Homeostasis of glucose in the rainbow trout (Oncorhynchus mykiss Walbaum): the role of serotonin

Studies performed in rodents suggest that 5-hydroxytryptamine (5-HT; serotonin) plays a role in the regulation of blood glucose. Thus, it has been reported that in rodents the injection of 5-HT or 5-HT receptor agonists results in hyperglycaemia (Chaouloff and Jeanrenaud, 1987; Yamada et al., 1995; Sugimoto et al., 1996a; Sugimoto et al., 1996b; Yamada et al., 1997). The 5-HT receptors that appear to mediate this hyperglycaemic action are the central 5-HT1A subtype (Uvnäs-Moberg et al., 1996; Sugimoto et al., 2001) and/or the central or peripheral 5-HT2 (Chaouloff et al., 1990b; Yamada et al., 1995; Yamada et al., 1997; Sugimoto et al., 1996a; Sugimoto et al., 1996b). However, the exact nature of the role that 5-HT plays in controlling blood glucose levels has still to be determined. Although some authors report that the effect of 5-HT appears to be related to changes in the plasma level of insulin (Chaouloff and Jeanrenaud, 1987; Uvnäs-Moberg et al., 1996; Yamada et al., 1998), others suggest that the effect is insulin independent (Chaouloff and Jeanrenaud, 1987; Yamada et al., 1997). Furthermore, according to some authors, it is the brain 5-HT that is involved in hyperglycaemia (Sugimoto et al., 1996a; Chaouloff et al., 1990a; Chaouloff et al., 1992), whereas others state that it is the peripheral 5-HT that is involved (Yamada et al., 1995; Sugimoto et al., 1996b; Yamada et al., 1997; Chaouloff et al., 1990b). Notwithstanding, the hyperglycaemia associated with activation of 5-HT receptors, whether central (5-HT1A and/or 5-HT2) or peripheral (5-HT2), is always associated with release of adrenaline (Yamada et al., 1995; Sugimoto et al., 1996b; Chaouloff et al., 1990b).

Although the rainbow trout ingests relatively small amounts of carbohydrates compared with other species and possesses a limited capacity to utilize dietary carbohydrate for energy purposes, the use of glucose as an oxidative substrate is essential for the functioning of the brain, and therefore the fish need a continuous supply of glucose for the blood (Aldegunde et al., 2000; Soengas and Aldegunde, 2002). Despite this, the mechanisms that control glucose homeostasis in fish have not been clarified, although there is some evidence that catecholamines (adrenaline and noradrenaline) induce a marked hyperglycaemic effect (Reid et al., 1998; Iwama et al., 2004), in a similar way to that observed in other vertebrates. In the rainbow trout, secretion of catecholamines is largely mediated by the release of acetylcholine from cholinergic preganglionic fibres of the sympathetic nerves, and subsequent activation of the nicotinic and muscarinic receptors present in the chromaffin cells (Montpetit and Perry, 1999). However, and as in other vertebrates, other mechanisms mediated by non-cholinergic neurotransmitters also modulate the secretion of catecholamines. It has been observed that 5-HT may induce the release of catecholamines from the chromaffin cells in both rainbow trout (Reid et al., 1998; Fritsche et al., 1993) and in the Atlantic hagfish (Bernier and Perry, 1996).

At present, nothing is known about the possible role of 5-HT in glucose homeostasis in fish. However, we can take the following into consideration: (1) the existence of serotoninergic modulation of catecholamine release; (2) in fish, 5-HT is involved in the physiological and behavioural responses to situations of stress associated with hyperglycaemia (Blanchard et al., 1993; Winberg and Nilsson, 1993); and (3) the abundance of peripheral 5-HT in the rainbow trout (Caamaño-Tubio et al., 2007). All of this appears to indicate that 5-HT may play an important role in the modulation of glucose homeostasis in fish. The purpose of the present study was therefore to investigate the effect of central and peripheral 5-HT on glucose homeostasis in the rainbow trout.

