hsa_circ_0006168 drives microglial activation in TNC via miR-99b-5p/KDM6B axis to promote central sensitization in migraine

Abstract

Background

Migraine is a disabling neurological disorder characterized by recurrent headache attacks and associated symptoms. The mechanisms underlying migraine remain unclear. This study aimed to identify differentially expressed circular RNAs (circRNAs) in migraine and elucidate their potential roles in migraine pathogenesis.

Methods

CircRNA expression was profiled in migraine patients using qRT-PCR. The circular structure and miRNA sponge function of hsa_circ_0006168 (orthologous to mmu_circ_0012878 in mice) were validated by Sanger sequencing, RNase R treatment assay, fluorescence in situ hybridization (FISH), and bioinformatics. Migraine mouse models were established by repeated nitroglycerin (NTG) injections. Lentiviral-mediated knockdown of mmu_circ_0012878, miR-99b-5p agomir, and GSK-J4 (KDM6B inhibitor) were administered to assess their effects on central sensitization, microglial activation, and M1/M2 polarization in the trigeminal nucleus caudalis (TNC) of mice by qRT-PCR, western blot, and immunofluorescence. In vitro, LPS-stimulated BV2 cells with mmu_circ_0012878 knockdown were analyzed for M1/M2 polarization using qRT-PCR, western blot, immunofluorescence, flow cytometry, multiplex flow cytometry assay, and ELISA. Molecular interactions within the mmu_circ_0012878/miR-99b-5p/KDM6B axis were confirmed by qRT-PCR, western blot and dual luciferase assay.

Results

hsa_circ_0006168 (mmu_circ_0012878) and KDM6B were upregulated, while miR-99b-5p was downregulated in the peripheral blood of migraine patients and the TNC of NTG-induced migraine mouse models. Silencing mmu_circ_0012878/KDM6B and overexpressing miR-99b-5p attenuated central sensitization, potentially through the PI3K/AKT signaling pathway. Additionally, mmu_circ_0012878 partially colocalized with Ionized calcium-binding adapter molecule 1 (Iba1), and its knockdown suppressed microglial activation and promoted M2 polarization in the TNC. In LPS-stimulated BV-2 cells, mmu_circ_0012878 knockdown reduced microglial activation, shifted polarization toward M2, and decreased M1-related cytokines (IL-6, IL-1β). Dual luciferase assays confirmed mmu_circ_0012878 regulated KDM6B expression by sponging miR-99b-5p.

Conclusions

hsa_circ_0006168 acted as a competitive endogenous RNA (ceRNA) by sponging miR-99b-5p, thereby blocking miR-99b-5p-mediated suppression of KDM6B. The hsa_circ_0006168/miR-99b-5p/KDM6B axis contributed to central sensitization by activating microglia of TNC, possibly via the PI3K/AKT signaling pathway. This study suggested a novel molecular mechanism underlying migraine development and identified potential therapeutic targets for intervention.

Graphical Abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s10194-026-02288-0.

Introduction

Migraine is a prevalent chronic neurovascular disorder that affects approximately 10% of the global population and affects more than 1 billion individuals worldwide [1]. As the second leading cause of neurological disability, migraine imposes a substantial burden on both quality of life and socioeconomic systems [2]. The pathogenesis of migraine is not fully understood, and current studies have focused mainly on the trigeminovascular hypothesis (peripheral and central sensitization), the cortical excitability imbalance hypothesis (cortical spreading depression and glutamatergic dysfunction), and the neuroglial regulation hypothesis [3–5]. However, no single pathological process can fully account for the diverse and complex clinical manifestations of migraine.

Central sensitization, characterized by increased responsiveness of nociceptive neurons in the central nervous system to normal or subthreshold afferent stimuli, serves as a fundamental mechanism in the pathogenesis of pain. This mechanism primarily contributes to the occurrence of hyperalgesia and chronification in migraine, which are correlated with clinical symptoms such as cutaneous allodynia, photophobia, and phonophobia [6, 7]. During migraine attacks, nociceptor activation triggers the release of proinflammatory neuropeptides, including calcitonin gene-related peptide (CGRP) and substance P (SP), which induce local inflammatory reactions. These reactions subsequently sensitize and activate nociceptive neurons at various levels, ultimately contributing to pain perception through cortical integration [8]. Furthermore, noxious stimuli activate microglia in the trigeminal nucleus caudalis (TNC). Once activated, microglia interact with neurons through chemotaxis, facilitating nociceptive signaling and driving aberrant neuronal discharges that contribute to central sensitization and pain onset [9]. Activated microglia also secrete a variety of inflammatory mediators and chemokines. On the basis of their functional state, microglia can polarize into proinflammatory M1 phenotypes (classically activated) or anti-inflammatory M2 phenotypes (alternatively activated), allowing them to play dual roles in the regulation of neuroinflammation [10]. Through dynamic phenotype switching and the secretion of inflammatory factors, activated microglia critically modulate neuroinflammation and contribute to the development and maintenance of central sensitization [11]. Moreover, neuroinflammation promotes central sensitization via integrated peripheral-central mechanisms, such as glutamate-CGRP cascade-mediated peripheral sensitization [12] and the remodeling of synaptic plasticity [13]. A self-perpetuating triad—activated microglia, neuroinflammation, and central sensitization—are interconnected and involved in the occurrence and progression of migraine.

Circular RNA (circRNA) is an endogenous RNA molecule generated through the back-splicing of precursor mRNAs (pre-mRNAs) [14]. Structurally, circRNAs are covalently closed-loop molecules lacking 5′-caps and 3′-poly(A) tails, which confer resistance to RNA exonuclease-mediated degradation and enable their stable expression [15]. Functionally, circRNAs act as competitive endogenous RNAs (ceRNAs) that sponge microRNAs (miRNAs) to regulate gene expression and participate in protein translation [16], making circRNAs promising diagnostic biomarkers and potential therapeutic targets for various diseases. Recent studies have demonstrated that mmu_circRNA_007893 sponges mmu-miR-485-5p, regulating CGRP and influencing interleukin-6 (IL-6) mRNA expression, thereby participating in inflammatory responses. Additionally, mmu_circRNA_003795 mediated the upregulation of FOS-like 2 (FOSL2) through the sequestration of miR-504-3p. FOSL2 is a key transcriptional regulator involved in pain sensitivity [17]. These findings highlight the emerging role of specific circRNAs in modulating migraine-related molecular pathways.

In our previous study, we identified four differentially expressed circRNAs, namely, hsa_circ_0006168, hsa_circ_0072088, hsa_circ_0000367, and hsa_circ_0003745, in migraine patients through high-throughput circRNA microarray analysis [18]. Building upon these findings, this study further screened and validated the differentially expressed circRNA hsa_circ_0006168 in migraine, investigating its potential downstream pathways involved in migraine pathogenesis. In addition, we explored the relationship between hsa_circ_0006168 and central sensitization, with particular attention to microglial activation in the TNC. In summary, this study elucidated the potential mechanisms by which hsa_circ_0006168 contributes to the development and progression of migraine.

