Acta Biomaterialia 7 (2011) 2163–2175

Contents lists available at ScienceDirect

Acta Biomaterialia

journal homepage: www.elsevier .com/locate /actabiomat

Organic/inorganic hybrid network structure nanocomposite scaffolds basedon grafted chitosan for tissue engineering

D. Depan, P.K.C. Venkata Surya, B. Girase, R.D.K. Misra ⇑Biomaterials and Biomedical Engineering Research Laboratory, Center for Structural and Functional Materials, University of Louisiana at Lafayette,P.O. Box 44130, Lafayette, LA 70504-4130, USA

a r t i c l e i n f o

Article history:Received 4 November 2010Received in revised form 12 January 2011Accepted 20 January 2011Available online 1 February 2011

Keywords:Hydroxypropyl chitosanAlginatePorous scaffoldsBiodegradationCell culture

1742-7061/$ - see front matter � 2011 Acta Materialdoi:10.1016/j.actbio.2011.01.029

⇑ Corresponding author. Tel.: +1 337 482 6430; faxE-mail address: dmisra@louisiana.edu (R.D.K. Misr

a b s t r a c t

We describe the first study of structure–processing–property relationship in organic/inorganic hybridnetwork structure nanocomposite scaffolds based on grafted chitosan for bone tissue engineering. Chito-san was first grafted with propylene oxide to form hydroxypropylated chitosan, which was subsequentlylinked with ethylene glycol functionalized nanohydroxyapatite to form an organic/inorganic networkstructure. The resulting scaffold was characterized by a highly porous structure and significantly superiorphysico-chemical, mechanical and biological properties compared to pure chitosan. The scaffolds exhib-ited high modulus, controlled swelling behavior and reduced water uptake, but the water retention abil-ity was similar to pure chitosan scaffold. MTT assay studies confirmed the non-cytotoxic nature of thescaffolds and enabled degradation products to be analyzed. The nanocomposite scaffolds were biocom-patible and supported adhesion, spreading, proliferation and viability of osteoblasts cells. Furthermore,the cells were able to infiltrate and colonize into the pores of the scaffolds and establish cell–cell inter-actions. The study suggests that hydroxypropylation of chitosan and forming a network structure with anano-inorganic constituent is a promising approach for enhancing physico-chemical, functional and bio-logical properties for utilization in bone tissue engineering applications.

� 2011 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

1. Introduction

Chitosan (CS) is the partially deacetylated form (poly-b(1,4)-2-amino-2-deoxy-D-glucose) of chitin, a natural polymer found inthe cell wall of fungi and microorganisms. It is a potential biomate-rial for tissue engineering applications such as the repair of osseousand chondral defects. Because of its biocompatibility, biodegrad-ability, hydrophilicity, good adhesion and non-toxicity, CS is gener-ally preferred in many biomedical applications, including tissueregeneration, as a drug carrier and as an antimicrobial agent [1].In spite of the benefits outlined above, there is a need to enhancethe mechanical properties and promote the biological response ofCS, especially for bone tissue engineering applications [2,3]. Chem-ical approaches involving synthesis of CS derivatives with diversechemical and molecular structures have been attempted in recentyears to enhance its physico-chemical properties and cellular re-sponse [4–7]. Polyethylene glycolization was observed to enhancethe water solubility of CS, while incorporation of hyaluronic acidimproved the biological response of pure CS [8,9].

The ability of CS to support cell attachment and proliferation isrelated to the structural and chemical properties of the polysaccha-

ia Inc. Published by Elsevier Ltd. A

: +1 337 482 1220.a).

ride backbone of CS, which is structurally similar to glycoseamino-glycans, the major component of the extracellular matrix of boneand cartilage. Other advantages of CS with respect to bone tissueengineering include the formation of highly porous scaffolds withinterconnected pores, osteoconductivity and the ability to enhancebone formation in vitro and in vivo [10,11].

CS is known for its biocompatibility in tissue engineering forcell propagation and is a potential candidate to repair osseousand chondral defects, but the mechanical properties and biologicalresponse of CS scaffolds can be improved further to make themsuitable for bone tissue engineering [12]. A tissue engineering scaf-fold should have adequate mechanical strength so as to transferthe applied load at the implant site and to participate in matrixmineralization. Furthermore, a tissue engineering scaffold shouldbe interconnected and possess sufficient porosity (preferably over90%) to promote cell adhesion, in-growth and reorganizationin vitro. Pore interconnections improve the nutrients diffusion,while providing better room for neovascularization. Many studieshave attempted to improve the mechanical strength of CS by incor-porating various nanofillers in the form of bioceramics, such ashydroxyapatite (HA), by various in situ hybridization methods[13]. In the recent past, while studies have been devoted towardthe improvement in physico-chemical and biological properties,the applied wet chemical methods to disperse HA have prevented

ll rights reserved.

2164 D. Depan et al. / Acta Biomaterialia 7 (2011) 2163–2175

uniform dispersion and coherent binding with chitosan matrix.This is expected to result in the migration of HA particles fromthe implanted site [14]. In this regard, the fabrication of nanohy-brids with tunable bioactivity and biodegradability and appropri-ate mechanical property for bone tissue engineering is currentlybeing explored [15,16].

The above-mentioned reports indicated significant improve-ment in mechanical properties, including bending strength(86 MPa) and modulus (3.6 GPa) of the nanohybrids, whileenhanced cell proliferation and spreading on the chitosan–nHAscaffolds in comparison to pure chitosan scaffolds was attributedto the presence of apatite [17,18]. A recent approach toward thesynthesis of HA crystals by a diffusion method also indicated thatthe nanohybrid scaffolds exhibited improved osteointegration[19]. However, a rational approach to correlate the process–structure–functional property relationship and biological responseis yet to be adequately addressed.

To accomplish the objective of enhancing both the biologicalresponse and the mechanical properties, we describe the firststudy of the structure–processing–property relationship in organ-ic/inorganic hybrid network structure nanocomposite scaffoldsbased on grafted chitosan for bone tissue engineering. Chitosanwas first grafted with propylene oxide to form hydroxypropylatedchitosan, which was subsequently linked with ethylene glycolfunctionalized nanohydroxyapatite (f-nHA) to form an organic/inorganic network structure. An accompanying objective was toexplore the effect of incorporation of sodium alginate into the or-ganic–inorganic hybrid network structure on the strength andstiffness of the hybrid scaffold. The synthesized organic–inorganichybrid network structure nanocomposite scaffolds were studiedfor microstructure, physico-chemical, mechanical and biologicalproperties. The suitability and cellular response of scaffolds werestudied using mouse pre-osteoblasts (MC3T3-E1).

