Molecular and Regulatory Mechanisms of Desensitization and Resensitization of GABAA Receptors with a Special Reference to Propofol/Barbiturate

It is known that desensitization of GABAA receptor (GABAAR)-mediated currents is paradoxically correlated with the slowdown of their deactivation, i.e., resensitization. It has been shown that an upregulation of calcineurin enhances the desensitization of GABAAR-mediated currents but paradoxically prolongs the decay phase of inhibitory postsynaptic currents/potentials without appreciable diminution of their amplitudes. The paradoxical correlation between desensitization and resensitization of GABAAR-mediated currents can be more clearly seen in response to a prolonged application of GABA to allow more desensitization, instead of brief pulse used in previous studies. Indeed, hump-like GABAAR currents were produced after a strong desensitization at the offset of a prolonged puff application of GABA in pyramidal cells of the barrel cortex, in which calcineurin activity was enhanced by deleting phospholipase C-related catalytically inactive proteins to enhance the desensitization/resensitization of GABAAR-mediated currents. Hump-like GABAAR currents were also evoked at the offset of propofol or barbiturate applications in hippocampal or sensory neurons, but not GABA applications. Propofol and barbiturate are useful to treat benzodiazepine/alcohol withdrawal syndrome, suggesting that regulatory mechanisms of desensitization/resensitization of GABAAR-mediated currents are important in understanding benzodiazepine/alcohol withdrawal syndrome. In this review, we will discuss the molecular and regulatory mechanisms underlying the desensitization and resensitization of GABAAR-mediated currents and their functional significances.


Introduction
Ligand-gated channels open in response to the neurotransmitter binding but also close (desensitize) for long periods with the agonist still bound [1,2]. It is demonstrated that desensitization of GABA A receptor (GABA A R)-mediated currents is paradoxically correlated with the slowdown of their deactivation, i.e., resensitization [3]. Desensitization tends to prolong inhibitory currents and keeps the transmitter in the bound state of GABA A Rs. The rate at which the receptors enter the desensitization state will affect the shape of inhibitory currents [4][5][6]. Table 1. Kinases and molecules involved in desensitization and slowdown of deactivation of GABA A R-mediated currents.

Neuron/Recombinant GABA A Rs Effects References
PKA Chick cortical neurons increases desensitization [7] Rat sympathetic ganglion neurons decreases peak amplitude and increases fast desensitization [8] Rat cerebellar granule cells decreases fast desensitization [9] α1β1γ2S, α1β3γ2LS increases desensitization and slows deactivation [10] PKC α1β1 decreases fast desensitization [11] PKG Rat cerebellar granule cells decreases fast desensitization [9] CaMKII Rat spinal dorsal horn neurons decreases desensitization [12] Calcineurin Rat hippocampal neurons increases desensitization and slows deactivation [4] PRIP Mouse cortical pyramidal neurons PRIP deletion increases desensitization and generates hump-like currents through increased calcineurin activity [15] GABARAP α1β2γ2L promotes clustering of GABA A Rs, facilitates deactivation, and slows desensitization [18] Hump-like GABA A R currents after a strong desensitization were also seen at the offset of propofol applications at a high concentration (600 µM) in hippocampal pyramidal neurons [19], etomidate applications at a high concentration (1 mM) in rat spinal dorsal horn neurons [20], pentobarbital applications at high concentrations (1-3 mM) in frog sensory neurons [21,22], rat hippocampal neurons [23], and recombinant GABA A Rs [24][25][26][27][28][29] or phenobarbital applications at a high concentration (10 mM) in rat hippocampal neurons [23], although these were not seen at the offset of GABA applications. Drugs that cause desensitization and resensitization of GABA A R-mediated currents are summarized in Table 2. It is believed that the generation of hump-like currents may be caused by the removal of the blockade by anesthetic agents as partial antagonists [24], although their mechanisms remain unclear and the involvement of desensitization is not necessarily denied. Propofol and barbiturate are clinically used for treatment of benzodiazepine/alcohol withdrawal syndrome [30][31][32]. Considering that hump-like GABA A R currents that are seen after a strong desensitization or blockade were evoked at the offset of propofol or barbiturate applications, the regulatory mechanisms of desensitization/resensitization of GABA A R-mediated currents might be important for understanding benzodiazepine/alcohol withdrawal syndrome. Here, we discuss the molecular and regulatory mechanisms underlying the desensitization and resensitization of GABA A R-mediated currents in neurons of PRIP-DKO mice and their functional significances.  [20] Barbiturate Pentobarbital Frog sensory neurons slows deactivation and increases apparent desensitization of GABA responses at low concentrations and directly elicits hump currents upon washout at high concentrations [21,22] Rat hippocampal neurons slows deactivation and increases apparent desensitization of GABA responses at low concentrations and directly elicits rebound currents upon washout at high concentrations [23] α1β2γ2L directly elicits tail currents upon washout at high concentrations [24,26] α1β3γ2L slows deactivation and increases apparent desensitization of GABA responses at low concentrations and directly elicits rebound currents upon washout at high concentrations [25] α1β2γ2S, α6β2γ2S directly elicits hump currents upon washout at high concentrations [27] β3 increases apparent desensitization of GABA responses and directly elicits rebound currents upon washout at high concentrations [28] α1β3γ2L directly elicits tail currents upon washout at high concentrations [29] Phenobarbital Rat hippocampal neurons slows deactivation and increases apparent desensitization of GABA responses at low concentrations and directly elicits rebound currents upon washout at high concentrations [23]

