Biodegradability and cell-mediated contraction of porouscollagen scaffolds: The effect of lysine as a novelcrosslinking bridge

Lie Ma, Changyou Gao, Zhengwei Mao, Jie Zhou, Jiacong ShenDepartment of Polymer Science and Engineering, Zhejiang University, Hangzhou 310027, China

Received 15 August 2003; revised 6 July 2004; accepted 9 July 2004Published online 15 September 2004 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jbm.a.30170

Abstract: A novel crosslinking method was adopted tomodify the porous collagen scaffolds by using a water-soluble carbodiimide, 1-ethyl-3-(3-dimethylaminopropyl)-carbodiimide (EDAC) and N-hydroxysuccinimide (NHS) inthe presence of lysine, which functions as a crosslinkingbridge. In vitro biodegradation tests proved that in the pres-ence of lysine the biostability of the EDAC crosslinked scaf-folds was greatly enhanced. The biostability of the resultantscaffolds was also elucidated as a function of the concentra-tions of lysine and EDAC/NHS. Compared to the Col-DHT,the ability of the Col-EDAC and the Col/Lys to resist cell-mediated contraction (CMC) was greatly enhanced. Yet noobvious difference between the Col-EDAC and the Col/Lyswas found with respect to CMC. SEM observations showedthat the microstructure of the crosslinked scaffolds could be

largely preserved after fibroblast seeding. As a result, MTTassays proved that the fibroblasts in the Col/Lys scaffoldsproliferated faster compared to the DHT-treated one on theassumption that the cell viability was preserved to a similarlevel. Histological section results indicated that the Col/Lysscaffolds had the ability to accelerate the cell infiltration andproliferation. All these results demonstrated that this novelcrosslinking method is an effective way to achieve a collagenscaffold with improved biostability and a more stable struc-ture, which can resist cell-mediated contraction. © 2004Wiley Periodicals, Inc. J Biomed Mater Res 71A: 334–342,2004

Key words: collagen; scaffold; lysine; crosslink; contraction;tissue engineering

INTRODUCTION

Because of the excellent biocompatibility of colla-gen, the use of porous collagen scaffolds as a regen-eration template for tissues such as skin,1–3 cartilage,4,5

bones,6 and nerve7,8 has been widely researched.However, the fast biodegradation rate and low me-chanical strength of the untreated collagen cannotmeet the demands of in vitro or in vivo applications inmany cases, and have been the crucial factors thatlimit the further use of this material.9,10

Crosslinking of the porous collagen scaffolds is aneffective way to slow the biodegradation rate andoptimize the mechanical properties.11,12 Currently,two kinds of crosslinking methods are frequently em-

ployed to improve the properties of the porous colla-gen scaffolds—physical methods and chemical meth-ods. The former includes the use of photooxidation,dehydrothermal treatments (DHT), and ultraviolet ir-radiation, which can avoid introducing potential cyto-toxic chemical residuals and sustain the excellent bio-compatibility of collagen materials.13 However, mostof the physical treatments cannot yield enoughcrosslinking to meet the demands of tissue engineer-ing. Therefore, treatments by chemical methods arestill necessary in most cases. Several crosslinkingagents, such as glutaraldehyde (GA), 1-ethyl-3-(3-dimethylaminopropyl)-carbodiimide (EDAC), poly-glycidyl ether, and polyepoxides have been usedwidely.11,14–17 Among them, the EDAC/N-hydroxy-succinimide (NHS) crosslinking system, which hasshown better biocompatibility than GA, is used fre-quently.12 This crosslinking takes place by reactionbetween carboxyl groups of glutamic or aspartic acidresidues and amine groups to form amide bonds.18–21

However, the crosslinking degree cannot match theapplications of tissue engineering in many cases underthe EDAC/NHS treatment. Therefore, diamine agentssuch as putrescine, hexane diamine, and so forth, have

Correspondence to: C. Gao; e-mail: cygao@mail.hz.zj.cnContract grant sponsor: Natural Science Foundation of

China; contract grant number: 50173024Contract grant sponsor: Science and Technology Program

of Zhejiang Province; contract grant number: 2004C21022Contract grant sponsor: Major State Basic Research Pro-

gram of China; contract grant number: G1999054305

© 2004 Wiley Periodicals, Inc.

been used to assist the crosslinking.15,22 Up to now theuse of amino acids as crosslinking bridges to enhancethe biostability of porous collagen scaffolds has rarelybeen reported. Amino acids are the basic componentsof proteins and naturally have good biocompatibility.A preliminary study has found that amino acids withneutral, acidic, or basic properties have different ef-fects on the crosslinking efficiency. Basic amino acidsare favorable for the improvement of biostability.