Immature rainbow trout (Oncorhynchus mykiss; 100±40 g) were obtained from a commercial trout farm (Soutorredondo, Noia, Spain). Fish were acclimated in 200 litre tanks for 3 weeks in running dechlorinated tap water (temperature 14±0.5°C, pH 6.3±0.1) with continuous aeration, and maintained under a 12 h:12 h light:dark photoperiod (lights on at 8:00 h). Fish were fed once daily, in the morning (at 12:00 h), with commercial dry pellets (ration equivalent to 1.5% of body mass per day), and were not fed in the 24 h prior to experiments. All experiments were carried out in the morning to avoid possible effects of circadian variations. Fish were anaesthetised before handling, injection or sampling, with 50 mg l⁻¹ tricaine methanesulphonate (MS-222; Sigma, Sigma Chemical Co., St Louis, USA) buffered to pH 7.4 with sodium bicarbonate. The anaesthetic was added directly to the tank where the fish were held, as this type of administration efficiently reduces the release of catecholamines induced by handling stress in the rainbow trout (Caamaño-Tubio et al., 2010). Replicate (triplicate or more) tanks were established for each experiment.

Serotonin (5-hydroxytryptamine creatinine sulphate), d-fenfluramine (d-FF; an indirect serotonergic agonist) and WAY 100635 maleate (WAY; a 5-HT1A receptor antagonist), were purchased from Sigma Chemical Co (St Louis, USA). Mianserin hydrochloride (MIAN; a 5-HT2 receptor antagonist), 5-carboxamidotryptamine maleate (5-CT; a 5-HT1 receptor agonist), 1-(3-trifluromethyphenyl) piperazine hydrochloride (TFMPP; a 5-HT1B receptor agonist), pimozide (PIM; a 5-HT7 receptor antagonist), α-methyl-5-hydroxytryptamine maleate (α-m-5-HT; a 5-HT2 receptor agonist), BRL 15572 hydrochloride (BRL; a 5-HT1D receptor antagonist) and SB 204741 (SB; a 5-HT2B receptor antagonist), were obtained from Tocris Cookson Ltd (Bristol, UK).

Drugs (5-HT and 5-HT agonists and antagonists) were administered by intraperitoneal (i.p.) injection in 125 μl of saline (SAL; 0.6% w/v NaCl), except for PIM, BRL and SB, which were dissolved in a vehicle (VEH) of propylenglycol and dimethylsuphoxide (PPG and DMSO) in proportions of 1:6, 1:9 and 1:2, respectively. The intracerebroventricular (i.c.v.) administration of 5-HT (in 2 μl of saline per fish, 0.6% w/v NaCl) was as previously described (Aldegunde and Mancebo, 2006). Briefly, the i.c.v. administration was carried out with a 10 μl Hamilton microsyringe. On the day of the experiment, the needle was introduced into the third ventricle at the level of the preoptic nuclei, just at the level of the anterior point of union of the lobes of the optic tectum; the microsyringe was held in place with a micromanipulator while the 2 μl dose was infused.

One series of experiments (dose–response experiments) was designed to investigate the effect of the i.p. administration of different doses of 5-HT (0.5, 1, 2, 5 and 10 mg kg⁻¹) 150 min after their administration (N=8 fish per dose). A second series of experiments (time–response experiments) examined the effect of the administration of a dose of 5-HT (2 mg kg⁻¹), with samples taken at 0, 50, 100, 150 and 200 min after administration (N=6–7 fish each time). The plasma levels of adrenaline (A) and noradrenaline (NA) were also determined at 0, 100, 150 and 200 min after administration of a dose of 5-HT (2 mg kg⁻¹; N=4–6). In all experiments samples of blood were taken for later determination of glucose or catecholamine levels.

