Biomaterials 61 (2015) 85e94

Contents lists available at ScienceDirect

Biomaterials journal homepage: www.elsevier.com/locate/biomaterials

Circumferentially aligned fibers guided functional neoartery regeneration in vivo Meifeng Zhu a, 1, Zhihong Wang a, 1, Jiamin Zhang a, Lina Wang a, Xiaohu Yang b, Jingrui Chen b, Guanwei Fan b, Shenglu Ji a, Cheng Xing a, Kai Wang a, c, Qiang Zhao a, Yan Zhu b, Deling Kong a, *, Lianyong Wang a, * a State Key Laboratory of Medicinal Chemical Biology, Key Laboratory of Bioactive Materials of Ministry of Education, Collaborative Innovation Center of Chemical Science and Engineering (Tianjin), College of Life Science, Nankai University, Tianjin 300071, China b Tianjin State Key Laboratory of Modern Chinese Medicine, Tianjin University of Traditional Chinese Medicine, Tianjin, China c State Key Laboratory of Molecular Engineering of Polymers, Fudan University, Shanghai 200433, China

a r t i c l e i n f o

a b s t r a c t

Article history: Received 22 January 2015 Received in revised form 2 May 2015 Accepted 14 May 2015 Available online 16 May 2015

An ideal vascular graft should have the ability to guide the regeneration of neovessels with structure and function similar to those of the native blood vessels. Regeneration of vascular smooth muscle cells (VSMCs) with circumferential orientation within the grafts is crucial for functional vascular reconstruction in vivo. To date, designing and fabricating a vascular graft with well-defined geometric cues to facilitate simultaneously VSMCs infiltration and their circumferential alignment remains a great challenge and scarcely reported in vivo. Thus, we have designed a bi-layered vascular graft, of which the internal layer is composed of circumferentially aligned microfibers prepared by wet-spinning and an external layer composed of random nanofibers prepared by electrospinning. While the internal circumferentially aligned microfibers provide topographic guidance for in vivo regeneration of circumferentially aligned VSMCs, the external random nanofibers can offer enhanced mechanical property and prevent bleeding during and after graft implantation. VSMCs infiltration and alignment within the scaffold was then evaluated in vitro and in vivo. Our results demonstrated that the circumferentially oriented VSMCs and longitudinally aligned ECs were successfully regenerated in vivo after the bi-layered vascular grafts were implanted in rat abdominal aorta. No formation of thrombosis or intimal hyperplasia was observed up to 3 month post implantation. Further, the regenerated neoartery exhibited contraction and relaxation property in response to vasoactive agents. This new strategy may bring cell-free small diameter vascular grafts closer to clinical application. © 2015 Elsevier Ltd. All rights reserved.

Keywords: Vascular grafts Vascular smooth muscle cells Poly(ε-caprolactone) Aligned microfiber Wet-spinning

1. Introduction The small diameter vascular grafts (SDVGs) are in considerable need for clinical replacement of damaged and blocked blood vessels [1]. The native blood vessels consist of tunica intima, media and adventitia from inner to outer with distinct patterns. The tunica intima is an antithrombogenic monolayer with longitudinally aligned endothelial cells (ECs). The tunica media contains multiple

* Corresponding authors. Institute of Molecular Biology, Nankai University, Weijin Road 94, Nankai District, Tianjin 300071, China. E-mail addresses: [email protected] (D. Kong), [email protected] (L. Wang). 1 These authors contributed equally to this work. http://dx.doi.org/10.1016/j.biomaterials.2015.05.024 0142-9612/© 2015 Elsevier Ltd. All rights reserved.

layers of vascular smooth muscle cells (VSMCs) and ECMs (mainly collagens and elastin) that are circumferentially aligned, which plays a crucial role in maintaining the mechanical strength and vasoactive responsiveness of the blood vessels [2]. Therefore, the ideal vascular grafts should have potential ability to regenerate new blood vessels with structure and function similar to those of the native blood vessels [3]. Cell alignment has been widely observed at various scales in native tissues and organs [4], which plays a critical role in maturation and regeneration of functional tissue. As reported, cell alignment could be achieved in vitro by means of the guidance role of various geometric cues including nano or micro-sized fibers [5,6] and channels [7,8]. For instance, based on the structural feature of native blood vessel, Jiang et al. developed a stress-induced rolling

86

M. Zhu et al. / Biomaterials 61 (2015) 85e94

membrane technique to fabricate tubular structures, in which VSMCs and ECs location and orientation could be precisely controlled. Using this approach, tubular grafts with longitudinally aligned endothelial cells and circumferentially aligned VSMCs were successfully obtained [7]. Moreover, Jone et al. prepared the circular microchannel embedded in the curvature surface of tubular structure by micro-fabrication technology, which realized circumferential alignment of human aortic smooth muscle cells (HASMCs) during in vitro culture [9]. Recently, Huang et al. prepared a composite tubular scaffold with circumferentially aligned poly (L-lactic acid) (PLLA) nanofibers in the inner surface. Their results indicated that the aligned PLLA fibers could guide VSMCs' orientation in vitro [10]. The above studies have demonstrated that VSMCs were aligned and elongated along the direction of geometric cues, and had significant advance in methodology for the design and fabrication of vascular graft which could regulate orientation and function of VSMCs in vitro. However, most of these studies were limited to in vitro model. Besides the VSMCs circumferential alignment, infiltration of VSMCs into the interior of the vascular scaffolds is another key factor for the regeneration of small-diameter blood vessels. For this objective, suitable pore size and highly interconnected porous structure of the scaffolds are essentially important [11,12]. Up to date, the vascular scaffolds with different pore structure have been developed by various techniques such as phase separation [13], salt leaching [14], electrospinning [15,16] and so on. For instance, He et al. fabricated a bi-layered elastomeric porous vascular scaffolds by thermally induced phase separation method and the VSMCs could infiltrate into the wall of the grafts in vivo [13.] Wu et al. reported that a porous tubular poly (glycerol-sebacate) scaffolds fabricated by salt leaching method could realize sufficient cell infiltration in vivo [14]. In our previous study, the electrospun poly (ε-caprolactone) (PCL) vascular grafts with thick fibers and large pore obviously enhanced cell infiltration into the grafts wall compared with the grafts of thin fibers [16]. Despite some encouraging results have been obtained, the regenerated VSMCs in these works arranged randomly due to the lack of well-defined aligned geometric cues. How to design and fabricate a vascular graft with well-defined aligned geometric cues to realize simultaneously VSMCs infiltration and circumferential alignment in vivo remains a great challenge [17]. To our knowledge, there are no reports on this issue, particularly in vivo [18]. Numerous reports have demonstrated that the host is not only a superior source of cells and but also an all-round “bioreactor” [18], the host selfremodeling capability could promote in vivo neoartery regeneration from the cell-free vascular grafts [14,19]. It suggests that the ideal vascular grafts should be well adapted to implantation environment, and coordinate with the host self-remodeling ability to enhance vascular regeneration. Wet-spinning is a simple and straightforward method for fabricating fibrous scaffolds with fibers in microscale [20]. In wetspinning process, the polymer solution is extruded into a coagulation bath to precipitate the polymer in the form of a fiber because of solvent diffusion. The polymer fibers are collected onto collectors to form fibrous scaffolds. Wet-spun microfibers have attracted considerable interest as a scaffold matrix to guide and direct the behavior of cells for a variety of applications, such as cartilage, tendon and ligament tissue engineering [21]. However, there are few reports about the in vivo application of vascular grafts fabricated using wet-spinning. Additionally, PCL is a non-toxic, biocompatible polyester biomaterial that has been used to fabricate small diameter vascular grafts [22]. Inspired by the results mentioned above, we hypothesize that simultaneous VSMCs infiltration and circumferential alignment could be achieved under the guidance of grafts with

