PRMT6 inhibition or deficiency attenuates diabetic neuropathic pain in male mice and is associated with reduced spinal neuroinflammation, microgliosis, and altered p53-p21 signaling

Abstract

Diabetic neuropathic pain (DNP) is a major contributor to chronic pain in adults, yet effective targeted therapies are still lacking, underscoring the need to elucidate its underlying mechanisms. Microglial proliferation and activation are key drivers of central sensitization and pain hypersensitivity. In a type 2 diabetes mouse model, protein arginine methyltransferase 6 (PRMT6) was markedly upregulated in spinal dorsal horn microglia in male mice, and high-glucose stimulation similarly increased PRMT6 expression in BV-2 cells, accompanied by enhanced proliferation and inflammatory activation. Genetic deletion of Prmt6 or pharmacological inhibition with EPZ020411 alleviated pain hypersensitivity and reduced spinal microgliosis and inflammation in male mice. Transcriptomic analysis revealed enrichment in cell proliferation-related processes and the p53 signaling pathway. In BV-2 cells, PRMT6 knockdown induced G0/G1 arrest and attenuated high-glucose-induced proliferation and inflammatory activation, whereas PRMT6 overexpression exerted opposite effects. Mechanistically, PRMT6 methylated p53 and decreased its transcriptional activity, leading to reduced p21 mRNA expression and enhanced cell-cycle progression. In contrast, in female DNP mice, spinal microgliosis was limited, PRMT6 expression remained unchanged, and Prmt6 deficiency did not significantly alter spinal microglial density, inflammatory markers, or nociceptive hypersensitivity. Collectively, our results uncover a previously unrecognized PRMT6–p53–p21 regulatory axis that may contribute to microglial proliferation and neuroinflammation under hyperglycemic conditions in male mice, highlighting PRMT6 as a potential therapeutic target for microglia-associated DNP.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12974-026-03893-1.

Introduction

Diabetes is one of the major diseases affecting human health worldwide. More than 50% of diabetic individuals get diabetic neuropathy as their condition worsens, and between 30 and 50% of them develop neuropathic pain [1, 2]. Clinically, DNP is characterized by spontaneous pain, allodynia, and paresthesia, predominantly in the distal limbs [3]. These symptoms are frequently accompanied by anxiety and depression, severely impairing patients’ quality of life. The primary preventive measures remain optimal glycemic control and lifestyle modification, but their effectiveness is limited, especially in type 2 diabetes [4]. There isn’t a disease-modifying treatment for DNP at the moment; the medications that are available, like NSAIDs, opioids, antidepressants, and anticonvulsants, mostly relieve symptoms and are frequently limited by their inadequate efficacy and side effects [5, 6]. Therefore, it is crucial to comprehend the molecular mechanisms underlying DNP and find new treatment targets.

Neuroinflammation mediated by spinal microglia is increasingly recognized as a central mechanism underlying central sensitization and chronic neuropathic pain [7–10]. As the resident immune cells of the central nervous system, microglia continuously monitor their local microenvironment and respond to metabolic stress, inflammatory cues, and damage-associated signals [11]. Following nerve injury, damaged sensory neurons trigger microglial activation and proliferation by releasing key signaling molecules, most notably CSF1, CX3CL1, and ATP. These mediators act on their respective microglial receptors—CSF1R, CX3CR1, and purinergic receptors—to orchestrate the neuroinflammatory response [12–14]. Activated microglia proliferate and release proinflammatory mediators, including IL-1β and TNF-α, which act on postsynaptic neurons to enhance excitability and reshape synaptic plasticity [15]. In turn, cytokines such as IL-1β [16], TNF-α [17], and CCL2 [18], further stimulate microglial proliferation, establishing a feed-forward loop that amplifies neuroinflammation and pain hypersensitivity. Consistent with this, microglial proliferation is considered a critical prerequisite for neuropathic pain development [12], and its pharmacological or genetic inhibition markedly attenuates pain hypersensitivity in several neuropathic pain models, including spared nerve injury (SNI) [19] and spinal nerve transection (SNT) [12]. Likewise, under diabetic conditions, spinal dorsal horn microglia display aberrant proliferative and activated phenotypes [20]. Therapeutic interventions such as minocycline [21] or gabapentin [22] primarily mitigate microglial activation and neuroinflammatory signaling—rather than directly inhibiting proliferation—thereby alleviating diabetic neuropathic pain. Nevertheless, the molecular mechanisms that trigger and sustain excessive microglial proliferation under hyperglycemic stress remain largely undefined.

Protein arginine methyltransferases (PRMTs), comprising nine known members (PRMT1–9), catalyze the methylation of both histone and non-histone proteins. This post-translational modification serves as a key epigenetic mechanism that regulates diverse biological processes [23]. Among them, the type I enzyme PRMT6 regulates cell metabolism [24, 25], proliferation [26, 27] and inflammatory responses [28, 29] across many diseases. For example, PRMT6 promotes tumor cell mitosis and proliferation in glioblastoma by methylating the regulator of chromatin condensation 1 (RCC1) [30]. Notably, our previous work demonstrated that PRMT6 promotes microglial activation via the HIF-1α pathway in chronic constriction injury [31] and exacerbates postherpetic neuralgia by suppressing the cGAS–STING pathway [32], positioning PRMT6 as a key regulator of neuroinflammatory signaling. Nevertheless, whether PRMT6 contributes to DNP and how it might modulate microglial proliferation remain unexplored.

Cell proliferation is ultimately governed by the orderly progression through the cell cycle, and disruption of cell cycle in microglia has been shown to attenuate neuropathic pain in preclinical models [33, 34]. The tumor suppressor p53 is a central regulator of the cell cycle, orchestrating cell-cycle arrest, cellular repair, or apoptosis in response to stress signals [35]. Through its regulation of downstream targets such as p21, p53 induces arrest at the G1/S or G2/M control points, thereby securing the necessary time for DNA repair [36]. Nevertheless, studies on the regulatory function of p53 in microglial proliferation are relatively limited.

In this study, we found that diabetic conditions increased PRMT6 expression in the spinal dorsal horn of male mice, particularly in microglia. Prmt6 deficiency alleviated pain hypersensitivity in diabetic male mice and was accompanied by reduced microgliosis and neuroinflammation. In BV-2 cells, PRMT6 promoted the G1/S phase transition and accelerated cell proliferation. Further mechanistic analyses revealed that PRMT6 methylated p53 and decreased its transcriptional activity, leading to reduced p21 mRNA expression. In addition, we examined a separate cohort of female mice and found that the in vivo effects of PRMT6 were sex dependent. Together, these findings suggest that PRMT6 may promote microglial proliferation under hyperglycemic stress via the p53-p21 signaling axis, thereby contributing to the development of diabetic neuropathic pain.

Materials and methods

Animals

The experimental subjects primarily comprised male C57BL/6J mice with a confirmed specific pathogen-free (SPF) status, sourced from the Experimental Animal Center of the Naval Medical University. In parallel, a genetic mouse model was established on this background, in which Prmt6 was completely ablated using CRISPR/Cas9-mediated genomic editing. Two guide RNAs (5’-ACGAATCCCAGCAGGCCCCG-3’ and 5’-GAGATCGCCTATGCAAGTTG-3’) and Cas9 mRNA were microinjected into fertilized wild-type C57BL/6J embryos. Founder (F0) mice carrying Prmt6 deletions were crossed with wild-type mice to produce heterozygous (Prmt6^+/−) offspring, which were intercrossed to obtain homozygous Prmt6^−/− mice. Given that no significant differences in body weight, blood glucose levels, or nociceptive sensitivity were observed between male Prmt6^+/+ and Prmt6^+/− littermates under our experimental conditions (Supplementary Fig. 4A–E), Prmt6^+/− littermates generated from Prmt6^+/− × Prmt6^−/− breeding pairs were used as controls in the male experiments. This breeding strategy enabled the efficient generation of sufficient littermate controls and knockout mice while minimizing unnecessary animal use. All mice were maintained in an SPF facility at 24 °C and 60% relative humidity with ad libitum access to food and water. Male littermates aged 6–8 weeks were used for the main mechanistic and intervention experiments. Additional validation experiments were independently performed in age-matched female littermates to assess potential sex-dependent effects of Prmt6 deficiency in DNP. All experimental procedures involving animals were performed in accordance with the guidelines of the International Association for the Study of Pain (IASP).

