Article pubs.acs.org/Biomac
Chondroitin Sulfate Coatings Display Low Platelet but High Endothelial Cell Adhesive Properties Favorable for Vascular Implants Pradeep K. Thalla,†,‡ Hicham Fadlallah,‡,§ Benoit Liberelle,∥ Pauline Lequoy,†,‡ Gregory De Crescenzo,∥ Yahye Merhi,§ and Sophie Lerouge*,†,‡ †
Laboratory of Endovascular Biomaterials (LBeV), Centre hospitalier de l’Université de Montréal (CRCHUM), 900 Saint Denis, Tour Viger, 11th Floor, Montreal, QC, H2X 0A9, Canada ‡ Department of Mechanical Engineering, École de technologie supérieure (ÉTS), 1100 Boulevard Notre-Dame West, Montreal, QC, H3C 1K3, Canada § Laboratory of Thrombosis and Haemostasis, Montreal Heart Institute, 5000 Belanger, Montreal, QC, H1T 1C8, Canada ∥ Department of Chemical Engineering, École Polytechnique de Montréal, P.O. Box 6079, Succ. Centre-Ville, Montreal, QC, H3C 3A7, Canada S Supporting Information *
ABSTRACT: This study highlights the advantages of chondroitin sulfate (CS) as a sublayer combining selective low-fouling properties, low-platelet adhesion and pro-adhesive properties on endothelial cells, making CS promising for vascular graft applications. These properties were evaluated by comparing CS with well-known low-fouling coatings such as poly(ethylene glycol) (PEG) and carboxymethylated dextran (CMD), which were covalently grafted on primary amine-rich plasma polymerized (LP) films. Protein adsorption studies by quartz crystal microbalance with dissipation monitoring (QCM-D) and fluorescence measurements showed that CS is as effective as PEG in reducing fibrinogen adsorption (∼90% reduction). CS also largely reduced adsorption of bovine serum albumin (BSA) as well as fetal bovine serum (FBS) but to a lower extent than PEG and CMD surfaces (72% vs 85% for BSA and 66% vs 89% for FBS). Whole blood perfusion assays indicated that, while LP surfaces were highly reactive with platelets, PEG, CMD, and CS grafted surfaces drastically decreased platelet adhesion and activation to levels significantly lower than polyethylene terephthalate (PET) surfaces. Finally, while human umbilical vein endothelial cell (HUVEC) adhesion and growth were found to be very limited on PEG and CMD, they were significantly increased on CS compared to that on bare PET and reached similar values as those for tissue culture polystyrene positive controls. Interestingly, HUVEC retention during perfusion with blood was found to be excellent on CS but poor on PET. Overall, our results suggest that the CS surface has the advantage of promoting HUVEC growth and resistance to flow-induced shear stress while preventing fibrinogen and platelet attachment. Such a nonthrombogenic but endothelial-cell adhesive surface is thus promising to limit vascular graft occlusion.
1. INTRODUCTION Biomaterials used in synthetic vascular grafts, mainly polyethylene terephthalate (PET) and polytetrafluoroethylene (PTFE), fail to promote endothelialization and induce thrombosis and inflammation due to platelet and neutrophil activation.1 This limits their use to large-diameter vessels because of the high risk of thrombosis in vascular grafts smaller than 6 mm. Adsorption of blood proteins, especially the clotting enzymes and fibrinogen, plays an important role in this material-induced clotting. Platelet adhesion and activation that follow protein adsorption also contribute to surface-induced thrombosis. Therefore, protein-resistant coatings such as those made of poly(ethylene glycol) (PEG) have been extensively investigated since they can reduce inflammatory response, leukocyte activation, platelet adhesion, and the risk of thrombus formation.2,3 The low-fouling properties of PEG coatings are © 2014 American Chemical Society
attributed to steric repulsion and a highly hydrated surface layer.2,4−6 Despite very good protein resistance in vitro, their benefit in vivo is limited, for several reasons. First, their stability was shown to be poor, which might be due to the rapid oxidization of PEG in the presence of oxygen and transition metal ions.7−9 Second, these coatings do not inhibit clotting actively per se. More generally, they do not allow the formation of a confluent endothelium lining which is the sole long-term nonthrombogenic surface that is known to date. Dextran-based coatings were also investigated as an alternative to PEG coatings since their low-fouling properties were demonstrated to be as good as PEG.10,11 However, they face the same Received: March 10, 2014 Revised: June 11, 2014 Published: June 13, 2014 2512
dx.doi.org/10.1021/bm5003762 | Biomacromolecules 2014, 15, 2512−2520
Biomacromolecules
Article
prevents platelet adhesion and activation while favoring a stable endothelial lining. It thus presents many advantages for vascular grafts.
