The Effects Of Repeated Physical Stress On The β-Adrenergic Response Of The Rainbow Trout Red Blood Cell

In response to severe acute stress, many teleost fish, including the rainbow trout (Oncorhynchus mykiss), release the catecholamine hormones adrenaline and noradrenaline into their circulation (reviewed by Randall and Perry, 1992). The elevation of circulating catecholamine levels is generally thought to promote an array of compensatory physiological and biochemical responses aimed at alleviating many of the deleterious consequences of acute stress. In rainbow trout, one of the most significant responses is the β-adrenergic activation of a red blood cell (rbc) Na⁺/H⁺ exchanger (reviewed by Nikinmaa and Tufts, 1989; Fiévet and Motais, 1991; Nikinmaa, 1992; Thomas and Perry, 1992). This response promotes the regulation of blood oxygen transport owing to the effects of Na⁺/H⁺ exchange activation on the intracellular levels of allosteric modifiers of haemoglobin oxygen-binding (Nikinmaa, 1992).

Activation of rbc Na⁺/H⁺ exchange commences with the binding of catecholamines to β-adrenoceptors on the cell surface. This interaction triggers, via the β-adrenergic signal transduction pathway, the formation of the second messenger cyclic AMP. The elevation of intracellular cyclic AMP levels ultimately initiates activation of the Na⁺/H⁺ exchanger via phosphorylation by cyclic-AMP-dependent protein kinase A (Guizouarn et al. 1993). The β-adrenergic response can be modified both by the degree of blood oxygenation (Motais et al. 1987) and by changes in the acid–base status of the blood (Nikinmaa et al. 1987; Borgese et al. 1987; Salama and Nikinmaa, 1989; Cossins and Kilbey, 1989). One well-documented mechanism underlying the acute modification of the β-adrenergic signal transduction pathway in fish is the rapid insertion/removal of -adrenoceptors from the rbc plasma membrane. But, since -adrenergic cell swelling is also enhanced by hypoxic conditions in vitro independently of intracellular cyclic AMP levels, regulation of the rbc adrenergic response is also likely to occur at the level of the Na⁺/H⁺ exchanger itself (Reid and Perry, 1994a).

Recent work in this laboratory has shown that elevation of plasma cortisol levels in rainbow trout can increase the size of the cytosolic receptor pool and that these additional receptors can be mobilized to the cell surface during an appropriate acute stress (e.g. hypoxia) to increase rbc adrenergic responsiveness (Reid and Perry, 1991). These observations were the basis for a model (Perry and Reid, 1993) in which, under chronic stress, elevated levels of cortisol (Schreck, 1990) ‘pre-adapt’ the rbc to respond to adrenergic stimulation during acute stress. However, chronic stress may also be associated with persistently elevated levels of catecholamines, especially if the stress consists of repetitive acute disturbances over the long term. In direct contrast to the effects of cortisol, increases in plasma catecholamine levels are known to reduce β-adrenoceptor numbers in rainbow trout (Gilmour et al. 1994) by a process termed down-regulation (e.g. Hausdorff et al. 1990). Therefore, during long-term repetitive stress in which both plasma cortisol and catecholamine levels are habitually elevated, rbc adrenergic responsiveness is likely to be influenced by the simultaneous and opposite effects of cortisol (up-regulation) and catecholamines (down-regulation). Thus, with this general background, the goal of the present study was to evaluate the effects of a 7-day period of daily physical disturbance on (i) the levels of plasma catecholamine hormones (adrenaline, noradrenaline), (ii) the initial steps of the rbc -adrenergic signal transduction pathway (cell surface β-adrenoceptor numbers, cyclic AMP accumulation) and (iii) Na⁺/H⁺ exchange activity (as estimated by cell water and Na⁺ accumulation).

Freshwater rainbow trout Oncorhynchus mykiss (Walbaum) were obtained from a local hatchery (Linwood Acres Trout Farm) and transported to the University of Ottawa, where they were maintained indoors in large rectangular tanks supplied with flowing and aerated City of Ottawa dechlorinated tap water at 10 °C (see Perry et al. 1992, for the chemical composition of the water). The fish used in these experiments consisted of individuals approximately 2 years of age weighing between 200 and 300 g (experimental N=124). All fish were fed daily to satiation with commercial salmonid pellets. Photoperiod was kept constant at 12 h:12 h light:dark. Fish were acclimated to these conditions for approximately 2 weeks prior to experimentation.