The contraction of cell-seeded matrices and in vivoimplants is a key factor that influences the construc-tion of tissue engineering organs. Because of the con-traction of cell-seeded scaffolds, changes in the sizeand microstructure of the implants take place fre-quently. A change in size will cause difficulty in fittinga specific implant site, or cause separation from thesurrounding tissue host.23,24 The crucial effect of thecontraction is that the pore size and porosity reductionof the scaffolds will impede the migration of cells fromthe surrounding tissue into the scaffolds and affect thefree exchange of the nutriments and the metabolites.Contraction of collagen-based matrices, such as gelsand scaffolds, caused by many types of connectivecells, has been reported in many studies.25 In order toresist the contraction, crosslinking treatments are ap-plied to enhance the mechanical properties of the scaf-folds.

In this study, a novel crosslinking strategy, that is,EDAC/NHS in the presence of lysine, was em-ployed to improve the biostability and the ability toresist cell-mediated contraction of porous collagenscaffolds. The factors controlling the biostability ofthe crosslinked scaffolds (Col/Lys) and the effect onthe cell-mediated contraction were researched indetail. Changes to the microstructure and the cyto-compatibility of the scaffolds were evaluated aswell.

MATERIALS AND METHODS

Materials

Collagen type I was isolated from fresh bovine tendon bya trypsin digestion and acetic acid dissolution method.26

Lysine (Lys) and hydroxyproline were purchased fromShanghai Amino Acid Company (China). Glutaraldehyde(GA), 25% water solution, was a product of Shanghai Phar-maceutical Company (China). Collagenase (278 U/mg),trypsin (250 U/mg), 1-ethyl-3-(3-dimethylaminopropyl)-car-bodiimide (EDAC), N-hydroxysuccinimide (NHS), and2-morpholinoethane sulphonic acid (MES) were purchasedfrom Sigma. All other reagents and solvents were of analyt-ical grade and were used as received.

Preparation of porous collagen scaffold

Collagen was dissolved in 0.5M acidic acid solution toobtain a 0.5% (w/v) solution. After being deaerated underreduced pressure to remove entrapped air bubbles, the col-lagen solution was injected into a homemade mold (diame-ter: 16 mm, depth: 2 mm), frozen in 70% ethanol bath at�20°C for 1 h, and then lyophilized for 24 h to obtain aporous collagen scaffold.

Crosslinking treatments

All the scaffolds were dehydrated at 105°C under vacuum(less 0.2 mbar, DHT) for 24 h prior to any additional chem-ical crosslinking. Collagen scaffolds weighted 2.5 mg wereincubated in 1 mL 50 mM MES (pH 5.5) solution containing(Col/Lys) or not containing (col-EDAC) 2.5 �M lysine for1 h at room temperature. Then 1 mL MES solution contain-ing 40 mM EDAC and 20 mM NHS was added. After incu-bation for 24 h at room temperature, the scaffolds werewashed with double-distilled water (10 min � 6 times) andlyophilized.

The porous collagen scaffolds were also crosslinked with20 mM GA in the presence or absence of lysine in order tocompare the crosslinking efficiency of these two kinds ofcrosslinking agents.

Characterization of collagen scaffolds

The microstructure of the scaffolds before or aftercrosslinking with EDAC/NHS in the presence or absence oflysine was observed under scanning electron microscopy(SEM, Cambridge stereoscan 260). The mean pore size of thescaffolds was determined by analysis of the SEM imagesfrom the cross-section positions. The porosity of each kindscaffold was measured by immersing the scaffolds (initialweight, W0) into ethanol at room temperature for 24 h (wetweight, W) and calculated by the formulas below.

porosity (%) �(W � W0)�1

�1W � ��2 � �1�W0� 100%,

where �1 and �2 represent density of collagen (1.21 g/mL)and ethanol (0.79 g/mL), respectively.

The swelling test was performed by incubating the scaf-folds in distilled water at room temperature and determin-ing the wet weight of the scaffolds 24 h later. The swellingratio of the scaffolds was defined as the ratio of weightincrease to the initial weight. Each value was averaged fromthree parallel measurements.

