Blood pressure control in the antarctic fish Pagothenia borchgrevinki

Antarctic fish of the sub-Order Notothenioidei live at constant low temperatures south of the Antarctic convergence. This group of fish includes both the haemoglobin-free ice fish (family Channichthyidae) and the red-blooded nototheniids (family Nototheniidae). Notothenioids maintain a high cardiac output despite a low heart rate. In a previous preliminary study, the cryopelagic nototheniid Pagothenia borchgrevinki (Boulenger) showed a heart rate of 11 beats min^-1 while maintaining a cardiac output estimated to be about 30 ml min^-1 kg^-1. Total vascular resistance was low (Axelsson et al. 1992). Pagothenia borchgrevinki has a low resting haematocrit, which increases substantially when demand for increased oxygen transport is induced by various types of stimuli, including handling, exercise and hypoxia (Macdonald et al. 1987; Davison et al. 1988; Franklin et al. 1993). The low resting haematocrit contributes to the low vascular resistance as it reduces blood viscosity, a major factor at low temperatures (Farrell, 1984; Macdonald and Wells, 1991; Axelsson et al. 1992).

The low heart rate depended largely on a tonic vagal cholinergic influence, with atropine producing a doubling of the heart rate. Atropine also produces a marked decrease in ventral aortic blood pressure (Axelsson et al. 1992), which could be interpreted as indicating a tonic cholinergic vasoconstrictor influence on the branchial vascular resistance (see Smith, 1977; Pettersson and Nilsson, 1979). However, as the study of Axelsson et al. (1992) did not measure dorsal aortic blood pressure, separate effects on the gill vasculature and systemic vasculature could not be determined.

The role of the sympathico-adrenal system in the regulation of blood pressure in vertebrates has received considerable attention. In teleost fish, circulating catecholamines, released from both chromaffin tissue and the adrenergic neurones of the autonomic nervous system, have been implicated as the cause of an adrenergic tonus, affecting the heart rate and vascular resistance both at rest and during exercise (e.g. Randall and Stevens, 1967; Cameron, 1979; Smith et al. 1985; Axelsson and Nilsson, 1986). In addition, there is now evidence for an involvement of several other groups of substances in cardiovascular control in fish, including adenosine derivatives, serotonin (5-hydroxytryptamine), numerous neuropeptides and the renin–angiotensin system (Nilsson and Holmgren, 1992; Olson, 1992). Although we have evidence of a high inhibitory cholinergic tonus and weak excitatory adrenergic tonus on the heart of two species (Axelsson et al. 1992), our knowledge of cardiovascular control in Antarctic fish is very limited (Montgomery and Wells, 1993). Nothing is known of the effects of other potential control substances.

Most Antarctic fish are benthic and will not swim unless provoked, making them unsuitable for looking at methods of cardiovascular control during exercise (Egginton et al. 1991; Axelsson et al. 1992). Pagothenia borchgrevinki is a pelagic labriform swimmer, using its large pectoral fins for propulsion and the tail as a rudder. It will swim readily in a swimming tunnel (Forster et al. 1987; Davison et al. 1988), making it ideal for cardiovascular studies. The present study is an investigation of patterns of blood pressure regulation in P. borchgrevinki at rest and during non-exhaustive swimming.

The cryopelagic red-blooded Antarctic nototheniid fish Pagothenia borchgrevinki (Boulenger) (commonly called ‘borch’) was caught by hook and line through holes cut in the fast ice covering McMurdo Sound, Antarctica. Fish were transferred to Scott Base (New Zealand Antarctic Programme) and kept in plastic tanks at –l to +1 °C. They were subsequently flown to Christchurch, New Zealand, where this study was conducted. Fish were maintained in a closed-circuit aquarium system at –0.5 °C until needed.

Fish of either sex (body mass 82.5–198 g, body length 21.5–27 cm; N=24) were anaesthetized in sea water containing benzocaine, 100 mg l^-1, until breathing movements stopped, and transferred to an operating table in a cold room (0 °C). A steady flow of sea water at 0 ° to +2 °C containing approximately 25 mg l^-1 benzocaine was maintained over the gills throughout surgery.

Fish were placed ventral side up in an operating tray, and a ventral incision was made to expose the ventral aorta. A cuff-type single-crystal Doppler flow probe (cuff diameter 1.5 mm) was fitted around the ventral aorta, and the leads were secured by skin sutures. The fish was placed on its side, and a polyethylene cannula (PE50, pulled to provide a smaller diameter at the tip, and slightly bent to fit the curve of the artery) filled with heparinized (approximately 100 i.u. ml^-1) 1.4 % NaCl was inserted occlusively into the efferent branchial artery of the third gill arch and pushed forward to about the level of the suprabranchial artery. This cannula was used to record dorsal aortic blood pressure (P_DA) and to inject drugs. The cannula was tied to the gill arch and secured by two skin sutures. A second similar cannula (PE50 tipped with about 1 cm of PE10) was inserted into the afferent branchial artery of the same gill arch and secured by two skin sutures.

