The small GTPase RhoA regulates the expression and function of the sodium channel Nav1.5 in breast cancer cells.

Voltage-gated Na+ channels (VGSCs) are highly expressed in several types of carcinomas including breast, prostate and lung cancers as well as in mesothelioma and cervical cancers. Although the VGSCs activity is considered crucial for the potentiation of cancer cell migration and invasion, the mechanisms responsible for their functional expression and regulation in cancer cells remain unclear. In the present study, the role of the small GTPase RhoA in the regulation of expression and function of the Nav1.5 channel in the breast cancer cell lines MDA-MB 231 and MCF-7 was investigated. RhoA silencing significantly reduced both Nav1.5 channel expression and sodium current indicating that RhoA exerts a stimulatory effect on the synthesis of an active form of Nav1.5 channel in cancer cells. The inhibition of Nav1.5 expression dramatically reduced both cell invasion and proliferation. In addition, a decrease of RhoA protein levels induced by Nav1.5 silencing was observed. Altogether, these findings revealed: i) the key role of the small GTPase RhoA in upregulation of Nav1.5 channel expression and tumor aggressiveness, and ii) the existence of a positive feedback of Nav1.5 channels on RhoA protein levels.

Collagens, stromal cell-derived factor-1alpha and basic fibroblast growth factor increase cancer cell invasiveness in a hyaluronan hydrogel.

OBJECTIVE: Beyond to control of cell migration, differentiation and proliferation, the extracellular matrix (ECM) also contributes to invasiveness of human cancers. As the roles of hyaluronan (HA) and collagens in this process are still controversial, we have investigated their involvement in cancer pathogenesis. MATERIALS AND METHODS: With this aim in view, we developed a three-dimensional matrix, as reticulate HA hydrogel alone or coated with different collagens, in which cells could invade and grow. RESULTS: We show that cancer cells, which were non-invasive in a single HA hydrogel, acquired this capacity in the concomitant presence of type I or III collagens. Both types of ECM compound, HA and collagens, possess the capacity to stimulate production of metalloprotease-2, recognized otherwise as a factor for poor cancer prognosis. HA-provoked cellular invasiveness resulted from CD44-mediated increase in cytosolic [Ca2+] and its subsequent hydrolysis due to ADAM (a disintegrin and metalloprotease) proteolytic activity. Interestingly, this mechanism seemed to be absent in non-invasive cancer cell lines. CONCLUSION: Furthermore, using basic fibroblast growth factor and stromal cell-derived factor-1alpha, we also show that this three-dimensional reticulate matrix may be considered as a valuable model to study chemokinetic and chemotactic potentials of factors present in tumour stroma.

Endostatin variations in childhood acute lymphoblastic leukaemia--comparison with basic fibroblast growth factor and vascular endothelial growth factor.

Angiogenic factors such as basic fibroblast growth factor (bFGF) or vascular endothelial growth factor (VEGF) were previously studied in childhood acute lymphoblastic leukaemia (ALL) but little is known concerning the anti-angiogenic response in ALL. At diagnosis, the plasma levels of the anti-angiogenic factor endostatin were significantly higher in 33 children with ALL than in controls (median values 17.7 and 7.6 ng/ml, respectively, p=0.0192) but no relationship was observed with plasma bFGF or VEGF levels. The highest levels were observed in patients with an hyperdiploïd karyotype. Expression of mRNA for collagen XVIII/endostatin in lymphoblasts was detected in 19/24 cases but protein secretion was found only in 14/28 supernatants of cultured lymphoblasts. No direct relationship appeared between secretion of endostatin by lymphoblasts and plasma levels. In addition, endostatin levels remained elevated in remission, suggesting that endostatin could have a stromal origin as well. No prognostic value of plasma endostatin could be assessed. In conclusion, the present data indicate that an anti-angiogenic response is observed in some ALL children, but its physiopathological importance remains to be established.

Sequential activation of constitutive and inducible nitric oxide synthase (NOS) in rat cerebellar granule neurons by pseudomonas fluorescens and invasive behaviour of the bacteria.

