In Vitro and In Vivo Biocompatibility Study of Fluorinated Polyphosphazene Coatings for Blood-Contacting Medical Devices

Abstract

Implant-induced thromboembolic events are the most common complication of blood contacting medical devices. Coatings are a promising approach to improve the biocompatibility of current biomaterials and devices. Poly[bis(trifluoroethoxy) phosphazene] (TFE) has been demonstrated to be biocompatible, anti-inflammatory, and antithrombogenic as a device coating over the past decades; however, its inherently poor mechanical properties make applications in medical devices challenging, especially regarding potential detachment from devices. Our previous work developed a new fluorinated polyphosphazene, poly[bis(octafluoropentoxy) phosphazene] (OFP), and incorporated allylphenoxy side groups to the P-N backbone to make the polymer crosslinkable (X-OFP). In this study, we applied this X-OFP coating on central venous catheters and investigated the surface properties and biocompatibility of the coatings. In vitro and in vivo studies demonstrated that X-OFP has a similar antithrombogenic performance as TFE, but its mechanical properties including adhesion strength of coating-to-substrate are significantly improved, thereby enhancing the stability of the coating. The success of X-OFP will provide a platform to incorporate other different side groups to the polymer backbones and generate new fluorinated polyphosphazene polymers having improved biocompatibility and mechanical properties for coating applications in blood-contacting medical devices.

Graphical abstract

1. Introduction:

Blood-contacting medical devices such as intravenous cannula, central venous catheters, cardiovascular stents, ventricular assist devices, extracorporeal circulation circuits, and cardiopulmonary bypass form a critical component of the health care system in modern clinical practice and play a crucial role in disease treatment. However, when these devices contact with blood, they inevitably trigger a complex series of events, including protein adsorption, platelet adhesion and activation, coagulation, and thrombosis, leading to early device failure and increased risk of patient complications. Such device-induced thrombosis is the main problem impeding the development of long-term blood contacting devices [1–3]. Despite decades of efforts by researchers to improve the biocompatibility of biomaterials, thrombosis remains a major challenge for blood-contacting medical devices [3–5].

Device-induced thrombosis arises from the triggering of the blood plasma coagulation cascade along with platelet adhesion/activation on the surface. Both of these adverse events can be influenced by the surface properties of biomaterials that are used in devices. Surface modification, such as coating with antithrombotic functionalities, is an important strategy for the development of biomaterials. One strategy for the development of such biocompatible coating is to incorporate active components, such as heparin, which has been popularly incorporated into surface coatings of medical devices to achieve biocompatibility [6–8]. However, the elution of active agents from these coatings may lead to loss of functionality during long-term use. In addition, any coating failure such as delamination and detachment may present as particulates and result in a stroke or heart attack. In 2015, the U.S. Food and Drug Administration (FDA) warned that the separation of coatings on intravascular medical devices had the potential to cause serious injuries to patients [9]. Therefore, the development of a mechanically robust, non-drug eluting functional polymer coating that can provide long-term thrombo-resistance properties are highly desirable for clinical applications.

Fluorinated polyphosphazene (PP) polymers show resistance to thrombosis, providing an alternative approach for applications in a wide variety of blood-contacting medical devices[10, 11]. Polyphosphazenes are macromolecules with a linear alternating phosphorus-nitrogen (P-N) backbone and two organic/inorganic side groups linked to each phosphorus (Scheme 1A) [12, 13]. The diversity of side groups linked to the backbone generates a variety of PP polymers with a wide range of properties that can be used in biomedical applications [14, 15]. Among the many PPs that have been developed and evaluated for biomedical applications, one class of PPs with fluorinated organic side groups have proven to be especially useful [16]. These fluorinated polymers are hydrophobic, flexible, and often elastomeric. They are also resistant to oxidation, ultraviolet light, and pathogenic microbes, viruses, and fungi [17]. One of the most deeply investigated fluorinated PP is poly[bis(trifluoroethoxy) phosphazene] (TFE) which possesses two trifluoroethoxy side groups on each phosphorus (Scheme 1B). TFE is a soft, high molecular weight fluoropolymer that exhibits biocompatibility, anti-inflammatory and antithrombogenic properties, and exceptional chemical stability. It attracts significant interest as a hydrophobic and biocompatible material for industrial and life science applications such as medical device coatings [18–20]. Prior studies reported successful application of TFE coatings to a coronary stent and showed good outcomes with reduced platelet adhesion, decreased clotting, reduced inflammation and accelerated healing compared with different surfaces [10, 21–24]. A TFE coated coronary stent (trademarked as COBRA PzF^™) has undergone preclinical and clinical testing over the past decade [25–30]. These testing showed accelerated endothelialization relative to drug-eluting stents along with reduced inflammation and thromboresistance. It was suggested that the new stent could be used to safely facilitate an abbreviated duration of dual antiplatelet therapy (DAPT), particularly for those with high bleeding risk. However, the latest COBRA-REDUCE trial to reduce the DAPT concluded that “the treatment with the COBRA PzF stent plus 14 days of DAPT was not superior with respect to bleeding events and was not noninferior with respect to thromboembolic events at 6 months compared with treatment with standard FDA approved drug-eluting stent plus 3 to 6 months of DAPT” [31]. Although no clinical incident was reported in those public documents regarding to TFE-coated devices, the structural integrity of the coating, application techniques and surface compatibility testing with various materials remain a high priority [10]. One limitation of the current TFE coating appears related to its limited mechanical properties which make its applications in devices challenging. For TFE coating on stainless steel stents, an adhesion promoter was used to cover the stent surfaces first to ensure reliable adhesion of the TFE coating to the stent, indicative of poor adhesion strength of the TFE coating to the substrate [21]. In our studies, TFE coatings applied on stainless steel or commercial polyurethane (PU) catheter surfaces were found to become separated from substrates during exposure to fluid under shear, causing coating failure. Therefore, the development of a new biostable PP having thrombogenesis prevention, superior mechanical properties, and biocompatibility is necessary.

Scheme 1.

Chemical structures of (A) general polyphosphazene polymer, (B) TFE, (C) OFP, and (D) X-OFP.