Following surgery, the fish were returned to fresh sea water to recover for at least 22 h before any experiments were conducted.

The cannulae were attached to Honywell model 156PC06GW2 disposable pressure transducers calibrated against a static column of water. The transducers were connected to Grass preamplifiers, and a tachograph preamplifier triggered by the pulsatile blood pressure signal was used to obtain heart rate (f_H) data. Blood flow in the ventral aorta (cardiac output) was recorded using a Doppler flowmeter system (Iowa University). This system allows accurate measurement of blood velocity; previous studies have demonstrated a linear relationship between blood velocity and instantaneous volume flow (e.g. Axelsson et al. 1990; Axelsson and Fritsche, 1991; Sundin and Nilsson, 1992). However, the size of the fish and the placement of the flow probes do not allow reliable and regular posthumous calibration of volume flow, and variations in cardiac output (Q̇) are therefore retained as ΔkHz Doppler shift and presented as percentage changes compared with initial values.

Pressure traces from the ventral and dorsal aortas (P_VA and P_DA) and cardiac output (Q̇) were displayed and recorded using a Yokogawa model 3701 LR8100 recorder, and all data (including f_H) were continuously sampled by a Toshiba model 5200 PC-based data-acquisition system (AD-DATA, Dr P. Thorén, Department of Physiology, Göteborg).

Individual fish (N=12) were placed in a Blazka-type swim tunnel, which allowed a reproducible work load to be set. After recovery from surgery, each fish was exposed to a 5 min (control) swim trial at 20 cm s^-1 (0.8–1.0 body lengths s^-1). During recovery from this swim, once the cardiovascular variables had stabilized, a dose of adrenaline (5 nmol kg^-1) was injected via the dorsal aortic cannula, and the effect recorded. Doses of drugs injected were based on the previous study (Axelsson et al. 1992). When the cardiovascular variables had again stabilized, the α-adrenoceptor antagonist prazosin (1 mg kg^-1) was injected and allowed to act for 1 h. When the recorded variables had stabilized, a second dose of adrenaline (5 nmol kg^-1) was injected to check the integrity of the α-adrenoceptor blockade. After a further 30 min, the fish was again exercised for 5 min.

To investigate a possible role of the angiotensin system in blood pressure control in P. borchgrevinki, angiotensin I (Ang I; from salmon; 1 nmol kg^-1) was injected into the fish and, when the pressor response to this drug had subsided, a dose of the angiotensin converting enzyme inhibitor enalapril (1 mg kg^-1) was injected. The effects of a second dose of Ang I were then recorded, and finally the response to angiotensin II (Ang II; human; 1 nmol kg^-1) was recorded. The effects of the angiotensins were initially tested over the range 0.1-100 nmol g^-1, with substantial differences in the magnitude of the response being noted in individual fish. A dose of 1 nmol g^-1 was chosen since most fish showed consistent blood pressure responses to this dose. Seven of the twelve fish in the group maintained patent cannulae and flow signals and swam for the full programme.

A second group of fish (N=12) was subjected to a 5 min swim trial as described for group I. After recovery from this control swim period, each fish was injected with Ang II (1 nmol g^-1). Although the group I fish had also been injected with Ang II, they received it after prior treatment with a number of other drugs, which might have influenced the response. In group II fish, 10 min after the injection of Ang II, atropine (1.0 mg kg^-1) was injected. One hour later, a second swim trial was carried out. The -adrenoceptor antagonist sotalol (3 mg kg^-1) was then injected, and 1 h later the fish swam one last time. The full protocol was completed in eight of the twelve fish.

The following drugs were used in this study: atropine base (neutralized using HCl) (Sigma); L-adrenaline base (neutralized using HCl) (Sigma); [Asn¹, Val⁵, Asn⁹]-angiotensin I, salmon, (Sigma); angiotensin II, human (Sigma); enalapril (Renitec, MSD); prazosin (Sigma) and sotalol hydrochloride (Hässle AB).

Cardiovascular variables were continuously recorded using the Yokogawa recorder, and data were also sampled using the computerized data-acquisition system for later retrieval and calculations. Mean values were created at 30 s intervals from continuous sampling in 5 s blocks at 5 samples s^-1, and data presented show means ± S.E.M. for N animals. During swimming exercise, the tachometer-derived f_H signal was irregular as a result of the movement of the animal. Therefore, this variable and stroke volume (V_S), calculated from f_H, were omitted for the swimming period in Figs 2 and 3.

