lable at ScienceDirect

Biomaterials 32 (2011) 2489e2499

Contents lists avai

Biomaterials

journal homepage: www.elsevier .com/locate/biomater ia ls

Autologous extracellular matrix scaffolds for tissue engineering

Hongxu Lu a,b, Takashi Hoshiba a, Naoki Kawazoe a,c, Guoping Chen a,b,c,*

aBiomaterials Center, National Institute for Materials Science, 1-1 Namiki, Tsukuba, Ibaraki 305-0044, JapanbGraduate School of Pure and Applied Sciences, University of Tsukuba, 1-1-1 Tennodai, Tsukuba, Ibaraki 305-8577, Japanc International Center for Materials Nanoarchitectonics (MANA), National Institute for Materials Science, 1-1 Namiki, Tsukuba, Ibaraki 305-0044, Japan

a r t i c l e i n f o

Article history:Received 22 November 2010Accepted 10 December 2010Available online 5 January 2011

Keywords:Autologous extracellular matrix scaffoldBiomimeticsTissue engineeringDecellularization

* Corresponding author. Biomaterials Center, NatScience, 1-1 Namiki, Tsukuba, Ibaraki 305-0044, Japanþ81 29 860 4714.

E-mail addresses: guoping.chen@nims.go.jp, guopi

0142-9612/$ e see front matter � 2010 Elsevier Ltd.doi:10.1016/j.biomaterials.2010.12.016

a b s t r a c t

Development of autologous scaffolds has been highly desired for implantation without eliciting adverseinflammatory and immune responses. However, it has been difficult to obtain autologous scaffolds bytissue decellularization because of the restricted availability of autologous donor tissues from a patient.Here we report a method to prepare autologous extracellular matrix (aECM) scaffolds by combiningculture of autologous cells in a three-dimensional template, decellularization, and template removal. TheaECM scaffolds showed excellent biocompatibility when implanted. We anticipate that “Full AutologousTissue Engineering” can be realized to minimize undesirable host tissue responses by culturing thepatient’s own cells in an aECM scaffold.

� 2010 Elsevier Ltd. All rights reserved.

1. Introduction

Tissue engineering and regenerativemedicine have been rapidlydeveloped as an encouraging alternative practice to restore orreplace lost or malfunctioning tissues and organs through the useof cells and biomaterial scaffolds [1e3]. In a common approach, thecells isolated from a patient are cultured in a biocompatible three-dimensional (3D) porous scaffold supplemented with growthfactors to regenerate new tissues or organs [4,5]. The scaffoldprovides necessary interim support for cell adhesion, proliferation,and phenotypic differentiation; offers biochemical and biophysicalcues to modulate the neo-tissue formation by mimicking thefunctional and structural characteristics of the native ECM [6,7],which play a crucial role in controlling and regulating cell behaviorand function [8,9].

Thus far, scaffolds have been developed from synthetic biode-gradable polymers such as polylactide (PLA), polyglycolide (PGA),and poly (lactide-co-glycolide) (PLGA) [10]; natural polymers suchas collagen, chondroitin, and hyaluronic acid [10,11]; and acellularmatrices derived from decellularized tissues and organs [12e15].However, the synthetic polymers are limited by their biologicalinertness [16] and the acidic moieties, residual catalysts, andmicroscale particulates that accompany degradation [17]. On the

ional Institute for Materials. Tel.: þ81 29 860 4496; fax:

ng-chen@aist.go.jp (G. Chen).

All rights reserved.

other hand, although the naturally derived polymer and acellularmatrices can provide abundant biological signals and degrade intophysiologically tolerable compounds [18], they are exhaustivelyxenogeneic or allogeneic. This situation adds potential risks ofpathogen transmission [19,20], and provocation of undesirableinflammatory and immunological reaction, leading to unfulfilledresults from the regenerated tissues and organs [21e26].

To avoid these problems, the autologous extracellular matrix(aECM) scaffold should be a safe and reliable biomaterial candidate[27]. The development of aECM scaffolds has been strongly antici-pated for use in tissue engineering and regenerative medicine[28,29]. The use of both autologous cells and autologous scaffoldswould eliminate negative host responses and lead to optimal tissueregeneration. However, the availability of autologous sources ofdonor tissues and organs is highly limited. It has been almostimpossible to use such acellular autologous matrices for tissueengineering. The ECM secreted by autologous cells would bea potential alternate to acellular autologous tissues and organsbecause some autologous cells can be expanded in vitro andmaintained under a pathogen-free condition.

In this study, we developed amethod of preparing ECM scaffoldsby the 3D culture of cells in a selectively removable template(Fig. 1). The intracellular components and the biodegradable poly-mer template were selectively removed after cell culture to obtainthe ECM scaffolds. To confirm the effectiveness of this method,three cell types, human bone marrow mesenchymal stem cells(MSCs), normal human articular chondrocytes (NHAC), and normalhuman dermal fibroblasts (NHDF) were used to prepare theirrespective ECM scaffolds. If autologous cells were used, the method

Fig. 1. Preparation scheme for ECM scaffolds.

