Evaluation of Antithrombogenic pHPC on CoCr Substrates for Biomedical Applications

Bare metal endovascular implants pose a significant risk of causing thrombogenic complications. Antithrombogenic surface modifications, such as phenox’s “Hydrophilic Polymer Coating” (pHPC), which was originally developed for NiTi implants, decrease the thrombogenicity of metal surfaces. In this study, the transferability of pHPC onto biomedical CoCr-based alloys is examined. Coated surfaces were characterized via contact-angle measurement and atomic force microscopy. The equivalence of the antithrombogenic effect in contact with whole human blood was demonstrated in vitro for CoCr plates compared to NiTi plates on a platform shaker and for braided devices in a Chandler loop. Platelet adhesion was assessed via scanning electron microscopy and fluorescence microscopy. The coating efficiency of pHPC on CoCr plates was confirmed by a reduction of the contact angle from 84.4° ± 5.1° to 36.2° ± 5.2°. The surface roughness was not affected by the application of pHPC. Platelet adhesion was significantly reduced on pHPC-coated specimens. The platelet covered area was reduced by 85% for coated CoCr plates compared to uncoated samples. Uncoated braided devices were completely covered by platelets, while on the pHPC-coated samples, very few platelets were visible. In conclusion, the antithrombogenic effect of pHPC coating can be successfully applied on CoCr plates as well as stent-like CoCr braids.


Introduction
Ischemic and hemorrhagic vascular diseases (e.g., myocardial infarct, stroke, atherosclerosis) are among the most frequent causes of death in Western countries [1]. Today, metallic implants are the mainstay of endovascular therapy, which is true for peripheral, cardiological, and neurovascular interventions. For the purpose of vessel occlusion, device-induced thrombus formation is desired, but when it comes to vessel reconstruction, the inherent thrombogenicity of stents and their derivates is an issue [2,3].
Extended research in the field of endovascular treatments of diseases has brought forth a large variety of devices designed for many purposes. While the designs of stents and flow diverters vary significantly, the most commonly used materials today are NiTi alloys, CoCr alloys, or medical-grade steel due to their excellent biocompatibility and mechanical properties [4].
However, once exposed to blood, endovascular implants with a bare metal surface trigger the adhesion of plasma proteins and activate both the coagulation cascade and

Coating Efficiency-Wettability
In the literature, the antithrombogenic pHPC coating is known to change the wettability of NiTi surfaces, and can therefore be used as an indicator for the coating efficiency [31]. Thus, in this study, the wettability and hydrophilicity of the plates were measured using a modified Wilhelmy Plate method in a tensiometer (DCAT 21, DataPhysics Instruments GmbH, Stuttgart, Germany) through immersion of the sample plates in pure water. The dynamic contact angle was derived from the weight change during immersion, as water will form a larger lamella when the specimens are more hydrophilic. The CAs of the NiTi (n = 10) and CoCr (n = 10) plates were measured before and after the coating of the samples.

Atomic Force Microscopy
AFM of coated and uncoated NiTi and CoCr plates was performed on a Dimension FastScan AFM (Bruker, Billerica, MA, USA) to characterize the topographic nature of the plates before and after coating. An area of 1 × 1 µm was scanned on each sample, and the surface analysis was assessed through determination of the mean arithmetic roughness value, Ra, from three randomly chosen line scans (length of 1 µm each).

