Neurosteroids are steroids that are locally produced in the central nervous system (CNS) independent of endocrine gland and that play an important role as endogenous neuromodulators [1,2]. Steroids can exert genomic effects by binding to nuclear receptors, whereas neurosteroids act on membrane receptors or ion channels to rapidly change the excitability of neurons and the activation of glial cells. This non-genomic effects of neurosteroids are mediated by gamma-aminobutyric acid type A (GABA_A) receptors, N-methyl-D-aspartate (NMDA) receptors, L- and T-type calcium channels, and sigma-1 receptors [1,3-5]. Brain contains a variety of neurosteroids, such as pregnane (allopregnanolone and pregnenolone), androstane (androstanediol and etiocholanolone), and estrogen classes [1,2,6,7]. Sulfated form of neurosteroids, including dehydroepiandrosterone sulfate (DHEAS), also exists in the CNS [8]. Numerous neurological disorders, such as persistent neuropathic pain, multiple sclerosis, and seizures, have altered the levels of neurosteroids in the CNS. Experimental and clinical evidence suggest that neurosteroids may have clinical applications in various neurological conditions, whereas the effects of neurosteroids are not always clear in clinical practice [6,9-11]. Thus, we aimed to review the basic mechanisms of biosynthesis, metabolism, and actions of endogenous neurosteroids in the CNS and describe therapeutic potentials of neurosteroids for the treatment of neurological disorders.

Neurosteroids are synthesized from a common cholesterol precursor. Human steroidogenic cells derive their cholesterol from two forms: uptake of cholesterol from low density lipoprotein (LDL) by means of receptor-mediated endocytosis or de novo cholesterol synthesis [12]. It is thought that the majority of the cholesterol used in neurosteroidogenesis begins with LDLs (Fig. 1). LDLs are processed by lysosomal acid lipase (LAL) to cholesterol in order to be used in neurosteroidogenesis [13]. On the other hand, cells may synthesize cholesterol from three acetyl CoA molecules [14]. A rate-limiting step is required in neurosteroidogenesis for the translocation of cholesterol across the mitochondrial membrane, such as steroidogenic acute regulatory protein (StAR) and 18 kDa translocator protein (TSPO) [15,16]. Once cholesterol comes to the outer mitochondrial membrane, its transport to the inner mitochondrial membrane is directed by the StAR [17]. StAR transfers cholesterol within the mitochondria, and it performs its role exclusively when anchored to the outer mitochondrial membrane. On the other hand, TSPO also plays a role in cholesterol transport within the mitochondria, and it is under constant study [18]. TSPO is an 18 kDa protein principally found on the outer mitochondrial membrane and is assumed to occupy 5% to 10% of proteins in the outer mitochondrial membrane of steroidogenic cells [12].

Figure 1. Schematic diagram of the steps in the steroidogenesis of progesterone. Cholesterol is synthesized from low density lipoprotein (LDL) by lysosomal acid lipase (LAL) or from acetyl CoA molecules. Steroidogenic acute regulatory protein (StAR) transfers cholesterol from cytoplasm into the mitochondria across the outer and inner mitochondrial membranes. Cholesterol is converted to pregnenolone by cytochrome P450side chain cleavage (P450scc), and pregnenolone is metabolized to progesterone by 3β-hydroxysteroid dehydrogenase (3β-HSD).

The cholesterol transported to the mitochondria is converted to pregnenolone by the mitochondrial side chain cleavage enzyme, which is referred to as P450 side chain cleavage (P450scc) [9,19]. A chain of three distinct chemical reactions causes the loss of the hydrophobic six-carbon side chain of cholesterol. A pair of electrons is needed for each of the three chemical reactions, so it requires two co-factors for P450scc to perform: 1) ferredoxin reductase, which gets electrons from NADPH, and 2) ferredoxin, which transports electrons from ferredoxin reductase to P450scc [12,20]. This process consequently involves cleavage of the side chain at the C20 and C22 positions with hydroxylation of the side chain [13]. Pregnenolone leaves the mitochondria and is metabolized to progesterone (4-Pregnene-3,20-dione) or 17OH-pregnenolone. 17OH-pregnenolone gives rise to dehydroepiandrosterone (DHEA) (Fig. 2). Both progesterone and DHEA can be metabolized to androstenedione [15]. In the next step, androstenedione can be converted to estradiol, estrone, and testosterone. Testosterone is converted to 17β-estradiol through aromatization [19]. The outline of Fig. 2 summarizes the process of neurosteroidogenesis.