2. Materials and methods

2.1. Materials

High molecular weight CS (400 kDa, 85% degree of deacetyla-tion, as provided by the supplier) was obtained from Aldrich (St.Louis, MO, USA). Nanohydroxyapatite (nHA) powder and hexam-ethylene diisocyanate (HMDI), dibutyltindilaurate, ethylene glycol,propylene oxide (PPO) and N,N-dimethyl formamide (DMF), cal-cium chloride and other reagents of analytical grade were obtainedfrom Aldrich, USA. The average diameter of crystalline nHA was�50 nm based on transmission electron microscopy. Alphaminimum essential medium and fetal bovine serum used werepurchased from Gibco, Invitrogen Corporation (USA). Penicillin–streptomycin (10,000 IU–10,000 lg ml�1), trypsin–EDTA (0.25%trypsin/0.53 mM EDTA) in Hank’s buffered salt solution and phos-phate-buffered saline (PBS) without calcium and magnesium wereobtained from American Type Cell Culture Collection (ATCC,Manassas, VA, USA).

2.2. Functionalization of nHA with ethylene glycol

Hydroxyapatite was functionalized by dispersing 3 g of nHA inDMF, followed by the addition of HMDI and dibutyltindilaurate[20]. The reaction was carried out in an inert atmosphere for 8 hat room temperature. To this, ethylene oxide was added and theresulting solution was heated overnight at 60 �C, with continuousstirring. After the completion of reaction, the f-nHA was collectedby washing with methylene chloride and centrifugation at2500 rpm, followed by drying in vacuum for 24 h.

2.3. Hydroxypropylation or grafting of chitosan and fabrication oforganic–inorganic hybrid network structure scaffolds

The four types of scaffolds prepared were: CS, CS–nHA, CS–grafted nHA and CS–grafted nHA with alginate (AL). Pure CS scaf-folds were prepared by dissolution of CS (2% w/v) in aqueous aceticacid (1 vol.%) and lyophilization. CS–nHA scaffolds were preparedby the dropwise addition of a dispersion of nHA (0.01 g) withdeionized water. To prepare the hydroxypropyl derivative of CS,a solution of CS was pre-heated at 70 �C, PPO (20 vol.%) was addedslowly to the CS solution and heating was continued for another1 h, followed by the dropwise addition of f-nHA. The resultingmixture was stirred for 2 h and the scaffold was prepared by lyoph-ilization (CS–g-nHA). The synthesis of the fourth type of scaffold(CS–grafted nHA–AL) involved addition of AL solution (1% w/v inNaOH) to the CS–g-nHA solution and stirring for 2 h, followed bylyophilization and cross-linking of scaffolds in the presence ofCaCl2 solution overnight. The cross-linked scaffolds were thenwashed several times with distilled water to remove unboundCaCl2 and again lyophilized to prepare CS–g-nHA–AL scaffolds.The CS–g-nHA and CS–g-nHA–AL scaffolds are denoted as CSderivative scaffolds hereafter in this article.

2.4. Chemical and morphological characterization

To confirm the grafting of PPO onto CS and the incorporation ofnHA and AL into the scaffold, Fourier transform infrared (FTIR)spectra were recorded on a JASCO FTIR-480 spectrophotometer inthe range between 4000 and 400 cm�1, with a resolution of2 cm�1 using a KBr pellet.

The surface morphology, pore morphology and pore-size distri-bution of the scaffolds was evaluated using scanning electronmicroscopy (SEM; Hitachi S-3000). Scaffolds were mounted on alu-minum stubs using adhesive tape, followed by sputter coating witha gold layer to minimize the accumulation of negative charge fromthe electron beam during SEM analysis. SEM micrographs wereused to estimate the average pore diameter (d):

d ¼ffiffi

lp� h ð1Þ

where l and h are the maximum and minimum pore lengths, respec-tively, while porosity was determined using Eq. (2) [21]:

porosityð%Þ ¼ V � ðW=qÞ=V � 100% ð2Þ

where V is the volume of the scaffold (cm3), W is the weight of thescaffold (g) and q is the density of a non-porous CS film (g cm�3).Confocal laser scanning microscopy (Leica, USA) was used to inves-tigate the three-dimensional structure of the scaffolds. Micrographswere obtained at the surface and at different depths approaching�500 lm. Scaffolds were also viewed at different rotations to com-plete a series of 360�.

2.5. In vitro swelling and degradation studies

The in vitro swelling behavior of the scaffolds was investigatedusing three different solutions at different pHs: 1 N HCl (pH 1.2),1 N NaOH (pH 14) and simulated body fluid (SBF) (pH 7.4). Theswelling behavior was quantified by measuring the change in thesample diameter as a function of the sample immersion time inthe medium. The water uptake, water retention ability andin vitro biodegradation properties of the scaffolds were studiedas follows: briefly, dry scaffolds were weighed (Wd) and immersedin distilled water, then taken out after 24 h, followed by blottingwith filter paper to remove the excess bulk water and weighedagain (Ww) to determine the water uptake. The water retentioncapability was estimated by centrifuging the wet scaffolds

D. Depan et al. / Acta Biomaterialia 7 (2011) 2163–2175 2165

(500 rpm, 3 min) followed by weighing (W 0W), in accordance with

ISO-62-1980.The percentage of the equilibrium water absorption (EA) and

water retention (ER) of the scaffolds at equilibrium were calculatedusing Eqs. (3) and (4), respectively:

EA ¼ ½ðWw �WdÞ=Wd� � 100 ð3Þ

ER ¼ ½ðW 0w �WdÞ=Wd� � 100 ð4Þ

Enzymatic degradation of the scaffolds was studied using alysozyme degradation test. The initial dry weight of the samples(W0) was recorded before they were incubated in the degradationmedium for hydrolysis (0.1 M PBS containing 500 lg ml�1 of lyso-zyme at 37 �C) in an incubator. The degradation medium (PBS) waschanged every 3 days to maintain a constant pH value of 7.4. Deg-radation was allowed to proceed for fixed time intervals, afterwhich they were taken out from the degradation medium, washedwith distilled water, lyophilized and weighed (Wt). Degradationwas quantified as the change in sample weight over time. The per-centage of weight remaining was estimated by:

%weight remaining ¼ 100� ½ðW0 �WtÞ=W0Þ � 100� ð5Þ

2.6. Mechanical properties

The modulus of pure CS and its grafted derivatives were evalu-ated in the dry state by the depth sensing indentation approachusing a nanoindentor (MTS Systems, Oak Ridge, TN) of 50 nm ra-dius indenter, calibrated with a fused silica standard. A maximumload of 0.15 mN was set and 15 indents were made at 35 lm inter-vals for each sample. The maximum indentation depth was set to1000 nm. The load–displacement data were recorded continuouslythrough one complete cycle of loading and unloading. The Pois-son’s ratio of 0.35 reported for HA [22] was used to calculate themodulus of the scaffolds.

2.7. Cell culture studies

2.7.1. Cell proliferation and viabilityCell culture studies were conducted using mouse pre-osteo-

blasts MC3T3-E1 cell line from ATCC. Pre-osteoblasts (80–85% con-fluence) were seeded onto the scaffolds and investigated for thecytocompatibility of CS and its derivative scaffolds. The scaffoldswere kept in a Petri dish and incubated with pre-osteoblasts for4 h, after which the numbers of the unbound cells were countedas follows. First, the scaffolds were taken out of the Petri dish, thenthe cells that were not bound to the scaffolds but attached to thePetri dish were detached enzymatically using trypsin and countedusing a hemocytometer (Hausser Scientific, Horsham, USA). Thecell attachment to the scaffolds was determined by subtractingthe number of unattached cells from the number of cells initiallyseeded [23].

The proliferation of pre-osteoblasts on CS and its derivative scaf-folds was studied by using fluorescence microscopy after stainingwith the nucleic acid staining dye Hoechst 33342 (bisbenzimide tri-hydrochloride) at predetermined times. The cell–scaffold constructswere washed twice with PBS, followed by incubation with dye(10 lg ml�1 of PBS) for 10 min at room temperature. After incuba-tion, the scaffolds were washed again with PBS to remove the excessdye and viewed under a fluorescence microscope (Nikon, ECLIPSE E600 FN) with ultraviolet (UV) excitation and emission at 346/442 nm. The cell nuclei appeared as blue fluorescence after staining.

The cell viability of the cell-seeded scaffolds was measuredusing MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazoliumbromide) assay (Sigma). For this, the scaffolds were washed withPBS, followed by incubation with fresh culture medium containing

MTT (0.5 mg ml�1 medium) at 37 �C for 4 h in darkness. After-wards, the unreacted dye was removed and dimethylsulfoxidewas added to dissolve the intracellular insoluble purple formazanproduct into a colored solution. The absorbance of the resultingsolution was measured at 570 nm on a spectrophotometric platereader (Bio TEK Instrument, EL 307 C).

Cell morphology was assessed using SEM. Cell seeded scaffoldswere rinsed twice with PBS, fixed with 2 vol.% glutaraldehyde in0.1 M cacodylate buffer for 30 min, then washed with PBS. Thenthe samples were dehydrated in an ascending ethanol series, crit-ically point dried, sputter coated with gold and finally examined bySEM. The amount of calcium phosphate produced by cells wasdetermined by energy-dispersive X-ray analysis (EDS) in an elec-tron microscope (EDS 2006 software, IXRF system).

In order to assess the cell compatibility of the solid particlesremaining from the scaffolds after 28 days of enzymatic biodegra-dation, the particles were isolated from the degradation mediumand dissolved in distilled water to provide 75,000 ppm stock solu-tions, followed by filtration (0.45 lm filters) and then sterilizingunder UV light for 3 h prior to cell culture [24]. MC3T3-E1 pre-osteoblasts (85–90% confluence) were added to tissue cultureplates at 75,000 cells per well. The appropriate volume of the deg-radation stock solution was pipetted into three wells of each plateto provide 10 and 1000 ppm concentration in three wells of sepa-rate plates.

The rest of the wells were used as controls, while the total vol-ume in each well was made up to 4 ml with Dulbecco’s minimumessential medium and the prepared plates were placed in a CO2

incubator for 72 h. After the incubation the cells were trypsinizedand stained with trypan blue, and the cell numbers in each wellwere counted using a hemocytometer.

The cell attachment (adhesion) study was performed as follows:briefly, 12 mm diameter scaffolds were packed into a 12 mm tubecoated with Teflon. Then 0.1 ml of a pre-osteoblast suspension(1.4 � 107 cells ml�1) was loaded onto the scaffold-loaded column.The cells were allowed to attach in a humidified incubator for 1 h.Each column was rinsed with 1 ml of PBS, and the unattached cellswere quantified by the spectroscopic observation of the rinsedsolution.

2.7.2. Data AnalysisFor all the experiments, a minimum of five samples were used.

Five tests were carried out for all the cell culture studies, while theobtained values in each experiment were normalized with the con-trol samples. The results are expressed as the mean value of at leastfive replicates ± standard deviation (SD). Statistical analysis wascarried out using a one-way analysis of variance with 95% confi-dence intervals. The error bars denote ± SD (n P 5).

3. Results and discussion

3.1. Fabrication of scaffolds

The synthesis of CS–g-nHA scaffolds involved the facile graftingof PPO onto the surface of CS by hydroxypropylation under acidicconditions and the subsequent reaction with f-nHA. This approacheliminated the need for initiators and a coupling agent. Further-more, the linkage of PPO with the amine group (–NH2) of CS gen-erated –OH groups that were subsequently employed to anchorthe f-nHA to prepare the CS–g-nHA scaffold. The reaction stepsare summarized in Scheme 1. Hydroxypropylation of CS is anetherification process achieved using PPO as the etherifying re-agent. The reaction is unique in that it introduces hydroxypropylgroups in the CS polymeric chain (step 1), which are utilized forestablishing a link with f-nHA. The f-nHA ensures covalent linkingwith the hydroxypropylated CS scaffold to form a network

O

OO

OH

HONH2

H3COCHN

OH

HO

CH3

CH-CH2

O

O

OO

OH

HO

NH CHCH2OH

H3COCHN

OH

HO

CH3

OH

O

OO

OH

HO

NH-R

H3COCHN

OH

HO

R =

CH3

CH-CH2-OH

= -O-CH2-CH2-O-

Step-1

Step-2

Chitosan (CS) Propylene oxide (PPO)

Grafted Chitosan (Cs-g)

nHA

Functionalized nHA (f-nHA)

Grafted Chitosan-Functionalized nHA (CS-g-nHA)

nHA

Scheme 1. Schematic illustration of the reaction scheme for the hydroxypropylation of CS by PPO and synthesis of grafted chitosan–functionalized nHA.

