Plasma Catecholamines in the Lesser Spotted Dogfish and Rainbow Trout at Rest and During Different Levels of Exercise

Tight control of the circulatory and respiratory systems during an increased demand for oxygen is of paramount importance for all animals and yet the details of this control during exercise are poorly understood even for mammals (Whipp & Ward, 1982, 1985), let alone for other vertebrates. When swimming at maximum sustainable speed (critical swimming speed) at 10 °C there is approximately an eightfold increase in oxygen uptake in the rainbow trout (Kiceniuk & Jones, 1977), and in spontaneously swimming dogfish at 18 °C there is almost a doubling of oxygen consumption (Piiper, Meyer, Worth & Willmer, 1977). It would seem reasonable that the nervous system is involved, at least to some extent, in initiating and/or maintaining the respiratory and cardiovascular responses that occur in fish during exercise (see Kiceniuk & Jones, 1977) in order to deliver the required oxygen to the active muscles. There is, however, little or no sympathetic innervation of the heart and systemic blood vessels in elasmobranchs (Gannon, Campbell & Satchell, 1972; Opdyke, McGreehan, Messing & Opdyke, 1972), and no innervation of the branchial blood vessels at all (Metcalfe & Butler, 1984b). So, in these fish at least, any adrenergic control of the cardiovascular system during exercise will be by way of circulating catecholamines released from the chromaffin tissue.

Ever since the work of Krawkow (1913) on pike and Keys & Bateman (1932) on eels it has been known that catecholamines cause overall vasodilatation in the gills of teleost fish, a response that would seem ideally suited to the increased oxygen demand of exercise. This effect of adrenaline and noradrenaline on the branchial blood vessels has been verified for teleosts (see, for example, Ostlund & Fange, 1962; Rankin & Maetz, 1971 ; Wood, 1974; Wahlqvist, 1980) and demonstrated to occur in elasmobranchs (Davies & Rankin, 1973: Metcalfe & Butler, 1984 b). Catecholamines also have positive inotropic and chronotropic effects on the hearts of teleosts (Holmgren, 1977; Cameron & Brown, 1981) and elasmobranchs, although in the latter noradrenaline has a negative chronotropic effect (Capra & Satchell, 1977).

In 1967 it was demonstrated by Nakano & Tomlinson that plasma adrenaline and noradrenaline increased 26-fold and eightfold, respectively, in rainbow trout in response to ‘physical disturbance’. This disturbance consisted of forcing the fish to swim vigorously ‘by repeatedly grasping them by the tail’ and the fish were close to exhaustion after 15 min of such activity. Largely as a result of their paper, however, it has been claimed that increased levels of circulating catecholamines occur during exercise in fish (elasmobranchs as well as teleosts) and that they may play an important role in the cardiovascular adjustments to exercise (Randall, 1970, 1982). Indeed, some authors have used a similar technique to that of Nakano & Tomlinson (1967) to induce activity yet have still referred to the effect of exercise on release of catecholamines (Opdyke, Carroll & Keller, 1982). In a recent review, though, Wood & Perry (1985), while attributing many of the compensatory responses that occur in fishes during exercise to circulating catecholamines, point out that their evidence is ‘circumstantial rather than direct’ as the levels of plasma catecholamines have rarely been measured during the ‘experimental treatments’.

In the present study the levels of plasma adrenaline and noradrenaline in an elasmobranch and a teleost were measured during different degrees of activity: at rest; while swimming either spontaneously or at different speeds in a water channel; and during repeated burst swimming. These levels of catecholamines were then compared with the concentrations used in other studies to produce changes in branchial blood flow, or to produce effects on gas transfer in isolated head preparations, and with the doses used to produce effects on gas transport in the whole animal.