H. Lu et al. / Biomaterials 32 (2011) 2489e24992490

could be used to prepare aECM scaffolds. To demonstrate thebiocompatibility of the aECM scaffolds, we implanted mousefibroblast-derived ECM scaffolds (ECM-mF) into ICR (Clrj:CD1)miceto evaluate the host tissue responses.

2. Materials and methods

2.1. ECM scaffold fabrication and characterization

MSC and NHAC, both at passage 2, were obtained from Lonza (Walkersville, MD).NHDF (derived from neonatal foreskin) were purchased from Cascade Biologics(Invitrogen, Portland, OR). The cells were seeded in 75 cm2 tissue culture flasks andcultured using their respective proliferation media under an atmosphere of 5% CO2

at 37 �C. MSC were cultured in MSC basal medium with MSC growth supplement;NHACwere cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplementedwith 10% fetal bovine serum; NHDF were cultured in Medium 106 with low serumgrowth supplement. The details were listed in Supplementary Tables 1 and 2. Thecells were further subcultured thrice after confluence. Cells at passage 5 werecollected by treatment with trypsin/EDTA solution and re-suspended in themediumfor scaffold fabrication.

A knitted PLGA mesh (Vicryl polyglactin 910, Johnson & Johnson, Somerville, NJ)was used as a template to fabricate the ECM scaffolds. The mesh was cut into smalldiscs with a diameter of 10.4 mm. Cells at passage 5 were seeded in discs of knittedPLGAmesh to form cell-PLGA constructs. The cell densities of MSC, NHAC, and NHDFwere 3 � 105, 5 � 105, and 5 � 105 cells/ml, respectively; 200 ml cell suspension wasseeded on one side of each PLGA disc. Glass rings (inner Ø ¼ 10 mm, outerØ ¼ 12 mm) were covered on PLGA discs to prevent cell leakage during cell seeding.After culture for 6 h, the cell-seeded PLGA mesh discs were turned over and theother sides of the mesh discs were also seeded with the same number of cells. Thecell-ECM-PLGA constructs were formed by culturing cell-PLGA constructs in DMEMsupplemented with 10% fetal bovine serum (FBS), 4 mM L-glutamine, 0.4 mM L-proline, 0.1 mM MEM non-essential amino acid, 1 mM sodium pyruvate, 100 U/mLpenicillin, and 100 mg/mL streptomycin. Ascorbic acid (50 mg/ml) and ascorbate-2-phosphate (150 mg/ml) (Wako, Tokyo, Japan) were added into the culturemedium tostimulate ECM deposition. A mixture of ascorbic acid and ascorbic acid phosphateprovides a constant concentration of ascorbate in the culture medium, and has beenproved nontoxic to cells and stimulated cell proliferation and ECM deposition[30,31].

After being cultured for 5 days (MSC) or 6 days (NHAC and NHDF), the cell-ECM-PLGA constructs were washed with phosphate buffered saline (PBS) and MilliQwater (Millipore, Bedford, MA). Freeze-thaw cycling plus NH4OH aqueous solutiontreatment was used for the decellularization. The cell-PLGA constructs were frozenat �80 �C in MilliQ water for 3 h, thawed at room temperature, and washed withMilliQ water. The freeze-thaw cycles were repeated 6 times. The constructs werethen immersed in an aqueous solution of ammonia (25 mM) for 20 min on a slowlymoving four-way shaker. After the ammonia solution treatment, the constructs werewashed with MilliQ water 6 times. The constructs were immersed in a 0.5 M aqueoussolution of trisodium phosphate and incubated at 37 �C for 48 h to remove the PLGAtemplate. The ECM scaffolds were obtained after being washed with MilliQ water.

Finally, the ECM scaffolds were freeze-dried for 24 h on an EYELA FDU-2200 freeze-drier (Tokyo Rikakikai, Tokyo, Japan) and sterilized with ethylene oxide gas (EOG)using an Eogelk SA-1000 sterilizer (Elk Corp., Osaka, Japan) for cell culture use.

2.2. Characterization of ECM scaffolds

The ECM scaffolds were coated with Pt and their porous structures wereobserved by a scanning electron microscope (SEM) (JSM-5610; JEOL, Tokyo, Japan).The removal of PLGA was examined by ATR-FTIR on an IR Prestige-1 FTIR spectro-photometer (Shimadzu, Kyoto, Japan). To evaluate the removal of cellular compo-nents after decellularization, the ECM scaffolds were stained with Hoechst 33258(for cell nuclei), Alexa Fluro 488-labeled phalloidin (for F-actin), and DiI (for cellmembrane). The cell-ECM constructs without decellularization and PLGA were alsostained as controls. The fluorescent staining was observed under an Olympus BX51fluorescence microscope with a DP-70 CCD camera (Olympus, Tokyo, Japan). Theimages were manipulated by DP controller software.

