The Biomaterials Silver Jubilee Compendium
The Best Papers Published in
BIOMATERIALS
1980-2004
Elsevier Internet Homepage- http://www.elsevier.com Consult the Elsevier homepage for full catalogue information on all books, major reference works, journals, electronic products and services.
All Elsevier journals are available online via ScienceDirect: www.sciencedirect.com
To contact the Publisher Elsevier welcomes enquiries concerning publishing proposals" books, journal special issues, conference proceedings, etc. All formats and media can be considered. Should you have a publishing proposal you wish to discuss, please contact, without obligation, the publisher responsible for Elsevier's materials and engineering programme:
Jonathan Agbenyega Publisher Elsevier Ltd The Boulevard, Langford Lane Kidlington, Oxford OX5 1GB, UK
Phone: Fax: E-mail:
+44 1865 843000 +44 1865 843987
[email protected]
General enquiries, including placing orders, should be directed to Elsevier's Regional Sales Offices - please access the Elsevier homepage or full contact details (homepage details at the top of this page).
The Biomaterials Silver Jubilee Compendium
The Best Papers Published in
BIOMATERIALS
1980-2004
Edited by D.F. Williams
2006
iiii!!i!!!iiiilDiii!!i!iiiiiiii!!!!!!!!7~, :~:i i~l ,~':!iii!!!!!!!iii!ii
Amsterdam Paris
. Boston
. San Diego
. Heidelberg . San Francisco
. London . Singapore
' New York ' Sydney
. Oxford . Tokyo
Elsevier Ltd is an imprint of Elsevier with offices at: Linacre House, Jordan Hill, Oxford OX2 8DP, UK The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, UK 84 Theobald's Road, London WC1X 8RR, UK Radarweg 29, PO Box 211, 1000 AE Amsterdam, The Netherlands 30 Corporate Drive, Suite 400, Burlington, MA 01803, USA 525 B Street, Suite 1900, San Diego, CA 92101-4495, USA First edition 2006 Copyright 9 Elsevier Ltd. All rights reserved No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means electronic, mechanical, photocopying, recording or otherwise without the prior written permission of the publisher Permissions may be sought directly from Elsevier's Science & Technology Rights Department in Oxford, UK: phone (+44) (0) 1865 843830; fax (+44) (0) 1865 853333; email:
[email protected]. Alternatively you can submit your request online by visiting the Elsevier web site at http://elsevier.com/locate/permissions, and selecting Obtaining permission to use Elsevier material Notice No responsibility is assumed by the publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein. Because of rapid advances in the medical sciences, in particular, independent verification of diagnoses and drug dosages should be made
British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloging-in-Publication Data A catalog record for this book is available from the Library of Congress For information on all biomaterials related publications visit our web site at boo ks. e lsevi er. corn Printed and bound in Great Britain 06 07 08 09 10 l0 9 8 7 6 5 4 3 2 1 ISBN-13:978-0-08-045154-1 ISBN-10:0-08-045154-3
Working together to grow libraries in developing countries www-elsevier.c~ I www.bookaid.org t w~sw.sabre.org
The Biomaterials Silver Jubilee Compendium
Table of C o n t e n t s Title, Author(s) and Reference
Page No.
Preface D. F. Williams Controlled release of biologically active compounds from bioerodible polymers J. Heller Biomaterials 1980 Jan; volume 1 issue 1:pp51-57 The response to the intramuscular implantation of pure metals A. McNamara, D.F. Williams Biomaterials 1981 Jan; volume 2 issue 1:pp33-40 Osseointegrated titanium fixtures in the treatment of edentulousness 17 P.I. Branemark, R. Adell, T. Albrektsson, U. Lekholm, S. Lundkvist, B. Rockler Biomaterials 1983 Jan;volume 4 issue 1."pp25-28 B iomaterial biocompatibility and the macrophage J.M. Anderson, K.M. Miller Biomaterials 1984 Jan," volume 5 issue 1."pp5-10
21
Systemic effects of biomaterials J. Black Biomaterials 1984 Jan; volume 5 issue 1: ppl 1-18
27
The in vitro response of osteoblasts to bioactive glass T. Matsuda, J.E. Davies Biomaterials 1987 Jul; volume 8 issue 4:pp275-284
35
Activation of the complement system at the interface between blood and artificial surfaces M.D. Kazatchkine, M.P. Carreno Biomaterials 1988 Jan; volume 9 issue 1:pp30-35
45
Dynamic and equilibrium swelling behaviour of pH-sensitive hydrogels containing 2-hydroxyethyl methacrylate L. Brannon-Peppas, N.A. Peppas Biomaterials 1990 Nov; volume 11 issue 9: pp635-644.
51
Macroencapsulation of dopamine-secreting cells by coextrusion with an organic polymer solution P. Aebischer, L. Wahlberg, P.A. Tresco, S.R. Winn. Biomaterials 1991 Jan; volume 12 issue 1" pp50-56
61
Interaction between phospholipids and biocompatible polymers containing a phosphorylcholine moiety M. Kojima, K. Ishihara, A. Watanabe, N. Nakabayashi Biomaterials 1991 Mar; volume 12 issue 2:pp121-124
69
The Biomaterials Silver Jubilee Compendium
vi
Quantitative assessment of the tissue response to implanted biomaterials D.G.Vince, J.A. Hunt, D.F. Williams Biomaterials 1991 Oct; volume 12 issue 8" pp731-736
73
Immune response in biocompatibility A. Remes, D.F. Williams Biomaterials 1992;volume 13 issue 11" pp731-743
79
Laminated three-dimensional biodegradable foams for use in tissue engineering A.G. Mikos, G. Sarakinos, S.M. Leite, J.P. Vacanti, R. Langer Biomaterials 1993 Apr; volume 14 issue 5" pp323-330
93
Late degradation tissue response to poly(L-lactide) bone plates and screws J.E. Bergsma, W.C. de Bruijn, F.R. Rozema, R.R. Bos, G. Boering Biomaterials 1995 Jan; volume 16 issue 1" pp25-31
101
Mechanism of cell detachment from temperature-modulated, hydrophilichydrophobic polymer surfaces T. Okano, N. Yamada, M. Okuhara, H. Sakai, Y. Sakurai Biomaterials 1995 Mar; volume 16 issue 4" pp297-303
109
Mechanisms of polymer degradation and erosion A. Gopferich Biomaterials 1996 Jan; volume 17 issue 2" pp 103-114
117
Stabilized polyglycolic acid fibre-based tubes for tissue engineering 129 D.J. Mooney, C.L. Mazzoni, C. Breuer, K. McNamara, D. Hem, J.P. Vacanti, et al Biomaterials 1996 Jan; volume 17 issue 2" pp 115-124 Poly(alpha-hydroxy acids): carriers for bone morphogenetic proteins J.O. Hollinger, K. Leong Biomaterials 1996 Jan; volume 17 issue 2:pp187-194
139
Response ofMG63 osteoblast-like cells to titanium and titanium alloy is 147 dependent on surface roughness and composition J. Lincks, B.D. Boyan, C.R. Blanchard, C.H. Lohmann, Y. Liu, D.L. Cochran, et al. Biomaterials 1998 Dec; volume 19 issue 23" pp2219-2232 Patterning proteins and cells using soft lithography R.S. Kane, S. Takayama, E. Ostuni, D.E. Ingber, G.M. Whitesides. Biomaterials 1999 Dec;volume 20 issue 23-24" pp2363-2376
161
Scaffolds in tissue engineering bone and cartilage D.W. Hutmacher Biomaterials 2000 Dec; volume 21 issue 24:pp2529-2543
175
The Biomaterials Silver Jubilee Compendium
vii
Topographical control of human neutrophil motility on micropatterned materials 191 with various surface chemistry. J. Tan, W.M. Saltzman. Biomaterials 2002 Aug; volume 23 issue 15" pp3215-3225. Photopolymerized hyaluronic acid-based hydrogels and interpenetrating networks Y.D. Park, N. Tirelli, J.A. Hubbell Biomaterials 2003 Mar; volume 24 issue 6" pp893-900
203
Cell sheet engineering for myocardial tissue reconstruction T. Shimizu, M. Yamato, A. Kikuchi, T. Okano Biomaterials 2003 Jun; volume 24 issue 13" pp2309-2316
211
Biomaterial-associated thrombosis: roles of coagulation factors, complement, platelets and leukocytes M.B. Gorbet, M.V. Sefton Biomaterials 2004 Nov; volume 25 issue 26" pp5681-5703
219
Author Index
243
The Biomaterials Silver Jubilee Compendium
viii
This Page Intentionally Left Blank
/x
The Biomaterials Silver Jubilee Compendium
Preface The journal Biomaterials was launched in 1980. The subject of biomaterials science was then in its infancy, being largely confined to the study of the characteristics of materials used for medical devices. In reality, few of those materials had ever been developed for this specific use and instead, were taken from other industrial applications, for example in aerospace, nuclear engineering or chemical processing, and experimented with in surgical or medical procedures. The science was largely observational, as the performance of these materials in their new surroundings was evaluated by combinations of physical, chemical, engineering, biological, pathological and clinical techniques. Over the ensuing decade, the subject evolved, as more became known about their performance and especially about the mechanisms of the interactions between the materials and the tissues that underpin the performance. This branch of biomaterials science has become associated with the term biocompatibility, a field that has been the driving force for the subject. With greater knowledge about these interactions, old serendipitous biomaterials were discarded, and new, intentionally designed, or at least modified, materials, introduced. Moreover, these materials started to find applications in related areas, and medical devices were no longer the sole home for biomaterials, as applications in pharmaceutical technology through drug and gene delivery, regenerative medicine and tissue engineering, and biotechnology have emerged and developed. Twenty-five years on, we can truly say that biomaterials science has matured at an incredible rate and now represents a formidable sector that bridges the materials sciences, advanced medical therapies, and molecular and cell sciences. This development could not have been achieved without high quality scientific journals, including those that represent the main parent disciplines and the interdisciplinary field of biomaterials science itself. Although by no means alone, the journal Biomaterials has taken centre stage here and, at the time of its silver jubilee in 2004 is widely considered to be the premier journal in this field. In order to celebrate 25 years of publishing biomaterials science, Elsevier decided to confer two awards, the Elsevier Biomaterials Gold Medal and the Elsevier Biomaterials Silver Medal. A panel of judges was established in 2005 in order to select the recipients. The Elsevier Biomaterials Gold Medal was awarded to the person judged to have made the most significant contribution to the subject of biomaterials science during
The Biomaterials Silver Jubilee Compendium
the 25 years from 1980 to 2004, irrespective of where the work was published. Elsevier were delighted to announce at the European Society for Biomaterials Annual Meeting in Sorrento, Italy, in September 2005, that the winner of the Gold Medal was Professor James Anderson, of Case Western Reserve University, Cleveland, Ohio, USA. The medal was presented at the Annual Meeting of the Tissue Engineering Society International in Shanghai, in October 2005. The Elsevier Biomaterials Silver Medal was awarded for the most significant paper published in the journal Biomaterials during the first 25 years. Over 60 papers were nominated and the judges had a very difficult time in making the selection since most of the world's leading biomaterials scientists were represented in the nomination list. The subject matter covered much of the seminal research that has set the foundation for the high quality science that is undertaken today and which will embrace the future. This Silver Jubilee Compendium consists of reprinted versions of the top 25 of these papers, arranged chronologically. The current Editor-in-Chief is both appreciative of and humbled by the decision of the panel of judges to select one of his earliest papers, with graduate student Anne McNamara, as the leading paper and recipient of the Silver Award. This Compendium is published as a landmark in biomaterials science and it is to be hoped that it will serve as a stimulus to young biomaterials scientists of the early twenty-first century for their pioneering work of the future. I have been proud to serve as Editor-in-Chief of the journal during this exciting period of its development. I place on record my thanks to previous editors of the journal, listed on a separate page, to editorial staff within Elsevier and their predecessor publishers during this 25 years and to colleagues who have served as Editorial Board members, referees and authors. I would particularly wish to thank Amanda Weaver, Publisher of the journal in Elsevier who skilfully steered this process of the medals through the company, to the panel of judges who had to work very hard on this process, and to Peggy O'Donnell, Managing Editor of the journal, who carried the full logistics burden of the medals procedure.
Professor David Williams Editor-in-Chief Liverpool
History of Editorial Appointments Stephen Bruck
Founding Editor
(1980-1983)
Garth Hastings
Founding Editor
(1980-1995)
Nicholas Peppas
Editor-in-Chief
(1983-2001)
Robert Langer
Editor-in-Chief
(1983- 2003)
David Williams
Editor-in-Chief
(1996-)
The Biomaterials Silver Jubilee Compendium
Controlled release of biologically active compounds from bioerodible polymers J. Heller
Polymer Sciences Department, SRI International, Menlo Park, CA94025, USA Received 8 June 1979; revised 1 October 1979
This article reviewsthe controlled releaseof biologically active agents by the erosion or chemical degradation of a polymer matrix into which the agent is incorporated. Chemically bound active agentsand work on steroid releasefrom cholesterol implants are not covered. The mechanismsof polymer erosion discussedare: cross4inkedscission;hydrolysis, ionization or protonation of pendant groupts; backbone cleavage. Drug releasestudiesare dealt with under each of these headings.
It is now generally recognized that the controlled release of biologically active agents to a local environment can be achieved by means of one of three general methodologies: (1) diffusion through a rate-controlling membrane, (2) use of osmotically regulated flow, and (3) release controlled by the erosion or chemical degradation of a matrix into which the active agent is incorporated 1. Each of these methodologies offers certain unique characteristics which determine the design of specific therapeutic systems. Thus, methodology (1) allows construction of drug delivery devices that release therapeutic agents by zero order kinetics and where rate of delivery can be readily adjusted by changing the rate-limiting membrane and/or memb[ane thickness and area. Methodology (2) allows construction of devices that not only release their contents by zero order kinetics but are also able to sustain high delivery rates not normally available with membranemoderated devices. Methodology (3) allows construction of drug delivery devices that have a predetermined life span and need not be removed from the site of action once their drug delivery role has been completed. Drug release from bioerodible polymers finds use in both topical applications and systemic applications. Both uses demand that the polymer degrade to nontoxic products, and polymers used in systemic applications must also degrade to low-molecular-weight fragments that can be readily eliminated or metabolized by the body. In topical applications, retainment of high molecular weight of the degradation products is desirable, since in this way no unnecessary systemic absorption of the polymer will occur, and toxicological hazards are thus reduced. The purpose of this article is to present a comprehensive review of methodology (3) where active agents are released to a surrounding aqueous environment by solubitization of the polymer matrix induced by the aqueous environment. The review is limited to devices in which the active agent is dissolved or dispersed in a polymer, and does not
cover the important work in which the active agent is chemically bound to the polymer and is released to the surrounding medium by hydrolysis of a bond between the active agent and the polymer chain2,3; nor does it cover the extensive work of Kincl and coworkers on the release of steroids from cholesterol implants 4. For this review, it is convenient to systematize polymer erosion according to the three mechanisms shown in Figure I, where ~) denotes a hydrolytically unstable bond 5. In general terms, Mechanism ] encompasses watersoluble polymers that have been insolubilized by hydrolytically unstable crosslinks; Mechanism ]! includes polymers that are initially water-insoluble and are solubilized by Mechonism T
l~,~h~ism
---'-
c
Mechanism ]]T
Figure 1 Schematic representation of degradation mechanisms; ~) denotes a hydrolyticaHy unstable bond A represents a hydrophobic substituen t and B - . C represents hydrolysis, ionization or protonation
0142-9612/80/010051-07 $02.00 9 1980 IPC BusinessPress Biomaterials 1980, Vol 1 January
51
The Biomaterials Silver Jubilee Compendium Bioerodible polymers: J. Heller
hydrolysis, ionization, or protonation of a pendant group; and Mechanism [[[ includes hydrophobic polymers that are converted to small water-soluble molecules by backbone cleavage. Clearly, these three mechanisms represent extreme cases, and erosion by a combination of mechanisms is possible.
recognized the utility of erodible hydrogels for providing sustained delivery of entrapped macromolecules. Both studies took advantage of the hydrolytic instability of crosslinks formed in vinyl polymers by using N, N~-methyl enebisacrylamide as a comonomer. Hydrolysis of the crosslinks proceeds as follows:
, 0
MECHANISM I
i-
, O~ H_I~_H i 2
= 2-R-_-NH +
Solubilization by crosslink cleavage In these systems, water-soluble polymers are insolubilized by means of hydrolytically unstable crosslinks. Consequently, the resulting matrix is highly hydrophilic and completely permeated by water. Since the active agent is in an aqueous environment, its water solubility becomes an important consideration, and compounds with appreciable water solubility will be rapidly leached out, independent of the matrix erosion rate. There are two general applications in which erodible hydrogets are useful in the controlled delivery of active agents. In the first the active agent has extremely low water solubility, and in the second the active agent is a macromolecule that is entangled in the hydrogel matrix and cannot escape until a sufficient number of crosslinks have cleaved and matrix crosslink density has been reduced. The first application is illustrated in Figure 2, which shows the release of a highly water-insoluble drug, hydrocortisone acetate, from a gelatin matrix crosslinked with formaldehyde 5. As indicated by the first-order dependence, release is by diffusion with little contribution by matrix erosion. Because the drug is very water insoluble, useful release over many days is achieved. Such a device could be used when zero-order kinetics are not important and removal of the expended device is not convenient or desirable 6. Illustrative of the second application are many examples in which water-soluble macromolecules have been immobilized in hydrogets by physical entanglement 7. However, the intent of most of these studies was to achieve long-term immobilization of enzymes or antigens; the slow diffusional escape and/or slow hydrolysis of the matrix with consequent liberation of the entrapped macromolecules was generally regarded as undesirable. Two studies, however,
40-] 3O E 2O-
I0
1
I
72 96 Time(days)
.........
J
120
t44
Figure 2 Releaseo f hydrocortisone acetate from a cross~inked gelatin matrix
52
Biornateriais 1980, Vol 1 January
where - R - represents the vinyl polymer chain. In the first study 8, bovine pancreatic insulin was immobilized in a hydrogel prepared from acrylamide and 2% N, N'-methylenebisacrylamide. Slow release of insulin from the hydrogel was inferred because insulin-containing hydrogel implants sustained diabetic animals for at least a few weeks. It is not clear from the study how much insulin was released by diffusion and how much by cleavage of crosslinks. In the second study 9, (~-chymotrypsin was immobilized in a hyclrogel prepared from N-vinyl pyrrolidone and N, N'-methylenebisacrylamide. It was found that, by varying the N, N'-methylenebisacrylamide concentration from 0.1 to 1.0w/w % with respect to N-vinylpyrrolidone, hydrogels with dissolution times of several days to practically insoluble could be prepared. However, release of =-chymotrypsin did not correlate well with hydrogel dissolution time, presumably because of diffusional escape. To prevent diffusional escape, e-chymotrypsin was acylated with acryloyl chloride and then chemically incorporated in the hydrogel by copolymerization with N-vinyl pyrrolidone and N, N'-methylenebisacrylamide. Release of e-chymotrypsin from the resulting hydrogels was then found to correlate more closely with hydrogel dissolution times.
MECHANISM II Solubilization by hydrolysis, ionization or protonation of pendant groups
Systems in this category include all polymers that are initially water-insoluble but become water-soluble as a consequence of hydrolysis, ionization, or protonation of pendant groups. Because no backbone cleavage takes place, the solubilization does not result in any significant changes in molecular weight. The major emphasis in the development of these materials has been on enteric coatings. These are coatings designed to be insoluble in a certain pH environment, usually the stomach, and then to dissolve abruptly in an environment of a different pH, such as the intestines. Usually these polymers are applied to pills as protective coatings and do not produce steady, sustained release of therapeutic agents. However, by using mixtures of enteric coatings, each with a different disintegration time, it is possible to prolong the action of therapeutic agents10. Literature on enteric coatings is much too voluminous for detailed review, but the coatings can be grouped into three categories according to their dissolution mechanism; (1) dissolution by side group hydrolysis, (2) dissolution by ionization of a carboxylic acid function and (3) dissolution by protonation of amine functions.
The Biomaterials Silver Jubilee Compendium Bioerodible polymers: J. Heller Dissolution b y side group hydrolysis. Materials in this
category are represented by copolymers of vinyl monomers and maleic anhydride'
XI HO -C H^C CHH -CHI~ I ~ 2~/~', 0
0
Xl -CH 2-CH - I H ~ CIHcoo~ cor
0
He
He
where X is OR or H. In the anhydride form, the polymers are waterinsoluble, but on exposure to water they become soluble because of the hydrolysis of the anhydride group. A number of variables affect the rate of polymer dissolution and lag time before initiation of the dissolution process11. In general, time before dissolution increases as the size of the alkyt substituent R in the vinyl ether portion of the copolymer increases, and decreases as the pH of the aqueous environment increases. The rate of polymer dissolution also increases as the pH of the aqueous environment increases. Dissolution by i o n i z a t i o n o f c a r b o x y l groups. Currently used enteric coatings represent this type, and can be represented generally as polyacids. While in unionized form they are water-insoluble, but on ionization of the carboxylic acid functions they become water-soluble. The most widely used enteric coatings are based on cellulose acetate phthalate 12. They are insoluble in aqueous acidic media but, because of the free acid groups on the phthalate radical, dissolve in aqueous bases. Enteric coatings based on cellulose acetate succinate have also been described 13. Partially esterified copolymers of methyl vinyl ether and maleic anhydride or partially esterified copolymers of ethylene and maleic anhydride have also been investigated 14-16. It was found that these materials characteristically exhibit a pH range above which they are soluble and below which they are insoluble. This pH range is quite sharp, about 0.25 pH units, and changes with the number of carbon atoms in the ester side group of the copolymers. This behaviour can be understood readily by considering the number of ionized carboxyls necessary to drag the polymer chain into solution. With relatively small ester groups, a low degree of ionization is sufficient to solubilize the polymer; hence the dissolution pH is low. As the size of the alkyl group increases, so does the hydrophobicity, and IO0
. . . . . . . . . . . . . . . .
progressively more ionization is necessary to solubilize the polymer, resulting in increasingly high dissolution pH. The same argument holds for polymers having the same ester grouping but different degrees of esterification. The higher the degree of esterification, the more hydrophobic the polymer and consequently the higher the dissolution pH. Recently it has been shown 17 that partially esterified copolymers of methyl vinyl ether and maleic anhydride can, in a constant pH environment, release hydrocortisone incorporated therein by excellent zero-order kinetics. Figure 3 shows polymer dissolution rate and the rate of hydrocortisone release for n-butyl half-ester polymer films containing the dispersed drug. Each pair of points represents a separate device in which the amount of drug released by the device into the wash solution was determined by u.v. measurements and the amount of polymer dissolved was calculated from the total weight loss of the device. The excellent linearity of both polymer erosion and drug release over the lifetime of the device provides strong evidence for a surface-erosion mechanism and for negligible diffusional release of the drug. The latter result was independently verified by placing a drug-containing film in water at a pH low enough that no dissolution of the matrix took place and periodically analysing the aqueous solution for hydrocortisone. None was found over several days. Figure 4 shows the effect of size of alkyl group on polymer erosion rate for a series of partial esters measured at pH 7.4. Because of the linear correlation between polymer erosion and drug release, distance eroded can be directly correlated with amount of drug released and was, in fact, derived from measuring drug release. All drug release rates again show excellent linearity and strong dependence on the size of the alkyl group. Since in all experiments drug depletion also coincided with total polymer dissolution, again it can be assumed that drug release and polymer erosion occur concomitantly. The effect of pH on rate of polymer erosion and hence release of hydrocortisone dispersed in the n-butyl partial ester is shown in Figure 5. The date show a clear dependence of erosion rate and drug release on the pH of the eroding medium and, as expected, a progressive decrease in rate as the critical dissolution pH is approached. The partial ester copolymer also has been used recently as a model for a bioerodible drug delivery system that releases a therapeutic agent in response to the presence of a specific external molecule 18. In this model, hydrocortisone was incorporated into an n-hexyt half ester of a methyl vinyl ether-maleic anhydride copolymer and the 3ooI
100
.
.
.
.
.
.
.
.
.
.
.
.
.
.
"~.&250I 60
-6o
cJ_ ~4,,~ 2 o 20 0 O
o
~0 E a. 20
~ tO
20
30 40 Time (h)
50
g
60
Figure 3 Rate of polymer dissolution and rate of release of hydrocortisone for the n-butyl half-ester or methyl vinyl ether-maleic anhydride copolymer containing 10 wt-% drug dispersion. (0), drug release; (A), polymer dissolution. Reproduced from J. Appl. Polym. Sci. 1978, 22, 1991 by permission of John Wiley and Sons Inc., New York
0
I0
!
20
J
:50
1
I
40 50 Time (ram)
'
60
70
80
90
Figure 4 Effect of size ot ester group in half-esters of methyl vinyl ether-maleic anhydride copotymers on rate of erosion at pH 7.4. Reproduced from J. AppL Polym. ScL 1978, 22, 1991 by permission of John Wiley and Sons Inc., New York
Biomateriais 1980, Vol 1 January
53
The Biomaterials Silver Jubilee Compendium Bioerodible polymers:
d. Heller
20c
300
c
17
25(-
15 A
2(30=L
10 G)
~5
150-
IOO-
25 i
%
~6
20 .....30
~ ..... 50
Time (rain)
so
~o-~
9o
~oo
Figure 5 Effect of erosion medium pH on erosion rate of half-esters of methyl vinyl ether~rnaleic anhydride copolymers. A = pH 5.5, B = pH 5. 75, C = pH 6.0, D = pH 6.5, E = pH 7. Reproduced from J. AppI. Polym. Sci. 1978, 22, 1991 with permission of the copyright owner
polymer
I I
.
l
!
I
I
,
1
t
I.
,.
I
1
1
I
20 40 60 80 I00 120 140 160 180 200 220 240 Time (h)
Figure 6 Rate o f hydrocortisone release at 35~ from a n-hexyl half.ester of a copolymer of methyl vinyl ether and maleic anhydride at p H 6.25 in the absence and presence of external urea. A = polymer + hydrogel, 10-1 m urea; B = polymer + hydrogel, 10-2m urea; C = polymer + hydrogel, no urea. Reproduced from J. Pharm. Sci. 1979, 68,919 with permission of the copyright owner
54
In vivo release of hydrocortisone from ocular inserts in
rabbits 5, Kinetic plots were obtained by placing weighed devices in eyes and removing the devices at desired time intervals. The amount of drug released was determined by measuring residual drug in the devices and subtracting that from the known original amount. The results expressed in terms of polymer erosion are shown in Figure 7. Clearly, the devices are highly functional and release hydrocortisone by excellent zero-order kinetics. Furthermore, considering that each point represents a different device and a different rabbit eye, there is very little scatter, indicating a high degree of reproducibility. P o l y m e r dissolution b y p r o t o n a t i o n o f amine functions.
Materials in this group are, in effect, reverse enteric coatings" they are insoluble in water and alkali but soluble in acids. Although no drug-release studies have been described, suggested uses have been as moisture-resistant medicament coatings that will release their contents in the stomach 19,) and as veterinary medicament coatings to allow passage through the stomach of ruminants and subsequent release in the abomasum 20. A material typical of this group is cellulose acetate N, N-diethylaminoacetate, prepared by the amination of cellulose acetate thloroacetate with diethylamine 21. Another material was prepared by the addition of amines to crotonic acid esters of cellulose 22.
Solubilization by backbone cleavage
c 6)
0
Figure 7 rabbits
8O
MECHANISM III
o
9O
Time(h)
Biomaterials 1980, Voi 1 January
This category includes all water-insoluble polymers that undergo hydrolytic backbone cleavage and are solubilized by conversion to small, water-soluble molecules. A major driving force for the development of these materials was a search for absorbable sutures superior to catgut. From these studies have evolved two synthetic materials that were suitable for surgical implants: poly (lactic acid) 23,24 and poly(glycolic acid) 25. The first demonstration of the utility of poly(lactic acid) as a bioerodible implant capable of sustained release of a therapeutic agent was described about 10 years ago26, and since then many publications have dealt with the sustained release of pharmaceuticals from bioerodible implants. For this review, the literature will be discussed according to the type of pharmaceutical agent delivered: (1) delivery of narcotic antagonists, (2) delivery of contraceptive steroids, and (3) delivery of others. D e l i v e r y o f narcotic antagonists. The release of cyclazocine from poly(lactic acid) 27 and the release of Naltrexone
The Biomaterials Silver Jubilee Compendium Bioerodible polymers: J. Heller
.ol
IOO 90 80
70
/
~Q;7 0
_=
30
2O I0
N
t
tO
O
..........t
J
20
t
30 40 Time (doys)
~_.....
50
1............ ~ r
60
70
Figure 9 /n vitro release of 3H-Na/trexone base from uncoated rods of 75/25 poly (L (+)-/actic~co~]lycolic acid) into 37~ buffered, pH7 solution as a function of drug loading, x, 80% w/w n. base; II, 70% w/w n. base," A, 60% w/w n. base," I , 50% w/w n. base. Reproduced with permission from Life Sci 1975, 17, 1877 9 1975 Pergam on Press Ltd.
!
i i ~
......
t
i0
.........
i
20
1
I
3o 40 Time (doys)
.....
i
5o
l
6o
.
Figure 8 Cumulative amounts of cyc/azocine excreted from composites of po/y (lactic acid) implanted or injected in rats. Amounts have been corrected to account for unrecovered drug, (E3), (O), and (A), film; (~), small particles; (~), film with drug sealed into poly(/actic acid) envelope. Reproduced with permission from J. Med. Chem. 1973, 16,897 9American Chemical Society
from poly(lactic acid) 28 have been described in two studies. In the first study, composites of poly(lactic acid) and labelled cyclazocine were prepared and implanted in rats as 4 cm 2 films, as ground films having particle sizes falling within No. 25/35 sieves, and as 6.2 cm 2 poty(lactic acid)-cyclazocine composites enveloped in poly(lactic acid) containing no drugs. In vivo release was followed by monitoring the radioactivity of excreted urine. The results, shown in Figure 8 were not those expected" cyclazocine is released by diffusion, and therefore release from small particles with their greater surface area should be much faster than release from films. It was postulated 27 that the in vivo results are influenced by varying degrees of inflammation and oedema. This may be reasonable, because in vitro release data show the expected much faster release from particles than from films. Since the polymer was said to bioerode in about 62 days and after 55 days, only 30% of the drug was released, a large burst in drug delivery should take place shortly thereafter. In the second study 28, tritiated Naltrexone was incorporated into poly(lactic acid), and the composite was ground and then injected into rats. Drug release, as measured by urinary excretion, levels off at about 25% after about 30 days. Again, there was a considerable disparity between in vivo and in vitro release, this time attributed to tritium exchange in the body. Release of cyctazocine microencapsulated in poly (lactic acid) has also been described. 29 However, because of macroscopic defects in the capsule wall, all cyclazocine was released in vivo between 14 and 17 days. The permeability of cyctazocine through solvent-cast poly(lactic acid) was measured to yield a value of 2.9 and 3.0 x 10 -11 cm2/sec. Using an average of these values, it was calculated that a defect-free capsule should release 50% of its contents in about 28.5 months.
The in vitro release of Naltrexone and Naltrexone pamoate from poly(lactic acid) microcapsules has also been described 30. Again, the microcapsutes contained defects, and about 50% of the contents were released after the first few hours of extraction. Copolymers of lactic and glycolic acid have also been used in a bioerodibte delivery system for Naltrexone and Naitrexone pamoate3-1~ The in vitro release rates of Naltrexone from rods of 75/25 poly(lactic/glycolic acid) copolymer as functions of drug loading are shown in Figure 9. As expected, increase in drug loading results in increased delivery rates. In vivo release of labelled Naltrexone from 90/10 poly(lactic/glycolic acid) beads is shown in Figure 10. The release was measured as urinary excretion; after about 90 days about 50% of the drug was accounted for. The remaining drug was assumed to have been excreted in the faeces. Other investigators 32 found that Naltrexone is excreted in approximately equal amounts in the urine and in the faeces. Delivery o f contraceptive steroids, d-Norgestrei was incorporated into poly(lactic acid) and its release rate followed both in vitro and in vivo 33. It was found that films of the polymer containing 33% Norgestrel released the steriod at a relatively constant rate of 3 #g/day per cm 2 for over 80 days. When similar films were implanted subcutaneously in rats, the initial release was about 5.5 #g/day per cm2; by 80 days release had declined to about 3 #g/day .~ 5o -
E
g
3
t0
20
30
40 50 60 70 BO 90 Time(days pastimpi~tion) Cumulative percent o f 3H, implanted as 3H-Nattrexone
Figure 10 base {33% w/w) in 90/10 po/y (Lf+)-/actic
Biomateria/s 1980, Vo/ I January
55
The Biomaterials Silver Jubilee Compendium Bioerodible polymers: J. Heller
50j ............ 50
1~ 40_=
-T
o 30
~, 3 0 -
~9 20 ~
2o-
Io
13 o
0-- ~ ~
Io
~
0
..~.o -~
20
!
4o
i
t
60 80 Time(doys)
.............................. 1...............
t00
1
120
Figure 11 Cumulative percent norethisterone re~easedin vitro from particles containing 20% by weight norethisterone. (0) 90-1801~; uncoated (n) 180-250 ~; uncoated (<>) 180-250p; coated (using a 10% poly-DL-lactic acid/benzene solution). From Contraception 1976, 13, 375. Reproduced with permission of the copyright owner
per cm 2. Biodegradation of the matrix was slower than drug release. In another study 34 14C-labelled 20 wt% Norethisterone was incorporated into poly (L (+)-tactic acid) and the composite was cryogenically ground and separated into 90- to 180-/J particles and 180- to 250-/J particles. The 180- to 250-/J particles were overcoated with the benzenesoluble poly(DL-lactic acid). The in vitro release of the steriod from all three particle-size formulations is shown in Figure 11. Release from the larger particles was faster than that from the smaller particles, which is contrary to what might be expected because rate of diffusional release is directly proportional to surface area. Release from coated samples is considerably lower than that from uncoated samples. This is as expected; the steriod must diffuse through a rate-controlling membrane. Studies of in vivo release, as measured by urinary excretion, show an apparent zero-order delivery with an initial burst and another burst around day 90, which is attributed to dissolution of the polymer matrix. Comparison of actual in vivo release kinetics for all three samples is difficult because the steriod is excreted in both urine and faeces, and in the uncoated samples only the urinary steroid was measured. Furthermore, either because of cumulative error or errors in original specific activity, mass balance is poor. However, the results indicate fairly linear and sustained steroid release for about three months; a typical plot is shown in Figure 12. In vivo and in vitro release of progesterone, ~-oestradioi, and dexamethasone from poly (lactic acid) beads and chips also has been described 35. A considerable amount of work has been done on release of contraceptive steroids from crosslinked poly (dimethyl siloxane) implants 36. However, silicone rubber is not degradable, so the expanded device must be surgically removed. This is not desirable; consequently it is desirable to develop devices which would release steroids by membrane diffusion and the expended devices would then later bioerode. in an effort to develop biodegradable materials with permeabilities comparable to those of silicone rubber, homo- and copolymers of e-caprolactone and DE-lactic acid were investigated as potential bioerodible membranes 37. It was found that poly (e-caprolactone) and
56
j
Biomateriais 1980, Vot 1 January
10
i
2o
L_.
3o
1
I
.I
4o 5o 60 Time (doys)
| .......................... I. . . .
70
80
t
90
Figure 12 Cumulativepercent 14C label from norethisterone excreted in vivo in the urine of rats from 90-180 # particles of poly-L (+)-lactic acid containing 20~ by weight norethisterone. From Contraception 1976, 13, 375. Reproduced with permission of the copyright owner
silicone rubber have comparable maximum steady state fluxes of progesterone at 37~ 2.2 x 10"10 and 0.6 x 10-10 g/cm sec, respectively). Poly (DE-lactic acid) had a maximum steady state flux of progesterone of 3.3 x 10-15 g/cm sec. In another study 38, release rates of several steroids from films and capsules of homopolymers and copolymers of e-caprolactone, DE-tactic acid and glycotic acid in vitro and in vivo were determined. Detailed interpretation of the data is difficult because various factors, such as simple drug diffusion, polymer erosion, morphological changes in the polymer, and changes in concentration gradient across capsule walls, all contribute to the measured erosion rate. The relaase of progesterone from bioerodible capsules based on gtutamic acid/leucine copolymers has also been investigated 39. Devices were fabricated by first preparing progesterone-filled polypeptide rods and then coating the rods with unfilled polypeptide. Because of various fabrication difficulties, it was not possible to determine meaningful release rates. Bioerosion of the copolymers was extremely slow and complete erosion for a 50/50 copotymer was estimated to take about four years. Delivery o f other agents. Although the greatest emphasis of drug delivery from bioerodible polymers has centered around narcotic antagonists and contraceptive steroids, the methodology has been applied to other drugs. A brief study described release of two anticancer drugs, cyclophosphamide and cis-dichlorodiammineplatinum, from poly(lactic acid) 40. Another study 41 described the release of the antimalarial drug, 2, 4-diamino-6-(2-naphthylsulphonyl)quinazoline (WR-158122) from a 25/75 poly(DL-lactic/ glycolic acid) copolymer in mice. Measuring radioactivity excreted into urine and faeces, sustained release through 14 weeks, was demonstrated.
DEGRADATION MECHANISMS The hydrolytic erosion of a solid polymer can be rationalized in terms of two extreme mechanisms. In one, referred to as homogeneous, the hydrolysis occurs at a uniform rate throughout the matrix. In the other, called heterogeneous, the process is confined to the surface of the device. Actual erosion can, of course, occur by some intermediate mechanism. In a purely homogeneous process, the matrix will remain essentially intact until all parts reach some critical
The Biomaterials Silver Jubilee Compendium Bioerodible polymers: J. Helter
degree of reaction, at which point the matrix dissolves. Drug release from such a matrix is complicated because it is a combination of diffusion and erosion. Diffusional drug release in the absence of erosion can be described by the Higuchi model 42 which proposes that the drug is initially removed from the surface regions of the polymer, and consequently a progressively thicker drugdepleted layer forms adjacent to the surface of the device. Because the remaining drug must diffuse through a progressively thickening polymer membrane, the rate of drug release declines continuously. In the presence of polymer erosion, the matrix progressively loosens up, and the permeability of the polymer to the drug increases with time. Consequently, the rate of drug release from polymers which erode either wholly or partially by bulk degradation at first shows the normal expected decline, but as the increasing polymer permeability gradually offsets this decline, the rate of drug release eventually accelerates. The heterogeneous process is a much more desirable degradation mode because it will lead to zero-order drug release, provided that diffusional release of the drug is minimal and the cwerall shape of the device remains essentially constant, thus maintaining constant surface area. Furthermore, such a process avoids deterioration of mechanical properties, which can take place when random chain cleavage occurs in the bulk material. In principle surface erosion can be achieved by constructing a highly hydrophobic polymer, so that water penetration and consequent bulk hydrolysis is much less probable than surface reaction. A much preferred approach is one in which a pH-sensitive reaction is selected and the interior of the matrix is buffered so that bulk hydrolysis is prevented and reaction can take place only at the surface of the device, where the buffer is neutralized. The only polymers thus far described in detail that undergo true surface erosion are the partial esters of methyl vinyl ether/mateic anhydride copolymers 17,18, However, these represent a special case because sotubitization does not involve backbone cleavage, and consequently the only polymer chains than can escape into the aqueous environment are at the surface of the device. None of the described backbone-degradable polymers undergo surface erosion; rather, they undergo bulk erosion. This is not surprising, since most of them were originally designed as bioerodible suture materials and were not intended as drug carriers capable of releasing a drug by zero-order kinetics. However, work in progress here 43 and elsewhere44, 45 is devoted specifically to the development of bioerodible polymers that do undergo surface erosion and are capable of releasing incorporated drugs by zeroorder kinetics.
REFERENCES 1 2
3 4 5
Baker,R.W. and Lonsdate, H.K., Chemtec. 1975, 5,668 Harris,F.W., Aulabaugh, A.E., Case, R.D., Dykes, M.K. and Fetal, W.A., 'Controlled Release Polymeric Formulations' iEds. D.R. Paul and F.W. Harris) 1976, ACS Symposium Series No. 33, p. 222 Neogi,A.N. and Allan, G.G., 'Controlled Releaseof Biologically Active Agents', (Eds. A.C. Tanquary and R.E. Lacey) 1976, Plenum Press, p. 195 Joseph,A.A., Hill, J.L., Patet, J., Patet, S. and Kincl, F.A., J. Pharm. Sci. 1977, 66, 490 Heller, J. and Baker, R.W., Paper presented at the University of Southern California Symposium on Applications of Polymers in Dentistry and Medicine, November 1973
6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35
36 37 38 39 40 41 42 43 44 45
Hussain,A.A., Higuchi, T. and Shell, J.W., US Pat 3960 150 (1976) Goldman,R., Goldstein, L. and Katchalski, E., 'Biochemical Aspects of Reactions on Solid Supports' (Ed. G,R. Stark) 1971 Academic Press, p. 13 Davis,B.K., Experentia 1972, 28, 348 Torchilin, V.P., Tischenko, E.G., Smirnov, V.N. and Chazov, E.i., J. Biomed. Mater. Res. 1977, 11,223 Ballard,B.E, and Nelson, E., 'Remington's Pharmaceutical Science', 1975, Mack, Easton PA, Ch. 91 Woodruff, C.W., Peck, G.C. and Banker, G.S., J. Pharm. Sci. 1972, 61, 1916 Wagner,j.G. and Long, S., J. Pharm. Sci. 1960, 49, 121 Wilken,L.O. Jr., Kochhar, M.M., Bennett, D.P. and Cosgrove, F.P., J. Pharm. Sci. 1962, 51,484 Lappas,L.C. and McKeehan, W.,J. Pharm. Sci. 1962, 51, 8O8 Lappas,L.C. and McKeehan, W., J. Pharm. Sci. 1965, 54, 176 Lappas,LC. and McKeehan, W., J. Pharm. Sci. 1967, 56, 1257 Heller,J., Baker, R.W., Gale, R .M. and Rodin, J.O,, J. Appl. Po/ym. Sci. 1978, 22, 1991 Heller,J. and Trescony, P.V., J. Pharm. Sci. 1979, 68,919 Hiatt, G.D., US Pat. 2940901 (1960) FarbenfabrikenBayer A.G., Ger. Pat. 1 219 175 (1965) Hiatt, G.D., Mench, J.W. and Fulkerson, B., Ind. Eng. Chem. Prod. Res. Develop. 1964, 3, 295 Mench,J.W. and Fulkerson, B., Ind. Eng. Chem. Prod. Res. Deve/op, 1968, 7, 2 Kulkarni, R.K., Pani, K.C., Neuman, C. and Leonard, F., Arch. Surg. 1966, 93, 839 Kulkarni, R.K., Moore, E.G., Hegyeli, A.F. and Leonard, F., J. Biomed. Mater. Res. 1971,5, 169 Frazza,E.J. and Schmitt, E.E., J. Biomed. Mater. Res. Syrup. 1971, 1,43 Yolles, S., Eldridge, J.E. and Woodland, J.H .R., Po/ym. News 1970, 1,9 Woodland,J.H.R., Yoiles, S., Blake, D.A., Helrich, M. and Meyer, F.J., J. Meal. Chem. 1973, 16, 897 Yolles,S., Leafe, T.D., Woodland, J.H.R. and Meyer, F.J., J. Pharm. Sci. 1975, 64,348 Mason,N., Thies, C. and Cicero, T.J., J. Pharm. Sci. 1976, 65,847 Thies,C., 'Controlled Release Polymeric Formulations', (Ed. D.R. Paul and F.W. Harris) ACS Symposium Series No. 33, 1976, p. 190 Schwope,A.D., Wise, D.L. and Howes, J.F., Life Sci. 1975, 17, 1877 Reuming,R.H. and Malspeis, L., Ohio State University reports on National Institute on Drug Abuse, Contract No. H5M42-73-182, 1974-t975 Jackanicz,T.M., Nash, H.A., Wise, C).L. and Gregory, J.B., Contraception 1973, 8, 227 Anderson,L.C., Wise, D.L. and Howes, J.F., Contraception 1976, 13, 375 Yolles,S., Leafe, T., Sartori, M., Torkelson, M., Ward, L. and Boettner, F., "Controlled Release Polymeric Formulations', (Eds. D.R. Paul and F.W. Harris) ACS Symposium Series No. 33, 1976, p. 123 Hiltier, S.G., Jha, P., Griffiths, K. and Lanmas,K.R., Contraception 1977, 15,473 Pitt, C.G., Jeffcoat, R.A., Zweidinger, R.A. and Schindler, A., J. Biomed. Mater. Res. 1979, 13,497 Pitt, C.G., Gratzl, M.M., Jeffcoat, R.A., Zweidinger, R. and Schindler, A., J. Pharrn. Sci., in press Sidman,K.R., Steban, W.D. and Burg, A.W., in Drug Delivery Systems (Ed. H.L. Gabelnick) 1976, DHEW Publication No. (NIH) 77-1238 Yolles,S., Leafe, T.D. and Meyer, F.J., J. Pharm. Sci. 1975, 64, 115 Wise,D.L., McCormick, G.J., Willet, G.P. and Anderson, L.C., Life Sci. 1976, 19, 867 Higuchi, T., J. Pharm. Sci. 1961,50,874 Hetler,J. and Baker, R.W., Work in progress Choi, N.S. and Heller, J., US Pat. 4093 709 (1978); 4131 648 (1978); 4138344 (1979) Capozza,R.C., Schmitt, E.E. and Sendelbeck, L.R., in National Inst. Drug Abuse Research Monograph 1976, No. 4 (Ed. R.E. Willette) DHEW Publication No. (ADM) 76-296
Biomateriats 1980, Vol 1 January
57
The Biomaterials Silver Jubilee Compendium
This Page Intentionally Left Blank
The Biomaterials Silver Jubilee Compendium
The response to the intramuscular implantation of pure metals A. McNamara and D.F. Williams Department of Dental Sciences, School of Dental Surgery, University of Liverpool, P.O. Box 147, Liverpool L69 3BX, UK (Received 30 September 1980)
Discs of five high-purity metals, cobalt, nickel, copper, aluminium and lead have be~n implanted intramuscularly in rats and the response observed histologically for period up to 52 weeks. A reproducible but different response was observed with each metal. Whenever corrosion occurred, as with copper, nickel and some cobalt specimens, the implants became loose. In the absence of corrosion, the implants were firmly held within a more confined capsule. A minimal response was seen with lead, implying normally toxic metals do not necessarily induce excessive responses on implantation. Evidence was found, however, that some metals elicit an immune response whilst some, especially copper and nickel appear to render the host more susceptible to disease. The implants appear to have a profound effect on the immediate vasculature, are able to cause a prolonged polymorphonuclear response in the same way as bacteria, are associated with varying amounts of haemosiderin laden macrophages but not with giant cells. The animals appear to be able to deal with bacteria introduced at surgery without hindrance from the metal.
Ever since the early experimentation with metallic implants by Zierold 1, Hey-Groves 2 and others, there has been a steady rationalisation of the metals and alloys used clinically and the development of a concise list of recognised and acceptable materials, These are stainless steel of the 316 type, titanium and titanium-aluminium-vanadium alloy and some cobaltbased alloys such as the c o b a l t - c h r o m i u m molybdenum casting alloy, the wrought cobaltc h r o m i u m - nickel, cobalt- nickel- chromiummolybdenum and cobalt- c h r o m i u m - nickel- i r o n molybdenum alloys. The basis for this selection has been largely that of corrosion resistance and apparently good biocompatibility, although naturally mechanical properties have also been an important consideration. The corrosion resistance of these alloys has been extensively investigated, both from in vitro experiments and in vivo observations 3' 4 and it is well established that titanium and its dilute alloy are essentially immune from corrosion in the physiological environment, the cobalt-based alloys are highly corrosion resistant with only a minimal susceptibility whilst stainless steel quite readily suffers crevice and pitting corrosion. The tissue response has been less well investigated. There have been some experimental studies using various animal models s,6 and observations on biopsy or autopsy samples 7"8 but little in the way of a detailed systematic study of the tissue response to metals. The general observation is that, in the absence of gross corrosion or wear there appears to be little difference in the reaction to the alloys 0142-9612/81/010033..08 $02.00
mentioned above, implants becoming enveloped in a relatively thin fibrous capsule, with no ultrastructionai features to distinguish the response to the different materials. There appears, therefore, to be little to choose between the various metals on the basis of this tissue response. However, with increasing concern over the biocompatibility of implant materials and especially in terms of the significance of corrosion, wear, hypersensitivity and carcinogenicity, it is important that a more detailed understanding of the effects of implanted metals be gained. A major programme of work is under way in the author's laboratory based on this objective and the present paper is concerned with part of this work. Although the ultimate aim must be related to the materials used clinically, with the exception of commercially pure titanium, these are all alloys with 3, 4 or even 5 major elements present. Since the tissue response will be largely dependent on the rate of release of metal ions (or particulate corrosion products) and the physiological activity of these ions, and since each element will be released at a different rate from a complex alloy and will have a different mechanism of toxicity, it would be difficult, if not impossible, to interpret the response to such alloys in terms of basic parameters until the precise effect of each constituent had been determined. In the present study, therefore, a series of pure metals has been investigated. Five metals were chosen for the inital study, cobalt, nickel, copper, aluminium and lead, These were selected to give a wide range of characteristics, both metallurgically and biologically.
91981 IPC Business Press Biomateria/s 198 I, Vo/ 2 January
33
The Biomaterials Silver Jubilee Compendium
10 Metal implants: A. McNamara and D.F. Williams
Other metals will be studied and reported in due course. The investigations have involved the implantation of discs and powders into animals with examination of the tissue by conventional histological techniques and also by scanning and transmission electron microscopy, EDAX analysis, enzyme histochemistry and local tissue and organ analysis using atomic absorption and polarographic techniques. The present paper is concerned only with the detailed microstructural examination of the soft tissue adjacent to intramuscular implants of these metals. Results of other studies will be reported in future papers.
MATERIALS AND METHODS High purity metals obtained from Metals Research Ltd., were used in this study, the specifications being given in Table 1. Discs of 5 mm diameter and approximately 2 mm thickness were prepared from the cobalt, nickel, aluminium, lead and copper rods by slow speed, continuously cooled cutting with a diamond wheel. The surfaces of the discs were ground and polished by conventional metallographic techniques to a 0.5/~m alumina finish. The discs were autoclaved individually at 1 20~ for 1 5 minutes. The discs were implanted intramuscularly in 4-6 month old male black and white hooded Lister rats of the Liverpool strain. The rats were anaesthetised using the neuroleptanalgesic Immobilon* and two implants were place in separate locations within the musculature of the dorso-lumbar region. The wounds were closed using Dexont and silk sutures and the area sprayed with a clear acrylic dressing. The rats were maintained on standard laboratory diet and water for varying periods between 10 and 52 weeks. At sacrifice, the implant site was located and
Table 1 Metals used .
.
.
.
.
.
.
.
.
.
.
:.
,~
_ _
Metal
Purity (%)
Form
Aluminium Cobalt Copper Nickel Lead
99.998 99.998 99.999 99.999 99.999
5 mm dia. rod 5 mm dia. rod 5 rnm dia. rod 5 mm dia. rod 5 mm dia. rod
the block of tissue surrounding the implant was removed. Swabs were taken and cultured aerobically and anaerobically. The tissue was secured between two metal discs to faciltate rapid heat transfer, perfused with 10% sucrose and quenched in isopentane which had been precooled with liquid nitrogen. The samples were transferred to a cryostat where 7/~m sections were cut onto glass coverslips at a temperature of--10~ Routine staining procedures, using Haematoxylin and Eosin and, in a small number of cases where it was necessary to confirm the presence of haemosiderin, Perl's stain, were carried out on the sections although histochemical methods were used on others for the demonstration of enzyme activity.
RESULTS Over the time course studied, that is 10-52 weeks, a morphologically distinct pattern developed for each metal studied. Certain features were common to some metals, of course, but the sequence of events and the histological picture at a given time interval was quite specific for the different metals and very reproducible. The results for each metal are presented here in the form of a drawing in which the main morphological features of the capsule surrounding the implant are given along with one or more photo-micrographs that illustrate specific features. Although the tissue response was different in each case, an overall pattern could be distinguished. The general features of this pattern are shown in Figure 1.
In no case was any trace of infection found.
Lead The general form of the capsule seen with lead is given in Figure 2a. This capsule may be regarded as the simplest of those observed in these experiments. A typical histological section is shown in Figure 2b, a dense collagenous capsule forms around the metal through fibroblast activity, but once the capsule has formed the situation remains static. Further away from the site some abnormal changes may be noted in the muscle (Figure 2c) indicating some degree of activity. This feature, which is of considerable potential significance is under further investigation.
Copper Figure I General structure of capsule around metal implant: A: implant. B: zone of necrosis. C: area of cellular infiltration, of varying intensity. D: oriented collagen. E: loosely woven collagen. F: blood vessels. G: muscle fibre island. H: normal muscle.
"Small animal Immobilon, Reckitt El' Coleman, Pharmaceutical Division. York. t Dexon, polyglycolic acid sutures, Davis and Geck Ltd., Southampton.
34
Biomaterials 1981, Vol 2 January
In contrast, there is an extensive reaction to copper, as indicated in Figure 3a, the most noticeable aspect being the numerous sprouts of blood vessels in the capsule. This capsule has a definite stratified appearance. Pus is normally present, adjacent to the implant, which has always proved sterile on culturing. Haemosiderin-laden macrophages are always in evidence, as are cells containing a black pigmented material.
11
The Biomaterials Silver Jubilee Compendium Metal implants: A. McNamara and D.F. Wilh'ams
Nickel As illustrated in F i g u r e 4a, the overall pattern with nickel is quite similar to that with copper, a sterile exudate, haemosiderin-taden macrophages ( F i g u r e 4 b j a high level of vascularity and muscle fibre 'islands' being found in both tissues. The latter are small groups of fibres, apparently healthy, which appear to become isolated from the main muscle blocks by an advancing front of collagen that is laid down by fibroblasts
Figure 2 (a) General structure of capsule around lead implant. Some activity is noted in distant muscle fx) but the capsule itself fy) is relatively unremarkable, consisting largely of collagen fibres fb) Normal muscle (leftj with thick collagen boundary adjacent to lead implant, X 62.5, H and E stain (c) Unusual array of very basophilic nuclei in muscle fibre at some distance from lead implantX 400 H and E stain
Large areas of a yellow-brown #~gmented tissue are seen { F i g u r e 3 b ) which appear to be in the process of degeneration when their progress through this time period is monitored. This pigmentation, the constitution of which is unknown at present, is unique to the tissue adjacent to copper. With time, progressive vacuolisation and resorption of the muscle fibres occurs within the affected parts ( F i g u r e 3 c j .
Figure 3 (a) General Structure of capsule around copper implant. There is an extensive reaction with pigmented material (x) contained within a dense cellular infiltrate and a rich vascular supply fy) fb) Extensive pigmentation of unknown nature farrowed) in the immediate vicinity of a copper implant, X 62.5, H and E stain (c) Vacuolisation and degeneration of muscle fibres occurring around a copper implant. The circular outline of the fibres is still visible, ~< 400, H and E stain
Biomateriais 1981, Vol 2 January
35
12
The Biomaterials Silver Jubilee Compendium Metal implants: A. McNamara and D.F. Wit/lares
(Figure 4c). The significance of this is not known but these islands are often seen to be degraded in preference to the main blocks of muscle.
Aluminium A very compact reaction is seen with aluminium implants, as represented in Figure 5a although it is, nevertheless, very severe. Ordered layers of collagen are present adjacent to the muscle fibres (Figure 5b).
Figure 5 (a) General structure of capsule around aluminium implants, showing compact but general reaction (b) Well structured but compact capsule around aluminium, with collagen layers clearly visible, X 63, H and E stare (c) Numerous cells visible between the necrotic regions (NJ and the collagen fibres (CO) around alummium, X62.5, H and E stain
Packed cells occupy a position between these layers and a zone of necrosis adjacent to the implant itself
(Figure 5cJ.
Cobalt Figure 4 (a) General structure of capsule around nickel implants, quire similar to that associated with copper but not so marked, (b) Numerous haemosiderin laden macrophages (arrowed) evident around the margin of the capsule of nickel specimen, X 80. H and E stain (C,t Groups of muscle fibres isolated from the remainder of the muscle by" interwoven collagen strands; nickel implant, X 62.5, H and E stain
36
Biomaterials 198 r, Vot 2 January
The reaction to cobalt is, perhaps, the most interesting of all ( F i g u r e 6a). With this metal there is no c/ear boundary or capsule edge, but rather a general gradation in appearance from the normal muscle fibres some distance from the implant to those which are breaking down closer to the metal (Figure 6b). Many
The Biomaterials Silver Jubilee Compendium
13
Metal imlMants: A. McNamara and D.F. Williams
usually completely blackened but the black surface layer could be peeled away to reveal normal metal. The copper discs exhibited patchy areas of corrosion. Two types of surface appearance could be distinguished with cobalt. In some cases the metal became dark in colour, almost black, but with a pink deposit on the surface. In other cases the metal appeared unchanged. Often two bilateral samples in the same animal yielded different surfaces on recovery. Aluminium discs became duller and more obviously grey, while the colour of the lead specimens deepened into a dark grey with no surface lustre. Adhesiveness between the implant and surrounding tissue varied with individual metals. Lead was the most difficult to remove, followed by aluminium. Both these metals were enveloped by a thin covering of protein which, if not completely removed, made implant extraction extremely difficult. Cobalt, when removed in a bright state, was also covered by a similar layer and was difficult to excise but the darker corroded discs were easily removed. Copper and nickel discs often appeared to be in a loose pulp of cells and were very easily extracted.
DISCUSSION
Figure 6 (a) General structure of capsule around cobalt implants showing more gradual change tn appearance moving out from the implant surface and the lymphoid clumps (x), and areas of muscle breakdown {!r {bj Muscle breakdown m tissue around cobalt, (arrowed) ~< 52.5, H and E stare /cj Lymphozd clumps between the muscle fibres around cobalt implant, :~:i 157, H and E stain
cells have infiltrated between the muscle fibres over this region and a number of lymphoid clumps are evident f F i g u r e 6c). These are composed mainly of plasma cells. In some areas, blood-borne lymphocytes have stuck to the endothelium of blood vessel walls as they are passing through to the site of the stimulus. tn most cases, the macroscopic appearance of the surface of the metal discs at recovery was different to that seen at implantation. The nickel implants were
As noted in the introduction, the morphology of the fibrous capsule around implanted metals has rarely been studied and no detailed differences in capsules formed around different metals and alloys have been reported. However, unless the fibrous capsule arises only as a result of the mere physical presence of the implants, which seems unlikely, it shou!d be expected that the capsular structure would vary with the actual metal involved, being dependent upon the rate of release of metal ions from the implant surface, their diffusivity in the tissue, their protein-binding capability and their cytotoxicity. Moreover it should be possible to correlate the biocompatibility of these metals determined in this way with the known physiological effects of the metals observed in other situations such as after oral ingestion of excesses of the metal, skin contact, inhalation and certain metabolic disorders. The present investigation has provided a detailed characterisation to be made of the capsular tissue following the intramuscular implantation of pure metal and certain comments can be made concerning the significance of these observations. The very fact that the appearance of the tissue was quite specific for each metal, and was consistently so, indicates the structure is dependent on the nature of the metal ions released. A number of short term controls were performed in which incisions were made but no implants inserted, the operational sites being located by the sutures. Long term controls proved technically difficult because of problems of distinguishing sites, nonresorbable sutures or dyes being inappropriate since they may influence the response. A relatively small number of muscle fibres would be damaged at incision and these could easily be overlooked at subsequent microscopical examination, even if the site of damage were known. There is, however, indirect evidence that the damage incurred at operation has been repaired at the time intervals reported here. A number of large well-formed muscle fibres with central nuclei have
Biomaterials 1981, Vol 2 Januarv
37
The Biomaterials Silver Jubilee Compendium
14 Metal implants: A. McNamara and D.F. Williams
been found in association with the implants. Reznik and Engei 27 found similar fibres in cold-damaged muscle and Carlson and Gutmann 2~ found similar fibres in regenerated muscle grafts. The fact that such fully regenerated fibres start to appear from one month onwards post-trauma in these cases implies that the underlying repair processes in the present situation should have ceased and that the observations are a reflection on the action of the metals. In this discussion some comments are made first about the individual metals prior to an assessment of the general features or the tissue response to metals.
Aluminium The response to the aluminium implants was quite remarkable in that immediately adjacent to the implant was a distinct necrotic region but only a short distance away the muscle was perfectly healthy with no abnormalities. The capsular boundary was marked by a few layers of collagen but these were far less extensive than those found with the lead. There was no polymorphonuclear leucocytic infiltration, but many macrophages were present, actively secreting hydrolytic enzymes. There was no major alteration in the vascular network, the area being poorly vascularised. This could account for the necrosis present since catabolic wastes, including dead cell fragments, accumulate, while the oxygen and nutrient supply to the tissues remains unchanged. These experiments provide no evidence of sarcoma induction, following the implantation of metallic aluminium, in contrast to those of O'Gara and Brown 9. These authors stated, however, that the preceding fibrous and inflammatory reactions had subsided before tumour development was noticed. In the present experiments, the reaction had shown no signs of subsiding, even at the longest time period and it is possible that a longer implantation period is necessary for this tumour induction.
It is possible that lead is readily excreted from the body when introduced intramuscularly since it is known that in both man and laboratory animals, the normal lead distribution pattern reveals very low muscle levels 12. During high lead intake, lead concentrations increase in all tissues except the muscles.
Copper The most noticeable feature of the response to copper is the very high degree of vascularity, which should, of course, correlate with the speed of systemic distribution of the metal ions. That such augmentation of the vascular supply results from the presence of the copper is an interesting observation and work is in progress in other laboratories to determine if copper is a true angiogenic factor, (personal communication) that is, if it has the capacity to induce directional growth and proliferation of blood vessels. Many tumours have such a capability and by secretion of tumour angiogenic factors are able to create favourable local conditions for increased growth or metastasis. The tissue pigmentation observed in the vicinity of the implant could result from a local deposition of copper ions, although the exact nature of the action here is not known. This hypothesis is supported by observations that systemic copper overload can result in yellow pigmentation of the liver ~3 and in Wilsons disease golden Kayser-Fleischer rings are visible on the cornea due to deposits of crystals TM. The presence of large vacuolated cells is not readily explained. These closely resembled the altered cells characteristic of many lipid and storage diseases, for example, those seen in Gaurcher's disease Ts or cholesterosis. These are large macrophages which have a foamy appearance. Indeed similar lipid-filled cells formed in atherosclerosis lesions are commonly called foam cells 18. However, the cells observed here were not characteristic of the lipid-filled macrophages often found in areas of cellular breakdown.
Lead
Nickel
The response to lead is, perhaps, the least spectacular of the series. Fibrobtastic proliferation occurs adjacent to the implant with subsequent encapsulation within a dense collagenous layer. Once this has occurred, no further fibroblastic action is seen. No abnormalities were seen in this immediate tissue, although, as noted earlier, there were some changes in the muscle at some distance from the implants. These observations will be reported at a later date. This relatively benign response appears to be in conflict with the known toxicity of this metal. However, although much is known about the effects of ingesting lead in reasonably larger quantities 1~ data is sparse concerning the biological effects of lead when present in small quantities in the tissue. Certainly none of the classical manifestations of lead poisoning have been observed in the present experiments. In order for the cells involved in the reactions to effect a repair of the original trauma and synthesise a capsule, the initial rate of lead ion release must have been low. The subsequent presence of the thick collagen layer would further decrease the possibility of deleterious effects and the avascular nature of the area would prevent rapid dissolution of the metal systemically.
The reaction to nickel had many similarities to that obtained with copper. The area of involvement was not so large and the degree of vascularity was not so extensive. Nevertheless a potentially similar angiogenic action to that proposed for copper could well operate. A loose cellular exudate was again present, with an identical composite of polymorphonuclear leucocytes, especially neutrophils, to that observed with copper. A situation exists here in which the metal cannot be effectively isolated from the surrounding tissues because of the toxic effects on the cells that come into contact. Normally in an inflammatory response polymorphonuclear cells arrive quickly and secrete enzymes to break down or render harmless the irritating agen( 7. Subsequently macrophages invade the area to clear away the debris. The continued presence over a lengthy period of time of a leucocytic exudate adjacent to both copper and nickel implants implies that fresh cells were constantly being recruited from the blood stream. H istochemical tests run parallel to the present studies have shown that there are constantly releasing enzymes but these are at best ineffective and, at worst, could lead to increased corrosion and greater toxicity. Monocyte invasion
38
Biomaterials 1981, Vof 2 January
The Biomaterials Silver Jubilee Compendium
15
Metal implants: A. McNamara and D.F. Williams
certainly occurred in the general area of the implants but few of these cells were found in the region to be cleared, that is, directly adjacent to the metal.
Cobalt The most noticeable feature here is the presence of many clumps of a lymphoid nature in between the muscle fibres near the implants. These were composed mainly of plasma cells, a type of activated lymphocyte normally seen in association with an antigenic stimulus. The presence of lymphoid cells indicates some form of immunological activity, and plasma cells in particular indicate immunoglobulin formation 18. Proteins native to the rats are rendered 'foreign' by contact with the metal surface, either being adsorbed onto the metal surface or suffering a change in their configuration. In this way the cobalt metal acts as a hapten and becomes antigenic. Crystalline cobaltprotein complexes have been observed on the surface of many of the cobalt implants in this series ~9. Both cell mediated and humoral responses are thought to occur in this type of immunological reaction. It is interesting to note from the above descriptions that a reproducible response pattern is evoked by each of the implanted metals. This implies that the nature of the ions released, and probably the rate of their release, in each case determines the response. The physical shape and dimensions of an implant are obviously important as well 2~ but their constancy in these experiments indicates that the chemical aspects play an equal if not greater role. A wide variety of serum proteins exist in vivo and it is reasonable to assume that many of these become rapidly adsorbed onto the metal surfaces upon implantation. Since different metals appear to interact differently with proteins in vitro it is also reasonable to assume that the properties of the metals will govern which proteins preferentially adsorb on the surfaces. It is worth noting, therefore, that for much of its term in the host, the surface presented to the environment by the implant will be altered from its original state, the extent of this alteration depending on the metal. Scanning electron microscopy of the specimens in this series, details of which will be reported elsewhere, confirm this variable but modified nature of the surfaces. As noted previously, in some of the metals, notably of copper, nickel and cobalt, extensive corrosion took place. In all cases examined, this resulted in loosening of the discs, which were freely mobile within the surrounding tissue and easily excised. Some cobalt specimens exhibited large scale corrosion and were always loose. Where there was no evidence of gross corrosion, as with aluminium, lead and some of the cobalt specimens, the discs were held firmly in place and proved difficult to excise, ine area of tissue affected by the reaction was smaller than that associated with corroding specimens. One might expect the implantation of a known toxic metal would lead to significant disturbances of the adjacent tissue. It is clear from these experiments, however, that this is not necessarily the case. In vitro lead is known to alter mitochondrial structure or oxidation processes, or both, in a number of systems 2~. In vivo damage to the kidneys occurs as a result of
chronic exposure to lead. Reduction in life span and growth, and impaired functioning of the haematopoetic and central nervous system are also common features in lead toxicity. Nevertheless animals bearing lead implants for up to 52 weeks showed no external signs of any toxic effects and displayed very little evidence of pathological changes in the area immediately adjacent to the implanted discs. it is possible, as noted earlier, that the above observation is merely a dose-related effect; that is either there are insufficient lead ions being released from the implant or the ions are rapidly removed from the site. In this context it is important to note the observations made with cobalt, copper and nickel, where higher rates of ion release were likely. Around some of the cobalt implants a number of lymphoid clumps were observed. It was apparent from the cell types present that immunoglobulin synthesis was taking place, probably as an indirect effect arising from metal-protein complexes formed near the implant. Presumably the long term effects of this process would be host sensitisation. Such a situation has been observed clinically 22 with cobalt-based alloys and cobalt is a known sensitizer. In this case, therefore, it would appear that the implanted metal is having an effect that one might expect from its usual behaviour. Also, in the later stages of the experiments some of the animals developed signs of respiratory distress which could be symptomatic of an impaired ability to deal with common pathogenic challenges. Although all laboratory rodents are susceptible to this condition in old age, the incidence amongst the experimental animals, especially those implanted with copper and nickel, was noticeably greater than in other animals of the same colony at the same age. These observations prompted a series of antigen challenge experiments to determine whether immune competence was impaired within this group. Preliminary results support this idea; the results of more detailed experiments will be reported later. It seems, therefore, that systemic distribution of the more rapidly corroding metals does result in an impaired response to antigenic challenge. Whether this is at the level of recognition or at the stage of antibody synthesis is not certain. Studies of the soft tissue vasculature associated with implanted materials are rare although it is a general concensus that modern prosthetic materials develop a virtually avascular fibrous capsule around them in vivo. Although this was the situation found with lead in the present study, an increase in the number of blood vessels became apparent with other metals, most noticeably with copper and nickel. There was no decrease in this number with time, implying a long term or permanent change in the vascutature associated with these particular metals. A common feature in the response to exogenous foreign bodies is the presence of giant cells 23. Large numbers of these were observed in association with the sutures used in the operative procedure. However, no such cells were produced in response to any of the implanted discs. In contrast, large numbers of haemosiderin-laden macrophages were present around all metals except lead. Winter 2~ reported haemosiderin-like granules in association with stainless steel implants; VernonRoberts and Freeman 2s found large amounts of iron in
Biomaterials 1981, Vot 2 January
39
16
The Biomaterials Silver Jubilee Compendium Metal implants: A. McNamara and D.F. Williams
macrophages of the tissue reaction to both cobaltchromium and stainless steel and haemosiderin-like granules in response to cobalt-chromium and Williams and Meachim 26 reported haemosiderin in the reaction to titanium and stainless steel implants. The most likely causes of iron deposition would be haemoglobin breakdown after bleeding at operation or repeated trauma to the tissue in the presence of the implant. In the present study the number of iron-containing macrophages has been found to decrease with time. Whilst this would imply that the actual insertion of the implant was the cause, the numbers did vary from one metal to another and, specifically, no haemosiderin was observed around the lead. It could be significant here that lead salts have been used medically to relieve the congestion of inflammation on the basis that lead ions have an astringent action upon the blood vessels, thus reducing bleeding. Over the time course studied, a persistent polymorphonuclear leucocyte exudate was present around the copper and nickel implants, in a similar manner to abscess formation with a persisting infection. With an infection an initially extensive necrosis occurs, accompanied by massive infiltration of polymorphonuclear leucocytes, many of which would be killed by the infective agent. Proteolytic enzymes released by these dead cells softens the necrotic material, resulting in a fluid. If left undisturbed, the watery part of the fluid becomes absorbed and the exudate becomes more viscous. The metals act in a similar way by supplying a source of constant irritation and a progressive series of events similar to the above has been observed in these experiments. There was no evidence of bacterial infection of the implant sites in any of the experimental animals when examined at sacrifice. Introduction of some bacteria into the wounds at operation would be expected in a proportion of the cases, in spite of the conventional aseptic methods used. It would seem, therefore that at this level normal post-operative bacterial clearance processes are functioning within the rats in spite of the presence of metal discs in the clearance zone.
CONCLUSIONS The following conclusions may be drawn from these experiments. 1. The response to each of the metals upon implantation was different but reproducible. 2. If corrosion visibly occurs on the metal the implant becomes loose, but if there is no corrosion, the reaction zone is more compact and the implant is firmly retained in the tissue. 3. It is possible for a known toxic metal to produce little response in vivo when implanted intramuscularly. 4. It is also possible, however, for implanted metals to elicit an immune response, either directly or indirectly and preliminary results indicate that chronic systemic metal ion release may cause the host to be more susceptible to disease. 5. Implanted metals may cause a long term or permanent change in the immediate vasculature. 6. Giant cells have not been a feature of the response to metal implants in this series.
40
Biomaterials 1981, Vot 2 January
However, haemosiderin-laden macrophages were present in the tissue adjacent to all metals except lead, but to varying degrees. 7. A persistent polymorphonuclear leucocyte response is possible in the absence of bacteria, especially in association with a high level of corrosion. 8. Any bacteria introduced at the time of surgery appeared to be cleared normally from the site without hindrance from the metal, implying that the metals do not impair the response to acute post-operative infection.
ACKNOWLEDGEMENT The work reported here forms part of a large programme of research funded by grants from the Medical Research Council (G 977/237/S) and the Science Research Council (GR/A63498), to whom our gratitude is expressed. One of us (A. McN.), is the recipient of a Medical Research Council Studentship.
REFERENCES 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
20 21 22 23 24 25 26 27 28
Zierold,A.A., Arch. Surg., 1924, 9, 364 Hey-Groves,E.W., Brit. J. Surg., 1913, 1,438 Williams,D.F.,Ann. Rev. Mater. ScL, 1976, 6, 237 Williams,D.F., Fundamental Aspects of Biocompatibility, Vol. 1, (Ed. D.F. Williams) CRC Press. Boca Raton, In press Laing,P.G., Ferguson, A.B. and Hodge, E.S.,J. Biomed. Mater. Res., 1967. 1, 135 Escalas,F., Galante, J.. Rostoker, W. and Coogan, P., J. Biomed. Mater. Res., 1976, 10, 175 Meachim,G. and Williams, D.F., J. Biomed. Mater. Res., 1973, 7, 555 Winter,G.D., In Biocompatibility of Implant Materials, (Ed. D.F. Williams), Pitman Medical, London 1976, p 28 O'Gara,R.W. and Brown, J.M., J. Natl. Cancer Inst., 1967, 38, 947 Waldren,H.A., Brit. J. Ind. Med., 1966, 23, 83 Schroeder,H.A., Balassa, J.J. and Vinton, W.H., J. Nutr., 1965, 86, 51 Gross,S.B., Pfitzer, E.A.. Yeager, D.W. and Kehoe, R.A., ToxicoL AppL Pharmacol. 1975, 32, 638 McNary,W.F., Anat. Rec,, 1963, 146, 193 Johnson,B.C.,Am. J. OphthalmoL, 1973, 76, 455 Brady,R.O., IN The Metabolic basis of inherited disease, (Ed. J.B. Stanbury, J.B. Wyngaarden and D.S. Fredrickson. McGraw Hill, New York, 1978, p 731 Balls,J.V., Haust, M.D. and More, R.H.. Expt. MoL PathoL, 1969, 3, 511 Meachim,G., In Fundamental aspects of Biocompatibility, Vol. 1, (Ed. D.F. Williams), CRC Press, Boca Raton, In press Weiss,L., The Cells and Tissues of the Immune System: Structure, Functions and Interactions, Prentice Hall, New York, 1972, p 151 McNamara, A. and Williams, D.F., Eng. in Med., 1981, In Press Bischoff,F. and Bryson, G., In Progress in Experimental Tumour Research, Vol. V.. (Ed. F. Hamburger), Karger, Basel and Hafner PubL, p 85 Fowler,B.A. and Mahaftey, K.R., Environ. Health. Perspectives, 1978, 25, 87 Evans,E.M., Freeman, M.A.R., Miller, A.J. and VernonRoberts, B., J. Bone Jr. Surg., 1974, 56(B), 626 Mariano,M. and Spector, W.G., J. Pathol., 1974, 113, 1 Winter,G.D.. J. Biomed. Mater. Res., Syrup. 8, 1974, 11 Vernon-Roberts,B. and Freeman, M.A.R.. In Scientific Basis of Joint Replacement, (Ed. S.A.V. Swanson, and M.A.R. Freeman), Pitman Medical, London, 1977, p 86 Williams,D.F. and Meachim, G., J. Biomed. Mater. Res., Symph. 5(1 ), 1974, 1 Reznik,M. and Engel, W.K., J. Neurol. Sci., 1970, 11, 167 Carlson,B.M. and Gutmann, E.,Anat. Rec., 1975. 183, 47
17
The Biomaterials Silver Jubilee Compendium
0sseointegrated titanium fixtures in the treatment of edentulousness P-I. Br nemark, R. Adell, T. Albrektsson, U. Lekholm, S. Lundkvist and B. Rockler Laboratory of Experimental Biology, Dept of Anatomy, Dept of Oral Surgery, the Dept of Stomatognathic Physiology & Dept of Oral Roentgen Diagnostics, University of Gothenburg and the Institute for Applied Biotechnology, Gothenburg, Sweden. This paper was first presented at a Symposium: Head, Neck and Dental Implants held in Amsterdam on 2-3 October 1981
A 91 per cent positive 5-9 year result has been reported when using titanium implants and gold bridges to restore edentulous jaws. About 400 consecutive patients have been operated. The reasons for the good results are believed to depend on the anchorage of the implants in the living bone without interposing soft tissue layers. Repeated X-rays ensuring a strict parallelism are used to indicate direct bone integration. Some implants had to be removed in spite of still being anchored in the bone. In these cases SEM and TEM provided direct evidence of an osseointegration. Keywords: Titanium, osseo-integration, dental implants, clinical study.
A review of the osseointegration system for clinical jaw reconstruction was published in 19771 . In recent publications the 5-9 year success rates of the total material of 3000 osseointegrated implants inserted into edentuious jaws of 400 consecutive patients in Gothenburg, Sweden, were determined as 81% in the maxilla and 91% in the mandible. The method has been evaluated at other Universities, e.g. in Toronto, where success rates slightly superior to those of the Gothenburg material have been reported 5. In fact, the osseointegration method is the only procedure described for dental restoration in edentulous cases which fulfills all demands suggested by the Harvard conference 8 for a functioning dental implant system. We believe there are two principal reasons for the favourable results arrived at: 1. The establishment of a biological seal around the abutments penetrating the soft tissue, preventing possible descending inflammatory reactions around the implant superstructure to reach down to the level of the bone. 2. The establishment of osseointegration which is defined as direct contact between living, haversian bone and implant. However, the concept of osseointegration is not accepted as valid for metallic implants by everyone. Many authors are of the opinion that osseointegration is possible only with non-metallic implants 7-1~ The purpose of this paper is to give a summary of indications and evidence that a direct bone to titanium implant contact in fact is possible and may last for years at load-bearing implants piercing the gingiva.
MATERIAL AND METHODS Clinical material The total material of the Gothenburg clinic today amounts to about 3250 fixtures which have been inserted into edentulous jaws of more than 400 patients.
The maximal follow-up time is 16 years. The patient characteristics are middle age (average 53 years), women to men ratio of 60 to 40 and presence of removable dentures in the opposite jaw in 60% of the cases. 94% of the patients were totally endentulous and the majority of those had severely resorbed alveolar ridges.
The implant and dental bridge The implant fixture (Figure 1) was manufactured from pure (99.7%) titanium. Using a gentle surgical technique, 4-6 fixtures were inserted into each jaw to be operated. The soft tissues were sutured over the fixtures. After a period of at least 3-4 months during which the patient wore ordinary dentures, the mucosa was again opened and abutments were connected to the fixtures. These abutments were aimed for permanent penetration of the mucous membrane and were later connected to each other via a dental bridge (Figure 2) made of gold and acrylic.
Clinical evaluation of osseointegration All fixtures were clinically checked for stability at the abutment operation. The stability check was performed by a manual test to reveal possible instability in the fixation of the fixture.
Radiographical evaluation of osseointegration X-rays were taken at the abutment operation and at the controls, at least once annually. A method for serialidentical stereopair roentgenograms with a parallel projection has been developed by Rockier 11. The interfacial region between the titanium and bone was carefully examined on all radiograms. Also, bone height measurements were performed as described by Hollender and Rockier 12. A detailed analysis of the X-ray image of the interface was performed using an IBAS I and il data unit.
9 1983 Butterworth & Co (Publishers) Ltd. 0142-9612/83/040025-04 $03.00 Biomaterials 1983, Vol 4 January
25
The Biomaterials Silver Jubilee Compendium
18 Titanium bone implants: P-I. Br&nemark et al.
Histology, SEM and TEM for evaluation of osseointegration Fifteen fixtures have had to be removed in spite of being clinically and roentgenologically osseointegrated. The reasons for removal have been patient factors such as psychiatric disorders or implant factors such as fracture of the fixture. After 2-4 years of loading 5 such fixtures were removed by aid of a trephine ensuring an intact bone cover around them and were then fixed in formaldehyde. The bone was decalcified with formic acid and staining was performed using the HTX-eosin technique. ground sections and cross sections were examined by light microscopy. Another 10 implants were cut in a similar manner after 2.5 to 7.5 years of loading. They were placed in 3% glutaraldehyde and after conventional treatment including gold sputtering the implants were analysed in a Scanning Electron Microscope (Jeol JSM-35). In one case it was possible to cut obliquely through the intact interface between bone and implant and perform observations by Trans Electron Microscopy, TEM. Ion probe analysis (EDAX) was used to examine the implant surface for any traces of foreign material on the fixture surfaces.
RESULTS Figure 1 The titanium fixture is a threaded implant which is to be anchored in the bone. There is a central screw which connects the abutment with the fixture.
Figure 2 in each j a w four to six implants are inserted and connected to a dental bridge
26
Biomaterials 1983, Vol 4 January
Clinical and radiological indications of osseointegration As a rule, the clinical control at the abutment operation revealed a stable fixture which was not moveable with the manual test. Such stable fixtures coincided with the X-ray appearance of a normal bone with an intimate contact with the fixtures at the resolution level of radiography. In those cases where the fixtures were manually moveable, there was a thin, roentgenologically verified, radiolucent zone around the implant. Figure 3 is a radiogram of a stable fixture and Figure 4 represents unstable ones. The functioning implants became surrounded by a bone tissue which was condensed immediately adjacent
Figure 3 Radiogram of stable implants four years after insertion. After the first year has passed the annual loss in bone height is <0.1 ram.
The Biomaterials Silver Jubilee Compendium
19
Titanium bone implants: P-I. Br&nemark et al.
(a)
few fixtures were lost also during the second year of implantation.
Histological-, SEM- and TEM-observations
(b)
Histological sections obtained from stable fixtures showed no interposed soft tissue, a remodelled cortical bone and a healthy marrow tissue. Mobile fixtures were found to be surrounded by a sheath of soft tissue. Scanning Electron Microscope analyses of stable fixtures, removed after 2.5-7.5 years of function, revealed adhering healthy cells in the interface zone. Calcified tissue (Figure 5) was seen in such an immediate contact with the bone that the interpositioning of soft tissue layers can be ruled out. This direct bone-toimplant contact occurred all around the implant surface. Ion probe analysis (EDAX) of the fixture and surrounding osseous tissues, showed calcium on the fixture surface and no titanium on the bone surface. By oblique cutting through the undecalcified bone and the margin of the implants it was possible to obtain TEM-pictures of the intact interface zone (Figure 6). The TEM-analysis also showed a direct bone-to-implant contact without interposed soft tissue layers. According to staining reactions to lanthanum and Alcian blue the very interface of osseointegrated implants seems to consist of a proteoglycan layer.
DISCUSSION Fixtures may be lost through three major mechanisms: 1. Initial healing may occasionally occur with a soft tissue capsule. 2. An achieved osseointegration may be lost due to repeated overloading. 3. A gradual apical migration of the marginal bone
Figure 4(a) The implant marked with an X is surrounded by a thin radiolucent zone, a clear radiological sign of failure. (b) Another implant (X) with radiological signs of loosening. Altogether 9 per cent of mandibular implants were lost over a follow-up period of 5-9 years.
tO the titanium surface. This bone corticalization was directly visible in radiograms of about 10% of the implants. Using the IBAS picture analyser, however, the presence of bone condensation adjacent to the implants was revealed on X-rays from the great majority of the inserted implants. This bone condensation around the implants was becoming more evident in radiograms obtained after longer times of implantation in comparison to short time followed-up implants. The full details of the IBAS-analysis will be published separately. The fixture losses, as a rule, occurred during the first year after implantation. Mandibular fixtures showing clinical and radiological evidence of osseointegration at the control one year after fixture installation were all still in place and stable 5 years later. In the maxilla a
Figure 5 SEM-image of calcified tissue (Ca) in intimate contact with the titanium (Ti) oxide surface.
Biomateriats 1983, Vol 4 January
27
20
The Biomaterials Silver Jubilee Compendium Titanium bone implants: P-I. Br&nemark et al.
woven bone should be present in the interface. Albrektsson et aLlS in a rabbit experimental study compared titanium coated implants with gold coated ones. In the
titanium group there was haversian bone in contact with the implants at a resolution level of 30,~. In the gold group there was a cellular layer separating the bone from the inorganic material, and, furthermore, the bone in a several microns wide zone of the interface was of a disordered, woven character. In conclusion, the resolution level of single radiograms, at least when examined by the naked eye, would not permit a definite statement as to whether osseointegration has occurred or not. However, the clinical and radiographical analyses of the same implants performed at various times after implantation, combined with the numerical data from the IBASinvestigation, are definitely indicative of osseointegration. Undisputable evidence for such an osseointegration were found in SEM- and TEM-analyses of removed stable fixtures. The favourable results observed in our study with the osseointegrated implants, which all have shown excellent clinical function, in comparison to implants clinically or radiographically diagnosed as anchored in soft tissue, which with no exceptions have been lost, provide strong support for the opinion that a direct bone anchorage is of great import,~nce for the long term function of dental implants. Figure 6 TEM-analysis o f intact bone (B) to titanium (Ti-d) interface after 30 months of implant function. To obtain this picture an oblique cutting was used through the bone and the marginal titanium layer.
REFERENCES 1 2
level may anchorage.
deprive
the
fixture
of
its
bone
In our clinical study we found radiographical and clinical signs of soft tissue anchorage to coincide with later implant loss. On the other hand, when there were clinical and radiological signs of an osseointegration excellent implant function did ensue. The radiographs showed a clear tendency towards an increasing bone density in the immediate surroundings of the implant with increasing time of implantation. We believe this gradual corticalization to be indicative of a successive load-adapted bone remodelling taking place in the interface region. In a study of the surface characteristics of titanium implants which had been inserted in the human jaw for periods of up to 8 years, McQueen et al. 13 found the oxide layer on the titanium implants to increase in width from around 50 A before implantation to about 2000 in width after 8 years of clinical function. Both these observations indicate that a titanium implant inserted in the human jaw may have an active interaction with the tissues. There is a remodelling not only of the bone, but also of the titanium oxide surface, the possible interrelationship of which, however, is presently difficult to define. The matter is further discussed in Albrektsson et
3
4
5 6 7 8 9 10 11
12
al. 14
A radiogram indicating osseointegration at one year after fixture installation gives a very good predictability of the long-term results. This is also indicated by the radiographically observed annual bone loss which, after the first year has passed, has been less than O.1 m m 2. It is our opinion that a direct bone contact per se is not a guarantee for long term implant function. The quality of the interface bone is probably also of great importance and preferrably haversian bone instead of
28
Biomaterials 1983, Vol 4 January
13 14 15
Br~nemark,P-I., Hansson, B-O., Adell, R., Breine, U., Lindstr6m, J., Hall~n, O. and Ohman, A., Osseointegrated implants in the treatment of the edentulous jaw. Scand. J. Plast. Reconstr. Surg. 1977, suppl 16, 1-132. Adell,R., Lekholm, U., Rockier, B. and Br&nemark, P-I., A 15-year study of osseointegrated implants in the treatment of the edentuIous jaw. Int. J. Oral Surg. 1981, 10, 387-416. Albrektsson,T., Br~nemark, P-I., Hansson, H-A. and Lindstr6m, J., Osseointegrated titanium implants. Requirements for ensuring a long-lasting, direct bone-to-implant anchorage in man. Acta Orthop. Scand. 1981, 52, 155-170. Br~nemark,P-I., Adell, R., Albrektsson, T., Carlsson, G-E., Haraldson, T., Lekholm, U., Lindkvist, L., Lindstr0m, J., Lundkvist, S. and Rockier, B., Osseointegratedtitanium implants in the rehabilitation of the endentulous patient. Advances in Biomaterials 1982, 4, 133-141. Br&nemark,P-I., Albrektsson, T., Skalak, R., Symington, J. and Zarb, G. Osseointegrated dental implants. Transactions 8th meeting of the Society for Biomaterials and the 14th International Biomaterials symposium 1982, V, 132.
Schnitman,P.A. and Schulman, LB. Dental implants: Benefit and risk. US Department of Health and Human Services 1980, No 81, 1531, 1-351. Jacobs,H.G. Implantologie und Zahnersatz, 1976, Hanser, MLinchen. Jacobs, H.G. Formgestaltung und Materialfrage bei enossealen Implantaten zur Aufnahme yon Zahnersatz. Dtsch. Zahn~rztL Z. 1977, 36, 63-69. Muster, D. and Champy, M., Le Probl~me d'interface osbiomat&iaux. Actu/itds Odonto-Stomat. 1978, 121, 109-124. Osborn,F. and Newesly, H., Dynamic aspects of the implant-bone interface, in Dental implants, materials and systems, (Ed G. Heimke) Hanser, Mfinchen 1980, 111-123. Rockier,B. A 15-year study of osseointegrated implants in the treatment of the edentulous jaw with special reference to roentgenological examinations. To be published as Thesis, University of Gothenburg.
Hollender, L. and Rockier, B., Radiographic evaluation of osseointegrated implants of the jaws. Dentomaxillofac. RadioL 1981, 9, 91-95. McQueen,D., Sundgren, J.E., Ivarsson, B., Lundstr0m, I., Af Ekenstare, B., Svensson, A., Br~nemark, P-I. and Albrektsson, T. Auger electron spectroscopic studies of titanium implants. Advances in Biomaterials 1982, 4, 179-185. Albrektsson,T., Br~nemark, P-I., Hansson, H-A., Kasemo, B., Larsson, K. and Lundstr~m, I. The interface zone of inorganic implants in vivo: Titanium implant in bone. Annals of Biomedical Engineering, in press 1982. Albrektsson,T., Br~nemark, P-I., Hansson, H-A., tvarsson, B. and J0nsson, U., Ultrastructural analysis of the interface zone of titanium and gold implants. Advances in Biomaterials 1982, 4, 167178.
21
The Biomaterials Silver Jubilee Compendium
Biomaterial biocompatibility and the macrophage J mes M. _Andersonand Eat een M. Miner
Departments of Pathology and Macromolecular Science, Case Western Reserve University, Cleveland, Ohio 44106, U.S.A.
The biocompatibility of biomaterials at implant sites is controlled by the tissue/material interaction. A major cell in the tissue reaction is the macrophage. A summary is presented on macrophage mediation of cellular and humoral regulatory pathways in inflammatory and immune responses. Keywords: Biocompatibility, macrophage, inflammation, mediators, cell-cell interaction
In general, the biocompatibility of a given material with tissue has been described in terms of the acute and chronic inflammatory responses and the fibrous capsule formation which is seen over various time periods following implantation of the respective material 1'3. H istologic evaluation of tissue adjacent to implanted materials as a function of implant time has been the most commonly used method of evaluating the biocompatibility of the material 4-7. Classically, the biocompatibility of an implanted material has been described in terms of the morphological appearance of the inflammatory reaction to the material but the inflammatory response itself is a series of complex reactions involving various types of cells whose densities, activities, and functions are controlled by various endogenous and autocoid mediators. In considering compatibility and its relationship to the in vivo elements that follow the implantation of a material, an appreciation of the temporal events of wound healing that follow implantation is necessary 8'9. Healing by second intention is initiated when blood clot formation occurs in this space around the biomaterial after implantation. With clot formation and contraction, the first phase of the acute inflammatory response is initiated by permeability changes in the adjacent vasculature which result influenced by other defence mechanisms such as the mediators into the area of injury. The initial clot formation and contraction and the permeability changes may be influenced by other defence mechanisms such as the extrinsic and intrinsic coagulation systems, the complement system, the fibrinolytic system, the kinin generating system, and platelets. Following the complex interaction of the components of these systems, neutrophil migration into the wound site and exudate occurs. The preferential migration of neutrophils mediated through a chemotactic stimulus, is a well-known characteristic of the acute inflammatory response. Monocyte migration into the wound site then occurs, and these cells differentiate into macrophages, the cells principally responsible for normal wound healing; fibroblast proliferation with collagen deposition and capillary proliferation then follow.
Our efforts over the past several years have been directed toward developing a better understanding of the biological phenomena which can occur with polymeric implants1~ In considering the events that follow implantation of a biomaterial, the macrophage and its role in the development of granulation tissue and subsequent wound healing is not well understood. Macrophages are longlived cells which may reside in tissues interfacing with implants for extended periods of time. An excellent example of this is the observation that implanted fabrics, i.e. vascular grafts, have an encapsulating macrophage/ granulation tissue response for years ~ . The following is a brief review of various mechanisms by which macrophages migrate to implant sites, oxygen metabolism in macrophages, effectors and inhibitors of macrophage activity and products of macrophage activation.
MAJOR MEDIATORS OF MACROPHAGE CHEMOTAXIS The accumulation of macrophages in implant sites is achieved through selective cellular and humoral mechanisms that both attract and immobilize macrophages and macrophage precursors where their functional activity can be directed at resolution and/or repair of tissue. Monocytes contact stimulatory factors within blood vessels around the implant site and adhere to the endothelial cells lining the vessels. These monocytes then pass through the vessel wall~and into the tissue, undergo transformation in response to the stimuli and differentiate into macrophages. Monocyte and macrophage movement within the tissue toward the site of injury is controlled and directed by various agents. This movement is defined as either chemotaxis or chemokinesis. Chemokinesis is the accelerated random locomotion in response to and the speed of turning of the cells toward chemical stimuli, whereas chemotaxis is the highly directed movement of the cells along a chemical gradient 12-14. Of the various agents which mediate chemotaxis (Table 1), most are integral features of the inflammatory process and can both sustain and control the severity of the inflammatory res-
91984 Butterworth Et Co (Publishers) Ltd. 0142-9612/84/O10005-06503.OO Biomaterials 1984, Vol 5 January
5
22
The Biomaterials Silver Jubilee Compendium Biocompatibility: J.M. Anderson and K.M. Miller Table 1. ==
Major Mediators of Macrophage Chemotaxis .
.
.
.
.
.
,
A. B. C. D.
Complement Fragments Lymphokines Fibronectin Fragments Leukotriene B4 .ll
ponse. In addition to dead or dying cells, chemotactic stimuli include fragments derived from the activation of complement protein, mediators of the kinin, clotting and fibrinolytic systems, and products produced by leukocytes themselves 15. The complement system is a major biological source of chemotactic activity. Complement itself is not a single protein but a complex and extensive ,series of glycoproteins and protein inhibitors. Its activation can be initiated by antigen-antibody complexes, bacterial polysaccharides, endotoxins and certain polymeric structures. Once initiated, activation proceeds along either of two pathways, classical or alternative, through a series of cascading enzymatic events generating components which mediate membrane damage directly or fragments which mediate inflammatory or immune processes. The best characterized complement derived chemoattractant is the C5a fragment. It is chemotactic for all leukocytes except lymphocytes. Although C3a has been reported to have chemotactic activity, it is now considered that this activity was due to contaminating C5a fragments in the preparation, since purified C3a cannot be shown to be chemotactic. The C567 complex is chemotactic for neutrophils, and a Factor B complex, generated via the alternative pathway, also has chemotactic activity 14'15. Lymphokines, soluble mediators derived from sensitized lymphocytes, can reduce leukocyte emigration and lead to accumulation of macrophages at an implant site. Monocyte chemotactic factor (MCF) attracts macrophages while macrophage migration inhibition factor (MIF) immobilizes the macrophages at the site of injury. These two factors may be different manifestations of the same molecule since high concentration of chemoattractants can immobilize cells. Other factors such as macrophage activating factor (MAF), macrophage fusion factor and specific macrophage arming factors then serve to activate the captive macrophages to better mediate the inflammatory response and interaction with the implant. It is not yet clear whether al of these lymphocyte factors are separate molecules or varying activities of a single molecule. Fibronectins are a group of glycoproteins derived from fibroblasts, monocytes and endothelial cells that are important for cellular adhesion to material surfaces and to biological surfaces such as fibrin and collagen ~6. As such, they are proteins prominently associated with inflammation and wound healing. Specific fibronectin fragments generated by endogenous proteases are potent chemoattractants for human blood monocytes, while the intact molecule is not chemotactically active ~7. These fragments show both chemotactic and chemokinetic activity only for peripheral blood monocytes and not for lymphocytes or polymorphonuclear leukocytes. The specificity of potency of these fragments suggest that the proteolytic cleavage of fibronectin during the inflammatory process produces moieties that selectively augment the recruitment of blood monocytes into tissue sites of inflammation. Oxidized lipid components of cell membranes have
6
Biomaterials 1984, Vol 5 January
long been known to be chemotactic 18. More recently, the stable products of arachidonic acid metabolism, prostaglandins, thromboxanes and leukotrienes, have been examined in greater detail and only Leukotriene B4 has been shown to exhibit chemotactic activity. Leukotriene B4, an intermediate in the lipoxygenase pathway of arachidonic acid metabolism, also provokes inflammation and augments leukocyte adhesion, aggregation and lysosomal enzyme release as well 19. In addition to these major chemotactic factors, there are several other products that have similar activity. Most are protein degradation products associated with the inflammatory process. Fibrinopeptide B, a fibrin degradation product, and Kallikrein, a clotting cascade enzyme, both have chemotactic activity. Synthetic chemotaxins based on bacterial peptides are convenient in in vitro systems. They are synthetic formyl methionine peptides such as f- Met-Leu-Phe and are chemotactic for macrophages at very low concentrations 15.
OXYGEN METABOLISM IN MACROPHAGES Phagocytosing macrophages undergo an associated 'respiratory burst' much as phagocytizing neutrophils do 2~ During this increase in metabolic activity molecular oxygen is taken into the cell from the surrounding medium and almost entirely converted by the action of an NADPHoxidase to H 2 0 2 through a superoxide anion intermediate. These oxygen metabolites and the products of their interaction, in particular the potent oxidant hydroxyl radical (OH), are essential to the principal physiological function of phagocytes: the elimination of invading microorganisms. These highly reactive oxygen metabolites are generated upon contact of the cell plasma membrane with any of a number of surface-active materials or soluble substances that will induce phagocytosis or membrane perturbation 15'2~ Some examples of such materials are particulates such as bacterial and other microorganisms, or latex, silica, carbon, or polymeric particles. Attachment of macrophages to large areas such as glass or polymer surfaces will trigger an attempt at phagocytosis culminating in all the manifestations of normal particle phagocytosis; i.e. toxic oxygen matabolite production and lysosomal enzyme release. Macrophages which undergo activation prior to phagocytosis demonstrate a markedly enhanced capacity to generate these highly reactive metabolites. This increase production of superoxide anion and hydroxyl radical as well as released lysomal hydrolases by activated macrophages will not only contribute to improved microbial activity of the macrophages, but also augment arly tissue injury due to the invading organisms. These highly reactive radicals generated by cellular mechanisms at or near the surface of implanted polymers may contribute to damage of the polymer surface in the same fashion as established polymer degradation reactions by rective radicals 21. Table 2
Oxygen Metabolism in Macrophages .
.
.
.
.
.
.
A. Phagocytosis-Associated Respiratory Burst Similar to Neutrophils B. Mediators of 02 Metabolism 1. Phagocytic particles or surface 2. Soluble membrane perturbers
The Biomaterials Silver Jubilee Compendium
23 Biocompatibihty: J.M. Anderson and K.M. Miller
EFFECTORS OF MACROPHAGEACTIVATION Cytokines are soluble hormone-like factors produced by a wide variety of cell types including lymphocytes, monocytes, platelets, fibroblasts and keratinocytes. These secreted proteins exert multiple biological effects on various target cells and regulate immunological and inflammatory host responses by serving as intracellular messengers that modulate cellular functions 22. Under the influence of lymphokines, i.e. cytokines released by activated T lymphocytes, macrophages show various features of activation. Morphologically, the cell increases in size and in the number of granules and the plasma membrane becomes more ruffled. Activated macrophages also increase their biosynthetic functions and release various materials such as lysosomal enzymes, Interferon, leukocytic pyrogen, cytotoxic agents and prostaglandins. The affinity and the number of membrane receptors for cytophilic antibodies is increased and changes in intracellular metabolism also occur ~5. The principle soluble mediator of macrophage activation is macrophage activating factor (MAF), also known as Interleukin 2 (11_2), a T cell derived lymphokine. In response to antigen binding and I L2 the macrophage will synthesize its own cytokine, Interleukin 1 (ILl). This further activates T cells to produce more I L2 to activate more macrophages in a feed-back cycle. Colony-stimulating factor (CSF) is a cytokine with more than one distinct activity. CSF is produced by activated T cells, macrophages, granulocytes and fibroblasts. Initially it was demonstrated that CSF stimulates the proliferation of granulocyte precursor stem cells in the bone marrow. At high concentrations, CSF not only stimulates the proliferation of the phagocyte precursors but favors their differen.iation into granulocytes, whereas at low concentrations differentiation into monocytes is favored. More recently, purified preparations of CSF have been shown to activate macrophages to produce other biological activating molecules including prostaglandins, neutral proteases, Interferon and lnterleukin 1 22,23 Phagocytosis by macrophages can proceed via a receptor mediated action. Macrophages have on their surfaces receptors for the complement cleavage fragment C3b and the Fc portion of the immunoglobulin molecule. Lymphokines have been described that enhance both the C3b receptor and the Fc receptor mediated phagocytosis 24. However, it is not clear whether this phagocytic activity is enhanced by the tymphokine through increased receptor expression or by some other mechanism. Leukocyte aggregation is a well described response of PMNs to chemotactic factors, especially CSa 25-29. Table 3.
Effectors of Macrophage Activation
_
A. Cytokines 1. Interleukin 2 [IL2) 2. Colony-Stimulating Factor (CSF) 3. Heat-StableLymphokine 4. MacrophageAggregating Factor B. Complement Cleavage Products C. Immune Complexes D. Interferon 1. Type 1 -- Macrophage Derived 2. Type2 - Lymphocyte Derived E. Mediators of Membrane Perturbation 1. Soluble Compounds 2. Particulates F. Prostaglandins
Leukocyte aggregates have been implicated in vascular injury in many non-immune inflammatory disease, especially in situations where complement becomes activated by contact with polymer surfaces as in hemodialysis 26'29. Macrophage aggregating factor, another lymphokine, has been described as inducing macrophage aggregation which results in a similar form of vascular injury 25. Macrophages are also activated by complement cleavage fragments, specifically C3b and C5a. Macrophages have receptors on their surface for both of these fragments; thus, activation is mediated by receptor-ligand binding. Binding of C3b to its receptor perturbs the macrophage membrane leading to consequences of activation. C3b binding also acts as a stimulus for increased procoagulant or thromboplastin activity in human peripheral monocytes 15. The existence of specific C5a receptors has been demonstrated in macrophage membranes. The binding of C5a to these receptors results in the augmentation of the primary humoral immune response through the induction of Interteukin 1 secretion by macrophages 27 Neutrophil adhesion and aggregation are enhanced by activated C5a fragment 29. Since macrophages are also known to aggregate and since they have specific C5a membrane receptors, it may be expected that C5a also mediates macrophage aggregation. Macrophage aggregation may lead to macrophage fusion and the formation of foreign body giant cells on the surface of implanted biomaterials 3~ It is not uncommon to observe foreign body giant cells at the interface between tissue and retrieved smooth surface implants 32. Macrophage aggregation may also contribute to the role of neutrophil aggregation in the pathogenesis of certain clinical disorders including pulmonary dysfunction that accompanies hemodialysis and the neutropenia of cardiopulmonary bypass. Exposure of circulating neutrophiis and macrophages to C5a stimulus results in the formation of occluding leukocyte aggregates within the microvasculature as well as the release of inflammatory mediators that are toxic to vascular endothelial 28. Among other characteristics, activated macrophages exhibit increased phagocytic activity and upon attachment to a surface, spread more rapidly and extensively than do normal resting macrophages. This spreading has been correlated with increased bactericidal activity in vivo. Activated complement Factor Bb of the alternative pathway will induce human monocytes to undergo increased rapid spreading 33. The attachment and phagocytosis of most particles is increased if the particles are coated with serum components termed opsonins. The opsonins then bind to specific receptors on the phagocyte cell surface. Since macrophages have receptors for both immunoglobulin molecules (Fc-Receptors) and complement fragments (C3b and C4b receptors), specific antibodies and the complement fragments C3b and C4b enhance the phagocytosis of foreign particles. Immune complexes are also major agents in the activation of macrophages 34-36. Immune complexes are aggregates of antibody molecules which mediate macrophage activation by binding to the macrophage surface receptors for immunoglobulin molecules (Fc receptors) and subsequently perturbing the membrane. In modulating the Fc receptors, immune complexes will either inactivate or remove them from the macrophage membrane for
Biomaterials 1984, Vol 5 January
7
The Biomaterials Silver Jubilee Compendium
24 8iocompatibility: J.M. Anderson and K.M. Miller
a time, thereby eliminating their availability for opsoninmediated phagocytosis. Immune complexes can also serve as stimuli for the increased production of thromboplastin activity by normal peripheral blood monocytes and for the enhanced synthesis of prostaglandin E2 by macrophages 15. In addition to its classical antiviral activity, Interferon has other regulatory effects including macrophage activation. There are two major types of Interferon produced, each from a different cell type. Interferon Type 1 is synthesized by activated macrophages and its production is regulated by prostaglandins and colony-stimulating factor (CSF). Prostaglandin E inhibits the synthesis of Interferon while CSF stimulates it is. Interferon Type 1 can feed back and further activate macrophages by enhancing their phagocytic capacity. Interferon Type 2 is produced by sensitized T lymphocytes and it increases the ability of macrophages to limit viral growth. Macrophages are also believed to become activated as a consequence of membrane perturbation. This change within the membrane bilayer can be due to a ligand binding to receptors and subsequent aggregation and/or internalization of receptor-ligand complex. Receptor internalization can then trigger the intracellular metabolic changes observed in activated cells. This mechanism is thought to be the major mediator of cellular activation. Rearrangement of components within the lipid bilayer upon external contact with an activating agent may be an additional mechanism. Most mediators that induce membrane perturbation usually involve receptor binding and phagocytosis. The consequence of phagocytosis of microorganisms by macrophages is well known. Particles such as silica and asbestos, after being phagocytosed, induce ILl production by the macrophage. The subsequent effects of ILl are enhanced fibroblast proliferation generating a fibrogenic effect. Polymeric particles would also be expected to enhance ILl synthesis upon phagocytosis by macrophages. Prostaglandins are stable products of the oxidation of cell-derived arachidonic acid'via the cyclooxygenase pathway. They are important positive and negative regulators of both the inflammatory and immune processes. The synthesis of prostaglandins by macrophages is increased by certain lymphokines, by immune complexes, by colony-stimulating factor (CSF) and by endotoxin. As positive regulators of macrophage activitation, prostaglandins will augment phagocytes and endotoxin-induced collagenase production. Prostaglandins have also been shown to increase the numbers of both immunoglobulinFc receptors and the lectin Concanavalin A receptors on macrophage membranes 37-39.
INHIBITORS OF MACROPHAGEACTIVITY Prostaglandins are important positive and negative regulators of macrophage function. They are elicited by soluble and particulate macrophage activators such as complement C3 components, immune complexes, endotoxin, the ionophore A23187, zymosan particles, colchicine and latex particles. Since prostaglandins activate suppressor lymphocyte mechanisms, lymphokine Table 4 ,,m
production is inhibited and the production of macrophage factors such as Interleukin 1 and Interferon are also inhibited. Exogenous prostaglandins also affect macrophage proliferation and other functions. Prostaglandin E2 inhibits clonal proliferation of the committed granulocyte-macrophage stem cell. Thus, prostaglandins produced by macrophages may negatively regulate their own production by inhibiting the proliferation of the macrophage progenitor cells. Prostaglandins will also inhibit macrophage spreading, adherence and migration 15'37. Glucocorticosteroids, potent anti-inflammatory and immunosuppresive drugs, are also potent inhibitors of macrophage function. In high doses, corticosteroids diminish the production of monocytes. High-affinity receptors for glucocorticosteroids h a v e been demonstrated on human monocytes. Corticosteroids will also inhibit synthesis and secretion of macrophage neutral proteases, expecially plasminogen activator. Interleukin 1 production will also be inhibited by corticosteroids, as will the numerous activities mediated by ILl 4~
PRODUCTS OF ACTIVATED MACROPHAGES Interleukin 1 is a macrophage-derived cytokine that was initially identified by its capacity to increase proliferation of thymocytes. It also induces the production of Interleukin 2 and other factors from activated T lymphocytes. Like endogenous pyrogen, Interleukin 1 can stimulate the hypothalamic cells of the fever centre and induce the synthesis of acute phase proteins (serum amytoid A, fibrinogen, C-reactive protein) from hepatocytes. Recent evidence suggests that ILl, endogenous pyrogen and serum amyloid A (SAA) inducer are biochemically and functionally related molecules 41-44. Interleukin 1 is also an important mediator of the inflammatory process because of its regulation of fibroblast growth and proliferation 45'46. By stimulating fibroblast activity, ILl induces the synthesis of the fibroblast product collagen. However, it has also been shown to stimulate collagenase production by cultured fibroblasts and human synovial cells45-s~ These findings suggest a multiple role for ILl in the regulation of normal tissue repair during chronic inflammation. There must be sensitive mechanisms capable of regulating the duration and extent of fibroblastic activity, and ILl seems to be a good control candidate. Macrophages produce Interferon Type 1. Interferon is a regulator of macrophage activation as well as a regulator of the immune response. Interferon activates macrophages by enhancing phagocytosis. Macrophages have been shown to synthesize numerous complement components in vitro. They are a source of classical components C2, C4, C3 and C5, as well as several alternative components: Factor B, Factor D, C3b-inactivator(I), and/~1 H-globulin (H). The most important of these are C3 and Factor B. C3 is cleaved to C3a and Table 5 .
Products of Activated Macrophages .
.
.
.
,~,j
, ,,
A. Prostaglandins B. Corticosteroids ,
8
,,,
~,=,,
,
Biomaterials 1984, Vol 5 January
.
.
A. B. C. D. E. F. G.
inhibitors of Macrophage Activity j
.
,,,,,,
,,
.
.
.
.
,,
Interleukin 1 Interferon Complement Components Plasminogen Activator Thromboplastin Activity Lysosomal Hydrolases Fibronectin ,
,
,
25
The Biomaterials Silver Jubilee Compendium Biocompatibility: J.M. Anderson and K.M. Miller
C3b. C3b further activates macrophages as well as acting as an opsonin to enhance phagocytosis. Factor Bb produces more C3b via the alternative pathway 13. C3a is a potent anaphylatoxin. The enzymatic cleavage of plasminogen to plasmin is catalysed by plasminogen activator released by macrophages 51. Plasmin, the product formed in the reaction catalysed by plasminogen activator, has multiple regulatory capacities. It degrades fibrin into soluble degradation products, some of which exhibit chemotactic activity and cleaves the complement components C3 and C5 to active fragments C3a, C3b, C5a and C5b. Plasmin can also activate Hageman factor which will then activate the clotting cascade. Once the clotting cascade is activated, kinins, mediators of inflammation, can be generated 52. In addition to plasminogen activator activity, macrophage plasma membranes contain a thromboplastin activity or procoagulant activity 15. Macrophages can be stimulated to increase this thromboplastin activity by activated complement component C3b, immune complexes, endotoxins, lectins or ionophores. Since monocytes and macrophages can release both thromboplastin and plasminogen activator, the in vivo result depends on the balance of synthesis of these activators in terms of time and quantity of synthesis. Macrophages will also increase the synthesis and secretion of lysosomal hydrolases in response to activation by inflammatory stimuli. In a low pH environment these enzymes can degrade carbohydrates and other tissue components. Macrophages also secrete proteinases that are active at neutral pH. These neutral proteinases are important in the pathogenesis of chronic inflammation because they degrade connective tissue constituents, including the proteoglycan matrix of cartilage, collagen and elastin, and they generate inflammatory mediators such as activated complement components and kinins. Fibronectin, a glycoprotein that mediates cell-tissue matrix interaction, is also elaborated by macrophages. When produced by macrophages, especially at sites of tissue injury requiring repair, fibronectin becomes a chemoattractant for fibroblasts 53-56. It also enhances the phagocytosis of opsonized particles by macrophages without acting as an opsonin itself 57. Thus, fibronectin may augment healing after an inflammatory response.
CELLULAR AND HUMORAL MACROPHAGE INTERACTIONS .
11
("PMN~
.
,- %
.
....
M,F i!:i MAF i [1.1
GM-CSF
Fibrobllst ~ ~ . ~ f ~
.
.;.:.
~,0 C,x,,B~'
( M(~ ~CSs-R: ,L~ Prod ...... C
\ "~0, Pro ~tagllndinS~o~ / staglznams ~.,r Arlchidonic ACiCl ~Leukolrienes
F'b act I ron m
Endogenous
~
Flbronectln Frsgment=
'~
•,
]:i C:~. C3tb. C3d:Pha.ocy .....
" Polymeric Particles [L I: Intorteukin 1 F brob as1 Prol feration
Figure 1. Cellular and humoral regulatory pathways in inflammatory and immune responses mediated by the macrophage.
ACKNOWLEDGEMENT This effort was partially supported by National Institutes of Health Grants No. HL-25239 and HL-27277, and a NIH Research Career Development Award, HL-00779, to J.M.A.
REFERENCES 1 2 3 4
5 6 7 8 9 10
SUMMARY The marked ability of activated macrophages to secrete products capable of modulating cell-cell interactions and cell-mediator interactions suggests that both nonadhesive and adhesive events are important in the macrophage response to implanted materials. Figure 1 summarizes t h e various cellular and humoral interactions in which the macrophage may be involved when present at an implant site. It is of interest to speculate how these interrelated mechanisms may be controlled by altering the surfaces of biomaterials or by varying the form, texture or porosity of biomaterials. A better understanding of how these mechanisms are controlled will lead to a better understanding of 'biocompatibility' and also to improved biomaterials.
.
11 12 13 14 15 16 17
Coleman, D.L., King, R.N. and Andrade, J.D., The foreign body reaction: A chronic inflammatory response, J. Biomed. Mater. Res., 1974, 8, 1 99-211 Coleman, D.L., King, R.N. and Andrade, J.D., The foreign body reaction: An experimental protocol, J. B/omed. Mater. Res. Syrup., 1974, 5(1 ), 65-76 Rigdon, R.H., Tissue reaction to foreign materials, Crit. Rev. Food ScL Nutr., 1975, 7, 435-476 Gourlay, S.J., Rice, R.M., Hegyeli, A.F., Wade, C.W.R., Dillon, J.G., Jaffe, H. and Kulkarni, R.K., Biocompatibility testing of polymers: In vivo implantation studies, J. Biomed. Mater. Res., 1978, 12, 219-232 Turner, J.E., Lawrence, W.H. and Autian, J., Subacute toxicity testing of biomaterials using histopatho{ogic evaluation of rabbit muscle tissue, J. Biomed, Mater. Res., 1973, 7, 39-58 Autian, J., Toxicological evaluation of biomaterials: Primary acute toxicity screening program, Artif. Organs, 1977, 1, 5356 Marion, L, Haugen, E. and Mjor, I.A., Methodologicat assessments of subcutaneous implantation techniques, J. Biomed. Mater. Res., 1980, 14, 343-357 Taussig, M.J., Processes m Pathology, Blackwell Scientific Publications, London, 1979 Hilt, R.B. and LaVia, M.F., Principles of Pathobiology, 3rd edn, Oxford University Press, New York, 1980 Merchant, R., Hiltner, A., Hamlin, C.. Rabinovitch, A.. Slobodkin, R and Anderson, J.M., In vivo biocompatibility studies, i. The cage implant system and a biodegradable hydrogel, J. Biomed. Mater. Res., 1983, 17, 301-325 Anderson, J.M., unpublished observations Robbins, S.L. and Cotran, R.S., Pathological Basis of Disease, 2nd edn, W.B. Saunders, Philadelphia, 1979, pp. 55-90 Atkinson, J.P. and Frank, M.M., Clinical Immunology, (Ed. C.W. Parker) Vol. 1, W.B. Saunders, Philadelphia, 1980, pp. 219271 Becker, E.L and Ward, P.A., Clinical Immunology, (Ed. C.W. Parker) Vol. 1, W.B. Saunders, Philadelphia, 1980, pp. 272-297 Allison, A.C., Cfinical Aspects of Immunology, (Eds Lachman, P.J. and Peters, D.K.) Voi. 1. 4th edn, Blackwell Scientific Publications, Boston, 1982, pp. 101-129 Grinnell, F., Fibronectin and wound healing, Am. J. DermatopathoL, 1982, 4, 185-187 Norris, D.A., Clark, R.A.F., Swigart, L, Huff, J.C., Weston, W. and Howell, S.E., Fibronectin fragment(s) are chemotactic for human peripheral blood monocytes, J. Immunol., 1982, 129, 16121618
Biomateriats 1984, Voi 5 January
9
The Biomaterials Silver Jubilee Compendium
26 Biocompatibility: J.M. Anderson and K.M. Miller 18
19 20 21 22 23
24
25
26 27
28
29 30 31 32 33 34
35
36 37 38
10
Serhan, C.N., Smolen, J.E., Korchak, H. and Weissmann, G., Leukotriene B4 is a complete secretagogue in human neutrophils: Ca++ translocation in liposomes and kinetics of neutrophil activation, Adv. in Prostaglandin, Thromboxane, and Leukotriene Res., 1983, 11, 53-63 Weissmann, G., Personal Communication Johnston, R.B., Oxygen metabolism and the microbicidal activity of macrophages, Fed. Proc., 1978, 37, 2759-2764 Liebert, T.C., Chartoff, R.P., Cosgrove, S.I_ and McCuskey, R.S., Subcutaneous implants of polypropylene filaments, J. Biomed. Mater. Res., 1976, 10, 939-951 Farrar,J.J. and Hilfiker, M.L., Antigen-nonspecific helper factors in the antibody response, Fed. Proc., 1982, 41, 263-268 Shikita, M., Tsuneoka, K., Hagiwara, S. and Tsurufuji, S., A granulocyte-macrophage colony-stimulating factor (GM-CSF) produced by ca rrageenin-induced inflammatory cells of mice, J. Cell PhysioL, 1981, 109, 161-169 Coleman, D.L, Root, R.K. and Ryan, J.L., Enhancement of macrophage Fc-dependent phagocytosis by resident thymocytes: Effect of a unique heat-stable lymphokine, J. ImmunoL, 1983, 130, 2195-2199 Rouveix, B., Larno, S., Badenoch-Jones, P. and Lechat, P., Lymphokine-induced rnacrophage aggregation: Involvement of cyclic-GMP and rnicrotubules, Agents and Action, 1981, 11, 622-624 Boggs, D.R. and Winkelstein, A., White Cell Manual, 4th edn, F.A. Davis Company, Philadelphia, 1983, pp. 46-48 Goodman, M.G., Chenoweth, D.E. and Weigle, W.O., Induction of lnterleukin 1 secretion and enhancement of humoral immunity by binding of human CSa to macrophage surface C5a receptors, J. Exp. Med., 1982, 156, 912-917 Abramson, S.B., Given, W.P., Edelson, H.S. and Weissmann, G., Neutrophil aggregation induced by sera from patients with active systemic lupus erythematosus, Arthritis and Rheumatism, 1983, 26, 630-636 Craddock, P.R., Hammerschmidt, D., White, J.G., Dalmasso, A.P. and Jacob, H.S., Complement (C5a)-induced granulocyte aggregation in vitro, J. C/in. Invest., 1977, 60, 260-264 Sutton, J.S. and Weiss, L, Transformation of monocytes in tissue culture into macrophage, epithelioid cell, and muttinucleate giant cells, J. Cell Biol., 1966, 28, 303-307 Kaplan, G. and Gaudernack, G., In vitro differentiation of human monocytes, J. Exp. Med., 1982, 156, 1101-1114 Marchant, R.E., Miller, K.M. and Anderson, J.M., unpublished results Sundsmo, J.S. and Gotze, O., Human monocyte spreading induced by factor Bb of the alternative pathway of complement activation, J. Exp. Med., 1981, 154, 763-777 Michl, J., Unkeless, J.C., Pieczonka, M.M. and Silverstein, S.C., Modulation of Fc receptors of mononuclear phagocytes by immobilized antigen-antibody complexes, J. Exp. Med., 1983, 157, 1746-1757 Michl, J., Pieczonka, M.M., Unkeless, J.C., Bell, G.I. and Silverstein, S.C., Fc receptor modulation in mononuclear phagocytes maintained on immobilized immune complexes occurs by diffusion of the receptor molecule, J. Exp. Med., 1983, 157, 21212139 Cohen, L, Sharp, S. and Kurczycki, A., Human monocytes, B lymphocytes, and non-B lymphocytes each have structurally unique Fc v receptors, J. Immunol., 1983, 131,373-383 Stenson, W.F. and Parker, C.W., Prostaglandins, macrophages, and immunity, J. Immunol., 1980, 125, 1-5 Rutherford, B. and Schenkein, H.A., C3 cleavage products stimulate release of prostaglandins by human mononuclear phagocytes in vitro, J. ImmunoL, 1983, 130, 874-877
Biomaterials 1984, Vo/ 5 January
39
40 41 42 43
44
45 46 47 48
49 50 51
52 53 54 55
56
57
Rouzer,C.A., Scott, W.A., Hamill, A.L, Fei-Tong Liu, Katz, D.H. and Cohn, Z.A., Secretion of leukotriene C and other arachidonic acid metabolites by macrophages challenged with immunoglobulin E immune complexes, J. Exp. Med., 1982, 156, 1077-1086 Snyder, O.S. and Unanue, E.R., Corticosteroids inhibit murine macrophage la expression and interleukin 1 production, J. Immunol., 1982, 129, 1803-1805 Oppenheim, J.J., Stadler, B.M., Siraganian, R.P., Mage, M. and Matheson, B., Lymphokines: Their role in lymphocyte responses, Fed. Proc., 1982, 4 1 , 2 5 7 - 2 6 2 Oppenheim, J.J. and Gery, I., Interleukin 1 is more than an interleukin, Immunology Today, 1982, 3, 113-119 Baracos,V., Rodemann, H.P., Dinarello, C.A. and Goldberg, A.L., Stimulation of muscle protein degradation and prostaglandin E2 release by leukocytic pyrogen (Interleukin 1), IV. Engl. J. Med., 1983, 308, 553-558 Sipe, J.D., Vogel, S.N., Sztein, M.B., Skinner, M. and Cohen, A.S., The role of Interleukin 1 in acute phase serum amyloid A (SAA) and serum amyloid P (SAP) biosynthesis, Ann. N. Y. Acad. ScL, 1982, 137-150 Schmidt J.A., Mizel, S.B., Cohen, D. and Green, I., Interleukin 1, a potential regulator of fibroblast proliferation, J. Immunol., 1982, 128, 21 77-2182 Leibovich, S.J. and Ross, R., A macrophage-dependent factor that stimulates the proliferation of fibroblasts in vitro, Am. J. PathoL, 1976, 84, 501-513 Postlethwaite, A.E., Lachman, L.B., Mainardi, C.L and Kang, A.H., Interleukin 1 stimulation of collagenase production by cultured fibroblasts, J. Exp. Med., 1983, 157, 801-806 Hibbs, M.S., Postlethwaite, A.E., Mainardi, C.L and Kang, Kang, A.H., Alterations in collagen production in mixed mononuclear leukocyte-fibroblast cultures, J. Exp. Med., 1983, 157, 47-59 Jimenez, S.A, McArthur, W. and Rosenbloom, J., Inhibition of collagen synthesis by mononuclear cell supernatants, J. Exp. Med., 1979, 150, 1421-1431 Jalkanen, M., Connective tissue activating macromolecules in macrophage culture medium, Conn. Tiss. Res., 1981, 9. 1924 Chapman, H.A., Vavrin, Z. and Hibbs, J.B., Macrophage fibrinolytic activity: Identification of two pathways of plasmin formation by intact cells and of a plasminogen activator inhibitor, Cell, 1982, 28, 653-662 Gordon, S., Handbook of Experimental Immunology (Ed. D.M. Weir) Vol. 2, 3rd edn, Blackwell Scientific Publications, London, 1978, pp. 33.1-33.14 Werb, Z., Banda, M.J. and Jones, P.A., Degradation of connective tissue matrices by macrophages, J. Exp. Med., 1980, 152, 1340-1357 Tsukamoto, Y., Helsel, W.E. and Wahl, S.M., Macrophage production of fibronectin, a chemoattractant for fibroblasts, J. lmmunol., 1981, 127, 673 Rennard,S.I., Hunninghake, G.W.. Bitterman, P.B. and Crystal, R.G., Production of fibronectin by the human alveolar macrophage: Mechanism for the recruitment of fibroblasts to sites of tissue injury in interstitial lung disease, Proc. Natl. Acad. ScL, 1981, 78, 7147-7151 Remold, H.G., Shaw, J.E. and David, J.R., A macrophage surface component related to fibronectin is involved in the response to migration inhibitory factor, Cellular ImmunoL, 1981, 58, 1 75187 Pommier, C.G., Inada, S., Fries, LF., Takahashi, T., Frank, M.M. and Brown, E.J., Plasma fibronectin enhances phagocytosis of opsonized particles by human peripheral blood monocytes, J. Exp. Med., 1983, 157, 1844-1854
27
The Biomaterials Silver Jubilee Compendium
Systemic effects of biomaterials Jonathan Black
University of Pennsylvania, Philadelphia, Pennsylvania, USA
Evaluation of the host response to implanted biomaterials usually focuses on the implant site tissue response. This may lead to erroneous conclusions in the same way that examination of battles outside of their historic context does. A broader view discloses a variety of possible and actual systemic effects of carcinogenic, metabolic, immunological and bacteriological nature. Recognition of these effects in patients is hampered by a lack of epidemiological studies. Keywords: Host response, biomaterials, carcinogenesis, metabolism, immune response, infection
Studies of the interaction of foreign materials, of manmade or natural origins, with biological systems have commonly been described as studies of biocompatibility. I have previously suggested 1 that this term is inappropriate, in that it contains a biased value of judgement, and I have suggested the use of an alternate formula. That is, I have proposed that the aspects of materials that relate to their interactions with biological environments be termed the 'biological performance' of those materials. Within this term, I have further suggested that 'biological performance' is composed of two complementary parts: 'material response', encompassing all effects on materials including degradation and 'host response', encompassing the earlier ideas of biocompatibility, but in an even handed way. It is to the second part of biological performance of materials that this paper is addressed; that is, to the proper scope of consideration of the term 'host response'. The traditional host response experiment involves the preparation of specimens, often without much materials' characterization, their surgical implantation in defined anatomical sites in test animals and their recovery at various later times, with examination of the surrounding tissue.