In Vitro Blood Contact
Blood sampling was performed under the approval of the ethics commission of the Faculty of Medicine of the Ruhr Universität Bochum, Germany (registration number: . Informed consent was obtained from all individual participants included in the study. All procedures performed in studies involving human participants were in accordance with  /or national research committee and with the  1964 Helsinki Declaration and its later amendments or comparable ethical standards. For in vitro testing, the plates were incubated under dynamic conditions on a shaker in heparinized human venous blood. The blood was collected from 10 voluntary and healthy donors. An abnormal blood count measured via a hematology analyzer (KM-21 N, Sysmex, Nordersted, Germany) and the intake of drugs interfering with blood coagulation were exclusion criteria for participation in the study. All steps were performed under sterile conditions. The plates were rinsed twice in phosphate-buffered saline (PBS) before blood contact. Afterwards, they were incubated under dynamic conditions in 24-well plates in 2 mL blood on a shaker (speed level 4; Titramax 100, Heidolph Instruments, Schwabach, Germany). After 10 min of incubation, non-adherent cells were removed by rinsing the specimens with PBS.
The stent devices were incubated in a modified Chandler loop flow model. Blood was collected with the "S-Monovette Blood collection system" (Sarsted, Nümbrecht, Germany). An abnormal blood count measured via a hematology analyzer (KM-21 N, Sysmex, Nordersted, Germany) and the intake of drugs interfering with blood coagulation were exclusion criteria for participation in the study. The devices were deployed in PVC tubes (Rehau, Erlangen, Germany) with an inner diameter of 3 mm. The tubes were first rinsed with PBS and filled with 2 mL of blood, leaving an air bubble of 1 mL inside the tube; then, the tube's ends were interconnected, forming loops. The loops were then rotated at a rate of 30 rotations per minute in a water bath at 37 • C for 120 min. The rotation speed resulted in a flow volume of 60 mL/min and a flow velocity of 90 mL/min (due to the air bubble), thereby resembling typical flow rates in the vertebral artery [38]. The blood count was again performed after incubation.

CD61 Immunohistochemistry
CD61 (integrin b-3) is a glycoprotein found on the surface of platelets. CD61+ adherent platelets were stained using CD61-PE antibody (BD Pharmingen, Heidelberg, Germany) fluorescence staining. After incubation in whole blood, the specimens were rinsed with PBS twice in order to remove non-adherent cells. The antibody was diluted 1:5 in PBS and added to the specimens. After 15 min incubation in the dark, specimens were rinsed with PBS twice, fixed with 1% paraformaldehyde (Sigma-Aldrich, Taufkirchen, Germany) in PBS, and analyzed using a fluorescence microscope (Olympus BX40, Olympus, Hamburg, Germany) equipped with a digital camera (Moticam 5, Motic, Wetzlar, Germany).

Semi-Quantitative Phase Analyses
Platelet adherence was assessed by analyzing the platelet-covered area of the plates using ImageJ software (National Institutes of Health, Bethesda, MD, USA). Four defined areas of interest per plate (0.06 mm 2 each, total 0.24 mm 2 ) were examined. The results were reported as the percentage of positively stained area by relating the positively stained area (CD61+ stained cells) with the absolute area of the regions of interest.

Scanning Electron Microscopy
To examine the level of platelet activation and clot formation on the stent devices, SEM examinations were performed after incubation in the Chandler loop on the same specimens used for fluorescence imaging.
After incubation, the specimens were prepared for SEM examination by rinsing with PBS twice, fixing with 3.7% glutaraldehyde (Sigma-Aldrich, Taufkirchen, Germany) in PBS for 15 min, dehydrating with an ascending series of alcohols, drying for 24 h at room temperature, and, finally, sputtering with gold (Edwards Sputter Coater S150B9, Edwards Limited, Crawley, UK).
SEM was performed on an LEO 1530 Gemini (Carl Zeiss AG, Jena, Germany).

Statistical Analysis
Data from platelet activation were analyzed with the Mann-Whitney test. If a significant difference was found, treated groups were compared with the control group by using Dunn's post-test. GraphPad Prism software (version 3.03; GraphPad Software Inc., La Jolla, CA, USA) was used for the analysis. p-values of less than 0.05 were considered statistically significant. The data of the surface roughness was assessed by determining the average Ra for three individual lines per scan. Results are presented as 'mean ± standard error of the mean' (mean ± SE).