In our prior studies, we developed a new fluorinated PP, poly[bis(octafluoropentoxy) phosphazene] (OFP) (Scheme 1C), based on the TFE-like platform, in which octafluoropentoxy side units are introduced into both functional sites along the backbone instead of trifluoroethoxy side units as in TFE. OFP is stable in biological media and has a high degree of stability against UV and gamma rays, but more importantly, it possesses similar antithrombogenic properties as TFE. To improve the mechanical properties of OFP, we further added an allylphenoxy crosslinking functional group to the P-N backbone to yield a new group of fluorinated PPs, crosslinkable poly[bis(octafluoropentoxy) phosphazene], termed as X-OFP (Scheme 1D). The new polymers contain higher fluorocarbon content and the allylphenoxy side groups which crosslink under UV or at a temperature of ~75°C under nitrogen [32, 33]. Our prior studies showed that X-OFP successfully improved the mechanical strength of OFP and reduced both human Factor XII activation and platelet adhesion, demonstrating that they are resistant to plasma coagulation and thrombosis in vitro. Bacterial adhesion testing on X-OFPs show that crosslinking enhanced surface mechanical properties and significantly reduced bacterial adhesion compared to the pre-crosslinked X-OFP surfaces, suggesting that X-OFPs are also resistant to bacteria in infection models[32]. In biofilm studies of TFE and X-OFP coatings on stainless steel surfaces, X-OFP showed significant reduction of biofilms compared to TFE, indicating that X-OFP is resistant to microbial infection [34]. This is important because healthcare-associated infections (HAIs) have placed a significant burden on the U.S. healthcare system each year [35].

Although our prior studies demonstrated great potential application of new fluorinated PP such as X-OFP in blood contacting medical devices, in vitro studies appear inadequate to provide the information necessary to determine whether specific devices effectively prevent blood clotting without causing systemic adverse effects. To gain a deeper insight into the fluorinated PP coatings that translate into clinical settings, in this study, we examined the hemocompatible characteristics of coatings in vivo using a 7-d rabbit model. We applied the fluorinated PP coatings, TFE and X-OFP, on commercial clinical catheters and characterized their surface properties and biological responses to coatings. In addition, we evaluated the cytocompatibility and blood compatibility of coated catheters in an in vitro blood flow loop system. Finally, the in vivo antithrombotic efficacy of the surfaces was evaluated through a 7-d rabbit thrombogenicity model.

2. Materials and Methods

2.1. Materials

The reagents for synthesis of PPs were pretreated as follows: 2-Allylphenol and 2,2,3,3,4,4,5,5-octafluoro-1-pentanol (Aldrich) were distilled over calcium hydride. Poly(dichlorophosphazene) was prepared via the thermal ring-opening polymerization of recrystallized and sublimed hexachlorocyclotriphosphazene (Fushimi Pharmaceutical Co., Japan) in evacuated Pyrex tubes at 250°C. Tetrahydrofuran (THF, EMD) was dried using solvent purification columns. Other chemicals including sodium hydride (60% dispersion in mineral oil, Aldrich) and 4-fluorophenol (TCI) and 4-(trifluoromethyl) phenol (Aldrich) were used as received.

Synthesis of TFE and X-OFP was carried out under a dry argon atmosphere using standard Schlenk line techniques that have been described in our previous publication [34]. The synthesized X-OFP contains 9.1% of allyphenoxy crosslinking side groups. The structures of PPs were confirmed by ¹H and ³¹P NMR using a Bruker NEO-400 (NANO-2) instrument operated at 400 and 162 MHz, respectively.

2.2. Preparation of PP coatings on catheters

TFE and X-OFP polymers were dissolved in methyl ethyl ketone (MEK) at ~10% concentration (w/v) with stirring for 12 h, and 2,2′-Azobis(2-methylpropionitrile) (5% w/w, Sigma-Aldrich) was added in X-OFP solution as the initiator for crosslinking reactions. The commercial Central Venous Catheterization Kits were purchased from ARROW International (Cleveland, Ohio, USA. Product No. AK-04210) and the polyurethane catheters (16 Ga, 20 cm) were cut into the desired length. For coating, the catheters were dipped in polymer solution twice with an interval time of 20 h between applications. After drying, the X-OFP coated catheters were thermally crosslinked in a sealed glass container with N₂ gas protection at ~75°C overnight.

2.3. Characterization of coating surfaces

An x-ray photoelectron spectroscope (XPS, Physical Electronics VersaProbe III) equipped with a monochromatic Al kα x-ray source (hν = 1,486.6 eV) and a concentric hemispherical analyzer was used to analyze the surface chemistry of catheter coatings. Charge neutralization was performed using both low energy electrons (<5 eV) and argon ions during XPS analysis. The binding energy axis was calibrated using sputter cleaned Cu (Cu 2p_3/2 = 932.62 eV, Cu 3p_3/2 = 75.1 eV) and Au foils (Au 4f_7/2 = 83.96 eV). Peaks were referenced to the CF₂ band in the carbon 1s spectra at 292.5 eV. Quantification was done using instrumental relative sensitivity factors (RSFs) that account for the x-ray cross section and the inelastic mean free path of the electrons.

A multimode atomic force microscope (AFM) with a Nanoscope IIIa control system (Veeco, Santa Barbara, CA) operated in air was used to characterize catheter coating surface topography. A short piece of catheter (1 cm) was carefully attached on a steel disc for imaging. AFM was operated in tapping mode using a sharp Si probe (NSG30-SS, TipsNano) to collect height and phase image data simultaneously. Each sample was measured at no less than 3 different random locations.

A short piece of catheter was carefully attached on the glass slide and a water drop (~ 4 μL) was put on the catheter. The water contact angles were continuously recorded by a ramé-hart contact angle goniometer over 20 min. The average of contact angles was calculated to determine the surface wettability of coatings. The contact angles were obtained by a minimum of six independent measurements and are presented as mean ± standard deviation.