THE PROBLEM OF THE OBSERVER Figure 1, reproduced from a study by McNamara and Williams 2 shows a scanning electron microscope view of a glutaraldehyde fixed section of the interface between a pure nickel implant and muscle after 1 5 weeks implantation in the rat. The authors describe this figure in the following words: '(We could see that) (n)umerous loosely packed cells were present, many of which appear damaged. Material was apparently being secreted or engulfed in some areas, and small objects similar to those protruding from copper crystallites (as seen with copper implantation) were seen. Many cells were partially buried among
cellular breakdown products and general debris, and in places, covered by small strips of deposited protein. Healthy macrophages were visible among the d e b r i s . . . and there were occasional platelet aggregates. '3 We are tempted to say, yes, we can understand and interpret this snapshot: there has been an acute coagulation/inflammation response but things appear to be in hand with a general cleanup underway and a more chronic response, perhaps leading to capsular maturation, beginning. The authors interpret this view, as part of a time series, as follows: ' . . . (A) classical chronic inflammatory response was seen with the nickel implants . . . . Later, many active macrophages became involved in clearing protein and cellular debris. With time, a more stable situation developed, macrophage activity having resulted in a great decrease in the surrounding fluid. '4 Let us now leave this example for the moment and view another scene. Figure 2 5 is a diagram of the battlefield of Waterloo, during the day of June 18, 1815. It is a summary, reflecting the fighting up until perhaps 9:00 pm, before the final assault by the Allies on Napoleon's lines. We have no photographs, but there are, as you might imagine, many eye witness accounts. In the late afternoon, a crisis occurred: Marshall Ney, at the head of a force of cavalry numbering an estimated 1 5 000, led a charge directly north into the centre of Wellington's line. Cotton 6 quotes a Captain Silborne: "When the tremendous cavalry f o r c e . . , moved forward to the attack, the whole space between La Haye-Sante and Hougoumont appeared one moving, glittering m a s s ; . . , and as it approached the Anglo-Allied position, undulating with the conformation of the ground, it resembled a sea in agitation". The crisis passed, the battle reversed and several days later a civilian, a Miss Charlotte Eaton, visiting the same site, remarked7: " O n top of the ridge in front of the British position, on the left of the road, we traced a long line of tremendous graves, or rather pits, into which hundreds of dead had been thrown as they had fallen in
91984 Butterworth 8" Co (Publishers) Ltd. 0142-9612/84/010011-O8503.00 Biomaterials 1984, Vol 5 January
11
The Biomaterials Silver Jubilee Compendium
28 Systemic effects: J. Black
Figure 1. Nickel-tissue interface (SEM) gluteraldehyde-fixed specimen, 15 weeks after implantation in the rat (Reproduced from Biomaterials, 1982, 3, 165-176)
their ranks . . . The ground was ploughed up in several places with the charge of the cavalry, and the whole field was literally covered with the soliders' caps, shoes, gloves, belts, and scabbards, broken feathers battered into the mud . . . . " What do these descriptions of the Battle of Waterloo reflect upon the study of host response to biomaterials? I would like to make two points: 1. Charlotte Eaton's account, several days after Ney's great charge, bears a remarkable resemblance to the histopathologist's view of the interface between biomateriai and tissue. Even with precise tools, like the scanning electron microscope, the view is that of days to weeks after the encounter. The players are dead or departed and we cannot hope to realise the sweep of combat, as in Silborne's graphic description of the moment. 2. Students of Waterloo concentrate on the battle itself, examining maps, diaries, artifacts, comparative technical studies of weapons and tactics, and so forth. Some may consider the military engagements of the day before or even the previous events of the Hundred Days, since Napoleon's landing in the south on March 1st. However, the tendency is to focus on the battle itself and to miss the larger social, political and economic setting that makes it possible to understand the conflict, its origins and its consequences. Similarly, studies of host
12
Biomaterials 1984, Vo! 5 January
response tend to focus almost solely on the local tissue responses to the implant. Since history has traditionally focused upon the role of individuals with little consideration until recently given to the "thermodynamics of the masses", the focal interest in specific battles is not unexpected. However, with the early understanding that the human body is made up of vast populations of nearly identical cells and constitutes a co-ordinated set of interacting physiological systems, the concentration on implant site phenomena is difficult to understand. One explanation for this bias, which will become clearer as this paper proceeds, is that remote or, more generally, systemic effects of implants are not out of the ordinary. That is to say, the phenomena to be discussed: infection, immune response, neoplastic transformation etc., all occur in any group of individuals. If we were to study one thousand individuals at random from the citizens of any country, we would find all of these conditions. Some of these individuals would have surgical implants and some of the conditions found might be referable to these implants. However, we would not expect to see any great qualitative disparity between the 'symptoms' of those with and those without implants. Thus, it is most likely that systemic effects of implants have gone largely unnoticed and unremarked.
The Biomaterials Silver Jubilee Compendium
29
Systemic effects: J. Black
Figure 2. General arrangement of the Battlefield of Waterloo, June 18, 1815 (Reproduced from "'Wellington at Waterloo" by Jac Weller. Jac Weller. Reprinted by permission of Harper and Row, Publishers, Inc. and Curtis Brown, Ltd.)
This does not diminish their importance. Much is made today about 'cost-benefit' calculations, in medicine as in business. The benefits of modern implantology are clear, graphic and dramatic. What is hidden is their true cost. Without a refined understanding of systemic phenomena, their severity and frequency of occurrence, we cannot complete the equation and make realistic assessments of the proper use of implants in humans.
EXAMPLES OF SYSTEMIC EFFECTS There is a general lack of recognition of the presence of systemic effects in patient populations. I would like to point out two examples of systemic sequellae of the introduction of foreign materials into the body, the first well known, the second only now coming into focus. These examples illustrate two points: systemic effects may be totally unforseen and are indeed seen 'away from' the implant. Poly(methyl methacrylate) cements are widely used in surgical applications for filling bony defects and stabilizing prosthetic devices. Early in the use of these cements it was observed that systemic hypotension frequently occurred within a minute or two of cement insertion into medullary spaces. Arterial pressure drops of more than 1 5 mmHg, occasionally observed in the company of transient cyanosis, and a number of cardiac arrests drew medical attention to this effect. It was eventually traced to a combination of a pharmacological effect of released monomer 8 and a possible secondary effect of intrusion of fat emboli into the venous return system. It was also noted in canine studies that the effects were accentuated by hypovolemia. Modern use of PMMA type cements, with normovolemic anaesthesia and, in some cases, venting of medullary cavities during cement insertion, has largely eliminated this unwanted acute systemic effect.
91967 by
A more complex problem has been the recognition of a pattern of encephalopathy, osteoporosis and anaemia in chronic dialysis patients awaiting kidney transplantation 9. The disease may progress to global dementia, inhalation pneumonia and death. Analytic post-mortem studies have shown that this condition is associated with 2 to 10 fold increases of aluminium content in grey and white matter in the brain. The source of this aluminium appears to be the dialysis solutions (dialysate). These symptoms are unknown in aialysis populations where diatysate aluminium content is less than 50/.~g/I 1~ but they appear and increase with aluminium concentration above this level. In general, concentrations of 200/xg/I are sufficient to cause an "'outbreak "'11. There are two additional fascinating aspects of this problem of inadvertent " ' i m p l a n t a t i o n " : 1. The osteoporosis that is part of this syndrome is diet and vitamin D resistant 1~ In patients who have normal renal function but receive aluminium from other sources, such as in antacids and anti-seizure drug formulations, periods of chelation treatment of up to one year are unable to markedly lower serum aluminium levels and restore vitamin D sensitivity i~. 2. The neuroactive species appears not to be ionic aluminium but a group of low molecular weight, size exclusion column separable species t3. These species, with molecular weights between 60 000 and 80 000, contain bound aluminium as well as proteins so they are properly termed organometallic ions. In the presence of increased dialysate and serum aluminium concentrations, the concentration of these species increases, as do the symptoms previously mentioned, while the aluminium content of albumin remains more or less constant. It may be argued that these patients are seriously ill, with a number of systemic disorders and that these findings may be incidental. However, there is a close correspondence between these findings and the correlation that
Biomaterials 1984, Vol 5 January
13
The Biomaterials Silver Jubilee Compendium
30 Systemic effects: J. Black
has been found TM between Alzheimer's Disease, a senile dementia, and increased aluminium content of brain tissue in its victims. Here the source of aluminium is presumably dietary and/or medicinal and many individuals involved are otherwise in normal health. In chronic dialysis, care is now being taken to provide dialysate fluids with low aluminium content. But these solutions have been shown is specifically still to increase serum levels of Cd, Co, Mn and Pb and chronic dialysis patients have been shown to have elevated serum levels (in comparison to normal controls) of three of these four metals (except Co) as well as of AI, As, Cu, Hg and Zn. We may well ask, what as yet unrecognized problems do these patients have associated with these elevated serum concentrations and their concomitant body burden increases?
CLINICAL KNOWLEDGE OF SYSTEMIC EFFECTS OF IMPLANTS IN PATIENTS Having, I think made the case for the existence of systemic effects in acute and chronic exposure conditions, let me now turn to a more general consideration of systemic effects related to the use of implants in man. It is important here to recognize that much of our understanding of these phenomena depends upon animal studies. The almost total lack of clinical epidemiological studies of patients with implants and the ubiquitous occurrence of symptoms proposed as indicating systemic effects have combined to leave us with little clinical knowledge. This former point is worth an additional comment. Studies of implants in patients tend to concentrate on three areas of evaluation, in order of perceived importance: function of the implant, change in pain status and, occasionally, change in quality of life. These studies are usually conducted by either the implanting surgeon or his or her colleagues. Thus, it is not surprising that orthopaedic surgeons, for instance, pay little attention to systemic infectious, oncologic or immunological conditions; these are outside of these investigators" normal field of view. Putting this problem another way, consider the following argument: It is probable that 1/4 of all individuals who received total hip replacement prostheses in the United States during 1 983 will eventually die of cancer. This prediction can be made, assuming no significant change in cancer incidence or survival during the next 35 years, since cancer currently accounts for approximately one quarter of all deaths in the United States. But, can one predict that these patients will have greater or less fractional rates of incidence, degrees of morbidity or cohort death rates than age, disease state, occupation and residence and life habits matched individuals without total hip replacements? The answer, of course, is no, since there have been no retrospective nor prospective studies addressing such issues. The absence of studies, as previously suggested, can be explained by the observation that orthopaedic surgeons do not treat patients for soft tissue tumours and oncologists and general surgeons are generally unaware of the implant status of patients. In light of the growing body of results from animal studies which do suggest long term effects of metallic implants, such epidemiological studies are imperative. Their need is rendered even more urgent by current trends of increasing volume of implant use, earlier implantation
14
Biomaterials 1984, Vol 5 January
and the use in orthopaedics of high specific surface area implants for fixation by biological ingrowth.
INTERACTION OF IMPLANTS WITH THE BODY Implants interact or 'couple' with the body in three principle ways: chemically, electrically and mechanically. Each of these modes has been shown separately to affect local tissue response and it is probable that synergistic effects exist. However, these phenomena are not of concern to us here. Systemic effects are primarily mediated by chemical and mass transport phenomena. Metallic implants release corrosion products which appear to be largely or wholly organometallic TM. Some of these are locally tissue bound but others are free to circulate, as seen by their diffusion gradients in tissue adjacent to the implant 17 and their presence in systemic circulation 18, remote tissue sites 19 and excretory products 2~ In man, systemic (serum) elevation 21'22, remote site accumulation 23 and excretion have been demonstrated 2~'22. Polymeric implants can release components by structural degradation (depolymerization or hydrolysis) 24, elution 25 or enzymatic attack 2s while ceramic implants release materials primarily by dissolution 27 with particulate release also possible. In absence of experimental findings, I will assume that materials released from ceramic implants behave in the same qualitative way as material released from metallic implants. Mass transport of small particles may prove to be an important route of distribution. These particles may move passively, through tissue and/or the circulatory system or can be actively transported by macrophages 28. These particles may accompany implantation, such as glove talc 29, or be released by wear 3~ corrosion 31 or dissolution processes or be produced by fatigue processes, such as graphite particle of fibre release from graphite-containing ligament prostheses 32. Charnley's observations 33 of 'teflonomas', abdominal granulomas associated with remote site accumulation of polymeric wear debris are a classic example of biological response to transported particles. Pierce and Boretos 34 illustrate our ignorance of the consequences of these phenomena, which they term collectively 'unintentional clinical administration of particles', by asking: "What size constitutes a particle? What are the safe limits of size and quantity? What are the biological risks?" I must pause at this point to reflect upon two peculiarities that set the systemic effects of implants apart from more normal concerns about pharmacological function. In the first place, there is the problem of the route and rate of administration. Most of what we know about drug action is based upon observation of the effects of a single dose or a series of doses given over a short period of time. These doses are sized to produce a specific effect, based upon titration with animal models. Studies focus on transport, distribution, and excretion. Thus it is not startling that efforts to study distribution of implant degradation products should have proceeded the same way 35. However, implants release materials on a chronic basis at very low rates. This is clearly a different situation and, in the presence of biological adaptability, may produce radically different results from acute administration. The route of administration is also important, as can be seen
The Biomaterials Silver Jubilee Compendium
31
Systemic effects: J. Black
from the experience with aluminium in dialysis patients, which we have discussed earlier. The extensive use of aluminium, in colloidal form, in oral medications, early on led investigators to reject the possibility that dialysate aluminium, at levels below 200/#1 could be of biological importance. However, the ability of sialysis to introduce ions directly into the circulatory system, bypassing the mucosal 'barrier' with its selective control of uptake and excretion, produced significant effects, as we have seen. It is worth reflecting on this latter point for a moment. Many, if not all, of the materials used in implants, are foreign to the experience of biological systems. Metals, in their fully reduced forms, are extremely rare in the natural environment. Most long chain polymers used in implants can only be produced by artificial synthetic pathways. Both reduced metals and long chain polymers as well as fine powders are normally effectively excluded from the bodies of animals and people by a complex series of mucosal membrane based defence systems. The act of implantation breaches these systems and introduces these novel material forms into biological systems that can be said to be naive. The second peculiarity of implants that must be pointed out follows from aspects of the first. The chronic, low level release of materials by implants should alert us to the possibility of long term or delayed effects. That is, the circulating level of degradation product may not, in itself, be sufficient to product a biological response and result in detectable symptoms. However, adaptive changes, such as acquired hypersensitivity or remote concentrations of degradation products 19 may change this situation.
TYPES OF SYSTEMIC EFFECTS Smith and Black 36 have previously discussed these effects, specifically with regard to distribution of corrosion products, and suggested a classification scheme that divides them into four types: carcinogenic, metabolic, immunologic and bacteriologic. Although it could be argued that the latter two types should properly be combined and that carcinogenic effects probably represent failures of the immune system, ! still find this is a useful scheme and will use it to organize the following discussion. It is not possible to be inclusive in such a discussion; references are included largely for illustrative purposes and the reader is referred to the bibliography appended at the end of this paper for further information.
CARCINOGENIC EFFECTS Concern about carcinogenesis secondary to implantation is a measure of a general societal concern about cancer. It is ironic that the success of implants is a part of the reason for this concern: it is probably the case that increasing life expectancy associated with the use of implants in correcting congenital and acquired conditions has contributed to the impression that cancer incidence rates are rising. Examination of age adjusted rates of tumour incidence for the forty years prior to 1 980 37 reveals that with the exception of lung and possibly pancreatic cancer, rates are not rising. Nevertheless, studies of implants in animals have revealed considerable reason to be suspicious of the
possibility of chemical carcinogenesis. Metal ions, such as Ni2+, 38, C02+, 39 and CrS+,4~ as well as monomers, such as vinyl chloride 41, all of which can be released by implants or blood contact devices, have been shown to be either primary carcinogens, co-carginogens or carcinogenesis promoters. The role of metals ions as carcinogenic agents is especially worrisome. Furst42'43 provided a stipulation that he felt metal ions should meet before being considered potential carcinogens in man: "Tumors must appear at a site distant from the point of application; more than one route (of application) must be effective; more than one species must respond; the growth should be transplantable; and, if malignant, invasion and/or metastasis must be noted "'43. Unfortunately, the metal ions listed previously meet all of these criteria. Chemical carcinogenesis secondary to implant use, if it exists, should probably be considered primarily a systemic effect. Implants are most frequently placed in tissues that are not particularly sensitive to primary tumour genesis (bone, heart, etc.) Implant site tumours in man are largely unknown or unrecognized. Orthopaedic devices, which may represent the largest such class of chronic metallic implants, are rarely associated with new, primary tumours; the literature contains only three reports 44. Animal studies suggest that the failure to study and observe correlations between implants and remote site tumours may be a real problem. Remote site tumours have been reported associated with implantation of clinical metal alloys in rats 45 and a subsequent study of these same animals showed an apparent relationship between nickel release and tumour incidence 46. These studies raise considerable concern about the long term use of nickel bearing alloys, especially those with high nickel content, and are being repeated 47 with larger numbers of animals. The other concern about relationships between implants and carcinogenesis has been with the possibility of foreign body tumourgenesis. An exception to this rule is the use of breast augmentation implants. A large retrospective and prospective study to determine if such implants are associated with higher rates of incidence, and in the case of reconstruction after partial breast excision, or recurrence of tumourgenesis, is now underway in the United States 49. However, since particles may be transported by macrophages to the lungs, we should be sensitive to future potential tumourgenesis problems of a systemic nature. A failure to excrete particles across the lung parenchyma into the airway, as is apparently usual 28 in mammals, may lead to particle accumulation in relatively sensitive tissue. Since many of these particles will exceed the apparent lower size limit (0.22 microns) for foreign body carcinogenesis 5~ their accumulation should be a matter of concern. A recent study 51'52 of high specific surface area titanium alloy implants in the baboon, which showed significant time-dependent aluminium and titanium accumulations in the lungs, but not in other tissues, is suggestive of such a particle transport effect.
METABOLIC EFFECTS Degradation products of implants are to one degree or
Biomaterials 1984, Vol 5 January
15
32
The Biomaterials Silver Jubilee Compendium Systemic effects: J. Black
another soluble. Thus, it is not unexpected that they should enter into metabolic processes. Let us briefly examine the two principal classes of simple soluble products: metallic ions and monomers. In a sense, these are artificial classes: metallic release from implants probably results in production of only organometallic ions 16 while polymer degradation products may be larger than monomers. Metals play a very large role in metabolic processes. It has been stated: "There probably does not exist a single enzyme-catalysed reaction in which either substrate, product, enzyme, or some combination within the triad is not influenced in a very direct and highly specific manner by the precise nature of inorganic ions which surround or modify it ''s3. It used to be the case that we could divide metals into three classes, based upon their normal activities: the physiological metals (Na, K and Ca), the essential metals (Co, Cu, etc.) and the toxic metals (Pb, Cd, etc.) with the recognition that there was a fourth, or leftover class: the neutral metals, which have no apparent biological activity, such as titanium. This view is giving way to a more general approach, motivated by the sense of the quotation in the previous paragraph, in which each metal is considered in a pharmacological manner, with a Bertrand model doseresponse curve characterized by a low concentration regime of low or no effect, an intermediate concentration regime encompassing dose related positive or beneficial effects and a high dose regime of inhibitory or toxic effect. The toxic metals were suspected to lack a region of positive effect; however, many, such as arsenic 54 are now being shown to be essential in trace concentrations. Finally, the extension of the study of trace elements now made possible by the advent of modern analytical techniques such as neutron activation analysis and plasma excitation atomic absorption spectroscopy will probably result in defining biological activities for most of the socalled neutral metals. Of the principal metals present in implant alloys, all but possibly titanium and some of the refractory metals (Nb, Ta) are now known to have biological effects. Thus consideration of the biological effects of metals released from implants depends upon a knowledge of two areas: the quantity and form of the released agent and the normal pathways for such an agent to act in. In the first area, we are beginning to recognize the depth of our ignorance. Even in the light of widely recognized complexing between metal ions and proteins, we have until recently assumed that such complexes, when formed by the addition of inorganic salts to protein solutions, such as serum, could produce the same complexes that occur when implants corrode in vivo. This appears not to be the general case. For instance, for chromium 55 the valence of the released metallic moiety is apparently influenced by the presence of serum proteins. The particular importance of this observation is highlighted by the strange, dual aspect of the biological effects of chromium 56. As a trivalent ion, it appears to be of great biological value and merit, is largely excluded from cells and is an essential element, forming part of the glucose tolerance factor. As a hexavalent ion, it penetrates cells easily and is known to be a potent mutagenic and carcinogenic agent at remarkably low concentrations 57. Even if we know unequivocally the form of the released species, we still need to know the kinetics of distribution, storage and excretion. As previously noted,
16
Biomaterials 1984, Vol 5 January
kinetic studies are defective, as a group, since they are based upon high dose, acute experiments rather than low dose, chronic studies. Furthermore, for many of the metals involved, the basic metabolic pathways are not currently well defined, even without the complication of the endogeneous source represented by the implant. I have found it most curious and interesting that the vast majority of studies of metal metabolism, while paying considerable attention to diseases involving defects in metal metabolism, such as Wilson's Disease, totally overlook the role of implants. Monomer introduction into metabolic pathways might be considered to be a relatively minor matter, so long as bizarre metal or halide bearing molecules were avoided in materials' fabrication, since polymers contain chemical bonds that are 'familiar' to mammalian biological processes. However, my attention was recently drawn to this issue 5e by the suggestion that polymer degradation products that resemble but are not identical to metabolic intermediates might enter into substrate pools and function in an adverse, competitive manner. An experimental model for this exists in the use of cis-hydroxyproline as an agent to control scar formation after tendon and nerve repair procedures 59. This amino acid analogue, which does not occur naturally, apparently cannot be distinguished from unhydroxylated proline during ribosomal procollagen synthesis. However, since proline is hydroxylated after molecular assembly, but in the trans form, the presence of cis hydroxyproline within procollagen would be expected to have significant effects. This is in fact the case, with both an inhibition of collagen synthesis and poor mechanical properties of tissue containing collagen with incorporated cis-hydroxyproline. These results are desired locally to reduce the effects of scarring but possible remote effects, secondary to systemic distribution and local incorporation of cis-hydroxyproline, stimulated the investigators to develop local drug release systems that can restrict the abnormal form to the vicinity of the surgical procedure 6~ Metabolic pathways for monomer or low molecular weight degradation products of implants are not generally known unless the structure is one previously investigated for its biological activity, either as a potential drug, or in fate assessment studies of toxic materials, such as insecticides. A further complication, probably not present in the consideration of the metabolic participation of organometallic materials, is that polymer degradation products may be hydrophobic pools in the body.
IMMUNOLOGICAL EFFECTS Here we are on much safer and less speculative ground. It is well established that implants can invoke both B-cell and T-cell mediated immune responses 6~. These can arise in clinical populations, in response to the use of both metallic 62 and polymeric 83 biomaterials. While there is some debate concerning the nature of implant site effects, these and other clinical reports document a range of remote site immune system responses, including urticaria and asthma, which are precipitated by implants and that resolve on removal of the challenge. Furthermore, screening of patients from whom metallic implants have been removed suggests 64, when compared with epidemiological studies of non-implant patients 65, that implants themselves can cause immune system sensitization; that is,
The Biomaterials Silver Jubilee Compendium
33
Systemic effects: J. Black
these symptoms are not solely the result of challenging already established hypersensitivities. All immune effects must of necessity be classified as systemic effects, whether the antigen-antibody complexing occurs at the implant site or elsewhere. However, even in the narrower sense of the this discussion, in which we are considering effects remote to the implant site, immunological effects must still be taken seriously. Despite clear clinical evidence of such effects, there has been scant attention given to this field. Thus, we know very little definitive about occurrence rates, sensitivity thresholds or severity of consequences, both acute and chronic. In light of the increasing use of implants and their predicted increased life expectancy, based both upon technical and surgical advances and upon declining age of patients at implantation, we need to know far more.
BACTERIOLOGICAL EFFECTS In the same way that trace elements play significant roles in mammalian metabolism, they are known to be important in bacterial metabolism. It is well recognized that one response to systemic infection in mammals is a rapid reduction in serum iron concentration 66. This is brought about by movement of transferrin-bound iron to storage depots, primarily rough endoplasmic reticulum in hepatocytes and other cells and its deposition as haemosiderin and related compounds. The net effect is to improve binding strength of the remaining serum iron to transferrin and thus deny bacteria access to it. Ths process, termed "nutritional immunity", is so strong that patients with chronic infections may be unable to mobilize sufficient iron to maintain normal levels of haemoglobin synthesis, even in the presence of normal body iron stores 67. Weinberg 68 goes so far as to suggest that, " . . . (I)ron is the metal whose concentration in host fluids appears to be most important (in the establishment/suppression of bacterial disease in animals)." Implantation of stainless steel in rabbits produces elevated serum iron concentrations 66. In this experiment, there was an apparent accompanying suppression of haemoglobin synthetic rate, secondary to chromium release. This might be expected to result in an elevated infectability, especially in the vicinity of implants, where concentrations should be higher 17, since chromium binds with transferrin in competition with iron. A recent experiment reported from our laboratory 2~ involving the implantation of cobalt-chromium alloy in rats, displayed a dose related incidence of remote site (lung) disease that could apparently be related temporily to chromium release. Clinical observation of infections in the vicinity of implants in patients, reveals no evidence to suggest that iron bearing implants suppress resistance to infection more than non-iron bearing implants, thus further supporting the possibility that chromium, common to both stainless steel and cobalt based alloys, is related to increased infectability of metallic implant sites. However, a more interesting possibility is that corrosion product release can produce a general suppression of the immune system. Rae69 has shown that corrosion products in vitro are cytotoxic. In lower concentrations 7~ cor rosion products also suppress chemotaxis, a vital step in the suppression of bacterial infection by host cells. Whether these effects are reflected in a greater systemic infectability of implant patients is unknown; this is, yet
again, another area devoid of clinical studies. Furthermore, the possible participation of polymer degradation products in immune system suppression is also unknown.
CONCLUSIONS I would summarize the situation with regard to systemic effects of implants as follows: 1. There are animal and in vitro experiments that suggest the possibility of a broad range of effects. 2. Clinical implant materials degrade to produce systemic and remote site concentrations sufficient to predict that these effects could occur in patients. 3. Isolated clinical observations support the presence of systemic effects, especially associated with immune responses and metal overload and accumulation conditions. 4. Large scale epidemiological studies are needed to reveal the full extent and importance of systemic effects in patient populations.
BIBLIOGRAPHIC NOTE This review is necessarily very brief. The reader is directed to the following for additional information: Comar, C.L., Bronner, F. (Eds.) Mineral Metabolism, Vol 1 and 2, Academic Press, New York, New York, 1960, 1964 Davies, I.J.T. The Clinical Significance of the Essential Biological Metals, C.C. Thomas, Springfield, Illinois, 1972 Luckey, T.D., Venugopal, B. Metal Toxicity in Mammals, Vol 1 and 2, Plenum Press, New York, New York, 1977 Underwood, E.J. Trace Elements in Human and Animal Nutrition, 4th edn., Academic Press, New York, 1977 Weller, J. Wellington at Waterloo, T.Y. Crowell Co., New York, New York, 1967 Williams, D.F. (Ed.) Systemic Aspects of Biocompatibility, Vol. I and II, CRC Press, Inc., Boca Raton, Florida, 1981
REFERENCES 1 2
3 4 5
6 7 8
9
10
11 12 13
Black, J. Biological Performance of Materials, Marcel Dekker, New York, New York, 1981, pp. 3 - 6 McNamara, A. and Williams, D.F., Scanning electron microscopy of the metal-tissue interface. I1. Observations with lead, copper, nickel, aluminium and cobalt, Biomaterials, 1982, 3, 165-1 76, figure 5d Op. cit., 2, p. 169 Op. cir., 2, p. 175 Wetler, J. Wellington at Waterloo, T.Y. Crowell Co., New York, New York, 1967, Map 5, p. 232 Cotton, E. A Voice from Waterloo, 5th edn., Printed privately, London, UK, 1854, p. 82 Eaton, C., Narrative of a Residence in Belgium During the Campaign of 1815, Printed privately, London, UK, 1817, cited in: Naylor, J. Waterloo, B.T. Batsford, London, UK, 1960, p. 190 Homsy, C.A., Tullos, H,S., Anderson, M.S., Differante, N.M., King, J.W., Some physiological aspects of prosthesis stabilization with acrylic polymer. Clin. Orthop. Rel. Res., 1972, 83,317-328 Parkinson, I.S., Ward, M.K., Kerr, D.N.S., Dialysis encephalopathy, bone disease and anaemia: the aluminium intoxication syndrome during regular haemodialysis. J. Clin. Path., 1981,34, 1285-1294 Parkinson, I.S., Ward, M.K., Feest, T.G., Fawcet, R.W.P., Kerr, D.N.S., Fracturing dialysis osteodistrophy and dialysis osteodistrophy and dialysis encephalopathy, An epidemiological survey. Lancet 1979, i, 4 0 6 - 4 0 9 Rozas,V.V., Port, F.K., Easterling, R.E., An outbreak of dialysis dementia due to aluminum in the dialysate, J. DiaL, 1978, 2, 459-470 Kaplan, F., Personal communication, 1 983 King, S.W., Savory, J., Wills, M.R., Aluminum distribution in
Biomaterials 1984, Vol 5 January
17
34
The Biomaterials Silver Jubilee Compendium Systemic effects: J. Black
14 15 16 17
18 19 20 21 22 23 24
25 26
27
28 29 30 31
32 33 34 35 36
37 38
18
serum following haemodialysis., Ann. Clin. Lab. ScL, 1982, 12, 143-149 Perl, D.P., Brody, A.R., Alzheimefs disease, X-ray spectrometric evidence of aluminum accumulation in neuro-fibrillary tanglebearing neurons, Science, 1980, 208, 297-299 SaIvadeo, A., Minoia, C., Segagni, S., Villa, G., Trace metal changes in dialysis fluid and blood of patients on haemodialysis, Int. J. Art. Org., 1981,2, 17-21 Woodman, J.L., Organometallic corrosion products: An in vivo and in vitro comparison. Ph.D. Thesis, University of Pennsylvania, 1980, xii + 428 Lux, F., Zeisler, R., Investigations of the corrosive deposition of components of metal implants and of the behaviour of biological trace elements in metallosis tissue by means of instrumental multi-element activation analysis, J. Radioanal. Chem., 1974, 19, 289-297 Woodman, J.L, Black, J., Nunamaker, D.N., Release of cobalt and nickel from a new total finger joint prosthesis made of vitallium, J. Biomed. Mater. Res., 1983, 17, 655-668 Ferguson, A.B. Jr., Akahoshi, Y., Laing, P.G.. Hodge, E.S., Characteristics of trace ions released from embedded metals implants in the rabbit, J. Bone Jt. Surg., 1960, 44A, 323-336 Wapner, K.L, Black, J., Morris, D., Chromium release by cast CoCr alloy: Ionic valence and its implications for morbidity, Trans. Orthop. Res. Soc., 1983, 8, 240 Coleman, R.F., Herrington, J., Scales, J.T., Concentration of wear products in hair, blood and urine after total hip replacement, Brit. Med. J., 1973, 1, 527-529 Black,J., Maitin, E.C., Gelman, H., Morris, D.M., Serum concentrations of chromium, cobalt and nickel after total hip replacement: A six month study, Biomaterials, 1983, 4, 160-164 Dobbs, H.S., Minski, M.J., Metal ion release after total hip replacement, Biomaterials, 1980, 1, 193-198 Gilding, D.K., Degradation of polymers: Mechanisms and implications for biomedical applications, in Fundamental Aspects of Biocompatibility, Vol I, (Ed. D.F. Williams) CRC Press, Inc., Boca Raton, Florida, 1981,43-65 Salthouse, T.N., Williams, J.A., Willigan, D.A., Relationship of cellular enzyme activity to catgut and collagen suture absorption, Surg., Gyn., Obstet, 1969, 129, 691-698 Homsy, C.A., Ansevin, K.D., O'Bannon, W., Thompson, S.A., Hodge, Ro, Estrella, M.E., Rapid in vitro screening of polymers for biocompatibility, J. Macromol. ScL Chem., 1970, A4, (3), 615-634 Hench, L.L., Stability of ceramics in the physiological environment, in Fundamental Aspects of Biocompatibility, Vol. II, (Ed. D.F. Williams), CRC Press, Inc., Boca Raton, Florida, 1981, 67-85 Styles, J.T., Wilson, J., Comparison between in vitro toxicity of two novel fibrous dusts and their tissue reactions in vivo., Ann. Occup. Hyg., 1976, 19, 63-68 Henderson, W.J., Maskell, A.V., Griffiths, K., Barr, W.T., Contamination of surgical gloves, Brit. Med. J., 1978, 1 (6109), 363 Walker, P.S., Bullough, P.G., The effects of friction and wear in artificial joints, Ortho. Clin. N.A., 1973, 4 (2), 275-294 Kruger, J., Fundamental aspects of corrosion of metallic implants, in Corrosion and Degradation of Implant Materials, STP 694, (Eds. B.C. Syrett and A.Acharya), American Society for Testing and Materials, Philadelphia, Pennsylvania, 1979, 107-127 Jenkins, D.H.R., The repair of cruciate ligaments with flexible carbon fibre, J. Bone Jr. Surg., 1978, 60B, 520-522 Charnley, J., Arthroplasty of the hip. A new operation, Lancet, 1961, i, 1129 Pierce,W.S., Boretos, J.W., The dilemma of patient exposure to ubiquitious foreign particles, J. Biomed. Mater. Res., 1983, 17, 389-391 Onkelinx, C., Whole-body kinetics of metal salts in rats, in Clinical Chemistry and Chemical Toxicology, (Ed. S.S. Brown), Elsevier/North Holland, Amsterdam, Netherlands, 1977, 37-40 Smith, G.K., Black, J., Models for systemic effects of metallic implants, in Retrieval and Analysis of Orthopaedic Implants, (Eds. A. Weinstein, E. Horowitz and A.W. Ruff), NBS Special Publication 472, US Government Printing Office, Washington, DC, 1977, 23-30 Anon., 1982 Cancer Facts and Figures, American Cancer Society, New York, New York; 1983 Committee on Medical and Biological Effects of Environmental Pollutants, National Academy of Sciences, Nickel, National Academy of Sciences, Washington, 1975
Biomaterials 1984, Vol 5 January
39 40
41 42 43 44 45 46 47 48 49 50 51 52
53 54 55 56 57 58 59
60 61 62 63 64
65 66 67 68 69 70
Gilman, J.P., Metal carcinogenesis I1: A study of the carcinogenic activity of cobalt, copper, iron and nickel compounds, Cancer Res., 1962, 22, 158--162 International Agency for Research on Cancer (IARC), Chromium and chromium compounds in, IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemical to Humans, Vol. 23, IARC, Switzerland, 1980, 205-324 Maltoni, C., Lefemine, G., Carcinogenicity assays of vinyl chloride, Ann. IV. Y. Acad. ScL, 1975, 246, 195-218 Furst,A., An overview of metal carcinogenesis, Adv. Exp. Med. Biol., 1978, 91, 1-12 Furst,A., Haro, R.T., A survery of metal carcinogenesis, Prog. Exp. Tumor. Res., 1969, 12, 102-133 Black, Op. cir., 1, p. 146 Memoli, V.A., Woodman, J.R., Urban, R.M., Galante, J.O., Malignant neoplasms associated with orthopaedic implant alloys, Trans. Orthop. Res. Soc., 1982, 7, p. 164 Woodman, J.R., Memoli, V.R., Urban, R.M., Galante, J.O., Nickel and titanium release from orthopaedic materials in rats with malignant neoplasms, Trans. Soc. Biomater., 1983, 6, 43 Galante, J.O., Personal communication, 1983 Brand,K.G., Foreign body induced sarcoma, in Cancer." A Comprehensive Treatise, (Ed. F.F. Becker), Plenum Press, New York, New York, 1975, 485-511 Brody, G., Personal communication, 1982 Brand, Op. cit., 48, p. 494 Woodman, J.R., Jacobs, J.J., Galante, J.O., Urban, R.M., Titanium release from fiber metal composites in baboons--A long term study, Trans. Orthop. Res. Soc., 1982, 7, 166 Woodman, J.R., Jacobs, J.J.. Urban, R.M., Galante, J.O., Vanadium and aluminium releae from fiber metal composites in baboons - - A long term study, Trans. Orthop. Res. Soc., 1983, 8, p. 238 Anon, Timely topics in clinical chemistry: Toxicology of trace metals II, Am. J. Med. Tech., 1969, 35, 652-655 Mertz, W, The essential trace elements, Science, 1981, 213, 1332-1338 Wortman, R.S., Merritt, K., Brown, S.A., Millard, M.M., XPS analysis of metal salts, corrosion products and protein interactions, Trans. Soc. Biomater., 1983, 6, 105 Mertz, W., Chromium: An ultra-trace element, Chemica Scripta, 1983, 21, 145-150 Roe, F.J.C., Carter, R.L., Chromium carcinogenesis: Calcium chromate as a potent carcinogen for the subcutaneous tissue of the rat, Brit. J. Cancer, 1969, 23, 172-176 Williams, D.F., Lecture, Systemic Effects of Implants, Society for Biomaterials, Birmingham, Alabama, April 28, 1983 Lane, J.M., Bora, F.W.Jr., Prockop, D.J., Heppenstall, R.B., Black, J., Reduction of scar formation by cis-hydroxyproline An inhibitor of collagen synthesis, J. Surg. Res., t 972, 3, 135137 Bora,F.W. Jr., Unger, A.S., Osterman, A.L., The local inhibition of nerve scar by the biodegradable vehicle, Chronomer, carrying cis-hydroxyproline, Trans. Orthop. Res. Soc., 1983, 8, p. 278 Merritt, K, Brown, S.N., Hypersensitivity to metallic biomaterials, in, Systemic Aspects of Biocompatibility, Vol. II, (Ed. D.F. Williams), CRC Press, Inc., Boca Raton, Florida, 1981, 33-48 Cramers,M., Lucht, L., Metal sensitivity in patients treated for tibial fractures with plates of stainless steel, Acta. Orthop. Scand., 1977, 48, 245-249 Stungis, T.E., Fink, J.N., Hypersensitivity to acrylic resin, J. Prosthet. Dent., 1969, 22, 245-248 Merritt, K., Mayor, M.B., Brown, S.A., Evaluation of sensitivity to metallic implants in, Evaluation of Biomaterials, (Advances in Biomaterials, Vol 1), (Eds. G.D. Winter, J.L Leray and K. deGroot), John Wiley Et Sons, Chichester, UK, 1980, 315-324 Fregert,S., Rorsman, H., Allergy to chromium, nickel and cobalt, Acta Dermatology-Venerology, 1966, 46, 144-148 Smith, G.K., Systemic transport and distribution of iron and chromium from stainless steel implants. Pd.D Thesis, University of Pennsylvania, Philadelphia, Pennsylvania, 1982, x + 334 Moore, C.V., Iron, in Modern Nutrition in Health and Disease, 4th ed., (Eds. R.S. Goodhart and M.G. Wohl), Lea 8 Febiger, Philadelphia, Pennsylvania, 1968, 339-364 Weinberg, E.D., Roles of iron in host-parasite interactions, J. Int. Diseases, 1971, 124, 401-410 Rae, T., A study of the effects of particulate metals of orthopaedic interest on murine macrophages in vitro, J. Bone Jr. Surg., 1975, 57B, 444-450 Rae, T., The toxicity of metals used in orthopaedic prostheses, J. Bone Jt. Surg., 1981, 63B, 435-440
35
The Biomaterials Silver Jubilee Compendium
The m tro response of osteoblasts to bioactive glass T. Matsuda* and J.E. Davies t
Department of Anatomy, University of Birmingham Medical School, Birmingham B 15 2TJ, UK (Received 5 February 1987; accepted 18 February 1987)
Osteoblasts from neonate rat calvaria migrated in culture from the endocranial surface of parietal bones onto fragments of bone-bonding 45S5 glass or non-bone-bonding quartz glass. These organ culture units were maintained for up to 4 wk. No significant production of extracellular matrix (ECM) was seen on the quartz glass samples. However, osteoblasts colonized the 45S5 samples in multUayers and produced abundant ECM as seen by light (LM), scanning electron (SEM) and transmission electron (TEM) microscopy. The interface developed on 45S5 glass was designated as either Type I (non-collagenbondir~g) or Type II (showing direct interdigitation of collagen with the calcium phosphate-rich glass surface). It was concluded that, since this in vitro method is capable of reproducing some aspects of the known in vivo behaviour of 45S5, such techniques may be developed as a means of batch-testing bioactive biomaterials and investigating bone cell/biomaterial interactions. Keywords: Implants, bioactive glass, biocompatibility, glass-ceramic, bone, tissue response
It has been shown by Hench and co-workers 1-3 that the creation of a biological bone bond with glass and glassceramic materials is due to tissue reactions provoked by the implanted material. By comparing such tissue reactions to bone-bonding and non-bone-bonding materials, Gross and Strunz 4' s reported that the implant is capable of releasing substances which may both influence steps in the mineralization process and affect differentiation of cells in the periimplant compartment. These early effects on the processes of bone healing are considered to be more important, in the assessment of an implant material, than long-term boneremodelling effects 5. Indeed, as a single cell type is responsible for the elaboration of bone tissue, an understanding of osteoblast reactions to artificial substrata is of central importance in explaining the bioreactive pathways engendered by a bone-substitute biomaterial 6. However it is known, from both in vivo and in vitro studies that osteoblasts, which may migrate7 to the wound site, can express a modulated phenotype e and elaborate a variable combination of fibrous, chondroid or bony tissue 5. This potential for variable phenotypic expression, which may be critically influenced by the presence of an artificial material, can seriously complicate the interpretation of experimental data and, in particular, in vitro findings. Thus, Gross et al. 9 have recently reported that osteoblast multiplication is inhibited on a surface reactive (bone-bonding) glass-ceramic, with concomitant increase in the appearance of isolated clumps of cells containing mineralized foci, with *visiting Research Fellow. Permanentaddress: Departmentof Orthodontics, Universityof KagoshimaDentalSchool,Usuki-cho,Kagoshima,890 Japan. tTo whom correspondence should be addressed.
respect to a non-bone-bonding material. They concluded that this in vitro behaviour was contradictory to known in vivo responses for the same glass-ceramic materials. In vitro assays, chosen to investigate the reaction of osteoblasts to biomaterials, must therefore critically depend upon an ability to mimic at least some of the known in vivo responses to the same material. A basis for such a method was proposed for biocompatibility testing by Jones and Boyde in 1979 lo who described the migration of osteoblasts, in vitro, from rat calvaria onto various artificial or natural overlay materials. However, this technique was not adapted to the examination of calcium phosphate bioceramics until 1 98411. We have recently described a variety of these methods in detail s and employed them to observe the response of osteoblasts to various inorganic 8' 11,12 and polymeric 13' 14 materials. In this paper we briefly outline one such method to observe the in vitro response of primary rat osteoblasts to a bioactive glass substrata and illustrate key similarities to the known in vivo reactivity of this material. In particular we demonstrate, by light (LM), scanning electron (SEM) and transmission electron (TEM) microscopic examination, that both the host bone surface/overlay interface and the interface created between the overlay surface and the extraceilular matrix elaborated by the migrated cells may be used to model aspects of the known in vivo behaviour of bioactive materials.
MATERIALS AND METHODS Bioactive glass The bioactive glass, known to promote direct bone-bonding 15, used in these experiments was of the 45S5 type (prepared
91987 Butterworth s Co (Publishers) Ltd. O142-9612/87/040275-10503.00 Biomaterials 1987, Vol 8 July
275
The Biomaterials Silver Jubilee Compendium
36 Tissue response to bioactive g/ass: T. Matsuda and J.E. Davies
and supplied by Nippon Kogaku k.k., Japan). The glass 16 batches were prepared as described by Fujui and Ogino from reagent grade Na2CO3, CaCO3, H3PO4 and SiO2 by melting in a platinum crucible at 1 2 0 0 - 1 4 0 0 ~ for 2 h, stirred, maintained at that temperature for 30 min and cast into steel moulds. Following annealling at 5OO~ for 4-6 h the glass was cut, polished and fractured to produce particles of approx. 1 mm 3 and irregular shape. The samples were then stored in evacuated plastic bags which were sealed until required to avoid surface hydration.
Control Silica glass (99.99 wt% SiO2) was used as a non-bonebonding 15 control material. Fragments of similar size and shape as the bioactive glass were also supplied by Nippon Kogaku k.k.
Tissue culture method Neonate Albino Wistar rats (1-4 d old) were killed and the calvaria removed, using asceptic techniques, following reflection of the cranial skin. While maintaining the calvaria in phosphate buffered saline, (PBS) pH 7.2, containing 10% foetal calf serum (FCS) both the endocranial and extracranial periostea were removed to minimize fibroblastic contamination of the organ culture and to expose, on the endocranial side, the surface osteoblast population. A central area of each parietal bone, 2-3 mm square was then removed, devoid of sutural areas, and transferred to Bigger's culture medium (Fitton Jackson modification - - Gibco) containing 10% FCS,
glutamine 20 i~l/ml, penicillin/streptomycin 10 IA/ml and Hepes 2 5 iA/ml -- (sFJm) and the remainder discarded. The resulting square of host bone, with an intact endocranial osteoblast population, was then used as a source of primary osteoblasts. Figure 1 a illustrates these preparatory steps and Figure I b, the SEM appearance of the resultant osteoblast population. The test material fragment was then placed on this bone cell surface and maintained in culture for up to 4 wk in Linbro multiwell dishes containing 2 ml of culture medium. Figure 1 c shows an SEM photomicrograph of a host bone/ overlay organ culture unit, while Figure I d illustrates, diagrammatically, a typical cross-section of such a culture arrangement where osteoblasts from the bone surface have migrated over the lateral surfaces (LS) of the overlay to colonize the dorsal overlay surface (DS).
Preparation for microscopy Following the culture period the sFJm was replaced by buffered paraformaldehyde (pH 7.4) and fixed for periods up to 12 h. The whole sample, parietal square and test material overlay, were then dehydrated in a graded ethanol series and embedded in L.R. White (hard grade) resin cold-cured at 4~ Undecalcified sections, 0.5 ~um thick, for light microscopy were cut with either a glass knife, for bioactive glass, or a tungsten carbide knife, for quartz glass samples, in a Reichert-Jung 1140/Autocut, stained with toluidine blue and examined using a Leitz Labolux microscope. Samples for TEM were then cut from selected areas of the same blocks on a Reichert Jung OMU2 ultramicrotome using a
Figure 1 Summary of the experimental procedure. (a) A square of parietal bone is removed from the neonate rat and stripped of both endocranial and extracranial periostea (b) SEM appearance of the typical tesse/ated arrangement of endocranial osteoblasts. (Field width = 135 ~m). (c) The test biomatena/ fragment (or "overlay') is placed on the endosteal host bone surface as seen in this low power SEM photomicrograph. (d) Diagrammatic representation of the host bone/overlay organ culture unit. Osteoblasts from the endocraniai surface of the host bone may migrate up the lateral surfaces of the overlay material (LS) to colonize the dorsal surface (DS~ Two interface arrangements can then be described. Firstly, the interface between the host bone and the ventral surface of the overlay as seen in Figures 3 (a, b and c) and 4 (a and b ). Secondly, the interface created between the dorsal overlay surface and the colonizing bone cells as seen in Figures 6(b, c and d~ 7(a and b), 8(a and b) and 9.
2 76
Biomaterials 198 7, Vo/8 July
37
The Biomaterials Silver Jubilee Compendium Tissue response to bioactive glass." T M,~tsuda and J.E Davies
Figure 2 SEMs of osteoblasts from the endocrania/ surface of rat catvaria which have migrated onto overlays of quartz glass (a) and 45S5 bioactive glass (b) after a 1 wk culture period. Note that the cell coverage and migratory ceil morphology is similar on both overlays. (Field widths = 0.9 ram).
Figure 3 SEMs of the host bone surface after a t wk culture period where in (a) an overlay of quartz glass (QG) and in (b) 45S5 bioactive glass (BG) has been removed from the bottom left and top righthand side of the field of view respectively. In (a) beneath QG, the cell layer is clearly seen while in (b) spherical foci of calcium phosphate crystallization have remained attached to the host cell surface. (c) This shows a similar appearance to (b) but at a highermagnification.(d) This shows a similar, but continuous, calcium phosphate crystallization layer attached to the ventral surface of the removed bioactive glass overlay fragment. (Field width = (a) 6 t 5 pro; (b) 784 ~m; (c) 28 ~m and (d) 21 ~m.
diamond knife, mounted on 400 mesh uncoated grids, stained with uranyl acetate and viewed using a JEOL 100 CX II TEM. Whole samples were also prepared separately for SEM by fixation in 2.5% cacodylate buffered glutaraldehyde (pH 7.2), dehydrated, critical point dried from carbon dioxide, and sputter coated with either gold for secondary imaging or carbon for X-ray analysis prior to examination in an ISl 100A SEM (accelerating voltage 1 5 - 2 0 kV).
RESULTS The organ culture units comprising calvaria and either bioactive glass or quartz glass oveday fragments were maintained for up to 4 w k . After either 1 or 4 w k , observations were made on the affect of the overlay on the host calvarial surface, the migration of cells from the calvarial surface to colonize the overlay fragment and the elaboration of extraceltutar matrix by the colonizing cells.
Biomateriats 1987, Vol 8 July
277
The Biomaterials Silver Jubilee Compendium
38 Tissue response to bioactive glass: T. Matsuda and J.E. Davies
The host bone surface/overlay interface After short culture periods of up to 1 wk both osteoblast morphology and migration rate, from the host bone surface onto quartz glass control samples and 45S5 samples, were similar as assessed by the degree of cellular colonization of the overlays. The SEM appearance of such 1 wk cultures is shown in Figure 2 (a and b). After this initial period some fragments were removed to allow examination of the surface of the host bone which had been in direct contact with the overlay material. Scanning electron micrographs such as those shown in Figure 3 (a and b) demonstrated marked differences in the 'appearance of the developing interface with the control glass and bioactive glass overlays respectively. The former were characterized by a smooth, uninterrupted cell layer. These cells showed little dorsal cell surface activity in the form of blebs, ruffles or evidence of pseudopodial or filopodial connection to the previously overlying material. However the surface previously in contact with the bioactive 45S5 glass was significantly different in appearance and revealed numerous residual calcium phosphate-rich crystalline masses still adherent to the underlying cell layer. In many areas such crystalline formations completely obliterated the osteoblast cell layer directly below the bioactive glass overlay. It was of interest to note that, on removal of the bioactive glass, the fracture plane invariably occurred within the surface active calcium phosphate-rich layer rather than at the substrate/cell interface. Although we have previously reported such an appearance using both in vitro and in vivo assays 1s. 17, the present results demonstrate that such a layer was elaborated from discrete foci of calcium phosphate crystal growth to form, initially, spherical masses as shown in Figure 3 (b and c). The calcium phosphate composition of these crystalline masses was confirmed by energy dispersive X-ray analysis. Confirmation of the fracture plane was obtained by examining the surface of the bioactive glass fragment which had been removed. This also showed a calcium phosphate-rich surface layer, Figure 3d, but no evidence of spheritic masses. After longer culture periods of up to 4 wk the host bone surfaces, below the overlay materials, still demonstrated significant differences. Figure 4 a shows the appearance, after 4 wk, of the host bone surface after removal of a quartz glass overlay. The outline of the overlay site was clearly
marked by a continuous raised edge of cells which were torn when the overlay was removed. The surface cells at this interface were of similar appearance to those in Figure 3 a and showed no evidence of contact with the ventral surface of the overlay material. In contrast, Figure 4 b is the equivalent appearance found beneath the 45S5 surface. Again the outline of the overlay was clearly marked by a torn tissue layer but which was considerably thicker than that observed around quartz glass, due to the greater number of cells which had migrated onto the bioactive glass during the 4 wk culture period. The interface exposed by removal of the overlay was also significantly different both from that of quartz glass and the appearance noted in 1 wk cultures of bioactive glass. Figure 4 b shows the cellular layer with clear indication of fibrous tissue formation between cells and cell processes torn as a result of the removal procedure. It is assumed that these cells were connected, by their cell processes, to others which remained adherent to the bioactive glass surface during removal of the overlay. Thus, the created fracture plane passed through biological tissue, rather than the calcium phosphate layers of the overlay material, and after this 4 wk period the bioactive glass/ tissue interface had resisted disruption.
Elaboration of extracellular matrix on the overlay surface We have recently shown, using polarized light microscopy of similar 4 wk culture samples, that the ECM elaborated on bioactive glass by migrated osteoblasts was birefringent 6. Due to the thickness of the cell and ECM layer formed on the dorsal bioactive glass overlay surface, the elaborated tissue was easily visible by LM, Figure 5 a. The quartz glass surface however, while colonized by migrated cells, showed a considerably thinner tissue layer when compared to 45S5 and no detectable intercellular birefringence at LM level, Figure 5 b . The critical point drying procedures employed to prepare calvaria and overlays for SEM observation commonly resulted in the tissue layers, formed during a 4 wk culture period on bioactive glass, fracturing due to differential contraction of the artificial and biological materials. These fracture sites allowed the elaborated tissue to be examined [Figure 6 (a-d)] and demonstrated the formation of sheets of material of approx. 1 IJm thickness interposed between successive cell layers. These sheets of ECM, shown in Figure 6 (c and d), were perforated by cell bodies or cell processes
Figure 4 SEMs of the host bone surface after a 4 wk culture period. (a) The surface beneath a quartz glass overlay. (b) The surface beneath a bioactive glass overlay. Note that the cell surface in(a) is smooth with no sign of previous cell contact with the overlay material in contrast to the appearance in (b). Also, the torn cell layer surrounding the overlay area in (b) is significantly thicker than in (a). (Field width = (a) 1107 pm and (b) 830/Jm.
2 78
Biomaterials 198 7, Vol 8 July
39
The Biomaterials Silver Jubilee Compendium Tissue response to bioactive glass: T. Matsuda and J.E. Davies
and possessed a granular surface. In limited areas a more typical fibrillar collagenous appearance was observed
(Figure 6 c).
Figure 5 Light micrographs of O.5 pm thick sections prepared from (a) 45S5 bioactive glass and (b ) quartz glass (see also Figure 1 d) after a 4 wk culture period. Note that the tissue elaborated on the dorsal BG surface is significantly thicker than that seen on QG. Neither sample was treated, prior to sectioning, to remove the glass and shards of BG can be seen both at the tissue interface and throughout the upper part of the section. (Field width = 2. 5 mm).
To confirm the presence of collagen in the extracellular matrix observed on 45S5 by both LM and SEM, sections were cut for TEM observation. These were prepared from the same blocks as had been used for LM (Figures 1 d and 5 a). Figure 7 (a and b) shows the TEM appearance of the tissue present on the dorsal surface of a bioactive glass fragment. It is clear from this Figure that the extraceilular compartment was composed mainly of collagen fibres which exhibited a layered appearance with fibre direction generally alternating in successive layers. In close proximity to the cells fibril orientation was random (Figure 7) while the bulk of the intercellular space was filled with a more ordered fibre organization. Interspersed between the thick, collagen containing ECM were cells of slightly flattened morphology which possessed some elongated cell processes. The surface calcium phosphate-rich layer of the bioactive glass, equivalent to that illustrated in Figure 3 d, could easily be seen on the shards of glass which remained at the glass/ tissue interface. At this interface two morphological arrangements were observed. The first, illustrated in Figure 7a demonstrated an arrangement, which we refer to as Interface Type I, where the immediate glass/tissue interface was occupied by a clear zone, approx. 0.1-0.2 IJm thick and which was, apparently, devoid of structured tissue. The second, illustrated in Figure 7b, exhibited direct contact
Figure 6 A ser~es of SEMs of increasing magnification of the tissue layer covering a bioactive glass overlay after a 4 wk culture period. (a) The overlay has been covered with multilayers of migrating cells which have elaborated an extracellular matrix. The fracturing of this tissue covering is due to differential contraction during the critical point drying preparation for SEM observation. Note the number of dead ceils (round cells) on the surface of the vital cell layer afterthis culture period and compare with Figure 2 (a and b), (b) The same specimen at higher magnification showing the sheets of tissue elaborated on the overlay material (c) Higher magnification of the central area of (b) showing a small group of cells between two parallel sheets of extracellular material Note: while the surface of this material is granular in appearance, a fibrous arrangement similar to that of collagen fibres is seen in the immediate vicinity of the cells (white arrows). (d} Another area of the same extracellular matrix clearly showing both its granular surface appearance and the presence within it of a smoothly outlined cavity containing a cell process. Some shrinkage of the cell has taken place, during preparation, as witnessed by the torn filopodia deep within the cavity. (Field width = (a) 1.1 ram; (b) 270 tJm; (c) 17 lzm and (d) 9 pro.
Biomateriats 198 7, Vol 8 July
279
40
The Biomaterials Silver Jubilee Compendium Tissae response to bioactive glass: T. Matsuda and J.E. Davies
Figure 7 TEMs of the dorsal cell-colonized surface of the 45S5 bioactive glass overlay showing the elaborated collagen-containing extracellular matrix (ECM). (a) The interfacial tissue arrangement, interface Type I, is characterized by a "clear'zone immediately adjacent to the calcium phosphate (CP)-rich surface of the overlay. The electron dense layer above the clear zone (white arrows) is the extension of a neighbouring cell body. (b) Interface Type ii. Here the collagen fibrils of the ECM are seen to be elaborated between the nearest cell and the bioactive glass surface. Some fibrils are seen to interdigitate with the latter (white arrows). (Original magn. = 5.8 K; Field height = 14.5 pm).
between and interdigitation of the collagenous ECM produced by the migrated osteoblasts and the calcium phosphate-rich surface layer of the bioactive glass. We refer to this morphological arrangement as Interface Type I1. Both these appearances were observed at the dorsal ECM/overlay interface of the same glass fragment. Figure 8 a shows Interface Type I at higher magnification where the clear zone was separated from the overlying collagenous ECM by an electron dense layer. In the field immediately adjacent to that shown in Figure 7 a this layer could be seen to be continuous with a cell body which was closely apposed to, but not touching, the glass surface. The
280
Biomateriais 1987, Voi 8 July
cellular nature of this laver is more easily seen in Figure 8 b which shows two neighbouring cell processes in contact and a dorsal filopodium extending into the fibril containing compartment. While the clear zone, of variable thickness, did contain an occasional collagen fibril (Figure 8 b), we saw no evidence of fibre organization in this zone or interdigitation with the 45S5 surface. Figure 9 shows Interface Type II at higher magnification. Here collagen fibres, produced by the overlying cell, were randomly orientated in the subcellular space and in some areas interdigitated with the calcium phosphate surface of the glass.
The Biomaterials Silver Jubilee Compendium
41
Tissue response to bioactive glass. T Matsuda and J.E. Davies
Figure 8 TEMs of the structure of Interface Type 1. The dense black material at the lower left side of each photomicrograph is the calcium phosphate-rich surface of the bioactive glass (CP).(a) The "clear'zone is immediately adjacent to the glass and separates the latter from an overlying cell process. Above the cell process arrays of collagen fibrils are easily seen. (b) The thickness of the "clear'zone is variable and here can be seen to contain an occasional collagen fibril (black arrow). The identification of the cell layer is facilitated here by the overlapping of two ceil processes and a dorsal fifopodiat extension from one cell (k~. (Original magn. = 48 K: field width = 1.79 urn).
DISCUSSION The aim of the present work was to observe the reaction of primary osteoblast cell cultures to a bioactive glass surface and attempt to reproduce in v i t r o some of the known in vivo characteristics of such materials. Such potential testing procedures are of considerable importance in the biomaterials field since, not only could they be adapted to provide a biological batch-testing assay for bioactive bone-substitute materials, but they also provide a means of investigating the intimate step-by-step interactions occurring at the tissue/material interface using relatively simple techniques when compared with in vivo implantation. Indeed, while Greenlee e t al. t8 showed that in vivo systems could be used to study the effect on tissue behaviour brought about by changing implant composition, they recommended that in v i t r o systems be devised, for evaluating materials variables, which would be less laborious than in vivo methods. A major problem with the in v i t r o use of osteoblasts, however, is that when they migrate, for example from a calvarial surface to colonize an overlay material, they express a modulated phenotype which results in not only a morphological change but also functional synthetic variations 8. This problem is exacerbated if suspended cell populations are seeded into tissue culture flasks which provide a low cell : recipient surface area ratio. Thus, migrated osteoblasts may, or may not, produce a collagenous extracellular matrix in vitro. Furthermore, this matrix production may, or may not, be calcified by the nucleation of calcification foci due to matrix vesicle release. To achieve the latter in v i t r o it is necessary to supplement the culture medium with both ascorbic acid and a source of organic phosphate19. Although we have recently described an alternative technique to longterm in v i t r o cultures using implanted diffusion chambers 8, in this present work we are only concerned with the -primary stage of ECM production i.e. collagen fibre formation. The use of rat calvaria as a source of bone cells in biomaterials testing is not new. Wilson e t al. 2~ listed a range of in v i t r o and i n vivo toxicity assays for bioactive glass systems. While they reported that rat calvarial bone cells both spread and divided on bioactive glass substrates, the
light micrographic appearance of their cells (at 29d) resembled the single cell layer which we have demonstrated may be achieved after only a 1 wk culture period. Although they gave no details of the experimental methods employed, it can be assumed that these, contrary to the method we report here, did not allow the osteoblast population to retain their phenotype and elaborate an extracellutar matrix. However, it is clear, from the present results, that the in v i t r o method we employ is capable of reproducing in vivo behavioural characteristics of 45S5 bioactive glass and thus may be used to investigate the nature of the created bone cell/biomaterial interface.
Bone surface/overlay interface The formation of a calcium phosphate-rich layer on the surface of bioactive glass on creation of a solid/liquid interface is a well known phenomenon which is considered to be of central importance in the biological-bonding capacity of such materials 1,2. The initiation of this layer has been described by Ogino and Hench 2t to occur in two stages; the rapid accumulation of Ca and P-rich species on the glass surface followed by a maturation phase. Approximately 1 h after the creation of the solid/liquid interface, a 200 nm thick calcium phosphate layer is formed which subsequently grows thicker due to migration of additional Ca and P species from the bulk glass in a time-dependent fashion. The rough surface which this imparts to the glass has recently been described by Matsuda e t ai. 17 as being composed of small pleats of about 70 nm thickness and 370-1 500 nm in height and which we confirm in the present work. We saw no evidence of the amorphous gel layer discussed by Clark et al. 22. However the presence of discrete foci of calcium phosphate crystal growth to form, initially, spherical masses as shown in F i g u r e 3 (b and c) indicated that while discrete nucleation sites may characterize the initiation of this surface layer, the individual calcification foci fuse to form a continuous layer. The formation of this surface layer will already be complete, in the present experimental system, before the osteoblasts from the host bone have had time to migrate onto the dorsal overlay surface. On the ventral overlay
Biomaterials 1987, Vol 8 July
281
The Biomaterials Silver Jubilee Compendium
42 Tissue response to bioactive glass: T. Matsuda and J.~. Uawes
surface, however, this chemical reaction will take place directly above the apposed osteoblast surface as soon as the overlay is introduced into the culture system. At this interface the bone cells would not be able to immediately adjust to the changed environment created by the overlay and thus there will be a delay before the open lattice work of the calcium phosphate layer is invaded by macromolecular components of the ECM. Thus it may be assumed that the maturation of this ventral surface calcium phosphate-rich layer occurs with concomitant production of ECM components by the underlying cells. The interdigitation of ECM and calcium phosphate is time-dependent and after 1 wk the tissue/glass interface is sufficiently matured to prevent separation of the components when the overlay is removed from culture. Instead, the created fracture plane occurs within the calcium phosphate layer which therefore must be weaker than the cell-to-cell binding forces on the host bone surface layer. With time, this inorganic layer stabilizes and becomes more infiltrated with the fibre component of the
elaborated extracellular matrix so that at 4 wk the junctional complex is resistant to disruption, when the overlay is removed from culture, and the created fracture plane occurs within the cellular layers on the host bone. As it is assumed that the initial calvarial preparation provides a single cell layer of osteoblasts, some cell proliferation must have taken place in this compartment of the culture unit as well as the more obvious proliferation associated with the cells migrating to colonize the dorsal overlay surface. The development of this bone surface/overlay interface which we describe reflects the similar sequence of events known to occur in vivo where the created interface is sufficiently stable to resist disruption during push-out testing and fracture takes place either within the biological tissue or between the calcium phosphate layer and the bulk glass 16.23,24 Thus, from these observations of the host bone surface alone, it is clear that comparisons can be made between the tissue reactions promoted by bioactive and
Figure 9 TEM showing the structure of Interface Type !1. The bioactive glass surface (BG) is separated from the overlying cell process (C) containing rough endop/asmic reticulum (RER) by a collagen-containing extraceilutar space. The 64A cross-banding typical of collagen is easily visible throughout the field of view and it is clear that some of these fibrils are intimately associated with the overlying cell process. At various points (white arrows) collagen fibrils are also intimately interdigitated with the surface of the 45S5 glass surface layer. (Original magn. = 19 I~" field width = 29.6 pro).
282
Biomaterials 198 7, Voi 8 July
The Biomaterials Silver Jubilee Compendium
43 Tissue response to bioactive glass: T Matsuda and J.E. Davies
control materials which may be correlated with their known
in vivo behaviour.
Elaboration of extracellular matrix Clearly, this in vitro method will permit observation of the elaborated ECM on the dorsal overlay surface by light microscopy. The birefringent nature of the ECM produced by migrated osteoblasts on bioactive glass has been previously reported, together with the demonstration of calcified ECM on other calcium phosphate ceramics using the von Kossa technique in similar experimental methods 6. In view of the marked differences at LM level seen between bioactive glass and the control quartz glass during the experimental time period, together with the time-dependent elaboration of tissue on the dorsal bioactive overlay surface, it is conceivable that quantitative, morphometric, methods could be established to enable different batches of bioactive substrates to be compared or the effect of changing the chemical composition of the substrate on tissue reaction be assessed at LM level. With regard to the SEM appearance of the elaborated ECM, the multilayering effect is clearly seen with cells lying in lacunae and surrounded by the matrix they have produced, as is typical for connective tissue cells and confirmed by our TEM observations. It is quite evident that the SEM appearance reported here is atypical of collagen and could, in part, be due to surface coating of collagen by other macromolecutar species combined with physical effects due to preparation procedures. In particular the disorganized fibrillar arrangement, seen in F i g u r e s 7 and 9 in the immediate vicinity of actively secreting cells would result in an SEM appearance of randomly orientated fibril ends in the walls of the cell lacunae. We assume that the granular appearance may either be an accentuation of these fibril terminations due to focal deposition of gold during the sputtering process or possibly the appearance of early calcification of the collagenous ECM. This calcification could be caused by the release of calcium and phosphate ions from the surface of the 45S5 rather than cellular activity; however, we have not, as yet, seen any evidence of calcification at TEM level. We are currently re-examining similar uncoated material at tow SEM accelerating voltages in order to increase the resolution of our photomicrographs at these magnifications. Beckham et ai. 25 stated that SEM was not an ideal tool to provide sufficient resolution to examine the tissue reactions at a glass implant interface following implantation. Nevertheless, it is clear from the present work that SEM does yield valuable information concerning tissue organization especially when combined with both light and transmission electron micrography. The TEM appearance of the bioactive glass/tissue interface reported by Beckham e t al. 25 (Ioc cit) demonstrated the formation, in vivo, of collagen fibres in contact with a bioactive glass surface. By comparison of Figure 3 in Reference 2 5 with F i g u r e s 7 b and 9, in the present work, it is clear that we have successfully reproduced this behaviour in vitro.
It is not possible from our present results to state whether the collagen/surface apatite interface, Interface Type Ii, was progressively developing or whether, given longer culture periods, a larger percentage of surface contiguity could be achieved up to the 95% level recorded in vivo 2, 26. It could be that Interface Type I represents an early developmental stage which, following future collagen deposition, will be converted to Interface Type I!. However
another possibility also exists i.e. that regional variations in the physicochemical nature of the biomaterial surface are the causative factors for these two morphologically distinct arrangements. The progressive fixation of the 45S5 series was anticipated by Beckham e t a / . 25 and subsequently described in detail by Greenlee e t al. 2~ who showed that implant stability was gained after 4 wk implantation. However, the mechanisms of these interfacial reactions are still not understood 15, in part due to the complex heterogeneous celt and tissue fluid environment which characterizes the in vivo situation. Thus we propose that the in vitro system described here is suitable for detailed investigation of the creation of a bioactive biomaterial/tissue interface.
CONCLUSIONS (1) This in vitro assay has enabled the direct collagen/ bioactive glass interface to be established. This interface is a critical component of the known in vivo bonebonding mechanism associated with bioactive bonesubstitute implant materials and we have shown here that this behaviour may be reproduced in vitro. (2) The creation of both collagen binding and non-collagen binding interfaces at bioactive glass surfaces can be demonstrated in vitro. (3) These two morphologically distinct interfaces have been designated as Interface Type I (no collagen binding to the bioactive glass surface) and Interface Type II (direct interdigitation of collagen with the bioactive glass surface). (4) This in vitro assay can be developed to facilitate the biological batch testing of bioactive bone-substitute materials.
ACKNOWLEDGEMENTS We are grateful to Yvonne Bovell, Mark Dallas and Alan Murdoch for technical assistance and Lyndsey Davies for manuscript preparation. We should also like to thank Dr Ulrich Gross for supplying a preprint manuscript of Reference 9. This work was financially supported by Kagoshima Prefecture, Japan and The Medical and Dental Faculty of the University of Birmingham. The development of the in vitro system described was supported by Grant No. G R / D / 5 3 9 1 3 from the UK Science and Engineering Research Council.
REFERENCES t 2 3 4 5
Hench,L.L., Splinter, R.J., Allen, W.C. and Greenlee,T.K., Bonding mechanisms at the interface of ceramic prosthetic materials, J. Biomed. Mater. Res. Syrup. 1971, 2, 11 7-141 Hench,L.L. and Paschall, H.A., Direct chemical bond of bioactive glass-ceramic materialsto boneand muscle J. Btomed. Mater. Res. Symp. 1973, 4, 25-42 Hench,L.L.and Ethridge,E.C.,Biomateriats, ,4n lnterfaciaiApproach, Academic Press,New York, 1982 Gross, U.M. and Strunz, V., The anchoring of glass-ceramics of different solubility in the femur of the rat, J. Biomed. Mater. Res. 1980, 14, 607-618 Gross,U., Strunz, V., The interface of various glasses and glass ceramicswith a bonyimplantationbed,J. Biorned. Mater. Res. 1985. 19, 251-27t
Biomaterials 1~,~87, Voi 8 July
283
The Biomaterials Silver Jubilee Compendium
44 Tissue response to bioactive glass." T. Matsuda and J.E. Davies
6
7 8
9
10 11
12
13 14
284
Davies, J.E., Tarrant, S.F. and Matsuda, T., Interaction between primary bone cell cultures and Biomaterials. Part 1. Method: The in vitro and in vivo stages, Presented at 6th Europ. Congr. Biomaterials, Bologna, Italy, September 14-1 7, 1986, Elsevier, Amsterdam (in Press) Jones, S.J. and Boyde, A., The migration of osteoblasts, Cell Tiss. Res. 1977, 184, 179-193 Tarrant,S.F. and Davies, J.E., Interactions between primary bone cell cultures and biomaterials. Part 2. Osteoblast modulation, Presented at 6th Europ. Congr. Biomaterials, Bologna, Italy, September 14-1 7, 1986, Elsevier, Amsterdam (in Press) Gross,U., Schmitz, H.J., Kinne, R., Fendler, F.R. and Strunz, V., Tissue or cell culture versus in vivo testing of surface-reactive biomateriats, Presented at 6th Europ. Congr. Biomaterials, Bologna, Italy, September 14-17, 1986, Elsevier, Amsterdam (in Press) Jones, S. and Boyde, A., Colonization of various natural substrates by osteoblasts in vitro, Scanning Electron Microscopy 1979, II, 529-538 Davies,J.E., Hurst, R.P. and Spooner, N.T., Surface emission and biological probes for inorganic interfaces, Presented at the Royal College of Surgeons IBMS Symposium, London, November 1984. In, Interactions of Cells with Natural and Foreign Surfaces, (Eds N. Crawford and D.E.M. Taylor), Plenum Press, New York, 1986, 95-108 Rout, P.G.J., Tarrant, S.F., Frame, J.W. and Davies, J.E., Interaction between primary bone cell cultures and biomaterials. Part 3. A comparison of dense and macroporous hydroxyapatite, Presented at 6th Europ. Congr. Biomaterials, Bologna, Italy, September 14-1 7, 1986, Elsevier, Amsterdam (in Press) Davies,J.E., Causton, B., Bovell, Y., Davy, K. and Sturt, C.S., The migration of osteoblasts over substrata of discrete surface charge, Biomaterials 1986, 7, 231-234 Shelton, R.M., Whyte, I.M. and Davies, J.E., Interaction between primary bone cell cultures and biomaterials. Part 4. Colonization of charged polymer surfaces, Presented at 6th Europ. Congr. Biomaterials, Bologna, Italy, September 14-17, 1986, Elsevier, Amsterdam (in Press)
Biomateria/s 1987, Vol 8 July
15 16 17 18 19 20 21 22
23
24
25 26
Ito, G., Matsuda, T., Inoue, N. and Kamegan, T., An histologic comparison of the tissue interface of Bioglass| and silica glass, J. Biorned. Mater. Res. (in Press) Fului,T. and Ogino, M., Difference of bone-bonding behaviour among surface active glasses and sintered apatite, J. Biorned. Mater. Res. 1984, 18, 845-859 Matsuda, T., Yamauchi, K. and Ito, G., The influence of B'ioglass| on the growth of fibroblasts, J. Biomed. Mater. Res. (in Press) Greenlee,T.K., Beckham, A.Ro, Crebo, A.R. and Malmorg, J.C., Glassceramic bone implants, J. Biomed. Mater. Res. 1972, 6, 235-244 Tenenbaum,H. and Heersche, J.M.M., Differentiation of osteoblasts and formation of mineralized bone in vitro, Calc. Tiss. Int. 1982, 34, 76-79 Wilson, J., Pigott, G.H., Schoen, F.J. and Hench, L.L., Toxicology and biocompatibility of Bioglass| J. Biomed. Mater. Res. 1981, 15, 805-817 Ogino,M. and Hench, L.L., Formation of calcium phosphate films on silicate glasses, J. Non-Cryst. Sol. 1980, 38 Et 39, 673-678 Clark,A.E., Hench, L.L. and Paschall, H.A., The influence of surface chemistry on implant interface histology: A theoretical basis for implant materials selection, J. Biomed. Mater. Res. 1976, 10, 161-174 Greenspan,D.C., Piotrowski, G. and Hench, L.L., Mechanical evaluation of bone-bioglass bonding, in An Investigation of Bonding Mechanisms at the Interface of a Prosthetic Material, US Army Medical Research and Development Command, Contract No. DAMD 1 7-77-C-6033, Report No. 7 1979, pp 24-39 Ogino, M. and Fatami, T., Chemical and biological properties of biological ly active glasses -- compositional dependence of the surface reaction and bone-bonding strength, (11-12), Trans. 2nd Sym. on Apatite, Tokyo, Japan, December 1-2, 1986 (in Japanese) Beckham,C.A., Greenlee, T.K. and Crebo, A.R., Bone formation at a ceramic implant interface, Ca/c. Tiss. Res. 1971, 8, 165-1 71 Gross,U., Brandes, J., Strunz, V., Bab, I. and Sela, J., The ultrastructure of the interface between a glass ceramic and bone, J. Biorned. Mater. Res. 1981, 15, 291-305
45
The Biomaterials Silver Jubilee Compendium
Activation of the complement system at the interface between blood and artificial surfaces M.D. Kazatchkine and M.P. Carreno
Unit~ d'tmmunopathologie and INSERM U28, H~pital Broussais, 96 rue Didot, 75014 Paris, France Presented at Biointeractions "8 7, Cambridge, UK in July 1987
The interaction between blood and the artificial devices used in haemodialysis results in the activation of the immune system. Complement activation is known to play a key role in the production of inflammatory mediators. This article describes the complement system and the way in which it is triggered in the patient undergoing haemodialysis. It highlights those factors thought to be of particular importance in the pathogenesis of adverse reactions and targets them as obstacles to be overcome in the future design of biocompatible materials. Keywords: Complement activation, haemodialysis, biocompatibility, artificial surfaces
The interaction between blood and artificial devices used in extracorporeai circulation results in the activation of a number of humoral and cellular processes involved in natural (i.e. "non-specific') and specific immunological recognition of foreign surfaces by the host 1. Among these, activation of the complement system appears as a key event that elicits the secondary production and release of inflammatory mediators by various leucocyte subsets which express specific membrane receptors for proteins derived from complement activation.
COMPLEMENT ACTIVATION -- THE 'CLASSICAL" AND 'ALTERNATIVE' PATHWAYS The human complement system comprises 19 component and regulatory plasma proteins and at least nine distinct cellular receptors for some of these proteins or their cleavage fragments 2'3. The contact of blood with an "activating' surface triggers complement activation through one of two pathways, 'the classical" or the "alternative'. Both pathways form specific enzymatic complexes named 'C3 convertases' that cleave the third component of complement, C3, generating the anaphylatoxin C3a and a major cleavage fragment, C3b (Figure 1). A labile binding site is transiently expressed on nascent C3b that allows C3b to bind covalently to bystander surfaces. This reaction transfers molecular interactions between complement proteins from the fluid Correspondenceto Dr M.D. Kazatchkine.
phase onto the target surface of complement activation. It occurs through a transesterification mechanism involving a glutamyl group from a cleaved internal thioester bond in the C3 molecule and OH- or NH 2 groups on the acceptor surface (Figure 2) 4'5. Surface-bound C3b may then interact with alternative pathway proteins to form the amplification C3 convertase which augments C3 cleavage and the subsequent deposition of C3b molecules on the activating surface. Accumulation of C3b molecules in the vicinity of C3 convertases on the target surface changes the specificity of the C3 cleaving enzymes into C5 convertases, resulting in the cleavage of C5, generation of C5a and recruitment of the terminal effector sequence C5-C9. The classical pathway of activation is primarily initiated by antigen-antibody complexes, although it may also be activated in the absence of antibody by some crystals or bacterial and virus surfaces, and by complexes between positively and negatively charged molecules such as those formed between heparin and protamineS. The alternative pathway is activated by surfaces which exhibit specific biochemical characteristics allowing bound-C3b to initiate the assembly of the amplification C3 convertase on that surface. Since activation may occur in the absence of antibodies, the alternative pathway is the complement pathway that is usually triggered when plasma is exposed to artificial surfaces. Activation of the alternative pathway involves the following sequence of events T: binding to the activating surface of few C3b molecules derived from the low grade cleavage of C3 that continuously occurs in normal plasma; Mg++-dependent binding of Factor
91988 ButterworthEl.Co (Publishers) Ltd. 0142-9612/88/010030-06503.00 30
Biomaterials 1988, Vot 9 January
46
The Biomaterials Silver Jubilee Compendium Complement system activation." M.D. Kazatchkine and M.P. Carreno
AMPLIFICATION CLASSICAL
PATHWAY
ACTIVATION
C1 C4 C 2
INITIAL
.
Clinh
ALTERNATIVE . . . . . . . .
.
C3b, B D
PATHWAY
ON
ACTIVATING
SURFACES
H, [
C4bp,[
PATHWAY
ACTIVATION
C3b_like C3,B.D
"~.~C3a
= C 3 a desArg
C3b
l
C5 C O N V E R T A S E
[
ACTIVATION
OF THE
COVALENT
BINDING
TO
SURFACES -
[
L---.---.~lnteraction
~lnact/vat
with
ion
CR1
into
on
cells
iC3b
FORMATION
TERMINAL
SEQUENCE
C5b _C9 Figure 1 Complement activation. Cleavage of C3 may result from the assembly of the classical pathway C3 convertase, the "in/t/a/" alternative pathway C3 convertase or the amplification C3 convertase which is formed on surfaces that activate the alternative pathway. The C3b fragment of C3 may covalently bind to surfaces expressing OH- or NH 2 "acceptor" groups and serve to assemble the C3 amplification convertase and the C5 convertase on the target surface. Surfacefixed C3b may also interactwith cells which express the C3b receptor, CR 1. C3b is irreversibly inactivated into iC3b when it is bound to non-activating surfaces.
B to surface-bound C3b; cleavage of B within the C3b, B complex by the serine protease, Factor D, resulting in formation of the amplification convertase C3b, Bb; cleavage of C3 by surface-bound C3b, Bb, followed by deposition of additional C3b molecules and formation of new convertase complexes on the surface. The enzymatic activity of the C3b, Bb amplification convertase is regulated by the spontaneous dissociation of the complex (which has a half-life of only 3 min at 37~ the stabilizing effect of protein P which decreases the rate at which the complex dissociates and the inhibitory effect of Factor H which accelerates the dissociation of C3b, Bb even when the convertase complex has been stabilized by P. Factor H decay dissociates the C3 amplification convertase by binding to C3b and displacing Bb from C3b, Bb. It also serves as a cofactor for irreversible cleavage-inactivation of C3b into C3bi by another plasma regulatory protein, Factor I. Discrimination between 'act/vat/ng' and 'non-activating' surfaces in whole blood depends on the outcome of the competition between B and H for binding to surface-fixed C3b. On a 'non-activating' surface, bound-C3b interacts with
f
r NATIVE
C3
H with an almost 100-fold greater affinity than that with which it interacts with B. In contrast, on an activating surface, the uptake of H is not favoured relative to the uptake of B, so that the amplification convertase C3b, Bb is allowed to form by "escaping' from the decay-dissociating action of H 8'9. The affinity of H for surface-bound C3b is influenced by biochemical characteristics of the particle surface such as its relative content in sialic acid 9, heparin10 or carboxymethyl groups (Figure 3) 11 and, with regard to circulating cells, in 100
-
c) 5 o -
D
"o
';
_
A
0
NASCENT
O-
C3b
I
I
I
CM
R--O----~
[ II 0
R--NH--C
I II 0
Figure 2 Schematic representation of the internal thioester bond in the alpha chain of C3 and the mechanisms of covalent binding of nascent C3b to surfaces bearing reactive OH or NH 2 acceptor groups.
|
group
[
I
1
t
i
0.5
per
unit
1
1
sugar
Figure 3 Dose-dependent suppression of the complement-activating capacity of Sephadex with increasing substitution of saccharidic units with carboxymethyl groups. Nine milligrams of unsubstituted or substituted Sephadex G25 were incubated with 1 m l of normal human serum diluted 1 : 4 in veronal buffered saline containing Ca ++ (0.15 mM) and Mg ++ (0.5 mM) for 60 min at 3 7 ~ Residual CH50 activity was measured and expressed as a percentage of the initial value (time zero of incubation}.
Biomateriafs 1988, Voi 9 January
31
47
The Biomaterials Silver Jubilee Compendium Complement system activation: M.D. Kazatchkine and M.P. Carreno
specific regulatory membrane proteins (e.g. decay accelerating factor or DAF) which normally protect the cells from attack by autologous complement. For example, cuprophane is a potent activator of the alternative pathway in v i t r o 12,13 and in vivo in patients undergoing haemodialysis with first-use cellulosic membranes 14'16. The membrane has a high density of exposed hydroxyl groups providing a high number of sites for the covalent binding of C3b. In addition, the C3b-binding sites on the polysaccharidic membrane are protected from inactivation by H ~7, which may trigger formation of alternative pathway amplification C3 convertase sites, in whole plasma, alternative pathway activation by cuprophane membranes may further be enhanced by natural or acquired antibodies. IgG anti-dextran antibodies have been shown to enhance alternative pathway activation by Sephadex particles of which the basic chemical structure is similar to that of cuprophane11' 18. Zymosan, another alternative pathway activating polysaccharide, also requires specific antibodies to efficiently activate the alternative pathway ~9. It has also been shown that the serum of patients who exhibit maximal complement activation during haemodialysis with cuprophane membranes show maximal reactivity to zymosan in v i t r o 2~ These observations suggest that the differences in the amount of complement activation observed between patients dialysed with cuprophane relate to the concentration of anti-dextran antibodies in their serum. Anti-polysaccharidic antibodies reactive with cellulosic membranes could be specifically acquired following repeated exposure of the individual to the dialysis membranes, or represent cross-reactive antibodies against irrelevant (e.g. bacterial) antigens, similar to those which induce acute hypersensitivity reactions to dextran derivatives. Cellulose acetate membranes and reused cuprophane membranes saturated with covalently bound C3b, induce less complement activation than cuprophane because they provide fewer C3b-binding sites/surface unit2~; polyacrylonitrile membranes do not activate complement because they do not express binding sites for C3b. Thus, in order to be a non-activating surface of the alternative pathway, a biomaterial should either be incapable of covalently binding C3b (e.g. polyacronitrile membranes), possess surface characteristics that will result in preferential binding of H over that of B to surface-bound C3b (e.g. heparin-coated surfaces) or express few or no antigenic sites that may be recognized by natural or acquired, specific or cross-reactive antibodies.
FACTORS IMPORTANT TG THE DESIGN OF BIOCOMPATIBLE SURFACES The capacity of surface-bound heparin to inhibit alternative pathway activation may be of particular interest for the design of biocompatible surfaces. Covalent coupling of heparin to zymosan suppresses the capacity of the polysaccharide to activate the alternative pathway 1~ The inhibitory effect of heparin depends on the increased ability of H to bind to C3b on the heparin-coated surface ( F i g u r e 4). Heparin in the fluid phase also inhibits alternative pathway activation; heparin exerts its inhibitory effect on the binding site of C3b for factor B and requires O-sulphation and N-substitution of the molecule22; the site which is responsible for the anticomplementary effect of heparin is distinct from that which binds anti-thrombin lit, as demonstrated by the similar inhibitory capacity on complement activation of heparins with high and low affinity for anti-thrombin and of synthetic pentasaccharides endowed with or devoid of anti-Xa activity.
32
Biomateriats 1988, Vo/ 9 January
F
,
O0 60
I--
I'-
(3
20
(3
& 10
I
I
t5
30 MINUTES
..a, ........................
45
,ill
......
60
Figure 4 Lack of complement activation by heparin-coated surfaces. The figure depicts the time course of inactivation by H and I of C3b bound to zymosan (/k ) and to zymosan to which heparin has been covalently coupled ( A ) (adapted from reference I 0).
The inhibitory effect of heparin on complement activation may be mimicked by synthetic dextrans substituted with sulphonated and carboxylic groups 23"24. C5 convertase is formed on activating surfaces following the deposition of C3b molecules on the target surface in the vicinity of C3 convertase complexes. The efficiency of C5 convertase formation relative to that of C3 convertase formation differs from one activating surface to another. The factors which determine the relative "coupling efficiency" between C3 and C5 cleavage on a surface have not yet been fully characterized. These factors are probably of importance in determining the clinical tolerance of an artificial surface because of the key role of C5a in the pathogenesis of adverse reactions occurring during extracorporeal circulation. Cleavage of C5 by the C5 convertase releases C5a and a larger fragment C5b. Nascent C5b may form a macromolecutar complex with C6, C7, C8 and several molecules of C9 which either inserts into biological membranes (cytolytic complex) or binds the S protein in plasma to form a fluid phase cytolytically inactive SC5b-9 complex ( F i g u r e 5) 25. Small amounts of terminal C5b-9 complexes are generated during haemodialysis since, in contrast with observations in cardiopulmonary bypass, no increase in the concentration of SCSb-9 complexes is found in the plasma of patients dialysed with cuprophane membranes 26. Of potential relevance, however, are the recent observations that low amounts of C5b-9 generated on leucocytes may induce neosynthesis and release of leukotrienes from the cells 27. Thus, the following proteins and complexes derived from complement activation may be involved in the initiation of bioincompatible reactions: 1. The soluble peptides C3a, C5a, C3adesArg and C5adesArg released by the cleavage of C3 and C5 (see below).
The Biomaterials Silver Jubilee Compendium
48
Complement system activation: M,D, Kazatchkine and M.P. Carreno C~)
CONVERT&SE
FOR'M&TION
ON
&CT I V ' A T t N G
SLI RF Jl, CE S
~e. ~ ........ ,,, C i l ~
C ~)ll dls&qr 9
CSl
ACTio,, w,r. RECEPTORS
ON
LEUCOCYTES ............... CSb
CSb.~
. . .C, . .St). . ,~ . .7. . . . . INS1ERTION INTO MEMORANES mC 5b_7 mC ~b_8 mCbl)_9 ~CylOtytlr s
F/gure 5
}
FLUID
PHASE
SCSb_/ S ~ 5b-6 SCSb-g {r162
inlctivtl
complls )
The terminal sequence of complement activation, C5-C9.
2. Surface-bound C3b which may trigger a number of effector functions from leucocytes following its interaction with C3b receptors on the cell membrane. These effects include the release of granular enzymes, generation of active oxygen species and the synthesis and release of the derivatives of arachidonic acid metabolism from neutrophils and monocytes. Similar effects may be mediated by surface-bound iC3b (which also forms on surfaces) and specific receptors for this fragment on leucocytes 28. 3. C5b-9. The most relevant of these products of complement activation with regard to bioincompatibility are the anaphylatoxins C3a and C 5 a 29. C3a and C5a are polypeptides with molecular weights of 9 0 0 0 and 11 000, respectively, which both have an arginine in the C-terminal position. The COOH terminal region is part of the site required for induction of smooth muscle cell contraction and histamine release by native anaphylatoxins. In plasma, C3a and C5a are rapidly transformed into stable desArg derivatives. C3adesArg and C5adesArg may trigger I L-1 production and release from monocytes at concentrations similar to those found in the plasma of patients undergoing haemodialysis with complement-activating membranes 3~ At nanomolar concentrations, C5adesArg elicits a variety of additional effects in neutrophils and monocytes, from which originate most complement-derived adverse reactions of extracorporeal circulation (Table 1). Both C5a and C5adesArg have chemotactic activity for neutrophils and monocytes and induce IL-1 release from monocytes 3~ Binding of C5a and CSadesArg to specific receptors on neutrophils induces aggregation of the cells 32 and their adherence to endothelial cells 33, the release of active oxygen radicals which may damage the endothelium to which activated neutrophils have bound 34, the release of lysosomal enzymes, and the neosynthesis and release of leukotrienes. The increased adhesiveness of granulocytes following incubation with C5a and C5adesArg is dependent on the enhancement, by the anaphylatoxins, of the membrane expression of the adhesion-
promoting glycoprotein Mo 1/Mac 1 on the cells 35. Mo 1/Mac 1 is the alpha chain of the iC3b complement receptor (CR3) which is non-covalently associated with a beta chain that is common to two other leucocyte membrane antigens involved in cell adhesion, LFA1 and p150,95 36. The M o l / M a c l molecule is involved in a number of granulocyte adhesion functions including adhesion to glass, fibronectin-coated surfaces, yeast cell walls, endothelial cells, leucocytes and target cells of antibody-dependent cellular cytotoxicity. C5a-mediated aggregation and pulmonary sequestration of activated granulocytes are believed to represent the main pathogenic mechanism responsible for haemodialysis-induced granulocytopenia and contribute to the hypoxaemia associated with the first use of complement-activating dialysis membranes 37,38.
THE CLINICAL SITUATION During haemodialysis with first-use cellulose derivatives, a rapid increase in the plasma concentration of C3a antigen is observed in the absence of increased plasma C4a concentrations and in the absence of demonstrable cleavage of Factor B, indicating that activation occurs via the alternative pathway. Higher concentrations of C3a antigen in the dialyser venous line as compared with the arterial line indicate that complement activation occurs at the bloodmembrane interface (Figure 6). Plasma concentrations of C3a antigen increase from the beginning of dialysis, reach a peak of approximately 1O- 1 5 fold the initial concentration at 1 5 min and then decrease to predialysis values after 4 h of dialysis. This decrease reflects the smaller rate of complement activation and C3a generation that follows the saturation of available C3b binding sites on the activating membrane; the catabolism of C3adesArg by the patient; the adsorption of the C3adesArg protein to the membrane 11. Generation of C5a/C5adesArg has also been documented during haemodialysis with first use cuprophane membranes, although at much lower concentrations than those of C3a/ C3adesArg. Low concentrations of C5a depend on the relatively low efficiency of C5 convertase sites relative to that of C3 cleavage, and to the rapid binding of newly generated C5a and C5adesArg molecules to high affinity receptors on neutrophils and monocytes. A direct relationship has been observed between the amount of C3a generated during haemodialysis with cuprophane membranes and the severity of leucopenia 39. Furthermore, the peak increase in C3a concentration coincides in time with the nadir of the
5.
;4. 33,
"""',
/
Table I Effects of binding of C5a and C5adesArg to CSa receptors on neutrophils Chemotaxis. Enhanced expression of C3b receptors (CRI). Enhanced expression of the adhesion-promoting glycoprotein Mo 1; reversible aggregation of neutrophils; enhanced adhesion of neutrophils to endothelial cells. Increased oxygen consumption and generation of free oxygen radicals, Stimulation of arachidonate metabolism with production of 5 HETE and LTB4. Specific granule secretion.
,, { MINUTES
r MINUTES
Figure 6 Time course of complement activation during haemodia/ysis (left panel) and cardiopu/monary bypass (CPBP) (right panel). C3adesArg antigen concentrations have been measured in samples taken from the dia/yser afferent (squares) and efferent (circles) line in patients on haemodiatysis with first use cuprophane (closed symbols) and polyacry/onitrile (open circles) membranes. For patients undergoing CPBP, blood was taken from a peripheral vein. Each point represents the mean of values obtained in five patients,
Biomaterials 1988, Vol 9 January
33
49
7"tie Biomaterials Silver Jubilee Compendium Complement system activation: M.D. Kazatchkine and M.P. Carreno
leucocytes count, and reused membranes and membranes which do not activate complement, do not induce leucopenia. An enhancement in the membrane expression of M o l / M a c 1 antigen on leucocytes has been observed during haemodialysis with cuprophane membranes4~ the peak increase in surface Mo 1/Mac 1 coincides with maximal leucopenia and with the peak increase in C3adesArg concentrations. Pulmonary sequestion of granulocyte microaggregates 12, perdialytic release of granulocytic elastase41 and neosynthesis and release of leukotrienes and PAF have been reproducibly observed during haemodialysis and contribute to the
hypoxaemia w h i c h occurs d u r i n g dialysis w i t h c o m p l e m e n t activating m e m b r a n e s as well as d u r i n g c a r d i o p u l m o n a r y bypass. C5a may also cause p u l m o n a r y and cardiac dysf u n c t i o n in vivo by directly acting on the lung vasculature, and other factors (e.g. the nature of the dialysate) may f u r t h e r enhance the severity of h y p o x a e m i c s y n d r o m e s d u r i n g b i o i n c o m p a t i b l e reactions w i t h artificial surfaces. The.pathogenic role of the I L- 1 that may be f o r m e d and released d u r i n g such reactions is not fully u n d e r s t o o d at the present time, A f t e r 5 h of haemodialysis on c u p r o p h a n e devices, m o n o cytes f r o m dialysed patients p r o d u c e biologically active intracellular IL- 113. In patients in w h o m an at least a t w o - f o l d increase in IL-1 activity is observed d u r i n g haemodialysis, a correlation is f o u n d b e t w e e n the perdialytic increase of I L - i p r o d u c t i o n and C 3 a d e s A r g generation, as measured after 1 5 rain of dialysis. C o m p l e m e n t activation may be one of the m e c h a n i s m s involved in the induction of IL-1 since no perdialytic increase in IL-1 p r o d u c t i o n is f o u n d in patients u n d e r g o i n g dialysis w i t h polyacrylonitrile m e m b r a n e s . H o w ever, a p p r o x i m a t e l y 5 0 % of patients dialysed w i t h polyacrylonitrile devices have high c o n c e n t r a t i o n s of intracellular IL-1 at the b e g i n n i n g of the dialysis session, indicating that m o n o c y t e s f r o m these patients are chronically stimulated to p r o d u c e I L- 1. One explanation for this observation could be t r a n s m e m b r a n e passage of rough LPS species t h r o u g h the h i g h - p e r m e a b i l i t y dialysis membrane13,
CONCLUSIONS The nature and severity of delayed or tong-term consequences of complement activation following the exposure of blood to an activating artificial surface are as yet poorly known. The degree of complement activation has been correlated with the severity of pulmonary, cardiac and renal dysfunction after open heart surgery 42. A relationship has been suggested, a l t h o u g h not proven in any ways, b e t w e e n c o m p l e m e n t activation, p u l m o n a r y endothelial damage, release of beta-2 m i c r o g l o b u l i n f r o m d a m a g e d cells and the b e t a - 2 - m i c r o g l o b u l i n - c o n t a i n i n g amyioid deposits w h i c h are observed in patients w h o have been dialysed w i t h cuprophane membranes for m o r e than 1 0 or 1 5 years. Designing b i o c o m p a t i b l e surfaces that do not activate the c o m p l e m e n t system should u n d o u b t e d l y contribute to decrease the m o r b i d i t y associated w i t h acute and chronic exposure of blood to synthetic polymers.
REFERENCES 1
2 3
34
Cazenave,J.P., Davies, J.A., Kazatchkine, M.D. and Van Aken, W.G., Blood-surface interactions, in BiotogicaI Principles Underlying Hemocompatibility withArtih'cial Materials, (Eds J.P. Cazenave,J.P. Davies, M.D. Kazatchkine and W.G. Van Aken), Elsevier, Amsterdam, 1986 Brown,F.J., Complement, in Fundamental Immunology (Ed.W.E. Paul), Raven Press, New York, 1985, pp 645-667 Kazatchkine, M.D. and Nydegger, U.E., Complement-Mediated Injury, in the Reticuloendothelial System. A Comprehensive Treatise:
Biomateria/s 1988, Vol 9 January
Hypersensitivity (Eds S.M. Phillips and M.R. Escobar), Plenum Press,
4 5
6 7 8
9
10
11
12
13
14 15 16 17
18 19
20 21 22
23 24
25 26
New York, 1986, pp 173-196 Tack, B.F., The alpha-Cys-/~ thiolester bond in human C3, C4 and alpha2 macroglobulin, Semin. Immunopathol. 1983, 6, 259-264 Law, S.K. and Levine, R.P., Interaction between the third complement protein and cell surface macromolecules, Proc. Natl. Acad. Sci. USA 1977, 74, 2701-2705 Cooper, N.R., The classical complement pathway: activation and regulation of the first complement component, Adv. lmmunoL 1985, 37, 151-2t6 Kazatchkine, M.D. and Nydegger, U.E., The human alternative complement pathway: biology and immunopathology of activation and regulation, Prog. Allergy. 1982, 30, 193-234 Fearon,D.T. and Austen, K.F., Activation of the alternative complement pathway due to resistance of zymosan-bound amplification convertase to endogeneous regulatory mechanisms, Proc. Natl. Acad. Sci. USA 1977, 74, 1683-1687 Kazatchkine,M.D., Fearon, D.T. and Austen, K.F., Human alternative complement pathway: membrane-associated sialic acid regulates the competition between B and beta- 1H for cell-bound C3b, J./mmunol. 1979, 122, 75-81 Kazatchkine,M.D., Fearon, D.T.,Siibert, J.E. and Austen, K.F.,Surfaceassociated heparin inhibits zymosan-induced activation of the human alternative complement pathway by augmenting the regulatory action of the control proteins on particle-bound C3b, J. Exp. Med. 1979, 150, 1202-1215 Carreno,M.P., Labarre, D., Jozefowicz, M. and Kazatchkine, M .D., The ability of Sephadex to activate human complement pathway is suppressed in specifically substituted functional Sephadexderivatives, Mo/. Immunol., in press Craddock,P.R., Fehr, J., Dalmasso, A,P., Brigham, K.L. and Jacob, H.S., Hemodialysis leukopenia: pulmonary vascular leukostasis resulting from complement activation by dialyzer cellophane membranes, J. C/in. Invest. 1977, 59, 879-888 Haeffner-Cavaillon,N., Fischer, E., Bacle, F., Carreno, M.P., Maillet, F., Cavaillon, J.M. and Kazatchkine, M.D., Complement activation and induction of interleukin-1 (IL-1) production during hemodialysis, Contr. NephroL, in press Chenoweth,D.E., Cheung, A.K. and Henderson, L.W., Anaphylatoxin formation during hemodialysis: effect of different dialyser membranes, Kidney Int. 1983, 24, 764-769 Chenoweth,D.E., Cheung, A.K, Ward, D.M. and Henderson, L.W., Anaphylatoxin during hemodialysis: comparison of new and re-used dialysers, Kidney Int. 1983, 24, 770-774 Wegmuller, E., Mondandon, A., Nydegger, U.E. and Descoeudres, C., Biocompatibility of different hemodialysis membranes: activation of complement and leukopenia, Int. J. Artif Organs. 1986, 9, 85-92 Maillet, F., Carreno, M.P., Labarre, D., Jozefowicz, M. and Kazatchkine, M.D., Carboxymethylation suppresses the ability of Sephadex to activate the human alternative pathway by facilitating inactivation of bound C3b with H and I, Complement, 1987, 4, 188 Carreno,M.P., Maillet, F., Jozefowicz, M. and Kazatchkine, M., The role of Sephadex antibodies in activation of the human alternative pathway by Sephadex, Complement, 1987, 4, 140 Schenkein, H.A. and Ruddy, S., The role of immunoglobulins in alternative complement pathway activation by zymosan. I. Human IgG with specificity for zymosan enhances alternative pathway activation by zymosan, J. Immunol. 1981, 126, 7-10 Hakim,R.M., Breillat, J., Lazarus, J.M. and Port, F.K., Complement activation and hypersensitivity reactions to dialysis membranes, N. Engl. J. Med. 1984, 311,878-882 Chenoweth,D.E.,Complement activation during hemodialysis: clinical observations, proposed mechanisms and theoretical implications, Artif Organs 1984, 8, 281-287 Kazatchkine,M.D., Fearon, D.T., Metcalfe, D.D., Rosenberg, R.D. and Austen, K.F., Structural determinants of the capacity of heparin to inhibit formation of the human amplification C3 convertase, J. C/in. Invest. 1981,67, 223-226 Mauzac, M., Maillet, F., Jozefonvicz, J. and Kazatchkine, M.D., Anticomplementary activity of dextran derivatives, Biomaterials 1985, 6, 61-63 Crepon, B., Maillet, F., Kazatchkine, M.D. and Jozefonvicz, J., Molecular weight dependence of the acquired anticomplementary and anticoagulant activities of specifically substituted dextran, Biomaterials, 1987, 8, 248-53 Bhakdi,S. and Tranum Jansen, J., Membrane damage by complement, Biochem. Biophys. Acta. 1983, 737, 343-354 Salama,A., Hugo, F., Heinrich, D., Kiefel, V., Mueller-Eckhardt, C. and Bhakdi, S., Deposition of C5b-9 complexes on erythrocytes and blood leukocytes during cardiopulmonary bypass, Complement, 1987, 4, 220
50
The Biomaterials Silver Jubilee Compendium Complement system activation: M.D. Kazatchkine and M.P. Carreno
27
28 29 30
31 32 33
34
Seeger, W., Suttorp, N, Hellwig, A. and Bhakdi, S., Non-cytolytic terminal complement complexes may serve as calcium gates to elicit leukotrienes by generation in human potymorphonuctear ieukocytes, J. Immunol. 1986, 137, 1286-1293 Fearon,D.T. and Wong, W.W., Complement iigand-receptor interactions that mediate biological responses, Ann. Rev. tmmunol, t 98 3, 1,243-271 Hugli, T.E., Biochemistry and biology of anaphylatoxins, Complement 1986, 3, 111-1:)7 Goodman, M.G., Chenoweth, D.E. and Weigle, W.O., Induction of interleukin-1 secretion and enhancement of humoral immunity by binding of human CSa to macrophages surface C5a receptors, J. Exp. Med. 1982, 156, 912-917 Haeffner-Cavaillon, N., Cavaitlon, J.M., Laude, M.D. and Kazatchkine, M.D., C3a(C3adesArg) induces production and release of interleukin-1 by cultured human monocytes, J. lmmunoL, t 987, 139, 794-799 Craddock, P.R., Hammerschmidt, D.E., White, J.G., Datmasso, A.P. and Jacobs, H.S., Complement (C5a)-induced granulocyte aggregation in vitro, J. Ciin. Invest. 1977, 60, 2 6 0 - 2 6 4 Tonnesen, M.G., Smedly, L.A, and Henson, P.M., Neutrophilendothelial cell interactions. Modulation of neutrophit adhesiveness induced by complement fragments C5a and C5adesArg and formylmethionyl-leucyl-phenylalanine m vitro, J. Clin. Invest. 1984, 74, 1581-1592 Sacks,T., Moldow, C.F., Craddock, P.R., Bouers, T.K. and Jacobs, H.S., Oxygen radicals mediate endothelial cell damage by complement
35
36'
37 38 39 40
41 42
stimulated granulocytes. An /n vitro model of immune vascular damage, J. Clin. Invest. 1978, 61, 1161-1167 Berger, M., O'Shea, J., Cross, A.S., Folks, T.M., Chused, T.M., Brown, E. and Frank, M.D., Human neutrophils increase expression of C3bi as well as C3b receptors upon activation, J. Ctin. Invest. 1984, 74, 1566-1571 Sanchez-Madrid, F,, Nagy, J.A., Robbins, E., Simon, P. and Springer, T.A., A human leukocyte differentiation antigen family with distinct alpha subunits and a common beta subunit. J. Exp. Med. 1983, 158, 1785-1803 Hakim, R.M. and Lowrie, E.G., Hemodialysis-associated neutropenia and hypoxemia: the effect of dialyzer membrane materials, Kidney Int. t 982, 32, 32-39 Cheung, A.K. and Henderson, L.W., Effects of complement activation by hemodialysis membranes, Am. J. Nephroi. 1986, 6, 81-91 Henderson, L.W. and Chenoweth, D.E., Cellulose membranes -- time for a change? Contr. NephroL 1985, 44, 112-126 Arnaout, M.A., Hakim, R.M., Todd III, R.F., Dana, N. and Colten, H.R., Increased expression of an adhesion-promoting surface gtycoprotein in the granulocytopenia of hemodialysis, N EngL J. Med. 1985, 312, 457-462 Horl, W.H., Rieget, W., Steinhauern, H.N,, Wanner, C., Thaiss, F., Bozkurt, F., Haag, M. and Schollmeyer, P., Granulocyte activation during hemodialysis, Clin. Nephrol. 1 986, 26, 530-534 Chenoweth, D.E., Anaphylatoxin formation in extracorporeal circuits, Complement 1986, 3, 162-165
Biomaterials 1988, Vot 9 January
35
The Biomaterials Silver Jubilee Compendium
51
Dynamic and equilibrium swelling behaviour of pH-sensitive hydrogels containing 2-hy oxyethyl methac lat Lisa Brannon-Peppas* and Nikolaos & Peppas
School of Chemical Engineering, Purdue University, West Lafayette, IN 47907, USA (Received 17 April 1990; accepted 16 May 1990)
The equilibrium and dynamic swelling behaviour of hydrogels containing methacrylic acid or various acrylamides was studied as a function of copolymer composition. In all cases, the comonomer was 2-hydroxyethyi methacrylate, methyl methacrylate or N-vinyl-2-pyrrolidone. It is shown that pH-sensitive behaviour with a wide range of swelling ratios could be obtained using a range of compositional changes. The dynamic swelling behaviour was a function of the acidity of the buffered solution. Keywords: Hydrogels, copolymers, swelling
Research over the past 25 years has shown that hydrogeis may be used as substitutes for natural materials. Structural, chemical and theoretical analyses of polymer structures have given scientists a clearer picture of the nature of hydrogels and ways to control their structure and behaviour ~-5. Hydrogels may be sensitive to their environment and their structure may change according to the conditions around them. They are potentially useful materials in biomedical applications. Hydrogels are particularly promising because of the similarities between their physical properties and those of living tissue1' 6.7 The most widely used hydrogel is water-swollen, cross-linked poly(2-hydroxyethyl methacrylate) p(HEMA). Its structure permits a water content similar to living tissue. p(HEMA) is inert to normal biological processes, shows resistance to degradation, is permeable to metabolites, is not absorbed by the body, withstands heat sterilization without damage, and can be prepared in a variety of shapes and forms. The swelling, mechanical, diffusional and biomedical characteristics of p(HEMA) gels have been studied extensively 1. The properties of these hydrogels are dependent upon their method of preparation, polymer volume fraction, degree of cross-linking, temperature and swelling agent. Many researchers have studied the transport of solutes through p(HEMA) hydrogels in recent years 8-13. The solutes used include insulin, urea, sodium chloride, glucose and others. The transport of these solutes depends on the Correspondence to Professor N.A. Peppas. *Present address: Lilly Research Laboratories, Eli Lilly and Co., Greenfield, IN 4 6 1 4 0 , USA.
structure of the hydrogel, the water content of the hydrogel, the temperature of the system TM and the nature of the solute 15'~6 Some work has been done on the relation between the permselectivity of the hydrogel and the pH of the system for membranes containing H EMA 17 but results of changes in swelling with changes in pH were not reported. Other biocompatible hydrogels include polyacrylamides lB. 19. Copolymers of H EMA with acrylamide and methacrylamide have been evaluated for swelling and mechanical behaviour by Du~ek and Jana6ek 2~ Numerous acrylamides have been studied for biocompatibility 21'22, drug delivery systems 23'24 as well as fundamental understanding of their behaviour 25-28. The extreme swelling behaviour of polyacrylamide gels has been attributed to the hydrolysis of some of the acrylamide groups to acrylic acid. Ilavsky 29 studied copolymers of acrylamide and sodium methacrylate and noticed discontinuous swelling behaviour in water-acetone mixtures with an increase in the discontinuity as the amount of sodium methacrylate was increased. There has been increased interest in recent years in evaluating the behaviour and bioapplicability of acrylic acidand methacrylic acid-containing polymers due to their ionic nature. Some basic characteristics, methods of preparation and commercial applications may be found in a report by Greenwald and Luskin 3~ The suitability of acrylic acidcontaining polymers for intraocular systems 31 and bioerodible drug delivery systems 32 has also been investigated. The presence of ionic groups in these polymers can also lead to polymer complexation 33-36, giving another option to the polymer's structure.
91990 Bunervvorth-Heinemann Ltd. 0 1 4 2 - 9 6 1 2 / 9 0 / 0 9 0 6 3 5 - 1 0
Biomateriais 1990, Vol 11 November
635
The Biomaterials Silver Jubilee Compendium
52 pH-sensitive hydrogels: L Brannon-Peppas and tV.A. Peppas
The degree of ionization and the nature of the medium surrounding these polymers is extremely important. Polymers containing acrylic acid and methacrylic acid groups show a sensitivity to the pH of their surroundings. For example, small amounts of maleic anhydride as a comonomer have been shown to increase dramatically the swelling of p(HEMA-co-MAA) polymers 37. In this work, we present new data of the dynamic and equilibrium swelling of a series of pH-sensitive hydrogels, and we investigate the nature of their swelling behaviour in relation to their hydrophilic/hydrophobic nature and composition.
EXPERIMENTAL METHODS Copolymer preparation The comonomers used were 2-hydroxyethyl methacrylate (HEMA), methyl methacrylate (MMA), methacrylic acid (MAA), maleic anhydride (MAH), n-vinyl-2-pyrrolidone (NVP, methacrylamide (MAc), N , N - d i m e t h a c r y l a m i d e (DMAc), N-isopropyl acrylamide (IPAc) and diacetone acrylamide (DAAc). Copolymers were prepared in 6 ml cylindrical vials at varying compositions. Each formulation contained 0.5 wt% benzoyl peroxide as initiator, 0.5 mol% ethylene glycol dimethacrylate (EGDMA) as a cross-linking agent, and two comonomers. Every comonomer feed solution contained HEMA of 0, 1 O, 20, 30, 40, 50, 60, 70, 80, 90 or 100 mol%. The remainder of the mixture was NVP, MMA, MAA, MAH, MAc, DMAc, IPAc or DAAc. In the cases of preparation of copolymers with MAA, 25 vol% distilled water was added to facilitate good mixing. The monomers, initiator and cross-linking agent were mixed at room temperature and then immersed in a water bath; the temperature was increased gradually according to Table 1
the following programme: 1 h at 40~ 2 h at 50~ 3 h at 60~ 11 hat65~ 11 hat 70~ 11 hat75~ and 10 hat 80~ After the polymerization was complete, the vials were cooled to room temperature and the polymers (in the form of cylinders) were removed. Those formulations made without water were then annealed at 80~ for 2 d. The polymer cylinders were cut into discs of 0.5-1 mm thickness, 12.7 mm diameter.
Polymer characterization The polymer discs were dried in a vacuum oven for 4 d or until constant weight. The weights were measured both in air and heptane. After the polymers reached equilibrium in a given swelling agent, their dimensions and their weights in air and heptane were remeasured. The discs were then redried in air and in a vacuum oven and their weights in air and heptane were measured. A select group of the polymers were then used to measure dynamic swelling. The weight uptake of the swelling agent was measured as a function of time by measuring the weight of the polymer at numerous times while it was swelling to equilibrium. The swelling solutions used included distilled-deionized water, methanol, acetone, 0.1 5 M HCI, artificial gastric fluid, artificial intestinal fluid and buffers of pH 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0 and 10.0 as shown in Table 1.
DYNAMIC AND EQUILIBRIUM SWELLING OF HYDROGELS To determine the volume of each polymer sample tested, their weights in air and heptane were measured and their
Swelling solutions used ....
Swelling solution/pH value
Components
Molar concentration
Ionic strength
Dilute HCI/pH 1.1 Simulated gastric fluid/pH 1.2
Hydrogen chloride Sodium chloride Hydrogen chloride Tartaric acid Potassium biphthalate Tartaric acid Potassium biphthalate Potassium biphthalate Sodium bicytrate Hydrogen chloride Disodium phosphate Potassium phosphate Disodium phosphate Potassium phosphate Disodium phosphate Potassium phosphate Sodium chloride Potassium sulphate Sodium phosphate Ammonium chloride Urea Monobasic potassium phosphate Sodium hydroxide Disodium phosphate Potassium phosphate Sodium carbonate Sodium borate Sodium carbonate monohydrate Sodium bicarbonate Glycine Sodium chloride Sodium hydroxide
0.1500 0.0342 0.1921 0.0546 0.0088 0.0233 0.0318 0.0490 0.6640 0.3360 0.0014 0.0720 0.0120 0.0610 0.0437 0.0279 0.1250 0.0375 0.0250 0.0750 0.3200 0.0500 0.0380 0.0655 0.0051 0.0113 0.0427 0.0524 0.0417 0.0625 0.0625 0.0375
O.1500 0.2263
Buffer/pH 2.0 Buffer/pH 3.0 Buffer/pH 4.0 Buffer/pH 4.4 Buffer/pH 5.0 Buffer/pH 6.0 Buffer/pH 7.0 Simulated urine/pH 7.2
Simulated intestinal fluid/pH 7.5 Buffer/pH 8.0 Buffer/pH 9.0 Buffer/pH 10.0 Buffer/pH 10.0
,,
636
Biomaterials 1990, Vol 11 November
0.1585 0.1378 0.1225 0.1332 0.1835 0.1825 0.1790 0.2938
0.0880 0.1765 O.1350 0.2353 0.1625
The Biomaterials Silver Jubilee Compendium
53 pH-sensitive hydrogels: L. 8rannon-Peppas and N.A. Peppas
difference was related to the polymer volume as shown in Equation 1.
v=
W A - WH
(1)
PH
8[
4
In this equation WA is the weight of the polymer in air, W H is the weight of the polymer in heptane and PH is the density of heptane. The swelling ratio, Q was then calculated from the swollen polymer volume, Vg, and the polymer volume after redrying, Vp. Q=Vg=
1
Vp
(2)
U2, s
The polymer volume fraction, u2, s, given by Equation 2, and the specific volume of the polymer, 5, calculated by Equation 3, were two parameters determined from the polymer swelling studies which were used in the swelling analysis. u=
4
The theoretical number average molecular weight between cross-links for a polymer network, Mc, th, was calculated from knowledge of the mole fraction of each monomer, x~, the molecular weight of the repeating unit of each monomer, Mr, i, and the molar cross-linking ratio, X. For a copolymer of two monomers, the value of Mo, th was calculated as shown in Equation 4. XlMr, 1 + x2Mr, 2
2X
M
(5)
n
O
,Q r
<>
M,=k't n
(6)
where M< is the mass uptake of swelling agent at equilibrium, k and k' are constants and n is an exponent describing the Fickian or anomalous swelling mechanism.
EQUILIBRIUM SWELLING RESULTS Effect of composition on swelling We will first discuss swelling studies of the prepared samples in methanol and acetone. For simplicity, the data have been grouped into non-acrylamide-containing (Figures 1 and 3) and acrylamide-containing (Figures 2 and 4) copolymers. All data shown here have an experimental error lower than 2.6% of the actual value. It can be seen that copolymers of p(HEMA-co-MAA), p(HEMA-co-NVP), p(HEMA-co-DAAc), p(HEMA-co-IPAc) and p(HEMA-coDMAc) have, in general, higher degrees of swelling in methanol than the pure homopolymer p(HEMA). Of these, p(HEMA-co-MAA) and p(HEMA-co-NVP)with more than
[3 O
o
0 A
A I
0
-.,
!
20 MolYo
-
,L
40 HEMA
,,,L
60 in
80
100
Copolymer
Figure 1 Methanol equilibrium swelling ratio of HEMA-containing copolymers at 25~ as a function of the mol% HEMA in the copolymer. The symbols represent the copolymer containing varying second comonomers as follows: NVP ( 9 MAA ([3), MMA (0), and MAH ( A ).
15
6
~
A
13 11
~P4
O0
9
[]
[]
7 5
,F-I ,--4
-
-
O
<>
O
<>
A 3
-
&
r
o
or
O
<>
~
tO:
- kt n
O
0
m
(4)
The equilibrium degree of swelling, Q, was calculated for all copolymers in swelling agents such as methanol, acetone and a number of buffered aqueous solutions of different pH values In addition, the dynamic swelling was measured for selected samples. The mass uptake of swelling solution, Mt, as a function of time, t, was analysed according Me
O
(3)
Wp
Mc,h =
O R
1
o s 0
, 20
Mol~
9
~
I in
~
....L
60
40 HEMA
[]
80
100
Copolymer
Figure2 Methanol equilibrium swelling ratio of HEMA-containing copolymers at 25~ as a function of the tool% HEMA in the copolymer. The symbols represent the copolymer containing varying second comonomers as follows: MAc ( 9 IPAc [n), DAAc (0), and DMAc (z~).
30 mol% HEMA show the highest degrees of swelling. The lowest degrees of swelling in methanol were observed with copolymers of p(HEMA-co-MAc). In addition, p(HEMA-coMAH), p(HEMA-co-MMA) and p(HEMA-co-MAc) exhibited degrees of swelling lower than those of p(HEMA), p(HEMAco-MAA) exhibited a maximum in the degree of swelling at about 60 mol% H EMA. This behaviour was indicative of the compatibility of MAA and HEMA in methanol. In general, a high H EMA content leads to lower methanol equilibrium swelling, probably due to hydrogen bonding. In acetone, p(H EMA-co-DAAc), p(H EMA-co-IPAc) and p(HEMA-co-DMAc) had the highest degrees of swelling,
Biomateriais 1990, Vol 11 November
637
The Biomaterials Silver Jubilee Compendium
54 pH-sensitive hydrogels: L. Brannon-Peppas and NJ~. Peppas
p(HEMA-co-MMA) with 50mo1% HEMA. This value increased to Q = 1.84 for pure p(HEMA). The swelling studies of p(HEMA-co-NVP) showed a decrease of the equilibrium swelling ratio as the HEMA content of the copolymer increased as shown in Figure 6. This was expected since NVP is a more hydrophilic moiety than HEMA. The results are in agreement with results reported by other researchers in our group 3e' 40. It must be noted that all of the studies presented in Figures 5 and 6 are for acidic and alkaline solutions, and that the general dependence on HEMA content is retained in both the acidic and the alkaline region. Copolymers of p(NVP-co-M MA) were previously tested for their swelling behaviour in water 41' 42. They showed a consistent increase in swelling as the amount of NVP in the copolymer increased. The conclusion of all of these studies is that significant
0
20
MolY~
40
60 in
HEMA
80
100
(Y
6
Copolyrner
Figure 3 Acetone equilibrium swelling ratio of HEMA-containing copolymers at 25~ as a function of the mol% HEMA in the copolymer. The symbols represent the copolymer containing varying second comonomers as follows: NVP (0), MAA (D), MMA (0), and M A l l (A).
Or
10
D
0
9 .P
8<
.I-I
0
1@
o
.0=l
1
i--(
0
0
.PI
-i-I v=4 .~=1
o'
0 A
Q 20 MolYo
n
0
L
0 [3
0 [3
l
20
40
MolY~
HEMA
0
I 60 in
I 80
100
Copolyrner
0
o
o
0
,~
o
~
40 HEMA
60 in
0
e .,,
d|
~
80
100
followed by p(HEMA-co-MMA) and p(HEMA-co-MAA). All these swelled more than pure p(HEMA). Swelling ratios lower than those of p(HEMA) were observed for p(HEMAco-MAc), p(HEMA-co-MAH) and p(HEMA-co-NVP). For copolymers containing methyl methacrylate (see Figure 5), the equilibrium swelling ratio increased with increasing H EMA content up to a maximum value of Q = 2.1 at 90 mol% H EMA. A similar observation had been made by Franson and Peppas 38, but studying only the range from 50 to 90 mol% HEMA. In addition, Davidson and Peppas 39 had determined an equilibrium swelling ratio of 1.39 for
Biomaterials 1990, Vol 11 November
Or
6 41,
Copolyrner
Figure 4 Acetone equilibrium swelling ratio of HEMA-containing copolymers at 25~ as a function of the mol% HEMA in the copolymer. The symbols represent the copolymer containing varying second comonomers as follows: MAc ( 9 IPAc ([3), DAAc (<>), and DMAc (ix).
638
0 D
Figure 5 Equilibrium swelling ratio of p(HEMA-co-MMA) in pH 6.0 ( 9 and 10.0 (D) at 25~ as a function of the mol% HEMA in the copolymer.
0 o
O 0
i,.=1
<>
@
a
GO
0
I.
o
o
o
i=,,I
o'
L
0
20 Mol~
I
~ !
4,0 HEMA
80 in
I
80
100
Copol~er
Figure 6 Equilibrium swelling ratio of p(HEMA-co-NVP) in pH 4.4 ( 9 7.2 (D) and 10.0 (<>) at 25~ as a function of the mol% HEMA in the copolymer.
The Biomaterials Silver Jubilee Compendium
55
pH-sensitive hydrogets.- L. 8rannon-Peppas and N.A. Peppas
changes in the swelling ratio of hydrogels are observed with changes in the amount of methyl methacrylate (MMA) or N-vinyl-2-pyrrolidone (NVP). This result indicates that the hydrophilicity of HEMA is relatively moderate. By adding the very hydrophilic NVP, one may achieve a highly swellable copolymer. By adding the very hydrophobic MMA, one obtains a very poorly swellable network. In both cases, the designer of novel devices utilizing the diffusive or related characteristics of the hydrogels recognizes that she can achieve a significant change in the degree of swelling with a small change in the hydrophilicity of the system. Previous work by several investigators has shown that acrylamide-containing gels are quite promising for the development of pH-sensitive systems. However, the literature presents only scattered data on the effect of the compo-
(Y .I,-4
10 9
8 o []
7 -,"4 ~4
D o
o o
6
o
[3 o
5 4
,n
Effect of pH on swelling
3 2
N
1
1
20
40
1
0 Mol~
HEMA
in
I
!
60
80
100
Copolyrner
Figure 7 Equilibrium swelling ratio of p(HEMA-co-MAA) in pH 8.0 ( 9 and 9.0 ([3) at 25~ as a function of the mol% HEMA in the copolymer.
(Y
12
0
11
oP4
sitional structure on the equilibrium degree of swelling. Our investigations showed a distinct change in the equilibrium swelling ratio with composition for copolymers of HEMA with both MAA ( F i g u r e 7) and DMAc ( F i g u r e 8). It is interesting that this dependence was observed in both acidic and alkaline solutions. For these copolymers, the swelling ratio increased dramatically with decreasing HEMA content up to 9.23 for a copolymer with 90 mol% MAA swollen at pH 10, and up to 1 1.37 for a copolymer with 90 mol% DMAc swollen at pH 6. These results may be used to evaluate the general contributions of HEMA, MAn,, NVP and DMAc to the copolymer swelling in water. By comparing F i g u r e s 6, 7 and 8, it can be seen that the water swelling ratio of copolymers containing HEMA as the first comonomer and DMAc or MAA or NVP as the second comonomer increases in the following way: DMAc > NVP > MAA. In contrast, addition of M MAto the copolymer decreases its equilibrium swelling. Considering that PMMA is the most hydrophobic of the homopolymers studied, this result was expected. Since copolymers of p(HEMA-co-MAH), p(H EMA-coIPAc) and p(HEMA-co-MAc) demonstrated no significant change in the swelling with composition, it can be inferred that these monomers contribute approximately equally to the degree of swelling of the copolymers. Copolymers of p(HEMA-co-DAAc) showed a decrease in swelling with an increase in the amount of DAAc used (see F i g u r e 9). However, this decrease was not as significant as that found for p(HEMA-co-MMA).
<>
Even more interesting is the change in polymer swelling with pH. For the polyacrylamides investigated here, there was no difference in the swelling with changes in pH for HEMA copotymers containing also DAAc, DMAc or MAc. This occurred because N-substituted polyacrylamides are resistant to hydrolysis. Equilibrium swelling of partially hydrolysed acrylamide gels increases with an increase in the pH of the swelling solution. This hydrolysis converts some of the acrylamide pendant chains into acrylic acid, thus
O'
3.0
6
10
2.5
l
.e4 .Pi r--4
5 .I-4
h
2.0
0
4
.r-I
h ,0
3 2 1
1.5 C
,r-4 p-q
0
1
1
I
1
20
40
60
80
Mol~
HE~L~
in
100
Copolyr~er
Figure 8 Equilibrium swelling ratio of p(HEMA-co-DMAc) in pH 1.6 ( 9 4.4 (D), 6.0 (0) and 7.5 ( A ) at 25~ as a function of the mol% HEMA in the copolymer.
ID
1.0
0
D O
L
{
20
40
Mol%
HEMA
0
O I......
60
in
I
80
100
Copolymer
Figure 9 Equilibrium swelling ratio of p(HEMA-co-DAAc) in pH 4.4 ((3) and 7.0 (E]) at 25~ as a function of the mol o~ HEMA in the copolymer.
Biomaterials 1990, Vol 11 November
639
The Biomaterials Silver Jubilee Compendium
56 pH-sensitive hydrogets: L 8rannon-Peppas and N.~ Peppas
increasing the polymer's response to changes in its environmental pH. These results support the conclusion that it is indeed the acid groups which contribute to a dependence of swelling on pH and not the acrylamide groups themselves. In our study, p(HEMA-co-MAA) copolymer samples containing 10 to 90 roof% MAA consistently showed an increase in swelling with an increase in pH as seen in F i g u r e s 10 and 1 1 . The equilibrium swelling ratio at high pH values increased as the amount of MAA in the copolymer increased, indicating an addition to the ionic contribution to the swelling. Specifically, the equilibrium swelling ratio increased from 5.7 for copolymers containing 10 mol% MAP, to 9.2 for copolymers containing 90 tool% MAA at pH 10. The
(Y
10
greatest increase in the swelling ratio occurred around pH 7. In addition to p(HEMA-co-MAA), the copolymers p(HEMA-co-tPAc) and p(HEMA-co-MAH) and pure p(HEMA) all showed an increase in swelling with an increase in the pH of the swelling solution as shown in Figures 12-14. The pH dependence of the swelling ratio of p(HEMA-co-MAH) samples is due to the ability of MAH to form a stable, but ionized ring structure as shown by Gaylord 43. Both p(HEMA) and p(HEMA-co-IPAc) may form partially ionized structures as well; this would explain the slight dependence of their swelling ratio on the pH of their environment. Copolymers of p(HEMA-co-NVP) showed no significant swelling dependence on pH.
(Y
6
3.0
.P4
~a
2.5 7 ~
03
.I-I w-4
6
-
5
-
4
-
3
-
2
-
1
9 0
K!
0
03
0
5~ o
-
~
6
............ !._
2 pH
0
4 of
0
~
l
........
6
Swelling
6
9
*e4
8
~0
,
V <>
of
I_. 6
S~relling
~ I 8
10
Medium
Figure 12 Equilibrium swelling ratio of p(HEMA-co-IPAc) containing 20 ( 9 and 40 (0) tool% IPAc at 25~ as a function of the pH of the swelling medium.
.... m
O'
-
D 0
0 E]
! ..... J.... 2 4
pH
Medium
Ill
III
Ill
m
=
III
mLII
IIIIIII
Ull
.
~a
V
( {)
03
03
2
4 w-I
3
a~
V
2
1
I ...............
2
0
pH
of
I
o
4
6
Sw'elling
J
0
,,
I
8
10
Medium
Figure 11 Equilibrium swelling ratio of p(HEMA-co-MAA) containing 50 (0}, 60 (O), 70 (0), 80 (A J and 90 (V) tool% MAA at 25~ as a function of the pH of the swelling medium.
