J Biol Chem, Vol. 274, Issue 46, 32762-32770, November 12, 1999

Intracellular Pathways of V₁ and V₂ Receptors Activated by Arginine Vasopressin in Rat Hippocampal Neurons^*

Tomohiro Omura^§, Junichi Nabekura, and Norio Akaike^¶

From the  Department of Physiology, Graduate School of Medicine, Kyushu University, Fukuoka 812-8582 and ^§ Research Laboratories, Nippon Chemiphar Co., Ltd., Saitama 341-0005, Japan

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

To explore the intracellular pathways activated by vasopressin receptors, the effects of arginine vasopressin (AVP) and its analogues mediating glycine (Gly)-induced Cl currents (I_Gly) were examined in acutely dissociated rat hippocampal CA1 neurons using the whole-cell patch recording technique. AVP and its analogues inhibited I_Gly in a concentration-dependent manner. The inhibitory actions of AVP(4-9) (AVP metabolite) and NC-1900 (AVP(4-9) analogue) were reversed by a V₁ receptor antagonist, or pretreatment with 1,2-bis(2-amino-5-fluorophenoxy)ethane-N,N,N',N'-tetraacetic acid. In contrast, these blocking procedures had no effect on the 1-desamino-8-D-AVP (DDAVP; V₂ agonist) action. A V₂ receptor antagonist did not block the inhibitory action of AVP(4-9) or NC-1900, but blocked that of DDAVP. The inhibitory action of AVP was completely blocked by the co-application of the V₁ and V₂ antagonists. The inhibitory action of NC-1900 was not affected by perfusion with a Ca²⁺-free external solution, but was strongly blocked by thapsigargin. The intracellular application of heparin or anti-inositol 1,4,5-triphosphate (IP₃) also blocked the NC-1900 action. Furthermore, Ca²⁺/calmodulin (CaM) inhibitors blocked the NC-1900 action, while a CaM-dependent kinase II inhibitor and PKC modulators had no effect. 2',5'-Dideoxyadenosine (an adenylate cyclase inhibitor), H-89, and Rp-cAMPS blocked the inhibitory actions of NC-1900 and DDAVP. These results suggest that the activation of the V₁ receptor in the hippocampal neurons induces the production of IP₃, which releases Ca²⁺ from the IP₃-sensitive Ca²⁺ storage sites. The Ca²⁺ binds to CaM, resulting in the activation of Ca^2+/CaM-sensitive adenylate cyclases. The activation of protein kinase A through the adenylate cyclase inhibits I_Gly.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Arginine vasopressin (AVP)¹ synthesized in the supraoptic and paraventricular nuclei of the hypothalamus exerts antidiuretic and vasopressor effects. AVP also has a neuroregulatory effect in the central nervous systems including learning and memory (1). Vasopressin receptors have been classified into two major subtypes, named V₁ and V₂ receptors, based on their intracellular transduction mechanisms. The V₁ receptor is associated with phosphoinositide (PI) turnover (2), while the V₂ receptor activates adenylate cyclase (3).

The hypothalamic vasopressin-containing neurons project to the hippocampal regions (4, 5). Autoradiographic ligand-binding studies found vasopressin receptors in the pyramidal cell layer of the hippocampus (6, 7). The vasopressin receptor in the hippocampus has been shown to be the V₁ subtype in histological studies using a ³H-labeled V₁ antagonist (8, 9). Moreover, the possible existence of V₂ receptor in the rat hippocampus has also been suggested at the mRNA level (10). Extracellular recording revealed an increase of firing rate of hippocampal neurons stimulated by AVP (11). In addition, AVP has been shown to induce an increase of spike discharge in the hippocampal CA1 pyramidal neurons in studies using microelectrode techniques (12). In both electrophysiological (9, 11) and behavioral (13, 14) studies, the effect of AVP was blocked by a selective V₁ receptor antagonist. It has also been reported that AVP induced the production of IP₃ through the V₁ receptor in rat cultured hippocampal neurons (15, 16). Brinton and McEwen (17) demonstrated that AVP potentiates the norepinephrine-induced cAMP accumulation in a calcium-dependent manner in the rat hippocampal slice preparation through the V₁ receptor. It has been demonstrated that the activation of V_1a receptor potentiated the V₂ receptor-mediated cAMP accumulation in Chinese hamster ovary (CHO) cells transfected with both receptor cDNAs (18). However, the specific mechanism by which cAMP accumulates as a consequence of V₁ receptor activation has yet been elucidated for hippocampal neurons.

The intracellular cAMP levels may be modified by calcium through regulation of the activity of some adenylate cyclase isoforms. Nine subtypes of adenylate cyclase have been identified by cloning and expression studies, and they have been divided into six subfamilies based on their functional properties and sequence similarities: type 1, type 2-like group (types 2, 4 and 7), type 3, type 5-like group (types 5 and 6), type 8, and type 9 (19-23). The presence of types 1, 2, 4, and 8 messenger RNAs was revealed by the polymerase chain reaction in pyramidal neurons in the CA1 region (21, 23-25). The types 1 and 8 adenylate cyclases are regulated by not only Ca²⁺/calmodulin (CaM) but also by G_s (21, 26). The type 2 adenylate cyclase is insensitive to Ca²⁺/CaM but is regulated by G_s (27).

The glycine (Gly)-induced response is modulated by either protein kinase A (PKA) or protein kinase C (PKC) (28-33). In addition, we and others have shown that freshly dissociated rat hippocampal neurons respond to Gly (34-36). Therefore, to explore the intracellular pathways activated by vasopressin receptors, we have examined the affects of AVP and its analogues on the Gly-induced inhibitory responses in acutely dissociated pyramidal neurons from the rat hippocampal CA1 region using the patch-clamp technique.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Preparation-- Hippocanpal pyramidal neurons were acutely dissociated from the rat CA1 region, as described elsewhere (37). Briefly, 2-3-week-old Wistar rats were anesthetized with ether and decapitated. The brain was quickly removed from the skull and was sliced at a thickness of 400 µm with a microslicer (DTK-1000; Dosaka, Kyoto, Japan). The slices were then treated with Pronase (1 mg in 6 ml) for 15 min at 31 °C and subsequently with thermolysin (1 mg in 6 ml) under the same conditions. The CA1 region of the slice was removed by micropunching and mechanically triturated with fire-polished Pasteur pipettes in a 35-mm plastic culture dish (Primaria 3801; Becton Dickinson, Franklin Lakes, NJ) under a phase-contrast microscope (BH-2; Olympus, Tokyo, Japan). The dissociated CA1 neurons adhered to the bottom of the dish within 20 min.

