engineering interaction between bone marrow derived endothelial cells and electrospun surfaces for...

Engineering Interaction between Bone Marrow Derived EndothelialCells and Electrospun Surfaces for Artificial Vascular GraftApplicationsFurqan Ahmed,† Naba K. Dutta,† Andrew Zannettino,‡,§ Kate Vandyke,‡,§ and Namita Roy Choudhury*,†

†Ian Wark Research Institute, University of South Australia, Mawson Lakes Campus, South Australia, Australia‡Myeloma Research Laboratory, School of Medical Science, Faculty of Health Science, University of Adelaide, Adelaide, SouthAustralia, Australia§Centre for Cancer Biology, SA Pathology, Adelaide, South Australia, Australia

ABSTRACT: The aim of this investigation was to understandand engineer the interactions between endothelial cells and theelectrospun (ES) polyvinylidene fluoride-co-hexafluoropropy-lene (PVDF-HFP) nanofiber surfaces and evaluate theirpotential for endothelialization. Elastomeric PVDF-HFPsamples were electrospun to evaluate their potential use assmall diameter artificial vascular graft scaffold (SDAVG) andcompared with solvent cast (SC) PVDF-HFP films. Weexamined the consequences of fibrinogen adsorption onto theES and SC samples for endothelialisation. Bone marrowderived endothelial cells (BMEC) of human origin wereincubated with the test and control samples and theirattachment, proliferation, and viability were examined. Thenature of interaction of fibrinogen with SC and ES samples was investigated in detail using ELISA, XPS, and FTIR techniques.The pristine SC and ES PVDF-HFP samples displayed hydrophobic and ultrahydrophobic behavior and accordingly, exhibitedminimal BMEC growth. Fibrinogen adsorbed SC samples did not significantly enhance endothelial cell binding or proliferation.In contrast, the fibrinogen adsorbed electrospun surfaces showed a clear ability to modulate endothelial cell behavior. Thissystem also represents an ideal model system that enables us to understand the natural interaction between cells and theirextracellular environment. The research reported shows potential of ES surfaces for artificial vascular graft applications.

■ INTRODUCTION

Formation of thrombus on the cardiovascular implantsrepresents a significant limitation to their long-term success.1,2

In natural blood vessels, anticoagulants like prostacycline, nitricoxide and surface bound heparan sulfate to endothelial cellssurface can prevent the formation of thrombus andatherosclerosis.3−5 A vascular graft without an adherent layerof endothelial cell lining can be rejected due to the nonspecificadsorption of biomolecules from the circulating blood, leadingto thrombosis/occlusion. To limit thrombus formation, anuninterrupted endothelial cell lining on the implant surface isrequired in order to make the graft compatible with the bloodand tissue environment.6 To date, the success of small diameterartificial vascular grafts (SDAVG) is in its formative stages dueto the low patency rate, poor reliability and thrombosis issues.Careful selection of synthetic materials based on propertieswhich improve blood and cytocompatibility has the potential toovercome many of these issues. As such, significant work hasbeen focused on the material structure and composition,surface modifications, topographic changes, and surface andbiochemical coatings to enhance endothelialization. In recent

years, one of the most important breakthroughs has been thedevelopment of fibrous scaffolds using electrospinning.7,8

The electrospinning technique has the advantage of beingable to generate fibrous scaffolds of micro to nanoscaletopography and high porosity that resemble the naturalextracellular matrix (ECM) found in blood vessels.9,10 Thistechnique provides the capacity to develop biomimeticmaterials that promote endothelial cell binding and prolifer-ation and allow for transfer of the nutritional and oxygen tomeet the demands of the engrafted cells.Besides the surface chemistry and topography of the original

SDAVG, the post-treatment changes to the SDAVG,particularly the adsorption of serum proteins plays animportant role in the success and rejection of the graft.Biomacromolecules such as proteins, polysaccharides, andproteoglycans can act as biological cues for adherent cells.Cell adhesion to ECM proteins, adsorbed from the environ-ment or secreted by the cultured cells, is primarily mediated by

Received: December 13, 2013Revised: February 22, 2014Published: February 24, 2014

Article

pubs.acs.org/Biomac

© 2014 American Chemical Society 1276 dx.doi.org/10.1021/bm401825c | Biomacromolecules 2014, 15, 1276−1287

pubs.acs.org/Biomac

integrins. ECM proteins, such as fibronectin, vitronectin,fibrinogen, and collagen, have the ability to support celladhesion. Therefore, if biomaterial surfaces are modified withthese bioactive macromolecules, the biocompatibility of thesurfaces can be significantly improved.Fibrinogen is one of the most important serum proteins and

displays both clotting and cell proliferation capabilities, with theconcentration and conformation dictating the mobilization andrecruitment of different cell; and, in turn, their adhesion andproliferation at the site of the implantation.11−13 The fibrinogenmolecule has strong binding sites for platelets and endothelialcells, which defines its role in clotting and endothelial cellproliferation. Various materials have been used to study the roleof fibrinogen on inhibition and proliferation of cells onsynthetic materials like polyethylene terephthalate (PET,Dacron), expanded polytetrafluoroethylene (e-PTFE), polyur-ethane (PU), polycaprolactone (PCL), and copolylactic acid/glycolic acid (PLGA),14−20 some of which are the subject ofSDAVG development. However, most of these materials haveshown very limited success either due to early thrombosis,antigen response, or loss of structural integrity due to poormechanical (e.g., elasticity) performance. To overcome thesedeficiencies, we have selected poly(vinylidene-fluoride)-co-hexafluoropropylene (PVDF-HFP) as it possesses robust,elastic, nonreactive, nontoxic, nonthrombogenic, resilient, andantibacterial properties.21 Moreover, PVDF-HFP can be easilyprocessed using ES techniques, allowing for the generation ofSDAVG with different diameter, morphology, porosity, andsurface characteristics. Because of its biostability and bio-compatibility, it also has the potential to minimize the issues ofthrombosis and aneurysm formation. We have recentlyoptimized the electrospinning process of PVDF-HFP andtheir physical/physicochemical characteristics.22 In this work,we used electrospinning to produce a morphologically relevantscaffold composed of PVDF-HFP fibres and examined in detailits capacity to support the attachment, proliferation, andviability of bone marrow-derived endothelial cells. Wehypothesized that the physical and chemical characteristics ofthe biomimetic scaffold plays an important role in promotingadhesion of cells. To assess this hypothesis, we evaluated theability of solvent cast (SC) and electrospun (ES) PVDF-HFPsamples to support adhesion and proliferation of bone marrowderived endothelial cells (BMECs). Furthermore, we alsoassessed the effect of preadsorbing the surfaces with fibrinogento BMEC adhesion and proliferation.

