Compromised PGF2α signaling in the paraventricular hypothalamic nucleus contributes to the central sensitization of the nitroglycerin-induced chronic migraine in male mice

Abstract

Impaired descending inhibitory controls are now understood to exacerbate central sensitization of chronic migraine, yet the underlying neural and molecular mechanisms remain largely elusive. Herein, a paraventricular hypothalamic nucleus (PVN) ^{oxytocin (OXT)} → trigeminal nucleus caudalis (TNC) ^GABA neural circuit was identified through the application of a rigorous anterograde tracing strategy and RNAscope in situ hybridization techniques, involved in regulating trigeminal nociceptive transmission. In both episodic and chronic migraine mouse models induced by nitroglycerin (NTG) injections, increased activity of the PVN ^OXT → TNC ^GABA circuit was observed. However, the activity of PVN ^OXT neurons decreased in chronic migraine mice when compared to episodic migraine mice. Chemogenetic activation of PVN ^OXT neurons alleviated migraine hyperalgesia and enhanced the release of OXT and GABA in TNC in chronic migraine mice, while these beneficial effects were abrogated by the intra-TNC administration of OXT receptor antagonist. Interestingly, the expression of prostaglandin F2α receptor (FP) in PVN ^OXT neurons decreased with the chronification of migraine despite upregulation of PVN prostaglandin F2α (PGF2α) levels. Targeting FP overexpression in PVN ^OXT neurons restored neuronal activity and ameliorated chronic migraine hyperalgesia. Overall, our study reveals a novel neural and molecular mechanism for descending modulation of trigeminal central sensitization, thereby providing a basis for treating chronic migraine.

Graphic abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s10194-025-02124-x.

Introduction

Migraine is a prevalent and recurrent primary headache disorder characterized by unilateral, throbbing, moderate-to-severe pain, often accompanied by symptoms such as nausea, vomiting, photophobia, and phonophobia. Its etiology is influenced by both genetic and environmental factors, with a global prevalence of approximately 15% [1]. The World Health Organization (WHO) ranks migraine as the second most disabling disease worldwide [2]. Epidemiological evidence suggests that approximately 3% of patients with episodic migraine progress to chronic migraine annually, a condition defined by experiencing migraine headaches on at least 15 days per month for more than 3 months [3]. A key mechanism in chronic migraine involves the occurrence of central sensitization in the trigeminal nucleus caudalis (TNC), the primary relay site for nociceptive trigeminal afferents [4]. Central sensitization is a pathophysiological process characterized by structural, functional, and neurochemical changes in the central nervous system (CNS), leading to increased sensitivity to both noxious and innocuous stimuli [5]. Specifically, atypical release of nociceptive molecules, aberrant microglial and astrocytic activation, and abnormalities in multiple brain regions promote the dysregulation of top-down pain modulation, which is essential for the establishment of trigeminal central sensitization [6, 7]. Notably, extensive research has shown that impaired descending inhibitory controls aggravate trigeminal central sensitization [8].

Clinical studies have revealed the involvement of the hypothalamus in migraine pathophysiology by observing its abnormalities during migraine attacks [9]. In the context of migraine regulation, the hypothalamus is widely thought to affect the activities of TNC neurons, thereby modulating the trigeminal nociceptive transmission [10, 11]. The paraventricular hypothalamic nucleus (PVN), a critical subregion of the hypothalamus, regulates metabolic homeostasis and other physiological functions through the orchestration of multiple neuroendocrine neurons [4]. Oxytocin (OXT) neurons constitute a major neuronal type within the PVN, projecting to widespread brain regions, where they secrete oxytocin, a neuropeptide that binds to oxytocin receptors (OXTR) and elicits analgesic and anxiolytic effects [12–14]. Recent studies have revealed that PVN ^OXT neurons exert a descending pain-inhibitory effect in migraine central sensitization by enhancing the activity of TNC ^GABA neurons through the release of OXT from their projection terminals [10, 15]. Therefore, PVN ^OXT neurons represent a potential target for preventing the development of chronic migraine.

Accumulating evidence demonstrates the critical role of prostaglandins (PGs) in migraine pathogenesis [16]. While vasodilating PGs such as PGE2, PGI2, and PGD2 are known to induce migraine attacks directly, the vasoconstricting PGF2α does not trigger migraine attacks in healthy volunteers [17]. However, elevated salivary and urinary PGF2α have been documented in migraine sufferers [18, 19], suggesting the potential involvement of PGF2α in migraine pathophysiology. The binding of PGF2α to prostaglandin F receptor (FP) facilitates the generation of second messengers, including diacylglycerol (DAG) and inositol-1,4,5-trisphosphate (IP3), through the activation of phospholipase C (PLC). The upregulation of these second messengers promotes the transcription of immediate-early response genes, including c-fos and c-jun, which are rapidly expressed in cells upon exposure to various extracellular stimuli and are considered indicators of neuronal activation [20–22]. Besides, PGF2α signaling accelerates the production of another second messenger, cAMP, which regulates cell activities in various biological processes [23]. Single-cell transcriptomic analysis has revealed high expression of FP in PVN ^OXT neurons [24]. Interestingly, PGF2α signaling has been implicated in the activation of PVN neurons by promoting the expression of immediate early genes [22].

Current evidence indicates that OXT neurons mediate descending pain inhibition through projections from the PVN to the TNC, where they suppress primary nociceptive inputs via GABAergic interneurons [10]. Accordingly, we hypothesized that PGF2α signaling could reduce migraine central sensitization by activating PVN ^OXT neurons, thereby enhancing the inputs of PVN ^OXT neurons to TNC ^GABA neurons. To validate this hypothesis, the activity of PVN ^OXT neurons and PGF2α signaling was assessed in a chronic migraine mouse model induced by repeated NTG injections. Using a Cre-dependent anterograde tracing strategy, the neural connectivity between the PVN ^OXT neurons and TNC ^GABA neurons was delineated. Next, we conducted a series of experiments, including targeted chemogenetic modulation and neuropharmacological interventions, to investigate the role of PVN ^OXT neurons in the sensitization of chronic migraine hyperalgesia. Finally, a neuronal-specific genetic modulation strategy was employed to investigate whether PGF2α signaling could influence chronic migraine through PVN ^OXT→TNC ^GABA circuit augmentation, thereby providing a theoretical basis and therapeutic target for the clinical treatment of chronic migraine.

