Adrenergic Inhibition of Carbon Dioxide Excretion by Trout Red Blood Cells In Vitro is Mediated by Activation of Na⁺/H⁺ Exchange

In the previous paper, we have developed a sensitive new in vitro assay to measure the rate of CO₂ excretion by trout blood under conditions comparable to those in vivo (Wood and Perry, 1991). Using this assay, we confirmed that, under the typical post-exercise condition of metabolic acidosis, catecholamines cause a transient inhibition of CO₂ excretion. This observation supports the CO₂ retention theory’ (Wood and Perry, 1985; Perry, 1986; Perry and Wood, 1989) to explain the increase in Pa_CO2 commonly observed after strenuous exercise or catecholamine infusion in salmonids. However, the mechanism of this inhibition remains unclear.

Our original idea was that catecholamines directly inhibited the HCO₃⁻ entry step, perhaps via a direct action on Cl⁻/HCO₃⁻exchange via band 3 (Wood and Perry, 1985; Perry, 1986). An alternative possibility, however, is that the inhibition is a consequence of β-adrenergic activation of the RBC Na⁺/H⁺ antiporter, the well-known RBC pHi regulatory response (see reviews by Nikinmaa and Tufts, 1989; Thomas and Motais, 1990). Indeed, in the previous paper, the magnitude of the inhibition was correlated with the magnitude of this RBC adrenergic response, though the time courses did not appear to be identical (Wood and Perry, 1991). In theory, adrenergic stimulation of Na⁺/H⁺ exchange and the attendant alkalization of the RBC could impede CO₂ excretion by reducing the gradient for HCO₃⁻ entry. In consequence, the net rate of dehydration of plasma-derived HCO₃⁻by erythrocytic carbonic anhydrase would be reduced. Furthermore, the abrupt alkalization of the RBC could reduce (or temporarily abolish) the normally outwardly directed gradient between RBC and plasma, as suggested previously by Thomas and Motais (1990).

The aim of the present study was to employ the new in vitro assay to elucidate the underlying mechanisms for adrenergic inhibition of RBC CO₂ excretion. In particular, we were interested in the dose-dependency of the response with the two naturally occurring catecholamines (adrenaline and noradrenaline), the pharmacological identification of the adrenoceptors involved, and the relationship of the inhibition (if any) to activation of Na⁺/H⁺ exchange and/or to a direct action on Cl⁻/HCO₃⁻exchange.

Rainbow trout [Oncorhynchus mykiss (Walbaum)] of either sex weighing between 175 and 225 g were obtained from Thistle Springs Trout Farm (Ashton, Ontario). Fish were maintained indoors in flowing dechlorinated and vigorously aerated City of Ottawa tapwater at 10°C, as described by Wood and Perry (1991). Prior to experiments, the trout were fitted with indwelling dorsal aortic catheters (Soivio et al. 1975) and allowed to recover for 48 h in opaque Perspex boxes.

The new technique for assessing CO₂ excretion within trout RBCs has been described in detail in the accompanying paper (Wood and Perry, 1991) and is therefore only briefly reiterated here. Approximately 20ml of whole blood was required for a typical single experimental run (i.e. N=1). Thus, it was necessary to use pooled blood obtained by slow withdrawal from the dorsal aortic cannulae of 4–5 fish (3–4ml per fish). The pooled blood (on ice) was assayed for haematocrit (17.5±1.5%), whole-blood pH (pHe=7.84±0.03) and total carbon dioxide content ⁠. Appropriate volumes of HC1 (140mmoll⁻¹) were added to simulate typical whole blood acidosis in trout immediately after exhaustive exercise (i.e. pHe=7.40±0.02; ⁠; Wood and Perry, 1985). In one experimental series the blood was not acidified and in this case an equivalent volume of 140 mmol l⁻¹ NaCl was added. Endogenous catecholamine levels (adrenaline plus noradrenaline) were always less than 7 nmol ⁻¹. Samples (1 ml) of acidified blood (see below) or true plasma were added to glass scintillation vials (20 ml), stoppered and gassed with a humidified gas mixture to yield (1.91 mmHg), (155 mmHg), remainder N₂, for 2h at ambient water temperature in a shaking water bath. The gas mixture was provided by a gas-mixing pump (Wösthoff model M 301a/f).

