Composite vascular scaffold combining electrospun

journal of the mechanical behavior of biomedical materials 61 (2016) 12 –25

Available online at www.sciencedirect.com

www.elsevier.com/locate/jmbbm

Research Paper

Composite vascular scaffold combining electrospun fibers and physically-crosslinked hydrogel with copper wire-induced grooves structure Yuanyuan Liua,b,n, Chen Jianga, Shuai Lia, Qingxi Hua,b a

Rapid Manufacturing Engineering Center, Mechatronic Engineering and Automation of Shanghai University, Shanghai 200444, China b Shanghai Key Laboratory of Intelligent Manufacturing and Robotics, Shanghai 200444, China

ar t ic l e in f o

abs tra ct

Article history:

While the field of tissue engineered vascular grafts has greatly advanced, many inade-

Received 12 October 2015

quacies still exist. Successfully developed scaffolds require mechanical and structural

Received in revised form

properties that match native vessels and optimal microenvironments that foster cell

23 December 2015

integration, adhesion and growth. We have developed a small diameter, three-layered

Accepted 4 January 2016

composite vascular scaffold which consists of electrospun fibers and physically-

Available online 13 January 2016

crosslinked hydrogel with copper wire-induced grooves by combining the electrospinning

Keywords:

and dip-coating methods. Scaffold morphology and mechanics were assessed, quantified

TEVGs

and compared to native vessels. Scaffolds were seeded with Human Umbilical Vein

Hydrogel

Endothelial Cells (HUVECs), cultured in vitro for 3 days and were evaluated for cell viability

Electrospinning

and morphology. The results showed that composite scaffolds had adjustable mechanical

Dip-coating method

strength and favorable biocompatibility, which is important in the future clinical application of Tissue-engineered vascular grafts (TEVGs). & 2016 Elsevier Ltd. All rights reserved.

1.

Introduction

2010). Synthetic grafts, like expanded polytetrafluoroethylene and Dacron, work well in replacing high-flow, large-diameter

Cardiovascular diseases are the leading cause of morbidity

arteries (inner diameter46 mm), but are not suitable for

and mortality throughout the world (Hu et al., 2012). Vascular

small vessel reconstructions (inner diametero6 mm). This

diseases are commonly treated with autografts or blood

is due to the low blood flow and high shear stress which can

vessel transplantations. Traditional grafting methods, how-

cause thrombosis and neointimal hyperplasia (Hu et al.,

ever, have several problems. These include lack of suitable

2012). Tissue-engineered vascular grafts (TEVGs) represent a

harvest sites, additional surgical costs for the harvesting

potential source of alternatives and have been the focus of

procedure, pain, infection and lack of donors (Rainer et al.,

extensive research (Pooyan et al., 2012).

n Corresponding author at: Rapid Manufacturing Engineering Center, Mechatronic Engineering and Automation of Shanghai University, Shanghai 200444, China. E-mail address: [email protected] (Y. Liu).

http://dx.doi.org/10.1016/j.jmbbm.2016.01.002 1751-6161/& 2016 Elsevier Ltd. All rights reserved.


When designing an artificial bio-scaffold, the ultimate goal is to produce an extracellular matrix that is identical to its native counterpart (Pooyan et al., 2012). Since the extracellular matrix is a composite structure with a combination of fibrous proteins within a gelatinous grounded substance (Bosworth et al., 2013) and native blood vessels made of several layers (Rayatpisheh et al., 2014), using only one construction material cannot meet the requirements for a complex vascular structure. Additionally, it was found that the initial, low stiffness response of the artery wall is dominated by elastic fibers, whereas the high stiffness response during increased strain is dominated by tense collagen fibers. This arterial mechanical response has proven difficult to reproduce with a single component (McMahon et al., 2011). A multicomponent scaffold of interpenetrating layers may result in a composite construct with similar mechanical and structural features of native vessels. Recently, studies have begun to focus on making composite scaffolds using different techniques and materials (Centola et al., 2010; Liu et al., 2015; Rayatpisheh et al., 2014; Seo et al., 2009). For many reasons, electrospinning is widely used for fabricating one of the composite scaffold components. This technique can produce nano- and micron-scale fiber constructs, which have a large surface-to-volume ratio, are highly permeable and have an interconnected pore structure, all of which are biologically desirable (Centola et al., 2010). Nevertheless, electrospun fibers do not perfectly replicate the native structure of extracellular matrix (ECM). Most soft tissues are a composite made of a fibrous phase embedded in a hydrogellike substance (Shapiro and Oyen, 2013). Consequently, composite scaffolds that combine the electrospun fibrous mats and hydrogels could overcome the obstacles associated with either material alone. Some research groups are beginning to investigate composite scaffolds made of electrospun fibers and hydrogels in an attempt to overcome the individual material shortcomings. A few review papers (Bosworth et al., 2013; Butcher et al., 2014; Shapiro and Oyen, 2013) are available that discuss the various approaches used to create these novel composite structures and their intended applications. Methods used to combine fibers with hydrogels include layering, mixing hydrogels and fibers together and concurrent electrospinning and electrospraying (Butcher et al., 2014). Unfortunately, there are only a small number of these composite scaffolds used as TEVGs. This may be due to the presence of a significant amount of water (495%) in hydrogels. While this feature improves biocompatibility and transport properties, it is at the expense of mechanical integrity (An et al., 2015; Shapiro and Oyen, 2013), thus restricting its operability, such as coiling into a tubular structure. Native blood vessels have several layers: the inner layer (tunica intima) made of endothelial cells (ECs), the medial layer (tunica media) made of smooth muscle cells (SMCs) and the outer layer (tunica adventitia) made of fibroblasts (Rayatpisheh et al., 2014). Keeping in mind the native vasculature, our aim was to create a three-layered composite scaffold. The outer and inner layers of the composite scaffold were formed with electrospun PCL. The middle layer was formed by physically cross-linking polyvinyl alcohol (PVA) and sodium alginate (SA), which were prepared using the freeze-thaw cycling method to avoid damage from traditional

13

chemical crosslinking (Cho et al., 2005; Kamoun et al., 2015). Additionally, we incorporated copper wire-induced grooves to the scaffold by curling copper wire around a mandrel that functioned as the electrospinning collector. Adding the grooves was an effort to facilitate adhesion and proliferation of the ECs guided by the patterned structures to enhance and expedite the endothelialization process. Before seeding the cells, the scaffolds were treated with poly-L-lysine (PLL). With plentiful active amino groups and positive surface charges, PLL is most commonly used to promote cell adhesion and cell growth (Qi et al., 2014). We successfully fabricated threelayered composite vascular scaffolds (5 mm inner diameter) with grooves in the lumen using the electrospinning and dipcoating methods. This simple approach resulted in muchimproved tensile strength and favorable biocompatibility, which makes it a strong candidate for use in smalldiameter blood vessel transplantation.