■ MATERIALS AND METHODSReagents. PVDF-HFP, with an inherent viscosity of 2300−2700

Pa and Mw ∼ 400000, was procured from Sigma Aldrich, Australia.The polymer was dissolved in 70/30 ratio of N,N-dimethylacetamide(DMAc) and acetone (from Sigma Aldrich, Australia) at 10% (w/v)concentration and left for 24 h mixing with a magnetic stirrer at roomtemperature.Fibrinogen, from bovine plasma type 1-S, 65−85% protein, was

purchased from Sigma Aldrich. To make 25 mg mL−1 aliquot, 450 mg(0.45 g) of the protein was diluted in PBS (the physiologicalconcentration, i.e., 2 mg/mL of fibrinogen, was used for endothelialcell study). A monoclonal antifibrinogen antibody and a goat anti-mouse IgG, HRP conjugate were purchased from Sigma Aldrich andMillipore Australia, respectively.Preparation of PVDF-HFP SC Samples. Uniform flat films of the

PVDF-HFP were prepared by casting of PVDF-HFP solution in a glasspetri dish. The PVDF-HFP solution in petri dish was left in controlled

environment for drying for 2 days at 65 °C. After solvent evaporation,the PVDF-HFP flat films were cut to make 1 cm discs.

Nanofiber Fabrication. The fabrication of PVDF-HFP electro-spun surfaces was performed using 10% solution, 13 kV voltage, 0.15mL/h flow rate and 14 cm tip capillary distance as described in ourprevious study.22

Quantification of Fibrinogen Adsorption by Enzyme LinkedImmunosorbent Assay (ELISA). The quantity of purified fibrinogen(Fg) on PVDF-HFP (SC and ES) samples was determined using amodified indirect ELISA method23 in 2-fold serial dilution offibrinogen. A standard curve of fibrinogen adsorption on 96-wellplates was established in two ways: first, surfaces were preincubatedwith Tris-buffered saline (TBS), followed by fibrinogen andantibodies; and second, surfaces were preincubated with 1% bovineserum albumin (BSA) blocking buffer, followed by incubation withfibrinogen and antibodies, as described below. The second standardcurve step was established in order to account for the potential forfibrinogen to adsorb to the polystyrene wells at the time of ELISAassay for SC and ES samples. In brief, polystyrene wells were blockedwith the addition of blocking buffer for 1 h, and then the PVDF-HFP(SC and ES) samples (after washing), which were cut to the size of thewell were placed into the blocked well and subsequently exposed topurified bovine fibrinogen for 2 h at room temperature. These surfaceswere rinsed with TBS to remove unattached proteins. The fibrinogenadsorbed PVDF-HFP samples were incubated with 1% BSA for 1 h toblock surfaces; and subsequently exposed to an antifibrinogenantibody (0.25 μg mL−1) in TBS for 2 h. The samples were washedfour times with TBS and then 100 μL of biotinylated goat anti-mouseIgG (0.25 μg mL−1 diluted in TBS) was added to all sample wells andincubated for 1 h. The samples were washed with TBS four times. Thesamples were exposed to 100 μL (at a 1:4000 dilution) of avidin-HRPfor 30 min. The samples were again washed with TBS and the surfacesincubated for 5 min with 200 μL of ortho-phenylene diamine (OPD,Sigma; 1 tablet OPD reagent with 1 tablet buffer in 20 mL of waterwrapped in foil for 20 min on a rocking platform). After thisincubation, the reaction was stopped by adding 50 μL of 2.5 Nsulphuric acid. Absorbance at 490 nm was measured on a 200 μLaliquot using an ELISA auto reader. The varying fibrinogenconcentration assay was performed using TBS as control and, foreach day, similar time intervals and concentrations were used with thesame conditions.

Characterization of ES and SC Samples. Water Contact Angle(WCA) Measurement. The static contact angle of water on pristineand fibrinogen preadsorbed samples (SC and ES) was measured at 25°C using a proprietary contact angle goniometer by sessile dropmethod by placing a 10 μL drop of distilled water on surfaces of ESand SC samples of PVDF-HFP. The droplet shape was imaged with avideo camera and contact angle calculated using proprietary software(Ian Wark Research Institute, Adelaide, Australia).24

Characterization by SEM. A Philips XL30 field emission gunScanning Electron Microscopy (FEGSEM) with Oxford CT1500HFCryo stage was used to characterize the morphology of cells on SC,polystyrene, and ES samples. To minimize the charging effect,platinum was deposited after cell fixation on each sample usingsputtering and examined at an accelerating voltage of 10 KV. Theanalysis of the diameter and porosity of the ES samples was performedfrom the SEM image.

Characterization by XPS. XPS analysis was performed with anAXIS HSi Spectrometer (Kratos Analytical Ltd., Manchester, U.K.),equipped with a monochromated Al Kα source at a power of 144 W(12 mA, 12 kV). The samples (∼6.5 mm) were mounted on amultisample holder stage. The charging of the samples duringirradiation was compensated for by an electron flood gun incombination with a magnetic immersion lens. A reference bindingenergy of 285.0 eV for the aliphatic hydrocarbon C1s component wasused to correct for any remaining offsets due to charge neutralizationof specimens under irradiation.25 The pressure in the main vacuumchamber during analysis was typically 5 × 10−6 Pa. The spectra wererecorded with the nominal photoelectron detection normal to thesample surface. The sampling depth was up to 10 nm depending on

