J Pharmacol Sci 100, 93 – 118 (2006) Journal of Pharmacological Sciences ©2006 The Japanese Pharmacological Society Critical Review

The Sigma1 Protein as a Target for the Non-genomic Effects of Neuro(active)steroids: Molecular, Physiological, and Behavioral Aspects François P. Monnet1 and Tangui Maurice2,* 1Unité 705 de l’Institut National de la Santé et de la Recherche Médicale, Unité Mixte de Recherche 7157 du Centre National de la Recherche Scientifique, Université de Paris V et VII, Hôpital Lariboisière-Fernand Widal, 2, rue Ambroise Paré, 75475 Paris cedex 10, France 2Unité 710 de l’Institut National de la Santé et de la Recherche Médicale, Ecole Pratique des Hautes Etudes, Université de Montpellier II, cc 105, place Eugène Bataillon, 34095 Montpellier cedex 5, France

Received December 15, 2005

Abstract. Steroids synthesized in the periphery or de novo in the brain, so called ‘neuro- steroids’, exert both genomic and nongenomic actions on neurotransmission systems. Through rapid modulatory effects on neurotransmitter receptors, they influence inhibitory and excitatory neurotransmission. In particular, progesterone derivatives like 3α-hydroxy-5α-pregnan-20-one (allopregnanolone) are positive allosteric modulators of the γ-aminobutyric acid type A (GABAA) receptor and therefore act as inhibitory steroids, while pregnenolone sulphate (PREGS) and dehydroepiandrosterone sulphate (DHEAS) are negative modulators of the GABAA receptor and positive modulators of the N-methyl-D-aspartate (NMDA) receptor, therefore acting as excitatory neurosteroids. Some steroids also interact with atypical proteins, the sigma (σ) receptors. Recent studies particularly demonstrated that the σ1 receptor contributes effectively to their pharmaco- logical actions. The present article will review the data demonstrating that the σ1 receptor binds neurosteroids in physiological conditions. The physiological relevance of this interaction will be analyzed and the impact on physiopathological outcomes in memory and drug addiction will be illustrated. We will particularly highlight, first, the importance of the σ1-receptor activation by PREGS and DHEAS which may contribute to their modulatory effect on calcium homeostasis and, second, the importance of the steroid tonus in the pharmacological development of selective σ1 drugs.

Keywords: neuro(active)steroid, sigma1 receptor, neurotransmission, neuronal plasticity, learning and memory

1. Neurosteroids and σ receptors ...... 94 3. Physiological aspects of σ receptor and neurosteroid functions ... 101 1.1. Neuro(active)steroids biosyntheses 3.1. Does σ binding protein interfere with (neuro)steroid synthesis? 1.2. Neurosteroids play a role in the excitatory/inhibitory balance 3.2. Neurosteroids and σ drugs share modulatory functions at both in the brain pre- and post-synaptic levels 1.3. The σ receptor, an atypical neuromodulatory system 3.2.1. Do neurosteroids and σ drugs exhibit similar effects 1.4. The σ1 receptor acts as an intracellular amplifier of signal on neuronal firing? transduction system involved in the formation 3.2.2. Neurosteroids and σ drugs affect neuronal excitability and recomposition of membrane lipid microdomains induced by the NMDA receptor 2. Neurosteroids and σ drugs apparently share the same binding sites 3.2.3. Neurosteroids and σ drugs may have a similar impact ...... 98 on neurotransmitter release 2.1. Steroids and neurosteroids bind both σ1 and σ2 sites 4. Behavioral effects of σ-receptor ligands and neurosteroids...... 105 in the central nervous system 4.1. Neurosteroids and σ drugs affect learning and memory 2.2. Steroids and neurosteroids bind both σ1 and σ2 sites processes in peripheral tissues 4.1.1. Pro-mnesic effects of neurosteroids and σ drugs 2.3. Do neurosteroids also bind atypical σ-receptor subtypes? 4.1.2. Anti-amnesic effects in cholinergic models of amnesia 4.1.3. Anti-amnesic effects in NMDA-receptor-dependent amnesia *Corresponding author. [email protected] 4.1.4. Anti-amnesic effects of neurosteroids and σ drugs Published online in J-STAGE: February 11, 2006 during aging DOI: 10.1254/jphs.CR0050032 4.2. Neurosteroids and σ drugs may influence abused drug intake 4.3. Cross-influences between neurosteroids and σ receptor Invited article 5. Conclusions...... 111

93 94 FP Monnet and T Maurice

1. Neurosteroids and σ receptors P450 side-chain cleavage (P450scc) enzyme, a conver- sion that constitutes the rate-limiting step in steroido- The first description of a biological activity for genesis (3). PREG could then be metabolized into steroids, the anesthetic properties of progesterone, dates progesterone by a 3β-hydroxysteroid dehydrogenase from the 1941 report by Selye (1). The subsequent (3βHSD). PREG could also be converted into 17α- characterization of central effects exerted by steroid hydroxypregnenolone by a cytochrome P450c17, which hormones, and their syntheses in the nervous system leads to dehydroepiandrosterone (DHEA) and andros- defining the concept of neurosteroids, dates from 25 tenedione by the scission of the c17,20 bond. These years ago. Neuroactive steroids include both steroids two pathways lead to the formation of pregnanes and from the periphery, which are transported through the androstanes steroids, respectively, and the most highly blood-brain-barrier and act within the brain, and locally expressed steroids in the brain are PREG, DHEA, their synthesized neurosteroids. Their physiological actions, sulphate esters (PREGS and DHEAS), progesterone, demonstrated from embryogenesis through adult life, and 3α-hydroxy-5α-pregnan-20-one (allopregnanolone) involve genomic actions, mediated by steroid receptors among the tetrahydroprogesterone isomers. The expres- translocating into the nucleus, and non-genomic neuro- sion, distribution, and ontogeny of most of the modulatory actions affecting directly several ion chan- steroidogenic enzymes and synthesized steroids have nels, neurotransmitter receptors, and second messenger been determined in the brain (3) and similarities with systems. The syntheses and effects of these neuro the peripheral synthetic pathway were identified. Most (active)steroids and their physiopathological conse- of the enzymes, including P450scc, P450c17, 3βHSD, quences have been extensively reviewed (2 – 13). In and 5α-reductase, which convert PREG into progester- particular, the mechanisms by which they act as allos- one, are co-expressed in limbic structures such as the teric modulators of the γ-aminobutyric acid type A hippocampus, caudate putamen, hypothalamic nuclei, (GABAA) receptor and N-methyl-D-aspartate (NMDA) cortex, olfactory bulb, and cerebellum (3). type of glutamate receptor is now extensively docu- Various physiological and pathological conditions are mented. They also affect acetylcholine systems through associated with changes in neurosteroid levels. Neuro- direct and indirect actions. More atypical is their re- steroid syntheses significantly vary during acute and ported interaction with the sigma1 (σ1) receptors (14 – chronic stress, pregnancy, neural development, and 17). In the present review article, we will detail the normal and pathological aging. In humans, plasma levels molecular, physiological, and behavioral data support- of DHEAS decline with age (22, 23). In rodents, ing the concepts that not only certain neurosteroids significant decreases of PREGS levels in aged Sprague interact at physiological concentration with σ1 receptors Dawley rat brain were reported to correlate with and may represent their endogenous ligands but also that impaired memory functions (24). Aged C57BL/6 mice the neurosteroid/σ1 receptor interaction may present or senescence-accelerated (SAM) mice also show major physiopathological consequences. We will parti- important decreases in the brain levels of PREG and cularly illustrate their involvement in memory processes PREGS (PREG/S), DHEA and DHEAS (DHEA/S), and vulnerability to drug addiction. or progesterone (25, 26). Moreover, brain structural abnormalities related to Alzheimer’s disease, like both 1.1. Neuro(active)steroids biosyntheses β-amyloid deposits and neurofibrillary tangles, which Steroids are synthesised in adrenal glands and gonads result from the aggregation of pathologic τ proteins, and exert their hormonal effects in peripheral organs and affect brain neurosteroid levels. Brown et al. (27) the brain. The identification of pools of steroids whose reported that β1–42-amyloid protein increased DHEA levels were higher in the brain than in plasma, indepen- levels in a human glia-derived cell line, after a 24 h dently of peripheral sources led to the concept of ‘neuro- application, a synthesis interpreted as a neuroprotective steroids’ (18 – 20). These neurosteroids are synthesized response to the β-amyloid-induced toxicity. Neuro- locally in the mitochondria of glial cells, such as oligo- steroid levels were measured in two in vivo models of dendrocytes and astrocytes, and neurons. Fifteen days β-amyloid toxicity, the intracerebroventricular injection after removing the sources of circulating steroids by of aggregated β25–35-amyloid peptide in mice and the adrenalectomy and gonadectomy (AdX/CX) in rodents, chronic infusion during a 2-week period of β1–40-amyloid no difference in the brain neurosteroid levels could be protein in rats. Decreased levels of PREG/S, DHEA/S, measured as compared to non-operated animals (18, 19, and progesterone were identified in the hippocampus, 21). Cholesterol is transported into the mitochondrion by cortex, and cerebellum as compared to the control the peripheral type of benzodiazepine receptor and then animals (28, 29). Neurosteroid levels were also mea- converted into pregnenolone (PREG) by a cytochrome sured post-mortem in individual brain regions of Neurosteroid Action at σ1 Receptor 95

Alzheimer’s disease patients and aged non-demented to rat cuneate nucleus slices and enhanced the binding of 3 controls, including the hippocampus, amygdala, frontal the GABAA-receptor agonist [ H]muscimol. Extensive cortex, striatum, hypothalamus, and cerebellum (30), electrophysiological studies have been performed, parti- supporting a general trend towards decreased levels of cularly using recombinant GABAA receptors (for a all steroids observed in brain regions of Alzheimer’s review, see ref. 34). Progesterone, its metabolites disease patients compared to controls. PREGS levels 3α,5α-tetrahydroprogesterone (allopregnanolone) and were also significantly lower in the striatum, and 3α,5β-tetrahydroprogesterone (alloepipregnanolone), and cerebellum, as were DHEAS levels in the hypothalamus, 3α,5α-THDOC are potent stereoselective positive striatum, and cerebellum. In contrast, progesterone and allosteric modulators of the GABAA receptor. Patch- allopregnanolone levels were reduced, but non-signifi- clamp studies showed that these steroids had no effect on cantly, brain structures, including the hypothalamus, the single-channel conductance of the receptor, but striatum, frontal cortex, or amygdala. Finally, a signifi- greatly promoted the open state of the GABA-gated ion cant negative correlation was found between the levels channel (35). At physiologically relevant concentra- of cortical β-amyloid peptides and those of PREGS in tions, that is, below 100 nM, these steroids directly the striatum and cerebellum and between the levels of activated the GABAA receptor–channel complex (35, phosphorylated τ proteins and DHEAS in the hypo- 36) and exerted a GABAmimetic effect sufficient to thalamus (30). Since high levels of key proteins suppress excitatory neurotransmission (36). Their implicated in the formation of plaques and neuro- effects affect synaptic as well as non-synaptic GABAA fibrillary tangles were correlated with decreased brain receptors and a clear heterogeneity of their interaction levels of PREGS and DHEAS, the authors, in agreement with GABAA receptors has been observed, the with Brown et al. (27), supported the concept of a selectivity depending on individual brain regions and possible neuroprotective role of these neurosteroids in even population of neurons (34). Therefore, progester- Alzheimer’s disease. one, allopregnanolone and its stereoisomers, and 3α,5α- THDOC are potent inhibitory neurosteroids. DHEAS 1.2. Neurosteroids play a role in the excitatory/inhi- has been reported to antagonize the GABAA receptor by bitory balance in the brain interacting with the barbiturate site. In particular, These neurosteroids, as well as circulating steroids DHEAS decreased the potency of pentobarbital to crossing the blood-brain-barrier and penetrating the potentiate the [3H]flunitrazepam binding and inhibited brain, can influence neuronal functions. Among neuro- GABA-induced currents in neurons (37). High micro- steroids, allopregnanolone and 3α,5α-tetrahydrodeoxy- molar concentrations of DHEAS inhibited [3H]musci- corticosterone (3α,5α-THDOC) can be oxydized into 5α- mol and [3H]flunitrazepam binding to rat brain mem- dihydroprogesterone or 5α-dihydrodeoxycorticosterone, branes, primarily by reducing the binding affinities. respectively, which have the ability to bind to the cyto- DHEAS also produced a concentration-dependent solic progesterone receptor and subsequently activate blockade of GABA-induced currents in cultured neurons transcription factors, regulate gene expression, and from ventral mesencephalon (38). DHEAS acts as a stimulate protein synthesis (10 – 12, 31, 32). Other non-competitive modulator of the GABAA receptor. neurosteroids bind to membrane-bound receptors to Indeed, DHEAS blocked the GABAA receptor in a exert rapid, non-genomic effects by binding to or indi- primary culture of ventral midbrain neurons of fetal rats rectly modulating the activity of neurotransmitter recep- by accelerating the desensitization and not by acting on tors or ion channels. Accordingly, two types of neuro- the conductance of the chloride channel (39). steroids can be differentiated in the brain on the basis of PREGS and DHEAS positively modulate several their pharmacological actions: excitatory steroids, which NMDA-receptor-mediated responses, and thus appear to include mainly PREG/S and DHEA/S, the sulphated be excitatory neurosteroids. In the nanomolar range, forms being more active than the free steroid; and PREGS specifically enhanced the NMDA-gated cur- inhibitory steroids, which include mainly progesterone rents in spinal cord neurons (40), intracellular Ca2+ and its reduced metabolites 3α/β,5α/β-tetrahydro- fluxes mediated through NMDA-receptor channels in progesterone, with allopregnanolone presenting the cultured rat hippocampal neurons or chick hippocampal highest brain levels. neurons (41, 42), convulsant potency of NMDA in mice The first report of an interaction between neuro- (43), and NMDA-induced phasic firing of vasopressin steroids and the GABAA receptor is from Harrison and neurons in the rat supraoptic nucleus (44). DHEAS Simmonds (33) and likely corresponded to the Selye’s potentiated the intracellular Ca2+ fluxes mediated observation in 1941. They demonstrated that alphaxa- through NMDA-receptor channels in mouse neocortical lone increased GABA action when the latter was applied neuronal cultures (45). However, several evidences 96 FP Monnet and T Maurice show that PREGS and DHEAS markedly differ in their established (52). At the same time, σ sites were also effect on the NMDA neurotransmission. PREGS, but demonstrated to be different from the high affinity not DHEAS, enhanced the NMDA-induced phasic firing phencyclidine (PCP) binding sites, located within the of vasopressin neurons in the rat supraoptic nucleus (44). ion channel associated with NMDA receptors (52). The On the contrary, DHEA/S, but not PREG/S, potentiated lack of selectivity between the σ and PCP binding the NMDA-evoked catecholaminergic release (15) and sites showed by several compounds, including benzo- firing activity of CA3 hippocampal neurons (16). More- morphans or PCP derivatives, led to a confusion that was over, the NMDA-stimulated [3H]norepinephrine release cleared up by the availability of new highly selective is inhibited by PREGS (15). Each neurosteroid thus drugs. Among them, the reference PCP non-competitive acts differently on the NMDA receptor complex. antagonist (+)MK-801 maleate (dizocilpine) fails to PREGS behaves also as a positive modulator of NMDA displace radioligands labeling the σ sites and selective σ receptors, whereas epipregnanolone sulphate behaves as agonists like 1,3-di-O-tolylguanidine (DTG), (+)N- a negative modulator by inhibiting NMDA-induced cyclopropylmethyl-N-methyl-1,4-diphenyl-1-ethyl-but- conductances. These two types of steroids act at specific, 3-en-1-ylamine hydrochloride (JO-1784, igmesine), 2- extracellularly directed sites that are distinct from one (4-morpholino)ethyl-1-phenylcyclohexane-1-carboxylate another and from the spermine, redox, glycine Mg2+, hydrochloride (PRE-084), and 1-(3,4-dimethoxy- PCP, and arachidonic acid sites (46, 47). Studies using phenethyl)-4-(3-phenylpropyl)piperazine dihydrochloride recombinant receptors showed that the NR2A subunit (SA4503) do not bind to the NMDA-receptor-associated controls the efficacy of the neurosteroids enhancement, PCP site. These compounds are now reference com- but not inhibition, which suggests in addition that pounds in terms of selectivity between σ and PCP potentiating and inhibitory neurosteroids act at distinct receptors. sites on the NMDA receptor (48). In other respects, there Multiple criteria are now used to identify σ receptors, is still no convincing evidence that DHEAS act directly especially their ability to bind several chemically unre- at the NMDA receptor. lated drugs with high affinity. These drugs include Progesterone was reported to attenuate by itself the psychotomimetic benzomorphans, PCP and derivatives, excitatory neuronal responses to local application of cocaine and derivatives, amphetamine, certain neuro- quisqualate, kainite, and NMDA of cerebellar Purkinje leptics, many atypical antipsychotic agents, anticonvul- cells (49). It appeared however that two classes of neuro- sants, cytochrome P450 inhibitors, monoamine oxidase steroids could be distinguished, based on their effects on inhibitors, histaminergic receptor ligands, peptides from the NMDA neurotransmission, namely excitatory neuro- the neuropeptide Y (NPY) or calcitonin gene-related steroids: PREGS and DHEAS, which potentiate the peptide (CGRP) families, substance P, and neuroactive NMDA-receptor activation through a direct or indirect steroids. The σ binding sites could be labeled by mechanism, and inhibitory neurosteroids: epipreg- various specific radioligands, including [3H](+)SKF- nanolone sulphate and progesterone, which inhibit the 10,047, [3H](+)3-(3-hydroxyphenyl)-N-(1-propyl)-piperi- NMDA-receptor activation. dine ([3H](+)3-PPP), [3H]haloperidol, [3H]DTG, and [3H](+)pentazocine. Pharmacological structure/activity 1.3. The σ receptor, an atypical neuromodulatory system studies led to the definition of two subclasses of σ sites, The σ receptor represents a unique binding site in named σ1 and σ2 (53). The two sites were distinguished mammalian brain and peripheral organs, distinct from based on their different drug selectivity patterns and any other known transmitter receptors. Historically, the molecular weights. The σ1 site shows a stereoselectivity σ receptor was identified by Martin et al. (50) as one of with high affinity for the dextrogyre isomers of benzo- the subtypes of opiate receptors, differentiating in the morphans, whereas σ2 sites show the reverse stereo- chronic spinal dog, the unique psychotomimetic effects selectivity with a lower affinity range (54). DTG, (+)3- induced by N-allylnormetazocine (SKF-10,047) (σ- PPP, and haloperidol are non-discriminating ligands syndrome), from the effects induced by morphine (µ- with high affinity on both subtypes. In addition, several syndrome) and ketocyclazocine (κ-syndrome). However, biochemical features were proposed to be selectively subsequent studies established that σ sites possess observed with σ1 receptors such as an allosteric modu- negligible affinity for naloxone or naltrexone and that lation by phenytoin (55) and sensitivity to pertussis toxin certain behaviors elicited by SKF-10,047 were resistant or G-protein modulators (56, 57). The σ1 receptor is a 29- to the blockade by classical opiate receptor antagonists kDa single polypeptide that has been cloned in several such as naloxone or naltrexone (51). A complete animal species and humans (58 – 60). The ligand bind- distinction between the non-opiate σ binding sites and ing profile for the cloned σ1 receptors were similar as the classical µ-, δ-, and κ-opiate receptors was therefore described in brain homogenates studies. The 223 amino Neurosteroid Action at σ1 Receptor 97 acid sequence of the purified protein is highly preserved, (25, 65). with 87% – 92% identity and 90% – 93% homology The σ2 site is not yet cloned. It was first characterized among tissues and animal species. The protein appeared in pheochromocytoma PC12 cells (69). It presents low identical in peripheral tissues and brain and also shares affinity for (+)benzomorphans and has an apparent a similarity, 33% identity and 66% homology, with a molecular weight of 18 to 21 kDa (54). Some selective sterol C8 –C7 isomerase (see paragraph 3.1), but no and high affinity σ2 site ligands are now available such as homology was evidenced with any other mammalian 1'-(4-(1-(4-fluorophenyl))-1H-indol-3-yl)-1-butyl)spiro protein, outlining the unicity of the σ1 receptor as (isobenzofuran-1(3H),4'piperidine (Lu 28-179) (70), N- compared with any other known receptor. The gene, [2-(3,4-dichlorophenyl)ethyl]-N-methyl-2-(1-pyrrolidi- located on chromosome 9 in human and 2 in rodents, is nyl)ethylamine (BD1008) (54), or ibogaine (71). Several 7-kbp-long and contains four exons and 3 introns (61). functions have been proposed for σ2 sites: regulation of Exon 2 codes for the single transmembrane domain, motor functions, induction of dystonia after in situ identified at present, but two other hydrophobic regions administration in the red nucleus (72), regulation of ileal exist and one of them may putatively constitute a second function (73), blockade of tonic potassium channels transmembrane domain (62). The σ1-receptor sequence (74), potentiation of the neuronal response to NMDA in contains a 22 amino acid retention signal for the the CA3 region of the rat dorsal hippocampus (75), or endoplasmic reticulum at its N-terminal region and two activation of a novel p53- and caspase-independent short C-terminal hydrophobic amino acid sequences that apoptotic pathway, distinct from mechanisms used by were suggested to be involved in sterol binding (58). some DNA-damaging, antineoplastic agents and other Amino acid substitutions in the transmembrane domain, apoptotic stimuli (76). S99A, Y103F, and LL105,106AA, did not alter the expression levels of the protein but suppressed ligand 1.4. The σ1 receptor acts as an intracellular amplifier of binding activity (62), suggesting that these amino acids signal transduction system involved in the formation belong to the binding site pharmacophore located within and recomposition of membrane lipid microdomains the transmembrane domain. In addition, anionic amino Acute activation of the σ1 receptor results in a direct 2+ acid residues were identified, D126 and E172, that also modulation of intracellular calcium ([Ca ]i) mobiliza- appeared critical for ligand binding (63). The promoter tions. Selective σ1 agonists, (+)pentazocine and also region sequence of the σ1 receptor contains several PREGS, potentiate the bradykinin-induced increase in 2+ consensus sequences for the liver-specific transcription [Ca ]i, mediated by activation of inositol-1,4,5 tris- factors nuclear factor (NF)-1/L, activator protein (AP)- phophate (InsP3) receptors in neuroblastoma cells (77). 1, AP-2, IL-6RE, NF-GMa, NF-GMb, NF-κB, steroid A similar observation was also carried out in primary response element, GATA-1, Zeste, and for the xeno- culture of hippocampal neurons by both (+)SKF-10,047 biotic responsive factor called the arylhydrocarbon and (+)pentazocine (78). After depletion of intracellular receptor (61, 64). Ca2+ from endoplasmic reticulum (ER) stores, the 2+ The σ1 protein is widely distributed in peripheral depolarization-induced increase in [Ca ]i in the cells organs, including the heart, lung, kidney, liver, intes- could also be modulated by σ1 agonists. Both effects tines, and sexual and immune glands. In the brain, the σ1 were blocked by an antisense oligodeoxynucleotide receptor is expressed in neurons, ependymocytes, oligo- targeting the σ1 receptor (77). Therefore, activation of dendrocytes and Schwan cells (65 – 68). It is particularly the σ1 receptor resulted in a complex, bipolar modulation concentrated in specific areas throughout limbic systems of calcium homeostasis. At the ER level, the σ1-receptor and brainstem motor structures. The highest levels of σ1 activation facilitates the mobilization of InsP3 receptor- immunostaining can be observed in the granular layer of gated intracellular calcium pools and at the plasma the olfactory bulb, hypothalamic nuclei, and pyramidal membrane level, the σ1-receptor activation modulates layers of the hippocampus (25, 65). Among other areas extracellular calcium influx through voltage-dependent exhibiting intense to moderate σ1 immunostaining are calcium channels. A co-immunoprecipitation study the superficial cortical layers, different striatal areas further revealed that the σ1 receptor could regulate the including the caudate putamen and nucleus accumbens coupling of the InsP3 receptor with the cytoskeleton core and shell, the midbrain, the motor nuclei of the via an ankyrin B anchor protein, a cytoskeletal protein hindbrain, Purkinje cells in the cerebellum, and the originally attached to ER membranes (79). Activation of dorsal horn of the spinal cord. At the subcellular level, the σ1 receptor dissociated ankyrin B from InsP3 receptor the σ1 receptor was found to be mostly present within in NG-108 cells, and this dissociation correlated with neuronal perikarya and dendrites, where it is associated the efficacy of each ligand in potentiating the Ca2+ efflux with microsomal, plasmic, nuclear, or ER membranes induced by bradykinin. These results, coherent with the 98 FP Monnet and T Maurice