PRIP-1/2 are Involved in Desensitization and Resensitization of GABA A R-Mediated Currents
PRIP-1/2 are involved in the membrane trafficking of GABA A Rs and the regulation of intracellular Ca 2+ stores [16,17]. Thus, it was investigated whether and how the deletion of PRIP-1/2 affects GABA A R-mediated currents evoked by puff applications of GABA in layer III pyramidal cells of the barrel cortex. It was found that the deletion of PRIP-1/2 enhanced the desensitization of GABA A R-mediated currents but paradoxically induced a hump-like tail-current at the offset of the GABA puff ( Figure 1) [15]. Thus, it is likely that PRIP-1/2 are involved in the desensitization and resensitization of GABA A R-mediated currents. Although similar tail-currents were observed following the removal of propofol [19], etomidate [20], pentobarbital [21][22][23][24][25][26][27][28][29], and phenobarbital [23], it was the first report on such hump-like tail-currents that were induced by GABA itself.

[Ca 2+ ] i Dependence of Desensitization and Resensitization of GABA A R-Mediated Currents and Their Abolishment by a Calcineurin Inhibitor
It is well known that the desensitization of GABA A R-mediated currents is accelerated by increases in [Ca 2+ ] i [33,34]. As expected, it was clearly demonstrated that both the acceleration of desensitization of GABA A R-mediated currents and the generation of the hump-like tail-currents were caused by increases in [Ca 2+ ] i [15]. Consistent with the idea that desensitization is mechanistically related to the deactivation of GABA A R-mediated currents [3], the progress of desensitization of GABA A R-mediated currents was invariably accompanied by the enhancement of the hump-like tail-currents [15]. These results suggested that the deletion of PRIP-1/2 results in an enhancement of the desensitization and resensitization of GABA A R-mediated currents through increases in [Ca 2+ ] i . The involvement of CICR and the following SOCE in both the desensitization of GABA A R-mediated currents and the generation of the hump-like tail-currents in PRIP-DKO pyramidal cells was also demonstrated by an intracellular application of ruthenium red [15].
It has been demonstrated that a calcineurin inhibitor, cyclosporin A-cyclophilin A complex, suppressed the desensitization of GABA A R-mediated currents in acutely dissociated hippocampal neurons [13]. It has also been reported that the inhibition of calcineurin increased the rate of GABA unbinding from GABA A Rs [4]. Consistent with these previous studies, the bath application of a calcineurin inhibitor, fenvalerate, alleviated the desensitization of GABA A R-mediated currents and markedly decreased the hump-like tail-currents [15]. Thus, it is likely that the hump-like tail-currents in PRIP-DKO pyramidal cells were generated as a result of an acceleration of desensitization of GABA A R-mediated currents coupled with a slowdown of the GABA unbinding, which was mediated by Ca 2+ -dependent activation of calcineurin. Furthermore, Ca 2+ imaging revealed that CICR and the following SOCE were more potent in PRIP-DKO pyramidal cells than in wild-type pyramidal cells [15]. Taken together, these results strongly suggest that the enhancement of desensitization and resensitization of GABA A R-mediated currents in PRIP-DKO pyramidal cells was largely mediated by the upregulation of Ca 2+ -dependent activity of calcineurin due to the potentiation of CICR followed by SOCE.

