Biomaterials 21 (2000) 2315}2322
Vascular cell responses to polysaccharide materials: in vitro and in vivo evaluations Janeen M. Chupa, Angela M. Foster, Stephanie R. Sumner, Sundararajan V. Madihally, Howard W.T. Matthew* Department of Chemical Engineering & Materials Science, Wayne State University, 5050 Anthony Wayne Drive, Detroit, MI 48202, USA
Abstract Chitosan has shown promise as a structural material for a number of tissue engineering applications. Similarly the glycosaminoglycans (GAGs) and their analogs have been known to exert a variety of biological activities. In this study we evaluated the potential of GAG}chitosan and dextran sulfate (DS)}chitosan complex materials for controlling the proliferation of vascular endothelial (EC) and smooth muscle cells (SMC). GAG}chitosan complex membranes were generated in vitro and seeded with human ECs or SMCs for culture up to 9 d. In addition, porous chitosan and GAG}chitosan complex sca!olds were implanted subcutaneously in rats to evaluate the in vivo response to these materials. The results indicated that while chitosan alone supported cell attachment and growth, GAG}chitosan materials inhibited spreading and proliferation of ECs and SMCs in vitro. In contrast, DS}chitosan surfaces supported proliferation of both cell types. In vivo, heparin}chitosan and DS}chitosan sca!olds stimulated cell proliferation and the formation of a thick layer of dense granulation tissue. In the case of heparin sca!olds the granulation layer was highly vascularized. These results indicate that the GAG}chitosan materials can be used to modulate the proliferation of vascular cells both in vitro and in vivo. 2000 Elsevier Science Ltd. All rights reserved. Keywords: Chitosan; Glycosaminoglycans; Heparin; Dextran sulfate; Endothelial cells; Smooth muscle; Sca!old; Tissue response
1. Introduction The expanding "eld of Tissue Engineering has accelerated the demand for materials which are tissue compatible, biodegradable, and with mechanical properties closely matched to the target tissues [1}4]. Molecular level control of biological activity is also a highly desirable feature. For many such materials, porous microstructures are also required to either allow tissue ingrowth in vivo or to provide a template for directed tissue assembly in vitro. The matrix polysaccharides termed glycosaminoglycans (GAGs) and composite materials derived from them are of interest for such applications since carbohydrate moieties interact with or are integral components of many cell adhesion molecules and matrix glycoproteins [5]. In this work, bioactive polysaccharide-based materials are being investigated as
* Corresponding author. Tel.: #1-313-577-5238; fax: #1-313-5773810. E-mail address:
[email protected] (H.W.T. Matthew).
potential solutions to a long-standing biomaterials problem, namely, the design of an e!ective small-diameter vascular graft. Incomplete endothelialization and smooth muscle cell hyperplasia are two of the problems contributing to the poor performance of existing small-diameter vascular grafts [6}8]. Improvements in long-term performance of these devices might be attained through the use of structural materials with speci"c biological activities. In one scenario, the bioactive sca!old would both enhance the rate of endothelialization and speci"cally inhibit the migration of smooth muscle to the graft lumen. GAG-based materials hold promise for this system because of their growth inhibitory e!ects on vascular smooth muscle cells and their anticoagulant activity. Furthermore, these molecules readily form complexes with the structural polysaccharide derivative chitosan [9]. In this study, the morphological and growth responses of vascular cells to polysaccharide complex surfaces were evaluated. In addition, porous bioactive materials were prepared by complexation of GAGs with porous chitosan sca!olds and the tissue response to these implants was examined.