σ1-receptor subcellular localization (65, 80), showed 2. Neurosteroids and σ drugs apparently share the that the σ1 receptor might act as a sensor/modulator for same binding sites the neuronal intracellular Ca2+ mobilizations and con- 2+ secutively for extracellular Ca influx. 2.1. Steroids and neurosteroids bind both σ1 and σ2 After the observation using classical immuno- sites in the central nervous system precipitation and subcellular fractionation techniques The initial link between steroids and σ sites was that activation of the σ1 receptor resulted in its trans- suggested by Su et al. (14) given the discovery that in location from the ER, Hayashi and Su (81 – 83) used guinea-pig brain progesterone was the most active confocal fluorescence microscopy to examine the inhibitor of [3H](+)SKF-10,047 binding to σ receptors, protein dynamics in NG108 cells and primary oligo- with a Ki value of 268 nM in membrane extracts. It is dendrocytes over-expressing tagged σ1 receptors. They noteworthy that this Ki value is close to the physiological observed that endogenously expressed σ1 receptors concentration of the steroid during pregnancy. In localize on the ER reticular network and nuclear enve- addition, the variation of the Bmax for the σ1 binding site lope. They are seen particularly as highly clustered within the rodent brain from 100 (striatum, cingular unique globular structures associated with the ER cortex, globus pallidus) to 600 fmol/mg of tissue (80 – 82). These σ1-receptor-enriched globules contain (cranial nerve nuclei and cerebellar Purkinje cells) moderate amounts of free cholesterol and neutral lipids parallels that of progesterone in the rodent brain (for a (81). review, see ref. 72). The Bmax in the presence or absence Therefore, on the one hand, σ1 receptors translocate of progesterone remained between 700 – 770 fmol/mg from the ER lipid droplets to plasmalemma membranes of protein, indicating that all σ1 sites available could when stimulated by agonists. The translocation of σ1 bind the steroid. receptors, associated with the ankyrin B protein, 2+ consequently affects Ca mobilization at the ER (78, 2.2. Steroids and neurosteroids bind both σ1 and σ2 79). On the other hand, lipid droplets are formed by sites in peripheral tissues coalescence of neutral lipids within the ER membrane Simultaneously with their study in the central nervous bilayer and may, when reaching a critical size, bud off to system, Su et al. (14) showed that human peripheral form cytosolic lipid droplets, serving as a new transport blood lymphocytes and rat spleen, ovaries, testis, and pathway of lipids between the ER and Golgi apparatus or pituitary contain high densities of [3H](+)SKF-10,047- plasma membrane (83 – 85). Indeed, Hayashi and Su sensitive σ1 receptors. The rationale for this outstanding (81) observed that when functionally dominant negative study was based on the observations that gonadal and σ1 receptors, which can not target ER lipid droplets and adrenal steroids share molecular weights, non-peptidic can not translocate, are transfected into NG108 cells, a nature, and influence both humoral and cell-mediated large amount of neutral lipids and cholesterol is retained immunity with σ-active brain material obtained by in the ER, causing the pathological aggregation of the partial purification from male guinea pigs. In their ER and decreases of cholesterol in the Golgi and plasma binding study performed with crude membrane fractions membrane. Therefore, σ1 receptors on the ER may play a of rat spleen, Su et al. (14) showed that progesterone role in the compartmentalization of lipids into the ER was by far the most active inhibitor of [3H](+)SKF- lipid storage sites and in the export of lipids to 10,047 binding to σ receptors, with a Ki value of 376 nM peripheries of cells (81). Lipid rafts play a role in a in the spleen, a value being close to the physiological variety of cellular functions including vesicle transport, concentration of the steroid during pregnancy. The Bmax receptor clustering and internalization, and coupling of in the presence or absence of progesterone remained receptor with proteins involved signal transduction (86). within the same order of magnitude of that in brain Over-expression of functional σ1 receptors increased (≈ 600 – 800 fmol/mg of protein), indicating that all cholesterol contents and altered glycosphingolipid σ1 sites available could label the steroid. In addition, components in lipid rafts of NG108 or PC-12 cells (83, Scatchard analysis indicated that the steroid was acting 87), suggesting that up-regulation of σ1 receptors in a competitive manner. Testosterone, desoxycortico- potentiates lipid raft formation. Since glycosylated sterone, and PREGS were about equipotent, Ki values moieties of gangliosides have been proposed to play a within the micromolar range, while PREG and estro- role in regulating, for instance, the localization of genic hormones (estriol, estrone and estradiol) remained growth factor receptors in lipid rafts (86), chronic inactive. The notion that the potent ligands for the activation of σ1 receptors may present substantial progesterone receptor 11β-hydroxy-progesterone, consequences in cell viability and differentiation. promegestone, and RU-27987 failed to modify 3 [ H](+)SKF-10,047 binding to σ1 sites underlined the Neurosteroid Action at σ1 Receptor 99 specificity of the labeling for progesterone to the σ1 B-lymphocyte proliferation (88, 89). More recently, binding site. Casellas et al. (93) have shown that SR-31747 inhibits Simultaneously with the initial work of Su et al. (14), mitogen (ConA, PWM)-stimulated human T-lympho- linking steroids and σ-binding protein from brain and cyte proliferation with a similar efficacy to cyclosporin spleen extracts, Wolfe et al. (88) have demonstrated A; interfere with the production of proinflammatory such a relationship in the immune and endocrine cytokines IL-1, IL-6, and TNF-α; and also inhibit systems. Indeed, they have shown the existence of high experimental acute graft-versus-host disease by sup- affinity binding sites for both σ-receptor subtypes, that pressing the production of IFN-γ by Th1 CD4+ T-cells 3 is, labeled with [ H](+)pentazocine for the σ1 subtype or (94, 95). Conversely to most σ drugs, haloperidol 3 3 [ H]DTG or [ H]haloperidol for the σ2 subtype, in affected only the ConA-mediated response, suggesting lymphocytes and thymocytes (88, 89). In addition, that σ receptors might differentially regulate T- and B- hypophysectomy increased σ binding in the adrenal cell-mediated proliferation. The non-selectivity of both gland and testis (88), indicating that in immune and haloperidol and DTG for differentiating σ1 and σ2 sites endocrine tissues, the interplay between steroids and σ and the discrepant modulation of lymphocyte prolifera- sites may also exist. Inasmuch as kinetic and pharmaco- tion induced by haloperidol and DTG do not permit any logical characteristics of σ receptors were similar to those conclusion to be drawn concerning the subtype of σ obtained with steroids in the endocrine and immune receptor involved in both effects. Further assessing the systems, it was thus logical to propose that at least some role of σ ligands in the immune system, Ganapathy et al. 3 physiological effects of steroids in the endocrine and (95) have shown that [ H]haloperidol-sensitive σ1 + immune systems might be mediated through σ1 recep- receptor is expressed in the Jurkat cell, a human CD4 tors. Here also, as in the brain and spleen, the rank order T-cell line able to release IL-2 and IFN-γ in response to for steroids was progesterone > 5α-dihydrotestosterone stimuli. Although (+)dextromethorphan and (+)SKF- > testosterone > corticosterone > estradiol ≈ cholesterol. 10,047 exhibited Ki values in the low micromolar (0.2 – Further support for this notion was provided by the 0.5) range, Northern blot analyses and RT-PCR using + observations that ovarian maturing follicules, which are poly(A) RNA with σ1-receptor-specific primers, have under the control of progesterone, are the richest supported the involvement of the receptor. These region for σ sites where hypophysectomy depleted σ authors also showed that progesterone inhibited the 3 binding, indicating the strength of the link between the Jurkat cell [ H]haloperidol-sensitive σ1-binding in a σ1 receptor and endocrine tissues. The hypothalamic- competitive manner (with a Ki value of 93 nM and a Bmax pituitary-adrenal (HPA) axis may possibly also be under of 5.7 pmol/106 cells). The cDNA-induced pro- the control of distinct σ-receptor subtypes or distinct gesterone binding was however unaffected by R-5020, a intracellular regulations triggered by the same σ-receptor steroid competitor of progesterone against the classical subtype. Evidence for a positive control of the HPA nuclear progesterone receptor. In addition, the Kd values axis by σ drugs was provided by the observations calculated directly from binding of progesterone were 3 that (±)SKF-10,047, (±)pentazocine, (±)3-PPP, and 88 nM against [ H]haloperidol-sensitive σ1 sites and 3 (±)butaclamol dose-dependently and stereoselectively 69 nM against [ H](+)pentazocine-sensitive σ1 sites. stimulate adrenocorticotropic hormone release in vivo. This would signify that whatever the phase of the In addition, the σ-receptor-mediated modulation of the menstrual cycle, the T-cell-mediated impact of pro- HPA axis has been demonstrated to be mimicked by gesterone would likely be dependent at least on the σ1 steroids, since (+)SKF-10,047 and (+)pentazocine receptor. increase, as did progesterone (after estrogen priming), Among peripheral tissues, the liver is the richest tissue testosterone and desoxycorticosterone, whereas (+)3- in σ binding sites. The concentration of binding protein PPP and (−)butaclamol decrease prolactin release (90, being appreciatively 20 – 100 times higher than that 91). This was however not the case in vitro for corti- found in the brain, with Kd values within the low cotrophin-releasing-factor-evoked adrenocorticotropic nanomolar range and Bmax values ≈ 10 pmol/mg of hormone release from primary culture cells of the protein (69, 96 – 98). The similar order of binding anterior pituitary (90, 92). potency for σ ligands in inhibiting [3H](+)SKF-10,047 Splenic lymphocytes and B-enriched lymphocytes binding suggests that central and hepatic σ sites might be possess higher amounts of σ receptors than T-enriched identical (72, 99, 100). However, in liver microsomes, lymphocytes, mesenteric node lymphocytes, or thymo- the [3H]haloperidol-sensitive σ site might also resemble cytes (89). Both (+)pentazocine and DTG have been a variant of the σ1-receptor subtype that, conversely reported to suppress mitogen (ConA, PWM)-stimulated to the classical σ1 receptor, exhibits high affinity for T- and B-lymphocyte proliferation and LPS-stimulated 4-(4-chlorophenyl)-α-4-fluorophenyl)-4-hydroxy-1-pipe- 100 FP Monnet and T Maurice ridinebutanol (reduced haloperidol) and arylethylene was found to bind to purified membranes from the diamine-related compounds, such as (+)cis-N-methyl- placental brush border with an apparent Kd of 3.5 nM N-[2-(3,4-dichlorophenyl)ethyl]-2-(1-pyrrolidinyl)cyclo- and Bmax of 1.16 pmol/mg protein (104). Similar results hexylamine (BD737), but low affinity for DTG and the were obtained with choriocarcinoma cells, with Kd value benzomorphans (+)SKF-10,047 and (+)pentazocine, that within the nanomolar range and Bmax ≈ 2.5 pmol/mg is, Kd values between 100 and 1000 nM (98). protein, indicating that σ proteins might play important Linking further steroids and σ sites, McCann and Su roles in this endocrine tissue whose main hormonal (97) assessed whether [3H]progesterone binds to the σ function is to secrete the progesterone mandatory for the receptor from rat liver membranes. Unexpectedly, no maintenance of pregnancy. The σ nature of the binding receptor binding could be detected using acutely isolated was ascertained by the effectiveness of DTG, (+)3-PPP, cells. However, when a solubilized receptor preparation and dextromethorphan, but the dopaminergic antagonist from liver was used, in lieu of a membrane preparation, spiperone or the serotonergic antagonist ketanserin did [3H]progesterone was found to compete with the σ drugs. not affect [3H]haloperidol binding. Homogenates from Ross (101) confirmed this observation by showing that placental syncytiotrophoblasts as well as from chorio- the binding of [3H]SKF-10,047, [3H](+)3-PPP, and carcinoma cells exhibited similar binding values to those [3H]haloperidol was inhibited by progesterone with obtained from fresh border cells and choriocarcinoma apparent IC50 values of 442, 640, and 146 nM, respec- extracts, indicating that the cell distribution of σ sites, tively, in rat brain membranes, and 31, >10000, and that is, within microsomes, mitochondria, nucleus, and 146 nM, respectively in rat liver membranes. Additional plasma membrane, are similar in placental cells and evidence further supports the link between steroids and brain. It also suggested that the σ protein present in 3 σ sites in liver since [ H]progesterone binding, which placenta is most likely of the σ1 subtype. However, it is saturable with a Kd value of 31 nM and Bmax <6 seems that this σ1-receptor subtype might differ slightly pmol/mg protein (98), exhibits a similar rank order of from the brain σ1 site since phenytoin and cocaine bound affinity for several high affinity and selective σ drugs to with less affinity in placenta than in the nervous system. [3H]haloperidol binding. Interestingly, a similar profile Following the characterization of the placental σ protein, for an atypical σ-related receptor has also been shown in Ramamoorthy et al. (104) assessed whether steroids the brain, where progesterone was acting as a powerful might affect the σ binding in placental tissue. In their antagonist (102). In addition, in the liver, 5β-pregnane- study, progesterone and testosterone at a concentration 3,20-dione, 5α-dihydrotestosterone, testosterone, and of 1 µM were indeed capable of reducing [3H]halo- 5α-pregnane-3,20-dione exhibited similar affinity to peridol binding by 40% and 24%, respectively, with 3 [ H]haloperidol (i.e., IC50 values between 270 and IC50 values of 1.2 and 3.8 µM, respectively, whereas in 3 600 nM) and [ H]progesterone binding, contrarily to placental syncytiotrophoblasts, the Ki values were 310 progesterone that shows different affinities for both and 970 nM, respectively. 17β-Estradiol, estrone, deoxy- sites (i.e., IC50 values between 25 and 150 nM), whereas corticosterone, 5-α/β-pregnan-3α/β-ol–20-one deriva- 21-hydroxyprogesterone, 3α-hydroxy-5α-pregnan-20- tives, or DHEA were at least two orders of magnitude one, and 17β-estradiol had micromolar affinities. In less active than the two previous steroid hormones. The this latter study, typical σ drugs showed a tendency to same order of potency was found in choriocarcinoma attain the uppermost limit at the level of 60% – 70% cells against [3H]haloperidol, with either progesterone, 3 inhibition of the [ H]progesterone binding, whereas testosterone, or 17β-estradiol, which exhibited Kd values progesterone and other steroids increased inhibition in a of 70 nM, 120 nM, and 15 µM, respectively. Finally, concentration-dependent manner with 85% inhibition at Scatchard analysis of the binding indicated that a single their maximal concentration used (10 µM). A further binding site was present in both human placental step supporting the functional link between steroids and syncytiotrophoblast and choriocarcinoma and that σ receptors concerns the membrane-bound progesterone progesterone was a competitive inhibitor of [3H]halo- binding site purified from liver. Most interestingly, peridol binding in both tissue membranes. Falkenstein et al. (103) described that part of the Primary breast carcinomas have also been used to 3 immunostaining for the [ H]progesterone-sensitive assess the putative interrelationship between the σ1 re- binding protein was localized to the endoplasmic ceptor and human sterol isomerase. Simony-Lafontaine reticulum and Golgi apparatus, which is quite unusual et al. (105) have indeed found a close positive correla- for the classical steroid binding proteins. tion between σ1 protein expression and progesterone Human placental syncytiotrophoblast and choriocar- receptor status, but found an inverse one between the σ1 cinoma have also been considered with respect to the receptor and human sterol isomerase in 95 patients link between steroids and σ receptors. [3H]Haloperidol using immunochemical analysis. Neurosteroid Action at σ1 Receptor 101