Electrical Measurements-- Electrical measurements were performed with the nystatin perforated patch recording technique (38). In some experiments, the conventional whole-cell patch recording mode was used (Fig. 6C). The resistance between the recording electrode filled with the internal pipette solution and the reference electrode was 5-8 megohms. The current and voltage were measured with a patch-clamp amplifier (CEZ-2300; Nihon Kohden, Tokyo, Japan), filtered at 1 kHz (E-3201B; NF Electronic Instruments, Tokyo, Japan) and monitored on both an oscilloscope (VC-11; Nihon Kohden) and a pen recorder (RJG-4124; Nihon Kohden). The data were stored on videotapes after digitalization with a pulse-coded modulation processor (RP-880; NF Electronic Instruments). The membrane potential was held at 40 mV throughout the experiment, except when examining the current-voltage (I-V) relationships (Fig. 4C). All experiments were performed at room temperature (22-24 °C).

Solutions-- The ionic composition of the incubation solution was (mM): 124 NaCl, 5 KCl, 1.2 KH₂PO₄, 24 NaHCO₃, 2.4 CaCl₂, 1.3 MgSO₄, and 10 glucose. The pH of the incubation solution was adjusted to 7.45 with 95% O₂ and 5% CO₂. The ionic composition of the standard external solution was (mM): 150 NaCl, 5 KCl, 1 MgCl₂, 2 CaCl₂, 10 HEPES, and 10 glucose. The pH was adjusted to 7.4 with tris(hydroxymethyl)aminomethane (Tris). The I-V relationship for I_Gly was made in the standard solution containing 0.3 µM tetrodotoxin and 10 µM CdCl₂. The ionic composition of the patch pipette (internal) solution was (mM): 20 N-methyl-D-glucamine-methanesulfonate, 60 potassium methanesulfonate, 5 MgCl₂, 60 KCl, and 10 HEPES. The pH was adjusted to 7.2 with Tris. Nystatin was dissolved in the perforated patch pipette solution at a final concentration of 200 µg ml^1. In the conventional whole-cell patch recording, the ionic composition of the patch pipette solution was (mM): 20 N-methyl-D-glucamine-methanesulfonate, 60 potassium methanesulfonate, 0.5 MgCl₂, 60 KCl, 0.5 ATP, 0.5 GTP, 0.1 EGTA, and 10 HEPES. The pH was adjusted to 7.2 with Tris.

Drugs-- The drugs purchased were AVP (Peptide Institute, Inc., Osaka, Japan; Sigma); [Pmp¹,Tyr(Me)²]AVP (Peptide Institute, Inc.; American Peptide Co., Sunnyvale, CA); glycine, ryanodine, 3-isobutyl-1-methylxanthine (IBMX), tetrodotoxin, and EGTA (Wako Pure Chemical Industries, Ltd., Tokyo, Japan); N-(6-aminohexyl)-5-chloro-1-naphthalenesulfonamide (W-7), strychnine, thermolysin, [pGlu⁴,Cyt⁶]AVP(4-9) (AVP(4-9)), 1-desamino-8-D-AVP (DDAVP), ATP, GTP, forskolin, dibutyryl cyclic AMP (db-cAMP), trifluoperazine, and chlorpromazine (Sigma); N-(6-aminohexyl)-1-naphthalenesulfonamide (W-5) and BAPTA-AM (Molecular Probes, Inc., Eugene, OR); N-[2-(p-bromocinnamylamino)ethyl]-5-isoquinoline sulfonamide (H-89), Rp-adenosine-3',5'-cyclic monophosphorothioate (Rp-cAMPS), (s)-5-isoquinolinesulfonic acid (KN-62), chelerythrine, and 2',5'-dideoxyadenosine (Biomol Research Laboratories, Inc., Plymouth Meeting, PA); phorbol 12-myristate 13-acetate (PMA) and thapsigargin (Research Biochemicals, Inc., Boston, MA); anti-inositol 1,4,5-trisphosphate (IP₃) and Pronase (Calbiochem, La Jolla, CA); 1-oleoyl-2-acetylglycerol (OAG) (Doosan Serdary Research Laboratories, Englewood Cliffs, NJ). OPC-31260 was a gift (Otsuka Pharmaceutical Co., Ltd., Tokushima, Japan). L-Pyroglutamyl-L-asparaginyl-L-seryl-L-prolyl-L-arginylglycinamide (NC-1900) was synthesized by Nippon Chemiphar Co., Ltd. (Saitama, Japan).

Stock solutions of db-cAMP, IBMX, forskolin, and PMA were prepared in dimethyl sulfoxide (Me₂SO) and diluted to their final concentrations in standard solution just before use. The final concentration of Me₂SO was always less than 0.1%. It did not induce any ionic current and had no effect on the Gly-induced response at the concentrations used. The other drugs were dissolved in the standard external solution just before use. Drugs were applied by the "Y-tube system," which enables a solution to exchange within 20 ms (39).

Statistics-- The data are presented as the mean ± S.E. The results were analyzed with the Student's t test, paired t test, or Dunnett's multiple comparison test, and a p value of < 0.05 was considered to be significant.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Effects of Arginine Vasopressin and Its Analogues-- Rapid application of 10⁴ M Gly at 2-min intervals induced constant peak inward currents for 60 min at a holding potential (V_h) of -40 mV under the voltage-clamp conditions. Strychnine, a competitive Gly receptor antagonist, inhibited strongly the Gly-induced current (I_Gly) at a concentration of 10⁷ M (Fig. 1B, a). AVP or DDAVP alone did not induce any noticeable current at concentrations up to 10⁶ M. AVP(4-9), a AVP metabolite in the brain (40), or NC-1900, a synthetic analogue of AVP(4-9), also did not induce currents. However, the I_Gly was gradually inhibited during the continuous application of these peptides for 10 min (Fig. 1A, a-d). The inhibition completely recovered to the control levels after washing out the peptides. Fig. 1B summarized the inhibitory actions of AVP analogues and strychnine on Gly-induced response. The half-inhibition dose (IC₅₀) for strychnine was 5.2 × 10⁸ M. The percentages of inhibition of the I_Gly by NC-1900, AVP(4-9), AVP, and DDAVP at a concentration of 10⁶ M were 46.4 ± 5.5% (n = 14), 31.2 ± 3.8% (n = 5), 29.2 ± 4.7% (n = 5), and 31.8 ± 6.8% (n = 4), respectively. NC-1900 was the most potent inhibitor. The block by strychnine appeared to reach steady state immediately. In contrast, AVP and its analogues required several minutes to attain steady state. The difference in time course suggests that AVP and its analogues act indirectly on the Gly receptor.