To study the effect of i.c.v. administration of 5-HT on plasma glucose and hepatic glycogen levels, a series of dose–response experiments was carried out with the following doses of 5-HT (5.12, 12.8, 32, 80 and 200 μg kg⁻¹). Blood samples were extracted from the fish 60 min after administration of 5-HT, then the fish were decapitated and a sample of liver taken from each (N=10 fish per dose). A second series, of time–response experiments, examined the effect of the i.c.v. administration of a dose of 5-HT (12.8 μg kg⁻¹), and in this case the samples (blood and liver) were taken 20, 60, 120 and 180 min after administration (N=6–10 fish per group and time).

This experiment was carried out specifically to determine the capacity of the blood-brain barrier (BBB) to block the passage of peripherally administered 5-HT to the brain. For this purpose, we used [¹⁴C]5-HT (56 mCi mmol l⁻¹; Amersham Biosciences UK Limited, UK). The final concentration of 5-HT in the working solution (non-radioactive 5-HT + [¹⁴C]5-HT) was 6.36 mmol l⁻¹, which is equivalent to i.p. administration of 2 mg kg⁻¹ in a volume of 125 μl per fish. Samples of blood and brain were taken 15 min after the i.p. administration of [¹⁴C]5-HT. The radioactive content was determined, 5 min per sample, in a liquid scintillation counter (LS-6500, Beckman, USA), and the number of counts per minute detected in the plasma and in the brain tissue were expressed in nanomoles of 5-HT, based on the counts in the working solution of known concentration (N=8). Times in the order of 10–15 min are commonly used in studies of this type in mammals (Buschiazzo et al., 1970; Vitte et al., 1988; Naylor et al., 1995; Preston et al., 1995; Perasso et al., 2003) and fish (Aldegunde et al., 2000). If a substance passes the BBB it will actually be detectable in the brain long before 15 min. At the end of the study the fish and associated material were disposed of appropriately for radioactive material.

The effect of two different doses of fenfluramine (5 and 15 mg kg⁻¹) were investigated, 150 min after i.p. administration (N=6–8 fish per group and dose). In both cases, blood samples were taken for determination of glucose levels. The plasma levels of catecholamines were also determined in fish administered the lowest dose of d-FF (N=7 fish per group and dose).

To study the effect of the 5-HT2 receptor antagonist (MIAN), four different groups were established: (1) SAL–SAL, (2) SAL–MIAN (5 mg kg⁻¹), (3) SAL–5-HT (2 mg kg⁻¹) and (4) MIAN (5 mg kg⁻¹)–5-HT (2 mg kg⁻¹; N=6–7 fish per group). In all experiments in which antagonists were administered, the time between the first (antagonist) and second (5-HT) injections was 30 min. Blood samples were taken, for later determination of glucose and/or of catecholamines, 150 min after the second i.p. injection.

In addition, to study the effect of the receptor agonists for 5-HT1 (5-CT), 5-HT1B (TFMPP) and 5HT2 (α-m-5-HT), two different experiments were carried out. In the first (N=6 fish per group) three groups were established: (1) SAL, (2) 5-CT (1 mg kg⁻¹) and (3) TFMPP (2 mg kg⁻¹). In the second experiment (N=8 fish per group), two groups were established: (1) SAL and (2) α-m-5-HT (5 mg kg⁻¹). Blood samples were always collected 150 min after i.p. administration. In the first experiment, blood samples were taken for later determination of glucose and catecholamine levels, whereas in the second group, only the glucose levels were determined.

