Electrospun scaffolds for multiple tissues regeneration in vivo through

Zi Yin , Xiao Chen , Hai-xWei-liang Shen d, Jia-lin Chen a, b

Hong-Wei Ouyang a, b, c, *

a Department of Sports Medicine, School of Medicine, Zhb Zhejiang Provincial Key Laboratory of Tissue Engineerc State Key Laboratory for Diagnosis and Treatment of IDiseases, The First Afliated Hospital, College of Medicind Department of Orthopedic Surgery, 2nd Afliated Hose Department of Rehabilitation, Sir Run Run Shaw Hospf Department of Biosystems Science & Engineering (D-B

ology and enhancedqRT-PCR analysis ofat MSCs cultured onompared with cellsension release abro-. Collectively, theseeir inductive role infunctionalized bio-

. All rights reserved.

Stem cell survival, self-renewal and differentiation are governedby local biochemical and mechanical factors within their

mponents includesoluble factors, other cells, and extracellular matrix molecules.Whilethe role of biochemical signals is well-documented, the importanceof biophysical cues has received more recognition and attention onlyin the last decade. Current advances inmicrofabrication technologieshave enabled the generation of substrates with nano/micro-scaletopographies to study the effects of biophysical signals on cellularfunction. A number of studies have demonstrated that the physicalproperties of substrata have profound effects on the cellular func-tions of pluripotent and multipotent stem cells, including cell

* Corresponding author. Center for Stem Cell and Tissue Engineering, School ofMedicine, Zhejiang University, 866 Yu Hang Tang Road, Hangzhou 310058, China.Tel./fax: 86 571 88208262.

E-mail address: hwoy@zju.edu.cn (H.-W. Ouyang).1

Contents lists availab

Biomat

journal homepage: www.elsev

Biomaterials 44 (2015) 173e185These authors contribute equally to this work.raphy. Furthermore, MSCs on the aligned substrate exhibited tenocyte-like morphtenogenic differentiation compared to cells grown on randomly-oriented scaffold.osteogenic marker genes and alkaline phosphatase (ALP) staining demonstrated thrandomly-oriented ber scaffolds displayed enhanced osteogenic differentiation ccultured on aligned ber scaffolds. Finally, it was demonstrated that cytoskeletal tgated the divergent differentiation pathways on different substrate topographyndings illustrate the relationship between topographic cues of the scaffold and thtissue regeneration; thus providing an insight into future development of smartscaffold design and its application in tissue engineering.

2014 Elsevier Ltd

1. Introduction microenvironmental niche [1]. The key niche coBone formationMesenchymal stem cells

bone tissue formation through ossication. Additionally, X-ray imaging and osteocalcin immunohisto-chemical staining also demonstrated that osteogenesis in vivo is driven by randomly oriented topog-a r t i c l e i n f o

Article history:Received 10 September 2014Accepted 20 December 2014Available online

Keywords:Bio-scaffold topographyTissue engineeringTendon regenerationhttp://dx.doi.org/10.1016/j.biomaterials.2014.12.0270142-9612/ 2014 Elsevier Ltd. All rights reserved.in Song , Jia-jie Hu , Qiao-mei Tang , Ting Zhu ,, Huanhuan Liu a, b, Boon Chin Heng f,

ejiang University, Hangzhou, Chinaing and Regenerative Medicine, Hangzhou, Chinanfectious Diseases, Collaborative Innovation Center for Diagnosis and Treatment of Infectiouse, Zhejiang University, Hangzhou, China

pital, School of Medicine, Zhejiang University, Hangzhou, Chinaital, School of Medicine, Zhejiang University, Hangzhou, ChinaSSE), ETH-Zurich, Basel, Switzerland

a b s t r a c t

Physical topographic cues from various substrata have been shown to exert profound effects on thegrowth and differentiation of stem cells due to their niche-mimicking features. However, the biologicalfunction of different topographic materials utilized as bio-scaffolds in vivo have not been rigorouslycharacterized. This study investigated the divergent differentiation pathways of mesenchymal stem cells(MSCs) and neo-tissue formation trigged by aligned and randomly-oriented brous scaffolds, bothin vitro and in vivo. The aligned group was observed to form more mature tendon-like tissue in theAchilles tendon injury model, as evidenced by histological scoring and collagen I immunohistochemicalstaining data. In contrast, the randomly-oriented group exhibited much chondrogenesis and subsequenta, b, 1 a, b, 1 e a, b a, b a, btopography dependent induction of lineage specic differentiationle at ScienceDirect

erials

ier .com/locate/biomater ia ls

rialsadhesion, morphology, proliferation, migration and differentiation[2e6]. Based on the concept of contact guidance, we designed abiomimetic aligned nanober scaffold modelled on the parallelcollagenous bers of tendon extracelluar matrix; and subsequentlydemonstrated that alignment within the scaffold regulate tendonstem cell orientation and induce specic teno-lineage differentiation[7]. Meanwhile, both random orientation and nanoscale disorderhave been demonstrated to induce ossication of human multi-potent stem cells in vitro, even in the absence of osteogenic media[6]. Although there are promising results in various studies that haveattempted to control stem cell fate in vitro through modication ofsubstrate physical properties, there is a dire need to move fromculture substrate to implantable scaffolds with direct applications intissue engineering [8].

Conventional scaffolds are designed and fabricated according tothe basic requirements of biocompatibility, structural support aswell as cell delivery, and have already been widely utilized invarious tissue engineering applications [9]. Modern bio-scaffoldsnot only just serve as a carrier for seed cells, but also provide anappropriate microenvironment for stem cells and mediates bio-logical functions. Further microstructural renement of currentscaffold biotechnology will enhance the progress of tissue engi-neering in the future [10]. However, the three-dimensionalmicroenvironment in vivo represents a much more complicatedmilieu that encompasses a much more diverse multitude ofsignaling cues compared to an in vitro culture system. Underphysiological conditions, stem cells naturally encounter a variety ofdifferent signaling cues that can potentially inuence cell fate. It isessential and necessary to use the results of in vitro studies to aidthe rigorous characterization of the functionality of tissue engi-neered scaffolds in vivo. This prompted our investigation on theinductive effects of scaffold topographic cues on stem cell differ-entiation pathways and lineage fate.

This study aims to characterize the biophysical effects of scaffoldtopography on tissue regeneration in vivo within a 3D microenvi-ronment, utilizing aligned and randomly-oriented brous scaffolds.We tested the hypothesis that topographic cues from the alignedbrous scaffold can enhance tendon-like tissue formation, and thatthere would be a higher degree of osteogenesis and tissue ossi-cation with the randomly-oriented ber scaffold. Additionally, therole of cytoskeletal organization in topography driven differentia-tion of mesenchymal stem cells was also investigated in vitro. Webelieve that the data presented here would be benecial to thedesign and application of future biomaterials.