Branchial and systemic vascular resistances (R_gill and R_sys, respectively) were calculated as (P_VA-P_DA)/Q̇ and P_DA/Q̇, respectively, assuming that all the cardiac output passed through the systemic vessels. Stroke volume was determined as Q̇/f_H. In all cases, Q̇ was expressed as ΔkHz Doppler shift, and for presentations these values have been converted to percentage changes.

Evaluation of statistically significant differences (P⩽0.05) in the observations was made using the Wilcoxon signed-ranks test. The sequentially rejective Bonferroni test described by Holm (1979) was used to eliminate, as far as possible, the possibility of discarding any true null hypothesis.

Resting values for P_VA (3.58±0.13 kPa; N=24) were similar to those described in the previous study (Axelsson et al. 1992), although mean f_H was considerably higher at 20.6±0.9 beats min^-1. Resting P_DA was 2.77±0.10 kPa; N=24. Resting branchial vascular resistance (R_gill) was approximately 25 % of the systemic vascular resistance (R_sys).

Prior to treatment with drugs, all of the fish in both groups swam steadily when the current speed was increased in the swim tunnel. During exercise at 20 cm s^-1, P_DA consistently fell by 10 %, while P_VA increased by about 18 %. Cardiac output increased dramatically by 78 %, as a result of increases in stroke volume of 36 % and in f_H of 33 %. The fall in P_DA was due to a decrease in R_sys, by more than 50 %, which could not be offset by the increase in Q̇. The branchial vascular resistance increased by about 30 % (Figs 1 and 2). Blood pressure and flow recovered to pre-swim values within about 1 h following the swimming period.

Adrenaline injection in group I fish produced marked elevation of both P_DA and P_VA. Cardiac output increased substantially as a result of an increase in V_S, while f_H remained largely unchanged. R_gill decreased slightly, while changes in R_sys, if any, were transient (see Fig. 4).

After treatment of the fish in group I with the α-adrenoceptor antagonist prazosin (1 mg kg^-1) there was a decrease in resting P_DA due to a decrease in R_sys (Table 1). Stroke volume increased by 73.3 %. Only seven fish were able to swim after the prazosin treatment. In these fish, there was a decrease in P_DA similar to that found in the control group. In contrast to the control fish, there was an initial drop in P_VA, which, however, returned towards pre-swim values before the swim had ended. Cardiac output increased, entirely because of an increase in f_H, since V_S did not change from its new, high level. Recovery of R_sys and P_DA following exercise was somewhat slower than in the drug-free fish (Fig. 3).

Following prazosin treatment, a second adrenaline injection still produced increases in P_DA and P_VA, although they were less pronounced and they were associated with smaller increases in V_S and Q̇. However, R_sys and R_gill did not change (Fig. 4).

Ang II produced consistent increases in R_sys, while R_gill decreased slightly. In fish that had not had any previous treatment with drugs (group II), P_DA and P_VA increased after Ang II injection (P_DA from 2.79±0.11 to 3.98±0.14 kPa; P_VA from 3.72±0.14 to 4.81±0.16 kPa) with peak pressure at 2–3 min (Fig. 5). This initial rise in P_DA and P_VA occurred without a change in Q̇ as bradycardia compensated for a rise in V_S. In the group I fish, the rise in pressure in response to Ang II was slowed somewhat (peak after more than 5 min), but here Ang II was given after pretreatment with a number of other drugs, including prazosin (Fig. 6).

The hypertensive effect of Ang I was abolished by pretreatment with the angiotensin converting enzyme (ACE) inhibitor enalapril. Ang I had a similar hypertensive effect to Ang II, but without the marked bradycardia associated with Ang II injection (Fig. 6). The pressor response was slower to develop. Both Ang I and Ang II injections caused an immediate increase in V_S.

A striking effect in the group II fish was the bradycardia induced by Ang II injection (Fig. 5). Atropine injection changed this to a tachycardia, f_H being 35 % greater than the pre-Ang II injection values. Following atropine injection, f_H remained high and regular (Figs 3, 5). Atropine injection produced a substantial decrease in resting R_sys and, despite an increase in Q̇ (due to elevated f_H), P_DA decreased (Table 1).

The increase in P_VA and R_gill observed during swimming in the control animals was absent in atropinized fish. The increase in Q̇ was due almost entirely to an increase in V_S, although a small increase in f_H did occur (Fig. 3).