Data were obtained from eight lesser spotted dogfish, Scyliorhinus canicula, of either sex and of mean mass (±S.E. of mean) 0·79 ± 0·024kg, and from 25 female rainbow trout, Salmo gairdneri, of mean mass 0·41 ± 0·015 kg and of mean length 33 ± 0·7 cm. The dogfish were obtained either from the Plymouth laboratories of the Marine Biological Association of the UK or from the Marine Biology Laboratory, UCNW, Menai Bridge. They were transported in oxygenated sea water to aquaria in Birmingham where they were held in aerated, recirculated sea water, maintained at 15 °C, for at least 2 weeks prior to experimentation. They were fed periodically on sprats obtained from a local fishmonger. The rainbow trout were obtained from Bibury Trout Farm, Gloucestershire, and transported in oxygenated water to aquaria in Birmingham where they were held in aerated, dechlorinated Birmingham tap water maintained at 12°C for at least 2 weeks prior to experimentation. They were fed once a day on trout pellets.

Attempts to train the dogfish to swim in a water channel, similar in design to that described by Johnston & Moon (1980), were unsuccessful. They merely angled their pectoral fins so that the water flowing over them generated a downward force, thus enabling the fish to remain motionless on the bottom. Even placement beneath the fish of a mesh inclined at an angle to the water flow was not effective. It was decided, therefore, to record data from dogfish during bouts of spontaneous swimming activity which occur predominantly at night (Metcalfe & Butler, 1984a). The ventral aorta was cannulated under MS 222 anaesthesia using a modification of the method described by Smith & Bell (1964). A length of polyethylene tubing (i.d. 0·76 mm, o.d. 1·22 mm) was positioned in the blood vessel with the aid of a stainless steel wire sharpened to a point at the end (Soivio, Nyholm & Westman, 1975). The cannula was led out through the spiracle and the fish left at least 24 h in an actograph (Metcalfe & Butler, 1984a) before measurements were taken. It was possible to take a mixed venous blood sample via this cannula in resting fish and immediately after periods of spontaneous swimming activity, as indicated by the actograph. Unfortunately, there was so much movement artefact on the pressure trace during swimming that it was not possible to record heart rate or blood pressure.

The rainbow trout would swim in the water channel. Water velocity could be varied between 0 and 0·9ms⁻¹ and was measured by a Braystoke BFM 002 current flowmeter. The dorsal aorta of the fish was cannulated under MS 222 anaesthesia, as described above for the dogfish. Each fish was left in the water channel overnight to recover from the anaesthesia and handling and swam for 30 min at approximately 1 body length per second (1 Ls⁻¹) on 3–4 occasions during the next 2 days. If the fish rested against the back grid during the swimming bouts it was encouraged to swim again by gently tapping the grid. Covering the anterior portion of the swimming chamber while brightly illuminating the rear also encouraged the fish to continue swimming. After each swimming period the fish was left for at least 6 h at rest in the water channel. The dechlorinated water in the channel was kept well-acrated and at 12 °C. On the fourth day, resting values of the measured variables were taken before the water flow in the channel was increased, either to a velocity approximately equivalent to ILs⁻¹ and left there for 30 min or progressively at increments of 10cms⁻¹ every 10 min until the fish ceased to swim and rested against the posterior grid. The only attempt that was made to encourage the fish to swim during the latter experiment was to reduce the water velocity slightly after the fish had stopped swimming. Invariably it began to swim again and the water velocity was then returned to its former value. This usually gave sufficient time for a blood sample (approximately 1·2ml) to be taken before the fish stopped swimming again. The fish performed little or no burst swimming during this procedure and apparent exhaustion occurred at approximately 2Ls⁻¹. This compares with a critical swimming speed of 0·5–l·5Ls⁻¹ for the (somewhat larger) instrumented rainbow trout used by Kiceniuk & Jones (1977). Blood pressure was measured continuously.

At least 1 h after swimming, either spontaneously (dogfish) or in the water channel (trout), the fish was induced into burst swimming by repeatedly touching its tail for a period of 2–3 min. This was sometimes sufficient to exhaust the fish, inasmuch as touching the tail no longer caused it to swim. However, the fish were not routinely completely exhausted by this procedure. A final blood sample was then taken.

