Active cortical networks promote shunting fast synaptic inhibition in vivo

Abstract

Fast synaptic inhibition determines neuronal response properties in the mammalian brain and is mediated by chloride-permeable ionotropic GABA-A receptors (GABA_ARs). Despite their fundamental role, it is still not known how GABA_ARs signal in the intact brain. Here, we use in vivo gramicidin recordings to investigate synaptic GABA_AR signaling in mouse cortical pyramidal neurons under conditions that preserve native transmembrane chloride gradients. In anesthetized cortex, synaptic GABA_ARs exert classic hyperpolarizing effects. In contrast, GABA_AR-mediated synaptic signaling in awake cortex is found to be predominantly shunting. This is due to more depolarized GABA_AR equilibrium potentials (E_GABAAR), which are shown to result from the high levels of synaptic activity that characterize awake cortical networks. Synaptic E_GABAAR observed in awake cortex facilitates the desynchronizing effects of inhibitory inputs upon local networks, which increases the flexibility of spiking responses to external inputs. Our findings therefore suggest that GABA_AR signaling adapts to optimize cortical functions.

Keywords: GABA-A receptor signaling; cortex; equilibrium potential; ionic driving force; network activity; population coupling; stimulus discrimination; synaptic inhibition.

Graphical abstract

Figure 1

Measuring synaptic E_GABAAR and synaptic GABA_AR driving forces in vivo (A) Setup for performing gramicidin perforated patch-clamp recordings in V1, in combination with optogenetic activation of local GABAergic synaptic inputs. (B) Series resistance (R_s) over time, shown for neurons that met the criteria for successful perforation (mean ± SEM; n = 10 neurons, 7 mice). (C) Optogenetic approach elicits presynaptic GABA release and activates postsynaptic GABA_ARs, which are mainly permeable to Cl⁻ and, to a lesser extent, bicarbonate (HCO₃^-). (D) Light-evoked currents (voltage-clamp, VC) and potentials (current-clamp, IC). (E) Ramp protocol in perforated configuration (top) and following breakthrough into whole-cell configuration (middle). The voltage protocol before R_s correction is also shown (bottom) and consisted of a control ramp (“baseline”) and a second ramp during the light-evoked synaptic GABA conductance (“light”). (F) IV plots for baseline (black) and light (cyan) ramps performed under perforated (top) and breakthrough (bottom) conditions. The reversal potential of the baseline current (“resting membrane potential [RMP]”) and E_GABAAR are indicated with circles. (G) IV plots showing synaptic E_GABAAR before (−VU, cyan, top) and after VU application (+VU, purple, bottom). (H) Population data (n = 6 neurons from 6 mice) showed a depolarizing shift in synaptic E_GABAAR after VU (−VU: −80.9 ± 1.6 mV vs. +VU: −70.9 ± 2.6 mV; p = 0.002, paired t test). (I) VU did not affect RMP (−VU: −68.5 ± 3.5 mV vs. +VU: −69.1 ± 2.5 mV; p = 0.73, paired t test). (J) VU caused a depolarizing shift in GABA_AR driving force (DF_GABAAR = RMP − E_GABAAR, −VU: 12.4 ± 2.9 mV vs. +VU: 1.4 ± 3.0 mV; p = 0.006, paired t test). ns, not significant; ^∗∗p < 0.01.

Figure 2

Awake cortex exhibits depolarized synaptic E_GABAAR and shunting fast synaptic inhibition (A) Current clamp (IC) recording of spontaneous activity from a L2/3 pyramidal neuron in an anesthetized (An., black, top) and awake mouse (Aw., blue, bottom). (B) Probability density function for V_m in anesthetized (n = 13 cells, 7 mice) and awake (n = 12 cells, 10 mice) cortex. Awake data are also used in Figure 3. (C) Mean V_m was more depolarized in awake cortex (An.: −69.1 ± 1.4 mV vs. Aw.: −60.3 ± 1.1 mV; p < 0.001, one-way ANOVA with Bonferroni correction). (D) Mean change in subthreshold V_m was greater in awake cortex (An.: 4.9 ± 0.2 mV/ms vs. Aw.: 6.6 ± 0.3 mV/ms; p < 0.001, one-way ANOVA with Bonferroni correction). (E) Illustration of how a more depolarized V_m and higher synaptic activity are conducive to greater GABA_AR-mediated Cl⁻ influxes. (F) Averaged light-evoked postsynaptic IC responses in anesthetized (n = 18 cells, 10 mice) and awake (n = 14 cells, 11 mice) cortex. Responses in awake cortex produced V_m deflections that remained close to the RMP and could be depolarizing or hyperpolarizing (An.: Depol. 0/18 vs. Aw.: Depol. 6/14; p = 0.003, Fisher-Exact test). (G) Summary IV plot of all anesthetized (n = 10 cells, 6 mice) and awake (n = 10 cells, 9 mice) light-evoked GABA currents reveal a more depolarized synaptic E_GABAAR in awake cortex. (H) Synaptic E_GABAAR is more depolarized in awake cortex (An.: −81.1 ± 1.7 mV vs. Aw.: −63.3 ± 1.4 mV; p < 0.001, one-way ANOVA with Bonferroni correction). (I) Voltage-clamp recordings confirmed a more depolarized RMP in awake cortex (An.: −70.7 ± 1.9 mV vs. Aw.: −64.2 ± 1.1 mV; p = 0.02, one-way ANOVA with Bonferroni correction). (J) GABA_AR driving force (DF_GABAAR = RMP − E_GABAAR) was more depolarized in awake cortex (An.: 10.4 ± 2.0 mV vs. Aw.: −0.9 ± 1.7 mV; p < 0.001, one-way ANOVA with Bonferroni correction). Awake DF_GABAAR was not different to zero, consistent with synaptic GABA_ARs exerting a predominantly shunting effect (An.: p = 0.0006, one-sample t test; Aw.; p = 0.61, one-sample t test). (K) Light-evoked synaptic GABA conductances did not differ between the anesthetized and awake cortex (An. 11.04 ± 1.8 nS: vs. Aw.: 11.4 ± 1.5 nS, p = 0.92, one-way ANOVA with Bonferroni correction). ns, not significant; ^∗p < 0.05; ^∗∗p < 0.01; ^∗∗∗p < 0.001.

