 49:272–277. 53. Bach FH, Fineberg HV. Call for moratorium on xenotransplants. Nature 1998; 391:326. 54. Patience C, Takeuchi Y, Weiss RA. Infection of human cells by an endogenous retrovirus of pigs. Nat Med 1997; 3:282–286. 55. Martin U, Keisig V, Blusch JH, et al. Expression of pig endogenous retrovirus by primary porcine endothelial cells and infection of human cells. Lancet 1998; 352:692–694. 56. Heheine W, Tibell A, Switzer WM, et al. No evidence of infection with porcine endogenous retrovirus in recipients of porcine islet-cell xenografts. Lancet 1998; 352:695–699. 57. Patience C, Patton GS, Takeuchi Y, et al. No evidence of pig DNA or retroviral infection in patients with short-term extracorporeal connection to pig kidneys. Lancet 1998; 352:699–701. 58. Oettinger HF, Rodrigue-Way A, Bousquet JJ, Edge ASB. Porcine repeat element DNA: in situ detection of xenotransplanted cells. Cell Transplant 1995; 4:253–256. 59. Gupta S, Rajvanshi P, Lee C. Integration of transplanted hepatocytes into host liver plates demonstrated with dipeptidyl peptidase IV-deficient rats. Proc Natl Acad Sci USA 1995; 92:5860–5864. 60. Stefan AM, Coulter S, Gray B, et al. Xenogeneic transplantation of porcine hepatocytes into the CCl4 rat model. Cell Transplant 1999; 8:649–659. 61. Gupta S, Gorla GR, Irani AN. Hepatocyte transplantation: emerging insights into mechanisms of liver repopulation and their relevance to potential therapies. J Hepatol 1999; 30:162–170. 62. Strom SC, Fisher RA, Thompson MT, et al. Hepatocyte transplantation as a bridge to orthotopic liver transplantation in terminal liver failure. Transplantation 1997; 63:559–569. 63. Bilir BM, Guenette D, Ostrowska A, et al. Percutaneous hepatocyte transplantation (PHT) in liver failure. Hepatology 1997; 26:252A. 64. Lane C, Boulton M, Marshall J. Transplantation of retinal pigment epithelium using a pars plana approach. Eye 1989; 3:27–32. 65. Gouras P, Algvere P. Retinal cell transplantation in the macula: new techniques. Vision Res 1996; 36:4121–4125. 66. Algvere PV, Berglin L, Gouras P, Sheng Y. Transplantation of fetal retinal pigment epithelium in age-related macular degeneration with subfoveal neovascularization. Graefe’s Arch Clin Exp Ophthalmol 1994; 232:707–716. 67. Li R-K, Jia Z-Q, Weisel RD, et al. Cardiomyocyte transplantation improves heart function. Ann Thorac Surg 1996; 62:654–661. 68. Steel KP. Progress in progressive hearing loss. Science 1998; 279:1870– 1871. 69. Ruth RA. Evaluation of sensorineural hearing loss. Comp Ther 1997; 23:742–749. 70. Starr A, Picton TW, Sininger Y, Hood LJ, Berlin CI. Auditory Neuropathy Brain 1996; 119:741–743. 71. Humes HD, Buffington DA, MacKay SM, Funke AJ, Weitzel WF. Replacement of renal function in uremic animals with a tissue-engineered kidney. Nat Biotechnol 1999; 17:451–455.
