Cooperative control of blood compatibility and re-endothelialization by immobilized heparin and substrate topography

Acta Biomaterialia 15 (2015) 150–163

Contents lists available at ScienceDirect

Acta Biomaterialia journal homepage: www.elsevier.com/locate/actabiomat

Cooperative control of blood compatibility and re-endothelialization by immobilized heparin and substrate topography Yonghui Ding a, Meng Yang a, Zhilu Yang b, Rifang Luo b, Xiong Lu b, Nan Huang b, Pingbo Huang c,d,e, Yang Leng a,⇑ a

Department of Mechanical and Aerospace Engineering, The Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong Key Laboratory of Advanced Technology of Materials, School of Material Science and Engineering, Southwest Jiaotong University, Chengdu 610031, Sichuan, China Division of Life Science, The Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong d Division of Biomedical Engineering, The Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong e State Key Laboratory of Molecular Neuroscience, The Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong b c

a r t i c l e

i n f o

Article history: Received 4 August 2014 Received in revised form 4 December 2014 Accepted 16 December 2014 Available online 23 December 2014 Keywords: Blood compatibility Endothelial cell Substrate topography Heparin

a b s t r a c t A wide variety of environmental cues provided by the extracellular matrix, including biophysical and biochemical cues, are responsible for vascular cell behavior and function. In particular, substrate topography and surface chemistry have been shown to regulate blood and vascular compatibility individually. The combined impact of chemical and topographic cues on blood and vascular compatibility, and the interplay between these two types of cues, are subjects that are currently being explored. In the present study, a facile polydopamine-mediated approach is introduced for immobilization of heparin on topographically patterned substrates, and the combined effects of these cues on blood compatibility and re-endothelialization are systematically investigated. The results show that immobilized heparin and substrate topography cooperatively modulate anti-coagulation activity, endothelial cell (EC) attachment, proliferation, focal adhesion formation and endothelial marker expression. Meanwhile, the substrate topography is the primary determinant of cell alignment and elongation, driving in vivo-like endothelial organization. Importantly, combining immobilized heparin with substrate topography empowers substantially greater competitive ability of ECs over smooth muscle cells than each cue individually. Moreover, a model is proposed to elucidate the cooperative interplay between immobilized heparin and substrate topography in regulating cell behavior. Ó 2015 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

1. Introduction Coronary artery disease, mostly caused by atherosclerosis, has been a leading cause of death and morbidity in developed countries. Vascular stents or grafts are the most preferred and commonly used prosthesis for treating severe cases of coronary artery disease. However, such interventions are associated with major complications, such as in-stent restenosis, caused by the proliferation of vascular smooth muscle cells (SMCs), and thrombosis, induced by inadequate re-endothelialization on the surface and poor blood compatibility of the vascular stents or grafts. Since the vascular endothelium prevents blood coagulation and SMC proliferation, rapid re-endothelialization is critical to the success of vascular stents or grafts [1]. Several researchers have successfully enhanced re-endothelialization via various surface modification approaches ⇑ Corresponding author. Tel.: +852 2358 7185; fax: +852 2358 1543. E-mail address: [email protected] (Y. Leng). http://dx.doi.org/10.1016/j.actbio.2014.12.014 1742-7061/Ó 2015 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

[2–7]. However, the clinical results have been unsatisfactory [8], which may be attributed to the ignorance of competitive growth of endothelial cells (ECs) over SMCs in vivo [9,10]. Therefore, there remains an urgent need for developing multifunctional stent/graft materials, which are able to improve the blood compatibility, favor rapid re-endothelialization and enhance the competitive ability of ECs over SMCs. In vivo, the vascular endothelium consisting of a monolayer of ECs attaches to the basement membrane (BM), which presents a variety of biophysical and biochemical cues. Recently, there is increasing evidence that the substrate topography, as one of the biophysical cues, is an important surface parameter in determining EC fates and functions [2–4]. In particular, anisotropic topography has been shown to induce an in vivo-like EC elongation and alignment, and potentially improve the EC function, e.g. by enhanced EC migration [2] and an athero-resistant phenotype [11]. Moreover, substrate topography also has a profound influence on the proliferation and differentiation of SMCs [12,13].

Y. Ding et al. / Acta Biomaterialia 15 (2015) 150–163

In addition to substrate topography, the biochemical cues presented on the BM also provide instructive signals to ECs. A number of biomolecules have been immobilized onto biomaterial surfaces to promote re-endothelialization [5–7], improve the blood compatibility [14,15] and inhibit SMC proliferation [16]. Heparin, the most widely used anticoagulant drug, has been immobilized on material surfaces to improve blood compatibility by increasing the affinity of antithrombin III to thrombin [14,17,18]. Moreover, the immobilized or released heparin has also been demonstrated to inhibit SMC adhesion and proliferation [19–22]. However, the effects of heparin on EC proliferation are ambiguous. While some groups have demonstrated the enhancement of EC growth induced by heparin [23,24], other groups did not detect significant effects of heparin on EC growth [25,26]. Despite continuing efforts in this field, an ideal system is yet to be proposed that confers multifunctionality in terms of blood compatibility and re-endothelialization. Few studies have modulated blood compatibility and re-endothelialization by integrating both topographic and biochemical cues on biomaterials, and little is known about the cooperative interplay between these two cues. In our previous study, we demonstrated that blood and vascular compatibility can be induced by rationally designed surface topography [27]. Topographic 1 lm grooves induced the maximum EC attachment and proliferation, while concurrently inhibiting SMC growth and platelet activation. In this study, we combine the topographic cue with the biochemical cue via polydopamine-mediated immobilization of heparin on topographically patterned substrates. We also explore the combined effects of substrate topography and immobilized heparin to elucidate their interplay in regulating vascular cell responses, and investigate how this ultimately impacts the blood compatibility and re-endothelialization. 2. Materials and methods 2.1. Fabrication of micropatterned substrates The micropatterned substrates with anisotropic grooves (ridge = groove = 1 lm, depth = 3.5 lm) were first fabricated on silicon wafers through standard photolithography and deep reactive-ion etching. The patterned silicon substrate was then coated with a thin layer of titanium using a radio frequency sputtering system (Model Explorer 14, Denton Vacuum). A post-sputtering heat treatment was conducted at 700 °C for 1 h under an air flow to transfer the coating layer to the titanium oxide. 2.2. Immobilization of heparin The heparin immobilization processes are illustrated in Fig. 1. Firstly, primary amine (–NH2) functional groups were introduced onto the substrate surface via co-conjugation of polydopamine (PDA) and poly(ethyleneimine) (PEI). The substrates were immersed in dopamine hydrochloride solution (2 mg ml1, Sigma–Aldrich, USA) in a tris-(hydroxymethyl) aminomethane (Tris) buffer (10 mM, 15 ml, pH 8.5) at 25 °C for 90 min in an open vessel. After that, PEI solution (20 mg ml1, average Mw 1300, 50 wt.% in H2O, Sigma–Aldrich) in a Tris buffer (10 mM, 15 ml, pH 8.5) was added and surface functionalization was further performed for another 30 min at 25 °C in an open vessel. The functionalized substrate (PEI/PDA) was vigorously washed with deionized (DI) water and blown dry under a weak stream of nitrogen. Secondly, to immobilize the heparin onto PEI/PDA films, the carboxylic groups of heparin were activated in advance by Nhydroxysuccinimide (NHS) and N-(3-dimethylaminopropyl)-N0 ethylcarbodiimide (EDC), and then were further bound to the PEI/PDA films via their surface –NH2 [18]. Briefly, 2 mg of heparin

