All experimental protocols were approved by the Animal Care and Use Committee of Zhejiang University (Hangzhou, Zhejiang Province, China) and were in accordance with the Declaration of the National Institutes of Health Guide for the Care and Use of Laboratory Animals (Publication No. 80-23, revised 1996). Eight- to twelve-week-old male C57BL/6J mice, weighing about 20 g, provided by Slac Laboratory Animal Co. (Cat No: 5807319, Research Resource Identifier (RRID):MGI: 5807391, Shanghai, China), were used in our experiments. The mice were housed in groups with a 12 h dark/light cycle at an appropriate temperature and humidity in specific pathogen-free cages (5 mice/cage). They had free access to food and water. After the animals had been acclimated in a quiet, temperature-controlled room for 1 h, behavioral studies were performed by other investigators, who were blinded to the drugs used. All behavioral studies were performed between 7 am and 7 pm. The study was not preregistered, and experimental animals were selected at random in this study. For the behavioral tests, there were 12 mice per group. For western blots, there were 6 samples per group, and each DRG sample came from 3 mice, while each spinal cord sample came from 1 mouse. For immunofluorescence, there were 3 mice per group. For quantitative real-time reverse transcription (RT)–PCR, there were 4 samples per group, and each DRG sample came from 2 mice, while each spinal cord sample came from 1 mouse. For pyrosequencing, there were 3 samples per group and each DRG sample came from 6 mice, while each spinal cord sample came from 3 mice. For chromatin immunoprecipitation (ChIP), there were 3 samples per group and each sample came from 10 mice. In addition, samples from 20 mice were used for co-immunoprecipitation (co-IP) experiments. In total, 7 mice died from anesthetic overdose, and 5 mice were excluded based on the exclusion criteria (e.g., normal activity, body weight, gait, and free of wound infection) during the experiment. The precise numbers of mice used are described in the figure legends. To minimize the number of mice used, the experiments for Figures 1, 2 shared the same samples for western blotting and qRT-PCR. All mice were ear tagged with unique numerical identifiers so that blinded experiments could be performed. The randomization function of Microsoft Excel 2013 was used to generate random numbers between 0 and 1 for each mouse. These random numbers were then sorted in ascending order, generating a list that categorized the mice into corresponding groups. Following all procedures, the mice were euthanized by an overdose of isoflurane followed by rapid decapitation with sharp scissors. The experimenter was unaware of each animal’s group membership during experiment.

In accordance with a previous study (Bennett and Xie, 1988), after sodium pentobarbital (40 mg/kg, intraperitoneal injection; Sigma–Aldrich, St Louis, MI, United States; Cat No: 76-74-4) anesthesia, mice were cut with a small incision beneath the mid-thigh to expose the left sciatic nerve. A unilateral constriction injury was produced with three loose ligatures (6-0 silk thread, spaced at 1-mm intervals) just proximal to the trifurcation. In the sham group, prior steps were identical to the CCI group except that the nerve remained unligated. Finally, both groups had the incised skin closed off with suture. Only mice without any abnormality (e.g., normal activity, body weight, gait, and free of wound infection) were included in this study.

Morphine (Sigma; Cat No: 25-27-2) and 5-aza-2′-deoxycytidine (5-aza-dC; Sigma; Cat No: 2353-33-5) were dissolved in saline; saline was set as the control. Suberoylanilide hydroxamic acid (SAHA; Selleck Chemicals, Houston, TX, United States; Cat No: S1047) was dissolved in 5% dimethyl sulfoxide (DMSO; Sigma; Cat No: D2650-100 mL); 5% DMSO was set as the control. Morphine was subcutaneously given with a dose of 1.5 mg per kg of body weight. The use of 5-aza-dC was continued for three consecutive days at 5 μg/5 μL daily, 30 min before surgery. SAHA was continued for 7 consecutive days, at a dose of 50 mg/kg in a volume of 5 μL, 30 min before surgery. All drug doses are based on preliminary experiments (Zhou et al., 2014; Cao et al., 2015; Garriga et al., 2018), and did not alter the nociceptive response per se. Intrathecal injections were performed manually between L5 and L6 spinal segments, through the theca of the spinal cord into the subarachnoid space (Hylden and Wilcox, 1980). On entry of the needle, a quick flicking motion of the mouse’s tail or leg was noticed. For anesthesia, sodium pentobarbital and isoflurane (Sigma; Cat No: 26675-46-7) were used.

