Epigenetic suppression of GAD65 expression mediates persistent pain

Abstract

Chronic pain is a common neurological disease involving lasting, multifaceted maladaptations ranging from gene modulation to synaptic dysfunction and emotional disorders. Sustained pathological stimuli in many diseases alter the output activities of certain genes through epigenetic modifications, but it is unclear how epigenetic mechanisms operate in the development of chronic pain. We show here that in the rat brainstem nucleus raphe magnus, which is important for central mechanisms of chronic pain, persistent inflammatory and neuropathic pain epigenetically suppresses Gad2 (encoding glutamic acid decarboxylase 65 (GAD65)) transcription through histone deacetylase (HDAC)-mediated histone hypoacetylation, resulting in impaired γ-aminobutyric acid (GABA) synaptic inhibition. Gad2 knockout mice showed sensitized pain behavior and impaired GABA synaptic function in their brainstem neurons. In wild-type but not Gad2 knockout mice, HDAC inhibitors strongly increased GAD65 activity, restored GABA synaptic function and relieved sensitized pain behavior. These findings suggest GAD65 and HDACs as potential therapeutic targets in an epigenetic approach to the treatment of chronic pain.

Figure 1

Persistent inflammatory pain induces time-dependent hyperacetylation of histones H3 and H4. (a) Time course for the development of persistent pain sensitization induced by complete Freund's adjuvant (CFA) and for saline controls, measured by the paw-withdrawal test (n = 6 rats in each group). (b,c) Western blot lanes (b) and summarized data (c, n = 5–9 rats for each group) of global acetylated histone H3 (AcH3) and total H3 proteins, normalized to β-actin, in tissues of rat nucleus raphe magnus (NRM) taken at various time points after CFA injection. (d) Western lanes of AcH3 and AcH4 3 d after CFA injection. (e,f) Western lanes (e) and summarized results (f, n = 7 rats for each group) of AcH4 and total H4 after CFA injection. Data are expressed as mean ± SEM. * p < 0.05, ** p < 0.01. BL, baseline. Sal, saline.

Figure 2

Persistent pain decreases GABAergic synaptic function by inhibiting presynaptic GABA release. (a) Representative traces of GABA inhibitory post-synaptic currents (IPSCs) evoked by various stimulation intensities in NRM neurons from a saline-injected rat and a CFA-injected rat at 4 h and 3 d post-injection. (b) A plot of input-output curves for IPSC amplitudes in neurons of the three treatment groups as in (a). Slopes: saline, 90.4 ± 11.5 pA 0.1 mA^–1, n = 13; CFA 4 h, 88.7 ± 12.9 pA 0.1 mA^–1, n = 26, p > 0.05; CFA 3 d, 53.6 ± 10.5 pA 0.1 mA^–1, n = 35, p < 0.01. (c) A similar input-output plot of IPSCs from hippocampal CA1 neurons. Slopes: saline, 98.1 ± 14.8 pA 0.1mA^–1, n = 15; CFA 3 d, 102.7 ± 16.2 pA 0.1mA^–1, n = 16, p > 0.05. (d) Representative IPSC pairs evoked by two stimuli (100 ms apart) in NRM neurons from the three groups of rats. (e) Two representative IPSC pairs, from the two indicated groups, superimposed and scaled to the amplitude of the first IPSC. (f) Group data of changes in the paired-pulse ratios in the three groups (n = 15–35 cells for each group). (g) Representative traces of spontaneous miniature IPSCs (mIPSCs) in neurons from the three groups. (h–j) Distribution plots of mIPSC frequency (h) and amplitude (i) from neurons of each group and their group data (j, n = 16–20 cells).

Figure 3

Persistent pain induces histone hypoacetylation at the gad65 promoter. (a) Summarized data (n = 5–6 rats each group) of acetylated H3 levels in the gad65 promoter region [(–646 to –484 bp upstream of the transcription start site (TSS)] in NRM tissues from the indicated groups of CFA-injected rats and controls. (b,c) Immunoblotting of acetylated H3 (b) and summarized results of Western analyses and ELISA (c) in saline- and CFA-injected rats (n = 5–7 for each group) treated with TsA or SAHA. (d) Normalized levels of acetylated H3 in the indicated sequence regions upstream of the TSS in the gad65 gene in NRM tissues from saline- and CFA-injected rats at 3 d post-injection (n = 5–6 rats each group). (e) Representative real-time PCR data from ChIP with HDAC-specific antibodies at the gad65 promoter region (–646 to –484 bp) in NRM tissues of rats (n = 5) at 3 d after CFA injection, showing amplification curves of input, immunoprecipitates with antibodies to HDAC1, HDAC2, HDAC4 and HDAC5, and non-immune IgG as negative control.