640
1.5
10
Z3
=-t w-I
O0 []
1.0
I
8
~ F-4 r-4
,.o
8
2.0
.r4 I-4 q-4
Figure 10 Equilibrium sweltiJ~g ratio of p(HEMA-co-MAA) containing 10 (0), 20 (D), 30 (~) and 40 ( A ) tool% MAA at 25 ~C as a function of the pH of the swelling medium.
10
0
.P4 e-4 ,-4
.H
0
(Y
[3
w
-
Biomateria/s 1990, Vo/ 11 November
0
2
pH
4
of
S'w'elltng
6
8
10
Medtuz~
Figure 13 Equilibrium swelling ratio of p(HEMA-co-MAH] containing 10 (0) and 20 ([3) tool% MAH at 25~ as a function of the pH of the swelling medium.
The Biomaterials Silver Jubilee Compendium
57 pH-sensitive hydrogels: L. Brannon-Peppas and N.A. Peppas 2.5
3.0
8 m
or
2.5
0" 0
2,0
:900
0
1.9
-
1.6
-
Zl
A
/k
0
0
0
A
~..4 i--I
~)
1.5
O~
l.O
9 t
!
+t
l
I
I
2
:3
4
5
6
7
1
pH
of
[. 8
Swelling
1 9
1 10
I.,
1 0
11
12
z,
0
.
50
Medium
L
L
100
Time
150
....
200
(rnin)
Figure 14 Equih'briumswelling ratio of poly(2-hydroxyethyl methacrylate) at 25~ as a function of the pH of the swelling medium.
Figure 15 Dynamic swelling of p(HEMA) in buffered solutions of pH 6.0 ( 9 7.0 (H), 8.0 (<>) and 9.0 (/k ) at 25~
DYNAMIC SWELLING RESULTS
indicates Fickian diffusion of the w a t e r into the samples, whereas a value of n around 1.0 indicates Case-II (relaxationcontrolled) transport. Values of the exponent n lying between 0.5 and 1.0 indicate a non-Fickian transport, controlled both by diffusion and relaxation. The swelling mechanism of p ( H E M A ) in water showed little dependence on pH, except at pH 9. At this high pH value the swelling mechanism w a s found to be non-Fickian, w i t h n = 0 . 7 2 2 , whereas for other pH solutions, the swelling mechanism was approximately Fickian w i t h n = 0.5. The early swelling ratios of all copolymers studied varied w i t h pH, as seen f r o m the studies of copolymers of p ( H E M A - c o - N V P ) , p ( H E M A - c o - D A A c ) and p ( H E M A - c o M A H ) (see F i g u r e s 1 6 - 1 8 ) . However, the mechanism of dynamic swelling remained p r e d o m i n a n t l y Fickian (as
Dynamic swelling studies were performed on selected samples to better understand the time dependence of their swelling behaviour w i t h pH. These polymers w e r e p ( H E M A ) ; p ( H E M A - c o - N V P ) containing 6 0 mol% or 7 0 mol% NVP; p ( H E M A - c o - D A A c ) containing 10 mol% DAAc; p ( H E M A c o - M A H ) containing 1 0 m o 1 % M A H ; and p ( H E M A - c o M A A ) containing 1 0 or 2 0 mol% M A A . The analysis of the first 6 0 % of the w a t e r uptake was accomplished using E q u a t i o n 5. The constant k and exponent n are given in Table 2 as calculated from the results s h o w n in F i g u r e s 1 5 - 2 0 . All samples analysed here w e r e thin discs w i t h aspect ratios usually greater than 1 0 : 1 . Thus, the results of onedimensional diffusional equations can be used, as discussed by Ritger and Peppas 44. Therefore, a value of n around 0 . 5 Table 2 .
Analysis of dynamic swelling in buffered aqueous solutions using Equation 5 (Mt/M.,
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
=
kt n)
.
Copolymer
Mol% HEMA
pH of swelling solution
k(s-')
p(HEMA)
100
2 3 4 6 7 8 9
0.0140 0.0126 0.0158 0.0139 0.01 26 0.01 78 0.0034
_+0.0020 + 0.0015 +0.0018 + 0.0013 + 0.0022 • 0.0018 • 0.0005
0.476 -+ 0.01 9 0.474 ~- 0.016 0.516+ 0.016 0.492 + 0.016 0.454 + 0.028 0.365 + 0.014 0.695 + 0.022
p(H EMA-co-NVP)
60
2 3 4
0.0132 + 0.0060 0.0063 + 0.0029 0.0091 + 0.0029
0.615 + O.058 0.712 + 0.058 0.61 9 § 0.040
p(HEMA+co-DAAc)
90
7 8 9
0.0332 + 0.0086 0.0040 • 0.0006 0.0048 • 0.0012
0.369 + 0.033 0.542 + 0.020 0.596 ~ 0.032
p(HEMA-co-MAH)
90
7 8
0.0034 • 0.0006 0.0037 + 0.0009
0.551 + 0.022 0.642 + 0.032
p(HEMA-co-MAA)
90
6 7 8
0.0038 _+_0.0007 0.0028 + 0.0004 0.0024 + 0.0003
0.491 + 0.025 0.51 3 + 0.015 0.634 + 0.01 5
p(HEMA-co-MAA)
80
6 7 8
0.0032 § 0.0007 0.0094 -~ 0.0012 0.0112 + 0.0003
0.514 + 0.029 0.556 + 0.01 7 0.579 § 0.004
.
.
.
n
.
.
.
.
.
.
.
.
Biomaterials 1990, Vol 11 November
641
The Biomaterials Silver Jubilee Compendium
58 pH-sensitive hydrogels: L. Brannon-Peppas and N.A, Peppas 3.5
<>
<> 3 . 0
p
d
2.5
.P4
-
d []
~-4
;2.0
1.0
o 0
-
1.5
2.0
~4
1.5
D
=.
[
_
[
5
0
1.0r
-,
10
Time
I
0
15
o
D
0
0 0
oo [3 Oo
-
n
o
l
L
50
100
(rain)
Time
Figure 16 Dynamicswelling of p(HEMA-co-NVP) with 70 mol% NVP in buffered solutions of pH 2.0 (O), 3.0 (D) and 4.0 (r at 25~
o
o
l 150
200
(rain)
Figure 18 Dynamicswelling of p (HEMA-co-MAH) with 10 mol% MAH in buffered solutions of pH 7.0 (0) and 8.0 (0) at 25~
2.5
O'
6
2.0 o
0
0
1.5
0
0 oo~ [3
0
d
<>
Ul
D ~
o
oQ
4
6
0
d
F-4
(:7
0
D o
0
3
0
0
F,4
o
.
O
r
0
0
^OO r 1.0
i o
t 50 Time
._. L
L
100
150
o
-
200
oo
o
~
~
~
~
~
0 ,,
0
25
50
75
100
,
125
(rnin)
Figure 17 Dynamicswelling of p(HEMA-co-DAAc) with 10 mol% DAAc in buffered solutions of pH 7.0 ( 9 8.0 (s and 9.0 (r at 25~
indicated by the values of n of Table 2), excluding perhaps some dynamic swelling studies at pH values of 8, which showed a slight deviation from Fickianism with values of n around 0.6. For copolymers of p(HEMA-co-MAA), the swelling mechanism became non-Fickian with increasing pH values (see Figure 19). Thus, not only did these copolymers swell more when placed in highly alkaline solutions, but the swelling phenomenon occurred under non-Fickian conditions, indicating an important chain relaxational phenomenon. To investigate the influence of abrupt changes of the pH on the swelling ratio, selected polymer samples were swollen in a buffered solution of pH 10, then placed in a buffered solution of pH 2, and finally returned to a buffered solution of pH 10. The observed experimental changes of the swelling ratio are plotted in Figure 2 0 .
642
8
Biomaterials1990, Vol 11 November
Figure 19 Dynamicswelling of p(HEMA-co-MAA) with 10 mol% MAA in buffered solutions of pH 6.0 (0), 7.0 ([]) and 8.0 (r at 25~
It can be seen that after an initial re-equilibration there was little swelling change caused by the pH change in the p(HEMA-co-NVP) copolymer containing 70mo1% NVP. However, p(HEMA), p(HEMA-co-MAH), and p(HEMA-coIPAc) samples showed significant decrease in their swelling ratios as they were transferred from a pH 10 solution to a solution with a pH 2. It must be noted that the changes in swelling ratio seem to be slightly dependent on the timehistory of the material. Therefore, a hydrogel that swelled once to pH 1 0 had a slightly different swelling ratio than a similar hydrogel that had been exposed to the cycle pH 1 0 - * 2--, 10. The most promising results of these swelling studies were those of the p(HEMA-co-MAA) copolymer. They demonstrated the most interesting swelling behaviour by
59
The Biomaterials Silver Jubilee Compendium pH-sensitive hydrogels: L. Brannon-Peppas and IV.A. Peppas 2
O'
6
5.0
•
4.0
wa
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3
.
4 5 6
~ 3.0 7
~ 2..0
8
9
~ 1.0_ 9 0
*
L
tO0
2,00
~
Time
i
L
300
_
4-00
I
500
600
(rain)
Figure 20 Dynamic swelling of HEMA-containing polymers with 10 tool% MAA (0), 20 mol% MAA ([3), 70 tool% NVP (0), 40mo1% IPAc (~), 20 mol% MAH (0) and 100 mol% HEMA ( i l l at 25~ All samples were placed in pH 10.0 buffer at t = 0 min, pH 2.0 buffer at t = 335 min and pH 10.0 buffer at t = 440 min.
expelling over 70% of the swelling agent initially present in the hydrogel within the first 1 5 min after the abrupt change of the pH from 10 to 2. The only drawback of the cycling process of these polymers was that this treatment required mechanical strength that some of these materials did not possess. Figure 20 shows that, after 600 min, only three samples, p(HEMA), p(HEMA-co-IPAc) and p(HEMA-co-MAA) containing 10 mol% MAA, had not broken. Each of these materials reached a final equilibrium swelling approximately equal to its first swelling; this result indicates reproducibility of their swelling behaviour. A conclusion of these swelling studies is that copolymers containing HEMA as the first comonomer and MAc, MMA, DAA, DMAc or IPAc as the second comonomer showed the least tearing upon swelling in water. However, methanol swelling of IPAc- and MMA-containing copolymers led to cracking of the samples.
10
11
12
13 14
15 16 17 18 19 20 2t
CONCLUSIONS In conclusion, it has been shown that pH-sensitive hydrogeis can be prepared by incorporation of moieties with carboxylic groups, or moieties with acrylamide, or simply by drastic change of the hydrophilicity of one moiety. These hydrogels show significant dependence on pH and ionic strength.
22 23
24
ACKNOWLEDGEMENTS This work was supported by a grant from the National Science Foundation.
25 26 27
REFERENCES Peppas, N.A., Hydrogels in Medicine and Pharmacy, CRC Press, Boca Raton, FL, USA, 1987
28 29
Du~ek,K. and Prins, W., Structure and elasticity of non-crystalline polymer networks, Adv. Polymer Sci. 1969, 6, 1- 102 Russo,P.S., A perspective on reversible gels and related systems, in Reversible Polymeric Gets and Related Systems, ACS Symposium Series, Vol, 350 (Ed. P.S. Russo), American Chemical Society, Washington DC, 1987, pp. 1-21 Bikales, N,M., Present and future directions of basic research on polymeric materials, Polym. J. 1987, 19, 11-20 Williams, C., Brochard, F. and Frisch, H.L., Polymer collapse, Annu. Rev. Phys. Chem. 1981, 32., 433-451 Peppas,N,A., Structure, testing, and applications of biomaterials, in Biomaterials: lntenfaciat Phenomena and Applications, ACS Symposium Series 199 (Eds S.L. Cooper and N.A. Peppas), American Chemical Society, Washington DC, 1982, pp. 465-473 Hoffman, A.S., Synthetic polymer biomaterials in medicine-a review, in The Past, Present and Future of Artificial Organs (Eds E. Piskin and T.M.S. Chang), Meteksan Publishing. Ankara, Turkey, 1983, pp. 32-51 lshihara, K., Kobayashi, M, and Shinohara, I., Insulin permeability through amphiphitic polymer membranes having 2-hydroxyethyl methacrylate moiety, Potym. J, 1984, 16, 647-651 Peppas,N.A. and Moynihan, H,J., Transport phenomena in polymers. PHEMA memranes: preparation and properties, Polymer News 1983, 9, 139-141 Robert, C.R., Buri, P.A. and Peppas, N.A., Effect of degree of crosstinking on water transport in polymer microparticles, J. Appt. Polym. Sci. 1985, 30, 301-306 Ratner,B.D. and Miller, I.F., Transport through crosslinked poly(2hydroxyethyl methacrylate) hydrogel membranes, J. Biomed. Mater, Res. 1 973, 7, 353-367 Wisniewski, S, and Kim, S.W,, Permeation of water-soluble solutes through poly(2-hydroxyethyl methacryfate)and poly(2-hydroxyethyl methacrylate) crosslinked with ethylene glycol dimethacrylate, J. Membr. Sci. 1980, 6, 299-308 Robert,C., Peppas, N.A. and Buri, P., Swelling of gets of PHEMA and release of phenylephrin HCI from this polymer, Proceed, Intern. Syrup. Control Rel. Bioact. Mater. 1 985, 12, 130-1 31 Yoon,S.C. and Jhon, M.S., Temperature effect on the permeation through poly(2-hydroxyethyl methacrytate) membrane, J. AppL Polym. Sci. 1982, 27, 4 6 6 1 - 4 6 6 8 C h e n , R.Y.S., Electrolyte transport through crosslinked poly(2hydroxyethyl methacrytate), Potym. Prepr. 1979, 20(1), 10051008 Sung, Y.K., Gregonis, D.E., Russell, G.A. and Andrade, J.A., Effect of water and tacticity of the glass transition temperature of poly(2hydroxyethyl methacrylate), Polymer, 1978, 19, 1 362-1363 Kudela,V., Va~ik, J. and Kope6ek, J., Strong-acid membranes with enhanced hydrophiticity, J. Membr. Sci. 1980, 6, 123-I 31 Thompson, R.A.M., Preparation of ionic polymers, in Developments in Ionic Polymers, (Eds A.D. Wilson and H.J. Prosser), Elsewer, New York, NY, 1986, pp. 1- 76 Hotliday,L., Classification and general properties of ionic polymers, in /omc Polymers (Ed. L. Holliday). Wiley. New York, NY, 1975, pp. 1-68 Du.~ek,K. and Jana~ek. J., Hydrophifrc gels based on copotymers of 2-hydroxyethyl methacrvtate w~th methacrylamtde and acrylamide, J_ AppL Potym. Sci 1975. 19, 3061-3075 Cartlidge, SA., Duncan, R., Lloyd, J.B.. Kopeckova-Rejmanova, P. and Kope~.ek, J., Soluble, crosslinked N-( 1-hydroxypropyI) methacrylamide copolymers as potentlal drug carriers, J. Contr. Ret. 1987, 4, 265-278 Kulicke,W.M. and Nottelmann, H., Rheological and swelling studies of synthetic polymer-networks in comparison to biopolymer-networks, Polym. Mater. Sci. Eng. Prepr. 1987, ,57, 265-269 Ulbrich, K., Kohak, C., Tuzar, Z, and Kope~ek, J., Solution properties of drug carriers based on polyIN-(2-hydroxypropyl)methacrylamide1 containing biodegradable bonds, Makromol. Chem. 1987, 188, 1261-1272 Yen, H.R., Kope(~ek, J. and Andrade, J.A., Synthetic water soluble copolymers for optically controlled ligand delivery, Polym. Mater. Sci. Eng. Prepr. 1987, 5 7 , 2 4 3 - 2 4 7 Plestit,J., Ostanevich, Y.M., 8orbely, S., Stejskai, J. and llavsky, M., Phase transition in swollen gets, Potym. Bull. 1987, 17, 465-472 McCormick, C.L., Hutchinson, B.H. and Morgan, S.E., Water-soluble copolymers, MakromoL Chem. 1987, 188, 357-370 Schuitz, J., The structure of liquid networks: polyacrylamide ~nwater, J. Polym. Sc/~, Polym. Lett. Ed. 1984, 22., 43-48 MacWilliams, D.C.. Acrytamide and other alpha, beta unsatured amides, in Functional Monomers (Eds R.H. Yocum and E.B. Nyquist), Dekker, New York, NY, 1973, pp. 1- 197 Itavsky, M,, Phase transition in swollen gets. 2. Effect of charge
Biomatenals t990, Vo/ t I November
643
60
The Biomaterials Silver Jubilee Compendium pH-sensitive hydroge/s: L. Brannon-Peppas and N.A. Peppas
30 31
32 33 34
35 36
644
concentration on the collapse and mechanical behavior of polyacrylamide networks, Macromolecules 1982, 15, 782-788 Greenwald, H.L. and Luskin, L.S., Poly(acrylic acid) and its homotogs, in Handbook of Water-Soluble Gums and Resins (Ed. R.L. Davidson), McGraw-Hill, New York, NY, 1980, pp. 1 7.1-17.19 Smetana, K. Jr, Sulc, J., Krcova, Z. and Pitrova, S., Intraocular biocompatibility of hydroxyethyl methacrylate and methacrylic acid copolymer/partially hydrolyzed poly(2-hydroxyethyl methacrylate), J. Biomed. Mater. Res. 1987, 21, 1247-1253 Ponchel,G., Touchard, F., Wouessidjewe, D., Duchene, D. and Peppas, N.A., Bioadhesive analysis of controlled-release systems, Inter. J. Pharm. 1987, 38, 65-70 Ricka,J. and Tanaka, T., Phase transition in ionic gels induced by copper complexation, Macromolecules 1985, 18, 83-85 Makushka, R.Y., Bayoras, G.I., Shulskus, Y.K., Bolotin, A.B., Roganova, Z.A. and Smolyanskii, A.L., Effect of complex formation on reactivity of acrylic and methacrylic acids in radical polymerization, Polyrn. ScL U.S.S.R. 1985, 27, 634-641 Eustace,D.J., Siano, D.B. and Drake, E.N., Polymer compatibility and interpolymer association in the poly(acrylic acid)-polyacrylamidewater ternary system, J. AppL Polym. ScL 1988, 35, 707-716 Chatterjee, S.K., Malhotra, A. and Yadav, D., Influence of preferential solvation on the interpolymer complex formation between methacrylic acid-methacrylamide copolymer and poly(vi nyl pyrrolidone), J. Polym. ScL, Polym. Chem. 1986, 24, 2591-2597
Biomaterials 1990, Vol 11 November
37 38 39
40
41 42
43 44
Pinchuk, L., Eckstein, E.C. and Van De Mark, M.R., Effects of low levels of methacrylic acid on the swelling behavior of poly(2-hydroxyethyl methacrylate), J. AppL Polym. Sci. 1984, 29, 1749-1760 Franson, N.M. and Peppas, N.A., Influence of copolymer composition on non-Fickian water transport through glassy polymers, J. AppL Polym. Sci. 1983, 28, 1299-1310 Davidson, G.W.R., Ill and Peppas, N.A., Solute and penetrant diffusion in swellable polymers. VI. The Deborah and swelling interface numbers as indicators of the order of biomolecular release, J. Contr. Re/. 1986, 3, 259-271 Korsmeyer, R.W. and Peppas, N.A., Macromolecular and modeling aspects of swelling-controlled systems, in Controlled Release Delivery Systems (Eds T.J. Roseman and S.Z. Mansdorf), Dekker, New York, NY, 1983, pp. 77-90 Turner, D.T., Water sorption of poly(methyl methacrylate). 1. Effect of molecular weight, Polymer 1987, 28, 293-296 Hosaka,S., Yamada, A,, Tanzawa, H., Momose, T., Magatani, H. and Nakajima, A., Mechanical properties of the soft contact lens of poly(methyl methacrylate-N-vinyl pyrrolidone), J. Biomed. Mater. Res. 1980, 14, 557-566 Gaylord, N.G., Poly(maleic anhydride), J. Macromol. ScL-Revs. MacromoL Chem., 1975, C13, 235-261 Ritger, P.L. and Peppas, N.A., A simple equation for description of solute release. Ill. Fickian and anomalous release from swellable devices, J. Contr. ReL 1987, 5, 37-42
61
The Biomaterials Silver Jubilee Compendium
Macroencapsulation of dopaminesecreting cells by coextrusion with an organic polymer solution P. Aebischer, L. Wahlberg, P.A Tresco and S.R. Winn
Art/f/c/a/Organ Laboratory, Brown University, Providence, Rt 02912 USA (Received 8 March 1990; accepted 4 June 1990)
A new method of coextruding living cells in the core of a forming hollow fibre is described. PC 12 cells, an immortalized cell line which secretes large amounts of dopamine, and dissociated bovine adrenal chromaffin cells, a non-dividing cell type which also secretes dopamine, were coextruded by a dry-jet wet spinning technique through a double-lumen spinneret from a 15% weight by volume solution of poly(acrylonitrile vinyl chloride) in either dimethylsulphoxide (DMSO), dimethylacetamide (DMAC) or dimethylformamide (DMF). Closure of the fibre was achieved by mounting polytetrafluoroethylene tubes on a rotating coaxial wheel system which squeezed the forming hollow fibre at regular intervals. Spontaneous and potassium-stimulated release of catecholamines from the macrocapsules were quantified under static conditions by ion-pair reverse-phase high-performance liquid chromatography equipped with electrochemical detection at 2, 4 and 6 wk. At all time periods, coextruded macrocapsules with either PC12 cells or adrenal chromaffin cells released dopamine under either unstimulated or stimulated conditions. An increase over time in dopamine release was observed from PC ! 2 cell coextruded macrocapsules with observable difference between capsules extruded with DMSO, DMAC or DMF as solvents. Well-preserved PC12 cells and adrenal chromaffin cells were present in coextruded macrocapsules with no observable difference between capsules extruded with DMS0, DMAC or DMF as inocuity of macroencapsulation by coextrusion from an organic polymer solution. Owing to the particular fluid dynamics of this technique, minimal potentially toxic ceil-solvent contact occurs allowing the use of a wider range of water-insoluble polymeric systems. The ability of encapsulated PC 12 cells and adrenal chromaffin cells to release spontaneously dopamine over time also suggests that polymer encapsulation may provide an alternative to the transplantation of dopamine-secreting cells in the treatment of Parkinson's disease. Keywords: Acrylate, encapsulation, coextrusion
Certain clinical conditions, such as diabetes or Parkinson's disease, may be ameliorated through the use of transplanted cells which secrete the missing molecule in the diseased state. Living cells can be transplanted from one species to another if they are immunoisolatecl from the host by a permselective membrane. Two techniques, microencapsulation and macroencapsulation, have been designed for this purpose, in microencapsulation, the cells are sequestered in a small permselective spherical container, whereas in macroencapsulation the cells are entrapped in a preformed tubular membrane closed at both ends. Microencapsulation through an interfacial adsorption technique has been reported with the use of polyelectrolyte gels 1' 2 In this Correspondenceto Dr P. Aebischer.
technique, gelation of polysaccharides, such as sodium alginate, around cell clusters is induced by exposure to multivalent cations. Next, a semipermeable membrane is layered on to the periphery of the gelled microspheres adding step-by-step various reagents including polycationic amino acid compounds such as polylysine. While this technique has the advantage of avoiding the use of organic solvents, the water solubility and mechanical fragility of the system makes it impractical for long-term use, which is a potential problem when encapsulating tumour cell lines or xenogeneic tissue. Sefton and collaborators first focussed on the application of water-insoluble polymeric systems, such as polyacrylates, to the production of durable, implantable, microencapsulated tissue cells through an interfacial precipitation technique 3-6. While the water insolubility of the 91991 Butterworth-HeinemannLtd. 0142-9612/91/010050-07
50
Biomaterials 1991, Vol 12 January
62
The Biomaterials Silver Jubilee Compendium Macroencapsulation of dopamine-secreting cells: P. Aebischer et ai.
polyacrylate systems is an advantage for maintaining capsule stability when exposed to body fluids, the exposure of the encapsulated cells to potential toxic organic solvents constitutes a major disadvantage. The technique of macroencapsulation involves loading cells into hollow fibres and then closing the extremities at both ends with a polymer glue 7. This procedure has been used to encapsulate a variety of mammalian tissue such as pituitary 8, islets of Langerhans 7' 9, parathyroid lO, thymus11 and embryonic mesencephalon12 In contrast to microcapsules, macrocapsules offer the advantage of easy retrievability, an important feature in neural implants. Unreliable closure with conventional macroencapsulation has provided, however, inconsistent results 7, 10. To try to correct this problem, we have developed a new technique of coextruding living cells in the core of forming hollow fibres and integrally sealing the partially coalescent organoget into discrete tubular macrocapsules. In the present study, we used dopamine-secreting cells as a test model for polymer encapsulation. Parkinson's disease is characterized by the lack of dopamine within the striatum following the degeneration of the dopaminergic nigrostriatal system 13' 14. Striatal implantation of polymer rods which release sustained amounts of dopamine has been reported to alleviate experimental parkinsonism in rodents, indicating that the sole release of dopamine in the proper target structure may be sufficient to correct this functional deficiency 15. In contrast to the finite capacity of a polymeric drug release system, polymer-encapsulated dopamine-releasing cells may provide a continuous supply of neurotransmitters. Besides their potential clinical relevance, the use of dopamine-secreting cells as a model to evaluate encapsulation techniques offers several advantages as compared with the use of insulin-secreting cells. There are a variety of cells secreting dopamine, including cell lines and engineered cells, which do not involve complex isolation procedures as is required for islets of Langerhans. Also, dopamine can be readily quantified with chromatographic techniques.
Polymers Poly(acrylonitrile vinyl chloride) (PAN-PVC) (65:33) containing 2% acrylamide 2-methylpropane sulphonate was obtained from Polysciences Inc., Warrington, PA. HPLC grade dimethylsulphoxide (DMSO), dimethylacetamide (DMAC) and dimethylformamide (DMF) were purchased from Sigma. The raw polymer was purified by four successive dissolution in DMSO and precipitation in distilled water. The DMSO was subsequently removed by lyophilization. Purified PAN-PVC solutions 15% (w/v) were prepared with either DMSO, DMAC or DMF and their viscosity measured at 40~ with a Cannon-Fenske viscometer (Cannon Instrument Co., State College, PA). The viscosity in centiPoises equaled 879 for DMSO, 823 for DMAC and 548 for DMF.
Macroencapsulation procedure The coextrusion apparatus consists of three electronically controlled, programmable infusion pumps (Infusion pump 22, Harvard Apparatus Inc., South Natick, MA); a stainless steel jet spinneret (inner tube: inner diameter 0.5 mm, outer diameter 0.8 mm; external tube: 1.1 mm, 1.4 mm); two motor-controlled, coaxial wheel systems (outer diameter 31 ram) on the perimeter of which occluding polytetrafluoroethylene (PTFE) tubes (outer diameter 3 mm, with a 1 3 mm spacing) are mounted; and an uptake quench bath (Figure 1 ). The macrocapsules are formed by the injection of a polymeric solution, at a flow of about 0.4 ml/min, into the outer tube of the spinneret. A coagulant, typically the encapsulated cells in their culture medium, is simultaneously injected in the spinneret inner tube at a flow of 0.8 ml/min. The encapsulating membrane is formed by a dry-jet, wetspinning process, i.e. the fast stabilization of the polymer solution emerging from the spinneret orifice by the internal quenching medium coupled with further stabilization in a physiologic saline quench bath. The closure of the fibre is performed by mechanically squeezing the forming hollow
MATERIALS AND METHODS
Dopamine-secreting cells PC12 cells, an immortalized cell line derived from a rat pheochromocytoma which secretes large amounts of dopamine16, were kindly provided by Dr L. Greene, Columbia University. PC1 2 cells were cultivated on collagen-coated tissue culture dishes in RPMI 1640 (Gibco, Grand Island Biological Co., Grand Island, NY) supplemented with 1 0% heat-inactivated horse serum and 5% fetal calf serum (Hazelton, Lenexa, KN) at 37~ in a water-saturated, ambient air atmosphere containing 7% C02. Dissociated bovine adrenal chromaffin cells, a non-dividing cell type which secretes dopamine were kindly provided by Dr J. Sagen, University of Illinois. These cells were maintained in Dulbecco's Modified Eagles Medium (DMEM)supplemented with 5% fetal calf serum. All media were supplemented with 50 units/ml of penicillin and 50 ng/ml of streptomycin (Sigma Chemicals Co., St Louis, MO). Before encapsulation, the cells were harvested by gently aspirating the culture dishes, followed by centrifugation at 800g. The cells were resuspended in the media and counted with a haemocytometer after treatment with trypan blue to determine viable cell concentration. The cells were then loaded at a concentration of 1 X 105 cells/ml in a 3 ml syringe.
E
r/////,l-..4..~ ~ = ~ i SOLVENT + POLYMER
A~
'"
Figure 1 Extrusion apparatus showing the three electronically controlled, programmable infusion pumps: (a) the jet spinneret, (b) the variable controlled coaxial wheel system on the perimeter of which PTFE tubes are mounted for pinching the capsules and (c} a quench bath. A magnified representation of the spinneret shows (d) the outer tube through which the polymer solution is injected, and (e) the inner tube through which the ceilcontaining aqueous solution is injected.
Biomateriais 1991, Vol 12 January
51
63
The Biomaterials Silver Jubilee Compendium Macroencapsulation of dopamine-secreting cells: P. Aebischer et al.
fibre with the coaxial wheel system before immersion in the quench bath. Near the spinneret head, 5 to 10 cm from the nozzle, the solvent concentration is sufficiently high to allow proper fusion of the hollow fibre wall. The number of PTFE tubes mounted on the coaxial wheels typically controls the capsule length. In the present report, 1 cm long capsules were extruded. Following each encapsulation procedure, pure solvent is flushed automatically through the lumen of the spinneret to avoid clogging the nozzle. A 15% (w/v) PAN-PVC solution in either DMSO, DMAC or DMF was loaded into a 5 ml glass syringe. When dopamine-secreting cells were coextruded, both the cell and the polymer solution were first run through the spinneret, and the capsules were then collected in a physiologic saline solution. The capsules were rinsed, cut at the midpoint of the squeeze, and placed in individual wells containing the appropriate culture medium for 2, 4 and 6 wk. When a blue dextran solution was coextruded, the cell suspension solution was replaced by a 1% of a 2 • 106 Daltons blue dextran (Polysciences inc., Warrington, PA) solution in distilled water. The blue dextran-loaded macrocapsules were then placed in sterile, physiologic saline solution and watched over time for potential leaks. The macrocapsule microgeometry was assessed with the aid of a scanning electron microscope (AM Ray IO00A). To assess the potentially toxic effect of solvent on the cells during encapsulation, the coextrusion technique was compared with the conventional macroencapsulation procedure as in this latter method no organic solvents are used. Both PC 12 cells and adrenal chromaffin cells were loaded at an initial concentration of 1 • 105 cells/ml into prespun acrylic copolymer fibres by a gentle injection of the appropriate cell solution through a 25 gauge needle. The hollow fibre's extremities were then closed by applying bone wax.
Membrane permeability determination The relative equilibration rate of the hollow fibre membrane was measured in a diffusion chamber consisting of a cylindrical plexigtass body 6 cm long and 1 cm internal diameter closed at both ends by plexiglass caps (1 cm long) on which silicone etastomer O-rings were mounted. A stainless steel tubing (0.8 mm inner diameter) 1 cm long was pressed and glued into the centre of the caps allowing a single hollow fibre 8 cm long to be threaded through the centre of the chamber. The fibre was secured to the stainless steel tubing with a polyurethane potting solution (Enka, Wuppertal, FRG). The body of the chamber possessed two additional ports which were used to either fill or empty the chamber. All ports were fitted with silicone rubber removable caps. For relative equilibration rate determination, the chamber was first filled with 3 ml of l OmM sodium phosphate buffer at pH 7.2. The hollow fibre was filled with approximately 20 p i of the appropriate radiolabelled market solution. The diffusion chamber was then placed in an incubator at 37~ for 24 h. The following molecular weight markers obtained from New England Nuclear (Wilmington, DE) were tested individually: D-glucose [3-3H] (Mw: 180), inulin [methyl-14C] (Mw: 5000), cytochrome C [methyl-14C] (Mw: 12 300), carbonic anhydrase [methyl-14C] (Mw: 30000), bovine serum albumin [methyl-14C] (Mw: 69000), phosphorylase B [methyl-14C] (Mw: 97 400), myosin [methyl-14C] (Mw: 2 0 0 0 0 0 ) . Three diffusion chambers were used for each marker. Following the 24 h incubation period, the 3 ml of phosphate buffer was removed from the
52
Biomaterials 1991, Voi 12 January
chamber. The fluid contained in the hollow fibre was flushed out with 1 ml of fresh phosphate buffer. The hollow fibre was removed from the diffusion chamber. All three components were assayed for radioactivity using a dual-channel liquid scintillation counter (model LS 3801, Beckmann, Futlerton, CA). Relative equilibration rate values were calculated as 1 - (radioactivity in the diffusion chamber/total radioactivity).
Catecholamine determination Spontaneous and potassium-evoked release of catecholamines was quantified under static incubation conditions by ion-pair reverse-phase high-performance liquid chromatography equipped with electrochemical detection (LCEC) at 2, 4 and 6 wk. Each capsule was first rinsed twice with 1 ml of Hank's Balanced Salt Solution (HBSS) containing 5.4 mM K § and then returned to the incubator in 5 0 0 p t of HBSS medium for 1 5 min. The medium was then removed, and replaced with 500 pl of HBSS containing 56.0 mM K§ for an extra 1 5 min. The medium was subsequently assayed for catecholamines after concentration with alumina. The LCEC system consisted of a 5700 model, solvent-delivery system, reverse-phase-H R-80 column, and a model 5100A Coulochem muttielectrode, electrochemical detector (ESA, Bedford, MA) operated at 1.5 mi/min. A 20 pl aliquot of each sample was injected on to the column (CA-HR 80; ESA). The mobile phase contained 6.9 g/I sodium monobasic phosphate, 80 mg/I EDTA, 212 mg/i heptane sutphonic acid, and 3% methanol, at a pH of 2.6. Total run time was approximately 8 min. The concentration of each compound was determined by comparison with the peak height of serial diluted standards run with each assay. The dopamine detection limit of the chromatographic system used was 50 pg.
Morphological analysis Ceil-loaded capsules were placed in 2.5% paraformaldehyde and 2% glutaratdehyde in phosphate buffered saline (PBS) for 24 h at the various retrieval times. The capsules were post-fixed in 1% osmium tetroxide in PBS, dehydrated in a graded series of alcohols, and embedded in Spurr's resin. Semithin sections (0.8/jm) were stained with totuidine blue and basic fuchsin while ultrathin sections (60 nm) were stained with uranyl acetate and lead citrate. Transmission electron microscopic analysis was performed with a Philips 410. in order to access their content, the capsules were fractured following immersion in liquid nitrogen and examined under a scanning electron microscope (SEM). Following post-fixation, the capsules were washed, dehydrated, critical point dried, and sputter-coated with gold palladium.
RESULTS
Capsule characterization The capsules' measurements were: length 10 + 0.1 mm, inner diameter 450 _+ 28 pm, outer diameter 652 + 22/Jm. The thickness of the inner skin was approximately 1/j m. The relative equilibration rate for all markers used is illustrated in F i g u r e 2. Only minimal diffusion occurred for substances larger than 50 000 Daltons. No leakage of blue dextran was observed when assessed by a DU-65 spectrophotometer (Beckman, Fullerton, CA) at 620 nm or by macroscopic observation either acutely or for periods up to 3 month in physiologic saline when pinching of the forming fibre was
64
The Biomaterials Silver Jubilee Compendium Macroencapsulation of dopamine-secreting cetls: P Aebischer et aL 1.0 1
0.8
.~
0.6
J~ l.IJ >~
0.4
0.2
0,0
50
100
150
200
Molecular Weight Figure 2 Relativeequilibration rate as a function of the molecular weight for a fibre spun from a 15% (w/v) PAN-PVC solution in DMSO as measured under diffusion for the following radiolabetled molecules: glucose, 180: inulin, 5000; cytochrome C, 12300; carbonic anhydrase, 30000; bovine serum albumin, 69000; phosphorytase B 97400; myosin, 200000. Bars represent SD; n = 3 for each molecule.
was suggested by the presence of numerous m~totic figures and by the observation that the capsule space occupied by cells increased over time. Cellular ghosts and debris were observed predominantly in the centre of the capsule's lumen, reflecting potential restrictive diffusive transport of oxygen and nutrients. At the transmission electron microscope, well-preserved PC 1 2 cells, with their typical electron-dense secretory granules, were seen. Although initially coextruded as a cell suspension, 1 wk post-encapsulation adrenal chromaffin cells reassembled in densely packed aggregates (Figure 7a). Four weeks post-encapsulation, all chromaffin cells were included in one or two pieces of dense tissue resembling the normal adrenal medulla tissue. Wellpreserved adrenal chromaffin cells with their typical electrondense secretory granules were seen at the transmission electron micrograph (Figure 7b). Occasionally, a monolayer of fibroblasts surrounded the packed chromaffin tissue as well as the capsule's inner lumen. No noticeable difference in cell morphology was observed between the hand-loaded and the coextruded macrocapsules with either PC 12 cells or adrenal chromaffin cells. At the scanning electron microscope, the microgeometry of the cell-coextruded capsules was identical to that observed with blue dextran-coextruded capsules.
performed at a distance smaller than 1 0 c m from the spinneret nozzle, At the SEM, the capsule wall featured the typical morphology of asymmetric membranes with an inner skin on the luminal side, supported by a trabecular network in the wall and a fenestrated outer membrane (outer skin). Fusion of the capsule wall was observed at the pinched area ( F i g u r e s 3a and 3 b ) . Leakage of blue dextran was observed and measured within the first hours following the encapsulation procedure when the squeezing of the membrane was performed at a distance greater than 1 0 cm from the nozzle. Incomplete fusion of the wall membrane was then evidenced with SEM.
Cell-loaded capsules Cell viability as assessed by trypan blue exclusion before and immediately following the coextrusion process was identical, suggesting that this encapsulation system is effective in controlling the exposure of the cells to organic solvents. All cell-loaded capsules released dopamine into the medium under basal conditions at 2, 4 and 6 wk post-encapsulation, High potassium treatment, a chemically induced depolarization procedure, increased dopamine release from both PC1 2 and adrenal chromaffin cells (Figure 4). A similar release profile was observed with coextruded macrocapsules of PC 1 2 and chromaffin cells as compared with hand-loaded macrocapsules (Figure 4). In contrast to chromaffin cells, dopamine release from coextruded PC12 cells, a turnout cell line, increased with time (Figure 5). This increase is believed to be related to cell proliferation within the polymer capsule. An increase in dopamine release was observed with the three solvent systems used to coextrude PC1 2 cells. At all time periods, no significant difference in dopamine release was observed from PC12 cells-loaded capsules extruded with DMSO, DMAC or DMF as solvents (Figure 5). Morphological analysis of the coextruded macrocapsules revealed the presence of small clusters of PC12 cells dispersed throughout the lumen of the capsule ( F i g u r e s 6a and 6 b ) . Cell division within the capsule space
Figure 3 Scanning electron micrographs of a macrocapsule coextruded with a t5% (w/v) PAN-PVC solution in DMSO. The forming fibre was pinched 5 cm from the nozzle with the coaxial wheel system. Note on a longitudinal section the area of the pinch (a). A higher power of a pinch area shows fusion of the tuner skin (b).
Biomaten'als 1991, Vol 12 January
53
65
The Biomaterials Silver Jubilee Compendium Macroencapsulation of dopamine-secreting cells: P, Aebischer et aL 700
250OO t-
A
.-=
E
E ,c=.=
tt~
2O0OO
G)
:3
600
l
500
i
D.
u
15000
tj
D. 10000
g:
< Q
400
r
5000
G)
300
n-
2OO
Q
CE
HL
T
t00
CE
HL
PC12
Chromaffin
Figure 4 Histogram of the dopamine (DA) release at 4 wk from PC 12 cells and adrenal chromaffin cells coextruded with a 15% (w/v) PAN-PVC solution in DMSO orhand-loaded in a prespun PAN-PVC hollow fibre. Dopamine release was measured under basal conditions ([3) and after a 15 rain exposure to a medium containing 56 mM potassium (•). Note the similarity in dopamine release between coextruded v e r s u s hand-loaded macrocapsufes for both PC 12 ceils and adrenal chromaffin cells. Bars represent SD; n = 6 capsules,
DISCUSSION The present study demonstrates that dopamine-secreting cells can be coextruded with an organic polymeric solution and still display good viability as well as adequate physiological response, such as increase in dopamine output, following potassium-induced depolarization. The PC 12 cells proliferate and maintain their phenotype over time but are not able to escape or burst the capsule, an absolute requirement if encapsulated tumour cell lines are to be considered clinically. The present study shows that more sensitive post-mitotic 500 A
e-
,===
E
It)
D, o v
400
300
c Q)
200
m i
n-
<
100
2
3
Time
4
5
6
7
(weeks)
Figure 5 Dopamine (DA) release from PC 12 celt-loaded capsules coextruded with a 15% (w/v) PAN-PVC solution using DMSO (F1), DMAC ( i ) or DMF (/k ) as a solvent. Note the increase and similarity of dopamine release over time independent of the solvent used in the coextrusion process, Dopamine release from each capsule was measured after a 15 rain exposure into media containing 56 mM potassium. Bars represent SD; n = 6 capsules.
54
Biomaterials 1991, Vol 12 January
cells, such as adrenal chromaffin cells, also survive and maintain their phenotype following coextrusion with an organic polymer solution. The similarity in dopamine release, as well as the absence of a noticeable difference in the morphological appearance between coextruded and handloaded PC 12 cells and adrenal chromaffin cells, indicate that macroencapsulation by a coextrusion technique is not harmful to cell suspensions. Over time, the similarity of dopamine release from PC12 cells enclosed in macrocapsules coextruded with either DMSO, DMAC or DMF suggests that this particular technique of encapsulation may prevent cell damage inflicted by organic solvents. Sefton et aL have implicated organic solvents as being one of the major drawbacks for the microencapsutation of mammalian cells using non-water-soluble systems 7-9. In microencapsulation by interfacial precipitation, solvents can diffuse inside the capsule and exert their potential deleterious effects. In macroencapsulation by coextrusion, due to the polymer precipitation syneresis effect and a presumably higher pressure of the inner bore system, the solvent is quickly driven toward the outside of the polymer capsule preventing, therefore, extended cell-solvent contact. This premise is suggested by the observation that pearls of solvent form on the outside wall of the capsule during the coextrusion process. The loading of capsules with blue dextran or the use of the scanning electron microscope demonstrates that adequate closure can be obtained by imposed mechanical pressure while some solvent is still entrapped in a forming hollow fibre. Two wheels with occluding elements on their periphery cooperate to pinch periodically the tubular extrudate and thereby seal it. This mechanical compression system can be replaced by a variety of other mechanical or pneumatic compression systems to seal the tubular extrudate at regular intervals. Alternatively, the extrudate can be sealed by interruption of the flow of non-solvent. In normal operation, the infusion pumps are controlled to maintain a pressure differential between the aqueous cell suspension and the polymeric solution, so that the polymer solvent is driven outward during coagulation. By periodically interrupting the
66
The Biomaterials Silver Jubilee Compendium Macroencapsulation of dopamine-secreting cells: P. Aebischer et al.
Figure 6 Histological profiles of PC 12 cells coextruded from a 15% (w/v) PAN-PVC solution in DMSO and maintained for 4 wk in vitro. {a) Cross-sectional scanning electron micrograph showing the relationship between the polymer capsule and the PC 12 cells. After removal of the polymer wall (b), note the homogenous PC 12 cell cluster arrangement over the capsule length. A light micrograph (c) reveals the cellular ghosts surrounded by intact PC 12 cells. A transmission electron micrograph (d) shows well-preserved PC 12 cells with their typical electron-dense secretory vesicles.
non-solvent flow, the tubular extrudate can be collapsed at intervals to define individual cell compartments. As compared to the hand-loaded technique, the macroencapsulation of cells by coextrusion allows a better control of sterility because of the high degree of automation. It should, therefore, be more efficient in terms of batch processing.
No difference was observed in the morphological appearance of PC12 cells in the polymer capsule whether the cells were coextruded or hand-loaded. This observation was also true for adrenal chromaffin cells. The PC1 2 cells were scattered in small cell clusters, whereas the adrenal chromaffin cells were found densely packed in one or two
Figure 7 Micrographs of adrenal chromaffin cells coextruded with a 15% (w/v) PAN-PVC solution in DMSO and maintained for 4 wk in culture. Note m the light micrograph (a) a large adrenal chromaffin cell aggregate. Note the presence in the electron micrograph (b) of the typical electron-dense secretory granules in the adrenal chromaffin cells.
Biomaterials 199 1, Vol 12 January
55
67
The Biomaterials Silver Jubilee Compendium Macroencapsulation of dopamine-secreting cells: P. Aebischer et al.
aggregates. Using the same coextrusion technique, macroencapsulation of LLC-PK1 cells, a kidney epithelial cell line derived from the dog proximal renal tubule, resulted in a confluent cell monolayer w h i c h entirely covered the internal surface of the capsule (unpublished observations). The same m o r p h o l o g y w a s reported for LLC-PK1 seeded on prespun PAN-PVC hollow fibres 17' 18 It appears, therefore, that the cell and not the encapsulation t e c h n i q u e dictates the spatial cellular organization in coextruded macrocapsules. Little is k n o w n about the tissue reaction of the brain to polymeric implants. In preliminary experiments, w e reported that PAN-PVC macrocapsules induce only minimal host rat brain reaction consisting mainly of o n e - t o - t w o layers of reactive astrocytes positive for the glial fibrillary acidic protein marker 19. Neurones w e r e found at an average of 3 0 p m f r o m the capsule wall, Provided careful implantation techniques are used, an inhibitive fibroblastic response to implanted PAN-PVC macrocapsules is not observed in the brain, s u g g e s t i n g that the brain may be a privileged site for the implantation of encapsulated neurotransmitter-secreting cells. Our results s h o w that both immortalized and differentiated d o p a m i n e - s e c r e t i n g cells survive macroencapsulation by coextrusion t h r o u g h the core of f o r m i n g h o l l o w fibres. Because of its particular fluid dynamics, coextrusion of macrocapsules may allow the use of a w i d e r range of p o l y m e r / s o l v e n t systems and avoid the difficulties intrinsic to loading living cells in p r e f o r m e d fibres w h i c h m u s t then be sealed. The ability of these capsules to release spontaneously d o p a m i n e over time suggests that polymer encapsulation may provide an alternative to the transplantation of dopamine-secreting cells in the treatment of Parkinson's disease.
3 4 5 6 7
8 9 10 11 12 13
14 15
ACKNOWLEDGEMENTS
16
W e w o u l d like to a c k n o w l e d g e the technical s u p p o r t of Georg Panol and Henry Ashley.
17
REFERENCES 1 2
56
Lim, F. and Sun, A.M., Microencapsulated islets as bioartificial endocrine pancreas, Science 1980, 210, 908-910 Goosen,M.F.A., O'Shea, G.M., Gharapetian, H.M., Chou, S. and Sun,
Biomaterials I99t, Voi 12 January
18 19
A.M., Optimization of microencapsulation parameters: semipermeable microcapsules as a bioartificial pancreas, Biotech. Bioeng. 1985, 27, 146-150 Lamberti,F.V.and Sefton, M.V., M icroencapsulation of erythrocytes in Eudragit RL: coated calcium atginate,Biochim. Phys. Acta 1983,759, 81-82 Dawson,R,M., Broughton, R.L., Stevenson, W.T.K. and Sefton, M.V., Microencapsulation of CHO cells in a hydroxyethyt methacrylatemethyl methacrylate copolymer, Biornaterials 1987, 8, 360-366 Stevenson,W.T.K. and Sefton, M.V., Graft copolymer emulsions of sodium alginatewith hydroxyalkylmethacrylatesfor microencapsulation, Biomaterials 198 7, 8, 449-457 Mallabone,C.L., Crooks, C.A. and Sefton, M.V., Microencapsulation of human diploid fibroblasts in cationic polyacrylates, Biomaterials 1989, 10, 380-386 Altman,J.J., HoulberL D., Callard, P., McMillan, P., Solomon, B.A., Rosen, J and Galletti, P.M., Long-term plasma glucose normalization in experimental diabetic rats with macroencapsulated implants of benign human insulinomas, Diabetes 1986, 35, 625-633 Hymer,W.C., Wilbur, D.L., Page, R., Hibbard, R., Kelsey, R.C. and Hatfield, J.M., Pituitary hollow fiber units in vivo and in vitro, Neuroendocrinology 1981, 32, 339-349 Archer,J., Kaye, R. and M utte, B.S., Control of streptozotocin diabetes in Chinese hamsters by cultured mouse islet cells without immunosuppression: a preliminary report, J, Surg. Res. 1980, 28, 77-85 Aebischer,P., Russell, P.C., Christenson, L., Panol, G., Monchik, J.M. and Galletti, P.M., A bioartificial parathyroid, Trans. Am. Soc. Artif Intern. Organs 1986, 32, 134-137 Christenson,L., Aebischer, P. and Galletti, P.M., Encapsulated thymic epithelial cells as a potential treatment for immunodeficiencies, Trans. Am. Soc. Artif Intern. Organs 1988, 34, 681-688 Aebischer,P., Winn, S.R. and Galletti, P.M., Transplantation of neural tissue in polymer capsules, Brain Res. 1988, 448, 364-368 Ehringer,H. and Hornykiewicz, O., Vetreilung yon Noradrenalin und Dopamin (3-Hydroxytyramin) im Gehirn des Menschen und ihr Verhatten bei Erkrankungen des extrapyramidalen Systems, K/in.Ther. Wochenschr. 1960, 38, 1236-1239 Hornykiewicz,O., Dopamine (3-hydroxytyramine) and brain function, Pharmacol. Rev. 1966, 18, 925-964 Winn, S.R., Wahlberg, L., Tresco, P.A. and Aebischer, P., An encapsulated dopamine-releasing polymer alleviates experimental parkinsonism in rats, Exp. NeuroL 1989, 105, 244-250 Greene,L.A. and Tischler, A.S., Establishment of a noradrenergic clonat line of rat adrenal pheochromocytoma cells which respond to nerve growth factor, Proc. Nat/ Acad. Sci., USA 1976, 73, 2424-2428 Aebischer,P., Ip, T.K., Miracoli, L. and Galletti, P.M., Renal epithelial cells grown on semi-permeable hollow fibers as a potential uttrafiftrate processor, Trans. Am. Soc. Artif Intern. Organs 1987, 33, 96-102 Ip, T.K. and Aebischer, P., Renal epithelial-cell-controlled solute transport across permeable membranes as the foundation for a bioartificial kidney, Artif Organs 1989, 13(1 ), 58-65 Winn, S.R., Aebischer, P. and Galletti, P.M., Brain tissue reaction to permselective polymer capsules, J. Biomed. Mater. Res. 1989, 23, 31-44
The Biomaterials Silver Jubilee Compendium
68
This Page Intentionally Left Blank
The Biomaterials Silver Jubilee Compendium
69
Interaction between phospholipids and biocompatible polymers containing a phospho lcholine moiety Masayoshi Kojima,
Ishihara, Akihik0 Watanabeand NobuoNakabayashi
Institute for Medical and Dental Engineering, Tokyo Medical and Dental University, 2-3-10 Kaoda-surugada~ Chiyoda-ku, Tokyo 101, Japan Presented at Biointeractions "90, Oxford, UK 21-23 August 1990
Random and block copolymers containing a phospholipid polar group in their side chain were synthesized by the copolymerization between 2-methacryloyloxyethyl phosphorylcboli,e and styre,e. These copolymers showed amphiphilic character, especially poly(methacryloylexyethyl phosphorylcholineblock-styrene) formed stable polymer micelles i, water. The interaction between natural phospholipid, dipalmitoylphosphatidylcholine and methacryloyloxyethyl phosphorylcholine copolymers was investigated. The amount adsorbed of dipalmitoylphosphatidylcholine from its liposomal solution on to the poly(methacryloyloxyethyl phosphorylcholine-co-styrene) sudace increased with increase of methacryloy|oxyethyl phospherylcholine composition. Moreover, when poly(methacryloyloxyethyl phosphorylcholineblock-styrene) was added to dipalmitoylphesphatidylcholine solution, organization of dipalmitoylphosphatidylcholine molecules and stabilization of bilayer structure of dipaimitoylphosphatidylcholine liposome were found. This means that methacryloyloxyethyl phosphorylcholine moieties in the copolymer have a strong affinity to dipalmitoylphosphatidylcholine molecules. The blood compatibility of methacryloyloxyethyl phosphorylcholine copo|ymers was also investigated with particular attention to the aggregation ability of platelets after contacting methacryloyloxyethyl phosphorylcholine copolymers; this ability decreased when platelets were put in co,tact with polymers without a methacryloyloxyethyl phosphorylcholine moiety. On the other hand, aggregatio, ability remained at almost the same level to that of original platelets after contact with methacryloyloxyethyl phosphoryIcheline copolymers. From these findings, we concluded that methacryleylexyethyl phospherylcholine copolymers show excellent blood compatibility due to adsorption of lipids from plasma a,d the formation of an organized adsorption layer of lipids on the surface of the methacryloyloxyethyl phosphorylcholine r Keywords: B/ocompetibitity, copofymefs, phospholipids
Since biomaterials are used in contact with blood or body fluid, their design must be based upon the normal contribution to the maintenance of a desirable interaction with living tissues. The simplest common feature of biomembranes is the high content of the electrically neutral phospholipids which contain phosphorylcholine head groups1. It would be interesting to prepare biomaterials with good affinity for phospholipids, which are then covered with lipid membrane by adsorbing natural lipids 2. We pursued the development and application of amphiphilic polymers with a phospholipid polar group which we expected to form a phospholipid adsorption layer on the copolymer surface. 2-methacryloyloxyethyl phosphorylcholine (MPC) was prepared to provide a new biomaterial showing a e.
Correspondenceto ProfessorN. Nakabayashi.
desirable interaction with living tissues 3. For example, on the surface of M PC copolymers with n-butyl methacrylate (BMA), platelet adhesion and activation were completely suppressed and protein adsorption was also prohibited 4' 5 This good biocompatibility was considered to be caused by a highly organized lipid adsorption layer on the surface which was confirmed by X-ray photoelectron spectroscopic analysis of the surface pretreated with phospholipid liposome solution 6. In this paper, we will describe the interaction between random and block-type M PC copolymers with hydrophobic styrene (St) and phospholipid molecules. Furthermore, the blood compatibility of these copolymers was evaluated with particular attention to one platelet function, aggregation induced by ATP or collagen, after bringing them in with the copoiymers to clarify the mechanism of the good biocompatibility of the MPC copolymer surface.
91991 Butterworth-HeinemannLtd. 0142-961 2/91/0201 21-04 Biomateriafs 1991, VoI 12 March
121
The Biomaterials Silver Jubilee Compendium
70 Phospholipids and phosphorylcholine polymers." M. Kojima et ai.
EXPERIMENTAL
Table 1
Materials
Abbreviation
M PC was synthesized by the method previously reported and purified by recrystallization from acetonitrile 3. St was purified by distillation under reduced pressure and a fraction of bp 48~ mm Hg was used. Polystyrene (PSt) containing peroxide groups (PPSt) in the main chain was supplied by Nippon Oil and Fats Co. Ltd. The mol wt of PPSt was 10600. D,L-dipalmitoylphosphatidylcholine (DPPC) was purchased from Sigma.
Preparation of MPC copolymer with St Random copolymerization of M PC with St was carried out by a conventional radical copolymerization in chloroformethanol mixture using Vo70 as an initiator 7. The polymers were purified by reprecipitation from their chloroformethanol solution into diethylether (Figure 1). Block-type copolymers of MPC and St were prepared by polymerization of MPC by PPSt in chloroform-methanol mixture at 60 ~ overnight 8. The synthetic route of poly(M PCblock-St) is shown in Figure 2. The poly(MPC-biock-St) obtained was purified by selective extraction of both poly(M PC) and PSt using water and cyclohexane, respectively. The MPC mole fraction in these copolymers was determined by 1HNMR, phosphate assay and X-ray photoelectron spectroscopy (XPS) on the surface of the casted membrane. Results of these copolymerization are shown in Table 1.
Measurement of adsorption amount of DPPC An aqueous solution of DPPC (0.5 wt%) was prepared by conventional methods. The M PC copolymer was coated on CH3 I ----(--C-- CH2 ) . . . . . . .
I
a
C=O
I
t C H - C H 2-)-
I
+
Poly(MPC-co-St) Structure of MPC copolymer with St.
Synthetic route of poly(MPC.block-St) i ....
!---(o-
+ MPC
PSt chain with peroxide group
~
Heat
I Poly(M PC- block-St) CH3 I
(c-cH, ~=O
~ - C H - CH2 O II
+
OCH2CH20~OCH2C~N(CH3)3 OFigure 2
122
MPC mole fraction Infeed
In copolymer Bulk a
Surface b
Yield (%)
Swelling degreec
Surface/bulk
Random rMS-1 rMS-2 rMS-3
0.063 0.092 0.11
0.10 0.17 0.29
0.17 0.20 0.23
1.7 1.2 0.81
52.0 54.0 54.0
0.18 0.20 0.24
Block bMS-1 bMS-2
0.40 0.50
0.051 0.11
0.27 0.54
4.3 4.7
54.0 43.5
-
aMPC mole fraction was determined by 1HNMR for random copolymers and phosphate assay for block copolymers. bThe values were determined by XPS analysis on the surface of cast film on EVAL for random copolymers and polyethylene for block copolymers. CThe values were obtained after immersion at 30~ for 7 d.
the poly(ethylene-co-vinyl alcohol) membrane (EVAL, Kuraray Co. Ltd, vinyl alcohol composition: 0.32) by a solvent evaporation technique. The membrane coated with poly(M PCco-St) was immersed in the DPPC liposome solution at 45 ~ for 10 min. After washing with water, DPPC molecules adsorbed on the membrane were desorbed completely by ethanol. The phosphate content based on DPPC in the ethanol solution was measured by phosphate assay and the amount of DPPC adsorbed was calculated.
Differential scanning calorimetry measurement The DPPC solution and polymer emulsion (0.5 wt%) were mixed and incubated at 45 ~ for 10 min. We placed 50 pl of the mixture into an aluminium pan which was then sealed. The differential scanning calorimetry (DSC) measurement was carried out using a Seiko DSC-IO0. The heating rate was 0.6 ~
Measurement of aggregation ability of platelets
O-
O C H 2 C H 2 O P O C H 2 C H 2 N ( C H 3) 3 I! 0 Figure I
Copolymerization of MPC with St
Synthetic route of poly(MPC-block-St)
Biomaterials 199 I, Vol 12 March
Platelet-rich plasma (PRP) was prepared from Japanese white rabbits weighing 3.0 kg. The carotid artery was cannulated using poly(vinyl chloride) tubing and 90 ml of fresh blood was collected in a disposable syringe containing 10 ml of a 3.8 wt% aqueous sodium citrate solution. The citrated blood was immediately centrifuged for 1 5 rain at 750 rev rain -1 to obtain citrated PRP. The number of PRP platelets was adjusted to approximately 1 X 108 cells/ml by dilution with platelet-poor plasma, which was prepared from the citrated blood by centrifugation for 1 5 min at 2 6 0 0 rev 9
mln
-1
.
The PRP was contacted on the surface of cellulose membrane coated with either poly(MPC-co-St) or poly(2hydroxyethyl methacrylate)(poly(H EMA)), then the aggregation ability of platelets induced by ADP(62.5 pg/ml PRP) was estimated using an aggregometer (Sienco, DP-247E) to measure the transmittance of the PRP. In block-type copolymer case, a definite amount of poly(MPC-block-St) was added to PRP and the aggregation induced by collagen(62.5 pg/ml PRP) was estimated with the aggregometer.
RESULTS AND DISCUSSION Property and structure of MPC copolymers with St The M PC molecules had a good copolymerization ability with methacrylates and St derivatives. In random copolymer case,
71
The Biomaterials Silver Jubilee Compendium Phospholipids and phosphorylcholine polymers." M. Kojima et al.
the MPC composition in the copolymer was greater than that in feed. The random copolymers were soluble in methanol and chloroform. Water did not dissolve them. Since M PC is extremely hydrophilic, poly(MPC-co-St) membrane absorbed water and became the hydrogel. The water content increased with the increase in MPC composition. The MPC composition in the block-type copolymer was quite low, as indicated in Table 1. It was considered that PPSt hardly copolymerizes with M PC in the chloroformmethanol mixture, because the chain length of original PSt chain was too long to continue the polymerization. However, on the cast film surface, M PC composition increased drastically and its ratio to the bulk polymer was 4.3. The block-type copolymer was effective in concentrating the M PC on the surface of substrate. The block-type copolymer also showed a unique solution property. The stable emulsion of poly(MPC-block-St) could be prepared in water. The PSt chain is hydrophobic and the poly(MPC) chain is hydrophitic, therefore, the poly(MPC) chain surrounds the PSt region in water just like a polymer micelle.
Adsorption of phospholipid on M PC copolymers F i g u r e 3 shows the relation between the amount of DPPC adsorbed on the surface of EVAL coated with poly(MPC-coSt) and the MPC composition in the copolymer. It is clearly observed that the adsorption amount of DPPC increased with the increase in MPC composition. In general, when water content of materials increases, adsorption amount of substances decreases if there are no specific interactions. Therefore, M PC moieties on the surface had a strong affinity to DPPC molecules. We have already found that the DPPC molecules adsorbed on MPC copolymer, poly(MPC-co-BMA), were organized on the surface 6. That is, when the poly(MPC-coBMA) membrane was treated with DPPC solution and the surface was analysed by XPS, the ratio of phosphorous atom versus carbon atom-based on DPPC molecule exceeded the theoretical value calculated from the number of atoms in DPPC. This means that the head group of DPPC which has a phosphorous atom concentrated on the surface. It is strongly suggested that the DPPC molecules adsorbed end-on to the MPC copolymer surface and made a biomembrane-like structure. F i g u r e 4 shows the DSC curves of DPPC solution containing poly(MPC-block-St) emulsion compared with
1.0-
T
o
C o x LU
l
35
,
1
40
......
Temp (~
I
45
....
I
50
Figure 4 DSC curves of DPPC solution containing polymers." (A) original DPPC solution (0.5 wt%); (B) with PSt latex (0.5 wt%); (C) with poly(MPCblock, St) (bMS-2) emulsion (0.5 wt%).
that containing PSt latex. The original DPPC solution showed endothermic absorption at 41.8~ which corresponded to the gel-liquid crystalline transition of the DPPC liposomal bilayer structure. When PSt latex was added to the DPPC solution, the endothermic absorption disappeared. However, addition of poly(M PC-block-St) induced a large endothermic absorption peak. These results indicated that the bilayer structure was destroyed by addition of PSt latex because of random adsorption of DPPC molecules, whereas the formation of an organized structure was prompted by the addition of poly(M PC-block-St) as shown in F i g u r e 5. We have confirmed that the M PC copolymers can organize the adsorbed phospholipid molecules with a specific interaction.
Blood compatibility
of M PC copolymers
It has been reported that platelets did not adhere and activate on the surface of poly(MPC-co-BMA) and showed an excellent blood compatibility even for human whole blood 9. When platelets are in contact with a foreign surface, platelets adhere on the surface and/or accept damage of their functions. Therefore, it is important to evaluate the blood compatibility of the materials with attention to the change in platelet functions after contacting the materials. We
E E~
.3
Q~ 0.5 o -o 0cL I%
~o
,
I
I . . . . . . .
0.2
MPC composition
i
o,3
Figure 3 MPC composition dependence of amount of DPPC adsorbed on EVAL surface coated with poly(MPC-co-St).
Figure 5 Schematic representation of adsorption state of DPPC on PSt latex and poly(MPC-block-St).
Biomaterials 1991, Vol 12 March
123
72
The Biomaterials Silver Jubilee Compendium Phospholipids and phosphorylchotine polymers: M. Kolima et aL
investigated the change in the aggregation ability of platelets by putting them in contact with the MPC copolymers by using an aggregometer, Platelet aggregation occurs as a result of the addition of ADP, collagen and calcium ion, etc. By measuring the transmission of PRP, aggregation process can be followed. Figure 6 shows the principle of measurement of aggregation ability. That is, the transmission of PRP increases when platelets are aggregated by addition of ADP. The value of Z~ A2 denotes the maximum aggregation amplitude. Figure 7 shows the time dependence of aggregation ability after putting them in contact with various polymers compared with original PRP. The mean values of duplicate samples are shown. It decreased even original PRP slightly with increase of incubation time. A greater decrease was found by placing them in contact with poly(HEMA) and cetlutose(Cuprophan), which are typical biomedical polymers. In the case of poty(M PC-co-St), the decrease of aggregation ability of platelets was remarkably suppressed. This means that the interaction between platelets and poly(MPC-co-St) was milder than between poly(HEMA) and cellulose. Figure 8 shows the effect of poly(MPC-block-St) addition on platelet aggregation induced by collagen. Even when various doses of the block copolymer were added to PRP, there was no significant difference in platelet aggregation in this experiment. Therefore we can say that amphiphilic block copolymers with MPC chain interact mildly to ptatetets. We believe that the mild interaction between MPC copotymers 100
~
100
0
lOOpg
. . . . . . . . . . 20Oug
v
4
I't~
E C C~
t--
lmin Figure8 Effect of poly(MPC-btock-St) (bMS-2) addition to PRP on aggregation of p/atelets induced by collagen.
and platelets can be attributed to the formation of organized lipid adsorption layer from plasma on the polymer surface,
CONCLUSION Copolymers with phospholipid polar group, MPC copolymers, have a strong affinity for natural phosphotipids and can organize the adsorption layer of the tipids on the surface. MPC copolymers interact with platetets less than other polymers widely used in the biomedical field.
ACKNOWLEDGEMENT v
Part of this investigation was supported by a Grant for Scientific Research (No. 0 2 2 0 5 0 3 3 ) from The Ministry of Education, Science and Culture, Japan, for which one of the authors (K.I.) expresses his appreciation.
ADP
o
/
E ~0 C
p.
REFERENCES 2min
Figure 6 Transmission change of PRP based on ptatetet aggregation induced by ADP.
70 60 50 04
<
40 30 20
._, .................... 140
--1............................ 190
Incubation
' :-.... 240
Time (rain)
Figure 7 Contact-time dependence of aggregation ability of platelets after contacting with various polymers. (0) Original PRP, contact with ( 9 poly(MPC-co-St) (rMS-3), t"t:3)po/y(HEMA), (A) cellulose (Cuprophan).
t24
Biomateriafs t99 I, Vol 12 March
Hayward,J.A. and Chapman,D, Biomembranesurfaceas modelsfor polymer design: the potential for haemocompatibility,Biomaterials 1984, 5, 135-142 Fukushima,S., Kadoma,Y. and Nakabayashi,N., Interaction between the polymercontainingphosphorytchotinegroupandceils,Koubunshi Ronbunshu 1983, 40, 785-793 Ishihara, K., Ueda.T. and Nakabayashi,N., Preparationof phospholipid polymersand their propertiesas polymerhydrogetmembrane.Polym. J. 1990, 22, 355-360 lshihara, K., Aragaki, R.. Ueda,T., Watanabe, A~and Nakabayashi,N., Reduced thrombogenicity of polymers having phospholipid polar group, J. Biomed. Mater. Res. 1990, 24, 1069 ishihara, K., Ueda.T. and Nakabayashi,N., Drugreleasefrom hydroget membrane having phospholipid structure, Koubunshi Ronbunshu 1989, 46. 591-595 Ishihara, K., Aragaki, R, Yamazaki,J., Ueda, T., Watanabe, A. and Nakabayashi, N., Organized adsorption of phospholipid on the polymer surface with phosphotipid polar group and its blood compatibility, Seitai Zain/o 1990, 8, 231-237 Kojima, M., Ishihara, K., Watanabe, A, and Nakabayashi, N., New biocompatible polymers with phosphotipid moiety, Preprints of IUPAC International Symposium 26 June 1989, SeouI, South Korea, p. 330 Kojima, M., Ishihara, K., Watanabe, A. and Nakabayashi, N., Biocompatibility of biomembrane-like surface composed of block copolymer containing phosphorytcholinemoieties, Polymer Preprints, Jpn. 1990, 39, 984 ishihara, K., Kojima, M., Watanabe, A. and Nakabayashi, N., Biocompatibilityof the surfaceof phosphotipidpolymers,Preprints of the 33rd IUPAC International Congress on Macromolecutes 11 July 1990, Montreal, Canada,Session 3.4.6
The Biomaterials Silver Jubilee Compendium
Quantitative assessment of the tissue response to implanted biomaterials D. Geoffrey Vince, John A. Hunt and David F. Williams
Institute of Medical and Dental Bioengineering, University of Liverpool, PO Box 14 7, Liverpool L69 3BX, UK (Received 25 August 1990; revised 20 November 1990; accepted 20 December 1990)
The tissue response to a small number of polymeric biomaterials was studied using monoclonal antibodies specific for certain inflammatory cell types, to develop a reliable and accurate method for the quantitative evaluation of biocompatibility. The sites of antibody binding were identifed using an avidin-biotin staining procedure and the sections evaluated using a computer-aided image analysis system. The staining technique successfully demonstrated both polymorphonuclear leucocytes and macrophages in tissue samples containing polymeric biomaterials. The image analysis system facilitated the measurement of up to 30 cell-related parameters and allowed a large number of cells to be analysed. Keywords: Image analysis, cell-material interactions, biocompatibitity
The concept of biocompatibility is based on the interactions between a material and a biological environment. The failure of a biomaterial, in a clinical situation, to display good biocompatibility is often revealed by a breakdown in the desired material properties or an unsatisfactory biological response. The most important aspect of biocompatibility, for the performance of the material, is the local tissue response, as this usually provides a clinical indication of a biocompatibility deficiency. Analysis of the local tissue response to a biomaterial has long been recognized to play an important r61e in biocompatibility testing. The normal wound-healing response is a dynamic phenomenon, in which cells and their products interact to repair damaged tissue. If an implant is present in the tissue, this sequence of events is disrupted to varying degrees, resulting in a visible change in tissue morphology. Many types of cell are involved in normal wound healing, including macrophages and polymorphonuclear leucocytes (PMNs). Wound healing in the presence of a biomaterial may induce a more complex reaction involving lymphoid and myeloid cells. Traditional staining techniques, such as haematoxylin/ eosin, van Giesen, or Periodic Acid-Schiff, are often used in the classification of inflammatory cell types. These cells, which include macrophages, PMNs, T-lymphocytes and B-lymphocytes, can then be identified by morphological characteristics. In the assessment of the local tissue response to implanted biomateriats, the observed distribution of these cells may be used to provide a qualitative description of the reaction. Attempts to provide a more quantitative Correspondenceto ProfessorD.F. Williams.
assessment have involved cell counting, using morphological criteria, and the allocation of a grading of the responseS, usually on a scale of 1 to 5. This method of evaluation relies on the subjective assessment by the operator, leading to errors such as misidentification of cell types or miscounting of cells. In an attempt to overcome some of these problems and to quantify accurately the complex interactions between cells and the surrounding tissue, a number of staining methods have been investigated and used in conjunction with a computer-aided image analysis system.
MATERIALS AND METHODS To study the biocompatibility of polymeric biomateriats, a series of polymers were implanted bilaterally into the dorsolumbar musculature of black and white hooded Lister rats. Two animals per material were used for each time period, with each rat implanted with two pieces of the same material. The materials used in this study comprised poly(glycolic acid) (PGA), (polylactic acid) (PLA), 50% PLA/ 50% PGA copolymer (Medisorb, DuPont, USA) and Biomer% After periods of up to 3 months, the rats were killed by cervical dislocation, and the tissue surrounding the implant carefully removed. The tissue, with the implant in situ, was frozen using isopentane and dry ice, and sectioned at 7 p m using a cryostat microtome. Each tissue block was trimmed down to the edge of the implant and serial sections taken for staining. Two staining methods were used. The first of these is a rapid enzymic technique to detect chloroacetate esterase. This enzyme is found almost exclusively in PMNs and mast
91991 Butterworth-HeinemannLtd. 0142-9612/91/080731-06 Biomaterials 1991, Vol 12 October
731
74
The Biomaterials Silver Jubilee Compendium Tissue response to biomaterials: D.G. Vince et aL
cells. These cells can be detected by incubating tissue sections with naphthol AS-D chloroacetate in the presence of freshly formed diazonium salt. Enzymic hydrolysis of ester linkages liberates free naphthol compounds which couple with the diazonium salt, forming highly cotoured deposits at sites of enzyme activity. This method can be carried out quickly and conveniently using a commercial test kit (Sigma, Code 91-A). Minor modifications were made to the method sheet provided. Fixation of the sections before staining was by using a citrate-acetone-methanol solution, followed by thorough rinsing. The citrate-acetone-formaldehyde fixative recommended for use with the kit was found to suppress staining of PMNs. A developing solution was prepared as follows: sodium nitrate solution (50pl), fast red violet LB base solution (50 pl), distilled water at 37~ (4 ml), trizmal buffer pH 6.3 (250 pl) and naphthol AS-D chloroacetate solution (50 pl). The sodium nitrate solution and fast red violet LB base solution were mixed gently by inversion and allowed to stand for 2 min. The distilled water, trizmal buffer and naphthol AS-D chloroacetate solution were then added to the mixture. This was applied directly to the slides and incubated for 60 min at 37~ protected from the light. Finally, the slides were rinsed in distilled water for 2 min. The slides were than coverslipped using aqueous mounting media before undergoing image analysis. If morphological examination was required, haematoxylin counter staining was performed before mounting (Figure 1). Identification of the other inflammatory cells was performed using an immunohistochemical staining method. The avidin-biotin technique represents one of the most recent developments in immunoperoxidase staining and is based on the ability of the egg-white glycoprotein avidin to bind non-immunologically four molecules of the vitamin biotin. Three primary reagents are used in this technique. The first is a primary antibody specific for the antigen to be localized. M urine monoclonal antibodies towards specific rat surface antigens were obtained. The monoclonals selected were as follows.
ED2. This recognizes a membrane antigen present pre-
dominantly on tissue macrophages of the rat. Monocytes, dendritic cells, lymphocytes and granulocytes are negative for ED2.
CD8 type. This recognizes a determinant on the majority of thymocytes (90-95%), a subset of peripheral T-cells, and
the majority of N K-cells. The antigen recognized is a complex of surface glycoproteins and is the rat homologue of the human CD8 antigen and the mouse Ly2. The antibody labels a T-subset which mediates suppression of antibody formation and the cytotoxic cell precursor.
IL2 receptor. This recognizes a gtycoprotein found on activated rat T-cells but not resting lymphocytes.
L-CA. This is a mouse monoclonal antibody to a subfraction
of rat B-cell leucocyte common antigen (L-CA). This monoclonal binds only to B-cells.
W3/13 HLI~ This displays specificity towards PMN, all thymocytes and T-lymphocytes and haemopoietic stem cells, but not B-lymphocytes. The ED2, CD8 type and the IL2 receptor monoclonal antibodies were obtained from Serotec and used at 1 : 100 dilution, with the exception of CD8 type which was used at 1:40. These dilutions were decided on the basis of titration data obtained from the suppliers. The L-CA and the W3/1 3 HLK were purchased from Seralab; these two antibodies were supplied as supernatant and therefore used neat (Table 1).
The second reagent used is antimouse polyclonal antibody (Dako, Code E413) and is covalently linked to one molecule of biotin. The third reagent is a complex of peroxidase conjugated biotin and avidin (Dako, Code K355). The free sites on the avidin molecule allow binding to the biotin on the second antibody. The peroxidase enzyme, and therefore the original antigen, are identified with an appropriate chromogen (Figure 2). The technique was performed as outlined below. Sections were cut at 7 p m and fixed in acetone for 10 min before air drying. To destroy any endogenous peroxidase activity, the slides were incubated for 20 min in 0.6% hydrogen peroxide in methanol. Following rinsing in distilled water, the slides were placed in a bath of phosphate buffered saline (PBS) for 10 rain. The sections were then dried by inverting on to paper towels and incubated for 25 rain with rabbit serum diluted 1:5 with PBS. This stage of the staining procedure blocks any rodent cross-reactivity antigenic sites. It is therefore important that immediately after this stage the slides are not rinsed but dried again by inversion on to paper towels. The serial sections were then incubated for 90 min with the monoclonat antibodies listed in Table 1, before rinsing in PBS for 5 min and drying by inversion. The rabbit antimouse polyclonal antibody was diluted to 1 : 100 and two drops of rat serum added to remove any antibodies which may cross-react with rat antigens. This solution was applied to the sections and incubated Table I
Figure I Chloroacetate esterase stain showing mast cell in counterstained muscle tissue.
732
Biomaterials 1991, Vol 12 October
Antigenic specificity and dilution of monoclonal antibodies
Antigen
Specificity
Form
Dilution
ED2 CD8 type
Macrophage T-lymphocytes (cytotoxic/ suppressor cells) Activated T-lymphocytes B-lymphocytes
Ascities Ascities
1 : 100 of 2-10 mg/ml 1:40 of 3.39 mg/ml
Ascities
1 : 100 of 3.80 mg/ml
Supernatant
Neat at 5-10 mg/ml
Supernatant
Neat at 5-10 mg/ml
Interleukin 2 receptors Leucocyte common antigen subfraction T-lymphocytes W 3 / 1 3 HLK
The Biomaterials Silver Jubilee Compendium
75
Tissue response to biomaterials: DG. Vince et al.
with these methods. Although the number of lymphocytes in the tissue was expected to be much lower than that found in the separations, they were included in the positive controls as they demonstrate the ability of this technique to identify these cell types. The avidin-biotin indirect immunoperoxidase method gave clear staining of sections of rat muscle with all the monoclonal antibodies described. An example is shown in Figure 3. The use of 0.6% H202, rat and rabbit serum reduced the background staining, thus increasing the sensitivity of the computer-aided image analysis system. All cell types were present in the tissue sections, although with these materials, and at the stated time periods, only macrophages and PMNs were present in large numbers ( F i g u r e s 3 and 4). A few lymphocytes were detected in the sections, although not in sufficiently large numbers to present reliable analyses in these experiments. In the data presented in this section, cell numbers are presented as a function of distance from the implant surface and time. The number of cells refers to those cells counted analysed over a 20 frame area, corresponding to a defined location within the section. The tissue response to Biomer produced the maximum number of PMNs (539 cells) by 2 d. The cell number then dropped sharply to 21 cells by 7 d. By 3 months, there were no PMNs in the tissue adjacent to the implant (Figure 5). In contrast, only 1 61 macrophages were present at 2 d, with the maximum number of cells found at 7 d (774 cells). At 1 4 d, 1 01 macrophages were present in tissue, with the Figure 2 Diagrammatic representation of the avidin-biotin-peroxidase staining technique
for 40 min. The slides were rinsed in PBS for 5 min before applying the avidin-biotin complex, incubated for 30 min and rinsed in PBS for 5 min. Sites of peroxidase activity were identified by incubating with a solution of diaminobenzidine (DAB). This was prepared by dissolving 10 mg of DAB in 1 5 ml of PBS. To 4 ml of this solution, 60 p l of 3% hydrogen peroxide was added. This was applied to the sections and incubated for 1 5 min. The slides were then rinsed in distilled water for at least 2 min before dehydrating through alcohol and mounting in DPX. Positive controls for the stains comprised samples of spleen, liver, blood and muscle tissue surrounding a range of copper/polymer materials. Ex vivo positive controls were also performed, in which rat lymphocytes were separated from whole blood and stained using the panel of antibodies.
Histomorphometry
Figure 3 Staining techmque.
of
macrophages
using
avtdm-biotm-peroxidase
The image analysis system used in this study was a JoyceLoebl mini-Magiscan. The slides were viewed under a Zeiss Jenaval photomicroscope and the image captured using a Hitachi KP 1 40 CCD monochrome video camera. The JoyceLoebl software contains menu-driven routines which can be linked to create task lists. When executed, these routines allow fully automated tissue analysis. A task routine was created that would analyse 20 fields of view per slide, to a distance of 6 4 0 p m from the implant, providing a total analysed area of 2.06 • 106 pm 2. A number of parameters were measured including cell number, cell area, distance from implant and circularity.
RESULTS The positive controls, involving either muscle tissue or lymphocyte separations, demonstrated successful staining
Figure 4 Staining of PMNs using the chloroacetate esterase staining technique.
Biomaterials 1991, Vo112 October
733
The Biomaterials Silver Jubilee Compendium
76 Tissue response to biomateriats." D. G Vince et af.
Figure 5 The distribution of PMNs surrounding implants of Biomer ~. Period: 1 = 2 d; 2 = 7 d; 3 = 14d.
Figure8 The distribution of macrophages surrounding implants copolymer. Period: 1 = 2 d; 2 = 7 d; 3 = 14 d; 4 = 3 months.
of
of 381 cells at 7 d. By 14 d, the number of macrophages had fallen to 319, and to 75 by 3 months (Figure 8). The PGA gave a similar response to the copolymer, with the number of macrophages reaching a peak of 963 cells at 7 d. The total number of macrophages decreased to 207 by day 14; data are not yet available for the 3 month time period (Figure 9). As expected, the number of PMNs reached a maximum at 2 d (690 cells), falling to onty 55 cells by 7 d. By day 14, the number of PMNs had increased to 207 cells (Figure 10). The samples of PLA produced the greatest number
Figure 6 The distribution of macrophages surrounding implants of Biomef e. Period: I = 2 d; 2 = 7 d; 3 = 14 d; 4 = 3 months.
majority occurring within 4 0 p m from the implant/tissue interface. By 3 months, the macrophages were evenly distributed throughout the section, with only 152 cells present in the tissue (Figure 6). The PMN response to the degradable PLA/PGA copolymer was quite different, with the number of cells increasing from 210 at 2 d to 427 at 14 d. As with the Biomer samples, no cells were found at the 3 month interval (Figure 7). The copolymer gave a similar response to Biomer with 327 macrophages present at 2 d, reaching a maximum
Figure 7 The distribution of PMNs surrounding implants of copolymer. Period." 1 = 2 d; 2 = 7 d; 3 = 14d.
734
Biomaterials 1991, Vol 12 October
Figure 9 The distribution of PMNs surrounding implants of PG~ Period: I =2d;2= 7 d ; 3 = 14d.
Figure t 0 The distribution of macrophages surrounding imp~ants of PG,~ Period: I = 2 d; 2 = 7 d ; 3 = 14 d.
The Biomaterials Silver Jubilee Compendium
77
Tissue response to biomaterials. D,G. Vince et al.
Figure I 1 The distribution of macrophages surrounding implants of PLA Period." 1 = 2 d; 2 = 7 d; 3 = 14 d; 4 = 3 months.
Figure 12 The distribution o f P M N s surrounding implants of PLA. Period: 1 = 2d; 2 = 7d; 3 = 3months,
of macrophages after 2 d (488 cells). There were 340 macrophages present in the tissue at the 7 d period, with the majority of the cells occurring within 8 0 p m from the implant. After 14 d, 346 macrophages were identified, with the greatest number of cells occurring between 51 2 and 624 pm. By 3 months, only 1 80 macrophages were present and those were evenly distributed within the tissue (Figure 11). The maximum number of PMNs occurred within 2 d (533 cells), and were evenly distributed throughout the tissue. By 7 d, the cell number had fallen to 2 50, with the majority of cells occurring within 64 p m of the implant
(Figure 12).
DISCUSSION There are many immunoenzymic staining methods which can be used to localize antigens. The selection of the appropriate technique depends on the individual requirements of the operator. The avidin-biotin technique utilizes a primary mouse antirat monoclonal antibody, a biotinytated polyclonal rabbit antimouse secondary antibody, preformed avidin-biotin complex and horseradish peroxidase (HRP)is used as an enzyme label. Biotinylation is a process whereby biotin is covalently attached to the antibody. Free sites on the avidin molecule from the avidin-biotin/H RP complex bind to biotin on the secondary antibody, with DAB used to identify sites of enzyme activity. The strong affinity of avidin for biotin and the mild biotinylation process makes the avidin-biotin method more sensitive than direct immunoenzymic tech-
niques. This method has also been shown to be more sensitive than other indirect techniques such as peroxidaseantiperoxidase (PAP)2. When this staining technique is used in conjunction with the image analysis system, we have a very powerful tool for the quantitative assessment of tissue inflammation and wound healing. The process of wound healing consists of a series of events, during which cells interact with each other and produce a new extracellular matrix. If a foreign material is placed in the wound, an interaction may occur between the material surface and the tissue, which may affect this sequence of events. The stimulus to the tissue may be small if the material is inert, non-toxic and static, and any deviation from normal wound healing may then only be due to the physical presence of the implant. Alternatively, the material may constitute a significant degree of irritation which could severely disrupt the normal wound-healing process and may result in a different tissue response. As the majority of biomaterials require surgical implantation, it is convenient and relevant to consider biocompatibility in the light of the mechanisms of the normal wound-healing response, and the influence the presence of an implanted biomaterial has on this process. The tissue reaction towards Biomer produced a similar response to normal wound healing. The PM Ns were the first cell type to appear after surgery, with the majority of cells found close to the implant surface. The r61e of the PMNs is largely phagocytic, although they are a significant source of platetet activating factor 3. PMNs survive in vivo for approximately 1 d after which they are engulfed by macrophages. This would explain the greatly reduced number of these cells observed at 7 d, 14 d and 3 months. The slower-moving macrophages do not reach a maximum number until day 7 and decrease sharply to 1 4 d. Macrophages play several major r6tes in the inflammatory response. They may assist in controlling acute inflammation by their secretion of monokines such as interleukin 14, and by the release of metabotitesS. Secondly, macrophages govern any ongoing inflammatory response by regulation of both T- and B-lymphocytes, and fibroblasts 6. Finally, macrophages phagocytose cellular and molecular debris, and also detoxify and/or sequester toxic materials. The sample of PGA invoked a greater tissue response than the Biomer, with a greater number of both macrophages and PMNs present in the sections. As with the Biomer, the maximum number of cells occurred on day 7. Without the data for the 3 month time period, the extent of any chronic inflammatory response may be difficult to judge, although at 1 4 d, the smallest number of macrophages were present,. As expected, the greatest number of PMNs occurred at day 2. The tissue response to PLA had some similarities with that of the Biomer, with the maximum number of PMNs occurring on day 2, falling to no cells present at 3 months. The macrophage response gave the maximum number of cells at day 2, although the maximum count occurred on day 7 (43 cells at 64 pm). By 3 months, the total cell number had fallen to its lowest value. in a consecutively running project investigating the in vitro performance of polymeric materials, the molecular weights of the PLA, PGA and copolymer were measured over specific time periods. These were found to fall considerably over a period of 14 d, when placed in distilled water. The materials appeared to hydrolyse, with no concomitant release of particulate debris. Since a lowering of pH was noted, it can be assumed that lactic acid and glycotic acid are the breakdown products. As these two compounds are
Biomateria/s 1991, Voi 12 October
735
The Biomaterials Silver Jubilee Compendium
78 Tissue response to biomateria/s: D.G Vince et aL
soluble there will be no physical stimulation of the cells. This may explain the normal wound-healing response observed with these materials. The only material which did not follow this pattern was the copolymer, with the PMN rising at the 14 d interval before falling to undetectable levels at 3 months. This finding is as yet unexplained. This method of tissue evaluation has many advantages over the more commonly used staining/analysis procedures. One problem frequently encountered in the assessment of the tissue response to a biomaterial is that of cell identification. Morphological characteristics such as shape, size and nuclear area can be unreliable parameters, as these are dependant upon the state of activation and the motility of the cell. The immunostaining procedure overcomes this identification problem by using monoclonal antibodies that bind to surface antigens which are specific for the cell to be identified. The image analyser, when used in conjunction with the staining techniques, allows up to 30 celt-related parameters to be measured. Many of these, such as cell area, distribution and circularity are very difficult and timeconsuming to determine manually. Finally, as these cell measurements are fully automated, a large number of cells can be analysed without counting errors being introduced. This regime of histochemical staining and image analysis is a great improvement on the more routinely used tissue assessment techniques, in that small changes in the tissue response towards a material, or a group of materials, can be quantitatively assessed.
( 1 ) Once the cells have been identified, up to 30 parameters such as cell size and distribution can be measured. (2) Errors such as misidentification of cell types will not occur, due to the high specificity of the mo~oclonal staining technique. (3) As cell measurements are fully automated, a large number of cells can be analysed without counting errors being introduced.
ACKNOWLEDGEMENTS This staining regime has been developed under contract to the Laboratory of the Government Chemist, UK, with the objective of facilitating the quantitative assessment of biocompatibility and biosafety. The authors wish to acknowledge Peter Johnson, Professor of Immunology at the University of Liverpool, and the financial support of the Science and Engineering Research Council of the United Kingdom for the research studentship for John A. Hunt.
REFERENCES ! 2 3
CONCLUSIONS The combination of histochemistry and image analysis greatly aids the quantitative assessment of this tissue response to biomaterials. This method of evaluation has many advantages over more commonly used procedures in that:
736
Biomaterials I991, Vo/ 12 October
4 5
6
Iskandar, S.S., Emancipator, S.N. and Pretiow, T.G., Enzyme histochemistry of monocytes/macrophages, J. Histochem. Cytochem. 1989, 37, 25-29 Hsu, S-M., Raine, L. and Fanger, H., Use of avidin-biotin-peroxidase complex in immunoperoxidase techniques, J. Histochem. Cytochem. 198t, 29, 577-580 Jouvin-Marche, E., Cerrina, J., Coeffier, E., Duroux, P. and Benveniste, J., The effect of the Ca2+ antagonist nifedipine on the release of platetet activating factor, stow reacting substance and ~-gtucuronidase from human neutrophiis, Eur. J. PharmacoL 1983, 89, 19-26 Oppenheim, J.J. and Gery, I., Interleukin 1 is more than an interleukin, Immunology Today 1982, 3, 113-119 Dawson, W., Boot, J.R., Walker, J.R. and Meade, C.J., The arachidonic metabolites, in Textbook of Immunopharmocology (Eds M.M. Dale and J.C, Foreman),Black Scientific Publications,Oxford, UK, pp 126-139 Dinarello, C.A., Interleukin 1, Rev. Infect. Dis. 1984. 6, 51-95
The Biomaterials Silver Jubilee Compendium
79
731
REVIEW Immune response in biocompatibility A. Remes and D.F. Williams
The Department of Clinical Engineering, University of Liverpool, PO Box 147, Liverpool L69 3BX, UK Biocompatibility is concerned with the interactions that occur between biomaterials and host tissues. As foreign objects in that host tissue these materials may initiate several types of response. It has often been postulated that the immune response, by which the host normally defends itself against invasion by foreign organisms, can be involved in the response to biomaterials. This review discusses the mechanisms by which this could occur and the evidence that suggests the immune response is indeed of significance in biocompatibility. Keywords: Review, biocompatibility, immune response Received 15 May 1991; accepted 16 December 1991
Biocompatibility has been recently defined in terms of 'the ability of a material to perform with an appropriate host response in a specific application'. This host response is clearly the principal issue in the complex set of phenomena that comprise biocompatibility and there is a vast amount of literature that describes and characterizes features of this response. A question raised on many occasions but not yet clearly answered in this respect is the role of the immune system in this response. This review attempts to address this question. Immunology is the study of the mechanisms by which a host defends itself against invasion by foreign organisms. Biomaterials, clearly, are not organisms, either living or non-living, and so the host response to biomaterials is not obviously related to these mechanisms of defence. However, these mechanisms themselves involve several different components and processes which can interact with biomaterials and which may be involved in this host response. The protective mechanisms of the body may be divided into two categories -- specific and non-specific. The processes in these groups are quite different although they may be influenced and mediated by the same agents. Cells such as macrophages play an important role in both specific and non-specific responses, as described later. The non-specific immune system is present in all individuals and operates on different foreign agents (e.g. bacteria, fungi) in the same way. It does not Correspondence to Professor D. F. Williams. 91992 Butterworth-HeinemannLtd 0142-9612/92/110731-13
distinguish between different infectious agents and does not alter intensity upon re-exposure. Examples include stomach acid, lysozyme in tears, the alternative pathway of complement activation and phagocytic cells (cells which engulf particles), e.g. neutrophils and macrophages. The initial requirement for antimicrobial activity, the synthesis and secretion of enzymes and high-energy oxygen radicals by phagocytic cells, involves recognition of the foreign object. Phagocytes may recognize a foreign intruder, either because it has a physiologically abnormal surface ~ or because it is covered in opsonins 2. 'Opsonin' is a term given to both the molecules C3b (produced when the complement cascade is activated] and IgG {an antibody produced by B cells). These molecules can bind to both the phagocyte and the foreign surface, greatly increasing the efficiency of phagocytosis 3. Other molecules may attract phagocytic cells to the inflammatory site, such as C5a which is also produced when complement is activated. Such products may modify phagocyte locomotion in two ways, by chemotaxis and chemokinesis 4. Chemotaxis is the reaction by which the orientated and directional movement of cells or organisms is determined by substances in their environment. Chemokinesis is the reaction by which the frequency or speed of random locomotion {orthokinesis} and/or the frequency and magnitude of turning (klinokinesis} is influenced by substances in their environment. The specific immune response is not actively present in all individuals. It requires stimulation or immunization in order to be activated and is mediated by white blood cells (lymphocytes}. Each lymphocyte has receptors on Biomaterials 1992, Vol. 13 No. 11
80
The Biomaterials Silver Jubilee Compendium 732
Immune response in biocompatibility: A. Remes and D.F. Wi//iarns
its surface which will recognize only a single structural conformation, termed an antigenic determinant. On recognition of a specific antigen, a lymphocyte will reproduce forming clones of itself. Thus, on reintroduction of the antigen, the host response is larger and is accelerated. Lymphocytes may be subdivided into T cells and B cells. T cells are responsible for the cell-mediated immune response and B cells for the humoral immune response which is mediated by antibodies. Antibodies {ab) are proteins produced by B cells in response to a foreign agent that reacts specifically with that agent. Activation of the classical complement cascade requires antibodies [IgG or IgM). Thus complement activation by the classical pathway is related to the specific immune response. The cell-mediated immune response is mediated by T cells. T cells differentiate in the thymus and develop into three distinct types, whose phenotype may be differentiated by antigens on their surface {Table 1}. Delayedtype hypersensitivity cells {TDTH}are responsible for the attraction and activation of macrophages. The major mechanism of tissue damage at the inflammatory site is phagocytosis and destruction of cells by macrophages. If macrophages, activated by antibody or lymphocytes, are unable to digest the material they phagocytose, masses of macrophages collect in the tissue forming a granuloma. This is an attempt to wall off the material from the rest of the body. Other T cells which may be present at the inflammatory site include suppressor cells (Ts) which suppress the immune response and cytotoxic cells {Tcyt) which act by ]ysing the target cells for which they are specific. For T cells to be stimulated by a foreign agent, the agent has to be presented to them, on the surface of an accessory cell (e.g. activated macrophage} along with a self-antigen called a histocompatibility antigen. A histocompatibility antigen has been described as 'a genetically determined isoantigen carried on the surface of nucleated cells of many tissues which, when tissue is grafted onto another individual of the same species whose tissues do not carry that antigen, may incite an immune response which leads to graft rejection 's. The most closely studied histocompatibility antigens are those of the H-2 system in mice and of the HL-A system in man. These systems are located on a single chromosome, which codes for both class I and class II major histocompatibility complex (MHC} antigens. Class I antigens are present on all nucleated cells. Class II antigens are found only in cells involved in the immune response, with the exception of polymorphonuclear leukocytes and sperm.
Table1
Lymphocyte surface antigens - -
m
-
Surface antigen
T cell type
Human, T4; mice, Ly1+23-
T cells responsible for the delayed-type hyper-sensitivity reaction (ToTH) Human, TS; mice, LY1-23 + (a) Cytotoxic T cells, which may kill neoplastic cells and virus infected cells (b) Suppressor T cells, which suppress the immunological response .
.
.
.
.
.
.
Biomaterials 1992, Vol. 13 No. 11
,
Zinkernagel and Doherty 6 showed that cytotoxic T lymphocytes would only kill virus-infected cells which have the same MHC antigens on their surface as they themselves have. This phenomenon is called MHC restriction and is exhibited by T cells but not B cells. MHC restriction allows the body to differentiate between self and non-self. T lymphocytes with the phenotype Lyl +23- {usually TDTH cells] recognize antigen presented with class II MHC, and cells with the phenotype Ly1-23 + {usually Tcyt cells) recognize antigen presented with class I MHC. T cells recognize, and are activated by, antigens and MHC presented together on the surface of accessory cells such as activated macrophages 7 (which have a higher concentration of class II MHC on their surface than nonactivated macrophages; see Table 2 for other properties of activated macrophages}, dendritic cells 8 and B cells 9. Activated macrophages and B cells are capable of digesting large complex antigens, e.8. Listeria monocytogenes 7, lo. The digested antigen fragments can then either combine with MHC for presentation on the accessory cell surface or may be released into the surroundings 11 where they may bind to MHC antigen on the surface of other accessory cells such as dendritic cells. Activation of T~TH cells requires two signals 12. One signal is recognition of antigen in the context of class II MHC on the surface of an antigen presenting cell. The second signal is IL-1. IL-1 is produced by activated rnacrophages and dendritic cells on interaction with T cells. IL-1 stimulates two TDTHcell second populations. One population secretes the T cell growth factor IL-2 and a second population has receptors for IL-2. The T cells divide, develop into large lymphoblast cells and finally into smaller effector T cells. On activation TDTH cells secrete soluble mediators which are collectively termed lymphokines. As well as IL-2, these include factors which activate and attract macrophages, B cell growth factors and B cell differentTable2
Macrophage activation .
.
.
.
.
.
.
.
.
Inducing agents
Biochemical changes (increases in)
Bacteria e.g. listeria Protozoa e.g. leishmania trypanosome Lipopolysaccharide
Plasminogen activator H202
o'~
Lysosomal enzymes collagenase elastase interferons (a,/~) Macrophage activation factor Oxygen consumption Purified protein derivative from Unchanged: Mycobacterium tuberculosis lysozyme complement components
Immunological changes (increases in)
Functions (increases in)
Morphological changes (increases in)
Fc receptors C3b receptors release of IL-1
Microbicidal activity Phagocytic activity
Granule size Rough ER DNA + Protein synthesis Spreading
81
The Biomaterials Silver Jubilee Compendium Immune response in biocompatibility: A. Remes and D.F. Williams
733
4. Immune cells are often noted at the inflammatory site. In the case of calcium phosphate ceramics, macrophages are important in the resorption process, where the ceramic implant may be replaced by bone TM. 5. Studies which used image analysis to examine the local tissue response to polymer implants quantitatively (poly(lactic acid), poly(glycolic acid], 50% PLA/50% PGA copolymer, Biomer) showed that the macrophage and neutrophil responses differ for different implants 2~
iation factor 13. Thus the humoral immune response is also dependent on activation of TDTH cells. Unlike TDTH cells, Tcyt cells recognize class I MHC and have a different mechanism of activation. Toyt cells require interaction of cells bearing antigen in context of class I MHC and also IL-2 derived from TDTH cells. Thus, TDT H cells must be activated in order for Tc~ cells to be activated.
BIOMATERIALS AND IMMUNE RESPONSE
This article examines the immunological phenomena of chemotaxis, neutrophil activation and complement activation to determine how these processes may influence biocompatibility. These phenomena are not mutually exclusive. When considering the role of the immune response in biocompatibility one must consider the influence that the biomaterial has on each of the above phenomena and the influence that these phenomena have on each other (Figure 1). For example, when complement is activated the chemotactic factor C5a which is released also stimulates an oxygen burst in neutrophils 21 and enzyme degranulation 22.
As a biomaterial is foreign to the host, it is not surprising that the immune response has been suggested as being an important factor in biocompatibility. Recent studies have shown that" 1. The complement pathway is activated during blood flow through extracorporeal circuits, e.g. kidney dialysis 14 and cardiopulmonary bypass 15. This results in neutrophil aggregation leading to pulmonary dysfunction. 2. Sensitivity to metals may be involved in the loosening of metallic implants 16' 1~. The metal ion as a hapten may bind to a protein carrier. This hapten-carrier complex may trigger a cell-mediated immune response. 3. Polymer surfaces have been shown to stimulate IL-1 release from monocytes TM. IL-1 is important for activation of T cells and growth and differentiation of B cells.
CHEMOTAXIS The locomotion of neutrophils is a critical determinant of their ability to fulfil their role in the first line of defence
A t t r a c t more p h a g o c y t e s Tissue damage Enzyme secretion l
Stimulates I L-I release from macrophages
~
/
Neutrophil activation
T
Platelet aggregation Complement activation A
C3b ~ /
IBiomaterial ]
C3a
Complement activation
~-C3a + C5a + C567
+
C5a
I Anaphylatoxins ~i,,.
+ secretion of high-energy oxygen radicals
cell
Histamine PAF Neutrophil chemotactic factor Prostaglandins etc. SRS-A
Anaphylations+Mast cell mediators cause~in vasodilation + ~ in vascular permeability
Neutrophils and monocytes attracted towards the sou rce of chemotactic factors
T
N N N N N
N
N
N N N N N
Figure I Biological activities of activated complement components. N, neutrophils; PAF, platelet-activating factor; SRS-A, slowreacting substance of anaphylaxis. Biomaterials 1992, Vol. 13 No. 11
82
The Biomaterials Silver Jubilee Compendium 734
Immune response in biocompatibility: A. Remes and D.F. Williams
against foreign pathogens. There may be many reasons for the accumulation of cells at the inflammatory site including contact guidance 23, increased cell stickiness 24, cell preference for hydrophobic rather than hydrophilic surfaces, rugophilia z5 [the accumulation of cells on rough rather than smooth surfaces), inhibition of migration of cells away from the inflammatory site by lymphokines 2e and chemotaxis 27. Chemotaxis is the most commonly studied, not because any of the other phenomena are less important, but because chemotaxis is more easily examined. Biomaterials may influence chemotaxis directly or through interaction of degradation products with cells. Cobalt, chromium and nickel, common components of metallic alloys, are released as ions or particles into tissue next to the implant site 28. Nickel ions, but not cobalt or chromium ions, were found to inhibit chemotaxis to FMLP in a chemotaxis under agarose assay 2g. A recent study has shown that nickel ions block the polymerization of actin in FMLP-activated cells and that this is responsible for causing the inhibition of chemotaxis to FMLP 3~ This may be of importance with regard to infection which may not be controlled if phagocytic cells are inhibited from migrating towards the site. Infection has frequently been suggested to be a cause of implant failure 3~. The neutrophil polarization assay for detecting chemotaxis and the chemotaxis under agarose assay indicate that these three metal ions are not chemotactic for human neutrophils in the absence of s e r u m 29. However, there is much evidence to suggest that metal ions exist in serum as organometallic complexes 32. In the case of nickel, van Soestbergen et al. 33, found that 24 h after injection of radioactive nickel chloride into rabbits, an average of 90% of activity was bound to albumin and 10% was ultrafilterable. The ultrafilterable fraction of serum consisted primarily of nickel complexes with serum proteins. Further studies should examine the chemotactic activity of metal ion-protein complexes. At high concentrations of cobalt, in the absence of serum, a cobalt phosphate precipitate caused neutrophils to aggregate, inhibited chemotaxis under agarose and caused neutrophils to release lysozyme 29' 34. This may be a result of the toxic effect which the cobalt phosphate precipitate has for neutrophils and may be of significance in biocompatibility by causing tissue damage. McNamara and Williams 35 noted a cobalt phosphate precipitate on the surface of cobalt discs implanted intramuscularly into rats. The cobalt phosphate precipitate may have contributed to the inflammatory response that was observed around the cobalt disc implants. When interpreting results from experiments involving metal ions, one must consider that metal ions may exist in different valency states. Cobalt and nickel exist in a single valency state. However, chromium released from implant alloys may be incorporated into organometallic complexes as CrIII or CrVI. CrVI is far more biologically active than CrIII. CrVI is considered to be the valency of chromium released from a metallic implant in viva. Chromium with a valence of 3+ was found almost exclusively in serum whereas chromium with a valence of 6+ was found strongly bound to red and white cells 36. Chromium ions from corrosion products were found strongly bound to cells. These results indicate that Cr6§ is being released from implants during corrosion in vivo. Biomaterials 1992, Vol. 13 No. 11
Studies which examined the biological response to chromium with a valence of 3+ may not therefore be relevant to the in viva situation. In the case of metal particles, cobalt and chromium powders in the presence and absence of serum, and silver and nickel powders in the absence of serum, did not stimulate neutrophils to polarize. It was concluded that, under these experimental conditions, these metal powders are not chemotactic 37. However, serum incubated with nickel and silver powder stimulated neutrophils to polarize 3~, suggesting that these metal powders alter serum perhaps rendering it chemotactic. This may be due to the metal powders denaturing serum proteins. Denatured proteins have been reported to be chemotactic 38' 39 and will therefore stimulate cells to polarize. Protein denaturation may be one of the factors involved in attracting cells to the inflammatory site around an implant. Cobalt and nickel discs cause an inflammatory response when implanted intramuscularly into rats. Cellular migration to the inflammatory site may at least in part result from denatured proteins stimulating chemotaxis. McNamara and Williams 35 have previously commented on deposits of protein, which appeared to be denatured on the surface of metal discs implanted intramuscularly into rats. Proteins may be denatured after adsorption on to biomaterial surfaces and/or if the biomaterial is toxic for tissue, resulting in death and lysis of cells. In 1905, Metchnikoff 4~ proposed that after death of phagocytic cells, proteolytic lysosomal enzymes escaped into the surrounding tissue. These proteolytic enzymes remain active in viva after the cells die 41 and can cause tissue damage. It has frequently been reported that cobalt and nickel, as particles and/or ions, are toxic for macrophages and fibroblasts, i.e.: 1. Cobalt and nickel (particle size 1 pro) caused damage to macrophages by 4 h of culture, reaching a maximum after 30 h 42. Lysed macrophages may release proteolytic enzymes into the tissue. These enzymes can degrade the tissue matrix, causing further damage to the tissue. Lysis of macrophages at an implant site may increase the risk of infection. 2. Cobalt, but not nickel particles (particle size l p m ) poisoned human synovial fibroblasts 43. Cell death was noted within one day of culture. 3. Cobalt and nickel {15-30 pg/ml] depressed the growth rate of mouse embryo fibroblasts 44. Nickel chloride at 100 and 200pg/ml decreased the viability of rabbit alveolar macrophages to 75 and 50%, respectively. The toxic effect that these metals have on the tissue, resulting from interaction of the metals with proteins and cells, may result in an inflammatory response, pH has been reported to be lowered at the inflammatory site 45. Acid pH has been reported to stimulate neutrophils to release a factor which is chemotactic for other neutrophils 48. An acidic pH in tissue adjacent to an implant may therefore amplify the chemotactic signal from the implant site. Indeed, the lowered pH of arthritic synovial fluid has been correlated with an increased white blood cell accumulation 47. However, other workers have reported that reduced pH impaired neutrophil migration in vitro 48
83
The Biomaterials Silver Jubilee Compendium 735
Immune response in biocompatibility: A. Remes and D.F. Williams .
.
.
.
.
.
.
.
.
_
.
_
Glucose+NADP+
.
.
.
.....
~ Pentose phosphate+ NADPH
Cytoplasm Azurophil granule containing myeloperoxidase NADPH
NADP + H +
Oxidase
02 ~ 021
+
~ 012
--
Myeloperoxidase + H202 + Halide
02 1
02 -
+
H202
-
Phagosome
O21
_-
H202
~
Toxic p r o d u c t e . g . hypochlorite ( H O C I ) =
OH*
+
OH
+
02
superoxide anion" v e r y toxic and v e r y unstable moiety which will react with many substances
OH*- h y d r o x y l radical- v e r y toxic and will kill bacteria r a p i d l y
Figure 2
Leukocyte respiratory burst.
Few studies have examined phagocyte activation by biomaterials. A study involving ceramic biomaterials noted that, in the absence of serum, CaHPO4 (Merck) powder and CaHPO4 [BDH) powder stimulated an oxygen burst in neutrophils indicating that the neutrophils had been activated 49. Neutrophils have also been shown to release superoxide anion after incubation with polyurethane 5~ (a commonly used elastomer) and Velcro pile (used in the Jarvik 7 heart). Biomaterials may also activate neutrophils indirectly by activating the complement cascade. The chemotactic complement activation product C5a is a potent activator of neutrophils. Metallic, ceramic and polymeric powder surfaces activate the complement cascade 5~'52. The neutrophil polarization assay for detecting complement activation 5~'52 suggests that chemotaxis is a secondary phenomenon which may follow complement activation by biomaterial powders. Lysosomal proteolytic enzymes and high-energy oxygen species that may be secreted when phagocytes are activated, or when phagocytes die, cause tissue necrosis (Figure 2). Denatured proteins which could result from tissue damage may be chemotactic. Constituents of lysosomes from rabbit and human neutrophils have also been reported to cleave C5 to yield fragments which have chemotactic activity and that also activate neutrophils 53-55. Lysosomes also contain cationic proteins with permeability-increasing activity, which w~en released will enhance neutrophil migration from blood vessels into tissue 4~. Prostaglandins and leukotrienes are synthesized in the cell membrane of activated phagocytic ceils 56. As well as regulating ]ysosomal enzyme release [both stimulating and inhibiting], these mediators can cause an increase in vascular permeability and leukotriene B4 is chemotactic for neutrophils and macrophages 5~. Results from a study which investigated the interaction of CaHPO4 (BDH) powder with neutrophils indicate that this powder may
stimulate neutrophils to secrete a factor which is chemotactic for other neutrophils 4Q.Other studies have reported that neutrophils can be stimulated to secrete a neutrophil chemotactic factor after incubation with immunoglobulins ~8, monosodium urate crystals 5~ and calcium pyrophosphate crystals 6~ NEUTROPHIL ACTIVATION Chemotactic agents may all help in attracting phagocytic cells to the implant site where these cells may be involved in the tissue response. Once phagocytes have reached the inflammatory site, high concentrations of chemotactic factor, i.e. C5a, may: (a] cause the cells to become more sticky 24, inhibiting them from leaving the site: or (b) stimulate an oxygen burst 61 (Figure 2] and enzyme secretion 62. This may cause tissue damage with resultant inflammation63.64. The significance of the release of lysosomal enzymes and high-energy oxygen radicals in response to implant materials remains to be established. However, opsonized zymosan which activates phagocytes in vitro cause a strong inflammatory response in vivo. There appears to be a link between the ability of various agents to release lysosomal enzymes from macrophage cultures in vitro and to elicit chronic inflammation in vivo 65, and most drugs that inhibit release of lysosomal enzymes (e.g. glucocorticoids, colchicine) also have anti-inflammatory activity B6. An inflammatory response may also result from tissue damage due to oxygen metabolites released from neutrophils into the environment. Toxic oxygen species released from C5a-activated neutrophils have been reported to injure the membranes of cultured human endothelial cells 67 and hydroxyl radicals which attack the neutrophil itself could be responsible for the death of these cells during phagocytosis and contribute to the inflammatory response by releasing hydrolytic enzymes ~8. Biomaterials 1992, Vol. 13 No. 11
84
The Biomaterials Silver Jubilee Compendium 736
Immune response in biocompatibility: A. Remes and D.F. Williams
Carp and Janoff 69 have reported functional inactivation of the antiproteinase, a-l-proteinase inhibitor by neutrophil-derived oxidants in vitro. The latter constitutes over 90% of the total trypsin-inhibiting capacity of normal plasma. Inactivation of this enzyme may render tissue susceptible to damage by lysosomal proteinases released into the microenvironment of activated neutrophils. CaHPO4 powder derived from Merck and BDH laboratories stimulated a respiratory burst in neutrophils in the absence of s e r u m 49. Coating these powders with serum prevented them from stimulating a respiratory burst. The relevance of an in vitro respiratory burst, in the absence of serum, to the in vivo situation may be regarded as being insignificant when one considers that protein adsorption to an implant surface occurs immediately after implantation of an implant. However, in that study, the serum-coated powders were not phagocytosed. In vivo, in the presence of the bloodclotting proteins and other immune cells, the neutrophils may be stimulated to phagocytose the ceramic powders. Once inside the cell, proteins on the powder surface may be digested by lysosomal enzymes exposing the surface of the ceramic powder which may then stimulate neutrophil activation. Stimulation of enzyme secretion from neutrophils by ceramic powders was not examined, but as many factors which stimulate a respiratory burst also stimulate enzyme degranulation, e.g. C 5 a 21' 22 and other crystals, e.g. hydroxyapatite TM,have been reported to stimulate enzyme secretion, it is possible that ceramic powders also stimulate neutrophils to secrete enzymes. In the case of resorbable ceramic implants, secretion of enzymes and oxygen radicals on to the surface of the implant may facilitate resorption. Lysosomal enzymes have been shown to provoke acute inflammation and tissue injury. Lysosomes contain a variety of enzymes, e.g. nucleases, phosphatases. Most of these have acid pH optima and probably cannot operate outside of the lysosome. The primary sources of enzyme activity from human polymorphonuclear leukocytes that are responsible for tissue degradation at physiological pH are the neutral proteases which have an optimal activity at neutral pH. The two best known are elastase and collagenase 41. Elastase is the only enzyme that can digest blood vessel walls and collagenase is an enzyme which initiates the degradation of collagen. Man but not rabbits has neutral proteases. Lysosomal enzymes from humans but not rabbits were found to digest vascular basement membrane at neutral pH 66. The mechanism controlling lysosomal enzyme release is dependent on the levels of the intracellular cyclic nucleotides cGMP and cAMP. Compounds which increase intracellular cGMP, e.g. cholinergic agonists, stimulate lysosomal enzyme release whilst compounds which increase intracellular cAMP, e.g. prostaglandin E, inhibit release. Cobalt, chromium and nickel ions from metal salts and from corrosion products of stainless steel have been reported to bind to white blood cells in vivo 32. These metal ions do not stimulate an oxygen burst in neutrophils and do not stimulate neutrophil degranulation 34. Thus, the tissue response to metallic implants such as the reported inflammatory response around cobalt and nickel discs implanted intramuscularly into rats 71, does not result from tissue damage due to metal ions Biomaterials 1992, Vol. 13 No. 11
activating neutrophils. However, for intramuscularly implanted copper discs histochemical observations indicated a massive release of lysosomal enzymes and an inflammatory response occurred in tissue next to the implant 7z. These results correlated well with in vitro studies in this laboratory which show that copper ions stimulate neutrophils to degranulate 34. Under these experimental conditions, neutrophils did not stimulate a respiratory burst. Further experiments are required to determine whether metal ions, metal ion-protein conjugates and metal particles activate phagocytic cells and whether phagocytic cells adhere to metal surfaces resulting in activation of these cells. Attachment of macrophages to a substratum may be sufficient to trigger release of lysosomal enzymes. After 16 h incubation on a substratum, untreated cells have been reported to release over 50% of their /~-glucuronidase and 20% of their N-acetylglucosiminidase73. High-energy oxygen species and lysosomal enzymes are an integral part of the neutrophil bactericidal killing mechanisms. Their importance is evident in a number of inherited deficiencies in the mechanisms for killing microorganisms. For example, neutrophils and macrophages from children with the often fatal chronic granulomatous disease (CGD), fail to show the normal respiratory burst and the production of the bacterial agents superoxide and hydrogen peroxide. Neutrophils from CGD patients can be stimulated to release granular enzymes as well as normal control neutrophils TM. This suggests that the mechanisms involved in stimulating a neutrophil oxygen burst and enzyme secretion are not necessarily dependent on one another. This has been confirmed by other workers B1. Rae 31 has reported that nickel and cobalt at concentrations of 0.05 and 0.01pM/ml, reduced the ability of polymorphonuclear leukocytes to phagocytose Staphylococcus epidermidis by 50% over an 18 h incubation period. In a separate study, Rae 42 reported a lowering of the phagocytic capacity of macrophages exposed to particulate cobalt, chromium and cobalt-chromium alloy. A decrease in the efficiency of phagocytosis could increase the risk of infection at the site of a metallic implant. This may account, at least in part, for the high infection rate after total joint replacement, where a mean rate of infection of 5% has been reported one month after surgery 43. Rae 42 provided no evidence to indicate that nickel or cobalt adversely affected any of the bactericidal processes. However, a recent study 75 has shown that activation of exocytosis and a respiratory burst in rabbit neutrophils by the chemotactic peptide FMLP is inhibited by Co 2+ and that this inhibition may be the result of interaction of Co 2+ with a Ca2+-dependent intracellular target. Other metal ions and particles have also been shown to affect macrophage function. The phagocytic and microbicidal activities of macrophages are affected by cadmium ions TMand zinc ions immobilize peritoneal macrophages 77. These functional alterations have been proposed as a reason for the interference with the host defence system after in vivo administration of cadmium and zinc. Nickel particles inhibited proliferation of murine lymphocytes 37 suggesting an inhibition of the cellular immune response. Thus, nickel particles may increase the risk of infection at the implant site.
85
The Biomaterials Silver Jubilee Compendium Immune response in biocompatibility: A. Remes and D.F. Wi//iams
C O M P L E M E N T ACTIVATION Complement is a system of at least 13 serum proteins that are activated by enzymatic cleavages and aggregations to produce components with biological activity. The complement system can be activated by either the classical or alternative pathways (Figure 3). It is well recognized that activation products of the complement system play an important role in acute inflammatory processes and in host defence against invasion by pathogenic organisms. This involvement may lead to changes in vascular permeability, histamine release from mast cells, activation and emigration of phagocytic leukocytes, enhancement of opsonization and phagocytosis and intracellular killing of pathogenic microbes and multicellular pathogens 78' 79. These processes are mediated mainly by C3a, C5a and C3b (Figure 1). C3a and C5a also stimulate monocytes and macrophages to synthesize and secrete interleukin-1 (IL-1)8~ Many features of the inflammatory response are mediated by the effect of IL-1 on the target tissue. Biological effects of IL-1 include an ability to stimulate proliferation of fibroblasts and pyrogenic activity 8z. The importance of the complement cascade is manifested in those people who are deficient in one or more complement proteins. Such people may be plagued with recurrent infections 83' 84. Experimental animals depleted of complement are unable to direct circulating neutrophils to diverse inflammatory sites 8s' 86. If the complement system is inadvertently activated by an implant, complement activation products could mediate an inflammatory response. Complement activation could therefore underlie the foreign body reaction to some materials. Experiments in our laboratory have shown that cobalt Classical pathway
Figure3 Diagram of complement activation.
737
powder and perhaps nickel powder but not chromium powder activate complement 51. Complement activation may play a role in mediating the inflammatory response which occurred after cobalt and nickel discs were implanted intramuscularly in rats 71. Complement may be activated as a result of either the chemical or morphological properties of the metals and this is discussed later, The ceramic powders CaHPO 4 (Merck), Ca3(PO4) 2 (fl-whitlockite) and coral were also observed to activate complement ~9'52. Complement activation may be the mechanism by which some ceramic implants are resorbed. In the case of biodegradable ceramic implants, which are used as bone substitutes, e.g. coral, fl-whitlockite (Ca3[P04~2~,macrophages and giant cells associated with the implant are involved in the biodegradation processes87, 88. Macrophages have receptors for C3b (Ref. 89), The possibility has arisen that the complement proteins of blood plasma mediate cellular adhesion following exposure of a synthetic material to blood. The evidence for this has been reviewed ~176 and indicates that the complement system mediates cellular adhesion to synthetic materials. These receptors cause cells to adhere to other cells or particles bearing C3b on their surfaces and are important in facilitating phagocytosis. C3b may also stimulate 'frustrated' phagocytosis by phagocytes 91, a process which would result in degradative enzymes and high-energy oxygen species being secreted over the implant surface, Both phagocytosis and frustrated phagocytosis may lead to resorption of a ceramic implant. Macrophages have been reported to phagocytose both hydroxyapatite ~2' 93 and/~-whitlockite 87' 88 material resulting in resorption of ceramic implants made of these materials (macrophages have also been shown to degrade hydroxyapatite crystals /n vitro941. Vacoules containing electran-dense implant-derived material were present in the cytoplasm of phagocytes around the degraded ceramic implant. In some cases, indentations were seen in the macrophage plasma membrane suggestive of exocytosis at the implant surface. C3b {as well as C3a and CSa) may also mediate resorption of ceramic implants through its ability to stimulate IL-1 release from monocytes 95. IL-1 has been reported to initiate bone-cartilage resorption 96. Since hydroxyapatite is the main constituent of the mineral matrix of bone, some resemblance between the degradation of hydroxyapatite material and of bone may be expected. Chemotactic complement activation products, i.e. C5a, may contribute towards the accumulation of phagocytic cells at the implant site. Previous workers have reported that complement may play an important role in the migration of neutrophils to an inflammatory site 85'97. Indeed, the neutrophil polarization assay for detecting complement activation indicates that, in the case of ceramic biomaterials, the percentage of neutrophils that assume a locomotor morphology is related to the quantity of C3 activation 52. As the percentage neutrophil polarization is related to chemotaxis, chemotaxis may therefore be related to the quantity of C3 activation. The phenomena of neutrophil polarization and chemotaxis which may occur when complement is activated are probably in response to C5a which is recognized to be the most important chemotactic factor released when complement is activated 98' 99. Indeed, the polarization response to purified C5a has previously Biomaterials 1992, Vo]. 13 No. 11
86
The Biomaterials Silver Jubilee Compendium 738
Immune response in biocompatibility: A. Remes and D.F. Williams
been shown to be dose dependent 1~176One way of determining whether CSa is responsible for stimulating neutrophils to polarize is to block the response with antiC5a antibody. A panel of routine monoclonal antibodies that recognize human complement C5a has been described TM. These antibodies blocked C5a-induced neutrophil polarization and chemiluminescence. Another method of examining whether C5a is the complement factor that stimulates chemotaxis is to use C5-deficient serum in complement activation studies. C5a also enhances neutrophil adhesion 24 and aggregation ~~ For granulocyte aggregation, the magnitude of the response is proportional to the log (concentration) of activated plasma 1~ Since macrophages are also known to aggregate and since they have specific C5a receptors, it may be expected that C5a also mediates macrophage aggregation. Macrophage aggregation may lead to macrophage fusion and the formation of giant cells TM. Giant cells are commonly observed in the vicinity of metallic and ceramic implants T M . C5a alters the surface charge of neutrophils ~~ Surface charge has previously been discussed as being an important factor in the interaction between cells and the surface of a biomaterial 4~ It is clear from the above discussion that the interaction of an implant with the complement system may modulate the tissue response to the implant by influencing the quantity of inflammatory mediators C3a, C3b and C5a that are released into the tissue. It is important therefore to elucidate the properties of a biomaterial which determine how the biomaterial surface will interact with the complement cascade. In a study by the authors, cobalt powder was: found to activate complement C3 in serum but serum has little or no chemotactic activity s~, This may be accounted for by either complement-derived chemotactic factors being adsorbed on to the surface of the cobalt powder or the surface of cobalt powder being inefficient in its capacity to form a C5 convertase. In haemodialysis with first-use cuprophan membranes, low concentrations of C5a compared with C3a are due to a lower efficiency of C5 convertase sites relative to that of C3~~ 108.The efficiency of C5 convertase formation relative to C3 convertase formation differs from one activating surface to another. The factors which determine the relative 'coupling efficiency' between C3 and C5 cleavage on a surface have not yet been fully characterized but may depend on adsorption of complement components on to the biomaterial surface. Adsorption of inflammatory mediators of complement activation may also influence the tissue response to an implant by preventing the mediators from binding to cells and blood vessels in the vicinity of the implant. Indeed, C3 is adsorbed by CaHPO4 (BDH) powder 5z. A biomaterial surface may modulate the tissue response through its interaction with C3b. Previous studies have reported that C3b and iC3b may stimulate an oxygen burst in neutrophils and monocytes and promote phagocytosis89.109, a~0. A study in the present authors' laboratory 53 on complement activation by ceramic biomaterial powders, determined that CaHPO 4 [Merck) powder activated serum C3. Despite this, the opsonized powder failed to stimulate an oxygen burst in neutrophils and did not associate with the cell membrane of Biomateria]s 1992, Vo|. 13 No. 11
neutrophils. Thus, under these experimental conditions the neutrophil C3b receptor fails to mediate phagocytosis or a respiratory burst. This is opposite to the situation with sodium urate crystals and zymosan. The latter is a potent activator of the alternative pathway of complement activation 111. Naff and Byers 112 noted that phagocytosis of sodium urate crystals by neutrophils was markedly stimulated by the addition of fresh, but not heated, serum and this implies that the presence of opsonizing complement protein enhanced phagocytosis through interaction with cell membrane receptors for complement. Zymosan opsonized with C3b stimulates a much larger respiratory burst in neutrophils and phagocytosis is greatly increased 37' 41,11s. However, it cannot be ruled out that zymosan particles were also coated with immunoglobulin, the Fc portion of which is a well-known inducer of the oxidative burst and phagocytosis el' 114. Indeed, recent studies have reported that specific antibodies enhanced alternative pathway activation of complement on the surface of opsonized zymosan 11~, and studies by Newman and Johnston ~6 showed that C3b on the surface of sheep erythrocytes did not elicit superoxide anion release from human neutrophils. Thus, in the case of Ca3(PO4) 2 powder and CaHPO 4 (Merck) powder, C3b on the powder surface may not be able to mediate a neutrophil respiratory burst or phagocytosis. However, C3b has previously been reported to enhance IgGmediated phagocytosis 89. As phagocytosis of complement opsonized ceramic powders was not observed in this study, it could be that neutrophils do not recognize IgG on the surface of the ceramic powders. It may also be the case that other serum proteins sterically inhibit neutrophil receptors from recognizing the surface-associated C3b molecule or that the powder surface changes the molecular configuration of the C3b molecule such that it is no longer recognized by the neutrophil C3b receptor. For egg albumin adsorbed to silica, the protein has been reported to be reversibly altered in its configuration, antigenically resembling heat-denatured protein 117. One of the major determinants of the biological activity of an artificial surface is the interaction of the surface with serum and other proteins 118. The adsorbed proteins then determine how the surface interacts with other blood components. Protein adsorption is controlled in part by the properties of the material surface, including chemical composition 1~9 and topography ~2~ Evidence has been obtained that the electrostatic charge of the protein influences adsorption 121'122. It has been demonstrated that negatively charged albumin molecules were avidly bound by positively charged asbestos or quartz 12~. Adsorption of serum proteins to surfaces has been suggested to influence complement activation. There are two possible ways in which complement factors can adsorb to a surface: 1. Complement factors adsorb directly to the crystal surface. However, it has been reported that there is no relationship between C3 adsorption and complement activationlQ' 52. 2. Complement is bound to the protein layer adsorbed to the surface. In this case complement activation may be antibody mediated. IgG or IgM may be adsorbed on to a foreign surface TM creating an antigen-antibody complex which activates the classical pathway.
