ARTICLE IN PRESS

0142-9612/$ - se

doi:10.1016/j.bi

�Correspondfax: +86571 87

E-mail addr

Biomaterials 26 (2005) 6305–6313

www.elsevier.com/locate/biomaterials

Collagen-coated polylactide microspheres aschondrocyte microcarriers

Yi Hong, Changyou Gao�, Ying Xie, Yihong Gong, Jiacong Shen

Department of Polymer Science and Engineering, Zhejiang University, Hangzhou 310027, China

Received 19 October 2004; accepted 23 March 2005

Available online 23 May 2005

Abstract

Polylactide (PLA) microspheres were coated with collagen for cell culture and injectable cell carriers. Utilizing a method of

emulsion-solvent evaporation, PLA microspheres with diameter ranging from 180 to 280mm were prepared, followed with

aminolysis in hexanediamine/n-propanol solution to introduce free amino groups on their surfaces. After the amino groups were

transferred into aldehyde groups by a treatment of glutaraldehyde, collagen type I was covalently coupled via Schiff base

formation between the aldehyde groups and the amino groups on collagen molecules. Meanwhile, physically entangled collagen

molecules were retained following a grafting-coating protocol to yield microspheres coated with larger amount of collagen.

Aminolysis resulted in weight loss of the microspheres following a linear relationship with the aminolysis time. The NH2 and

collagen contents existed on the microsphere surface were quantitatively determined by ninhydrin and hydroproline (Hyp) analyses,

respectively. Larger amount of collagen was immobilized on the microspheres with higher content of NH2. In vitro chondrocyte

culture revealed that the cells could attach, proliferate and spread on these PLA microspheres, in particular on the ones having

higher content of collagen. These results show that the collagen-coated PLA microspheres are promising candidate as cell

microcarriers.

r 2005 Elsevier Ltd. All rights reserved.

Keywords: Microspheres; Polylactide; Chondrocytes; Collagen; Cell microcarriers

1. Introduction

Regeneration of wound tissue under assistance of apolymeric template has been attempted in the earlier1980s [1]. Later on it was recognized as a tissueengineering technology [2,3]. Nowadays a number ofdamaged or disabled tissues/organs such as bone,cartilage and skin are being treated in a tissueengineering way.Great progress has been made in fabricating various

three-dimension scaffolds, which basically can beclassified into non-injectable scaffolds, such as porousfoams [4–6] and non-woven meshes [7], and injectable

e front matter r 2005 Elsevier Ltd. All rights reserved.

omaterials.2005.03.038

ing author. Tel.: +86571 87951108;

951948.

ess: cygao@mail.hz.zj.cn (Changyou Gao).

ones [8–10], typically hydrogels [11,12]. More recently,attention was paid to the injectable scaffold becauseminimal incision is incurred during the transplantation.The transplanted cells proliferate and differentiate insitu in a normal histological condition as well [13].Hence, this kind of scaffold is more strongly recom-mended from viewpoints of both operational conve-nience and tissue regeneration with normal functions.Recently, microspheres have gained much attention in

tissue engineering because of their unique properties forscaling up 3D cell culture, loading capacity for drugs orgrowth factors, and most importantly, for tissueregeneration as injectable scaffolds. Cell culture onmicrospheres may produce a larger amount of cellsduring a relatively short period. The trypsinizationprocedure, which is more or less harmful to cells, can bebasically avoided since expansion of cells can be

ARTICLE IN PRESSY. Hong et al. / Biomaterials 26 (2005) 6305–63136306

achieved by simple addition of new microspheres. It isalso much convenient if the cell-attached microspheresare used directly as injectable cell microcarriers. Whenused in vivo, biodegradable microspheres are especiallyappreciated since traditional cell microcarriers, such ascollagen-coated Sephadex beads (Cytodex-3) and Culti-spher series, are less or even non resorbable. Forexample, bioresorbable porous polylactide (PLA) mi-crospheres have been used to culture rat bladder smoothmuscle cells or aortic smooth muscle cells [14,15].Moreover, retinoic acid-loaded microcarriers of poly(-lactide-co-glycolide) (PLGA) have shown a function tomodulate the differentiation of pluripotent stem cells[16].Though PLA has many advantages such as non-