2.

Material and methods

2.1.

Materials

Polycaprolactone (PCL) (Mn¼80,000), dichloromethane (DCM), N,N-dimethylformamide (DMF), Polyvinyl alcohol (PVA-124, degree of polymerization: 2400, degree of alcoholysis: 98– 99.8%) and sodium alginate (SA) were purchased from Sinopharm Chemical Reagent Co., Ltd. For electrospinning, PCL was dissolved in a 2:1 ratio of DCM: DMF (v/v) at a total concentration of 10% (wt). PVA was dissolved in deionized water at a 13% (wt) concentration. To prepare a homogeneous PVA solution, PVA was stirred and heated in a 90 1C water bath, then cooled to room temperature. SA was dissolved in water at a 3% (wt) concentration. The PVA and SA solutions were uniformly mixed at a 1:1 ratio (PVA-SA). The solutions were kept at room temperature until the bubbles disappeared. The PVA and sodium alginate mixtures were compatible and did not have any phase separation. Poly-L-lysine (PLL) solution was purchased from Shanghai Fusheng Co., Ltd, (China) and had an average molecular weight of 150–300 kDa.

2.2.

Scaffold preparation

To create the inner vascular scaffold layer, the PCL solution was delivered using a 20-gauge needle. The nanofibers were spun at a fixed electrical potential of 15 kV across 15 cm between the tip and a grounded collector. For the grounded collector, a 5-mm-diameter mandrel with copper wire wound around it was connected to a stepping motor with a 300 rpm rotating speed. The solution delivery rate was maintained at 20 μl/min with a micro-pump attached to the syringe. After electrospinning for a fixed time, the mandrel was removed from the stepping motor and placed in the PVA-SA solution for 20 min. This step ensured the PVA-SA solution completely filled the interstitial space of the nanofiber membrane. When no droplets dripped after removal from the PVASA solution, the mandrel was put in a refrigerator ( 50 1C) for 5 min to allow a very thin film of hydrogel to develop around the electrospun mesh. Then, the mandrel was dipped again in the PVA-SA solution for a few seconds to add another thin

14


hydrogel coat. This layer was also solidified at 50 1C for 5 min. Three freeze-thaw cycles ( 50 1C for 12 h, 20 1C for 6 h) were then conducted to physically crosslink the PVA. To fabricate the outer layer of the vascular scaffold, the mandrel was reconnected to the stepping motor to repeat the electrospinning process as described above. Finally, the scaffold was immersed in a 3% CaCl2 solution (wt) for 2 h to allow complete calcium ionic interaction between alginate and CaCl2. The interior copper wire was then easily pulled out because of its thinness and flexibility. The entire fabrication process can be seen in Fig. 1. For the control groups, we used composite scaffolds prepared without using copper wire and scaffolds made only from electrospun PCL prepared with and without using copper wire. We planned to prepare a scaffold of only hydrogel for a control group, but the hydrogel could not adhere to the mandrel without pre-deposited electrospun nanofibers. All scaffold thicknesses were made the same by adjusting the electrospinning and dipping times. After fabrication, samples could be gently pulled from the mandrel. The scaffolds were then lyophilized at 35 1C in a freeze-dryer (LGJ-10D, Beijing Sihuan, China) to form a well-connected pore structure.

2.3. Microstructure and surface wettability of the composite scaffolds Scaffold morphology was observed by scanning electron microscope (SEM, HITACHI SU-1510) and optical microscope (EV3020, Easson, China). For SEM detection, samples were transected into short segments, mounted on a specimen holder and then coated with gold to reduce electron overcharging effects. The optical microscope was used to examine the inner scaffold grooves (Fig. 2E) and outer scaffold layer delamination caused by electrospinning after CaCl2 immersion (Fig. 2F). To investigate the surface wettability of scaffolds, the static water contact angles (WCA) were captured by ChargeCoupled Device (CCD) camera and quantified using ImageJ software (Image Processing and Analysis in Java). 3.0 μl water droplets were carefully dropped onto scaffold surfaces. The average WCA value was calculated from water droplet measurements at 10 randomly distributed positions.

2.4.

Uniaxial tensile testing

To prepare for the uniaxial stretching test (Z2.5, ZWICK, Germany), each tubular construct was cut longitudinally into

Fig. 1 – (A) Fabrication process for composite scaffolds with copper wire-induced grooves. (B) The electrospinning collector, a 5-mm-diameter mandrel with copper wire wound around it. (C) Interior copper wire being easily pulled out.


15

Fig. 2 – Morphology of the tubular scaffolds. (A) Cross-sectional morphology of the composite scaffold showing a three-layer structure without voids between the layers. (B) Cross-sectional morphology of the electrospun scaffold without three-layer structure. (C) Inner surface morphology of the composite scaffold before freeze-drying showing that the electrospun fibers are embedded in the hydrogel. (D) Inner surface morphology of the composite scaffold after freeze-drying which shows rough lumps. (E) Grooves on the inner side of scaffold prepared with copper wire. (F) Delamination of the composite scaffold caused by electrospinning the outer layer after CaCl2 immersion.