In vitro collagenase degradation

The biological stability of the crosslinked collagen scaf-folds was evaluated by in vitro collagenase biodegradationtest. Each kind of scaffold was incubated in phosphate-

LYSINE AS A NOVEL CROSSLINKING BRIDGE 335

buffered saline (PBS, pH 7.4) containing a given concentra-tion of collagenase (type I, Sigma) at 37°C. The degradationwas terminated at the given time interval by incubating theassay mixture in an ice bath immediately. Following centrif-ugation at 1000 rpm for 10 min, the clear supernatant washydrolyzed with 6M HCl at 120°C for 12 h. The amount ofhydroxyproline released from the collagen molecules in thescaffolds was measured by absorbance spectroscopy (CARY100 BIO, America) at a wavelength of 560 nm.27 The biodeg-radation degree is defined as the percentage of the releasedhydroxyproline from the scaffolds to the completely de-graded one with the same composition and weight.

Fibroblast isolation and scaffold seeding

Fibroblasts used in this study were isolated from humandermis by collagenase digestion.28 The cell suspensionswere cultured at 37°C, 5% CO2, and 95% humidity in DMEMsupplemented with penicillin (100 U/mL), streptomycin(100 U/mL), and 10% fetal bovine serum (FBS) (completemedium), which was changed every 3 days.

Cells were passaged at confluence and the fourth–eighthpassage fibroblasts were used for the in vitro seeding. Thescaffolds were sterilized in 75% ethanol for 12 h, followedwith solvent exchange by PBS six times. Each scaffold wasthen placed on a 24-well polystyrene plate (Falcon) andseeded with 200 �L human dermal fibroblast suspension ata density of 5 � 106 cells/mL. After incubation for 4 h, thecell-seeded scaffold was moved to another well in the 24-well plate with 1 mL complete medium. Then the scaffoldswere cultured in a 5% CO2 incubator at 37°C with mediumchange every 3 days. Cultures were terminated 3, 7, 14, and21 days after seeding. The unseeded scaffolds were incu-bated in the same culture conditions and employed as con-trols.

Measurement of the scaffold contraction

The diameters of the seeded and unseeded scaffolds weredetermined with a ruler with 0.5-mm intervals (n � 3). Thecontraction of the scaffolds was calculated by dividing thediameter change of each scaffold at the termination of theculture time by its original diameter. Herein, cell-mediatedcontraction (CMC) was determined by subtracting the con-traction of the unseeded scaffolds from the contraction of theseeded scaffolds.23,29

CMC � �original diameter � diameteroriginal diameter �

Seeded

� �original diameter � diameteroriginal diameter �

Unseeded

.

Morphology of the fibroblast-seeded scaffold

At 7 and 21 days after seeding, the culture was discontin-ued and the scaffolds were washed with PBS and fixed with

2.5% glutaraldehyde at room temperature for 24 h. Afterbeing washed with PBS to remove the remaining glutaral-dehyde, the scaffolds were dehydrated with a graded seriesof ethanol. Then the scaffolds were further dehydrated withacetone and treated with isoamyl acetate. After being driedby the critical point dry method, the scaffolds were coatedwith an ultrathin gold layer and observed under scanningelectron microscopy (SEM, Cambridge stereoscan 260).

Behavior of cell proliferation

The cell proliferation as a function of culture time wasevaluated by 3-(4,5-dimethythiazol-2-yl)-2,5-diphenyltetra-zoliumbromide (MTT) assay according to the methods ofMosmann with minor modification.30 At each culture inter-val, the media in the plates were removed and 200 �L MTT(5 mg/mL, Sigma)/complete medium (volume ratio 1:4)solution was injected into the scaffolds. After being incu-bated for 4 h, the scaffolds were transferred into a flask,where dimethylsulfoxide (DMSO) was added up to 500 �L.Then the solution was mixed thoroughly with a vortexmixer. Finally, the absorbance of violate substance was mea-sured at 570 nm by using a microplate reader (Biorad, model550).

