This study was conducted to describe the cardiovascular responses to intra-arterial injections of serotonin in the Antarctic fish Pagothenia borchgrevinki and to elucidate the underlying mechanisms. Immunohistochemistry was used to localise serotonin-containing cells within the gills. Simultaneous and continuous recordings of ventral and dorsal aortic blood pressure, heart rate and ventral aortic blood flow (cardiac output) were made using standard cannulation procedures in combination with Doppler flow measurement. An extracorporeal loop with an in-line oxygen electrode allowed continuous measurements of arterial oxygen pressure Pre-branchial injection of serotonin (5-hydroxytryptamine, 5-HT) or the 5-HT2 receptor agonist α-methylserotonin increased the branchial vascular resistance and ventral aortic pressure, while the 5-HT1 receptor agonist piperazine was without effect. The branchial vasoconstriction produced by serotonin injection was completely blocked by the 5-HT1/5-HT2 receptor antagonist methysergide and the branchial vasoconstriction produced by α-methylserotonin injection was completely blocked by the specific 5-HT2 receptor antagonist LY53857. The results suggest that the 5-HT2 receptor alone mediates the branchial vasoconstriction. Serotonin also mediated a methysergide-sensitive reduction in , the reduction being greatest when the pre-injection value was high. 5-HT-immunoreactive cells and nerve fibres were present within the gill tissues. All the 5-HT-immunoreactive cells were located on the efferent side of the filaments, but 5-HT-immunoreactive nerve fibres were found lining both of the branchial arteries. Our findings demonstrate a potential serotonergic control system for the gills in Pagothenia borchgrevinki. In contrast to its effects on the branchial vasculature, serotonin produced a methysergide-insensitive decrease in the systemic vascular resistance. However, neither the specific 5-HT1 nor 5-HT2 receptor agonists produced a decrease in the resistance of the systemic vasculature. The nature of the serotonergic receptor(s) inducing vasodilation in teleost fish is uncertain.

The presence of both cholinergic and adrenergic control mechanisms in the branchial vasculature of the notothenioid Antarctic fish Pagothenia borchgrevinki (‘borch’) has been demonstrated (Axelsson et al. 1994), with α-adrenergic vasoconstrictor activity predominating over β-adrenergic vasodilatation. This prevalence of α-adrenoceptors in the branchial vasculature in borch has also been confirmed by experiments on isolated perfused gill arches (Forster et al. 1998). The dominance of α-adrenoceptor mechanisms over β-adrenoceptor effects on the cardiovascular system is interesting because, in teleost fish from temperate latitudes, β-adrenoceptor-mediated responses are dominant in summertime, while α-adrenoceptor-mediated responses are relatively more important during the winter (Pärt et al. 1982; Randall and Perry, 1992).

In addition, results from a recent study by Forster et al. (1998) show that serotonin (5-hydroxytryptamine, 5-HT) is more potent than both adrenaline and acetylcholine in constricting the branchial circulation in the borch. Serotonergic storage sites have been identified within the gills of every fish species so far studied (e.g. Dunel-Erb et al. 1982, 1989; Bailly et al. 1989, 1992; Zaccone et al. 1992; Sundin, 1996). The monoamine-induced branchial vasoconstriction can be blocked by the 5-HT1/5-HT2 receptor antagonist methysergide (Atlantic cod Gadus morhua, Östlund and Fänge, 1962; rainbow trout Oncorhynchus mykiss, Katchen et al. 1976; Fritsche et al. 1992; Sundin et al. 1995; eel Anguilla anguilla, Janvier et al. 1996b). Therefore, it is possible that serotonergic mechanisms are a component of the excitatory non-adrenergic and non-cholinergic branchial vascular control previously suggested by Pettersson and Nilsson (1979).

Measurements of oxygen partial pressure in the blood have shown that injection of serotonin is accompanied by a decrease in arterial in rainbow trout Oncorhynchus mykiss (Thomas et al. 1979; Fritsche et al. 1992). This is presumably due to the observed constriction of filamental arteries and arterioles induced by serotonin in the same species (Sundin et al. 1995). Preliminary results indicate that serotonin reduces the in borch as well (Forster et al. 1998). Nevertheless, the physiological significance of the serotonergic haemodynamic effects and the impact on gas exchange are unclear (Fritsche et al. 1992; Sundin et al. 1995).

It has been shown that branchial 5HT1/5HT2 receptors are involved in the 5-HT-induced increase in branchial resistance in teleosts; however, no attempts have been made to distinguish further the receptor mechanism involved. Furthermore, the in vitro study by Forster et al. (1998) on isolated borch gill arches and gill vessels indicates a branchial vasculature very sensitive to serotonin. No in vivo studies have been performed to elucidate the functional significance of the serotonergic control mechanisms, and there is also a lack of information regarding the presence and localization of serotonin in the branchial vasculature of borch. The present study was undertaken to investigate the branchial and systemic responses to intravascularly injected serotonin in the Antarctic fish P. borchgrevinki, to identify the receptor type involved and to localise branchial storage sites for serotonin.

