Delivery of rivaroxaban and chitosan rapamycin microparticle with dual antithrombosis and antiproliferation functions inhibits venous neointimal hyperplasia

Abstract Neointimal hyperplasia is a complex process after vascular interventions, acute platelet deposition and smooth muscle cell proliferation both contributed to this process. There are still no perfect solutions to solve this problem. Rivaroxaban is a novel anticoagulant that has been widely used in clinic, it has a good pharmacological effects both in vivo and in vitro. Chitosan microparticle rapamycin (MP-rapa) was fabricated, interspaces of polyglycolic acid (PGA) scaffold were used as a reservoir of MP-rapa, and the scaffold was coated with hyaluronic acid rivaroxaban (MP-rapa-riva). Scanning electronic microscopy (SEM) photographs were taken and water contact angles were measured, rat inferior vena cava (IVC) patch venoplasty model was used; patches were harvested at day 14 and examined by immunohistochemistry and immunofluorescence. SEM photographs showed the microparticles rapamycin were inside the interspace of the scaffold, hyaluronic acid rivaroxaban was also successfully coated onto the surface of the scaffold. There was a thinner neointima, fewer proliferating cell nuclear antigen (PCNA) positive cells, fewer macrophages in the MP-rapa and MP-rapa-riva grafts compared to the control PGA graft. The result showed that this scaffold with dual anticoagulation and antiproliferation functions can effectively inhibit venous neointimal hyperplasia, although this is an animal experiment, it showed promising potential clinical application in the future.


Introduction
Neointimal hyperplasia is still the leading cause of graft failure after vascular interventions (Goel et al., 2012), it is a complex process including acute platelet deposition, inflammatory cell accumulation, and smooth muscle cell proliferation (Hristov and Weber, 2008;Lee and Roy-Chaudhury, 2009;Muto et al., 2007). Heparin, rapamycin or paclitaxel coated stents or balloons have been widely used in clinic; although these instruments contributed to a better patency rate, but none of these grafts showed a long term success in clinic (Kandzari et al., 2017;Cochrane Vascular Group, 2016;Tepe et al., 2017). Current protocols are mainly focusing on one step in the process of neointimal hyperplasia, like anti-thrombosis or anti-proliferation, this maybe the cause of the unsatisfactory result of the drug coated balloon and stent (Mehta et al., 2011). So, new methods are needed to decrease neointimal hyperplasia, especially in basic research. We recently showed a three-layered hydrogel patch with hierarchical releasing of PLGA nanoparticle drugs decrease neointimal hyperplasia in a rat inferior vena cava (IVC) venoplasty model, this hierarchical releasing system can significantly and effectively inhibit venous neointimal hyperplasia (Wei et al., 2022). But that drug delivery system is complex, so a simpler drug delivery system is needed.
Different grafts have been used in vascular surgeries, like the autologous vein, ePTFE and polyester, and pericardial patch . We previously showed therapeutic drugs can be conjugated to the pericardial patch , decellularized human saphenous vein patch . Polyglycolic acid (PGA) is widely used in vascular research, PGA felt is composed of PGA fibers and interspaces between them (Quint, 2020, these interspaces can be a reservoir for drug delivery (Liu et al., 2016). Chitosan particles are easy to make and are commonly used to delivery drugs (Liang et al., 2021, Hojnik Podrepsek et al., 2020, the chitosan particles can be stored into the interspaces of PGA fibers and slowly released. Rivaroxaban is a direct FXa (Xa factor) factor inhibitor and now is widely used in the treatment of venous thromboembolism disease (Kaplovitch et al., 2020;Kearon et al., 2016;Perzborn et al., 2005;Roehrig et al., 2005), rivaroxaban can also protect the oxysterol-induced damage and inflammatory activation of the vascular endothelium (Gorzelak-Pabis et al., 2021); but there is still no research on whether rivaroxaban can be coated to the vascular graft. A previous research showed rivaroxaban plays a role in mediating tissue remodeling and inflammation (Krupiczojc et al., 2008). Rivaroxaban treatment (15 mg/kg/d) can significantly reduce the maximal aortic diameter and suppress experimental abdominal aortic aneurysm progression in ApoE-deficient (ApoE -/-) mice (Ding et al., 2021). It can also attenuated calcium chloride-induced abdominal aortic aneurysms in mice by inhibiting abdominal aortic remodeling and reducing peripheral inflammation (Ding et al., 2021). In a rabbit model, rivaroxaban inhibits intimal hyperplasia and smooth muscle cell proliferation at the carotid anastomosis site (Akkaya et al., 2017).
Based on these previous researches, we hypothesized that chitosan microparticle rapamycin can be loaded into the interspaces of PGA scaffold, and rivaroxaban can be coated onto the surface of the PGA scaffold to decrease venous neointimal hyperplasia in a rat inferior vena cava (IVC) venoplasty model.

