V_1a- and V₂-type vasopressin receptors mediate vasopressin-induced Ca²⁺ responses in isolated rat supraoptic neurones

Abstract

1. The pharmacological profile of receptors activated by vasopressin (AVP) in freshly dissociated supraoptic magnocellular neurones was investigated using specific V_1a- and V₂-type AVP receptor agonists and antagonists.

2. In 97 % of AVP-responding neurones (1–3000 nm) V_1a or V₂ receptor type agonists (F-180 and dDAVP, respectively) elicited dose-dependent [Ca²⁺]_i transients that were suppressed by removal of external Ca²⁺.

3. The [Ca²⁺]_i response induced by 1 μm F-180 or dDAVP was selectively blocked by 10 nm of V_1a and V₂ antagonists (SR 49059 and SR 121463A, respectively). The response to V_1a agonist was maintained in the presence of the V₂ antagonist, and the V₂ agonist-induced response persisted in the presence of the V_1a antagonist.

4. The [Ca²⁺]_i response induced by 1 μm AVP was partially (61 %) blocked by 10 nm SR 121463A. This blockade was increased by a further 31 % with the addition of 10 nm SR 49059. Similarly, the AVP-induced response was partially (47 %) decreased by SR 49059, and a further inhibition of 33 % was achieved in the presence of SR 121463A.

5. We demonstrate that AVP acts on the magnocellular neurones via two distinct types of AVP receptors that exhibit the pharmacological profiles of V_1a and V₂ types. However, since V₂ receptor mRNA is not expressed in the supraoptic nucleus (SON), and since V_1b receptor transcripts are observed in the SON, we propose that the V₂ receptor agonist and antagonist act on a ‘V₂-like’ receptor or a new type of AVP receptor that remains to be elucidated. The possibility that V₂ ligands act on the V_1b receptor cannot be excluded.

Vasopressin (AVP) magnocellular neurones in the hypothalamus exhibit phasic electrical activity that depends on intrinsic membrane properties and is controlled by extrinsic factors such as plasma osmolarity, blood volume and pressure (see review by Armstrong, 1995), and also by AVP itself. Indeed, a somato-dendritic release of AVP in the extracellular space of hypothalamic nuclei has been demonstrated by several studies (Leng & Mason, 1982; Di Scala-Guenot et al. 1987; Pow & Morris, 1989; Landgraf & Ludwig, 1991). Recently, we showed that AVP favours the expression of a specific phasic activity known to optimize the systemic release of AVP (Gouzènes et al. 1998a). This autocontrol was complex and depended on the initial state of activation of AVP neurones: AVP being excitatory for quasi-silent neurones or inhibitory for highly active neurones. These opposite effects may suggest the involvement of different types of AVP receptors in the control of the phasic pattern. Indeed, results from extracellular recordings performed in vivo or from rat hypothalamic slice preparation suggests that AVP-induced changes in firing rate of AVP neurones are mediated through the V₁ receptor (Inenaga & Yamashita, 1986; Dayanithi et al. 1995; Ludwig & Leng, 1997; Gouzènes et al. 1998b). On the other hand, in guinea-pig hypothalamic slices, AVP applied to neurones from the supraoptic nucleus (SON) inhibited spike generation by membrane depolarization (Abe et al. 1983). This inhibitory effect of AVP was mimicked by cAMP and enhanced by phosphodiesterase inhibitors (Abe et al. 1983), which suggests the involvement of V₂ receptors. Microspectrofluorimetric studies on dissociated supraoptic magnocellular neurones have shown that AVP induces an increase in [Ca²⁺]_i (Dayanithi et al. 1996) that requires an influx of external Ca²⁺ via voltage-dependent channels (Sabatier et al. 1997). The [Ca²⁺]_i response results from the activation of multiple intracellular transduction mechanisms involving the phospholipase C and adenylyl cyclase pathways (Sabatier et al. 1998), i.e. the classical second messenger systems activated by V₁- and V₂-type receptors, respectively (Orloff & Handler, 1967; Stephens & Logan, 1986). The mediation of AVP-induced [Ca²⁺]_i response by V_1a receptors has been previously suggested by the use of a specific V_1a receptor antagonist, SR 49059 (Dayanithi et al. 1995, 1996). In addition, the expression of V_1a and V_1b but not V₂ mRNAs in magnocellular neurones has recently been reported (Hurbin et al. 1998).