Fish were anaesthetised in a 0.1 g l^-1 solution of ethyl-m-aminobenzoate (MS 222; Sigma Chemical Company), adjusted to pH 7.5 with NaHCO₃, and placed onto an operating table to allow continuous retrograde irrigation of the gills with anaesthetic solution. To permit blood sampling, an indwelling cannula was implanted into the dorsal aorta (Soivio et al. 1975) using flexible polyethylene tubing (Clay-Adams PE 50; internal diameter 0.580 mm, external diameter 0.965 mm). Rainbow trout were revived on the operating table by irrigation of the gills with aerated water, then transferred to individual opaque acrylic experimental chambers (volume 3 l) supplied with aerated, flowing water. Cannulae were flushed daily with freshwater teleost saline (Wolf, 1963) containing 50 i.u. ml^-1 ammonium heparin (Sigma Chemical Company).

Fish were divided into two groups and held under identical conditions. One group served as the controls while the other group was subjected to daily physical stress (Reid et al. 1994) for a period of 7 days. Two fish were removed from the holding tank and placed into a separate tank where they were chased with a net for a period of 5–6 min until they were visibly exhausted and could be handled without resistance. This procedure was repeated until all fish had been chased until exhaustion. This stress regime was performed each morning for 7 days. Owing to poor feeding in the stressed fish (see also Reid et al. 1994), neither group was fed during the 7-day experimental period.

The levels of plasma catecholamines were measured in cannulated fish after 1, 3 and 7 days of chasing, as follows. Fish were cannulated 18 h prior to chasing (i.e. on days 0, 2 and 6) and then placed into opaque acrylic individual holding boxes for recovery. Previous work in this laboratory has shown that plasma catecholamine levels are restored to normal within 4 h of recovery from dorsal aortic cannulation (S. F. Perry and A. Tsay, unpublished results; see also Fig. 1). The next morning, a pre-stress blood sample (300 μl) was removed and the fish then was transferred to the chasing tank and chased until exhaustion. A blood sample was withdrawn immediately upon exhaustion and, after returning the fish to its holding box, additional blood samples were taken at 0.25, 0.5, 1, 4, 8, 12 and 24 h. Plasma was obtained by centrifugation (12 000 g for 30 s), quick-frozen in liquid N₂ and then stored for less than 2 weeks at −80 °C for subsequent determination of plasma adrenaline and noradrenaline levels.

The numbers (B_max) and binding affinities (1/K_D) of rbc surface β-adrenoceptors were determined in vitro in control fish and in fish subjected to 7 days of physical stress. After 7 days, a cannula was implanted into the dorsal aorta (see above) and the fish were allowed to recover in individual holding boxes for 24 h. A 2.5 ml blood sample was withdrawn from the dorsal aortic cannula and divided equally between two tonometry flasks (Eschweiler; 5 ml volume) containing 50 i.u. ml^-1 ammonium heparin. The blood was equilibrated for 45 min at 10 °C either with normoxic (⁠ remainder N₂) or with hypoxic (⁠remainder N₂) gas mixtures.

The protocol used to determine B_max and K_D was similar to that employed by Reid et al. (1991). Briefly, this technique uses a hydrophilic radioligand binding assay (Marttila and Nikinmaa, 1988) that specifically targets high-affinity cell surfaceβ -adrenoceptors. Radioligand binding was initiated by the addition of 40 μl of blood to 160 μl of physiological teleost saline (Wolf, 1963) containing 100 mg ml^-1 ascorbic acid to which the radioligand (±)-3,4-(3-t-butylamino-2-hydroxy-propoxy)-[5,7-³H]benzimidazol-2-one (³H-CGP 12177; 5–40 nmol l^-1) had been added alone or in combination with 200 μmol l^-1 (-)-isoproterenol. Incubations were performed in triplicate. The number of rbcs added to the incubation solution was determined by diluting 10 μl of blood in 10 ml of saline and then counting the number of rbcs present using a haemocytometer.

Incubations were stopped, after 45 min, by transferring the rbcs to borosilicate filters (no. 32, Mandel Scientific) using a cell membrane harvester (Brandell 24R) and five subsequent washings with 5 ml of ice-cold saline (0.9 % w/v NaCl). The filters were placed into glass scintillation vials containing 8 ml of fluor (ACS II, Amersham) and then stored in the dark for 24 h before being counted. The sample radioactivity was determined using a liquid scintillation counter (Canberra Packard model 2500 TR) with all counts automatically corrected to disints min^-1 using an external standard technique. The maximal number of isoproterenol-displaceable binding sites (B_max, in disints min^-1) and the apparent dissociation constants (K_D) were determined using Scatchard plot analysis (Scatchard, 1949). Binding site density was then converted to and expressed as binding sites per rbc by multiplying B_max (disints min^-1 cell^-1) by the specific activity of the radioligand and Avogadro’s number. In accordance with previously reported findings (Reid et al. 1991), the isoproterenol-displaceable CGP 12177 binding sites are hereafter referred to as rbc surface β-adrenoceptors.