To evaluate the adhesion strength of coating-to-substrate, TFE and X-OFP polymer solutions were applied on 316 stainless steel (316SS) plates and dried. The X-OFP coating was thermally crosslinked at ~75°C with nitrogen overnight. The coating adhesion test was completed in accordance with ASTM D3359-23[36] (Standard Test Methods for Rating Adhesion by Tape Test) using the standard “cross-cut” Method B. To assess the adhesion of coating to polymer surfaces, we used a nano-scratch method to study the adhesion and delamination of coatings. Samples were evaluated by Covalent Meterology (Sunnyvale, CA, USA). Briefly, TFE and X-OFP were coated on flat polyurethane sheets (PU, 1 cm × 2 cm, Carbosil 20 80A, DSM Biomedical Inc., Exton, PA) and treated with the same method as above. An Anton Paar NST3 was used to study the nano-scratch with a 20 μm length wedge blade indenter. Samples were superglued to aluminum stubs and clamped on the instrument stage. The scratching was performed using a linear progressive load over a 1 mm length at a rate of 1 mm/min, and the progressive load range was 0.3 mN to 124 mN. Adhesion is reflected by critical load at delamination. SEM equipped with EDS (energy dispersive X-ray spectroscopy) was used to reveal the coating failure modes.

2.4. In vitro assessment of hemocompatibility of catheter coatings

The hemocompatibility of catheter coatings was assessed by the measurement of plasma coagulation time, platelet activation, and red blood cell (RBC) lysis when blood/plasma was incubated with catheters. Plasma coagulation time is defined as the time from the initiation of the plasma coagulation cascade by contact with materials to the appearance of visible clot. Human platelet poor plasma (PPP) was purchased from Biochemed Services (Winchester, VA, USA) and stored in a −80°C freezer. Prior to use, plasma was thawed in a water bath and centrifuged at 1500g for 20 min to remove any platelets or cells. To assess plasma coagulation time, a catheter was cut into pieces with lengths of 1 cm and placed into 2-mL polystyrene tubes, to which 500 μL of PPP and 400 μL of PBS were added. Then, 100 μL of 0.1 M CaCl₂ was added to allow the plasma coagulation cascade to move forward and a timer was started. The polystyrene tube was capped with parafilm and rotated at 8 rpm on a hematology mixer until a visible clot appeared, and the time was recorded.

Platelet activation in human blood was measured by flow cytometry. Briefly, human blood was collected in ACD Vacutainer^® blood collection tubes (Becton Dickinson, Franklin Lakes, NJ) from healthy volunteers in accordance with institutional policies. The catheter pieces (~1 cm length) were incubated in blood and rotated at 8 rpm at 37°C for 1 and 4 h. All platelets in blood were labeled with CD61 monoclonal antibody (APC Mouse Anti-human CD61, BD, #564174) and the activated platelets were stained with CD62P monoclonal antibody (BV421, Mouse Anti-Human CD62P, BD, #564038). Platelets were assessed by flow cytometry (10 color BD FACSCanto, BD Bioscience) and the platelet activation was calculated by the ratio of activated/total platelets (%).

The catheter coatings were also assessed for hemolytic properties in accordance with ASTM F756-17 [37]. Briefly, catheter pieces (~1 cm length) were incubated in 15-mL centrifuge tubes containing 1 mL human whole blood and 7 mL PBS for 3h at 37°C. To prevent the precipitation of red blood cells (RBC), the tubes were gently inverted every 30 min to mix the RBCs. The complete lysis of RBCs was attained by adding 100 μL blood in 900 μL H₂O to define the 100% lysis point, and blood without polymer contact was used as the negative control (minimal lysis). After 3h, the blood samples were centrifuged at 10,000 rpm for 10 min and the free hemoglobin concentration in supernatant was measured by Drabkin’s reagent kit. The hemolytic index is calculated as:

$$hemolysis=\frac{Free\mathit{\text{hemoglobin ceoncetration in supernatant}}}{Total\mathit{\text{hemoglobin concentration}}}\times 100\%.$$

2.5. MTT cytotoxicity test of catheter coatings

The cytotoxicity of catheter coatings was tested in vitro using 3T3 Mouse fibroblast cells (ATCC CRL - 1658). 3T3 fibroblast cells were grown and maintained in a 75 cm² T-flask and nourished with Dulbecco’s modified Eagle’s medium (DMEM, Gibco, USA) containing 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin. The cytocompatibility of the PU, TFE and X-OFP catheter coatings was tested using the yellow tetrazolium salt (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT, Sigma-Aldrich, St. Louis, MO). This assay evaluates the metabolically active cells to determine cell viability after exposure to samples. To prepare the cells for the assay, the cells were seeded at a seeding density of 5× 10⁴ cells mL⁻¹ into a 96-well flat bottom plate. The plates were incubated at 37°C with 5% CO₂ for 24 h to achieve confluency before exposure to the leachate solution. Subsequently, the leachates were prepared following the complete submersion of the samples in cell culture media for 24 h. After the incubation period, the culture media was removed from the wells and supplemented with 100 μL of the sample leachates. The plates were incubated at 37°C with 5% CO₂ for 24 h. After exposure, the media was removed from the wells and replaced with 100 μL MTT reagent at 0.5 mg mL⁻¹. The MTT was left to react with the cells for 3 h at 37°C. The remaining reagent was removed, and the crystals were dissolved in 100 μL of dimethyl sulfoxide (DMSO). The absorbance was recorded at 570 nm and 690 nm using a BioTek Cytation^™ 5 Cell-Imaging Multi-Mode Reader (Winooski, VT). The percent cellular viability is calculated as followed:

$$\mathrm{Relative}\text{Cell Viability}(\%)=\frac{\text{O}{\text{D}}_{570,\mathrm{Treatment}}-\text{O}{\text{D}}_{690,\mathrm{Treatment}}}{\text{O}{\text{D}}_{570,\mathrm{Control}}-\text{O}{\text{D}}_{690,\mathrm{Control}}}\times 100\%$$

Final data are reported as the mean cellular viability ± standard deviation.