When the atropinized fish were treated with sotalol, resting f_H fell (Table 1). Sotalol restored the exercise-induced increase in P_VA, in association with an increase in R_gill (Fig. 3).

An incidental observation of spleen status was made during the later part of this study in the freshly killed atropinized and atropine/sotalol-treated fish. These fish showed dark blood-engorged spleens weighing 0.68±0.17 % of body mass (N=4).

The surgical procedure used in the present study involves occlusion of one gill arch to allow access to both afferent and efferent branchial arteries for blood pressure recording. The procedure is not vastly different from that used in a previous study of the same species under otherwise nearly identical conditions, yet the heart rate appears to be substantially higher in the present work (mean f_H was 20.6 beats min^-1 compared with 11.3 beats min^-1; Axelsson et al. 1992). One difference between the two studies, apart from the slight difference in cannulation procedure, was the use of MS222 (tricaine) (Axelsson et al. 1992) or benzocaine (present work) anaesthesia. In a parallel study of the benthic nototheniid Trematomus (Pagothenia) bernacchi, a similarly high f_H was observed, compared with the values obtained by Axelsson et al. (1992). Switching to MS222 anaesthesia had no effect on the high heart rate in that species. The other difference between the two studies was that the fish in the present study were held in the aquarium system in Christchurch for a considerably shorter period (less than 3 weeks in this study compared with more than 6 weeks in the previous). The range of heart rates observed in the present study was large, indicating that, in P. borchgrevinki, small differences in experimental protocol may cause variations in the cholinergic tonus regulating f_H (Axelsson et al. 1992). Any of these three factors (surgical procedure, anaesthesia and holding time), alone or in combination, may be responsible for the discrepancy in the resting f_H but, until further analysis can be made, we cannot offer a final explanation to the observed difference.

Swimming produced an increase in Q̇ resulting from both tachycardia and increased V_S. Despite this, swimming fish showed a clear and remarkably consistent decrease in P_DA resulting from a marked drop in R_sys. All swimming episodes produced a drop in R_sys. The α-adrenoceptor antagonist prazosin lowered resting P_DA, demonstrating the existence of an α-adrenergic tonus affecting resting systemic blood pressure in borchs. Atropine also reduced R_sys, indicating a cholinergic vasomotor tonus, the function of which is clearly in need of further investigation. With more than one possible controlling system affecting systemic resistance, it is perhaps not surprising that R_sys fell even further after exercise in both the prazosin-and atropine-treated fish.

Prazosin reduced the increases in R_gill and V_S seen during exercise. In addition to the well-known -adrenoceptor-mediated vasodilation of the arterio-arterial pathway in teleosts, adrenergic nerve stimulation and exogenous catecholamines produce an α-adrenoceptor-mediated vasoconstricion of the arterio-venous pathway and efferent lamellar arterioles (Nilsson and Pettersson, 1981; Pettersson, 1983; Laurent, 1984; Nilsson, 1984). This vasoconstrictor mechanism appears to be present in borchs. The significance of this α-adrenoceptor-mediated vasoconstricion is not fully understood, but an increased lamellar recruitment has been suggested which, in turn, may facilitate oxygen uptake (Pettersson and Johansen, 1982; Pettersson, 1983).

In addition to the adrenoceptor-mediated vasomotor control of the branchial vasculature, there is good histochemical evidence from several teleost species for cholinergic vasoconstrictor innervation of the sphincter at the base of the efferent filamental artery (Bailly and Dunel-Erb, 1986; Dunel-Erb et al. 1989). Physiological studies corroborate the structural findings, showing a functional cholinergic vasoconstrictor control of the filamental sphincters (Smith, 1977, 1978; Pettersson and Nilsson, 1979; Nilsson and Pettersson, 1981; Nilsson, 1984). In the present study, atropine reduced or abolished the branchial vasoconstriction seen during exercise, suggesting the presence of a similar cholinergic vasoconstrictor control in borchs. A β-adrenoceptor-mediated vasodilator mechanism may also be present, since the branchial vasoconstrictor response to exercise, which was abolished in atropinised fish, returned when they were treated with the β-adrenoceptor antagonist sotalol. However, treatment with prazosin and adrenaline did not reduce R_gill, which may indicate that there are few -adrenoceptors in the gill. It would appear that the branchial vasculature of the borchs is mainly influenced by cholinergic and α-adrenergic vasoconstrictor mechanisms. In the present study, we did not test the effects of adrenaline after β-adrenoceptor blockade, but the apparent dominance of α-adrenoceptor mechanisms over β-adrenoceptor effects on the cardiovascular system of these cold-adapted fish is interesting. In teleost fish from temperate latitudes, β-adrenoceptor-mediated responses are dominant in summer, while α-adrenoceptor-mediated responses are relatively more important in winter (Randall and Perry, 1992).