Biomacromolecules Article

dx.doi.org/10.1021/bm401825c | Biomacromolecules 2014, 15, 1276−12871277

the kinetic energy of the measured photoelectrons. The elementalcomposition of the samples was obtained from survey spectra (320 eVpass energy) using sensitivity factors supplied by the manufacturer.High resolution spectra of individual peaks were recorded at 40 eVpass energy.Characterization by PA-FTIR. The fibrinogen adsorbed surfaces

were also investigated using Photoacoustic Fourier TransformsInfrared spectroscopy (PA-FTIR). PA-FTIR was performed using aNicolet Magna Spectrometer (Model 750) equipped with a MTEC(model 300) photo acoustic cell. The sample was placed in a circularstainless steel cup, 3 mm deep and 10 mm in diameter, and sealed in adevice with a potassium bromide salt window and a heliumatmosphere to promote good heat transfer. A carbon black film wasused as reference, as it absorbs all wavelengths of the infrared radiationand produces a spectrum that mirrors both the energy characteristicsof the detector and optical performance of the instrument. The

resolution of 4 cm−1, 256 scans, and a mirror velocity of 0.158 cm s−1

was used.Three SC and ES samples of PVDF-HFP were analyzed for each

characterization technique and the standard deviation of these sampleswas interpreted.

Cell Culture. Human bone marrow derived endothelial cells(BMECs)26,27 were a gift from Professor Babette Weksler (CornellUniversity Medical College, New York, New York). The cells wereplated in monolayer in T-75 tissue culture flasks and cultured toconfluence in M199 culture medium containing 10% fetal bovineserum, growth factor, and heparin. The medium was replaced every 2days and cultures were maintained in a tissue culture incubator at 37°C with 5% CO2.

Cell Seeding. For the cell culture studies, a preoptimized solutionconcentration (10% w/v) and electrospinning conditions (0.15 mLh−1 flow rate, 13 kV applied voltage and 14 cm tip capillary distance)for the production of ES samples were used. Control samples (∼1 cm)

Figure 1. (A) Standard curve of fibrinogen adsorption on 96-well plates and scheme of ELISA assay on TCP (B) percentage of relative fibrinogenbinding on SC and ES samples with scheme of ELISA assay (n = 3).



of TCP, SC, and optimized ES PVDF-HFP samples (after 2 hsterilization with ethanol) were placed at the bottom of a 24-well tissueculture plate and BMECs in M199 culture medium were seeded ontoeach surface at a density of 3 × 104, 6 × 104, and 12 × 104 cells cm−2.

After a seeding time of 5 h, each culture sample was analyzed underSEM and cell seeding efficiency was calculated with reference to TCP.

Assessment of Cell Adhesion. The adhesion of BMECs todifferent samples was evaluated using cell counting. The SC and ES

Figure 2. X-ray photoelectron spectra of PVDF-HFP (A) pristine solvent cast, (B) pristine electrospun, (C) preadsorbed fibrinogen SC, and (D)preadsorbed fibrinogen ES, (E) C1s core level spectra of pristine SC, (F) C1s core level spectra of preadsorbed fibrinogen SC, (G) C1s core levelspectra of pristine ES, and (H) C1s core level spectra of preadsorbed fibrinogen ES (n = 3).



samples were placed in a 24-well cell culture plate. All samples weresterilized in 70% ethanol for 2 h, followed by overnight drying. BMECswere detached with 0.05% trypsin-EDTA and their number wascounted using a hemocytometer. Cell concentrations were adjusted inthe culture medium to the corresponding plating densities.Cell Fixation. After 5 h of incubation, BMEC attached samples

were washed with PBS to remove nonadherent cells and then fixedwith 4% glutaraldehyde for 30 min at room temperature, rewashedwith PBS for 5 min, and postfixed with 2% osmium tetraoxide for 30min. The biomaterial/BMEC constructs were dehydrated with 70%ethanol for 10 min, with 90% ethanol for 10 min and 100% ethanol for10 min followed by rinsing with hexamethyldisilazane (HMZ) dilutedwith equal amount of 100% ethanol for 10 min, then undiluted HMZfor 10 min, and left to dry overnight. After critical point drying, thebiomaterial/BMEC constructs were sputter coated with platinum andobserved under the SEM at an accelerating voltage of 10 kV. Cellcount was determined by SEM image on 10 randomly selectedlocations using Image J software.Assessment of Cell Proliferation and Viability. For the

proliferation studies, samples were sterilized as described above.BMECs were seeded separately on TCP and PVDF-HFP (SC and ES)samples, at a plating density of 6 x104 cells cm−2. The BMECs wereallowed to proliferate for 1, 3, 7, and 9 days, after which proliferationand viability were determined using a WST-1 cell proliferation assaykit (Roche), as described below. Cell counting of SEM images was alsoperformed in order to confirm our findings with the WST-1 assay.Experiments were performed in triplicate. The SEM images were takenfrom 10 randomly chosen locations for each sample condition. Thecell densities were determined using Image J software.WST-1 Viability Assay. WST-1 reagent was diluted to 1:10 with

phenol red free DMEM/10% FCS (prewarmed at 37 °C). BMEC/biomaterial composites were transferred to adjacent wells to avoidcounting cells which had adhered to the base and wall of the wells. Atotal of 400 μL of diluted WST-1 reagent was added to each wellcontaining samples. The same amount of reagent was also added to acontrol well. The samples were incubated for 40 min at 37 °C/5%CO2 and triplicates of 100 μL from each well were transferred to a flatbottom 96-well plate for absorbance measurement on a ELISA platereader with a wavelength of 490 nm.Cell Culture Study on Fibrinogen Adsorbed Samples. After

proper sterilization with ethanol, the TCP, SC, and ES samples weresoaked in 2 mg mL−1 fibrinogen solution for 2 h and rinsed with MilliQ water three times. All the samples (n = 3) were placed in a 24-wellculture plate for adhesion and proliferation assays, as described above.Data Analysis. Statistical comparisons from different cell counts

were performed in excel and data were presented as mean ± standarddeviation (SD). The differences were considered statistically significantwith <5% SD of mean.