Deletion of PRIP-1/2 Prolongs eIPSCs in Layer II/III Pyramidal Cells
The differences in the kinetic properties of GABA A R-mediated currents between pyramidal cells of wild-type and PRIP-DKO mice should be reflected in the difference in inhibitory postsynaptic responses. Then, it was investigated how inhibitory postsynaptic responses reflect the changes in the kinetic properties of the GABA A R-mediated currents in layer III pyramidal cells of the PRIP-DKO barrel cortex.
It was found that the deletion of PRIP-1/2 resulted in the prolongation of the decay phase of inhibitory postsynaptic currents/potentials (IPSCs/IPSPs) in layer II/III pyramidal cells evoked by stimulation of layer III (Figure 2), leaving the overall features of miniature IPSCs unchanged [35]. These observations suggest that the prolongation of inhibitory synaptic actions is likely to result from an enhancement of desensitization followed by an enhanced resensitization of GABA A R-mediated currents. It has been reported that the PRIP-DKO mice exhibited a reduced expression of synaptic GABA A Rs containing γ2 subunits by 40% in hippocampal neurons [36] and by 18% in cerebellar granule cells [37] as a consequence of the lack of binding between PRIP-1/2 and GABA A R-associated protein [38]. The mean peak amplitudes of the IPSCs and IPSPs in the PRIP-DKO pyramidal cells were not significantly different from those in the wild-type pyramidal cells. In any case, the amplitude of eIPSPs would not be increased by deletion of PRIP-1/2 [35]. Then, an increase in duration instead of amplitude of eIPSPs is likely to be caused in PRIP-DKO mice.