2. Materials and methods

2.1. Fabrication of PLLA scaffolds

Both aligned (1068 190 nm) and randomly-oriented PLLA scaffolds(739 129 nm) were fabricated using the electrospinning technique as previousreported. The polymer solution was prepared by dissolving PLLA (Ji'nan DaigangBiomaterial Co., Ltd) in a mixture of chloroform/ethanol (3:1) at a concentration of4% (aligned) or 3% (random). The solution was then fed into a 12-ml plastic syringe,which was controlled by a syringe pump at a rate of 2 ml/h. A high voltage (12 kV)was applied to the needle tip, which was placed 10 cm above the collector. A ataluminum plate was used to collect the random bers. The collector for aligned -bers was a disk rotating at 4000 rpm. The resulting scaffolds were then transferredto cover slips and sterilized with ethanol and UV overnight before theywere utilizedfor cell culture. Nanobers were collected for 2e3 h, resulting in a ber mat rangingin thickness from 0.14 to 0.17 mm. The aligned and randomly-oriented scaffoldsutilized in this study were of similar thickness and distribution.

2.2. Morphology of PLLA scaffolds

The scaffold samples were sputter-coated with gold, and then their structurewas observed under scanning electron microscopy (SEM) (Hitachi S3000N) at anaccelerating voltage of 15 kV. After the micrographs were obtained, image analysis

Z. Yin et al. / Biomate174software (Image-Pro Plus) was used to measure the average diameter of the nano-bers (n 3). For each sample, an average of 50 bers were counted.2.3. SEM imaging

C3H10T1/2 cells (mouse multipotent mesenchymal stem cell line) were ob-tained from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China).Cells were seeded onto PLLA scaffolds at 2104 cells/cm2 and cultured in Dulbecco'smodied Eagle's medium (DMEM, low glucose; Gibco, Grand Island, NY, http://www.invitrogen.com) with 10% (v/v) fetal bovine serum (FBS; Invitrogen, Carls-bad, CA, http://www.invitrogen.com-Gibco) and 1% (v/v) penicillin-streptomycin(Gibco). The medium was changed once every 3 days. Three days after seeding,the cell morphology and distribution were visualized using SEM. Specimens werexed in 0.25% glutaraldehyde solution, and then rinsed 3 times in PBS, 30 min eachtime. The specimens were immersed in OsO4 for 40 min and then rinsed 3 times inPBS, 30 min each time, followed by dehydration in increasing concentrations ofacetone (30e100% v/v). After drying, the specimens were mounted on aluminumstubs and coated with gold, then viewed under a Hitachi S-3000N SEM at anaccelerating voltage of 15 kV. For quantication of cell morphology on differentscaffolds, aminimum of fty cells for each SEM imagewere selected randomly as ROI(Regions of Interest). The area and radius ratio were then quantied using theImage-Pro Plus software.

2.4. Alkaline phosphatase (ALP) staining

C3H10T1/2 cells (104/cm2) were seeded onto scaffolds and cultured in osteo-genic induction medium, in the presence of 10 mM b-glycerol phosphate (Sigma),0.1 mM dexamethasone (Sigma), and 50 mg/ml ascorbic acid (Sigma) supplementedin DMEM-high glucose medium containing 10% (v/v) FBS and 1% (v/v) pen-icillinestreptomycin. After 7 days, ALP activity was assayed using a BCIP/NBT alka-line phosphatase color development kit (Beyotime Institute of Biotechnology). DAPI(Beyotime Institute of Biotechnology) was used to stain nuclei and observed under alight microscope (Olympus IX71).

2.5. Quantitative PCR

Total cellular RNA was isolated by lysis in TRIzol (Invitrogen). The expressionlevels of tendon-specic genes and osteogenic markers in cells cultured on alignedand randomly-aligned brous scaffolds were assessed by quantitative PCR. PCR wasperformed using a Brilliant SYBR Green qPCR Master Mix (TakaRa) on a Light Cyclerapparatus (ABI 7900HT). The PCR cycling consisted of 40 cycles of amplication ofthe template DNA with primer annealing at 60 C. The relative expression levels ofeach target gene was then calculated using the 2-DDCt method. The amplicationefciencies of primer pairs were validated to enable quantitative comparison of geneexpression. All primers (Generay) were designed using primer 5.0 software and aresummarized in the Supplementary Table 1.

2.6. Animal model

The Zhejiang University Institutional Animal Care and Use Committee approvedthe study protocol. In situ rat Achilles tendon repair model: Twenty hind limbs ofskeletally mature female rats weighing 200e220 gwere utilized for this experiment.Under general anesthesia, a gap wound was created and the Achilles tendon wasremoved to create a defect of 6 mm in length. Aligned and random brous scaffolds(8 mm 8 mm, thickness 100 um) were folded about 2 mm from the bottom ofthe membrane upwards. This was followed with fan-folding the next 2 mm to theback. Fan-folding of the scaffold was continued until it was completely folded. Thiswas followed by binding the center of the strip using a suture and subsequentplacement into the gap wound. Suturing to the remaining Achilles tendon was thencarried out using a non-resorbable suture material (Nylon6). The wound was thenirrigated and the skin was closed. The animals were allowed free cage activity aftersurgery. At 2, 4, and 8 weeks post-implantation, samples from each group wereharvested for the evaluation of histology, transmission electronmicroscopy imaging,mechanical testing, as well as collagen content determination (SupplementaryTable 2).

2.7. Immunouorescence

Briey, cells were xed in 4% (w/v) paraformaldehyde for 10 min at roomtemperature, permeabilized, and blocked for 30 min with 1% (w/v) bovine serumalbumin, and then permeabilized with 0.1% (w/v) Triton X-100. Fixed cells werewashed and incubated with a primary antibody against SCX (Abcam Inc.), Vinculin(Millipore), or control IgG (BD) at 4 C overnight. Cells were then incubated withAlexa uor 488-conjugated secondary antibody (Invitrogen) for 2 h and the nucleiwere stained with DAPI. TRITC-phalloidin (Millipore) staining was used to visualizethe cytoskeleton. The imaging was then performed with confocal microscopy (ZeissLSM-510).

2.8. Histological evaluation and staining

Harvested specimens were immediately xed in 10% (v/v) neutral bufferedformalin, dehydrated through an alcohol gradient, cleaned, and then embeddedwithin parafn blocks. Histological sections (7 um) were prepared using a micro-

44 (2015) 173e185tome and subsequently stained with hematoxylin and eosin. In addition, Massontrichrome staining was performed according to standard procedures to examine the

rialsgeneral appearance of the collagen bers. To detect proteoglycan synthesis as anindicator of cartilage formation, sections were stained with 0.1% Safranin O (Sangon,Shanghai, China) for 8 min, followed by counter staining with 0.02% Fast Green for4 min. Polarizing microscopy was employed to detect mature collagen brils. Gen-eral histological scoring was performed using hematoxylin and eosin staining. Sixparameters (ber structure, ber arrangement, rounding of nuclei, inammation,vascularity, cell population) were semi-quantitatively assessed. These six parame-ters were semi-quantitatively graded on a four-point scale (0eIII), with 0 beingnormal and 3 being maximally abnormal. Therefore, a normal tendonwould score 0,while a maximally abnormal tendon would score 18. The detailed scoring systemwas summarized in Supplementary Table 3.