Previous studies have shown that Pseudomonas fluorescens and its lipopolysaccharide (LPS) exert dose-related cytotoxic effects on neurons and glial cells. In the present work, we investigated the time course effect of P. fluorescens MF37 and its LPS on cultured rat cerebellar granule neurons. The kinetics of binding of P. fluorescens to cerebellar granule neurons is rapid and reaches a mean of 3 bacteria/cell after 5 h. As demonstrated by measurement of the concentration of nitrite in the culture medium, P. fluorescens induces a rapid stimulation (3 h) of the nitric oxide synthase (NOS) activity of the cells. In contrast, LPS extracted from P. fluorescens requires a long lag phase (24 h) before observation of an activation of NOS. Measurement of the membrane resting potential of granule neurons showed that within 3 h of incubation there was no difference of effect between the action of P. fluorescens and that of its endotoxin. Two complementary approaches allowed to demonstrate that P. fluorescens MF37 presents a rapid invasive behaviour suggesting a mobilisation of calcium in its early steps of action. The present study reveals that P. fluorescens induces the sequential activation of a constitutive calcium-dependent NOS and that of an inducible NOS activated by LPS. Our results also suggest that in P. fluorescens cytotoxicity and invasion are not mutually exclusive events.

Pseudomonas fluorescens as a potential pathogen: adherence to nerve cells.

In order to determine the infectious potential of the psychrotrophic bacterium Pseudomonas fluorescens, a species closely related to the opportunistic pathogen P. aeruginosa, we investigated the binding activity of this bacterium on primary cultures of rat neonate cortical neurons and glial cells, adrenal paraneurons and NG108-15 neuroblastoma cells. Incubated at concentrations of 10(6) and 10(8) CFU/mL, P. fluorescens MF37 exhibited a high binding activity on neurons in the same range as that of P. aeruginosa PAO1. A significant, but lower, adherence of P. fluorescens was also detected on glial cells and adrenal paraneurons. In contrast, when P. fluorescens MF37 or P. aeruginosa PAO1 were incubated with neuroblastoma cells, no binding was observed. In neurons, the association of P. fluorescens with the plasma membrane occurred both on neurites and cell body. Leakage of the cytoplasmic content was frequently noted. Studies performed using the fluorescent probe Hoechst 33258 revealed that in 10% of neurons, P. fluorescens induced the appearance of densely stained clusters of DNA that was typical of an early step of apoptosis. In glial cells exposed to P. fluorescens, marked changes in the morphology of the nucleus, including fragmentation into lobular structures and aggregation of DNA, were also reminiscent of the existence of a possible apoptotic mechanism. Taken together, these results reveal that P. fluorescens can bind to nerve cells and affect their physiology and, in agreement with recent clinical observations, suggest that P. fluorescens could behave as a pathogen.

In vivo regulation of vasomotricity by nitric oxide and prostanoids during gestation.

Pharmacological studies using the Doppler technique revealed that pregnancy decreases the systemic blood pressure and enhances uterine blood velocity in rats. The reactivity of the uterine artery to alpha-adrenoceptor and muscarinic receptor agonists was higher than that of systemic arteries. Sodium nitroprusside increased uterine arterial blood velocity slightly during gestation and markedly in non-pregnant rats. N(G)-L-Arginine methyl ester (L-NAME) decreased the uterine blood velocity mainly in gravid animals. The effect of diclofenac on uterine blood velocity was also more pronounced during pregnancy. The actions of sodium nitroprusside, L-NAME and diclofenac on systemic blood pressure were similar in pregnant and virgin rats. Altogether, these results indicate that pregnancy enhances nitric oxide (NO) and vasodilatory prostanoid production in the uterine vascular muscle which becomes less sensitive to exogenous NO. The uterine vasodilated status appears to be determined by conjugated actions of endothelial NO and vasodilator prostanoids of which the synthesis and the effects are weakly modified in systemic arteries during gestation.

Developmental changes of (3)H-labelled mu-opioid receptors in brainstems of intra-uterine growth-restricted rats.

The opioid mu-system is involved in brainstem-mediated respiratory control. Infants with intra-uterine growth restriction (IUGR) have more respiratory disorders in the early postnatal period. Using [(3)H]DAGO, a mu-selective ligand, and a computer-based image analysis of autoradiography, we compared the ontogeny and distribution of mu-opioid binding sites in the brainstem of IUGR and control rats in utero (E21), at birth (P0) and on postnatal days 1 (P1), P7, P10, P14 and P21. The ontogeny pattern was found to be similar in both groups. The density of the binding sites, which was low in E21, increased at P0, slightly declined at P1 and remained relatively constant thereafter. The distribution of DAGO-binding sites, also similar in both groups, was heterogeneous and was much denser in the dorsal areas of medulla and pons. In particular, binding sites were highly concentrated in nuclei involved in the cardio-respiratory function. However, DAGO-binding density was higher at all ages (except for P0 and P1) in IUGR than in control rats. Taken together, these results give at least a partial explanation for the effects of IUGR which lowers the Apgar score at birth and raises the incidence of respiratory disorders in infants.