In the present study, we have investigated the physiological significance of AVP receptors on freshly dissociated neurones from SON by evaluating the [Ca²⁺]_i responses evoked by specific V_1a and V₂ receptor agonists, F-180 and dDAVP, respectively. A preliminary report of this work has appeared in abstract form (Gouzènes et al. 1998c).

METHODS

Preparation of dissociated magnocellular neurones

For each experiment, two adult male Wistar rats (Iffa-Credo, France; 150–300 g) were killed by decapitation with a guillotine following the guidelines laid down by the French/European ethical committee (licence no. 005039/04-08-1992/G. Dayanithi), and supraoptic neurones were isolated as previously described (Lambert et al. 1994) with modifications (Sabatier et al. 1997). Briefly, the isolated SONs were incubated for 45 min at 25°C in oxygenated Locke buffer supplemented with deoxyribonuclease I (0.5 mg ml⁻¹), proteases X and XIV (1 mg ml⁻¹ each). All enzymes and other standard chemicals were purchased from Sigma (USA). The Locke buffer contained (mM): 140 NaCl, 5 KCl, 1.2 KH₂PO₄, 1.2 MgSO₄, 1.8 CaCl₂, 10 glucose and 10 Hepes; pH 7.2, and the osmolarity was 295–300 mosmol l⁻¹. Tissue pieces were then rinsed with Locke buffer and submitted to mechanical trituration. Dissociated cells were plated onto glass coverslips and incubated for 1 h at 37°C with Locke buffer supplemented with 1.5 μm fura-2 AM (acetoxymethyl ester form of fura-2) plus 0.1 % (w/v) Pluronic F-127 (Molecular Probes Inc., Eugene, OR, USA). [Ca²⁺]_i measurements were performed with the fast fluorescence photometer system (FFP; Zeiss, Oberkochen, Germany; for details, see Dayanithi et al. 1996). Only cells with dendritic processes and a soma diameter more than 12 μm were used, since these large neurones have been previously demonstrated to contain either AVP or OT (Oliet & Bourque, 1992; Lambert et al. 1994). Control and test solutions were applied in the vicinity of the recorded neurone (1 mm) using a gravity-driven perfusion system with a flow rate of 100 μl min⁻¹. Complete solution change around the neurone was achieved within 2 s. The free Ca²⁺/EGTA buffer contained (mM): 0.1 EGTA, 140 NaCl, 5 KCl, 1.2 KH₂PO₄, 1.2 MgSO₄, 10 glucose and 10 Hepes; pH 7.2, and the osmolarity was 295–300 mosmol l⁻¹. The Ca²⁺ concentration in the Ca²⁺-free EGTA buffer was approximately 100 nm, which corresponds to the resting [Ca²⁺]_i typically observed in neurones.

Test compounds

The V₂ receptor agonist dDAVP (1-deamino-8-D-AVP), the V_1a receptor agonist [Phe², Orn⁸]vasotocin and the V₂ receptor antagonist desGlyNH₂-d(CH₂)₅-[D-Ile², Ile⁴] AVP were generously given by Dr C. Barberis (INSERM U-469, Centre de Pharmacologie-Endocrinologie, Montpellier, France). The V_1a receptor antagonist, SR 49059 ((2S)1[(2R, 3S)-(5-chloro-3-(2-chlorophenyl)-1-(3, 4-dimethoxybenzene-sulphonyl))-3-hydroxy-2, 3-dihydro-1H-indole-2-carbonyl]-pyrrolidine-2-carboxamide), and V₂ receptor antagonist SR 121463A (1-[4-(N-tert-butylcarbamoyl) −2-methoxybenzene sulphonyl]-5-ethoxy-3-spiro-[4-(2-morpholino ethoxy)cyclohexane]indol-2-one, fumarate; equatorial isomer) were gifts from Sanofi Recherche (Centre de Toulouse, Toulouse, France). The V_1a agonist F-180 ({underline 1}Hmp-Phe-Ile-Hgn-Asn-Cys{underline}-Pro-Dab(Abu)-Gly-NH₂) was kindly supplied by Dr J. L. Junien (Ferring France, Gentilly, France). AVP and OT were purchased from Boehringer Mannheim (Meylan, France). Concentrated stock solutions of substances were prepared in DMSO (SR 49059 and SR 121463A) or distilled water (dDAVP) and stored at −20°C.

