Pharmacological Analysis of GABAA Receptor and Sigma1R Chaperone Interaction: Research Report I―Investigation of the Anxiolytic, Anticonvulsant and Hypnotic Effects of Allosteric GABAA Receptors’ Ligands

Two groups of facts have been established in previous drug development studies of the non-benzodiazepine anxiolytic fabomotizole. First, fabomotizole prevents stress-induced decrease in binding ability of the GABAA receptor’s benzodiazepine site. Second, fabomotizole is a Sigma1R chaperone agonist, and exposure to Sigma1R antagonists blocks its anxiolytic effect. To prove our main hypothesis of Sigma1R involvement in GABAA receptor-dependent pharmacological effects, we performed a series of experiments on BALB/c and ICR mice using Sigma1R ligands to study anxiolytic effects of benzodiazepine tranquilizers diazepam (1 mg/kg i.p.) and phenazepam (0.1 mg/kg i.p.) in the elevated plus maze test, the anticonvulsant effects of diazepam (1 mg/kg i.p.) in the pentylenetetrazole-induced seizure model, and the hypnotic effects of pentobarbital (50 mg/kg i.p.). Sigma1R antagonists BD-1047 (1, 10, and 20 mg/kg i.p.), NE-100 (1 and 3 mg/kg i.p.), and Sigma1R agonist PRE-084 (1, 5, and 20 mg/kg i.p.) were used in the experiments. Sigma1R antagonists have been found to attenuate while Sigma1R agonists can enhance GABAARs-dependent pharmacological effects.


Introduction
The evidence that the effect of the non-benzodiazepine anxiolytic fabomotizole depends on the interaction with Sigma1R [1] became a key stimulus for setting up the study of GABA A receptor's (GABA A R) interplay with Sigma1R chaperone.
The rationale for the work was also based on previously established neurochemical data, which determined the direction of the search for fabomotizole. GABA A Rs are the major mammalian CNS receptors mediating rapid inhibitory neurotransmission. Structurally, they are pentameric ligand-linked chloride channels activated by the endogenous agonist γ-aminobutyric acid (GABA) [2]. In 1994, S. Deutsch et al. summarized the evidence of a decrease in the binding capacity of the benzodiazepine site of GABA A Rs upon stress exposure [3]. We confirmed these results. It was found that emotional stress causes a decrease in 3 H-labeled benzodiazepine binding in open-field (OF) stress-sensitive BALB/c mice [4]. In OF stress-resistant C57Bl/6 mice, no changes were registered [5][6][7]. More potent predator exposure stress caused a decrease in specific benzodiazepine binding in BALB/c and C57Bl/6 mice as well [7,8]. Similar data were obtained for Maudsley reactive (MR) and Maudsley nonreactive (MNRA) rats [9].
Keeping in mind that anxiolytic effect depends on the binding capacity of the GABA A R allosteric benzodiazepine site, we focused on the repeatedly confirmed phenomenon of stress-induced specific benzodiazepine binding decrease in the search for a new anxiolytic [10,11]. In other words, we believed that a substance preventing the stress-induced drop in specific benzodiazepine binding would possess anxiolytic effects.
Thus, the ability of fabomotizole to prevent stress-induced drop in specific benzodiazepine binding, and its interaction with chaperone Sigma1R, defining the anxiolytic properties, were established. These results determined the task of investigating the dependence of drugs effects mediated through GABA A Rs on the interaction of GABA A Rs with Sigma1R.
In the present work, using Sigma1R antagonists and agonists, we studied anxiolytic, anticonvulsant, and hypnotic effects.