Immunofluorescence staining was performed to confirm the existence of ECMcomponents in the scaffolds, including type I collagen, type II collagen, type IIIcollagen, fibronectin, vitronectin, laminin, aggrecan, versican, decorin, and biglycan.The antibodies are listed in Supplementary Table 3. Briefly, the ECM scaffolds wereincubated with 2% BSA/PBS for 30 min for blocking. The ECM scaffolds weresequentially incubated with the first antibodies for 1 h and second antibodies for40 min at 37 �C in dark. Then the ECM scaffolds were mounted with Vectashield�

mounting medium (Vector, Burlingame, CA) and observed.For DNA and GAG analysis, the ECM scaffolds were digestedwith papain solution

at 60 �C for 6 h. The cell-ECM-template constructs before decellularization andtemplate removal were digested in papain solution as controls. Papain (Sigma, St.Louis, MO) was dissolved at 400 mg/ml in 0.1 M phosphate buffer (pH 6.0), with 5 mM

cysteine hydrochloride, and 5 mM EDTA. The lysates were used for detection of theDNA and GAG amount. A Hoechst 33258-based DNA quantification kit (Sigma) wasused to measure the DNA. Five mL lysate was added to 2 ml Hoechst 33258 solutionand the fluorescence intensity was recorded by a FP-6500 spectrofluorometer(JASCO, Tokyo, Japan). The excitation wavelength was set at 360 nm and the emis-sion wavelength was set at 460 nm. The DNA amount was calculated based ona standard curve obtained with the standard DNA supplied with the kit (n ¼ 3). Thesulfated glycosaminoglycan (GAG) amounts were measured by a Blyscan� sGAGassay kit (Biocolor, Newtonabbey, UK) according to the manufacturer’s manual. Thespecimen lysate was mixed with Blyscan� dye to bind the GAG. The GAG-dyecomplex was then collected by centrifugation. After the supernatant was removedand the tube drained, the dissociation reagent was added. Then 200 ml solution wastransferred into a 96-well plate. Absorbance against the background control wasobtained at a wavelength of 656 nm on a Benchmark Plus� microplate spectro-photometer (Bio-Rad, Tokyo, Japan) and the GAG amount (n ¼ 3) was calculatedbased on a standard curve obtained with the standard GAG supplied with the kit.

Collagen contents in the ECM scaffolds were determined using the Sircolcollagen assay (Biocolor) according to the manufacturer’s instructions. In brief, ECMscaffolds were incubated for 48 h at 48 �C in 0.5 M acetic acid containing 0.1 mg/mlpepsin. The samples were added to Sircol dye reagent, and collagen-dye complexesformed and precipitated out from the soluble unbound dye. After centrifugation, thepellet was washed once with Acid-Salt Wash Reagent and suspended in alkali

Fig. 2. SEM images of PLGA mesh (a), MSC-ECM-PLGA constructs (b), NHAC-ECM-PLGA constructs (c) and NHDF-ECM-PLGA constructs (d). MSC were cultured in the PLGA disc for 5days; NHAC and NHDF were cultured for 6 days. Scale bar, 200 mm.

Fig. 3. Gross appearance of a mesh-like ECM-M (a), ECM-C (b), and ECM-F (c) and SEM images of the ECM scaffolds (dei). (gei), High magnification of (d), (e), and (f), respectively.Scale bar ¼ 500 mm in (def) and 50 mm in (gei).

H. Lu et al. / Biomaterials 32 (2011) 2489e2499 2491

H. Lu et al. / Biomaterials 32 (2011) 2489e24992492

reagent. The solution (200 ml) was transferred to a 96-well plate and the absorbancewas read at 555 nm. The amount of collagen (n ¼ 3) was calculated based ona standard curve obtained with the standard bovine type I collagen supplied withthe kit.

2.3. Fabrication of ECM scaffolds from mouse fibroblasts and in vivo implantation

The animal experiment was conducted according to the committee guidelines ofthe National Institute for Materials Science of Japan for Animal Experiments.Surgical plane anesthesia in each animal was induced and maintained with 2%isoflurane in oxygen. Mouse fibroblasts (mF) were isolated from the biopsies of6-week-old ICR (Clrj:CD1) mice and expanded in vitro according to the protocoldescribed by A. Takashima [32]. P2 mouse fibroblasts were seeded (5 � 105 cells/ml,200 ml/side) and cultured in PGLA meshes for 10 days. ECM-mF scaffolds werefabricated by culturing the mouse fibroblasts in the PLGA mesh using the samemethod described above.

Thirty mice were randomly divided into 5 groups. The aECM groups wereimplanted with aECM derived from autologous cells. The Allo groups wereimplanted with ECM scaffolds derived from the allogeneic cells. After being sterilelycut into Ø 8mm discs, bovine collagen sponge (BCS) (Koken, Tokyo, Japan) and PLGAknitted mesh (Johnson & Johnson) were separately implanted into the mice. Mice

Fig. 4. Fluorescence staining of F-actin, cell nuclei, and cell membrane for the confirmation oconstruct, and ECM-M (b). Scale bars ¼ 100 mm in (a).

that were opened but without implantation (sham operation) were used as controls.The central back was shaved and opened for implantation in a sterile fashion. Thematerials were subcutaneously implanted in the dorsum of each mouse. Theopenings were then sutured with nylon sutures.