2.9. Immunohistochemistry

Parafn sections (7 mm) were incubated in antigen retrieval buffer (100 mM Tris,5% (w/v) urea, pH 9.5) at 95 C for 10 min for antigen retrieval. Endogenousperoxidase was blocked by incubation with 3% (v/v) hydrogen peroxide in methanolfor 10 min. Non-specic protein binding was blocked by incubation with 10% (v/v)goat serum. After overnight incubation at 4 C with primary antibodies againstCollagen (Abcam Inc.) or osteocalcin (Millipore), sections were washed and thenincubated with goat anti-mouse (Beyotime Institute of Biotechnology Inc., Jiangsu,China) or goat anti-rabbit (Beyotime Institute of Biotechnology Inc., Jiangsu, China)secondary antibodies for 2 h at room temperature. The DAB substrate system (Zsbio,Beijing, China) was used for color development. Hematoxylin staining was used toreveal the nuclei.

2.10. Mechanical testing

Mechanical testing was performed using an Instron tension/compression sys-tem with Fast-Track software (Model 5543, Instron, Canton, MA). Measurements ofthe tendon cross-sectional area were performed using two Vernier calipers at 5 mmproximal to the conjunction of bone and tendon. The bone end of the tendon wassecured by a specially designed restraining jig and the tendon end was pinched witha clamp [11]. The AT-calcaneus complex (ACC) was then rigidly xed to custom-made clamps. After applying a preload of 0.1 N, each ACC underwent pre-conditioning by cyclic elongation of between 0 and 0.5 mm for 20 cycles at 5 mm/min. This was followed by a load to failure test at an elongation rate of 5 mm/min.The loadeelongation behavior of the ACCs and failure modes were recorded. Thestructural properties of the ACCwere represented by stiffness (N/mm), ultimate load(N), energy absorbed at failure (mJ) and stress at failure. For each ACC, the greatestslope in the linear region of the loadeelongation curve over a 0.5 mm elongationinterval was used to calculate the stiffness.

2.11. Determination of collagen content

The amount of deposited collagen in the scaffold was quantied by using acollagen assay kit according to the manufacturer's protocol (Jiancheng Ltd., Nanjing,China). This kit is a hydroxyproline assay, which is widely used to determine totalcollagen content, and a conversion factor of 1:7.46 was used to convert hydroxy-proline to collagen [12], lyophilized tendons were digested with a hydrolysis regentat 95 C for 20min. Serial dilutions of acid-soluble collagen type I provided by the kitwere utilized as standards. Following the assay, collagen concentration was deter-mined through absorbance measurements at 550 nm using a microplate reader(Molecular Devices).

2.12. Transmission electron microscopy

Tissue specimens were xed by standard procedures for TEM to assess collagenbril diameter and alignment. Briey, samples were pre-xed in 2% (w/v) glutar-aldehyde for 2 h at 4 C and then washed twice in PBS at 4 C followed by post-xation with 1% (v/v) osmic acid for 2 h at 4 C. After two washes in PBS, thesamples were dehydrated with an ethanol gradient and dried to a critical point. Thesamples were then mounted and sputter-coated with gold for viewing under TEM(Quanta 10 FEI). Approximately 500 collagen brils were measured for each sampleto obtain an accurate representation of the bril diameter distribution.

2.13. Radiographic evaluation

The X-ray photographs of whole animals were captured with a non-invasiveKodak-FX in vivo imaging system (Kodak, Inc.) to evaluate ectopic bone formation.At 8 weeks post-implantation, captured images were analyzed with the Image-proplus software to quantify the area, mean density, max density, and sum density ofectopic bone formation.

2.14. Statistical analysis

All quantitative data sets are expressed as mean SD. The Student's t-test wasperformed to assess statistically signicant differences in the results of different

Z. Yin et al. / Biomateexperimental groups. Values of p < 0.05 were considered to be signicantlydifferent.3. Results

3.1. Fabrication and morphological characterization of scaffolds

Electrospinning is used to fabricate aligned brous scaffolds, aswell as randomly-oriented brous membranes with similar berdiameters, which served as controls. The surface topography ofboth aligned and randomly-oriented brous scaffolds was exam-ined using SEM (Fig. 1A and B). Majority of the bers in the alignedscaffolds were parallel to each other and formed angles from 0 to10 with respect to the horizontal axis, while the randomly-oriented nanober scaffold exhibited nearly equal distributions atall angles.

3.2. The effects of aligned and randomly-oriented scaffolds on neo-tissue formation

3.2.1. Histology of repaired tendonsAfter 2 weeks post-surgery, histological analysis showed that

there was a greater number of cells exhibiting spindle-shapedmorphology on the aligned scaffolds (Fig. 1C). While, the cellmorphology in the randomly-oriented groups displayed relativelyround shapes (Fig. 1D). Masson trichrome staining showed that thetissue matrix was denser in the aligned versus randomly-orientedscaffold group, which means that more collagen bers have beendeposited compared to the control group (Fig. 1 E and F). Collagen Iimmunohistochemical staining showed organized collagen depo-sition on the aligned bers (Fig. 1 G and H), while the randomly-oriented scaffolds were lled with loose, disarranged matrix.

At 4weeks post-surgery, the histological results showed that thealigned scaffold-induced spindle-shaped cells and tendon-like tis-sue formation in vivo, as evidenced by the Masson trichromestaining. In addition, the aligned scaffold implantation showedimproved tendon repair quality was compared to the histology ofblank groupwithout scaffold implantation, which displayed limitedtissue regeneration (Supplementary Fig.S1). The collagen I immu-nohistochemical staining reveals enhanced collagen I matrix pro-duction and arrangement in the aligned scaffold group (Fig. 2A).Histology scoring also conrmed the histological results that thequality of repaired tendons in the aligned scaffold group weresignicantly better than the control group (p < 0.05, Fig. 2E),particularly in ber structure, ber arrangement and nuclei shapeaspects. The ultrastructural morphology based on TEM imagingfrom transection revealed larger bril formation within the alignedscaffold, as compared to the thin brils observed within therandomly-oriented scaffold (Fig. 2B). The average diameter ofcollagen brils in the aligned group was 52.88 nm (373 bers),which was 121.4% of the random group (43.57 nm, 266 bers) at 4weeks post-implantation (Fig. 2D). Analysis of the distribution ofbril diameters within the two groups (Fig. 2C) revealed that theATs in the aligned scaffold treatment group formed signicantlylarger brils compared to ATs in the random group at 4 weeks post-implantation. The collagen content assay demonstrated that thealigned group had signicantly more collagen deposition,compared to the random group (0.281 0.039 mg/mg vs.0.198 0.072 mg/mg, p < 0.05, Supplementary Fig.S2).