circumferentially oriented microfibers, and the regenerated VSMCs could facilitate endothelium formation and its maturation upon implantation in rat abdominal aorta (Fig. 1a). In this study, we designed a bi-layered vascular graft. The internal layer is composed of circumferentially aligned microfiber prepared by wet-spinning, which could provide a topographic guidance for the regeneration of circumferentially aligned VSMCs. The external layer is composed of random nanofibers prepared with electrospining method in order to enhance the mechanical property and prevent bleeding during implantation. Furthermore, we evaluated cell infiltration and alignment in the scaffolds in vitro, and assessed the feasibility of the circumferentially oriented regeneration of VSMCs in vivo by implantation of the bi-layered grafts into rat abdominal aorta. The goal of this study was to develop a small-diameter vascular graft with the ability to induce the regeneration of circumferentially aligned VSMCs and thus mimic the organization of native blood vessels. 2. Materials and methods 2.1. Materials Poly (ε-caprolactone) (PCL) pellets (Mn ¼ 70,000e90,000) were purchased from Sigma (USA). Analytical reagents including chloroform, tetrahydrofuran, trifluoroethanol and hexane were

Fig. 1. Schematic illustration showed the hypothesis and fabrication process in the present study. (a) Our hypothesis is that the circumferentially aligned microfibers of the grafts could guide VSMCs regeneration in circumferential orientation within the internal layer of the grafts in vivo, and complete endothelialization could be achieved rapidly. VSMC is abbreviation of vascular smooth muscle cell; EC is abbreviation of endothelial cell. (b) The fabrication process of vascular grafts. The internal layer composed of circumferentially oriented fibers which provide a topographic guidance for VSMCs regeneration was firstly prepared by wet-spinning method. Then the external layer composed of randomly oriented nanofibers was prepared by electrospinning method, which could enhance the mechanical property of the grafts and prevent from bleeding in implantation.

M. Zhu et al. / Biomaterials 61 (2015) 85e94

obtained from Tianjin Chemical Reagent Company (Tianjin, China). Triton X-100 was purchased from Alfa Aesar (London, England). Soybean Salad oil was purchased from Walmart Shop. Sprague Dawley rats (male, weight 280e320 g) were purchased from the Laboratory Animal Center of the Academy of Military Medical Sciences (Beijing, China). 2.2. Graft fabrication Vascular grafts consisting of an external layer and an internal layer were fabricated by the process shown in Fig. 1b. Firstly, the internal layer with circumferentially orientated PCL fibers was prepared by a wet spinning technique. PCL solution (10%, w/v) in chloroform/tetrahydrofuran (3:1, v/v) was transferred into a 10 ml glass syringe equipped with a 7-G injection needle which was submerged into a coagulation bath of oil/hexane solution (3:2, v/v). Driven by a syringe pump, the spin trickle of PCL solution was extruded into the coagulation bath at flow rate of 2 ml/h. The spin trickle transformed into micro-sized fibers as the solvent diffuses through the rotating coagulation bath. The fibers were collected by a rotating stainless steel rod (1.8 mm in diameter) at 2000 rpm for 8 min. Then the fibers wrapped stainless steel rods were washed 3 times in hexane to remove excess oil, and dried in a vacuum desiccator for 2 days to remove residual hexane. The fibers wrapped rod was installed in electrospinning apparatus, and the external layer with randomly arranged PCL nanofibers was prepared by the following electrospinning process. PCL solution (14%, w/v in trifluoroethanol) was loaded into a 10 ml syringe with a 21-G needle. Electrospinning was performed under 16 kV voltage with a receiving distance of 10 cm for 3 min. The prepared grafts were dried in vacuo for 2 days to remove the residual organic solvent. Prior to implantation, the grafts were sterilized by immersing in 75% ethanol for 1 h, and washed 5 times with physiological saline. 2.3. Mechanical tests Mechanical properties of the grafts in longitudinal and radial directions were measured on a tensile-testing machine (Instron 5865, USA) with a load capacity of 100 N. Grafts with 3 cm in length were clamped with a 1 cm inter-clamp distance and pulled longitudinally at a rate of 1 mm/s until rupture (n ¼ 3). For radial direction mechanical properties measurement, grafts with 0.3 cm in length were fixed on two steel rings which were clamped by machine chuck, and then pulled radially at a rate of 1 mm/s until rupture (n ¼ 3). The radial direction mechanical properties of the explanted grafts at 12 weeks (n ¼ 3) with 0.3 cm in length were also measured using the same test conditions. Maximum stress and strain at rupture were measured. The Young's modulus representing the elasticity was obtained by measuring the slope of the stressestrain curve in the elastic region. Burst pressure was measured with a self-made instrument by filling a graft of 1 cm in length with Vaseline, fastening one end and hermetically closing the other end to a vascular graft. A constant filling rate of 0.1 ml/min was applied, and the filling pressure was recorded until the graft wall burst. Grafts with（n ¼ 3）or without（n ¼ 3）external layer were both tested. 2.4. Scaffold porosity, fiber and pore size measurements Scaffold porosity was measured by a liquid intrusion method [16]. Dry grafts were weighed and then immersed in 100% ethanol for 2 h. Grafts were then gently wiped to remove excess ethanol and weighed again. Graft porosity was calculated with the following equations:

Veth ¼




Wwet  Wdry

87

. . Deth ; VPCL ¼ Wdry DPCL ;

. Scaffold porosity ¼ Veth ðVPCL þ Veth Þ  100% [Veth is the volume of ethanol entrapped in the graft; Wdry and Wwet are the dry and wet weights of the grafts; DPCL is the density of PCL (1.145 g/mm3) and Deth is the density of ethanol (0.789 mg/ mm3)]. The average fiber diameter were measured based on scanning electron microscopy (SEM) images. For each sample, five SEM images were analyzed, and at least 50 fibers and pores were manually measured on each image and analyzed using Image J software. In order to measure the pore size precisely, the bare grafts were embedded and frozen in optimum cutting temperature compound (OCT), then were sectioned to 10 mm in thickness. Pore sizes were analyzed by measuring the distance among the fibers (n ¼ 50) (Supplementary Fig. 2). Results are expressed as mean ± standard deviation. 2.5. Cell culture and evaluation in vitro The scaffolds were cut into circular discs and sterilized by immersing in 75% ethanol for 1 h, then washed 3 times with PBS. VSMCs (ScienCell, USA) were seeded onto the circular discs and cultured in DMEM-F12 medium (Gibco Invitrogen, USA) with 10% fetal bovine serum. After culture 3 days, cell morphology was observed by SEM. The cytoskeleton organization was tested by staining with rhodamine-phalloidin (SigmaeAldrich, CA, USA) and DAPI (SigmaeAldrich, CA, USA). Finally, images were taken by laser scanning confocal microscope (Leica, Germany). The cell viability analysis was carried out by live/dead (Invitrogen, American) staining of seeded cells after culture for 3 days. To evaluate the cell infiltration inside the scaffolds, VSMCs were labeled with DiI (orangeered fluorescent dye, Invitrogen) prior to seeding onto the scaffolds. The cell-seeded samples were harvested, embedded and cut into frozen sections with 10 mm thickness. Cell infiltration was observed using upright microscope (Leica, Germany). 2.6. In vivo implantation Twentyeone Sprague Dawley rats were used in this study. Animal experiment was approved by the Animal Experiments Ethical Committee of Nankai University and complied with the Guide for Care and Use of Laboratory Animals. In brief, the rats were anesthetized with chloral hydrate (300 mg/kg) by an intraperitoneal injection. Heparin (100 unit/kg) was used for anticoagulation by tail vein injection before surgery. A midline laparotomy incision was made and the abdominal aorta was isolated, clamped, and transected. The tubular PCL grafts (inner diameter 1.8 mm, length 1.0 cm) were sewed in an end-to-end fashion with 8e10 interrupted stitches using 9-0 monofilament nylon sutures (Yuan Hong, Shanghai, China). The wound was closed with 3-0 monofilament nylon sutures. No anticoagulation drug was administered after surgery. At the predetermined time points (2, 4 and 12 weeks), the patency of the grafts was visualized by high-resolution ultrasound (Vevo 2100 System, Visualsonics, Canada) after the rats were anesthetized with isoflurane. Grafts were explanted under anesthetization, and then animals were sacrificed by injection of overdose chloral hydrate. For endothelialization analysis, three explanted grafts at each time points were longitudinally cut into two parts and observed by stereomicroscope, one part was fixed with 4% paraformaldehyde for longitudinal frozen section. The other part was fixed with 2.5%

88

M. Zhu et al. / Biomaterials 61 (2015) 85e94

glutaraldehyde for SEM examination. For other histological staining analysis, the residual grafts were rinsed with saline and radially cut into two parts from the middle, each part was first observed by stereomicroscope, and then fixed with 4% paraformaldehyde for frozen cross-section. 2.7. Histological analysis of the explanted vascular grafts For SEM analysis, the explants were rinsed with PBS and fixed with 2.5% glutaraldehyde overnight, and dehydrated in ascending series of ethanol. Samples were sputter-coated with gold and observed by SEM (Quanta200, Czech). For sectioning and staining, the explants were fixed with 4% paraformaldehyde, dehydrated by 30% sucrose solution until the grafts sank to the bottom. The explants were sectioned to 6 mm in thickness after embedded in OCT. Subsequently, the sections were stained with hematoxylin and eosin (H&E), Masson's trichrome, Verhoeffevan Gieson (VVG), Safranin O and Sirius red (Zhongshan Golden bridge Biotechnology, China). Images were observed under upright microscope (Leica DM3000). To perform immunofluorescent staining, the frozen sections were rinsed once with 150 mM PBS. Then the slides were incubated in 5% normal goat serum (Zhongshan Golden bridge Biotechnology, China) for 30 min at room temperature. For intracellular antigen staining, 0.1% Triton-PBS was used to permeate the membrane before incubation with serum. Then the sections were incubated with primary antibodies in PBS overnight at 4  C, followed by incubation with secondary antibody in PBS for 2 h at room temperature. The nuclei were counterstained with 4,6-diamidino-2phenylindole (DAPI) containing mounting solution (Dapi Fluoromount G, Southern Biotech, England). The spreading area (S) and perimeter (L) of the nuclei at each time points were measured with Image J software. The nuclear shape index (NSI) was calculated as (4ᴫS/L2). Nuclei with a linear and elongated morphology have NSI approaching 0. While nuclei with nearly circular shape have an NSI close to 1. Endothelial cell staining was performed using rabbit anti-von Willebrand factor (vWF, 1:200, Dako, USA) primary antibody. The smooth muscle cells were stained using mouse anti-a-SMA (aSMA, 1:100, Boster, China) and mouse anti-smooth muscle myosin heavy chain I (SM-MHC, 1:200, Abcam, USA) primary antibody. To observe inflammatory cells in the explanted grafts, anti-CD68 antibody (CD68, 1:100, Abcam, USA) was used. To visualize M2 macrophages, anti-Mannose Receptor antibody (CD206, 1:200, Abcam, USA) was used. Elastin was stained using rabbit polyclonal anti-rat elastin (1:200, Abcam, USA) primary antibody. Alexa Fluor 488 goat anti-rabbit IgG (1:200, Invitrogen) and Alexa Fluor 594 goat anti-mouse IgG (1:200, Invitrogen) were used as the secondary antibodies, respectively. The sections without incubation with primary antibodies were used as negative controls. Slides were observed under a fluorescence microscope (Zeiss Axio Imager Z1, Germany). 2.8. Assessment of vascular function Aortic ring bioassay was performed as described below. After the grafts were dissected and cleaned from connective and fat tissues, the rings of 3 mm in length were obtained and bathed in the standard Krebs buffer (composition in mM: NaCl, 118.4; KCl, 4.7; CaCl2, 2.5; MgSO4, 1.2; KH2PO4, 1.2; NaHCO3, 25; dextrose, 11.1; Na2Ca EDTA, 0.029; pH 7.4) at 37  C and gassed with carbogenic mixture (95% O2 and 5% CO2). All preparations were stabilized under a resting tension of 2.00 g for 1 h with the buffer changed every 15 min. The presence of functional smooth muscle cells was indicated by the contractile responses induced by adding KCL

