Anxiolytics targeting GABA_A receptors: Insights on etifoxine

Abstract

Objectives: Anxiety and adjustment disorders are among the most prevalent mental health conditions. This review focuses on γ-aminobutyric acid receptor type A (GABA_AR)-mediated anxiolysis, describing the action of both endogenous and exogenous modulators of GABA_AR. Future directions and innovative strategies to alleviate anxiety symptoms are discussed, with a particular emphasis on etifoxine.

Methods: We used available data from the recent literature to update the mode of action of anxiolytics. We focussed our search on anxiolytics acting at GABA_ARs, as well as on the pharmacological properties of formerly and currently prescribed anxiolytics.

Results: Considering the adverse effects of current treatments aimed at increasing inhibitory controls, optimisation of existing pharmacotherapies is of crucial importance. Among the alternative compounds targeting the GABAergic system, translocator protein (TSPO) ligands, such as etifoxine (EFX), which promote endogenous neurosteroidogenesis, are emerging as promising candidates for anxiety relief. In several papers comparing the efficacy of benzodiazepines and EFX, EFX showed interesting properties with limited side effects. Indeed, neurosteroids are potent GABA_AR modulators with highly underrated anxiolytic properties.

Conclusions: Novel therapeutic strategies have been emerging following the recognition of neurosteroids as potent anxiolytics. Featured at the top of the list for well-tolerated anxiety relief, TSPO ligands such as etifoxine appear promising.

KEYWORDS:

• Anxiety
• neurosteroids
• benzodiazepines
• barbiturates
• inhibition

1. Introduction

Anxiety is an emotional experience that is characterised by a state of arousal and the expectation of danger (Gutierrez-Garcia and Contreras 2013). It has an adaptive role, preparing the body to deal with future threatening stimuli arising from the environment. Nevertheless, processing of such aversive information by the brain, particularly by the amygdala circuits, is associated with a negative subjective state. When excessive, then a pathological anxious condition may be diagnosed. In DSM-5, such states include panic disorder, generalised anxiety disorder, separation anxiety disorder, specific phobias, social anxiety disorder and adjustment disorder with anxiety.

Anxiety and adjustment disorders are excessively prevalent worldwide, and frequently comorbid with other medical conditions. Although the neurobiological mechanisms underlying anxiety are still poorly understood, treatment of anxiety conditions has largely improved over the last decades. In the first half of the twentieth century, barbiturates were widely used for their sedative properties, but major drawbacks, such as dependence and lethal overdose, led to their progressive replacement by benzodiazepines (Johns 1977). Recent investigation has led to the discovery of new circuits involved in anxiety, thus broadening potential treatment strategies and giving rise to new therapeutic opportunities (Calhoon and Tye 2015). Optimisation of treatment is of crucial importance, considering the substantial side effects of current anxiolytics.

Since anxious states are classically associated with neuronal hyperexcitability in structures processing emotions, the first line of treatment was thus aimed at increasing inhibitory control, mainly by targeting γ-aminobutyric acid receptor type A (GABA_AR) in the central nervous system. In this paper, we review the molecular pharmacology of GABA_AR and its most potent modulators, with a particular emphasis on etifoxine, an interesting anxiolytic with double-targeting properties.

2. Molecular pharmacology of GABA_ARs

GABA_AR is a heteropentamer made of five subunits delimiting a chloride/bicarbonate-permeable channel (Olsen 2018). It is activated by at least two molecules of γ-aminobutyric acid (GABA), an amino acid neurotransmitter synthesised from glutamate in neurons expressing glutamate decarboxylase enzymes (i.e. GABAergic neurons). As for many ligand-gated receptor-channels the apparent affinity of GABA_ARs for GABA varies from low to high micromolar range. This parameter strongly depends on the subunit composition and, more specifically, on α-β subunit dimers forming together the agonist binding site for GABA.

Although about 19 subunits have been identified (Figure 1), there is a limited set of receptors identified in the whole nervous system. Molecular genetics and pharmacological experiments suggest the actual existence of less than 10 different GABA_AR subtypes, which is legitimate since two αβ dimers per receptor are at least required to activate GABA_AR-channels. Among the identified native GABA_AR-channels (Olsen and Sieghart 2009), those composed of α₁β₂γ₂ are presumed to account for about 60% of all GABA_ARs, while minor forms are likely to be constituted by the following subunit combinations: α₂β₃γ₂ (15–20%), α₃β_Xγ₂ (10–15%), α₄β_Xγ_X or α₄β_Xδ (5%), α₅β₂γ₂ (less than 5%), α₆β_2/3γ₂ (less than 5%) (Mohler et al. 2002).