We also utilized polarized light microscopy to compare thecollagen ber maturation levels within the two groups. The alignedscaffold group exhibited more continuous collagen bers at therepair site (Fig. 3B). Furthermore, the expression of tendon-specicmarker scleraxis (Scx)was signicantly higher in the aligned versusrandomly-oriented group at 2 weeks and 8 weeks post-surgery(Fig. 3A). The other teno-lineage marker gene tenomodulin(Tnmd) also displayed signicantly higher levels in the aligned

44 (2015) 173e185 175scaffold group, thus indicating that aligned topographic cues have

Fig. 1. (A) and (B) SEM micrographs (1000 ) of electrospun PLLA with aligned (A) and randomly-oriented (B) brous scaffold surface topography. Scale bars, 50 mm. Histologicalresults of repaired rat Achilles tendon in section of aligned group (upper panel) and randomly-oriented group (lower panel) at 2 weeks post-surgery, (C) and (D) are typicalhematoxylin and eosin staining, (E) and (F) are Masson trichrome staining, (G) and (H) are immunohistochemical staining of collagen type I, respectively. Arrows indicated theremaining brous scaffolds. Scale bars, 50 mm, 100 mm (inset).

Fig. 2. Repaired rat Achilles tendon at 4 weeks post-surgery. (A) Typical hematoxylin and eosin staining, Masson trichrome staining, immunohistochemical staining of collagen typeI in repaired zones within sections of aligned group (upper panel) and randomly-oriented group (lower panel). (B) Transmission Electron Microscopy images show ultrastructure ofrepaired tendons after 4 weeks post-surgery. (C) Histogram and distribution of collagen bril diameters of aligned group (upper panel) and randomly-oriented group (lower panel).(D) Collagen bril diameters. Data are mean SD, n 3. (E) The overall histology score is the sum of six parameters (ber structure, ber arrangement, rounding of nuclei,inammation, vascularity, cell population). Statistically signicant at *p < 0.05. Scale bars, 100 mm (HE and Masson stained images), 200 mm (immunohistochemical staining ofcollagen type I, left images), 50 mm (immunohistochemical staining of collagen type I, right images).

Z. Yin et al. / Biomaterials 44 (2015) 173e185176

rialsZ. Yin et al. / Biomatemore potential in tendon tissue repair and regeneration.Msx-2 [13],which has been reported to play a central role in preventing ten-dons from mineralizing, exhibited elevated expression in thealigned versus randomly-oriented groups at all time-points(Fig. 3A).

3.2.2. Histology of bone formationOn the other hand, chondrocyte-like cells appeared within the

injury site implanted with the randomly-oriented scaffold at 4weeks post-surgery (Fig. 2A). Furthermore, Safranin O stainingshowed signicantly more chondrocyte-like cells in the randomly-oriented versus aligned scaffold group (Fig. 4A). It is obvious withthe randomly-oriented group that the ossied deposits surroundedby chondrocyte-like cells were localized at the tendon mid-sectioninside the wound, which caused disruption to the organization ofcollagen bers (Fig. 4A). Additionally, bone marrow was alsoformed in the randomly-oriented scaffold group at 8 weeks(Fig. 4A) Nevertheless, no ossication was detected in all sampleswith X-ray scanning at 4 weeks post-surgery (data no shown),whereas all samples of the randomly-oriented group exhibitedspontaneous ectopic bone formation at 8 weeks post-surgery(Fig. 5A). Upon quantication of the area and density of ectopicbone formed, it was found that the randomly-oriented scaffoldgroup displayed signicantly larger area (Fig. 5B, p < 0.05), highermean density (Fig. 5B), max density (Fig. 5B, p < 0.05) and inte-grated optical density (IOD) (Fig. 5B, p < 0.05) of ectopic bone, ascompared to the aligned scaffold group. The results thus indicated

Fig. 3. (A) Gene expression levels of tendon-related markers assessed by quantitative PCR(upper panel) and randomly-oriented group (lower panel) at 2, 4 and 8 months post-surge44 (2015) 173e185 177that randomly-oriented topographic scaffolds had signicantlyhigher potential to induce ectopic bone formation compared toaligned scaffolds.

Expression of osteochondral-lineage marker genes wereanalyzed to further evaluate tissue formation. It was observed thatthe expression of growth factors BMP4, chondrogenic transcriptionfactor Sox9 and chondrocyte specic matrix Col IIwere signicantlyhigher in the randomly-oriented versus aligned groups at 2 weekspost-surgery (Fig. 4B). The matrix osteocalcin expression wassignicantly much higher in the randomly-oriented versus alignedgroups, while expression of the osteogenic transcription factorRunx2 exhibited the largest difference at 8 weeks post-surgery(Fig. 4B). This suggests that randomly-oriented topographic cuesplay a critical role in initiating cartilage and bone formation at theearly repair stage (2 weeks) with neo-bone being formed as a resultof cartilage ossication at the later repair stage (8 weeks). Theexpression of collagen type X, a representative marker of chon-drocyte hypertrophy was detected by immunohistochemicalstaining. Comparison of the two groups showed signicantlydenser and larger positively-stained areas within the randomly-oriented versus aligned group (Fig. 6), which is consistent withincreased Safranin-O staining (Fig. 4A). This indicated that theossied deposits were formed by endochondral ossication.Expression of the bone formation marker osteocalcin was exam-ined by immunohistochemical staining as well, and the expressionlevels were obviously much higher in the randomly-orientedversus aligned group at 8 weeks post-surgery (Fig. 6). These data

at 2, 4 and 8 months post-surgery. (B) Polarized microscopy images of aligned groupry. Statistically signicant at *p < 0.05, **p < 0.01. Scale bars, 100 mm.

rialsZ. Yin et al. / Biomate178collectively suggested that randomly-oriented scaffolds can inducemore endochondral bone tissue formation by 8 weeks post-implantation.

3.2.3. Mechanical properties of repaired tendonsTo further correlate tissue structural features with their me-

chanical properties, harvested tendons (n 5 for each group)were subjected to mechanical testing at 8 weeks post-surgery. Thealigned group had better mechanical properties than therandomly-oriented controls (Fig. 7 and Supplementary Fig.S3). Themodulus (24.42 2.20 MPa vs. 20.86 3.56 MPa) of the alignedgroup were better than the control group (Fig. 7). The energy inthe aligned group was 20% higher than that of the control group(179.65 9.29 mJ vs. 149.23 38.49 mJ, Fig. 7). The maximumforce in the aligned group was higher than that of the controlgroup (87.12 5.61 N vs. 77.09 13.27 N, p > 0.05). The stress atfailure and stiffness were consistently higher in the aligned groupthan that of the control group, but these difference was not sta-tistically signicant (Fig. 7), probably due to the small samplingsize.

