Introduction

Top Abstract Introduction Methods Results Discussion References

Serotonin has been shown to be a potent vasoconstrictor of the vasculature. The contractile action of serotonin is mainly mediated by two distinct serotonergic receptor groups: the 5-HT₁ and 5-HT₂ receptors. The vascular 5-HT₂ receptors have been well characterized (Bradley et al., 1986; Feniuk and Humphrey, 1989), but the nature of vascular 5-HT₁ receptors remains the subject of ongoing investigation (Humphrey et al., 1988; Shimamoto et al., 1993; Movahedi et al., 1995).

5-HT₁ receptors have been divided into 5-HT_1A, 5-HT_1B, 5-HT_1D, 5-HT_1E and 5-HT_1F subtypes. Recent nomenclature changes (Hartig et al., 1996) are used as follows. The rat 5-HT_1B and human 5-HT_1D receptor subtypes are classified together as 5-HT_1B. The rat 5-HT_1B receptor subtype has a unique pharmacology (Hoyer, 1989) and is prefaced by the word "rat" when referring specifically to this receptor subtype. The human 5-HT_1D receptor subtype is referred to as 5-HT_1D. Based on both pharmacological and molecular biological (Northern) analyses, there is increasing evidence the 5-HT_1B (formerly 5-HT_1D) receptor subtype is the most common 5-HT₁ receptor subtype found in blood vessels (Hamel et al., 1993a, 1993b; Skingle et al., 1996). Thus, the term 5-HT_1B-like is used in the present study to refer to the vascular 5-HT_1B receptor subtype when it has been characterized by pharmacological but not molecular biological (Northern), analysis. Recently, the 5-HT_1E (Leonhardt et al., 1989) and 5-HT_1F (Adham et al., 1993) subtypes have been cloned. Sumatriptan has a high affinity for the 1A, 1B and 1D subtypes, all of which also mediate inhibition of adenylate cyclase (Hoyer, 1989). The former 5-HT_1C subtype, now reclassified 5-HT_2C (Hoyer et al., 1994), displays low affinity for sumatriptan and is coupled to metabolism of membrane phosphoinositides.

The pharmacological profile of vascular 5-HT_1B-like receptors generally resembles that of the neuronal and brain 5-HT_1B receptors. This is the case in rabbit basilar (Parsons and Whalley, 1989; Clark and Garland, 1993), renal (Choppin and O'Connor, 1994) and mesenteric arteries (Choppin and O'Connor, 1995; Yildiz and Tuncer,1995a); rabbit (Martin et al., 1991) and canine saphenous vein (Humphrey et al., 1988); and canine coronary artery (Cushing and Cohen, 1992). In addition, the second-messenger systems associated with the vascular 5-HT_1B-like receptors also resemble those of the brain 5-HT_1B receptors. For example, both brain 5-HT_1B and vascular 5-HT_1B-like receptors have been shown to mediate inhibition of adenylate cyclase (Waeber et al., 1990). Vascular 5-HT_1B receptors also mediate extracellular Ca⁺⁺ mobilization (Sumner et al., 1992; Ebersole et al., 1993). Interestingly, mediation of Ca⁺⁺ influx appears to be the main factor for the initiation of vasoconstriction in blood vessels, whereas the inhibition of adenylate cyclase may have little to do with vasoconstriction (Sumner et al., 1992).

Vascular 5-HT_1B-like receptors have been found in a limited number of blood vessels such as the bovine cerebral vessels (Hamel et al., 1993a, 1993b), rabbit (Parsons and Whalley, 1989), canine (Connor et al., 1989) and human (Parson et al., 1989) basilar arteries and canine saphenous vein (Humphrey et al., 1988). However, vascular 5-HT_1B-like receptors have been found in additional vessels under unusual experimental circumstances, namely, in vessels precontracted with one of several agonists such as histamine, angiotensin II or prostaglandin F₂ (Choppin and Connor, 1995; Yildiz and Tuncer, 1995b). In the present study, receptors that become enabled by precontraction are referred to as "silent" 5-HT_1B-like receptors.