limitations as PEG. More generally, it is now quite accepted that bioactive coatings that actively inhibit thrombosis (heparin, antithrombin−heparin complex (ATH) etc.) with very positive short-term effects, also face durability and specificity issues. Thus, although heparin coatings reduce thrombus formation, both in vitro12 and in vivo,13 they were also shown to interact with platelets and macrophages14 and to induce neointima formation in vivo.15 The sole completely antithrombogenic surface that is known so far is a stable confluent endothelium lining. Since, in vivo, such a lining does not form spontaneously at the surface of a vascular implant in humans, in vitro preseeding of endothelial cells (ECs) or progenitor cells prior to implantation has been proposed in order to increase the patency of synthetic vascular prostheses.16,17 However, this approach is limited by poor cell adhesion and retention under blood flow.18 Many recent efforts were thus focused on immobilizing bioactive molecules to increase the strength of EC adhesion as well as their growth and survival. Once again, PEG and carboxymethylated dextran (CMD) raised great interest as an under-layer coating/linker to graft peptides and growth factors promoting specific interactions with ECs.19−21 Indeed, PEG and CMD may increase the bioactivity of grafted biomolecules by preventing nonspecific adsorption and biomolecule denaturation. However, it was shown to be difficult to achieve complete cell coverage on such surfaces, probably due to steric hindrance of the nonfouling background.22 Recently, chondroitin sulfate (CS)-based surfaces have been proposed and developed in our laboratory to tether growth factors (GF) in a random23−25 or oriented fashion.25 CS, a member of the glycosaminoglycan family, is a complex sulfated polysaccharide containing repetitive units of glucuronic acid and galactosamine, which was recently shown to enhance resistance to apoptosis in vascular cells.26 When we compared cell behavior on CS and CMD surfaces on which GF had been grafted, CS surfaces presented a huge advantage.25 We hypothesize that it is mainly due to CS cell pro-adhesive properties despite good protein resistance. Moreover, others have shown that sulfated polysaccharides can improve hemocompatibility by means of electrostatic repulsion toward negatively charged blood components.27 This suggests that CS could combine several advantages as a coating on vascular implants. The purpose of the present work was to study the advantages and limitations of CS for the creation of low-fouling and nonthrombogenic surfaces by comparing CS to well-known low-fouling polymers, i.e., multi arm-PEG and CMD. In order to achieve this objective, we took advantage of stable nitrogenrich plasma polymerized thin film coatings (hereafter referred as LP) that can be deposited on any biomaterial surface.28,29 LP presents a high concentration of primary amine functional groups that allowed covalent coupling of the three (PEG, CMD, and CS) carboxyl-functionalized molecules. Fouling properties were first assessed by following adsorption of fibrinogen, bovine serum albumin (BSA), and fetal bovine serum (FBS) in real time using quartz crystal microbalance with dissipation monitoring (QCM-D) and confirmed using labeled BSA static incubation. The functionalized surfaces were also evaluated for their ability to promote endothelial cell adhesion, to promote in vitro endothelial cell adhesion, growth and retention under flow. Finally, platelet adhesion during blood perfusion was also assessed on bare and endothelial cell grown surfaces. Altogether, our results show that a CS coating
2. EXPERIMENTAL SECTION 2.1. Chemicals and Reagents. Amino-coated glass slides (10 × 10 mm2) were purchased from Erie Scientific Co. (Portsmouth, NH, USA). PET film was purchased from Goodfellow (Huntingdon, England). Albumin Texas Red conjugate, bovine serum albumin (BSA), fibrinogen, phosphate-buffered saline (PBS), chondroitin sulfate (CS), anhydrous ethanol (EtOH), 1-ethyl-3-(3-dimethylaminopropyl)-carbodiimide (EDC), N-hydroxysuccinimide (NHS), paraformaldhehyde, 2-morpholinoethanesulfonic acid (MES), and Hoechst 33342 were obtained from Sigma-Aldrich Canada Ltd. (Oakville, ON, Canada). A 4-arm PEG with N-hydroxysuccinimide terminal functional groups (PEG-NHS; Mw = 10 kDa) was purchased from Creative PEG Works Inc. (Winston Salem, NC, USA). Carboxymethylated dextran (CMD) chains were generated from commercially available dextran (Pharmacosmos, Holbaek, Denmark) by adapting previously reported protocols.30 Under our experimental conditions, the carboxymethylation degree of the dextran chains (70 kDa) was estimated to be approximately 60%. Tygon flexible surgical tubes (Tygon R-100) and chloroform (99% purity) were obtained from Fisher Scientific Co. (Ottawa, ON, Canada). D-Phenylalanyl-Lprolyl-L-arginine chloromethyl ketone (PPACK) was obtained from Calbiochem (Quebec, Canada). Antivinculin antibody was purchased from Millipore (Mouse antivinculin MAB3574, Millipore, USA). Secondary antibody (Alexa Fluor 546 Goat Anti-Mouse IgG) and Alexa 488 phalloidin were purchased from Molecular probes (Invitrogen, Canada). 