(60 mM). The function of the neo-endothelium was confirmed by the relaxation by acetylcholine (Ach, 10 mM) in pre-constricted segments by adrenaline (AD, 1 mM) and the vascular smooth muscle function was evaluated by vascular relaxation in response to sodium nitroprusside (SNP, 1  107 mol/L). Isometric forces were recorded with force transducers connected to a PowerLab/870 Eight-channel 100 kHz A/D converter (AD Instruments, Sydney, Australia). Results were obtained from 6 individual rings derived from the explants from 3 animals. 2.9. Statistical analysis All quantitative results were obtained from at least three samples for analysis. Data were expressed as the mean ± standard error of the mean (SEM). The two-tailed unpaired t-test was used to compare the differences. A value of p < 0.05 was considered to be statistically significant. 3. Results 3.1. Structure and properties of the grafts A schematic which illustrates our proposed concept and fabrication process is presented in Fig. 1. The bi-layered grafts were fabricated using PCL first by wet-spinning which produced the internal layer (Fig. 2a), then followed by electrospinning to generate the external layer. No kinking formed when the bi-layered grafts were folded at 180 (Fig. 2b), while the electrospun grafts with randomly arranged PCL nanofibers could form dead folding when it was bent (Supplementary Fig. 1). There was no visible separation between two layers (Fig. 2c, e). The internal layer and external layer consisted of circumferentially oriented microfibers (Fig. 2d) and randomly oriented nanofibers (Fig. 2f), respectively. The statistical analysis from SEM images showed that the thickness is 360.82 ± 33.21 mm for the internal layer and 40.21 ± 8.32 mm for the external layer. The fiber diameter was 18.26 ± 6.13 mm in the internal layer, and 0.65 ± 0.46 mm in the external layer. Crosssection of the internal layer showed that the macropores among the microfibers were highly interconnected (Fig. 2e， Supplementary Fig. 2), and the average pore size was 27.04 ± 9.05 mm. These interconnected macropores were beneficial to efficient cell infiltration and orientation. The external layer largely improved transverse maximum stress of grafts from 2.39 ± 0.14 MPa to 3.10 ± 0.26 MPa (Fig. 2g, h) (p < 0.05), and significantly increased longitudinal maximum stress from 0.38 ± 0.06 MPa to 1.89 ± 0.10 MPa (Fig. 2i, j) (p < 0.05). Besides, the elastic modulus and burst pressure of the grafts were significantly increased after adding the external layer (Fig. 2k, l). 3.2. VSMCs behaviors in vitro on the microfibers scaffolds Cell alignment, morphology, infiltration and live/dead assay in the scaffolds were performed in vitro. Confocal microscopy images showed that the cytoskeleton of VSMCs stained by rhodaminephalloidin were aligned after cultured for 3 days (Supplementary Fig. 3a); SEM images clearly demonstrated that VSMCs spread along the oriented fibers (Supplementary Fig. 3b); Cross section images showed that DiI labelled VSMCs could infiltrate inside the scaffolds (Supplementary Fig. 3c); Dead/live staining showed high viability of VSMCs in the scaffold (Supplementary Fig. 3d). 3.3. Cellularization and oriented VSMCs regeneration in vivo Cell infiltration and VSMCs circumferential regeneration were detected by H&E and immunofluorescent staining. The distribution

M. Zhu et al. / Biomaterials 61 (2015) 85e94

89

Fig. 2. Characterization of structure and mechanical properties of the grafts. (a), Gross morphology of the internal layer of graft by stereomicroscope; (b), no kinking formation when folded at 180 ; (c), SEM images of the grafts showing the microstructure of the two layers in the cross section; (d), Luminal surface of the graft showing the circumferentially oriented microfibers in the internal layer; (e), SEM image of longitudinal section of the graft wall showing macropores generated among the microfibers, and there was no visible separation between the two layers (red arrow); (f), The external layer of the graft with randomly distributed nanofibers. The stressestrain curves (g) and the maximum stress (h) of various grafts in the transverse direction; The stressestrain curves (i) and the maximum stress (j) of the grafts in the longitudinal direction. Arrows indicates the direction of force (F); The two layered grafts exhibited much greater elastic modulus (k) and burst pressure (l) than the grafts without the external layer. Scale bar for (a) 1 mm; (b) 2 mm; (c) 400 mm; (d) 300 mm; (e) 100 mm; (f) 50 mm. I þ E: grafts composed of internal layer and external layer; I: grafts composed of only internal layer; NEO refers to the explanted grafts at 12 weeks; NAT refers to native artery. The asterisks indicate statistical significant: p < 0.05 (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.).

of cells and ECM in grafts resembled that of native arteries (Fig. 3a, b). At 2 weeks, the orientated porous structure of the grafts allowed extensive cell penetration into the whole graft wall (Supplementary Figs. 4 and 5). Most of the cells were organized into circumferential layers containing a-SMAþ and a-SMA cells, but few cells express anti-smooth muscle myosin heavy chain I (SM-MHC) at this early stage(Fig. 3d). At 4 weeks, more cells infiltrated into the graft wall, and the orientation arrangements were more compactly and orderly (Fig. 3b, c, d). The number of SM-MHCþ cells rapidly increased within the graft wall. By 12 weeks, the number of aligned SM-MHCþ cells maintained a similar level to that at 4 weeks (Fig. 3g). Cell alignment inside the internal layer of grafts correlated well with nuclear elongation (Fig. 3e). The nuclear shape index of VSMCs at 2, 4, 12 weeks were 0.53 ± 0.12, 0.48 ± 0.11, 0.49 ± 0.11, which were slightly lower than the value 0.55 ± 0.16 in the native artery (Fig. 3h). The percentage of aligned VSMCs nuclei at 2, 4 and 12 weeks were 42.34 ± 6.23%, 47.12 ± 5.67%, 48.09 ± 5.34%, which increased over time and approached the value 52.83 ± 4.94 in the native artery (Fig. 3i). In addition, there were no change of luminal diameters occurred over the implantation time, which suggests the absence of aneurysm and restenosis within the grafts (Fig. 3a, f). Taken together, these data revealed a significant effect of the circumferentially oriented fibers in driving cell elongation. Macrophages play a critical role in the neo-artery regeneration process. Thus, we also detected the distribution and phenotypes of

