CXCL10 knockdown attenuates vestibular migraine in rats by inhibiting PI3K-mediated neuroinflammation and central sensitization Authors: Mao-mei Song, Gianluca Coppola, Ying-jie Gao, Shi-na Song, Chang-xin Li, Sui-yi Xu Published: 2026 ABSTRACT BackgroundVestibular migraine (VM) is characterized by recurrent episodes of headache and vertigo, and its pathogenesis is closely associated with neuroinflammation and central sensitization. C-X-C motif chemokine ligand 10 (CXCL10) plays a critical role in neuroinflammation and pain modulation; however, its specific involvement in VM remains unclear.MethodsA rat model of VM was established by repeated intraperitoneal nitroglycerin injections combined with intratympanic kainic acid administration. To investigate the role of CXCL10, adeno-associated virus encoding CXCL10-targeted shRNA (CXCL10-shRNA-AAV) was delivered intracerebroventricularly prior to model induction. A rescue experiment was further performed using the PI3K agonist 740 Y-P to reactivate PI3K/AKT signaling. Mechanical pain thresholds (hind paw and periorbital), head scratching and grooming behavior, and vestibular function scores were assessed. Expression levels of CXCL10, CXCR3, PI3K/AKT pathway components, inflammatory cytokines (pro-IL-1β, IL-6, TNF-α), and central sensitization markers (CGRP, c-fos) in the trigeminal nucleus caudalis (TNC) and vestibular nuclei (VN) were examined by Western blot, qPCR, and immunofluorescence.ResultsVM rats exhibited hyperalgesia, vestibular dysfunction, and upregulated CXCL10 expression in both TNC and VN. Intracerebroventricular delivery of CXCL10-shRNA-AAV effectively knocked down CXCL10 expression and significantly ameliorated pain hypersensitivity and vestibular deficits. Mechanistically, CXCL10 knockdown suppressed PI3K/AKT pathway activation, reduced pro-inflammatory cytokine production (pro-IL-1β, IL-6, TNF-α), and downregulated the central sensitization markers CGRP and c-fos in both TNC and VN. Notably, reactivation of the PI3K/AKT pathway by 740 Y-P partially reversed these effects.ConclusionsCXCL10 contributes to VM pathophysiology by activating PI3K/AKT-mediated neuroinflammation and promoting central sensitization. Targeted knockdown of CXCL10 attenuates both pain and vestibular symptoms in a rat VM model, highlighting CXCL10 as a potential novel therapeutic target for VM.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1186/s10194-026-02437-5. Introduction Vestibular migraine (VM), characterized by recurrent episodes of headache and vertigo, substantially impairs patients’ daily functioning and quality of life [1]. A retrospective study of patients presenting with vertigo identified VM as one of the most common causes of vertigo in women under 60 years of age, with its prevalence gradually declining with advancing age [2]. Notably, among individuals younger than 30 presenting with vertigo, the prevalence of VM reaches as high as 69%, underscoring the urgent clinical need to elucidate its underlying pathophysiology. Neuroinflammation plays a critical role in the pathogenesis of various neurological disorders, including multiple sclerosis, stroke, Alzheimer’s disease, and Parkinson’s disease [3]. Emerging evidence indicates that neuroinflammation plays a pivotal role in migraine pathogenesis, manifested by inflammatory cells activation and upregulation of pro-inflammatory cytokines [4, 5]. As a pivotal relay center in the trigeminovascular pathway, the trigeminal nucleus caudalis (TNC) plays an essential role in integrating nociceptive inputs from craniofacial regions during migraine attacks. The existence of neural connections between the TNC and vestibular nuclei (VN), as demonstrated in animal models, implies a common neuroanatomical substrate underlying both migraine and VM, pointing to shared pathophysiological processes [6, 7]. Neuroinflammation contributes to central sensitization in a rat VM model, underscoring its significance in VM pathophysiology [8]. Neuroinflammation is characterized by inflammatory cells activation and the recruitment of peripheral immune cells, a process tightly regulated by chemokines that direct cell trafficking to inflamed sites [3, 9]. In particular, C-X-C motif chemokine ligand 10 (CXCL10) functions both as a mediator of neuroinflammation and as a canonical pain-associated chemokine, acting via its cognate receptor C-X-C motif chemokine receptor 3 (CXCR3) [10, 11]. CXCL10 has been implicated in various pain-related disorders, including cluster headache, trigeminal neuralgia, and neuropathic pain [12–14]. We previously reported that serum CXCL10 levels were markedly upregulated in VM patients and correlated with the