87
The Biomaterials Silver Jubilee Compendium
Immune response in biocompatibility: A. Remes and D.F. Wi//iarns
Ig absorption to a surface has been reported to be strongly influenced by the ionic charge of both Ig and the surface ~z~.Antibody-coated surfaces are known to activate complement in a manner similar to that induced by immune complexes 12~. Experiments described whereby IgG was bound to glycerol methacrylate microspheres showed that the more IgG was bound, the more complement was activated ~26. Alternative pathway activation of several activators, e.g. sephadex ~27 and zymosan ~15, is greatly enhanced by specific antibodies. However, the enhancing effect of specific IgG on alternative pathway function does not require the Fc region. It is reported to be mediated by an increased rate of C3 cleavage and subsequent C3 deposition on the target surface I28. Previous reports suggest that biomaterials may activate complement by the classical and/or alternative pathways14,129. The classical pathway may be activated in the absence of immunoglobulin. Giclas et al. 1~~demonstrated that low concentrations of sodium urate crystals activated C1 directly in solutions of purified macromolecular C1. Most polymers activate complement by the alternative pathway. The alternative pathway activating capacity of polymers such as sephadex and cellulose is dependent on a high density of exposed nucleophilic hydroxyl groups providing sites for covalent binding of C3b and on the protection of polymer bound C3b from inactivation by the regulatory proteins H and I TM. The mechanisms by which metals and ceramic powders activate complement is unknown. The initial step in complement activation by a material is thought to be covalent binding of the labile thioester group of C3 to nucleophilic groups on the material surface as described earlier for polymers. In the case of ceramic and metal powders, after adsorption of proteins to the material surface, the proteins may undergo a conformational change and nucleophilic groups within the protein may be exposed. For activation of the alternative pathway an ordered surface configuration of repeating polysaccharities, e.g. dextran, or other polymeric units appear to be important. In the case of ceramic powders, the crystals with their repetitive surface structure appear to be suited for alternative pathway complement activation. CaHPO 4 (Merck] powder activates complement, whereas CaHPO4 (BDH) powder does not $2. The different ways in which these two ceramic powders interact with the complement cascade suggest that the physical, as well as the chemical, nature of a biomaterial surface is an important factor in determining the mechanism of complement activation. The charge on the surface of CaHPO4 powders has previously been discussed as being an important factor in determining the interaction of CaHPO4 powders with neutrophils 49. The surface charge of ceramic powders may also be a factor in determining how these powders interact with the complement cascade.
CELLULAR IMMUNE RESPONSE Metallic implants may sensitize the patient to one or more of the metal ions in a prosthesis TMand macrophages and giant cells are found in tissue next to ceramic and
739
metallic implants 88'1~ These findings suggest that biomaterials may stimulate a cell-mediated immune response [CMI), also called a hypersensitivity reaction. Additional evidence for this comes from recent reports that the tissue around a loose total joint replacement prosthesis displays a synovial-like lining comprised of cells that produce IL-1132 and that polymers stimulate IL-1 release from monocyte cultures TM. IL-1 is a very important factor in mediating a CMI and has been suggested to play a major role in determining the biocompatibility of surgical biomaterials 133. Biomaterials have also been shown to elicit a differential host response in terms of cytokine production. Macrophages adherent to expanded PTFE were stimulated to produce more IL-1 than macrophages adherent to silicone elastomer 1~3. Nickel, cobalt and chromium are important constituents of many implants, Hypersensitivity to metals has been reported to be the cause of failure of some of these implants16.17 possibly through the binding of metal ions to proteins or cells. Metal ions from both metal salts and corrosion products have been reported to bind to white cells 32'36 and injection of syngeneic macrophages haptenized with nickel ions into guinea pigs induces nickel hypersensitivity TM. These complexes could stimulate a CMI by activating T cells. T cell clones have been isolated from peripheral blood of nickel contactsensitive patients and these clones respond to nickel presented with MHC class II antigens on the surface of antigen presenting cells such as macrophages TM. In other studies, lymphocytes from patients with chromium and nickel sensitivity have been shown to be activated in the presence of the metal ions, as evidenced by lymphocyte proliferation assays 13e' I3~. In vitro studies to determine the effects of metal ions on the expression of lymphocyte surface antigens showed that Fe 3+ or Co 2+ caused inhibition of CD2 only, whereas Ni 2§ caused inhibition of both CD2 and CD3 antigens ~38.These findings suggest that Fe2*, Co 2§ and Ni 2+ ions may interfere with T cell activation since both CD2 and CD3 are involved in that process. Wear products of biomaterials may be either particulate or ionic. Nickel powder did not stimulate but inhibited in vitro proliferation of mouse lymph node lymphocytes 37. This suggests that nickel powder is toxic for mouse lymphocytes. Previous studies have shown particulate nickel to be toxic for macrophages 42 and strongly haemolytic 13~ Haemolysis did not appear to be caused mechanically as other particles, in similar size ranges had no effect. Other work 1~~ has shown that nickel discs implanted intramuscularly are toxic for tissue next to the implant. Thus, nickel particles may not interact with lymphocytes in the same way as nickel salts. Nickel ions have been reported to stimulate the proliferation of murine splenocytes 141, a response which is mediated by cells with the phenotype of TDTH cells. The different modes of interaction with lymphocytes for metal salts and nickel powder is perhaps not surprising when one considers that the mechanism of sensitization to metal salts is based on metal ions forming a complex with protein or cells. Clearly metal particles cannot interact with pj:otein in the same way as metal salts. Biomaterials 1992, Vol. 13 No. 11
88
The B i o m a t e r i a l s Silver J u b i l e e C o m p e n d i u m
740
Immune response in biocompatibility: A. Remes and D.F. Williams
A ceU-mediated immune response has been reported after ~-whitlockite was implanted intramuscularly either as a powder ~4z or as a porous semicylindrical implant 143. Lymphocytes and plasma cells were present at the implant site. In the case of the semicylindrical implant ceramic particles were found in the popliteal and inguinal lymph nodes. In the case of the implanted ~-whitlockite powder, the authors suggested that 'fl-whitlockite m a y act as a hapten and stimulate an immunoresponse w h e n coupled to protein'. It could be that proteins are altered in conformation after attachment to the CaHPO4 powder. This protein-powder combination may be phagocytosed by macrophages and carried to a local lymph node where the altered self protein is presented on the surface of macrophages to specific T cells. These T cells would then proliferate and mediate a cellular immune response. Other workers have shown that macrophages phagocytose ceramic and metallic biomaterial powders 42' 88, 144 and have observed powder particles in the regional lymph nodes 143' 144.
9
10
11 12 13 14
15 CONCLUSIONS Cells and mediators of the immune system are often noted at the site of implantation of biomaterials. It is important to determine the influence that these agents have on biocompatibility as this could influence the length of time which a biomedical device can remain in the host as well as the biofunctionality of such a device. It is clear from this review that the immune system does play an important role in determining the biocompatibility of a biomaterial and further research is essential to understand the precise mechanisms involved.
16 17
18
19 REFERENCES Jones, T.C. and Hirsch, The interaction in vitro of Mycoplasma pulmonis with mouse peritoneal macrophages and L-cells, J. Exp. IVied. 1971, 133, 231-259 2 Jasin, H.E., Human heat labile opsonins: evidence for their mediation via the alternative pathway of complement activation, ]. Immunol. 1972, 109, 26-31 3 Hart, P.H., Spencer, L.K., Nikoloutsopoulos, A., Lopez, A.F., Vadas, M.A., McDonald, P.J. and Finlay, Jones, J.J., Role of cell surface receptors in the regulation of intracellular killing of bacteria by murine peritoneal exudate neutrophils, Infect. Immun. 1986, 52, 245-251 4 Wilkinson, P.C., Random locomotion, chemotaxis and chemokinesis. A guide to terms defining cell locomotion, Immunol. Today 1985, 9, 273-278 5 Herbert, W.J. and Wilkinson, P.C., Dictionary of immunology, 3rd edn, Blackwell Scientific, Oxford, UK, 1985 6 Zinkernagel, R.M. and Doherty, P.C., H-2 compatibility requirement for T-ceU mediated lysis of target cells infected with lymphocytic choriomeningitis virus: different T-cell specificities are associated with structures coded for in H-2K or H-2D, J. Exp. Med. 1975, 141, 1427-1436 7 Chestnut, R.W., Colon, S.M. and Grey, H.W., Requirements for the processing of antigens by antigen-presenting cells. I. Functional comparison of B cell tumors and macrophages, J. Immunol. 1982, 129, 2382-2388 8 Guidos, C., Wong, M. and Lee, K., A comparison of the stimulatory activities of lymphoid dendritic cells and
20
1
Biomaterials 1992, Vol. 13 No. 11
21
22
23 24 25 26
27 28
macrophages in T proliferative responses to various antigens, J. Immunol. 1984, 133, 1179-1183 Chestnut, R.W., Colon, S.M. and Grey, H.M,, Antigen presentation by normal B-cells, B cell tumors and macrophages: functional and biochemical comparison, J. ImmunoL 1982, 128, 1764-1768 Ziegler, H.K. and Unanue, E.R., Identification of a macrophage antigen processing event required for I-region-restricted antigen presentation to T lymphocytes, J. Immunol. 1981, 125, 1869-1875 Calderon, J. and Unanue, E.R., The release of antigen molecules from macrophages: characterization of the phenomena, J. Immunol. 1974, 112, 1804-1814 Klaus, G.G.B. and Hawrylowicz, C.M., Cell-cycle control in lymphocyte stimulation, Immunol. Today 1984, 5, 15-19 ]'andl,R.C., Flanagan, R.G. and Schur, P.H., Interleukin-1 stimulation of human B-lymphoblast differentiation, Clin. Immuno]. Immunopathol. 1988, 46, 115-121 Craddock, P.R., Fehr, J,, Dalmasso, A.P., Brigham, K.L. and Jacob, H.S., Haemodialysis leukopenia: pulmonary vascular leukostasis resulting from complement activation by dialyser membranes, J. Clin. Invest. 1977, $9, 879-888 Hammerschmidt, D.E., Stroncer, D.F. and Bowers, T.K., Complement activation and neutropenia occurring during cardiopulmonary bypass, J. Thorac. Cardiovasc. Surg, 1981, 81, 370-377 Elves, M.W., Wilson, ].M., Scales, J.T. and Kemp, H.B.S., Incidence of metal sensitivity in patients with total joint replacements, Br. Med. J. 1975, 4, 376-378 Evans, E.M., Freeman, M,A.R., Muller, A,J. and VernonRoberts, B., Metal sensitivity as a cause of bone necrosis and loosening of the prosthesis in total joint replacement, I. Bone Joint Surg. (Br.) 1974, 56, 626-642 Miller, K.M. and Anderson, J.M., Human monocyte/ macrophage activation and interleukin 1 generation by biomedical polymers, J. Biomed. Mater. Res. 1988, 22, 713-731 Klein, C.P.A.T., de Groot, K. and van Kamp, G., Activation of complement C3 by different calcium phosphate powders, Biomaterials 1983, 4, 181-184 Vince, D.G., Hunt, J.A. and Williams, D.F., Quantitative assessment of the tissue response to biomaterials, Biomaterials, in press McPhail, L,C. and Syderman, R., Activation of the respiratory burst enzyme in human polymorphonuclear leukocytes by chemoattractants and other soluble stimuli, J. Clin. Invest. 1983, 72, 192-200 Becker, E.L., Showell, H.J., Henson, P.M. and Hsu, L.S., The ability of chemotactic factors to induce lysosomal enzyme release. I. The characteristics of the release, the importance of surfaces and the relation of enzyme release to chemotactic responsiveness, J. lmmunol. 1974, 112, 2047-2054 Wilkinson, P.C. and Lackie, ].M., The influence of contact guidance on chemotaxis of human neutrophil leukocytes, Exp. Cell. Res. 1983, 145, 255-264 O'Flaherty, J.T,, Craddock, P.R. and Jacob, H.S., Effect of intravascular complement activation on granulocyte adhesiveness and distribution, Blood 1978, 51, 731-739 Rich, A. and Harris, A.K., Anomalous preferences of cultured macrophages for hydrophobic and roughened substrata, J. Cell. Sci. 1981, 50, 1-7 Astor, S.H., Stifler, L.E., Frick, O.L. and Fudenberg, H.H., Human leukocyte migration inhibition in agarose, using four antigens -- correlation with skin reactivity, J. Immunol. 1973, 110, 1174-1179 Allan, R.B. and Wilkinson, P.C., A visual analysis of chemotactic and chemokinetic locomotion of human neutrophil leukocytes, Exp. Cell. Res. 1978,111,191-203 Williams, D.F. in Fundamental Aspects of Biocompati-
89
The Biomaterials Silver Jubilee Compendium Immune response in biocompatibility: A. Remes and D.F. Williams
bility Vol. 1 (Ed. D.F. Williams), CRC Press, Boca Raton, FL, USA, 1981, p. 11 29 Remes, A. and Williams, D.F., Chemotaxis and the inhibition of chemotaxis of human neutrophils in response to metal ions, J. Mater. Sci.: Materials in Medicine 1990, 1, 26-32 30 Simchowitz, L. and Cragoe, E.J., Polyvalent cations inhibit human neutrophil chemotaxis by interfering with the polymerization of actin, J. Biol. Chem. 1990, 9.65, 13457-13463 31 Roe, T., The action of cobalt, nickel and chromium in phagocytosis and bacterial killing by human polym0rphonuclear leukocytes: its relevance to infection after total joint arthroplasty, Biomaterials 1983, 4, 175-180 32 Merritt, K., Brown, S.A. and Sharkey, N.A., The binding of metal salts and corrosion products to cells and proteins in vitro, J. Biomed. Mater. Res. 1984, 18, 1005-1015 33 Van Soestbergen, M. and Synderman, F.W., 63Ni complexes in rabbit skin serum and urine after injection of 63NIC12, Clin. Chem. 1972, 18, 1478-1484 34 Hunt, J.A., Remes, A. and Williams, D.F., The effect of metal ions on neutrophil degranulation, J. Mater. Sci.: Materials in Medicine, in press 35 McNamara, A. and Williams, D.F., Scanning electron microscopy of the metal-tissue interface. II. Observations with lead, copper, nickel, aluminium and cobalt, Biomaterials 1982, 3, 165-176 36 Merritt, K., Brown, S.A. and Sharkey, N.A., Blood distribution of nickel, cobalt and chromium following intramuscular injection into hamsters, jr. BJomed. Mater. Res. 1984, 18, 991-1004 37 Remes, A., PhD Thesis 38 Bessis, M. and Butte, B., Positive and negative chemotaxis as observed after the destruction of a cell by U.V. or laser microbeams, Tex. Rep. Biol. IVied. 1965, 9.3, 204-212 39 Wilkinson, P.C. and McKay, I.C., The chemotactic activity of natured and denatured serum albumin, Int. Arch. Allergy Appl. Immunol. 1971, 41, 237-247 40 Metchnikoff, E., Immunity in Infective Diseases, Cambridge University Press, London, UK, 1905 41 Gleisner, J.M., Lysosomal factors in inflammation, in Handbook of Inflammation, Vol. 1 (Ed. J.C. Houck) 1979, pp. 229-260. 42 Roe, T., A study of the effects of particulate metals of orthopaedic interest on murine macrophages in vitro, J. Bone Joint Surg. 1975, 57-8, 444-450 43 Rae, T., Localised tissue infection and the influence of foreign bodies, in Fundamental Aspects of Biocompatibility, Vol II (Ed. D.F. Williams) CRC Press, FL, USA, 1981, pp. 139-158 44 Bearden, L.J. and Cooke, F.W., Growth inhibition of cultured fibroblasts by cobalt and nickel, J. Biomed. Mater. Res. 1990, 14, 289-309 45 Eldow, D.W. and Sheldon, W.H., The pH of inflammatory exudates, Proc. Soc. Exp. Biol. Med. 1971, 137,1328-1332 46 Zigmond, S.H. and Hargrove, R.L., Orientation of PMN in a pH gradient: acid induced release of a chemotactic factor, J. Immunol. 1990, 128, 478-481 47 Ward, T.T. and Steigbigel, R.T., Acidosis of synovial fluid correlates with synovial fluid leukocytosis, Am. ]. Med. 1978, 64, 933-936 48 Rotstein, O.D., Fiegel, V.D., Simmons, R.L. and Knighton, D.R., The deleterious effect of reduced pH and hypoxia on neutrophil migration in vitro, J. Surg. Res. 1988, 45, 298-303 49 Remes, A. and Williams, D.F., Stimulation of a neutrophil respiratory burst by calcium hydrogen phosphate powder, Clin. Mater., in press 50 Kaplan, S.S., Basford, R.E., Kormos, R.L., Hardesty, R.L., Simmons, R.L., Mora, E.M., Cardona, M. and Grifith, B.L., Biomaterial associated impairment of local
51
52 53
54
55 56 57
58
59
60
61
62 63 64
65
66 67
68 69
741
neutrophil function, A S A I O Trans. 1990, 36, M172-175 Remes, A. and Williams, D.F., Neutrophil polarization and immunoelectrophoresis assays in the study of complement activation by biomaterials, Biomaterials, in press Remes, A. and Williams, D.F., The relationship between chemotaxis and complement activation by ceramic biomaterials, Biomaterials, in press Wright, D.G. and Gallin, J.I., A functional differentiation of human neutrophil granules: generation of C5a by a specific (secondary) granule product and inactivation of C5a by azurophil (primary] granule products, J. lmmunol. 1977, 119, 1068-1076 Goldstein, I.M. and Weissman, G., Generation of C5derived lysosomal enzyme-releasing activity (C5a) by lysates of leukocyte lysosomes, J. lmmunol. 1974, 113, 1583-1588 Ward, P.A. and Hill, J.H., C5 chemotactic fragments produced by an enzyme in lysosomal granules of neutrophils, I. Immuno]. 1970, 104, 535-543 Higgs, G.A., Arachidonic acid metabolism in leukocytes, in Phagocytes, Past and Future [Eds M.L. Karnovsky and L. Bolis), Academic Press, London, UK, 1982, pp. 105-129 Ford-Hutchinson, A.W., Bray, M.A., Doig, M.V,, Shipley, M.E. and Smith, M.J.H., Leukotriene B, A potent chemokinetic and aggregating substance released from polymorphonuclear leukocytes, Nature [Lond.} 1980, 286, 264-265 Zigmond, S.H. and Hirsch, J.G., Leukocyte locomotion and chemotaxis. New methods for evaluation and demonstration of a cell derived chemotactic factor, J. Exp, IVied. 1973, 137, 387-410 Phelps, P., Polymorphonuclear leukocyte motility in vitro. III. Possible release of a chemotactic substance after phagocytosis of urate crystals by polymorphonuclear leukocytes, Arthritis Rheum. 1969,12,197-203 Tse, R.L. and Phelps, P., Polymorphonuclear leukocyte motility in vitro. V. Release of chemotactic activity following phagocytosis of calcium pyrophosphate crystals, diamond dust and urate crystals, f. Lab. Clin. Med. 1970, 76, 403-415 Goldstein, I.M., Roos, D., Kaplan, H.B. and Weissman, G., Complement and immunoglobulins stimulate superoxide production by human leukocytes independently of phagocytosis, J. Clin. Invest. 1975, 56, 1155-1163 Becker, E.L. and Showell, H.J., The ability ofchemotactic factors to induce lysosomal enzyme release. II. The mechanisms of release, J. lmmunol. 1974, 112, 2055-2062 Till, G.O. and Ward, P.A., Lung injury secondary to chemotactic factor induced leukocyte activation. Agents Actions {Suppl.) 1983, 12, 383-396 Henson, P.M., Secretion of lysosomal enzymes induced by immune complexes and complement, in Lysosomes in Biology and Pathology, Vol. 5, (Eds J.T. Dingle and R.T. Dean), North Holland, The Netherlands, 1976, pp. 99-126 Schorlemmer, H.K., Davies, P., Hylton, W., Gugig, M. and Allison, A., The selective release of lysosomal acid hydrolases from mouse peritoneal macrophages by stimuli of chronic inflammation, Br. J. Exp. Pathol. 1977, 58, 315-326 Ignarro, L.J., Regulation of lysosomal enzyme secretion: role in inflammation, Agents Actions 1974, 414, 241-258 Sacks, T., Moldow, C.F., Craddock, P.R., Bowers, T.K. and Jacobs, H.S., Oxygen radicals mediate endothelial cell damage by complement stimulated granulocytes. An in vitro model of immune vascular damage, J. Clin. Invest. 1978, 61, 1161-1167 Bawdey, J.A. and Karnovsky, M.L., Active oxygen species and the functions of phagocytic leukocytes, Ann. Rev. Biochem. 1980, 49, 695-726 Carp, H. and Janoff, A., In vitro suppression of serumBiomaterials 1992, Vol. 13 No. 11
The Biomaterials Silver Jubilee Compendium
90
Immune response in biocompatibility: A. Remes and D.F. Williams
742
70 71 72 73 74
75 76
77
78 79
80
81
82 83 84
85
86
87
88
elastase-inhibitory capacity by reactive oxygen species generated by phagocytosing polymorphonuclear leukocytes, I. Clin. Invest. 1979, 63, 793-797 Elferink, J.G.R. and Deierkauf, M., A biochemical study of hydroxyapatite crystal induced enzyme release from neutrophils, Ann. Rheum. Dis. 1987, 46, 590-597 McNamara, A. and Williams, D.F., The response to the intramuscular implantation of pure metals, Biomaterials 1981, 2, 33-40 McNamara, A. and Williams, D.F., Enzyme histochemistry of the tissue response to pure metal implants, J. Biomed. Mater. Res. 1984, 18, 185-206 Rae, T., The biological response to titanium and titaniumaluminium-vanadiurn alloy particles. I. Tissue culture studies, Biomaterials 1986, 7, 30-36 Ulevitch, R.J., Henson, P., Holmes, B. and Good, R.A. An in vitro study of exocytosis of neutrophil granule enzymes in chronic granulomatous disease neutrophils, J. Immunol. 1974, 119., 1383-1386 Elferink, ].G. and Deierkauf, M., Suppressive action of cobalt on exocytosis and respiratory burst in neutrophils, Am. J. Physiol. 1989, 257{5 pt 1}, C859-864 Loose, L.D., Silkworth, J.B. and Simpson, D.W., Influence of cadmium on the phagocytic and microbicidal activity or murine peritoneal macrophages and polymorphonuclear neutrophils, Infect. lmmun. 1978, 22, 378-381 Zukosi, C.F., Chvapil, M., Carlson, E., Hattler, B. and Ludwig, J., Functional immobilization of peritoneal macrophages by zinc, J. Reticuloendothel. Soc. 1974, 16, 6a Vogt, W., Activation of the complement system, Agents Actions 1983, 13, 391-397 Sobel, A., Blanc, C., Cattaneo, A., Moisy, M., LopezTrascasa, M., Bourgarit, ].]. and Gobay, Y., Receptors for complement components in inflammation, Agents Actions 1983, 13, 398-405 Goodman, M.G., Chenoweth, D.E. and Weigle, W.O., Induction of interleukin-1 secretion and enhancement of humoral immunity by binding of human C5a to macrophage surface C5a receptors, J. Exp. Med. 1982, 156, 912-917 Haeffner-Cavaillon, N., Cavaillon, J.M., Laude, M. and Kazatchkine, M.D., C3a (C3a des Arg] induces production and release of interleukin I by cultured human monocytes, ]. Immunol. 1987, 139, 794-799 Dinarello, C.A., Interleukin I, Rev. Infect. Dis. 1984, 6, 51-95 Ruddy, S., Gigli, I. and Austin, K., The complement system of man, N e w Engl. J. Med. 1972, 287, 489-495, 545-549, 592-596, 642-646 Alper, C.A., Colten, H.R., Rosen, F.S., Rabson, A.R., McNab, G.M. and Gear, ].S.S., Homozygous deficiency of C3 in a patient with repeated infections, Lancet 1972, 2, 1179 Cochrane, C., Muller-Eberhard, H. and Aiken, B., Depletion of plasma complement in vivo by a protein of cobra venom: its effect on various immunological reactions, J. Immunol. 1970, 105, 55-69 Wiener, S., Lendvai, S., Rogers, B., Urivetsky, M. and Meilman, E., Non immune chernotaxis in vivo: inhibition by complement depletion with cobra venom factor, Am. J. Pathol. 1973, 73, 807-816 Klein, C.P.A.T., de Groot, K., Priessen, A.A. and van der Lubbe, H.B.M., A comparative study of different fl-whitlockite ceramics in rabbit cortical bone with regard to their biogradation behaviour, Biomaterials 1986, 7, 144-146 Van Blitterswijk, C.A., Grote, J.J., Koerton, H.K. and Kuijpers, W., The biological performance of calcium ceramics in an infected implantation site. III. Biological performance of ~-whitlockite in the noninfected and
Biomaterials 1992, Vol. 13 No. 11
89 90 91
92
93
94
95
96 97 98 99 100
101
102
103
104 105
106
107
infected rat middle ear,J. Biomed. Mater. Res. 1986, 20, 1197-1217 Ehlenberger, A.G. and Nussenzweig, V., The role of membrane receptors for C3b and C3d in phagocytosis, J. Exp. Med. 1977, 145, 357-371 Ward, C.A. and Kalman, P.G., Complement activation mediates cellular adhesion to synthetic biomaterials, Med. Prog. Technol. 1989, lS, 63-75 Henson, P.M.,The immunologicalrelease of constituents from neutrophil leukocytes. I. The role of antibody and complement on nonphagocytosable particles,J. Immuno]. 1971, 107, 1535-1540 Van Blitterswijk, C.A., Grote, J.J., Kuijpers, W., Blok-van Hock, C.J.G. and Daems, W.Th., Bioreactions at the tissue/hydroxyapatite interface, Biomaterials 1985, 6, 243-251 Van Blitterswijk, C.A., Grote, J.J., Kuijpers, W., Daems, W.Th. and de Groot, K., Macropore tissue ingrowth: a quantitative and qualitative study on hydroxyapatite ceramics, Biomaterials 1986, 7, 137-143 Kwong, C.H., Burns, W.B. and Cheung, H.S., Solubilization of hydroxyapatite crystals by murine bone cells, macrophages and fibroblasts, Biomaterials 1989, 10, 578-584 Couturier, C., Haeffner-Cavaillon, N. and Kazatchkine, M.D., C3 receptor-mediated production and release of interleukin 1 {IL-1}by human monocytes, in Biochemistry of Acute Allergic Reactions (Eds A. Tauber, B.U. Wintroub and A. Stoppler}, Alan Riss, N e w York, USA, 1989, pp. 233-245 Gowen, M., Wood, D.D., lhrie,E.J.,McGuire, M.K.B. and Russel, R.G.G., An interleukin 1-1ike factor stimulates bone resorption in vitro, Nature 1983, 306, 378-380 Elmgreen, J., Neutrophil recruitment in skin window chambers -- activation by complement, Acta. Pathol. Microbiol. Irnmunol. Scand. Sect. C, 1985, 93, 139-142 Hartung, H.P. and Hadding, U., Complement components in relation to macrophage function, Agents Actions 1983, 13, 415-428 Elmgreen, J,, Berkowicz, A. and Sorensen, H., Defective release of C5a related chemo-attractant activity from complement in Crohn's disease, Gut 1983, 24, 525-531 Haston, W.S. and Shields, J.M., Neutrophil leukocyte chemotaxis: a simplified assay for measuring polarizing responses to chemotactic factors, J. lmmunol. Methods 1985, 81, 229-237 Larrick, ].W., Wang, J., Fondly, B.M., Chenoweth, D.E., Kunkel, S.L. and Deinhart, T., Characterization of murine rnonoclonal antibodies that recognize neutralizing epitopes on human C5a, Infect. Immun. 1987, $5, 1867-1872 Craddock, P.R., Hammerschmidt, D.E., White, J.G., Dalmasso, A.P. and Jacob, H.S., Complement (C5a}induced 8ranulocyte aggregation in vitro, J. Clin. Invest. 1977, 60, 260-264 Hammerschmidt, D.E., Bowers, T.K., Lammi-Keefe, C.J., Jacob, H.S. and Craddock, P.R., Granulocyte aggregometry: a sensitive technique for the detection of C5a and complement activation, Blood 1980, 5S, 898-902 Sutton, J.S. and Weiss, L., Transformation of monocytes in tissue culture into macrophages, epithelioid cell and multinucleate giant cells, ]. Cell. Biol. 1966, 28,303-307 Spector, M., Cease, C. and Xong-Li, X., The local tissue response to biomaterials, in Critical Reviews in Biocompatibility {Ed. D.F. Williams}, CRC Press, Boca Raton, FL, USA, 1989, pp. 269-295 Gallin, J.I., Durocher, J. and Kaplan, A., Interaction of leukocyte chemotactic factors with the cell surface. I. Chemotactic factor-induced changes in human granulocyte surface charge, J. Clin, Invest. 1975, 55, 967-974 Chenoweth, D.E., Cheung, A.K. and Henderson, L.W.,
91
The Biomaterials Silver Jubilee Compendium Immune response in biocompatibility: A. Remes and D.F. Williams Anaphylatoxin formation during hemodialysis: effect of different dialyser membranes, Kidney Int. 1983, 24, 764-769 108 Wegmuller, E., Montandon, A., Nydegger, U.E. and Des Coeudres, C., Biocompatibility of different hemodialysis membranes: activation of complement and leukopenia, Int. f. Artif. Organs 1986, 9, 85-92 109 Goldstein, I.M., Kaplan, H.B., Radin, A. and Frosch, M., Independent effects of IgG and complement upon human polymorphonuclear leukocyte function, f. Immunol. 1976, 117, 1282-1287 110 Cook, J. and Kazatchkine, M.D., Complement and cell interactions in inflammatory reactions, in Enzyme Activation in Blood Perfused Artificial Organs (Eds H. Schmidt-Schonbein, L.]'. Wurzinger and R.E. Zimmerman), Martinus Nijhoff, Dordrecht, 1985, pp. 122-134 111 Heideman, M., Complement activation in vitro induced by endotoxin and injured tissue, J. Surg. Res. 1979, 26, 670-673 112 Naff, G.B. and Byers, P.W., Possible implication of complement in acute gout, J. Clin. Invest. 1967, 46, 1099-1100 113 Hill, H.R., Hogan, N.A., Bale, J.F. and Hemming, V.G., Evaluation of nonspecific (alternative pathway) opsonic activity by neutrophil chemiluminescence, Int. Arch. Allergy Appl. Immunol. 1977, 53, 490-497 114 Mantovani, B., Different roles of IgG and complement receptors in phagocytosis by polymorphonuclear leukocytes, ]. Immunol. 1975, 15, 15-17 115 Schenkein, H.A. and Ruddy, S., The role of immunoglobulins in alternative complement activation by zymosan. I. Human IgG with specificity for zymosan enhances alternative pathway activation by zymosan, J. Immunol. 1981, 126, 7-10 116 Newman, S.L. and Johnston, R.B., Role of binding through C3b and IgG in polymorphonuclear neutrophil function: studies with trypsin generated C3b, J. Irnmunol. 1979, 123, 1839-1846 117 Scheel, L.D., Smith, B., Van Riper, J. and Fleisher, E., Toxicity of silica. II. Characteristics of protein films absorbed by quartz, Arch. Indust. Hyg. 1954, 9, 29-36 118 Grinnell, F., The role of fibronectin in the bioreactivity of material surfaces, in Biocompatible Polymers, Metals and Composites (Ed. M. Szycher] Technomic, Lancaster, PA, USA, 1983, pp. 673-699 119 Laing, P.G., Ferguson, A.R. and Hodge, E.S., Tissue reaction in rabbit muscle exposed to metallic implants, J. Biomed. Mater. Res. 1967, 1, 135-149 120 Taylor, S.R. and Gibbons, D.F., Effect of surface texture on the soft tissue response to polymer implants, J. Biomed. Mater. Res. 1983, 17, 205-227 121 Hasselbacher, P., Binding of IgG by monosodium urate monohydrate is a function of the ionic charge of the protein, Clin. Res. 1978, 26, 623 122 Hasselbacher, P., Binding of IgG and complement protein by monosodium urate and other crystals, jr. Lab. Clin. ivied. 1979, 94, 532-541 123 Morgan, A., Absorption of human serum albumin by asbestiform minerals and its application to the measurement of surface areas of dispersed samples of chryostile, Environ. Res. 1974, 7, 330-341 124 Klein, C.P.A.T., de Groot, K., Vermeiden, J.P.W. and van Kemp, G., Interaction of some serum proteins with hydroxyapatite and other materials, J. Biomed. Mater. Res. 1980, 14, 705-712 125 Ishizaka, T., Ishizaka, K. and Boros, T., Biological activity of aggregated gamma-globulin. IV. Mechanism of complement fixation, J. Immunol. 1961, 89, 709-716 126 Uchida, T., Hosaka, S. and Murao, Y., Complement
127
128
129 130
131
132
133
134
135
136
137
138
139 140 141 142
143
144
743 activation by polymer binding IgG, Biomaterials 1984, 5, 281-283 Carreno, M.P., Maillet, F., Labarre, D., jozefowicz, M. and Kazatchkine, M.D., Specific antibodies enhance sephadex-induced activation of the alternative complement pathway in human serum, Biomaterials 1988, 9, 514-518 Moore, F.D. Jr, Fearon, D.T. and Austen, K.F., IgG on mouse erythrocytes augments activation of the human alternative complement pathway by enhancing deposition of C3b, J. lmmunol. 1981, 126, 1805-1809 Shepard, A.D., Gelfand, J.A., Callow, A.D. and O'Donnell, T.F., Complement activation by synthetic vascular prosthesis, J. Vasc. Surg. 1984, 1, 829-838 Giclas, P.C., Ginsberg, M.H. and Cooper, N.R., Immunoglobulin G independent activation of the classical complement pathway by monosodium urate crystals, ]. Clin. Invest. 1979, 63, 759-764 Haeffner-Cavaillon, N., Fischer, E,, Bacle, F,, Carreno, M.P., Maillet, F., Cavaillon, J.M. and Kazatchkine, M.D., Complement activation and induction of interleukin 1 production during hemodialysis, Contrib. Nephrol. 1988, 62, 86-98 Thornhill, T.S., Ozuna, R.M., Shortkroff, S., Keller, K., Sledge, C.B. and Spector, M., Biochemical and histological evaluation of the synovial-like tissue around failed (loose) total joint replacement prosthesis in human subjects and a canine model, Biomaterials 1990, 11, 69-72 Krause, T.J., Robertson, F.M., Liesch, J.B., Wasserman, A.J. and Grecos, R.S., Differential production of IL-1 on the surface of biomaterials, Arch. Surg. 1990, 125, 1158-1160 Von Blomberg-Van der Flier, M., Scheper, R,J., Boerrighter, G.H. and Polak, L. Induction of contact sensitivity to a broad variety of allergens with haptenized macrophages, J. Invest. Dennatol. 1984, 83, 91-95 Sinigaglia, F., Scheidegger, D., Garotte, G., Schepter, R., Pletscher, M. and Lanzavecchia, A., Isolation and characterization of Ni-specific T cell clones from patients with Ni contact dermatitis, J. lmmunol. 1985, 13s, 3929-3932 Gimenez-Camarasa, J.M., Garcia-Calderon, P., Asensio, J. and De Moragas, J.M., Lymphocyte transformation test in allergic contact nickel dermatitis, Br. J. Dermatol. 1975, 92, 9-15 Grosfeld, J.C., Penders, A.J., de Grood, R. and Verwilghen, L., In vitro investigations of chromium and nickel hypersensitivity with culture of skin and peripheral lymphocytes, Dermatologica (Basel] 1966, 132,189-198 Bravo, I., Carvalho, G.S., Barbosa, M.A. and de Souse, M., Differential effects of eight metal ions on lymphocyte differentiation antigens in vitro, J. Biomed. Mater. Res. 1990, 24, 1059-1068 Rae, T., The haemolytic action of particulate metals, ]. Pathol. 1978, 125, 81-89 Vince, D.G., PhD Thesis, 1989 Warner, G.L. and Lawrence, D.A., Stimulation of murine lymphocyte response by cations, Cell. Immunol. 1986, lol, 425-439 Semmelink,/.M., Klein, C.P.A.T., Vermeiden, J.p.w. and Althuis, A.L., Granuloma and plasma cell formation induced by the subcutaneous implantation of betawhitlockite particles, BiomaterJals 1986, 7, 152-154 Klein, C.P.A.T., Patka, P. and van der Hollander, W., Macroporous calcium phosphate bioceramics in dog femora, a histological study of interface and biodegradation, Biomaterials 1989, 10, 59-62 Meachim, G. and Brooke,-E., The synovial response to intra-articular Co-Cr-Mo particles in guinea pigs, Biomaterials 1983, 4, 153-159 Biomaterials 1992, Vol. 13 No. 11
The Biomaterials Silver Jubilee Compendium
92
This Page Intentionally Left Blank
93
The Biomaterials Silver Jubilee Compendium 323
Laminated three-dimensional biodegradable foams for use in tissue engineering Antonios G. Mikos** Georgios Sarakinos** Susan M. Leite* Joseph P. Vacantit, and Robert Langer*
*Department of Chemical Engineering and **Department of Chemistry, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, MA 02139. tOepartment of Surgery, The Children's Hospital, Harvard Medical School, 300 Longwood Avenue, Boston, MA 02115, USA A novel processing technique is reported to construct three-dimensional biodegradable polymer foams with precise anatomical shapes, The technique involved the lamination of highly-porous membranes of porosities up to 90%. Implants with specific shapes were prepared made of poly(L-lactic acid) and copolymers of poly(oL-lactic-co-glycolic acid) to evaluate feasibility. The biomaterials produced have pore morphologies similar to those of the constituent membranes. The pores of adjacent layers of laminated devices are interconnected, resulting in continuous pore structures. The compressive creep behaviour of multilayered devices is also similar to that of the individual layers. Recent discoveries from our group and others that organs and tissues can be regenerated and reconstructed, using cells cultured on synthetic biodegradable polymers, renders this method useful in creating polymer-cell grafts for use in cell transplantation. Keywords: Biodegradation, poly(L-lactic acid), poly(oL-lactic-co-glycofic acid), tissue engineering Received 12 June 1992; revised 28 August 1992; accepted 31 August 1992
Cell transplantation using synthetic biodegradable polymer substrates was recently proposed as a new means of tissue reconstruction and repair 1' 2 The creation of an autologous implant requires that donor tissue is harvested and dissociated into individual cells. These cells are attached and cultured into a proper polymer scaffold which is ultimately implanted at the desired site of the functioning tissue "~'4. The scaffold is used to mimic its natural counterparts, the extracellular matrices {ECM} of the body. It serves as both a physical support and an adhesive substrate for isolated parenchymal cells during in vitro culture and subsequent implantation. As the transplanted cell population grows and the cells function normally, they begin to secrete their own ECM support. Concurrently, the scaffold continuously degrades and is eliminated as the need for an artificial support diminishes. Biodegradable polymer templates for cell transplantation must be adhesive substrates for cells, promote cell growth and allow retention of differentiated cell function 5. High porosity provides adequate space for cell seeding, growth and ECM production. A uniformly distributed Correspondence to Professor R. Langer. =Present address: A.G. Mikos, Department of Chemical Engineering and Institute of Biosciences and Bioengineering, Cox Laboratory for Biomedical Engineering, Rice University, PO Box 1892, Houston, TX 77251, USA 91993 Butterworth-Heinemann Ltd 0142-9612/93/050323-08
and interconnected pore structure is important, so that an organized network of tissue constituents can be formed. In the reconstruction of structural tissues like cartilage and bone, tissue shape is integral to function, Therefore, these scaffolds must be processable into devices of varying thickness and shape s' 7. We recently reported a that chondrocytes cultured in vitro on to fibrous poly(glycolic acid) for 6 wk yielded cell densities of the same order of magnitude as reported for normal articular cartilage and produced cartilage matrix (i.e. sulphated glycosaminoglycan, collagen) at a high, steady rate. Chondrocytes attached on to porous poly{L-lactic acid) {PLLA) for the same time period grew to half the cellularity of normal cartilage. We also demonstrated that chondrocytes grown on to these biodegradable polymers in vivo for a period of up to 6 months maintained the shape of the original scaffold and resulted in cartilage formation a. To utilize biodegradable and biocompatible polymers in reconstructive or plastic surgery as templates for chondrocyte attachment and transplantation, it is essential to process them into foams resembling the desired implants. Furthermore, for the generation of metabolic organs like liver and pancreas, we need large transplantation devices to accommodate a ceil mass sufficient for functional replacement. However, to date no processing techniques exist to prepare three-dimensional bioBiomaterials 1993, Vol. 14 No, 5
94
The Biomaterials Silver Jubilee Compendium 324
Laminated three-dimensional biodegradable foams: A.G. Mikos et al.
degradable foams with complex and delicate shapes, which limits the potential of organ regeneration by cell transplantation. In this paper, a new processing technique is presented to construct highly porous biomaterials with anatomical shapes by a lamination procedure evaluated by preparing PLLA and poly(DL-lactic-co-glycolic acid) [PLGA) cell transplantation devices.
PROCESSING TECHNIQUE The proposed methodology to process biodegradable polymer foams with anatomical shapes follows three steps: 1. Drawing the contour plot of the particular threedimensional shape. 2. Preparing highly porous biodegradable membranes with the shapes of the contours. 3. Laminating the constituent membranes with the proper order to form a structure with the desirable shape. For example, the sequence of events involved in the preparation of an implant with a nose-like shape is illustrated in Figure 1.
EXPERIMENTAL Materials The polymers poly(L-lactic acid] (PLLA], poly(DI,-lacticco-glycolic) {PLGA} 85/15 and PLGA 50/50 were supplied by Medisorb (Cincinnati, OH, USA). The ratios 85/15 and 50/50 correspond to the copolymer ratio of lactic to glycolic acid. The polymer molecular weights were determined by gel permeation chromatography (Perkin-Elmer, Series 10, Newton Centre, MA, USA) as Mn = 104 800 (Mw/Mn = 1.13) for PLLA, as Mn = 121 100 [Mw/Mn = 1.16) for PLGA 85/15 and a s M n = 82 800 [Mw/Mn = 1.14) for PLGA 50/50. Here, Mn stands for the number average molecular weight and M w for the weight average. Granular sodium chloride [Mallinckrodt, Paris, KY, USA) was ground with an analytical mill (model A-10, Tekmar, Cincinnati, OH, USA). The ground particles were sieved with ASTM sieves placed on a sieve shaker (model 18480, CSC Scientific, Fairfax, VA, USA). Chloroform was furnished by Mallinckrodt. The mercury
Figure I Schematic presentation of the preparation procedure of a biodegradable foam with nose-like shape.
Biomaterials 1993, Vol. 14 No. 5
used in the porosimetry studies was triple-distilled (Bethlehem Apparatus, Hellertown, PA, USA),
Methods Contour mapping For simple geometries with equations f(x, y, z) = 0 (with z ~ 0) describing the surface coordinates, the contours are defined by the family of equations {f(x, y, i Az) - O, i = 0, 1, 2 . . . . J. The parameter Az is equal to the thickness of the membranes to be laminated. For complex shapes, such as that of a human ear, the contours can be defined by sectioning a solid mould into slices of equal thickness. Preparation of porous membranes Biodegradable polymer membranes were prepared with a novel solvent-casting particulate-leaching technique 9. Briefly, 4.5 g of sieved sodium chloride particles (with sizes in the range of 250-500 pro) were added to a solution of 0.5 g polymer in 8 ml of chloroform, and the vortexed dispersion was cast in a 5 cm petri dish. The solvent was allowed to evaporate from the covered petri dish for 48 h. Residual amounts of chloroform were removed by vacuum-drying at 13 Pa for 24 h. For PLLA, the resulting composite membranes were immersed in 250 ml distilled, deionized water at 25~ for 48 h; to leach out the salt, the water was changed every 6 h. For PLGA 85/15 and PLGA 50/50, the composite membranes were immersed in 250 ml distilled, deionized water at 37~ for 96 h; to leach out the salt, the water was changed every 12 h. Afterwards, the salt-free membranes were air-dried for 24 h, vacuum-dried at 13 Pa for 48 h, and stored in a desiccator under vacuum until use. Membranes of PLLA were also prepared using half of the above quantities. The membranes cast with 0.25 g polymer, 2.25 g NaC1 and 4 ml of chloroform were designated as PLLA/2. The membranes prepared with PLLA were semicrystalline, with a degree of crystallinity of 24.5%, measured by differential scanning calorimetry [7 Series, Perkin-Elmer, Newton Centre, MA, USA). The membranes of the copolymers PLGA 85/15 and 50/50 were amorphous. Preparation of non-porous films Polymer films were prepared and used to measure the contact angle between the different polymers and mercury. In a typical experiment, a solution of 0.5 g polymer in 4ml of chloroform was cast in a 5 c m petri dish. The chloroform evaporated from the covered petri dish in a fume hood over 48 h before vacuum-drying at 13 Pa for 24 h. The produced films were stored in a desiccator under vacuum until use. Lamination of porous membranes Porous biodegradable membranes were cut to the contour shapes. A small quantity of 2-:3 mt of chloroform was poured on to a few paper tissues (Kimwipes, Kimberly-Clark, Roswell, GA, USA) to wet them sufficiently. Two membranes to be joined were placed on the wet tissues. Light pressure was applied on their top faces with the aid of forceps for 4--5 s, to ensure that they
95
The Biomaterials Silver Jubilee Compendium Laminated three-dimensional biodegradable foams: A.G. Mikos et a/.
became adequately wet. Subsequently, one membrane was carefully removed and inverted on a dry surface, wet face up. The other membrane was quickly picked up with forceps and was placed carefully on top of the first, the wet faces touching. Then, the resulting structure was squeezed gently, to ensure that the two portions were well glued together. Every additional membrane layer was attached to the incomplete three-dimensional shape by following the procedure just described, until the desired structure was obtained. The final laminated structure was vacuum dried at 13 Pa for 24 h to ensure complete solvent evaporation. This method was used to prepare foams with anatomical shapes. For the characterization of laminated structures, circular discs of diameter 13.45 mm were cut with the aid of a cork borer and were laminated to make devices of two and three layers, which were tested by the experimental techniques described below.
Characterization Scanning electron microscopy The samples were coated with gold using a sputter-coater (model Desk II, Denton Vacuum, Cherry Hill, NJ, USA). The gas pressure was set at 7 Pa and the current was 40 mA for a coating time of 75 s. A Hitachi (model S-530) scanning electron microscope was used in our studies and operated at 15 kV.
Mercury porosimetry The pore size distribution of membranes and laminated devices of two and three layers of PLLA, PLLA/2, PLGA 85/15 and PLGA 50/50 was determined by mercury porosimetry (model Poresizer 9320, Micromeritics, Norcross, GA, USA}. A solid penetrometer with 5 cm 3 bulb volume {model 920-61707-00, Micromeritics, CA, USA} was used with samples in the range 0.02-0.1 g. The values of void volume and pore area were calculated from measurements of the mercury intrusion volume at different pressures 1~ 11. The filling pressure, P=i,, of the penetrometer was 3.45 kPa and the maximum pressure was 207 kPa. At 207 kPa, the total intrusion volume had reached a plateau value. Thus, the reported values of cumulative void volume and pore area refer to pores with diameters smaller than that given by the Washburn equation1~ d _~.
-47cos0 Pmin
C1}
where y is the surface tension of mercury (485 dyn/cm at 25~ and 0 is the advance contact angle between the mercury and the pore wall. The diameter calculated from Equation 1 corresponds to an equivalent cylindrical pore and provides an approximation of the interstitial distance. The values of 0 were measured at 25~ by a contact angle goniometer {model 100-00, Rome-Hart, Mountain Lakes, NJ, USA} using flat films of the polymers. They were 160 ~ for PLLA, 132 ~ for PLGA 85/15 and 135~ for PLGA 50/50. Then, the values of the diameter d calculated from Equation I were 529 pm for PLLA, 377 pm for PLGA 85/15 and 398 pm for PLGA 50/50. The polymer densities used to calculate the porosity
325
from the total intrusion volume were measured by micropycnometry {model Accupyc 1330, Micromeritics} as 1.26 g/cm 3 for PLLA, 1.29 g/cm 3 for PLGA 85/15 and 1.35 g/cm 3 for PLGA 50/50. The porosity of the membranes and the laminated devices was also calculated from measurements of their surface density and their thickness. The thickness was measured with a micrometer, with accuracy +10pm. The surface density, in units of mass per area, was estimated from the weight of single-layered or multilayered discs of diameter of 13.45 mm.
Thermomechanical analysis A 7 Series thermal analysis system (Perkin-Elmer} was used for compressive creep studies of the laminated devices at 37~ A constant force of 100 mN was applied to a sample using a quartz expansion probe with a circular base of 3.66 mm diameter for 60 rain, yielding a compressive stress of 9.5 kPa. The force was then removed, and the sample was allowed to recover for 30 rain. The sample strains at 60 and 90 min had reached a plateau value and were used to evaluate the creep behaviour of laminated devices versus that of the constituent layers.
RESULTS AND D I S C U S S I O N Highly porous membranes were produced by the solventcasting particulate-leaching technique. The physical properties of these membranes are presented in Table 1. The membrane thickness depended on the polymer itself and also on its initial amount. The thickness of the PLLA/2 membranes was approximately half that of the PLLA ones, prepared with twice as much polymer. The porosity of all membranes measured by mercury porosimetry was smaller than that calculated from surface density. The difference between these two values corresponds to the fraction of pores with diameter larger than that calculated by Equation 1 which was not detected by mercury porosimetry. The difference was minimal for PLLA {and PLLA/2} and maximal for PLGA 85/15 characterized with the largest {d = 529 pm} and the smallest diameter (d = 377 pm}, respectively. In addition to the void volume, the pore area and median pore diameter were examined by mercury porosimetry for single membranes and for laminated devices of two and three membranes. A decrease was observed in the values of the void volume {Figure 2a} and pore area {Figure 2b}, and was explained by the dissolution of porous polymer in the vicinity of the glued surfaces followed by its solidification during solvent evaporation. By measuring the thickness of laminated devices (Figure 3}, we found that there was a small decrease in the total thickness of two-layer and threelayer devices compared to the sum of the values of the original layers. This total decrease was dependent on the particular polymer composition and scaled to the number of laminated layers. Nevertheless, the decrease of the void volume and pore area diminished as the number of laminated layers increased. No appreciable weight decrease was recorded during each lamination (Table 1 }and the surface density, defined as weight per unit area, was proportional to the number Biomaterials 1993, Vol. 14 No. 5
96
The Biomaterials Silver Jubilee Compendium 326
Laminated three-dimensional biodegradable foams: A.G. Mikos et at.
Rgure3 Thickness decrease due to the lamination pro-
cedure of devices of two and three layers for the four types of foams shown. The error bars correspond to average +_the standard deviation of five samples. II, Two layers; B, three layers.
Figure 2 a, Void volume; b, pore area; and c, median pore diameter of devices of one, two and three layers for the four types of foams shown, measured by mercury porosimetry. For PLLA, the error bars designate average + the range of two measurements; for the other foams, the reported values correspond to single measurements. II, One layer; ~, two layers; I~, three layers. Biomaterials 1993, Vol. 14 No. 5
of laminated layers. Thus, we inferred that the polymers close to the joined surfaces had dissolved in the presence of chloroform vapours and intermingled upon contact. Furthermore, this decrease of the void volume was not large enough to affect the calculated values of the porosity significantly. For example, for PLGA 50/50, the decrease of the void volume from 4930 cmZ/g [one layer) to 3360 cm2/g {two layers) resulted in a decrease of the porosity from 0.84 to 0,77. Also, no trend of increase or decrease in the median pore diameter could be identified (Figure 2c). Finally, by inspecting the cumulative size distribution of PLLA-laminated devices shown in Figure 4, we concluded that the pore morphology of multilayered devices was similar to that of the constituent layers. These findings are crucial to the evaluation of the lamination technique for creating thick implantation devices. The lamination procedure is useful only if it preserves the uniform porous structure of the original foams. The boundary between each two layers must be indistinguishable from the bulk of the device. These results show that bulk properties of the devices are indeed preserved during the lamination procedure. Scanning electron microscopy [SEM) photomicrographs of a PLLA three-layer laminated device and one of its layers before lamination are shown in cross-section in Figure 5. For this device, no discontinuity of the porous structure was observed, nor could the junctions between adjacent devices be located. This is also true for a PLGA 85/15 three-layer laminated device, an SEM photomicrograph of which is shown in Figure 6 together with that of one of its constituent layers. These photomicrographs further support the gentleness of this lamination process. In addition, from cell seeding studies 4 into similar
The Biomaterials Silver Jubilee Compendium
97
Laminated three-dimensional biodegradable foams: A.G. Mikos et a/. Table I
327
Properties of highly porous membranes and two-layer laminated foams _
Polymer type
PLLA PLLA/2 PLGA 85/15 PLGA 50/50 -
Thickness* (pro)
Surface density* (mg/cm 2)
_
_.
Foam
Membrane
.
Weight
decrease*
Porosity**
Porosity t
Thickness* (pro)
Surface density* (mg/cm 2)
Porosity**
Porosity t
1727 • 92 27.1 + 1.1 904 _+ 84 12.5 + 0.5 1297 +_ 119 24.0 + 1.4
0.88 0.89 0.86
0.83 0.84 0.64
3168 + 223 1596__+85 2427 + 258
51.1 + 2.2 23.9+0.7 48.8 + 1.8
0.87 0.88 0.84
0.82 0.83 0.58
2.1 + 0.4 0.3 + 0.1 0,0 + 0.5
1867 +_ 185 26.3 + 2.0
0.90
0.84
3269 __+274 52.6 + 4.5
0.88
0.77
-0.1 + 0.1
_
(mg/cm 2)
......
*Average value + standard deviation of five samples.
**As measured from the surface density. tAs measured by mercury porosimetry.
-E
#
E
3
O1
U
:3
O > "13
>
10
100
1000
Pore diameter (~m] Figure 4 Cumulative pore size distribution of PLLA devices of one, two and three layers, measured by mercury porosimetry. 13, One layer; O, two layers; ,4, three layers.
multilayered devices, it was evident from the unhindered transport of fluids and cell suspensions across the interface of adjacent layers that the communication from layer to layer was not obstructed by the lamination process and that the interconnected pore structure was preserved. The effect of the lamination process on the creep behaviour of the polymer devices was evaluated by thermomechanical analysis. Figure 7a shows the strain measured for each device after 60 min of loading with 9.5 kPa of compressive stress. The strain measured for each of the PLLA, half-thickness PLLA (PLLA/2) and PLGA 85/15 devices was about 0.1, while the strain measured for the PLGA 50/50 devices was in the range 0.4-0.5, The number of membranes which made up the device had no effect on the measured strain. After 60 min, the stress was removed and the devices were allowed to recover unloaded for 30 min. Figure 7b shows the strain after the recovery period. For each material, nearly 50% of the deformation was recovered. In addition, the rates of strain change during a compressive creep cycle were identical for devices with different numbers of layers (Figure 8). These results are encouraging, in that once
again there was no observed effect of the lamination process on the properties of the membranes. The lamination process did not cause a weakening of the polymer foam in response to compressive forces. Laminated devices with anatomical shapes were constructed for potential use in reconstructive or orthopaedic surgery. The implant with a nose-like shape (Figure 1) was created by lamination of six PLLA membranes of an average thickness of 1727 pm and total porosity of 88%. A photograph of the implant is shown in Figure 9. We also processed laminated foams from similar PLLA membranes with shapes of metacarpalphalangeal pieces for joint repair (Figure 10). The head of each foam was prepared by lamination of four layers (discs) with orientation perpendicular to the axis of symmetry of the hemisphere and the stem was created separately by lamination of two strips. The two pieces were joined together to form a foam with the desired pinlike shape. Transplantation devices were also prepared for a variety of cells. Examples include devices for hepatocyte transplantation 12shown in Figure 11. Here, the lamination procedure was necessary to produce thick devices to accommodate a large number of hepatocytes for functional replacement. The devices were made of three layers with a catheter inserted in the centre of each device as a route for injection of hepatocytes into the bulk of the polymer. No delamination or failure of any devices due to the development of shear stresses from surrounding tissues was detected 12 from histological sections of a large number of harvested devices implanted in the mesentery of rats for a period of 35 d. The distribution of cells seeded into these devices via injection was recently modelled by our group 4 to maximize the device volume effectively employed in cell transplantation and determine the optimal surgical injection conditions. In conclusion, a new method was developed to laminate highly porous membranes and produce threedimensional foams with continuous pore structure and morphology. This method can be used to process biodegradable polymers into custom-made shaped devices with potential use in cell transplantation. With the aid of computer-assisted modelling to contour-map tissues and organs 13, we can easily construct templates with the desired implant shape. With further study of: (1] the procedures to uniformly seed large devices with cells, (2) the cell and tissue culture techniques to eliminate any mass transfer limitations of nutrients to the whole cell Biomaterials 1993, Vol. 14 No. 5
98
The Biomaterials Silver Jubilee Compendium 328
Laminated three-dimensional biodegradable foams: A.G. Mikos et al.
Figure 5 SEM photomicrographs of cross-sections of: a, three-layer laminated PLLA foam; and b, one of its constituent layers before lamination.
Figure 6 SEM photomicrographs of cross-sections of: a, three-layer laminated PLGA 85/15 foam' and b, one of its constituent layers before lamination. Biomaterials 1993, Vol. 14 No, 5
The Biomaterials Silver Jubilee Compendium
99
Laminated three-dimensional biodegradable foams: A.G. Mikos et al.
329
Figure8 Creep behaviour of PLLA (open symbols) and PLGA 50/50 (filled symbols) devices of N, !:3, one layer; O, O, two layers; and &, A, three layers as measured by thermomechanical analysis at 37~ At time zero, a constant compressive stress of 9.5 kPa was applied for 60 min. Afterwards, the stress was removed and the sample was allowed to recover for an additional time of 30 min. The strain is defined as the thickness change divided by the initial thickness.
Rgure9 Photograph of a laminated PLLA foam with noselike shape.
Figure 7 Compressive strain of devices of one, two and three layers for the four types of foams shown: a, after 60 rain under a stress of 9.5 kPa and b, after 30 min of recovery from the previous stress, measured by thermomechanical analysis at 37~ B, One layer; t~, two layers; I~, three layers.
mass and ensure the viability, growth and function of the attached cells, and (3} the surgical approaches to implant polymer-cell devices to the sites of the functioning tissues, this method could become important in the development of novel cell-based artificial organs for clinical use.
Figure lO Photograph of metacarpal-phalangeal pieces made of laminated PLLA foams (A, B) similar to a nondegradable medical implant (C).
Biomaterials 1993, Vol. 14 No. 5
100
The Biomaterials Silver Jubilee Compendium Laminated three-dimensional biodegradable foams: A.G. Mikos et al.
330
4
Wald, H,L., Sarakinos, G., Lyman, M.D., Mikos, A,G., Vacanti, ].P. and Langer, R., Cell seeding in porous transplantation devices, Bioraaterials [in press} Cima, L,G., Ingber, D.E., Vacanti, ].P. and Langer, R., Hepatocyte culture on biodegradable polymeric substrates, BiotechnoL Bioeng. 1991, 38, 145-158 Vacanti, C.A., Langer, R., Schloo, B. and Vacanti, ].P., Synthetic polymers seeded with chondrocytes provide a template for new cartilage formation, Plast. Reconstr. Surg, 1991, 88, 753-759 Cima,L.G., Vacanti, ].P., Vacanti, C., Ingber, D., Mooney, D. and Langer, R., Tissue engineering by cell transplantation using degradable polymer substrates, ]. Biomech. Eng. 1991, 113, 143-151 Freed, L.E., Marquis, I.C., Nohria, A., Emmanual, 1., Mikos, A.G. and Langer, R., Neocartilage formation in vitro and in vivo using ceils cu(tured on synthetic biodegradable polymers, ]. Biomed. Mater. Res. 1993, 27, 11-23 Mikos, A,G., Thorsen, A.]., Czerwonka, L.A., Bao, Y., Winslow, D.N., Vacanti, ].P. and Langer, R., Preparation and characterization of poly[L-lactic acid) foams for cell transplantation, Polymer {submitted} Winslow, D.N., Advances in experimental techniques for mercury intrusion porosimetry, in Surface and Colloid Science {Eds E. Matt}eric and R.]. Good}, Plenum Press, NY, USA, 1984, pp 259-282 Tsakiroglou, C.D. and Payatakes, A.C., A new simulator of mercury porosimetry for the characterization of porous materials, ]. Colloid Interface Sci. 1990, 137, 315-339 Mikos, A.G., Sarakinos, G., Lyman, M.D., Ingber, D.E., Vacanti, ].P. and Langer, R., Prevascularization of porous biodegradable polymers, Biotechnol. Bioeng. [in press] ]imenez, ]., Santisteban, A., Carazo, ].M. and Carrascosa, I.L., Computer graphic display method for visualizing three-dimensional biological structures, Science 1986, 232, 1113-1115
5 6
7
Figure 11 Photograph of hepatocyte transplantation devices made by lamination of three layers of PLLA (A), PLLAI2 (B), PLGA 85/15 (C) and PLGA 50/50 (D), with a catheter positioned in the middle of each device.
8
9
ACKNOWLEDGEMENTS Many thanks to Ms Michelle D. Lyman for excellent technical assistance. This work was supported by a grant from Advanced Tissue Sciences.
10
REFERENCES
i1
1
2 3
Vacanti, ].P., Morse, M.A., Saltzman, W.M., Domb, A.|,, Perez-Atayde, A. and Langer, R,, Selective cell transplantation using bioabsorbable artificial polymers as matrices, ]. Pediatr. Surg. 1988, 23, 3-9 Vacanti, ].P., Beyond transplantation, Arch. Surg. 1988, 123, 545-549 Mikos, A.G., Bao, Y., Cima, L.G,, Ingber, D.E., Vacanti, I.P. and Langer, R., Preparation of poly{glycolic acid] bonded fiber structures for cell attachment and transplantation, ]. Biomed. Mater. Res. 1993, 27, 183-189
.....
~
12
13
9
UJII
-
__
.
.
.
.
.
. . . .
I
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
- - ~ L
-
-
Biomedical Materials & Technologies
Biocompatibility of Medical Devices
Centro de Citologia Experimental, Porto, Portugal 28-30 June 1993 This course aims to give a comprehensive survey of current knowledge in the fields of biomaterials and biocompatibility, and will appeal to a multidisciplinary audience. The structure and properties of biornaterials will be reviewed, and particular attention paid to their uses in orthopaedics, dentistry and cardiovascular surgery. The biocompatibility of different materials will be fully examined. The interactions of host cells with biomaterials will be described and discussed, with emphasis being placed on the importance of biocompatibility to implant survival.
For further information and registration details please contact: COMETT Course Secretary, Department of Clinical Engineering, University of Liverpool, PO Box 147, Liverpool L69 3BX, UK. Fax" + 44 051 706 5803 _ .
_
-
9
, ......
,
,,.
J
L,J
- -
Biomaterials 1993, Vol. 14 No. 5
,,,
.
.
.
.
.
.
,, ,,~,,,__..,
101
The Biomaterials Silver Jubilee Compendium
.B U . T . T. E.R .W. O. R T H E I N E M A N
~
Biomaterials16 (1995)25-31 9 1995 ElsevierScience Limited Printed in Great Britain. All rights reserved 0142-9612/95/$10.00
N
Late degradation tissue response to poly(L-lactide) bone plates and screws J.E. Bergsma, W.C. de Bruijn*, F.R. Rozema, R.R.M. Bos and G. Boering Department of Oral and Maxillofacial Surgery, University Hospital Groningen, PO Box 30.001, 9700 RB Groningen, The Netherlands; *AEM Unit, Clinical Pathological Institute I, Erasmus University of Rotterdam, The Netherlands
Patients with fractures of the zygomatic bone were treated with high molecular weight poly(L-lactic) acid (PLLA) bone plates and screws. Three years after implantation four patients returned to our department with a swelling at the site of implantation. At the recall of the remaining patients we found an identical type of swelling after the same implantation period. To investigate the nature of the tissue reaction, eight patients were reoperated for the removal of the swelling. The implantation period of the PLLA material varied from 3.3 to 5.7 years. Microscopic evaluation and molecular weight measurements were performed. The excised material showed remnants of degraded PLLA material surrounded by a dense fibrous capsule. Ultrastructural investigation showed crystal-like PLLA material internalized by various cells. The results of this investigation suggest that the PLLA material slowly degrades into particles with a high crystallinity. The intra- and extracellular degradation rate of these particles is very low. After 5.7 years of implantation, these particles were still not fully resorbed. Biomaterials (1995), 16 (1), 25--31 Keywords: Poly (L-lactide), biodegradation, tissue response, enzyme activity
Received 28 November 1993; accepted 25 April 1994
In a study on rats the biocompatibility and degradation characteristics of PLLA were investigated TM. The histological reaction to the implanted PLLA material was very mild: only a slight foreign body reaction was observed after a follow-up of 2.8 years. The implanted PLLA material showed a rapid decrease of molecular weight but only a small mass loss. Total resorption of the PLLA material was not observed in this study on rats, but was estimated to be about 3.5 years. Based on the positive results in animal studies, a limited investigation in humans was set up. PLLA bone plates and screws were used for the fixation of unstable zygomatic fractures 17. Three years after implantation, four patients returned to our department spontaneously because of a swelling at the site of implantation TM. Another five patients were recalled and all showed identical swellings. The aim of this study is to characterize the remainder of the PLLA material after an implantation period of 3.3 and 5.7 years, in order to gain an insight to the nature and course of the swelling at a light microscopical, ultrastructural and cytochemical level.
Metallic bone plates and screws are commonly used in oral and maxillofacial surgery for internal fracture fixation. Although good fracture healing is obtained, the disadvantages of metallic plates and screws are the possibility of bone atrophy due to stress-shielding and the obligation to remove these devices in a second operation 1-3. Bone plates and screws made of a biodegradable material are considered to be a good alternative for metallic ones. The main advantage of biodegradable plates and screws is that they lose their mechanical properties due to degradation so that loads are gradually retransferred to the bone, preventing stress-shielding of the healed bone. Moreover, if the material fully degrades, a reoperation for the removal of the plate and screws can be avoided 4-6. Because of these advantages, biodegradable polyesters such as poly(L-lactide) or polyglycolide have been studied extensively during the last two decades. These biodegradable polyesters have been used as internal fixation devices in the shape of rods, screws and bone plates 7-9. At our department high molecular weight aspolymerized poly(L-lactide) (PLLA) has been a material of special interest because of its gaod mechanical properties ~~ To gain insight into the mechanical behaviour during degradation, PLLA was used for fracture fixation of the mandible of dogs 13 and sheep 14, and for orbital floor reconstructions in goats 15. In all cases the PLLA plates gave sufficient stability to enable undisturbed fracture healing 13-15.
MATERIALS AND METHODS
Patients From 1986 to 1988, ten patients (mean age 39.6 yr; range 20.2-61.8 yr) with solitary displaced unstable fractures of the zygoma were treated with high molecu-
Correspondence to Dr J.E. Bergsma. 25
Biomaterials 1995, Vol. 16 No. 1
102
The Biomaterials Silver Jubilee Compendium Late degradation tissue response to poly(L-lactide) bone plates and screws' J.E. B e r g s m a
26
lar weight as-polymerized poly0.-lactide) bone plates and screws (My = 1 x 106, calculated using the formula [~1] = 5.45 • 104 l~//~73; and Mn = 7.6 x 105, calculated using the formula [77]= 3.25 x 104 ~'~n0"77).Unfortunately one patient died of a cerebrovascular accident one year after implantation. Three years after implantation, four out of nine patients returned to our department at their own initiative because of a painless swelling at the site of implantation TM. At a recall, the remaining five patients showed a similar swelling. In a period of two years, seven patients agreed to have a reoperation for the exploration of the area of swelling. The postoperative implantation period varied from 3 years 4 months to 5 years 8 months. The reoperation was performed under general anaesthesia. The tissue in the area of the swelling was excised via an incision laterally in the eyebrow. Samples of the screw-holes were taken by trephination of the bone.
Characterization of the degraded PLLA From a part of the excised tissue the remainder of the PLLA material was mechanically removed and was subsequently treated with trypsin 2.5% in Hank's balanced salt solution and collagenase ]a for the removal of organic components. The PLLA particles were then dried under vacuum for 17 h at 1 0 -3 bar to constant weight. Nuclear magnetic resonance (NMR) measurements for the determination of the molecular weight were carried out on a Varian 300 NMR spectrometer. The 1H NMR spectra were obtained from polymer solutions in deuterated chloroform in 5 m m tubes. For scanning electron microscopy (SEM) analysis, dried PLLA particles were gold-palladium sputter-coated and photographed in a DS 130 (ISI) scanning electron microscope.
Histological procedures Slices 2 m m thick were cut perpendicular to the long axis of the excised tissue mass and fixed in 2% v/v glutaraldehyde in 0.1 M phosphate buffer of pH 7.4 for at least 48 h at 4~ For light microscopy, sections were dehydrated in graded series of ethanol. The sections were embedded in glycol-methacrylate, polymerized for 24 h at -20~ Sections of 2 pm were made (]ung microtome 1140/autocut), which were stained with toluidine blue and basic fuchsin. For electron microscopy, postfixation was performed with 1 wt% OsO4 to which K4Fe(CN)~. 3H20 was added to a final concentration of 0.05 M. Subsequently, the material Table 1
et al.
was dehydrated in series of 70, 80, 90 and 100% acetone. The material was embedded in LX 112 epoxy resin and polymerized for 24 h at 60~ Based on light microscopic observations, ultrathin 70 nm sections were acquired at selected sites, around the screw-head and at the periphery of the bone plate. These ultrathin sections were stained with uranyI acetate/lead citrate. For transmission electron microscopy a Zeiss EM 902 was used, operating at 80 kV.
Histochemical procedures To investigate possible enzymatic activity towards the membrane-bound PLLA conglomerates, cytochemical reactions were performed on the material implanted for 5.7 years. For the demonstration of acid phosphatase and alkaline phosphatase, aldehyde-fixed tissue was used. For the demonstration of lactate dehydrogenase (LDH), fresh 5 m m tissue slices were quickly frozen at -80~ Perpendicular to the bone plate axis, sections were cut of 50 gm thickness in a cryostate microtome at -28~ The histochemical methods for the enzymes investigated are summarized in Table 1.
RESULTS Material characterization The mean number molecular weight (~/,) of the PLLA particles as determined by NMR was respectively 5600 and 5400 for the 3.3 and 5.7 years implanted PLLA. The plates and screws were machined out of a block of as-polymerized with an N/n of 7.6 X 105. Ultrastructural SEM examination of the 3.3 years material revealed particles varying in size from 1 to 1500 pm (Figure 1). The larger PLLA fragments showed numerous whitish cracks. Many smaller particles seemed to be adhered to the surface. Higher magnification of a particle showed an irregular surface structure. The SEM appearance of the material implanted for 5.7 years showed at some parts a microporous structure, again with many smaller particles attached to it. The mean particle size seemed to be smaller compared with those of the 3.3 years implanted material.
Histological analysis The material excised after an implantation period of 3.3 years showed a firm consistency and the contours of some of the screw-heads could still be seen. Light
Histochemical methods for the enzymes investigated. .
.
.
.
.
.
.
.
Enzyme
Substrate, cofactors, coupling agent
Buffer and pH
Incubation (time, temp)
Reference
Acid phosphatase
1.5 mg ml-1 Sodium-~-glycerophosphate 1 mM Cerium chloride
0.08 M Tris-mateate pH 5.0
30 min, 37~'C
Hulstaert et al. 2~
Alkaline
1.5 mg m1-1 Sodium-,8-glycerophosphate 1 mM Cerium chloride 4 mM Magnesium chloride
0.1 M Tris-maleate pH 9.2
30 min, 37:C
Hulstaert et al. 2~
Lactate dehydrogenase
2.5 mg m1-1 Lactic acid lithium salt 0.05 M potassium ferricyanide 0.5 mg m -~ NAD +
0.1 M Phosphate buffer pH 7.2
60 min, 37~C
Hanker et al. 21
phosphatase
(LDH) .
..
9
B i o m a t e r i a l s 1995, Vol. 16 No. I
,
.
,,
103
The Biomaterials Silver Jubilee Compendium Late degradation tissue response to poly(L-lactide) bone plates and screws: J.E. Bergsma et a/.
27
Figure 1 Scanning electron micrograph of the remnants of the 3.3 year implanted poly(L-lactic) acid material (scale bar = 500/zm, 40 kV).
microscopic analysis of this material revealed a fibrous tissue capsule that enveloped large areas of foreign body material. Under crossed Nicol prisms the areas with the foreign body material were brilliantly birefringent representing the PLLA material. At low magnification of a section through a screw-head and bone plate, areas with densely packed birefringent PLLA-like material were seen. These areas were separated by fibrous tissue that varied in thickness throughout the section. In parts where the capsule measured about 150 #m, a sharp boundary between the closely packed PLLA material and the fibrous tissue was seen (Figure 2a). Here, no birefringent material was situated in the capsule. These parts of the capsule, with a sharp interface PLLA/fibrous tissue, were investigated ultrastructurally. Transmission electron microscopy (TEM) revealed densely packed foreign body material with a lamellar or needle-like structure in close contact with orientated bundles of collagen fibres (Figure 2b). These non-electron dense needle-like particles represented the PLLA material. Between these fibres long slender cells were present that could be characterized on morphological grounds to be fibrocytes. Virtually no other ceils like macrophages, foreign body giant cells or lymphocytes were seen in this area. No PLLA was situated in the cytoplasm of cells or in the extracellular space between the collagen fibres. Light microscopically it was observed that in other parts of the section the fibrous tissue spread out, and PLLA particles were seen between bundles of collagen and cells. In these parts the cells most frequently present were long slender fibrocytes. Only a small number of macrophages or lymphocytes were seen. Electron microscopy revealed that fibrocytes possessed well developed organelles like rough endoplasmic reticulum and golgi apparatus. In a number of cells with internalized PLLA, a clear deposition of glycogen around phagosomes was seen. In the plasma membranes of the fibrocytes a high number of endocytotic vesicles was observed. Intracellularly, fibrocytes showed a profuse amount of PLLA material which was mostly packed in membrane-b~ vacuoles which could be described as phagosomes (Figure 3). Fusion
Figure 2 a, Micrograph of the 3.3 year implanted material, taken under crossed Nicols prisms, of the fibrous capsule with centrally removed poly(L-lactic) acid (PLLA) particles (RP). The arrows indicate a part of the capsule with a sharp interface PLLA/fibrous tissue. In other parts birefringent PLLA particles (P) were situated between bundles of collagen (C) (original magnification x40). b, Transmission electron microscopic (TEM) photograph of the 3.3 year implanted material. The arrows indicate a sharp interface of densely packed needle-like PLLA material (P) and orientated bundles of collagen fibres (C). These orientated bundles of collagen are intermingled wffh fibrocytic (F) cells (scale bar = 2.5 #m).
of a lyosome with a packed vacuole forming a phagolysosome was observed only infrequently. In a small number of fibrocytes the incorporated PLLA particles were also situated apparently freely in the cytoplasm. The cells with incorporated PLLA material had swollen parts of endoplasmic reticulum, which could indicate an increased protein synthesis, and swollen mitochondria that lacked the cristae suggesting physiological damage. Although PLLA particles were seen between collagen fibres and cells, the bulk of the PLLA material was still situated extrae~llularly and was not interlaced with collagen fibres and cells. Macroscopical investigation of the 5.7 years material showed a tissue mass which lacked the firm consistency of the 3.3 years material. The contours of some of the screw-heads were still visible. Microscopic examinatiort at low magnification showed a sharp outline of a thin fibrous capsule. At the peripheral Biomaterials 1995, Vol. 16 No. 1
104
The Biomaterials Silver Jubilee Compendium 28
Late degradation tissue response to poly(L-lactide) bone plates and screws: J.E. Bergsma et al.
Figure 3 Transmission electron micrograph of a fibrocyte with glycogen accumulation (arrows) around phagosomes and swollen mitochondria (M). The cell is situated in fields of poly(L-lactic) acid particles (P) and sheets of collagen (C) (scale bar = 1.7 #m).
Figure 4
Birefringent poly(L-lactic) acid particles intraceliularly in foamy macrophages (arrows), 5.7 years after implantation (original magnification • taken under crossed Nicol prisms).
parts of the screw-head region, blood capillaries, nerve fibres and fat deposition were observed. More centrally in the screw-head region, randomly orientated bundles of collagen were seen amidst various kinds of cells. In this section, no large areas of densely packed extracellular PLLA particles were found. A section through the bone plate showed large fields of cells with intracellularly positioned birefringent PLLA material (Figure 4). This section was composed of mainly foamy macrophages surrounded by fibrous tissue. Electron microscopic observations revealed that most of the PLLA material was situated intracellularly in various cells. These phagocytizing cells formed clusters that were encapsulated by mature and randomly orientated bundles of collagen. The number of elongated fibrocytic cells with internalized PLLA material had diminished as compared to the situation after 3.3 years. The number of macrophages and foreign body giant cells had increased and represented the Biomaterials 1995, Vol. 16 No. 1
Figure 5 Membrane-bound conglomerates of poly (L-lactic) acid particles (arrows) described as phagosomes in the cytoplasm of a phagocytic cell. All cells are embedded in a mature fibrous capsule (C) (scale bar - 2.5/~m).
major phagocytizing cellular component. In these cells the PLLA particles were no longer found freely in the cytoplasm, but entirely as membrane-bound conglomerates (Figure 5). The morphology of phagocytizing cells showed minimal signs of cell damage. The cytoplasmic organelles, such as mitochondria and rough endoplasmic reticulum, were of normal appearance. A sample of trephined bone was obtained from a patient after an implantation period of 5.7 years. The tapped screw-holes were still visible and not fully filled in with cortical bone. On a section perpendicular to the long axis of the screw-hold, birefringent PLLA particles were still densely packed in the shape of the screw-thread. These densely PLLA particles were not interlaced with collagen fibres or cells. A fibrous capsule was situated between cortical bone and the bulk of the PLLA particles and had at some parts spread out into the lacunae of the cortical bone (Figure 6). Fields of PLLA particles were seen up to 0.5 mm from the original implant site showing birefringent PLLA material between sheets of collagen and in various cells. Ultrastructural investigation showed that the PLLA material had the same lamellar or needlelike structure as observed in the soft tissue. As an indicator of cell damage or high lactic acid concentrations, possibly released from the PLLA particles, the presence of lactate dehydrogenase (LDH) was investigated. The presence of LDH in mitochondria was demonstrated by a cytochemical reaction: Hatchett's brown depositions amplified by treatment with 3,3'-diaminobenzidine (DAB) and osmication were seen in close relation to the cristae, the intracristate spaces and the intermembrane spaces of mitochondria (Figure 7). There was no evidence of LDH-related precipitates in the membrane-bound conglomerates, or in close contact with the PLLA particles; nor were extracellular LDH-related precipitates, possibly released by damaged or lytic cells demonstrated. Control specimens, without nicotinamide adenine dinucleotide (NAD) and/or DAB/osmium showed no Hatchett's brown depositions in mitochondria. Acid phosphatase could be demonstrated in lysosomes in a
105
The Biomaterials Silver Jubilee Compendium Late degradation tissue response to poly(L-lactide) bone plates and screws: J.E. Bergsma et al.
Figure 6 Densely packed poly(L-lactic) acid (PLLA) particles (P) in the shape of the screw-thread surrounded by cortical bone (B). Fields of PLLA particles were seen at some distance of the original implant in the cortical bone (P') (original magnification x40).
Figure 7 Hatchett's brown depositions amplified by treatment with 3,3'-diaminobenzidine and osmication were seen in close relation to the cristae, the intracristate spaces and the intermembrane spaces of mitochondria (scale bar = 0.25/~m).
limited number of macrophages that had internalized PLLA particles. Some of these lysosomes were seen in close contact with the PLLA-bearing phagosomes. Fusion of a primary lysosome and a phagosome forming a phagolysosome was rarely seen. DISCUSSION The total resorption time of as-polymerized PLLA was estimated in previous studies to be 3.5 years 15' 16. The results of this experiment show that the PLLA bone plate and screws, implanted for 3.3 years, had degraded into fragments and disintegrated into particles that have a needle-like structure on TEM. Ultrastructural TEM analysis of the PLLA material with an implantation period of 5.7 years shows a comparable morphology. SEM analysis would suggest that the average particle size of the materia] implanted
29
for 5.7 years is much smaller. Between 3.3 and 5.7 years the PLLA material degrades from fragments into particles that have a needle-like structure an TEM. Light microscopic observations suggested that the number of PLLA particles that were internalized by cells had increased with longer implantation periods. The molecular weight, about 5000, is identical for both implantation periods. Rozema 21 described that an M, of 5000 may be a break-even point as a start of relative high disintegration. However, the PLLA particles have a rather high crystallinity 21 which is probably one of the factors that makes them very stable and not very susceptible to hydrolysis. This may explain the very limited progression of the degradation of PLLA particles in the period from 3.3 to 5.7 years. Substantial mass loss or total resorption had not taken place up to 5.7 years. If a PLLA particle degrades, it is probably in non-detectable oligomers that are washed away with tissue fluids and are not detected in the material analysis. This mechanism may account for the same values of molecular weight and crystallinity for both implantation periods. The origin of the described swelling is not quite clear. Maybe the swelling is initiated by a gradual disintegration of the PLLA bone plate and screws into fragments. Bergsma et al. TM described how during degradation the PLLA plates and screws disintegrate into small fragments which may lead to an increased volume in comparison with the volume of the intact bone plate and screws. In a cross-section of tissue implanted for 3 years, the surface area occupied by the a-cellular PLLA particles was estimated to be 65% of the total surface area. The remaining 35% of the cross-section was occupied by the enveloping fibrous capsule. B6stman eta]. 22, in a study with intraosseously placed polyglycolide screws and pins, suggest that an increased osmotic intracavital pressure associated with the degradation of polyglycolide and the resistance of the surrounding tissue may determine the formation of a sinus. The origin of the described swelling may possibly be explained by a combination of factors such as the disintegration of the PLLA material into small particles, and an increased osmotic pressure caused by these fragments and the, compared to bone, low resistance of the subcutaneous tissue. Another mechanism that may induce or maintain the swelling is given by Fornasier et oi. 23, who described a correlation between the presence of birefringent polyethylene particles, the density of histiocytes and the thickness of a fibrohistiocytic membrane all of which showed an increase with time. A section obtained from the material with an implantation period of 5.7 years consists of a thin fibrous capsule and sheets of collagen interlaced with various cells. In contrast to the material that was implanted for 3.3 years, scarcely any PLLA material can be found in the extracellular space. The majority of the PLLA crystals has been internalized by phagocytizing ceils in membrane-bound vacuoles. These results may lead to the conclusion that with longer implantation periods there is a gradual shift of PLLA particles from extra- to intracellular in phagocytic cells that are imbedded in a fibrous matrix. The presence of macrophages and fibrocytes in response to the PLLA particles can be Biomaterials 1995, Vo]. 16 No. 1
106
The Biomaterials Silver Jubilee C o m p e n d i u m
30
Late degradation tissue response to poly(L-lactide) bone plates and screws: J.E. Bergsma et al.
expected since macrophages are known to phagocytize and remove foreign body material. As a response to internalization of the foreign body material macrophages can activate and attract fibroblast-like ceils. The extracellular degradation of the PLLA particles is probably a hydrolytic process. However, phagocytizing cells, especially macrophages, can release a number of lysosomal hydrolytic enzymes that may influence the degradation. If this is the case then an increased concentration of the lysosome guide enzyme, acid phosphatase, would be expected. Acid phosphatase is present in all lysosomes and its easy identification makes it an excellent marker. In the tissue with implantation periods of 5.7 years the presence of acid phosphatase was demonstrated, although not in abundance. Another enzyme that has been studied is lactic dehydrogenase (LDH). LDH converts lactic acid into pyruvate that can be metabolized in the citric acid cycle. If a substantial amount of intracellular PLLA particles degrades into lactic acid an increase might be expected. Again, the presence of enzyme-related precipitates were demonstrated but not in large amounts. Although a very limited number of enzymes were investigated these results may lead to the conclusion that the PLLA particles are eventually all internalized by phagocytizing cells that cannot actively degrade the PLLA particles. Hydrolysis is probably the only degradation mechanism and the highly crystalline particles seem to degrade very slowly. This implies that there is a long lasting presence of intracellular PLLA particles or that the particles are egested into the extracellular space because the cell cannot actively degrade the particles. Indigestible foreign body particles may cause a continuous attraction of macrophages that may again phagocytize the PLLA particles and thus repeat the intracellular cycle. Based on the literature on silicone implants another possibility may be that PLLA particles, or macrophages with PLLA particles, migrate to nodal tissue from the implant site 24. In this study no lymph nodes were excised, but perhaps in future studies the possibility of migration of PLLA particles to lymph nodes should be investigated. In the orthopaedic literature many studies have been published about aseptic loosening of prosthetic joints due to the presence of particulate polymer debris found within fibrous tissue, macrophages and foreign body cells. Horowitz et al. 25 described in an in vitro study that exposure to polymethylmethacrylate (PMMA) particles inhibits macrophage DNA synthesis, impairs their cytotoxic ability and eventually kills the cells. In our study cells that had internalized the lamellar or needle-like PLLA particles showed signs of mild cell damage such as enlarged rnitochondria and accumulation of glycogen. Human fibroblasts in culture accumulate glycogen in their cytoplasm as they approach senescence. In the 5.7 year specimens no signs of cell damage were observed. When an implanted material causes cellular damage, an increase in the leakage of intracellular lactate dehydrogenase may be expected. The damaging effect of the PLLA particles seems to be very low, no increased amounts Biomaterials 1995, Vol. 16 No. 1
of mitochondrial LDH could be demonstrated, so it may be assumed that the internalized PLLA crystals do not cause severe cell injury or cell death. The PLLA particles will probably induce a macrophage and fibrocyte response. The time needed for total hydrolytic degradation of the PLLA crystals will probably determine the duration of the swelling. The results of the trephined bone from the patient with an implantation period of 5.7 years, show a number of differences compared to the results of subcutaneously implanted material. The degradation of the PLLA screw-thread resembles the degradation of the PLLA bone plate, but the screw remnants are not interlaced with collagen fibres and internalization of PLLA particles by phagocytic cells is very limited. These results may indicate that there can be a variation in the degradation mechanism between subcutaneous and intraosseous PLLA implants and the histological reaction the implant induces. These differences may be explained by the fact that perhaps cortical bone can withstand the osmotic pressure of the degrading material and thus prevent swelling of the PLLA material. The PLLA material remains densely packed which perhaps prevents cellular ingrowth and internalization of PLLA particles. In summary, the disintegration of PLLA into particles with the accompanying increase in volume of the PLLA material itself and the fibrous tissue, may explain the origin of the described swelling. The PLLA particles with a very slow hydrolytic degradation rate, although not very irritable to the cell, do induce a cellular reaction. These are processes that resemble those seen in aseptic bone loosening in orthopaedic applications. The biocompatibility of the non-degraded PLLA material has been established in a number of studies. The degraded PLLA particles do not cause major cell injury but they can induce and maintain a clinically detectable swelling which could imply that these PLLA particles can no longer be considered to be fully biocompatible. Future research has to focus on biodegradable polymers that do not disintegrate into highly crystalline particles to avoid very long degradation periods, and in some applications a clinically detectable swelling.
REFERENCES 1
2
3
4
Akeson WH, Woo SL-Y, Coutts RD, Matthews JV, Gonsalves M, Amiel D. Quantitative histological evaluation of early fracture healing of cortical bones immobilized by stainless steel and composite plates. Calcif Tiss Res 1975; 19: 27-37. Paavolainen P, Karaharju E, Slatis P, Ahonen J, Holmstorm T. Effect of rigid plate fixation on structure and mineral content of cortical bone. Clin Orthop Rel Res 1987; 136: 287-293. Simon BR, Woo SL-Y, McCarthy M, Lee S, Akeson WH. Parametric study of bone remodeling beneath internal fixation plates of varying stiffness. J Bioeng 1978; 2: 543-556. Tunc DC, Rohovsky MW, Lehman WB, Strongwater A, Kummer F. Evaluation of body absorbable bone fixation devices. Proc 31st Ann Orthop Soc, Las Vegas, Jan 1985: 165.
107
The Biomaterials Silver Jubilee Compendium Late degradation tissue response to poly(L-lactide) bone plates and screws: J.E. Bergsma et ai.
5
6
7
8
9
10
11 12
13
14
15
Christel P, Vert M, Chabot F, Abols Y, Leray JL. Polylactic acid for intramedullary plugging. In: Advances in Biomaterials Vol. 5 (ed Ducheyne P}. Amsterdam: Elsevier Science; 1984: 1-6. Vert M, Christel P, Chabot F, Leray J. Bioresorbable plastic materials for bone surgery. In: Macromolecular Biomaterials, (eds Hastings GW, Ducheyne P). Boca Raton, FL: CRC Press; 1984: 120-142. B6stman O, M~ikel~i EA, T6mtil~i P, Rokkanen P. Transphyseal fracture fixation using biodegradable pins in children. J Bone Joint Surg [Br] 1989; 17-B: 7067O7. Suuronen R. Comparison of absorbable self-reinforced poly-L-lactide screws and metallic screws in the fixation of mandibular condyle osteotomies: An experimental study in sheep. J Oral Maxi]lofac Surg 1991; 49: 989-995. Leenslag JW, Pennings AJ, Bos RRM, Rozema FR, Boering G. Resorbable materials of poly(L-lactide). VI. Plates and screws for internal fracture fixation. Biomaterials 1987; 8: 70-73. Gogolewski S, Pennings AJ. Resorbable materials of poly(L-lactide) 2. Fibers spun from solutions of poly(Llactide) in good solvents. ! Appl Polym Sci 1983, 28: 1045-1061. Gogolewski S, Pennings AJ. Resorbable materials of poly{L-lactide). 3. Porous materials for medical application. Colloid Sci 1983; 261: 477-484. Leenslag JW, Gogolewski S, Pennings AJ. Resorbable materials of paly(t.-lactide). 5. Influence of secondary structure on the mechanical properties and hydrolyzability of poly(u-lactide) fibers produced by a dryspinning method. J Appl Polym Sci 1984; zg: 2829-2842. Bos RRM, Rozema FR, Boering G, Nijenhuis AJ, Pennings AJ, Verwey AB. Bioabsorbable plates and screws for internal fixation of mandibular fractures. A study in six dogs. Int J Oral Maxillofac Surg 1989; 18: 365-369. Bos RRM, Rozema FR, Boering G, Nijenhuis AJ, Pennings AJ, Jansen HWB. Bone plates and screws of bioabsorbable poly(L-lactide). An animal pilot study. Br J Oral Maxillofac Surg 1989; 27: 467-476. Rozema FR, Bos RRM, Pennings AJ, Jansen HWB. Poly(L-lactide) implants in repair of defects of the orbital floor. An animal study. ] Oral Maxillofac Surg 1990; 48: 1305-1309.
16
17
18
19
20
21
22
23
24
25
31
Bos RRM, Rozema FR, Boering G et al. Degradation of and tissue reaction to biodegradable poly(L-lactide) for use as internal fixation of fractures. A study in rats. Biomateriols 1991; lZ: 32-36. Bos RRM, Rozema FR, Boering G, Nijenhuis AJ, Penniags AJ, Verwey AB. Resorbable poly(L-lactide) plates and screws for the fixation of unstable zygomatic fractures. J Oral Maxillofac Surg 1987; 45: 751-753. Rozema FR de Bruijn WC, Bos RRM, Boering G, Nijenhuis AJ, Pennings AJ. Late tissue response to bone-plates and screws of poly(L-lactide) used for fracture fixation of the zygomatic bone. In: Doherty PJ, Williams RL, Williams DF, eds, Biomaterial - Tissue Interfaces, Advances in Biomaterials vol. 10. Amsterdam: Elsevier, 1992: 349-355. Bergsma JE, Rozema FR, Bos RRM, de Bruijn WC. Foreign body reactions to resorbable poly(L-lactide) bone plates and screws used for the fixation of unstable zygomatic fractures. J Oral Maxillofac Surg 1993; 51: 666-670. Hulstaert CE, Kalicharan D, Hardonk MJ. Cytochemical demonstration of phophatases in the rat liver by a cerium-based method in combination with osmiumtetroxide and potassium ferrocyanide post-fixation. Histochemistry 1983; 78: 71-79. Hanker JS, Kusyk CJ, Bloom FE. The demonstration of dehydrogenases and monoamine oxidase by the formation of osmium blacks at the sites of Hatchett's brown. Histochemie 1973; 33: 205-230. BSstman O, P~iiv~irinta U, Manninen M, Rokkanen P. Polymeric debris from absorbable polyglycolide screws and pins. Acta Orthop Scand 1992; 63: 555559. Fornasier V, Wright J, Seligman J. The histomorphologic and morphometric study of asymptomatic hip arthroplasty. A postmortem study. Clin Orthop 1991; 271: 272-282. Dolwick MF, Aufdemorte TB. Silicone-induced foreign body reaction and lymphadenopathy after temperomandibular joint arthroplasty. Oral Surg Oral Med Oral Pathol 1985; 59: 449-452. Horowitz SM, Gautsch TL, Frondoza CG, Riley Jr L. Macrophage exposure to polymethyl methacrylate leads to mediator release and injury. J Orthop Res 1991; 9: 406-413.
Biomaterials 1995, Vol. 16 No. 1
108
This Page Intentionally Left Blank
The Biomaterials Silver Jubilee Compendium
109
The Biomaterials Silver Jubilee Compendium
Biomaterials 16 (1995) 297-303 9 1995 Elsevier Science Limited Printed in Great Britain. All rights reserved 0142-9612/95/$10.00
r~u TT ERWO R T H ~l~E I N E M A N N _
Mechanism of cell detachment from temperature-modulated, hydrophilichydrophobic polymer surfaces Teruo Okano, Noriko Yamada, Minako Okuhara, Hideaki Sakai and Yasuhisa Sakurai
Institute of Biomedical Engineering, Tokyo Women's Medical College, 8-1 Kawada, Shinjuku, Tokyo 162, Japan
Poly(N-isopropylacrylamide) (PIPAAm), exhibiting a lower critical solution temperature (LCST) at 25 ~C in physiological phosphate buffered saline solution (pH 7.4) and at 32~ in pure water, was grafted onto the surfaces of commercial polystyrene cell culture dishes. This PIPAAm-grafted surface exhibited hydrophobic surface properties at temperatures over the LCST and hydrophilic surface properties below the LCST. Endothelial cells and hepatocytes attached and proliferated on PIPAAmgrafted surfaces at 37~ C, above the LCST. The cultured cells were readily detached from these surfaces by lowering the incubation temperature without the usual damage associated with trypsinization. In this case, the optimum temperature for cell detachment was 10~ for hepatocytes and 20~ for endothelial cells. Cell detachment was partially inhibited by sodium azide treatment, suggesting that cell metabolism directly affects cell detachment. Morphological changes of the adherent cells during cell detachment experiments indicated further involvement of active cellular metabolic processes. Cells detached from hydrophobic-hydrophilic PIPAAm surfaces not only via reduced cellsurface interactions caused by the spontaneous hydration of grafted PIPAAm chains, but also by active cell morphological changes which were a function of cell metabolism. Biomaterials (1995) 16 (4), 297-303
Keywords: Thermoresponsive polymer surface, cell culture, cell detachment, hepatocyte, endothelial cell Received 21 December 1993; accepted 25 April 1994
Poly(N-isopropylacrylamide) (PIPAAm), a thermoresponsive polymer, exhibits a lower critical solution temperature (LCST) of about 32 ~ in water a.2. PIPAAm is fully hydrated with an extended chain conformation in aqueous solutions below 32~ and is extensively dehydrated and compact above 32~ These unique thermosensitive polymers and their copolymers have therefore been utilized in temperature-modulated bioconjugates constructed by a bioactive molecule and a stimuli-responsive PIPAAm chain. Thermally modulated to induce soluble-insoluble changes in solution, these new bioconjugates have attracted considerable attention in both fundamental research and practical application, such as bioactivity control and bioseparations in protein engineering 3-~. Further, cross-linked PIPAAm and its copolymers have been developed as thermal on-off switching polymers for drug permeation and release 7-1~ We have studied thermal on-off modulation of hydrophilic-hydrophobic changes on PIPAAm-grafted surfaces 11'12. Cells cultured on hydrophobic PIPAAmgrafted surfaces at 37~ (above the LCST of 32~ were prompted to detach spontaneously by lowering
the medium temperature and changing the hydration of the PIPAAm chains. Cultured cells generally will adhere to hydrophobic surfaces but not on highly hydrated hydrophilic surfaces 13'14. Our study clearly demonstrated the feasibility of a new recovery strategy for harvesting cultured cells by external modulation of thermoresponsive surfaces. In fact, PIPAAm-grafted surfaces demonstrate very effective thermal switching to reverse hepatocyte and endothelial cell attachment and detachment without cell damage 11'15. Cell adhesion onto a material surface can be arbitrarily classified as a two-step mechanistic process: the first stage is controlled by complex combinations of physicochemical interactions including hydrophobic, coulombic, and van der Waals forces between the cell membrane and the material surface. This process might be termed 'passive adhesion' according to this adsorption mechanism. The second stage might be considered as 'active adhesion', because of the participation of cellular metabolic processes. Attached ceils are well-known for changing their shapes and expending metabolic energy in order to stabilize the interface between their membrane and the underlying materials, by both physicochemical and biological mechanisms16,17
Correspondence to Dr Teruo Okano. 297
Biomaterials 1995, Vol. 16 No. 4
110
The Biomaterials Silver Jubilee Compendium 298
When surface properties of the material are changed from hydrophobic to hydrophilic, cells often attempt to detach themselves from the surface as mentioned above. In this case, cells may not be able to detach from the surface without involving cellular metabolic processes which actively change membrane morphology. This paper attempts to clarify the influence of cell metabolic processes on cell detachment. No metabolic effects on cell detachment would be manifested by increasingly rapid cell detachment with decreasing temperature as PIPAAm hydration is enhanced when temperature is reduced. Mechanisms of cell detachment were discussed with regard both to effects of temperature-modulated surface changes and cell metabolic changes.
MATERIALS AND M E T H O D S Preparation of poly(IPAAm)-grafted surfaces The procedures for the preparation of PIPAAm-grafted cell culture dishes are described elsewhere 11'15. NIsopropylacrylamide monomer (IPAAm) (Eastman Kodak, Rochester, NY, USA) was dissolved in isopropyl alcohol. IPAAm solution (45wt%, 0.1ml) was added to each polystyrene tissue culture dish (Falcon 3001, diameter 35mm, Falcon Becton Dickinson Labware, Oxnard, CA, USA) and then irradiated with a 0.25 MGy electron beam (200kV, under 1.3 x 1 0 -4 Pa) using an Area Beam Electron Processing System (Nisshin High Voltage, Kyoto, Japan). IPAAm was polymerized and grafted onto the surfaces of the dishes using an electron beam. The PIPAAm-grafted dishes were rinsed with cold distilled water to remove non-grafted IPAAm, dried under nitrogen gas and gassterilized by ethylene oxide before use in cell culture experiments. Untreated Falcon 3001 dishes were used as controls. Homogeneous coverage of PIPAAm-grafted dishes was confirmed using field emission scanning electron microscopy. The amount of grafted PIPAAm polymer can be controlled by the concentration of IPAAm solution in each preparation. Surfaces of these PIPAAm-grafted dishes change reversibly between hydrophilic and hydrophobic by controlling temperature, as previously reported '5.
Cell culture Endothelial cells were isolated from bovine thoracic aorta by a dispase digestion method described previously' 5. Endothelial cells were cultured in Dulbecco's modified Eagle's Minimal Essential Medium (DMEM) (Gibco, Grand Island, NY, USA) supplemented with 10% fetal bovine serum (FCS) (Gibco), lOOUm1-1 of penicillin (Gibco), 100~gm1-1 of streptomycin (Gibco) and 2.Spgm1-1 of fungizone (Gibco) at 37~ in a fully humidified atmosphere of 5% CO2 in air. The cells were subcultured by treatment with 0.05% trypsin and 0.02% ethylenediaminetetraacetic acid (EDTA) solution (Gibco) for 5 min after confluent cell monolayers had formed. Endothelial ceils from the third to sixth passages were used in all experiments. Rat hepatocytes were isalated from 5-week-old male Biomaterials 1995, Vol. 16 No. 4
Cell detachment from thermoresponsive surface: T. Okano et al.
Wistar rats, weighing about 150g, using an in situ collagenase perfusion method previously reported '5"18. More than 98% of the cells obtained were parenchymal cells as determined by phase-contrast microscopy, and more than 90% were viable as measured by trypan blue dye exclusion. The hapatocytes were cultured in Williams E medium (Gibco) supplemented with 5% FCS, 10 ngm1-1 human epidermal growth factor (hEGF) (Wakunaga Pharmaceutical, Osaka, Japan), 10mM nicotinamide (Wako, Pure Chemical Industries, Osaka, Japan), 5 U m l - ' aprotinin (Wako), 1 0 - 7 M insulin (Sigma Chemical, St Louis, MO, USA), lO-aM dexamethasone (Sigma) and 50mgml-1 canamaicin sulphate (Gibco) at 37 ~ under a humidified atmosphere of 5% CO2 in air.
Measurement of DNA Cell numbers were calculated by DNA content in cultured cells. The amount of DNA was assayed fluorometrically with calf thymus DNA (type 1, Sigma) as the standard 15,19. Briefly, cells were solubilized with 10mM EDTA solution, pH 12.3 (Kanto Chemical, Tokyo, Japan) for 30min at 37~ and neutralized by the addition of 1M potassium dihydrogen phosphate (KH2PO4) (Kanto) solution and then mixed with 2'-(4hydroxyphenyl)-5-(4-methyl-l-piperazinyl)-2, 5'-bi-lHbenzimidazole (Hoechst 33258) solution (Sigma). DNA content was determined by fluorimetry (JASCO FP-770 spectrofluorometer; Japan Spectroscopic Co, Tokyo, Japan) at 360 nm excitation and 450 nm emission.
Influence of temperature on cell detachment Rat hepatocytes and endothelial cells were seeded onto PIPAAm-grafted and control dishes at a density of 4 x 1 0 4 cells cm -2 and cultured in their respective culture medium at 37 ~ under a humidified atmosphere of 5% CO2 in air. After 2 d, the temperature of the cell culture system was decreased from 37~ to T~ (4, 10, 15, 20 and 27 ~C) by changing with medium of T~ and cooling the culture dishes. Culture dishes containing cells were incubated at T~ for 30 min and cells detached from cell culture dish surfaces were estimated after 5min additional incubation at 25~ Detached cells from cell culture dishes were collected and cell number was estimated from DNA measurement.
Effect of sodium azide on hepatocyte detachment Sodium azide (Nacalai Tesque, Kyoto, Japan) was dissolved in culture medium for hepatocytes and adjusted to concentrations of 0, 0.2, 1.0 and 2.0ruM. Rat hepatocytes were seeded on PIPAAm-grafted and control dishes at an initial density of 4 x 1 0 4 cells c m -2 and cultured for 2 d in culture medium for hepatocytes at 37~ under a humidified atmosphere of 5% CO2 in air. Culture dishes were then changed with a culture medium containing sodium azide at each concentration, and incubated for 60rain at 37 C,C. After treatment with sodium azide, the culture dish incubation temperature was reduced and incubated at 10~ for 30min followed by additional incubation at 25~ for 5min. Cells detached from cell culture dishes were then collected and cell number was determined by
111
The Biomaterials Silver Jubilee Compendium Cell detachment from thermoresponsive surface: T. Okano et al.
DNA measurement as described above. Moreover, to clarify the influence of cell metabolism in cell detachment, the effect of sodium azide on the temperature dependence of cell detachment was also investigated. Hepatocytes were cultured on PIPAAm-grafted dishes at 37~ for 2 d as described above. After pretreatment with or without 2 mM sodium azide solution for 60min at 37 ~ C, the culture dishes were incubated at T~ (4, 10, 15, 20 and 27~ for 30min and an additional 5 m i n at 25 ~ The percentage of detached cells was calculated and cell numbers were determined.
Cell morphology by optical and scanning electron microscopy Hepatocytes were cultured on PIPAAm-grafted dishes as described above. After 2 d culture at 37~ the culture dishes were incubated at 10~ for 30min and an additional 5 m i n at 25~ The morphological changes of detaching hepatocytes on dish surfaces were directly and continuously observed by phasecontrast microscopy (Nikon Diaphot-TMD, Tokyo, Japan) using a micro-cool plate (Kitazato Supply, Sizuoka, Japan). For scanning electron microscopy (SEM), detaching hepatocytes on dish surfaces were fixed at 10~ for 6 0 m i n with 2% glutaraldehyde (EM Sciences, Fort Washington, USA) in 0.1M cacodylatebuffered solution (EM Science), pH 7.4. The fixed cells were washed with cacodylate-buffered solution and then lyophilized. After sputter-coating with gold, the samples were observed using a scanning electron microscope (JEOL JSM5300LV, Tokyo, Japan).
RESULTS AND DISCUSSION
Influence of temperature on hepatocyte detachment For primary rat hepatocytes, cell growth curves were observed to be similar on PIPAAm-grafled and control dishes as reported previously 15. After hepatocytes were cultured for 2 d at 37 ~C, the temperature of the cell culture systems was decreased from 37 to T~ by both cooling the culture dishes and exchanging the medium with fresh medium at T~ After 30min incubation at T ~C, an additional 5 min incubation was performed at 25~ (Figures 1, curve A) and compared with results at constant temperature of T~ (curve B). Figure I shows the correlation between the percentage of detached cells and incubation temperature, T~ Cells remained over 85% attached at both 30 and 35 rain incubation times after the cell culture systems were decreased from 37 to T ~C. At lower temperatures, 4 and 10 ~C, the number of detached cells was smaller than at higher temperature, 15 and 20 ~C, at both 30 min (C) and 35 min (B) incubation times. These results show that cell detachment is not directly correlated with reduced temperature. Grafted PIPAAm chains are assumed to be hydrated and maintain expanded conformations at lower temperatures resulting in reduced interactions between the cells and grafted surfaces of the cell culture dishes. Hydration of PIPAAm at the cell-material interface, therefore, does not completely govern cell detachment from culture dish surfaces. As cell metabo-
299
100 80 A tj "{3 t3 to
60
40
Q
20
_ ~
0
..
9
.... !
L
10
_
I .
20
L
II
,,
30
_
40
Temperature (~ Figure 1 Hepatocyte detachment from PIPAAm-grafted surfaces by reducing temperature, and subsequent additional 5-min incubation at 25~ (A) and T~ (B) after
30-min incubation at T~
(C) (T = 4, 10, 15, 20, 27 ~C).
lism is suppressed by decreasing temperature, influences of the cell metabolic processes as well as hydration of the culture surface on cell detachments are implicated. In fact, 30 min incubation at T~ followed by a temperature change to 25~ in order to increase cell metabolism drastically enhances cell detachment, as shown in Figure 1 (curve A). In this case, numbers of detached cells show a maximum at 10 ~C. These results demonstrate that cell detachment is controlled not only by the hydration of grafted PIPAAm on the culture dishes but also by active cellular metabolism. Morphologies of detaching cells were observed by both optical and electron microscopies over time after 30 rain incubation at 10 ~C followed by a temperature increase to 25 ~C as shown in Figures 2 and 3, respectively. Cells start to change their shape from a spread to a rounded form, and finally cells are observed to detach completely from the surface. After 10min incubation at 25~ 100% of cells were detached. These results clearly demonstrate that the cell detachment process involves cell shape changes accompanying a consumption of cellular metabolic energy. Hydration changes of grafted PIPAAm at the cell-material interface is an important initial stimulus to induce active cell detachment mediated by cellular processes.
Effect of sodium azide on hepatocyte detachment Sodium azide is a known inhibitor of cytochrome C oxidase in mitochondoria and decreases ATP generat i o n 2~ resulting in the disruption of cellular activities which require ATP. The effect of sodium azide on cell detachment was investigated to clarify the role of cell metabolism in cell detachment. Cultured cells were damaged and detached from Biomaterials 1995, Vol. 16 No. 4
112
The Biomaterials Silver Jubilee Compendium 300
Cell detachment from thermoresponsive surface: T. Okano et al.
Figure 2 Phase contrast micrographs showing the process of hepatocyte detachment from PIPAAm-grafted surfaces at 25 ~ C after 30-min incubation at 10 ~ C.
culture surfaces when sodium azide over 10raM was added. At concentrations below 2mM sodium azide, cells were not observably detached from culture dish surfaces and cell metabolism was only partially inhibited. After 2 d cultured cells were treated with sodium azide for 60min at 37~ then incubated at 10~ for 30min followed by additional incubation at 25~ for 5 min. Cell detachment was scarcely observed with and without additions of up to 2mM sodium azide, as shown in Figure 4. By contrast, the temperature-responsive cell recovery system shows significant inhibition of cell detachment at reduced temperatures. This inhibition on PIPAAm-grafted surfaces was increased with increasing sodium azide. Since sodium azide treatment under these conditions may not completely inhibit cellular ATP generation, the inhibition of cell detachment observed is partial but not complete, as shown in Figure 4. Figure 5 shows detached cell percentages from PIPAAm-grafted dishes with or without sodium azide treatment followed by incubation at T~ for 30min and additional 5 min at 25 ~C. The inhibition effects of 2ram sodium azide treatment on cell detachment was observed over all temperatures. These results strongly suggest that active cellular processes enabling cell morphological changes are essential procedure to complete cell detachment from hydrated surfaces.
Mechanism of endothelial cell detachment Endothelial cells are readily cultured on the PIPAAmgrafted surfaces and proliferate the same as on commercial dishes, as shown in a previous paper 15. Detachment of endothelial cells was investigated using the same experiments for temperature-modulated cell detachment on PIPAAm-grafted surfaces. More significant cell Biomaterials 1995, Vol. 16 No. 4
detachment maxima for endothelial cells are observed, as shown in Figure 6. The maximum at 20 ~C - - a significantly higher temperature than in the case of hepatoc y t e s - - i s evident. As discussed above with regard to hepatocytes, the detachment of endothelial cells is also controlled by two steps: the initial temperature-responsive PIPAAm surface hydration, and active processes of cell detachment accompanying cell shape changes. Cell detachment is not significant at lower temperatures, even if PIPAAm swelling is more significant at lower temperatures. The maximum peak for endothelial cell detachment at 20~ suggests that endothelial cell metabolism is inhibited more significantly at lower temperatures compared with hepatocytes. Since different cell lines exhibit different temperature sensitivities, the maximum peak for detachment of endothelial cells at a higher temperature than that for hepatocytes is another important result to support a metabolically related cell detachment mechanism. After the culture dish was changed from 37 to 25 c' C, the temperature-induced increase in PIPAAm swelling was no enough to initiate cell detachment. However, dishes reduced from 37 to 20~ increased PIPAAm hydration sufficiently to initiate cell detachment. A 30-min incubation at 20~ allows PIPAAm chains to hydrate and expand their conformations. However, this alone is not sufficient to induce cell detachment because the temperature is not high enough to allow cellular metabolism and accompanying morphological changes. Therefore, cell detachment is significantly enhanced by increasing the temperature at 25~ sufficient to induce observable cellular shape changes. Even if PIPAAm chains dehydrate slightly by increasing this temperature from 20 to 25~C, metabolic changes of cells at this higher temperature seem to be much more significant than hydration changes of PIPAAm-grafted chains.
113
The Biomaterials Silver Jubilee Compendium 301
Cell detachment from thermoresponsive surface: T. Okano et al.
Figure 4 Effect of sodium azide concentration on hepatocyte detachment. Hepatocytes were pretreated with sodium azide for 60min at 37~ and then detached by 30-rain incubation at 10~ and an additional 5min at 25 ~
Figure 3 Scanning electron micrographs showing the process of hepatocyte detachment from PIPAAm-grafted surfaces at 25~ after 30-rain incubation at 10~ C_ a, 0 rain. b, 3 min. c, 10 rain.
This result demonstrates that cells are very sensitive to hydratian changes on surfaces of the culture dishes. Effective cell detachment requires an increase in temperature to recover cell metabolism after increasing the hydration of PIPAAm by decreasing temperature. As different cells have different metabolic requirements, the optimum temperature is different for different cell lines.
Mechanism of the cell detachment from thermoresponsive polymer surfaces Figure 7 represents cell adhesion and detachment data on material surfaces. Cells are small particles but are
Figure 5 Effect of sodium azide on hepatocyte detachment by reducing temperature, 30-rain incubation at T~C and an additional 5rain at 25 ~ Hepatocytes on PIPAAm-grafted dishes were pretreated with ( 9 and without (0) 2 m i sodium azide. Hepatocytes on control dishes were pretreated with 2 m u sodium azide (I-l).
distinctly different from small artificial particles in adhesion because cells have metabolism. After cells contact surfaces (passive adhesion), cells are always dynamically altering their cell membrane and its morphology to optimize interactions and to stabilize the cell-material surface interface (active adhesion), both physicochemically and biologically. Therefore, cell adhesion should be divided into two stages: passive adhesion and active adhesion, as shown in
Figure 7. Biomaterials 1995, Vol. 16 No. 4
114
The Biomaterials Silver Jubilee Compendium Cell detachment from thermoresponsive surface: T. Okano et al.
302
Figure 7
Mechanism of the cell attachment to and detachment from material surfaces.
100
different temperature-sensitive cell metabolic efficiencies.
80
A
m m
0
0 t~
60
ACKNOWLEDGEMENTS The authors gratefully acknowledge Dr David Grainger for valuable discussions. This research was supported by the Ministry of Education (grant no. 04453108), Japan.
A
40 20
00
REFERENCES
1
B, C 10 20 30 Temperature (~
2
40
3
Figure 6 Endothelial cell detachment from PIPAAm-grafted
surfaces by reducing temperature and subsequent additional 5-min incubation at 25~ (A) and T~C (B) after 30-min incubation at T~ (T = 4, 10, 15, 20, 27).
When cultured cells in active adhesion seek to detach themselves from a surface, cell shape changes which consume energy are necessary as shown in this paper. W h e n the temperature of culture dishes originally incubated at 37~ is decreased, PIPAAm chains start to hydrate below 32 ~ This remarkable hydration change initiates cell detachment. Further cell d e t a c h m e n t from these temperature-responsive surfaces requires adherent cells to change their m e m b r a n e shape, consuming internal metabolic energy. Lower temperatures provide more hydrated PIPAAm chains but reduce cell metabolism. Therefore, o p t i m u m temperatures are observed to recover cells fully self-detached from temperature-responsive surfaces. As different cells have different temperature sensitivities for cellular metabolism, hepatocytes and endothelial cells require different o p t i m u m temperatures for their detachment. Also, subtle thermal control of cell d e t a c h m e n t is an important basis for advanced technologies, not only for cell culture but for purification, sorting and separation of ceils on the basis of Biomaterials 1995, Vol. 16 No. 4
4
5
6
7 8
9
10 11
Bae YH, Okano T, Kim SW. Temperature dependence of swelling of crosslinked poly(N,N'-alkyl substituted acrylamide) in water. J Polym Sci: Polym Phys 1990; 28: 923-936. Heskins M, Guillent JE, James E. Solution properties of poly(N-isopropylacrylamide). J Macromol Sci Chem 1968; A2: 1441-1455. Takei YG, Aoki T, Sanui K, Ogata N, Okano T, Sakurai Y. Temperature-responsive bioconjugates. 1. Synthesis of temperature-responsive oligomers with reactive end groups and their coupling to biomolecules. Bioconjugate Chem 1993; 4: 42-46. Takei YG, Aoki T, Sanui K, Ogata N, Okano T, Sakurai Y. Temperature-responsive bioconjugates. 2. Molecular design for temperature-modulated bioseparations. Bioconjugate Chem 1993; 4: 341-348. Chen JP, Yang HJ, Hoffman AS. Polymer-protein conjugates I. Effect of protein conjugation on the cloud point of poly(N-isopropylacrylamide). Biomaterials 1990; 11: 625-630. Chen JP, Hoffman AS. Polymer-protein conjugates II. Affinity precipitation separation of human immunogammaglobulin by a poly(N-isopropylacrylamide)protein A conjugate. Biomaterials 1990; 11: 631-634. Okano T, Bae YH, lacobs H, Kim SW. Thermally on-off switching polymers for drug permeation and release. J Control Rel 1990; 11: 255-265. Yoshida R, Sakai K, Okano T, Sakurai Y, Bae YH, Kim SW. Surface-modulated skin layers of thermal responsive hydrogels as on-off switches: I. Drug release. J Biomater Sci Polym Edn 1991; 3:155-162. Yoshida R, Sakai K, Okano T, Sakurai Y. Surfacemodulated skin layers of thermal responsive hydrogels as on-off switches: II. Drug permeation. / Biomater Sci Polym Edn 1991; 3: 243-252. Yoshida R, Sakai K, Okano T, Sakurai Y. Pulsatile drug delivery systems using hydrogels. Adv Drug Delivery Rev 1993; 11: 85-108. Yamada N, Okano T, Sakai H, Karikusa F, Sawasaki Y,
115
The Biomaterials Silver Jubilee Compendium Cell detachment from thermoresponsive surface: T. Okano et al.
12
13
14
15
16
Sakurai Y. Thermo-responsive polymeric surfaces: Control of attachment and detachment of cultured cells. Makromol Chem, Rapid Commun 1990; 11: 571576. Takei YG, Aoki T, Sanui K, Ogata N, Okano T, Sakurai Y. Dynamic contact angle measurement of temperature-responsive surface properties for poly(Nisopropylacrylamide) grafted surface. Macromolecules (in press). Ratner BD, Horbett T, Hoffmen AS. Cell adhesion to polymeric materials; implications with respect to biocompattibility. J Biomed Mater Res 1975; 9: 407422. McAuslan BR, Johnson G. Cell response to biomaterials I: Adhesion and growth of vascular endohelial cells on poly(hydroxyethyl methacrylate) following surface modification by hydrolytic etching. J Biomed Mater Res 1987; 21: 921-935. Okano T, Yamada N, Sakai H, Sakurai Y. A novel recovery system for cultured cells using plasma-treated polystyrene dishes grafted witth poly(N-isopropylacrylamide). J Biomed Mater Res 1993; 27" 1243-1251. Kataoka K, Okano T, Sakurai Y, Maruyama A, Tsuruta
303
17 18 19
20
21 22
T. Controlled interactions of cells with multiphasestructured surfaces of block and graft copolymers. In: Tsuruta T, Nakajima A, eds. Multiphase Biomedical Materials. Tokyo: VSP; 1989: 1-19. Baier RE, DePalma VA, Goupil DW, Cohen E. Human platelet spreading on substrata of known surface chemistry. J Biomed Mater Res 1985; 19: 1157-1167. Seglen PO. Preparation of isolated liver cells. Method Cell Biol 1976; 13: 29-83. West DC, Sattar A, Kumar S. A simplified in situ solubilization procedure for determination of DNA and cell number in tissue culture. Anal Biochem 1985; 147: 289-295. Yonetani T, Ray GS. Studies on cytochrome oxidase VI. Kinetics of the aerobic oxidation of ferrocytochrome c by cytochrome oxidase. J Biol Chem 1965; 240: 33923398. Wilson DF, Chance B. Reversal of azide inhibition by uncoupler. Biochim Biophys Res Commun 1966; 23: 751-756. Palmieni F, Klingenberg M. Inhibition of respiration under tthe control of azide uptake by mitochondria. Eur J Biochem 1967; 1: 439-446.
Biomaterials 1995, Vol. 16 No. 4
116
This Page Intentionally Left Blank
The Biomaterials Silver Jubilee Compendium
117
The Biomaterials Silver Jubilee Compendium
Biomateria]s 17 (1996) 103-114 9 1996 Elsevier Science Limited Printed in Great Britain. All rights reserved 0142-9612/96/$15.00
ELSEVIER
Mechanisms of polymer degradation and erosion Achim G6pferich
Department of Pharmaceutical Technology, University of Erlangen-NOrnberg, CauerstraBe 4, 91058Erlangen, Germany The most important features of the degradation and erosion of degradable polymers in vitro are discussed. Parameters of chemical degradation, which is the scission of the polymer backbone, are described such as the type of polymer bond, pH and copolymer composition. Examples are given how these parameters can be used to control degradation rates. Degradation leads finally to polymer erosion, the loss of material from the polymer bulk. The resulting changes in morphology, pH, oligomer and monomer properties as well as crystallinity are illustrated with selected examples. Finally, a brief survey on approaches to polymer degradation and erosion is given.
Keywords: Polymers, erosion, degradation, modelling, mechanisms Received 29 October 1994; accepted 30 December 1994
At present, tremendous progress is being made in the medical sciences toward the advancement of medical therapies through the application of degradable polymers. Degradable materials are used for the local treatment of cancer 1, the development of vaccines 2'3, the manufacture of nanoparticles with increased plasma half.life 4, 5, self-regulated drug delivery systems 6'7, orthopaedic fixing devices 8 and the fight against organ failure 9. Concomitantly, investigations of these sophisticated applications, however, have raised serious questions about the suitability of degradable polymers in some cases. Examples are the stability of sensitive compounds such as protein and peptide drugs, or the survival of living cells, in the constantly changing chemical environment of an eroding polymer. Other concerns are related to the loss of mechanical stability of polymers during erosion TM, which can be undesirable when occurring too fast, or the toxicity of high concentrations of degradation products. A physical chemical understanding of polymer degradation and erosion processes is the key for a better understanding of these problems and maybe also for their solution. Polymer degradation and erosion play a role for all polymers. The distinction between degradable and non-degradable polymers is, therefore, not clean-cut and is in fact arbitrary, as all polymers degrade. It is the relation between the time-scale of degradation and the time-scale of the application that seems to make the difference between degradable and non-degradable polymers. We usually assign the attribute 'degradable' to materials which degrade during their application, or immediately after it. Non-degradable polymers are those that require a substantially longer time to
degrade than the duration of their application. Degradation and erosion are investigated in many fields of science, such as waste management I1'12 and space science 13. Therefore, many different definitions for degradation and erosion exist in the current literature and sometimes vary markedly from one another 14. The following definitions are adapted for this review. The process of 'degradation' describes the chain scission process during which polymer chains are cleaved to form oligomers and finally to form monomers. 'Erosion' designates the loss of material owing to monomers and oligomers leaving the polymer 15 There are different types of polymer degradation such as photo-, thermal-, mechanical and chemical degradation 16'17. All polymers share the property that they erode markedly under the influence of UV light or 7-radiation TM. For polymer biomaterials, such effects are of minor importance, unless they are submitted to 7-sterilization, after which a significant loss of molecular weight can be observed 19. Thermal degradation plays a greater role for non-degradable polymers 2~ Mechanical degradation affects those biodegradable polymers that are subjected to mechanical stress, such as non-degradable polymers 21 or biodegradable polymers used as fixture or suture material 22. All biodegradable polymers contain hydrolysable bonds. Their most important degradation mechanism is, therefore, chemical degradation via hydrolysis or enzyme-catalysed hydrolysis. The latter effect is often referred to as biodegradation, meaning that the degradation is mediated at least partially by a biological system TM. The processes involved in the erosion of a degradable polymer are complicated. Water enters the polymer bulk, which might be accompanied by swelling. The
Correspondence to Dr A. G6pferich. 103
Biomaterials 1996, Vo]. 17 No. 2
The Biomaterials Silver Jubilee Compendium
104
intrusion of water triggers the chemical polymer degradation, leading to the creation of oligomers and monomers. Progressive degradation changes the microstructure of the bulk through the formation of pores, via which oligomers and monomers are released. Concomitantly, the pH inside pores begins to be controlled by degradation products, which typically have some acid-base functionality. Finally, oligomers and monomers are released, leading to the weight loss of polymer devices. The development of biodegradable polymers during the last two decades has increased exponentially, going hand in hand with new applications for such materials. In early applications, degradable polymers were used as resorbable suture materials 2a. Poly(lactic acidJ and poly(glycolic acid) served as raw materials for such applications. Since the 1970s, these polymers have been used for drug delivery 24 because of their excellent biocompatibility. Soon it was realized, however, that these polymers would not fit the needs of a growing number of applications. Therefore, new polymers were synthesized. Poly(ortho esters) 2'~ poly(anhydrides) 26 and many other polymers emergeci as new materials in the early 1980s. Since then, numerous polymers have been manufactured to keep pace with a steadily increasing demand 27'2~. It is not possible to elucidate the details of erosion for all these polymers in this article. The intention of this review is rather to summarize the most important features of chemical polymer degradation and erosion in vitro and to show how these effects may be described by theoretical simulations.
POLYMER DEGRADATION Polymer degradation is the key process of erosion. There are two principal ways by which polymer bonds can be cleaved: passively by hydrolysis or actively by enzymatic reaction 2~. The latter option is only effectively available for naturally occurring biopolymers like polysaccharides, proteins (gelatin and collagen :~~ and poly(fl-hydroxy acids) 31, where appropriate enzymes are available. A detailed list of enzymatically degradable polymers can be found in Ref. 32. For most biodegradable materials, especially artificial polymers, passive hydrolysis is the most important mode of degradation. There are several factors that influence the velocity of this reaction: the type of chemical bond, pH, copolymer composition and water uptake are the most important. Chemical and physical changes go along with the degradation of biodegradable polymers, like the crystallization of {~ oligomers:: and monomers :~4 or pH changes 5 Some of these factors can have a substantial feedback effect on the degradation velocity. The most important parameter for monitoring degradation is molecular weight. Besides loss of molecular weight, other parameters have been proposed as a measure for degradation, like loss of mechanical strength, complete degradation into monomers or monomer release. All of these are related but need not necessarily obey the same kinetics. For example, complete degradation of poly(L-lactic acid) is known to take B i o m a t e r i a l s 1996, Vol. 17 No. 2
118 Mechanisms of polymer degradation and erosion A. Gopferich
substantially more time than the loss of tensile strength as. Aqueous solutions of lactic acid form spontaneously poly(lactic acid) oligomers that might affect molecular weight measurements :~6, and monomers from copolymers need not be released with identical kinetics during erosion~~. The specific relation between the erosion parameters varies according to the type of polymer. There are, however, basic principles, according to which degradation proceeds, and how degradation can be influenced.
The importance of the type of chemical bond for polymer degradation It is mainly the type of bond within the polymer backbone that determines the rate of hydrolysis :~7. Several classifications for ranking the reactivity exist which are either based on hydrolysis kinetics data for polymers :~':~~ or are extrapolated from low-molecularweight compounds containing the same flmctional group 3z. A brief list is given in Table 1. Anhydrideand ortho-ester bonds are the most reactive ones, followed by esters and amides. Such rankings must be viewed, however, with circumspection. Reactivities can change tremendously upon catalysis 4~J'4j or by altering the chemical neighbourhood of the functional group 4z through steric and electronic effects. The substitution of hydrogen by chlorine in the acid :~position of ethyl acetate, for example, increases the reaction rate constant for hydrolysis in neutral media from 2.5 x 1 0 ~ (s ~) to 1.1x I0 ~ (s ~) through a negative inductive effect 42. The influence of steric effects on degradation can be seen with poly(~.hydroxy esters). The slower degradation of poly(lactic acid) is partially due to the steric effects 4:~. because the voluminous alkyl group hinders the attack of water.
Table 1 Classes of hydrolysable bonds with half-lives according to References 32 and 39 Polymer class o
Half-life poly(anhydrides)
0.1 h
poly(ortho esters)
4h
poly(esters)
3.3 yrs
poly(amides)
83 000 y rs
II
R~C--O~
O~(~C CH:~
i i HH i N--C--
I
R
The Biomaterials Silver Jubilee Compendium
119
Mechanisms of polymer degradation and erosion: A. Gdpferich
The effect of pH on polymer degradation The pH affects reaction rates through catalysis. After shifts in pH, reaction rates of esters, for example, may change some orders of magnitude due to catalysis 42. Ester hydrolysis can, thereby, be either acid or base catalysed 4~ The effect of pH on degradation has been investigated carefully for most biodegradable polymers. For poly(glycolic acid) and poly(lactic-coglycolic acid) sutures, the breaking strength was found to depend markedly on the pH of the degradation medium and was found to be highest at neutral pH 44, reflecting the fastest degradation at low and high pH. This faster chain scission at low pH explains the heterogeneous erosion of poly(lactic acid) due to autocatalysis. The generated monomers, which are carboxylic acids, accelerate polymer degradation by lowering pH 45. For poly(bis-(p-carboxyphenoxy)propane anhydride) cylinders, for example, the degradation rate increases by a factor of 10 when increasing the pH of the degradation medium from 7.4 to 10 (Ref. 46). Poly(ortho esters), in contrast, are resistant against basic pH and degrade substantially faster at acidic compared to neutral pH 47. By using acidic or basic excipients, the degradation rate of polymer hydrolysis can be varied in a controlled way. The internal pH can, thus, effectively be used to influence the degradation rate of polymers 48.
