Male adult Sprague–Dawley rats weighing 300–350 g were used. 18 rats were used in the experiment. Under standard laboratory conditions on a 12-h light/dark cycle, rats were housed and maintained. Tap water and rat chow were used ad libitum. A week before the experiment, rats were habituated to the experimental environment. The experiments were conducted with the approval of the Committee on Animal Use for Research and Education of the Renmin Hospital of Wuhan University (Wuhan, China). Our evaluation ethics were guided by the guidelines of the International Association for the Study of Pain in Conscious Animals (25). Efforts were made to reduce the number of animals and the degree of suffering.

Rats were anesthetized with isoflurane (3% concentration, 300 ml/min flow). The skin and soft tissues of the rats were sequentially incised to expose the skull. The skull near the sagittal sinus was drilled with a dental drill to expose the dura without penetrating it (26). Then we implanted the guide cannula into the previously drilled cranial window. To minimize trauma to adjacent tissue in an attempt to reduce the chance of infection during surgery. Antibiotics and analgesics were administered post-operatively. After the operation, the state of each rat was observed. And each rat was placed in a single cage to prevent the guide cannula from falling off due to mutual bite. All rats rested for 1 week after surgery.

Rats were randomly divided into three groups: control group (CON), inflammatory soup model group (IS), and 2-DG pretreatment + IS group (2-DG). The experimental model of migraine was constructed by repeated IS (20 μL, including 2 mM histamine, 2 mM serotonin, 2 mM bradykinin, and 0.2 mM prostaglandin E2 in PBS) dural infusion at the same time each day for at least 7 days. In the control group, rats were injected with the same dose of PBS. In the 2-DG group, 2-DG (500 mg/kg; Meryer) was intraperitoneally injected 30 min before IS infusion.

The rat was placed in a plastic tube restraint device, and after the rat was quiet (about 30 min), the periorbital pain threshold of the rat was measured with a von Frey wire and the results were recorded. Baseline pain thresholds were measured before the first IS infusion and after dural infusion. The standard range of Von Frey filaments is 0.008 to 100 g. Von Frey filaments were placed on the periorbital skin of rats. Each test lasted 2 s, with five consecutive tests. Each with a 5 s interval, and a waiting period of at least 30 s between different stimulation intensities (27). The pain threshold is the minimum force that causes the rat to rapidly withdraw its head at least three times out of five stimuli.

The head scratching times of the rats were recorded within 1 h after the IS was injected via the tube. Scratching the periorbital and top of the head with the front paws was recorded as the effective number of times.

Two hours after the last administration, the rats were sacrificed after deep anesthesia with sodium pentobarbital. For immunoflurescence, we first exposed rat hearts, flushed systemic blood with PBS through the left ventricle, and then perfused them with paraformaldehyde until they became rigid. Finally, the brain tissue and trigeminal ganglion were removed and fixed in 4% paraformaldehyde solution overnight. The brain tissue and trigeminal ganglion were embedded in paraffin and sliced. For western bloting (WB), the head was placed on ice to remove the brain tissue, and the medullary dorsal horn tissue was separated. The tissue was frozen in liquid nitrogen and stored in −80°C refrigerator.

BV-2 murine microglia were cultured as described in previous studies (n = 3/group) (28). BV-2 cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) containing 1% 100 U/ml penicillin/streptomycin and 10% fetal bovine serum at 37°C. LPS was diluted in the medium so that the final concentration was 1,000 ng/ml. In six-well plates, 2 ml medium was added to each well of LPS administration group. After 12 h of LPS activation of BV-2 cells with or without 2-DG (1 mM, 2 mM, 4 mM), the medium was harvested and checked for glucose by a glucometer and pH by a pH indicator. CGRP, IL-1β, IL-6, Iba-1, pro-BDNF, BDNF were detected by Western blotting.