Histological observation

At the termination of the cultures, the cell-seeded Col/Lysscaffolds were fixed with 10% neutral buffered formalin atroom temperature for 24 h and embedded in paraffin. Thespecimens were cross sectioned at a thickness of 5 �m andstained with hematoxylin and eosin (H&E). The morphologyand distribution of cells in the scaffolds were observed witha light microscope (Axiovert 200, Zeiss)

RESULTS

Microstructure and swelling property

Figure 1 shows the cross-section morphology of thescaffolds before and after crosslinking treatment. Thecross-section images reveal that the porous structureof the collagen scaffolds was largely preserved aftercrosslinking either in the presence or absence of lysine.However, more sheet-like structures could be ob-served after crosslinking. Table I lists the analyzingdata on the scaffold pore size, porosity, and PBS swell-ing ratio. As seen in the table, the typical pore size issmaller than 100 �m in the Col-DHT scaffolds. Ineither the presence or absence of lysine, the mean poresize became slightly enlarged under the EDAC/NHStreatment. No significant changes in the porosity wereobserved before or after the crosslinking treatment.All the scaffolds have porosities larger than 99%.

336 GAO ET AL.

However, the crosslinking leads to an obvious de-crease in swelling ratios. The value was lowered from105.9 � 6.2 for the Col-DHT, to 79.3 � 4.6 for theCol-EDAC, and further to 68.8 � 1.9 after crosslinkingin the presence of lysine (Col/Lys).

In vitro biodegradation

Figure 2 compares the biodegradation degree of thescaffolds by the collagenase digest test. The figureshows that the Col-DHT has been fully degraded afterincubation in the collagenase solution for 12 h. After

Figure 1. SEM images to show the cross sections of the collagen scaffolds treated with EDAC/NHS in the presence/absenceof lysine (original magnification �200). (a) Col-DHT, (b) Col-EDAC, (c) Col/Lys.

Figure 2. The effect of lysine on the biodegradation degreeof the collagen scaffolds crosslinked with EDAC/NHS orGA. All scaffolds were incubated in 100 �g/mL (30 units)collagenase for 12 h. Values are mean � standard deviation(n � 3).

TABLE IComparison of the Microstructure Parameters and

Swelling Ratio between the Col-DHT and theCross-Linked Scaffolds

Pore Size(�m) Porosity

SwellingRatios

Col-DHT 99 � 19 99.46 � 0.02% 105.9 � 6.2Col-EDAC 106 � 17 99.35 � 0.03% 79.3 � 4.6Col/Lys-EDAC 107 � 16 99.44 � 0.03% 68.8 � 1.9

LYSINE AS A NOVEL CROSSLINKING BRIDGE 337

being treated with EDAC/NHS, the crosslinked colla-gen scaffold (Col-EDAC) was more resistant to colla-genase digestion, and the biodegradation degree wasdecreased from 100% to 23.3 � 4.3%. With the addi-tion of lysine (Col/Lys), a lower biodegradation de-gree, 9.1 � 2.4%, was achieved. Compared to the valueof Col-EDAC, the decrease in the biodegradation de-gree was highly significant (P � 0.007). The biodegra-dation degree of the Col-GA and the Col/Lys-GA was7.7 � 4.6% and 6.7 � 2.8%, respectively, after the samedigestion procedure. No significant difference was de-tected between these two scaffolds (P � 0.76). TheCol-GA showed a lower biodegradation degree com-pared to Col-EDAC (P � 0.01). The addition of lysineto the scaffolds crosslinked with EDAC/NHS resultedin biostability similar to that of the GA crosslinkedmaterial.

The effect of the lysine concentration on biostabilityis illustrated in Figure 3. The biodegradation degreedecreased as a function of the concentration of lysinewhen the concentration was lower than 20 �M. From0 to 10 �M, the biodegradation degree decreasedsharply with the increase of lysine concentration. Thenthe biodegradation degree had a slight decrease whenthe lysine concentration was increased to 20 �M. Withfurther increase of the lysine concentration, the bio-degradation degree was kept nearly constant up to 100�M.

Figure 4 shows the biodegradation behavior of thescaffolds crosslinked with different concentrations ofEDAC/NHS in the presence of 2.5 �M lysine. Whenthe EDAC and NHS concentrations were 5 and 2.5�M, respectively (abbreviated as 5/2.5), the scaffoldhad been degraded 26.4% over the 12-h digestion pe-riod. After 96 h of digestion, the biodegradationdegree was increased to 60%. The degradation rate

became slower with increasing EDAC/NHS concen-tration. Only 9.9% of scaffolds were degraded afterdigestion for 96 h when the concentrations of EDACand NHS were 50/25. It is worth noting that concen-tration increase higher than 30/15 has no significanteffect on the decrease of the biodegradation degree.

Scaffold contraction

The cell-mediated contraction (CMC) of the scaf-folds as a function of culture time is shown in Figure5. For the Col-DHT, the CMC value had increased to0.34 after culture for only 3 days. Twenty-one days

Figure 3. The biodegradation degree of the collagen scaf-folds as a function of the concentrations of lysine under thetreatment of EDAC/NHS (20/10). All scaffolds were incu-bated in 500 �g/mL (150 units) collagenase for 12 h.

Figure 4. The biodegradation degree of the collagen scaf-folds as a function of the concentration of EDAC and NHS inthe presence of 2.5 �M lysine. All scaffolds were incubatedin 100 �g/mL (30 units) collagenase solutions.

Figure 5. Cell-mediated contraction (CMC) of the Col-DHT and Col/Lys scaffolds as a function of the culture time.Contraction was normalized to the shrinkage of the un-seeded scaffolds. Values are mean � standard deviation (n� 3).

338 GAO ET AL.

later, the CMC value was 0.7, meaning that the scaf-fold had been largely contracted. In contrast to theCol-DHT, no obvious contraction was observed for theCol-EDAC and Col/Lys scaffolds during the 21-dayculture period. The addition of lysine has no signifi-cant enhancement to resist the cell-mediated contrac-tion.

The cross-section images of the scaffolds after 21days of culture are presented in Figure 6. In compar-ison to Figure 1, the pores in the Col-DHT are hardlydistinguished at the same magnification [Figure 6(a)].It indicates that the pore size was reduced to a largeextent because of the cell-mediated contraction. How-ever, the three-dimensional structure of the Col/Lyswas still preserved with a pore size similar to that ofthe unseeded one because of the crosslinking treat-ment [Figure 6(b)].

Cytocompatibility

Figure 7 shows the cell-proliferation behavior byMTT assay. For the Col-DHT, the cell viability in thescaffold increased in the first week. Then, with pro-longed culture time, the cell viability decreased. How-ever, the cell viability of the Col/Lys increased overthe whole culture period, and the Col/Lys alwaysbehaved with higher viability than the Col-DHT (P 0.005 at all intervals).

The SEM images of the cell-seeded Col/Lys scaf-folds are displayed in Figure 8. After being culturedfor 7 days, several fibroblasts could be detected in thescaffold with a typical shuttle-like morphology [Fig-ure 8(a)]. When the culture time was prolonged, morecells were observed in the scaffold. The cells spreadwell along the collagen fibers and had been confluentto some extent [Figure 8(b)]. Histological results showthat the cells only existed at the edge of the scaffold

during the initial culture stage. After being culturedfor 21 days, many fibroblasts could be observed bothon the surface and in the inside of the scaffold anddistributed evenly (Figure 9). This proves that the cellshad infiltrated into the Col/Lys and proliferated there.

DISCUSSION

The effects of the lysine used as a novel crosslinkingbridge on enhancing the ability of collagen scaffolds toresist the collagenase biodegradation and the cell-me-diated contraction are demonstrated. It is well knownthat the three-dimensional porous structure of a scaf-fold is rather important for cell infiltration and prolif-eration.31 Therefore, the preservation of the porousstructure is a crucial aspect to evaluate a crosslinkingmethod. In general, the crosslinking treatment always

Figure 6. SEM images showing the cross sections of (a) the Col-DHT and (b) the Col/Lys scaffolds after culturing for 21days.

Figure 7. The comparison of the cell viability between theCol-DHT and the Col/Lys scaffolds at different culture ter-minations. Values are mean � standard deviation (n � 3).

LYSINE AS A NOVEL CROSSLINKING BRIDGE 339

results in some extent of collapse because of the rehy-dration process. After crosslinking by EDAC/NHSunder the assistance of lysine, the porous structureswere largely preserved, with a slight reduction inporosity compared to the Col-DHT one. Meanwhile,the crosslinking treatments have some effect on en-larging the pore size. These results can be explainedby the recombination of the collagen fibers during therehydration and refreeze–drying procedures of thecrosslinking treatment.28,32

Swelling ratio is another key factor of a scaffold,representing the ability to hold the nutriments to ac-celerate cell infiltration and proliferation as well as totransfer the metabolites through the scaffold.33 Thecrosslinking treatment led to about 30% shrinkage inswelling ratio. This decrease should be mainly attrib-uted to the partial diminishing of the hydrophilicgroups and the collapse of the scaffold after crosslink-

ing. The initially existing hydrophilic groups, such asamino groups and carboxyl groups, were partiallytransformed into hydrophobic amide groups underthe EDAC/NHS treatment. Nevertheless, the swellingratio of the crosslinked scaffolds is still big enough tomeet the demand of the matrices used in tissue engi-neering.

To the best of the authors’ knowledge, a highercrosslinking efficiency can be achieved under GAtreatment than under EDAC/NHS treatment. How-ever, EDAC/NHS has been used more frequently inthe treatment of porous collagen scaffolds, because ofbetter biocompatibility. In this study, the collagenscaffolds crosslinked by EDAC/NHS in the presenceof lysine had biostability similar to that of the GAcrosslinked one. The added lysine could function as acrosslinking bridge between collagen molecules. Pre-vious study has proved that the biodegradation de-gree depends on the NH2/COOH ratio in the EDAC/NHS crosslinking system. In the presence of aminoacids, there are two complete reactions. One is thefunctional group’s blocking reaction by amino acids;the other is the bridging reaction between two colla-gen molecules. The predominant reaction can be tai-lored by controlling the total NH2/COOH ratio in thecrosslinking system. Below an optimal ratio, the bio-degradation degree decreases with augmented NH2/COOH ratios; then a reversed function can be ob-tained. Therefore, the further addition of lysine resultsin the NH2/COOH ratio closing to a constant, forwhich the biodegradation degree tends to be a con-stant with increasing lysine concentration. It is easy tounderstand that the biodegradation degree will de-crease with increasing EDAC/NHS concentration atthe beginning. When EDAC/NHS is too excessive,further addition will have little effect on the crosslink-ing degree.

Along with enhanced biostability, the crosslinkingtreatment yielded a more stable structure that can

Figure 8. SEM images of the morphology of fibroblasts in the Col/Lys scaffolds after culturing for (a) 7 days and (b) 21 days.

Figure 9. Light micrograph of the H&E-stained sections ofthe fibroblast-seeded Col/Lys scaffolds at 21 days (originalmagnification, �200). Bar indicates 100 �m.

340 GAO ET AL.

better resist cell-mediated contraction. The scaffold’scontraction may be due to the contraction of the myo-fibroblasts, a major contractile phenotype of fibro-blasts that expresses a smooth muscle actin isoform,-smooth muscle actin (SMA).34 In general, the con-tractile cells are inevitable in the seeding population,because of the process of the cell isolation and thetwo-dimensional culture conditions.35 Hence, the ob-vious contraction always occurred in the Col-DHTcollagen matrix implanted in vitro or in vivo for its lowmechanical properties. Although the addition of ly-sine could bring a great enhancement in biostability,there is no obvious difference in the ability to resistcontraction between the Col-EDAC and the Col/Lys.This would mean that the contractile force caused bySMA is not big enough to distinguish the difference ofCMC between these two kinds of scaffolds.

The effects of the contraction will reduce the porediameter of the scaffolds and compress the space forcell proliferation and the diffusion of the nutrients.23 Itis important to determine the successive proliferationof the seeding fibroblasts. In order to evaluate cyto-compatibility of the scaffolds, the MTT assay was ap-plied to analyze the cell proliferation because the ab-sorbance value in 570 nm can be correlated to the cellnumber in the scaffolds on the assumption that the cellviability was preserved to a similar level.33 When theculture time was prolonged, the viability of the Col/Lys increased gradually. It can be observed that theviolating substance was only on the surface in the firstweek, then on the entire scaffold after 21 days ofculturing. The low CMC of the Col/Lys resulted in thethree-dimensional structure being largely preserved.Therefore, the Col/Lys scaffolds have enough spacefor the cell proliferation. In contrast, the Col-DHTscaffolds contracted 50% in the first week, so that thecell viability did not increase further, but decreasedwith increasing culture time. This can be explained bythe limited amount of available nutriments and thelimited amount of available space caused by the con-traction. Meanwhile, both the SEM images and thehistological sections indicated that the fibroblastscould spread along the wall of the pores and prolifer-ate well in the scaffolds. It is attributed to the originalcytocompatibility of lysine, the addition of which doesnot deteriorate the excellent cytocompatibility of col-lagen. All these results prove that this novel crosslink-ing strategy is an effective way to modify the porouscollagen scaffold.