Anxiolytic Effect of Benzodiazepines
Anxiolytic effects of diazepam and phenazepam were reproduced on BALB/c mice in the elevated plus maze test (EPM). Diazepam at a dose of 1.0 mg/kg and phenazepam at a dose of 0.1 mg/kg administered 30 min before the animals' exposure to the EPM statistically significantly increased number of entries and time spent in the open arms (Figures 1 and 2,  Tables S2 and S3 Table S2). The Sigma1R antagonist NE-100 at a dose of 1.0 mg/kg administered 30 min before diazepam did not influence its anxiolytic effect. However, when the dose was increased to 3.  Table S2). Administration of BD-1047 prior to diazepam resulted in a peculiar feature of increased number of entries into closed arms (N closed) compared to NE-100 (p < 0.01) ( Figure S1, Table S2). Pretreatment with the Sigma1R agonist PRE-084 at a dose of 1.0 mg/kg had no influence on diazepam effect in the EPM (Table S4).
Sigma1R  Table S3), consequently decreasing the total entries parameter (N total p < 0.05, p < 0.01) ( Figure S2, Table S3).  Administration of vehicle before the EPM exposure had no effect on the behavior of BALB/c mice except for a slight increase in the % of time spent in the open arms (%T open; p = 0.048) compared with intact animals. The Sigma1R antagonists BD-1047 at a dose of 1.0 mg/kg and NE-100 at doses of 1.0 and 3.0 mg/kg, injected 60 min before testing, had no effect on mouse behavior in the EPM (Table S1).
To exclude the association of Sigma1R antagonists' effects on the anxiolytic action of diazepam and phenazepam with the affinity to the benzodiazepine site of GABA A Rs, the competitive displacement of [N-methyl-3 H] flunitrazepam in P2 fractions of BALB/c mouse brain homogenates was evaluated. BD-1047 and NE-100 were found to have no affinity to the benzodiazepine site of GABA A R ( Figure S3).

Anticonvulsant Effect of Benzodiazepines
Clonic jerks, generalized clonic seizures, generalized tonic seizures, and their time of onset were recorded during the experiment (Figure 3).  Tables S5 and  S6). In control experiments, BD-1047 at doses of 1, 10, and 20 mg/kg had no effect on PTZ-induced seizure thresholds (Figures 4 and S4, Table S5). Diazepam administered at a 1 mg/kg dose 30 min before PTZ increased the clonic jerk threshold 2.9-fold and generalized clonic and tonic seizures 1.8-fold. Pre-injected BD-1047 (1-20 mg/kg) attenuated the effect of diazepam in a dose-dependent manner, statistically significantly reducing the PTZ dose inducing generalized clonic and tonic seizures by 21% and 26% at the 20 mg/kg dose ( Figure 4, Table S5). The findings indicate the ability of a 20 mg/kg BD-1047 dose to inhibit the anticonvulsant effect of diazepam.
PRE-084 showed no independent effect on PTZ-induced seizure activity ( Figure S5, Table S6). Pretreatment of mice with 5 mg/kg PRE-084 statistically significantly enhanced the anticonvulsant effect of diazepam at a 1 mg/kg dose, increasing all seizure reaction thresholds ( Figure 5, Table S6). Raising the dose of PRE-084 to 20 mg/kg increased the anticonvulsant effect of diazepam only on the thresholds of clonic jerks and generalized tonic seizures (Figure 5a,c, Table S6) but not generalized clonic seizures (Figure 5b, Table S6). PRE-084 increased the anticonvulsant effect of diazepam in a dose-dependent manner when registering the threshold of generalized tonic seizures ( Figure 5c, Table S6). The obtained data indicate the ability of high doses of Sigma 1R agonist PRE-084 to enhance the anticonvulsant effect of diazepam.  (c) generalized tonic seizures. The dotted line represents the threshold for PTZ-induced seizures. The upward arrow indicates that the compound increases the seizure threshold and has anticonvulsant effects. PRE-084 (5 and 20 mg/kg) was administered i.p. 90 min before the PTZ infusion. Diazepam (1 mg/kg) was administered i.p. 30 min before the PTZ infusion. Data are presented as mean ± S.E.M. Statistically significant differences vs. diazepam (1 mg/kg) according to one-way ANOVA with Dunnett post-hoc test: * p < 0.05; ** p < 0.01; **** p < 0.0001.