toxicity, processibility and biodegradability, a surfacetreatment is still required to further improve itscytocompatibility, because the polymer is rather hydro-phobic and lack of cell binding site [17]. Many naturalbiomacromolecules such as collagen have domains intheir molecules recognized as ligands that can specifi-cally bind with integrin on cell membranes, thus caneffectively accelerate cell attachment and spread. Wehave recently developed several methods such ascovalent bonding [18], layer-by-layer assembly [19],and grafting-coating [20] to anchor biomacromoleculesonto synthetic material surfaces. Among them, grafting-coating can yield the largest introduction of biomacro-molecules, which in turn may induce better cell growthand differentiation.We report here the collagen-coated PLA microspheres

as cell carriers for chondrocyte proliferation, whichcan be expected to find important application asinjectable construct for tissue regeneration. For thispurpose, emulsion-solvent evaporation method isemployed to fabricate PLA microspheres with adiameter ranging from 180 to 280 mm. Surfaceaminolysis is performed to introduce reactive NH2

groups, which are then transferred into aldehydegroups by reaction with glutaraldehyde. By Schiffbase formation between the aldehyde groups andthe amino groups on collagen, a layer of collagen iscreated on the microspheres following a grafting-coating protocol. In vitro chondrocyte culture is carriedout to assess the cell growth behaviors and cellmorphology.

2. Experiments

2.1. Materials

PLA (average Mn ¼ 99; 000, average Mw ¼ 212; 000)was synthesized according to the method describedpreviously [21]. Methylene chloride (CH2Cl2) andisopropanol were supplied by Hangzhou Shuanglin

Chemical Company, China. Chloramine T, n-propanoland poly(vinyl alcohol) 124 (PVA 124, average Mw

85,000–124,000, 98–99% hydrolyzed) were supplied byShanghai Medicine and Chemical Company, China.Ninhydrin and p-dimethylaminobenzaldehyde weresupplied by Shanghai San’aisi Chemical Company,China. Collagen type I was extracted from bovinetendons according to the method described previously[22]. Hexanediamine was purified by distillation underreduced pressure.

2.2. Preparation of PLA microspheres

PLA microspheres were prepared by an emulsion-solvent evaporation technique [23]. Briefly, 1 g PLAwas dissolved in 20ml methylene chloride (CH2Cl2)to obtain transparent PLA/CH2Cl2 solution. Thesolution was then poured into 100ml deionizedwater containing 0.5% (w/V) PVA under agitation witha rate of 400 rpm by a mechanical stirrer. The O/Wratio was 1:5. The agitation was continued for 24 h at25 1C to evaporate the organic solvent. The producedmicrospheres were collected by a membrane filtration,extensively washed with deionized water, anddried under reduced pressure at 35 1C for 3 d. PLAmicrospheres with a diameter of 180–280 mm wereseparated by standard sieves and used for the followingstudies.

2.3. Surface aminolysis of PLA microspheres

PLA microspheres were immersed in 6% hexanedia-mine/n-propanol solution at 60 1C in water bath for agiven period. They were then extensively washed withdeionized water and dried at 35 1C under reducedpressure. The amount of introduced NH2 groups onthe aminolyzed PLA microspheres was quantified byninhydrin analysis. Briefly, �6mg dried microsphereswere placed into a glass tube containing �3ml 1mol/lninhydrin/isopropanol solution. The tube was thenheated at 100 1C for 5min to accelerate the reactionbetween the ninhydrin and the amino groups on thePLA microspheres. After discarding the residue solu-tion, 4ml 1,4-dioxane was added into the tube todissolve the microspheres. Another 2ml isopropanolwas added to stabilize the formed blue compound. Theabsorbance at 450–650 nm of this mixture was measuredon a UV–Vis spectrophotometer (CARY 100 BIO,America). A calibration curve was obtained with 1,6-hexanediamine in 1,4-dioxane/isopropanol (2:1, V/V)solution.The weight loss ratio of the microspheres after

aminolysis is defined as 100%*(W1�W2)/W1, whereW1 and W2 represent the initial and final weights of themicrospheres, respectively.