CONCLUSIONS

Herein lysine is applied as a novel crosslinkingbridge to enhance the biostability of the porous colla-gen scaffolds. Under the crosslinking treatment, the

microstructures of the scaffolds were largely pre-served. Biostability was much improved aftercrosslinking in the presence of lysine, resulting in asimilar crosslinking degree as the GA crosslinked one.There exists an optimal concentration of lysine andEDAC/NHS to increase the biostability. Compared tothe Col-DHT, the ability of the Col-EDAC and theCol/Lys to resist the cell-mediated contraction wasgreatly enhanced. Yet no obvious difference betweenthe Col-EDAC and the Col/Lys was found with re-spect to CMC. A porous structure could remain after21 days of culture, which could accelerate the cellinfiltration. Cell MTT assay and histological resultsproved that the cells could proliferate well and dis-tribute evenly in the Col/Lys scaffolds. In conclusion,crosslinking by EDAC/NHS in the presence of lysineis an effective method to fabricate collagen scaffoldswith enhanced biostability and good cytocompatibil-ity.

References

1. Yannas IV, Lee E, Orgill DP, Skrabut EM, Murphy GF. Syn-thesis and characterization of a model extracellular matrix thatinduces partial regeneration of adult mammalian skin. ProcNatl Acad Sci USA 1989;86:933–937.

2. Yannas IV, Burke JF. Design of an artificial skin I. Basic designprinciples. J Biomed Mater Res 1980;14: 65–81.

3. Bell E, Ehrlich HP, Battle DJ, Nakatsuji T. Living tissue formedin vitro and accepted as skin-equivalent tissue of full thickness.Science 1981;211:1052–1054.

4. Speer DP, Chvapil M, Volz RG, Holmes MD. Enhancement ofhealing in ostechondral defects by collagen sponge implanta-tion. Clin Orthop 1979;144:326–335.

5. Nehrer S, Breinan HA, Ramappa A, Hsu HP, Minas T, Short-kroff S, Sledge CB, Yannas IV, Spector M. Chondrocyte-seededcollagen matrices implanted in a chondral defect in a caninemodel. Biomaterials 1998;19:2313–2328.

6. Pohunkova H, Stehlik J, Vachal J, Cech O, Adam M. Morpho-logical features of bone healing under the effect of collagen-graft-glycosaminoglycan copolymer supplemented with thetripeptide Gly-His-Lys. Biomaterials 1996;17:1567–1574.

7. Yannas IV. Regeneration templates. In: Bronzino JD, editor.The biomedical engineering handbook. Boca Raton, FL: CRCPress; 1995. p 1633

8. Chang AS, Yannas IV, Krarup C, Sethi R, Norregaard TV,Zervas NT. Polymeric templates for peripheral never regener-ation: Electrophysiological study of functional recovery. ProcACS Div Polymeric Mater Sci Eng 1988;59:906–910.

9. Freyman TM, Yannas IV, Gibson LJ. Cellular materials asporous scaffolds for tissue engineering. Prog Mater Sci 2001;46:273–282.

10. Yannas IV, Burke JF, Gordon PL, Huang C, Rubenstein RH.Design of an artificial skin II. Control of chemical composition.J Biomed Mater Res 1980;14:107–131.

11. Zeeman R, Dijkstra PJ, Wachem PB van, Luyn MJA van, Hen-driks M, Cahalan PT, Feijen J. Successive epoxy and carbodi-imide cross-linking of dermal sheep collagen. Biomaterials1999;20:921–931.

12. Rault I, Frei V, Herbage D, Abdul-Marak N, Hue A. Evaluationof different chemical methods for cross-linking collagen gel,films and sponges. J Mater Sci Mater Med 1996;7:215–221.

LYSINE AS A NOVEL CROSSLINKING BRIDGE 341

13. Lee JE, Park JC, Hwang YS, Kim JK, Kim JG, Suh H. Charac-terization of UV-irradiated dense/porous collagen mem-branes: Morphology, enzymatic degradation and mechanicalproperties. Yonsei Med J 2001;42:172–179.

14. Khor E. Methods for the treatment of collagenous tissues forbioprostheses. Biomaterials 1997;18:95–105.

15. Osborne CS, Reid WH, Grant MH. Investigation into the bio-logical stability of collagen/chondroitin-6-sulphate gels andtheir contraction by fibroblasts and keratinocytes: the effect ofcross-linking agents and diamines. Biomaterials 1999;20:283–290.

16. Courtman DW, Errett BF, Wilson GJ. The role of cross-linkingin modification of the immune response elicited against xeno-genic vascular acellular matrices. J Biomed Mater Res 2001;55:576–586.