0142-9612/00/$ - see front matter 2000 Elsevier Science Ltd. All rights reserved. PII: S 0 1 4 2 - 9 6 1 2 ( 0 0 ) 0 0 1 5 8 - 7
2316
J.M. Chupa et al. / Biomaterials 21 (2000) 2315}2322
2. Materials and methods Unless otherwise noted, all reagents were obtained from Sigma Chemical Co. (St. Louis, MO). 2.1. Preparation of immobilized GAG-complex culture membranes The GAGs evaluated in this study were hyaluronic acid (HA) from human umbilical cord, heparin from porcine intestinal mucosa, chondroitin sulfate A (CSA, chondroitin-4-sulfate) from bovine articular cartilage, chondroitin sulfate B (CSB, dermatan sulfate) from bovine intestinal mucosa, and chondroitin sulfate C (CSC, chondroitin-6-sulfate) from shark cartilage. Dextran sulfate (DS, MW"500 000), a semi-synthetic GAG analog, was also evaluated. Chitosan with a 90% degree of deacetylation was obtained from Carbomer Inc. (Westborough, MA). Polysaccharide complex membranes were deposited onto tissue culture polystyrene dishes using the ionic interaction between a GAG solution and a pre-deposited chitosan-acetate membrane. To prepare sterile chitosan solution (2 wt% in 0.2 M acetic acid), 2 g of chitosan was autoclaved in 100 ml of water, and then dissolved by adding 1.2 g of sterile "ltered glacial acetic acid and stirring for 2 h. Sterile glycosaminoglycan (GAG) solutions (1 wt%) were prepared by dissolving the GAG in saline (0.9% NaCl) bu!ered with 50 mM HEPES (pH 7.4) and then autoclaving. Cell cultures were conducted in 24-well culture plates, with all culture surfaces represented on each plate by 3 replicate wells. Untreated culture wells were used as controls. Pure chitosan culture surfaces were prepared by covering the bottom of each well with 50 ll of sterile chitosan solution. The solution was then evaporated to dryness at room temperature in a biological laminar #ow hood. This process produced a coating of dry chitosan acetate. The coated wells were then washed with 0.5 ml of 0.1 M NaOH to neutralize the chitosan and create a hydrogel membrane. This was followed by four washes with 1.5 ml of 50 mM HEPES-bu!ered saline (pH 7.4). GAG}chitosan complex surfaces were prepared by "rst coating wells with 50 ll of sterile chitosan solution and again evaporating to yield a chitosan acetate membrane. The coated wells were then covered with 0.5 ml of sterile GAG solution and allowed to equilibrate for a minimum of 6 h. This procedure resulted in the deposition of a GAG}chitosan complex membrane on the dish. The formed complex membranes were then washed 5 times with 1.5 ml of HEPES-bu!ered saline. 2.2. Cell seeding and culture Human coronary artery endothelial cells and human coronary artery smooth muscle cells were obtained from
Clonetics (Walkersville, MD) and used between passages 2 and 6. Endothelial and smooth muscle cultures were conducted separately. Cells were seeded onto several 24-well culture plates coated with GAG}chitosan membranes. Seeding density was 4000 or 5000 cells/cm, and 0.5 ml of culture medium was used per well. The culture medium was changed 24 h post-seeding and every 48 h thereafter. For endothelial cells, the culture medium consisted of MCDB 131 base medium supplemented with 2% fetal bovine serum, 2.4 lg/ml vascular endothelial growth factor, 10 ng/ml human epidermal growth factor (hEGF), 1.0 lg/ml hydrocortisone, 50 lg/ml of gentamycin, and 50 ng/ml amphotercin B. The smooth muscle culture medium consisted of MCDB 131 supplemented with 2% FBS, 2 lg/ml human "broblast growth factor (hFGF-b), 10 ng/ml hEGF, 10 lg/ml insulin, 50 lg/ml of gentamycin, and 50 ng/ml amphotercin B. 2.3. Morphological analysis After 5 d of culture, phase contrast microscopic images of the cells on each surface were digitally captured. Morphometric analyses were conducted using image analysis software (Sigma Scan Pro, Jandel Scienti"c, CA). The parameters determined included mean values of projected cell area, shape factor and area mean diameter. 2.4. Evaluation of growth kinetics Cell growth was evaluated by the MTT-formazan assay. Culture plates were sacri"ced at 48 h intervals. The culture medium was aspirated and replaced with 500 ll of MTT solution (thiazolyl blue, 2 mg/ml in Krebs}Ringer bicarbonate bu!er, pH 7.4). The plate was incubated for 2 h at 373C. The solution was then aspirated and 500 ll of DMSO was added to dissolve the formazan crystals. After 10 min of rotary agitation, the absorbance of the DMSO solution at 540 nm was measured using a spectrophotometer. Exponential cell growth was assumed and a speci"c growth rate (k) was determined by "tting the following equation to the absorbance data:
ln
A "k(t!t ), A
where A is the absorbance at t , A the absorbance at t, k the speci"c growth rate (h\), t the time (h), and t the initial time (h). 2.5. Implantation of porous polysaccharide scawolds Porous chitosan sca!olds were fabricated by freezing and lyophilizing chitosan solutions as previously described [10]. Brie#y, chitosan solution (2 wt% in 0.2 M acetic acid) was poured into a #at-bottomed polyethylene
J.M. Chupa et al. / Biomaterials 21 (2000) 2315}2322
weighing dish to a depth of 2 mm. Freezing was accomplished by placing the dish on a surface maintained at !783C. The frozen sample was then lyophilized until dry. The lyophilized chitosan sca!olds were cut into 0.75 cm squares. Pure chitosan sca!olds were rehydrated and sterilized for implantation by immersion in absolute ethanol for 1 h followed by 70% isopropanol for 2 d. GAG}chitosan complex sca!olds were prepared by immersing freshly lyophilized chitosan sca!olds in 1% heparin or 1% dextran sulfate for 3 h. The hydrated complex sca!olds were rinsed extensively with water and then sterilized by immersion in 70% isopropanol for 2 d. All sterilized sca!old samples were thoroughly rinsed and equilibrated with PBS prior to implantation. The samples were implanted in dorsal subcutaneous pockets of male Sprague Dawley rats, using standard surgical procedures. Animals were sacri"ced 1, 2 and 3 weeks post-implantation. Retrieved sca!old samples were "xed in 4% paraformaldehyde in PBS (pH 7.4), and processed for para$n embedding. Sections (10 lm) were stained with hematoxylin and eosin and evaluated by light microscopy.
3. Results 3.1. Polysaccharide complex surfaces In solution, chitosan is a positively charged polymer and is capable of forming insoluble ionic complexes with negatively charged polymers such as the GAGs. Similarly, dehydrated chitosan salts can be complexed in situ by rehydrating with an aqueous solution of a negatively charged polymer. We have previously used this ionic interaction to microencapsulate primary hepatocytes within hydrogel membranes [9,11,12]. In this study, pure chitosan membranes were smooth and completely transparent. Most GAG}chitosan membranes appeared smooth or slightly textured under the phase contrast light microscope. In the case of heparin, dextran sulfate and hyaluronate complexes, the surface appeared granular and highly textured. GAG}chitosan complex membranes exhibited uniform staining with toluidine blue, indicating uniform GAG distribution. After several days of extensive washing with PBS, toluidine blue staining was still evident, indicating that measurable quantities of GAG remained throughout the culture period. 3.2. Cell morphology on polysaccharide surfaces Both endothelial (EC) and smooth muscle cells (SMC) attached well ('90% attachment) to all the test surfaces. However, large di!erences in cell morphology were observed on the various surfaces. On the chitosan surfaces (Fig. 1c and d), EC and SMC both exhibited a normal morphology comparable to the control polystyrene
2317
(Fig. 1a and b). On the GAG}chitosan surfaces however, quite di!erent morphologies were observed. On the chondroitin sulfate surfaces (CSA, CSB, CSC), both cell types exhibited a clear stellate appearance. The SMCs, in particular, appeared to be extending several narrow pseudopodal structures from a poorly spread central cell body (Fig. 1e and f). On the heparin and hyaluronate surfaces, both cell types exhibited a rounded morphology, though all cells were well attached (Fig. 1g and h). On the dextran sulfate surface, EC exhibited a stellate morphology. SMCs in contrast were able to spread enough to establish a cellular network (Fig. 1i and j). In order to quantify cell spreading, projected cell areas for each cell type were normalized to the mean area seen on tissue culture plastic (control). This approach allowed direct comparison between the responses of the two cell types to each surface. The normalized area data are shown in Fig. 2. On chitosan, ECs exhibit a 20% increase in spreading while SMCs exhibited a 40% reduction. On the GAG surfaces (heparin, hyaluronate, chondroitin sulfates), both cell types exhibited reduced cell spreading, with SMCs being more severely a!ected on a fractional basis. For both cell types hyaluronate surfaces produced the minimum spreading (i.e. almost spherical cells). In one experiment, the projected cell area for smooth muscle cells was measured at intervals over the course of a 9 d culture. A 25}100% increase in cell area was observed on the GAG surfaces (Fig. 3). Since a corresponding increase was not seen on the control or chitosan surfaces, this fact suggested that the cell}surface interactions were changing with time. We previously determined that the GAG}chitosan surfaces release most of the GAG over the course of culture [13]. Therefore, it is likely that the reduction in GAG content was leading to an increased cellular interaction with the remaining insoluble chitosan component, and an accompanying increase in spreading. Taken together, the morphological data show that polysaccharide surfaces can in#uence the morphological characteristics of vascular cells in vitro. In addition, the magnitude of the e!ect is a function of GAG type and density. 3.3. Cell growth kinetics Since the degree of cell spreading is often correlated with the magnitude of a growth response, the proliferation rates on the polysaccharide surfaces was evaluated. Fig. 4 illustrates the speci"c growth rates calculated. Chitosan membranes supported EC proliferation, but in contrast to the increased spreading seen on this surface, the growth rate was 50% lower than on tissue culture polystyrene. Similarly, SMCs exhibited a 70% reduction in speci"c growth rate on chitosan compared to control. No cell growth was detected on the GAG}chitosan surfaces for either cell type. The negative rates shown in Fig. 4
2318
J.M. Chupa et al. / Biomaterials 21 (2000) 2315}2322
Fig. 1. Photomicrographs of endothelial and smooth muscle cells (respectively) on polysaccharide surfaces: (a, b) polystyrene; (c, d) chitosan; (e, f ) chondroitin sulfate B (dermatan sulfate); (g, h) heparin; (i, j) dextran sulfate.
J.M. Chupa et al. / Biomaterials 21 (2000) 2315}2322
2319
represent a net loss of cells due to poor adhesion. This idea is consistent with the observation that the highest cell loss rates were observed on those surfaces which also produced the lowest cell spreading: namely CSB, CSC and hyaluronate. Cell loss over the course of culture was clearly evident with a low rate of cell detachment. However, since the MTT assay actually measures mitochondrial activity, the declines may also have been partially due to some reduction in oxidative capacity on the GAG surfaces. In stark contrast to the GAG}chitosan surfaces, the DS}chitosan surface produced substantial cell growth. This again appears to correlate with the spreading result which was equivalent to control in both cases. Fig. 2. E!ect of polysaccharide surfaces on EC and SMC spreading. Projected cell areas were normalized to the mean of the control for a given cell type. Values are mean and standard deviation from two independent experiments with a minimum of 20 cells measured per point. Asterisk indicates statistically signi"cant di!erences (p(0.05) between the two cell types.
Fig. 3. Variation of SMC projected area over the course of culture for CSB and CSC. Values are mean and standard deviation from one culture experiment with a minimum of 10 cells measured per point. Asterisk indicates a statistically signi"cant di!erence compared to day 1 values (p(0.02 for CSB and p(0.001 for CSC).
Fig. 4. E!ect of polysaccharide surfaces on the speci"c growth rates of EC and SMC. Values are mean and standard error determined from linear regression and ANOVA of culture data.