ARTICLE IN PRESSY. Hong et al. / Biomaterials 26 (2005) 6305–6313 6307

2.4. Grafting-coating and grafting of collagen type I on

the aminolyzed microspheres

The aminolyzed microspheres were immersed in 1%glutaraldehyde solution at room temperature for 3 h totransfer the NH2 groups into CHO groups. Afterwashing extensively, the microspheres were immersedin 0.5% collagen/3% acetic acid solution at 4 1C for 24 hwith occasional shaking. After filtering, two indepen-dent protocols were adopted to treat the microspheres,yielding different collagen content. In the grafting-coating protocol, the microspheres were washed withpure water directly. Since the collagen type I used here isonly acidic soluble, the physically adsorbed or entangledcollagen was also largely retained on the microspheresurfaces. Thus, comparatively larger amount of collagencan be introduced. As a comparison, extensive washingof the microspheres with 1% acetic acid solution wasalso performed to remove the physically adsorbedcollagen, remaining only the covalently grafted collagenmolecules (thus designated as grafting protocol).The surface morphology of the PLA microspheres

before and after collagen immobilization was observedunder a scanning electron microscope (JSM 5000).

2.5. Measurement of collagen content

The hydroproline (Hyp) analysis was employed toquantitatively detect the amount of collagen on themicrospheres. Briefly, the microspheres were placed in aglass tube containing 2ml 6mol/l HCl solution. Aftersealed at reduced pressure, the tube was heated at 120 1Cfor 24 h to degrade the collagen and the PLAcompletely. After removal of HCl at 70 1C, the residueswere dissolved in 2ml water. 1ml 0.05mol/l chloramine-T solution was added and reacted with this solution at25 1C for 20min, followed by addition of 1ml 3.15mol/lperchloric acid solution. Five minutes later, the mixturewas treated with 1ml 10% dimethylaminobenzaldehyde/ethylene glycol monomethyl ether solution at 60 1C for20min. The absorbance at 560 nm was measured on aUV–Vis spectrophotometer. The collagen content wasquantified by referring to a calibration curve obtainedwith pure collagen at same conditions.

2.6. Chondrocyte culture

Chondrocytes were isolated from cartilage tissue ofrabbit ears (Japanese big ear white). Briefly, cartilagetissue was cut into small pieces. Chondrocytes wereisolated by incubating the cartilage pieces in F-12HAM’s (Hyclone) culture medium containing 0.2%collagenase type II (Sigma) at 37 1C for 6 h underagitation. The isolated chondrocytes were centrifuged,resuspended in F-12 HAM’s supplemented with 20%fetal calf serum (FBS), 300mg/l glutamine, 50mg/l

vitamin C, 100U/ml penicillin and 100 mg/ml strepto-mycin. The cell suspension was then seeded in a 11 cmplastic tissue culture dish (Falcon, seeding density2� 104 cells/cm2) and incubated in a humidified atmo-sphere of 95% air and 5% CO2 at 37 1C. After aconfluent cell layer was formed (about 3–4 days), thecells were detached using 0.25% trypsin in PBS and wereresuspended in the supplemented culture medium asdescribed above, and used for the experiments.

2.7. Cell seeding, cell adhesion and proliferation

A known weight of PLA microspheres (controlsample, without any treatment) and collagen-coatedmicrospheres was sterilized by 75% ethyl alcoholsolution and UV irradiation. After washed by PBS(pH ¼ 7:4), they were placed on a 25ml glass cultureflask, which was previously treated with dimethyldichlorosilane to prohibit cell adhesion. On each speci-men 450,000 chondrocytes were seeded. The finalconcentration of the microspheres for all the sampleswas controlled as 20mg/ml. In order to accelerate celladhesion onto the microspheres, the flask was shookevery 10min in the first 1 h. The culture conditions weresame as described above.The attached cell number was determined after seeded

for 24 h. When the microspheres were suspendeduniformly in the medium via pipette blowing, 200 mlsolution (including 4mg microspheres) was taken outwith a pipette and placed into 96-well culture plate.Before harvesting the attached cells by trypsinization,twice-gentle washings with PBS were performed. Thedetached cells were then counted under a haemocyt-ometer. Cell proliferation was similarly determined aftercultured for 7 d. All data are averaged from 3 parallelexperiments.