151

(sodium salt from porcine intestinal mucosa, potency P180 units mg1, Sigma–Aldrich) was dissolved in a 20 ml MES (50 mmol l1, Sigma–Aldrich) buffer solution (pH 5.4) in order to minimize hydrolysis of EDC and then mixed with 20 mg of EDC (Sigma–Aldrich) and 4.8 mg of NHS (Sigma–Aldrich). The PEI/PDA substrates were subsequently immersed in the heparin solution at 25 °C for 12 h. After reaction, the heparin-immobilized substrates (Hep-PEI/PDA) were washed with phosphate-buffered saline (PBS, pH 7.4) and DI water, then blown dry under a weak stream of nitrogen. 2.3. Physiochemical characterization 2.3.1. Film thickness The thickness of the deposited PEI/PDA film before and after heparin immobilization was measured by a spectroscopic ellipsometer (M-2000V, J.A. Woollam, USA). D and W values measured at a wavelength of 370–1000 nm were chosen for data analysis, and the Cauchy model was used to determine the thickness of the deposited PEI/PDA film. 2.3.2. Surface wettability and morphology The surface wettability of various surfaces was examined by water contact angle (WCA) measurements with a contact angle instrument (Digidrop, France) and applying the sessile drop method. The surface topographies of various surfaces were analyzed by an atomic force microscope (AFM; NanoScope IIIa/Dimension 3100, Digital Instruments, CA). The AFM images were obtained in tapping mode and the root mean square (RMS) was used to evaluate the surface roughness on the basis of a 20 lm  20 lm scan area. The surface morphologies of various surfaces were observed by scanning electron microscopy (SEM; JSM-6700F, JEOL, Japan). 2.3.3. Zeta potential The zeta potentials of samples were examined with a commercial electrokinetic analyzer (EKA, Anton Paar GmbH, Graz, Austria). A 0.001 M KCl solution (pH 7.4) was chosen as the electrolyte for the test, which was conducted at room temperature. For each sample, the zeta potential was measured five times and the average value was reported. 2.3.4. Chemical compositions The quantification of surface elemental composition of samples was measured by X-ray photoelectron spectroscopy (XPS; Kratos, Axis Ultra DLD). A monochromatic Al Ka X-ray was used as the excitation source (hm = 1486.6 eV), running at 15 kV and 150 W. The atomic percentages of the various elements were derived from broad range spectra, using the Al source in a low-resolution mode (pass energy 160 eV), while a pass energy of 20 eV was used for the high-resolution spectra of N1s and S2p. The C1s peak (binding energy 285.0 eV) was used as a reference for charge correction. 2.4. Stability test of PEI/PDA films and immobilized heparin The PEI/PDA and Hep-PEI/PDA substrates were immersed into PBS solution (pH 7.4) at 37 °C in dynamic state (orbital shakers at 250 rpm) for 3, 7, 15 and 30 days. The surface morphology and atomic concentrations of S2p for both untreated and treated substrates were then evaluated, by SEM and XPS, respectively. 2.5. In vitro blood compatibility The fresh human whole blood used in our experiments was obtained legally from the central blood station of Chengdu, China. Whole blood obtained from healthy human volunteers, who were aspirin-free for a minimum of 2 weeks prior to donating, was

152

Y. Ding et al. / Acta Biomaterialia 15 (2015) 150–163

Fig. 1. Schematic illustration of the heparin immobilization process on topographically patterned substrates and the possible manner of PEI/PDA conjugation on the TiO2 surface.

anti-coagulated with trisodium citrate in a 9:1 (v/v) ratio. The analysis was performed within 12 h after the blood donation. 2.5.1. Platelet adhesion and activation The protocol used to evaluate platelet adhesion is described in our previous report [28]. Briefly, to prepare platelet-rich plasma (PRP), whole blood was centrifuged (200g, 15 min) and the supernatant was collected. The PRP (50 ll) was placed on the substrate surface and incubated at 37 °C for 2 h. After washing in PBS, the substrates with adherent platelets were fixed with 2.5% glutaraldehyde for 1 h at room temperature. The samples were rinsed three times with PBS and then dehydrated using an ethanol series (30, 50, 75, 90 and 100%; 15–30 min each), and subsequently maintained in 100% ethyl alcohol until drying. After critical point drying, the samples were mounted on copper stubs and sputter-coated with gold for examination by SEM (JSM-7100F, JEOL, Japan). 2.5.2. Activated partial thromboplastin time (APTT) assay For APTT measurement, the samples were placed to a 24-well culture plate. The whole blood was centrifuged (400g, 15 min) and the supernatant was collected to get human platelet-poor plasma (PPP). The fresh PPP (200 ll) and actin-activated cephthaloplastin reagent (100 ll) were added onto the sample surface, followed by the addition of 0.03 M CaCl2 solution (100 ll) after incubation for 30 min at 37 °C. The clotting time of the plasma solution was measured by a coagulometer (ACL 200, Instrumentation Laboratory Co. USA). 2.6. Evaluations of Human umbilical vein endothelial cell (HUVEC) responses and EC selectivity HUVECs (Lonza, Baltimore, MD; passages 2–7) were cultured in 0.2%-gelatin-coated T75 flasks with a Lonza endothelial growth medium bullet kit containing 100 U ml1 penicillin and 100 U ml1 streptomycin. Human umbilical artery smooth muscle cells (HUASMCs) were obtained through the slow outgrowth of the cells from small pieces of umbilical artery in the medium, and were cultured in DMEM/F12 medium containing 10% fetal bovine serum. All cells

were cultured at 37 °C in a humidified incubator containing 5% CO2. HUVECs were plated on the samples at a density of 1  104 cells cm2, and incubated for predetermined time periods. The cell culture medium was changed every 2–3 days. 2.6.1. HUVEC attachment, morphology and proliferation Cell attachment, morphology and proliferation were evaluated based on cell number counting and cell cytoskeleton observation on the surfaces after 2 h, 1 day and 3 days in culture. To perform these evaluations, HUVECs cultured on the samples were fixed with 4% paraformaldehyde (PFA; Sigma–Aldrich) in PBS for 20 min at room temperature. The cells were then permeabilized using 0.5% Triton X-100 in PBS for 5 min and blocked with 1% bovine serum albumin (BSA; Sigma–Aldrich) in PBS for 60 min to reduce nonspecific staining. Subsequently, to examine cytoskeletal formation, the cells were labeled for F-actin using phalloidin conjugated to tetramethyl rhodamine isothiocyanate (2 lg ml1; Sigma–Aldrich) for 60 min at room temperature. The cells were also counter-stained with 40 -6-diamidino-2-phenylindole (DAPI, 1 lg ml1; Sigma–Aldrich) to observe the nuclei. The samples were finally washed three times in PBS and mounted on microscope slides for examination using confocal laser scanning microscopy (LSM710, Zeiss, Germany). For quantitation, images were collected using 10 and 40 objectives. The number and morphology of both cells were analyzed using ImageJ software (NIH, USA). 2.6.2. HUVEC viability HUVEC metabolic activity (viability) was investigated by cell counting kit-8 (CCK-8) (Dojindo Laboratories) after incubation for 1 and 3 days, respectively. A 300 ll aliquot of fresh medium containing 10% CCK-8 reagent was added to each sample and incubated for 2 h in standard culture conditions. Next, 200 ll of each of these solutions was transferred to a 96-well plate. The absorbance was measured at 450 nm by a microplate reader. 2.6.3. HUVEC focal adhesion After incubation of HUVECs for 1 day, the cells were fixed with 4% PFA in PBS, permeabilized using 0.5% Triton X-100, blocked with

