Epigenetic upregulation of metabotropic glutamate receptor 2 in the spinal cord attenuates estrogen-induced visceral hypersensitivity

Animals

Experimental protocols were approved by the University of Maryland School of Medicine Institutional Animal Care and Use Committee and adhered to guidelines for experimental pain in animals published by the International Association for the Study of Pain. Female Sprague-Dawley rats weighing 225–250 g were obtained from Harlan. Rats were housed in pairs with free access to food and water at 25°C with 12 h/12 h alternating light/dark cycle.

Surgery

Rats were anesthetized with 55 mg/kg ketamine, 5.5 mg/kg xylazine, and 1.1 mg/kg acepromazine. Ovariectomies (OVx) were performed by a dorsolateral approach. In order to direct drugs to the spinal cord, the atlanto-occipital membrane was slit and a polyethylene catheter (32 g, ReCathCo, Allison Park, PA, USA) was inserted in the subdural space to the level of the lumbosacral spinal cord (L6-S2). After catheter placement, electromyogram (EMG) electrodes made from Teflon-coated 32 gauge stainless steel wire (Cooner Wire Company, Chatsworth, CA, USA) were stitched into the ventrolateral abdominal wall. The electrode leads were tunneled subcutaneously and exteriorized at the back of the neck with the catheter. Rats were individually housed and allowed 10–14 days to recover from surgery.

Visceromotor response (VMR)

Rats were fasted for 18–24 hours prior to recording the VMR. Water was available ad libitum. On the day of testing, rats were briefly sedated with isoflurane and a 5–6 cm balloon attached to Tygon tubing was inserted into descending colon and rectum through the anus. The secured end of the balloon was at least 1 cm proximal to the external anal sphincter and the tubing was taped to the tail. Rats were loosely restrained in acrylic tubes and given 30 min to recover from sedation. The EMG signals were recorded with a CED 1401 and analyzed using Spike 2 for windows software (Cambridge Electronic Design, Cambridge, UK). Colorectal distention trials (CRD; each trial was 5 distentions to 60 mmHg CRD, 20 sec duration, 3 min interstimulus interval) was produced by inflating the distention balloon with air under computer control. The recorded EMG signal was rectified in Spike 2 and the area under the curve (AUC) calculated. The response for each distention stimulus was the AUC of the EMG for the 20 s before distention (background) subtracted from the AUC during the 20 s distention. The responses of the 5 distentions during a trial were averaged for the response for that trial.

Tissue collection

Rats were euthanized with CO₂ and decapitated at selected time points after treatment. The spinal cord was removed by pressure ejection with ice cold saline as described previously ¹⁴. Using the lumbar enlargement as a landmark, the L6 to S2 section of the spinal cord was isolated. The dorsal part of isolated spinal cord was dissected, frozen on dry-ice and saved at −80°C until use.

Western blots analysis

Tissues were homogenized in RIPA buffer (50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 1 mM EDTA, 1% NP-40, 0.5% deoxycholic acid, 0.1% SDS) supplemented with 1 mM Na₃VO₄ and protease inhibitor cocktail (Roche, Branford, CT, USA). The homogenates were centrifuged at 14,000×g for 10 min at 4°C and the supernatant was collected. Protein contents in supernatants were measured using the Bradford method. After denaturing protein samples were fractionated 25 µg per lane on 4–12% SDS-NuPAGE gel and blotted to nitrocellulose membrane. The membranes were incubated with primary antibody (1:1,000) directed against acetylated lysine 9 on histone 3 (H3K9ac, Cell Signaling Technology, Danvers, MA, USA) or mGluR2 (Abcam Inc, Cambridge, MA, USA) at 4°C overnight. The membranes were further incubated for 1 hour in blocking buffer with IRDye 800CW goat anti-rabbit secondary antibody (Li-Cor Biosciences, Lincoln, NE, USA). Washed membranes were scanned and analyzed with the Li-Cor Odyssey System (Li-Cor Biosciences, Lincoln, NE, USA). Blots were then stripped for 30 min and reprobed with antibodies for pan-histone 3 (1:1000, Cell Signaling Technology) or β-actin primary antibody (1:5,000, Sigma, St. Louis, MO, USA) and IRDye 800CW goat anti-rabbit secondary antibody or IRDye 680LT goat anti-mouse secondary antibody (1:20,000, Li-Cor Biosciences) with same processes as above.

