Graefe's Arch Clin Exp Ophthalmol (1996) 234:193-198 © Springer-Verlag 1996

Traian V. Chirila Dawn E. Thompson-Wallis Geoffrey J. Crawford Ian J. Constable Sarojini Vijayasekaran

Production of neocollagen by cells invading hydrogel sponges implanted in the rabbit cornea

Received: 22 November 1994 Revised version received: 9 May 1995 Accepted: 29 May 1995

T.V. Chirila (~) • D.E. Thompson-Wallis G.J. Crawford • I.J. Constable S. Vijayasekaran Lions Eye Institute, 2 Verdun Street, Block A, 2nd Floor, Nedlands, Western Australia 6009, Australia

Abstract • Background: Poly(2- hydroxyethyl methacrylate) sponges are artificial tissue-equiva- lent matrices with potential value as materials for the peripheral zone of artificial corneas. A keratoprosthet- ic device was developed incorporat- ing a poly(HEMA) spongy skirt which allowed cellular invasion. The present in vivo study investigat- ed the biosynthetic activity of stro- real fibroblasts growing within a poly(HEMA) sponge implanted into the rabbit cornea. • Methods: A porous poly(HEMA) hydrogel was synthesized by polymerization in a large excess of water. Specimens with a pore size larger than 10 gm were impregnated with collagen type I and then implanted into the limbal region of cornea in four rab- bits. The animals were followed clinically for 28 days, when they were anaesthetized and new sponge specimens were implanted in their

second eye. After 2 h, both eyes were enucleated. The 28-day and 2- h explants were subjected to autora- diographic analysis following la- belling with tritiated proline and to an immunostaining technique using antibodies to collagen types I- VI. • Results: The autoradiograph- ic analysis showed that the fibrob- lasts within the 28-day explants continued to be synthetically active and deposited proteins. Using the immunostaining technique, the de- position was most clearly demon- strated by the localization of colla- gen type III in the tissue invading the sponge. Both techniques failed to indicate any cellular activity in the short-time implants. • Conclu- sions: The presence of collagen type III is consistent with a normal healing response of the stromal fi- broblasts and indicates that poly(- HEMA) sponges are able to func- tion as tissue-equivalent matrices.

Introduction

The introduction of the so-called "core-and-skirt" kera- toprostheses made from synthetic polymers was proba- bly the most significant advance in the long history of artificial cornea. The past decade further witnessed the development of prosthetic devices in which the skirts were made from porous polymers [5]. It was expected that a porous rim would allow the invasion and prolifera- tion of stromal fibroblasts and consequently a permanent and tight union between prosthesis and host cornea

would be achieved. This process may prevent the extru- sion of keratoprosthesis, and some encouraging results have indeed been reported in human patients [18, 24, 25, 361.

A major problem with the core-and-skirt keratopros- theses designed so far lay in the difficulty of creating a reliable attachment of the peripheral porous polymer to the central transparent polymer, since they are quite dif- ferent in their chemical nature and physical properties. To overcome this, we proposed and developed [8] a kera- toprosthesis in which both parts were made from crosslinked polymers of 2-hydroxyethyl methacrylate

194

( H E M A ) and were therefore chemica l ly identical. A long the boundary between the core and the skirt a permanent joint was achieved by creat ing an in terpenet ra t ing poly- mer network, as demons t ra ted by e lectron mic roscop ic techniques [9]. The skirt was made f rom p o l y ( H E M A ) sponges p roduced by phase-separa t ion po lymer iza t ion in aqueous solution [4, 6]. It was shown, both in vi tro and in vivo (animals) , that the sponges synthesized in more than 75% water, displaying pores larger than 10 gm, were b iocolonized th rough cel lular invasion both subcu- taneously [7] and in t racorneal ly [14]. However, so far we have not provided p roof that the stromal f ibroblasts are synthetical ly active after they have co lonized the sponge.