Animals

The red-blooded Antarctic notothenioid fish Pagothenia borchgrevinki (Boulenger) (commonly called ‘borch’) was caught by line and hook in McMurdo Sound, Antarctica, and transported back to Scott Base (New Zealand Antarctic Programme) where they were kept in plastic tanks at −1.3 to −0.8 °C. Animals were also flown to Christchurch, New Zealand, where part of this study was undertaken. In Christchurch, the animals were held in plastic tanks in a closed-circuit sea water aquarium system at −0.5 °C.

Surgery

Twenty-one fish with body mass of 114.3±5.8 g (mean ± S.E.M.) were used in these studies. Prior to surgery, animals were given at least 48 h to recover from the effects of capture, and they were subsequently allowed 24 h to recover from surgery before measurements were started.

The surgery, including cannulation of the afferent and efferent branchial arteries and placement of a Doppler flow probe around the ventral aorta, was as described by Axelsson et al. (1994), as was the equipment used to record pressures and flows. An extracorporeal loop was established by pumping dorsal aortic blood via the efferent branchial cannula into the afferent branchial cannula via a flow-through oxygen electrode, allowing continuous recording of dorsal aortic oxygen tension. Blood was pumped around the extracorporeal loop using a Minipuls 3 pump (Gilson, Villiers-le-Bel, France). The volume of blood in the loop was approximately 180 μl, and the total circulation time was approximately 2 min. The Clark-type flow-through electrode (type 16-73) and meter (type OM-4) used to measure arterial oxygen partial pressure were manufactured by Microelectrodes Inc. (Londonderry, NH, USA). The oxygen electrode was submerged in the sea water at the same temperature (−0.5 to 0 °C) as the fish and was zeroed using sodium sulphite solution in sodium borate. The electrode was calibrated using air-saturated sodium chloride solution (0.9 %) and proved to be stable during the experimental period; it had a 95 % response time of less than 45 s at the experimental temperature. When the data were analysed, allowance was made both for the time taken for sampled blood to reach the oxygen electrode and for the response time of the electrode itself.

Experimental protocol

Individual fish were kept in the experimental chambers (approximately 170 mm×250 mm and 120 mm deep) for at least 24 h postsurgery to allow sufficient time for recovery from implantation of cannulae and the flow probe and to let the animals become accustomed to the environment. Since it is impossible, with the present arrangement, to record cardiovascular variables and oxygen partial pressures simultaneously, the animals were first connected for recordings of ventral (PVA) and dorsal (PDA) pressures, heart rate (fH) and cardiac output (Q). This allowed us to check whether the ventral and dorsal aortic cannulae were patent and that the animals had recovered sufficiently from surgery. Fish without apparent beat-to-beat heart rate variation were not used since a reduced cholinergic tone on the heart, leading to a low beat-to-beat variation in heart rate, is indicative of stress.

In preliminary trials, we tested for suitable dosages of agonist drugs. The doses of agonists chosen gave clear and consistent submaximal effects, and the selected doses of antagonists were sufficient to cause blockage.

Series 1

This part of the study was conducted at Scott Base, Antarctica. In this group, 0.1 nmol kg−1 (100 μl kg−1, 10−5 mol l−1) serotonin was administered twice before and twice after treatment with methysergide (1 mg kg−1, 2 μmol kg−1, a non-selective 5-HT1/5-HT2 receptor antagonist).

During the first injection, effects on cardiovascular parameters were measured; during the second, was measured. At the beginning of each experiment, resting values of the measured variables (Q, PVA, PDA, fH, ) were recorded, and after methysergide treatment the animals were left to recover for at least 30 min or until the cardiovascular variables had stabilised to pre-treatment values.

Series 2

This part of the study was conducted in Christchurch, New Zealand. To evaluate the serotonergic receptors involved in the cardiovascular responses obtained in series 1, specific agonists and antagonists for 5-HT1 and 5-HT2 receptors were administered sequentially. When the cardiovascular variables were stable, α-methylserotonin (a 5-HT2 receptor agonist; 0.1 nmol kg−1, 100 μl kg−1, 10−5 mol l−1) was injected and the effects recorded. After all the recorded variables had returned to resting values, piperazine (0.1 nmol kg−1, 100 μl kg−1, 10−5 mol l−1), a 5-HT1 receptor agonist, was administered and the effect was recorded. Then the fish were pre-treated with the 5-HT2 receptor antagonist LY53857 (0.5 mg kg−1, 1 μmol kg−1) and, when the cardiovascular parameters were stable, a final injection of α-methylserotonin was given.

Drugs

The following drugs were used in this study: serotonin hydrochloride (Sigma), chlorophenyl piperazine (Sigma), α-methylserotonin maleate (Research Biochemicals International, RBI), LY53857 maleate (RBI) and methysergide dimaleate (Sandoz). All drug injections were made into the efferent arterial cannula.

Data acquisition and statistics

The cardiovascular variables and were recorded continuously using a Yokogawa recorder (model 3701 LR8100), and the data were also sampled using LabView (version 4.0, National Instruments). Sampling frequency was 30 Hz, and mean values were subsequently created at 15 s intervals. Data are presented as means ± S.E.M. for N animals. Branchial and systemic vascular resistance (RGill and RSys, respectively) were calculated as (PVAPDA)/Q and PDA/Q, respectively, assuming that the entire cardiac output passed through the systemic vessels and that venous pressure was close to zero and did not change substantially during the experiments. Cardiac output was measured using the pulsed Doppler technique. Since no calibration of the flow signal was performed, the resistance changes described above are expressed as a percentage change from the value for the control period.

