The TN animal model in this study was established as previously described (Luo et al., 2012, 2020; Lin et al., 2018). Briefly, rats were anesthetized with pentobarbital sodium (40 mg/kg, i.p.), and a curved skin incision was made above the right eye. Subsequently, the orbital fascia was gently stripped laterally to reveal the infraorbital groove of the maxillary bone to expose the right infraorbital nerve. A small plastic filament (1 mm in diameter) was slowly and carefully inserted into the intracalvarium through the inferior orbital fissure at a depth of 1.2 cm to compress the trigeminal nerve root. Rats in the sham operation group underwent the same procedure without nerve compression. The incision was closed using 6–0 silk sutures.

The baseline orofacial mechanical stimulation thresholds of rats were tested before the operation, and after the operation, rats were tested for mechanical allodynia using von Frey filaments (Aesthesio, Ugo Basile, Italy; Figure 1A). All rats were habituated to behavioral testing for 3 days prior to baseline testing. von Frey filaments were applied to the vibrissa pad of the rats to determine the orofacial mechanical allodynia threshold. Each von Frey filament was applied five times. Stimulation always began with the filament that produced the lowest force and was stopped when the threshold was found within the vibrissa pad.

On postoperative days (PODs) 7, 14, 21, and 28, rats in the two groups were injected with 2 ml/kg 2% Evans blue (Merck Ca, NJ, United States) through the caudal vein. An hour later, rats in each group (n = 3 or 4/group) were anesthetized with sodium pentobarbital (40 mg/kg, i.p.) and perfused with cold 0.1 M PBS through the left ventricle to remove intravascular Evans blue. The TREZ was separated and homogenized in 400 μl of formamide. The samples were then incubated at 60°C for 24 h, followed by centrifugation at 10,000 × g for 20 min. The optical density value of the supernatant was measured at a wavelength of 620 nm by a microplate reader (BioTek, VT, United States). The extravasation of Evans blue in the tissue was calculated using a standard curve.

Furthermore, Evans blue dye distribution in TREZ tissue sections was observed by confocal microscopy. Rats were anesthetized with sodium pentobarbital (40 mg/kg, i.p.) and perfused with 0.1 M PBS. TREZ segments were dissected and cryoprotected with 30% (W/V) sucrose in 0.1 M PBS overnight at 4°C. Trigeminal nerve roots were horizontally sectioned using a cryostat (Leica CM1950, Heidelberger, Germany) to obtain consecutive 10 μm sections. Then, the sections were incubated with 3% bovine serum albumin (BSA) for 30 min to reduce autofluorescence. The extravasated Evans blue-albumin complex yielded red fluorescence in unstained TREZ tissue sections. The processed sections were examined and photographed under a Leica laser confocal microscope (Leica TCS SP8, Heidelberger, Germany).

Rats in both groups (n = 6) were deeply anesthetized with sodium pentobarbital (200 mg/kg, i.p.) on PODs 7, 14, 21, and 28. Total RNA was extracted in TRIzol reagent (Invitrogen, TX, United States, Cat# 15596018). Subsequently, a High-Capacity cDNA Reverse Transcription Kit (Vazyme, Nanjing, China, Cat# R223) was used to convert RNA into cDNA, while a SYBR-Green Real-Time PCR Kit (Vazyme, Cat# Q311) and a Bio-Rad MiniOption thermocycler (Bio-Rad, CA, United States) were used for detection. Primer sequences are provided in Table 1. The relative expression levels of RNA were evaluated using the 2^−ΔΔCt method.

Rats in both groups were deeply anesthetized with sodium pentobarbital (200 mg/kg, i.p.) on PODs 7, 14, 21, and 28 and were then perfused through the left ventricle with 4% paraformaldehyde phosphate buffer (pH 7.4). The trigeminal nerve segment from the junction of the trigeminal nerve root with the brainstem to the trigeminal ganglion was dissected and cryoprotected with 30% (W/V) sucrose in 0.1 M PBS overnight at 4°C. Trigeminal nerve roots were horizontally sectioned using a cryostat (Leica CM1950, Heidelberger, Germany) to obtain consecutive 10 μm sections.

For immunohistochemical staining, sections were washed three times in 0.1 M PBS for 10 min and blocked in 3% BSA for 30 min. The sections were subsequently incubated with primary antibodies following the removal of blocking solution without washing. The following primary antibodies were used: rabbit polyclonal anti-Caspr (1:100; Abcam, Cambridge, United Kingdom, Cat# ab34151, RRID: AB_869934), mouse monoclonal anti-myelin basic protein (MBP; 1:1,000; Abcam, Cat# ab62631, RRID: AB_956157), rabbit polyclonal anti-glial fibrillary acidic protein (GFAP; 1:1,000; Proteintech, IL, United States, Cat# 16825-I-AP), and mouse monoclonal anti-PAR1 (1:200; Santa Cruz, Texas, United States, Cat# sc-13,503, RRID: AB_2101175). Biotinylated goat anti-mouse IgG (1:200; Vector, CA, United States, Cat# BA-9200, RRID: AB_2336171), donkey anti-rabbit Alexa Fluor 488 (1:1,000; Invitrogen, CA, United States, Cat# A-21206, RRID: AB_2535792), CyTM3-conjugated streptavidin (1:500; Jackson ImmunoResearch, PA, United States, Cat# 016-160-084, RRID: AB_2337244), and DAPI (1:1,000; Beyotime, Shanghai, China, Cat# C1002) were also used. The immunohistochemical sections were imaged and analyzed with a Leica laser confocal microscope.