Previous work in our laboratory (Movahedi et al., 1995) has shown that the rabbit ear artery possesses silent 5-HT_1B-like receptors that become enabled by precontractions with alpha adrenergic agonists. In the present study, evidence is presented showing that the "silent" 5-HT_1B-like receptors of rabbit ear artery resemble both the fully functional vascular 5-HT_1B-like receptors and the brain 5-HT_1B receptor subtype as described above. This was addressed using two methods of identification: (1) pharmacological receptor analysis using agonists and antagonists selective for the subtypes of the 5-HT₁ receptors and (2) analysis of the second-messenger systems involved using drugs active at the levels of G proteins and membrane-bound enzymes.

	Methods

Top Abstract Introduction Methods Results Discussion References

Male New Zealand White rabbits (2-3 kg) were killed by exposure to 100% CO₂ to produce deep anesthesia (Glen and Scott, 1973), followed by rapid decapitation. Central ear arteries were isolated and placed in Krebs' solution at room temperature, cleaned and cut into 3-mm rings. No effort was made to remove the endothelium because this cell layer was found not to contribute to or influence the amplified response to serotonin in the presence of PHE (Movahedi et al., 1995). These rings were mounted for the measurement of isometric contraction (Bevan and Osher, 1972) in tissue baths containing 30 ml of 95% O₂-5% CO₂ gassed Krebs-bicarbonate solution at 37°C. The composition of the Krebs' solution in millimoles per liter was NaCl 119.2, KCl 4.9, CaCl₂ 1.3, MgSO₄ 1.2, NaHCO₃ 25, glucose 11.1, ascorbic acid 0.114 and tetrasodium ethylenediamine tetraacetate 0.03. The rings were placed under 1.5 × g resting force, a value found in preliminary experiments to provide optimal active force development (i.e., the largest repeatable contraction to the standard stimulus of 68 mM KCl). After 60 min under resting force, baths were drained and refilled with Krebs' solution containing 68 mM KCl with equimolar reduction of Na⁺. When tissues achieved steady-state contraction, baths were drained and refilled three times over 5 min with normal Krebs' solution, and tissues were allowed to relax to base line. Resting force was adjusted as needed. Thirty minutes later, this potassium exposure was repeated, and the resulting contraction was expressed as 100% and used to normalize contractile responses to serotonin and other serotonergic agonists. Isometric contractions were recorded using Grass FT03C strain gauges (Grass, Quincy, MA) connected to a Maclab Electronic Data Acquisition System (Castle Hill, Australia).

Amplification studies. Amplification was defined as an increase in sensitivity, signified by a leftward shift of the CRC, for a serotonergic receptor agonist in the presence of a threshold concentration of an alpha-1 adrenergic receptor agonist. Alpha receptor agonist threshold concentration was that which produced a 0.2 to 0.4 g contraction, equal to 2% to 6% of the contraction elicited by 68 mM K⁺. The threshold concentration of PHE ranged between 30 and 60 nM. In a typical experiment, a threshold contraction was obtained by the addition of an alpha receptor agonist; subsequently, a serotonin CRC was obtained by cumulative addition in 0.5-log increments. Calcium-free Krebs' solution was prepared by leaving out the calcium. All competitive antagonists and some functional antagonists (e.g., nifedipine) were added to the tissue at least 30 min before the addition of agonist. In experiments in which nifedipine or forskolin was used, the concentration of PHE that induced a threshold contraction was raised to 600 to 900 nM. This was necessary to achieve contractions to PHE equivalent to those in the absence of these antagonists, 0.2 to 0.4 g.

Experiments conducted with forskolin and sodium nitropruside. Forskolin, a direct activator of adenylate cyclase, was used in two different experimental protocols. First, ear artery rings were contracted equivalently with either a submaximal concentration of the alpha-1 receptor agonist or a combination of a threshold concentration of the alpha-1 receptor agonist plus serotonin or 5-HT₁-selective agonist. When the tissues reached steady-state contraction, forskolin was added to the tissue baths in increasing concentration. As a control, separate rings of ear artery, contracted as above, received sodium nitropruside instead of forskolin. In the second protocol, a fixed concentration of forskolin was added to the bath 5 min before the initiation of the agonist CRC.