Blood was withdrawn from the dorsal aortic cannula of control or stressed fish and pooled to obtain a sufficient volume for a single experiment (i.e. N=1). In practice, 6.0 ml was obtained from two fish and this was diluted with 2.0 ml of teleost saline to yield the volume (8.0 ml) required for an experiment. Blood sampling was halted immediately at the first sign of agitation or struggling. Blood was stored on ice in 25 ml round-bottomed tonometer flasks (final heparin concentration 50 i.u. ml^-1) for 2 h prior to experimentation. Plasma catecholamine levels were measured immediately in the pooled blood, and the noradrenaline levels were found to be 5.6±1.2 nmol l^-1 and 4.1±0.7 nmol l^-1 (means ± S.E.M., N=8) in the control and chased fish, respectively. Although plasma catecholamine levels were not determined after the 2 h waiting period, previous studies in this laboratory have shown that the concentration of noradrenaline decreases at a rate of 2 nmol l^-1 h^-1 in resting blood samples over such a time period (S. G. Reid, unpublished results). Furthermore, Tetens et al. (1988) have reported that catecholamines decay with a half-life of 24 min when stored at 15 °C under conditions of continuous tonometry and that this decay is insensitive to oxygenation status. Thus, in the present study, we have assumed noradrenaline levels of 1 nmol l^-1 in the blood prior to experimentation.

Adrenergic cell swelling, Na⁺ accumulation and cyclic AMP accumulation were determined at 10 °C in normoxic (⁠⁠, remainder N₂, pHe=7.8; 30 min pre-incubation) and hypoxic (⁠ remainder N₂, pHe=7.8; 30 min pre-incubation) blood samples (500 μl) after exposure to saline and seven doses of noradrenaline to yield final nominal concentrations of 10⁻⁹ mol l^-1 (saline addition) and 5×10⁻⁹ to 10⁻⁵ mol l^-1. After 5 min, a sample (100 μl) was removed to determine rbc cyclic AMP levels and cell solids, and after 30 min the remaining blood (400 μl) was removed to determine rbc water and Na⁺ contents. These sampling times were selected on the basis of previous studies showing that cyclic AMP accumulation was maximal 5 min after addition of agonist and that changes in cell water levels were detectable 30 min after addition of agonist (Salama and Nikinmaa, 1990; Salama, 1993).

To determine cyclic AMP levels, 50 μl of the 100 μl blood sample was deproteinized with 50 μl of 0.6 mol l^-1 perchloric acid and then neutralized with 1.0 mol l^-1 KHCO₃. A supernatant was obtained by centrifugation (10 000 g for 2 min) and then stored at −80 °C ( for 2–4 weeks) prior to analysis. Red cell dry mass was determined on the remaining 50 μl of sample by drying the pellet to a constant mass.

To determine cell water content, half of the 400 μl blood sample was centrifuged (10 000 g for 2 min) and the supernatant and upper layer of cells were discarded. The red cell pellet was weighed, dried to a constant mass overnight and re-weighed. The water content of the cells (%H₂O) was calculated as the percentage water of the total mass and was not corrected for any water contained within trapped plasma. To determine rbc [Na⁺], the remaining 200 μl of blood was centrifuged and the cell pellet was weighed, deproteinized with 0.6 mol l^-1 perchloric acid and stored at 4 °C until analysis a few days later.

Blood from control or stressed fish was sampled, stored and equilibrated under normoxic and hypoxic conditions as described above. Red cell water and Na⁺ contents were determined after a 30 min exposure to saline and to a range of 8-bromo cyclic AMP concentrations (10⁻⁴ to 10⁻² mol l^-1 final nominal concentrations).

Plasma catecholamine levels were determined on alumina-extracted samples (200 μl) using HPLC in conjunction with electrochemical detection (Woodward, 1982). 3,4-Dihydroxybenzylamine was used as an internal reference standard in all analyses. Red cell cyclic AMP levels were measured according to the protocol of a commercial kit (Amersham catalogue no. TRK.432). Red cell [Na⁺] was determined by flame emission spectrophotometry (Varian Spectraa 250 Plus).