2166 D. Depan et al. / Acta Biomaterialia 7 (2011) 2163–2175

structure (step 2). Thus, nHA is not present as a reinforcement fillerbut is an integral part of CS network structure. This is anticipatedto impart high modulus and strength. Freeze drying provides aninterconnected pore structure for cell penetration and prolifera-tion. In an attempt to further increase the modulus, AL was addedto CS–g-nHA. AL can be directly added to the CS–g-nHA complex ina single step to prepare CS–g-nHA–AL scaffolds.

a

4000 3000 2000 1000

Tran

smit

tan

ce (

arb

.un

its)

Wavenumber (cm-1)

1596 1420

1654

1153 10703459

3454 1642 1590 1420 1153

1070 1031

955 635

2970

1570 1385

11601031

1650 1424

b

c

d

Fig. 1. FTIR spectrum of (a) CS, (b) CS–nHA, (c) CS–g-nHA and (d) CS–g-nHA–AL.

3.2. Chemical and morphological characterization

FTIR spectroscopy is considered an appropriate tool to study theinteraction between polymer and nanoparticles. The FTIR spectrumof pure CS has the typical absorption peaks assigned to the saccha-ride moiety at 3459 cm�1 corresponding to the stretching vibrationof N–H, while peaks at 1654 and 1574 cm�1 represent the presenceof amide-I and amide-II, respectively (Fig. 1). The double amidepeaks for CS correspond to the partial N-deacetylation of chitin[25]. Furthermore, the sharp peaks at 1420 cm�1 correspond tothe CH3 symmetrical deformation, while peaks at 1153, 1070 and1031 cm�1 are assigned to CAO stretching vibrations (m(CAOAC)).

The FTIR spectrum of CS–nHA suggests the characteristic bandsof both nHA and CS, as shown in Fig. 1 and summarized in Table 2.In comparison to pure CS, CS–nHA scaffold is characterized bytwo new absorption bands, at 955 and 635 cm�1, which corre-spond to the stretching vibration bands of P–O from PO4

3� andthe bending deformation mode of O–H from nHA, confirmingthe incorporation of nHA and subsequent formation of CS–nHAnanocomposite scaffold. Furthermore, the 3570 cm�1 peak corre-sponding to m(O–H) of nHA modifies the broad ms(N–H)3459 cm�1 peak of CS.

In the case of CS–g-nHA, two new absorption bands appeared inthe IR spectrum. The peak at 2970 cm�1 corresponds to CAHstretching and the other, at 1385 cm�1, represents the asymmetricdeformation of the CH3 group. These absorption peaks indicatedthat the CH3 group was introduced into the CS molecule afterreacting with PPO, thus confirming the successful completion ofthe hydroxypropylation reaction under acidic conditions [26].

D. Depan et al. / Acta Biomaterialia 7 (2011) 2163–2175 2167

Furthermore, the absorption peaks at 1031 and 1160 cm�1, whichcorrespond to the primary and secondary AOH groups of CS poly-saccharide moiety, were still present in the FTIR spectra of CS–g-nHA, confirming that the hydroxypropylation occurred with theamine (NH2) group of CS. Furthermore, the linking of f-nHA withhydroxypropylated CS was confirmed by the presence of HA peaksat 1630 and 1570 cm�1 [20]. The above observations corroboratethe synthesis of organic–inorganic hybrid nanocomposite scaffold,where nHA is an integral part of the network structure.

Fig. 2. SEM micrographs of (a–c) pure CS, (d–f) CS–nHA, (g–i) CS–g-nHA and (j–l) CS–dispersion of nHA. The nanoparticles are encircled.

Table 1Summary of physical, chemical, mechanical and biological properties of pure CS and its d

Scaffold/property CS

Average pore diameter (lm) 100 ± 20Average diameter of interconnecting pores (lm) 40 ± 15% Porosity 93 ± 2% Water uptake/water retention 900/800Modulus (GPa) 0.061 ± 0.017% Adhesivity of osteoblasts 72 ± 5

The results presented are means ± SD.

In the spectrum of CS–g-nHA–AL scaffold, the peak at1424 cm�1 corresponds to carboxyl –COOH group in AL, and theamide-I peak shifted from 1654 to 1650 cm�1 suggests the forma-tion of the CS–AL complex as a result of the ionic interaction be-tween the negatively charged carbonyl group (–COOH) of AL andthe positively charged amino group (–NH2) of CS [27].

The physical appearance of lyophilized chitosan and its deriva-tive scaffolds was stiff and inelastic. The representative SEM micro-graphs of all the scaffolds revealed a highly porous, homogeneous

g-nHA–AL scaffolds. (m) SEM micrograph of CS–nHA showing the homogeneous

erivative scaffolds.

CS–nHA CS–g-nHA CS–g-nHA–AL

110 ± 20 120 ± 30 110 ± 3045 ± 10 50 ± 12 50 ± 1494 ± 3 96 ± 2 95 ± 3730/680 620/560 640/5700.068 ± 0.013 0.075 ± 0.025 0.096 ± 0.03176 ± 4 81 ± 4 82 ± 4

Table 2Assignment of FTIR spectra of CS, CS–nHA, CS–g-nHA and CS–g-nHA–AL scaffoldspresented in Fig. 1.