Quantitative reverse transcription polymerase chain reaction (qRT-PCR)

Total RNA was extracted from the dissected L6–S2 dorsal spinal cord using absolutely RNA miniprep kit (Stratgene, La Jolla, CA, USA) and reverse transcribed into cDNAs using SuperScript II Reverse Transcriptase kit with random primers (Invitrogen, Grand Island, NY, USA). Quantification of rat mGluR2 and mGluR3 mRNAs were completed by SYBR green-based real-time PCR (qPCR). PCR primers were designed based on mRNA sequences to cross at least one Exon/Exon junction (Genbank accession number of NM_001105711 for mGluR2 and NM_001105712 for mGlu3) using Primer3 software. Three pairs of primers per mRNA were evaluated and one each (mGluR2, forward, 5′-CGTGAGTTCTGGGAGGAGAG, reverse, 5′- GCGGACCTCATCGTCAGTAT; mGluR3, forward, 5′-GTGGTCTTGGGCTGTTTGTT, reverse, 5′-GCAGCATGTGAGCACTTTGT) with comparable PCR efficiency to that of the GAPDH mRNA (forward, 5′-TCACCACCATGGAGAAGGC, reverse, 5′-GCTAAGCAGTTGGTGGTGCA) was chosen. qPCR was completed in Maxima SYBR Green/Rox qPCR Master Mix (Thermo Scientific, Waltham, MA, USA) on the Eppendorf Mastercycler Real-plex system (Eppendorf, Hauppauge, NY, USA) and C_T values were obtained from the system. The efficiency of PCR was calculated from the slope of the standard curve of serially diluted cDNA and was found within the range of 90–110%. Relative level of mRNA was calculated using the ΔΔC_T method after normalization to GAPDH mRNA as described previously ¹⁵.

Chromatin immunoprecipitation (ChIP)

ChIP was performed using a previously described method with the following modifications ¹⁶. Half of the dissected spinal dorsal horn was minced in 1xPBS (pH7.4) supplemented with protease inhibitors and subjected to DNA-protein cross-link by incubation with 1.5% paraformaldehyde for 20 min at room temperature. After homogenization on ice, the homogenate was centrifuged and the resulting pellet incubated in swelling buffer (25 mM Hepes, pH 7.9, 10 mM KCl, 1.5 mM MgCl₂, 0.5% NP-40) on ice for 15 min and subjected to a second homogenization. Nuclei were isolated by centrifugation, broken in lysis buffer (50 mM Tris-HCl, pH 8.0, 10 mM EDTA, 1% SDS, protease inhibitor cocktail, 100 µg/ml valproic acid) and sonicated in the Covaris M2 system to yield DNA fragments around 500 bp. For immunoprecipitation, 3 µg of polyclonal antibodies specific to H3K9ac or H3K18ac (Cell Signaling Technology, Danvers, MA, USA) or to ERα (Millipore, Temecula, CA, USA) were incubated overnight with 200 µg protein at 4°C. The antibody-antigen complexes were pulled down by protein A/G-agarose followed by successive washes once each with low salt buffer (20 mM Tris-HCl, pH 8.0, 150 mM KCl, 2 mM EDTA, 0.1% SDS, 1% Triton X-100), high salt buffer (the same as low salt buffer except 500 mM KCl), LiCl buffer (10 mM Tris-HCl, pH 8.0, 250 mM LiCl, 1 mM EDTA, 1% NP-40, 1% sodium dexoycholate) and twice in TE buffer (pH 8.0). The complex was eluted into solution of 100 mM NaHCO₃-1% SDS. After reverse cross-link in 250 mM NaCl at 65°C for 6 ~12 hours, DNA was purified by a Qiaquick column (Qiagen, Frederick, MD, USA). Lysate equal to 5% of the immunoprecipitate was directly subjected to reverse cross-link and Qiaquick column purification without antibody immunoprecipitation and served as input. Two percent of eluted DNA was subjected to qPCR with SYBR green (Thermo Scientific, Waltham, MA, USA) in duplicate using primers covering five different regions in the GRM2 promoter (figure 4A). qPCR was completed on the Eppendorf RealPlex2 system with a program of initial denaturing at 95°C for 4 min followed by cycling at 95°C for 15 sec and 57~64°C for 30 sec.

Two negative controls were performed. First, in a mock ChIP, IgG purified from non-immunized rabbit serum (VWR, Philadelphia, PA, USA) was added to replace antibody during immunoprecipitation for initial studies of ChIP specificity. Second, a 187-bp intergenic locus in rat chromosome 4 was examined by qPCR from ChIP elutes and inputs of every sample, which was then used for calculation. Primers used for this negative locus are forward, 5′-TGCAAACTTCACACGTCCTC, and reverse, 5′-GGGGTGGAAATTTCTGGTTT. Antibody-enriched experimental DNA over the negative control was calculated using the ΔΔCt method with a normalization of immunoprecipitation data to relevant input of each sample for the promoter regions and the intergenic locus ¹⁷ (supplemental figure 2). Six rats per group were tested for each pair of primers in ChIP.