The EC₅₀ values for noradrenaline-evoked and 8-bromo-cyclic-AMP-evoked responses were calculated in three different ways. First, sigmoidal dose–response curves of averaged data were constructed using iterative curve-fitting software (Sigmaplot Windows ver. 2.0) in which the EC₅₀ values were generated automatically. Second, Hill plots were constructed from each individual dose–response curve, using responses between 5 and 95 %, to yield an EC₅₀ value from each experiment. These EC₅₀ values were averaged and an appropriate statistical analysis was performed on the means (see below). Third, Hill plots were constructed using data from the averaged dose–response curves. Finally, an EC₅₀ value was calculated by averaging the results of the three techniques.

Data in tables and figures are presented as mean values ±1 standard error of the mean (S.E.M.). The data were statistically analysed by parametric analysis of variance (ANOVA) followed by a Fisher’s least squares difference (LSD) multiple-comparison test (comparison of all points) or Bonferroni’s comparison with a single control point. When parametric test assumptions were violated, the data was analysed using Kruskal–Wallis ANOVA followed by Dunn’s multiple-comparison test (comparison of all points) or Dunnett’s comparison with a single control point. All statistical testing, including determinations of normality and variance, was performed using commercial software (Sigmastat Windows). The fiducial limit of significance was set at 5 %.

Plasma catecholamine levels were elevated immediately upon exhaustion of the fish and remained significantly elevated for 15–30 min (Fig. 1). The post-exhaustion levels of catecholamines were statistically greater on day 1 when compared with levels after 3 and 7 days of chasing. The predominant circulating catecholamine after each acute stress episode was adrenaline. Typically, adrenaline accounted for more than 85 % of the total circulating catecholamine concentration; the peak values for adrenaline ranged between 50 and 275 nmol l^-1, whereas the peak values for noradrenaline ranged between 20 and 40 nmol l^-1. Serial blood sampling, in itself, did not contribute to a rise in plasma catecholamine levels as indicated by the stable and low (less than 5 nmol l^-1) levels of adrenaline and noradrenaline in the control fish (Fig. 1A).

The number of high-affinity cell surface -adrenoceptors was reduced significantly by 20 % after 7 days of physical stress, when measured under normoxic incubation conditions (Fig. 2A). Incubation of the blood under hypoxic conditions did not cause a statistically significant increase in the number of surface β-adrenoceptors in either group of fish, but further widened the difference in numbers of receptor between the control and stressed fish (30 % difference; Fig. 2A).

The apparent receptor affinity was unaltered by repeated physical stress when measured under normoxic conditions (Fig. 2B). Under hypoxic conditions, however, the receptor affinity was significantly decreased (K_D increased), but only in the rbcs from control fish. Thus, during hypoxia, the receptor affinity was higher in the rbcs obtained from stressed fish (K_D values of 4.2 and 2.1 nmol l^-1 in the control and stressed groups, respectively; Fig. 2B).

Red cells exhibited a dose-dependent rise in water content (Fig. 3A,D) in response to noradrenaline, which is a characteristic of β-adrenergic cell swelling induced by activation of rbc Na⁺/H⁺ exchange. Resting rbc water content and maximal β-adrenergic cell swelling were greatest under hypoxic conditions (compare Fig. 3A,D) and, moreover, there was an obvious leftward shift in the dose–response curve caused by hypoxia (compare Fig. 3C,F). Table 1 shows that, in the control fish, the average EC₅₀ value was lowered from 2.85×10⁻⁷ mol l^-1 (normoxia) to 2.55×10⁻⁸ mol l^-1 (hypoxia). In the chased fish, hypoxia was associated with a lowering of the average EC₅₀ from 8.09×10⁻⁸ mol l^-1 to 1.62×10⁻⁸ mol l^-1 (Table 1).

Repeated chasing had a marked impact on both the resting levels of rbc water content (Fig. 3A,D) and -adrenergic cell swelling (Fig. 3B–F; Table 1). The water content of the rbcs from chased fish was significantly lower at nearly all concentrations of noradrenaline under normoxic or hypoxic conditions. Furthermore, repeated chasing caused an obvious leftward shift of the dose–response curve (Fig. 3B–F). Under normoxic conditions, the average EC₅₀ was reduced from 2.85×10⁻⁷ mol l^-1 to 8.09×10⁻⁸ mol l^-1, a statistically significant difference based on the analysis of individual dose–response curves (Table 1). Under hypoxic conditions, the reduction in the EC₅₀ caused by repeated stress was smaller but, nevertheless, there was a significant lowering (statistical analysis performed on individual dose–response curves) of the average EC₅₀ from 2.55×10⁻⁸ mol l^-1 to 1.62×10⁻⁸ mol l^-1 (Table 1).