Figure 2. Outline of neurosteroidogenesis. Cytochrome P450side chain cleavage (P450scc) converts cholesterol to pregnenolone, which is metabolized to progesterone by 3β-hydroxysteroid dehydrogenase (3β-HSD) or to dehydroepiandrosterone (DHEA) by cytochrome P450c17 (P450c17). Progesterone is metabolized to allopregnanolone by 5α-reductase (5α-R) and 3α-hydroxysteroid oxidoreductase (3α-HSOR) or to androstenedione by P450c17. Aromatase converts androstenedione to estrone, which is metabolized to estradiol by 17β-hydroxysteroid dehydrogenase (17β-HSD), and synthesizes estradiol from testosterone. DHP, dihydroprogesterone; DHT, dihydrotestosterone; 17-OH-PREG, 17-OH-pregnenolone; 17-OH-PROG, 17-OH-progesterone.

The conversion of pregnenolone to progesterone is catalyzed by 3β-hydroxysteroid dehydrogenase (3β-HSD). 3β-HSD exists in two isoforms: 3β-HSD1 and 3β-HSD2 [21]. While both of the isoforms likely have a similar mechanism of action considering the high similarity in structure, the 3β-HSD2 has been extensively studied and is closely related to neurosteroid biosynthesis [12]. Progesterone crosses the outer mitochondrial membrane into the cytosol via passive diffusion after synthesis by 3β-HSD. Progesterone, in turn, is metabolized by 5α-reductase and this process produces allopregnanolone. Progesterone is also able to synthesize androstenedione after conversion to 17α-hydroxyprogesterone, and this conversion is mediated by cytochrome P450c17 enzyme [9,22]. In the next step, androstenedione is reduced by 17β-hydroxysteroid dehydrogenase (17β-HSD) to produce testosterone [22]. 17β-HSD catalyzes the redox reactions of sex steroids such as estrone and estradiol. The enzyme aromatase converts testosterone into estradiol.

Neurosteroids are synthesized in the CNS and control nerve excitement quickly [23]. Inhibitory neurosteroids are a type of neurosteroid produced in the brain. These neurosteroids have the ability to influence neurotransmission. They act as positive allosteric binders of the GABA_A receptor and exhibit the following effects: antidepressant, anxiolytic, stress-reducing, rewarding, prosocial, antiaggressive, prosexual, sedative, pro-sleep, cognitive- and memory-imparing, and cited as having analgesic, anesthetic, anticonvulsant, neuroprotective, and neurogenic effects [24]. Some notable examples of this class include progesterone metabolite allopregnanolone (3α,5α-tetrahydroprogesterone), 5β-dihydroprogesterone, and androstane 3α-androstanediol [25,26].

By contrast, excitatory neurosteroids stimulate the nervous system. These neurosteroids modulate the GABA_A, NMDA, and sigma-1 receptors. They also have antidepressant, anxiogenic, cognitive- and memory-enhancing, convulsant, neuroprotective, and neurogenic effects. In addition, these neurosteroids have strong effects on neurotransmission. Excitatory neurosteroids may act as weak positive allosteric modulators of the NMDA receptor, agonists of the sigma-1 receptor, potent negative allosteric modulators of the GABA_A receptor, and mostly have excitatory effects on neurotransmission [5]. Neurosteroids, depending on the type, interact with ion channels or receptors to control brain excitability.

Sigma-1 receptor is a chaperone protein residing in the mitochondrion-associated endoplasmic reticulum (ER) membrane (MAM), which regulates inositol 1,4,5-triphosphate (IP₃) receptors to control calcium signaling between ER and mitochondrion. Sigma-1 receptor’s recent crystallographic structures only show one portion of the receptor with membrane penetration. The C-terminal portion has been identified and is believed to be in the cytoplasm. The C-terminal portion contains the ligand-binding sites as well as two steroid binding domain-like, SBDL1 and SBDL2 [5,27]. Sigma-1 receptor forms a complex with an endoplasmic protein called binding immunoglobulin protein (BiP). The dissociation of the bonds of this complex is mainly determined by the endogenous ligand neurosteroid, which affects calcium ion channels. Neurosteroids that act as agents of sigma-1 receptor include pregnenolone sulfate (PREGS) and DHEAS, and representative neurosteroids that act as antagonists include progesterone [5]. However, these neurosteroids also modulate NMDA and GABA_A receptors.