2.2. Coating Preparation. 2.2.1. Plasma Polymerization. Nitrogen-rich plasma polymerized ethylene coatings (LP) were deposited on microscope glass slides, PET, and on gold-plated QCM-D crystal surfaces using a low-pressure radio frequency (r.f.) glow-discharge plasma reactor, as previously described in detail.28,29 In brief, a mixture of anhydrous ammonia (NH3) and ethylene (C2H4) (99.9% and 99.5% purity, respectively; Air Liquide Canada Ltd., Montreal, QC, Canada) was admitted into a cylindrical aluminum/ steel reactor chamber at flow rates of 15 and 20 standard cubic centimeters per minute (sccm), respectively. The low-power (10 W) plasma was created at a pressure of 80 Pa, resulting in a negative d.c. bias voltage of −40 V. This gas ratio (R = FNH3/FC2H4 = 0.75) was chosen since it was previously found to create coatings with the best compromise in terms of primary amine concentration and stability in aqueous media. XPS analysis showed a high concentration of nitrogen ([N] = 14%) and primary amines ([NH2] = 7.5%, determined by chemical derivatization using 4-[trifluoromethyl] benzaldehyde [TFBA] followed by F content measurement) in the resulting coating, as detailed previously.29 Moreover, these coatings were found to be very stable in aqueous solutions, with less than 10% decrease in thickness when immersed in Milli-Q water or PBS solution for up to 1 week, as assessed by Dektak profilometry.29 LP deposition using these conditions led to a smooth and homogeneous layer, without detectable porosity and with a roughness (Rq) of 0.2−0.3 nm, as observed by scanning electron microscopy (SEM) and atomic force microscopy (AFM).29 In the present study, a plasma deposition time of 10 min was chosen for all experiments, leading to 80 to 90 nm-thick LP coatings. 2.2.2. Chemical Grafting. PEG, CMD, and CS grafting were performed on LP-coated substrates as well as on amino-coated glass slides, both containing NH2 functional groups. Amino-coated glass surfaces were first cleaned with chloroform for 2 min using an ultrasonic bath followed by rinsing twice with Milli-Q water and drying with a stream of nitrogen gas. Chemical grafting was performed according to previously optimized protocols24,30,31 that are briefly summarized below. 2.2.2.1. Star PEG Grafting. The multiarm PEG-NHS solution of 5% (w/v) was prepared in 25 mM phosphate buffer (pH 8.5) and left to react with the NH2-displaying surface for 2 h at room temperature 2513
dx.doi.org/10.1021/bm5003762 | Biomacromolecules 2014, 15, 2512−2520
Biomacromolecules
Article
(RT).31 The slides were finally rinsed once with PBS (10 mM, pH 7.4) and two times with Milli-Q water for 2 min in an ultrasonic bath. Finally, surfaces were dried with a nitrogen gas stream. 2.2.2.2. CMD Grafting. CMD was grafted using water-soluble carbodiimide chemistry.30 The CMD solution was prepared in Milli-Q water (2 mg/mL). Once dissolved, CMD was activated by preparation of a solution containing (for a total volume of 1 mL) 800 μL of CMD solution, 100 μL of 0.4 M EDC in Milli-Q water, and 100 μL of 0.1 M NHS in Milli-Q water. The NHS-activated CMD solution was allowed to react with the NH2-displaying surface for 1 h at RT. The surfaces were finally rinsed and dried as described above. 2.2.2.3. CS Grafting. The CS solution was prepared by dissolving 0.1 g in 1 mL of Milli-Q water and was filtered (0.2 μm PTFE filter) to remove aggregates. CS was then activated by the preparation of a solution containing (for 1 mL) 400 μL of EtOH, 347 μL of Milli-Q water, 50 μL of 1 M MES, 57 μL of 0.4 M EDC, 46 μL of 0.1 M NHS, and 100 μL of CS solution.24,25 The solution was reacted with the NH2-displaying surfaces for 1 h at RT. Rinsing steps and drying were performed as described above. 2.3. Surface Characterization. CS, PEG, and CMD coatings on aminated surfaces have been investigated by complementary surface analysis methods, as presented in our previous publications.23,24,30,31 Grafting of each biomolecule was confirmed by X-ray photoelectron spectroscopy (XPS), and the coatings’ dry thicknesses were determined by ellipsometry. Time-of-flight secondary ion mass spectrometry (ToF-SIMS) analysis was also conducted to confirm the uniformity of the coating. The main results are summarized in the Results section. In this article, the wettability of each coating was compared with static water contact angle measurements using a visual contact angle (VCA) Optima (AST Products, Billerica, MA) automated system. A 2 μL Milli-Q water (pH ≈ 6) drop was placed on each sample, and the static contact angle measurements were performed on both sides of the drop within 5 s. Three measurements were taken on each surface, and experiments were repeated on three independent samples. 