Methods and materials

Animals

All experiments in this study were conducted using 8-12-week-old, wild-type or OXT-Cre C57BL/6J male mice (weighing 20–25 g, purchased from Charles River or Jackson Laboratory), which were housed in groups of 2–4 per cage under controlled conditions (temperature: 23–25℃, humidity: 50%, and a 12-hour light/dark cycle with the light phase beginning at 8:00 AM). All animals had ad libitum access to food and water. Before each experiment, mice were numbered in ascending order of their body weight and randomly assigned to groups using the random number table method. All experimental procedures were conducted in accordance with institutional guidelines and approved by the Institutional Animal Care and Use Committee (IACUC) of Qingdao University.

Establishment of the chronic migraine mouse model

NTG, a nitric oxide donor, was systemically administered to establish both episodic and chronic migraine mouse models, as previously described in the literature [25]. To prepare the solution for administration, NTG (Yimin Pharmaceutical Co., Ltd., Beijing, China) was diluted to 1 mg/ml in 0.9% saline. For the episodic migraine model, mice were administered a single intraperitoneal injection of NTG (10 mg/kg). For the chronic migraine model, mice received five intraperitoneal injections of NTG (10 mg/kg) every other day over a 9-day period (Fig. 1A). Behavioral tests were conducted 1 h before the first to fifth NTG injection. Significant periorbital mechanical and thermal allodynia, characterized by the decreased periorbital mechanical threshold and increased acetone response time, were confirmed before the third to fifth NTG injections on days 5–9 (Fig. 1B and C). Since sustained craniofacial hyperalgesia is a hallmark of chronic migraine [26], the behavioral test results confirmed successful establishment of the chronic migraine mouse model by repeated NTG injections.

Fig. 1.

Repeated NTG injections established chronic migraine mouse model. A Experimental schedule of establishment of chronic migraine in mice by repeated NTG injections. B and C Von Frey and acetone tests showed the effects of NTG or NS injection on mechanical and thermal allodynia. The data are presented as the mean ± SEM. *P<0.05, **P<0.01, vs. NS group, Bonferroni post-hoc test after Repeated measure ANOVA, n = 8 (B: NTG, F_1,14=29.963, P < 0.001, ges = 0.682; time: F_1,14=41.609, P < 0.001, ges = 0.748; interaction, F_1,14=39.693, P < 0.001, ges = 0.739; 95%CI 0.723–0.851. C: NTG, F_1,14=85.698, P < 0.001, ges = 0.860; time: F_1,14=69.259, P < 0.001, ges = 0.832; interaction, F_1,14=81.703, P < 0.001, ges = 0.854; 95%CI 3.449–4.012)

Cell culture and transfection

Primary mouse hypothalamic neurons (CPM242, Procell, Wuhan, China) were seeded into 24-well plates at a density of 5–8 × 10⁴ cells per well in a mixture of DMEM/F-12 (1:1, Invitrogen) supplemented with 10% fetal bovine serum (Gibco). The cells were then randomly divided into two groups: the Control group and the Overexpression group. Each group was tested in three independent trials, with two replicates per trial to ensure reproducibility. In the Overexpression group, the Cre virus (AAV-hSyn-Cre-EGFP, titer: 1.00 × 10¹³ vg/mL, OBiO, Shanghai, China) was mixed with the Cre-dependent PTGFR gene (encoding FP) overexpression virus (AAV-CMV-DIO-PTGFR, titer: 5 × 10¹² vg/mL, OBiO) at a 1:2 ratio, and 5 µL of the mixture was added to each well. In the Control group, the AAV-hSyn-Cre-EGFP was mixed with a non-encoding control virus (AAV-CMV-DIO-mCherry, titer: 5 × 10¹² vg/mL, OBiO) and added to the wells at equivalent titers and volumes as the Overexpression group. After 48 h of viral transfection, the culture medium was refreshed, and the cells were cultured for an additional 5 days. The transfection efficiency was assessed by quantifying the expression of FP at the end of the culture period.

Behavioral assessments

Behavioral tests were conducted 2 h before NTG injection on days 1, 3, 5, 7, and 9, and at 1, 2, and 4 h after NTG injection on days 1 and 9. Mice were habituated to a quiet and dim (20 lx) testing environment for 30 min before the behavioral tests. Given that craniofacial hyperalgesia is a hallmark clinical manifestation in chronic migraine patients [26], mechanical and thermal allodynia were evaluated using Von Frey filaments and acetone assays, respectively, with emphasis on the trigeminal-innervated region (periorbital area). The experimental procedures were conducted in a blind manner.

Acetone test

The acetone test was used to assess thermal allodynia. In brief, mice were acclimated in a transparent glass chamber (9 cm × 5 cm × 5 cm) for 30 min following the periorbital hair removal. A 20 µL drop of acetone was applied to the periorbital area, and scratching or shaking the head was considered a positive response. Following acetone application, a high-resolution camera was placed in front of the chamber to record the mice’s performance immediately following the acetone application. The duration of positive responses within 1 min was calculated as the acetone response time.

Von Frey test

The Von Frey test was performed to detect mechanical pain sensitivity. The periorbital mechanical threshold was determined by applying Von Frey filaments (Stoelting, IL, USA) (from 0.04 to 2 g) perpendicularly to the periorbital skin with increasing pressure until the filament bent into a “C” shape, held for 3 s, using the up-down method to determine the threshold [27, 28]. The initial stimulus intensity was 0.4 g, and force was gradually increased until the mouse exhibited head withdrawal or shaking (positive response). The corresponding force value was recorded as the periorbital mechanical threshold. Measurements were repeated at 5-minute intervals, and the average value of three trials was taken as the final threshold. To avoid stress-induced false positives, mice underwent a 5-day habituation period prior to the periorbital mechanical threshold measurements. Mice with a baseline periorbital mechanical threshold below 0.16 g were excluded from the following measurements.

Stereotaxic AAV injection and cannula implantation

Mice were anesthetized with an intraperitoneal injection of 50 mg/kg pentobarbital (1% solution) and placed in a stereotactic frame (Kopf Instruments). The viral injections were performed using pulled glass microelectrodes (GC-3.5, RWD Life Science, Shenzhen, China) connected with a 10-microliter syringe. A volume of 100 nl of virus was injected into the PVN using an infusion pump (LEGATO 130, Kd Scientific) at a constant rate of 30 nl/min. The stereotactic injection coordinates were defined based on The Mouse Brain in Stereotaxic Coordinates (2nd edition, Paxinos and Franklin). The coordinates were identified by three indicators: dorsal-ventral (DV) from the brain surface, anterior-posterior (AP) from bregma, and medial-lateral (ML) from the midline. The corresponding coordinates for PVN were: AP, −0.7 mm; ML, ± 0.2 mm; DV, − 4.85 mm. For chemogenetic activation of PVN ^OXT neurons, OXT Cre mice received bilateral microinjections of either AAV-EF1α-DIO-hM3Dq-mCherry (AAV2/9, titer: 1.00 × 10¹² vg/mL, OBiO, Shanghai, China) or AAV-EF1α-DIO-mCherry (AAV2/8, titer: 1 × 10¹² vg/mL, OBiO, Shanghai, China) into the PVN. For overexpression of PTGFR in PVN ^OXT neurons, the designed PTGFR overexpression AAV vector, AAV-CMV-DIO-PTGFR (AAV2/9, titer:1 × 10¹² vg/mL, OBiO, Shanghai, China), was bilaterally injected into the PVN of OXT Cre mice. The AAV-CMV-DIO-mCherry (AAV2/9, titer:1 × 10¹² vg/mL, OBiO, Shanghai, China) was injected as the non-coding control.