2μCi (10 μl of 200 μCi ml⁻¹) of sodium [¹⁴C]bicarbonate (in teleost Ringer) was added to each 1 ml of blood or plasma. The vial was then immediately sealed with a rubber septum, from which was suspended a plastic well containing a filter paper trap (150 μl of hyamine hydroxide) for CO₂, and shaking was started. After exactly 3 min of shaking, the filter was removed and assayed for ¹⁴C activity. Whole-blood or true plasma pH was determined and the remaining blood centrifuged (12000 g for 2min). The pellet was utilized to determine RBC pHi according to the freeze-thaw method (Zeidler and Kim, 1977). Samples of true plasma were assayed for¹⁴C activity (50 μl) and (100 μl) to determine plasma HCO₃⁻specific activity (disintsmin⁻¹μmol⁻¹). Analytical techniques and calculations were identical to those described by Wood and Perry (1991). RBC CO₂ excretion rate was calculated by subtracting the plasma CO₂ excretion rate from the wholeblood rate. This procedure, therefore, allows an assessment of the rate of CO₂ excretion arising only from the dehydration of plasma HCO₃⁻.

Assays were performed 5 min after addition of 10 μl of freshly prepared catecholamine (L-adrenaline bitartrate, L-noradrenaline bitartrate; Sigma) or adrenoceptor agonist (L-isoproterenol bitartrate, phenylephrine HO; Sigma) in 140 mmol I⁻¹ NaCl to yield nominal final concentrations of 10–1000 nmol l⁻¹ (adrenaline or noradrenaline) or 1000 nmol I⁻¹ (isoproterenol, phenylephrine). In controls, 10 μl of 140 mmol I⁻¹ NaCl was added to each vial. In one experimental series, 50μ1 of the adrenoceptor antagonists (2×10⁻⁴mol⁻¹) propranolol (β-adrenoceptor antagonist; Sigma) or phentolamine (β-adrenoceptor antagonist; Ciba-Geigy) was added to samples of blood 30 min prior to addition of 1000 nmol I⁻¹ noradrenaline to achieve final nominal antagonist levels of 1×10⁻⁵ mol 1⁻¹ ; 50 μl of 140 mmol I⁻¹ NaCl was added to control vials. In another experimental series, utilising 1000 nmol I⁻¹ noradrenaline, blood was pre-incubated for 30 min with 50 μl of the antiporter inhibitors amiloride HC1 (Na⁺/H⁺ exchange blocker; Sigma) or SITS (4-acetamido-4-isothiocyanatostilbene-2,2-disulphonic acid; Cl⁻/HCO₃⁻exchange blocker; Sigma) dissolved in 2 % dimethyl sulphoxide (DMSO) to yield final nominal concentrations of l×10⁻⁴moll⁻¹ in 0.1 % DMSO; 50 μl of 2 % DMSO was added to control vials.

An additional series of experiments was performed using non-acidified blood. In this instance a simple comparison was made between the effects of 1000 nmol l⁻¹ noradrenaline on HCO₃⁻ dehydration in acidified and non-acidified blood.

All values shown are means±l standard error of the mean (S.E.M.). The results have been statistically analyzed using factorial analysis of variance followed by Fisher’s LSD multiple-comparison test; 5% was taken as the fiducial limit of significance.

The acute addition of catecholamines to acidified whole blood caused a dosedependent inhibition of CO₂ excretion within RBCs. The effect of noradrenaline was more pronounced than that of adrenaline (Fig. 1). A concentration of 10 nmol I⁻¹ noradrenaline caused a significant 23% inhibition of CO₂ excretion, whereas a 10-fold higher concentration of adrenaline was required to elicit a similar degree of inhibition. At the highest concentrations (1000 nmol l⁻¹) utilised, RBC CO₂ excretion was reduced by 54 % and 45 % by noradrenaline and adrenaline, respectively. The catecholamines also induced pronounced and dosedependent reductions in the RBC transmembrane pH gradient (pHe—pHi), owing to simultaneous changes in both pHe and pHi (Table 1).

Fig. 2 illustrates the effects of a variety of adrenoceptor antagonists and agonists on RBC CO₂ excretion. Pre-incubation of whole blood for 30min with the β-adrenoceptor antagonist propranolol (10⁻⁵mol I⁻¹) entirely abolished the inhibitory effect of noradrenaline (1000 nmol I⁻¹) on CO₂ excretion as well as preventing the usual changes in pHe, pHi and pHe—pHi (Table 2). In contrast, preincubation with an equivalent concentration of the β-adrenoceptor antagonist phentolamine did not significantly alter the effects of noradrenaline on either CO₂ excretion or pHe, pHi or pHe-pHi (Fig. 2; Table 2). Addition of these antagonists alone was without effect on any measured or calculated variable (Fig. 2; Table 2). The )3-adrenoceptor agonist isoproterenol (1000 nmol l⁻¹) mimicked the effects of an equivalent dose of noradrenaline, whereas the α-adrenoceptor agonist phenylephrine (1000nmoll⁻¹) had no effect on CO₂ excretion (Fig. 2), although it did significantly affect blood acid-base status (Table 2).