2.8. Cell viability detected by MTT assay

Cell viability assay was performed after cultured for7 d. For this purpose, 6mg microspheres were taken outas mentioned above and placed into 96-well cultureplate. After supplemented with 40 ml MTT (3-(4,5-dimethyl) thiazol-2-yl-2,5-dimethyl tetrazolium bro-mide, 5mg/ml), the cells were continually cultured foranother 4 h. During this period, viable cells could reducethe MTT to formazan pigment, which was dissolved by100 ml dimethyl sulphoxide (DMSO) after removal of theculture medium. The absorbance at 490 nm wasrecorded under a microplate reader (Bio-Rad 550).

2.9. Cell distribution and morphologies

Cell distribution on the microspheres was observedafter cultured for 7 d. To visualize the viable cells underconfocal laser scanning microscopy (CLSM, Bio-Rad

ARTICLE IN PRESSY. Hong et al. / Biomaterials 26 (2005) 6305–63136308

Radiance 2100), the cells were incubated in 5 mg/mlfluorescein diacetate (FDA)/PBS solution for 10min. Inthis process, FDA (no fluorescence) could penetratethrough cell membranes and was hydrolyzed intofluorescein by viable cells, which was then excited at488 nm under CLSM [24].After cultured for 7 d, cell morphology on the

microspheres was also studied under SEM after fixedby 2.5% glutaraldehyde at 4 1C for 2 h. To dehydrate,the cells were sequentially treated in a series of ethanolsolution. The microspheres were then treated by acetoneand isoamylacetate successively, each for 15min at roomtemperature. Finally, critical point drying was per-formed, and the cells were observed under SEM(Stereoscan 260, Cambridge).

40

2.10. Statistical analysis

Data from all studies except for weight loss of themicrospheres were analyzed using ANOVA. Results arereported as mean7standard deviation.

00 10 20 30 40 50 60

10

20

30

Wei

ght l

oss

ratio

(%

)

Aminolysis time (min)

Fig. 1. Weight loss ratio of the PLA microspheres as a function of

aminolysis time. The aminolysis was conducted in 6% hexanediamine/

n-propanol solution at 60 1C.

1

0 2 4 6 8 10

2

3

NH

2 co

nten

t (*1

0 -7 m

ol/m

g)

Aminolysis time (min)

Fig. 2. NH2 content on the PLA microspheres as a function of

aminolysis time.

3. Results and discussion

3.1. Aminolysis of PLA microspheres

To stably immobilize collagen onto PLA micro-spheres’ surfaces, pre-treatment of their surfaces isnecessary because of the intrinsic hydrophobicity andabsence of active sites in their molecules. We haverecently developed a convenient and effective method tointroduce free NH2 groups onto polyester materials,accompanying with enhancement of their surfacehydrophilicity. The method makes use of the reactionbetween the ester groups (–COO–) of polyester anddiamine, e.g. hexanediamine. The amino groups can beintroduced onto the polyester surface in case that oneamino group reacts with –COO– group to form acovalent bond, –CONH–, while the other amino groupis unreacted. It is worth to note that hydroxyl-terminated chains will also be yielded on the polyestersurface during this process [25]. Hence, aminolysis inprinciple is an alkaline catalyzed degradation of PLAmacromolecules. Thus weight loss is unavoidable evenat rather short reaction time, as shown in Fig. 1. Thisweight loss is understood as a result of surface erosion,i.e. dissolution or dissociation of the degraded oligomerswhen their size is small enough. Fig. 1 shows that theweight loss was almost linearly proportional to theaminolysis time, except for slight deviation at the initialregion. The weight loss rate was rather fast. Forexample, a weight loss of over 5% had been achievedat reaction time of 10min. From practical consideration,the reaction time should be controlled as short as

possible since longer time was not favorable to increaseNH2 content (Fig. 2).The existence of free amino groups on the aminolyzed