severity of both migraine and vertigo, implicating CXCL10 in VM pathophysiology. This observation was corroborated in a rat VM model, where elevated CXCL10 expression in brain tissue was linked to activation of the PI3K/AKT pathway [15]. Here, we investigate the causal role of CXCL10 using targeted gene knockdown. Specifically, we employed intracerebral delivery of adeno-associated virus (AAV) encoding CXCL10-specific shRNA to suppress CXCL10 expression and elucidate its mechanistic role in VM. Methods Animals This experiment was approved by the Ethics Committee of the First Hospital of Shanxi Medical University (KYLL-2024-129). Male Sprague-Dawley rats (6–8 weeks old, 180–220 g) were obtained from the Experimental Animal Center of Shanxi Medical University and housed in a controlled environment with regulated humidity, temperature, a regular light/dark cycle, and ad libitum access to food and water. All studies were complied with the ARRIVE guidelines. All animals were randomly assigned to experimental groups. Drug administration The VM model was constructed as previously described [15, 16]. Nitroglycerin (NTG, Beijing Yimin, China) with an original concentration of 5 mg/mL was diluted with 0.9% saline to 1 mg/mL before each injection. Rats were intraperitoneally injected with NTG (10 mg/kg) on days 1, 3, 5, 7, and 9. On day 10, rats were anesthetized with isoflurane using a compact small animal anesthesia machine (RWD, China; induction: 5%, maintenance: 1.5–2.5% via face mask). Bilateral injections of kainic acid (KA, Abcam, UK; 12.5 mM, 10 µL per side) were then administered into the tympanic cavities. After each injection, rats were kept in a lateral position for 15–20 min. A normal control group was additionally included in the rescue experiment. To investigate the role of CXCL10 in VM, three different CXCL10-shRNA sequences (shRNA1-CXCL10: 5’-ACT TGA AAT CAT TCC TGC AAG TCT A-3’, shRNA2-CXCL10: 5’-GAA TCT GAG GCC ATC AAG AGC TTA T-3’, shRNA3-CXCL10: 5’-AGC GGT GAG CCA AAG AAG GTC TAA A-3’) and one negative control sequence (shRNA-NC: 5’-TTC TCC GAA CGT GTC ACG TAA-3’) were designed and synthesized by HanHeng Biotechnology, which also performed the packaging of AAV. Fourteen days prior to the first NTG injection, the rats received bilateral intracerebroventricular injections of the AAV. The procedure for the stereotaxic intracerebroventricular injection was as follows: rats were anesthetized with isoflurane and placed on a stereotaxic frame. Bilateral injections targeting the lateral ventricles were performed using stereotaxic coordinates derived from The Rat Brain in Stereotaxic Coordinates, 6th Edition. Coordinates were referenced to bregma and were as follows: AP, -1.0 mm; ML, ± 1.5 mm; and DV, -3.5 mm from the skull surface [17]. A volume of 2 µL of viral solution per side was slowly injected using a 2 µL Hamilton microsyringe at a rate of 0.2 µL/min. After the injection, the syringe was left in place for an additional 10 min before being slowly withdrawn to ensure adequate absorption. Behavioral assessments Behavioral assessments were conducted in a quiet environment with appropriate lighting and temperature. All evaluations were performed at the same time each day. The experimenters were blinded to the group assignments of the rats. To reduce stress and ensure habituation, rats were placed in the testing chamber for adaptation prior to the behavioral testing. To evaluate pain hypersensitivity in rats, the mechanical pain thresholds of the hindpaw and periorbital region were measured by Von Frey filament (Touch-Test™, USA) [18], and the duration of head scratching and grooming behavior was recorded. Baseline pain thresholds were assessed 30 min before each drug administration, while post-treatment thresholds and grooming times were measured 2 h after each injection. Rats were placed in the testing chamber 30 min prior to each assessment to allow them to calm down. Vestibular function in rats was assessed 2 h after tympanic injection by evaluating several behavioral indicators, including head movement, circling behavior, backward walking, tail-hanging reflex, surface righting, and air righting. Each behavior was scored on a scale from 0 to 4 (0 = normal, 4 = severe), with higher scores indicating greater impairment of vestibular function [19, 20]. Western blot After completion of the vestibular function assessment, rats were deeply anesthetized and euthanized. The TNC and VN tissues were dissected according to The Rat Brain in Stereotaxic Coordinates, 6th Edition [17]. The TNC was collected from 14.28 to 15.96 mm posterior to the bregma, 7.2–9.2 mm beneath the dura mater, and 1.6–3.4 mm lateral to