To study the effect of the receptor antagonists for 5-HT1A (WAY), 5-HT1D (BRL), 5-HT2B (SB) and 5-HT7 (PIM), the same timing was used for the injection and collection of blood samples for glucose analysis, as described at the beginning of this section. For the 5-HT1A antagonist, three treatments were considered: (1) SAL–SAL, (2) SAL–5-HT (2 mg kg⁻¹) and (3) WAY (2 mg kg⁻¹)–5-HT (2 mg kg⁻¹; N=8 fish per group). The experiment was repeated with the following modifications: 5-HT and WAY were injected consecutively, one immediately after the other, and the blood samples were collected after 75 min. The dose of WAY was 2.5 mg kg⁻¹ (N=12–16). For the other antagonists, different groupings were established, as follows; for 5-HT1D: (1) VEH–SAL, (2) VEH–5-HT (2 mg kg⁻¹) and (3) BRL (5 mg kg⁻¹)–5-HT (2 mg kg⁻¹); for the 5-HT2B antagonist: (1) VEH–SAL, (2) VEH–5-HT (2 mg kg⁻¹) and (3) SB (2 mg kg⁻¹)–5-HT (2 mg kg⁻¹); and for the 5-HT7antagonist: (1) VEH–SAL, (2) VEH–5-HT (2 mg kg⁻¹) and (3) PIM (10 mg kg⁻¹)–5-HT (2 mg kg⁻¹). The number of fish per group was 8–12.

Finally, to study the combined effect of the antagonists of 5-HT2B (SB) and 5-HT1A (WAY), three groups were considered: (1) VEH/SAL–SAL (2) VEH/SAL–5-HT (2 mg kg⁻¹), and (3) SB (2 mg kg⁻¹)/WAY (2 mg kg⁻¹)–5-HT (2 mg kg⁻¹). The VEH/SAL and the antagonists were injected consecutively, one immediately after the other, and 5-HT was injected i.p. 30 min later. Blood samples were collected, for determination of glucose levels, 150 min after the second injection (N=11–12 fish per group).

As most of the drugs were being used for the first time in fish (rainbow trout), we based the doses on those used in mammals. We decided to use doses within the medium to high range of those used in mammals (Niesink and Van Ree, 1982; Scotti de Carolis et al., 1986; Schechter, 1988; Hagan et al., 1997; Guscott et al., 2003; Morita et al., 2005; Freitas et al., 2006; Egashira et al., 2008).

Blood was obtained by caudal puncture with ammonium-heparinized syringes (Sigma, Sigma Chemical Co., St Louis, USA), and plasma was obtained after centrifugation of the blood (11,600 g for 5 min at 2–3°C). Aliquots of plasma (250 μl) for posterior catecholamine and glucose quantification, and tissue liver samples, for glycogen quantification, were stored at −80°C until their respective analysis. Liver samples (200 mg) were placed on ice and homogenized at 12,500 r.p.m. for 15 s, with ice-cooled 6 mol l⁻¹ perchloric acid (HClO₄), neutralized (with 1 mol l⁻¹ NaHCO₃), centrifuged (2 min at 9,000 g), and the supernatant stored at −80°C until glycogen assay (Figueroa et al., 2000). Glucose obtained after glycogen breakdown, and plasma glucose levels were determined by the glucose-oxidase–peroxidase method (SPINREACT, S.A., Girona, España). Plasma A and NA levels were determined on alumina (aluminium oxide)-extracted samples, by high-pressure liquid chromatography with electrochemical detection (HPLC-EC) (Caamaño-Tubio et al., 2010). Briefly, aliquots of plasma (250 μl) were mixed with DHBA (3,4-dihydroxybenzylamine hydrobromide, Sigma Chemical Co., St Louis, USA), added as an internal standard (40 μl of 100 ng ml⁻¹ of an aqueous solution of 0.1 mol l⁻¹ HClO₄), and 125 μl of 0.4 mol l⁻¹ HClO₄. After stirring for 1 min, samples were centrifuged for 5 min (11,600 g, 2–3°C). The supernatant was then transferred to a tube containing 50 mg acid alumina (Bio-Rad Laboratories, Richmond, California, USA), 250 μl of Tris-EDTA buffer 1.5 mol l⁻¹ (pH 8.6) and 100 μl of EDTA 0.8 mol l⁻¹. The tubes were then vortexed for 10 min during which time the catecholamines were adsorbed onto the alumina. This was followed by centrifugation at 11,600 g and subsequent washing of the alumina three times with 1 ml of chilled ultrapure water. For elution of catecholamines, 100 μl of 0.1 mol l⁻¹ HClO₄ was added to the alumina and stirred for 3 min. After centrifugation for 5 min (11,600 g), aliquots of the supernatant were stored at −80°C until A and NA analysis by HPLC-EC. Briefly, the chromatographic system consisted of a Waters M510 solvent delivery pump, a Supelcosil LC-18-DB reverse-phase analytical column (5 μm particle size, 150 mm×4.6 mm), and a Coulochem M5100 A detector, which included an analytical cell set at −0.50 mV (first electrode) and +330 mV (second electrode). The mobile phase, composed of a mixture of sodium acetate (93.3 mmol l⁻¹), citric acid (11.34 mmol l⁻¹), octanesulphonic acid (70 mg l⁻¹) and Na₂-EDTA (10 mg l⁻¹; final pH 5.25), was pumped isocratically at a flow rate of 1.5 ml min⁻¹ at room temperature. The concentrations of A and NA were estimated from data obtained from sample aliquots (20 μl) injected into the chromatographic system. Adrenaline (hydrated epinephrine tartrate) and noradrenaline (hydrated norepinephrine tartrate) were purchased from Sigma Chemical Co., St Louis, USA).