Rats in the TN and sham groups were sacrificed after anesthetization with sodium pentobarbital (200 mg/kg, i.p.), after which the trigeminal nerve root was quickly removed and rapidly frozen in liquid nitrogen. The samples from different groups were homogenized in extraction buffer (100 mM Tris, pH 7.4) containing 2 mM phenylmethanesulfonylfluoride and 10 mg/ml aprotinin. The solutions were centrifuged at 12,000 × g for 10 min at 4°C. The protein concentration of the supernatants was determined using a BCA Protein Assay Kit (Beyotime, Cat# P0010). Protein samples (10–20 μg) were loaded on 8 or 10% polyacrylamide gels for SDS–PAGE and were then transferred onto polyvinylidene difluoride membranes. The membranes were blocked with 5% nonfat milk in Tris-buffered saline containing 0.1% Tween 20 for 1 h at room temperature and incubated with the following primary antibodies overnight at 4°C: rabbit monoclonal anti-NF155 (1:1,000; Cell Signaling Technology, MA, United States, Cat# 15035, RRID: AB_2798693), mouse monoclonal anti-PAR1 (1:500; Santa Cruz, Cat# sc-13,503, RRID: AB_2101175), mouse polyclonal anti-β-tubulin (1:10,000; Bioworld, MN, United States, Cat# BS1482M), and rabbit polyclonal anti-GAPDH (1:10,000; Bioworld, Cat# AP0066, RRID: AB_2797448). Then, the tissues were incubated with goat anti-mouse IgG (H + L)-HRP (1:5,000; Bioworld, Cat# BS12478, RRID: AB_2773727) or goat anti-rabbit IgG (H + L)-HRP (1:10,000, Bioworld, Cat# BS13278, RRID: AB_2773728) secondary antibody for 2 h at room temperature. The bands were detected using Immobilon Western Chemiluminescent reagent (Bio-Rad, CA, United States).

The PAR1 antagonist SCH79797 (25 μg/kg; Abcam, Cat# ab120858) was selected for in vivo intervention experiments (Figure 1B). SCH79797 was dissolved in dimethyl sulfoxide (DMSO) and diluted in sterile 0.01 M PBS. Rats were randomly divided into four groups: the sham group, TN group, TN + SCH79797 group, and the TN + DMSO group. Rats in the TN + SCH79797 and TN + DMSO groups were given an intraperitoneal injection of SCH79797 (25 μg/kg) and an equivalent volume of DMSO, respectively. After testing the mechanical pain thresholds on PODs 7, 14, and 21, rats in the TN + SCH79797 and TN + DMSO groups were repeatedly injected with SCH79797 and DMSO, respectively, once a week.

All data presented represent the mean of at least three replicates of independent samples. Data are presented as the mean ± SD and were analyzed using two-way ANOVA and Sidak’s multiple comparisons test. Statistical analysis and graphing were performed using R language and GraphPad Prism v8.0 (GraphPad Software, Inc., CA, United States). Significant differences are denoted by asterisks, and p < 0.05 was considered statistically significant.

According to the results of orofacial mechanical behavioral tests in the sham and TN groups, no significant difference was observed in the baseline mechanical pain thresholds between the two groups (F_1,27 = 1.64, p > 0.05; Figure 2A). Interestingly, the mechanical pain thresholds of rats in the sham group dramatically declined on POD 1 and then gradually increased and approached baseline levels from PODs 3 to 7. However, rats in the TN group exhibited a converse trend of mechanical pain thresholds, i.e., the thresholds markedly increased and peaked on POD 1 and then gradually decreased and approached the baseline levels from PODs 3 to 7. In addition, the mechanical pain thresholds on PODs 14, 21, and 28 in the sham group were maintained at baseline levels and were significantly higher than those in the TN group. These results indicate that acute injury following the operation transiently facilitated orofacial mechanical allodynia in the sham group but caused temporary orofacial numbness in the TN group. Rats in the TN group exhibited obvious orofacial mechanical allodynia on PODs 14, 21, and 28, which suggests that the TN animal model induced by chronic compression of the trigeminal root was successfully constructed.

To examine whether compression injury of the TREZ induces disruption of the BNB, an Evans blue extravasation test was conducted. The results showed that the permeability of the BNB increased slightly in the TN group on PODs 7, 14, and 21 and was markedly upregulated on POD 28 compared with the sham group (Figures 2B,C). Notably, Evans blue exudation in the sham group was mainly restricted to the vascular wall and surrounding area (Figure 2C, white arrow), while endoneurial exudation was not obvious. In contrast, more Evans blue exudation was observed in the nerve bundle (Figure 2C and Supplementary Figure 1, yellow arrow) in the TN group, which indicated that the integrity of the BNB was impaired upon compression injury of the TREZ.