Y. Ding et al. / Acta Biomaterialia 15 (2015) 150–163

1% BSA, treated with primary antibodies of monoclonal antihuman vinculin antibodies (No. ab18058, 1:50 dilution; Abcam, Cambridge, UK) overnight at 4 °C, incubated with secondary antibodies of goat polyclonal secondary antibodies against mouse IgG (H&L, fluorescein isothiocyanate (FITC)-linked; No. ab6785, 1:500 dilution; Abcam) for 60 min at room temperature, and finally counter-stained with DAPI. The samples were then examined by confocal laser scanning microscopy, and images were collected using 63 objectives. The number, size and morphology of focal adhesions were analyzed using ImageJ software. 2.6.4. HUVEC phenotype To analyze the expression of two key endothelial cell markers – platelet endothelial cell adhesion molecule (PECAM-1/CD31) and von Willebrand factor (vWF) – HUVECs were plated on the samples at a higher density of 2.5  104 cells cm2. After incubation for 1 day, the cells were fixed with 4% PFA in PBS, permeabilized using 0.5% Triton X-100, blocked with 1% BSA, treated with primary antibodies overnight at 4 °C, incubated with secondary antibodies for 60 min at room temperature, and finally counter-stained with DAPI. Antibodies were used at the following dilutions: 1:20 for monoclonal mouse anti-human CD31 antibody (Dako, Copenhagen, Denmark), 1:200 for polyclonal rabbit anti-human vWF antibody (Dako) and 1:500 for FITC-linked goat polyclonal secondary antibodies against mouse IgG (Abcam). 2.6.5. HUVEC and HUASMC co-culture The co-culture of HUVECs and HASMCs was performed to evaluate the selectivity of ECs over SMCs. HUVECs were labeled with

153

Cell Tracker Green CMFDA and the HASMCs were labeled with Orange CMTMR according to the manufacturer’s instructions (Molecular Probes, America). In brief, HUVECs were incubated in DMEM/F12 medium supplemented with 5 mM Cell Tracker Green CMFDA for 20 min, then washed with PBS and cultured in fresh medium free of dye for 30 min. HASMCs were labeled in the same manner using Cell Tracker Orange CMTMR. Subsequently, the fluorescently labeled HUVECs and HASMCs were isolated by 0.25% trypsin–EDTA solution and centrifuged (130 g, 2 min). The cells were resuspended separately in DMEM/F12 medium with 10% fetal bovine serum. Finally, the HUVEC and HASMC suspensions were mixed in a volume ratio of 1:1 and the cells were seeded at a density of 2  104 cells cm2 of each cell type. The competitive cell attachment and growth were examined after incubation for 2 h and 1 day. The cells were observed by a Leica DMRX fluorescence microscope (Leica, Germany). The amount of adherent cells was calculated from at least 12 images for each sample. 2.7. Statistical analysis All data are expressed as mean ± standard deviation. All the cell data were quantified from at least three independent experiments, each with at least three replicates. One-way analysis of variance was used to measure differences for experiments with multiple data sets, with a post hoc Tukey multiple comparison test performed between groups having significant differences. A value of p 6 0.05 was considered statistically significant. Within the figures, the significance is denoted by the following marks: ⁄ or # for p < 0.05; ⁄⁄ or ## for p < 0.01; and ⁄⁄⁄ or ### for p < 0.001.

Fig. 2. Physicochemical properties: (A) film thickness, (B) water contact angles, (C) zeta potentials, (D) topographic AFM images with measured RMS values of surface roughness and (E) surface morphological SEM images of PEI/PDA and Hep-PEI/PDA modified surfaces as well as the pristine TiO2 surface. Statistically significant differences are marked as follows: ⁄ vs. TiO2 surface.

154

Y. Ding et al. / Acta Biomaterialia 15 (2015) 150–163

3. Results 3.1. Physicochemical properties To immobilize the heparin to the patterned substrate, we utilized the co-conjugation of PDA and PEI on the TiO2 surface to functionalize the patterned substrate with amine groups while preserving pattern fidelity. During the PEI/PDA co-conjugation (Fig. 1), the PDA may first bind to underlying TiO2 substrates through the catecholic OH group of the PDA [29], then the PEI with amine would bind to the pre-deposited PDA with quinone via Schiff base and/or Michael addition reactions [30]. The deposition of the PEI/PDA films did not significantly alter the geometry or size of the underlying patterns (Fig. S1). The thickness of the deposited PEI/PDA film was 9.3 ± 1.0 nm, as determined by ellipsometry (Fig. 2A). The immobilization of heparin did not cause any noticeable change in the film thickness, which may be a result of the formation of multiple binding sites along the polysaccharide backbone of heparin and the subsequent ‘‘side-on’’ orientation of the heparin immobilized on the PEI/PDA films. WCA measurement suggested the successful deposition of PEI/ PDA film and immobilization of heparin. As shown in Fig. 2B, the WCA was increased from 46.3 ± 0.7° for pristine TiO2 to 61.8 ± 0.7° after the deposition of the PEI/PDA film, which was similar to the WCA of the PDA coating measured in our previous study [31]. In contrast, the WCA was dramatically decreased to 30.6 ± 1.1° after the immobilization of heparin, which may be due to the addition of hydrophilic groups, such as –COOH and – SO3, present on the heparin. The zeta potentials of various substrates also confirmed the successful modification of the substrate (Fig. 2C). The zeta potential of the PEI/PDA was less negative compared to pristine TiO2, which may be ascribed to the addition of –NH2, which provides some positive charge (–NH+3) on the PEI/PDA. However, it became substantially more negative after immobilization of heparin, since heparin contains a high content of negatively charged sulfur and carboxyl groups. The surface morphology of modified substrates was examined by AFM and SEM (Fig. 2D and E). AFM images showed that the deposition of PEI/PDA generated a rough surface, with the RMS roughness being increased to 8.0 ± 2.7 nm on the PEI/PDA and 6.7 ± 2.8 nm on the Hep-PEI/PDA. SEM images showed the addition of some particles on the surface of PEI/PDA, which may result from