All data are presented as means ± standard error of the mean (s.e.m.). Statistical significance was evaluated with Student's t-test for comparison between two groups. In all other cases, one-way analysis of variance (ANOVA) followed by the Student–Newman–Keuls multiple comparison test were used. All analyses were carried out with commercial software (SigmaStat v2.0, SPSS Scientific Inc.).

In fish given i.p. injections of different doses of 5-HT (0.5, 1, 2, 5 and 10 mg kg⁻¹) there was a dose-dependent (P<0.001) increase in plasma glucose levels 150 min after administration of the compound (Fig. 1A). There was also a time-dependent (P<0.001) increase in plasma glucose levels 0, 50, 100, 150 and 200 min after 5-HT administration (2 mg kg⁻¹; Fig. 2B). In fish injected with saline solution no significant variations in plasma levels of glucose were observed. In fish administered 5-HT (2 mg kg⁻¹), there were significant increases (P<0.05) in plasma levels of A and NA, although the levels of NA were only significant after 150 and 200 min (Fig. 1C,D).

In fish given i.c.v. injections of different doses of 5-HT (5.12, 12.8, 32, 80 and 200 μg kg⁻¹) there were no significant variations in plasma glucose (Fig. 2A) or liver glycogen (Fig. 2C) levels 60 min after administration. Likewise, no significant variations in plasma glucose levels were observed 20, 60, 120 and 180 min after i.c.v. administration of one dose of 5-HT (12.8 μg kg⁻¹; Fig. 2B), or in liver glycogen (Fig. 2D).

The i.p. administration of two different doses of d-FF (5 mg kg⁻¹ and 15 mg kg⁻¹), a drug capable of inducing 5-HT release, led to a significant (P<0.05) dose-dependent increase in plasma levels of glucose (Fig. 3A). Intraperitoneal administration of d-FF (5 mg kg⁻¹) induced increases in A and NA levels, which was only statistically significant (P<0.05) in the case of A (Fig. 3B).

After i.p. administration of [¹⁴C]5-HT, 0% radioactivity was detected in the brain (data not shown). The radioactivity of the solution of [¹⁴C]5-HT injected into the fish, should be more than enough to detect any 5-HT in the brain if present.

The i.p. administration of 5-HT (SAL–5-HT) and MIAN–5-HT produced a significant level of hyperglycaemia (P<0.05) relative to the control (MIAN–SAL; Fig. 4A). The increases in glucose levels caused by these two treatments were not significantly different from each other. The treatments (SAL–5-HT or MIAN–5-HT) resulted in significant increases (P<0.05) in the levels of A (Fig. 4B) relative to the control (MIAN–SAL). Administration of 5-HT (SAL–5-HT) led to a significant increase (P<0.05) in the levels of NA, but administration of MIAN–5-HT, did not result in significantly different levels of NA (Fig. 4C) relative to the control (MIAN–SAL). No significant differences between the MIAN–SAL group (used as a control) and the SAL–SAL group were observed in any of the cases (data not shown).