The interactive effects of hypoxia and repeated stress on - adrenergic rbc Na⁺ accumulation were similar to the changes in rbc water content (see Fig. 4; Table 2). In both the control and the stressed fish, hypoxia elicited an increase in maximal Na⁺ accumulation (Fig. 4B,E) and a leftward shift of the dose–response curves (Fig. 4C,F). Hypoxia lowered the EC₅₀ values from 2.67×10⁻⁷ mol l^-1 to 6.77×10⁻⁸ mol l^-1 in the control fish and from 1.49×10⁻⁷ mol l^-1 to 4.01×10⁻⁸ mol l^-1 in the stressed fish (Table 2).

At low concentrations of noradrenaline (5×10⁻⁸ to 10⁻⁹ mol l^-1), the intracellular [Na⁺] was lower in the stressed fish under normoxic conditions only (Fig. 4A,D). Maximal Na⁺ accumulation was unaffected by chasing, but the EC₅₀ was significantly reduced in the normoxic blood (2.67×10⁻⁷ mol l^-1versus 1.49 × 10⁻⁷ mol^-1; Table 2) although not significantly in the hypoxic blood (P=0.07, Table 2) by chasing.

Red cell cyclic AMP accumulated in a dose-dependent manner with significant increases first observed at 10⁻⁷ and 10⁻⁶ mol l^-1 in normoxic and hypoxic blood, respectively (Fig. 5). Neither hypoxia nor repeated chasing significantly affected the pattern of cyclic AMP accumulation during adrenergic stimulation (Fig. 5B,D). Because maximal responses were not observed in these experiments, Hill plots were not constructed and EC₅₀ values were not calculated.

To investigate further whether repeated physical stress was influencing the responsiveness of the Na⁺/H⁺ exchanger to cyclic AMP, rbcs were incubated with the permeable analogue 8-bromo cyclic AMP under normoxic or hypoxic conditions. Hypoxia alone caused a marked increase in the maximal cell swelling (compare Fig. 6B,D) and appeared to induce a leftward shift of the dose–response curve. Although it was not possible to calculate EC₅₀ values in the normoxic blood owing to the lack of a typical sigmoidal dose–response relationship (Fig. 6B), the data shown in Fig. 6A,C clearly indicate that cell swelling commenced at lower concentrations of 8-bromo cyclic AMP in the hypoxic blood. Under hypoxic conditions, the blood from chased fish displayed a significantly lower EC₅₀ compared with that of control fish (Fig. 6E; Table 3).

The changes in rbc [Na⁺] during incubation with 8-bromo cyclic AMP are illustrated in Fig. 7 and Table 4. Hypoxia, by itself, increased maximal Na⁺ accumulation (Fig. 7A–E) and evoked a leftward shift of the dose–response curve in both the control and the chased fish (Fig. 7C,F). Repeated stress did not alter the maximal accumulation of Na⁺ but did cause a leftward shift in the dose–response curve in the hypoxic blood (Fig. 7F); the average EC₅₀ was lowered 1.6-fold from 5.16×10⁻⁴ to 3.20×10⁻⁴ mol l^-1 (Table 4). In the normoxic blood, the average EC₅₀ values were 1.23×10⁻³ and 8.34×10⁻⁴ mol l^-1 in the control and stressed fish, respectively (P=0.06; Table 4).

The pattern of elevation of plasma catecholamine levels observed in fish after physical stress was similar to that observed in other studies on rainbow trout employing violent exhaustive exercise as the mode of stress (e.g. Tang and Boutilier, 1988; Van Dijk and Wood, 1988; Wood et al. 1990; see also review by Gamperl et al. 1994). It is clear from these studies that maximal catecholamine levels are achieved immediately upon exhaustion and the levels then decline rapidly to reach baseline values within 0.5–2 h. The rapid decline in plasma catecholamine levels presumably arises from the combined effects of tissue accumulation and metabolic degradation (Nekvasil and Olson, 1986a,b; for a review, see Randall and Perry, 1992). Indeed, Nekvasil and Olson (1986a) have estimated the biological half-life of an injected catecholamine dose to be less than 10 min.

