Trt

PDa

b

c

d

e

a

ARR2AA

KRPEHGPH

1

crmyb

aec

h0

Colloids and Surfaces B: Biointerfaces 163 (2018) 19–28

Contents lists available at ScienceDirect

Colloids and Surfaces B: Biointerfaces

journa l homepage: www.e lsev ier .com/ locate /co lsur fb

he in vitro and in vivo biocompatibility evaluation of electrospunecombinant spider silk protein/PCL/gelatin for small caliber vascularissue engineering scaffolds

ing Xiang a,b, Shan-Shan Wang a,c, Meng He a,d, Yong-He Han c, Zhi-Hua Zhou a,eng-Long Chen c, Min Li a,∗, Lena Q. Ma e

College of Life Sciences, Fujian Normal University, Fuzhou, Fujian 350108, People’s Republic of ChinaState Key Laboratory of Pollution Control and Resource Reuse, School of the Environment, Nanjing University, Nanjing 210023, People’s Republic of ChinaQuangang Petrochemical Research Institute, Fujian Normal University, Quanzhou 362801, People’s Republic of ChinaDepartment of Physics and Astronomy, Macquarie University, Sydney, New South Wales 2122, AustraliaSoil and Water Science Department, University of Florida, Gainesville, FL 32611, United States

r t i c l e i n f o

rticle history:eceived 16 August 2017eceived in revised form5 November 2017ccepted 11 December 2017vailable online 13 December 2017

eywords:ecombinant spider silk proteinCLlectrospun vascular scaffoldsemocompatibilityenotoxicity

a b s t r a c t

Recombinant spider silk protein (pNSR32) and gelatin (Gt) were demonstrated to enhance cytocom-patibility of electrospun pNSR32/PCL/Gt scaffold. However, its potential pro-inflammatory effects andinteractions with tissue and blood are unknown. In this study, the physicochemical properties and in vitroand in vivo biocompatibility of such scaffolds were evaluated. The results showed that the pNSR32/PCL/Gtscaffold possessed larger average fiber diameters, wider fiber diameter distribution and faster degrada-tion rate than that of pNSR32/PCL and PCL scaffolds. The addition of pNSR32 and Gt had little influenceon the hemolysis and plasma re-calcification time, but prolonged kinetic clotting time and reduced theplatelet adhesion. The Il-6 and Tnf-˛ mRNA expression levels were up-regulated in macrophages seededon the PCL and pNSR32/PCL scaffolds. The lowest release of IL-6 and TNF-� appeared in the pNSR32/PCL/Gtscaffold. Histological results revealed that the PCL and pNSR32/PCL scaffolds elicited severe host tissueresponses after implantation, while prominent ingrowth of host cells were observed in the pNSR32/PCLand pNSR32/PCL/Gt scaffolds. The comet assay and bone marrow micronucleus test demonstrated that

ro-inflammatory responsesost tissue response

the pNSR32/PCL/Gt scaffold did not increase the frequency of DNA damage or bone marrow micronu-cleus. In short, this study confirmed that the pNSR32/PCL/Gt scaffold exhibited better blood and tissuecompatibility than pNSR32/PCL and PCL scaffolds. No induction of genotoxicity and inflammatory factorreleases makes the pNSR32/PCL/Gt scaffold a good candidate for engineering small diameter vascular

tissue.

. Introduction

Cardiovascular diseases pose a serious threat to human health,ausing 17.5 million deaths worldwide each year [1,2]. Blood vesseleplacement is an effective treatment for these diseases. As such,

ore than 1.4 million arterial bypass grafting were conducted perear in the USA [3]. The annual cost in the USA could reach 3.5illion dollars [4].

Autologous or allografts is widely utilized to replace diseased

nd damaged native blood vessels in clinical therapy [5]. How-ver, inadequate sources or ineligible donor vessels have limitedlinical success [6]. Recently, synthetic vascular grafts made from

∗ Corresponding author.E-mail address: mli@fjnu.edu.cn (M. Li).

ttps://doi.org/10.1016/j.colsurfb.2017.12.020927-7765/© 2017 Published by Elsevier B.V.

© 2017 Published by Elsevier B.V.

expanded polytetrafluoroethylene (ePTFE) or polyethylene tereph-thalate (PET) are successfully applied in large diameter syntheticvascular grafts with encouraging results for durability, long-termpatency rate, and safety [7]. However, in a femoropopliteal bypassgrafting trial, PET and ePTFE showed 36–47% patency after 2 yearsof implantation [8]. Additionally, rapid thrombosis, occlusion, inti-mal hyperplasia, and infection make them less acceptable in smalldiameter blood vessel replacement [9]. Therefore, it is urgent todevelop a feasible small diameter vascular graft for clinical appli-cations.

Recently, small diameter tissue-engineered vascular scaffoldholds promise for potential applications. A large number of

biodegradable synthetic polymers, including polycaprolactone(PCL) and poly(lactic-co-glycolic acid) (PLGA), with sufficientmechanical strength have been fabricated into tubular scaffoldsand seeded with vascular cells to develop functional blood vessel

2 ces B:

ssbphhhih

tfitptmltocitIa[tIgngp

wsnpmpr

2

2

GCswbipMiFoo

2

[wo

0 P. Xiang et al. / Colloids and Surfa

ubstitutes [10,11]. However, mounting studies demonstrated thatynthetic polymers are not ideal since most of them are hydropho-ic and lack bioactive sites [12]. Natural polymers (e.g., spider silkroteins and gelatin) provide a scaffold with bioactive sites andydrophilicity, facilitating further functionalization [13]. Therefore,ybrid materials combining both synthetic and natural polymersave been widely explored [6]. Previous studies found that compos-

te scaffolds with PCL and silk proteins and/or gelatin (Gt) showedigher cytocompatibility than single-material scaffolds [14,15].

The prosthetic scaffolds will be in direct and continuous con-act with blood. Poor hemo-compatibility can activate coagulationactors and complement systems to initiate immunological andnflammatory reactions, which may result in failure of vascularransplants [16]. Therefore, the interaction between blood com-onents and scaffold surfaces should be assessed before preclinicalrials [17]. In addition, existing data show that the nanostructure

aterials may disrupt the mitochondrial respiratory chain in a bio-ogical system to generate an excessive reactive oxygen specieso induce genotoxic activities [18]. The International Organizationf Standardization (ISO 10,993-1/EM 30, 993-1) also recommendsonducting the genotoxicity assessment of medical biomaterials ast is associated with carcinogenesis [19]. However, the informa-ion on genotoxicity of electrospun vascular scaffolds is limited.nflammatory responses after exposure to prosthetic scaffolds havelso attracted increasing attentions in vascular tissue regeneration6]. One study found that electrospun mats with different struc-ure showed distinct impacts on pro-inflammatory mediators (i.e.,l-6 and Tnf-a) mRNA expression in macrophages [20]. Collectively,iven the considerable uncertainties about the safety of electrospunanomaterials, it is imperative to study their hemocompatibility,enotoxicity and inflammatory response in vitro and in vivo beforere-clinical and clinical application as vascular grafts.

In the present study, a hybrid small-diameter vascular graftith inner diameter 5 mm was fabricated using recombinant spider

ilk protein (pNSR32), PCL and Gt by the electrospinning tech-ique. The blood-scaffold interactions and genotoxicity as well asro-inflammatory responses were evaluated. Furthermore, intra-uscular implantation of such electrospun scaffolds was also

erformed in Sprague Dawley (SD) rats to study in vivo host tissueesponses.

. Materials and methods

.1. Materials

PCL was purchased from Daicel Chemical Industries, Ltd, Japan.elatin (Gt), chemical reagents were brought from Sinopharmhemical Reagent Co., Ltd (Shanghai, China). Recombinant spiderilk protein (pNSR32, 102 KD containing Arg-Gly-Asp peptide sites)as prepared according to our previous description [21]. Fetal

ovine serum (FBS) was purchased from Hangzhou Sijiqing Biolog-cal Engineering Materials Co., Ltd. (Hangzhou, China); Cell culturelates, dishes and flasks were from Corning Incorporated (NY, USA);edical collagen sponges were from the Shanghai Qisheng Biolog-

cal Preparation Co Ltd. SD rats and Kunming (KM) mice were fromujian Medical University. Animal operations and care were carriedut according to the guidelines of the Animal Research Ethics Boardf Fujian Normal University, China.

.2. Electrospinning of small diameter vascular scaffolds

Electrospinning was performed in line with our previous work13]. Briefly, polymer solutions with a concentration of 30% (wt/v)ere prepared by dissolving pNSR32, PCL and Gt with a weight ratio

f 0:100:0 (A), 5:95:0 (B), 5:85:10 (C) in 98% formic acid, respec-

Biointerfaces 163 (2018) 19–28

tively. In the electrospinning process, the positive voltage (80 kV)was applied to a rotating metallic mandrel (5 mm outer diameter),while the syringe was grounded. The polymer solution was loadedinto a 5 mL disposable syringe with a blunt-end needle (0.6 mminternal diameter) at a constant flow rate of 80 �L/min. Fibers werecollected and deposited on the external surface of the mandrel at adistance of 10 cm from the tip of the needle to the mandrel (rotationrate at 2 rpm) to construct nanofiber tubular scaffolds. After elec-trospinning, the steel mandrel was treated with 70% (v/v) ethanolto take scaffolds down easily. The as-spun scaffolds were dried in afume cupboard at room temperature.

2.3. Scaffold characterization: morphology, FTIR and degradation

To investigate the morphology, tubular scaffolds were sputtercoated with gold and observed under Field Emission Scanning Elec-tron Microscopes (FE-SEM; Model JSM-7500F, JEOL LTD., Japan).Images were captured using a FE-SEM operating at an accelerat-ing voltage of 5 kV with a 9 mm working distance. Smile View Ver.2.1 (JEOL LTD., Japan) was used to calculate the mean and standarddeviation (sd) of the fibers diameters by measuring fifty randomfibers per image.

To identify the chemical structure, the attenuated totalreflectance Fourier transform infrared (ATR-FTIR) spectroscopicanalysis of electrospun scaffolds was conducted on Nicolet 5700spectrometer (Thermo Nicolet, US) over a range of 4000–500 cm−1

at a resolution of 0.09 cm−1.Electrospun scaffolds were dried to constant weight and

recorded the initial weight (W0). After that, they were placed in20 mL vials each containing 10 mL of 0.1 M, pH 7.4 phosphate-buffered saline (PBS) or multi-enzyme solution (MES, 2.0 × 106 Ulysozyme, 2.5 × 103 U trypsin and 4.0 × 102 U pancreatic lipase dis-solved in 0.1 M PBS). The vials were placed into an incubatorchamber at 37 ◦C and observed up to 20 weeks. Every two weeks,triplicate specimens for each scaffold were taken out and washed3 times with distilled water and vacuum freeze-dried. The PBS andMES were renewed every two weeks. The weight of each specimen(W1) was carefully measured and the rate of residual mass wascalculated as follow:

The rate of residual mass (%) = W1

W0× 100%

2.4. Hemocompatibility analysis: hemolysis, plasmarecalcification time, kinetic clotting time and platelet adhesion

Blood was drawn from SD rats into heparin anticoagulation vac-uum tubes. Electrospun scaffolds were cut into 5 mm × 5 mm piecesand placed into clean 10 mL conical glass test tubes. After that,5 mL saline was added into each tube of tested and negative con-trol, whereas positive control was added with 5 mL sterile distilledwater. All tubes were pre-incubated for 30 min at 37 ◦C. Anticoag-ulated blood was diluted (4 mL blood in 5 mL saline) and added toeach tubes (100 �L per tube). All tubes were incubated under gen-tle agitation for 60 min at 37 ◦C. Tested samples were taken out andeach tube was centrifuged at 2000 rpm for 5 min. The supernatantfrom each test was aspirated into a cuvette for measurement ofabsorbance at 545 nm. The hemolysis percentage (HP) was evalu-ated by the following equation:

HP (%) = At − AncA − A

× 100%

pc nc

Where At is the absorbance of tested samples; Anc and Apc are theabsorbance of the negative and positive control, respectively. TheHP was calculated by mean of three measures for each scaffold.

P. Xiang et al. / Colloids and Surfaces B: Biointerfaces 163 (2018) 19–28 21

Table 1qRT-PCR primer sequences for pro-inflammatory genes analysis.

Genes Forward primer (5′-3′) Reserve primer (5′-3′) Accession no.

pwsbc3tf

wte(ai7po

bfAwteac

2

i(s1tpFTitPqet2TlL

2h

5lD

Il-6 CTGCAAGAGACTTCCATCCAGTT

Tnf-˛ AGCACAGAAAGCATGATCCG

ˇ-actin ATCACTATTGGCAACGAGCG

To evaluate the possible delay induced by scaffolds in plateletoor plasma (PPP) in the presence of Ca2+. Anticoagulated bloodas centrifuged at 2000 rpm for 20 min to isolate PPP. Electrospun

caffolds were cut into 5 mm × 5 mm pieces and placed into theottom of glass tubes. Tubes without scaffolds were served as aontrol. PPP (200 �L) was dropped onto each tube and incubated at7 ◦C for 30 min. Subsequently, 200 �L of 0.025 M CaCl2 was addedo each tube except negative control. The time of onset of fibrinormation was recorded.

To compare the kinetic clotting time, three types of scaffoldsere cut into small pieces (5 mm × 5 mm) in triplicate and posi-

ioned into the bottom of 50 mL Erlenmeyer flasks. Coverslips weremployed as a control. Then, 100 �L anticoagulant citrate dextroseACD) anticoagulated blood was dropped onto the surface of piecesnd coverslips, followed by injecting 12.5 �L of 0.2 M CaCl2 solutionnto each blood drop and keeping at 37 ◦C for 5, 10, 20, 30, 50 and0 min. After each fixed interval of time, 50 mL distilled water wasut into each flask and carefully shaken for 10 min. The absorbancef hemoglobin was detected by a spectrophotometer at 540 nm.

The platelet solution was prepared from ACD anticoagulatedlood by centrifugation at 2000 rpm for 20 min. Electrospun scaf-olds were incubated with platelet solution at 37 ◦C for 30 min.fter that, the suspension was removed and scaffolds were washedith PBS for 3 times. Adherent platelets were fixed with 2.5% glu-

araldehyde for 2 h, followed by dehydration in a graded series ofthanol, and lyophilization. The samples were then sputter coatednd observed under FE-SEM. The number of adherent platelets wasalculated by measuring five random views per image.