macrophages at the different time points by immunofluorescence staining. CD68þ macrophages were observed in the graft wall with uniform distribution over time (Supplementary Fig. 6a). CD206þ macrophages, which stand for the M2 phenotype macrophage, were found within the graft wall at 2, 4 and 12 weeks (Supplementary Fig. 6b). The M1 phenotype macrophages, detected by marker CCR7 and iNOS, were observed within the graft wall at all times (Supplementary Fig. 6c and 6d). 3.4. ECM composition and organization In order to detect the organization and composition of ECM in the regenerated neoartery, histological staining and immunofluorescent staining for collagen and elastin were performed. The Sirius red and Masson staining for collagen showed that fibrillar collagen had an obvious circumferential orientation, which was similar to the native arteries' arrangement, although its compactness was lower than that in native arteries (Fig. 4a, Supplementary Fig. 7a). The composition of collagen including types I and III were observed in the internal layer of grafts over time (Fig. 4b, c). The elastin could be detected by immunofluorescent staining at 4 and 12 weeks, while very few elastin fibers were observed at 2 weeks (Fig. 4d, Supplementary Fig. 7b). Safranin O staining exhibited that the glycosaminoglycan distributed within the internal layer of grafts over time (Fig. 4e). Consistently, the transverse maximum stress of

90

M. Zhu et al. / Biomaterials 61 (2015) 85e94

Fig. 3. VSMCs regeneration in vivo with circumferential orientation within the graft. (a), H&E staining showed that cells quickly infiltrated into the entire graft wall as early as 2 weeks without a sign of regression after 12 weeks; (b), Magnified views of the graft wall showed defined cell alignment along the circumferentially oriented fibers. (c, d), Circumferential regeneration of smooth muscle cells was detected by a-SMA antibody (c, green) and SM-MHC antibody (d, green) staining. SM-MHC is the marker of contractile smooth muscle cells; (e), DAPI staining showed the shape of cell nuclei in the explanted graft; (f), Luminal diameter of the explanted grafts showed no restenosis and aneurysm formation at all the time points; (g), The number of SM-MHC positive cells dramatically increased from week 2 to week 4, and maintained in the same level until week 12. (h), Nuclear shape index (circularity) and (i) Percentage of aligned cell nuclei of the explanted graft. All the immuofluorescent micrographs were counterstained for nuclei by DAPI (blue). Data represent mean ± standard deviation for g, h, i and j. Scale bar: (a) 500 mm; (b) 100 mm; (c,d, f) 50 mm. The red dotted line denoted the interface between internal layer and external layer of the graft. For native artery, it denoted the interface between media and adventitia (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.).

explanted grafts at 12 weeks was significantly increased compared with the unimplanted grafts (Fig. 2h). 3.5. Rapid endothelialization Under stereomicroscope, the luminal surfaces of the grafts at 2, 4 and 12 weeks were clean and smooth (inserts in Fig. 5a, Supplementary Fig. 8). Furthermore, the endothelialization of the grafts was examined by SEM and vWF immunofluorescent staining. As shown in SEM images, the luminal surfaces were free of platelet aggregates and thrombosis. At 2 weeks, about 70% of the lumen surface was covered by cobble stone-like cells which were vWFpositive endothelial cells (Supplementary Fig. 8). At 4 weeks, the endothelial cells became more organized along the direction of blood stream and complete endothelialization was achieved, which is similar to the native one (Fig. 5a, b, c). Immunostaining confirmed that the lumen of the grafts was completely covered by vWF positive cells (Fig. 5d). Similar results were observed at 12 weeks (Supplementary Fig. 8). New capillary formation was observed in the internal layer of the grafts at all the time points (Fig. 5e, f, Supplementary Fig. 8). 3.6. Doppler ultrasound and function of the regenerated neoartery Laser Doppler ultrasound imaging of the grafts indicated excellent patency, which confirmed the absence of thrombus and restenosis (Fig. 6a). In addition, regular and synchronous pulsation

was detected from the implanted grafts at 12 weeks (Supplementary Video 1). Immunostaining of ECs and VSMCs indicated a confluent endothelial monolayer covering the entire lumen and separating the smooth muscle layer from the blood, which is similar to native artery (Fig. 6b, c). Velocity of blood flow in the grafts was similar to that in the native abdominal aorta (Fig. 6d). Supplementary data related to this article can be found online at http://dx.doi.org/10.1016/j.biomaterials.2015.05.024. The physiological functions of the regenerated smooth muscle and endothelium layers were examined by standard dual pin wire myography after 12 weeks. The explanted grafts displayed apparent vascular function of contraction and relaxation, indicating that the explants were responsive to various vasoactivators (Fig. 6e). The contraction of regenerated VSMCs was in response to the vasomotor agonists, KCL and AD, though the magnitude of constriction was less than that of native neoartery (0.29 ± 0.02 g vs. 2.20 ± 0.04 g for KCL, and 0.51 ± 0.01 g vs. 2.82 ± 0.10 g for AD, p < 0.05) (Fig. 6f). The explanted grafts also displayed vasodilation to endothelium-independent vasodilators including Ach and SNP (Fig. 6g). Besides, the regenerated neoartery displayed vasodilation to the endothelial-specific activator acetylcholine (Ach), suggesting the presence of signaling capability between ECs and contractile VSMCs. These data demonstrated that VSMCs and ECs layers were sensitive to specific physiologically relevant activators and can translate these signals into vasoactivity, suggesting that the regenerated neoarteries were healthy and functional.

M. Zhu et al. / Biomaterials 61 (2015) 85e94

91

Fig. 4. ECM content and organization within the graft over time. (a), Sirius red (red) staining shows the organization and distribution of collagen; Immunofluorescent staining shows the distribution of (b) collagen I (green), (c) collagen III (green), and (d) elastin (green); (e) Safranin O staining shows the presence of glycosaminoglycan (red). Scale bar: (a) 100 mm; (b, c, d) 50 mm and (e) 100 mm (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.).

