Colloids and Surfaces B: Biointerfaces 123 (2014) 181–190

Contents lists available at ScienceDirect

Colloids and Surfaces B: Biointerfaces journal homepage: www.elsevier.com/locate/colsurfb

Regulation of the migration of endothelial cells by a gradient density of vascular endothelial growth factor Pian Wu a , Ya Fu a,b,∗ , Kaiyong Cai a,∗∗ a Key Laboratory of Biorheological Science and Technology, Ministry of Education, College of Bioengineering, Chongqing University, Chongqing 400044, PR China b College of Chemistry and Chemical Engineering, Chongqing University of Science & Technology, Chongqing 401331, PR China

a r t i c l e

i n f o

Article history: Received 2 July 2014 Received in revised form 3 September 2014 Accepted 5 September 2014 Available online 16 September 2014 Keywords: Gradient biomaterials Endothelial cells VEGF Orientation Cell migration

a b s t r a c t To investigate the effect of the surface gradient density of growth factor on the migration of endothelial cells (ECs), an approach to fabricate a gradient density of vascular endothelial growth factor (VEGF) onto silicon slides has been developed in this study. Our approach involves gradual injection of 3-glycidoxypropyltrimeth oxysilane (GPTMS) and then back filling with 3-triethoxysilylpropyl succinicanhydride (TESPSA) to produce a gradient density of carboxyl groups (–COOH) onto the silicon slides. The –COOH moieties were then activated for the immobilization of VEGF, which leading to a surface gradient density of VEGF. The successful formation of both carboxyl and VEGF gradient densities were confirmed by contact angle measurement, confocal laser scanning microscopy (CLSM) and X-ray photoelectron spectroscopy (XPS), respectively. The treated silicon slide displayed a gradient density of VEGF from 54 to 132 ng/cm2 with a slope of 7.8 ng/cm2 /mm. ECs cultured on the surface gradient density of VEGF demonstrated preferential orientation and an enhanced directional migration behavior. Up to 72% of cells migrated towards the region with high surface density of VEGF. However, the gradient density of VEGF had no significant effect on the cell migration rate. The study provides an alternative to explore chemical-directing cells migration, which is essentially important for understanding cell migration/in-growth behavior for angiogenesis involved in implant technology. © 2014 Elsevier B.V. All rights reserved.

1. Introduction The migration of endothelial cells (ECs) is essentially important for angiogenesis, which is involved in many physiological and pathological processes, including embryonic development, wound healing, tissue regeneration, and tumor growth etc. [1]. It is a complex process regulated by a tight balance between proand anti-angiogenic agents. Cells migration is generally involved with six sequential events: (1) sensing external motile stimuli by cell filopodia; (2) protrusion of lamellipodia and pseudopodia; (3) attachment of the protrusion to the extracellular matrix (ECM) via focal adhesion;(4) stress fiber-mediated contraction of the cell body to allow forward movement; (5) rear release by stress fiber-mediated traction forces; and (6) repeating the adhesion and

∗ Corresponding author at: Key Laboratory of Biorheological Science and Technology, Ministry of Education, College of Bioengineering, Chongqing University, Chongqing 400044, PR China. ∗∗ Corresponding author. Tel.: +86 23 65102507; fax: +86 23 65102877. E-mail addresses: [email protected] (Y. Fu), kaiyong [email protected] (K. Cai). http://dx.doi.org/10.1016/j.colsurfb.2014.09.013 0927-7765/© 2014 Elsevier B.V. All rights reserved.

signaling materials [2]. The process is closely related to the structures of actin, filopodia and lamellipodia of cells. The migration of ECs could be regulated through three major mechanisms, namely chemotaxis (directional migration towards a gradient of soluble chemoattractants), haptotaxis (directional migration towards a gradient of immobilized ligands) and mechanotaxis (directional migration induced by mechanical forces) [3]. ECs migration is an integrated process of these three mechanisms in vivo. Inspired by the knowledge, biomaterials with gradient chemical and/or physical cues could be constructed to spatiotemporally guide cell migration in vitro [4]. The physical cues include gradient stiffness of substrates, gradient pore size and gradient pH value etc. Those cues demonstrated profound influence on cell migration, proliferation, and differentiation [5–8]. For instance, Jorge et al. constructed polyelectrolyte multilayer (PEM) with gradient stiffness by using layer-by-layer assembly technique, which guided the cell adhesion and spreading [9]. The gradient densities of ECM proteins such as laminin, fibronectin (FN), RGD (Arg-Gly-Asp), and collagen as well as other chemicals have been widely used to regulate cell behaviors[9–14]. Oju et al. fabricated three-dimensional gradients of growth factor concentration,