Current Applications of Cellular Xenografts
263
72. Oberpenning F, Meng J, Yoo JJ, Atala A. Nat Biotechnol 1999; 17:149–155. 73. Beauchamp JR, Morgan JE, Pagel CN, Partridge TA. Dynamics of myoblast transplantation reveal a discrete minority of precursors with stem cell-like properties as the myogenic source. J Cell Biol 1999; 144:1113–1121. 74. Qu Z, Balkir L, JDeutekom CTV, Robbins PD, Pruchnic R, Huard J. Development of approaches to improve cell survival in myoblast transfer therapy. J Cell Biol 1998; 142:1257–1267. 75. Mignon A, Guidotti JE, Mitchell C, et al. Selective repopulation of normal mouse liver by Fas/CD95-resistant hepatocytes. Nat Med 1998; 10:1185–1188. 76. Gussoni E, Blau HM, Kunkel LM. The fate of individual myoblasts after transplantation into muscles of DMD patients. Nat Med 1997; 3:970–977.
vi
Index
265
Index anti-HLA, 162 anti-human C5, 161 anti-Pk antibodies, 77 antigen-presenting cell (APC), 5, 64–65, 71–75, 177–179, 180, 184, 186 ATP diphosphohydrolase, 220, 233, 237
A A and B blood groups, 10, 24, 58, 66, 74–75, 120, 124, 156, 162, 201 ABO-incompatible kidney, 23 accommodation, 2, 24 acute vascular rejection (AVR), 2, 17, 20–22, 25, 50, 148, 217–222, 224, 226, 229, 231–232, 234–238 ADP, 233 affinity, 88, 90 AIDS, 1 Alexandre, G., 26 _-Gal, 57–58, 60–62, 66–67, 200, 202, 208 _-galactose, 10 _-galactosidase, 129 _1,2fucosyltransferase, 127–128, 222 _1,3glactosyltransferase (_1,3GT), 60–61, 119–120, 123–125, 127, 200, 220 _1,3glactosyltransferase gene, 4, 10, 13, 125–126 _2,3sialytransferase, 128 anti-C5 antibodies, 16, 159 anti-Gal_1-3Gal (anti_Gal) 11, 58–59, 66, 74–77, 79, 82, 90, 92, 97, 104, 108, 109–110, 112, 123, 126, 204, 229 antibody-dependent cell cytotoxicity (ADCC), 62, 65, 67, 200, 203, 205–206
B B cell, 49, 105, 108, 111, 140, 145, 179, 202, 229 B1-B cell, 107, 112 B2-B cell, 107 B7, 175, 178 baboon, 10 baboon-to-human, 1 Bach, F., 30, 49 Bb, 222 Bcl-2, 258 bcl-xl, 177 binding constant, 88, 90 binding kinetics, 96 biplaner irreversible binding, 95 blood group, 125 blood vessels, 8, 17–18 bone marrow, 1, 250 bone marrow transplantation (BMT), 226, 228 bovine, 182 C C1 inhibitor (C1 inh), 14, 143, 159–160 C1q, 48, 97, 140, 151 265
266
C2, 14, 48, 140 C3, 16, 140, 142, 149, 161–162, 218 C3a, 16, 142, 146–147, 149, 151–153, 158, 221–222 C3aR, 140 C3b, 16, 142, 144–146, 161, 221 C3bi, 145, 149–152, 160 C3 convertase, 15, 24, 141, 143–144 C3d, 221 C4, 14, 16, 48, 140 C4b, 144–146 C4b2a3b, 142 C4-binding protein, 140, 143 C5, 140–142, 149, 162 C5a, 16, 18, 146, 149–153 C5aR, 140 C5b, 142–143 C5b6, 18 C5b7, 143, 153 C5b8, 143–144, 153 C5b9, 152–153, 155 C5b67, 18 C5 convertase, 133, 143 C6, 13, 21, 140, 143, 235 C7, 140, 143 C8, 140 C9, 140, 143–144, 160 CAB-2, 158–160 Calne, R., 46 cAMP, 18, 19, 236 cartilage, 256 CD2, 176, 181, 185 CD3, 173–174, 176–177, 186 CD4, 27, 29, 175, 183, 186 CD5, 105, 107, 201 CD8, 27, 175 CD11b, 161 CD16+, 205
Index
