All experiments were approved by the Ethics Committee of West China Hospital of Stomatology (Chengdu, China). All experiments were performed in accordance with relevant guidelines and regulations and executed in compliance with the ARRIVE guidelines (https://arriveguidelines.org). Briefly, 8–10-week-old, male, C57BL/6 mice (20–25 g) were provided by Dossy Experimental Animals Company (Chengdu, China). The animals were exposed to a 12-h light-and-dark cycle. Mice were allowed for a standard diet with free access to food and water. The mice were randomly allocated into the following four groups: control group that consisted of 31 mice (n = 31) receiving phosphate-buffered saline injection; capsaicin group that consisted of 31 mice (n = 31) receiving capsaicin injection; capsazepine group that consisted of 31 mice (n = 31) receiving capsazepine injection; capsazepine + CGRP group that consisted of 28 mice (n = 28) receiving capsazepine and CGRP overexpression lentivirus injection.

A mouse model of bone defect was established in accordance with a previous study (Monfoulet et al., 2010). General anesthesia was administered using a mixture of 80 mg/kg ketamine-HCl and 10 mg/kg xylazine HCL intraperitoneally. The animal was fixed with the prone position, and the fur on distal surface of right and left legs was shaved and disinfected with povidone-iodine. Reaching the distal surface of the femur, the soft tissue was separated, the periosteum was released, and the bone was exposed. A standard bone defect measuring 1-mm in diameter was created using round drills in both the right and left femur of each mouse. The mice were killed post-surgery either on day 7 or day 14. The femurs were isolated for subsequent analysis.

The drug was administered intrathecally to mice by direct lumbar puncture as described previously (Pan et al., 2018). Control group received topical applications of PBS. For mice in capsaicin group, lumbar puncture was performed using a 5-μL Hamilton microsyringe (Hamilton, United States) fitted with a 30-G needle to deliver 2 μL capsaicin (2 mg/mL, Abmole, United States) into the subdural space of the spinal L4–L6 levels in non-anaesthetized mice before surgery (Jiang et al., 2023). For mice in capsazepine group, 0.5 μL capsazepine (2 mg/mL, Abmole, United States) was injected by lumbar puncture per day. Successful puncture was indicated by a reflexive lateral flick of the tail. For mice in capsazepine + CGRP group, CGRP overexpression lentivirus 10 μL (2.88 × 10⁸ TU/mL, GeneCopoeia Co., China) was injected around the femoral bone defect locally.

Using a 30-gauge needle, we injected 0.5-µL Fluoro-gold (FG, Fluorochrome, CO, United States) into the bone defect area to retrogradely label DRG neurons by a Hamilton microsyringe (Hamilton, United States). After 48 h, L4-L6 DRGs were harvested for immunofluorescence (IF) staining of DRG sections.

After general anesthesia, the skin was separated from the neck down to the glabella. The chest was opened and expose the heart. The right atrium was clipped. 100 U/mL heparin solution, 4% paraformaldehyde and Microfil^® reagent (MV-122, Flow Tech Inc, United States) were perfused successively via the left ventricle and incubation overnight at 4°C. The femurs were stripped, fixed in 4% paraformaldehyde for 48 h and decalcified in 10% EDTA for 4 weeks, and then scanned for micro-CT.

For Micro-CT analysis, four animals were used at each timepoint in each group. The perfused femurs were imaged on a micro-CT (SCANCO 50, Switzerland) using 10-μm resolution with a voltage of 90 kV and a current of 88 μA. The region of interest (ROI) was defined as the defect of the femurs to evaluate the vascular microstructure. Vascular volume (VV), vascular volume fraction (VVF), vascular number (V*N), vascular thickness (V*Th), and vascular separation (V*Sp) were calculated to evaluate vascularization of the defect.

For Immunohistochemistry evaluation, five animals were used at each timepoint in each group. Femurs were fixed in 4% PFA for 48 h and decalcified in 10% EDTA (pH 7.0) for 4 weeks. The samples were dehydrated in ethanol concentrations (70%–95%), washed in xylene and embedded in paraffin blocks to obtain 5 μm thick cross-sections. After deparaffinization, immunohistochemistry staining was done. To prevent non-specific binding, 3% bovine serum albumin (BSA) was incubated with all sections for 30 min. Sections were incubated with primary antibodies, anti-CGRP (1:100, Santa Cruz, United States), anti-CD31 (1:100, Santa Cruz, United States), and anti-alpha smooth muscle actin (α-SMA) (1:100, Santa Cruz, United States) overnight at 4°C. Sections were rinsed three times with PBS before being incubated for 1 h at room temperature with HRP labeled secondary antibody (1:500, Abcam, United States). Finally staining was shown with diaminobenzidine (DAB) substrate, followed by counterstaining with a hematoxylin. The mean optical densities (Integral optical density/area of interest) were calculated by ImageJ 1.46 software (NIH, United States) for further semi-quantitative analysis of CD31, a-SMA and CGRP expression.

For immunofluorescence, L4-L6 DRGs were extracted after injecting retrograde tracer for 72 h and fixed in 4% paraformaldehyde for 24 h and embedded in paraffin. Sections were incubated for 20 min at room temperature in a blocking solution containing 3% BSA. Sections were incubated with primary antibody anti-TRPV1 (Santa Cruz, United States) overnight at 4°C and incubated with Cy3 labeled secondary antibody (1:100, Servicebio, China) for 50 min at room temperature, then incubate in TSA (1:500, Servicebio, China) for 10mins at room temperature in the dark. After antigen retrieval with microwave oven for 10 min, second primary antibody, anti-CGRP (Santa Cruz, United States) was applied overnight at 4°C and incubated with FITC labeled secondary antibody (1:100, Servicebio, China) for 50 min at room temperature.

Bone tissue from the distal and proximal 1 mm of the femoral defect and L4-L6 DRGs from five animals in each group were collected at days 7 and 14. Total RNA was isolated from tissues and DRGs using the FastPure Cell/Tissue Total RNA Isolation Kit V2 (Vazyme, China). RNA was reversely transcribed into cDNA using HiScript III RT SuperMix for qPCR (+gDNA wiper) (Vazyme, China) and amplified using Taq Pro Universal SYBR qPCR Master Mix (Vazyme, China). Real-time PCR was carried out through a LightCycler 480 system (Roche Diagnostics, Germany). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as the internal control for normalization. The primer sequences were listed in Table 1.

Proteins of specimens from three mice in each group were extracted by RIPA buffer (Sigma-Aldrich, United States) followed by high-speed centrifugation (12,000 g/min at 4°C for 10 min) at day 7. The protein concentration was determined using a BCA protein quantitative kit (Beyotime, China). 20 μg of total protein was separated by SDS-PAGE, electrophoresed, and transferred onto a PVDF membrane (Sigma-Aldrich, United States). The membrane was blocked with 5% BSA/TBST for 1 h, then incubated with the following primary antibodies: anti-CGRP (1:1,000, Santa Cruz, United States), anti-VEGFA (1:1,000, Abcam, United States) and anti-GAPDH (1:5,000, Signalway Antibody, United States) at 4°C overnight. The membrane was washed with TBST thrice and then incubated with HRP labeled secondary antibody (1:5,000, Abclonal, China) for 2-3 h at room temperature. The membrance was treated with the ECL blotting reagents (Zenbio, China). The protein bands were visualized with QuantityOne software (Bio-Rad, United States). Before incubating other primary antibody, the blots were rinsed in distilled water for 5 min, then soaked in the antibody stripping buffer (Servicebio, China) for 60 min to remove ECL substrates.

All data were expressed as mean ± SD. Statistical differences were determined using one-way ANOVA with Tukey’s multiple comparison test, p < 0.05 was considered statistically significant. Graphs were produced by GraphPad Prism Software (GraphPad Prism 9.0, United States).

Capsaicin serves as a potent agonist of the TRPV1 channel. This experiment found the same results in that capsaicin activated the TRPV1+ sensory neurons in the DRG (Caterina et al., 2000; Li et al., 2019; Lee et al., 2020). To further investigate the potential co-expression of TRPV1 and CGRP in DRG nerve fibers, Fluoro-gold (FG) was injected at the defect to retrograde label the upstream neurons in the DRG and we conducted immunofluorescence staining (Figure 1). Our findings indicated that TRPV1 (Red) immunolabeling and CGRP (Green) expression significantly overlapped in the FG-positive neurons (Blue), indicating that DRG nerve fibers express both CGRP and TRPV1. After activating through capsaicin, both TRPV1 and CGRP expression were elevated in the neuron. Furthermore, Capsazepine, a TRPV1 antagonist, the expression of TRPV1 and CGRP were weakened simultaneously in comparison to the control group. Taken together, these findings suggest that TRPV1 and CGRP are co-expressed in DRG nerve fibers, and capsaicin enhanced the expression of both TRPV1 and CGRP.

To observe the vascular formation in a femur defect, mice were perfused with Microfil^® reagent. Multi-perspective 3D-reconstructed images revealed that the capsazepine group had less and thinner vessels, with decreased interconnectivity and increased separation compared to the control group; however, the capsaicin group showed the opposite result (Figure 2A). Quantification of the microstructural parameters revealed that the VVF, V.N, V.Th, and VV in the defect were higher in the capsaicin group than the control group, while the V.Sp was significantly lower than the control group V.Sp, lower V.Sp represented that vessels were closer to each other, and the area was more densely vascularized. The capsazepine group showed lower levels of the VVF, V.N, V.Th, and VV and a higher level of V.Sp than the control group. We restored CGRP levels in localized bone defect by injecting CGRP lentiviral vector after capsazepine injection. The local CGRP rescue treatment enhanced vascularization in femur defect. The level of VVF, V.N, and V.Th was improved and V.Sp was lower than the capsazepine group (Figures 2B–F). These results suggest that activation of TRPV1 by capsaicin improves vascular formation, while capsazepine reduces it. These findings have significant implications for understanding the role of TRPV1 in vascular formation.

The immunohistochemical assay was conducted to observe the expression of CD31, α-SMA, and CGRP in newly formed tissue (Figures 3A–C). The intensities of CD31 and α-SMA immunostaining were substantially greater in the capsaicin group than in the control group, but significantly lower in the capsazepine group. Local rescue by CGRP promoted CD31 and α-SMA expressions. Moreover, the CD31 and α-SMA intensities continuously increased from 7 to 14 days in each group, while the trend of differences between groups remained consistent (Figures 3D, E). In addition, CGRP was strongly expressed in the capsaicin group, but showed lower expression in the capsazepine group than the control group. Interestingly, CGRP expression was lower in the capsaicin group at day 14 than at day 7 but was still the highest among all three groups (Figure 3F). These findings suggest that the TRPV1 agonist regulates angiogenesis of the bone defect area through its downstream gene CGRP in vivo.

To evaluate the expression of angiogenesis-related genes, we extracted mRNA and protein from bone tissue surrounding the defect. Specifically, we analyzed the expression of CGRP, VEGF, and CD31 in these samples. PCR results showed a significant upregulation in the capsaicin group, whereas the capsazepine group exhibited a downregulation of CGRP, VEGF, and CD31 mRNA expression. After local injection of CGRP lentivirus, the VEGF and CD31 mRNA expression level was elevated in the capsazepine group (Figures 4A–C). The expression of CGRP, VEGF, and CD31 continuously increased from 7 to 14 days in each group, while the trend of differences between groups remained consistent. The result of western blot demonstrated that CGRP and VEGF protein expressions were simultaneously increased in the capsaicin group and both decreased in the capsazpine group (Figure 4F). Meanwhile, sensory neurons in DRG were also extracted. The results demonstrated that TRPV1 inhibition exhibited a co-repressive effect on mRNA expression of TRPV1 and CGRP, while the activation of TRPV1 showed the positive effect (Figures 4D, E). Taken together, this suggested that neuronal TRPV1 could regulate the expression of angiogenic genes in defect and possibly relate to CGRP expression.