Scanning electronic microscopy (SEM) photograph and water contact angles (WCA) measurement
PGA scaffold, PGA scaffold filled with collagen-1, PGA scaffold filled with chitosan microparticle rapamycin and collagen-1 (MP-rapa), PGA scaffold filled with chitosan microparticle rapamycin and collagen-1 and coated with hyaluronic rivaroxaban (MP-rapa-riva) were observed their morphology under scanning electronic microscopy (SEM), and the water contact angles (WCA) were also measured as previously described (Wang et al., 2019).

PGA scaffold modification
The absorbable polyglycolic acid felt (0.5 mm) was bought from GuNZE medical, Japan; the felt was cut in to 1X1cm 2 for implantation or modification.
Chitosan (F1905050, Aladdin, China) was dissolved at 0.5 wt% in 1.5% acetic acid (H1808008, Aladdin, China), once dissolved adjust pH to 4.7. 0.5% tripolyphosphate (TPP, H1808008, Aladdin, China) crosslinker solution (1 mL, with 0.5 mg rapamycin) was added into chitosan solution (3 mL) at room temperature with constant stirring as previously described. Microparticles was then stabilized by a 30 min incubation at room temperature and collected by centrifugation at 13,000 × g at 10 °C for 30 min. Collagen-1 (C8061, Solarbio, China) solution was prepared at the concentration of 0.1%. The coating of hyaluronic acid rivaroxaban was carried out as we previously described . The samples were immersed in a hyaluronic acid (HA) solution for 15 minutes, then washed with phosphate-buffered saline. The HA coated samples were then immersed in a rivaroxaban solution (1 mL, 2 mg/mL) that had also been advance-activated in water-soluble carbodiimide solution for 15 minutes, then incubated for 6 hours. Rivaroxaban was given at a dose determined by previous studies (Gigi et al., 2012;Seki et al., 2017;Álvarez et al., 2018).

Animal model
The First Affiliated Hospital of Zhengzhou university authorized all research, and the National Institutes of Health guidelines for the care and use of laboratory animals (NIH Publication #85-23 Rev. 1985) were followed. For patch implantation, male Sprague Dawley (SD) rats (6-8 weeks old) were used as previously described . The inferior renal inferior vena cava (IVC) was exposed through a midline incision; on the anterior wall of the IVC, a 3 mm longitudinal venotomy was made approximately 2 mm below the level of the renal vein.
Microsurgical procedures were performed aseptically using a dissecting microscope (Nikon, Japan). Control patch (PGA scaffold filled with collagen-1), MP-rapa patch (PGA scaffold filled with chitosan nanoparticle rapamycin and collagen-1), MP-rapa-riva patch (PGA scaffold filled with chitosan nanoparticle rapamycin and collagen-1 and coated with hyaluronic rivaroxaban) were trimmed to 3 mm × 1.5 mm × 0.5 mm and implanted to the rat IVC using running 11-0 nylon sutures. The abdomen was closed with 5-0 polyester sutures after patch implantation. The patches were harvested for analysis at day 14 after patch implantation, since the neointimal thickness was consistent after day 7 in this model Bai et al., 2016). No immunosuppressive, antibacterial, or antiplatelet medications were administered at any time. Patches were also implanted subcutaneously in rat.