View larger version (23K): [in this window] [in a new window]   Fig. 1.   The inhibitory effects of AVP and its analogues on the 104 M Gly-induced current (IGly). A, typical examples of IGly inhibited by AVP (a), AVP(4-9) (b), NC-1900 (c), and DDAVP (d) at a concentration of 106 M. Gly was applied at 2-min intervals at a holding potential (Vh) of -40 mV. B, a, a typical example of IGly inhibited by 107 M strychnine. b, concentration-inhibition curves for strychnine, AVP, AVP(4-9), NC-1900, and DDAVP. All currents were normalized to the peak current amplitude induced by 104 M Gly alone. Each point and vertical bar show the mean ± S.E. of 4-14 neurons.

Effect of a V₁ Receptor Antagonist-- To determine which vasopressin receptor subtypes mediate the inhibition of I_Gly, the effect of [Pmp¹,Tyr(Me)²]AVP, a selective V₁ receptor antagonist (41), was examined. Pretreatment with 10⁵ M [Pmp¹,Tyr(Me)²]AVP alone induced no current and had no effect on I_Gly. However, the inhibitory actions of 10⁶ M NC-1900 and 10⁶ M AVP(4-9) on I_Gly were completely blocked by the pretreatment with the V₁ antagonist (Fig. 2A, a and b). Fig. 2B summarizes the blocking effects of the V₁ antagonist on the inhibitory actions of AVP and its related peptides on I_Gly. In this figure, the independent points on the left-hand side (arrow) indicate the control percentage of inhibition of I_Gly by these peptide alone. The V₁ receptor antagonist blocked the inhibitory actions of 10⁶ M NC-1900 and 10⁶ M AVP(4-9) in a concentration-dependent manner with a near-complete block at a concentration of 10⁵ M (p < 0.01; Dunnett's multiple comparison test). In contrast, the inhibitory effect of 10⁶ M AVP was only partially blocked (p < 0.05), while the inhibitory effect of DDAVP was not affected by the V₁ antagonist.

View larger version (22K): [in this window] [in a new window]   Fig. 2.   Effect of a selective V1 receptor antagonist, [Pmp1, Tyr(Me)2]AVP, on the inhibitory actions of AVP and its analogues on IGly. A, the blockade of NC-1900 (a) and AVP(4-9) (b) actions on IGly by 105 M V1 receptor antagonist. B, concentration-inhibition curves for the V1 antagonist on the inhibitory actions of these peptides. The independent points indicated by the arrow are the control percentage of inhibition of IGly without the antagonist (n = 4-14). All currents in the presence of the V1 antagonist were normalized to the corresponding control currents. Each point and vertical bar represent the mean ± S.E. of 4 to 5 neurons. *, p < 0.05; **, p < 0.01 (Dunnett's multiple comparison test).

Effect of a V₂ Receptor Antagonist-- To explore the contribution of V₂ receptors, the effect of OPC-31260, a selective V₂ receptor antagonist (42), was examined. Pretreatment with 10⁶ M OPC-31260 alone induced no current and had no effect on I_Gly. The inhibitory action of 10⁶ M DDAVP on I_Gly was completely blocked by pretreatment with OPC-31260 (Fig. 3A, a), whereas the action of NC-1900 on I_Gly was not affected (the percentage of inhibition of I_Gly was still 48.3 ± 5.2%, n = 4) (Fig. 3A, b). The concentration-inhibition relationships between the peptides and the V₂ antagonist are shown in Fig. 3B, in which the independent points on the left-hand side (arrow) indicate the control inhibition percentage of I_Gly by these peptides. The inhibitory action of 10⁶ M DDAVP on I_Gly was blocked in a concentration-dependent manner by OPC-31260, and the V₂ antagonist considerably blocked the action of DDAVP at concentrations above 10⁷ M (Dunnett's multiple comparison test). The results indicate that the inhibitory action of DDAVP on I_Gly is mediated through the V₂ receptor. The inhibitory action of 10⁶ M NC-1900 or 10⁶ M AVP(4-9) on I_Gly was not affected by the V₂ receptor antagonist. The inhibitory action of 10⁶ M AVP was partially blocked by the V₂ antagonist, but the co-application of 10⁵ M V₁ antagonist and 10⁶ M V₂ antagonist completely blocked this peptide action (Fig. 3C).

View larger version (26K): [in this window] [in a new window]   Fig. 3.   Effect of a selective V2 receptor antagonist, OPC-31260, on the inhibitory actions of these peptides on IGly. A, the effect of 106 M V2 receptor antagonist on the inhibitory actions of DDAVP (a) and NC-1900 (b) on IGly. B, concentration-inhibition curves for the V2 antagonist on the inhibitory actions of peptides. The independent points indicated by the arrow are the percentage of inhibition of IGly without the antagonist (n = 4-14). All currents in the presence of the V2 antagonist were normalized to the corresponding control currents. Each point and vertical bar represent the mean ± S.E. of 4 to 5 neurons. *, p < 0.05; **, p < 0.01 (Dunnett's multiple comparison test). C, the effect of co-application of 105 M [Pmp1,Tyr(Me)2]AVP and 106 M OPC-31260 on the effect of AVP on IGly.