Materials and methods

Sample collection

Peripheral blood samples from 45 migraine patients and 30 healthy controls were collected from outpatients at the First Affiliated Hospital of Harbin Medical University (Harbin, China). This study was designed to conform with the 1964 revision of the Helsinki Declaration and was approved by the Ethics Committee of the First Affiliated Hospital of Harbin Medical University (IRB-AF/SC-12/03.0). All patients provided written informed consent. Patients were diagnosed with migraine without aura on the basis of the 3rd edition of the headache classification standard developed by the International Headache Society (IHS) [19]. Blood samples were collected from all patients during the interictal period, none of the enrolled migraine patients had taken acute analgesic medications within 48 hours prior to enrollment or received preventive treatments within one week before sampling.

Animals

All animal procedures and ethical considerations for this study received formal approval from the Institutional Animal Care and Use Committee at the First Affiliated Hospital of Harbin Medical University, under reference numbers YS166. C57BL/6J mice, with an average weight of approximately 25 g, were obtained from Liaoning Changsheng Biotechnology Co., Ltd. (Benxi, Liaoning, China). All the animals were housed under controlled conditions, with a constant temperature of 23 ± 1 °C and a 12-hour light‒dark cycle. They had free access to food and water.

Cell culture and treatment

The murine BV2 microglial cell line was obtained from Procell (Wuhan, China). The cells were cultured in high-glucose Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin‒streptomycin (all from Procell). In this study, BV2 cells were stimulated with 1 μg/ml lipopolysaccharide [20] (LPS; Solarbio, L8880) for 12 h.

qRT-PCR

Total RNA was extracted from individual samples via TRIzol Reagent (Invitrogen, 15596–026), cDNA was synthesized via the ReverTra Ace qPCR RT Kit (TOYOBO, FSQ-101), and qRT‒PCR was performed with FastStart Universal SYBR Green Master Mix (Rox, 4913914001) and specific primers (partial sequences shown in Table 1; others synthesized by RiboBio, Guangzhou, China). The cycling conditions were 95 °C for 10 min, followed by 40 cycles of 95 °C for 5s, 60 °C for 20s, and 72 °C for 10s. β-actin served as the reference gene. All reactions were conducted in triplicate. Relative expression was calculated via the 2^^−ΔΔCT method and reported as the fold change.

Table 1.

Primers and probes used in this study

Name	Sequence	Application
hsa_circ_0072088	F:ATTTTCCAAGCTGGCCCTTAC R:AATGGTGGCATGTTTTGTCATT	qRT-PCR
hsa_circ_0000367	F:CTTGTCCAGGTCCATTCCAT R:CCCCTTTCAGAGTCATATTGC	qRT-PCR
hsa_circ_0003745	F:GCTCCCTGGCTACTACATGACC R:AAGAAGATGCTCTCCACCACG	qRT-PCR
hsa_circ_0006168-Divergent	F:CCACCTGATATTGCCAAGCTTC R:CCTTTGGCATCCCTATTAGTCTTTC	qRT-PCR, Agarose gel electrophoresis
hsa_circ_0006168-Convergent	F:TGCACCTAAATGACAATTACCTT R:GACACCATGTTTCCTAGTTCTGC	qRT-PCR, Agarose gel electrophoresis
Human-U6-Divergent	F:GAAGCGTTCCATATTTTGTGCTC R:ACGAATTTGCGTGTCATCCTTG	qRT-PCR, Agarose gel electrophoresis
Human-U6-Convergent	F:CTCGCTTCGGCAGCACA R:AACGCTTCACGAATTTGCGT	qRT-PCR, Agarose gel electrophoresis
mmu_circ_0012878 probes	5’-TGCGAGGATCTGGAGGATCA-3’	FISH

Sanger sequencing and RNase R treatment assay

Total RNA and genomic DNA (gDNA) were extracted via the TRIzol Reagent and the HiPure Blood DNA Mini Kit (Magen, 020-89857862), respectively. The purity and integrity of the RNA and DNA were assessed via a NanoDrop spectrophotometer (Thermo, USA) and agarose gel electrophoresis. Total RNA was treated with RNase R to degrade linear RNAs. Divergent primers were used in qRT‒PCR to specifically amplify the back-splice junctions of hsa_circ_0006168, and the resulting PCR products were verified by 2.5% agarose gel electrophoresis. The primers used are shown in Table 1.

Fluorescence in-situ hybridization (FISH)

Cy3-labeled mmu_circ_0012878 probes were designed and synthesized by Boerfu (Wuhan, China), and the probe sequences are listed in Table 1. Paraffin sections of brain tissues containing the TNC region were prepared. The sections were permeabilized, blocked, and incubated with mmu_circ_0012878 probes at 42 °C overnight. Signal amplification was performed via the use of enzyme-labeled antibodies and tyramide reagents. Nuclei were counterstained, and fluorescence images were acquired via a fluorescence microscope (Nikon, Japan).

Enzyme-linked immunosorbent assay (ELISA)

An ELISA kit (Jianglai, JL54733-96; JL20268; JL18442; JL20246) was used to measure KDM6B, IL-6, IL-1β and IL-12p70 expression in the peripheral blood of patients according to the manufacturer’s instructions.

Migraine model

To induce migraine in mice, nitroglycerin (NTG; Shandong Shenglu Pharmaceutical Co., Ltd., China) was administered intraperitoneally at a dose of 10 mg/kg every other day for 9 d, with a total of five injections [21].

Periorbital and plantar mechanical thresholds

The mechanical thresholds were assessed following 3 d of habituation to establish baseline responses. Von Frey filaments were applied to specific regions of the mice: the periorbital area (adjacent to the posterior part of the eyes) and the central plantar surface (avoiding the toe pads). The initial filament forces were set at 0.6 g for the periorbital region and 1 g for the plantar region. Each filament was applied perpendicularly and maintained for 3–6 s. Behaviors such as asymmetric withdrawal, escape, or attack of the stimulus were considered positive responses. The thresholds were determined via the updown method [22].

Paw heat pain threshold

The paw heat pain threshold was measured after 3 d of habituation. The mice were placed on a heated platform maintained at 50 °C, and the latency to paw withdrawal, shaking, or licking was recorded as a positive response. Each mouse was tested three times, with an interval of at least 5 min between trials.

Western blot (WB)

Proteins were extracted from TNC tissues and BV2 cells via RIPA lysis buffer (Seven, 104–02) supplemented with PMSF (Seven, 106–02) and phosphatase inhibitors (Rox, 4906845001). Protein concentrations were measured with a NanoDrop spectrophotometer. Proteins were then transferred onto polyvinylidene difluoride (PVDF; PALL, USA) membranes. The membranes were blocked with 5% nonfat milk (Seven, 128–03) for 2 h at room temperature, incubated with primary antibodies (shown in Table 2) overnight at 4 °C and then with secondary antibodies (shown in Table 2) for 1 h. The bands were visualized via enhanced chemiluminescence (ECL; Biosharp, BL520A) reagents and quantified with ImageJ.

