Generation of human umbilical cord vein CD146+ perivascular cell origined three-dimensional vascular construct

Microvascular Research 118 (2018) 101–112

Contents lists available at ScienceDirect

Microvascular Research journal homepage: www.elsevier.com/locate/ymvre

Generation of human umbilical cord vein CD146+ perivascular cell origined three-dimensional vascular construct Beyza Gökçinar-Yagcia,b, Nilgün Yersalc, Petek Korkusuzc, Betül Çelebi-Saltika,b,

T

⁎

a

Department of Stem Cell Sciences, Hacettepe University Graduate School of Health Sciences, 06100, Sihhiye, Ankara, Turkey Center for Stem Cell Research and Development, Hacettepe University, 06100, Sihhiye, Ankara, Turkey c Department of Histology and Embryology, Hacettepe University, Faculty of Medicine, 06100, Sihhiye, Ankara, Turkey b

A R T I C LE I N FO

A B S T R A C T

Keywords: CD146+ perivascular cell Collagen type I Elastin Fibrin Cell sheet engineering Vascular tissue engineering

Small-diameter vascular grafts are needed for the treatment of coronary artery diseases in the case of limited accessibility of the autologous vessels. Synthetic scaffolds have many disadvantages so in recent years vascular constructs (VCs) made from cellularized natural scaffolds was seen to be very promising but number of studies comprising this area is very limited. In our study, our aim is to generate fully natural triple-layered VC that constitutes all the layers of blood vessel with vascular cells. CD146+ perivascular cells (PCs) were isolated from human umbilical cord vein (HUCV) and differentiated into smooth muscle cells (SMCs) and fibroblasts. They were then combined with collagen type I/elastin/ dermatan sulfate and collagen type I/fibrin to form tunica media and tunica adventitia respectively. HUCV endothelial cells (ECs) were seeded on the construct by cell sheet engineering method after fibronectin and heparin coating. Characterization of the VC was performed by immunolabeling, histochemical staining and electron microscopy (SEM and TEM). Differentiated cells were identified by means of immunofluorescent (IF) labeling. SEM and TEM analysis of VCs revealed the presence of three histologic tunicae. Collagen and elastic fibers were observed within the ECM by histochemical staining. The vascular endothelial growth factor receptor expressing ECs in tunica intima; α-SMA expressing SMCs in tunica media and; the tenascin expressing fibroblasts in tunica adventitia were detected by IF labeling. In conclusion, by combining natural scaffolds and vascular cells differentiated from CD146+ PCs, VCs can be generated layer by layer. This study will provide a preliminary blood vessel model for generation of fully natural small-diameter vascular grafts.

1. Introduction Cardiovascular diseases are leading cause of death or deterioration of the quality of life in worldwide (World Health Organization, 2017). Coronary artery disease constitutes the most common type of cardiovascular diseases and is the leading cause of mortality (Mozaffarian et al., 2015; Wilkins et al., 2017). Today, autologous vessels or prosthetic grafts are used during coronary bypass surgery. Although the autologous vessels are the gold standard for the treatment, they have limited accessibility because of the disease progression and the age of the patient (Bajpai and Andreadis, 2012). On the other hand, clinical use of small-diameter synthetic vascular grafts is inadequate because of the inflammatory response (Grandi et al., 2011), neointimal

hyperplasia (Zilla et al., 2007), and thrombosis (Yahagi et al., 2016) resulted in the delay of healing the damaged areas (Pashneh-Tala et al., 2015). Limitations of autologous and synthetic grafts create an urgent need of generating small diameter vascular grafts, completely composed of natural materials, for clinical use. Vascular tissue engineering approaches aim to mimic vascular layers using natural or synthetic materials with vascular cells (Pashneh-Tala et al., 2015). The small-diameter vascular grafts were first generated by Weinberg and Bell (1986), using purified extracellular matrix (ECM) proteins, vascular smooth muscle cells (SMCs), fibroblasts and, the endothelial cells (ECs) that were embedded in collagen gel. Similarly, Loy et al. (2016), cultured vascular cells in collagen gel (tunica media and adventitia-like layers) and EC monolayer was formed to mimic the intima. Moreover, fibrin was used in gel

Abbreviations: VCs, vascular constructs; HUCV, human umbilical cord vein; SMCs, smooth muscle cells; ECs, endothelial cells; SEM, scanning electron microscopy; TEM, transmission electron microscopy; IF, immunofluorescent; ECM, extracellular matrix; α-SMA, alpha-smooth muscle actin; UC, umbilical cord; PNIPAAm, poly N‑isopropylacrylamide; MSCs, mesenchymal stem cells; PCs, perivascular cells; HUVEC, human umbilical cord vein endothelial cells; 3D, three dimensional; PGM, Pericyte Growth Medium; MACS, magnetic activated cell sorting; RT, room temperature; FACS, fluorencent activated cell sorting; PE, phycoerythrin; FITC, fluorescein isothiocyanate; APC, allophycocyanin; DMEM-LG, Dulbecco's modified Eagle's medium – Low glucose; CTGF, Connective Tissue Growth Factor; L-AA, L‑ascorbic acid; FBS, fetal bovine serum; Pen-Strep, penicillin-streptomycin; WST-1, Water Soluble Tetrazolium-1; EGM, endothelial growth medium; FM, fibroblast medium; VEGF, vascular endothelial growth factor; vWF, von Willebrand Factor ⁎ Corresponding author at: Hacettepe University, Center for Stem Cell Research and Development, 06100, Sihhiye, Ankara, Turkey. E-mail address: [email protected] (B. Çelebi-Saltik). https://doi.org/10.1016/j.mvr.2018.03.005 Received 5 December 2017; Received in revised form 13 March 2018; Accepted 13 March 2018 Available online 14 March 2018 0026-2862/ © 2018 Elsevier Inc. All rights reserved.

Microvascular Research 118 (2018) 101–112

B. Gökçinar-Yagci et al.

previously described protocol (Gokcinar-Yagci et al., 2016). UC samples were cut into 3–4 cm pieces in length and Wharton's jelly and arteries were removed with sterile forceps. Separated cord vein pieces were sutured at both ends and were incubated in collagenase solution (1 mg/mL, Sigma, USA) at 37 °C for 16–18 h. After washing and centrifugation steps, isolated cells were cultured in Pericyte Growth Medium (PGM, Promocell, Germany) at 37 °C in a 5% CO2 incubator. Cells that reached 75–80% confluence were trpsinized with a TrypLE (Gibco, USA) solution. CD146+ cells were sorted by magnetic activated cell sorting (MACS) technique using CD146 microbead Kit (Miltenyi Biotec, Bisley, UK) and cultured in PGM. At 80–85% confluence, adherent cells were trypsinized with TrypLE solution and cell viability was checked by trypan blue staining. Passage three CD146+ PCs were characterized according to their multilineage differentiation capacity requested for MSC characterization and specific marker expression. First, their adipogenic and osteogenic differentiation capacity were evaluated by incubating with adipogenic and osteogenic differentiation medium for 21 days at 37 °C in a 5% CO2 incubator (Celebi et al., 2010). Differentiation media was refreshed in every three days, osteogenic and adipogenic differentiation capacities were evaluated at 21st day with Alizarin red (Sigma, USA) and Oil red-O (Sigma, USA) staining respectively. Cells cultured in PGM for 21 days were used as control group for differentiation analysis. Surface markers of passage three CD146+ PCs were analyzed by flow cytometry using our previously described protocol (Gokcinar-Yagci et al., 2016). In brief, the cells (2–5 × 105 cells/mL) were suspended in fluorescent activated cell sorting (FACS) buffer and centrifuged at 200g for five minutes (min). Then they were incubated for 20 min at room temperature (RT) in the dark with the combination of following fluorescent-conjugated mouse anti-human antibodies: CD146-phycoerythrin (PE), CD73-fluorescein isothiocyanate (FITC), and CD105-allophycocyanin (APC). IgG1-PE, IgG1-FITC and IgG1-APC were the isotype controls. All antibodies were obtained from Becton Dickinson Pharmingen (Mississauga, ON, Canada). After three times of washing steps with FACS buffer, cells were analyzed with FACS Aria using FACS Diva Analysis Software v6.1.2 (BD Biosciences, USA).