a flat sheet (10 mm wide 50 mm long). Scaffold thickness was measured using cross-sectional SEM images. To ensure a firm but delicate hold between the metal tensile clamps, each specimen was sandwiched between two pieces of cardboard (Fig. 4C). The distance between the two clamps was used as the initial length. At room temperature, the specimens were pulled at a 2 mm/min crosshead speed until rupture. Assuming incompressibility and taking into account length and cross-sectional area, load-displacement curves were computed to determine stress–strain relationships (Soletti et al., 2010). Briefly, strain was calculated from the initial length of the sample (L1) and the stretched sample length when force (F) was loaded on the sample (LF) using the equation ε¼ (LF L1)/L1. The tensile stress was calculated using s ¼F/S, where S was the cross-sectional area of the

sample. In this case, the cross-sectional area was calculated as S ¼tw, where t is the thickness of the scaffold, and w is the width of the sample (10 mm). For the circumference stress strain analysis, scaffolds were cut into 5 mm-wide rings and clasped between mechanical clamps with two iron wires (Fig. 4D). The rings were loaded to failure at a displacement rate of 2 mm/min. Cross sectional area was calculated as S¼ 2th where t is the thickness of the scaffold and h is the width of the ring (5 mm). The tensile stress was calculated using equation above. From the stress–strain curves, ultimate tensile stress (UTS) and strain to failure (STF) were calculated from the maximum stress value before failure and the corresponding strain value, respectively. The linear modulus was defined as the slope of

16


the linear portion of the stress–strain curve between 25–80% of the sample's UTS (Bourget et al., 2012; Gauvin et al., 2011). Burst pressure was estimated using an adaptation of ÞðtÞ , where Laplace's law for intraluminal pressure: Pburst ¼ ðUCTS R UCTS is the circumferential UTC, t is the thickness and R is the initial internal diameter of the scaffold (Bourget et al., 2012; Gauvin et al., 2011; Ghezzi et al., 2012). The thickness was measured from cross-sectional SEM images, and R was 5 mm (the diameter of the tubular mandrel used to prepare the scaffold).

2.5.

Suture retention strength

Suture retention strength (SRS) is commonly used to measure the ability of sutures to adhere implants to surrounding tissues. Suture retention testing was performed using the same apparatus that was used for uniaxial testing. Each tubular scaffold was cut to obtain rectangular specimens (length¼ 30 mm, width¼10 mm). The short edge of each specimen was originally oriented on the circumference of the tubular scaffold. One end of the sample was clamped in the testing device. The other end of the sample was sutured 2 mm from the end using a 5-0 nylon loop (Ningbo Clinical Suture Needle Co., Ltd, China). The distance between two stitches was 2 mm. The sutures were attached to holes in a printed circuit board (PCB) connected to the testing device clamp (Fig. 7B). Sutures were pulled at an extension rate of 2 mm/s. SRS was calculated as the maximum force recorded prior to the sutures pulling out divided by the number of sutures.

2.6.

Scaffold PLL treatment and cell seeding

Before seeding the cells, the scaffolds were treated with PLL to promote cell adhesion. The PLL solution was prepared following the manufacturer's protocol. Briefly, a 1 mg/ml stock solution was prepared by dissolving 10 mg PLL in 10 ml phosphate-buffered saline (PBS). The stock solution was diluted to create a 0.1 mg/ml working solution. Before use, the PLL solution was filtered through a 0.22-μm filter. The scaffolds were cut along the axis and soaked in 75% alcohol for 2 h. After the surfaces were dried, scaffolds were immersed in the 0.1 mg/ml PLL solution for 30 min. The samples were remove from solution by aspiration and the surfaces were allowed to dry. Scaffolds were exposed to ultraviolet light for 30 min before seeding cells. Passage 3 HUVECs were harvested and suspended in culture medium at a density of 5.0 107 cells/ml. To form the cell-scaffold construct, 50 ml of cell suspension was added to each flat inner surface. Seeding was done in the bottom of a 24-well culture plate. After scaffolds were incubated for 4 h to let cells attach, 2 ml growth medium was added to each well. The cell-scaffold constructs were cultured in a humidified incubator at 37 1C with 5% CO2, and media was changed daily.

2.7.

Live-Dead cell staining assay

At 1 and 3 days after seeding, cell viability was checked. According to the manufacturer's protocol, samples were submerged in the Live-Dead Cell Staining Kit (BioVision,

Inc.) and viewed under a laser scanning confocal microscope (FV1000, Olympus Corporation, Japan).

2.8.

Cell attachment

After a 3-day incubation, the cell-scaffold constructs were fixed with 2.5% glutaraldehyde for 30 min. Then, samples were dehydrated using a graded ethanol series and dried. The samples were sputter-coated with gold and observed by SEM (HITACHI SU-1510).

2.9.

Statistical analyses

Statistical analyses were performed using SPSS v. 22 software (SPSS Inc., Chicago, IL). All values are reported as means7standard deviation. One-way ANOVA was performed to compare groups with different treatments, followed by multiple pairwise comparison procedure (Tukey test). Student's ttest was used for analyzing groove dimensions, contact angles and mechanical properties. Three independent experiments were carried out, and at least five samples per each test were taken for statistical analysis. Differences were considered significant when po0.05 (*) or po0.01 (**).

3.

Results

3.1.

Scaffold characterization

Composite and electrospun scaffold morphology is shown in Fig. 2. The cross-sectional morphology of the composite scaffold (Fig. 2A) showed a three-layer structure without voids between the layers. This demonstrates the feasibility of using electrospinning and dip-coating to fabricate vascular scaffolds with multilayered structures similar to native vessels. By adjusting the electrospinning and dipping times, the composite and electrospun scaffolds were made the same thickness. In this study, tubular scaffolds with 0.1 mm wall thickness were fabricated by coating with hydrogel two times. The cross-sectional morphology of an electrospun scaffold without three-layer structure is shown in Fig. 2B. As shown in Fig. 2C, the electrospun fibers were embedded in the hydrogel. This was because the nanofiber membrane's capillary action facilitated the wicking of hydrogel. Consequently, the two distinct structures completely merged together during the fabrication process. The nanofibers were coated with the more biocompatible hydrogel, and the soft hydrogel penetrated the non-woven nanofibers to coat the outer electrospun wall. This composite structure with the advantages of both nanofibers and hydrogel was better able to form a tubular structure. As an additional benefit, morphology of the composite scaffold after freeze-drying showed rough lumps (Fig. 2D) which could facilitate cell adhesion. Fig. 2E shows the grooves on the inner side of the scaffold prepared with 0.3 mm diameter copper wire. Calculations from scaffolds indicated that the average groove dimension was 0.317 mm, slightly larger than the copper wire. It should be noted that the distance between two grooves could be controlled. On the one hand, the positon of the groove is mainly determined by the position the copper wire. On the