Pertussis toxin treatment. Rabbits received an i.v. injection of pertussis toxin (5 µg/kg b.wt.). The toxin was dissolved in 0.9% saline solution, and the control rabbits received the same volume of the vehicle. Animals were killed on the fourth day after injection, and responses of ear artery rings from the PTX- and vehicle-treated animals were compared in the same experiment.

Chemicals. The drugs used were forskolin, 5-HT, phenylephrine HCl, L-propranolol and sodium nitropruside (Sigma Chemical, St. Louis, MO); pertussis toxin (List Biochemical, Campbell, CA); 5-carboxamidotryptamine, sumatriptan and GR127935T (gifts from Glaxo Pharmaceutical, Stevenage, UK); and rauwolscine (Carl Roth, Karlsruhe, Germany).

Data analysis. The n values are given in each figure, indicating both the number of artery rings and animals used for each CRC; for example, n = 12/5 means that the given CRC is based on 12 artery rings from 5 rabbits. Statistical analysis was based on the number of experiments (animals). Contractile responses are expressed as mean ± SEM and EC₅₀ values were determined from regression analysis of the linear portions of the CRC. Geometric means of these values are expressed ± SEM. Using SuperANOVA statistical software (Abacus Concepts, Berkeley, CA) CRC values were compared by repeated-measures, two-way analysis of variance. Differences between individual points on different curves were analyzed by one-way analysis of variance followed by a post hoc Scheffé test. Differences were considered significant at P < .05.

	Results

Top Abstract Introduction Methods Results Discussion References

First-Messenger Experiments

The effects of the vascular 5-HT_1-like receptor-selective agonist sumatriptan (Humphrey et al., 1988) were assessed in the rabbit ear artery; results are shown in figure 1. In the absence of PHE, sumatriptan failed to induce vasoconstriction, whereas in the presence of a threshold concentration of PHE (30-60 nM), sumatriptan caused a strong, concentration-dependent contraction. L-Propranolol (1 µM) failed to displace this amplified sumatriptan curve (fig. 1A). However, 0.1 µM rauwolscine inhibited the PHE-amplified sumatriptan CRC, causing a depression of maximum response (fig. 1B). Neither rauwolscine nor L-propranolol had an inhibitory effect on the threshold response induced by PHE. GR127935, a selective 5-HT_1B antagonist (Skingle et al., 1996), inhibited the sumatriptan CRC obtained in the presence of PHE in a concentration-dependent manner (fig. 2). At the concentrations used, 1, 10 and 100 nM, GR127935 displayed no inhibition of alpha-1 adrenergic receptors (data not shown).

View larger version (14K): [in this window] [in a new window]   Fig. 1.   CRCs for sumatriptan alone (SUMA), in the presence of a threshold concentration of PHE (PHE/SUMA) and in the presence and absence of (A) 1 µM L-propranolol (PHE/SUMA/PROP) or (B) 0.1 µM rauwolscine (PHE/SUMA/RAUW B). Responses are presented as a percent of maximal response induced by KCl. A, n = 3/9, 100% = 4.2 g; B, n = 5/10, 100% = 4.4 g. *PHE/SUMA different from PHE/SUMA/RAUW, P < .05.

View larger version (22K): [in this window] [in a new window]   Fig. 2.   CRCs for the sumatriptan alone (SUMA), in the presence of a threshold concentration of PHE (PHE/SUMA), with and without 1, 10 and 100 nM GR127935, a selective 5-HT1D antagonist (GR 1 nM, GR 10 nM, GR 100 nM). Responses are presented as a percent of maximal response induced by KCl. n = 4/8, 100% = 3.6 g. *PHE/SUMA different from GR 1 nM, P < .05.