Hypnotic Effect of Pentobarbital
To evaluate the influence of Sigma1R ligands on the pharmacological effects dependent on the barbiturate binding site of GABA A Rs, an experiment was performed on a pentobarbital-induced sleep model in mice ( Figure 6).  Table S7) and reducing pentobarbital-induced sleep duration ( Figure 7b, Table S7). At a dose of 10 mg/kg, BD-1047 statistically significantly reduced only pentobarbital-induced sleep duration ( Figure 7b, Table S7).  Table S7). Increasing the dose of PRE-084 to 5 mg/kg revealed a statistically significant decrease in falling asleep time ( Figure 8a, Table S7) and an increase in pentobarbital-induced sleep duration (Figure 8b). The results indicate opposite effects of Sigma1R antagonists and agonists on the hypnotic properties of pentobarbital.
PTZ was shown to act as a GABA A R competitive antagonist through the binding with the picrotoxin site, decreasing the frequency of Cl − channel opening rather than an openchannel state duration [27,28]. We tested the effect of selective Sigma1R agonist PRE-084 and antagonist BD-1047 in a mouse PTZ-induced seizure model to prove that Sigma1R is engaged in the regulation of anti-convulsive activity of diazepam. Single administration of PTZ at a submaximal dose is generally used to model epileptic seizures in rodents, which is believed to reflect myoclonic seizures in humans [29].
In the present study, the GABA A R-dependent anxiolytic and anticonvulsant effects of benzodiazepines (diazepam, phenazepam) and the hypnotic effect of pentobarbital were attenuated by Sigma1R ligands with antagonistic activity.
The negative influence of Sigma1R antagonists on the anxiolytic-like effect of benzodiazepines is consistent with the anxiety-like behavior of Sigmar1 −/− mice in standard tests [47]. In our study, the Sigma1R agonist PRE-084 (1 mg/kg) did not alter the anxiolyticlike effect of diazepam, but in in vivo experiments, compounds with Sigma1R agonist properties can interfere with anxiety-like behavior [48][49][50][51][52].
The attenuation of diazepam anticonvulsant action by Sigma1R antagonist BD-1047 established in our study is consistent with the results of E. Vavers et al. [53], who showed a decrease in the threshold of seizures induced by PTZ [28] and GABA A Rs antagonist bicuculine [54] in mice with Sigmar1 gene inactivation. These effects of PTZ and bicuculine were independent of the expression levels of α4, α5, β3, γ2, and δ GABA A R subunits, which did not change in Sigmar1 −/− mice [53,55]. Similar to [56], our study showed that Sigma1R antagonists had no effect on the pro-convulsive activity of PTZ at doses less than 20 mg/kg. It is important to note that the independent pro-convulsive activity of NE-100 detected at doses above 25 mg/kg was attenuated by Sigmar1 gene inactivation [53,56]. The effects of Sigma1R antagonists on GABA A R-dependent effects were recorded not only in vivo but also in vitro. Thus, NE-100 increased synaptosomal transporter-mediated [ 3 H]-GABA uptake but enhanced the negative effect of flumazenil, a benzodiazepine GABA A R site antagonist, in the process. Moreover, NE-100 reduced stimulated transporter-mediated and exocytotic release of [ 3 H]-GABA from nerve terminals [57]. The above data together with the lack of affinity of BD-1047 and NE-100 to the benzodiazepine site of GABA A Rs and competition of NE-100 for [ 3 H]-muscimol binding site [53] indicate the ability of Sigma1R antagonists to prevent pharmacological effects of PAMs of GABA A Rs by changing functional activity of Sigma1R at relatively low doses.