Surface Analysis-Wettability and Surface Topography
The results of the CA measurement are given in Table  0.383 ± 0.094 0.214 ± 0.020 0.029 ± 0.002 0.036 ± 0.001 Figure 1 shows representative AFM images of the uncoated (Figure 1a,c) and coated specimens (Figure 1b,d). Microscopically, the surface of the NiTi plates had a homogeneous surface with a honeycomb-like structure and minimal waviness. The evenly distributed peaks were about 50 nm apart and had an altitude of about 8 nm. The CoCr plates had a homogenous surface profile with peaks of less than 0.2 nm height and a spacing of less than 10 nm, with some slight waviness. The corresponding results of a nanoscopic analysis of the mean arithmetic roughness values (Ra) are given in Table 1. The coating of the specimens did not alter the homogenous surface profile compared to the corresponding uncoated specimens.
After incubation, the specimens were prepared for SEM examination by rinsing with PBS twice, fixing with 3.7% glutaraldehyde (Sigma-Aldrich, Taufkirchen, Germany) in PBS for 15 min, dehydrating with an ascending series of alcohols, drying for 24 h at room temperature, and, finally, sputtering with gold (Edwards Sputter Coater S150B9, Edwards Limited, Crawley, UK).
SEM was performed on an LEO 1530 Gemini (Carl Zeiss AG, Jena, Germany).

Statistical Analysis
Data from platelet activation were analyzed with the Mann-Whitney test. If a significant difference was found, treated groups were compared with the control group by using Dunn's post-test. GraphPad Prism software (version 3.03; GraphPad Software Inc., La Jolla, CA, USA) was used for the analysis. p-values of less than 0.05 were considered statistically significant. The data of the surface roughness was assessed by determining the average Ra for three individual lines per scan. Results are presented as 'mean ± standard error of the mean' (mean ± SE).