Second-Messenger Experiments

Mediation of increase in cytoplasmic Ca⁺⁺ concentration. Figure 3 displays the effect of Ca⁺⁺ deprivation on the PHE-amplified response of the rabbit ear artery to serotonin. Nifidepine (0.1 µM), a blocker of the voltage-sensitive Ca⁺⁺ channels, decreased the maximal response of the PHE-amplified serotonin curve by 55%, whereas the removal of the extracellular Ca⁺⁺ completely abolished the response. The concentration of PHE was increased in calcium-deprived tissues to achieve an equivalent degree of contraction as that obtained in the presence of calcium.

View larger version (18K): [in this window] [in a new window]   Fig. 3.   CRCs for the serotonin-induced contraction in the presence of a threshold concentration of PHE (PHE/5-HT), in the presence of 100 nM nifidipine (PHE/5-HT/NIF) and in a Ca++-free Krebs' solution. In all treatments, the concentration of PHE was increased until equivalent threshold contractions were achieved. Responses are presented as a percent of maximal response induced by KCl. n = 6/12, 100% = 3.8 g.

Inhibition of adenylate cyclase. Because the activation of 5-HT_1B receptors have been associated with inhibition of adenylate cyclase (Ebersole et al., 1993; Zgombick et al., 1993), the presence of such inhibition in the rabbit ear artery was assessed. Figure 4 shows representative electronic tracings of a series of nearly equivalent submaximal contractions, induced by either PHE or a combination of a threshold concentration of PHE plus a submaximal concentration of either serotonin or sumatriptan. The addition of low concentrations of forskolin (0.1-0.3 µM) rapidly relaxed the PHE-contracted rings of ear artery. However, rings contracted with a combination of either PHE/serotonin or PHE/sumatriptan were completely resistant to these low concentrations of forskolin and did not relax until 10-fold higher concentrations (1-3 µM) were present (see fig. 7 for serotonin pooled results). This was in contrast to the effect of sodium nitroprusside, which induced the relaxation of rings contracted with PHE or PHE/serotonin at the same concentrations (0.3-3.0 µM) and with identical relaxation rates (fig. 4B).

View larger version (18K): [in this window] [in a new window]   Fig. 4.   Computer-generated pen tracings of submaximal contractions induced by PHE (0.3 µM), PHE (0.03 µM) plus serotonin (0.03 µM) (PHE/5-HT) and PHE (0.03 µM) plus sumatriptan (0.01 µM) (PHE/SUMA). When the contracted ear artery rings reached a contraction steady state, increasing concentrations of either forskolin (FSK; A) or sodium nitrprusside (SNP; B) were added to induce vasorelaxation.

View larger version (27K): [in this window] [in a new window]   Fig. 7.   The forskolin-induced relaxation of rabbit ear artery rings contracted with either PHE or a combination of PHE plus serotonin (PHE/5-HT). Some ear artery rings were from rabbits pretreated with 5 µg/kg i.v. PTX. Responses are presented as a percent of maximal response induced by agonists in the absence of forskolin. n = 3/9, *P < .05 significant difference between PHE/5-HT and PHE/5-HT/PTX; **P < .05 between PHE/5-HT and all other treatments at the given concentration of forskolin.

View larger version (25K): [in this window] [in a new window]   Fig. 5.   The forskolin-induced relaxation of rabbit ear artery rings contracted with a combination of PHE plus sumatriptan (PHE/SUMA) in the absence (3 µM SUMA) and presence (10 µM SUMA) of 10 nM GR127935 (GR 10 nM). Responses are presented as a percent of maximal response induced by the combination of agonists in the same ear artery ring. Maximal contractile responses in the absence and presence of GR127935 were 3.01 and 2.27 g, respectively. n = 3/6, *P < .05.