Hormones and drugs

E2 was dissolved in safflower oil to a concentration of 50 µg/100 µl. E2 (100 µl) or safflower oil was injected subcutaneously. SAHA (Cayman Chemical, Ann Arbor, MI, USA), made fresh, was dissolved in 20% dimethyl sulfoxide (DMSO) to a final concentration of 5–80 µg/20 µl. The HDAC inhibitor trichostatin A (TSA) and mGluR2/3 antagonist LY341495 (Tocris Bioscience, Bristol, UK) were also dissolved in 20% DMSO. For intrathecal injection (i.t.), 20 µl volume of drug was injected followed by a 10 µl saline flush.

Data analysis

All data are expressed as mean ± SEM. Data were analyzed by one or two way ANOVA as appropriate. Significant ANOVA results were further analyzed using Bonferroni’s multiple comparison test and reported in the figures. t-test was used for two group comparison. p < 0.05 was considered significant.

HDAC inhibitors attenuate the E2-faciliated VMR and increase histone acetylation

Ten to fourteen days following ovariectomy the baseline VMR was recorded. Rats were subsequently injected with E2 or safflower oil and the VMR recorded again 4 hours later. Rats were then injected with SAHA or vehicle and the VMR recorded over the next 2 hours. There was no difference in the baseline response (labeled OVx in figure 1A, one way ANOVA, p > 0.05) between treatment groups. Following E2, but not oil injection, there was a significant increase in the magnitude of the VMR as previously reported (t-test, p < 0.05; figure 1C) ^3,4.

The HDAC inhibitor SAHA or its vehicle DMSO was then intrathecally injected. The VMR trial was recorded starting 30, 60, 90, and 120 min after injection of SAHA/vehicle. SAHA at doses of 20 and 80 µg reversed the E2-facilitated VMR for at least 2 hours (one way RM ANOVA for each treatment group, p < 0.0001 and 0.01, respectively; figure 1A). A lower dose of SAHA (5 µg) or vehicle (20% DMSO) had no effect on the E2-facilitated VMR (one way RM ANOVA, p > 0.05 for both). In the absence of E2 (safflower oil treated ovariectomized rats), SAHA showed no effect on the magnitude of the VMR compared to the OVx or post oil injected state (one way RM ANOVA, p > 0.05).

Within each treatment group the response post SAHA remained stable over the 2 hour recording period so the response was averaged and normalized to the E2 or oil response prior to SAHA/DMSO injection. Comparison between treatment groups show that 20 and 80 µg SAHA significantly attenuated the magnitude of the VMR compared to 5 µg SAHA or DMSO in E2-treated rats or 80 µg SAHA in oil-treated rats (one way ANOVA, p < 0.005, figure 1D).

The reversal of the E2-faciliated VMR by HDAC inhibition was confirmed by i.t. administration of another HDAC inhibitor, TSA (20 µg, one way RM ANOVA, p < 0.005, supplemental figure 1). A lower dose of TSA (5 µg) had no effect (one way RM ANOVA, p > 0.05). Similar to SAHA, TSA had no effect on the VMR in the absence of E2 (one way RM ANOVA, p > 0.05).

Western blots were used to confirm that SAHA inhibited HDAC and increased global histone acetylation in the spinal cord. Rats were treated with E2 or safflower oil for 4 hours followed by SAHA (80 µg) or DMSO treatment for 30 min or 2 hours before tissues were collected for Western blot analysis. SAHA significantly increased H3K9ac in E2-treated animals compared to the E2 + DMSO (two way ANOVA, p < 0.001 at 30 min post SAHA; p < 0.01 at 2 hours post SAHA, figure 2). SAHA had no effect on histone acetylation in safflower oil treated animals compared to the oil + DMSO (p > 0.05 at both 30 min and 2 hours post SAHA). Interestingly, E2 alone did not alter the level of histone acetylation (E2 + DMSO group vs oil + DMSO group, p > 0.05 at both 30 min and 2 hours post DMSO).

These data show that E2-induced visceral hypersensitivity is attenuated by spinal administration of HDAC inhibitors. In addition, the HDAC inhibitor increased global histone acetylation in the presence of E2.

