Colloids and Surfaces B: Biointerfaces 130 (2015) 192–198

Contents lists available at ScienceDirect

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

Athrombogenic hydrogel coatings for medical devices – Examination of biological properties Beata A. Butruk-Raszeja a,∗ , Ilona Łojszczyk a , Tomasz Ciach a , Magdalena Ko´scielniak-Ziemniak b , Karolina Janiczak b , Roman Kustosz b , Małgorzata Gonsior b a Laboratory of Biomedical Engineering, Faculty of Chemical and Process Engineering, Warsaw University of Technology, Wary´ nskiego 1, 00-645 Warsaw, Poland b Artificial Heart Laboratory, Foundation for Cardiac Surgery Development, Wolno´sci 345A, 41-800 Zabrze, Poland

a r t i c l e

i n f o

Article history: Received 2 January 2015 Received in revised form 2 April 2015 Accepted 5 April 2015 Available online 11 April 2015 Keywords: Polyurethane Surface modification Blood compatibility Platelet Fibrinogen Hydrogel

a b s t r a c t In the article the authors present hydrogel coatings prepared from polyvinylpyrrolidone (PVP) macromolecules, which are chemically bonded to polyurethane (PU) substrate. The coating is designed to improve the surface hemocompatibility of blood-contacting medical devices. The coating was characterized in terms of physical properties (swelling ratio, hydrogel density, surface morphology, coating thickness, coating durability). In order to examine surface hemocompatibility, the materials were contacted with whole human blood under arterial flow simulated conditions followed by calculation of platelet consumption and the number of platelet aggregates. Samples were also contacted with platelet-poor plasma; the number of surface-adsorbed fibrinogen molecules was measured using ELISA assay. Finally, the inflammatory reaction after implantation was assessed, using New Zealand rabbits. The designed coating is characterized by high water content and excellent durability in aqueous environment – over a 35-day period, no significant changes in coating thickness were observed. Experiments with blood proved twice the reduction in adsorption of serum-derived fibrinogen together with a moderate reduction in the number of platelet aggregates formed during the contact of the material with blood. The analysis of an inflammatory reaction after the implantation confirmed high biocompatibility of the fabricated materials – studies have shown no toxic effects of the implanted material on the surrounding animal tissues. © 2015 Elsevier B.V. All rights reserved.

1. Introduction The formation of thrombus on biomaterial surface still remains a serious problem during the applications of implants controlling the blood flow, i.e. in mechanical heart assist devices. The highly adverse and dangerous effects include clot formation inside the prosthesis cavity, followed by the blood clots release to the patient’s arteries, which entails the risk of serious complications for the patient’s life and health. Another undesirable effect is platelets contact activation, resulting in platelets micro-particles distribution to the patient’s circulation, playing significant role in late blood coagulation activation in the patient’s peripheral vessels. To prevent blood-clotting process and improve the prostheses

∗ Corresponding author. Tel.: +48 22 234 64 19; fax: +48 22 825 14 40. E-mail address: [email protected] (B.A. Butruk-Raszeja). http://dx.doi.org/10.1016/j.colsurfb.2015.04.008 0927-7765/© 2015 Elsevier B.V. All rights reserved.

hemocompatibility, various methods of surface modifications have been proposed. It is worth noticing, that the layer of proteins adsorbed on the surface of the material determines all interactions between the biomaterial and various blood components. The number, type and spatial conformation of the adsorbed proteins is, in turn, determined by the physicochemical properties of the surface material, in particular its wettability, morphology and chemical composition. There are many reports on the influence of the surface wettability on the process of protein adsorption and cell adhesion [1–4]. Generally, it is believed that surfaces with low wettability (hydrophobic) promote the adhesion and conformational changes of proteins, whereas the surfaces with high wettability (hydrophilic) reduce protein adsorption and minimize the possibility of denaturation. Therefore, the majority of modification methods are based on increasing the hydrophilicity of the material. One of the techniques for biocompatibility improvement is based on biopassive surfaces

B.A. Butruk-Raszeja et al. / Colloids and Surfaces B: Biointerfaces 130 (2015) 192–198