Administration of 5-CT, but not TFMPP, induced a significant hyperglycaemic effect (P<0.05) (Fig. 5A). Administration of 5-CT induced a significant increase (P<0.05) in plasma A compared with the control (Fig. 5B). The TFMPP treatment did not significantly alter the levels of A (Fig. 5B). Neither 5-CT nor TFMPP significantly altered plasma levels of NA (Fig. 5C). Treatment with α-m-5-HT (5-HT2 receptor agonist) induced a slight, non significant increase in plasma levels of glucose (data not shown).

Analysis of the effect of the pretreatment with the antagonist 5-HT1A revealed that the i.p. administration of 5-HT (VEH–5-HT) and WAY-5-HT induced a significant level of hyperglycaemia (P<0.05) relative to the control (VEH–SAL). The increases in the levels of glucose caused by these two treatments were not significantly different (Fig. 6A). However, under the modified conditions, there appeared to a tendency for WAY to counteract the hyperglycaemic effect of 5-HT (Fig. 6B). Following i.p. administration of the antagonists of 5-HT1D (BRL), 5-HT2B (SB) and 5-HT7 (PIM) receptors, in all cases the levels of glucose in the groups subjected to the treatment (VEH–5-HT and 5-HT–antagonist) were higher than in the respective controls, but there were no significant differences between treated groups in any cases (Fig. 7A–C).

Finally, the administration of 5-HT (VEH/SAL–5-HT) and of the combined treatment with SB/WAY-5-HT induced a significant hyperglycaemic effect (P<0.05). However, there were no significant differences between the levels of glucose in fish administered VEH/SAL–5-HT and those administered SB/WAY-5-HT (Fig. 8).

The above results demonstrate, for the first time in a teleost, that peripheral (but not central) 5-HT exerts a hyperglycaemic effect. This effect is mediated by activation of 5-HT1-like receptors and release of adrenaline.

The first set of experiments investigated the possible modulatory role of 5-HT on blood glucose. The results demonstrate that the i.p. administration of 5-HT induces a clear hyperglycaemia, similar to that observed in mammals after i.p. administration of similar doses of 5-HT (Yamada et al., 1995). Given that in mammals the hyperglycaemic effect of 5-HT may be mediated by its action on the central and/or peripheral system, we investigated whether this was also true in fish. For this purpose, we first tested whether the peripherally administered 5-HT is capable of passing the BBB. As no radioactivity was detected in the brain 15 min after the i.p. administration of [¹⁴C]5-HT, we deduced that 5-HT does not pass the BBB, or that it only does so in very small quantities. This is consistent with findings in mammals (Axelrod and Inscoe, 1963; Davson and Segal, 1996) and with the fact that in most studies carried out on teleosts, the BBB has been found to be effective, although to different degrees (Lundquist, 1942; Davson and Segal, 1996). Further evidence for the lack of transfer of peripheral 5-HT to the brain is that although central administration of 5-HT has a strong anorexigenic effect, peripheral administration (i.p.) does not have any effect on feeding in the goldfish (De Pedro et al., 1998). By contrast, i.p. administration of d-FF, an indirect 5-HT agonist that passes the BBB, is capable of inducing a strong anorexigenic action (Ruibal et al., 2002). As 5-HT does not pass the BBB, we suggest that in the rainbow trout, the hyperglycaemic effect resulting from the i.p. administration of 5-HT is mediated by its action on the peripheral system. However, it is important to determine whether or not the hyperglycaemic action is a specific physiological response, i.e. if the release of endogenous 5-HT from its peripheral deposits is capable of replicating the action. The results do not leave any room for doubt: the i.p. administration of d-FF, a drug capable of inducing the release of 5-HT in the CNS (Ruibal et al., 2002) and from the peripheral stores of 5-HT in the rainbow trout (Caamaño-Tubio et al., 2007), mimics the hyperglycaemic action of 5-HT in a dose-dependent fashion. At the moment it is not possible to determine the specific origin of the stores of 5-HT involved.