Developmental changes in cardio-respiratory responses to hypoxia and hypercapnia in anesthetized low-birth-weight rats.

The present study compared the developmental changes in the cardio-respiratory responses to hypoxia and hypercapnia between full-term low-birth-weight (LBW) and control rats during the postnatal period. The heart rate (HR), respiratory frequency (fR) and amplitude (aR) were measured during hypoxia (10% O(2) for 10 min) and hypercapnia (5% CO(2) for 10 min) in rats aged 7, 14 and 21 days, anesthetized with urethane. During hypoxia, HR was not significantly modified in the younger rats of both groups. In the older rats, aged 14 and 21 days, HR was markedly diminished, with a more pronounced decrease in LBW rats. The HR recovery was never observed in the older LBW rats. The fR and aR showed an age-related increase in both groups: a biphasic fR pattern observed on day 7 was replaced by a sustained increase on days 14 and 21. In contrast to controls, LBW rats never displayed a fR recovery during reoxygenation. In controls, aR shifted from a biphasic pattern in the younger rats to a sustained increase in the older ones. The LBW rats only displayed a decrease of aR in the younger, while in the older ones, a transient and slight increase preceded this decrease. During hypercapnia, the only significant difference detected between these two groups was that aR increased in LBW rats to a greater extent than in controls on days 14 and 21. Altogether, our results revealed a markedly attenuated cardio-respiratory response to hypoxia in LBW rats, but no such effect in response to hypercapnia.

Contribution of changes in the chloride driving force to the fading of I(GABA) in frog melanotrophs.

Chloride redistribution during type A gamma-aminobutyric acid (GABA(A)) currents (I(GABA)) has been investigated in cultured frog pituitary melanotrophs with imposed intracellular chloride concentration ([Cl(-)](i)) in the whole cell configuration or with unaltered [Cl(-)](i) using the gramicidin-perforated patch approach. Prolonged GABA exposures elicited reproducible decaying currents. The decay of I(GABA) was associated with both a transient fall of conductance (g(GABA)) and shift of current reversal potential (E(GABA)). The shift of E(GABA) appeared to be time and driving force dependent. In the gramicidin-perforated patch configuration, repeated GABA exposures induced currents that gradually vanished. The fading of I(GABA) was due to persistent shifts of E(GABA) as a result of g(GABA) recovering from one GABA application to another. In cells alternatively clamped at potentials closely flanking resting potential and submitted to a train of brief GABA pulses, a reversal of I(GABA) was observed after 150 s recording. It is demonstrated that, in intact frog melanotrophs, shifts of E(GABA) combine with genuine receptor desensitization to depress I(GABA). These findings strongly suggest that shifts of E(GABA) may act as a negative feedback, reducing the bioelectrical and secretory responses induced by an intense release of GABA in vivo.

Regulation of GABAA receptor by protein tyrosine kinases in frog pituitary melanotrophs.

The effects of protein tyrosine kinase (PTK) and PTK inhibitors on the GABAA receptor function were studied in cultured frog pituitary melanotrophs by using the patch-clamp technique. Extracellular application of the PTK inhibitors genistein (10-9 to 10-5 M) or lavendustin A (10-12 to 10-7 M) provoked a bell-shaped potentiation of the whole-cell current induced by GABA (3x10-6 M). In contrast, at high concentrations, genistein (10-4 M) and lavendustin A (10-5 M) reversibly reduced the GABA-evoked current. Daidzein and lavendustin B, the inactive analogs of genistein and lavendustin A, respectively, did not modify the current induced by GABA. In the inside-out configuration, bath application of the recombinant PTK pp60c-src (75 U/ml) inhibited the GABA-activated chloride current, and the inhibitory effect of pp60c-src was prevented by genistein (10-7 M). Immunoblotting revealed that genistein, at doses of 10-7 M or 10-4 M, markedly inhibited tyrosine phosphorylation of the beta2/beta3 subunits of the GABAA receptor. Extracellular application of the PKA activator Bt2cAMP (10-3 M), the PKA/PKC inhibitor H7 (10-5 M) and the Cam KII inhibitor W7 (10-5 M) reversibly diminished the whole-cell GABA-induced current. Internal application of H7 and W7 (10-4 M) did not modify the dose-dependent effects of genistein. Internal application of sodium orthovanadate (10-4 M), a protein tyrosine phosphatase inhibitor, decreased the GABA-evoked current and markedly reduced the potentiating effect of genistein. The present study provides the first evidence that, in frog pituitary melanotrophs, the GABAA receptor is phosphorylated at least on its beta2/beta3 subunits by an endogenous PTK. Our data also demonstrate that tyrosine phosphorylation exerts an inhibitory effect on GABAA receptor function.