The synthesis o f neocol lagen and other connect ive tis- sue proteins is an essential funct ion o f f ibroblasts dur ing the process o f wound heal ing and organizat ion o f new connect ive tissue [17, 33, 37, 42]. Col lagen is involved in each phase o f the heal ing process , be ing the major re- pa i r -p romot ing agent. Tissue regenerat ion, t issue re- model l ing and organ morphogenes i s are all dependent on the deposi t ion o f collagen. Whi ls t the biosynthet ic func- tion o f f ibroblasts in monolayer cultures on various solid substrates has been investigated to some extent [28, 33, 35, 41], the studies o f this process when the f ibroblasts were embedded in three-d imensional ar t i f icial mat r ices have been mainly prompted by the tentative use of vari- ous porous materials as prosthet ic devices [16, 23, 26, 43 -45 ] . F ibroplas ia and neocol lagen deposi t ion was demons t ra ted by histological and /o r h is tochemical tech- niques in porous poly te t raf luoroethylene [23] and in po lybu ty l ene /po lyp ropy lene f ibrous webs [43-45] , both p roposed as per ipheral componen t s in core-and-sk i r t keratoprostheses .

The present study was des igned to prove the biosyn- thetic act ivi ty in vivo of s t romal f ibroblasts g rowing into p o l y ( H E M A ) hydrophi l ic sponges implanted in the l im- bal region of rabbit corneas. Col lagen type I was incor- porated into sponges prior to implantat ion. This is a con- venient method to induce cel lular invasion in H E M A - based polymers , both homogeneous [2, 11, 13, 19, 38, 39] and he terogeneous [7, 14], o therwise well known as mater ia ls which do not suppor t adhesion, spreading and prol iferat ion o f the cells. For the sake o f a convenient ly rapid p re l iminary evaluation, only two time points were cons idered for the examinat ion o f tissues,i.e., 2 h and 28 days postoperatively. The identi ty o f col lagen types was based entirely on immunosta in ing .

Materials and methods

Sterile, stable collagen type I was isolated from tendons of Wistar rat tails and solutions (3.7 mg/ml) were prepared in aqueous acetic acid (0,04 M) using previously described~methods [3]. The prove- nience of other agents is indicated whenever they are first men- tioned.

Synthesis of polymer sponge

A porous hydrogel was synthesized by polymerizing HEMA in the presence of 80% water in the monomer mixture as described in our previous papers [7, 14]. A cutting device consisting of two scalpel blades separated by a spacer was used to cut 1-mm thick sheets from a button of sponge. Discs (4 mm in diameter) were then trephined from the sheets, washed extensively in deionized water, and sterilized in an autoclave at 130 ° C for 20 min.

Prior to their implantation, the sponge discs were impregnated with collagen type I. They were first stirred in a sterile, acidic (pH 2.9) solution of collagen (3.7 mg/ml) for 5 h at 4 ° C. The pH was then adjusted to neutral and the discs were incubated for 30 min at 37 ° C without stirring, to allow the crosslinking of colla- gen. The collagen-impregnated discs were transferred to sealed containers with phosphate-buffered saline (PBS) and stored at 4 ° C until implantation.

Sponge implantation

Four Dutch belted rabbits were used in these experiments, which were conducted in accordance with The Australian Code of Prac- tice for the Care and Use of Animals for Scientific Purposes (1990).

The animals were anaesthetized with halothane. The right eye of each rabbit was proptosed and two 4-ram, half-depth incisions were made at diametrically opposed positions in the clear cornea adjacent to the limbus. An ophthalmic crescent knife was inserted into each wound and a pocket was dissected. One sponge disc was inserted in each pocket, and the wound was closed with 10-0 nylon sutures. Soframycin eye drops (Roussel Pharmaceutical, Aus- tralia) were applied to the operated eyes. We chose the limbal region for the surgical insertion of the sponge specimens because in a real situation (i.e. when a full keratoprosthesis is implanted into the cornea) the spongy skirt will actually reside in this region.

During the next 4 weeks, the eyes were checked periodically. All eyes were quiet, with only minimum erythema around the site of implant. The anterior chambers showed no cellular activity. At no stage was there any evidence of stromal melting or implant extrusion.