Surface Analysis-Wettability and Surface Topography
The results of the CA measurement are given in Table 1. The CAs were 78.1° ± 4.8° for the uncoated NiTi specimens and 84.4° ± 5.1° for the uncoated CoCr specimens. After functionalization with pHPC, the CAs of CoCr and NiTi plates were decreased by roughly 45°-50° (CA of 31.6° ± 2.2° for the NiTi plates, 36.2° ± 5.2° for the CoCr plates).  Figure 1 shows representative AFM images of the uncoated (Figure 1a,c) and coated specimens (Figure 1b,d). Microscopically, the surface of the NiTi plates had a homogeneous surface with a honeycomb-like structure and minimal waviness. The evenly distributed peaks were about 50 nm apart and had an altitude of about 8 nm. The CoCr plates had a homogenous surface profile with peaks of less than 0.2 nm height and a spacing of less than 10 nm, with some slight waviness. The corresponding results of a nanoscopic analysis of the mean arithmetic roughness values (Ra) are given in Table 1. The coating of the specimens did not alter the homogenous surface profile compared to the corresponding uncoated specimens.    Figure 2 shows the representative fluorescence microscopic images of pHPC-coated (Figure 2b,d) and uncoated (Figure 2a,c) NiTi and CoCr plates, which were incubated in vitro in human blood and stained with CD61 antibody. Few CD61+ platelets were detected on the coated specimens (yellow fluorescence), whereas the uncoated specimens were homogenously covered with cells in each experiment.  Figure 3 shows the results of the quantitative phase analysis of the platelet-covered area on the plates. The amount of surface area covered with adhering thrombocytes was significantly lower on the pHPC-coated NiTi and CoCr plates compared to the uncoated ones. Overall, the CoCr plates did elicit more platelet activation than the NiTi plates. Uncoated CoCr specimens were covered with CD61+ platelets by 85.5% ± 10.7%, while the uncoated NiTi specimens were covered by 66.9% ± 21.5%. The pHPC-coated CoCr plates were covered by 12.8% ± 9.4% and the coated NiTi plates by 6.7% ± 8.8%. On average, the platelet adherence was reduced by 90% on coated NiTi and by 85% on the pHPC-coated CoCr specimens. . Quantitative phase analysis of the area coated with CD61+ cells from experiments with ten different donors (mean ± SE; asterisks denote significance at *** p ≤ 0.001; Mann-Whitney and Dunn's post-test; sample size uncoated vs. coated n = 10). While significantly fewer CD61+ platelets adhered to the coated specimens than to the uncoated ones, there is no significant difference between the NiTi and the CoCr specimens.  Figure 3 shows the results of the quantitative phase analysis of the platelet-covered area on the plates. The amount of surface area covered with adhering thrombocytes was significantly lower on the pHPC-coated NiTi and CoCr plates compared to the uncoated ones. Overall, the CoCr plates did elicit more platelet activation than the NiTi plates. Uncoated CoCr specimens were covered with CD61+ platelets by 85.5% ± 10.7%, while the uncoated NiTi specimens were covered by 66.9% ± 21.5%. The pHPC-coated CoCr plates were covered by 12.8% ± 9.4% and the coated NiTi plates by 6.7% ± 8.8%. On average, the platelet adherence was reduced by 90% on coated NiTi and by 85% on the pHPC-coated CoCr specimens. Figure 2 shows the representative fluorescence microscopic images of pHPC-coated (Figure 2b,d) and uncoated (Figure 2a,c) NiTi and CoCr plates, which were incubated in vitro in human blood and stained with CD61 antibody. Few CD61+ platelets were detected on the coated specimens (yellow fluorescence), whereas the uncoated specimens were homogenously covered with cells in each experiment.  Figure 3 shows the results of the quantitative phase analysis of the platelet-covered area on the plates. The amount of surface area covered with adhering thrombocytes was significantly lower on the pHPC-coated NiTi and CoCr plates compared to the uncoated ones. Overall, the CoCr plates did elicit more platelet activation than the NiTi plates. Uncoated CoCr specimens were covered with CD61+ platelets by 85.5% ± 10.7%, while the uncoated NiTi specimens were covered by 66.9% ± 21.5%. The pHPC-coated CoCr plates were covered by 12.8% ± 9.4% and the coated NiTi plates by 6.7% ± 8.8%. On average, the platelet adherence was reduced by 90% on coated NiTi and by 85% on the pHPC-coated CoCr specimens. . Quantitative phase analysis of the area coated with CD61+ cells from experiments with ten different donors (mean ± SE; asterisks denote significance at *** p ≤ 0.001; Mann-Whitney and Dunn's post-test; sample size uncoated vs. coated n = 10). While significantly fewer CD61+ platelets adhered to the coated specimens than to the uncoated ones, there is no significant difference between the NiTi and the CoCr specimens. . Quantitative phase analysis of the area coated with CD61+ cells from experiments with ten different donors (mean ± SE; asterisks denote significance at *** p ≤ 0.001; Mann-Whitney and Dunn's post-test; sample size uncoated vs. coated n = 10). While significantly fewer CD61+ platelets adhered to the coated specimens than to the uncoated ones, there is no significant difference between the NiTi and the CoCr specimens.