2.6. Assessment of human fibrinogen adsorption responses

Human fibrinogen adsorption was carried out on polymer film surfaces rather than catheter coatings in order to provide sufficient surface areas for protein adsorption and analysis. Polyurethane (PU) Carbosil 20 80A (DSM Biomedical Inc., Exton, PA) was used as the control. PU, TFE and X-OFP solutions were cast on poly(dimethylsiloxane) (PDMS) flat molds and dried to form films. X-OFP was thermally crosslinked at ~75°C under N₂ for overnight. The polymer films were carefully separated from PDMS mold and punched to pieces with diameter of 11 mm. The polymer pieces were incubated in 1 mL of human fibrinogen (2 mg/mL in PBS) solution in 24-well plate at 37°C with shaking at 120 rpm for 2 h. Then, polymer films were rinsed with PBS for 3 times and transferred to 15-mL centrifuge tubes containing 1 mL of PBS and 2% sodium dodecyl sulfate. Samples were ultrasonicated for 20 min and centrifuged at 1500g for 15min. The supernatant was collected and measured by Micro BCA^™ Protein Assay Kit.

To further understand the protein interaction with polymer surfaces, we measured the adhesion forces between human fibrinogen and polymer surfaces using an AFM technique that was described in our prior publication[38]. Briefly, Si₃N₄ AFM probes were first bonded with reactive amine groups by incubation of probes in a 1% (v/v) solution of aminopropyltriethoxysilane (Gelest Inc., PA) in ethanol for 1 h. The probes were then reacted with 10% glutaraldehyde in aqueous solution for 1 h followed by thoroughly rinsing with Millipore water. This process provides a link of the aldehyde group for covalent bonding with the primary amine group on the protein. Finally, probes were incubated in a fibrinogen solution (50 μg/ml) for 1 h to immobilize the proteins on probes.

To measure the adhesion forces of protein and polymer coating surfaces, polymers were cast on glass slides to form coatings. Adhesion force measurements were performed with a JPK NanoWizard^® 4 XP AFM (Bruker) using force mode in PBS at a scan rate of 1 Hz with a z ramp size of 0.5 μm. The applied force was set up at 0.2 nN for all measurements. The forces were collected using force map mode to obtain an 8×8 array of force curves over a scan area of 20×20 μm². All AFM force data were analyzed offline with JPKAFM data processing software.

2.7. In vitro hemocompatibility assessment in blood loop.

A custom-made blood loop setup was used to assess the blood thrombosis response to catheter coatings. The setup was described in our previous publication with a small modification [39] (Scheme 2). Briefly, the catheter was attached on a stainless-steel rod and both ends of the rod were sealed with polyurethane before being inserted into a Tygon tube (Ø 8mm). Bovine blood with CPDA-1 was selected for in vitro studies. Bovine blood was purchased from Lampire biological laboratories (Pipersville, PA, USA) and was recalcified by adding CaCl₂ to yield a final concentration of 2 mM of Ca²⁺. The blood was aliquoted into multiple blood bags containing 120 mL blood each. Blood was circulated at 37°C for 20 h by a Masterflex L/S pump drive (model 07557-02) with a multichannel pump (model 7535-08) at a flow rate of 100 mL min⁻¹. Eight channels were run simultaneously using a mixture of test samples each time. After 20 h, the catheters were carefully removed and briefly placed into 100 mL PBS buffer to remove any nonadherent blood. The exterior blood-contacting surface of the catheters were photographed around 360° and examined by the Image J program to measure the thrombus area on catheter surfaces.

Scheme 2.

In vitro blood loop is used for testing blood coagulation on catheter surfaces. (Left) A picture of blood loop set up. (Right) The catheter alignment in Tygon tube.

2.8. In Vivo hemocompatibility assessments by a 7-day rabbit model

2.8.1. Catheter implantation

In vivo hemocompatibility of catheter coatings was assessed by a New Zealand rabbit model under the standards of the University Committee on Use and Care of Animals IACUC at the University of Georgia. The experimental procedures were described in our previous publication[39]. Briefly, a total of 9 New Zealand white rabbits (2.5-3.5 kg, Charles River Laboratories, Wilmington, MA) were used in this study. All surgical instruments were sterilized via steam sterilization and catheters were sterilized through vaporized hydrogen peroxide at the University of Georgia College of Veterinary Medicine prior to surgery. All rabbits were initially anesthetized with 4 mg kg⁻¹ xylazine and 15 mg kg⁻¹ ketamine hydrochloride. Maintenance anesthesia was administered via isoflurane gas inhalation at a rate of 1.5-3% via chemical ventilation through an ADS 2000 ventilator (Engler Engineering Corp. Hialeah, FL).

Subcutaneous lidocaine (Lidocaine 2%, Vet One, Boise, ID) was infused on the right lateral cervical area, and a 2 cm skin incision was made over the bifurcation of the right external jugular vein, and the linguofacial vein branch was isolated for the catheter insertion. The facial vein was ligated proximally. A small venotomy was made under distal occlusion, followed by the introduction of the vein pick. The catheter (~8 cm length) was then introduced into the jugular vein and advanced into the cranial vena cava under the vein pick. Approximately 6 cm of the catheter length is inserted and fixed to the vein at its entrance by sterile silk sutures (4-0, Ethicon). Through the facial vein, the external jugular vein blood flow was maintained over the catheter, which aided thrombosis assessment. The remaining catheter was placed in the subcutaneous tissue, and the incision site was closed with an intradermal suture pattern.

2.8.2. Postoperative Recovery

Animals were continuously checked during the recovery period until they were able to maintain sternal recumbency before being transferred back to the animal facility. The rabbits were housed individually within their respective cages. Routine checks include temperature, respiratory rate, general mentation, feed consumption, and urine and feces monitoring. The site of the skin incision was checked daily for inflammation or infection. Meloxicam 0.6 mg kg⁻¹ (Metacam, Boehringer Ingelheim, Duluth, GA) was given for a total of 3 doses and 5 mg kg⁻¹ of enrofloxacin (Enroflox, Norbrook Inc, Lenexa, KS) was given daily for 4 d after surgery per IACUC requirements. After 7d, the rabbits were given 400 IU kg⁻¹ sodium heparin prior to euthanasia to prevent necrotic thrombosis. Administration of heparin before euthanasia is necessary to ensure the patency of the intravenous catheter (IV catheter in the auricular vein for short-term access) and is also required by the acceptable use policy (AUP). Lack of heparin administration prior to euthanasia can lead to an extended euthanasia process and distress in the animal. In this study, the animals were euthanized using a dose of 1 mL of pentobarbital sodium and phenytoin sodium (Euthasol, Virbac, Fort Worth, TX).