For cannula implantation, custom-made stainless steel cannulae (O.D.0.41×I.D.0.25 mm, RWD Life Science, Shenzhen, China) were bilaterally implanted into the TNC (coordinates: −6.5 mm, ± 1.65 mm, − 4.00 mm) and secured on the skull using dental cement.

Drug administration

For chemogenetic manipulation of neuronal activity, Clozapine-N-oxide (CNO, 1 mg/kg [29], 20230718, Brain Case, Shenzhen, China), a synthetic ligand known to activate hM3Dq designer receptors in chemogenetic activation of neuronal activity, was intraperitoneally administered 30 min prior to the behavioral tests. To investigate the role of TNC OXTR in PVN ^OXT neuronal activation, mice received bilateral microinjections of vehicle or OXTR antagonist L-368,899 (1ug/200nl per side [30], HY-15008, Medchemexpress) into bilateral TNC through the cannula 15 min before CNO administration.

Western blot

Immediately following the completion of behavioral tests, mice were anesthetized to facilitate the immediate collection of PVN tissues. Briefly, mice underwent intracardiac perfusion with ice-cold PBS. Next, the brains were quickly extracted and frozen in liquid nitrogen. The frozen brains were placed in a rodent brain mold to dissect the coronal brain slices (1 mm) containing the PVN region (−0.4 mm to 1.4 mm relative to bregma) based on the mouse brain atlas. The slices were then quickly placed under a microscope to isolate the PVN regions. After viral transfection, the culture medium was centrifuged to harvest samples of cultured primary mouse hypothalamic neurons. The PVN tissues or cell samples were lysed using the ice-cold RIPA buffer. The samples were centrifuged to obtain the supernatant containing total protein. The concentrations of total protein were identified using a Bicinchoninic acid (BCA) kit. The proteins were eluted using SDS sample buffer, followed by protein separation via SDS-PAGE. Rabbit anti-FP antibody (1:1000, NBP2-89045, Novus) and mouse anti-Caveolin-1 (1:2000, ab32577, Abcam) were separately used as primary antibodies. Goat anti-rabbit secondary antibodies (1:5000, ab6721, Abcam) were used to react with the corresponding primary antibodies. Immunoreactive bands were visualized using enhanced chemiluminescence reagents. Immunoblotting densities of the bands were analyzed using the ImageJ software. Target protein expression levels were determined by normalizing band densities to Caveolin-1.

RT‑PCR

Total RNA of cell samples was extracted using an RNA extraction reagent (Tiangen, Beijing, China) according to the manufacturer’s protocol. Complementary DNA (cDNA) was synthesized from the isolated total RNA using PrimeScript™ RT Master Mix (Takara). qRT-PCR was performed using a Roche RT-PCR system (Switzerland) under standard conditions. Target gene expression levels were analyzed via the 2^−ΔΔCt method, normalized to the internal control GAPDH. The primer sequences used were as follows: PTGFR: sense 5′- GGAAAGAGAGGTGGAACCCG-3′, anti-sense 5′- CAACTGTGCAGTCTCGGAGT-3′; GAPDH: sense 5′- AAGAGGGATGCTGCCCTTAC-3′, antisense 5′- TACGGCCAAATCCGTTCACA-3′.

Immunofluorescence staining

The harvested brain tissues were fixed in 4% paraformaldehyde for 8–12 h and dehydrated in 30% sucrose solution for 24 h. Transverse brain sections (30 μm) were obtained using a microtome (CM1900; Leica, Wetzlar, Germany). Subsequently, the brain sections were blocked and permeabilized by incubating with 10% donkey serum containing 0.5% Triton X-100 for 1 h at room temperature. The sections were then incubated with the primary antibodies overnight at 4 °C. The primary antibodies were listed as follows: rabbit OXT antibody (1:500, ab212193, Abcam), mouse c-fos antibody (1:1000, ab208942, Abcam). On the next day, the sections were incubated with donkey anti-rabbit Alexa Fluor^® 488 secondary antibodies (1:1000, ab150113, Abcam) or donkey anti-mouse Alexa Fluor^® 568 secondary antibodies (1:1000, ab175700, Abcam) at room temperature for 1 h. Finally, the sections were incubated with 4’,6-diamidino-2-phenylindole (DAPI) and visualized using a scanning system (VS120, Olympus, Tokyo, Japan). For quantification, every fourth section within the series of coronal sections containing the PVN region was averaged per mouse. The quantification of c-fos expression in PVN ^OXT neurons was calculated as (c-fos + OXT + cells/OXT + cells) × 100, representing co-labeled cells in the population. The specificity of chemogenetic intervention was validated by assessing the co-localization of OXT in hM3Dq-expressing neurons in the PVN.