Effect of PTX treatment. The inverse coupling of the 5-HT_1B receptors to adenylate cyclase has been shown to be mediated through a PTX-sensitive G protein (Zgombick et al., 1993). Thus, experiments were carried out to assess the effect of PTX treatment on both the PHE-amplified response of ear artery to serotonin and the forskolin-induced relaxation of that response. First, it was found that PTX treatment had no effect on the PHE CRC in ear artery (fig. 6A). Likewise, PTX pretreatment had no effect on the serotonin CRC obtained in the presence of a threshold concentration of PHE (fig. 6B). Second, ear artery rings contracted with PHE/serotonin were exposed to increasing concentrations of forskolin, and relaxation was measured. Control ear artery rings, contracted with a combination of PHE/serotonin, were resistant to the effects of forskolin, requiring higher forskolin concentrations to produce relaxation compared with ear artery rings contracted with PHE alone (fig. 7). In contrast, ear artery rings from PTX-treated rabbits relaxed significantly more than control rings to every concentration of forskolin studied (fig. 7).

View larger version (11K): [in this window] [in a new window]   Fig. 6.   CRCs for PHE- (A) and PHE plus serotonin-(B) induced contractions of rabbit ear artery rings from control vs. PTX-treated rabbits. Responses are presented as a percent of maximal response induced by KCl. n = 3/6.

Effect of forskolin exposure on PHE and PHE/serotonin contractile CRCs. Experiments were also carried out using forskolin as a physiological antagonist (i.e., vasoconstrictor agonist CRCs were obtained in the presence and absence of 1 µM forskolin). As shown in figure 8, forskolin caused a 100-fold rightward shift of the PHE CRC. In contrast, forskolin had almost no effect on the serotonin CRC obtained in ear arteries precontracted with a threshold concentration of PHE (i.e., the serotonin CRCs in the presence and absence of forskolin were superimposable except for a small but significant paradoxical forskolin-induced enhancement of the maximal response).

View larger version (12K): [in this window] [in a new window]   Fig. 8.   CRCs for vasoconstriction induced by PHE (A) and serotonin in the presence of a threshold concentration of PHE (B), in the presence and absence of 1 µM forskolin. Responses are presented as a percent of maximal response induced by KCl. n = 5/12, 100% = 3.4 g. *PHE/5-HT different from PHE/5-HT/FSK, P < .05.

	Discussion

Top Abstract Introduction Methods Results Discussion References

The present results confirm our previous finding (Movahedi et al., 1995) that the PHE-amplified serotonin-induced contraction of rabbit ear artery is mediated by 5-HT_1B-like receptors. Further pharmacological characterization of these vascular 5-HT_1B-like receptors in the present study indicated that these receptors resemble the 5-HT_1B subtype, as discussed below.

Because sumatriptan has been shown to be a selective 5-HT₁ receptor ligand (Peroutka, 1990), it was used as an agonist in our receptor identification studies. Binding studies have shown that sumatriptan displays a high affinity for the 5-HT_1A, rat 5-HT_1B, 5-HT_1B and 5-HT_1D receptor subtypes (Hoyer, 1989).

To differentiate among the 5-HT_1A, rat 5-HT_1B and the 5-HT_1B subtypes, we used 1 µM L-propranolol as a 1A/rat 1B antagonist (Hamon et al., 1990). The pK_B values of L-propranolol for the 5-HT_1A, rat 5-HT_1B and 5-HT_1B receptor subtypes are 6.8, 7.3 and 5.5 respectively (Hoyer, 1989). Rauwolscine (0.1 µM) was used as a 5-HT_1B antagonist (Hoyer and Schoeffter, 1988). The respective pK_B values of rauwolscine for these same receptors are 6.9, 5.3 and 7.7. Sumatriptan was used as a nonselective 5-HT₁ receptor agonist. Rauwolscine inhibited PHE-amplified sumatriptan-induced vasoconstriction, but L-propranolol was ineffective against this response. These results support the view that the amplified response of sumatriptan was mediated by vascular 5-HT_1B-like receptors that closely resemble the 5-HT_1B receptor subtype. Last, potent antagonism of the amplified sumatriptan response by GR127935, a selective 5-HT_1B antagonist (Skingle et al., 1996), further supported this suggestion. The 5-HT_1-like receptor of rabbit ear artery shall henceforth be referred to as a 5-HT_1B-like receptor.

