All animal experiments were approved by the Animal Ethics Committee of Anhui Medical University (Approval No. LLSC20160052). 20–25 g mice were purchased from the Animal Experiment Center of Anhui Medical University. All mice were bred in the Research and Experimental Center of the Second Hospital of Anhui Medical University. The temperature and humidity were controlled in the feeding environment, with a 12 h cycle of light and darkness, and food and water were readily available. Mouse mononuclear macrophage leukemia cells RAW 264.7 were cultured in DMEM (Hyclone, Invitrogen, United States) containing 10% FBS (Gibco, United States), 1% Glutamax and 1% Sodium Pyruvate (Invitrogen, United States), and were kindly provided by Stem Cell Bank, Chinese Academy of Sciences (Shanghai, China). Mouse brain vascular pericytes MBVP were kindly provided by Mingzhou Biological Technology (Ningbo, China), and were cultured in DMEM containing 10% FBS. The cells were cultured at 5% CO₂ and 37°C. MBVP cells were identified as PDGFRβ⁺ cells via immunocytochemistry and could be used as a model of PDGFRβ⁺ pericytes in vitro.

Each mouse was given water fasting before surgery. The mice were put in prone position after satisfactory anesthesia with isoflurane (induction 4%, maintenance 2%). After shaving the dorsal surface of the mouse and disinfecting the skin, T10 segment of spinal cord was exposed. The T10 spinal cord was clamped with the calibrated Dumont #5 forceps (11252-20, Fine Science Tools, Germany) for 5 s (McDonough et al., 2015). The incisions were sterilized and stitched layer by layer. The mice were routinely treated with anti-infection and auxiliary urination nursing after SCI.

RAW 264.7 cells were plated in 6-well plates at a density of 2 × 10⁵/ml and cultured overnight. After serum deprivation for 24 h, RAW 264.7 cells were treated with LPS (100 ng/ml, Beyotime Biotechnology, China) plus IFNγ (20 ng/ml, Beyotime Biotechnology, China) for M1 polarization or IL-4 (20 ng/ml, Beyotime Biotechnology, China) for M2 polarization, and the cells were cultured for 24 h. Then, the cell culture medium was replaced with serum-free DMEM for another 24 h. Supernatant was collected and centrifuged (1200 rpm, 5 min) as conditioned medium including conditioned medium of M0 (CM0), CM1, and CM2. WB and immunocytochemistry were used to identify the macrophage polarization model, and ELISA was used to identify the conditioned medium.

The mice underwent thoracotomy after satisfactory anesthesia. After cardiac perfusion with PBS and 4% paraformaldehyde (4% PFA, Servicebio, China), spinal cord containing the injured core (5 mm) was removed and fixed in 4% PFA overnight. Then, the tissue blocks were dehydrated, made transparent, dipped in wax, embedded, and sliced to 6 µm thickness serial sections using a microtome (RM2235, Leica, Germany). The sections were deparaffinized and rehydrated in turn. Antigen retrieval was performed using the microwave method, followed by washing with PBS and blocking in 5% donkey serum (SL050, Solarbio, China) in PBS with 0.3% Triton-X100 for 1 h at room temperature. The primary antibodies were incubated at 4°C overnight: Goat anti-PDGFRβ (5 µg/ml, AF1042-SP, R&D Systems, United States), Mouse anti-Mac2 (1:100, GB12246, Servicebio, China), Rabbit anti-iNOS (1:100, AF0199, Affinity, United States) and Rabbit anti-CD206 (1:100, ab64693, Abcam, China). The secondary antibodies were incubated at 37°C in dark for 1 h: Donkey anti-Rabbit Alexa Fluor 488, Donkey anti-Mouse Alexa Fluor 594, Donkey anti-Goat Alexa Fluor 488 and Donkey anti-Goat Alexa Fluor 594 (1:500, A-21206, A-21203, A-11055, A-11058, Invitrogen, United States). Finally, the anti-fluorescence quencher (P0126, Beyotime Biotechnology, China) was used to seal the sections. The representative images were obtained using fluorescence microscope (Axio Scope A1, Zeiss, Germany) by comparing and processing all images of the same light intensity and filtering.

Cells were fixed with 4% PFA for 10–15 min and blocked for 30 min at room temperature. Primary antibodies were diluted in blocking solution and applied for overnight at 4°C: Goat anti-PDGFRβ (5 µg/ml, AF1042-SP, R&D Systems, United States), Rabbit anti-iNOS (1:100, AF0199, Affinity, United States), Rabbit anti-CD206 (1:100, ab64693, Abcam, China) and Rabbit anti-Laminin (1:100, 23498-1-AP, Proteintech, China). Secondary antibody incubation as shown above, and blue counterstaining (DAPI, C0065, Solarbio, China) showed nuclei. Finally, the anti-fluorescence quencher was used to seal the sections. The representative images were obtained as above. Image J (NIH, United States) was used for quantitative analysis.

For cell proteins, cells were washed with PBS and lysed using RIPA buffer (P0013B, Beyotime Biotechnology, China) supplemented with Protease Inhibitor Cocktail and Phosphatase Inhibitor (04693124001, 04906845001, Roche, Switzerland). For spinal cord tissue proteins, after cardiac perfusion with PBS, the spinal cord (3 mm) containing the injury core was removed and lysed with lysis buffer mentioned above. The supernatant was collected as total protein. Protein concentrations were determined using bicinchoninic acid (BCA) kit (P0012S, Beyotime Biotechnology, China) and diluted to 1 mg/ml for each sample. 20 μg of protein was denatured at 100°C for 10 min in LDS Sample Buffer (NP0007, Invitrogen, Unites States) before being resolved on SDS-PAGE electrophoresis and transferred onto a Polyvinylidene difluoride (PVDF) membranes. Blots were blocked in 5% nonfat milk and probed with primary antibodies against at 4°C overnight: Rabbit anti-iNOS (1:2000, AF0199, Affinity, United States), Rabbit anti-CD206 (1:1,000, ab64693, Abcam, China), Goat anti-PDGFRβ (0.25 µg/ml, AF1042-SP, R&D Systems, United States), Mouse anti-PDGFB (1:500, SC-365805, Santa Cruz, United States), Mouse anti-PDGFD (1:500, 14075-1-AP, Proteintech, China), Rabbit anti-Laminin (1:1000, 23498-1-AP, Proteintech, China), Mouse anti-β-Tubulin (1:10000, T0023, Affinity, Unites States) and Mouse anti-GAPDH (1:1000, ab8245, Abcam, China). The membranes were incubated with secondary antibodies for 1 h at room temperature: Goat anti-Mouse (1:10000, A4416, Sigma, United States), Goat anti-Rabbit (1:10000, A0545, Sigma, United States) or Rabbit anti-Goat (1:10000, A8919, Sigma, United States). The protein band signals were obtained using Tanon 5200 system (Tanon, China). Image J (NIH, United States) was used for quantitative analysis.

The PDGFB ELISA kit (KE10034, Proteintech, China) was used to detect the concentration of PDGFB in the conditioned medium of different polarized macrophages. The supernatant was collected and centrifuged (1200 rpm, 5 min) for detection immediately, according to the kit’s instruction.

The marker pen was used to draw 3 horizontal lines evenly on the back of the 6-well plate at intervals of 1 cm. MBVP cells were plated in 6-well plates at a density of 2 × 10⁵/ml and cultured until the cells covered the entire plate bottom. The wound was created using a sterile 200 μl pipette tip. After washing off the exfoliated cells, macrophages conditioned medium, serum-free DMEM containing recombinant mouse PDGFB protein (10 ng/ml, NBP2-53416, Novus, United States) or serum-free DMEM containing SU16f (10 μm, 3304, R&D Systems, United States) were added to the 6-well plate to regulate MBVP cells migration for 72 h. The pictures of scratch wounds were taken at 0 and 72 h and assessed using the Image J software.

The 24-well plate containing 8 μm chamber (Corning, United States was used to detect the migration ability of cells. MBVP cells were inoculated to the upper chamber at a density of 1 × 10⁶/100 μl. According to the experimental requirements, macrophages conditioned medium or PDGFB (10 ng/ml, NBP2-53416, Novus, United States) were added to the lower chamber to regulate MBVP cells migration, or SU16f (10 μm, 3304, R&D Systems, United States) was added to the upper chamber to block PDGFRβ. After 24 h in an incubator at 37°C and 5% CO₂, a cotton swab was used to wipe off the cells in the inner layer of chamber. After washing with ddH₂O water twice, the chambers were fixed by 4% PFA at room temperature for 20 min. After washing with ddH₂O water twice, crystal violet (C0121, Beyotime Biotechnology, China) staining was performed for 15 min. The cells were observed under the microscope and counted.

The data were presented as means ± standard deviation. Multiple-group comparisons were statistically analyzed with one-way analysis of variance (ANOVA) followed by the Tukey method or nonparametric test using the Prism 8.3.0 software (GraphPad, United States). p < 0.05 was considered significant.

To further confirm the spatiotemporal distribution of macrophages and PDGFRβ⁺ pericytes after SCI, the goat anti-PDGFRβ antibody was used to specifically label PDGFRβ⁺ pericytes while the mouse anti-Mac2 antibody was used to specifically label blood-derived macrophages rather than microglia (Wang et al., 2015). The immunofluorescence staining showed that there was no obvious aggregation of macrophages and PDGFRβ⁺ pericytes in the normal spinal cord. Macrophages and PDGFRβ⁺ pericytes were scattered at the injury core at 3 dpi while significantly aggregated at the injury core at 7 dpi. Finally, PDGFRβ⁺ pericytes surrounded macrophages to form fibrotic scar at 14 dpi (Figure 1). During 3–14 dpi, PDGFRβ⁺ pericytes gradually migrated and aggregated to the injury core, and the spatiotemporal distribution between macrophages and PDGFRβ⁺ pericytes was closely related (Figure 1). These suggested that 3–7 dpi is a critical period for PDGFRβ⁺ pericytes to migrate and aggregate to the injury core, and macrophages may be involved in the migration of PDGFRβ⁺ pericytes after SCI.

To investigate the effect of macrophages polarization on PDGFRβ⁺ pericytes migration, anti-iNOS antibody and anti-CD206 antibody were used to specifically label M1 and M2 macrophages respectively to observe the macrophages polarization after SCI. The immunofluorescence staining showed that the macrophages (Mac2⁺) were mainly of type M2 with CD206⁺ and iNOS⁻ in injury core at 3 and 7 dpi (Figure 2B) while the macrophages were mainly of type M1 with iNOS⁺ and CD206⁻ in injury core at 14 dpi (Figure 2A). It was interesting to note that macrophages were mainly type M2 in the critical period of 3–7 dpi during when PDGFRβ⁺ pericytes migrated and aggregated to the injury core (Figures 1, 2B). These suggested that M2 macrophages may be involved in the migration of PDGFRβ⁺ pericytes after SCI.

To further investigate the effect of M2 macrophages on PDGFRβ⁺ pericytes migration, RAW 264.7 cells were used as the blood-derived macrophages model and MBVP cells as the PDGFRβ⁺ pericytes model in vitro. The results of immunocytochemistry suggested that MBVP cells were PDGFRβ⁺ cells and could be used as the PDGFRβ⁺ pericytes model in vitro (Figure 3A). In addition, we induced and identified M1 or M2 polarization of macrophages according to previous reports (Kigerl et al., 2009; Hiroi et al., 2013; Fan et al., 2020). The results of Western blot showed that compared with other groups, M1 macrophages significantly expressed iNOS while M2 macrophages significantly expressed CD206 (p < 0.0001 for iNOS, p < 0.001 for CD206, Figures 3B–D). Besides, immunocytochemistry was used to further confirm the reliability of macrophages polarization model. The results showed that compared with the other two groups, the fluorescence intensity of iNOS of M1 macrophages was significantly increased while that of CD206 of M2 macrophages was significantly increased (Figure 3E). These suggested that the PDGFRβ⁺ pericytes model and macrophages polarization model could be used for the follow-up study.