The fact that 5-HT exerts a peripherally mediated hyperglycaemic effect does not exclude the possibility that it may also exert a regulatory effect on blood glucose by acting centrally. However, in light of the lack of alterations in the levels of hepatic glycogen (an index of the release of catecholamines) and of plasma glucose after the i.c.v. administration of 5-HT, this possibility can be discounted. The results are not consistent with those observed in mammals and show that in the rainbow trout, cerebral 5-HT does not act as a neuroregulator of the sympathetic–chromaffin cell nervous system. In relation to the above, we have previously observed that the cerebral hypoglycaemia induced by i.c.v. administration of 2-deoxyglucose is incapable of stimulating the sympathetic nervous system and inducing an increase in plasma glucose levels in rainbow trout (Soengas and Aldegunde, 2004), in contrast with the observations made in mammals (Muller et al., 1971; Weidenfeld et al., 1994). The available data indicate differences in the central mechanisms for regulating blood glucose levels in mammals and in rainbow trout.

It is known that in the rainbow trout, 5-HT may induce the release of catecholamines (Fritsche et al., 1993). Indeed, in the present study, i.p. administration of 5-HT induced an increase in levels of A and NA, and moreover the effect was partially mimicked by the endogenous 5-HT released by the action of d-FF. In both cases, the increases in plasma catecholamines were associated with an increased in blood glucose. The possible catecholaminergic mediation of the hyperglycaemic effect of the 5-HT is consistent with findings in mammals, in which: (1) the hyperglycaemic effect of 5-HT, whether via its central or peripheral action, may be mediated by the release of A from the adrenal medulla (Yamada et al., 1995; Sugimoto et al., 1996b; Chaouloff et al., 1990b), and (2) d-FF produces hyperglycaemia, mediated by the release of adrenal catecholamines (Chaouloff et al., 1992). However, interestingly, the increase in levels of 5-HT within physiological limits (treatment with d-FF) induces, in addition to a hyperglycaemic effect, differential action on the levels of A and NA, with a clearly predominant effect on A. The effect of d-FF on A but not NA lends support to the idea that A and not NA stimulates glucose release. The possible differential regulation of the release of A and NA is totally consistent with the fact that catecholamines are stored in different populations of chromaffin cells in fish (Gallo and Civinini, 2003) and that regulation of their release is also different in these animals (Reid et al., 1994; Bernier and Perry, 1996; Al-Kharrat et al., 1997).

Considering that in mammals the 5-HT2 and 5-HT1A receptors are closely involved in the hyperglycaemic effect of 5-HT (Sugimoto et al., 1996b; Chaouloff et al., 1990b), we attempted to determine possible mediation of this effect in the rainbow trout. This investigation was initiated by studying the effect of MIAN, a generic antagonist of 5-HT2 receptors, on the hyperglycaemic action of 5-HT. The results show that this antagonist was not capable of blocking the hyperglycaemic action of 5-HT and of the release of A, but that it was capable of blocking the release of NA. Interpretation of these results leads us to three important conclusions: (1) that the hyperglycaemic action of 5-HT must be mediated mainly by A, (2) that the 5-HT2-type receptors may mediate the release of NA but not of A, and (3) that the 5-HT2 type receptors do not appear to play an important role in mediating the 5-HT-dependent hyperglycaemia. The relative unimportance of the 5-HT2-type receptors in mediating the effect is confirmed by the fact that neither the administration of α-m-5-HT (5-HT2 agonist) or of SB (selective 5-HT2B antagonist) was capable of increasing blood glucose levels or of blocking the hyperglycaemic action of 5-HT, respectively. By contrast, when 5-CT (a non selective 5-HT1 agonist) was administered, the results were conclusive. Administration of 5-CT increased the levels of both glucose and of A, but not of NA. It appears evident that in the rainbow trout, the hyperglycaemic action of 5-HT is mediated by activation of 5-HT1-type receptors and the subsequent release of adrenaline. Obviously, in addition to the role of A, we cannot rule out a possible role for 5-HT (via 5-HT receptors expressed in the interrenal cells) in increasing the secretion of cortisol, which would result in an increase in glucose. It would be interesting to explore this possibility in future studies.