2.8.3. Analysis of catheters post in vivo experiment

The analysis of catheter after in vivo experiment was performed. The jugular vein was first isolated, then cut longitudinally to expose and remove the catheter carefully. Pictures were taken of the interior of the vein and the exterior of the catheter post-removal. Pictures of the blood-contacting part of the jugular vein were examined using Image J software to measure the thrombus area caused by catheter implantation.

The catheter was then cut into 1 cm sections for scanning electron microscopy (SEM). A section of the catheter was rinsed in PBS and placed in 2% glutaraldehyde overnight, and then dehydrated via sequential alcohol drying and Bis(trimethylsilyl)amine (HMDS) chemical fixation. The fixed catheter segments were imaged using a Thermo Fisher Scientific Teneo Field Emission SEM. Specimens were sputter-coated with 10 nm Au – Pd via a Leica EM ACE200 sputter coater (Buffalo Grove, IL). Surface imaging of the catheters was conducted with an accelerating voltage of 5 kV.

2.9. Statistical Analysis

All experiments were performed in multiple replicates as noted and results are presented as mean ± standard deviation and analyzed by one-way ANOVA using OriginPro 2020b software. The differences between samples were considered statistically significant if p < 0.05. Significance is denoted with symbols (*), with one symbol denoting p < 0.05, two symbols denoting p < 0.01, and three symbols denoting p < 0.001, NS=no significance.

3. Results and Discussion

3.1. Characterization of catheter coatings

The surface chemistry of coatings was characterized at the exterior surfaces of the catheter by XPS (Supporting information, Figure S1). Table 1 shows the elemental compositions of surfaces. In general, the control PU catheter surface without coating was composed of C, N, O and Si, and no F and P elements were detected, as expected. For PP-coated surfaces, elements F and P were detected, while the Si amounts were significantly reduced on both interior and exterior surfaces, suggesting that both surfaces were successfully coated with fluorinated PPs. Looking at the detailed elemental compositions of F, N and P, the PP samples contained nearly equal amounts of P and N, which is consistent with a polyphosphazene backbone, while the F contents were dependent on the polymers. X-OFP has a higher F level than TFE since X-OFP contains higher amounts of fluorocarbon content with the octafluoropentoxy side group, while TFE contains the trifluoroethoxy side groups.

Table 1.

Atomic elemental compositions (atom%) of coating surfaces analyzed by XPS.

Coating surfaces	Elemental Composition (atom%)
	C	F	N	O	P	Si
Bare PU	76.6	0.0	2.6	15.2	0.0	5.6
TFE	39.9	33.2	4.9	14.4	5.9	1.8
X-OFP	34.4	50.3	3.0	8.0	3.9	0.5

Surface wettability of catheter coatings was measured on the exterior surfaces and is presented as water contact angles (Table 2, Figure S2). The PU catheter is hydrophobic with contact angle of 96.7 ± 4.1° due to the silicone in the polymer. Fluorinated PP coatings are significantly more hydrophobic than the bare PU surfaces, with a contact angle of ~109° due to the fluorocarbon characteristics. No significant difference was observed among the polyphosphazene coatings. Surface wettability is influenced by many factors including surface chemistry, surface charge, chemical heterogeneity, and topography. As expected, both TFE and X-OFP are hydrophobic because of fluorocarbon groups in the polyphosphazene backbone. However, we did not observe significant differences in wettability between TFE and X-OFP. This is likely due to the similar fluorine contents in the two polymers. Based on chemical structures, the fluorine contents were calculated to be 33.3% (atom%) in TFE and 44.4% (atom%) in X-OFP. Considering that approximately 10% octafluoropentoxy groups were replaced with allylphenoxy side groups for crosslinking in X-OFP, the fluorine contents are therefore close in TFE and X-OFP. This should result in similar wettability.

Table 2.

Surface wettability and roughness of coating surfaces on catheters.

Coating surfaces	Water contact angle (°)	Roughness (nm)
		Rq	Ra
Bare PU	96.7 ± 4.1	22.2 ± 2.5	17.5 ± 2.2
TFE	109.2 ± 6.1	48.1 ± 4.9	36.9 ± 3.7
X-OFP	106.2 ± 5.0	188.5 ±15.9	130.0 ± 8.2

The surface topography of coating surfaces on catheters was characterized by AFM. The height images show that bare PU control is smoother than TFE or X-OFP, and no specific features were observed on bare PU surfaces. However, a significant difference in surface morphology was observed between TFE and X-OFP coatings. Small spheroidal crystal morphologies (spherulite) were observed in TFE coatings while no such features were seen on X-OFP surfaces (Figure 1). In the early study of physical properties of TFE, Masuko et al. observed similar crystal structures in TFE films using a 6-order Berek compensator and regarded that the crystallization of TFE was the result of the molecular chain orientation [40]. The roughness of surfaces was analyzed from AFM images using scan areas of 10×10 μm². Results showed that roughness values (R_q and R_a) of TFE and X-OFP coatings are about 2 and 9 times higher than the values of bare PU surfaces (Table 2, Figure S3). The high roughness of X-OFP coating is likely due to its chemical structure, which contains high fluorocarbon content with the octafluoropentoxy side group and crosslinkable allylphenoxy side group. The other reason for high roughness may be related to the viscosity of polymer solutions. TFE is easily dissolved in methyl ethyl ketone (MEK) at a concentration of 10% (w/v). X-OFP contains a high content of fluorocarbon groups and crosslinkable allylphenoxy side groups in the backbone and takes longer to be dissolved in MEK. X-OFP polymer solution with a concentration of 10% (w/v) is visibly more viscous than TFE. When dipping the catheters in polymer solution, TFE on the catheters drained off quickly, resulting in relatively smooth surfaces, while the slow draining of X-OFP coating results in a relatively higher roughness after dry.

Figure 1.

AFM topography images of catheter surfaces. (A) bare PU control, (B) TFE coated catheter, and (C) X-OFP coated catheter. Scale bar =2 μm. Height color scale: 500 nm.

The nanostructures of polymer coatings were further analyzed by AFM tapping mode over a scan area of 1×1 μm². AFM height images show similar topographies for TFE and X-OFP coating surfaces within the small scan area, and no specific features were observed (Figure 2). However, a significant difference in mechanical properties can be seen from phase images. Generally, positive shifts of phase angles of polymers represent hard domains in polymers and negative shifts indicate the soft matrix [41]. TFE shows narrow, linear shaped structures, while X-OFP shows wide, linear shaped domains. We attribute this to the different sizes of side groups on PP backbones affecting chain packing, where larger molecular fluorocarbon octafluoropentoxy side groups are attached on X-OFP backbone, compared to trifluoroethoxy side groups on TFE. More bright areas with high phase angle shifts were observed in X-OFP coating, representing the harder domains that indicate the linear shaped chains were crosslinked. Nanoindentation AFM measured the elastic modulus of coatings. Results show that the surface modulus of TFE is 0.037 ±0.09 GPa while the moduli of X-OFP are 0.14 ± 0.04 GPa and 0.40 ± 0.19 GPa before and after crosslinking, respectively. This demonstrates that the incorporation of octafluoropentoxy and allylphenoxy groups on the backbone significantly increased the mechanical properties of the polyphosphazene coatings, and crosslinking further increased the surface modulus. In our previous study of PP coatings cast on stainless steel, the increase in the surface stiffness significantly reduced Staphylococcus bacterial adhesion and inhibited biofilm formation [34], indicating the importance of surface mechanical properties on bacterial adhesion responses.

Figure 2.

AFM height and phase images of catheter coating surfaces show the topography and structure of PPs within scan area of 1 ×1 μm². Scale bar = 200 nm.

The adhesion of coating-to-substrate is another important characteristic of coatings to be evaluated on medical devices because any coating failure such as delamination or detachment will generate particulates with the potential to result in a stroke or embolic event. To evaluate the adhesion of TFE and X-OFP coatings, polymer solutions were cast on 316SS (stainless steel) surfaces. A standard test method described in ASTM D3359 was used to test the adhesion. The coatings on 316SS surfaces were crosscut and adherence measured with tape. The TFE coating shows flakes along the edges of cuts in large ribbons and some whole squares were detached as the tape was pulled off. The X-OFP coating, before being crosslinked, was also found to be flaked along the cutting edges, and about 15 to 35 % of the lattice was affected. However, the X-OFP coatings after crosslinking showed edges of the cuts that were completely smooth and none of the squares of the lattice were detached (Figure S4). In accordance with ASTM D3359, the adhesion strengths of TFE, X-OFPs before crosslinking and after crosslinking can be rated as 1B, 2B and 5B, respectively, where 5B represents the best adhesion-to-substrate. This result is consistent with our experience during experiments, where we observed some pieces of TFE coatings were separated from substrates, coming off both 316SS and PU catheters during soaking in solution, indicating poor adhesion to substrate. Meanwhile, the crosslinked X-OFP coatings remained intact on surfaces and no debris of X-OFP coatings was observed during long term exposure to physiological solutions.

The ASTM D3359 standard method provides a qualitative evaluation of the adhesion strength of the coating to substrate at the visual detection level. To further characterize the adhesion properties of the coating to the substrate, we studied the adhesion and delamination of coatings using nano-scratch method. Here, we used a polyurethane sheet as a substrate, similar to our eventual catheter material, to carry out the nano-scratch study. TFE coating showed radial cracking during the scratch test, while no cracking was observed on X-OFP coating (Figures 3A and 3B). The optical images show the delamination of the polymer coating, which is identified by the first appearance of a distinct uniform contrast in the images. This correlated well with observations in SEM images and EDS fluorine maps, as the polyurethane substrate does not contain fluorine (Figure S5). The critical load for delamination (Lc) was measured at 63.9 ± 6.5 mN for X-OFP coating, approximately 3 times of the Lc for TFE coating (Figure 3C). The higher the critical load, the better the adhesion. In summary, nano-scratching results show that X-OFP has better adhesion than TFE, consistent with the observation of coatings on 316SS plates using the ASTM standard method (ASTM D3359).

Figure 3.

Optical microscopy images of (A) TFE and (B) X-OFP coatings after nano-scratching, and (C) critical load at delamination for polymer coatings. The green arrows show the delamination and cracks along the edges of the scratches in TFE coating. Scale bar = 200 μm.

Polymer coating adhesion to material is a complex, multiscale phenomenon that is related to polymer chemistry and physical properties. Crosslinking, molecular bonding, chain entanglement, domain and network morphology, fracture or debonding may affect adhesion to materials [42]. In this study, X-OFP contains allylphenoxy side groups for crosslinking. After UV or thermal crosslinking, the chemical bonds form between the polymer chains of the coating and the substrate, or between polymer chains within the coating. These bonds create a three-dimensional network, effectively locking the coating in place and preventing it from separating from the substrate, improving the adhesion and cohesion strength of the coating. Both TFE and X-OFP are fluorinated polyphosphazene polymers with similar chemistry (see XPS data in Table 1) except for the different fluorocarbon side groups on backbones and crosslinking process. The AFM images show crystal-like morphology in TFE coating surface, but no such morphology observed in X-OFP surface (Figure 1). We hypothesize that the spherulite morphology of TFE may cause its poor adhesion of coating-to-substrate, however a deeper understanding of this is beyond the topic of this study. The other important reason for the strong adhesion of X-OFP coating to substrate is due to its crosslinking process which has been a main strategy to increase adhesion for many polymers [42]. The allylphenoxy group in X-OFP crosslinks the polymer chains and increases mechanical properties but also enhances the adhesion to substrate to improve the stability of catheter coatings. Results strongly suggest that X-OFP containing a long octafluoropentoxy side group as well as the allylphenoxy crosslinking side group significantly improves the adhesion of coating-to-substrate, thereby eliminating the risk of coating failure caused by delamination or detachment.