Hematoxylin and eosin (H&E) staining
Rats were anesthetized with 10% chloral hydrate (intraperitoneal injection), and tissues were fixed with transcardial perfusion of phosphate buffered saline (PBS) followed by 10% formalin. Patch was removed and fixed in 10% formalin for 12 hours. Tissues were then paraffined and sectioned (4-μm thickness). Tissue slices were deparaffinized and stained with H&E staining kit (Baso, Zhuhai, China). The neointimal thickness was the mean of measurements in three different areas.

Immunostaining
Tissue sections were deparaffinized and were heated in a citric acid buffer (pH 6.0, Beyotime, Shanghai, China) at 100 °C for 10 min for antigen retrieval, and then incubated with primary antibodies overnight at 4 °C. After 1 hour of incubation at room temperature with secondary antibodies, the sections were stained with DAPI (Solarbio, Beijing, China) or hematoxyline to identify cellular nuclei.

Statistics
The mean SEM is used to express the data. ANOVA and t-tests were used to establish statistical significance for these analyses (Prism 6; GraphPad Software, La Jolla, CA). P-values of less than 0.05 were considered significant.

Results
The rivaroxaban coated PGA scaffold soaked with chitosan microparticle was made as showed in the illustration photograph ( Figure 1). The PGA scaffold was soaked in the mixture of collagen-1 solution and chitosan microparticle rapamycin for 5 minutes, the mixture can be recognized clearly under microscope (Figure 2A), hyaluronic acid rivaroxaban can be successfully coated to the PGA scaffold ( Figure 2A). Scanning electron microscope (SEM) photographs showed chitosan microparticle rapamycin and rivaroxaban both on the surface and inside the PGA scaffold ( Figure 2A). Water contact angle (WCA) showed a large angle in the control group, PGA with chitosan microparticle (PGA + chitosan) group and PGA with chitosan microparticle rapamycin (PGA + rapa) group, but after coated with hyaluronic acid rivaroxaban (MP-rapa-riva), there was a much smaller water contact angle ( Figure 2B).
These scaffolds were then implanted subcutaneously in rats and harvested at day 14, in the control group, there was a thick new tissue capsuled the graft, while there was a much thinner tissue capsuled the MP-rapa and MP-rapa-riva grafts ( Figure 3A), chitosan microparticle rapamycin remained in the interspaces of PGA scaffold in the MP-rapa and MP-rapa-riva grafts ( Figure 3A). There was a smaller amount of new formed capillaries ( Figure 3A, 3C), smaller amount of CD68 positive cells ( Figure 3A, 3D) and PCNA positive cells ( Figure 3A, 3E) in the MP-rapa and MP-rapa-riva grafts; but there was no different between MP-rapa and MP-rapa-riva grafts ( Figure 3A, 3E).
We then implanted these patches to the rat IVC, all the animals survived and all IVCs were patent with no stenosis. Patches were harvested at day 14, hematoxylin and eosin (H&E) staining showed a thick new formed adventitia and neointima in the control grafts; there was a significantly thinner neointima in both of the MP-rapa and MP-rapa-riva grafts compared to the control grafts ( Figure 4A, 4B). The chitosan microparticle rapamycin can still be found in the interspaces in the MP-rapa and MP-rapa-riva grafts, there was also fewer cells infiltrated into the interspaces in the MP-rapa and MP-rapa-riva grafts ( Figure 4A, 4C). There was also a significantly thinner adventitia in the MP-rapa and MP-rapariva grafts ( Figure 4A, 4D). Immunofluorescence showed CD34 and nestin dual positive progenitor cell in the neointima in these three grafts ( Figure 4E).   There were CD31 positive cells on the luminal side of the neointima, immunohistochemistry showed a similar reendothelialization rate in these three grafts ( Figure 5A, 5B); there were also α-actin positive cells in the neointima ( Figure 5A). Immunofluorescence showed phospho-mTOR (p-mTOR) expression in all three grafts, and there was a significantly smaller amount of PCNA positive cells in the neointima in the MP-rapa and MP-rapa-riva grafts ( Figure 5C, 5D). M1 and M2 macrophages play a role in neointima hyperplasia, there were macrophages in the neointima in these three grafts, M2 type (CD68 and IL-10 dual positive, CD68 and TGM2 dual positive), M1 type (CD68 and iNOS dual positive, CD68 and TNF α dual positive) macrophages were interspersed in the neointima (Figure 6).