Evaluation of statistically significant changes (P<0.05) in the recorded variables was made using the Wilcoxon signed-rank test. To test the correlation between resting levels and the change in induced by serotonin (see Fig. 2), a Spearman rank correlation test was used.

Immunohistochemistry

The histochemical procedure followed that outlined in Karila et al. (1995). Four whole gill arches from either the left or the right side were excised from four fish killed by a sharp blow to the head. The gill arches were fixed in 15 % saturated picric acid and 2 % formaldehyde in 0.1 mol l−1 phosphate-buffered saline (PBS, pH 7.3) for 24 h at 4 °C. They were then rinsed and dehydrated in an ethanol series, treated with xylene for 30 min, rehydrated in ethanol and stored in PBS containing 30 % sucrose as a cryoprotectant until cut sections containing 3–4 filament pairs from the middle portion of each gill arch were quick-frozen in isopentane (2-methylbutane), cooled in liquid nitrogen, cut into 10 μm thick sections using a cryostat and dried onto Vectabond-(Vector) coated slides. Since the bone of borch is less calcified than that of most other teleosts (Eastman and DeVries, 1981, 1982), it was possible to cut through the whole gill arch to obtain sections containing the branchial arteries. The sections were incubated in a moist chamber at room temperature (22–24 °C) with the primary antiserum (NA-08-324, dilution 1:500, CRB, Cambridge, UK, host rabbit) for single staining for 24 h. The preparations were rinsed three times in hypertonic PBS plus 2.0 % NaCl for 5 min and incubated with secondary antibodies (DaR-DTAF, dilution 1:100, Jackson Immunores., USA, host donkey) for 1–2 h. The secondary antibodies were conjugated to dichlorotriazinyl amino fluorescein (DTAF). They were diluted in hypertonic PBS to reduce non-specific binding to the tissue sections (Grube, 1980). Preparations were viewed with an Olympus Vanox fluorescence microscope and photographed with a Leitz Orthomat camera using Kodak T-MAX 400 film rated at 1000 ASA.

To control for the specificity of the secondary antibodies, sections were either pre-incubated with normal donkey serum (1:10), or the primary antiserum was omitted from the staining procedure. We use the term ‘immunoreactive’ (abbreviated as ‘IR’) when referring to the immunohistochemical localization of serotonin.

Cardiovascular experiments

Cardiac output measurements were used only to calculate branchial and systemic resistance; since none of the agonists or antagonists used had any significant effect on cardiac output, values of this variable are not included below.

Series 1

Injections of serotonin caused a significant increase in RGill, which resulted in a significant elevation of PVA (Fig. 1A). The vascular events were accompanied by a reduction in arterial (Fig. 1B). Pre-treatment with the 5HT1/5HT2 receptor antagonist methysergide abolished these effects (Fig. 1). In contrast, PDA decreased significantly as a result of the reduced RSys (Fig. 1A) and pre-treatment with methysergide did not alter this systemic effect of serotonin (Fig. 1B).

Fig. 1.

The effects of intra-arterial injections of serotonin (0.1 nmol kg−1) before (N=11, left panel) and after (N=8, right panel) treatment with 1 mg kg−1 methysergide (a 5-HT1/5-HT2 receptor antagonist) on cardiovascular parameters (A) and arterial oxygen tension (B). P, pressure; PVA and PDA, ventral and dorsal aortic pressure, respectively; RGill, gill resistance; RSys, systemic resistance;

PaO2
arterial oxygen tension. Values are means ± S.E.M. An asterisk indicates a significant difference between values determined 2 min (A) and 4 min (B) min after injection versus pre-injection values. A dagger indicates a statistical difference between the effects of the agonist before (left panel) and after (right panel) methysergide treatment. Injections are indicated by upward-pointing arrowheads. The discontinuous traces in A are due to injection artefacts in the dorsal aortic pressure trace affecting the resistance calculations.

Fig. 1.

The effects of intra-arterial injections of serotonin (0.1 nmol kg−1) before (N=11, left panel) and after (N=8, right panel) treatment with 1 mg kg−1 methysergide (a 5-HT1/5-HT2 receptor antagonist) on cardiovascular parameters (A) and arterial oxygen tension (B). P, pressure; PVA and PDA, ventral and dorsal aortic pressure, respectively; RGill, gill resistance; RSys, systemic resistance;

PaO2
arterial oxygen tension. Values are means ± S.E.M. An asterisk indicates a significant difference between values determined 2 min (A) and 4 min (B) min after injection versus pre-injection values. A dagger indicates a statistical difference between the effects of the agonist before (left panel) and after (right panel) methysergide treatment. Injections are indicated by upward-pointing arrowheads. The discontinuous traces in A are due to injection artefacts in the dorsal aortic pressure trace affecting the resistance calculations.

Fig. 2.

The correlation between resting arterial oxygen tension

(PaO2)
and the reduction in
PaO2
caused by injection of serotonin (0.1 nmol kg−1, N=11).

Fig. 2.

The correlation between resting arterial oxygen tension

(PaO2)
and the reduction in
PaO2
caused by injection of serotonin (0.1 nmol kg−1, N=11).