2.4. Evaluation of host responses

One week after implantation, the mice were scarified. The back skin and thesample-harboring tissues were excised and fixed in 10% neutral buffered formalinsolution. The specimens were trimmed and embedded in paraffin, and thensectioned at 7-mm intervals. The tissue sections were stained with hematoxylin andeosin. Because the ECM scaffolds were invisible after implantation (SupplementaryFig. 1), we excised large enough areas to ensure that the scaffolds were held in thespecimens. During sectioning, each sample was labeled for several domains andrepresentative slices of each domain to locate the implanted materials. The sectionsin positive domains were used for further exploration.

Immunocytochemical staining was performed to identify immunocytes by usinganti-F4/80, anti-Gr-1, anti-CD3, anti-MHC II monoclonal antibodies witha Vectastain� Elite� ABC kit (Vector, Burlingame, CA). The detailed information ofantibodies was listed in Supplementary Table 4. Briefly, the slides were deparaffi-nized and hydrated. Then a heat-mediated antigen retrieval technique that included

f decellularization in ECM-M (a) and ATR-FTIR spectra of PLGA template, cell-ECM-PLGA

Table 1DNA contents (ng/mesh) in the cell-ECM-PLGA constructs and the ECM scaffolds.

Cell-ECM-PLGA constructs ECM

ECM-M 670.11 � 51.54a 4.38 � 1.67ECM-C 2677.89 � 511.35a 2.79 � 0.63ECM-F 2698.33 � 567.10a 4.35 � 0.81

Data represent means � S.D. (n ¼ 3).a Significant difference compared with ECM scaffold, P < 0.001.

Table 2sGAG (ng/mesh) contents in the cell-ECM-PLGA constructs and the ECM scaffolds.

Cell-ECM-PLGA construct ECM

ECM-M 916.85 � 41.39a 54.72 � 31.79ECM-C 1937.41 � 104.57a 61.21 � 26.74ECM-F 2436.67 � 120.38a 51.93 � 10.89

Data represent means � S.D. (n ¼ 3).a Significant difference compared with ECM scaffold, P < 0.001.

H. Lu et al. / Biomaterials 32 (2011) 2489e2499 2493

20 min incubation at 95 �C in 0.01 M citrate buffer, pH 6.0, was first used. Aftercooling down, two separate 5minwashes in Tris-buffered saline (TBS)/Tween 20, pH7.4, followed by one wash in PBS were performed. Sequentially, 0.3% H2O2 in waterwas used to quench the endogenous peroxidase and 2% BSA/PBS was used for non-specific antigen blocking. The tissue sections were incubated with the first anti-bodies for 30min followed by a 30min’ incubation of biotinylated second antibodies(1:200 in PBS containing 10% serum). Then the ABC reagent was added for 30 minand 3, 30 diaminobenzidine (DAB) chromogen solutionwas used to develop the colorfor 5 min. Finally, the slides were counterstained with hematoxylin andmounted forobservation. All incubation was conducted at room temperature (RT). The macro-phage numbers were counted by a cell counter plugin of ImageJ 1.6. Three views, ata magnification of 400�, were randomly selected to count the positively stained andtotal cell numbers for each sample. Each group had three parallel samples (n ¼ 3).

To further investigate the host tissue response to the ECM scaffolds, theexpressions of genes encoding immune relative cytokines were analyzed. Theconnective tissues surrounding the materials were isolated and washed with PBS.They were frozen in liquid nitrogen and crushed into powder after the washing byPBS. The powders were dissolved in 1 ml Isogen (Nippon Gene, Toyama, Japan) andthe total RNA was extracted following the manufacturer’s protocol. One microgramtotal RNA was reversely transcribed into cDNA using random hexamer primers(Applied Biosystems, Forster City, CA) in 20 mL reaction on a GeneAmp PCR Systems9700 (Applied Biosystems). An aliquot (1 mL) of 10-times diluted reaction solutionwas used for each 25 mL real-time PCR reaction together with 300 nM forward andreverse primers, 150 nM probes, and qPCR Master Mix (Eurogentec, Seraing,Belgium). Real-time PCR analysis was conducted and monitored using a 7500 Real-Time PCR System (Applied Biosystems). After an initial incubation step of 2 min at50 �C and denaturation for 10 min at 95 �C, 40 cycles (95 �C for 15 s, 60 �C for 1 min)of PCR were performed. Reactions were performed in triplicate. The expressionlevels of Gapdh were used as endogenous controls and gene expression levelsrelative to 18S rRNA were calculated using the comparative Ct method. Threesamples under each condition were used for measurement to calculate the meansand standard deviations (n ¼ 3). TaqMan Gene Expression assay kits (Applied Bio-systems) were used for transcription levels of tumor necrosis factor-alpha (Tnf,reference sequence NM_013693.2), interleukin-2 (Il2, reference sequenceNM_008366.2), interleukin-4 (Il4, reference sequence NM_021283.1), interleukin-10(Il10, reference sequence NM_010548.1), GAPDH (Gapdh, reference sequenceNM_008084.2), and 18S rRNA (reference sequence X03205.1).

2.5. Statistical analysis

All results are presented as the mean � standard deviation (SD). A two-tailedt e test was used to examine the statistical differences of DNA, sGAG, and soluble

Fig. 5. Immunofluorescence images of ECM mole

collagen contents between decellularized and un-decellularized samples. A one-wayanalysis of variance (ANOVA) was used to analyze the statistical differences in hosttissue response. If the overall ANOVA test was significant, a Tukey’s post hoc test wasperformed for a pairwise comparison. A P-value <0.05 was considered statisticallysignificant.

