Abstract
Transient receptor potential ankyrin 1 (TRPA1) and vanilloid 1 (TRPV1) channels are crucial for detecting and transmitting nociceptive stimuli. Inflammatory pain is associated with sustained increases in TRPA1 and TRPV1 expression in primary sensory neurons. However, the epigenetic mechanisms driving this upregulation remain unknown. G9a (encoded by Ehmt2) catalyzes H3K9me2 and generally represses gene transcription. In this study, we found that intrathecal administration of UNC0638, a specific G9a inhibitor, or G9a-specific siRNA, substantially reduced complete Freund's adjuvant (CFA)-induced pain hypersensitivity. Remarkably, CFA treatment did not induce persistent pain hypersensitivity in male and female mice with conditional Ehmt2 knock-out in dorsal root ganglion (DRG) neurons. RNA sequencing and quantitative PCR analyses showed that CFA treatment caused a sustained increase in mRNA levels of Trpa1 and Trpv1 in the DRG. Ehmt2 knock-out in DRG neurons elevated baseline Trpa1 and Trpv1 mRNA levels but notably reversed CFA-induced increases in their expression. Chromatin immunoprecipitation revealed that CFA treatment reduced G9a and H3K9me2 levels while increasing H3K9ac and H3K4me3—activating histone marks—at Trpa1 and Trpv1 promoters in the DRG. Strikingly, conditional Ehmt2 knock-out in DRG neurons not only diminished H3K9me2 but also reversed CFA-induced increases in H3K9ac and H3K4me3 at Trpa1 and Trpv1 promoters. Our findings suggest that G9a in primary sensory neurons constitutively represses Trpa1 and Trpv1 transcription under normal conditions but paradoxically enhances their transcription during tissue inflammation. This latter action accounts for inflammation-induced TRPA1 and TRPV1 upregulation in the DRG. Thus, G9a could be targeted for alleviating persistent inflammatory pain.
- bivalent histone modifications
- chromatin plasticity
- dorsal root ganglion
- epigenetics
- histone lysine methyltransferase
- histone methylation
Significance Statement
This study demonstrates for the first time that G9a, a histone methyltransferase, in sensory neurons is crucial for the persistent pain development following inflammation. Inhibiting or knockdown of G9a at the spinal cord level reduced inflammation-induced pain hypersensitivity. Remarkably, mice lacking G9a in sensory neurons failed to develop persistent pain hypersensitivity after inflammation. Ablating G9a in sensory neurons increased baseline expression of the TRPA1 and TRPV1 ion channels but reversed their upregulation induced by inflammation. Additionally, G9a deletion blocked the inflammation-driven enrichment of activating histone marks at Trpa1 and Trpv1 promoters. These findings highlight a dual role of G9a in sensory neurons: suppressing Trpa1 and Trpv1 transcription under normal conditions while promoting their transcription during inflammation through bivalent histone modifications.
Introduction
Tissue injury and inflammation stimulate and sensitize nociceptive primary sensory neurons, resulting in pain. Current treatments for persistent inflammatory pain are unsatisfactory, highlighting the need to better understand the mechanisms underlying sustained pain after tissue inflammation. Nociceptors express receptors and ion channels that relay nociceptive information from primary sensory neurons to the spinal cord and brain for pain processing. Among these, transient receptor potential (TRP) channels—particularly ankyrin 1 (TRPA1) and vanilloid 1 (TRPV1)—which are expressed in dorsal root ganglion (DRG) neurons and their peripheral and central terminals, play pivotal roles in nociceptive detection and transmission during inflammatory pain (Caterina et al., 2000; Davis et al., 2000; Obata et al., 2005; Huang et al., 2019). For instance, TRPA1 and TRPV1 expression increases in the DRG during inflammatory pain induced by complete Freund's adjuvant (CFA; Davis et al., 2000; Breese et al., 2005; Obata et al., 2005). Additionally, TRPV1 antagonists or genetic knock-out of TRPV1 reduce CFA-induced pain hypersensitivity (Davis et al., 2000; Cui et al., 2006). Similarly, genetic depletion of TRPA1 or the TRPA1 antagonist decreases pain hypersensitivity induced by CFA (Fernandes et al., 2011; Horváth et al., 2016; Huang et al., 2019). However, the mechanisms underlying sustained high expression of TRPA1 or TRPV1 in the DRG during persistent inflammatory pain remain unclear.
Epigenetic mechanisms, such as reversible modifications to histones and DNA on chromatin, control the accessibility of gene transcription regulatory elements (Jenuwein and Allis, 2001; Kouzarides, 2007). Acetylation/deacetylation or methylation/demethylation at the lysine and arginine residues of histone proteins significantly impacts chromatin architecture. These epigenetic modifications critically influence the transcription of genes involved in the development of chronic neuropathic pain (Laumet et al., 2015; Zhang et al., 2016, 2022; Garriga et al., 2018; Ghosh et al., 2022). For instance, nerve injury-induced hyperacetylation of histones H3 and H4 at the α2δ-1 gene promoter leads to sustained α2δ-1 upregulation in DRG neurons (Zhang et al., 2022). Although epigenetic mechanisms have been extensively studied in the context of chronic neuropathic pain, the epigenetic basis of persistent inflammatory pain remains poorly understood. The histone lysine methyltransferase G9a, also known as euchromatic histone lysine N-methyltransferase 2 (EHMT2, encoded by the Ehmt2 gene), and its partner G9a-like protein (GLP) catalyzes the mono- and dimethylation of histone H3 lysine 9 (H3K9me1/2), a modification typically associated with transcriptional repression (Tachibana et al., 2001, 2005). Nerve injury-induced chronic pain is associated with the downregulation of voltage-gated K+ channels, μ-opioid receptors, and cannabinoid CB1 receptors in the DRG, driven by the enrichment of G9a-mediated H3K9me2 at their gene promoters (Laumet et al., 2015; Zhang et al., 2016; Luo et al., 2020). Notably, mice with conditional Ehmt2 knock-out in DRG neurons do not develop chronic pain after traumatic nerve injury (Laumet et al., 2015). Nevertheless, G9a can also act as a transcriptional coactivator, promoting gene expression depending on its interacting partners and post-translational modifications (Purcell et al., 2011; Shankar et al., 2013; Poulard et al., 2017). The potential role of G9a in regulating the development of persistent inflammatory pain has not been investigated previously.
In this study, we investigated the potential role of G9a in primary sensory neurons in relation to persistent inflammatory pain. We found that conditional G9a knock-out in primary sensory neurons abolishes inflammation-induced persistent pain and the associated increase in Trpa1 and Trpv1 transcription. Interestingly, conditional G9a knock-out not only increases the baseline transcription of Trpa1 and Trpv1 but also attenuates the inflammation-induced enrichment of activating histone marks at Trpa1 and Trpv1 promoters in the DRG. Our findings uncover a dual role of G9a in the epigenetic regulation of Trpa1 and Trpv1 transcription and the development of persistent inflammatory pain.