form as an alternative to collagen type I to increase elastin and collagen synthesis (Grassl et al., 2003; Gui et al., 2014). Buijtenhuijs et al. (2004) used a 3D tubular scaffold with insoluble elastin and collagen, in which umbilical cord (UC)-derived SMCs proliferate and protect their viability. Researchers have tried to coat the basement membrane with proteins like collagen type IV (Coelho et al., 2010), fibronectin (De Visscher et al., 2012) and heparin (Hoshi et al., 2013; Zhou et al., 2009) in order to facilitate the endothelialisation. Cell sheet engineering method that was introduced by Okano et al. (Owaki et al., 2014), generating cell sheets using smart polymer (poly N‑isopropylacrylamide; PNIPAAm) coated thermoresponsive surface; has been used for vascular graft production with SMCs and mesenchymal stem cells (MSCs) recently (Backman et al., 2017; Sekiya et al., 2010; Williams et al., 2012). We assume this method can be used as an alternative to coating techniques for endothelialisation. The CD146+ perivascular cells (or pericytes, PCs) located at the periphery of the vessel walls (Crisan et al., 2008; Nees et al., 2013) are involved in vascular development, maturation, stability, and regulation of blood flow and blood pressure (Kutcher and Herman, 2009; Mendes et al., 2012; Stefanska et al., 2013). Their interaction with the ECs by special junctions and paracrine signalling provides proliferation of ECs (Armulik et al., 2011) and efficient endothelialisation of vascular grafts (Avolio et al., 2017). The PCs have MSC-like properties thus they are good alternative cell sources to terminally differentiated cells for vascular tissue engineering approaches (Crisan et al., 2012; Gokcinar-Yagci et al., 2015). Perivascular cells are isolated from several vascularized tissues and organs like bovine retina, brain, skeletal muscle, adipose tissue, bone marrow, kidney, liver and fetal tissues (placenta and UC) (Gokcinar-Yagci et al., 2015). Among all these sources, human UC is the most favorable source of perivascular cells because it is noninvasive and ethical for being a medical waste (Kajiyama et al., 2015). Human UC perivascular cells expressing SMC markers (Chong et al., 2009), human embryonic stem cell derived pericytes (van der Meer et al., 2013) and, human skeletal muscle derived CD 146+ pericytes (He et al., 2010) have been combined with synthetic or natural scaffolds in a limited number of vascular tissue engineering studies and gave promising results (Chong et al., 2009; He et al., 2010; van der Meer et al., 2013). The human umbilical cord vein (HUCV) CD146+ PC derived vascular cells have not been studied for vascular tissue constructs yet. We suggest that CD 146+ PCs may be good candidates for generating three layered small diameter vascular constructs when combined with human collagen type I, fibrin, elastin, dermatan sulfate, heparin and fibronectin constituting the human natural vascular components. In our previous study, we gave positive answer to the hypothesis that the CD146+ PCs can differentiate to mesodermal origined SMCs of tunica media (Gokcinar-Yagci and Celebi-Saltik, 2017). Here, we hypothesized that CD146+ PCs can give rise to fibroblasts, that are the main cells of the adventitia by culturing with connective tissue growth factor supplemented differentiation medium. Our second hypothesis is that the vascular tunicae (tunica media and tunica adventitia) can be constructed by combining human collagen type I, elastin, fibrin and dermatan sulfate as natural scaffolds; with SMCs and fibroblasts, differentiated from human CD146+ PCs. We also hypothesized that tunica intima can be generated with HUVEC monolayer obtained with cell sheet engineering method and transferred on human heparin and fibronectin coated tunica media surface. In this study, our objective is to put the three tissue-engineered vascular layers together and culture in vitro to generate an intact three dimensional (3D) vascular construct.

2.2. Smooth muscle cell and fibroblast differentiation of CD146+ perivascular cells CD146+ PCs (n = 3, passage three) were seeded in culture plates as 5 × 103 cells/cm2 and cultured with PGM until confluence. SMC differentiation was induced by culturing the cells for seven days in SMC differentiation medium according to our previously described protocol (Gokcinar-Yagci and Celebi-Saltik, 2017). Medium changes were performed two times weekly. Fibroblast differentiation of CD146+ PCs was performed with a differentiation medium (Tong et al., 2011). CD146+ PCs (n = 3, passage three) were seeded in culture plates at 5 × 103 cells/cm2 and cultured with PGM until confluence. Fibroblast differentiation was induced by culturing the cells for 21 days in differentiation medium [Dulbecco's modified Eagle's medium – Low glucose (DMEM-LG), 100 ng/mL Connective Tissue Growth Factor (CTGF, Sigma Aldrich, USA), 50 μg/mL L‑ascorbic acid (L-AA), 10% fetal bovine serum (FBS), 1% penicillin-streptomycin (Pen-Strep) at 37 °C in a 5% CO2 incubator. Medium changes were performed two times weekly. Fibroblast differentiation of CD146+ PCs was characterized phenotypically by immunofluorescent (IF) labeling with primary mouse anti-human antibodies: 1/60 dilution of anti-tenascin-C (Abcam, UK) and 1/60 dilution of anti-collagen type I (Abcam, UK). Quantitative real-time polymerase chain reaction (qRT-PCR) was performed to analyze expression of fibroblast-specific markers (tenascin-C and collagen type I) by incubation of CD146+ PCs with fibroblast differentiation medium for 21 days (n = 3). PCs cultured with PGM (n = 3) were used as negative control, foreskin fibroblasts (ATCC, USA) cultured with fibroblast growth medium (FM, Promocell, Germany) were used as positive control cells. “EZ-10 Spin Column Total RNA Mini-Preps Kit” (Bio Basic Inc., Canada) was used to isolate cellular RNA. Concentration of the RNA was evaluated with NanoDrop ND-1000 spectrophotometer (Thermo Ficher Scientific Inc., USA) at 260 nm. The quality of the extraction

2. Materials and methods 2.1. Isolation and characterization of CD146+ perivascular cells Human UC samples were obtained immediately after delivery from non-complicated pregnant women in Hacettepe University Faculty of Medicine, Department of Obstetrics and Gynecology (n = 3). Hacettepe University Local Non-Interventional Clinical Researches Ethics Committee approved human material use (GO 13/417-27). CD146+ PCs were isolated from human UC samples according to our 102

Microvascular Research 118 (2018) 101–112

B. Gökçinar-Yagci et al.

Fig. 1. Overview of important steps in generation of three-layered vascular construct. A) CD146+ perivascular cells were isolated from umbilical cord vein and differentiated to smooth muscle cells (SMCs) and fibroblasts. SMCs, HUVECs and fibroblasts were mixed with human ECM proteins and cultured together for indicated time periods. B) Tunica media (TM) and tunica intima (TI) were cultured together as two-layered patch for 7 days. C) After addition of tunica adventitia (TA) layer to TM + TI layer, the shrunk three-layered patch was cultured for 15 days as planar form. D) The three-layered construct (TM + TI + TA) was wrapped around a mandrel and cultured for additional one day at this tubular form.

sample − OD value of the blank) × 100 / OD value of the sample (OD = optical density).

was determined with the A260/A280 ratio. “iScript cDNA Synthesis Kit” (Bio Rad, USA) was used to synthesize complementary DNA (cDNA). Gene expression was identified by using LightCycler_480 II (Roche, Germany) in 384well qPCR plate with 10 μL reaction mix [8 μL master mix: 19 primer–probe (Taqman Gene expression assays, Thermo Fisher Scientific Inc., USA), 19 Light Cycler 480 probes master (Roche, Germany), nuclease free water, and 2 μL cDNA] in each well. The condition of the PCR was the same as it was described in our previous study (Gokcinar-Yagci and Celebi-Saltik, 2017). The crossing point (Cp, or threshold cycle: Ct) value was calculated for target genes of three biological replicates and reference gene (β‑Actin) by using LightCycler 480 II software.