Figure 1. Schematic representation of GABA_A receptors (GABA_ARs) as heteropentameric complexes, made up of a combination of 19 subunits, delimiting a chloride-permeable channel when activated by at least two molecules of GABA. Extracellular sites for benzodiazepines and etifoxine, two anxiolytics, are indicated on the top view representation of the receptor channel, composed in the vast majority by αβγ subunits.

GABA_AR activation is associated with an increase in chloride permeability through the plasma membrane, which is responsible for neuronal inhibition. This inhibition is generally seen as a membrane hyperpolarisation but, alternatively, may also consist of a shunting inhibition due to a local drop in membrane resistance. Both mechanisms reduce the capacity of neurons to produce action potentials or to regenerate these along their progression to synaptic endings. Recent findings have highlighted the importance of chloride gradients and their plasticity in physiological and pathological conditions. Indeed, destruction of these chloride gradients seems to contribute to several hyperexcitable neuropathologies such as epilepsy or neuropathic pain (Coull et al. 2003; Sammler et al. 2015). As illustrated in Figure 1 and Table 1, a large variety of compounds can bind to GABA_ARs. This includes endogenous modulators (e.g. endozepines and neuroactive steroids such as allopregnanolone) as well as anxiolytics such as benzodiazepines (e.g. diazepam, lorazepam), barbiturates or etifoxine.

Table 1. List of some ligands exerting a positive modulation on GABA_AR function. Binding sites, if identified, are indicated in the table with their specific properties. In the case of steroids, access to the binding site is presumed to be intracellular or via membrane lateral mobility. All other compounds bind to extracellular binding sites.

Compound	Binding site	Remark
Benzodiazepines  –diazepam  –flunitrazepam, …	α1γ subunits: high affinity α2γ, α3γ, α5γ subunits: lower affinity α6γ subunits: no affinity	Flumazenil: site antagonist β-Carbolines: negative modulation
Imidazopyridine  –zolpidem	α1γ subunits (high affinity)	Same site as benzodiazepines
Etifoxine (benzoxazines)	β subunit (β3 > β2)	Distinct site from that of benzodiazepines
(neuro)Steroids  –allopregnanolone	High-affinity site on α	GABAAR agonist if concentration > 100 nM (low-affinity site on αβ)
Barbiturates  –pentobarbital  –thiopental	α/β and γ/β subunit interfaces?	GABAAR agonist if concentration > 100 nM
Etomidate	β subunit
Propofol	β subunit (β3 > β2 > β1)
Volatile anaesthetics –isoflurane, …	α,β subunit?	Binding is poorly specific compared to that of K^+ channels of the K2P family

3. Endogenous modulators of GABA_AR function with anxiolytic properties

3.1. Endozepines

A diversity of endogenous compounds, including endozepines, can modulate GABA_AR activity. Endozepines bind to the benzodiazepine site of GABA_AR, as demonstrated by their ability to displace diazepam from its binding site (Costa and Guidotti 1991). DBI (diazepam binding inhibitor, recently renamed ACBD1-acyl-coenzyme A binding domain containing 1 protein (Fan et al. 2010)) and its two major bioactive cleavage products, TTN (triakontatetraneuropeptide) and ODN (octadecaneuropeptide), are the main endozepines. Their physiological role is controversial, with the majority of research suggesting that endozepines are negative allosteric modulators (Guidotti et al. 1983; Bormann 1991), but some authors have accumulated evidence implying otherwise (Christian et al. 2013). Although little is known about endozepines, astrocytes seem to play a major role in the release of DBI and its conversion into active peptides (Loomis et al. 2010). Endozepines also bind to a mitochondrial translocator protein (TSPO), formerly known as the peripheral benzodiazepine receptor, which will be detailed more thoroughly later in this review.