Data analysis

In the figures, the [Ca²⁺]_i traces represent the ratio values between two fluorescence wavelengths (A: 340 nm/B: 380 nm). The dose- response relationship (Fig. 3) was constructed with peak amplitude values all normalized to that of the response induced by 100 nm agonist. The dDAVP dose-response curve was fitted with a single Hill equation:

Figure 3. Dose-dependent effect of V_1a and V₂ agonists.

The [Ca²⁺]_i traces in A and B illustrate the responses to increasing concentrations of F-180 and dDAVP, respectively. The graphs in C and D show the dose-response relationships of the mean [Ca²⁺]_i transients induced by F-180 (n = 24) and dDAVP (n = 22), respectively. For each cell tested, values of peak amplitude responses were expressed as a percentage of that induced by 100 nm agonist (normalized peak amplitude, %).

The dose-response curve for F-180 was fitted with a double Hill equation:

where Y is the normalized peak amplitude (%), Y_max is the maximum normalized peak amplitude (%), EC₅₀ is the concentration required to obtained the half-maximum peak amplitude (nm), X is the concentration of the agonist (nm) and n_H is the Hill coefficient.

Fits were made using Origin software (Microcal, Northampton, USA). All values are expressed as means ±s.e.m., and the results were analysed using Student's paired t test.

RESULTS

This study was performed on a total of 206 AVP-responding neurones (from 40 preparations) that displayed stable resting [Ca²⁺]_i levels. Those neurones initially displaying spontaneous [Ca²⁺]_i oscillations were discarded. Consistent with our previous study (Dayanithi et al. 1996), we also observed that neurones responding to OT (n = 13) were insensitive to AVP (1 μm), F-180 (V_1a receptor agonist, 1 μm) and dDAVP (V₂ receptor agonist, 1 μm).

Characteristics of the agonist-induced [Ca²⁺]_i increase

The [Ca²⁺]_i responses to F-180 and dDAVP (1 μm) were compared with those induced by AVP (1 μm) in 58 neurones (Fig. 1). Most of these AVP-responding neurones (72 %) exhibited [Ca²⁺]_i responses to both F-180 and dDAVP, 16 % responded to F-180 only, 9 % responded to dDAVP only, and 3 % responded neither to F-180 nor to dDAVP. Responses to F-180 and AVP displayed a short latency (8.4 ± 0.4 s and 7.4 ± 0.5 s, n = 51), a comparable amplitude (peak [Ca²⁺]_i rise evoked by F-180 represented 119 ± 17 % of that induced by AVP; n = 51), and similar profiles, although individual variations were noted. Conversely, the response to dDAVP markedly differed from that observed in response to AVP (see Fig. 1), with a significantly longer latency (29.3 ± 7.3 vs. 7.3 ± 0.5 s, P < 0.005, n = 47), a smaller peak amplitude (48 ± 6 % of the AVP response; n = 47) and more varied profiles (see examples in Fig. 4).

Figure 1. The traces represent the [Ca²⁺]_i responses induced by F-180 (V_1a agonist) and dDAVP (V₂ agonist) in a selected neurone responding to AVP.

The three substances were tested at 1 μm. Traces are expressed as the ratio between two fluorescence wavelengths: 340 nm (A)/380 nm (B).

Figure 4. Specificity of the responses to V_1a and V₂ receptor agonists.

The [Ca²⁺]_i transient elicited by F-180 (1 μm) was totally blocked by 10 nm of the specific V_1a receptor antagonist (SR 49059, A), but was maintained in the presence of 10 nm V₂ receptor antagonist (SR 121463A, B). Similarly, response to dDAVP (1 μm) was abolished after pre-incubation with 10 nm SR 121463A (C) but persisted in the presence of 10 nm SR 49059 (D). A and B show the reversibility of the blockade induced by each antagonist.

