Chronic benzodiazepine treatment triggers gephyrin scaffold destabilization and GABAAR subsynaptic reorganization

Abstract

Benzodiazepines (BZDs) are important clinical drugs with anxiolytic, anticonvulsant, and sedative effects mediated by potentiation of inhibitory GABA type A receptors (GABA_ARs). Tolerance limits the clinical utility of BZDs, yet the mechanisms underlying tolerance after chronic exposure have not been thoroughly investigated. Here, we assessed the impact of chronic (7-day) treatment with the BZD diazepam (DZP) on the dynamic plasticity and subsynaptic organization of the gephyrin scaffold and γ2 subunit-containing GABA_ARs in primary neurons. After functional confirmation of diminished BZD sensitivity, we provide the first super-resolution analysis of inhibitory nanoscale plasticity induced by chronic BZD exposure: gephyrin subsynaptic domains were smaller and the inhibitory postsynaptic area was overall diminished by DZP treatment, resulting in a condensation of synaptic γ2-GABA_ARs into smaller synaptic areas. Using a novel fluorescence-based in situ proximity ligation assay and biochemical fractionation analysis, the mechanism for gephyrin downregulation was revealed to be dependent on phosphorylation and protease cleavage. Accordingly, DZP treatment impaired gephyrin synaptic stability, demonstrated by live-imaging photobleaching experiments. Despite the loss of BZD sensitivity and stable synaptic gephyrin, 7-day DZP treatment did not reduce the surface or total protein levels of BZD-sensitive γ2-GABA_ARs, as shown in prior short-term BZD treatment studies. Instead, chronic DZP treatment induced an accumulation of γ2-GABA_ARs in the extrasynaptic membrane. Surprisingly, γ2-GABA_AR interactions with gephyrin were also enriched extrasynaptically. An identified rise in extrasynaptically-localized gephyrin cleavage fragments may function to confine receptors away from the synapse, as supported by a decrease in extrasynaptic γ2-GABA_AR mobility. Altogether, we find that chronic BZD treatment triggers several subtle converging plasticity events at inhibitory synapses which effectively restrict the synaptic renewal of BZD-sensitive GABA_ARs via mechanisms distinct from those observed with short-term treatment.

Keywords: GABAA receptor; benzodiazepine; chronic; gephyrin; inhibition; plasticity; subsynaptic; tolerance.

Figure 1

GABA_AR potentiation by BZDs is impaired after chronic DZP treatment in primary neurons. Miniature inhibitory postsynaptic currents (mIPSCs) were measured by whole-cell electrophysiology to assess baseline inhibitory function and sensitivity to BZDs in neurons treated with vehicle (Veh) or 1 μM DZP for seven days. (A) Representative mIPSC traces from 7-day Veh- or DZP-treated cortical neurons (i) before and (ii) after acute application of 1 μM diazepam. (B) Baseline mIPSC amplitude (Veh = 31.1 ± 9.47 pA, DZP = 57.2 ± 12.7 pA; p = 0.1379), frequency (Veh = 1.8 ± 0.57 Hz, DZP = 2.3 ± 0.64 Hz; p = 0.5615), and τ_decay (Veh = 50.0 ± 2.87 ms, DZP = 43.7 ± 2.92 ms; p = 0.1618) are unchanged by 7-day DZP treatment. (C) Representative mIPSC averaged traces before and after acute diazepam application. (D–G) BZD sensitivity in cultured neurons is severely diminished by 7-day DZP treatment. (D) mIPSC amplitude measured before and after application of acute diazepam (Veh-treated: before acute diazepam = 31.1 ± 9.47 pA, +diazepam = 48.9 ± 15.3 pA, p = 0.0480; DZP-treated: before acute diazepam = 57.2 ± 12.7 pA, +diazepam = 61.6 ± 14.2 pA, p = 0.0849). (E) The percent potentiation of mIPSC amplitude by application of acute diazepam is lost in chronic DZP-treated neurons (Veh = 55.9 ± 16.2%, DZP = 6.79 ± 2.96%; p = 0.0176). (F) mIPSC τ_decay in 7-day Veh vs. DZP neurons before and after acute application of diazepam (Veh-treated: before acute diazepam = 50.0 ± 2.87 ms, +diazepam = 86.2 ± 6.97 ms, p = 0.0042; DZP-treated: before acute diazepam = 43.7 ± 2.92 ms, +diazepam = 54.3 ± 3.00 ms, p = 0.0197). (G) The percent potentiation of mIPSC τ_decay by acute diazepam is significantly diminished in chronic DZP-treated neurons (Veh = 72.5 ± 12.9%, DZP = 25.6 ± 7.48%; p = 0.0125). n = 5 cells per treatment, N = 3 independent cultures; mean ± SEM. (B, E, G) unpaired t-test; (D, F) paired t-test; *p ≤ 0.05, **p ≤ 0.01.