Thrombogenicity-Fluorescence and SEM Imaging
Additionally, stent devices were incubated in a modified Chandler loop flow model with whole human blood for 120 min, and adherent cells were again stained with a CD61 antibody and analyzed using fluorescence microscopy. Figure 4 shows representative fluorescence micrographs of uncoated and pHPC-coated p48 MW (NiTi) and CoCr braids. While the uncoated devices were completely covered with Additionally, stent devices were incubated in a modified Chandler loop flow model with whole human blood for 120 min, and adherent cells were again stained with a CD61 antibody and analyzed using fluorescence microscopy. Figure 4 shows representative fluorescence micrographs of uncoated and pHPCcoated p48 MW (NiTi) and CoCr braids. While the uncoated devices were completely covered with adherent platelets, there were considerably fewer fluorescent cells on the pHPCcoated braids. On both the p48 MW HPC and the coated CoCr braids, the adherent cells tended to adhere to contact points where the wires of the braids crossed over each other.  (Figure 5a,c,e,g) were completely covered by a dense layer of fibrin and embedded platelets. In contrast, only a few platelets were visible on the pHPC-coated samples (Figure 5b,d,f,h). Again, the predominant adherence of platelets at the crossings of the wires on the coated devices was observed.    (Figure 5a,c,e,g) were completely covered by a dense layer of fibrin and embedded platelets. In contrast, only a few platelets were visible on the pHPC-coated samples (Figure 5b,d,f,h). Again, the predominant adherence of platelets at the crossings of the wires on the coated devices was observed. Additionally, stent devices were incubated in a modified Chandler loop flow model with whole human blood for 120 min, and adherent cells were again stained with a CD61 antibody and analyzed using fluorescence microscopy. Figure 4 shows representative fluorescence micrographs of uncoated and pHPCcoated p48 MW (NiTi) and CoCr braids. While the uncoated devices were completely covered with adherent platelets, there were considerably fewer fluorescent cells on the pHPCcoated braids. On both the p48 MW HPC and the coated CoCr braids, the adherent cells tended to adhere to contact points where the wires of the braids crossed over each other.  (Figure 5a,c,e,g) were completely covered by a dense layer of fibrin and embedded platelets. In contrast, only a few platelets were visible on the pHPC-coated samples (Figure 5b,d,f,h). Again, the predominant adherence of platelets at the crossings of the wires on the coated devices was observed.