Effect of NC-1900 on I_Gly-- Since NC-1900 had the most potent inhibitory action on I_Gly, the concentration-response relationship was examined. All currents induced by Gly at various concentrations with or without 10⁶ M NC-1900 were normalized to the peak current induced by 10⁴ M Gly alone (asterisk in Fig. 4A). NC-1900 inhibited the maximum value of the concentration-response relationship for Gly without affecting the apparent K_D (8.5 × 10⁵ M and 6.5 × 10⁵ M for with and without NC-1900, respectively) or the hill coefficient (1.01 and 1.02, respectively), indicating that the inhibition is non-competitive. Fig. 4B (a and b) shows the I-V relationships for I_Gly in the presence or absence of 10⁶ M NC-1900. The reversal potential of the I_Gly (E_Gly) was 19.0 ± 0.5 mV (n = 5) in the control and 18.7 ± 0.8 mV (n = 5) in the presence of 10⁶ M NC-1900. These E_Gly values were close to the Cl equilibrium potential (E_Cl) of -21.0 mV (arrow) calculated from the Nernst equation based on the given external and internal Cl concentrations (161 and 70 mM, respectively). The inhibition of I_Gly by NC-1900 did not show any voltage dependence. The desensitization rates also did not appear to vary with NC-1900 application (Fig. 4B, a).

View larger version (20K): [in this window] [in a new window]   Fig. 4.   Effect of NC-1900 on IGly. A, concentration-response relationship for IGly with (closed circle) or without (open circle) 106 M NC-1900. All currents induced by Gly at various concentrations were normalized to the current induced by 104 M Gly (*) alone. Each point and the vertical bar show the mean ± S.E. of 4 to 5 neurons. B, a, representative IGly recordings with or without 106 M NC-1900 at Vh levels of -40, 0, and +40 mV. b, the I-V relationships for IGly in the presence (closed circle) and absence (open circle) of 106 M NC-1900. The arrow indicates the Cl equilibrium potential (ECl). All currents were normalized to the peak amplitude induced by 104 M Gly alone at a Vh of -40 mV (*). Each point indicates the average of 4 to 6 neurons.

Effect of Intracellular Ca²⁺ Concentration ([Ca²⁺]_i)-- Since the activation of the V₁ receptor stimulates the hydrolysis of PI (2) and results in the elevation of [Ca²⁺]_i, the effect of [Ca²⁺]_i on the I_Gly inhibitory actions of these peptides was examined. The CA1 neurons were loaded with 5 × 10⁶ M BAPTA-AM, a membrane-permeable Ca²⁺ chelator, for 2 h. The inhibitory action of 10⁶ M NC-1900 on I_Gly was completely abolished in BAPTA-AM-treated neurons (Fig. 5A, a). The inhibition by 10⁶ M AVP(4-9) on I_Gly was also blocked in the BAPTA-AM-treated neurons (Fig. 5A, b). On the other hand, the inhibitory action of 10⁶ M AVP was only partially blocked in the BAPTA-AM-treated neurons, while that of 10⁶ M DDAVP was not affected at all (Fig. 5A, c and d). Fig. 5B summarizes the effect of BAPTA-AM for these peptide actions on I_Gly. The percentages of inhibition of NC-1900, AVP(4-9), AVP, and DDAVP in the BAPTA-AM-treated neurons were 9.9 ± 1.2% (n = 4), 6.8 ± 3.7% (n = 5), 17.2 ± 1.8% (n = 6), and 27.2 ± 1.7% (n = 4), respectively. The inhibitory effects of NC-1900, AVP(4-9) and AVP were significantly removed by BAPTA-AM treatment (p < 0.01 for NC-1900 and AVP(4-9), and p < 0.05 for AVP; Student's t test).

View larger version (18K): [in this window] [in a new window]   Fig. 5.   Effect of pretreatment with BAPTA-AM. A, the effects of AVP and its analogues on IGly in the neurons loaded with 5 × 106 M BAPTA-AM for 2 h. B, the percentages of inhibition of IGly by these peptides at a concentration of 106 M in the control (open column) and the BAPTA-AM-loaded (closed column) neurons. Each column and vertical bar represent the mean ± S.E. of 4 to 6 neurons. *, p < 0.05; **, p < 0.01 (Student's t test).

Calcium Dependence of NC-1900 Inhibitory Action-- Since the NC-1900 action on I_Gly through the V₁ receptor was most potent among the examined peptides (Fig. 1), the following experiments were carried out. First, the Ca²⁺ dependence for the inhibitory action of NC-1900 on I_Gly was investigated. After the I_Gly was inhibited by the pretreatment with 10⁶ M NC-1900 in a standard external solution for 10 min, the external solution was changed to Ca²⁺-free external solution containing 2 mM EGTA. The inhibitory action of NC-1900 persisted (Fig. 6A), suggesting that extracellular Ca²⁺ ([Ca²⁺]_o) is not involved in the inhibition of I_Gly by NC-1900.

View larger version (18K): [in this window] [in a new window]   Fig. 6.   Effects of extra- and intracellular Ca2+ on the inhibitory action of NC-1900. A, effect of a Ca2+-free external solution. After the IGly was inhibited in the standard solution by the application of 10-6 M NC-1900, the external solution was changed to a Ca2+-free external solution with 2 mM EGTA. B, a, the effect of 105 M thapsigargin. b, the effect of 106 M ryanodine treatment. C, effect of heparin (5 mg ml1) (closed circle) or anti-IP3 (5 µg ml1) (open triangle) on the inhibitory action of IGly by NC-1900. Recordings were made with the conventional whole-cell patch recording mode. All currents were normalized to the respective initial IGly (n = 4-5).

A possible involvement of intracellular Ca²⁺ storage sites in the NC-1900 action was confirmed by following experiments. The application of 10⁵ M thapsigargin, a Ca²⁺-ATPase inhibitor which depletes Ca²⁺ in the IP₃-sensitive Ca²⁺ storage sites (IICR), did not affect the peak amplitude of I_Gly, but facilitated slightly the desensitization phase of I_Gly· The inhibitory action of NC-1900 on I_Gly was completely blocked by the pretreatment with thapsigargin (n = 4) (Fig. 6B, a), suggesting the Ca²⁺ release from IICR in the presence of NC-1900. Ryanodine depletes the Ca²⁺ release from caffeine-sensitive intracellular Ca²⁺ storage sites (CICR). Even after the 10² M caffeine-induced outward current was considerably reduced by a continuous perfusion with 10⁶ M ryanodine, the inhibitory effect of NC-1900 on I_Gly was still observed, though the inhibition ratio of I_Gly by NC-1900 was reduced from 44.3 ± 4.6% of control (n = 5) to 35.0 ± 2.3% (n = 4) in the presence of ryanodine (Fig. 6B, b).