Twelve AVP-responsive neurones were subjected to four successive applications of F-180 (1 μm given every 2–3 min; Fig. 2A). In all neurones tested, F-180 induced reproducible [Ca²⁺]_i responses. Similarly, the [Ca²⁺]_i responses to successive applications of AVP (1 μm given every 2–3 min) were generally stable in each of the five neurones tested (Fig. 2B). In contrast, the responses to successive applications of dDAVP (1 μm, given every 2–3 min) showed a strong tachyphylaxis, as the third application of dDAVP induced a response with a peak amplitude reaching only 21 ± 9 % of that to the first test (n = 10; Fig. 2C). However, tachyphylaxis could be prevented by increasing the delay between applications to 10 min (n = 6; see Fig. 2D). Consequently, all subsequent experiments with dDAVP were performed with one application every 10 min.

Figure 2. The [Ca²⁺]_i responses to repeated applications of 1 μm F-180 (A), AVP (B) and dDAVP (C and D).

In each case, the upper Ca²⁺ traces show individual examples of responses. The lower histograms correspond to the mean values of the responses obtained in several neurones. The peak amplitude of the successive responses were expressed as a percentage of the first peak amplitude (% normalized peak amplitude). A and B, reproducibility of the responses to successive application of F-180 (n = 15 neurones) and AVP (n = 5 neurones). C, tachyphylaxis obtained with dDAVP applications every 2 min (n = 10 neurones). D, a longer interval (8–10 min) between successive applications of dDAVP prevented tachyphylaxis (n = 6 and 3 for the second and the third applications, respectively).

Dose dependence of the responses

To evaluate the dose-response relationships for the two agonists (F-180 and dDAVP), the neurones were systematically tested with each compound at 100 nm. Two or three other concentrations (from 1 nm to 3 μm) were applied to the same neurone in a random order. Threshold concentration was 3 nm for F-180 and 30 nm for dDAVP. The dose-response curve obtained with F-180 (Fig. 3B) was best fitted with the sum of two Hill equations (EC₅₀= 2.1 ± 0.5 and 184 ± 18 nm, n_H= 2.64 ± 1.24 and 1.49 ± 0.21, respectively), whereas the dose-response curve with dDAVP (Fig. 3D) presented a single slope only (EC₅₀= 95 ± 13 nm and n_H= 1.15 ± 0.15).

Specificity of the [Ca²⁺]_i responses

The specificity of the two agonists was studied using specific V_1a and V₂ receptor antagonists (Fig. 4). The [Ca²⁺]_i response to F-180 was inhibited by 97 ± 2 % by 10 nm SR 49059, a specific V_1a receptor antagonist (n = 6; Fig. 4A). SR 49059 did not affect the resting [Ca²⁺]_i level and inhibition of the response to F-180 recovered after wash-out of the antagonist. Similarly, the responses to 100 nm[Phe², Orn⁸]vasotocin, another V_1a receptor agonist, were suppressed by 10 nm SR 49059 (by 97 ± 1 %; n = 6; figure not shown). In contrast, the mean amplitude of the response induced by F-180 was unaffected by the presence of 10 nm SR 121463A, a specific V₂ receptor antagonist, (105 ± 5 % of respective control response; n = 10; Fig. 4B; P = 0.59). The same concentration of SR 121463A, reversibly inhibited (by 91 ± 5 %) the response to dDAVP (n = 9; Fig. 4C; P < 0.001). The mean amplitude of the [Ca²⁺]_i rise induced by dDAVP was unchanged in the presence of the V_1a antagonist, SR 49059 (127 ± 15 % of the respective control response, n = 5; Fig. 4D; P = 0.09).

The effects of the V_1a and V₂ receptor antagonists were also tested on the [Ca²⁺]_i rise induced by 1 μm AVP (Fig. 5). The V₂ receptor antagonist SR 121463A (10 nm) and desGlyNH₂-d(CH₂)₅-[D-Ile², Ile⁴] AVP (100 nm) reduced the response to AVP by 61 ± 6 % (n = 9; see example in Fig. 5A; P < 0.001) and by 53 ± 15 % (n = 4; figure not shown), respectively. A near complete inhibition (92 ± 3 %; n = 9; P < 0.001) of the AVP response was obtained by combined application of SR 49059 and SR 121463A (see example in Fig. 5A). Similarly, the response to AVP was reduced by 46 ± 7 % (n = 5) with SR 49059 given alone, and further decreased up to 79 ± 4 % (n = 5) by subsequent addition of SR 121463A (see example in Fig. 5B).

Figure 5. Simultaneous activation of V_1a- and V₂-type receptors by AVP.

