RE1-silencing transcription factor controls the acute-to-chronic neuropathic pain transition and Chrm2 receptor gene expression in primary sensory neurons

Neuropathic pain is associated with persistent changes in gene expression in primary sensory neurons, but the underlying epigenetic mechanisms that cause these changes remain unclear. The muscarinic cholinergic receptors (mAChRs), particularly the M2 subtype (encoded by the cholinergic receptor muscarinic 2 (Chrm2) gene), are critically involved in the regulation of spinal nociceptive transmission. However, little is known about how Chrm2 expression is transcriptionally regulated. Here we show that nerve injury persistently increased the expression of RE1-silencing transcription factor (REST, also known as neuron-restrictive silencing factor [NRSF]), a gene-silencing transcription factor, in the dorsal root ganglion (DRG). Remarkably, nerve injury–induced chronic but not acute pain hypersensitivity was attenuated in mice with Rest knockout in DRG neurons. Also, siRNA-mediated Rest knockdown reversed nerve injury–induced chronic pain hypersensitivity in rats. Nerve injury persistently reduced Chrm2 expression in the DRG and diminished the analgesic effect of muscarine. The RE1 binding site on the Chrm2 promoter is required for REST-mediated Chrm2 repression, and nerve injury increased the enrichment of REST in the Chrm2 promoter in the DRG. Furthermore, Rest knockdown or genetic ablation in DRG neurons normalized Chrm2 expression and augmented muscarine's analgesic effect on neuropathic pain and fully reversed the nerve injury–induced reduction in the inhibitory effect of muscarine on glutamatergic input to spinal dorsal horn neurons. Our findings indicate that nerve injury–induced REST up-regulation in DRG neurons plays an important role in the acute-to-chronic pain transition and is essential for the transcriptional repression of Chrm2 in neuropathic pain.

Chronic neuropathic pain is a debilitating condition caused by damage to the somatosensory system. Persistent up-regulation of pronociceptive genes and down-regulation of antinoci-ceptive genes account for the long-lasting aberrant excitability of dorsal root ganglion (DRG) 2 neurons in chronic neuropathic pain (1)(2)(3). For example, nerve injury causes up-regulation of ␣2␦-1 in the DRG (4,5), which leads to increased synaptic N-methyl-D-aspartate receptor activity and nociceptive input to spinal dorsal horn neurons. On the other hand, nerve injury reduces the expression of voltage-gated K ϩ channels (2, 6), contributing to increased excitability of DRG neurons. However, the mechanisms underlying the transition from acute to chronic pain after nerve injury remain unclear. Thus, identifying the epigenetic and transcriptional factors responsible for altered gene expression in injured DRG neurons is important for understanding the mechanisms of neuropathic pain development and for improving the treatment of chronic neuropathic pain.
The repressor element-1 (RE1)-silencing transcription factor (REST; also known as neuron-restrictive silencer factor (NRSF)) is a zinc-finger DNA-binding protein and a master regulator of gene expression during neuronal development (7). REST binds to the RE1 site, which is composed of specific consensus DNA sequences, and inhibits transcription of its target genes by recruiting the cosuppressors mSin3 and coREST. These co-repressor complexes epigenetically modify target genomic regions through several chromatinmodifying enzymes (8 -10). In nerve injury-induced neuropathic pain, REST is involved in silencing the expression of certain K ϩ channels and -opioid receptors (Oprm1) in the DRG (6,11). Nevertheless, the function of REST and its target genes in the injured DRG remain to be determined.
The muscarinic acetylcholine receptors (mAChRs) are G protein-coupled receptors that produce powerful analgesic effects upon activation. Spinally administered mAChR agonists or acetylcholinesterase inhibitors reduce pain in both animal models and humans (12)(13)(14). Furthermore, blocking mAChRs at the spinal cord level causes pain hypersensitivity, indicating that mAChRs tonically control nociceptive transmission (15). Among the five mAChRs, encoded by Chrm1-5, the odd-number group (M1, M3, and M5) and the even-number group (M2 cro ARTICLE and M4) preferentially bind to G q/11 and G i/o proteins, respectively (16,17). In the DRG, about 75% of mAChR transcripts are Chrm2 (2). The M2 subtype is also abundantly expressed at primary afferent terminals in the spinal dorsal horn (12,18) and predominantly mediates the analgesic effect of mAChR agonists (12, 14, 19 -21). However, little is known about how the expression of Chrm2 in the DRG is regulated in neuropathic pain.
In this study, we provide new evidence that REST in DRG neurons not only contributes to the transition from acute to chronic pain after nerve injury but is also responsible for nerve injury-induced Chrm2 repression and diminished muscarinic analgesia in neuropathic pain. Thus, REST-dependent epigenetic remodeling is a central mechanism involved in regulating Chrm2 expression in primary sensory neurons and in neuropathic pain development.