In the conventional whole-cell patch recording mode, the inhibitory action of I_Gly by NC-1900 was also observed (42.4 ± 5.6%, n = 5) (Fig. 6C, open circles). Therefore, the direct modification of IP₃ was examined by the intracellular perfusion with heparin (5 mg ml¹) or anti-IP₃ (5 µg ml¹) through the recording patch pipette (43). The inhibitory action of NC-1900 on I_Gly was completely blocked by either heparin or anti-IP₃ (closed circle and open triangle, respectively), suggesting that IP₃ plays an important role in the I_Gly inhibition by NC-1900. These results strongly support the hypothesis that Ca²⁺ release from the IICR is involved in the inhibitory action of NC-1900 on I_Gly.

Effects of Ca²⁺/CaM Inhibitors-- In order to clarify the signal transduction pathway following the intracellular free Ca²⁺ on the NC-1900 action, the possible contribution of Ca²⁺/CaM and CaM-dependent kinase II was investigated. W-7, trifluoperazine (TFP), and chlorpromazine (CPZ) are known Ca²⁺/CaM inhibitors. The inhibitory action of NC-1900 on I_Gly was blocked in a concentration-dependent fashion by treatment with 10⁵ M W-7, TFP, or CPZ, although these compounds alone had no effect on the I_Gly (Fig. 7, A, a-c, and B). In Fig. 7B, the open circle represents the percentage of inhibition of the I_Gly by 10⁶ M NC-1900 in the absence of these inhibitors. W-5 used as a negative control for W-7 had no effect. On the other hand, 10⁶ M KN-62, a potent CaM-dependent kinase II inhibitor, did not affect the inhibitory action of NC-1900 on I_Gly (the inhibition of I_Gly was 41.7 ± 7.9%, n = 8) (Fig. 7A, d). The results indicate that Ca²⁺/CaM is involved in the inhibition of I_Gly by NC-1900.

View larger version (19K): [in this window] [in a new window]   Fig. 7.   Effects of Ca2+/CaM and CaM-dependent kinase II inhibitors. A, the effects of 105 M W-7 (a), 105 M TFP (b), 105 M CPZ (c), and 106 M KN-62 (d) on the inhibition of IGly by NC-1900. B, concentration-inhibition curves for Ca2+/CaM inhibitors. The open circle in the left side is the percentage of inhibition of IGly by 106 M NC-1900 in the absence of these inhibitors. Each point represents the average values from 6 neurons. W-5 is a negative control of W-7. *, p < 0.05; **, p < 0.01 (Dunnett's multiple comparison test).

Effects of Protein Kinase A and C Modulators-- The hydrolysis of PI stimulated by the activation of V₁ receptor produces second messengers such as IP₃ and diacylglycerol (DAG), which activate PKC. The PKC would then phosphorylate the Gly receptor and modulates Gly response (30, 31, 33). The activation of PKA reduces the Gly response in the substantia nigral (29) and the ventromedial hypothalamic neurons (28). Therefore, the modulatory effect of either PKC or PKA on I_Gly was examined to clarify whether these protein kinases were involved in the inhibitory action of NC-1900 on I_Gly.

OAG (3 × 10⁶ M), a membrane permeable DAG analogue, and 10⁷ M PMA, a PKC activator, significantly potentiated the I_Gly to 1.21 ± 0.1 (n = 5) and 1.17 ± 0.05 (n = 4) times that of the control, respectively. Chelerythrine (3 × 10⁶ M), a PKC inhibitor, hardly affected the I_Gly (0.98 ± 0.02, n = 5). Forskolin (10⁶ M), an activator of adenylate cyclase, inhibited the I_Gly by 0.37 ± 0.03 (n = 4). The mixture of 10⁶ M IBMX, a phosphodiesterase inhibitor, and 10⁴ M db-cAMP, a membrane-permeable cAMP analogue, also inhibited the I_Gly by 0.45 ± 0.09 (n = 5). Rp-cAMPS (10⁴ M), a potent PKA inhibitor, potentiated the I_Gly to 1.42 ± 0.08 (n = 4) (Fig. 8A). Another PKA inhibitor, H-89 (10⁶ M) also potentiated the I_Gly to 1.42 ± 0.04 (n = 5). These results were the same as those reported previously (28, 29, 32).

View larger version (31K): [in this window] [in a new window]   Fig. 8.   Effects of PKC and PKA modulators on IGly. A, effects of PKC and PKA modulation on IGly. All currents were normalized to the respective control current amplitude. Each column and vertical bar represent the mean ± S.E. of 4 to 7 neurons. B, 3 × 106 M OAG (a) and 107 M PMA (b) potentiated the IGly· Chelerythrine of 3 × 106 M had no effect on IGly (c). NC-1900 (106 M) could inhibit the IGly regardless of the presence of these modulators (a-c). Effects of OAG, PMA, and chelerythrine on the inhibitory action of 106 M NC-1900 on IGly (d). All currents were normalized to the individual control currents in the presence of modulators without NC-1900. Each column and vertical bar show the mean ± S.E. of 4-5 neurons. *, p < 0.05; **, p < 0.01 (Dunnett's multiple comparison test in A and a paired t test in B, d).

NC-1900 still could inhibit the I_Gly to a similar extent in the presence of PKC modulators (Fig. 8B, a-c). The inhibitory effects of NC-1900 on I_Gly in the presence of PKC modulators are summarized in Fig. 8B (d). The relative values of I_Gly inhibition by 10⁶ M NC-1900 in the presence of OAG, PMA, and chelerythrine were 0.43 ± 0.06 (n = 4), 0.44 ± 0.02 (n = 4), and 0.42 ± 0.05 (n = 5), respectively. The results indicate that the PKC pathway is unlikely to participate in the inhibition of I_Gly by NC-1900.