Materials and Methods
Animal models of inflammatory pain
All experimental procedures were approved by the Institutional Animal Care and Use Committee at The University of Texas MD Anderson Cancer Center and followed the Guide for the Care and Use of Laboratory Animals (National Institutes of Health). Male Sprague Dawley rats (8–11 weeks old; Envigo) were used in this study. Also, both male and female mice (8–10 weeks old) were used in this study, and the genetic background of all mice was C57BL/6. Animals were housed (three rats per cage; 4–5 mice per cage) in a controlled environment at 24°C under a 12 h light/dark cycle. All animals were provided with ad libitum access to food and water. Inflammatory pain was induced by subcutaneous injection of 50 µl in rats or 20 µl in mice of complete Freund's adjuvant solution (CFA, #F5881, Sigma-Aldrich) into the plantar surface of the left hindpaw, as previously reported (Huang et al., 2019; Sun et al., 2019). Animals in the control group received vehicle (mineral oil, #M5904, Sigma-Aldrich) injection into the plantar surface of the left hindpaw.
Generation of conditional and inducible G9a knock-out mice
To ablate Ehmt2 in primary sensory neurons, we crossed male Avil-Cre mice and female Ehmt2flox/flox mice (Hasegawa et al., 2007; Lau et al., 2011; Laumet et al., 2015) to generate G9a conditional knock-out (Ehmt2-cKO). Three weeks after birth, mice were ear tagged and genotyped using ear tissues with polymerase chain reaction (PCR) using the primers listed in Table 1.
Genotyping PCR primers used in the study
For the generation of inducible conditional Ehmt2 KO (Ehmt2-iKO) mice, we crossed Ehmt2flox/flox mice with tamoxifen-dependent inducible Avil-iCreERT2/AJwo mice (stock #032027, The Jackson Laboratory). Tamoxifen (#T5648, Sigma-Aldrich) was dissolved in corn oil (#C8267, Sigma-Aldrich) and injected intraperitoneally (75 mg/kg/day for 5 consecutive days) in 8–10-week-old Avil-iCreERT2::G9aflox/flox mice. Age- and sex-matched Cre-negative Ehmt2flox/flox littermates were used as wild-type (WT) controls.
Intrathecal injections
Rats were anesthetized with 2–3% isoflurane for surgical implantation of intrathecal catheters. A small incision was made at the back of the neck, and a PE-10 catheter (8 cm) was inserted via a small opening made on the atlanto-occipital membrane of the cisterna magna so that the catheter tip reached the lumbar enlargement (Chen and Pan, 2001). The animals were allowed to recover for 5 d before intrathecal injections. UNC0638 (#HY-15273, MedChemExpress) were dissolved in 10% dimethyl sulfoxide (DMSO). UNC0638 was injected via the intrathecal catheter in a 10 µl volume followed by a 5 µl flush with saline.
Rat G9a-specific siRNA (#SASI-Rn01-00041925) and control siRNA (MISSION siRNA universal negative control #1; #SIC001) were obtained from Sigma-Aldrich. The specificity of G9a siRNA (5′-AGUAACGGGCAUCAAUGC-3′) was validated previously (Laumet et al., 2015; Zhang et al., 2022). Prior to intrathecal injection, 4 µg of G9a-specific siRNA or control-siRNA was mixed with 10 µl of i-Fect solution (#NI35150, Neuromics) for the siRNA delivery.
Assessment of nociceptive behaviors
Rats and mice were acclimated to the testing environment for 30 min before assessing baseline thresholds. Baseline nociceptive thresholds were measured over 3 consecutive days prior to CFA or vehicle injection, with the thresholds on the third day considered as the baseline. The tactile withdrawal threshold was measured by using the “up-down” method (Chaplan et al., 1994; Chen et al., 2001). We stimulated the plantar surface of the left hindpaw vertically with a series of calibrated von Frey filaments (Stoelting). The filament was applied to the plantar surface with a sufficient force and bent for 6 s. A brisk withdrawal or paw flinching was considered a positive response. When a positive response was observed, the filament of the next lower force was applied. In the absence of a response, the filament of the next greater force was applied. The tactile withdrawal threshold test was repeated twice, and the mean value was calculated.
To quantify the mechanical nociceptive threshold, we used a Pincher Analgesia Meter (model 2450, IITC Life Science) to apply increasing pressure to the midplantar glabrous surface of the hindpaw, as previously described (Jin et al., 2022; Zhang et al., 2022). A brisk withdrawal or vocalization was considered a positive response, and the reading on the scale was recorded as the withdrawal threshold. A cutoff of 400 g was set to minimize potential tissue injury.
To determine the heat sensitivity of the hindpaw, animals were tested on a Plantar Analgesia Meter (model 400, IITC Life Science). Animals were placed individually in the observation chamber on the glass surface, which was maintained constantly at 30°C. A mobile radiant heat source positioned beneath the glass was targeted at the plantar surface of the hindpaw, as we described previously (Chen and Pan, 2001, 2006). The cutoff of 30 s was set to minimize potential tissue injury. The paw withdrawal latency was recorded with a timer, and the hindpaw was tested twice to obtain a mean withdrawal latency.
RNA isolation and real-time quantitative PCR assay
Total RNA was isolated from the lumbar DRG and spinal cord tissues using RNeasy Plus Universal Mini Kit (#73404; Qiagen), as described previously (Ghosh et al., 2022). The genomic DNA was removed using quick DNase treatment, and the purified total RNA was then reverse transcribed to cDNA by using the SuperScript IV VILO Master Mix (#11766050; Invitrogen). A 100 ng of cDNA was used for the real-time qPCR using QuantStudio 7 Flex Real-Time PCR System (Applied Biosystems) with the PowerUp SYBR Green Master Mix (#A25776; Thermo Fisher Scientific). The fast-cycling mode conditions (50°C for 2 min, 95°C for 2 min, 40 cycles of 95°C for 1 s, and 60°C for 30 s) were used to perform the RT-qPCR reactions. Relative mRNA levels of the target genes were quantified using the comparative 2−ΔΔCT method, first normalized to the Gapdh mRNA level in the same sample, and then to its level in the WT or vehicle-treated group. The mean value in the WT or vehicle group was set to 1. The primers used are listed in Table 2.