17. Sung WW, Hsu HL, Shih CC, Lin DS. Cross-linking character-istics of biological tissues fixed with monofunctional or muti-functional epoxy compounds. Biomaterials 1996;17:1405–1410.

18. Olde Damink LHH, Dijkstra PJ, Luyn MJA van, Wachem PBvan, Nieuwenhuis P, Feijen J. Cross-linking of dermal sheepcollagen with the use of water-soluble carbodiimide. Biomate-rials 1996;17:765–774.

19. Lee JM, Edwards HL, Pereira CA, Samii SI. Cross-linking oftissue-derived biomaterials in 1-ethyl-3- (3-dimethyl amin-opropyl)-carbodiimide. J Mater Sci Mater Med 1996;7:531–542.

20. Olde Damink LHH, Dijkstra PJ, Luyn MJA van, Wachem PBvan, Nieuwenhuis P, Feijen J. In vitro degradation of dermalsheep collagen cross-linked using a water-soluble carbodiim-ide. Biomaterials 1996;17:679–684.

21. Pieper JS, Hafmans T, Veerkamp JH, van Kuppevelt TH. De-velopment of tailor-made collagen-glycosaminoglycan matri-ces: EDC/NHS cross-linking, and ultrastructural aspects. Bio-materials 2000;21:581–593.

22. Human P, Bezuidenhout D, Torrianni M, Hendriks M, Zilla P.Optimization of diamine bridges in glutaraldehyde treated bio-prosthetic aortic wall tissue. Biomaterials 2002;23:2099–2103.

23. Torres DS, Freman TM, Yannas IV, Spector M. Tendon cellcontraction of collagen-GAG matrices in vitro: effect of cross-linking. Biomaterials 2000;21:1607–1619.

24. Mueller SM, Shortkroff S, Schneider TO, Breinan HA, YannasIV, Spector M. Meniscus cells seeded in type I and type IIcollagen-GAG matrices in vitro. Biomaterials 1999;20:701–709.

25. Elsdale T, Bard J. Collagen substrata for studies on cell behav-ior. J Cell Biol 1972;54:626–637.

26. Gao CY, Wang DY, Shen JC. Fabrication of porous collagen/chitosan scaffolds with controlling microstructure for dermalequivalent. Polym Adv Technol 2003;14:373–379.

27. Pieper JS, Oosterhof A, Dijkstra PJ, Veerkamp JH, van Kuppe-velt TH. Preparation and characterization of porouscrosslinked collagenous matrices containing bioavailable chon-droitin sulphate. Biomaterials 1999;20:847–858.

28. Ma L, Gao CY, Mao ZW, Shen JC, Hu XQ, Han CM. Collagen/chitosan porous scaffold with improved biostability for skintissue engineering. Biomaterials 2003;24: 4833–4841.

29. Lee CR, Grodzinsky AJ, Spector M. The effects of cross-linkingof collagen-glycosaminoglycan scaffolds on compressive stiff-ness, chondrocyte-mediated contraction, proliferation and bio-synthesis. Biomaterials 2001;22:3145–3154.

30. Mossmann T. Rapid colorimetric assay for cellular growth andsurvival, application to proliferation and cytotoxicity assay.J Immunol Methods 1983;48:631–639.

31. Hutmacher DW, Goh JCH, Teoh SH. An introduction to bio-degradable materials for tissue engineering applications. AnnAcad Med Singapore 2001;30:183–191.

32. Ma L, Gao CY, Mao ZW, Shen JC, Hu XQ, Han CM. Thermaldehydration treatment and glutaraldehyde cross-linking to im-prove the biostability of collagen-chitosan porous scaffoldused as dermal equivalent. J Biomater Sci Polym Ed 2003;14:861–874.

33. Mao JS, Zhao LG, Yao KD, Shang QX, Yang GH, Cao YL. Studyof novel chitosan-gelatin artificial skin in vitro. J Biomed MaterRes 2003;64:301–308.

34. Hirschel BJ, Gabbiani G, Ryan GB, Majno G. Fibroblasts ofgranulation tissue: immunofluorescent staining with anti-smooth muscle serum. Proc Soc Exp Biol Med 1971;138:466–469.

35. Masur SK, Dewal HS, Dinh TT, Erenburg I, Petridou S. Myo-fibroblasts differentiate from fibroblasts when placed at lowdensity. Proc Natl Acad Sci USA 1996;93:4219–4223.

342 GAO ET AL.