The inhibitory action of NC-1900 on I_Gly was, however, completely removed in the presence of PKA activators or inhibitors, suggesting the PKA pathway is involved in the NC-1900 action. The percentage of inhibition of the I_Gly by NC-1900 (10⁶ M) were 3.0 ± 2.7% (n = 5), 7.2 ± 5.5% (n = 4), 2.7 ± 2.1% (n = 4) and 1.8 ± 2.1% (n = 5) in the presence of forskolin, mixture of IBMX and db-cAMP, Rp-cAMPS, and H-89, respectively (Fig. 9, A and B, a). The inhibitory effect of DDAVP was also reversed in the presence of Rp-cAMPS or H-89. The relative value of I_Gly in the presence of 10⁴ M Rp-cAMPS was 0.98 ± 0.04 (n = 4) (Fig. 9B, b), and that in the presence of 10⁶ M H-89 was 1.01 ± 0.03 (n = 4). Furthermore, the application of 2',5'-dideoxyadenosine (DDA), a membrane-permeable adenylate cyclase inhibitor (44), completely blocked the inhibitory actions both of NC-1900 and DDAVP at a concentration of 10³ M. The percentage of inhibition of the I_Gly by NC-1900 (10⁶ M) and DDAVP (10⁶ M) in the presence of DDA (10³ M) were 7.2 ± 3.4% (n = 5) and 8.3 ± 3.7% (n = 4), respectively (Fig. 9C, a and b). The results in the presence of DDA at various concentrations are summarized in Fig. 9C (c). The open symbols represent the percentage of inhibition of the I_Gly by NC-1900 and DDAVP in the absence of DDA (45.1 ± 6.7% and 31.1 ± 2.9%, respectively, n = 4 to 5). DDA blocked the inhibitory actions of NC-1900 and DDAVP in a concentration-dependent manner.

View larger version (22K): [in this window] [in a new window]   Fig. 9.   Effects of PKA modulators on the inhibitory action of NC-1900 on IGly. A, forskolin (106 M) (a) and a mixture of 106 M IBMX and 104 M db-cAMP (b) inhibited the IGly. NC-1900 at 106 M had no effect in the presence of these modulators. B, the IGly enhanced by 104 M Rp-cAMPS was not affected by additional application of NC-1900 (a) and DDAVP (b). C, the effect of 103 M DDA on the inhibitory actions of IGly by NC-1900 (a) and DDAVP (b). DDA itself had no effect on IGly. c, concentration-inhibition curve for DDA on the inhibitory actions of NC-1900 and DDAVP on IGly at a concentration of 106 M. The open symbols in the left side are the percentages of inhibition of IGly by NC-1900 and DDAVP in the absence of DDA. All currents were normalized to the control current amplitude in the absence of NC-1900. Each point and vertical bar show the mean ± S.E. of 4 neurons. **, p < 0.01 (Dunnett's multiple comparison test).

Effect of Co-application of AVP(4-9) and DDAVP-- To clarify the interaction between the V₁ and V₂ receptors, we examined the effect of co-application of 10⁶ M AVP(4-9) and DDAVP, which act on V₁ and V₂ receptors, respectively. Each peptide inhibits the I_Gly to a similar extent at this concentration (Fig. 1A, b and d). The mixture of AVP(4-9) and DDAVP inhibited the I_Gly by 32.1 ± 2.5% (Fig. 10, n = 4). Thus, the inhibitory effect of the mixture was similar to that of each peptide alone.

View larger version (9K): [in this window] [in a new window]   Fig. 10.   Effects of co-application of AVP(4-9), a V1 agonist, and DDAVP, a V2 agonist, on IGly. The mixture of 106 M AVP(4-9) and 106 M DDAVP inhibited the IGly by 32.1 ± 2.5% (n = 4).

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

V₁ or V₂ Receptor-mediated Inhibitory Actions on I_Gly-- The present study has demonstrated that AVP and its analogues inhibit the I_Gly in the pyramidal neurons of the hippocampal CA1 region. NC-1900 had the most potent inhibitory action among these peptides. In previous electrophysiological (11, 12) and biochemical (15-17) studies using the hippocampal preparations, AVP exerted its effect at concentrations between 10⁹ M and 10⁶ M. In the present study, AVP inhibited the I_Gly at the concentration between 10¹⁰ M and 10⁶ M. The administration of AVP, AVP(4-9) or NC-1900 to mice reduced the onset time of strychnine- or picrotoxin-induced convulsion, though these peptides themselves did not induce any convulsion.² Unlike strychnine, these peptides may not inhibit the I_Gly completely even at higher concentration.

The inhibitory action of AVP on I_Gly was partially blocked by either 10⁵ M [Pmp¹,Tyr(Me)²]AVP, a V₁ antagonist (Figs. 2 and 3), or 10⁶ M OPC-31260, a V₂ antagonist. The mixture of both antagonists completely blocked the AVP action (Fig. 3). In addition, the AVP action was partially removed in the BAPTA-AM loaded neurons (Fig. 5). The results indicate that the AVP action on I_Gly is mediated not only by V₁ receptor but also by V₂ receptor. It has been suggested that all the effects of AVP on the neurons in hippocampal slices are mediated through V₁ receptors (9, 11, 16). However, the possible existence of V₂ receptor in the rat hippocampus has been suggested by the presence of V₂ receptor mRNA (10). In addition, we demonstrated that DDAVP, a selective V₂ agonist, inhibited the I_Gly, and its action was blocked by V₂ antagonist. Thus, present data indicate the existence of both V₁ and V₂ receptors in the hippocampal CA1 neurons.

Mühlethaler et al. (45) reported that AVP and oxytocin (OT) exerted an electrophysiological effect via uterine-type OT receptor on non-pyramidal neurons in the hippocampus. However, Mizuno et al. (12) showed, using intracellular recording techniques, that 10⁶ M AVP increased the spike discharge of hippocampal pyramidal neurons. Chepkova et al. (46) also demonstrated that 10¹⁰ M AVP increased the amplitude and slope of EPSPs of the hippocampal pyramidal neurons, using intracellular recording techniques. Brinton and McEwen (17) demonstrated that AVP enhanced the cAMP accumulation induced by norepinephrine in cultured hippocampal neurons, whereas OT did not. Similarly, it has been reported that 10⁷ M AVP leads to IP₃ production in rat hippocampal slices via V₁ receptors, whereas OT has no affect (15). Both AVP and OT at the concentration of 2.5 × 10⁷ M induced IP₃ production in cultured hippocampal neurons, but these affects were exerted through the V₁ receptor (16). In addition, in autoradiographic studies, the AVP receptors were observed in the pyramidal cell layer (6, 7). These results suggest that the inhibitory effect of AVP on the I_Gly observed in this study was exerted through the V₁ receptors in the hippocampal pyramidal neurons.

