The treatment of femoral neck fracture using VEGF-loaded nanographene coated internal fixation screws

Purpose Previous studies have proved that vascular endothelial growth factor (VEGF) has a dual role in the promotion of new bone formation and blood vessel repair during fracture healing. However, how to introduce VEGF to a fracture site safely and effectively is still a challenge. This study aimed to prepare a VEGF-loaded nanographene coated internal fixation screw and to evaluate its effects in the treatment of femoral neck fracture. Methods Nanographene coated screws were prepared by direct liquid-phase exfoliation of the graphite method, and the surface characteristics were observed through scanning electron microscopy (SEM). VEGF was loaded on nanographene coatings through physical adsorption, and the VEGF controlled release was examined by ELISA. Then a canine femoral neck fracture model was built to examine both the angiogenic and osteogenic properties of the VEGF-loaded coated screws. X-ray, micro-CT-based microangiography, and histopathologic evaluation were used to assess the fracture healing progress. Results The results demonstrated that nanographene could load VEGF effectively, and the accumulative release of VEGF clearly increased during the entire testing period (9 days) without burst release. In canine fracture models, the results of X-ray, microangiography, and histopathologic examination proved that the speed of fracture healing, new bone formation area, and revascularization of the fractured femoral heads in the VEGF-loaded coated screws groups were significantly higher than in the control groups. Conclusion Our study proved that VEGF-loaded nanographene coated screws were effective in the treatment of femoral neck fracture and prevention of avascular necrosis of femoral head.


Results
The results demonstrated that nanographene could load VEGF effectively, and the accumulative release of VEGF clearly increased during the entire testing period (9 days) without burst release. In canine fracture models, the results of X-ray, microangiography, and histopathologic examination proved that the speed of fracture healing, new bone formation area, and revascularization of the fractured femoral heads in the VEGF-loaded coated screws groups were significantly higher than in the control groups.

Conclusion
Our study proved that VEGF-loaded nanographene coated screws were effective in the treatment of femoral neck fracture and prevention of avascular necrosis of femoral head. PLOS

Introduction
Femoral neck fracture has two major complications, avascular necrosis of femoral head and the non-union of the fracture. While the rate of non-union recently decreased apparently with the improvement of the internal fixation techniques, the rate of ANFH remains the same [1]. The occurrence of ANFH is primarily related to its unique anatomical structure, the artery circle of branches of the medial circumflex femoral artery and lateral femoral circumflex artery [2]. ANFH may occur if the femoral head blood supply is not restored in a timely manner after fracture [3]. The application of vascular repair-related growth factors is now considered to be a more effective way to treat femoral neck fracture and reduce the incidence of ANFH because of its ability to promote early recovery of the blood supply of femoral heads after fracture [4,5]. Vascular endothelial growth factor (VEGF) is currently considered the most important regulatory factor of angiogenesis involved in various vascular repair processes [6,7]. It plays an essential role in fracture healing [5,[8][9][10][11]. In the inflammatory phase of fracture healing, VEGF can recruit macrophages to the fracture site and regulate angiogenesis. VEGF participates in the regulation of the blood vessels and bone formation in intramembranous ossification and endochondral bone ingrowth in endochondral ossification. Given the important role of VEGF in both bone repair and vascular repair, we plan to load VEGF safely and effectively into the femoral head and hope it can be effective in the treatment of femoral neck fracture and the prevention of ANFH.
As the thinnest nanomaterial, graphene has only a single atomic layer of thickness and can be applied as a coating on any material surface [12]. Compared with other nanomaterials, graphene has an extremely high specific surface area, and both sides can absorb drugs and achieve high-dose drug loading [13][14][15]. In our study, we applied VEGF-loaded nanographene coated screws to the treatment of femoral neck fracture and evaluated whether it can simultaneously promote osteogenesis and angiogenesis in fracture healing.
In the evaluation of experimental results, we used a new contrast agent for the first timesubnanometer barium sulfate for canine femoral microangiography. After the image reconstruction of micro-CT, the evaluation of bony and vascular repair in fracture healing can be analyzed from two-dimensional and three-dimensional levels.

Preparation of nanographene coated internal fixation screws
Nanographene was prepared by the direct liquid-phase exfoliation of graphite. Fifteen milligrams of natural graphite was mixed with 5 ml disperse solvent (1:1 isopropyl alcohol/H 2 O) in a 7.5-ml glass flask. After agitation for 10 minutes, the suspension underwent ultrasonic treatment for 4 hours (ultrasonic frequency, 40 kHz). The dispersion was centrifuged twice for 10 minutes at 1000 rpm at room temperature. The supernatant was collected and a stable graphene suspension was finally obtained. The internal fixation screws (Ti6Al4V, Wego, China) were immersed in nitric acid solution for 1 week in advance and then soaked in graphene supernatant for another week. After that, screws were removed and adequately rinsed by deionized water. The nanographene coated screws were prepared completely. The morphology of the surface of the coated screws was examined by scanning electron microscopy (SEM, JEOL Japan).