The sigma-ligand (+)-pentazocine depresses M current and enhances calcium conductances in frog melanotrophs.

Gramicidin-perforated patch clamp experiments and microfluorimetric measurements were performed to study the ionic mechanisms involved in the sigma-receptor-mediated stimulation of frog (Rana ridibunda) pituitary melanotrophs. The sigma-ligand (+)-pentazocine (50 microM) depressed a sustained outward K(+) current. The kinetic properties of this K(+) component, investigated by analyzing tail currents, were reminiscent of those of the M current (I(M)), with an activation threshold close to -60 mV, a -21-mV half-maximal activation potential, and two-component exponential deactivation kinetics at -90 mV. (+)-Pentazocine (20 microM) produced a 12-mV rightward shift of the activation curve and accelerated the deactivation rate of the tail current. It is also demonstrated that (+)-pentazocine (20 microM) reversibly increased both voltage-dependent calcium conductances and internal calcium level. Altogether, these results suggest that the sigma-receptor-induced modulation of I(M) and calcium currents likely underlies the increase of intracellular [Ca(2+)].

A-Current down-modulated by sigma receptor in frog pituitary melanotrope cells through a G protein-dependent pathway.

Gramicidin perforated patch-clamp recordings were used to study the effects of two sigma 1 receptor ligands, (+)-N-cyclopropylmethyl-N-methyl-1, 4-diphenyl-1-ethyl-but-3-en-1-ylamine hydrochloride (JO 1784) and (+)-pentazocine, on the transient outward potassium current (IA) in cultured frog melanotrope cells. (+)-Pentazocine reversibly decreased the current amplitude in a dose-dependent manner. The effects of (+)-pentazocine were mimicked by JO 1784 and were markedly reduced by the sigma 1 receptor antagonist, N, N-dipropyl-2-[4-methoxy-3-2(2-phenylethoxy)phenyl]-ethylamine monohydrochloride (NE 100). Inactivation rate of IA was best fitted with a double exponential function, yielding time constants of 23.7 and 112.5 ms. (+)-Pentazocine (20 microM) accelerated the current decay, decreasing the time constants to 10.7 and 59 ms, respectively. Current-voltage experiments revealed that (+)-pentazocine (20 microM) did neither modify the open-state I/V curves nor the voltage dependence of IA. However, (+)-pentazocine (20 microM) shifted the steady-state inactivation curve toward more negative potentials and increased the time constant of the time-dependent removal of inactivation. In whole-cell experiments, internal dialysis of guanosine-5'-O-(3-thiophosphate) (100 microM) irreversibly prolonged the response to (+)-pentazocine. In addition, cholera toxin pretreatment (1 microgram. ml-1; 12 h) suppressed the inhibition of IA by (+)-pentazocine (20 microM). It is concluded that in frog melanotrope cells, a cholera toxin-sensitive, G protein-dependent inhibition of IA through a sigma 1 receptor activation, at least partially, underlies the excitatory effect of sigma ligands.

Sigma ligands stimulate the electrical activity of frog pituitary melanotrope cells through a G-protein-dependent inhibition of potassium conductances.

We have investigated the effects of sigma ligands [1,3-di(2-tolyl)guanidine (DTG) and (+)-pentazocine] on the electrical activity of cultured frog pituitary melanotrope cells by using the patch-clamp technique. DTG and (+)-pentazocine (10 microM each) induced a reversible depolarization associated with an increase in membrane resistance and action potential firing. In voltage-clamp experiments, DTG and (+)-pentazocine elicited inward currents whose intensity augmented with membrane depolarization. The currents vanished or reversed between -90 and -100 mV, at values close to the K+ equilibrium potential (E(K)+ = -102 mV). DTG (2-500 microM) and (+)-pentazocine (0.2-200 microM) reduced the outward delayed rectifier K+ current [IK (V)] in a dose-dependent manner with EC50 of 64 and 37 microM, respectively. In contrast, naloxone (50 microM) and pirenzepine (10 microM) did not affect the sigma ligand-induced inhibition of IK (V). Addition of guanosine-5'-O-(3-thiophosphate) in the pipette solution irreversibly sustained the DTG-induced current whereas guanosine-5'-O-(2-thiodiphosphate) virtually suppressed the response. Cholera toxin-pretreatment (1 microgram/ml; 18 hr) abolished the inward current and the inhibition of IK (V) induced by sigma ligands. In contrast, pretreatment with pertussis toxin (1 microgram/ml; 18 hr) had no effect. Taken together, these data indicate that DTG and (+)-pentazocine activate the electrical activity of cultured frog melanotrope cells by reducing both a tonic K+ current and a voltage-dependent [IK (V)] K+ conductance through the activation of a cholera toxin-sensitive G-protein.

Adenosine potentiates the delayed-rectifier potassium conductance but has no effect on the hyperpolarization-activated Ih current in frog melanotrophs.

The effects of adenosine on the voltage-sensitive delayed-rectifier K+ (IK) currents and hyperpolarization-activated cationic inward current (Ih) were studied in cultured frog melanotrophs using the whole-cell configuration of the patch-clamp technique. The A1 receptor agonist R-N6-phenylisopropyl-adenosine (R-PIA; 50 microM) reversibly increased IK. Perfusion of dibutyryl-cAMP (1 mM) in the external solution did not modify the R-PIA-induced enhancement of IK. Pretreatment of melanotrophs with pertussis toxin (1 microg/ml; 12 h) totally abolished the R-PIA-evoked response. Application of hyperpolarizing voltage pulses from -60 to -120 mV to melanotrophs induced a two-component inward current corresponding to an Ih-like conductance. This conductance was characterized by a high K+ selectivity and a low Na+ permeability and was resistant to tetrodotoxin (1 microM). R-PIA had no effect on Ih. The present study demonstrates that in frog melanotrophs adenosine inhibits the electrical activity by activating IK through an A1 receptor subtype coupled to a pertussis toxin-sensitive pathway independent of the cAMP/PKA system. This study also demonstrates the existence of a Ih conductance in frog melanotrophs which is not modulated by A1 receptors.

Gramicidin-perforated patch revealed depolarizing effect of GABA in cultured frog melanotrophs.

1. In frog pituitary melanotrophs, GABA induces a transient stimulation followed by prolonged inhibition of hormone secretion. This biphasic effect is inconsistent with the elevation of cytosolic calcium and the inhibition of electrical activity also provoked by GABA in single melanotrophs. In the present study, standard patch-clamp configurations and gramicidin-perforated patches were used to investigate the physiological GABAA receptor-mediated response and intracellular chloride concentration ([Cl-]i) in cultured frog melanotrophs. 2. In the gramicidin-perforated patch configuration, 1 microM GABA caused a depolarization associated with an action potential discharge and a slight fall of membrane resistance. In contrast, at a higher concentration (10 microM) GABA elicited a depolarization accompanied by a transient volley of action potentials, followed by a sustained inhibitory plateau and a marked fall of membrane resistance. Isoguvacine mimicked the GABA-evoked responses, indicating a mediation by GABAA receptors. 3. In gramicidin-perforated cells, the depolarizing excitatory effect of 1 microM GABA was converted into a depolarizing inhibitory action when 0.4 microM allopregnanolone was added to the bath solution. 4. After gaining the whole-cell configuration, the amplitude and/or direction of the GABA-evoked current (IGABA) rapidly changed before stabilizing. After stabilization, the reversal potential of IGABA followed the values predicted by the Nernst equation for chloride ions when [Cl-]i was varied. 5. In gramicidin-perforated cells, the steady-state I-V relationships of 10 microM GABA- or isoguvacine-evoked currents yielded reversal potentials of -37.5 +/- 1.6 (n = 17) and -38.6 +/- 2.0 mV (n = 8), respectively. These values were close to those obtained by using a voltage-ramp protocol in the presence of Na+, K+ and Ca2+ channel blockers. The current evoked by 1 microM GABA also reversed at these potentials. 6. We conclude that, in frog pituitary melanotrophs, chloride is the exclusive charge carrier of IGABA. In intact cells, the reversal potential of IGABA is positive to the resting potential because of a relatively high [Cl-]i (26.5 mM). Under these conditions, GABA induces a chloride efflux responsible for a depolarization triggering action potentials. However, GABA at a high concentration or in the presence of the potentiating steroid allopregnanolone exerts a concomitant shunting effect leading to a rapid inhibition of the spontaneous firing.

[The mechanism on hyperpolarization of membrane potential induced by adenosine in frog melanotrophs].

It has been previously shown that the activation of A1 adenosine receptors in frog melanotroph induced a hyperpolarization accompanied by blockage of spontaneous action potentials. In the present report, we explored mechanisms underlying the above phenomenon using the patch-clamp technique in whole-cell and cell-attached configuration. The result showed that adenosine could increase the opening of non-voltage-activated potassium channels leading to hyperpolarization, but an inward cation current activated by hyperpolarizing pulse was not involved.

Multiple modulatory effects of the neuroactive steroid pregnanolone on GABAA receptor in frog pituitary melanotrophs.

1. The effects of the neuroactive steroid pregnanolone (5 beta-pregnan-3 alpha-ol-20-one) on the electrical response to GABA were investigated in cultured frog pituitary melanotrophs using the patch-clamp technique. 2. Low concentrations of pregnanolone (0.01-1 microM) in the extracellular solution enhanced the current evoked by submaximal concentrations of GABAA receptor agonists and prolonged the GABA-induced inhibition of the spontaneous action potentials in a dose-dependent manner. 3. Pregnanolone augmented the opening probability of the single GABA-activated channels but did not modify the conductance levels. 4. Pregnanolone (1 microM) shifted the GABA dose-response curve towards the low GABA concentrations, reducing the EC50 from 4.2 to 1.8 microM. 5. Internal cell dialysis with pregnanolone (1 or 10 microM) did not alter the GABA-evoked current. 6. Pregnanolone accelerated the desensitization of both the current and conductance increases caused by GABA. 7. High concentrations of pregnanolone (30 microM) markedly and reversibly diminished the current evoked by 10 microM GABA. 8. At high concentrations (10-30 microM), pregnanolone induced an outward current which reversed at the chloride equilibrium potential. 9. It is concluded that, in frog pituitary melanotrophs, pregnanolone exerts a dual inverse modulation and a direct activation of the GABAA receptor-channel depending on the concentrations of both GABA and steroid. Pregnanolone acts on an extracellular site on the GABAA receptor inducing conformational changes of the receptor-channel complex, resulting in a desensitized less-conducting state.

Electrophysiological effects of various neuroactive steroids on the GABA(A) receptor in pituitary melanotrope cells.

The action of steroids on the bioelectrical response to gamma-aminobutyric acid (GABA) has never been studied in pituitary cells. In the present study, we have thus investigated the effects of a series of neuroactive steroids on the GABA-activated current in frog melanotrope cells in primary culture, using the patch-clamp technique in the whole-cell configuration. Bath perfusion of 3alpha-isomers of pregnanolone or tetrahydrodeoxycorticosterone (1 microM) significantly enhanced the current evoked by short pulses of GABA (3 microM) and accelerated its desensitization. In contrast, the 3beta-isomers (30 microM) had no effect on the GABA-activated current. Addition to the bath solution of dehydroepiandrosterone or dehydroepiandrosterone sulfate (10 microM) inhibited the GABA-activated current without modifying its kinetics while pregnenolone sulfate (10 microM) both inhibited the GABA-activated current and accelerated its decay rate. The effects of pregnane steroids were not impaired by the central-type benzodiazepine receptor antagonist flumazenil (10 microM). In conclusion, the present study reveals that neuroactive steroids may exert multiple modulatory activities on the GABA(A) receptor borne by melanotrope cells. The effect of steroids on the current evoked by GABA is rapid, reversible, stereospecific and not mediated through the benzodiazepine binding site of the GABA(A) receptor.