3.4. Tissue response and vascularization of polysaccharide scawolds Implantation of chitosan and GAG}chitosan porous sca!olds was conducted to glean preliminary information on the e!ects of GAG release from the complex sca!olds in an in vivo environment. The controlledtemperature freezing and lyophilization produced an open pore microstructure which was maintained after rehydration and GAG complexation. Mean pore diameters were &80 lm. Rehydrated chitosan sca!olds were soft and #exible, but maintained their size shape during implantation. However, during histological processing, the tissue component of the sample exhibited some shrinkage. Since the chitosan material did not shrink, this resulted in compression and distortion of the pore structure (Fig. 5a). After 2 weeks, chitosan sca!olds exhibited signi"cant tissue ingrowth and no major foreign body response. A thin "brous layer, 5}10 cell layers thick and containing many capillaries, covered the surface of the sca!old. Highly cellular, granulation tissue had penetrated the pores to an average depth of &200 lm (Fig. 5b). Penetration depth was dependent on pore size, with larger pores supporting greater penetration. No signs of material degradation were observed. In comparison to the chitosan sca!old, the heparin}chitosan and DS}chitosan sca!olds were much softer and lacked su$cient sti!ness to resist compression after implantation. As a result no tissue ingrowth was observed in these samples. However, both complex materials induced formation of a thick (500}800 lm) layer of granulation tissue around the implants (Fig. 5c). Higher magni"cation examination revealed that the heparin granulation layer contained numerous capillaries in all areas. These neo-vessels were all oriented perpendicular to the sca!old surface (Fig. 5d). Similar vascularity could not be con"rmed within the DS granulation layer. However, in both cases, the surface of the granulation layer was particularly well supplied with small blood vessels.
2320
J.M. Chupa et al. / Biomaterials 21 (2000) 2315}2322
Fig. 5. Photomicrographs of chitosan and GAG}chitosan implants. (a, b) Chitosan implant at day 14. Note the pore structure and tissue ingrowth from both surfaces. The convoluted pore appearance is a tissue shrinkage induced artifact. (c) DS}chitosan implant and associated granulation tissue. The collapsed DS structure is located on the left. (d) High magni"cation of the heparin}chitosan granulation layer (the scale bar should read 60 l not 120 l). Note the numerous capillaries running from the surface (upper left) towards the implant (lower right).
4. Discussion Polysaccharides are a class of materials which, while they have enjoyed popularity for certain specialty applications, have generally been underutilized in the Biomaterials "eld. Recognition of the potential utility of this class of materials is however growing and the "eld of polysaccharide biomaterials is poised to experience rapid growth. In particular, recent advances in the "eld of polysaccharide synthesis, together with an improved understanding of polysaccharide-mediated molecular recognition and signaling have raised awareness of the potential of these molecules. In this study, we conducted preliminary investigations on the response of vascular cells to polysaccharide-based ionic complex materials. We determined that complexes of GAGs or DS with chitosan can exert control over the adhesion, spreading and proliferation of both endothelial and smooth muscle cells. In addition, chitosan proved to be a promising material for tissue sca!old applications and release of heparin and DS was found to enhance neo-tissue formation and angiogenesis around the polysaccharide implants. We previously reported on the use of chitosan to fabricate prototypes for vascular graft sca!olds [10]. Our results show that chitosan supported proliferation of
both vascular cell types, but also retarded smooth muscle cell growth to a greater extent than endothelial growth. Since SMC hyperplasia is one of the primary causes of failure in small-diameter vascular grafts, this result suggests that the sca!old material may have intrinsic biological activity which could contribute to a retardation of SMC growth relative to ECs. The mechanism of this growth reduction is not clear but it may be related to the relative levels of cell spreading. Spreading, in turn, may be related to the components and conformations of the adsorbed protein layer. It should also be noted that chitosan's glucosamine group is also found in GAGs and may thus have direct integrin or receptor interactions. The results of our cell spreading measurements together with morphological observations suggest that the limited cell spreading seen on the GAG}chitosan surfaces was a result of poor adhesion. This is further supported by the observation that there was a gradual loss of cells on those surfaces which also did not support proliferation. We previously showed that GAG}chitosan complexes release the GAG component over time [13]. This continuous release of GAG was probably a signi"cant contributor to the overall poor adhesion. GAGs are known to bind to a variety of matrix and serum proteins, and the desorption may have prevented the adsorption of serum components such as "bronectin or cell-synthesized