The effect of copolymer composition on polymer degradation By introducing a second monomer into the polymer chain, many properties of the original polymer can be influenced, such as crystallinity or glass transition temperature 49. Such changes have been observed for poly(anhydrides) 5~ where degradation also depends on the copolymer composition. It was shown for poly(1,3-bis-p-carboxyphenoxypropane-co-sebacic acid) (p(CPP-SA)) that degradation depends markedly on the CPP content. Increasing the content of the aromatic monomer from 50 to 100% was reported to increase the time of erosion substantially 51. Other examples are poly(lactic-co-glycolic acid) copolymers, where the decrease of molecular weight during degradation was found to be accelerated with increasing glycolic acid content 52'53. Other factors that depend on the copolymer composition, such as the glass transition temperature and the crystallinity of a polymer, can have additional indirect effects on degradation rates. In general, it can be concluded that the degradation rates of degradable polymers depend on the prevailing type of bond.
The effect of water uptake an degradation Hydrolysis is a bimolecular reaction in which water and the functional group possessing the labile bond are involved. The reaction velocity is determined by the 'concentration' of both reaction partners 54. Lipophilic polymers cannot take up large quantities of water and decrease, thereby, their degradation velocity 46'55. Hydrophilic polymers, in contrast, take up large quantities of water and increase, thereby, degradation rates. The uptake of water is especially important in the area of drug delivery. Hydrogels, for example, may undergo
105
substantial swelling, which for some polymers is the decisive parameter for controlling the release of drugs, and may be more important than polymer degradation.
Influencing polymer degradation In cases where polymers tend to degrade too slowly for a specific application, one might choose to regulate the velocity of chemical degradation. In most cases this is achieved by adding excipients that regulate pH. In drug delivery applications, these can be the drugs themselves that are incorporated into the polymers, such as alkaloids and other bases 56'57 or acids 58. For poly(ortho esters), magnesium hydroxide47 and carboxylic acid anhydrides 59 have been used to modify the degradation of the polymer. The anhydrides accelerate hydrolysis through acid catalysis, whereas magnesium hydroxide decreases degradation rates due to the increased stability of orthoesters in basic media. In the case of poly(e-caprolactone) low-molecular-weight compounds, like ethanol, pentanol, oleic acid, decylamine and tributylamine, were also reported to enhance degradation 6~ Besides adding pH regulating substances, changing the polymer matrix structure has been shown to be a useful tool in controlling degradation rates. There are two principal ways by which this can be achieved: copolymerization and polymer blending. Copolymerization of lactic acid with glycolic acid, for example, increases degradation rates 61. Heller and co-workers have shown that the introduction of acidic and hydrophilic monomers increases water uptake and enhances the autocatalytic degradation 62. Pitt and coworkers observed an increase in degradation rates for soluble blends of poly(vinyl alcohol) and poly(lacticco-glycolic acid) 63. A completely different approach to influencing degradation rates has been proposed by Kost and Langer using ultrasound 64. For poly(anhydrides) a strong dependence of degradation rates on the application of ultrasound w a s f o u n d 65'66, which might be useful for the external regulation of polymer degradation in vivo.
POLYMER EROSION All degradable polymers share the property of eroding upon degradation. Degradation and erosion are the decisive performance parameters of a device made of such materials. To classify degradable polymers a distinction is made between surface (or heterogeneous) and bulk (or homogeneous) eroding materials 67, which is illustrated in Figure 1. During an application, surface eroding polymers lose material from the surface only. They get smaller but keep their original geometric shape. For bulk eroding polymers, degradation and erosion are not confined to the surface of the device. Therefore, the size of a device will remain constant for a considerable portion of time during its application 68. The advantage of surface eroding polymers is the predictability of the erosion process 69. This is desirable when using such polymers for drug delivery, where the release of drugs can be related directly to the rate of polymer erosion 7~ Surface and Biomaterials 1996, Vol. 17 No. 2
The Biomaterials Silver Jubilee Compendium
120 Mechanisms of polymer degradation and erosion" A. G6pferich
106
surface erosion
y
v
/
bulk erosion
,..._ ,...--
Figure I erosion.
pores: macropores with a diameter of approximately 100 ~m which stem from the formation of cracks and micropores with a diameter of approximately 0.1~m that stem from the erosion of polymer bulk. Figure 3 shows that for p(CPP-SA) 20:80 the number of smaller pores increases during erosion, while the number of macropores remains the same. Well defined erosion zones are visible under the light microscope for surface eroding polymers like poly(anhydrides) 77 and poly(ortho esters) 47. Figure 4(a) illustrates an erosion front in a p(CPP-SA) 20:80 disc. Inversely moving erosion fronts have been observed for the autocatalytic degradation of poly(D,L-lactic acid) and poly(D,l.-lacticco-glycolic acid) rods and discs TM that move from the inside of the polymer outward. The preferential erosion of amorphous compared to crystalline polymer parts was observed for enzymatic 79 as well as non-
Schematic illustration of surface erosion and bulk
bulk erosion are ideal cases to which most polymers cannot be unequivocally assigned. Polymer erosion is far more complex than degradation, because it depends on many other processes, such as degradation, swelling, the dissolution and diffusion of oligomers and monomers, and morphological changes. Even more parameters apply to some special types of polymers like electrically erodible materials 71, or during in viva applications 72. Although degradation is the most important process of erosion, depending on the type of polymer, other parameters may also become critical in controlling erosion behaviour. The knowledge of the erosion mechanism is, therefore, most important for the successful application of a degradable polymer. In tissue engineering, surface properties or porosity determine the performance of implantable scaffolds 73. In drug delivery, swelling and porosity are critical to the release behaviour of drugs 68. As with degradation, many different indicators of erosion have been proposed, such as molecular weight loss, sample weight loss and changing geometry. These parameters need not change at the same velocity as the example of poly(anhydrides) illustrates. The molecular weight loss of poly(anhydrides) can be substantial during the first 12 h TM, while there is almost no loss of weight and no change in geometry 34. Erosion is, like degradation, again an individual process for each polymer.
Figure 2 Picture of a p(CPP-SA) 20:80 polymer matrix disc surface after 18.5h in phosphate buffer, pH 7.4, at 37~ taken by scanning confocal microscopy (scale bar m 100/~m). (Reproduced with permission from Ref. 34, ',~(~: 1993, Wiley & Sons.)
Morphological changes during erosion The first morphological changes during erosion are confined to the polymer surface. For poly(anhydrides) the formation of cracks can be observed immediately after contact with buffer 34. Figure 2 shows the surface of a poly(anhydride) after 18.5 h in phosphate-buffered saline, pH 7.4 taken by scanning confocal microscopy, which is covered with cracks. The surface of poly(ortho ester) investigated by atomic force microscopy shows an increasing surface roughness 75. With proceeding erosion, polymers change to more porous structures. Such changes can be detected by mercury intrusion porosimetry TM. The investigation of poly(anhydrides) revealed that there are two types of Biomaterials 1996, Vol. 17 No. 2
Figure 3 Pore size distribution of eroding p(CPP-SA) 20:80 polymer matrix discs determined by mercury intrusion porosimetry (model Poresizer 9 2 2 0 , Micromeritics, Norcross, GA, USA). Penetrometer volume 6ml, size of sample discs: 7mm diameter, 2mm height. Pressure 0.530000psi (3.45-207000 kPa). Pore sizes calculated from Washburn equation for a guessed contact angle of 160'; between mercury and polymer. 9 Day 1; A, day 2; [], day 4.
121
The Biomaterials Silver Jubilee Compendium 107
Mechanisms of polymer degradation and erosion: A. GOpferich ....... . . . . . . .
Thus, anhydrides are cleaved into carboxylic acids, esters and orthoesters into alcohols and carboxylic acids. The degradation products, therefore, influence pH in the degradation medium as well as inside pores. Anhydrides, for example, were shown to affect the pH of the erosion medium substantially. Using pH sensitive dyes in combination with fluorescence scanning confocal microscopy, pH gradients in nonstirred buffered degradation media were detected when approaching the surface of eroding p(CPP-SA) 20:80 discs 34. Figure 5 shows such a profile. It was found that the pH inside the pores of eroding anhydrides is between 4 and 5, which is identical with the pKa of the monomers and far less than the pH of the degradation medium, which was 7.4 (Ref. 34). The findings agree with the results from earlier studies where the pH inside eroding anhydrides was measured using a glass electrode 84. Even more severe deviations of the pH inside the eroding polymer from the pH in the degradation medium were observed for poly(lactic acid) and its copolymers 78, which is due to the higher solubility and the low pKa compared to the poly(anhydride) monomers, pH values as low as 1.8 were measured inside eroding polymer rods 78.
The behaviour of oligomers and monomers during erosion
Figure 4 SEM picture of eroding p(CPP-SA) 20:80 polymer matrix discs, a, Erosion front (middle of the picture) separating eroded (left part) from non-eroded (right part) polymer. b, Eroded spherulite. (Reproduced with permission from Ref. 34, 9 1993, Wiley & Sons.)
enzymatic degradation a~ For poly(3-hydroxybutyrate) this is visible from the appearance of the crystalline spherulitic skeleton of the material a2. The same was observed for poly(anhydrides), where the amorphous parts of p(CPP-SA) 20:80 were less resistant ta erosion than the amorphous ones 34. Figure 4(b) shows the crystalline skeleton of an eroded spheru1Re. The amorphous regions of the spherulite have been eroded, while the crystalline skeleton is still largely in place. These findings were also confirmed for other poly(anhydrides) based on 1,6-bis(p-carboxyphenoxy)hexane, (carboxyphenoxy)methane and 5-(pcarboxyphenoxy]-valeric acid 83.
Changes in pH As already mentioned, the degradation rate depends strongly on pH. Through the chain scission, polymers are transformed into oligomers and monomers, which have different functional groups than the polymer.
During the degradation of polymer chains, oligomers and monomers are created which need not necessarily be released immediately. Lactic acid oligomers, for example, have been reported to form salts that have properties different from those of the protonated compounds aS. Li and Vert observed that poly(D,L-lactic acid) was able to crystallize during the degradation of the polymeric chains 86'a7. They identified the crystals as an oligomeric stereocomplex consisting of poly(Dlactic acid) and poly(L-lactic acid) chains 33, which has been identified and characterized earlier as-92. Monomers created by degradation have also been reported to crystallize during erosion. Differential scanning calorimetry and X-ray diffraction data suggest that the monomers of poly(sebacic acid) and 7.00
J ,
6.75 6.50 6.25 6.00 5.75 5.50 w 9I ' I'--" I ' I -200 -175 -150 -125 -It3() -75
"
]
w"
-50
I
-25
'
l
()
distance from surface [yml Figure 5 pH profile above an eroding p(CPP-SA) 20:80 polymer matrix disc determined by scanning confocal microscopy using fluorescein-5- (and 6)-sulphonic acid as a pH-sensitive fluorescent probe. (Reproduced with permission from Ref. 34, 9 1993, Wiley & Sons.)
Biomaterials 1996, Vol. 17 No. 2
The Biomaterials Silver Jubilee Compendium
122 Mechanisms of polymer degradation and erosion: A. G6pferich
108
p(CPP-SA) polymers crystallize inside eroding polymer matrix d i s c s 34' 93. The release of oligomers and monomers from the polymer bulk has been studied for many polymers. Oligomers have been reported to be released from poly(D,u-lactic acid) microspheres 94 and to increase drug release rates 95. Monomer release profiles for poly(sebacic acid) have a short induction period. The release rate is highest at early times and declines in a concave manner 96. More complex release profiles were obtained for L-lactic acid release from poly(Du-lactic acid). Induction periods are increased to 8 weeks, giving the release profile of L-lactic acid a sigmoidal shape 97, a clear sign of the lower reactivity of the ester bond compared to the anhydride bond. Similar profiles were obtained for the release of lactic and glycolic acids from poly(L-lactic-co-glycolic acid) 75:25 (Ref. 53) and poly(L-lactic-co-glycolic acid) 50:50 (Ref. 15), whereby glycolic acid left the devices approximately twice as fast as lactic acid. Remarkable are the monomer release profiles obtained from poly(anhydride) copolymers 34. The release profiles of sebacic acid are again concave whereas the release profiles of CPP monomer are sigmoidal, as shown in Figure 6. For such differences in the release of individual monomers from copolymers, two mechanisms have been proposed.
degradation was shown to determine the pH inside pores, thereby limiting the solubility of CPP. Whenever sebacic acid has left the device, which according to Figure 6 is after approximately 8 days, the solubility of CPP increases tremendously, visible from the increase in release rate. This example illustrates how intricate the erosion mechanism of biodegradable polymers can be.
1. Erosion controlled release: assuming that the different types of possible bonds in the copolymer backbone are cleaved at different rates, the monomers are set free and, therefore, released at divergent rates. 2. Diffusion controlled release: differences in solubility and diffusivity account for different release rates.
MODELLING OF POLYMER DEGRADATION AND EROSION
For poly(anhydride) copolymers it was proposed that different hydrolysis rates of SA-SA bonds and CPPCPP bonds might lead to different rates at which the monomers are created 15. More important, however, seems to be the solubility of monomers. In the case of poly(anhydrides) this is at any pH approximately 1:10 in favour of sebacic acid 34. Sebacic acid created by t20 10() "~,
80
_~
6o
There are two general sources of crystallinity changes during polymer erosion. One is the generation of crystallized oligomers and monomers. The other stems from the behaviour of partially crystalline polymers during erosion. Due to the faster erosion of amorphous compared to crystalline polymer regions, the overall crystallinity of samples increases, and has been measured for poly(L-lactic acid) 81 and poly(fl-hydroxy butyrate) derived materials 98. Crystallinity also increases during the erosion of intrinsically amorphous polymers like quenched samples of poly(Llactic acid). When introducing these samples to erosion media, their glass transition temperature is lowered due to the uptake of water, which leads to the recrystallization of the polymer 8~
There are many reasons for trying to model polymer degradation and erosion. It would, for example, be very useful if one could predict pH changes on the surface of polymeric scaffolds used in tissue engineering to ensure it is tolerable to attached cells. In drug delivery, proteins and peptides incorporated into polymers might become unstable at extreme pH values, which could be avoided if pH were predictable. The formation of crystallites due to the preferential erosion of amorphous polymer parts might decrease the biocompatibility of implants. All of these problems can only be partially addressed at present because none of the existing models takes into account all of these parameters. In addition, degradation and erosion are often simplified as separate events in modelling schemes, which is not generally the case.
Modelling of polymer degradation J.
4() {}
Changes in crystallinity
2O ()
9
()
,
14
7
9
,
,' ....
r
9
2
time [daysl Figure 6 Release of CPP and sebacic acid (SA) monomer from p(CPP-SA) 20:80. 9 1,3-Bis-p-Carboxyphenoxypropane (CPP); Q, (SA). (Reproduced with permission from Ref. 34, :i" 1993, Wiley & Sons.) B i o m a t e r i a l s 1996, Vol. 17 No. 2
Degradation modelling is not trivial. Major problems arise from investigating the process experimentally, which is necessary to obtain data on which a prospective model can be based. Most degradable polymers are not water soluble and their degradation is influenced by additional factors like swelling, or the kinetics of water uptake. Nevertheless, degradation data were obtained by investigating water soluble oligomers~},.l(,o by using polymer solutions in organic solvent-water mixturesl~ or by investigating degradation in bulk at elevated temperatures ~~ ao4. All these approaches have disadvantages. For example, the degradation of water soluble oligomers can be in equilibrium with the formation of oligomers from monomers in aqueous solutions 36, or the degradation mechanisms
123
The Biomaterials Silver Jubilee Compendium Mechanisms of polymer degradation and erosion: A. G6pferich
109
can be changed by the addition of organic solvents. In most modelling approaches degradation is regarded as a random scission process 1~ assuming first- or secondorder kinetics 1~176To describe the formation of oligomers, random theory has recently been applied in more depth to the description of degradation 1~ and allows the description of the formation of oligomers upon degradation. Little attention, however, has been given to the degradation of copolymers so far.
Modelling of polymer erosion Erosion modelling is even more complex than degradation modelling because of the multitude of involved processes. There are only few appraaches to erosion modelling but none of them covers all processes that are involved in erosion. In early approaches, only heterogeneous erosion was modelled. It was assumed to advance at constant velocity 1~ Similar assumptions were made to investigate spheres and cylinders with concentric bores l~ Next, diffusion theory was introduced to describe the diffusion of low-molecularweight compounds from eroding polymers 1~ Later, moving erosion fronts as well as dissolution fronts for crystalline matter were introduced 11~ A substantial improvement was made when combining the diffusion equation with a reaction term accounting for the degradation of the polymer 112. The degradation of polymer was included into the models under the premise of first-order kinetics for the chain scission 1~3. Recently, the formation and release of oligomers and molecular weight changes were taken into account, also using a diffusion/reaction equation 114'115. All these approaches relied on differential equations for describing erosion. A completely new way of modelling erosion takes advantage of random theory. The erosion of small polymer pieces is regarded as a random event, that cannot be predicted when it will occur, but the likelihood of which is known for any time 116. Similar approaches have been used before for modelling the erosion of controlled release devices 117 and have been developed recently for the optimization of drug release from bioerodible materials 118. The advantage is the inclusion of parameters such as shape, crystallinity, porosity and tortuosity. First, polymer matrices are partially covered using a twodimensional computational grid, as shown in Figure 7(a). The grid divides cross-sections into individual pixels representing crystalline and amorphous polymer areas. Figure 7(b) shows such a grid, on which dark pixels represent crystalline polymer areas and white pixels represent amorphous areas. As poly(anhydrides) are surface eroding, it was assumed that only pixels in contact with the buffer medium can erode. Erosion was assumed to be a Poisson process. The lifetime of a pixel, i.e. the time between the first contact with the erosion medium and the erosion, for such a process is distributed according to a first-order Erlang distribution. Crystalline and amorphous pixels differ by their erosion rate constants, which provide higher likelihood for amorphous pixels to erode. By removing eroded pixels continuously from the grid, time series like the one shown in Figure 8 are obtained 1~6. From such simulations, many experimen-
Figure 7 a, Schematical representation of a polymer matrix cut-out by a computational grid. b, Theoretical representation of a polymer matrix prior to erosion (black pixels, crystalline areas; white pixels, amorphous areas). (Reproduced with permission from Ref. 116, '~ 1993, ACS.)
tally measurable parameters can be calculated, like porosity or weight loss. Figure 9 shows the fit to experimental data for the erosion of p(CPP-SA) 20:80. The fit allows the determination of the erosion rate constants and illustrates that the model is quite well able to adjust to the experimental data. The Monte Carlo model, unfortunately, does not account for the release of incorporated drugs, oligomers or monomers from Biomaterials 1996, Vol. 17 No. 2
The Biomaterials Silver Jubilee Compendium
124 Mechanisms of polymer degradation and erosion: A. G6pferich
110
dependence of solubility on pH (Ref. 120). Such comprehensive models can then be used to describe the complex behaviour of monomer release. Figure 10 shows the fit of the model to the data of Figure 6. As a better quality criterion than the apparently good fit, the model's ability to predict experimental data was tested by predicting functions other than monomer release. Figure 11 shows the prediction of the suspended mass of monomers from the same erosion experiment as in Figure 10. It is apparent that such modelling approaches can be used for predicting data for eroding systems. Despite some progress in the area of modelling, much more data and more sophisticated models are needed to apply these approaches to other degradable polymers. A S P E C T S OF F U T U R E R E S E A R C H The future research in this area will have to focus on experimental aspects of eroding systems. More information on the processes are needed for a better Figure 8 Simulation of polymer erosion using a Monte Carlo model (black pixels, non-eroded areas; white pixels, eroded areas). (Reproduced with permission from Ref. 116, ,~) 1993, ACS.)
1.2 1.0 9.
0.8 0.6 0.4
0.2 0.0 0
7
14
21
time[days] Figure 10 Fit of erosion model to experimental data. 9 1,3-bis-p-Carboxyphenoxypropane (CPP); I , sebacic acid (SA). (Reproduced with permission from Ref. 120, :.~:ii 1994, Elsevier.)
time
60
[daysl
Figure 9 Fit of Monte Carlo model to experimental data. 9 Erosion front position; Q, relative polymer matrix disc mass. (Reproduced with permission from Ref. 116, :~ 1993, ACS.)
50,,.-..,
40 30
eroding polymer matrix discs. For the description of such transport phenomena, diffusion theory has to be applied. Equation 1 describes the one-dimensional diffusion equation in porous media119:
0 0 OC(x,t) O--tC(x, t)e(x, t) - - ~ DeffC(x,t)~:(x, t) Ox
o 6,
20 1() ()
(1)
where C, e, D~ff, t and x are concentration of the diffusant inside pores, porosity, effective diffusivity, time- and space-variables, respectively. The function can be expanded to describe additional phenomena such as the dissolution of suspended drug, or the Biomaterials 1996, Vo|. 17 No. 2
9
()
1
2
3
4
5
6
7
time Idays] Figure 11 Predicted and experimentally measured mass of suspended monomers contained in an eroding p(CPP-SA) 20:80 poly(anhydride) disc. 9 1,3-bis-p-Carboxyphenoxypropane (CPP); Q, sebacic acid (SA). (Reproduced with permission from Ref. 120, (~:~1994, Elsevier.)
The Biomaterials Silver Jubilee C o m p e n d i u m
125
Mechanisms of polymer degradation and erosion: A. G6pferich
understanding of the b e h a v i o u r of d e g r a d a b l e polymers. I n f o r m a t i o n on pH, osmotic pressure, the fate of m o n o m e r s , c h a n g e s in crystallinity as well as m o n o m e r a n d oligomer release and solubility are vital for a better design of devices m a d e of d e g r a d a b l e polymers. Once such information is available, i m p r o v e d m o d e l s can be d e v e l o p e d that m i g h t h e l p to p r e d i c t d e g r a d a t i o n and erosion of p o l y m e r s m o r e accurately. C o n c e p t s a d v a n c e d by such m o d e l s m a y be n e c e s s a r y to fully explore the e n o r m o u s potential of b i o d e g r a d a b l e polymers. REFERENCES
1
2
3 4 5
6 7 8
9 10 11
12 13 14 15 16 17
111
18 19
20 21
22
Brem H, Walter K, Langer R. PoLymers as controlled drug delivery devices for the treatment of malignant brain tumors. Eur ] Pharmacol Biopharmacol 1993;
23
39: 2-7.
24
Alonso MJ, Cohen S, Park, TG, Gupta R, Siber G, Langer R. Determinants of release rate of tetanus vaccine from polyester microspheres. Pharmaceut Res 1993; 10', 945-953. Esparza I, Kissel T. Parameters affecting the immunogenicity of microencapsulated tetanus toxoid. Vaccine 1992; 10' 714-720. Gref R, Minamitake Y, Peracchia MT, Trubetskoy V, Torchilin V, Langer R. Biodegradable long-circulating polymeric nanospheres. Science 1994; 383: 1600-1603. G6pferich A, Gref R, Minamitake J e t al. Drug delivery from bioerodible polymers: systemic and intravenous administration. In: Cleland J, Longer R, eds. Protein Formulations and Delivery (ACS Symposium Series 567). Washington, DC: ACS, 1994; 242-277. Heller J. Chemically self-regulated drug delivery systems. J Controlled Rel 1988; 8: 111-125. Kost J, Langer R, Responsive polymer systems for controlled delivery of therapeutics. Trends Biotechnol 1992; 10: 127-131. Leenslang JW, Pennings AJ, Ruud RM, Rozema FR, Boering G. Resorbable materials of poly(L-lactide). VI. Plates and screws for internal fracture fixation. Biomateria]s 1987; 8: 70-73. Langer R, Vacanti JP. Tissue engineering. Science 1993; 260: 920-926. Engelberg I, Kohn J. Physico-mechanical properties of degradable polymers used in medical applications. Biomaterials 1991; 12: 292-304. Garton A, Henry JL, McLean PD, Stevenson WT. Laboratory simulation of space and salt water environment effects on polymers. Polym Engng Sci 1989; 29: 23-28. Doi Y, Kanesawa Y, Tanahashi N. Biodegradation of microbial polyesters in the marine environment. Polym Degrad Stab 1992; 36:173-177. Golub MA. Reactions of atomic oxygen [O(3P)] with polymer films. Makromol Chem Makromo] Syrup 1992; 53: 379-391. Vert M, Feijen J, Albertson A, Scott G, Chiellini E. Degradable Polymers and Plastics. Melksham: Redwood Press Ltd, 1992: 73-92. Tamada J, Langer R. Erosion mechanism of hydrolytically degradable polymers. Proc Nat Acad Sci USA 1993; 90: 552-556. Banford CH, Tipper CFH. Comprehensive Chemical Kinetics (Vol 14: Degradation of Polymers). New York: Elsevier, 1972. Grassie N, Scott G. Polymer Degradation and Stabilization. New York: Cambridge University Press, 1985.
25
26 27 28 29 30 31 32 33 34
35
36 37 38 39
Sepp/il/i J, Linko Y-Y, Su T, Photo- and biodegradation of high volume thermoplastics. Acta Polytech Scand 1991; 198: 10-12. Spenlehauer G, Vert M, Benoit JP, Boddaert A. In vitro and in rive degradation of poly(D,U lactide/glycolide) type microspheres made by solvent evaporation method. Biomaterials 1989; 10: 557-563. Chatfield DA, Einhorn IN. Stepwise thermal degradation of a polybenzimidazole foam. J Polym Sci Chem 1981; 19: 601-618. Brandwood A, Noble KR, Schindhelm K et al. Influences of stress on biodegradation of novel biomedical polyurethanes. Adv Biomater 1992; 10: 413-420. Miller ND, Williams DF. The in rive and in vitro degradation of PGA suture material as a function of applied strain. Biomaterials 1984; 5: 365-368. Herrmann JB, Kelly RJ, Higgins GA. Polyglycolic acid sutures. Arch Surg 1970; 100: 486-490. Wise DL, Fellmann TD, Sanderson JE, Wentworth RL. Lactic/glycolic acid polymers. In: Gregoriadis G, ed. Drug Carriers in Biology and Medicine. New York: Academic Press, 1979' 237-270. Heller J, Himmelstein KJ. Biodegradable poly(ortho esters) as drug delivery forms. In: Borchardt RT, Arnold J, Stella vJ, eds. Directed Drug Delivery. Clifton, NJ: Humana, 1985: 171-188. Tamada J, Langer R. The development of polyanhydrides for drug delivery applications. J Biomater Sci Polym Edn 1992; 3.4: 315-353. Shalaby SW. Biomedical Polymers --Designed-toDegrade Systems. Munich: Hanser Publishers, 1994. Shalaby SW, Ikada Y, Langer R, Williams J. Polymers of biological and biomedical significance. ACS Symposium Series 1994; 540. Lenz R. Biodegradable polymers. In: Peppas NA, Langer RS, eds. Advances in Polymer Science (No 107: Biopolymers I). Berlin: Springer Verlag, 1993: 1-40. Tomihata K, Burczak K, Shiraki K, Ikada Y. Crosslinking and biodegradation of native and denaturated collagen. Polym Prepr 1992; 33: 534-535. Cox MK. The effect of material parameters on the properties and biodegradation of 'Biopol'. Biodegrad Polym Plastics 1992; 109: 95-105. Park K, Shalaby WSW, Park H. Biodegradable Hydrogels for Drug Delivery. Lancaster: Technomic Publ., 1993. Li S, Vert M. Crystalline oligomeric stereocomplex as an intermediate compound in racemic poly(DL-lactic acid) degradation. Polym Int 1994; 33: 37-41. G6pferich A, Langer R. The influence of microstructure and monomer properties on the erosion mechanism of a class of polyanhydrides. J Polym Sci 1993; 31: 24452458. Laitinen O, T6rm/il~ P, Taurio R, Skutnabb K, Saarelainen K, Iivonen T, Vainionp~/i S. Mechanical properties of biodegradable ligament augmentation device of poly(L-lactide) in vitro and in vivo. Biomaterials 1992; 13: 1012-1016. Holten CH. Lactic Acid, Properties and Chemistry of Lactic Acid and Derivatives. Weinheim: Verlag Chemie, 1971. Fan LT, Singh SK. Controlled Release--A Quantitative Treatment. Berlin: Springer-Verlag, 1989: 105-106. Baker R. Controlled Release of Biologically Active Agents. New York: John Wiley & Sons, 1987: 84-131. St. Pierre T, Chiellini E. Biodegradability of synthetic polymers used for medical and pharmaceutical applications: Part 1--principles of hydrolysis mechanisms. J Bioact Compatible Polym 1986; 1: 467---497. Biomaterials 1996, eel. 17 No. 2
The Biomaterials Silver Jubilee Compendium Mechanisms of polymer degradation and erosion: A. G6pferich
112 40 41 42
43
44 45 46
47 48 49 50
51
52
53
54
55 56 57
58 59
126
Sykes P. A Guidebook to Mechanism in Organic Chemistry (4th edn). London: Longman Group Ltd, 1975: 232-239. Shih C, Higuchi T, Himmelstein KJ. Drug delivery from catalysed erodible polymer matrices of poly(ortho ester)s. Biomaterials 1984; 5: 237-240. Kirby AJ. Hydrolysis and formation of esters of organic acids. In: Bamford CH, Tipper CFH, eds. Comprehensive Chemical Kinetics (Vol 10: Ester Formation and Hydrolysis and Related Reactions). Amsterdam: Elsevier, 1972: 57-202. Pitt CG. Non-microbial degradation of polyesters. In: Vert M, Feijen J, Albertson A, Scott G, Chiellini E, eds. Biodegradable Polymers and Plastics. Melksham: Redwood Press Ltd, 1992: 7-20. Chu CC. A comparison of the effect of pH on the biodegradation of two synthetic absorbable sutures. Arch Surg 1982; 1913: 55-59. Vert M, Li S, Garreau H. More about the degradation of LA/GA-derived matrices in aqueous media. J Controlled Rel 1991; 16: 15-26. Leong KW, Brott BC, Langer R. Bioerodible polyanhydrides as drug-carrier matrices. I: Characterization, degradation and release characteristics. J Biomed Mater Res 1985; 19: 941-955. Heller J. Controlled drug release from poly(orthoesters}. A surface eroding polymer. J Controlled Ret 1985; 2: 167-177. Heller J. Poly(ortho esters). In: Peppas N, Langer R, eds. Advances in Polymer Science (Vol 107). Berlin: Springer Vertag, 1993: 41-93. Hiemenz PC. Polymer Chemistry. New York: Marcel Dekker, 1984: 423-504. Mathiowitz E, Ron E, Mathiowitz G, Amato C, Langer R. Morphological characterization of bioerodible polymers. 1. Crystallinity of polyanhydride copolymers. Macromolecules 1990; 23: 3212-3218. Chasin M, Domb A, Ron E et al. Polyanhydrides as drug delivery systems. In: Chasin M, Langer R, eds. Biodegradable Polymers as Drug Delivery Systems. New York: Marcel Dekker, 1990: 43-71. Miller R, Brady J, Cutright D. Degradation rates of oral resorbable implants (polylactates and polyglycolares): rate modification with changes in PLA/-GA copolymer ratios. J Biomed Mater Res 1977; 11: 711719. Li SM, Garreau H, Vert M. Structure-property relationships in the case of the degradation of massive poly(a-hydroxy acids) in aqueous media. Part 2: degradation of lactide-glycolide copolymers: PLA37.5 GA25 and PLA75GA25. J Mater Sci Mater Med 1990; 1: 131-139. Pitt CG, Chasalow FI, Hibionada YM, Klimas DM, Schindler A. Aliphatic polyesters. I. The degradation of poly(e-caprolactone) in vivo. J Appl Polym Sci 1981; 26" 3779-3787. Ron E, Turek T, Mathiowitz E et al. Controlled release of polypetides from polyanhydrides. Proc Nat Acad Sci USA 1993; 90: 4176-4180. Cha Y, Pitt CG. The acceleration of degradation o controlled drug delivery from polyester microspheres. J Controlled Rel 1989; 8: 259-265. Yoshioka S, Kishida A, Izumikawa S et al. Baseinduced polymer hydrolysis in poly(fl-hydroxybutyrate/fl-hydroxyvalerate) matrices. J Controlled Rel 1991; 16: 341-348. McGinity JW. Influence of drugs and enzymes on the stability of bioerodible polymers. Polym Prepr (Am Chem Soc, Div Polym Chem) 1990; 31: 187-188. Shih C, Higuchi T, Himmelstein KJ. Drug delivery from
Biomaterials 1996, Vol. 17 No. 2
60
61 62
63
64
65
66
67
68 69 70 71 72 73
74 75
76 77 78 79
catalysed erodible polymeric matrices of poly(ortho ester)s. Biomaterials 1984; 5: 237-240. Pitt CG, Gu Z-W. Modification of the rates of chain cleavage of poly(~:-caprolactone) and related polyesters in the solid state. J Controlled ReI 1987" 4: 283-292. Dahlmann I, Rafter G, Fechner K, Mehlis B. Synthesis and properties of biodegradable aliphatic polyesters. Br Polym J 1990; 23: 235-240. Heller J, Penhale DW, Fritzinger BK, Ng SY. The effect of copolymerized 9,10-dihydroxystearic acid on erosion rates of poly(ortho esters) and its use in the delivery of levonorgestrel. J Controlled Rel 1987' 5' 173-177. Pitt CG, Cha Y, Shah SS, Zhu KJ. Blends of PVA and PGLA: control of the permeability and degradability of hydrogels by blending. J Controlled Rel 1992' 19: 189200. Kost J, Leong K, Langer R. Ultrasound-enhanced polymer degradation and release of incorporated substances. Proc Nat Acad Sci USA 1989" 86: 76637666. Liu L-S, D'Emanuele A, Kost J, Langer R. A study on the mechanism of polymer degradation enhanced by low intensity ultrasound. Proc Int Syrup Control Rel Bioact Mater 1991" 18: 225-226. Liu L-S, D'Emanuele A, Kost J, Langer R, Experimental approach to elucidate the mechanism of ultrasoundenhanced polymer erosion and release of incorporated substances. Macromolecules 1992; 25: 123-128. Langer R, Peppas N. Chemical and physical structure of polymers as carriers for controlled release of bioactive agents' a review. J Macromol Sci-Rev Macromol Chem Phys 1983-C23: 61-126. Longer R. New methods of drug delivery. Science 1990' 2 4 9 : 1527-1532. GSpferich A, Langer R. Predicting drug release from cylindric polyanhydride matrix discs. Eur J Pharmacol Biopharmaco] 1995; 41" 81-87. GSpferich A, Langer RS. Modeling of polymer erosion in three dimensions - - Rotationally symmetric devices. AIChE J 1995' 41: 2292-2299. Kwon IC, Bae YH, Kim SW. Electrically erodible polymer gel for controlled release of drugs. Nature 1991' 394" 291-293. Williams DF. Mechanisms of biodegradation of implantable polymers. Clin Mater 1992" 10: .9-12. Vacanti CA, Vacanti JP, Langer R. Tissue engineering using synthetic biodegradable polymers. In' Shalaby SW, Ikada Y, Longer R, Williams J, eds. Polymers of Biological and Biomedica] Significance (ACS Symposium Series 540). Washington, DC: ACS" 1994. D'Emanuele A, Hill j, Tamada ]A, Domb A. Longer R. Molecular weight changes in polymer erosion. Pharmaco] Res 1992; 9: 1279-1283. Shakesheff KM, Davies MC, Domb A et al. Visualizing the degradation of polymer surfaces with an atomic force microscope. Proc Int Symp Control Rel Bioact Mater 1994; 21' 618-619. Mikos AG, Thorsen AJ, Czerwonka LA, Bao Y, Langer R. Preparation and characterization of poly(I.-lactic acid) foams. Polymer 1994; 35: 1067-1078. Mathiowitz E, Ron E, Mathiowitz G et al. Surface morphology of bioerodible polyanhydrides. Polym Prepr 1989" 30:460-461. Herrlinger M. In Vitro Polymerabbau und Wirksttofffeigabe yon Poly-oL-Laktid-Formlingen. PhD thesis, Heidelberg, 1994. Nishida H, Tokiwa Y. Effects of higher-order structure
127
The Biomaterials Silver Jubilee Compendium Mechanisms of polymer degradation and erosion- A. G6pferich
80
81 82
83
84
85 86
87
88 89
90
91
92
93
94 95
of poly(3-hydroxybutyrate) on its biodegradation. II. Effects of crystal structure on microbial degradation. J Environ Polym Degrad 1993; 1" 65-80. Li SM, Garreau H, Vert M. Structure-property relationships in the case of the degradation of massive poly(~hydroxy acids) in aqueous media. Part 3: influence of the morphology of poly(L-lactic acid). ] Mater Sci Mater Med 1990; 1: 198-206. Pistner H, Bendix DR, Mfihlig J, Reuther jF. Poly(Llactide): a long-term degradation study & viva. Biomateriols 1993; 4: 291-298. Doi Y, Kumagai Y, Tanahashi N, Mukai K. Structural effects on biodegradation of microbial and synthetic poly(hydroxyalkanoates). Biodegrad Polym Plastics 1992; 109: 139-148. Mathiowitz E, Jacob J, Pekarek K, Chickering DIII. Morphological characterization of bioerodible polymers. 3. Characterization of the erosion and intact zones in polyanhydrides using scanning electron microscopy. Macromolecules 1993; 28: 6756-6765. Laurencin C. Novel bioerodible polymers for controlled release analyses of in vitro~in viva performance and characterizations of mechanism. PhD thesis. MIT, 1987. Wada R, Hyon S-H, Ikada Y. Salt formation of lactic acid oligomers as matrix for sustained release of drugs, l Pharm PharmacoI 1991; 43: 605-608. Li S M, Vert M. Morphological changes resulting from the hydrolytic degradation of stereocopolymers derived from L- and ~3L-|actides. Macromolecules 1994; 27: 1307-1310. Li SM, Garreau H, Vert M, Therin M, Christel P e t al. In viva degradation mechanism of massive aliphatic polyesters derived from lactic and glycolic acids. In: Vert, M, Feijen J, Albertson A, Scott G, Chiellini E, eds. Biodegradable Polymers and Plastics. Redwood Press Ltd, 1992: 7-20. Ikada Y, Jamshidi K, Tsuji H, Hyon S-H. Slereocomplex formation between enantiomeric poly(lactides). Macromo]ecules 1987; 20: 904-906. Tsuii H, Horii F, Hyon S-H, Ikada Y. Stereocomplex formation between enantiomeric poly(L-lactic acid)s. 2. Stereacomplex formation in concentrated solutions. Macromolecules 1991; 24: 2719-2724. Tsuji H, Hyon S-H, Ikada Y. Stereocomplex formation between enantiomeric poly(L-lactic acid)s. 3. Calorimetric studies on blend films cast from dilute solution. Macromolecules 1991; 24: 5651-5656. Tsuji H, Hyon S-H, Ikada Y. Stereocomplex formation between enantiomeric poly(L-lactic acid)s. Differential scanning calorimetric studies on precipitates from mixed solutions of poly(o-lactic acid) and poly(L-lactic acid). Macromolecules 1991; 24: 56575662. Tsuji H, Hyon S-H, Ikada Y. Stereocomplex formation between enantiomeric poly(IAactic acid)s. 5. Calorimetric and morphological studies on the stereocomplex formed in acetonitrile solution. Macromolecules 1992; 25: 2940-2946. G6prerich A, Langer R. The effect of erosion on the microstructure of biodegradable polymers. Proc Int Syrup Control Rel Bioact Mater 1993; 20: 133134. Park TG. Degradation of poly(O-L-lactic acid) microspheres: effect of molecular weight. ] Controlled Rel 1994; 30: 161-173. Bodmeier R, Oh KH, Chen H. The effect of addition of low molecular weight poly(D,L-lactide} on drug release from biodegradable poly(D,h-lactide) drug delivery systems. Int 1 Pharmacol 1985; 51: 1-8.
113
96 97
98
99
100
101 102 103
104
105
106
107
108
109 110
111
112
113
Tamada J, Longer R. Mechanism of the erosion of polyanhydride drug delivery systems. Proc lnt Syrup Rel Bioact Mater 1990; 17: 156-157. Li SM, Garreau H, Vert M. Structure-property relationships in the case of the degradation of massive aliphatic poly(~-hydroxy acids) in aqueous media, Part 1: poly(BL-lactic acid). ] Mater Sci Mater Med 1990; 1: 123-130. Yasin M, Tighe BJ. Strategies for the design of biodegradable polymer systems: manipulation of polyhydr0• materials. Plastic Rubber Camp Proc Appl 1993; 19: 15-27. Maniar ML, Kalonia DS, Simonelli AP. Use of liquid chromatography and mass spectroscopy to select an oligomer representative of polyester hydrolysis pathways. J Pbarmacol Sci 1989; 78" 858-862. Maniar ML, Kalonia DS, Simonelli AP. Determination of specific rate contents of specific oligomers during polyester hydrolysis. ] Pharmacol Sci 1991; 80" 778782. Domb A, Langer R. Solid-state and solution stability of poly(anhydrides} and poly(esters). Macromolecules 1989; 22: 2117-2122. Zhang X, Wyss UP, Pichora D, Goosen MFA. An investigation of poly(lactic acid) degradation. J Bioact Compat Polym 1994; 9: 80-100. Buchholz B. Accelerated degradation test on resorbable polymers. In: Planck H, Dauner M, Renardy M, eds. Degradation Phenomena on Polymeric Biomaterials. Heidelberg: Springer Verlag, 1992: 67-75. Wallis KH, Miiller RH. Comparative measurement of nanoparticle degradation velocity using an accelerated hydrolysis test. Pharmacol Ind 1993; 55: 168170. Pitt CG, Gratzel MM, Kimmel JL et a]. Aliphatic polyesters. II. The degradation of poly(D-L-lactide), poly(e-caprolactone), and their copolymers in viva. Biomaterials 1981; 2: 215-220. Granger G. Stochastic modeling for oligomers produced by degradation of linear polymers. In: Vert M e t al. eds. Degradable Polymers and Plastics. Melksham: Redwood Press Ltd, 1992:191-199. Hopfenberg HB. Controlled release from bioerodible slabs, cylinders and spheres. In: Paul DR, Harris FW, eds. Control]ed Release Polymeric Formulations (ACS Symp Set No 33). Washington, DC: ACS, 1976: 26-32. Cooney DO. Effect of geometry on the dissolution of pharmaceutical tablets and other solids: surface detachment kinetics controlling. AIChE J 1972; 18: 446-449. Baker RW, Lonsdale HK. Erodible controlled release system. Am Chem Sac Div Org Coat Plast Chem Prepr 1976; 3: 229. Lee PI. Diffusional release of a solute from a polymeric matrix--approximate analytical solutions. ] Membr Sci 1980; 7: 255-275. Thombre AG, Himmelstein KJ. Modeling of drug release kinetics from a laminated device having an erodible drug reservoir. Biomaterials 1984; 5: 250254. Thombre AG. Theoretical aspects of polymer biodegradation: mathematical modeling of drug release and acid-catalyzed poly(ortho-ester) biodegradation. Biodegrod Polym Plastics 1992; 109: 214-225. Heller J, Baker RW. Theory and practice of controlled drug delivery from bioerodible polymers. In: Baker RW, ed. Controlled Release of Bioactive Materials. New York: Academic Press, 1980: 1. Biomaterials 1996, Vol. 17 No. 2
The Biomaterials Silver Jubilee Compendium Mechanisms of polymer degradation and erosion: A. G6pferich
114
114 115 116 117
128
Himmelstein KJ, George J. Erosion of polymer with autocatalysis. Proc Int Syrup Control Rel Bioact Mater 1993; 20: 53-54. Himmelstein KJ, Heterogeneous mechanisms of polymer erosion. Polym Prepr (Am Chem Soc, Div Polym Chem 1992; 33: 48-49. G6pferich A, Langer R, Modeling polymer erosion. Macromolecules 1993; 16: 4105-4112. Zygourakis K. Development and temporal evolution of erosion fronts in bioerodible controlled release
Biomaterials 1996, Vol. 17 No. 2
118 119 120
devices. Chem Engng Sci 1990; 45: 2359-2366. Zygourakis K. Design and optimization of bioerodible devices with optimal release characteristics. Mater Res Soc Proc 1993; 331: 79-84. Saltzmann WM, Langer R. Transport rates of proteins in porous materials with known microgeometry. Biophys J 1989; 55: 163-171. G6pferich A, Langer R. Modeling monomer release from bioerodible polymers. ] Controlled Rel 1995: 33: 55-69.
129
The Biomaterials Silver Jubilee Compendium
Biomaterials 17 (1996) 115-124 (~) 1996 Elsevier Science Limited Printed in Great Britain. All rights reserved 0142-9612/96/$15.00
ELSEVIER
Stabilized polyglycolic acid fibrebased tubes for tissue engineering D.J. Mooney *t* C L. Mazzoni* C. Breuer* K. McNamara* D. Hern*, J.P. Vacanti* and R. Langer* 9
"
~
9
9
*Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge, MA 02139, USA; tDepartment of Surgery, Harvard Medical School and Children's Hospital, Boston, MA 02115, USA
Polyglycolic acid (PGA) fibre meshes are attractive candidates to transplant cells, but they are incapable of resisting significant compressional forces. To stabilize PGA meshes, atomized solutions of poly(L-lactic acid) (PLLA) and a 50/50 copolymer of poly(D,L-lactic-co-glycolic acid) (PLGA) dissolved in chloroform were sprayed over meshes formed into hollow tubes. The PLLA and PLGA coated the PGA fibres and physically bonded adjacent fibres. The pattern and extent of bonding was controlled by the concentration of polymer in the atomized solution and the total mass of polymer sprayed on the device. The compression resistance of devices increased with the extent of bonding, and PLLA bonded tubes resisted larger compressive forces than PLGA bonded tubes. Tubes bonded with PLLA degraded more slowly than devices bonded with PLGA. Implantation of PLLA bonded tubes into rats revealed that the devices maintained their structure during fibrovascular tissue ingrowth, resulting in the formation of a tubular structure with a central lumen. The potential of these devices to engineer specific tissues was exhibited by the finding that smooth muscle cells and endothelial cells seeded onto devices in vitro formed a tubular tissue with appropriate cell distribution.
Keywords: Tissue engineering, polyglycolic acid, polylactic acid, smooth muscle cells, endothelial cells Received 26 October 1994; accepted 5 January 1995
to engineer a variety of tissues, including liver, cartilage and intestine 3. This class of polymers degrades by a simple hydrolysis mechanism, and by varying the ratio of lactic and glycolic acids in the polymer one can control the crystallinity of the polymer, and thus its degradation rate and mechanical properties 4. Furthermore, these polymers can be processed to yield a variety of different structures, including fibres, hollow tubes and porous sponges 5-7. Non-woven meshes of polyglycolic acid (PGA) fibres have been particularly attractive materials for use as cell delivery devices as they are highly porous, permitting diffusion of nutrients throughout the device following implantation while allowing subsequent neovascularization of the developing tissue, and they can be easily fabricated into devices with varying geometry. However, this material lacks the structural stability to withstand compressive forces in vivo, and external supports are necessary if one desires to form a stable three-dimensional structure (e.g. a tube) from this material 8' 9. In this study, we investigated whether threedimensional structures capable of resisting large compressive forces and guiding the formation of a desired tissue structure could be formed from PGA fibre meshes by physically bonding adjacent fibres using a spray casting method. Poly(L-lactic acid)
While organ transplantation and tissue reconstruction are highly successful therapies for a variety of maladies, a shortage of donor tissue limits their application to a percentage of those who could potentially benefit from these therapies. For example, over 83 000 people either died or were maintained on less-thanoptimal therapies due to a lack of donated organs in the USA in 19901. To aid these people, a variety of investigators have proposed to engineer new tissues by transplanting isolated cell populations on biomaterial scaffolds to create functional new tissues in vivo 2. To engineer complex tissues such as blood vessels or intestine, cells must be localized to a specific site in vivo, and the formation of an appropriate tissue structure from the implanted cells and the host tissue must be promoted. Biodegradable materials are particularly attractive for fabricating the devices utilized ta transplant cells and engineer new tissues because they can be designed to erode after tissue development is complete, leaving a completely natural tissue 2'3. Templates synthesized from polymers of the lactic and glycolic acid family have previously been utilized tCurrent address: Departments of Biological and Materials Sciences and Chemical Engineering, University of Michigan, Ann Arbor, MI 48109, USA. Correspondence to Prof. R. Longer. 115
Biomaterials 1996, Vol. 17 No. 2
130
The Biomaterials Silver Jubilee Compendium 116
(PLLA) or a 50/50 copolymer of lactic and glycolic acids (PLGA) was dissolved in chloroform, atomized and sprayed over a PGA mesh formed into a tubular structure. Following solvent evaporation, a physically bonded structure resulted, and the pattern and extent of PGA fibre bonding was controlled by the processing conditions. These tubular devices were capable of withstanding large compressive forces in vitro (50200mN) and maintained their structure in vivo. The specific mechanical stability was dictated by the extent of physical bonding and the polymer utilized to bond the PGA fibres.
MATERIALS The PGA mesh (fibre diameter approximately 121~m; mesh thickness=0.3 ram, specific gravity= 80.2mgcm-3, porosity= 97%) was purchased from Albany Int. (Taunton, MA, USA), PLLA and the poly(D,L-lactic-co-glycolic acid) from Medisorb (Cincinnati, OH, USA), the lactic dehydrogenase kit, glycolic and lactic acid standards, and 4,5-dihydroxy2,7-naphthalenedisodium salt were purchased from Sigma Chemical Co. (St Louis, MO, USA), chloroform from Mallinckrodt (Paris, KY, USA), phosphatebuffered saline and DMEM medium from Gibco (Grand Island, NY, USA), Tmax film from Kodak, Lewis rats (250-300g) from Charles River (Wilmington, MA, USA), calf serum from Hyclone Lab. Inc. (Logan, UT, USA), penicillin and strepromycin from Irvine Scientific (Santa Ana, CA, USA), and methoxyflurane from Pitman-Moore Inc. (Mundelein, IL, USA).
METHODS Tube fabrication Rectangles (1.3 x 3.0cm) of the non-woven mesh of PGA fibres were wrapped around a Teflon cylinder (outside d i a m e t e r - 3.0 mm) to form a tube, and the two overlapping ends were manually interlocked to form a seam. The Teflon cylinders were then rotated at 20rpm using a stirrer (Caframo; Wiarton, Ontario, Canada). Solutions of PLLA and PLGA dissolved in chloroform (1-15%, w/v) were placed in a dental atomizer (Devilbus Corp.) and sprayed over the rotating PGA mesh from a distance of 6in (.~15cm) using a nitrogen stream (18 psi (.~.124.2kPa)) to atomize the polymer solution. The PLGA and PLLA had molecular weights (Mw) of 43 400 (M,,/M,, = 1.43) and 74100 (M,./M~ = 1.64), respectively. Molecular weights were determined by gel permeation chromatography as described previously 7. While PLLA and copolymers of lactic and glycolic acids are soluble in chloroform, PGA is very weakly soluble in this solvent. Thus, the PGA fibres are largely unchanged by the process. After spraying was completed, the tubes were lyophilized to remove residual solvent, removed from the Teflon cylinder and cut into specific lengths. The tubes were sterilized by exposure to ethylene oxide for 24h, followed by degassing for 24 h. Biomaterials 1996, Vol. 17 No. 2
Stabilized PGA tubes D.J, Mooney et al.
Device characterization The mass of PLLA and PLGA that bonded to the PGA scaffolds was determined by weighing PGA devices before and after spraying. For scanning electron microscopic examination, samples were gold coated using a Sputter Coater (Desk II, Denton Vacuum, Cherry Hill, NJ, USA). An environmental scanning electron microscope (Electro Scan , Wilmington, MA, USA) was operated at 30 kV with a water vapour environment of 5 torr (~665 Pa) to image samples. Photomicrographs were taken with Polaroid 55 film. Thermal mechanical analysis was performed with a TMA 7 (Perkin Elmer Corp, Norwalk, CT, USA) using a compression probe with a circular tip (d -- 3.0 mm). All testing was done at a constant temperature of 37"C. Tubes were placed on their sides for testing (axis of tube lumen perpendicular to the axis of force application), and the change in device diameter (parallel to the direction of force application) was followed during and after force application. The compressional forces applied to the tubes in vivo will presumably also be in a radial direction. The resulting deformations were normalized to the initial device diameter. Some samples were pre-wet by placing them in a vial containing phosphate-buffered saline and incubating at 37"C for 24 h. All tests were performed in triplicate, and representative data are given. The erosion characteristics of bonded devices were assayed by' placing individua| tubes in 5 ml of phosphate-buffered saline, pH 7.4, and incubating under static conditions at 37'C. The mass loss was analysed by weighing lyophilized devices before and after the incubation period. The release of lactic acid was assayed enzymatically with lactic dehydrogenase using a kit from Sigma. The release of glycolic acid was quantitiated with a colorimetric assay ~~ which involves decarboxylating glycolic acid in the presence of concentrated sulphuric acid to form formaldehyde, followed by reaction of formaldehyde with chromotropic acid to yield a coloured product which can be quantitated spectrophotometrically.
Implantation of tubes Polymer constructs were implanted into the omentum of syngeneic Lewis rats as described previously ~. NIH guidelines for the care and use of laboratory animals (NIH Publication No. 85-23 Rev. 1985) have been observed in all experiments involving animals. Inhalation anaesthesia with methoxyflurane was always utilized. The omental tissue was rolled around the devices to promote tissue invasion and neovascularizalion of the implants from all sides. Implants were secured in place with sutures of 7-0 Maxon (Davis and Geck). Recipients of polymer devices were killed on post-implantation days 3 and 18. The implants were removed, fixed in 10% buffered formalin and thin sections were cut from paraffin-embedded tissue. Histological sections were stained with haematoxylin and eosin. Photomicrographs were taken with Kodak Tmax fihn.
Cell seeding on devices To introduce bovine aortic smooth muscle cells (passage 6-9) into the polymeric: delivery devices,
The Biomaterials Silver Jubilee Compendium
131
Stabilized PGA tubes: D.J. Mooney et al.
l ml of a cell suspension containing 5 - 2 0 x 1 0 5 cellsm1-1 was injected into the interior of each tube using a l m l syringe and a 22-gauge needle. The cell suspension was retained in the tubes by placing a small plug of the PGA fibres at both ends of the tubes during the cell adhesion period. Devices were incubated at 37~ in an atmosphere of 10% CO2 to allow for cell adhesion and proliferation. The tubes were manually rotated periodically using sterile forceps during the period of cell adhesion to promote even cell seeding. Cell-polymer devices were kept in DMEM medium, containing 5% calf serum, 100Um1-1 penicillin and 100mgm1-1 streptomycin, during this time. The seeding protocol was repeated 7 days later to ensure even seeding of cells within the devices. Ten days later, a cell suspension of bovine aortic endothelial cells (passage 6-9) was similarly seeded onto the tubes. After 4 more days the devices were fixed in formalin, embedded in paraffin, sectioned and stained (haematoxylin and eosin) using standard techniques. Sections were stained for the presence of desmin (a smooth muscle specific protein) and Factor 8 (specific for endothelial cells) using standard immunohistochemical protocols. Antibodies for this analysis were purchased from Shandon (Pittsburgh, PA, USA). The endothelial cells and smooth muscle cells were isolated from bovine aortas using a collagenase digestion, and were a gift from Dr Judah Folkman.
RESULTS Bonding tubes with PLLA To determine whether PGA scaffolds could be stabilized by physically bonding adjacent fibres, chloroform containing dissolved PLLA (1-15% w/v) was sprayed over the exterior surface after the PGA mesh was wrapped around a Teflon cylinder to form a tube. The PLLA formed a coating over the exterior PGA fibres after the solvent evaporated, and physically bonded adjacent fibres. The tubes formed in this manner could be easily removed from the Teflon cylinder for characterization and use, The pattern of bonding was controlled by the concentration of the PLLA in the atomized solution (Figure I), even though the time of spraying was adjusted to maintain an approximately constant mass of PLLA on the devices under the various conditions (Table 1). Spraying with a solution containing 1 or 5% PLLA resulted in extensive bonding of PGA fibres without significantly blocking the pores of the PGA mesh. Spraying with a 10% solution of PLLA also bonded fibres, but resulted in the formation of a PLLA film on the exterior surface of the PGA mesh that contained only small pores. Spraying with a solution containing 15% PLLA had a similar effect, although the polymer film that formed was less organized. In all cases, the PLLA coated and bonded fibres only on the exterior surface of the PGA mesh, as no coating or bonding of fibres was observed on the interior surface of the PGA mesh (Figure 2). The compression resistance of bonded tubes was assessed in vitro to determine which patterns of
117
bonding resulted in the most stable devices. Unbonded tubes were completely crushed by a force of 5 mN, but banded tubes were capable of resisting forces in excess of 200raN. However, the ability of bonded tubes to resist a given compressional force was dependent on the pattern of bonding (Figure 3). For example, tubes bonded with 1 or 15% PLLA were significantly compressed by a force of 200mN, while tubes bonded with a solution of 5 or 10% PLLA were only slightly compressed by this force. The compression was viscoelastic in all cases, as the devices only partially decompressed after the force was removed. Uniform properties were observed with respect to the position along and around a tube. To determine if the extent, as well as the pattern, of bonding could vary the compression resistance of tubes, an atomized dispersion of 5% PLLA was then sprayed over the devices for different times. Lengthening the spraying time from 10 to 60s increased the mass of PLLA on the devices (Table 2). Infrequent bonds between adjacent fibres resulted from spraying for 10 s. Spraying for more extended periods increased the PLLA coating over the PGA fibres, and the extent of bonding (Figure 4). The ability of these tubes to resist compressional forces and maintain their shape was quantitated again using thermal mechanical analysis. The compression resistance strongly depended on the extent of bonding, as tubes that were more extensively bonded had a greater resistance to deformation (Figure 5A). The compression that did occur under these conditions was again a combination of a reversible, elastic strain and an irreversible deformation. Some tubes were also exposed to an aqueous environment before testing to determine whether this environment for 24h would destabilize the tubes. The aqueous environment had a slight detrimental effect on the stability of bonded tubes, but they were still capable of resisting large compressive forces (Figure 5B).
Bonding tubes with PLGA To determine whether this technique of stabilizing PGA devices could be utilized with a variety of polymers, the previous study was repeated using a 50/50 copolymer of lactic and glycolic acids. The mass of polymer bonded to the devices and the extent of physical bonding were again regulated by the time an atomized dispersion of the bonding polymer was sprayed over the PGA fibres (Table 2; Figure 6). Once again, bonding increased the compression resistance of devices formed into a tubular structure (Figure 7A). However, these devices were not able to resist the same compressional forces as PLLA bonded devices. Tubes bonded with PLLA were capable of resisting forces up to 200mN, while tubes bonded with PLGA were only capable of resisting forces slightly greater than 50 rnN. The difference between devices stabilized with PLLA and PLGA was even more striking when the devices were tested after immersion in phosphate-buffered saline for 24h. PLGA bonded tubes, in contrast to PLLA bonded tubes, were significantly weakened by this treatment
(Figure 7t3).
Biomaterials 1996, Vol. 17 No. 2
132
The Biomaterials Silver Jubilee Compendium Stabilized PGA tubes" D.J. Mooney et ai.
118
Figure 1
P h o t o m i c r o g r a p h s of the exterior surface of PGA m e s h e s formed into tubular structures and sprayed with solutions
containing (A) 1%, (B) 5%, (C) 10% and (D) 15% PLLA. The spraying time was varied to yield an approximately constant mass of sprayed PLLA in all conditions. The original magnifications and size bars are shown in the photomicrographs.
Table 1 PGA mesh sprayed with solutions of varying PLLA concentration .
.
.
.
-
PLLA concentration (w/v)
Spraying time (s)
Mass of PLLA on device* (% initial PGA mass)
1 5 10 15
150 30 15 10
115+20 168 • 16 145 _L. 12 108 4- 73
*VaLues represent the mean 4-s.d. of three devices,
Tube degradation in vitro The time course for erosion of the tubes was determined by quantitating the mass loss and monomer release from tubes immersed in a pH balanced, isotonic saline solution. Devices bonded with PLGA were completely degraded by 11 weeks, while devices bonded with PLLA only lost 30% of their mass after 10 weeks (Figure 8A). The degradation of the PLLA bonded tubes was solely due to erosion of the PGA fibres, as glycolic acid was released from the Biomaterials 1996, Vol. 17 No. 2
tubes, but virtually no lactic acid was released over this time from the tubes (Figure 8B). PLLA degrades slowly, and no significant loss of PLLA mass is expected until 1-2 years. Erosion of tubes bonded with PLGA was due to erosion of both the PLGA fibres and the PLGA, as both glycolic acid and lactic acid were released from the tubes over this time flame (Figure
8c).
Compression resistance in vivo To confirm that stabilized tubes were capable of resisting compressional forces in vivo as well as in vitro, devices bonded with PLLA (5% PLLA; 30s spraying time) were implanted into the omentum of laboratory rats. The initial (3 day) host response was characterized by fibrin deposition and scattered inflammatory cells throughout the devices. A mature fibrovascular tissue was evident throughout the devices by 7 days, and the devices maintained their tubular structure with a central lumen for the 18 day duration of the experiment (Figure 9A). The invading fibroblasts and the newly deposited matrix were aligned with the lumens of the tubes (Figure 9B).
133
The Biomaterials Silver Jubilee Compendium Stabilized PGA tubes: D.J. Mooney et al.
119
Cell adhesion and organization in vitro on bonded tubes
Figure 2 A photomicrograph of the interior surface of PGA mesh formed into a tubular structure and sprayed with a solution of 5% PLLA for 30s. The interior surface, in contrast to the exterior surface (see Figure 1), was largely unaffected by this process. The original magnification and size bar are shown in the photomicrograph.
PLLA bonded tubes (5% PLLA; 30s spraying time) were subsequently seeded with smooth muscle cells and endothelial cells to investigate the suitability of these devices to serve as cell delivery vehicles. Blood vessels are largely comprised of these two cell types. The smooth muscle cells adhered to the polymer fibres (Figure I OA and B), and proliferated to fill the void space present between polymer fibres (Figure lOB). Endothelial cells also adhered to the devices, and over time formed a lining on the interior section of the devices (Figure I OA and C). Immunohistochemical staining for desmin confirmed that the cells filling the interstices between polymer fibres were smooth muscle cells, and staining for Factor 8 confirmed that the cells lining the luminal surface were endothelial in nature (not shown). This organization of the muscle and endothelial cells is similar to that observed in blood vessels.
DISCUSSION
Figure 3 Representative strain diagrams of tubes formed from the PGA mesh after spraying with a solution containing r-I, 1%; B, 5%; O, 10%; and O, 15% PLLA. Devices were subjected to a compressive force of 200 mN applied in a direction perpendicular to the axis of the device lumen starting at 0min. The force was removed at 10 min, and the change in the diameter of the tube (parallel to the direction of force application) was monitored both during and after the time of force application, and normalized to the initial diameter.
Table 2 Spraying PGA scaffolds for various times with a 5% solution of PLLA or PLGA Spraying time (s)
Mass of PLLA on device* (% initial PGA mass)
Mass of PLGA on device* (% initial PGA mass)
10 20 30 60
43:t: 160 • 165 + 390 +
54+9 59 + 40 140 + 10 313 • 51
11 55 22 37
*Values represent the mean • s.d. of three devices.
Three-dimensional tubes can be formed from PGA fibre scaffolds by physically bonding adjacent fibres. The compression resistance and degradation rate of these devices were controlled by the pattern and extent of physical bonding, and the type of polymer utilized to bond the PGA fibres. Fibrovascular tissue invaded the devices following implantation, leading to the formation of a tubular tissue with a central lumen. The potential of these devices to engineer tissues was exhibited by the finding that endothelial cells and smooth muscle cells adhered to the devices and formed a new tissue in vitro with appropriate tissue organization. The compression resistance of devices was monitored by applying a constant force on the tubes. The resulting changes in the device diameters were partially elastic, as indicated by the partial decompression following removal of the applied force. The irreversible changes in the device diameters were likely caused by both crushing and bending of fibres, and by rearrangement of fibres. Contact between the compression tip and the tubes was not analysed, and will likely change as the tubes compress and fibres rearrange. For this reason, results were reported for compressional forces, not stresses. Calculation of stresses using the entire contact area of the compression probe would give the most conservative estimate of mechanical moduli. Tubes which were bonded with PLLA were more resistant to compressional forces than tubes bonded with PLGA. This finding is not surprising, as crystalline PLLA is typically much stiffer than amorphous PLGA 4. Additionally, while the compression resistance of PLLA bonded devices was not greatly changed after exposure to an aqueous environment, PLGA bonded devices were markedly weakened after the same treatment. PLGA is more hydrophilic than PLLA 4 due to the presence of the glycolic acid residues, and the absorbed water likely acts as a plasticizer, weakening Biomaterials 1996, Vol. 17 No. 2
134
The Biomaterials Silver Jubilee Compendium 120
Stabilized PGA tubes: D.J. Mooney et ai.
Figure 4 Photomicrographs of the exterior surface of PGA meshes formed into tubular structures and sprayed with solutions containing 5% PLLA for (A) 10, (B) 20, (C) 30 and (D) 60 s. The original magnifications and size bars are shown in the photomicrographs,
the PLGA. The PLLA bonded devices were slightly weakened after this treatment, indicating that the PLLA was also somewhat plasticized. The erosion of the devices was also dependent on the polymer utilized for bonding. PLLA is hydrolysed very slowly, and virtually no lactic acid release was observed over the 10 weeks of the erosion study. The erosion of PLLA bonded devices was entirely due to hydrolysis of the glycolic acid bonds in the fibres. In contrast, both the PGA fibres and the PLGA used to bond the fibres eroded completely over 11 weeks. The release of glycolic acid from these devices occurred more rapidly than the release of lactic acid. This was likely caused by the more rapid erosion of the PGA fibres, followed by the slower release of both lactic acid and glycolic acid from the PLGA. Biodegradable devices are attractive for cell transplantation and tissue engineering since they can be designed to erode once tissue development is complete, leaving a completely natural tissue. The approach described in this report to mechanically stabilize fibre-based scaffolds was performed with PGA, PLGA and PLLA because of the long history of these polymers in medical devices, and the range of degradation rates that can be obtained with this class Biomaterials 1996, Vol. ] 7 No. 2
of polymers (Figure 8). However, this technique could potentially be used with a variety of other polymers, both erodible and non-erodible, for medical or nonmedical applications. Various approaches have previously been taken to mechanically stabilize structures formed from PGA fibres. PGA fibres can be physically bonded with a second polymer in a similar manner as described here by simply dipping the PGA scaffold into a solution of PLLA dissolved in chloroform, and allowing the chloroform to evaporate ~. Alternatively, a thermal processing technique that results in temporary melting and subsequent bonding of PGA fibres has been reported ~2. The bonding approach described in this report is simple, permits a variety of bonding polymers to be utilized and allows the fabrication of various threedimensional scaffolds. It also results in bonding only of the outermost fibres of the device (Figure 2), in contrast to the other methods. This preserves the desirable features of the PGA mesh (high porosity, high surface area/polymer mass ratio) throughout the interior sections. This approach also allows both the extent and pattern of bonding to be easily controlled. Extensive coating and bonding of fibres resulted when the polymer concentration in the atomized solution
The Biomaterials Silver Jubilee Compendium
135
Stabilized PGA tubes: D.J. Mooney et al.
A
100 80
k,,
q)
e9~
|
u >
sec
30 20
sec sec
10
sec
B
100 ~
~
Control
80
!._
(D
60
E m N9
40
==,,
Pre-wet
60
E
E I,.
40
>
o.,.
.,m
C3
60
A
E q) N r .-a
121
20 a
0
0
20
--.
Time
~.
t0
..
. ....
(minutes)
i
20
0
0
....
1
i
,
5
~0
~s
Time
--"1
z0
(minutes)
Figure 5 Representative strain diagrams of (A) PGA tubes sprayed for various times with a 5% PLLA solution and subjected to a compressive force of 200 mN starting at 0 rain. The force was removed at 10min. The force application and the change in the diameter of the tube (normalized to the initial diameter) were monitored, as described in the legend for Figure 3, both during and after the time of force application. (B) Devices sprayed with a 5% PLLA solution for 30s and tested dry (Control) or after pre-wetting for 24 h in a saline solution (Pre-wet). The compressional force was again 200 mN.
Figure 6 Photomicrographs of the exterior surface of PGA meshes formed into tubular structures and sprayed with solutions containing 5% PLGA for (A) 10, (B) 20, (C) 30 and (D) 60s. The original magnifications and size bars are shown in the photomicrographs. Biomaterials 1996, Vol. 17 No. 2
136
The Biomaterials Silver Jubilee Compendium Stabilized PGA tubes: D.J. Mooney et al.
122 100
"-"
|
.u_>,.-c
80
60
sec
30
sec
10
100 b ~,
,,...o._r
Control
80
sec
60
3 ~
60
40
u ... c ";
|
g,
40
Pre-wet
g, 20
200 -
0
u
1'0
Time
2'0
(minutes)
0
1'0
TIME
(minutes)
20
Figure 7
Representative strain diagrams of (A) PGA tubes sprayed for various times with a 5% PLGA solution and subjected to a compressive force of 50 mN starting at 0 min. The force was removed at 10 minutes, and the change in the diameter of the tube (normalized to the initial diameter) was monitored both during and after the time of force application. (B) Devices sprayed with a 5% PLGA solution for 30s and tested dry (Control) or after pre-wetting for 24h in a saline solution (Pre-wet), The compressional force in this test was again 50 mN.
A
B 100
W W cu
Q
6
80
W 0
PLLA
80
(I}
m
60 o
E~
PLGA
'-"
0
e~
40 "" 20
0
20
4
6
10
8
cld
m
q)
"gae
a
100
0
12
/ 0
.ot,o .c,, 2
4
Time
(wk)
Time
6
8
10
12
(wk)
C (I)
w r (I) ~
Se
IX: =
_s
t.
9 ii
E E
o~ C ,-0 5
100
Olycollc
80
acid
60 4o
Lactic
acid
2O
0
2
4
Tlme
6
8
10
12
(wk)
The degradation of devices bonded by spraying with PLLA or PLGA (5% solution; spraying time .... 30s), as measured by (A) quantitating the change in device mass over time, or (B) the release of glycolic and lactic acids from PLLA bonded devices, or (C) PLGA bonded devices. Devices were incubated at 37"C under static conditions in buffered saline and removed at various times for analysis, Values in (A) represent the mean and standard deviation calculated from three samples. Figure 8
Biomaterials 1996, Vol. 17 No. 2
137
The Biomaterials Silver Jubilee Compendium Stabilized PGA tubes: D.J. Mooney et al.
123
Figure 9 (A) Low-power and (B) high-power photomicrographs of a histological section from a bonded tube (5% PLLA; 30s) implanted for 17 days in the omentum of a Lewis rat. These cross-sections of the implanted device were cut perpendicular to the axis of the tube's lumen. (A) The central lumen (I) is visible, along with numerous polymer fibres (arrows), the host omental tissue (o), and the ingrown fibroblasts and fibrous tissue they deposited. (B) The fibroblasts which invaded the device and the fibrous tissue deposited by these cells aligned in parallel with the central lumen. The original magnifications of these photomicrographs were (A) • 16 and ( B ) x 158.
was low (1-5%) (Figure 1A and B). Increasing the concentration of polymer in the atomized solution to 10% resulted in the formation of a relatively smooth film over the external surface of PGA meshes, and utilizing a 15% solution resulted in the formation of a fibrous, non-homogeneous film over the PGA meshes (Figure 1C and D). Increasing the polymer concentration raises the viscosity of this solution and this likely increases the droplet size which is formed during the atomization process. This will effect how these droplets penetrate the PGA mesh, how they aggregate on the PGA mesh, and the rate of solvent evaporation. All of these factors will affect the pattern of bonding. To engineer a tissue with a desired three-dimensional structure, the cell delivery device must maintain a preconfigured geometry in the face of external forces during the process of tissue development. While the magnitude of the compressive forces that are exerted on implanted devices by the surrounding tissue are unclear, they are significant and will vary depending on the implant site. The magnitude of forces utilized in the present study to quantitate the compression resistance of devices in vitro was 50-200mN. This
Figure 10 (A) Low-power photomicrograph of a histological section of a bonded tube (5% PLLA; 30s) seeded with smooth muscle cells and endothelial cells in vitro as described in the Methods section. This cross-section was cut perpendicular to the axis of the tube's lumen. Highpower photomicrographs of (B) an interior section of the device and (C) a section adjacent to the lumen. Smooth muscle cells readily adhered to polymer fibres (p) and filled the interstices between polymer fibres (A and B), while endothelial cells formed a lining on the luminal surface (A and C; arrows). The original magnifications of these photomicrographs were (A) • (B) • and (C) • 158.
results in pressures ranging from approximately 50 to 200mmHg (6.65-26.6kPa) (assuming complete and continuous contact between the TMA compression tip and the tube). These pressures are in the same range Biomaterials 1996, Vol. 17 No. 2
138
The Biomaterials Silver Jubilee Compendium
Stabilized PGA tubes: D.J. Mooney et ai.
124
observed in blood vessels. Devices which were stable to high forces (PLLA bonded devices) were also stable after implantation into the ornentum of laboratory rats. The omentum was chosen as the implant site because it is highly vascularized, easily accessed and manipulated surgically, and its anatomic location makes it a preferred site to engineer a variety of gastrointestinal tissues (e.g. small intestine). The compressional forces exerted by the surrounding tissue are likely not as great as other potential implant sites (e.g. popliteal space). The formation of fibrovascular tissue in implanted tubes was not surprising, as it is well documented that this type of ingrowth occurs in porous, synthetic materials13,14. The ingrowth and organization of the fibrovascular tissue will also exert compressional forces on the forming tissue, although the magnitude of these forces is unclear. It is anticipated that the ingrowing fibrovascular tissue would have eventually filled the central lumen of the implanted tubes since there was no epithelial cell lining of the lumen. An endothelial cell lining would likely prevent this outcome. While large diameter synthetic blood vessels (>5 m m diameter) have been successfully utilized for years, prosthetic small diameter blood vessels (<5ram diameter) have been unsuccessful. Various investigators have attempted to improve the performance of small diameter grafts by either lining them with endothelial cells before i m p l a n t a t i o n15, promoting endothelial cell migration from the adjacent vessels 16, or engineering blood vessels using extracellular matrix molecules as the template for tissue organization 17. This last approach showed that the cell types utilized in the present study (endotheliai and smooth muscle cells) have the ability to reform a tissue with the appropriate structure if placed on an appropriate matrix. The synthetic, biodegradable tubes described in this paper may provide a means to provide appropriate mechanical properties to an engineered blood vessel while also promoting the formation of a complete and natural replacement from the appropriate cell types. It may also be possible to combine the advantages of synthetic polymers (tailored mechanical and degradative properties, reproducible synthesis) with the biological specificity of extracellular matrix molecules such as collagen by producing templates from synthetic, biodegradable polymers which contain biologically active amino acid side chains TM. Using the appropriate cell signalling molecules on these polymers may allow one to promote endothelial cell adhesion in desirable spatial locations while preventing other cell types from adhering TM. The approach outlined in this report to stabilize fibre-based cell delivery devices could also be utilized to engineer a variety of other tubular tissues (e.g. intestine, trachea) and non-tubular tissues 2~
REFERENCES
1 2 3 4 5 6 7 8
9
10 11
12
13 14
15 16 17 18
19 ACKNOWLEDGEMENTS
The authors would like to acknowledge Dr Betsy Schloo for the preparation of histological sections. Financial support for this work was provided by the National Science Foundation (BCS-9202311) and Advanced Tissue Sciences.
Biomaterials 1996, Vol. 17 No. 2
20
The National Cooperative Transplantation Study. Sponsored by the United Network for Organ Sharing, Richmond, VA, USA. Langer R, Vacanti JP. Tissue engineering. Science 1993; 260: 920-932. Mooney DJ, Vacanti JP. Tissue engineering using cells and synthetic polymers. Transplant Rev 1993; 7: 153162. Gilding DK. Biodegradable polymers. In: Williams DF, ed. Biocompatibility of Clinical Implant Materials. Boca Raton, FL: CRC Press, 1981. Frazza EJ, Schmitt EE. A new absorbable suture. J Biomed Mater Res 1971; 1: 43. Mooney DJ, Organ G, Vacanti JP, Langer R. Design and fabrication of cell delivery devices to engineer tubular tissues. Cell Transplant 1994; 3:203-210. Mikos AG, Thorsen AJ, Czerwonka LA, Bao Y, Langer R. Preparation and characterization of poly(r.-lactic acid) foams. Polymer 1994; 35: 1068-1077. Organ GM, Mooney DJ, Hanson LK, Schloo B, Vacanti JP. Transplantation of enterocytes utilizing polymerceil constructs to produce neointestine. Transplant Proc 1992; 24: 3008-3009. Organ GM, Mooney DJ, Hansen LK, Schloo B, Vacanti JP. Enterocyte transplantation using cell-polymer devices to create intestinal epithelial-lined tubes. Transplant Proc 1993; 25: 998-1001. Tan S. A coloroinetric assay for glycolic acid. Clin Chim Acta 1978; 89: 13-23. Vacanti CA, Cima LG, Ratkowski D, Upton J, Vacanti JP. Tissue engineered growth of new cartilage in the shape of a human ear using synthetic polymers seeded with chondrocytes. In: Cima L(;, Ron ES, eds. Tissue Inducing Biomateria]s. Materials Research, Pittsburgh, PA, 1992; 252: 367-374. Mikos AG, Bao Y, Cima LG, lngber DE, Vacanti JP, Langer R. Preparation of poly(glycolic acid) bonded fiber structures for cell attachment and transplantation. J Biomed Mater Res 1993; 27: 183. Wesloski SA, Fries CC, Karlson KE, Bakey M, Sawyer PN. Porosity: primary determinant of ultinlate fate of synthetic vascular grafts. Surgery 1961; 50: 91-96. White RA, Hirose FM, Sproat RW, Lawrence RS, Nelson RJ. Histopathologic observations after short-term implantation of porous elastomers in dogs. Biomaterials 1981 ', 2: 171-176. Herring MB. Endothelial cell seeding. J Vascular Surg 1991; 5: 731-732. Clowes AW. Graft endothelialization: the role of angiogenic mechanisms. [ Vascular Surg 1991: 5: 734736. Weinberg CB, Bell E. A blood vessel model constructed from collagen and cultured vascular cells. Science 1986; 231: 397-400. Barrera DA, Zylstra E, Lansbury PT, Langer R. Synthesis and RGD peptide modification of a new biodegradable copolymer: poly(lactic acid-co-lysine). J Am Chem Soc 1993; 115: 11010-11011. Hubbell JA, Massia SP, Desai NP, Drumheller PC. Endothelial cell-selective materials for tissue engineering in the vascular graft via a new receptor. Biotechnology 1991; 9: 568-572. Puelacher WC, Mooney D, Langer R, Upton J, Vacanti JP, Vacanti CA. Design of nasoseptal cartilage replacements synthesized from biodegradable polymers and chondrocytes. Biornaterials 1994; 15: 774-778.
139
The Biomaterials Silver Jubilee Compendium
Biomaterials 17 (1996) 187-194 9 1996 Elsevier Science Limited Printed in Great Britain. All rights reserved 0142-9612/96/$15.00
ELSEVIER
Poly( -hydroxy acids): carriers for bone morphogenetic proteins Jeffrey O. Hollinger and Kam Leong*
Division of Plastic and Reconstructive Surgery, Oregon Health Sciences University, 3181 SW Sam Jackson Park Road, Portland, OR 97201-3098, USA; *Department of Biomedical Engineering, The Johns Hopkins University, 731 Ross Building, Baltimore, MD 21205, USA
A broad spectrum of cells and cell products is associated with bone homeostasis and the renewal of bone following injury. The coupled interactions among cells provide the power behind sculpting of bone, sustaining form, and ensuring functionality. Local and systemic regulatory molecules (e.g. growth factors, hormones) direct cellular interactions through autocrine, paracrine, and hormonal pathways. Recently, genes for a class of osteogenic regulatory molecules have been cloned, and gene product expression has enabled investigators to assess safety and efficacy in animal studies. The molecules are known as bone morphogenetic proteins (BMPs). Therapeutic applications of BMPs depend on a carrier system. A carrier could spatially and temporally localize BMP for regional needs and be custom-tailored for acute craniofacial applications or for recalcitrant extremity non-unions. The poly(~-hydroxy acids) (PHAs) may be suitable for these applications. Therefore, the purposes of this paper are (i) to mention, briefly, basic concepts of the bone wound continuum and the possible therapeutic roles of BMPs; (ii) to outline several properties of selected PHAs relevant to bone regeneration dynamics; and (iii) to review selected preclinical studies with PHAs. Keywords: Poly(2-hydroxy acids), bone morphogenetic protein, bone repair Received 7 November 1994; accepted 11 January 1995
with these modalities has been emphasized16. Engineering bone regeneration constructs mandates recapitulation of the bone wound continuum of cell recruitment, mitogenesis, differentiation, cell product expression, and the dynamic interactions with extracellular matrix repair 3' 7.17, 18. The bone induction cascade of allograft-induced bone formation typifies the bone wound continuum, and was chronicled in the seminal work by Reddi and Anderson TM, later validated by o t h e r s 2~ A pivotal molecule deduced from autograft and extracted from allogeneic bank bone, and touted as a 'trigger' for the cascade, was described first by Urist 23, its role in bone induction posited and its name bestowed: bone morphogenetic protein (BMP) 23-25. A legion of bone researchers, clinicians, and patients will be indebted forever to Urist's persistence and dedication to the basic biology and clinical applications of BMP. BMP is a component of the transforming growth factor beta 'superfamily', and consists of nine members: BMP-1 to BMP-9 (Ref. 26 and Wozney, J.M., personal communication). The BMPs are marphogenetic molecules that provide cues for bone development and regeneration, and are needed for morphodifferentiation in species from fruit fly to Xenopus to m a n 27-29. BMPs do not exist in solitude, a porridge of collegially dependent growth factors are operant in bone maintenance and repair 3~ What distinguishes BMP-2 to BMP-9 from the other growth
THE BONE W O U N D C O N T I N U U M The chronobiological fabric of cells and cell products is woven immutably into bone homeostasis and its renewal after injury. Events follow a strict taskmaster: time. The events are the local and systemic regulatory mechanisms maintaining skeletal physiology and promoting bone regeneration of fractures following reduction and fixation. Cells, cell products, and their spatial and temporal localization through instructive interactions with extracellular matrix have been described in several penetrating studies on fracture repair 1-7. However, regardless of time, bony wounds often will not regenerate spontaneously and require autografts or bank bone allografts to restore form and function. These procedures are performed between 100 000 and 200 000 times annually in the U S A 8'9. For various reasons, autografts fail 13-30% of the time 1~ whereas the failure rate for bank bone is higher 11. Bone grafting is increasing and the spectre of failures is unacceptable a2. Moreover, the risk from allogeneic tissue is a compelling incentive to develop immunologically privileged alternatives 13-15. Development of alternatives requires understanding why current therapies are, in general, successful. Autograft and allografl repair has been reviewed, and the chronobiological progression of bone restoration Correspondence to Dr J.O. Hollinger. 187
Biomaterials 1996, Vol. 17 No. 2
140
The Biomaterials Silver Jubilee Compendium Poly(,~-hydroxy acids): J.O. Hollinger and K. Leong
188
factors is that only BMP-2 to BMP-9 fulfill the osteoinductive principle. In a skeletal muscle, pluripotential cells are induced by BMP to differentiate to cartilage-forming and bone-forming cells in a time-dependent cascade to yield an ossicle with periosteum and haematopoietic m a r r o w 19'24'25. Therefore, BMPs appear to be logical contenders to replace autografts and allografls. However, a means to deploy BMP is key to successful clinical application. A carrier is needed. Ideally, a 'buffet' of carriers could be engineered that spatially and temporally position BMP for optimum bone regeneration. A number of candidate materials may fulfill these demanding criteria, one being the class of polymers known as the poly(:~-hydroxy acids) (PHAs). POLY(~-HYDROXY ACIDS)
(PHAs)
Garnering the spotlight for biomedical applications have been PHA homopolymers of polylactide (PL), polyglycolide ( P G ) , and their copolymers of poly(lactide-co-glycolide) (PLG). While other PHAs have been and continue to be developed for biomedical applications, the emphasis for this review will be on PL, PG, and PLG. PG and PLG have a 25-year history of clinical efficacy and safety as sutures 32. Consequently, as potential carriers for BMPs, they 'enjoy regulatory favour'. However, PHA suture biocompatibility does not ensure biocompatibility of PHAs formatted as bulk devices, such as carrier systems for BMPs, growth factors, or bone fixation. Some of the biocompatibility problems with PHA devices will be described later in this section. PHAs are appealing because they biodegrade in the body (reviewed by Hollinger and Battistone33). 'Biodegradation', the 'removal process', has evoked an array of terms: resorbable, erodible, absorbable, degradable and a subset of terms with the prefix 'bio'. Our preference is to include the prefix 'bio' to denote that the process of PHA removal occurs within a biological environment 34. For consistency within this paper, we will refer to PHAs as biodegradable. PL and PG are usually synthesized by ring opening of the appropriate six-membered lactone monomer (i.e. lactide and glycolide35). Lactide and glycolide can be prepared by dimerization of hydroxy acid precursors, lactic and glycolic acids, respectively 3'~. Lactic acid (LA) has a chiral carbon, therefore, it may polarize light to the left (laevorotatory:l) or right (dextrorotatory:d). Glycolic acid (GA) lacks a chiral centre and is optically inactive. Configurations of optically active molecules are standardized against glyceraldehyde, and there are two possible structures (enantiomers) for LA, identified by small capital L and D36'37, Frequently, upper case L and lower case 1 have been used indiscriminately to describe PL. They denote different molecular properties. L-LA is a metabolite in mammalian tissue, polarizes light to the right, and its configuration is L (referenced to glyceraldehyde) 38. Consequently, the designation L(+)-LA describes this molecule. There is a D enantiomer of LA, specifically, t3(-)LA. When the D Biomaterials 1996, Vol. 17 No. 2
and L enantiomers are polymerized, combinations of six-membered lactone rings result. Nieuwenhuis :~5 describes these six-membered rings as D,D-lactide, L,Llactide, and D,L-lactide (mesolactide), whereas an equimolar composition of D,D- and L,L-lactide is known as racemic lactide. Meso- and racemic lactides are optically inactive, whereas the D,D- and [.,Lstereoforms rotate polarized light in concurrence with their configuration: n,D- rotates polarized light to the right, and L,L- to the left. Chemical properties of homopolymers of D,D-, L,L-, D,L(meso)-, and D,D,L,L (racemic)-lactides are different35.3.q. Therefore, confusing observations can result from improperly identified homopolymers used in wound-healing studies. Intensifying the confllsion are the permutations of copolymers that can be synthesized from various combinations of lactide and glycolide precursors. Synthesis of linear copolymers from the sixmembered lactide and glycolide rings can be accomplished by four copolymerization techniques' bulk, melt, suspension, and solution. The simplest and most commonly used technique is melt-polymerization ring opening which takes place above the melting point of the polymer. To obtain ultra-high molecular weight, the polymerization temperature needs to be lowered to minimize depolymerization and side reactions. Solution and suspension techniques can produce polymers that, in general, are more reproducible than those from other methods because of better control of polymerization temperatures. However, solution and suspension techniques are used infrequently. Bulk-polymerization is best suited to produce high-molecular-weight polymers. This process is accomplished at temperatures between the melting point of the cyclic monomer precursor and the softening point of the polymer. It is not advocated for large-scale polymer synthesis production. PHA synthesis and application for devices will not be explored in this review, but are mentioned relevant to biocompatibility issues that will be addressed. Polymer syntheses involve catalysts for initiating the ringopening reactions. Commonly used catalysts are metal oxides and carboxylates. As residuals of the polymerization reaction, they may lead to adverse tissue responses during polymer biodegradation. Moreover, biomedical polymeric devices fabricated by solvent casting may contain residual solvent, causing unfavourable reactions at the implant site. Furthermore, thermally, processed polymers can cause different crystallinities within the same polymer and, therefore, biocompatibility and biodegradation will be affected. Different stereoforms (enantiomers) of lactides will yield PLs with different properties. Moreover, copolymers of PL and PG, the PLGs, can be difficult to synthesize with the same physical and chemical characteristics from batch to batch. Potential copolymer variability and improperly and/or incompletely characterized homo- or copolymers of PHAs used in animal wound model studies will produce diverse, unpredictable responses. However, one response is predictable" biodegradation. The interaction with water heralds the process of non-specific hydrolysis, the first phase of PHA
The Biomaterials Silver Jubilee Compendium Poly(~-hydroxy acids): J.O. Hollinger and K. Leong
biodegradation. PG is more hydrophilic than PL and its biodegradation is more rapid. In contrast, the pendant methyl groups of either D,D-PL or L,L-PL render it more than hydrophobic. Therefore, its biodegradation is slower than PG. PLG copolymers will biodegrade according to the constituent molar ratio. For example, copalymers with a molar ratio having a predominance of one species over another will be more resistant to hydrolysis than intermediary compositions 39. In addition, the constituent sequence of individual moieties comprising each macromolecular chain (of the PLG copolymer) will influence biodegradation. Alternating diads versus blocks of similar components will impact significantly on the diversity of host responses to PHAs, regardless of whether the molar ratios are identical 33. Furthermore, amorphous and crystalline regions of PL homopolymers will biodegrade differently: the amorphous component preferentially biodegrades first by hydrolysis 4~ In their insightful review, Vert et a]. 41 mention that, besides chemical structures of PHAs, 19 other variables can influence biodegradation. The authors state: "In spite of the rather large number of animal and human clinical experiments carried out with LA/ GA polymers .... one of the most striking features is... the.., remarkable discrepancies between data obtained apparently under similar conditions...". Vert eta]. 4 2 observed that PHAs with the same name often exhibit significantly different properties. This problem has been emphasized by Suganuma and Alexander 43, who have remarked: "The biocompatibility of poly(lactic acid) and poly(glycolic acid) with bone is... controversial. One possible reason for the discrepancy of results is the difference in models. Moreover, implantation site as well as the structure and molecular weight.., influence results". There is a great bounty of opportunities for PHAs in bone regeneration. Unfortunately, promiscuous applications with poorly and incorrectly defined polymers have soiled an unblemished record derived from suture-based applications. Therefore, a focussed mission to define PHAs formatted as homo- and copolymers for bone regeneration must be undertaken. A number of suggestions are proposed. Pre-eminent is the appropriate nomenclature to describe polymer configurations. In addition, polydispersity (derived from the ratio of weight-average and number-average molecular weight) must be divulged: the smaller the ratio, the greater the assurance that polymer distribution will be similar, thereby ensuring a predictable physico-chemical profile. Chemical structure can be verified with proton nuclear magnetic resonance (1HNMR) by identifying sequence distribution of different repeat units of PLG copolymer. Validation of glass transition temperature (Tg) and melting temperature (Tm) by differential scanning calorimetry must be reported to ensure that the homopolymer and copolymer are what they are supposed to be. The design of PHA carriers for bone regeneration must integrate a number of features required by cells to attach, differentiate, and express the appropriate products leading to tissue regeneration. Important features of PHA design include surface texture, charge, and wettability. Most eukaryotic cells are electronega-
141 189
tive and respond favourably for attachment when the substratum is electropositive 44. In general, hydrophilicity and a low contact angle promote cell attachment. However, trivializing substratum surface angles and charge can be misleading. For example, polystyrene is electropositive, yet is a poor substratum for cell attachment. Moreover, cross-linked pHEMA, a hydrogel that is hydrophilic, does not favour cell attachment. Furthermore, surface energetics of PHAs continually change during degradation. A strategy proposed to enhance cell attachment has been to prewet hydrophobic PHA 45'46. This procedure purportedly will facilitate cell anchorage. However, the extracellular milieu fluids may be capable of both wetting and providing an endogenous protein coating (e.g. fibronectin) to promote attachment. Dimensions of the carrier delivering a cargo (e.g. BMPs) to bony wound sites will be an important determinant of biodegradation rate. PLG suture will biodegrade more rapidly and differently than PLG engineered as a fixation device (described below). Integrated with dimension is internal architecture, an example is porosity. This property will be mentioned in more detail in the next section. Briefly, porosity may be viewed in terms of void volume, defined by void size and distribution. Carrier interactions with tissue fluids and cells are maximized by increasing surface area (voids/PHA structure), thus hastening biodegradation. However, PHAs are 'bulk eroders' 47-49 and polymer scientists have not yet mastered a technique that will produce PHAs that synchronously 'erode' with bone regeneration. There are several studies that describe PHA biodegradation and a general biodegradation profile may be derived. Unfortunately, poorly characterized PHAs were described (reviewed in Hollinger and Battistoneaa). Tightly defined temporal biodegradation profiles for PHAs are not available. Unpredictable biodegradation is not conducive to the ordered cellular and biochemical sequence of bone repair: the bone wound continuum. For example, after an extracellular matrix substratum and angiogenic environment are established by resident cells, osteoblasts can elaborate a matrix at a rate of approximately 0.7-0.8pm day -1. Designing a bulk eroder to biodegrade in register with these dynamic, interactive events will not be trivial. Moreover, regional differences, such as vascularity and biofunctionality, are additional challenges to polymer technology's capacity to craft bulk PHA carriers for bone repair. An aspect of device biodegradation involves shedding of fragments. It has been reported that pieces of PL and PG homopolymers and their copolymers < 2 pm in size are phagocytosed by macrophages 5~ and internalization of nanoparticles by human leucocytes has been reported 51. Moreover, slowly biodegrading crystalline PL nanoparticles have been observed over a year following implantation 52'53. Also, a foreign-body giant cell response to racemic D,L-PLG microspheres has been observed 54. An abundance of macrophages and foreign-body giant cells and their sustained presence can mitigate against tissue repair and regeneration. A number of disturbing reports have emerged concerning adverse tissue reactions from bulk PHA Biomaterials 1996, Vol. 17 No. 2
142
The B i o m a t e r i a l s Silver J u b i l e e C o m p e n d i u m
190
devices configured for bone fixation 55-5a. At the implantation site, swelling, redness, pain, and aseptic sinus tract formation have been unexpected sequelae observed 2-4 years after treatment, in vitro data indicate that degradation of bulk devices (i.e. fixation plates and screws) proceeds more rapidly in the centre than at the surface 47-4"*~. An outer layer or shell of slowly degrading polymer develops and entraps degrading macromolecules within the centre. Some oligomeric species diffuse through the shell; however, entrapped carboxylic moieties accumulate within the centre, accelerating internal decay of the bulk device until the central mass breaks the shell and dissects through soft tissue, manifesting as an aseptic sinus 5"~. Clinical reports of PL or combinations of PLG fixation devices for radius, maleolar, and zygoma fractures noted sinus formation, and histologically, foreign-body giant cells 58' 5.~. A recent study in rats involved implantation of poly(L-lactide) plates 6~ Over the 2 year study, tumours developed within the fibrous tissue capsule surrounding the plates. However, the control devices of medical-grade polyethylene were equally as tumorigenic 6~ No other reports have implicated PHAs to be tumorigenic. Humans are far less susceptible than rodents to tumour induction from foreign-body implantation (the solid-state, Oppenheimer effect) 61'62. Furthermore, a 2 year chronicity study in the rat represents a substantial spectrum of their life span. A biodegradable device would not be designed to endure for a commensurate period in humans. Moreover, in a study virtually identical to that of Nakamura et a]. 6~ Bos et al. 6:~ observed neither tumorigenicity nor PL remnants (after 33 months). Interestingly, Campbell et al. ~4 reported inhibition of carcinoma cell growth in vitro from PL. The authors speculated that metabolites of the degrading polymer may have affected cell response. Recently, Devereux e t a ] . 65 reported that implantation of PG mesh as an intestinal sling had a less than expected prevalence of pelvic infections. They believe the effect may be due to functional activation of leucocytes, priming them for response. Despite a compelling record of efficacy and safety as sutures, one wonders whether the marvellous properties of PHA sutures can be reproduced in PHAs engineered for bone regeneration applications, for example, fracture fixation and BMP carriers. Polymeric fixation devices are high molecular weight and dense, properties that retard biodegradation and may contribute to adverse tissue sequelae associated with the two-phased biodegradation profile of bulk PHAs 47-49. In contrast, a carrier for BMP does not need to meet the same biofunctional demands of fixation devices; therefore, a low-molecular-weight, porous carrier should be devised. The astonishing versatility of PHAs may allow for clever modifications fulfilling demanding criteria needed from a carrier for bone repair. However, from a practical, clinical perspective, chemical modification of PHAs will probably no longer confer 'regulatory favour', thus initiating a laborious, time-consuming, and expensive journey through the myriad of testing protocols required by the FDA. Biomaterials 1996, Vol. 17 No. 2
Poly(~-hydroxy acids): J.O. Hollinger and K. Leong
PRECLINICAL STUDIES Studies have noted that the combination of monolithic discs of PHAs with either partially purified allogeneic or xenogeneic BMP did not regenerate deficient osseous contour in either craniotomy defects in rhesus monkeys 66 or ostectomized extremity wounds in dogs 67, respectively. However, recombinant human (rh) BMP-2 and allogeneic bone collagen together have regenerated form and function to ablative mandibular wounds in dogs 68, femoral ostectomies in rats 69, and craniotomies in rats 7~ There are several possible explanations why monolithic PHA discs and native, partially purified BMP were not as effective as allogeneic bone collagen plus rhBMP-2 for promoting osseous regeneration. Monolithic devices may impede osteoconduction if they are engineered with inappropriate internal architecture 71'72. In contrast, a collagen carrier system may serve several roles, including (i) facilitating cellBMP interaction leading to osteob|ast expression 73, and (ii) providing for coordinated bone growth. Spatial orientation of signalling molecules for cell receptor interaction by a carrier will be discussed later. At this time, the concept of internal architecture focussing on porosity, merits attention. Porosity embodies several elements: pore size, range, volume, and distribution. Collectively, these elements contribute to the macromolecular internal space available for bone ingrowth (i.e. osteoconduction). The relationship between one component of porosity, pore size range, and osteoconduction (bone ingrowth) has been investigated for calcium phosphates and polyethylene materials 74-81. For these materials, a pore size range of 200-400#m appeared optimal for osteoconduction. However, polyethylene and calcium phosphates behave differently in tissue than PHAs. Moreover, in contrast to these materials, the more hydrophilic PHAs (e.g. copolymers of PLG), imbibe water following implantation, swell, and bulk erode through non-specific hydrolysis82. The physicochemical properties of PHAs, polyethylene, and ceramics are different. Therefore, it may not be valid to design PHA porosity for osteoconduction according to data from these materials. Currently, one study has explored porosity of a PHA for osteoconduction 8a. According to Robinson et a]. 83, calvarial wounds implanted with raceInic D,L-PL devices with 300-350#m sized pores and 60% void volume promoted osteoconduction, whereas smaller sized pores and void volumes did not. Pore sizes for osteoconduction with this PHA may be equivalent to certain ceramics and polyethylene. However, a successful carrier must do more than support bone ingrowth. Biodegradatio:n must be coordinated with bone formation and temporally calibrated with antecedent and subsequent events, otherwise bone repair will be delayed, inhibited, or corrupted. The PHA carrier should not begin to biodegrade until after initial deployment and spatial orientation of vulnerary factors, i.e. BMPs. Pharmacokinetics of BMP-release must be in concert with the local needs of the bone wound continuum. Pending hurdles for PHA
The Biomaterials Silver Jubilee C o m p e n d i u m
143
Poly(~-hydroxy acids)" J.O. Hollinger and K. Leong
carrier design, questions as
therefore,
include
answering
191
such
9 What are the local needs for BMP during bone regeneration? 9 What dose or doses are needed? 9 At what time period or periods during the bone regeneration cascade is BMP required? 9 Should BMP delivery be pulsatile or sustained? Answering these questions is occupying the intellectual energies of many investigators. Recombinant technology has led to the cloning and expressing of human BMPs in quantities suitable for therapeutic applications in preclinical animal studies 26'27'84-86. Preclinical studies with rhBMPs have included craniotomy and long bone repair in r a t s 69' 7~ a7' 88 , mandibular midbody continuity regeneration in dogs 89, alveolar cleft repair in dogs 9~ long bone reconstruction in sheep 91, rabbits 92, and rhesus non-human primates 93. In most of these studies, allogeneic type I collagen was used as a carrier for rhBMP-2. Allogeneic collagen is not ideal due to its immunological potential 94-96. Consequently, an experiment was executed with PLG (1"1 molar ratio of lactide:g|ycolide), prepared as porous particles (void volume approximately 60%), polydispersivity ~<2, and particle size range 75-250/~m 88. At the time of surgery, PLG was mixed with a known amount of allogeneic blood and rhBMP-2 and was allowed to clot in a syringe. The composition was relatively easy to manage for insertion into critical-sized intraosseous defects. However, the consistency was not sufficient to retain the mass" oozing fluid and soft tissue manipulation at the wound bed frequently caused displacement. A more clinically useful alternative is needed. Properties of a carrier system for rhBMP-2 will comprise a robust portfolio. Moreover, anatomical regional peculiarities will mandate unique modifications of that carrier. There are several properties inherent in a successful carrier system that are expressed by a familiar polymer: type I collagen. Based on preclinical animal studies, it appears to be the pre-eminent delivery system for rhBMP 6~' 70.87-89.91.92. Moreover, irrespective of the anatomical locale where it is placed, collagen is an effective carrier for rhBMP. Collagen offers cells a permissive substratum for attachment 97'98 and is the major component and fibrous backbone of the extracellular matrix (ECM). The ECM plays an instructive role influencing the biology of cells through cell-surface receptors (i.e. the integrins 9") that bind with growth-promoting factors, such as BMPs, thus positioning them for cell interaction 7'q7'98'1~176 The ECM can be the linchpin for engineering designs of PHA carriers. Mindful of the virtues of the ECM, research is under way exploring PHA modifications, including surface charge, void volume, surface texture, biodegradation rates, hydrophobicity and hydrophilicity, surface chemistry, crystallinity, and release kinetics. These features have been discussed in more detail and will be presented in an upcoming publication 101. Spatial orientation of rhBMP-2 and temporally
sequencing its availability with the dynamics of bone formation are formidable challenges that may not be satisfied by PG, PL, or PLGs. These PHAs lack pendant groups that can be derivatized to allow coupling with bioactive molecules such as BMPs or cell-specific ligands. A strategy to overcome this liability is chemical modification of the PHAs, but this jeopardizes 'regulatory favour'. A simple option may be to eliminate the carrier and to sprinkle rhBMP-2, as a powder, directly into bony defects. This concept has appeal, but there are obvious limitations: soft tissue prolapse can displace the powder and bleeding from the surgical site can dilute the delivered dose. Handling properties of rhBMP-2 at surgery will be difficult without a carrier. Unmodified PHAs may not be ideal carriers for BMPs; consequently, there are remarkable long-term opportunities for clever polymer chemists and chemical engineers working with bone physiolagists to develop modified PHA carriers that embody the virtues of collagen without its potential as an immunogen. These modifications will usher in a new and exciting phase of PHA applications that must fulfill efficacy and safety statutes of governing agencies. However, for short-term considerations, postsynthesis strategies may be used to yield highly porous PHA foams to carry and deliver exogenous BMP and 'to capture' the wound bed haematoma (within the voids of the foam). This composition may offer an immediate option for BMP delivery. This review highlighted some issues relevant to custom-tailoring PHAs. The temporal sequence of bone repair was mentioned and BMPs were described as marquis molecules for its regeneration. Several features of PHAs were presented, including variables affecting biodegradation. A number of preclinical studies were noted where bone regeneration has been established with combinations of rhBMP and either allogeneic collagen or PHA. Additional preclinical studies must be accomplished to determine regional dose-response requirements, biodegradation profiles, and release kinetics. The landscape for investigation is speckled with pitfalls, but is lush with rewards. REFERENCES
1 2 3 4 5 6 7
McKibbin B. The biology of fracture healing in long bones. ] Bone Joint Surg 1978; 60B: 150-162. Brighton C. Principles of fracture healing. Part I. The biology of fracture repair. Instructional Course 1984; 32: 60. Reddi A. Extracellular matrix and development. In: Piez KA, Reddi AH, eds. Extracellular Matrix Biochemistry. New York: Elsevier, 1984: 375-412. Simmons DJ. Fracture healing perspectives. Clin Orthop Rel Res 1985; 200: 100-113. Brightan CT, Hunt RM. Early histological and ultrastructnral changes in medullary fracture callus. J Bone Joint Surg 1991; 73A: 832-847. Bolander ME. Regulation of fracture repair by growth factors. Proc Soc Exper Biol Med 1992; 200: 165-170. Vukicevic S, Paralkar VM, Reddi AH. Extracellular matrix and bone morphogenetic proteins in cartilage and bone development and repair. Adv MoI Cell BioI 1993; 6: 207-224. Biomaterials 1996, Vol. 17 No. 2
144
The B i o m a t e r i a l s Silver J u b i l e e C o m p e n d i u m
192 8 9 10 11 12 13
14 15
16 17
18 19 20 21 22 23 24 25 26 27 28 29 30
31 32
Poly(~z-hydroxy acids): J.O. Hollinger and K. Leong Lane J, Sandhu H. Current approaches to experimental bone grafting. Orthop Clin North Am 1987; 18: 213-225. Horowitz MC, Friedlaender GE. Induction of specific T-cell responsiveness to allogenic bone. J Bone Joint Surg 1991; 73A: 1157-1168. Gregory CF. The current status of bone and joint transplants. Clio Orthop Rel Res 1972; 87: 156-166. Enneking WF, Mindell ER. Observations on massive retrieved h u m a n allografts. J Bone Joint Surg 1991; 73A: 1123-1142. Goldberg VM, Stevenson S. Natural history of autografts and allografts. Clio Orthop Rel Res 1987; 225: 7-16. Buck BE, Malinin TI. Bone transplantation and human immunodeficiency virus. An estimate of risk of acquired immunodeficiency syndrome (AIDS). Clio Orthop Rel Res 1989; 240: 129-136. Buck BE, Resnick L, Shah SM. Human immunodeficiency virus cultured from human bone. Clio Orthop Rel Res 1990; 251: 249-252. Asselmeier MA, Caspari RB, Bottenfield S. A review of allograft processing and sterilization techniques and their role in transmission of the human immunodeficiency virus. Am J Sports Med 1993; 21: 170-175. Hollinger JO. Factors for osseous repair and delivery: Part 1. J Craniofac Surg 1993; 4: 115-121. Madri JA, Basson MD. Extracellular matrix-cell interactions: dynamic modulators of cell, tissues, and organisms structure and function. Lab Invest 1992; 66: 519-521. Raghow R. The role of extracellular matrix in postinflammatory wound healing and fibrosis. FASEB J 1994; 8: 823-831. Reddi AH, Anderson WA. Collagenous bone matrixinduced endochondral ossification and hemopoiesis. J Cell Biol 1976; 69: 557-572. Triffitt JT. Initiation and enhancement of bone formation. Acta Orthop Scand 1987; 58: 673-684. Bernick S, Paule W, Ertl D, Nishimoto SK, Nimni ME. Cellular events associated with the induction of bone by demineralized bone. J Orthop Res 1989; 7: 1-11. Sawyer JR, Myers MA, Rosier RN, Puzas JE. Heterotopic ossification: clinical and cellular aspects. Calcif Tissue Int 1991; 49: 208-215. Urist MR. Bone: formation by autoinduction. Science 1965; 150: 893-899. Urist MR, Silverman MF, Buring K, Dubuc FL, Rosenburg JM. The bone induction principle. Clio Orthop Rel Res 1967; 53: 243. Urist MR, Strates BS. Bone morphogenetic protein. J Dent Res 1971; 50: 1392-1406. Wozney JM. Bone morphogenetic proteins. Prog Growth Factors 1989; 1: 267-280. Wozney JM. The bone morphogenetic protein family and osteogenesis. Mol Reprod Dev 1992; 32: 160-167. Kingsley D. What do BMPs do in mammals? Clues from the mouse short-ear mutation. TIG "1994; 10: 16-21. Kingsley DM. The TGF-beta superfamily: new members, new receptors, and new genetic tests of function in different organisms. Genes Dev 1994; 8: 133-146. Hauschka PV, Mavrakos AE, Iafrati MD, Doleman SE, Klagsbrun M. Growth factors in bone matrix. Isolation of multiple types with affinity chromatography and heparin sepharose. J Biol Chem 1986; 261: 1266512674. Hauschka PV. Growth factor effects in bone. In: Hall BK, ed. The Osteoblast and Osteocyte. West Caldwell, NJ: The Telfor Press, Inc., 1990: 103-170. Chu CC. The degradation and biocompatibility of suture materials. In: Williams DF, ed. CRC Critical Reviews in
Biomaterials 1996, Vol. 17 No. 2
33 34
35 36 37 38 39
40
41 42 43 44 45
46 47
48
49
50
51
52
Biocompatibility. Boca Raton, FL: CRC Press, 1985: 261-322. Hollinger JO, Battistone GC. Biodegradable bone repair materials. Synthetic polymers and ceramics. Clio Orthop Rel Res 1986; 207: 290-305. Ron E, Langer R. Erodible systems. In: Kydonieus A, ed. Treatise on Controlled Drug Delivery. Fundamentals, Optimization, Applications. New York: Marcel Dekker, 1992:199-224. Nieuwenhuis J. Synthesis of polylactides, polyglycolides and their copolymers. Clio Mater 1992; 10: 59-67. Morrison RT, Boyd RN. Organic Chemistry. Boston: Allyn and Bacon, 1980. Eliel EL. Stereochemistry of Carbon Compounds. New York: McGraw-Hill, 1962: 1-128. Brandt BB, Waters GM, Rispler MJ, Kline ES. D- and LLactate catabolism to carbon dioxide in rat tissues. Proc Soc Exp Biol Med 1984; 175: 328-335. Holland SJ, Tighe BJ, Gould PL. Polymers for biodegradable medical devices. 1. The potential of polyesters as controlled macromolecular release systems. J Contr Rel 1986; 4: 155-180. Pistner H, Bemdox DR, Muhling J, Reuther JF. Poly(Llactide): a long-term degradation study in vivo. Part 3. Analytical characterization. Biomaterials 1993; 14: 291-304. Vert M, Li S, Garreau H. New insights on the degradation of bioresorbable polymeric devices based on lactic and glycolic acids. Clio Mater 1992; 10: 3-8. Vert M, Chabot F, Leray J, Christel P. Biodegradable polyesters for bone surgery. Makromol Chem Suppl 1981; 5: 30-36. Suganuma J, Alexander H. Biological response of intramedullary bone to poly-L-lactic acid. J Appl Biomater 1993; 4: 13-27. Reddi AH. Bone matrix in the solid state: Geometric influence on differentiation of fibroblasts. Adv Biol Med Phys 1974; 15: 1. Mikos AG, Bao Y, Cima LG, Ingber DE, Vacanti JP, Longer R. Preparation of poly(glycolic acid) bonded fiber structures for cell attachment and transplantation. J Biomed Mater Res 1993; 27: 183-189. Mikos AG, Lyman MD, Freed LE, Langer R. Wetting of poly(t.-lactic acid) and poly(Dr.-lactic-co-glycolic acid) foams for tissue culture. Biomaterials 1994: 15: 55-58. Li S, Garreua H, Vert M. Structure-property relationships in the case of the degradation of massive aliphatic poly(x-hydroxy-acids) in aqueous media. Part 2: Degradation of lactide-glycolide PLA37.5;GA25 and PLA75:GA25 Poly(DL-lactic acid). J Mater Sci Mater Med 1990; 1: 131-139. Li S, Garreua H, Vert M. Structure-property relationships in the case of the degradation of massive aliphatic poly(:~-hydroxy-acids) in aqueous media. Part 1: Poly(Dl.-lactic acid). J Mater Sci Mater Med 1990; 1: 123-130. Li S, Garreua H, Vert M. Structure-property relationships in the case of the degradation of massive aliphatit: poly(~-hydroxy-acids) in aqueous media. Part 3: Influence of the morphology of poly(L-lactic acid). J Mater Sci Mater Med 1990; 1: 198-206. Tabata Y, Ikado Y. Macrophage phagocytosis of biodegradable microspheres composed of I.-lactic acid/ glycolic acid homo- and copolymers. J Biomed Mater Res 1988; 22: 837-858. Lam KH, Schakenraad JM, Esselbrugge H. Feijen J, Nieuwenhuis P. The effect of phagocytosis of poly(l,lactic acid) fragments on cellular morphology and viability. J Biomed Mater Res 1993; 27: 1569-1577. Verheyen CCPM, De Wijn JR, Van Blitterwijk CA,
The Biomaterials Silver Jubilee Compendium
145 193
Poly(~-hydroxy acids): J.O. Hollinger and K. Leong
53
54 55
56 57
58
59
60
61 62 63
64 65
66 67
68
69
70
Rozing PM, De Groot K. Examination of efferent lymph nodes after 2 years of transcortical implantation of poly(L-lactide) containing plugs: a case report. J Biomed Mater Res 1993; 27: 1115-1118. Leroux JC, Gravel P, Balant L e t al. Internalization of poly(D,L-lactic acid) nanoparticles by isolated human leukocytes and analysis of plasma proteins adsorbed onto the particles. J Biomed Mater Res 1994; 28: 471481. Yamaguchi K, Anderson J. In vivo biocompatibility studies of medisorb 65/35 D,L-lactide/glycolide copolymer microspheres. J Contr Rel 1993; 24: 81-93. Bostman O, Hirvensalo E, Makinen J, Rokkanen P. Foreign-body reactions to fracture fixation implants of biodegradable synthetic polymers. J Bone Joint Surg 1990; 72B: 592-596. Bostman O. Absorbable implants for fixation of fractures. J Bone Joint Surg 1991; 73A: 148-153. Bostman O, Paivarinta U, Patio E et al. Absorbable polyglycolide screws in internal fixation of femoral osteotomies in rabbits. Acta Orthop Scand 1991; 62: 587-591. Bergsma EJ, Rozema FR, Bos RRM, De Bruijn WC. Foreign body reactions to resorbable poly(L-lactide) bone plates and screws used for fixation of unstable zygomatic fractures. J Oral Maxillofac Surg 1993; 51: 666-670. Bostman O. Intense granulomatous inflammatory lesions associated with absorbable internal fixation devices made of polyglycolide in ankle fractures. Clin Orthop Rel Res 1992; 778: 178-199. Nakamura T, Shimizo Y, Okumura N, Matsui T, Hyon S, Shimamoto T. Tumorigenicity of poly-L-lactide (PLLA) plates compared with medical-grade polyethylene. J Biomed Mater Res 1994; 28: 17-25. Schoen F. Biomaterials-associated infection, neoplasia, and calcification. Trans Am Soc Artif Intern Organs 1987; XXXIII: 8-23. Schoen FJ. Biomaterial-associated tumorigenesis. In: Rather B, ed. Biomaterials Science: An Introductory Text. Orlando, FL: Academic Press, 1990. Bos R, Boering G, Rozema F et al. Degradation of and tissue reaction to biodegradable poly(L-lactide) for use as internal fixation for fractures: a study in rats. Biomaterials 1991; 12: 32-36. Campbell JP, Edsberg L, Meyer AE. Polylactide inhibition of carcinoma cell growth in vitro. ] Oral Maxillofac Surg 1994; 52:49-51. Devereux DF, O'Connell SM, Liesch JB, Weinstein M, Robertson FM. Induction of leukocyte activation by meshes surgically implanted in the peritoneal cavity. Am J Surg 1991; 162: 243-246. Hollinger JO, Schmitz JP, Mark DE, Seyfer AE. Osseous wound healing with xenogeneic bone implants with a biodegradable carrier. Surgery 1990; 107: 50-54. Heckman JD, Boyan BD, Aufdemorte TB, Abbott JT. The use of bone morphogenetic protein in the treatment of non-union in a canine model. J Bone Joint Surg 1991; 73A: 751-763. Toriumi DM, Larrabee WFJ, Walike JW, Millay DJ, Eisele DW. Demineralized bone. Implant resorption with long-term follow-up. Arch Otolaryngol Head Neck Surg 1990; 116: 676-680. Yasko AW, Lane JM, Fellinger EJ, Rosen V, Wozney JM, Wang EA. The healing of segmental bone defects induced by recombinant human bone morphogenetic protein (rhBMP-2). ] Bone Joint Surg 1992; 74A: 659671. Marden LJ, Hollinger JO, Chaudhari A, Turek T, Schaub R, Ron E. Recombinant bone morphogenetic protein-2 is
71 72 73
74
75
76
77 78
79
80 81 82
83 84 85 86
87 88
89
superior to demineralized bone matrix in repairing craniotomies defects in rat. J Biomed Mater Res 1994; 28: 1127-1138. Kleinschmidt J, Hollinger JO, Quigley N, Marden LJ. A multiphase system implant for regenerating calvaria. ] Plast Reconstr Surg 1993; 91: 581-588. Hollinger JO. Factors for osseous repair and delivery: Part 11. J Craniofac Surg 1993; 4: 135-141. Reddi AH, Sampath TK, Hand AR. Biosynthesis of extracellular matrix in bone: approaches and prospects. In: Veis A, ed. The Chemistry and Biology of Mineralized Connective Tissue. NY: Elsevier North Holland, 1981: 417-426. Hulbert S, Young F, Mathews R, Klawitter J, Talbert C, Stelling F. Potential of ceramics materials as permanently implantable skeletal prosthesis. J Biomed Mater Res 1970; 4: 433-442. Klawitter J, Hulbert S. Application of porous ceramics for the attachment of load bearing internal orthopedic applications. ] Biomed Res Syrup No. 2 (Part 1) 1971; 161-229. Klawitter J, Bagwell J, Weinstein A, Sauer B, Pruitt J. An evaluation of bone ingrowth into porous high density polyethylene. ] Biomed Mater Res 1976; 10: 311-323. Spector M, Flemming WR, Kreutner A. Bone ingrowth into porous high density polyethylene. ] Biomed Mater Res 1976; 7: 595-603. Spector M, Harmon S, Kreutner A. Characteristics of tissue ingrowth into porous proplast and porous polyethylene implants in bone. J Biomed Mater Res 1979; 13: 677-692. Hollinger JO, Schmitz JP, Mizgala JW, Hassler C. An evaluation of two configurations of tricalcium phosphate for treating craniotomies. J Biomed Mater Res 1989; 23: 17-29. Holmes RE. Bone regeneration within a coralline hydroxyapatite implant. J Plast Reconstr Surg 1979; 63: 626-633. Holmes RE, Bucholz RW, Mooney V. Porous hydroxyapatite as a bone-graft substitute in metaphyseal defects. J Bone Joint Surg 1986; 68A: 904-911. Lewis D. Controlled release of bioactive agents from lactide/glycolide polymers. In: Chasin M, Langer R, ed. Biodegradable PoIymers as Drug Delivery Systems. New York: Marcel Dekker, 1990: 1-42. Robinson B, Hollinger JO, Szachowicz E, Brekke j. Calvarial bone repair with porous D,L-polylactide. Otolaryngol Head Neck Surg 1995; 112: 707-713. Wozney JM, Rosen V, Celeste AJ et al. Novel regulators of bone formation: molecular clones and activities. Science 1988; 242: 1528-1534. C}zkaynak E, Rueger DC, Drier EA et al. OP-1 cDNA encodes an osteogenic protein in the TGF-fl family. J EMBO 1990; 9: 2085-2093. Ozkaynak E, Schnegelsberg PN, Jin DF et al. Osteogenic protein-2. A new member of the transforming growth factor-beta superfamily expressed in early embryogensis. J Biol Chem 1992; 267:25 220-25 227. Kenley R, Yim K, Abrams J et al. Biotechnology and bone graft substitutes. Pharmacol Res 1993; 10: 13931401. Kenley R, Marden L, Turek T, Jin L, Ron E, Hollinger J. Osseous regeneration in the rat calvarium using novel delivery systems for recombinant bone morphogenetic protein 2 (rhBMP-2). ] Biomed Mater Res 1994; 28: 1139-1147. Toriumi DM, Kotler HS. Luxunberg DP, Holtrop ME, Wang EA. Mandibular reconstruction with a recombinant bone-inducing factor. Functional, histologic, and Biomaterials 1996, Vol. 17 No. 2
The Biomaterials Silver Jubilee Compendium
146
Poly(o~-hydroxy acids): J.O. Hollinger and K. Leong
194
90 91
92
93
94 95
biomechanical evaluation. Arch Otolaryngol Head Neck Surg 1991; 117: 1101-1112. Mayer MM, Hollinger JO. Repair of alveolar clefts in dogs with recombinant bone morphogenetic protein and poly(~-hydroxy acid). J Plast Reconstr Surg (in press). Gerhart TNK Kirker-Head CA, Kriz MJ et aI. Healing segmental femoral defects in sheep using recombinant human bone morphogenetic protein (rhBMP-2). Clin Orthop Rel Res 1993; 293: 317-326. Cook SD, Baffes GC, Wolfe MW, Sampath K, Rueger DC, Whitecloud TS. The effect of human recombinant osteogenic protein-1 on healing of large segmental bone defects. J Bone Joint Surg 1994; 76A: 827-838. Hollinger JO, Zegzula D, Buck D, Smith J, Wozney J. Reconstruction and regeneration of critical-sized segmental defects in radii of rhesus nonhuman primates. In preparation. Ellingsworth LR, DeLustro F, Brennan JE, Sawamura S, McPherson J. The human immune response to reconstituted bovine collagen. J lmmunol 1986; 136: 877-882. DeLustro F, Dasch J, Keefe J, Ellingsworth L. Immune responses to allogeneic and xenogeneic implants of collagen and collagen derivatives. Clin Orthop Rel Res 1990; 260: 263-279.
Biomaterials 1996, Vol. 17 No. 2
96
97
98 99 100
101
van Luyn MJA, van Wachem PB, Olde Damink LHH, Dikstra PJ, Feijen J, Nieuwenhuis P. Secondary cytotoxicity of cross-linked dermal sheep collagens during repeated exposure to human fibroblasts. Biomaterials 1992; 13: 1017-1024. Sampath TK, Reddi AH. Dissociative extraction and reconstitution of extracellular matrix components involved in local bone differentiation. Proc Nat] Acad Sci USA 1981; 78: 7599-7603. Reddi AH. Regulation of cartilage and bone differentiation by bone morphogenetic proteins. Curt Opin Cell Bio] 1992; 4: 850-855. Lin CQ, Bissel MJ, Multi-faceted regulation of cell differentiation by extracellular matrix. FASEB J 1993; 7: 737-743. Paralkar VM, Nandedkar AKN, Pointer RH, Kleinman HK, Reddi AH. Interaction of osteogenin, a heparin binding bone morphogenelic protein, with type IV collagen. J Biol Chem 1990; 265: 17 281-17 284. Hollinger JO, Jamiolkowski D, Shalaby SW. Poly(~hydroxy acids) and bone repair. In: Hollinger JO, ed. Biomedical Applications of Synthetic Biodegradable Polymers. Boca Raton, FL: CRC Press, 1995: 197-222.
147
The Biomaterials Silver Jubilee Compendium
Biomaterials Biomaterials 19 (1998) 2219-2232
ELSEVIER
Response of MG63 osteoblast-like cells to titanium and titanium alloy is dependent on surface roughness and composition J. Lincks a'b B D. B o y a n b'c'd'* C.R. Blanchard e C.H. L o h m a n n c Y. Liu c, D.L. C o c h r a n b, D . D . D e a n c, Z. Schwartz b'~'f 9
~
~
a Wilford Hall Medical Center, Lackland AFB b Department of Periodontics, University of Texas Health Science Centre, San Antonio, TX, USA c Department of Orthopaedics, University of Texas Health Science Centre, San Antonio, TX, USA dDepartment of Biochemistry, University of Texas Health Science Centre, San Antonio, TX, USA eSouthwest Research Institute, San Antonio, Texas, USA f Department of Periodontics, Hebrew University Hadassah Faculty of Dental Medicine, Jerusalem, Israel
Abstract The success of an implant is determined by its integration into the tissue surrounding the biomaterial. Surface roughness and composition are considered to influence the properties of adherent cells. The aim of this study was to determine the effect of chemical composition and surface roughness of commercially pure titanium (Ti) and Ti-6A1-4V alloy (Ti-A) on MG63 osteoblast-like cells. Unalloyed and alloyed Ti disks were machined and either fine-polished or wet-ground, resulting in smooth (S) and rough (R) finishes, respectively. Standard tissue culture plastic was used as a control. Surface topography and profile were evaluated by cold field emission scanning electron microscopy and profilometry, while chemical composition was determined using Auger electron spectroscopy and Fourier transform infrared spectroscopy. The effect on the cells was evaluated 24 h postconfluence by measuring cell number, [3H]-thymidine incorporation into DNA, cell and cell layer alkaline phosphatase specific activity (ALPase), osteocalcin and collagen production, [-35S]-sulfate incorporation into proteoglycan, and prostaglandin E 2 (PGE2) and transforming growth factor-/~ (TGF-/~) production. When compared to plastic, the number of cells was reduced on the pure Ti surfaces, while it was equivalent on the Ti-A surfaces; [3H]-thymidine incorporation was reduced on all surfaces. The stimulatory effect of surface roughness on ALPase in isolated cells and the cell layer was more pronounced on the rougher surfaces, with enzyme activity on Ti-R being greater than on Ti-A-R. Osteocalcin production was increased only on the Ti-R surface. Collagen production was decreased on Ti surfaces except Ti-R; [35S]-sulfate incorporation was reduced on all surfaces. Surface roughness affected local factor production (TGF-/~, PGE2). The stimulatory effect of the rougher surfaces on PGE2 and TGF-/~ was greater on Ti than Ti-A. In summary, cell proliferation, differentiation, protein synthesis and local factor production were affected by surface roughness and composition. Enhanced differentiation of cells grown on rough vs. smooth surfaces for both Ti and Ti-A surfaces was indicated by decreased proliferation and increased ALPase and osteocalcin production. Local factor production was also enhanced on rough surfaces, supporting the contention that these cells are more differentiated. Surface composition also played a role in cell differentiation, since cells cultured on Ti-R surfaces produced more ALPase than those cultured on Ti-A-R. While it is still unknown which material properties induce which cellular responses, this study suggests that surface roughness and composition may play a major role and that the best design for an orthopaedic implant is a pure titanium surface with a rough microtopography. 9 1998 Published by Elsevier Science Ltd. All rights reserved Keywords: Osteoblasts; Titanium; Titanium alloy; Surface roughness; PGE2; TGF-/~; In vitro
1. Introduction The m o r p h o l o g y of an implant surface, including m i c r o t o p o g r a p h y and roughness, has been shown to be *Corresponding author. Tel.: (210) 567-6326; fax: (210) 567-6295; internet:
[email protected]
related to successful bone fixation [1, 2]. In addition, the m a n u f a c t u r i n g process used to achieve the surface texture, either chemical [3] or mechanical [4], also influences clinical success. At present, titanium implants in clinical use vary with respect to surface roughness and composition, with consensus being limited to the fact that bone forms m o r e readily on a rough surface whereas
0142-9612/98/$--See front matter ~ 1998 Published by Elsevier Science Ltd. All rights reserved. PII SO 1 42-96 1 2(98)00 1 44-6
148 2220
The Biomaterials Silver Jubilee Compendium J. Lincks et al./Biomaterials 19 (1998) 2219-2232
fibrous connective tissue is found more frequently on a smooth surface [5]. In vitro studies have provided some insight into the response of specific cell types to surface properties. It is clear that surface roughness affects cell response. In particular, osteoblast-like cells exhibit roughness-dependent phenotypic characteristics. They tend to attach more readily to surfaces with a rougher microtopography [6, 7]. Moreover, they appear to be more differentiated on rougher surfaces with respect to morphology, extracellular matrix synthesis, alkaline phosphatase specific activity and osteocalcin production, and response to systemic hormones such as 1,25-(OH)zD3 [8, 9]. The degree of roughness also affects production of local factors such as transforming growth factor beta (TGF-fl) and prostaglandin E2 (PGE2) 1-10], both of which can act on the osteoblastic cells as autocrine regulators [-11, 12], and can modulate the activity of bone resorbing cells via paracrine mechanisms [13, 14]. The morphology of the surface also plays a role. A variety of cells can orient themselves in the grooves of micromachined surfaces [15-17]. Depending on the degree of roughness, these cells may actually see the groove as smooth. On a randomly rough surface as is created by grit blasting or chemical etching, cells may form different focal attachments which result in a phenotype that is distinct from that seen on the grooved surface with the same degree of roughness. Titanium implants which are currently in clinical use in dentistry and orthopaedics, vary with respect to surface roughness and composition. In dentistry, commercially pure titanium (Ti) has become one of the most commonly used implant materials whereas in orthopaedics Ti alloys have virtually replaced Ti because of strength requirements [18, 19]. Both Ti and Ti-6A1-4V (Ti alloy) develop a surface oxide layer due to the natural passivation of Ti [20,21]. However, differences in the crystallinity of the underlying metal as well as the segregation of alloy components, may cause the oxide that forms on Ti to be quite different from the oxide that forms on Ti alloys. Several studies have shown that even subtle differences in surface composition, including Ti oxide crystallinity, can modify cell response, even when surface roughness is held constant [-6, 22-28]. We previously showed that when MG63 osteoblastlike osteosarcoma cells are cultured on Ti discs with average surface roughness values (Ra) varying from <0.1 lum (smooth) to 3-4 gm (rough) to > 6 gm (very rough), there are distinct differences in phenotypic expression [8, 10]. For these studies, the smooth surfaces were obtained by electropolishing following chemical etching; rough surfaces were obtained by coarse grit blasting; and very rough surfaces were achieved via Ti plasma spray. The results showed that as surface roughness increased, expression of a differentiated osteoblastic phenotype increased, including reduced cell number and
DNA synthesis (proliferation), and increased alkaline phosphatase specific activity (ALPase), osteocalcin production, collagen synthesis, proteoglycan sulfation, and production of latent TGF-/~ and PGE2. The optimal surface appeared to be those with R a values around 4 lam; cell proliferation was reduced but not blocked and phenotypic differentiation was enhanced. In contrast, cells on the smooth surface had high proliferation rates but ALPase and osteocalcin production were low, indicative of a loss of a differentiated osteoblastic phenotype. To determine whether the composition of the surface or microtopography are more important variables in determining osteoblastic phenotype, we examined the response of MG63 cells to machined surfaces with smooth R a values as well as with rough R a values that were prepared from Ti and Ti alloy. The results of the present study using machined surfaces were compared to those of our previous work using grit-blasting to obtain similar R a values.