A Possible Kinetic Mechanism Underlying the Generation of the Hump-Like Tail-Currents and the Prolongation of eIPSCs
To understand the kinetic mechanisms underlying the generation of the hump-like tail-currents and the prolongation of eIPSCs, these currents were simulated using a previously proposed model [3] ( Figure 3). It was examined whether the possible increase in the fast desensitization rate (d 2 ) and the possible decrease in the unbinding rate (k off ) can lead to a generation of the hump-like tail-current at the offset of the GABA puff. It is known that GABA binding affinity was much larger in the desensitized GABA A Rs compared to the non-desensitized GABA A Rs and the binding affinity of the desensitized GABA A Rs increased depending on the concentration of the pre-applied GABA as was the case with the degree of desensitization of GABA A R-mediated currents [39]. Then, when the probability of being in the desensitized state (D fast ) for GABA A Rs was increased by increasing GABA concentration ([GABA]) or during the 2 s puff application of GABA, D fast would be further recruited, leaving Open 2 unchanged. Thus, it is reasonable to assume that the d 2 , but not β 2 , increase in a manner dependent on [GABA] [15,39]. Because Bound 2 , which is bifurcated into Open 2 and D fast , increases in a manner dependent on [GABA], the idea was incorporated in this model by defining d 2 as follows; n where d max is the maximum desensitization rate, K h is the [GABA] that yields the half maximum desensitization rate, and n is the Hill coefficient [15]. It was assumed that calcineurin increased d 2 by increasing its [GABA] dependency through a reduction of k h , and the d 2 and k off were changed between the simulated wild-type and PRIP-DKO pyramidal cells. These changes were comparable to those caused by the activation of calcineurin reported previously [4,13].
In this simulation, the onset and offset of the 2 s puff application of GABA were assumed to be attenuated with a time constant raging between 0.1 and 0.3 s. In the simulated wild-type pyramidal cell, GABA A R-mediated currents were induced without a hump-like tail-current in response to 2 s GABA puff at 50 µM [15]. In contrast, in the simulated PRIP-DKO pyramidal cell, GABA A R-mediated currents displayed a prominent desensitization and were followed by a prominent hump-like tail-current [15]. Thus, a slowdown of k off and an acceleration of d 2 resulted in a generation of a hump-like tail-current. Following a sharp decrease in [GABA] at the offset of GABA puff, a sharp decrease in d 2 to a level smaller than the fast de-desensitization (i.e., resensitization) rate constant (r 2 ) occurred to subsequently induce a hump-like tail-current. Indeed, decreases in the decay time constant at the offset of GABA puff pulse from 0.3 to 0.1 sec decreased the half-duration of the hump-like tail-current, leaving its amplitude almost unchanged [15]. Only PRIP-DKO pyramidal cells, but not wild-type pyramidal cells, displayed hump-like tail-currents in response to the same GABA puff that may have decayed slowly. These observations clearly indicate that the generation of the hump-like tail-current reflects kinetic differences between GABA A R-mediated currents in wild-type and PRIP-DKO pyramidal cells. Taken together, it can be concluded that a higher calcineurin activity in PRIP-DKO layer III pyramidal cells might have caused a slowdown of k off and an acceleration of d 2 through the modulation of its GABA concentration dependency, leading to a generation of hump-like tail-currents in PRIP-DKO pyramidal cells.
Because there were no significant differences in the single-channel current and the number of GABA A Rs between eIPSCs in PRIP-DKO and wild-type pyramidal cells [35], it can be investigated whether the increase in d 2 and the decrease in k off can also lead to the prolongation of eIPSCs. Simulated IPSCs in PRIP-DKO and the wild-type pyramidal cells that have half-durations similar to those obtained in the real experiments [35] revealed that a prolongation of eIPSCs/eIPSPs in PRIP-DKO pyramidal cells results from resensitization of GABA A R-mediated currents, which is brought about by an acceleration of d 2 through the modulation of its [GABA] dependency together with a slowdown of k off . The finding of a negative skewness coefficient in PRIP-DKO eIPSCs obtained by the nonstationary variance analysis [35] is consistent with the occurrence of de-desensitization (resensitization) of GABA A R-mediated currents during the decay phase of PRIP-DKO eIPSCs.
Based on the experimental and simulation studies, the regulatory mechanisms of GABA A Rs are schematically depicted (Figure 4). In response to a prolonged application of GABA, GABA A Rs are desensitized (d) by increased calcineurin activity due to potentiated Ca 2+ -induced Ca 2+ release (CICR) followed by store-operated Ca 2+ entry (SOCE) [15]. GABA A Rs are resensitized through de-desensitization (r) at the offset of the GABA puff. PRIP outcompetes the PLCδ in binding to GABA A R β subunits [40]. d: desensitization, r: resensitization, RYR: ryanodine receptor, SOCC: store-operated Ca 2+ channel, IP 3 R: inositol trisphosphate receptor.

Physiological Significance of Desensitization and Resensitization of GABA A R-Mediated Currents
A single whisker deflection elicits an excitation in a subset of layer IV neurons within a single barrel-related column [41], which subsequently causes an excitation in layer II/III in the same column and then spreads horizontally into neighboring columns [42,43]. The spatio-temporal profile of the excitation spread in layer II/III evoked by stimulation of layer IV was narrower and faster in the barrel cortex of the PRIP-DKO mice compared to the wild-type mice [35].
Such a horizontal excitation spread in layer II/III seems to be strictly controlled by GABA A R-mediated lateral inhibition [42,44,45]. Indeed, bicuculline application abolished such a difference in the spatio-temporal profile of the excitation spread in layer II/III between the two genotypes [35]. It is reported that the PRIP-DKO mice exhibited a greater decrease in performance in the rotarod test [36], which is commonly used to assess the sensorimotor integration [46]. Then, the enhanced phasic inhibition caused by the PRIP-1/2 deletion would suppress the inter-columnar integration in the barrel cortex, consequently decreasing spatial recognition. Further studies are required to clarify the roles of PRIP-1/2 in sensorimotor processing in the barrel cortex.