Samples IR absorption bands (cm�1) Descriptiona

3459 ms(N–H)

CS 1654 m(AC@OA) amide I1574 amine1420 d(CAH)1153,1070, 1031 mas(CAOAC) and ms(CAOAC)3454 ms(NAH)1642 m(AC'OA) amide I

CS–nHA 1586 amine1420 d(CAH)1153,1070, 1031 mas(CAOAC) and ms(CAOAC)955 m(PAO) for PO4

3�

635 d(OAH)2970 ms(CAH)1570 ANHCOA

CS–g-nHA 1385 mas(CAH)1153, 1031 mas(CAOAC) and ms(CAOAC)1424 ACOOH

CS–g-nHA–AL 1650 m(AC@OA) amide I

a m = stretching vibration, ms, symmetric stretching vibration; mas, asymmetricstretching vibration; d, bending vibration.

Fig. 3. Three-dimensional confocal scanning micrographs of (a) pure CS, (b) CS–nHA, (c) Cof slices were taken along the z-axis for CS–g-nHA scaffold. The three-dimensional strcorrespond to 45� intervals. (f) Confocal images of CS–g-nHA scaffold obtained at the surthe z-axis. Randomly selected section numbers 1, 3, 9 and 12 are presented.

2168 D. Depan et al. / Acta Biomaterialia 7 (2011) 2163–2175

and interconnected structure (Fig. 2). Based on the quantitativeanalysis of a number of micrographs, the average pore diameterand average diameter of interconnecting pores were �100 and�40 lm, respectively, with 93% porosity (Table 1). While the incor-poration of nHA nanoparticles into the CS matrix did not alter themicrostructure of the scaffold, the SEM micrograph of CS–nHA athigh magnification (Fig. 2m) suggests the uniform dispersion ofnHA nanoparticles in the CS matrix. This uniform distribution isattributable to the cationic and hydrophilic nature of CS, whichfacilitates the homogeneous distribution in an aqueous solution.The pore diameter and the diameter of the interconnecting poresfor the CS–g-nHA scaffold were 120 and 50 lm, respectively.

The small increase in pore size is presumably related to thegrafting of PPO onto CS, which results from the swelling of the ma-trix required for the PPO to diffuse into the CS matrix and then tobecome cross-linked with the amine group of CS (Scheme 1). Apore diameter in the range of 110–150 lm is appropriate for tissueengineering applications [28]. A number of researchers have re-ported CS-based scaffolds with large pore sizes. For instance,scaffolds of cross-linked CS with pore diameters in the range of

S–g-nHA and (d) CS–g-nHA–AL scaffolds at x = �27�, y = 21� and z = 29�. (e) A seriesucture of the scaffold is presented as a series of 360� rotations; micrographs 1–9face and up to a depth of 100 lm viewed as a series with 5 lm thick sections along

D. Depan et al. / Acta Biomaterialia 7 (2011) 2163–2175 2169

30–325 lm have been synthesized [29,30]. In comparison to thesereports, the grafted CS scaffolds synthesized by us have the desiredpore size of 100 ± 20 lm and porosity of 90% for bone tissueregeneration.

The three-dimensional interconnected pore structure of CS andits derivative scaffolds were confirmed by confocal scanningmicroscopy (Fig. 3). For instance, the micrographs of CS–g-nHAscaffold were taken at different depths and directions to revealthe uniform porous structure along all the axes.

3.3. Shape retention and swelling studies

Equilibrium swelling, water absorption and retention was mea-sured for CS and its derivative scaffolds. These properties areimportant from the viewpoint of the absorption of fluids and exu-dates in bone tissue engineering. The scaffolds were immersed insolutions of different pHs, including SBF (pH 7.4), HCl (pH 1.2)and NaOH (pH 14), for 2 weeks. Fig. 4 summarizes the shape reten-tion data in terms of the diameter of scaffold as a function ofimmersion time. Pure CS scaffold disintegrated in the acidic med-ium, while CS–g-nHA and CS–g-nHA–AL scaffolds experiencedswelling, with diameters increasing by �20% within 2 h, and re-tained their overall size and shape in all three solutions. The swell-ing behavior of the CS–g-nHA and CS–g-nHA–AL scaffolds wasfound to be stable regardless of the pH value of the solution. The

Time (h)0 1 2 3 4 5 6

Time (h)0 1 2 3 4 5 6

12

13

14

15

16

17

18

19

20

21

22

Dia

met

er (m

m)

SBFHClNaOH

c. CS-g-nHA

13

14

15

16

17

18

19

20

21

22

Dia

met

er (m

m)

SBFHClNaOH

a. CS

Fig. 4. Shape retention of (a) pure CS, (b) CS–nHA, (c) CS–g-nHA and (d) CS–g-nHA–AL scaare means ± SD (n = 5).

swelling of pristine CS generally involves the protonation of ami-no/imine groups and the mechanical relaxation of coiled CS chains[31]. It is proposed that in the case of CS–g-nHA and CS–g-nHA–ALscaffolds –NH2 groups are used as the graft-point and the remain-ing amino groups interact with the –COOH groups of AL, prevent-ing protonation.

In tissue engineering, water uptake and retention of scaffoldsare considered to be important parameters because the absorp-tion of physiological fluid and the transfer of nutrients andmetabolites occurs through the scaffolds. Water uptake andretention data (Table 1) shows that the synthesized scaffoldshad the ability to retain more water than their respective weightas the obtained values were greater than 100%. This is advanta-geous for biomedical applications because water retention abilitysuggests that the scaffold can hold water to an extent determinedby the test, such that the tight aggregation of the polymericchains might make the scaffolds stable in size and shape duringcell culture or during contact with the implanted site. It was ob-served that the drying of swollen scaffolds took a longer time todry to a constant weight, indicating that the derivative scaffoldsheld the moisture for a longer time (Table 1).

This is attributed to nHA, which acts as a physical barrier for themoisture to exude out from the scaffolds and intact macromolecu-lar chains that makes the scaffolds stable in size and shape. Thewater holding property of the scaffold is beneficial to bone repair

SBFHClNaOH

b. CS-nHA

SBFHClNaOH

d. CS-g-nHA-AL

Time (h)0 1 2 3 4 5 6

Time (h)0 1 2 3 4 5 6

12

13

14

15

16

17

18

19

20

21

22

Dia

met

er (m

m)

12

13

14

15

16

17

18

19

20

21

22

Dia

met

er (m

m)

ffolds as a function of scaffold immersion time in media of different pHs. The values

2170 D. Depan et al. / Acta Biomaterialia 7 (2011) 2163–2175

because the scaffold can prevent exudate accumulation and wounddehydration [32]. Furthermore, in the case of grafted scaffolds, thewater uptake was inhibited by the rigid network of macromolecu-lar chains. The water retention behavior is favorable for cell adhe-sion. In tissue engineering applications, during the course of cellproliferation, the retained hydrophilicity of the scaffolds is ex-pected to enhance cell attachment and proliferation at the surfaceof the bone implant site.