3. Results and discussion

3.1. Fabrication and characterization of ECM scaffold

Bone marrow mesenchymal stem cells (MSCs), dermal fibro-blasts, and articular chondrocytes, were used to prepare theirrespective ECM scaffolds. MSCs, fibroblasts, and chondrocytes wereseeded and cultured in templates of PLGA knitted mesh (Fig. 2a),which were used as temporary skeletons. The cells adhered,proliferated, and secreted ECM in the templates. Fig. 2bed show thedistribution of cells and their secreted ECM in the PLGA mesh afterculture for several days. After culture for 5 (MSC) or 6 (NHAC andNHDF) days, the cellular components were removed by the decel-lularization method of freeze-thaw cycling plus the treatment withammonium hydroxide. The PLGA mesh templates were selectivelyremoved by the treatment of immersion in 0.5 M Na3PO4 aqueoussolution at 37 �C for 48 h. The freeze-thaw cycling causes cryoin-juries to cells through osmotic response, extracellular and intra-cellular ice formation [33]. After repeated freeze-thaw cycling, cellswere broken and released intracelluar components. After NH4OH isadded, the connections between intracellular and extracellularmolecules will be broken. Therefore, cellular debris could beremoved after complete washing. The weak alkaline environmentof 0.5 M Na3PO4 could also contribute to the removal of cellularcomponents (such as DNA). The resulting ECM scaffolds preparedfrom MSC, NHAC, and NHDF were designated as ECM-M, ECM-C,and ECM-F, respectively.

The ECM scaffolds exhibited a mesh-like appearance similar tothat of the temporary skeletal PLGA knitted mesh (Fig. 3aec). The

cules composing ECM-M. Scale bar, 100 mm.

Table 3Soluble collagen contents (mg/mesh) in the cell-ECM-template constructs and theECM scaffolds.

Cell-ECM-PLGA construct ECM

ECM-M 5.83 � 0.88a 2.78 � 0.34ECM-C 4.34 � 0.39a 2.01 � 0.44ECM-F 3.43 � 0.51a 2.04 � 0.49

Data represent means � S.D. (n ¼ 3).a Significant difference compared with ECM scaffold, P < 0.05.

H. Lu et al. / Biomaterials 32 (2011) 2489e24992494

porous 3D structure, microscale and nanoscale ECM fibers, wasobserved by a scanning electron microscope (SEM) (Fig. 3dei).However, there was no obvious difference among ECM-M, ECM-C,and ECM-F in terms of morphology. The geometrical properties,porosity, interconnectivity, and nanoscaled fibrous structure aremeant to support cell proliferation and differentiation, and benefittissue regeneration [34,35].

The cell nuclei, cell membrane, and F-actinwere removed by thedecellularization treatment (Fig. 4a, and Supplementary Fig. 2a andb). Most of the DNA was removed and only a very small amountremained in the scaffolds. The amount of DNA in ECM-M, ECM-C,

Fig. 6. ECM scaffolds derived frommouse fibroblasts. (a) Gross appearance of a mesh-like ECrevealed that ECM-mF was free of cellular components. Scale bar, 100 mm (d) ATR-FTIR spectmouse cell-ECM-PLGA constructs and ECM-mF. Data represent means � SD. ***,significant

and ECM-F was 0.65%, 0.10%, and 0.16% of the values before decel-lularization, respectively (Table 1). AIR-FTIR spectra verified thecomplete removal of the PLGA templates from the ECM scaffoldsbecause the typical band for the ester carbonyl stretch at 1740 cm�1

disappeared in the ECM scaffolds but was present in the cell-ECM-PLGA complexes (Fig. 4b, and Supplementary Fig. 2c and d). ThePLGA template was removed because its degradation mightprovoke adverse inflammation. Decellularization could facilitatethe removal of the PLGA template and avoid undesired effects suchas calcification induced by the cellular components of DNA andlipids [36].

The compositional biomolecules in ECM-M, ECM-C, and ECM-Fwere analyzed by immunohistological staining. There were somedifferences among the components of the ECM-M, ECM-C, andECM-F (Supplementary Table 5). ECM-M consisted of type Icollagen, type III collagen, fibronectin, vitronectin, laminin, aggre-can, decorin, and biglycan (Fig. 5). ECM-C consisted of type Icollagen, type III collagen, fibronectin, vitronectin, laminin, aggre-can, versican, decorin, and biglycan (Supplementary Fig. 3a). ECM-Fconsisted of type I collagen, type III collagen, fibronectin, vitro-nectin, laminin, decorin, and biglycan (Supplementary Fig. 3b).

M-mF. Scale bar, 2 mm (b) SEM image of the ECM-mF. Scale bar, 500 mm (c) HE stainingra showed the PLGA template was removed from the ECM-mF. (e) DNA contents in thedifference, P < 0.001.