Blood pressure was measured in the trout by a Druck pressure transducer and recorded on an Ormed pen recorder writing on rectilinear coordinates. Heart rate could be determined from the traces. Partial pressures of oxygen in the blood and in water and pH in trout blood were measured by an appropriate Radiometer electrode and a PHM71 blood gas analyser. Haematocrit was determined with a Hawksley microhaematocrit centrifuge. Plasma lactate was assayed enzymatically using a Sigma kit (no. 826-UV) and a Beckman 25 spectrophotometer, measuring at 340 nm.

The levels of adrenaline and noradrenaline in plasma were measured using reverse phase, ion-pair, high-performance liquid chromatography with electrochemical detection from 500μl plasma samples following extraction of the catecholamines with alumina (Hallman, Farnebo, Hamberger & Jonsson, 1978). Plasma samples were obtained from approximately 800 μl of whole fish blood following 1–2 min centrifugation (Beckman microfuge B) to precipitate the red blood cells. Plasma samples were quickly frozen and kept at below –20 °C for no longer than 7 weeks before analysis (Butler, Taylor & Davison, 1979).

Subsequently 250 μl of Tris/EDTA buffer (1·5 mol l⁻¹ Tris, 0·48 mol l⁻¹ EDTA, pH 8-6) and 26 mg of acid-washed alumina (AAO, Bioanalytical Systems) were added to each 500 μl plasma sample. The mix was briefly vortex-mixed for 30–40 s and then gently shaken for 5 min. The mix was centrifuged for 1 min to precipitate the AAO and the supernatant aspirated off. The AAO was washed twice, each time with 500 μl of 1/100 Tris buffer, gently shaken, centrifuged and the supernatant aspirated off. Finally the catecholamines were eluted in 300μl of perchloric acid/EDTA (0·1 moll⁻¹ HClO₄, 1·0 mmoll⁻¹ EDTA). This mix was briefly vortex-mixed, shaken for 5 min and centrifuged. 100 μl of the eluent was injected onto the analytical column. Recovery was calculated using paired plasma samples. One sample of each pair received 50pmol of catecholamine (‘spiked’); the other did not (‘unspiked’). Percentage recovery was calculated as: (‘spiked’-’unspiked’/50) × 100. Recovery was always better than 75 %.

Separation of the catecholamines in the perchloric acid eluent was performed on a reverse phase, ion-pair liquid chromatography column (Ultrasphere IP, 250 mm × 4·66mm) using an aqueous mobile phase (90% 0·1 moll⁻¹ KH2PO4, pH3·0, containing 37 mg EDTA and 47 mg sodium octyl sulphonic acid per litre, 10% methanol) pumped (Altex 110A) at a flow rate of l·2mlmin⁻¹. Catecholamines were measured using electrochemical detection (Bioanalytical Systems LC-4A) at +0·72 V using a carbon paste working electrode. The output from the detector was displayed on a chart recorder (Gilson). This system was sensitive enough to detect catecholamine levels as low as l–2 × 10⁻⁹moll⁻¹.

Mean values are expressed ±S.E. of mean and compared using a paired t-test (Bailey, 1981). The term ‘significant’ refers to the 95% confidence level.

Mean values of all the measured variables are given in Table 1 and changes in the levels of plasma catecholamines are illustrated in Fig. 1.

The resting values from the two groups of trout (i.e. those swimming at 1 Ls⁻¹ and those swimming to apparent exhaustion) were not significantly different from each other, so they have been combined. Noradrenaline was the predominant catecholamine in the blood of both species of fish at a concentration of approximately 10⁻⁸moll⁻¹. In resting dogfish the concentration of adrenaline was 42% that of noradrenaline, whereas in trout it was 14% of the level for noradrenaline. In both animals, plasma lactate was at a concentration of approximately 10⁻³ moll⁻¹.