Induction of diabetic neuropathic pain model

Mice were randomly assigned to control and DNP groups. In both male and female cohorts, DNP mice were fed a high-fat diet (60.2% fat, 19.6% carbohydrate, and 20.2% protein; XTHF60, Xietong Bio-Engineering, China), whereas control mice received a standard diet (10% fat, 69.9% carbohydrate, and 20.1% protein; XTCON50J, Xietong Bio-Engineering, China) using a 7-day food replacement protocol. After 8 weeks of high-fat feeding, mice were fasted overnight and administered streptozotocin (STZ, intraperitoneally, S1312, Selleck, USA) once daily for 4 consecutive days to induce type 2 diabetes, while control mice received an equal volume of vehicle (0.1 mol/L citrate buffer, Thermo Scientific). The STZ dose was 35 mg/kg in male mice and 50 mg/kg in female mice, as female mice are relatively less sensitive than male mice to high-fat diet/STZ-induced type 2 diabetes [37, 38]. Three days after the final injection, fasting blood glucose was measured from the tail vein using a glucometer (Sinocare) following an 8-h fast (8:00–16:00). Mice with fasting blood glucose ≥ 11.1 mmol/L were considered diabetic. Blood glucose levels, thermal withdrawal latency, and mechanical withdrawal thresholds were assessed on days 0, 3, 7, 14, and 21 to confirm DNP model establishment. Mice that did not meet the criteria were excluded.

Assessment of pain sensitivity

Thermal hyperalgesia

Following an hour of acclimation in clear enclosures on a smooth floor, thermal nociception was tested. A designated area on each hind paw’s plantar surface was stimulated by a focused radiant heat source from an IITC Model 33 apparatus. Paw withdrawal latency was measured up to three times per paw with a 10-minute interval, applying a 20-second safety cutoff to prevent tissue harm.

Mechanical hyperalgesia

After a one-hour habituation, mechanical nociception was quantified using von Frey filaments (Stoelting Co., USA). Tactile allodynia was tested with a 0.07 g filament, and mechanical hyperalgesia was gauged using a 0.4 g filament [39]. Each stimulus was delivered for one second perpendicular to the hind paw’s central plantar region. A withdrawal response was recorded as positive (1) or negative (0). Each filament was applied 10 times per paw with 5-min intervals, and the withdrawal frequency (%) was calculated.

Western blotting

Spinal cord dorsal horn tissues from L4-L6 segments or cultured cells were lysed using a commercially available RIPA lysis buffer (PC101, Epizyme) with a proprietary cocktail designed to inhibit both protease and phosphatase activity. Tissues were homogenized in a cryogenic grinder, and cells were scraped after washing. Lysates were sonicated and centrifuged. Protein content in the lysates was quantified using the standard bicinchoninic acid protein assay (ZJ102, Epizyme). Equal amounts of protein were denatured in loading buffer at 95 °C for 5 min, separated by SDS–PAGE, and transferred to PVDF membranes. After blocking, membranes were incubated overnight at 4 °C with primary antibodies, followed by HRP-conjugated secondary antibodies. The resultant protein bands were visualized using a standard enhanced chemiluminescence detection system (SQ201, Epizyme). Densitometric analysis was performed using ImageJ, and the corresponding quantitative data are presented in Supplementary Fig. 5. The following antibodies were used: PRMT6 (14641, CST, 1:1000), Iba1 (17198, CST, 1:1000), β-actin (AC026, ABclonal, 1:200000), NLRP3 (15101, CST, 1:1000), phospho-NF-κB (3033, CST, 1:1000), NF-κB (8242, CST, 1:1000), p53 (2524, CST, 1:1000), p21 (37543, CST, 1:1000), ADMA (13522, CST, 1:1000), phospho-Rb (8516, CST, 1:1000), Rb (ab181616, Abcam, 1:1000), anti-Rabbit IgG HRP Ab (Ab205718, Abcam, 1:10000), anti-Mouse IgG HRP Ab (ab6789, Abcam, 1:5000).

Quantitative real-time PCR (qPCR)

Total RNA was isolated from dorsal horn tissue sections or harvested cell cultures using FreeZol (R711, Vazyme) following the manufacturer’s protocol for protein removal, isopropanol precipitation, and repeated 75% ethanol washing. RNA concentration and purity were determined by NanoDrop spectrophotometry. For complementary DNA (cDNA) synthesis, one microgram of total RNA was subjected to reverse transcription with an All-in-One RT SuperMix reagent (R333-01, Vazyme). Quantitative PCR reactions were carried out in triplicate employing the manufacturer’s instructions for SYBR Green Master Mix (Q711-03, Vazyme) in a QuantStudio 5 system. Gene-specific primer sequences are detailed in Table 1. Expression levels of target genes were computed according to the 2^−ΔΔCt method after normalization to β-actin.

Table 1.

Primer sequences for RT-qPCR

gene	Forward (5’-3’)	Reverse (5’-3’)
Prmt6	GATGGGCTACGGACTTCTGC	GCATCTGGTCGCTAATCGGG
Actin-β	CACTGTCGAGTCGCGTCC	TCATCCATGGCGAACTGGTG
Il-1β	TGCCACCTTTTGACAGTGATG	AAGGTCCACGGGAAAGACAC
Tnf-α	TTTCACTCACTGGCCCAAGG	AGGCCATTTGGGAACTTCTCATC
Trp53	ACCGCCGACCTATCCTTACCATC	GGCACAAACACGAACCTCAAAGC
Cdkn1a	CCTGGTGATGTCCGACCTG	CCATGAGCGCATCGCAATC
Ki-67	ATCATTGACCGCTCCTTTAGGT	GCTCGCCTTGATGGTTCCT

Immunohistochemistry and immunofluorescence

Mice were deeply anesthetized with sevoflurane and perfused with phosphate-buffered saline (PBS) followed by 4% solution of paraformaldehyde. Lumbar spinal cords were isolated, post-fixed, dehydrated, embedded in paraffin, and sliced at 5 μm. After deparaffinization, rehydration, and antigen retrieval in EDTA buffer using microwave heating, the slides were blocked with 3% bovine serum albumin (BSA).

For immunohistochemistry (IHC), slides were incubated overnight with anti-PRMT6 antibody (anti-rabbit, 15395-1-AP, Proteintech, 1:100). After thorough washing, tissue sections were exposed to a horseradish peroxidase-conjugated secondary antibody (GB23303, Servicebio; 1:200). Subsequently, chromogenic reaction was allowed to proceed using a DAB substrate kit (G1212, Servicebio). Nuclei were then counterstained, and images were acquired under a light microscope.

For immunofluorescence (IF), sections were incubated overnight at 4 ℃ with primary antibodies including PRMT6 (anti-mouse, sc-271744, Santa Cruz, 1:100), Iba1(17198 S, Cell Signaling Technology, 1:200), GFAP (3670 S, Cell Signaling Technology, 1:200), NeuN (ab177487, Abcam, 1:200), CD68 (ab283654, Abcam, 1:50). Following several washes, the prepared sections were exposed to Alexa Fluor-tagged secondary antibodies for detection. Cell nuclei were visualized with fluorescent DAPI dye before being sealed under glass coverslips using a specialized anti-quenching solution. All fluorescence images were documented using a designated Nikon microscope system.

Transmission electron microscopy (TEM)

Euthanasia was performed by cervical dislocation after deep inhalation anesthesia with sevoflurane. Small lumbar spinal cord dorsal horn blocks (≤ 1 mm³) were excised and immediately placed in 2.5% glutaraldehyde in 0.1 M PBS for 2 h. After washing, samples were fixed in 1% osmium tetroxide (Ted Pella Inc., #18456) at 4 °C for 2 h and then washed again. After dehydrating in graded ethanol, tissue blocks were infiltrated and embedded in EPON 812 resin and polymerized for 48 h at 60 °C. Ultrathin sections (~ 60 nm) were cut, mounted on copper grids, and stained in the dark with 2% uranyl acetate for 8 min. Sections were rinsed in 70% ethanol, then in ultrapure water, followed by staining with 2.6% lead citrate for 8 min in a CO₂-free environment and a final rinse. Prepared grids were examined by transmission electron microscopy (HT7800, Hitachi), and representative images were captured at appropriate magnifications for analysis.