RNAscope in situ hybridization

RNAscope in situ hybridization was performed in accordance with the manufacturer’s protocol of RNAscope Fluorescent Multiplex Kit (320850, Advanced Cell Diagnostics). The protocol for the preparation of the brain sections used for RNAscope in situ hybridization was the same as immunofluorescence staining. Endogenous peroxidase activity was blocked by hydrogen peroxide for 10 min. Following three washes with distilled deionized water, the sections were treated with 1× Target Retrieval Agent at 100 °C for 5 min. After target retrieval, the sections were immediately immersed in distilled deionized water and washed three times. Next, the sections were successively dehydrated in ethanol with increasing concentrations of 50%, 70%, and 100% for 5 min each. After sample pretreatment, the sections were incubated with protease III for 30 min at 40 °C in a HybEZ II oven. Following three washes with 1× wash buffer, the sections were incubated with Slc17a7-C3 (1:50, 416631, Advanced Cell Diagnostics), Slc32a1-C2 (1:50, 319191, Advanced Cell Diagnostics), and PTGFR-C1 (501841, Advanced Cell Diagnostics) mRNA hybridization antisense probes for 2 h at 40 °C. Signal amplification was conducted using AMP1, AMP2, and AMP3 for 30 min each at 40 °C. Next, the sections were conjugated with HRP-C1, HRP-C2, or HRP-C3, respectively, stained with Opal Dye 520 or 570, and blocked with HRP-blocker (15 min, 30 min, and 15 min at 40 °C, respectively). After the RNAscope fluorescent staining, the sections were processed for immunofluorescence staining. The antibodies used for immunofluorescence staining were as follows: rabbit anti-mCherry primary antibodies (1:100, 43590, Cell Signaling Technology), rabbit EGFP primary antibodies (1:100, 2555, Cell Signaling Technology), donkey anti-rabbit Alexa Fluor^® 568 secondary antibodies (1:1000, ab175470, Abcam), and donkey anti-rabbit Alexa Fluor^® 488 secondary antibodies (1:1000, ab150113, Abcam). A confocal microscope (FV3000, Olympus, Tokyo, Japan) was used to capture images of the brain sections. Images were processed using the NIH Image J software. Specifically, image thresholds were standardized using identical parameters to optimize the detection of the target signal for quantification. The percentages of GABAergic and glutamatergic population in PVN ^OXT neuron innervating TNC neurons were calculated as: Slc32a1 mRNA + mCherry + cells/mCherry + cells × 100 or Slc17a7 mRNA + mCherry + cells/mCherry + cells × 100. The overexpression efficiency of the AAV-CMV-DIO-PTGFR was validated by quantifying the PTGFR mRNA relative intensity, calculated as the ratio of the total PTGFR mRNA + EGFP + pixel value to the total EGFP + pixel value within the region of interest (ROI).

Electrophysiology

Mice were anesthetized with 4% isoflurane and transcardially perfused with oxygenated (95% O₂/5% CO₂) artificial cerebrospinal fluid (ACSF) containing 126 mM NaCl, 2.5 mM KCl, 1.25 mM NaH₂PO₄, 25 mM NaHCO₃, 2 mM CaCl₂, 2 mM MgSO₄, and 10 mM glucose. Acute coronal slices (300 μm) containing the PVN region were obtained using a Leica VT1200s vibratome (Leica, Germany). The slices were then transferred to oxygenated ACSF and incubated at 35 °C for 30 min. Subsequently, the slices were kept at room temperature for an additional 30 min before recordings. After recovery, the slices were transferred to the recording chamber, where the hm3Dq-labeled neurons were selected using a microscope (BX51WI, Olympus, Japan). The patch pipettes (4–8 MΩ) were pulled and filled with an intracellular solution containing 133 mM K-gluconate, 8 mM NaCl, 0.6 mM EGTA, 2 mM MgATP, 0.3 mM Na3GTP, and 10 mM HEPES. Patch-clamp electrophysiology was performed using pClamp 10.0 software (Molecular Devices, USA) and a MultiClamp 700B amplifier. To validate the activation effects of the chemogenetic virus on the neurons, 10 µM CNO was added to the ACSF. hM3Dq-labeled neurons were selected and recorded for at least 2 min during CNO perfusion.

ELISA

1 mg of PVN or TNC tissue was homogenized and centrifuged to obtain the supernatant. The supernatant was then processed according to the manufacturer’s protocol of the ELISA kit (Saipei Biotechnology Co., Ltd., Wuhan, China) to detect the levels of PGF2α in PVN tissue, OXT and GABA in TNC tissue.

Statistical analysis

Data were presented as mean ± SEM and analyzed using the Statistical Package for the Social Sciences software (version 17.0; SPSS Inc., Chicago, IL, USA). Normality and homogeneity of variance were assessed using the Shapiro-Wilk test and Brown-Forsythe test, respectively. Normally distributed data with homogeneous variance were analyzed using the Unpaired t tests, One-way or Repeated measure ANOVA followed by Bonferroni’s post hoc tests. Other datasets were analyzed using a nonparametric test. The level of statistical significance was set at P < 0.05.

Results

PVN ^OXT neurons project to TNC ^GABA neurons

While the direct neural projections from the PVN to the TNC have been well established [31], the exact neural circuit underlying these connections remains unknown. Functionally, PVN stimulation has been established to enhance the activities of TNC ^GABA neurons, an effect abolished by intra-TNC administration of OXTR antagonist [10], indicating the potential link between PVN ^OXT neurons and TNC ^GABA neurons. Accordingly, an HSV (H129)-based anterograde transsynaptic tracing strategy was employed to delineate the structural connections between PVN ^OXT neurons and TNC neurons. Initially, a helper virus consisting of a mixture of AAV-UL26.5p-DIO-cmgD and AAV-hSyn-DIO-EGFP-T2A-Her2CT9 was injected into the PVN of OXT-Cre mice. 2 weeks following this initial injection, the anterograde tracer H129ΔgD-hUbC-mCherry-P2A-scHer2::gd was injected into the same site. Perfusion for assessment was conducted 7 days later (Fig. 2A–C). The mCherry reporter, encoded by the anterograde viral tracer, was expressed in TNC (Fig. 2D), indicating the presence of PVN ^OXT neurons-innervating neurons in the TNC. We further performed RNAscope in situ hybridization to confirm the neuronal type of these PVN ^OXT neurons-innervating neurons, which involved assessment of the expression of Slc17a6 mRNA (encoding vesicular glutamate transporter 2 (vGluT2), pan-glutamatergic marker) and Slc32a1 mRNA (encoding Vesicular GABA transporter (vGat), pan-GABAergic marker) within mCherry-immunoreactive cells. As a result, the mCherry-labeled PVN ^OXT neurons-innervating TNC neurons were predominantly GABAergic (Fig. 2E). Collectively, these mapping results confirmed a transsynaptic circuit of PVN ^OXT→TNC ^GABA.

Fig. 2.

Identification of the PVN ^OXT → TNC ^GABA circuit. A–C Experimental schedule and diagram of anterograde tracing virus injections in the PVN of OXT Cre mice. D Representative images showing the expression of H129 (mCherry) and helper (EGFP) virus in PVN (left) and TNC (right). E RNAScope and immunofluorescent co-staining showing the expression of Slc32a1 mRNA or Slc17a6 mRNA in PVN ^OXT neurons-innervating TNC neurons. Scale bar = 100/50µm. The data are presented as the mean ± SEM. **P < 0.01, vs. Slc32a1 + mCherry + neurons/mCherry + neurons, Unpaired t test, n = 8. t = 12.143, df = 14, P <0.001, cohen’d = 6.071, 95%CI 3.624–8.483

Persistent periorbital hyperalgesia and decreased activity of PVN ^OXT neurons in repeated NTG-induced chronic migraine mice

The episodic and chronic migraine mouse model was established by intraperitoneal injections of NTG, a nitric oxide donor. The behavioral tests, including the Von Frey and acetone tests, were performed to evaluate the mechanical and thermal allodynia, respectively. These tests were conducted at 1 h before and 1, 2, and 4 h after the first or 5th NTG injection on day 9, respectively (Fig. 3A). Significant periorbital mechanical and thermal allodynia, characterized by the decreased periorbital mechanical threshold and increased acetone response time, were confirmed at 1 and 2 h and disappeared at 4 h post-single NTG injection. This allodynia occurred pre- and post-fifth NTG injection (Fig. 3B–D). These behavioral results indicated that single and repeated NTG injections induced temporary and sustained craniofacial hyperalgesia in episodic and chronic migraine, respectively. It is well documented that PVN ^OXT neurons play a pivotal role in pain modulation by releasing the endogenous analgesic neuropeptide OXT [15, 32]. To investigate the involvement of these neurons in episodic and chronic migraine conditions, we performed c-fos (neuronal activation marker [33]) immunostaining to investigate the neuronal activity of PVN ^OXT neurons 4 h after single or repeated vehicle or NTG injections on day 9. Compared with the NS-injected control, single or repeated NTG-injected mice both exhibited significantly increased c-fos expression of PVN ^OXT neurons (Fig. 3E and F), indicating that NTG induced robust activation of PVN ^OXT neurons. However, compared with mice with single NTG-induced episodic migraine, mice with repeated NTG-induced chronic migraine showed a reduction of c-fos expression (Fig. 3E and F), suggesting the decreased activity of PVN ^OXT neurons in chronic migraine mice. Consistently, the levels of OXT and GABA in the TNC mirrored the activity changes in PVN ^OXT neurons (Fig. 3G and H). Based on our established neural circuit of PVN ^OXT → TNC ^GABA circuit, these findings suggested that decreased PVN ^OXT neuronal activity disrupted the equilibrium between the PVN ^OXT neurons and TNC ^GABA neurons.

Fig. 3.

Decreased activity of PVN ^OXT neurons in repeated NTG-induced chronic migraine mice. A Experimental schedule for the establishment of episodic and chronic migraine mouse models by single and repeated NTG injections, respectively. B–D Von Frey (C) and acetone (D) tests showing the mechanical and thermal allodynia among the four groups. E Representative images showing the expression of c-fos in PVN ^OXT neurons, scale bar = 50 μm. F Quantification of c-fos PVN ^OXT neurons after NTG or vehicle injections. G and H Quantification of OXT (G) and GABA (H) levels in TNC after NTG or vehicle injections. The data are presented as the mean ± SEM. ^#P < 0.05, ^##P < 0.01, vs. Single NS group, ^#P < 0.05, ^##P < 0.01, vs. Repeated NS group, ^&P < 0.05, ^&&P < 0.01, vs. Single NTG group, Bonferroni post-hoc test after Repeated measure ANOVA, n = 8 (C: NTG, F_3,28=81.935, P < 0.001, ges = 0.898; time, F_3,28=2.135, P = 0.155, ges = 0.071; interaction, F_3,28=2.487, P = 0.081 ges = 0.210; 95%CI 0.671–0.761; D: NTG, F_3,28=102.394, P < 0.001, ges = 0.916; time, F_3,28=2.272, P = 0.143, ges = 0.075; interaction, F_3,28=1.512, P = 0.233 ges = 0.139; 95%CI 4.047–4.761), Bonferroni post-hoc test after One way ANOVA, n = 8 (F: F_3,28=205.543, P < 0.001, eta-sq = 0.957, CI 0.910–0.969; G: F_3,28=25.397, P < 0.001, eta-sq = 0.731, CI 0.488–0.808; H: F_3,28=20.382, P < 0.001, eta-sq = 0.686, CI 0.416–0.775)

Repeated NTG injections decreased FP expression in PVN ^OXT neurons

PGF2α signaling is known to generate several important second messengers, including diacylglycerol (DAG), inositol trisphosphate (IP3), and cyclic adenosine monophosphate (cAMP), which are closely associated with neuronal calcium signaling, synaptic transmission, and other neuronal functions [34–38]. Indeed, PGF2α signaling is involved in the activation of PVN neurons in response to external stimuli [20–22]. Thus, we aimed to investigate the PGF2α signaling activity in PVN ^OXT neurons during the chronification of migraine. Compared to the NS-injected control, both episodic and chronic migraine mice showed a significant upregulation of PGF2α levels in the PVN (Fig. 4A). Moreover, chronic migraine mice exhibited a higher level of PVN PGF2α compared to episodic migraine mice (Fig. 4A), suggesting that PVN PGF2α levels increased with the frequency of migraine episodes. Given that PGF2α-induced cellular alterations are mediated by FP, which is highly expressed in PVN ^OXT neurons, we subsequently detected its expression in PVN. Interestingly, compared with the single NTG-injected or NS-injected mice, repeated NTG-injected mice showed a significant reduction in PVN FP expression (Fig. 4D and E). Next, the FP expression profiles in PVN ^OXT neurons were detected using an immunohistochemistry assay. Accordingly, repeated NTG injections significantly reduced PTGFR mRNA (encoding FP) expression in PVN ^OXT neurons compared to single NTG-injected mice (Fig. 4B and C). It is well-established that G-protein-coupled receptors (GPCR) undergo internalization and degradation in response to persistent high-level ligand binding, a process known as receptor desensitization [39, 40]. Given that FP belongs to GPCR family, we further investigated the dynamic changes in PGF2α levels and FP receptor expression in the PVN during migraine chronification. Thus, we measured the baseline PGF2α and FP levels in PVN 1 h before first to fifth NTG injections on days 1 to 9 (Fig. S1A). A significant downregulation of FP expression in PVN ^OXT neurons was observed on day 5 (Fig. S1C–F), coinciding with the onset of significant upregulation in PGF2α levels (Fig. S1B). The upregulation of PGF2α and downregulation of FP continued to day 7 and persisted at least to day 9 (Fig. S1B–F). Therefore, these findings suggested the impaired PGF2α signaling in PVN ^OXT neurons of chronic migraine mice, and elevated PGF2 levels might be a potential contributor to the downregulation of FP.

Fig. 4.