H. Lu et al. / Biomaterials 32 (2011) 2489e2499 2495

Therefore, the composition of the ECM scaffolds depends on the celltype and cell phenotype used to prepare the scaffolds. Quantitativeanalysis of sulfated glycosaminoglycans (sGAG) showed that sGAGremained in the ECM scaffolds (Table 2). However, the sGAGcontents decreased after decellularization and template removal.The amount of soluble collagen also decreased after decellulariza-tion and PLGA removal (Table 3). These results indicate that thedecellularization and PLGA removal processes partially removedsome of the extracellular matrices.

The combination of in vitro three-dimensional cell culture ina temporaryskeleton and selective removal of the template provides

Fig. 7. HE staining of the implanted materials and their surrounding tissues. (b) is the high mbovine collagen sponge; Mu, panniculus carnosus muscle; arrows and areas within the two

a useful method for preparing ECM scaffolds. The ECM scaffoldsshowed a subtle porous structure constructed by various ECMbiomolecules. Cellular components and PLGA template wereremoved by decellularization and Na3PO4 treatments. The compo-sitions depended on the cell sources used to fabricate ECM scaffolds.Other thanMSCs, chondrocytes andfibroblasts, a large variety of celltypes, are thought to be available for preparing their respective ECMscaffolds by the present method. A combination of different cellsmight offer a better recapitulation of the tissue (organ)-specific ECM[37] and satisfy the complexity requirements of tissue engineeringand regenerative medicine [38].

agnification of (a). aECM, autologous ECM scaffold; Allo, allogeneic ECM scaffold; BCS,dashed lines indicate implanted materials. Scale bar ¼ 200 mm in (a) and 100 mm in (b).

H. Lu et al. / Biomaterials 32 (2011) 2489e24992496

3.2. Host tissue responses to aECM scaffold

Mouse autologous ECM (aECM-mF) scaffolds were preparedusing autologous mouse fibroblasts, which were isolated frombiopsies of each mouse. The ECM-mF scaffolds showed mesh-likethree-dimensional structures (Fig. 6a and b). The removal of thecellular components and PLGA templates was confirmed (Fig. 6cand d). The amount of DNA in the ECM-mF was 0.42% of the valuebefore decellularization (Fig. 6e). They primarily consisted of type Icollagen, type III collagen, fibronectin, vitronectin, laminin, decorin,and biglycan (Supplementary Fig. 4).

The aECM-mF scaffolds were implanted into a subcutaneouspocket of the respective ICR mouse where the fibroblasts wereisolated. The host tissue responses to the aECM scaffolds werecompared with those to allogeneic ECM scaffolds (ECM scaffoldsprepared from allogeneic mouse cells), xenogeneic bovine collagensponge (BCS), and PLGA mesh. After implantation for one week,

Fig. 8. Micrographs of immunocytochemically stained macrophages (a) and macrophage/toECM scaffold; BCS, bovine collagen sponge; Mu, panniculus carnosus muscle; arrows and arData represent mean � SD (n ¼ 3). ***,significant difference, P < 0.001.

the materials were harvested with the surrounding tissues(Supplementary Fig. 1) for analysis of inflammatory and immunehost responses. From the hematoxylin and eosin (HE) staining(Fig. 7a and b), it was found that the host cells had penetrated intothe aECM-mF and allogeneic ECM-mF scaffolds. No dense fibrouslayers surrounding the implanted materials were observed.However, the BCS and PLGA were “wrapped” by dense fibrouslayers without being remodeled [39].

Immunocytochemical staining of the 1-week implants was per-formed to investigate cellular responses in the host tissue reactions.Themacrophages in the implantswere stained andcounted (Fig. 8a).The sham operation group had very few macrophages. Among thefour implantedmaterials, the lowest number ofmacrophageswas inthe aECM-mF scaffolds. The percentages of macrophages/total cellsin different groups were compared (Fig. 8b). The lowest proportionof macrophages was found in the aECM group. The macrophagepercentageof the allogeneic groupwas significantlyhigher than that

tal cell percentages in each group (b). aECM, autologous ECM scaffold; Allo, allogeneiceas within the two dashed lines indicate implanted materials. Scale bar ¼ 100 mm (b),

H. Lu et al. / Biomaterials 32 (2011) 2489e2499 2497

of the aECM scaffolds (P < 0.001). The macrophage percentageselicited by the BCS and PLGA scaffoldswere significantly higher thanwere those of the ECM scaffolds (P< 0.001). Some neutrophils wereobserved only in the PLGA, BCS, and allogeneic groups (Fig. 9a).When biomaterials are implanted, neutrophils and macrophageshave been reported to be involved in the inflammatory responses atthe implantation site. Neutrophils serve to remove foreignmaterialsand trigger other host responses by secreting factors that summonother immunocytes [23]. And macrophages become activated andact as themainmediators of the host tissue responses. The excretionof soluble mediators by macrophages can influence the behavior of

Fig. 9. Neutrophils (a) and MHC II-presenting cells (b) in the implanted materials and suscaffold; Allo, allogeneic ECM scaffold; BCS, bovine collagen sponge; Mu, panniculus carnosucircles indicate positive cells. Scale bar ¼ 100 mm.

other leukocytes [22,40]. The absence of neutrophils and lowestmacrophage percentage indicate aECM-mF induced the lowestinflammatory responses. A fewMHCclass II antigen-presenting cellswere observed in the allogeneic ECM-mF and BCS implantationgroups (Fig. 9b). These immunogenic antigen-presenting cells aresupposed to exacerbate inflammation and immunological reactions.No T cells were detected in any of the groups after 1-weekimplantation.