Cell culture

BV-2 microglial cells, HEK-293T cells, and primary mouse microglia were used for the in vitro experiments. BV-2(QuiCell-B151) and HEK-293T(QuiCell-H1360) cells were kindly provided by Shanghai QuiCell Biotechnology Co., Ltd. Both cell lines were authenticated by the supplier and routinely tested for mycoplasma contamination. BV-2 cells and HEK-293T cells were cultured in complete DMEM (11965092, Gibco) with 10% fetal bovine serum (FBS, F8318, Merck) and 1% penicillin–streptomycin (15140122, Gibco). The cells were incubated at 37℃ in an atmosphere of 5% CO₂. All experiments were performed using cells between passages 5 and 15 after thawing. Primary microglia were isolated from neonatal (P1–P2) male C57BL/6J mice. Pups were euthanized by decapitation under institutional guidelines. Cortices were dissected in ice-cold HBSS (24020117, Gibco), meninges removed, and tissues enzymatically digested (0.25% trypsin-EDTA, 100 U/mL DNase I, 15 min, 37 °C, #R001100 and #89836, Thermo Scientific). After quenching with DMEM/F12(11330057, Gibco) containing 10% FBS, cells were dissociated, filtered, and seeded into T25 flasks pre-coated with poly-D-lysine (one pup per flask). Mixed glial cultures were maintained for 14 days with medium changes every 3–4 days. Microglia were collected by shaking on an orbital shaker (180 rpm, 2 h, 37 °C), replated on coated plates, and cultured in complete medium. The isolated primary microglia were allowed to recover/adapt for 24 h before stimulation. Purity (> 90%) was confirmed by Iba1/CD11b co-immunostaining.

For glucose modulation, low- or high-glucose DMEM (11885084; 11965092, Gibco) was supplemented with a 50% D-glucose stock solution (prepared from D-glucose powder, HY-B0389, MCE, dissolved in PBS) to achieve the desired final concentrations.

Cell transfection

Mouse siNC, siPrmt6, and control plasmids (sequences listed in Table 2), as well as Prmt6 and p53 overexpression plasmids (pcDNA3.1 backbone), were constructed and validated by OBIO Biotechnology (Shanghai, China). siRNA was dissolved in TE buffer to 20 µM. BV-2 or HEK-293T cells were seeded into 6-well plates (1 × 10⁵ cells/well for siRNA; 1.5 × 10⁵ cells/well for plasmids). The next day, for each well, siRNA (final concentration 80 nM) or plasmids (1 µg) were diluted in 200 µL jetPRIME buffer, mixed with 2–3 µL jetPRIME reagent, and incubated for 10–15 min at room temperature before being added dropwise to cells. After 4–8 h, the medium was replaced with DMEM supplemented with 10% FBS. Cells were harvested 24 h after transfection for expression analysis or subjected to high-glucose stimulation.

Table 2.

Prmt6 siRNA sequences

siRNA	Forward (5’-3’)	Reverse (5’-3’)
siPrmt6-689	GAUAUGAGCUGCAUGGAGATT	UCUCCAUGCAGCUCAUAUCTT
siPrmt6-1111	CUACAAAGUGGGAGACCAUTT	AUGGUCUCCCACUUUGUAGTT
siPrmt6-249	GCUUAGGCAUCCUGAAGAATT	UUCUUCAGGAUGCCUAAGCTT
Negative control	UUCUCCGAACGUGUCACGUTT	ACGUGACACGUUCGGAGAATT
FAM Negative control	UUCUCCGAACGUGUCACGUTT	ACGUGACACGUUCGGAGAATT
Positive control (GAPDH)	UUGAUGACAAGCUUCCCAUUCUTT	AGAAUGGGAAGCUUGUCAUCAATT

Cell proliferation and cycle assays

CCK-8 assay

BV-2 cells were seeded into 96-well plates (4000 cells/well) and cultured for 24 h. We replaced the original medium with medium containing the specified concentrations of glucose or EPZ020411 and incubated for the indicated time. Cells were then incubated with 100 µL of medium containing 10% CCK-8 solution (A311, Vazyme) for 1 h, and absorbance at 450 nm was measured using a microplate reader.

EdU incorporation assay

Cells were incubated with 1× EdU (10 µM) for 2 h, harvested by trypsinization, and fixed/permeabilized using a commercial kit (C0071, Beyotime, China). Click-iT reaction was performed with azide-488 for 30 min. After washing, cells were resuspended in PBS and analyzed by flow cytometry to determine the percentage of EdU-positive cells.

Cell cycle analysis

Cells were fixed, permeabilized, and stained with PI solution (containing RNase A, CCS012, MultiSciences, China) for 30 min. Cell cycle distribution was analyzed by flow cytometry using the PE channel, and data were processed with FlowJo (v10.8.1) to quantify the proportions of cells in each cell-cycle phase.

RNA-seq analysis

Trizol reagent was used to extract total RNA from the tissues of the dorsal horns of the mouse spinal cord, and RNA integrity was confirmed. mRNA was purified with oligo(dT) beads, fragmented, and reverse-transcribed into cDNA for library creation utilizing Illumina adapters. Libraries were measured, size-verified, and sequenced using the Illumina NovaSeq X Plus platform (PE150). Raw reads underwent quality correction via FastP, were aligned to the reference genome utilizing HISAT2, and transcript abundance (FPKM) was determined using StringTie. Differential expression analysis was conducted using DESeq2 (|log2FC| > 0.4, p < 0.05). Functional enrichment analysis of differentially expressed genes was performed utilizing TopGO for Gene Ontology keywords and KEGG pathways.

Co-immunoprecipitation (Co-IP)

Protein G magnetic beads (10004D, Invitrogen) were equilibrated by washing three times with PBST (PBS + 0.02% Tween 20) on a rotator. For antibody coupling, beads were incubated with 5 µg of primary antibody in 200 µL IP lysis buffer (abs9239, Absin, China) at 4 °C for 4 h with rotation. After being cleaned with PBS, the cells were lysed on ice in a cocktail IP buffer that contained phosphatase and protease inhibitors and collected using a cell scraper. Lysates were sonicated (30% amplitude, 5 s × 3) and centrifuged. An aliquot of supernatant was reserved as input. For immunoprecipitation, antibody-coupled beads were mixed with the lysates and incubated overnight at 4 °C with rotation. Beads were washed three times with PBST and resuspended in 200 µL PBS. Target proteins were eluted either with 100 µL 3×Flag peptide (for Flag-tagged proteins) at 4 °C for ≥ 2 h or by adding 20 µL 1× SDS-PAGE loading buffer and heating at 95 °C for 5 min. Eluates were collected for subsequent Western blot analysis. The following antibodies were used: PRMT6 rabbit (14641, CST, 1:50), Rabbit IgG(2729, CST), HA mouse(2367, CST, 1:50), DYKDDDDK Tag mouse(8146, CST, 1:50) for IP; p53 mouse(2524, CST, 1:1000), DYKDDDDK Tag rabbit(14793, CST, 1:1000), HA rabbit (3724, CST, 1:1000), ADMA(13522, CST, 1:1000) for IB.

Dual-luciferase reporter assay

The p21 promoter (~ 2.2 kb) was cloned into pGL3-Basic (firefly luciferase), with pRL-TK (Renilla luciferase) as internal control (Hanbio Biotechnology, Shanghai, China). HEK-293T cells were seeded in 24-well plates to 80% confluence. For each well, 0.4 µg p21 promoter plasmid, 0.4 µg Prmt6 and/or p53 plasmid, and 0.2 µg pRL-TK were mixed with 2 µL jetPRIME reagent in 200 µL DMEM, incubated 20 min at room temperature, and added to cells. After 48 h, cells were lysed with 500 µL Passive Lysis Buffer, centrifuged, and 40 µL supernatant was transferred to a 96-well plate. Firefly and Renilla luciferase activities were measured sequentially using LAR II and Stop & Glo^® reagents.

Statistical analysis

All data are presented as mean ± standard deviation (SD). Image analyses were performed using ImageJ (version 1.54p, NIH, USA). Statistical analyses and graph generation were performed using GraphPad Prism 9 (GraphPad Software, LLC, USA). An unpaired two-tailed Student’s t-test was used to compare two groups with normal distribution and equal variance; otherwise, the nonparametric Mann–Whitney U test was applied. For ≥ 3 groups, ANOVA with Tukey’s post hoc test was applied if assumptions were met; otherwise, the Kruskal–Wallis test was used. Repeated-measures data, such as pain thresholds over time, were analyzed by two-way repeated-measures ANOVA with Tukey’s post hoc test. Statistical significance was defined as P < 0.05. In all figures, significance is indicated as follows: P < 0.05 (*), P < 0.01 (**), P < 0.001 (***), P < 0.0001 (****) and not significant (ns).

Results

PRMT6 is upregulated in the spinal dorsal horn of male mice with DNP

To investigate the role of PRMT6 in DNP, a type 2 diabetes model was established using high-fat diet combined with low-dose streptozotocin (STZ) injections [40]. Body weight, blood glucose, and pain behaviors were monitored on days 0, 3, 7, 14, 21, and 28 post-STZ administration (Fig. 1A). DNP mice exhibited sustained hyperglycemia and mild weight loss (Fig. 1B and Supplementary Fig. 1A). Behavioral tests revealed thermal hyperalgesia on days 7, 14, and 21, as well as tactile allodynia and mechanical hypersensitivity from day 7 to 28, with the most pronounced hypersensitivity observed on day 21 (Fig. 1C-E).