Figure 3

High network activity raises synaptic E_GABAAR and promotes shunting fast synaptic inhibition in awake cortex (A) Setup (top) and current-clamp (IC) recording (bottom) showing the effects of reducing local network activity by acute, local delivery of NBQX (100 μM) in awake cortex. (B) Probability density function for V_m in control awake cortex and following NBQX (Aw. blue: n = 12 neurons, 10 mice vs. +NBQX, turquoise: n = 10 neurons, 7 mice). Control data from awake group presented in Figure 2. (C) Reducing local network activity caused a hyperpolarizing shift in mean V_m (Aw.: −60.3 ± 1.1 mV vs. +NBQX: −72.6 ± 1.0 mV; p < 0.001, one-way ANOVA with Bonferroni correction). (D) Reducing local network activity caused a decrease in the mean change in subthreshold V_m (Aw: 6.6 ± 0.3 mV/ms vs. +NBQX: 0.7 ± 0.1 mV/ms; p < 0.001, one-way ANOVA with Bonferroni correction). (E) Averaged light-evoked postsynaptic IC responses (top; Aw., blue: n = 14 neurons, 11 mice vs. +NBQX, turquoise: n = 10 neurons, 7 mice). Reducing local network activity caused a hyperpolarizing shift in the polarity of light-evoked GABA currents (bottom; Aw.: Depol. 6/14 vs. +NBQX: Depol. 0/10; p = 0.03, Fisher-Exact test). (F) Summary IV plots of light-evoked GABA currents reveal more hyperpolarized synaptic E_GABAAR values following local NBQX (Aw.: n = 10 neurons, 9 mice vs. +NBQX: n = 11 neurons, 7 mice). (G) Reducing local network activity led to more hyperpolarized synaptic E_GABAAR (Aw: −63.3 ± 1.4 mV vs. +NBQX: −78.1 ± 1.3 mV; p < 0.001, one-way ANOVA with Bonferroni correction). (H) Reducing local network activity caused a hyperpolarizing shift in RMP (Aw.: −64.2 ± 1.1 mV vs. +NBQX: −71.4 ± 1.4 mV; p = 0.009, one-way ANOVA with Bonferroni correction). (I) Reducing local network activity caused a hyperpolarizing shift in DF_GABAAR (Aw.: −0.9 ± 1.7 mV vs. +NBQX: 6.7 ± 1.5 mV; p = 0.03, one-way ANOVA with Bonferroni correction). DF_GABAAR became different from zero (Aw.: p = 0.61, one-sample t test; +NBQX: p = 0.001, one-sample t test). (J) Reducing local network activity did not affect light-evoked synaptic GABA conductances (Aw.: 11.4 ± 1.5 nS vs. +NBQX: 10.4 ± 0.7 nS; p = 0.78, one-way ANOVA with Bonferroni correction). ns, not significant; ^∗p < 0.05; ^∗∗p < 0.01; ^∗∗∗p < 0.001.

Figure 4

Shunting inhibition promotes local network desynchronization and response flexibility (A) High-density Neuropixels (NPXs) recordings were used to compare spiking activity in the same cortical neurons under awake and anesthetized conditions. (B) Example raster plots show the spiking activity of a population of neurons in somatosensory cortex (SS). Spiking activity is shown across three stimulation trials for each condition. (C) Neuronal synchrony on a trial-to-trial basis (top) and the entropy of the peri-stimulus histogram (bottom) were calculated for each of the 662 recorded neurons under awake and anesthetized conditions (17 probe recordings in 16 mice). Each dot corresponds to a single neuron and the dashed line indicates the line of equality. (D) Neurons in the awake condition exhibited lower synchrony (top; Aw.: 0.325 ± 0.005 vs. An.: 0.393 ± 0.007, p < 0.001, paired t test) and higher entropy (bottom; Aw.: 1.347 ± 0.018 nats vs. An.: 1.111 ± 0.020 nats, p < 0.001, paired t test). (E) Schematic of network model consisting of interconnected excitatory pyramidal neurons (Pyr.) and inhibitory interneurons (IN). E_GABAAR in the pyramidal neurons was adjusted relative to the RMP to create two different conditions: a shunting (Shunt.) and a hyperpolarizing (Hyperpol.) E_GABAAR condition. Spiking activity was evoked by delivering brief depolarizing currents (input) of varying amplitudes to each neuron in the network. (F) Raster plots for the same population of pyramidal neurons (n = 50) in the shunting (left) and hyperpolarizing (right) E_GABAAR conditions. (G) Synchrony (top) and entropy (bottom) for pyramidal neurons in the shunting and hyperpolarizing E_GABAAR conditions (n = 1,000 randomly selected). Dashed line indicates the line of equality. (H) Neurons in the shunting E_GABAAR condition exhibited lower synchrony (Shunt: 0.784 ± 0.001 vs. Hyperpol: 0.881 ± 0.001, p < 0.001, paired t test) and higher entropy (Shunt: 0.451 ± 0.002 nats vs. Hyperpol: 0.197 ± 0.002 nats, p < 0.001, paired t test) (n = 16,000 pyramidal neurons from the model). ^∗∗∗p < 0.001. MO, motor cortex; Stim., electrical stimulation; TH, thalamus.