NMDA receptors are ion channels that open when a molecule mimicking calcium binds to them. They perform important functions such as synaptogenesis, synaptic plasticity, memory, and learning. NMDA receptors are heterotetramers made of two GluN1-subunits and two other subunits, which are consisted of GluN2 or GluN3 [28]. NMDA receptors have two distinct agonist binding sites: the glycine/D-serine-binding site on GluN1/GluN3 subunits and the glutamate-binding site on GluN2 subunits. The NMDA receptor has unique properties that must be activated. It needs bindings of the agonist glutamate and the co-agonist glycine or D-serine. In addition, it needs the release of Mg²⁺, blocking the influx of sodium, potassium, and calcium ions, which is induced by substantial positive charge of the inner membrane.

There are two kinds of NMDA receptor antagonists. One is a competitive antagonist such as D-2-amino-5-phosphonopentanoate, which binds on the glutamate binding site of the NMDA receptor instead of glutamate, preventing the NMDA receptor’s activity [29]. In other words, the activation of itself is prevented by blocking the glutamate from attaching itself to the binding site of NMDA receptors. The other one is non-competitive antagonists, which are also called open channel blockers [30]. After activation of NMDA receptors, Mg²⁺ is removed, and sodium ions can flow into the inner membrane freely, which generates a positive charge. Open-channel blockers, such as ketamine and phencyclidine, work here. They go to GluN2 subunits and block them, which stops NMDA receptor current. In other words, they don’t bind to glutamate binds, but they actually block open channels.

The sulfated neurosteroids PREGS and DHEAS have been shown to be potent allosteric agents in the NMDA receptor complex [31]. In general, high micromolar concentrations in PREGS and DHEAS are required to achieve action on NMDA receptor-mediated currents. PREGS can enhance the NMDA-mediated response when evaluated by electrophysiological records. Alternatively, measurement of NMDA-induced increase in intracellular Ca²⁺ in cultured neurons. PREGS inhibits GABA, glycine, and non-NMDA reactions.

Neuropathic pain is regarded as an uninvited guest that can be caused by a variety of injuries to the central or peripheral nervous system. Peripheral nerve injury can change the neurosteroidogenesis in the CNS by modulation of the activity and/or expression of neurosteroid synthesizing enzymes. It has been demonstrated that chronic constriction injury (CCI) of the sciatic nerve increases the expression of P450scc and P450c17 enzymes in the glial fibrillary acidic protein (GFAP)-positive astrocytes of the lumbar spinal cord dorsal horn in nerve-injured mice contributing to the development of neuropathic pain [32,33]. These findings suggest that neurosteroidogenesis is increased by nerve injury and plays an important role in the induction of neuropathic pain. In addition, inhibition of spinal nitric oxide synthase type II, which was increased in spinal microglial cells after nerve injury, reduced the mRNA and protein levels of P450c17 expression in CCI rats [34]. Intrathecal administration of interleukin-1β during the early phase of nerve injury inhibited the expression of P450c17 as well as sigma-1 receptor in the spinal astrocytes of CCI mice [35,36]. Thus, nerve injury-induced increased neuromodulators and/or cytokines can modulate neurosteroidogenesis in the CNS.

It has been demonstrated that neurosteroids can affect the activation of NMDA or sigma-1 receptor via indirect mechanisms. Choi et al. [32] suggested that inhibition of P450scc suppressed the CCI-induced increases in serine racemase expression and concomitant D-serine production in the spinal cord of peripheral nerve injured mice. In addition, intrathecal DHEAS administration increased protein kinase A- and protein kinase C-dependent phosphorylation of NMDA receptor GluN1 subunit in a rodent model of neuropathic pain [34]. Moreover, inhibition of the DHEA synthesizing enzyme, P450c17 reduced the expression of sigma-1 receptors in the spinal astrocytes of CCI mice [33]. Sigma-1 receptors also play an important role in phosphorylation of NMDA receptor GluN1 subunit resulting in increased activity of NMDA receptors. These findings suggest the possibility that initiation of neurosteroidogenesis can increase the expression of sigma-1 receptors and the activity of NMDA receptors by the increase in co-agonist D-serine and phosphorylation of the subunits.