4. Discussion In general, tissue engineering scaffold could provide a microenvironment to support in vitro cell attachment, proliferation and differentiation. In addition, the scaffolds play an important role as a template to guide in vivo regeneration of new tissues which possess

similar structure and function to native tissues [23]. For the vascular regeneration in vivo, the vascular grafts should have the ability to regenerate neovessels mimicking the microstructural organization of VSMCs and ECs in the native arteries [2,24]. The circumferential alignment of VSMCs is crucial for maintaining appropriate mechanical properties and vasoacitivity for the long-

Fig. 5. Endothelialization of the grafts at week 4. (a), SEM of explant which was cut longitudinally into two pieces. Higher magnification micrographs showed the luminal surface of native artery (b) and the middle part of the graft (c). (d), Coverage of lumen by endothelial cells (vWF, green). Higher magnification micrographs showed vWF positive cells and capillary formation in the middle part of the graft (e), and near the suture site in the graft (f). Nuclei counterstained by DAPI (blue). White arrows and red arrows in (e) and (f) denoted endothelial cells layer and capillary within the graft, respectively. Scale bar: for (a, d) 500 mm; (b) 100 mm; (c,e,f) 50 mm (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.).

92

M. Zhu et al. / Biomaterials 61 (2015) 85e94

Fig. 6. Graft patency and vascular physiological functions at week 12 after implantation. (a) The patency and synchronization of pulsation with adjacent host aorta were assessed by Laser Doppler ultrasound imaging. Synchronization of pulsation was also shown in Supplementary Video 1. Counterstaining images showed the distribution of endothelial cells (vWF, red) and vascular smooth muscle cells (a-SMA, green) in the explanted graft (b) at week 12 which are similar to that in the native thoracic aorta (c); (d), The values of blood velocity at distal end, graft and proximal end were close to each other, which indicated the graft was patent; (e), The examination of physiological functions of the neoarteries displayed a certain extent of sensitivity to vasodilators and vasoconstrictors; (f), Constriction of the neoarteries in response to both KCL and adrenaline, though the magnitude of constriction was less than native arteries; (g), Physiologic response of the explanted grafts to vasodilators. The explanted grafts were pre-constricted with adrenaline, and then assessed in their response to vasodilators. Neoarteries showed relaxation in response to the endothelial-specific activator acetylcholine and the vascular smooth muscle cell specific activator sodium nitroprusside (SNP). Scale bar: (b, c) 400 mm (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.).

term implantation of vascular grafts [25]. To achieve VSMCs regeneration, several kinds of vascular grafts with different architecture were designed and fabricated [7e9]. However, the evaluations of these grafts were mostly performed in cell culture experiments, and no reports demonstrated that fibrous vascular grafts could support simultaneous VSMCs infiltration and circumferential orientation in vivo. In the present study, we developed a new vascular graft with a two-layered structure. The internal layer consists of circumferentially aligned microfibers prepared by wetspinning, which could enhance cell infiltration and guide oriented VSMCs regeneration. The external layer consists of dense and random nanofibers prepared by electrospinning method, which provide adequate mechanical strength and prevent bleeding. The in vivo experiment of rat abdominal aorta replacement demonstrated that this kind of vascular grafts could maintain high patency without thrombosis and intimal hyperplasia. Circumferentially oriented VSMCs and longitudinally aligned ECs were successfully generated. Further, the regenerated neoarteries exhibited vascular physiological function of VSMCs and ECs at 3 months postimplantation. Up to date, two strategies including the typical tissue engineered vascular grafts (TEVGs) and cell-free vascular grafts, have been developed for the regeneration of small diameter blood vessels based on either decellularized matrix or biodegradable natural and synthetic material. The TEVGs were produced by seeding VSMCs and ECs on scaffold materials and cultured in a bioreactor in vitro. These TEVGs populated with VSMCs and ECs have demonstrated good patency upon in vivo implantation [18].

Nevertheless, it takes several weeks to make TEVGs, and need extra bioprocesses for storage, delivery, and handling, which limited its application. Compared with the TEVGs, cell-free vascular grafts could realize vascular regeneration in vivo by taking the advantage of host self-remodeling ability which possesses more remarkable advantages: avoidance of donor site morbidity, devoid of extensive in vitro cell culture, easy storage and transport, ready availability, and potentially closer clinical adoption. Electrospinning is a popular method to fabricate cell-free vascular grafts. However, cell infiltration into the grafts was hampered by the thin and dense electrospun nanofibers, which resulted in little VSMCs regeneration and little ECM secretion in the grafts [15,26]. To improve cell infiltration, a rapidly degradable vascular graft with pores about 20 mm in size has been prepared by salt leaching method using elastic poly (glycerol sebacate). The grafts successfully regenerated into new arteries within 90 days. However, the regenerated VSMCs were not well organized into circumferential layers at early stage and the generated neoarteries were not uniform in wall thickness [14]. More recently, an electrospun PCL vascular graft with thick fibers and large pores was prepared. Cell infiltration into the graft wall was greatly improved compared to the grafts with thin fibers. However, the regenerated VSMCs grew in a random arrangement on the luminal surface of grafts due to the lack of aligned cues [16]. Taken together, these findings suggest that VSMCs with circumferential alignment were hard to regenerate in vivo only by employing the host remodeling capacity. In the present work, we fabricated a scaffold with circumferentially aligned fibers in the internal layer (Fig. 2d). Our results demonstrated that a VSMCs

M. Zhu et al. / Biomaterials 61 (2015) 85e94

layer with circumferential orientation fashion was successfully regenerated upon implantation into rat abdominal aorta (Fig. 3). Interestingly, the regenerated VSMCs distributed evenly within the wall of the grafts instead of on luminal surface as reported by other studies [16,19]. Accordingly, we propose that two key factors may have played a key role in VSMCs circumferential regeneration in the scaffolds: large and interconnected pores in the scaffolds and the topographical guiding effect of the grafts. Large pores provided sufficient spaces for VSMCs infiltration, while the circumferential fibers could serve as cues for VSMCs circumferential growth. Another key issues on constructing a functional vascular tunica media is to realize SMCs appropriate phenotype transition at the proper stage during the regeneration process [27]. In general, there are two phenotypes of VSMCs in vascular tunica media: contractile phenotype and synthetic phenotype. VSMCs with a synthetic phenotype can rapidly proliferate and produce ECM, while VSMCs with a contractile phenotype contribute to maintain the function of tunica media. It is critical to achieve a contractile VSMCs phenotype at proper development stage. Otherwise, uncontrolled continuing proliferation of VSMCs in the implanted grafts will thicken the vessel wall and narrow the lumen, leading to failure of the implanted grafts [28]. In this study, a great number of a-SMAþ cells which are normally identified as early and mid-staged VSMCs grew along with the circumferentially oriented fibers at week 2 (Fig. 3c). While a large amount of SM-MHCþ cells which are defined as mature contractile VSMCs were observed at week 4 (Fig. 3d). Furthermore, the regenerated neoarteries showed notable VSMCs vasoactivity at week 12 (Fig. 6). These data indicated that the VSMCs phenotype switch was achieved in vivo in the grafts with aligned fibers. It was reported that the architecture of the scaffolds significantly regulate the phenotype and function of cells [2]. Many in vitro studies have proved that the gene expression associated to the contractile VSMCs was up-regulated when VSMCs were cultured on the aligned topography such as fibers or channels [6,8,29]. Our results gave further evidence that the circumferential alignment of fibers also could regulate VSMCs phenotype transition from synthetic to contractile in vivo. The organization, density and spatial geometry of ECM, especially collagen and elastin determine the mechanical strength of arteries [2]. Elastin offers compliance and elasticity, whereas collagen contributes mainly to arterial tensile strength. The aligned collagen type I/III and elastin in the regenerated neoarteries had phenotypic and structural similarities with the native arteries (Fig. 4). The transverse maximum stress of the explanted grafts at 3 months was greater than that of native arteries and the unimplanted grafts owing to the aligned ECM regeneration within the graft wall (Fig. 2h). The absence of rupture and aneurysmal dilatation in all grafts are most likely due to the circumferential regeneration of VSMCs and ECM. Other studies showed that aneurismal dilation often occurred if the porous structure and fibers of the grafts was randomly distributed [13,19]. A healthy, confluent endothelial layer is a prerequisite for a graft to resist thrombosis and smooth muscle cell proliferation that can lead to intimal hyperplasia. In order to enhance the endothelialization of small diameter vascular grafts, a variety of strategies have been employed including both physical and chemical modification to graft surface [30]. In this study, complete endothelialization was achieved after 4 weeks, and ECs exhibited mature and cobble stonelike morphology, which is similar to the native endothelium (Fig. 5b, c). Meanwhile, efficient cell infiltration and capillaries were observed within the graft wall after 2 weeks (Supplementary Figs. 4 and 8) due to the high interconnectivity of the scaffolds. Such a sufficient cellularization and capillary formation provided a suitable environment for rapid endothelium formation and maturation [31,32]. In addition, we found that the regenerated ECs were

93

organized in parallel to the flow direction although the orientation of fibers was perpendicular to the flow direction. It indicates that the fluid shear stress contributed greater effect than the fibers orientation on the alignment of the endothelial cells in vivo. Although our two-layered vascular grafts greatly facilitated the regeneration of VSMCs with a circumferential orientation, and the regenerated new arteries exhibited apparent vascular function of contraction and relaxation, it is just a study of small animal model. Given the slower vascular regeneration capability in large animals and human (especially those aged and diseased ones) than the small animals [33], tests in large animal models are needed to further evaluate this two-layered vascular grafts. Bioactive molecules (such as SDF-1, heparin) should be considered to be incorporated into the grafts to improve the vascular regeneration capability in large animals. Additionally, the source of VSMCs and ECs during the neoartery regeneration should be further investigated. Recent findings support the assumption that the differentiation of multipotent vascular stem cells, rather than the dedifferentiation of VSMCs, contributes to vascular remodeling [34]. ECs may come from the connected native arteries or endothelial progenitor cells in the blood [35]. It is particularly important that making clear the source of VSMCs and ECs will guide the design and construction of vascular grafts with optimal structure and bioactive modification to improve vascular regeneration and long-term patency. Additionally, we speculated that regeneration of complete new arteries could be realized if the residual PCL fibers degrade completely over time. In this regards, exploring grafts using polymers with faster degradation rate such as Poly (L-lactideco-ε-caprolactone) need further investigation. Overall, in the present study, a new electrospun PCL graft composed of circumferentially oriented microfibers was prepared. The aligned microstructure could induce rapid regeneration of tunica media in vivo, in which the regenerated VSMCs were well organized in circumferential orientation that is similar to the native arteries. More importantly, the regenerated neoarteries exhibited a notable vascular function of contraction and relaxation in response to vasoactive agents. This new strategy may bring cell-free small diameter vascular grafts closer to clinical application. Acknowledgment This study was supported by NSFC project (51173084, 81171478, 81274128, 31400833), National Basic Research Program of China (2011CB964903), and Program for Changjiang Scholars and Innovative Research Team in University (No. IRT13023), China Postdoctoral Science Foundation (No. 2014M561177, No. 2014M560183). Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.biomaterials.2015.05.024. References [1] D. Mozaffarian, E.J. Benjamin, A.S. Go, D.K. Arnett, M.J. Blaha, M. Cushman, et al., Heart disease and stroke statisticse2015 update: a report from the American Heart Association, Circulation 131 (2015) e29e322. [2] M.B. Chan-Park, J.Y. Shen, Y. Cao, Y. Xiong, Y. Liu, S. Rayatpisheh, et al., Biomimetic control of vascular smooth muscle cell morphology and phenotype for functional tissue-engineered small-diameter blood vessels, J. Biomed. Mater. Res. Part A 88A (2009) 1104e1121. [3] D.G. Seifu, A. Purnama, K. Mequanint, D. Mantovani, Small-diameter vascular tissue engineering, Nat. Rev. Cardiol. 10 (2013) 410e421. [4] Y. Li, G. Huang, X. Zhang, L. Wang, Y. Du, T.J. Lu, et al., Engineering cell alignment in vitro, Biotechnol. Adv. 32 (2014) 347e365. [5] C.Y. Xu, R. Inai, M. Kotaki, S. Ramakrishna, Aligned biodegradable nanotibrous structure: a potential scaffold for blood vessel engineering, Biomaterials 25

94

M. Zhu et al. / Biomaterials 61 (2015) 85e94

(2004) 877e886. [6] S. Rayatpisheh, D.E. Heath, A. Shakouri, P.-O. Rujitanaroj, S.Y. Chew, M.B. Chan-Park, Combining cell sheet technology and electrospun scaffolding for engineered tubular, aligned, and contractile blood vessels, Biomaterials 35 (2014) 2713e2719. [7] B. Yuan, Y. Jin, Y. Sun, D. Wang, J. Sun, Z. Wang, et al., A strategy for depositing different types of cells in three dimensions to mimic tubular structures in tissues, Adv. Mater. 24 (2012) 890e896. [8] C. Williams, X.Q. Brown, E. Bartolak-Suki, H. Ma, A. Chilkoti, J.Y. Wong, The use of micropatterning to control smooth muscle myosin heavy chain expression and limit the response to transforming growth factor beta 1 in vascular smooth muscle cells, Biomaterials 32 (2011) 410e418. [9] J.S. Choi, Y. Piao, T.S. Seo, Circumferential alignment of vascular smooth muscle cells in a circular microfluidic channel, Biomaterials 35 (2014) 63e70. [10] Y. Wang, H. Shi, J. Qiao, Y. Tian, M. Wu, W. Zhang, et al., Electrospun tubular scaffold with circumferentially aligned nanofibers for regulating smooth muscle cell growth, ACS Appl. Mater. Interfac. 6 (2014) 2958e2962. [11] Q.L. Loh, C. Choong, Three-dimensional scaffolds for tissue engineering applications: role of porosity and pore size, Tissue Eng. Part B: Rev. 19 (2013) 485e502. [12] B. Guo, B. Lei, P. Li, P.X. Ma, Functionalized scaffolds to enhance tissue regeneration, Regen. Biomater. 2 (2015) 47e57. [13] W. He, A. Nieponice, L. Soletti, Y. Hong, B. Gharaibeh, M. Crisan, et al., Pericytebased human tissue engineered vascular grafts, Biomaterials 31 (2010) 8235e8244. [14] W. Wu, R.A. Allen, Y. Wang, Fast-degrading elastomer enables rapid remodeling of a cell-free synthetic graft into a neoartery, Nat. Med. 18 (2012) 1148e1153. [15] S. de Valence, J.-C. Tille, D. Mugnai, W. Mrowczynski, R. Gurny, M. Moeller, et al., Long term performance of polycaprolactone vascular grafts in a rat abdominal aorta replacement model, Biomaterials 33 (2012) 38e47. [16] Z. Wang, Y. Cui, J. Wang, X. Yang, Y. Wu, K. Wang, et al., The effect of thick fibers and large pores of electrospun poly (epsilon-caprolactone) vascular grafts on macrophage polarization and arterial regeneration, Biomaterials 35 (2014) 5700e5710. [17] S.V. Murphy, A. Atala, 3D bioprinting of tissues and organs, Nat. Biotechnol. 32 (2014) 773e785. [18] S. Li, D. Sengupta, S. Chien, Vascular tissue engineering: from in vitro to in situ, Wiley Interdiscip. Rev.-Syst. Biol. Med. 6 (2014) 61e76. [19] S. Tara, H. Kurobe, M.W. Maxfield, K.A. Rocco, T. Yi, Y. Naito, et al., Evaluation of remodeling process in small-diameter cell-free tissue-engineered arterial graft, J. Vasc. Surg. (2014), http://dx.doi.org/10.1016/j.jvs. [20] E. Mathiowitz, D.M. Lavin, R.A. Hopkins, Wet spun microfibers: potential in the design of controlled-release scaffolds? Ther. Deliv. 4 (2013) 1075e1077.

[21] A. Tamayol, M. Akbari, N. Annabi, A. Paul, A. Khademhosseini, D. Juncker, Fiber-based tissue engineering: progress, challenges, and opportunities, Biotechnol. Adv. 31 (2013) 669e687. [22] T.K. Dash, V.B. Konkimalla, Poly-small je, Ukrainian-caprolactone based formulations for drug delivery and tissue engineering: a review, J. Control. Release: Off. J. Control. Release Soc. 158 (2012) 15e33. [23] S.J. Hollister, Porous scaffold design for tissue engineering, Nat. Mater. 4 (2005) 518e524. [24] P.W. Alford, A.P. Nesmith, J.N. Seywerd, A. Grosberg, K.K. Parker, Vascular smooth muscle contractility depends on cell shape, Integr. Biol. Quant. Biosci. Nano Macro 3 (2011) 1063e1070. [25] M. Parizek, K. Novotna, L. Bacakova, The role of smooth muscle cells in vessel wall pathophysiology and reconstruction using bioactive synthetic polymers, Physiol. Res. 60 (2011) 419e437. [26] A. Hasan, A. Memic, N. Annabi, M. Hossain, A. Paul, M.R. Dokmeci, et al., Electrospun scaffolds for tissue engineering of vascular grafts, Acta Biomater. 10 (2014) 11e25. [27] J. Qiu, Y. Zheng, J. Hu, D. Liao, H. Gregersen, X. Deng, et al., Biomechanical regulation of vascular smooth muscle cell functions: from in vitro to in vivo understanding, J. R. Soc. Interfac. 11 (2014), http://dx.doi.org/10.1098/rsif. [28] G.K. Owens, M.S. Kumar, B.R. Wamhoff, Molecular regulation of vascular smooth muscle cell differentiation in development and disease, Physiol. Rev. 84 (2004) 767e801. [29] Y. Cao, Y.F. Poon, J. Feng, S. Rayatpisheh, V. Chan, M.B. Chan-Park, Regulating orientation and phenotype of primary vascular smooth muscle cells by biodegradable films patterned with arrays of microchannels and discontinuous microwalls, Biomaterials 31 (2010) 6228e6238. [30] A.J. Melchiorri, N. Hibino, J.P. Fisher, Strategies and techniques to enhance the in situ endothelialization of small-diameter biodegradable polymeric vascular grafts, Tissue Eng. Part B-Rev. 19 (2013) 292e307. [31] L.P. Neff, B.W. Tillman, S.K. Yazdani, M.A. Machingal, J.J. Yoo, S. Soker, et al., Vascular smooth muscle enhances functionality of tissue-engineered blood vessels in vivo, J. Vasc. Surg. 53 (2011) 426e434. [32] L.E. Niklason, Building stronger microvessels, Nat. Biotechnol. 29 (2011) 405e406. [33] P. Zilla, D. Bezuidenhout, P. Human, Prosthetic vascular grafts: wrong models, wrong questions and no healing, Biomaterials 28 (2007) 5009e5027. [34] Z. Tang, A. Wang, F. Yuan, Z. Yan, B. Liu, J.S. Chu, et al., Differentiation of multipotent vascular stem cells contributes to vascular diseases, Nat. Commun. 3 (2012), http://dx.doi.org/10.1038/ncomms1867. [35] T. Merkulova-Rainon, D. Broqueres-You, N. Kubis, J.-S. Silvestre, B.I. Levy, Towards the therapeutic use of vascular smooth muscle progenitor cells, Cardiovasc. Res. 95 (2012) 205e214.