The magnitude of the reduction in after serotonin injection was significantly correlated with resting levels (Fig. 2).

Series 2

The branchial effects of injected α-methylserotonin (a 5-HT2 receptor agonist) resembled those obtained with serotonin: a significant increase in RGill and PVA (Fig. 3). Pre-treatment with a specific 5-HT2 receptor antagonist (LY53857) blocked the increase in RGill and RSys (Fig. 3). In contrast to the effects of serotonin, α-methylserotonin did not significantly reduce RSys (Fig. 3). Piperazine (a 5-HT1 receptor agonist) lacked effect on any of the cardiovascular variables measured (Fig. 4).

Fig. 3.

The effects of intra-arterial injections of α-methylserotonin (a 5-HT2 receptor agonist, 0.1 nmol kg−1) before (N=10, left panel) and after (N=10, right panel) treatment with LY53857 (a 5-HT2 receptor antagonist) on cardiovascular variables. Abbreviations as in Fig. 1. Values are mean ± S.E.M. An asterisk indicates a significant difference between values determined 2 min after injection and pre-injection values. A dagger indicates a significant difference between the effect of α-methylserotonin before and after treatment with LY53857. Injections are indicated by the upward-pointing arrowheads. The discontinuous traces in A are due to injection artefacts in the dorsal aortic pressure trace affecting the resistance calculations.

Fig. 3.

The effects of intra-arterial injections of α-methylserotonin (a 5-HT2 receptor agonist, 0.1 nmol kg−1) before (N=10, left panel) and after (N=10, right panel) treatment with LY53857 (a 5-HT2 receptor antagonist) on cardiovascular variables. Abbreviations as in Fig. 1. Values are mean ± S.E.M. An asterisk indicates a significant difference between values determined 2 min after injection and pre-injection values. A dagger indicates a significant difference between the effect of α-methylserotonin before and after treatment with LY53857. Injections are indicated by the upward-pointing arrowheads. The discontinuous traces in A are due to injection artefacts in the dorsal aortic pressure trace affecting the resistance calculations.

Fig. 4.

The effects of intra-arterial injections of piperazine (a 5-HT1 receptor agonist, 0.1 nmol kg−1, N=7) on cardiovascular variables. Abbreviations as in Fig. 1. Values are mean ± S.E.M. There were no significant differences in the variables before and after injection of piperazine (upward-pointing arrowhead). The discontinuous traces in A are due to injection artefacts in the dorsal aortic pressure trace affecting the resistance calculations.

Fig. 4.

The effects of intra-arterial injections of piperazine (a 5-HT1 receptor agonist, 0.1 nmol kg−1, N=7) on cardiovascular variables. Abbreviations as in Fig. 1. Values are mean ± S.E.M. There were no significant differences in the variables before and after injection of piperazine (upward-pointing arrowhead). The discontinuous traces in A are due to injection artefacts in the dorsal aortic pressure trace affecting the resistance calculations.

Immunohistochemistry

Two morphologically different types of 5-HT-immunoreactive (5-HT-IR) cell could be identified on the efferent side of the filament on all four gill arches. While round cells without short processes or with only a few short processes were concentrated in the distal third of the filament, bipolar neurones with long processes ran parallel to the filament along the efferent side (Figs 5A, 6A). Varicose nerve fibres were also seen along the filaments (Fig. 6B). At the individual filamental tips, the round cells formed discrete groups that were invariably located outside the efferent filamental artery close to the external medium (Fig. 5B). The cells in the afferent filamental artery in Fig. 5B are autofluorescent red blood cells and are not 5-HT-immunoreactive. These cells resemble the neuroepithelial cells (NECs) described by previous authors in other species (e.g. Dunel-Erb et al. 1982; Laurent, 1984; Bailly et al. 1992). In addition, there was a rich 5-HT-IR innervation of the basal parts of the efferent filamental arteries of the sphincter area (Fig. 7A). No 5-HT-IR cells were distinguished along the afferent filamental arteries (non-specifically stained red blood cells can be seen in Fig. 5B in the afferent filamental artery). However, 5-HT-IR fibres were found innervating the outer walls of both the afferent and efferent branchial arteries (Fig. 7B). Peculiar 5-HT-IR cells were commonly observed on the opposite side of the gill arch towards the buccal cavity. They formed groups of at least two cells (sometimes more cells could be seen) oriented like flower-cups at the ends of branch-like structures in the gill rakers (Fig. 8).

Fig. 5.

Longitudinal (A) and transverse (B) sections through a gill filament of Pagothenia borchgrevinki showing 5-HT-immunoreactive (5-HT-IR) cells (indicated by arrowheads) on the efferent side of the filament. A group of 5-HT-IR cells forms a cluster near the filament tip (A); these cells are located on the outside of the efferent filamental artery (EFA) close to the external medium. AFA, afferent filamental artery. Scale bars, 100 μm.

Fig. 5.

Longitudinal (A) and transverse (B) sections through a gill filament of Pagothenia borchgrevinki showing 5-HT-immunoreactive (5-HT-IR) cells (indicated by arrowheads) on the efferent side of the filament. A group of 5-HT-IR cells forms a cluster near the filament tip (A); these cells are located on the outside of the efferent filamental artery (EFA) close to the external medium. AFA, afferent filamental artery. Scale bars, 100 μm.