182

P. Wu et al. / Colloids and Surfaces B: Biointerfaces 123 (2014) 181–190

density and substrate stiffness in hydrogels and found that the three-dimensional gradients could regulate the migration and osteochondral differentiation of the encapsulated human mesenchymal stem cells (hMSCs) [15]. The studies proved that gradient biomaterials could affect cells behaviors, such as adhesion, spreading and differentiation. It is well known that growth factors are efficient agents for the regulation of cells behaviors. For example, three different types of growth factors, i.e. basic fibroblast growth factor (bFGF), epidermal growth factor (EGF) and single-chain vascular endothelial growth factor (scVEGF121) were immobilized to collagen gel (scaffold) and enhanced the proliferation and differentiation of human umbilical vein endothelial cells (HUVECs) [16]. Shen et al. found that the immobilization of VEGF was beneficial for promoting the invasion and proliferation of ECs grown onto collagen scaffolds [17]. In other studies, gradient concentration of nerve growth factor (NGF)efficiently induced axonal growth on patterned surfaces and nerve regeneration on poly(␧-caprolactone)-blockpoly(l-lacticacid-co-␧-caprolactone) (PCLA) conduits [18,19]. As a result, a variety of growth factors have been applied in the field of tissue regeneration. However, it is difficult to quantitatively/semiquantitatively elucidate how the immobilized growth factors would affect the cell migration in details (e.g. migration distance, rate, etc.) under a three-dimensional environment. From this point of view, two-dimensional model surfaces provide great convenience for such investigations, however, needs to be further developed. Previously, some studies proved the profound effect of surface gradient densities of growth factors on cell migration. Liu et al. investigated the effect of surface gradient densities of FN, VEGF, as well as combined FN/VEGF on ECs and found that cells adhering to those three surface gradient densities moved faster along the gradient direction than those on uniform surface [20]. Won et al. found that mesenchymal stem cells (MSCs) preferred to migrate towards an SDF-1 gradient [21]. In another study, Gao et al. found that vascular smooth muscle cells (VSMCs) exhibited preferential orientation and enhanced directional migration on bFGF gradient surface. Around 70% of cells migrated towards the region with high density of bFGF on the gradient [4]. Nevertheless, more works should be investigated on how the gradient densities of growth factors affect the migration of ECs, since it is the fundamental basis of the formation of blood vessels (angiogenesis) of microcirculation. Actually, the lack of angiogenesis in tissue-engineered structures is one of the main challenges that hinder the wide clinical applications of tissue-engineered products so far [22]. The purpose of this study was to develop an approach for the fabrication of gradient density of VEGF onto silicon slides and to investigate the influences of surface-anchored VEGF gradient density on the migration behaviors of ECs in vitro. 2. Materials and methods 2.1. Materials Toluidine blue (TBO) was purchased from Aladdin Ltd Co. (Shanghai, China). TESPSA and GPTMS were purchased from J&K Company (Beijing, China).1-(3-Dimethyla-minopropyl)-3ethylcarbodiimide hydrochloride (EDC·HCl), N-hydroxysuccinimide(NHS), recombinant human VEGF, bovine serum albumin (BSA), rhodamine–phalloidin and Hoechst 33258 were obtained from Sigma Chemical Co. (MO, USA). Rabbit monoclonal antibody against human recombinant VEGF and Alex flour 488-labeled goat anti-rabbit IgG were purchased from Beyotime (Shanghai, China). Human VEGF enzyme-linked immunosorbent assay (ELISA) kit was purchased from NeoBioscience Technology Co. (Shenzhen, China). Other chemicals were obtained from Oriental Chemicals

Co Ltd. (Chongqing, China). All chemicals were of analytical grade and used without further treatment if otherwise mentioned. Ultrapure water (>18.2 MG cm, Millipore Milli-Q system) was used in this study. 2.2. Silane treatments GPTMS and TESPSA were dissolved into dry toluene to obtain 4.43 mM and 1.78 mM solutions, respectively. The silicon slides were cut into 10 mm × 10 mm pieces, and then cleaned with toluene, acetone, alcohol and water with ultrasonic treatment each for 15 min. Next, the samples were treated with “piranha” solution (a mixture of 30% hydrogen peroxide and 70% sulfuricacid (v/v) at 80 ◦ C for 2 h [23]. Caution: the “piranha” solution is strongly corrosive and should be handled very carefully! The slides were washed with distilled water for 6 times and then dried under nitrogen flow. Next, the cleaned silicon slides were treated with two different silane coupling agents to form a material ending up with carboxyl groups. To prepare graded carboxyl density samples, however, with uniform carboxyl density on each substrate, different slides were firstly immersed into GPTMS solution at 80 ◦ C for 0, 10, 20, 40, 60 and 80 min, respectively. After washing each sample with toluene for 3 times, the slides were back filled with TESPSA at 80 ◦ C for 8 h (Fig. 1A). The treated slides were then sequentially washed by toluene, ethanol and water each for 6 times. The samples were heated at 60 ◦ C for 4 h to stabilize the silane layer. To produce gradient density of carboxyl on a single silicon substrate, the silicon slide was uprightly placed in a glass vessel with a water bath at 80 ◦ C. Next, GPTMS containing toluene solution was continuously injected into the vessel via a pump at a constant rate of 80 ␮L/min. The inner diameter of the injection syringe was 15.5 mm. A 10 mm slide with gradient density of GPTMS could be generated in 80 min. The backfilling procedure was the same as above. Thus, a gradient density of carboxyl along the slide from bottom (0 mm) to top (10 mm) could be formed for further immobilization of VEGF (Fig. 1C). 2.3. Immobilization of VEGF VEGF was covalently immobilized onto silanized silicon slides with both uniform density and gradient density of carboxyl groups. The slides containing carboxyl groups were first activated with a mixture solution of EDC/NHS (EDC: 4.8 mg/mL; NHS: 12 mg/mL, PBS, pH 5.5) at room temperature for 1 h, followed by rinsing with PBS (pH 7.4) for 3 times [23]. Then, those activated silicon slides were incubated with 1 ␮g/mL of VEGF solution at 4 ◦ C for 24 h, and rinsed with PBS for 3 times to remove the physically absorbed VEGF. The silicon slides were sterilized with 75% (v/v) ethanol for 1 h and washed with PBS before immobilization of VEGF when they were for cell experiments. In order to facilitate the following description, the silicon slides reacted with GPTMS, TESPSA, and VEGF following the above steps were named as GPTMS-Si, TESPSA-Si, and VEGF-Si, respectively. 2.4. Determination of carboxyl density Different densities of carboxyl groups were formed onto silicon slides after reaction with GPTMS for different times (0, 10, 20, 40, 60 and 80 min) and further back filing with TESPSA, respectively. The uniform density of carboxyl groups on each substrate was determined by reaction with TBO, a cationic dye, which could be combined with anions by ionic reaction. Briefly, samples with dimensions of 1 cm × 1 cm were immersed into0.5 mM TBO aqueous solution (pH 10) and reacted at room temperature for 12 h. The samples were rinsed with 0.1 mM NaOH solution for 3 times