Table 2.

Antibodies used in this study

Name	Catalog NO.	Manufacturer	Dilution	Application
CGRP	ab283568	Abcam	1:1000/1:50	WB, IF
c-Fos	2250	CST	1:1000	WB
GAPDH	YN5585	Immunoway	1:10000	WB
KDM6B	55354–1-AP	proteintech	1:1000	WB
β-Tubulin	AB0039	Abways	1:10000	WB
P-AKT	YP0006	Immunoway	1:2000	WB
AKT	YT0185	Immunoway	1:2000	WB
PI3K	YT6156	Immunoway	1:2000	WB
iNOS	18985–1-AP	proteintech	1:1000	WB
Arg-1	82975–1-RR	proteintech	1:10000	WB
Goat Anti Rabbit IgG (H+L)	RS0002	Immunoway	1:10000	WB
Iba1	ab178846	Abcam	1:200	IF
GFAP	16825–1-AP	proteintech	1:500	IF
NeuN	26975–1-AP	proteintech	1:1000	IF
CD206	24595s	CST	1:400	IF
KDM6B	ab38113	Abcam	1:200	IF
Goat Anti Mouse IgG (H+L) (AbFlour 488)	RS3208	Immunoway	1:200	IF
APC anti-mouse CD86 Antibody	159215	Biolegend	5 μl/106 cells	Flow cytometry
FITC anti-mouse CD206 Antibody	141703	Biolegend	5 μl/106 cells	Flow cytometry

Immunofluorescence staining (IF)

Paraffin-embedded brain tissues, including the TNC region, were sectioned at a thickness of 5 μm. The sections were dewaxed in xylene, rehydrated through a graded ethanol series and treated with EDTA (Boerfu, Wuhan, China) for antigen retrieval. The sections were subsequently blocked with 3% BSA (Solarbio, SW3015) for 30 min and incubated with primary antibodies (listed in Table 2) overnight at 4 °C. Secondary antibodies (listed in Table 2) were applied for 1 h at room temperature in the dark. Nuclei were counterstained with DAPI (Beyotime, C1005) for 10 min, and slides were mounted with antifade medium (Beyotime, P0126). Fluorescence images were acquired with a fluorescence microscope (Nikon, Japan). To quantify immunofluorescence-positive cells, five fields within the core region of each nuclear section were randomly selected (150 × 150 μm per field). Images were acquired under identical microscope settings.

Intra-TNC microinjection of LV-mmu_circ_0012878-RNAi

The mice were randomly divided into four groups: (1) the Sham group; (2) the NTG group; (3) the NTG + LV-NC group; and (4) the NTG + LV-circRNA-RNAi group. On d 10 following NTG administration, the mice were anesthetized with isoflurane and secured in a stereotaxic apparatus (RWD, Shenzhen, China) with the skull positioned in the horizontal plane. The TNC was targeted at X: ±1.6 mm, Y: −6.1 mm, and Z: −5.3 mm [23]. A total volume of 1 × 10⁶ TU lentivirus (LV-mmu_circ_0012878-RNAi or LV-NC, 1–1.5 μl, GeneChem, Shanghai, China) [24] was slowly injected at a rate of 0.3 μl/min, which was maintained for 10 min to prevent backflow. Tissue samples were collected for further analysis one week after injection.

Intra-TNC microinjection of miR-99b-5p agomir

The mice were randomly divided into four groups: (1) the Sham group; (2) the NTG group; (3) the NTG + agomir NC group; and (4) the NTG + miR-99b-5p agomir group. Figure 4A shows a schematic of the procedure. The microinjection was performed as described above. A total volume of 0.5 nmol of miR-99b-5p agomir or agomir NC solution (1.5 μl, RiboBio, Guangzhou, China) [25] was delivered into the TNC. Tissue samples were collected for further analysis 3 d after injection.

Fig. 4.

Upregulation of miR-99b-5p alleviated central sensitization and downregulated KDM6B in the TNC of migraine mice. A Flowchart of the animal experimental procedure. B Behavioral tests of mechanical and paw heat pain thresholds (n = 10). C Western blot was performed to detect the expression of CGRP and c-Fos in the TNC (n = 6). D Representative immunofluorescence image of CGRP (red) in the TNC of NTG+miR-99b-5p agomir group. Scale bar = 500 μm. E CGRP (red) expression in the TNC was examined by immunofluorescence in each group (n = 4). Scale bar = 200 μm. F KDM6B expression was assessed by western blot (n = 6). Data are shown as the means ± SD. The p values were calculated using two-way ANOVA followed by Tukey’s post hoc tests (B) and one-way ANOVA (C and E-F). ^##p < 0.01, *p < 0.05, **p < 0.01, ***p < 0.001

GSK-J4 treated mice

GSK-J4 functions as an inhibitor of the KDM6 family [26]. The mice were randomly divided into four groups: (1) the Sham group; (2) the NTG group; (3) the NTG + DMSO group; and (4) the NTG + GSK-J4 group. On the basis of the literature [27] and our findings, GSK-J4 was administered intraperitoneally at a dose of 20 mg/kg once daily for 9 d. GSK-J4 was given 30 min before NTG injection. The NTG + DMSO group received an equal volume of vehicle solution (10% DMSO, 40% PEG300, 5% Tween-80, or 45% 0.9% saline). Tissue samples were collected on d 9 for subsequent experiments.

Cell transfection

BV2 cells were transfected with LV-mmu_circ_0012878-RNAi or LV-NC (GeneChem, Shanghai, China) along with transfection reagent (GeneChem, Shanghai, China). After 72 h, puromycin (3 μg/ml; Biosharp, BL528A) was added for 48 h to select successfully infected cells for subsequent experiments. BV2 cells were transfected with 100 nM miR‑99b‑5p inhibitor or inhibitor NC and 50 nM miR‑99b‑5p mimic or mimic NC (RiboBio, Guangzhou, China) via Lipofectamine^TM 2000 (Invitrogen, 11668019). The transfection efficiency was confirmed by qRT-PCR at 24 h post infection for further analyses. KDM6B overexpression was achieved by transfecting 3 μg of KDM6B overexpression plasmid or negative control (Miaoling, Wuhan, China) using Lipofectamine^TM 2000, and overexpression efficiency was confirmed by qRT-PCR and WB after 48 h.

Dual luciferase assay

A dual luciferase assay was performed via a commercial kit (KeyGEN, KGE3308-10) following the manufacturer’s instructions. The wild-type or mutant sequences of mmu_circ_0012878 and KDM6B containing the predicted miR-99b-5p binding sites were constructed and cotransfected with the miR‑99b‑5p mimic or mimic NC into cells. Relative luciferase activity was calculated by normalizing firefly luciferase activity to Renilla luciferase activity.

