When Are Depolarizing GABAergic Responses Excitatory?

Abstract

The membrane responses upon activation of GABA(A) receptors critically depend on the intracellular Cl^- concentration ([Cl^-]_i), which is maintained by a set of transmembrane transporters for Cl^-. During neuronal development, but also under several pathophysiological conditions, the prevailing expression of the Cl^- loader NKCC1 and the low expression of the Cl^- extruder KCC2 causes elevated [Cl^-]_i, which result in depolarizing GABAergic membrane responses. However, depolarizing GABAergic responses are not necessarily excitatory, as GABA(A) receptors also reduces the input resistance of neurons and thereby shunt excitatory inputs. To summarize our knowledge on the effect of depolarizing GABA responses on neuronal excitability, this review discusses theoretical considerations and experimental studies illustrating the relation between GABA conductances, GABA reversal potential and neuronal excitability. In addition, evidences for the complex spatiotemporal interaction between depolarizing GABAergic and glutamatergic inputs are described. Moreover, mechanisms that influence [Cl^-]_i beyond the expression of Cl^- transporters are presented. And finally, several in vitro and in vivo studies that directly investigated whether GABA mediates excitation or inhibition during early developmental stages are summarized. In summary, these theoretical considerations and experimental evidences suggest that GABA can act as inhibitory neurotransmitter even under conditions that maintain substantial depolarizing membrane responses.

Keywords: KCC2; NKCC1; SLC12A2; SLC12A5; chloride homeostasis; gaba receptor; neuronal development.

Figure 1

Stoichiometry and typical operation of secondary active Cl⁻ transporters. The GABA_A receptor (GABA_A-R) mediates mainly Cl⁻ fluxes and to a lesser extent HCO₃⁻ fluxes. NKCC1 mediates uptake of two Cl⁻ ions with of one K⁺ and one Na⁺ ion. KCC2 mediates extrusion of one Cl⁻ with one K⁺ ion. The anion-exchanger (AE3) is supposed to mediate uptake of one Cl⁻ ion in antiport with one HCO₃⁻ ion. The Na⁺-dependent Cl⁻/HCO₃⁻ exchanger (NDCBE) utilizes the Na⁺ gradient to extrude Cl⁻ ions.

Figure 2

Dependency between [Cl⁻]_i and GABAergic actions. (A) Schematic diagrams illustrating the two exemplary effects of GABAergic inputs (blue traces) on glutamatergic inputs (red traces) of different intensities. A GABAergic hyperpolarization augments the distance between peak glutamate depolarization and the AP threshold (hyperpolarizing inhibition). At a passive Cl⁻-distribution GABA does not affect E_m, but the decreased membrane resistivity induced by GABA reduced the peak glutamate depolarization (shunting inhibition). The dashed line represents a hypothetical action potential threshold. (B) Schematic diagrams illustrating that the combination of the membrane potential shift with the shunting effect caused by the decreased membrane resistivity augments the effect of a hyperpolarization inhibition (left traces) and can lead to inhibition even at depolarizing GABAergic membrane responses (right traces). (C) [Cl⁻]_i-dependency of the membrane potential (E_m) calculated for five different GABAergic conductances (g_GABA, normalized to g_Input) under stationary conditions (see main text for details). The gray plane represents AP threshold. Note that considerable g_GABA in combination with high [Cl⁻]_i is needed for a suprathreshold GABAergic depolarization. (D) Dependency of g_AMPA^Rheoshift (normalized to g_Input, ; g_AMPA^Rheoshift _{= gAMPA} in presence of GABA minus _gAMPA in absence of GABA) on [Cl⁻]_i and g_GABA. Red traces indicate excitatory and blue traces inhibitory GABAergic effects. Note that g_GABA^Rheoshift becomes negative at identical [Cl⁻]_i independent of g_GABA.

Figure 3

Temporal profile of GABAergic shunting and GABAergic depolarizing effects on excitatory glutamatergic inputs. (A) The upper traces illustrate GABAergic (blue line) and glutamatergic (orange and gray lines) currents provided at latencies between −150 and +150 ms. The lower traces illustrate the postsynaptic potentials (PSPs) evoked by these currents. Note that the PSPs outlast the synaptic currents. (B) Compound PSPs induced by the co-stimulation of GABA and glutamate synapses, with the glutamatergic inputs provided at latencies between −150 and +150 ms. (C) Peak amplitude of compound PSP, normalized to the glutamatergic PSP in the absence of GABA, plotted against the latency between AMPA and GABA stimuli, as shown in (B). Note that the compound PSP amplitude drops if glutamatergic synapses are activated within a narrow interval around coincident stimulation, but increases when AMPA receptors are stimulated several ms after the GABA input.

Figure 4

Spatial profile of GABAergic depolarization and GABAergic shunting effects on excitatory glutamatergic inputs. (A) Relative amplitude of GABAergic PSPs, as measured at the soma, upon activation of a depolarizing GABA synapse (E_GABA = −52 mV) at different dendritic positions. The voltage traces above graph illustrate GABAergic PSPs at 0%, 25%, 50%, 75%, and 100% of the dendritic length. Scale bar in (A–C) is 5 mV/500 ms. (B) Profile of the GABAergic shunting effect on glutamatergic inputs, calculated by normalizing the amplitude of the compound PSPs obtained in the presence of GABA (purple traces) to the EPSC amplitude obtained in the absence of GABA inputs (orange trace). In these experiments the shunting effect was isolated by maintaining E_GABA at resting membrane potential. Note that GABA synapses located proximally to the AMPA synapse (“on-path”) mediate a stable shunting effect, while for GABA synapses distal to the AMPA synapse (“off-path”) the shunting effect declines rather fast. (C) Effect of a depolarizing GABAergic input (E_GABA = −52 mV) at different positions along the dendrite on the peak compound PSP amplitude during co-stimulation The blue traces represent purely GABAergic PSPs, the orange trace the glutamatergic PSP, and the purple traces the compound PSPs upon co-activation of AMPA and GABA synapses. Note that GABA inputs mediate an inhibitory effect when co-localized with the AMPA synapse, while at more distant on-path and off-path synapses an excitatory effect is observed.