Compromised PGF2α signaling in PVN ^OXT neurons of repeated NTG-induced chronic migraine mice. A Quantification of PGF2α level in PVN after NTG or vehicle injections. B and C RNAScope and immunofluorescent co-staining showing the expression of PTGFR mRNA in PVN ^OXT neurons, scale bar = 50 μm. D and E Western blot showing the expression of FP in PVN after NTG or vehicle injections. The data are presented as the mean ± SEM. **P < 0.01, vs. Single NS group; ^##P < 0.01, vs. Repeated NS group; ^&&P < 0.01, vs. Single NTG group. Bonferroni post-hoc test after One way ANOVA, n = 8 (A: F_3,28=52.905, P < 0.001, eta-sq = 0.850, 95%CI 0.700-0.893; B: F_3,28=25.591, P < 0.001, eta-sq = 0.733, 95%CI 0.491–0.809; E: F_3,28=19.744, P < 0.001, eta-sq = 0.679, 95%CI 0.405–0.770)

Chemo-activation of PVN ^OXT neurons attenuated migraine hyperalgesia and promoted TNC GABA release

To investigate the role of PVN ^OXT neurons in chronic migraine, we employed chemo-activation and pharmacological strategies to examine their functional contribution in repeated NTG-induced chronic migraine mice. Specifically, OXT-Cre transgenic male mice received bilateral PVN injections of a Cre-dependent chemogenetic activation virus (AAV-EFα-DIO-hM3Dq-mCherry) or a scramble control virus (AAV-EFα-DIO-mCherry), with cannulas bilaterally injected into TNC. Four weeks later, repeated NTG injections were initiated to establish the chronic migraine model. Behavioral tests were performed on day 10 of chronic migraine establishment. 45 min before behavioral tests, the OXTR antagonist L-368,899 or vehicle was microinjected into bilateral TNC through the cannula, followed 15 min later by intraperitoneal injection of CNO to specifically activate the PVN ^OXT neurons, and another 30 min later for behavioral tests (Fig. 5A–C). The OXT-Cre Mice with virus injection showed confined hM3Dq- or mCherry-labeled neurons in the PVN (Figs. 5C and S2A). Moreover, these viral-targeted neurons were highly colocalized with OXT immunoreactive neurons (Fig. S2A and B), indicating a specific expression of the AAV virus in PVN ^OXT neurons. Whole-cell recordings in acute brain slices confirmed that CNO significantly increased the firing rates of hM3Dq-expressing PVN ^OXT neurons (Fig. S2C–E). In OXT-Cre mice, significant c-fos immunoreactivity was observed in the hM3Dq-expressing neurons compared with the mCherry-expressing neurons 3 h after CNO administration (Fig. D and E). These data suggested that our chemogenetic modulation strategy successfully activated PVN ^OXT neurons. Compared with the NTG + mCherry + Vehicle group, mice in the NTG + hM3Dq + Vehicle group showed significant alleviation of periorbital mechanical and thermal allodynia (Fig. 5H–J), consistent with upregulation of OXT and GABA levels in TNC (Fig. 5F and G). In contrast, mice in the NTG + hM3Dq + L-368,899 group, with intra-TNC administration of OXTR antagonist L-368,899, showed significantly decreased periorbital mechanical threshold and GABA levels, along with increased acetone response time compared with the NTG + hM3Dq + Vehicle group. This indicated that OXTR inhibition in the TNC reversed the protective effects of PVN ^OXT neuronal activation on both migraine hyperalgesia and TNC GABA release (Fig. 5H–J). Although activation of PVN ^OXT neurons induced a significant upregulation in the release of OXT and GABA levels in TNC (Fig. 5F and G), there were no significant differences in behavioral changes between the NS + hM3Dq + Vehicle group and the NS + mCherry + Vehicle group (Fig. 5H–J), suggesting the activation of PVN ^OXT neurons did not affect the basal nociceptive responses in normal mice. Overall, these findings suggested that the PVN ^OXT→TNC ^GABA circuit played a counteracting role in chronic migraine hyperalgesia.

Fig. 5.

Chemo-activation of PVN ^OXT neurons alleviated chronic migraine hyperalgesia through TNC GABA enhancement. A and B Experimental schedule and diagram of chemo-activation of PVN ^OXT neurons and TNC local OXTR inhibition. C Representative images showing the expression of hM3Dq- or mCherry-labelled virus in bilateral PVN and trace of cannula in bilateral TNC, scale bar = 200 μm. D and E Immunofluorescent staining showing the expression of c-fos in PVN ^OXT neurons after administration of CNO, scale bar = 100 μm. F and G Effects of chemo-activation of PVN ^OXT neurons and TNC local OXTR inhibition on TNC OXT (F) and GABA (G) release. H–J Effects of chemo-activation of PVN ^OXT neurons and TNC local OXTR inhibition on mechanical (I) and thermal (J) allodynia of chronic migraine mice. The data are presented as the mean ± SEM. *P < 0.01, **P < 0.01 vs. NS + mCherry + Vehicle group, ^#P < 0.01, ^##P < 0.01, vs. NTG + mCherry + Vehicle group, ^&&P < 0.01, vs. NTG + hM3Dq + Vehicle group, Bonferroni post-hoc test after One way ANOVA, n = 8 (E: F_4,35=438.471, P < 0.001, eta-sq = 0.980, 95%CI 0.962–0.985; F: F_4,35=32.420, P < 0.001, eta-sq = 0.787, 95%CI 0.608–0.841; G: F_4,35=47.104, P < 0.001, eta-sq = 0.843, 95%CI 0.706–0.883; I: F_4,35=35.237, P < 0.001, eta-sq = 0.801, 95%CI 0.632–0.851; J: F_4,35=50.247, P < 0.001, eta-sq = 0.852, 95%CI 0. 721-0.889)

Overexpression of FP restored the activity of PVN ^OXT neurons in chronic migraine mice

Having confirmed the descending inhibitory effect of PVN ^OXT neurons in the trigeminal central sensitization of chronic migraine mice via enhanced TNC GABA release, we then examined whether the reduction of FP led to the decreased activity of PVN ^OXT neurons. For this purpose, we intended to selectively enhance FP expression in PVN ^OXT neurons using a gene overexpression strategy. Thus, a plasmid containing the coding sequence of the PTGFR gene (encoding FP) was embedded in a Cre-dependent AAV vector: AAV-CMV-DIO-PTGFR, and the AAV-CMV-DIO-mCherry was used as the non-encoding scramble control. To validate the efficacy of this overexpression system, the overexpression or scramble control AAV was mixed with a Cre recombinase encoding AAV (AAV-hSyn-Cre-EGFP) and then delivered into cultured primary hypothalamic neurons. After culturing for 7 days, the neurons were successfully transfected with the AAV virus, characterized by more than 90% of neurons expressing the EGFP fluorescent protein (Fig. 6A). As expected, the PTGFR mRNA and protein levels in the Overexpression group were significantly increased compared with the Control group (Fig. 6B–D). Next, AAV-CMV-DIO-PTGFR (AAV-CMV-DIO-mCherry as control) was directly delivered into the bilateral PVN of OXT-Cre mice, followed 4 weeks later by repeated NTG injections and subsequent behavioral tests (Fig. 6E). RNAscope results also confirmed efficient PTGFR overexpression in PVN ^OXT neurons (Fig. 6F and G). According to c-fos immunofluorescence staining, the NTG + OE group showed a significant enhancement in the activation of PVN OXT neurons compared with the NTG + NC group (Fig. 6H and I), indicating that FP overexpression restored the activity of PVN ^OXT neurons in chronic migraine mice. As expected, FP overexpression in PVN ^OXT neurons attenuated the periorbital hyperalgesia of chronic migraine mice (Fig. 6J and K). Moreover, FP overexpression promoted the release of OXT and GABA in the TNC (Fig. 6L and M). These findings collectively suggest that enhancing PGF2 signaling in PVN ^OXT neurons, specifically by overexpressing FP, attenuates chronic migraine hyperalgesia by restoring the activity of these neurons.