Fig. 4. (A) Safranin O staining images of aligned group (upper panel) and randomly-orientchondrogenic (aggrecan, Sox9, collagen type II) and osteogenic markers (Bmp4, Ocn, Runx2) a*p < 0.05, **p < 0.01. Scale bars, 200 mm (100X), 100 mm (200X).44 (2015) 173e1853.3. The effects of topographical cues on MSCs

The SEMmicrographs indicated thatMSCswerewell attached toboth scaffolds, but displayed distinct morphology. On the alignednanober scaffold, the cells exhibited an elongated spindle-shapedmorphology and were oriented parallel to the substrate alignment,whereas cells cultured on the randomly-oriented scaffolds werespread out and exhibited a polygonal phenotype (Fig. 8A and B).The bar graphs illustrate the size and aspect ratios of MSCs culturedon different topographic substrates. MSCs grown on the alignedscaffold displayed relatively smaller size, but with a signicantlymuch higher aspect ratio than their counterparts on the randomly-oriented scaffold (Fig. 8C and D). Under confocal uorescence mi-croscopy, alignment of cell orientation was also apparent underTRITC-phalloidin staining, with the cytoskeleton being more uni-formly oriented towards the alignment of nanobers within thealigned scaffold (Fig. 9A). However, the MSCs cultured on therandomly-oriented scaffolds exhibited different arrangement of theF-actin network (Fig. 9A). The immunouorescence staining ofvinculin, which is a component of the focal adhesion complex,showed the presence of focal adhesions and their distribution

ed group (lower panel) at 4 and 8 months post-surgery. (B) Gene expression levels ofssessed by quantitative PCR at 2, 4 and 8 months post-surgery. Statistically signicant at

rialsZ. Yin et al. / Biomatewithin cells cultured on scaffolds. Vinculin was localized mainly atthe peripheral region of MSCs cultured on the randomly-orientedscaffold, at the end of F-actin ber bundles in the lopodia orlamellipodia. A higher density of vinculinwas observed at the polesof elongated spindle-shaped MSCs cultured on the aligned scaffold.Consistent with the great differences in cell morphology andorientation observed on the two scaffolds, the focal adhesion andcytoskeleton distribution also displayed distinctively differentpatterns between the two groups (Fig. 9A). The expression of thetenogenic transcription factor gene Scx was signicantly morehighly expressed by MSCs on aligned versus randomly-orientednanobers (Fig. 9C). The expression of the osteogenic transcrip-tion factor Runx2 was signicantly lower in MSCs cultured onaligned versus randomly-oriented scaffold (Fig. 9D).

3.4. Topography-induced lineage commitment of MSCs isdependent on cytomyosin cytoskeleton

To further understand the phenomena of topographic induceddifferentiation of MSCs, we examined the effects of two smallmolecules on cytoskeletal reorganization andmechanotransduction- cytochalasin D (cyto D) and the Rho kinase (ROCK) inhibitor Y-27632. The addition of cyto D to the medium attenuated the contactguidance response by suppressing cell elongation in the alignedscaffold group, while reducing projected surface area of cells in therandomly-oriented group (Fig. 9B). We observed that cytochalasin D

Fig. 5. (A) X-ray images of repaired tendons of mice at 8 weeks post-transplantation and evmax density, sum density and integrated optical density of ectopic bone in aligned group (ectopic bone formation. Statistically signicant at *p < 0.05.44 (2015) 173e185 179treatment caused cells to become rounded without any notabledifferences in cell morphology between the two groups (Fig. 9B).Meanwhile, we also found that random orientation-inducedosteogenesis and alignment-induced tenogenesis were attenuatedas well by cyto D, as evidenced by downregulation of osteogenicmarker Runx2 and tenogenic mark Scx expression levels (Fig. 9C andD).These results indicated that the actin cytoskeleton might beimportant to the MSC lineage commitment process. Furthermore,we treated the cell with Rho kinase (ROCK) inhibitor Y-27632 toinhibit myosin-generated cytoskeletal tension. In the presence of Y-27632, cells remained well-spread and morphologically similar onthe two different topographical substrates. Additionally, vinculinexpression was diminished signicantly in both groups (Fig. 9A).Moreover, the quantitative PCR results revealed that there were nosignicant differences in either tenogenesis or osteogenesis of MSCscultured on different topographic substrates (Fig. 9C andD). The SCXimmunouorescence images showed that before addition of Y-27632, SCX expression was higher and more concentrated in thenuclei ofMSCs cultured on aligned bers. Subsequent exposure to Y-27632 not only blocked cell alignment but also eliminated differ-ences in SCX expression between the two groups (Fig. 10). Similarly,the ALP staining results also demonstrated that osteogenesisinduced by randomly-oriented nanober topographywas abrogatedupon treatment with Y-27632 (Fig. 11). These ndings collectivelysuggest that topography-induced MSC differentiation andmorphological change are dependent on cytoskeletal tension.

aluation of the extent of ossication (n 5). (B) The distribution and medians of area,dot) and randomly-oriented group (square). Arrows depict exactly the locations of the

Z. Yin et al. / Biomaterials 44 (2015) 173e1851804. Discussion

This study is based on our previous report on alignment-induced tenogenesis of tendon stem cells, as well as osteo-genesis being enhanced by randomly-aligned ber scaffolds [7].

Fig. 6. Immunohistochemical staining of collagen type X and osteocalcin in repaired zoneweeks post-surgery. Scale bars, 100 mm (200X), 50 mm (400X).

Fig. 7. Mechanical properties of repaired teThis inspired us to evaluate the long term efcacy of aligned andrandomly-oriented brous scaffolds for tissue engineering in vivoand to investigate the effects of topographic cues on mesenchymalstem cells, which are more likely to be utilized clinically. In thisstudy, both the aligned and randomly-oriented brous scaffolds

s within sections of aligned group and randomly-oriented group after 4 weeks and 8

ndons at 8 weeks post-surgery. n 5.

us sc

rialsFig. 8. (A) and (B) showMSCs cultured on the aligned and randomly-oriented nanobro(D) on different substrates. Statistically signicant at *p < 0.05, **p < 0.01.