3.4. In vitro biodegradation and cell toxicity study of the degradedproduct

The long-term functioning of tissue engineered cell–materialconstructs depends on the enzymatic biodegradation behavior ofthe materials. The porous scaffolds in tissue engineering are ex-pected to naturally disintegrate as the tissue grows. Thus, the deg-radation time of the scaffold affects the implantation site. It wasobserved that pure CS scaffolds exhibited more degradation thanits derivative scaffolds. The weight loss of CS and CS–nHA scaffoldsafter 28 days of immersion in the degradation medium were 25%and 19%, respectively. In the case of CS–g-nHA and CS–g-nHA–ALscaffolds, the weight loss was 15% and 16%, respectively.

CS primarily degrades by lysozyme, present in the physiologicalfluid and tissue, while nHA is bioresorbable, eventually dissolvingand releasing calcium and phosphate ions which assist in boneregeneration [33]. The results suggested that grafting of PPO im-proved the stability of CS scaffolds towards enzymatic degradation.The CS–g-nHA and CS–g-nHA–AL scaffolds degraded slowly, whileCS degraded rapidly.

The degradation led to s change in the structural morphology interms of collapse of the cell wall. The SEM micrographs taken after

CS

100 µm

a

CS-nHA

b

100 µm

e

12 24 360.0

0.2

0.4

0.6

0.8

1.0

1.2

Cel

l Den

sity

(OD

570

)

Time

Fig. 5. Scanning electron micrographs of (a) CS, (b) CS–nHA, (c) CS–g-nHA and (d) CS–g-nof the degradation products of scaffolds after 72 h. Mean ± SD (n = 5).

28 days of enzymatic degradation indicated that the CS–g-nHA andCS–g-nHA–AL (Fig. 5c and d) exhibited the least porous structuresduring degradation, compared to the CS and CS–nHA scaffolds(Fig. 5a and b), which helps to explain the slower degradation pro-cess. The pores permit diffusion of the lysozyme into the matrixand with the collapse of the pores less area is exposed to enzymedegradation.

The toxicity of the degradation products from the scaffolds wereassessed by incubating the scaffolds with MC3T3 cell suspensionsfor a period of 72 h and studying them for possible cellular toxicity.In all the cases, the percentage cell viability values were found tobe higher (Fig. 5e), implying that the degradation products frompure CS and its derivative scaffolds were biocompatible and didnot impart significant level of toxicity.

3.5. Mechanical properties

Bone tissue engineering requires adequate mechanicalstrength to retain the initial shape of the scaffold at the implantsite. During fabrication of grafted scaffolds, our prime objectivewas to improve the mechanical strength of CS. The moduli ofCS and its derivative scaffolds are presented in Table 1. The mod-ulus of the CS–g-nHA scaffold was found to be increased by 23%in relation to pure CS. It can be assumed that the grafting of PPOonto CS contributed to the enhancement of the modulus by theincorporation of cross-linking points within the grafted scaffold,hence increasing the cross-link density of the scaffold. The ob-tained results are in close agreement with an earlier report, inwhich the effect of cross-linking on the mechanical propertiesof collagen scaffolds was studied [34]. This increase in moduluswas observed to be even higher for the CS–g-nHA–AL scaffold,

CS-g-nHA CS-g-nHA-AL

c d

100 µm 100 µm

48 60 72 (h)

CS CS-nHA CS-g-nHA CS-g-nHA-AL

HA–AL scaffolds after 28 days of in vitro enzymatic degradation, and (e) MTT assay

D. Depan et al. / Acta Biomaterialia 7 (2011) 2163–2175 2171

where the modulus was increased by 57%. The obtained mechan-ical properties of the grafted scaffolds indicate adequate mechan-ical strength and could serve as an ideal scaffold for boneregeneration. The high mechanical properties are due to the pres-ence of the carboxyl group (–COOH) of AL, which cross-links withthe amine (–NH2) group of CS.

3.6. Cell adhesion

As shown in Table 1, the numbers of attached cells in the CS–g-nHA and CS–g-nHA–AL scaffolds were significantly greaterthan in the pure CS scaffold. This demonstrates that the adhesiv-ity of osteoblasts is significantly higher on the CS–g-nHA and

Fig. 6. Fluorescence micrographs illustrating proliferation of pre-osteoblasts on CS, CS–nscaffolds at days 7 and 28, respectively. The fluorescence micrographs of cell–scaffold con

CS–g-nHA–AL scaffolds. Madihally and Matthew [35] reported thatthe cationic nature of CS allows electrostatic interactions with an-ionic glycoseaminoglycans, proteoglycans and other negativelycharged species. These ionic interactions may serve as a mecha-nism for retaining and recruiting cells, growth factors andcytokines in a tissue scaffold. Consequently, CS–g-nHA and CS–g-nHA–AL hybrid scaffolds exhibit potential as a practically relevantbiomaterial for tissue engineering scaffolds.

3.7. Osteoblasts integration and function

To corroborate the cell proliferation and distribution, the CS andits derivative scaffolds were stained with the nucleic acid dye

HA, CS–g-nHA and CS–g-nHA–AL scaffolds at similar locations (e.g. center) on purestructs were obtained after staining with nucleic acid staining dye (Hoechst 33342).

2172 D. Depan et al. / Acta Biomaterialia 7 (2011) 2163–2175

Hoechst 33342 and analyzed using fluorescence microscopy. Thecell nuclei of pre-osteoblasts cultured on CS and its derivative scaf-folds appear as blue spots after staining. The cells were presenteither individually or as aggregates on the scaffolds. Moreover,the proliferation and migration of cells were also apparentthroughout the porous structure on observing the saggital planeof CS and its derivative scaffolds (Fig. 6). The following can be sum-marized for Fig. 6:

(1) Cell density increased with incubation time for pure CS andits derivative scaffolds, and were observed to be embedded in thethin pore wall membrane and attached to the pores. The sampleson days 7 and 28 indicated that the cells attach predominantly inthe vicinity of pores and within the pores, which indicates thatthe scaffolds were completely covered with cells.

(2) After 7 or 28 days of incubation only a very small increase incell density was observed for pure CS and CS–nHA scaffolds, whilefor CS–g-nHA and CS–g-nHA–AL scaffolds the increase in cell pro-liferation was clearly evident, as shown in Fig. 6.

(3) There was an appreciable increase in cell density on CS–g-nHA and CS–g-nHA–AL scaffolds, implying greater cytocompati-bility of grafted scaffolds compared to pure CS scaffolds. Theaddition of AL to the CS–g-nHA scaffold further enhances theproliferation of cells.

Fig. 7. Scanning electron micrographs illustrating the morphology of pre-osteoblasts cell21 days, respectively.

The differences in cell proliferation can be explained on thebasis of the interconnectivity of pores. The lower interconnectiv-ity of CS restricts the medium flow containing cells to the super-ficial regions, and as a consequence more cells end up in the samepores of the superficial area. In contrast, the cells seeded onto CSderivative scaffolds predominantly attached to the inner surfaceof the pores with a homogeneous distribution, and only few cellsattached to the outmost surface, as shown in Fig. 6. The even celladhesion onto the inner pores in this case is believed to be theresult of identical medium flow through each repeating unit ofthe prototyped scaffold, whereas the high sessile flow existingon the outmost surface limits the cell adhesion. These results ob-tained from fluorescence microscopy provide an insight into theenhanced cytocompatibility and biological response of graftedCS scaffolds, in accordance with the interconnectivity and poros-ity of the grafted scaffolds.

The SEM micrographs in Fig. 7 show the cell attachment on pureCS and its derivative scaffolds after 1, 3, 7 and 21 days. It can beseen that the cell morphology was different on pure CS and itsderivative scaffolds. Furthermore, all the scaffolds were effectivein promoting cellular adhesion, spreading and proliferation, irre-spective of the type of scaffold. After 1 day of seeding, homoge-neously distributed viable cells can be seen along the polymericwalls of the scaffolds, while cell spreading becomes more evidentafter 7 days of culture, and a layer of cytoplasmic extensions can

s seeded on pure CS, CS–nHA, CS–g-nHA and CS–g-nHA–AL scaffolds after 1, 3, 7 and

D. Depan et al. / Acta Biomaterialia 7 (2011) 2163–2175 2173

be seen covering the walls of the scaffold pores. After 21 days ofcell culture, the morphology of the scaffold–cell construct resem-bled a sheet. The development of this sheet across the porous scaf-folds involves the clustering of cells, followed by bridge formationbetween the pore walls with subsequent formation of a multilay-ered structure. The CS derivative scaffolds exhibit better cellattachment and development of cytoplasmic processes in the formof numerous elongated pseudopodia, suggesting a high affinity to-wards pre-osteoblasts. Our results show that grafting of chitosanwith PPO and linkage with f-nHA increase the interconnected sur-face area such that cell adhesion and spreading is promoted.

The cells developed cytoplasmic processes with lengths rangingfrom about 20 to 50 lm that attached to the scaffold surfaces andinside the pores, as depicted in Fig. 8. These cytoplasmic extensionsare regions of the cell plasma membrane that contain a mesh orbundles of actin-containing microfilaments that permit the pene-trating cells to move along a substratum [36]. When the osteoblastcells are confined to a small architectural feature, like a pore, theytend to adopt a spherical morphology with cellular projectionsreaching out to the pore walls. Since, osteoblasts have a morethree-dimensional cuboidal morphology in vivo rather than a flat-tened one, this may be advantageous to the osteoblasts’ phenotypeand therefore bone formation.

The cell surface on the pure CS scaffold was relatively smoothcompared to its derivative scaffolds and can be distinguishedby a layer of small particles covering the cells grown on theCS–g-nHA and CS–g-nHA–AL scaffolds. These particles werecharacterized by EDS. The deposition of calcium phosphate ormineralization is very important from the viewpoint of the

Fig. 8. Scanning electron micrographs representing cell–cell interactions in a saggital secafter 7 days of cell culture. (c) and (d) are the saggital sections of CS–g-nHA–AL scaffold

appropriateness of the scaffolds [37]. The EDS pattern of osteo-blasts on the CS scaffold indicates little evidence of calcium andphosphorous after 7 days of cell culture, while they were presentto a significant extent on the derivative scaffolds as well as theextracellular matrix (Fig. 9). These results suggest that both theindividual cells and the cell clusters on the CS derivative scaffoldscontributed to the production of calcium and phosphate. Theosteoblast cells proliferated well on the fabricated scaffolds andformed a three-dimensional cell–scaffold construct in manner con-sistent with the proliferation process of bone regeneration, typi-cally described for pre-osteoblasts in tissue culture in vivo [38].From the above-mentioned results, we can conclude that theosteoconductivity of the fabricated scaffolds can be attributed tothe formation of a bone-like apatite layer on the surface, which en-ables the formation of a direct bond with the host bone and is aconsequence of a dynamic dissolution/precipitation process.

The results of a cell proliferation test carried out for 2 weeksafter osteoblasts cell were seeded onto each scaffold are presentedin Fig. 10. The number of cells increased significantly in 2 weekswhen cultured on the CS derivative scaffolds. This behavior canalso be related to superior cell adhesion. CS is not cytotoxic, butit does retard cell proliferation. This reduction in cell proliferationis potentially rectified by using a combination of CS with osteocon-ductive nHA and AL.

4. General conclusions

We describe here the successful synthesis of organic–inorganichybrid network structure nanocomposite scaffolds. Grafting of

tion of (a) CS–g-nHA scaffold, while the arrow in (b) shows pore-infiltration of cellsrepresenting cell spreading and pore infiltration, respectively.

Fig. 9. EDS spectra of the pre-osteoblasts seeded on (a) pure CS, (b) CS–nHA, (c) CS–g-nHA and (d) CS–g-nHA–AL scaffolds after 7 days of cell culture. Note the significantpresence of P and Ca.