the conjugation of PDA aggregates from the dopamine solution. There was no apparent change in surface morphology after heparin immobilization. The successful modification of the PEI/PDA and heparin was further confirmed by XPS (Fig. 3). After deposition of PEI/PDA, the peak intensity of C1s and N1s significantly increased (Fig. 3B), while the peak intensity of Ti2p3 dramatically decreased. The successful immobilization of heparin was evidenced by the presence of a new peak of S2p, as shown in Fig. 3C (Hep-PEI/PDA). Moreover, the S atomic concentrations on both flat and grooved Hep-PEI/PDA were the same (data not shown), suggesting the same surface chemical composition was obtained on both flat and grooved substrates. 3.2. Surface stability The long-term stability of any surface modification of an implantable device is vital to ensure its safe performance in vivo. SEM images of the deposited PEI/PDA film before and after incubation in PBS for various periods of time showed no apparent differences in surface morphology and no phenomena of swelling observed on the PEI/PDA even after incubation for 30 days, which indicates good dynamic stability of PEI/PDA films (Fig. S2). Moreover, the S content was quantified, confirming the good stability of immobilized heparin after incubation in PBS (Fig. 4). There was no burst release of heparin and almost 80% of the heparin was retained after incubation for 30 days, which may be attributed to the good stability of the PEI/PDA film and the robust heparin binding. The good stability of the immobilized heparin in this study could reduce the risk of hemorrhage and guarantee safe application in vivo [32,33]. 3.3. Cooperative control of blood compatibility 3.3.1. Platelet adhesion and activation The morphology of adherent platelets was inspected with SEM (Fig. 5A). On the flat substrates, most adherent platelets were in a dendritic state, with a few pseudopodia, on the TiO2, while more adherent platelets, in a spreading or fully spread state, were observed on the PEI/PDA. In contrast, most adherent platelets remained spherical, without pseudopodia, on the Hep-PEI/PDA. In addition, the presence of grooves did not apparently affect the platelet morphology.

Fig. 3. Chemical composition analysis. (A) XPS wide spectra of the substrates and high-resolution spectra of (B) N1s and (C) S2p.

Y. Ding et al. / Acta Biomaterialia 15 (2015) 150–163

155

15,833 platelets mm2 on the TiO2. Moreover, PEI/PDA enhanced platelet activation, while Hep-PEI/PDA totally suppressed it, regardless of the substrate topography (Fig. 5C). On the other hand, grooves decreased the number of adherent platelets compared to the flat substrates, to different extents on the various surface chemistries. The most remarkable decrease (57%) induced by grooves was observed on the Hep-PEI/PDA, compared to an 40% decrease on the TiO2 and an 22% decrease on the PEI/PDA, suggesting that the immobilized heparin and the substrate topography had synergistic effects on platelet adhesion.

Fig. 4. Relative quantification of heparin, represented by S atomic concentration (at.%), retained on Hep-PEI/PDA surfaces after incubation in PBS (pH 7.4) at 37 °C in a dynamic state for various periods of time.

Quantitative analysis (Fig. 5B) indicated that on the flat substrates, immobilized heparin dramatically decreased the number of adherent platelets by 98% compared to the TiO2, from 27,332 platelets mm2 on the TiO2 to 532 platelets mm2 on the Hep-PEI/PDA. On the grooved substrates, an even greater decrease in the number of adherent platelets (99%) was observed on the Hep-PEI/PDA, which had 230 platelets mm2, compared to the

3.3.2. APTT As shown in Fig. 6, the APTTs of the PEI/PDA were almost identical to those of the TiO2 and untreated plasma, irrespective of the substrate topography. However, the APTTs of the Hep-PEI/PDA were prolonged for about 20 s on the flat substrates and for 23 s on the grooved substrates with respect to the TiO2. This indicates that activation of the intrinsic blood coagulation system is effectively suppressed by immobilized heparin, and the good bioactivity of the immobilized heparin is maintained on both flat and grooved substrates. 3.4. Cooperative control of HUVEC responses 3.4.1. Attachment and proliferation HUVEC attachment and proliferation were sensitive to both immobilized heparin and substrate topography (Fig. 7). On the flat substrates, the cell attachment number was increased on the HepPEI/PDA compared to the PEI/PDA and TiO2 after 2 h of culture. In

Fig. 5. (A) SEM images of adherent platelets on various surfaces after culturing for 2 h in PRP. Quantitative analysis of (B) number of adherent platelets and (C) relative quantification of activated platelets (number of platelets with spreading or fully spread morphology/total number of adherent platelets) on various surfaces, which were analyzed from at least eight random SEM images at 500 magnification. Statistically significant differences are marked as follows: ⁄ vs. TiO2; # vs. Hep-PEI/PDA.

156

Y. Ding et al. / Acta Biomaterialia 15 (2015) 150–163

Similar comparisons and analyses were performed for cell alignment and elongation. The grooves induced dramatic decreases in the average cell alignment angle and minor/major axis ratio of cells compared to the flat substrates, indicating that the strong cell alignment and elongation was induced by substrate grooves regardless of the surface chemistry (Fig. 8D and E). No significant changes in the cell alignment angles or minor/major axis ratio of cells were observed among different surface chemistries over time, whether on the flat or grooved substrates.

Fig. 6. APTT of various surfaces as well as original plasma.

contrast, on the grooved substrates, it was decreased on the HepPEI/PDA, the highest cell attachment number being observed on the TiO2 (Fig. 7A). In agreement with our previous findings, grooves promoted cell attachment compared to the flat substrates for the TiO2. However, such promotion induced by grooves was reduced and almost vanished after PEI/PDA and heparin modification. After 1 day of culture, the cell density showed a similar trend to the cell attachment number among different groups. After 3 days of culture, the immobilized heparin dramatically increased cell density compared to the TiO2 and PEI/PDA; interestingly, the increase was more noticeable on the flat substrates than on the grooved substrates (Fig. 7B). On the other hand, grooves also significantly increased cell densities, which was more evident for the TiO2 and the PEI/PDA than the Hep-PEI/PDA. 3.4.2. Morphology The cellular spreading and cytoskeletal development in HUVECs were primarily determined by the substrate topography, while scarcely being influenced by the surface chemistry (Fig. 8A). The cell projected areas in the different groups were statistically similar (Fig. 8B). Noticeably, the cell projected area consistently decreased from day 1 to day 3. This phenomenon was more visible on the grooved substrates than on the flat ones, indicating a better cell proliferation on the grooved substrates. In addition, immobilized heparin induced a much larger cell area coverage compared to PEI/PDA and TiO2, irrespective of the substrate topography (Fig. 8C). However, the percentage of cell area coverage was comparable between the flat and grooved substrates.

3.4.3. Viability As shown in Fig. 9, cell viability was maintained at similar levels among different surface chemistries after 1 day of culture. However, the Hep-PEI/PDA remarkably enhanced cell viability compared to the TiO2 after 3 days of culture, and this enhancement was greater on the flat substrates than on the grooved ones. On the other hand, considering the topographic effects, grooves significantly increased the cell viability level compared to the flat substrate for the TiO2, though such topographic effects were not significant for either the PEI/PDA or the Hep-PEI/PDA.