2.4. Generation of the three-dimensional vascular construct In order to mimic the middle layer of human blood vessel (tunica media), human collagen type I, elastin and dermatan sulfate were mixed to form a 3D gel. Collagen type I 3-D gel was prepared according to the manufacturer's instruction (VitroCol®, Advanced BioMatrix, San Diego, CA). One part of 10× DMEM-LG (Sigma Aldrich, USA) was slowly added to 8 parts of human collagen type I (Advanced BioMatrix) and the pH of the suspension was adjusted to 7.2–7.4 with 0.1 M NaOH. Collagen type I suspension was mixed with human elastin (Elastin Products Company, Inc., USA) and recombinant human dermatan sulfate (R&D Systems, USA) with final concentrations of 1.6 mg/mL, 0.8 mg/mL and 0.01 mg/mL respectively (Madhavan et al., 2010; Miwa et al., 1993). All of these steps were carried out on ice. Collagen type I, elastin and dermatan sulfate mixture was placed in petri dish after addition of SMCs (differentiated from CD146+ PCs) (Fig. 1A) at a density of 5 × 105 cells/mL and allowed gelation at 37 °C in 2–3 h. After gel formation SMC growth medium was added on the tunica media construct and cultured for 15 days at 37 °C in a 5% CO2 incubator at static conditions. Culture medium was changed two times a week. In order to generate the luminal part of the vascular construct (tunica intima), endothelial cells were obtained from Lonza (Switzerland) instead of differentiating CD146+ PCs. According to our previous data (not shown), CD146+ PCs did not differentiate into endothelial cells with vascular endothelial growth factor (VEGF)-supplemented culture medium. So, for the generation of intimal layer, we used HUVECs which came from the same source with CD146+ PCs. They were seeded on PNIPAAm-coated cell culture plate (Nunc UpCell, Thermo Fisher Scientific, USA) and cultured with EGM (Promocell) at 37 °C in a 5% CO2 incubator until confluency. EC layer was obtained according to the instructions of Nunc UpCell, Thermo Fisher Scientific. The HUVEC monolayer was separated from the Nunc UpCell surface with the supplied membrane by decreasing the temperature to 24–25 °C for 15–40 min. Then the membrane was transferred on to the tunica media construct which was cultured for 15 days and coated with fibronectin (250 μg/mL, Sigma, USA) and heparin (3,78 mg/mL, Mustafa Nevzat, Turkey) for 2 h at 37 °C (Fig. 1A). After 30 min incubation at 37 °C,

2.3. Cell viability assay Attachment and viability of HUVECs, SMCs and fibroblasts in their corresponding matrix protein mixtures were assayed with “Water Soluble Tetrazolium-1 (WST-1) Cell Proliferation Assay Kit” (Cayman Chemical, USA). HUVECs were seeded at 5000 cells/well with endothelial growth medium (EGM, Promocell, Germany) in 96-well culture plates after they were coated with fibronectin (250 μg/mL) and heparin (3,78 mg/mL). Cells were cultured at 37 °C in a 5% CO2 incubator. Medium changes were performed twice weekly. SMCs were mixed with collagen type I (1.6 mg/mL), elastin (0.8 mg/mL) and dermatan sulfate (0.01 mg/mL) at a density of 5 × 105 cells/mL and allowed gelation at 37 °C in 2–3 h in 96-well culture plates. After gelation SMC growth medium was added on the cell-matrix mixture and refreshed two times weekly for the culture at 37 °C in a 5% CO2 incubator. Fibroblasts were mixed with collagen type I (1.6 mg/mL) and, fibrin (1.6 mg/mL) at a density of 5 × 105 cells/mL and allowed gelation at 37 °C in 2–3 h in 96-well culture plates. After gelation fibroblast medium (FM, Promocell, Germany) was added on the cell-matrix mixture and refreshed two times weekly for the culture at 37 °C in a 5% CO2 incubator. Cell proliferation and viability was tested at 1st, 2nd and 3rd weeks of cultures according to the manufacturer instructions. Briefly, 10 μL of WST-1 mixture was added to every 100 μL of culture medium. After shaking the plate gently for 1 min, cells were incubated at 37 °C in a 5% CO2 incubator for 2 h. In order to distribute the color homogenously, the plate was shaken again for 1 min. The absorbance of each sample was measured with microplate reader at a wavelength of 450 nm. Data were expressed as (OD value of the 103

Microvascular Research 118 (2018) 101–112

B. Gökçinar-Yagci et al.

For thrombocyte adhesion test, the three-layered vascular patch in the petri dish was cultured with thrombocytes differentiated from CD34+ cells for two days. Then, immunofluorescent staining was performed with primary mouse anti-human antibodies: anti-CD61-PE and anti-CD41-APC (BD) to evaluate the adherence of thrombocyte on endothelial layer. The 3D ring shaped vascular construct were allocated to immune phenotypic and morphologic analyses. One half of each 3D patch was embedded in OCT compound and frozen in liquid nitrogen for immune labeling studies. The other half was immersed in the gluteraldehyde or formaldehyde chemical fixative solution in order to process for light (paraffin blocks for histochemistry) or electron microscopy (plastic blocks for TEM or sputter coating for SEM).

membrane was detached leaving the HUVEC layer on the tunica media. Tunica media + tunica intima were cultured together for seven days with EGM (Promocell) (Fig. 1B). Culture medium was changed two times a week. In order to mimic the outer layer of human blood vessel (tunica adventitia), human collagen type I and, fibrin were mixed with fibroblasts to form 3D gel (Fig. 1A). Collagen type I solution was prepared as described above. Fibrin gel was prepared according to previously described protocol (Isenberg et al., 2006). Briefly, 2 mL of 6 mg/mL fibrinogen (Sigma, USA) and 500 μL of thrombin solution [37,5 μL of 2.5 units thrombin (Sigma, USA) and 7.5 μL of 2 N calcium chloride (Sigma, USA) in 2 mL of DMEM-LG] was mixed. Collagen type I solution was mixed with fibrinogen/thrombin mixture with final concentrations of 1.6 mg/mL and 1.6 mg/mL respectively (Cummings et al., 2004). All of these steps were carried out on ice. Collagen type I and fibrinogen/thrombin mixture was placed on the tunica intima + media construct in the petri dish after addition of fibroblasts (differentiated from CD146+ PCs) at a density of 1 × 106 cells/mL and allowed gelation at 37 °C in 2–3 h. After gel formation FM (Promocell) was added on the three-layered (tunica intima + tunica media + tunica adventitia) construct and cultured for additional 15 days at static conditions (Fig. 1C). Culture medium was changed two times a week. After 15-day of culture, the three-layered vascular patch in the petri dish was wrapped around a metal mandrel (30 mm) and a 4 mm diameter tubular graft was formed (Fig. 1D). One day later, this tubular construct was cut into pieces for further characterization analyses.