Pre-incubation with 10 nm SR 49059 or SR 121463A (middle panels in A and B) partially inhibited the [Ca²⁺]_i response normally induced by AVP (left panels in A and B). At the same dose, each antagonist was shown to prevent the [Ca²⁺]_i transients elicited by the respective agonist (see Fig. 4). Blockade of AVP response was nearly complete when the two antagonists were pre-incubated simultaneously (right panels in A and B).

Origin of the Ca²⁺ rise induced by the agonists

The responses induced by V_1a or V₂ receptor agonists were also tested in Ca²⁺-free EGTA buffer to investigate the dependence upon external Ca²⁺ (Fig. 6A and B). The [Ca²⁺]_i responses induced by 1 μm F-180 or dDAVP were completely abolished in a Ca²⁺-free EGTA buffer (n = 10 and 4, respectively), suggesting that extracellular Ca²⁺ is necessary for the action of either agonist.

Figure 6. Block of [Ca²⁺]_i responses to V_1a and V₂ receptor agonists by suppression of extracellular Ca²⁺.

In A and B, the left panel represents the [Ca²⁺]_i response induced by 1 μm F-180 and dDAVP, respectively, in the presence of normal Locke buffer containing 1.8 mM external Ca²⁺. Pre-incubation in a Ca²⁺-free EGTA buffer (see Methods) for 10 min abolished the response to each agonist (right panels).

DISCUSSION

In this study, the use of specific agonists and antagonists allowed us to define the pharmacological properties of AVP receptors expressed by AVP-responding neurones of the SON (i.e. vasopressinergic neurones, see Dayanithi et al. 1996). In most of these neurones, both the V_1a and V₂ receptor agonists, F-180 and dDAVP, induced dose-dependent rises in [Ca²⁺]_i that were totally abolished in a Ca²⁺-free EGTA buffer and inhibited by their respective selective antagonist. The response to AVP was decreased by both V_1a and V₂ receptor antagonists, suggesting the simultaneous activation of V_1a- and V₂-type receptors by AVP. Because AVP was previously shown to induce an increase in [Ca²⁺]_i in AVP-sensitive neurones (Dayanithi et al. 1996), and because F-180 and dDAVP elicited responses only in AVP-sensitive neurones, these agonists are most likely to act on AVP neurones.

Characteristics of the responses to V_1a and V₂ receptor agonists

The [Ca²⁺]_i responses to the V_1a receptor agonist, F-180, and to AVP were similar, i.e. the responses occurred with a short latency and did not desensitize. The reproducibility of the AVP responses could be due to the use of a rapid and short duration of application of the peptide rather than static incubation over long periods, which might have caused desensitization as previously reported (Dayanithi et al. 1996). The dose-response relationship to F-180 was best described by a double Hill function, which suggests the activation of two receptor types with different affinities. Such a hypothesis is plausible since the existence of low- and high-affinity states for the V_1a receptor has been demonstrated in smooth muscle cells (Stassen et al. 1987; Gopalakrishnan et al. 1991) and in the liver (Gopalakrishnan et al. 1988). The biphasic dose-response to F-180 could result from complex interactions between different second messenger pathways (see Sabatier et al. 1998), which could modulate the activation of Ca²⁺ channels. Indeed, it has been shown that the rise in [Ca²⁺]_i activates the neurospecific type I calmodulin-sensitive adenylyl cyclase, which in turn exerts a positive feedback regulation of Ca²⁺ channels by cAMP-dependent protein kinase (Choi et al. 1993). In the present work, [Ca²⁺]_i responses induced by F-180 were also abolished in a free Ca²⁺ buffer, suggesting the involvement of a Ca²⁺ influx but not excluding an additional participation of Ca²⁺ from internal stores, as has been previously demonstrated for AVP (Dayanithi et al. 1996; Sabatier et al. 1997). In cultured hippocampal (Brinton et al. 1994) and cortical (Son & Brinton, 1998) neurones, the absence of Ca²⁺ in the extracellular medium has been shown to abolish the rise in [Ca²⁺]_i induced by a V₁ vasopressin receptor agonist, but the inositol-1-phosphate formation persisted. These studies indicate that the absence of extracellular Ca²⁺ does not affect the binding characteristics of the V₁ vasopressin receptor. They further showed that the activation of phospholipase C, and thus the production of the phosphatidylinositols, was involved in the response to the V₁ receptor agonist, but was not sufficient to increase the [Ca²⁺]_i.