REST expression is increased in the injured DRG and maintains chronic pain induced by nerve injury
REST is a key transcriptional repressor that is normally expressed at a low level in differentiated neurons (22). To determine how nerve injury affects Rest expression in the DRG, we performed L5 and L6 spinal nerve ligation (SNL) in rats, a commonly used rodent model of neuropathic pain (23,24). Quantitative PCR analysis showed that the mRNA level of Rest in the DRG was significantly higher in SNL rats than in sham-operated control rats on days 5, 10, and 21 after surgery (p Ͻ 0.0001, F(1,30) ϭ 120.5, n ϭ 6 rats/group; Fig. 1A). In contrast, SNL had no significant effect on the mRNA level of Rest in the dorsal spinal cord (Fig. 1B).
To determine whether REST plays a role in nerve injuryinduced pain hypersensitivity, we first used an siRNA to knock down Rest in rats subjected to SNL. Three weeks after SNL, intrathecal injection of 2 g/day of Rest-specific siRNA, but not control siRNA, for 5 days fully restored the mRNA level of Rest in the injured DRG to that observed in the DRG of rats that underwent sham surgery (p Ͻ 0.0001, F(2,15) ϭ 56.64, n ϭ 6 rats/group; Fig. 1C). However, treatment with Rest-specific siRNA had no effect on the mRNA level of REST-corepressor 1 (Rcor1) in the DRG (Fig. 1C). Western blotting showed that treatment with Rest-specific siRNA restored the protein level of REST in the injured DRG to that found in sham control rats (p Ͻ 0.0001, F(2,15) ϭ 23.24, n ϭ 6 rats/group; Fig. 1D).
SNL rats treated with control siRNA displayed a sustained reduction in the withdrawal threshold of the ipsilateral hind paw in response to tactile and noxious pressure stimuli (Fig. 1, E and F). By comparison, treatment with Rest-specific siRNA (2 g/day) significantly attenuated tactile allodynia and mechanical hyperalgesia caused by SNL (n ϭ 9 rats/group; Fig. 1, E and F). These results suggest that nerve injury increases REST expression in the DRG and that REST up-regulation is involved in maintaining chronic neuropathic pain.

REST in DRG neurons contributes to chronic but not acute pain induced by nerve injury
To further elucidate the role of REST in DRG neurons in the development of chronic pain after nerve injury, we generated Rest conditional knockout mice (Rest-cKO) by crossing Restfloxed mice (25) with a primary sensory neuron-specific Cre mouse line, Advillin-Cre (26,27), such that Rest was selectively ablated from DRG neurons. The ambulatory behavior and motor coordination, measured with the rotarod test, were similar in Rest-cKO mice and WT littermate controls ( Fig. 2A). The protein and mRNA levels of REST in the DRG were diminished in Rest-cKO mice compared with those in WT controls (n ϭ 8 mice/group; Fig. 2, B and C). Furthermore, spared nerve injury (SNI) significantly increased the mRNA level of Rest in the DRG in WT (p ϭ 0.0004, F (1.84,9.20) ϭ 179, n ϭ 6) but not in Rest-cKO mice (Fig. 2D).
In WT mice, SNI induced a persistent reduction in the tactile and pressure withdrawal thresholds of the ipsilateral hind paw that lasted for at least 4 weeks (n ϭ 8 mice; Fig. 2, E and F). The withdrawal thresholds did not differ significantly between WT and Rest-cKO mice during the first week after SNI. Remarkably, the nerve injury-induced reduction in the withdrawal thresholds was significantly attenuated in Rest-cKO mice 8 to 29 days after SNI (Fig. 2, E and F). These data indicate that REST-dependent epigenetic remodeling in DRG neurons contributes to the transition from acute to chronic pain after nerve injury.