the midline. The VN was collected from 10 to 12 mm posterior to the bregma, 7–8 mm beneath the dura mater, and 1–3 mm lateral to the midline. Samples were obtained from six individual animals per group (n = 6). The tissues were thoroughly homogenized in RIPA lysis buffer (Boster, China) supplemented with protease and phosphatase inhibitors (Boster, China) on ice, followed by incubation for 30 min. The lysates were then centrifuged at 12,000 rpm for 30 min at 4 ℃. The supernatant containing the extracted proteins was collected and quantified using the BCA protein assay kit (Boster, China). Protein samples (30 µg) were separated on a 10% SDS-PAGE gel and transferred onto a PVDF membrane. The membrane was then blocked with 5% BSA (Boster, China) at room temperature for 2 h, followed by overnight incubation at 4 °C on a shaker with the following primary antibodies: rabbit anti-CXCL10 (1:1000, Proteintech, Cat#10937-1-AP, China), rabbit anti-CXCR3 (1:500, Boster, Cat#PB9079, China), rabbit anti-PI3K (1:1000, Abclonal, Cat#A11402, China), mouse anti-p-AKT (1:5000, Proteintech, Cat#66444-1-Ig, China), mouse anti-AKT (1:5000, Proteintech, Cat#60203-2-Ig, China), rabbit anti-pro-IL-1β (1:1000, Proteintech, Cat#26048- 1- AP, China), rabbit anti-TNF-α (1:1000, Abclonal, Cat#A23264, China), rabbit anti-IL-6 (1:500, Abclonal, Cat#A0286, China), rabbit anti-c-fos (1:1000, Abclonal, Cat#A17351, China), rabbit anti-CGRP (1:1000, Abclonal, Cat#A5542, China), and rabbit anti-GAPDH (1:5000, Boster, Cat#BM3874, China). The next day, after washing with TBST, the membrane was incubated with secondary antibodies (goat anti-rabbit IgG-HRP, 1:5000, Cat#BA1054; goat anti-mouse IgG-HRP, 1:5000, Cat#BA1050; Abclonal, China) on a shaker at room temperature for 1.5 h. Protein bands were visualized using ECL detection reagent (Boster, China) and imaged with a Chemidoc imaging system. Band intensities were quantified using ImageJ software. Quantitative real-time polymerase chain reaction (qPCR) After dissection of the TNC and VN tissues, total RNA was extracted using the M5 Universal RNA Mini Kit (Mei5bio, China). The purity and concentration of the RNA were measured using a NanoDrop spectrophotometer. Genomic DNA was removed, and reverse transcription into complementary DNA was performed using the M5 Super Plus qPCR RT Kit with gDNA Remover (Mei5bio, China). qPCR was carried out using the Real-Time Quantitative PCR SYBR Green Kit on the LightCycler 480 II system. Primers were designed and synthesized by Sangon Biotech (Shanghai, China), and their sequences are as follows: rat CXCL10 F: 5′-CAA GTG CTG CTG TCG TTC TCT G-3′, R: 5′-GCG GCT GTT CAT GGA AGT CG-3′; GAPDH F: GTA TGA CTC TAC CCA CGG CA, R: AAG ACG CCA GTA GAC TCC AC). Immunofluorescence staining After the assessment of vestibular function, rats were anesthetized and transcardially perfused with 0.9% saline (4 °C) followed by 4% paraformaldehyde (Servicebio, China, 4 °C), initially at a rapid rate and then more slowly, until the rats exhibited limb rigidity and oral frothing, indicating successful fixation. The brains were carefully removed and post-fixed in 4% paraformaldehyde at 4 °C for 24 h. The following day, the brain tissues were sequentially dehydrated in 15% and 30% sucrose solutions at 4 °C until they sank to the bottom. Coronal sections of the TNC and VN regions were cut at a thickness of 12 μm using a cryostat, three sections were randomly selected from each rat for immunofluorescence analysis. Sections were permeabilized with 0.3% Triton X-100 (Beyotime, China) and blocked with 5% BSA (Solarbio, China), followed by incubation with a primary antibody (rabbit anti-CXCL10, 1:100, Proteintech, Cat#10937-1-AP, China) overnight at 4 °C. The next day, sections were incubated at room temperature for 1 h with a secondary antibody (Alexa Fluor 555-labeled donkey anti-rabbit IgG, 1:500, Beyotime, Cat#A0428, China). Nuclei were stained with DAPI (Beyotime, China) for 3 min. Images were acquired using a fluorescence microscope (Olympus, Japan) under identical imaging settings. For quantification analysis, images were acquired from the left and right sides of the target region in each section. Mean fluorescence intensity was measured using ImageJ software and averaged for each rat. Data from six rats per group were included in the analysis. PI3K/AKT pathway rescue experiment To determine whether the protective effects mediated by CXCL10 knockdown could be reversed by activation of the PI3K/AKT pathway, a rescue experiment was conducted using PI3K agonist 740 Y-P. Twenty-four male Sprague-Dawley rats (6–8 weeks old, 180–220 g) were randomly assigned to four groups: the control (CTR) group, VM group, VM + shRNA group, and VM + shRNA + 740 Y-P group. Rats in the VM + shRNA + 740 Y-P group received intraperitoneal injections of 740 Y-P (10 mg/kg) 1 h before each NTG administration [21]. 