In our study, PRE-084 at doses of 5 and 20 mg/kg increased the anticonvulsant effect of diazepam. It also shortened the falling asleep time (5 mg/kg PRE-084) and prolonged pentobarbital-induced sleep duration (1 and 5 mg/kg PRE-084). Similarly, prior single administration of the non-selective Sigma1R ligand opipramol, which has anxiolytic properties [58], also increased the latency of the PTZ-induced clonic seizures [59]. The data on the attenuation of seizures by Sigma1R ligands with agonist properties [60] are consistent with the results obtained in the anhedonia modeling with the proconvulsive GABA A R antagonist picrotoxin. Only fluvoxamine and S-(+)-fluoxetine, high-affinity Sigma1R ligands with agonist properties [61,62], were shown to attenuate picrotoxin-induced anhedonia [63], unlike antidepressants with lower affinity for Sigma1R [64]. Pre-administration of Sigma1R antagonist BD-1047 prevented this effect of fluvoxamine. However, the anti-anhedonic effect of paroxetine interacting with Sigma1R in the micromolar concentration range was manifested in this experimental model only when administered together with the selective Sigma1R agonist (+)-SKF-10.047 [64]. In PTZ-induced seizures and kainic-acid-induced status epilepticus models, compounds with agonist activity to Sigma1R (SKF-10.047, dextromethorphan, and carbetapentane) and Sigma1R positive modulators (SKF83959, SOMCL-668, and E1R) showed anticonvulsant properties. The effects of Sigma1R positive modulators were eliminated by prior administration of BD-1047 or NE-100 [56,65,66]. The opposite influence of Sigma1R antagonists and agonists on the effects of PAMs of GABA A R found in our work is consistent with the study results where PRE-084 attenuated NE-100-induced seizures in mice [56]. The effects of Sigma1R ligands and modulators in various seizure models are discussed in the review [67]. Thus, most experimental data point to attenuation of seizures by compounds with Sigma1R agonist activity. However, in Zebrafish, the scn1a mutant model of Dravet syndrome antagonistic ligand regulation of Sigma1R attenuates epileptiform behavior [68]. Seizure disorders in Dravet syndrome mouse models caused by mutations in the Scn1a gene are accompanied by disturbances in GABAergic firing in hippocampal neurons [69], which may also indicate attenuation of epileptiform behavior due to Sigma1R-dependent modulation of GABA A Rs activity.
In our study, the Sigma1R antagonist BD-1047 attenuated the hypnotic effect of pentobarbital. The current literature does not provide sufficient data on the relationship of Sigma1R with sleep regulation and the effects of barbiturates. However, orexin-systemmediated/controlled sleep-wake cycle effects may be dependent on Sigma1R [70,71]. The role of chaperone-dependent reduction of ER stress in sleep normalization is also known [72].
Based on the data on the relationship between the subunit composition of GABA A Rs and the effects mediated by them, we might predict which subtypes of GABA A Rs could be involved in the regulatory effects of Sigma1R. Diazepam binds with the benzodiazepine site of GABA A Rs at the α+/γ2interfaces. Synaptic subtypes of GABA A Rs that mediate the sedative (α1βγ2) [73], anticonvulsant (α1-containing GABA A Rs) [74], hypnotic (α2-, α3-, or α5-containing GABA A Rs) [19] and anxiolytic (α2βγ2 as well as α2/α3-containing GABA A Rs) [75] effects were shown to be the main targets of diazepam. Pentobarbital predominantly allosterically modulates α4β2γ2, α4β2δ [76], and α1β3δ [77] GABA A Rs subtypes. Sigma1R-mediated effects may also depend on the interactions with extrasynaptic GABA A Rs subtypes, predominantly represented by α4-6, β2/3, and δ subunits [21,[78][79][80].
However, given the opposite influence of Sigma1R antagonists and agonists on the effects of GABA A R PAMs interacting with different GABA A R binding sites, we cannot rule out other Sigma1R-dependent mechanisms of GABA A R regulation modulating the functional activity of the latter in general. Possible mechanisms of Sigma1R chaperone influence on GABA A R-dependent effects can be discussed based on the knowledge of Sigma1R physiological properties.