As described by Yuan et al. (2019), morphine was injected at a predefined time point subcutaneously. Heat tests were conducted before surgery (baseline latency) and 30 min after morphine injection (response latency). The percentage of the maximal possible analgesic effect (% MPAE) = ([response latency − baseline latency]/[cut-off time − baseline latency]) × 100%.

After sodium pentobarbital anesthesia was induced, we transcardially perfused the mice with 0.9% saline and then with 4% paraformaldehyde. After their removal, the ipsilateral DRGs (L3–L5) were post-fixed in 4% paraformaldehyde for 2 h, and then cryoprotected at 4°C in 30% sucrose until they sank to the bottom of the tube. Sample sections (16 μm) were cut using a cryostat. Next, the sections were incubated in phosphate buffered saline (PBS) containing 5% normal goat serum and 0.3% Triton X-100 at room temperature for 1 h. The sections were then incubated overnight with primary antibody at 4°C. The following primary antibodies were used in this study: mouse anti-NeuN (1:200; Abcam, Cambridge, United Kingdom; Cat No: ab104224, RRID:AB_10711040), mouse anti-GS (1:500; Millipore, Cat No: MAB302, RRID:AB_2110656), rabbit anti-MOR (1:200; Abcam; Cat No: ab10275, RRID:AB_2156356), rabbit anti-MeCP2 (1:200; #3456; Cell Signaling Technology, Danvers, MA, United States; Cat No: 1205-MeCP2, RRID:AB_2632401), rabbit anti-HDAC1 (1:200; Abcam; Cat No: ab19845, RRID:AB_470299), rabbit anti-HDAC2 (1:200; Cell Signaling Technology; Cat No: #2540, RRID:AB_2116822), rabbit anti-HDAC4 (1:200; Cell Signaling Technology; Cat No: #2072, RRID:AB_2232915), and rabbit anti-HDAC5 (1:200; Cell Signaling Technology; Cat No: #20458, RRID:AB_2713973). The sections were then incubated with the appropriate secondary antibody. The secondary antibodies used in the present study included Alexa Fluor^TM 488 donkey anti-mouse IgG (H+L) (1:500; Invitrogen, Carlsbad, CA, United States; Cat No: A21202, RRID:AB_141607) and Alexa Fluor^TM 594 donkey anti-rabbit IgG (H+L) (1:500; A21207; Invitrogen; Cat No: A21207, RRID:AB_141637). 40, 6-diamidino-2-phenylindole (DAPI) (Abcam, Cat No: ab104139, LOT: GR3281289-1). The sections were examined in a blinded manner using a fluorescence microscope.

Western blotting was performed as we detailed previously (Zhou et al., 2014). First, after conducting sevoflurane anesthesia, DRGs (L3–5) (from 3 mice) or 1 spinal cord (the injured side of L3–L5 segments) were pooled together to yield a sufficient protein concentration and then homogenized in lysis buffer (100 μL/sample) containing phenylmethanesulfonyl fluoride (PMSF; Beyotime, Shanghai, China; Cat No: ST506-2). After resting on ice for 30 min, the samples were centrifuged at 13,000 rpm for 10 min at 4°C. We then measured total protein concentrations in supernatants using an Enhanced BCA Protein Assay Kit (Beyotime, Cat No: P0009). Equal amounts of protein (∼60 μg) were loaded into each well for separation by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE; 10% gels; Genshare Biology, Shanxi, China; Cat No: JC-PE022). After electrophoresis, proteins were transferred to polyvinylidene fluoride membranes (Millipore, Burlington, MA, United States; Cat No: IPVH00010). We then blocked membranes with nonfat dry milk for 1 h at room temperature, followed by overnight incubation at 4°C with the following primary antibodies: rabbit anti-MOR (1:1000, Abcam, Cat No: ab10275, RRID:AB_2156356), rabbit anti-MeCP2 (1:1000, #3456, Cell Signaling Technology, Cat No: #3456, RRID:AB_2632401), rabbit anti–acetyl-histone H3 (1:1000, Cell Signaling Technology, Cat No: #9649, RRID:AB_823528); rabbit anti-histone H3 (1:1000, Cell Signaling Technology, Cat No: #9715, RRID:AB_331563); rabbit anti-HDAC1 (1:200, Abcam, Cat No: ab19845, RRID:AB_470299), rabbit anti-HDAC2 (1:200, Cell Signaling Technology, Cat No: #2540, RRID:AB_2116822), rabbit anti-HDAC4 (1:200, Cell Signaling Technology, Cat No: #2072, RRID:AB_2232915), rabbit anti-HDAC5 (1:200, Cell Signaling Technology, Cat No: #20458, RRID:AB_2713973); and mouse anti-GAPDH (1:1000, Beyotime, Cat No: AF0006). The membranes were then incubated with horseradish peroxidase-conjugated secondary antibody for 1 h at room temperature. The acquired blots were analyzed accordingly. Membranes were exposed to X-ray film, and band densities were calculated using ImageJ 2.0.