Fig. 6.

Longitudinal sections through the middle of the efferent part of a Pagothenia borchgrevinki gill filament showing 5-HT-immunoreactive (5-HT-IR) bipolar neurones (A, arrowhead) and varicose 5-HT-IR nerve fibres (B, arrowhead). Scale bars, 100 μm.

Fig. 6.

Longitudinal sections through the middle of the efferent part of a Pagothenia borchgrevinki gill filament showing 5-HT-immunoreactive (5-HT-IR) bipolar neurones (A, arrowhead) and varicose 5-HT-IR nerve fibres (B, arrowhead). Scale bars, 100 μm.

Fig. 7.

Longitudinal sections through the basal parts of the gill filament (A) and the branchial arteries (B) of Pagothenia borchgrevinki. The section in A passes cut transversely through the efferent filamental artery (EFA). Note the rich innervation by 5-HT-immunoreactive (5-HT-IR) nerve fibres surrounding the EFA in A. The outer walls of the two branchial arteries (BA) in B are innervated by 5-HT-IR nerve fibres. Scale bars, 100 μm.

Fig. 7.

Longitudinal sections through the basal parts of the gill filament (A) and the branchial arteries (B) of Pagothenia borchgrevinki. The section in A passes cut transversely through the efferent filamental artery (EFA). Note the rich innervation by 5-HT-immunoreactive (5-HT-IR) nerve fibres surrounding the EFA in A. The outer walls of the two branchial arteries (BA) in B are innervated by 5-HT-IR nerve fibres. Scale bars, 100 μm.

Fig. 8.

Longitudinal section through the gill arch of Pagothenia borchgrevinki showing 5-HT-immunoreactive (5-HT-IR) cells in the gill rakers directed towards the buccal cavity (BC). Note that the stained cells are situated at the ends of the branches of a tree-like structure (A) and that the grouped cells (at least two, sometimes more) are connected like flower-cups (B). Scale bars, 100 μm.

Fig. 8.

Longitudinal section through the gill arch of Pagothenia borchgrevinki showing 5-HT-immunoreactive (5-HT-IR) cells in the gill rakers directed towards the buccal cavity (BC). Note that the stained cells are situated at the ends of the branches of a tree-like structure (A) and that the grouped cells (at least two, sometimes more) are connected like flower-cups (B). Scale bars, 100 μm.

The present investigation demonstrates clearly that the branchial vasoconstriction caused by serotonin in P. borchgrevinki is mediated by a 5-HT2 receptor, since the serotonin-activated increase in RGill and PVA (Fig. 1A) was mimicked by α-methylserotonin (a specific 5-HT2 receptor agonist) (Fig. 3) but not by piperazine (a specific 5-HT1 receptor agonist) (Fig. 4). In addition, the constriction of the gill vessels by α-methylserotonin was effectively blocked by LY53857 (Fig. 3), a specific 5-HT2 receptor antagonist, which further supports this conclusion. Because the vascular responses to serotonin, which are suppressed by methysergide (a general 5-HT1/5-HT2 receptor antagonist) (Fig. 1), are similar to those obtained in all other species investigated (Atlantic cod Gadus morhua, Östlund and Fänge, 1962; rainbow trout Oncorhynchus mykiss, Katchen et al. 1976; Fritsche et al. 1992; Sundin et al. 1995; eel Anguilla anguilla, Janvier et al. 1996b), it seems very likely that the 5-HT2 receptor is a ‘general’ mediator of serotonergic vasoconstriction in the teleost branchial vasculature. It is also interesting to note the difference in the concentration of serotonin required to elicit similar significant cardiovascular responses between this species (0.1 nmol kg−1) and rainbow trout (100 nmol kg−1). The high potency of serotonin in borch gill vessels has also been demonstrated in vitro (Forster et al. 1998). The approximately 1000-fold higher sensitivity to serotonin in borch compared with rainbow trout may be due to species differences, but it could also represent adaptation to a cold environment. Cold adaptation has been suggested as an explanation for the dominance of α-adrenoceptor over β-adrenoceptor responses in borch (Axelsson et al. 1994) and in winter fish from temperate waters (Pärt et al. 1982; Randall and Perry, 1992). Interestingly, cooling enhances, presumably by an increasing in receptor affinity, both 5-HT2-receptor-and α2-adrenoceptor-mediated vasoconstriction in mammals (Vanhoutte and Flavahan, 1986; Bodelsson et al. 1990a,b; Harker et al. 1991). If a direct effect of temperature on receptor affinity applies to 5-HT2 receptors in fish and if there is an equivalent receptor density among species, then fish should display different sensitivities to serotonin depending on the ambient water temperature.

A consequence of the serotonin-induced branchial vasoconstriction was an impaired gas transfer and, as for the vascular responses, the reduction in arterial could be blocked by methysergide treatment (Fig. 1B). Interestingly, the degree of reduction in depended on the resting (Fig. 2), i.e. animals with high resting levels displayed a larger reduction in following serotonin administration. A close correlation between an increased pressure differential (PVAPDA) and has been noted in a study of the borch cardiovascular system (Forster et al. 1998). Assuming that the pressure difference across the gills represents gill vascular resistance, there appears to be a relationship between branchial resistance and dorsal aortic oxygen tension. In the present study, the correlation between the magnitude of the serotonin-induced decrease in and the resting may again indicate the importance of branchial vascular resistance for the oxygenation of arterial blood. Exogenous injection of serotonin similarly diminishes gas transfer in rainbow trout (Thomas et al. 1979; Fritsche et al. 1992).

The physiological significance of the effects of serotonin on gas transfer in teleosts is unclear, since serotonin is released from neuroepithelial cells in rainbow trout gills during hypoxic exposure (Dunel-Erb et al. 1982), and Bailly et al. (1989) suggested that serotonin-mediated constriction of the efferent arterial vasculature would cause an increase in perfusion pressure which would improve perfusion of the more distal lamellae, a mechanism thought to be essential during hypoxia. Therefore, the present results (Fig. 1B), like those in rainbow trout (Thomas et al. 1979; Fritsche et al. 1992), seem maladaptive. This conflict initiated a study on rainbow trout (Sundin et al. 1995) in which simultaneous observations of the microcirculation within the filaments and measurements of cardiovascular parameters were made. Serotonin induced a vasoconstriction in the filaments that redistributed blood flow towards the base of the filaments. Olson (1979) has suggested that the most efficient way to increase oxygen transfer across the gill is to recruit additional lamellae; however, the occlusion of the distal lamellae induced by serotonin would have the opposite effect. It has been shown that a serotonergic mechanism is not involved in hypoxia-induced branchial vasoconstriction in rainbow trout (Sundin and Nilsson, 1997).

Moreover, the water south of the Antarctic convergence is cold and well-oxygenated throughout the year; therefore, it is unlikely that Antarctic fish would be exposed to hypoxia. This, together with the fact that borch seem to be more sensitive to serotonin than are the temperate species tested, further argues against a hypoxic serotonergic vasoconstriction of the arterio-arterial pathway in fish. Other possible roles for serotonin in fish gills are in ventilatory (Fritsche et al. 1992; Janvier et al. 1996a), sensory (Dunel-Erb et al. 1982; Burleson and Milsom, 1995; Sundin, 1996) and acid–base (Thomas et al. 1979; L. Sundin and S. Nilsson, unpublished data) systems. It should also be remembered that 5-HT is normally released locally and, when injected into the circulation, it can bind to all available 5-HT receptors in the circulatory system causing more generalised effects than when it is released locally.

As in other teleosts (Dunel-Erb et al. 1982, 1989; Bailly et al. 1989, 1992; Zaccone et al. 1992; Sundin, 1996), and corroborating the physiological experiments discussed above, 5-HT-IR cells and nerve fibres were found within the gill tissues (Figs 58). As has been described in all other teleosts studied so far, the 5-HT-IR cells and nerve fibres were located on the efferent side of the filaments. 5-HT-IR nerves lining both arteries were detected (Fig. 7B). Therefore, unlike previous reports of an ‘exclusive’ location of serotonin storage sites on the efferent side of the filaments in fish gills (Dunel-Erb et al. 1982, 1989; Bailly et al. 1989, 1992; Zaccone et al. 1992; Sundin, 1996), the present study also provides evidence for a serotonergic innervation of the afferent branchial artery. Such innervation supports recent observations of serotonin-induced constriction of afferent branchial artery ring preparations in borch (Forster et al. 1998). Furthermore, despite the supposed exclusive location of efferent filamental serotonin storage sites, serotonin-induced constriction of an afferent filamental artery in rainbow trout has also been observed (Sundin et al. 1995). Evidently, there are receptors for serotonin on both sides of the branchial vasculature. If the finding of afferent branchial 5-HT-IR fibres in borch applies to other teleost species, the source providing serotonin to act on afferent receptors does not necessarily have to be extrabranchial, as has been discussed previously (Sundin et al. 1995). Although a direct vasoregulatory function of serotonin released from the different 5-HT-IR cells can be assumed, the localisation of the grouped neuroepithelial cells at the filamental tips in borch (Fig. 5A) implies a function unrelated to branchial vasomotor control. In accordance with the concept of water-sensing branchial oxygen receptors in rainbow trout (Bailly et al. 1992) and Atlantic cod Gadus morhua (Sundin, 1996), these grouped neuroepithelial cells may represent a site for environmental oxygen sensors. In addition, the peculiar 5-HT-IR cells on the gill rakers near the buccal cavity (Fig. 8) may also be part of a chemoreceptive mechanism.

As opposed to the branchial events, intra-arterial injections of serotonin induced a methysergide-insensitive decrease in RSys (Fig. 1), a response identical to that obtained in other teleosts (Sundin et al. 1995; Janvier et al. 1996b). This insensitivity to methysergide and the fact that specific 5-HT1 and 5-HT2 receptor agonists did not reduce RSys to a similar extent (Figs 3, 4) show that the mechanism responsible is not mediated by 5-HT1/5-HT2 receptors. Furthermore, the serotonin-induced systemic vasodilation in the European eel Anguilla anguilla is also insensitive to 5-HT3 and 5-HT4 receptor antagonists and cardiac vagotomy (Janvier et al. 1996b). In contrast to the lack of blockade of the inhibitory effect of serotonin on RSys in fish, methysergide is a potent antagonist of both endothelium-dependent and endothelium-independent inhibitory serotonergic mechanisms in the rabbit and rat jugular vein (Leff et al. 1987; Martin et al. 1987; Bodelsson et al. 1993). This controversy adds complexity to the mechanisms behind the vasodilatory action of serotonin in fish.