Figure 2

Altered subsynaptic organization of gephyrin and γ2-GABA_ARs by chronic DZP treatment. The subsynaptic organization of gephyrin and γ2-GABA_AR was analyzed by DNA-PAINT super-resolution localization microscopy. (A) DNA-PAINT Schematic. Primary antibodies recognizing surface γ2-GABA_AR or intracellular gephyrin are targeted by DNA-coupled secondary nanobodies (docking strands), while imager strands containing the fluorophore-bound complementary oligonucleotide remain freely available. Transient binding between the imager and docking strands produces a bright localization event. (B) Example snapshot of a VGAT-stained neuron used in DNA-PAINT microscopy. The zoomed region is overlayed with localizations (locs) of gephyrin (red) and γ2-GABA_AR (blue); scale bar: 1 μm. Colocalization with VGAT confirmed high-density localization clusters as synaptic. (C) Representative synapses from a 7-day Veh- or DZP-treated neuron captured by DNA-PAINT; each point represents a localization of either surface γ2-GABA_AR (blue squares) or gephyrin (red circles). (D, E) Localization analysis of γ2-GABA_AR and gephyrin total synapse area (D) and localization density (E). (D) Chronic DZP treatment reduced the total synapse area of both γ2-GABA_AR (Veh = 50.1 ± 2.4 × 10³ nm², DZP = 35.4 ± 1.8 × 10³ nm², p = 0.0005) and gephyrin (Veh = 130.1 ± 3.7 × 10³ nm², DZP = 98.9 ± 3.9 × 10³ nm²; p < 0.0001). (E) γ2-GABA_AR synapse localization density was increased in DZP-treated neurons (Veh = 4.3 ± 0.13 × 10⁻³ locs/nm², DZP = 6.5 ± 0.56 × 10⁻³ locs/nm²; p = 0.0019). Gephyrin synapse localization density was unchanged by DZP treatment. (F) Representative synapses from (C) with SSD localizations highlighted. (G–J) Analysis of γ2-GABA_AR and gephyrin SSD numbers per synapse (G), SSD area (H), localization density within SSDs (I), and SSD/Synapse Area (J). Gephyrin SSD area was reduced after chronic DZP treatment (Veh = 5.2 ± 0.26 × 10³ nm², DZP = 4.2 ± 0.25 × 10³ nm²; p = 0.0265); SSDs were otherwise similar between vehicle- and DZP-treated neurons. n = 7–9 cells, N = 2 independent cultures; mean ± SEM. (D, E, G–I) Mann-Whitney test, (J) Mann-Whitney test (γ2-GABA_AR) or unpaired t-test (gephyrin). *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001.