Fig. 1.

Upregulation of PRMT6 in spinal dorsal horn microglia during diabetic neuropathic pain in male mice. (A) Schematic timeline of type 2 diabetic neuropathic pain model establishment and experimental procedures. (B) Blood glucose levels in wild-type mice before (day 0) and at 3, 7, 14, 21, and 28 days after STZ injection, (C–E) Thermal withdrawal latency (C), response frequency to 0.07 g von Frey filament (D), and response frequency to 0.4 g von Frey filament (E) in wild-type mice (n = 8 mice per group, P values from two-way ANOVA with Tukey’s multiple comparisons at each time point). (F) Western blot analysis of PRMT6 and Iba1 expression in spinal dorsal horn tissues at baseline and at 7, 14, and 21 days post-modeling, with β-actin as loading control. (G) Relative mRNA expression of PRMT6 in spinal dorsal horn of control and DNP groups on days 7, 14, and 21 by qPCR (n = 3 mice per group, P values from Student’s t-test). (H) Immunohistochemical staining of PRMT6 in spinal dorsal horn of control and DNP groups on day 21. Scale bar = 50 μm. (I) Immunofluorescence co-staining of Iba1 (green), PRMT6 (red), and DAPI (blue) in spinal dorsal horn on day 21. Scale bar = 100 μm; inset scale bar = 50 μm. The bottom row shows representative magnified microglial images. (J, K) Quantification of relative fluorescence intensity of PRMT6 in Iba1+ or NeuN+ cells (n = 3 mice per group; 10 randomly selected dorsal horn cells per mouse were quantified and normalized to the control group, P values from Student’s t-test). Data are presented as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001; ns, not significant

Western blot analyses showed upregulation of Iba1 protein in the spinal cord dorsal horn on days 7, 14, and 21 in the DNP group (Fig. 1F) with statistical significance (Supplementary Fig. 5A). PRMT6 mRNA and protein levels were also significantly elevated on days 7, 14 and 21 (Fig. 1F, G and Supplementary Fig. 1B), as confirmed statistically (Supplementary Fig. 5B). Immunohistochemistry on day 21 indicated predominant nuclear localization of PRMT6, with increased expression in DNP mice (Fig. 1H). Immunofluorescence co-staining showed that PRMT6 co-localized with the neuronal marker NeuN and the microglial marker Iba1, but not with the astrocytic marker GFAP (Fig. 1I and Supplementary Fig. 2A, B). Quantitative analysis confirmed that PRMT6 immunoreactivity was significantly increased in microglia, while remaining largely unchanged in neurons in the DNP group (Fig. 1J, K), suggesting that microglia may be more closely associated with PRMT6 upregulation in the spinal dorsal horn under diabetic conditions in male mice.

PRMT6 deficiency attenuates DNP and reduces microgliosis and neuroinflammation in male mice

To further illustrate the function of PRMT6 in DNP, we generated Prmt6-deficient male mice (Prmt6^−/−) using CRISPR/Cas9 technology (genotyping verification in Supplementary Fig. 1C), with littermate Prmt6^+/− male mice serving as controls. HFD induced significant weight gain in both groups, although Prmt6^−/− mice exhibited lower overall body weight (Supplementary Fig. 1D, E). No significant differences in blood glucose levels were observed between genotypes at this stage (Supplementary Fig. 1F). After STZ injection, both groups showed mild weight loss and a pronounced increase in blood glucose (≥ 11.1 mmol/L) (Supplementary Fig. 1G, H). Both genotypes developed diabetic neuropathic pain following high-fat diet (HFD) feeding and streptozotocin (STZ) administration. Behavioral tests revealed that Prmt6^−/− mice displayed significantly attenuated thermal hyperalgesia, tactile allodynia, and mechanical hyperalgesia compared with Prmt6^+/− controls, especially on day 21 post-modeling (Fig. 2A-C).

Fig. 2.

PRMT6 deficiency alleviates DNP, attenuates spinal microgliosis and reduces neuroinflammation in male mice. (A–C) Thermal withdrawal latency (A), response frequency to 0.07 g von Frey filament (B), and response frequency to 0.4 g von Frey filament (C) in Prmt6^+/− and Prmt6^−/− mice before and on days 3, 7, 14, and 21 after STZ injection. n = 8 mice per group, P values from two-way repeated-measures ANOVA with Tukey’s multiple comparisons at each time point, Prmt6^+/− Con vs. Prmt6^+/− DNP: P < 0.01 (**), and P < 0.001 (***), P < 0.0001 (****); Prmt6^+/− DNP vs. Prmt6^−/− DNP: P < 0.05 (#), and P < 0.01 (##), P < 0.001 (###). (D) Western blot analysis of PRMT6, NLRP3, p-p65/p65, and Iba1 in spinal dorsal horn of control and DNP groups on day 21 post-modeling, with β-actin as loading control. (E, F) Relative mRNA levels of IL-1β and TNF-α in spinal dorsal horn (n = 3 mice per group, P values from two-way ANOVA with Tukey’s multiple comparisons). (G) Immunofluorescence co-staining of Iba1 (red), and DAPI (blue) in the lumbar spinal cord of Prmt6^+/− and Prmt6^−/− mice from control and DNP groups on day 21 post-modeling. Top: transverse spinal section; bottom: magnified dorsal horn. Scale bars = 100 μm. (H) Quantification of microglial density (n = 5 mice per group, P values from two-way ANOVA with Tukey’s multiple comparisons). (I) Immunofluorescence co-staining of Iba1 (red), CD68 (green), and DAPI (blue) in the spinal cord. Yellow indicates Iba1⁺/CD68⁺ cells. Top: transverse spinal section; bottom: magnified dorsal horn; inset: single microglial cell. Scale bars = 200 μm (top), 100 μm (bottom). (J) Quantification of relative fluorescence intensity of microglial CD68 (n = 8 mice/group; 10 randomly selected dorsal horn microglia/mouse were quantified and averaged per mouse; P values from two-way ANOVA with Tukey’s multiple comparisons). (K) TEM of spinal dorsal horn. Yellow arrows indicate mitochondria. Scale bars = 1 μm (main), 500 nm (inset). (L) TEM of spinal dorsal horn. Blue arrows indicate myelin sheaths. Scale bars = 5 μm (main), 2.5 μm (inset). Data are presented as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001; ns, not significant

Western blot analysis revealed increased Iba1 expression in spinal dorsal horn of Prmt6^+/− mice, along with elevated levels of NLRP3 and phosphorylated p65 on day 21 (Fig. 2D), with statistical significance (Supplementary Fig. 5C-F). Consistently, qPCR analysis showed a significant upregulation of proinflammatory cytokines IL-1β and TNF-α. In contrast, these changes were markedly attenuated in Prmt6^−/− mice (Fig. 2D-F and Supplementary Fig. 5C-F). Immunofluorescence staining indicated increased density and hypertrophy of Iba1⁺ microglia in the dorsal horn of Prmt6^+/− DNP mice, whereas Prmt6^−/− DNP mice showed reduced microglial numbers with a more ramified morphology (Fig. 2G, H). Co-staining with the microglial activation marker CD68 [41] confirmed attenuated microglial activation in Prmt6^−/− mice (Fig. 2I, J). Moreover, transmission electron microscopy demonstrated pronounced ultrastructural abnormalities in Prmt6^+/− DNP mice, including mitochondrial swelling and myelin vacuolation, which were markedly alleviated in Prmt6^−/− DNP mice (Fig. 2K, L).

Collectively, these findings indicate that PRMT6 deficiency mitigates DNP in male mice and is associated with reduced microgliosis, suppressed neuroinflammatory signaling, and preservation of neuronal ultrastructure.

High glucose promotes cell proliferation, inflammatory responses, and PRMT6 expression in BV-2 cells

To establish optimal conditions for in vitro studies, we first examined the effects of glucose concentration and exposure time on BV-2 cell viability using the CCK-8 assay. Cell viability peaked at 25 mM glucose but declined sharply at concentrations ≥ 75 mM, indicating cytotoxicity (Fig. 3A). Time-course analysis showed that cytotoxic effects of 75 mM glucose appeared at 24 h and persisted through 72 h, while cells were most sensitive to glucose elevation at 48 h, with reduced viability observed at 40 mM (Fig. 3B). Based on these findings, 75 mM glucose was selected for subsequent experiments, and proliferation assays were performed at 48 h. EdU incorporation assays revealed a concentration-dependent increase in microglial proliferation (Fig. 3C), in contrast to the biphasic response detected by CCK-8.