Fig. 6.

Overexpression of FP restored the activity of PVN ^OXT neurons and attenuated chronic migraine hyperalgesia. A Representative images showing the expression of overexpression virus in cultured hypothalamic neurons, scale bar = 100 μm. B–D Expression of PTGFR mRNA (B) and FP protein (C, D) in cultured hypothalamic neurons transfected with overexpression or control virus. E Experimental schedule and diagram of specific overexpression of FP in PVN ^OXT neurons. F and G RNAscope and immunofluorescent co-staining showing the expression of PTGFR mRNA in PVN ^OXT neurons of NS- or NTG-injected mice transfected with overexpression or control virus, scale bar = 50 μm. H and I Double immunofluorescent staining showing the expression of c-fos in PVN ^OXT neurons of NS- or NTG-injected mice transfected with overexpression or control virus, scale bar = 100 μm. J and K Von Frey (J) and acetone (K) tests showing the effects of FP overexpression in PVN ^OXT neurons on mechanical (J) and thermal (K) allodynia. L and M ELISA analysis showing the effects of FP overexpression in PVN ^OXT neurons on TNC OXT (L) and GABA (M) release. The data are presented as the mean ± SEM. **P < 0.01, vs. Control group, Unpaired t-test, n = 8 (B: t = 12.232, df = 14, P < 0.01, cohen’d = 6.116, 95%CI 3.654–8.543; D: t = 11.132, df = 14, P < 0.01, cohen’d = 5.566, 95%CI 3.285–7.809). *P < 0.05, **P < 0.01, vs. NS + NC group, ^#P < 0.05, ^##P < 0.01, vs. NTG + NC group, Bonferroni post-hoc test after One way ANOVA, n = 8 (G: F_2,21=47.618, P < 0.001, eta-sq = 0.819, 95%CI 0.610–0.879; I: F_2,21=119.344, P < 0.001; eta-sq = 0.919, 95%CI 0.817–0.946; L: F_2,21=48.671, P < 0.001; M: F_2,21=25.741, P < 0.001); Bonferroni post-hoc test after Repeated measure ANOVA, n = 8 (J: Overexpression, F_2,21=162.079, P < 0.001, ges = 0.939; time, F_2,21=1.363, P = 0.256, ges = 0.061; interaction, F_2,21=0.176, P = 0.840, ges = 0.016; 95%CI 0.536–0.613; K: Overexpression, F_2,21=153.849, P < 0.001, ges = 0.936; time, F_2,21=4.134, P = 0.010, ges = 0.164; interaction, F_2,21=1.562, P = 0.173, ges = 0.130; 95%CI 5.031–4.598)

Discussion

This study revealed a neural circuit involving OXT projections from the PVN to GABAergic neurons in the TNC, which is critically involved in regulating trigeminal nociceptive transmission. Decreased activity of PVN ^OXT neurons was observed in repeated NTG-induced chronic migraine mice, while chemo-activation of these neurons significantly alleviated periorbital hyperalgesia by OXTR-dependent enhancement of TNC GABA release, suggesting that the impaired PVN ^OXT neuronal activity was responsible for the trigeminal central sensitization and migraine hyperalgesia. Consistent with the essential role of PGF2α signaling for neuronal activation, our study revealed that chronic migraine mice exhibited reduced expression of FP in PVN ^OXT neurons. Overexpression of FP in PVN ^OXT neurons restored their neuronal activity, which in turn promoted the release of OXT and GABA in TNC, thereby resulting in the amelioration of migraine-like hyperalgesia. These findings collectively establish the pivotal role of PVN ^OXT neurons in regulating the central sensitization of trigeminal pain in chronic migraine and its underlying mechanisms.

Tools that accurately mimic the characteristics of sensory abnormalities experienced by migraine sufferers are essential for elucidating pathophysiological mechanisms and discovering novel therapies for chronic migraine. It is well-established that NO, by triggering vasodilation and promoting CGRP release, is a potent initiator of migraine attacks [41–43]. Indeed, repeated administration of NTG, a widely used NO donor, induces the pathophysiological alterations in TNC such as microglial and astrocytic overactivation, CGRP accumulation, and neuronal excitatory-inhibitory imbalance, which contribute to the central sensitization of trigeminal nociceptive transmissions [44–46]. Therefore, repeated NTG injection has been employed to model human chronic migraine in mice [25]. Our behavioral findings revealed that a single NTG injection induced a reversible migraine-like periorbital hyperalgesia, whereas five repeated injections resulted in persistent periorbital hyperalgesia. Since persistent hyperalgesia is a hallmark of central sensitization and its occurrence indicates that migraine transitions from an episodic to a chronic state [26], the repeated administration of NTG successfully established a chronic migraine mouse model. In a rodent model of migraine with aura, dural inhibition of persistent sodium current was found to activate PVN ^OXT neurons, closely associated with the duration of mechanical allodynia in the trigeminal nerve innervating area [15]. Therefore, we hypothesized that NTG injection might elicit a similar effect. Indeed, our c-fos immunostaining results demonstrated that NTG injection activated PVN ^OXT neurons, indicating that these neurons also play a role in NTG-evoked migraine pathophysiology. A recent study revealed that PVN electrical stimulation could block the firing of TNC second-order trigeminal nociceptive transduction neurons, and this effect was eliminated by an OXTR antagonist [10]. Moreover, OXTR agonism has been shown to exert analgesic effects by suppressing central sensitization at the spinal level in various chronic pain conditions [47–49], including chronic migraine [50]. Given that the PVN is one of the primary sites for OXT secretion, these findings suggest that PVN ^OXT neurons may participate in regulating trigeminal central sensitization through the release of OXT from their terminals. However, due to the lack of direct neuroanatomical evidence, it remains unclear whether and how PVN ^OXT neurons regulate trigeminal central sensitization in migraine. To address this, we employed a rigorous anterograde transmonosynaptic tracing strategy [51], which confirmed the direct neural connection between PVN ^OXT neurons and TNC neurons. Besides, the downstream TNC neurons innervated by PVN ^OXT neurons were classified as GABAergic by our RNAscope in situ hybridization assay. Given that the imbalance between glutamatergic and GABAergic activities in the TNC is a key mechanism contributing to trigeminal central sensitization [44], our identification of a neural circuit from PVN ^OXT neurons to TNC ^GABA neurons provides a neuroanatomical substrate linking PVN ^OXT neurons and trigeminal central sensitization.