In vitro cytotoxicity evaluation
The extracts of the nanographene coated screws were prepared according to ISO 10993-12.
Steps are briefly described as follows. Sterile nanographene coated fixation screws were completely immersed in 6 ml endothelial cell medium (ECM) with 5% fetal bovine serum (FBS, Gibco, Thermo Fisher Scientific, Waltham, MA, USA) and osteoblast medium (Scien-Cell, San Diego, CA, USA) and placed in an incubator at 37˚C for 72 hours. Coated screws were removed from the ECM and osteoblast medium, and the extracts were centrifuged for 5 minutes at 1000 rpm. The supernatants were sterilized through a 0.22-μm filter and refrigerated at 4˚C and used within 1 week. The human umbilical vein endothelial cells (HUVECs) (Allcells, Alameda, CA, USA) and osteoblasts (Type culture collection of the Chinese Academy of Sciences, Shanghai, China) were used to evaluate the cytotoxicity of the coated screws. HUVECs and osteoblasts were seeded at a density of 3 × 10 3 /well in 96-well plates and cultured with 100 μl coated screw extract or special medium as control groups, for 4 days. The cells were then stained every 24 hours with 10 μl Alamar Blue dye and examined by microplate reader (wavelength, 450 nm).Each group was measured with n = 6 samples for the in vitro cytotoxicity evaluation, with 3 repeat measurements.

VEGF loading and in vitro release kinetics
The 1-cm thread side of the nanographene coated screws and screws without coating were immersed in 20 ug/ml VEGF solution (dissolved in phosphate-buffered saline [PBS]) for 12 hours at 4˚C. As the control group, the same length nanographene coated screws were immersed in phosphate-buffered saline (PBS) for 12 hours at 4˚C.The screws were then washed three times with distilled water, lyophilized, and placed in a 15-ml sterile tube. The tube was added with 10 ml PBS (pH 7.4) and placed in a shaking water bath at 37˚C and shaken at 100 rpm/min.

Released VEGF biological activity test
A VEGF-loaded nanographene coated screw was dissolved in 10 ml ECM (ScienCell, Carlsbad, CA, USA) with 5% FBS (Gibco) and placed in the same VEGF releasing condition in step 2.3 for 9 days. The supernatant was collected, sterilized by a 0.22-μm filter, and stored at 4˚C (used within 1 week); 11.6% Matrigel (Becton, Dickinson, and Company, UK) was thawed at 4˚C overnight. Ninety-six-well plates and micropipette tips were precooled on ice. In each well, 50 μl Matrigel was added and incubated at 37˚C for 30 minutes to solidify; 4,000/100 μl HUVECs were added on the Matrigel and incubated at 37˚C in an atmosphere of 5% CO 2 and 95% air. After 12 hours of incubation, the HUVECs were imaged using an inverted microscope (Leica, Wetzlar, Germany). Three fields were chosen at random, and the length of the capillary-like tubes was measured by Image-Pro Plus 6.0 software.

Animal experiment design and surgical procedure
All experimental animal surgery procedures and feeding standards were approved in accordance with the guidelines of the Care and Use of Laboratory Animals of the Zhongshan Hospital of Fudan University, Shanghai, China. Ten healthy adult (male and female) Beagle dogs, weighing 15-18 kg, were provided by the Laboratory Animal Center of Zhongshan Hospital of Fudan University. They were housed individually in the same room and could move freely in cages under conditions at 18-25˚C,25-75% humidity, and 12h lighting (6:00-18:00). The room and cages were cleaned daily. And during the cage cleaning time, the dogs were taken outside of cages the reduce their stress. Animals were provied with 230-250g of solid food daily and filtered tap water and libitum. The operations would only be performed while the animal had no fear appearance, ate regularly and slept well. The left and right posterior limbs were treated with VEGF-loaded nanographene coated screws (experimental side) and nanographene coated screws without VEGF (control side) separately on the same dog. All surgery procedures were performed under basic anesthesia by injecting 10% pentobarbital sodium (25 mg/kg) intravenously. The fracture procedures are briefly described as follows. After general sterilization, a curved incision, approximately 10 cm, was made in the hip followed by the subcutaneous soft tissue separated step by step. The hip joint capsule was exposed and opened. Then a chisel was used to create a fracture at the midpoint of the femoral neck. The fracture was reduced and fixed temporarily with a 1.5-mm Kirschner wire. Then a VEGF-loaded nanographene coated screw was implanted into the femoral head through femoral torque and the other femoral head with a nanographene coated screw that was not VEGF loaded. All operative procedures were performed by the same surgeons. Postoperatively, all experimental dogs received routine care and were injected penicillin (800,000 units) intramuscularly for 3 successive days to prevent infection. For pain relief, meloxicam (0.2 mg/kg on the day of surgery, and there after the dose of 0.1 mg / kg every day for 1 week). General well-being, wound healing, and postoperative complications were observed daily. Sutures were removed after 2 weeks.