3.4.4. Focal adhesion and stress fiber development Immunostaining of vinculin indicated that the sparse focal adhesions (FAs) in ECs induced by the TiO2 were restricted to the periphery of the cells (Fig. 10A). In contrast, the PEI/PDA and Hep-PEI/PDA resulted in densely distributed FAs all over the cell, particularly on the flat substrates. F-actin staining revealed that more prominent stress fibers were distributed throughout the cells on the PEI/PDA and Hep-PEI/PDA compared to the TiO2. In addition, both the FAs and stress fibers were anisotropic and aligned along groove directions on the grooved substrates, in sharp contrast to their random distribution on the flat substrates, regardless of the surface chemistry. Furthermore, the quantification of FA number, area fraction, size and aspect ratio indicated that FA development depended on both surface chemistry and substrate topography (Fig. 10B–D). On both the flat and grooved substrates, a twofold increase in the FA number, a 1.5-fold increase in FA area fraction and a decrease in FA size were observed on the Hep-PEI/PDA compared to the TiO2. However, a similar FA aspect ratio was found among different surface chemistries, suggesting that the surface chemistry had little effect on FA elongation. On the other hand, considering the topographic effects, much fewer FAs and a smaller FA area fraction were induced by grooves, while no apparent differences in FA size and aspect ratio were observed between the grooved and flat substrates, irrespective of the surface chemistry.

Fig. 7. EC attachment and density on various substrates after culturing for 2 h, 1 day and 3 days. Statistically significant differences are marked as follows: ⁄ vs. TiO2; # vs. Hep-PEI/PDA.

Y. Ding et al. / Acta Biomaterialia 15 (2015) 150–163

157

Fig. 8. EC morphology on various substrates. (A) The morphology of ECs revealed by immunofluorescence staining of cytoskeletal F-actin (red) and nuclei (DAPI, blue). Scale bars: 50 lm. (B) Projected area per cell, (C) area coverage by ECs on various substrates, (D) alignment angle and (E) minor/major axis ratio of ECs over time. The projected area per cell, alignment angle and minor/major axis ratio were calculated from at least 120 cells from six different images of each substrate. The area coverage was analyzed from at least 12 images of each substrate. Statistically significant differences are marked as follows: ⁄ vs. TiO2; # vs. Hep-PEI/PDA.

3.4.5. Cell phenotype To analyze the cell phenotype of ECs on various substrates in our culture conditions, two important endothelial cell markers, PECAM-1/CD31 and vWF, were immunostained and observed

(Fig. 11A). The CD31 expression was up-regulated by both the PEI/PDA and the Hep-PEI/PDA, which was more evident on the flat substrates than on the grooved ones (Fig. 11B). Considering the topographic effects, grooves up-regulated the expression levels of

158

Y. Ding et al. / Acta Biomaterialia 15 (2015) 150–163

featuring two topographic geometries, i.e. groove and pillar, and five topographic sizes, ranging from 1 to 50 lm. We then determined the relative cell number increase, representing the chemical effect, induced by the immobilized heparin compared to the TiO2 on different substrate topographies. As shown in Fig. 13, on both grooved and pillared substrates, the relative increase in cell number, i.e. chemical effects, monotonically increased with pattern size. Interestingly, the chemical effects on small-sized patterns were less than on a flat substrate, while large-sized patterns showed greater chemical effects than a flat substrate. As a result, the overall chemical effects induced by immobilized heparin could be finely tuned via the underlying topographies.

4. Discussion

Fig. 9. The viability of ECs on various substrates after culturing for 1 and 3 days. Statistically significant differences are marked as follows: ⁄ vs. TiO2; # vs. Hep-PEI/ PDA.

CD31 for both the TiO2 and PEI/PDA, but not for the Hep-PEI/PDA. In addition, no significant difference in vWF expression was observed among the different groups (Fig. 11C). 3.5. Cooperative control of EC selectivity in co-culture The co-culture of HUVECs and HASMCs was performed on various substrates to further investigate the competitive ability of ECs over SMCs (Figs. 12 and S3). As shown in Fig. 12B, on the flat substrates, both the PEI/PDA and Hep-PEI/PDA substantially suppressed the attachment of SMCs compared to the TiO2, particularly for the Hep-PEI/PDA, although they only induced a small increase in the attachment number of ECs. On the grooved substrates, the Hep-PEI/PDA dramatically increased the attachment number of ECs while decreasing the attachment number of SMCs compared to the TiO2. It is worth noting that the attachment number of ECs was much higher than SMCs on the PEI/PDA and Hep-PEI/PDA irrespective of the substrate topography, suggesting that the PEI/PDA and immobilized heparin endowed the attachment selectivity toward ECs. Considering the topographic effects, it was noticeable that, for the Hep-PEI/PDA, grooves substantially promoted EC attachment compared to the flat substrates. The ratio of the EC attachment number to the SMC attachment number was also calculated (Fig. 12C). The ratio of EC/SMC was 1.0 on both the flat and grooved TiO2. It was increased by 2.5-fold on the PEI/PDA compared to the TiO2 regardless of the substrate topography. Notably, 3.4- and 3.8-fold increases were observed on the flat and grooved Hep-PEI/PDA, respectively, indicating that the immobilized heparin profoundly enhanced the competitive ability of ECs over SMCs. The combination of immobilized heparin and substrate topography seemed to empower more significant enhancement in EC selectivity. 3.6. Tuning chemical effects by varying substrate topographies It was noteworthy that differences in chemical effects on cell behavior were observed on the flat and grooved substrates, suggesting the existence of cooperative interplay between immobilized heparin and substrate topography in regulating cell behavior. Inspired by this observation, we wondered whether it was possible to tune chemical effects by simply varying substrate topographies. We therefore examined the influence of changing substrate topographies on cellular response to the immobilized heparin. For this, heparin was immobilized on a patterned platform

Our focus in this study is on the combined effects of immobilized heparin and substrate topography on blood compatibility and endothelialization, with the aim of elucidating the relative contribution of each cue as well as their interactions. The immobilized heparin substantially inhibited platelet adhesion and activation and prolonged the APTT for more than 20 s compared to the pristine TiO2. These findings corroborate previous reports of the excellent anti-coagulant ability of immobilized heparin [18,26]. However, one study reported that the immobilization of thiol-modified heparin on the surface did not significantly increase anti-coagulant activity, which might be due to the depleted carboxylic acid groups and diminished bioactivity of heparin caused by random thiolation [34]. The excellent anti-coagulant ability we observed in this study is attributed to the good stability and high bioactivity of heparin immobilized via the PEI/ PDA film. We previously found that substrate topography, i.e. 1 lm grooves, also played an important role in platelet adhesion [27]; however, it was not as significant as that of immobilized heparin. In addition, we found that substrate topography had little effect on the APTT. By combining immobilized heparin with substrate topography, the experimental results show that topographic effects, particularly on platelet adhesion, become more potent for the Hep-PEI/PDA than for the pristine TiO2. These findings suggest that immobilized heparin is the primary determinant of blood compatibility in vitro, and synergistic interactions of these two cues occur in modulating platelet adhesion. The rapid re-endothelialization is critical to the success of vascular stents or grafts. However, the effects of heparin on EC growth have long been controversial [23–26]. This study demonstrates that immobilized heparin prominently promotes EC attachment and proliferation, which is consistent with some recent studies [23,24]. Moreover, it was found that the CD31 expression was significantly up-regulated by immobilized heparin. CD31 not only contributes to EC monolayer formation, but is also related to wound healing, neo-vascularization and angiogenesis [35]. The up-regulation of CD31 indicates the better functional development of ECs induced by heparin. The topographic grooves also demonstrate the ability to promote more EC attachment and proliferation compared to flat substrates, which is consistent with our previous findings [27]. When immobilizing heparin on the grooves, however, the EC densities were statistically similar between the flat and grooved Hep-PEI/PDA. In other words, the stimulatory effect of heparin on EC growth was less evident on the grooved substrates than on the flat substrates, which implies that there is a complex interaction between the immobilized heparin and the substrate topography. The findings suggest that EC attachment and proliferation can be cooperatively regulated by both immobilized heparin and substrate topography. Apart from regulating EC attachment and proliferation, our results demonstrate that substrate topography is the determining