Flow cytometry

M1- and M2-polarized microglia were labeled with CD86 and CD206, respectively [28]. The cells were resuspended, and the corresponding antibodies (listed in Table 2) were added at 5 μl/10⁶ cells. The mixture was gently mixed and incubated at room temperature in the dark for 30 min, with mixing every 10 min. The mean fluorescence intensities (MFIs) of CD86 and CD206 were analyzed via flow cytometry (Beckman Coulter, USA).

Multiplex flow cytometry assay

Proteins were extracted from BV2 cells after intervention, and their concentration and purity were assessed. Inflammatory cytokine expression levels were detected via the RayPlex^® Mouse Inflammation Array Kit 1 (RayBiotech, USA) according to the manufacturer’s instructions.

Statistical analysis

All the experiments were repeated at least three times. In the animal experiments, each group contained 6 or more mice. The data were analyzed via GraphPad Prism 9.1.5. For the behavioral tests in the mice, two-way ANOVA followed by Tukey’s post hoc test was performed. One-way ANOVA was used for comparisons among multiple groups, and Student’s t test was used for comparisons between two groups. For data that were not normally distributed, comparisons between two groups were performed using the Mann-Whitney U test, and the results are presented as the median and interquartile range (P25–P75). p < 0.05 was considered statistically significant. The data are presented as the means ± SDs.

Results

hsa_circ_0006168 was highly expressed in the peripheral blood of migraine patients and may be involved in migraine pathogenesis via the miR-99b-5p/KDM6B axis

On the basis of our previous study [18], 4 differentially expressed circRNAs were identified in migraine patients (Fig. 1A, B). This study enrolled 45 patients with migraine without aura and 30 age- and sex-matched healthy controls. The clinical characteristics of the migraine group are shown in Table 3. Further qRT‒PCR analysis demonstrated that hsa_circ_0006168 was the most significantly upregulated circRNA in the peripheral blood of migraine patients compared with healthy controls (Fig. 1C). Sanger sequencing confirmed that hsa_circ_0006168 was generated by backsplicing of exons 2–4 of the CNOT6L pre-mRNA (Fig. 1D). Amplification via divergent primers was detected only in cDNA, verifying that the circular structure of hsa_circ_0006168 formed through back-splicing (Fig. 1E). FISH revealed that mmu_circ_0012878 (a mouse homolog of hsa_circ_0006168) was highly expressed and predominantly localized in the cytoplasm of the TNC in NTG-induced migraine mice (Fig. 1F), suggesting that hsa_circ_0006168 may be involved in migraine pathogenesis through a ceRNA mechanism. A ceRNA network based on hsa_circ_0006168 was constructed, and the hsa_circ_0006168/miR-99b-5p/KDM6B axis was identified as a potential pathway contributing to migraine (Fig. 1G). Similarly, hsa_circ_0006168 and KDM6B were upregulated, whereas miR-99b-5p was downregulated in the peripheral blood of migraine patients (Fig. 1H–J), suggesting that this axis may play a pivotal role in the development of migraine.

Fig. 1.

Identification and validation of the hsa_circ_0006168 (mmu_circ_0012878)/miR-99b-5p/KDM6B axis in migraine. A-B 8 candidate circRNAs were identified by high-throughput circRNA microarray and confirmed by qRT-PCR. 4 circRNAs exhibiting consistent expression trends in both microarray and qRT-PCR analyses were selected for further investigation. C qRT-PCR validation of the 4 selected circRNAs in patients (n = 25). D Schematic diagram showing the back-splicing junction of hsa_circ_0006168, validated by Sanger sequencing. E Agarose gel electrophoresis of cDNAs and gDNAs from a migraine patient’s peripheral blood. “►◄” represents divergent primers, “►◄” represents convergent primers. U6 was used as a negative control. F FISH was used to assess the expression and subcellular localization of mmu_circ_0012878 in the TNC of mice. Scale bar = 10 μm. G A ceRNA network predicted using TargetScan and miRanda, with the migraine-related regulatory axis highlighted. H-I The expression levels of hsa_circ_0006168 and miR-99b-5p were detected by qRT-PCR analysis (n = 23). J ELISA analysis of KDM6B expression levels (n = 23). Data are shown as the means ± SD. The p values were calculated using Student’s t-test (C and H-J). *p < 0.05, **p < 0.01, ***p < 0.001

Table 3.

General characteristics of enrolled patients

Characteristics	Control (n = 30)	Migraine (n = 45)	P value
Age(years)	41.50 (26.00, 56.50)	45.00 (31.00, 55.50)	0.7612
Sex			0.4701
Male	5 (16.67%)	4 (8.89%)
Female	25 (83.33%)	41 (91.11%)
Clinical Features
Disease Duration (years)	-	6 (3, 10)
Attack Frequency (/month)	-	4 (3, 5)
Attack Duration (hours)	-	8 (5, 24)

Central sensitization in the TNC of NTG-induced migraine mice with upregulated mmu_circ_0012878 and KDM6B, and downregulated miR-99b-5p

Cellular proto-oncogene (c-Fos) serves as an indicator of neuronal activation and mediates pain-related central sensitization [20]. In this study, hyperalgesia was utilized as the behavioral phenotype, and the expression levels of CGRP and c-Fos were assessed as molecular markers to comprehensively evaluate the functional state of central sensitization. The procedures for migraine model induction and behavioral testing are illustrated in Fig. 2A. There were no significant differences in baseline periorbital or plantar mechanical thresholds or paw heat pain thresholds between the two groups. Following the first NTG administration, the mechanical threshold decreased significantly, whereas the heat pain threshold of the paw began to decrease gradually after the second NTG administration, indicating the development of hyperalgesia in the NTG group (Fig. 2B). The TNC is a key region involved in the initiation and maintenance of central sensitization [9]. Western blot and immunofluorescence analyses revealed significantly elevated CGRP and c-Fos expression levels in the TNC of NTG-treated mice compared with those in the sham group (Fig. 2C–F). These results confirmed the successful establishment of the migraine mouse model and the occurrence of central sensitization. Furthermore, in the TNC of migraine model mice, mmu_circ_0012878 (Fig. 2G) and KDM6B (Fig. 2I, J) were markedly upregulated, whereas miR-99b-5p was downregulated relative to that in the sham group (Fig. 2H), indicating the potential role of the mmu_circ_0012878/miR-99b-5p/KDM6B axis in migraine pathogenesis.

Fig. 2.