CD19, 145 CD25, 186 CD28, 175–177, 178, 180, 184, 190 CD39, 233, 237 CD40, 112, 174, 177, 179–180, 252 CD40L, 107, 112, 187 CD46, 219 CD55, 217, 222, 224 CD56+, 205 CD59, 6, 15–16, 140, 144, 155–157, 159–160, 163, 217, 219, 222, 224 CD80, 176, 177–179, 184 CD86, 176, 177–179, 184, 188 CD154, 174, 177, 179–180 CDR3, 109, 190 cell-mediated cytotoxicity, 54 cell therapy, 247 cell transplantation, 247, 254 cellular rejection, 2, 4 chimpanzee-to-human, 46 chronic rejection, 2, 29 53cGMP, 236 CNS, 257 cobra venom factor (CVF), 49, 147–149, 159, 162, 200, 219, 231, 236–237 complement (C), 9, 14, 21, 139–140, 145–146, 202, 216, 224, 230, 232–233, 238 complement activation, 143, 155 complement receptor, 143 complement regulatory proteins, 217 Compstatin, 159 concordant, 8 COS cells, 121, 127, 129, 233 Coxsackie B viruses, 157 CR1, 140, 144–146
Index
CR2, 140, 145 CR3, 140 Crigler–Najjar, 247 CTL, 180, 183 CTLA4-Ig, 187–189, 252 cytotoxic T-lymphocyte antigen (CTLA-4), 178, 180, 184, 190 D D’Apice, A., 222 decay accelerating factor (DAF), 6, 15, 140, 143–144, 155–157, 217 delayed xenograft rejection (DXR), 148, 199, 204, 217–220, 222, 224, 226, 229, 231–232, 234, 236–238 diabetes, 248 discordant, 8, 199 disseminated intravascular coagulation (DIC), 220–224 DNA, 79 dog, 182 Dorling, A., 28 E E-selectin, 181 EBV, 109 Echo viruses, 157 ELISA, 80, 81 endothelial cell (EC or VEC), 8, 18, 22, 24, 50–52, 199, 203, 205, 207–208, 216, 218, 225, 228, 230–234 endothelial cell activation, 218 endothelium, 17, 23, 119, 181, 216
267
enthalpy, 91–93 entropy, 90, 92–93 enzyme-linked immunosorbant assay, 78 Epstein Barr virus (EBV), 59 equilibrium, 94, 97 F Factor I, 160 Factor V, 232 Factor VIII, 257 Factor IX, 248 Factor D, 140 Factor H, 140, 143 familial hypercholesterolemia, 247 Fas ligand, 190, 208, 252 Fc receptors, 23, 160 FcgRIII, 200, 206–208 Food and Drug Administration (FDA), 254 fetal cells, 255 fibrinogen receptor, 220 fibroblasts, 175 Frank, 16 free energy, 90 functional avidity, 89–90, 94 G Gal_1-3Gal 4, 6, 10–13, 25–26, 29, 57–58, 63, 67, 74–82, 94–95, 104–107, 119–123, 125, 127–131, 190, 200, 217, 220 Galili, U., 10, 54, 120 gamma globulin, 49 gap, 19 genetic engineering, 1 Gewurz, H., 14, 147 glial cells, 256 glycolipid, 74
268
glycosyltransferase, 122 gonadotrophin releasing hormone (GnRH), 189 GPIIbIIIa, 236–237 guinea pig, 5, 6 guinea pig-to-rat, 26, 148 H H transferase, 25, 49, 156 hair cells, 256 Hammer, C., 30 hamster, 5 hamster-to-rat, 139 heparan sulfate (HS), 18, 217, 230 heparin, 159 hepatocytes, 2, 248 HLA, 82, 183, 186 HLA antibodies, 24 human, 182, 220, 248, 249, 254 Huntington’s disease, 254 hyperacute rejection (HAR), 2, 5, 8–11, 16–17, 20, 46–47, 52, 119–122, 126, 139, 147–148, 150, 155, 163, 187, 189–199, 202, 204, 208, 216–217, 220, 229, 234, 236 hyper-IgM syndrome, 112 I iC3b, 221 ICAM-1, 154, 174, 176, 181, 183, 185, 188, 189, 207, 219, 225 IFN-a, 207 Ig variable (V) genes, 108 IgG, 11, 52, 54, 59, 62–63, 77, 80, 82, 96, 98, 109–110, 112, 119–120, 130, 141, 148,
Index