Z. Yin et al. / Biomatewere utilized as tissue engineering platforms in vivo. The alignedbrous scaffolds promoted tendon-like tissue formation and yiel-ded much more mature tendon tissue at 4 weeks post-implantation. In contrast, we observed substantial chondro-genesis and subsequent tissue ossication at the implantationsites of the randomly-oriented brous scaffold group. Theobserved signicant up-regulation of tendon-specic markers inthe aligned versus randomly-oriented brous scaffold group,provided further evidence that aligned topography initiated teno-lineage differentiation. Consistent with the topographic effectin vivo, the random orientation of bers within the scaffold pro-moted osteogenesis of MSCs in vitro, as evidenced by elevatedosteogenic marker expression and positive ALP staining. Finally,the results of further experiments showed distinct differences inthe distribution of focal adhesion complexes and cytoskeletal or-ganization of MSCs on the aligned and randomly-oriented brousscaffolds. Treatment with cyto D and Y-27632 led to the loss ofspindle-shaped morphology of MSCs cultured on the alignedbrous scaffold and abrogated the effects of topography-induceddifferentiation. These results thus highlight the important role ofthe actomyosin cytoskeleton in lineage commitment of MSCs.Collectively, these ndings illustrate the relationship betweentopographic cues of the scaffold and tissue formation, as well assuggesting the possible risks of scaffold-induced ectopic ossica-tion in tendon tissue repair.

Although there are numerous studies demonstrating thatvarious physical characteristics of scaffolds can profoundly inu-ence stem cell biological functions, particularly cell fate decision[14e17], it is still unknown whether any topography-induced ef-fects observed in vitro can be faithfully recapitulated in vivo. Thisstudy compared the function of two microstructurally distinctbrous scaffolds in an Achilles tendon injury model to evaluateneo-tissue regeneration potential. Our in vivo study showed thataffolds respectively. Scale bars, 100 mm. Quantitative data of cell size (C) and radius ratio

44 (2015) 173e185 181tendon-like tissue formation, assessed by histological examinationand immunohistochemical analysis, were signicantly increased inthe aligned versus randomly-oriented brous scaffold group. Incontrast, histological analysis clearly revealed much tissue ossi-cation and bone formation in the randomly-oriented brous scaf-fold group. At two weeks post-implantation, there were signicantincreases in the expression levels of chondro-lineage specic genessuch as collagen type II, Sox9 and aggrecan in the randomly-oriented versus aligned brous scaffold group, thus suggestingthat chondrogenesis was initiated during the early repair stage. Thebiological microenvironment of injured tendons is very different tothat of normal healthy tendons. Transforming growth factor-b2(TGF-b2), TGF-b3, bone morphogenetic protein-2 (BMP-2), BMP-4BMP-7 and vascular endothelial growth factor (VEGF) were signif-icantly up-regulated [18]. These growth factors have been impli-cated in the regulation of cartilage and bone development and havebeen widely utilized in various chondrogenic and osteogenic dif-ferentiation protocols [19,20]. We reported that alignment-inducedlineage commitment could even override the effects of osteogenicinduction medium, in contrast to the synergetic effect of randomly-aligned topographic cues on osteogenesis [7]. Consistent with thein vitro data, the randomly-oriented brous scaffold group dis-played more mature bone tissue formation during healing.Regarding the possible source of endogenous progenitor cells fortendon repair, it is likely to be either MSCs that have migrated intothe wound or tendon-derived stem cells, which was identiedrecently [21]. These cells possess multipotent differentiation ca-pacity and would likely undergo aberrant differentiationwithin thetendon injury niche. A comparative study of tenocytes andmesenchymal stem cells seeded on polyglycolic acid (PGA) andcollagen type I scaffolds in a full-size tendon defect model showedthat the transplantation of tenocytes results in a lower degree oftissue ossication and better extracellular matrix organization, in

rialsZ. Yin et al. / Biomate182comparison to the use of MSCs alone or just scaffold materials [22].Our ndings indicated that ectopic tendon ossication, which maydevelop following surgical trauma deserves more attention, as thiscould shed light on the development of novel bio-scaffolds withappropriate micro/nanoscaled structures.

In recent years, the physical properties of scaffolds have beengiven more attention and were explored by a variety of differenttechniques, such as lithography, nano/micro-pattern, electro-spinning, as well as decellularization. Electrochemically alignedcollagen threads have also been used to keep the cells orientedparallel to aligned collagen bers and it was also found thatanisotropic orientation promotes tenogenic differentiation of hu-man MSCs in the absence of bio-inductive cues in vitro [14]. Zhuet al. reported that simply maintaining the cultured tenocytes in an

Fig. 9. (A) Immunouorescence staining of vinculin and F-actin in MSCs cultured on the aligwith the right column being the merged images with DAPI staining. Scale bars, 20 mm. (B)scaffolds, with or without exposure to cyto D for 48 h, with the right column being the merg(C) and osteogenic-specic gene (D) expression of MSCs seeded on different scaffolds withhousekeeping gene, Gapdh. n 5. Statistically signicant at *p < 0.05.44 (2015) 173e185elongated form by culturing them on microgrooved siliconemembrane could maintain their phenotype [23]. Besides physicalcues from surface topography, Sharma et al. also investigatedhydrogels with varying gradients of mechanical compliances andfound the appropriate stiffness range that was conducive fortenogensis [24]. Furthermore, the comparative study of collagenousdecellurized matrices of different origins showed that tendon-derived matrix possessed native mechanical properties andtopography, which is conducive for tenogenesis of tendon stemcells [25]. Although the decellularized matrix is more complicatedthan articial scaffolds, the role of physical architecture was suc-cessfully investigated by comparing the inductive effects of cross-cut and longitudinally-cut tendon sections on MSCs, so as todiscern the biochemical cues within bio-matrix [26]. The difference

ned and randomly-oriented nanobrous scaffolds with or without Y-27632 for 3 days,F-actin staining in MSCs cultured on the aligned and randomly-oriented nanobroused images with DAPI staining. Scale bars, 20 mm. Quantitative PCR analysis of tenogenicor without Y-27632 and cyto D on day 3. Gene expression levels are normalized to the

rialsZ. Yin et al. / Biomatebetween nanometer scale and micrometer scale architecture in-uence cell behavior signicantly in terms of proliferation anddifferentiation [27e29]. The diameters of around 700 nm and1000 nm bers used in this study are belongs to similar micro-scaleand would not cause signicant effect on cell activities. Furthercomparison studies of topography need to be carried out onnanometer scale bers if they can be manufactured with sophisti-cate control of diameters and alignment. Based on the aforemen-tioned studies, elongated morphology and alignment orientation isalways associated with tenogenic differentiation, whereas spreadcell shape and random orientation is associated with osteogenesis[6]. Since the cytoskeleton is known to play important roles inmaintaining cell morphology, this study demonstrated that cyto Dand Y-27632 treatment caused signicant cell morphologicalchanges followed by loss of lineage commitment. Divergent pat-terns of MSC adhesion on different substrate topography reectedvarying cytoskeletal tension, as focal adhesions form the anchorpoints of the cytoskeleton [30]. The ber topography altered thepattern of cellesubstrate interaction and such changes in cyto-skeletal rearrangements will in turn lead to changes in intracellularmechanotransductive pathways, such as demonstrated by changesin the Rho A-ROCK pathway of stem cells in response to material