Y. Ding et al. / Acta Biomaterialia 15 (2015) 150–163

159

Fig. 10. The focal adhesion (FA) and stress fiber development of ECs on various substrates. (A) Fluorescent images showing distinct development of vinculin (green) and Factin (red) as well as nuclei (DAPI, blue) in cells. Scale bars: 20 lm. (B) Number of FAs per cell, (C) FA area fraction, (D) FA size and (E) FA aspect ratio in cells, which were analyzed from at least 50 cells on each surface. Statistically significant differences are marked as follows: ⁄ vs. TiO2; # vs. Hep-PEI/PDA.

factor in the cellular morphological responses, while immobilized heparin plays a negligible role. In vivo, EC morphology has been linked to atherosclerosis, where well-aligned and elongated ECs appear to be resistant to vascular inflammation while the nonaligned or cuboidal ECs at a branch or bifurcation seem to promote atherosclerosis [36–38]. Importantly, a recent study demonstrated

that the elongation and alignment of ECs induced by micropatterns decreased EC immunogenic gene expression and function compared with cobblestone ECs, indicating that the elongated shape of ECs might be the key contributor to EC-mediated atheroprotection [39]. In addition, the groove pattern was demonstrated to enhance cell retention under shear flow conditions [40]. Our

160

Y. Ding et al. / Acta Biomaterialia 15 (2015) 150–163

Fig. 11. The phenotype of ECs on various substrates. (A) Fluorescent images showing different development of CD31 (green) and vWF (red) as well as nuclei (DAPI, blue) in cells. Scale bars: 50 lm. Quantitation of relative expression of (B) CD31 and (C) vWF (cells with positive CD31 or vWF staining/total cell number), which were analyzed from at least 12 images of each substrate. Statistically significant differences are marked as follows: ⁄ vs. TiO2; # vs. Hep-PEI/PDA.

Fig. 12. Competitive attachment of ECs over SMCs on various substrates during the co-culture. (A) Fluorescent images showing ECs (green) and SMCs (red) attached on various substrates. Scale bars: 50 lm. (B) Cell attachment number of ECs and SMCs, and (C) attachment ratio of ECs/SMCs on various substrates, which were analyzed from at least eight images of each substrate. Statistically significant differences are marked as follows: ⁄ vs. TiO2; # vs. Hep-PEI/PDA.

results show that the alignment and elongation of ECs were manipulated primarily by substrate topography regardless of surface chemistry. This is broadly in agreement with previous reports combining groove patterns with the RGD peptide [41]. Yet the mechanism by which substrate topography regulates cell morphology remains unclear [42]. Previous studies indicated

that FA formation and positioning play an important role in the stimulation of contact guidance [43,44]. Our findings in this study suggest that the FA alignment is more indicative of cellular morphological response to topography, rather than FA number, size or elongation. Our results show that both surface chemistry and substrate topography had profound influences on FA number and

Y. Ding et al. / Acta Biomaterialia 15 (2015) 150–163

Fig. 13. The relative increase in cell number induced by immobilized heparin on variable substrate topographies after culturing HUVECs for 3 days. All the values were normalized to the mean value of flat substrate. G represents groove pattern, P represents pillar pattern, 1–50 represents pattern size (pattern width = spacing) in micrometers.

size. However, the sharp difference between the chemistry and topography is in their ways of manipulating FA alignment: topographic grooves rigidly aligned and orientated FAs along the groove direction regardless of the surface chemistry, while the surface chemistry did not influence FA alignment on either flat or grooved substrates. In addition, FA formation has been shown to be sensitive to the level of mechanical forces that cells generate and exert on substrates [45]. One possible reason why topographic grooves (rather than the surface chemistry) exhibit contact guidance is that the anisotropic positioning of FAs generates an anisotropic force field inside the cell, which will continuously promote realignment of the actin cytoskeleton until the stress fibers are oriented along the same groove direction, and subsequently aligned with the whole cell body. Note that an alternative proposed by Teixeira et al. [46] cannot be excluded at this point, in which anisotropic development of lamellipodia and filopodia generates a similar anisotropic force field inside the cell and, in turn, induces the contact guidance. Nevertheless, our results suggest that the determinant role of substrate topography in cellular alignment and elongation can probably be ascribed to FA alignment and subsequently the anisotropic force profile generated inside the cell. More importantly, in addition to modulating EC behavior in separate culture, the EC/SMC co-culture results demonstrate that immobilized heparin and substrate topography cooperatively regulate competitive attachment and growth of ECs over SMCs. Most previous studies focused on enhancing EC attachment and growth on various modified material surfaces; however, little attention has been paid to the competitive growth of ECs over SMCs in vivo [23,26]. Recent studies demonstrated that the competitive ability of ECs over SMCs played a more important role than the quantity of ECs in the formation of a pure and confluent endothelial layer in vivo, which is essential for preventing restenosis [9,10]. Herein, our study demonstrates for the first time that immobilized heparin can empower the strong competitive ability of ECs over SMCs. Meanwhile, the substrate topography itself seems to have little impact on the competitive ability of ECs over SMCs for TiO2 and PEI/PDA. When combining immobilized heparin with substrate topography, however, a substantially higher EC attachment number was observed, resulting in greater competitive ability of ECs over SMCs than that induced by each cue alone. Interestingly, the enhancement of EC selectivity induced by cooperative action of heparin and topography was more apparent in the co-culture system than in the separate culture system, which was consistent

161

Fig. 14. Model of the cooperative interplay between immobilized heparin and substrate topography in regulating cell behavior.