other hand, the nanofibers are soft, so the fibers can be well deposited on the copper wire, which will guarantee the groove dimensions well. In this experiment, we set the distance between two grooves at 1 mm. This is a common issue that can cause delamination in multilayer scaffolds. During this experiment, the delamination phenomenon occurred in the scaffolds when the outer layer was electrospun after CaCl2 immersion to crosslink the alginate (Fig. 2F). However, we did not find delamination (Fig. 2A) when the scaffold was prepared following the steps in Fig. 1A. Fig. 3 shows the contact angle measurements from the electrospun and composite scaffolds. Fig. 3, panels A and C, show water droplets on the scaffolds immediately after deposition. Fig. 3, panels B and D, show the water droplets 120 s after deposition. Contact angles measured using ImageJ are listed in Table 1. The electrospun scaffold had an initial contact angle of 115741, which decreased to 65761 after 120 s. The composite scaffold had an initial contact angle of 85741, which decreased to 20731 after 120 s. These results confirm that introducing hydrogel into the electrospun fibers can increase the hydrophilicity of PCL, thus increasing its biocompatibility.

3.2.

Mechanical characterization of scaffolds

Load-displacement recordings were used to determine scaffold stress–strain curves in longitudinal and circumferential directions, as shown in Fig. 4. The results show that the composite scaffold tensile strength in the longitudinal

17

direction was greater than in the circumferential direction, indicating anisotropic properties in the scaffolds. The circumferential and longitudinal UTS and STF comparisons are shown in Fig. 5. Significant differences in UTS and STF were seen when comparing electrospun scaffolds to composite scaffolds, but no significant differences were detected between scaffolds with grooves and scaffolds without grooves. From this data, we can conclude that hydrogel increased the mechanical strength of the electrospun fibers, but grooves in the scaffolds did not change the structural mechanical properties. Electrospun scaffold longitudinal UTS values were significantly weaker than circumferential UTS values. This is because the circumferential direction is parallel to the nanofiber alignment, and the longitudinal direction is perpendicular to the nanofiber alignment. The weak longitudinal strength can be attributed to the weak connection between the neighboring parallel fibers. The composite scaffolds were stronger than the electrospun scaffolds, regardless of the measurement direction. Longitudinally, the composite scaffold UTS values were approximately 17-fold higher when compared to the Table 1 – Contact angles measured using ImageJ software. Data are from 3 independent experiments and are represented as mean7SD, n ¼ 3.

Electrospun scaffold contact angles Composite scaffold contact angles

0S

120 S

115741 85741

65761 20731

Fig. 3 – Water droplet contact angles on electrospun (A, B) and composite scaffolds (C, D). The decrease in contact angle is indicative of water absorption.

18


Fig. 4 – Stress–strain curves from uniaxial tensile testing of scaffolds (po0.05). (A) Longitudinal uniaxial tensile properties of composite and electrospun scaffolds. (B) Circumferential uniaxial tensile properties of composite and electrospun scaffolds. (C) Scaffold sandwiched between two cardboard pieces to ensure a firm grip for testing longitudinal mechanical properties. (D) Scaffold clasped between mechanical clamps with two iron wires for testing circumferential properties. electrospun scaffolds (Fig. 5A). When measured in the cir-

The estimated burst pressure results are presented in

cumferential direction, the composite scaffold UTS values

Fig. 6B. Estimated burst pressure values in the electrospun

were approximately 1.4-fold higher than the electrospun

scaffolds without and with grooves and the composite scaf-

scaffolds (Fig. 5B). The electrospun scaffolds withstood tensile

folds without and with grooves were: 993776, 936769,

loading parallel to the longitudinal axis at less than 65%

1377780, and 1293770 mmHg, respectively. These results

strain, while the composite scaffolds withstood up to 150%

are all lower that the burst pressure of the human saphenous

(Fig. 5C). Along the circumferential axis, the composite scaf-

vein (16807307 mmHg). However, our results were derived

fold STF values were also significantly higher than the

from scaffolds with only 0.1 mm thickness. Better mechanical

electrospun scaffolds (Fig. 5D). Importantly, we compared the longitudinal and circumfer-

strength could be achieved by increasing the dipping time to augment the thickness of the scaffold's middle layer.

ential tensile strengths of the composite scaffolds with grooves to native abdominal aorta, native coronary artery and a Teflon graft (TF-208) (Table 2). The table shows that

3.3.

Suture retention strength

the tensile properties of the composite scaffolds were closer to the natural human arteries than the commercial Teflon grafts. Grooves in the scaffolds made no significant differences in

In order to determine the potential performance of the scaf-

the linear moduli. Linear modulus values for composite

measured. Composite scaffold SRS values were significantly

scaffolds, however, were significantly lower than the electro-

higher than electrospun scaffolds (Fig. 7A), human mammary

spun scaffolds (Fig. 6A). These results show that introducing

arteries (1.4770.49 N) (Konig et al., 2009) and human saphe-

hydrogel to the nanofibers can reduce scaffold elasticity.

nous veins (1.7670.45 N) (Schaner et al., 2004).

folds in clinical tissue repair, the suture retention strength was


19

Fig. 5 – Scaffold mechanical properties calculated from uniaxial tensile testing. (A) Longitudinal UTS, (B) Circumferential UTS, (C) Longitudinal STF and (D) Circumferential STF for composite and electrospun scaffolds with and without grooves. * means po0.05, ** means po0.01. Table 2 – Longitudinal and circumferential tensile strengths of composite scaffold, Teflon (e-PTFE) graft, native abdominal aorta and native coronary artery. Material

Longitudinal Directions (MPa)

Composite scaffold 5.9 with grooves Teflon TF-208 (He 85.2 et al., 2009) Abdominal aorta 1.47 (He et al., 2009) Coronary artery 1.30 (Xu et al., 2015)

3.4.

Circumferential Directions (MPa) 4.3 66.7 5.29 1.43

Cell viability

HUVEC viability on the scaffolds at 1 and 3 days post-seeding is shown in Fig. 8. In the fluorescent images, green indicates live cells and red indicates dead cells. The quantitative results are shown in Fig. 9. Between days 1 and 3, the cell viability on electrospun scaffolds decreased from 8375% to 5773%, while cell viability on composite scaffolds increased from 8674% to 9076%. These results show that the composite scaffold of hydrogel and nanofibers is better for cell growth than the nanofibers alone.

Although the cells grew better on the composite scaffold, the copper wire-induced grooves had undesirable effects on the cells. As shown in Fig. 10A, the cells lined the grooves, indicating that HUVECs tended to migrate toward them on the seeding matrix. However, most of the cells concentrated on the groove edges (Fig. 10B), which suggests they had difficulty crossing the grooves.

3.5.

Cell attachment

In order to determine the cellular responses to the different scaffold structures, morphology of the cells at 1 and 3 days post-seeding was observed under SEM (Fig. 11). Compared to samples 1 day post-seeding (Fig. 11, panels A and D), samples at 3 days post-seeding had more cells that adhered and spread out on the scaffold surfaces. This shows that cells were able to grow on both types of scaffolds, but more cells adhered to the composite scaffold than the electrospun scaffold. The electrospun scaffold surface could still be seen 3 days post-seeding (Fig. 11B). While the composite scaffold surface could be seen 1 day post-seeding (black arrows shown in Fig. 11D), the cells covered almost all the scaffold's surface 3 days post-seeding (Fig. 11E). By increasing the magnification, the cell morphology differences could be observed. Cells morphologies were different on the electrospun and composite scaffolds. Most cells were ball-shaped on the electrospun scaffold (Fig. 11C), while most cells stretched into polygonal

20


shapes on the composite scaffold (Fig. 11F). These results suggest that the composite scaffolds are better for supporting HUVEC attachment.

4.

Discussion

Electrospinning is widely used to generate TEVGs, because the electrospun nanofibers can mimic the nanostructure of

Fig. 6 – (A) Linear modulus and (B) Estimated burst pressure of all scaffold groups. * means po0.05.

native blood vessels. Limitations like poor mechanical properties and limited cellular infiltration, however, can impede its applications. The introduction of hydrogel dip-coating could enhance the scaffold strength and allow better cell affinity to the electrospun nanofibers. By combining the electrospinning and dip-coating methods, this study has developed a reconstituted vascular model much closer to native arterial composition and architecture. Morphologically, our scaffold is biomimetic and has three layers – a PVA/SA layer mimicking the media and two electrospun nanofiber layers mimicking the intima and adventitia of vessels. We speculate that these features may be able to guide new tissue formation toward developing native-like features upon in vivo remodeling. Delamination is a common issue in developing multilayer scaffolds. The composite method we have used can efficiently address this issue. During our immersion process, the middle and inner layers are well merged because the hydrogel can penetrate into the nanofibers. The middle and outer layer coalescence can be attributed to SA crosslinking after the outer layer has been electrospun. When we crosslinked SA with CaCl2 before the outer layer was electrospun, delamination could be seen between the middle and outer layers. This phenomenon occurs because the calcium ions react with alginate to form an insoluble calcium alginate gel, decreasing SA binding ability and cohesiveness (Roopa and Bhattacharya, 2008). Thus, fabricating the outer layer before crosslinking SA can facilitate nanofiber embedding in the aqueous alginate. Congruently, the soft hydrogel is reinforced by the nanofibers, making the tubular scaffold easier to handle. We planned to prepare scaffolds of only hydrogel as a control group, but hydrogel could not adhere to the mandrel without pre-deposited electrospun nanofibers. Without capillary-action wicking to drive the hydrogel into the nanofibers, the hydrogel could not adhere to the surface of the mandrel. For both groups in the water contact angle experiment, the decrease in contact angle is indicative of water absorption. However, the electrospun scaffold contact angle was reduced

Fig. 7 – (A) Suture retention force of all scaffold groups. ** means po0.01. (B) Scaffold connected to printed circuit board with nylon sutures for testing suture retention strength.


21

Fig. 8 – Fluorescent images of cell viability. Green indicates live cells and red indicates dead cells. (A) Cells on electrospun scaffold without grooves 1 day post-seeding. (B) Cells on electrospun scaffold without grooves 3 days post-seeding. (C) Cells on composite scaffold without grooves 1 day post-seeding. (D) Cells on composite scaffold without grooves 3 days postseeding.

less than the composite scaffold contact angle. These results confirm the relatively high hydrophobicity of PCL. Thus, integrating the hydrogel with PCL can improve scaffold hydrophilicity. From the mechanical testing results, we found that grooves in the scaffolds did not change the mechanical properties of the structures. After hydrogel incorporation, the scaffold mechanical properties were significantly enhanced, especially in the longitudinal direction. These results confirm the advantages of introducing hydrogel into the electrospun nanofibers to enhance its mechanical properties. In our experiment, we used two crosslinking processes. First, the freeze-thaw method was used to induce PVA crosslinking. Then, the SA chains in PVA-cross-linked gels were chemically cross-linked in CaCl2 aqueous solution. However, the mechanical properties appeared to be determined by PVA. It has been recognized that the freeze-thaw process induces physical cross-linking of PVA chains by hydrogen bonding and crystallite formation (Cho et al., 2005). When starting with fresh polymer solution, the initial freezing step forms some ice crystals, which concentrates the polymer solution between the ice crystals and promotes PVA crystallite

formation. These crystallites act as crosslinking sites for polymer chains. Upon thawing, the ice crystals melt and create one region of polymer-rich gel surrounded by another region of polymer-poor solution. During successive freezethaw cycles, additional formation of ice crystals takes place in the polymer-poor region. The newly generated crystalline sites further enhance polymer crystallization in the gel, resulting in greater stiffness (Liu et al., 2009). Like most synthetic hydrogels, however, the use of PVA hydrogels has been limited by lack of inherent recognition sites required for protein and cell interactions, which are key determinants of biocompatibility (Xie et al., 2012). Alginate, a polysaccharide derived from brown seaweed, has been widely used as a cell carrier since the 1970s. Alginate can be rapidly cross-linked with divalent ions like calcium to become a gel. The 3D porous alginate scaffolds are usually fabricated by freeze-drying the calcium-cross-linked gel (Cho et al., 2005). Although alginate, as well as other natural polymers, has better biocompatibility than synthetic polymers, its practical application in porous cell scaffolds is limited due to brittleness.

22


Fig. 9 – % cell viability of electrospun and composite scaffolds without grooves. From day 1 to day 3, the cell viability on electrospun scaffolds decreased from 8375% to 5773%, while cell viability on composite scaffolds increased from 8674% to 9076%. * means po0.05.

Fig. 10 – Cell distribution on the composite scaffold with grooves 3 days post-seeding. (A) Fluorescent image (B) Merged fluorescent and morphology images. The images suggest that the cells have difficulty crossing the grooves. We used PVA and SA as the middle layer of the vascular scaffold because combining a synthetic polymer, PVA, with a natural polymer, alginate, can overcome the individual shortcoming of PVA and SA. But this didn’t mean that the middle layer is limited to only these two materials. Since SA and PVA have been widely used as cell, drug, protein carriers, it can be

expected that the performance of the scaffold can be greatly enhanced by adding some drug, protein, etc. Nevertheless, the composite scaffold linear moduli values were lower than the electrospun scaffolds. These results show that introducing hydrogel into the nanofibers can reduce the scaffold elasticity. Previous work (Holloway et al., 2013, 2011) has evaluated the impact of freeze-thaw cycling on compressive and tensile moduli on a range of PVA concentrations. Those results revealed a linear increase in the hydrogel modulus during the first six cycles at all polymer concentrations. With increasing numbers of freeze-thaw cycles, the degree of polymer phase separation, crystallite formation and hydrogen bonding were all increased (Holloway et al., 2011). Scaffold mechanical properties can thus be tailored based on the number of freeze-thaw cycles. The estimated burst pressures for the composite scaffolds were higher than the electrospun scaffolds. Although the estimated burst pressures for composite scaffolds with grooves (1293770 mmHg) are slightly lower than the human saphenous vein (16807307 mmHg), these results must analyzed in light of the geometry of the test specimens utilized. With respect to geometry, Laplace's law states that burst pressure increases linearly with increasing scaffold thickness ÞðtÞ ). Our if the internal diameter is kept constant (Pburst ¼ ðUCTS R results were obtained from scaffolds with 0.1 mm thickness. When the dipping time is increased to double the thickness to 0.2 mm, the burst pressure will be more than twice as high as a 0.1 mm scaffold. In this case, the scaffold UCTS value will also increase with the thicker hydrogel coating. Hence, better mechanical strength can be achieved through longer dipping times to increase the middle layer thickness of the scaffold. The fluorescent images clearly show the advantage of composite scaffolds over electrospun scaffolds. Cell viability was decreased at 3 days post-seeding on the electrospun scaffolds due to the PCL hydrophobicity. The increase of cell viability at 3 days post-seeding on the composite scaffolds could be explained by the fact that the cells first contacted the intimal surface of PCL and then contacted the media layer of hydrogel which was more biocompatible through cells’ proliferation. Although the HUVECs had a tendency to spread toward the grooves, the size of the cells made it difficult to cross the wide grooves. The relatively large depth and width of the gaps (copper wire diameter was 0.3 mm, distance between two grooves was 1 mm), compared to the cell size (roughly 20 μm), made it difficult for the cells to traverse the gaps to reach confluence. This was a significant limitation of the wire-induced-groove method. We also observed cell morphology of the electrospun and composite scaffolds by SEM. Images showed that the cells could adhere and spread on the surface of both scaffolds. However, more cells adhered to the surface of the composite scaffold, which facilitated the formation of the underlying basal lamina on the vascular scaffold lumen. HUVECs adopted a spindle or polygonal shape with fine filopodia on the composite scaffold. In contrast, cells on the electrospun scaffold had spherical or ellipsoidal shapes and showed minimal spreading. These observations could be explained by the stiffness and hydrophilicity of the scaffolds. In order to truly mimic the ECM, it is necessary to develop materials with mechanical properties conducive to cell


23

Fig. 11 – SEM images of cell attachment. (A) HUVECs on the surface of electrospun scaffold without grooves 1 day postseeding. (B) HUVECs on the surface of electrospun scaffold without grooves 3 days post-seeding. (C) Higher magnification of (B) showing that most cells are ball-shaped. (D) HUVECs on the inner surface of composite scaffold without grooves 1 day postseeding (the scaffold’s surface can be seen where black arrows point). (E) HUVECs on the inner surface of composite scaffold without grooves 3 days post-seeding in which the cells cover almost all the scaffold’s surface. (F) Higher magnification of (E) showing that most cells are stretched into polygonal shapes. development (Chen et al., 2014; Tibbitt and Anseth, 2009). For example, the mechanics-sensitive receptors on the cell surface receive signals and internalize them using special cell surface molecules. Those mechanical signals are translated into biological signals and can then affect cell biological behavior (Chen et al., 2014). The composite scaffolds were stiffer than the electrospun scaffolds, and cells were more likely to attach to the stiffer surface. This generally agrees with previously reported work showing that hydrogel mechanical properties can greatly affect cell adhesion, spreading and proliferation (Blacklock et al., 2010; Chen et al., 2014; Li et al., 2013). One reason our composite scaffold was conducive to cell development could be attributed to the water contact angle. As evidenced by the smaller water contact angle, introducing hydrogel promoted the hydrophilicity of PCL, thus providing enough specific binding sites for the integrin receptors to bind to (Legeay et al., 2010; Liu et al., 2014). After adhesion, cells can change their shape to fit the contact surface. In other words, after the cells connect to the scaffold by adhesion plaque, the scaffold rigidity determines the dynamic behavior of the cellular skeleton, directly affecting morphology (Chen et al., 2014). Cells can produce various shaped extensions, like filopodia, that enable them to consolidate their position on the receptor site and initiate colonization (Legeay et al., 2010). Cells can sense their surroundings through protrusions and receptor responses in order to bind to specific sites (Liu et al., 2014; Tibbitt and Anseth, 2009). In our experiment, the differences in cell morphology could be attributed to differences in surface topography. The polygonal shaped cells on the composite

scaffold were likely results of the rough lumps after freezethaw cycles (Fig. 2D). As this is an initial study, HUVEC-seeded tubular scaffolds are not to be used for implantation. The composite scaffolds we developed merely demonstrate the feasibility of nanofiber and hydrogel integration to prepare tubular scaffolds with adjustable mechanical strength and favorable biocompatibility. Microgrooves can provide a surface track to induce directional cell migration. It has been reported that patterns with 45–80 μm spacing and depths of no more than 1–2 mm may be optimal for EC adhesion, proliferation, and spreading (Melchiorri et al., 2013). Although the groove size is limited by the copper wire size, this method provides a more feasible and easy way to induce grooves when compared to the electron beam physical vapor deposition method. Future studies are underway to optimize the groove size for cell infiltration, differentiation and proliferation.

5.

Conclusion

To overcome the limitations of electrospun scaffolds, we developed a biomimetic scaffold with a hydrogel media between two layers of PCL nanofibers by combining the electrospinning and dip-coating methods. The composite scaffold we produced had a three-layer structure with no voids between the layers. The scaffold mechanical strength can be adjusted by changing the hydrogel dip times. Biological characterization after HUVEC seeding demonstrated that the scaffolds facilitated the attachment and migration of cells. The composite vascular recapitulated the microenvironment and architecture of native

24


cardiovascular tissues. These advances are a step toward integrating hydrogels into vascular scaffolds to provide a more suitable environment for cells to remodel into a native-like vascular substitute.

Acknowledgments This work was funded by National Natural Science Foundation of China (No. 51475281 and No. 51375292), National Youth Foundation of China (No. 51105239) and the open funded project of Jiangsu Key Laboratory of 3D Printing Equipment and Manufacturing (No. 3DL201504).

Appendix A.

Supplementary material

Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.jmbbm. 2016.01.002.

references

An, D., Ji, Y., Chiu, A., Lu, Y.-C., Song, W., Zhai, L., Qi, L., Luo, D., Ma, M., 2015. Developing robust, hydrogel-based, nanofiberenabled encapsulation devices (NEEDs) for cell therapies. Biomaterials 37, 40–48, http://dx.doi.org/10.1016/j. biomaterials.2014.10.032. Blacklock, J., Vetter, A., Lankenau, A., Oupicky´, D., Mo¨hwald, H., 2010. Tuning the mechanical properties of bioreducible multilayer films for improved cell adhesion and transfection activity. Biomaterials 31, 7167–7174, http://dx.doi.org/10.1016/ j.biomaterials.2010.06.002. Bosworth, L.A., Turner, L.-A., Cartmell, S.H., 2013. State of the art composites comprising electrospun fibres coupled with hydrogels: a review. Nanomed. Nanotechnol. Biol. Med. 9, 322–335, http://dx.doi.org/10.1016/j.nano.2012.10.008. Bourget, J.-M., Gauvin, R., Larouche, D., Lavoie, A., Labbe´, R., Auger, F.A., Germain, L., 2012. Human fibroblast-derived ECM as a scaffold for vascular tissue engineering. Biomaterials 33, 9205–9213, http://dx.doi.org/10.1016/j.biomaterials.2012.09.015. Butcher, A.L., Offeddu, G.S., Oyen, M.L., 2014. Nanofibrous hydrogel composites as mechanically robust tissue engineering scaffolds. Trends Biotechnol. 32, 564–570, http://dx.doi.org/ 10.1016/j.tibtech.2014.09.001. Centola, M., Rainer, A., Spadaccio, C., De Porcellinis, S., Genovese, J.A., Trombetta, M., 2010. Combining electrospinning and fused deposition modeling for the fabrication of a hybrid vascular graft. Biofabrication, 2, http://dx.doi.org/10.1088/ 1758-5082/2/1/014102. Chen, H., Liu, Y., Jiang, Z., Chen, W., Yu, Y., Hu, Q., 2014. Cellscaffold interaction within engineered tissue. Exp. Cell Res. 323, 346–351, http://dx.doi.org/10.1016/j.yexcr.2014.02.028. Cho, S.H., Oh, S.H., Lee, J.H., 2005. Fabrication and characterization of porous alginate/polyvinyl alcohol hybrid scaffolds for 3D cell culture. J. Biomater. Sci. Polym. Ed. 16, 933–947, http: //dx.doi.org/10.1163/1568562054414658. Gauvin, R., Guillemette, M., Galbraith, T., Bourget, J.-M., Larouche, D., Marcoux, H., Aube´, D., Hayward, C., Auger, F.A., Germain, L., 2011. Mechanical properties of tissue-engineered vascular constructs

produced using arterial or venous cells. Tissue Eng. – Part A 17, 2049–2059, http://dx.doi.org/10.1089/ten.tea.2010.0613. Ghezzi, C.E., Marelli, B., Muja, N., Nazhat, S.N., 2012. Immediate production of a tubular dense collagen construct with bioinspired mechanical properties. Acta Biomater. 8, 1813–1825, http://dx.doi.org/10.1016/j.actbio.2012.01.025. Holloway, J.L., Lowman, A.M., Palmese, G.R., 2013. The role of crystallization and phase separation in the formation of physically cross-linked PVA hydrogels. Soft Matter 9, 826–833, http://dx.doi.org/10.1039/c2sm26763b. Holloway, J.L., Spiller, K.L., Lowman, A.M., Palmese, G.R., 2011. Analysis of the in vitro swelling behavior of poly(vinyl alcohol) hydrogels in osmotic pressure solution for soft tissue replacement. Acta Biomater. 7, 2477–2482, http://dx.doi.org/ 10.1016/j.actbio.2011.02.016. Hu, J.-J., Chao, W.-C., Lee, P.-Y., Huang, C.-H., 2012. Construction and characterization of an electrospun tubular scaffold for small-diameter tissue-engineered vascular grafts: A scaffold membrane approach. J. Mech. Behav. Biomed. Mater. 13, 140–155, http://dx.doi.org/10.1016/j.jmbbm.2012.04.013. Kamoun, E.A., Kenawy, E.-R.S., Tamer, T.M., El-Meligy, M.A., Mohy Eldin, M.S., 2015. Poly (vinyl alcohol)-alginate physically crosslinked hydrogel membranes for wound dressing applications: Characterization and bio-evaluation. Arab. J. Chem. 8, 38–47, http://dx.doi.org/10.1016/j.arabjc.2013.12.003. Konig, G., McAllister, T.N., Dusserre, N., Garrido, S.A., Iyican, C., Marini, A., Fiorillo, A., Avila, H., Wystrychowski, W., Zagalski, K., Maruszewski, M., Jones, A.L., Cierpka, L., de la Fuente, L.M., L’Heureux, N., 2009. Mechanical properties of completely autologous human tissue engineered blood vessels compared to human saphenous vein and mammary artery. Biomaterials 30, 1542–1550, http://dx.doi.org/10.1016/j.biomaterials.2008.11.011. Legeay, G., Coudreuse, A., Poncin-Epaillard, F., Herry, J.M., BellonFontaine, M.N., 2010. Surface engineering and cell adhesion. J. Adhes. Sci. Technol. 24, 2301–2322, http://dx.doi.org/10.1163/ 016942410X508037. Li, P., Dou, X.-Q., Feng, C.-L., Zhang, D., 2013. Mechanical reinforcement of Coinf42o/inf4-phenyl-derived hydrogels for controlled cell adhesion. Soft Matter 9, 3750–3757, http://dx. doi.org/10.1039/c3sm27727e. Liu, W., Zhan, J., Su, Y., Wu, T., Wu, C., Ramakrishna, S., Mo, X., AlDeyab, S.S., El-Newehy, M., 2014. Effects of plasma treatment to nanofibers on initial cell adhesion and cell morphology. Colloids Surf. B: Biointerfaces 113, 101–106, http://dx.doi.org/ 10.1016/j.colsurfb.2013.08.031. Liu, Y., Vrana, N.E., Cahill, P.A., McGuinness, G.B., 2009. Physically crosslinked composite hydrogels of PVA with natural macromolecules: Structure, mechanical properties, and endothelial cell compatibility. J. Biomed. Mater. Res. – Part B Appl. Biomater. 90B, 492–502, http://dx.doi.org/10.1002/jbm.b.31310. Liu, Y., Xiang, K., Chen, H., Li, Y., Hu, Q., 2015. Composite vascular repair grafts via micro-imprinting and electrospinning. AIP Adv., 5, http://dx.doi.org/10.1063/1.4906571. McMahon, R.E., Qu, X., Jimenez-Vergara, A.C., Bashur, C.A., Guelcher, S.A., Goldstein, A.S., Hahn, M.S., 2011. Hydrogelelectrospun mesh composites for coronary artery bypass grafts. Tissue Eng. – Part C Methods 17, 451–461, http://dx.doi. org/10.1089/ten.tec.2010.0427. Melchiorri, A.J., Hibino, N., Fisher, J.P., 2013. Strategies and techniques to enhance the in situ endothelialization of small-diameter biodegradable polymeric vascular grafts. Tissue Eng. – Part B Rev. 19, 292–307, http://dx.doi.org/10.1089/ten.teb.2012.0577. Pooyan, P., Tannenbaum, R., Garmestani, H., 2012. Mechanical behavior of a cellulose-reinforced scaffold in vascular tissue engineering. J. Mech. Behav. Biomed. Mater. 7, 50–59, http://dx. doi.org/10.1016/j.jmbbm.2011.09.009.


Qi, W., Yuan, W., Yan, J., Wang, H., 2014. Growth and accelerated differentiation of mesenchymal stem cells on graphene oxide/ poly-l-lysine composite films. J. Mater. Chem. B 2, 5461–5467, http://dx.doi.org/10.1039/c4tb00856a. Rainer, A., Centola, M., Spadaccio, C., De Porcellinis, S., Abbruzzese, F., Genovese, J.A., Trombetta, M., 2010. A biomimetic three-layered compartmented scaffold for vascular tissue engineering. In: Proceedings of Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC’10 839–842. doi:10.1109/IEMBS.2010.5626795. Rayatpisheh, S., Heath, D.E., Shakouri, A., Rujitanaroj, P.-O., Chew, S.Y., Chan-Park, M.B., 2014. Combining cell sheet technology and electrospun scaffolding for engineered tubular, aligned, and contractile blood vessels. Biomaterials 35, 2713–2719, http://dx.doi.org/10.1016/j.biomaterials.2013.12.035. Roopa, B.S., Bhattacharya, S., 2008. Alginate gels: I. Characterization of textural attributes. J. Food Eng. 85, 123–131, http: //dx.doi.org/10.1016/j.jfoodeng.2007.07.012. Schaner, P.J., Martin, N.D., Tulenko, T.N., Shapiro, I.M., Tarola, N.A., Leichter, R.F., Carabasi, R.A., Dimuzio, P.J., 2004. Decellularized vein as a potential scaffold for vascular tissue engineering. J. Vasc. Surg. 40, 146–153, http://dx.doi.org/ 10.1016/j.jvs.2004.03.033.

25

Seo, H.J., Yu, S.M., Lee, S.H., Choi, J.B., Park, J.-C., Kim, J.K., 2009. Effect of PLGA Nano-Fiber/Film Composite on HUVECs for Vascular Graft Scaffold. IFMBE Proc. 23, pp. 2147–2150. doi:10.1007/978-3-540-92841-6_536. Shapiro, J.M., Oyen, M.L., 2013. Hydrogel composite materials for tissue engineering scaffolds. JOM 65, 505–516, http://dx.doi. org/10.1007/s11837-013-0575-6. Soletti, L., Hong, Y., Guan, J., Stankus, J.J., El-Kurdi, M.S., Wagner, W.R., Vorp, D.A., 2010. A bilayered elastomeric scaffold for tissue engineering of small diameter vascular grafts. Acta Biomater. 6, 110–122, http://dx.doi.org/10.1016/j. actbio.2009.06.026. Tibbitt, M.W., Anseth, K.S., 2009. Hydrogels as extracellular matrix mimics for 3D cell culture. Biotechnol. Bioeng. 103, 655–663, http://dx.doi.org/10.1002/bit.22361. Xie, L., Jiang, M., Dong, X., Bai, X., Tong, J., Zhou, J., 2012. Controlled mechanical and swelling properties of poly(vinyl alcohol)/sodium alginate blend hydrogels prepared by freezethaw followed by Ca2þ crosslinking. J. Appl. Polym. Sci. 124, 823–831, http://dx.doi.org/10.1002/app.35083.