In the present study, plasma adrenaline was by far the predominant circulating catecholamine after stress. This also suggests that adrenaline was the primary catecholamine released from its site of storage, the chromaffin cells associated with the wall of the posterior cardinal vein (Nandi, 1961; Nilsson, 1983; Randall and Perry, 1992; Reid and Perry, 1994b; Reid et al. 1994). The prevalence of adrenaline in the plasma is also consistent with the results of in situ perfusion studies on rainbow trout showing that adrenaline is the principal catecholamine released from the posterior cardinal vein during stimulation with the chlolinoceptor agonist carbachol (Fritsche et al. 1993; Reid and Perry, 1994b). With a few exceptions, the results of previous studies investigating the effects of acute stress on teleost fish also demonstrate that plasma adrenaline is the primary circulating catecholamine (e.g. see Table 2 in Gamperl et al. 1994). However, this does not necessarily exclude an important role for noradrenaline in mediating adrenergic physiological responses. For example, thorough analyses of the dose-dependency of rbc Na⁺/H⁺ exchange in rainbow trout blood in vitro have revealed a much higher sensitivity to noradrenaline than to adrenaline (Tetens et al. 1988; Salama, 1993). The higher sensitivity of the rainbow trout rbc to noradrenaline has led to speculation that the rbc β-adrenoceptor in this species is of the ₁ subtype (Tetens et al. 1988). In the present study, the maximal plasma concentrations of noradrenaline achieved immediately after stress were approximately 20–40 nmol l^-1. On the basis of the in vitro dose–response curves for cell swelling and Na⁺ accumulation and a previous study (Tetens et al. 1988), significant activation of Na⁺/H⁺ exchange probably did occur in the chased fish upon release of catecholamines, despite the fact that the release occurred under normoxic conditions when adrenergic responsiveness is presumably lower (e.g. see Fig. 3B). Chasing the fish, however, probably caused a significant depression of blood pH (Wood et al. 1990) and this is known to enhance the adrenergic responsiveness of rainbow trout rbcs in vitro under normoxic conditions (e.g. Nikinmaa et al. 1987). An important caveat is that the dose–response curves were constructed after 7 days of repeated chasing, at which time the EC₅₀ values for adrenergic Na⁺/H⁺ exchange were markedly lowered. It is uncertain when exactly during the 7-day stress regime this leftward shift of the dose–response curve occurred. Although EC₅₀ values for adrenaline were not determined, the results of previous studies documenting high EC₅₀ values (e.g. 7.6×10⁻⁷ mol l^-1, Tetens et al. 1988; 2× 10⁻⁶ mol l^-1, Salama, 1993) do not support a significant role for adrenaline in the activation of rbc Na⁺/H⁺ exchange in rainbow trout, at least under normoxic conditions. Interestingly, Berenbrink and Bridges (1994) showed that the rbcs of Atlantic cod (Gadus morhua) possess β-adrenoceptors that appear to have a higher affinity for adrenaline than for noradrenaline, and thus that adrenaline may play a more important role in this species.

An important observation of the present study was an attenuation of the catecholamine release response on days 3 and 7 of the repetitive stress protocol; the maximal levels of plasma catecholamines were significantly higher on day 1 (Fig. 1). These data imply a desensitization of the catecholamine release process with repeated stress and thus confirm the results of a previous study (Reid et al. 1994) on catecholamine secretion in an in situ perfused posterior vein preparation. In that study, a pronounced desensitization of carbachol-evoked catecholamine release was noted following 5 days of repeated physical stress. Because the sensitivity of release, rather than the maximal quantities of catecholamines released, was lowered, it was suggested by Reid et al. (1994) that the desensitization was caused by a reduced affinity of chromaffin cell cholinoceptors for cholinergic agonists. Catecholamine release in vivo is probably not causedexclusively by stimulation of cholinoceptors, but is presumably the net result of numerous factors including a variety of neural inputs (Reid et al. 1995), humoral influences (Opdyke et al. 1981; Fritsche et al. 1993; Kloas et al. 1994; Epple et al. 1994) and localized chemical changes (Opdyke et al. 1983; Perry et al. 1991). Thus, desensitization of the catecholamine release process, which was apparent in the present study, could have arisen in many ways. It is unlikely that the reduced levels of catecholamines immediately post-exhaustion on days 3 and 7 were caused by accelerated catecholamine degradation or clearance, given that the blood samples were taken within seconds after the stress period. The physiological significance of reduced levels of circulating catecholamines with repeated stress may be to prevent excessive stimulation of target cells. This idea is consistent with general models of desensitization of G-protein-coupled receptors (Collins et al. 1990; Hausdorff et al. 1990; Gershengorn, 1994).

In a preliminary study (S. F. Perry and S. G. Reid, unpublished observations), plasma cortisol levels were measured on uncannulated juvenile trout (mass 20–115 g; N=124) after 1, 3 and 7 days of chasing. Plasma levels rose from 11.6±6.3 nmol l^-1 to peak levels of approximately 415 nmol 1^-1 and remained elevated for 1–12 h depending upon the number of days of prior chasing. Although these data obtained from juvenile fish may differ from the pattern of cortisol elevation in the larger fish that were used in the in vitro experiments, they are consistent with the hypothesis that acute physical disturbance causes pronounced and prolonged elevation of plasma cortisol levels in trout (see Gamperl et al. 1994). Thus, increases in plasma cortisol levels in the stressed fish could conceivably have influenced rbc β-adrenoceptors (see below), as demonstrated previously (Reid and Perry, 1991).