The present study provided evidence that the 5-HT_1B-like receptor subtype in rabbit ear artery is coupled to both adenylate cyclase inhibition and calcium influx. This is in agreement with other studies of the 5-HT_1B receptor subtype in the canine saphenous vein (Sumner et al., 1992), bovine basilar artery (Ebersole et al., 1993) and murine fibroblasts (Zgombick et al., 1993).

The 5-HT_1B receptor-mediated inhibition of adenylate cyclase has been shown directly through the measurement of the effect of receptor activation on cAMP levels (Zgombick et al., 1993). Moreover, 5-HT_1B receptor activation was shown to reduce the effectiveness of forskolin, a direct activator of adenylate cyclase, in elevating cAMP (Zgombick et al., 1993). According to the authors' interpretation (Zgombick et al., 1993), inhibition of adenylate cyclase by 5-HT_1B receptor activation made this enzyme resistant to direct stimulation by forskolin. In the present study, an analogous approach was taken to test for 5-HT_1B-like receptor-mediated inhibition of adenylate cyclase in rabbit ear artery. The ability of forskolin to relax a precontracted artery was used as an indirect measure of forskolin-induced cAMP formation. Forskolin, in low concentrations, caused complete relaxation of PHE-contracted ear artery. However, these concentrations of forskolin were ineffective in ear artery rings contracted with PHE/serotonin, in which 3- to 10-fold higher concentrations caused only partial relaxation.

Vascular relaxation can be mediated by the products of both adenylate and guanylate cyclase enzymes. Sodium nitroprusside is an activator of guanylate cyclase. The relaxing effects of this agent on ear arteries contracted with either PHE or PHE/serotonin were compared, and no differences were found. This observation suggests that the resistance of PHE/serotonin-contracted arteries to forskolin-induced relaxation derives exclusively from 5-HT_1B-like receptor-mediated inhibition of adenylate cyclase.

Three additional experiments gave further support to the proposal that eliciting contractions with PHE plus serotonergic agonist in rabbit ear artery activated 5-HT_1B-like receptors and that these receptors mediated inhibition of adenylate cyclase. The first experiment was carried out to determine whether GR127935 blocked the inhibitory effect of PHE/sumatriptan on forskolin-mediated relaxation. The results shown in figure 5 demonstrate such blockade. Because GR127935 also blocked contraction (see fig. 2), these results show that both contraction to sumatriptan and inhibition of relaxation to forskolin appear to be mediated by the same or similar receptor subtype, the vascular 5-HT_1B-like receptor. In the second experiment, advantage was taken of the fact that 5-HT_1B receptors mediate the inhibition of adenylate cyclase through a PTX-sensitive G protein (Ebersole et al., 1993; Zgombick et al., 1993). Thus, ear artery rings from PTX-treated rabbits were contracted with PHE/serotonin and were found to be much more sensitive to the relaxing effects of forskolin compared with control ear arteries; this is consistent with a PTX-mediated uncoupling of the 5-HT_1B-like receptors from adenylate cyclase.

The first- and second-messenger work described above supports the view that contraction of ear artery rings with PHE/serotonergic agonist activated 5-HT_1B-like receptors and that these receptors mediated inhibition of adenylate cyclase through a PTX-sensitive G protein. We then explored whether this receptor/second-messenger pathway mediated the ear artery contraction to PHE/serotonergic receptor agonist. This was done using artery rings from control and PTX-treated rabbits. It was found that PTX treatment had no effect on ear artery contractions to either PHE or PHE/serotonergic agonist. Thus, we conclude that the 5-HT_1B-like receptor-mediated inhibition of adenylate cyclase neither mediated nor influenced the amplified ear artery contraction to PHE/serotonergic receptor agonist. This leads us to suggest that it was calcium influx, the other second-messenger pathway associated with the 5-HT_1B-like receptor of rabbit ear artery, that mediated the amplified response.