2.4. Protein Adsorption Studies. 2.4.1. Quartz Crystal Microbalance with Dissipation Monitoring (QCM-D). QCM-D measurements were performed with a Q-sense E4 system (Q-Sense AB, Sweden). QCM-D allows to probe both the adsorption kinetics and the viscoelastic properties of the adsorbed protein layer in real-time by measuring the changes in frequency and energy dissipation of an oscillating quartz crystal.32 LP was deposited on the gold plated sensor surfaces, followed by chemical grafting of PEG, CMD, and CS. Protein adsorption kinetics were studied on the three different surfaces and on an LP surface as the control. First, adsorption of human fibrinogen (0.5 mg/mL; 340 kDa) and BSA (0.2 mg/mL; 66 kDa) were studied independently. Then the surfaces were subjected to FBS (10% v/v), which contains a wide range of proteins of various sizes. Since the protein solutions were prepared in PBS, stable baselines for frequency and dissipation were first obtained in PBS solution. Then, each protein solution was allowed to flow for 2 h, followed by rinsing with PBS for 30 min to remove reversely adsorbed proteins. All experiments were performed at 50 μL/min, and each surface was tested at least three times. Protein adsorption was monitored by recording changes in resonance frequency (Δf) and energy dissipation (ΔD) of the crystal. The adsorbed mass of thin and nondissipative layers can be calculated using the Sauerbrey relationship (Δm = C/n Δf; where Δf is linearly related to the adsorbed mass (Δm), n is the harmonic number, and C is the mass sensitivity constant of the crystal).33 In practice, most surface-adsorbed protein layers are hydrated, viscous, and cause significant energy dissipation. In such cases, the dissipation factor (D) can be considered, which is defined as the ratio of the energy dissipated during one period of oscillation and the energy stored in the oscillating system (D = Edissipated/2πEstored).34 In this study, depending on dissipation values, the adsorbed mass was calculated via the Sauerbrey equation33 or by applying a viscoelastic, single-layer Voigt model implemented in the QCM software (Q tools, Q-sense AB, Sweden).34 The Voigt model was applied when dissipation shift was at least 5% of the frequency shift of the respective surface.32 QCM-D
data analyzed with the Voigt model were normalized using all the overtones but the first one. The fitting parameters, layer density (1200 kg/m3), layer viscosity (between 0.001 and 0.01 kg/ms), and layer thickness (between 10−10 and 10−7 m) were used in the Voigt model as described previously.31 2.4.2. Fluorescence Measurements. To confirm QCM-D data and enable the comparison with bare PET surfaces, Texas Red conjugated BSA was also used to measure protein adsorption on PEG, CMD, and CS grafted LP surfaces, as well as on bare PET. Each surface was covered with labeled BSA solution (0.2 mg/mL in PBS) for 2 h in static condition at RT (surfaces were protected from light exposure). The surfaces were washed thoroughly with PBS to remove unbound proteins, dried with a stream of nitrogen, and examined under a fluorescence microscope (Nikon Eclipse E600) at 10× magnification. The fluorescence intensity was assumed to be directly proportional to the amount of adsorbed albumin on the surface. The background autofluorescence was subtracted for each sample. 2.5. Endothelial Cell Adhesion, Growth, and Focal Adhesion. The various surfaces (bare PET, PET coated with LP ± CS, PEG, and CMD; 1 cm2 each) were placed at the bottom of 24-well plates. Cloning cylinders (internal surface area = 0.5 cm2; Corning, Lowell, MA, USA) were used to limit the area and to retain the films at the bottom of the wells. HUVECs (Lonza, Shawinigan, Canada) were routinely cultured in Lonza EGM-2 medium (2% (v/v) FBS) and used between passages 2 and 6. For cell adhesion and growth assays, cells were suspended in EGM2 culture medium at a density of 0.75 × 105 cells/mL. A volume of 200 μL (15,000 cells) was deposited on each surface and left to adhere for 4 h. The cloning cylinders were then removed, and surfaces were washed with PBS (1×) to remove nonadherent cells. Cells were stained with a 0.75% (w/v) crystal violet solution (Sigma-Aldrich, Oakville, Canada) for 15 min and rinsed three times with Milli-Q water before drying. For growth assays, after removal of the cloning cylinders, cells were left to grow for 2 days