formation. Biopassive surfaces are characterized by low interfacial energy, which results in a decreased number of surface-adsorbed proteins and surface-adhered cells. In addition, the strength of interaction between the protein and the surface is small, which limits the conformational changes occurring in the protein molecules and enables maintenance of their biological activity [5,6]. As a result of our study, we have developed the process of fabrication hydrogel coatings on the surface of polyurethanes (PU). The coating aims to increase the hydrophilicity of the basis polymer and create a biopassive surface. As the coating materials we used polyvinylpyrrolidone (PVP). PVP displays a high biocompatibility [7–9], antibacterial activity [10,11] and high hemocompatibility – the polymer was used as a human plasma substitute during World War II [12]. Further studies have shown that supplementation with PVP biomaterials can lead to a reduction in hemolysis of erythrocytes [13] as well as a reduction in adsorption of albumin [14], fibronectin [15] and fibrinogen [16]. It has been proven that the PVP coating significantly improves the hemocompatibility of the intravascular catheters [15] and membranes for plasma separation [17]. The great advantage of PVP is its resistance to hydrolysis, which allows to apply it in contact with highly hydrolytic environment such as blood. In our previous papers, we described the method of coating fabrication and explained in detail the mechanism of hydrogel formation [18,19]. The coating is formed due to free radical macromolecular grafting – crosslinking. Polymer surface was first immersed in an organic solution containing radical source: cumene hydroperoxide (CHP) with addition of a branching and anchoring agent: ethylene glycol dimethylacrylate (EGDMA). In the second step, the substrate was immersed in a water solution containing given concentration of PVP and Fe2+ . The novelty of the process consists in the fact that free radicals are formed mostly at the polymer/solution interphase, what assures high grafting efficiency together with formation of covalent bonds between polymer substrate and modifying layer. In this paper, we present the results of testing the chosen physicochemical and biological properties of the material.

193

RT. After the coating procedure, the polymer discs were washed with 0.1% (w/v) SDS water solution for 5 min and subsequently in water (overnight). In order to compare coatings fabricated with the proposed method, the control samples (PVP S) were also prepared. These samples were fabricated using standard dip-coating technique, without the presence of free radicals or cross-linking agent in the coating solution. PU discs were immersed in 10% solution of PVP for 15 min and subsequently washed with 0.1% (w/v) SDS water solution for 5 min. 2.4. Coating characterization Physical properties of hydrogel coatings were evaluated by measuring the equilibrium swelling ratio (EWC, %) and hydrogel density (HG, g/m2 ) as described previously [18]. The coating thickness and adhesion to the substrate were analyzed microscopically: materials were incubated in PBS at 37 ◦ C to obtain the equilibrium hydration followed by microscopic analysis of cross-sections. To ensure a better visualization of the coating, the samples were stained with rhodamine (0.002%, w/v). Coating thickness measurements were carried out for six replicates of the material, the thickness of each coating was measured at four randomly selected locations. The morphology of the freeze-dried hydrogel coatings was analyzed using a scanning electron microscope (Phenom, Phenom World). 2.5. Coating durability The stability of hydrogel coating in an aqueous medium was determined by the analysis of changes in coating thickness. The samples were incubated in PBS with 0.5% (w/v) sodium azide at 37 ◦ C. After a certain time of degradation (7, 14, 21 and 28 days), the samples were removed from the solution, rinsed with PBS and the cross-sections were analyzed using optical microscopy. Coating thickness measurements were carried out for six replicates of the material, the thickness of each coating was measured at four randomly selected locations.

2. Materials and methods 2.6. Fibrinogen adsorption 2.1. Materials Polyurethane (ChronoFlex C, 75D, AdvanSource Biomaterial, Wilmington, U.S.) was purchased in the form of pellets. Reagents, namely polyvinylpyrrolidone (PVP) powder with average molecular weight of 360 kDa, iron (II) chloride (FeCl2 ), ascorbic acid (AA), cumene hydroperoxide (CHP), ethylene glycol dimethacrylate (EGDMA) and sodium dodecyl sulfate (SDS) were obtained ´ Poland). Hexane was purchased from from Sigma–Aldrich (Poznan, Chempur, Poland. 2.2. PU film preparation PU disc were injection-moulded from PU granulates to a form of discs with 40-mm diameter. The discs were cut out with a metal stamp and subsequently modified as follows. Before modification PU discs were washed with 5% alcohol solution and dried. 2.3. PU modification with hydrogel coating PU samples were modified according to the previously presented free-radical-based method [18]. Briefly, the PU discs were immersed in a hexane solution containing EGDMA (5%, v/v) and CHP (3%, v/v) for 5 min at RT. The samples were then placed in a water solution containing given amounts of PVP (5%, w/v, for PVP 5 or 10%, w/v, for PVP 10), FeCl2 (0.1%, w/v) and 0.1% (w/v) AA for 15 min at