Figure 3

Chronic DZP treatment increases gephyrin Ser270 phosphorylation at synapses. Phosphorylation of gephyrin at Ser270 was assessed by proximity ligation (PL) assay (PLA) in 7-day Veh- vs. DZP-treated neurons. (A) PLA Schematic; fluorescent PL signal (yellow) is only observed when the phospho-Ser270-specific mAb7a antibody is within 40 nm of the total gephyrin antibody, indicating Ser270 phosphorylation. (B) Representative images of Veh- or chronic DZP-treated neurons with PLA signals (yellow); MAP2 (green) and VGAT (pink) counterstaining were used to label neuronal dendrites and inhibitory synapses, respectively. (C–F) Quantification of the number (C, D) or intensity (E, F) of mAb7a–gephyrin PL signals in the whole field or at synaptic sites. Chronic DZP treatment significantly increased synaptic PL signal intensity, indicating increased gephyrin Ser270 phosphorylation (Veh = 100.0 ± 11.6%, DZP = 157.8 ± 20.0%; p = 0.0263). n = 36–37 cells, N = 3 independent cultures; median (solid line) and quartiles (dashed lines) are shown. (C–F) Mann-Whitney test; *p ≤ 0.05. Scale bars are 20 μm for neurons and 2 μm for dendrite zoom images.

Figure 4

Enhanced proteolytic gephyrin cleavage decreases full-length gephyrin expression in 7-day DZP-treated neurons. Full-length or cleaved gephyrin protein expression was assessed by subcellular fractionation and western blotting. (A) Representative western blots. Each lane represents a biological replicate. (B–D) Quantifications of full-length gephyrin (B), cleaved gephyrin (C), and the ratio of cleaved/full-length gephyrin (D) in the synaptic, extrasynaptic, and total protein fractions from 7-day Veh- or DZP-treated neurons. Immunoreactivities were normalized to GAPDH. (B) Full-length gephyrin was near-significantly reduced in the synaptic fraction (Veh = 100.0 ± 7.01%, DZP = 75.16 ± 9.64%; p = 0.0559) and significantly reduced in the total fraction (Veh = 100.0 ± 3.14%, DZP = 82.02 ± 5.26%; p = 0.0109), while extrasynaptic full-length gephyrin was unchanged. (C) Chronic DZP treatment increased cleaved gephyrin levels only at extrasynaptic sites (100.9 ± 6.04%, DZP = 295.0 ± 65.19%, p = 0.0003). (D) The proportion of cleaved/full-length gephyrin was elevated by chronic DZP treatment in the synaptic and extrasynaptic fractions (synaptic: Veh = 1.31 ± 0.099, DZP = 1.90 ± 0.20, p = 0.0220; extrasynaptic: Veh = 0.89 ± 0.19, DZP = 2.74 ± 1.87, p = 0.0003). n = 2 replicates per treatment from N = 4 independent cultures; mean ± SEM. B-D: unpaired t-test or Mann-Whitney test; *p ≤ 0.05, ***p ≤ 0.001.

Figure 5

BZD-sensitive γ2-GABA_ARs are redistributed from synaptic to extrasynaptic sites after chronic DZP treatment without loss of surface expression. Surface expression and synaptic or extrasynaptic localization of γ2-GABA_ARs was assessed in 7-day Veh- or DZP-treated neurons. (A) Representative immunofluorescence (IF) images. Cells were first surface stained for endogenous γ2-GABA_AR then subsequently permeabilized and stained for GAD65 to mark presynaptic inhibitory terminals. (B–E) Quantification of IF results, including cluster density, signal area (μm²), and signal intensity (% Veh) of synaptic γ2-GABA_ARs (B), extrasynaptic γ2-GABA_ARs (C), total surface γ2-GABA_ARs (D), or GAD65 (E). Synaptic γ2-GABA_AR signal was defined by binary intersection with GAD65. (B) Chronic DZP treatment reduced γ2-GABA_AR clustering density (Veh = 3.33 ± 0.192, DZP = 2.74 ± 0.155; p = 0.0196) and synaptic area (Veh = 1.74 ± 0.126 μm², DZP = 1.40 ± 0.109 μm²; p = 0.0411) without altering signal intensity. (C) Chronic DZP treatment enriched the extrasynaptic accumulation of γ2-GABA_ARs (binary area: Veh = 0.391 ± 0.041 μm², DZP = 0.536 ± 0.064 μm², p = 0.0657; sum intensity: Veh = 100.0 ± 9.66%, DZP = 163.1 ± 20.4%, p = 0.0091). (D, E) Total surface γ2-GABA_AR and GAD65 staining were unchanged by DZP treatment. (F, G) Surface biotinylation experiments confirm that chronic DZP treatment does not alter surface or total protein expression of γ2-GABA_AR subunits. (F) Representative western blots of the surface and total fractions collected by surface biotinylation. Each lane represents a biological replicate. The lack of GAPDH signal in the surface fraction confirms isolation of surface proteins. (G) Quantification of γ2-GABA_AR subunit surface and total protein expression. (B–E) n = 42–47 cells, N = 3 independent cultures; (G) n = 15–22 replicates, N = 6–9 independent cultures; mean ± SEM. (B–E, G) unpaired t-test; *p ≤ 0.05, **p ≤ 0.01. Scale bars are 20 μm for neuron images and 2 μm for dendrite zoom images.