We further investigated the cytokine profiles, which are the keymediators in the host tissue response [22]. The expressions of genesencoding interleukin-10 (Il10), interleukin-2 (Il2), interleukin-4

rrounding tissues revealed by immunocytochemical staining. aECM, autologous ECMs muscle; arrows and areas within the two dashed lines indicate implanted materials;

Fig. 10. Il10 (a) and Tnf (b) expression analyzed by real-time RT PCR. Data representmean � SD (n ¼ 3). *significant difference, P < 0.05.

H. Lu et al. / Biomaterials 32 (2011) 2489e24992498

(Il4), and tumor necrosis factor-a (Tnf) in the connective tissuesunder the 1-week implanted materials were analyzed by real-timeRT-PCR. The expression of Il10 was lower in the aECM and controlgroups than it was in the allogeneic ECM, BSC, and PLGA groups(Fig.10a). Among the samples, therewas no significant difference inthe transcription level of Tnf although it was lower in the aECMgroup (Fig. 10b). No Il2 and Il4 gene expressions were detected.These results indicate aECM-mF induced minimal cytokine medi-ation to modulate the host responses.

All of the results from the host tissue response analyses indicatedexcellent biocompatibility of the aECM scaffolds, which is essentialfor ideal tissue engineering scaffolds [17]. Furthermore, the use ofautologous serum [41] and serum-free culture [42] is technicallypossible to reduce the use of animal serum and to minimize thepotential side effects induced by exogenous materials [43].

4. Conclusions

A method was developed to prepare aECM scaffolds bycombining culture of autologous cells in a three-dimensionaltemplate, decellularization, and template removal. The aECMscaffolds showed excellent biocompatibility when implanted. Byusing aECM scaffolds for the culture of autologous cells, “fullautologous tissue engineering” could be realized to make the tissueengineered construct more biocompatible with the host. Further-more, ECM scaffolds may constitute a potent tool for otherbiomedical researches such as artificial stem cell niches. Theselective removal methods of biodegradable polymer templates bytreating with weak alkaline aqueous solution without the use oforganic solvent could also be applied in the other fields of materialsscience and biomedical engineering.

Acknowledgement

This work was supported by World Premier InternationalResearch Center Initiative on Materials Nanoarchitectonics fromthe Ministry of Education, Culture, Sports, Science and Technology,Japan.

Appendix. Supplementary data

Supplementary data associated with this article can be found inthe online version, at doi:10.1016/j.biomaterials.2010.12.016.

Appendix

Figure with essential color discrimination. Figs. 1, 3, 6 and 7 inthis article have parts that are difficult to interpret in black andwhite. The full color images can be found in the online version, atdoi:10.1016/j.biomaterials.2010.12.016.

References

[1] Langer R, Vacanti JP. Tissue engineering. Science 1993;260:920e6.[2] Griffith LG, Naughton G. Tissue engineering - current challenges and expanding

opportunities. Science 2002;295:1009e14.[3] Shin H, Jo S, Mikos AG. Biomimetic materials for tissue engineering. Bioma-

terials 2003;24:4353e64.[4] Hollister SJ. Porous scaffold design for tissue engineering. Nat Mater 2005;4:

518e24.[5] Engelmayr Jr GC, Cheng M, Bettinger CJ, Borenstein JT, Langer R, Freed LE.

Accordion-like honeycombs for tissue engineering of cardiac anisotropy. NatMater 2008;7:1003e10.

[6] Lutolf MP, Hubbell JA. Synthetic biomaterials as instructive extracellularmicroenvironments for morphogenesis in tissue engineering. Nat Biotechnol2005;23:47e55.

[7] Hutmacher DW. Scaffolds in tissue engineering bone and cartilage. Biomate-rials 2000;21:2529e43.

[8] Giancotti FG, Ruoslahti E. Integrin signaling. Science 1999;285:1028e32.[9] Discher DE, Mooney DJ, Zandstra PW. Growth factors, matrices, and forces

combine and control stem cells. Science 2009;324:1673e7.[10] Nair LS, Laurencin CT. Biodegradable polymers as biomaterials. Prog Polym Sci

2007;32:762e98.[11] Chen G, Ushida T, Tateishi T. Hybrid biomaterials for tissue engineering:

a preparative method for PLA or PLGA-collagen hybrid sponges. Adv Mater2000;12:455e7.

[12] Kaushal S, Amiel GE, Guleserian KJ, Shapira OM, Perry T, Sutherland FW, et al.Functional small-diameter neovessels created using endothelial progenitorcells expanded ex vivo. Nat Med 2001;7:1035e40.

[13] Gilbert TW, Sellaro TL, Badylak SF. Decellularization of tissues and organs.Biomaterials 2006;27:3675e83.

[14] Ott HC, Matthiesen TS, Goh SK, Black LD, Kren SM, Taylor DA, et al. Perfusion-decellularized matrix: using nature’s platform to engineer a bioartificial heart.Nat Med 2008;14:213e21.