In conclusion, the cardiovascular system of borch is extremely sensitive to serotonin and the branchial vasoconstriction is mediated by 5-HT2 receptors. These facts, in conjuction with the immunohistochemical findings of both 5-HT-IR cells and 5-HT-IR nerve fibres within the gill tissues, demonstrate the presence of a serotonergic control system in the branchial vasculature of the Antarctic fish P. borchgrevinki. However, we cannot, at present, identify the mechanism responsible for the systemic serotonin-induced vasodilation.

This work was supported by grants from the New Zealand Antarctic Programme (NZAP), the Swedish Natural Science Research Council (NFR), Swedish Agricultural and Forestry Reseach Council, Knut and Alice Wallenberg Foundation, Längmanska Kulturfonden and Helge Ax:son Johnsons Stiftelse. We wish to thank the staff of Scott Base, Antarctica, for their support.

Axelsson
,
M.
,
Davison
,
B.
,
Forster
,
M.
and
Nilsson
,
S.
(
1994
).
Blood pressure control in the Antarctic fish Pagothenia borchgrevinki
.
J. exp. Biol
.
190
,
265
279
.
Bailly
,
Y.
,
Dunel-Erb
,
S.
,
Geffard
,
M.
and
Laurent
,
P.
(
1989
).
The vascular and epithelial serotonergic innervation of the actinopterygian gill filament with special reference to the trout, Salmo gairdneri
.
Cell Tissue Res
.
258
,
349
363
.
Bailly
,
Y.
,
Dunel-Erb
,
S.
and
Laurent
,
P.
(
1992
).
The neuroepithelial cells of the fish gill filament: Indolamine-immunocytochemistry and innervation
.
Anat. Rec
.
233
,
143
161
.
Bodelsson
,
M.
,
Arneklo-Nobin
,
B.
,
Nobin
,
A.
,
Owman
,
C. H.
,
Sollerman
,
C.
and
Törnebrandt
,
K.
(
1990a
).
Cooling enhances alpha2-adrenoceptor-mediated vacoconstriction in human hand veins
.
Acta physiol. scand
.
138
,
283
291
.
Bodelsson
,
M.
,
Törnebrandt
,
K.
and
Arneklo-Nobin
,
B.
(
1990b
).
Effect of cooling on smooth muscle response to 5-hydroxytryptamine in human hand veins
.
Acta physiol. scand
.
140
,
331
339
.
Bodelsson
,
M.
,
Törnebrandt
,
M.
and
Arneklo-Nobin
,
B.
(
1993
).
Endothelial relaxing 5-hydroxy-tryptamine receptors in the rat jugular vein – similarity with the 5-hydroxy-tryptamine1c receptor
.
J. Pharmac. exp. Ther
.
264
,
709
716
.
Burleson
,
M. L.
and
Milsom
,
W. K.
(
1995
).
Cardio-ventilatory control in rainbow trout. I. Pharmacology of branchial, oxygen-sensitive chemoreceptors
.
Respir. Physiol
.
100
,
231
238
.
Dunel-Erb
,
S.
,
Bailly
,
Y.
and
Laurent
,
P.
(
1982
).
Neuroepithelial cells in fish gill primary lamellae
.
J. appl. Physiol
.
53
,
1342
1353
.
Dunel-Erb
,
S.
,
Bailly
,
Y.
and
Laurent
,
P.
(
1989
).
Neurons controlling the gill vasculature in five species of teleosts
.
Cell Tissue Res
.
255
,
567
573
.
Eastman
,
J. T.
and
Devries
,
A. L.
(
1981
).
Buoyancy adaptations in a swim-bladderless Antarctic fish
.
J. Morph
.
167
,
97
102
.
Eastman
,
J. T.
and
Devries
,
A. L.
(
1982
).
Buoyancy studies of notothenioid fishes in McMurdo Sound, Antarctica
.
Copeia
1982
,
385
393
.
Forster
,
M. E.
,
Forster
,
A. H.
and
Davision
,
W.
(
1998
).
Effects of serotonin, adrenaline and other vasoactive drugs on the branchial blood vessels of the Antarctic fish Pagothenia borchgrevinki
.
Fish Physiol. Biochem. (in press)
.
Fritsche
,
R.
,
Thomas
,
S.
and
Perry
,
S. F.
(
1992
).
Effects of serotonin on circulation and respiration in the rainbow trout Oncorhynchus mykiss
.
J. exp. Biol
.
173
,
59
73
.
Grube
,
D.
(
1980
).
Immunoreactivities of gastrin (G-) cells. II. Non-specific binding of immunoglobulins to G-cells by ionic interactions
.
Histochemistry
87
,
149
167
.
Harker
,
C. T.
,
Taylor
,
L. M.
and
Porter
,
J. M.
(
1991
).
Vascular contractions to serotonin are augmented by cooling
.
J. cardiovasc. Pharmac
.
18
,
791
796
.
Janvier
,
J.-J.
,
Peyraud-Waitzenegger