2. Materials and methods
2.1. Titanium disk preparation and characterization 2.1.1. Disk preparation Titanium disks (14.75 mm diameter; 0.8 mm thick) were fabricated from sheets of either commercially pure titanium (Ti: medical grade 2, ASTM F67, 'unalloyed Ti for medical applications') or titanium-6 wt% aluminum4 wt% vanadium alloy (Ti-6A1-4V; Ti-A) obtained from Timet, Inc. (O'Fallon, MO). Chemical composition was provided by the supplier and was not verified prior to surface preparation. Each sheet was sectioned into one foot by one foot plates for ease of handling and to ensure a consistent finish. The disks were either polished or ground to acquire the desired surface finishes. Polishing to create the smooth surface was performed by lapping with 18T grit (oil based 500-600 grit aluminum oxide) followed by polishing with 4.0 paper (1200 grit aluminum oxide) by French Grinding Service, Inc. (Houston, TX). The rough surface was prepared by wet sanding using a carborundum brand zirconium oxide/aluminum oxide resin bonded to a cloth belt by Metal Samples, Inc. (Mumford, AL). Disks were stamped using an automated metal punch and cleaned in an acetone bath using an ultrasonic cleaner for one hour. The disks were then washed in Jet-A fuel (grade AL-24487-F; Diamond Shamrock, San Antonio, TX) in an ultrasonic cleaner for one hour and was followed by four washes with Versa Clean (Fisher Scientific, Pittsburgh, PA). Between each wash with Versa Clean the disks were rinsed twice with deionized, distilled water. After the final wash, the disks were rinsed with 70% ethanol and then dried in vacuo. Prior to use each disk was washed again three times with ethanol and rinsed three times with deionized, distilled water. The
The Biomaterials Silver Jubilee Compendium J. Lincks et al./Biomaterials 19 (1998) 2219-2232
disks were individually wrapped in gauze to prevent damage and then sterilized by autoclaving.
2.1.2. Surface characterization Representative disks from each group were subjected to surface analysis. The surface microtopography of the disks was examined using an Amray 1645 cold field emission scanning electron microscope (Amray, Bedford, MA) with a nonthermally assisted tip and secondary and backscattered electron capability. Two samples from each group were examined at 100 to 500 x. Surface roughness was measured by profilometry using a Taylor-Hobson Surtronic 3 profilometer (Leicester, UK). Average surface roughness (Ra) measurements were taken at ten different locations on each one foot x one foot sheet to obtain an accurate assessment. For the smooth surfaces, measurements were made in all directions, whereas on the rough surfaces, measurements were taken perpendicular to the machine markings. Following the punching operation, four disks from each sheet were randomly sampled to confirm the R a values obtained earlier. Auger electron spectroscopy was performed to analyze the Ti oxide layer using a Perkin-Elmer Model 595 scanning Auger microprobe (Perkin Elmer, Physical Electronics Division, Eden Prairie, MN). Spectra were obtained from two representative disks from the two groups with a smooth surface (Ti and Ti-A) to determine the chemical profile of the subsurface layer. Rough disks were not examined to avoid artifacts associated with rough morphologies; further, the thickness and composition of the surface oxides on the rough and smooth disks for each material would be expected to be identical since all disks were machined and cleaned using the same protocol. The spectra were obtained at regular sputtering intervals at a sputtering rate of 400 A min- ~. Comparing spectra and relative peak heights at given surface depths provided information about the chemistry of the oxide layer. Fourier transform infrared spectroscopy (FTIR) was performed to determine if an organic residue remained on the disk surfaces after cleaning. Spectra were obtained from four disks (two from the smooth Ti group and two from the smooth Ti-A group) using a Nicolet Magna FTIR in reflection mode. Spectra were collected using 32 scan summations at a resolution of 16 cm- 1. FTIR spectroscopy was not performed on the rough surfaces, because artifactual measurements are obtained on rough samples.
149 2221
a relatively immature osteoblast, including the stimulation of alkaline phosphatase activity and osteocalcin synthesis and inhibition of proliferation in response to treatment with l~,25-(OH)zD3 [29,30]. As a result they are a good model for examining the early stages of osteoblast differentiation. However, the culture conditions under which MG63 cells will mineralize their matrix have not been defined, so terminal differentiation cannot be studied using these cells. Despite this limitation, we selected this model in preference to fetal rat calvarial cells since the latter are derived from embryonic rat bone which may differ significantly from adult human bone. We recognize that MG63 cells are not normal osteoblasts and data interpretation must take this into consideration. MG63 cells were obtained from the American Type Culture Collection (Rockville, MD). For all experiments, cells were cultured on disks placed in 24 well plates (Corning, Corning, IL). Controls consisted of cells cultured directly on the polystyrene surface of the 24 well plate. Cells were plated at 9300 cells cm-2 in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal bovine serum (FBS) and 0.5% antibiotics (diluted from a stock solution containing 5000 U m l - 1 penicillin, 5000 U ml-1 streptomycin; GIBCO, Grand Island, NY) and cultured at 37~ in an atmosphere of 100% humidity and 5% CO2. Media were changed every 48 h until the cells reached confluence. Because of the opacity of the Ti disks, there was no practical way to assess confluency of the cultures. As a result, when cells reached visual confluence on plastic, cultures on all other surfaces were treated exactly as those grown on plastic.
2.3. Cell morphology To determine whether cell morphology varied as a function of surface roughness, the cultures were examined by scanning electron microscopy. At harvest, the culture media were removed and the samples rinsed three times with phosphate-buffered saline (PBS) and fixed with 1% OsO4 in 0.1 M PBS for 15-30 min. After fixation, the disks were rinsed with PBS, sequentially incubated for 30-45rain each in 50, 75, 90 and 100% ter-butyl alcohol, and vacuum dried. A thin layer of goldpalladium was sputter-coated onto the samples prior to examination in a JEOL 6400 FEC cold field emission scanning microscope (JEOL USA, Inc. Peabody, MA).
2.4. Cell proliferation 2.2. Cell culture MG63 osteoblast-like cells were used for these experiments because they were obtained from a human osteosarcoma [29] and have been well-characterized. They display numerous osteoblastic traits that are typical of
2.4.1. Cell number At harvest, cells were released from the culture surface by addition of 0.25% trypsin in Hank's balanced salt solution (HBSS) containing 1 mM ethylenediamine tetraacetic acid (EDTA) for ten minutes at 37~ and this
150 2222
The Biomaterials Silver Jubilee Compendium J. Lincks et al./Biomaterials 19 (1998) 2219-2232
was followed by addition of DMEM containing 10% FBS to stop the reaction. Previous studies demonstrated that two trypsinizations are necessary to quantitatively harvest MG63 cells from rough Ti surfaces [8]. Accordingly, a second trypsinization was performed to ensure that any remaining cells had been removed from the surface. Cell suspensions from both trypsinizations were combined and centrifuged at 500 x 9 for 10 min. Cell pellets were washed with PBS and resuspended in PBS. Cell number was determined by use of a Coulter Counter (Coulter Electronics, Hialeah, FL). Cells harvested in this manner exhibit > 95 % viability based on trypan blue dye exclusion.
2.4.2. [~H]-thymidine incorporation DNA synthesis was estimated by measuring [3H]thymidine incorporation into trichloroacetic acid (TCA) insoluble cell precipitates as previously described by Schwartz et al. [31]. MG63 cells were cultured on the plastic surface or Ti disks until the cells on plastic reached visual confluence. Media were changed and the incubation continued for an additional 24 h. Four hours prior to harvest, 50 ~1 [3H]-thymidine (from a 1 ~Ci ml- 1 stock solution) was added to the cultures. At harvest, the cell layers were washed twice with cold PBS, twice with 5% TCA, and then treated with ice-cold saturated TCA for 30 min. TCA-precipitable material was dissolved in 0.25 ml 1% sodium dodecyl sulfate (SDS) at 20~ and radioactivity measured by liquid scintillation spectroscopy.
2.5. Cell differentiation 2.5.1. Alkaline phosphatase specific activity At harvest, either cell layers, as described below, or isolated cells, as described above, were prepared and their protein content determined by use of commercially available kits (Micro/Macro BCA, Pierce Chemical Co., Rockford, IL). Alkaline phosphatase [orthophosphoric monoester phosphohydrolase, alkaline; E.C. 3.1.3.1] activity was assayed as the release of p-nitrophenol from p-nitrophenylphosphate at pH 10.2 as previously described [32] and specific activity determined. Cell layers were prepared following the method of Hale et al. [33]. At harvest, culture media were decanted, cell layers washed twice with PBS, and then removed with a cell scraper. After centrifugation, the cell layer pellets were washed once more with PBS and resuspended by vortexing in 0.5 ml deionized water plus 25 l.tl 1% Triton-X-100. Pellets were further disrupted by freeze/thawing three times. Isolated cells were harvested as described above for the determination of cell number, except that after the cell pellets had been washed twice with PBS, the cells were resuspended by vortexing in 0.5 ml of deionized water with 25 t.tl of 1% Triton-X-100.
Enzyme assays were performed on both cell and cell layer lysates.
2.5.2. Osteocalcin production The production of osteocalcin by the cultures was measured using a commercially available radioimmunoassay kit (Human Osteocalcin RIA Kit, Biomedical Technologies, Stoughton, MA). Culture media were concentrated five-fold by lyophilization and reconstituted in 100 ~1 normal rabbit serum, 10 btl rabbit anti-human osteocalcin antibody, 100 ~1 [125I]-human osteocalcin, and 200 ~1 Tris-saline buffer and placed overnight on an orbital platform shaker (approximately 80 rpm) at room temperature. Goat anti-rabbit antibody and polyethylene glycol (100 l.tl each) were added to each tube the following morning. After vortexing, the samples were placed on an orbital shaker for 2 h at room temperature. One ml of Tris-saline buffer was added to each sample. The solution was then vortexed and centrifuged at 500 x 9 for 20 min at 4~ The supernatant was decanted and the pellet placed in scintillation cocktail and counted. Osteocalcin concentrations were determined by correlating the percentage bound over unbound counts to a standard curve.
2.6. Matrix production 2.6.1. Collagen production Matrix protein synthesis was assessed by measuring the incorporation of [3H]-proline into collagenase digestible (CDP) and noncollagenase digestible (NCP) protein [34]. When the cells reached confluence on plastic, the media in all cultures were replaced with 500 gl DMEM containing 10% FBS, antibiotics, and 50 ~gm1-1 /%amino proprionitrile (Sigma, St. Louis, MO), and 10 gCim1-1 of L[GaH]-proline (New England Nuclear, Boston, MA). After 24 h, media were discarded. Cell layers (cells and matrix) were obtained by scraping and resuspending in two 0.2 ml portions of 0.2 N NaOH. Proteins were precipitated with 0.1 ml 100% TCA containing 1% tannic acid, washed three times with 0.5 ml 10% TCA + 1% tannic acid, and then twice with icecold acetone. The final pellets from the cell layers were dissolved in 500 l.tl 0.05 N NaOH. Digestion of the cell layer pellet was performed using highly purified clostridial collagenase (Calbiochem, San Diego, CA; 138 U mg-1 protein) as described previously [8]. NCP synthesis was calculated after multiplying the labeled proline in NCP by 5.4 to correct for its relative abundance in collagen [34]. Percent collagen production was calculated by comparing CDP production with total CDP + NCP production (i.e.: [CDP/(CDP + NCP)] x 100). The protein content of each fraction was determined by miniaturization of the method of Lowry et al. [35]. This assay does not take into account any
The Biomaterials Silver Jubilee Compendium
151
J. Lincks et al. /Biomaterials 19 (1998) 2219-2232
degradation that may have occurred during the incubation or during sample preparation.
2.6.2. Proteoglycan sulfation Proteoglycan synthesis was assessed by [-35S]-sulfate incorporation according to the method of O'Keefe et al. [36]. Previously, we found that the amount of radiolabeled proteoglycan secreted into the media by MG63 cells was less than 15 % of the total radiolabeled proteoglycan produced. Because more than 85% of the radiolabeled proteoglycan was in the cell layer, we examined the incorporation of [35S]-sulfate only in the cell layer. At confluence, 50 ~tl D M E M containing 90 ~tCiml-1 [35S]-sulfate were added to the media to make a final concentration of 9 ~tCiml-1. Four hours later, the media were discarded and the wells washed one time with 500 ~tl PBS. The cell layer was collected in two 0.25 ml portions of 0.25 M NaOH. The protein content was determined by the method of Lowry et al. [35]. To measure [-3SS]sulfate incorporation into the cell layers, the total volume was adjusted to 0.7 ml by the addition of 0.15 M NaC1 and the sample dialyzed in a 12000-14000 molecular weight cut-off membrane against buffer containing 0.15 M NaC1, 20 mM NazSO4, and 20 mM NazHPO4 at pH 7.4 and 4~ The dialysis solution was changed until the radioactivity in the dialysate reached background levels. The amount of [35S]-sulfate incorporated was determined by liquid scintillation spectrometry and was calculated as dpm mg-1 cell layer protein.
2. 7. Local factor production 2. 7.1. Prostaglandin E: The amount of PGE2 produced by the cells and released into the media was assessed using a commercially available competitive binding radioimmunoassay kit (NEN Research Products, Boston, MA). In this assay, unlabeled PGE2 in the sample was incubated overnight with radiolabeled PGE2 and unlabeled PGE2 antibody. Antigen-antibody complexes were separated from free antigen by precipitation with polyethylene glycol. Sample PGE2 concentrations were determined by correlating the percentage bound over unbound counts to a standard curve.
2. 7.2. Transforming growth factor-beta (TGF-fi) In order to measure the level of total TGF-~ production by the cells, a commercially available enzyme-linked immunoassay (ELISA) kit (Promega Corp., Madison, WI) specific for human TGF-~I was used. Immediately prior to assay, conditioned media were diluted 1:10 in D M E M and the 1:10 dilution further diluted by adding four volumes of PBS. The media were then acidified by the addition of 1 M HC1 for 15 min to activate latent TGF-fi (LTGF-fl), followed by neutralization with 1 M
2223
NaOH. The assay was performed according to the manufacturer's directions. Intensity measurements were conducted at 450nm using a BioRad Model 2550 EIA Reader (Hercules, CA). Sample concentrations were determined by comparing the absorbance value to a known standard curve. The amount of TGF-fil in the cell layer was not examined because of difficulties associated with quantitatively extracting this cytokine from the matrix.
2.8. Statistical analysis Experiments were conducted at least twice and the data shown are from one representative experiment. For any given experiment, each data point represents mean + SEM of six individual cultures. Data were first analyzed by analysis of variance; when statistical differences were detected, the Student's t-test for multiple comparisons using Bonferroni's modification was used. P-values <0.05 were considered to be significant. m
3. Results
3.1. Disk characteristics 3.1.1. Morphology When the Ti-S and Ti-A-S disks were examined by scanning electron microscopy, the surfaces were found to be very similar (Fig. 1A and C). Morphologically, the disks had small pits (2 jam in diameter) and randomly oriented scratches from the polishing operation, which were only evident at high magnification (data not shown). The Ti-R and Ti-A-R disks also had a similar appearance (Fig. 1B and D) and contained parallel, longitudinal grooves with both sharp and serrated edges, resulting from the grinding operation. Parallel grooves of varying heights were prominent; in addition, the distance between the grooves varied. On both rough surfaces, curved sheets of material were observed occasionally at the apex of the grooves. Additionally, the Ti-A-R surface contained areas with pits that were 10-20 jam in diameter.
3.1.2. Surface roughness Based on profilometry (Table 1) the smooth surfaces, Ti-S and Ti-A-S, had similar R, values of 0.22 and 0.23 ~tm, respectively. The Ti-R surface was the roughest and had a n R a of 4.24 ~tm, while the Ti-A-R surface had a n Ra of 3.20 ~tm. Both rough surfaces were significantly rougher than both smooth surfaces.
3.1.3. Auger electron spectroscopy Both smooth surfaces (Ti-S and Ti-A-S) were found to contain Ti, O, and C by Auger electron spectroscopy before sputtering. In the alloyed surface, A1 was also found. After 10 s of sputtering, the C signal was virtually gone at a depth of 67 A in both Ti and Ti-A disks. In
152 2224
The B i o m a t e r i a l s Silver J u b i l e e C o m p e n d i u m J. Lincks et al./Biomaterials 19 (1998) 2219-2232
Fig. 1. Scanning electron micrographs of the different disk surfaces used in this study. Panel A: Ti-S; Panel B: Ti-R; Panel C: Ti-A-S; Panel D: Ti-A-R. Bar = 200 ~tm. Original magnification: 100x.
Table 1 Average surface roughness values for the Ti and Ti-alloy disks used in this study
the O signal virtually disappeared. N o v a n a d i u m was found in the disks.
Surface
Ra value
Ti-S Ti-R Ti-A-S Ti-A-R
0.22 + 0.00a 4.24 + 0.13 b 0.23 + 0.00a 3.20 + 0.12
3.1.4. Fourier transform infrared spectroscopy F T I R analysis of the disks confirmed that no organic residue was left on the surface of either the Ti-S or Ti-A-S disks.
Note: Ti and Ti alloy (Ti-A) disks were prepared with either a smooth (S) or rough (R) surface as described in the Materials and Methods. The Ra value for each disk type was determined by profilometry. Data shown in the table represent the mean +_SEM for four disks in each group; each disk was measured in four areas. a p < 0.05, smooth vs. rough surface. b p < 0.05, T i - R vs. T i - A - R .
addition to Ti and O, A1 was also present in the alloy. Twenty seconds of sputtering to a depth of 134 ,~ produced a continuously decreasing O signal while sputtering t h r o u g h the oxide layer, and an increasing Ti signal. After one minute, the Ti signal became very strong, and
evidence of
3.2. Cell morphology The a p p e a r a n c e of the cells varied with surface roughness and chemical composition of the disks. Cells grown on the Ti-S surface were spread out across the surface and grew as a monolayer, but this m o n o l a y e r was not continuous (Fig. 2C and D). The cells had a dendritic appearance, with extensions that were up to 10 ~tm in length and had ruffled m e m b r a n e s on their surfaces. Cells cultured on Ti-R (Fig. 2A and B) and Ti-A-S (Fig. 3C and D) disks grew as a continuous, thin m o n o l a y e r across the surface. O n the Ti-R surface, all cracks and fissures were covered by a m o n o l a y e r of cells (Fig. 2A and B). Cultures on the Ti-A-R surface induced the cells to grow as a multilayer (Fig. 3A and B), with m a n y cells p r o d u c i n g
153
The Biomaterials Silver Jubilee Compendium J. Lincks et al./Biomaterials 19 (1998) 2219-2232
2225
Fig. 2. Scanning electron micrographs of MG63 osteoblast-like cells cultured on smooth and rough Ti surfaces. Panel A: Ti-R, magnification: 100• bar = 10 ~tm; Panel B: Ti-R, magnification: 500x, bar = 1 ~tm; Panel C: Ti-S, magnification: 100• bar = 10 ~tm; Panel D: Ti-S, magnification: 500x, bar = 1 ~tm.
extensions that covered distances of up to 10 lam. In addition, the cells were oriented along the parallel cracks and grooves and grew over the sharp edges, forming a multilayer.
3.3. Cell proliferation 3.3.1. Cell number Cell number was affected by both chemical composition and surface roughness (Fig. 4). Compared to plastic, cell number was reduced by 36% on Ti-R. Although not statistically significant, cell number was also reduced by 20% on Ti-S. Fewer cells were present on the Ti-R surfaces than on Ti-A-R as well. The numbers of the cells grown on the Ti-A-S and Ti-A-R surfaces were similar to that seen on the plastic.
3.3.2. [3H]-thymidine incorporation [3H]-thymidine incorporation was reduced on all metal surfaces when compared to plastic (Fig. 5). The effect was comparable on the alloyed Ti surfaces (49%) and the Ti-R surface (48%). However, the decrease seen on the Ti-S surface was significantly less than on the other surfaces (19%).
3.4. Cell differentiation 3.4.1. Alkaline phosphatase specific activity Enzyme activity varied with surface roughness and composition (Fig. 6). Cell layers from cells cultured on all different surfaces contained significantly more alkaline phosphatase specific activity than on the plastic control (1.6 fold to 2.2 fold). Activity on Ti-R was 20% greater than on Ti-A-R. Activity on the rough surfaces was consistently greater than on smooth surfaces. Alkaline phosphatase on Ti-R was 1.5-fold greater than on Ti-S; on Ti-A-R, alkaline phosphatase was 1.3-fold greater than on Ti-A-S. When enzyme activity of isolated cells was measured, similar observations were made (Fig. 7). Cells grown on Ti-R surfaces exhibited a 1.8-fold increase in enzyme activity over that seen on plastic. On Ti-R, the increase was 1.4-fold, and on the smooth surface disks, there was a 1.3-fold increase. Activity was greater on Ti-R in comparison to Ti-S and in comparison to Ti-A-R. These results also showed that the effects of surface roughness and composition on alkaline phosphatase specific activity were primarily due to enzyme present in the matrix. Specific activity of the cell layer was consistently
154 2226
The B i o m a t e r i a l s S i l v e r J u b i l e e C o m p e n d i u m J. Lincks et al./Biomaterials 19 (1998) 2219-2232
Fig. 3. Scanning electron micrographs of MG63 osteoblast-like cells cultured on smooth and rough Ti-A-surfaces. Panel A: Ti-A-R, magnification: 100• bar = 10~m; Panel B: Ti-A-R, magnification: 500• bar = 1 ~m; Panel C: Ti-A-S, magnification: 100• bar = 10 ~tm; Panel D: Ti-A-S, magnification: 500• bar = 1 ~tm.
Fig. 4. Number of MG63 osteoblast-like cells released by two trypsinizations of the Ti disks 24 h after they had reached confluence on the plastic. Values are the mean ___ SEM of six cultures. *P < 0.05, Ti disk vs. plastic; # P < 0.05, Ti-A-R vs. Ti-R. Data are from one of two replicate experiments. t w o t i m e s t h a t of t h e i s o l a t e d cells, d e s p i t e t h e l a r g e r d e n o m i n a t i o n d u e to t h e p r e s e n c e of m a t r i x p r o t e i n . T h e f o l d - i n c r e a s e s n o t e d as a f u n c t i o n of e i t h e r r o u g h n e s s or c o m p o s i t i o n w e r e g r e a t e r w h e n a s s a y i n g cell layers, res u l t i n g in s i g n i f i c a n t l y g r e a t e r real e n z y m e a c t i v i t y t h a n w a s seen in t h e i s o l a t e d cells. T h i s w a s p a r t i c u l a r l y evid e n t for cell l a y e r s c u l t u r e d o n Ti-R.
Fig. 5. [3H]-Thymidine incorporation by MG63 osteoblast-like cells during culture on plastic or Ti disks. When the cells reached confluence on plastic, the media were changed and culture continued for another 24 h. Four hours prior to harvest, [3H]-thymidine was added and incorporation into TCA insoluble cell precipitates measured. Values are the mean _+ SEM of six cultures. *P < 0.05, Ti disk vs. plastic: # P < 0.05, Ti-S vs. Ti-R. Data are from one of two replicate experiments.
3.4.2. Osteocalcin production Cell c u l t u r e s g r o w n o n t h e T i - R surface s h o w e d a significant i n c r e a s e (1.9 fold) in o s t e o c a l c i n p r o d u c t i o n c o m p a r e d to p l a s t i c (Fig. 8). T h e o s t e o c a l c i n p r o d u c t i o n by
155
The Biomaterials Silver Jubilee Compendium J. Lincks et al./Biomaterials 19 (1998) 2219-2232
2227
Fig. 6. Alkaline phosphatase specific activity of cell layers produced by MG63 osteoblast-like cells during culture on Ti disks. After cells had reached confluence on plastic, cultures were continued for an additional 24 h and then harvested by scraping. Enzyme activity was measured in the cell layer lysate. Values are the mean + SEM of six cultures. * P <0.05, titanium vs. plastic; # P <0.05, Ti-A-R vs. Ti-R; 9 P < 0.05, smooth vs. rough surface of same material. Data are from one of two replicate experiments.
Fig. 8. Osteocalcin production by MG63 osteoblast-like cells during culture on Ti disks. After cells reached confluence on plastic, the media were changed and the culture continued for an additional 24 h. At harvest, the media were collected, and osteocalcin content measured by RIA. Values are the mean • SEM of six cultures. * P < 0.05, titanium vs. plastic. Data are from one of two replicate experiments.
Fig. 7. Alkaline phosphatase specific activity of trypsinized MG63 osteoblast-like cells after culture on Ti disks. After cells had reached confluence on plastic, cultures were continued for an additional 24 h and then harvested by trypsinization. Enzyme activity was measured in lysates of the cells. Values are the mean _+ SEM of six cultures. * P <0.05, titanium vs. plastic; # P <0.05, Ti-A-R vs. Ti-R; 9 P < 0.05, smooth vs. rough surface of same material. Data are from one of two replicate experiments.
Fig. 9. Percent collagen production by MG63 osteoblast-like cells during culture on Ti disks. Values were derived from CDP and NCP production and are the mean _+ SEM of six cultures. * P < 0.05, titanium vs. plastic; # P < 0.05, Ti-A-R vs. Ti-R; 9 P < 0.05, smooth vs. rough surface of same material. Data are from one of two replicate experiments.
cells g r o w n plastic.
o n all t h e o t h e r
s u r f a c e s w a s s i m i l a r to
p r o d u c t i o n b y t h e cells w a s s i g n i f i c a n t l y d e c r e a s e d ( 1 5 % ) o n r o u g h T i - A - R s u r f a c e s c o m p a r e d to T i - R surfaces. M o r e o v e r , cells o n T i - S p r o d u c e d 3 1 % less c o l l a g e n t h a n
3.5. Matrix production
o n T i - R , a n d cells o n T i - A - S p r o d u c e d 1 7 % less c o l l a g e n than on Ti-A-R.
3.5.1. Collagen production
3.5.2. Proteoglycan sulfation
C o l l a g e n s y n t h e s i s w a s a l s o affected b y s u r f a c e c o m -
Compared
to
plastic,
[-35S]-sulfate
p o s i t i o n a n d r o u g h n e s s (Fig. 9). W h i l e c o l l a g e n s y n t h e s i s
by M G 6 3
w a s u n a f f e c t e d in cells c u l t u r e d o n T i - R , cells g r o w n o n
surfaces e x a m i n e d ( 3 5 - 4 8 % )
Ti-S, T i - A - R a n d T i - A - S s u r f a c e s s y n t h e s i z e d 1 4 - 3 0 %
least p r o n o u n c e d
less c o l l a g e n c o m p a r e d
surface. N o
t o plastic. T h e p e r c e n t c o l l a g e n
incorporation
cells w a s s i g n i f i c a n t l y r e d u c e d
o n all d i s k
(Fig. 10). T h i s effect w a s
in cells g r o w n o n t h e s m o o t h T i - A - R
significant
difference in t h e
[35S]-sulfate
156
The Biomaterials Silver Jubilee Compendium J. Lincks et al./Biomaterials 19 (1998) 2219-2232
2228
Effect of Disk Surface on [3ss]-sulfate Incorporation
Fig. 10. [35S]-Sulfate incorporation by MG63 osteoblast-like cells during culture on Ti disks. When the cells reached confluence on plastic, the media were changed and culture continued for another 24 h. Four hours prior to harvest, [35S]-sulfate was added and incorporation into the cell layer measured. Values are the mean + SEM of six cultures. * P < 0.05, titanium vs. plastic. Data are from one of two replicate experiments.
Fig. 12. Latent transforming growth factor fl (LTGFfl) production by MG63 osteoblast-like cells during culture on Ti disks. After the cells reached confluence on plastic, the media were changed and the culture continued for an additional 24 h. At harvest, the media were collected, and LTGFfi content measured by ELISA. Values are the mean _+ SEM of six cultures. * P < 0.05, titanium vs. plastic; # P < 0.05, Ti-A-R vs. Ti-R; 9 P < 0.05, smooth vs. rough surface. Data are from one of two replicate experiments. those on the Ti-A-S surface. T h e levels on b o t h s m o o t h surface p r e p a r a t i o n s were n o t significantly different f r o m plastic.
3.6.2. Transforming growth factor-~
Fig. 11. Prostaglandin E2 (PGE2) production by MG63 osteoblast-like cells during culture on Ti disks. After cells reached confluence on plastic, the media were changed and the culture continued for an additional 24 h. At harvest, the media were collected, and PGE2 content measured by RIA. Values are the mean _ SEM of six cultures. * P <0.05, titanium vs. plastic; # P <0.05, Ti-A-R vs Ti-R; 9 P < 0.05, smooth vs. rough surface. Data are from one of two replicate experiments. i n c o r p o r a t i o n a m o n g the different surface r o u g h n e s s e s a n d c o m p o s i t i o n s was observed.
3.6. Local factor production 3.6.1. Prostaglandin E2 T h e level of P G E 2 p r o d u c t i o n by the cells was affected by the different surface t r e a t m e n t s (Fig. 11). Significantly m o r e P G E 2 was p r o d u c e d by cells c u l t u r e d on Ti-R w h e n c o m p a r e d to plastic (3.9-fold) a n d to Ti-S surface (2.0 fold). Cells on the Ti-A-R surface synthesized 2.9-fold m o r e P G E 2 t h a n those on plastic a n d 3.3-fold m o r e t h a n
T h e level of latent T G F - / / i n the c o n d i t i o n e d m e d i a was also influenced by culture on the different surfaces (Fig. 12). L a t e n t TGF-/~ levels were increased by 1.7-fold on the Ti-A-R a n d 2.7-fold on the Ti-R surfaces. L a t e n t TGF-/~ p r o d u c t i o n was greater on the Ti-R surface comp a r e d to cultures g r o w n on the Ti-A-R surface (1.6-fold) a n d 2.1-fold greater w h e n c o m p a r e d to Ti-S. T h e r e was a slight, b u t insignificant, increase in L T G F / ~ levels prod u c e d by cells g r o w n on b o t h s m o o t h surfaces c o m p a r e d to plastic.
4. Discussion This study confirms previous o b s e r v a t i o n s t h a t osteoblast-like cells r e s p o n d in a differential m a n n e r to b o t h surface r o u g h n e s s [7, 37-39] a n d m a t e r i a l comp o s i t i o n [ 2 5 , 4 0 - 4 2 ] . As n o t e d previously [ 8 - 1 0 ] , M G 6 3 cells g r o w n on Ti-R surfaces exhibited a m o r e differentiated p h e n o t y p e as evidenced by r e d u c e d cell proliferation a n d increased alkaline p h o s p h a t a s e specific activity a n d osteocalcin p r o d u c t i o n . Cells g r o w n on Ti-S surfaces also exhibited r e d u c e d cell proliferation, a n d they h a d elevated alkaline p h o s p h a t a s e in c o m p a r i s o n with cultures g r o w n on plastic, b u t the effects were less r o b u s t t h a n those seen on Ti-R. M o r e o v e r , osteocalcin p r o d u c t i o n was u n a l t e r e d in these M G 6 3 cells, indicating t h a t they were n o t as differentiated as those cells g r o w n on Ti-R.
157
The Biomaterials Silver Jubilee Compendium J. Lincks et al. / Biomaterials 19 (1998) 2219-2232
Although [3H]-thymidine incorporation was reduced in cells cultured on Ti-A-R, total cell number was unaffected. The latter value is a cumulative measure of the viable cells in the culture, whereas the former value is an indication of the rate of DNA synthesis, and therefore, cell replication during the radiolabeling period, in our case, the last four hours of culture. This indicates that the cells grown on Ti-A-R ceased to proliferate and initiated expression of the mature osteoblastic phenotype at a slower rate than cells cultured on Ti-R, since proliferation is negatively correlated with phenotypic expression [43]. This hypothesis is supported by the fact that alkaline phosphatase activity on Ti-A-R was elevated, but to a lesser degree than seen on Ti-R, and the MG63 cells on Ti-A-R did not exhibit elevated osteocalcin production. Even for the alloy disks, however, the cells cultured on the rougher surfaces were more differentiated than the cells cultured on the smoother surfaces. Other aspects of osteoblast function were sensitive to the substrate, either with respect to roughness or to the bulk composition of the material. Production of extracellular matrix vesicles was affected by the nature of the substrate based on differences in cell layer alkaline phosphatase, where matrix vesicles are present, when compared to enzyme activity in isolated cells. Alkaline phosphatase is an early marker of osteogenic differentiation. While this enzyme activity is present in all cell membranes, it is found in higher levels in cells which mineralize their matrix such as osteoblasts [44]. As osteoblasts mature, they produce extracellular matrix vesicles which are enriched in alkaline phosphatase specific activity; because of this specific enrichment, alkaline phosphatase is the marker enzyme for this extracellular organelle [45]. Matrix vesicles are associated with the onset of calcification and they contain enzymes necessary for matrix modification necessary for crystal deposition and growth [46, 47]. The results of the present study show clearly that the effects of surface roughness were targeted to the matrix vesicles, whether the cells were cultured on Ti-R or Ti-A-R, since the fold increases in enzyme activity in the cell layer were significantly greater than the fold increases observed in the isolated cells. In addition, the effects of material composition were also found predominately in the matrix vesicle compartment, supporting previous in vivo and in vitro observations. Studies examining endosteal healing adjacent to various implant materials demonstrate that matrix vesicle production and function are sensitive to the type of material used [-41, 48, 49]. Similarly, when cells were cultured on thin films of various implant materials which had been sputtered onto tissue culture plastic, the effects of material composition were targeted to the matrix vesicles [26]. In comparison to plastic, proteoglycan sulfation was reduced in all of the cultures to a comparable extent. In
2229
contrast, collagen production was differentially affected by the nature of the surface topography and the material used. In general, synthesis on rougher surfaces was greater than seen on smoother surfaces, correlating with the production of latent TGF-/~. The expression of this growth factor is associated with the collagen deposition in the extracellular matrix of osteoblasts [50]. Similarly, production of PGE2 was greater on the rougher surfaces, supporting our previous observation that there is a positive correlation of latent TGF-/~ and PGE2 production with increasing surface roughness [10]. Both latent TGF-/3 and PGE2 are produced by osteoblasts as paracrine and autocrine regulators of cell function and differentiation. Their release by the MG63 cells cultured on Ti-S and Ti-A-S was essentially identical to the basal levels seen on plastic, another smooth surface. However, on the rougher Ti-R and Ti-A-R surfaces, their production was markedly enhanced, although in a material-specific manner, with the greatest production being observed in cells grown on Ti-R. This supports the contention that these cells exhibit a more differentiated osteoblastic phenotype. Whether more differentiated cells produce higher levels of these local factors, or whether the cells are more differentiated because they produce and respond to higher levels of these factors, is not known. The amounts of PGE2 produced per culture are well within the limits of prostaglandin known to be osteogenic and not inflammatory [ 11]. Since all of the TGF-/~ released into the media was in latent form, it is difficult to comment on its contribution to the differentiation of the MG63 cells. However, recent studies in our lab [51] and others [52] indicate that the latent TGF-/3 which is incorporated into the matrix may be activated locally via the action of matrix vesicles and may regulate the phenotypic expression of the cells. There is some indication that this is the case in the present study. In cells cultured on Ti-R surfaces, both ALPase and osteocalcin production were increased, whereas on Ti-A-R surfaces, ALPase was stimulated and osteocalcin production was not. When osteoblasts are treated with TGF-/3, alkaline phosphatase, an early marker of osteoblastic differentiation, is stimulated [12], whereas production of osteocalcin, a marker of terminal differentiation, is inhibited [12]. Whether TGF-/~ is modulating the differential expression of osteoblastic phenotypic markers in the MG63 cells is certainly not established by this study but the potential for regulation of this type is evidenced by the fact that production of local regulatory factors is sensitive to the material used. The results presented here also support our previous observation that roughness may play a more important role in determining cell response than the type of topography, as long as the R a values can be sensed by the cells. For practical purposes, the distance between peaks
The Biomaterials Silver Jubilee Compendium
158 2230
J. Lincks et al./Biomaterials 19 (1998) 2219-2232
should not exceed the ability of the cell to form focal attachments on two or more peaks; otherwise, the cell would sense a rough surface as smooth. In the present study, surface roughness was achieved by machining, resulting in parallel grooves, whereas our previous studies used commercially pure Ti disks that were roughened by grit-blasting and acid-etching, resulting in random peaks and valleys. In general, the MG63 cells responded to smooth surfaces in a manner similar to their behavior on tissue culture plastic and to machined Ti-R surfaces in a manner similar to grit-blasted Ti surfaces with comparable R a values. The morphology of the cells on the Ti-R and Ti-A-R surfaces demonstrates that they have assumed a more cuboidal shape with dendritic extensions, similar to the morphology noted on rough cpTi surfaces achieved by grit-blasting, and typical of a more differentiated osteoblast. Similar observations have been noted with chick embryonic osteoblasts [37]. In contrast, cells on the smoother surfaces appear more flattened and fibroblastic. Our data also show that MG63 cells are sensitive to the bulk composition of the material, whether the surface is smooth or rough. Even though a titanium oxide layer formed on both the Ti and Ti-alloy surfaces, it is unlikely that the oxides were identical. Certainly mosaicism of the alloy components would result in a more complex surface chemistry. This would have a direct effect on the nature of the conditioning film that forms as the material surface interacts with the culture medium [53-55]. In addition, ions released from the alloy could also modulate cellular response. Recently, studies using fibroblast cultures demonstrated that locally released vanadium ions from Ti-6A1-4V alloy surfaces negatively impacted cell adhesion [56]. Thompson and Puleo [57] have also shown that Ti-6A1-4V ion solutions can inhibit expression of the osteogenic phenotype by bone marrow stromal cells, suggesting that ions released from implants could also impair normal bone formation. Despite the differences in cellular response due to material composition, roughness remains the overriding variable in promoting osteogenic differentiation. As strength requirements of orthopaedic implants necessitate the need for alloyed titanium preparations, it is essential that the optimal surface characteristics be determined, potentially mitigating any negative effects of the bulk material on bone formation and function.
Acknowledgements The authors gratefully acknowledge the expert assistance of Sandra Messier, Monica Luna, Kimberly Rhame, and Roland Campos in the preparation of the manuscript. Jack Lincks is a fellow in the Air Force Institute of Technology. This work does not necessarily reflect the views of the United States Air Force. Funding
for this research was provided by the Center for the Enhancement of the Biology/Biomaterials Interface at the University of Texas Health Science Center at San Antonio. Support for Dr. Lohmann was provided by a grant from the B. Braun Foundation. Melsungen, Germany.
References [1] Rich A, Harris AK. Anomalous preferences of cultured macrophages for hydrophobic and roughened substrata. J Cell Sci 1981;50:1-7. 1-2] Thomas K, Cook SD. An evaluation of variables influencing implant fixation and direct bone appostition. J Biomed Mater Res 1985;19:875-901. [-3] Schroeder A, Van der Zypen E, Stich H, Sutter F. The reactions of bone, connective tissue and epithelium to endosteal implants with titanium sprayed surfaces. J Maxillofac Surg 1981;9:15-25. [-4] Buser D, Schenk R, Steinemann S, Fiorellini J, Fox C, Stich H. Influence of surface characteristics on bone integration of titanium implants. A histomorphometric study in miniature pigs. J Biomed Mater Res 1991;25:889-902. 1-5] Cochran DL, Simpson J, Weber H, Buser D. Attachment and growth of periodontal cells on smooth and rough titanium. Int J Oral Maxillofac Implants 1994;9:289-97. [6] Michaels CM, Keller JC, Stanford CM, Solursh M. In vitro cell attachment of osteoblast-like cells to titanium. J Dent Res 1989;68:276. [7] Bowers KT, Keller JC, Randolph BA, Wick DG, Michaels CM. Optimization of surface micromorphology for enhanced osteoblast responses in vitro. Int J Oral Maxillofac Implants 1992;7:302-10. [8] Martin JY, Schwartz Z, Hummert TW, Schraub DM, Simpson J, Lankford J, Dean DD, Cochran DL, Boyan BD. Effect of titanium surface roughness on proliferation, differentiation, and protein synthesis of human osteoblast-like cells (MG63). J Biomed Mater Res 1995;29:389-401. [-9] Boyan BD, Batzer R, Kieswetter K, Liu Y, Cochran DL, Szmuckler-Moncler S, Dean DD, Schwartz Z. Titanium surface roughness alters responsiveness of MG63 osteoblast-like cells to l~,25(OH)zD3. J Biomed Mater Res 1998;39:77-85. [10] Kieswetter K, Schwartz Z, Hummert TW, Cochran DL, Simpson J, Dean DD, Boyan BD. Surface roughness modulates the local production of growth factors and cytokines by osteoblast-like MG63 cells. J Biomed Mater Res 1996;32:55-63. [11] Raisz LG, Fall PM. Biphasic effects of prostaglandin E2 on bone formation in cultured fetal rat calvariae: interaction with cortisol. Endocrinology 1990;126:1654-9. [12] Bonewald LF, Kester MB, Schwartz Z, Swain LD, Khare AG, Johnson TL, Leach RJ, Boyan BD. Effects of combining transforming growth factor/~ and 1,25-dihydroxyvitamin D3 on differentiation of a human osteosarcoma (MG63). J Biol Chem 1992;267:8943-9. [13] Akatsu T, Takahashi N, Udagawa N, Imamura K, Yamaguchi A, Sato K, Nagata N, Suda T. Role of prostaglandins in interleukin1-induced bone resorption in mice in vitro. J Bone Miner Res 1991;6:183-90. [14] Bonewald LF, Mundy GR. Role of transforming growth factor-beta in bone remodeling. Clin Orthop Rel Res 1990;250: 261-76. [15] Cheroudi B, Gould TRL, Brunette DM. Titanium-coated micromachined grooves of different dimensions affect epithelial and connective tissue cells differently in vivo. J Biomed Mater Res 1990;24:1203-19.
The Biomaterials Silver Jubilee Compendium
159
J. Lincks et al./Biomaterials 19 (1998) 2219-2232
[16] Cheroudi B, Gould TRL, Brunette DM. Effects of grooved titanium-coated implant surface on epithelial cell behavior in vitro and in vivo. J Biomed Mater Res 1987;23:1067-85. [17-] Brunette DM. The effects of implant surface topography on the behavior of cells. Int J Oral Maxillofac Implants 1988;3:231-46. [18] Van Noort R. Review titanium: the implant material of today. J Mater Sci 1987;22:3801-11. [19] Grabowsky KS, Gossett CR, Young FA, Keller JC. Cell adhesion to ion implanted titanium. Mater Res Soc Symp Proc 1989; 110:697-701. [20] Rostoker W, Galante JO. Materials for human implantation. J Biomech Eng 1979;101:2-14. [21] Lausmaa J, Mattson L, Rolander U, Kasemo B. Chemical composition and morphology of titanium surface oxides. Mater Res Soc Symp Proc 1986;55:351-9. [22] Golijanin L, Bernard G. Biocompatibility of implant metals in bone tissue culture. J Dent Res 1988;67:367. [23] Golijanin G, Bernard G, Tuck M, Davlin L. Comparative study of the canine bone/implant interface in vitro and in vivo. J Dent Res 1989;68:307. [24] Nowlin P, Carnes D, Windeler A. Biocompatibility of dental implant materials sputtered onto cell culture dishes. J Dent Res 1989;68:275. [25] Arai T, Pilliar M, Melcher AH. Growth of bone-like tissue on titanium and titanium alloy in vitro (Abstract #3). Trans Soc Biomet Ann Meeting 1989. [26] Hambleton JC, Schwartz Z, Windeler SW, Luna MH, Brooks BP, Khare AG, Dean DD, Boyan BD. Culture surfaces coated with various implant materials affect chondrocyte growth and metabolism. J Orthop Res 1994;12:542-52. [27] Evans EJ, Benjamin M. The effect of grinding conditions on the toxicity of cobalt-chrome-molybdenum particles in vitro. Biomaterials 1987;8:377-84. [28] Merritt K, Brown SA. Biological effects of corrosion products from metals. In: Fraker AC, Griffer CD, editor. Corrosion and degradation of implant materials: 2nd Symposium, ASTM STP 859P. Philadelphia, PA: Am Soc Testing Materials, 1985:195. [29] Franceschi RT, James WM, Zerlauth G. 1,~,25-dihydroxyvitamin D3 specific regulation of growth, morphology, and fibronectin in a human osteosarcoma cell line. J Cell Physiol 1985;123:401-9. [30] Boyan BD, Schwartz Z, Bonewald LF, Swain LD. Localization of 1,25-(OH)2D3 responsive alkaline phosphatase in osteoblast-like cells (ROS 17/2.8, MG63, and MC3T3) and growth cartilage cells in culture. J Biol Chem 1989;264:11 879-86. [31] Schwartz Z, Schlader DL, Ramirez V, Kennedy MB, Boyan BD. Effects of vitamin D metabolites on collagen production and cell proliferation of growth zone and resting zone cartilage cells in vitro. J Bone Miner Res 1989;4:199-207. [32] Bretaudiere JP, Spillman T. Alkaline phosphatases. In: Bergmeyer HU, editor. Methods of enzymatic analysis. Weinheim: Verlag Chemica, 1984:75-92. [33] Hale LV, Kemick ML, Wuthier RE. Effect of vitamin D metabolites on the expression of alkaline phosphatase activity by epiphyseal hypertrophic chondrocytes in primary cell culture. J Bone Miner Res 1986;1:489-95. [34] Peterkofsky B, Diegelmann R. Use of a mixture of proteinase-free collagenases for the specific assay of radioactive collagen in the presence of other proteins. Biochemistry 1971;10:988-94. [35] Lowry OH, Rosebrough NJ, Farr AL, Rano RI. Protein measurement with the folin phenol reagent. J Biol Chem 1951; 193:265-75. [36] O'Keefe RJ, Puzas JE, Brand JS, Rosier RN. Effects of transforming growth factor-beta on matrix synthesis by chick growth plate chondrocytes. Endocrinology 1988;122:2953-61. [37] Groessner-Schreiber B, Tuan RS. Enhanced extracellular matrix production and mineralization by osteoblasts cultured on titanium surfaces in vitro. J Cell Sci 1992;101:209-17.
2231
[38] Ong JL, Prince CW, Raikar GN, Lucas LC. Effect of surface topography of titanium on surface chemistry and cellular response. Implant Dentistry 1996;5:83-8. [39] Windeler AS, Bonewald LF, Khare AG, Boyan BD, Mundy GR. The influence of sputtered bone substitutes on cell growth and phenotypic expression. In: Davies JE, editor. The bone-biomaterial interface. Toronto: Toronto Press, 1991:205-13. [40] Vrouwenfelder WC, Groot CG, deGroot K. Better histology and biochemistry for osteoblasts cultured on titanium-doped bioactive glass: bioglass 45S5 compared with iron-, titanium-, fluorine-, and boron-containing bioglasses. Biomaterials 1994;15:97-106. [41] Braun G, Kohavi D, Amir D, Luna MH, Caloss R, Sela J, Dean DD, Boyan BD, Schwartz Z. Markers of primary mineralization are correlated with bone-bonding ability of titanium or stainless steel in vivo. Clin Oral Implants Res 1995;6:1-13. [42] Bordij K, Jouzeau JY, Mainard D, Payan E, Netter P, Rie KT, Stucky T, Hage-Ali M. Cytocompatibility of Ti-6A1-4V and Ti-5A1-2.5Fe alloys according to three surface treatments, using human fibroblasts and osteoblasts. Biomaterials 1996;17: 929-40. [43] Rickard DJ, Gowen M, MacDonald BR. Proliferative responses to estradiol, IL-1 alpha and TGF beta by cells expressing alkaline phosphatase in human osteoblast-like cell cultures. Calcif Tissue Int 1993;52:227-33. [44] Owen TA, Aronow M, Shalhoub V, Barone LM, Wilming L, Tassinari MS, Kennedy MB, Pockwinse S, Lian JB, Stein GS. Progressive development of the rat osteoblast phenotype in vitro: reciprocal relationships in expression of genes associated with osteoblast proliferation and differentiation during formation of the bone extracellular matrix. J Cell Physiol 1990;143:420-30. [45] Ali SY. Mechanisms of calcification. In: Owen R, Goodfellow J, Bollough P, editors. Scientific foundation of orthopaedics and traumatology. London: Heinemann, 1984:175-95. [46] Dean DD, Schwartz Z, Bonewald LF, Muniz OE, Morales SM, Gomez R, Brooks BP, Qiao M, Howell DS, Boyan BD. Matrix vesicles produced by osteoblast-like cells in culture become significantly enriched in proteoglycan-degrading metalloproteinases after addition of/3-glycerophosphate and ascorbic acid. Calcif Tissue Int 1994;54:399-408. [47] Anderson HC. Matrix vesicle calcification: review and update. In: Pick WA, editor. Bone and mineral research. Amsterdam: Excerpta Medica, 1984:109-49. [48] Boyan BD, Schwartz Z, Dean DD, Hambleton JC. Response of bone and cartilage cells to biomaterials in vivo and in vitro. J Oral Implantol 1993;19:116-22. [49] Schwartz Z, Amir D, Boyan BD, Cochavy D, Muller-Mai C, Swain LD, Gross U, Sela J. Effect of glass ceramic and titanium implants on primary calcification during rat tibial bone healing. Calcif Tissue Int 1991;49:359-64. [50] Choi JY, Lee BH, Song KB, Park RW, Kim IS, Sohn KY, Jo JS, Ryoo HM. Expression patterns of bone-related proteins during osteoblastic differentiation in MC3T3-E1 cells. J Cell Biochem 1996;61:609-18. [51] Boyan BD, Schwartz Z, Park-Snyder S, Dean DD, Yang F, Twardzik D, Bonewald LF. Latent transforming growth factor-/~ is produced by chondrocytes and activated by extracellular matrix vesicles upon exposure to 1,25-(OH)zD3. J Biol Chem 1994;269:28,374-81. [52] Dallas SL, Miyazono K, Skerry TM, Mundy GR, Bonewald LF. Dual role for the latent transforming growth factor-/~ binding protein in storage of latent TGF-/3 in the extracellular matrix and as a structural matrix protein. J Cell Biol 1995;131: 539-49. [53] Norde W. Behavior of proteins at the interfaces, with special attention to the role of the structure stability of the protein molecule. Clin Mater 1998;11:85-91.
160 2232
The Biomaterials Silver Jubilee Compendium J. Lincks et al./Biomaterials 19 (1998) 2219-2232
[54] Hench LL, Paschall HA. Direct chemical bond of bioactive glassceramic materials to bone and muscle. J Biomed Mater Res 1973;7:25-42. [55] Jarcho M, Kay JK, Gumaer RH, Doremus RH, Drobeck HP. Tissue, cellular, and subcellular events at a bone-ceramic hydroxylapatite interface. J Bioeng 1977;1:79-92.
[56] Eisenbarth E, Meyle J, Nachtigall W, Breme J. Influence of the surface structure of titanium materials on the adhesion of fibroblasts. Biomaterials 1996;17:1399-403. [57] Thompson GJ, Puleo DA. Ti-6A1-4V ion solution inhibition of osteogenic cell phenotype as a function of differentiation time course in vitro. Biomaterials 1996;17:1949-54.
The B i o m a t e r i a l s S i l v e r J u b i l e e C o m p e n d i u m
161
Biomaterials ELSEVIER
Biomaterials 20 (1999) 2363-2376
Patterning proteins and cells using soft lithography Ravi S. Kane a, Shuichi Takayama a, Emanuele Ostun?, Donald E. Ingber b, George M. Whitesides a'* aDepartment of Chemistry and Chemical Biology, Harvard University, 12 Oxford Street, Cambridge, MA 02138, USA bDepartments of Pathology and Surgery, Children's Hospital and Harvard Medicals School, Boston, MA 02115, USA
Abstract
This review describes the pattering of proteins and cells using a non-photolithographic microfabrication technology, which we call 'soft lithography' because it consists of a set of related techniques, each of which uses stamps or channels fabricated in an elastomeric ('soft') material for pattern transfer. The review covers three soft lithographic techniques: microcontact printing, patterning using microfluidic channels, and laminar flow patterning. These soft lithographic techniques are inexpensive, are procedurally simple, and can be used to pattern a variety of planar and non-planar substrates. Their successful application does not require stringent regulation of the laboratory environment, and they can be used to pattern surfaces with delicate ligands. They provide control over both the surface chemistry and the cellular environment. We discuss both the procedures for patterning based on these soft lithographic techniques, and their applications in biosensor technology, in tissue engineering, and for fundamental studies in cell biology. 9 1999 Elsevier Science Ltd. All rights reserved. Keywords: Microfabrication; Microfluidics; Patterning; Cell biology; Tissue engineering; Self-assembled monolayers
1. Introduction
This review describes techniques for patterning the properties and structures of surfaces at the molecular level, and for using these patterns to control both the adsorption of proteins to these surfaces and the attachment of cells to them. The ability to generate patterns of proteins and cells on surfaces is important for biosensor technology [1-4] for tissue engineering [5], and for fundamental studies of cell biology [6-8]. The placement of biological ligands at well-defined locations on substrates is required for certain biological assays, for combinatorial screening, and for the fabrication of biosensors. Biosensors based on living cells [3,9-12] can also be used for environmental and chemical monitoring; accurate positioning of the cells used for sensing on these devices is critical for monitoring the status of the cells. Control over the positioning of cells is also important for cell-based screening, in which individual cells need to be accessed repeatedly to perturb them and to monitor their re-
* Corresponding author. Tel.: 1-617-495-9430;fax: 1-617-495-9857.
E-mail address:
[email protected] (G.M. White-
sides)
sponse. Tissue engineering may require that cells be placed in specific locations to create organized structures. Patterning techniques that control both the size and shape of the cell anchored to a surface, and the chemistry and topology of the substrate to which the cell is attached, are also extremely useful in understanding the influence of the cell-material interface on the behavior of cells [7,8,13,14]. Photolithography is the technique that has been used most extensively for patterning proteins and cells. For example, photolithography can be used to generate patterns by photoablating proteins preadsorbed to a silicon or glass surface [15], by immobilizing proteins on thiolterminated siloxane films that have been patterned by irradiation with UV light [ 16], and by covalently linking proteins to photosensitive groups [17]. Although photolithography is a technique that is highly developed for patterning, the high costs associated with photolithographic equipment, and the need for access to clean rooms, make this technique inconvenient for biologists. Photolithography is not well suited for introducing either specific chemical functionalities, or delicate ligands required for bio-specific adsorption, onto surfaces. Photolithography cannot be used to pattern non-planar substrates. While photolithography can be used to
0142-9612/99/S-see front matter 9 1999 Elsevier Science Ltd. All rights reserved. PII: S0 142-9612(99)001 65-9
162 2364
The Biomaterials Silver Jubilee Compendium R.S. Kane et al. /Biomaterials 20 (1999) 2363-2376
Table 1 Techniques that have been used to pattern proteins and cells (other than the soft lithographic techniques covered in this review) Technique
Reference
Photolithography and photochemistry Laser lithography Laser photoablation Laser vapor deposition Microwriting and micromachining 3D printing Ion implantation
[ 15,16,63-79] [80] [81,82] [83] [42] [84,85] [86,87]
produce patterns with features smaller than 1 jam, this resolution may be unnecessary for many applications of patterning in cell biology. We have developed a set of microfabrication techniques that is an alternative to photolithography for patterning surfaces used in biochemistry and biology. We call this set 'soft lithography' [ 18-23], because each of the techniques uses stamps or channels fabricated in an elastomeric ('soft') material for pattern transfer or modification. Soft lithographic techniques are not expensive, are procedurally simple, can be used to pattern a variety of different planar and non-planar substrates, and do not require stringent control over the laboratory environment for their successful application. This review will focus on the patterning of proteins and cells using soft lithographic techniques, in programs carried out in our group and by others. The patterning of proteins and cells by other techniques will not be discussed in this review, other than by reference (Table 1).
2. Key elements of soft lithography 2.1. Elastomeric stamps Soft lithographic methods use an elastomeric stamp or mold, prepared by casting the liquid prepolymer of an elastomer against a master that has a patterned relief structure (Fig. 1). Photolithography is used only for the fabrication of the masters. Most of the research based on soft lithography has used poly(dimethylsiloxane) (PDMS) as the elastomer. PDMS has several properties [24] that make it well suited for patterning proteins and cells. PDMS is biocompatible, permeable to gases, and can be used for cell culture. It is optically transparent down to about 300 nm. Because it is elastomeric, it can contact non-planar surfaces conformally. Its interfacial properties can be readily modified by treating the surface with plasma and subsequently forming self-assembled siloxane monolayers on the oxidized surface [25]. It is a durable elastomer. We have used the same PDMS stamp approximately 100 times over a period of several
Fig. 1. Schematic illustration of the procedure used to fabricate a PDMS stamp from a master having relief structures in photoresist on its surface.
months without noticeable degradation in its performance.
2.2. Masks and rapid prototyping An advantage of soft lithography as a method for patterning cells is that at the feature sizes required for this application--2-500 Hm--it is often possible to use procedures for making the initial patterns that are substantially more rapid and less expensive than those commonly used to make chrome masks for conventional photolithography. For the fabrication of masters having feature sizes greater than or equal to 20 Hm, the patterned mask that is used in the initial photolithographic step can be generated using high-resolution laser printing technology [26] that is inexpensively available commercially. The patterns used are generated using computer programs such as Freehand or AutoCAD, and are printed onto flexible transparencies. The masks can be made in a few hours, at a cost as low as $0.25 per square inch (at the time of writing the paper). For feature sizes greater than 10 Hm, the optical reduction of images printed onto transparencies generates patterns in microfiche [27]. Microfiche is then used as the photomask. For smaller feature sizes (between 20 and 2 jam), commercial laser writing can be used to fabricate PDMS stamps relatively inexpensively [28].
163
The Biomaterials Silver Jubilee Compendium
2365
R.S. Kane et al. / Biomaterials 20 (1999) 2363-2376
The capability to produce features larger than 20 ~tm rapidly and inexpensively has had a strong impact on the ability of researchers to prototype and produce small numbers of simple microstructures and microsystems. This capability, which we call 'rapid prototyping' [26,29,30], has greatly reduced the barriers to the use of lithographic techniques by researchers in the biological sciences.
x
I
x
X
r
3. Soft lithographic techniques 3.1. Microcontact printing
(
Since many of the studies involving the patterning of proteins and cells using microcontact printing have used self-assembled monolayers (SAMs) of alkanethiols on gold, we begin with a brief discussion of these SAMs. While SAMs can also be formed on silver, and are more ordered on silver than on gold, the cytotoxicity of the Ag + released from the silver films when they are exposed to air or other oxidants limits the use of these films in biological experiments involving living cells.
3.1.1. SAMs of alkanethiols on gold SAMs of alkanethiols on gold are formed by exposing a gold surface to a solution of, or to the vapors of, an alkanethiol (RSH) [31,32]. The nature of the gold-sulfur bond is not yet completely understood [33]. We adopt the view that the species present at the surface is a gold(I) thiolate. R-SH + Au(0), ~ RS-Au(I)Au(0)n_ 1 + 89 T.
X
(1)
Substrates used for experiments are prepared on glass cover slips or silicon wafers by evaporating a thin layer of titanium (1-5 nm) to promote the adhesion of gold, followed by a thin (10-200 nm) film of gold [34]. Gold substrates covered with SAMs are compatible with the conditions used for cell culture and they are not toxic to living cells [35]. While SAMs of alkanethiols on gold can be used in cell culture for periods of days, they desorb on heating above 70~ when irradiated with UV light in the presence of oxygen, or when exposed to atmospheric ozone. The substrates therefore need to be protected from intense light or excessive temperatures prior to and during experiments. The structure of the SAMs is shown in Fig. 2. The sulfur atoms of the alkanethiols coordinate to the gold surface, while the alkyl chains are close-packed and tilted by 30 ~ with respect to the surface normal [36]. The terminal functional group (X) of an co-substituted alkanethiol determines the properties of the interface between the SAM and the contacting liquid, and enables control over the interracial interactions at a molecular level. The exposure of a gold surface to a solution containing a mixture of alkanethiols forms mixed SAMs
Fig. 2. Schematic diagram of a self-assembled monolayer of an alkanethiolate on gold. The alkyl chains are oriented 30 ~ from the surface normal and packed with nearly crystalline densities; the interfacial properties of the film are largely determined by the chemical properties of the terminal group X.
(monolayers composed of a mixture of gold(I) thiolates), and allows the density of functional groups on the surface to be varied, although phase segregation might affect the surface properties of certain mixed SAMs.
3.1.2. Adsorption of proteins on SAMs Understanding and controlling both the specific and non-specific adsorption of proteins to surfaces is important for designing biomaterials and for fundamental studies of biology. The ability to control the nature and density of functional groups presented at the surface of SAMs makes them excellent model surfaces with which to study protein adsorption at surfaces. Sigal et al. [37] studied the non-specific adsorption of several proteins to surfaces presenting different functional groups such as alkyls, amides, esters, alcohols, and nitriles. The adsorption of proteins to uncharged SAMs showed a general correlation with the tendency of water to wet the surface (as determined by the contact angle of water on the SAM under cyclooctane), and on the size of the proteins. While the smaller proteins tested (ribonuclease A and lysozyme) adsorbed only on the least wettable surfaces tested, larger proteins such as pyruvate kinase and fibrinogen adsorbed to some extent on almost all of the surfaces tested. There were exceptions to the general trend of increased adsorption with decreased wettability. For example, fibrinogen adsorbed to a greater extent on SAMs presenting - C N groups than on SAMs presenting-CH3 groups, although the surface p r e s e n t i n g - C N groups has a greater wettability. Adsorption of proteins on hydrophobic surfaces was usually kinetically irreversible.
The Biomaterials Silver Jubilee Compendium
164 2366
R.S. Kane et al. / Biomaterials 20 (1999) 2363-2376
3.1.3. SAMs that resist the adsorption of proteins Surfaces that resist the adsorption of proteins are essential for certain applications in biomaterials and for research in biosurface chemistry. SAMs presenting oligomers of ethylene glycol, prepared using the alkanethiols H S ( C H z ) I 1 ( O C H z C H z ) , , O H ( E G n ) , resist the adsorption of proteins to surfaces [38,39]. The EG,terminated SAMs are not unique in their ability to resist the adsorption of proteins. SAMs presenting tripropylene sulfoxide groups also prevent the non-specific adsorption of proteins [40]. The existence of SAMs that resist the non-specific adsorption of proteins, when coupled with the ability to pattern SAMs using soft lithographic techniques, enables the facile patterning of proteins on surfaces. 3.1.4. Patterning SAMs by microcontact printing Microcontact printing [18,41] (pCP) is a technique that uses the relief pattern on the surface of an elastomeric PDMS stamp to form patterns of SAMs on the
surfaces of substrates (Fig. 3). The stamp is 'inked' with a solution of an alkanethiol in ethanol, dried, and brought into conformal contact with a gold substrate for 10-20 s. The alkanethiol is transferred to the gold substrate only in the regions where the PDMS stamp contacts the substrate. Subsequent exposure of the remaining bare gold substrate to a second alkanethiol generates a surface patterned into regions presenting different terminal groups.
3.1.5. Patterning proteins by microcontact printing Patterned SAMs generated by microcontact printing can be used to control the adsorption of proteins on surfaces. Lopez et al. [42] used microcontact printing to pattern gold surfaces into regions terminated in oligo(ethylene glycol) groups and methyl groups. Immersion of the patterned SAMs in solutions of the proteins such as fibronectin, fibrinogen, pyruvate kinase, streptavidin, and immunoglobulins resulted in adsorption of the proteins on the methyl-terminated regions. The
Fig. 3. Microcontact printing (gCP) and fabrication of contoured substrates using soft lithography. (A) A stamp is inked with an alkanethiol and placed on a gold surface; the pattern on the stamp is transferred to the gold by the formation of an SAM on the regions that contacted the substrate. The bare areas of the gold are exposed to a different alkanethiol to generate a surface patterned with an SAM that presents different chemical functionalities in different regions. (B) The PDMS stamp can also be used as a master to mold harder polymers and generate contoured surfaces. After evaporation of a layer of gold, these surfaces can be functionalized by pCP of one alkanethiol with a flat stamp. The grooves of the substrate can then be exposed to an alkanethiol presenting a different functional group to produce a contoured surface with patterned chemical reactivity.
165
The Biomaterials Silver Jubilee Compendium
2367
R.S. Kane et al. / Biomaterials 20 (1999) 2363-2376
and benzene sulfonamide printed on the reactive SAMs subsequently bound the proteins streptavidin and carbonic anhydrase, respectively. Microcontact printing has also been used to generate patterns on different substrates such as glass, silicon, and polystyrene, by using solutions of proteins as the ink [46,47]. A number of proteins printed using elastomeric stamps had activities indistinguishable from those of proteins adsorbed from solutions. Antibody gratings generated by microcontact printing the antibody onto silicon substrates were capable of detecting bacteria by a diffraction-based assay [48].
adsorption of proteins could be characterized by scanning electron microscopy, and the layers of adsorbed protein appeared to be homogeneous. Microcontact printing can also be used to form patterns of oligopeptide ligands by using self-assembling oligopeptides. Zhang et al. [43] have synthesized oligopeptides containing a cell adhesion motif at the N-terminus connected by an oligo(alanine) linker to a cysteine residue at the C-terminus. The thiol group of cysteine allowed the oligopeptides to form monolayers on gold-coated surfaces. A combination of microcontact printing and these self-assembling oligopeptide monolayers can be used to pattern gold surfaces into regions presenting cell adhesion motifs and oligo(ethylene glycol) groups that resist protein adsorption. Surfaces can also be patterned by microcontact printing ligands onto reactive SAMs (Fig. 4). Yan et al. [44] generated surfaces presenting interchain carboxylic acid anhydrides by treating SAMs that present terminal carboxylic acid groups with trifluoroacetic acid. Microcontact printing of ligands containing amino groups onto these activated SAMs resulted in the covalent attachment of the ligands to the SAM through amide bonds. Lahiri et al. [45] prepared mixed SAMs from mixtures of thiols presenting terminal tri(ethylene glycol) groups ( H S ( C H 2 ) I 1( O C H 2 CH2)3 OH) and terminal hexa(ethylene glycol)-CHzCOOH groups ( H S ( C H z ) l l ( O C H z C H 2 ) 6 OCHzCOOH). The carboxylic acid groups were converted to reactive pentafluorophenyl esters. Microcontact printing of amine-terminated ligands onto these activated surfaces also resulted in the covalent attachment of ligands to the SAM through amide bonds. Biotin
All
3.1.6. Patterning cells by microcontact printing Most mammalian cells are anchorage-dependentthey must adhere to and spread on a substrate in order to live. Extracellular matrix proteins such as fibronectin, vitronectin, and laminin promote the adhesion of anchorage-dependent cells to substrates. The ability to pattern SAMs by microcontact printing, and the resulting control over the adsorption of adhesive proteins, enables the patterning of cells on substrates [13]. Mrksich et al. [35] used microcontact printing to pattern gold or silver substrates into regions presenting oligo(ethylene glycol) groups and methyl groups. After coating the substrates with fibronectin, bovine capillary endothelial cells attached only to the methyl-terminated, fibronectin-coated regions of the patterned SAMs. The cells remained attached in the pattern defined by the underlying SAMs for 5-7 d. In addition to confining cells to specific regions of a substrate, microcontact printing can also be used to change the size and shape of cells (Figs. 5 and 6).
O
Au~S(C.~),,c%. (CF~CO)~OAu~S(C.~),,C,o R-,.~ .;~--S(0H2)15002H (02H5)3N
~L-S(CH2)150~O
.••
=- Au
S(CH2) 15CO2H S(CH2)15CONHR
a -n-ClsH37 t-C4H9
N c~,~-di-Ac.L.Lys.D_Ala.D_Ala n-C7F15
F.
B 9Au.~_S(CH2)1I(OCH2CH2)6OCH2COOH .~S(CH2)I I(OCH2CH2)3OH
R-NH2
Pentafluorophenol/EDC
Au[~-- S(CH2)11(OCH2CH2)6OCH2CONHR ,~--S(CH2)1I(OCH2CH2)3OH
F
A u ~ S(CH2)11(OCH2CH2)6OCH2COO~ F ~'- S(CH2)11(OCH2CH2)3OH / \ F F H~~]
R=
CF3COO"~+H3N ~ ' ~ ~ ~ I N
tSO2N H2
O
Fig. 4. Formation of mixed SAMs by performing reactions on pre-formed SAMs. Patterns can be generated by stamping the reactive ligand. (A) Alkylamines (in 1-methyl-2-pyrrolidinone) react with interchain carboxylic anhydride groups formed on SAMs of SH(CHz)IsCOOH after reaction with trifluoroacetic anhydride and triethylamine in anhydrous N, N-dimethylformamide. This procedure generates a surface that presents an equimolar mixture of the ligand R and carboxylic acid groups on the alkanethiolate molecules of the SAM. (B) SAMs are formed that comprise a mixture of carboxylic acid-capped hexa(ethylene glycol)- and tri (ethylene glycol)-terminated undecanethiolate. Pentafluorophenol and N-ethyl-N'(dimethylaminopropyl)carbodiimide (EDC) then react with carboxylic acid groups on this surface to form the pentafluorophenyl ester; this active ester subsequently reacts with alkyl amines carrying the ligand of interest, R.
166 2368
The Biomaterials Silver Jubilee Compendium R.S. Kane et al. / Biomaterials 20 (1999) 2363-2376
Fig. 6. A single bovine capillary endothelial cell conforming to a 40 ~m square adhesive island generated by microcontact printing using the procedures described in Fig. 5.
Fig. 5. (A)A gold surface was patterned into regions of hexadecanethiolate and undecanethiolate terminated with tri(ethylene glycol) by ~tCP. Fibronectin (light) adsorbed on the hydrophobic squares of hexadecanethiolate but not on the tri(ethyleneglycol) terminated alkanethiolate (dark). Patterned substrates were soaked in a solution of fibronectin (50 ~tg/ml in phosphate-buffered saline (PBS)) for 2 h, fixed using 20% paraformaldehyde (v/v) in PBS buffer and then immersed in a solution of anti-human fibronectin IgG (5 ~tg/ml) for 1 h followed by extensive rinsing. The substrates were then placed in contact with 100~tl of Texas Red| goat anti-rabbit IgG
(50 ~g/ml) for 1 h, followed by mounting in fluoromount-G (Southern Biotechnology Inc.). (B) Bovine capillary endothelial (BCE) cells patterned by culturing on a substrate presenting hydrophobic squares of varying sizes that were coated with fibronectin, prior to incubation with cells using the procedure described in (A).
Microcontact printing has also been used to pattern astroglial cells on silicon substrates [49]. Astroglial cells attached selectively to 50 lam wide adhesive bars generated by microcontact printing N-l[3-(trimethoxysilyl) propyl]diethylenetriamine (DETA) on silicon surfaces. Mrksich et al. developed a simple technique based on microcontact printing and micromolding to control the attachment of cells on contoured surfaces [14] (Fig. 3). Replica molding in P D M S molds having micron-scale relief patterns on their surfaces formed a contoured flim of polyurethane supported on a glass slide. After evaporating a thin film of gold onto these substrates, the raised plateaus of the contoured surface were derivatized with an SAM by stamping with an flat P D M S stamp; a different SAM was formed in the grooves by immersing the substrate in a solution of another alkanethiol. On modifying the raised plateaus with an SAM of hexadecanethiolate, and the grooves with an SAM terminated in oligo(ethylene glycol) groups, and coating the substrates with fibronectin, bovine capillary endothelial cells attached only to the methyl-terminated, fibronectincoated raised plateaus. A complementary procedure confined fibronectin adsorption and cell attachment to the grooves in the substrate.
3.1.7. Fundamental studies in cell biology using microcontact printed substrates 3.1.7.1. Effect of cell shape on cell function. Singhvi et al. [7] used the ability to control cell shape by microcontact
167
The Biomaterials Silver Jubilee Compendium R.S. Kane et al. /Biomaterials 20 (1999) 2363-2376
printing to explore the effect of cell shape on cell function. They plated primary rat hepatocytes on substrates consisting of square and rectangular islands of laminin surrounded by non-adhesive regions. Cells attached preferentially to the laminin-coated islands, and in most cases conformed to the shape of the island. The size and shape of the cells could be controlled by changing the size of the adhesive islands, without changing the density of the adhesive protein, laminin. Singhvi et al. [-7] found that the synthesis of DNA was highest on unpatterned surfaces where cells could spread without restriction, and a decrease in the size of the adhesive islands resulted in a progressive reduction in growth. For the smallest islands (less than 1600 ~tm2), less than 3% of the adherent cells entered S (DNA synthesis) phase. The differentiated function of hepatocytes cultured on islands of different sizes was also assessed by measuring the secretion of albumin in the culture supernatant. Hepatocytes cultured on unpatterned substrata rapidly lost the ability to secrete high levels of albumin. Albumin secretion rates increased as the size of the adhesive island was decreased. This study demonstrated that the modulation of cell shape provided control over cell growth and protein secretion independent of the density of the adhesive protein laminin.
3.1.7.2. Effect of cell shape on cell life and cell death. Chen et al. [-8] used micropatterned substrates to control the shape of human and bovine capillary endothelial cells. Cells were shifted from growth to apoptosis by using substrates that contained extracellular matrix-coated adhesive islands of decreasing size. These results are compatible with two hypotheses: (i) that growth increases with the area of the surface that is in adhesive contact with the cell, and (ii) that growth depends on the extent of cell spreading rather than the area of adhesive contact. To discriminate between these two hypotheses, Chen et al. [-8] varied the extent of cell spreading while keeping the total cell-matrix contact area constant by culturing cells on substrates that presented islands with diameters of 20, 5, and 3 lam, separated by 40, 10, and 6 ~tm respectively. They found that the extent of spreading (the projected surface area of the cell) and not the area of the adhesive contact controlled cell life and death (Fig. 7). Cell shape was found to determine cellular choice between division and apoptosis regardless of the type of matrix protein used to mediate adhesion. Because of the power of the micropatterning technique, it was also possible to define how shape regulates the cell cycle machinery in these cells and thus, finally, to translate the long recognized phenomenon of shape-dependent control of growth into specific molecular terms [50].
3.1.7.3. Effect of cell shape on cell differentiation. Dike et al. [88] found that bovine capillary endothelial cells cultured on fibronectin-coated, micropatterned substra-
2369
Fig. 7. Graphs of the projected cell area, extracellular matrix (ECM) contact area, growth index, and the apoptotic index (percentage of cells that enter apoptosis) for cells cultured on circular islands having diameters of 20, 5, and 3 ~m, separated by 40, 10, and 6 ~tm, respectively. The results illustrate that the apoptotic index depends on the projected area of the cell and not on the ECM contact area.
tes containing 10 ~tm wide lines formed extensive cell-cell contacts while cell spreading was restricted to approximately 1000 ~tm2. Within 72 h, these cells shut off growth and apoptosis programs and underwent differentiation, as indicated by the formation of capillary tube-like structures containing a central lumen (Fig. 8). Cells cultured on wider (30 ~tm) lines formed cell-cell contacts, but these cells continued to proliferate and did not form tubes. The use of substrates prepared by microcontact printing revealed that bovine capillary endothelial (BCE) cells could be switched between the three genetic programs of growth, apoptosis, and differentiation, by altering the geometry of spreading.
3.1.7. 4. Effect of surface chemistry on lamellipod extension during chemotaxis. Bailly et al. [51] used micropatterned substrates in studies of the regulation of protrusion shape during chemotactic responses of mammalian carcinoma cells. Since tumor cell motility and protrusive activity are generally correlated with tumor metastatic potentials, such studies increase our fundamental understanding of oncogenic progression. Bailly et al. [51] plated rat mammary carcinoma cells on gold-coated glass coverslips having 10 ~tm wide adhesive lanes. The
168 2370
The Biomaterials Silver Jubilee Compendium R.S. Kane et al. / Biomaterials 20 (1999) 2363-2376
Fig. 8. Capillary tube formation by bovine capillary endothelial cells cultured on 10 ~tm lines. A larger unpatterned region was included as an internal control. Pattern generation and cell attachment were performed according to procedures described in Fig. 5.
cells attached only to the adhesive lanes. On stimulating the cells with epidermal growth factor (EGF), the cells could extend lamellipods laterally, over the non-adhesive part of the substrate. These results showed that lamellipod extension could occur independent of any contact with the substratum. Contact formation was, however, necessary for stabilizing the protrusion.
3.2. Patterning using microfluidic channels Microcontact printing relies on the transfer of material (thiols, proteins, etc.) from an 'inked' elastomeric stamp to select regions of a substrate. Patterning can also be carried out by restricting fluid flow to desired regions of a substrate. Kim et al. [21] developed a technique called micromolding in capillaries (MIMIC) for fabricating three-dimensional structures by allowing solutions to flow into microfluidic channels formed by bringing
a PDMS mold into conformal contact with a substrate. MIMIC is not restricted to patterning curable prepolymers, and has also been used to pattern a wide variety of materials such as precursor polymers to glassy carbon or ceramics [52,53], sol-gel materials [54], inorganic salts [55], polymer beads [56], and colloidal particles [55]. Delamarche et al. [57,58] extended MIMIC to the patterning of biological molecules such as immunoglobulins. They patterned biomolecules with submicron resolution on a variety of substrates including gold, glass, and polystyrene, by allowing solutions of the biomolecules to flow through microfluidic channels. Only microliters of reagent were required to cover square millimeter-sized areas. The technique enables simultaneous and highly localized immunoassays for the detection of different IgGs. Patel et al. [59] developed a method of generating micron-scale patterns of any biotinylated ligand on the surface of a biodegradable polymer using microfluidic channels. A biotin molecule was introduced into the end group of the poly(ethylene glycol) (PEG) block of a polylactide-poly(ethylene glycol) copolymer to produce the biodegradable polymer, PLA-PEG-biotin. Films of PLA-PEG-biotin were patterned by allowing solutions of avidin to flow over them through elastomeric microfluidic channels. The avidin moieties bound to the biotin groups on the surface, and served as a bridge between the biotinylated polymer and biotinylated ligands. Patel et al. [59] used their method to achieve spatial control over the adhesion and spreading of bovine aortic endothelial cells and PC12 nerve cells on films of PLA-PEG-biotin. Neurite extension on the polymer surface was found to be directed by patterned features composed of peptides containing the IKVAV sequence. Folch et al. [60,61-] also used microfluidic channels to produce patterns of cells on biocompatible substrates. They created protein templates on surfaces by the adsorption of proteins from solutions that were passed through elastomeric channels. Micropatterns of collagen or fibronectin were used to cause cells to adhere selectively on various biomedical polymers and on heterogeneous or microtextured substrates. On removing the elastomeric stamp, the bare substrate areas could be seeded with more adhesive cell types such as fibroblasts, thereby producing micropatterned co-cultures. By allowing different cell suspensions to flow through different microchannels, patterns of cells could be generated on surfaces (Fig. 9).
3.3. Laminar flow patterning Laminar flow patterning, a method recently developed by our group [62], adds a new capability to the patterning of microfluidic channels. Microfluidic systems have distinctive properties as a result of their small
169
The Biomaterials Silver Jubilee Compendium R.S. Kane et al. /Biomaterials 20 (1999) 2363-2376
Fig. 9. Illustration of the procedure used to pattern proteins and cells using microfluidic channels.
dimensions. One notable feature is that the flow of liquids in capillaries often has a low Reynolds number (Re) and is laminar (Re is a dimensionless parameter relating the ratio of inertial to viscous forces in a specific fluid flow
2371
configuration, and is a measure of the tendency of the liquid to develop turbulence). Operating in laminar flow conditions allows two or more layers of fluid to flow next to each other without any mixing other than that taking place by diffusion of their constituent molecular and particulate components across the interface. This ability to generate and sustain parallel streams of different solutions in capillaries provides a unique opportunity to pattern cells and their environment--that is, the molecular structure of the surface to which the cells are attached, the nature and position of other cells in their vicinity, and the composition of the fluid medium surrounding them. Although we have discussed methods that allow the patterning of surfaces and cells, laminar flow patterning is the only method that allows patterning of the culture medium itself in a highly controlled fashion. A typical setup for laminar flow patterning experiments, together with images from representative patterning experiments, is outlined in Fig. 10. A network of capillaries with multiple inlets that converge into a single main channel is fabricated by bringing a PDMS membrane with the pattern of channels molded into its surface into contact with the flat surface of a Petri dish. By allowing different patterning components to flow from the inlets, patterns of parallel stripes are created in the main channel (Fig. 10A). Micrographs show the patterns
Fig. 10. Schematic representation of a laminar flow patterning experiment. (A) Top view of the capillary network. A poly(dimethylsiloxane) (PDMS) membrane containing micron-sized channels molded in its surface was placed on the flat surface of a Petri dish to form a network of capillaries. Micrographs were obtained for the area of the capillary system at which the inlet channels converge into a single main channel. (B) Two different cell types patterned next to each other. Chick erythrocytes and Escherichia coli were deposited selectively in their designated lanes by patterned flow of cell suspensions. Adherent cells were visualized with a fluorescent nucleic acid stain (Syto 9). (C) Pattern of selectively stained BCE cells. A suspension of BCE cells was introduced into the capillary network (pretreated with fibronectin) and allowed to attach and spread. After removing non-adherent cells by washing with media, Syto 9 and media were allowed to flow from the designated inlets. (D) Patterned detachment of BCE cells by treatment with trypsin/EDTA. Cells were allowed to adhere and spread in a fibronectin-treated capillary network, and non-adherent cells removed by washing. Trypsin/EDTA and media were allowed to flow from the designated inlets. Pictures B and C are fluorescence micrographs taken from the top through PDMS. Picture D is a phase contrast image observed by an inverted microscope looking through the polystyrene Petri dish. White dotted lines identify channels not visible with fluorescence microscopy.
Table 2 Examples of patterning proteins and cells using soft lithographic techniques
b,.)
Technique
Substrate
Feature
Initial component patterned
Proteins patterned
Microcontact printing
Gold
5 lam lines
Patterned adsorption of ribonuclease A, pyruvate kinase, fibrinogen, fibronectin, streptavidin, and immunoglobulins
Adsorbed proteins were visualized [42] using scanning electron microscopy
SAM of interchain carboxylic acid anhydride on gold
10 lam squares
Glass, polystyrene, silicon, various SAMs on gold
Arbitrary patterns of 1 gm and larger
Alkane thiols terminated in various functional groups including oligoethylene glycol n-Hexadecylamine, cystamine, 3-amino- 1propanesulfonic acid and other amine containing molecules Proteins
Silicon
10 gm stripes separated by 30 gm
Antibodies
Anti-E. coli O157"H7 antibodies
Mixed SAMs of terminal oligo(ethylene glycol) and oligoethylene carboxylate thiols on gold Gold
5-50 ~tm squares
Amine compounds (derivatives of biotin or benzene sulfonamide)
Anti-biotin antibodies or streptavidin bound to regions patterned with biotin
Hexadecane thiol, hexa(ethylene glycol) terminated alkane thiol
Laminin adsorbed to patterned hexadecane thiol SAMs
Gold
Thin gold film ( < 12 nm) on polyurethane
2-80 gm rectangles
60 gm lines separated by 120 gm
25-50 gm ridges and grooves
Hexadecane thiol, tri(propylene sulfoxide) terminated alkane thiol Hexadecane thiol, tri(ethylene glycol) terminated alkane thiol
Gold
5-40 gm squares and circles
Hexadecane thiol, tri(ethylene glycol) terminated alkane thiol
Gold and silicon
Circles and squares of 5-80 gm
Hexadecane thiol, tri(ethylene glycol) terminated alkane thiol
Antibodies, phosphatase, cytochrome c, bovine serum albumin, streptavidin, protein A, proteinase K, peroxidase, chymotrypsin, NgCAM (cell adhesion molecule)
Fibronectin adsorbed to patterned hexadecane thiol SAM
Fibronectin adsorbed to patterned hexadecane thiol SAMs
Fibronectin, vitronectin, collagen, anti-integrin/31 antibody, or anti-integrin ~v/~3 antibody adsorbed to patterned hexadecane thiol SAMs Fibronectin adsorbed to patterned hexadecane thiol SAMs
"--.I
Ref.
Effect observed (cells patterned)
Gold nanoparticles (~ 20 nm) adhered selectively to regions patterned with cystamine
[44]
The amount of protein adsorbed onto surfaces is similar to adsorption from solutions, however, the non-solution environment inevitable with stamping sometimes compromises protein function Diffraction-based detection of E. coli O157 :H7 cells using an antibody grating created by microcontact printing Patterned binding of streptavidin. Bound streptavidin can be used to capture any biotinylated ligand or protein in so-called 'sandwich' experiments Primary rat hepatocytes adhere and spread only on laminincoated islands. Cell shape controls cell growth and function BCE cells were confined to the hexadecane thiol SAM regions for 1-2 d then started to spread into other areas Patterned adsorption of fibronectin. Cells are confined to tri(ethylene glycol) terminated alkane thiol SAM regions for at least 5 d Extent of cell spreading and not the area of adhesive contact controls life and death of bovine capillary endothelial (BCE)cells Cell cycle progression of BCE cells were controlled by cell shape and cytoskeletal tension
[47]
[48]
[45]
2. t,,a
[7]
t~ ta, a
t,,~~
"q
[40]
t~ c,z t.,~.
[14]
t.,~.
t~ [8]
[50]
Gold
10 jam lines
Hexadecane thiol, hexa(ethylene glycol) terminated alkane thiol
Silicon
Features of 1 jam and larger
Octadecyltrichlorosilane (OTS), N(3-(trimethoxysilyl)propyl)diethylenetriamine (DETA), polylysine
Gold
Microfluidic networks
Laminar flow patterning
Squares and lines of 20 jam and larger
Peptides containing cell adhesion motifs and cysteine, hexa(ethylene glycol) terminated alkane thiol Hexadecane thiol, tri(ethylene glycol) terminated alkane thiol
Vitronectin adsorbed to patterned hexadecane thiol SAM
Gold
10 or 30 jam width lines and 5 or 10 jam squares
Gold, glass, silicon, polystyrene
Lines with widths of 3 jam and bigger
Immunoglobulins
Immunoglobulins and bovine serum albumin (BSA)
Polylactidebiotinylated poly(ethylene glycol) block copolymer Various polymers
12-70 jam lines
Avidin
Biotinylated ligands bind to regions with avidin attached to it
20 jam lines
Protein
Fibronectin or collagen
Tissue culture-grade plastic
200 jam channels separated by 200 jam
Cells
Polystyrene, glass, gold, silicon
Lines with widths of 10-100 jam
Proteins, cells, and media
Fibronectin adsorbed to patterned hexadecane thiol SAMs
Wheat germ agglutinin, BSA, and fibronectin were patterned onto the substrate Trypsin was a component of patterned media flow.
Chemotactic response of MTLn3 metastatic rat mammary adenocarcinoma cells. Lamellipod extension is independent of contact with the substratum LRM 55 cells (astroglial cell line) selectively attached to DETA patterned surfaces in the presence of serum containing media. Rat hippocampal neurons attach to polylysine patterned regions Human epidermoid carcinoma A431 cells, primary human embryonic kidney 293 cells, bovine aorta fibroblasts
[51]
[49]
t,,~~
[43]
BCE cells can be switched between [88] growth, apoptosis, and differentiation by altering the geometry of spreading [57,58] Simultaneous delivery of functionally distinct proteins onto targeted regions of a surface Biotinylated RGD peptide pro[59] motes bovine aortic endothelial cells attachment and IKVAV peptides promote PC12 nerve cells attachment Selective attachment of hepato[61] cytes and 3T3-J2 fibroblasts onto protein patterned areas Fluorescently labeled 3T3-J2 [60] fibroblast cells attached selectively in the areas over which they flowed. Spatially separated co-cultures of two different cell types were created E. coli RB 128 cells, chick [62] erythrocytes, and BCE cells. Patterning of substrate, cell position, and media.Media can be patterned to flow selectively over half of a cell
t,,~~
b.a
t.,a tao
to, a ..q
"-,1
172
2374
The Biomaterials Silver Jubilee C o m p e n d i u m R.S. Kane et al. /Biomaterials 20 (1999) 2363-2376
created at the junction where the inlet channels converge into a single main channel. Fig. 10B shows patterning where cells (E. coli and erythrocytes) are used directly as the patterning component. The cells adhere only to those areas over which they were allowed to flow. Fig. 10C shows an example of using patterned culture media to selectively deliver chemicals to cells. In this experiment, BCE cells covered the entire bottom face of the capillary; this fluorescence micrograph shows selective staining only of those cells over which a medium containing a fluorescent dye was allowed to flow. Fig. 10D demonstrates patterned release of attached BCE cells using trypsin/ethylenediaminetetraacetic acid (EDTA). Digestion of fibronectin on the channel surface, and sequestering of calcium by EDTA, causes cells to detach and contract. When a solution of trypsin/EDTA flowed over only a portion of a cell, the treated part of the cell detached and contracted; the untreated part remained spread (for example, see inside dotted square in Fig. 10D). Since no physical barriers are required to separate the different liquid streams, different liquids can flow over different portions of a single cell. Laminar flow patterning has some features that makes it complementary to other patterning techniques used for biological applications. It takes advantage of the easily generated, multiphase laminar flows to pattern fluids and to deliver components for patterning. This mild delivery method allows the use of cells themselves as the patterning component. Laminar flow patterning is experimentally simple. Multiple-component patterns can be made in a short sequence of steps, without the need for multiple stages of pattern transfer with registration required by other methods. Neither patterning of the growth medium itself, nor patterning over delicate structures such as mammalian cells, is possible by other techniques.
4. Conclusions The soft lithographic techniques described in this review are a powerful set of tools for controlling the cell-material interface. These techniques offer several advantages over conventional photolithographic techniques. They are inexpensive, and are accessible to chemists and biologists. They allow the patterning of delicate ligands on a variety of substrates, including biocompatible substrates. They can be used to pattern non-planar substrates and to make three-dimensional microstructures. Soft lithographic techniques can be used to control not only the surface chemistry, but also the cellular environment. The soft lithographic techniques complement each other well. Microcontact printing provides the highest resolution; it has been used to make features smaller than 1 ~tm. It also provides the greatest flexibility in the shapes of the pattern generated. It is the best technique for
controlling the chemistry of the surface at a molecular level. It is most useful when one only needs to pattern two types of ligands. Patterning multiple ligands requires sequential registered stamping steps with different inks, and is more complicated. Microfluidic channels are well suited for patterning delicate objects like proteins and cells on a variety of substrates. They are useful when multiple ligands need to be patterned, although the range of possible patterns is limited. They can also be used to pattern multiple cell types. Laminar flow patterning has greatly enhanced the capabilities of patterning using microfluidic channels. The technique enables the patterning of multiple ligands or cells without the problem of registration. It is best suited for patterning parallel lines, although more complicated patterns can be created with additional steps. The ability to pattern over the surface of delicate materials like mammalian cells is unique to this technique (Table 2). The use of substrates patterned using microcontact printing has already led to fundamental insights into the effects of cell shape on cell function. The other soft lithographic techniques, used either by themselves, or in conjunction with microcontact printing, enhance our ability to control the cellular environment, and should help increase our understanding of fundamental cell biology. The capabilities described in this review should pave the way for engineering cells and tissues for use in biosensors and other 'hybrid' systems that combine living and non-living components.
Acknowledgements This work was supported by NIH GM30367, NIH HL 57669, NIH CA 55833, NSF ECS 9729405, NSF DMR98-09363 MRSEC, and by D A R P A / S P A W A R (Space and Naval Warfare Systems Center San Diego--the content of the information does not necessarily reflect the position or the policy of the Government, and no official endorsement should be inferred). S.T. is a Leukemia Society of America Fellow and thanks the society for a fellowship.
References [1] Thomas CA, Springer PA, Loeb GE, Berwald-Netter Y, Okun LM. A miniature microelectrode array to monitor the bioelectric activity of cultured cells. Exp Cell Res 1972;74:61-6. [2] Gross GW, Rhoades BK, Azzazy HME, Wu MC. The use of neuronal networks on multielectrode arrays as biosensors. Biosens Bioelec 1995;10:553-7. [3] Jung DR, et al. Cell-based microelectrode array characterized by imaging X-ray photoelectron spectroscopy, scanning electron microscopy, impedance measurements, and extracellular recordings. J gac Sci Technol A 1998;16:1183-8.
173
The Biomaterials Silver Jubilee Compendium R.S. Kane et al. /Biomaterials 20 (1999) 2363-2376
[-4] Mrksich M, Whitesides GM. Patterning self-assembled monolayers using microcontact printing: a new technology for biosensors? Trends Biotech 1995;13:228-35. [5] Merritt MV, Mrksich M, Whitesides GM, editors. Using selfassembled monolayers to study the interactions of man-made materials with proteins. Principles of tissue engineering. Austin: R.G. Landes Co., 1997. p. 211-23. [6] Mrksich M, Whitesides GM. Using self-assembled monolayers to understand the interactions of man-made surfaces with proteins and cells. Ann Rev Biophys Biomol Struct 1996;25:55-78. [-7] Singhvi R, et al. Engineering cell shape and function. Science 1994;264:696-8. [8] Chen CS, Mrksich M, Huang S, Whitesides GM, Ingber DE. Geometric control of cell life and death. Science 1997;276:1425-8. [-9] Pacrazio JJ, et al. Description and demonstration of a CMOS amplifier-based-system with measurement and stimulation capability for bioelectrical signal transduction. Biosens Bioelec 1998; 13:971-9. [10] Kovacs GT, Storment CW, Rosen JM. Regeneration microelectrode array for peripheral nerve recording and stimulation. IEEE Trans Biomed Eng 1992;39:893-902. [-11] Kewley DT, et al. Plasma-etched neural probes. Sens Actuators A 1997;58:27-35. [12] Borkholder DA, Bao J, Maluf NI, Perl ER, Kovacs GTA. Microelectrode arrays for stimulation of neural slice preparations. J Neurosci Methods 1997;77:61-6. [13] L6pez GP, et al. Convenient methods for patterning the adhesion of mammalian cells to surfaces using self-assembled monolayers of alkanethiolates on gold. J Am Chem Soc 1993;115:5877-8. [14] Mrksich M, et al. Controlling cell attachment on contoured surfaces with self-assembled monolayers of alkanethiolates on gold. Proc Natl Acad Sci USA 1996;93:10775-8. [15] Hammarback JA, Palm SL, Furch LT, Letourneau PC. Guidenace of neurite outgrowth by pathways of substratum-adsorbed laminin. J Neurosci Res 1985;13:213-20. [16] Bhatia SK, Hickman JJ, Ligler FS. New approach to producing patterned biomolecular assemblies. J Am Chem Soc 1992; 114:4432-3. [17] Matsuda T, Inoue K, Tani N. Control of cell arrangement. US Patent 1993, 5,202,227. [18] Kumar A, Whitesides GM. Features of gold having micrometer to centimeter dimensions can be formed through a combination of stamping with an elastomeric stamp and an alkanethiol ink followed by chemical etching. Appl Phys Lett 1993;63:2002-4. [19] Xia Y, et al. Complex optical surfaces by replica molding against elastomeric masters. Science 1996;273:347-9. [20] Zhao X-M, Xia Y, Whitesides GM. Fabrication of three-dimensional micro-structures: microtransfer molding. Adv Mater 1996; 8:837-40. [21] Kim E, Xia Y, Whitesides GM. Polymer microstructures formed by moulding in capillaries. Nature 1995;376:581-4. [22] Xia Y, Whitesides GM. Soft lithography. Angew Chem Int Ed Engl 1998;37:550-75. [23] Xia Y, Whitesides GM. Soft lithography. Ann Rev Mater Sci 1998;28:153-84. [24] Clarson SJ, Semlylen JA. Siloxane polymers. Englewood Cliffs, NJ: Prentice Hall, 1993. [25] Ferguson GS, Chaudhury MK, Biebuyck HA, Whitesides GM. Monolayers on disordered substrates: self-assembly of alkyltrichlorosilanes on surface-modified polyethylene and poly(dimethylsiloxane). Macromolecules 1993;26:5870-5. [26] Qin D, Xia Y, Whitesides GM. Rapid prototyping of complex structures with feature sizes larger than 201am. Adv Mater 1996;8:917-9. [27] Deng T, Tien J, Xu B, Whitesides GM. Using patterns in microfiche as photomasks in 10 gm scale microfabrication. Langmuir 1999, submitted for publication.
2375
[28] Grzybowski BA, Haag R, Bowden N, Whitesides GM. Generation of micrometer-sized patterns for microanalytical applications using a laser direct-Write method and microcontact printing. Anal Chem 1998;70:4645-52. [29] Duffy DC, McDonald JC, Schueller OJA, Whitesides GM. Rapid prototyping of microfluidic systems in poly(dimethylsiloxane). Anal Chem 1998;70:4974-84. [30] Duffy DC, Schueller OJA, Brittain ST, Whitesides GM. Rapid prototyping of microfluidic switches in poly(dimethyl siloxane) and their actuation by electroosmotic flow. J Micromech Microeng 1999, submitted for publication. [31] Dubois LH, Nuzzo RG. Ann Rev Phys Chem 1992;43:437-63. [32] Whitesides GM, Gorman CB, editors. Self-assembled monolayers: models for organic surface chemistry. Handbook of surface imaging and visualization. Boca Raton: CRC Press, 1995. p. 713-33. [33] Fenter P, Eberhardt A, Eisenberger P. Science 1994;266:1216-8. [34] DiMilla P, et al. Wetting and protein adsorption of self-assembled monolayers of alkanethiolates supported on transparent films of gold. J Am Chem Soc 1994;116:2225-6. [35] Mrksich M, Dike LE, Tien J, Ingber DE, Whitesides GM. Using microcontact printing to pattern the attachment of mammalian cells to self-assembled monolayers of alkanethiolates on transparent films of gold and silver. Exp Cell Res 1997;235:305-13. [36] Laibinis PL, et al. A comparison of the structures and wetting properties of self-assembled monolayers of n-alkanethiols on the coinage metal surfaces. Cu, Ag, Au. J Am Chem Soc 1991; 113:7152-67. [37] Sigal GB, Mrksich M, Whitesides GM. Effect of surface wettability on the adsorption of proteins and detergents. J Am Chem Soc 1998;120:3464-73. [38] Prime KL, Whitesides GM. Adsorption of proteins onto surfaces containing end-attached oligo(ethylene oxide): a model system using self-assembled monolayers. J Am Chem Soc 1993; 115:10714-21. [39] Prime KL, Whitesides GM. Self-assembled organic monolayers: model systems for studying adsorption of proteins at surfaces. Science 1991;252:1164-7. [40] Deng L, Mrksich M, Whitesides GM. Self-assembled monolayers of alkanethiolates presenting tri(propylenesulfoxide) groups resist the adsorption of protein. J Am Chem Soc 1996;118:5136-7. [41] Jackman R, Wilbur J, Whitesides GM. Fabrication of submicron features on curved substrates by microcontact printing. Science 1995;269:664-6. [42] L6pez GP, Biebuyck HA, Hiirter R, Kumar A, Whitesides GM. Fabrication and imaging of two-dimensional patterns of proteins adsorbed on self-assembled monolayers by scanning electron microscopy. J Am Chem Soc 1993;115:10774-81. [43] Zhang S, et al. Biological surface engineering: a simple system for cell pattern formation. Biomaterials 1998;20:1213-20. [44] Yan L, Zhao X-M, Whitesides GM. Patterning a preformed, reactive SAM using microcontact printing. J Am Chem Soc 1998;120:6179-80. [45] Lahiri J, Ostuni E, Whitesides GM. Patterning ligands on reactive SAMs by microcontact printing. Langmuir 1999;15:2055-60. [46] James CD, et al. Patterned protein layers on solid substrates by thin stamp microcontact printing. Langmuir 1998;14:741-4. [47] Benard A, et al. Printing patterns of proteins. Langmuir 1998;14:2225-9. [48] St John PM, et al. Diffraction-based cell detection using a microcontact printed antibody grating. Anal Chem 1998;70:1108-11. [49] St John PM, et al. Preferential glial cell attachment to microcontact printed surfaces. J Neurosci Methods 1997;75:171-7. [50] Huang S, Chen CS, Ingber DE. Control of cyclin D1, p27 KipI, and cell cycle progression in human capillary endothelial cells by cell shape and cytoskeletal tension. Mol Bio Cell 1998; 9:3179-93.
174 2376
The Biomaterials Silver Jubilee Compendium R.S. Kane et al. / Biomaterials 20 (1999) 2363-2376
[51] Bailly M, Yan L, Whitesides GM, Condeelis JS, Segall JE. Regulation of protrusion shape and adhesion to the substratum during chemotactic responses of mammalian carcinoma cells. Exp Cell Res 1998;241:285-99. [52] Schueller OJA, Brittain ST, Whitesides GM. Fabrication of glassy carbon microstructures by pyrolysis of microfabricated polymeric precursors. Adv Mater 1997;9:477-80. [53] Schueller OJA, Brittain ST, Marzolin C, Whitesides GM. Fabrication and characterization of glassy carbon MEMS. Chem Mater 1997;9:1399-406. [54] Trau M, et al. Microscopic patterning of oriented mesocopic silica through guided growth. Nature 1997;390:674-6. [55] Xia Y, Kim E, Whitesides G M. Micromolding of polymers in capillaries: applications in microfabrication. Chem Mater 1996; 8:1558-67. [56] Kim E, Xia Y, Whitesides GM. Two- and three-dimensional crystallization of polymeric microspheres by micromolding in capillaries. Adv Mater 1996;8:245-7. [57] Delamarche E, Bernard A, Schmid H, Michel B, Biebuyck H. Patterned delivery of immunoglobulins to surfaces using microfluidic networks. Science 1997;276:779-81. [58] Delamarche E, et al. Microfluidic networks for chemical patterning of substrates: design and application to bioassays. J Am Chem Soc 1998;120:500-8. [59] Patel N, et al. Spatially controlled cell engineering on biodegradable polymer surfaces. FASEB J 1998;12:1447-54. [60] Folch A, Ayon A, Hurtado O, Schmidt MA, Toner M. Molding of deep polydimethylsiloxane microstructures for microfluidics and biological applications. J Biomech Eng 1999;121:28-34. [61] Folch A, Toner M. Cellular micropatterns on biocompatible materials. Biotech Prog 1998;14:388-92. [62] Takayama S, et al. Patterning cells and their environments using multiple laminar fluid flows in capillary networks. Proc Natl Acad Sci USA 1999;96:5545-8. [63] Britland S, et al. Micropatterning proteins and synthetic peptides on solid supports: a novel application for microelectronics fabrication technology. Biotech Prog 1992;8:155-60. [64] Britland S, Clark P, Moores G. Micropatterned substratum adhesiveness: a model for morphogenic cues controlling cell behavior. Exp Cell Res 1992;198:124-9. [65] Healy KE, et al. Kinetics of bone cell organization and mineralization on materials with patterned surface chemistry. Biomaterials 1996;17:195-208. [66] Lom B, Healy KE, Hockberger PE. A versatile technique for patterning biomolecules onto glass coverslips. J Neurosci Methods 1993;50:385-97. [67] Pritchard DJ, Morgan H, Cooper JM. Patterning and regeneration of surfaces with antibodies. Anal Chem 1995;67:3605-7. [68] Mooney JF, et al. Patterning of functional antibodies and other proteins by photolithography of silane monolayers. Proc Natl Acad Sci USA 1996;93:12287-91. [69] Kleinfeld D, Kahler KH, Hockberger PE. Controlled outgrowth of dissociated neurons on patterned substrates. J Neurosci 1988;8:4098-120.
[70] Flounders AW, Brandon DL, Bates AH. Patterning of immobilized antibody layers via photolithography and oxygen plasma exposure. Biosens Bioelec 1997;12:447-56. [71] Hengsakul M, Cass AEG. Protein patterning with a photoactivatable derivative of biotin. Bioconj Chem 1996;7:249-54. [72] Ravenscroft MS, et al. Developmental neurobiology implications from fabrication and analysis of hippocampal neuronal networks on patterned silane-modified surfaces. J Am Chem Soc 1998; 120:12169-77. [73] Stenger DA, et al. Coplanar molecular assemblies of amino- and perfluorinated alkylsilanes: characterization and geometric definition of mammalian cell adhesion and growth. J Am Chem Soc 1992;114:8435-42. [74] Vargo TG, et al. Monolayer chemical lithography and characterization of fluoropolymer films. Langmuir 1992;8:130-4. [75] Bhatia SK, et al. Fabrication of surfaces resistant to protein adsorption and application to two-dimensional protein patterning. Anal Biochem 1993;208:197-205. [76] Bhatia SN, Yarmush ML, Toner M. Controlling cell interactions by micropatterning in co-cultures: hepatocytes and 3T3 fibroblasts. J Biomed Mater Res 1997;34:189-99. [77] Bhatia SN, Balis UJ, Yarmush ML, Toner M. Probing heterotypic cell interactions: hepatocyte function in microfabricated co-cultures. J Biomater Sci--Polym Ed 1998;9:1137-60. [78] Fodor S, et al. Light-directed spatially addressable parallel chemical synthesis. Science 1991;251:767-72. [79] Dulcey CS, et al. Deep UV photochemistry of chemisorbed monolayers: patterned coplanar molecular assemblies. Science 1991; 252:551-4. [80] Shivshankar GV, Libchaber A. Biomolecular recognition using submicron laser lithography. Appl Phys Lett 1998;73:417-9. [81] Schwarz A, et al. Micropatterning of biomolecules on polymer substrates. Langmuir 1998;14:5526-31. [82] Vaidya R, et al. Computer-controlled laser ablation: a convenient and versatile tool for micropatterning bifunctional synthetic surfaces for applications in biosensing and tissue engineering. Biotech Prog 1998;14:371-7. [83] Morales P, et al. A laser assisted deposition technique suitable for the fabrication of biosensors and molecular electronic devices. Biosens Bioelec 1995;10:847-52. [84] Park A, Wu B, Griffith L. Integration of surface modification and 3D fabrication techniques to prepare patterned poly(L-lactide) substrates allowing regionally selective cell adhesion. J Biomater Sci--Polym Ed 1998;9:89-110. [85] Kim SS, et al. Survival and function of hepatocytes on a novel three-dimensional synthetic biodegradable polymer scaffold with an intrinsic network of channels. Ann Surg 1998;228:8-13. [86] Lee JS, et al. Cultured cell patterning for bio-electronics. Mol Cryst Liq Cryst Sci Tech Sect A 1994;247:365-72. [87] Lee JS, et al. Selective adhesion and proliferation of cells on ion-implanted polymer domains. Biomaterials 1993;14:958-60. [88] Dike LE, et al. Geometric control of switching between growth, apoptosis, and differentiation during angiogenesis using micropatterned substrates. Curr Biol 1999, submitted for publication.
175
The Biomaterials Silver Jubilee C o m p e n d i u m
Biomaterials ELSEVIER
Biomaterials 21 (2000) 2529-2543
Scaffolds in tissue engineering bone and cartilage Dietmar W. Hutmacher Laboratory for Biomedical Engineering, Institute of Engineering Science, Department of Orthopedic Surgery, National University of Singapore, 10 Kent Ridge Crescent, Singapore 119260, Singapore
Abstract Musculoskeletal tissue, bone and cartilage are under extensive investigation in tissue engineering research. A number of biodegradable and bioresorbable materials, as well as scaffold designs, have been experimentally and/or clinically studied. Ideally, a scaffold should have the following characteristics: (i) three-dimensional and highly porous with an interconnected pore network for cell growth and flow transport of nutrients and metabolic waste; (ii) biocompatible and bioresorbable with a controllable degradation and resorption rate to match cell/tissue growth in vitro and/or in vivo; (iii) suitable surface chemistry for cell attachment, proliferation, and differentation and (iv) mechanical properties to match those of the tissues at the site of implantation. This paper reviews research on the tissue engineering of bone and cartilage from the polymeric scaffold point of view. 9 2000 Elsevier Science Ltd. All rights reserved. Keywords" Tissue engineering of bone and cartilage; Design and fabrication of 3-D scaffold; Biodegradable and bioresorbable polymers
1. Introduction Bone and cartilage generation by autogenous cell/tissue transplantation is one of the most promising techniques in orthopedic surgery and biomedical engineering [1]. Treatment concepts based on those techniques would eliminate problems of donor site scarcity, immune rejection and pathogen transfer [-2]. Osteoblasts, chondrocytes and mesenchymal stem cells obtained from the patient's hard and soft tissues can be expanded in culture and seeded onto a scaffold that will slowly degrade and resorb as the tissue structures grow in vitro and/or in vivo [3]. The scaffold or three-dimensional (3-D) construct provides the necessary support for cells to proliferate and maintain their differentiated function, and its architecture defines the ultimate shape of the new bone and cartilage. Several scaffold materials have been investigated for tissue engineering bone and cartilage including hydroxyapatite (HA), poly(~-hydroxyesters), and natural polymers such as collagen and chitin. Several reviews have been published on the general properties and design features of biodegradable and bioresorbable polymers and scaffolds [4-12]. The aim of this paper is to complete the information collected so far, with special emphasis on the evaluation of the material and design characteristics which are of specific interest in tissue engineering the mesenchymal tissues bone and cartilage.
The currently applied scaffold fabrication technologies, with special emphasis on the so-called solid-free form fabrication technologies, will also be bench marked. Finally, the paper discusses the author's research on the design and fabrication of 3-D scaffolds for tissue engineering an osteochondral transplant.
2. Polymer-based scaffold materials The meaning and definition of the words biodegradable, bioerodable, bioresorbable and bioabsorbable (Table 1)--which are often used misleadingly in the tissue engineering literature--are of importance to discuss the rationale, function as well as chemical and physical properties of polymer-based scaffolds. In this paper, the polymer properties are based on the definitions given by Vert [13]. The tissue engineering program for bone and cartilage in the author's multidisciplinary research curriculum has been classified into six phases (Table 2). Each tissue engineering phase must be understood in an integrated manner across the research program--from the polymer material properties, to the scaffold micro- and macroarchitecture, to the cell, to the tissue-engineered transplant, to the host tissue. Hence, the research objectives in each phase are cross-disciplinary and the sub-projects are linked horizontally as well as vertically.
0142-9612/00/$- see front matter 9 2000 Elsevier Science Ltd. All rights reserved. PII: S0 142-96 12(00)001 2 1-6
176 2530
The Biomaterials Silver Jubilee Compendium
D. W. Hutmacher / Biomaterials 21 (2000) 2529-2543
Table 1 Definitions given by Vert Biodegradable are solid polymeric materials and devices which break down due to macromolecular degradation with dispersion in vivo but no proof for the elimination from the body (this definition excludes environmental, fungi or bacterial degradation). Biodegradable polymeric systems or devices can be attacked by biological elements so that the integrity of the system, and in some cases but not necessarily, of the macromolecules themselves, is affected and gives fragments or other degradation by-products. Such fragments can move away from their site of action but not necessarily from the body. Bioresorbable are solid polymeric materials and devices which show bulk degradation and further resorb in vivo; i.e. polymers which are eliminated through natural pathways either because of simple filtration of degradation by-products or after their metabolization. Bioresorption is thus a concept which reflects total elimination of the initial foreign material and of bulk degradation by-products (low molecular weight compounds) with no residual side effects. The use of the word 'bioresorbable' assumes that elimination is shown conclusively. Bioerodible are solid polymeric materials or devices, which show surface degradation and further, resorb in vivo. Bioerosion is thus a concept, too, which reflects total elimination of the initial foreign material and of surface degradation by-products (low molecular weight compounds) with no residual side effects. Bioabsorbable are solid polymeric materials or devices, which can dissolve in body fluids without any polymer chain cleavage or molecular mass decrease. For example, it is the case of slow dissolution of water-soluble implants in body fluids. A bioabsorbable polymer can be bioresorbable if the dispersed macromolecules are excreted.
Table 2 The research program for tissue engineering bone and cartilage classified into six phases I--Fabrication of bioresorbable scaffold II--Seeding of the osteoblasts/chondrocytes populations into the polymeric scaffold in a static culture (petri dish) III--Growth of premature tissue in a dynamic environment (spinner flask) IV--Growth of mature tissue in a physiologic environment (bioreactor) V--Surgical transplantation VI--Tissue-engineered transplant assimilation/remodeling
The first stage of tissue engineering bone or cartilage begins with the design and fabrication of a porous 3-D scaffold, the main topic of this review paper. In general, the scaffold should be fabricated from a highly biocompatible material which does not have the potential to elicit an immunological or clinically detectable primary or secondary foreign body reaction [9]. Furthermore, a polymer scaffold material has to be chosen that will degrade and resorb at a controlled rate at the same time as the specific tissue cells seeded into the 3-D construct attach, spread and increase in quantity (number of cells/per void volume) as well as in quality. Currently, the design and fabrication of scaffolds in tissue engineering research is driven by three material categories: I. Regulatory approved biodegradable and bioresorbable polymers (Table 3), such as collagen, polyglycolide (PGA),
polylactides (PLLA, PDLA), polycaprolactone (PCL), etc. II. A number of non-approved polymers, such as polyorthoester (POE), polyanhydrides, etc. which are also under investigation. III. The synthesis of entrepreneurial polymeric biomaterials, such as poly (lactic acidco-lysine), etc., which can selectively shepherd specific cell phenotypes and guide the differentiation and proliferation into the targeted functional premature and/or mature tissue. In general, polymers of the poly(~-hydroxy acids) group undergo bulk degradation. The molecular weight of the polymer commences to decrease on day one (PGA, PDLA) or after a few weeks (PLLA) upon placement in an aqueous media [12]. However, the mass loss does not start until the molecular chains are reduced to a size which allows them to freely diffuse out of the polymer matrix [ 14]. This phenomenon described and analyzed in detail by a number of research groups [15-18], results in accelerated degradation and resorption kinetics until the physical integrity of polymer matrix is compromised. The mass loss is accompanied by a release gradient of acidic by-products. In vivo, massive release of acidic degradation and resorption by-products results in inflammatory reactions, as reported in the bioresorbable device literature [19-22]. If the capacity of the surrounding tissue to eliminate the by-products is low, due to the poor vascularization or low metabolic activity, the chemical composition of the by-products may lead to local temporary disturbances. One example of this is the increase of osmotic pressure or pH manifested through local fluid accumulation or transient sinus formation from fiber reinforced polyglycolide pins applied in orthopedic surgery [21]. Potential problems of biocompatibility in tissue engineering bone and cartilage, by applying degradable, erodable, and resorbable polymer scaffolds, may also be related to biodegradability and bioresorbability. Therefore, it is important that the 3-D scaffold/cell construct is exposed at all times to sufficient quantities of neutral culture media, especially during the period where the mass loss of the polymer matrix occurs. The incorporation of a tricalciumphosphate (TCP) [23], hydroxyapatite (HA) [24] and basic salts [15] into a polymer matrix produces a hybrid/composite material. These inorganic fillers allow to tailor the desired degradation and resorption kinetics of the polymer matrix. A composite material would also improve biocompatibility and hard tissue integration in a way that ceramic particles, which are embedded into the polymer matrix, allow for increased initial flash spread of serum proteins compared to the more hydrophobic polymer surface [9]. In addition, the basic resorption products of HA or TCP would buffer the acidic resorption by-products of the aliphatic polyester and may thereby help to avoid the formation of an unfavorable environment for the cells due to a decreased pH [15,23,24].
t~
Table 3 Properties of bioresorbable and bioerodable polymers Polymer
Comparison of mechanical properties of bioerodable and bioresorbable polymers
t,,,,~ o
Degradation and resorption process via hydrolysis
Molecular weight loss/loss of mechanical properties (in month) a
Mass loss (in month) a
References (scaffolds)
References Area of (medical device) application
Products with regulatory approval ,~,~o
Poly(L-lactide)
+ + +
Bulk erosion
9-15
36-48
Poly(L-lactide-co-D, L-lactide) 70/30
+ +
Bulk erosion
5-6
12-18
Poly(L-lactide-coglycolide) 10/90 Polyglycolide Poly(D,L-lactide) Poly(D,L-lactideco-glycolide) 85/15 Poly(D,L-lactideco-glycolide) 75/25 Poly(D,L-lactideco-glycolide) 50/50 Polycaprolactone
+ +
Bulk erosion
1-2
3-4
28, 63
+ + + + +
Bulk erosion Bulk erosion Bulk erosion
0.5-1 1-2 1-2
3-4 5-6 4-5
6, 30-35, 41, 65 6, 31, 34, 49, 60 53-56, 60, 70
+
Bulk erosion
1-2
4-5
49, 61
+ +
Bulk erosion
1-2
3-4
28
+
Polyorthoester Polyanhydrides
+ + + +
Bulk and surface erosion Surface erosion Surface erosion
46, 48, 49, 60-64, 70-72
19, 20, 24
22, 23
Orthopedic Surgery, Oral and Maxillofacial Surgery Oral and Maxillofacial Surgery, Orthopedic Surgery Suture Periodontal Surgery, Surgery, Orthopedic Surgery
FixSorb System (screws, nails, pins) Neofix (screws, nails, pins) ResorPin, Leadfix MacroSorb System (screws and plates, mesh, nails, pins) PolyPin Vicryl Suture, Vicryl Mesh Biofix t..j
t~ t-.a
I
9-12
24-36
29
Drug delivery
4-6 4-6
12-18 12-18
59
Animal experiments
Capranor
4~
a Molecular weight and mass loss vary depending on factors such as chemical structure and composition; presence of ionic groups and of side group defects; configuration of the structure molecular weight and molecular weight distribution (polydispersity); presence of low molecular weight components (monomers, oligomers, solvents, softeners, drugs, growth factors, etc.); production and manufacturing procedures and their process parameters, implant design, sterilization method, morphology (amorphous versus semi-crystalline, presence of microstructures and stress within the components), tempering, storage, implant site. + + + , good; + + , average; +, poor.
-.q t.J t.a.a
"q
178
The Biomaterials Silver Jubilee Compendium
2532
D. W. Hutmacher / Biomaterials 21 (2000) 2529-2543
Control of the hydrodynamic and biochemical environment is essential for the successful in vitro engineering of 3-D scaffold/tissue constructs for potential clinical use [25]. Computer-controlled bioreactors that continuously supply physiological nutrients and gases, serve to regulate the required cell/tissue culture conditions for a long period of time. After the in vitro culturing of the 3-D scaffold/tissue construct, the degree of remodeling and cell/tissue replacement of the bone/cartilage transplant by the host tissue has to been taken into consideration [26]. Cell and tissue remodeling is important for achieving stable mechanical conditions and vascularization at the host site. Hence, the 3-D scaffold/tissue construct should maintain sufficient structural integrity during the in vitro and/or in vivo growth and remodeling process. The degree of remodeling depends on the host anatomy and physiology [26]. The polymer selection from a material science point of view is based on two different strategies in regard to the overall function of the scaffold.
the scaffold matrix must serve an additional function; it must provide sufficient temporary mechanical support to withstand in vivo stresses and loading. In Strategy I research programs, the material must be selected and/or designed with a degradation and resorption rate such that the strength of the scaffold is retained until the tissue engineered transplant is fully remodeled by the host tissue and can assume its structural role. Bone is able to remodel in vivo under so-called physiological loading [27]. It is a requirement that the degradation and resorption kinetics have to be controlled in such a way that the bioresorbable scaffold retains its physical properties for at least 6 months (4 months for cell culturing and 2 months in situ). Thereafter, it will start losing its mechanical properties and should be metabolized by the body without a foreign body reaction after 12-18 months (Fig. 1). The mechanical properties of the bioresorbable 3-D scaffold/tissue construct at the time of implantation should match that of the host tissue as closely as possible [7]. It should posses sufficient strength and stiffness to function for a period until in vivo tissue ingrowth has replaced the slowly vanishing scaffold matrix. Thompson et al. [28] studied a poly(D,L-lactide-coglycolide) matrix under cyclic compressive loading. They concluded that changes in surface deformation and morphology suggest that the compressive loading initially collapses and stiffens the polymer matrix. The de-
2.1. Strategy I
In the first strategy (Fig. 1), the physical scaffold structure supports the polymer/cell/tissue construct from the time of cell seeding up to the point where the hard tissue transplant is remodeled by the host tissue. In the case of load-bearing tissue such as articular cartilage and bone,
I
100%
I
II
I
III
1 i
i
I
i
I
I IV
1
i
!
I
l
I Hydration II Hydration and Degradation III Degradation and Mass Loss IV Resorption and Metabolisation V Metabolisation
V
,' Mdle~zular w~ight ios~ , i I , ' ~D Scat~fold t =.-.K,.r-.~, ~, . . , . ~ . . ~ . , . - , olOSS Ol , \ , , \/ , lvlass ' i\ i i '1 i I ' . . . . i [ x~ I b,oresor~able 3I~,, scaffol , .. . .. . . . ", \]......I- ,, ', /
/\.
50%
l
,
I
i
i
,
I I '
,
,
i
'
i
', I ',I |
!,
',i
!
!
i
~
11
I
i
~
I
/ /
"~[',
]
,'.\
I
~
~I
I
/
f,
9
i
i
~
,, - .
~ , Remodelin", g of tissue '
e r e d ~trans henn g~ ~ mt ee;s:ed ; ; ; P ~P lant ;intn bbo i
_
tissue in vi
'i
i
i
/
,
~ k e l e t ~ l
I
II X".\
!i
/
i
i
~
i
0% !
i
i
I
!
i
i
i0 1 2 ~
6
9
I
i
12
15
24
,,
,
,' I i
,
', !
B
C
D
A
!
',
I iT M
Numb, er of weeks
I f-~,
~
F
'o I
-,q
E
Fig. 1. Graphical illustration of the complex interdependence of molecular weight loss and mass loss of a strategy 1 3D scaffold matrix plotted against the time frame for tissue engineering a cartilage/bone transplant. (A) Fabrication of bioresorbable scaffold; (B) seeding of the osteoblast/cartilage populations into the polymeric scaffold in a static culture (petri dish); (C) growth of premature tissue in a dynamic environment (spinner flask); (D) growth of mature tissue in a physiologic environment (bioreactor); (E) surgical transplantation; (F) tissue-engineered transplant assimilation/remodeling.
179
The Biomaterials Silver Jubilee Compendium
2533
D.W. Hutmacher / Biomaterials 21 (2000) 2529-2543
7 month in a bioreactor reached 40% of the mechanical properties of natural cartilage. The next phase of those research programs will be to evaluate how these tissueengineered cartilage transplants assimilate and remodel in vivo [35]. In an in vivo model, one of the major problems from a biomechanical and clinical view point, is the primary mechanical stabilization of cartilage transplants [28]. This aspect will be discussed below, under scaffold design, in more detail.
crease in molecular weight is slowed down due to the reduction of surface area from hydrolysis, until the matrix architecture no longer accommodates the mechanical loading and begins to lose its integrity. Conclusively, in Strategy I the scaffold architecture has to withstand mechanical loading in vitro and in vivo.
2.2. Strategy H For Strategy II (Fig. 2), the intrinsic mechanical properties of the scaffold architecture templates the cell proliferation and differentiation only up to the phases where the premature bone or cartilage is placed in a bioreactor. The degradation and resorption kinetics of the scaffold are designed to allow the seeded cells to proliferate and secrete their own extracellular matrix in the static and dynamic cell seeding phase (weeks 1-12), while the polymer scaffold gradually vanishes leaving sufficient space for new cell and tissue growth. The physical support by the 3-D scaffold is maintained until the engineered tissue has sufficient mechanical integrity to support itself. Different research groups [29-33] have shown in a number of studies that a nonwoven mesh made of polyglycolide fibers offers degradation and resorption kinetics for Strategy II. However, the challenge for the grown cell/tissue construct is to have similar mechanical properties to the host bone and cartilage. Ma and Langer [34] showed that cartilage which was cultured for [ Hydration [ 9
|
I
9
.
.
.
.
.
.
Skeletal tissue, such as bone and cartilage, is usually organized into 3-D structures in the body [36]. For the repair and generation of hard and ductile tissue, such as bone, scaffolds need to have a high elastic modulus in order to be retained in the space they were designated for; and also provide the tissue with adequate space for growth [37]. If the 3-D scaffold is used as a temporary load-bearing device (Strategy II), the mechanical properties would maintain that load for the required time without showing symptoms of fatigue or failure. Therefore, one of the basic problems from a scaffold design point of view, is that to achieve significant strength the scaffold material must have sufficiently high interatomic and intermolecular bonding, but must have at the same time a physical and chemical structure which allows for hydrolytic attack and breakdown.
Hydration .] Bulk Erosion .] Bulk Erosion .] Metabolisation i Degradation [ Mass loss i Metabolisation .
I
00% i_
3. Scaffold design
.
',
I
i i
I
9
,:
.
: Molelzula
I w;igl~t los; : ] ' I9
,,
50~176i
i
9 I)% :i I I~ v
! /ii i / I !9/ 1
/
,
i ~ / !
~
,
6
!
'
,
I
I
D
\ //;\
9construct r e n i o d e l i n g a n d ~ell/tissue repla[~ement by the 9i host tissue \',: ":
Mass l~ ~f'~3 D / s c a f f ~
I 3
i
!
~
9 construct in:vitro 'i, '1
"i ,
IB C',
I
9
',
F
I
:
:
I 12
Ii ',15
Number of mo~hs ', i
w"i'~
'E
Fig. 2. Graphical illustration of the complex interdependence of molecular weight loss and mass loss of a strategy II 3D scaffold matrix plotted against the time frame for tissue engineering a cartilage/bone transplant. (A) Fabrication of bioresorbable scaffold; (B) seeding of the osteoblast/cartilage populations into the polymeric scaffold in a static culture (petri dish); (C) growth of premature tissue in a dynamic environment (spinner flask); (D) growth of mature tissue in a physiologic environment (bioreactor); (E) surgical transplantation; (F) tissue-engineered transplant assimilation/remodeling.