P. Wu et al. / Colloids and Surfaces B: Biointerfaces 123 (2014) 181–190

183

Fig. 1. Schematic illustration to show (A) surface conjunction of VEGF via silanization agents of GPTMS/TESPSA; (B) chemical reaction routes for the immobilization of VEGF and (C) the fabrication of VEGF gradient density via an injection approach.

to remove the unbound TBO molecules. The bound TBO on the TESPSA-Si was detached by incubation with 2 mL of 50% acetic acid solution for 10 min. Then the absorbance of the solution was measured with a microplatereader (Bio-Rad 680, USA) at a wavelength of 630 nm. The amount of the carboxyl groups was calculated by referring to a calibration curve. It was obtained by measuring the

solution absorbance with different concentrations of TBO dissolved into 50% acetic acid solution at the same condition. The equation of the calibration curve was y = 0.023x (R2 = 0.999). The calculation of –COOH density was based on the assumption that TBO had combined with equivalent of carboxyl groups [24]. The mean value of four replicates was used as the final result.

184

P. Wu et al. / Colloids and Surfaces B: Biointerfaces 123 (2014) 181–190

2.5. Contact angle measurement Static water contact angle measurement was performed with a Model 200 video-based optical system (Future Scientific Ltd. Co., Taiwan, China). Different GPTMS-Si, TESPSA-Si, and VEGFSi substrates with uniform density were measured, respectively. Four replicates were measured for each group and two different areas were measured for each sample. Each water drop (5 ␮L) was deposited to the sample surface and kept for 15 s before measurement. Then, an image of the drop was recorded by a camera and analyzed with the software supplied by the manufacturer. All measurements were conducted at room temperature. 2.6. Measurement of surface density of VEGF The uniform surface density of bound VEGF to Si-slides was measured by a VEGF ELISA kit according to a previous study [25]. After immobilization of VEGF onto the slides; the soaking and rinsing solutions were collected and measured by an ELISA reader. The immobilized amount of VEGF onto the VEGF-Si was determined by subtracting the amount of VEGF remaining in the soaking and rinsing solutions from the original amount of VEGF [26]. The mean value of four replicates was used as the final result. 2.7. Immunofluorescence observation

with 4% glutaraldehyde at 4 ◦ C for 20 min, followed by rinsing with PBS for 3 times. The cells were further permeabilized with 0.2% Triton X-100/PBS at 4 ◦ C for 2 min. After washing with PBS for 3 times, the cells were stained with 5 U/mL of rhodamine–phalloidin at 4 ◦ C overnight. Next, the treated cells were counterstained with 10 ␮g/mL of Hoechst 33258 at room temperature for 5 min. The stained samples were finally observed with CLSM (Leica DMI 6000, Germany).

2.11. Cell migration analysis ECs were seeded onto VEGF-Si at an initial density of 5 × 103 cells/cm2 in order to minimize the influence of potential contacts between cells and cells. After culture for 12 h, the cells migrations were monitored using a live cell station equipped with an incubation chamber (37 ◦ C and 5% CO2 humidified atmosphere) and a time-lapse digital camera [29]. Images for cell migration were taken every 15 min within 6.5 h. The ECs trajectories were reconstructed according to the center positions of individual cell over the observation period. 30 cells were recorded for each sample. Various parameters such as migration trajectories, net displacement, total migration distance, chemotactic index and percentage of cells move towards gradient regarding cell migration were calculated according to previous studies [11,30].