in complete medium at 37 °C and 5% CO2 before being stained as described above. Three images per sample (corresponding to 4.7 mm2, about 10% of the total area) were captured by optical microscopy to evaluate cell density. At least four samples were tested per condition, and the experiment was repeated three times. In addition, focal contact formation was evaluated by immunostaining for vinculin and actin fibers after 24 h adhesion of HUVEC (15,000 cells/well) on the various surfaces (PET, LP, LP + CS, and LP + CMD) in the presence as well as in the absence of serum. To that purpose, substrates were rinsed with PBS, fixed with 4% paraformaldehyde for 10 min, permeabilized with 0.4% Triton X-100 for 10 min, and blocked for 1 h with 2% BSA in PBS. Samples were then incubated overnight with a mouse antivinculin antibody (MAB3574; diluted 1:200 in PBS/2% BSA) at 4 °C, followed by incubation with a secondary Goat Anti-Mouse IgG Alexa 546 (diluted 1:600 in PBS/2% BSA) for 1 h at RT. To stain actin, samples were then incubated with Phalloidin-Alexa 488 (1:40 in PBS) for 1 h and finally counterstained for the nucleus with Hoechest33258 (0.5 μg/mL in PBS), which was added for the last 10 min of incubation. Samples were then mounted on glass slides using a DABCO mounting medium and images of stained samples were acquired using confocal microscopy (Olympus FV1000MPE). 2.6. Platelet Adhesion Assay. Platelet adhesion was evaluated by perfusion assays using fresh human blood. This part of the study has been approved by the human ethical committee of the Montreal Heart Institute. All subjects gave informed consent and were free from drugs interfering with platelet function for at least 2 weeks before blood sampling. The experiments were conducted as follows. A 60 mL sample of venous blood from each subject was anticoagulated with 6 mL of D-phenylalanyl- L -prolyl-L -arginine chloromethyl ketone (PPACK) in saline (50 nM final concentration). The platelet adhesion assay was conducted using a slightly modified setup of the Badimon chamber,35 consisting of Plexiglas perfusion chambers that mimic the tube-like cylindrical shape of blood vessels, connected to a peristaltic pump using nontoxic and nonpyrogenic Tygon flexible surgical tubes (see Figure 1S in Supporting Information). The samples (bare PET, 2514
dx.doi.org/10.1021/bm5003762 | Biomacromolecules 2014, 15, 2512−2520
Biomacromolecules
Article
PET coated with LP ± PEG, CMD, and CS) were mounted in the perfusion chambers that contained a window of 2 mm internal diameter and 10 mm in length permitting direct exposure of the samples to the blood.36 A thermostatically controlled water bath was used to maintain the perfusion system at 37 °C. Samples were exposed to blood for 15 min at a flow rate of 40 mL/min, which corresponded to a shear stress of about 850 s−1 (29 dyn/cm2), as calculated using Poiseuille’s law with the following approximation: the blood is an ideal Newtonian fluid, the blood flow is steady and laminar, and the blood vessel is straight, cylindrical, and inelastic (eq 1).
τ = 32·μ·
Q π ·d3
tions.23,30,31 In brief, XPS analysis confirmed the presence of sulfur (0.3%) indicative of CS on the surface,24 while ToFSIMS demonstrated its uniform distribution on the surface (see Figure 2S in Supporting Information).23 PEG grafting was confirmed with XPS survey and high resolution scans of C 1s and O 1s.33 The dry thickness of the CS layer was estimated by ellipsometry to be around 0.8 nm (n ≥ 5) and slightly higher for CMD (about 1.3 nm; n = 5).30 In both cases, the surface was almost completely covered since adsorption of a 6 kDa molecule (epidermal growth factor) was reduced by 93% according to ELISA measurements, as published by Lequoy et al.25 In the present work, contact angle measurements were performed to rapidly confirm chemical graftings and compare the three surfaces. As shown in Figure 1, all modified surfaces
(1)
where τ corresponds to the shear stress, μ to the viscosity of blood (3.5 × 10−3 Pa s), Q to the mean volumetric flow rate (40 mL/min), and d to the chamber’s internal diameter (2 mm). Blood exposure was followed by rinsing with saline for 10 s to remove nonadhered blood elements. 2.6.1. Platelet Staining with CD61/FITC. At the end of the perfusion, each sample was removed from the chambers and rinsed three times with PBS. Adherent platelets were then stained with fluorescein isothiocyanate (FITC)-labeled anti-CD61 antibody, which is specific for CD61, a platelet membrane glycoprotein.37 The surfaces were covered with 20 μg/mL of antibody in 1% BSA for 15 min, washed three times with PBS, and fixed with 1% paraformaldhehyde for 25 min. The surfaces were then mounted on microscope slides containing DABCO (Fisher Scientific, Canada) as a mounting medium. Confocal microscopy images (Carl Zeiss LSM 510) were taken in the central zone of the exposed surface, while areas out of the edges were excluded. The percentage of surface area covered by platelets was calculated using ImageJ software. 