Central sensitization with altered mmu_circ_0012878/miR-99b-5p/KDM6B expression in the TNC of migraine mice. A Flowchart of the migraine model induction and behavioral tests. B The mechanical thresholds were assessed using von frey filaments. The heat pain threshold was evaluated by recording thermal withdrawal latency (n = 10). C CGRP and c-Fos protein levels in the TNC analyzed by western blot (n = 6). D The red-marked area in the schematic brain section represented the TNC expression region. E Representative immunofluorescence image showing CGRP (red) expression in the TNC of migraine group. Scale bar = 500 μm. F Immunofluorescence staining of CGRP (red) in the TNC of each group (n = 3). Scale bar = 200 μm. G-I qRT-PCR analysis of mmu_circ_0012878, miR-99b-5p, and KDM6B mRNA in the TNC (n = 10). J KDM6B protein expression of TNC was examined by western blot (n = 6). Data are shown as the means ± SD. The p values were calculated using two-way ANOVA followed by Tukey’s post hoc tests (B) and Student’s t-test (C and F-J). *p < 0.05, **p < 0.01, ***p < 0.001

mmu_circ_0012878 promoted central sensitization and regulated miR-99b-5p and KDM6B in the TNC of migraine mice

To further investigate the role of mmu_circ_0012878 in central sensitization, LV-mmu_circ_0012878-RNAi was stereotaxically microinjected into the TNC of the migraine group (Fig. 3A). The downregulation of mmu_circ_0012878 significantly attenuated hyperalgesia, as reflected by the gradual recovery of the mechanical and paw heat pain thresholds (Fig. 3B). Western blot and immunofluorescence staining revealed that the expression levels of CGRP and c-Fos in the TNC of migraine model mice were significantly reduced following lentiviral injection (Fig. 3C–E). Taken together, these findings indicate that mmu_circ_0012878 affects hyperalgesia and CGRP/c-Fos expression in the TNC, suggesting its involvement in central sensitization. In addition, the silencing of mmu_circ_0012878 resulted in the upregulation of miR-99b-5p (Fig. 3F) and the downregulation of KDM6B (Fig. 3G) in the TNC of migraine model mice. These findings indicated that the inhibition of mmu_circ_0012878 reversed the miR-99b-5p suppression and KDM6B overexpression observed in migraine model mice.

Fig. 3.

Silencing mmu_circ_0012878 alleviated central sensitization, upregulated miR-99b-5p and downregulated KDM6B in the TNC of migraine mice. A LV-mmu_circ_0012878-RNAi was delivered via stereotactic injection into the TNC on day 10. B Assessment of mechanical and heat pain thresholds in each group of mice (n = 10). C Western blot analysis of CGRP and c-Fos in the TNC of migraine mice following mmu_circ_0012878 knockdown (n = 6). D Representative immunofluorescence image of CGRP (red) in the TNC of NTG+LV-circRNA-RNAi group. Scale bar = 500 μm. E Immunofluorescence staining of CGRP (red) in the TNC of various groups (n = 4). Scale bar = 200 μm. F-G After knockdown of mmu_circ_0012878, miR-99b-5p and KDM6B expression levels in the TNC assessed by qRT-PCR and western blot (n = 6). Data are shown as the means ± SD. The p values were calculated using two-way ANOVA followed by Tukey’s post hoc tests (B) and one-way ANOVA (C and E-G). ^#p < 0.05, ^##p < 0.01, *p < 0.05, **p < 0.01, ***p < 0.001

Overexpression of miR-99b-5p suppressed central sensitization and KDM6B expression in the TNC of migraine mice

The effect of miR-99b-5p on migraine was assessed by stereotaxic injection of a miR-99b-5p agomir into the TNC of migraine model mice (Fig. 4A). The PWP threshold was significantly increased on d 13, and the mechanical threshold tended to increase (Fig. 4B), demonstrating that miR-99b-5p overexpression partially relieved hyperalgesia in migraine model mice. Western blot and immunofluorescence analyses revealed that CGRP and c-Fos expression levels in the TNC were lower in the NTG + miR-99b-5p agomir group than in the NTG + agomir NC group (Fig. 4C–E), confirming the protective role of miR-99b-5p in migraine. Moreover, KDM6B expression in the TNC was decreased in migraine mice overexpressing miR-99b-5p (Fig. 4F), revealing that miR-99b-5p regulated KDM6B expression.

KDM6B potentially mediated central sensitization through the PI3K/AKT pathway in migraine mice

GSK-J4 was intraperitoneally injected to suppress KDM6B expression to explore the mechanisms by which KDM6B contributes to migraine (Fig. 5A). Western blot analysis revealed that CGRP expression in the TNC decreased in a dose-dependent manner following GSK-J4 treatment, with the lowest expression observed at 20 mg/kg/d (Fig. 5B). Accordingly, 20 mg/kg/d was selected as the optimal dose for subsequent experiments. Migraine mice exhibited significant increases in the mechanical and paw heat pain thresholds after GSK-J4 administration (Fig. 5C). Both c-Fos and CGRP protein expression in the TNC were markedly reduced, as confirmed by western blot and immunofluorescence, respectively (Fig. 5D, E). These findings suggest that KDM6B is strongly linked to the development of central sensitization in migraine patients. Additionally, GO and KEGG enrichment analyses were performed on the mRNAs associated with hsa_circ_0006168, and the PI3K-AKT signaling pathway was highly enriched (Fig. 5F). Western blot analysis further demonstrated that inhibition of KDM6B blocked activation of the PI3K/AKT pathway (Fig. 5G), indicating that KDM6B may mediate migraine pathogenesis via the PI3K/AKT signaling cascade.

Fig. 5.

KDM6B regulated central sensitization through activation of the PI3K/AKT pathway in migraine mice. A Schematic diagram of the GSK-J4 intervention procedure. B CGRP protein expression levels in the TNC of migraine mice treated with different concentrations of GSK-J4 detected by western blot (n = 6). C Behavioral tests were conducted every other day (n = 10). D Western blot detection of c-Fos expression in the TNC of mice following administration of 20 mg/kg/d GSK-J4 (n = 6). E Immunofluorescence visualization of CGRP (red) in the TNC across different treatment groups (n = 3). Scale bar = 200 μm. F GO functional and KEGG pathway enrichment analysis of hsa_circ_0006168, with significantly enriched pathway highlighted in red box. G Western blot analysis of PI3K/AKT pathway proteins in the TNC of mice (n = 6). Data are shown as the means ± SD. The p values were calculated using one-way ANOVA (B, D-E and G) and two-way ANOVA followed by Tukey’s post hoc tests (C). ^#p < 0.05, ^##p < 0.01, ^###p < 0.001, *p < 0.05, **p < 0.01, ***p < 0.001

Inhibition of mmu_circ_0012878 alleviated microglial activation and promoted M2 polarization in the TNC of migraine mice, and LPS-stimulated microglia showed increased mmu_circ_0012878/KDM6B and decreased miR-99b-5p expression

The cellular mechanisms of mmu_circ_0012878 were investigated via FISH combined with immunofluorescence staining to assess its colocalization with the microglial marker ionized calcium-binding adapter molecule 1 (Iba1), the astrocyte marker glial fibrillary acidic protein (GFAP), and the neuronal marker neuronal nuclei (NeuN) in the TNC of mice. Compared with the sham group, mmu_circ_0012878 partially colocalized with Iba1 and NeuN, showing more pronounced colocalization with Iba1 (Fig. 6A, C, D and F), whereas no apparent colocalization was observed with GFAP (Fig. 6B, E) in the TNC of the migraine group, suggesting that mmu_circ_0012878 may exert its effects on migraine through microglia in the TNC. Next, we performed immunofluorescence costaining of Iba1 with CD206 and CD86 to evaluate the specific effects of mmu_circ_0012878 on microglia. In the migraine model mice, microglia were markedly activated, characterized by hypertrophic somata and shortened, thickened processes (Fig. 7A, B). However, the inhibition of mmu_circ_0012878 partially suppressed microglial activation, significantly reduced the number of Iba1+/CD86+ double-positive cells, and markedly increased the number of Iba1+/CD206+ double-positive cells (Fig. 7A–E). These findings indicated that mmu_circ_0012878 contributed to microglial activation in the TNC and that its silencing suppresses M1 polarization and promotes microglial polarization toward the M2 phenotype. Subsequently, in in vitro experiments, BV2 cells were activated with 1 μg/ml LPS. Western blot analysis revealed that KDM6B expression began to increase at 3 h following LPS treatment and peaked at 12 h (Fig. 7F). Furthermore, mmu_circ_0012878 was significantly upregulated (Fig. 7G), whereas miR-99b-5p was downregulated after 12 h of LPS stimulation (Fig. 7H), suggesting that the mmu_circ_0012878/miR-99b-5p/KDM6B axis was closely associated with microglial activation.

Fig. 6.

Co-localization between mmu_circ_0012878 with cell-specific markers in the TNC of mice. A Immunofluorescence labeling of mmu_circ_0012878 (red) and Iba1 (green) in the TNC of control and migraine mice. Co-localized cells were marked by white arrows. B Representative immunofluorescence images showing co-localization of mmu_circ_0012878 (red) with GFAP (green). C shows co-labeling of mmu_circ_0012878 (red) with NeuN (green). D-F show the quantification of double-positive cells. Scale bar = 500 μm, 100 μm, 10 μm

Fig. 7.

The effect of mmu_circ_0012878 on microglia in the TNC of mice and LPS-treated BV2 cells. A Representative immunofluorescence images of Iba1 (green) and CD206 (red) co-localization in the TNC following mmu_circ_0012878 knockdown (n = 3). Scale bar = 20 μm. B Representative microglial morphology from each group in the areas marked by white boxes in A. Scale bar = 5 μm. C Qulification of Iba1+/CD206+ double-positive cells (n = 3). D Representative immunofluorescence images showing co-localization of Iba1 (green) and CD86 (red) in the TNC following mmu_circ_0012878 knockdown (n = 3). Scale bar = 20 μm. E Qulification of Iba1+/CD86+ double-positive cells (n = 3). F Western blot analysis of KDM6B at different time points following 1 μg/ml LPS stimulation (n = 3). G-H the relative expression of mmu_circ_0012878 and miR-99b-5p in LPS-stimulated BV2 cells were measured using qRT-PCR (n = 4). Data are shown as the means ± SD. The p values were calculated using one-way ANOVA (C-F). *p < 0.05, **p < 0.01, ***p < 0.001

mmu_circ_0012878 functioned as a sponge for miR-99b-5p to regulate KDM6B

mmu_circ_0012878 was successfully knocked down in BV2 cells (Fig. 8A). Further experiments revealed that the inhibition of mmu_circ_0012878 reversed the LPS-induced downregulation of miR-99b-5p (Fig. 8B), upregulation of KDM6B (Fig. 8C, E) and activation of the PI3K/AKT pathway (Fig. 8D). In addition, LPS stimulation resulted in hypertrophic somata in BV2 cells, whereas these microglial activation-associated morphological changes were mildly attenuated following mmu_circ_0012878 knockdown (Fig. 8E). These findings demonstrated that mmu_circ_0012878 may promote microglial activation by regulating the downstream miR-99b-5p/KDM6B axis.

Fig. 8.

mmu_circ_0012878 acted as a miR-99b-5p sponge to modulate KDM6B expression. A qRT-PCR analysis confirmed the knockdown efficiency of mmu_circ_0012878 (n = 3). B Relative expression levels of miR-99b-5p in each group were assessed by qRT-PCR (n = 3). C Western blot analysis of KDM6B expression in BV2 cells (n = 3). D Western blot analysis of PI3K/AKT pathway protein expression in BV2 cells following mmu_circ_0012878 knockdown (n = 3). E Immunofluorescence staining showing KDM6B expression and morphological changes in microglia (n = 3). Scale bar = 50 μm. F Western blot was performed to investigate the regulatory relationships among mmu_circ_0012878, miR-99b-5p, and KDM6B (n = 3). G Expression of PI3K/AKT pathway proteins in LPS-treated BV2 cells co-transfected with LV-mmu_circ_0012878-RNAi and KDM6B overexpression plasmids (n = 3). H-I Predicted binding sites among mmu_circ_0012878, miR-99b-5p, and KDM6B are shown schematically. The predicted interactions were validated using dual luciferase assays (n = 3). Data are shown as the means ± SD. The p values were calculated using one-way ANOVA. *p < 0.05, **p < 0.01, ***p < 0.001

To validate the interaction among mmu_circ_0012878, miR-99b-5p, and KDM6B, BV2 cells with mmu_circ_0012878 knockdown were transfected with the miR-99b-5p inhibitor or mimic. Compared with the inhibitor NC, the miR-99b-5p inhibitor significantly increased KDM6B expression in LPS-treated BV2 cells with mmu_circ_0012878 knockdown. Conversely, compared with the mimic NC group, the miR-99b-5p overexpression group presented a decreasing trend in KDM6B expression, although the difference was not statistically significant (Fig. 8F). Meanwhile, in BV2 cells with mmu_circ_0012878 knockdown, further overexpression of KDM6B markedly restored the suppression of the PI3K/AKT pathway caused by mmu_circ_0012878 inhibition (Fig. 8G). The dual-luciferase assay results revealed that the overexpression of miR-99b-5p suppressed the luciferase activity of mmu_circ_0012878-WT and KDM6B-WT-UTR but had no notable effect on mmu_circ_0012878-MUT or KDM6B-MUT-UTR (Fig. 8H, I). These findings indicate that mmu_circ_0012878 directly binds to miR-99b-5p and that miR-99b-5p specifically targets KDM6B.

Knockdown of mmu_circ_0012878 suppressed LPS-induced M1 polarization and promoted M2 polarization in BV2 cells

We further examined the effect of mmu_circ_0012878 knockdown on microglial polarization following LPS stimulation. Inducible nitric oxide synthase (iNOS) and CD86 were used to indicate M1 polarization, whereas arginase-1 (Arg-1) and CD206 served as markers for M2 polarization [28]. Compared with the LV-NC group, the mmu_circ_0012878-silenced group presented a nonsignificant decrease in iNOS protein expression (Fig. 9A); however, the mean fluorescence intensity of CD86 was markedly decreased (Fig. 9B). In contrast, Arg-1 protein expression (Fig. 9A) and CD206 fluorescence intensity (Fig. 9C) were both elevated. Moreover, a multiplex flow cytometry assay was conducted to detect M1 proinflammatory and M2 anti-inflammatory cytokines. Compared with the sham group, the LPS-treated BV2 cells presented significantly elevated levels of MCP-1, IL-6, G-CSF, IL-12p70, IL-1β, IFN-γ, IL-10, IL-17, IL-4, IL-2, IL-23p19, KC, and TNF-α (fold change ≥ 1.2, p < 0.05; Fig. 9D). Upon knockdown of mmu_circ_0012878, the expression of IL-12p70, IL-6, IL-1β, MCP-1, and G-CSF was markedly reduced (fold change ≥ 1.2, p < 0.05; Fig. 9E). Next, we measured the levels of IL-6, IL-1β, and IL-12p70 in each group using ELISA. We found that inhibition of mmu_circ_0012878 suppressed the expression of IL-6 and IL-1β (Fig. 9F. G), whereas, in contrast to the microarray results, the expression of IL-12p70 did not show a significant change (Fig. 9H). IL-6 and IL-1β are known as M1-associated proinflammatory cytokines, in summary, these results confirmed that mmu_circ_0012878 knockdown promoted a shift toward M2 polarization and inhibited M1 polarization in LPS-stimulated BV2 cells.

Fig. 9.

Silencing mmu_circ_0012878 attenuated LPS-induced M1 polarization while enhancing M2 polarization in BV2 cells. A The protein expression levels of iNOS and Arg-1 were examined by western blot in each group (n = 3). B-C Flow cytometry was performed to analysis the MFl of CD86 and CD206 (n = 3). D-E The quantification of cytokines from multiplex flow cytometry assay (n = 3). F-H ELISA analysis of IL-6, IL-1β and IL-12p70 expression levels (n = 4). Data are shown as the means ± SD. The p values were calculated using one-way ANOVA. *p < 0.05, ***p < 0.001

Discussion

Based on gene microarray and bioinformatics validation, this study identified the hsa_circ_0006168/miR-99b-5p/KDM6B axis was associated with migraine. We further demonstrated that this axis contributed to migraine pathogenesis by promoting central sensitization through microglial activation in the TNC and was closely linked to the PI3K/AKT pathway. This study provides new insights into the pathophysiological mechanisms of migraine and suggests potential biomarkers and therapeutic targets for its diagnosis and treatment.

hsa_circ_0006168 is a 395 bp circular RNA derived from the 4q21.1 chromosomal region and generate from exons of the CNOT6L mRNA. In vivo experiments we confirmed its cytoplasmic localization. Cytoplasmic circRNAs typically exert their regulatory effects through the ceRNA mechanism. CeRNAs regulate gene expression by competitively binding to shared miRNAs through microRNA response elements [29, 30]. Substantial evidence has supported the crucial role of the ceRNA mechanism in various neurological disorders. In migraine research, a NEAT1/miR-196a-5p ceRNA network was identified in the trigeminal ganglia of photophobic mice, where NEAT1 promoted photophobia via the NEAT1/miR-196a-5p/Trpm3 axis, offering a novel mechanistic explanation for this migraine-associated symptoms [31]. Furthermore, in a rat model of chronic pain-induced depression, the lncRNA XR_351665 sponged miR-152-3p, leading to the upregulation of DNMT1 and contributing to the development of depressive-like behaviors [32]. Our study identified the migraine-associated hsa_circ_0006168/miR-99b-5p/KDM6B axis. Migraine is increasingly recognized as a systemic disorder characterized by coordinated dysregulation across the central nervous, peripheral nervous, immune, and endocrine systems. Consequently, the altered expression of the hsa_circ_0006168/miR-99b-5p/KDM6B axis observed in peripheral blood likely represents a secondary manifestation of systemic disease responses during migraine attacks rather than a direct indicator of central pathological activity. Accumulating evidence supports a bidirectional interaction between peripheral and central systems in migraine, wherein peripheral molecular changes can modulate central nociceptive processing [33]. Thus, peripheral blood alterations alone are inadequate to fully elucidate disease - specific central mechanisms. In this context, our functional validation in the TNC of migraine mouse model provides critical mechanistic evidence that this molecular axis directly regulates central nociceptive nuclei, thereby reinforcing its causal role in migraine pathophysiology. Through in vivo and in vitro experiments, we demonstrated that mmu_circ_0012878 functions via the ceRNA mechanism by acting as a sponge for miR-99b-5p, thereby alleviating miR-99b-5p-mediated repression of its target KDM6B, which contributes to migraine pathogenesis.

Studies have indicated that hsa_circ_0006168 may contribute to esophageal carcinogenesis by sponging miR-100 and modulating mTOR [34]. Notably, miR-100 has anti-inflammatory effects in chronic inflammatory conditions [35]. As a critical regulator of autophagy, mTOR is a key component of the PI3K/AKT/mTOR pathway, and our previous study confirmed that mTOR inhibition attenuated migraine-associated central sensitization [36]. Accordingly, hsa_circ_0006168/miR-100/mTOR axis may represent a potential alternative pathway by which hsa_circ_0006168 contributes to migraine pathogenesis. Wang et al. [37] reported that hsa_circ_0006168 acts as a ceRNA for miR-628-5p, modulating the IGF1R/Ras/Erk signaling pathway to suppress proliferation and migration in glioblastoma. The downstream target IGF-1R is implicated in migraine pathogenesis, as it contributes to central sensitization and activates the PI3K/AKT/mTOR signaling pathway [38]. Overall, considering the regulatory effect of mmu_circ_0012878 on central sensitization found in this study, it can be concluded that central sensitization and the PI3K/AKT/mTOR pathway are pivotal mechanisms and downstream effectors underlying the role of hsa_circ_0006168 in migraine. However, the precise mechanistic relationship between hsa_circ_0006168 and the PI3K/AKT pathway remains to be further elucidated.

miR-99b-5p expression is downregulated in patients with headache associated with reversible cerebral vasoconstriction syndrome [39]. However, its specific function in migraine remains unexplored. In spinal cord injury (SCI) models, miR-99b-5p has been shown to regulate synaptic plasticity and exert neuroprotective effects at the synaptic level, thereby alleviating central sensitization and slowing the progression of SCI [40]. Moreover, miR-99b-5p has been confirmed to target IGF-1R [41]. Gujrati et al. [42] further identified the miR-99b-5p/mTOR regulatory axis as a critical determinant in prostate cancer. Collectively, evidence from previous studies and our findings suggests that miR-99b-5p contributes to migraine pathogenesis by regulating central sensitization, potentially via the PI3K/AKT/mTOR pathway.

KDM6B is a histone demethylase that catalyzes the removal of methyl groups from di- and trimethylated histone H3 lysine 27 (H3K27me2/me3) via its iron-containing JmjC domain, thereby promoting gene expression [43]. Its dysfunction has been implicated in a range of conditions, including neurodevelopmental disorders and hematologic malignancies [44]. Qiao et al. [45] reported that upregulation of KDM6B exacerbates TNF-α-dependent inflammatory pain in a rat model. This finding indicates that KDM6B functions in regulating the expression of proinflammatory cytokines and chemokines. In addition, the pan-KDM inhibitor JIB-04 has been demonstrated to inhibit the PI3K/AKT pathway, validating the regulatory role of KDM6B in this signaling cascade [46]. The PI3K/AKT pathway is a central regulator of cell growth, proliferation, motility, metabolism, and survival and is critically involved in the pathogenesis of numerous human diseases [47]. Specifically, this pathway has been implicated in TNC microglial activation and subsequent modulation of migraine-associated central sensitization [48]. These findings, consistent with our results, support the involvement of KDM6B in migraine-associated central sensitization via modulation of the PI3K/AKT pathway. Notably, KDM6B functions as epigenetic regulator through histone demethylation and transcriptional alterations, which generally require more time to manifest at the protein and behavioral levels. In this study, the mechanical and paw heat pain thresholds in the NTG-induced migraine mice showed improvement as early as 3 days after mmu_circ_0012878 lentiviral injection. We posit that these early changes may be mediated by other downstream effectors triggered by mmu_circ_0012878 suppression and the rapid restoration of miR-99b-5p levels, such as mTOR and IGF-1R. In contrast, KDM6B-mediated effects are more prominent at later time points. Collectively, these results suggest that the observed behavioral changes result from a combination of early hsa_circ_0006168/miR-99b-5p axis actions and subsequent epigenetic regulation by KDM6B, highlighting the temporal dynamics of the hsa_circ_0006168/miR-99b-5p/KDM6B axis in central sensitization.

Microglia are the primary innate immune cells involved in neuroinflammation. The activation of microglia in the TNC promotes central sensitization, which is a major contributor to the development of hyperalgesia and the chronification of migraine [25]. Activated microglia undergo morphological changes, including soma enlargement and process retraction, enhancing their ability to interact neurons and influence central pain transmission [20]. Functionally, activated microglia exhibit increased phagocytic activity and migratory capacity, however, excessive uptake of pathological proteins can impair their phagocytic function, thereby triggering neuroinflammation and exacerbating central sensitization [49].

Researches on mmu_circ_0012878 are primarily cancer-related. He et al. [50] reported that in ischemic stroke, hsa_circ_0006168 acts as a miRNA sponge for miR-99a-5p, thereby modulating serpin expression to attenuate astrocyte apoptotic signaling and mitigate posthypoxic injury, suggesting its potential role in glial cells. Regarding miR-99b-5p, study indicated that high preoperative plasma levels of miR-99b-5p correlate with reduced pain one year after surgery in lumbar spinal stenosis patients, implying that miR-99b-5p may exert anti-inflammatory effects [51]. Futhermore, a recent study revealed that miR-99b-5p is downregulated in the peripheral blood of schizophrenia patients and in the prefrontal cortex of model mice. Inhibition of miR-99b-5p in microglia increases phagocytic activity and reduces synaptic density, while miR-99b-5p regulates inflammation by targeting Zbp1 [52], supporting its anti-inflammatory role mediated through microglia. Previous study has demonstrated that KDM6B is involved in the development and maintenance of neuropathic pain following peripheral nerve injury [53]. It mediates neuropathic pain in a lumbar spinal stenosis animal model by inducing macrophage infiltration and activation [54]. KDM6B serves as a regulatory factor in the expression of pro-inflammatory mediators and inflammatory pain [55]. Moreover, KDM6B is implicated in microglial activation. Studies have shown that silencing KDM6B in LPS-stimulated microglia suppresses the expression of IL-6 and iNOS [56]. Similarly, GSK-J4 alleviates the upregulation of IL-1β and Iba1 in the hippocampus of LPS-treated mice [57]. These findings suggest that KDM6B may be closely associated with microglia-mediated inflammatory responses. Taken together with the results of our study, these observations indicate that the hsa_circ_0006168/miR-99b-5p/KDM6B axis may mediate inflammation through microglial activation.

Limitations

Several limitations warrant consideration. (1) Behavioral assessments of mechanical pain thresholds may have been affected by gaze avoidance behaviors from the third drug administration, potentially skewing data despite methodological adjustments. (2) The maximal tested dose of GSK-J4 (20 mg/kg/d) demonstrated therapeutic efficacy, but higher doses are needed to fully delineate its pharmacodynamic profile. (3) This study did not perform in vivo rescue experiments, and thus the causal role of the hsa_circ_0006168/miR-99b-5p/KDM6B axis in central sensitization and TNC microglial activation cannot be fully validated in animal models. Further investigations in migraine mice are required to clarify and confirm the functional mechanisms of this axis. (4) Additionally, heterogeneity in patient medication use and variability in lentivirus transfection efficiency could introduce interpretive constraints. (5) The expression of the mmu_circ_0012878/miR-99b-5p/KDM6B axis was not examined in the peripheral blood of migraine model mice.

Conclusion

This study identified the hsa_circ_0006168/miR-99b-5p/KDM6B axis as a novel regulatory mechanism underlying migraine. By promoting PI3K/AKT-mediated microglial activation in the TNC, this axis facilitates central sensitization and disease progression. These findings lay a foundation for future exploration of circRNA-based biomarkers and therapeutic targets in migraine pathophysiology.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

CGRP

Calcitonin gene-related peptide

SP

Substance P

TNC

Trigeminal nucleus caudalis

CircRNA

Circular RNA

MiRNA

microRNA

IL6

Interleukin-6

FOSL2

FOS-like 2

Gdna

Genomic DNA

NTG

Nitroglycerin

c-Fos

Cellular proto-oncogene

Iba1

Ionized calcium-binding adapter molecule 1

GFAP

Glial fibrillary acidic protein

NeuN

Neuronal nuclei

iNOS

Inducible nitric oxide synthase

Arg-1

Arginase-1

SCI

Spinal cord injury

Author contributions

QHC and QLZ contributed equally to this work. YHP, JHL and QHC designed this study. Experiments were performed by QHC, YNH, HYL and CMZ. Statistical analysis and manuscript preparation were conducted by QLZ, QHC and QJY. YHP and TTS provided supervision and final review. All the authors read the final version of this paper and approved it.

Funding

Project funded by General Program of National Natural Science Foundation of China (No. 82071549, 82371483); Key Research & Development Program of Heilongjiang (No. 2023ZX06C02); and Youth Foundation of the first Affiliated Hospital of Harbin Medical University (No. 2024YQ22).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

This study was approved by the Ethics Committee of the First Affiliated Hospital of Harbin Medical University (IRB-AF/SC-12/03.0), and informed consent was obtained from all participants. All animal experiments and related ethical considerations were approved by the Institutional Animal Care and Use Committee of the First Affiliated Hospital of Harbin Medical University (YS166).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.