740 Y-P was dissolved in sterile water before use. Rats in the VM and VM + shRNA groups received the corresponding treatments as described above. Rats in the CTR group received no intervention. After the interventions, behavioral tests and Western blot analyses were conducted as previously described. Statistical analysis Data analysis was performed using GraphPad Prism 9.0. Normally distributed data are presented as mean ± standard deviation. One-way analysis of variance and repeated measures analysis of variance were used for comparisons among multiple groups. When significant differences were detected, post hoc multiple-comparison analyses were conducted using Dunnett’s or Tukey’s test, depending on the experimental design. Dunnett’s test was employed when comparing each treatment group with the control group, while Tukey’s test was used for pairwise comparisons among all groups. Differences were considered statistically significant at p < 0.05. Results Screening of shRNA sequences targeting CXCL10 To identify the most effective shRNA for silencing CXCL10, three shRNA sequences (shRNA1, shRNA2, and shRNA3) targeting CXCL10 were designed and synthesized. The knockdown efficiency of each shRNA was evaluated by qPCR and Western blot (n = 6). The results showed that shRNA2 exhibited the most significant inhibitory effect on CXCL10 expression (Fig. 1). Therefore, shRNA2 was selected for use in subsequent experiments. Fig. 1Expression level of CXCL10 in TNC and VN. (A) Schematic diagram of bilateral TNC site sampling. (B) qPCR analysis of CXCL10 in TNC. (C) and (D) Western blot analysis of CXCL10 in TNC. (E) Schematic diagram of bilateral VN site sampling. (F) qPCR analysis of CXCL10 in VN. (G) and (H) Western blot analysis of CXCL10 in VN. (n = 6, **p < 0.01, ***p < 0.001, ****p < 0.0001) Expression level of CXCL10 in TNC and VN. (A) Schematic diagram of bilateral TNC site sampling. (B) qPCR analysis of CXCL10 in TNC. (C) and (D) Western blot analysis of CXCL10 in TNC. (E) Schematic diagram of bilateral VN site sampling. (F) qPCR analysis of CXCL10 in VN. (G) and (H) Western blot analysis of CXCL10 in VN. (n = 6, **p < 0.01, ***p < 0.001, ****p < 0.0001) CXCL10 knockdown improves hyperalgesia and vestibular dysfunction in VM rats Previous studies have demonstrated that repeated intraperitoneal injection of NTG combined with intratympanic injection of KA can successfully establish a VM rat model, characterized by hyperalgesia and vestibular dysfunction [15]. In the present study, we found that CXCL10 knockdown significantly improved both pain hypersensitivity and vestibular function in VM rats (n = 6, Fig. 2). Fig. 2Behavioral assessment. (A) Schematic diagram of drug administration in rats. (B) and (C) Baseline mechanical thresholds of the hind paw and periorbital region in rats on days 14, 16, 18, 20 and 22. (D) and (E) Hind paw and periorbital region mechanical thresholds in rats 2 h after NTG injection on days 14, 16, 18, 20 and 22. (F) Head scratching and grooming time (20 min) was recorded 1.5 h after NTG injection on days 14, 16, 18, 20 and 22 in rats. (G) Vestibular function scores in rats 2 h after KA injection (n = 6, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001) Behavioral assessment. (A) Schematic diagram of drug administration in rats. (B) and (C) Baseline mechanical thresholds of the hind paw and periorbital region in rats on days 14, 16, 18, 20 and 22. (D) and (E) Hind paw and periorbital region mechanical thresholds in rats 2 h after NTG injection on days 14, 16, 18, 20 and 22. (F) Head scratching and grooming time (20 min) was recorded 1.5 h after NTG injection on days 14, 16, 18, 20 and 22 in rats. (G) Vestibular function scores in rats 2 h after KA injection (n = 6, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001) The expression of CXCL10/CXCR3 in TNC and VN after CXCL10-shRNA-AAV treatment To confirm the successful establishment of the CXCL10 knockdown VM rat model, the expression of CXCL10 in brain tissues was examined by Western blot and immunofluorescence (n = 6). Compared with the VM group, VM + shRNA group showed significantly reduced CXCL10 expression in both the TNC and VN (Fig. 3), indicating that the model was successfully constructed. In contrast, CXCR3 expression showed a slight decrease in the VM + shRNA group, but the difference did not reach statistical significance. Fig. 3Expression level of CXCL10 and CXCR3 in TNC and VN. (A) and (E) Immunofluorescence detection of CXCL10 in TNC (Bar = 100 μm). (B) Stained with secondary antibody only as a negative control (Bar = 100 μm). (C) and (F) Western blot analysis of CXCL10 in TNC. (D) and (G) Western blot analysis of CXCR3 in TNC. (I) Stained with secondary antibody only as a negative control (Bar = 100 μm). (H) and (L) Immunofluorescence detection of CXCL10 in VN (Bar = 100 μm). (J) and (M) Western blot analysis of CXCL10 in VN. (K) and (N) Western blot analysis of CXCR3 in TNC. (n = 6, **p < 0.01, ***p < 0.001, ****p < 0.0001). 