Fig. 10. Immunouorescence staining of SCX and F-actin in MSCs cultured on the aligned andays, with the right column being the merged images with DAPI staining. Scale bars, 20 mm44 (2015) 173e185 183stiffness and alignment [23,31]. Just like cells in tendon tissue, webelieve that spindle-shape morphology is essential for full teno-genesis but is not sufcient by itself. Tong et al. used nano-imprinting to replicate the physical topography and elasticity oftendon matrix so that the resulting shape and alignment ofcultured MSCs were similar to that seeded on longitudinally-cutsection of tendons [26]. Nevertheless, this was not accompanied bysignicantly increased Tnmd expression [26]. However, collagen Icoating of the bioimprint could effectively induce Tnmd expression.It indicated that alignment directly inuenced early teno-lineagecommitment but an additional second signal is required to com-plete the full differentiation process [32]. In our study, the in vivomicroenvironment provides signals for further differentiation andregeneration. The synergistic benecial effect of growth factors,such as broblast growth factor-2 (FGF2), bone morphogeneticprotein-12 (BMP12), platelet-derived growth factors (PDGFs),together with mechanical stimulation can potentially be put togood use in tendon differentiation [14,32e35]. Collectively, thesedata strongly suggests that the integration of topographical cueswith chemical stimuli can facilitate novel scaffold fabrication andtheir application in tissue engineering to achieve functionalhealing.

d randomly-oriented nanobrous scaffolds with or without exposure to Y-27632 for 3.

rialsZ. Yin et al. / Biomate1845. Conclusion

This study contributes to understanding of the biological effectsof physical cues from aligned and randomly-oriented nanobrousscaffolds, not only on the aspect of cellular behavior, but also ontissue formation in vivo. The aligned brous scaffold displayspromising results in tendon-like tissue regeneration at early repairstage, while in the randomly-oriented brous scaffold group, weobserved the development of bone formation at the injury site. Thetwo topographically-different scaffolds not only support MSCadhesion and spreading, but also induced tenogenesis and osteo-genesis respectively, both in vitro and in vivo. Moreover, we foundthat this topography-induced lineage commitment is dependent oncytoskeleton-mediated mechanotransduction. These ndings thusprovide vital information for the development of the next-generation of stem cell and bio-scaffold interfaces in future tissueengineering applications.

Acknowledgment

This work was supported by NSFC grants (81330041, 81125014,31271041, 81401781, 81201396, J1103603). The Project Supportedby Zhejiang Provincial Natural Science Foundation of China

Fig. 11. ALP staining of MSCs cultured on the aligned and randomly-oriented nanobrous scashows MSC nuclei being stained with DAPI to quantify the cell number. The right columns44 (2015) 173e185(LR14H060001). The National Key Scientic Program(2012CB966604), the National High Technology Research andDevelopment Program of China (863 Program)(No.2012AA020503). Sponsored by Regenerative Medicine inInnovative Medical Subjects of Zhejiang Province. Medical andhealth science and technology plan of Department of Health ofZhejiang Province (2013RCA010). The Postdoctoral Foundation ofChina (2014M561775, 2014M551759). The Technology Develop-ment project (CXZZ20130320172336579) from the Science Tech-nology and Innovation Committee of Shenzhen Municipality. Weare grateful to The Core Facilities of Zhejiang University School ofMedicine for technical assistance.

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.biomaterials.2014.12.027.

References

[1] Discher DE, Mooney DJ, Zandstra PW. Growth factors, matrices, and forcescombine and control stem cells. Science 2009;324(5935):1673e7.

ffolds with or without exposure to Y-27632 for 7 days, respectively. The middle columnare merged images of the left and middle images. Scale bars, 100 mm.

[2] Chowdhury F, Na S, Li D, Poh YC, Tanaka TS, Wang F, et al. Material propertiesof the cell dictate stress-induced spreading and differentiation in embryonicstem cells. Nat Mater 2010;9(1):82e8.

[3] Gentleman E, Swain RJ, Evans ND, Boonrungsiman S, Jell G, Ball MD, et al.Comparative materials differences revealed in engineered bone as a functionof cell-specic differentiation. Nat Mater 2009;8(9):763e70.

[4] Engler AJ, Sen S, Sweeney HL, Discher DE. Matrix elasticity directs stem celllineage specication. Cell 2006;126(4):677e89.

[5] McMurray RJ, Gadegaard N, Tsimbouri PM, Burgess KV, McNamara LE, Tare R,et al. Nanoscale surfaces for the long-term maintenance of mesenchymal stemcellphenotype and multipotency. Nat Mater 2011;10(8):637e44.

[6] Dalby MJ, Gadegaard N, Tare R, Andar A, Riehle MO, Herzyk P, et al. Thecontrol of human mesenchymal cell differentiation using nanoscale symmetryand disorder. Nat Mater 2007;6(12):997e1003.

[7] Yin Z, Chen X, Chen JL, Shen WL, Hieu NTM, Gao L, et al. The regulation oftendon stem cell differentiation by the alignment of nanobers. Biomaterials2010;31(8):2163e75.

[8] Chen W, Villa-Diaz LG, Sun Y, Weng S, Kim JK, Lam RH, et al. Nanotopographyinuences adhesion, spreading, and self-renewal of human embryonic stemcells. ACS Nano 2012;6(5):4094e103.

[9] Cen L, Liu W, Cui L, Zhang W, Cao Y. Collagen tissue engineering: developmentof novel biomaterials and applications. Pediatr Res 2008;63(5):492e6.

[10] Liu Y, Ramanath HS, Wang DA. Tendon tissue engineering using scaffoldenhancing strategies. Trends Biotechnol 2008;26(4):201e9.

[11] Shen W, Chen X, Chen J, Yin Z, Heng BC, Chen W, et al. The effect of incor-poration of exogenous stromal cell-derived factor-1 alpha within a knittedsilk-collagen sponge scaffold on tendon regeneration. Biomaterials2010;31(28):7239e49.

[20] O'Brien EJ, Frank CB, Shrive NGB, Hallgrimsson, Hart DA. Heterotopic miner-alization (ossication or calcication) in tendinopathy or following surgicaltendon trauma. Int J Exp Pathol 2012;93(5):319e31.

[21] Bi Y, Ehirchiou D, Kilts TM, Inkson CA, Embree MC, Sonoyama W, et al.Identication of tendon stem/progenitor cells and the role of the extracellularmatrix in their niche. Nat Med 2007;13(10):1219e27.