2.3. Do neurosteroids bind atypical σ-receptor sub- sterol transporters to translocate cholesterol from the types? outer to the inner mitochondrial membrane, with a Ki 6 Using chromatographic procedures designed to purify value of 4 nM and a Bmax of 20 – 50 pmol/10 cells (108, 3 the σ proteins on a DAPE-containing column, Tsao and 109). The [ H](+)SKF-10,047-sensitive σ1 site, for 3 Su (106) proposed that [ H](+)SKF-10,047 might also which DHEA and progesterone exhibit a Ki values of 6 bind to another protein bearing certain similarities to 0.4–4µM and a Bmax of 10 – 15 pmol/10 cells (98), is the σ receptor sensitive to (±)SKF-10,047 but distinct that associated with limiting plasma membranes, mito- from the classical opiate-insensitive receptor since the chondria, synaptic vesicles, and endoplasmic reticulum labeling appeared to be sensitive to naloxone, the proto- membranes (65). typic opiate antagonist. Furthermore, its Kd value for In addition, DHEA and its sulphate derivative as well 3 [ H](+)SKF-10,047 was 165 nM with Bmax values of as pregnenolone sulphate have been shown to compete 8,064 pmol/mg. It was however underlined that pro- in a concentration-dependent manner with (+)pentazo- gesterone exhibited a 537 nM affinity for the CHAPS- cine for stimulating the [35S]GTPγS binding from solubilized liver preparation, versus [3H](+)SKF-10,047, synaptic membranes from mouse frontal cortex. The while the steroid exhibited a much weaker competitive notion that this action was mediated by the σ1 receptor activity against the affinity-purified protein labeled was provided by the following: i) the inhibition of 3 with [ H](+)SKF-10,047, as shown by the drop in Kd neurosteroid-induced effects by the prototypic σ1 value to 18 µM (106). Although the significance of antagonist NE-100 as well as by progesterone, and 3 this atypical [ H](+)SKF-10,047-binding protein has ii) the fact that inactivation of Gi/o proteins by ADP- remained elusive, these data further link steroids and σ ribosylation with pertussis toxin has prevented this drugs in the liver. action. Meyer et al. (107) have clearly established, with pig We previously mentioned that Ganapathy et al. (95) liver crude membrane preparations and solubilized demonstrated the presence of [3H]haloperidol-sensitive fractions, that haloperidol, (+)3-PPP, DTG, and rimca- σ receptor in the Jurkat cell, which exhibits Ki values in zole compete with progesterone for [3H]progesterone the low micromolar range (0.2 – 0.5 µM). Nevertheless, + binding, with a Ki value of 20, 290, 310, and 510 nM, Northern blot analysis and RT-PCR using poly(A) respectively, while (±)pentazocine, (+)SKF-10,047, RNA with σ1-receptor-specific primers, have supported and phenytoin exhibited low micromolar affinity. In the involvement of a splice variant of the cloned σ1 this study, the most active steroids were progesterone receptor characterized by the deletion of 31 amino acids. > corticosterone ≈ testosterone > cortisol >> 17β-estradiol. The only cytochrome P450 inhibitor active against the 3. Physiological aspects of σ receptor and neuro- 3 [ H]progesterone binding was SKF-525A, with a Ki steroid functions value of 140 nM, while methyrapone and cimetidine remained inactive even at concentrations >35 µM. The 3.1. Does σ binding protein interfere with (neuro)steroid unusual rank order of potency for these drugs does not fit synthesis? completely with the typical order proposed for either the Thereafter, Klein and Musacchio (110) also focused 3 σ1- or σ2-receptor subtype (53). This atypical binding on the putative link between steroids and [ H](+)3-PPP- protein might thus correspond to that found by Tsao and [3H]dextromethorphan-sensitive σ sites, using for and Su (106), who reported a [3H](+)SKF-10,047-sensi- this purpose guinea-pig brain membranes. Progesterone, tive σ protein with a molecular mass of 31 kDa, that is, deoxycorticosterone, testosterone, 17α-hydroxy-pro- similar to the cloned 28-kDa σ1 protein (58), which was gesterone, 5α-androstane-3,17-dione, 4-androstene-3,17- preferentially sensitive to dextrorotary benzomorphans dione, and dehydroisoandrosterone were the most active and naloxone but not to DTG or (+)3-PPP and exhibited compounds with an IC50 of 0.65, 3.5, 4, 5, 7, 9, and only high micromolar affinity for progesterone. It is thus 15 µM against [3H]dextromethorphan and 0.26, 0.68, tempting to speculate that in the liver but likely also in 0.45, 5.6, 0.7, 8.5, and 3.7 µM against [3H](+)3-PPP, other tissues, atypical σ proteins might exist with respectively. Conversely to Su and co-workers, Klein specific binding activity and specific physiological and Musacchio suggested that the steroids were labeling properties. a σ2-receptor subtype or that they bind to a σ1-receptor- Apart from the cell surface central benzodiazepine- like site with no affinity for dextromethorphan. This associated chloride channel and the nuclear gluco- notion also emerges from the displacement experiments corticoid receptor, DHEA binds to both the peripheral of [3H](+)3-PPP by steroids. Interestingly, pregnenolone benzodiazepine receptor and the σ1 protein. The benzo- and its 17-hydroxy-derivative, which were reported to be diazepine binding protein is the one that cooperates with totally inactive in the study of Su et al. (14), displaced 102 FP Monnet and T Maurice

3 [ H]dextromethorphan with IC50 values over 50 µM. correlated negatively with the human sterol isomerase However, they questioned whether σ1 binding proteins protein. Although the conclusion that the σ1 receptor might interfere with steroidogenesis. Having tested represents a sub-form of sterol isomerase still remains in both metabolites and precursors of their most active question, it is noteworthy that Hayashi and Su (81 – 83) steroids (progesterone, testosterone, and corticosterone), have very recently proposed that the σ1 protein might Klein and Musacchio (110) found that the former ones take up cholesterol from the intracellular medium to lacked affinity for the σ-related binding sites, thus carry it to the plasma membrane and even outside the concluding that σ-binding proteins were unlikely to be cell. Steroids have been known for years to activate the functionally associated with the steroidogenic enzyme HPA axis. responsible for C17 hydroxylation. This was nevertheless Although very abundant in liver, the functional reconsidered by Moebius et al. (111), who cloned the importance of the σ receptor in this tissue still remains σ1 protein from guinea-pig liver microsomes. They have unexplored. This aspect reinforces the “σ Enigma”, indeed suggested that the σ1 protein corresponded to a especially in the liver, and the exact nature and role variant or mutant of a sterol isomerase having a role in of the receptor have been questioned. A decade ago, postsqualene sterol biosynthesis. Focusing particularly taking into consideration that the sub-cellular distri- on the yeast ERG2 gene, which encodes the C8-C7 bution of the σ binding protein was predominant in the isomerase of ergosterol biosynthesis pathway that microsomal fraction, representing 48% of [3H](+)SKF- 3 catalyzes a shift of the C8(9) double bond in the B-ring 10,047 or [ H]dextromethorphan binding; most xeno- of sterols to the C7(8) position, whose equivalent in biotics metabolized in the liver bind to σ sites; and mammals is the emopamil binding protein that catalyzes several cytochrome P450 inhibitors compete with σ the reductases involved in the biosynthesis of cholesterol, binding, Musacchio and co-workers (110, 113) have Moebius et al. (111) claimed that the σ1 protein was a thus proposed that σ binding sites might represent sterol enzyme. This assertion was based on i) a similar microsomal enzymes related to the cytochrome P450 binding profile of the σ drugs haloperidol, ifenprodil, hemoproteins. They showed that SKF-525A, the and opipramol for the cloned σ1 protein, the fungal classical inhibitor of liver cytochrome P450 oxygenases; ERG2 protein and the emopamil binding protein; ii) a l-lobeline, GBR-12909, and sparteine, which exhibit 30% homology of the gene of the σ1 protein with those cytochrome P450 inhibitor activities; or cytochrome of the fungal ERG2 gene product and the emopamil P450 substrates, such as the β-adrenergic antagonist binding protein; iii) from identical transmembrane bufaralol, perhexiline, miconazole, or ajmalicine, 3 3 topologies of the mammalian σ1 protein, the emopamil displaced [ H]dextromethorphan, [ H](+)3-PPP, and binding protein, the fungal ERG2 protein, as well as [3H]DTG from the σ-binding proteins with very similar their similar aminoterminal membrane anchor in the affinities, ranging from 2.4 to 3.7 µM for SKF-525A. membrane of the endoplasmic reticulum. This anchor This observation, reproduced by Ross (101) with was indeed composed of two additional stretches of [3H](+)SKF-10,047 and [3H]haloperidol, supported the hydrophobic residues involved in the substrate binding notion of close interrelationships between cytochrome by proteins. Moreover, Moebius et al. (112) have shown P450 and σ-binding proteins. The binding was equally 3 that both the [ H]haloperidol-sensitive σ1 receptor and distributed between σ1- and σ2-receptor subtypes. Since the ERG2 binding site were affected by (+)benzomor- [3H](+)3-PPP and [3H]dextromethorphan bindings were phans, although (+)pentazocine elicited highly different not affected by either phenobarbital or phenytoin, known potencies for both bindings, with Ki values being 1.7 nM to bind to σ1 sites (72) and to recruit cytochrome P-450B for the σ1 site and 1 µM for the ERG2 protein, respec- isoforms, Klein and Musacchio (110) preferred to pro- tively. It is noteworthy that in mammalian cells, pose a preferential association of σ proteins with non-B conversely to yeast, sterol signaling is not restricted to isoforms of the enzymes. Interestingly, haloperidol, the the nucleus but also involves the endoplasmic reticulum prototypic σ1 antagonist (114, 115), was acting in human membrane. This provides further interest to the hypo- liver microsomes as a competitive inhibitor of 1'- thesis of Moebius and co-workers. A functional comple- hydroxybufaralol formation with a Ki of 1.2 µM, ment to this hypothesis was recently provided by suggesting that the drug was binding to the active site of Simony-Lafontaine et al. (105). Indeed, these authors the enzyme. However, this could not definitively allow have failed to show any correlation between the expres- the conclusion that σ proteins exhibit enzymatic activity. sion and the human sterol isomerase from human In this regard, Klein and Musacchio (110) also focused primary breast carcinomas. They did report however a on the purported link between steroids and cytochrome positive correlation between the σ1 status and both P450/σ sites. Their attention was stimulated following estrogen and progesterone receptor densities which both reports that most steroidogenic enzymes belong to the Neurosteroid Action at σ1 Receptor 103 cytochrome P450 superfamily of oxidases and favored 3.2. Neurosteroids and σ drugs share modulatory func- the notion that steroids were active on σ sites rather than tions at both pre- and post-synaptic levels on cytochrome P450-related binding proteins. Simulta- 3.2.1. Do neurosteroids and σ drugs exhibit similar neously, but conversely to Klein and Musacchio (110), effects on neuronal firing? Yamada et al. (98), working with liver microsomes, The initial statement that σ ligands potentiate the excluded the possibility that cytochrome P450 iso- NMDA neuronal response originates with Monnet et al. enzymes or other steroid/drug-metabolizing enzymes (119, 120) who showed that only drugs active on σ participate directly in the progesterone/σ binding since receptors exhibited this ability to affect the NMDA- they failed to show any correlation between the induced neuronal activation. Therefore, the paradigm of inhibitory action of the σ drugs as well as that of SKF- in vivo microiontophoretic application of NMDA and σ1 525A and GBR-12909 on the binding and oxidative drugs has been extensively used in the rat hippocampus metabolism of liver microsomes. Further support for to recognize the agonistic or antagonistic profile of doubting that σ sites may be solely linked with action of the ligands. Accordingly, σ drugs that aug- cytochrome P450 isoenzymes are their distinct cellular mented the excitatory action of NMDA were denoted as distribution since the former but not the latter are close agonist, whereas those devoid of effect on micro- to the plasma and nuclear membranes, as suggested by iontophoretic application of NMDA but capable of their sensitivity to 5'-nucleotidase, a plasma membrane preventing the action of σ agonists, for example, the marker (116), their coupling to G proteins (114), and as antipsychotic haloperidol, were denoted as σ antagonists shown by both confocal microscopy (80) and electronic (120). Subsequently, whole-cell patch clamp recordings microscopy (65). Altogether, this has ruined the asser- confirmed these modulatory actions of σ drugs showing tion that σ-binding proteins represent cytochrome P450 that NMDA-induced currents were modified in a similar isoforms or co-factors. Additional interest for the inter- manner as neuronal firing while non-NMDA responses relationships between liver and σ-binding proteins has were only slightly affected by huge doses of the σ drugs been suggested by the observation that microsomal anti- (121). estrogen-binding proteins bound arylethylene-diamine- Bergeron et al. (16) extended this observation using related compounds (117). In fact, microsomal anti- the in vivo microiontophoretic approach by showing estrogen-binding proteins share a high affinity for BD that DHEA potentiated the NMDA response in a pro- compounds with the emopamil-binding protein. How- gesterone-, testosterone-, and haloperidol-sensitive ever, the microsomal antiestrogen-binding protein, manner. However, in this paradigm, neither PREG nor unlike the emopamil protein, lacks affinity for 17β- PREGS were acting as σ agonists or antagonists (16, estradiol and other estrogens. 122). To further complement their study, Bergeron et al. Liver and brain mitochondria are known to contain (16) have also investigated the electrophysiological high levels of σ-binding sites as well as steroidogenic behavior of σ drugs in spayed rats (i.e., progesterone- enzymes (118). Although it has long been known that free rats). Following a 2 – 3-week ovariectomy, the cytochrome P450scc, which converts cholesterol into prototypic σ agonist DTG further potentiated the NMDA pregnenolone, is located on the outer membrane of the response in the hippocampal CA3 pyramidal layer. In mitochondria, it is only recently that Klouz et al. (118) addition, in pregnant rats and 3-week progesterone- 3 have demonstrated that a [ H](+)pentazocine-sensitive treated rats, the σ1 agonists were ten times less effective σ1 protein is present on the outer membranes of purified on the NMDA response than in control conditions. rat liver and brain mitochondria. The analogous distri- Furthermore, at day five post-partum, the neuronal bution of both the steroidogenic precursor enzyme and a response to NMDA following σ1 agonist administration σ-binding protein on the outer mitochondrial membrane was not only restored but again enhanced (123), indicat- would be compatible with their interaction, although ing that σ1 receptor was most likely tonically inhibited definitive data are still lacking. The observation that by endogenous progesterone (16, 123). Altogether, these progesterone modulates the mitochondrial binding of in vivo data supported the initial statement that pro- 3 [ H](+)pentazocine further strengthens the intimate links gesterone was acting as a σ1 antagonist (15). Farb et al. between steroids and the σ receptor. The physiological (124), using whole cell recordings from voltage- significance of such binding is emphasized by the clamped spinal cord neurons, have observed no enhanc- importance of mitochondria in the regulation of the ing effect of DHEAS (at concentrations up to 10 µM) on intracellular calcium homeostasis, the ATP production, the basal transmembrane potential, and the spontaneous as well as the initial phases of the steroid metabolism firing activity and no modulatory effect of DHEAS on from cholesterol. the neuronal response to NMDA, that is, a direct modulatory action. However, Meyer et al. (125) showed 104 FP Monnet and T Maurice that DHEAS, in the concentration range of 10 – 100 µM, tory effect of neurosteroids using the intracellular and weakly facilitated the activation of CA1 neurons in patch-clamp techniques. Bowlby et al. (133) were the hippocampal slices after stimulation of the Schaffer first to demonstrate in outside-out and cell-attached collaterals. This enhancement was related however to a configurations the facilitatory role of PREGS on the concomitant antagonistic activity of the neurosteroid on NMDA-mediated current from primary culture of rat GABA-mediated inhibitory postsynaptic potentials as hippocampal neurons. This action was concentration- well as an indirect augmentation of the glutamatergic dependent and rapidly reversible, supporting a co- excitatory postsynaptic potentials. allosteric modulation (47, 133). The PREGS-induced PREG and PREGS have no effect on spontaneous fir- increase corresponded to a facilitation of open pro- ing (124), but allosterically potentiate at micromolar bability attributed to an increase in both the frequency of concentrations NMDA-evoked currents in rat hippo- opening and mean open time of the NMDA receptor. campal neurons in culture (see above; and refs. 40 – 42, PREGS was the most active neurosteroid and its effect 124, 125). More recently, Partridge and Valenzuela appeared selective to the NMDA receptor (133), as soon (126) have shown that PREGS, acting on both NMDA as the NMDA- subunit NR2A was constitutively or and AMPA ionotropic receptors, enhances paired-pulse transiently expressed in the cell (48, 134). facilitation of EPSPs with an EC50 <1 µM, giving support to the notion that the neurosteroid acts pre- 3.2.2. Neurosteroids and σ drugs affect neuronal excit- synaptically to modulate neuronal excitability. ability induced by the NMDA receptor Since the initial studies assessing the intracellular Although the capacity of σ receptor ligands to modu- impact of σ drugs on the membrane potential, potassium late NMDA-mediated glutamatergic neuronal firing in conductance was considered as the prominent target. the mammalian central nervous system has been docu- Indeed, from rat cortical synaptosomes, C6 glioma cells mented for a decade, the exact nature of this interaction (74) or NCB-20 cells (127), DTG, (+)3-PPP, and remains still elusive, biochemical paradigms supporting haloperidol have been shown to facilitate hyperpolari- their close interrelationships. In particular, the selective zation with a reversal potential corresponding to that of σ ligands igmesine and DTG potentiated and inhibited, K+ or to block tonic outward K+ currents. In rat neuro- respectively, the NMDA response in a concentration- hypophysis, Wilke et al. (128) have confirmed this dependent manner, using the in vivo approach combin- notion by providing evidence that (+)pentazocine and ing microiontophoresis and extracellular recordings of (+)SKF-10,047 as well as DTG and haloperidol, but hippocampal pyramidal neurons (120). Haloperidol, not DHEAS or progesterone, elicited a marked inhibi- which also displays high affinity for σ1-binding sites, but tion of K+ currents. This latter study then suggested that not spiperone, another butyrophenone devoid of such σ drugs and steroids may act distinctly at the molecular affinity, prevented the effects of igmesine and DTG. The level. Pursuing the elucidation of the molecular target ability of progesterone to bind both [3H](+)SKF-10,047- 3 of σ drugs, Soriani et al. (129, 130) have characterized and [ H]haloperidol-sensitive σ1 sites under equilibrium + that σ1 ligands affected at least K conductance of IA, binding conditions on rat brain (see above) prompted us 2+ + Ca -activated K current, and IM types using perforated to assess whether neurosteroids might mimic σ drugs, patches of frog melanotropic cells. Using reconstituting that is, modulate NMDA-evoked [3H]noradrenaline 3 responses in transfected Xenopus oocytes, Aydar et al. ([ H]NE) overflow via action on σ1 receptors. Accord- (131) documented that σ1 receptor, activated by ingly, DHEAS, at nanomolar concentrations and in a (+)benzomorphans and inactivated by antisense con- concentration-dependent manner, potentiated the release structs, modulated voltage-gated K+ channels (Kv1.4 of [3H]NE induced by NMDA (15). The lowest effective and Kv1.5) depending on the presence or absence of σ concentration of DHEAS (30 nM) enhanced the ligands. According to their results, σ1 protein likely response by NMDA by 37%. Conversely, PREGS forms a stable complex with the K+ channels, acting as a inhibited the NMDA-induced release also in a concen- co-allosteric modulatory protein with variable function- tration-dependent manner. The lowest effective concen- alities, depending on the presence or absence of labeling. tration of PREGS (100 nM) induced a 60% inhibition Interestingly, Wang et al. (132), using Chinese hamster of the NMDA response. Haloperidol and 1-[2-(3,4- ovary cells, found that PREGS as well as 3α-hydroxy- dichloro-phenyl)ethyl]-4-methylpiperazine (BD1063) 5α-pregnan-20-one and 3α-hydroxy-5β-pregnan-20-one (100 nM), but not spiperone, completely prevented both increased current amplitude and decreased time con- the potentiating effect of DHEAS and the inhibitory stants for the voltage-gated K+ channels Kv1.1 and effect of PREGS. Conversely to sulphated steroids, Kv2.1. progesterone, DHEA, PREG, and allopregnanolone Several groups have investigated the neuromodula- did not affect NMDA-evoked [3H]NE release in the Neurosteroid Action at σ1 Receptor 105 concentration range of 10 nM to 1 µM. In addition, and N-type voltage-sensitive calcium channel (VSCC) progesterone concentration-dependently inhibited (in blockers, supporting a contribution of both neurosteroids the 10 nM to 1 µM range) both the potentiation and into the Ca2+ modulation during the neurotransmitter inhibition of NMDA-evoked release induced by DHEAS release. According to Hayashi et al. (77), the involve- and PREGS but also by DTG, respectively. At 100 nM, ment of the σ1 receptor could however be excluded from progesterone decreased the enhancing effect of DHEAS this action of neurosteroids since PREGS inhibits via the 2+ by 69% and abolished the reducing effect of PREGS. σ1-receptor-mediated KCl-induced increase in Ca , The pre-treatment with pertussis toxin, injected in the supporting the previous observation of Monnet et al. dorsal hippocampus 3 to 11 days prior to sacrifice, (15), who have demonstrated that NMDA-evoked totally abolished the effects of both DHEAS and PREGS release of [3H]NE was reduced in the presence of nano- on NMDA-evoked release of [3H]NE, indicating that the molar concentrations of PREGS. neurosteroids were most likely acting via the σ1-receptor subtype. From that initial observation, progesterone 3.2.3. Neurosteroids and σ drugs may have a similar has been thereafter proposed as a potential endogenous impact on neurotransmitter release σ1-receptor antagonist. Consistently, the σ1-receptor- Murray and Gillies (136) have shown that DHEAS, mediated antagonist-like activity of progesterone has but not PREGS, concentration-dependently (1 pM – since been supported by in vivo experiments showing 10 nM) stimulated endogenous dopamine release from that stereotaxically administered, progesterone, inactive in vitro perfused rat hypothalamic cell cultures. Further- on induced neuronal activation in the CA3 dorsal more, intracerebral microdialysis of PREGS (100 – 400 hippocampus, counteracted the DTG-induced facilita- pmol) enhanced in vivo meso-accumbens dopamine tion of the response of pyramidal neurons to NMDA release from dopaminergic terminals, and further (16). However, in the release experiments, the addition enhanced morphine-induced dopamine release (137) as of haloperidol (100 nM) partially reversed the inhibitory well as acetylcholine release from rat hippocampus effect of PREGS (in the concentration range of 0.1 to (138). Recently, Meyer et al. (139) reported that PREGS 1 µM) on NMDA-evoked [3H]NE overflow and induced (10–50µM), DHEAS (0.1 – 1 µM) as well as (+)penta- a robust potentiation of the NMDA response following zocine (5 – 50 µM) caused a robust and transient Gi/o protein inactivation. In such conditions, the enhancement of the frequency and amplitude of AMPA- occurrence of an inhibitory effect of PREGS most likely mediated miniature excitatory postsynaptic currents in allows the conclusion that there is an indirect σ1- primary mixed hippocampal cell cultures. This effect receptor-mediated modulation of the NMDA response involved σ-receptor activation per se since haloperidol since haloperidol and BD1063 blocked the PREGS- and BD1063 blocked the PREGS-induced increase in mediated response. Thus, PREGS would exert two miniature excitatory postsynaptic current frequency. opposite effects on NMDA-induced neuronal activation: Moreover, the σ1-receptor subtype was most likely the direct potentiation of NMDA-receptor interaction involved since pre-treatment with pertussis toxin pre- and an indirect σ1-receptor-mediated inhibition of the vented the PREGS response. They also showed with NMDA response, which seems to predominate under single pre-synaptic elements that the miniature excita- physiological conditions. In the light of the data of tory postsynaptic currents, corresponding to the most Farb et al. (124) on isolated neurons, the data obtained in elementary forms of synaptic transmission due to the NMDA-evoked [3H]NE overflow paradigm with postsynaptic responses to action potential-independent hippocampal slices point to an indirect effect of DHEAS spontaneous glutamate release, was sensitive to the 2+ on the NMDA response, that is, via the σ1-receptor modulation of [Ca ]i with BAPTA-AM, suggesting that subtype. Although, the affinity of DHEAS for σ1-binding neurosteroids as well as σ1 ligands may interfere with a sites is within the micromolar range, the effect of membrane-bound receptor that initiates a signal BD1063 and haloperidol, which interact with the σ1- transduction cascade involving a desensitization process binding sites but not with the NMDA receptor indicate that occurs simultaneously with the translocation of that DHEAS most likely acted on σ1 receptors at such σ1 proteins and activation of the Gi/o protein- nanomolar concentrations depending when the drug phospholipase C (PLC)-protein kinase C (PKC) cascade, application lasted 20 min. as Morin-Surun et al. (80) have shown. Monnet (135) has recently shown that the co-super- fusion of either DHEAS or PREGS, with a D2 dopamine 4. Behavioral effects of σ-receptor ligands and neuro- receptor antagonist, spiperone or sulpiride, produced a steroids facilitation of the KCl-evoked [3H]NE release from rat hippocampal slices, which can be suppressed by both L- The neuromodulatory effects of neurosteroids and 106 FP Monnet and T Maurice

σ drugs on excitatory and inhibitory neurostransmission learning-induced activation of physiological responses have been linked to several physiopathological behav- could be enhanced by an increase in neuroactive steroid ioral responses. In particular, exogenous, systemic as levels. PREGS also enhanced performances in a spatial well as central, administration of steroids influences recognition memory task after post-acquisition intra- response to stress, depression, anxiety, sleep, epilepsy, cerebroventricular infusion (138). Interestingly the and memory formation (6 – 13). Some drugs, including active dose, 12 nmol, induced a mild increase in extra- σ ligands and the selective serotonin reuptake inhibitor cellular acetylcholine (ACh) level measured by micro- fluoxetine, were also demonstrated to affect neuro- dialysis, whereas higher doses (up to 192 nmol) pro- transmission systems by modifying intracerebral neuro- voked robust increases in ACh levels without behavioral steroid levels (140, 141). Systemic administration of improvement. DHEAS, administered immediately after selective σ1-receptor agonists and antagonists has also training systemically or centrally or given in the notable effects on response to stress, depression, anxiety, drinking water during two weeks, facilitated memory drug addiction and memory formation. A crossed retention in a step-down passive avoidance test in pharmacology between both systems was clearly mice but did not improve acquisition (149). Moreover, evidenced, reinforcing the concept that an interaction when administered subcutaneously or intracerebroventri- with the σ1 receptor is a main component in the rapid cularly, the steroid improved the percentage of correct neuromodulatory effect of neurosteroids. We will focus responses in a delayed alternation task in a Y maze and more specifically on two behavioral aspects, learning improved the performances in a water-maze task as and memory and drug addiction, which are particularly compared to vehicle (oil)-treated rats (150). Therefore, representative of the recent achievements in the field. memory enhancing effects have been demonstrated for The reader is encouraged to refer to recent and both PREGS and DHEAS. exhaustive reviews covering similar aspects relevant to In most of the studies examining the anti-amnesic depression and mood disorders (142, 143). potentials of σ1-receptor agonists in either pharmaco- logical or pathological models of amnesia, a putative 4.1. Neurosteroids and σ drugs affect learning and pro-mnesic effect was examined in short-term or long- memory processes term memory tests, by injecting the drugs alone. 4.1.1. Pro-mnesic effects of neurosteroids and σ drugs Consistently, none of the compounds, (+)SKF-10,047, The involvement of neurosteroids in learning and (+)pentazocine, PRE-084, igmesine, SA4503 or DTG, memory processes has been demonstrated, first, by the tested in large dose range, facilitated learning in observation of pro-mnesic and anti-amnesic effects of control animals (for reviews, see refs. 6 – 8). Activation exogenously administered neuroactive steroids and, of the σ1 receptor failed to improve learning capacities second, by a parallel decrement between cognitive in control animals. Conversely, σ1 antagonists such as functions and neurosteroid levels during normal and α-(4-fluorophenyl)-4-(5-fluoro-2-pyrimidinyl)-1-pipera- pathological aging. Neurosteroids have been shown to zine butanol (BMY-14,802), haloperidol, N-[2-(3,4- affect memory performances by themselves in control dichlorophenyl)ethyl]-N,N',N'-trimethylethylenediamine animals and to alleviate the deficits in several pharmaco- (BD1047), or N,N-dipropyl-2-[4-methoxy-3-(2-phenyl- logical models of amnesia. For instance, DHEA, PREG, ethoxy)phenyl]ethylamine monohydrochloride (NE-100) and their sulphate esters enhanced, after central admin- failed to show any amnesic effect. Moreover, down- istration, memory retention in an active avoidance regulation of the σ1-receptor expression using in vivo learning task in mice after central or oral administration antisense strategies also failed to affect the learning (144 – 146). PREGS appeared as the most potent ability of mouse submitted to a passive avoidance test, compound and its long-duration effect suggested that confirming the lack of involvement of the receptor in PREG may serve as a precursor for the formation of normal memory functions (151, 152). The behavioral other different steroids, ensuring a near-optimal modula- phenotyping of σ1-receptor knockout animals (153) is tion of transcription of immediate-early genes required expected to confirm this observation. The lack of for the facilitation of the plastic changes in memory consequence of σ1-receptors blockade on learning pro- processes (145). PREGS also enhanced memory forma- cesses is presently understood as a consequence of its tion when administered after the first training session in neuromodulatory role. As usually observed in most of an alternation task in rats (147) or in an appetitive the physiological or pharmacological tests used to reinforced Go - No go visual discrimination task in mice evidence the σ1-receptor pharmacology, σ1 compounds (148). The rapid effect of the sulphated steroid clearly are devoid of effect alone, but exert some action only evokes its non-genomic interaction with both the when the transmission is perturbed. Therefore, the direct GABAA and/or NMDA receptors and suggests that the effects of neurosteroids on learning capacities may not Neurosteroid Action at σ1 Receptor 107 exclusively involve their interaction with the σ1 receptor, activation, indirectly potentiates acetylcholine release. but also imply their modulatory actions at GABAA and At the behavioral level, a crossed pharmacology NMDA receptors. Indeed, both receptors are expressed between neurosteroids and σ1-receptor ligands was on cholinergic neurons of the medial septum and reported for this amnesia model. DHEAS and PREGS diagonal band projecting into the hippocampal forma- attenuated the the scopolamine-induced learning impair- tion and glutamatergic and GABAergic responses regu- ments in mice submitted to a series of spatial and late the activity level of cholinergic inter-neurons. contextual tests. These effects could be blocked by the Cholinergic systems sustain the learning-induced plas- selective σ1-eceptor antagonist NE-100 or progesterone ticity and cholinergic deficiency, particularly during (155). Similar crossed pharmacology studies have been neurodegenerative diseases, is directly responsible for performed in amnesia models induced by blockade of learning deficits (24). Therefore, the cholinergic activity the NMDA-receptor activation. of neurosteroids, shown on physiological studies (138), is directly responsible for their anti-amnesic properties 4.1.3. Anti-amnesic effects in NMDA-receptor-depen- (12, 13). dent amnesia Impairment of the NMDA-receptor activation, by 4.1.2. Anti-amnesic effects in cholinergic models of either competitive or non-competitive antagonists, also amnesia results in marked learning deficits. Administration of The beneficial effects of neuroactive steroids were PREGS dose-dependently reduced the learning deficit tested in experimental models of amnesia. Several and motor impairment induced by pre-training admin- studies examined the effects of steroids on amnesia istration of the competitive antagonist 3-((±)2-carboxy- induced in rodents by the cholinergic muscarinic antago- piperazin-4-yl)-propyl-1 phosphonic acid (CPP) in a nist scopolamine. PREGS and DHEAS reversed its step-through passive avoidance task in the rat (165) amnesic effects in mice submitted to a foot-shock active or induced by (−)-2-amino-5-phosphonopentanoic acid avoidance test (149), step-through type passive avoid- (D-AP5) in an active avoidance test in mice (166). In ance test (154), Go - No go visual discrimination task addition, administration of PREGS in rats prevented (148), and spontaneous alternation and place learning in the cognition deficits induced by dizocilpine, a non- a water-maze (155). The beneficial effects induced by competitive NMDA-receptor antagonist (167). both sulphated steroids could be related to their acetyl- The σ1-receptor agonists were extensively studied in choline release properties resulting from negative this last amnesia model. (+)SKF-10,047, (+)pentazo- modulation of GABAA receptors and/or positive cine, igmesine, DTG, PRE-084, or SA4503 attenuated modulation of NMDA receptors. However, an involve- the dizocilpine-induced learning deficits in rats and ment of the σ1 receptor could not be definitively mice submitted to mnesic tasks involving spontaneous excluded. The learning impairment induced by scopol- alternation, passive avoidance, place learning in the amine, measured using spontaneous alternation, passive water-maze, three panel runway, or 8-radial arms maze avoidance, or water-maze procedures, could be attenuated (160, 168 – 171). Moreover, the involvement of the σ1 or reversed by σ1 agonists. This was indeed described for receptor in the anti-amnesic effect induced by the DTG, (+)3-PPP, (+)SKF-10,047, pentazocine, igmesine, steroids was indicated by the blockade of both DHEA or SA4503 (155 – 160). These effects could be fully sulphate and pregnenolone sulphate effects by the blocked by σ1 antagonists or an in vivo antisense strategy selective σ1-receptor antagonists BMY-14,802 and NE- (152). In addition, administration of SA4503 has also 100 (17, 172) and the observation that progesterone been reported to attenuate the learning impairment in behaved as a clear antagonist of the efficacy of the σ1 rats with cortical cholinergic dysfunction, such as agonists (17, 172). ibotenic acid injection of the basal forebrain, that is, Interestingly, the in vivo antisense strategy revealed lesioning cholinergic ascending pathways, in the passive some discrepancies regarding the anti-amnesic effects avoidance and water-maze tests (161). In parallel, mediated by neurosteroids. The antisense probe treat- physiological studies showed that (+)SKF-10,047, ment led to a complete blockade of the anti-amnesic pentazocine, DTG, igmesine or SA4503 potentiated effect mediated by DHEAS, confirming that the anti- [3H]acetylcholine release from hippocampal slices in amnesic effect of the steroid involves primarily an vitro or extracellular acetylcholine levels, measured in interaction with σ1 receptors. On the contrary, PREGS vivo by microdialysis, in the rat frontal cortex and still induced a potent anti-amnesic effect in antisense- hippocampus (162 – 164). Therefore, activation of the treated animals. This result must be analyzed in line σ1 receptor either directly affects cholinergic systems or, with other observations. Both steroids act as negative through a positive modulation of the NMDA-receptor allosteric modulators of the GABAA-receptor-mediated 108 FP Monnet and T Maurice responses (37, 173, 174). They also potentiate several mances in a delayed alternation task in the Y-maze and responses mediated through the NMDA receptor. a place learning task in the water-maze. Administration However, when PREGS acts through a specific extra- of PREGS directly into the hippocampus temporarily cellularly directed modulatory site located on the corrected the memory deficits of aged rats when NMDA-receptor complex, but distinct from either the administered immediately after the acquisition trial (24). spermine, glycine, phencyclidine, arachidonic acid, Cholinergic systems in the basal forebrain are known to Mg2+ or redox sites (47), DHEAS failed to affect the be altered during aging and degenerative changes in NMDA receptor through a direct interaction (46, 175). cholinergic nuclei are correlated with memory impair- However, through its interaction with the σ1 receptor, ment in aged rats. The central administration of PREGS DHEA/S have been reported to indirectly potentiate stimulated acetylcholine release in the adult rat hippo- several NMDA-mediated physiologic responses in vitro campus. Moreover, a correlation between PREGS levels and in vivo, and at the behavioral level (15, 16, 155, 160, and learning and memory performance was evidenced in 176). Morever, PREGS acts as an inverse σ1-receptor 24-month-old rats (24). In parallel, a single systemic agonist in vitro, when DHEA/S is a full agonist (see injection of DHEAS immediately after training im- above, and ref. 15). Parallel to this observation, it must proved the impairment of memory in middle-aged and be noted that both steroids differently affected the extent old mice submitted to a footshock active avoidance of excitotoxic insults resulting from over-activation of test, up to the levels observed in young mice (149). the NMDA receptor. PREGS was reported to facilitate Furthermore, the neurosteroid plays a physiological role the NMDA-receptor-mediated excitotoxic cell death in in preserving and/or enhancing cognitive abilities in old several in vitro models of neurodegeneration (177) animals, possibly via an interaction with the central and in mice exposed to a hypoxic insult in vivo (178). cholinergic systems. Such observations suggest that the The selective and efficient potentiation of the NMDA- neuromodulatory action of PREGS and/or DHEAS receptor activation induced by PREGS led to major reinforce tonically neurotransmitter systems. Further- consequences in the case of over-activation of this more, since their blood concentrations decrease with receptor. Indeed, the steroid potentiated the excitotoxic age, neurosteroids might be the rate-limiting endo- neurodegeneration and worsened the resulting behavioral genous substances for correct memory capacities (182). deficits. DHEAS, on the contrary, prevented or reduced The σ1-receptor expression and behavioral efficacy of the neurotoxic effects in primary hippocampal cultures selective σ1 drugs has been extensively studied in aged exposed to NMDA (175, 179, 180) and blocked both animals. Using radioligand binding, positon-emission the appearance of neurodegeneration and the resulting tomography (PET) scan imaging, and immunohisto- learning deficits in mice exposed to carbon monoxide chemical and molecular approaches, a remarkable (178). This neuroprotective effect seemed only partly preservation of the σ1-receptor expression levels has related to the interaction of the steroid with the σ1 been observed in the different brain structures, including receptor, since it was poorly sensitive to σ1 antagonists, limbic and cortical structures of SAM, aged mice or NE-100, rimcazole, or BD1063 (178, 179). However, aged monkeys (25, 26, 183, 184). In parallel, σ1-receptor these different results confirmed and brought physio- agonists attenuated after both acute and repeated pathological consequences for the differential pharmaco- treatments, the learning deficits in SAM, aged mice or logical profiles presented by both steroids. aged rats (25, 176, 185, 186). Decrease of central levels of neurosteroids was identified in these studies. 4.1.4. Anti-amnesic effects of neurosteroids and σ drugs Moreover, a preserved expression of σ1 receptors was during aging documented by RT-PCR and immunohistochemical Finally, the anti-amnesic effects of neuroactive techniques in numerous brain structures, whereas the steroids and σ1-receptor agonists have been tested behavioral efficacy of σ1 drugs was maintained and against learning deficits induced by normal or patho- inversely correlated with the decrease in progesterone logical aging. Robel et al. (181) observed a significant contents (25, 26). Although no direct link could be correlation between the PREGS levels in the hippo- formalized, these observations are in agreement with the campus of aged rats and memory performances in a endocrine manipulations studies detailed in paragraph water-maze and a two-trial recognition task; the animals 4.3. with better performances had greater levels of PREGS. Furthermore, Vallée et al. (24) reported significant and 4.2. Neurosteroids and σ drugs may influence abused selective decreases of hippocampal PREGS levels in drug intake aged rats as compared with young adults, and a positive The recent demonstration by Romieu et al. (187) that correlation between these levels and cognitive perfor- neuroactive steroids are able to modulate the acquisition Neurosteroid Action at σ1 Receptor 109 of cocaine-induced conditioned place preference (CPP), neurosteroids in mood and motivation. a behavioral procedure measuring the appetitive pro- The involvement of the σ1 receptor in several cocaine- perties of drugs, suggest their influence on acquisition induced behavioral effects has also been extensively of drug addiction. PREGS and DHEAS potentiated examined (for a review, see ref. 196). Selective σ1 cocaine-induced CPP acquisition, as do selective σ1- antagonists, like haloperidol, BMY-14,802 or rimcazole, receptor agonists, like igmesine or PRE-084, and the blocked the locomotor stimulant effect of cocaine after effect was blocked by the σ1 antagonist BD1047 (188). acute administration (197 – 199) and the development of Exogenous administration of progesterone, or its endo- behavioral sensitization after repeated cocaine injections genous accumulation by a finasteride treatment, blocked and withdrawal (200). Selective σ1 antagonists, like the cocaine rewarding effect (187). This may be BD1047 or NE100, or antisense oligodeoxynucleotide mediated through either GABAA- and NMDA-receptor- probes targeting the σ1 receptor also blocked acquisition modulation, since mesolimbic dopamine neurons in the or expression of cocaine-induced CPP in mice (188, nucleus accumbens are under the control of inhibitory 201). GABA neurons and excitatory glutamatergic neurons It is noteworthy that the modulating effect of neuro- originating from the frontal cortex or hippocampus. steroids was less observable on cocaine-induced loco- Indeed, diazepam, a benzodiazepine acting as a positive motor sensitization and toxicity. Some effects could GABAA-receptor modulator, reduced the release of however be observed on locomotor sensitization such as dopamine in the nucleus accumbens measured by in DHEA sensitizing to a delayed injection of cocaine and vivo microdialysis (189) and, in turn, attenuated progesterone attenuating the short-term sensitization to cocaine-induced CPP (190). Similarly, pre-synaptic cocaine (P. Romieu & T. Maurice, unpublished observa- NMDA receptors facilitate dopamine release and tions). However, σ1 antagonists potently prevented NMDA-receptor antagonists decreased cocaine-induced cocaine-induced locomotor stimulation and sensitization locomotor stimulation, sensitization, or CPP (191, 192). (197, 200) and contribution of the other targets of Since the respective pharmacological profiles of DHEA, neurosteroids, namely, the NMDA and GABAA recep- PREG, and progesterone on GABAA and NMDA tors, but also of genomic effects may also contribute to receptors are highly consistent with such actions, these an effective modulation of locomotor effects. Observa- interactions could be involved in their effects on tion of a clear pharmacological response on reward as cocaine-induced CPP. However, most of GABAA- and compared to other behavioral responses suggests that NMDA-receptor modulators showed CPP or place the σ1 receptor and the resulting interaction of neuro- aversion (CPA) by themselves. Indeed, CPP was active steroids appear to be particularly sensitive in observed with the direct GABAA agonist meprobamate, specific brain structures and pathways. Indeed, a 4-day the GABA metabolite γ-hydroxybutyric acid (GHB), or cocaine treatment resulted in high increase of the σ1- several benzodiazepine compounds. Moreover, benzo- receptor mRNA levels only in the nucleus accumbens, a diazepine antagonists or inverse agonists produced CPA key structure for drug reward (188), while self-admin- (for an exhaustive review, see ref. 194). It is noteworthy istered methamphetamine increased expression of σ1- that 3α-hydroxy-5α-pregnan-20-one (allopregnanolone), receptor mRNA in the hippocampus and decreased its devoid of affinity for the σ1 receptor but acting as a expression in the frontal cortex, in contrast to yoked highly efficient GABAA-receptor positive modulator administration (202). induced CPP after exogenous administration (193). A recent study extended the behavioral evidences by While NMDA-receptor competitive antagonists have showing that neurosteroids, administered at very low also been reported to induce CPP, more inconsistent doses also modulate cocaine’s effects. A modified results were reported with noncompetitive antagonists, passive avoidance procedure was used in mice to particularly dizocilpine and both CPP or CPA has been examine whether cocaine induces state-dependent learn- described (194). Curiously, neuroactive steroids failed to ing (StD) (203). Cocaine administration, at the low dose induce CPP or CPA alone, with the notable exception of of 0.1 mg/kg before training, produced a chemical state testosterone (195). in the brain that served as an endogenous cue. Animals Interestingly, a recent report by Barrot et al. (137) must be treated with the same dose of the drug before showed that intracerebroventricular injection of PREGS retention to ensure optimal retention. Among neuro- increased dopamine efflux measured in the rat nucleus active steroids, pregnanolone and allopregnanolone accumbens by in vivo microdialysis and potentiated the sustained StD by themselves. However, steroids also morphine-induced dopamine release. Consequently, the acting as σ1 agonists (e.g, DHEA and PREG) or as authors suggested that this potentiation of mesolimbic antagonist (i.e., progesterone) failed to induce StD but dopamine concentration may explain the involvement of modified the cocaine state. The σ1 agonist igmesine or 110 FP Monnet and T Maurice antagonist BD1047 also failed to induce StD but the function of σ1 receptors during pregnancy and post- modified the cocaine state. Furthermore, optimal reten- partum, that is, in case of drastic variations of steroid tion was noted in mice trained with (igmesine or levels, and suggested that the steroid tonus may impede DHEA) + cocaine and tested with a higher dose of the behavioral efficacy of σ1 agonists. cocaine or in mice trained with (BD1047 or pro- This question was also addressed through a series of gesterone) + cocaine and tested with vehicle. From these endocrine manipulations using adrenalectomized (AdX) studies, it is concluded that a low dose of cocaine and castrated (CX) mice, to deplete the peripheral induces a chemical state/memory trace that sustains biosyntheses in circulating steroids, and treated with StD. Modulation of σ1-receptor activation is not suffi- trilostane or finasteride, in order to deplete or accumu- cient to provoke StD, but it alters the cocaine state, in late progesterone in the brain. Accordingly, the in vivo 3 line with the pure neuromodulatory role of the σ1 [ H](+)SKF-10,047 binding levels to σ1 sites in the receptor. Neurosteroids have a unique impact on state- mouse forebrain and hippocampal formation appeared dependent versus state-independent learning, via significantly increased in AdX/CX mice and further GABAA- or σ1-receptor modulation, and can affect the increased after the trilostane treatment. Conversely, the cocaine-induced mnesic trace at low brain concentra- finasteride treatment led to a significant decrease of tions. binding as compared to AdX/CX animals. The attenuat- Finally, similar observations could be presented for ing effect of PRE-084 against dizocilpine-induced other drugs of abuse, since methamphetamine, mor- learning impairment was examined using spontaneous phine, and ethanol also involve σ1-receptor activation alternation behavior, step-down passive avoidance, and (202, 204 – 206). In particular, morphine-induced contextual latent learning in the elevated plus-maze in hyperlocomotion and CPP could be modulated by σ1- the different endocrine conditions. The learning deficits receptor ligands (T. Maurice, unpublished observations). induced by dizocilpine were not affected by the treat- Moreover, the naloxone-precipitated morphine with- ments. The anti-amnesic effect of PRE-084 was drawal symptoms could be blocked by administration facilitated in AdX/CX mice and even more after of the σ1 agonist igmesine (T. Maurice, unpublished trilostane treatment since several parameters for (PRE- observations) as well as DHEA in a NE-100-sensitive 084 + dizocilpine)-treated animals returned to control manner, thus confirming the involvement of the σ1- values. The PRE-084 effect was blocked after finasteride receptor subtype, as stated by Ren et al. (207). (208). Moreover, PREGS infusion in AdX/CX rats was previously reported to prevent the learning deficits 4.3. Cross-influences between neurosteroids and σ induced by dizocilpine and (−)3-[2-carboxypiperazin-4- receptor yl]-propyl-1-phosphonic acid (CPPene) (209). It was The interaction of endogenous neuro(active)steroids also demonstrated that infusion of PREGS markedly with the σ1 receptor has direct physiological conse- increased the brain contents of PREG, progesterone, quences. The variations of neuro(active)steroids levels, 5α-pregnane-3,20-dione, and allopregnanolone (209). for example, during pregnancy, stress, mood disorders, Interestingly, blockade of 5α-reductase activity led to or normal or pathological ageing, directly affects the the disappearance of the PREGS effect, leading the efficacy of σ1-receptor agonists in vivo. authors to suggest that an increase in allopregnanolone The direct demonstration of the physiological impor- level could mediate the observed effect. As an alterna- tance of the steroidal tonus was provided by Bergeron tive, 5α-reductase activity inhibition led to sustained and co-workers (123) who examined the effects of σ1 increase in progesterone levels and the steroid, acting as agonists in control female rats, at day 18 of pregnancy a σ1-receptor antagonist, which may attenuate the and day 5 post-partum, and in ovariectomized rats PREGS effect. following a 3-week treatment with a high dose of We also observed that an acute swim stress induced progesterone. In pregnant rats and following a 3-week an increase in the level of progesterone and a decrease of treatment with progesterone, tenfold higher doses of in vivo [3H](+)-SKF-10,047 binding in the hippocampus DTG, (+)-pentazocine, and DHEA were required to of control animals (21). These effects were enhanced elicit a potentiation of the NMDA response of hippo- in AdX/CX mice, but completely blocked following campal neurons comparable to that obtained in control treatment with trilostane. A significant inverse correla- females. Conversely, at day 5 post-partum and following tion was observed between progesterone increase and a 3-week treatment with progesterone and after a 5-day the inhibition of in vivo binding to σ1 sites after stress washout, the potentiation of the NMDA response among the different endocrine conditions. These induced by the σ agonist DTG was greater than in observations not only confirmed that neurosteroids control females (123). These data showed changes in modulate the efficacy of σ1-receptor agonists in the Neurosteroid Action at σ1 Receptor 111 response to stress and depression, but also showed that duration at 30 and 60 mg/kg only in β25–35-treated mice. endogenous neurosteroid levels directly and tonically In comparison, desipramine reduced the immobility regulate the in vivo bioavailability of σ1 receptors. duration similarly among the groups and fluoxetine These observations documented that progesterone appeared less potent in β25–35-treated animals (28). In could be a major endogenous effector of the σ1 receptor rats, igmesine and (+)-SKF-10,047 significantly reduced and stressed that physiopathological variations of this the stress-induced motor suppression at 30 and 6 mg/kg, neurohormone must be taken into consideration before respectively, in β40–1-amyloid-treated control rats. Active designing any pharmacological strategy targeting the σ1 doses were decreased to 10 and 3 mg/kg, respectively, receptor. Moreover, a different behavioral efficacy of in β1–40-amyloid-treated rats. The DHEAS effect was σ1-receptor ligands was observed among mouse strains, also facilitated, both in dose (10 vs 30 mg/kg) and that could be related to the endogenous tonus in neuro- intensity, in β1–40-amyloid-treated rats (29). These steroids (210). In particular, the C57BL/6 mouse results suggested that σ1-receptor agonists, due to their appears to present the lowest progesterone levels in enhanced efficacy in a nontransgenic animal model, may basal and stressful conditions, in line with a higher be particularly interesting drugs to alleviate not only efficacy of σ1 drugs and a highest vulnerability to Alzheimer’s disease-associated memory impairments cocaine’s appetitive effects (210). Extrapolated to but also depressive symptoms (8, 28, 29, 176, 185). humans, patient populations showing important levels in steroids acting negatively on σ1 receptors may impede 5. Conclusions the use of selective σ1 ligands. On the contrary, physiopathological conditions with The understanding of the action of steroids within disturbed steroidal tonus may be highly sensitive to σ1- the brain functions and their inter-relationships must be ligand treatments. This is of importance during ageing currently considered within the double framework of and in Alzheimer’s disease-like neurodegenerative patho- endocrine mechanisms, as responses elicited by hor- logies. With this regard, we examined the pharmaco- mones secreted by the gonads and adrenal glands, and logical efficacy of several σ1-receptor agonists in two regulation of neural functions by autocrine and/or nontransgenic rodent models of Alzheimer’s disease: paracrine actions of neurosteroids, that is, as responses mice injected acutely and intracerebroventricularly with elicited by steroids synthesized and metabolized within β25–35-amyloid peptide (28) and rats infused intra- the nervous system. Whatever their origins, several lines cerebroventricularly during 14 days with the β1–40- of evidence support the notion that apart from their amyloid protein delivered using an Alzet minipump genomic actions, neuroactive steroids, as well as neuro- (29). Neurosteroid levels were measured in several brain steroids, play a pivotal role in cognition and adaptive structures after β-amyloid injection in mice or infusion responses to neuronal damage. Moreover, the σ1 receptor in rats, in basal and stress conditions. The β25–35 mice constitutes one of the key targets in their trophic, exhibited decreased progesterone levels in the hippo- neuromodulatory and behavioral effects. Although campus. Progesterone levels, both under basal and partially unveiled using pharmacological, biochemical, stress-induced conditions, were also decreased in the and genetic approaches, experiments outlined above hippocampus and cortex of β1–40-amyloid-treated rats. have hence enlightened and documented the role of this The levels in PREG/S and DHEA/S appeared less orphan endoplasmic reticulum-bound receptor in the affected by the β-amyloid infusion (29). The behavioral regulation of physiologic responses. These include + 2+ efficacy of σ1 agonists is known to depend on the levels modulation of transmembrane K and Ca fluxes, of neuro(active)steroids synthesized by glial cells and NMDA-sensitive glutamatergic, catecholaminergic, and neurons, which are affected by the β-amyloid toxicity. cholinergic pathways, to the immediate and robust The antidepressant-like response of animals was recruitment of Ca2+-dependent intracellular signaling examined using either the forced swim test in mice or and protein (de)phosphorylation. These neuromodulatory the conditioned fear stress test in rats. The β25–35-peptide- effects accounted for the molecular mechanisms of injected mice developed memory deficits after 8 days in interaction between σ1 sites and neuro(active)steroids in contrast to the controls injected with scrambled β25–35 relationship with most of their physiological properties, peptide or vehicle solution. In the forced swim test, the detailed here in terms of cognitive and addictive out- β-amyloid treatment failed to affect the immobility comes. The merely simultaneous impact of σ1-receptor duration, but the antidepressant effect of the σ1 agonists activation at the level of both cytosol and nuclear and was facilitated in β25–35-treated mice: igmesine reduced plasmic membranes, due to its unique translocation immobility duration at 30 mg/kg versus 60 mg/kg in process, provides further insight into the identification control groups and PRE-084 decreased immobility of the commonality of action of σ1 drugs/neurosteroids 112 FP Monnet and T Maurice involved in these behavioral tests and further points out 5 Lambert LL, Belelli D, Hill-Venning C, Callachan H, Peters JA. the interest for the σ1 receptor in cognition and ageing. Neurosteroid modulation of native and recombinant GABAA This should facilitate the elaboration of efficient receptors, Cell Mol Neurobiol. 1996;16:155–174. strategies for preventing and treating cognitive and 6 Maurice T, Phan VL, Urani A, Kamei H, Noda Y, Nabeshima T. Neuroactive neurosteroids as endogenous effector for the sigma1 memory deficits during ageing, stimulating neural (σ1) receptor: pharmacological evidences and therapeutic oppor- regeneration and correcting behavioral disorders in tunities. Jpn J Pharmacol. 1999;81:125–155. response to stress. These above-mentioned findings 7 Maurice T, Urani A, Phan VL, Romieu P. The interaction support the notion that the recruitment of membrane- between neuroactive steroids and the sigma1 (σ1) receptor func- bound second messenger cascades (i.e., G-protein/PLC tion: behavioral consequences and therapeutic opportunities. /PKC pathway) via activation of a single transmembrane Brain Res Rev. 2001;37:116–132. 8 Maurice T. Neurosteroids and σ1 receptors, biochemical and domain intracellular receptor constitutes a novel mode behavioral relevance. Pharmacopsychiatry. 2004;350:S171–S182. of action for medicines. 9 Mellon SH. Neurosteroids: biochemistry, modes of action, and Finally, behavioral studies focusing on the impact of clinical relevance. J Clin Endocrinol Metab. 1994;78:1003– variations in the central level of neurosteroids in 1008. AdX/CX mice demonstrated that progesterone levels in 10 Rupprecht R, Hauser CA, Trapp T, Holsboer F. Neurosteroids: basal and stressful conditions determine the behavioral molecular mechanisms of action and psychopharmacological significance. J Steroid Biochem Mol Biol. 1996;56:163–168. efficacy of σ1-receptor drugs. Endocrine conditions must therefore be taken into consideration for the 11 Schumacher M, Akwa Y, Guennoun R, Robert F, Labombarda F, Desarnaud F, et al. Steroid synthesis and metabolism in the σ1 development of -receptor-based therapies. Extra- nervous system: trophic and protective effects. J Neurocytol. polated to humans, patient populations showing impor- 2000;29:307–326. tant levels in steroids acting negatively on σ1 receptors 12 Schumacher M, Weill-Engerer S, Liere P, Robert F, Franklin RJ, may impede the use of selective σ1 ligands. On the Garcia-Segura LM, et al. Steroid hormones and neurosteroids in contrary, physiopathological conditions with disturbed normal and pathological aging of the nervous system. Prog steroidal tonus may be particularly responding to σ1 Neurobiol. 2003;71:3–29. ligand treatments. 13 Vallee M, Mayo W, Koob GF, Le Moal M. Neurosteroids in learning and memory processes. Int Rev Neurobiol. 2001;46: Indeed, σ1-receptor-based therapies were already 273–320. proposed to be of physiological and therapeutic impor- 14 Su TP, London ED, Jaffe JH. Steroid binding at σ receptors tance in neurology and psychiatry since the σ1 receptor is suggests a link between endocrine, nervous, and immune abundant in the brain, where it regulates these functions. systems. Science. 1988;240:219–221. Although this σ1 receptor plays a crucial role in the 15 Monnet FP, Mahe V, Robel P, Baulieu EE. Neurosteroids, via σ 3 adaptative changes mandatory for memory processes or receptors, modulate the [ H]norepinephrine release evoked by addiction, it remains however to determine whether and N-methyl-D-aspartate in the rat hippocampus. Proc Natl Acad how these molecular mechanisms involved in learning Sci U S A. 1995;92:3774–3778. σ 16 Bergeron R, De Montigny C, Debonnel G. Potentiation of and recruited by 1 drugs operate synergically (most of neuronal response induced by dehydroepiandrosterone and its them usually proceeding in parallel) and whether the suppression by progesterone: effects mediated via sigma interplay between cytosolic and membrane-bound receptors. J Neurosci. 1996;16:1193–1202. molecular events occurring in response to σ1 drugs is 17 Maurice T, Junien JL, Privat A. Dehydroepiandrosterone sulfate consistent with the roles of cytoskeleton in mediating attenuates dizocilpine-induced learning impairment in mice via σ these changes. There is no doubt that the identification 1 receptors. Behav Brain Res. 1997;83:159–164. of the nature of the proteins synthesized in response to 18 Corpechot C, Robel P, Axelson M, Sjovall J, Baulieu EE. Characterization and measurement of dehydroepiandrosterone σ1 drugs will also need to be addressed soon. sulfate in rat brain. Proc Natl Acad Sci U S A. 1981;78:4704– 4707. References 19 Corpechot C, Synguelakis M, Talha S, Axelson M, Sjovall J, Vihko R, et al. Pregnenolone and its sulfate ester in the rat brain. 1 Selye H. The anesthetic effect of steroid hormones. Proc Soc Brain Res. 1983;270:119–125. Exp Biol Med. 1941;46:116–121. 20 Baulieu EE. Steroid hormones in the brain: several mechanisms? 2 Baulieu EE, Robel P. Neurosteroids: a new brain function? J In: Fuxe K, Gustafson JA, Wettenberg L, editors. Steroid Steroid Biochem Mol Biol. 1990;37:395–403. hormone regulation of the brain. Pergamon Press: Oxford. 3 Compagnone NA, Mellon SH. Neurosteroids: biosynthesis 1981. p. 3–14. and function of these novel neuromodulators. Front Neuro- 21 Urani A, Roman FJ, Phan VL, Su TP, Maurice T. The anti- endocrinol. 2000;21:1–56. depressant-like effect induced by sigma1-receptor agonists and 4 Gasior M, Carter RB, Witkin JM. Neuroactive steroids: potential neuroactive steroids in mice submitted to the forced swimming therapeutic use in neurological and psychiatric disorders. Trends test. J Pharmacol Exp Ther. 2001;298:1269–1279. Pharmacol Sci. 1999;20:107–112. 22 Orentreich N, Brind JL, Rizer RL, Vogelman JH. Age changes Neurosteroid Action at σ1 Receptor 113

and sex differences in serum dehydroepiandrosterone sulfate 40 Wu FS, Gibbs TT, Farb DH. Pregnenolone sulfate: a positive concentrations throughout adulthood. J Clin Endocrinol Metab. allosteric modulator at the N-methyl-D-aspartate receptor. Mol 1984;59:551–555. Pharmacol. 1991;40:333–336. 23 Orentreich N, Brind JL, Vogelman JH, Andres R, Baldwin H. 41 Irwin RP, Maragakis NJ, Rogawski MA, Purdy RH, Farb DH, Long-term longitudinal measurements of plasma dehydro- Paul SM. Pregnenolone sulfate augments NMDA receptor epiandrosterone sulfate in normal men. J Clin Endocrinol Metab. mediated increases in intracellular Ca2+ in cultured rat hippo- 1992;75:1002–1004. campal neurons. Neurosci Lett. 1992;141:30–34. 24 Vallée M, Mayo W, Darnaudéry M, Corpéchot C, Young J, 42 Irwin RP, Lin SZ, Rogawski MA, Purdy RH, Paul SM. Steroid Koehl M, et al. Neurosteroids: deficient cognitive performance potentiation and inhibition of N-methyl-D-aspartate receptor- in aged rats depends on low pregnenolone sulfate levels in the mediated intracellular Ca++ responses: structure-activity studies. hippocampus. Proc Natl Acad Sci U S A. 1997;94:14865–14870. J Pharmacol Exp Ther. 1994;271:677–682. 25 Phan VL, Urani A, Sandillon F, Privat A, Maurice T. Preserved 43 Maione S, Berrino L, Vitagliano S, Leyva J, Rossi F. sigma1 (σ1) receptor expression and behavioral efficacy in the Pregnenolone sulfate increases the convulsant potency of N- aged C57BL/6 mouse. Neurobiol Aging. 2003;24:865–881. methyl-D-aspartate in mice. Eur J Pharmacol. 1992;219:477– 26 Phan VL, Miyamoto Y, Nabeshima T, Maurice T. Age-related 479. expression of sigma1 receptors and antidepressant efficacy of a 44 Wakerley JB, Richardson CM. The neurosteroid pregnenolone selective agonist in the senescence-accelerated (SAM) mouse. sulphate enhances NMDA-induced phasic firing of vasopressin J Neurosci Res. 2005;79:561–572. neurones in the rat supraoptic nucleus. Neurosci Lett. 1997;226: 27 Brown RC, Cascio C, Papadopoulos V. Pathways of neuro- 123–126. steroid biosynthesis in cell lines from human brain: regulation 45 Compagnone NA, Mellon SH. Dehydroepiandrosterone: a of dehydroepiandrosterone formation by oxidative stress and potential signaling molecule for neocortical organization during β-amyloid peptide. J Neurochem. 2000;74:847–859. development. Proc Natl Acad Sci U S A. 1998;95:4678–83. 28 Urani A, Romieu P, Roman FJ, Maurice T. Enhanced anti- 46 Park-Chung M, Wu FS, Farb DH. 3α-Hydroxy-5β-pregnan-20- depressant effect of sigma1 agonists in β25–35-amyloid peptide- one sulphate: a negative modulator of the NMDA-induced treated mice. Behav Brain Res. 2002;134:239–247. current in cultured neurons. Mol Pharmacol. 1994;46:146–150. 29 Urani A, Romieu P, Roman FJ, Yamada K, Noda Y, Kamei H, 47 Park-Chung M, Wu FS, Purdy RH, Malayev AA, Gibbs TT, et al. Enhanced antidepressant efficacy of σ1 receptor agonists Farb DH. Distinct sites for inverse modulation of N-methyl-D- in rats after chronic intracerebroventricular infusion of β- aspartate receptors by sulfated steroids. Mol Pharmacol. amyloid-(1-40) protein. Eur J Pharmacol. 2004;486:151–161. 1997;52:1113–1123. 30 Weill-Engerer S, David JP, Sazdovitch V, Liere P, Eychenne B, 48 Yaghoubi N, Malayev A, Russek SJ, Gibbs TT, Farb DH. Pianos A, et al. Neurosteroid quantification in human brain Neurosteroid modulation of recombinant ionotropic glutamate regions: comparison between Alzheimer’s and nondemented receptors. Brain Res. 1998;803:153–160. patients. J Clin Endocrinol Metab. 2002;87:5138–5143. 49 Smith SS. Progesterone administration attenuates excitatory 31 Rupprecht R, Reul JM, Trapp T, van Steensel B, Wetzel C, amino acid responses of cerebellar Purkinje cells. Neuroscience. Damm K, et al. Progesterone receptor-mediated effects of 1991;42:309–320. neuroactive steroids. Neuron. 1993;11:523–530. 50 Martin WR, Eades CG, Thompson JA, Huppler RE, Gilbert 32 McEwen BS, Coirini H, Westlind-Danielsson A, Frankfurt M, PE. The effects of morphine- and nalorphine-like drugs in the Gould E, Schumacher M, et al. Steroid hormones as mediators of nondependent and morphine-dependent chronic spinal dog. J neural plasticity. J Steroid Biochem Mol Biol. 1991;39:223–232. Pharmacol Exp Ther. 1976;197:517–532. 33 Harrison NL, Simmonds MA. Modulation of the GABA receptor 51 Vaupel DB. Naltrexone fails to antagonize the sigma effects of complex by a steroid anaesthetic. Brain Res. 1984;323:287–292. PCP and SKF-10,047 in the dog. Eur J Pharmacol. 1983;92:269– 34 Belelli D, Lambert JJ. Neurosteroids: endogenous regulators of 274. the GABAA receptor. Nature Rev Neurosci. 2005;6:565–575. 52 Quirion R, Chicheportiche R, Contreras PC, Johnson KM, 35 Callachan H, Cottrell GA, Hather NY, Lambert JJ, Nooney JM, Lodge D, Tam SW, et al. Classification and nomenclature of Peters JA. Modulation of the GABAA receptor by progesterone phencyclidine and sigma receptor sites. Trends Neurosci. metabolites. Proc R Soc Lond B Biol Sci. 1987;231:359–369. 1987;10:444–446. 36 Shu HJ, Eisenman LN, Jinadasa D, Covey DF, Zorumski CF, 53 Quirion R, Bowen WD, Itzhak Y, Junien JL, Musacchio JM, Mennerick S. Slow actions of neuroactive steroids at GABAA Rothman RB, et al. A proposal for the classification of sigma receptors. J Neurosci. 2004;24:6667–6675. binding sites. Trends Pharmacol Sci. 1992;13:85–86. 37 Majewska MD, Demirgören S, Spivak CE, London ED. The 54 Hellewell SB, Bruce A, Feinstein G, Orringer J, Williams W, neurosteroid dehydroepiandrosterone sulfate is an allosteric Bowen WD. Rat liver and kidney contain high densities of σ1 antagonist of the GABAA receptor. Brain Res. 1990; 526:143– and σ2 receptors: characterization by ligand binding and photo- 146. affinity labeling. Eur J Pharmacol. 1994;268:9–18. 38 Demirgoren S, Majewska MD, Spivak CE, London ED. 55 Musacchio JM, Klein M, Santiago LJ. High affinity dextro- Receptor binding and electrophysiological effects of dehydro- methorphan binding sites in guinea pig brain: further character- epiandrosterone sulfate, an antagonist of the GABAA receptor. ization and allosteric interactions. J Pharmacol Exp Ther. Neuroscience. 1991;45:127–135. 1988;247:424–431. 39 Spivak CE. Desensitization and noncompetitive blockade of 56 Itzhak Y. Multiple affinity binding states of the sigma receptor: GABAA receptors in ventral midbrain neurons by a neurosteroid effect of GTP-binding protein-modifying agents. Mol Pharma- dehydroepiandrosterone sulfate. Synapse. 1994;16:113–122. col. 1989;36:512–517. 114 FP Monnet and T Maurice

57 Itzhak Y, Stein I. Sigma binding sites in the brain; an emerging 73 Kinney GG, Harris EW, Ray R, Hudzik TJ. σ2 Site-mediated concept for multiple sites and their relevance for psychiatric inhibition of electrically evoked guinea pig ileum longitudinal disorders. Life Sci. 1990;47:1073–1081. muscle/myenteric plexus contractions. Eur J Pharmacol. 1995; 58 Hanner M, Moebius FF, Flandorfer A, Knaus HG, Striessnig J, 294:547–553. Kempner E, et al. Purification, molecular cloning, and expres- 74 Jeanjean AP, Mestre M, Maloteaux JM, Laduron PM. Is the σ2 sion of the mammalian sigma1-binding site. Proc Natl Acad Sci receptor in rat brain related to the K+ channel of class III U S A. 1996;93:8072–8077. antiarrhythmic drugs? Eur J Pharmacol. 1993;241:111–116. 59 Kekuda R, Prasad PD, Fei YJ, Leibach FH, Ganapathy V. 75 Couture S, Debonnel G. Modulation of the neuronal response to Cloning and functional expression of the human type 1 sigma N-methyl-D-aspartate by selective σ2 ligands. Synapse. 1988;29: receptor (hSigmaR1). Biochem Biophys Res Commun. 1996; 62–71. 229:553–558. 76 Crawford KW, Bowen WD. Sigma-2 receptor agonists activate a 60 Pan YX, Mei J, Xu J, Wan BL, Zuckerman A, Pasternak GW. novel apoptotic pathway and potentiate antineoplastic drugs in Cloning and characterization of a mouse σ1 receptor. J Neuro- breast tumor cell lines. Cancer Res. 2002;62:313–322. 2+ chem. 1998;70:2279–2285. 77 Hayashi T, Maurice T, Su TP. Ca signaling via sigma1- 61 Prasad PD, Li HW, Fei YJ, Ganapathy ME, Fujita T, Plumley receptors: novel regulatory mechanism affecting intracellular LH, et al. Exon-intron structure, analysis of promoter region, Ca2+ concentration. J Pharmacol Exp Ther. 2000;293:788–798. and chromosomal localization of the human type 1 sigma 78 Monnet FP, Morin-Surun MP, Leger J, Combettes L. Protein receptor gene. J Neurochem. 1998;70:443–451. kinase C-dependent potentiation of intracellular calcium influx 62 Yamamoto H, Miura R, Yamamoto T, Shinohara K, Watanabe by sigma1 receptor agonists in rat hippocampal neurons. J M, Okuyama S, et al. Amino acid residues in the transmembrane Pharmacol Exp Ther. 2003;307:705–712. domain of the type 1 sigma receptor critical for ligand binding. 79 Hayashi T, Su TP. Regulating ankyrin dynamics: Roles of FEBS Lett. 1999;445:19–22. sigma-1 receptors. Proc Natl Acad Sci U S A. 2001;98:491–496. 63 Seth P, Ganapathy ME, Conway SJ, Bridges CD, Smith SB, 80 Morin-Surun MP, Collin T, Denavit-Saubié M, Baulieu EE, Casellas P, et al. Expression pattern of the type I sigma receptor Monnet FP. Intracellular σ1 receptor modulates phospholipase C in the brain and identity of critical amino acid residues in the and protein kinase C activation in the brain stem. Proc Natl ligand-binding domain of the receptor. Biochim Biophys Acta. Acad Sci U S A. 1999;96:8196–8199. 2001;1540:59–67. 81 Hayashi T, Su TP. σ-1 Receptors (σ1 binding sites) form raft-like 64 Seth P, Leibach FH, Ganapathy V. Cloning and structural analy- microdomains and target lipid droplets on the endoplasmic sis of the cDNA and the gene encoding the murine type 1 sigma reticulum: roles in endoplasmic reticulum lipid compart- receptor. Biochem Biophys Res Commun. 1997;241:535–540 mentalization and export. J Pharmacol Exp Ther. 2003;306:718– 65 Alonso G, Phan VL, Guillemain I, Saunier M, Legrand A, Anoal 725. M, et al. Immunocytochemical localization of the σ1 receptor in 82 Hayashi T, Su TP. Intracellular dynamics of sigma-1 receptors the adult rat central nervous system. Neuroscience. 2000;97: (σ1 binding sites) in NG108-15 cells. J Pharmacol Exp Ther. 155–170. 2003;306:726–733. 66 Palacios G, Muro A, Verdu E, Pumarola M, Vela JM. Immuno- 83 Hayashi T, Su TP. The potential role of sigma-1 receptors in histochemical localization of the σ1 receptor in Schwann cells of lipid transport and lipid raft reconstitution in the brain: implica- rat sciatic nerve. Brain Res. 2004;1007:65–70. tion for drug abuse. Life Sci. 2005;77:1612–1624. 67 Palacios G, Muroa A, Vela JM, Molina-Holgadoc E, Guitart X, 84 Murphy DJ, Vance J. Mechanisms of lipid-body formation. Ovalle S, et al. Immunohistochemical localization of the σ1- Trends Biochem Sci. 1999;24:109–115. receptor in oligodendrocytes in the rat central nervous system. 85 Ohashi M, Mizushima N, Kabeya Y, Yoshimori T. Localization Brain Res. 2003;961:92–99. of mammalian NAD(P)H steroid dehydrogenase-like protein on 68 Hayashi T, Su TP. Sigma-1 receptors at galactosylceramide- lipid droplets. J Biol Chem. 2003;278:36819–36829. enriched lipid microdomains regulate oligodendrocyte differen- 86 Simons K, Ikonen E. Functional rafts in cell membranes. Nature. tiation. Proc Natl Acad Sci U S A. 2004;101,14949–14954. 1997;387:569–572. 69 Hellewell SB, Bowen WD. A σ-like binding site in rat 87 Takebayashi M, Hayashi T, Su TP. Sigma-1 receptors potentiate pheochromocytoma (PC12) cells: Decreased affinity for epidermal growth factor signaling towards neuritogenesis in (+)benzomorphans and lower molecular weight suggest a PC12 cells: potential relation to lipid raft reconstitution. different σ receptor form from that of guinea pig brain. Brain Synapse. 2004;53:90–103. Res. 1990;527:244–253. 88 Wolfe SA, Culp SG, De Souza EB. Sigma-receptors in endo- 70 Moltzen EK, Perregaard J, Meier E. Sigma ligands with crine organs: identification, characterization, and autoradio- subnanomolar affinity and preference for the σ2 binding site. II. graphic localization in rat pituitary, adrenal, testis, and ovary. Spiro-joined benzofuran, isobenzofuran, and benzopyran Endocrinology. 1989;124:1160–1172. piperidines. J Med Chem. 1995;38:2009–2017. 89 Carr DJ, De Costa BR, Radesca L, Blalock JE. Functional 71 Bowen WD, Vilner BJ, Williams W, Bertha CM, Kuehne ME, assessment and partial characterization of [3H](+)-pentazocine Jacobson AE. Ibogaine and its congeners are σ2 receptor-selec- binding sites on cells of the immune system. J Neuroimmunol. tive ligands with moderate affinity. Eur J Pharmacol. 1995;279: 1991;35:153–166. R1–R3. 90 Iyengar S, Wood PL, Mick S, Dilworth V, Gray NM, Farah JM, 72 Walker JM, Bowen WD, Walker FO, Matsumoto RR, De Costa et al. (+)3-[3-Hydroxyphenyl-N-(1-propyl) piperidine] selec- B, Rice KC. Sigma receptors: biology and function. Pharmacol tively differentiates effects of sigma ligands on neurochemical Rev. 1990;42:355–402. pathways modulated by sigma receptors: Evidence for subtypes, Neurosteroid Action at σ1 Receptor 115

in vivo. Neuropharmacology. 1991;30:915–922. 106 Tsao LI, Su TP. Naloxone-sensitive, haloperidol-sensitive, 91 Su TP. σ Receptors. Putative links between nervous, endocrine, [3H](+)SKF-10047-binding protein partially purified from rat and immune systems. Eur J Biochem. 1991;200:633–642. liver and rat brain membranes: an opioid/sigma receptor? 92 Iyengar S, Mick S, Dilworth V, Michel J, Rao TS, Farah JM, Synapse. 1997;25:117–124. et al. Sigma receptors modulate the hypothalamic-pituitary- 107 Meyer C, Schmieding K, Falkenstein E, Wehling M. Are high- adrenal (HPA) axis centrally: evidence for a functional inter- affinity progesterone binding site(s) from porcine liver micro- action with NMDA receptors, in vivo. Neuropharmacology. somes members of the sigma receptor family? Eur J Pharmacol. 1990;29:299–303. 1998;347:293–299. 93 Casellas P, Bourrie B, Canat X, Carayon P, Buisson I, Paul R, 108 Besman MJ, Yanagibashi K, Lee TD, Kawamura M, Hall PF, et al. Immunopharmacological profile of SR 31747: in vitro and Shively JE. Identification of des-(Gly-Ile)-endozepine as an in vivo studies on humoral and cellular responses. J Neuro- effector of corticotropin-dependent adrenal steroidogenesis: immunol. 1994;52:193–203. stimulation of cholesterol delivery is mediated by the peripheral 94 Carayon P, Bouaboula M, Loubet JF, Bourrie B, Petitpretre G, benzodiazepine receptor. Proc Natl Acad Sci U S A. 1989;86: Le Fur G, et al. The sigma ligand SR 31747 prevents the 4897–4901. development of acute graft-versus-host disease in mice by 109 Sprengel R, Werner P, Seeburg PH, Mukhin AG, Santi MR, blocking IFN-gamma and GM-CSF mRNA expression. Int J Grayson DR, et al. Molecular cloning and expression of cDNA Immunopharmacol. 1995;17:753–761. encoding a peripheral-type benzodiazepine receptor. J Biol 95 Ganapathy ME, Prasad PD, Huang W, Seth P, Leibach FH, Chem. 1989;264:20415–20421. Ganapathy V. Molecular and ligand-binding characterization of 110 Klein M, Musacchio JM. Effects of cytochrome P-450 ligands the sigma-receptor in the Jurkat human T lymphocyte cell line. on the binding of [3H]dextromethorphan and sigma ligands to J Pharmacol Exp Ther. 1999;289:251–260. guinea-pig brain. In: Itzhak Y, editor. Sigma receptors. San 96 Samovilova NN, Nagornaya LV, Vinogradov VA. (+)- Diego: Academic Press; 1994. p. 243–262. [3H]SK&F 10,047 binding sites in rat liver. Eur J Pharmacol. 111 Moebius FF, Bermoser K, Reiter RJ, Hanner M, Glossmann H. 1988;147:259–264. Yeast sterol C8-C7 isomerase: identification and characteri- 97 McCann DJ, Su TP. Solubilization and characterization of zation of a high-affinity binding site for enzyme inhibitors. haloperidol-sensitive (+)-[3H]SKF-10,047 binding sites (sigma Biochemistry. 1996;35:16871–16878. sites) from rat liver membranes. J Pharmacol Exp Ther. 112 Moebius FF, Reiter RJ, Hanner M, Glossmann H. High affinity 1991;257:547–554. of σ1-binding sites for sterol isomerization inhibitors: Evidence 98 Yamada M, Nishigami T, Nakasho K, Nishimoto Y, Miyaji H. for a pharmacological relationship with the yeast sterol C8-C7 Relationship between sigma-like site and progesterone-binding isomerase. Br J Pharmacol. 1997:121:1–6. site of adult male rat liver microsomes. Hepatology. 1994;20: 113 Klein M, Musacchio JM. Computer-assisted analysis of 1271–1280. dextromethorphan and (+)-3-(-3-hydroxyphenyl)-N-(1-propyl) 99 Ross SB. Is the sigma opiate receptor a proadifen-sensitive piperidine binding sites in rat brain. Allosteric effects of subform of cytochrome P-450? Pharmacol Toxicol. 1990;67:93– ropizine. Life Sci. 1990;47:1625–1634. 94. 114 Monnet FP, Blier P, Debonnel G, de Montigny C. Modulation by 100 Klein M, Canoll PD, Musacchio JM. SKF 525-A and cyto- sigma ligands of N-methyl-D-aspartate-induced [3H]noradrena- chrome P-450 ligands inhibit with high affinity the binding of line release in the rat hippocampus: G-protein dependency. [3H]dextromethorphan and sigma ligands to guinea pig brain. Naunyn Schmiedebergs Arch Pharmacol. 1992;346:32–39. Life Sci. 1991;48:543–550. 115 Su TP. Evidence for sigma opioid receptor: binding of [3H]SKF- 101 Ross SB. Heterogeneous binding of σ radioligands in the rat 10,047 to etorphine-inaccessible sites in guinea-pig brain. J brain and liver: possible relationship to subforms of cytochrome Pharmacol Exp Ther. 1982;223:284–290. P-450. Pharmacol Toxicol. 1991;68:293–301. 116 McCann DJ, Su TP. Haloperidol-sensitive (+)[3H]SKF-10,047 102 Monnet FP, de Costa BR, Bowen WD. Differentiation of σ binding sites (σ sites) exhibit a unique distribution in rat brain ligand-activated receptor subtypes that modulate NMDA-evoked subcellular fractions. Eur J Pharmacol. 1990;188:211–218. [3H]-noradrenaline release in rat hippocampal slices. Br J 117 Kedjouar B, Daunes S, Vilner BJ, Bowen WD, Klaebe A, Faye Pharmacol. 1996;119:65–72. JC et al. Structural similitudes between cytotoxic antiestrogen- 103 Falkenstein E, Schmieding K, Lange A, Meyer C, Gerdes D, binding site (AEBS) ligands and cytotoxic sigma receptor Welsch U, et al. Localization of a putative progesterone ligands. Evidence for a relationship between cytotoxicity and membrane binding protein in porcine hepatocytes. Cell Mol affinity for AEBS or sigma-2 receptor but not for sigma-1 Biol. 1998;44:571–578. receptor. Biochem Pharmacol. 1999;58:1927–1939. 104 Ramamoorthy JD, Ramamoorthy S, Mahesh VB, Leibach FH, 118 Klouz A, Sapena R, Liu J, Maurice T, Tillement JP, Ganapathy V. Cocaine-sensitive sigma-receptor and its inter- Papadopoulos V et al. Evidence for σ1-like receptors in isolated action with steroid hormones in the human placental syncytiotro- rat liver mitochondrial membranes. Brit J Pharmacol. 2002;135: phoblast and in choriocarcinoma cells. Endocrinology. 1995; 1607–1615. 136:924–932. 119 Monnet FP, Debonnel G, Junien JL, De Montigny C. N-methyl- 105 Simony-Lafontaine J, Esslimani M, Bribes E, Gourgou S, D-aspartate-induced neuronal activation is selectively modulated Lequeux N, Lavail R, et al. Immunocytochemical assessment of by sigma receptors. Eur J Pharmacol. 1990;179:441–445. sigma-1 receptor and human sterol isomerase in breast cancer 120 Monnet FP, Debonnel G, De Montigny C. In vivo electrophysio- and their relationship with a series of prognostic factors. Br J logical evidence for a selective modulation of N-methyl-D- Cancer. 2000;82:1958–1966. aspartate-induced neuronal activation in rat CA3 dorsal hippo- 116 FP Monnet and T Maurice