3.2. Neuroactive steroids

Neuroactive steroids have long been known to produce GABA-mediated anxiolysis (Longone et al. 2011; Schule et al. 2011); however, their precise mechanism of action is still being investigated. More precisely, 3α-reduced metabolites of progesterone (allopregnanolone), testosterone (3α-androstanediol) and corticosterone (tetrahydrodeoxycorticosterone), are recognised as the most potent endogenous GABA_AR-positive allosteric modulators. When present at high nanomolar concentrations, they increase the opening burst duration of GABA_AR chloride channels as well as the mean open time (Callachan et al. 1987; Twyman and Macdonald 1992). Synthesis of 3α-reduced neurosteroids starts with the stimulation of cholesterol transport into the mitochondria through an 18-kDa translocator protein (TSPO), expressed on the outer mitochondrial membrane (Figure 2, see also (Rupprecht et al. 2010)). Interestingly, TSPO was previously called ‘peripheral benzodiazepine receptor’ since TSPO protein complexes possess binding sites for benzodiazepines as well as binding sites for other endogenous (e.g. protoporphyrin IX) and exogenous substances (e.g. etifoxine). The molecular action of these substances stimulates neurosteroidogenesis and the production of anxiolytic neurosteroids (Wolf et al. 2015). It is important to mention here that some metabolites, namely sulphated steroids and 3β-hydroxysteroids, may also have a negative modulatory role, sometimes as antagonist at neurosteroid binding sites or as allosteric negative modulators (Wang et al. 2002; Chisari et al. 2010).

Figure 2. Simplified view of neurosteroidogenesis focussing on the synthesis of 3α-reduced neurosteroid anxiolytics targeting GABA_ARs. This production is fully ensured by nerve cells if cholesterol is used as a precursor or can be made possible by using circulating sex and stress steroids, if they reach nerve cells. Note that 3α-reduced neurosteroids are likely to reach their GABA_AR binding sites by lateral membrane diffusion. Cholesterol transport by translocator protein (TSPO) has been shown to be stimulated by several compounds including etifoxine, which also exerts an allosteric modulation of GABA_ARs. DHDOC: dehydro-deoxycorticosterone; DHP: dihydroprogesterone; DHT: dihydrotestosterone; DOC: deoxycorticosterone; THDOC: tetrahydrodeoxycorticosterone; p450scc, cytochrome p450 enzyme side chain cleavage; 5αR, 5α-reductase; 3α-HSOR, 3α-hydroxysteroid dehydrogenase.

Recent work has demonstrated the existence of two binding sites for neurosteroids on GABA_AR, distinct from those of other modulatory compounds. For instance, neurosteroids modulate GABA activity by binding to α-subunit M1 and M4 transmembrane segments, whereas direct activation of the receptor requires binding to the interface between α-subunit M1 and β-subunit M3 (Hosie et al. 2006). Since neurosteroids are lipophilic molecules, they are likely reaching their site of action through membrane lateral diffusion (Akk et al. 2005). The strong affinity of neurosteroids for extrasynaptic δ-expressing GABA_AR is also worth mentioning, considering their strong influence on neuronal excitability (Stell et al. 2003). At concentrations above 100 nM, neurosteroids develop GABA mimetic properties and can directly activate the receptor (Shu et al. 2004).

Besides their rapid effects on membrane receptors, neurosteroids also are part of the broader steroid group, and as such possess a genomic effect, impacting the transcriptional activity of numerous genes of interest. For example, allopregnanolone has been shown to inhibit corticotropin-releasing hormone (CRH) gene expression, therefore limiting CRH-induced anxiety (Patchev et al. 1994). It is also worth mentioning that sex steroids such as progesterone and oestradiol also have an impact on anxiety, through action in the amygdala (Frye and Walf 2004). Apart from the direct modulation of GABA_AR function, increase in receptor expression mediated by sex hormone genomic modulation may also account for the anxiolytic effect of neurosteroids, as shown for α₄ and δ subunits (Carver and Reddy 2016).

4. Pharmacological properties of prescribed anxiolytics acting at GABA_ARs

4.1. Barbiturates

Besides facilitating chloride channel opening, barbiturates also increase opening duration, and can act as GABA mimetics at high concentrations (ffrench-Mullen et al. 1993). Therefore, their inhibitory potential is much greater than that of benzodiazepines, which accounts for a higher risk of depression of the central nervous system. Exact binding sites of barbiturates on GABA_AR have yet to be identified, although recent studies suggest the importance of the α/β and γ/β subunit interfaces (Chiara et al. 2013; Olsen 2018). Barbiturates can be classified according to their pharmacokinetic properties, with the intermediate-acting derivatives being used as anxiolytics. Despite the discontinuation of their use for anxiety disorders, barbiturates are still indicated in various conditions, such as seizure disorder and preoperative sedation.