■ RESULTS AND DISCUSSION

Fibrinogen Adsorption Study by ELISA, XPS, and FTIR.Figure 1 shows the adsorption of fibrinogen (Fg) to control(Figure 1A) and PVDF-HFP (both SC and ES) surfaces(Figure 1B) as a function of Fg concentration after incubationfor 2 h. Indirect ELISA assay was used to quantify the degree ofFg binding. A standard curve of Fg adsorption was establishedas described in the Materials and Methods. A stock of 4000 μgmL−1 Fg was serially 2-fold diluted down to a concentration of0.24 μg mL−1 and added to both SC and ES samples. As seen inFigure 1B, both surfaces displayed a similar adsorption trend,with increasing Fg concentrations leading to increased Fgbinding until a “plateau” was reached. As it can be seen inFigure 1B, the percentage of maximum Fg adsorbed on SCsamples was ∼85.4% compared to that on ES samples, whichwas ∼74.6%. The EC50 (half maximal effective concentration ofa protein that can produce a response equal to 50% of themaximum dose response28) of each surface was measured to

determine the Fg response on SC and ES surfaces. The SCsurfaces had an EC50 of 0.687 μg mL−1, with a confidenceinterval (CI) of 95%; while the ES surface had an EC50 of 2.30μg mL−1 and 95% CI, suggesting that SC had a greater capacityto adsorb Fg than that on the ES surfaces.The quantitative effect of Fg adsorption on various samples

was evaluated using XPS. Figure 2A−D shows representativeXPS wide-scan survey spectra of all elements for two differentsamples of PVDF-HFP and has been used as the basis fordetermining the elemental composition of adsorbed protein.The observed peaks at 286 and 688 eV was attributable to C1sand F1s, respectively. The high-resolution scan of C1s core-level spectrum of pristine PVDF-HFP, for both ES (Figure 2G)and SC (Figure 2E) samples, can be curve fitted with fivecomponents, with binding energies at about 286.4, 289.6, 290.9,293.9, and 284.8, which may be attributed to CH2, CF, CF2,CF3, and CC species, respectively. After fibrinogen adsorption,both SC and ES samples display a distinctive peak of nitrogenat 400 eV, indicating protein adsorption. The atomicpercentage of nitrogen was calculated to be approximately11% on SC surfaces; compared to only 6% on ES surfaces,further highlighting the low adsorption of Fg on ES sample.The high-resolution C1s core level spectrum of SC and ESsamples after Fg adsorption (Figures 2F,H), were curve fittedwith six peak components by fitting Gaussian function to theexperimental curve at the binding energy level of 284.8, 286.2,288, 290.1, 293.4, and 285.6 eV and has been attributed to theCH2, COH, CHF, CF2, CF3, and CN species, respectively,which further indicates the adsorption of Fg. The results ofcurve fitting of carbon, peak position of each element and their% are reported in Table 1. The level of nitrogen is almost

double on the SC sample than that on ES sample (10.94 vs5.86%). A similar trend was observed into the CN peak as wellinto SC sample showing highest 8.6% vs ES 5.74%. Overall, theXPS results are in line with the ELISA data for physiologicalconcentration (2000 μg mL−1) of fibrinogen on SC and ESsamples, confirming the reliability of ELISA results. Figure 3shows a close relationship of ELISA and XPS results forphysiological concentration of fibrinogen on SC and ESsamples.

Table 1. Relative Atomic Percentages of Different Elementsfrom XPS before and after Fibrinogen Adsorption on SC andES Surfaces

Atomic %

survey spectraSC

pristineES

pristineSC preadsorbed

fibrinogenES preadsorbed

fibrinogen

F1s 62.88 61.43 11.18 31.26C1s 33.86 31.84 57.56 40.37O1s 0.71 0.38 20.31 22.49N1s 10.94 5.86

C1s Core LevelSpectraCH2 41.55 37.53 40.75 41.46CF2 40.38 38.80 6.66 11.33C−C 9.33 16.50CF3 4.69 3.98 0.66 0.96CF 4.03 3.17COH 27.90 27.39CHF 15.40 13.05CN 8.60 5.78



The PA-FTIR spectra of pristine solvent cast and electrospunsamples (A and B) and Fg adsorbed samples (C and D) areshown in Figure 4, and provides clear evidence of Fg adsorption

in the form of amide I at 1650 cm−1 and amide II band at 1550cm−1. In proteins, amide I and amide II bands are observed dueto CO stretching and combination of N−H and C−Nstretching, respectively. Similar observation has also been madeby Clarke et al. for fibrinogen adsorption on polyurethane andperfluorinated polymer.29 In the present case, the overallamount of fibrinogen adsorbed on SC samples is relativelyhigher than that on ES samples as observed by ELISA and XPS.Fibrinogen is a 340 kDa plasma glycoprotein (conc. in blood

is ∼2−4 mg/mL)30 and has an elongated three nodular domainstructure (called D−E−D domains), as shown schematically inFigure 5. Biochemically, Fg molecule is a disulfide-bondedhexameric protein consisting of two sets of three differentpolypeptide chains (namely, Aα, Bβ, and cγ, respectively),which are interconnected by disulfide bonds and forms a coiled-coil structure and further folded to form three-globular domains(∼47.5 nm long and 6.5 nm wide). More recently, TEM and X-ray crystallography results demonstrated that fibrinogen hasother αC domains formed by the folding of Aα chains. Ingeneral, D and E domains are hydrophobic while the αCdomain is hydrophilic.31 Being a nonspecific protein, there can

be four possibilities of fibrinogen interaction with polymersurfaces, namely, ionic or electrostatic, hydrogen bonding,charge transfer, and hydrophobic interaction.32 Because twodifferent samples of a hydrophobic polymer have been used inthis study, so, there are many possibilities of hydrophobicinteractions. The driving force for the deposition of the proteinchains of Fg molecule (Figure 5) on the SC surface can beattributed to hydrophobic−hydrophobic interaction; however,in the case of ES material, the surface is porous andsuperhydrophobic in nature due to fiber-induced roughness,and porosity leading to less spreading and wettability of the Fgmolecule on the fibres. With the passage of time, the voids aremore accessible due to release of air from the pockets and thefibrinogen molecule can subsequently spread (size of pores ofES sample, which is around 1−2 μm and the size of thefibrinogen molecule is 47 nm). Therefore, the response offibrinogen adsorption is lower and slower for ES than for SCsample. As shown in Figure 6, this is further confirmed bycontact angle results after Fg adsorption. Both the surfaces

Figure 3. Co-relation of physiological serum concentration offibrinogen adsorption by ELISA and XPS.

