Inhibitory and excitatory synaptic neuroadaptations in the diazepam tolerant brain

Abstract

Benzodiazepine (BZ) drugs treat seizures, anxiety, insomnia, and alcohol withdrawal by potentiating γ2 subunit containing GABA type A receptors (GABA_ARs). BZ clinical use is hampered by tolerance and withdrawal symptoms including heightened seizure susceptibility, panic, and sleep disturbances. Here, we investigated inhibitory GABAergic and excitatory glutamatergic plasticity in mice tolerant to benzodiazepine sedation. Repeated diazepam (DZP) treatment diminished sedative effects and decreased DZP potentiation of GABA_AR synaptic currents without impacting overall synaptic inhibition. While DZP did not alter γ2-GABA_AR subunit composition, there was a redistribution of extrasynaptic GABA_ARs to synapses, resulting in higher levels of synaptic BZ-insensitive α4-containing GABA_ARs and a concomitant reduction in tonic inhibition. Conversely, excitatory glutamatergic synaptic transmission was increased, and NMDAR subunits were upregulated at synaptic and total protein levels. Quantitative proteomics further revealed cortex neuroadaptations of key pro-excitatory mediators and synaptic plasticity pathways highlighted by Ca²⁺/calmodulin-dependent protein kinase II (CAMKII), MAPK, and PKC signaling. Thus, reduced inhibitory GABAergic tone and elevated glutamatergic neurotransmission contribute to disrupted excitation/inhibition balance and reduced BZ therapeutic power with benzodiazepine tolerance.

Keywords: Benzodiazepine; GABA(A) receptor; NMDA receptor; Proteomics; Sedation; Tolerance.

Fig. 1.

Repeated DZP treatment rapidly leads to sedative tolerance in mice and reduced DZP sensitivity of synaptic inhibition. (A) Diazepam-induced hypolocomotion was assayed with the open field behavioral test on the days indicated. All animals received a Veh injection at day 0 to record basal locomotor activity. Mice received either a DZP (10 mg/kg) or Veh injection on days 1–7. DZP-treated mice demonstrated sedation and a significant decrease in distance traveled compared to Veh-treated animals on day 1. By day 3, the distance traveled was not significantly different between groups, indicating emergence of DZP sedative tolerance. Two-way ANOVA found a significant effect of time (p < 0.0001), treatment (p = 0.0261), and a significant interaction (p = 0.0020), with a post-hoc Sidak test indicating a significant difference between Veh- and DZP-treated animals at day 1 (p < 0.001); n = 16 Veh-treated mice, n = 14 DZP-treated mice. (B) Baseline mIPSCs were recorded (V_hold = −70 mV) in slices from mice treated for seven days with Veh or DZP and following bath application of DZP (3 μM). (C) Basal mIPSC amplitude, frequency and decay time constant (tau) were not different in DZP-treated mice compared to Veh-treated mice. (D-E) mIPSC amplitude (D) and tau (E) were enhanced by application of 3 μM DZP in Veh-treated animals (amplitude, p < 0.0001; tau, p < 0.0001; paired t-test), but this effect was largely diminished in DZP-treated animals (amplitude, p = 0.5396; tau, p = 0.0018; paired t-test), representing a significant overall loss of DZP potentiation (amplitude, p < 0.0001; tau, p = 0.0020). (**p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001, Student’s t-test unless otherwise stated; n = 9 mice per treatment group; error bars ± S.E.M.).

Fig. 2.

Synaptic and extrasynaptic GABA_A receptor and gephyrin redistribution with seven-day DZP treatment in vivo. (A-F) Mice were treated i.p. once daily for seven days with Veh or DZP. Western blot analysis of collected cortical tissue was performed to assess protein levels of GABA_AR subunits and gephyrin from total (A, B), synaptic (C, D), and extrasynaptic (E, F) fractions. Representative blots show five mice from each treatment group. Proteins quantified by western blot were normalized to GAPDH or KIR3.2 loading control. (B) Measurements of total protein levels revealed a significant increase in α1 (p = 0.0018), β3 (p = 0.0005), and γ2 (p = 0.0002) subunits. The amount of synaptic α1 (p = 0.0019), α4 (p = 0.0028), and γ2 (p = 0.0007) subunits also increased (D), while extrasynaptic α1 (p = 0.0034) and α4 (p = 0.0010) subunits were decreased and gephyrin (p = 0.0040) was increased (F) (*p ≤ 0.05, **p < 0.01, ***p < 0.001, Student’s t-test; n = 5–10 mice per treatment; error bars ± S.E.M.).

Fig. 3.