4.2. Benzodiazepines

Benzodiazepines are one of the most used groups of prescription drugs around the world. They produce a potent anxiolysis but also possess sedative, muscle relaxant and anticonvulsant properties. Unfortunately, their use is impaired by major unwanted side effects such as retrograde amnesia, functional tolerance and a high dependence risk. In hippocampal neurons, it has been shown, for example, that withdrawal from a 5-day treatment with flunitrazepam in adult rats was sufficient to silence GABAergic synapses and produce severe withdrawal symptoms (Poisbeau, Williams et al. 1997). These side effects are particularly hard to manage in the long term, and this is critical in the elderly since benzodiazepines may also lead to psychomotor impairment and promote cognitive decline.

Benzodiazepine sites are found at the αγ subunit interface, and benzodiazepine affinity depends on α subunit identity (Olsen 2018). For example, α₁ subunit-containing receptors are highly sensitive to most benzodiazepines (and to zolpidem) whereas those containing α₂ or α₃ are less sensitive to benzodiazepines. If benzodiazepines act as positive allosteric modulators, i.e. potentiating GABA_AR function by increasing the affinity for GABA, modulation at this αγ site can be blocked by the antagonist flumazenil (also called a silent agonist) or negatively modulated by the β-carbolines.

The recent development of α subunit-selective benzodiazepines with possible limited adverse side effects showed that targeting the α₂-containing GABA_ARs was the most efficient way to promote anxiolysis but also analgesia and sedation (Table 2). These studies also showed that potentiating α₁-containing GABA_ARs could lead to major sedative and amnesic effects, while cognitive decline could be mostly attributed to the potentiation of α₅-containing GABA_ARs. This opens a new avenue for the development of symptom-specific benzodiazepines but, so far, refinement of benzodiazepine molecules has not proven its efficacy to prevent functional tolerance or benzodiazepine dependence upon cessation of their use.

Table 2. Summary of the effects of benzodiazepines, either clinically relevant or adverse side effects, and the presumed α subunit composition of GABA_ARs (Adapted from (Tan et al. 2011)). Note that α4 or α6 subunit-containing GABA_ARs containing are not shown since they are benzodiazepine insensitive.

	α subunits	Ref.
Clinically relevant effects
Sedation	α1βXγX	(Rudolph et al. 1999; McKernan et al. 2000)
Anxiolysis	α2βXγX	(Low et al. 2000; McKernan et al. 2000; Crestani et al. 2001)
Muscle relaxation	α2βXγX α3βXγX α5βXγX	(Low et al. 2000; McKernan et al. 2000; Crestani et al. 2001; Dias et al. 2005)
Anti-convulsive	α1βXγX	(Rudolph et al. 1999)
Side effects
Amnesia	α1βXγX α5βXγX	(Rudolph et al. 1999; McKernan et al. 2000; Collinson et al. 2002; Crestani et al. 2002;Cheng et al. 2006)
Addiction	α1βXγX	(Heikkinen et al. 2009; Tan et al. 2010)

Ligand binding at αγ benzodiazepine sites on GABA_ARs may rapidly be associated with an uncoupling between the GABA and benzodiazepine site, likely explaining its loss of function (i.e. functional tolerance). Apart from this acute tolerance explanation, long-term use of benzodiazepine is associated with changes in the membrane expression of GABA_ARs. These newly expressed receptors often display a lower affinity to benzodiazepines or even a full pharmacoresistance, as seen sometimes in temporal lobe epilepsy (Leroy et al. 2004). The molecular mechanisms explaining these changes are still poorly understood but likely involve intracellular phosphorylation processes, as suggested by in vitro studies (Ali and Olsen 2001). This intracellular pathway seems to be strongly recruited when benzodiazepines promote sedation, i.e. after activation of α₁γ-containing GABA_ARs. Taken together, it is very clear that high-affinity benzodiazepine binding sites contribute to a large part to functional tolerance and the main challenge to treat anxiety states is to design new benzodiazepine compounds that target α₂γ-containing receptors.

An alternative to benzodiazepines is to use non-benzodiazepine compounds with distinct binding sites on GABA_ARs and limited adverse side effects. Etifoxine (Stresam) is one of the rare anxiolytics that is available on the market and that displays this property.

5. Etifoxine, a non-benzodiazepine anxiolytic

Etifoxine is a non-benzodiazepine (benzoxazine-derived) anxiolytic prescribed in several countries, including in France since 1979, for the treatment of adjustment disorders with anxiety (Servant et al. 1998; Nguyen et al. 2006). Its anxiolytic effect involves potentiation of GABA_AR function, but etifoxine’s mechanism of action is quite original as it seems to use both direct and indirect pathways.

The direct positive allosteric modulation is observed after binding to a selective site, distinct from that of benzodiazepines (Verleye et al. 1999). Indeed, etifoxine binds to GABA_ARs on β subunits (Hamon et al. 2003), with a marked preference for β2 and β3 (Figure 1 and Table 1). As illustrated in Figure 3, this binding leads to a positive allosteric modulation of chloride currents resulting from the opening of GABA_ARs (Schlichter et al. 2000). Increase in the affinity for the GABA binding site results in an overall decrease of the half maximal amplitude (EC₅₀) (i.e. concentration of GABA producing a 50% modulation in the dose–response curve), an increase in the current amplitude in the case of non-saturating GABA concentrations and an increase in the current duration (see examples in Figure 3).

Figure 3. Positive allosteric modulation of GABA_AR currents by the anxiolytic etifoxine (EFX). Graph on the left illustrates the GABA_AR current amplitude (in %) when spinal neurons are submitted to increasing concentrations of GABA (red). In the presence of etifoxine (EFX, 60 µM), the dose–response curve is shifted to the left, indicating an increase in the apparent affinity for GABA. Indeed, the concentration of GABA inducing GABA_AR-mediated current of half the maximal amplitude (EC₅₀) is reduced from 20 to 7 µM. Bar chart in inset indicates that this modulation is more selective for GABA_AR containing β₂ and β₃ subunits (adapted from (Hamon et al. 2003)). Another representation of this effect is shown on the right with representative traces. Etifoxine increased the amplitude and duration of GABA_AR-mediated currents evoked by non-saturating concentration of GABA (a, 10 µM). In the case of saturating concentration (b, 200 µM), EFX only increased the duration of the current, as predicted by the dose–response curve.

As shown in Figure 2, etifoxine also binds to the mitochondrial TSPO. This specific protein has recently received some attention since it may be a promising target for the development of novel anxiolytics with minimal adverse side effects (Nothdurfter et al. 2012). Binding of etifoxine to TSPO favours the entry of cholesterol into the mitochondria and its conversion by several enzymes to the potent endogenous anxiolytic called allopregnanolone (Verleye et al. 2005; do Rego et al. 2015; Wolf et al. 2015). Anxiolysis is then achieved by a neurosteroid-based potentiation of GABA_AR function. Together, the direct and indirect actions of etifoxine on GABA_ARs lead to synergistic and long-lasting anxiolytic effects, at least in animal models (Verleye and Gillardin 2004; Ugale et al. 2007; Verleye et al. 2009, 2011).

The potentiation of GABA_AR function is not exclusively restricted to a global anxiolytic action of etifoxine. Besides its initial description as an anticonvulsant (Kruse and Kuch 1985, 1986), etifoxine’s preclinical efficacy has been proven to limit acute pain expression (Juif et al. 2015), to alleviate pain symptoms resulting from an inflammatory or neuropathic insult (Aouad et al. 2009; Aouad, Petit-Demouliere, et al. 2014; Aouad, Zell, et al. 2014; Zeitler et al. 2016; reviewed in Poisbeau et al. 2014; Choi and Kim 2015), to exert a neuroprotective action in various models (Girard P et al. 2009; Ravikumar et al. 2016; Simon-O'Brien et al. 2016) and to promote nerve regeneration (Girard C et al. 2008; Zhou et al. 2013, 2014). As a follow-up of several preclinical reports from our laboratory (Aouad et al. 2009; Aouad, Petit-Demouliere, et al. 2014; Aouad, Zell, et al. 2014; Poisbeau et al. 2014; Juif et al. 2015; Zeitler et al. 2016) and others, a recent review has also highlighted the potential interest of using etifoxine for pain patients with anxiety (Choi and Kim 2015).

6. Future directions

Considering the risks associated with benzodiazepines, tremendous effort has been devoted to the development of novel anxiolytic strategies. These include compounds specifically targeting the α₂ and α₃ GABA_AR subunits, which are the most relevant for anxiolysis. Although promising results were found in preclinical research (Griebel et al. 2003), clinical trials so far have failed to validate the anxiolytic potential of these drugs, and this line of research has been discontinued (de Haas et al. 2009). Alternative methods also include the use of anxiolytics targeting neurotransmitter systems other than the GABAergic one. Indeed, enhancement of serotoninergic activity through buspirone, a 5-HT_1A receptor partial agonist, or antidepressants such as selective serotonin uptake inhibitors (SSRIs), comprise a commonly employed method to reduce anxiety symptoms (Lucki 1996; Reinhold et al. 2011). However, given the relatively slow onset of action of antidepressants, they are often prescribed in combination with other anxiolytics in the first weeks of treatment. Evidence has also shown a role for the glutamatergic system in anxiety, with antagonism of metabotropic glutamate receptors inducing anxiolysis (Dunayevich et al. 2008). The antihistamine hydroxyzine, an H₁ receptor inverse agonist as well as a 5-HT₂A receptor antagonist, also seems effective in the treatment of anxiety (Lader and Scotto 1998). Overall, these molecules still appear to carry a GABAergic component, as serotoninergic fibres act on GABA neurons, and as reducing glutamate activity favours general inhibition.

Currently, perhaps the most promising pharmacological alternative for the alleviation of anxiety conditions lies with the neurosteroids. Although they constitute one of the most potent groups of GABA_AR modulators, their anxiolytic properties have long been underrated, and they may represent a promising option for treatment with limited side effects. It has been pointed out that many anxiolytics currently prescribed have the ability to promote neurosteroid synthesis, which may mediate part of their efficacy. For instance, diazepam and other benzodiazepines can bind to the TSPO (hence its former appellation ‘peripheral benzodiazepine receptor’), and SSRIs appear to correct the neurosteroid imbalance seen in patients with anxiety (Uzunov et al. 1996; Pinna et al. 2006).

In addition to being a direct modulator of GABA_AR function, EFX is also a TSPO ligand, and, as such, the pleiotropic effects of 3α-reduced neurosteroids are intimately involved in its benefits. In several papers comparing the efficacy of EFX and benzodiazepines, EFX showed interesting properties with limited side effects, which opens the door for a new generation of anxiolytics targeting the TSPO (Nguyen et al. 2006; Stein 2015). Mitochondrial TSPO ligands may act in almost every cell type of the organism, and the stimulation of neurosteroidogenesis in nervous and peripheral tissues is far from being trivial. For example, rare severe liver injuries and toxidermia have been suspected to be possibly associated with EFX treatment in two recent studies published by French groups (Moch et al. 2012; Cottin et al. 2016). Although the double-targeting properties of drugs such as EFX show considerable potential, further investigation on the molecular mechanisms underlying TSPO ligand efficacy is still needed. Their better understanding could, inter alia, facilitate the development of synthetic neurosteroid analogues.

7. Conclusion

Despite the variety of agents approved for the relief of anxiety conditions, the search for the optimal drug is still ongoing. Although it is unlikely that one unique compound would have beneficial effects on the whole spectrum of anxiety and adjustment disorders, the recognition of neurosteroids as potent anxiolytics gave rise to novel therapeutic strategies. It is, for example, interesting to note that the production of pregnenolone after TSPO stimulation has been recently suggested to exert a potent inhibition of cannabinoid receptor type 1 (CB1), thus limiting psychotic-like states for this drug of abuse (Busquets-Garcia et al. 2017). Following the innovative trail of etifoxine, TSPO ligands are emerging as a promising method for well-tolerated anxiety relief.

Statement of interest

Pierrick Poisbeau received financial support from Biocodex to investigate the molecular mechanisms of action of etifoxine.

Acknowledgements

None.

Additional information

Funding

The following institutions gave their support to the project: Centre National de la Recherche Scientifique and Université de Strasbourg, Région Alsace. GG is a fellow from the French Ministère de la Recherche et de l’Enseignement Supérieur.