At 28 days after implantation, the same rabbits were anaes- thetized again and sponge discs were inserted into the left eye of each animal, using the same surgical procedure. The rabbits were kept under anaesthesia for 2 h and then killed by intravenous ad- ministration of sodium pentobarbital. Both eyes of each rabbit were immediately enucleated and placed into sterile PBS.

Autoradiography

Autoradiography was employed to ascertain the biosynthesis of proteins. The labelling of newly synthesized collagenous proteins by an isotope is indicated by the presence of autoradiographic granules. Tritiated proline (L-[2,3-3H]proline; 31 Ci/mmol, Amer- sham Australia), which labels collagens, was the precursor of choice for this study.

Pieces of 2-h and 28-day sponges explanted from two rabbits were incubated with L-[2,3-3H]proline (at 20 gCi/ml in proline- free Dulbecco's modification of Eagle's medium, supplied by Flow Laboratories, Scotland) in 5% CO2/air at 37 ° C for 20 h. The pieces were then washed six times in PBS over 6 h, to remove the proline that was not incorporated, and fixed overnight in 2.5% glutaraldehyde in 0.1 M phosphate buffer. The pieces were post- fixed in osmium tetroxide, dehydrated in graded ethanols, and embedded in epoxy resin (Durcupan ACM, supplied by Fluka, Switzerland). Semithin (2 gm) sections were cut using an ultrami- crotome. The sections were dipped in K2 emulsion (Ilford, UK) diluted (1:3) with deionized water, air-dried and finally incubated

195

in a light-proof box with dessicant at 4 ° C for 3 weeks. The sec- tions were counterstained with toluidine blue and then examined and photographed in a Olympus BH-2 light microscope. Negative controls were also prepared by subjecting to the same autoradio- graphic procedure pieces of 28-day implants that had not been exposed to tritiated proline.

The method used by Trinkaus-Randall et al. [44], where the quantitative assessment of protein synthesis was made by incubat- ing the explanted polybutylene/polypropylene discs with tritiated proline and then performing scintillation counting, was found un- suitable for our study. It was impossible to remove completely, after explanation, the scar tissue adhered to the sponge discs, and thus the results of scintillation counting would probably have been influenced by the remaining tissue.

Immunostaining

Since proline can be incorporated into noncollagenous proteins too, affinity-purified goat antibodies to human/bovine collagen types I-VI (supplied with ELISA-confirmed specificity by South- ern Biotechnology Associates, Birmingham, Ala., USA), were used to detect the newly produced collagen within the explanted sponges, as well as to determine its type.

Pieces of 2-h and 28-day sponges explanted from the other two rabbits, with adherent scar tissue and scleral tissue, were fixed in 4% paraformaldehyde in 0.1 M phosphate buffer at room tempera- ture for 30 min. They were then dehydrated in graded methanols

and embedded in LR Gold resin (London Resin, UK). Sections (4 ~tm thick) were cut and floated in PBS.

The sections were treated successively with: (a) 3% hydrogen peroxide to quench endogenous peroxidases, in methanol, for 10 rain; (b) 10% normal rabbit serum (Dako, Glostrup, Denmark) in PBS containing 0.1% bovine serum albumin (PBS/BSA), for 30 rain; (c) 2.5 mg/ml IgG (for the negative controls), or the specific primary antibody, in PBS/BSA, for 2 h; (d) biotinylated secondary antibody diluted (1: 200) in PBS/BSA, for 1 h; (e) Vec- tastain Elite ABC reagent, for 30 rain; and (f) 0.05% 3,3'-di- aminobenzidine and 0.015% hydrogen peroxide in TRIS buffer (pH 7.6), for 10 rain. (The biotinylated rabbit anti-goat IgG and Vectastain Elite ABC immunoperoxidase kit were supplied by Vector Laboratories, Burlinghame, Calif., USA). The sections were washed thoroughly with PBS or TRIS buffer following each of the above stages.

The treated and washed sections were then counterstained with methyl green, DEPEX-mounted onto glass slides, and pho- tographed through the light microscope. Positive antibody la- belling was observed as brown staining of the sections.