The inhibitory effects of both NC-1900 and AVP(4-9) on I_Gly were not blocked by the V₂ antagonist but by the V₁ antagonist, and these inhibitory actions were completely reversed by the pretreatment with BAPTA-AM. These results indicate that NC-1900 and AVP(4-9) activate the V₁ receptor in hippocampal neurons. Blockade of the AVP(4-9) effect by the V₁ antagonist has also been reported in other studies using both neurochemical (16, 47) and behavioral paradigms (48). However, the autoradiographic studies indicated that binding sites for AVP(4-9) are distinct from that of putative V₁ receptors in the hippocampus or other brain regions (49, 50). NC-1900 failed to inhibit the [³H]AVP binding, but it partially inhibited the [³H]V₁ antagonist binding in a rat hippocampal membrane preparation.² The discrepancies among these studies suggest two alternative explanations. First, AVP, AVP(4-9), and NC-1900 recognized different regions of the V₁ receptor, while V₁ antagonist recognizes all of them. Second, these peptides recognize the different subtypes of V₁ receptor while V₁ antagonist recognizes all of them. The recent molecular cloning of peripheral vasopressin receptors has revealed the presence of 5-9 different bands in the human and rat genomic DNA, which hybridized with a probe for the rat V₁ receptor cDNA (51). Additional members of the vasopressin receptor gene family may well exist. Further functional investigations are necessary to confirm this hypothesis.

Intracellular Mechanisms of V₁ Receptor-mediated Action-- The inhibitory action of NC-1900 on I_Gly was completely abolished (Fig. 6C) in the conventional whole-cell recording using the patch pipette filled with heparin or anti-IP₃. This result suggests that IP₃ is involved in vasopressin transduction. Both AVP and AVP(4-9) induced the production of IP₃ in cultured hippocampal neurons through activation of the V₁ receptor (16). Thus, these results suggest that the inhibitory actions of NC-1900, AVP(4-9) and AVP on I_Gly through the V₁ receptor are mediated by IP₃. In addition, the NC-1900 action was completely blocked by the treatment with thapsigargin (Fig. 6B, a), suggesting the involvement of Ca²⁺ release from IICR in the inhibitory action of NC-1900. However, the inhibitory effect of NC-1900 was also suppressed slightly by treatment with ryanodine (Fig. 6B, b). Such a minor effect of ryanodine might be due to the close coupling or interaction between IICR and CICR (52) rather than a partial involvement of CICR. Also, the inhibitory action of NC-1900 on I_Gly was not affected in Ca²⁺-free external solution (Fig. 6A) but was blocked by chelating [Ca²⁺]_i with BAPTA-AM (Fig. 5A, a, and B). These results clearly indicate that the main source of the intracellular free Ca²⁺ increase in the presence of NC-1900 is IICR.

The inhibitors of Ca²⁺/CaM block the inhibitory action of NC-1900 on I_Gly (Fig. 7). Interestingly, the action of NC-1900 was not affected by PKC activators or inhibitors but was suppressed by PKA modulators (Figs. 8 and 9). Brinton and McEwen (17) demonstrated that AVP facilitated the norepinephrine-induced cAMP accumulation in the hippocampal slices, and this facilitation was blocked by TFP. These results may support the present findings that the Ca²⁺/CaM-sensitive adenylate cyclase (type 1 or 8) mediates the inhibitory actions of NC-1900 and AVP(4-9) on I_Gly. Previous biochemical studies have shown that AVP alone does not elevate cAMP level in the hippocampal slice preparation (17, 53, 54). In this study, however, AVP and its analogues exerted an inhibitory action on I_Gly that was mediated by PKA activation. This result strongly suggests that these peptides can induce cAMP accumulation in dissociated hippocampal neurons. Brinton and Brownson (55) have demonstrated that AVP(4-9) induces cAMP accumulation in cultured hippocampal neurons. Brinton et al. (16) also indicated that AVP(4-9) elevated IP₃ levels in cultured hippocampal neurons, and its effect was blocked by V₁ antagonist. In the present study, AVP(4-9) inhibited I_Gly through the V₁ receptor (Figs. 1 and 2). These results indicate that AVP(4-9) alone, which stimulates the V₁ receptor, can induce cAMP accumulation through IP₃ production in the hippocampal neurons. In vascular smooth muscle cells, AVP alone induces cAMP accumulation by activating the Ca²⁺/CaM-sensitive adenylate cyclase (type 3) expressed in the smooth muscle cells through the V₁ receptors and modulates the isoproterenol-stimulated cAMP accumulation (56).

Brinton and McEwen (17) have also reported that the facilitative effect of AVP on norepinephrine-induced cAMP accumulation depends on [Ca²⁺]_o. In cultured hippocampal neurons, the activation of V₁ receptors induced Ca²⁺ influx (16), suggesting the necessity of extracellular Ca²⁺ for eliciting V₁ receptor-mediated responses. Furthermore, the activation of N-methyl-D-aspartate (NMDA) receptors in the CA1 region increased the level of cAMP through the activation of a Ca²⁺/CaM-sensitive adenylate cyclase (57, 58). The increase of cAMP induced by NMDA depended on the presence of extracellular Ca²⁺ (57). Based on these results, they suggested that Ca²⁺ influx was required for the activation of Ca²⁺/CaM-sensitive adenylate cyclase. In the present study, however, the inhibitory effect of NC-1900 on I_Gly was not mediated by Ca²⁺ influx but by the rise of intracellular free Ca²⁺ released from IICR, suggesting the possible activation of Ca²⁺/CaM-sensitive adenylate cyclase in the absence of Ca²⁺ influx. Interestingly, it has been reported that -aminobutyric acid-induced current was also modulated by an increase of intracellular free Ca²⁺ released from intracellular Ca²⁺ storage sites but not by Ca²⁺ influx in cultured porcine pituitary intermediate lobe neurons (59). In addition, Zhang et al. (56) demonstrated that influx of Ca²⁺ from extracellular medium appears to contribute little to the AVP-induced enhancement action of isoproterenol-stimulated cAMP accumulation in the smooth muscle cells.