γ2 containing GABA_A receptor composition is unchanged and tonic inhibition is reduced in DZP mice. (A) Immunoprecipitation of γ2-GABA_AR from seven-day Veh- or DZP-treated mouse cortex was analyzed by DIA mass spectrometry to assess changes in receptor subunit composition (n = 4 mice per treatment group). The intensity of α1–5 subunit-specific peptides are shown. Inset: Relative abundance (%) of α and β subunits associated with γ2 after seven-day DZP treatment. (B,C) (B) Left: representative traces with mIPSCs from seven-day DZP-treated animals before (dark red) and after (gray) 300 nM Ro 15–4513 application. Right: averaged mIPSCs before and after Ro 15–4513. (C) Quantification shows inverse agonist activity of Ro 15–4513, consistent with predominant receptors composed of γ2 with α1, α2, α3, α5-GABA_AR subunits (n = 5 cells; amplitude, p = 0.0191; frequency, p = 0.0179; tau, p = 0.0026). (D) GABA_AR-mediated tonic current was measured in acute cortical slices from mice treated i.p. once daily for seven days with Veh or DZP. Picrotoxin-sensitive changes in holding current (V_hold = −70 mV) were used to measure tonic inhibition in cortical slices from seven-day Veh- or DZP-treated mice. (E) Quantification revealed that GABA_AR-mediated tonic current was significantly reduced (p = 0.0084) in DZP-treated mice (n = 8) relative to Veh-treated mice (n = 6). (E: **p ≤ 0.01, Student’s t-test; C: *p ≤ 0.05, **p ≤ 0.01, paired t-test; error bars ± S.E.M.).

Fig. 4.

Enhanced protein expression, synaptic levels and functional activity of excitatory NMDARs was observed following seven-day DZP treatment in vivo. Total (A,B) and synaptic (C,D) levels of cortical NMDAR subunit expression were assessed by western blot analysis in mice treated i.p. once daily for seven days with Veh or DZP. Representative blots show five mice from each treatment (A,C). Quantification revealed a significant increase in the overall levels of the GluN2A (p = 0.0489) and GluN2B subunits (p = 0.0033) (B). GluN2A synaptic levels were also found to be significantly enhanced (p = 0.0110) after DZP exposure (D) (n = 7–10 mice per treatment). (E) Representative mEPSC traces recorded from acute cortical slices of mice treated i.p. once daily for seven days with Veh (black) or DZP (red) (V_hold = −70 mV). (F) mEPSC amplitude, frequency and decay were unchanged after seven days of DZP treatment (n = 6). (G) Representative evoked EPSC traces (V_hold = +40 mV). (H) The AMPA/NMDA ratio was significantly lower (p = 0.0357) in DZP-treated (n = 7) compared to Veh-treated mice (n = 8) (*p ≤ 0.05, **p ≤ 0.01, Student’s t-test; error bars ± S. E.M.).

Fig. 5.

SRM and TMT quantitative proteomics reveals enhancement of key kinases involved in synaptic receptor signaling and long-term potentiation pathways following DZP treatment. (A, B) Volcano plot of DZP-Veh differences by quantitative mass spectrometry for brain cortical homogenate peptide levels with charts of the −log₁₀(p-value) vs log2 fold change of DZP/Vehicle. (A) SRM proteomics and (B) TMT proteomics. (C) DZP/Veh ratio for shared key proteins identified by both methods as significantly changed included CaMKII subunits α/β/δ, PKCγ, Dlgap2, Rab3A, Mapk3, Nsf, and Vglut1. N = 10 mice per treatment group for both SRM and TMT proteomic analyses.

Fig. 6.

Bioinformatics analysis of DZP treatment-induced cortical changes. (A) Canonical pathway analysis using Ingenuity Pathway Analysis (IPA) identified predicted enriched activated pathways within SRM and TMT proteomic data from DZP vs Veh treated mice. Left chart shows pathway z-score status, while right chart shows −log₁₀(p-values), of the most significantly enriched pathways shared between SRM (black bar) and TMT (hatched bar) analyses. A positive z-score predicts pathway upregulation, while a negative z-score predicts pathway inhibition. An absolute z-score ≥ 2 or a −log₁₀(p-value) ≥ 1.3 as calculated by Fisher’s exact test is considered significantly activated or enriched (dashed lines in left and right charts, respectively). (B) IPA Diseases and Functions analysis using proteins which were found to be increased or decreased with a p < 0.05 predicted significant activation of long-term potentiation (LTP) (z-score = 2.428, SRM; 3.063, TMT). Red-hued icon = increased measurement; green-hued icon = decreased measurement; orange dashed line = activation of pathway; yellow dashed line = findings inconsistent with state of downstream molecule; gray dashed line = effect not predicted.