150, 200–202, 204–208, 220 IgG1, 82, 202 IgG2, 82, 202 IgG3, 202 IgG4, 202 IgM, 11, 13–14, 52, 58–59, 77, 82, 94–95, 98, 105, 112, 119–120, 141, 145, 147–148, 150, 199–202, 219–220, 225 IL-1, 24 IL-1_, 154 IL-2, 176–177, 183, 184 IL-3, 184 IL-8, 154, 208, 219, 225 integrins, 121 intercellular gaps, 18 intrabody, 190 intravenous IgG (IVIG), 159–160 islets of Langerhans, 26, 200, 208 K kidney, 222–223 kinetics, 96–97 knock-out pigs, 13 knocked-out, 25 L L-selectin, 181 LDL, 255 LFA-1, 174, 176, 181, 183, 185, 187 LFA-3, 176, 181, 185, 188, 189 LFA-4, 176 Liver, 247 LPS, 225 M MAC 142–144, 146, 149–150,
Index
153–154, 161–163 macrophages, 8, 23, 53, 207 MASP-1, 140, 142 MASP-2, 140, 142 MBL, 140, 142 MCP, 140, 144, 155–157 MCP-1, 154, 219 melibiose (Gal_1-Gal), 58 membrane-associated regulators of C (RCA), 217 membrane attack complex, 22, 141 membrane co-factor protein (MCP; CD46), 15 mesenchymal stem cells, 250 MHC, 3, 5, 27–29, 31, 57, 111, 173–175, 179, 229, 252–253 MHC class I, 17, 23, 52, 111 MHC class II, 52, 173, 177, 193, 253 mixed lymphocyte reaction (MLR), 177 Miyagawa, S., 14–15 monocytes, 209 Morgan, P., 15 mouse, 125 mouse-to-rabbit, 148 multiple sclerosis, 248 murine, 182 myoblasts, 175, 256 myocytes, 51, 52 N Najarian, J., 10, 14, 46 natural antibodies, 9, 65 natural killer cells (NK), 17, 23, 62, 130, 148, 203–209, 217, 253 natural antibodies, 200, 216 neovascularization, 3
269
neural cell, 248, 256 New world (NW) monkey, 5, 10–11 nuclear factor-kappa B (NF-gB), 154, 177, 207, 225 O Old world (OW) monkey, 4–5, 60–61, 65, 120, 123, 125, 200, 220 P P-selectin, 19, 181, 220, 233, 236 PAI-1, 232 PAF, 220, 236 Parkinson’s disease, 248, 254 PGE2, 154 pig, 5, 125, 186, 189, 202, 217, 223, 249 pig-to-baboon, 79, 221 pig-to-human, 131 pig-to-primate, 17, 47-8, 53–54, 148 pig-to-primate xenografts, 23 pig-to-rat, 14, 47, 160 pig-to-rhesus monkey, 79 plasminogen activator inhibitor type 1, 22 platelet, 8–9, 22, 51, 225, 232, 236 Platt, J., 47–49, 80 polymorphonuclear (PMN), 52, 204, 207–209 polyreactive antibodies, 9, 79–80, 104 porcine, 249, 253, 254–256 porcine endogenous retrovirus (PERV), 26, 30 primary non-function, 2, 8, 27 primate, 249 protein C, 6, 231 prothrombinase, 22
270
pseudoequilibrium 90 R rabbit-to-newborn pig, 148 rat, 6, 125, 182 rat-to-cynomolgous monkey, 148 rat-to-fetal sheep, 148 RCA, 219 Reemtsma, K., 45 renal tubule cells, 257 retinal pigment epithelial (RPE), 255 S Saadi, S., 18, 225 Sachs, D., 11, 221, 226 Sandrin, M., 10–11 serotonin, 233 sFv, 190 soluble complement receptor-1 (sCR1), 200, 219 soluble human CRI-1 (sCr-1), 49, 158–160, 221 Springer, G., 110 Starzl, T., 45 stem cells, 250 sulfatide, 77 Sykes, M., 226 T T cell, 4, 8, 28, 30, 53, 57, 64, 108, 173–175, 180, 181, 184, 186, 190, 200, 207–208, 252 T cell receptor (TCR), 173–174, 176 thrombin, 6, 223–232 thrombotic thrombocytopenic purpura (TTP), 226 thrombocytopenia, 221
Index