2174 D. Depan et al. / Acta Biomaterialia 7 (2011) 2163–2175

chitosan and functionalization of nHA were used as an approach tofabricate three-dimensional scaffolds. The derivative scaffoldsexhibited greater modulus, controlled degradation rate and re-duced water uptake and retention ability compared to pure CSscaffolds. The grafting of chitosan involved hydroxypropylationby reaction with the –NH2 group of CS. Propylene oxide was cova-lently incorporated into the chitosan matrix, followed by grafting

0

0.2

0.4

0.6

0.8

1

1.2

1 3 7 9 14

Cel

l Den

sity

(OD

570)

Time (days)

MTT ASSAY CS

CS-nHA

CS-g-nHA

CS-g-nHA-AL

Fig. 10. Viability of pre-osteoblasts seeded on pure CS and its derivative scaffolds asa function of time, measured by MTT assay, which represents the activemitochondrial activity of living cells. Cell density was significantly higher on CS–g-HAP and CS–g-HAP–AL scaffolds on day 7 onwards compared to pure CS and CS–HAP scaffolds (p < 0.05) (n = 5).

of f-nHA, rendering a high content of –OH functionality for furtherphysio-chemical modifications. The organic–inorganic hybridnanocomposites can be considered as potential biomaterials fortissue engineering applications because of their significantly supe-rior physico-chemical, mechanical and biological responses overchitosan or chitosan–nHA scaffolds.

Acknowledgement

The authors acknowledge support and funding from the Centerfor Structural and Functional Materials, University of Louisiana atLafayette, USA.

Appendix A. Figures with essential colour discrimination

Certain figures in this article, particularly Figures 3 and 6, aredifficult to interpret in black and white. The full colour imagescan be found in the on-line version, at doi:10.1016/j.actbio.2011.01.029).

References

[1] Pillai CKS, Paul W, Sharma CP. Progress Polym Sci 2009;34(7):641–78.[2] Larkin AC, Davis RM, Rajagopalan P. Biomacromolecules doi:10.1021/

bm100867h.[3] Amsden BG, Sukarto A, Knight DK, Shapka SN. Biomacromolecules

2007;8(12):3758–66.[4] Mourya VK, Inamdar NN. React Funct Polym 2008;68:1013–51.[5] Depan D, Pratheep AK, Singh RP. Acta Biomater 2009;5:93–100.[6] Liang YY, Zhang LM. Biomacromolecules 2007;8(5):1480–6.[7] Nicholas B, Stephane G, Veronique C. Biomacromolecules 2010;11(1):88–96.[8] Yan S, Changsheng L, Yuan Y, Xinyi T, Fan Y, Xiaoqian S, et al. Biomaterials

2009;30(12):2340–8.

D. Depan et al. / Acta Biomaterialia 7 (2011) 2163–2175 2175

[9] Guak-Kim T, Donna LMD, Lauren NB, Justin JCW. Biomaterials2010;31(23):6104–18.

[10] Wang L, Stegemann JP. Biomaterials 2010;31:3976–85.[11] Seol YJ, Lee JY, Park YJ, Lee YM, Young K, Rhyu IC. Biotechnol Lett

2004;26:1037–41.[12] Li J, Chen Y, Yin Y, Yao F, Yao K. Biomaterials 2007;28:781–90.[13] Muzzarelli RAA. Carbohydr. Polym. 2011;83:1433–45.[14] Han E, Shan D, Xue H, Cosnier S. Biomacromolecules 2007;8:971–5.[15] Rusu VM, Ng CH, Wilke M, Tiersch B, Fratzl P, Peter MG. Biomaterials

2005;26(26):5414–26.[16] Kim SB, Kim YJ, Yoon TL, Park SA, Cho IH, Kim EJ, et al. Biomaterials

2004;25:5715–23.[17] Kong L, Gao Y, Cao W, Gong Y, Zhao N, Zhang X. J Biomed Mater Res

2005;75A:275–82.[18] Moreau JL, Xu HHK. Biomaterials 2009;30(14):2675–82.[19] Manjubala I, Scheler S, Bossert J, Jandt KD. Acta Biomater 2006;2(1):75–84.[20] Lee HJ, Choi HW, Kim KJ, Lee S. Chem Mater 2006;18:5111–8.[21] Wang YC, Lin MC, Wang DM, Hsieh HJ. Biomaterials 2003;24:1047–57.[22] Park JB, Lakes RS. Biomaterials: an introduction. New York: Plenum Press;

1992.

[23] Dhiman HK, Ray AR, Panda AK. Biomaterials 2004;25:5147–54.[24] Shi X, Sitharam B, Pham Q-P, Spicer PP, Hudson JL, Wilson LJ, et al. J Biomed

Mater Res 2008;86A:813–23.[25] Sigomoto M, Morimoto M, Sashiwa H, Saimoto H, Shigemasa Y. Carbohydr

Polym 1998;36:49–59.[26] Peng Y, Han B, Liu W, Liu X. Carbohydr Res 2005;340:1846–51.[27] Wang L, Khor E, Lim LY. J Pharm Sci 2001;90:1134–42.[28] Shoichet MS. Macromol 2010;43(2):581–91.[29] Yin Y, Ye F, Cui J, Zhang F, Li X, Yao K. J. Biomed. Mater. Res. 2003;24:844–55.[30] Ma L, Gao C, Mao Z, Zhou J, Shen J, Hu X, et al. Biomaterials 2003;24:4833–41.[31] Darnell J, Lodish H, Baltimore D. Molecular Cell Biology. vol. 41, 2nd ed. New

York: Freeman; 1990. pp. 1118–1122.[32] Mi F, Shyu S, Wu Y, Lee S, Shyong J, Huang R. Biomaterials 2001;22:165–73.[33] Murugan R, Ramakrishna S. Biomaterials 2004;25:3829–35.[34] Harley BA, Leung JH, Silva ECCM, Gibson LJ. Acta Biomater 2007;3:463–74.[35] Madihally SV, Matthew HWT. Biomaterials 1999;20:1133–42.[36] Dalby MJ, Gadegaard N, Curtis AS, Oreffo RO. Curr Stem Cell Res Ther

2007;2:129–38.[37] Vachoud L, Zydowicz N, Domard A. Carbohydr Res 2000;326:295–304.[38] Stein GS, Lian JB, Owen TA. FASEB J 1990;4:3111–8.