After sevoflurane anesthesia, DRGs from 2 mice or 1 spinal cord were pooled together. We concentrated total RNA with TRIzol reagent (Invitrogen, Cat No: 15596-026) and synthesized cDNA with a Thermo Scientific Verso cDNA synthesis kit (ABgene; Thermo Scientific, Waltham, MA, United States; Cat No: RR036A). We performed qRT–PCR with a DyNAmo Flash SYBR Green qPCR kit (Thermo Scientific, Cat No: RR820). Thermal cycle conditions were set as the following: 95°C for 3 min; 40 cycles of 95°C for 15 s, 58°C for 30 s, 72°C for 30 s; and 72°C for 30 min. A StepOnePlus real-time PCR system (Thermo Fisher Scientific, Cat No: 2725213) was used for reactions. β-actin was used as a house-keeping gene for data normalization. For all primers, please refer to Supplementary Table 1.

After sevoflurane anesthesia, L3–L5 DRGs were harvested 7 days after CCI or sham surgery. DRGs from 6 mice and spinal cord from 3 mice, were separately pooled for DNA extraction (Qiagen, Hilden, Germany; Cat No: 69506). Methylation treatment of genomic DNA was carried out using a QiagenEpiTect Bisulfite Kit (Qiagen, Cat No: 59104) in accordance with the manufacturer’s instructions. A region of the Oprm1 gene promoter from –85 to +634 bp, which consists of 17 CpG sites, was divided into eight regions and amplified. For the pyrosequencing assay, after amplifying the bisulfite DNA with eight pairs of primers of the Oprm1 gene (Supplementary Table 2), binding buffer, streptavidin-sepharose beads, and PCR products were prepared for the binding reaction. The plate was then placed on a shaker for at least 15 min, but no longer than 1 h. Pyrosequencing was then performed. For the clone-sequencing assay, the same promoter fragment was amplified. PCR products were transformed into competent cells. Finally, PCR amplification and direct sequencing were carried on after an overnight bacterial culture (Taylor et al., 2007; Tost and Gut, 2007).

For siRNA injection, MeCP2 siRNA and its negative control (NC) siRNA were purchased from Genomeditech company. The three sequences were dissolved using Entranster^TM in vivo transfection reagents (Engreen Biosystem Co., Ltd., Auckland, New Zealand; Cat No: 18668-11-1) at a volume of 10 μL (100 pmol in total) for four continuous days intrathecally 30 min before surgery. In vivo transfection reagents served as a vehicle. 7 days after surgery, DRGs were collected. The siRNAs of MeCP2 and related primer duplexes are listed in Supplementary Table 3.

Recombinant adeno-associated virus 5 (AAV5)–MeCP2 [(Cat No: pSE3556, titer: 3.23 × 10¹³ VG/mL) (VG/mL means viral genome copy number per ml of virus solution)] and negative control AAV5-enhanced green fluorescence protein (EGFP; Cat No: Pmt423, Titer: 2.52 × 10¹³ VG/mL) were supplied by Shanghai SunBio Medical Biotechnology Co., Ltd. (Shanghai, China) and infused intrathecally (5 × 10¹⁰ VG in total) through the theca of the L5 and L6 segments of the spinal cord into the subarachnoid space. As previously reported (Lisowski et al., 2014), 30 days were needed to highly express full-length MeCP2 in the DRGs of mice.