Discussion
In this research, we showed that rivaroxaban and chitosan microparticle rapamycin coated polyglycolic acid (PGA) scaffold inhibits venous neointimal hyperplasia after patch angioplasty in a rat IVC venoplasty model. Although this is a small animal research, this broadens the application of rivaroxaban.
After vascular interventions, antiplatelet and anticoagulation drugs are used to prevent platelet accumulation and thrombus formation (Liu et al., 2019, Haybar et al., 2019. Neointimal cell, especially smooth muscle cells proliferation also plays an important role in the neointimal hyperplasia, so rapamycin or paclitaxel coated grafts are also widely used to inhibit neointimal cell proliferation . Recent studies showed programmed death-1 antibody or inhibitor coated vascular graft can also inhibit neointimal hyperplasia after patch angioplasty in rat . IL-33 antibody absorbed plant leaf patch can be used to inhibit venous neointimal hyperplasia in rat . But these interventions were all focused on one step in the neointimal hyperplasia; this may be the cause that the treatment of neointimal hyperplasia is not satisfactory in clinic. Although we developed a hierarchical hydrogel releasing system, but this system was complex. In this research, we fabricated this PGA scaffold patch which has dual antithrombosis and antiproliferation functions to decrease neointimal hyperplasia; this dual functional scaffold is much easier to make compared to the hierarchical hydrogel patch.   Figure 6. M1 and M2 macrophages in the neointima after rat ivC patch venoplasty at day 14. Photographs of the immunofluorescence of the neointima, first row, CD68(green), il10 (red) and DAPi (blue); second row, CD68(green), TgM2(red) and DAPi (blue); third row, CD68 (green), iNOS (red) and DAPi (blue); forth row, CD68(green), TNFα (red) and DAPi (blue); scale bar, 100 μm; n = 3.
Heparin is still a widely used parenteral anticoagulant for prevention and treatment of thrombotic events, but heparin-induced thrombocytopenia (HIT) is a severe complication in clinic (Arepally and Ortel, 2006;Warkentin et al., 1998). Several clinical studies showed rivaroxaban may be an alternative option to heparin in patients with a history of HIT (Krauel et al., 2012;Sartori et al., 2015;Tardy-Poncet et al., 2015). Consistent with our research, other studies also showed that thrombosis and neointimal proliferation were significantly inhibited with antithrombotic agents such as FXa inhibitors (Iwatsuki et al., 2004) and P2Y12 inhibitors (Patil et al., 2010). These findings suggest that coagulation and platelet activation after vascular injury are critical for neointimal hyperplasia in animal models (Morishima and Honda, 2018).
Theoretically, hyaluronic acid rivaroxaban coating on the surface of the scaffold would release first after implantation to play a role of anticoagulation, we previously showed heparin coated decellularized human saphenous vein patch can decease neointimal thickness in a rat patch angioplasty model . Then the chitosan microparticle rapamycin would release from the interspaces of PGA scaffold to inhibit cell proliferation, this chitosan microparticle rapamycin has been widely investigated in basic researches (Chiesa et al., 2018, Yuan et al., 2008, we also used collagen as an adjuvant to secure chitosan microparticle rapamycin in the interspaces of the scaffold (Sareethammanuwat et al., 2020, Păun et al., 2020. Since we have previously showed rapamycin can inhibit venous neointimal hyperplasia in several researches, so we did not spend too much efforts on the function of rapamycin . We used rat IVC patch venoplasty model in this experiment because rivaroxaban is commonly used in the venous system, and the neointima is much thicker in the rat IVC venoplasty model compared to the aortic patch angioplasty model. There were also limitations in this research: Firstly, the observation time is short (14 days) in this preliminary study, longer time of observation is needed to confirmed this dual functional vascular patch; secondly, the function of this coating technique need to be tested in large animal (pig or sheep) models; thirdly, more commonly used ePTFE and Dacron grafts should also be tested.

Conclusion
We developed a new dual functional vascular patch to decrease thrombus formation and cell proliferation in a rat IVC venoplasty model. This is a preliminary small animal experiment, but this idea may have a potential clinical application.