,
M.
and
Soulier
,
P.
(
1996a
).
Mediation of serotonin-induced hyperventilation via 5-HT3-receptor in European eel Anguilla anguilla
.
J. comp. Physiol. B
165
,
640
646
.
Janvier
,
J.-J.
,
Peyraud-Waitzenegger
,
M.
and
Soulier
,
P.
(
1996b
).
Effects of serotonin on the cardio-circulatory system of the European eel (Anguilla anguilla) in vivo
.
J. comp. Physiol. B
166
,
131
137
.
Karila
,
P.
,
Axelsson
,
M.
,
Franklin
,
C. E.
,
Fritsche
,
R.
,
Gibbins
,
I. L.
,
Grigg
,
G. C.
,
Nilsson
,
S.
and
Holmgren
,
S.
(
1995
).
Neuropeptide immunoreactivity and co-existence in cardiovascular nerves and autonomic ganglia of the estuarine crocodile, Crocodylus porosus and cardiovascular effects of neuropeptides
.
Regul. Peptides
58
,
25
39
.
Katchen
,
M. S.
,
Olson
,
K. R.
and
Wayne
,
C.
(
1976
).
Effects of histamine and serotonin on isolated perfused gill of rainbow trout (Salmo gairdneri)
.
Fedn Proc. Fedn Am. Socs exp. Biol
.
35
,
528
.
Laurent
,
P.
(
1984
).
Gill internal morphology
.
In Fish Physiology
(ed.
W. S.
Hoar
and
D. J.
Randall
), pp.
73
183
.
Orlando, FL
:
Academic Press
.
Leff
,
P.
,
Martin
,
G. R.
and
Morse
,
J. M.
(
1987
).
Differential classification of vascular smooth muscle and endothelial cell 5-HT receptors by use of tryptamine analogues
.
Br. J. Pharmac
.
91
,
321
331
.
Martin
,
G. R.
,
Leff
,
P.
,
Cambridge
,
D.
and
Barrett
,
V. J.
(
1987
).
Comparative analysis of two types of 5-hydroxytryptamine receptor mediating vasorelaxation: different classification using tryptamines
.
Naunyn-Schmiedeberg’s Arch. Pharmac
.
336
,
365
373
.
Olson
,
K. R.
(
1979
).
The linear cable theory as a model of gill blood flow
.
J. theor. Biol
.
81
,
377
388
.
östlund
,
E.
and
Fänge
,
R.
(
1962
).
Vasodilation by adrenaline and noradrenaline and the effects of some other substances on perfused fish gills
.
Comp. Biochem. Physiol
.
5
,
307
309
.
Pärt
,
P.
,
Kiessling
,
A.
and
Ring
,
O.
(
1982
).
Adrenaline increases vascular resistance in perfused rainbow trout (Salmo gairdneri Rich.) gills
.
Comp. Biochem. Physiol
.
72C
,
107
108
.
Pettersson
,
K.
and
Nilsson
,
S.
(
1979
).
Nervous control of the branchial vascular resistance of the Atlantic cod, Gadus morhua
.
J. comp. Physiol
.
129
,
179
183
.
Randall
,
D. J.
and
Perry
,
S. F.
(
1992
).
Catecholamines
.
In Fish Physiology, vol XIIB
(ed.
W. S.
Hoar
,
D. J.
Randall
and
A. P.
Farrell
), pp.
255
300
.
San Diego, CA
:
Academic Press
.
Sundin
,
L.
(
1996
).
Control of gill blood flow in teleosts with emphasis on mechanisms activated by hypoxia
.
PhD thesis, Zoological Institute, Göteborg University
, ISBN 91-628 2229-2.
Sundin
,
L.
and
Nilsson
,
G. E.
(
1997
).
Neurochemical mechanisms behind gill microcirculatory responses to hypoxia in trout: in vivo microscopy study
.
Am. J. Physiol
.
41
,
R576
R585
.
Sundin
,
L.
,
Nilsson
,
G. E.
,
Block
,
M.
and
Löfman
,
C. O.
(
1995
).
Control of gill filament blood flow by serotonin in the rainbow trout, Oncorhynchus mykiss
.
Am. J. Physiol
.
37
,
R1224
R1229
.
Thomas
,
S.
,
Belaud
,
A.
and
Peyraud
,
C.
(
1979
).
Arguments for serotoninergic adjustments in gill blood circulation in fish
.
IRCS Medical Science: Cardiovascular System; Environmental Biology and Medicine; Experimental Animals; Physiology; Respiratory System
7
,
543
.
Vanhoutte
,
P. M.
and
Flavahan
,
N. A.
(
1986
).
Effects of temperature on alpha adrenoceptors in limb veins: role of receptor reserve
.
Fedn Proc. Fedn Am. Socs exp. Biol
.
45
,
2347
2354
.
Zaccone
,
G.
,
Lauweryns
,
J. M.
,
Fasulo
,
S.
,
Tagliafierro
,
G.
,
Ainis
,
L.
and
Licata
,
A.
(
1992
).
Immunocytochemical localization of serotonin and neuropeptides in the neuroendocrine paraneurons of teleost and lungfish gills
.
Acta zool
.
73
,
177
183
.