Radiological examination
Under basic anesthesia, hip anteroposterior X-ray examinations were performed on all experiment animals at day 1 and at 8 and 12 weeks post surgery. The radiographic results were assessed by two radiologists who were blinded to the study.

Microangiography
Preparation of the contrast agent consisted of 4% gelatin-30% subnanometer barium sulfate suspension. 1) 4% gelatin solution:120 g gelatin (Sigma) was added in 3000 ml PBS and placed in a shaking water bath at 37˚C for 12 hours; 2) 900 g subnanometer barium sulfate was added in 3000 ml 4% gelatin solution and placed in a shaking water bath at 45˚C for 1 hour; and 3) the suspension in step 2 was filtered three times by sterile gauze and used within 24 hours.
Under basic anesthesia, the dogs were fixed on the operating table in the dorsal position after skin preparation. A median abdominal incision (approximately 15 cm) was made followed by the skin, linea alba, and retroperitoneal tissue cut in a step-by-step fashion. The abdominal aorta and inferior vena cava were exposed and inserted with a vessel cannula, respectively. Two hundred and fifty milliliters of 0.9% saline with 12,500 U heparin was injected into the abdominal aorta for the heparinization of lower limb blood supply. Then, approximately 3000-5000 ml saline was injected into the abdominal aorta to replace the lower limb blood. After that, the experiment animals were sacrificed by injecting an overdose of anesthetics. One thousand milliliters of 4% paraformaldehyde was inputted into the abdominal aorta slowly. The 4% gelatin-30% subnanometer barium sulfate suspension was injected into the abdominal aorta rapidly with certain pressure. The infusion was stopped when there was an outflow of white liquid on the vena cava side. Then the abdominal aorta and vena cava were closed. To solidify the contrast agent sufficiently, all of the specimens were stored at 4˚C overnight. The femoral heads were removed and fixed in 4% paraformaldehyde for 1 week at 4˚C before micro-CT scanning.

Micro-CT scanning and image reconstruction
After calibration, micro-CT (Healthcare Locus SP, General Electric, Fairfield, Connecticut, USA) scanning (46 μm resolution, scanning time 18 minutes) was performed on specimens.
Reconstructed images, maximum-intensity projection (MIP), and Isosurface images were generated by Micro View software (General Electric). MIP is an overall projection of the femoral head that can display the whole two-dimensional distribution of vessels within the femoral head, whereas Isosurface exhibits a three-dimensional image. By selecting the appropriate threshold (3000 HU), the data of vessels within the femoral head can be extracted and reconstructed to a three-dimensional angiographic image through Isosurface.
After micro-CT scanning, the specimens were cut in half and then embedded in paraffin for histopathological analysis.
Analysis of microstructure and bone mineral density of femoral head By manually choosing the uniform region of interest (ROI) (Fig 1), the trabecular thickness (Tb.Th) and bone mineral density (BMD) of femoral head, excluding the screw canal and cortical bone, were quantitatively analyzed by advanced bone analysis (ABA) in Micro View software. Each group was measured with 5 ROI for further analysis, with 3 repeat measurements.

Histopathological evaluation
Dogs were sacrificed at 12 weeks after microangiography examination by injecting an overdose of pentobarbital sodium intravenously. Samples of bilateral femoral heads were harvested and fixed in 4% neutral-buffered formalin, decalcified in 10% ethylenediaminetetraacetic acid (EDTA) for further histopathological measurements after the surrounding soft tissue was removed clearly. Then the samples were sectioned and stained using hematoxylin and eosin staining (HE), Masson's trichrome, and CD31 immunohistochemical staining. Stained sections were examined under an inverted microscope. Average Tb.Th, newly formed bone area (%), and microvessel density (MVD) [16] were obtained using Image-Pro Plus 6.0 and analyzed by two blinded pathologists.

Statistics
All data are expressed as mean ± standard deviation (x±s). Statistical analysis was performed using SPSS 16.0 statistical software. Differences between the experimental and control groups were compared using an independent samples t test. P<0.05 was defined as statistically significant.

Results
Nanographene coated screws morphological observation SEM images of internal fixation screws with and without nanographene coating were obtained at 1000x magnification (Fig 2). The surface of screws without coating was performed as a loose and porous structure. After coating of the screws with nanographene, a thin layer of laminated material could be observed on the surface of the screws. Nanographene coating did not have an obvious change on the morphology of the screws.

In vitro cytotoxicity evaluation
A comparison of the cytotoxicity of the coated screw extracts with ECM or osteoblast medium were carried out by Alamar blue assay. As shown in Figs 3 and 4, there was no significant difference between the proliferation rate of the study groups (P>0.05). The results indicated that nanographene coating did not increase the cytotoxicity of the internal fixation screws.

In vitro release kinetics
The cumulative release rates of VEGF from nanographene coated screws in PBS are shown in Fig 5 and Table 1. The release of VEGF was sustained for 9 days with a cumulative release amount up to 9.51 ± 0.23 ng. There was no burst release observed in the whole process. In the control group, there had very little burst release of VEGF in the uncoated screws in the first few hours and degraded within 3 days.

Released VEGF promotes angiogenesis in vitro
To further elucidate the bioactivity of VEGF released by nanographene coated screws, Matrigel tube formation was investigated in vitro. When cultured with cumulative 9-day extracts of VEGF-loaded nanographene coated screws for 12 hours, the average length of tubular  The release of VEGF was sustained for 9 days with a cumulative release amount up to 9.51 ± 0.23 ng. There was no burst release observed in the whole process. In the control group, there had very little burst release of VEGF in the uncoated screws in the first few hours and degraded within 3 days. structures formed by HUVECs was significantly increased (P<0.05), as compared with the control groups (Figs 6 and 7). The result demonstrated that the released VEGF still had bioactivity.

X-ray examination
After surgery, all objects had clear fracture lines and received good reduction (Fig 8A and 8B). By 8 weeks postoperatively, the fracture lines in the experimental groups disappeared and the Femoral neck fracture treatment and nanographene coated screws bone density was similar to the bone tissues around the fracture site (Fig 8C) compared with the fracture line blurred in the control groups and a lower bone density in the fracture site ( Fig  8D). By 12 weeks, the fracture was completely healed in the experimental groups with no apparent difference in the bone density of the fracture site as compared with the nearby normal bone tissue (Fig 8E). In the control groups, the fracture lines were almost invisible as well, and the bone density of the fracture site was still lower than the normal bone tissue around it (Fig 8F).

2D reconstructed images
After 12 weeks of surgery, the MIP images showed that there were several microvessels in the femoral heads of the experimental groups (Fig 9A) connected with the vessels of great trochanter of the femur through the femoral neck, and they distributed more evenly and broadly compared with the control groups. In the control groups (Fig 9B), the numbers and branches of microvessels in the femoral heads were lower than the experimental groups, and there was no obvious microvessels of the femoral heads connected with the surrounding vessels.

3D reconstructed images
The 3D images of the microvessels within the femoral heads were reconstructed and customized to red by Isosurface of Micro View. The spatial relationship between the bone microstructure and microvessels could be observed from the coronal and sagittal planes (Fig 10A and  10C). By choosing the appropriate threshold (3000 HU), microvessels could be extracted and displayed directly without the surrounding bone structure (Fig 10B and 10D). The results of the 3D-reconstructed images of the microvessels further demonstrated that the numbers and branches of microvessels in the femoral heads of the experimental groups (Fig 10A and 10B) were more than the control groups (Fig 10C and 10D).

Advanced bone analysis of femoral heads
Bone structural parameters are shown in Table 2. The BMD of the femoral heads of the experimental groups was significantly higher than the control groups (408.6 ± 39.9 vs. 310.6 ± 28.7, P<0.05). The Tb.Th of the two groups had no statistically significant difference despite the fact that there was an increasing trend in the experimental groups.