2.6.2. Scanning Electron Microscopy (SEM). SEM was used to assess platelet adhesion and morphology. The specimens were dehydrated through a series of graded ethanol solutions (30%, 50%, 70%, 95%, and 2 × 100% v/v) and subjected to CO2 critical point drying (E3000, Polaron, Quorum Technologies). The dried specimens were sputter-coated with gold for 2 min (∼15 nm layer) and then observed using a Hitachi S-3600N (Hitachi High-Technologies, Canada) operating at an acceleration voltage of 5 kV under high vacuum. 2.6.3. Platelet Adhesion in the Presence of HUVECs. Platelet adhesion and activation were also investigated on bare PET, LP, and CS surfaces in the presence of HUVECs. In this case, HUVECs were first left to adhere and grow on each surface for 7 days at the initial density of 105 cells/mL (20,000 cells/well) in EGM-2 (2% FBS). The surfaces were then rinsed once with EGM-2 and stained with CellVue Maroon (M.Target, West Chester, USA) (1.5 μM in EGM-2) for 10 min at 37 °C. The staining was stopped by adding FBS, followed by three consecutive washes with EGM-2. Surfaces were transferred onto perfusion chambers and exposed to whole blood for 15 min. Platelet staining and calculation of the percentage of surface area covered by the platelet were carried out as described above. In addition, the cell density on each surface was calculated. Cell density on similar surfaces that were not submitted to blood perfusion was also determined for comparison purposes to estimate the percentage of HUVECs detached from each type of surface during perfusion. 2.7. Statistical Analysis. Data are expressed as the mean ± standard deviation. Statistical analysis was carried out using one-way ANOVA followed by Tukey’s posthoc analysis. Student’s t test was used when comparing two groups. p < 0.05 was considered to be statistically significant.
Figure 1. Static water contact angle measurements on bare PET, LP, and LP coated with PEG, CMD, and CS (mean ± SD; n ≥ 10); * p < 0.0001 vs LP.
(PEG, CMD, and CS) exhibited contact angles significantly lower than those observed on LP (62 ± 1°) as well as PET (66 ± 3°) surfaces, the lowest contact angle corresponding to CMD (40 ± 2°). The same trend was observed when grafting these polymers on aminated glass surfaces (see Figure 3S in Supporting Information). 3.2. Protein Adsorption. Low-fouling properties of PEG-, CMD- and CS-coated LP surfaces were assayed by QCM-D and compared to those of LP. First, fibrinogen injections were performed on the different surfaces. Examples of frequency and dissipation curves obtained for LP and CMD coatings are presented in Figure 2. Fibrinogen adsorption occurred rapidly on LP surfaces, yielding a significant negative frequency shift (Figure 2a) and positive dissipation shift (Figure 2b), compared to the PBS stable baseline. As expected, these shifts were largely reduced on CMD-coated surfaces, indicating almost complete prevention of fibrinogen adsorption. Similar adsorption assays were performed for each surface using BSA and FBS. In each case, frequency and dissipation values were used to calculate the mass of adsorbed material (Figure 3). The average reduction of adsorbed mass on PEG, CMD, and CS compared to that on LP surfaces is presented in Table 1. LP surfaces induced a large amount of fibrinogen adsorption that was almost suppressed by CMD (98%) and PEG (90%) coatings, as expected. Interestingly, results for CScoated surfaces were comparable to those of PEG (87%). In the case of BSA, PEG and CMD also exhibited very good resistance to adsorption, while CS was somewhat less efficient (72% compared to 85% and 84% for PEG and CMD, respectively; Table 1 and Figure 3a). The small number of samples that were assayed did not permit us to draw conclusions on the significance of these differences. Finally, when FBS was injected
3. RESULTS 3.1. Physical Characterization of Coatings. The chemical grafting protocols for PEG, CMD, and CS on aminated surfaces had been recently optimized to maximize surface density. These surfaces were characterized using various techniques such as XPS, ToF-SIMS, and ellipsometric dry thickness measurements, as detailed in previous publica2515
dx.doi.org/10.1021/bm5003762 | Biomacromolecules 2014, 15, 2512−2520
Biomacromolecules
Article
Figure 2. QCM-D frequency (Δf, panel a) and dissipation (ΔD, panel b) versus time for fibrinogen (0.5 mg/mL) adsorption on CMD modified and LP surfaces. The arrows refer to the injection of the protein solution (left) and PBS rinsing (right).
Figure 3. Adsorption and desorption kinetics of BSA (a) and 10% (w/v) FBS (b) on PEG, CMD, and CS modified and unmodified LP surfaces. Arrows indicate the start of the protein (left) and PBS (right) injections.
reductions of BSA adsorption mediated by PEG and CMD coatings (85 ± 3% and 84 ± 4% reduction compared with LP, respectively) was significantly (p < 0.0001) better than that mediated by a CS coating (71 ± 4%). The difference was also significant compared with the bare PET surface (p < 0.0001), with a reduction of 81 ± 4%, 80 ± 3%, and 64 ± 8% on PEG, CMD, and CS, respectively. 3.3. Cell Adhesion and Growth. HUVEC adhesion and growth were investigated on PEG, CMD, and CS surfaces and compared with those of LP, bare PET, and on polystyrene culture plates (PCP) for the sake of comparison (Figure 5 and typical crystal violet pictures in Supporting Information). HUVEC adhesion and growth were excellent on LP and the PCP positive control but relatively poor on PET as already observed in previous studies.38 Grafting of PEG and CMD to the surfaces led to even lower adhesion and no growth. Cells displayed a round shape, and spreading was limited. These
Table 1. Mean Percentage Reduction of Protein Adsorbed Mass Compared to that of the LP Surface, Based on QCM-D Results (mean ± SD; n ≥ 3) coating
% reduction of fibrinogen
% reduction of BSA
% reduction of FBS (10% v/v)
PEG CMD CS
90 ± 3 98 ± 0 87 ± 9
85 ± 3 84 ± 4 72 ± 4
88 ± 5 89 ± 5 66 ± 9
(Figure 3b), the same trend was observed (Figure 3b and Table 1). For validation purposes, under static conditions, BSA adsorption on various surfaces was also directly evaluated using fluorescent labeled albumin. Results (Figure 4) were in excellent agreement with those obtained by QCM-D. The
Figure 4. Fluorescence intensity of adsorbed Texas Red labeled albumin (0.2 mg/mL) on PEG-, CMD-, and CS-modified and unmodified LP surfaces, as well as on bare PET control (mean ± SD; n ≥ 8); *, p < 0.0001 vs PET; #, p < 0.0001 vs CS.
Figure 5. HUVEC density after 4 h (adhesion) and 2 days (growth) (mean ± SD; n = 12); *, p < 0.0001. Typical images after crystal violet staining are presented in Supporting Information. 2516
dx.doi.org/10.1021/bm5003762 | Biomacromolecules 2014, 15, 2512−2520
Biomacromolecules
Article
Figure 6. Immunostaining of vinculin (red), actin (green), and the nucleus (blue) after 24 h of HUVEC adhesion on PET, LP, LP-CS, and LP-CMD surfaces.
observations are consistent with previous investigations with low-fouling polymer surfaces.4,39 In stark contrast, cells adhered and spread well on CS, and showed rapid growth similar to that observed on LP and PCP surfaces. The presence of HUVEC focal adhesion after 24 h on CS was confirmed by actin and vinculin immunostaining, as shown in Figure 6. HUVECs on CS as well LP surfaces exhibited a spread morphology, with well-defined actin fibers (in green) throughout the cytoplasm and termini clustered at vinculin rich sites (red spots mainly at the edges). These observations are indicative of integrin-mediated anchorage on CS and LP surfaces and suggest strong cell attachment and adequate signaling for cell growth and survival. On bare PET, fewer cells were observed. Some exhibited a rounded morphology (unshown), and others were spread but exhibited only a few stress fibers. The actin cytoskeleton did not appear to be well organized. Only rare focal adhesion points could be observed. Finally, as expected, very few cells were observed after 24 h on CMD, and they generally present a rounded morphology, without any focal contact formation. The absence of spreading and focal adhesion of HUVEC on CMD suggests that there is a strong risk of cell detachment and cell death and correlates well with the very low cell number after 2 days of growth in Figure 5. 3.4. Platelet Adhesion. Platelet adhesion was assayed on the different coatings. Figure 7 presents the average percentages
Figure 8. Representative SEM (a, c, e; scale bar = 10 μm) and confocal microscopy (b, d, f: labeling with CD 61/FITC; scale bar = 20 μm) images of platelet adhesion on LP (a,b), PET (c,d), and CS (e,f) surfaces after perfusion with whole blood.
the variability in platelet reactivity between the different blood donors. However, on PET surfaces, platelets presented filopodia typical of an activated phenotype (Figure 8c). In contrast, platelet adhesion was almost abolished on PEG, CMD, and CS surfaces. Interestingly, CS-grafted surfaces were as efficient as PEG and CMD surfaces in resisting platelet adhesion as no significant difference was observed between these three surfaces. Since CS surfaces exhibited very low platelet adhesion and favorable HUVEC growth, they appeared as promising surfaces to create a complete endothelial layer on a nonthrombogenic under-layer. To that purpose, HUVECs must present a nonthrombogenic phenotype and must be resistant to the shear induced by blood flow. As a first assessment of these properties, cells grown for 7 days on the various surfaces were subjected to blood flow for 15 min. HUVEC density and platelet adhesion were measured. Figure 9 presents the HUVEC density on each type of surface that had been submitted or not to blood flow, thus enabling one to estimate HUVEC retention. Despite relatively
Figure 7. Percentage of surface area covered by platelets after perfusion with whole blood for 15 min (mean ± SD; n ≥ 5); *, p < 0.0001 vs LP; #, p < 0.0001 vs PET.