During spontaneous exercise in the dogfish, which lasted for a mean duration of 7·5 ± 2·5 min, noradrenaline increased to 2·3 times the resting value while adrenaline rose to 3⁻3 times resting. In trout swimming at ILs⁻¹ for 30min there were reductions in plasma adrenaline and noradrenaline to approximately 25% of their resting values. In neither species was there a significant change in the level of plasma lactate under these swimming conditions. In trout there were no significant changes in arterial pH and dorsal aortic blood pressure but there were significant increases in arterial and in the heart rate. When trout swam to apparent exhaustion, at approximately 2Ls⁻¹, there was a 2·2-fold increase in plasma noradrenaline and a tenfold increase in plasma adrenaline above the resting values. The levels reached when trout swam to apparent exhaustion were similar to those recorded in spontaneously swimming dogfish. Plasma lactate, however, was 2·6 times the resting value in trout swimming to apparent exhaustion, although arterial pH was similar to the resting value. This could be, at least in part, the result of a respiratory alkalosis counteracting the potential metabolic acidosis. Arterial was similar to that recorded in fish swimming steadily at 1 L s⁻¹, whereas heart rate was slightly (but not significantly) higher and dorsal aortic blood pressure was significantly lower. Haematocrit was some 12–15 % below the resting value in both groups of swimming trout whereas in spontaneously swimming dogfish haematocrit was the same as in resting fish.

It was only when the fish were close to physical exhaustion as a result of repeated burst swimming that plasma catecholamines increased by any substantial amount. In dogfish, this increase was to approximately seven and 16 times the resting value for noradrenaline and adrenaline, respectively. In trout, noradrenaline increased to eight times the resting value whereas adrenaline increased to an impressive 150 times resting. The highest levels of plasma catecholamines were recorded in one particular trout. Plasma adrenaline reached 1·2 × 10⁻⁶moll⁻¹ and noradrenaline was 0·8 × 10⁻⁶molI⁻¹ after repeated burst swimming. In trout close to physical exhaustion, plasma lactate was 4·4 times the resting value and arterial pH was significantly below the resting value, indicating an uncompensated metabolic acidosis. Arterial ⁠, however, was significantly above the resting value and similar to the values recorded during steady swimming at 1 L s ⁻¹ and when swimming to apparent exhaustion at approximately 2Ls⁻¹.

When trying to measure the levels of hormones that may be released in response to exercise, it is important that the resting animal is as unstressed as possible. The resting levels of plasma adrenaline and noradrenaline found here for the dogfish are similar to those reported by Opdyke et al. (1982) but are somewhat lower than those reported previously for fish left 72 h after implantation of the ventral aortic catheter (Butler et al. 1979). This may relate to the less traumatic method used for inserting the catheter in the present experiments. In resting trout, the observed level of plasma adrenaline is similar to that reported by Primmett, Randall, Mazeaud & Boutilier (1986) but somewhat lower than that reported by Woodward (1982). On the other hand, noradrenaline is several times higher than that measured by Woodward and by Primmett et al.

It is interesting to note that both adrenaline and noradrenaline decreased from the resting value in trout swimming at ILs⁻¹ for 30 min. This could be the result of reduced production of these hormones during sustained swimming (i.e. the fish were less stressed) and/or of increased removal of the hormones from the blood, particularly by the gills. The catecholamines were measured in blood from the dorsal aorta of trout in the present study, and it has recently been demonstrated for rainbow trout (Nekvasil & Olson, 1985) and for Atlantic cod (Ungell, 1985) that the gills are able to remove noradrenaline from the blood (to a greater extent than adrenaline in cod and apparently selectively in trout). As both adrenaline and noradrenaline decreased by similar proportions from the resting values in trout during sustained swimming, increased removal from the blood by the gills would not seem to be an adequate explanation. Unfortunately, we were not able successfully to cannulate the ventral aorta in the trout or the dorsal aorta in the dogfish using techniques involving little or no surgery, so it was not possible to determine whether differences in concentrations of catecholamines do exist in venous and arterial blood and whether these concentrations vary independently during exercise. This is clearly a point worthy of further study.