Elevated levels of stress hormones have been shown to influence rbc β-adrenoceptors in rainbow trout, although in opposing ways. Increased levels of cortisol in vivo or in vitro are known to increase the numbers of internalized β-adrenoceptors and can result in an increase in the numbers of cell surface β-adrenoceptors when these cells are subsequently exposed to acute hypoxia (Reid and Perry, 1991). Additionally, the number of surface -adrenoceptors in rainbow trout hepatocytes was markedly increased after in vivo elevation of plasma cortisol levels (Reid et al. 1992). In contrast, increased levels of catecholamines are known to lower the numbers of rbc surface -adrenoceptors in rainbow trout (Gilmour et al. 1994). In the present study, the number of rbc surface β-adrenoceptors measured in the blood of the repeatedly stressed fish was lower than in the control fish when assayed under both normoxic and hypoxic conditions. These results suggest that the effects of cortisol on increasing rbc -adrenoceptor numbers were more than offset by the down-regulation caused by the elevated catecholamine levels and any other potential negative modifiers of rbc β-adrenoceptors that were not measured in these experiments. Thus, the model of Perry and Reid (1993), which proposes that chronic stress may ‘pre-adapt’ the teleost rbc to acute stress, is tenable only if plasma catecholamine levels do not rise concurrently with cortisol levels.

The protocol employed in this study required that all fish be cannulated 18 h prior to blood withdrawal. Thus, even the non-stressed group probably experienced post-surgery elevations of plasma cortisol and catecholamine levels (Gingerich and Drottar, 1989). Typically, in this laboratory, plasma catecholamine levels remain elevated for less than 4 h (S. F. Perry and A. Tsay, unpublished data) and, although the cortisol levels probably remained high in the 24 h following surgery, there were nevertheless pronounced differences (also see below) between the repeatedly stressed and non-stressed groups.

Previous studies have demonstrated varied results with respect to the effects of acute hypoxia in vitro on the numbers of rbc β-adrenoceptors. Although there are several reports of increased numbers of β-adrenoceptors during acute hypoxia (Marttila and Nikinmaa, 1988; Reid and Perry, 1991; Reid et al. 1993), this study and a previous study (Perry and Reid, 1992) were unable to demonstrate statistically significant increases. The inability of hypoxia to modify the rbc -adrenoceptor population in the present study is consistent with the identical pattern of cyclic AMP accumulation that was observed in normoxic and hypoxic blood (see below). Thus, while increased number of rbc surface β-adrenoceptors may contribute to enhanced adrenergic responsiveness under certain conditions (Reid and Perry, 1991; Reid et al. 1993; Perry and Reid, 1993), the relatively small magnitude of the response and its inconsistency raise doubts as to its physiological importance. Indeed, the stimulatory effects of hypoxia observed in the present study can be explained solely on the basis of post-adrenoceptor modifications (see below). Although, the apparent affinity of the rbc β-adrenoceptors decreased in the control fish during hypoxia, this did not appear to influence the pattern of cyclic AMP accumulation during adrenergic stimulation.

In this study, rbc Na⁺ and water accumulation were used as indices of β-adrenergic Na⁺/H⁺ exchange activity. It is important to point out that the net accumulation of Na⁺ and water within the cell after application of catecholamines will also be influenced by secondary mechanisms, including Na⁺ pump activity and volume regulatory responses. Thus, while we believe that the predominant factor determining rbc Na⁺ and water levels is activity of the Na⁺/H⁺ exchanger, we cannot entirely exclude involvement of these other factors.

Although several previous studies have demonstrated a stimulatory effect of acute hypoxia on adrenergic responsiveness in rainbow trout (e.g. Motais et al. 1987; Reid and Perry, 1991; Salama, 1993), this is the first study to show a clear positive influence on both the maximal activity of the adrenergic response and the sensitivity of this response to applied catecholamines. These results were obtained by a rigorous analysis of dose–response curves. Owing to the importance of obtaining reliable EC₅₀ values, three separate methods of determination were employed and the results were then averaged (see Materials and methods). Statistical evaluation, however, could be performed only on the EC₅₀ values obtained from averaging individual dose–response curves. Thus, the statistical differences noted in this paper are based on such values.