2.6. Histochemistry and immunofluorescence analysis of paraffin and frozen sections For histochemistry, vascular constructs were fixed in 10% phosphate buffered formalin (pH 7.0) at RT, rinsed in buffer, and dehydrated in a graded series of ethanol before embedding in paraffin by using an automated tissue processor. Five to seven micrometer thick serial sections were stained with Haematoxylin & Eosin, Verhoeff Van Gieson, Mallory and Masson's Trichrome then the micrographs were analyzed with a light microscope attached digital camera (Leica DMR, DC500, Wetzlar, Germany). For immunofluorescence analysis, the samples were immediately frozen in liquid nitrogen and stored at −80 °C. Five to seven micrometer thick cryostat sections on poly‑L‑Lysine-coated slides underwent indirect immune fluorescent labeling procedure (Gokcinar-Yagci et al., 2016). In brief, sections were fixed in 4% paraformaldehyde for 5 min, air-dried and blocked for nonspecific binding with 10% rabbit serum and 10% human serum for 10 min at RT. After washing in PBS for 5 min, primary antibody incubation was carried out with the following monoclonal antibodies at optimal concentrations in a humidified chamber overnight at 4 °C (1/10 dilution of mouse anti-human tenascin (Abcam, UK), 1/10 dilution of mouse anti-human α-SMA (Abcam, UK), mouse anti-human von Willebrand factor (vWF, R&D). After washing with 0.05% tween solution, sections were incubated with Alexa Fluor® 488conjugated rabbit anti-mouse IgG (Thermo Fisher Scientific, USA) containing 10% rabbit serum and 10% normal human serum. A counterstaining for nuclei was performed with DAPI. Control staining was performed by omitting the initial primary antibody step and using a control mouse IgG. Data were documented with a light microscope attached digital camera (Leica DMR, DC500, Wetzlar, Germany).

2.5. Characterization of the three-dimensional vascular patch and construct 2.5.1. The dead cell apoptosis analysis in three-dimensional vascular patch The Annexin V-fluorescein isothiocyanate (FITC)/Propidium Iodid (PI) Dead Cell Apoptosis Detection Kit (Invitrogen) was used to evaluate live, apoptotic and death cells in 3D vascular patch. After rinsed with ice cold PBS, the patch was resuspended in 1× of Annexin binding buffer. Then, 62.5 μL of Annexin V stock solution and 5 μL of propidium iodide (PI) in 250 μL Annexin binding buffer was added on the patch and incubated for 15 min at room temperature according to the manufacturer's instructions. The vascular patch which had the size as “x: 921,36 μm, y: 921,36 μm, z: 197,25 μm” examined under a Confocal Laser-Scanning Microscope (LSM Pascal, Zeiss, Germany) for fluorescent imaging. Through its z position, 36 layers of the vascular patch were analyzed in terms of live/dead cell ratio. The cells separate into three groups; live, apoptotic and death. Live cells show only weak membranous Annexin V labeling, while apoptotic cells show as significantly higher intensity of surface labeling. The dead cells show both membrane labeling by Annexin V and strong nuclear labeling with propidium iodide. The ratio of dead cells to the total number of cells within the 3D patch were measured and reported.

2.7. Transmission and scanning electron microscopy (TEM and SEM) For TEM analysis; vascular constructs were fixed in 2.5% glutaraldehyde in Sorensons' phosphate buffer (pH 7.0), rinsed in buffer, post-fixed in 1% osmium tetroxide at 4 °C for 2 h and, dehydrated in graded series of ethanol for embedding in araldite (EMS, Germany) by using an automated tissue processor. Semi-thin sections were stained with methylene blue-Azure II; thin sections were stained with uranyl acetate and lead citrate, analyzed on a transmission electron microscope (Jeol, JEM1400, Japan) attached digital camera (Orius, Germany) to evaluate ultrastructural integrity. For SEM analysis, the samples were fixed with 2.5% gluteraldehyde in PBS (pH 7.0) over night at 4 °C, washed, dehydrated in graded series of ethanol and, sputter-coated by 10 nm thick palladium gold (Precision etching coating system, GATAN 682, USA) after drying with a critical point of dryer (Tousimis, Taiwan). The ultramicrographs were captured with High Voltage Scanning Electron Microscope (Fei Quanta 200F, USA), attached digitalized image analysis system (GATAN, Germany).

2.5.2. Anti-thrombogenic properties of vascular patch Human umbilical cord blood mononuclear cells (MNC) were isolated with Ficoll Paque Plus (1.077 g/mL; GE Healthcare Life Sciences, Buckinghamshire, UK) density gradient method. Fresh or thawed cord blood samples were pooled before manipulation (4–5 samples). MNC concentration was adjusted to 2 × 107 cells/mL, and CD34+ cells were enriched by negative selection using the Easy Sep Cell Separation System and the Human Progenitor Enrichment Cocktail (StemCell Technologies, Vancouver, BC, Canada). Purity of CD34+-enriched cells was confirmed by flow cytometry (purity 75.0 ± 5.0%). Then, CD34+ cells were grown in serum-free thrombocyte differentiation medium; IMDM (Invitrogen) supplemented with 20% serum substitute solution (BIT, Stem Cell Technologies), 40 μg/mL of LDL (Sigma), 5 × 10−5 M 2‑mercaptoethanol (Sigma) and different combinations of human cytokines (TPO, SCF and Flt-3 ligand were purchased from Peprotech, Rocky Hill, USA). In the expansion phase, 2 × 104 cord blood CD34+ cells/mL were cultured for 12 days with the OMPC cocktail (TPO 35 ng/mL, Stem Cell Factor 10 ng/mL, Flt-3 ligand 11 ng/mL). Flow cytometry analysis of differentiated cells was performed as previously described. The antibodies used were; anti-CD41a (GPIIb)-APC (BD), anti-CD61-PE and corresponding control antibodies (mouse isotype APC or PE controls (BD)).

2.8. Statistical analysis The data were analyzed with SPSS 16.0 statistical software (SSPS Inc., Chicago, IL, USA). Statistical significance was evaluated based on Student paired t-test, value of P < 0.05 was considered significant. All experiments were repeated three times. 104

Microvascular Research 118 (2018) 101–112

B. Gökçinar-Yagci et al.

3. Results

PCs cultured with PGM (control group) maintained their undifferentiated state. Flow cytometry analysis of CD146+ PC population indicated that, 77.6 ± 4.6% of cells was CD146 positive (Fig. Suppl. 1). The MSC-specific surface markers like CD73 and CD105 were highly expressed by CD146+ PCs (≥90.0%, Fig. Suppl. 1). These cells did not express hematopoietic (CD34 and CD45) and EC (CD31) markers (data not shown), which revealed that there is no contamination of hematopoietic or ECs.

3.1. Isolation and characterization of CD146+ perivascular cells CD146+ PCs isolated from perivascular region of HUCV were successfully cultured and expanded with PGM. At the end of their first week at culture, they displayed spindle-shaped, fibroblast-like morphology in vitro (Fig. Suppl. 1). These irregularly shaped cells reached approximately 80% confluency within two weeks. In order to assess the multilineage differentiation capacity of CD146+ PCs, they have cultured under defined differentiation conditions. Upon adipogenic induction, lipid droplets stained with Oil Red O were examined to be accumulated in the cytoplasm of CD146+ PCs after 21-day culture (Fig. Suppl. 1). Alizarin Red staining after 21-day culture with osteogenic differentiation medium was used to evaluate osteogenic differentiation of CD146+ PCs. But the calcium deposits stained with Alizarin red in the cell aggregates were observed very rarely (Fig. Suppl. 1). CD146+

3.2. Smooth muscle cell and fibroblast differentiation of CD146+ perivascular cells Human SMCs were successfully obtained by differentiating HUCV CD146+ PCs as reported before (Gokcinar-Yagci and Celebi-Saltik, 2017). Tenascin-C and collagen type I expression of non-differentiated PCs and fibroblasts differentiated from PCs were shown in Fig. 2A–B. According to immunofluorescent staining, fibroblasts differentiated from PCs were positive

Fig. 2. Immunofluorescence and qRT-PCR analysis of non-differentiated perivascular cells and fibroblasts differentiated from CD146+ perivascular cells at passage three. Tenascin-C and Collagen type I expression of non-differentiated perivascular cells and fibroblasts differentiated from CD146+ perivascular cells with Alexa 488 (DAPI (blue) was used as nuclear staining) (A). qRT-PCR results of fibroblasts differentiated from CD146+ perivascular cells and foreskin fibroblasts (positive control). Fold exchange of the Tenascin-C and Collagen type I expression was normalized compared to non-differentiated perivascular cells (negative control) (B).