To qualitatively characterize the relative density of VEGF on different positions of gradient VEGF surface, immunofluorescence staining was performed. The slides were firstly incubated with 1% BSA/PBS solution for 1 h to block the non-specific interactions, and then treated with a rabbit monoclonal antibody against human recombinant VEGF at 37 ◦ C for 1 h. After washing with PBS for 3 times, the slides were incubated with Alex flour 488-labeled goat anti-rabbit IgG at 37 ◦ C for 1.5 h in dark. The samples were then washed with PBS for 3 times before observation. Fluorescence images at the position of 0, 2.5, 5, 7.5, and 10 mm from the bottom of the slides were recorded with confocal laser scanning microscopy (CLSM, Leica DMI 6000, Germany). 2.8. XPS The chemical compositions at different positions of the VEGF gradient surfaces were detected by XPS. It was performed by using a PHI-5400model system (PerkinElmer, USA) with Mg K␣ X-ray source (1253.6 eV). The hydrocarbon peak maximum in the C 1s spectra was set as 285.0 eV to refer the binding energy scales for samples. The detected positions were chosen consistent with immunofluorescence observation. Data were analyzed with Avantage software (version 4.46, Thermo Fisher Scientific Inc., USA). 2.9. Cell culture Human endothelial cells were kindly provided by the Third Military Medical University (Chongqing, China). Cells were cultured according to previous reports [9,11]. Briefly, cells were cultured in polystyrene flasks with medium of RPMI1640(Hyclone), supplemented with 10% (v/v) fetal bovine serum (FBS, Gibco, USA), 100 U/mL penicillin and 100 ␮g/mL streptomycin at 37 ◦ Cunder 5% CO2 atmosphere. Cell culture media was changed every 2 days. 2.10. Cell staining The F-actin and nuclei of cells were stained according to previous studies [27,28]. Briefly, ECs were seeded onto VEGF-Si at an initial density of 5 × 103 cells/cm2 . After culture for 2 days, the cell layers were rinsed with PBS for 3 times. The cells were then fixed

Fig. 2. Physical property characterization: (A) carboxyl density on TESPSA-Si substrates; (B) contact angles of GPTMS, TESPSA and VEGF conjugated silicon slides with different reaction times of GPTMS. Error bars represent mean ± SD for n = 4.

P. Wu et al. / Colloids and Surfaces B: Biointerfaces 123 (2014) 181–190

185

Fig. 3. VEGF gradient density characterization: (A) gradient density of VEGF on VEGF-Si with different GPTMS reaction times. Error bars represent means ± SD for n = 4; (B) fluorescence images of (a) native silicon slides and (b) gradient density of VEGF at different positions with immunochemical staining, scale bar: 50 ␮m;and (C) fluorescence intensity analysis based on image of (b).

186

P. Wu et al. / Colloids and Surfaces B: Biointerfaces 123 (2014) 181–190

2.12. Statistical analysis

O1s Si2s

All results were reported as mean ± standard deviation (SD) and analyzed using t-test and one-way analysis of variance (ANOVA) for differences. The statistical analysis was performed with the software Origin (version 7.5). The significant levels were set as *p < 0.05 and **p < 0.01.

C1s (A) N1s

Intensity(a.u.)

(B)

3. Results and discussion

(C) (D) (E)

3.1. Relationship between GPTMS reaction time and carboxyl density The fabrication methodology of this study was illustrated in Fig. 1A. It was involved in sequential self-assembly monolayer of GPTMS and TESPSA onto silicon slides, thus resulting in gradient density of carboxyl for further conjugation of VEGF. Briefly, after treatment with “piranha” solution, plenty of hydroxyl groups (–OH) were formed onto the surfaces of silicon slides. The hydroxyl groups were firstly reacted with GPTMS for desired time intervals, resulting in different densities of non-reacted –OH, which were further reacted with TESPSA to generate different densities of –COOH (Fig. 1B) [31]. That is to say, a higher –COOH density could be obtained with shorter GPTMS reaction time, since more –OH groups left to be further reacted with TESPSA. The density of –COOH onto the slides surfaces was quantified by a TBO method. The cationic in TBO combined with –COOH by ionic reaction, the density of –COOH could be calculated by measuring the concentration of TBO. The calculation was based on the assumption that TBO had combined with equivalent of –COOH. Fig. 2A shows the relationship between –COOH density and GPTMS reaction time. It shows that the density of –COOH was decreased as time extended. The reason was that with extending GPTMS reaction time, less –OH groups left onto the slides surfaces for the generation of –COOH groups. The maximum density of the –COOH was approximately 5 × 10−10 mol/cm2 , which was about 5 times higher than the minimum density. It was consistent with a previous study [32]. The result suggests that –COOH density could be tuned by controlling the GPTMS reaction time. It in turn provided feasibility to generate tunable VEGF density on the surfaces of silicon slides via EDC/NHS coupling agents to mediate the reactions between –COOH groups and –NH2 groups within VEGF molecules (Fig. 1B). One of the important advantages of the present strategy was that the density of VEGF was strictly controlled by the reaction time of GPTMS. To investigate the wettable properties of GPTMS-Si, TESPSA-Si and VEGF-Si substrates, water contact angle measurement was performed. Water contact angle measurement is one of the techniques to reveal the wettable property (hydrophobicity or hydrophilicity) of a measured sample. Fig. 2B shows the wettability variation of different slides as a function of the reaction times of GPTMS, TESPSA and VEGF. The longer the reaction time of GPTMS the higher contact angle was. It could be interpreted that less –OH was left when increasing the GPTMS reaction time, leading to the increase of contact angles. After reaction with TESPSA, –OH groups were converted into –COOH groups, the contact angles of TESPSA-Si substrates were lower than GPTMS-Si substrates at different time points. The phenomenon could be interpreted as follows: on the one hand, the ending groups of –COOH were inherently more hydrophilic than those of –OH groups [31], leading to the decreased contact angles of GPTMS-Si substrates; on the other hand, –OH groups could not be completely converted into –COOH groups, resulting in the higher contact angles of TESPSA-Si substrates than those of GPTMS-Si at time point of 0 min. The contact angles of TESPSA-Si substrates steadily increased along the reaction time, implying the decrease of –COOH density along the process. VEGF-Si substrates