Nerve injury causes a persistent reduction in Chrm2 expression and increases REST occupancy at the Chrm2 promoter in the DRG
Chrm2 is the most dominant mAChR subtype expressed in the DRG (2,19). Quantitative PCR analysis showed that 5 days after surgery, the mRNA level of Chrm2 in the DRG was significantly lower in SNL rats than in control rats subjected to sham surgery (p Ͻ 0.001, F(1,10) ϭ 619.4, n ϭ 6 rats/group, Fig. 3A). This nerve injury-induced Chrm2 down-regulation lasted for at least 3 weeks after surgery (Fig. 3A). However, the mRNA level of Chrm2 in the dorsal spinal cord was not significantly different between SNL and sham control rats on days 5, 10, and 21 after surgery (Fig. 3B). Also, the mRNA level of Chrm4 in the DRG did not differ significantly between the SNL and sham control groups (Fig. 3C).
We next used ChIP followed by quantitative PCR to determine whether nerve injury affects the amount of REST in the promoter region of Chrm2 in the DRG. In DRG tissues from sham control rats, the REST occupancy at the Chrm2 promoter was very low and similar to that precipitated by the IgG-negative control (Fig. 3D). However, SNL caused a large increase in the enrichment of REST in the promoter region of Chrm2 in the DRG (p ϭ 0.0002, t(10) ϭ 5.65, n ϭ 6 rats/group; Fig. 3D). In contrast, the level of REST in the promoter region of Gapdh, a highly expressed gene in the DRG, did not differ significantly between the SNL and control groups (Fig. 3E). These results indicate that nerve injury increases REST enrichment at the Chrm2 promoter in the DRG.

The RE1 site is essential for transcriptional repression of Chrm2 by REST
DNA sequences containing the RE1 site typically recruit the binding of REST and are associated with repression of gene expression (7). We initially used Matinspector, a software tool (28), to search for transcription factor binding sites to the pro-

Transcriptional regulation in neuropathic pain
moter region of rat Chrm2. We found an RE1-binding site (taTTCAGgAtacagaagaaGGACtGggaagt; uppercase letters indicate conserved sequences) near the transcription start site (between nucleotides Ϫ150 and Ϫ119) of Chrm2 (Fig. 4A). We then used a luciferase reporter assay to determine directly whether the RE1 site is involved in suppression of Chrm2 expression by REST. In HEK293 cells transduced with Chrm2 promoter/luciferase expression vectors, luciferase activity was significantly lower in cells with the Chrm2 promoter containing the RE1 sequence than in cells without the RE1 sequence (p ϭ 0.0286, n ϭ 4; Fig. 4B).
We repeated the luciferase reporter assay using the F11 cell line, which is a hybrid of a rat embryonic DRG neurons with mouse neuroblastoma cells (29). Luciferase activity was also significantly lower in cells transduced with the Chrm2 promoter containing the RE1 sequence than in F11 cells with the RE1 sequence-deficient Chrm2 promoter (p ϭ 0.0286, n ϭ 4; Fig. 4B). These data indicate that the Chrm2 RE1 site is necessary and sufficient to repress Chrm2 promoter activity.

REST is critically involved in nerve injury-induced repression of Chrm2 in the DRG and the diminished analgesic effect of muscarine
We used a Rest-specific siRNA in SNL rats to determine whether REST is responsible for nerve injury-induced transcriptional repression of Chrm2 in the DRG. Intrathecal treatment with the Rest-specific siRNA fully restored the mRNA Gapdh was used as an endogenous control, and the mean value in the sham group was set to 1 (n ϭ 6 rats/group). C, real-time PCR data showing the mRNA levels of Rest and Rcor1 in the DRG of sham control and SNL rats treated intrathecally with control (CT) siRNA or Rest-specific siRNA for 5 days (n ϭ 6 rats/group). ***, p Ͻ 0.001 (one-way ANOVA). D, original immunoblotting images and quantification of the REST protein level in the DRG in sham control and SNL rats treated with control siRNA or Rest-specific siRNA for 5 days (n ϭ 6 rats/group). GAPDH was used as an internal control. M, molecular weight marker. E and F, time course of changes in the tactile (E) and noxious pressure (F) withdrawal thresholds in sham control rats and SNL rats treated with control siRNA or Rest-specific siRNA for 5 days (n ϭ 9 rats/group). Data are shown as means Ϯ S.D. (A-D) or means Ϯ S.E. (E and F). *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001 compared with the sham group at the same time point (two-way ANOVA).
Because M2 plays a major role in mediating mAChR agonistinduced analgesia (19,21), we examined whether nerve injuryinduced Chrm2 repression in DRG neurons influences the analgesic effect of muscarine in neuropathic pain. Treatment with Rest-specific siRNA only slightly augmented the analgesic effect of intrathecal injection of muscarine in sham control rats (Fig. 5, D and E). Compared with the sham control group, intrathecal injection of 2 g of muscarine in SNL rats produced a minimal analgesic effect on tactile allodynia and mechanical hyperalgesia (n ϭ 6 rats/group; Fig. 5, D and E). Strikingly, the effect produced by intrathecal injection of 2 g of muscarine on pain hypersensitivity was much larger in SNL rats treated with Rest-specific siRNA than in SNL rats treated with control siRNA (Fig. 5, D and E).

REST in DRG neurons is required for the nerve injury-induced repression of Chrm2 and the diminished analgesic effect of muscarine
We then used Rest-cKO mice to confirm the role of REST in DRG neurons in the diminished analgesic effect of muscarine in nerve injury-induced neuropathic pain. In WT mice, SNI significantly reduced the mRNA (p ϭ 0.0003, F(3,20) ϭ 9.93, n ϭ 6 mice/group) and protein (p ϭ 0.0025, F(3,20) ϭ 6.77, n ϭ 6 mice/group) levels of M2 in the DRG (Fig. 6, A and B). In contrast, in Rest-cKO mice, SNI failed to significantly reduce the mRNA and protein levels of M2 in the DRG (Fig. 6, A and B). The mRNA and protein levels of M2 in the injured DRG of Rest-cKO mice were comparable with those in the DRG of sham-operated WT mice.
In WT mice subjected to SNI, intrathecal injection of 0.2 g and 0.5 g of muscarine produced only a minimal effect on pain

Transcriptional regulation in neuropathic pain
hypersensitivity. By comparison, the analgesic effect of intrathecally administrated muscarine at both 0.2 g and 0.5 g was markedly enhanced in Rest-cKO mice subjected to SNI (n ϭ 8 mice/group; Fig. 6, C-F). Collectively, these findings indicate that REST in DRG neurons is responsible for nerve injuryinduced transcriptional repression of Chrm2 and the diminished analgesic effect of muscarine in neuropathic pain.

REST in DRG neurons mediates the reduced muscarinic inhibition of glutamatergic input to spinal dorsal horn neurons in neuropathic pain
At the spinal cord level, M2 is predominately expressed at the central terminals of primary afferent nerves and plays a critical role in regulating glutamate release to spinal dorsal horn neurons (14,18,30). We performed electrophysiological recordings in spinal cord slices to determine the role of REST-dependent expression of presynaptic M2 in the control of glutamate release from primary afferent nerve terminals after nerve injury. We recorded excitatory postsynaptic currents (EPSCs) monosynaptically evoked by electrical stimulation of the dorsal root, which reflect glutamate release from central terminals of primary afferent nerves (18,30). Bath application of muscarine at concentrations of 0.1 to 10 M reduced the amplitude of evoked EPSCs of lamina II neurons of sham control rats treated with control siRNA or Rest-specific siRNA in a concentration-dependent manner (n ϭ 12 neurons per group; Fig. 7, A and B). The baseline amplitude of evoked EPSCs in lamina II neurons was significantly higher in SNL than in sham control rats, as reported previously (4,31). The inhibitory effect of muscarine (at concentrations of 1-10 M) on the amplitude of evoked monosynaptic EPSCs was significantly lower in neurons of SNL rats treated with control siRNA (n ϭ 11 neurons) than in neurons of sham rats treated with control siRNA (Fig. 7, A and B). Remarkably, treatment with Rest-specific siRNA in SNL rats completely restored the inhibitory effect of muscarine on the amplitude of evoked EPSCs (n ϭ 12 neurons; Fig. 7, A-C) to the level observed in sham control rats treated with control siRNA.
In addition, we conducted electrophysiological recording in spinal cord slices from WT and Rest-cKO mice subjected to SNI or sham surgery. In both WT and Rest-cKO mice subjected to sham surgery, bath application of 0.1 to 10 M muscarine similarly reduced the amplitude of monosynaptically evoked EPSCs of lamina II neurons in a concentration-dependent manner (n ϭ 15 neurons/group; Fig. 8, A and B). Muscarine produced only a small inhibitory effect on the amplitude of evoked monosynaptic EPSCs in WT mice subjected to SNI (n ϭ 16 neurons/group; Fig. 8, A and B). In contrast, in Rest-cKO mice subjected to SNI, muscarine potently reduced the amplitude of evoked EPSCs of lamina II neurons (n ϭ 15 neurons; Gapdh was used as an endogenous control, and the mean value in the sham group was set to be 1 (n ϭ 6 rats/group). ***, p Ͻ 0.001 compared with the sham group at the same time point (two-way ANOVA). C, real-time PCR data showing the mRNA level of Chrm4 in the rat DRG on days 5, 10, and 21 after SNL or sham surgery (n ϭ 6 rats/group). D and E, ChIP-PCR quantification data showing the enrichment of REST at the Chrm2 promoter (D, Ϫ197 to Ϫ98 bp upstream of ATG) and Gapdh promoter (E, Ϫ120 to Ϫ23 bp upstream of ATG) in DRG tissues from sham and SNL rats 21 days after surgery. IgG was used as a negative control for the REST antibody. Data are shown as means Ϯ S.D., and ChIP-PCR values were calculated relative to the input. **, p Ͻ 0.01 (n ϭ 6 rats/group, paired Student's t test). Fig. 8, A-C). These data indicate that REST in DRG neurons plays an essential role in the nerve injury-induced reduction in the ability of mAChR activation to inhibit glutamatergic transmission from primary afferent nerves to spinal dorsal horn neurons.

Discussion
Our study provides new evidence that REST-dependent transcriptional repression in injured DRG neurons plays an important role in the transition from acute to chronic pain after nerve injury. Long-term changes in gene expression profiles in DRG neurons are mediated by epigenetic mechanisms that include DNA methylation, histone modifications, and transcriptional activation and repression (1, 2, 6, 32, 33). By binding to RE1 sites, REST serves as a molecular scaffold that recruits modulatory and epigenetic cofactors to itself. The REST macromolecular repressor complex has both transient and longterm effects on gene transcription in neurons. REST is expressed at higher levels in embryonic and adult neural progenitor cells in neuronal tissues than in nonneuronal tissues (25,34). In nonneuronal mammalian cells and tissues, REST silences neuron-specific genes (34,35). REST is expressed at low levels in mature neurons in several brain regions (22,36). We found in this study that peripheral nerve injury caused a large and sustained increase in REST expression in the DRG. Although previous studies have reported that REST expression in the DRG is increased after sciatic nerve ligation (6,11,37), the functional significance of REST up-regulation in neuropathic pain development is still unclear. Interestingly, we found that, although the acute pain phase after nerve injury was similar in Rest-cKO and WT mice, genetic ablation of Rest in DRG neurons significantly impaired the development of chronic pain 1 week after nerve injury. Furthermore, we showed that knockdown of REST with siRNA attenuated the chronic pain hypersensitivity induced by nerve injury. These findings suggest that nerve injury-induced up-regulation of REST in DRG neurons predominantly contributes to the maintenance of chronic pain. It has been reported that knockdown of Rest with antisense oligonucleotides does not reduce nerve injuryinduced pain hypersensitivity (11). However, the antisense treatment was conducted only for the first 6 days (i.e. the acute pain phase) after nerve injury (11), which could explain why no effect of Rest knockdown on neuropathic pain was observed. As a transcriptional regulator, REST may maintain the chronic phase of neuropathic pain by targeting a select group of genes in the DRG, which may include various voltage-gated K ϩ channels and Oprim1 (2,6,11,38). Nevertheless, the exact REST target genes in the DRG that are involved in the acute-tochronic pain transition remain to be identified.
Another important finding of our study is that nerve injury causes a sustained reduction in Chrm2 expression in the DRG and diminishes the analgesic effect of muscarine at the spinal cord level in chronic neuropathic pain. M2 is the most abundant mAChR subtype in the DRG and spinal cord; it mainly mediates mAChR agonist-produced analgesia (12,19,20,39). It has been shown that M2 is down-regulated in facial motoneurons after axotomy (facial nerve transection) (40) and in the hippocampus after traumatic brain injury (41). Glutamate is the major excitatory neurotransmitter that conveys nociceptive information from primary sensory nerves to spinal dorsal horn neurons (42,43). The M2 sub-

Transcriptional regulation in neuropathic pain
type expressed at central terminals of primary sensory nerves in the spinal dorsal horn plays a critical role in regulating nociceptive glutamatergic input to dorsal horn neurons (14,18,19,21,39). Thus, activation of presynaptic M2 in the spinal dorsal horn produces analgesia by inhibiting glutamatergic transmission from primary sensory nerves to dorsal horn neurons (14,18,39). Repression of Chrm2 expression in the DRG after nerve injury can lead to a reduction in the amount of M2 expressed at presynaptic terminals, which accounts for the attenuation of muscarine-mediated inhibition of nociceptive transmission at the spinal cord level.
Another salient finding of our study is that REST plays an obligatory role in nerve injury-induced repression of Chrm2 expression in the DRG. Our in vitro study showed that the RE1 binding site in the Chrm2 promoter region is required for silencing of Chrm2. Also, nerve injury caused a large increase in enrichment of REST at the Chrm2 promoter in the DRG. Importantly, we showed that knockdown of Rest using siRNA or ablation of Rest in DRG neurons completely normalized the expression level of Chrm2, which had been diminished by nerve injury. In addition, Rest knockdown or Rest ablation in DRG neurons fully restored the reduced muscarinic inhibition of glutamate release from primary afferent nerves to spinal dorsal horn neurons, restoring the muscarinic analgesic effect. These results strongly support the notion that REST is essential for nerve injury-induced Chrm2 silencing and the diminished analgesic effect of muscarine in neuro-

Figure 5. Nerve injury represses Chrm2 expression in the DRG and diminishes the analgesic effect of muscarine at the spinal cord level through REST.
A and B, real-time PCR data showing the mRNA level of Chrm2 (A) and Chrm4 (B) in the DRG of rats that underwent sham surgery or SNL and were treated intrathecally with control siRNA or Rest-specific siRNA (2 g/day, 5 days) 3 weeks after surgery. ***, p Ͻ 0.001 (n ϭ 6 rats/group, one-way ANOVA). C, original blotting images and quantification data showing the protein level of M2 in the DRG of sham and SNL rats treated with control (CT) siRNA or Rest-specific siRNA. GAPDH was used as an internal control. M, molecular weight marker. **, p Ͻ 0.01 (n ϭ 6 rats/group, one-way ANOVA). D and E, time course of the effects of intrathecal injection of muscarine (2 g) on the tactile (D) and pressure (E) withdrawal thresholds in sham and SNL rats treated with control siRNA or Rest-specific siRNA 3 weeks after sham or SNL surgery. Data are shown as means Ϯ S.D. (A-C) or means Ϯ S.E. (D and E). *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001 compared with the respective baseline (time 0). #, p Ͻ 0.05; ##, p Ͻ 0.01; ###, p Ͻ 0.001 compared with SNL rats treated with control siRNA (n ϭ 6 rats/group, two-way ANOVA).

Transcriptional regulation in neuropathic pain
pathic pain. G9a, a histone methyltransferase, may function as a REST partner (44) in the injured DRG; knockdown of REST or G9a can restore the nerve injury-induced reduction in Oprm1 expression in the DRG and opioid-induced analgesia in neuropathic pain (11,38). However, RNA sequencing data show that the reduced mRNA level of Chrm2 in the DRG caused by nerve injury is not restored by G9a inhibition (2). Thus, other REST-interacting epigenetic enzymes and cofactors, such as histone demethylases and mSin3 (45), may cooperate with REST to create a relatively condensed chromatin feature and thus silence the expression of Chrm2 in the injured DRG.
Although it has been reported that the RE1 site is required for the promoter activity of Chrm4 in several nonneuronal cell lines (46 -48), we found in this study that neither nerve injury nor Rest-specific siRNA knockdown had any effect on the expression level of Chrm4 in the DRG. It is not fully known what determines the specificity of interaction between REST and its targets as well as the enrichment of REST at the promoter of a given target gene. It should be noted that experiments performed in cell lines do not reflect the cellular environment and epigenetic landscape, which vary with developmental stages, cell types, and disease conditions. For instance, the REST target genes in the hippocampal CA1 altered by ischemia are distinct from those in the hippocampal CA3 related to seizure activity (49,50). Other transcription factors and epigenetic marks may also influence the REST binding affinity of target genes. Thus, REST likely controls the expression of different ensembles of target genes in different cell types and disease conditions.

Transcriptional regulation in neuropathic pain
In summary, our study provides important new evidence that REST up-regulation in the injured DRG contributes to the transition from acute to chronic pain after nerve injury. Furthermore, Chrm2 expression in the DRG and the muscarinic analgesic effect are diminished in neuropathic pain through a transcriptional mechanism involving REST. This information suggests new strategies for treating patients with chronic neuropathic pain. The potentiated analgesic effects of muscarine induced by Rest knockdown or ablation in neuropathic pain suggest that a combination of REST inhibitors and mAChR agonists may be effective for the treatment of chronic neuropathic pain. REST inhibitors may increase the analgesic efficacy of mAChR agonists, thereby reducing the dose of mAChR agonists and improving their therapeutic windows. In addition, reduced cholinergic activity/M2 levels and increased REST expression have been detected in the brain tissues of patients with Alzheimer and Huntington diseases (51)(52)(53)(54)(55). Thus, our findings regarding the transcriptional control of Chrm2 by REST may also be relevant for understanding the molecular mechanisms underlying the development of these neurodegenerative diseases.

Animal models
The surgical procedures and experimental protocols were approved by the Animal Care and Use Committee at the University of Texas MD Anderson Cancer Center. All animals were housed (2-3 rats/cage) on a standard 12:12 h light-dark cycle and received food and water ad libitum. Male Sprague-Dawley rats (7-8 weeks old, Harlan, Indianapolis, IN) were subjected to SNL as described previously (23,24). Briefly, rats were anesthetized with 2-3% isoflurane, and the left L5 and L6 spinal nerves were exposed and ligated separately with 6-0 silk sutures under a surgical microscope. Sham surgery (without nerve ligation) was performed as the control.
Rest conditional knockout mice (Rest cKO) were generated by crossing female mice carrying loxP sites flanking the first coding exon 4 of the Rest gene (Rest-loxP ϩ/ϩ , stock no. 024549, The Jackson Laboratory, Bar Harbor, ME) (25) and male mice with primary sensory neuron-specific Advillin-Cre (Advillin-Cre ϩ/Ϫ ) (26,27). Advillin-Cre ϩ/Ϫ : Rest-loxP ϩ/Ϫ mice obtained from the first generation were crossed with female Rest-loxP ϩ/ϩ mice to get Advillin-Cre ϩ/Ϫ : Rest-loxP ϩ/ϩ mice (Rest-cKO mice). Littermates without Cre (Advillin-Cre Ϫ/Ϫ : Rest-loxP ϩ/ϩ ) were used as controls. Mice were ear-marked at the time of weaning (3 weeks after birth), and tail biopsies were used for PCR genotyping. Both male and female adult mice (8 -10 weeks of age) were used for electrophysiological and behavioral studies. SNI was performed as described previously (2,56). Mice were anesthetized with 2-3% isoflurane, and the sciatic nerve and its three terminal branches (the sural, common peroneal, and tibial nerves) of the left leg were exposed under a surgical microscope. The tibial and common peroneal nerves were tightly ligated with a 6-0 silk suture and sectioned distal to the ligation sites, leaving the sural nerve intact.

Intrathecal catheter placement and treatment
Rats were anesthetized with 2-3% isoflurane. A small incision was made at the back of the neck, and a PE-10 catheter (8 EPSCs of dorsal horn neurons in sham control rats treated with Rest-specific siRNA (n ϭ 12 neurons) or control siRNA (n ϭ 12 neurons) and SNL rats treated with Rest-specific siRNA (n ϭ 12 neurons) or control-siRNA (n ϭ 11 neurons). Because the baseline amplitude of evoked EPSCs was different in SNL rats treated with control siRNA and those treated with Rest-specific siRNA, the amplitude of evoked EPSCs was normalized to the baseline for each recorded neuron in C. Data are shown as means Ϯ S.E. *, p Ͻ 0.05; **, p Ͻ 0.01 compared with the respective baseline. #, p Ͻ 0.05; ##, p Ͻ 0.01 compared with SNL rats treated with control siRNA (two-way ANOVA).

Transcriptional regulation in neuropathic pain
cm) was inserted via a small opening made in the atlanto-occipital membrane of the cisterna magna so that the catheter tip reached the lumbar spinal enlargement (57). The rostral end of the catheter was exteriorized, and the wound was closed with sutures. The animals were allowed to recover for 5-7 days before intrathecal injections.

Nociceptive behavioral tests
For tactile sensitivity measurements, rats or mice were habituated on a wire grid panel for 30 min before testing. The plantar surface of the hind paw was subjected to a series of von Frey filaments (Stoelting, Wood Dale, IL). Brisk withdrawal or flinching of the paw was considered a positive response. In the absence of a response, we applied the filament of the next greater force. We calculated the tactile threshold that produced a 50% likelihood of a withdrawal response using the "up-down" method (24).
To measure the mechanical nociceptive threshold, the paw pressure test was performed using an analgesiometer (Ugo Basile, Varese, Italy). A foot pedal was pressed to activate a motor that applied a linearly increasing force to the hind paw (19). The pedal was immediately released when the animal displayed pain by either withdrawing the paw or vocalizing. Each trial was repeated two or three times at 2-min intervals, and the mean value was used as the force needed to produce a withdrawal response.

Rotarod test
The rotarod test was carried out on a rotarod apparatus (LE8205, Harvard Apparatus, Holliston, MA) programmed to accelerate from 4 to 40 rpm during a 120-s window. Mice were placed on a moving rotarod (4 rpm) and habituated for 5 min. After accelerating, the time on the rotarod was recorded, and the mean value from three repeated trials was obtained. The test was conducted on three consecutive days.

Quantitative PCR
Total RNAs were extracted from the lumbar DRG and spinal cord tissues using TRIsure (BIO-38032, Bioline, Taunton, MA). After treatment with RNase-free DNase (79254, Qiagen, Hilden, Germany), 1 g of RNA was used for reverse transcription with the RevertAid RT reverse transcription kit (K1619, Thermo Fisher Scientific, Waltham, MA). 2 l of complementary DNA diluted five times was added to a 20-l reaction volume with SYBR Green real-time PCR mixture (A25780, Thermo Fisher Scientific). Real-time PCR was performed using the QuantStudio 7 Flex real-time PCR system (Applied Biosystems, Waltham, MA). The thermal cycling conditions were as follows: 1 cycle at 95°C for 10 min and 40 cycles at 95°C for 15 s and at 60°C for 45 s. The following primers were used: rat

Vector constructs
A fragment (ϳ770 bp) upstream of ATG in the Chrm2 gene that contains RE1 binding sites was amplified using paired primers: GCG AAG CTT GGC CAG AGC TAC CTC TTC TGC AA and GCG CCA TGG TTT GTG TTC AGT AGT CAA GTG GCC A. After digestion with HindIII and NcoI, the sequence was ligated to pGL3 luciferase reporter vectors (E1751, Promega, Madison, WI) to generate a proM2::Luc (WT version) vector. The vector for proM2-del-Luc (RE1-deleted version) was obtained by deleting the RE1 binding site from vector proM2⌬RE1::Luc using the In-Fusion HD Cloning Plus kit (638916, Clontech Laboratories, Inc., Mountain View, CA) and the paired primers TAT ACT GTT CCC TGT CTG ACC AGG C and A GGG AAC AGT ATA ta TTC TCA AAA GGA AGA AAT CTG ATG TGT T. All constructs were confirmed by DNA sequencing.

ChIP
ChIP assays were performed as described previously (2,61). Briefly, lumbar DRG (L5/L6) tissues were isolated and crosslinked with 2 mM disuccinimidyl glutarate for 35 min, followed by 2% formaldehyde for another 20 min at 25°C. The DRG tissues were lysed with lysis buffer and sonicated to fragments of around 200 to 1000 bp using a water bath sonicator (Qsonica, Newtown, CT) at 4°C (40 cycles of 30 s on and 30 s off). Chromatin was pulled down using Dynabeads Protein G magnetic beads (10003D, Thermo Fisher Scientific) conjugated with an anti-IgG antibody (ab124055, Abcam, Cambridge, MA) or anti-REST antibody (07-579, Millipore, Burlington, MA). After decross-linking, DNA was recovered using the QIAquick PCR purification kit (58106, Qiagen). Data were analyzed and corrected by input (10% of the amount used for ChIP). Primers used for the ChIP assay were as follows: rat Chrm2 forward, TT CCA TCT AGA GTT TTT GCC TCT TGC C; rat Chrm2 reverse, CAG ATT TCT TCC TTT TGA GAA ACT TCC C; rat Gapdh forward, CAT TGC AGA GGA TGG TAG AGG A; and rat Gapdh reverse, GGA GAA GAA AGT CAG ATT AGC GTG G.
Neurons from the lamina II outer region were selected for whole-cell patch clamp recordings because they are involved in processing nociceptive input from primary afferent nerves (43,62). A glass pipette (5-10 megaohm) was filled with an internal solution containing 135 mM potassium gluconate, 5 mM tetraethylammonium chloride, 2 mM MgCl 2 , 0.5 mM CaCl 2 , 5 mM HEPES, 5 mM EGTA, 5 mM magnesium ATP, 0.5 mM sodium GTP, 1 mM guanosine 5Ј-O-(2-thiodiphosphate) trilithium salt, and 10 mM lidocaine N-ethyl bromide (pH adjusted to 7.2-7.4 with 1 M KOH and osmotic pressure to 290 -300 mosmol). EPSCs were elicited by electrical stimulation (0.2 ms, 0.6 mA, and 0.1 Hz) of the dorsal root and recorded at a holding potential of Ϫ60 mV. Monosynaptic EPSCs were identified on the basis of the constant latency and absence of conduction failure of evoked EPSCs in response to a 20-Hz electrical stimulation (18,30). The input resistance was continuously monitored, and the recording was terminated when the input resistance changed by more than 15%. Signals were recorded using an amplifier (MultiClamp700B, Axon Instruments Inc., Union City, CA), filtered at 1 to 2 kHz, digitized at 10 kHz, and stored for off-line analysis. All drugs were freshly prepared in artificial cerebrospinal fluid before the experiments and delivered via syringe pumps to reach their final concentrations.

Statistical analysis
Data are presented as means Ϯ S.E. or means Ϯ S.D., as specified in the figure legends. Although no statistical methods were used to predetermine sample sizes, our sample sizes were similar to those generally employed in the field. Data collection was randomized, and the investigators performing behavioral tests and electrophysiological recordings were blinded to the treatment. In electrophysiological experiments, only one neuron was recorded from each spinal cord slice, and at least six animals were used for each group. The amplitude of the evoked EPSCs was quantified by averaging 10 consecutive EPSCs with Clampfit 10.0 software (Axon Instruments). We used Student's t test to compare the mRNA and protein levels in the two groups. Mann-Whitney nonparametric test was used for luciferase reporter assay data analysis. To determine the effects of treatment on the withdrawal thresholds, we used repeated measures analysis of variance followed by Dunnett's post hoc test. We used one-way and two-way analysis of variance followed by Tukey's post hoc test to compare values in more than two groups. Statistical analysis was performed using Prism version 7 (GraphPad Software Inc., La Jolla, CA). A p value of less than 0.05 was considered to be statistically significant.