Following the manufacturer’s instructions (EZ-Magna ChIP^TM A/G; Cat No: #17-10086) and a previous study (Huang et al., 2019), we performed ChIP assays. A homogenate of the injured side of L3–L5 DRGs of 10 mice was cross-linked with 1% formaldehyde, and glycine was then added to quench the reaction. After centrifugation and resuspension, the collected pellet was sonicated with eight 30-s pulses and 30-s intervals on ice to shear cross-linked DNA. After further centrifugation at 13,000 rpm for 10 min at 4°C, the supernatant was collected, and 10 μL was used as an input sample. Chromatin was added with protein G magnetic beads plus anti-MeCP2 antibody (1:20, Abcam, Cat No: ab2828, RRID:AB_2143853), anti-HDAC1 antibody (1:20, Abcam, Cat No: ab19845, RRID:AB_470299), or anti-HDAC2 antibody (1:20, Cell Signaling Technology, Cat No: #2540, RRID:AB_2116822), and incubated at 4°C overnight. Then protein G magnetic beads were collected and washed sequentially with low-salt buffer, high-salt buffer, LiCl buffer, and Tris–EDTA buffer. Protein/DNA complexes were eluted following the breaking of cross-links and reversed to free DNA. DNA fragments were identified using PCR with the appropriate primers. For all primers, please refer to Supplementary Table 4.

Co-immunoprecipitation was carried out as described previously (Sharma et al., 2018). L3–L5 DRGs from wild-type mice were collected in lysis buffer containing 50 mmol Tris (pH 8.0), 150 mmol NaCl, 0.1% SDS, 1.0% NP-40, protease inhibitor cocktail, and then homogenized with glass tissue grinders. After centrifugation, the supernatant was collected and then pre-incubated with 50 μL protein A/G mouse agarose (BBI Life Sciences, Shanghai, China; Cat No: 9012-36-6). After overnight incubation, 400 μg protein was incubated with rabbit anti-MeCP2 antibody (Abcam; Cat No: ab2828, RRID:AB_2143853) serum plus Protein A/G mouse agarose and mixed for 4 h at 4°C. After the mixture was washed 4 times with lysis buffer, proteins were eluted by boiling in loading buffer, then separated by SDS-PAGE (Genshare Biological, Cat No: JC-PE022), and finally transferred to a polyvinylidene difluoride membrane (Merck Millipore, Burlington, MSA, United States; Cat No: ISEQ00010) for western blotting. HDACs were detected as described above.

293T cells were purchased from American type culture collection (ATCC, Manassas). Cells were cultured in DMEM medium (Corning, Cat No: 10-013-CVR) with 10% FBS (Ausbian, Cat No: VS500T). 5% CO2 incubator at 37 °C with water-saturated atmosphere, and 12–12 h light/dark cycle environment was used for cell culture. Cells were transfected with MeCP2-OE or the corresponding negative control (OE-NC) via X-tremeGENE HP DNA transfection (ROCHE, Cat No: 06366236001).

As described in previous literatures (Zhang et al., 2018), the oprm1 promoter region (1316bp) was amplified using standard PCR protocol. PCR products were directionally inserted into the pGL3-Basic luciferase vector, and the DNA sequences were confirmed by direct sequencing. HEK293 cells (3 × 10⁴/well) were seeded and co-transfected MeCP2-OE or the corresponding negative control (OE-NC) with pGL3-Oprm1. Cells transiently transfected with the luciferase constructs were harvested at 48 h. The activities of firefly and Renilla luciferase in cell lysates were measured by a dual-luciferase reporter assay system (Promega, E1910) according to the manufacturer’s instructions. The firefly luciferase to renilla luciferase ratios were determined and defined as the relative luciferase activity. All experiments were performed in triplicate.

Both data analyses and chart production were performed using GraphPad Prism 7.00 and ImageJ 2.0. The results are expressed as mean ± standard error of the mean (SEM). Statistical analyses for 2 sets of data were performed using two-tailed unpaired Student’s t-test. A one-way or 2-way analysis of variance (ANOVA) with post hoc Tukey testing was used to analyze data from multiple groups. A statistical test result was considered significant when p < 0.05. Data was statistically analyzed in blinded studies. To estimate the group size required for the behavioral experiments, a pilot study was conducted to measure the PWL in 6 mice 7 days after surgery. The standard deviation was 1.62 in the sham (S₁) group and 2.22 in the CCI (S₂) group. The deviation (δ) between the two groups was 3.12 s. With α = 0.05 (Z_α/2 = 1.96), two-tailed test and power of 80% (Z_β = 0.842), following the formula below, we calculated the need for 12.2 (≈12) mice per group. The sample sizes in the molecular experiments are based on previous reports and pilot studies in this field (Li et al., 2020; Ma et al., 2021).

First, we examined the expression of MOR and MeCP2 after peripheral nerve injury in the DRG. The level of MeCP2 protein increased significantly while MOR decreased in the injured DRG from day 5 after unilateral chronic constrictive injury surgery. As expected, sham surgery did not change the basal expression of either MOR or MeCP2 proteins during the observation period (Figure 1A). Correlating with the changes in protein level, we also detected a higher level of MeCP2 mRNA and a lower level of MOR mRNA in the CCI group than those in the sham group on day 7 post-CCI (Figure 1B). Immunostaining further confirmed that the MOR protein was dramatically decreased in the ipsilateral DRG, but not in the sham DRG 7 days after CCI (Figure 1C). Additionally, by using triple labeling for MeCP2 or MOR, neuron (NeuN, a specific neuronal marker) or glutamine synthetase (GS, a marker for satellite glial cells), and DAPI (a marker for cellular nuclei), we found that MOR was predominantly located in neuronal cytoplasm (Figure 1Di), and undetected in the nuclei of GS-labeled DRG cells (Figure 1Dii), while MeCP2 was mostly co-expressed with neurons in the nuclei (Figure 1Diii) and a smaller minority of MeCP2 localized in the nuclei of GS-labeled DRG cells (Figure 1Div). We also analyzed the expression of MOR and MeCP2 on the injured side of L3–5 spinal cord 7 days after surgery. Unexpectedly, only protein and mRNA of MeCP2 showed overt increases, while MOR protein showed a decreasing tendency compared with that in the sham group (Figure 1E–G). Immunostaining also showed the reduced intensity of MOR signal in the ipsilateral spinal dorsal horn, but not in the sham dorsal horn (Figure 1H). Our findings indicate the reduction of MOR and increased expression of MeCP2 in the injured DRG and the ipsilateral side of the spinal cord.

Our prior work showed that a methylation inhibitor rescued the analgesia of morphine and that methylation of the Oprm1 gene promoter increased after nerve injury (Zhou et al., 2014). Therefore, we conducted pyrosequencing to analyze the Oprm1 gene promoter’s methylation status. We divided the 17 CpG sites in the 5′-flanking regions (from −85 to +634 bp) of the Oprm1 gene into 8 regions (Table 1) from three groups of animals: Sham plus vehicle group, CCI plus vehicle group, and CCI plus 5-aza-2′-deoxycytidine (5-aza-dC) group. We checked both the ipsilateral DRG and the spinal cord 7 days after surgery. Our work revealed that the methylation status of three of these eight regions (−40 to −39, 49 to 50, and 195 to 196) was higher in the CCI group, while these three regions were demethylated in the CCI plus 5-aza-dC group (Figure 3A). However, there were no obvious hypermethylation changes in the region of the Oprm1 gene promoter in the ipsilateral spinal cord after nerve injury, with or without 5-aza-dC treatment (Figure 3B). The eight sites are shown in Table 1 (the sites detected here are marked with superscript numbers 1–8 in the rectangular frame).

Considering that the methylation of CpG islands was evident in the DRG, we further focused on this area in our subsequent experiments. First, to ascertain whether blocking the increase in DRG MeCP2 could rescue MOR expression, MeCP2-specific short interfering RNA (MeCP2-siRNA, 100 pmol/10 μL, intrathecal infusion, i.t.) was used to knockdown MeCP2 via intrathecal infusion on day 0. Negative scramble siRNA was used in the control group. The analgesic effect of morphine was confirmed on day 7 via subcutaneous infusion with a dose of 1.5 mg per kg of body weight. As expected, intrathecal infusion of MeCP2 siRNA markedly increased the levels of both protein and mRNA of MOR, compared with the injection of scramble siRNA (Figures 4A,B,D). Meanwhile, the increased expression of MeCP2 in the mouse DRG was absent after the intrathecal infusion of MeCP2 siRNA (not scramble siRNA) at both transcription and translation levels on day 7 post-CCI (Figures 4A,C,E). Consistent with previous studies, CCI led to pain allodynia, but neither siRNA altered paw response to heat on the ipsilateral side of the CCI or sham group (Figure 4F). In addition, morphine analgesia was compromised 7 days after CCI surgery compared with that in the mice of the sham group; however, the injection of mice with MeCP2 siRNA, but not scramble siRNA, prevented this reduction of the analgesic effect of morphine, which was quantified by MPAE (Figure 4G). We further overexpressed MeCP2 in the DRG of normal mice, mimicking the process during neuropathic pain. AAV5 was injected intrathecally to express full-length MeCP2 (5 × 10^10VG, i.t.) into the subarachnoid space of L5 and L6 segments of the spinal cord. As expected, the overexpression of MeCP2 decreased the expression of MOR protein on day 36 post injection (Figures 4H–J). Simultaneously, inducing the expression of MeCP2 also inhibited MOR transcription (Figures 4K,L). From there findings, we concluded that MeCP2 could serve as a key regulator in reducing MOR in DRG after nerve injury.