Conversely, [Ca²⁺]_i responses induced by the V₂ receptor agonist, dDAVP, were smaller and delayed compared with those elicited by either AVP or F-180, and the dose-response curve showed a single slope. Figure 3 clearly illustrates the fact that the latency of the responses to dDAVP and F-180 does not depend on the concentrations that induced a significant response (i.e. over 30 nm). Indeed, the latency to dDAVP response is similar from 30 nm (i.e. below the half-maximal dose) to 300 or 1000 nm (illustration not shown), and is longer than the latency to F-180 responses from 100 to 1000 nm. The threshold of the dDAVP response was 30 nm, i.e. much higher than the concentration (0.7 nm) required to elicit [Ca²⁺]_i changes in rat medullary collecting tubules (Champigneulle et al. 1993). The tachyphylaxis obtained with a short interval between applications of dDAVP seems to be directly linked to the agonist characteristics rather than the receptor itself, since AVP showed reproducible responses. In the present study, dDAVP triggered a rise in [Ca²⁺]_i that was abolished in a Ca²⁺-free buffer, as has been demonstrated in rat inner medullary collecting tubes (Ishikawa & Saito, 1990; Naruse, 1992). This result supports the activation of Ca²⁺ influx by V₂ receptors, although the possibility of an additional mobilization of Ca²⁺ from internal stores is not excluded, as is the case for AVP (Dayanithi et al. 1996). Indeed, mobilization of Ca²⁺ from intracellular stores by V₂-like receptors has been observed in a subpopulation of neurones from circumventricular organs (Jursak et al. 1995). Whatever the origin of the Ca²⁺, the transduction pathways responsible for the [Ca²⁺]_i rise induced by dDAVP in the present work are still unknown. The involvement of the cAMP pathway is an interesting issue, since typically the V₂ receptor has been reported to be linked to activation of adenylyl cyclase (see Zingg, 1996).

Which AVP receptors subtypes are present on magnocellular neurones?

Based on their pharmacological profile and transduction signals, three types of vasopressin receptors have been described in the periphery: V_1a- and V_1b-types, both coupled to the phospholipase C, and V₂-type, coupled to the adenylate cyclase (see Zingg, 1996). The V_1a receptor agonist, F-180, is a selective vasoconstrictor and appears to be the most specific ligand of the V_1a receptor available to date (Aurell et al. 1991; Bernadich et al. 1998). Similarly, SR 49059 is a potent, non-peptidic antagonist of rat and human vasopressin V_1a receptors (Serradeil-Le-Gal et al. 1993; Guillon et al. 1995). The dDAVP and SR 121463A are respectively considered as a specific agonist (Manning et al. 1976) and a selective antagonist of the V₂ receptor (Serradeil-Le-Gal et al. 1996).

The affinities of AVP for peripheral vasopressin receptors in the rat are: K_d= 1.7 nm for the V_1a receptor and K_d= 0.4 nm for the V₂ receptor. The affinities of the V₂ agonist dDAVP are: K_d= 250 nm for the V_1a receptor and K_d= 0.3 nm for the V₂ receptor. Concerning the V_1a agonist, F-180, the relative affinities studied in COS cells are: K_d= 5.8 nm for the V_1a receptor and K_d > 10 000 nm for the V₂ receptor. Therefore, the V_1a agonist presents a very good selectivity between V_1a and V₂ receptors (see Barberis & Tribollet, 1996).

The antagonists we used are very potent and selective. The affinities of the V_1a antagonist for peripheral receptors in the rat are: K_d= 2 nm for the V_1a receptor and K_d= 275 nm for the V₂ receptor. The affinities of V₂ antagonist SR 121463A are: K_d= 10 600 nm and K_d= 1.42 nm for V_1a and V₂ receptors, respectively (see Serradeil-Le-Gal et al. 1996). At the concentrations we used, the V₂ agonist (1 μm) is able to bind to V_1a receptors, but in our experiments the response to V₂ agonist is not affected by the V_1a antagonist (10 nm), while it is totally blocked by the V₂ antagonist (10 nm). However, we also have to consider that the relative affinities and even the nature of vasopressin receptors expressed in the central nervous system are not well known and remain to be clearly elucidated.

In the present work, 10 nm SR 49059 abolished the [Ca²⁺]_i response to F-180, but not that to dDAVP. Similarly, 10 nm SR 121463A completely suppressed the [Ca²⁺]_i response to dDAVP but not that to F-180. These results argue for the presence of two distinct types of AVP receptors on the cell body of the large majority of AVP-responding neurones. These two AVP receptor types appear to be simultaneously activated by AVP, since the selective V_1a and V₂ receptor antagonists show additivity in the blockade of the AVP-induced response. In agreement with these results, we recently demonstrated that the [Ca²⁺]_i rise induced by AVP in AVP magnocellular neurones involved both the phospholipase C and adenylyl cyclase intracellular pathways (Sabatier et al. 1998), the established transduction pathways of V_1a and V₂ receptors, respectively. Furthermore, AVP has been reported to induce a rise in cAMP levels in neurones from the SON, and cAMP to mimic the electrophysiological effect of AVP on magnocellular neurones (Abe et al. 1983).

Taken together, these results clearly suggest a possible expression of V_1a and V₂ receptors on AVP magnocellular neurones. However, a different picture arose from a recent study from our laboratory using RT-PCR, which showed the presence of V_1a and V_1b receptor mRNAs, but not the presence of the V₂ transcript, in the SON (Hurbin et al. 1998). By in situ hybridization, these authors further visualized V_1a and V_1b mRNAs in AVP magnocellular neurones. It is of interest to note that a similar observation was made by Vaccari and collaborators (Vaccari et al. 1998) that the supraoptic nucleus expresses V_1b transcripts. With regard to the V_1a receptor, these results fit with the present study. For the V_1b receptor, the lack of specific ligands limits its pharmacological characterization (Barberis & Tribollet, 1996). Nevertheless, since a residual AVP response persisted in the presence of both V_1a and V₂ receptor antagonists, we suggest the presence of another receptor type on AVP magnocellular neurones, which could be the V_1b receptor type. With regard to V₂ receptors, the results obtained with RT-nested PCR (Hurbin et al. 1998) are in apparent contradiction with the present pharmacological data. Similarly, the V₂-type receptor has not been identified in the brain by either autoradiography or immunocytochemistry (see Burbach et al. 1995), whereas several pharmacological studies attest for its central involvement in different functions such as the proconvulsive effect of AVP (Croiset & De Wied, 1997), analgesia (You et al. 1995), cardiovascular regulation (Lowes et al. 1993) and Ca²⁺ responses in circumventricular organs (Jursak et al. 1995). Thus the nature of the AVP receptor mimicking the V₂-type pharmacology or activating the cAMP pathway remains to be elucidated.

In conclusion, our results demonstrate that AVP-responding neurones express at least two distinct AVP receptors that have the pharmacological profile of V_1a and V₂ receptors. The existence of functional V_1a-type receptors is corroborated by the presence of V_1a receptor mRNA in AVP neurones, but this is not the case for the receptor exhibiting the pharmacological profile of the V₂ type. The nature of this V₂-like receptor has yet to be defined. There is no direct solution offered in this paper for the discrepancies between the data (absence of V₂ receptor mRNA vs. V₂ agonist-induced [Ca²⁺]_i response) obtained within our research group. However, we can clearly hypothesize that the V₂-agonist could act on a different type (V₂-like?) or a new type of receptor that may have a similar pharmacology to ‘classical’ V₂-type receptors. Nevertheless, it appears that AVP magnocellular neurones express multiple AVP receptor types that could account for the complex effects of AVP on the electrical activity of AVP neurones observed in vivo (Gouzènes et al. 1998a). The inhibitory effect of AVP seems to be mediated via the V_1a receptor, as revealed by the excitatory effect of the V_1a receptor antagonist on the phasic pattern of AVP neurones (Dayanithi et al. 1995; Ludwig & Leng, 1997; Gouzènes et al. 1998b). In contrast, Inenaga & Yamashita (1986) have suggested that the activation of V₁ receptors results mostly in neuronal excitation. However, our preliminary results suggest that the excitatory effect could be mediated by V₂-type receptors (Gouzènes et al. 1998b). A pharmacological characterization of all receptors activated by AVP expressed on AVP magnocellular neurones, as well as intracellular pathways activated by the different agonists, is under investigation. This will help in defining the exact nature of the receptor types activated by AVP on AVP magnocellular neurones.