The lowering of the EC₅₀ values (increased response affinity) and the increase in the apparent maximal Na⁺/H⁺ exchange activity during hypoxia were clearly unrelated to modification of the cell surface β -adrenoceptors because the B_max of these receptors was unchanged during hypoxia and the K_D values were either unchanged (chased fish) or actually increased (control fish). Consequently, the pattern of cyclic AMP accumulation was also unaltered by hypoxia (see also Salama, 1993). Thus, unlike in previous studies (e.g. Reid et al. 1993), there was no evidence of hypoxia-mediated modification of the β-adrenergic signal transduction pathway leading to cyclic AMP accumulation. Instead, the dual effects of hypoxia on increasing the affinity and the maximal activity of Na⁺/H⁺ exchange are presumably related to regulation at the level of the exchanger itself, as proposed initially by Motais et al. (1987). This could involve a direct effect of hypoxia on the numbers and/or affinities of the Na⁺/H⁺ exchanger (Reid and Perry, 1994a) or an indirect effect mediated by the oxygenation state of haemoglobin (Motais et al. 1987; Nikinmaa and Jensen, 1992). The results of experiments using 8-bromo cyclic AMP revealed that the responsiveness of the Na⁺/H⁺ exchanger to cyclic AMP was modified by hypoxia since there were significant increases in maximal Na⁺/H⁺ exchange activity (Figs 6, 7) and reductions in EC₅₀ values (Table 4).

We are unable to explain the apparent lack of an effect of hypoxia on cyclic AMP accumulation in the present study, given the previous results obtained from this laboratory (Reid and Perry, 1991; Reid et al. 1993). Salama (1993) also did not observe an effect of hypoxia on cyclic AMP formation in rainbow trout rbcs, yet a previous study from that same laboratory (Salama and Nikinmaa, 1990) demonstrated a marked stimulatory effect of hypoxia on cyclic AMP formation in carp (Cyprinus carpio) rbcs. Thus, increases in adrenergic responsiveness during acute hypoxia in teleost rbcs may, or may not, involve modification of the -adrenergic signal transduction pathway leading to cyclic AMP formation, while regulation at the level of the Na⁺/H⁺ exchanger appears to be a more pronounced and prevalent response.

The rbcs from the physically stressed fish had lower water contents under resting conditions in vitro compared with those of the control fish. The underlying explanation for the difference is unclear but it may reflect overcompensation of the rbc volume regulatory response in the stressed fish, which probaby experienced daily adrenergic rbc swelling in response to the chasing.

This is the first study to demonstrate that chronic repeated stress can markedly influence the responsiveness of the rainbow trout rbc to acute adrenergic stimulation. Specifically, chronic stress caused significant decreases in the noradrenaline EC₅₀ values without affecting maximal Na⁺/H⁺ exchange activity. The increased affinity of the Na⁺/H⁺ exchanger for exogenous noradrenaline appeared to be unrelated to changes in the - adrenergic signal transduction pathway because the pattern of cyclic AMP accumulation during adrenergic stimulation was unaltered. Thus, the significant reduction in the numbers of cell surface β-adrenoceptors associated with repeated stress clearly had no negative impact on signal transduction. Under hypoxic incubation conditions, the rbc β-adrenoceptors from the stressed fish displayed a higher binding affinity, yet this also did not affect the pattern of cyclic AMP formation. Thus, the enhanced rbc adrenergic responsiveness in the stressed fish must reflect post-cyclic-AMP modifications. This conclusion is supported by the results of the experiments demonstrating a lowering of the 8-bromo cyclic AMP EC₅₀ values in hypoxic stressed fish (Tables 3, 4). The mechanisms underlying the increased sensitivity of the Na⁺/H⁺ exchanger to adrenergic stimulation following repeated stress are unknown. Regardless of the underlying mechanisms, the results reveal that the prior history of a fish can markedly influence rbc adrenergic responses in vitro, even over a relatively short period (e.g. 7 days). Furthermore, the impact of prior stress on the rbc adrenergic response may vary according to the nature of the stress. For example, Cossins and Kilbey (1989) documented that stressful episodes (e.g. rain storms or rapid temperature changes) at a commercial fish farm were associated with a reduced rbc adrenergic response. Therefore, in the light of these results, experiments designed to assess rbc adrenergic responses must carefully control for the potential effects of chronic or prior stress on the acute adrenergic stress response.

This work was supported by NSERC of Canada operating and equipment grants to S.F.P. S.G.R. was the recipient of an NSERC Postgraduate Scholarship. A.S. was the recipient of an NSERC International Postdoctoral Fellowship.