The total RNA was extracted from tissues using TRIzol reagent according to manufacturer’s instructions. Following the manufacturer’s instructions, 1 μg of total RNA was used to synthesize cDNA using the SweScript All-in-One First-Strand cDNA Synthesis SuperMix Kit (servicebio). The Lightcycler 4800II instrument was used for quantitative RT-PCR of HK2, PKM2, PFKP, LDH and HIF-1α genes. TaqMan primers were purchased from Sangon Biotech. Gene expression was measured by comparative CP. Standardize the expression level in the test sample as an endogenous reference β-actin level. The primer sequence of HK2 was: forward, 5′-TGATCGCCTGCTTATTCACGG-3′ and reverse, 5′-AACCGCCTAGAAATCTCCAGA-3′. The primer sequence of PKM2 was: forward, 5′-GCCGCCTGGACATTGACTC-3′ and reverse, 5′-CCATGAGAGAAATTCAGCCGAG-3′. The primer sequence of PFKP was: forward, 5′-CGTCCAGCACCTCCTTT-3′ and reverse, 5′-ACGGACAGCAGCATTCA-3′. The primer sequence of LDH was: forward, 5′-AGGAGAAACACGCCTTGATTTAG-3′ and reverse, 5′-ACGAGCAGAGTCCAGATTACAA-3′. The primer sequence of HIF-1α was: forward, 5′-ACCTTCATCGGAAACTCCAAAG-3′ and reverse, 5′-ACTGTTAGGCTCAGGTGAACT-3′. The primer sequence of β-actin was: forward, 5′-AGGGAAATCGTGCGTGAC-3′ and reverse, 5′-CGCATTGCCGATAGTG-3′. Data are expressed as mean ± SD.

The medulla dorsal horn was cut into small pieces, and the tissues were washed three times by adding an appropriate amount of cold PBS after weighing. Add 100 μL of lysis buffer and corresponding protease inhibitors per 50 mg of tissue. Protein concentration was measured by using the BCA protein assay kit (servicebio). After heating the mixture with loading buffer at 100°C for 5 min, the collected proteins were loaded and separated on a 10% SDS PAGE gel (Epizyme) and electrically transferred to PVDF membranes. Membranes were incubated with the primary antibody overnight at 4°C following 5% nonfat milk blocking for 2 h at room temperature. The primary antibody was diluted as follows: PFKP (1:1,000; servicebio), HK1 (1:1,000; servicebio), G6PD (1:1,000; servicebio), COXIV (1:1,000; proteintech), CGRP (1:1,000; Santa Cruz), c-Fos (1:1,000; Santa Cruz), Iba-1 (1:1,000; woko), IL-6 (1:1,000; proteintech), IL-1β (1:1,000; proteintech), β-actin (1:2,000; servicebio), pro-BDNF (1:2,000; abcam), BDNF (1:2,000; abcam). β-actin served as respective controls. Incubation with the secondary antibodies followed for 2 h at room temperature. Enhanced chemiluminescence (ECL; biosharp) luminescent solution was used to visualize bands. Images were acquired using a chemiluminescence imaging system (BIO-RAD ChemiDoc Touch). The acquired images were analyzed by using Image J software (Image J 1.48v).

Trigeminal ganglion slices with a thickness of 4 μm were then prepared. Sections were dewaxed in xylene. Then the alcohol grade was lowered for rehydration. Sections were incubated with 0.3% Triton X-100 for 15 min, blocked with 5% BSA for 60 min at room temperature, and then incubated with primary antibody overnight at 4°C. PBS was washed three times after each operation. The primary antibody was diluted as follows: CGRP (1:100; Santa Cruz), Iba-1 (1:500; woko), PFKP (1:200; servicebio), GFAP (1:200; GeneTex), NeuN (1:500; proteintech). After washing with PBS, sections were incubated with mouse secondary antibody (488 nm, green; servicebio). Anti-mouse (488 nm, green; servicebio), anti-rabbit (CY3, red; servicebio) and anti-goat (FITC, green; servicebio) were used. DAPI (servicebio) was incubated for 10 min for staining nuclei. Capture images under the objective of an upright microscope.

GraphPad Prism 8 was used for statistical analysis and graph generation. The variance homogeneity test and the normality test were performed. Unpaired t test was used to assess the difference between the two groups and a one-way analysis of variance (ANOVA) was used to assess the difference among multiple groups. Graphical representation of data is made with bar and line charts. All results are expressed as mean  ±  SD. Differences were considered statistically significant at p < 0.05.

To explore whether glycolysis increased during the experimental models of migraine, we detected the gene expression of important enzymes related to glycolysis, such as PFKP, HK2, HIF-1α, LDH and PKM2. Compared with CON group, quantitative real-time polymerase chain reaction showed that PFKP (2.02 ± 0.51), HK2 (2.55 ± 0.72), HIF-1α (3.05 ± 0.61) and PKM2 (1.97 ± 0.49) mRNA expression increased in migraine model group (Figures 1A–E). In addition, we examined the key enzymes of the glycolysis and pentose phosphate pathways, and COXIV protein levels in medullary dorsal horn (Figure 1F). The expression of COXIV (0.59 ± 0.13) in the experimental model of migraine was significantly reduced (Figure 1G). In the medullary dorsal horn of the experimental migraine model, the key glycolysis enzyme PFKP (2.09 ± 0.32) was significantly higher than that in the control group (1.10 ± 0.38; Figure 1H). However, another key glycolysis enzyme, HK1 (1.11 ± 0.14), did not change significantly (Figure 1I). Moreover, G6PD (1.01 ± 0.08), a key enzyme in the pentose phosphate pathway, was almost unchanged (Figure 1J). These results indicated that oxidative phosphorylation is impaired but glycolysis is enhanced in migraine animal models.

Daily intraperitoneal injection of 2-DG 0.5 h before IS dural infusion for 7 days significantly inhibited the development of facial friction (Figure 2A) and tactile allodynia (Figure 2B). In rodents, stimulatory factors in the trigeminal region produce characteristic directional grooming behaviors. We used this migraine-like feature as a tool for assessing distress. To investigate the effect of glycolysis inhibitor 2-DG on the rat model of migraine, 2-DG was administered intraperitoneally. The mechanical thresholds of the periorbital region significantly decreased in the IS group (1.23 ± 0.41) compared to the CON group (6.00 ± 1.27). The administration of 2-DG (4.67 ± 1.03) increased the mechanical thresholds of migraine rats. Compared with those of the CON group (9.00 ± 5.33), the number of head scratchings in the IS group (119.83 ± 11.37) significantly increased. The administration of 2-DG (17.00 ± 3.52) decreased the head scratching times of migraine rats. These results indicated that glycolysis may participates in migraine-associated allodynia.

We further explored the effects of 2-DG treatment on astrocyte proliferation and neuron loss (Figure 3A). We found that the astrocytes proliferated in the medullary dorsal horn after IS stimulation in rats (1.08 ± 0.21). The treatment of 2-DG (0.71 ± 0.06) can inhibit the proliferation of astrocytes (Figure 3B). However, no significant neuronal loss was observed in the medullary dorsal horn after IS stimulation (3.07 ± 0.16). No significant difference was observed between the rats treated with 2-DG (3.42 ± 0.57) and the control group (3.33 ± 0.28; Figure 3C).

To investigate the contribution of glycolysis to the experimental models of migraine, we measured the expression of c-Fos in the medullary dorsal horn by WB (Figure 4A). Compared with the control group, the IS group (1.41 ± 0.14) showed higher c-Fos protein levels. However, intraperitoneal administration of 2-DG (1.10 ± 0.03) reduced the level of c-Fos protein expression (Figure 4B).

To investigate the effect of inhibition of glycolysis on the alleviation of migraine, we measured the expression of CGRP in the medulla dorsal horn by western blotting (Figure 4C). The protein levels of CGRP were elevated in the IS group (1.34 ± 0.15) compared with the CON group (1.00 ± 0.15), whereas 2-DG (0.51 ± 0.04) treatment decreased CGRP levels (Figure 4D). Immunofluorescence staining were also conducted to detect the expression of CGRP in trigeminal ganglion and medulla dorsal horn. The fluorescence intensity of CGRP in the trigeminal ganglia of the IS group (2.57 ± 0.27) was increased compared with that of the CON group (1.00 ± 0.21). However, 2-DG (1.21 ± 0.24) treatment attenuated the increase in CGRP immunofluorescence intensity induced by inflammatory soup infusion (Figures 4E,G). Similarly, the fluorescence intensity of CGRP in the medulla dorsal horn Figures 4F, H in the IS group (2.83 ± 0.42) was significantly higher than that in the CON group (1.09 ± 0.20), and this increase was suppressed in the 2-DG group (0.82 ± 0.23).

Iba-1 is a measure of microglia reactivity after central nervous system injury. And when microglia are activated, Iba-1 protein expression increases (29). To investigate the effect of inhibiting glycolysis on migraine microglia activation, we used WB to detect the expression of Iba-1 in the medulla dorsal horn (Figure 5A). WB analysis showed that compared with the control group, the level of Iba-1 in the IS group (1.75 ± 0.44) was significantly increased. Furthermore, intraperitoneal injection of 2-DG (1.07 ± 0.24) significantly decreased Iba-1 levels (Figure 5B). This suggests that the inhibition of glycolysis may inhibit the activation of microglia. Likewise, the expression of Iba-1 in the medulla dorsal horn was detected using immunofluorescence. Compared with the CON group, immunofluorescence showed that the expression of Iba-1 in the IS group (1.48 ± 0.16) was significantly increased and could be inhibited by 2-DG (0.71 ± 0.26). In addition, the double-labeled immunofluorescence studies further revealed we found that Iba-1 co-stained with PFKP in the microglia cytoplasm, and the co-expression of both was increased in activated microglia.

The activation of microglia and subsequent release of various pro-inflammatory cytokines such as IL-6 and IL-1β lead to hyperalgesia or allodynia in experimental migraine models (30, 31). To investigate the effect of inhibiting glycolysis on the release of pro-inflammatory factors in migraine, we used WB to examine the expression of IL-1β and IL-6 in the medulla dorsal horn (Figures 6A,C). WB analysis showed that compared with the control group, the levels of IL-1β (1.20 ± 0.09) and IL-6 (1.38 ± 0.16) in the IS group were significantly increased. In addition, intraperitoneal injection of 2-DG significantly decreased IL-1β (0.55 ± 0.15) and IL-6 (0.97 ± 0.20) levels (Figures 6B,D). This suggests that inhibition of glycolysis may inhibit the release of pro-inflammatory factors following inflammatory cell activation.

To explore whether there is enhanced glycolysis in activated microglia, we investigated the metabolism of LPS-stimulated BV2 cells. Following microglia activation, the glucose concentration (8.20 ± 0.46) in the medium, which reflects glucose consumption, dropped dramatically (Figure 7A), and the pH (7.88 ± 0.04), which reflects lactate levels, dropped significantly (Figure 7B). Both glucose consumption and changes in pH indicated enhanced glycolysis in activated microglia. Glucose consumption and acidity of the medium were significantly reduced after 2-DG treatment (Figures 7A,B).

To explore the effect of 2-DG on pro-inflammatory cytokine and migraine-related protein expression, IL-6, IL-1β and CGRP production were analyzed by western blotting. Our results showed that the protein levels of IL-6 (1.24 ± 0.15), IL-1β (1.26 ± 0.05) and CGRP (1.24 ± 0.10) were upregulated in activated microglia, whereas 2-DG treatment inhibited it (Figures 7C–F). Taken together, our results suggest that 2-DG has a dose-dependent anti-inflammatory effect and reduces the expression of CGRP (0.82 ± 0.16; 0.63 ± 0.06; 0.45 ± 0.17) in LPS-induced BV2 cells. In order to investigate whether changes in proBDNF/BDNF levels are associated with activation of microglia, we stimulated microglia with endotoxin, a well-recognized activator of microglia. The levels of proBDNF (1.17 ± 0.13) and BDNF (1.33 ± 0.07) secreted by activated microglia were significantly increased, compared with resting microglia. Iba-1 levels (1.26 ± 0.03) were also shown to be higher than the CON group (1.00 ± 0.08). However, the treatment of glycolysis inhibitor 2-DG (0.65 ± 0.23) significantly reduced the expression of Iba-1. This suggests that inhibition of glycolysis inhibits activation of microglia. With the increase of 2-DG concentration, the expression of Iba-1(0.92 ± 0.15; 0.67 ± 0.24; 0.38 ± 0.15) decreased in a dose-dependent manner. In addition, after 2-DG treatment, the expressions of proBDNF (0.85 ± 0.08; 0.80 ± 0.07; 0.48 ± 0.09) and BDNF (0.67 ± 0.07; 0.37 ± 0.11; 0.34 ± 0.10) decreased significantly, also in a dose-dependent manner (Figures 7G–J). This finding proves that inhibition of glycolysis may have a role in migraine through the BDNF pathway.