Results

The autoradiographic pho tomic rographs of the negative controls (Fig. l a ) and of the sponges explanted after 2 h (Fig. l b ) ind ica ted the absence of au toradiographic gran-

Fig. l a -d Representative au- toradiograms of explanted poly(HEMA) sponges. a Sponge explanted after 28 days, not labeled. Fibroblasts are indicated by open arrows. Bar 10 gin. b Sponge explant- ed after 2 h, labeled in vitro with tritiated proline for 20 h. No cells can be seen. Polymer particles are indicated by as- terisks. Bar 12 gin. e Peripher- al pigmentation in the sponge explanted after 2 h and labeled in conditions described. Bar 12 gin. d Autoradiographic granules (arrows) in the sponge explanted after 28 days and labeled in conditions de- scribed. Fibroblasts are indi- cated by open arrows. Bar 10 ~tm

196

Fig. 2a, b Immunostaining patterns in poly(HEMA) sponges explanted after 28 days. Bars 15 gin. a Negative control sections (primary anti- body omitted), b Positive staining for collagen type III

ules. Some peripheral pigmentation was noted (Fig. lc), not of a radiographic nature. In contrast, the sponges explanted after 28 days showed consistently autoradio- graphic granules throughout the matrix (Fig. ld).

Immunostaining photomicrographs of sponges ex- planted after 2 h showed no positive staining for colla- gens types I-VI. The IgG negative control showed a non- specific background (Fig. 2a). The sponges explanted after 28 days stained positive only for collagen type III (Fig. 2b).

Several sponge discs impregnated with collagen type I were immediately fixed and processed for collagen type I immunostaining, without being implanted. As ex- pected, all these samples stained positive for collagen type I.

Discussion

Autoradiographic analysis of long-term in vitro labeling with tritiated proline or sponge explants indicated specific incorporation in the 28-day explants, but not in the negative controls (sponges explanted after 28 days but not exposed to tritiated proline) or in the 2-h ex- plants. This clearly proves the active biosynthesis of proteins by cells within the sponges implanted into the cornea for 4 weeks. The peripheral pigmentation ob- served in the controls and in the short-term implants is probably due to the apposition of the implants to the iris, with the consequent adherence of pigmented cells to the implants.

Immunohistochemical studies were carried out with antibodies against collagen types I-VI. We confirmed prior to their use the species cross-reactivity of these antibodies by observation of positive staining in rabbit femoral artery and scleral scar tissue. The results of im- munohistochemical analysis were somewhat surprising. Collagen type I, added on purpose to the sponges, was

detected in the specimens before implantation; however, after only 2 h in the rabbit cornea the immunostaining method indicated the absence of any trace of collagen type I. This could be due to a rapid degradation of colla- gen by collagenolytic enzymes. Although the degenera- tion rate may seem too fast, it is hard to imagine an alter- native decay mechanism.

Neocollagen was detected only in the sponges ex- planted after 28 days, and it was exclusively collagen type III. In the light of the controversy surrounding the presence of collagen type III in the mammalian corneal stroma, our finding warrants some comments. Not very long ago the existence of collagen type III in the stroma was seriously doubted, as best illustrated in the words of one recognized expert [17]: "In my laboratory we were unable to confirm the presence of type III collagen in any significant quantity in rabbit, bovine or human tissue, although we do not exclude the possibility of the pres- ence of tiny amounts of this molecular species at levels detectable only by type-specific antibodies." Soon after, however, its presence was demonstrated in the corneas of humans [32], bovines [22] and rabbits [12]. Some inves- tigators reported the presence of collagen type III in the human cornea only in early fetal life [1]; after 27 weeks of gestation, type III could not be detected except in the vascular limbus. Similarly, the presence of collagen type III in the rabbit cornea was detected after birth, but not in mature animals [21]. Such findings led to the tentative generalization that collagen type III may represent an embryonic collagen which disappears after birth [20]. However, other studies reported the presence of collagen type III throughout the stroma in aged human cornea [29, 30]. It appears that collagen type III is present only in the human corneal stroma after maturity and not as a result of trauma or disorders [30].