In both species of fish, two blood samples were taken before the final procedure, amounting to less than 2ml for the dogfish and approximately 2·5 ml for the trout. If blood volume/body weight ratio for dogfish is 7% (Satchell, 1971) then no more than 4% of total blood volume was removed before the final procedure and this would not be expected to have any significant effect on haematocrit and hence on oxygen-carrying capacity of the blood. In trout, blood volume is approximately 5·6% of body weight (Milligan & Wood, 1982) so that about 12% of total blood volume was removed before the final procedure. This could well, therefore, have had a significant effect on haematocrit and on oxygen-carrying capacity of the blood in these fish (cf. Soivio et al. 1975). The relatively large amount of blood removed from the trout could, therefore, at least partly explain the reduced haematocrit during swimming in both groups of these fish, a trend which was not seen in the larger fish used by Kiceniuk & Jones (1977).

Resting plasma lactate values were similar in both species to those measured by other authors (Heisler, 1984). Arterial and haematocrit of resting fish were somewhat less than those measured by Kiceniuk & Jones (1977) for larger rainbow trout, but were almost identical to those measured by Primmett et al. (1986) for fish of similar size to those used in the present study. The fish used were, therefore, in similar physiological condition to those used by other groups of workers.

The levels of catecholamines measured during the different forms of exercise can be related to the concentrations that are required to cause physiological effects on the gills and gas transporting system of the particular species of fish. Studies on elasmobranchs are sparse. Davies & Rankin (1973) looked at the effects of adrenaline and noradrenaline on the flow of filtered saline perfusate through the isolated fifth gill of the dogfish, Scyliorhinus canicula. Unfortunately, the exact concentrations of the catecholamines in the perfusates are not known, but the amounts of the drugs given were in approximately 1ml of perfusate (J. C. Rankin, personal communication). Thus, the molar concentrations were approximately 1000 times greater than the values given in the paper. On this basis, it would appear from the dose/response curves in fig. 3 of Davies & Rankin (1973) that the concentration of adrenaline present in the blood of the dogfish during spontaneous swimming in the present study has near maximum effect (80%) on the gill blood vessels. The level of noradrenaline present during spontaneous swimming in the dogfish has approximately 50% of maximum effect. Perfusion of the whole head of the rainbow trout with saline containing a colloid substitute for plasma proteins (polyvinyl pyrrolidone) enabled Wood (1974) to construct concentration/response curves for the effects of adrenaline and noradrenaline on the gill vasculature. The levels of plasma catecholamines present in rainbow trout swimming to apparent exhaustion in the water channel would have approximately 20% of maximum effect on the gill vasculature, while the levels present after repeated burst swimming would have 40% of maximum effect on the blood vessels in the gills. The apparent difference in the responses of the branchial blood vessels of the dogfish and trout to the levels of plasma catecholamines present during spontaneous or sustained swimming could well be related to the absence of adrenergic innervation in the former (Metcalfe & Butler, 1984b).

It is not really possible to predict the physiological significance of these levels of catecholamines from their in vitro effects on overall resistance of the gill vasculature. This is partly because other hormones such as cortisol are also released during sustained swimming activity in the trout (Zelnick & Goldspink, 1981), and these may potentiate the β-adrenergic effects of the catecholamines (Newsholme & Leech, 1983), and partly because we do not know the relationship between overall branchial vascular resistance and the effectiveness of gas exchange across the gills.