Figure 6

Gephyrin associations with γ2-GABA_ARs are enhanced by chronic DZP treatment in the extrasynaptic membrane. Gephyrin associations with γ2-GABA_ARs were assessed by PLA in neurons treated with Veh or 1 μM DZP for 7 days. (A) PLA Schematic; fluorescent PL signal is only observed when γ2-GABA_AR and gephyrin are within 40 nm, indicating association; experiments were performed under permeabilized conditions. (B) Representative neuron images; MAP2 and VGAT counterstaining was included to label neuronal dendrites and inhibitory synapses, respectively; PLA signals are shown in yellow. (C–E) Quantification of the number of whole-field (C), synaptic (D), or extrasynaptic (E) gephyrin-GABA_AR PL signals. Chronic DZP treatment resulted in a near-significant increase in the number of whole-field PL signals (C; Veh = 100.0 ± 6.015%, DZP = 121.1 ± 9.219%; p = 0.0561) and significantly higher numbers of extrasynaptic PL signals (E; Veh = 100.0 ± 5.328%, DZP = 125.0 ± 9.508%; p = 0.0270), while synaptic PL signals were similar between vehicle- and DZP-treated neurons. n = 38–47 cells, N = 3 independent cultures; median (solid line) and quartiles (dashed lines) are shown. (C–E) Mann-Whitney test; *p ≤ 0.05. Scale bars are 20 μm for neurons and 2 μm for dendrite zoom images.

Figure 7

Chronic DZP treatment destabilizes synaptic gephyrin and impairs the mobility of extrasynaptic γ2-GABA_ARs. Fluorescence Recovery After Photobleaching (FRAP) experiments were performed in hippocampal neurons co-transfected with γ2^pHFAP and mScarlet-Gephyrin.FingR and treated with Veh or 1 μM DZP for 7 days. (A) Representative images of synaptic γ2^pHFAP and mScarlet-Gephyrin.FingR (geph.FingR, gephyrin.FingR) before bleaching (pre-bleach), immediately after bleaching (t = 0 min), and 30 min post-bleach (t = 30 min). White boxes indicate regions of photobleaching. (B) Analysis of fluorescence recovery after photobleaching for synaptic mScarlet-gephyrin.FingR and γ2^pHFAP. mScarlet-gephyrin.FingR fluorescence recovery was elevated by chronic DZP treatment, while synaptic γ2^pHFAP recovery was unchanged by DZP treatment. (C) Representative images of extrasynaptic regions of γ2^pHFAP and mScarlet-Gephyrin.FingR before bleaching (pre-bleach), immediately after bleaching (t = 0 min), and 30 min post-bleach (t = 30 min). White boxes indicate regions of photobleaching. (D) Analysis of fluorescence recovery after photobleaching of extrasynaptic regions. Chronic DZP treatment did not affect extrasynaptic trafficking of mScarlet-gephyrin.FingR, while γ2^pHFAP extrasynaptic mobility was reduced. (B, D) n = 16–17 cells, N = 3 independent cultures; mean ± SEM. Analyses by multiple unpaired t-tests; *p ≤ 0.05, **p ≤ 0.01. Scale bars: 2 μm.