In in vitro culture studies, bovine stromal fibroblasts produced exclusively collagen type I [10], but organ cell cultures of normal, keratoconus, and healed grafted

197

(penetrating keratoplasty) human corneal specimens synthesized collagen type III in addition to collagen type I [31]; the amount of type I I I was larger in the scarred regions of the keratoconus corneas and in the healed graf t /host junction of operated corneas.

The latter observation suggests that stromal collagen type III is rather a "healing" species, and recent investi- gations [27, 40] conf i rmed this assertion. Anterior ex- c imer laser keratectomies were performed in monkey eyes, and the distribution of extracellular matr ix proteins was determined periodically up to 1 months postopera- tively [27]. At that stage, collagen type III was present in the healing region, continuing to be synthesized by the stromal fibroblasts long after reepithelialization. In an- other monkey eye study [40], the healing responses to the excimer laser ablation was followed for up to 18 months, when collagen type III was found still to be a major com- ponent of the newly generated stromal tissue.

Our results are clearly consistent with the above ob- servations and show that the presence of collagen type III within the artificial sponge matr ix represents a normal healing response of the host s t roma to the wound associ- ated with the implantation of a foreign body. By acting as a biocompatible matr ix for the deposition of collagen, the poly(HEMA) spong appears able to secure a f irm bond between the corneal tissue and the per iphery of keratoprosthesis, thus preventing the complications aris-

ing f rom lack of healing at the interface between tissue and prosthesis, including erosive necrosis ("melt ing") and subsequent extrusion of the implant.

The tissue-equivalent matr ices are important because they mimic the in vivo state of cells significantly better than monolayer cell cultures. A study in a tissue-equiva- lent matr ix made of rat tail tendon collagen has found [34] that (i) neocollagen is bound to the matr ix, not passed into the culture medium; (ii) total protein synthe- sized per cell is 2 times larger, but the total output of collagen 6 -8 times lower, than in monolayer cultures; (iii) collagenolytic activity of cells is much higher than in monolayer cultures. When corneal fibroblasts f rom embryonic chickens were cultured in vitro within three- dimensional bovine type I collagen gels, they formed layers and deposited smal l -diameter collagen fibrils contining collagen types I, V and VI [15]. (No attempts were made to localize other types of collagen.) Our study shows that po ly(HEMA) sponges are able to function as tissue-equivalent matr ices by tightly retaining the de- posited neocollagen and by withstanding the presumably enhanced enzymatic activity of fibroblasts.

Acknowledgements This work was supported by a grant from National Health and Medical Research Council of Australia (No. 910167).

References

1. Ben-Zvi A, Rodrigues MM, Krachmer JH, Fujikawa LS (1986) Immunohisto- chemical characterization of extracel- lular matrix in the developing human cornea. Curt Eye Res 5:105-117

2. Bergethon PR, Trinkaus-Randall V, Franzblau C (1989) Modified hydrox- yethylmethacrylate hydrogels as a modelling tool for the study of cell- substratum interactions. J Cell Sci 92:111-121

3. Bockxmeer FM van, Martin CE (1982) Measurement of cell prolifera- tion and cell mediated contraction in 3-dimensional hydrated collagen ma- trices. J Tissue Cult Methods 7: 163- 167

4. Chen YC, Chirila TV, Russo AV (1993) Hydrophilic sponges based on 2-hydroxyethyl methacrylate. II. Ef- fect of monomer mixture composition on the equilibrium water content and swelling behaviour. Mater Forum 17:57-65

5. Chirila TV (1994) Modern artificial corneas: the use of porous polymers. Trends Polym Sci 2:296-300

6. Chirila TV, Chen YC, Griffin B J, Constable IJ (1993) Hydrophilic sponges based on 2-hydroxyethyl methacrylate. I. Effect of monomer mixture composition on the pore size. Polym International 32:221-232

7. Chirila TV, Constable IJ, Crawford GJ, Vijayasekaran S, Thompson DE, Chert YC, Fletcher WA (1993) Poly(2- hydroxyethyl methacrylate) sponges as implant materials: in vivo and in vitro evaluation of cellular invasion. Biomaterials 14:26-38