Discussion
pHPC was originally developed for surface modification of NiTi devices and is approved and available on endovascular stent devices made from NiTi. However, the most commonly used materials for similar applications, apart from NiTi alloys, are CoCr alloys and medical-grade steel [4]. While NiTi exhibits a surface oxide that mainly consists of TiO 2 , for CoCr alloys, the high amount of Cr is responsible for the formation of a thick oxide layer that mostly consists of Cr 2 O 3 [33]. Hence, the purpose of this study was to evaluate the compatibility of pHPC with CoCr-alloy substrates to expand the range of antithrombogenic biomaterials that are commonly used in blood contact. Therefore, the coating efficiency, quality, and (anti)thrombogenicity of pHPC on CoCr substrates was examined by means of dynamic contact-angle (CA) measurement, atomic force microscopy (AFM), in vitro testing, fluorescence microscopy, and scanning electron microscopy (SEM).
The platelet activation and adhesion were significantly reduced on pHPC-coated CoCr specimens compared to the corresponding uncoated specimens (Figures 2-4). The surface area covered by adherent platelets was reduced by 85% on the pHPC-coated CoCr plates compared to a reduction by 90% on coated NiTi (Figure 3). This is also true for pHPCcoated NiTi and CoCr braids when tested under dynamic conditions in a Chandler loop experiment (Figures 4 and 5). It is known that the thrombogenicity of a surface depends on the preceding adhesion of proteins [39]. In turn, the adhesion of proteins depends on the surface topography and wettability [40].
The surface topography of a biomaterial's surface has been shown to have a great impact on the adherence and proliferation of proteins and, thereby, cells; therefore, it is of interest regarding platelet activation and endothelialization of vascular implants [40]. The microscopic and nanoscopic surface structure was apparently unaffected by the coating (Table 1, Figure 1). The NiTi plates were very smooth and had a consistent surface structure, typical for electropolished NiTi [41], while the CoCr plates were even smoother. The surface structures of the NiTi and CoCr samples were not microscopically or macroscopically affected by the coating. The coating is known to be extremely thin (<10 nm) on NiTi [31] and apparently formed the same way on CoCr, as the coating procedure did not affect the original surface. The smooth surfaces of the plates resembled those of typical bare metal stents, which are also usually electropolished to deburr the edges and to create a smooth surface finish after laser cutting, thereby lowering the likelihood of a foreign body response by blood or vessel cells [42,43].
While the pHPC coating does not affect the topography of the NiTi and CoCr plates, it certainly does affect the wettability ( Table 1).
The CAs of the uncoated plates can be considered as weakly hydrophilic, with 78.1 • ± 4.8 • for the NiTi surface and 84.4 • ± 5.1 • for the CoCr surface (Table 1) [44]. After functionalization of the surfaces using the pHPC coating, the decrease of the CA by roughly 45 • to 50 • resulted in a moderately hydrophilic surface (CA of 36.2 • ± 5.2 • for the CoCr plates and 31.6 • ± 2.2 • for the NiTi plates-see Table 1), thereby indicating successful coating of both materials.
A biomaterial's wettability is known to influence the adhesion of proteins and, thereby, to influence the interaction with cells [45]. However, the wettability of devices is of interest for further reasons. The adhesion of proteins is known to be the start of cell-surface interaction [46]. Firstly, thrombocyte activation and adhesion might be affected by a changed wettability on the surface of implants designed to treat vascular disease, and secondly, the proliferation of endothelial cells might be enhanced or decreased due to the same effect. A moderate wettability with CAs of around 35 • can be considered optimal for devices that are in contact with blood. Protein adsorption to surfaces is the beginning of all cell-surface interactions. Extremely hydrophilic (CA < 25 • ) or hydrophobic surfaces (CA > 85 • ) do reduce protein adsorption and, thereby, reduce cell-surface interaction in general [45]. Platelet activation and adherence are known to be decreased by increasing wettability [47][48][49]. However, endothelialization, which is important for the long-term biocompatibility of endovascular devices, is decreased on extremely hydrophilic or hy-drophobic surfaces due to the reduced adhesion of extracellular matrix proteins (EMPs). Moderately hydrophilic surfaces, like pHPC-coated NiTi or CoCr surfaces, however, do allow the adherence of extracellular matrix proteins (and, therefore, endothelialization), with a maximum endothelial cell adhesion occurring at a CA of 35 • to 50 • [45,50]. This is in line with the observations from our three animal studies. In those studies, pHPC-coated and uncoated specimens were implanted in the carotid arteries of New Zealand white rabbits for 30 and 120 d. It was shown that pHPC coating of NiTi surfaces does not delay the ingrowth of stent devices into the vessel wall [51,52]. The antithrombogenic effect of pHPC has been shown to last long enough after implantation to keep patients under ASA monotherapy until the device is completely incorporated by endothelial cells [32]. If this is true for CoCr substrates as well still needs to be examined in further research.
Overall, the antithrombogenic effect of pHPC was successfully applied to two different medical CoCr alloys (plates from L-605 and flow-diverter-like stent devices from 35NLT wires). The transfer onto other materials for biomedical applications might be of interest for further research. The remaining platelets on the coated stent devices were predominantly observed at the contact points of the crossings of wires. This has been observed before on braided devices after implantation in animals and humans, and is mainly caused by additional turbulence at these points, as platelets are activated by additional sheer stress [53][54][55].
However, the surfaces of medical-grade CoCr alloys and stainless steels are effectively alike. While NiTi exhibits a surface oxide mainly consisting of TiO 2 , for CoCr alloys and stainless steel, the high amount of Cr is responsible for the formation of a thick oxide layer that mostly consists of Cr 2 O 3 [33]. pHPC might, therefore, be effortlessly transferred onto stainless steel devices as well.
Even though the results are promising, further experiments regarding the coating's mechanical and biological durability on CoCr surfaces might be of further interest. However, these properties need to be evaluated under consideration of the actual application because, e.g., different implant designs might generate different loads, have different surface properties, or hold other challenges.

Conclusions
The conducted experiments demonstrate the transferability of the pHPC technology onto two different CoCr alloys for biomedical applications. The coating was successfully applied to plates and braided wires without negatively affecting the materials' surface quality. Furthermore, the platelet adhesion was significantly reduced on coated NiTi and CoCr devices compared to uncoated devices. The results of this study show that antithrombogenic pHPC coating is a promising candidate for the development of endovascular devices of different materials with antithrombogenic surface properties.