The Biomaterials Silver Jubilee Compendium
180 2534
D. W. Hutmacher / Biomaterials 21 (2000) 2529-2543
Ingber and his group [38,39] design their scaffolds based on a concept which they named tensegrity. Geodesical 3-D constructs were designed by applying tensegrity so that the entire scaffold structure evenly distributes and balances mechanical stresses. The walls, layers or struts that make up the interconnecting scaffold framework are connected into triangles, pentagons or hexagons, each of which can bear tension or compression. However, the mechanical rational of the tensegrity design concept has been known for centuries in the area of civil engineering. Gothic architects used a stone skeleton structure of stone columns and ribs supported by arches and buttresses to build cathedrals (Fig. 3). Another point, which has to be focused on is the diffusion of nutrients into the 3-D scaffold. Although, an interconnected macropore-structure of 300-500 ~tm enhances the diffusion rates to and from the center of a scaffold, transportation of the nutrients and by-products is not sufficient for large scaffold volumes. A fluiddynamic microenvironment provided by a bioreactor can mimic the interstitial fluid conditions present in natural bone and cartilage in a macroporous scaffold architecture. Bioreactors permit in vitro culture of larger and better organized 3-D cell communities than can be achieved using standard tissue culture techniques [40].
For tissue engineering a bone transplant, the creation of a vascularized bed ensures the survival and function of seeded cells, which have access to the vascular system for nutrition, gas exchange, and elimination of by-products [41]. The vascularization of a scaffold may be compromised by purely relying on capillary ingrowth into the interconnecting pore network from the host tissue. In situ, the distance between blood vessels and mesenchymal cells are not larger then 100 gm [42]. Therefore, the time frame has to be taken into account for the capillary system to distribute through larger scaffold volume. It may also be possible to control the degree and rate of vascularization by incorporating angiogenic and anti-angiogenic factors in the degrading matrix of the scaffold. From a biomechanical and clinical point of view, the tissue-engineered bone or cartilage transplant should allow for a mechanically secure and stable fixation on or to the host tissue [29]. For bone, the currently available medical devices, such as pins, screws, and plates might be used. However, the integration of a device-like part into the 3-D scaffold design can be advantageous. The rationale for such an innovative design concept will be for the tissue engineering of an osteochondral bone transplant is described below.
4. Scaffold fabrication
Fig. 3. The Cologne cathedral was built in the 18th and 19th century by gothic architects who used a stone skeleton structure of stone columns and ribs supported by arches and buttresses.
A number of fabrication technologies have been applied to process biodegradable and bioresorbable materials into 3-D polymeric scaffolds of high porosity and surface area. The conventional techniques for scaffold fabrication include fiber bonding, solvent casting, particulate leaching, membrane lamination and melt molding (Table 4). Several papers have reviewed the past and current research on scaffold fabrication techniques [43-45]. However, none of those papers has directly compared the 3-D scaffold-processing technologies for the tissue engineering community. From a scaffold design and function view point each processing methodology has its pro and cons. It is the aim of this paper to aggregate the compiled information and to present this data in a comprehensive form. Table 4 summarizes the key characteristics and parameters of the techniques currently used. The aim of this part of the review paper is to assist research teams with their choice for a specific 3-D scaffold-processing technology by providing the information needed to determin the critical parameters. As discussed above, at present the challenge in tissue engineering bone and cartilage is not only to design, but also to fabricate reproducible bioresorbable 3-D scaffolds, which are able to function for a certain period of time under load-bearing conditions. Solvent casting, in combination with particle leaching, works only for thin membranes or 3-D specimens with
e~
t,,~~
Table 4 Currently applied 3D scaffold fabrication technologies c,z Fabrication technology
Processing
Material properties required for processing
Scaffold design and reproducibility
Achievable pore size in Jam
Porosity in %
Architecture
Reference
Solvent casting in combination with particular leaching Membrane lamination
Casting
Soluble
User, material and technique sensitive
30-300
20-50
Spherical pores, salt particles remain in matrix
47
Solvent bonding
Soluble
30-300
<85
Irregular pore structure
48
Fabrication of non-woven
Carding, Needling, Plate pressing Moulding Extrusion through dies Casting
Fibres
User, material and technique sensitive Machine controlled
20-100
<95
Insufficient mechanical properties
30-35, 41, 63-65
Thermoplastic Thermoplastic
Machine controlled Machine controlled
50-500 < 100
< 80 < 84
49
Soluble
< 200
<97
Casting
Soluble
< 200
<97
Casting
Amorphous
< 100
10-30
Casting
Amorphous
User, material and technique sensitive User, material and technique sensitive Material and technique sensitive Material and technique sensitive
Solid free form fabrication
Soluble
Solid free form fabrication
Thermoplastic
Spherical pores, salt particles remain in matrix High volume of inter-connected micropore structure High volume of inter-connected micropore structure High volume of non interconnected micropore structure Low volume of non-interconnected micropore structure combined with interconnected macropore structure 100% interconnected macropore (triangles, pentagons, honey comb, etc.), design and fabrication layer by layer, by use of water-based binder incorporation of biological agents into matrix possible 100% interconnected macropore structure (triangles, pentagons, honey comb, etc.), design and fabrication layer by layer
Melt moulding Extrusion in combination with particular leaching Emulsion freeze drying Thermally induced phase separation Supercritical-fluid technology Supercritical-fluid technology in combination with particle leaching 3-D printing in and without combination of particle leaching
Fused deposition modelling
Micropores < 50 macropores < 400
<97
Machine and computer controlled
45-150
< 60
Machine and computer controlled
> 150
< 80
t...~ o t~
r~
54-56 58-61 t..a
53 54
69-72
t-.a t~ t-,a t..a t~ 4~ t~
29
c~ t,3
182
2536
The Biomaterials Silver Jubilee C o m p e n d i u m D.W. Hutmacher/ Biomaterials 21 (2000) 2529-2543
very thin wall sections: otherwise, it is not possible to remove the soluble particles from within the polymer matrix [46]. Mikos et al. [47], using the above-described technology, fabricated porous sheets and laminated them to 3-D structures. Chloroform was used on the attachment interface for the lamination process. This fabrication technology is time consuming because only thin membranes can be used. Another disadvantage is that the layering of porous sheets allows only a limited number of interconnected pore networks. Solvent-casted polymersalt composites have also been extruded into a tubular geometry [48]. The disadvantages of the above technologies include the extensive use of highly toxic solvents, time required for solvent evaporation (days-to-weeks), the labor intensive fabrication process, the limitation to thin structures, residual particles in the polymer matrix, irregularly shaped pores, and insufficient interconnectivity. The supercritical fluid-gassing process has been known for many years in the non-medical polymer industry [49] as well as in the pharmaceutical community [50]. This technology is used to produce foams and other highly porous products. The polymers which can be used for this technology have to have an high amorphous fraction. The polymer granules are plasticized due to the employment of a gas, such as nitrogen or carbon dioxide, at high pressures. The diffusion and dissolution of the gas into the polymer matrix results in a reduction of the viscosity, which allows the processing of the amorphous bioresorbable polyesters in a temperature range of 30-40~ [51]. The supercritical fluid-gassing technology allows the incorporation of heat sensitive pharmaceuticals and biological agents. However, on average only 10-30% of the pores are interconnected [51,52]. Harris et al. [53] combined this technology with particulate leaching to gain a highly interconnected void network. The researchers could control porosity and pore size by varying the particle/polymer ratio and particle size. Whang et al. [54,55] developed a protocol for the fabrication of aliphatic polyester-based scaffolds by using the emulsion freeze-drying method. Scaffolds with porosity greater than 90%, median pore sizes ranging from 15 to 35 gm with larger pores greater than 200 gm were fabricated. The scaffold pore architecture was highly interconnected which is necessary for tissue ingrowth and regeneration [54]. Based on their results from an animal experiment, the interdisciplinary group proposed a scaffold design concept which results in in vivo bone regeneration based on hematoma stabilization [56]. The authors compare their in vivo bone engineering concept to the induction phase of fracture healing. The osteoprogenitor cells which are in the blood of the osseous wound are embedded in the scaffold microarchitecture via the hematoma. The multipotent cells differentiate to osteoblasts due to the presence of growth factors which are released by the host bone. However, the emulsion freeze-drying method is user and technique sensitive. The
fabrication of a truly interconnecting pore structure depends on the processing method and parameters as well as on the used equipment [57,58]. Several groups [57-60] studied thermally induced phase separation technology to process polymeric 3-D scaffolds. This technique has been used previously to fabricate synthetic membranes for non-medical applications. The method has been extensively applied in the field of drug delivery to fabricate microspheres, which allows the incorporation of pharmaceutical and biological agents, such as bone morphogenetic proteins (BMPs) into the polymer matrix. In general, the microand macro-structure is controlled by varying the polymer material, polymer concentration, quenching temperature, and solvents. However, current research shows that the method, similar then emulsion freeze-drying technique, is user and technique sensitive and that the processing parameters have to be well controlled. Nam and Park [57] as well as Zhang and Ma [58] fabricated polymer and polymer/HA specimens with a porosity of
Fig. 4. Polymerdisks with a diameter of 500 ~tm and 40 gm thickness allow to stack a 3D scaffold with a porosity of 98%.
Fig. 5. Schematicillustration of the fused deposition modeling(FDM) process.
The B i o m a t e r i a l s Silver J u b i l e e C o m p e n d i u m D. W. Hutmacher/ Biomaterials 21 (2000) 2529-2543 up to 9 5 % . At present, only p o r e sizes of up to 100 g m can be r e p r o d u c i b l y f a b r i c a t e d by t h e r m a l l y i n d u c e d phase-separation technology. A n u m b e r of textile t e c h n o l o g i e s h a v e the p o t e n t i a l to design a n d fabricate highly p o r o u s scaffolds [61]. Yet, only so-called n o n - w o v e n mesh-like designs h a v e been
183 2537
used to tissue e n g i n e e r b o n e a n d cartilage [62-64]. Excellent results in tissue e n g i n e e r i n g cartilage h a v e been achieved by using n o n - w o v e n m e s h e s c o m p o s e d of polym e r fibers of P G A , P G A / P D L A , a n d P G A / P L L A . This w o r k has been reviewed by F r e e d et al. [65] a n d will n o t be discussed here. In general, n o n - w o v e n c o n s t r u c t s can
Fig. 6. (a) 3D scaffold systems of various porosity and pore geometry fabricated by FDM. Magnification x 7.5, scale bar represents 1 mm. (i)-(iii) lay-down pattern: 0/90~ nozzle tip: 0.016"; porosity: 50, 68, 75%; (iv)-(vi)0/90~ 0.010"; 50, 68, 75%; (vii)-(viii)0/60/120~ 0.016"; 68, 75%; (ix) 0/60/120~ 0.010"; 80%; (x)-(xii) 0/60/120~ 0.010"; 50, 68, 75%. (b) Left: cross-sectional view of freeze-fractured PCL scaffold with lay-down pattern 0/72/144/36/108 ~ Right: plan view of same specimen. (c) Left: top view of PCL-HA scaffold with lay-down pattern 0/60/120 ~ The material composition consists of 80% PCL and 20% HA by weight. Right: a close-up view of the same specimen at a higher magnification, shows that HA particles are at the scaffold surface. (d) Freeze-fracture cross-sectional surface ( x 23) of a bioerodable scaffold designed for a bone/cartilage interface. A 0/60/120 ~ and a 0/90 ~lay-down pattern of the roads shown in the upper and lower portion respectively of the SEM picture. Despite being fabricated with different lay-down patterns, the porosity (67%) of both portions could be made identical. However, by applying FDM the porosity of each portion of the 3D scaffold can be varied according to the road spacing for individual layers.
184 2538
The Biomaterials Silver Jubilee Compendium
D.W. Hutmacher / Biomaterials 21 (2000) 2529-2543
Fig. 6. (Continued).
be only used for Strategy II since their physical properties do not allow load-bearing applications. All the above-described technologies except the membrane-lamination method, do not allow the fabrication of a 3-D scaffold with a varying multiple layer design. Such a matrix architecture is advantageous in instances where tissue engineers want to grow a bi- or multiple tissue interface, e.g. an articular cartilage/bone transplant.
Rapid prototyping technologies as well as so-called 'wafer stacking systems' [66] (Fig. 4) have the potential to design a 3-D construct in a multi-layer design within the same gross architectural structure. In engineering literature [67] Rapid prototyping Technologies (RP) also called Solid Free Form fabrication (SFF) methods are defined as a set of manufacturing processes that are capable of producing complex-free
185
The Biomaterials Silver Jubilee Compendium D.W. Hutmacher / Biomaterials 21 (2000) 2529-2543
form parts directly from a computer-aided design (CAD) model of an object without part specific toolin g or knowledge. Unlike machining processes such as milling, drilling, etc., which are subtractive in nature, RP systems join together liquid, powder and sheet materials to form parts. Layer by layer, RP machines fabricate plastic, wood, ceramic and metal objects using thin horizontal cross sections directly from a computer-generated model. Rapid prototyping technologies, such as 3-D printing (3-DP) and fused deposition modeling (FDM) allow the development of manufacturing processes to create porous scaffolds that mimic the microstructure of living tissue. Three-dimensional printing--developed at the Massachusetts Institute of Technology [68]--is also a rapid prototyping technology which has been used to process bioresorbable scaffolds for tissue engineering applications [69,70]. The technology is based on the printing of a binder through a print head nozzle onto a powder bed, with no tooling required. The part is built sequentially in layers. The binder is delivered to the powder bed producing the first layer, the bed is then lowered to a fixed distance, powder is deposited and spread evenly across the bed, and a second layer is built. This is repeated until the entire part, e.g. a porous scaffold, is fabricated. Following treatment, the object is retrieved from the powder bed and excess unbound powder is removed. The speed, flow rate and even drop position can be computer controlled to produce complex 3-D objects [63]. This printing technique permits CAD) and custom-made fabrication of bioresorbable hybrid scaffold systems. The entire process is performed under room-temperature conditions. Hence, this technology has great potential in tissue engineering applications. Biological agents, such as cells, growth factors, etc., can be incorporated into a porous scaffold without inactivation if non-toxic binders, e.g. water can be used [71]. Unfortunately, aliphatic polyesters can be only dissolved in highly toxic solvents, such as chloroform, methylene chloride, etc. To date, only bioresorbable scaffolds without biological agents within the polymer matrix and in combination with particle leaching have been processed by 3-D printing. In addition, the mechanical properties and accuracy of the specimen manufactured by 3-D printing have to be significantly improved [72]. The FDM process forms 3-D objects from a CAD file as well as digital data produced by an imaging source such as computer tomography (CT) or magnetic resonance imaging (MRI) [73]. The process begins with the design of a conceptual geometric model on a CAD workstation. The design is imported into a software, which mathematically slices the conceptual model into horizontal layers. Toolpaths are generated before the data is downloaded to the FDM hardware. The FDM extrusion head (Fig. 5) operates in the X and Y axes while the platform lowers in the Z-axis for each new layer to form.
2539
Fig. 7. Schematic illustration of the surgical placement of a tissue engineered a bone/cartilage interphase. The dental cylinder implantlike design of the bony part of the scaffold allows to apply a press-fit between the host bone and the cylindrical portion of the scaffold.
In effect, the process draws the designed model (scaffold) one layer at a time. Thermoplastic polymer material, 1.78 mm in diameter, feeds into the temperature-controlled FDM extrusion head where it is heated to a semi-liquid state. The head extrudes and deposits the material in ultra-thin layers onto a fixture less base. The head directs the material precisely into place. The material solidifies, laminating to
186 2540
The Biomaterials Silver Jubilee C o m p e n d i u m D.W. Hutmacher/ Biomaterials 21 (2000) 2529-2543
the preceding layer. Parts are fabricated in layers, where a layer is built by extruding a small bead of material, or road, in a particular lay-down pattern, such that the layer is covered with the adjacent roads. After a layer is completed, the height of the extrusion head is increased and the subsequent layers are built to construct the part. In the past, non-medical and medical F D M users could only use a few non-resorbable polymeric materials, such as polyamide, ABS, and other resins. The broadening of
the research and application scope of F D M results in a need/demand to evaluate the critical factors for using the new polymeric and composite materials on F D M systems [74]. At present, the author's multidisciplinary group has been able to evaluate the parameters to process a number of potential scaffold materials, such as P C L and P C L / H A by F D M . To our knowledge, we are the first group to report the processing of bioresorbable scaffolds
Fig. 8. (a) Schematic illustration of a perfusion culture chamber for tissue engineering a bone/cartilage interphase; (b) schematic sketch of the bioreactor concept for the tissue engineering of bone and cartilage simultaneously, in the one device like scaffold architecture.
The Biomaterials Silver Jubilee C o m p e n d i u m D.W. Hutmacher/ Biomaterials 21 (2000) 2529-2543 (Fig. 6a-d) for tissue engineering applications using FDM.
5. Tissue engineering of a articular cartilage-bone interface Articular cartilage repair, regeneration and generation have been reviewed by Buckwalter and M a n k i n [75] as well as N e w m a n [76]. Both the reports have concluded that in the last two decades, clinical and basic scientific investigations have shown that the i m p l a n t a t i o n of artificial matrices, growth factors, perichondrium, and periosteum, can stimulate the f o r m a t i o n of cartilaginous tissue in o s t e o c h o n d r a l and chondral defects in synovial fluids. However, the available evidence indicates that the results vary considerably a m o n g the individuals, and that the tissues formed using these t r e a t m e n t regimes do not duplicate the composition, structure, and mechanical properties of n o r m a l articular cartilage [77-81]. In addition, none of those matrix designs could be securely fixed under load-bearing conditions to the o s t e o c h o n d r a l bone which is a conditio sine qua non from a biomechanical and clinical point of view. The author's multidisciplinary g r o u p has conceptualized a rationale for tissue engineering a load-bearing o s t e o c h o n d r a l t r a n s p l a n t (Fig. 7). The scaffold design, material and fabrication technology, as well as the bioreactor design (Fig. 8a and b), enables the tissue engineering of bone and cartilage simultaneously, in the one scaffold architecture. The device-like design of the bony scaffold part allows a secure b o n e - t o - b o n e fixation of the articular b o n e - c a r t i l a g e interface. The concept has been discussed in detail elsewhere E82].
6. Conclusions The application of regulatory a p p r o v e d biomaterials to design and fabricate 3-D scaffolds has strongly s u p p o r t e d the drive for the establishment of tissue engineering research. Reviewing the experimental and clinical studies, it can be concluded that the ideal scaffold and matrix material for tissue engineering bone and cartilage has not yet been developed. In general, the tissue engineers do not i m p l e m e n t innovative scaffold fixation features into their design concept. However, from a clinical point of view, the secure and user friendly transplant fixation is a conditio sine qua non. The shape of a m e s e n c h y m a l tissue such as a bone and articular cartilage is often critical to its function. Ideally, a polymeric scaffold material should permit application of a solid-free form fabrication technology, so that a porous scaffold with any desired 3-D g e o m e t r y can be designed and fabricated by using CT and M R I data.
187 2541
References [1] Patrick Jr CW, Mikos AG, McIntire. Prospectus of tissue engineering. In: Patrick Jr CW, Mikos AG, McIntire LV, editors. Frontiers in tissue engineering. New York, USA: Elsevier Science, 1998. p. 3-14. [-2] Naughton GK, et al. Emerging developments in tissue engineering and cell technology. Tissue Eng 1995;1(2):211-9. [3] Langer R, Vacanti JP. Tissue engineering. Science 1993;260: 920-6. [4] Agrawal C, Niederauer G, Micallef D, Athanasiou K. The use of PLA-PGA polymers in orthopaedics. In: Encyclopedic handbook of biomaterials and bioengineering. New York: Marcel Dekker, 1995. p. 2081-115. [5] Bajpai PK. Biodegradable scaffolds in orthopedic, oral and maxillofacial surgery. In: Rubin LR, editor. Biomaterials in reconstructive surgery. Mosby: St. Louis, 1983. p. 156. [6] Chaignaud BE, Langer R, Vacanti JP. The history of tissue engineering using synthetic biodegradable polymer scaffolds and cells. In: Atala A, Mooney DJ, editors. Synthetic biodegradable polymer scaffolds. Boston: Birkhauser, 1997. p. 1-14. [-7] Hollinger JO, Chaudhari A. Bone regeneration materials for the mandibular and craniofacial complex. Cells Mater 1992;2:143-51. [-8] Hutmacher D, Hurzeler M, Schliephake H. A review of material properties of biodegradable and bioresorbable polymers and devices for GTR and GBR applications. Int J Oral Maxillofac Implants 1996;11:667-78. [9] Hutmacher D, Kirsch A, Ackermann KL, Huerzeler MB. Matrix and carrier materials for bone growth factors--state of the art and future perspectives. In: Stark GB, Horch R, Tancos E, editors. Biological matrices and tissue reconstruction. Heidelberg, Germany: Springer, 1998. p. 197-206. [10] Pachence JM, Kohn J. Biodegradable polymers for tissue engineering. In: Lanza RP, Langer R, Chick WL, editors. Principles of tissue engineering. Austin, Texas, USA: R.G. Landes Co., 1997. p. 273-93. [11] West JL, Hubbell JA. Bioactive polymers. In: Atala A, Mooney DJ, editors. Synthetic biodegradable polymer scaffolds. Boston: Birkhauser, 1997. p. 83-95. [12] Wong WH, Mooney DJ. Synthesis and properties of biodegradable polymers used as synthetic matrices for tissue engineering. In: Atala A, Mooney DJ, editors. Synthetic biodegradable polymer scaffolds. Boston: Birkhauser, 1997. p. 51-82. [13] Vert M, Li MS, Spenlehauer G, Guerin P. Bioresorbability and biocompatibility of aliphatic polyesters. J Mater Sci 1992;3: 432-46. [-14] Kronenthal RL. Biodegradable polymers in medicine and surgery. Polym Sci Technol 1975;8:119-37. [15] Agrawal CM, Athanasiou KA. Techniques to control pH in vicinity of biodegrading PLA-PGA implants. J Biomed Mater Res Appl Biomater 1997;38(2):105-14. [-16] Li S, Garreau H, Vert M. Structure-property relationships in the case of the degradation of massive aliphatic poly(~-hydroxy-acids) in aqueous media. Part 1: poly(D,L-lacticacid). J Mater Sci: Mater Med 1990;1:123. [17] Li S, Garreau H, Vert M. Structure-property relationships in the case of the degradation of massive aliphatic poly(~-hydroxy-acids) in aqueous media. Part 2: degradation of lactide-glycolide PLA37.5 : GA25 and PLA75 :GA poly(D,L-lactic acid). J Mater Sci: Mater Med 1990;1:131. [18] Li S, Garreau H, Vert M. Structure-property relationships in the case of the degradation of massive aliphatic poly(~-hydroxy-acids) in aqueous media. Part 3: influence of the morphology of poly(Llactic acid). J Mater Sci: Mater Med 1990;1:198. [19] Bergsma EJ, Brujn W, Rozema FR, Bos RM, Boering G. Late tissue response to poly(L-lactide) bone plates and screws. Biomaterials 1995;16(1):25-31.
The Biomaterials Silver Jubilee Compendium
188 2542
D.W. Hutmacher / Biomaterials 21 (2000) 2529-2543
[20] Bergsma EJ, Rozema FR, Bos RM, Brujn W. Foreign body reactions to resorbable poly(L-lactide) bone plates and screws used for the fixation of unstable zygomatic fractures. J Maxillofac Surg 1993;51:666-70. [21] BiSstmann O, Hirvensalo E, M~ikinen J, Rokkanen P. Foreign body reactions to fracture fixation implants of biodegradable synthetic polymers. J Bone Jt Surg 1990;B72:592. [22] Rehm KE, Claes L, Helling HJ, Hutmacher D. Application of a polylactide pin. An open clinical prospective study. In: Leung KS, Hung LK, Leung PC, editors. Biodegradable implants in fracture fixation. Hong Kong: World Scientific, 1994. p. 54. [23] Hutmacher DW, Kirsch A, Ackermann KL, Huerzeler MB. A tissue engineered cell occlusive device for hard tissue regeneration--a preliminary report. Int J Periodontics Restorative Dent, accepted for publication. [24] Shikinami Y, Okuno M. Bioresorbable devices made of forged composites of hydroxyapatite (HA) particles and poly-L-lactide (PLLA): part I. Basic characteristics. Biomaterials 1998;20: 859-77. [25] Freed LE, Vunjak-Novakovic G. Tissue culture bioreactors: chondrogenesis as a model system. In: Lanza RP, Langer R, Chick WL, editors. Principles of tissue engineering. Austin. Texas, USA: R.G. Landes Co, 1997. p. 151-65. [26] Young JH, Teumer J, Kemp PD, Parenteau NL. Approaches to transplanting engineered cells and tissues. In: Lanza RP, Langer R, Chick WL, editors. Principles of tissue engineering. Austin. Texas, USA: R.G. Landes Co., 1997. p. 297-307. [27] Hillsley MV, Frangos JA. Review: bone tissue engineering: the role of interstitial fluid flow. Biotechnol Bioeng 1994;43:573-81. [28] Thompson DE, Agrawal CM, Athanasiou K. The effects of dynamic compressive loading on biodegradable implants of 50-50% polylactic acid-polyglycolic acid. Tissue Eng 1996;2(1):61-74. [29] Dunkelman NS, Zimber MP, LeBaron RG, Pavelec R, Kwan M, Purchio AF. Cartilage production by rabbit articular chondrocytes on polyglycolic acid scaffolds in a closed bioreactor system. Biotechnol Bioeng 1995;46:299-305. [30] Freed L, Vunjak-Novakovic G, Biron RJ, Eagles DB, Lesnoy DC, Barlow SK, Langer R. Biodegradable polymer scaffolds for tissue engineering. Biotechnology 1994;12:689-93. [31] Grande DA, Halberstadt C, Naughton G, Schwartz R, Manji R. Evaluation of matrix scaffolds for tissue engineering of cartilage grafts. J Biomed Mater Res 1997;34:211-20. [32] Martin I, Padera RF, Langer R, Vunjak-Novakovic G, Freed LE. In vitro differentiation of chick embryo bone marrow stromal cells into cartilaginous and bone-like tissues. J Orthop Res 1998;16:181-9. [33] Puelacher WC, Vacanti JP, Ferraro NF, Schloo, Vacanti CA. Femoral shaft reconstruction using tissue engineered growth of bone. Int J Oral Maxillofac Surg 1996;25:223-8. [34] Ma PX, Langer R. Morphology and mechanical function of long-term in vitro engineered cartilage. J Biomed Mater Res 1999;44:217-21. [35] Freed LE, Hollander AP, Martin I, Barry JR, Langer R, VunjakNovakovic G. Chondrogenesis in a cell-polymer-bioreactor system. Exp Cell Res 1998;240(1):58-65. [36] Reddi AH, Wientrob S, Muthukumaran N. Biological principles of bone induction. Orthopedic Clinics North America 1987;18:207-12. [37] Brekke JH. A rationale delivery of osteoconductive proteins (a review). Tissue Eng 1996;2(2):97-114. [38] Chen CS, Milan Mrksich, Sui Huang, Whitesides GM, Ingber D. Geometric control of cell life and death. Science 1997;276:1425-8. [39] Ingber DE. Tensegrity: the architectural basis of cellular mechanotransduction. Ann Rev Physiol 1997;59:575-99. [40] Freed LE, Hollander AP, Martin I, Barry JR, Langer R, VunjakNovakovic G. Chondrogenesis in a cell-polymer-bioreactor system. Exp Cell Res 1998;240(1):58-65.
[41] Reece GP, Patrick Jr CW. Tissue engineered construct design principles. In: Patrick Jr CW, Mikos AG, McIntire LV, editors. Frontiers in tissue engineering. New York, USA: Elsevier Science, 1998. p. 166-96. [42] Vander AJ, Shermann JH, Luciano DS. Human physiology. New York: McGraw-Hill, 1985. p. 341-66. [43] Lu L, Mikos AG. The importance of new processing techniques in tissue engineering. MRS Bull 1996;11:28-32. [44] Thomson RC, Yaszemski MJ, Mikos AG. Polymer scaffold processing. In: Lanza RP, Langer R, Chick WL, editors. Principles of tissue engineering. Austin, TX, USA: R.G. Landes Co., 1997. p. 263-72. [45] Widmer MS, Mikos AG. Fabrication of biodegradable polymer scaffolds for tissue engineering. In: Patrick Jr CW, Mikos AG, McIntire LV, editors. Frontiers in tissue engineering. New York, USA: Elsevier Science, 1998. p. 107-20. [46] Mikos AG, Thorsen AJ, Czerwonka LA, Bao Y, Langer R, Winslow DN, Vacanti JP. Preparation and characterization of poly(L-lactic acid) foams. Polymer 1994;35(5):1068-77. [47] Mikos AG, Sarakinos G, Leite SM, Vacanti JP, Langer R. Laminated three-dimensional biodegradable foams for use in tissue engineering. Biomaterials 1993;14:323-30. [48] Widmer MS, Gupta PK, Lu L, Meszlenyi RK, Evans GRD, Brandt K, Savel T, Gurlek A, Patrick CW, Mikos AG. Manufacture of porous biodegradable polymer conduits by an extrusion process for guided tissue regeneration. Biomaterials 1998;19: 1945-55. [49] Vieth WR. Diffusion in and through polymers: principles and applications. M finchen: Carl Hanser Verlag, 1991. [50] Tom JW, Debenedetti PG. Particle formation with supercritical fluids--a review. J Aerosol Sci 1991;22(5):555-84. [51] Michaeli W, Seibt S. Molding of resorbable polymers at low temperatures. ANTEC 1995. p. 3397-9. [52] Mooney DJ, Baldwin DF, Suh NP, Vacanti JP, Langer R. Novel approach to fabricate porous sponges of poly(D,L-lactic-coglycolic acid) without the use of organic solvents. Biomaterials 1996;17:1417-22. [53] Harris LD, Kim BS, Mooney DJ. Open pore biodegradable matrices formed with gas foaming. J Biomed Mater Res 1998; 42:396-402. [54] Whang K, Thomas CH, Healy KE, Nuber G. A novel method to fabricate bioabsorbable scaffolds. Polymers 1995;36:837-42. [55] Whang K, Tsai DC, Nam EK, Aitken M, Sprague SM, Patel PK, Healy KE. Ectopic bone formation via rhBMP-2 delivery from porous bioresorbable polymer scaffolds. J Biomed Mater Res 1998;42:491-9. [56] Whang K, Healy KE, Elenz DR, Nam EK, Tsai DC, Thomas CH, Nuber MD, Glorieux FH, Travers R, Sprague SM. Engineering bone regeneration with bioabsorbable scaffolds with novel microarchitecture. Tissue Eng 1999;5(1):35-51. [57] Nam SY, Park TG. Porous biodegradable polymeric scaffolds prepared by thermally induced phase separation. J Biomed Mater Res 1999;47(1):8-16. [58] Zhang R, Ma PX. Poly(~-hydroxyl acids)/hydroxyapatite porous composites for bone-tissue engineering. I. Preparation and morphology. J Biomed Mater Res 1999;44(4):446-55. [59] Schugens C, Maguet V, Grandfils C, Jerome R, Teyssie P. Polylactide macroporous biodegradable implants for cell transplantation. 2. Preparation of polylactide foams by liquid-liquid phase separation. J Biomed Mater Res 1996;30:449-61. [60] Lo H, Ponticiello MS, Leong KW. Fabrication of controlled release biodegradable foams by phase separation. Tissue Eng 1995;1:15-28. [61] Hutmacher DW, Teoh SH, Feng SS, Ramakrishna S, Zein I. Tissue engineering research--the engineer's role. Medical Dev J, accepted for publication.
189
The Biomaterials Silver Jubilee Compendium D.W. Hutmacher / Biomaterials 21 (2000) 2529-2543
[62] Rotter N, Aigner J, Naumann A, Planck H, Hammer C, Burmester G, Sittinger M. Cartilage reconstruction in head and neck surgery: comparison of resorbable polymer scaffolds for tissue engineering of human septal cartilage. J Biomed Mater Res 1998;42:347-56. [63] Sittinger M, Reitzel D, Dauner M, Hierlemann H, Hammer C, Kastenbauer E, Planck H. Resorbable polyesters in cartilage engineering: affinity and biocompatibility of polymer fiber structures to chondrocytes. J Biomed Mater Res Appl Biomater 1996;33(2):57-63. [64] Puelacher WC, Vacanti JP, Ferraro NF, Schloo, Vacanti CA. Femoral shaft reconstruction using tissue engineered growth of bone. Int J Oral Maxillofac Surg 1996;25:223-8. [65] Freed LE, Vunjak-Novakovic G. Culture of organized cell communities. Adv Drug Deliver Rev 1998;33(1-2):15-30. [66] Hutmacher DW, Burdet E, Schantz TJ, Robotic micro-assembly fabrication of three-dimensional bioresorbable scaffolds for tissue engineering, in press. [67] Beaman JJ. Backround and definitions. In: Beamann JJ, Barlow JW, Bourell DL, Crawford RH, Marcus HL, McAlea KP, editors. Solid free-form fabrication: a new direction in manufacturing. Boston, USA: Kluwer Academic Publishers, 1997. p. 1-20. [68] Sachs E, Cima M, Williams P, Brancazio D, Cornie J. Threedimensional printing: rapid tooling and prototypes directly from a CAD model. J Eng Industry 1992;114:481-8. [69] Cima LG, Vacanti JP, Vacanti C, Ingber D, Mooney D, Langer R. Tissue engineering by cell transplantation using degradable polymer substrates. J Biom Eng 1991;113:143-51. [70] Park A, Wu B, Griffith LG. Integration of surface modification and 3-D fabrication techniques to prepare patterned poly(L-lactide) substrates allowing regionally selective cell adhesion. J Biomater Sci Polym Ed 1998;9(2):89-110. [71] Wu BM, Borland SW, Giordano RA, Cima LG, Sachs EM, Cima MJ. Solid free-form fabrication of drug delivery devices. J Contr Rel 1996;40:77-87. [72] Giordano RA, Wu BM, Borland SW, Cima LG, Sachs EM, Cima MJ. Mechanical properties of dense polylactic acid structures
[73]
[74]
[75]
[76] [77]
[78]
[79]
[80]
[81]
[82]
2543
fabricated by three-dimensional printing. J Biomater Sci Polym Ed 1996;8(1):63-75. Agarwala MK, Jamalabad VR, Langrana NA, Safari A, Whalen PJ, Danforth SC. Structural quality of parts processed by fused deposition. Rapid Prototyping J 1996;2(4):4-19. Gray IV RW, Baird DG, Bohn JH. Effects of processing conditions on short TCLP fiber reinforced FDM parts. Rapid Prototyping J 1998;1(4):14-25. Buckwalter JA, Mankin HJ, Articular cartilage. Part II: degeneration and osteoarthrosis, repair, regeneration, and transplantation. J Bone Jt Surg 1997;A79(4):612-32. Newman AP. Articular cartilage repair. Amer J Sports Med 1998;26(2):309-24. Glowacki J, Yates K, Little G, Mizuno S. Induced chondroblastic differentiation of human fibroblasts by three-dimensional culture with demineralized bone matrix. Mater Sci Eng 1998; C6:199-203. de Chalain T, Phillips JH, Hinek A. Bioengineering of elastic cartilage with aggregated porcine and human auricular chondrocytes and hydrogels containing alginate, collagen, and •-elastin. J Biomed Mater Res 1999;44:280-8. Lee EH, Chen F, Chan JWK, Bose K. Treatment of growth arrest by transfer of cultured chondrocytes into physeal defects. J Paediatric Orthoped 1998;18(2):155-60. ten Koppel PG, van Osch GJ, Verwoerd CD, Verwoerd-Verhoef HL. Efficacy of perichondrium and a trabecular demineralized bone matrix for generating cartilage. Plast Reconstruct Surg 1998; 102(6):2012-20. Bruns J, Kahrs J, Kampen J, Behrens P, Plitz W. Autologous perichondral tissue for meniscal replacement. J Bone Jt Surg-British Volume 1998;80(5):918-23. Hutmacher DW, Zein I, Teoh SH, Ng KW, Schantz JT, Leahy JC. Design and fabrication of a 3D scaffold for tissue engineering bone. In: Agrawal CM, Parr JE, Lin ST, editors. Synthetic bioabsorbable polymers for implants, STP 1396. American Society for Testing and Materials, West Conshohocken, PA, 2000. p. 152-67.
190
This Page Intentionally Left Blank
The Biomaterials Silver Jubilee Compendium
191
The B i o m a t e r i a l s S i l v e r J u b i l e e C o m p e n d i u m
Biomaterials ELSEVIER
Biomaterials 23 (2002) 3215-3225 www.elsevier.com/locate/biomaterials
Topographical control of human neutrophil motility on micropatterned materials with various surface chemistry Jian Tan, W. Mark Saltzman* School of Chemical Engineering, Cornell University, 120 Olin Hall, Ithaca, NY 14853, USA
Received 29 November 2001; received in revised form 12 February 2002; accepted 20 February 2002
Abstract
Controlling cell responses to an implantable material is essential to tissue engineering. Because the surface is in direct contact with cells, both chemical and topographical properties of a material surface can play a crucial role. In this study, parallel ridges/grooves were micropatterned on glass surfaces using photosensitive polyimide to create transparent substrates. The migratory behavior of live human neutrophils on the patterned surfaces was observed using a light microscope with transmitted light source. The width (2 lam) and length (400 lam) of the ridges were kept constant. The height (5 or 3 lam) and the repeat spacing (6-14 lam) of the ridges were systematically changed to investigate the effect of microgeometry on neutrophil migration. In addition, the effect of surface chemistry on neutrophil migration was studied by deposition of a thin layer of "inert", biocompatible metal such as Au-Pd alloy and titanium on patterned substrates. More than 95% of neutrophils moved in the direction of the long axis of ridges/grooves regardless of the topographical geometry and chemistry, consistent with a phenomenon termed "contact guidance". Therefore, cell migration was characterized using a one-dimensional persistent random walk. The rate of cell movement was strongly dependent on the topographical microgeometry of the ridges. The random motility coefficient/t, 9.8 x 10-9 cm2/s, was the greatest at a ridge height of 5 btm and spacing of 10 lam, about 10 times faster than on smooth glass surface. The Au-Pd coating did not change neutrophil migratory behavior on patterned surfaces, whereas titanium decreased cell motility substantially. The results of this study suggest that optimization of both surface chemistry and topography may be important when designing biomaterials for tissue engineering. In addition, parallel ridges/grooves can be used to control the direction and rate of cell migration on the surface. 9 2002 Published by Elsevier Science Ltd.
1. Introduction
Controlling cell responses to a material is of interest with regard to tissue culture in the laboratory or implantable devices in body [1]. Because the surface is in direct contact with cells, both chemical and topographical properties of the material surface can play a crucial role in determining cell responses. Topographical effects on cell behavior were observed as early as the beginning of last century, when nerve cells were cultured on spider webs and coverslips [2]. Weiss did extensive studies of nerve cell orientation using various substrates including plasma clots [3], glass fibers [4] and fish scales [5] and introduced the concept of "contact guidance" to describe the shape and orientation of cultured cells on such surfaces. However, early studies of the effect of substrate topography on cell behavior lacked precise *Corresponding author. E-mail address."
[email protected] (W.M. Saltzman). 0142-9612/02/$- see front matter 9 2002 Published by Elsevier Science Ltd. PII:S0142-9612(02)00074-1
control of the structure. Microfabrication technology, which has been developed in the electronics industry in recent years, provides tools for producing well-defined and precisely controlled topographical features on various materials. With this technology, investigators have demonstrated that the shape (such as ridges/ grooves, spikes, holes and spirals), the dimension and the distribution of the features can have a significant effect on cell behavior (see reviews [6,7]). In our previous report [8], we directly patterned regular arrays of pillars and holes on glass surfaces using photosensitive polyimide. Our study demonstrated that human neutrophils were almost immobilized by the presence of pillars (2 x 2 x 5 lam) spaced at 10 l.tm. However, migration was significantly enhanced by holes (only 210 nm deep) with a similar distribution and size. In contrast, the focus of the present study was the migration of human neutrophil on microfabricated anisotropic patterns. Many types of cells have been found to react to micron-scale ridges/grooves; the effect
192
The Biomaterials Silver Jubilee Compendium
3216
J. Tan, W.M. Saltzman / Biomaterials 23 (2002) 3215-3225
Table 1 Effect of parallel ridges/grooves on cell behavior Cells
Material
h/d (jam)
w (gm)
Chick heart fibroblasts
Glass
*
2
*
4
4-12
Aligned
36-78
36-78
Aligned
92
100-162
101-162
Not aligned
s (gm)
Result Not aligned
[19]
Human gingival fibroblasts [27]
Epon
Teleost fin mesenchymal cells [28]
Quartz
0.8-1.1
1-4
1-4
Aligned with increasing width
BHK, M D C K and chick embryo neurites [ 2 0 ]
Silicon or ECM protein-coated silicon
0.2-1.9
2-12
2-12
Aligned with increasing depth and alignment dependent on depth
Hippocampal neurons
Quartz
0.014
Not aligned
1.1
Aligned
[21]
Xenopus spinal cord neurons [21]
Quartz
0.014-1.1
1-4
1-4
Aligned
Epithelial tissue and cells
Polystyrene
1 or 5
1-10
1-10
Migration was enhanced along the grooves; more significant effect on deeper grooves
Osteoblasts [29]
Ti or Ca-P-coated silicon
3, 10 or 30
5
42
Aligned and increased bone-like nodule formation
Murine macrophage
Fused silica or ECM proteincoated silica
0.03-0.282
2 or 10
[181
P388D1 [30]
Aligned with increasing depth or decreasing width
Abbreviation used: BHK, baby hamster kidney; MDCK, Madin Darby canine kidney; ECM, extracellular matrix; h, height of ridges; d, depth of grooves; w, width of ridges; s, spacing between ridges; *, data not specified.
on cell behavior is dependent on cell type, substratum chemistry and pattern size (see Table 1). Neutrophils are important immune cells. The accumulation of neutrophils at inflammation sites is the result of directional movement of individual cells. Therefore, it is possible that the alignment of structures within the tissue may facilitate neutrophil migration towards targets in vivo. Neutrophils show significant contact guidance to aligned substratum in vitro; for example, previous studies demonstrated that neutrophils migrated preferentially in the direction of aligned collagen gels and protein-coated grooves on glass, but not on aligned, dried fibrin surfaces [9]. To our knowledge, quantitative analysis of neutrophil migration on parallel grooved surfaces with precisely controlled microgeometry has not been documented. Such knowledge could enable us to guide and optimize neutrophil motility in the development of new biomaterials. In this report, parallel ridges/grooves were micropatterned on glass using a photosensitive polyimide to create transparent substrates. Live human
neutrophil migration on patterned surfaces was directly observed using a light microscope with transmitted light source. The width and length of the ridges were kept constant while the height and the repeat spacing of the ridges were systematically varied to investigate the effect of microgeometry on neutrophil motility (Fig. 1). Chemical composition of a substrate can have a profound effect on cell migration. In this study, the heterogeneous chemistry on patterned surfaces, which were composed of glass and polyimide, may cause cellular responses that interfere with drawing conclusions on a strictly topographical effect. Therefore, modification to produce a homogeneous surface chemical composition was necessary. To accomplish this, a thin film of "inert" and biocompatible metal was sputter-deposited on the patterned surfaces. During sputter deposition, a conformal layer of metal was coated upon the top as well as the side of ridges. The thickness of coating was controlled to be about 10nm so that the substrate was still transparent to optical light and the microgeometry of grooves was not
193
The Biomaterials Silver Jubilee Compendium J. Tan, W.M. Saltzman / Biomaterials 23 (2002) 3215-3225
3217
Briefly, a block of parallel ridges of 4001.tm • 21am spaced with constant edge-to-edge distance (repeat spacing) was designed in CAD. The spacing was varied between 6 and 14 l.tm at an increment of 2 ~tm. A GCA PG 3600F optical pattern generator was used to create images on chrome masks through which many polyimide patterns were produced on glass substrates. Microscope slides (VWR) were cleaned with Nanostrip (Cyantek Corporation Fairmont, CA) and oxygen plasma (Plasma Therm 72 RIE system). Photosensitive Probimide Polyimide 7005 (Olin Microelectronic Materials, Norwalk, CT) was spun on glass at various speed to obtain the desired film thickness (5 or 8 ~tm). The thin film was soft baked in a convective oven to remove the solvent slowly and evenly. Then, the film was exposed through the mask using a GCA 6300 10:1 i-line stepper and post-exposure baked immediately before development. The exposed area was crosslinked and insoluble; the unexposed regions remained soluble and were washed away during development. The patterns were cured in a nitrogen gas-purged oven (Yield Engineering System Polyimide oven) at 350~ for 1 h to convert the Probimide photosensitive polyimide precursor into the final product (polyimide) with a reduced thickness. Surface structure was confirmed by SEM. TM
Fig. 1. SEM images of parallel polyimide ridges patterned on glass surfaces. The magnification for (a) is 3000 and (b-d) is 2000, both scale bars represent 10 ~tm.
altered to any significant degree. An alloy of Au-Pd (with a ratio of 60:40) that is widely used to provide a conductive layer for SEM examination was employed in this study. Both Au and Pd are noble metals and they have been used in prosthetic materials for many years because of their attractive chemical and mechanical properties. In addition, titanium was used in this study to compare neutrophil migration on surfaces with similar microgeometry but different chemistry (glass/ polyimide, Au-Pd vs. titanium). Titanium has been widely used in aircraft and missiles where weight is a prime concern. Recently, it has attracted much attention in orthopedic implantation because of its excellent corrosion resistance under many conditions, which can be attributed to the formation of a passive-oxide surface film. Therefore, studies of neutrophil behavior on Au-Pd alloy and titanium surfaces patterned with ridges/grooves would provide valuable information in the design of orthopedic materials.
2. Materials and methods
2.1. Microfabrication of parallel ridges~grooves on glass substrate Long and thin parallel ridges of photosensitive polyimide were directly patterned on clean glass substrates using standard photolithography techniques at the Cornell Nanofabrication Facility (CNF) [8].
2.2. Characterization of micropatterned surfaces A high magnification scanning electron microscope (SEM, Zeiss LEO2, DSM 982 Field Emission Scanning Electron Microscope) at low voltage (< 5 keV) was used to observe polyimide patterns on glass substrates, which were coated with Au-Pd conductive layer. The width, height, length and spacing between ridges were quantified from these images.
2.3. Neutrophil separation Neutrophils were separated and purified by a centrifugation method described previously [10]. Briefly, about 7ml of fresh human whole blood taken by venipuncture was layered onto 4ml MPRM media (Mono-Poly Resolving Media; ICN Flow, Irvine, CA) and centrifuged at 1700 rpm for 30 min. The lower band containing neutrophils was transferred into a clean tube and underlayered with 1 ml fresh MPRM. Residual erythrocytes were removed by centrifuging with 1 ml underlayered fresh MPRM at 2900 rpm for 15 min. The upper fraction was collected and cells were washed twice with serum-free medium 199 (GibcoBRL, Grand Island, NY).
2.4. Cell morphology of micropatterned surfaces Live cell morphology was first observed using an inverted phase contrast microscope (Diaphot, Nikon,
194 3218
The Biomaterials Silver Jubilee Compendium J. Tan, W.M. Saltzman / Biomaterials 23 (2002) 3215-3225
Garden City, NY). Neutrophils at a density of about 3.0 • 104cells/cm 2 were allowed to settle on micropatterned surfaces for 30min in an incubator at 37~ and the digitized images were captured in the computer using N I H Image 1.61. Next, neutrophils were fixed, dehydrated and dried for SEM examination [11]. Cells on the surfaces were rinsed three times with PBS before being fixed with 2% glutaraldehyde in PBS and dehydrated with a series of graded ethanol solutions. The cells were dried in a critical point dryer (EM 850, Fort Washington, PA). The resulting samples were sputter coated with Au-Pd immediately and examined by SEM (Zeiss LEO2).
2.5. Coating of micropatterned surfaces A thin layer (< 10nm in order to remain transparency) of chemically stable material was deposited on micropatterned glass to control surface chemistry. Either a CVC 601 sputter deposition system or BioRad Polaron sputter coating system (CNF) was used to generate a more uniform and conformal coverage than can be obtained with vapor evaporation technique, e.g. the side walls of parallel ridges were also coated with an equivalent layer of the material. Surfaces were coated with Au-Pd (BioRad Polaron) or Ti (CVC 601).
2.6. Contact angle measurement of surfaces with different compositions Goniometer Model 100-00 (Ram6-Hart, Mountain Lakes, NJ) was used to evaluate the surface energy of various metal-coated surfaces at room temperature. Advancing contact angle (0a) was determined by placing a drop of distilled water (5 ~tl from a microliter syringe) onto the surface and measuring the contact angle between liquid, vapor and substrate.
2.7. Cell motility on micropatterned surfaces using direct visual assay [12] Micropatterned surfaces were used after rinsing with deionized water. Neutrophils (1.5 x 104cells/cm 2) were allowed to settle on patterned substrates in a cell culture plate at 37~ for 15min. The plate was placed on the stage of an inverted light microscope (Diaphot, Nikon) that was maintained at 37+0.5~ throughout the experiment. Cell movement was monitored by a computer-based image analysis system (NIH Image 1.61). Digitized still images of the field of view were collected every 90 s for 15 min. The positions (x, y) of cells were determined on each digitized image by tracing the cell outline and calculating the center of mass. The squared displacement of more than 50 cells, obtained from a variety of healthy donors, was calculated for every possible time interval. From these observations,
more than 95% of cells moved in the direction of parallel ridges (as discussed in Results). The mean squared displacement as a function of time is given by the following equation for a one-dimensional random walk:
D 2(t) - 2 l~(t - P + Pe-t/P),
(1)
where D2(t) is the mean squared displacement over a time interval of length t, ~t is the random motility coefficient, and P is the persistence time. When time is sufficiently large, t~>P, Eq. (1) reduces to D2(t)= 2 # ( t - P). The values of/~ and P were estimated from the slope and intercept of the line; this method for estimating /~ and P from experimental data was described in more detail in a previous publication [12].
3. Results
3.1. Microfabrication and characterization of parallel ridges~grooves Parallel polyimide ridges were successfully patterned on clean glass surfaces (Fig. l a-d). The ridges were 2 gm wide and 400 gm long. These two parameters were kept constant throughout the study. The height (h) was varied--either 5 lain (Fig. l a) or 3 lain (Fig. l b-d)--by altering the speed of rotation during spin coating. The repeat spacing (s) between ridges (both 5 and 3 gm tall) was varied from 6 to 14 lain in increment of 2 gm (in Fig. lb-d, h = 3 gm, only 6, 10 and 14 gm spacings are shown). SEM examination confirmed the scale and reproducibility of the patterns.
3.2. Cell morphology on microfabricated surfaces Neutrophil morphology on surfaces patterned with ridges/grooves was observed using either light microscopy (LM) (Fig. 2) or SEM (Fig. 3). LM observations revealed that cells consistently appeared to be within the grooves, rather than on the top of ridges, regardless of surface chemistry (Fig. 2). Uncoated and Au-Pd-coated surfaces supported similar cell morphology in the patterned region (compare cells within the patterned region of Fig. 2a-b). In contrast, cells on titaniumcoated patterned surfaces appeared more spread and elongated along the ridges (Fig. 2c). SEM studies of fixed cells on patterned surface with different geometry were consistent with LM observations; images of cells on the patterned surface revealed that almost 99% of cells were inside the grooves, regardless of ridge height and spacing (Fig. 3a-f). Cells appeared more elongated and confined in narrower grooves (6 lam) than in wider ones (14 ~tm). For example, most cells contacted two sides of ridges in narrow grooves (Fig. 3a and d) but cells contacted only one
195
The Biomaterials Silver Jubilee C o m p e n d i u m J. Tan, W.M. Saltzman / Biomaterials 23 (2002) 3215-3225
3219
Fig. 2. Light microscope images of neutrophils on microfabricated substrates (h = 5 btm, s = 10 btm) with various surface chemistry.
ridge in wide grooves (Fig. 3c and f). There appeared to be more opportunities for the cell surface to interact with the top of ridges in narrow and shallow grooves (Fig. 3a). The extent of cell-surface contact might be important in modifying cell motility, as described in Discussion. 3.3. Cell motility on microfabricated surfaces
Typical paths of neutrophil movement on surfaces with various microgeometry (h = 5 or 3 lam and s = 6, 10, 141.tm) and chemistry are shown in Figs. 4 and 5. More than 95% of neutrophils moved in the direction of the long axis of ridges/grooves (that is, movement was predominantly in the x-direction with IAyl<41am), regardless of the topographical geometry and chemistry. The few cells that crossed one ridge did not continue motion in the y-direction but moved along the ridge afterwards (in the x-direction). Cells did turn into the opposite direction occasionally (i.e. from moving in + x
to - x direction) but they moved in one direction for relatively long periods. Thus, we employed a onedimensional random walk model to describe cell motion on these anisotropic patterns (cell migration on unpatterned regions of the material was still characterized using a more common 2-D model, see Tan et al. for details [8]). The 1-D directional movement was independent of the geometry of topography and the surface chemistry so that the effect of these factors on neutrophil migration could be accurately compared using the same model. A typical plot of the mean squared displacement, D 2, as a function of time for neutrophil migration on patterned substrate (h = 5 btm, d = 10 pm) is compared with plain glass and polyimide surfaces in Fig. 6; the presence of patterns significantly enhanced cell movement. The random motility coefficient (Table 2) and persistence time (Table 3) for cell migration on each substrate was estimated using the linear region of the curve as described in Materials and Methods.
The Biomaterials Silver Jubilee Compendium
196 3220
J. Tan, W.M. Saltzman / Biomaterials 23 (2002) 3215-3225
Fig. 3. SEM images of neutrophils on microfabricated substrates with various microgeometry.
3.4. Topoyraphical control of neutrophil motility On surfaces with ridges of h -- 5 pm, cell motility was faster in patterned regions than on either smooth glass or polyimide (Fig. 7a). The dependence of cell motility on ridge spacing was biphasic; the random motility coefficient, /~, was the greatest at a spacing of 10 ~tm, about 9.8 x 10-9cm2/s. A 2 ~tm change in spacing (to 8 or 12~tm) decreased the motility about 40% to 5-6 x 10-9cmZ/s (Fig. 7a, Table 2). However, there was no significant difference between motility coefficients as spacing was changed further to 6-14 ~tm. Similar experiments were performed on surfaces patterned with ridges of h - 3 ~tm. Even though the presence of ridges generally increased cell motility compared to smooth regions, cell migration was
considerably slower compared to the surfaces with 5 ~tm tall ridges (Fig. 7a). The cell motility coefficient (5.1 x 10 - 9 c m 2 / S ) w a s the greatest at the narrowest spacing (6 ~tm) and decreased about 40% at a spacing of 8~tm. However, the motility coefficient increased slightly at a spacing of 10 ~tm and decreased slightly with further change in spacing (12 and 14pm) (Table 2, Fig. 7a). The persistence time for neutrophil migration was calculated in order to estimate how frequently cells changed the direction of movement (Table 3). Even though relatively large errors were associated with these calculations, P was about 2 min for patterned surfaces with various spacing and height (except d = 14~tm, h - 3 lain), significantly larger than P on plain glass and polyimide (about 1 min).
197
The Biomaterials Silver Jubilee Compendium J. Tan, W.M. Saltzman I Biomaterials 23 (2002) 3215-3225 500
..,.,,~
2 0 0 0 .....!,
+
"a ~600
400
O L
E
...,.
glass
c~
g 200
Q]m
100 I llllllllllll
"~ "ID
800 ....
u~
4 0 0 ..!.
E
0~
II
011 0
o 1O0
o
(a) h=5 gm
200
300
X positon (micron)
400
d (pm)
b 300
E e-
~1~ ~
200
Smooth 6 8 10 12 14
A 4~
100
o
0
(b) h=3 gm
100
15
Table 2 Random motility coefficient # (10 -9 cm2/s) for neutrophil movements on patterned substrates with various geometry and chemistry
OB
.2
10
Fig. 6. Typical plots of mean squared displacement as a function of time for a patterned surface, smooth glass and polyimide.
400 A e" O
Time (min)
500
500
O e~
"
o~ 1 2 0 0
.9 300
8
patterns polyimide
C:
u~
"E"
3221
200 300 x position (micron)
400
No coating
Au-Pd
Ti
No coating
0.9• 5.5• 6.0• 9.8• 5.2• 5.2•
0.9• 5.5• 5.5• 9.7• 4.7• 3.8•
0.08• 2.4• 2.8• 3.2• 3.2• 2.5•
0.9• 5.1• 2.8• 4.4• 3.3• 1.5•
Table 3 Persistence time P (min) for neutrophil movements on patterned substrates with various geometry and chemistry d (pm)
500 [I-II !11111 II II I
Smooth 6 8 10 12 14
400
i
o 300
h = 3pro
500
Fig. 4. Comparison of cell migration paths on microfabricated surface with different repeat spacing: circle for 6 l.tm; square for 10 pm and triangle for 14 pm.
A C O L
h = 5pro
E
h = 5pm
h = 3pro
No coating
Au-Pd
Ti
No coating
1.0• 1.1• 2.2• 2.5• 2.0• 2.7•
2.5• 2.9• 2.8• 3.5• 2.8• 2.7•
* 2.6• 2.1• 2.3• 2.4• 1.8•
1.0• 2.3• 1.7• 1.8• 2.0• 0.7•
*Larger error, not listed.
C . 0_ 0
200
3.5. Chemical control of neturophil motility on microfabricated surface
cx~Q~ 100
0
100
200
300
x position (micron)
4oo
500
Fig. 5. Comparison of neutrophil migration paths on microfabricated surfaces with different chemistry: (a) circle, Au-Pd coating, (b) square, no coating and (c) triangle, Ti coating. The dimension of the bars was h = 5pm, w = 2pm, s - 10pm.
The chemical composition of the substrate influenced the rate of cell motility. Surface energy was evaluated by water contact angle measurement (Table 4). The advancing angles of smooth glass surface (treated with the same procedure as the patterned one) and Au-Pd-coated surface were almost the same, 47.0 ~ and 45.7 ~ respectively. However, advancing angle was significantly smaller for titanium-coated surfaces, only
The Biomaterials Silver Jubilee Compendium
198
J. Tan, W.M. Saltzman / Biomaterials 23 (2002) 3215-3225
3222
14-
14-
---I-[] O
12 10 E
h=5micron h=3 micron glass polyimide
12 -~
10 ~E o?, o
_
v
6
DO Sm~176 "
A
I
_
v
_,
4 -~
--....
2-
6 Z
o g (a)
~]~
O
O
o
,I, no coating ---I-- AuPd +Ti
, 6
~
,
8 10 Spacing (micron)
2
, 12
0
o F1 Z
14
~
~
o g (b)
a
.
.
.
.
6
8 10 Spacing (micron)
12
14
Fig. 7. Effect of (a) microgeometry of substrates (without any surface coating) and (b) surface chemistry (with h = 5 gm on all surfaces) on neutrophil motility.
Table 4 Advancing water contact angle measurement (0a) Substrate
Glass
Polymide
Au-Pd
Ti
0a
47.0_+ 2.6
73.4_+ 5.0
45.7 _+7.3
13.8 _+1.7
occurred at around 10 and 12 lam and there was only a slight decrease in motility at narrower or wider spacing (Table 2, Fig. 7b).
4. Discussion
13.8 ~ indicating a high surface energy. The contact angle of polyimide, 73.4 ~ was substantially greater than either glass or metal. Regardless of surface treatment, cells moved along the direction of Au-Pd-coated grooves, with < 5% crossing ridges (Fig. 5). The presence of A u - P d did not change the random motility coefficient (0.9 x 10 .9 cm2/s) for cell movement on smooth surface or have any significant effect on neutrophil migration on fabricated surfaces (Table 2 and Fig. 7b). The dependence of random motility coefficient on spacing was biphasic and almost identical to that on un-coated surfaces with the greatest motility (9.7 x 10-9cm2/s) occurring at a spacing of 10gm (Table 2 and Fig. 7b). The persistent time on Au-Pd-coated grooves was similar to uncoated grooves, P > 2 min. The presence of titanium did not effect the direction of cell movement but significantly reduced cell motility and eliminated the biphasic dependence on spacing. On unpatterned, titanium-coated surfaces, cells were almost immobilized with/~ only about 0.08 x 10 .9 cm2/s (Table 2). On patterned surfaces, cells were able to move a short distance (Fig. 5), but the motility was substantially slower than motility observed on uncoated or Au-Pd-coated surface with the same microgeometry (Table 2, Fig. 7b). The effect of spacing on cell motility was not as significant as those on un-coated and Au-Pdcoated surfaces. The greatest motility (3.2 x 10 .9 cm2/s)
Our studies revealed that more than 95% of human neutrophils migrated along the major axis of surfaces with parallel ridges/grooves of various spacing, depth and chemistry; this phenomenon was termed "contact guidance" by Weiss [3]. It has been suggested that neutrophils probably do not form focal contacts with substrata during migration, instead, cell membrane extensions probably determine cell speed [13-16]. On surfaces with the regularly arranged pillars or holes employed in our previous studies, cell membrane extensions occurred freely in both the x- and y-directions; cell motion was observed as a 2-D persistent random walk [8]. However, when cells were moving inside grooves (anisotropic topography) as in the experiments performed here, cell migration was observed along one axis and a 1-D persistence random walk was appropriate to describe such motion. In this situation, membrane extension appeared to be inhibited in the direction perpendicular to grooves and favored in the direction parallel to grooves [17,18]. In addition to guidance of direction of movement, neutrophil motility was enhanced for cells moving inside the grooves with particular characteristics. Our observations are consistent with studies using other cell types [18]. It was also interesting to note that neutrophils did change their direction of movement, but the frequency of directional change was less than that on smooth
The Biomaterials Silver Jubilee Compendium
199
J. Tan, W.M. Saltzman / Biomaterials 23 (2002) 3215-3225
3223
Fig. 8. Comparison of neutrophil motility on surfaces with isotropic and anisotropic features: holes vs. ridges/grooves.
surfaces. The mechanism is not clear at present time; we speculate that it is related to the inefficiency of reorganization of cytoskeleton structure upon contact with an adhesive vertical wall [19]. The speed of movement of a cell population (i.e. the cell motility coefficient) depended on groove spacing (6-14gm) and depth (3 or 5 gm). When the depth of grooves was 5gm, the behavior of cell motility as a function of repeat spacing could be described as biphasic. This biphasic behavior is similar to that observed in our previous studies of neutrophil migration on surfaces patterned with holes, in which the maximal cell motility was also observed at a spacing of ~ 10 gm [16]. However, there were noticeable differences in the cellular response to spacing. On ridge-patterned surfaces, cell motility decreased significantly with a spacing change of _+2~tm from 10gm whereas no significant change was observed with further changes in spacing; on hole patterned surfaces, the variation in motility with spacing was less pronounced (Fig. 8). We speculate that the mechanisms for cell migration on the two patterned substrates must be different. On surfaces with holes, cells appear to use the holes as mechanical edges to
"grab and pull" and the frequency of cell interaction with the edges is directly related to density (spacing) of the holes. However, this mechanism may not apply to neutrophil migration on parallel grooved surfaces, in which mechanical ridges are only available in one direction. SEM images suggest that cell surfaces interact with material of the groove bottom, side wall, and ridge top. The accumulative effect of these interactions probably determines the rate of cell motility. When the spacing was wide, 12 or 14 gin, cells only contacted one groove. Once cells were attached to this groove, they did not appear to dissociate easily. As a result, 12 and 14 gm spaced ridges produce a similar degree of "contact enhancement" for cell migration. When the repeat spacing was narrow (6 or 8 lam), cells were squeezed inside grooves, contacting materials on both sides, which might lead to some degree of restriction upon the cytoskeleton rearrangements necessary for movement [19]. The combination of "contact enhancement" and "restriction" probably produced no further change in cell motility as the spacing was changed from 8 to 6gm. An optimal spacing that was close to a cell diameter (10 lain) was required to achieve maximal cell
The Biomaterials Silver Jubilee Compendium
200 3224
J. Tan, W.M. Saltzman / Biomaterials 23 (2002) 3215-3225
[]
[]
[]
[]
I
"11
II
II
l
I
I
.... I '
I
I
[]
[]
[]
I
II
II
II
I
I
I
I
I
I
[]
[]
[]
I
II
1[
II
I
I
I
I
I
!
[]
[]
[]
(b) r ' - i
J i
i i
i
(c) l
I
I
i
I
Y
(a) L D x
I
t
,
I
I
I
i
i
(d) l ....
I
Fig. 9. Topographical transformation of pillars into ridges/grooves: (a) regularly distributed pillars; (b) pillars become wider in the horizontal direction (x-direction); (c) pillars contact each other in the horizontal direction to form continuous ridges (d).
motility; we speculate that this spacing maximizes contact with grooves but minimizes restrictions. Neutrophil migration was sensitive to groove depth, as observed for other cells (Table 1) [18,20,21]. Neutrophil motility decreased with decreasing groove depth probably due to less "contact enhancement". However, the effect of spacing for 3 ~m high features was not a simple biphasic function of the depth of grooves. We speculate that this behavior may be related to cell interaction with the top of the ridges and such interactions could play a role in modifying cell motility. Cells have more opportunities to interact with the top surfaces of narrower and shallower grooves. As a result, cell motility was the greatest on the surface with narrowest groove (6 ~tm). Surface chemistry was another important factor influencing cell migration speed. Surface energy (hydrophilicity/hydrophobicity) is often used to characterize materials with respect to cell behavior, although no general relationships between surface energy and cell behavior have emerged [22-25]. Contact angle measurements demonstrated that polyimide was more hydrophobic than glass, suggesting the possibility that neutrophils might respond to chemical anisotropy with fast motility on polyimide-patterned surfaces. The deposition of a thin layer of Au-Pd alloy converted these glass/polyimide patterns to an isotropic surface chemistry. Neutrophil migration on Au-Pd-coated surfaces did not differ from those on un-coated surfaces, either smooth or patterned ones. Together, these results indicate that (1) the chemical heterogeneity of polyimide and glass had a negligible impact on cell motility; and (2) the directional movement and the increased motility of neutrophils were indeed caused by the physical-rather than the chemical--patterns in the materials. On titanium-coated surfaces, for which the surface energy was much higher than glass and Au-Pd, cell migration was significantly slower. Other studies have
shown that neutrophil adhesion to metals generally increases with increasing surface energy (decreasing water contact angle) [26]. Therefore, the slower movement of neturophil on titanium-coated surfaces was probably due to stronger adhesion between the cells and the material. Apparently, neutrophil movement was also dependent on surface chemical property. The heterogeneity of polyimide and glass had a small effect on cell migration, probably because cell motility on these two materials was similar, even though the surface energy, as well as cell adhesion, was significantly different as we have previously shown [8]. Because the surface energy and cell motility of Au-Pd was similar to that of plain glass, the dependence of cell motility coefficient on spacing was almost identical on Au-Pd-coated and uncoated substrates patterned with ridges/grooves. Even though cell motility coefficient was increased 30-40 times due to the presence of ridges/grooves on titaniumcoated substrates, the effect of spacing (microgeometry) was not great. These results indicate that both surface topography and chemistry play important roles in the regulation of cell migration, and the effect of microgeometry can be overridden by a strong chemical property, such as the high adhesiveness of titanium. We recognize that protein adsorption is a key mediator in the interaction of cells with surfaces, so we intentionally performed the experiments in serumfree medium. Therefore, the only proteins that were present during cell migration are either trace amounts left from the whole blood during the separation (we estimate this to be less than a few nano gram per square centimeter in each experiment) or protein produced and secreted by the cells. Future experiments should examine the relationship of protein adsorption in the changes that we observe. Previous studies showed that neutrophil migration was significantly hindered by the presence of isotropic pillars [8]. But we note that pillar connected along one
201
The Biomaterials Silver Jubilee Compendium J. Tan, W.M. Saltzman / Biomaterials 23 (2002) 3215-3225
direction become a ridge (Fig. 9). Such topographical manipulation produces a substantial change in cell migration: not only in direction but speed and persistence time were also demonstrated to change in this study. Microfabrication technology provides a powerful tool for achieving precise surface topography. In addition, surface chemistry of a substrate can also influence neutrophil motility; this property can also be manipulated by microfabrication techniques. Therefore, the combination of both surface chemistry and topography should be considered to control and optimize neutrophil migration when designing implantable materials.
Ackno wledgements
We thank the CNF staff for their help in microfabrication technology. We thank Dr. S.H. Kang and Prof. Christopher K. Ober's laboratory for the assistance in contact angle measurement. We thank Hong Shen for many useful discussions, Thomas M. Yung and Sheryl Parker for the help in tracking cell positions in motility study. This work was supported by a grant from the National Science Foundation (BES-9710313) to WMS and it was performed in part at CNF (a member of National Nanofabrication Users Network) which is supported by the National Science Foundation under Grant ECS-9319005, Cornell University and industrial affiliates.
References [1] Saltzman WM. Cell interactions with polymers. In: Lanza RP, Langer R, Vacanti J, editors. Principles of tissue engineering. San Diego: Academic Press, 2000. [2] Harrison RG. The cultivation of tissues in extraneous media as a method of morphogenetic study. Anat Rec 1912;6:181-93. [3] Weiss P. In vitro experiments on the factors determining the course of the outgrowing nerve fiber. J Expt Zool 1934;68:348-93. [4] Weiss P. Experiments on cell and axon orientation in vitro: the role of colloidal exudates in tissue organization. J Expt Zool 1945;100:353-86. [5] Weiss P, Taylor AC. Fish scales as substratum for uniform orientation of cells in virto. Anat Rec 1956;124:381. [6] Curtis A, Wilkinson C. Topographical control of cells. Biomaterials 1997;18:1573-83. [7] Brunette DM, Chehroudi B. The effects of the surface topography of micromachined titanium substrata on cell behavior in vitro and in vivo. J Biomech Eng 1999;121:49-57. [8] Tan J, Shen H, Carter KL, Saltzman WM. Controlling human polymorphonuclear leukocytes motility using microfabrication technology. J Biomed Mater Res 2000;51:694-702. [9] Wilkinson PC, Shields JM, Haston WS. Contact guidance of human neutrophil leukocytes. Exp Cell Res 1982;140:55-62. [10] Tan J, Saltzman WM. Influence of synthetic polymers on neutrophil migration in three-dimensional collagen gels. J Biomed Mater Res 1999;46:465-74.
3225
[11] Glauert AM. Fixation, dehydration and embedding of biological specimens. Amsterdam: North-Holland Publishing Company, 1975. [12] Parkhurst MR, Saltzman WM. Quantification of human neutrophil motility in three-dimensional collagen gels. Effect of collagen concentration. Biophys J 1992;61:306-15. [13] Lackie JM. Aspects of the behaviour of neturophil leukocytes. In: Bellairs R, Curtis A, Dunn G, editors. Cell behaviour. New York: Cambridge University Press, 1982. p. 319-48. [14] Brown AF. Neutrophil and monocyte behavior in three-dimensional collagen matrices. Scanning Electron Microsc 1984;2: 747-54. [15] Mandeville JT, Lawson MA, Maxfield FR. Dynamic imaging of neutrophil migration in three dimensions: mechanical interactions between cells and matrix. J Leukoc Biol 1997;61:188-200. [16] Tan J, Shen H, Saltzman WM. Micron-scale positioning of features influences the rate of polymorphonuclear leukocyte migration. Biophys J 2001;81:2569-79. [17] Dunn GA, Heath JP. A new hypothesis of contact guidance of tissue cells. Expt Cell Res 1976;101:1-14. [18] Dalton BA, Walboomers XF, Dziegielewski M, Evans MDM, Taylor S, Jansen JA, Steele JG. Modulation of epithelial tissue and cell migration by microgrooves. J Biomed Mater Res 2001;56:195-207. [19] Dunn GA. Contact guidance of cultured tissue cells: a survey of potentially relevant properties of the substratum. In: Bellairs R, Curtis A, Dunn GA, editors. Cell behaviour. New York: Cambridge University Press, 1982. [20] Clark P, Connolly P, Curtis AS, Dow JA, Wilkinson CD. Topographical control of cell behaviour: II. Multiple grooved substrata. Development 1990;108:635-44. [21] Rajnicek AM, Britland S, McCaig CD. Contact guidance of CNS neurites on grooved quartz: influence of groove dimensions, neuronal age and cell type. J Cell Sci 1997;110:2905-13. [22] Horbett TA, Waldburger JJ, Ratner BD, Hoffman AS. Cell adhesion to a series of hydrophilic-hydrophobic copolymers studied with a spinning disc apparatus. J Biomed Mater Res 1988;22:384-404. [23] Tamada Y, Ikada Y. Fibrobalst growth on polymer surfaces and biosynthesis of collagen. J Biomed Mater Res 1994;28: 783-9. [24] Hallab NJ, Bundy KJ, O'Connor K, Moses RL, Jacobs JJ. Evaluation of metallic and polymeric biomaterial surface energy and surface roughness characteristics for directed cell adhesion. Tissue Eng 2001;7:55-71. [25] Tegoulia VA, Cooper SL. Leukocyte adhesion on model surfaces under flow: effects of surface chemistry, protein adsorption, and shear rate. J Biomed Mater Res 1999;50:291-301. [26] Nygren H, Hrustic E, Karlsson C, Oster L. Respiratory burst response of peritoneal leukocytes adhering to titanium and stainless steel. J Biomed Mater Res 2001;57:238-47. [27] Brunette DM. Fibroblasts on micromachined substrate orient hierarchically to grooves of different dimensions. Exp Cell Res 1986;164:11-26. [28] Wood A. Contact guidance on microfabricated substrata: the response of teleost fin mesenchyme cells to repeating topographical patterns. J Cell Sci 1988;90:667-81. [29] Perizzolo D, Lacefield WR, Brunette DM. Interaction between topography and coating in the formation of bone nodules in culture for hydroxyapatite- and titanium-coated micromachined surfaces. J Biomed Mater Res 2001;56:494-503. [30] Wojciak-Stothard B, Curtis A, Monaghan W, Macdonald K, Wilkinson C. Guidance and activation of murine macrophages by nanometric scale topography. Exp Cell Res 1996;223: 426-35.
202
This Page Intentionally Left Blank
The Biomaterials Silver Jubilee Compendium
The B i o m a t e r i a l s S i l v e r J u b i l e e C o m p e n d i u m
203
Available online at www.sciencedirect.com
8ClENCE~DIRECT e
ELSEVIER
Biomaterials
Biomaterials 24 (2003) 893-900 www.elsevier.com/locate/biomaterials
Photopolymerized hyaluronic acid-based hydrogels and interpenetrating networks Yong Doo Park, Nicola Tirelli*, Jeffrey A. Hubbell Department of Materials, Institute for Biomedical Engineering, Swiss Federal Institute of Technology and University of Zi;trich, Moussonstrasse 18, CH-8044 Zi;trich, Switzerland
Received 3 May 2002; accepted 6 September 2002
Abstract
Hyaluronic acid (HA) was derivatized with methacrylic esters used for the preparation of hydrogels via photopolymerization. Poly(ethylene glycol) diacrylate (PEG-DA) with a molecular weight of 570 was also used as a comacromonomer to improve elastic modulus and swelling behavior. The hydrogels were readily degraded by hyaluronidase and their mechanical properties could be modulated by HA molecular weight and concentration of PEG-DA. The incorporation of RGD peptides allowed modulation of the HA properties from cell non-adhesive to adhesive. Human dermal fibroblasts were cultured on the RGD, RDG, and nonfunctionalized HA hydrogels for up to 7 d, showing adhesion and proliferation only with incorporated RGD. 9 2002 Elsevier Science Ltd. All rights reserved. Keywords." Hyaluronic acid; Hydrogels; Photopolymerization; RGD; Cell adhesion
I. Introduction
The physiological significance of hyaluronic acid (HA), the linear, very high molecular weight (up to 1-2 million Da) glycosaminoglycan copolymer of D-glucuronic acid and N-acetyl-D-glucosamine that is found in all connective tissues [1], is largely attributable to its unique viscoelastic properties [2]. HA is also known to play a role in promoting cell motility and proliferation [3]. At least three HA cell surface receptors, namely CD44, R H A M M , and ICAM-1 [4-6], have been identified and influence HA activity in processes such as morphogenesis, wound repair, inflammation, and metastasis [7-10]. Much of the attention devoted to HA in the biomaterials field has been motivated based upon its specific chemical properties: (a) it can be obtained in wide range of molecular weights (by controlled hyaluronidase degradation); (b) it is enzymatically remodeled in vivo and in culture in presence of selected cell types (e.g. chondrocytes); (c) it can be functionalized with reactive groups, undergo cross-linking reactions and *Corresponding author. Tel.: + 41-1-632-63-48; fax: + 41-1-632-1214. E-mail address:
[email protected] (N. Tirelli).
produce materials in the form of hydrogels; (d) in the native form, it is substantially non-adhesive to cells [11], but can be functionalized or blended with cell-adhesive materials to tailor its adhesive properties [12]. Unmodified [13-15] and derivatized [16,17] HAs have been used for a variety of clinical applications such as ocular surgery, viscosupplementation for arthritis, wound healing, and plastic surgery, where HA is generally used as anti-adhesive component. In many cases, HA has been chemically modified, exploiting the reactivity of its carboxy and hydroxy groups. The carboxy groups have been modified by esterification with various techniques [18] and hydrogels have been obtained via biscarbodiimide coupling and hydrazide cross-linking [19-22]. The hydroxyl group has also been used for bisepoxide or divinyl sulfone cross-linking [23]. The overall goal of our work was to produce materials, e.g. for cartilage repair, through the synthesis of HA derivatives that can be cured in situ and can then promote cell in-growth; the cross-linked materials should not swell substantially after curing, should possess mechanical properties that are useful in a surgical setting, should be selectively degradable by hyaluronidase, and finally should promote cell adhesion. A mild cross-linking technique is necessary for the in vivo application of such a process; we have chosen
0142-9612/02/$- see front matter 9 2002 Elsevier Science Ltd. All rights reserved. PII: SO 142-96 12(02)00420-9
The Biomaterials Silver Jubilee Compendium
204 894
Y.D. Park et al. / Biomaterials 24 (2003) 893-900
CH20H COONa
0
00
0 NHCOCH. HN
~
+
(EDC) -..--~]~ H~
o
\ -]
HN...
NH 2
0 HO _~
4-o'
CH=OH 0 NHCOCH=
!.
Fig. 1. Scheme of modification of HA using a methacrylating agent. N-(3-amonopropyl)-methacrylamide hydrochloride was mixed with HA in water and EDC was added for the coupling reaction.
photopolymerization, a technique previously used in our laboratory [24]. In this study, HA-based hydrogels were prepared via photopolymerization of pendent methacrylic esters, previously introduced through functionalization of the carboxylic groups (see Fig. 1). Poly(ethylene glycol) diacrylate (PEG-DA) was copolymerized with HA, in order to improve elastic modulus and swelling behavior. HA-based hydrogels were then made cell-adhesive, introducing integrin-binding peptide via Michael-type addition of a cysteine-containing RGD sequence onto the PEG acrylates [25,26]; in this way, only a fraction of the double bonds was functionalized with negligible effect on the mechanical properties, while at the same time a sufficient number of RGD groups was immobilized to influence cell adhesion. The presence of the RGD peptide showed a dramatic effect on fibroblast adhesion and proliferation on the gels.
2. Materials and methods
was then boiled and the white protein precipitate was filtered using 0.45 gm nylon filter. The MW of degraded HA was determined using gel permeation chromatography (GPC) using pullulan as a standard and phosphate buffered saline (PBS, 10 mM in normal saline, pH 7.4) as a mobile phase.
2.3. Methacrylation of HA To a solution of 0.25mmol of degraded or nondegraded HA (based on the mer MW) in 100 ml distilled water were added 0.096g of EDC (0.5mmol) and 0.089 g of N-3-aminopropyl methacrylamide (0.5 mmol). The reaction mixture was incubated for 2 h at pH 6.5. The same amounts of EDC and N-(3-aminopropyl) methacrylamide were added after 2 h and were further incubated for 2 more hours. The solution was filtered through a 0.45gm nylon filter, then dialyzed against 10 mM sodium chloride for 1 d and distilled water for 2 d, and finally lyophilized for 4d to give a white cake of solid methacrylated HA (HA-Ac). The degree of acrylation was examined using 1H-NMR.
2.1. Materials HA sodium salt (MW 1-2 million Da) and N-(3aminopropyl) methacrylamide were purchased from Genzyme Inc. and Polysciences Inc., respectively. Hyaluronidase and N-(3-dimethylpropyl)-N-ethylcarbodiimide hydrochloride (EDC) were from Sigma (St. Louis, MO). Poly(ethylene) glycol diacrylate (MW 570) and N-vinyl pyrrolidone were purchased from Merck.
2.2. Degradation of HA HA sodium salt was dissolved in 100 ml distilled water to a concentration of 0.5 mg/ml. After complete HA dissolution, 1000 U of hyaluronidase was added to the solution and incubated for 16 h. The reaction mixture
2.4. Introduction of RGD peptide into the polymerization mixture The peptides G C G Y G R G D S P G and GCGYGRDGSPG were synthesized using standard solid-phase synthesis. The peptides were dissolved in 10 mM HEPES saline buffer (10 mM in normal saline, pH 7.4). PEG-DA was then added in a 10:1 acrylate/thiol ratio. The reaction mixture was incubated for l h at room temperature, allowing Michael-type addition to take place, with the thiol on the cysteine residues in the peptide as the Michael donor and the acrylate groups on the PEG termini as the Michael acceptors. The partially peptide-functionalized PEG-DA was not isolated, but used for photopolymerization experiments according to the protocol below. Because of the high ratio of
205
The Biomaterials Silver Jubilee Compendium Y.D. Park et al. / Biomaterials 24 (2003) 893-900
acrylates to thiols, only a very small functionalized PEGs will be unreactive tionalization of the PEG diacrylate, peptide will be incorporated dangling functionalized PEG acrylate.
number of the due to difunci.e. almost all from a singly
2.5. Photopolymerization In a typical photopolymerization experiment, 10mg HA-Ac or native HA was dissolved in 0.1 ml of 10 mM HEPES saline buffer (10mM in normal saline, pH 8.0) containing 100mM triethanolamine, 1 mM eosin Y, and 1% w/v N-vinyl pyrrolidone (NVP). A Xenon arc lamp was utilized to provide illumination for l min at 480-520 nm and 75 W/cm 2. To modulate the mechanical properties of the HA-based hydrogel, a variable quantity (generally 0.006 or 0.012mmol) of PEG-DA or peptide-modified PEG-DA mixture was also added before photopolymerization.
2.6. Rheology of gel formation The mechanical properties of the HA-based hydrogels were measured on a 120CVOHR Bohlin Rheometer, using a parallel plate geometry; a quartz bottom plate allowed the use of an optical fiber to perform photopolymerization experiments. In a typical experiment, 50 ~tl of a solution containing the photopolymerization reagents was placed between the plates, at a distance set to 0.1 mm. The frequency of oscillation was set to 10 Hz. During the photopolymerization, the changes of elastic and viscous moduli and of the phase angle were monitored.
2. 7. Swelling of the hydrogels Hydrogels samples with different HA-Ac MW, acrylation degree, and concentration of PEG-DA were prepared in Eppendorf tubes. The wet weight was measured, incubating the gels in water overnight at room temperature for swelling. The swelling ratio was measured by comparing change of weight of hydrogel before and after incubation.
895
merization and subsequence initiation of the rheological study was kept short relative the kinetics of degradation. In a different set of experiments the degradation products were studied using GPC analysis. The gels were prepared and incubated in HEPES saline buffer (10mM in normal saline, pH 8.0) overnight for swelling, then placed in 5 ml tubes, containing 200~tl HEPES saline buffer (10 mM in normal saline, pH 8.0). Hyaluronidase (1000 U) was added to these solutions and incubated at 37~ for 2 d. Samples of the solutions were collected at regular intervals and analyzed.
2.9. Cellculture Three different types of hydrogel were prepared, namely one with no adhesion peptide (as a control), one with the RGD-containing peptide, and one with a nonadhesive peptide, namely with the inactive sequence R D G (as a second control). Hydrogels were incubated overnight to swell in PBS. Human dermal fibroblasts (5000 cells/well) were seeded on hydrogels in D M E M with 10% fetal bovine serum. Cells were cultured in the same medium in a humidified incubator (5% CO2, 37~ Cell adhesion was documented by photomicrography for periods up to 1 week. The proliferation of fibroblasts on the hydrogel was measured using WST-1 kit, in which the derivatives of tetrazolium are formed and converted to colored formazan by mitochondrial dehydrogenases (Roche, Indianapolis, IN). Briefly, fibroblasts (5000 cells/well) were seeded in a 96-well tissue culture plates, which were already covered with the different types of pre-swelled hydrogels. Cells were incubated for 3 d in DMEM with 10% fetal bovine serum at 37~ in a humidified incubator (5% CO2, 37~ After incubation, 10~tl of cell proliferation reagent WST-1 was added to each well. The reagent was also added to the wells containing hydrogel without cells for the control. The culture plates were incubated for 3 h in a humidified incubator. After incubation, the multiwell plates were shaken 1 min on a shaker. The formation of red formazan was measured at 420 nm using ELISA reader.
3. Results and discussion
2.8. Enzymatic degradation of hydrogels 3.1. Modification of HA In a rheological study, 100U of hyaluronidase was added to 100 ~tl of solution before polymerization. The solution was photopolymerized in the rheometer as already described at a pre-set temperature of 37~ After illumination, the evolution of the elastic and viscous moduli was monitored for 10min. The enzyme was incorporated before gelation to prohibit gradients of the enzymes within the hydrogel; the time between mixing of the enzyme with the HA-Ac solution and photopoly-
HA of MW 1-2 million Da was used as a starting material, and the molecular weight of this material was reduced enzymatically by treatment with hyaluronidase. Analysis by GPC demonstrated reduction to an MS of approx. 50kDa after 16h enzymatic treatment (see Fig. 2). Both native and enzymatically degraded HA were subsequently used for chemical modification. Both HAs
206
896
The Biomaterials Silver Jubilee C o m p e n d i u m Y.D. Park et al. / Biomaterials 24 (2003) 893-900
were modified by using N-(3-aminopropyl) methacrylamide as an acrylating agent and the water soluble EDC as a coupling agent, incubating the mixture for 4 h at room temperature. The degree of acrylation of roughly 10% was determined by 1H-NMR, comparing the singlet peak of methyl group in HA acetamide (2.7 ppm) and the multiplet peaks of the acrylic double bond (5.3 and 5.6 ppm). 3.2. Gel preparation and mechanical properties
In the photopolymerization experiments, eosin Y was used as a visible light sensitizer, triethanolamine was employed as an initiator, and PEG-DA and NVP were used as comacromonomer and comonomer, respectively, according to a technique developed in our laboratory for hydrogel synthesis [27]. The use of a low MW PEG-DA allows the material to be toughened and permits reduction of its hydrophilicity and degree of
Fig. 2. GPC results of HA before and after treatment with hyaluronidase. Hyaluronidase was added to the HA solution at a concentration of 10U (A), 50U (11), and 100U (O) per mg HA. The reaction mixture was incubated for 20 h.
swelling (PEG hydrophilicity increases with MW), without reduction in biocompatibility; NVP was incorporated for the primary reason of enhancing the rate of the gelation reaction. After 1 min of illumination with a Xenon arc lamp, clear and soft hydrogels were obtained. Both HA-Ac and native HA were used in gel formation generating respectively, a cross-linked copolymeric network and a pseudo-interpenetrating polymer network (pIPN: one domain is a network and the other a polymer dispersed in the network; a real IPN is constituted by two or more networks), in which the native HA is present within the network as a physically entrapped non-cross-linked polymer. The gel points (defined as crossing points between viscous and elastic modulus, see Fig. 3) were detected in both cases; the rate of gelation of HA-Ac and P E G D A was faster than was observed with the formation of the pIPN, i.e. with HA and PEGDA. Covalent incorporation of HA-Ac within the gel was demonstrated by the fact that the complex modulus reached a plateau after 1 min and stabilized at around 10kPa, whereas the pIPN gel containing native HA yielded a complex modulus of approximately 1 kPa after the same duration of irradiation. In exploration of hydrogel preparation, we focused our attention on some of the factors affecting mechanical strength and post-gelation swelling, namely MW of the HA-Ac and degree of acrylation of the derivatized HA, as well as concentration and composition of the comacromonomer mixture. As expected (see Fig. 4), the complex modulus increased as the number of reactive groups in multifunctional monomers were increased, i.e. with the degree of HA acrylation (0% vs. 10%), and the PEG-DA concentration (2% vs. 5% (w/v)); furthermore, higher HA-Ac MW (1-2 million Da) resulted in gels with a higher complex modulus than was observed in gels formed from lower M W HA-Ac (50kDa), namely 5 x 104 vs. 106 Pa. The equilibrium extent of swelling of the hydrogel may be of significant interest in two regards: first, the
Fig. 3. Viscous and elastic modulus evolution during photopolymerization. Panel A: native HA with 5% PEG-DA. Panel B: HA-Ac with 5% PEG-DA.
207
The Biomaterials Silver Jubilee Compendium Y.D. Park et al. / Biomaterials 24 (2003) 893-900
Fig. 4. Influence of the hydrogel composition on the mechanical properties. The complex modulus of the hydrogel was measured by changing the acrylation (solid bar; acrylated HA, hatched bar; native HA), molecular weight of HA (degHA; 50,000, and HA; 500,000) and PEG-DA concentration (2% and 5%).
mechanical and cell invasion characteristics of the hydrogel would be expected to relate to swelling, and second it may be desirable to limit the post-gelation swelling, to enable a surgeon to apply the material by photopolymerization in a size and shape that will be retained in vivo, after an opportunity for equilibration from the as-gelled state. It should be possible to manipulate both equilibrium swelling and post-gelation incremental swelling, in that they should be dependent upon cross-linking density and hydrophobicity, as well as the concentration of precursors in the macromonomer solution. The absolute level of swelling is in a general sense related to the complex modulus results: the higher the modulus, the higher the cross-linking density and the lower the swelling. An exception is that an increase in the MW of HA increases the cross-linking density, but also the internal osmotic pressure (the volume occupied is exponentially increasing with the MW [28,29]), and the second factor can overwhelm the first one. In fact, the lowest swelling was shown by the formulation having the higher cross-linking density compatible with a low MW of HA, that is highest PEG-DA content, lowest MW of HA-Ac (see Fig. 5). The final aspect of gel synthesis consisted of the introduction of a cell-adhesive peptide, comprising the RGD sequence, and this was accomplished by the means of a Michael-type addition reaction to attach the peptide to one end of a PEG-DA chain, the other acrylate terminus then remaining to participate in the photopolymerization reaction. The peptide was designed so as to contain a cysteine residue, bearing a thiol group that is able to rapidly react with the acrylic groups in PEGDA at pH 7.4. This scheme for incorporation of
897
Fig. 5. Influence of the hydrogel composition on the swelling behavior. The swelling of the hydrogel was measured as a function of the acrylation, molecular weight of HA (degHA; 50,000, and HA; 500,000) and PEG-DA concentration (2% and 5%).
adhesion peptides into the precursors of photopolymerizable gels has been described in detail elsewhere [25]; it has been previously demonstrated that the rate of the Michael-type addition reaction is substantially higher than the rate of the competitive disulfide bonding reaction under these reaction conditions. After incubation of the cysteine-containing peptide with an excess of PEG-DA (in order to obtain a mixture mostly constituted by PEG-DA with some singly coupled chains, with one acrylate group remaining), assuming the reaction to be quantitative approximately 10% of acrylic groups were functionalized with peptides. The remaining acrylate groups in the peptide-grafted PEG monoacrylate (the minor component) and the remaining PEG-DA (the major component) were then used for the photopolymerization reaction. The mechanical properties of the RGD-containing hydrogels were statistically indistinguishable from the hydrogels not containing the RGD-peptide (data not shown), presumably because such a small fraction of the acrylate groups had been consumed by the Michael-type addition reaction that was used to incorporate the peptide.
3.3. Deoradation of HA-based hydrogels The acrylation of HA and the presence of nondegradable PEG-DA introduce new features on the HA backbone, possibly interfering with the enzymatic degradation. The effect of hyaluronidase on the HAAc, measured by GPC, and on the formed hydrogels, monitored theologically and by characterization of the MW of released soluble products, was examined. The hyaluronidase degradation rate of the acrylated polymer
208
The Biomaterials Silver Jubilee Compendium
898
Y.D. Park et al. / Biomaterials 24 (2003) 893-900
(HA-Ac) was somewhat slower, by about 70%, than was observed with native HA; the derivatization may cause some steric hindrance to the enzyme. In the rheological characterization of the cross-linked gels, the change of complex modulus was monitored by incubating the hydrogels at 37~ with an excess of hyaluronidase (100U/50 ~tl solution) inside; the enzyme was mixed in the sample before polymerization, and it could safely be assumed to produce a negligible effect over the time scale of photopolymerization. In a typical experiment
10000
mmm mn mm
13..
v
(,t)
1000
3.4. Cell behavior
_= '--I
mi
0
E
mm:
x
E
(see Fig. 6), the complex modulus increased from roughly 200Pa for the precursor solution to 10kPa during a 1 min photopolymerization. After illumination for 4 min, the value started fluctuating because of loss of mechanical integrity, and decreased down to 500Pa. After 10 min, the value reached about 20 Pa indicating a complete degradation of the gel. GPC analysis was performed on the degradation products of HA-Ac hydrogels formed with and without the incorporation of PEG-DA in the precursor mixture, after incubation of the resulting gel with hyaluronidase (1000 U). The same retardation effect observed in the enzymatic degradation of HA-Ac seems to occur also in the case of the pIPN, but the somewhat slower degradation did not give products of appreciable different molecular weight (see Fig. 7).
9
mm
100
l~/llil
0
nm n
c)
mm 9 mUm
I
I
0
200
,
I
'
400
I
600
;
mUunm_ 9
I
800
'
I
1000
Time (sec) Fig. 6. Rheological study of the preparation and degradation of a hyaluronidase-containing hydrogel. Hyaluronidase (100 U) was added to the 100mg/ml HA-Ac solution containing the photopolymerizing reagents. The gel was obtained via illumination for 1 min.
Human dermal fibroblasts were cultured on the surface of hydrogels containing no peptide (as a control), the active RGD peptide, or the inactive peptide RDG (as an additional control, with similar overall physicochemical properties of the RGD peptide, but without any cell-binding bioactivity). After washing and swelling, fibroblasts were seeded on the three different hydrogel surface, respectively. Cells on the HA-Ac gel and RDG peptide-containing hydrogel did not spread at all after 1 or 2 d in culture. By contrast, cells on RGDcontaining hydrogel spread well, comparably to human dermal fibroblasts on tissue culture plates, and
Fig. 7. Molecular weight determination of the degradation products released from HA-based hydrogels. Preswelled hydrogels (wet weight 0.1 g) were incubated by adding 1000 U hyaluronidase in 200 gl HEPES buffer. After collecting the sample (50 gl), the molecular weight of the degraded product was measured by GPC.
209
The Biomaterials Silver Jubilee Compendium Y.D. Park et al. / Biomaterials 24 (2003) 893-900
899
Fig. 9. Proliferation of fibroblasts as determined by mitochondrial activity. Human dermal fibroblasts were cultured for 3d on four different substrates: (a) reagent only (blank), (b) HA-Ac hydrogels (negative control), (c) RDG-containing hydrogels (negative control), (d) RGD-containing hydrogels, and (e) tissue culture polystyrene (positive control) in 96 well plates. After adding 10 Ftl WST-1 reagent, culture plates were incubated for 3 h and formation of red formazan was measured at 420 nm using an ELISA plate reader.
Fig. 8. Fibroblasts cultured on different HA-based hydrogels. Human dermal fibroblasts were cultured for 7 d on three different substrates: acrylated HA-based hydrogel (a), RDG-modified HA-based hydrogel (b), and RGD-modified HA-based hydrogel (c).
useful for anti-adhesive treatments [11,30]. The cell proliferation on the RGD hydrogel was by contrast found to be around 70% of that observed on the positive control material, namely tissue culture-modified polystyrene (see Fig. 9). HA-based gels containing the inactive RDG peptide offered a much less favorable condition for cell proliferation, with about one-half of the proliferation observed with the active RGD peptide.
4. Conclusions
proliferated and eventually covered the surface of the hydrogel throughout a 7 d period of culture (Fig. 8). Cells on the RGD-containing hydrogel were stained with rhodamine-phalloidin to label F-actin stress fibers of the cytoskeleton, which would be expected if cell spreading morphology were normal. The observed stress fiber morphology was qualitatively consistent with that observed on tissue culture plates and was, by contrast, lacking in cells cultured on RDG-containing hydrogels (data not shown). Cell proliferation was measured after the third day of culture, using the cell proliferation assay reagent WST1, which reflects the mitochondrial dehydrogenase activity of the cell. Cells seeded on tissue culture plates or HA-Ac hydrogels (lacking peptides) were used as a positive and negative control, respectively. The HA hydrogel did not provide conditions that were suitable for cell proliferation. Indeed, HA is known from the literature to provide cell-resistant surfaces and to be
A process for the in situ and possibly in vivo preparation of hydrogels based on HA has been developed. The rationale for the use of this natural polymer was its well-known biocompatibility and very low protein absorption, which make it an excellent antiadhesive material and a kind of natural and enzymatically degradable analog of PEG. The goals that have been achieved in this study are that (a) the functionalization with methacrylic groups and an appropriate formulation of the comacromonomer mixture did produce hydrogels, using photopolymerization; (b) with this technique, a polymer solution can be applied to any shape and form according to a defect region and after a short illumination can produce a gel with negligible post-gelation swelling and interesting mechanical properties; (c) the gels obtained with this technique were demonstrated to be still degradable by hyaluronidase; in other words, HA was an essential part of the network. These HA-based hydrogels showed the
210 900
The Biomaterials Silver Jubilee Compendium Y.D. Park et al. / Biomaterials 24 (2003) 893-900
typical anti-adhesive properties of HA materials; in our case, fibroblasts on the gel surface did not spread at all and their proliferation was as low as control, blank values. It seems that H A alone cannot provide a sufficient signal for cell adhesion and spreading, although may cell types do possess a HA receptor [31]. Finally, the introduction of an R G D peptide showed dramatic changes in terms of cell adhesion and proliferation. Cells cultured on the RGD-modified hydrogel proliferated and grew to confluence, whereas cells on the R D G control hydrogel did not show any sign of spreading over long durations in culture and in the presence of serum proteins.
References [1] Laurent TC, Laurent UB, Fraser JR. The structure and function of hyaluronan: an overview. Immunol Cell Biol 1996;74:A1. [2] Lee B, Litt M, Buchsbaum G. Rheology of the vitreous body: Part 3. Concentration of electrolytes, collagen and hyaluronic acid. Biorheology 1994;31:339-51. [3] Trochon V, Mabilat C, Bertrand P, Legrand Y, Smadja-Joffe F, Soria C, Delpech B, Lu H. Evidence of involvement of CD44 in endothelial cell proliferation, migration and angiogenesis in vitro. Int J Cancer 1996;66:664-8. [4] Entwistle J, Hall CL, Turley EA. HA receptors: regulators of signalling to the cytoskeleton. J Cell Biochem 1996;61:569-77. [5] Yang B, Zhang L, Turley EA. Identification of two hyaluronanbinding domains in the hyaluronan receptor RHAMM. J Biol Chem 1993;268:8617-23. [6] Lesley J, Hyman R, Kincade PW. CD44 and its interaction with extracellular matrix. Adv Immunol 1993;54:271-335. [7] Chen WY, Abatangelo G. Functions of hyaluronan in wound repair. Wound Repair Regen 1999;7:79-89. [8] Menzel EJ, Farr C. Hyaluronidase and its substrate hyaluronan: biochemistry, biological activities and therapeutic uses. Cancer Lett 1998;131:3-11. [9] Lesley J, Hyman R, English N, Catterall JB, Turner GA. CD44 in inflammation and metastasis. Glycoconj J 1997;14:611-22. [10] King SR, Hickerson WL, Proctor KG. Beneficial actions of exogenous hyaluronic acid on wound healing. Surgery 1991;109: 76-84. [11] Pavesio A, Renier D, Cassinelli C, Morra M. Anti-adhesive surfaces through hyaluronan coatings. Med Device Technol 1997; 8(20-1):24-7. [12] Liu LS, Thompson AY, Heidaran MA, Poser JW, Spiro RC. An osteoconductive collagen/hyaluronate matrix for bone regeneration. Biomaterials 1999;20:1097-108. [13] Balazs EA, Denlinger JL. Clinical uses of hyaluronan. Ciba Found Symp 1989;143:265-75.
[14] Wen DY. Intra-articular hyaluronic acid injections for knee osteoarthritis. Am Fam Physician 2000;62:565-70. 572. [15] Rosier RN, O'Keefe RJ. Hyaluronic acid therapy. Instr Course Lect 2000;49:495-502. [16] Vercruysse KP, Prestwich GD. Hyaluronate derivatives in drug delivery. Crit Rev Ther Drug Carrier Syst 1998;15:513-55. [17] Radice M, Brun P, Cortivo R, Scapinelli R, Battaliard C, Abatangelo G. Hyaluronan-based biopolymers as delivery vehicles for bone-marrow-derived mesenchymal progenitors. J Biomed Mater Res 2000;50:101-9. [18] Campoccia D, Doherty P, Radice M, Brun P, Abatangelo G, Williams DF. Semisynthetic resorbable materials from hyaluronan esterification. Biomaterials 1998;19:2101-27. [19] Prestwich GD, Marecak DM, Marecek JF, Vercruysse KP, Ziebell MR. Controlled chemical modification of hyaluronic acid: synthesis, applications, and biodegradation of hsdrazide derivatives. J Controlled Rel 1998;53:93-103. [20] Benedetti L, Cortivo R, Berti T, Berti A, Pea F, Mazzo M, Moras M, Abatangelo G. Biocompatibility and biodegradation of different hyaluronan derivatives (Hyaff) implanted in rats. Biomaterials 1993;14:1154-60. [21] Kuo JW, Swann DA, Prestwich GD. Chemical modification of hyaluronic acid by carbodiimides. Bioconj Chem 1991;2:232-41. [22] Tomihata K, Ikada Y. Crosslinking of hyaluronic acid with water-soluble carbodiimide. J Biomed Mater Res 1997;37:243-51. [23] Larsen NE, Pollak CT, Reiner K, Leshchiner E, Balazs EA. Hylan gel biomaterial: dermal and immunologic compatibility. J Biomed Mater Res 1993;27:1129-34. [24] Hill-West JL, Chowdhury SM, Slepian MJ, Hubbell JA. Inhibition of thrombosis and intimal thickening by in situ photopolymerization of thin hydrogel barriers. Proc Natl Acad Sci USA 1994;91:5967-71. [25] Elbert DL, Hubbell JA. Conjugate addition reactions combined with free-radical cross-linking for the design of materials for tissue engineering. Biomacromolecules 2001;2:430-41. [26] Lutolf M, Tirelli N, Cerritelli S, Cavalli L, Hubbell JA. Systematic modulation of Michael-type reacti,~ity of thiols through the use of charged amino acids. Bioconi Chem 2001; 12:1051-6. [27] Hubbell JA. Bioactive biomaterials. Curr Opin Biotechnol 1999; 10:123-9. [28] Kobayashi Y, Okamoto A, Nishinari K. Viscoelasticity of hyaluronic acid with different molecular weights. Biorheology 1994;31:235-44. [29] Bothner H, Wik O. Rheology of hyaluronate. Acta Otolaryngol Suppl 1987;442:25-30. [30] Schier F, Danzer E, Bondartschuk M. Hyaluronate, tetrachlorodecaoxide, and galactolipid prevent adhesions after implantation of Gore-Tex and dura mater into the abdominal wall in rats. Pediatr Surg Int 1999;15:255-9. [31] Hu M, Sabelman EE, Lai S, Timek EK, Zhang F, Hentz VR, Lineaweaver WC. Polypeptide resurfacing method improves fibroblast's adhesion to hyaluronan strands. J Biomed Mater Res 1999;47:79-84.
The Biomaterials Silver J u b i l e e C o m p e n d i u m
211
Available online at www.sciencedirect.com
SCIENCE~DIRECT e
ELSEVIER
Biomaterials
Biomaterials 24 (2003) 2309-2316 www.elsevier.com/locate/biomaterials
Cell sheet engineering for myocardial tissue reconstruction Tatsuya Shimizu, Masayuki Yamato, Akihiko Kikuchi, Teruo Okano* Institute of Advanced Biomedical Engineering and Science, Tokyo Women's Medical University, 8-1 Kawada-cho, Shinjuku-ku, Tokyo 162-8666, Japan
Received 28 October 2002; accepted 9 December 2002
Abstract
Myocardial tissue engineering has now emerged as one of the most promising treatments for the patients suffering from severe heart failure. Tissue engineering has currently been based on the technology using three-dimensional (3-D) biodegradable scaffolds as alternatives for extracellular matrix. According to this most popular technique, several types of 3-D myocardial tissues have been successfully engineered by seeding cardiomyocytes into poly(glycolic acid), gelatin, alginate or collagen scaffolds. However, insufficient cell migration into the scaffolds and inflammatory reaction due to scaffold biodegradation remain problems to be solved. In contrast to these technologies, we now propose novel tissue engineering methodology layering cell sheets to construct 3-D functional tissues without any artificial scaffolds. Confluent cells on temperature-responsive culture surfaces can be harvested as a viable contiguous cell sheet only by lowering temperature without any enzymatic digestions. Electrical communications are established between layered cardiomyocyte sheets, resulting in simultaneous beating 3-D myocardial tissues. Layered cardiomyocyte sheets in vivo present long survival, macroscopic pulsation and characteristic structures of native heart tissue. Cell sheet engineering should have enormous potential for fabricating clinically applicable myocardial tissues and should promote tissue engineering research fields. 9 2003 Elsevier Science Ltd. All rights reserved. Keywords." Myocardial tissue engineering; Cell sheet; Cardiac myocyte; Transplantation; Temperature-responsive culture surface
1. Introduction
Recently, alternative treatments for cardiac transplantation have been strongly requested to repair damaged heart tissue, because the utility of heart transplantation is limited by donor shortage. Cell therapy is now considered to be one of the most effective treatments for impaired heart tissue [1,2]. Direct transplantation of cell suspension has been researched since the early 1990s [3]. In these studies, survival of transplanted cells, integration of native and grafted cells, and improvement of host cardiac function have been reported. It is a critical point how to isolate and expand clinically transplantable myocardial cell source. Autologous myoblast transplantation has been performed clinically and the contraction and viability of grafted myoblasts have been confirmed [4]. Multipotent bone marrow cells or embryonic stem cells have been
*Corresponding author. Tel.: + 81-3-3353-8111x30234; fax: + 81-33359-6046. E-mail address."
[email protected] (T. Okano).
now aggressively investigated as possible candidates for human implantable myocardial cell source [5-8]. In direct injection of dissociated cells, it is difficult to control shape, size and location of the grafted cells. Additionally, isolated cell transplantation is not enough for replacing congenital defects. To overcome these problems, research on fabricating three-dimensional (3D) cardiac grafts by tissue engineering technology has also now begun [9]. Tissue engineering has currently been based on the concepts that 3-D biodegradable scaffolds are useful as alternatives for extracellular matrix (ECM) and that seeded cells reform their native structure in according to scaffold biodegradation [10]. This context has been used for every type of tissue. In myocardial tissue engineering, poly(glycolic acid) (PGA), gelatin and alginate have been used as prefabricated biodegradable scaffolds. Papadaki et al. engineered 3-D cardiac constructs by using PGA scaffolds processed into porous meshes and rotating bioreactors [11]. Li et al. have demonstrated that transplantation of tissue-engineered cardiac grafts using biodegradable gelatin sponges replaced myocardial scar and right ventricular outflow track defect [12,13].
0142-9612/03/S-see front matter 9 2003 Elsevier Science Ltd. All rights reserved. doi: 10.1016/S0142-9612(03)00110-8
212 2310
The Biomaterials Silver Jubilee Compendium T. Shimizu et al. / Biomaterials 24 (2003) 2309-2316
Furthermore, Leor et al. reported that bioengineered heart grafts using porous alginate scaffolds attenuated left ventricular dilatation and heart function deterioration in myocardial infarction model [14]. As the technique premixing cells and ECM alternatives instead of seeding cells into preformed scaffolds, Zimmermann et al. engineered 3-D heart tissue by gelling the mixture of cardiomyocytes and collagen solution [15,16]. The construct has allowed direct measurement of isometric contractile force as heart tissue model. In spite of these desirable results, insufficient cell migration into scaffolds and inflammatory reaction due to scaffold biodegradation remain problems to be solved [13,14]. In native myocardial tissue, cells are considerably dense (Fig. 1A) in comparison with other tissues including cartilage, vascular, and heart valve, which are cell-sparse tissues and have been successfully engineered by using biodegradable scaffolds (Fig. 1B). Cardiomyocytes are also tightly interconnected with gap junctions, which mediated the reciprocal exchange of small molecules and ions resulting in electrically synchronous beating [17]. In myocardial tissue engineering, biodegradable scaffolds themselves attenuate cell-to-cell connections and scaffold biodegradation leads to fibrous tissues containing excessive amount of ECM, which is shown in pathological states including ischemic heart disease or dilated cardiomyopathy. Investigators are now trying to fabricate more porous structure of biodegradable scaffolds and to develop new techniques seeding more cells into the scaffolds. In particular, structural balance between cells and ECM should be controlled to fabricate native heart-like tissues. By contrast, we now propose novel tissue engineering methodology that is to construct 3-D functional tissues by layering 2-D cell sheets without any biodegradable alternatives for ECM. To obtain viable cell sheets, we have exploited intelligent culture surfaces, from which cultured cells detach as a cell sheet simply by reducing temperature. In this paper, we present the new technology "cell sheet engineering" and its application to myocardial tissue reconstruction.
2. Temperature-responsive culture surfaces
Temperature-responsive culture surfaces were developed among the research to control cell adhesion to biomaterials. Cells adhere to culture surfaces via membrane receptors and cell adhesive proteins, including fibronectin, that reside in serum or are secreted from the cells in culture (Fig. 2A). The interaction between adhesive proteins and culture surfaces depends on the wettability of the surface. Normal tissue culture polystyrene (TCPS) dishes are hydrophobic and absorb ECM proteins resulting in cell attachment and proliferation. To harvest cells from the surfaces, enzymatic
digestion including trypsin and dispase are usually utilized. In that case, both adhesive proteins and membrane receptors are disrupted, then cells detach with considerable damages (Fig. 2B). On the other hand, we graft temperature-responsive polymer, poly(N-isopropylacrylamide)(PIPAAm) to TCPS dishes covalently by electron beam. The surfaces are hydrophobic and cells adhere and proliferate under culture condition at 37~ By lowering temperature below 32~ the surfaces change reversibly to hydrophilic and not cell adhesive due to rapid hydration and swelling of the grafted PIPAAm. This unique surface change allows cultured cells to detach spontaneously from these grafted surfaces simply by lowering temperature [18]. As against using enzymatic digestion, only the interaction between adhesive proteins and material surfaces is released and cells detach together with intact membrane proteins and adhesive proteins (Fig. 2C) [19]. As a result, cells recovered by using PIPAAm-grafted surfaces maintain their differentiated functions more strongly than the cells recovered by protease digestion [20]. For example, trypsin-treated hepatocytes decrease albumin production, on the other hand, those cells harvested from PIPAAm-grafted surface preserve albumin secretion [21]. In addition to the passive mechanism of the surface change from hydrophobic to hydrophilic, cell-mediated active processes have been ascertained as cell detachment mechanisms [22]. Sodium azide, an ATP synthesis inhibitor, considerably retarded cell release from PIPAAm-grafted surfaces, indicating that energy-dependent metabolic process is one of major mechanisms. The active processes are also mediated by intracellular signal transduction, including tyrosine phosphorylation and cytoskeltal reorganization and lead to the cell morphological change from spread to round after surface property change [23].
3. Cell sheet engineering
When cells are cultured confluently, they connect to each other via cell-to-cell junction proteins and ECM (Fig. 3A). With enzymatic digestions, these proteins are disrupted and each cell is released separately (Fig. 3B). In the case using PIPAAm-grafted surfaces, cell-to-cell connections are not disrupted and cells are harvested as a contiguous cell sheet by decreasing temperature (Fig. 3C). Furthermore, adhesive proteins underneath cell sheets are also maintained and they play a desirable role as an adhesive agent in transferring cell sheets onto other culture materials or other cell sheets [24]. These viable cell sheets are composed of cells and biological ECM without any artificial scaffolds. Various types of cell sheets have been successfully lifted up and transferred on other surfaces [25-32].
The Biomaterials Silver Jubilee Compendium T. Shimizu et al. / Biomaterials 24 (2003) 2309-2316
213 2311
Fig. 1. Histological comparison between cell-dense and cell-sparse tissues. Hematoxilin and eosin staining shows that cells are dense and tightly connected in myocardial tissue (A) On the other hand, cartilage tissue includes sparse cells and large amount of ECM (B).
Fig. 2. Cell harvest mechanism by using temperature-responsive culture surfaces. (A) Cells attach to hydrophobic culture surfaces via cell membrane proteins and ECM, which reside in serum or are secreted from the cells. (B) When enzymatic digestion is used, both membrane and ECM proteins are disrupted, resulting in cell detachment. (C) When cells are cultured on temperature-responsive culture surfaces, the interconnection between ECM and hydrophilic culture surfaces is released only by lowering temperature. Then the cells detach together with intact proteins.
Fig. 3. Cell sheet release from temperature-responsive culture surfaces. (A) When cells are cultured confluently, the cells connect to each other via cell-to-cell junction proteins. (B) When harvested by protease treatments, cell-to-cell connections are disrupted and cells are released separately. (C) When PIPAAm-grafted surfaces are used, cell-to-cell connections are completely preserved and the cells are released as a contiguous cell sheet. ECM retained underneath the cell sheets play a role as a adhesive agent.
214
2312
The B i o m a t e r i a l s S i l v e r J u b i l e e C o m p e n d i u m T. Shimizu et al. /Biomaterials 24 (2003) 2309-2316
As cell sheet manipulation, two techniques have been performed according to cell types and objects. One is to manipulate cell sheets directly with forceps or pipetting after the sheets are completely harvested resulting in proportionally shrunk and thicker constructs due to active cytoskeletal reorganization. As indicated by synchronized beating of shrunk cardiomyocyte sheets, cell-to-cell connections are preserved after this procedure [31]. The other is to use support membranes including a hydrophilically modified poly(vinylidene difluoride)(PVDF) membrane for preserving cell sheet morphology without any shrinkage. Before cell sheets release, support membranes are placed over the confluent cells. Then the cell sheets physically attached to the support membranes are harvested from PIPAAmgrafted surfaces below 32~ and transferred onto other surfaces. Incubation at 37~ causes reattachment of the cell sheets to new surfaces via remaining adhesive proteins. Finally, only the support membranes are removed. The latter technique has realized the cell sheet manipulation preserving their structure and function [26-30]. These cell sheet manipulation techniques without using any biodegradable scaffolds have been applied to tissue engineering in three types of contexts (Fig. 4). First is transplanting single cell sheet for skin and
cornea reconstruction. Advantages of skin epithelial cell sheets harvested by using PIPAAm-grafted surfaces have been confirmed in comparison with those harvested by dispase treatments. E-cadherin, which is an essential protein for skin cell-to-cell junctions, and laminin 5, which is a major component of epithelial basement membranes, were retained in skin cell sheets released from PIPAAm-grafted surfaces [27]. It should attenuate the risk of infection after artificial skin transplantation. Second is to layer same cell sheets for reconstructing homogeneous tissues including myocardium. Third is to layer several types of cell sheets for fabricating laminar structures including liver, kidney and vascular. Layered co-culture comprising a hepatocyte sheet and an endothelial cell sheet has revealed the differentiated cell shape and extensive albumin expression of hepatocytes, which have never been seen in hepatocyte mono-culture [32]. We have been now applied these technologies "cell sheet engineering" to reconstructing various types of tissues. Among them, myocardial tissue engineering based on the second context is described below.
4. Myocardial tissue reconstruction by layering cardiomyocyte sheets [28,30,31] Cardiomyocytes are tightly interconnected with gap junctions and pulsate simultaneously in native heart tissue. It is also well-known that confluent cultured cardiomyocytes on culture surfaces connect via gap junctions and beat simultaneously [33]. Therefore, in myocardial tissue engineering by layering cell sheets, it is a crucial point whether electrical and morphological communications are established between bilayer cell sheets. Chick embryo or neonatal rat cardiomyocyte sheets released from PIPAAm-grafted surfaces presented synchronized pulsation. To examine the electrical communication, two cardiomyocyte sheets were overlaid partially as schematically illustrated in Fig. 5. Two
Fig. 4. Three contexts in cell sheet engineering. (A) Single cell sheet is useful for skin or cornea transplantation. (B) Same cell sheets are layered to reconstruct homogeneous 3-D tissues including myocardium. (C) Several types of cell sheets are co-layered to fabricate laminar structures including liver and kidney.
Fig. 5. Schematic illustration of electrical analysis of layered cardiomyocyte sheets. To examine the electrical synchronization, two cardiomyocyte sheets (A, B) are overlaid partially. Two electrodes are set over monolayer parts of both cell sheets to detect the electrical potentials separately.
215
The Biomaterials Silver Jubilee Compendium
2313
T. Shimizu et al. / Biomaterials 24 (2003) 2309-2316
electrodes were set over monolayer parts of both cell sheets. Detected electrical potentials of the two sheets completely synchronized (Fig. 6). Furthermore, electrical stimulation to the single-layer region of one sheet was transmitted to the other cell sheet and the two cell sheets pulsated simultaneously. Histological analysis showed that bilayer cardiomyocyte sheets contacted intimately resulting in homogeneous tissue. Cell-to-cell connections including desmosomes and intercalated disks were confirmed by transmission electron microscopic images. These data indicate that electrical and morphological communications are established between layered cardiomyocyte sheets. Under conventional culture conditions, cardiac myocytes are fixed to rigid material surfaces and their motion is highly limited. To minimize the interaction between cell sheets and culture materials, the sheets were overlaid on several types of materials including polyethylene meshes, elastic polyurethane meshes or framelike collagen membranes. In any cases, the constructs pulsated simultaneously with higher amplitude than the cells fixed on rigid culture surfaces. When cardiomyocyte sheets were layered on frame-like collagen mem-
branes, the center part of them is free from any culture materials. In result, 4-layer cardiac constructs on the frame-like collagen membranes pulsated spontaneously in macroscopic view. To examine in vivo survival and function of layered cardiomyocyte sheets, the constructs were transplanted into dorsal subcutaneous tissues of nude rats. Surface electrograms originating from transplanted constructs were detected independently from host electrocardiograms, in the earliest case, at 2 weeks after the operation (Fig. 7). When transplantation sites were opened, macroscopic simultaneous graft beatings were observed at the earliest period, 3 days after the transplantation. Furthermore, graft survival was confirmed at least up to 1 year. Morphological analysis demonstrated that neovascularizations occurred in a few days and that vascular network was organized within a week (Fig. 8A). Cross-sectional views revealed stratified celldense myocardial tissues (Fig. 8B), well-differentiated sarcomeres and diffuse formation of gap junctions. In comparison between 2-layer and 4-layer cardiac tissue grafts, fractional shortening increased depending on the number of layered cell sheets. Thus, the basic technology has been established to fabricate electrically communicative, pulsatile myocardial tissues by using cell sheets both in vitro and in vivo.
Sheet A
5. Future perspectives Sheet B
1 sec
Fig. 6. Synchronization of layered cardiomyocyte sheets. Representative tracings of electrical potentials of sheet A and sheet B show complete synchronization.
Recently, research on myocardial tissue engineering has been accelerated to develop further advanced therapy for severe heart failure. Transplantation of layered cardiomyocyte sheets on the myocardial scar may be more beneficial than that of bioengineered heart tissue including biodegradable scaffolds in the point of scaffold-mediated disadvantages. However, there are several common problems in myocardial tissue
1 sec
Fig. 7. Skin surface electrogram of transplanted cardiomyocyte sheets. Representative tracings of the host electrocardiogram (upper) and the electrical potential detected via the electrode set at the skin just above the transplanted heart graft (lower) are shown. Skin surface electrogram originating from the graft is detected independently from host electrocardiogram.
216 2314
The Biomaterials Silver Jubilee Compendium T. Shimizu et al. / Biomaterials 24 (2003) 2309-2316
Fig. 8. (A) Macroscopic view of the transplanted cardiac graft. Multiple neovascularization is shown in the square-designed cardiac graft transplanted into dorsal subcutaneous tissue. (B) Azan staining shows a stratified cardiac tissue graft including elongated cardiomyocytes and microvasculars (arrows).
Fig. 9. Schematic illustration of myocardial reconstruction based on cell sheet engineering. We now propose the application of "cell sheet engineering" to myocardial tissue reconstruction. Cell sourcing remains a crucial problem. Neovascularization for oxygen and nutrition supply is also critical to fabricate human applicable myocardial tissue. Growth factors, gene delivery and the utility of gene-modified cells or endothelial cells may be helpful. Mechanical load by using bioreactors should strengthen the engineered myocardial tissues. Transplantation of engineered tissue into myocardial infarction model is now in progress.
engineering. As described in Section 1, myocardial cell sourcing remains a crucial problem. Further advance in stem cell biology for cardiomyocytes will be needed to realize clinical application of bioengineered myocardial tissues. Vascular reconstruction is also one of the most critical issues in myocardial tissue engineering. Sufficient supply of oxygen and nutrition is required for functionally beating heart tissue. It has been reported that cells are dense in the graft periphery, but sparse in the interior part due to insufficient oxygen perfusion in scaffoldbased heart tissue grafts [34]. Although, in our studies, multiple neovascularization arose in transplanted cardiac grafts in a few days, primary insufficient oxygen and nutrition permeation also limit the number of transplanted cardiomyocyte sheets. Hence, new methods to accelerate blood vessel formation are now requested to engineer larger or thicker constructs for heart tissue
repair. As examined in isolated cell injection, genemodified cells may be also applicable for engineering more vascularized heart tissues [35]. Using cell sheet technology, it has been reported that a single layer of endothelial cell sheet enhances the capillary formation in vivo [36]. Therefore, heterogenous layering of endothelial cell sheets between cardiomyocyte sheets may promote neovascularization. Further research and development will be needed to engineer vascular networks sufficient for fabricating clinically applicable heart tissues. In native heart, cardiomyocytes are gradually elongated and hypertrophied by mechanical load increase in accordance with the growth of the body. Therefore, some investigators have attempted to strengthen bioengineered heart tissues by using mechanical devices. Carrier et al. used a rotating bioreactor for culturing cardiomyocytes on PGA scaffolds [37]. Fink et al.
217
The Biomaterials Silver Jubilee Compendium T. Shimizu et al. / Biomaterials 24 (2003) 2309-2316
clearly demonstrated that application of stretch devices to engineering heart tissues strengthened the contraction power and oriented the cells unidirectionally [38]. We are now trying to stretch layered cardiomyocyte sheets to fabricate more powerful cardiac constructs in vitro. Finally, our concept of myocardial tissue engineering is schematically illustrated in Fig. 9. Although further interdisciplinary research will be needed to clear the existing several problems, cell sheet engineering should have enormous potential for constructing clinically applicable heart grafts and should promote tissue engineering research fields.
[12]
[13]
[14]
[15]
Acknowledgements [16]
The present work was supported by the Japan Society for the Promotion of Science, Grant-in-Aid for Scientific Research (A) (13308055) and Grant-in-Aid for Encouragement of Young Scientists (13780693). It was also supported in part by the Open Research Grant from the Japan Research Promotion Society for Cardiovascular Diseases.
[17]
[18]
[19]
References [1] Fuchs JR, Nasseri BA, Vacanti JP. Tissue engineering: a 21st century solution to surgical reconstruction. Ann Thorac Surg 2001;72:577-91. [2] Mann BK, West JL. Tissue engineering in the cardiovascular system: progress toward a tissue engineered heart. Anat Rec 2001;263:367-71. [3] Reinlib L, Field L. Cell transplantation as future therapy for cardiovascular disease. A workshop of the National Heart Lung and Blood Institute. Circulation 2000;101:E182-7. [4] Menasche P, Hagege AA, Scorsin M, Pouzet B, Desnos M, Duboc D, Schwartz K, Vilquin JT, Marolleau JP. Myoblast transplantation for heart failure. Lancet 2001;357:279-80. [5] Kehat I, Kenyagin-Karsenti D, Snir M, Segev H, Amit M, Gepstein A, Livne E, Binah O, Itskovitz-Eldor J, Gepstein L. Human embryonic stem cells can differentiate into myocytes with structural and functional properties of cardiomyocytes. J Clin Invest 2001;108:407-14. [6] Min JY, Yang Y, Converso KL, Liu L, Huang Q, Morgan JP, Xiao YF. Transplantation of embryonic stem cells improves cardiac function in postinfarcted rats. J Appl Physiol 2002;92:288-96. [7] Makino S, Fukuda K, Miyoshi S, Konishi F, Kodama H, Pan J, Sano M, Takahashi T, Hori S, Abe H, Hata J, Umezawa A, Ogawa S. Cardiomyocytes can be generated from marrow stromal cells in vitro. J Clin Invest 1999;103:697-705. [8] Orlic D, Kajstura J, Chimenti S, Jakoniuk I, Anderson SM, Li B, Pickel J, McKay R, Nadal-Ginard B, Bodine DM, Leri A, Anversa P. Bone marrow cells regenerate infarcted myocardium. Nature 2001;410:701-5. [9] Akins RE. Can tissue engineering mend broken hearts? Circ Res 2002;90:120-2. [10] Langer R, Vacanti JP. Tissue engineering. Science 1993;260:920-6. [11] Papadaki M, Bursac N, Langer R, Merok J, Vunjak-Novakovic G, Freed LE. Tissue engineering of functional cardiac muscle:
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
2315
molecular, structural, and electrophysiological studies. Am J Physiol Heart Circ Physiol 2001 ;280:H 168-78. Li RK, Jia ZQ, Weisel RD, Mickle DA, Choi A, Yau TM. Survival and function of bioengineered cardiac grafts. Circulation 1999;100:II63-9. Sakai T, Li RK, Weisel RD, Mickle DA, Kim ET, Jia ZQ, Yau TM. The fate of a tissue-engineered cardiac graft in the right ventricular outflow tract of the rat. J Thorac Cardiovasc Surg 2001;121:932-42. Leor J, Aboulafia-Etzion S, Dar A, Shapiro L, Barbash IM, Battler A, Granot Y, Cohen S. Bioengineered cardiac grafts: a new approach to repair the infarcted myocardium. Circulation 2000; 102:III56-61. Eschenhagen T, Fink C, Remmers U, Scholz H, Wattchow J, Weil J, Zimmermann W, Dohmen HH, Schafer H, Bishopric N, Wakatsuki T, Elson EL. Three-dimensional reconstitution of embryonic cardiomyocytes in a collagen matrix: a new heart muscle model system. FASEB J 1997;11:683-94. Zimmermann WH, Schneiderbanger K, Schubert P, Didie M, Munzel F, Heubach JF, Kostin S, Neuhuber WL, Eschenhagen T. Tissue engineering of a differentiated cardiac muscle construct. Circ Res 2002;90:223-30. Luque EA, Veenstra RD, Beyer EC, Lemanski LF. Localization and distribution of gap junctions in normal and cardiomyopathic hamster heart. J Morphol 1994;222:203-13. Yamada N, Okano T, Sakai H, Karikusa F, Sawasaki Y, Sakurai Y. Thermo-responsive polymeric surfaces; control of attachment and detachment of cultured cells. Makromol Chem Rapid Commun 1990;11:571-6. Yamato M, Konno C, Kushida A, Hirose M, Utsumi M, Kikuchi A, Okano T. Release of adsorbed fibronectin from temperatureresponsive culture surfaces requires cellular activity. Biomaterials 2000;21:981-6. Nakajima K, Honda S, Nakamura Y, Lopez-Redondo F, Kohsaka S, Yamato M, Kikuchi A, Okano T. Intact microglia are cultured and non-invasively harvested without pathological activation using a novel cultured cell recovery method. Biomaterials 2001;22:1213-23. Okano T, Yamada N, Sakai H, Sakurai Y. A novel recovery system for cultured cells using plasma-treated polystyrene dishes grafted with poly(N-isopropylacrylamide). J Biomed Mater Res 1993;27:1243-51. Okano T, Yamada N, Okuhara M, Sakai H, Sakurai Y. Mechanism of cell detachment from temperature-modulated, hydrophilic-hydrophobic polymer surfaces. Biomaterials 1995;16:297-303. Yamato M, Okuhara M, Karikusa F, Kikuchi A, Sakurai Y, Okano T. Signal transduction and cytoskeletal reorganization are required for cell detachment from cell culture surfaces grafted with a temperature-responsive polymer. J Biomed Mater Res 1999;44:44-52. Kushida A, Yamato M, Konno C, Kikuchi A, Sakurai Y, Okano T. Temperature-responsive culture dishes allow nonenzymatic harvest of differentiated Madin-Darby canine kidney (MDCK) cell sheets. J Biomed Mater Res 2000;51:216-23. Kikuchi A, Okuhara M, Karikusa F, Sakurai Y, Okano T. Twodimensional manipulation of confluently cultured vascular endothelial cells using temperature-responsive poly(N-isopropylacrylamide)-grafted surfaces. J Biomater Sci Polym Ed 1998;9:1331-48. Hirose M, Kwon OH, Yamato M, Kikuchi A, Okano T. Creation of designed shape cell sheets that are noninvasively harvested and moved onto anothersurfaces. Biomacromolecules 2000;1: 377-81. Yamato M, Utsumi M, Kushida A, Konno C, Kikuchi A, Okano T. Thermo-responsive culture dishes allow the intact harvest of
218 2316
[28]
[29]
[30]
[31]
[32]
The Biomaterials Silver Jubilee Compendium T. Shimizu et al. / Biomaterials 24 (2003) 2309-2316
multilayered keratinocyte sheets without dispase by reducing temperature. Tissue Eng 2001;7:473-80. Shimizu T, Yamato M, Kikuchi A, Okano T. Two-dimensional manipulation of cardiac myocyte sheets utilizing temperatureresponsive culture dishes augments the pulsatile amplitude. Tissue Eng 2001;7:141-51. Kushida A, Yamato M, Kikuchi A, Okano T. Two-dimensional manipulation of differentiated Madin-Darby canine kidney (MDCK) cell sheets: the noninvasive harvest from temperatureresponsive culture dishes and transfer to other surfaces. J Biomed Mater Res 2001;54:37-46. Shimizu T, Yamato M, Akutsu T, Shibata T, Isoi Y, Kikuchi A, Umezu M, Okano T. Electrically communicating three-dimensional cardiac tissue mimic fabricated by layered cultured cardiomyocyte sheets. J Biomed Mater Res 2002;60:110-7. Shimizu T, Yamato M, Isoi Y, Akutsu T, Setomaru T, Abe K, Kikuchi A, Umezu M, Okano T. Fabrication of pulsatile cardiac tissue grafts using a novel 3-dimensional cell sheet manipulation technique and temperature-responsive cell culture surfaces. Circ Res 2002;90:e40-8. Harimoto M, Yamato M, Hirose M, Takahashi C, Isoi Y, Kikuchi A, Okano T. Novel approach for achieving doublelayered cell sheets co-culture: overlaying endothelial cell sheets onto monolayer hepatocytes utilizing temperature-responsive culture dishes. J Biomed Mater Res 2002;62:464-70.