8. Chirila TV, Constable IJ, Crawford G J, Russo AV (1994) Keratoprosthe- sis. US Patent 5,300,116

9. Chirila TV, Vijayasekaran S, Horne R, Chen YC, Dalton PD, Constable IJ, Crawford GJ (1994) Interpenetrating polymer network (IPN) as a perma- nent joint between the elements of a new type of artificial cornea. J Biomed Mater Res 28:745-753

10. Church RL (1980) Procollagen and collagen produced by normal bovine corneal stroma fibroblasts in cell cul- ture. Invest Ophthalmol Vis Sci 19:192-202

11. Chvapil M, Holuga R, Kliment K, Stol M (1969) Some chemical and bi- ological characteristics of a new colla- gen-polymer compound material. J Biomed Mater Res 3:315-332

12. Cintron C, Hong BS, Covington MI, Macarak EJ (1988) Heterogeneity of collagens in rabbit cornea: type III collagen. Invest Ophthalmol Vis Sci 29:767-775

13. Civerchia-Perez L, Faris B, LaPointe G, Beldekas J, Leibowitz H, Franzblau C (1980) Use of collagen- hydroxyethylmethacrylate hydrogels for cell growth. Proc Natl Acad Sci USA 77:2064-2068

14. Crawford GJ, Constable IJ, Chirila TV, Vijayasekaran S, Thompson DE (1993) Tissue interaction with hy- drogel sponges implanted in the rabbit cornea. Cornea 12:348-357

15. Doane KJ, Babiarz JR Fitch JM, Lin- senmayer TF, Birk DE (1992) Colla- gen fibril assembly by corneal fibrob- lasts in three-dimensional collagen gel cultures: small-diameter heterotypic fibrils are deposited in the absence of keratan sulfate proteoglycan. Exp Cell Res 202:113-124

198

16. Freed LE, Marquis JC, Nohria A, Era- manual J, Mikos AG, Langer R (1993) Neocartilage formation in vitro and in vivo using cells cultured on synthetic biodegradable polymers. J Biomed Mater Res 27:11-23

17. Freeman IL (1982) The eye. In: Weiss JB, Jayson MIV (eds) Collagen in health and disease. Churchill Living- stone, Edinburgh, pp 388-403

18. Girard LJ (1993) Girard keratopros- thesis with flexible skirt: 28 years ex- perience. Refract Corneal Surg 9:194-195

19. Jeyanthi R, Panduranga Rao K (1990) In vivo biocompatibility of collagen- poly(hydroxyethyl methacrylate) hy- drogels. Biomaterials 11:238-243

20. Kay EP (1988) Molecular anatomy of the vertebrate eye: distribution of col- lagen in ocular tissues. In: Nimni ME (ed) Collagen, vol I: Biochemistry. CRC Press, Boca Raton, Fla, pp 207- 224

21. Lee RE, Davison PF (1981) Collagen composition and turnover in ocular tissues of the rabbit. Exp Eye Res 32:737-745

22. Lee RE, Davison PF (1984) The colla- gens of the developing bovine cornea. Exp Eye Res 39:639-652

23. Legeais JM, Renard G, Salvodelli M, Keller N (1988) Etude de la colonisa- tion tissulaire du polytetrafluo- roethylene expans6 en corn6e saine en vue de son utilisation comme support de k6ratoproth6se. J Fr Ophthalmol 11 : 727-732

24. Legeais JM, Renard G, Rossi C, Salvodelli M, D'Hermies F, Pouliquen Y (1991) Keratoprosthesis: a compar- ative study of three different micropo- rous polymers and first application in human eyes. Invest Ophthalmol Vis Sci 32 [Suppl]:778

25. Legeais JM, Renard G, Pouliquen Y (1993) A novel colonizable kerato- prosthesis. Refract Corneal Surg 9:205-206

26. LiVecchi AB, Tombes RM, LaBerge M (1994) In vitro chondrocyte colla- gen deposition within porous HDPE: substrate microstructure and wettabil- ity effects. J Biomed Mater Res 28:839-850