Hypothetically, for inhibition of adenylate cyclase to have contributed to contraction, this enzyme would have had to be constitutively active, producing the vasodilator cAMP. Inhibition of adenylate cyclase by 5-HT_1B-like receptor activation would have lowered cAMP, removing a vasodilator mechanism and allowing contraction to occur. We propose that adenylate cyclase was not active in the ear artery and that cellular cAMP was very low. In this setting, inhibition of adenylate cyclase would have had no effect.

It was found in the present study that the ear artery contraction to PHE/serotonin was completely dependent on the availability of extracellular calcium. Thus, we propose that the 5-HT_1B-like receptors of rabbit ear artery are coupled to two second-messenger pathways: directly to calcium entry and inversely to adenylate cyclase. This is in agreement with the earlier findings of Ebersole et al. (1993) in cultured bovine basilar artery cells. Hypothetically, these two pathways could be complementary, in vivo, as follows. It is conceivable that the ear artery could be exposed to both serotonin and norepinephrine, leading to contraction mediated by 5-HT_1B-like receptor-induced calcium entry. Any activation of beta adrenoceptors by norepinephrine could oppose this contraction by activating adenylate cyclase and elevating cAMP. However, this would be prevented by the 5-HT_1B-like receptor-mediated inhibition of adenylate cyclase, making this enzyme resistant to beta adrenoceptor-mediated activation.

When no precontraction is used in in vitro experiments, the 5-HT_1B-like receptor subtype has been found to exist in a fully functional state in only a few blood vessels, such as the cerebral arteries of several species (Parsons and Whalley, 1989), canine saphenous vein (Humphrey et al., 1988) and canine coronary arteries (Cushing and Cohen, 1992). However, several recent studies have reported that precontraction with a variety of agents yields an enhanced contractile response mediated by previously inactive 5-HT_1B-like receptors in rabbit iliac (Yildiz and Tuncer, 1995b), mesenteric (Choppin and O'Connor, 1995) and renal (Choppin and O'Connor, 1994) arteries and canine mesenteric artery (Shimamoto et al., 1993).

Yildiz and Tuncer (1995b) found that the rabbit iliac artery failed to contract to sumatriptan unless the vessel was precontracted with an agent such as prostaglandin F₂, histamine, PHE or angiotensin II. These authors proposed that the contraction to sumatriptan depended on the inverse coupling of the 5-HT_1B-like receptor to adenylate cyclase according to the following scheme. The agents used to precontract the iliac artery are coupled to phospholipase C and, therefore, elevate cytosolic calcium via inositol-1,4,5-triphosphate-induced calcium release. In this setting, sumatriptan, activating 5-HT_1B-like receptors, would inhibit adenylate cyclase and reduce cAMP formation; in turn, this would allow cytosolic calcium levels to increase, yielding a contraction. Although this proposal is plausible, it is not consistent with the present results in rabbit ear artery. We found evidence that although PTX treatment uncoupled the 5-HT_1B-like receptor from adenylate cyclase, this treatment had no effect on the serotonin-induced contraction in PHE-precontracted artery rings. Further experiments are required to clarify what role, if any, the adenylate cyclase second-messenger pathway has in the response of precontracted iliac and other arteries to serotonergic agonists.

If it is correct to rule out a role for adenylate cyclase in the contractile response of rabbit ear artery to PHE/serotonin, this implicates the other second-messenger pathway for the 5-HT_1B-like receptor suggested by the present results, namely, calcium entry. The question that remains is why the 5-HT_1B-like receptor and this pathway appear inoperative unless alpha adrenoceptors are simultaneously activated. It is possible that alpha adrenoceptors themselves are closely associated with 5-HT_1B-like receptors and, when activated, exert an allosteric influence on 5-HT_1B-like receptors. Alternatively, there may be a convergence of one or more second-messenger steps for the two receptors. In this case, alpha adrenoceptor activation may amplify a critical second-messenger step for the 5-HT_1B-like receptor, enabling this receptor. This second explanation seems more likely because many classes of vasoconstrictor agents have been shown to uncover previously inactive 5-HT_1B-like receptors (Shimamoto et al., 1993; Choppin and O'Connor, 1994, 1995).