Double immunofluorescence staining for Caspr and MBP allowed clear visualization of the CNS-PNS interface in the TREZ (Figure 3). A higher density of Caspr protein was observed in the TREZ (Figures 3, 4A1–H1), and two adjacent Caspr protein expression regions were used to measure the gap between the nodes of Ranvier (Figures 5A–H). In line with previous reports (Henry et al., 2005), two virtually node-free domains were seen, i.e., node-depleted zones were observed flanking the CNS-PNS interface (Figure 3). For further statistical analysis, Leica TCS SP 8 confocal microscopy with ultrahigh resolution was used. To measure the length of node-depleted zones, images of 2,048 × 2,048 × 1 pixels of TREZ were obtained (pixel size = 0.284 μm) and six lines were selected for measuring (Supplementary Figure 2). To measure the diameter of the nerve fibers and nodal gap length at the CNS-PNS interface, images were obtained at higher magnification (logical size, X = 2048 pixels, Y = 2048 pixels, Z = 1 pixel, and pixel size = 0.114 μm). All images were measured and analyzed by a blinded expert. We also calculated and analyzed the nodal ratio, which is the ratio of nodal gap length to the axon diameter, as previously described (Tavares et al., 2019; Figures 4A–H). Our results showed that the length of the node-depleted zones in the TN group was shorter than that in the sham group (Figures 4I,J), while the average lengths of nodal gaps at the CNS-PNS interface in the TN group were significantly wider than those in the sham group (Figure 5I). No difference was found in axon diameters between the two groups (Figure 5J). However, the nodal ratio in the TN group was significantly increased, and further linear correlation regression analysis showed that in the TN group, larger diameter myelinated nerve fibers had a wider nodal gap at the CNS-PNS interface (Figures 5K–O). These results indicated chronic compression caused distinct structural changes at nodes of Ranvier in the TREZ.

To determine if the coagulation/anti-coagulation system in the TREZ was activated during the process of TN, the expression levels of prothrombin and PN-1 were measured by quantitative real-time PCR (qRT–PCR). According to qRT–PCR, no differences were observed in the transcription level of prothrombin between the two groups on PODs 7 and 21, while the transcription level of prothrombin in the TN group was significantly increased compared with that in the sham group on PODs 14 and 28 (Figure 6A). In addition, the relative expression of PN-1 mRNA in the TN group was significantly upregulated on PODs 14 and 28 compared with that in the sham group (Figure 6B).

Due to the activation of the coagulation/anti-coagulation system in the TREZ, we speculated the paranodal protein NF155 in the TREZ may undergo proteolysis during TN pathogenesis, to validate the hypothesis; a Western blotting assay was performed. As shown by Western blotting, two bands of NF155 protein (155 and 125 kDa) were detected (Figure 6C), which suggests that some NF155 proteins may have been hydrolyzed into NF125 and NF30. The expression of NF155 protein in the TN group was upregulated compared with that in the sham group (Figure 6D). Additionally, the protein expression of NF125 in the TREZ in the TN group was increased on PODs 14 and 21, but the difference between the two groups on POD 28 was not significant (Figure 6E).

To define the spatiotemporal expression pattern of PAR1 in the TREZ of the TN rats, an immunohistochemistry assay was performed. Immunofluorescence staining showed that PAR1 was expressed on both sides of the TREZ transitional zone and was accompanied by GFAP-immunopositive astrocytes that extended from the central to the peripheral side of the TREZ (Figures 7A-H). Subsequent immunoblotting assays revealed that PAR1 protein expression in the TREZ in the TN group was significantly higher than that in the sham group on POD 7 and then gradually decreased to a level similar to that of the sham group (Figure 7I).

To further analyze the effect of PAR1 on orofacial hyperalgesia, the PAR1 antagonist SCH79797 was administered to TN rats. According to the behavioral test results, rats in all groups had similar baseline orofacial mechanical pain thresholds before surgery. In contrast to the sham group, rats in the other three groups all displayed a markedly higher orofacial mechanical pain threshold on POD 1, which then gradually decreased to the baseline thresholds from PODs 3 to 7. However, the mechanical pain thresholds of the sham group showed a trend opposite that of the other three groups, i.e., they declined to the nadir on POD 1, then gradually increased and approached the baseline values from PODs 3 to 7, which was similar to the behavior test results described above. Notably, on PODs 14, 21, and 28, the mechanical pain thresholds of the sham group were maintained at baseline values and were higher than those of the other three groups (F_3,48 = 172.8, p < 0.0001; Figure 7J). In addition, rats in the TN + SCH79797 group exhibited higher mechanical pain thresholds than those in the TN and TN + DMSO groups from POD 14, especially on PODs 14 and 28 (F_2,33 = 10.54, p < 0.0001; Figure 7J). These results indicate that SCH79797 ameliorated orofacial mechanical allodynia induced by compression injury of the TREZ.