Having established the direct descending projection from PVN ^OXT neurons to TNC ^GABA neurons, we subsequently focused on the functional role of the PVN ^OXT neurons in chronic migraine. Our c-fos immunostaining showed reduced neuronal activity in PVN ^OXT neurons in chronic migraine mice compared to those in acute migraine mice. These differences suggest that decreased activity of PVN ^OXT neurons may contribute to central sensitization and persistent hyperalgesia in chronic migraine. As expected, chemo-activation of PVN ^OXT neurons in chronic migraine mice not only attenuated repeated NTG-induced preorbital hyperalgesia but also accelerated OXT and GABA release in TNC. Specifically, pharmacologic OXTR blockade (using L-368,899) in the TNC reversed PVN ^OXT neuron activation-induced beneficial effects in chronic migraine mice. Overall, our results suggest a descending pain inhibitory pathway from PVN ^OXT neurons to TNC ^GABA neurons, which modulates trigeminal central sensitization in chronic migraine.

In this study, we provide preliminary evidence suggesting that PGF2α signaling plays an important role in modulating the activation of PVN ^OXT neurons, which are involved in regulating central sensitization in chronic migraine. Unlike other vasodilative PGs, the administration of PGF2α fails to trigger migraine-like attacks in humans or animals [17, 28]. Given the elevated levels of PGF2α found in the urine or saliva of migraine sufferers [18, 19], PGF2α may play a role in migraine pathophysiology. Our observations revealed that NTG injections led to a significant increase in PGF2α in the PVN. This elevation may be attributed to the entry of peripheral PGF2α into the PVN via the blood-brain barrier or the dysregulated local secretion of PGF2α within the PVN [52, 53]. Moreover, mice with repeated NTG injections (chronic migraine) exhibited higher levels of PGF2α than mice with a single NTG injection (episodic migraine), indicating that PVN PGF2α levels increase with the number of migraine attacks. Genomic profiling of the PVN revealed that PVN ^OXT neurons exhibit high FP expression [24]. PGF2α signaling is closely related to neuronal activity, primarily through the rapid generation of second messengers such as IP3, DAG, and cAMP, which play key roles in mobilizing intracellular calcium stores and activating downstream signaling pathways, thereby influencing neuronal function and plasticity [20–23]. Therefore, heightened PGF2α signaling may be a potential cause of PVN ^OXT neuron activation during NTG-induced migraine pathological conditions. In contrast to PGF2α, FP expression in PVN ^OXT neurons significantly decreased in mice with repeated NTG-induced chronic migraine. Unlike previous studies that examined all neurons in a specific brain region, AAVs with transgenic Cre mice were used to overexpress FP in the PVN ^OXT neurons. As expected, specific FP overexpression in the PVN ^OXT neurons improved preorbital hyperalgesia and promoted OXT and GABA release in the TNC. Persistent activation of GPCR upon agonist binding is known to lead to both desensitization and internalization, processes characterized by the loss of receptor expression and function [39, 40]. As one of the GPCRs, the expression level of FP in the PVN decreased along with the increase in PGF2α levels during the chronification of migraine. So, the decreased expression of FP in PVN ^OXT neurons might be due to the high level of PGF2α in the PVN. Given that PGF2α signaling can regulate the activity of PVN ^OXT neurons, it is highly conceivable that the compromised PGF2α signaling pathway disrupted the equilibrium between the PVN ^OXT neurons and TNC ^GABA neurons, thereby impairing their ability to suppress trigeminal central sensitization in chronic migraine mice.

In conclusion, we identified a neural circuit: PVN ^OXT →TNC ^GABA, which plays a significant role in regulating trigeminal central sensitization. Our findings suggest that decreased activity of PVN ^OXT neurons results in a reduction in GABA secretion of the TNC ^GABA neurons, thereby contributing to central sensitization and the chronification of migraine. Finally, we investigated the regulatory influence of PGF2α signaling on OXT neurons within the PVN. Notably, targeted FP overexpression in PVN ^OXT neurons counteracted their functional suppression, restoring neuronal activity and mitigating migraine hyperalgesia. In brief, this study presents novel findings on the neural mechanisms underlying top-down regulation in chronic migraine.

Limitations

Only male mice were tested, which is a limitation of this study. The reason we chose male mice was that the pain perception of females is more sensitive and variable than that of males, primarily due to the influence of the estrous cycle and fluctuating sex hormone levels [54, 55]. As OXT has been shown to be a critical mediator in the descending inhibitory effect of the PVN, sex differences in OXT secretion may lead to different effects on migraine hyperalgesia. Although the current evidence regarding the role of PVN ^OXT neurons in migraine has been collected from both male and female rodents, it has not yet clarified the sex differences in their function in regulating migraine hyperalgesia [10, 14, 15]. A previous study has shown that the analgesic effect of spinal OXT administration varied between sexes in the context of neuropathic pain [56]. Therefore, further investigation is needed to determine whether a sex-specific effect exists in the descending inhibitory modulation of PVN ^OXT neurons in chronic migraine. Beyond the pain processing, PVN ^OXT neurons also project to many brain regions to modulate various physiological functions, including social interaction [57], emotion [58] and feeding [59]. More than half of migraine patients have comorbid psychiatric disorders, particularly anxiety and depression [60]. Therefore, it is meaningful to explore whether PVN ^OXT neurons could modulate the psychiatric comorbidities of chronic migraine in our future work.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Author contributions

This study was designed by JB, YHZ, and LY. The experiments were completed by XRH, LW, LQW, CXP, LN and YYT. QZ and HWZ performed statistical analysis. XRH and LQW wrote the manuscript. JB, YHZ, and LY provided supervision and final check. All the authors read the final version of this paper and approved it.

Funding

This work was supported by the Health Commission of Sichuan Province Medical Science and Technology Program (No: 24QNMP019 to Jiang Bian), Innovative Program for the Integration of Meteorological Medicine and Engineering and Application Transformation in Panzhihua (No: 2023ZD-C-1 to Li Yin).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

All animal experiments performed in this study were approved by the Ethics Committee for Animal Experimentation of Qingdao University.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.