Adrenaline injection produced marked increases in P_VA, P_DA and Q̇. f_H remained largely unaffected, indicating that the change in Q̇ was due entirely to an increase in V_S. In contrast to most other teleosts and, indeed, to most vertebrates, the hypertension produced by adrenaline was caused almost entirely by an elevated Q̇, rather than by increased vascular resistance. After prazosin treatment, part of the blood pressure response to adrenaline persisted, and the main action of prazosin appeared to be to weaken the effect of adrenaline on V_S. It is not clear whether the action of prazosin on the heart is due to nonspecific effects or whether the adrenergic control of the borch heart involves α-adrenoceptors.

The renin–angiotensin system has been shown to possess strong effects on blood pressure in teleost fish and may be a major ‘anti-drop’ factor in teleost blood pressure control (Taylor, 1977; Bailey and Randall, 1980; Nishimura and Bailey, 1982; Gray and Brown, 1985; Olson, 1992; Platzack et al. 1993). As in other vertebrates, Ang II acts directly on the teleost vasculature and a substantial part of its vasomotor effect may be caused by secondary release of catecholamines (Nishimura et al. 1978; Carroll, 1981; Carroll and Opdyke, 1982; Olson, 1992). Ang II produced a consistent increase in both P_VA and P_DA as a result of a dramatic elevation in R_sys. In the experiments in group I, only the direct effects of angiotensin were recorded, since the α-adrenoceptors had been blocked by prazosin.

The blood pressure rise after Ang I injection suggests the presence of endogenous angiotensin converting enzyme (ACE), which is necessary for the conversion of Ang I to the vasoactive Ang II. Treatment with the ACE inhibitor enalapril impaired or abolished the effects of Ang I, while Ang II was still active. Treatment with Ang I did not produce the large increase in R_sys and the bradycardia seen in response to Ang II. The slow development of pressure following Ang I injection indicates that Ang II was being produced fairly slowly, allowing other mechanisms to keep R_sys and f_H constant. The 10–20 % increase in V_S following injection of a number of drugs, including Ang I and enalapril, may well be an artefact of volume loading (Figs 4 and 6). It can, therefore, be concluded that P. borchgrevinki possess vital elements of a renin–angiotensin system and that these mechanisms are part of the system controlling blood pressure in the borchs.

In some of the fish in group II, the bradycardia observed after Ang II injection was spectacular, with f_H reaching levels as low as 3–4 beats min^-1. The bradycardia was due to a cholinergic vagal barostatic reflex, since f_H increased dramatically after atropine treatment (Fig. 3). The marked barostatic reflex was not observed when adrenaline was used to increase blood pressure, where it is possible that the inhibitory effect of a vagal reflex was offset by a direct adrenergic stimulation of the heart. Certainly, sotalol treatment reduced heart rate in atropinised fish.

Incidental observation of spleen status in four freshly killed atropine-and atropine/sotalol-treated fish, showed a mean spleno-somatic index of 0.68±0.17 %. This is somewhat higher than that observed in a previous systematic study (0.50±0.21 % in resting controls; Franklin et al. 1993). In contrast to that of other vertebrates, the teleost spleen is controlled in part by cholinergic excitatory nerve fibres (cf. Nilsson and Grove, 1974). In view of this, a swelling of the spleen due to sequestering of erythrocytes could be expected to occur after atropine treatment. However, no systematic evaluation of atropine effects on the borch spleen was made in the present study, and the observed increase needs further clarification.

We conclude that P. borchgrevinki regulates its blood pressure more through modulation of cardiac performance (cardiac output) than by changing vascular resistance. This may be a consequence of total vascular resistance being unusually low in borchs compared with most teleost species (Axelsson et al. 1992). The scope of heart rate control is large: from as low as 3–4 beats min^-1 in individual fish during hypertensive bradycardia to about 28 beats min^-1 after atropine treatment. During exercise, a consistent decrease in R_sys occurred, possibly because of a reduction in a cholinergic as well as an α-adrenergic tonus or because of an exercise-related ‘active hyperaemia’ (Randall and Daxboeck, 1982). The prerequisites for angiotensin-related vasomotor control mechanisms appear to be present.

This work was supported by grants from the New Zealand Antarctic Programme (NZAP), the Swedish Natural Science Research Council (NFR) and the Swedish Forestry and Agriculture Research Council (SJFR). We want to express our most cordial thanks to the staff of Scott Base, Antarctica, for their friendship and support.