qPCR primers used in the study
Immunoblotting
Total proteins were extracted from lumbar DRG tissues (L3–L5 in mice and L5–L6 in rats) using RIPA lysis and extraction buffer (#89900, Thermo Fisher Scientific) in the presence of Halt protease and phosphatase inhibitor cocktail (#78442; Thermo Fisher Scientific). The concentration of extracted total proteins was determined using the DC Protein Assay kit (Bio-Rad Laboratories) with bovine serum albumin as a reference standard. Equal amounts of proteins from all samples were subjected to reducing and denaturing gel electrophoresis on NuPAGE 4–12% Bis-Tris Mini Protein Gels (1.5 mm, #NP0335BOX, Invitrogen). Subsequently, the proteins on the gel were transferred to a 0.2 μm polyvinylidene difluoride membrane using a Trans-Blot Turbo Transfer System (Bio-Rad) for 10 min. The membrane was immediately submerged in SuperBlock T20 (Tris-buffered saline, TBS) blocking buffer (#37536, Thermo Scientific) and blocked for 10 min at 23°C under constant agitation. The membrane was then incubated overnight at 4°C with one of the following primary antibodies diluted in SuperBlock T20 (TBS): rabbit anti-G9a antibody (1:1,000, #3306, Cell Signaling Technology), rabbit anti-GAPDH antibody (1:1,000, #5174S, Cell Signaling Technology). The GAPDH was used as the protein loading control. On the next day, the membrane was washed twice with TBS containing 0.05% Tween 20 (TBST) buffer, and subsequently a goat anti-rabbit HRP-linked antibody (1:5,000, #7074S, Cell Signaling Technology) or a horse anti-mouse HRP-linked antibody (1:5,000, #7076S, Cell Signaling Technology) was applied to the immunoblots for 1 h at 23°C. After three washes with TBST, the protein bands on the membrane were visualized using the SuperSignal West Femto Maximum Sensitivity Substrate (#34094, Thermo Scientific) on the Odyssey Fc Imager (LI-COR Biosciences). The protein bands were then normalized to the corresponding GAPDH protein band on the same blot. ImageJ software (https://imagej.net/ij/) was used to quantify the intensity of the protein bands. The amount of target proteins in each sample was first normalized to the level of GAPDH on the same gel and then to its expression level in WT or vehicle-treated animals. The mean value for the WT or vehicle control group was set to 1.
Chromatin immunoprecipitation-qPCR
Chromatin immunoprecipitation (ChIP) was performed using the SimpleChIP Plus Sonication Chromatin IP Kit (#56383S, Cell Signaling Technology) following the similar methods described previously (Ghosh et al., 2022; Zhang et al., 2022). The DRG tissues (pooled from two mice for each sample) were placed in ice-cold phosphate-buffered saline (PBS) containing a protease inhibitor cocktail and cross-linked with 2% methanol-free formaldehyde for 10 min followed by quenching using glycine for 5 min on ice. After three washes with PBS, the tissue was homogenized for 15 min in ChIP Sonication Cell Lysis Buffer supplemented with a protease inhibitor cocktail, and the cells were pelleted at 5,000 × g for 5 min at 4°C followed by nuclei isolation using ChIP Sonication Nuclear Lysis Buffer containing a protease inhibitor cocktail. The chromatin fragmentation was performed in a Covaris M220 AFA-focused ultrasonicator (Covaris) at a 5% duty factor for 6 min. The efficacy of shearing (200–800 bp) was cross-checked by resolving on a 1.2% agarose gel at 30 volts for 4 h. The lysate was clarified by centrifugation at 21,000 × g for 10 min at 4°C, and the supernatant with cross-linked fragmented chromatin was diluted at 1:4 ratio with ChIP buffer containing a protease inhibitor cocktail. A 10% (v/v) of the fragmented, diluted chromatin was stored at −20°C for input DNA extraction. For each immunoprecipitation reaction, a 250 µl of diluted chromatin was incubated overnight at 4°C with rotation by adding 5 μg of a primary antibody against a target protein of interest. The following primary antibodies were used for ChIP: rabbit anti-G9a (#3306, Cell Signaling Technology), rabbit anti-H3K9me2 (#4658, Cell Signaling Technology), rabbit anti-H3K9ac (#9649, Cell Signaling Technology), and rabbit anti-H3K4me3 (#9751S, Cell Signaling Technology). The specificity of the primary antibodies has been validated previously (Laumet et al., 2015; Zhang et al., 2016, 2022; Kato et al., 2020; Ghosh et al., 2022). For the negative control, 1 μg of normal rabbit IgG (#2729S, Cell Signaling Technology) was used.
Following immunoprecipitation, 30 µl of ChIP-grade protein G magnetic beads was added to each IP reaction tube and incubated for 2 h at 4°C with rotation. Subsequently, the beads were washed twice with low salt buffer, followed by a wash with high salt buffer. The bound chromatin was eluted using ChIP Elution Buffer for 2 h at 65°C in a thermomixer. The eluted chromatin was de-cross-linked with 5 M NaCl and Proteinase K treatment for 8 h at 65°C in a thermomixer, followed by DNA purification using the QIAquick PCR Purification Kit (#28104, Qiagen). The purified DNA was subjected to real-time PCR using the promoter-specific primers (Table 3) mixed with the PowerUp SYBR Green Master Mix (Thermo Fisher Scientific) in a QuantStudio 7 Flex Real-Time PCR System (Applied Biosystems). The fast-cycling mode conditions (50°C for 2 min, 95°C for 2 min, 40 cycles of 95°C for 1 s, and 60°C for 30 s) were used to perform RT-PCR reactions. The mouse negative control primer set 2 (#71012, Active Motif) was used as a negative control for each antibody immunoprecipitation reaction because these primers are specific for a gene desert region located on mouse chromosome 17. The cycle threshold (CT) value in each group was normalized to the input using the following formula: (2−ΔCT) × 100%, where ΔCT represents CT [ChIP] − (CT [Input] − Log2 (Input Dilution Factor).
ChIP-qPCR primers used in the study
RNA sequencing
RNA sequencing of total RNAs isolated from L3–L5 DRGs of WT and G9a-iKO mice injected with CFA or vehicle was performed, as we described previously (Laumet et al., 2015; Zhang et al., 2022). A library was constructed and sequenced using the Illumina Platform with paired-end, 150 bp reads (Novogene). The raw FASTQ reads (average of 20 million reads/sample) were mapped to the mouse genome assembly (GRCm39/mm39) using the HISAT2 (version 2.2.1) aligner program as a reference (Kim et al., 2019). The aligned reads in sequence alignment/map format were then sorted and indexed into binary alignment/map format using SAMtools (version 1.19.2; Li et al., 2009). The number of mapped reads for each gene in the mouse genome was then counted by featureCounts (version 2.0.6) program (Liao et al., 2014) of the Subread (http://subread.sourceforge.net) package. The count data were used to estimate fold changes and identify differentially expressed genes (DEGs) using DESeq2 function in R environment (Love et al., 2014). Differences in log 2 fold change with a p value <0.05 were considered statistically significant.
To compare groups and generate a Venn diagram, the Venny tool (version 2.1) was used (https://csbg.cnb.csic.es/BioinfoGP/venny.html). A heatmap was produced using the ClustVis web tool (https://biit.cs.ut.ee/clustvis/). A volcano plot was generated using the −log10 function of the p value for each gene. Gene set enrichment analysis for the DEGs was performed online using the Enrichr tool (https://maayanlab.cloud/Enrichr/). The raw sequencing datasets have been deposited in the Gene Expression Omnibus (GEO accession GSE272517).
Study design and statistical analysis
Data are expressed as mean ± standard deviation (SD). The sample sizes of behavioral, electrophysiological, and biochemistry experiments were similar to those we published previously (Laumet et al., 2015; Huang et al., 2019; Ghosh et al., 2022). The animals were randomly assigned (1:1 allocation) to the control and treatment groups as they become available. The investigators performing the behavioral experiments were blinded to the drug treatment and genotypes. No animal died during the final experiments, and no test for outliers in the data were conducted. Data from male and female mice were pooled because no sex differences were observed in the biochemical and behavioral data in this study. Likewise, both male and female mice displayed comparable CFA-induced pain hypersensitivity, which is similarly reduced by TRPA1 and TRPV1 antagonists (Fernandes et al., 2011) or by Trpa1 knock-out (Horváth et al., 2016). Normality of datasets was evaluated using the Shapiro–Wilk test. The significance between the two groups was determined using a two-tailed Student's t test. Comparisons involving more than two groups were analyzed using one-way or two-way ANOVA followed by Tukey's post hoc test. All statistical analyses were conducted using Prism software (version 10, GraphPad Software). A p value of <0.05 was considered statistically significant.