with a previous study [9]. This finding suggests that SMC may play a role in stimulating EC adhesion in the co-culture system even over a short time. Nevertheless, further study of the interplay between SMCs and ECs in the co-culture system is needed to better understand this observation. The findings of this study have broad implications for the design of novel vascular stents or grafts. For blood-contacting devices, thrombosis is a major complication after implantation, putting emphasis on the importance of designing surfaces with excellent anticoagulant properties. The combination of immobilized heparin and substrate topography produces surfaces with sufficient anticoagulation ability. On the other hand, ensuring the rapid generation of functional endothelium is critical to the long-term success of vascular stents or grafts. The immobilized heparin substantially promoted EC proliferation, ensuring the rapid re-endothelialization, while the topographic grooves stimulated in vivo-like alignment and elongation of ECs, implying an anti-inflammatory and atheroresistant EC phenotype [38,39]. Importantly, the immobilized heparin and topographic grooves cooperatively empowered a stronger competitive ability of ECs over SMCs, favoring the formation of a pure and confluent endothelial layer. Therefore, the combination of immobilized heparin and substrate topography provides a promising strategy to obtain a pure and fully confluent EC monolayer with in vivo-like EC alignment and elongation, which might be more capable of maintaining vascular homeostasis. It should further be noted that the substrate topography seemed to play an essential role in how the cells responded to immobilized heparin. For instance, the enhancement in EC density of immobilized heparin was stimulated to different levels by different substrate topographies (Fig. 13). Given the observation that an individual cell could bridge multiple ridges on small-sized patterns, it is certainly conceivable that different chemical effects on varied substrate topographies likely stem from the differing effective cell–substrate contact area, as suggested to explain the different cellular responses to nanopatterns of various size described previously [47]. Fig. 14 depicts a model to elucidate the cooperative interplay between immobilized heparin and substrate topography. On the flat substrates, cells can contact the entire surface with their whole cell body, and the effective cell–substrate contact area is regarded as the 100% reference point. After heparin immobilization, 100% heparin within the scope of cell coverage would have an effect on stimulating EC growth. In contrast, on the small-sized patterns, e.g. 1 lm grooves, cells do not entirely envelop the grooved surface but bridge multiple ridges, so that the effective cell–substrate contact area is only 50% with respect to the flat substrates. After heparin immobilization, only 50% heparin within the scope of cell coverage would have an effect

162

Y. Ding et al. / Acta Biomaterialia 15 (2015) 150–163

on stimulating EC growth, lessening the chemical effects on the small-patterned substrates compared to those on the flat substrates. Moreover, with increasing pattern gap size, cells could penetrate more into the pattern gaps to maximize the contact area with the substrate, resulting in a larger effective cell–substrate contact area and subsequently greater chemical effects, which may be greater than the chemical effects on a flat substrate, considering the extra contact area provided by the 3-D topography. This model agrees well with our experimental observations, and we furthermore confirmed the cell penetration into grooves via SEM. We indeed observed the cell penetration into pattern gaps, and the degree of cell penetration increased with increasing pattern size (Fig. S4). 5. Conclusions This study demonstrates that immobilized heparin and substrate topography cooperatively contribute to maintaining excellent anti-coagulant activity, stimulating EC attachment and proliferation, and enhancing the competitive ability of ECs over SMCs, although the immobilized heparin seems to play the more prominent roles. The substrate topography is the primary determinant of cellular alignment and elongation, which is ascribed to the stimulatory effect of the topographic groove on the alignment of FAs and stress fibers. Collectively, the experimental results indicate that combining immobilized heparin with substrate topography provides a promising strategy to achieve excellent blood compatibility and accelerate healthy re-endothelialization. In addition, the study indicates that the chemical effects can be readily tuned to different extents by underlying substrate topographies. Acknowledgements This work was financially supported by the Research Grants Council of Hong Kong (FSGRF13EG58), the Program for New Century Excellent Talents in University (NCET-10-0704) and the Sichuan Youth Science Technology Foundation (2011JQ0010). Samples were fabricated and characterized at the Nanoelectronics Fabrication Facility (NFF) and the Materials Characterization and Preparation Facility (MCPF) at the Hong Kong University of Science and Technology (HKUST). Appendix A. Figures with essential colour discrimination Certain figures in this article, particularly Figs. 1, 2, 3, 8, and 10– 14, are difficult to interpret in black and white. The full colour images can be found in the on-line version, at http://dx.doi.org/ 10.1016/j.actbio.2014.12.014. Appendix B. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.actbio.2014.12. 014. References [1] Gimbrone MA, Topper JN, Nagel T, Anderson KR, García-Cardeña G. Endothelial dysfunction, hemodynamic forces, and atherogenesisa. Ann N Y Acad Sci 2000;902:230. [2] Liliensiek SJ, Wood JA, Yong J, Auerbach R, Nealey PF, Murphy CJ. Modulation of human vascular endothelial cell behaviors by nanotopographic cues. Biomaterials 2010;31:5418. [3] Lu J, Rao MP, MacDonald NC, Khang D, Webster TJ. Improved endothelial cell adhesion and proliferation on patterned titanium surfaces with rationally designed, micrometer to nanometer features. Acta Biomater 2008;4:192.

[4] Park J, Bauer S, Schmuki P, Von Der Mark K. Narrow window in nanoscale dependent activation of endothelial cell growth and differentiation on TiO2 nanotube surfaces. Nano Lett 2009;9:3157. [5] Poh CK, Shi Z, Lim TY, Neoh KG, Wang W. The effect of VEGF functionalization of titanium on endothelial cells in vitro. Biomaterials 2010;31:1578. [6] Iuliano DJ, Saavedra SS, Truskey GA. Effect of the conformation and orientation of adsorbed fibronectin on endothelial cell spreading and the strength of adhesion. J Biomed Mater Res 1993;27:1103. [7] Lord MS, Yu W, Cheng B, Simmons A, Poole-Warren L, Whitelock JM. The modulation of platelet and endothelial cell adhesion to vascular graft materials by perlecan. Biomaterials 2009;30:4898. [8] Aoki J, Serruys PW, Van Beusekom H, Ong ATL, McFadden EP, Sianos G, et al. Endothelial progenitor cell capture by stents coated with antibody against CD34: the HEALING-FIM (Healthy Endothelial Accelerated Lining Inhibits Neointimal Growth-First in Man) registry. J Am Coll Cardiol 2005;45:1574. [9] Wei Y, Ji Y, Xiao LL, Lin QK, Xu JP, Ren KF, et al. Surface engineering of cardiovascular stent with endothelial cell selectivity for in vivo reendothelialisation. Biomaterials 2013;34:2588. [10] Yang Z, Xiong K, Qi P, Yang Y, Tu Q, Wang J, et al. Gallic acid tailoring surface functionalities of plasma-polymerized allylamine-coated 316L SS to selectively direct vascular endothelial and smooth muscle cell fate for enhanced endothelialization. ACS Appl Mater Interfaces 2014;6:2647. [11] Huang NF, Lai ES, Ribeiro AJS, Pan S, Pruitt BL, Fuller GG, et al. Spatial patterning of endothelium modulates cell morphology, adhesiveness and transcriptional signature. Biomaterials 2013;34:2928. [12] Thakar RG, Ho F, Huang NF, Liepmann D, Li S. Regulation of vascular smooth muscle cells by micropatterning. Biochem Biophys Res Commun 2003;307:883. [13] Yim EKF, Reano RM, Pang SW, Yee AF, Chen CS, Leong KW. Nanopatterninduced changes in morphology and motility of smooth muscle cells. Biomaterials 2005;26:5405. [14] Li G, Yang P, Qin W, Maitz MF, Zhou S, Huang N. The effect of coimmobilizing heparin and fibronectin on titanium on hemocompatibility and endothelialization. Biomaterials 2011;32:4691. [15] Yang Z, Tu Q, Maitz MF, Zhou S, Wang J, Huang N. Direct thrombin inhibitorbivalirudin functionalized plasma polymerized allylamine coating for improved biocompatibility of vascular devices. Biomaterials 2012;33:7959. [16] Weng Y, Song Q, Zhou Y, Zhang L, Wang J, Chen J, et al. Immobilization of selenocystamine on TiO2 surfaces for in situ catalytic generation of nitric oxide and potential application in intravascular stents. Biomaterials 2010;32:1253. [17] Gong F, Cheng X, Wang S, Zhao Y, Gao Y, Cai H. Heparin-immobilized polymers as non-inflammatory and non-thrombogenic coating materials for arsenic trioxide eluting stents. Acta Biomater 2010;6:534. [18] Yang Z, Wang J, Luo R, Maitz MF, Jing F, Sun H, et al. The covalent immobilization of heparin to pulsed-plasma polymeric allylamine films on 316L stainless steel and the resulting effects on hemocompatibility. Biomaterials 2010;31:2072. [19] Beamish JA, Geyer LC, Haq-Siddiqi NA, Kottke-Marchant K, Marchant RE. The effects of heparin releasing hydrogels on vascular smooth muscle cell phenotype. Biomaterials 2009;30:6286. [20] Lindner V, Olson NE, Clowes AW, Reidy MA. Inhibition of smooth muscle cell proliferation in injured rat arteries: interaction of heparin with basic fibroblast growth factor. J Clin Invest 1992;90:2044. [21] Gu Z, Rolfe BE, Xu ZP, Thomas AC, Campbell JH, Lu GQM. Enhanced effects of low molecular weight heparin intercalated with layered double hydroxide nanoparticles on rat vascular smooth muscle cells. Biomaterials 2010;31:5455. [22] Stewart EM, Liu X, Clark GM, Kapsa RMI, Wallace GG. Inhibition of smooth muscle cell adhesion and proliferation on heparin-doped polypyrrole. Acta Biomater 2012;8:194. [23] Yang Z, Tu Q, Wang J, Huang N. The role of heparin binding surfaces in the direction of endothelial and smooth muscle cell fate and re-endothelialization. Biomaterials 2012;33:6615. [24] Lu Y, Shen L, Gong F, Cui J, Rao J, Chen J, et al. Polycarbonate urethane films modified by heparin to enhance hemocompatibility and endothelialization. Polym Int 2012;61:1433. [25] Azizkhan RG, Azizkhan JC, Zetter BR, Folkman J. Mast cell heparin stimulates migration of capillary endothelial cells in vitro. J Exp Med 1980;152:931. [26] Hoshi RA, Van Lith R, Jen MC, Allen JB, Lapidos KA, Ameer G. The blood and vascular cell compatibility of heparin-modified ePTFE vascular grafts. Biomaterials 2013;34:30. [27] Ding Y, Yang Z, Bi CW, Yang M, Xu SL, Lu X, et al. Directing vascular cellselectivity and hemocompatibility on patterned platforms featuring variable topographic geometry and size. ACS Appl Mater Interfaces 2014;6:12062. [28] Ding Y, Leng Y, Huang N, Yang P, Lu X, Ge X. Effects of microtopographic patterns on platelet adhesion and activation on titanium oxide surfaces. J Biomed Mater Res, Part A 2013;101A:622. [29] Dalsin JL, Lin L, Tosatti S, Vörös J, Textor M, Messersmith PB. Protein resistance of titanium oxide surfaces modified by biologically inspired mPEG-DOPA. Langmuir 2005;21:640. [30] Xu LQ, Yang WJ, Neoh KG, Kang ET, Fu GD. Dopamine-induced reduction and functionalization of graphene oxide nanosheets. Macromolecules 2010;43:8336. [31] Ding Y, Yang ZL, Bi CWC, Yang M, Zhang J, Xu SL. Modulation of protein adsorption, vascular cell selectivity and platelet adhesion by mussel-inspired surface functionalization. J Mater Chem B 2014;2:3819.

Y. Ding et al. / Acta Biomaterialia 15 (2015) 150–163 [32] Nunes GL, Thomas CN, Hanson SR, Barry JJ, King III SB, Scott NA. Inhibition of platelet-dependent thrombosis by local delivery of heparin with a hydrogelcoated balloon. Circulation 1995;92:1697. [33] Silver D, Kapsch DN, Tsoi EKM. Heparin-induced thrombocytopenia, thrombosis, and hemorrhage. Ann Surg 1983;198:301. [34] Joshi P, Schilke KF, Fry A, McGuire J, Bird K. Synthesis and evaluation of heparin immobilized ‘‘side-on’’ to polystyrene microspheres coated with end-group activated polyethylene oxide. Int J Biol Macromol 2010;47:98. [35] Albelda SM, Muller WA, Buck CA, Newman PJ. Molecular and cellular properties of PECAM-1 (endoCAM/CD31): a novel vascular cell–cell adhesion molecule. J Cell Biol 1991;114:1059. [36] Davies PF. Flow-mediated endothelial mechanotransduction. Physiol Rev 1995;75:519. [37] Barakat AI, Lieu DK. Differential responsiveness of vascular endothelial cells to different types of fluid mechanical shear stress. Cell Biochem Biophys 2003;38:323. [38] Flow Cooke JP. Flow, NO, and atherogenesis. Proc Natl Acad Sci USA 2003;100:768. [39] Vartanian KB, Berny MA, McCarty OJT, Hanson SR, Hinds MT. Cytoskeletal structure regulates endothelial cell immunogenicity independent of fluid shear stress. Am J Physiol Cell Physiol 2010;298:C333. [40] Zorlutuna P, Rong Z, Vadgama P, Hasirci V. Influence of nanopatterns on endothelial cell adhesion: enhanced cell retention under shear stress. Acta Biomater 2009;5:2451.

163

[41] Wang PY, Wu TH, Chao PHG, Kuo WH, Wang MJ, Hsu CC. Modulation of cell attachment and collagen production of anterior cruciate ligament cells via submicron grooves/ridges structures with different cell affinity. Biotechnol Bioeng 2013;110:327. [42] Yim EKF, Leong KW. Significance of synthetic nanostructures in dictating cellular response. Nanomed Nanotechnol Biol Med 2005;1:10. [43] Xia N, Thodeti CK, Hunt TP, Xu Q, Ho M, Whitesides GM, et al. Directional control of cell motility through focal adhesion positioning and spatial control of Rac activation. FASEB J 2008;22:1649. [44] Kim DH, Han K, Gupta K, Kwon KW, Suh KY, Levchenko A. Mechanosensitivity of fibroblast cell shape and movement to anisotropic substratum topography gradients. Biomaterials 2009;30:5433. [45] Teo BKK, Wong ST, Lim CK, Kung TYS, Yap CH, Ramagopal Y, et al. Nanotopography modulates mechanotransduction of stem cells and induces differentiation through focal adhesion kinase. ACS Nano 2013;7:4785. [46] Teixeira AI, Abrams GA, Bertics PJ, Murphy CJ, Nealey PF. Epithelial contact guidance on well-defined micro- and nanostructured substrates. J Cell Sci 2003;116:1881. [47] Kim DH, Lipke EA, Kim P, Cheong R, Thompson S, Delannoy M, et al. Nanoscale cues regulate the structure and function of macroscopic cardiac tissue constructs. Proc Natl Acad Sci USA 2010;107:565.