J.M. Chupa et al. / Biomaterials 21 (2000) 2315}2322
collagen. This idea is supported by the fact that cell spreading increased later in culture when the GAG #ux would have been signi"cantly reduced. The growth inhibition seen on the GAG surfaces may have resulted in part from the reduction in spreading, since cell spreading has been shown to correlate with DNA synthesis in some cell types [14]. However, the growth factor binding capability of GAGs in general may also have contributed to the growth inhibition. GAGs are known to bind a wide variety of peptide growth factors. Heparin in particular has been extensively studied for its interactions with factors in the "broblast growth factor (FGF) family [15}18]. The FGF family includes factors with stimulatory e!ects on proliferation of both endothelial cells and smooth muscle cells. Their binding interactions with heparin can result in either enhancement or inhibition of growth factor activity depending on the relative concentrations of the two species and the nature of the ionic environment. Thus it is likely that the high levels of soluble GAG present early in our cultures may have bound and sequestered both mediumsupplied and endogenously secreted growth factors. The observation that DS}chitosan surfaces produced results quite di!erent from the GAG}chitosan surfaces is particularly intriguing. In this study, we initially treated DS as simply a semi-synthetic GAG analog because of its sulfated sugar residue structure. The molecule has also been reported to have growth factor binding activity similar to heparin [19,20]. However, contrasting proliferation e!ects seen here suggest that an alternative mechanism may have been in e!ect. Further study is clearly needed, but the possible mechanisms include speci"c matrix protein binding and direct receptor interactions similar to that reported for heparin [20]. It should also be noted that the molecular weight disparity between the DS and the other GAGs may have played a role. Heparin and the chondroitin sulfates have a maximum molecular weight of &50 000 Da, whereas the DS had an average molecular weight of 500 000 Da. At this molecular weight the DS would be less likely to desorb to the degree known to occur with GAGs. A more stable surface may therefore have allowed more secure cell attachment, greater spreading and positive growth kinetics. Interestingly, hyaluronate (MW&1 000 000 Da) produced results similar to the other GAGs. However, its known low protein binding and direct interactions with its receptor may have played a role in this scenario. The results of our implantation study indicate that chitosan has very good potential for in vivo tissue scaffold applications. Furthermore, the rapid and extensive tissue growth response to heparin and DS sca!olds suggests that the GAG desorption from such sca!olds can be harnessed to help stimulate tissue regeneration within the tissue engineering paradigm. The strong angiogenic response to the heparin sca!old holds particular promise for the use of heparin}chitosan complexes for vascular
2321
graft applications as well as other tissue applications requiring neo-vascularization. The fact that this angiogenic response was obtained without added growth factors suggests that the desorbed heparin served to enhance the activity of endogenously secreted growth factors in the vicinity of the implant. This e!ect may have been synergistic with the known ability of chitosan to attract neutrophils and activate macrophages [21}25]. In conclusion, our results indicate that chitosan and chitosan complexes with GAGs and DS have signi"cant potential for the design of new biologically active biomaterials which can modulate the activities of vascular endothelial and smooth muscle cells in vitro and in vivo.
Acknowledgements This work was partially supported by grants from the National Science Foundation (BES-9624151) and the Whitaker Foundation for Biomedical Engineering (970361).