of LP, PET, PEG, CMD, and CS surfaces covered by platelets after whole blood perfusion for 15 min. Representative images from confocal microscopy after CD61/FITC labeling and from SEM for LP, PET, and CS are shown in Figure 8. LP surfaces presented numerous adhered and activated platelets, most likely due to the positively charged amine groups present in aqueous media. Although platelet adhesion on PET was significantly lower than that on LP surfaces, the adhesion on PET surfaces varied a lot from one experiment to another, probably due to
Figure 9. Cell density on the different surfaces (bare PET and LP ±CS coating) not perfused (−Perfusion) and after perfusion (+Perfusion) of whole blood (mean ± SD; n = 7); *, p < 0.0001 vs all other surfaces. 2517
dx.doi.org/10.1021/bm5003762 | Biomacromolecules 2014, 15, 2512−2520
Biomacromolecules
Article
Figure 10. (a) Percentage of LP and CS surfaces covered by platelets after perfusion with whole blood in the absence (−HUVEC) and presence (+HUVEC) of previously seeded HUVEC (mean ± SD; n = 7; *, p < 0.0001 vs other surfaces). (b) Representative images of HUVECs and platelets on LP and CS surfaces after perfusion (HUVEC membranes colored with CellVue Maroon (blue) and platelets stained with anti-CD61/FITC antibody (green)). Scale bar corresponds to 200 μm.
identifying the adsorbed proteins nor determining if these proteins are denatured once adsorbed. Other methods would be required to identify the proteins selectively recruited by CS and thus gain a better understanding of its mode of action. Interestingly, although protein adsorption was not completely abrogated on the CS surface, platelet adhesion was almost abolished and comparable to that observed on PEG and CMD surfaces. The perfusion assay in this work mimicked physiological conditions;43 it is believed to give a better estimate of the antiplatelet property of our surfaces as compared to static assays because the shear stress induced by blood flow causes qualitative and quantitative differences in platelet adhesion.44 Moreover, platelets have a relatively low density and do not settle easily in static conditions. One limitation of our platelet adhesion test is the short perfusion time (15 min), which was chosen due to the limited lifetime of the anticoagulant as well as to allow for comparing several surfaces with blood from the same donor. However, this time was sufficient to observe a significant decrease in platelet adhesion on PEG, CMD, and CS surfaces as compared to that on PET. Almost no platelets were observed on CS, and none were activated. These observations are consistent with some previous results27 and are probably due to the electrostatic repulsion of the negative charges on sulfated CS and the additive effect of highly hydrophilic properties of CS surfaces. Indeed, platelet adhesion and complement system activation are known to decrease with increasing surface hydrophilicity.45 Overall, resistance to fibrinogen adsorption and absence of platelet adhesion on CS suggest that it can prevent thrombus formation. Yet, additional studies (blood clotting time, leukocyte activation, and longer time points) would be required to confirm the low-thrombogenicity of CS coatings. More importantly, in stark contrast with PEG and CMD, CS was shown to support and promote strong endothelial cell adhesion and growth leading to the formation of a complete and flowresistant endothelium, whereas most cells detached from the PET surface during perfusion. Focal adhesion points and strong cell attachment suggest integrin-mediated anchorage. CS probably promotes the adsorption of some proteins and/or growth factors present in the culture medium (containing 2% FBS), which are favorable to HUVEC adhesion and proliferation. Indeed, CS by itself was not shown to exhibit proliferative properties,24 and much lower cell adhesion was observed when cells were incubated on the surface in serum free medium.
similar initial HUVEC densities on all three surfaces (PET, LP, and CS), cell density on PET was shown to decrease dramatically during perfusion, with only 25% retention on the surface. In contrast, cells strongly adhered on LP and CS, on which cell densities before and after perfusion were not significantly different. As expected, platelet adhesion on CS was very low, both on HUVEC-grown and bare-CS surfaces (Figure 10). This suggests that HUVEC growth on CS exhibited a low platelet reactivity phenotype. However, further studies would be required to confirm the functional properties of HUVECs, including the assay of released pro- and antithrombogenic molecules40 during longer perfusion times. On HUVEC-grown LP surfaces, the presence of numerous cells significantly reduced platelet adhesion compared to that of LP alone (p < 0.0001). However, platelet adhesion was still visible in regions where no cells were present (Figure 10b), which might be either due to the lack of complete coverage after 7 days of growth or to cell detachment during perfusion. Altogether, even with HUVECs, LP surfaces tended to be more thrombogenic than CS, though the difference was not statistically significant (p = 0.16; n = 7).
4. DISCUSSION Modifying surface physical and chemical properties of vascular implants has shown some success in reducing the associated risks of thrombosis. However, the long-term patency of smalldiameter (