Based on the fact that in mammals 5-CT shows affinity for 5-HT1B, 5-HT1D, 5-ht5A, 5-HT7 and particularly, subtype 5-HT1A receptors (Yamada et al., 1998), we used a series of agonists and antagonists in an attempt to determine the subtype of 5-HT receptor involved. Under the experimental conditions of the present study, there was an absence of any role of the 5-HT1A, 5-HT1B, 5-HT1D and 5-HT7 receptors in the hyperglycaemic action of 5-HT. This is not consistent with the findings for 5-HT1A (Chaouloff et al., 1990c; Chaouloff et al., 1990a; Uvnäs-Moberg et al., 1996; Sugimoto et al., 2001) and 5-HT7 (Yamada et al., 1998), although it is consistent with the findings for 5-HT1B (Chaouloff et al., 1990b; Uvnäs-Moberg et al., 1996) in relation to mammals. The role of the 5-ht5A receptor was not evaluated in the present study, and therefore the possibility that it may at least partly mediate the hyperglycaemic action of 5-HT cannot be ruled out. Nevertheless, taking into account previously reported data in mammals (Hannon and Hoyer, 2008; Jonnakuty and Gragnoli, 2008), it can be concluded that this receptor is highly unlikely to play any role in the hyperglycaemic action of 5-HT.

The results demonstrate that in the rainbow trout, peripheral (but not central) 5-HT modulates homeostasis of peripheral glucose, with a hyperglycaemic effect. This effect is associated with the release of A and activation of 5-HT1-like receptors, which as far as could be ascertained in the present study, do not possess the pharmacological characteristics of subtypes 5-HT1A, 5-HT1B and 5-HT1D of mammals. The subtypes of 5-HT receptors in the fish nervous system have not yet been fully defined (Wikstrom et al., 1995; Winberg and Nilsson, 1996; Agrawal and Omeljaniuk, 2000) and even less is known about them in the peripheral system and in non-neuronal tissue, although subtypes that are clearly different from those found in mammals have been reported (Bakker et al., 1993; Janvier et al., 1996; Cerda et al., 1997; Sundin et al., 1998). Finally, the 5-HT2 type receptors that may mediate the release of NA, do not appear to play an important role in the hyperglycaemic effect induced by 5-HT.

This study was supported by a Research Grant from the Xunta de Galicia (PGIDIT03PXIB20001PR and INCITE08PXIB2000053PR) to M.A. J.P.M. was the recipient of a predoctoral fellowship from Xunta de Galicia (programa Maria Barbeito). The experiments described comply with the principles of animal care established by the Spanish government and the European Union (L358/1, 18/12/1986).

 
• A

  adrenaline

 
• BRL

  BRL 15572 hydrochloride

 
• 5-CT

  5-carboxamidotryptamine maleate

 
• d-FF

  d-fenfluramine

 
• DMSO

  dimethylsulphoxide

 
• HPLC

  high performance liquid chromatography

 
• 5-HT

  5-hydroxytryptamine or serotonin

 
• α-m-5-HT

  α-methyl-5-hydroxytryptamine maleate

 
• i.c.v.

  intracerebroventricular

 
• i.p.

  intraperitoneal

 
• MIAN

  mianserin hydrochloride

 
• NA

  noradrenaline

 
• PIM

  pimozide

 
• PPG

  propylenglycol

 
• SAL

  saline

 
• SB

  SB 204741

 
• TFMPP

  1-(3-trifluromethyphenyl)piperazine hydrochloride

 
• VEH

  vehicle

 
• WAY

  WAY 100635 maleate