Neurosteroid allopregnanolone has analgesic and neuroprotective effects. It has been suggested that allopregnanolone alleviated nerve injury- or chemotherapy-induced neuropathic pain as well as repaired oxaliplatin-induced functional alterations in peripheral nerves and intra-epidermal nerve fibers [37,38]. In addition, in vivo knockdown of the allopregnanolone synthesizing enzyme, 3a-hydroxysteroid oxido-reductase (3a-HSOR) expression in rat dorsal root ganglia increased mechanical and thermal nociception, while allopregnanolone had a potent antinociceptive effects [37]. Moreover, allopregnanolone prevented not only high glucose-induced apoptosis but also diabetic neuropathic pain [39]. Allopregnanolone is one of the most potent endogenous positive allosteric modulators of GABA_A receptor function as well as an effective ligand for membrane progesterone receptor [11]. Progesterone also has analgesic effects, but it could be due to its conversion to allopregnanolone. Thus, the effects of progesterone can be different depending on the activation of the metabolizing enzyme, 5α-reductase that catalyzes the production of allopregnanolone [40].

Multiple sclerosis is a complex disease of the CNS that has aspects of both inflammatory and degenerative [41,42]. It is a debilitating chronic inflammatory neurological condition that causes myelin and axonal injury when myelin-reactive lymphocytes infiltrate the white matter of the CNS [41]. The pathogenic process of multiple sclerosis is thought to begin with the immune system's breakdown of immunological tolerance to CNS antigens caused by genetic or environmental factors, which causes the activation of peripheral immune system and proliferation of neuroantigen-reactive T cells in susceptible individuals [43]. Activated T cells and macrophages then infiltrate the CNS and become reactivated, and they result in intraparenchymal production of chemokines and inflammatory cytokines as well as local microglial activation [44]. Subsequent waves of lymphocytic and monocytic cell infiltration into the CNS cause myelin degradation and axonal destruction [41].

An alternate theory proposes that early neurodegenerative events, such as oligodendrocyte death or structural changes to myelin, may occur and then result in inflammatory symptoms [45-47]. The breakdown of myelin and axonal degradation are the final pathogenic results that explain symptoms and indications of the disease, regardless of whether the initial event in the multiple sclerosis disease process was an innate or adaptive immunological dysregulation or a neurodegenerative/cell death phenomenon. After neurosteroids were found in the CNS, researchers worked to understand how these substances might be dysregulated in the context of multiple sclerosis and neuroinflammation [44].

Progesterone treatment of neuroantigen-reactive CD4+ T cells isolated from multiple sclerosis patients has been shown to affect their cytokine production [44,48]. Progesterone and its reduced form dehydroprogesterone (DHP) have an impact on the expression of progesterone-responsive genes by binding to intracellular progesterone receptors, whereas allopregnanolone lacks this genomic effect [44,49].

However, allopregnanolone exerts various effects on cells involved in multiple sclerosis pathogenesis [44]. Monocytoid cells, lymphocytes, oligodendrocytes, and neurons all express functional GABA_A receptors. In order to protect oligodendrocytes from harmful stimuli, allopregnanolone encourages myelin gene expression [44]. It has also been noted that neurons are protected from harm by this mechanism. The generation of inflammatory mediators by monocytoid cells is reduced as a result of allopregnanolone's binding to GABA_A receptors on these cells. In the context of autoimmune demyelination, these actions, along with decreased blood-brain barrier permeability and putative influence on lymphocytes, support allopregnanolone's positive activities [44].

Evidence for potential therapeutic roles of allopregnanolone in multiple sclerosis also comes from animal studies. Neuroinflammation and disease burden in the experimental autoimmune encephalomyelitis (EAE) animal model of disease reduced after treatment with allopregnanolone or TSPO ligands, which lead to induction of allopregnanolone-synthesizing enzymes [10,44,50]. In addition, recent studies show that ganaxolone (3α-hydroxy-3β-methyl-5α-pregnan-20-one), a 3β-methylated synthetic analog of allopregnanolone, has anti-inflammatory effects on neuroinflammation in EAE [51].