105

Microvascular Research 118 (2018) 101–112

B. Gökçinar-Yagci et al.

differentiated PCs (negative control cells). Hereby, both immunofluorescent staining and qRT-PCR results showed that fibroblasts differentiated from HUCV CD146+ PCs expressed fibroblast-specific markers as tenascin-C and collagen type I (Fig. 2A–B).

Table 1 The percentages of proliferative and viable cells cultured with their corresponding matrix proteins and glycosaminoglycans at the end of 1st, 2nd and 3rd week (mean ± SD of independent experiments is presented. Significant differences were determined by Student t-test; *: P < 0.05, n = 3). Day-7 HUVECs on heparin/ fibronectin matrix SMCs in collagen type I/ elastin/DS matrix Fibroblasts in collagen type I/ fibrin matrix

Day-14

Day-21 ⁎

87.79 ± 0.17

84.20 ± 0.30

63.10 ± 1.15

78.15 ± 5.88

81.71 ± 0.76

74.17 ± 2.31

91.95 ± 1.93

90.84 ± 1.21

3.3. Viability of cells cultured with ECM proteins

86.78 ± 3.78

Cells attached and maintained their viability for three weeks in 3D culture models (Table 1). According to the results of WST-1 assay, the viability of HUVECs on fibronectin and heparin coated surface remained approximately the same for three weeks (Table 1). For SMCs cultured in collagen type I - elastin - dermatan sulfate scaffold and fibroblasts cultured in collagen type I - fibrin scaffold, there was an increase of cell proliferation during 21 days. Their viability in the presence of matrix proteins didn't decrease in time.

for both markers whereas non-differentiated PCs was only tenascin-C positive. qRT-PCR results revealed that differentiated PCs and foreskin fibroblasts (positive control cells) were positive for both markers (Fig. 2B). Fold exchange of the protein expression was normalized compared to non-

Fig. 3. Three-dimensional vascular construct in petri dish. A macroscopic side view of three-layered vascular construct in petri dish (A), confocal microscopy of vascular patch stained with Annexin-V- FITC/PI. The three-dimensional vascular patch was incubated with Annexin-V FITC and PI then analyzed by using laser confocal scanning microscopy, light green fluorescence represents live/early apoptotic cells, while red fluorescence represents dead cells (10X). Note the low percentage of the dead cells on micrograph (B).

106

Microvascular Research 118 (2018) 101–112

B. Gökçinar-Yagci et al.

and necrotic cells (Fig. 3B). The assessment of 3D vascular patch for the presence of live and dead/late apoptotic cells by Annexin-V-FITC-PI labeling on laser confocal scanning microscopy revealed the presence of vascular cells along the whole thickness of the construct. There were increasing numbers of apoptotic and necrotic cells among all vascular z stack cross-sections, but the ratio for the PI positive death cells to all cells remained low (22%) in whole thickness of the 3D patch (Fig. 3B). To determine the anti-thrombogenic property of vascular patch, adhesion of thrombocytes (differentiated from hematopoietic stem cells) on endothelial cell layer was tested. For this purpose, we first examined whether CD34+ cells differentiated into megakaryocyte/ thrombocyte lineages. Flow cytometry analysis indicated that the frequencies of the CD41+ CD61+ cells were 93.7% (Fig. 4A). Differentiated cells were cultured with vascular patch and after two days of culture their adherence was evaluated by immunofluorescent staining. Low number of cells expressing CD61 (≤1.0%) was seen on the vascular patch (Fig. 4B). The morphologic analysis of 3D vascular constructs revealed the

3.4. Generation and characterization of three-dimensional vascular construct Smooth muscle cells and fibroblasts successfully integrated with natural human ECM proteins and glycosaminoglycans. In 24 h, the three-layered patch in the petri dish shrunk to one third of its size (Fig. 3A). The length of the triple layered vascular patch after it was wrapped around mandrel was approximately 1.5 to 2 cm. The images of the histologic, electron microscopic and IF analyses of the vascular patch belonged to this tubular construct. The three-dimensional vascular patch was incubated with Annexin-V-FITC and PI then analyzed by laser confocal scanning microscopy. The green fluorescence represented Annexin V, while red fluorescence represented cell nuclei stained by PI. According to confocal images in Fig. 4B, early apoptotic cells were Annexin V-positive and PI-negative, whereas late (end-stage) apoptotic cells were Annexin V/PI-double-positive (should be yellow but not detected) and PI positive cells were necrotic. Among all vascular cross-sections provided by confocal microscope, there were apoptotic

Fig. 4. Anti-thrombogenic activity of vascular patch. Determination of differentiated CD34+ cells surface markers by flow cytometry with anti-human CD41-APC and anti-human CD61PE (A). Representative immunofluorescent images of differentiated cells on vascular patch (B). Scale bar = 100 μm.

107

Microvascular Research 118 (2018) 101–112

B. Gökçinar-Yagci et al.

Fig. 5. Histologic analysis of three-dimensional vascular construct. Low power micrographs of the paraffin (A, C) and plastic semithin (B) sections of 3D vascular construct presenting three tunicae of the vessel wall with a thin tunica intima containing endothelial lining (Et with arrow), tunica media consisting of fusiform shaped myocytes and; the thick, loose tunica adventitia where the big fibroblasts are embedded in the ECM. Note the intact endothelial cells with their centrally located nuclei and the continuous lining in inset at high magnification (inset A) Collagen fibers appear pale pink with H&E, blue with methylene blue and green with light green (Masson's trichrome) (A–C).

pictures revealed the typical ultrastructure of the collagen fibril with their gap regions and the irregularity of the elastic fibrils within the ECM (Fig. 7G, H, I and J).

presence of three histologic tunicae (tunica intima, media and adventitia) at light microscopic levels (Fig. 5). Low power micrographs of the paraffin (A, C) and plastic semithin (B) sections of 3D vascular construct is presenting three tunicae of the vessel wall with a thin tunica intima containing endothelial lining (Et with arrow), tunica media consisting of fusiform shaped myocytes and, the thick, loose tunica adventitia where the big fibroblasts are embedded in the ECM. Note the intact endothelial cells with their centrally located nuclei and the continuous lining in inset at high magnification. Collagen fibers appear pale pink with H&E, blue with methylene blue and green with light green (Masson's trichrome) (Fig. 5A–C). Tunica intima consisted of a thin single lining of ECs (sometimes sloughed off during sample preparation for microscopic processing) (Figs. 5A and 6B). The subendothelial layer was not apparent; but the media and adventitia consisted of several layers of cells intermingled with ECM molecules and the fibers (Figs. 5B, C and 6). Collagen and elastic fibers were observed within the ECM by histochemical staining (Fig. 6C and E). The vWF expressing ECs in tunica intima (Fig. 6B); α-SMA expressing SMCs in tunica media (Fig. 6D) and; the tenascin expressing fibroblasts in tunica adventitia (Fig. 6F) were detected by immune fluorescent labeling. The 3D vascular constructs presented all the three tunicae that were well integrated with each other resembling the natural vessel wall at the ultrastructural analysis by scanning and transmission electron microscopy (Fig. 7). ECs and the collagen fibrils were observed in tunica intima (Fig. 7A and B). Fusiform smooth muscle-like cells with pinocytotic vesicles were embedded in an amorphous ECM containing elastic and collagen fibers in tunica media (Fig. 7C, D, G and H). The spindle or ovoid shaped fibroblastic cells exhibited high endosomal activity, being generally rich in lysosomal vesicles in their cytoplasms. Those fibroblasts were dispersed between collagen fibrils and the matrix particles in tunica adventitia (Fig. 7E, F, I and J). The TEM and SEM

4. Discussion In this study, we designed an in vitro vascular model histologically mimicking the vascular tissue with human SMCs and fibroblasts derived from UC CD146+ PCs and HUVECs by using human ECM proteins and glycosaminoglycans. We first isolated CD146+ PCs from HUCV samples by using the procedure that we described in our previous study (Gokcinar-Yagci et al., 2016). The fibroblast-like morphology of CD146+ PCs is consistent with the microscopic results of other studies (Montemurro et al., 2011; Tsang et al., 2013). Similar with our results, Murray et al. (2014a, 2014b) reported that the CD146+ PCs express MSC markers. They presented adipogenic (based on Oil Red O staining), but not osteogenic differentiation capacity based on Alizarin Red staining on day 21 of culture. Montemurro, Tsang and Ennis et al. reported osteogenic capacity of UC PCs (Ennis et al., 2008; Montemurro et al., 2011; Tsang et al., 2013). Greenwood-Goodwin et al. (2016) stated that perivascular progenitor cells have no multipotential differentiation capacity but the potential of giving rise to other progenitors that have osteogenic potential or pericyte-like function. Murray et al. (2014a, 2014b) reported that the osteogenic differentiation capacity of CD146+ perivascular MSC precursors was maintained and accelerated by co-culturing with ECs. An et al. (2015) showed high homogeneity and increased osteogenic potential for UC PCs when isolated using nonenzymatic technique. In our study, we used collagenase-based enzymatic isolation technique that may cause a heterogeneous population of CD146+ PCs and other vascular cells. We assume that PC populations may have variable differentiation capacity among themselves according

108

Microvascular Research 118 (2018) 101–112

B. Gökçinar-Yagci et al.

Fig. 6. Histochemical and immunofluorescent staining of vascular layers. The left and the right columns show histochemically (A, C and E) and immunohistochemically (B, D and F) labeled micrographs of different histological layers for the 3D vascular construct respectively. The collagen fibers (Col with arrow) appear pink with eosin in A, green with light green in C, pink to ten with Verhoeff in E. Elastic fibers (El with arrow) appear blue to black in E. Note the fibroblasts (Fb with arrow) stained with their purple nuclei in A, red nuclei in C. The myocyte (My with arrow) have brown nuclei in E. B presents vWF expressing ECs (Et with arrow) in tunica intima; D presents alpha smooth muscle actin (α-SMA) expression of myocytes (My with arrow) in tunica media and; F presents tenascin expressing fibroblasts (Fb with arrow) in tunica adventitia; all in green with FITC. Nuclei appear blue with DAPI. HE: Heematoxylin Eosin, MT: Masson's Trichrome, VvG: Verhoeff von Gieson.

culture in the presence of human ECM proteins and glycosaminoglycans for three weeks. Number of studies using PCs for vascular tissue engineering is very limited in the literature (Chong et al., 2009; He et al., 2010; van der Meer et al., 2013). He et al. (2010) cultured human skeletal muscle pericytes with elastomeric, tubular, and porous synthetic scaffolds. van der Meer et al. (2013) constructed micro-engineered 3D vascular tissue by mixing HUVECs, pericytes and the rat tail collagen type I. This is the only study that highlights the interaction of pericytes with collagen type I as a model of vascular graft. van der Meer et al. (2013) used human embryonic stem cell-derived pericytes and xenogeneic collagen type I, differently from our work. Chong et al. (2009) generated biphasic vascular model containing synthetic polymer seeded with HUCV PCs in the media layer and ECs in the intima layer. Similar to this research, we isolated PCs from HUCV, and we were able to differentiate them into vascular cell types for tunica media and adventitia layer construction. In our study, vascular cells differentiated from HUCV CD146+ PCs were successfully combined with human collagen, elastin, fibrin and dermatan sulfate for generation of biocompatible vascular media and adventitia tunicae. Ryan and O'Brien (2015) have used insoluble elastin in combination with collagen type I for the vascular construct, and evaluated the enhanced viscoelastic properties of the structure by increasing contractile properties of SMCs. In our study, collagen type I, elastin and dermatan sulfate efficaciously mimicked tunica media layer with SMCs differentiated from PCs. The SMCs successfully integrated

to the site, method of isolation and, in vitro culture conditions (An et al., 2015; Gokcinar-Yagci et al., 2016; Greenwood-Goodwin et al., 2016; Murray et al., 2014a, 2014b; Vezzani et al., 2016). In vitro conditions for differentiation of CD146+ PCs into fibroblasts and SMCs have not been defined previously. In our study, we used the media defined for differentiation of MSCs into SMCs and fibroblasts (Narita et al., 2008; Yang et al., 2011). TGF-β1 was found to be the key factor for the SMC lineage differentiation of CD146+ PCs in our previous work (Gokcinar-Yagci and Celebi-Saltik, 2017). In vitro fibroblastic differentiation of MSCs has been induced using CTGF by Tong and Lee et al. (Lee et al., 2010; Tong et al., 2011). In our study, CD146+ PCs expressed fibroblast-specific markers collagen type I and, tenascin-C after 21-day of culture with CTGF supplemented differentiation medium accordingly (Lee et al., 2010). The differentiated CD146+ PCs were used for generating the tunica media and adventitia layer of our novel construct. In our study, we tried to generate totally natural triple-layered vascular construct by combining human ECM proteins and glycosaminoglycans as scaffolds with vascular cells (SMCs and fibroblasts) differentiated from HUCV CD146+ PCs for tunica media + adventitia layers and HUVECs for tunica intima layer. The presence of these three layers was successfully confirmed by light and electron microscopy. Before generation of triple-layered construct, the viability and proliferation capacity of SMCs, fibroblasts and HUVECs were evaluated with WST-1 and were shown to be maintained (≥75.0%) during their 109

Microvascular Research 118 (2018) 101–112

B. Gökçinar-Yagci et al.

Fig. 7. Electron micrographs of vascular cells and extracellular matrix proteins. The left and the right columns show scanning (A, C, E, G and I) and transmission electron (B, D, F, H and J) micrographs of different histological layers for the 3D vascular construct respectively. Endothelial cells (Et with arrow) and the collagen fibrils (Col with arrow) are observed in tunica intima (A, B). Fusiform smooth muscle-like cells (My with arrow) with pinocytotic vesicles (Pv with arrow) are embedded in an extracellular matrix (ECM) containing elastic (El with arrow) and collagen (Col with arrow) fibers in tunica media (C, D, G and H). Fibroblastic cells (Fb with arrow) rich in lysosomal vesicles (Ly with arrow) in nucleus (N with arrow) are adjacent to collagen fibrils and fibrin (F with arrow) particles in tunica adventitia (E, F, I and J). B, D, F, H, J: Uranyl Acetate, lead citrate.

generating an adventitial layer that resembles to the native state. Thus, the spindle shaped, tenascin-C expressing fibroblastic cell layers effectively dispersed between collagen fibrils and fibrin were successfully observed among them. Recently Loy et al. (2016) designed a very convenient planar vascular model that is close to the native form, by combining collagen type I with three types of vascular cells (fibroblasts, SMCs and HUVECs). Although having some similarities with Loy et al., our tubular, three layered construct consists of fibrin for the adventitia, elastin and dermatan sulfate for the media layer in addition to collagen

with the ECM proteins and glycosaminoglycans in the media layer according to the light (H&E and van Gieson staining) and electron (TEM, SEM) microscopic data. Daamen et al. (2003) used collagen type I, elastin and chondroitin sulfate for their vascular graft scaffold, similar to our tunica media equivalent and their characterization data is in line with ours. Grassl and Aper et al. used only fibrin for their vascular graft and showed the viability of SMCs in their scaffold (Aper et al., 2016; Grassl et al., 2003). Different from their construct, we added collagen type I to the fibrin and mixed the fibroblasts differentiated from PCs for 110

Microvascular Research 118 (2018) 101–112

B. Gökçinar-Yagci et al.

order to strengthen the construct mechanically, feed the cells dispersed in the scaffolds and control the integrity of the intimal layer. Supplementary data to this article can be found online at https:// doi.org/10.1016/j.mvr.2018.03.005.

type I. Loy et al. (2016) characterized the planar vascular construct by histological and immunofluorescence analyses and showed the presence of vascular cells in each layer. Conformably, using histochemical staining and immunofluorescent labeling, we detected the vWF expressing HUVECs, α-SMA expressing SMCs and tenascin-C expressing fibroblasts, all embedded in ECM proteins and glycosaminoglycans we used. Our triple-layered patch shrunk to one third of its size likewise (Loy et al., 2016). In our study, we were able to combine the ECM components with vascular cells differentiated from CD146+ PCs and obtained an intact 3D vascular patch by using a new endothelialisation method. Endothelialisation is as important as materials and cells used for the tunica media and adventitia of the vascular construct; since it mediates main hemodynamic physiological functions in human blood vessels (Feletou, 2011). In our study, we used cell sheet engineering method with thermoresponsive smart surfaces for getting HUVEC monolayer with its intact ECM for endothelialisation. We first coated the patch with heparin and fibronectin before HUVEC seeding in order to enhance the efficacy of endothelialisation (De Visscher et al., 2012; Zhou et al., 2009). Thus, a thin single lining of vWF expressing ECs that were integrated with collagen type I fibrils were successfully observed in tunica intima layer. The morphologic assessment of endothelial lining is a challenging intervention since the cells may easily slough off during histologic tissue processing or immune labeling procedures. This is the first study in which cell sheet engineering method was used for generation of EC lining of vascular grafts. Williams and Backman et al. used PNIPAAm coated surfaces for obtaining SMC and MSC layers in vascular tissue engineering (Backman et al., 2017; Williams et al., 2012). They succeeded to obtain aligned single-layer cell sheets that are promising for engineering cell-based vascular grafts. On the other hand, our study demonstrated the effective use of cell sheet engineering for endothelialisation of these vascular constructs. Taking together, three types of vascular cells used in our vascular construct maintained their viability in fully natural scaffolds, and with proper endothelialisation, these cell-ECM units generated three-layered vascular construct similar to native blood vessel. Study results are limited due to lack of functional analyses for the biomechanical stability of the construct in bioreactor systems. This limitation however does not constrain future in vitro bioreactor and in vivo animal studies, as the essential histologic integrity was validated at this preliminary study similar to previous published works (Kaessmeyer et al., 2017; Loy et al., 2016; Muraoka et al., 2013). Transplantation of autologous and allogeneic blood vessels is the most convenient treatment for vascular disease. Developments in tissue engineering are contributing greatly with regard to this urgent need for blood vessels. According to a recent clinical survey, 3686 patients have received cellular or tissue-engineered therapies in Europe and Eurasian countries. Among them, 2513 patients (68%) have treated with autologous cells and 1173 patients (32%) have treated with allogeneic cells. In this context, indications for cardiovascular disorders were 12% of which 69% was due to autologous transplantation (Ireland et al., 2018). Thus, the development and in vitro- in vivo testing of biomimetic allogeneic vascular constructs are the first step for the design of new tissueengineered therapies in clinic. Here, we propose a new allogeneic, biomimetic 3D vascular construct for this purpose.

Conflict of interest The authors declare no potential conflicts of interest. Funding This study was supported by The Scientific and Technological Research Council of Turkey (Project Number: 113S815), and Hacettepe University, Scientific Research Project Coordination Unit (Grant Number: THD-2016-10504). Acknowledgement The authors wish to thank Prof. Nuhan Purali for confocal microscope analysis and The Scientific and Technological Research Council of Turkey, and Hacettepe University, Scientific Research Project Coordination Unit for their financial support. References An, B., et al., 2015. Non-enzymatic isolation followed by supplementation of basic fibroblast growth factor improves proliferation, clonogenic capacity and SSEA-4 expression of perivascular cells from human umbilical cord. Cell Tissue Res. 359, 767–777. Aper, T., et al., 2016. Novel method for the generation of tissue-engineered vascular grafts based on a highly compacted fibrin matrix. Acta Biomater. 29, 21–32. Armulik, A., et al., 2011. Pericytes: developmental, physiological, and pathological perspectives, problems, and promises. Dev. Cell 21, 193–215. Avolio, E., et al., 2017. Perivascular cells and tissue engineering: current applications and untapped potential. Pharmacol. Ther. 171, 83–92. Backman, D.E., et al., 2017. A robust method to generate mechanically anisotropic vascular smooth muscle cell sheets for vascular tissue engineering. Macromol. Biosci. 17. Bajpai, V.K., Andreadis, S.T., 2012. Stem cell sources for vascular tissue engineering and regeneration. Tissue Eng. Part B Rev. 18, 405–425. Buijtenhuijs, P., et al., 2004. Tissue engineering of blood vessels: characterization of smooth-muscle cells for culturing on collagen-and-elastin-based scaffolds. Biotechnol. Appl. Biochem. 39, 141–149. Celebi, B., et al., 2010. Proteome analysis of rat bone marrow mesenchymal stem cell differentiation. J. Proteome Res. 9, 5217–5227. Chong, M.S., et al., 2009. Development of cell-selective films for layered co-culturing of vascular progenitor cells. Biomaterials 30, 2241–2251. Coelho, N.M., et al., 2010. Different assembly of type IV collagen on hydrophilic and hydrophobic substrata alters endothelial cells interaction. Eur. Cell. Mater. 19, 262–272. Crisan, M., et al., 2008. A perivascular origin for mesenchymal stem cells in multiple human organs. Cell Stem Cell 3, 301–313. Crisan, M., et al., 2012. Perivascular cells for regenerative medicine. J. Cell. Mol. Med. 16, 2851–2860. Cummings, C.L., et al., 2004. Properties of engineered vascular constructs made from collagen, fibrin, and collagen-fibrin mixtures. Biomaterials 25, 3699–3706. Daamen, W.F., et al., 2003. Preparation and evaluation of molecularly-defined collagenelastin-glycosaminoglycan scaffolds for tissue engineering. Biomaterials 24, 4001–4009. De Visscher, G., et al., 2012. Improved endothelialization and reduced thrombosis by coating a synthetic vascular graft with fibronectin and stem cell homing factor SDF1alpha. Acta Biomater. 8, 1330–1338. Ennis, J., et al., 2008. Isolation, characterization, and differentiation of human umbilical cord perivascular cells (HUCPVCs). Methods Cell Biol. 86, 121–136. Feletou, M., 2011. The Endothelium: Part 1: Multiple Functions of the Endothelial CellsFocus on Endothelium-derived Vasoactive Mediators, San Rafael (CA). Gokcinar-Yagci, B., Celebi-Saltik, B., 2017. Comparison of different culture conditions for smooth muscle cell differentiation of human umbilical cord vein CD146+ perivascular cells. Cell Tissue Bank. 18, 501–511. Gokcinar-Yagci, B., et al., 2015. Pericytes: properties, functions and applications in tissue engineering. Stem Cell Rev. 11, 549–559. Gokcinar-Yagci, B., et al., 2016. Isolation, characterisation and comparative analysis of human umbilical cord vein perivascular cells and cord blood mesenchymal stem cells. Cell Tissue Bank. 17, 345–352. Grandi, C., et al., 2011. ECM-based triple layered scaffolds for vascular tissue engineering. Int. J. Mol. Med. 28, 947–952. Grassl, E.D., et al., 2003. A fibrin-based arterial media equivalent. J. Biomed. Mater. Res. A 66, 550–561. Greenwood-Goodwin, M., et al., 2016. A novel lineage restricted, pericyte-like cell line

5. Conclusions In conclusion, we generated triple-layered vascular construct with natural human ECM proteins/glycosaminoglycans mixed with SMCs and fibroblasts differentiated from HUV CD146+ PCs and with HUVECs in assistance with cell sheet engineering method. For the treatment of coronary artery diseases, this vascular construct is an important step for generation of fully-natural small-diameter (≤5 mm) vascular graft that has the structure closest to the native blood vessel. Our future efforts are directed toward the creation of intact small-diameter tubular vascular grafts cultivated in pulsatile flow bioreactor, in 111

Microvascular Research 118 (2018) 101–112

B. Gökçinar-Yagci et al.

Narita, Y., et al., 2008. Effects of transforming growth factor-beta 1 and ascorbic acid on differentiation of human bone-marrow-derived mesenchymal stem cells into smooth muscle cell lineage. Cell Tissue Res. 333, 449–459. Nees, S., et al., 2013. Focus on cardiac pericytes. Pflugers Arch. 465, 779–787. Owaki, T., et al., 2014. Cell sheet engineering for regenerative medicine: current challenges and strategies. Biotechnol. J. 9, 904–914. Pashneh-Tala, S., et al., 2015. The tissue-engineered vascular graft-past, present, and future. Tissue Eng. Part B Rev. 22 (1), 68–100. Ryan, A.J., O'Brien, F.J., 2015. Insoluble elastin reduces collagen scaffold stiffness, improves viscoelastic properties, and induces a contractile phenotype in smooth muscle cells. Biomaterials 73, 296–307. Sekiya, S., et al., 2010. Three-dimensional cell-dense constructs containing endothelial cell-networks are an effective tool for in vivo and in vitro vascular biology research. Microvasc. Res. 80, 549–551. Stefanska, A., et al., 2013. Renal pericytes: multifunctional cells of the kidneys. Pflugers Arch. 465, 767–773. Tong, Z., et al., 2011. Controlling the fibroblastic differentiation of mesenchymal stem cells via the combination of fibrous scaffolds and connective tissue growth factor. Tissue Eng Part A. 17, 2773–2785. Tsang, W.P., et al., 2013. CD146+ human umbilical cord perivascular cells maintain stemness under hypoxia and as a cell source for skeletal regeneration. PLoS One 8, e76153. Vezzani, B., et al., 2016. Not all pericytes are born equal: pericytes from human adult tissues present different differentiation properties. Stem Cells Dev. 25 (20), 1549–1558. Weinberg, C.B., Bell, E., 1986. A blood vessel model constructed from collagen and cultured vascular cells. Science 231, 397–400. Wilkins, E., L., W., Wickramasinghe, K., Bhatnagar, P., Leal, J., Luengo-Fernandez, R., Burns, R., Rayner, M., Townsend, N., 2017. European Cardiovascular Disease Statistics 2017. European Heart Network, Brussels. Williams, C., et al., 2012. A comparison of human smooth muscle and mesenchymal stem cells as potential cell sources for tissue-engineered vascular patches. Tissue Eng. Part A 18, 986–998. World Health Organization, 2017. Cardiovascular Diseases (CVDs). Yahagi, K., et al., 2016. Pathophysiology of native coronary, vein graft, and in-stent atherosclerosis. Nat. Rev. Cardiol. 13, 79–98. Yang, H., et al., 2011. Transforming growth factor-beta increases the expression of vascular smooth muscle cell markers in human multi-lineage progenitor cells. Med. Sci. Monit. 17, BR55–61. Zhou, M., et al., 2009. Development and validation of small-diameter vascular tissue from a decellularized scaffold coated with heparin and vascular endothelial growth factor. Artif. Organs 33, 230–239. Zilla, P., et al., 2007. Prosthetic vascular grafts: wrong models, wrong questions and no healing. Biomaterials 28, 5009–5027.

isolated from human embryonic stem cells. Sci. Rep. 6, 24403. Gui, L., et al., 2014. Construction of tissue-engineered small-diameter vascular grafts in fibrin scaffolds in 30 days. Tissue Eng. Part A 20, 1499–1507. He, W., et al., 2010. Pericyte-based human tissue engineered vascular grafts. Biomaterials 31, 8235–8244. Hoshi, R.A., et al., 2013. The blood and vascular cell compatibility of heparin-modified ePTFE vascular grafts. Biomaterials 34, 30–41. Ireland, H., et al., 2018. The survey on cellular and tissue-engineered therapies in Europe and neighboring Eurasian countries in 2014 and 2015. Cytotherapy 20, 1–20. Isenberg, B.C., et al., 2006. Endothelialization and flow conditioning of fibrin-based media-equivalents. Ann. Biomed. Eng. 34, 971–985. Kaessmeyer, S., et al., 2017. Subcellular interactions during vascular morphogenesis in 3D cocultures between endothelial cells and fibroblasts. Int. J. Mol. Sci. 18. Kajiyama, S., et al., 2015. Bone formation by human umbilical cord perivascular cells. J. Biomed. Mater. Res. A 103, 2807–2814. Kutcher, M.E., Herman, I.M., 2009. The pericyte: cellular regulator of microvascular blood flow. Microvasc. Res. 77, 235–246. Lee, C.H., et al., 2010. CTGF directs fibroblast differentiation from human mesenchymal stem/stromal cells and defines connective tissue healing in a rodent injury model. J. Clin. Invest. 120 (9), 3340. Loy, C., et al., 2016. A planar model of the vessel wall from cellularized-collagen scaffolds: focus on cell-matrix interactions in mono-, bi- and tri-culture models. Biomater. Sci. 5, 153–162. Madhavan, K., et al., 2010. Evaluation of composition and crosslinking effects on collagen-based composite constructs. Acta Biomater. 6, 1413–1422. van der Meer, A.D., et al., 2013. Three-dimensional co-cultures of human endothelial cells and embryonic stem cell-derived pericytes inside a microfluidic device. Lab Chip 13, 3562–3568. Mendes, L.F., et al., 2012. Perivascular-like cells contribute to the stability of the vascular network of osteogenic tissue formed from cell sheet-based constructs. PLoS One 7, e41051. Miwa, H., et al., 1993. An in vitro endothelialized compliant vascular graft minimizes anastomotic hyperplasia. ASAIO J. 39, M501–5. Montemurro, T., et al., 2011. Differentiation and migration properties of human foetal umbilical cord perivascular cells: potential for lung repair. J. Cell. Mol. Med. 15, 796–808. Mozaffarian, D., et al., 2015. Heart disease and stroke statistics—2015 update: a report from the American Heart Association. Circulation 131, e29–322. Muraoka, M., et al., 2013. Control of the formation of vascular networks in 3D tissue engineered constructs. Biomaterials 34, 696–703. Murray, I.R., G., Z., West, C.C., Corselli, M., Miranda-Carboni, G., Peault, B., 2014a. Osteogenic Differentiation of Perivascular MSC Precursors is Enhanced by Endothelial Cells. ORS 2014 Annual Meeting. Poster No: 1464. Murray, I.R., et al., 2014b. Natural history of mesenchymal stem cells, from vessel walls to culture vessels. Cell. Mol. Life Sci. 71, 1353–1374.

112