First, Sigma1R-dependent influence on the effects of GABA A R PAMs may be mediated by chaperone properties toward cytoplasmic membrane proteins involved in the regulation of anxiety and seizures [44]. Under misfolded proteins' accumulation conditions (ER stress) or ligand activation, Sigma1R is capable of intracellular translocation within lipid domains, including the plasma membrane region [12,38,81,82]. On the other hand, antagonistic action on the chaperone is able to reduce Sigma1R levels in the cell surface membrane and the expression of α and β subunits of GABA A Rs in vivo [83]. Examples of such target proteins of the chaperone are cannabinoid (CB 1 R) [67,[84][85][86] and NMDA glutamate (GluN1, GluN2) receptors [87][88][89][90][91], ligands of which inhibit anxiolytic effects of benzodiazepine tranquilizers [92,93]. To date, no data on protein-protein interactions of Sigma1R with GABA A Rs have been found in the available literature. Clarification of these issues requires further investigation.
Second, it is possible that the effects of PAMs of GABA A Rs are dependent on the dissociation of activated Sigma1R from the main chaperone ER BiP (GRP78, HSPA5) [32,38,46], which contributes to the increased activity of both chaperones [32,38,[94][95][96], enhanced protein folding [43,97], and regulation of ER stress sensor IRE1α, which triggers the unfolded protein response (UPR) signaling [39,42,98]. This assumption is supported by data on the interaction of the BiP chaperone with GABA A Rs subunits [99][100][101][102] and the enhancement of their folding and cell trafficking properties [103][104][105]. According to [106] the anticonvulsant drug valproate [107,108] is able to enhance BiP expression without activating ER stress.
Third, Sigma1R plays an important role in the maintenance of Ca 2+ cell homeostasis [38,[109][110][111]. Interestingly, Sigma1R activation is also accompanied by PKC translocation to the membrane vicinity [112], which may contribute to the phosphorylation of GABA A Rs and enhance the potentiating effect of PAMs on GABA A Rs [113]. The influence of Sigma1R antagonists on the effects of PAMs of GABA A Rs detected in our study is consistent with the down-regulation of protein kinase C (PKC) in Sigmar1 −/− mice [114]. E. Sallard et al. [20] discuss the possibility of activation and inhibition of GABA A Rs under conditions of Ca 2+ release from the ER at low and high concentrations, respectively [115,116]. A similar regulation of GABA A Rs may depend on the activated Sigma1R ability to increase intracellular calcium, eliminated by chaperone antagonists [117][118][119]. The contribution of Sigma1Rdependent regulation of L-and N-type voltage-gated Ca 2+ channels [120] to the effects mediated by GABA A Rs [121,122] is not excluded.
Fourth, it is plausible that chaperone Sigma1R contribution to the regulation of GABA A R-dependent effects is mediated by Sigma1R action on cell membranes. The influence of Sigma1R on the lipid environment of GABA A Rs is consistent with the established role of Sigma1R in cholesterol metabolism in cell membranes and reorganization of lipid raft proteins [42,[123][124][125][126][127][128]. This property of Sigma1R can influence the functional activity of GABA A R, which depends on the lipid environment [129,130]. This assumption is supported by the fact that the stimulatory effect of GABA and diazepam on GABA A R chloride current depends on lipid raft integrity [131]. In turn, pentobarbital anesthesia contributes to disruption of lipid-protein stability [132].
Therefore, our study demonstrates for the first time the opposite action of Sigma1R antagonists and agonists on the effects of allosteric modulators of GABA A R. The findings demonstrate that Sigma1R chaperone contributes to the mechanisms of GABA A Rsdependent effects. Further studies will be aimed at revealing the mechanisms of interaction between Sigma1R and GABA A Rs.

Ethical Approval
All experimental procedures were approved by the bioethics committee of the FSBI "Zakusov Institute of Pharmacology", protocols #09 of 29 October 2022 and #03 of 31 January 2023. All applicable national [133] and international [134] guidelines for the care and use of experimental animals were followed.

In Vivo Experimental Design
In vivo experimental design was developed in compliance with the 3R principles [135].
(3) Barbiturate-induced sleep: The Sigma1R agonist PRE-084 and antagonist BD-1047 were dissolved in saline solution for injection immediately before administration. Diazepam was dissolved in 20% PEG 400 solution [136][137][138]. Pentobarbital sodium salt (pentobarbital) was dissolved in saline solution [147]. The studied substances were injected intraperitoneally into mice at the rate of 0.1 mL per 10 g of animal weight. All animals were injected with pentobarbital in the dose of 50 mg/kg 60 min after the first injection of the studied drugs ( Figure 6). Immediately after pentobarbital injection, the mice were placed in individual transparent Plexiglas boxes with openings for ventilation.

Elevated Plus Maze Test
The accordingly, abiding by recommendations [148].

Pentylenetetrazole-Induced Seizures
Injection of PTZ into the lateral caudal vein of experimental mice was performed according to previously described methods with some modifications [56,[149][150][151][152]. The animals were kept in transparent Plexiglas boxes with holes for ventilation and tail. During the intravenous infusion, the tail was kept outside the box to access the lateral caudal vein, so the animal could move freely in the box without strain on the catheterized tail. PTZ at a 10 mg/mL concentration was injected at a constant rate of 6 µL/s, set on an MD-1020-K BASi Bee Hive Controller 240 V/50 Hz (BASi Corporate Headquarters, West Lafayette, IN, USA) connected to an MD-1001 BASi Bee Baby Bee syringe drive (BASi Corporate Headquarters, West Lafayette, USA). A 27G needle with an attached infusion cannula was injected intravenously into the lateral caudal vein [146]. The insertion site was preheated with an infrared lamp according to the recommendations [56,[149][150][151][152]. Accuracy of vein penetration and absence of thrombosis was confirmed by the presence of blood in the catheter. The minimum dose required to induce a seizure was considered the seizure threshold. The infusion was stopped when a generalized tonic seizure was observed. Animal behavior was recorded via video camera and evaluated according to previously described criteria [153]. In this experiment, clonic jerks, generalized clonic seizures, and generalized tonic seizures were recorded.

Barbiturate-Induced Sleep
The test was performed according to previously described methods with some modifications [154][155][156][157][158]. After intraperitoneal infusions, the animals were kept in a transparent Plexiglass box with ventilation holes and sawdust bedding. The animals could move freely without restrictions. Injection of test substances was performed according to the scheme shown above (Figure 6). After intraperitoneal injection of pentobarbital (50 mg/kg), animal behavior was recorded on video. Falling asleep time in seconds was recorded by loss of the righting reflex, sleeping time in seconds was recorded from the moment of falling asleep to the moment of recovery of spontaneous righting reflex.

Statistical Analysis
To evaluate the experimental data distribution, D'Agostino-Pearson, Shapiro-Wilk and Kolmogorov-Smirnov tests were used. Statistical significance was calculated using one-way ANOVA (Sidak or Dunnet post hoc tests) or the Kruskal-Wallis test (Dunn's post hoc test). The data are presented as a mean with standard deviation (mean ± S.D.), mean with standard error of the mean (mean ± S.E.M.), or median with interquartile range (Mdn (q25-75)). A value of p < 0.05 was considered to be statistically significant. Statistical analysis and visualization were performed using GraphPad Prism software version 8.0.1 for Windows (GraphPad, La Jolla, CA, USA, www.graphpad.com (accessed on 28 May 2023)).

Conclusions
The study shows for the first time the multidirectional effects of compounds with Sigma1R antagonist and agonist properties on GABA A Rs-dependent in vivo effects. Sigma1R antagonists inhibited anxiolytic-like, anticonvulsant, and hypnotic effects of GABA A Rs PAMs, whereas Sigma1R agonist enhanced their anticonvulsant and hypnotic effects. The obtained results may be a consequence of the ligand influence on the functional activity of Sigma1R, including chaperone interactions with receptors, enzymes and ion channels, regulation of BiP-dependent signaling, and cell membranes remodeling, which requires further studies.   Rodent sodium channel protein type 1 subunit alpha gene scn1a Zebrafish sodium channel protein type 1 subunit alpha gene