Results
G9a at the spinal cord level promotes persistent inflammatory pain in rats
We first examined whether G9a activity at the spinal cord level contributes to persistent inflammatory pain. Injection of CFA into the hindpaw induces inflammatory pain, characterized by mechanical allodynia and thermal hypersensitivity, lasting for at least 2 weeks (Nagakura et al., 2003; Huang et al., 2019). We injected 50 µl of either CFA or vehicle into the left hindpaw of rats. Within 24 h after CFA injection, hindpaw withdrawal thresholds in response to von Frey filaments (F(9,90) = 24.5, p < 0.001), noxious pressure (F(9,90) = 17.3, p < 0.001), and heat stimuli (F(9,90) = 11.9, p < 0.001) were profoundly reduced (n = 6 rats per group; Fig. 1A). Seven days later, we administered intrathecally UNC0638 (10 µg per day), a specific G9a inhibitor (Vedadi et al., 2011), for 8 consecutive days. The effective intrathecal dose of UNC0638 has been previously validated (Laumet et al., 2015). Intrathecal administration of UNC0638 gradually elevated hindpaw withdrawal thresholds over the treatment period in CFA-treated rats. However, treatment with UNC0638 had no significant effect in vehicle-injected rats (n = 6 rats per group; Fig. 1A).
G9a inhibition or knockdown at the spinal cord level attenuates CFA-induced persistent pain hypersensitivity in rats. A, Time course of the effect of intrathecal injection of UNC0638 (10 µg per day) for 8 consecutive days on tactile, pressure, and heat withdrawal thresholds in rats injected with CFA or vehicle (Veh) into the left hindpaw (n = 6 rats per group). B, Time course of the effect of intrathecal injection of G9a-specific siRNA (4 µg per day) or control siRNA (Ctrl-siRNA) for 7 consecutive days on tactile, pressure, and heat withdrawal thresholds in rats injected with CFA or vehicle (Veh) into the left hindpaw (n = 6 rats per group). The arrows indicate the timing of CFA or vehicle injection. *p < 0.05, **p < 0.01, ***p < 0.001 versus pre-UNC0638 treatment or pre-siRNA treatment (day 1). #p < 0.05, ##p < 0.01, ###p < 0.001 versus the CFA + control siRNA group at the same time point. Two-way ANOVA with Tukey's post hoc test. Data are shown as mean ± SD.
To further substantiate the role of G9a at the spinal cord level in persistent inflammatory pain, we administered G9a-specific siRNA or control siRNA (4 µg per day) intrathecally to rats 7 d after CFA or vehicle injection into the left hindpaw. The efficacy of G9a-specific siRNA has been demonstrated in our previous studies (Luo et al., 2005; Laumet et al., 2015). Intrathecal injection of either G9a-siRNA or control siRNA had no significant effect on the paw withdrawal thresholds in vehicle-injected rats (n = 6 rats per group; Fig. 1B). In contrast, intrathecal administration of G9a-specific siRNA for 7 consecutive days gradually reversed tactile, pressure, and heat withdrawal thresholds in CFA injected rats (n = 6 rats; Fig. 1B). Intrathecal treatment with the control siRNA had no significant effect on the withdrawal thresholds of CFA-injected rats (n = 6 rats; Fig. 1B). These data suggest that G9a at the spinal cord level plays a critical role in maintaining persistent inflammatory pain.
G9a in primary sensory neurons mediates the development of persistent inflammatory pain in mice
Because intrathecal delivery of agents targets both the spinal cord and DRG, we next aimed to specifically determine whether G9a expressed in DRG neurons plays a role in the development of persistent inflammatory pain. To generate Ehmt2 conditional knock-out (Ehmt2-cKO) mice in which G9a is genetically ablated from primary sensory neurons, we crossed Ehmt2flox/flox mice (Sampath et al., 2007) with AvilCre/+ mice (Zhou et al., 2010). Advillin is predominantly expressed in primary sensory neurons, and this Avil Cre-recombinase driver targets 84–87% of DRG neurons (Hasegawa et al., 2007; Lau et al., 2011). As previously reported (Laumet et al., 2015; Zhang et al., 2016), the paw withdrawal thresholds did not differ significantly between adult wild-type (WT) and Ehmt2-cKO mice under baseline conditions. In WT mice, injection of CFA, but not the vehicle, into the left hindpaw caused a rapid and sustained reduction in tactile, pressure, and heat withdrawal thresholds (n = 6 mice per group; Fig. 2A). Strikingly, CFA injection failed to induce a reduction in the pressure and heat withdrawal thresholds in Ehmt2-cKO mice (n = 6 mice per group; Fig. 2A). Although Ehmt2-cKO mice initially showed a reduced tactile threshold 3 d after CFA injection, it returned to the pre-CFA baseline level by 9 d postinjection.
G9a genetic ablation in primary sensory neurons prevent CFA-induced persistent pain hypersensitivity in mice. A, Time course of changes in tactile, pressure, and heat withdrawal thresholds in Ehmt2 conditional knock-out (Ehmt2-cKO) and wild-type (WT) mice (n = 6 mice per group) after CFA or vehicle (Veh) injection into the left hindpaw. B, Time course of changes in tactile, pressure, and heat withdrawal thresholds in inducible Ehmt2 conditional knock-out (Ehmt2-iKO) and WT mice (n = 6 mice per group) after CFA or vehicle (Veh) injection into the left hindpaw. The arrows indicate the timing of CFA or vehicle injection. In A and B, *p < 0.01, ***p < 0.001 versus respective baselines (time 0). #p < 0.05, ##p < 0.01, ###p < 0.001 versus WT mice at the same time point. Two-way ANOVA followed by Tukey's post hoc test. C, Quantitative PCR data show the mRNA level of Ehmt2 in the DRG and spinal cord tissues from WT and Ehmt2-iKO mice 2 weeks after CFA or vehicle treatment (n = 6 mice per group). ***p < 0.001 (two-way ANOVA followed by Tukey's post hoc test). D, Representative blotting images and quantification show the G9a protein levels in the DRG from WT and Ehmt2-iKO mice 2 weeks after the last injection of tamoxifen (n = 6 mice per group). GAPDH was used as the loading control. ***p < 0.001 (two-tailed Student’s t test). Data are expressed as mean ± SD.