[33] Oyamada M, Kimura H, Oyamada Y, Miyamoto A, Ohshika H, Mori M. The expression, phosphorylation, and localization of connexin 43 and gap-junctional intercellular communication during the establishment of a synchronized contraction of cultured neonatal rat cardiac myocytes. Exp Cell Res 1994;212:351-8. [34] Bursac N, Papadaki M, Cohen RJ, Schoen FJ. Eisenberg SR, Carrier R, Vunjak-Novakovic G, Freed LE. Cardiac muscle tissue engineering: toward an in vitro model for electrophysiological studies. Am J Physiol 1999;277(2 Part 2):H433-44. [35] Suzuki K, Murtuza B, Smolenski RT, Sammut IA, Suzuki N, Kaneda Y, Yacoub MH. Cell transplantation for the treatment of acute myocardial infarction using vascular endothelial growth factor-expressing skeletal myoblasts. Circulation 2001;104: I207-12. [36] Soejima K, Negishi N, Nozaki M, Sasaki K. Effect of cultured endothelial cells on angiogenesis in vivo. Plast Reconstr Surg 1998;101:1552-60. [37] Carrier RL, Papadaki M, Rupnick M, Schoen FJ, Bursac N, Langer R, Freed LE, Vunjak-Novakovic G. Cardiac tissue engineering: cell seeding, cultivation parameters, and tissue construct characterization. Biotechnol Bioeng 1999:64:580-9. [38] Fink C, Ergun S, Kralisch D, Remmers U, Weil J, Eschenhagen T. Chronic stretch of engineered heart tissue induces hypertrophy and functional improvement. FASEB J 2000;14:669-79.
219
The Biomaterials Silver Jubilee Compendium Available online at www.sciencedirect.com
8CIENCB~DIRECT e
ELSEVIER
Biomateriais
Biomaterials 25 (2004) 5681-5703 www.elsevier.com/locate/biomaterials
Review
Biomaterial-associated thrombosis" roles of coagulation factors, complement, platelets and leukocytes Maud B. Gorbet, Michael V. Sefton* Department of Chemical Engineering and Applied Chemistry, Institute of Biomaterials and Biomedical Engineering, University of Toronto, 4 Taddle Creek Road, Room 407D, Toronto, Ont., Canada M5S 3G9 Received 3 September 2003; accepted 19 January 2004
Abstract
Our failure to produce truly non-thrombogenic materials may reflect a failure to fully understand the mechanisms of biomaterialassociated thrombosis. The community has focused on minimizing coagulation or minimizing platelet adhesion and activation. We have infrequently considered the interactions between the two although we are generally familiar with these interactions. However, we have rarely considered in the context of biomaterial-associated thrombosis the other major players in blood: complement and leukocytes. Biomaterials are known agonists of complement and leukocyte activation, but this is frequently studied only in the context of inflammation. For us, thrombosis is a special case of inflammation. Here we summarize current perspectives on all four of these components in thrombosis and with biomaterials and cardiovascular devices. We also briefly highlight a few features of biomaterial-associated thrombosis that are not often considered in the biomaterials literature: 9 9 9 9
The importance of tissue factor and the extrinsic coagulation system. Complement activation as a prelude to platelet activation and its role in thrombosis. The role of leukocytes in thrombin formation. The differing time scales of these contributions.
9 2004 Elsevier Ltd. All rights reserved. Keywords." Leukocytes; Tissue Factor; CD1 l b; Platelets; Biomaterials; Complement activation; Thrombosis; Coagulation
Contents
1.
Introduction
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.
Coagulation cascade . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. The intrinsic pathway . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. The extrinsic pathway . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3. Physiologic inhibitors of coagulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4. Biomaterials and coagulation pathways . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5. Anticoagulants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5683 5683 5684 5684 5685 5686
3.
Complement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Classical pathway . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Alternative pathway . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. Regulatory molecules of complement activation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4. Interactions of complement and coagulation cascade . . . . . . . . . . . . . . . . . . . . . . . . . 3.5. Complement activation and biomaterials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5686 5686 5687 5688 5688 5688
*Corresponding author. Tel.: + 1-416-978-3088; fax: + 1-416-978-4317. E-mail address."
[email protected] (M.V. Sefton). 0142-9612/$- see front matter 9 2004 Elsevier Ltd. All rights reserved. doi: 10.1016/j.biomaterials.2004.01.023
5682
220
The Biomaterials Silver Jubilee Compendium
5682
M.B. Gorbet, M.V. Sefton / Biomaterials 25 (2004) 5681-5703
4. Platelets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. Platelet biology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2. Platelets and biomaterials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5689 5689 5690
5. Leukocytes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1. Leukocyte biology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2. Leukocyte activation and biomaterials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3. Leukocytes, platelets and coagulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5692 5692 5693 5694
6. Other important factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.1. Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.2. Endotoxin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5694 5694 5695
7. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5696
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5697
I. Introduction
Biocompatibility is defined as "the ability of a material to perform with an appropriate host response in a specific application". Biocompatibility of blood contacting devices relates mainly to the thrombotic response induced by the materials. Although no material has been found truly biocompatible, many cardiovascular devices function with low or acceptable risks of complications [1]. Hemolytic, toxic and immunologic reactions have usually been dealt with earlier in the development of a material to be used for cardiovascular devices and are rarely an issue with their use; an exception may be immunological reactions to tissue engineered constructs. However, thrombotic and thromboembolic complications, as well as bleeding risks associated with the necessary anticoagulant therapy remain of serious concern with cardiovascular devices. Clinical manifestations of the bioincompatibility of cardiovascular devices are numerous: sudden and complete obstruction of stents within weeks [2]; acute and subacute thrombotic occlusion in medium sized grafts (4-6 mm) [3]; embolic complications with artificial hearts [4], catheters [5] and prosthetic valves [6,7]; thrombotic complications during cardiopulmonary bypass [3] and angioplasty [8]. Larger vascular grafts remain thrombogenic for many years, but fewer thrombotic complications are observed as high flows disperse the clotting factors. However, occasional embolic episodes occur as high flows dislodge the thrombotic deposits. Even if the risk of thrombotic complication appears to be low (varying between 2% and 10% depending on the device), they may have fatal outcomes and the cost associated with the follow-up intervention is not negligible. Furthermore, these thrombotic complications with cardiovascular devices occur despite the use of antiplatelet and anticoagulant therapies reinforcing the inherent thrombogenicity of the materials. Material thrombogenicity is further illustrated by the acute failure of small diameter vascular
grafts despite the strong anticoagulant regimen. Many years of intensive research on biomaterials have not yet produced a material, which has proven suitable for this last application. To improve the blood compatibility of cardiovascular devices, surface modifications, such as attachment of antithrombotic agents or the immobilization of polyethylene oxide (PEO) have been considered but their success has been limited. Treating surfaces with PEO to reduce protein adsorption and prevent platelet adhesion has remained unproven in terms of thrombogenicity. Different heparin coatings have been developed and some have actually been able to reach the commercial stage in cardiopulmonary bypass [9] and in coronary stents. However, reports on the improvement of in vivo biocompatibility have been mixed [10-16]. Heparinized cardiopulmonary bypass circuits appear to partially reduce the inflammatory response associated with cardiopulmonary bypass [17,18]. But to date, heparin coatings have not yet been shown to significantly reduce the number of postoperative complications, improve patient outcome, or reduce hospital stay [16,19,20]. This illustrates another limit of our understanding of bloodmaterial interactions: we do not know how much of an inflammatory and thrombosis response is tolerable or whether any of the changes in normal hemostasis induced by the device result in harmful consequences. Since biomaterial strategies have not resolved the problem, different pharmacological approaches are being investigated. Complement inhibition with the use of sCR1 [21] or anti-C5a antibody [22], serine protease inhibitors such as aprotinin [23,24], platelet receptor antagonists such anti-GPIIb/IIIa [25] and cytokine antibody [26] are newer approaches to reduce thrombotic complications with cardiopulmonary bypass. While the results are promising--a partial reduction of the inflammatory or the thrombotic response to cardiovascular devices--it is as yet not possible to make conclusions with respect to the overall improvement of device biocompatibility. Unfortunately, most agents
221
The Biomaterials Silver Jubilee Compendium
5683
M . B . Gorbet, M . V . S e f t o n / B i o m a t e r i a l s 25 ( 2 0 0 4 ) 5 6 8 1 - 5 7 0 3
(Fig. 2). Thrombin is formed following a cascade of reactions where an inactive factor becomes enzymatically active following surface contact or after proteolytic cleavage by other enzymes; the newly activated enzyme then activates another inactive precursor factor. Initiation of clotting occurs either by surface-mediated reactions, or through tissue factor (TF) expression by cells. The two systems converge into the common pathway resulting in the formation of fibrin clot upon action of thrombin on fibrinogen. Then, Factor XIII, activated by thrombin, crosslinks and stabilizes the fibrin clot into an insoluble fibrin gel. Fig. 1. Overview of blood-material interactions showing the components relevant to thrombosis. While complement and leukocytes are normally considered under "inflammation" we consider them as participants in thrombosis along with platelets and coagulation.
affect only one of the players in the blood compatibility response and this is not likely to be sufficient to result in clinical benefits. On the other hand, these inhibitors and antibodies provide valuable information on the mechanisms involved in the thrombotic complications associated with cardiovascular devices. Under normal conditions, blood contacts an endothelium with anticoagulant and antithrombotic properties. The use of a cardiovascular device represents the introduction of a foreign surface in the circulation, without the properties of the endothelium. Bloodmaterial interactions trigger a complex series of events including protein adsorption, platelet and leukocyte activation/adhesion, and the activation of complement and coagulation; they are highly interlinked (Fig. 1). This review outlines the current state of understanding of these phenomena with particular reference to the biomaterial or cardiovascular device as an agonist of these thrombotic reactions. This review is not, however, a catalog of surface modification or biomaterial design strategies that have been used in an attempt to control this aspect of the host response. Rather the focus is on the mechanism of the thrombotic response to a biomaterial, how each component is thought to interact with a biomaterial and how these may interact to produce the observed thrombosis. We are particularly interested in the role of complement and leukocytes, which are not often considered by the biomaterials community to be contributors to thrombosis. Our approach is to treat thrombosis as a special case of inflammation.
2. Coagulation cascade Blood coagulation involves a series of proteolytic reactions resulting in the formation of a fibrin clot
2.1. The intrinsic pathway The classic picture (Fig. 2) shows the intrinsic pathway being initiated by contact activation of high molecular weight kininogen (HMWK), prekallikrein and Factor XII: it is commonly said that these molecules require contact with (negatively charged) surfaces for zymogen activation in vitro [27]. Factor XII is activated by adsorption, FXIIa converts prekallikrein into kallikrein and with H M W K as a cofactor activates Factor XI to Factor FXIa. Factor XIa activates Factor IX to Factor IXa. Following a cascade of reactions involving among others the intrinsic tenase complex (Factor IXaFactor VIIIa), prothrombin is cleaved into thrombin. The importance of the intrinsic pathway to normal blood coagulation remains speculative, as the occurrence of negatively charged surfaces in viva is limited.
EXTRINSIC SYSTEM
INTRINSIC SYSTEM Surface Contact
Factor :X]:I~--~:X~ o Factor XI ~-~--~XIo Factor
~ FactorV]~
',~,,,%/..Je_.le_.ts_ __ _ I . . . . Factor X
COMMON PATHWAY Prothrombin
- Factor
YJIo --~
IXo++ IX- C!++= i~/"a
F••lat ~Xo
Co~
Tissue/
-z.
,,
~247 ctor V
elets
Fibrinogen (monomer)
Factor X F~zto r XI]]
Thrombin
L
l(lllo Fibrifl (polymer)
- Fibrin
(slable polymer
)
Fig. 2. Simplified blood-coagulation cascade as viewed by biomaterial textbooks [117] and older hematology books. The intrinsic system, starting with Factor XII, is shown as a linear cascade of zymogen activation steps in parallel with the extrinsic system that involves TF. TF and platelets are shown as 'cofactors' of the process (similar to Ca +2) rather than as central participants. The intrinsic and extrinsic systems converge on the common pathway to produce thrombin and fibrin [used with copyright approval from Elsevier/Academic Press].
222
The Biomaterials Silver Jubilee Compendium
5684
M.B. Gorbet, M.V. Sefton / Biomaterials 25 (2004) 5681-5703
Collagen present in the subendothelium after vessel injury could be the surface required for this reaction [28]. Under physiologic conditions, the lack of relevance of the contact activation system is consistent with the fact that deficiencies of the contact proteins, HMWK, prekallikrein and Factor XII, have not been associated with abnormal bleeding [1,27,29].
2.2. The extrinsic pathway The hematology textbooks center their view of the blood-coagulation cascade, on the TF-dependent pathway (Fig. 3); this perspective has replaced the older perspective that is still used in biomaterials textbooks. The physiological initiator of coagulation is TF, which is expressed on damaged cells at the site of vascular injury. Plasma Factor VII (FVII) binds to TF on the cell membranes and requires activation to FVIIa to form the extrinsic tenase complex: TF-VIIa complex. FVII is activated by trace amounts of thrombin, FIXa, FXa and TF-VIIa complex, the latter being more likely physiologically relevant as small amounts of TF-FVIIa are present extravascularly in vivo [30,31]. Picomolar concentrations of FVIIa circulate normally in blood and are also thought to serve as a primer in the initiation of the coagulation cascade by allowing direct formation of TF-FVIIa complex upon TF exposure [32]. TF-FVIIa complex on cell membranes cleaves Factor X into Factor Xa in the presence of calcium. The prothrombinase complex can then assemble on the membrane and generate thrombin (common pathway of the coagulation cascade). FX is not the only physiological substrate of the TF-FVIIa complex. The TF-FVIIa complex also activates FIX [33].
Extrinsic tenase . ,,.
,~.~
"/
TF § VII
~~176 TF-VII
~.~c,
II . . . . . .
. . . . . . .
PROTHROMBIN Fig. 3. Revised (simplified) blood-coagulation cascade as given in a standard hematology textbook [30]. Unlike Fig. 2, there is no intrinsic system or Factor XII. The cascade is not linear with several feedback loops and most importantly it begins with TF and the tenase complex. Platelet involvement is still shown as a 'cofactor' and the reactions past thrombin are not shown [used with copyright approval from McGrawHill].
The extrinsic and intrinsic pathways are not independent of each other. When coagulation is initiated by a TF-dependent pathway, the intrinsic tenase remains important, since production of FXa by FIXa-FVIIIa complex has been shown to significantly contribute to thrombin generation [34]. It appears that extrinsic tenase TF-FVIIa is responsible for the onset of coagulation while the intrinsic tenase is the major player in the propagation phase [35]. The activation of FX by FIXa is all the more important since tissue factor pathway inhibitor (TFPI) will reduce the production of FXa by TF-VIIa complex [30,31].
2.3. Physiologic inhibitors of coagulation Most activated coagulation factors are serine proteases. Plasma contains several protease inhibitors, such as C~l-protease inhibitor, c~2-macroglobulin, heparin cofactor II and antithrombin III, to modulate and inhibit their activity. Antithrombin III is the most important [36]: it neutralizes its target enzymes, FXa and thrombin, by forming a complex with the enzyme in which the enzyme's active site is blocked. In the absence of heparin, complex formation occurs at a relatively slow rate. However, in the presence of the polysaccharide heparin or naturally occurring heparan sulfate (on the endothelium), inhibition rates rise significantly. Antithrombin is also able to inhibit FIXa, FXIa and FXIIa [37]. While TF-VIIa is not efficiently inhibited by antithrombin, it has its own inhibitor: the TFPI. Plasma TFPI is not the major intravascular pool of TFPI. A larger pool exists on the luminal surface of the vascular endothelium, which is released by a bolus injection of heparin. Platelets also carry 10% of the total TFPI in blood and release their TFPI following stimulation by thrombin and other agonists [38]. Activated monocytes may also release TFPI [39]. TFPI has two inhibitory sites, one for FXa and one for TF-VIIa complex. Inactivation of FXa through binding to TFPI in solution is required for inhibition of the complex TF-VIIa. The initial binding of FXa to TFPI is relatively slow [40] and may not be able to prevent thrombosis when TF-FVIIa complex are being formed at a high rate [31]. Leukocyte elastase also cleaves TFPI and impairs its ability to inhibit both FXa and TF-VIIa complex. Upon exposure of a complex FXa/TFPI/TFVIIa to elastase, FXa and TF-VIIa activities are restored [41]. The endothelium also participates in the regulation of thrombosis via thrombomodulin. Thrombin binds to thrombomodulin on endothelial cells and this complex activates protein C [42]. Protein C is a vitamin Kdependent protein and activated protein C inactivates FVa and FVIIIa. Protein S, also a vitamin K-dependent protein, is a necessary cofactor for activated protein C.
223
The Biomaterials Silver Jubilee Compendium M.B. Gorbet, M.V. Sefton / Biomaterials 25 (2004) 5681-5703
More than half of the protein S in plasma is bound to the C4b-binding protein (a regulatory protein of the complement system) and is not functionally active.
2.4. Biomaterials and coayulation pathways While under physiological conditions the role of FXII activation is questionable, in the presence of a cardiovascular device activation of FXII may occur. The artificial surface, with its adsorbed protein layer, then represents the required ("negatively charged") surface. Protein adsorption is the first event in blood-material interactions, and adsorption of the contact phase proteins may result in activation of the intrinsic cascade. Earlier studies had focused on protein adsorption on glass or biomaterial surfaces with isolated protein solutions or diluted plasma, and showed how fibrinogen is replaced over time by H M W K (the Vroman effect), suggesting a possible role of the intrinsic pathway in material thrombosis [43]. Recent studies using whole blood or plasma have provided new insights on the adsorption and activation of contact phase proteins. FXII adsorption has been observed in moderate amounts on materials used in vascular grafts [44] and hemodialysers [45]; however, it was not found in its activated form [46]. Although H M W K and prekallikrein may be adsorbed on the material surface, the lack of FXIIa on the material surface will stall the initiation of coagulation through the intrinsic pathway. In some instances, in vitro activation of FXII and kallikrein has been reported with biomaterials [47-49]; however, no test was performed to determine if such activation resulted in any significant activation of coagulation. Other studies have actually shown that only minute amounts of thrombin or thrombin-antithrombin III complex (TAT) are generated when biomaterials are incubated with undiluted plasma alone [50-52]. Furthermore, higher levels of adsorbed kallikrein and Factor XII on biomaterials do not correlate with TAT formation [50], suggesting that the contact phase proteins, by themselves, play little role in the activation of coagulation. In fact, a study by Hong et al. [51] suggested that the presence of leukocytes is required for activation of the coagulation cascade; the requirement of a TF-dependent pathway of initiation of coagulation may thus also apply to biomaterials. In vivo, a small increase of FXIIa is observed during cardiopulmonary bypass [53,54], but it appears to be in response to the surgical intervention and the establishment of extracorporeal circulation (i.e., exposure to the biomaterials of the circuit) does not further increase FXIIa levels [54]. Furthermore, no significant correlation has been observed between FXIIa and thrombin generation [54]. In vivo results with hemodialysis also failed to show any significant increase of FXIIa [55]. That a patient with a severe FXII deficiency showed levels of thrombin
5685
generation during cardiopulmonary bypass similar to normal patients [56] casts further doubt on the role of Factor XII in the initiation of coagulation with biomaterials. Taken together, these studies do not support the view that activation of the contact phase proteins is important in the activation of coagulation by biomaterials. While the TF pathway of blood coagulation has become the focus in hematology and has led to the revised version of the coagulation cascade (compare Figs. 2 and 3), the biomaterials community has been slow to recognize its importance. The textbooks refer to the older model of coagulation with the separated intrinsic and extrinsic pathways with the thought that the extrinsic pathway is not directly related to bloodmaterial interactions [1,57-59]. However, blood contact with a material represents a potential stimulus to induce TF expression by monocytes, resulting in blood coagulation through the extrinsic system. Indeed, monocyte TF expression has been observed in vivo during or after cardiopulmonary bypass [60-62]. Further details of TF expression in the presence of biomaterials are given elsewhere [63,64] and summarized below in the section on leukocytes. The role of leukocytes (most likely due to TF on monocytes) in activation of the coagulation by biomaterials was highlighted in the study by Hong et al. [51] that was referred to earlier. Thrombin-antithrombin formation (TAT) on PVC was found to be negligible in both plasma and platelet-rich plasma, while significant levels of TAT were observed in whole blood; i.e., only in the presence of leukocytes was there significant thrombin formation. More research is needed, however, to define the relative importance of the extrinsic and intrinsic pathways of coagulation in the overall picture of thrombosis on biomaterials. The time scales associated with initiation of the coagulation cascade by contact phase activation and with TF expression will have an impact on their relative importance. As it is part of the protein adsorption "reaction", contact phase activation occurs during the first few seconds/minutes of blood contact with a material. On the other hand, to be expressed on monocytes, TF requires synthesis and thus a minimum of 60rain (in a normal patient) would elapse before this pathway could significantly contribute to thrombin generation. Thrombin generation by FXIIa on materials is also dependent on flow [65], because of the effects of flow on mass transfer as well as direct effects on platelet/leukocyte phenotype. One manifestation of this is the relevant time scales of the coagulation reactions will vary with the flow situation. Thus, the relative roles of the intrinsic and extrinsic pathways in thrombin generation will likely depend on both the flow situation and the relevant time frame. This adds another level of complexity to the understanding/
224 5686
The Biomaterials Silver Jubilee Compendium M.B. Gorbet, M.V. Sefton / Biomaterials 25 (2004) 5681-5703
characterization of thrombin generation with cardiovascular devices.
2.5. Anticoagulants To prevent the formation of thrombi during cardiopulmonary bypass, hemodialysis and angioplasty, anticoagulants are routinely administered. Heparin is the most common [37]. Heparin binds to antithrombin III (ATIII) via a unique pentasaccharide sequence, and causes a conformational change in the reactive center of ATIII, thereby accelerating the rate of ATIIImediated inactivation of several clotting enzymes (thrombin, FXIa, FXa and FIXa). Heparin promotes the formation of the complex thrombin-ATIII (TAT) by binding to both proteins. On the other hand, to inactivate FXa, heparin needs to bind only to ATIII. ATIII forms a 1:1 irreversible complex with coagulation enzymes, the heparin then dissociates and can be reused. A limitation of heparin is that it is unable to inactivate thrombin bound to fibrin or to surfaces, or to inhibit FXa within the prothrombinase complex [66]. Heparin can also cause thrombocytopenia. Heparin binds to platelet factor 4 (PF4) and in some patients, antibodies will develop against this heparin-PF4 complex. The antibody-PF4-heparin complex then binds to platelets and induces platelet activation, aggregation and activates the blood-coagulation pathways, resulting in both a loss of circulating platelets and a thrombotic state [67]. Following implantation of cardiovascular devices, such as vascular grafts and artificial valves, anticoagulants such as warfarin is used. It can be taken orally and it interferes with the vitamin K cycle, thus impairing the biological function of vitamin K-dependent coagulation proteins (prothrombin, FVII, FIX and FX) [37]. For stents, anticoagulants are not needed except during placement. Aspirin and Plavix (clopidogrel, antiplatelet aggregation) are used after placement (aspirin forever; Plavix for 1 month or 1 year) to control thrombosis, while heparin and a GPIIb/IIIa antagonist (e.g., ReoPro) is only needed perioperatively. Other anticoagulants of interest for use of cardiovascular devices [66] are the tick anticoagulant peptide (TAP) and antistatin which binds to FXa even within the prothrombinase complex; hirudin, a leech-derived protein, a potent thrombin inhibitor; and D-Phe-ProArgCHzC1 (or PPACK), a peptide that directly inactivates thrombin by interacting with the active site of thrombin. Both PPACK and hirudin are able to inactivate fibrin-bound and surface-bound thrombin [66,68] which is a significant advantage. Ximelagatran is also a new oral direct thrombin inhibitor and clinical trials have been very successful: similar or superior efficacy relative to warfarin in some scenarios with reduced bleeding, obviating the need for monitoring. It
has many advantages over warfarin and will likely soon replace it [69]. When studying blood-material interactions in vitro, heparin is usually the anticoagulant of choice as it is the most widely used with cardiovascular devices. However, heparin also possesses some anticomplement activity [70]. PPACK and hirudin, which appear to be more specific than heparin, may then be used especially when the mechanisms of cell activation are studied. However, the associated high cost restricts their use.
3. Complement The complement system plays an important role in the body's defense mechanisms against infection and "non-self' elements. The complement system consists of more than 20 plasma proteins that function either as enzymes or as binding proteins. Complement activation is initiated via the classical or alternative pathways and the terminal pathway is common to both (Fig. 4). Both pathways contain an initial enzyme that catalyses the formation of the C3 convertase, which in turn generates the C5 convertase allowing the assembly of the terminal complement complex (TCC). Various complement products (C3b, C4b and iC3b) bind to particles, surfaces, bacteria and immune complexes in a process called opsonization [71], which facilitates their uptake by inflammatory cells. Activation of complement results in cell lysis when the terminal attack complex is inserted into the cell membrane. Complement activation also releases C3a, C4a and C5a, which are anaphylatoxins. These peptides are humoral messengers that bind to specific receptors on neutrophils, monocytes, macrophages, mast cells and smooth muscle cells. They induce a variety of cellular responses such as chemotaxis, vasodilatation, cell activation and cell adhesion [72]. At high enough concentration, they are responsible for the many systemic effects of complement activation.
3.1. Classicalpathway The classical pathway is normally triggered by antigen-antibody complexes that bind the C1 complex (Clq, Clr, C1 s) through the Clq component. This activates C ls, which is then able to cleave the C4 complement protein into C4a and C4b. C4b attaches to its target surface via its exposed metastable thioester binding site. It is important to note that C4b does not bind efficiently to membrane surfaces and the fluid phase C4b is rapidly inactivated by the loss of its binding site. C2 binds to the attached C4b and is cleaved by C ls, releasing C2a. The classical C3 convertase, C4bC2b, is thus formed and can cleave C3 into C3a (anaphylatoxin) and C3b. The combination of C4bC2b
225
The Biomaterials Silver Jubilee Compendium M.B. Gorbet, M.V. Sefton / Biomaterials 25 (2004) 5681-5703
5687
Fig. 4. Pathways of (a) complement activation [256], copyright JB Lippincott and (b) inhibition [257], copyright Elsevier. In (a), the classical and alternative pathways are shown leading to the production of the C3 convertase and C5 convertase complexes, both of which are part of amplification loops. In (b), various natural and synthetic inhibitors are shown inhibiting downstream portions (C3 and C5) of the complement cascades; only C1 inhibitor is shown as being effective at a level prior to the formation of the C3 or C5 convertases.
and C3b becomes the C5 convertase, which cleaves C5 into C5a (anaphylatoxin) and C5b. C5b is the first component of the terminal complex and has high affinity for C6. C5bC6 then binds C7, C8 and up to 12 molecules of C9 and this forms the TCC C5b-9. If C5b is attached to a biological surface, the TCC (also called the membrane attack complex, mC5b-9) inserts itself into the lipid layers resulting in cell damage and/or lysis. In the absence of a biological membrane, the complex binds to S protein (also known as vitronectin) to create SC5b-9 in the fluid phase.
3.2. Alternative pathway The activation of the alternative pathway does not require antibody or immune complexes and is activated by any foreign surfaces, such as fungal, bacterial polysaccharides, lipopolysaccharides (LPSs), particles and biomaterial surfaces. Complement activation via the alternative pathway occurs spontaneously at a low rate.
Spontaneous hydrolysis of the internal thioester group of C3 occurs in the fluid phase, generating C3. H20. This hydrolyzed C3 can bind and activate Factor B and cleave another C3 molecule into C3a and C3b. The alternative C3 convertase, C3bBb, is formed. In the absence of a surface to support binding of C3b, little complement activation occurs. In the presence of a surface, covalent binding of C3b to hydroxyl or amine groups on the surface may occur via the carbonyl group in the C3b thioester binding site. Attachment of C3b to a surface favors binding of Factor B and Factor D to C3b. Factor D cleaves Factor B into Ba and Bb, and the alternative C3 convertase, C3bBb, is formed again. This attached C3 convertase is able to generate more C3b, resulting in a positive amplification loop. Properdin acts to stabilize the C3 convertase. The clustering of C3b on the surface allows the formation of the alternative C5 convertase, C3bBbC3b, and C5 can be cleaved. The assembly of the TCC follows as for the classical pathway.
226 5688
The Biomaterials Silver Jubilee Compendium M.B. Gorbet, M.V. Sefton / Biomaterials 25 (2004) 5681-5703
3.3. Reyulatory molecules of complement activation The various molecules involved in the regulation of complement activation are illustrated in Fig. 4b. The classical pathway is regulated by two specific mechanisms [29]: C1 inhibitor, a plasma protein, binding Cls or C lq and inhibiting the enzymatic activity of the C1 complex; and the C4b-binding protein inhibiting C4b bound to a membrane or a surface. Factor I is a proteolytic enzyme that binds C4b and C3b, generating iC4b and iC3b, which are further degraded into smaller fragments. C4b-binding protein is a cofactor of Factor I and augments the degradation of C4b by Factor I. Factor H is a cofactor of Factor I for the degradation of C3b. Factor H is also able to displace Bb in the C3 convertase to promote C3b inactivation by Factor I [75]. On cell surfaces, the C3 and C5 convertase activity are regulated by three integral membrane proteins [75]. Decay-accelerating factor (DAF), found on all peripheral blood cells, destabilizes the C3 convertase by promoting the release of factor Bb. The membrane cofactor protein (MCP), expressed on leukocytes and platelets, favors the dissociation of Factor B and promotes C3b association with Factor I. The complement receptor type 1 (CR1), present on all blood cells except platelets, acts like Factor H and displaces Bb from the C3 convertase and facilitates inactivation by Factor I. To prevent lysis of "bystander" blood cells, membrane proteins called homologous restriction factors limit the ability of the terminal complex to properly assemble on autologous cells [75]. Two proteins have been characterized: CD59 and the C8-binding protein (also called MAC inhibiting protein). CD59, found not only on blood cells but on many cells such as hepatocytes and epithelial cells, binds to C8 and C9 and inhibits C9 polymerization. Little is known about the C8-binding protein, which is believed to bind C8 and C9.
3.4. Interactions of complement and coagulation cascade Although the coagulation and complement cascades are discussed as separate entities, the two cascades appear to interact significantly to modulate each other's activity [29,75]. Factor XIIa and kallikrein are known to cleave C ls [29] and thus have the capacity to trigger classical complement activation. Thrombin activates C3, C5, C6 and Factor B; kallikrein cleaves C5 and factor B; and Factor XIIa also cleaves C3. The activity of thrombin on C3 and C5 may actually explain the higher background levels of C3a and C5a in serum versus plasma. Table 1 outlines the various interactions between complement and coagulation factors.
Table 1 Interactions between complement and coagulation systems [29,75] Protein
Type of interaction
Thrombin Factor XIIa Kallikrein Antithrombin III Bb C3bBb C 1 inhibitor S protein (vitronectin) C4b-binding protein
Proteolysis of C3, C5, C6 and factor B Proteolysis of Clr, Cls and C3 Proteolysis of C1, C5 and Factor B Protect RBC from lysis by mC5b-9 Proteolysis of prothrombin Proteolysis of prothrombin Inactivates FXIIa and kallikrein Stabilizes plasminogen activator inhibitor 1 Binds to the vitamin K-dependent protein S
3.5. Complement activation and biomateria& Complement activation is generally treated as if it is part of the inflammatory response induced by biomaterials. For example, complement activation is known to occur during cardiopulmonary bypass and hemodialysis [75-78], and with catheters and prosthetic vascular grafts [79,80]. It is recognized that, both in the short and long term, complement activation plays a role in the leukocyte related clinical sequelae associated with the use of cardiovascular devices such as leukopenia, hypotension and pulmonary injury [81,82]. The thrombotic consequences of leukocyte activation are discussed below. The presence of a biomaterial is believed conventionally to activate complement via the alternative pathway. Biomaterials are usually classified as "activating" or "non-activating" surfaces [74]. On a non-activating surface, negatively charged groups such as carboxyl and sulfate, sialic acid and bound heparin appear to promote high-affinity association between bound C3b and Factor H. On the other hand, an activating surface is usually characterized by the presence of nucleophiles such as hydroxyl and amino groups: these groups will allow covalent binding of C3b and promote formation of the C3 and C5 convertase [2,73]. However, even in the absence of these activating groups on the surface, some biomaterials, such as polyacrylonitrile, are able to activate complement, suggesting that the mechanisms of material-induced complement activation due to nucleophilic groups is not the whole story. A newer hypothesis places emphasis on interaction of Factor H with the surface: an activating material is then defined by its capacity to bind Factor B rather than Factor H [73]. As noted earlier, binding of Factor H would lead to C3b inactivation by Factor I and thus terminate the propagation of the complement cascade. Some activating materials generate high levels of both C3b and C5b-9, while others will generate high C3b level but little C5b-9. Why the efficiency of C5 convertase formation relative to that of C3 formation differs from one activating surface to another is not well understood.
227
The Biomaterials Silver Jubilee C o m p e n d i u m
5689
M.B. Gorbet, M.V. Sefton / Biomaterials 25 (2004) 5681-5703
However, even low amounts of C5b-9 are able to activate leukocytes [74] and thus a low terminal complement activating material may still induce a significant inflammatory response. The question remains as to which are the appropriate levels of complement activation that the host can accept without deleterious effect. We also have to determine if, in the long term, the host will be able to differentiate/discriminate between moderate and high complement activating surfaces. The hypothesis that material-induced complement activation occurs exclusively via the alternative pathway has also been challenged. Reports of complement activation via the classical pathway during cardiopulmonary bypass [18,83] and a delay in complement activation observed with C4-deficient patients undergoing hemodialysis [84] suggest that classical activation plays a role in material-induced activation. The presence of immune complexes may allow for a rapid onset of complement activation, and then subsequently the alternative pathway becomes activated. In vitro studies have also demonstrated activation of the classical pathway by some biomaterials [85-87]. C1 inhibitors such as pentamidine and benzamidine were effective in lowering platelet adhesion and activation by polystyrene beads and polyethylene tubes while sCR1, an inhibitor of both pathways at the level of C3, had no effect [88,89]. Pentamidine was also effective in a canine chronic shunt in eliminating the thrombocytopenia seen with a platelet activating material (a polyvinyl alcohol (PVA) hydrogel)[90]. While some of the consequences of complement activation are well understood, more work is needed to fully understand how a material activates complement. Questions to be resolved include selecting inhibitors that block both pathways and at an early enough stage to inhibit the local (rather than systemic) effects of complement activation. Similarly, controlling the differential adsorption of C3b, Factor D and Factor H may be more important than preparing low adsorption, so-called "non-fouling" surfaces. The latter may lower the adsorption of all proteins but it may be more important to alter the composition of the protein adsorbate.
Platelets respond to minimal stimulation and become activated when they contact any thrombogenic surface such as injured endothelium, subendothelium and artificial surfaces. Platelet activation is initiated by the interaction of an extracellular stimulus with the platelet surface. This interaction involves the coupling of the agonist to specific receptors on the platelet plasma membrane [91]. Plasma proteins such as thrombin and fibrinogen; vascular wall products such as collagen; and molecules derived from inflammatory cells (i.e., leukocytes) or platelets, such as platelet activating factor (PAF) or cathepsin G, are all potent platelet activators. A list of known platelet receptors and their specific agonist/ligand is presented in Table 2. Activation results in at least five physiologic responses [92]. (1) A platelet release reaction in which biologically active compounds stored in intracellular platelet granules are secreted into the microenvironment, such as platelet factor 4, thrombospondin, fl-thromboglobulin, ADP and serotonin. (2) P-selectin (previously referred to as GMP-140 or P A D G E M ) is released and expressed on the platelet membrane after e granule secretion. Table 2 Platelet receptors [91,102] Receptor
(a) Receptors leadin9 to platelet activation
Thrombin receptor Thromboxane A2 receptor V1A receptor PAF receptor 5HT2 receptor ~2 receptor ADP receptor PGE2 receptor C 1q receptor
4.1. Platelet biology
The platelet's main role in hemostasis is to preserve the integrity of the vascular wall through formation of a platelet plug. Platelets are anuclear, disc-shaped cells with a diameter of 3-4gm. They are derived from megakaryocytes in the bone marrow and circulate at an average concentration of 200 x 106 cells/ml, with individual platelet concentrations ranging from 150 to 400 • 106 platelets/ml.
Thrombin TxA2, PGH2, PGG2 Vasopressin Platelet activating factor Serotonin or 5-OH tryptamine Epinephrine ADP, ATP PGE2 Clq
(b) Receptors leadin9 to platelet inhibition A2 receptor Adenosine
PGI2 receptor PGD2 receptor
PGI2, PGE~ PGD2
(c) Platelet adhesion receptors (bindin9 may also result in platelet activation)
GPIa/IIa or VLA-2 GPIb/IX or GPIb
4. Platelets
Ligand/agonist
GPIc/IIa or VLA-5 GPIc'/IIa or VLA-6 GPIIb/IIIa GPIV or GPIIIb GPVI Vitronectin receptor PECAM-1 FcT-RII ICAM-2 P-selectin Leukosialin, sialophorin
Collagen Von Willebrand factor (vWF), thrombin Fibronectin Laminin Collagen, fibrinogen, fibronectin, vitronectin, vWF Collagen, thrombospondin, Collagen Vitronectin, thrombospondin Heparin Immune complexes LFA-1 Sialyl-Lex ICAM-1
GP: glycoprotein; VLA: very late antigen.
228 5690
The Biomaterials Silver Jubilee Compendium M.B. Gorbet, M.V. Sefton / Biomaterials 25 (2004) 5681-5703
P-selectin is a cell-surface glycoprotein belonging to the selectin family and plays an important role in mediating adhesion of activated platelets to neutrophils, monocytes and a subset of lymphocytes [93,94]. (3) The platelet eicosanoid pathway is initiated, resulting in the liberation of arachidonic acid from platelet phospholipids and in the synthesis and release of prostaglandins and thromboxane B2. (4) Platelet activation is characterized by a drastic shape change, which promotes platelet-platelet contact and adhesion. The rearrangement of the platelet membrane during activation also promotes the association of the tenase and prothrombinase complexes on its phospholipids. (5) Platelet activation results in the formation of platelet microparticles (PMPs), which are particularly rich in factor Va, platelet factor 3 and phospholipid-like procoagulant activity (phosphatidylserine) [95,96]. PMPs are formed from the surface membrane through exocytotic budding. Their physiologic role remains unclear but in vitro results have shown that they can bind and adhere to fibrinogen and fibrin, and coaggregate with platelets [97,98]. The procoagulant activity of PMPs, generated both in vitro and in vivo, has also been demonstrated [99-101]. Among the different platelet adhesion receptors (Table 2), GPIb and GPIIb/IIIa have the highest density on platelets. GPIb (CD42) is a leucine-rich glycoprotein receptor and approximately 25,000 receptors are present on the platelet surface [92,102]. It is complexed one to one with GPIX but the function of the latter remains unknown. GPIb is a long molecule, making it more susceptible to conformational change upon shear stress. GPIb mediates platelet interaction with von Willebrand factor (vWF). It will not bind plasma vWF unless the antibiotic ristocetin or the snake venom botrocetin is present. On the other hand, GPIb will bind to adsorbed or immobilized vWF on a surface. Shear stress is an important factor in platelet adhesion to vWF as it induces the required conformational change of vWF to bind GPIb. Thrombin also binds to GPIb but the functional significance of this binding has not been elucidated. GPIIb/IIIa (CD41/CD61) is an integrin receptor and is constitutively expressed on platelets. GPIIb/IIIa is the dominant platelet receptor with 40-80,000 receptors present on the surface of a resting platelet. Another 20-40,000 are present inside the platelets, in c~ granule membranes and in the membranes lining the open canalicular system [102]. They are translocated to the platelet membrane during the release reaction. On resting platelets, GPIIb/IIIa is in an inactive form and has a low-affinity binding site for adsorbed fibrinogen [103]. Upon platelet activation, a conformational change occurs leading to the exposure of the high-affinity binding site for soluble fibrinogen. Binding of fibrinogen to GPIIb/IIIa leads to platelet aggregation as well as
formation of platelet-leukocyte aggregates, via crosslinking of GPIIb/IIIa on two different platelets by fibrinogen and crosslinking between GPIIb/IIIa and Mac-1 (on leukocyte) by fibrinogen. Other adhesive glycoproteins containing RGD sequences can also bind to activated GPIIb/IIIa: vWF, thrombospondin, fibronectin and vitronectin. Since GPIIb/IIIa mediates platelet aggregation, its inhibition has generated much interest in the control of prothrombotic states [104,105]. An antibody against GPII/IIIa (7E3 also called Reopro or Abciximab) has been developed and has entered clinical trials with angioplasty [106], myocardial infarction [107] and unstable angina [107,108]. All clinical trials have shown a significant improvement of longterm survival [109,110]. However, higher complication rates such as bleeding and thrombocytopenia have also been observed [111,112]. New clinical trials are underway to determine appropriate regimens [110].
4.2. Platelets and biomaterials Platelet activation (platelet release, PMP formation, P-selectin expression, aggregation) and adhesion is known to occur [1,58] during cardiopulmonary bypass, hemodialysis, as well as with vascular grafts and catheters. The thrombotic complications associated with cardiovascular devices are linked clearly to their ability to activate platelets. Adherent platelets [113] and circulating PMPs generated by material contact [100,101] have been shown to be procoagulant in nature. Association between platelets and leukocytes via Pselectin also occurs in the presence of cardiovascular devices [8] and such associations have become a relatively new parameter to study biocompatibility. However, the implications of this association are mostly unknown: they may directly or indirectly contribute to thrombin generation (via monocyte TF) or participate in the removal of platelets from the circulation since several platelets can be bound to each neutrophil or monocyte. While it is intuitive to suggest that a non-thrombogenic surface should not support platelet adhesion it has not, unfortunately, been that simple. It has been found that platelet contacts with some biomaterial surfaces results in activation leading to high consumption (removal from circulation) characterized by high platelet turnover rather than adherence and by the formation of microemboli rather than occlusive thrombi. Following contact with the layer of adsorbed proteins on the artificial surface, platelets will either adhere or "bounce off" [114], most likely depending on their state of activation and the ligands present at the interface [115]. Platelet adhesion on surfaces is mediated by GPIIb/IIIa and fibrinogen and interaction with GPIb/IIa and vWF can also occur [116-119]. However, the absence of significant platelet adhesion does not preclude platelet
229
The Biomaterials Silver Jubilee Compendium M.B. Gorbet, M.V. Sefton / Biomaterials 25 (2004) 5681-5703
activation as shown by the generation of PMPs with PVA hydrogel both in vitro [120] and ex vivo [90]. Indeed, animal studies have shown that, despite the absence of platelet adhesion, blood contact with some hydrogel surfaces [121,122], like contact with NHLBI reference materials and Silastic | [123] appears to activate platelets, resulting in their removal from the circulation. Furthermore, Hanson et al. [122] demonstrated a direct linear relationship between the water content of hydrogels and the rate of platelet consumption in a baboon AV shunt model. They also noted that platelet consumption was reduced by the antiplatelet agent, dipyridamole [124], similar to its effectiveness in normalizing platelet survival in-patients with artificial heart valves [125]. While there is agreement that the rapidly adsorbed proteins, especially fibrinogen, play a critical role in platelet adhesion, it is not clear, for example, how effective adsorbed fibrinogen is as a platelet agonist in vivo and what other mechanisms are involved in supporting or initiating platelet adhesion. While adsorbed fibrinogen is likely the critical ligand for adhesion, what activates the platelet to adhere in the first place? Further it has not been clear that inhibition of platelet adhesion will inhibit the generation of PMPs or reduce platelet consumption. The strategies available to minimize platelet adhesion (e.g., polyethylene glycol immobilization) have not been sufficient to prevent platelet consumption [126]. The mechanism of materialinduced platelet activation is often presumed to be via the generation of thrombin due to activation of the intrinsic coagulation cascade or the release of ADP from damaged red blood cells or platelets. Even in the presence of heparin, small levels of thrombin generation are generated and may activate platelets. However, the inability of thrombin and kallikrein inhibitors to reduce platelet activation suggests that platelet activation is at least in part mediated by other agonists [127]. For example, a correlation between complement activation and thrombocytopenia has been noted during dialysis [128,129]. Complement activation can lead to platelet activation in many ways. Platelets possess a receptor for C lq that has been shown to induce GPIIb/IIIa activation, P-selectin expression and procoagulant activity [130]. It is currently not known how classical complement activation leads to activated platelets, but in vitro results support a role for C1 [89]. Insertion of C5b-9 in platelets has also been associated with increased platelet procoagulant activity [95]. In vitro studies using human cells have been conducted to probe the effectiveness of various agents to inhibit platelet activation [88] using a microsphere based immunoassay. Platelet adhesion to polystyrene microspheres was found to be unaffected by the complement inhibitor, sCR1 which is otherwise capable of inhibiting material-induced SC5b-9 produc-
5691
tion. In contrast, classical pathway complement inhibitors, pentamidine, benzamidine, pyridoxal-5-phosphate (P5P) and cysteine were able to inhibit platelet adhesion to the polystyrene surface. These agents also inhibited platelet loss and microparticle levels in whole blood after contact within polyethylene tubing. Benzamidine and a derivative, pentamidine, are able to competitively inhibit C ls enzyme (ionic interaction with active site), although pentamidine has an IC50 10 times lower [131-133]. The antiplatelet effects of pentamidine have been documented but the mechanism of action remains unresolved. It has been reported that pentamidine has no disruptive effect on GPIIb/IIIa receptors, intracellular cAMP levels, calcium ion influxes or intracellular pH changes [88,134,135]. P5P, a major coenzyme from Vitamin B6 is also a known inhibitor of C 1 fixation [136,137]. Cysteine is known to inhibit both C ls and the alternative complement pathway [138]. Since sCR1 (in vitro) was not able to block platelet adhesion and activation, it is our hypothesis that C1 is playing a role in material-induced platelet activation and that these agents are effective because of their ability to inhibit C ls. These agents also effectively inhibit other serine proteases such as thrombin, trypsin and plasmin [132] and thus unequivocal delineation of the mechanism is not yet possible. However, during in vitro blood-material contact, inhibiting complement activation has led to conflicting results on the role of the terminal complement pathway in material-induced platelet activation. Monoclonal antibodies to C5 and C8 inhibited platelet activation during simulated extracorporeal circulation (SECC) [139,140]. On the other hand, sCR1 had no effect on platelets in our microsphere assay [88,141] and in a different extracorporeal circulation model [142]. The difference in experimental conditions such as higher flow rate, blood dilution and the presence of mannitol (a hydroxyl scavenger) between the studies may account for the difference in the apparent efficacy of terminal pathway inhibition. There are likely multiple pathways whereby platelets are activated, some more relevant during conditions of high complement and leukocyte activation and some more relevant where these pathways are less well developed. Further research is needed to fully understand which complement protein triggers material-induced platelet activation. In our arterio-venous canine shunt model with a platelet consumptive PVA hydrogel tubing segment, systemic low molecular weight heparin did not reduce material-induced platelet damage: loss of platelets, microparticle generation or reduced lifespan, indicating that thrombin production did not appear to play a role. On the other hand, pentamidine (12 mg/kg, daily, IM) dramatically inhibited thrombocytopenia during the connection of PVA hydrogel test segments (Gemmell, 2001, pers. comm.). This supports our hypothesis that
230 5692
The Biomaterials Silver Jubilee Compendium M.B. Gorbet, M.V. Sefton / Biomaterials 25 (2004) 5681-5703
such inhibitors are useful for blocking biomaterialassociated platelet activation. We have also used ReoPro T M (up to 0.Smg/kg, GPIIb/IIIa antagonist) and not observed any effect on thrombocytopenia caused by PVA. The ReoPro T M dosage was sufficient to inhibit agonist induced platelet aggregation. This preliminary finding could suggest that platelet adhesion (and vesiculation)--both blocked by IIb/IIIa inhibitors--are not responsible for premature platelet consumption and clearance, at least on these smooth surfaces that have few platelet deposits in the absence of any therapeutic agents.
5. Leukocytes
5.1. Leukocyte biology Circulating leukocytes comprise neutrophils, monocytes, lymphocytes, basophils and eosinophils. Only neutrophils and monocytes in blood, but not after they emigrate into tissues, will be addressed in this review as they are the major players in the inflammatory response with cardiovascular devices. Neutrophils are the most abundant white blood cells, representing 40-60% of the leukocyte population (3-5 • 106 neutrophils/ml), while monocytes represent 5% with a concentration of 0.2-1 x 106 monocytes/ml. Under normal circumstances, neutrophils have a very short half-life in blood (8-20 h) but after an inflammatory stimulus, such as LPS or cytokines, their lifespan can increase up to three-fold [143] and their role may be even more active. Monocytes enter the circulation for a short period (36-104h) and they migrate into tissues where maturation and differentiation occur and they become macrophages [144]. They may also be deposited on injured blood vessels and later differentiate into
macrophages. It is important to note that when neutrophils and monocytes are recruited in tissues during inflammation, their half-life increases to several days. Monocytes and neutrophils possess receptors for different complement products and other pro-inflammatory mediators such as PAF and cytokines. Platelet release, such as/~-thromboglobulin and PDGF, has also been reported to activate neutrophils in vitro [145-147]. Other neutrophil and monocyte activating stimuli include bacteria and their products and cell adhesion. A list of the most relevant receptors involved in the inflammatory response is presented in Table 3. As for platelet activation, leukocyte activation results in several physiological responses. (1) Upon activation, changes in expression of membrane receptors occur on neutrophils and monocytes: upregulation of CD1 l b by translocation from intracellular granules [148], shedding of L-selectin by shedding [148], synthesis and expression of TF [149]. TF expression by leukocytes is the subject of another review [63]. (2) Leukocyte activation results in release of inflammatory mediators. Neutrophils contain three types of granules (gelatinase, specific and azurophil granules) and their contents may be released upon activation: among others, elastase, cathepsin G and lactoferrin are important inflammatory mediators. Cytokines such as IL-1, IL-6, IL-8, TNFe, G-CSF and GM-CSF are also released. Arachidonic acid metabolites, such as leukotriene B4 and PAF, are produced and released by activated neutrophils. The released inflammatory mediators have various properties: they may be chemoattractant for leukocytes, promote adherence to endothelial cells, and further activate platelets or leukocytes. (3) Activation may also result in the onset of the oxidative burst whereby neutrophils and monocytes release oxidants, such as O2 and H 2 0 2 . These products damage tissues and activate cells [150].
Table 3 Leukocyte receptors in acute/immediate inflammatory response [254,255] Receptor
(a) Complement receptors Clq R CR1 CR3 or CD 1lb CR4 or CD1 l c C3a R C5a R
(b) Other receptors (R) TNF~, IL-1 IL-8 PAF, LTB4 GM-CSF, G-CSF, IFN7
Ligand
Function
Clq, MBP
Enhance phagocytosis, respiratory burst Immune adherence, phagocytosis
C3b > C4b >iC3b iC3b, fibrinogen, FX, ICAM- 1 iC3b, fibrinogen C3a C5a
Phagocytosis, respiratory burst, adhesion Adhesion, phagocytosis Chemotaxis, degranulation, respiratory burst Chemotaxis, degranulation, respiratory burst
Degranulation, respiratory burst Chemotaxis, degranulation, respiratory burst Strong activation Weak activation, priming
231
The Biomaterials Silver Jubilee Compendium M.B. Gorbet, M.V. Sefton / Biomaterials 25 (2004) 5681-5703
5.2. Leukocyte activation and biomater&ls
Table 4 Leukocyte adhesion receptors Receptor
(a) Selectin family L-selectin
5693
Ligand Mucin-like ligand, lymph node addressin
(b) Integrin family CD 11a/CD 18 (or LFA- 1) CD 11b/CD 18 (CR3 or Mac- 1) CD 11c/CD 18 (CR4 or p 150,95) VLA-1 VLA-2 VLA-3 VLA-4 VLA-5
ICAM-1, ICAM-2, ICAM-3 iC3b, fibrinogen, FX, ICAM-1 iC3b, fibrinogen Laminin, collagen Collagen Fibronectin, laminin, collagen VCAM-1 Fibronectin
(c) Immunoglobulin family PECAM-1 FcT-RI, FcT-RII, FcT-RIII
PECAM- l, heparin Immune complexes
VLA: very late antigen.
(4) Activated neutrophils and monocytes also have an increased adhesive capacity on endothelium and other surfaces [151]. Leukocyte adhesion to the endothelium is an important means by which neutrophils and monocytes participate in the inflammatory response. Adherent leukocytes have been shown to be more activated than their counterpart in the bulk [152,153] but the level of activation of adherent leukocytes depends on the surface [154]. Leukocyte adhesion molecules are divided into three main families [155]: the selectins, the integrins and the immunoglobulin superfamily (Table 4). The mechanism of leukocyte adhesion to endothelial cells, a threestep mechanism, has been well characterized [156]. Step 1: L-selectin is involved in the initial rolling of leukocytes on endothelium. Step 2: The rolling stage enables leukocytes to slow their movement and sample the local environment, and they may become activated due to local stimulation and additional interaction between receptor/ligand. Step 3 : C D l l / C D 1 8 mediates firm adhesion. With activated neutrophils and monocytes, a functional upregulation is observed for CD 11a, while for CD1 l b and CD1 l c both a quantitative and functional upregulation occurs, the upregulation for CD1 l b being more rapid and important than for CDllc. It is important to note that the functional change of CD1 l b upon leukocyte activation and/or adhesion can occur despite no measurable increase in C D l l b surface expression. The functional change is conformational involving receptor phosphorylation and resulting in increase binding affinity for certain ligands (such as fibrinogen and Factor X) [157]. On the other hand, a quantitative increase in CD 11b expression on leukocytes does not imply increased adhesion, unless i t is accompanied by functional change in the receptor [158].
Contact with cardiovascular devices in vivo activates both neutrophils and monocytes. Indicators of leukocyte activation such as L-selectin shedding and CD1 l b upregulation on leukocytes have been widely observed following angioplasty [159-162], hemodialysis [163-165] and cardiopulmonary bypass [20,22,166-169] (for an extensive review of studies on the expression of leukocyte adhesion molecules with in vitro cardiopulmonary bypass, see Asimakopoulos and Taylor [170]). Degranulation with the release of elastase and lactoferrin [171-175] and the presence of cytokines [169,176,177], such as IL-1 and TNF~, have been associated with extracorporeal circulation and further demonstrate leukocyte activation. Activation of the respiratory burst is also a common trait with hemodialysis [164,178]. Material-induced leukocyte activation also results in increased adhesion. As the biomaterial is larger than a micro-organism and cannot then be engulfed by leukocytes, adherent neutrophils and monocytes undergo a frustrated phagocytosis whereby they release their array of potent oxygen metabolites and proteolytic enzymes [144]. Material characteristics and proteins at the interface appear to modulate the level of activation of adherent leukocytes [153,179,180]. In vivo studies have found activated leukocytes adhering to stents [181,182], oxygenators [183] and hemodialysis membranes [184,185]. Material-activated leukocytes also adhere to the endothelium, such as at the anastomoses of a vascular graft or in the lung during extracorporeal circulation. Heparinized human whole blood contact with a PVA hydrogel surface in vitro for 1 h lead to a two-fold upregulation of C D l l b , typical of the degree of leukocyte activation induced by phorbol esters [186]. In contrast, whole blood contact with PE and Silastic T M surfaces resulted in minimal CD1 l b upregulation. We have also reported that many clinical materials can activate isolated neutrophils (without platelets) suspended in plasma and that fibrinogen adsorption (plasma but not serum pretreatment) enhances activation [153]. Longer exposures (2h) to whole blood with polystyrene beads and polystyrene beads grafted with polyethylene glycol (PS-PEG) lead to expression of monocyte TF [187] as well as C D l l b upregulation. Activation was dependent on the surface area to volume ratio, but there was no difference in the extent of activation in comparing PS beads with PS-PEG (or PSPEG-NH2) beads. On the other hand, monocyte TF expression and adherent platelet density were all greater on PS than PS-PEG; there were no differences in leukocyte adhesion densities. The mechanisms of leukocyte adhesion on artificial surfaces are not clear, but it appears to be mediated in
232 5694
The Biomaterials Silver Jubilee Compendium M.B. Gorbet, M.V. Sefton / Biomaterials 25 (2004) 5681-5703
part by the complement product iC3b [72,144]. This is supported by in vitro work showing that inhibition of complement activation in vitro significantly reduced leukocyte adhesion [153,188-190] while fibrinogen also appears to play an important role in leukocyte adhesion to materials [191,192]. In experiments with isolated leukocytes, however, it appears that monocyte TF expression is only partly dependent on complement inhibition but appears to be also dependent on the presence of platelets [64]. The removal of platelets or the blocking of GPIIb/IIIa by monoclonal antibody (7E3) partially lowered the degree of material-induced TF expression while having no effect on CDllb. The presence of platelets on the surface may also mediate leukocyte adhesion via the interaction between P-selectin and PSGL-1 and/or GPIIb/IIIa and CD1 lb [193,194]. Conflicting reports also exist on the requirement for platelets in leukocyte adhesion on artificial surfaces [195,196]. The mechanisms of material-induced leukocyte activation as distinct from adhesion also remain unknown. Whether they are directly activated by contact with a foreign surface, via complement activation, kallikrein or platelet activation has not been fully determined. In vitro and in vivo investigations with protease inhibitors [127,175,197-199], complement inhibitors [21,22,186,200-203] and antiplatelet agents [106] suggest that they all play a role, but no one inhibitor has led to consistent results with a significant reduction of material-induced leukocyte activation. Material-induced leukocyte activation may be mediated by several factors and inhibition of one pathway of activation may not be sufficient to result in a significant impact on leukocyte activation. For example, we have shown that complement inhibition via sCR1 was only partially effective in reducing leukocyte activation [186]. On the other hand, a combination of sCR1 and antiGPIIb/IIIa (which blocks platelet activation) reduced material-induced leukocyte activation to almost background levels. The complexities of leukocyte activation by the PEG modified materials and the inhibitory effects of sCR1 and pentamidine are discussed elsewhere [204]. In summary, the mechanisms that regulate materialinduced leukocyte activation are as yet not well understood, precluding a clear scientific basis for strategies for a significant reduction in the inflammatory response induced by cardiovascular devices.
5.3. Leukocytes, platelets and coagulation Circulating monocytes and neutrophils normally roll on the endothelium. They will however adhere to damaged or stimulated endothelial cells or adherent platelets and further contribute to localized thrombogenesis. We have evidence [64,187] that similar phenom-
ena apply to biomaterials. The different procoagulant activities of leukocytes may be classified as: 9 Membrane-associated procoagulant activity: via TF
expression on the cell membrane (TF-dependent coagulation pathway) or via TF-independent mechanisms through factor X binding to C D l l b receptors leading to factor Xa generation or fibrinogen binding to CD1 l b; or binding of the prothrombinase complex on the membrane. 9 Release of procoagulant mediators: degranulation and oxidative products have the capacity to neutralize various anticoagulant proteins and activate platelets. 9 Association between platelets and neutrophils or monocytes: their interactions may lead to mutual activation and to a microenvironment protected from inhibitors. For example, following blood contact with cardiopulmonary bypass circuits or ventricular assist devices, TF expression on monocytes has been observed in vitro [205-207] and in vivo [60,61,206]. It has also been shown that C D l l b , upregulated on monocytes by cardiopulmonary bypass, was able to directly activate factor X [208] and platelet-leukocyte aggregates have been observed in several scenarios [106,209-211] as noted above. The potential role of leukocytes in thrombogenesis is underscored by the number of studies that have tried to minimize thrombus formation by the administration of drugs specifically targeted at leukocytes. Antibodies to block leukocyte adhesion may prove to be a reasonable therapeutic approach in the prevention of thrombus formation as illustrated in in vivo baboon models [212,213]. Overall, however, there is relatively little known on the potential contribution of expression of leukocyte procoagulant activities to thrombogenesis and thrombotic complications associated with the use of biomaterials and cardiovascular devices.
6. Other important factors
6.1. Flow Fluid dynamics affects the growth of thrombi and the deposition of fibrin. The composition difference between arterial and venous thrombi is one old example of this, although the underlying mechanisms are still not well understood. Thorough reviews are available [5,65, 214-216]. Flow determines the rates of transport of cells and proteins to the surface; it can also change the level of receptor expression on platelets and leukocytes. As platelets are an important part of the thrombus, the effect of shear on platelets has been studied extensively. Higher shear results in higher platelet deposition and lower fibrin deposition, while at lower shear the inverse
233
The Biomaterials Silver Jubilee Compendium M.B. Gorbet, M.V. Sefton / Biomaterials 25 (2004) 5681-5703
is true [214]. High shear, such as the ones observed at stenotic plaques, is also able to induce platelet aggregation even in the absence of any other exogenous factors [217,218]. Conflicting results have been obtained on the effect of flow on leukocyte adhesion while little is known of its effect on leukocyte activation. High shear has been shown to either reduce, increase or leave unchanged leukocyte adhesion on different substrates [195, 219-225]. These conflicting results may be explained by differences in experimental conditions: the presence of red blood cells [220], platelets [195,223,226] and plasma proteins [195]; the surface studied [223] and the state of leukocyte activation [224]. As for the effects of flow on the coagulation cascade, it has been studied less. Current knowledge is limited to Factor Xa generation initiated by the extrinsic pathway and thrombin generation initiated by the intrinsic pathway (with biomaterials). Factor Xa generation by the complex TF:VIIa increases with shear rate (and shear stress) [227,228]. For thrombin generation by the intrinsic pathway, modeling has identified three types of reactions [65]: at low flow, a significant amount of thrombin is produced after a long lag time (over 10 h); at moderate flow, significant thrombin generation is produced in a short time (within minutes); at high flow, low levels of thrombin are produced within seconds. Turbulent flow can be present at anastomoses, joints, and bifurcations of cardiovascular devices and such turbulence also contributes to the observed thrombosis. It is believed to play a significant role in the failure of mechanical heart valves, for example. Turbulent flow (in distinction to recirculation and stagnation zones, which may or may not have the characteristics of turbulence) results in hemolysis and/or cell activation but the mechanisms leading to thrombus formation are still poorly understood. Platelet deposition remains the focus of most studies [229] but the literature on turbulent flow and thrombus formation is more limited. Nonetheless, much effort is done to design devices so that recirculation or stagnation zones are avoided, since these are known niduses for thrombus growth. While the importance of flow has been recognized, our current understanding of its mechanisms is limited mostly to platelets. Many in vitro and in vivo flow models are available and have been successfully used to assess antithrombotic drugs in whole blood [216]. More fundamental research is required on blood coagulation, leukocytes and flow. Previous research with flow has focused on isolated cells or proteins, which is far from the in vivo situation. The critical role of red blood cells in the in vitro study of mechanisms of leukocyte adhesion was recently demonstrated by Melder et al. [230]. In the absence of erythrocytes, blocking L-selectin had no effect on lymphocyte adhesion to activated endothelial cells, which was in contradiction with their in vivo observation. Upon addition of erythrocytes to
5695
the lymphocytes, L-selectin was then shown to play a significant role in adhesion especially under high shear rate. The use of more physiological experimental conditions (e.g., presence of red blood cells, plasma proteins) should result in significant advances in our knowledge on the effect of mechanical factors on thrombosis and hemostasis.
6.2. Endotoxin Endotoxins, also called LPS, are the component of the outer membrane of gram-negative bacteria and are released into the circulation upon disruption of the intact bacteria (death, cell lysis) [231]. Endotoxin is commonly found everywhere in our environment and it is the most significant pyrogen in parenteral drugs and medical devices. Endotoxins are also present in the digestive system. Their presence in the blood stream may cause septic reactions with a variety of symptoms such as fever, hypotension, nausea, shivering and shock [232]. High concentrations can lead to serious complications such as disseminated intravascular coagulation (DIC), endotoxin shock and adult respiratory distress syndrome (ARDS). Endotoxins are known to activate complement, the kinin system, leukocytes, platelets and endothelial cells [231,232] and are the "enemy" of both in vitro and in vivo study of blood-material interactions. In vivo, they may lead to the complications mentioned above, while in vitro, the presence of this contaminant may affect the results and compromise the conclusions. FDA regulates the acceptable level of endotoxin contamination with medical devices to be 0.5 endotoxin units/ml [233]. There have been few reports of endotoxin contamination with the use of cardiovascular devices. During cardiopulmonary bypass and extracorporeal membrane oxygenation, the presence of endotoxins has been observed in vivo [234,235]. They appear to originate mostly from the gut [236-238] rather than from the materials and are believed to be a reaction to the surgical procedure. During hemodialysis, endotoxin contamination is also an issue and the dialysate is usually the source [232,239]. While endotoxin contamination may be present in vivo in some patients and studies, there has been no investigation showing a significant correlation between the magnitude of endotoxin contamination and postoperative complications [235,240]. On the other hand, endotoxin contamination during in vitro work may be much more common, as sterile conditions are not always available and the laboratoryworking environment contributes to their presence. The most conspicuous source of endotoxin may actually be the water since distillation and deionizing columns do not remove endotoxin. Endotoxin has an effect on platelets only at high concentration (over l~tg/ml
234 5696
The Biomaterials Silver Jubilee Compendium M.B. Gorbet, M.V. Sefton / Biomaterials 25 (2004) 5681-5703
equivalent to 5000 EU/ml) [241], while leukocytes have been reported to be activated by endotoxin concentrations as low as 0.01 ng/ml (equivalent to 0.05 EU/ml) [242,243]. When studying blood-material interactions, endotoxin may be contained in buffers and/or on materials, and its priming and activating effect on leukocytes may affect the observed results. However, it is important to consider that all the studies performed on the effect of endotoxin on leukocyte activation used purified strains of endotoxin while the endotoxin present in laboratory materials and buffers are of an environmental nature. Purified endotoxins are much more potent than environmental endotoxins [244] and even among purified endotoxins, their activity might vary [242,245]. Contrary to a purified strain of endotoxin [239], the presence of relatively high levels of environmental endotoxins (100 EU/ml) was shown to have little impact on the leukocyte response to hemodialysis [246] in vivo. But when tested in vitro, environmental contamination of a material may have a dramatic effect on the results. In the orthopedic area, recent studies [258,259] have focussed on endotoxin contamination of microparticles used to assess the effect of wear debris in vitro and have confirmed the significant effect of environmental endotoxin contamination on cytokines. Many washing procedures are now available to ensure endotoxin removal from materials [247,248] and should ensure that the study of blood-material interactions is not impaired by the presence of endotoxin.
7. Conclusions
cytes; the ability of C D l l b to bind Factor X and fibrinogen; the ability of released inflammatory mediators to activate platelets and block inhibitors of coagulation; and by promoting the association between leukocytes and platelets. In the last 5 years, many leukocyte investigators have discussed the participation of inflammatory cells in coagulation [249-252]. Thrombosis is viewed now more as a multicellular event rather than just a platelet event [253]. In certain situations, blocking leukocyte contributions to thrombin generation may appear to be a reasonable means to reduce the occurrence of thrombotic complications. Such nontraditional approaches to thrombosis control with biomaterials may be a useful opportunity for further study. The mechanism of biomaterial-associated thrombosis is not fully clear. The role of Factor XII is uncertain while that of TF has not been directly assessed. Both the mechanisms of leukocyte and platelet activation by materials remain to be further elucidated. As noted above in the context of coagulation, the timing of the events contributing to thrombin formation is also a complex issue. Both Factor XII activation and platelet activation are able to generate thrombin formation within minutes while thrombin generation via leukocyte TF requires hours since TF has to be synthesized. The contribution of leukocyte proteases will also be affected by time since their effect will be dependent on the presence of inhibitors and other inflammatory mediators that can potentiate their action. As illustrated in Fig. 5, the time course of the underlying steps in biomaterialassociated thrombosis may need more consideration
The complexity of blood-material interactions explains our failure to design a material that is entirely blood-compatible. Our current stage of knowledge is far from providing us with a complete mechanism of material-induced thrombin generation. One issue has been the natural scientific tendency to focus on individual aspects of the whole problem rather than considering the various interactions. For example, the biomaterials community has typically looked at platelet interactions in platelet-rich plasma and so is unable to explore interactions between leukocytes and platelets. Alternatively, an anticoagulant prevents thrombin effects from being considered. Of course, without anticoagulants or with whole blood, the experiments get too complicated or sometimes impossible to perform or analyze. Unfortunately, simplifying the system has not allowed us to make real progress. We have also separated thrombosis from its normal context in inflammation and wound healing. The molecular links between inflammation and thrombosis are undeniable. Inflammation, as characterized by a leukocyte response to a stimulus, may contribute to thrombin generation by the TF expression on mono-
Fig. 5. Time scales of biomaterial-associated thrombogenicity. Each component is associated with a different time scale. Protein adsorption and Factor XII activation (assuming it is relevant) occurs within seconds of blood-material contact, producing low levels of thrombin. Platelet activation occurs within minutes creating the phospholipid surface required for assembly of the platelet bound coagulation enzymes and the production of enough thrombin to cause substantial fibrin formation. Leukocyte activation ( C D l l b upregulation) also occurs within minutes leading to adhesion while TF expression occurs over hours (as transcription and translation must occur first). Complement activation occurs at all these time scales, but whether it is the prelude to leukocyte activation and/or platelet activation is not clear; we suspect that these three are highly interlinked. The interplay between these components will vary depending on the time scale of the situation: a low time constant situation (e.g., high flow, straight tube) would involve different components than a high time constant (e.g., low flow, stagnation zone) situation.
235
The Biomaterials Silver Jubilee Compendium M.B. Gorbet, M.V. Sefton / Biomaterials 25 (2004) 5681-5703
than has hitherto been the case. Experimental studies already have difficulty making the transition from the first few seconds of protein adsorption to the first few minutes of coagulation and cell adhesion. What happens over hours to days as leukocytes synthesise TF and thrombotic deposits become remodeled is almost beyond current experimental capacity. Whether thrombosis leads to passivation or embolization or some other long-term consequence is still largely unknown. The solution to thrombotic complications associated with cardiovascular devices may not be to try to create a new material that will elicit the proper blood response: the inert cardiovascular biomaterial may be impossible. Rather, more success may be achieved by preventing the adverse effects of a biomaterial by actively blocking the pathway responsible for the inherent thrombogenicity of the materials. Rather than minimizing non-specific biomaterial-associated activation, active inhibition may be the only recourse. Despite more than 50 years of research on blood-material interactions, many questions remain unanswered. The intent of this review was to summarize what we know and to highlight what we have yet to learn.
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
References [1] Hanson SR. Device thrombosis and thromboembolism. Cardiovasc Pathol 1993;2:157S-65S. [2] Bittl JA. Coronary stent occlusion: thrombus horribilis. J Am Coll Cardiol 1996;28:368-70. [3] Clagett GP, Eberhart RC. Artificial devices in clinical practice. In: Colman RW, Hirsh J, Marder VJ, Salzman EW, editors. Hemostasis and thrombosis: basic principles and clinical practice. Philadelphia: Lippincott; 1994. p. 1486-505. [4] Bick RL. Hemostasis defects with cardiac surgery, general surgery and prosthetic devices. In: Bick RL, editor. Disorders of hemostasis and thrombosis. Chicago: American Society of Clinical Pathologist Press; 1992. p. 195-222. [5] Eberhart RC, Clagett CP. Catheter coatings, blood flow, and biocompatibility. Semin Hematol 1991;28:42-8. [6] Edmunds LHJ. Is prosthetic valve thrombogenicity related to design or material? Tex Heart Inst J 1996;23:24-7. [7] Geiser T, Sturzenegger M, Genewein U, Haeberli A, Beer JH. Mechanisms of cerebrovascular events as assessed by procoagulant activity, cerebral emboli and platelet microparticles in patients with prosthetic heart valves. Stroke 1998;29: 1770-7. [8] Mickelson JK, Lakkis NM, Villarreal-Levy G, Hughes BJ, Smith CW. Leukocyte activation with platelet adhesion after coronary angioplasty: a mechanism for recurrent disease? J Am Coll Cardiol 1996;28:345-53. [9] Wendel HP, Ziemer G. Coating techniques to improve the hemocompatibility of artificial devices used for extracorporeal circulation. Eur J Cardiothorac Surg 1999; 16:342-50. [10] Defraigne JO, Pincemail J, Larbuisson R, Blaffart F, Limet R. Cytokine release and neutrophil activation are not prevented by heparin-coated circuits and aprotinin administration. Ann Thorac Surg 2000;69:1084-91. [11] Moen O, Hogasen K, Fosse E, et al. Attenuation of changes in leukocyte surface markers and complement activation with
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27] [28] [29] [30]
5697
heparin-coated cardiopulmonary bypass. Ann Thorac Surg 1997;63:105-11. Baufreton C, Jansen PG, Le BP, et al. Heparin coating with aprotinin reduces blood activation during coronary artery operations. Ann Thorac Surg 1997;63:50-6. Muehrcke DD, McCarthy PM, Kottke-Marchant K, et al. Biocompatibility of heparin-coated extracorporeal bypass circuits: a randomized, masked clinical trial. J Thorac Cardiovasc Surg 1996;112:472-83. Gorman RC, Ziats N, Rao AK, et al. Surface-bound heparin fails to reduce thrombin formation during clinical cardiopulmonary bypass. J Thorac Cardiovasc Surg 1996;111:1-11. Redmond JM, Gillinov AM, Stuart RS, et al. Heparin-coated bypass circuits reduce pulmonary injury. Ann Thorac Surg 1993; 56:474-8. Levy M, Hartman AR. Heparin-coated bypass circuits in cardiopulmonary bypass: improved biocompatibility or not. Int J Cardiol 1996;53:$81-7. Fosse E, Thelin S, Svennevig JL, et al. Duraflo II coating of cardiopulmonary bypass circuits reduces complement activation, but does not affect the release of granulocyte enzymes: a European multicentre study. Eur J Cardiothorac Surg 1997;11: 320-7. Videm V, Mollnes TE, Bergh K. Heparin-coated cardiopulmonary bypass equipment. II. Mechanisms for reduced complement activation in vivo. J Thorac Cardiovasc Surg 1999;117: 803-9. Janvier G, Baquey C, Roth C, Benillan N, Belisle S, Hardy J-F. Extracorporeal circulation, hemocompatibility and biomaterials. Ann Thorac Surg 1996;62:1926-34. Videm V, Mollnes TE, Fosse E. Heparin-coated cardiopulmonary bypass equipment. I. Biocompatibility markers and development of complications in high risk population. J Thorac Cardiovasc Surg 1999;117:794-802. Gillinov AM, DeValeria PA, Winkelstein JA, et al. Complement inhibition with soluble complement receptor type 1 in cardiopulmonary bypass. Ann Thorac Surg 1993;55:619-24. 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; 100:2499-506. Murkin JM. Cardiopulmonary bypass and the inflammatory response; a role for serine protease inhibitor. J Cardiothorac Vasc Anesth 1997;11:19-23. Sundaram S, Gikakis N, Hack CE, et al. Nafamostat mesilate, a broad spectrum protease inhibitor, modulates platelet, neutrophil and contact activation in simulated extracorporeal circulation. Thromb Haemost 1996;75:76-82. Rao AK, Sun L, Gorman JH, Edmunds Jr LH. GPIIb/IIIa receptor antagonist Tirofiban inhibits thrombin generation during cardiopulmonary bypass in baboons. Thromb Haemost 1999;82:140-4. Vertress RA, Tao W, Kramer GC, et al. Tumor necrosis factor monoclonal antibody prevents alterations in leukocyte populations during cardiopulmonary bypass. ASAIO J 1994;40: M554-9. Schmaier AH. Contact activation: a revision. Thromb Haemost 1997;77:101-7. Furie B, Furie BC. The molecular basis of blood coagulation. Cell 1988;53:505-18. Blajchman MA, Ozge-Anwar AH. The role of the complement system in hemostasis. Prog Hematol 1986;XIV:149-82. Jetsy J, Nemerson Y. The pathways of blood coagulation. In: Beutler E, Lichtman MA, Coller BS, Kipps TJ, editors. Williams hematology. 5th ed. New York: McGraw-Hill; 1995. p. 1227-38.
236 5698
The Biomaterials Silver Jubilee Compendium M.B. Gorbet, M.V. Sefton / Biomaterials 25 (2004) 5681-5703
[31] Rapaport SI, Rao VM. Initiation and regulation of the tissue factor-dependent blood coagulation. Arterioscler Thromb 1992; 12:1111-21. [32] Morrissey JH, Macik BG, Neuenschwander PF, Comp PC. Quantitation of activated factor VII levels in plasma using a tissue factor mutant selectively deficient in promoting factor VII activation. Blood 1993;81:734-44. [33] Osterud B, Rapaport SI. Activation of Factor IX by the reaction product of tissue factor and Factor VIIa; additional pathway for initiating blood coagulation. Proc Natl Acad Sci USA 1977;74: 5260-4. [34] Lawson JH, Kalafatis M, Stram S, Mann KG. A model for the tissue factor pathway to thrombin. An empirical study. J Biol Chem 1994;269:23357-66. [35] Rand MD, Lock JB, van't Veer C, Gaffney DP, Mann KG. Blood clotting in minimally altered whole blood. Blood 1996;88: 3432-45. [36] Bauer KA, Rosenberg RD. Control of coagulation factors. In: Beutler E, Lichtman MA, Coller BS, Kipps TJ, editors. Williams hematology. 5th ed. New York: McGraw-Hill; 1995. p. 1239-50. [37] Hirsh J. Oral anticoagulant drugs. N Engl J Med 1991;324: 1865-75. [38] Novotny WF, Girard TJ, Miletich JP, Broze Jr GB. Platelets secrete a coagulation inhibitor functionally and antigenetically similar to the lipoprotein associated coagulation inhibitor. Blood 1988;72:2020-5. [39] McGee MP, Foster S, Wang X. Simultaneous expression of tissue factor and tissue factor pathway inhibitor by human monocytes. A potential mechanism for localized control blood coagulation. J Exp Med 1994;179:1847-54. [40] Cooke ED, Bowcock SA, Lloyd M J, Pilcher MJ. Intravenous lignocaine in prevention of deep venous thrombosis after elective surgery. The Lancet 1977;ii:797-9. [41] Higuchi DA, Wun TC, Likert KM, Broze GJ. The effect of leukocyte elastase on tissue factor pathway inhibitor. Blood 1992;79:1712-9. [42] Esmon CT. Regulatory mechanisms in hemostasis: natural anticoagulants. In: Colman RWet al,, editor. Hemostasis and thrombosis. Philadelphia: Lippincott; 1997. p. 1597-9. [43] Horbett TA. Principles underlying the role of adsorbed plasma proteins in blood interactions with foreign materials. Cardiovasc Pathol 1993;2:137S-48S. [44] Ziats N, Pankowsky DA, Tierney BP, Ratnoff OD, Anderson JM. Adsorption of Hageman Factor (Factor XII) and other human plasma proteins to biomedical polymers. J Lab Clin Med 1990;116:687-96. [45] Mulzer SR, Brash JL. Identification of plasma proteins adsorbed to hemodialyzers during clinical use. J Biomed Mater Res 1989; 23:1483-504. [46] Cornelius RM, Brash JL. Identification of proteins adsorbed to hemodialyser membranes from heparinized plasma. J Biomater Sci Polym Ed 1993;4:291-304. [47] Matata BM, Courtney JM, Sundaram S, et al. Determination of contact phase activation by the measurement of the activity of supernatant and membrane surface-adsorbed factor XII: its relevance as a useful parameter for the in vitro assessment of haemodialysis membranes. J Biomed Mater Res 1996;31:63-70. [48] Elam J-H, Nygren H. Adsorption of coagulation proteins from whole blood on to polymer materials: relation to platelet activation. Biomaterials 1992;13:3-8. [49] Van der Kamp KWHJ, van Oweren W. Factor XII fragment and kallikrein generation in plasma during incubation with biomaterials. J Biomed Mater Res 1994;28:349-52. [50] Van der Kamp KWHJ, Kauch KD, Feijen J, Horbett TA. Contact activation during incubation of five different polyurethanes or glass in plasma. J Biomed Mater Res 1995;29:1303-6.
[51] Hong J, Nilsson Ekdahl K, Reynolds H, Larsson R, Nilsson B. A new in vitro model to study interaction between whole blood and biomaterials. Studies of platelet and coagulation activation and the effect of aspirin. Biomaterials 1999;20:603-11. [52] Blezer R, Willems GM, Cahalan PT, Lindhout T. Initiation and propagation of blood coagulation at artificial surfaces studied in a capillary flow reactor. Thromb Haemost 1998;79:296-301. [53] Irvine L, Sundaram S, Courtney JM, Taggart DP, Wheatley DJ, Lowe GDO. Monitoring of Factor XII activity and granulocyte elastase release during cardiopulmonary bypass. ASAIO Trans 1991 ;XXXVII: 569-71. [54] Boisclair MD, Lane DA, Philippou H. Mechanisms of thrombin generation during surgery and cardiopulmonary bypass. Blood 1993;82:3350-7. [55] Boisclair MD, Philippou H, Lane DA. Thrombogenic mechanisms in the human: fresh insights obtained by immunodiagnostic studies of coagulation markers. Blood Coagul Fibrinolysis 1993;4:1007-21. [56] Burman JF, Chung HI, Lane DA, et al. Role of Factor XII in thrombin generation and fibrinolysis during cardiopulmonary bypass. Lancet 1994;344:1192-3. [57] Colman RW. Mechanisms of thrombus formation. Cardiovasc Pathol 1993;2:23S-31S. [58] Ratner BD. The blood compatibility catastrophe. J Biomed Mater Res 1993;27:283-7. [59] Edmunds Jr LH. Is prosthetic valve thrombogenicity related to design or material? Tex Heart Inst J 1996;23:24-7. [60] Chung JH, Gikakis N, Rao K, Drake TA, Colman RW, Edmunds LHJ. Pericardial blood activates the extrinsic coagulation pathway during clinical cardiopulmonary bypass. Circulation 1996;93:2014-8. [61] Wilhelm CR, Ristich J, Kormos RL, Wagner WR. Monocyte tissue factor expression and ongoing complement generation in ventricular assist devices patients. Ann Thorac Surg 1998;65: 1071-6. [62] Ernofsson M, Thelin S, Siegbahn A. Monocyte tissue factor expression, cell activation and thrombin formation during cardiopulmonary bypass: a clinical study. J Thorac Cardiovasc Surg 1997;113:576-84. [63] Gorbet MB, Sefton MV. Leukocyte activation and procoagulant activities, in preparation. [64] Gorbet MB, Sefton MV. Material-induced tissue factor expression but not CD1 l b upregulation depends on the presence of platelets. J Biomed Mater Res 2003;67A:792-800. [65] Basmadjian D, Sefton MV, Baldwin SA. Coagulation on biomaterials in flowing blood: some theoretical considerations. Biomaterials 1997; 18:1511-22. [66] Hirsh J, Weitz JI. Antithrombin therapy. In: Califf RM, editor. Acute myocardial infarction and other acute ischemic syndromes. St Louis: Mosby; 1996. [67] Hirsh J, Raschke R, Warkentin TE, Dalen JE, Deykin D, Poller L. Heparin: mechanism of action, pharmacokinetics, dosing considerations, monitoring, efficacy and safety. Chest 1995;108: 258S-75S. [68] Rubens FD, Weitz JI, Brash JL, Kinlough-Rathbone RL. The effect of antithrombin III-independent thrombin inhibitors and heparin on fibrin accretion onto fibrin-coated polyethylene. Thromb Haemost 1993;69:130-4. [69] Gustafsson D, Elg M. The pharmacodynamics and pharmacokinetcis of the oral direct thrombin inhibitor ximelagatran and its active metabolite melagatran: a mini-review. Thromb Res 2003;109:$9-15. [70] Edens RE, Lindhart RJ, Weiler JM. Heparin is not just an anticoagulant anymore: six and one-half decades of studies on the ability of heparin to regulate complement activity. Complement Profiles 1993;1:96-120.
The Biomaterials Silver Jubilee Compendium
237
M.B. Gorbet, M.V. Sefton / Biomaterials 25 (2004) 5681-5703
[71] Misoph M, Schwender S, Babin-Ebell J. Response of the cellular immune system to cardiopulmonary bypass is independent of the applied pump type and of the use of heparin-coated surfaces. Thorac Cardiovasc Surg 1998;46:222-7. [72] Remes R, Williams DF. Immune response in biocompatibility. Biomaterials 1992;13:731-43. [73] Hakim RM. Complement activation by biomaterials. Cardiovasc Pathol 1993;2:187S-97S. [74] Kazatchkine MD, Carreno M-P. Activation of the complement system at the interface between blood and artificial surfaces. Biomaterials 1988;9:30-5. [75] Johnson RJ. Complement activation during extracorporeal therapy: biochemistry, cell biology and clinical relevance. Nephrol Dial Transplant 1994;9:36-45. [76] Chenoweth DE. Complement activation during hemodialysis: clinical observations, proposed mechanisms, and theoretical implications. Artif Organs 1984;8:281-90. [77] Agostini A, Gardinali M. Complement activation during hemodialysis. J Biomater Appl 1989;4:102-22. [78] Videm V, Fosse E, Mollnes TE, Garred P, Svennevig JL. Time for new concepts about measurement of complement activation by cardiopulmonary bypass? Ann Thorac Surg 1992;54:725-31. [79] Kottke-Marchant K, Anderson JM, Miller KM, Marchant RE, Lazarus H. Vascular graft-associated complement activation and leukocyte adhesion in an artificial circulation. J Biomed Mater Res 1987;21:379-97. [80] Shepard AD, Gelfand JA, Callow AD, O'Donnell Jr TF. Complement activation by synthetic vascular grafts. J Vasc Surg 1984;1:829-38. [81] Johnson RJ. Immunology and the complement system. In: Ratner BD, Hoffman AS, Schoen FJ, Lemons JE, editors. Biomaterials science. San Diego: Academic Press; 1996. p. 173-88. [82] Craddock PR, Fehr J, Dalmasso AP, Brighan KL, Jacob HS. Hemodialysis leukopenia. Pulmonary vascular leukostasis resulting from complement activation by dialyzer cellophane membranes. J Clin Invest 1977;59:879-88. [83] Tulunay M, Demiralp S, Tatsan S, et al. Complement (C3, C4) and C-reactive protein responses to cardiopulmonary bypass and protamine administration. Anaesth Intens Care 1993;21:50-5. [84] Lhotta K, Wurzner R, Kronenberg F, Opperman M, Konig P. Rapid activation of the complement system by cuprophane depends on complement component C4. Kidney Int 1998;53: 1044-51. [85] Thylen P, Fernvik E, Lundhal J, Hed J, Jacobson SH. Modulation of CD1 lb/CD18 on monocytes and neutrophils following hemodialysis membrane interaction in vitro. Int J Artif Organs 1996; 19:156-63. [86] Nilsson UR, Larm O, Nilsson B, Storm KE, Elwing H, Nilsson Ekdahl K. Modification of the complement binding properties of polystyrene; effects of end-point heparin attachment. Scand J Immunol 1993;37:349-54. [87] Sefton MV, Gemmell CH, Gorbet MB. What really is blood compatibility. J Biomater Sci Polym Ed 2000;11:1165-82. [88] Gemmell CH. Platelet adhesion onto artificial surfaces: inhibition by benzamidine, pentamidine, and pyridoxal-5-phosphate as demonstrated by flow cytometric quantification of platelet adhesion to microspheres. J Lab Clin Med 1998;131:84-92. [89] Gemmell CH. Activation of platelets by in vitro whole blood contact with materials: increases in microparticle, procoagulant activity, and soluble P-selectin blood levels. J Biomater Sci Polym Ed 2001;12(8):933-43. [90] Gemmell CH, Yeo EL, Sefton MV. Flow cytometric analysis of material-induced platelet activation in a canine model: elevated microparticle levels and reduced platelet life span. J Biomed Mater Res 1997;37:176-81.
5699
[91] Blockmans D, Deckmyn H, Vermylen J. Platelet activation. Blood Rev 1995;9:143-56. [92] Marcus AJ. Platelet activation. In: Fuster V, Ross R, Topol EJ, editors. Atherosclerosis and coronary artery disease. Philadelphia: Lippincott-Raven Publishers; 1996. p. 607-37. [93] Rinder HM, Tracey JL, Rinder CS, Leitenberg D, Smith BR. Neutrophil but not monocyte activation inhibits P-selectinmediated platelet adhesion. Thromb Haemost 1994;72: 750-6. [94] Rinder HM, Bonan JL, Rinder CS, Ault KA, Smith BR. Dynamics of leukocyte-platelet adhesion in whole blood. Blood 1991;78:1730-7. [95] Sims PJ, Wiedmer T. Induction of cellular procoagulant activity by the membrane attack complex of complement. Semin Cell Biol 1995;6:275-82. [96] Jy W, Mao W-W, Horstman LL, Tao J, Ahn YS. Platelet microparticles bind, activate and aggregate neutrophils in vitro. Blood Cells Mol Dis 1995;21:217-31. [97] Siljander P, Carpen O, Lassila R. Platelet derived microparticles associate with fibrin during thrombosis. Blood 1996;87:4651-63. [98] Holme PA, Solum NO, Brosstad F, Pedersen T, Kveine M. Microvesicles bind soluble fibrinogen, adhere to immobilized fibrinogen and coaggregates with platelets. Thromb Haemost 1998;79:389-94. [99] Tans G, Rosing J, ChristeUa M, et al. Comparison of anticoagulant and procoagulant activities of stimulated platelets and platelet-derived microparticles. Blood 1991;77:2641-8. [100] Nieuwland R, Bercmans RJ, Rotteveel-Eijkman RC, et al. Cellderived microparticles generated in patients during cardiopulmonary bypass are highly procoagulant. Circulation 1997;96: 3534-41. [101] Gemmell CH. Assessment of material-induced procoagulant activity by a modified Russell viper venom coagulation time test. J Biomed Mater Res 1998;42:611-6. [102] Ware JA, Coller BS. Platelet morphology, biochemistry and function. In: Beutler E, Lichtman MA, Coller BS, Kipps TJ, editors. Williams hematology. New York: McGraw-Hill; 1995. p. 1161-225. [103] Calvete JJ. Clues for understanding the structure and function of a prototypic human integrin: the platelet glycoprotein IIb/IIIa complex. Thromb Haemost 1994;72:1-15. [104] van der Wieken LR. Stents and IIb/IIIa receptor blockers combined: usefulness in various types of coronary artery disease. Semin Interv Cardiol 1999;4:77-83. [105] Ellis SG, Bates ER, Schaible T, Weisman HF, Pitt B, Topol EJ. Prospects for the use of antagonists to the platelet glycoprotein IIb/IIIa receptor to prevent postangioplasty restenosis and thrombosis. J Am Coll Cardiol 1991;17:89B-95B. [106] Mickelson JK, Ali MN, Kleiman NS, et al. Chimeric 7E3 Fab (Reopro) decreases detectable C D l l b on neutrophils from patients undergoing coronory angioplasty. J Am Coll Cardiol 1999;33:97-106. [107] Coller BS, Anderson KM, Weisman HF. The anti-GPIIb/IIIa agents: fundamental and clinical aspects. Haemostasis 1996;26: 285-93. [108] The EPIC Investigators. Use of a monoclonal antibody directed against the platelet glycoprotein IIb/IIIa receptor in high risk coronary angioplasty. N Engl J Med 1994;330:956-61. [109] Topol EJ, Lincoff AM, Kereiakes DJ, et al. Multi-year follow-up of Abciximab therapy in three randomized, placebo controlled trials of percutaneous coronary revascularization. Am J Med 2002;113:1-6. [110] Spinler SA, Hilleman DE, Cheng JWM, et al. New recommendations from the 1999 American College of CArdiology/ American Heart Association acute myocardial infarction guidelines. Ann Pharmacotherapy 2002;35:589-617.
238 5700
The Biomaterials Silver Jubilee Compendium M.B. Gorbet, M.V. Sefton / Biomaterials 25 (2004) 5681-5703
[111] Cote AV, Berger PB, Holmes Jr DR, Scott CG, Bell MR. Hemorrhagic complications after percutaneous coronary interventions with adjunctive Abciximab. Mayo Clin Proc 2002;76: 890-6. [112] Makoni SN. Acute profound thrombocytopenia following angioplasty: the dilemma in the management and a review of the literature. Heart 2002;86:e18-9. [113] Grunkemeier JM, Tsai WB, Horbett TA. Hemocompatibility of treated polystyrene substrates: contact activation, platelet adhesion, and procoagulant activity of adherent platelets. J Biomed Mater Res 1998;41:657-70. [114] Godo MN, Sefton MV. Characterization of transient platelet contacts on a polyvinyl alcohol hydrogel by video microscopy. Biomaterials 1999;20:1117-26. [115] Sheppard JI, McClung WG, Feuerstein IA. Adherent platelet morphology on adsorbed fibrinogen: effects of protein incubation time and albumin addition. J Biomed Mater Res 1994;28:1175-86. [116] Chinn JA, Ratner BD, Horbett TA. Adsorption of baboon fibrinogen and the adhesion of platelets to a thin film polymer deposited by radio-frequency glow discharge of allylamine. Biomaterials 1992;13:322-32. [117] Hanson SR, Harker LA. Blood coagulation and blood-materials interactions. In: Ratner BD, Hoffman AS, Schoen FJ, Lemons JE, editors. Biomaterials science. San Diego: Academic Press; 1996. p. 193-9. [118] Tsai WB, Grunkemeier JM, Horbett TA. Human plasma fibrinogen adsorption and platelet adhesion to polystyrene. J Biomed Mater Res 1999;44:130-9. [119] Bailly AL, Laurent A, Lu H, et al. Fibrinogen binding and platelet retention: relationship with the thrombogenicity of catheters. J Biomed Mater Res 1996;30:101-8. [120] Gemmell CH, Ramirez SM, Yeo EL, Sefton MV. Platelet activation in whole blood by artificial surfaces: identification of platelet-derived microparticles and activated platelet binding to leukocytes as material-induced activation events. J Lab Clin Med 1995;125:276-87. [121] Cholakis CH, Zingg W, Sefton MV. Effect of heparin-PVA hydrogel on platelets in a chronic canine AV shunt. J Biomed Mater Res 1989;23:417-41. [122] Hanson SR, Harker LA, Ratner BD, Hoffman AS. In vivo evaluation of artificial surfaces with a nonhuman primate model of arterial thrombosis. J Lab Clin Med 1980;95: 289-304. [123] Ip WF, Sefton MV. Platelet consumption by NHLBI reference materials and silastic. J Biomed Mater Res 1991;25:1321-4. [124] Hanson SR, Harker LA. Effects of platelet-modifying drugs on arterial thromboembolism in baboons. Aspirin potentiates the antithrombotic actions of dipyridamole and sulfinpyrazone by mechanisms independent of cyclooxygenase inhibition. J Clin Invest 1999;75:1591-9. [125] Weily HS, Steele PP, Davies H, Pappas G, Genton E. Platelet survival inpatients with substitute heart valves. N Eng J Med 1974;290:534-6. [126] Llanos GR, Sefton MV. Immobilization of poly(ethylene glycol) onto poly(vinyl alcohol) hydrogel: 2. Evaluation of thrombogenicity. J Biomed Mater Res 1993;27:1383-91. [127] Wachtfogel YT, Bischoff R, Bauer R, et al. Alpha 1-antitrypsin Pittsburgh (Met358~Arg) inhibits the contact pathway of intrinsic coagulation and alters the release of human neutrophil elastase during simulated extracorporeal circulation. Thromb Haemost 1994;72:843-7. [128] Simon P, Ang KS, Cam G. Enhanced platelet aggregation and membrane biocompatibility: possible influence on thrombosis and embolism in haemodialysis patients. Nephron 1987;45: 172-3.
[129] Hakim RM, Schafer A1. Hemodialysis-associated platelet activation and thrombocytopenia. Am J Med 1985;78:575-80. [130] Peerschke EIB, Ghebrehiwet B. Platelet receptors for the complement component Clq: implication for hemostasis and thrombosis. Immunobiology 1998;199:239-49. [131] Bing DH. Inhibition of guinea pig complement by aromatic amidine and guanidine compounds. J Immunol 1970;105: 1289-91. [132] Andrew JM, Roman Jr DP, Bing DH. Inhibition of four human serine proteases by substituted benzamidines. J Medicinal Chem 1978;21:1202-7. [133] Hauptmann J, Markwardt F. Inhibition of the haemolytic complement activity by derivatives of benzamidine. Biochem Pharmacol 1977;26:325-9. [134] Cox D, Motoyama Y, Seki J, Aoki T, Dohi M, Yoshida K. Pentamidine: a non-peptide GPIIb/IIIa antagonist--in vitro studies on platelets from humans and other species. Thromb Haemost 1992;68:731-6. [135] Cox D, Aoki T, Seki J, Motoyama Y, Yoshida K. Pentamidine is a specific, non-peptide, GPIIb/IIIa antagonist. Thromb Haemost 1996;75:503-9. [136] Allan R, Rodrick M, Knobel HR, Isliker H. Inhibition of the interaction between the complement component C lq and immune complexes. Int Arch Allergy Appl Immunol 1979;58: 140-8.
[137] Subbarao K, Kuchubhotla J, Kakkar VV. Pyridoxal 5'phosphate--a new physiological inhibitor of blood coagulation and platelet function. Biochem Pharmacol 1979;28:531-4. [138] Takada Y, Arimoto Y, Mineda H, Takada A. Inhibition of the classical and alternative pathways by amino acids and their derivatives. Immunology 1978;34:509-15. [139] Rinder CS, Rinder HM, Smith M J, et al. Selective blockade of membrane attack complex formation during simulated extracorporeal circulation inhibits platelet but not leukocyte activation. J Thorac Cardiovasc Surg 1999;118:460-6. [140] Rinder CS, Rinder HM, Smith BR, et al. Blockade of C5a and C5b-9 generation inhibits leukocyte and platelet activation during extracorporeal circulation. J Clin Invest 1995;96: 1564-72. [141] Gorbet MB, Sefton MV. Role of complement and platelets in leukocyte activation induced by polystyrene and PEG-immobilized polystyrene beads in whole blood. J Biomed Mater Res, in preparation. [142] Larsson R, Elgue G, Larsson A, Ekdahl KN, Nilsson UR, Nilsson B. Inhibition of complement activation by soluble recombinant CR1 under conditions resembling those in a cardiopulmonary circuit: reduced up-regulation of CD1 l b and complete abrogation of binding of PMNs to the biomaterial surface. Immunopharmacology 1997;38:119-27. [143] Cassatella MA. The production of cytokines by PMN. Immunol Today 1995;16:21-6. [144] Anderson JM. Mechanisms of inflammation and infection with implanted devices. Cardiovasc Pathol 1993;2:33S-41S. [145] Del Maschio A, Corvazier E, Maillet F, Kazatchkine MD, Maclouf J. Platelet dependent induction and amplification of polymorphonuclear leukocyte lysosomal enzyme release. Br J Haematol 1989;72:329-35. [146] Bazzoni G, Dejana E, Del Maschio A. Platelet-dependent modulation of neutrophil function. Pharmacol Res 1992;26: 269-72. [147] Del Maschio A, Dejana E, Bazzoni G. Bidirectional modulation of platelet and polymorphonuclear leukocyte activities. Ann Hematol 1993;67:23-31. [148] Rebuck N, Finn A. Polymorphonuclear granulocyte expression of CDlla/CD18, CDllb/CD18 and L-selectin in normal individuals. FEMS Immunol Med Microbiol 1994;8:189-96.
239
The Biomaterials Silver Jubilee Compendium M.B. Gorbet, M.V. Sefton / Biomaterials 25 (2004) 5681-5703
[149] Camerer E, Kolsto A-B, Pridz H. Cell biology of tissue factor, the principal initiator of coagulation. Thromb Res 1996;81:1-41. [150] Smith JA. Neutrophils, host defense, and inflammation; a double-edged sword. J Leukocyte Biol 1994;56:686. [151] Kuijipers TW, Hakkert BC, Van Mourik JA, Roos D. Distinct adhesive properties of granulocytes and monocytes to endothelial cells under static and stirred conditions. J Immunol 1990; 145:2588-94. [152] van Ginkel CJW, van Haken WG, Oh JIH, Vreeken J. Stimulation of monocyte procoagulant activity by adherence to different surfaces. Br J Haematol 1977;37:35-45. [153] Gorbet MB, Yeo EL, Sefton MV. Flow cytometric study of in vitro neutrophil activation by biomaterials. J Biomed Mater Res 1999;44:289-97. [154] Ginis I, Tauber AI. Activation mechanisms of adherent human neutrophils. Blood 1990;76:1233-9. [155] Jones DA, Smith CW, McIntire LV. Leukocyte adhesion under flow conditions: principles important in tissue engineering. Biomaterials 1996;17:337-47. [156] Butcher EC. Leukocyte-endothelial cell recognition: three or (more) steps to specificity and diversity. Cell 1991;67:1033-6. [157] Arnaout MA. Structure and function of the leukocyte adhesion molecules CD11/CD18. Blood 1990;75:1037-50. [158] Philips MR, Buyon JP, Winchester R, Weissmann G, Abramson SB. Upregulation of the iC3b receptor is neither necessary nor sufficient to promote neutrophil aggregation. J Clin Invest 1988;82:495-501. [159] Takala AJ, Jousela IT, Takkumen OS, et al. Time course of/~2 integrin CD1 lb/CD18 upregulation on neutrophils and monocytes after coronary artery bypass grafting. Scand J Thorac Cardiovasc Surg 1996;30:141-8. [160] Serrano CV, Ramires JA, Venturinelli M, et al. Coronary angioplasty results in leukocyte and platelet activation with adhesion molecules expression. J Am Coll Cardiol 1997;29: 1276-83. [161] Inque T, Sakai Y, Morooka S, Hayashi T, Takayanagi K, Takabatake Y. Expression of polymorphonuclear adhesion molecules and its clinical significance in patients treated with percutaneous transluminal coronary angioplasty. J Am Coll Cardiol 1996;28:1127-33. [162] Neumann FJ, Ott I, Gawaz M, Puchner G, Sch6mig A. Neutrophil and platelet activation at balloon-injured coronary artery plaque in patients undergoing angioplasty. J Am Coll Cardiol 1996;27:819-24. [163] Rousseau Y, Carreno M-P, Poignet J-L, Kaztchkine MD, Haeffner-Cavaillon N. Dissociation between complement activation, integrin expression and neutropenia during hemodialysis. Biomaterials 1999;20:1959-67. [164] Cristol JP, Canaud B, Rabesandratana H, Gaillard I, Serre A, Mion C. Enhancement of reactive oxygen species production and cell surface markers expression due to hemodialysis. Nephrol Dial Transplant 1994;9:389-94. [165] Von Appen K, Goolsby C, Mehl P, Goewert R, Ivanovich P. Leukocyte adhesion molecule as biocompatibility markers for hemodialysis membranes. ASAIO J 1994;40:M609-15. [166] Rinder C, Fitch J. Amplification of the inflammatory response: adhesion molecules associated with platelet/white cell responses. J Cardiovasc Pharmacol 1996;27(Suppl. 1):$6-12. [167] E1 Habbal MH, Carter H Smith L, Elliot MJ, Strobel S. Neutrophil activation in paediatric extracorporeal circuits: effect of circulation and temperature variation. Cardiovasc Res 1995;29:102-7. [168] Gillinov AM, Bator JM, Zehr KJ, et al. Neutrophil adhesion molecule expression during cardiopulmonary bypass with bubble and membrane oxygenators. Ann Thorac Surg 1993;56: 847-53.
5701
[169] Cameron DE. Initiation of white cell activation during cardiopulmonary bypass: cytokines and receptors. J Cardiovasc Pharmacol 1996;27:S1-5. [170] Asimakopoulos G, Taylor KM. Effects of cardiopulmonary bypass on leukocyte and endothelial adhesion molecules. Ann Thorac Surg 1998;66:2135-44. [171] Horl WH, Riegel W, Steinhauer HB, et al. Granulocyte activation during hemodialysis. Clin Nephrol 1986;26: $30-4. [172] Grooteman MP, van TA, van HA, et al. Hemodialysis-induced degranulation of polymorphonuclear cells: no correlation between membrane markers and degranulation products. Nephron 2000;85:267-74. [173] Haag-Weber M, Schollmeyer P, Horl WH. Beta-2-microglobulin and main granulocyte components in hemodialysis patients. Artif Organs 1989; 13:92-6. [174] Stegmayr BG, Esbensen K, Gutierrez A, et al. Granulocyte elastase, /~ thromboglobulin, and C3d during acetate or bicarbonate hemodialysis with hemophan compared to a cellulose acetate membrane. Int J Artif Organs 1992;15:10-8. [175] Wachtfogel YT, Kucich U, Greenplate J, et al. Human neutrophil degranulation during extracorporeal circulation. Blood 1987;69:324-30. [176] Peek GJ, Firmin RK. The inflammatory and coagulative response to prolonged extracorporeal membrane oxygenation. ASAIO J 1999;45:250-63. [177] Frering B, Philip I, Dehoux M, Rolland C, Langlois JM, Desmont JM. Circulating cytokines in patients undergoing normothermic cardiopulmonary bypass. J Thorac Cardiovasc Surg 1994;108:636-41. [178] Himmelfarb J, Lazarus JM, Hakim R. Reactive oxygen species production by monocytes and polymorphonuclear leukocytes during dialysis. Am J Kidney Dis 1991;XVII:271-6. [179] Kaplan SS, Basford RE. Mechanisms of biomaterial induced superoxide release by neutrophils. J Biomed Mater Res 2000;28: 377-86. [180] Kuwahara T, Markert M, Wauters JP. Proteins adsorbed on hemodialysis membranes modulate neutrophil activation. Artif Organs 1989;13:427-31. [181] Serruys PW, Strauss BH, van Beusekom HM, van der Giessen WJ. Stenting of coronary arteries: has a modern Pandora's box been opened. J Am Coll Cardiol 1991;17:143B-54B. [182] van Beusekom HMM, van der Giessen WJ, Van Suylen RJ, Bos E, Bosman FT, Serruys PW. Histology after stenting of human saphenous vein bypass grafts: observations from surgically excised grafts 3-320 days after stent implantation. J Am Coll Cardiol 1993;21:45-54. [183] Jobes DR. Safety issues in heparin and protamine administration for extracorporeal circulation. J Cardiothorac Vasc Anesth 1998;12:17-20. [184] Grooteman MPC, Bos JC, Van Houte AJ, van Limbeek J, Schoorl M, Nube MJ. Mechanisms of intra-dialyser granulocyte activation: a sequential dialyser elution study. Nephrol Dial Transplant 1997;12:492-9. [185] Castiglione A, Pagliaro P, Romagnoni M, Beccari M, Cavaliere G, Veneroni G. Flow cytometric analysis of leukocytes eluted from haemodialysers. Nephrol Dial Transplant 1991;(Suppl. 2):31-35. [186] Gemmell CH, Black JP, Yeo EL, Sefton MV. Material-induced up-regulation of leukocyte CD1 l b during whole blood contact: material differences and a role for complement. J Biomed Mater Res 1996;32:29-35. [187] Gorbet MB, Sefton MV. Expression of procoagulant activities on Leukocytes following contact with polystyrene and PEG grafted polystyrene beads. J Lab Clin Med 2001;137: 345-55.
240 5702
The Biomaterials Silver Jubilee Compendium M.B. Gorbet, M.V. Sefton / Biomaterials 25 (2004) 5681-5703
[188] Kao WJ, Sapatnekar S, Hiltner A, Anderson JM. Complementmediated leukocyte adhesion on poly(etherurethane ureas) under shear stress in vitro. J Biomed Mater Res 1996;32:99-109. [189] McNally AK, Anderson JM. Complement C3 participation in monocyte adhesion to different surfaces. Proc Natl Acad Sci USA 1994;91:10119-23. [190] Cheung AK, Hohnholt M, Gilson J. Adherence of neutrophils to haemodialysis membranes: roles of complement receptors. Kidney Int 1991;40:1123-33. [191] Tang L, Lucas AH, Eaton JW. Inflammatory responses to implanted polymeric biomaterials: role of surface adsorbed immunoglobulin G. J Lab Clin Med 1993;122:292-300. [192] Tang L, Eaton JW. Fibrinogen mediates acute inflammatory responses to biomaterials. J Exp Med 1993;178:2147-56. [193] Eriksson O, Nygren H. Polymorphonuclear leukocytes in coagulating whole blood recognize hydrophilic and hydrophobic titanium surfaces by different adhesion receptors and show different patterns of receptor expression. J Lab Clin Med 2001; 137:296-302. [194] Gemmell CH. Flow cytometric evaluation of material-induced platelet and complement activation. J Biomater Sci Polym Ed 2000;11(11):1197-210. [195] Morley DJ, Feuerstein IA. Adhesion of polymorphonuclear leukocytes to protein-coated and platelet adherent surfaces. Thromb Haemost 1989;62:1023-8. [196] Herzlinger GA, Cumming RD. Role of complement activation in cell adhesion to polymer blood contact surfaces. ASAIO Trans 1980;XXXVI: 165-71. [197] Wachtfogel YT, Kettner C, Hack CE, et al. Thrombin and human plasma kallikrein inhibition during simulated extracorporeal circulation block platelet and neutrophil activation. Thromb Haemost 1998;80:686-91. [198] Wachtfogel YT, Kucich U, Hack CE, et al. Aprotinin inhibits the contact, neutrophil, and platelet activation systems during simulated extracorporeal perfusion. J Thorac Cardiovasc Surg 1993;106:1-9. [199] Himmelfarb J, Holbrook D, McMonagle E. Effects of aprotinin on complement and granulocyte activation during ex vivo hemodialysis. Am J Kidney Dis 1994;24:901-6. [200] Himmelfarb J, McMonagle E, Holbrook D, Toth C. Soluble complement receptor 1 inhibits both complement and granulocyte activation during ex vivo hemodialysis. J Lab Clin Med 1995;126:392-400. [201] Gillinov AM, Redmond JM, Winkelstein JA, et al. Complement and neutrophil activation during cardiopulmonary bypass: a study in the complement-deficient dog. Ann Thorac Surg 1994;57:345-52. [202] Finn A, Morgan BP, Rebuck N, et al. Effects of inhibition of complement activation using recombinant soluble receptor 1 on neutrophil CD 11b/CD 18 and L-selectin expression and release of interleukin-8 and elastase in simulated cardiopulmonary bypass. J Thorac Cardiovasc Surg 1996;111:451-9. [203] Rinder CS, Rinder HM, Johnson K, et al. Role of C3 cleavage in monocyte activation during extracorporeal circulation. Circulation 1999;100:553-8. [204] Gorbet MB, Sefton MV. The importance of platelets and complement in material-induced Leukocyte activation in vitro. PhD Thesis, University of Toronton, 2001. [205] Kappelmayer J, Bernabei A, Edmunds Jr LH, Edgington TS, Colman RW. Tissue factor is expressed on monocytes during simulated extracorporeal circulation. Circ Res 1993;72: 1075-81. [206] Barstad RM, Hamers MJAG, Moller A-S, Sakariassen KS. Monocyte procoagulant activity induced by adherence to an artificial surface is reduced by end-point immobilized heparin coating of the surface. Thromb Haemost 1998;79:302-5.
[207] Stahl RF, Fisher CA, Kucich U, et al. Effects of simulated extracorporeal circulation on human leukocyte elastase release, superoxide generation and procoagulant activity. J Thorac Cardiovasc Surg 1991;101:230-9. [208] Parratt R, Hunt BJ. Direct activation of factor X by monocytes occurs during cardiopulmonary bypass. Br J Haematol 1998; 101:40-6. [209] Rinder CS, Bonan JL, Rinder HM, Mathew J, Hines R, Smith BR. Cardiopulmonary bypass induces leukocyte-platelet adhesion. Blood 1992;79:1201-5. [210] Gawaz M, Bogner C. Changes in platelet membrane glycoproteins and platelet leukocyte interactions during hemodialysis. J Clin Invest 1994;72:424-9. [211] May AE, Neumann F-J, Gavaz M, Ott I, Walter H, Sch6mig A. Reduction of monocyte-platelet interaction and monocyte activation in patients receiving antiplatelet therapy after coronary stent implantation. Eur Heart J 1997;18:1913-20. [212] Palabrica T, Lobb R, Furie BC, et al. Leukocyte accumulation promoting fibrin deposition is mediated in vivo by P-selectin on adherent platelets. Nature 1992;359:848-51. [213] Toombs CF, De Graaf GL, Martin JP, Geng JG, Anderson DC, Shebuski RJ. Pretreatment with a blocking antibody to P-sel accelerates pharmacological thrombolysis in a primate model of arterial thrombosis. J Pharmacol Exp Ther 1995;275:941. [214] Turitto VT, Weiss HJ, Baumgartner HR. The effect of shear rate on platelet interaction with subendothelium exposed to citrated human blood. Microvasc Res 1980;19:352-65. [215] Turitto VT, Hall CL. Mechanical factors affecting hemostasis and thrombosis. Thromb Res 1998;92:$25-31. [216] Hanson SR, Sakariassen KS. Blood flow and antithrombotic drug effects. Am Heart J 1998;135:S132-45. [217] O'Brien JR. Shear induced platelet aggregation. Lancet 1990;335:711-3. [218] Oda A, Yokoyama K, Murata M, et al. Protein tyrosine phosphorylation in human platelets during shear-stress induced aggregation (SIPA) is regulated by GPIb/IX as well as GPIIb/ IIIa and requires intact cytoskeleton and endogenous ADP. Thromb Haemost 1995;74:736-42. [219] Kujiper PHM, Gallardo Torres HI, van der Linden JAM, et al. Platelet-dependent primary hemostasis promotes selectin- and integrin-mediated neutrophil adhesion to damaged endothelium under flow conditions. Blood 1996;87:3271-81. [220] Yeo EL, Sheppard J-OI, Feuerstein IA. Role of P-selectin and leukocyte activation in polymorphonuclear cell adhesion to surface adherent activated platelets under physiological shear conditions (an injury vessel wall model). Blood 1994;83: 2498-507. [221] Weber C, Springer TA. Neutrophil accumulation on activated, surface-adherent platelets in flow is mediated by interaction of MAc-1 with fibrinogen bound to ~IIb/~3 and stimulated by platelet activating factor. J Clin Invest 1997;100: 2085-93. [222] Kujiper PHM, Gallardo Torres HI, Lammers J-WJ, Sixma JJ, Koenderman L, Zwaginga JJ. Platelet and fibrin deposition at the damaged vessel wall: cooperative substrates for neutrophil adhesion under flow conditions. Blood 1997;89:166-75. [223] Bruil A, Sheppard JI, Feijen J, Feuerstein IA. In vitro leukocyte adhesion to modified polyurethane surfaces: III Effects of flow, fluid medium and platelets on PMN adhesion. In: Cooper SL, Bamford CH, Tsuruta T, editors. Polymer biomaterials in solution, as interfaces and as solid. Zeist, The Netherlands: VSP; 1995. p. 357-71. [224] Kujiper PHM, Gallardo Torres HI, van der Linden JAM, et al. Neutrophil adhesion to fibrinogen and fibrin under flow conditions is diminished by activation and L-selectin shedding. Blood 1997;89:2131-8.
241
The Biomaterials Silver Jubilee Compendium M.B. Gorbet, M.V. Sefton / Biomaterials 25 (2004) 5681-5703 [225] Hernandez MR, Escolar G, Bozzo J, Galan AM, Ordinas A. Inhibition of fibrin deposition on the subendothelium by a monoclonal antibody to polymorphonuclear leukocyte integrin CDllb. Studies in a flow system. Haematologica 1997;82: 566-71. [226] Bahra P, Nash GB. Sparsely adherent platelets support capture and immobilization of flowing neutrophils. J Lab Clin Med 1998;132:223-8. [227] Gemmell CH, Nemerson Y, Turitto VT. The effects of shear rate on the enzymatic activity of the tissue factor-factor VIIa complex. Microvasc Res 1990;40:327-40. [228] Nemerson Y, Contino PB. Tissue factor, flow and the initiation of coagulation. In: Seghatchian MJ, Samana MM, Hecker SP, editors. Hypercoagulable states. Boca Raton, FL: CRC Press; 1996. p. 21-8. [229] Bluestein D, Rambod E, Gharib M. Vortex shedding as a mechanism for free emboli formation in mechanical heart valves. J Biomech Eng 2000; 122:125-34. [230] Melder RJ, Munn LL, Yamada S, Ohkubo C, Jain RK. Selectin and integrin-mediated T lymphocyte rolling and arrest on TNFc~activated endothelium: augmentation by erythrocytes. Biophys J 1995;69:2131-8. [231] Lynn WA, Golenbock DT. Lipopolysaccharide antagonists. Immunol Today 1992;13:271-6. [232] Lemke HD. Methods for the detection of endotoxins present during extracorporeal circulation. Nephrol Dial Transplant 1994;9:90-5. [233] US Department of Health and Human Services/Public Health Services/Food and Drug Administration. Guideline on validation of the limulus amebocyte lysate test as an end-product endotoxin test for human and animal parental drugs, biological products and medical devices. 1987. p. 1-30. Available at http:// www.fda.gov/cber/gdlns/lal.pdf [234] Hirthler M, Simoni J, Dickson M. Elevated levels of endotoxin, oxygen-derived free radicals and cytokines during extracorporeal membrane oxygenation. J Pediatr Surg 1992;27:1199-202. [235] Nilsson L, Kulander L, Nystr6m S-O, Eriksson O. Endotoxin in cardiopulmonary bypass. J Thorac Cardiovasc Surg 1990; 100:777-80. [236] Butler J, Rocker GM, Westaby S. Inflammatory response to cardiopulmonary bypass. Ann Thorac Surg 1993;55:552-9. [237] Khabar KS, elBarbary MA, Khouqeer F, Devol E, al-Gain S, alHalees Z. Circulating endotoxin and cytokines after cardiopulmonary bypass: differential correlation with duration of bypass and systemic inflammatory response/multiple organ dysfunction syndrome. Immunol Immunopathol 1997;85:97-103. [238] Rocke DA, Gaffin SL, Wells MT, Koen Y, Brock Utine JG. Endotoxemia associated with cardiopulmonary bypass. J Thorac Cardiovasc Surg 1987;93:832-7. [239] Laude-Sharp M, Caroff M, Simard L, Pusineri C, Kazatchkine MD, Haeffner-Cavaillon N. Induction of IL-1 during hemodialysis: transmembrane passage of intact endotoxin (LPS). Kidney Int 1990;38:1089-94. [240] Khazazmi A, Andersen LW, Baek L, Valerius NH, Laub M, Rasmussen JP. Endotoxemia and enhanced generation of oxygen radicals by neutrophils from patients undergoing cardiopulmonary bypass. J Thorac Cardiovasc Surg 1989;98:381-5.
5703
[241] Casko G, Suba EA, Elin RJ. Endotoxin-induced platelet activation in human whole blood in vitro. Thromb Haemost 1988;59:375-82. [242] Weingarten R, Sklar LA, Mathison JC, et al. Interactions of lipopolysaccharide with neutrophils in blood via CD14. J Leukocyte Biol 1993;53:518-24. [243] Wright SD, Ramos RA, Hermanowski-Vosatka A, Rockwell P, Detmers PA. Activation of the adhesive capacity of CR3 on neutrophils by endotoxin: dependence on lipopolysaccharide binding protein and CD14. J Exp Med 1991;173:1281-6. [244] Pearson FC. A comparison of the pyrogenicity of environmental endotoxins and lipopolysaccharides. In: Ten Cate JW, Bfiller HR, Sturk A, Levin J, editors. Bacterial endotoxins: structure, biomedical significance, and detection with the Limulus Amebocyte Lysate test. New York: Alan Riss, Inc; 1985. p. 251-63. [245] Enders G, Brooks W, Von Jan N, Lehn N, Bayerd6rfer E, Hatz R. Expression of adhesion molecules on human granulocytes after stimulation with Helicobacter pylori membrane proteins: comparison with membrane proteins from other bacteria. Infect Immun 1995;63:2473-7. [246] Tielemans C, Husson C, Schurmans T, et al. Effects of ultrapure and non-sterile dialysate on the inflammatory response during in vitro hemodialysis. Kidney Int 1996;49:236-43. [247] Ragab AA, Van de Motter R, Lavish SA, et al. Measurement and removal of adherent endotoxin from Titanium particles and implant surfaces. J Orthop Res 1999;17:803-9. [248] Hitchins VM, Merritt K. Decontaminating particles exposed to bacterial endotoxin (LPS). J Biomed Mater Res 1999;46:434-7. [249] Gillis S, Furie BC, Furie B. Interactions of neutrophils and coagulation proteins. Semin Hematol 1997;34:336-42. [250] Altieri DC. Inflammatory cell participation in coagulation. Semin Cell Biol 1995;6:269-74. [251] Esmon CT. Introduction: cellular regulation of blood coagulation. Semin Cell Biol 1995;6:257-8. [252] May AE, Neumann FJ, Preissner KT. The relevance of bloodcell wall adhesive interactions for vascular thrombotic disease. Thromb Haemost 1999;82:962-70. [253] Marcus AJ. Thrombosis and inflammation as multicellular processes: significance of cell-cell interactions. Semin Hematol 2000;31:261-9. [254] Sengelov H. Complement receptors in neutrophils. Crit Rev Immunol 1995;15:107-31. [255] Pabst MJ. Priming of neutrophils. In: Hellewell PG, Williams TJ, editors. Immunopharmacology of the neutrophil. London: Academic Press; 1994. p. 195-222. [256] Sims PJ. Interaction of platelets with complement system. In: Kunicki TJ, Georges JN, editors. Platelet immunobiology: molecular and clinical aspects. Philadelphia: Lippincott; 1989. p. 354-83. [257] Kirschfink M. Controlling the complement system in inflammation. Immunopharmacology 1997;38:51-62. [258] Hitchins VM, Merritt K. Decontaminating particles exposed to bacterial endotoxin (LPS). J Biomed Mater Res 1999;46:434-7. [259] Ragab AA, Van de Motter R, Lavish SA, Goldberg VM, Ninomiya JT, Carlin CR, Greenfield EM. Measurement and removal of adherent endotoxin from titanium particles and implant surfaces. J Orthop Res 1999;17:803-9.
242
This Page Intentionally Left Blank
The Biomaterials Silver Jubilee Compendium
243
The Biomaterials Silver Jubilee Compendium
Author Index Author Adell, R. Aebischer, P. Albrektsson, T. Anderson, J.M. Bergsma, J.E. Black, J. Blanchard, C.R. Boering, G. Bos, R.R. Boyan, B.D. Branemark, P.I. Brannon-Peppas, L. Breuer, C. Bruijn de, W.C. Carreno, M.P. Cochran, D.L. Davies, J.E. Gopferich, A. Gorbet, M.B. Heller, J. Hem, D. Hollinger, J.O. Hubbell, J.A. Hunt, J.A. Hutmacher, D.W. Ingber, D.E. Ishihara, K. Kane, R.S. Kazatchkine, M.D. Kikuchi, A. Kojima, M. Langer, R. Leite, S.M. Lekholm, U. Leong, K. Lincks, J. Liu, Y. Lohmann, C.H. Lundkvist, S. Matsuda, T. Mazzoni, C.L. McNamara, A. McNamara, K. Mikos, A.G. Miller, K.M. Mooney, D.J.
Page No. 17 61 17 21 101 27 147 101 101 147 17 51 129 101 45 147 35 117 219 1 129 139 203 73 175 161 69 161 45 211 69 93 93 17 139 147 147 147 17 35 129 9 129 93 21 129
244
The Biomaterials Silver Jubilee Compendium
Nakabayashi, N. Okano, T. Okuhara, M. Ostuni, E. Park, Y.D. Peppas, N.A. Remes, A. Rockler, B. Rozema, F.R. Sakai, H. Sakurai, Y. Saltzman, W.M. Sarakinos, G. Sefton,M.V. Shimizu, T. Takayama, S. Tan, J. Tirelli, N. Tresco, P.A. Vacanti, J.P. Vince, D.G. Wahlberg, L. Watanabe, A. Williams, D. F. Winn, S.R Whitesides, G.M. Yamada, N. Yamato, M.
69 109, 211 109 161 203 51 79 17 101 109 109 191 93 219 211 161 191 203 61 93, 129 73 61 69 v, 9, 73, 79 61 161 109 211
The Biomaterials Silver Jubilee Compendium
This Page Intentionally Left Blank
245
246
This Page Intentionally Left Blank
The Biomaterials Silver Jubilee Compendium