HDAC inhibitor attenuation of the E2-facilitated VMR is mGluR2 mediated

The simplest explanation for the inhibitory effect of SAHA on visceral sensitivity is an increase in spinal inhibitory processing. Considering that HDAC inhibitors attenuated nociceptive responses in the formalin test through upregulation of mGluR2 ¹¹, we tested the hypothesis that SAHA upregulates mGluR2 to strengthen inhibitory processing and in turn attenuate visceral hypersensitivity. Results of qRT-PCR revealed that the mGluR2 mRNA level significantly increased 2 hours after SAHA (80 µg) injection in E2-treated rats (two way ANOVA, p < 0.05; figure 3A). In comparison, injection of SAHA following oil treatment at that time point had no effect on mGluR2 mRNA. There were no significant differences in mGluR2 mRNA among groups at 30 min and 1 hour after SAHA injection (two way ANOVA, p > 0.05). Consistent with the mRNA change, mGluR2 protein expression also significantly increased 2 hours after SAHA injection in E2-treated rats in comparison to DMSO treated rats (two way ANOVA, p < 0.05, figure 3B). However, no difference in the expression level of mGluR3 mRNA at any time point (figure 3C) or protein at 2 hours (not shown) was revealed after SAHA administration (two way ANOVA, p > 0.05).

Since the transcription rate is sensitive to histone acetylation and directly impacts mRNA levels, we applied ChIP to examine the interaction between acetylated histones to the GRM2 promoter following E2 + SAHA treatment. The rat GRM2 promoter was defined for the region upstream of and surrounding the transcriptional start site (TSS) or the first base of exon 1 which has 397 bp sequences as provided by the University of California Santa Cruz genome database (genome.ucsc.edu). We searched putative cis-elements within ±2Kb of the TSS using TRANSFAC software (BIOBASE Biological Database, Qiagen, Frederick, MD, USA) and designed primers to cover five regions containing motifs for potential positive transcription factors (figure 4A). After optimizing the PCR program, we were able to obtain a single band of the expected size from each pair of primers from gDNA precipitated with an antibody against H3K9ac in ChIP of naïve animals (figure 4C). The specificity of these primers was also confirmed by the single peak in melting curves of SYBR green qPCR (supplemental figure 3). In comparison, these primers were unable to generate signal or band from negative control of IgG mock ChIP elute (figure 4C). We also examined an intergenic locus as negative control for calculation of ChIP results. This locus does not contain binding sites for estrogen receptors, and exhibited a low binding to acetylated histone 3 (Supplemental figure 2). Our results demonstrate that E2 + 80 µg SAHA treatment significantly increased binding of H3K9ac to three regions (A1, 4, 5) of the GRM2 promoter (t-test, p < 0.05) while binding of H3K9ac to the other two regions and binding of H3K18ac to all five regions revealed no significant change compared to E2 + DMSO (figure 4D).

Interestingly, atypical or half estrogen response elements (aERE) were found in the A2 (TGACC, -1106/-1102), A4 (GTGAC, -503/-499) and A5 (GGGTCA, 1133/1138) regions of the GRM2 promoter. It has been reported that more than 50% of estrogen receptor-bound sequences are atypical sites ¹⁸. To examine whether these aEREs are functional, we performed ChIP with a polyclonal antibody against ERα and found that regions A2, A4 and A5 were precipitated by this antibody (figure 4E). After normalization to the input, A5 had the highest binding capacity. Furthermore, similar to the increase in H3K9ac binding, SAHA significantly increased ERα binding to A4 and A5 (t-test, p < 0.05). In comparison, the low binding level of ERα to A2 was not altered by SAHA, consistent with no change in H3K9ac binding to this region. These data indicate that hyperacetylation of K9 in histone 3 is associated with increased binding of ligand-bound ERα to the GRM2 promoter and with upregulation of GRM2 transcription.

Due to the observed increase in mGluR2 expression and the binding of H3K9ac and ERα to the GRM2 promoter in E2-SAHA treated animals, we next determined whether there was any functional significance of these changes. Thirty minutes following SAHA (80 µg) injection and establishing attenuation of the E2-faciliated VMR, the mGluR2/3 antagonist LY341495 (20 nmol, i.t.) was injected (figure 5). Starting 5 minutes later the VMR was recorded over the next hour. LY341495 gradually reversed the attenuation of the VMR evoked by SAHA, and the magnitude of the VMR was significantly increased at 35 and 50 min after LY341495 application (i.e., 105 and 120 min after SAHA application) compared with the VMR at same time point in E2 plus SAHA treated animals receiving DMSO (vehicle for LY341495)(two way RM ANOVA, p < 0.01 for treatment; figure 5).