.5. Pro-inflammatory response to electrospun scaffolds

To test whether the electrospun scaffolds perturb pro-nflammatory factors expression, murine RAW264.7 macrophagesATCC, USA) were seeded on the surface of sterilized electrospuncaffolds in Dulbecco’s Modified Eagles Media supplemented with0% FBS and 1% penicillin-streptomycin solution (DMEM). Cellsreated with 10 ng/mL lipopolysaccharide (LPS) were used as aositive control. After incubation in a 5% CO2 incubator (Thermoisher, USA) at 37 ◦C for 24 h, the total RNA was isolated usingrizol reagent (Life Technologies, USA) and reversely transcribed

nto cDNA using AMV Reverse Transcriptase (Promega Corpora-ion, USA). Specific primers (Table 1) and Power SYBR

®Green

CR Master Mix (Applied Biosystems, USA) were employed foruantitative real-time PCR (qRT-PCR) analysis of Il-6 and Tnf-axpression. Housekeeping gene ˇ-actin was served as a reference,he fold change for target genes was calculated by the formula−��Ct. Cell-free supernatants were harvested to detect IL-6 andNF-a concentrations in a protein level using mouse enzyme-

inked immunosorbent assay (ELISA) kits (Yi Fei Xue Biotech. Co.,td.,Nanjing, China) based on the manufacturer’s manual.

.6. In vivo biocompatibility testing: scaffold implantation andistopathological analysis

Electrospun scaffolds were cut into strips with a size of mm × 2 mm and then immersed in 75% ethanol for 30 min fol-

owed by washing 3 times with sterilized PBS and neutralized withMEM prior to implantation. Fifteen SD rats were randomly divided

AGGGAAGGCCGTGGTTGT NM 031168CTGATGAGAGGGAGGCCATT NM 013693TCAGCAATGCCTGGGTACAT NM 007393

into five groups (n = 3), designated blank control, negative control,PCL, pNSR32/PCL, and pNSR32/PCL/Gt. For intramuscular implanta-tion, the SD rats were anesthetized by an intraperitoneal injectionof chloral hydrate with a dose of 5 mg/kg·bw. After the skin wasshaved and sterilized with iodine, the longitudinal skin incisionswere made using blunt dissection near the spine. The length anddepth of the incision into the muscle was ∼8 and 4 mm, respec-tively. Each sample was placed in each pouch, two strips per rat.Subsequently, the fascia and skin of the rats were sutured. The sur-gical procedure was performed under general sterile conditions.Medical collagen sponge with similar size was used as the negativecontrol.

After two or four weeks’ implantation, the rats were sacri-ficed for the histopathological evaluation. Briefly, the implantedareas were dissected, the scaffolds with surrounding tissue wereretrieved and immediately fixed in Bouin’s solution for 24 hfollowed by dehydration in 70%, 80%, 90%, and 100% ethanol.They were then embedded in paraffin and cut the paraffin blockwith 7 �m section on a rotation microtome (MLCROM-HM340,Germany). The sections were stained with hematoxylin and eosin(H&E) and visualized under BX51 light microscope (Olympus,Japan).

2.7. Genotoxic assessment: comet assay and micronucleus test

To study whether degradation products of electrospun scaffoldsinduce DNA damage of vascular endothelial cells, the comet assaywas conducted in primary SD rat aortic epithelial cells (SDRAECs)as described by Braghirolli et al. [22]. Briefly, scaffold extracts wereprepared by immersion of electrospun scaffolds into DMEM at37 ◦C for 72 h, and then SDRAECs were treated with these extractsfor 48 h. Positive and negative controls were fed with 0.1% H2O2(diluted by DMEM) and DMEM, respectively. Subsequently, 10 �Lof cell suspension was mixed with 0.8% low-melting point agaroseand immediately placed onto a glass slide, pre-coated with normalmelting point agarose (1%). After polymerization at 4 ◦C, the slideswere treated by ice-cold lysis buffer (2.5 M NaCl, 100 mM Na2EDTA,and 10 mM Tris, pH 10.0, with 1% Triton X-100 and 10% dimethylsulfoxide) at 4 ◦C. The slides were then put into electrophoresisbuffer solution (300 mM NaOH, 1 mM Na2EDTA, pH > 13) for 20 minto unfold the DNA, followed by electrophoresis for 20 min at 25 V.After these, the slides were immersed into neutralized solution(Tris 0.4 M, pH 7.4) for 10 min and stained with ethidium bromide(20 �g/mL). The slides were allowed to dry at room temperatureand observed under a BX51 fluorescence microscope. One hun-dred cells per group (50 cells per slide, two slides per group) werecaptured and analyzed by Comet Assay Software Project (CASP) inrelation to DNA migration. Parameters calculated by CASP are tailmoment (TM), tail length (TL) and percentage of tail DNA (TD).

The KM mice with weight of 20 ± 2 g were randomly assigned tothree groups (5 females and 5 males, 10 for each): a tested group, apositive control and a negative control. They were intraperitoneallyinjected with 1 mL extracts, cyclophosphamide (1.2 mg/mL), orsaline, respectively twice at an interval of 24 h. After 6 h of the

last administration, the mice were sacrificed by cervical disloca-tion. Slides were prepared from the bone marrow according toa previous report [23]. After smearing, all slides were air-dried,fixed in a mixture of methanol and glacial acetic acid (3:1, v/v) for

22 P. Xiang et al. / Colloids and Surfaces B: Biointerfaces 163 (2018) 19–28

CL/Gt

1dz(wo

2

cUD

3

3s

tfwbweoscot(

3

fsis(mib3b

Fig 1. Macroscopic view of the electrospun tubular pNSR32/P

0 min, stained with 20% Giemsa solution for 30 min, washed withistilled water, and covered with coverslips after dimethylben-ene treated. For each mouse, 1.0 × 103 polychromatic erythrocytesPCE) were observed under microscope (Olympus, Japan) and PCEith micronucleus were recorded and represented as per thousand

f the total PCE number.

.8. Statistical analysis

Results are expressed as mean ± sd. Statistical analyses wereonducted using ANOVA by Origin Pro 8.5 (OriginLab Corporation,SA) and Graphpad Prism Version 6 (Graphpad Software, USA).ifferences were considered statistically significant at the p < 0.05.

. Results

.1. Morphology and fiber diameter distribution of electrospuncaffolds

The electrospun tubular scaffold, with a length of 2 cm, wallhickness of 0.3 mm and an inner diameter of 5 mm, was success-ully fabricated (Fig. 1). Interestingly, different tubular scaffolds

ith different wall thickness and inner diameter can be madey controlling the electrospinning time and using steel mandrelsith different external diameter. The micro-topography of the

lectrospun PCL, pNSR32/PCL and pNSR32/PCL/Gt nanofibers wasbserved by FE-SEM (Fig. 2). Randomly oriented fibers, smoothurface, and interconnected pore structure appeared in all threeases. Incorporation of pNSR32 did not affect the fiber diameterf PCL scaffolds (Fig. 2B, b). However, the average fiber diame-er of pNSR32/PCL/Gt scaffolds (171 ± 23 nm) was larger than PCL116 ± 30 nm) and pNSR32/PCL (112 ± 23 nm) scaffolds.

.2. ATR-FTIR analysis and scaffolds degradation

ATR-FTIR spectra of PCL, pNSR32/PCL and pNSR32/PCL/Gt scaf-olds are shown in Fig. 3. Representative spectrum peaks for PCLcaffolds were observed at 2944 cm−1 (asymmetric CH2 stretch-ng), 2860 cm−1 (symmetric CH2 stretching), 1720 cm−1 (C Otretching), 1290 cm−1 (C O and C C stretching) and 1240 cm−1

asymmetric C O C stretching). The pNSR32-related stretchingodes appeared at 1640 cm−1 (amide I) and 1545 cm−1(amide II)

n pNSR32/PCL and pNSR32/PCL/Gt scaffolds [24]. Several typicalands of Gt were also observed in composite scaffolds, including445 cm−1 (N H stretching of amide bond) and common amideands of Gt at 1650 cm−1 (amide I) and 1545 cm−1 (amide II).

scaffold: (A) the inner diameter and length (B) of the scaffold.

The amide I band at 1640 cm−1 implied �-sheet of pNSR32 and1545 cm−1 (amide II) was attributable to both �-helix and randomcoil conformation. The main conformation of Gt in composite scaf-folds included both �-helix and random coil due to the amide I(1650 cm−1) and amide II (1545 cm−1) peaks [25].

Bedsides ATR-FTIR analysis, we evaluated the electrospunscaffolds degradation in phosphate-buffered saline (PBS) andmulti-enzyme solution (MES). Results showed that electrospunscaffolds in MES degraded faster than that in PBS (Fig. 4). No signifi-cant degradation variation was observed among all scaffolds duringthe first 6 weeks in PBS (Fig. 4A). However, pNSR32/PCL/Gt experi-enced a slightly fast mass loss after 6 weeks compared to PCL andpNSR32/PCL (Fig. 4A). The residual mass of PCL, pNSR32/PCL andpNSR32/PCL/Gt after 20 weeks was 84%, 80% and 78%, respectively(Fig. 4A). Unlike PBS, pNSR32/PCL and pNSR32/PCL/Gt scaffoldsdisplayed similar degradation rate, but it was faster than PCLscaffolds during the initial 8 weeks in MES (Fig. 4B). After 12weeks, the residual mass of pNSR32/PCL/Gt scaffold decreased tobelow 80% (Fig. 4B). Moreover, the mass loss of pNSR32/PCL andpNSR32/PCL/Gt scaffolds was accelerated thereafter. The residualmass was 78% (PCL), 74% (pNSR32/PCL) and 68% (pNSR32/PCL/Gt)after 16 weeks (Fig. 4B). After 20 weeks, pNSR32/PCL/Gt scaffoldsexhibited the lowest residual mass, followed by pNSR32/PCL andPCL scaffolds (Fig. 4B).

3.3. Electrospun scaffolds hemocompatibility

The HP of ACD blood with electrospun scaffolds is shown inFig. 5A. The HP for all electrospun scaffolds was less than 1%, whichwas lower than the 5% requirement. Plasma re-calcification time(PRT) was measured to evaluate the activation of the intrinsic coag-ulation system. The onset of fibrin formation in three cases andcontrol was similar, indicating a close clotting time among elec-trospun scaffolds and TCP (Fig. 5B). For the negative control (PPPin the absence of Ca2+), it didn’t form any clots. Kinetic clottingtime is usually employed to assess the antithrombotic propertiesof biomaterials [26]. In this study, at each time, blood incu-bated with coverslips had a lower absorbance than pNSR32/PCL/Gt,pNSR32/PCL and PCL scaffold (Fig. 5C). However, there was a sig-nificant difference in the degree of clotting among pNSR32/PCL,PCL and pNSR32/PCL/Gt scaffolds. Excluding 10 min, blood incu-bated with pNSR32/PCL/Gt scaffold had a lower degree of clotting

than that with pNSR32/PCL and PCL scaffolds, whereas no differ-ence in absorbance was observed between pNSR32/PCL and PCLscaffolds after 30 min incubation (Fig. 5C). FE-SEM micrographsshowed that platelets incubated with scaffolds for 30 min, few

P. Xiang et al. / Colloids and Surfaces B: Biointerfaces 163 (2018) 19–28 23

ograms of (A, a) PCL, (B, b) pNSR32/PCL, and (C, c) pNSR32/PCL/Gt.

w(sP

3c

mmti(BppuRsIcs

Fig. 2. SEM micrographs and fiber diameter distribution hist

ere deposited on the surface of scaffolds and still remained roundFig. 5DEF). However, the platelet counts on the pNSR32/PCL/Gtcaffold (Fig. 5F) were less than that on pNSR32/PCL (Fig. 5E) andCL (Fig. 5D) scaffolds.

.4. The mRNA expression and secretion of pro-inflammatoryytokines

The acute pro-inflammatory response was evaluated by deter-ining the level of the inflammatory cytokines generated by theacrophages cultured on the electrospun scaffolds. Data showed

hat the mRNA expression of Il-6 and Tnf- was up-regulatedn RAW264.7 cells seeded on the PCL and pNSR32/PCL scaffoldsp < 0.05), but not on the pNSR32/PCL/Gt scaffold (Fig. 6A and), suggesting that PCL and pNSR32/PCL scaffolds might elicitro-inflammatory responses. To further verify our results, theroduction of IL-6 and TNF-a at protein level was measuredsing commercial ELISA kits. Background level was tested fromAW264.7 cells grown on TCP, with activated level being mea-

ured from cells treated with 10 ng/mL LPS. Higher increases ofL-6 and TNF-a were observed from LPS stimulated RAW264.7 cellsompared to background control (Fig. 6C and D). Interestingly, allcaffolds did not affect IL-6 secretion (Fig. 6C). However, TNF-a

Fig. 3. FTIR spectra of electrospun PCL (A), pNSR32/PCL (B), and pNSR32/PCL/Gt (C)scaffolds.

secretion was elevated when RAW264.7 cells cultured on the PCLscaffold (Fig. 6D).

3.5. In vivo host tissue response and genotoxicity assessment

All electrospun scaffolds and medical collagen sponge were suc-cessfully intramuscularly implanted into SD rats. After 4 weeks,

24 P. Xiang et al. / Colloids and Surfaces B: Biointerfaces 163 (2018) 19–28

he mu

ntiatasAimsiostitpa

asaittfHt(

i(tat

Fig. 4. Mass loss profile of electrospun scaffolds in PBS (A) or in t

o abnormal conditions were observed and all scaffolds were wellolerated in the host rats. Two weeks after implantation, notablenflammatory signs (i.e., accumulation of phagocytes, lymphocytesnd multinucleated foreign body giant cells) were visible adjacento PCL (Fig. 7C) and pNSR32/PCL (Fig. 7D) scaffolds. However, just

few of inflammatory cells appeared around the pNSR32/PCL/Gtcaffolds (Fig. 7E) and medical collagen sponge implants (Fig. 7B).dditionally, all scaffolds showed no invading host cells inside the

mplants. The signs of degradation were clearly seen in implantededical collagen sponge (Fig. 7B) while all scaffolds still held

tructural integrity (Fig. 7CDE). Four weeks after implantation, thenflammatory responses to all implants were attenuated basedn the disappearance of inflammatory cells in surrounding tis-ues (Fig. 7G–J). Besides, degradation became more apparent inhe medical collagen sponge (Fig. 7G), it disappeared 4-week aftermplantation. However, all scaffolds were visible at the implan-ation sites (Fig. 7HIJ). In addition, the pNSR32/PCL (Fig. 7I) andNSR32/PCL/Gt (Fig. 7J) scaffolds showed prominent infiltrationnd ingrowth of host cells.

To determine whether electrospun scaffolds induces DNA dam-ge of vascular cells, SDRAECs were exposed to pNSR32/PCL/Gtcaffold extract for 48 h. The results of comet assay are presenteds% migrated DNA, tail lengths and tail moments (Fig. 8A). A signif-cant difference of DNA damage between the scaffold extract andhe positive control was observed. The pNSR32/PCL/Gt extract ledo less DNA damage (1.10 ± 0.15%, 3.69 ± 0.73%, and 0.11 ± 0.02%or DNA migration, tail length and tail moment, respectively) than

2O2 treatment (18.3 ± 1.4%, 14.4 ± 2.9% and 2.4 ± 0.7%, respec-ively) (p < 0.001), while it was similar to that of the negative control1.2 ± 0.2%, 3.7 ± 0.8% and 0.13 ± 0.02%, respectively).

The bone marrow micronucleus test was performed afterntraperitoneal injection of the scaffold extract, cyclophosphamide

1.2 mg/mL, positive control) and saline (negative control) respec-ively twice at an interval of 24 h into KM mice. There were 10 foldss much as micronuclei production in positive control comparedo the scaffold extract (p < 0.001) (Fig. 8B). However, there was no

ltiple enzyme solution (MES) (B) with in vitro degradation time.

significant difference between the scaffold extract and the negativecontrol.

4. Discussion

In our previous study, valuable tubular vascular scaffolds werefabricated using a combination of PCL, pNSR32 and Gt, whichexhibited many desirable characteristics, including high porosity,hydrophilicity, and good cytocompatibility. In present study, ourinvestigations focused on other key aspects of biochemical proper-ties and in vitro and in vivo biocompatibility.

The surface morphology of pNSR32/PCL/Gt and pNSR32/PCLwas similar to the PCL scaffold, which consisted of randomly ori-ented fibers and thoroughly interconnected pore structure (Fig. 2).However, the pNSR32/PCL/Gt scaffold showed larger average fiberdiameter and wider fiber diameter distribution (70–300 nm) thanthat of pNSR32/PCL (60–180 nm) and PCL (60–180 nm) scaffolds(Fig. 2). This could be due to an increase in the viscosity of theelectrospun solution after addition of Gt [27]. In addition, our pre-vious study demonstrated that Gt also improved the hydrophilicity,which affected the adherence and growth of rat vascular endothe-lial cells on PCL scaffolds [13].

Degradation properties of scaffolds are of crucial importance,which influence not only the mechanical and structural integrity,but also cellular processes including cell proliferation, tissue regen-eration, and host responses over time [28]. Lam et al. (2009) foundthat the PCL scaffold fabricated by in-house built extruder justshowed 2.1% mass loss after 24 weeks’ degradation in PBS [29].However, in our study, mass loss of PCL-scaffold reached to 16%after 20 weeks’ degradation in PBS, which might be attributedto higher porosity (83% vs. 70%) that increased exposure area inour PCL-scaffold [13,29]. However, biodegradation of PCL-scaffold

occurred at a slow rate, it could be between 6 months to 2 yearsin vivo, limiting its utilizations in tissue engineering scaffold [30].In an attempt to improve the biodegradation rate, PCL was blendedwith other natural polymers (e.g., collagen, Gt, or silk protein)

P. Xiang et al. / Colloids and Surfaces B: Biointerfaces 163 (2018) 19–28 25

F lasmao eprese

wTpcditoteooiPt(anr

embI

ig. 5. Hemocompatibility of electrospun scaffolds. (A) Hemolysis percentage, (B) Pn PCL (D), pNSR32/PCL (E), and pNSR32/PCL/Gt (F) scaffolds. The different letters r

ith high susceptibility to hydrolysis and/or enzymolysis [15,31].he present study developed a PCL-based scaffold blended withNSR32 and Gt. As expected, the mass loss of scaffolds signifi-antly increased both in PBS and MES (Fig. 4). This was possiblyue to the addition of pNSR32 and Gt to PCL-scaffold, produc-

ng higher porosity, hydrophilicity and crystallinity susceptibilityo hydrolysis [13,32], which contributed to the faster degradationf pNSR32/PCL/Gt scaffold. In addition, most of in vitro degrada-ion studies were performed in PBS [28,33], which fail to considernzymatic digestion. The PCL degradation in vivo depends on notnly hydrolysis, but also enzymatic actions to caproic acid and itsligomers [34]. Our results found that mass loss rate increased

n the presence of enzymes, indicating that degradation study inBS may not accurately reflect physiological condition. Besides,he ratio of Gt in scaffolds also affected the biodegradation ratedata not shown), impling that scaffold degradation rate could bedjusted by altering its constituents. However, more studies areeeded to further balance the degradation of scaffold and tissueegeneration.

As one of the most crucial criteria, hemocompatibility is anssential requirement for vascular grafts. HP assay, a recommended

ethod by ISO, has been used to determine the extent of red

lood cells broken by scaffolds contact with peripheral blood [16].n present study, the HP values of all scaffolds were much lower

recalcification time, (C) Kinetic clotting time, and SEM image of attached plateletsnt significant difference at p < 0.05. The arrows represent platelets.

than the recommended value of 5% based on ISO standard, sug-gesting that PCL, pNSR32/PCL and pNSR32/PCL/Gt scaffolds hadno destruction to erythrocyte. Plasma coagulation property of anelectrospun scaffold implies its potential to activate the blood com-ponents and elicit thrombosis [35]. The PRT was tested to comparescaffolds-induced delay in clotting of PPP following activation ofprothrombin (Factor II) in the presence of Ca2+ [36]. Our datashowed that pNSR32/PCL and pNSR32/PCL/Gt scaffolds had slightlylonger PRT value than PCL scaffolds (p > 0.05), indicating that theaddition of pNSR32 and Gt had limited improvement on PRT ofthe PCL scaffolds. Kinetic clotting time, another simple and reliablemethod, is also frequently used to measure the activation extentof coagulation factors and the clotting time influenced by the scaf-folds [37]. In present study, the kinetic clotting time was differentamong these scaffolds (Fig. 5C). Accordingly, the blood incubatedwith pNSR32/PCL/Gt scaffolds had the highest absorbance and itscurve of coagulation time sloped downward more slowly thanother groups (Fig. 5C), indicating that the pNSR32/PCL/Gt scaf-fold had better anticoagulant property. This was because gelatinincorporation can potentially improve anticoagulant properties[38]. Besides, platelet attachment and activation are indicators

of platelet activation and a predominant mechanism, by whichbiomaterial thrombogenicity is elicited [39]. Zhang et al. foundthat improvement of hydrophilicity in biomaterials benefited plate

26 P. Xiang et al. / Colloids and Surfaces B: Biointerfaces 163 (2018) 19–28

Fig. 6. The pro-inflammatory factors expression of macrophages at the mRNA (A, B) and protein levels (C, D) were detected using real-time PCR and enzyme-linkedimmunosorbent assay. The different letters represent significant difference at p < 0.05.

F en spw . The s

aptwa

ckesmvm

ig. 7. Micrograph of H&E stained sections from blank control (A, F), medical collageeks (A-E) or 4 weeks (F-J) post-implantation. White asterisks represent scaffolds

dhesion [40]. Interestingly, in our study, platelets were morereferably adhered to hydrophobic PCL and pNSR32/PCL scaffoldshan that to the hydrophilic pNSR32/PCL/Gt scaffolds (Fig. 5D–F),hich may be due to the anionicity of the pNSR32/PCL/Gt scaffold,

parameter limiting platelet adhesion and aggregation [41]Implanted biomaterials often fail since they often induce con-

omitant fibrotic encapsulation and inflammatory reactions alsonown as foreign body responses [42]. The inflammatory response,specially by the macrophages, plays a pivot role in inducing steno-

is and intimal hyperplasia [43]. It has been widely accepted thatacrophages are a crucial source of inflammatory signals [44]. Acti-

ated macrophages at the site of vascular injury generate a serial ofediators such as interleukins (e.g., IL-6) and tumor necrosis factor

onge (B, G), PCL (C, H), pNSR32/PCL (D, I), pNSR32/PCL/Gt scaffold (E, J) implants 2cale bar is 200 mm.

(TNF-�), which stimulate the propagation and migration of vas-cular smooth muscle cells (VSMCs) [45]. Subsequently, excessiveextracellular matrix components produced by VSMCs may even-tually lead to intimal hyperplasia [46], whereas upregulation ofIL-6 and TNF-� are also associated with the occurrence of coronaryartery diseases [47]. Additionally, electrospun scaffolds topography(e.g., fiber diameters) influences the macrophage-activated inflam-matory responses [48]. A recent study found that electrospun PCLscaffolds with thin fibers induced higher mRNA level of Il-6 and Tnf-

in macrophage [20]. Similarly, in our study, the qRT-PCR analysisconfirmed the up-regulation of both Il-6 and Tnf- in macrophageson the PCL scaffold (Fig. 6AB). However, the pNSR32/PCL/Gt scaf-fold did not affect Il-6 or Tnf- expressions (Fig. 6AB). This was

P. Xiang et al. / Colloids and Surfaces B: Biointerfaces 163 (2018) 19–28 27

F one ml ositiv

btwftpii

mptmspTtiatcpiaiaiccddhdcip

btsteigfmn

ig. 8. Results of DNA damage in SDRAECs (A) and mocronuclei formation rate in betters represent significant difference at p < 0.05. N.C. and P.C. mean negative and p

ecause the pNSR32/PCL/Gt scaffold had a larger fiber diameterhan that of PCL scaffold [49]. Interestingly, after 24 h incubation,e observed a similar increase in TNF-� secretion on the PCL scaf-

old, whereas protein levels of IL-6 and TNF-� did not increase onhe pNSR32/PCL and pNSR32/PCL/Gt scaffolds, suggesting that theNSR32/PCL/Gt scaffold did not activate macrophages to induce

nflammatory response in vitro, making it a good candidate forn vivo use.

An innate immune response is often found at the site of bio-aterial implantation [50], therefore, it is critical to assess the

otential inflammatory reactions provoked by implanted bioma-erials in vivo. For in vivo studies of the host tissue response,

edical collagen sponge, PCL, pNSR32/PCL and pNSR32/PCL/Gtcaffolds were implanted intramuscularly into SD rats, and tissuearaffin sections were evaluated 2 or 4 weeks after implantation.hese times were selected since a previous study implied thathe inflammatory response attenuated as implantation durationncreased [50]. Histological results demonstrated that the tissuesround the PCL and pNSR32/PCL scaffolds exhibited a notable hostissue response, with dense recruitment of many inflammatoryells (Fig. 7CD), suggesting that the implantation of the PCL andNSR32/PCL scaffolds triggered severe inflammatory reactions dur-

ng the first 2 weeks, which may be attributed to their topographynd strong hydrophobicity [20,49]. Furthermore, host cells deepnto the scaffolds were only observed in the pNSR32/PCL (Fig. 7I)nd pNSR32/PCL/Gt (Fig. 7J) scaffolds 4 weeks after implantation,ndicating that the addition of pNSR32 and/or Gt facilitated hostell infiltration and decreased the accumulation of inflammatoryells in the tissues surrounding these scaffolds. This was probablyue to the increasing pore diameter of scaffolds after componentsegradation and incorporation of cell-binding motifs (i.e., RGD) andydrophilic groups [51]. Collectively, the pNSR32/PCL/Gt scaffoldid not trigger obvious inflammatory responses but facilitated hostell ingrowth compared to PCL and pNSR32/PCL scaffolds, indicat-ng that the addition of pNSR32 and Gt to PCL may be viable toroduce tissue-engineering scaffolds.

DNA lesions were examined using the comet assay, which haseen utilized to test the genotoxicity of cells exposed to nanofiberso visualize breaks in DNA [52]. In our studies, to prevent DNAelf-repair, DNA was immediately extracted and examined afterreatments. Our results demonstrated that cells incubated withxtract did not show obvious DNA damage, similarly to cells grownn DMEM, suggesting that the pNSR32/PCL/Gt scaffold did not cause

enotoxicity to SDRAECs. In addition, bone marrow micronucleiormation, another sensitive biomarker of genotoxic events and

anifestations of chromosomal instability that is often seen inanoparticles induced toxicity [53]. No notable difference in the

arrow PCE cells of KM mice (B) after exposure to the scaffold extract. The differente control.

frequencies of micronuclei was observed between the culturestreated with extract and negative control, suggesting little geno-toxic effect of pNSR32/PCL/Gt scaffold.

5. Conclusion

Our study showed that the electrospun pNSR32/PCL/Gt scaf-fold exhibited larger average fiber diameter, wider fiber diameterdistribution, and faster degradation rate than that of pNSR32/PCLand PCL scaffolds. All scaffolds showed excellent hemcompatibil-ity. Little notable pro-inflammatory response, DNA damage andmicronuclei formation were triggered by the pNSR32/PCL/Gt scaf-fold. Histological results revealed that implantation of the PCL andpNSR32/PCL scaffolds triggered obvious host tissue responses in SDrats, while pNSR32/PCL/Gt scaffold did not. More interestingly, thepNSR32/PCL/Gt scaffold facilitated host cells ingrowth. Based on thefact that the pNSR32/PCL/Gt scaffold possessing excellent in vitroand in vivo biocompatibility, it is a good candidate as a small calibervascular tissue engineering scaffold.

Acknowledgements

This work was sponsored partly by the Research Project of FujianNormal University (No. 5c1005), Key Program of the Fujian Provin-cial Department of Science and Technology (No. 2010Y0020) andthe Program B for Outstanding Ph.D. Candidates of Nanjing Univer-sity (No. 201702B060). P. Xiang is grateful for the support by thescholarship from China Scholarship Council (No. 201606190135).

References

[1] S.C. Smith, A. Collins, R. Ferrari, D.R. Holmes, S. Logstrup, D.V. McGhie, J.Ralston, R.L. Sacco, H. Stam, K. Taubert, Our time: a call to save preventabledeath from cardiovascular disease (heart disease and stroke), Eur. Heart J.(2012) ehs313.

[2] P. Xiang, R.Y. Liu, H.J. Sun, Y.W. Yang, X.Y. Cui, L.Q. Ma, Effects of novelbrominated flame retardant TBPH and its metabolite TBMEHP on humanvascular endothelial cells: implication for human health risks, Environ. Res.156 (2017) 834–842.

[3] M. Browning, D. Dempsey, V. Guiza, S. Becerra, J. Rivera, B. Russell, M. Höök, F.Clubb, M. Miller, T. Fossum, Multilayer vascular grafts based oncollagen-mimetic proteins, Acta Biomater. 8 (2012) 1010–1021.

[4] K.C. Yen, C.Y. Chen, J.Y. Huang, W.T. Kuo, F.H. Lin, Fabrication of keratin/fibroinmembranes by electrospinning for vascular tissue engineering, J. Mater.Chem. B 4 (2016) 237–244.

[5] M. Olausson, P.B. Patil, V.K. Kuna, P. Chougule, N. Hernandez, K. Methe, C.Kullberg-Lindh, H. Borg, H. Ejnell, S. Sumitran-Holgersson, Transplantation of

an allogeneic vein bioengineered with autologous stem cells: aproof-of-concept study, Lancet 380 (2012) 230–237.

[6] Y. Naito, T. Shinoka, D. Duncan, N. Hibino, D. Solomon, M. Cleary, A. Rathore,C. Fein, S. Church, C. Breuer, Vascular tissue engineering: towards the nextgeneration vascular grafts, Adv. Drug Deliv. Rev. 63 (2011) 312–323.

2 ces B:

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[C.C. Smith, R. Stetina, The comet assay: topical issues, Mutagenesis 23 (2008)143–151.

[53] B. Trouiller, R. Reliene, A. Westbrook, P. Solaimani, R.H. Schiestl, Titaniumdioxide nanoparticles induce DNA damage and genetic instability in vivo in

8 P. Xiang et al. / Colloids and Surfa

[7] L. Xue, H.P. Greisler, Biomaterials in the development and future of vasculargrafts, J. Vasc. Surg. 37 (2003) 472–480.

[8] B.I. Robinson, J.P. Fletcher, Australian and New Zealand Femoropopliteal GraftTrial Participants, Fluoropolymer coated Dacron or polytetrafluoroethylenefor femoropopliteal bypass grafting: a multicentre trial, SANZ J. Surg. 73(2003) 95–99.

[9] M.A. Cleary, E. Geiger, C. Grady, C. Best, Y. Naito, C. Breuer, Vascular tissueengineering: the next generation, Trends Mol. Med. 18 (2012) 394–404.

10] D. Motlagh, J. Yang, K.Y. Lui, A.R. Webb, G.A. Ameer, Hemocompatibilityevaluation of poly(glycerol-sebacate) in vitro for vascular tissue engineering,Biomaterials 27 (2006) 4315–4324.

11] S. de Valence, J.C. Tille, D. Mugnai, W. Mrowczynski, R. Gurny, M. Moller, B.H.Walpoth, Long term performance of polycaprolactone vascular grafts in a ratabdominal aorta replacement model, Biomaterials 33 (2012) 38–47.

12] J.M. Goddard, J.H. Hotchkiss, Polymer surface modification for the attachmentof bioactive compounds, Prog. Polym. Sci. 32 (2007) 698–725.

13] P. Xiang, M. Li, C.Y. Zhang, D.L. Chen, Z.H. Zhou, Cytocompatibility ofelectrospun nano fi ber tubular scaffolds for small diameter tissueengineering blood vessels, Int. J. Biol. Macromol. 49 (2011) 281–288.

14] Z. Ma, W. He, T. Yong, S. Ramakrishna, Grafting of gelatin on electrospunpoly(caprolactone) nanofibers to improve endothelial cell spreading andproliferation and to control cell orientation, Tissue Eng. 11 (2005) 1149–1158.

15] P. Xiang, K.C. Wu, Y. Zhu, L. Xiang, C. Li, D.L. Chen, F. Chen, G. Xu, A. Wang, M.Li, Z.B. Jin, A novel Bruch’s membrane-mimetic electrospun substrate scaffoldfor human retinal pigment epithelium cells, Biomaterials 35 (2014)9777–9788.

16] N. Alexandre, J. Ribeiro, A. Gartner, T. Pereira, I. Amorim, J. Fragoso, A. Lopes, J.Fernandes, E. Costa, A. Santos-Silva, M. Rodrigues, J.D. Santos, A.C. Mauricio,A.L. Luis, Biocompatibility and hemocompatibility of polyvinyl alcoholhydrogel used for vascular grafting–in vitro and in vivo studies, J. Biomed.Mater. Res. A 102 (2014) 4262–4275.

17] C. Fornaguera, G. Caldero, M. Mitjans, M.P. Vinardell, C. Solans, C. Vauthier,Interactions of PLGA nanoparticles with blood components: proteinadsorption, coagulation, activation of the complement system and hemolysisstudies, Nanoscale 7 (2015) 6045–6058.

18] A. Nel, T. Xia, L. Madler, N. Li, Toxic potential of materials at the nanolevel,Science 311 (2006) 622–627.

19] E. Dybing, J. O’Brien, A.G. Renwick, T. Sanner, Risk assessment of dietaryexposures to compounds that are genotoxic and carcinogenic–an overview,Toxicol. Lett. 180 (2008) 110–117.

20] Z. Wang, Y. Cui, J. Wang, X. Yang, Y. Wu, K. Wang, X. Gao, D. Li, Y. Li, X.L.Zheng, Y. Zhu, D. Kong, Q. Zhao, The effect of thick fibers and large pores ofelectrospun poly(epsilon-caprolactone) vascular grafts on macrophagepolarization and arterial regeneration, Biomaterials 35 (2014) 5700–5710.

21] C.R. Ruan, J.X. Huang, M.H. Wei, M. Li, Construction, fermentation andpurification of high polymer spider dragline silk protein containing RGDpeptide, Sheng Wu Gong Cheng Xue Bao 23 (2007) 858–861.

22] D.I. Braghirolli, F. Zamboni, P.C. Chagastelles, D.J. Moura, J. Saffi, J.A. Henriques,D.A. Pilger, P. Pranke, Bio-electrospraying of human mesenchymal stem cells:an alternative for tissue engineering, Biomicrofluidics 7 (2013) 44130.

23] D. Chen, Q. Tang, X. Li, X. Zhou, J. Zang, W.Q. Xue, J.Y. Xiang, C.Q. Guo,Biocompatibility of magnetic Fe(3)O(4) nanoparticles and their cytotoxiceffect on MCF-7 cells, Int. J. Nanomed. 7 (2012) 4973–4982.

24] P. Xiang, Study on RGD-spider Silk Protein/polycaprolacton/gelatin CompositeNanofibrous Small Diameter Vascular Scaffolds, Fujian Nornal University,2011.

25] S.W. Cai, B.R. Singh, Identification of beta-turn and random coil amide IIIinfrared bands for secondary structure estimation of proteins, Biophys. Chem.80 (1999) 7–20.

26] Y.M. Li, Y.F. Wang, J.J. Ye, J. Yuan, Y.H. Xiao, Fabrication ofpoly(epsilon-caprolactone)/keratin nanofibrous mats as a potential scaffoldfor vascular tissue engineering, Mat. Sci. Eng. C-Mater. 68 (2016) 177–183.

27] C.S. Wong, X. Liu, Z. Xu, T. Lin, X. Wang, Elastin and collagen enhanceselectrospun aligned polyurethane as scaffolds for vascular graft, J. Mater. Sci.Mater. Med. 24 (2013) 1865–1874.

28] L. Lu, S.J. Peter, M.D. Lyman, H.-L. Lai, S.M. Leite, J.A. Tamada, S. Uyama, J.P.Vacanti, L. Robert, A.G. Mikos, In vitro and in vivo degradation of porouspoly(dl-lactic-co-glycolic acid) foams, Biomaterials 21 (2000) 1837–1845.

29] C.X. Lam, D.W. Hutmacher, J.T. Schantz, M.A. Woodruff, S.H. Teoh, Evaluation

of polycaprolactone scaffold degradation for 6 months in vitro and in vivo, J.Biomed. Mater. Res. A 90 (2009) 906–919.

30] C.L. Salgado, E.M. Sanchez, C.A. Zavaglia, A.B. Almeida, P.L. Granja, Injectablebiodegradable polycaprolactone–sebacic acid gels for bone tissueengineering, Tissue Eng. Part A 18 (2011) 137–146.

Biointerfaces 163 (2018) 19–28

31] B.W. Tillman, S.K. Yazdani, S.J. Lee, R.L. Geary, A. Atala, J.J. Yoo, The in vivostability of electrospun polycaprolactone-collagen scaffolds in vascularreconstruction, Biomaterials 30 (2009) 583–588.

32] X.F.L. Christopher, M.S. Monica, T. Swee-Hin, W.H. Dietmar, Dynamics ofin vitro polymer degradation of polycaprolactone-based scaffolds: acceleratedversus simulated physiological conditions, Biomed. Mater. 3 (2008) 034108.

33] L. Wu, J. Ding, In vitro degradation of three-dimensional porouspoly(d,l-lactide-co-glycolide) scaffolds for tissue engineering, Biomaterials 25(2004) 5821–5830.

34] L.S. Nair, C.T. Laurencin, Biodegradable polymers as biomaterials, Prog. Polym.Sci. 32 (2007) 762–798.

35] M.B. Gorbet, M.V. Sefton, Biomaterial-associated thrombosis: roles ofcoagulation factors, complement, platelets and leukocytes, Biomaterials 25(2004) 5681–5703.

36] L.P. Amarnath, A. Srinivas, A. Ramamurthi, In vitro hemocompatibility testingof UV-modified hyaluronan hydrogels, Biomaterials 27 (2006) 1416–1424.

37] Z.X. Meng, W. Zheng, L. Li, Y.F. Zheng, Fabrication and characterization ofthree-dimensional nanofiber membrance of PCL–MWCNTs byelectrospinning, Mater. Sci. Eng. C 30 (2010) 1014–1021.

38] J. Kucinska-Lipka, I. Gubanska, H. Janik, Gelatin-modified polyurethanes forsoft tissue scaffold, Sci. World J. 2013 (2013).

39] L.K. Jennings, Mechanisms of platelet activation: need for new strategies toprotect against platelet-mediated atherothrombosis, Thromb. Haemost. 102(2009) 248–257.

40] E. Zhang, F. Shen, Blood compatibility of a ferulic acid (FA)-eluting PHBHHxsystem for biodegradable magnesium stent application, Mater. Sci. Eng. CMater. Biol. Appl. 52 (2015) 37–45.

41] M.M. Amiji, Surface modification of chitosan membranes bycomplexation-interpenetration of anionic polysaccharides for improved bloodcompatibility in hemodialysis, J. Biomater. Sci. Polym. Ed. 8 (1997) 281–298.

42] E. Dondossola, B.M. Holzapfel, S. Alexander, S. Filippini, D.W. Hutmacher, P.Friedl, Examination of the foreign body response to biomaterials by nonlinearintravital microscopy, Nat. Biomed. Eng. 1 (2016) 0007.

43] N. Hibino, T. Yi, D.R. Duncan, A. Rathore, E. Dean, Y. Naito, A. Dardik, T.Kyriakides, J. Madri, J.S. Pober, T. Shinoka, C.K. Breuer, A critical role formacrophages in neovessel formation and the development of stenosis intissue-engineered vascular grafts, FASEB J. 25 (2011) 4253–4263.

44] K. Garg, N.A. Pullen, C.A. Oskeritzian, J.J. Ryan, G.L. Bowlin, Macrophagefunctional polarization (M1/M2) in response to varying fiber and poredimensions of electrospun scaffolds, Biomaterials 34 (2013) 4439–4451.

45] P. Libby, D. Schwartz, E. Brogi, H. Tanaka, S.K. Clinton, A cascade model forrestenosis. A special case of atherosclerosis progression, Circulation 86 (1992)III47–52.

46] A.C. Newby, A.B. Zaltsman, Molecular mechanisms in intimal hyperplasia, J.Pathol. 190 (2000) 300–309.

47] J.S. Yudkin, M. Kumari, S.E. Humphries, V. Mohamed-Ali, Inflammation,obesity, stress and coronary heart disease: is interleukin-6 the link?Atherosclerosis 148 (2000) 209–214.

48] Z. Wang, Y. Cui, J. Wang, X. Yang, Y. Wu, K. Wang, X. Gao, D. Li, Y. Li, X.-L.Zheng, Y. Zhu, D. Kong, Q. Zhao, The effect of thick fibers and large pores ofelectrospun poly(�-caprolactone) vascular grafts on macrophage polarizationand arterial regeneration, Biomaterials 35 (2014) 5700–5710.

49] E. Saino, M.L. Focarete, C. Gualandi, E. Emanuele, A.I. Cornaglia, M. Imbriani, L.Visai, Effect of electrospun fiber diameter and alignment on macrophageactivation and secretion of proinflammatory cytokines and chemokines,Biomacromolecules 12 (2011) 1900–1911.

50] M.S. Kim, H.H. Ahn, Y.N. Shin, M.H. Cho, G. Khang, H.B. Lee, An in vivo study ofthe host tissue response to subcutaneous implantation of PLGA- and/orporcine small intestinal submucosa-based scaffolds, Biomaterials 28 (2007)5137–5143.

51] T.A. Telemeco, C. Ayres, G.L. Bowlin, G.E. Wnek, E.D. Boland, N. Cohen, C.M.Baumgarten, J. Mathews, D.G. Simpson, Regulation of cellular infiltration intotissue engineering scaffolds composed of submicron diameter fibrilsproduced by electrospinning, Acta Biomater. 1 (2005) 377–385.

52] A.R. Collins, A.A. Oscoz, G. Brunborg, I. Gaivao, L. Giovannelli, M. Kruszewski,

mice, Cancer Res. 69 (2009) 8784–8789.