thrombomodulin (TM), 6, 22, 217, 231–232, 235 thyroglobulin, 79–80 tissue factor (TF), 219, 221, 231–233 tissue factor pathway inhibitor (TFPI), 220, 231–232 TNF-_, 185, 207 tumor necrosis factor (TNF), 225 TXA2, 154 V V genes, 108 valency, 96 vascular ATP diphosphohydrolase (ATPDase/CD39), 217 VCAM, 187–188, 189 VCAM-1, 154, 174, 176, 181, 185, 207, 219 VEC (endothelial cells), 199, 203, 205, 207–208 VH3 genes, 110 VLA4, 174 von Willebrand factor (vWF), 12, 19, 154, 217 vWF, 226, 228, 233, 236 W White, D., 221 X xenoantibodies, 219 xenogeneic natural antibodies (XNA), 199–201, 203, 207–208, 216–217, 221, 224–225, 229, 232, 238 xenoreactive natural antibodies, 9–10, 20, 49, 73, 79, 81, 87–88, 90, 94, 97, 103, 119
XENOTRANSPLANTATION BASIC RESEARCH AND CLINICAL APPLICATIONS Edited by
JEFFREY L. PLATT, MD Transplantation Biology, Mayo Clinic, Rochester, MN
Xenotransplantation might well provide a revolutionary way of augmenting the function of diseased tissues or replacing organs. This possibility has been advanced by newly acquired understanding of the biological and immunological obstacles to conducting transplantation between species. In Xenotransplantation: Basic Research and Clinical Applications, internationally recognized scientists, clinicians, and technologists review and explain the fundamental molecular and cellular biology that has been applied to the emerging field of transplant immunology and xenotransplantation and what impact these advances might optimally have on medicine and science. The authoritative experts writing here—many of whom made the basic discoveries underlying the recent advances—examine the biological and immunological hurdles to xenotransplantation, illuminating how the immune system interacts with the xenograft, and laying a practical foundation for the use of genetic engineering and animal transplants in the treatment of human disease. Comprehensive and authoritative, Xenotransplantation: Basic Research and Clinical Applications provides basic and clinical investigators, as well as transplant surgeons, with today’s most thorough and up-to-date compendium of vital information on all the scientific and technological aspects of xenotransplantation. FEATURES 䉴 Cutting-edge review of the basic and applied aspects of xenotransplantation 䉴 Authoritative contributors who have made the basic discoveries underlying recent advances
䉴 Comprehensive treatment of how the immune system interacts with the xenograft
CONTENTS Molecular and Cellular Hurdles to Xenotransplantation. Pathological Responses to Xenotransplantation. Natural Xenoreactive Antibodies. Specificity of Xenoreactive Natural Antibodies. Biophyscial Properties of Xenoreactive Natural Antibodies. The Origin of Xenoreactive Natural Antibodies. Synthesis of Carbohydrate Antigens Recognized by Xenoreactive
Antibodies. The Complement Barrier to Xenotransplantation. Defects and Amplification of Costimulation Across the Species. Antibody-Dependent Effects on Cellular Immunity. Disordered Regulation of Coagulation and Platelet Activation in Xenotransplantation. Current Applications of Cellular Xenografts. Index.
90000
XENOTRANSPLANTATION BASIC RESEARCH AND CLINICAL APPLICATIONS ISBN: 0-89603-674-X humanapress.com
9 780896 036741