As transcriptional regulators, HDACs are involved in regulating the acetylation of histones and can be recruited by MeCP2 to facilitate transcriptional repression (Nan et al., 1998; Haberland et al., 2009; Beumer and Tawbi, 2010). However, it was unclear whether the same situation would occur in the regulation of MOR expression and which HDAC(s) is involved in this process. We first measured the expressions of HDAC1, –2, –4, and 5, which were reported to be abundant in the nervous system (de Ruijter et al., 2003). Significant increases in the levels of HDAC1, –2, and 5 proteins were observed 7 days after nerve injury, whereas HDAC4 remained unchanged in the ipsilateral DRG (Figure 2A). Similar results were verified regarding the mRNA levels of HDACs on day 7 post-CCI (Figure 2B). Immunostaining further confirmed that the HDAC1, –2 and 5 proteins were dramatically increased in the ipsilateral DRG 7 days after nerve injury (Figure 2C). Interestingly, we observed overt nuclear translocation of HDAC5 after nerve injury (Figure 2Civ). Additionally, immunofluorescence images revealed that HDAC2 was predominantly localized in the nuclei of neurons (Figure 2Dii), whereas other HDACs localized in both nuclei and cytoplasm of DRG neurons in naïve mice (Figures 2Di,iii,iv). Furthermore, using a co-IP assay, we found that MeCP2 antibody could immunoprecipitate HDAC1 and HDAC2, but not HDAC4 or HDAC5, suggesting a specific interaction of MeCP2 with HDAC1 and HDAC 2 in the DRG (Figure 2E). We next performed ChIP assay and revealed that MeCP2 directly bound to sites of Oprm1 promoter regions (21 to 112, 100 to 234, 206 to 343, and 398 to 511 bp) as the immunoprecipitation of this region (regions from −85 to +634 bp) with MeCP2 antibody showed (Figures 2F,G). Consistently, the dual luciferase assay confirmed the activity of that MeCP2 inhibited the activity of Oprm1 promoter by about 29% (Figure 2H). We further examined the acetylation level of histone H3, as well as HDAC1 and HDAC2 levels, in the promoter region of MOR in both CCI and sham groups. As presented in Figure 2I, HDAC1 but not HDAC2 was detected in this region. The binding activities of HDAC1 in three of the four regions (21 to 112, 100 to 234, and 398 to 511 bp) increased by 1.26-, 1.31-, and 1.21-fold, respectively, in the injured DRG on day 7 after CCI, compared with the levels after sham surgery (Figures 2I,J), while acH3 decreased in the CCI group in these three regions, which was consistent with our expectations (Figures 2I,K).

Next, we determined whether HDAC blockade could alter the expression of MOR and the analgesic effects of morphine. We corroborated these findings by intraperitoneally administering of SAHA (50 mg/kg, i.p.), a pan-inhibitor of HDACs Neither SAHA nor vehicle altered paw responses to heat stimuli on the ipsilateral side of CCI mice and sham mice (Figure 5A). Like in MeCP2 siRNA treated mice, SAHA inhibition was associated with a greater analgesic effect of morphine than in the vehicle group on the ipsilateral side of CCI mice (Figure 5B). Additionally, both protein and mRNA levels of MOR expression were reduced in injured DRG on day 7 post-CCI, which were rescued by SAHA injection (Figures 5C,D,G). The increased expression of HDAC1 was also observed, which was not affected by the infusion of SAHA, as SAHA could only inhibit its activity rather than its expression (Figures 5C,E,H). Concurrently, we detected a global decrease of acetylation of histone H3 in the injured DRG, while SAHA prevented this reduction (Figure 5F). Our results indicate that MOR downregulation could be rescued by SAHA, leading to an enhanced analgesic effect of morphine in neuropathic pain mice.