(F)

0

100

200

300

400

500

600

Binding energy (eV) Fig. 4. XPS spectra of native silicon and VEGF-Si substrates at different positions: (A) native silicon; (B) 0 mm; (C) 2.5 mm; (D) 5.0 mm; (E) 7.5 mm; and (F) 10 mm. Table 1 The chemical compositions of native silicon and VEGF-Si substrates at different positions detected by XPS. Substrates

O (at %)

C (at.%)

Si (at.%)

N (at.%)

A B C D E F

36.36 24.99 27.97 27.45 28.52 29.19

13.58 52.77 49.54 49.41 58.47 53.31

50.05 18.48 18.14 18.38 7.77 11.82

0 3.76 4.35 4.76 5.24 5.68

displayed similar water contact angles to those of TESPSA-Si substrates at different time of points, which might contribute to the similar hydrophilic nature of carboxyl and VEGF molecules [33]. This result indicates that gradient density of VEGF could be formed on the surfaces of silicon slides with this method. 3.2. Characterization of VEGF gradient density To determine the VEGF gradient density corresponding with different GPTMS reaction times on VEGF-Si, the quantity of VEGF in the loading solution and the combined washing solution was measured using an ELISA kit. Fig. 3A shows the relationship between VEGF density and reaction time of GRTMS, suggesting a linear equation of Y = −0.995X + 133 (R2 = 0.995). The result matches well with the results of –COOH density and contact angle measurement (Fig. 2). It implies that a linear gradient density of VEGF could be formed onto the surface of a single silicon slide. From the above results, we know that the VEGF density could be indirectly controlled by the reaction time of GPTMS. Next, we fabricated a gradient density of VEGF on a single slide by controlling the reaction time of GPTMS via an injection method. GPTMS solution was slowly injected into the vessel with a pump (Fig. 1C). Thus, the exposure time of the slides at the bottom to GPTMS solution was longer than that of the top of the slide. Meanwhile, the reaction was a time dependent process. The longer the expose time the lower the –COOH density and VEGF density was produced. Therefore, a gradient density of VEGF along the length of the slide was formed with the highest density at top and the lowest density at bottom. To verify the formation of VEGF gradient density, fluorescence microscopy and XPS were employed both for qualitative and quantitative analysis, respectively. The immobilized VEGF was conjugated with VEGF antibody and then stained with Alex Flour 488 for fluorescence visualization. Fig. 3B displays the fluorescence images of the immobilized VEGF along the slide. In this study, we

P. Wu et al. / Colloids and Surfaces B: Biointerfaces 123 (2014) 181–190

187

Fig. 5. Cell morphology and distribution characterization: (A) CLSM images of ECs cultured onto: (a–c) surfaces with uniform VEGF density and (d–f) surfaces with gradient density of VEGF. (a and d) Surfaces with low VEGF density; (b and e) Surfaces with middle VEGF density; (c and f) surfaces with high VEGF density. The arrows indicate the membrane protrusion and lamellipodia of cells, scale bar: 20 ␮m; and (B) angular distribution of ECs orientation cultured onto VEGF gradient surfaces(black) and uniform VEGF surface (white). Average data was obtained from 150 cells for both uniform slides and gradient slides.

evenly selected 5 positions at 0, 2.5, 5.0, 7.5 and 10 mm along the slide for observation. No obvious physical adoption of VEGF on native silicon slides was observed (Fig. 3B(a)). For gradient samples, the fluorescence images gradually became bright from the bottom end to the top end of the slides (Fig. 3B(b), left to right). It suggests that higher density of VEGF was immobilized on the top end than that of bottom end of the slides (Fig. 1C). Fig. 3C shows the semi-quantitative analysis of fluorescence intensity variance along

the gradient surface based on the images of Fig. 3B(b). The result suggests that gradient density of VEGF was successfully fabricated onto the surfaces of the silicon slides. XPS measurements were further performed to indirectly investigate the formation of VEGF gradient density (Fig. 4). The measured points were consistent with those for fluorescence observations. TESPSA-Si substrates displayed typical signal for Si, C and O elements, corresponding with binding energies of around 100 eV,