The plasma-derived fibrinogen adsorption to the test materials was analyzed using platelet poor plasma (PPP). In order to prepare PPP, 50 ml of blood was drawn from a healthy, aspirinfree donor (female, 28 years) in the K2EDTA tubes using the BD Vacutainer vacuum system. Blood was centrifuged at 300 × g for 30 min. The obtained supernatant was transferred to clean tubes and centrifuged at 2000 for 20 min. After the centrifugation, the supernatant (PPP) was transferred to a new sterile tube. The samples (disks with a diameter of 18 mm) were placed in 24-well polystyrene plates and equilibrated with PBS at 37 ◦ C overnight. Next, the investigated materials were contacted with 100% PPP for 1 h at 37 ◦ C. Materials with surface-adsorbed fibrinogen were analyzed with ELISA assay. The following procedure was applied: after incubation with PPP, the samples were rinsed (3× 5 min) with washing buffer (PBS supplemented with 0.05% Tween 20, Sigma–Aldrich, Poland), blocked with non-fat dry milk (5% solution in PBS, 1 h, RT), rinsed with washing buffer (3× 5 min), incubated with primary antibodies (1 h, RT), rinsed with washing buffer (3× 5 min), incubated with secondary antibodies (1 h, RT), rinsed with washing buffer (6× 5 min) and transferred to fresh plates to eliminate the influence of the protein adsorbed to the well walls. Materials were then incubated with peroxidase substrate solution – SigmaFast OPD (o-phenylenediamine dihydrochloride, Sigma–Aldrich, Poland) in the dark at RT for 30 min. After the reaction, a part of the solution (200 ␮l) from each well was transferred

194

B.A. Butruk-Raszeja et al. / Colloids and Surfaces B: Biointerfaces 130 (2015) 192–198

to 96-well plate; the optical density of the solution was read at the 450 nm. During the test the following antibodies were used: primary polyclonal anti-fibrinogen produced in goat (Sigma–Aldrich, Poland, dilution 1:1000), secondary anti-goat IgG conjugated to peroxidase (Sigma–Aldrich, Poland, dilution 1:20,000). Antibodies were diluted with a 2% solution of dry non-fat milk in washing buffer. A set of samples (K PU, K PVP 5 and K PVP 10) were subject to all ELISA assay steps omitting first step (materials incubation with PPP). These control materials were used to evaluate the possibility of the non-specific binding of the antibodies to the surfaces of the materials. As a blank, a solution of o-phenylenediamine dihydrochloride (a substrate for the antibody-conjugated peroxidase) was used. 2.7. Hemocompatibility under dynamic conditions Blood–biomaterial interactions under flow conditions were evaluated using a platelet analyzer (Impact-R, DiaMed) as described elsewhere [20]. Briefly, ETO-sterilized materials in the form of discs with a diameter of 14.5 mm were placed in a PTFE cones (part of the ImpactR analyzer) and covered with 130 ␮l of blood. Arterial flow simulating shear stress (720 rpm) was applied for 5 min. After the test, the blood samples were collected and analyzed with flow cytometry. Each material was tested in triplicate. Two additional blood samples were also studied: blood samples stored under static conditions (negative control) and blood sample activated by adding ADP at a concentration of 20 mM (positive control). Blood samples incubated with materials were analyzed in order to calculate the percentage of platelets remaining in blood after the shear-stress test in relation to free platelets present in negative control: RP [%] =

RPX × 100%, RPKN

where: RP – the remaining platelets [%], RPX – number of platelets present in blood sample after the shear stress test, RPKN – number of platelets present in the negative control blood sample. Additionally, we quantified the percentage ratio between free platelets and platelet aggregates present in blood after shear stress test. 2.8. In vivo biocompatibility Animal model (New Zealand rabbit) was used for the evaluation of in vivo biocompatibility of the fabricated material. Implantation was carried out according to ISO 10993-2 and ISO 10993-6. The samples (PVP 10) in the form of a film with dimensions of 5 mm × 5 mm and thickness of 1 mm were ETO sterilized and aired for 3 weeks. The animals were divided into groups of different implantation period (7, 14 and 30 days); with 4 rabbits in every group. There were two control groups. On one group of rabbits the surgery was carried out without implantation – to evaluate the inflammatory response derived from the surgical wounds (4 rabbits). The other group of rabbits did not undergo any treatment – to assess the body’s response to environmental conditions (1 rabbit). The rabbits of the test group were anesthetized (general anesthesia with 15 mg/ml of ketamine hydrochloride), blood samples were taken from the ear artery and the sterile single implant was implanted inside the muscle of the back of each animal. After the surgery, the wound was stitched, the animals were placed in cages and observed. The scar appearance and behavior of animals were observed thorough the implantation period. After the given period of implantation, blood samples were taken from a vein in the lower limb and the animals received an injection of xylazine hydrochloride and ketamine hydrochloride for premedication, and

Table 1 The physical properties of hydrogel coatings. #

Coating thickness [␮m]

HG [g/m2 ]

EWC [%]

PVP 5 PVP 10

90 ± 14 144 ± 30**

19 ± 2 28 ± 7*

650 ± 90 870 ± 300*

MV ± SD, n = 48 for thickness measurement, n = 16 for HG and EWC measurement. * p < 0.05 vs. PVP 5, ** p < 0.001 vs. PVP 5.

subsequently a morbital intracardiac injection. Tissue fragments obtained from the dorsal muscle, heart, thymus, liver, spleen, kidneys and lymph nodes in the vicinity of the implantation site were fixed (4% formaldehyde), dehydrated in a series of alcohol solutions and stained with hematoxylin, eosin and trichrome The peripheral blood samples collected during the experiment were used to assess blood morphology. 2.9. Statistical analysis Results of fibrinogen adsorption and hemocompatibility studies were expressed as means ± SD. Statistical significance of differences was analyzed using single factor analysis of variance (ANOVA) for p < 0.05 with post hoc Tukey’s test (OriginPRO 8.0). 3. Results 3.1. Coating characterization Lyophilized hydrogel-modified surfaces were subject to microscopic analysis. The surfaces displayed a highly porous structure characteristic for all xerogels (Fig. 1). The pore diameter was in the range of 50–100 microns. No significant differences in the xerogel structure between PVP 5 and PVP 10 were observed. The control material, obtained using a standard dip coating method (PVP S) showed no porous structure – the majority of PVP coating was washed away during the washing step. The morphology analysis confirmed that the coating produced as a result of the proposed modification process is chemically bonded to the substrate (the hydrogel layer remains on the substrate surface after washing in detergent solution). It can be also concluded that PVP chains were cross-linked in the course of the free radical reaction – the coating presents porous structure, which is not present in PVP coating produced using a dip coating technique. For the prepared materials coatings the basic physical parameters (HG, EWC, coating thickness) were calculated. The measured parameters depend on the concentration of PVP in the modifying solution (Table 1). The thickness of the coating equals to 90 ± 30 ␮m for PVP 5. In the case of PVP 10 coating thickness increased up to 144 ± 30 ␮m. The higher coating thickness for PVP 10 resulted in a higher content of PVP (HG) and water (EWC) in the material. The microscopic analysis of cross-sections confirmed good adhesion of the coating to the PU along the entire surface. 3.2. Coating durability Durability of the coatings was analyzed by measuring the changes in coating thickness during a 35-day degradation in an aqueous medium. The study revealed a slight change in the hydrogel coating thickness during the whole process of degradation (Fig. 2A). The initial thickness of the coating and the thickness of the coating after a 35-day degradation equaled, respectively, 90 ± 14 ␮m (t = 0) and 84 ± 17 ␮m (t = 35) for PVP 5. In the case of PVP 10 these values were as follows: 144 ± 30 ␮m (t = 0) and 124 ± 40 ␮m (t = 35). Microscopic examination of the cross-sections of the tested materials performed at 7-day intervals confirmed the

B.A. Butruk-Raszeja et al. / Colloids and Surfaces B: Biointerfaces 130 (2015) 192–198

195

Fig. 1. Morphology of materials modified using proposed free radical method after lyophilization, magnification 600×.

homogeneity of the coatings; no local discontinuities or defects in the structure of the coatings were observed. 3.3. Fibrinogen adsorption ELISA assay revealed a significant (50%, p < 0.001, Fig. 2B) reduction of surface-adsorbed fibrinogen after surface modification with hydrogel. The results showed no significant differences in fibrinogen surface adsorption between PVP 10 and PVP 5. The response of control materials (materials incubated with PBS instead of PPP) was low and comparable to the results obtained for blank (data not shown). Thus, we excluded the possibility of false positive response of ELISA assay as a result of a non-specific interaction of antibodies with the materials surfaces. 3.4. Hemocompatibility under dynamic conditions Hydrogel-coated materials were incubated with human whole blood under dynamic conditions. After the test, blood samples contacted with material samples were collected and analyzed. The platelet consumption was determined (expressed as a ratio of the number of free platelets present in the blood sample to the initial number of platelets). The results are presented in Fig. 3A. The positive control (blood sample activated with ADP) showed a significant platelet consumption – the percentage of free platelets was 15 ± 3%. The value of RP obtained for unmodified material (PU) was twice higher and equaled 35 ± 20%. For the PVP 5 sample the

value was similar: 37 ± 24%. However, for the PVP 10 sample the percentage of free platelets was significantly higher and equaled 80 ± 25%, which means that the vast majority of the initial number of platelets did not aggregate or adhere to the material sample. Statistical analysis showed that the results obtained for PVP 10 significantly differed from the values obtained for unmodified PU (p < 0.05). The percentage ratio between free platelets and platelet aggregates with distinction between small aggregates, consisting of 2–3 platelets and large aggregates, consisting of more than 3 platelets (Fig. 3B) was also determined. The percentage of free platelet content in the case of the negative control was 93 ± 2%. For the positive control, this percentage decreased and reached a value of 63 ± 6%. In the case of blood samples contacted with polymer materials the values were as follows: 86 ± 15% for the PU, 80 ± 17% for PVP 5 and 97 ± 3% for PVP 10. The number of big aggregates present in blood samples was as follows: 10 ± 13 for PU, 8 ± 9 for PVP 5 and