Germline constitutive knock-out of Ehmt2 may induce compensatory changes in primary sensory neurons. Therefore, we generated mice with inducible Ehmt2 deletion in primary sensory neurons by crossing Ehmt2flox/flox mice and Avil-CreERT2 mice (Lau et al., 2011). This approach produces G9a inducible knock-out (Ehmt2-iKO) mice upon tamoxifen treatment at the adult stage. Before CFA injection, tamoxifen treatment had no significant effect on tactile, pressure, or thermal withdrawal thresholds in WT or Ehmt2-iKO mice (n = 6 mice per group; Fig. 2B). In WT mice, CFA injection led to a rapid and sustained reduction in tactile, pressure, and heat withdrawal thresholds (n = 6 mice per group; Fig. 2B). In contrast, in Ehmt2-iKO mice, CFA injection did not significantly reduce the pressure or heat withdrawal thresholds. The tactile withdrawal threshold was reduced on Day 3 after CFA injection but gradually returned to the pre-CFA baseline level within 14 d in Ehmt2-iKO mice (n = 6 mice per group; Fig. 2B).
The mRNA level of Ehmt2 in the DRG, but not the spinal cord, was markedly lower in Ehmt2-iKO than in WT mice (F(1,20) = 274, p < 0.001; n = 6 mice per group; Fig. 2C). We also observed a substantial reduction in G9a protein levels in the DRG of Ehmt2-iKO mice compared with WT mice (t(10) = 6.14, p < 0.001; n = 6 mice per group; Fig. 2D). Collectively, these findings provide compelling evidence for the crucial role of G9a in primary sensory neurons in the development of persistent inflammatory pain.
Transcriptomic analyses of Ehmt2-regulated genes in the DRG of mice treated with CFA
G9a, as an epigenetic regulator, may influence the expression of numerous genes in the DRG in persistent inflammatory pain. Transcriptomic profiling of the DRG can reveal genome-wide expression changes in gene expression affected by experimental conditions (Laumet et al., 2015; Zhang et al., 2022). To uncover genome-wide expression changes caused by tissue inflammation and G9a ablation, we conducted RNA sequencing on RNA isolated from L3–L5 DRGs of WT and Ehmt2-iKO mice 14 d after CFA or vehicle injection. In WT mice, CFA treatment significantly upregulated 1,869 genes and downregulated 3,538 genes compared with vehicle-treated mice (Extended Data Table 3-1, Fig. 3A). In Ehmt2-iKO mice, CFA treatment significantly upregulated 4,269 genes and downregulated 4,746 genes (Extended Data Table 3-2, Fig. 3A).
Effects of CFA treatment and G9a ablation in primary sensory neurons on global gene expression in the DRG. A, Volcano plots show CFA-induced upregulated and downregulated (p < 0.05) genes in the RNA pool in the DRG from WT and Ehmt2-iKO mice. Total RNA was isolated from the DRG of WT and Ehmt2-iKO mice injected with CFA or vehicle and was subjected to RNA sequencing analysis to identify differentially expressed genes. B, Venn diagram shows the overlap and number of significantly (p < 0.05) upregulated and downregulated genes in the DRG of WT and Ehmt2-iKO mice injected with CFA. C, Heat map shows the clustering pattern of the significantly (p < 0.05) upregulated 659 genes and downregulated 322 genes in the DRG from WT and Ehmt2-iKO mice injected with CFA or vehicle. D, Top 10 biological processes identified by the gene ontology (GO) enrichment analysis of the significantly (p < 0.05) upregulated 322 genes and downregulated 659 genes in the DRG from WT and Ehmt2-iKO mice injected with CFA or vehicle (see Extended Data Tables 3-1–3-4).
Table 3-1
List of upregulated and downregulated genes by CFA treatment in wild-type mice (Excel file). Download Table 3-1, XLSX file.
Table 3-2
List of upregulated and downregulated genes by CFA treatment in Ehmt2-iKO mice (Excel file). Download Table 3-2, XLSX file.
Table 3-3
List of upregulated and downregulated genes in vehicle-treated wild-type and vehicle-treated Ehmt2-iKO mice (Excel file). Download Table 3-3, XLSX file.
Table 3-4
Gene ontology pathway analyses of differentially expressed genes in the DRG of CFA-treated wild-type and Ehmt2-iKO mice (Excel file). Download Table 3-4, XLSX file.
Comparison of vehicle-treated Ehmt2-iKO and vehicle-treated WT mice revealed that ablating Ehmt2 in DRG neurons significantly increased the expression of 4,576 genes and reduced the expression of 5,603 genes (Extended Data Table 3-3). Overlapping analysis showed that the expression of 322 genes upregulated by CFA in WT mice was reversed by Ehmt2 ablation (Extended Data Tables 3-1–3-3, Fig. 3B,C). Additionally, 659 genes downregulated by CFA in WT mice were reversed by Ehmt2 ablation (Extended Data Tables 3-1–3-3, Fig. 3B,C).
Gene ontology pathway enrichment analysis revealed that the 659 genes downregulated by CFA in WT mice (but reversed by Ehmt2-iKO) are associated with several biological processes, including the positive regulation of cytokine production, enhanced transcription, and key signaling pathways involved in the inflammatory response (Extended Data Table 3-4, Fig. 3D). Additionally, the 322 genes upregulated by CFA in WT mice (but reversed by Ehmt2-iKO) were enriched in processes related to synaptic and transmembrane transport, nervous system development, apoptosis, and anterograde trans-synaptic signal transduction, all of which are critical for neuronal regulation (Extended Data Table 3-4, Fig. 3D).
Interestingly, RNA sequencing analysis identified that 22 TRP channel genes were differentially expressed in the DRG after CFA treatment in WT and Ehmt2-iKO mice (Extended Data Tables 3-1–3-3). In WT mice, CFA treatment significantly increased the expression of Trpa1, Trpv1, Trpc7, and Trpm8, while reducing the expression of Trpm1 in the DRG. Notably, in vehicle-treated Ehmt2-iKO mice, the expression level of Trpa1, and Trpv1, Trpm8, and Trpc7 was already elevated compared with vehicle-treated WT mice. However, CFA treatment failed to further increase the expression of Trpa1, Trpv1, and Trpm8 in the DRG of Ehmt2-iKO mice (Extended Data Tables 3-1–3-3).
G9a in DRG neurons augments CFA-induced transcription of Trpa1 and Trpv1
Because the contribution of TRPA1 or TRPV1 to CFA-induced pain hypersensitivity has been well documented through genetic deletion and antagonist studies (Davis et al., 2000; Cui et al., 2006; Fernandes et al., 2011, 2016; Horváth et al., 2016), we focused subsequent experiments on the role of G9a in inflammation-induced transcriptional activation of Trpa1 and Trpv1 in the DRG. We used quantitative PCR to measure Trpa1 and Trpv1 mRNA levels in the L5 and L6 DRGs at 3, 10, and 14 d after CFA or vehicle injection in rats. CFA injection caused a sustained increase in Trpa1 mRNA levels in the DRG at Days 3 (t(10) = 4.65, p < 0.001), 10 (t(10) = 5.05, p < 0.001), and 14 (t(10) = 9.12, p < 0.001; n = 6 rats per group; Fig. 4A). Additionally, CFA injection substantially increased Trpv1 mRNA levels in the DRG at days 10 (t(10) = 5.13, p < 0.001) and 14 (t(10) = 6.86, p < 0.001; n = 6 rats per group; Fig. 4B).