More detailed studies of the role of circulating catecholamines on gas transport have been made almost exclusively on teleosts and particularly on the rainbow trout. In resting trout in well-aerated water some 40% of the secondary lamellae are not perfused with blood (Booth, 1978) and even in those lamellae that are perfused, up to 20% of the blood may be in basal channels which are probably non-respiratory because of their greater distance from the water than the other lamellar blood spaces (Tuurala, Part, Nikinmaa & Soivio, 1984). Adrenaline causes interlamellar recruitment in vivo and in vitro (Booth, 1979; Holbert, Boland & Olson, 1979) and an increase in perfusion pressure (which could be mediated by catecholamines) causes inter- and intralamellar recruitment in vitro (Farrell, Daxboeck & Randall, 1979; Farrell, Sobin, Randall & Crosby, 1980). During intralamellar recruitment a greater proportion of blood perfuses the central lamellar spaces, i.e. there is a move away from the basal channels. There may also, therefore, be a thinning of the epithelium of the lamellae, thus reducing the diffusion distance between blood and water. It has been demonstrated in vitro that adrenaline increases the permeability of the gills of freshwater-adapted rainbow trout to butanol and water, and it has been stated that it has the same effect on oxygen (Isaia, 1984) although direct experimental evidence is lacking.

Thus, there are a number of ways in which plasma catecholamines could affect oxygen transfer across the gills. Indeed, addition of adrenaline to the perfusate does increase (Wood, McMahon & McDonald, 1978) and oxygen uptake (Perry, Daxboeck & Dobson, 1985) in isolated saline-perfused head preparations of rainbow trout. The concentrations used were 10^−s and 10⁻⁷moll⁻¹, respectively, the latter level being almost an order of magnitude greater than that recorded in the plasma of trout swimming to apparent exhaustion, but similar to the level reached following repeated burst swimming. Similarly, the initial concentration of adrenaline required to cause an increase in the concentration of oxygen in the arterial blood of intact trout is approximately 10⁻⁷moll⁻¹ (calculated from Nikinmaa, 1982). Even though adrenaline from a single injection disappears from the plasma very quickly (Ungell & Nilsson, 1979) there is still an effect, albeit reduced, on haematocrit 10 min after the injection (Nikinmaa, 1982).

With the possible exception, therefore, of increasing the concentration of the oxygen in arterial blood, the level of adrenaline in the plasma of rainbow trout swimming to apparent exhaustion is not likely to cause an increase in oxygen delivery to the active muscles by any of the mechanisms discussed above. However, if repeated burst swimming occurs, as a result of some sort of physical disturbance (perhaps being chased by a predator), then plasma catecholamines could well reach an effective level. This would be the classical ‘fight or flight’ response. During burst swimming, however, it is the white, anaerobic muscle fibres that generate the high level of power necessary for the quick darting movements. So, the functional significance of the increased levels of circulating catecholamines would not be to enhance the supply of oxygen to the locomotor muscles, at least not during the period of the exercise itself. Enhanced oxygen supply would, presumably, be of some importance after the period of burst swimming and, in addition to the mechanisms mentioned previously, hyperventilation may be sustained as a result of CO2 retention mediated by the elevated plasma catecholamines (Wood & Perry, 1985). These authors point to other areas in which circulating catecholamines may be important after exhaustive exercise, e.g. in controlling the release of lactate from the white muscle into the blood and in maintaining intracellular pH of the red blood cells, thus preventing a decrease in oxygen affinity of the haemoglobin.

Under experimental conditions it is necessary to use some form of physical contact with the fish to induce repeated burst swimming, so it is not possible to dissociate the effect of the physical contact from the effect of the burst activity itself on the release of the catecholamines. It is clear that the levels of plasma catecholamines are quite different between fish that have refused to swim any longer in a water channel and those that have performed repeated bouts of burst swimming, and the inference is that the difference is attributable to the different type of swimming activity. There is a high level of mortality in rainbow trout within 4–8 h of experimentally induced burst swimming for a period of 6 min (Wood, Turner & Graham, 1983). This raises the questions: do such levels of mortality occur following more naturally induced burst swimming of similar duration, or indeed do fish only rarely perform such repeated bouts of burst swimming under natural conditions, for example when they are negotiating currents during spawning migration? Even then they are able to rest at will.

This study was supported by the SERC. The authors wish to thank an anonymous referee for some particularly useful comments.