[15] Uygun BE, Soto-Gutierrez A, Yagi H, Izamis M, Guzzardi MA, Shulman C, et al.Organ reengineering through development of a transplantable recellularizedliver graft using decellularized liver matrix. Nat Med 2010;16:814e20.

[16] Chan G, Mooney DJ. New materials for tissue engineering: towards greatercontrol over the biological response. Trends Biotechnol 2008;26:382e92.

[17] Williams DF. On the mechanisms of biocompatibility. Biomaterials 2008;29:2941e53.

[18] Badylak SF. The extracellular matrix as a biologic scaffold material. Biomate-rials 2007;28:3587e93.

[19] Tomford WW. Transmission of disease through transplantation of musculo-skeletal allografts. J Bone Joint Surg Am 1995;77:1742e54.

[20] Ota T, Taketani S, Iwai S, Miyagawa S, Furuta M, Hara M, et al. Novel method ofdecellularization of porcine valves using polyethylene glycol and gammairradiation. Ann Thorac Surg 2007;83:1501e7.

[21] Sefton MV, Babensee JE, Woodhouse KA. Innate and adaptive immuneresponses in tissue engineering. Semin Immunol 2008;20:83e5.

[22] Anderson JM, Rodriguez A, Chang DT. Foreign body reaction to biomaterials.Semin Immunol 2008;20:86e100.

[23] Babensee JE, Anderson JM, McIntire LV, Mikos AG. Host response to tissueengineered devices. Adv Drug Deliv Rev 1998;33:111e39.

[24] Badylak SF, Gilbert TW. Immune response to biologic scaffold materials.Semin Immunol 2008;20:109e16.

[25] Simon P, Kasimir MT, Seebacher G, Weigel G, Ullrich R, Salzer-Muhar U, et al.Early failure of the tissue engineered porcine heart valve SYNERGRAFT� inpediatric patients. Eur J Cardiothorac Surg 2003;23:1002e6.

[26] Meyer SR, Nagendran J, Desai LS, Rayat GR, Churchill TA, Anderson CC, et al.Decellularization reduces the immune response to aortic valve allografts inthe rat. J Thorac Cardiov Sur 2005;130:469e76.

[27] Cozzolino DJ, Cendron M, Devore DP, Hoopes PJ. The biological behavior ofautologous collagen-based extracellular matrix injected into the rabbitbladder wall. Neurourol Urodyn 1999;18:487e95.

[28] Hench LL, Polak JM. Third-generation biomedical materials. Science 2002;295:1014e7.

[29] Taylor PM. Biological matrices and bionanotechnology. Philos Trans R SocLond B; 2007:1313e20.

H. Lu et al. / Biomaterials 32 (2011) 2489e2499 2499

[30] Chepda T, Cadau M, Girin P, Frey J, Chamson A. Monitoring of ascorbate ata constant rate in cell culture: effect on cell growth. Vitro Cell Dev Biol Anim2001;37:26e30.

[31] Cimini M, Boughner DR, Ronald JA, Johnston DE, Rogers KA. Dermal fibroblastscultured on small submucosa: conditions for the formation of a neotissue.J Biomed Mater Res A 2005;75:895e906.

[32] Takashima A. Establishment of fibroblast cultures. Curr Protoc Cell Biol2001;2:2.1.1e2.1.12.

[33] Mazur P. Freezing of living cells: mechanisms and implications. Am J Physiol1984;247:C125e42.

[34] Chen G, Ushida T, Tateishi T. Scaffold design for tissue engineering. MacromolBiosci 2002;2:67e77.

[35] Smith LA, Ma PX. Nano-fibrous scaffolds for tissue engineering. Colloids Surf BBiointerfaces 2004;39:125e31.

[36] Schmidt CE, Baier JM. Acellular vascular tissues: natural biomaterials for tissuerepair and tissue engineering. Biomaterials 2000;21:2215e31.

[37] Singelyn JM, DeQuach JA, Seif-Naraghi SB, Littlefield RB, Schup-Magoffin PJ,Christman KL. Naturally derived myocardial matrix as an injectable scaffoldfor cardiac tissue engineering. Biomaterials 2009;30:5409e16.

[38] Place ES, Evans ND, Stevens MM. Complexity in biomaterials for tissue engi-neering. Nat Mater 2009;8:457e70.

[39] Jones KS. Effects of biomaterial-induced inflammation on fibrosis and rejec-tion. Semin Immunol 2008;20:130e6.

[40] Babensee JE. Interaction of dendritic cells with biomaterials. Semin Immunol2008;20:101e8.

[41] Stute N, Holtz K, Bubenheim M, Lange C, Blake F, Zander AR. Autologousserum for isolation and expansion of human mesenchymal stem cells forclinical use. Exp Hematol 2004;32:1212e25.

[42] Ahlfors JE, Billiar KL. Biomechanical and biochemical characteristics of a humanfibroblast-produced and remodeled matrix. Biomaterials 2007;28:2183e91.

[43] Martin MJ, Muotri A, Gage F, Varki A. Human embryonic stem cells express animmunogenic nonhuman sialic acid. Nat Med 2005;11:228e32.