4 V: fourth ventricle Expression level of CXCL10 and CXCR3 in TNC and VN. (A) and (E) Immunofluorescence detection of CXCL10 in TNC (Bar = 100 μm). (B) Stained with secondary antibody only as a negative control (Bar = 100 μm). (C) and (F) Western blot analysis of CXCL10 in TNC. (D) and (G) Western blot analysis of CXCR3 in TNC. (I) Stained with secondary antibody only as a negative control (Bar = 100 μm). (H) and (L) Immunofluorescence detection of CXCL10 in VN (Bar = 100 μm). (J) and (M) Western blot analysis of CXCL10 in VN. (K) and (N) Western blot analysis of CXCR3 in TNC. (n = 6, **p < 0.01, ***p < 0.001, ****p < 0.0001). 4 V: fourth ventricle CXCL10 knockdown alleviates neuroinflammation in the VM model Previous studies have demonstrated that CXCL10 exerts its effects in VM by specifically binding to CXCR3, thereby activating the PI3K/AKT signaling pathway and promoting the expression of downstream inflammatory cytokines. To validate this mechanism, we examined the expression of PI3K/AKT pathway related proteins and inflammatory cytokines in the brain tissues of each group by Western blot (n = 6). As shown in Figs. 4 and 5, the levels of PI3K, AKT, phosphorylated AKT (p-AKT), and downstream pro-inflammatory cytokines pro-IL-1β, IL-6, and TNF-α were significantly decreased in the VM + shRNA group compared to those in the VM group, suggesting that CXCL10 plays a critical role in PI3K/AKT pathway activation in VM. Fig. 4Expression level of PI3K/AKT pathway proteins and inflammatory cytokines in TNC. (A) Schematic diagram of bilateral TNC site sampling. (B) to (H) Western blot analysis of PI3K, p- AKT, AKT, IL- 6, pro- IL- 1β, TNF- α expression in TNC (n = 6, **p < 0.01, ***p < 0.001, ****p < 0.0001) Expression level of PI3K/AKT pathway proteins and inflammatory cytokines in TNC. (A) Schematic diagram of bilateral TNC site sampling. (B) to (H) Western blot analysis of PI3K, p- AKT, AKT, IL- 6, pro- IL- 1β, TNF- α expression in TNC (n = 6, **p < 0.01, ***p < 0.001, ****p < 0.0001) Fig. 5Expression level of PI3K/AKT pathway proteins and inflammatory cytokines in VN. (A) Schematic diagram of bilateral VN site sampling. (B) to (H) Western blot analysis of PI3K, p- AKT, AKT, IL- 6, pro- IL- 1β, TNF- α expression in VN (n = 6, *p < 0.05, ***p < 0.001, ****p < 0.0001) Expression level of PI3K/AKT pathway proteins and inflammatory cytokines in VN. (A) Schematic diagram of bilateral VN site sampling. (B) to (H) Western blot analysis of PI3K, p- AKT, AKT, IL- 6, pro- IL- 1β, TNF- α expression in VN (n = 6, *p < 0.05, ***p < 0.001, ****p < 0.0001) CXCL10 knockdown alleviates central sensitization in the VM model In addition, we evaluated the expression of central sensitization markers, including CGRP and c-fos (n = 6). As shown in Figs. 6, the levels of CGRP and c-fos in the TNC and VN were markedly reduced in the VM + shRNA group compared with the VM group, indicating that CXCL10 may also regulate the expression of central sensitization markers. Fig. 6Expression level of CGRP and c-fos in TNC and VN. (A) to (C) Western blot analysis of CGRP and c-fos in TNC. (D) to (F) Western blot analysis of CGRP and c-fos in VN. (n = 6, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001) Expression level of CGRP and c-fos in TNC and VN. (A) to (C) Western blot analysis of CGRP and c-fos in TNC. (D) to (F) Western blot analysis of CGRP and c-fos in VN. (n = 6, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001) Activation of PI3K/AKT signaling partially reverses the protective effects of CXCL10 knockdown To determine whether the anti-inflammatory and anti-central sensitization effects of CXCL10 knockdown could be reversed by activation of the PI3K/AKT pathway, a rescue experiment was performed. As shown in Fig. 7, administration of the PI3K agonist 740 Y-P significantly reversed the CXCL10 knockdown-mediated alleviation of mechanical hyperalgesia and vestibular dysfunction of CXCL10 knockdown in VM rats. At the molecular level, as shown in Figs. 8 and 9, treatment with 740 Y-P reactivated PI3K/AKT pathway in VM rats subjected to CXCL10 knockdown. Consistent with this effect, the expression levels of downstream inflammatory cytokines and central sensitization markers were significantly increased compared with those in the VM + shRNA group. These findings indicate that pharmacological activation of the PI3K/AKT pathway partially reversed the protective effects of CXCL10 knockdown, supporting a causal role of PI3K/AKT signaling in mediating the effects of CXCL10 in VM. Notably, administration of 740 Y-P did not significantly affect the expression levels of CXCL10 or CXCR3, indicating that PI3K/AKT signaling acts downstream of the CXCL10/CXCR3 axis. Fig. 7Behavioral assessment. (A) Schematic diagram of drug administration in rats. (B) and (C) Baseline mechanical thresholds of the hind paw and periorbital region in rats on days 14, 16, 18, 20 and 22. (D) and (E) Hind paw and periorbital region mechanical thresholds in rats 2 h after NTG injection on days 14, 16, 18, 20 and 22. (F) Vestibular function scores in rats 2 h after KA injection. (G) Head scratching and grooming time (20 min) was recorded 1.5 h after NTG injection on days 14, 16, 18, 20 and 22 in rats (n = 6; &, CTR vs. VM; +, CTR vs. VM + shRNA + 740 Y-P; ^, VM vs. VM + shRNA; #, VM + shRNA vs. VM + shRNA + 740 Y-P. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001) Behavioral assessment. (A) Schematic diagram of drug administration in rats. (B) and (C) Baseline mechanical thresholds of the hind paw and periorbital region in rats on days 14, 16, 18, 20 and 22. (D) and (E) Hind paw and periorbital region mechanical thresholds in rats 2 h after NTG injection on days 14, 16, 18, 20 and 22. (F) Vestibular function scores in rats 2 h after KA injection. (G) Head scratching and grooming time (20 min) was recorded 1.5 h after NTG injection on days 14, 16, 18, 20 and 22 in rats (n = 6; &, CTR vs. VM; +, CTR vs. VM + shRNA + 740 Y-P; ^, VM vs. VM + shRNA; #, VM + shRNA vs. VM + shRNA + 740 Y-P. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001) Fig. 8Expression level of CXCL10/CXCR3, PI3K/AKT pathway proteins and inflammatory cytokines in TNC. (A) Representative Western blot bands of PI3K, c-fos, p-AKT, AKT, IL-6, CXCR3, pro-IL-1β, TNF-α, CGRP and CXCL10. (B) to (K) Corresponding quantitative analysis of band intensity of PI3K, c-fos, p-AKT, AKT, IL-6, CXCR3, pro-IL-1β, TNF-α, CGRP and CXCL10 (n = 6, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001) Expression level of CXCL10/CXCR3, PI3K/AKT pathway proteins and inflammatory cytokines in TNC. (A) Representative Western blot bands of PI3K, c-fos, p-AKT, AKT, IL-6, CXCR3, pro-IL-1β, TNF-α, CGRP and CXCL10. (B) to (K) Corresponding quantitative analysis of band intensity of PI3K, c-fos, p-AKT, AKT, IL-6, CXCR3, pro-IL-1β, TNF-α, CGRP and CXCL10 (n = 6, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001) Fig. 9Expression level of CXCL10/CXCR3, PI3K/AKT pathway proteins and inflammatory cytokines in VN. (A) Representative Western blot bands of PI3K, c-fos, p-AKT, AKT, IL-6, CXCR3, pro-IL-1β, TNF-α, CGRP and CXCL10. (B) to (K) Corresponding quantitative analysis of band intensity of PI3K, c-fos, p-AKT, AKT, IL-6, CXCR3, pro-IL-1β, TNF-α, CGRP and CXCL10 (n = 6, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001) Expression level of CXCL10/CXCR3, PI3K/AKT pathway proteins and inflammatory cytokines in VN. (A) Representative Western blot bands of PI3K, c-fos, p-AKT, AKT, IL-6, CXCR3, pro-IL-1β, TNF-α, CGRP and CXCL10. (B) to (K) Corresponding quantitative analysis of band intensity of PI3K, c-fos, p-AKT, AKT, IL-6, CXCR3, pro-IL-1β, TNF-α, CGRP and CXCL10 (n = 6, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001) Discussion Emerging studies have established neuroinflammation as a key contributor to migraine pathophysiology, manifesting as enhanced vascular permeability, recruitment of leukocytes, activation of glial cells, and upregulation of pro-inflammatory cytokines and chemokines [22–24]. This inflammatory cascade drives central sensitization, thereby perpetuating and amplifying nociceptive signaling. CXCL10 functions as a critical pro-inflammatory mediator that facilitates leukocyte trafficking and exacerbates focal inflammation, playing a significant role in a broad spectrum of neurological diseases [10, 25, 26]. CXCL10 levels are elevated in the cerebrospinal fluid (CSF) of patients with multiple sclerosis and are significantly correlated with both the number of T cells in the CSF and the disease duration [27]. Beyond its role in neuroinflammation, CXCL10 functions as a canonical pain-related chemokine implicated in multiple pain conditions. Clinical evidence from traumatic spinal cord injury reveals increased serum CXCL10 concentrations, which correlate positively with the intensity of pain symptoms [28]. Elevated serum CXCL10 has been observed in patients with lumbar degenerative disc disease presenting with low back pain, with levels showing positive correlations with the severity of clinical symptoms and the extent of imaging-detected disease progression [29]. In addition to its well-established role in neuropathic pain, CXCL10 is also highly expressed in patients with cancer pain and various inflammatory disease-associated pain conditions [30–32]. We have previously shown that CXCL10 is upregulated in the serum of VM patients and correlates with symptom severity [15]. Collectively, these clinical findings support the potential of CXCL10 as a candidate molecular biomarker for assessing pain conditions. In parallel, preclinical evidence from animal models has revealed that CXCL10 contributes significantly to the pathogenesis of neuropathic, cancer-related, and inflammatory pain [14, 33, 34]. We previously found that CXCL10 is upregulated in the brain tissue of VM rats and may exert its effects via PI3K/AKT signaling. Inhibition of this pathway attenuated behavioral symptoms and inflammation in VM animals, underscoring the importance of PI3K/AKT in VM pathogenesis. However, no evidence of PI3K/AKT-mediated downregulation of CXCL10 was observed in our study [15]. Our results demonstrate that CXCL10 knockdown ameliorates pain hypersensitivity and vestibular deficits in a rat model of VM. Concurrently, PI3K/AKT signaling was suppressed and inflammatory cytokine expression was reduced, indicating that CXCL10 exerts its pro-inflammatory effects through PI3K/AKT activation to drive VM pathogenesis. Multiple studies have confirmed that CXCL10 contributes to the pathogenesis of pain disorders through activation of the PI3K/AKT signaling pathway. In a mouse model of trigeminal neuralgia, intra-ganglion injection of CXCL10 exacerbated hyperalgesia and promoted PI3K/AKT pathway activation [13]. Another study demonstrated that intrathecal administration of recombinant CXCL10 reversed the analgesic effect of morphine, concomitantly activating the PI3K/AKT pathway and upregulating downstream inflammatory mediators [35]. Furthermore, CXCL10-induced activation of the PI3K/AKT pathway has also been observed in animal models of cancer pain and inflammatory pain [34, 36]. Notably, our results demonstrated that CXCL10 knockdown led to a marked decrease in CGRP and c-fos levels—two widely recognized indicators of central sensitization [37, 38]. Central sensitization in migraine is known to localize predominantly to the TNC. The anatomical connectivity between the TNC and VN via fiber projections provides a rationale for similar sensitization processes in VM. Consistent with this, VM rat models exhibit marked upregulation of CGRP and c-fos, confirming the presence of central sensitization in VM pathophysiology [6]. Building on evidence that CXCL10 silencing suppresses PI3K signaling and downstream inflammation, we next examined its effect on CGRP and c-fos expression. CXCL10 knockdown reduced both inflammatory markers and the upregulation of CGRP and c-fos, indicating a deceleration of central sensitization. These findings align with prior migraine studies showing that anti-inflammatory treatments attenuate central sensitization, reinforcing the pivotal role of inflammation in this process [39, 40]. Although the present study demonstrates a critical role of CXCL10 in VM pathogenesis, the cellular source of CXCL10 remains incompletely understood. In our previous study, double-immunofluorescence staining revealed that CXCL10 was expressed in neurons, microglia, and astrocytes within the TNC and VN, suggesting that multiple cell types may contribute to CXCL10-mediated neuroinflammation [15]. Interestingly, studies of neuropathic pain have shown that CXCL10 is constitutively expressed in neurons, whereas its expression in astrocytes and microglia is markedly upregulated following spinal cord injury [11]. This dynamic and cell type-specific expression pattern suggests that CXCL10 may exert distinct functions in different cellular populations, highlighting the need for future studies to identify the principal cellular source and target of CXCL10 in VM. In addition, although CXCR3 is the primary receptor for CXCL10, CXCR3 expression was not significantly altered following CXCL10 knockdown in the present study, suggesting that suppression of CXCL10 does not substantially affect receptor expression. However, accumulating evidence indicates that CXCR3 plays an important role in pain-related disorders [41]. Previous studies have shown that administration of CXCR3 antagonists or knockdown of CXCR3 expression significantly attenuates pain-related behaviors and suppresses neuroinflammatory responses in experimental models of neuropathic pain [13, 14, 42]. Therefore, the lack of change in CXCR3 expression observed in the present study does not exclude its involvement in VM. Future studies are needed to determine the cellular distribution, functional significance, and therapeutic potential of CXCR3 in VM. To further verify the mechanistic involvement of PI3K/AKT signaling, a rescue experiment was performed using the PI3K agonist 740 Y-P. Reactivation of the PI3K/AKT pathway partially reversed the beneficial effects of CXCL10 knockdown. These findings provide direct evidence that PI3K/AKT signaling is an important downstream mediator of CXCL10 in VM. Previous studies have similarly demonstrated the involvement of PI3K/AKT pathway in pain modulation. In an osteoarthritis model, pharmacological activation of PI3K/AKT with 740 Y-P was reported to reverse the analgesic effects of tuina therapy, accompanied by opposite changes in pain behavior and PI3K/AKT/mTOR pathway activity compared with inhibition of this pathway [43]. Together with our previous findings, the present rescue experiment strengthens the causal relationship between CXCL10 and PI3K/AKT signaling in VM and further supports CXCL10 as a potential therapeutic target. Nevertheless, the protective effects of CXCL10 knockdown were not completely abolished by PI3K/AKT activation, suggesting that additional downstream pathways may also participate in CXCL10-mediated pathological processes. Previous studies have indicated that, in addition to PI3K/AKT, MAPK-related pathways such as p38 and ERK are also involved in CXCL10/CXCR3-associated inflammatory and pain models, highlighting a potential multi-pathway regulatory network underlying its biological effects [44]. Given that neuroinflammatory signaling and central sensitization are also recognized as fundamental mechanisms in migraine with or without aura, the CXCL10-mediated inflammatory pathway identified in this study may not be confined to vestibular migraine. It is conceivable that CXCL10 may also participate in the pathogenesis of non-vestibular migraine subtypes via similar neuroinflammatory and sensitization mechanisms. Further research is required to determine whether CXCL10 represents a broader therapeutic target across different migraine phenotypes. Several limitations of this study should be acknowledged. First, our investigation was largely confined to the PI3K signaling cascade and related molecules, leaving other potentially relevant pathways, including p38 MAPK, ERK, and NF-κB signaling, unexamined. Second, although CXCL10 has been reported to be expressed in multiple cell types, its cell-specific functions in VM remain unclear. Third, the mechanistic conclusions were based exclusively on in vivo experiments, and additional in vitro studies are needed to clarify the direct cellular actions of CXCL10. Forth, molecular analyses were performed only at the experimental endpoint, multiple time-point studies are needed to clarify the temporal relationship between CXCL10 expression and the development of VM-related symptoms. Fifth, the present study assessed pro-IL-1β expression, which serves as an indicator of inflammatory priming and cytokine production [45, 46]; however, the biologically active, mature form of IL-1β was not specifically evaluated. Future investigations that quantify mature IL-1β and assess inflammasome activation will help to further clarify the inflammatory mechanisms underlying VM. Sixth, while the present study focused on CXCL10, the role of its receptor CXCR3 was not investigated in depth and warrants further exploration. Finally, to avoid variability associated with the estrous cycle, only male rats were included, potential sex-specific differences should be evaluated in future investigations. Conclusion CXCL10 knockdown ameliorated pain hypersensitivity and vestibular deficits in VM rats by suppressing PI3K pathway activation, attenuating the release of inflammatory mediators (IL-6, pro-IL-1β, TNF-α), and downregulating central sensitization markers (CGRP, c-fos). Notably, reactivation of the PI3K/AKT pathway by 740 Y-P partially abolished these protective effects, further supporting a causal role of PI3K/AKT signaling in CXCL10-mediated VM pathology. These results highlight CXCL10 as a key modulator of VM and a potential therapeutic target. Considering the recognized role of neuroinflammation in migraine with and without aura, CXCL10 may also be involved in the pathogenesis of non-vestibular migraine, which deserves further investigation. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 1