[22] Pietschmann MF, Frankewycz B, Schmitz P, Docheva D, Sievers B, Jansson V,et al. Comparison of tenocytes and mesenchymal stem cells seeded onbiodegradable scaffolds in a full-size tendon defect model. J Mater Sci MaterMed 2013;24(1):211e20.

[23] Zhu J, Li J, Wang B, Zhang WJ, Zhou G, Cao Y, et al. The regulation of phenotypeof cultured tenocytes by microgrooved surface structure. Biomaterials2010;31(27):6952e8.

[24] Sharma RI, Snedeker JG. Biochemical and biomechanical gradients for directedbone marrow stromal cell differentiation toward tendon and bone. Bio-materials 2010;31(30):7695e704.

[25] Yin Z, Chen X, Zhu T, Hu JJ, Song HX, Shen WL, et al. The effect of decellu-larized matrices on human tendon stem/progenitor cell differentiation andtendon repair. Acta Biomater 2013;9(12):9317e29.

[26] Tong WY, Shen W, Yeung CW, Zhao Y, Cheng SH, Chu PK, et al. Functionalreplication of the tendon tissue microenvironment by a bioimprinted sub-strate and the support of tenocytic differentiation of mesenchymal stem cells.Biomaterials 2012;33(31):7686e98.

[27] Tong HW, Wang M, Lu WW. Electrospinning and evaluation of PHBV-basedtissue engineering scaffolds with different bre diameters, surface topographyand compositions. J Biomater Sci Polym Ed 2012;23(6):779e806.

[28] Bashur CA, Shaffer RD, Dahlgren LA, Guelcher SA, Goldstein AS. Effect of berdiameter and alignment of electrospun polyurethane meshes on mesen-chymal progenitor cells. Tissue Eng Part A 2009;15(9):2435e45.

[29] Hodgkinson T, Yuan XF, Bayat A. Electrospun silk broin ber diameter in-uences in vitro dermal broblast behavior and promotes healing of ex vivo

Z. Yin et al. / Biomaterials 44 (2015) 173e185 185[12] Neidert MR, Lee ES, Oegema TR, Tranquillo RT. Enhanced brin remodelingin vitro with TGF-beta1, insulin and plasmin for improved tissue-equivalents.Biomaterials 2002;23(17):3717e31.

[13] Yoshizawa T, Takizawa F, Iizawa F, Ishibashi O, Kawashima H, Matsuda A, et al.Homeobox protein MSX2 acts as a molecular defense mechanism for pre-venting ossication in ligament broblasts. Mol Cell Biol 2004;24(8):3460e72.

[14] Kishore V, Bullock W, Sun X, Van Dyke WS, Akkus O. Tenogenic differentiationof human MSCs induced by the topography of electrochemically alignedcollagen threads. Biomaterials 2012;33(7):2137e44.

[15] Czaplewski SK, Tsai TL, Duenwald-Kuehl SE, Vanderby Jr R, Li WJ. Tenogenicdifferentiation of human induced pluripotent stem cell-derived mesenchymalstem cells dictated by properties of braided submicron brous scaffolds.Biomaterials 2014;35(25):6907e17.

[16] Teh TK, Toh SL, Goh JC. Aligned brous scaffolds for enhanced mechanores-ponse and tenogenesis of mesenchymal stem cells. Tissue Eng Part A2013;19(11e12):1360e72.

[17] Kolambkar YM, Bajin M, Wojtowicz A, Hutmacher DW, Garcia AJ, Guldberg RE.Nanober orientation and surface functionalization modulate humanmesenchymal stem cell behavior in vitro. Tissue Eng Part A 2014;20(1e2):398e409.

[18] Lin L, Shen Q, Xue T, Yu C. Heterotopic ossication induced by Achillestenotomy via endochondral bone formation: expression of bone and cartilagerelated genes. Bone 2010;46(2):425e31.

[19] Yee LPP, Wong YM, Rui YF, Lee YW, Chan LS, Chan KM. Expression of chondro-osteogenic BMPs in ossied failed tendon healing model of tendinopathy.J Orthop Res 2011;29(6):816e21.wound models. J Tissue Eng 2014;5. 2041731414551661.[30] McBeath R, Pirone DM, Nelson CM, Bhadriraju K, Chen CS. Cell shape, cyto-

skeletal tension, and RhoA regulate stem cell lineage commitment. Dev Cell2004;6(4):483e95.

[31] Nikukar H, Reid S, Tsimbouri MP, Riehle MO, Curtis AS, Dalby MJ. Osteogenesisof mesenchymal stem cells by nanoscale mechanotransduction. ACS Nano2013;7(3):2758e67.

[32] Ker ED, Nain AS, Weiss LE, Wang J, Suhan J, Amon CH, et al. Bioprinting ofgrowth factors onto aligned sub-micron brous scaffolds for simultaneouscontrol of cell differentiation and alignment. Biomaterials 2011;32(32):8097e107.

[33] Wang W, Deng D, Li J, Liu W. Elongated cell morphology and uniaxial me-chanical stretch contribute to physical attributes of niche environment forMSC tenogenic differentiation. Cell Biol Int 2013;37(7):755e60.

[34] Caliari SR, Harley BA. Composite growth factor supplementation strategies toenhance tenocyte bioactivity in aligned collagen-GAG scaffolds. Tissue EngPart A 2013;19(9e10):1100e12.

[35] Cheng X, Tsao C, Sylvia VL, Cornet D, Nicolella DP, Bredbenner TL, et al.Platelet-derived growth-factor-releasing aligned collagen-nanoparticle berspromote the proliferation and tenogenic differentiation of adipose-derivedstem cells. Acta Biomater 2014;10(3):1360e9.

Electrospun scaffolds for multiple tissues regeneration in vivo through topography dependent induction of lineage specific ...1. Introduction2. Materials and methods2.1. Fabrication of PLLA scaffolds2.2. Morphology of PLLA scaffolds2.3. SEM imaging2.4. Alkaline phosphatase (ALP) staining2.5. Quantitative PCR2.6. Animal model2.7. Immunofluorescence2.8. Histological evaluation and staining2.9. Immunohistochemistry2.10. Mechanical testing2.11. Determination of collagen content2.12. Transmission electron microscopy2.13. Radiographic evaluation2.14. Statistical analysis

3. Results3.1. Fabrication and morphological characterization of scaffolds3.2. The effects of aligned and randomly-oriented scaffolds on neo-tissue formation3.2.1. Histology of repaired tendons3.2.2. Histology of bone formation3.2.3. Mechanical properties of repaired tendons

3.3. The effects of topographical cues on MSCs3.4. Topography-induced lineage commitment of MSCs is dependent on cytomyosin cytoskeleton

4. Discussion5. ConclusionAcknowledgmentAppendix A. Supplementary dataReferences