campus by sigma ligands. J Pharmacol Exp Ther. 1992;261: 138 Darnaudery M, Koehl M, Piazza PV, Le Moal M, Mayo W. 123–130. Pregnenolone sulfate increases hippocampal acetylcholine 121 Ishihara K, Sasa M. The function of sigma receptors – Electro- release and spatial recognition. Brain Res. 2000;852:173–179. physiological approach. Folia Pharmacol Jpn (Nippon 139 Meyer DA, Carta M, Partridge LD, Covey DF, Valenzuela CF. Yakurigaku Zasshi). 1999;114:69–76. (text in Japanese with Neurosteroids enhance spontaneous glutamate release in hippo- English abstract) campal neurons. Possible role of metabotropic sigma1-like 122 Debonnel G, Bergeron R, de Montigny C. Potentiation by receptors. J Biol Chem. 2002;277:28725–28732. dehydroepiandrosterone of the neuronal response to N-methyl- 140 Uzunova V, Sheline Y, Davis JM, Rasmusson A, Uzunov DP, D-aspartate in the CA3 region of the rat dorsal hippocampus: an Costa E, et al. Increase in the cerebrospinal fluid content of effect mediated via sigma receptors. J Endocrinol. 1996;150: neurosteroids in patients with unipolar major depression who are S33–S42. receiving fluoxetine or fluvoxamine. Proc Natl Acad Sci U S A. 123 Bergeron R, de Montigny C, Debonnel G. Pregnancy reduces 1998;95:3239–3244. brain sigma receptor function. Brit J Pharmacol. 1999;127: 141 Kumar S, Fleming RL, Morrow AL. Ethanol regulation of γ- 1769–1776. aminobutyric acidA receptors: genomic and nongenomic mecha- 124 Farb DH, Gibbs TT, Wu FS, Gyenes M, Friedman L, Russek SJ. nisms. Pharmacol Ther. 2004;101:211–226. Steroid modulation of amino acid neurotransmitter receptors. 142 Van Broekhoven F, Verkes RJ. Neurosteroids in depression: a Adv Biochem Psychopharmacol. 1992;47:119–131. review. Psychopharmacology. 2003;165:97–110. 125 Meyer JH, Lee S, Wittenberg GF, Randall RD, Gruol DL. 143 Bermack JE, Debonnel G. The role of sigma receptors in depres- Neurosteroid regulation of inhibitory synaptic transmission in sion. J Pharmacol Sci. 2005;97:317–336. the rat hippocampus in vitro. Neuroscience. 1999;90:1177–1183. 144 Roberts E, Bologa L, Flood JF, Smith GE. Effects of dehydro- 126 Partridge LD, Valenzuela CF. Neurosteroid-induced enhance- epiandrosterone and its sulfate on brain tissue in culture and on ment of glutamate transmission in rat hippocampal slices. memory in mice. Brain Res. 1987;406:357–362. Neurosci Lett. 2001;301:103–106. 145 Flood JF, Morley JE, Roberts E. Memory-enhancing effects in 127 Morio Y, Tanimoto H, Yakushiji T, Morimoto Y. Characteriza- male mice of pregnenolone and steroids metabolically derived tion of the currents induced by sigma ligands in NCB20 from it. Proc Natl Acad Sci U S A. 1992;89:1567–1571. neuroblastoma cells. Brain Res. 1994;637:190–196. 146 Flood JF, Morley JE, Roberts E. Pregnenolone sulfate enhances 128 Wilke RA, Mehta RP, Lupardus PJ, Chen Y, Ruoho AE, Jackson post-training memory processes when injected in very low doses MB. Sigma receptor photolabeling and sigma receptor-mediated into limbic system structures: the amygdala is by far the most modulation of potassium channels in tumor cells. J Biol Chem. sensitive. Proc Natl Acad Sci U S A. 1995;92:10806–10810. 1999;274:18387–18392. 147 Mayo W, Dellu F, Robel P, Cherkaoui J, Le Moal M, Baulieu 129 Soriani O, Vaudry H, Mei YA, Roman F, Cazin L. Sigma EE, et al. Infusion of neurosteroids into the nucleus basalis ligands stimulate the electrical activity of frog pituitary magnocellularis affects cognitive processes in the rat. Brain melanotrope cells through a G-protein-dependent inhibition of Res. 1993;607:324–328. potassium conductances. J Pharmacol Exp Ther. 1998;286:163– 148 Meziane H, Mathis C, Paul SM, Ungerer A. The neurosteroid 171. pregnenolone sulfate reduces learning deficits induced by 130 Soriani O, Foll FL, Roman F, Monnet FP, Vaudry H, Cazin L. scopolamine and has promnestic effects in mice performing an A-Current down-modulated by sigma receptor in frog pituitary appetitive learning task. Psychopharmacology. 1996;126:323– melanotrope cells through a G protein-dependent pathway. 330. J Pharmacol Exp Ther. 1999;289:321–328. 149 Flood JF, Roberts E. Dehydroepiandrosterone sulfate improves 131 Aydar E, Palmer CP, Klyachko VA, Jackson MB. The sigma memory in aging mice. Brain Res. 1988;448:178–181. receptor as a ligand-regulated auxiliary potassium channel 150 Frye CA, Sturgis DJ. Neurosteroids affect spatial/reference, subunit. Neuron. 2002;34:399–410. working, and long-term memory of female rats. Neurobiol 132 Wang Q, Wang L, Wardwell-Swanson J. Modulation of cloned Learning Mem. 1995;64:83–96. human neuronal voltage-gated potassium channels (hKv1.1 and 151 Maurice T, Phan VL, Privat A. The anti-amnesic effects of hKv2.1) by neurosteroids. Pflugers Arch. 1998;437:49–55. sigma1 (σ1) receptor agonists confirmed by in vivo antisense 133 Bowlby MR. Pregnenolone sulfate potentiation of N-Methyl-D- strategy in the mouse. Brain Res. 2001;898:113–121. aspartate receptor channels in hippocampal neurons. Mol 152 Maurice T, Phan VL, Urani A, Guillemain I. Differential Pharmacol. 1993;43:813–819. involvement of the sigma1 (σ1) receptor in the anti-amnesic 134 Ceccon M, Rumbaugh G, Vicini S. Distinct effect of preg- effect of neuroactive steroids, as demonstrated using an in vivo nenolone sulfate on NMDA receptor subtypes. Neuropharmaco- antisense oligodeoxynucleotide strategy in the mous. Brit J logy. 2001;40:491–500. Pharmacol. 2001;134:1731–1741. 135 Monnet FP. Sigma receptors and intracellular signaling: Impact 153 Langa F, Codony X, Tovar V, Lavado A, Gimenez E, Cozar P on synaptic plasticity? XXIII CINP Meeting Abstr 2002;S.29.3. et al. Generation and phenotypic analysis of sigma receptor type 136 Murray HE, Gillies GE. Differential effects of neuroactive I (σ1) knockout mice. Eur J Neurosci. 2003;18:2188–2196. steroids on somatostatin and dopamine secretion from primary 154 Li PK, Rhodes ME, Jagannathan S, Johnson DA. Reversal of hypothalamic cell cultures. J Neuroendocrinol. 1997;9:287–295. scopolamine induced amnesia in rats by the steroid sulfatase 137 Barrot M, Vallee M, Gingras MA, Le Moal M, Mayo W, Piazza inhibitor estrone-3-O-sulfamate. Cognit Brain Res. 1995;2:251– PV. The neurosteroid pregnenolone sulphate increases dopamine 254. release and the dopaminergic response to morphine in the rat 155 Urani A, Privat A, Maurice T. The modulation by neurosteroids nucleus accumbens. Eur J Neurosci. 1999;11:3757–3760. of the scopolamine-induced learning impairment in mice Neurosteroid Action at σ1 Receptor 117

involves an interaction with sigma1 (σ1) receptors, Brain Res. 171 Zou LB, Yamada K, Nabeshima T. Sigma receptor ligands (+)- 1998;799:64–77. SKF10,047 and SA4503 improve dizocilpine-induced spatial 156 Earley B, Burke M, Leonard BE, Gouret CJ, Junien JL. memory deficits in rats. Eur J Pharmacol. 1998;355:1–10. Evidence for an anti-amnesic effect of JO 1784 in the rat: a 172 Zou LB, Yamada K, Sasa M, Nakata Y, Nabeshima T. Effects of potent and selective ligand for the sigma receptor. Brain Res sigma1 receptor agonist SA4503 and neuroactive steroids on 1991;546:282–286. performance in a radial arm maze task in rats. Neuropharmaco- 157 Matsuno K, Senda T, Matsunaga K, Mita S, Kaneto H. Similar logy. 2000;39:1617–1627. ameliorating effects of benzomorphans and 5-HT2 antagonists 173 Majewska MD, Schwartz RD. Pregnenolone-sulfate: an endo- on drug-induced impairment of passive avoidance response in genous antagonist of the γ-aminobutyric acid receptor complex mice: comparison with acetylcholinesterase inhibitors. Psycho- in brain? Brain Res. 1987;404:355–360. pharmacology. 1993;112:134–141. 174 Majewska MD, Mienville JM, Vicini S. Neurosteroid 158 Matsuno K, Senda T, Matsunaga K, Mita S. Ameliorating effects pregnenolone sulfate antagonizes electrophysiological responses of sigma receptor ligands on the impairment of passive to GABA in neurons. Neurosci Lett. 1988;90:279–284. avoidance tasks in mice: involvement in the central acetyl- 175 Kimonides VG, Khatibi NH, Svendsen CN, Sofroniew MV, cholinergic system. Eur J Pharmacol. 1994;261:43–51. Herbert J. Dehydroepiandrosterone (DHEA) and DHEA 159 Matsuno K, Senda T, Kobayashi T, Okamoto K, Nakata K, sulphate (DHEAS) protect hippocampal neurons against excita- Mita S. SA4503, a novel cognitive enhancer, with σ1 receptor tory amino acid-induced neurotoxicity. Proc Natl Acad Sci agonistic properties. Behav Brain Res. 1997;83:221–224. U S A. 1998;95:1852–1857. 160 Maurice T, Privat A. SA4503, a novel cognitive enhancer with 176 Maurice T, Su TP, Privat A. Sigma1 (σ1) receptor agonists and σ1 receptor agonist properties, facilitates NMDA receptor- neurosteroids attenuate β25–35-amyloid peptide-induced amnesia dependent learning in mice. Eur J Pharmacol. 1997;328:9–18. in mice through a common mechanism. Neuroscience. 161 Senda T, Matsuno K, Kobayashi T, Nakazawa M, Nakata K, 1998;83:413–428. Mita S. Ameliorative effect of SA4503, a novel cognitive 177 Weaver CE, Wu FS, Gibbs TT, Farb DH. Pregnenolone sulphate enhancer, on the basal forebrain lesion-induced impairment of exacerbates NMDA-induced death of hippocampal neurons. the spatial learning performance in rats. Pharmacol Biochem Brain Res. 1998;803:129–136. Behav. 1998;59:129–134. 178 Maurice T, Phan VL, Sandillon F, Urani A. Differential effect of 162 Junien JL, Roman FJ, Brunelle G, Pascaud X. JO-1784, a novel dehydroepiandrosterone and its steroid precursor pregnenolone σ ligand, potentiates [3H]acetylcholine release from rat hippo- against the behavioural deficits in CO-exposed mice. Eur J campal slices. Eur J Pharmacol. 1991;200:343–345. Pharmacol. 2000;390:145–155. 163 Matsuno K, Matsunaga K, Mita S. Increase of extracellular 179 Kurata K, Takebayashi M, Morinobu S, Yamawaki S. β- acetylcholine level in rat frontal cortex induced by (+)N-allyl- Estradiol, dehydroepiandrosterone, and dehydroepiandrosterone normetazocine as measured by brain microdialysis. Brain Res. sulfate protect against N-methyl-D-aspartate-induced neuro- 1992;575:315–319. toxicity in rat hippocampal neurons by different mechanisms. 164 Matsuno K, Matsunaga K, Senda T, Mita S. Increase in extra- J Pharmacol Exp Ther. 2004;311:237–245. cellular acetylcholine level by sigma ligands in rat frontal cortex. 180 Mao X, Barger SW. Neuroprotection by dehydroepiandro- J Pharmacol Exp Ther. 1993;265:851–859. sterone-sulphate: role of an NFκB-like factor. Neuroreport. 165 Mathis C, Paul SM, Crawley JN. The neurosteroid pregnenolone 1998;9:759–763. sulfate blocks NMDA antagonist-induced deficits in a passive 181 Robel P, Young J, Corpéchot C, Mayo W, Perché F, Haug M, avoidance memory task. Psychopharmacology. 1994;116:201– et al. Biosynthesis and assay of neurosteroids in rats and 206. mice: functional correlates. J Steroid Biochem Mol Biol. 166 Mathis C, Vogel E, Cagniard B, Criscuolo F, Ungerer A. The 1995;53:355–360. neurosteroid pregnenolone sulphate blocks deficits induced by a 182 Roberts E. Pregnenolone: From Selye to Alzheimer and a model competitive NMDA antagonist in active avoidance and lever- of the pregnenolone sulfate binding site on the GABAA receptor. press learning tasks in mice. Neuropharmacology. 1996;35: Biochem Pharmacol. 1995;49:1–16. 1057–1064. 183 Wallace DR, Mactutus CF, Booze RM. Sigma binding sites 167 Romeo E, Cheney DL, Zivkovic I, Costa E, Guidotti A. identified by [3H]DTG are elevated in aged Fisher-344 × Brown Mitochondrial diazepam-binding inhibitor receptor complex Norway (F1) rats. Synapse. 2000;35:311–313. agonists antagonize dizocilpine amnesia: putative role for 184 Kawamura K, Kimura Y, Tsukada H, Kobayashi T, Nishiyama allopregnanolone. J Pharmacol Exp Ther. 1994;270:89–96. S, Kakiuchi T, et al. An increase of sigma receptors in the aged 168 Maurice T, Hiramatsu M, Itoh J, Kameyama T, Hasegawa T, monkey brain. Neurobiol Aging. 2003;24:745–752. Nabeshima T. Behavioral evidence for a modulating role of 185 Maurice T. Improving Alzheimer’s disease-related cognitive σ ligands in memory processes. I. Attenuation of dizocilpine deficits with sigma1 (σ1) receptor agonists. Drug News Perspect. (MK-801)-induced amnesia. Brain Res. 1994;647:44–56. 2002;15:617–625. 169 Maurice T, Su TP, Parish DW, Nabeshima T, Privat A. PRE- 186 Tottori K, Nakai M, Uwahodo Y, Miwa T, Yamada S, Oshiro Y, 084, a sigma selective PCP derivative, attenuates MK-801- et al. Attenuation of scopolamine-induced and age-associated induced impairment of learning in mice. Pharmacol Biochem memory impairments by the sigma and 5-hydroxytryptamine-1A Behav. 1994;49:859–869. receptor agonist OPC-14523 (1-{3[4-(3-chlorophenyl)-1- 170 Ohno M, Watanabe S. Intrahippocampal administration of (+)- piperazinyl]propyl}-5-methoxy-3,4-dihydro-2[1H]-quinolinone SKF-10,047, a sigma ligand, reverses MK-801-induced impair- monomethanesulfonate). J Pharmacol Exp Ther. 2002;301:249– ment of working memory in rats. Brain Res. 1995;684:237–242. 257. 118 FP Monnet and T Maurice

187 Romieu P, Martin-Fardon R, Bowen WD, Maurice T. Sigma1 Exp Ther. 1993;266:473–482. (σ1) receptor-related neuroactive steroids modulate cocaine- 199 McCracken KA, Bowen WD, Matsumoto RR. Novel σ receptor induced reward. J Neurosci. 2003;23:3572–3576. ligands attenuate the locomotor stimulatory effects of cocaine. 188 Romieu P, Phan VL, Martin-Fardon R, Maurice T. The sigma1 Eur J Pharmacol. 1999;365:35–38. receptor involvement in cocaine-induced conditioned place 200 Ujike H, Kuroda S, Otsuki S. σ Receptor antagonists block the preference is consequent upon dopamine uptake blockade. development of sensitization to cocaine. Eur J Pharmacol. Neuropsychopharmacology. 2002;26:444–455. 1996;296:123–128. 189 Invernizzi R, Pozzi L, Samanin R. Release of dopamine is 201 Romieu P, Martin-Fardon R, Maurice T. Involvement of the σ1 reduced by diazepam more in the nucleus accumbens than in receptor in the cocaine-induced conditioned place preference. the caudate nucleus of conscious rats. Neuropharmacology. Neuroreport. 2000;11:2885–2888. 1991;30:575–578. 202 Stefanski R, Justinova Z, Hayashi T, Takebayashi M, Goldberg 190 Meririnne E, Kankaanpaa A, Lillsunde P, Seppala T. The effects SR, Su TP. Sigma1 receptor upregulation after chronic of diazepam and zolpidem on cocaine- and amphetamine- methamphetamine self-administration in rats: a study with yoked induced place preference. Pharmacol Biochem Behav. 1999;62: controls. Psychopharmacology. 2004;175:68–75. 159–164. 203 Romieu P, Lucas M, Maurice T. σ1 Receptor ligands and related 191 Karler R, Calder LD, Chaudhry IA, Turkanis SA. Blockade of neuroactive steroids interfere with the cocaine-induced memory “reverse tolerance” to cocaine and amphetamine by MK-801. state. Neuropsychopharmacology. 2005; in press. Life Sci. 1989;45:599–606. 204 Sabino V, Zhao Y, Steardo L, Cottone P, Koob GF, Bowen WD, 192 Cervo L, Samanin R. Effects of dopaminergic and glutamatergic et al. Sigma receptor modulation of voluntary ethanol intake in receptor antagonists on the acquisition and expression of cocaine rats. Program No. 489.8. 2004 Abstract Viewer/Itinerary conditioning place preference. Brain Res. 1995;673:242–250. Planner. Washington DC, Society for Neuroscience. 193 Finn DA, Phillips TJ, Okorn DM, Chester JA, Cunningham CL. 205 Maurice T, Casalino M, Lacroix M, Romieu P. Involvement of Rewarding effect of the neuroactive steroid 3α-hydroxy-5α- the sigma1 receptor in the motivational effects of ethanol in pregnan-20-one in mice. Pharmacol Biochem Behav. 1997;56: mice. Pharmacol Biochem Behav. 2003;74:869–876. 261–264. 206 Meunier J, Demeilliers B, Celerier A, Maurice T. Compensatory 194 Tzschentke TM. Measuring reward with the conditioned place effect by sigma1 (σ1) receptor stimulation during alcohol with- preference paradigm: a comprehensive review of drug effects, drawal in an object recognition task in mice. Behav Brain Res. recent progress and new issues. Prog Neurobiol. 1998;56:613– 2006;166:166–176. 672. 207 Ren X, Noda Y, Mamiya T, Nagai T, Nabeshima T. A neuro- 195 Packard MG, Schroeder JP, Alexander GM. Expression of active steroid, dehydroepiandrosterone sulfate, prevents the testosterone conditioned place preference is blocked by development of morphine dependence and tolerance via c-fos peripheral or intra-accumbens injection of α-flupenthixol. expression linked to the extracellular signal-regulated protein Horm Behav. 1998;34:39–47. kinase. Behav Brain Res. 2004;152:243–250. 196 Maurice T, Martin-Fardon R, Romieu P, Matsumoto RR. 208 Phan VL, Su TP, Privat A, Maurice T. Modulation of steroidal Selective sigma1 (σ1) receptor antagonists as a new promising levels by adrenalectomy/castration and inhibition of neuro- strategy to prevent cocaine-induced behaviors and toxicity. steroid synthesis enzymes affect σ1 receptor-mediated behaviour Neurosci Biobehav Res. 2002;26:499–527. in mice. Eur J Neurosci. 1999;11:2385–2396. 197 Menkel M, Terry M, Pontecorvo M, Katz JL, Witkin JM. 209 Cheney DL, Uzunov D, Guidotti A. Pregnenolone sulfate Selective σ ligands block stimulant effects of cocaine. Eur J antagonizes dizocilpine amnesia: role for allopregnanolone. Pharmacol. 1991;201:251–252. Neuroreport. 1995;6:1697–1700. 198 Witkin JM, Terry M, Menkel M, Hickey P, Pontecorvo M, 210 Phan VL, Urani A, Romieu P, Maurice T. Strain differences in σ1 Ferkany J, et al. Effects of the selective sigma receptor ligand receptor-mediated behaviours are related to neurosteroid levels. 6-[6-(4-hydroxypiperidinyl)hexyloxy]-3-methylflavone (NPC Eur J Neurosci. 2002;15:1523–1534. 16377), on behavioral and toxic effects of cocaine. J Pharmacol