Figure 4

Persistent inflammatory and neuropathic pain reduces GAD65 expression. (a) Effects of CFA and TsA on GAD65 mRNA levels in NRM tissues from control and CFA-injected rats (n = 5–7 each group). (b,c) Western lanes (b) and group data (c, n = 5– 7 each group) of GAD65 proteins in saline- and CFA-injected rats 4 h and 3 d post-injection without or with the TsA treatment in vivo. (d) Western lanes of GAD65 proteins in NRM tissues from normal rats. (e) Micrographs of immunohistochemical staining for GAD65 (red), the synaptic terminal protein synapsin I (green), their merged images and co-localization of GAD65 and synapsin I (yellow) in saline- (top row) and CFA- (bottom two rows) injected rats without or with the TsA treatment (n = 5–6 rats each group and 4–6 randomly selected sections were analyzed and compared). Scale bars = 50 μm (left three columns) and 2 μm (right column). (f) Time course of changes in pain threshold in rats after spinal nerve ligation (SNL, n = 6) or sham surgery (n = 5) performed on day 0 (arrow). (g) Normalized changes in levels of acetylated H3 at gad65 promoter (–646 to – 484 bp), GAD65 mRNA and protein in NRM tissues from sham and SNL rats at 1 d and 21 d after surgery (n = 5–6 each group). (h) Representative Western lanes of GAD65 and β-actin proteins from a sham rat and an SNL rat at 1 d and 21 d.

Figure 5

Histone hyperacetylation overcomes pain-induced reduction in GABA synaptic function through GAD65. (a) Spontaneous mIPSC traces (top) and summarized data (bottom) of effects of TsA and SAHA on mIPSC frequency in NRM neurons from saline- and CFA-injected rats (n = 12–17 cells each group). (b) Representative traces of GABA IPSCs evoked by stimuli of increasing intensities in an NRM neuron from a wild type (WT) and from a gad65^–/– mouse (top), and an input-output plot of GABA IPSCs in neurons from WT (n = 23 cells) and gad65^–/– mice (n = 15 cells) (bottom). (c,d) Representative mIPSC traces (c) and group data (normalized to WT) of mIPSC frequency and amplitude (d) in NRM neurons from WT and gad65^–/– mice without or with the TsA or SAHA treatment (n = 15–26 cells for each group). (e) Summarized data of the normalized levels of acetylated H3 in the gad67 promoter region (–374 to –273 bp), GAD67 mRNA and GAD67 protein in NRM tissues from saline- and CFA-injected rats at 3 d post-injection (n = 4–6 for each group). (f) Representative Western blots of GAD67 protein from a saline- or CFA-injected rat. (g, h) Representative membrane current traces with response to the mu-opioid receptor (MOR) agonist DAMGO (1 μM) (top) and graphs of evoked action potential firing (bottom) in MOR-expressing NRM neurons (g) and in MOR-lacking NRM neurons (h) from saline- and CFA-injected rats. * p < 0.05, ** p < 0.01. Veh, vehicle.

Figure 6

HDAC inhibitors relieve pain through GAD65. (a, b) Changes in pain threshold in CFA-injected rats with repeated NRM infusions (arrowheads) of vehicle (n = 5) or TsA (a, n = 8), and vehicle (n = 5) or SAHA (b, n = 6 each group). (c) Effects of TsA at indicated doses (n = 5–6 for each data point). (d) Effect of TsA pretreatment by NRM infusions (arrowheads) on CFA-induced acute hyperalgesia. (e) Averaged baseline pain threshold in WT (n = 10) and gad65^–/– mice (n = 7), and in gad65^–/– mice treated with NRM infusions of vehicle (n = 7) or TsA (n = 7). (f) Pain threshold changes in saline- and CFA-injected WT and gad65^–/– mice. (g) Effects of single NRM infusion of saline or the GABA_A receptor agonist muscimol (50 ng) on CFA-induced hyperalgesia (n = 5 rats each group). (h) Pain threshold changes following repeated NRM infusions of saline (n = 5) or the proinflammatory cytokine interleukin 1 (IL-1β, 3 μg, n = 5) (arrows), and following single NRM infusion of IL-1β plus the IL-1 receptor antagonist IL-1Ra (17 μg, n = 6). (i,j) Western blots (i) and summarized data (j) of GAD65 protein normalized to β-actin in NRM tissues from normal rats after repeated NRM infusions of saline or IL-1β, and from CFA-injected rats treated with similar 4-day NRM infusions of saline or IL-1Ra (17 μg). *, compared to saline group; #, compared to IL-1β + L-1Ra group. MPE, maximum possible effect.