3.2. In vitro assessment of hemocompatibility and cytocompatibility of polyphosphazene coatings.

Hemocompatibility of blood-contacting biomaterials is one of the most important criteria for their successful application in clinics. In this study, we focused on the evaluation of thrombosis resistance of polyphosphazene coatings. Biomaterial-induced thrombosis is due to the activation of the intrinsic pathway of plasma coagulation cascade along with platelet-mediated reactions [43, 44]. Thus, we analyzed the coagulation time of plasma and platelet activation as the catheters were incubated with plasma or blood. We also measured the hemolysis of RBC which is another important aspect of hemocompatibility of biomaterials in contact with blood.

Figure 4A shows the coagulation time of human plasma incubated with catheters. The clots in control plasma (blank) appeared at ~25.8 min after triggering the plasma coagulation cascade, while the clots were formed significantly faster when PU control catheters were added (p <0.05). It is interesting to see that the clotting times for TFE and X-OFP coatings were longer than the blank control, particularly the coagulation time of plasma in contact with X-OFP, which increased to ~30.6 min, significantly longer than blank (p <0.01). The increase of coagulation time for TFE and X-OFP is likely due to the hydrophobicity of fluorocarbon chemistry of coatings which adsorbs proteins and inhibit the activation/propagation of coagulation factors such as FXII. Our previous study showed that the hydrophobic surface inhibited the apparent activity of FXII_a due to the adsorption of kallikrein, which activates additional FXII molecules to FXII_a in the reciprocal-activation reactions in the intrinsic pathway of plasma coagulation[45]. These results suggest that fluorinated polyphosphazene coatings are resistant to plasma coagulation.

Figure 4.

Hemocompatibility and cytocompatibility of TFE and X-OFP coated catheters, (A) human plasma coagulation time (B) platelet activation after blood contact with catheters for 1 h and 4 h, (C) hemolysis of red blood cells in blood incubated with catheters for 3h at 37°C, and (D) relative cell viability response to catheters as measured by MTT assay. A dashed line was drawn (70% viability) to indicate the reference line for cytotoxicity.

The platelet activation assessment was performed in human whole blood incubated with catheters at 37°C for 1h and 4h. The activation of platelets was analyzed by flow cytometry. Results show that platelet activation levels in blood were ~1.5% after 1h of interaction for all samples. No significant difference was observed among the samples. After 4h of incubation, PU catheters increased the platelet activation to ~10%, significantly higher than the Control (p<0.001), while the platelet activation for TFE and X-OFP coatings appeared higher than the Control, but no significant difference was observed (Figure 4B).

For assessment of hemolytic properties of coatings, catheters were incubated with human blood for 3 h at 37°C. The analysis of free hemoglobin shows that the hemolysis of RBC was low for all samples at ~0.32%, similar to the Control (without catheters) (Figure 4C). No significant differences were observed between blank control and the catheters, demonstrating that all catheters were nonhemolytic and polyphosphazene coatings did not induce RBC lysis.

The cytocompatibility of coatings was measured using 3T3 mouse fibroblast cells with MTT assay. Results show that the cell viability of all catheters was in the range of 90-105% (Figure 4D), higher than the cytotoxic reference line (70% viability), suggesting fluorinated polyphosphazene coatings are cytocompatible in accordance with ISO-10993-5 standards[46].

3.3. Human fibrinogen adsorption on polyphosphazene surfaces.

Protein adsorption to the biomaterial surface is the initial step after blood contact and mediates the subsequent platelet adhesion and activation. While blood contains a multitude of proteins with a wide range of different biological functions and activities, fibrinogen has been identified as one of the most important proteins involved in platelet adhesion/activation and thrombus formation[47]. Understanding fibrinogen adsorption is an important step in the mechanistic understanding of thrombosis caused by the implanted surfaces. Protein adsorption with human fibrinogen shows that the fluorinated PP materials reduced fibrinogen adsorption by ~25% compared to the PU control (Figure 5A). This is consistent with the early report of protein adsorption on TFE surfaces [19]. Welle et al. measured plasma protein adsorption including albumin, fibrinogen, and fibronectin on a variety of material surfaces and found TFE had the highest human serum albumin adsorption and the lowest adsorption of fibrinogen and fibronectin[19]. These last 2 proteins are related to thrombus formation and cellular attachment, respectively.

Figure 5.

(A) Human fibrinogen (2 mg/mL) adsorption on polymer surfaces for 2h at 37°C, (B) adhesion forces of human fibrinogen and polymer surfaces in PBS.

To further understand the interactions of fibrinogen and polymer surfaces, we measured adhesion forces of fibrinogen and polymer coatings using the AFM technique. The adhesion forces of fibrinogen and PU surface varied over a small range with the mean value at 1.56 ± 0.15 nN. However, the adhesion forces of fibrinogen and TFE or X-OFP varied over a larger range with the mean forces at 1.18±0.52 and 1.21±0.49 nN, respectively, both of which are statistically significantly smaller than the forces measured on PU surfaces (Figure 5B). This is consistent with the observation in protein adsorption (Figure 5A). There was no significant difference observed between TFE and X-OFP. Results strongly suggest that fluorocarbon chemistry, both the trifluoroethoxy or octafluoropentoxy side groups on polyphosphazene backbones, enhances the resistance to fibrinogen adsorption, thereby reducing platelet adhesion and activation, and finally inhibiting thrombosis formation.

3.4. In vitro test of blood thrombosis on catheter coatings in a blood flow loop.

A custom-made blood loop system was used to simulate the blood circulation and thrombosis formation on catheters in dynamic blood flow. A plethora of thrombi formations were observed on the PU control catheter surface, while significant reductions were seen on both TFE and X-OFP catheters. The catheter surfaces were imaged and analyzed to calculate the clot area using Image J software and the data was normalized against the PU control. As shown in Figure 6, the clot areas on TFE and X-OFP surfaces were significantly smaller than the PU controls, and the fluorinated PP coatings reduced approximately 80% of the clot area, suggesting that PP coatings improved the resistance to thrombus formation and significantly reduced clots on catheters. There was no significant difference in clot area observed between TFE and X-OFP, indicating the introduction of the octafluoropentoxy side group (instead of trifluoroethoxy in TFE) and the allylphenoxy for crosslinking didn’t decrease the hemocompatibility in vitro.

Figure 6.