PLA microspheres has also evidenced by ninhydrinanalysis (Fig. 2). The blue reaction product of ninhydrinand free NH2 groups has a maximum absorbance at550 nm in solvent of 1,4-dioxane/isopropanol (2:1 V).Fig. 2 shows that the NH2 content increased with theprolongation of aminolysis time at first few minutes,then reached almost a constant value after 6minðp40:05Þ. Similar alteration tendency has also beenobserved previously [25,26] when this method wasapplied onto flat PLA or polycaprolactone (PCL) films.However, taking into account the average diameter ofthe microspheres, from a simple calculation one can findthat the density of the NH2 groups per area on themicrospheres is �5 times higher than that on a flat film.Apart from the rougher surface texture of the micro-spheres (Fig. 3a) which may lead to deeper penetration

ARTICLE IN PRESS

Fig. 3. SEM images to show the surface morphology of the PLA microsphere before (a) and after aminolysis for (b) 2min, and (c) 8min. (d) CLSM

image taken on the cross-section of a hemisected PLA microsphere, which was incubated in 6% rhodamine/n-propanol solution at 60 1C for 8min,

followed by extensively washing.

Fig. 4. SEM images to show the collagen-coated PLA microspheres which were initially aminolyzed for (a), (c) 2min, and (b), (d) 8min, respectively.

(c) and (d) are higher magnification of (a) and (b), respectively. Noting that the collagen-discontinued areas were chosen in these images to

qualitatively identify the existence of the collagen.

Y. Hong et al. / Biomaterials 26 (2005) 6305–6313 6309

and reaction of free hexanediamine molecules, existenceof cavities in the microspheres arising from thefabrication technique should also be responsible forthis higher NH2 value. To confirm this deduction

and to estimate the penetration depth of thehexanediamine, the PLA microspheres were incubatedin rhodamine solution at same conditions as aminolysis.After hemisected, they were subjected to confocal

ARTICLE IN PRESS

1

2

3

4

5

6

7

NH2 content (*10 -7 mol/mg)

Grafted collagen (µg/mg) Grafted-coated collagen (µg/mg)

a b

Fig. 5. Collagen and NH2 contents on the PLA microspheres which

were initially aminolyzed for (a) 2min and (b) 8min, respectively

(po0:05).

0

2

4

6

8

10

Cel

l num

ber(

*104 /2

0mg)

Adhesive cell number, 24h Cell number, 7days

a b c

Fig. 6. Chondrocyte numbers on PLA microspheres at culture time of

24 h and 7 d. (a) Control PLA microspheres without treatment; (b) and

(c) collagen-coated PLA microspheres which were initially aminolyzed

for 2 and 8min, respectively. Cell seeding density 10� 104/ml;

microsphere density 20mg/ml.

0.0

0.1

0.2

0.3

Abs

orba

nce

a b c

Fig. 7. MTT viability of chondrocytes on PLA microspheres at culture

time of 7 d. (a) Control PLA microspheres without treatment; (b) and

(c) collagen-coated PLA microspheres which were initially aminolyzed

for 2 and 8min, respectively. Cell seeding density 10� 104/ml;

microsphere density 20mg/ml.

Y. Hong et al. / Biomaterials 26 (2005) 6305–63136310

characterization (Fig. 3d). A penetration depth of50–90 mm along the microsphere surface can be clearlyidentified, which is 2–3 times deeper than that on a flatfilm [26]. This would mean that the outer part of themicrospheres is sparse enough for the penetration ofsmall molecules.We took the microspheres before and after amino-

lyzed for 2 and 8min for further studies. Their surfacemorphologies were firstly observed under SEM asshown in Fig. 3. Hill-like surfaces having plenty ofgrains and occasional holes in the range of several tenmicrometers could be clearly seen on the controlmicrospheres (Fig. 3a). This surface morphology wasthe result of inhomogeneous particle shrinking duringevaporation of the organic solvent. Visible alterationwas hardly found for the microspheres aminolyzed for2min (Fig. 3b), although weight loss already occurred atthis time (Fig. 1). After aminolyzed for 8min, however,

needle-like cracks with length of several micrometersemerged on the entire particle surface (Fig. 3c). Largercracks could also be observed at this stage. Thisobservation is in accordance with the apparent weightloss detected in Fig. 1. Together with the result of NH2

content, one can conclude again that this reaction timeis already long enough to yield sufficient binding sitesfor the next collagen immobilization.