G9a mediates inflammation-induced transcription of Trpa1 and Trpv1 in the DRG. A, B, Quantitative PCR data show the mRNA levels of Trpa1 (A) and Trpv1 (B) in the DRG of rats 3, 10, and 14 d after CFA or vehicle injection (n = 6 rats per group). The mRNA of Gapdh was used for data normalization. ***p < 0.001 (two-tailed Student's t test). C, D, Quantitative PCR data show the mRNA levels of Trpa1 (C) and Trpv1 (D) in the DRG of WT and Ehmt2-iKO mice 3, 10, and 14 d after CFA or vehicle injection (n = 6 mice per group). *p < 0.05, **p < 0.01, ***p < 0.001 (two-way ANOVA followed by Tukey's post hoc test). E, Quantitative PCR data show the mRNA levels of Mrgpra3, Mrgprb4, Oprd1, and Cnr1 in the DRG of WT and Ehmt2-iKO mice (n = 6 mice per group). ***p < 0.001 (two-tailed Student's t test). Data are expressed as mean ± SD.
In wild-type mice, CFA treatment similarly increased Trpa1 mRNA levels in the DRG on Days 3 (F(1,20) = 73.2, p < 0.001), 10 (F(1,20) = 27.4, p < 0.01), and 14 (F(1,20) = 79.5, p < 0.001; n = 6 mice per group; Fig. 4C). CFA injection also significantly increased Trpv1 mRNA levels in the DRG of WT mice at Days 3 (F(1,20) = 22.3, p < 0.001), 10 (F(1,20) = 33.1, p < 0.001), and 14 (F(1,20) = 90.7, p < 0.001; n = 6 mice per group; Fig. 4D).
Notably, the mRNA levels of both Trpa1 and Trpv1 in the DRG were significantly higher in vehicle-treated Ehmt2-iKO mice than in vehicle-treated WT mice (n = 6 mice per group; Fig. 4C,D). Remarkably, CFA injection substantially reduced the mRNA levels of Trpa1 and Trpv1 in the DRG of Ehmt2-iKO mice at Days 3, 10, and 14 compared with vehicle-injected Ehmt2-iKO mice (n = 6 mice per group; Fig. 4C,D). In contrast, Trpv1 mRNA levels were notably higher in the DRG of vehicle-treated Ehmt2-iKO mice compared with vehicle-treated WT mice (n = 6 mice per group; Fig. 4C,D). These findings suggest that G9a in DRG neurons plays a critical role in the sustained transcriptional activation of Trpa1 and Trpv1 in response to tissue inflammation.
The RNA sequencing data (Extended Data Table 3-3) revealed that Ehmt2 ablation in DRG neurons led to increased expression of several antinociceptive genes, including MAS-related G-protein-coupled receptor subtypes (Mrgpra3 and Mrgprb4), the δ-opioid receptor (encoded by Oprd1 gene), and the cannabinoid CB1 receptor (encoded by Cnr1 gene). Quantitative PCR analysis confirmed significant increases in the mRNA levels of Mrgpra3 (t(10) = 8.48, p < 0.001), Mrgprb4 (t(10) = 9.12, p < 0.001), Oprd1 (t(10) = 6.78, p < 0.001), and Cnr1 (t(10) = 10.4, p < 0.001) in the DRG of Ehmt2-iKO mice compared with WT mice (n = 6 mice per group; Fig. 4E).
Tissue inflammation diminishes G9a enrichment and activity at Trpa1 and Trpv1 promoters in the DRG
Transcriptional activation of nociceptive genes is dynamically regulated by chromatin-level epigenetic mechanisms, which determine the accessibility of promoter elements on DNA (Penas and Navarro, 2018; Ghosh and Pan, 2022). This regulation can occur either through the loss of repressive histone marks (such as H3K9me2 introduced by G9a) or through the gain of activating histone marks (such as H3K9ac and H3K4me3) at the gene promoter (Laumet et al., 2015; Luo et al., 2020; Ghosh et al., 2022; Zhang et al., 2022). We first measured the total protein level of G9a in the DRG and spinal cord of rats 14 d after injection of CFA or vehicle. CFA did not significantly affect the total protein levels of G9a in either the DRG or spinal cord (n = 6 rats per group; Fig. 5A).
CFA-induced inflammation diminishes the enrichment of G9a and its repressive H3K9me2 mark at Trpa1 and Trpv1 promoters in the DRG. A, Representative blotting images and quantification show the G9a protein level in the DRG and spinal cord (SC) of rats 14 d after CFA or vehicle injection (n = 6 rats per group). GAPDH was used as the loading control. B, C, ChIP-qPCR analyses show the enrichment of G9a and H3K9me2 in the indicated regions at the promoter of Trpa1 (B) and Trpv1 (C) in the DRG from rats 14 d after injection with vehicle or CFA (n = 6 rats in each group). ***p < 0.001 (two-tailed Student's t test). D, ChIP-qPCR analyses show the H3K9me2 enrichment in the indicated region within the promoter of Trpa1 and Trpv1 in the DRG from WT and Ehmt2-iKO mice 14 d after injection with vehicle or CFA (n = 6 mice per group). ***p < 0.001 (two-way ANOVA followed by Tukey's post hoc test). Data are expressed as mean ± SD.
Using ChIP-qPCR, we next determined whether CFA treatment affects the enrichment of G9a and H3K9me2 at Trpa1 and Trpv1 promoters in the DRG. CFA injection profoundly reduced G9a occupancy in the −201 to −112 bp region (t(10) = 9.53, p < 0.001) and the −14 to +97 bp region (t(10) = 23.6, p < 0.001) across the transcription start site of the rat Trpa1 gene (located on chromosome 5) in the DRG (n = 6 rats per group; Fig. 5B). CFA injection caused a similar reduction in the enrichment of H3K9me2 in the −201 to −112 bp region (t(10) = 17.7, p < 0.001) and the −14 to +97 bp region (t(10) = 21.1, p < 0.001) of the Trpa1 promoter in the DRG (n = 6 rats per group; Fig. 5B).
The Trpv1 promoter was similarly enriched with G9a and H3K9me2 in the DRG of vehicle-treated rats. CFA injection caused a large reduction in G9a abundance in the −126 to −22 bp region (t(10) = 11.8, p < 0.001) and +9 to +105 bp region (t(10) = 20.8, p < 0.001) within the Trpv1 promoter (on chromosome 10) in the DRG (n = 6 rats per group; Fig. 5C). Furthermore, CFA injection markedly reduced the enrichment of H3K9me2 in the −126 to −22 bp region (t(10) = 9.00, p < 0.001) and +9 to +105 bp region (t(10) = 11.3, p < 0.001) of the Trpv1 promoter in the DRG (n = 6 rats per group; Fig. 5C).
In WT mice, CFA treatment significantly reduced H3K9me2 levels at the promoters of Trpa1 (−72 to +34 bp region; F(1,20) = 76.2, p < 0.001) and Trpv1 (−16 to +78 bp region; F(1,20) = 94.6, p < 0.001) in the DRG (n = 6 mice per group; Fig. 5D). Notably, the H3K9me2 enrichment at Trpa1 (F(1,20) = 145, p < 0.001) and Trpv1 (F(1,20) = 166, p < 0.001) promoters in the DRG was diminished in vehicle-treated Ehmt2-iKO mice compared with vehicle-treated WT mice (n = 6 mice per group; Fig. 5D). These results suggest that tissue inflammation reduces the levels of G9a and its repressive H3K9me2 histone mark at the Trpa1 and Trpv1 promoters in the DRG.