Absence seizures, which occur suddenly and last only a few seconds, are characterized by a lack of voluntary movements and unique electrographic spike-and-wave discharges (SWDs) at 2.5–4 Hz [52-55]. The majority of these generalized non-convulsive seizures are polygenic in origin and can coexist with other seizure types in a variety of age-dependent and age-independent epilepsies with varying rates of remission and expected outcomes [52,56-58]. Brief (seconds) generalized seizures with abrupt onset and resolution characterize typical absences. Clinically, they consist of the impairment of consciousness, and the electroencephalogram (EEG) shows generalized 3 to 4 Hz spike/polyspike and slow wave discharges. They are pharmacologically distinct from other seizures and fundamentally different from them. Their clinical and EEG symptoms are associated with specific syndromes. Consciousness impairment can range from severe to moderate to mild to barely noticeable. This is frequently linked to autonomic abnormalities, automatisms, and motor symptoms. Motor symptoms include clonic, tonic, and atonic components alone or in combination; the most prevalent is myoclonia, which mostly affects the face muscles [59]. This disease has a high incidence of pharmaco-resistance and the persistence of comorbidities even after full control of the seizures [60,61]. Therefore, emphasize the necessity of better and more comprehensive mechanistic knowledge to encourage the creation of new drugs and other therapeutic modalities that efficiently treat the full disease, including seizures and concomitant diseases [62]. We decided to see if neurosteroids, which have recently gained attention for their potential to treat epilepsy, are also applicable to absence seizures.

GABAergic systems related to thalamocortical interaction contribute to the generation of spike-and wave discharges, which are a hallmark of generalized absence epilepsy [52,63]. It has been suggested that intraperitoneal administration with the neurosteroids allopregnanolone and PREGS (3β-hydroxy-5α-pregnen-20-one 3-sulfate) dose-dependently promote SWDs in a genetic animal model of absence epilepsy, the Wistar Albino Glaxo Rijswijk (WAG/Rij) rat [63,64]. Consistently, a recent study by Pisu et al. [65] found that in this rodent model of absence seizures, there was an increase in allopregnanolone and allotetrahydrodeoxycorticosterone along with overexpression of α4 and δ-GABA_A subunits [63,65]. However, Citraro et al. [66] had demonstrated that local microinjection of neurosteroid in the same animal model had both dose-dependent and region-specific effects on generalized spike-and-wave activity [63,66]. Moreover, these investigators found opposite effects when comparing allopregnanolone with PREGS, which rely on the ensuing interaction with both GABA_A and glutamate receptor function. These results demonstrate how the actions of neurosteroids varies in absence seizures and suggest that careful selection should be made when considering neurosteroid therapy for this type of epilepsy [63].

A naturally occurring neurosteroid allopregnanolone and a synthetic derivative ganaxolone act as positive allosteric modulators of the GABA_A receptor, and they bind on a specific steroid recognition site. Both medications inhibit generalized tonic-clonic seizures in numerous animal epilepsy models. Citraro et al. [66] had evaluated the effects of compounds like PREGS, which acts as a positive modulator of the NMDA receptors and a negative allosteric modulator of GABA_A receptors, in the WAG/Rij rat. The occurrence of SWDs in WAG/Rij rats was typically greatly worsened by focal and bilateral microinjections of the GABA_A-positive modulators into some thalamic nuclei (nucleus ventralis posteromedialis, nucleus reticularis thalami, and nucleus ventralis posterolateralis). In contrast, these substances were effective at reducing the frequency and length of SWDs when microinjected into peri-oral region of the primary somatosensory cortex [66]. The effects of PREGS varied depending on the dose and the location of injection, but in general, at low doses in the cortex and thalamic nuclei, PREGS caused an increase in absence activity and a decrease at larger doses [66]. Collectively, neurosteroids can both act to worsen or alleviate seizures depending on the site and dose of action. Therefore, it seems important to set an appropriate dosage point and dose to be used as a drug to alleviate seizures.

This study offers new insights into a potential role of neurosteroids in the physiology and pathophysiology of neurological disorders. Neurosteroid therapy produces positive effects in experimental animal models or humans without causing harmful side effects. Thus, it could be important to develop safe neurosteroid-based analgesic or neuroprotective strategy against neurological disorders. However, the effects of synthetic neurosteroids are not always apparent in clinical practice. Thus, it further suggests that it is important to develop an effective neurosteroid therapy for the treatment of neurological diseases, and that to do this the site of action, dose level, and the activity of the neurosteroid-metabolizing enzymes in the CNS need to be considered.

None.

The authors declare no conflicts of interest.

This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (No. 2021R1F1A1060919 and RS-2023-00262398).