The Biomaterials: Silver Jubilee Compendium

Citation preview

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

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 (