Tissue inflammation enhances activating histone marks at Trpa1 and Trpv1 promoter via G9a in DRG neurons
Reduced G9a levels at the Trpa1 and Trpv1 promoters due to inflammation cannot explain why the inflammation-potentiated expression of TRPA1 and TRPV1 was reversed by Ehmt2 ablation in DRG neurons. H3K4me3 and H3K9ac are well-known active histone marks associated with active gene transcription (Wang et al., 2008; Laumet et al., 2015; Ghosh et al., 2022; Zhang et al., 2022). Trimethylation at H3K4 facilitates acetylation of neighboring H3K9 residues, with both marks colocalizing on actively transcribing gene promoters (Gates et al., 2017). ChIP-qPCR assays revealed that CFA injection substantially enhanced the abundance of H3K9ac (F(1,20) = 46.3, p < 0.001) and H3K4me3 (F(1,20) = 63.5, p < 0.001) in the −72 to +34 bp region of the Trpa1 promoter in the DRG of WT mice (n = 6 mice per group; Fig. 6A). Similarly, CFA injection significantly increased the occupancy of H3K9ac (F(1,20) = 160, p < 0.001) and H3K4me3 (F(1,20) = 289, p < 0.001) in the −16 to +78 bp region of the Trpv1 promoter in the DRG of WT mice (n = 6 mice per group; Fig. 6B).
Inflammation increases the enrichment of H3K9ac and H3K4me3 at Trpa1 and Trpv1 promoters via G9a in DRG neurons. A, B, ChIP-qPCR analyses show the enrichment of H3K9ac and H3K4me3 marks in the indicated region within the promoter of Trpa1 (A) and Trpv1 (B) in the DRG from WT and Ehmt2-iKO mice 14 d after injection with vehicle or CFA (n = 6 mice per group). *p < 0.05, **p < 0.01, ***p < 0.001 (two-way ANOVA followed by Tukey's post hoc test). Data are expressed as mean ± SD.
H3K9ac (F(1,20) = 46.3, p < 0.001) and H3K4me3 (F(1,20) = 63.5, p = 0.003) levels at the Trpa1 promoter in the DRG were much greater in Ehmt2-iKO mice than in WT mice (n = 6 mice per group; Fig. 6A). Notably, Ehmt2-iKO alone significantly enriched H3K9ac (F(1,20) = 160, p < 0.001) and H3K4me3 (F(1,20) = 289, p < 0.001) at the Trpv1 promoter in the DRG (n = 6 mice per group; Fig. 6B). Strikingly, in Ehmt2-iKO mice, CFA injection significantly reduced the occupancy of H3K9ac (F(1,20) = 46.3, p = 0.026) and H3K4me3 (F(1,20) = 63.5, p = 0.012) at the Trpa1 promoter in the DRG (n = 6 mice per group; Fig. 6A). Additionally, CFA treatment markedly lowered the enrichment of both H3K9ac (F(1,20) = 160, p < 0.001) and H3K4me3 (F(1,20) = 289, p < 0.001) in the −16 to +78 bp region of the Trpv1 promoter in the DRG in Ehmt2-iKO mice (n = 6 mice per group; Fig. 6B). These data suggest that in response to tissue inflammation, G9a in primary sensory neurons has a major role in recruiting activating histone marks to Trpa1 and Trpv1 promoters, thereby favoring their transcriptional activation.
Discussion
Our study, using complementary loss-of-function approaches, provides the first evidence that G9a in primary sensory neurons is crucial for the development of persistent inflammatory pain. Although the epigenetic mechanisms of chronic neuropathic pain have been extensively studied, our understanding of the epigenetic basis of persistent inflammatory pain remains limited. Histone modifications, including acetylation and methylation, can modulate chronic pain by regulating inflammatory mediators such as chemokines and cytokines (Leoni et al., 2002; Zhou et al., 2020). For instance, inhibition of histone deacetylases (HDACs), particularly HDAC1/2/6, has been shown to reduce inflammation-induced pain hypersensitivity (Chiechio et al., 2009; Bai et al., 2010; Xu et al., 2024). Also, lysine-specific demethylase 6B (KDM6B) may exacerbate inflammatory pain by enhancing TNF-α expression in the DRG and spinal cord (Qiao et al., 2023). In this study, we found that that inhibiting G9a activity or knocking down G9a via siRNA at the spinal cord level diminished CFA-induced persistent pain hypersensitivity in rats. Remarkably, both constitutive and inducible conditional knock-out of Ehmt2 in primary sensory neurons effectively blocked the development of CFA-induced persistent pain in mice. These results underscore the critical role of G9a as an epigenetic regulator in primary sensory neurons, driving persistent inflammatory pain.
TRPA1 and TRPV1 are coexpressed in DRG neurons and their central and peripheral nerve terminals (Story et al., 2003; Nagata et al., 2005; Bautista et al., 2006; Chen and Pan, 2006). Both channels are crucial for detecting nociceptive physical and chemical stimuli (Caterina et al., 2000; Davis et al., 2000; Obata et al., 2005) and in increasing glutamatergic input to spinal dorsal horn neurons during inflammatory pain (Zhou et al., 2009; Huang et al., 2019). Our RNA sequencing analysis showed that CFA-induced inflammation altered the transcription of several TRP channel members. Among them, Ehmt2 deletion in DRG neurons reduced CFA-induced upregulation of Trpa1, Trpv1/2, Trpm2/4/8, and Trpc1/4/7. While the role of TRPA1 and TRPV1 in persistent inflammatory pain is well established, the contributions of other TRP channels remain less clear. Therefore, we selected Trpa1 and Trpv1 as representative pronociceptive genes for further promoter analysis to investigate the mechanisms underlying their transcriptional activation. In the present study, we found that ablation of Ehmt2 in DRG neurons increased baseline mRNA levels of Trpa1 and Trpv1, indicating that G9a normally acts as a constitutive repressor of these genes in primary sensory neuron. However, despite the enhanced expression of Trpa1 and Trpv1 in the DRG, Ehmt2-iKO and Ehmt2-cKO mice did not exhibit any pain hypersensitivity, consistent with our previous reports (Laumet et al., 2015; Zhang et al., 2016). The absence of pain hypersensitivity could be attributed to the concurrent upregulation of certain antinociceptive genes, such as Oprd1 and Cnr1 (Luo et al., 2020; Jin et al., 2022), in the DRG of these mice. Additionally, our RNA sequencing data analysis and quantitative PCR assay revealed that Ehmt2 deletion in DRG neurons increased the expression levels of Mrgpra3 and Mrgprb4 in the DRG, both of which are primarily antinociceptive (Guan et al., 2010). Intriguingly, we found that CFA-induced inflammation reduced the enrichment of G9a and its repressive substrate mark, H3K9me2, at the Trpa1 and Trpv1 promoters in the DRG. This inflammation-induced dissociation of G9a from the Trpa1 and Trpv1 promoters could facilitate their transcription in the context of tissue inflammation.
A striking finding of our study is that G9a in DRG neurons has a noncanonical role in promoting the transcription of Trpa1 and Trpv1 in inflammatory pain. We showed that CFA-induced inflammatory pain is associated with a sustained increase in TRPA1 and TRPV1 expression in the DRG of WT mice. Remarkably, in mice with Ehmt2 ablation in DRG neurons, CFA injection failed to increase the mRNA levels of Trpa1 and Trpv1 in the DRG. This finding suggests that while G9a in DRG neurons normally restricts the transcription of Trpa1 and Trpv1 under noninflammatory conditions, it actively enhances their transcription in response to tissue inflammation. Thus, under inflammatory pain conditions, G9a mainly facilitates the transcription of Trpa1 and Trpv1 in the DRG. It should be noted that G9a inhibition or cKO in DRG neurons may also attenuate inflammatory pain by reducing the expression of nociceptive genes in addition to Trpa1 and Trpv1. For example, the expression of Mrgprd, which is involved in persistent inflammatory hyperalgesia (Lan et al., 2020), in the DRG was increased in WT but not in Ehmt2-iKO mice.
Our promoter analyses reveal that ablating Ehmt2 in DRG neurons reversed inflammation-induced increases in the enrichment of H3K9ac and H3K4me3—two activating histone marks—at the promoters of Trpa1 and Trpv1 in the DRG. This suggests that G9a in DRG neurons enhances the expression of these genes by acting as a transcriptional activator during tissue inflammation. However, the precise mechanism by which G9a enhances activating histone marks at these promoters in response to inflammation remains unclear. The lysine methyltransferase mixed-lineage leukemia 2 (MLL2) contributes to H3K4 trimethylation on bivalent promoters (Denissov et al., 2014). Additionally, H3K4me3 may recruit lysine acetyltransferases, such as GCN5 and PCAF, which catalyze the formation of H3K9ac (Bian et al., 2011; Jin et al., 2011). Future studies are needed to determine whether G9a promotes the recruitment of MLL2 and CREB-binding protein at Trpa1 and Trpv1 promoters in the DRG to enhance their transcription during tissue inflammation. The dual role of G9a as both a repressor and an activator of gene expression has been observed in other tissues and cell types (Shankar et al., 2013). For instance, in adult erythroid cells, G9a regulates β-globin gene transcription through both repressive and activating mechanisms (Chaturvedi et al., 2009). Similarly, in naive CD4+ T helper cells, G9a knock-out or inhibition impairs differentiation into Th2 cells, leading to reduced expression of Th2-specific cytokines, such as IL-4, IL-5, and IL-13 (Lehnertz et al., 2010). Additionally, SUMOylation of G9a has been implicated in the recruitment of PCAF to the promoters of E2F1-target genes, resulting in elevated H3K9ac and enhanced transcriptional activation, which contributes to myoblast proliferation (Srinivasan et al., 2019).
G9a may regulate Trpa1 and Trpv1 transcription either directly or indirectly by balancing activating and repressive histone modifications at their promoters in the DRG. The dissociation of G9a from the Trpa1 and Trpv1 promoters alone cannot fully explain why CFA-induced upregulation of TRPA1 and TRPV1 was reversed in mice lacking Ehmt2 in DRG neurons. Typically, the colocalization of repressive and activating histone marks at gene promoters is associated with transcriptional activation (Gates et al., 2017). The concept of bivalent chromatin domains, which are marked by both activating and repressive histone modifications (Bernstein et al., 2006; Kumar et al., 2021), may be involved in G9a's action in DRG neurons in response to tissue inflammation. These bivalent domains maintain a “poised” state at gene promoters, allowing for rapid activation or silencing based on cellular signals (Blanco et al., 2020). In Ehmt2-iKO mice, CFA-induced inflammation failed to increase the abundance of H3K4me3 and H3K9ac at Trpa1 and Trpv1 promoters in the DRG. This suggests that G9a in DRG neurons is required for the sustained increase in transcription of Trpa1 and Trpv1 during inflammatory pain. Tissue inflammation reduces G9a binding and activity at Trpa1 and Trpv1 promoters, which may produce a permissive epigenetic landscape, allowing their transcriptional upregulation. We have previously reported similar bivalent modifications at the promoter of Cnr2, with increased levels of active marks (H3K4me3 and H3K9ac) and diminished repressive marks (H3K9me2 and H3K27me3) in the injured DRG, contributing to increased CB2 receptor expression in DRG neurons during neuropathic pain (Ghosh et al., 2022).
The dynamic regulation of active and repressive states at bivalent chromatin regions often involves a complex interplay of various histone modifying enzymes. Histone modifying “writers” and “erasers” can colocalize at bivalent regions to modulate gene expression. Depending on their interacting partners, G9a can function either as a repressor or activator. For example, by interacting with the H3K4 demethylase lysine-specific demethylase 1 A (LSD1), G9a can act as a transcriptional activator through the demethylation of H3K9me2 (Metzger et al., 2005; Lee et al., 2006; Shankar et al., 2013). Conversely, G9A may interact with Jarid1a (lysine demethylase 5A) to demethylate H3K4me3, contributing to gene silencing (Chaturvedi et al., 2012). It remains unclear how inflammation affects the enrichment of LSD1 or Jarid1a at the Trpa1 and Trpv1 promoters and alters the methylation status of H3K9 and H3K4 in the DRG. Further studies are warranted to investigate how these enzymes interact with G9a in the control of inflammation-induced Trpa1 and Trpv1 transcription in the DRG.
In summary, our study reveals that the histone methyltransferase G9a has a dual role in regulating the Trpa1 and Trpv1 transcription in primary sensory neurons. Under normal conditions, G9a acts as a repressor, suppressing the transcription of these genes. However, in response to tissue inflammation, G9a switches to an activator, promoting Trpa1 and Trpv1 transcription by increasing the presence of activating histone marks at their promoters. These bivalent histone modifications lead to the sustained upregulation of TRPA1 and TRPV1 in the DRG during inflammation (Fig. 7). This new information advances our understanding of the epigenetic mechanisms underlying persistent inflammatory pain and underscore the multifaceted role of G9a in primary sensory neurons. Targeting G9a could offer a potential therapeutic strategy to reduce TRPA1 and TRPV1 expression and alleviate persistent inflammatory pain.
The schematic illustrates G9a-mediated transcriptional regulation of Trpa1 and Trpv1 in primary sensory neurons in inflammatory pain. Under normal conditions, G9a and its repressive H3K9me2 mark are abundantly present at the Trpa1 and Trpv1 promoters, suppressing transcription in primary sensory neurons. Following inflammation, the enrichment of H3K9me2 and repressive G9a complexes decreases, while levels of H3K9ac and H3K4me3 increase due to the recruitment of activating G9a complexes at these promoters. Through these bivalent histone modifications, G9a promotes the transcription of Trpa1 and Trpv1 in primary sensory neurons, contributing to sustained inflammatory pain
Footnotes
This work was supported by grants (NS132398 and NS101880) from the National Institutes of Health and by the Pamela and Wayne Garrison Distinguished Chair Endowment.
The authors declare no competing financial interests.
- Correspondence should be addressed to Hui-Lin Pan at huilinpan{at}mdanderson.org.