27. Malley DS, Steinert RF, Puliafito CA, Dobi ET (1990) Immunofluorescence study of corneal wound healing after excimer laser anterior keratectomy in the monkey eye. Arch Ophthalmol 108:1316-1322

28. Marinkovich MR Rocha V (1986) Collagen synthesis and deposition dur- ing mammary epithelial cell spread- ing on collagen gels. J Cell Physiol 128:61-70

29. Marshall GE, Konstas AG, Lee WR (1991) Immunogold fine structural lo- calization of extracellular matrix components in aged human cornea. I. Types I-IV collagen and laminin. Graefe's Arch Clin Exp Ophthalmol 229:157-163

30. Marshall GE, Konstas AGR Lee WR (1993) Collagens in ocular tissues. Br J Ophthalmol 77:515-524

31. Newsome DA, Foidart JM, Hassell JR, Krachmer JH, Rodrigues MM, Katz SI (1981) Detection of specific collagen types in normal and kerato- conus corneas. Invest Ophthalmol Vis Sci 20:738-750

32. Newsome DA, Gross J, Hassell JR (1982) Human corneal stroma con- tains three distinct collagens. Invest Ophthalmol Vis Sci 22:376-381

33. Nimni ME, Harkness RD (1988) Molecular structure and functions of collagen. In: Nimni ME (ed) Colla- gen, vol I: Biochemistry. CRC Press, Boca Raton, Fla, pp 1-77

34. Nusgens B, Merrill C, Lapiere C, Bell E (1984) Collagen biosynthesis by cells in a tissue equivalent matrix in vitro. Collagen Rel Res 4:351-363

35. Olmo N, Lizarbe MA, Turnay J, Mfiller KR Gavilanes JG (1988) Cell morphology, proliferation and colla- gen synthesis of human fibroblasts cultured on sepiolite-collagen com- plexes. J Biomed Mat Res 22: 257- 270

36. Py DC, DeBiccari A, Caldwell D (1993) Keratoprosthesis: engineerig and safety assessment. Refract Corneal Surg 9:206-208

37. Shoshan S, Gross J (1974) Biosynthe- sis and metabolism of collagen and its role in tissue repair processes. Israel J Med Sci 10:537-561

38. Stol M, Tolar M, Adam M (1985) Poly(2-hydroxyethyl methacrylate)- collagen composites which promote muscle cell differentiation in vitro. Biomaterials 6:193-197

39. Stol M, Cifkov~i I, Tyrfickovfi V, Adam M (1991) Structural alterations of p(HEMA)-collagen implants. Bioma- terials 12:454-460

40. SundarRaj N, Geiss MJ, Fantes F, Hanna K, Anderson SC, Thompson KP, Thoft RA, Waring GO (1990) Healing of excimer laser ablated mon- key corneas: an immunohistochemical evaluation. Arch Ophthalmol 108:1604-1610

41. Tamada Y, Ikada Y (1994) Fibroblast growth on polymer surfaces and biosynthesis of collagen. J Biomed Mater Res 28:783-789

42. Trelstad RL, Birk DE, Silver FH (1982) Collagen fibrillogenesis in tis- sues, in solution and from modeling: a synthesis. J Invest Dermatol 79 [Suppl 1]:109-112

43. Trinkaus-Randall V, Capecchi J, Sam- mon L, Gibbons D, Leibowitz HM, Franzblau C (1990) In vitro evaluation of fibroblasia in a porous polymer. In- vest Ophthalmol Vis Sci 31:1321- 1326

44. Trinkaus-Randall V, Banwatt R, Capecchi J, Leibowitz HM, Franzblau C (1991) In vivo fibroplasia of a porous polymer in the cornea. Invest Ophthalmol Vis Sci 32:3245-3251

45. Trinkaus-Randall V, Banwatt R, Wu XY, Leibowitz HM, Franzblau C (1994) Effect of pretreating porous web on stromal fibroplasia in vivo. J Biomed Mater Res 28:195-202