In an attempt to elucidate the post-receptor mechanism by which catechol-amines inhibit CO₂ excretion, the effects of noradrenaline (1000nmoll⁻¹) were assessed in the presence or absence of the Na⁺/H⁺ exchange blocker amiloride (10⁻⁴moll⁻¹) or the C1⁻/HCO₃⁻ exchange blocker SITS (10⁻⁴mol I⁻¹). Incubation of whole blood with amiloride prevented the usual adrenergic inhibition of CO₂ excretion (Fig. 3). In addition, the changes in pHe and pHe—pHi were significantly reduced by amiloride (Table 3), indicating that this treatment was indeed effective in blocking (at least partially) the Na⁺/H⁺ exchanger. Amiloride alone did not affect the rate of RBC CO₂ excretion (Fig. 3). The addition of SITS caused a marked 70% inhibition of CO₂ excretion. However, this reduced rate was further substantially inhibited by the presence of noradrenaline. Indeed, on a relative basis, the percentage inhibition (53%) by noradrenaline in SITS-treated blood was identical to that (54%) in control blood (Fig. 3).

Fig. 4 illustrates the curvilinear relationship between RBC CO₂ excretion and pHe—pHi which was generated using a range of concentrations (10–1000 nmol I⁻¹) of noradrenaline and adrenaline and 1000nmoll⁻¹ isoproterenol. These data support the hypothesis that activation of the Na⁺/H⁺ exchanger (as indicated by the changes in pHe-pHi) is the mechanism underlying the adrenergic inhibition of CO₂ excretion. It is noteworthy that, in the presence of amiloride, noradrenaline did not elicit a reduction of CO₂ excretion, despite a small reduction of pHe—pHi (Fig. 4). Indeed, the rate of CO₂ excretion obtained in the presence of noradrenaline plus amiloride was significantly greater than that predicted from the curvilinear relationship. Possible reasons for this deviation are discussed below. Predictably, SITS (in the presence or absence of noradrenaline) caused a more pronounced inhibition of CO₂ excretion than predicted (Fig. 4), owing to direct inhibition of the C1⁻/HCO₃⁻ exchanger.

All the experiments described above were performed using acidified blood, 5–8 min after catecholamine addition, based on the protocol of Wood and Perry (1991), because the adrenergic inhibition of RBC CO₂ excretion was well-defined under these conditions. Furthermore, the situation was appropriate because was were testing the hypothesis that the effect is linked to activation of the Na⁺/H⁺ exchanger, and acidification is known to enhance the responsiveness of this antiporter to catecholamines. Preliminary experiments by Wood and Perry (1991) suggested that the adrenergic inhibition of CO₂ excretion was absent if the blood was kept at normal pHe. However, those trials were performed at inappropriate times (30–120min after catecholamine addition).

Therefore, in a separate series of experiments, we tested whether acidification of the blood was indeed required to evoke adrenergic inhibition of RBC CO₂ excretion. The trials were run 5-8min after addition of noradrenaline (1000nmoll⁻¹). The same original blood (N=6), half pre-equilibrated for 2h at normal acid-base status (pHe=7.88±0.02, plasma [HCO₃⁻]=5.5±0.1 mmoll⁻¹) the other half pre-equilibrated for the same period under metabolic acidosis (pHe=7.46±0.01, plasma [HCO₃⁻]=2.1±0.1 mmoll⁻¹), was used in these experiments. The results (Fig. 5) clearly demonstrated that RBC CO₂ excretion is totally insensitive to noradrenaline when the blood is non-acidified, thereby supporting our contention of linkage to Na⁺/H⁺ exchange. However, the absence of acidosis only reduced and did not abolish the Na⁺/H⁺ exchange response. Whole-blood pH (pHe) still fell (to 7.58±0.03) and pHi still increased slightly (from 7.37±0.03 to 7.42±0.02) in response to noradrenaline in non-acidified blood. Interestingly, on Fig. 4, this point lies above the predicted relationship in an identical position to the other anomalous point (noradrenaline after amiloride).

The results of the present study unequivocally demonstrate adrenergic inhibition of RBC CO₂ excretion in vitro and suggest that the underlying mechanism is linked to β-adrenoceptor-mediated activation of the Na⁺/H⁺ exchanger rather than via a direct action on the CF/HCO₃⁻ exchanger. The involvement of RBC β-adrenoceptors is indicated by the experiments utilizing β-adrenoceptor agonists or antagonists. Although there is conflicting experimental evidence (Bennett and Rankin, 1985; Tetens et al. 1988), the current consensus is that the RBC β-adrenoceptors are of the β₁ subtype (see review by Nikinmaa and Tufts, 1989). Our observation that noradrenaline is a more potent inhibitor than adrenaline of CO₂ excretion further supports this consensus.