To further investigate the effect of macrophages polarization on PDGFRβ⁺ pericytes migration, the conditioned medium of different polarized macrophages (CM0, CM1, and CM2) were extracted for culture of pericytes. The scratch test showed that after adding M2 conditioned medium, the wound closure rate of PDGFRβ⁺ pericytes was significantly higher than that of other groups (p < 0.001 or p < 0.0001, Figures 4A,C). Besides, transwell test was used to further show that the number of pericytes passing through membrane increased significantly after adding M2 conditioned medium compared with other groups (p < 0.001 or p < 0.0001, Figures 4B,D). Compared with Sham group and CM1 group, the number of pericytes passing through membrane increased significantly after adding M0 conditioned medium (p < 0.001), but this effect was much weaker than M2 conditioned medium. These suggested that M2 macrophages could promote PDGFRβ⁺ pericytes migration in vitro, possibly due to the secretion of certain cytokines.

PDGFRβ⁺ pericytes and extracellular matrix including Collagen, Fibronectin and Laminin together surround macrophages to form fibrotic scar after SCI (Soderblom et al., 2013). It has been shown that Collagen and Fibronectin are mainly secreted by PDGFRβ⁺ pericytes after SCI while the main cellular origin of Laminin is unclear (Soderblom et al., 2013; Zhu et al., 2015b). The results of Western blot showed that after adding M2 conditioned medium, the expression level of Laminin in pericytes was significantly higher than that in other groups (p < 0.001, Figures 5A,B). Besides, immunocytochemistry was used to further show that the fluorescence intensity of Laminin in pericytes cultured in M2 conditioned medium was significantly higher than that of other groups (p < 0.01, Figures 5C,D). These suggested that Laminin may also be mainly secreted by PDGFRβ⁺ pericytes, and M2 macrophages can promote PDGFRβ⁺ pericytes to secrete extracellular matrix in vitro.

PDGFRβ is a specific marker for PDGFRβ⁺ pericytes while PDGFB can activate pericyte surface PDGFRβ and regulate pericyte migration in the retinal tissue. Therefore, we further investigated whether PDGFB/PDGFRβ pathway also plays an important role in the process of M2 macrophages promoting PDGFRβ⁺ pericyte migration. Our results showed that the expression level of PDGFB increased significantly at 3 and 7 dpi compared with Pre (p < 0.01 and p < 0.05, respectively), while decreased significantly at 14 dpi compared with 3 and 7 dpi (p < 0.001 and p < 0.01, respectively, Figures 6A,B). The expression level of PDGFD was significantly decreased at 3, 7, and 14 dpi compared with Pre (p < 0.0001, Figures 6A,C). In addition, the expression level of PDGFRβ was increased significantly at 7 and 14 dpi compared with Pre and 3 dpi (p < 0.001, Figures 6A,D), consistent with immunofluorescence staining that PDGFRβ⁺ pericytes aggregated to the injury core (Figure 1). These preliminarily suggested that PDGFB rather than PDGFD may be involved in PDGFRβ⁺ pericytes migration after SCI.

Secondly, the expression level of PDGFB was detected after macrophage polarization in vitro. The results of Western blot showed that the expression level of PDGFB in M2 macrophages was significantly higher than that in other groups (p < 0.01 and p < 0.001, respectively, Figures 6E,F). Besides, the results of ELISA showed that the concentration of PDGFB in M2 macrophages culture medium was significantly higher than that in other groups (p < 0.01, Figure 6G). These furtherly suggested that PDGFB may be involved in M2 macrophages promoting PDGFRβ⁺ pericytes migration after SCI.

To investigate the effect of M2 macrophages secretion of PDGFB on PDGFRβ⁺ pericytes migration and its possible mechanism, SU16f (Andersen et al., 2015; Jiang et al., 2017; Chatterjee et al., 2019) was used to specifically inhibit PDGFRβ and then the effect of exogenous PDGFB and M2 macrophages conditioned medium on PDGFRβ⁺ pericytes migration were observed in vitro. The scratch test showed that compared with Sham group, M2 conditioned medium and exogenous PDGFB significantly promoted the wound closure rate of pericytes (p < 0.0001) while the promotion effect was obviously eliminated by SU16f (p < 0.0001, Figures 7A,C). Besides, transwell test also showed that after adding M2 conditioned medium or exogenous PDGFB, the number of pericytes passing through membrane increased significantly compared with other groups while the promotion effect was significantly eliminated by SU16f (p < 0.0001, Figures 7B,D). These suggested that M2 macrophages promote PDGFRβ⁺ pericytes migration via PDGFB/PDGFRβ pathway in vitro.