Histopathological evaluation
Compared with the control groups, the width and area of the bone trabecula of the experimental groups showed in HE staining were higher at 12 weeks postoperatively (Fig 11A and 11B). However, the results had no statistically significant difference (P>0.05) ( Table 3). The images of Masson's trichrome stain (Fig 11C and 11D) showed that there was more red-stained  The BMD of the femoral heads of the experimental groups was significantly higher than the control groups.
The Tb.Th of the two groups had no statistically significant difference despite the fact that there was an increasing trend in the experimental groups. https://doi.org/10.1371/journal.pone.0187447.t002 Femoral neck fracture treatment and nanographene coated screws mature bone observed in the femoral heads of the experimental groups than the control groups. New bone formation area (%) in the experimental femoral heads ( Table 4) was more than the control groups (p<0.05). CD31 immunohistochemical staining (Fig 11E and 11F) demonstrated that the number of CD31 positive cells and microvessels of experimental groups ( Table 5) was significantly higher than the control groups (P<0.05).

Discussion
Previous studies have proven that the severity of fracture displacement and internal fixation avascular necrosis incidence were closely related: the higher the degree of fracture displacement, the more complete damage to the blood supply of the femoral head, and the higher the incidence of postoperative ANFH [17,18]. A number of recent studies have shown that the choice of surgical treatment for femoral neck fracture should be based on the remaining blood supply of the femoral head after injury rather than relying solely on the severity of the fracture displacement [19,20]. This suggests that early recovery of the femoral head blood supply could be the key to preventing ANFH after internal fixation. We chose VEGF as the delivery drug for the treatment of femoral neck fracture. VEGF, as the most important regulatory factor of angiogenesis, is involved in all processes of angiogenesis [6,7,21]. VEGF also played a crucial role in the progress of bone growth as well as fracture repair [10,11,22]. Our experiment results demonstrated that the microvessel density and BMD of the femoral heads in the experimental sides increased compared with the control sides, indicating that VEGF-loaded nanographene coated screws could accelerate fracture healing and promote blood vessel repair. To load VEGF into the fracture site effectively, we prepared nanographene coating as the drug delivery system. Nanographene, as a novel biomaterial, has been proved to have good biocompatibility [23]. Also, Kim et al. [24] and Li et al. [25] reported that nanographene could be conductive of the adhesion and differentiation of adipose-derived MSCs and neural stem cells. Nanographene material can be applied to the surface of any shape of materials as a single atom sheet to enhance the materials' performance, such as biocompatibility and cell adhesion [26]. As reported nanographene has shown the advantage of little cytotoxicity while with the ability sustained release for many days [27,28].Our experimental results showed that the nanographene could be successfully applied as a coating on the surface of the internal fixing screws and that it did not affect the biocompatibility and surface characteristics of the screws. The nanographene coating could absorb a certain amount of VEGF, demonstrating effective sustained release of VEGF. As a simple and effective drug delivery system, nanographene coating could deliver VEGF to the fracture site directly in the treatment of femoral neck fracture. Microangiography is one of the most intuitive methods in the assessment of the blood supply recovery of the femoral head after injury. We carried out microangiography by abdominal aorta and inferior vena cava catheter, which is a clear, simple approach that is easy to repeat. The diameter of subnanometer barium sulfate particles is smaller than 500 nm and can be easily injected into microvessels of bone. The CT value of the subnanometer barium sulfate, as the vascular contrast agent, is greater than 3000 HU. However, the CT value of cancellous bone CT is generally less than 1500 HU. This CT value difference could be distinguished between microvessels and bone microstructure in femoral heads. Thus, samples could be scanned by micro-CT without decalcification, so as to keep all the vascular and bony information in the femoral heads. After image reconstruction, microvessels and bone microstructure within the femoral head could be displayed simultaneously and analyzed from any two-dimensional and three-dimensional level through Micro View software. In addition, compared with Microfil [29] and Angiofil [30], subnanometer barium sulfate is a highly economical and suitable agent for the angiography of medium and large animals. Our experiment results showed that subnanometer barium sulfate could be successfully used as a contrast agent for the microangiography of beagles. After micro-CT scanning, the distribution of microvessels could be displayed clearly from two-and three-dimensional aspects.
The main limitations in our study were that we reduced the fixation time to 12 weeks. If it could be extended to 6 months, there might be a better result. In our study, we testified that nanographene coating had good biocompatibility, but we did not study the rest of material characterisation of it, such as adhesivity.Otherwise,this animal model could not reflect the real condition that is seen in human fracture healing process and may account for the excellent healing of experimental animals in this study.

Conclusion
In conclusion, our experiment proved that VEGF-loaded nanographene coated internal fixation screws could promote new bone formation and the recovery of blood supply in the fracture healing process. As a new contrast agent, combined with micro-CT scanning, subnanometer barium sulfate, as a new contrast agent, could clearly show microvessels in femoral heads and make the quantitative analysis of microvessels and bone microstructure possible.