188

P. Wu et al. / Colloids and Surfaces B: Biointerfaces 123 (2014) 181–190

Fig. 6. Representative migration trajectories of ECs cultured onto: (A–C) surfaces with uniform VEGF density and (D–F) surfaces with gradient density of VEGF. (A and D) Surfaces with low VEGF density; (B and E) surfaces with middle VEGF density; and (C and F) surfaces with high VEGF density. The initial position was defined as the 0 point in the X–Y plane (n = 30).

285 eV and 532 eV, respectively. After immobilization of VEGF, additional peak around 400 eV was observed from all other samples (Fig. 4B–F), which was assigned to N element. The presence of N element was contributed to the large amount of amine and amide groups derived from the VEGF molecules, since no N element was presented at the surface of TESPSA-Si (Fig. 1B). Thus, the appearance of N element in these samples would indirectly reveal that VEGF was successfully immobilized onto the surfaces of TESPSA-Si substrates. Moreover, the relative atomic percentage of O, N, Si and C elements was quantified in this study. A gradual decrease of the relative N atomic percentage was observed along the slide from top to bottom (Table 1). The result indirectly indicates that a gradient density of VEGF was formed onto the surface of the silicon slide. 3.3. Cytoskeleton morphology Since the gradient density of VEGF was gradually distributed, for analysis convenience, the gradient area on VEGF-Si slide was evenly divided into three groups based on their durations of

exposure to GPTMS solution, i.e., reaction time of 80–53 min, 53–26 min, 26–0 min, corresponding with the distance from bottom of the slide of 0–3.3 mm, 3.3–6.6 mm and 6.6–10 mm, respectively. According to the relatively linear relationship between GPTMS reaction time and VEGF density (Fig. 3A), the VEGF densities of three groups were roughly estimated as ranging from 54 to 80 ng/cm2 , 80 to 106 ng/cm2 and 106 to 132 ng/cm2 , respectively. The gradient area with low VEGF density, middle VEGF density and high VEGF density were donated as L, M and H groups in the following text, respectively. The control slides with uniform density of VEGF throughout the surface was denoted as UL, UM and UH groups, respectively. Cell migration is a complex process involving sequential steps of cell adhesion, polarization, contraction and forward movement [34]. Actin is the major structure in the microfilaments whose polymerization is a key step that regulates cell movement [35]. Therefore, fluorescent staining of F-actin and nuclei of ECs were performed to visualize the cells responses to the immobilized gradient density of VEGF at different areas. Fig. 5A shows the

P. Wu et al. / Colloids and Surfaces B: Biointerfaces 123 (2014) 181–190

189

Fig. 7. Motility of ECs cultured onto surfaces with uniform and gradient density of VEGF: (A) total cell migration distance; (B) net displacement; (C) CI of ECs; and (D) percentage of ECs moving towards gradient. Error bars represent mean ± SD for n = 30. *p < 0.05, **p < 0.01.

representative fluorescent images of nuclei and cytoskeleton actin networks of ECs grown onto uniform and gradient surfaces. Cells adhered to samples with uniform VEGF surfaces (high, middle or low density) displayed random cytoskeleton organization with various orientations (Fig. 5A(a–c), arrows). In comparison, most cells cultured onto gradient surfaces (high, middle and low VEGF density) demonstrated an actin orientation tendency along gradient direction of increasing VEGF density (Fig. 5A(d–f), arrows), respectively. About 42.6% of ECs on gradient density surfaces were aligned within 20◦ along the gradient direction, whereas around 73.3% of cells were aligned within 40◦ parallel to the gradient direction (Fig. 5B). In contrast, endothelial cells cultured on uniform density surfaces randomly distributed in any directions. The results indicate that the gradient density of VEGF was the direct cue that mediated the cell orientation. 3.4. Cell migration on gradient density of VEGF In order to avoid cell-cell interactions, the ECs were seeded at a low density. In this case, the mobility of cells could be dominated by the cell-substrate interactions. However, it depends on the surface properties of the substrates and the type of cells [14]. The ECs were monitored for 6.5 h with a live cell station. The cells migration trajectories on various surfaces were rebuilt by analyzing the sequential images. For the ease of discussion, X-axis refers to the direction of gradient density of VEGF, while Y-axis refers to the direction perpendicular to gradient density of VEGF. The ECs moved randomly without a preferential direction on the uniform VEGF surfaces regardless of their densities (Fig. 6A–C). This is reasonable since the cells would not directionally polarize in a uniform environment having no directional chemical and/or physical cues. In contrast, most cells moved towards −X direction on the

VEGF gradient, i.e. cells moved towards the higher VEGF density (Fig. 6D–F). To elucidate the influence of VEGF gradient density on cell migration, several cells mobility parameters were calculated according to the migration trajectories, including total migration distance, net displacement, chemotactic index (CI) and the percentage of cells moving towards gradient [36]. At the same culture time, cell total migration distance reflected the migration rate. Total migration distance was calculated by counting up the straight-line distance that a cell migrated between two sequential images. As shown in Fig. 7A, the VEGF gradients (high, middle or low density) had no significant influence on cell migration distance (migration rate) when compared to those of the corresponding uniform surfaces. The result was consistent with a previous study [20]. Next, we measured the net cell displacement (Fig. 7B). The net cell displacement was defined as the straight distance of cell displacement between initial and final points, reflecting the potential of cell movement. ECs cultured on VEGF gradient density surfaces displayed average higher net displacement than those of corresponding uniform density samples. Cells adhered to the gradient area with middle VEGF density displayed significant higher (p < 0.01) net displacement than those grown onto samples with uniform VEGF density. However, there were no significant difference was observed in the groups with either high or low VEGF density. Subsequently, we characterized CI of ECs grown onto different samples. CI is the net displacement of cell towards gradient divided by its total distance. CI reflects the cell migration tendency towards the gradient direction; value of 1 means a cell moves in parallel with gradient and value of 0 means a cell moves perpendicular to gradient. As shown in Fig. 7C, cells adhered to slides with VEGF gradient density displayed an average higher values of CI compared with

190

P. Wu et al. / Colloids and Surfaces B: Biointerfaces 123 (2014) 181–190

those of the corresponding uniform samples. Cells adhered to areas with middle and low gradient density of VEGF displayed significant higher (p < 0.01 or p < 0.05) CI values than those of uniform slides, respectively. Surprisingly, the CI of ECs was not obviously enhanced by high gradient density of VEGF. It could be attributed to the fact that high VEGF density overwhelmed the relatively small impact of VEGF density difference (∼7.8 ng/cm2 /mm) at this range. The result of CI is consistent with that of net cell displacement. Finally, we quantified the cells percentage moved towards various gradient densities of VEGF. Around 88.9%, 83.3% and 72.2% of ECs moved towards the VEGF gradients with high, middle and low density, respectively (Fig. 7D). The result reveals strong preferential orientation tendency towards the gradient density of VEGF. Taken together, the migration rates of ECs adhered to gradient densities of VEGF (high, middle or low density) and uniform surfaces had no difference, and the average net displacements of ECs on gradients (especially middle density) were higher than those of corresponding uniform density samples. It indicated that the migration trajectory were more zigzag on uniform surfaces than those of gradient density surfaces, in particular of VEGF gradient with middle density, leading to chemotaxis effect [1,2]. In this work, we found that ECs cultured on gradient density of VEGF promoted cells to move towards the gradient. ECs adhering to gradient area with middle VEGF density displayed the highest motility when compared with other gradient areas or uniform samples. This result was consistent with previous studies [37,38]. Barkefors et al. found that the gradient of VEGF efficiently induced chemotaxis of ECs [37]. Odedra et al. immobilized VEGF onto a porous scaffold and found that the VEGF gradient promoted the migration of ECs [38]. 4. Conclusion In this study, the gradient density of VEGF was successfully fabricated onto silicon slides by using an injection method. The VEGF density gradually increased from 54 to 132 ng/cm2 with a slope of 7.8 ng/cm2 /mm. The ECs exhibited preferential orientation and an enhanced migration potential on the surfaces with gradient density of VEGF, in particular of middle VEGF density. Up to 72% cells migrated towards high density of VEGF gradient. However, the gradient density of VEGF had no effect on the cell migration rate. Acknowledgments This work was financially supported by Natural Science Foundation of China (31170923 and 51173216), Natural Science Foundation of Chongqing Municipal Government (CSTC2013jjB50004, CSTC2011JJJQ10004), National Key Technology R&D Program of the Ministry of Science and Technology (2012BAI18B04)

and Fundamental Research Funds for the Central Universities (CQDXWL-2013-Z002). References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14]

[15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38]