3.2. Grafting-coating of collagen on the aminolyzed PLA

microspheres

After treated the aminolyzed microspheres with alarge amount of glutaraldehyde, the NH2 groups on themicrospheres were transferred into aldehyde (–CHO)groups. Upon incubation of these –CHO enrichedmicrospheres in collagen solution the reaction betweenthe –CHO groups and amino groups on collagenmolecules (so-called Schiff base linkages) would thenyield covalent grafting of the latter. During this process,the un-grafted collagen molecules may also be physicallyentangled and intertwined with the covalently graftedcollagen layer. Consequently, the physically adsorbedcollagen molecules can be partly remained on themicrospheres after rinsing with water since collagen isnot soluble in solution of neutral pH. The as-coated(grafted-coated) collagen layer is comparatively thickand rather stable due to the existence of the precursorgrafted collagen molecules [20]. As a comparison, thephysically adsorbed collagen molecules were thoroughlyrinsed off with acetic acid solution to yield graftedcollagen only.

ARTICLE IN PRESS

Fig. 8. CLSM images to show cell distribution on PLA microspheres at culture time of 7 d. (a) Control PLA microspheres without treatment; (b) and

(c) collagen-coated PLA microspheres which were initially aminolyzed for 2 and 8min, respectively. (d) A higher magnification of (c). Cell seeding

density 25� 104/ml; microsphere density 20mg/ml. Viable cells were stained by FDA.

Y. Hong et al. / Biomaterials 26 (2005) 6305–6313 6311

SEM observations qualitatively reveal that thereindeed exist collagen fibers on the surfaces of themicrospheres, which were initially aminolyzed eitherfor 2 or for 8min and followed by collagen grafting (Fig.4). The small cracks were partly covered by the collagenfibers. Although a larger amount of collagen on themicrospheres aminolyzed for 8min could be roughlyconcluded according to these micrographs, the collagenamount was quantified by a hydroproline colorimetryanalysis. Hydroproline detection revealed that thegrafting-coating had significantly increased the collagenamount by a factor of 3.5 compared with the purelygrafting regardless of the aminolysis time (Fig. 5)ðpo0:05Þ. For example, 1.270.2 and 4.270.4 mg col-lagen/mg microspheres were measured for the graftingand the grafting-coating samples aminolyzed for 2min,respectively. On the other hand, the collagen amount forboth the grafting and the grafting-coating samples wasalso increased when the initial aminolysis time wasprolonged ðpo0:05Þ. Since longer aminolysis timeresults in larger amount of NH2 groups, it is under-standable that the grafted collagen is increased. Larger

amount of grafted collagen provides more chance forphysical adsorption of collagen through interactions ofboth physical entanglement and hydrogen bondingbetween collagen molecules.

3.3. Chondrocyte growth on the microspheres coated with

collagen

Collagen immobilization endows the microsphereswith better compatibility to cells, as demonstrated by theresults of in vitro chondrocyte culture (Figs. 6–9).Compared to the control sample (Fig. 6a, untreatedPLA microspheres), both the cell attachment numberand the cell proliferation rate were improved by a factorof 2 for the microspheres with higher collagen content(Fig. 6c) ðpo0:05Þ. However, less difference was foundbetween the control (Fig. 6a) and the microsphereshaving lower collagen content (Fig. 6b) ðp40:05Þ. MTTassay recorded a consistent tendency of cytoviability onthese microspheres (Fig. 7). Again, compared to thecontrol (Fig. 7a), significant (Fig. 7c) ðpo0:05Þ and less(Fig. 7b) ðp40:05Þ difference could be concluded for the

ARTICLE IN PRESS

Fig. 9. SEM images to show cell morphology on the PLA microspheres at culture time of 7 d. (a) Control PLA microspheres without treatment; (b)

and (c) collagen-coated PLA microspheres which were initially aminolyzed for 2 and 8min, respectively. (d) Same sample as (c) to show the cells

bridging between the microspheres. Cell seeding density 10� 104/ml; microsphere density 20mg/ml.

Y. Hong et al. / Biomaterials 26 (2005) 6305–63136312

microspheres having higher and lower collagen content,respectively.Observations under CLSM confirmed the existence of

viable chondrocytes on the surfaces of the microspheres(Fig. 8). Noting that the round black regions are PLAmicrospheres, while the bright regions represent viablechondrocytes stained by FDA. Fig. 8 shows that thechondrocytes could grow on all the PLA microspheresregardless of the existence of collagen, but the cellnumbers are different. Relatively more cells wereobserved on the microspheres with higher collagencontent (Fig. 8c,d), while a fewer chondrocytes werefound on the control (Fig. 8a) and the microsphereshaving lower collagen content (Fig. 8b). Moreover, thechondrocytes distributed more densely on the collagencoated microspheres, especially on the ones with highercollagen content (Fig. 8c,d). This is consistent with thecell proliferation and viability results (Figs. 6 and 7), inparticular for those microspheres coated with higheramount of collagen.SEM observations provided further proof to demon-

strate the different cell morphology on the control andthe collagen coated microspheres (Fig. 9). The chon-drocytes on the microspheres with higher collagencontent had spread very well, leading to the difficultyto distinguish mono cells (Fig. 9c). Interestingly, cellscould bridge between microspheres with elongatedshapes (Fig. 9d). As a result, these microspheres were

bound together and were not easy to be separated. Bycontrast, only few separate cells were observed on boththe control (Fig. 9a) and the microspheres coated withlower content of collagen (Fig. 9b). Together with thecell proliferation and viability results, one can concludethat these microspheres coated with higher content ofcollagen are more favorable for chondrocyte attachmentand growth.It is worth noting that these microspheres are not

restricted to chondrocytes. Other kinds of cells such asosteoblasts, hepatocytes and stem cells can also becultured. Together with other advantages of the micro-spheres, such as larger surface area and injectablefeature, these microspheres may find important applica-tions in fields of biomaterials and tissue engineering,particularly as injectable cell carriers.

4. Conclusions

PLA microspheres having larger amount of collagenon their surfaces have been successfully fabricated by amethod of aminolysis and grafting-coating. The intro-duced NH2 content and collagen content on the surfacesof the microspheres have been quantified by ninhydrinand hydroproline (Hyp) analyses, respectively. With ahigher NH2 content, larger amount of both the graftedand the grafted-coated collagen was found. In vitro

ARTICLE IN PRESSY. Hong et al. / Biomaterials 26 (2005) 6305–6313 6313

chondrocyte culture revealed better cell attachment,proliferation, viability, and distribution on the micro-spheres immobilized with larger amount of collagen.These results demonstrate that these collagen-coatedPLA microspheres could effectively support the attach-ment and proliferation of chondrocytes.

Acknowledgements

This study is financially supported by the Science andTechnology Program of Zhejiang Province(2004C21022) and the National Science Fund forDistinguished Young Scholars of China (No. 50425311).

References

[1] Yannas IV, Burke JF, Orgill DP, Skrabut EM. Wound tissue can

utilize a polymeric template to synthesize a functional extensional

extension of skin. Science 1982;215:174–6.

[2] Nerem RM. Cellular engineering. Ann Biomed Eng 1991;19:

529–45.

[3] Langer R, Vacanti JP. Tissue engineering. Science 1993;260:920–6.

[4] Watanabe J, Eriguchi T, Ishihara K. Cell adhesion and morpho

logy in porous scaffold based on enantiomeric poly(lactic acid)

graft-type phospholipid polymers. Biomacromolecules 2002;3:

1375–83.

[5] Ma ZW, Gao CY, Gong YH, Shen JC. Paraffin spheres as

porogen to fabricate poly(L-lactic acid) scaffolds with improved

cytocompatibility for cartilage tissue engineering. J Biomed Mater

Res Part B 2003;67B:610–7.

[6] Nam YS, Park TG. Porous biodegradable polymeric scaffolds

prepared by thermally induced phase separation. J Biomed Mater

Res 1999;47:8–17.

[7] Kim K, Yu M, Zong XH, Chiu J, Fang DF, Seod YS, Hsiaoa BS,

Chua B, Hadjiargyrou M. Control of degradation rate and

hydrophilicity in electrospun non-woven poly(D,L-lactide) nano-

fiber scaffolds for biomedical applications. Biomaterials

2003;24:4977–85.

[8] Schmitt M, Weiss P, Bourges X, Amador del Valle G, Daculsi G.

Crystallization at the polymer/calcium-phosphate interface in a

sterilized injectable bone substitute IBS. Biomaterials 2002;23:

2789–94.

[9] Iooss P, Le Ray AM, Grimandi G, Daculsi G, Merle C. A new

injectable bone substitute combining poly (�-caprolacteone)microparticles with biphasic calcium phosphate granules. Bioma-

terials 2001;22:2785–94.

[10] Salem AK, Rose FRAJ, Oreffo ROC, Yang XB, Davies MC,

Mitchell JR, Roberts CJ, Stolnik-Trenkic S, Tendler SJB,

Williams PM, Shakesheff KM. Porous polymer and cell

composites that self-assemble in situ. Adv Mater 2003;15:210–3.

[11] Lee KY, Kong HJ, Larson RG, Mooney DJ. Hydrogel formation

via cell crosslinking. Adv Mater 2003;15:1828–32.

[12] Saim AB, Cao YL, Weng YL, Chang CN, Vacanti MA, Vacanti

CA, Eavey RD. Engineering autogenous cartilage in the shape of

a helix using an injectable hydrogel scaffold. Laryngoscope

2000;110:1694–7.

[13] Hoffman AS. Hydrogels for biomedical application. Adv Drug

Deliver Rev 2002;43:3–12.

[14] Senuma Y, Franceschin S, Hilborn JG, Tissieres P, Bisson I, Frey

P. Bioresorbable microspheres by spinning disk atomization as

injectable cell carrier: from preparation to in vitro evaluation.

Biomaterials 2000;21:1135–44.

[15] McGlohorn JB, Grimes LW, Webster SS, Burg KJL. Character-

ization of cellular carriers for use in injectable tissue-engineering

composites. J Biomed Mater Res 2003;66A:441–9.

[16] Newman KD, McBurney MW. Poly(D,L lactic-co-glycolic acid)

microspheres as biodegradable microcarriers for pluripotent stem

cells. Biomaterials 2004;25:5763–71.

[17] van Wachem PB, Beugelling T, Feijen J, Bantjes A, Detmers JP,

van Aken WG. Interaction of cultured human endothelial cells

with polymeric surfaces of different wettabilities. Biomaterials

1985;6:403–8.

[18] Zhu YB, Gao CY, He T, Shen JC. Endothelium regeneration on

luminal surface of polyurethane vascular scaffold modified with

diamine and covalently grafted with gelatin. Biomaterials 2004;25:

423–30.

[19] Zhu YB, Gao CY, He T, Liu XY, Shen JC. Layer-by-layer

assembly to modify poly(L-lactic acid) surface toward improving

its cytocompatibility to human endothelial cells. Biomacromole-

cules 2003;4:446–52.

[20] Ma ZW, Gao CY, Gong YH, Ji J, Shen JC. Immobilization of

nature macromolecules on poly-L-lactic acid membrane surface in

order to improve it cytocompatibility. J Biomed Mater Res 2002;

63:838–47.

[21] Schindler A, Harper D. Polylactide II. viscosity–molecular weight

relationships and unperturbed chain dimensions. J Polym Sci

Polym Chem Ed 1979;17:2593–9.

[22] Pieper JS, Oosterhof A, Dijkstra PJ, Veerkamp JH, van

Kuppevelt TH. Preparation and characterization of porous

crosslinked collagenous matrices containing bioavailable chon-

droitin sulphate. Biomaterials 1999;20:847–58.

[23] Bodmeier R, McGinity JW. Solvent selection in the preparation

of poly(D,L-lactide) microspheres prepared by the solvent eva-

poration method. Int J Pharm 1986;43:179–86.

[24] Bancel S, Hu WS. Confocal scanning microscopy examination of

cell distribution in macroporous microcarriers. Biotechnol Prog

1996;12:398–402.

[25] Zhu YB, Gao CY, Liu XY, He T, Shen JC. Immobilization of

biomacromolecules onto aminolyzed poly(L-lactic acid) toward

acceleration of endothelium regeneration. Tissue Eng 2004;10:

53–61.

[26] Zhu YB, Gao CY, Liu XY, Shen JC. Surface modification of

polycaprolactone membrane via aminolysis and biomacromole-

cule immobilization for promoting cytocompatibility of human

endothelial cells. Biomacromolecules 2002;3:1312–9.