(A) Representative images of catheters showing the clots formed on catheter surfaces and (B) clot areas on catheter surfaces in in vitro blood loop experiments (normalized to PU bare catheters)

3.5. In Vivo hemocompatibility assessments by a 7-day rabbit model

The above in vitro studies demonstrated the biocompatibility of fluorinated PP coatings on catheters. In order to demonstrate effectiveness in preventing blood clotting without causing systemic adverse effects in human, in vivo animal models can be used to evaluate devices and techniques in environments that simulate human physiology. In this study, a 7-day rabbit model was chosen to further study the hemocompatibility of catheters with polyphosphazene coatings and using the bare PU catheter as the control. The selection of rabbits as a model is useful because they are well suited from the standpoints of physiology and coagulation status. Using the same rabbit model, various polymeric materials have been analyzed and evaluated for hemocompatibility and resistance to microbial infections [48, 49].

Catheters were implanted in animals for 7 days, and then both the catheters and veins were explanted for antithrombogenicity evaluation. Images of the interior morphology of the vein with the catheter were taken (Figure 7A). The extent of thrombus formation was assessed by calculating the surface area of the clots that had been formed on the inner surface of the jugular vein. As shown in Figure 7A, the red color indicates the presence of clot formation. As analyzed by Image J, results show that the catheterization of the bare PU control resulted in a large clot in the vein, indicating the increased recruitment and entrapment of red blood cells in the clot. Comparing to PU controls, TFE and X-OFP coated catheters showed significantly reduced clot formation within the jugular vein by 72.6% and 78.5%, respectively (Figure 7B). In this study, the animals were given heparin prior to euthanasia. Heparin administration does not break down an existing clot already formed in the jugular, but lack of heparin administration can lead to false-positive thrombus formation results, as clot formation and adhesion can occur during the euthanasia process. All sample groups received the same heparin injection during termination, and we still observed a significant increase in clot formation for the Control groups.

Figure 7.

In vivo thrombosis assessment on catheter surfaces in a 7-day rabbit model. (A) Representative images of catheters of PU control, X-OFP, and TFE coatings after explantation. The jugular vein was cut longitudinally to show thrombus formation on the interior of the vessel (highlighted region), (B) Quantitative analysis of total clot area formed on the internal jugular vein for catheters (N=3, actual power value = 0.95)

In addition, we also measured the platelets, white blood counts, neutrophils and lymphocytes to assess whether the catheters caused any systemic adverse effects. In Figure S6, a reduction in platelet count can be observed in the PU control group after catheter implantation. Meanwhile, the other sample groups did not experience significant platelet count changes. This suggests the PU control group experiences more clot formation and therefore recruits and activates circulating platelets. No difference in WBC count was observed across the sample groups. We tested the neutrophil and lymphocyte composition in whole blood to see if any groups experienced heightened inflammation or immune response due to the implanted catheter. However, the PU control group showed a slight decrease in number of neutrophils. Meanwhile, a higher number of lymphocytes was observed for PU control, indicating potential elevation in inflammatory response. We can conclude that TFE and X-OFP coatings improved the hemocompatibility of the catheters and did not induce adverse effects.

Taken together, this study demonstrated the improved hemocompatibility and cytocompatibility of fluorinated PP coatings applied on clinical catheters. As discussed in the introduction, TFE has been extensively studied in vitro and in vivo in the past decades. Literature shows TFE was biocompatible and antithrombogenic in in vivo and pre-clinical studies [10, 21–24]. Through the above in vitro and in vivo measurements, we found that X-OFP is superior in terms of mechanical properties, the stability of coatings, and the adhesion strength of coating-to-substrate, while the hemocompatibility is comparable with TFE.

X-OFP utilizes the characteristics in the chemical structure of TFE and attaches the octafluoropentoxy side groups on backbones, which result in good thrombogenesis resistance and biocompatibility, but also incorporates the allylphenoxy functional groups for crosslinking to improve the adhesion strength to substrate and mechanical properties. PPs are hybrid polymers with flexible side groups linked to PP backbones. This provides various properties of polyphosphazene materials that are suitable for the applications in healthcare. The success of X-OFP will provide a platform to incorporate other different side groups to the backbones and generate new fluorinated polyphosphazene polymers having improved biocompatibility and mechanical properties for coating applications in blood-contacting devices such as catheters and stents.

4. Conclusion

A fluorinated polyphosphazene, X-OFP, was applied to central venous catheters as a coating. X-OFP contains a high amount of fluorocarbon content and is crosslinkable. The crosslinked X-OFP coating improved the mechanical properties including adhesion strength to substrates. Both in vitro and in vivo studies demonstrated that X-OFP has similar antithrombogenic performance to TFE but is better in terms of adhesion strength of coating-to-substrate. X-OFP can be a new generation of fluorinated polyphosphazene coatings for blood contacting medical devices with improved thrombogenesis resistance and importantly, improved stability under physiological conditions.

Supplementary Material

XPS survey spectra of PU control, TFE, and X-OFP coating catheters; Water contact angles of coated catheter surfaces; Surface roughness of coated catheters measured by AFM over a scan area 10×10μm². Adhesion test of TFE and X-OFP coatings on 316SS in accordance with ASTM D3359; Optical image, SEM, and EDS fluorine map of scratches on (A) TFE and (B) X-OFP coatings during nano-scratching; Platelet count, WBC count, Neutrophil % and Lymphocyte % in blood of rabbits before and after catheter implantation.

Statement of Significance.

Implant-induced thrombosis is a major complication of blood-contacting medical devices. This study demonstrated a new fluorinated polyphosphazene coating suitable for the medical device with the significant improvement of the biocompatibility of catheters. Compared to the traditional fluorinated polyphosphazene coating, poly[bis(trifluoroethoxy) phosphazene] (TFE), crosslinkable poly[bis(octafluoropentoxy) phosphazene] (X-OFP) contains a higher amount of fluorocarbon content with the octafluoropentoxy side group and is crosslinkable with the allylphenoxy side group. TFE and X-OFP were applied on central venous catheters as coatings. In vitro and in vivo studies demonstrated that X-OFP has a similar antithrombogenic performance as TFE, but its mechanical properties including adhesion strength of coating-to-substrate are significantly improved, thereby enhancing the stability of the coating.