Figure 4. PA-FTIR spectra of pristine SC (A), ES (B), andpreadsorbed fibrinogen SC (C) and ES (D) surfaces (n = 3).

Figure 5. Schematic representation of the fibrinogen molecule.

Figure 6. Measured contact angle of SC and ES samples before andafter fibrinogen adsorption (n = 3, std = <5%).



show change in contact angle, that is, SC shows a change from95 to 77° and ES samples show a change from 147 to 113°,indicating the hydrophilic nature of the modified surface.The difference in wettability of SC and ES sample surfaces

may be a good justification for difference in fibrinogenadsorption at macroscopic level. A model fibrinogen studywas performed by Siegismund et al.33 who postulated that theinteractive energy of a material surface for fibrinogenadsorption is different according to surface wettability ofmaterial. Siegismund and co-workers33 observed high inter-active energy for hydrophobic surfaces than on hydrophilic,which shows strong and irreversible affinity of fibrinogen onhydrophobic than on hydrophilic surface that needs less energyto detach fibrinogen. The model discusses four different stagesof fibrinogen interaction with the interface such as fibrinogentransfer, adsorption, surface diffusion, and inner clusterdiffusion. It has been shown that fibrinogen conformation/clustering, denaturation, and adsorbed amount depends onhydrophobicity of the materials’ surface. In this study, the SCsample has shown higher surface energy and shows strongbinding of fibrinogen compared to that of the ES sample.Figure 7A,C,E show the adsorption of fibrinogen molecules to

the SC interface at different stages. The process of diffusionstarts in response to difference in concentration gradient offibrinogen at the interface and solution (Figure 7A). Thenonadsorbed Fg remained in solution when saturation levelswere achieved. There is a critical contact angle between the Fgmolecule and SC surface ranging from 0 to 90°, and within thisrange, an adsorption process will proceed (Figure 7C). Figure7E indicates the clustering of the fibrinogen molecule, which

occurs due to translative and rotatory movements within the Fgmolecule. However, on ES samples, the different nature of thesurface has led to a different adsorption trend (lower amount),as proposed in Figures 7B,D,F. The voids among the fibers inES samples may affect the critical angle of fibrinogeninteraction compared to solid (SC) sample. A zero criticalangle of the fibrinogen molecule on the ES sample is leastpossible; therefore, there is more possible diffusion offibrinogen molecules back toward the solution and lessadhesion on the surface. Also the air pockets among thepores generate a negative force against the Fg molecules tokeep them off the track, as shown in Figure 7F.An alternative model of Fg adsorption has been proposed by

Zbigniew et al.,34 termed the random sequential adsorptionmodel (RSA). The authors speculate that there are twopossibilities of Fg adsorption, including (i) irreversibly boundside-on interaction and (ii) reversibly bound end-on interactionas shown in Figure 8. They observed more adsorption and

clustering of Fg on hydrophobic surfaces compared tohydrophilic surfaces. In the present study, SC surfaces displayhydrophobic behavior, and according to the above model, therewould be more chances of irreversible side-on interaction. Incomparison, as the ES surfaces have an abundance of pores andair pockets, there is a possibility of reversibly bound end-oninteractions. This model also shows the conformation offibrinogen molecule interaction with SC and ES samples andalso supports the model of Siegismund et al.33

Endothelial Cell Response on Pristine SC and ESSamples. BMEC Adhesion. The interaction of BMEC duringadhesion on hydrophobic PVDF-HFP (Scheme 1) SC, ES

surfaces and on TCP (control) samples are shown in Figure 9Afor different plating densities. BMEC attached to all samples ina dose dependent manner over a period of 5 h. The influence ofsamples’ topography is summarized in Figure 9A. As shown inFigure 10A, a gradual increase in the number of cells attachedwas proportional to the increase in the seeding density.However, more cell adhesion to TCP and SC surfaces wasobserved at any cell concentration at any specific time point.The overall adhesion of BMEC was greater on SC sample;however, the attachment of BMEC to the sample wasdependent on sample morphology/topography. Morphologi-cally, most of the cells on flat samples were rounded to oval

Figure 7. Diffusion of fibrinogen molecules on (A) SC and (B) ES,adsorption of fibrinogen molecule on (C) SC and (D) ES, andclustering of fibrinogen molecule on (E) SC and (F) ES samples (notto scale).

Figure 8. (A) Side-on and (B) end-on adsorption of the fibrinogenmolecule on SC and ES samples.

Scheme 1. Structure of PVDF-HFP



shaped and moderately spread; however, on the ES sample, thecells observed to be elongated in shape.BMEC Proliferation. To examine the effect of topography on

proliferation, the cell number was assessed on days 1, 3, 5, 7,and 9, respectively. The proliferation kinetics of BMEC on SC,ES as well as TCP samples is presented in Figure 9B and showsthat the proliferation occurred on all samples with increasedduration of culture. BMEC on TCP and SC samples wassignificantly increased from 80 to 98% between 5 and 9 days;while BMEC proliferation was delayed with 40 to 65% of cellson ES samples (Figure 10B). BMEC proliferation on TCP wasshown to be significantly higher than ES but marginally higherthan that on the SC samples by day 9. True confluence was notreached on ES samples by day 9. The low proliferation ofBMEC on ES samples is most likely related to the lower levelsof BMEC adhesion (Figure 10A) and lower rates of BMECproliferation (Figure 10C).Many authors have provided evidence that surface top-

ography, roughness, and wettability influence the adhesion ofdifferent cell types, including fibroblast cells,35,36 platelets,37 ratbone marrow derived cells (RBMCs),38 and endothelialcells.39−41 It is evident that the process of adhesion betweencells and sample surfaces is guided by proteins within theserum.42−45 In the native tissue environment the cells bind tothe extracellular matrix via proteins through receptor−ligandinteractions. In vitro cell adhesion to the samples is guided by

proteins in the culture media and serum. Serum proteinfacilitates cell binding to ligands, while cell attachment toextracellular matrix is mediated by integrins: transmembranecellular receptors.46−48 Following the engagement of the ligand,integrins arrange themselves to form clusters, which furtherform focal adhesion complexes that contain structural andsignalling proteins. These focal adhesions make contact withcytoskeleton of the cells to control migration, survival,proliferation, and differentiation of the cells.48

Lower cell adhesion or reduced rates of adhesion andproliferation process on pristine ES samples have beenattributed to their ultrahydrophobic to superhydrophobicnature as can be seen in Figure 5 (contact angle values).Ishizaki et al. reported that the cells take more time toproliferate on superhydrophobic surfaces compared to that onsuperhydrophilic surfaces.49 This was attributed to high surfaceroughness values of ES samples making them superhydropho-bic in nature, which can influence the cell adhesion andproliferation on pristine ES samples.Thus, the sample surface topography plays an important role

in establishing adhesion of cells.50−52 The contact angle valuesare 94.2 ± 4.1° and 147.6 ± 4.5° for SC and ES PVDF-HFPsamples, respectively, representing diversity in topography ofthese samples. A previous study has shown that when a waterrepellent superhydrophobic surface was immersed in cellculture liquid it glistens with a silvery sheen, indicating that a

Figure 9. SEM micrographs of BMECs (A) adhesion and (B) proliferation seeded on pristine TCP, SC, and ES surfaces at different densities andtime of incubation (n = 3, std = <5%).



film of air covered the entire surface beneath the liquid.35 In thepresent case, this behavior of superhydrophobic surfaces is bestexplained by Cassie−Baxter and Wenzel53−55 as BMECattachment and proliferation is favored by hydrophilic (lesshydrophobic) or relatively high surface energy substrates.Overall, the results of WST-1 assay are in agreement with the

findings from proliferation experiments. The metabolic activityof BMEC on SC and ES samples is statistically comparable withTCP surfaces. There was delayed and 30−40% lessproliferation of BMEC on ES samples and their metabolicactivity was delayed when compared to positive control and SCsample.Endothelial Cell Response on Pre-Fibrinogen Ad-

sorbed SC and ES Samples. BMEC Adhesion. A parallel cellculture study was also performed on fibrinogen adsorbed SCand ES samples. A dynamic adsorption process of fibrinogenwas performed by immersing all three samples in normal serumconcentration (2 mg/mL) of fibrinogen solution in PBS for 2 h.The fibrinogen was quantified by ELISA (Figure 1) and atomicpercentage of nitrogen by XPS (Figure 2), which represents theadsorption of fibrinogen on SC higher than ES samples. Theadhesion profile of BMEC, seeded on fibrinogen preadsorbedSC, ES, and TCP samples as a function of different platingdensities is shown in Figure 11A. BMEC adhesion wasperformed for 5 h. However, the preadsorption of fibrinogenhas surface-regulation effect, which has direct influence on

adhesion of BMEC to the samples. Morphologically, most ofthe cells on flat samples are rounded to oval and moderatelyspread cells, but on ES samples the cells are elongated in shape,which indicates that cells are well spread on the modified ESsamples with adsorbed fibrinogen molecules and their numberalso increased, as shown in Figure 11A. We observed a gradualincrease in number of cell attachment with increasing seedingdensity on all the samples including a 2-fold increase in numberof cells on ES samples after the adsorption of fibrinogen(Figure 12A) compared to pristine samples.

BMEC Proliferation. To observe the effect of fibrinogenadsorbed molecules on proliferation, the cell number wasassessed on days 1, 3, 5, 7, and 9. The proliferation kinetics ofBMEC on preadsorbed fibrinogen SC and ES as well as TCPsamples is presented in Figure 11B, which shows that theproliferation occurred on all the samples. A surprising effect ofadsorbed fibrinogen was observed on the ES sample in the formof 98% proliferation of BMEC by day 9, which was 30−40%higher than on pristine ES samples. However, no apparentinfluence on proliferation of SC and TCP samples’ on BMECafter the adsorption of Fg was observed. Thus, under theinfluence of the Fg molecule, true confluence of BMEC wasachieved. A marginally high number of BMEC has beenobserved on ES sample than on SC sample by day 9 ofproliferation (Figure 12B) due to specific morphologicalfeatures of ES surfaces (Figure 11B). All samples displayed

Figure 10. Bar graph depicting (A) BMECs adhesion on pristine TCP (control), SC, and ES samples. (B) BMECs proliferation on pristine TCP(control), SC, and ES samples and (C) the WST-1 assay representing the viability, metabolic activity, and proliferation of cultured BMECs onpristine TCP (control), SC, and ES surfaces at days 1, 3, 5, 7, and 9 (n = 3, std = <5%).



the ability to support BMEC viability with adsorbed fibrinogen,which was also in line with WST-1 assay as shown in Figure12C.It is important to understand the process of Fg adsorption,

which triggers the adhesion and proliferation of BMECs. The Dand C domains of the Fg (Figure 5) molecule interact with thesample’s surface, which plays a critical role in fibrinogenconformation. Surface roughness, surface energy, surfacechemistry, and topography are the key players that affect thestructure of adsorbed protein.56 The flat samples or TCP andSC PVDF-HFP surfaces are hydrophobic in nature and thehydrophobic domain of fibrinogen molecule, which wasshielded by hydrophilic corona attached to both surfaces byhydrophobic interactions. This hydrophobic interaction andunfolding of fibrinogen molecule exposes the hydrophiliccorona, allowing another layer of Fg to form by hydrophilicinteractions among protein molecules due to limited surfacesarea of flat sample. Therefore, no significant change wasobserved in adhesion and proliferation of cells on pristine andpreadsorbed Fg flat surfaces. Due to the hydrophobic nature ofFg adsorbed-SC sample, there are many chances of Fg clusterformation as discussed above. But in the case of ES PVDF-HFPsamples, there is a large surface area available for the adhesionof more Fg molecules, which may contribute to thedevelopment of monolayer on ES sample and cause more cell

adhesion and proliferation. Due to voids among fibres, there aremany possibilities of end-on interaction and less chances ofside-on Fg molecules attachment as discussed above. Thisadhesion behavior is presented in the schematic model inFigure 13, demonstrating that when the Fg molecule formsmultilayers on flat samples, there is no significant effect on celladhesion and proliferation. However, on ES samples, Fg formsa monolayer due to the large available surface area, which couldresult in a change in the surface nature from superhydrophobicto less hydrophobic due to exposed hydrophilic region of Fgmolecule. Using contact angle measurement, we furtherconfirmed that the wettability of ES and SC samples hadchanged after fibrinogen adsorption. The morphology of cellson ES samples also changed after Fg adsorption with the cellsdisplaying a well spread phenotype, compared with pristine ESand flat surfaces. Herrick et al reported that there are integrinand nonintegrin receptors that bind to specific sites on Fgmolecule that mediate cell adhesion and spreading. In the caseof endothelial cells, the αv β3, α5 β1, 130 kd receptors, andCD44 surface receptors with Aα572−574, Aα572−574, Bβ15−42, and Bβ recognition sites are responsible for adhesion andproliferation.57 Therefore, a different conformation of Fgmolecule on ES and SC surfaces may influence the exposureof endothelial cells and platelet adhesion sites in different ways.

Figure 11. SEM micrographs of BMECs (A) adhesion and (B) proliferation seeded on preadsorbed fibrinogen TCP, SC, and ES surfaces at differentdensities and times of incubation (n = 3, std = <5%).



The data presented in Figure 12B reveals that after theadsorption of fibrinogen there is more facilitation of BMECproliferation on ES sample compared to SC samples. Although,the fibrinogen conformation can recruit the endothelial cells on

ES PVDF-HFP surface but least chances of in vivo bloodcoagulation can be overcome by establishing in vitroendothelial cell proliferation. Once there is establishedconfluent endothelial cell lining, the hazard of platelet adhesionand blood coagulation can be eliminated.

■ CONCLUSIONSTuning surface morphology using electrospinning can directlyinfluence the wettability parameters, which indirectly affect theadsorption, adhesion, and proliferation of proteins and cells.Highly rough surfaces or superhydrophobic surfaces are notdesirable for BMECs adhesion and proliferation. By changingthe surface chemistry using dynamic adsorption of Fg showeddramatic improvement of BMEC adhesion and proliferation onES surfaces. Thus, using a simple and dynamic surfacemodification procedure it is possible to enhance the BMECadhesion and proliferation on superhydrophobic ES PVDF-HFP surfaces. The consequences of fibrinogen adsorption onthe model electrospun systems also allow us to understand thenatural interaction between endothelial cells and theirextracellular environment.

■ AUTHOR INFORMATION

Corresponding Author*E-mail: [email protected]. Tel.: 61 8 83023719.

Figure 12. Bar graph depicting (A) BMECs adhesion on preadsorbed fibrinogen TCP (control), SC, and ES samples. (B) BMECs proliferation onfibrinogen preadsorbed TCP (control), SC, and ES samples and (C) the WST-1 assay representing the viability, metabolic activity, and proliferationof cultured BMECs on fibrinogen preadsorbed TCP (control), SC, and ES surfaces at days 1, 3, 5, 7, and 9 (n = 3, std = <5%).

Figure 13. Plausible interaction of BMECs with surfaces of (A)pristine SC, (B) preadsorbed fibrinogen SC, (C) pristine ES surface,and (D) preadsorbed fibrinogen ES.



mailto:[email protected]

NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSThe authors are thankful to ARC for financial support to carryout this research work through ARC-Discovery grant. One ofthe authors (F.A.) is thankful for the scholarship supported byNED University of Engineering and Technology, Karachi,Pakistan.

■ REFERENCES(1) Lee, A. Y.; Mahler, N.; Best, C.; Lee, Y. U.; Breuer, C.K.. Transl.Res. 2014, DOI: 10.1016/j.trsl.2014.01.014.(2) Sigler, M. J. Clin. Exp. Cardiol. 2013, 4, e125.(3) Awumey, E. M.; Bridges, L. E.; Williams, C. L.; Diz, D. I. J.Pharmacol. Exp. Ther. 2013, 346, 38−47.(4) Ramadoss, J.; Pastore, M. B.; Magness, R. R. Clin. Exp. Pharmacol.Physiol. 2013, 40, 753−764.(5) Williamson, K. A.; Hamilton, A.; Reynolds, J. A.; Sipos, P.;Crocker, I.; Stringer, S. E.; Alexander, Y. M. Aging Cell 2013, 12, 139−147.(6) Glynn, J. J.; Hinds, M. T. Tissue Eng., Part B 2013, DOI: 10.1089/ten.teb.2013.0285.(7) Grey, C. P.; Newton, S. T.; Bowlin, G. L.; Haas, T. W.; Simpson,D. G. Biomaterials 2013, 34, 4993−5006.(8) Sill, T. J.; von Recum, H. A. Biomaterials 2008, 29, 1989−2006.(9) Huang, N. F. Tissue Engineering and Regenerative Medicine;Springer: Berlin, 2013; Vol. 9, pp 313−323.(10) Bourget, J. M.; Guillemette, M.; Veres, T.; Auger, F.o.A.;Germain, L. Adv. Biomater. Sci. Biomed. Appl. 2013, DOI: 10.5772/56420.(11) Gugutkov, D.; Gonzalez-García, C.; Altankov, G.; Salmeron-Sanchez, M. J. Bioact. Compat. Polym. 2011, 26, 375−387.(12) Lindon, J. N.; McManama, G.; Kushner, L.; Merrill, E. W.;Salzman, E. W. Blood 1986, 68, 355−362.(13) Weijers, E.; Van Wijhe, M.; Joosten, L.; Horrevoets, A.; DeMaat, M.; Van Hinsbergh, V.; Koolwijk, P. J. Thromb. Haemost. 2010,8, 2800−2809.(14) Schmidt, S. P.; Hunter, T. J.; Falkow, L. J.; Michelle Evancho,M.; Sharp, W. V. J. Vasc. Surg. 1985, 2, 898−906.(15) Bos, G. W.; Poot, A. A.; Beugeling, T.; Van Aken, W.; Feijen, J.Arch. Physiol. Biochem. 1998, 106, 100−115.(16) Bots, J.; Van der Does, L.; Bantjes, A. Biomaterials 1986, 7,393−399.(17) Venugopal, J.; Zhang, Y.; Ramakrishna, S. Nanotechnology 2005,16, 2138.(18) E Niklason, L.; S Langer, R. Transpl. Immunol. 1997, 5, 303−306.(19) Dejana, E.; Colella, S.; Languino, L. R.; Balconi, G.; Corbascio,G. C.; Marchisio, P. C. J. Cell Biol. 1987, 104, 1403−1411.(20) Weimar, B.; Delvos, U. Aterioscl. Throm. Vas. 1986, 6, 139−145.(21) Yao, C.; Li, X.; Neoh, K.; Shi, Z.; Kang, E. Appl. Surf. Sci. 2009,255, 3854−3858.(22) Ahmed, F.; Choudhury, N. R.; Dutta, N. K.; Zannettino, A.;Knott, R. Biomacromolecules 2013, 14, 3850−3860.(23) Tang, Z.; Wu, H.; Du, D.; Wang, J.; Wang, H.; Qian, W.-j.;Bigelow, D. J.; Pounds, J. G.; Smith, R. D.; Lin, Y. Talanta 2010, 81,1662−1669.(24) Liu, D. P.; Majewski, P.; O’Neill, B. K.; Ngothai, Y.; Colby, C. B.J. Biomed. Mater. Res., Part A 2006, 77, 763−772.(25) Beamson, G.; Briggs, D. Mol. Phys. 1992, 76, 919−936.(26) Shi, Q.; Rafii, S.; Hong-De Wu, M.; Wijelath, E. S.; Yu, C.;Ishida, A.; Fujita, Y.; Kothari, S.; Mohle, R.; Sauvage, L. R. Blood 1998,92, 362−367.(27) Gronthos, S.; Zannettino, A. C. W.; Hay, S. J.; Shi, S.; Graves, S.E.; Kortesidis, A.; Simmons, P. J. J. Cell Sci. 2003, 116, 1827−1835.(28) Mouton, J. W.; Dudley, M. N.; Cars, O.; Derendorf, H.;Drusano, G. L. Int. J. Antimicrob. Agents 2002, 19, 355−358.

(29) Clarke, M. L.; Wang, J.; Chen, Z. J. Phys. Chem. B 2005, 109,22027−22035.(30) Kamath, S.; Lip, G. QJM Int. J. Med. 2003, 96, 711−729.(31) Fuss, C.; Palmaz, J. C.; Sprague, E. A. J. Vasc. Interv. Radiol.2001, 12, 677−682.(32) Sheinerman, F. B.; Norel, R.; Honig, B. Curr. Opin. Struct. Biol.2000, 10, 153−159.(33) Siegismund, D.; Keller, T. F.; Jandt, K. D.; Rettenmayr, M.Macromol. Biosci. 2010, 10, 1216−1223.(34) Adamczyk, Z.; Barbasz, J.; Ciesla, M. Langmuir 2011, 27, 6868−6878.(35) Ranella, A.; Barberoglou, M.; Bakogianni, S.; Fotakis, C.;Stratakis, E. Acta Biomater. 2010, 6, 2711−2720.(36) Koufaki, N.; Ranella, A.; Aifantis, K. E.; Barberoglou, M.;Psycharakis, S.; Fotakis, C.; Stratakis, E. Biofabrication 2011, 3, 045004.(37) Sun, T.; Tan, H.; Han, D.; Fu, Q.; Jiang, L. Small 2005, 1, 959−963.(38) Alves, N. M.; Shi, J.; Oramas, E.; Santos, J. L.; Tomas, H.; Mano,J. F. J. Biomed. Mater. Res., Part A 2009, 91, 480−488.(39) Toes, G.; Van Muiswinkel, K.; Van Oeveren, W.; Suurmeijer, A.;Timens, W.; Stokroos, I.; Van den Dungen, J. Biomaterials 2002, 23,255−262.(40) Lee, J.; Chu, B. H.; Chen, K. H.; Ren, F.; Lele, T. P. Biomaterials2009, 30, 4488−4493.(41) Kim, M. S.; Khang, G.; Lee, H. B. Prog. Polym. Sci. 2008, 33,138−164.(42) Schakenraad, J.; Busscher, H.; Wildevuur, C. R. H.; Arends, J. J.Biomed. Mater. Res. 1986, 20, 773−784.(43) Hersel, U.; Dahmen, C.; Kessler, H. Biomaterials 2003, 24,4385−4415.(44) Deligianni, D.; Katsala, N.; Ladas, S.; Sotiropoulou, D.; Amedee,J.; Missirlis, Y. Biomaterials 2001, 22, 1241−1251.(45) Maheshwari, G.; Brown, G.; Lauffenburger, D. A.; Wells, A.;Griffith, L. G. J. Cell Sci. 2000, 113, 1677−1686.(46) Buck, C. A.; Horwitz, A. J. Cell Sci. Suppl. 1987, 8, 231−250.(47) Choi, C. K.; Breckenridge, M. T.; Chen, C. S. Trends Cell. Biol.2010, 20, 705−714.(48) Clark, E. A.; Brugge, J. S. Science 1995, 268, 233−239.(49) Ishizaki, T.; Saito, N.; Takai, O. Langmuir 2010, 26, 8147−8154.(50) Anselme, K.; Ploux, L.; Ponche, A. J. Adhes. Sci. Technol. 2010,24, 831−852.(51) Lord, M. S.; Foss, M.; Besenbacher, F. Nano Today 2010, 5, 66−78.(52) Le Saux, G.; Magenau, A.; Bocking, T.; Gaus, K.; Gooding, J. J.PloS One 2011, 6, e21869.(53) Wenzel, R. N. J. Phys. Chem. 1949, 53, 1466−1467.(54) Kwon, Y.; Choi, S.; Anantharaju, N.; Lee, J.; Panchagnula, M.;Patankar, N. Langmuir 2010, 26, 17528−17531.(55) Andrews, H.; Eccles, E.; Schofield, W.; Badyal, J. Langmuir2011, 27, 3798−3802.(56) Norde, W. Wiley Online Library 1996, 5−18.(57) Herrick, S.; Blanc-Brude, O.; Gray, A.; Laurent, G. Int. J.Biochem. Cell Biol. 1999, 31, 741−746.



engineering interaction between bone marrow derived endothelial cells and electrospun surfaces for...

Documents