Clinical Significance of Desensitization and Resensitization of GABA A R-Mediated Currents
Central nervous system depressants slow brain activity, making them useful for treating anxiety, panic, and sleep disorders. Alcohol and benzodiazepine are useful to mitigate anxiety through enhancing GABA A R-mediated inhibition. However, alcohol and benzodiazepine are known as abused drugs. Alcohol or benzodiazepine withdrawal syndrome appears following a reduction in alcohol or benzodiazepine use after a period of excessive use [47][48][49][50]. The alcohol or benzodiazepine withdrawal symptoms typically include anxiety, sweating, hand tremor, and sleep disturbance. The underlying mechanisms involve neuronal adaptations, which are revealed as decreased GABAergic responses [51] and enhancement of NMDA responses [52][53][54][55]. Although the exact mechanism for the reduced responsiveness of GABA A Rs remains uncertain, changes in surface GABA A R protein level and subunit composition, changes in turnover, recycling, and production rates, degree of phosphorylation, and decreased coupling mechanisms between GABA and alcohol/benzodiazepine sites are thought to be involved in the reduced responsiveness [56][57][58][59]. It has recently been demonstrated that the benzodiazepine diazepam caused downregulation of GABAergic inhibition through the phospholipase C (PLCδ)/Ca 2+ /calcineurin signaling pathway [40]. The study showed that overexpression of PRIP-1 suppressed diazepam-dependent activation of PLCδ and diazepam-dependent downregulation of GABA A Rs in HEK293 cells [40], indicating that PRIP-1 acts as an inhibitor by outcompeting the PLCδ binding to GABA A Rs. Because intracellular Ca 2+ and calcineurin activity are increased in PRIP-DKO mice [15], these findings suggest that the diazepam-induced long-term downregulation of GABAergic inhibition is mediated by the PLCδ/Ca 2+ /calcineurin signaling pathway. Nevertheless, it is also true that calcineurin causes resensitization of GABA A R-mediated currents by facilitating their desensitization [4,15]. Given the apparently contradictory behaviors of GABA A R-mediated currents by calcineurin activation, the two different behaviors of GABA A R-mediated currents may depend on whether calcineurin activation occurs before or after activation of GABA A Rs.
As for the treatment of benzodiazepine/alcohol withdrawal syndrome, propofol and barbiturate which enhance GABA A R-mediated inhibition are useful. Indeed, it was demonstrated that propofol and barbiturates (pentobarbital and phenobarbital) were effective for the treatment of alcohol withdrawal syndrome [30,32] and barbiturate (pentobarbital) was effective for the treatment of benzodiazepine withdrawal syndrome [60]. However, it remains unclear how propofol and barbiturate ameliorate reduced GABA responsiveness in patients with benzodiazepine/alcohol withdrawal syndrome. Although the concentrations of propofol and barbiturates that generated the hump-like current are very high [19,21,22] compared to the dose used for treatment of the withdrawal syndrome [30,32], the generation of hump-like GABA A R currents itself may suggest the occurrence of resensitization of GABA A R-mediated currents. Indeed, the desensitization and deactivation of GABA A R-mediated currents are facilitated and slowed, respectively, by propofol/barbiturate at much lower concentrations [19,22]. Then, propofol and barbiturate may improve the reduced GABA responsiveness through the resensitization of GABA A R-mediated currents. Therefore, the regulatory mechanisms of desensitization/resensitization of GABA A R-mediated currents are important to better understand benzodiazepine/alcohol withdrawal syndrome and to develop the treatment method.
Author Contributions: Y.K. and H.T. conceptualized, drafted, reviewed, and revised the manuscript. M.S. performed a simulation study. All authors have read and agreed to the published version of the manuscript Funding: This research was funded by Japan Society for the Promotion of Science (17K07055 to Y.K. and 17K08538 to H.T.).

Conflicts of Interest:
The authors declare no conflict of interest.