References [1] Peter SJ, Miller MJ, Yasko AW, Yaszemski MJ, Mikos AG. Polymer concepts in tissue engineering. J Biomed Mater Res 1998;43:422}47. [2] Hubbell JA. Biomaterials in tissue engineering. Biotechnology (NY) 1995;13:565}76. [3] Cima LG, Vacanti JP, Vacanti C, Ingber D, Mooney D, Langer R. Tissue engineering by cell transplantation using degradable polymer substrates. J Biomech Eng 1991;113:143}51. [4] Langer R, Vacanti JP. Tissue engineering. Science 1993;260: 920}6. [5] Lindahl U, Hook M. Glycosaminoglycans and their binding to biological macromolecules. Annu Rev Biochem 1978;47:385}417. [6] Bos GW, Poot AA, Beugeling T, van Aken WG, Feijen J. Smalldiameter vascular graft prostheses: current status. Arch Physiol Biochem 1998;106:100}15. [7] Massia SP, Hubbell JA. Tissue engineering in the vascular graft. Cytotechnology 1992;10:189}204. [8] Schoen FJ, Castellot Jr JJ. Vascular graft intimal "brous hyperplasia: prospects for pharmacological inhibition. J Vasc Surg 1991;13:758}61. [9] Matthew HW, Salley SO, Peterson WD, Klein MD. Complex coacervate microcapsules for mammalian cell culture and arti"cial organ development. Biotechnol Prog 1993;9:510}9. [10] Madihally SV, Matthew HW. Porous chitosan sca!olds for tissue engineering. Biomaterials 1999;20:1133}42. [11] Matthew HW, Salley SO, Peterson Jr WD, Deshmukh DR, Mukhopadhyay A, Klein MD. Microencapsulated hepatocytes. Prospects for extracorporeal liver support. ASAIO Trans 1991;37: M328}30. [12] Matthew HW, Basu S, Peterson WD, Salley SO, Klein MD. Performance of plasma-perfused, microencapsulated hepatocytes: prospects for extracorporeal liver support. J Pediatr Surg 1993;28: 1423}7; discussion 7}8. [13] Madihally SV, Flake AW, Matthew HWT. Maintenance of CD34 expression during proliferation of CD34# cord blood cells on glycosaminoglycan surfaces. Stem Cells 1999;17:295}305.
2322
J.M. Chupa et al. / Biomaterials 21 (2000) 2315}2322
[14] Hansen LK, Mooney DJ, Vacanti JP, Ingber DE. Integrin binding and cell spreading on extracellular matrix act at di!erent points in the cell cycle to promote hepatocyte growth. Mol Biol Cell 1994;5:967}75. [15] DeBlois C, Cote MF, Doillon CJ. Heparin}"broblast growth factor}"brin complex: in vitro and in vivo applications to collagen-based materials. Biomaterials 1994;15:665}72. [16] Guimond S, Maccarana M, Olwin BB, Lindahl U, Rapraeger AC. Activating and inhibitory heparin sequences for FGF-2 (basic FGF). Distinct requirements for FGF-1, FGF-2, and FGF-4. J Biol Chem 1993;268:23906}14. [17] Maccarana M, Casu B, Lindahl U. Minimal sequence in heparin/heparan sulfate required for binding of basic "broblast growth factor. J Biol Chem 1993;268:23898}905. [18] Zhou FY, Kan M, Owens RT, McKeehan WL, Thompson JA, Linhardt RJ, Hook M. Heparin-dependent "broblast growth factor activities: e!ects of de"ned heparin oligosaccharides. Eur J Cell Biol 1997;73:71}80. [19] Gao G, Goldfarb M. Heparin can activate a receptor tyrosine kinase. Embo J 1995;14:2183}290.
[20] Amiji MM. Surface modi"cation of chitosan membranes by complexation*interpenetration of anionic polysaccharides for improved blood compatibility in hemodialysis. J Biomater Sci Polym Ed 1996;8:281}98. [21] Usami Y, Okamoto Y, Minami S, Matsuhashi A, Kumazawa NH, Tanioka S, Shigemasa Y. Chitin and chitosan induce migration of bovine polymorphonuclear cells. J Vet Med Sci 1994;56:761}2. [22] Kosaka T, Kaneko Y, Nakada Y, Matsuura M, Tanaka S. E!ect of chitosan implantation on activation of canine macrophages and polymorphonuclear cells after surgical stress. J Vet Med Sci 1996;58:963}7. [23] Usami Y, Okamoto Y, Takayama T, Shigemasa Y, Minami S. Chitin and chitosan stimulate canine polymorphonuclear cells to release leukotriene B4 and prostaglandin E2. J Biomed Mater Res 1998;42:517}22. [24] Usami Y, Okamoto Y, Minami S, Matsuhashi A, Kumazawa NH, Tanioka S, Shigemasa Y. Migration of canine neutrophils to chitin and chitosan. J Vet Med Sci 1994;56:1215}6. [25] Peluso G, Petillo O, Ranieri M, Santin M, Ambrosio L, Calabro D, Avallone B, Balsamo G. Chitosan-mediated stimulation of macrophage function. Biomaterials 1994;15:1215}20.