Fig. 3.

High glucose promotes BV-2 cell proliferation and upregulates the expression of proinflammatory cytokines and PRMT6. (A) Optical density (OD450) of BV-2 cells after 24 h treatment with glucose at 5–150 mM (n = 3 biological replicates). (B) OD450 of BV-2 cells cultured with high glucose (25–150 mM) for 24, 48, and 72 h, measured by CCK-8 assay (n = 3 biological replicates, P values from one-way ANOVA with Fisher’s LSD test). (C) Percentage of EdU-positive BV-2 cells after 48 h culture with glucose at 5–200 mM (n = 3). (D–F) Relative mRNA levels of IL-1β, TNF-α, and PRMT6 in BV-2 cells treated with 25, 50, or 75 mM glucose for 24 h, determined by qPCR (n = 4 biological replicates). (G, H) Western blot analysis (left) and quantification (right) of PRMT6 protein expression in BV-2 cells treated with 25, 50, or 75 mM glucose for 24 h; β-actin served as a loading control (n = 3 biological replicates). Data are presented as mean ± SD, P values from one-way ANOVA with Dunnett’s multiple comparisons, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001; ns, not significant

To assess microglial activation under high-glucose stimulation, BV-2 cells were exposed to 50 mM or 75 mM glucose for 24 h, and qPCR analysis demonstrated upregulated expression of proinflammatory cytokines IL-1β and TNF-α (Fig. 3D, E). Moreover, both mRNA and protein levels of PRMT6 increased in parallel with glucose concentration (Fig. 3F-H), consistent with our in vivo observations.

PRMT6 knockdown induces cell-cycle arrest and inhibits proliferation and inflammation in BV-2 cells

To validate the functional role of PRMT6 in microglia, BV-2 cells were transfected with siRNA targeting Prmt6. The most efficient siRNA sequence achieved > 70% knockdown, as confirmed by qPCR and Western blotting (Supplementary Fig. 3A-C, E). EdU incorporation assays revealed that high glucose accelerated proliferation in both control (si-NC) and knockdown (si-Prmt6) cells, but the proliferative response was markedly reduced in si-Prmt6 cells compared with si-NC cells (Fig. 4A, B). Flow cytometric analysis following PI staining was performed to assess cell-cycle distribution. High glucose treatment decreased the proportion of G0/G1-phase cells while increasing S–M-phase cells in both groups. However, compared with si-NC cells, si-Prmt6 cells exhibited a greater proportion of cells retained in G0/G1 (Fig. 4C-E), indicating that PRMT6 knockdown induces G0/G1 arrest and slows cell-cycle progression under high-glucose conditions. To assess inflammatory activation, expression of NLRP3, phosphorylated NF-κB, IL-1β, and TNF-α was examined under normal (25 mM) and high-glucose (75 mM) conditions. si-Prmt6 cells displayed significantly reduced NLRP3 and p-NF-κB protein levels and decreased mRNA expression of IL-1β and TNF-α compared with si-NC cells under high-glucose stimulation (Fig. 4F-H and Supplementary Fig. 5G-I). These results demonstrate that PRMT6 knockdown induces G0/G1 arrest, reduces microglial proliferation, and attenuates inflammatory responses under high-glucose conditions.

Fig. 4.

PRMT6 knockdown attenuates high glucose-induced BV-2 cell proliferation and inflammatory responses. (A, B) Representative flow cytometry histograms (left) and quantification (right) of EdU incorporation in BV-2 cells transfected with non-targeting control siRNA (si-NC) or PRMT6-targeting siRNA (si-Prmt6) and exposed to 25mM or 75mM glucose for 48 h (n = 3 biological replicates). (C–E) Cell cycle profiles assessed by PI staining and flow cytometry under the same treatment conditions. (C) Histogram overlays indicating DNA content: purple – G0/G1 phase; yellow – S phase; green – G2/M phase. (D) Proportion of cells in each cell cycle phase. (E) Percentage of cells in G0/G1 phase (n = 3 biological replicates). (F) Western blot analysis of PRMT6, NLRP3, and phospho-p65 (p-p65) / total p65 in transfected BV-2 cells under 25mM or 75mM glucose for 24 h; β-actin was used as a loading control (n = 3 biological replicates). (G, H) Relative mRNA expression of IL-1β and TNF-α in transfected BV-2 cells treated with 25mM or 75mM glucose for 24 h (n = 3 biological replicates). Data are presented as mean ± SD. P values from two-way ANOVA with Tukey’s multiple comparisons, **P < 0.01, ***P < 0.001, ****P < 0.0001; ns, not significant

PRMT6 overexpression promotes cell cycle and enhances proliferation and inflammation in BV-2 cells

BV-2 cells were transfected with a Prmt6-overexpression plasmid, and successful overexpression was confirmed by qPCR (Supplementary Fig. 3D). EdU assays showed that while high glucose (75 mM) increased proliferation in both groups, the acceleration was more pronounced in Prmt6-overexpressing cells (OE-Prmt6) compared with mock controls (Fig. 5A, B). Flow cytometric analysis of PI-stained cells after 48 h of high-glucose exposure revealed a significant decrease in G0/G1-phase cells and an increase in S–M-phase cells in both groups, consistent with the knockdown results. This shift, however, was more pronounced in the OE-Prmt6 group (Fig. 5C-E). Under high-glucose conditions, PRMT6 overexpression enhanced inflammatory activation, as evidenced by increased NLRP3 expression and NF-κB phosphorylation (Fig. 5F and Supplementary Fig. 5J-L) and elevated IL-1β and TNF-α mRNA levels (Fig. 5G, H) compared with controls. These results demonstrate that PRMT6 overexpression accelerates cell-cycle progression, enhances microglial proliferation, and amplifies inflammatory activation under high-glucose conditions.

Fig. 5.

PRMT6 overexpression exacerbates high glucose-induced BV-2 cell proliferation and inflammatory responses. (A, B) Representative flow cytometry histograms (left) and quantification (right) of EdU incorporation in BV-2 cells transfected with mock or PRMT6-overexpressing (OE-Prmt6) plasmid and cultured under 25mM or 75mM glucose for 48 h (n = 3 biological replicates). (C–E) Cell cycle distribution analyzed by PI staining and flow cytometry. (C) Histograms showing DNA content: purple – G0/G1 phase; yellow – S phase; green – G2/M phase. (D) Proportion of cells in each cell cycle phase. (E) Percentage of cells in G0/G1 phase (n = 3 biological replicates). (F) Western blot analysis of PRMT6, NLRP3, and phospho-p65 (p-p65) / total p65 in transfected BV-2 cells under 25mM or 75mM glucose for 24 h; β-actin served as loading control (n = 3 biological replicates). (G, H) Relative mRNA levels of IL-1β and TNF-α in transfected BV-2 cells treated with 25mM or 75mM glucose for 24 h (n = 3 biological replicates). Data are presented as mean ± SD. P values from two-way ANOVA with Tukey’s multiple comparisons, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001; ns, not significant

Pharmacological inhibition and genetic deletion of PRMT6 suppress microglial proliferation and inflammatory responses in vitro

To determine whether the effects of PRMT6 on microglial activation and proliferation depend on its enzymatic activity, BV-2 cells were treated with the selective inhibitor EPZ020411. CCK-8 assays showed that EPZ020411 reduced cell viability in a dose-dependent manner, with significant effects observed at 10 µM, which was selected for subsequent experiments (Fig. 6A). Pretreatment with EPZ020411 for 2 h followed by 75 mM glucose stimulation significantly suppressed microglial proliferation as assessed by EdU incorporation (Fig. 6B, C), and qPCR analysis revealed reduced expression of IL-1β and TNF-α under high-glucose conditions (Fig. 6D, E).

Fig. 6.

PRMT6 inhibition or deficiency suppresses microglial proliferation and inflammation in vitro. (A) OD450 of BV-2 cells treated with EPZ020411 (0–100 µM) for 48 h (n = 3 biological replicates). (B, C) EdU incorporation in BV-2 cells pretreated with DMSO or EPZ020411 (10 µM) under 25 mM or 75 mM glucose for 48 h: representative histograms (B) and quantitation of EdU⁺ cells (C) (n = 3 biological replicates). (D, E) IL-1β (D) and TNF-α (E) mRNA levels in BV-2 cells under 25mM or 75mM glucose with DMSO or EPZ020411 for 24 h (n = 3 biological replicates). (F–I) Relative mRNA expression of Prmt6 (F), Ki67 (G), IL-1β (H), and TNF-α (I) in primary microglia from Prmt6^+/− and Prmt6^−/− male mice under 25 mM or 50 mM glucose for 24 h (n = 3 biological replicates). Data are presented as mean ± SD. P values from two-way ANOVA with Tukey’s multiple comparisons, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001; ns, not significant

To further confirm these findings in primary microglia, cells were isolated from Prmt6^+/− and Prmt6^−/− neonatal male mice and validated for purity (> 90%) by CD11b/Iba1 co-staining (Supplementary Fig. 3F, G). Upon 24 h stimulation with 50 mM glucose [42], Prmt6^+/− microglia showed increased Prmt6 mRNA expression along with elevated IL-1β and TNF-α levels. In contrast, Prmt6^−/− microglia displayed no induction of Prmt6, reduced expression of the proliferation marker Ki-67 [43], and lower proinflammatory cytokine production, as determined by qPCR (Fig. 6F-I).

Together, these results indicate that both pharmacological inhibition and genetic deletion of PRMT6 suppress high glucose-induced microglial proliferation and proinflammatory cytokine expression in vitro.

PRMT6 regulates the p53–p21 axis to promote microglial proliferation under high glucose conditions

Transcriptomic analysis of spinal dorsal horn tissue was performed as an initial unbiased screening approach to identify biological processes and pathways associated with PRMT6-mediated regulation of DNP. RNA-seq of spinal dorsal horn tissue from male Prmt6^+/− and Prmt6^−/− mice on day 21 post-modeling revealed 154 upregulated and 119 downregulated genes. Notably, Prmt6 was among the most significantly downregulated transcripts, confirming knockout efficiency. Principal component analysis (PCA) plot, volcano plot of differentially expressed genes, and heatmap of the top differentially expressed genes are shown in Fig. 7A–C. GO enrichment highlighted processes related to cell population proliferation and regulation of cell population proliferation (Fig. 7D), while KEGG analysis identified enrichment in p53 signaling, immune pathways, and cancer-related processes (Fig. 7E). Heatmaps showed marked upregulation of Cdkn1a (p21), a canonical downstream target of p53 (Fig. 7C).

Fig. 7.

PRMT6 represses p21 mRNA transcription by methylating p53. (A–E) Transcriptomic profiling of spinal cords from Prmt6^+/− and Prmt6^−/− DNP mice on day 21 after modeling (n = 3 mice per group): (A) Principal component analysis (PCA). (B) Volcano plot of differentially expressed genes (|log2FC| > 0.4, p < 0.05). (C) Heatmap of top differentially expressed genes. (D) Significantly enriched GO Biological Processes. (E) Significantly enriched KEGG pathways. (F–H) Relative mRNA expression of p53, p21 and Ki67 in BV-2 cells transfected with si-NC or si-Prmt6 and treated with 25 mM or 75 mM glucose for 24 h (n = 3 biological replicates, P values from two-way ANOVA with Tukey’s multiple comparisons). (I) Western blots of PRMT6, p53, p21, phospho-Rb (Ser807/811), and total Rb in transfected BV-2 cells under 25 mM or 75 mM glucose for 24 h (β-actin loading control). (J, K) ADMA immunoblot of spinal dorsal horn tissues from control and DNP Prmt6^+/− and Prmt6^−/− mice on day 21 and quantitation of a prominent ∼53 kDa band (β-actin loading control; n = 3 mice per group, P values from two-way ANOVA with Tukey’s multiple comparisons). (L) Dose-dependent PRMT6-Flag overexpression in HEK-293T cells modulates global ADMA modification and p21 protein levels (β-actin loading control). (M) Molecular docking between p53 (PDB:1KZY) and PRMT6 (PDB:4HC4). (N) Co-immunoprecipitation (Co-IP) in BV-2 cells using anti-PRMT6 antibody or IgG control, immunoblotted for p53 (Input: whole cell lysate). (O) Co-IP in HEK-293T cells expressing PRMT6-Flag, p53-HA, or both: Anti-FLAG immunoprecipitates were blotted for FLAG and HA; anti-HA precipitates were blotted for FLAG, HA, and ADMA (left); Input lysates were blotted for FLAG and HA (β-actin loading control; right). (P) Dual-luciferase reporter assay of Cdkn1a promoter activity. Constructs: promoter-less control; p21 promoter + empty vector; + p53; + PRMT6; + p53 and PRMT6. Data are mean ± SD of Firefly/Renilla luciferase ratio (n = 3 biological replicates; last four groups analyzed by two-way ANOVA with Tukey’s test). Data are presented as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001; ns, not significant

Because the transcriptomic analysis was performed on whole spinal dorsal horn tissue rather than purified microglia, these data do not allow microglia-specific transcriptional conclusions. Nevertheless, the enrichment of cell-proliferation-related processes and p53 signaling at the tissue level, together with the reduced spinal Iba1-positive microglial density observed in Prmt6^−/− mice and the effects of PRMT6 on BV-2 cell proliferation and cell-cycle progression in vitro, prompted us to examine whether PRMT6 regulates microglial cell cycle progression via the p53–p21–Rb axis, a central regulator of the G1/S checkpoint. In this pathway, p53 transactivates the cyclin-dependent kinase inhibitor p21, which in turn inhibits cyclin-CDK complexes, prevents Rb phosphorylation, and thereby blocks E2F-dependent transcription and S-phase entry [36]. Using siRNA-mediated knockdown of PRMT6 in BV-2 cells, we observed that under high-glucose conditions, PRMT6 knockdown did not affect p53 expression but significantly upregulated p21 at both mRNA and protein levels (Fig. 7F, G,I; protein quantified in Supplementary Fig. 5M, N). This was accompanied by a reduction in Rb phosphorylation (Ser807/811) (Fig. 7I and Supplementary Fig. 5O) and a consequent decrease in the proliferation marker Ki-67 (Fig. 7H). These findings indicate that PRMT6 facilitates G1/S transition, at least in part, by suppressing p21 expression and maintaining Rb phosphorylation, whereas PRMT6 knockdown disrupts this axis and reinforces cell cycle arrest.

Additionally, Western blot analysis revealed altered asymmetric dimethylarginine (ADMA), a modification catalyzed specifically by type I PRMTs [44], across a range of molecular weights in spinal tissues of model mice, with a particularly pronounced increase at approximately 53 kDa in Prmt6^+/− male mice (Fig. 7J). In contrast, Prmt6^−/− male mice showed markedly reduced methylation signals especially at ~ 53 kDa (Fig. 7J, K). These changes indicate that PRMT6 mediates asymmetric dimethylation near 53 kDa, which may contribute to DNP. Gradient overexpression of PRMT6-Flag in HEK-293T cells enhanced ~ 53 kDa ADMA signals while reducing p21 protein levels in a dose-dependent manner (Fig. 7L).

Guided by the macroscopic transcriptomic clues and the observed changes in methylation, we hypothesized that PRMT6 regulates p21 transcription by interacting with p53. To model the interaction between PRMT6 and p53, we performed molecular docking using HDOCK. Docking between PRMT6 and p53 yielded a docking score of − 283.34 and a confidence score of 0.9350, indicating the docking is reliable and the two proteins are very likely to interact. Site prediction indicates that p53 residues Ser241, Met243, Arg248, Arg280, Asp281, Asn288 and Thr284 form hydrogen-bond interactions with PRMT6 residues Cys229, Glu235, Gln239, Arg204 and Lys319. These residues are located in p53’s DNA-binding domain and PRMT6’s catalytic core domain, respectively, suggesting that the PRMT6–p53 interaction may impair p53’s ability to bind DNA, thereby affecting its transcriptional activity (Fig. 7M).

To verify the binding between PRMT6 and p53, we first performed Co-IP in BV-2 cells, showing that PRMT6 specifically interacted with p53 (Fig. 7N). Co-expression of PRMT6-Flag and p53-HA confirmed their interaction and demonstrated that PRMT6 increased p53 methylation (Fig. 7O). Dual-luciferase reporter assays revealed that p53 alone activated the Cdkn1a promoter, whereas PRMT6 alone suppressed it; co-expression markedly reduced p53-induced activation of Cdkn1a transcription. Notably, two-way ANOVA revealed a significant interaction between PRMT6 and p53 (P = 0.0002), supporting our initial hypothesis (Fig. 7P). Collectively, these data imply that PRMT6 promotes microglial cell cycle progression and proliferation, at least in part, by methylating p53, thereby repressing its transcriptional activity and downregulating p21. These findings support a mechanism by which high glucose conditions enhance microglial proliferation and inflammatory responses.

EPZ020411 ameliorates DNP and suppresses microgliosis and neuroinflammation in male mice

To evaluate the therapeutic potential of PRMT6 inhibition, wild-type DNP male mice were administered the PRMT6 inhibitor EPZ020411 [45] (10 mg/kg, i.p.) or normal saline for three consecutive days [32] beginning on day 8 after model induction. Behavioral tests showed that EPZ020411 significantly attenuated thermal hyperalgesia, tactile allodynia, and mechanical hypersensitivity on days 14 and 21 (Fig. 8A-C). Immunofluorescence staining for Iba1 in spinal cord tissues collected on day 21 revealed that EPZ treatment markedly reduced microglial density compared with saline-treated DNP mice (Fig. 8D, E). qPCR analysis further showed that EPZ020411 treatment inhibited the expression of proinflammatory cytokines IL-1β and TNF-α (Fig. 8F, G). Transmission electron microscopy further revealed that EPZ020411 attenuated mitochondrial damage and axonal demyelination (Fig. 8H, I), indicating a protective effect against neuroinflammation-associated structural injury. These findings suggest that pharmacological inhibition of PRMT6 with EPZ020411 may represent a potential therapeutic strategy for diabetic neuropathic pain.

Fig. 8.

EPZ020411 attenuates nociceptive hypersensitivity, spinal microgliosis, and neuroinflammation in male DNP mice. (A–C) Thermal withdrawal latency, response frequency to 0.07 g von Frey filament, and response frequency to 0.4 g von Frey filament in wild-type (WT) control, DNP + normal saline (NS), and DNP + EPZ020411(EPZ) groups before and on days 3, 7, 14, and 21 after STZ injection. NS or EPZ020411 (10 mg/kg) was administered intraperitoneally on days 8–10. n = 6 mice per group, P values from two-way repeated-measures ANOVA with Tukey’s multiple comparisons at each time point, Con vs. DNP + NS: P < 0.05 (*), P < 0.01 (**), P < 0.001 (***), P < 0.0001 (****); DNP + NS vs. DNP + EPZ: P < 0.05 (#), P < 0.01 (##), P < 0.001 (###). (D) Immunofluorescence co-staining of Iba1 (red), and DAPI (blue) in the lumbar spinal cord on day 21 post-modeling. Scale bars = 100 μm. (E) Quantification of microglial density (n = 3 mice per group, P values from one-way ANOVA with Tukey’s multiple comparisons). (F, G) Relative mRNA expression of IL-1β and TNF-α in the spinal dorsal horn on day 21 (n = 3 mice per group, P values from one-way ANOVA with Tukey’s multiple comparisons). (H, I) Representative TEM images of mitochondrial (H; yellow arrow; scale bar = 1 μm) and myelin sheath (I; blue arrow; scale bar = 5 μm) ultrastructure in the spinal dorsal horn on day 21. Data are presented as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001; ns, not significant

PRMT6 deficiency has no significant effect on pain hypersensitivity, microglial density, or neuroinflammation in female DNP mice

To determine whether PRMT6 exerts similar effects in females, we established a T2DM-associated DNP model using female Prmt6^+/+ and Prmt6^−/− littermates. Following 8 weeks of a high-fat diet and subsequent STZ administration at 50 mg/kg, both genotypes exhibited comparable increases in body weight and fasting blood glucose that met the criteria for diabetes (Fig. 9A, B). Western blot analysis showed no significant elevation in PRMT6 protein levels in the spinal dorsal horn of Prmt6^+/+ mice on day 21 post-modeling (Fig. 9C). Behavioral assessments performed on days 0, 3, 7, 14, and 21 after modeling showed that both Prmt6^+/+ and Prmt6^−/− female mice developed significant thermal, tactile, and mechanical hypersensitivity by day 14. However, unlike the protective phenotype observed in males, Prmt6 deficiency did not significantly alter pain behaviors in female DNP mice at any tested time point (Fig. 9D-F). Immunofluorescence analysis of the lumbar spinal dorsal horn on day 21 revealed no significant increase in Iba1-positive microglial density in female Prmt6^+/+ DNP mice relative to controls. In addition, Iba1-positive microglial cell counts did not differ significantly between Prmt6^−/− and Prmt6^+/+ mice in any group (Fig. 9G, H). Although microglial CD68 immunoreactivity and spinal IL-1β and TNF-α mRNA levels were increased after DNP induction, these changes were not significantly modified by Prmt6 deficiency (Fig. 9G, I-K). Together, these findings suggest that spinal microgliosis is less prominent in female DNP mice than in males, and that the protective effects of Prmt6 deficiency on spinal neuroinflammation and pain hypersensitivity observed in males were not detected in the female cohort examined here.

Fig. 9.

Limited spinal microgliosis in female DNP mice and no significant effect of Prmt6 deficiency on spinal microglial density, neuroinflammation, or nociceptive hypersensitivity. (A) Body weights of Prmt6^+/+ and Prmt6^−/− female mice in control and HFD groups before (week 0) and after 8 weeks of HFD feeding. (B) Fasting blood glucose of Prmt6^+/+ and Prmt6^−/− female mice in control and DNP groups before (day 0) and after 3 days of STZ injection (A-B, P values from two-way ANOVA with Tukey’s multiple comparisons at each time point). (C) Western blot analysis of PRMT6 protein levels in the spinal dorsal horn of Prmt6^+/+ mice from the control and DNP groups on day 21 post-modeling (β-actin loading control, n = 3 mice per group, P values from Student’s t-test). (D–F) Thermal withdrawal latency and response frequencies to 0.07 g and 0.4 g von Frey filaments in Prmt6^+/+ and Prmt6^−/− female mice at the indicated time points after STZ injection (n = 8 mice per group, P values from two-way repeated-measures ANOVA with Tukey’s multiple comparisons at each time point). (G) Immunofluorescence co-staining of Iba1 (green), CD68 (red) and DAPI (blue) in the lumbar spinal cord dorsal horn of Prmt6^+/+ and Prmt6^−/− female mice on day 21 post-modeling. Scale bars = 100 μm; inset scale bar = 50 μm. (H) Quantification of microglial density (n = 3 mice per group). (I) Quantification of relative fluorescence intensity of microglial CD68 (n = 3 mice/group; 15 randomly selected dorsal horn microglia/mouse were quantified and averaged per mouse; G-H, P values from two-way ANOVA with Tukey’s multiple comparisons). (J, K) Relative mRNA expression of IL-1β and TNF-α in the spinal dorsal horn of Prmt6^+/+ and Prmt6^−/− female mice on day 21 after STZ injection (n = 3 mice per group, P values from one-way ANOVA with Tukey’s multiple comparisons). Data are presented as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001; ns, not significant

Discussion

This study supports a role for PRMT6 in spinal neuroinflammation under hyperglycemic conditions in male mice. In male DNP mice, PRMT6 expression was markedly upregulated in spinal microglia. Mechanistically, our in vitro findings demonstrate that elevated PRMT6 levels promote BV-2 cell proliferation and inflammatory responses by methylating p53, which in turn suppresses p21 transcription and accelerates the G0/G1–S phase transition. Furthermore, in vivo genetic ablation or pharmacological inhibition of PRMT6 effectively alleviated pain hypersensitivity in male mice, accompanied by attenuated microgliosis and suppressed spinal inflammatory signaling. To our knowledge, these findings support the involvement of the PRMT6–p53–p21 axis in microgliosis-associated spinal neuroinflammation during DNP in male mice and suggest that hyperglycemia may promote neuroimmune dysregulation through this pathway.

We selected a type 2 diabetes model because the prevalence of painful diabetic peripheral neuropathy is consistently higher in type 2 than in type 1 diabetes [2, 46], and neuropathic pain in type 2 diabetes often exhibits limited improvement with glycemic control [3]. To better recapitulate this clinical scenario and its therapeutic challenges, we employed a high-fat diet combined with low-dose streptozotocin to induce insulin resistance and partial β-cell dysfunction characteristic of type 2 diabetes.

Although prior mechanistic studies of DNP have largely focused on canonical inflammatory cascades—including pERK [10], pSTAT3, pAKT [47] and nuclear factor kappa B (NF-κB) signaling [48]—neuroinflammation does not arise in isolation. An underexplored yet biologically consequential facet is the aberrant re-entry of quiescent microglia into the cell cycle, which drives a proliferative state that sustains chronic immune activation. Emerging evidence from other neuropathologies indicates that dysregulated cell-cycle progression serves not only as a trigger for proliferation but also as a key determinant of inflammatory severity. Indeed, arresting cell-cycle progression via HDAC3 deficiency [49] or Cyclin D1 inhibition [34] has been shown to effectively dampen neuroinflammation, highlighting an untapped therapeutic avenue for DNP.

PRMT6 has been reported to promote proliferation in tumor cells [26, 50–52], and our in vivo and in vitro findings suggest that it may play a similar role in microglia. Prmt6 deletion was associated with a reduced abundance of Iba1-positive microglia in the spinal dorsal horn of diabetic mice. To further explore the underlying mechanisms, we performed transcriptomic profiling of Prmt6^+/− and Prmt6^−/− mice with diabetic neuropathic pain. GO analysis indicated enrichment of proliferation-related processes, while KEGG enrichment highlighted the p53 signaling pathway, with notable changes in its downstream effector p21. p53, the “guardian of the cell,” orchestrates stress responses by controlling cell growth, apoptosis, and senescence, primarily through cell-cycle arrest [35]. p21, a critical downstream target of p53, blocks the G1/S transition [53] by inhibiting CDK4, CDK6 and CDK2 activity [54], reducing Rb phosphorylation, and stabilizing the Rb–E2F complex [36]. This p53–p21 axis thus acts as a checkpoint to allow cellular repair under oxidative stress, DNA damage, or inflammation [55, 56]. Our cellular findings, together with tissue-level transcriptomic data, suggest that PRMT6 may be involved in regulating microglial proliferation and functional activation, potentially through the p53–p21 pathway in a methyltransferase-dependent manner.

PRMT6 has previously been reported to negatively regulate p53 transcription in mouse embryonic fibroblasts by catalyzing histone H3 arginine 2 methylation, thereby reducing p53 protein levels [57]. However, in our study of microglia, we observed no change in p53 protein abundance. Instead, our data suggest that PRMT6 modulates the post-translational methylation of p53, which in turn alters p21 transcription and promotes cell-cycle progression. Although PRMT5, a type II PRMT, has been shown to methylate p53 and modify its transcriptional selectivity [58], our findings provide the first evidence that the type I PRMT family member PRMT6 may exert a similar regulatory effect. We therefore speculate that PRMT6 regulation of p53 is both cell-type- and context-dependent. Our cellular data suggest that, under chronic diabetic-like metabolic conditions, PRMT6 fine-tunes p53 function in microglia primarily through post-translational modification rather than by altering its protein abundance, thereby affecting microglial cell-cycle dynamics and functional state.

PRMT6 deficiency, we observed, does not simply restrain microglial proliferation or suppress inflammation in isolation; rather, it exerts a synchronized dual inhibitory effect. Previous work by Rackov et al. demonstrated p21 as a molecular switch that transitions macrophages from a proinflammatory to a hyporesponsive state by curbing NF-κB signaling [59]. Consistent with this literature, our data place PRMT6 upstream of the p53–p21 axis and indicate that PRMT6 activity is associated with concurrent cell-cycle dysregulation and activation of NF-κB and NLRP3 signaling in microglia.

Identifying PRMT6 as a potential contributor to microgliosis and neuroinflammation highlights it as a candidate therapeutic target for DNP. Current standard-of-care pharmacotherapies, including gabapentinoids, tricyclic antidepressants, and serotonin/norepinephrine reuptake inhibitors, often have limited efficacy and significant adverse effects [60]. Our data demonstrated that the small‑molecule PRMT6 inhibitor EPZ020411 potently suppressed neuroinflammation and reversed pain hypersensitivity in a mouse DNP model. Given that several type I PRMT inhibitors have advanced into clinical evaluation and selective PRMT6 inhibitors are increasingly being developed [61], PRMT6‑targeting agents may represent a promising therapeutic approach for microgliosis-associated conditions such as diabetic neuropathic pain.

An additional finding of this study is that the in vivo effects of PRMT6 in DNP appear to be sex dependent. In male mice, DNP was associated with marked spinal microgliosis, and both genetic deletion and pharmacological inhibition of PRMT6 attenuated microglial expansion, inflammatory signaling, and pain hypersensitivity. In contrast, female DNP mice showed little change in spinal Iba1-positive microglial density, PRMT6 expression remained unchanged, and Prmt6 deficiency did not significantly alter spinal microglial markers or nociceptive behaviors. These findings are broadly consistent with previous studies suggesting sexual dimorphism in neuropathic pain mechanisms [62–65]. In particular, Sorge et al. reported that pain hypersensitivity is predominantly mediated by microglia in males, whereas females may engage alternative pathways, including T cell-related signaling [66]. The observed sex differences in microglial responses to metabolic stress might also be influenced by sex hormones [67] or differential regulation of signaling pathways, such as CX3CR1 signaling [68]. However, whether microglia contribute differently to DNP in female mice remains to be further determined, for instance by using approaches such as microglial depletion. Moreover, given that PRMT6 can act as a direct transcriptional coregulator for sex hormone receptors [69, 70], the sex-dependent effects observed here may reflect its participation in distinct signaling networks in males and females. Our findings therefore suggest that PRMT6 primarily participates in a male-biased, spinal microgliosis-associated mechanism in this DNP model. However, this does not indicate that PRMT6 is irrelevant in females; rather, it suggests that Prmt6 deficiency does not substantially affect the spinal microglial and behavioral endpoints examined here. Whether PRMT6 contributes to female DNP through other spinal cell populations, peripheral immune pathways, or distinct metabolic and neuroimmune mechanisms will require further study.

Several limitations should be acknowledged. First, the in vivo experiments were performed using a global Prmt6 knockout model and systemic pharmacological inhibition; therefore, we cannot definitively attribute the observed in vivo phenotypes specifically to PRMT6 loss in microglia. It remains possible that PRMT6 deficiency in neurons or other cell types indirectly modulates microglial responses through established intercellular signaling pathways. Moreover, based on our breeding strategy, male Prmt6^+/− littermates were used as controls for male Prmt6^−/− mice in the experiments, and this control design should be considered when interpreting the Prmt6^−/− phenotypes. Thus, although our BV-2 and primary microglial experiments support a direct regulatory role of PRMT6 in microglia, future studies using microglia-specific conditional Prmt6 deletion, together with single-cell/nucleus transcriptomic approaches, will be required to establish its cell-autonomous role in vivo. Second, BV-2 and primary culture models have inherent limitations in fully recapitulating the spinal dorsal horn microenvironment. Third, although we included a cohort of female mice, these experiments were not conducted in parallel with those in male mice, partly because different STZ doses were required for diabetes induction. The analyses were also restricted to selected spinal and behavioral endpoints and were not designed to comprehensively define the specific differences between male and female mice or the underlying cellular and molecular mechanisms. In addition, as an epigenetic modifier, PRMT6 may affect additional targets, and the therapeutic implications of PRMT6 inhibition in DNP therefore require cautious interpretation; validation in human DNP tissues and assessment of safety and pharmacokinetics of PRMT6 inhibitors in metabolic disease are warranted.

In summary, our findings indicate that diabetic/hyperglycemic conditions upregulate PRMT6 expression in the spinal dorsal horn of male mice, predominantly within microglia. Our cellular experiments further support that PRMT6 promotes microglial proliferation and activation, at least in part, by methylating p53, thereby suppressing p21 transcription and facilitating cell-cycle progression. In male mice, PRMT6 deficiency or inhibition alleviated spinal microgliosis, neuroinflammation, and pain hypersensitivity in DNP. In contrast, female DNP mice exhibited relatively limited spinal microgliosis and less pronounced PRMT6 changes, and Prmt6 deficiency did not significantly alter the spinal microglial or behavioral endpoints examined. Together, these findings implicate a PRMT6-p53-p21 axis in spinal microgliosis and neuroinflammation under diabetic conditions and suggest that PRMT6 may represent a potential therapeutic target for DNP, with possible sex-dependent effects.

Authors’ contributions

HY conceived the study and designed experiments, and supervised the whole study. CH and HY performed the research. HS and XF performed animal experiments. YC, HS and MD performed experiments in vitro. YF, RD and KH performed immunostaining, enzyme-linked immunosorbent assay and behavior test. XF, YC and LP analyzed data and drafted the manuscript. XF, LP, YY, KH and HW contributed to interpretation of data and critical revision of manuscript. All the authors have read and approved the final manuscript.

Funding

This work was supported by the National Natural Science Foundation of China (82171220, 81971046, 82471239, 32170878, 82471792) and National Key R&D Program of China (2024YFA0918400).

Data availability

The data supporting the findings of this study are available within the article and its Supplementary Information files. Further inquiries can be directed to the corresponding author.

Declarations

Ethics approval and consent to participate

The experimental protocols were reviewed and approved in advance by the Animal Care and Use Committee of Naval Medical University and authorized by the Medical Ethics Committee of Changzheng Hospital, Shanghai, China (approval no. 2025SLYS1).

Competing interests

The authors declare no competing interests.