A pathway that is consistent with the present data is summarized in Fig. 11. Both AVP(4-9) and NC-1900 produce IP₃ through the activation of V₁ receptors. The IP₃ induces Ca²⁺ release from IICR, and the increased intracellular free Ca²⁺ is bound to CaM. The Ca²⁺/CaM activates the Ca²⁺/CaM-sensitive adenylate cyclase. Activation of adenylate cyclase results in the activation of PKA, which inhibits the Gly-induced Cl current. However, AVP activates both V₁ and V₂ receptors, and DDAVP inhibits the I_Gly through the V₂ receptor by activating G_s. In CHO cells transfected with the V_1a and V₂ receptor cDNA, V₂-induced cAMP accumulation was synergistically potentiated by the stimulation of V_1a receptor transduction pathway. However, this is not the case in the rat hippocampal CA1 neurons because the co-application of V₁ and V₂ receptor selective agonists, AVP(4-9) and DDAVP, had no additive or synergistic effect in the I_Gly inhibition (Fig. 10). The stimulation of V₁ receptor activates Ca²⁺/CaM-sensitive adenylate cyclase (type 1 or 8). On the other hand, the stimulation of V₂ receptor activates Ca²⁺-insensitive adenylate cyclase (type 2 or 4). The type 2 adenylate cyclase is also activated by PKC (26). We demonstrated, however, that PKC was not involved in the actions of AVP and its analogues, suggesting that only type 2 adenylate cyclase is not involved in the intracellular pathways stimulated by these peptides. The existence of type 4 adenylate cyclase in the mouse hippocampus has been demonstrated (25). To explain the lack of additive effect in the inhibitory action of I_Gly by co-application of V₁ and V₂ agonists (Fig. 10), two alternative possibilities can be raised. 1) Activation of V₁ and V₂ receptors share a common intracellular pathway. The amounts of intracellular mediators such as PKA and glycine receptors might be the limiting factors for the inhibitory actions of these peptides. However, this is not likely because 10⁶ M NC-1900 inhibited I_Gly more than co-application of 10⁶ M V₁ and V₂ agonists employed in Fig. 10 (Fig. 1B). 2) the reduction of type 1 adenylate cyclase activity by the activation of V₂ receptor. Type 1 adenylate cyclase is inhibited by the -subunit (19). Further investigation using specific inhibitors is needed to reveal the adenylate cyclase isoforms related to the intracellular pathways stimulated by these receptors.

View larger version (31K): [in this window] [in a new window]   Fig. 11.   Schematic illustration of the intracellular signal transduction pathways through the V1 and V2 receptors activated by AVP-related peptides. *AC, Ca2+/CaM-sensitive adenylate cyclase is suspected.

Recently, it has been demonstrated that AVP induces a [Ca²⁺]_i increase in vasopressinergic magnocellular supraoptic nucleus neurons through both phospholipase C (PLC)- and adenylate cyclase-linked signal transduction pathways (60). They hypothesized three possibilities: first, that AVP activates both V₁ and V₂ receptors; second, that AVP activates a unique vasopressin receptor, which could be coupled to both PLC and adenylate cyclase; and third, that AVP activates V₁ receptor uniquely coupled to PLC whose intracellular cascade could stimulate the cAMP second messenger system. Our results suggest that AVP activate both V₁ and V₂ receptors in the hippocampal pyramidal neurons, consistent with the first hypothesis for magnocellular supraoptic nucleus neurons. In addition, AVP(4-9) and NC-1900 activation of the V₁ receptor results in the accumulation of cAMP through intracellular pathways, consistent with the third hypothesis. In any case, interestingly, AVP and its analogues require no Ca²⁺ influx for their inhibitory actions on I_Gly in the hippocampal CA1 neurons.

	ACKNOWLEDGEMENTS

We thank Dr. M. Uki (Otsuka Pharmaceutical Co., Ltd.) for the kind gift of OPC-31260 and Drs. M. C. Andresen and Dan H. Sanes for their helpful comments and critical reading.

	FOOTNOTES

^* This work was supported by Grants-in-aid for Scientific Research 10044301 and 10470009 (to N. A.) and 11670044 (to J. N.), and Grant-in-aid for Priority Areas 11170240 (to J. N.) from the Ministry of Education, Science and Culture, Japan.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

^¶ To whom all correspondence should be addressed: Dept. of Cellular and System Physiology, Graduate School of Medicine, Kyushu University, 3-1-1 Maidashi, Higashi-ku, Fukuoka 812-8582, Japan. Tel.: 81-92-642-6086; Fax: 81-92-642-6094; E-mail: akaike@mailserver.med. kyushu-u.ac.jp.

² O. Maeda and K. Hirate, unpublished data.

	ABBREVIATIONS

The abbreviations used are: AVP, arginine vasopressin; PI, phosphoinositide; CHO, Chinese hamster ovary; CaM, calmodulin; PKA, protein kinase A; PKC, protein kinase C; IBMX, 3-isobutyl-1-methylxanthine; W-7, N-(6-aminohexyl)-5-chloro-1-naphthalenesulfonamide; AVP(4-9), [pGlu⁴,Cyt⁶]AVP(4-9); DDAVP, 1-desamino-8-D-AVP; W-5, N-(6-aminohexyl)-1-naphthalensulfonamide; H-89, N-[2-(p-bromocinnamylamino)ethyl]-5-isoquinoline sulfonamide; Rp-cAMPS, Rp-adenosine-3',5'-cyclic monophosphorothioate; BAPTA-AM, 1,2-bis(2-amino-5-fluorophenoxy)ethane-N,N,N',N'-tetraacetic acid; KN-62, (s)-5-isoquinolinesulfonic acid; DDA, 2',5'-dideoxyadenosine; PMA, phorbol 12-myristate 13-acetate; IP₃, inositol-1,4,5-triphosphate; OAG, 1-oleoyl-2-acetylglycerol; db-cAMP, dibutyryl cyclic AMP; V_h, holding potential; I_Gly, glycine-induced current; IICR, IP₃-sensitive Ca²⁺ storage sites; CICR, calcium-sensitive intracellular Ca²⁺ storage sites; TFP, trifluoperazine; CPZ, chlorpromazine; NMDA, N-methyl-D-aspartate; PLC, phospholipase C; OT, oxytocin.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES