GPER/β-Alanine Positive Interaction in The Dorsal Root Ganglion Uncovers Potential Mechanisms: Mediating Continuous Neuronal Sensitisation and Neuroinammation Responses in Neuropathic Pain

Background: The pathogenesis of neuropathic pain and the reasons for the prolonged unhealing are still unknown. Increasing evidence suggests that estrogen sex differences play a role in pain sensitivity, but few studies focused on the role of estrogen receptor which maybe an important molecular component contributing to peripheral pain transduction. We aimed to investigate the impact of oestrogen receptors in nociceptive neuronal response in the dorsal root ganglion (DRG) and spinal dorsal horn using a spared nerve injury (SNI) rat model of chronic pain. Methods: We used a class of estrogen receptors antagonists and agonists intrathecal (i.t.) administrated to male rats with SNI or normal rats to identify the main receptor. Moreover, we applied genes identied through genomic metabolic analysis to determine the key metabolism point and elucidate potential mechanisms mediating continuous neuronal sensitisation and neuroinammation responses in neuropathic pain. The excitability of DRG neurons was detected using the patch clamp technique. Immunohistochemistry, Western blotting, qPCR and behavioral testing were used to assess the expressions, cellular distributions, and actions of main receptor and its related signaling molecules. Results: Increasing the expression and function of G protein-coupled estrogen receptor (GPER), but not estrogen receptor-α (ERα) and estrogen receptor-β (ERβ), in the DRG, but not the dorsal spinal cord, contributed to SNI-induced neuronal sensitisation. Inhibiting GPER expression in the DRG alleviated SNI-induced pain behaviours and neuroinammation by downregulating IL-1β and IL-6 expression as well as restoring GABAα2 expression simultaneously. Additionally, the positive interaction between GPER and β-alanine, β-alanine accumulation enhances pain sensation and promotes chronic pain development. Conclusion: GPER activation in the DRG causes a positive interaction of β-alanine with IL-1β and IL-6 expression and represses GABAα2 involved in post-SNI neuropathic pain development. Blocking GPER and eliminating β-alanine in the DRG may prevent neuropathic pain development.


Introduction
There is increasing evidence of the critical roles of estrogen in the peripheral and central nervous systems [1][2][3]. Estrogen targets neurons to regulate neural development, plasticity, and neuroprotective actions in behavioural and cognitive functions [4]. Estrogen exerts its effects through its receptors, which are widely distributed in the nervous system [5,6]. However, there have been few studies on the role of estrogen receptors (ERs), which may be crucially involved in peripheral pain transduction. It remains unclear whether modulating ERs could be an effective therapeutic strategy for neuropathic pain. Anatomical studies have con rmed that pain perception results from nociceptive signals produced by nociceptive neuron activation in the peripheral sensory nerves [7]. Subsequently, the nociceptive signals are transduced via the dorsal root ganglion (DRG) neurons, which synapse to spinal dorsal horn (SDH) neurons and nally project to the thalamus and cerebral cortex [8]. Neuropathic pain is a clinically common chronic refractory pain syndrome with a global prevalence of approximately 7-10% [9,10].
Neuropathic pain has a complex pathogenesis with an unclear molecular mechanism, which impedes its clinical treatment. Recent studies have demonstrated increased intracellular Cl − levels in the DRG neurons after sciatic nerve section or in ammation [11][12][13]. There has been extensive research on chloride channels in primary sensory neurons since chloride channel activation in sensory neurons may cause chloride e ux and depolarisation due to high intracellular chloride concentrations [14][15][16]. Anion channels, especially chloride channels, may be involved in neuropathic pain pathogenesis [17,18].
Numerous studies have investigated GABAα2-mediated analgesia in neuropathic pain, whose absence is a classic neuropathic pain sign, accompanied by neuroin ammation [19,20]. Spared nerve injury (SNI) is the most recognised method for simulating neuropathic pain [21,22]. Numerous studies on the pathogenesis, prevention, and treatment of neuropathic pain have suggested that the pain threshold is sex-speci c [23,24]. Notably, ERs are present in the DRG and SDH. Few studies have investigated pain regulation by estrogen or ERs through ion channels. GPER is a potential 17β-E2 target that is modulated by the ERs in breast cancer cells [25]. Further, oestradiol was found to modulate the e cacy of synaptic inhibition by decreasing the dwell time of GABA A receptors at inhibitory synapses [26]; however, the receptors involved remain unclear. The possible function of ERs in pain modulation is further supported by the ubiquitous expression of ERs in nociceptive system neurons [27]. There have been no studies focusing on the estrogen effects on pain modulation or the role of ERs in mediating continuous neuronal sensitisation and neuroin ammation responses in neuropathic pain. Additionally, estrogen can exert its effects in the absence of ERs. This study aimed to identify the main ERs involved in neuropathic pain and determine its related genomics and metabolomics to explore the role of ER-related metabolites in neuropathic pain. Speci cally, we aimed to examine the metabolic mechanism of ERs in pain and elucidate the potential relationship of ERs with neuronal sensitivity and neuroin ammation.

Animals
We purchased adult male Sprague-Dawley rats (8-10 weeks old, 180-200 g, n = 400) from the Animal Centre of the Charles River (Charles River Laboratories, US). The use of animals was approved by The Institutional Animal Care and Use Committee at Tongji Medical College, Huazhong University of Science and Technology, China. The animals were housed in plastic boxes with controlled temperature (24 ± 2°C), humidity (40-50%), and a 12:12 h light:dark cycle. We selected rats with relatively uniform and stable baseline responses to cold, mechanical, and hot stimuli. All protocols were approved by the Animal Ethics Committee of the Tongji Medical College, Huazhong University of Science and Technology (approval No. 2405) on December 30, 2020. Further, they were consistent with the Guidelines for the Care and Use of Laboratory Animals, published by the US National Institutes of Health.

Surgical Procedure for Establishing a Neuropathic Pain Model through Spared Nerve Injury (SNI)
We established a model of neuropathic pain using SNI as previously reported [28]. Experimental procedures were performed on animals under anaesthesia through intraperitoneal administration of 40 mg/kg sodium pentobarbital (Sigma-Aldrich, St. Louis, MO, USA). Care was exercised to prevent infection and reduce the in ammation impact. The skin was cut to expose the sciatic nerve and its three terminal branches directly through the part formed by the biceps muscle: the lateral side, common bular nerve, and tibial nerves. The tibial and common peroneal nerves were cut and ligated through SNI, while the sural nerve was preserved. Given that the common peroneal and tibial nerves are closely connected, we subsequently removed 3-5 mm of the distal nerve ends. Care was taken to avoid damaging the nearby sural nerve. All wounds were postoperatively irrigated with sterile saline and closed in layers. Experimenters blinded to the treatments performed the behavioural tests, electrophysiologic recordings, and Western blot and immunohistochemical experiments. The primary outcome was the effect of pharmacologic and genetic interventions involving different ERs in the DRG and SDH on pain-related aversion. We did not use analgesics.

Measurement of β-alanine
BioVision's Alanine Assay Kit allows simple and sensitive detection of alanine. In the kit, alanine is converted to pyruvate, which is speci cally detected based on proportional colour (λ = 570 nm: 0-10 nmol) or uorescence (Ex/Em 535/587 nm: 01 nmol) development. The serum alanine levels are ~2476 µg/ml (~3-9 nmol/10 µl) [30]. Brie y, tissues or cells (1 × 10 6 ) were homogenised in a 100 µl Assay Buffer centrifuge to remove insoluble material at 13,000 g for 10 min. Subsequently, we directly diluted 10-50 µl of deproteinised serum samples in the Assay Buffer and added them to a 96-well plate at 50 µl/well. Next, su cient reagents were prepared according to the number of assays to be performed. We added 50 µl reaction mix [containing assay buffer (44 µl), alanine converting enzyme (2 µl), alanine development mix (2 µl), and alanine probe (2 µl)] to each well containing alanine standard, test, and background control samples. They were mixed well and incubated for 60 min at 37°C with protection from light. Subsequently, the optical density was measured at 570 nm in a microplate reader or uorescence using Ex/Em 535/587 nm. Correction of the background was performed by subtracting the value derived from the alanine control sample from all the sample readings. This is because the background reading can be signi cant and must be subtracted from all sample readings. Finally, we plotted the alanine standard curve and calculated the alanine levels as follows: C = Sa/Sv nmol/µl or mM.

Mechanical allodynia
Mechanical allodynia was assessed by measuring the paw withdrawal threshold in response to a calibrated series of von Frey hairs (Stoelting) [31]. The rats were individually placed in cages with a wire mesh bottom. A series of calibrated von Frey hairs were applied to the plantar surface of the hind paw in an ascending order (2, 4, 6, 8, 10, 15, and 26 g) with su cient force to bend the hair for 2 s or until paw withdrawal. Withdrawal responses were considered valid only if the hind paw was completely removed from the customised platform. Each hair was applied ve times; moreover, the minimum value causing at least three responses was recorded as the paw withdrawal mechanical threshold (PWMT).

Heat hyperalgesia (hot plate test)
Thermal hyperalgesia was assessed as previously described [32]. The thermal withdrawal latency after radiant heat stimulation was measured using an analgesia meter (Ugo Basile, Stoelting, IL, USA). Animals were placed in the chamber and allowed to acclimatise for 30 min before testing. Subsequently, a radiant heat source was focused under the glass oor beneath the hind paws. We adjusted the thermal-stimulus intensity to obtain a baseline thermal withdrawal latency of approximately 20 s. The digital timer automatically recorded the duration between stimulus initiation and the thermal withdrawal latency, with a 30-s cut-off being applied to prevent tissue damage. Each rat was tested at 5-min intervals, and the paw withdrawal thermal latency (PWTL) was calculated as the average of six trials.

Cold allodynia (acetone drop method)
Cold sensitivity was measured by applying an acetone drop to the plantar surface of the hind paw as previously described [33,34]. Rats were housed and habituated for 30 min in transparent plastic boxes using a wire-mesh oor. After the adaptation period, we gently applied acetone against the plantar skin of the left hind paws with an acetone bubble formed with a 0.1-ml syringe alternately three times to hind paws at 5-min intervals. Moreover, we recorded the duration of licking/biting and remaining in the air.
Each rat was tested at 5-min intervals, and the paw withdrawal cold duration (PWCD) was calculated as the average of six trials.

Sample Preparation
At predetermined time points after behavioural testing, the animals were deeply anaesthetised using sodium pentobarbital (40 mg/kg, intraperitoneal; Sigma) and sacri ced. Subsequently, ipsilateral L [4][5][6] DRGs and SDH tissues were collected. Samples for reverse transcription polymerase chain reaction (RT-PCR) and western blot experiments were snap-frozen in liquid nitrogen and stored at -80°C. We perfused samples used for immuno uorescence imaging through the ascending aorta with saline, followed by 4% paraformaldehyde in 0.1 M phosphate buffer (4°C, pH 7.4), as previously reported [35].

Immuno uorescence
The L4-6 DRG on the surgical side was removed and xed in 4% paraformaldehyde overnight, followed by dehydration in 20%/30% sucrose in phosphate buffer at 4°C. The tissue was cut into 5-µm thick sections with a cryostat (Leica CM1950, Nussloch, Germany). The sections were blocked using 20% bovine serum albumin for 1 h in a 37°C incubator (303-0S; Beijing Ever Bright Medical Treatment Instrument Co., Ltd., Beijing, China), washed using phosphate-buffered saline (PBS), and incubated with primary antibody (rabbit anti-GPER polyclonal antibody; 1:100, 260033, Abcam, rabbit anti-ERα polyclonal antibody; 1:100, ab32063; Abcam, rabbit anti-ERβ polyclonal antibody; 1:100, ab187291; Abcam) overnight at 4°C. After washing with PBS, sections were incubated with secondary antibody (TRITCconjugated anti-rabbit secondary antibody; 1:100; Santa Cruz Biotechnology, Heidelberg, Germany) for 1 h at 37°C. For double immuno uorescence staining, we incubated tissue sections with a mixture of anti-GPER antibody and antibodies against neuro lament-200 (NF-200, a marker for myelinated A-bres, 1:100, ab82259; Abcam, Cambridge, UK), calcitonin gene-related peptide (CGRP, a marker of peptidergic Ctype neurons, 1:100; ab81887; Abcam) for two nights at 4°C or IB4 (FITC-conjugated; a marker for nonpeptidergic C-type neurons, 5 µg/ml; L2895; Sigma). Except for IB4-treated tissue sections, the other sections were treated using a mixture of FITC-and TRITC-conjugated secondary antibodies at a 1:100 dilution for 1 h at 37°C. IB4 was 1:750 mixed using TRITC-conjugated secondary antibody. Sections were rinsed using 0.01 M PBS three times, mounted on gelatin-coated slides, and air dried. We visualised immunoreactivity using uorescence microscopy; further, a negative control was used by omitting the primary antibody to con rm the immunoreaction speci city. Sections were observed at 200 × magni cation using a confocal laser scanning microscope (LSM710; Carl Zeiss AG, Oberkochen, Germany). Optical density measurements and data analysis of GPER-positive cells for both types of DRG neurons were performed using Image-Pro Plus 6.0 (Media Cybernetics, MD, USA). We recorded the percentage uorescence results of the positive neurons of the three independent experiments.

Isolation of DRG Neurons
We used enzyme digestion to dissociate L 4-6 DRG neurons from the ipsilateral side of the operation as previously described [35]. DRGs in the drug-intervention group were treated using 17β-oestradiol and Βalanine. Brie y, we freed the excised ganglia from their connective tissue sheaths and cut them into pieces using a pair of sclerotic scissors in DMEM/F12 medium (GIBCO; Thermo Fisher Scienti c, Waltham, MA, USA) under low temperature on ice. The fragments were transferred into 5 mL of DMEM/F12 medium containing trypsin (0.4 mg/mL, Sigma) and collagenase (type IA, 0.6 mg/mL, Sigma), followed by incubation for 5 min at 37ºC. Subsequently, the ganglia were gently triturated using ne re-polished Pasteur pipettes. The suspension was dissociated in DMEM/F12 medium and supplemented with 10% foetal bovine serum; moreover, DRG neurons were plated on glass cover slips coated using Poly-L-Lysine (Sigma). We maintained the cells in a humidi ed atmosphere (5% CO 2 , 37°C) and used them for electrophysiological recordings 6-24 h after plating.

Transcriptomics Studies Library Preparation for Transcriptome Sequencing
DRG tissues of three rats in the sham, SNI, and SNI+G15 groups were randomly selected for transcriptome analysis. We extracted total RNA from approximately 150 mg of DRG tissue using a MirVana total RNA extraction kit (Ambion, Carlsbad, CA, United States). RNA levels were measured using the Qubit® RNA Assay Kit and a Qubit® 2.0 Fluorometer (Life Technologies, Carlsbad, CA, United States). We assessed RNA integrity using an RNA Nano 6,000 Assay Kit of the Bioanalyzer 2,100 system (Agilent Technologies, CA, United States). We used 1.5 µg of RNA per sample as the follow-up test material for RNA sample preparation. Whole transcriptome pro ling was performed using NEBNext® Ultra™ RNA Library Prep Kit for Illumina® (NEB, Ipswich, MA, United States) following the manufacturer's instructions.

Clustering, Sequencing, and Quanti cation of Gene Expression
We performed clustering of index-coded samples using a CBOT Cluster Generation system using HiSeq 4,000 PE Cluster Kit (Illumina). Next, we performed RNA sequencing (150 bp, pair-ends) using standard Illumina HiSeq 4000 platform protocols. All downstream analyses were based on high-quality clean data.
Gene FPKM was calculated by adding the FPKM of each genome transcript. The differential expression of both conditions was analysed using DESeq2R software package (1.26.0). For genes with FPKM values ≥1 in at least one sample, signi cant between-group differences in expression were determined according to the criteria as follows: p < 0.05 and |log2FoldChange| ≥ 0.5.

Gene Ontology (GO) and KEGG Enrichment Analysis of Differentially Expressed Genes
We used the Functional Annotation Bioinformatics Microarray Analysis (DAVID) database [37] to perform GO enrichment analysis of DEGs. GO terms (p < 0.05) were considered signi cantly enriched by DEGs. DEG enrichment in KEGG pathways was examined using KEGG Orthology Based Annotation System v3.0 software. P-values were calculated using one-way ANOVA based on the normalised dataset. KEGG terms (p < 0.05) were considered signi cantly enriched by DEGs.

Metabolomics Analysis
We randomly selected the DRG samples of 36 rats in the sham, SNI, and SNI+G15 groups for metabolomics analysis. Tissues were treated using methanol containing 2-chloro-L-phenylalanine (10 µg/ml) in a 4:1 ratio (methanol: plasma supernatant, v/v, 250 µl). After centrifugation at 4°C (4,000 rpm, 10 min), the supernatant was collected for analysis. We uniformly mixed a 50-µl aliquot of each centrifuged supernatant to yield a quality control (QC) sample. LC-MS analysis was performed using a Shimadzu UHPLC (ESI) system and an AB Sciex quadrupole time-of-ight mass spectrometer (TripleTOF 5600). All analyses were performed using the ACQUITY UPLC C18 column (100 mm × 2.1 mm, 1.7 µm, Waters). The mobile phase was a mixture of 0.1% formic acid in water (A) and acetonitrile (B). The proportion of mobile phase B was as follows: 2-30%, 0-3 min; 30-60%, 3-5 min; 60-80%, 5-15 min; 80-100%, 15-16 min; 100%, 16-19 min; 100-2%, 19-20 min; and 2%, 20-25 min. The injection volume, ow rate, and column temperature were 2 µl, 0.3 ml/ min, and 40°C, respectively. Time-of-ight mass spectrometry (TOF-MS) and TOFMS/ MS were synchronously performed. Mass spectra were acquired in the positive and negative ion modes. The parameters of the positive ion mode were as follows: ion spray voltage oating, 5,500 V, and collision gas, 40 eV, while those in the negative ion mode were as follows: ion spray voltage oating, −4,500 V, and collision gas, −40 eV. The turbo spray temperature was 550°C, while the TOF-MS mass ranged from m/z 50-1,250 Da. During analysis of the sample sequence, we run a QC sample after every ve injections to validate the analytical methodology . QC (n = 6) and plasma samples were analysed using LC-MS. We evaluated data quality based on the relative SDs of the retention times and intensities of 10 typical peaks (including internal standard) of the QC samples.

Electrophysiological Recordings
All recordings were performed on small-and medium-diameter (20-35 µm) neurons as previously described [38]. Coverslips with DRG neurons were mounted in a small ow-through chamber positioned on the stage of an inverted microscope (Nikon Eclipse Ti, Tokyo, Japan) to select DRG cells with smooth membrane surfaces and good translucency. We continuously perfused coverslips with gravity-driven bath solution. We performed standard whole-cell patch-clamp recordings of isolated DRG neurons at room temperature (22°C) using an EPC-10 ampli er and the PULSE program (HKA Electronics, Lambrecht, Germany). Membrane capacitance was read from the ampli er using PULSE to measure the size of cells and current densities. Glass pipettes (3-5 MΩ) were prepared using a Sutter P-87 puller (Sutter Instruments, Novato, CA, USA). Action potentials were elicited by a series of depolarising currents from 0 to 500 pA (150 ms) in 50-pA step increments under the current clamp mode to measure the current threshold (rheobase) in the vicinity of the explosive action potential current. The current was altered at 10-pA steps, i.e., the minimal current that evoked an action potential, which is a parameter for excitability. The recorded signal was ampli ed using a MultiClamp 700B ampli er (Molecular Devices, LLC, Sunnyvale, CA, USA), ltered at 10 kHz, and converted by an Axon Digidata 1550A D/A converter (Molecular Devices) at a 10-20 kHz sampling frequency. Voltage errors were minimised using 80-90% series resistance compensation; further, we used linear leak subtraction for all recordings. For the current clamp experiments, the bath solution contained the following (in mM): 140 NaCl, 5 KCl, 2 CaCl 2 , 2 MgCl 2 , 10 D-glucose, and 10 HEPES. The pH was adjusted to 7.4 using NaOH. The pipette solution contained the following (in mM): 30 KCl, 100 K-aspartate, 5 MgCl 2 , 2 Mg-ATP, 0.1 Na-GTP, and 40 HEPES, with adjustment of the pH to 7.2 using KOH. All chemicals were obtained from Sigma.

Statistical Analysis
All data are expressed as mean ± SEM of three independent experiments. Prior to further statistical analysis, the normal distribution hypothesis of the test data and homogeneity of variance were examined. Statistical analyses were performed using SPSS 20.0 (SPSS Inc., Chicago, IL, USA) and Graphpad 9.0 (California, USA). We analysed PWMT, PWCD, and PWTL using repeated-measures analysis of variance.
Multiple between-group comparisons at each time point were conducted using Bonferroni's post-hoc tests. Among-group comparisons of the western blot, PCR, and patch-clamp data were conducted using one-way analysis of variance followed by Tukey's post-hoc tests. Between-group comparisons were performed using Student's t-test. Statistical signi cance was set at P < 0.05.

Result
3.1 Development of cold and mechanical allodynia in SNI rats had no effect on cold and thermal hypersensitivity. Time courses of the withdrawal latency to thermal stimuli were determined in the sham and SNI groups. Radiant heat stimulus (IR 50) was applied to the left paws of the rats. BTime courses of the withdrawal duration to cold stimuli were measured in the sham and SNI groups. Acetone (100 µl) was used on the left hind paw of the rats. Time courses of withdrawal thresholds in the sham and SNI groups to von Frey hair stimuli on day 1 before and on days 3, 5, 7, 14, and 21 after the surgery. Mechanical stimuli were applied to the left hind paw of rats with von Frey hairs. (Fig. 1A), n = 6, compared with the sham group at the same time point, *** p < 0.001.
3.2 CGRP loss in the SDH as well as co-expression of the three ERs in the missing area and mainly expressed in pain-related small and medium-sized neurons on the SNI injury side. On the SNI injury side, there was signi cant CGRP loss in the SDH. (Fig. 1B, C). To determine the function of ERs in DRG neurons, we initially used immuno uorescence double staining to characterize the expression of ERs in DRG neurons. The percentages of IB4 (a marker of nonpeptidergic C-type neurons), CGRP (a marker of peptidergic C-type neurons)-,and NF-200 (a marker of A-type neurons). (Fig. 1D). Immuno uorescent double staining experiments showed that ERs colocalized with IB4, CGRP, and NF-200 (Fig. 1E). The percentages of IB4-, CGRP-, and NF-200-positive neurons relative to the percentage of ERs -positive cells respectively (n = 6 in each group; Fig. 1G). These results showed that ERs was mainly located in A-and Ctype neurons in the DRG. The neuronal diameter size ranges of IB4, CGRP, and NF-200 were 31.00 ± 1.13, 17.75 ± 0.87, and 42.75 ± 1.917, respectively (n = 10 in each group) (Fig. 1F).

ERα and GPER was signi cantly increased GPER and ERα expression in the DRG but not the SDH.
Immuno uorescence staining and PCR detection of rat ipsilateral L4-6 DRGs at spared nerve injury (SNI) showed that the protein and mRNA levels of GPER and ERα were higher than those in sham-operated rats on day 5 after the operation (p < 0.01) in each group ( Fig. 2A and 3C-E). Notably, the ERs uorescence intensity protein and mRNA levels in SDH had no change. (Fig.s 2B, 3C-E).Thus, GPER and ERα is expressed at a higher level in small and medium-sized neurons associated with pain signaling in DRG. (Fig.s 2B, 3C-E).
3.5 GPER can be involved in neuropathic pain by regulating α2-GABAA. Double immuno uorescence staining showed that GPER and α2-GABAA are colocalized in the dorsal root ganglia, (Fig.s 3 G), indicated that GPER may be involved in the regulation of super cial sensations such as pain like α2-GABAA. α2-GABAA in the DRG plays an important role in pathophysiology of NPP caused by sciatic nerve injury [39]. (Fig. 4F) GPER and ERα both upregulated post SNI operation, therefore, normal rats were administered with G1, PPT, and DPN, which are GPER-speci c, ERα-speci c, and ERβ-speci c agonists, respectively. Only GPER-speci c agonists could induce hyperalgesia similar to the SNI model. (Fig.s 3A, B). SNI induced loss of GABAα2 restored after G15 in sheath. The mRNA levels were consistent with the western blotting results, which indicates that GPER can be involved in neuropathic pain by regulating α2-GABAA. (Fig.s 3F, 4C, D).
3.6 Transcriptomics testing found that GPER involved in neuropathic pain regulation focus on metabolic pathways while regulating neuroin ammation. We examined the gene expression pro les with a focus on mRNA in neurons of Sham, SNI, and SNI+G15 rats through RNA-Seq. A total of 18439 genes were expressed in the three groups (Fig. 5A). There are a total of 154 genes with differences between groups (Fig. 5B). Compared with the SNI+G15 group, 36 genes were up-regulated in the SNI group, and 24 genes were down-regulated, there were signi cant statistical differences (Fig. 5C, D). KEGG analysis of upregulated and down-regulated genes shows that the metabolic pathway has the most signi cant statistical difference (Fig. 5E, F). From this we can perform metabolomics testing on the three groups to nd the key metabolites that cause GPER changes. At the same time, we unexpectedly discovered that GPER can regulate the expression of in ammatory factors IL-6 and IL-1β, and GPER blockers can signi cantly reduce neuroin ammation (Fig. 5G).

3.7
Metabolomics tests showed that alanine could cause GPER upregulation. According to the results of genomics, targeted metabolomics has detected more than 300 metabolites, and found that 14 substances have differences between groups, β-alanine, Aminocaproic acid, 1H-Indole-3-acetamide, Nicotinic acid, Aminoadipic acid, N-Acetylglutamine, Acetic acid, alpha-Hydroxyisobutyric acid, Methylsuccinic acid, Lithocholic acid, L-Carnitine, 2,2-Dimethyladipic acid, 2,2-Dimethylsuccinic acid, Butyric acid (Fig. 6A, B). The Sham group compared with the SNI group, the SNI group compared with the SNI+G15 group, the substances that have changed are β-alanine and Nicotinic acid (Fig. 6C). We were especially interested in pathways involved in pain and in ammation responses, since these pathways may get involved in mediating neuroin ammation and pain mechanisms in SNI model rats. Therefore, we continued to perform PPI analysis of the metabolites that are related to process we identi ed from KEGG analysis, β-alanine may be a key factor causing the increase in GPER (Fig. 6D, E).
3.8 β-alanine and GPER can interact positively and co-cause the excitability of DRG neurons and neuroin ammation similar to SNI-induced. To further determine the contribution of β-alanine to the increased GPER in neuropathic pain, β-alanine 10 mM or 1% DMSO as vehicle, 10 µL, was administered intrathecally to normal rats for 24h mechanical and cold hyperalgesia, itching signi cantly altered in Rats (Fig. 7A). Post-SNI operation, the alanine content increased, and the alanine content continued to accumulate until 14 days later (Fig. 7D). After administration of GPER blocker, the content of alanine was signi cantly lower than that after SNI, and the content of alanine was not reduced after administration of GABA, so it can be inferred that alanine is upstream of GABA (Fig. 7B). We speculated that there was an interaction between GPER and β-alanine. In normal rats, β-alanine content was signi cantly increased after administration of GPER agonist G1 (Fig. 7C). After the administration of β-alanine, the western-blot technique was used to detect a signi cant increase in the GPER protein content, which can con rm that there is a positive interaction between GPER and β-alanine, and a cascade effect is formed after SNI (Fig.   7E). Increased of GPER and accumulation of β-alanine simultaneously up-regulate downstream IL-6, IL-1β, and down-regulate GABA to cause hyperalgesia and neuroin ammation (Fig. 7F). Normal rats given GPER blockers G15 have no obvious effect on neuronal excitability and neuroin ammation, suggesting that GPER plays a role in pathological conditions (Fig. 7G). Temperature gradient experiments show that in DRG tissues incubated with β-alanine, GPER degrades more slowly, indicating that alanine can combine with GPER and play a role (Fig. 8). From a functional point of view, the patch clamp technique veri ed that the increase of GPER and β-alanine caused the excitability enhancement of DRG neurons similar to that induced by SNI (Fig. 7H, I), and revealed the pathogenesis of neuropathic chronic pain (Fig.  9).

Discussion
Currently, neuropathic pain cannot be effectively treated since its exact cause remains unclear [10,40].
Sensitisation is crucial for neuronal hyperactivity and is characterised by overreacting to normal stimuli [41]. Ectopic discharge is the electrophysiological basis of allodynia; moreover, the change of ion channels is the main reason for abnormal discharges and α2/α3 GABAA-mediated analgesia in neuropathic pain [11,42]. Increasing studies have shown that estrogen can affect pain by regulating ion channels; however, there are con icting opinions regarding pain and analgesia [42]. Estrogen mainly acts through its three receptors GPER, ERα, and ERβ [5]. However, few studies have investigated the role of ER in peripheral pain transduction. We observed signi cant post-SNI GPER upregulation; moreover, blocking GPER can relieve hyperalgesia. These ndings demonstrated the possible mechanism of neuropathic pain onset and persistence from a perspective of material metabolism.
Oestradiol modulates the e cacy of synaptic inhibition by decreasing the dwell time of GABAA receptors at inhibitory synapses [26]. Mouse models of in ammatory and neuropathic pain have revealed that the novel α2-GABAA receptor is antihyperalgesic [20]. Therefore, we chose other receptors as the reference molecular biological indicators of pain. In this study, after establishing the SNI model, there were behavioural changes in mechanical and cold hyperalgesia; moreover, there was no signi cant change in thermal pain. In the follow-up study, we chose PWMT and PWCD as the behavioural reference indicators.
On the SNI injury side, there was signi cant CGRP loss in the SDH as well as co-expression of the three ERs in the missing area. Moreover, our ndings demonstrated that GPER, ERα, and ERβ are mainly expressed in pain-related small and medium-sized neurons. DRG neurons of various sizes have different responses to various stimuli [43]. Small-and medium-sized neurons speci cally respond to harmful thermal, chemical, and mechanical stimuli; therefore, they are considered multimodal nociceptors. Furthermore, the DRG and SDH are closely associated with pain transmission; moreover, we found that the three ERs are closely related to pain. Estrogen could in uence P2X3 expression via the ERα and GPER ERα to affect neuropathic pain, which may be mediated through the ERK pathway [44]. Deleting the ER subunit α (ERα) in TRPV1 nociceptors abolishes IL-23-and IL-17-induced pain in females [45]. ER-β and G protein-coupled ER-1, but not ER-α, in the rostral anterior cingulate cortex are involved in pain-related aversion by modulating N-methyl-D-aspartate receptor-mediated excitatory synaptic transmission [46]. Behavioural tests revealed that the PWMT and PWCD peaked on the fth day and lasted > 14 days in SNI animals. Therefore, we choose the fth day as the timepoint for selecting animals for the experiment and drug intervention. SNI animals have been reported to maintain strong mechanical allodynia throughout an 85-day observation period [47].
We used immuno uorescence to detect the expression of three ERs in DRG neurons and SDH areas with CGRP deletion. On the fth day of SNI, there was signi cantly increased ERα and GPER expression in the DRG but not the SDH. Western blot and PCR results were consistent with those of immuno uorescence.
G15, which is a speci c blocker of GPER, can relieve hyperalgesia in SNI rats; however, MMP, which is a speci c ERα blocker, cannot. Therefore, normal rats were administered with G1, PPT, and DPN, which are GPER-speci c, ERα-speci c, and ERβ-speci c agonists, respectively. Only GPER-speci c agonists G1 could induce hyperalgesia similar to the SNI model. Our ndings suggest that GPER could be involved in SNI-induced neuropathic pain development; moreover, GPER is an important molecular component contributing to peripheral pain transduction. A study using a rat model of visceral pain demonstrated that spinal ERα mediates oestradiol-induced pronociception [48]. The ERβ and ERα have speci c advantages regarding analgesic effects [49]. Our ndings suggest that GPER up-regulation in the DRG, but not the SDH, may be an important component in SNI-induced neuropathic pain. The roles of GPER downstream signalling pathways remain unclear. This study provided evidence regarding chronic pain by using the α2-GABAA receptor to evaluate pain behaviour and the role of GPER in pain. SNI allodynia is followed by downregulating the α2-GABAA receptor, which is improved by intrathecal G15 administration; moreover, mechanical pain and cold hyperalgesia are alleviated. Additionally, we tested the expression changes of other GABA subunits. The mRNA levels were consistent with the Western blotting results, which indicates that GPER can be involved in neuropathic pain by regulating α2-GABAA. Moreover, there was high α2-GABAA and GPER expression on small and medium DRG neurons. Therefore, we will conduct follow-up research on GPER expression on the DRG.
We successfully established the rat SNI model and intrathecally administered GPER blockers. We examined the gene expression pro les with a focus on mRNA in neurons of SNI rats, sham rats, and SNI+G15 rats through RNA-Seq. We found several differentially expressed mRNAs. Further, we validated their expressions via the protein assay and examined the molecular functions, cellular components, and enriched biological processes of these DEGs by applying bioinformatics analysis. A sizeable proportion of these DEGs were exclusively involved in in ammatory processes; furthermore, IL-6 and IL-1β are of signi cant characteristic signi cance. Injury to the peripheral sensory nerves causes a neuroin ammatory response in the somatosensory pathway, from the DRG to the spinal cord, which contributes to neuropathic pain [50]. Activating nociceptor sensory neurons through noxious stimuli triggers pain as well as increases capillary permeability and blood ow to yield neurogenic in ammation, which results from the antidromic activation of nociceptor peripheral terminals [51]. Additionally, ceRNA analysis further identi ed the direct or indirect regulatory relationship that may exist in the GPER metabolic pathway under SNI conditions. Chronic pain, especially high-intensity chronic pain, is associated with an increased prevalence of cardiovascular risk factors and metabolic syndrome [52]. GPER upregulation is crucially involved in promoting neuroin ammation and DRG sensitisation during chronic pain. We performed metabolomic analysis in neurons of rats in all three groups. Compared with the sham and SNI+G15 groups, the SNI group showed changes in β-alanine and nicotinic acid. β-alanine showed the most signi cant changes through enrichment analysis among a series of metabolites. Further, administering G15, which is a GPER blocker, reverses the SNI-induced increase in alanine and alleviates behavioural problems. A local intradermal injection of β-alanine directly induces itching in humans, which con rms that β-alanine induces itching through a peripheral, cutaneous mechanism [53,54]. Additionally, certain mechanical stimuli mediated by recently discovered circuits contribute to the itch sensation in the spinal cord and brain. These circuits that mediate touch, pain, and itching could engage in cross modulation [55]. We administered β-alanine into the upper sheath of normal rats, and the behavioural test revealed that rats showed mechanical and cold hyperalgesia upon itching occurrence. After SNI, there was a continuous increase in the β-alanine levels of the rat DRG until 14 days. Western blot experiments also revealed an increased GPER levels, increased in ammation, decreased α2-GABAA levels, and excitation of DRG neurons after SNI. The high excitability state of DRG neuron cells after the G15 SNI model can be relieved. Moreover, there was enhanced excitability of DRG neurons in normal rats after incubation with G1 and βalanine. Supplementation with β-alanine has been shown to delay accumulation of lactate during exercise through buffering the formation of lactate from pyruvate [56]. In this study, we selected smalland medium-sized neurons. Therefore, our ndings demonstrate that the post-SNI GPER upregulation in small DRG neurons increases the levels of the in ammatory cytokines IL-1β and IL-6, as well as decreases α2-GABAA levels. In addition, β-alanine accumulation can be reversed through G15 administration. Both the secretion of in ammatory factors and the ring of neurons require energy consumption [57]. It is also possible that the energy generated during the continuous accumulation of alanine may cause the DRG cells to continuously produce in ammation and nerve excitability. β-alanine can also positively regulate GPER formation similar to the waterfall effect. TGR7 functions as a speci c membrane receptor for β-alanine, which corresponds to MrgD. Mas-related GPCR D is speci cally expressed in small-diameter nociceptive DRG neurons and is implicated in pain modulation[58]. We con rmed that β-alanine could ligate the GPER, which could explain the chronicity of neuropathic pain.
Notably, it remains unclear whether the effect of estrogen is the main contributor to sex differences in pain [59][60][61]. We can choose to locally deplete alanine and block GPER as an effective clinical treatment for chronic neuropathic pain

Conclusions
In conclusion, we speculate that GPER activation in the DRG causes a positive interaction of β-alanine with activation of IL-1β and IL-6 production and represses GABAα2 involved in post-SNI neuropathic pain development. Therefore, GPER blocker and β-alanine elimination may be an underlying molecular target for controlling neuropathic pain without affecting normal sensation.

Consent for publication
Not applicable.

Availability of data and materials
The key data are contained in the gures, tables, and additional les. The datasets used and/or analyzed during this study can be further obtained from the corresponding author on reasonable request.

Competing interests
The authors declare that the research was conducted in the absence of any commercial or nancial relationships that could be construed as a potential con ict of interest.       (two-tailed Student's t test); n = 3. (E) Western blot images for GPER, ERα, ERβ, IL-6, IL-1β, GABAα2 in the DRG after intrathecal injection of G1 and β-alanine in rats. (F) Both GPER, IL-6, IL-1β were increased, GABAα2 downregulated in the dorsal root ganglion (DRG) of G1 group rats 6 in total. **P < 0.01, ***P < 0.001 versus Vehicle, (two-tailed Student's t test); n = 3 and 3. (G) Both GPER, IL-6, IL-1β, GABAα2 had no signi cant. (H) Current threshold (rheobase) was determined as the current required for activating the rst action potential. On the right, representative traces of action potentials (APs) evoked by current injections into DRG neurons from Sham, SNI, SNI+G15, G1, β-alanine groups; n = 4-6 neurons per group; On the right, twice in the Fig.,    The mechanisms diagram of GPER/β-alanine mediating neuronal sensitisation and neuroin ammation in modulating neuropathic pain. Show that upregulating GPER in DRG cells signals to sensory neurons upon peripheral nerve damage, inducing neuronal β-Alanine increase while positive promote GPER, the combined effect of them further activates IL-1β and IL-6 ,regulates GABAα2 in the DRG. The GPER/β-Alanine positive interaction in DRG uncovers enhances pain sensation and promotes chronic pain development.