S. Rousseau, F. Houle, J. Huot, Trends Cardiovasc. Med. 10 (2000) 321. L. Lamalice, F.L. Boeuf, J. Huot, Circ. Res. 100 (2007) 782. S. Li, N.F. Huang, S. Hsu, J. Cell. Biochem. 96 (2005) 1110. J.D. Wu, Z.W. Mao, L.L. Han, Y.Z. Zhao, J.B. Xi, C.Y. Gao, Colloid Surf. B 117 (2014) 290. I. Hopp, A. Michelmore, L.E. Smith, D.E. Robinson, A. Bachuka, A. Mierczynska, K. Vasilev, Biomaterials 34 (2013) 5070. K. Kreppenhofer, J. Li, R. Segura, L. Popp, M. Rossi, P. Tzvetkova, B. Luy, C.J. Kahler, A.E. Guber, P.A. Levkin, Langmuir 29 (2013) 3797. R.K. Paradise, M.J. Whitfield, D.A. Lauffenburger, K.J. Van Vliet, Exp. Cell Res. 319 (2013) 487. R. Sunyer, A.J. Jin, R. Nossal, D.L. Sackett, PLoS ONE 7 (2012) e46107. ˇ Selimovic, ´ T. Boudou, A. Khadehosseini, C. Picart, J. Almodóvar, T. Crouzier, S. Lab Chip 13 (2013) 1562. K.Y. Cai, H.D. Dong, C. Chen, L. Yang, K.D. Jandt, L.H. Deng, Colloid Surf. B 72 (2009) 230. A. Tiwari, J.J. Jung, S.M. Inamdar, C.O. Brown, A. Goel, A. Choudhury, J. Biol. Chem. 286 (2011) 36749. K.Y. Cai, T.T. Kong, L. Wang, P. Liu, W.H. Yang, C. Chen, Colloid Surf. B 79 (2010) 291. X.T. Shi, S. Ostrovidov, Y.W. Shu, X.B. Liang, K. Nakajima, H.K. Wu, A. Khademhosseini, Langmuir 30 (2014) 832. C.R. de Almodovar, C. Coulon, P.A. Salin, E. Knevels, N. Chounlamountri, K. Poesen, K. Hermans, D. Lambrechts, K.V. Geyte, J. Dhondt, T. Dresselaers, J. Renaud, J. Aragones, S. Zacchigna, I. Geudens, D. Gall, S. Stroobants, M. Mutin, K. Dassonville, E. Storkebaum, B.F. Jordan, U. Eriksson, L. Moons, R. D’Hooge, J.J. Haigh, M.F. Belin, S. Schiffmann, P.V. Hecke, B. Gallez, S. Vinckier, A. Chedotal, J. Honnorat, N. Thomasset, P. Carmeliet, C. Meissirel, J. Neurosci. 30 (2010) 15052. O. Jeon, D.S. Alt, S.W. Linderman, E. Alsberg, Adv. Mater. 25 (2013) 6366. Y. Assal, M. Mie, E. Kobatake, Biomaterials 34 (2013) 3315. Y.H. Shen, M.S. Shoichet, M. Radisic, Acta Biomater. 4 (2008) 477. B. Joddar, A.T. Guy, H. Kamiguchi, Y. Ito, Biomaterials 34 (2013) 9593. S. Tang, J. Zhu, Y. u, A.P. Xiang, M.H. Jiang, Biomaterials 34 (2013) 7086. L.Y. Liu, B.D. Ratner, E.H. Sage, S.Y. Jiang, Langmuir 23 (2007) 11168. Y.W. Won, A.N. Patel, D.A. Bull, Biomaterials 35 (2014) 5627. L.G. Griffith, G. Naughton, Science 295 (2002) 1009. J.D. Wu, Z.W. Mao, C.Y. Gao, Biomaterials 33 (2012) 810. Y.X. Liu, T. He, C.Y. Gao, Colloid Surf. B 46 (2005) 117. W.X. Shen, K.Y. Cai, Z.X. Yang, Y. Yan, W.H. Yang, P. Liu, Colloids Surf. B 94 (2012) 347. C.K. Poh, Z.L. Shi, T.Y. Lim, K.G. Neoh, W. Wang, Biomaterials 31 (2010) 1578. K.B. Vartanian, S.J. Kirkpatrick, S.R. Hanson, M.T. Hinds, Biochem. Biophys. Res. Commun. 371 (2008) 787. M. Lai, K.Y. Cai, Y. Hu, X.F. Yang, Q. Liu, Colloid Surf. B 97 (2012) 211. T.C. Ren, Z.W. Mao, J. Guo, C.Y. Gao, Langmuir 29 (2013) 6386. S.J. Wang, W. Saadi, F. Lin, N.C. Minh-Canh, J.N. Li, Exp. Cell Res. 300 (2004) 180. G.K. Toworfe, R.J. Composto, I.M. Shapiro, P. Ducheyne, Biomaterials 27 (2006) 631. P. Ye, R.B. Wan, X.P. Wang, J. Mol. Catal. B: Enzym. 61 (2009) 296. X.F. Hu, K.G. Neoh, J.Y. Zhang, E.T. Kang, W. Wang, Biomaterials 33 (2012) 8082. C.C. Zimmer, L.F. Shi, Y.P. Shih, L.J. Ren, L. Jin, S. Lo, G. Liu, Sci. China Chem. 55 (2012) 1922. D.A. Lauffenburger, A.F. Horwitz, Cell 84 (1996) 359. B. Mosadegh, W. Saadi, S.J. Wang, N.L. Jeon, Biotechnol. Bioeng. 100 (2008) 1205. I. Barkefors, S.L. Jan, L. Jakobsson, E. Hejll, G. Carlson, H. Johansson, J. Jarvius, J.W. Park, N.L. Jeon, J. Kreuger, J. Biol. Chem. 283 (2008) 13905. D. Oderdra, L.L.Y. Chiu, M. Shoichet, M. Radisic, Acta Biomater. 7 (2011) 3027.