RBC CO₂ excretion can be modified not only by direct antagonism of the C1⁻/HCO₃^_ exchange on the RBC membrane (e.g. the inhibition caused by SITS; see also Perry et al. 1982), but also by a linkage between the C1⁻/HCO₃⁻ and Na⁺/H⁺ exchange pathways. It is the latter mechanism that is the basis of the adrenergic inhibition of RBC HCO₃⁻ dehydration. Catecholamines cause a β-adrenergic, cyclic-AMP-dependent stimulation of the Na⁺/H⁺ antiporter in the RBCs of trout and many other teleost species (Nikinmaa and Tufts, 1989). The consequent extrusion of H⁺ from the RBC causes a pronounced reduction of pHe, a smaller rise in RBC pHi and therefore a marked alteration of the transmembrane pH gradient (pHe—pHi). The change in pHe—pHi after adrenergic stimulation is directly related to the activity of the Na⁺/H⁺ exchanger and is largely blocked by amiloride. Thus, the inhibitory effects of amiloride on adrenergic impairment of CO₂ excretion and the observed relationship between CO₂ excretion rates and pHe—pHi (Fig. 4) both support the idea that adrenergic activation of Na⁺/H⁺ exchange is responsible for the inhibition of RBC CO₂ excretion.

Although Wood and Perry (1991) portrayed a simple correlation between RBC CO₂ excretion and pHe—pHi, the more extensive data of the present study suggest that the true relationship is curvilinear (Fig. 4). The inhibition by SITS was far greater than that predicted by this curvilinear relationship, and reveals the additional effect of direct inhibition of the C1⁻/HCO₃⁻ exchanger. The finding that noradrenaline caused an additional 53% reduction in CO₂ excretion after SITS treatment (equal, on a relative basis, to its effect before SITS; Fig. 3), correlated with a large decrease in pHe—pHi (Table 3; Fig. 4), provides further evidence that the adrenergic effect is not directly on the C1⁻/HCO₃⁻ exchanger, but rather via Na⁺/H⁺ exchange.

It is noteworthy that there was significant deviation from the predicted relationship between CO₂ excretion and pHe—pHi in adrenergically stimulated blood pre-treated with amiloride (Fig. 4), and also in non-acidified blood treated with noradrenaline (identical position on Fig. 4). Specifically, the rate of CO₂ excretion was greater than predicted if activation of Na⁺/H⁺ exchange was the solitary determinant of the change in CO2 excretion rates after catecholamine addition. The most likely explanation is that these data may reflect a minor adrenergic stimulation of CO₂ excretion which was unmasked by these treatments.

This apparent stimulation of CO₂ excretion may simply reflect the pHe-depen-dence of the C1⁻/HCO₃⁻ exchanger, the activity of which (as measured using SO₄^2-) increases as extracellular pH is reduced (Romano and Passow, 1984; Fievet et al. 1988). The observation of stable rates of CO₂ excretion in blood treated with the a-adrenoceptor agonist phenylephrine (Fig. 2), despite a significant reduction of pHe-pHi (Table 2), also supports this hypothesis. The net effect of catecholamines on CO₂ excretion may be the result of two opposing phenomena; an inhibitory effect linked to activation of the Na⁺/H⁺ exchanger as well as a stimulatory component linked to the pHe-dependence of the C1⁻/HCO₃⁻exchanger. Clearly, under most conditions, the inhibitory response is dominant.

The proposed mechanisms underlying the linkage between adrenergic activation of Na⁺/H⁺ exchange and inhibition of RBC CO₂ excretion are as follows. Activation of the Na⁺/H⁺ exchanger and concurrent alkalization of the RBC must cause an immediate rise in intracellular HCO₃⁻ levels, owing to physicochemical buffering and the presence of intra-erythrocytic carbonic anhydrase (Thomas and Motais, 1990). This results in a diminution of the electrochemical gradient for the inward movement of HCO₃⁻ and hence reduces the rate of HCO₃⁻ dehydration within the RBC. Unfortunately, the extracellular acid-base disequilibrium after adrenergic stimulation prevents us from calculating the changes in the transmembrane HCO₃⁻ gradient. It is obvious, nonetheless, that the pronounced changes in pHe-pHi must be associated with significant reductions in this gradient. Owing to the pronounced sensitivity of the RBC CO₂ excretion rate to acute changes in the transmembrane HCO₃⁻ gradient (see Fig. 3 in Wood and Perry, 1991), it is possible to achieve significant inhibition with relatively minor changes in the HCO₃⁻ gradient. The other proposed mechanism causing the adrenergic reduction of CO₂ excretion is CO₂ recycling between the plasma and the interior of the RBC immediately after activation of the Na⁺/H⁺ exchanger (Motais et al. 1989; Thomas and Motais, 1990). Under normal conditions, decreases from RBC to plasma to water, which permits effective excretion of CO₂ according to existing partial pressure gradients. In our experimental apparatus, the decreases from RBC to plasma to gas phase (see Wood and Perry, 1991), which is essentially analogous to the in vivo situation. The massive stimulation of proton extrusion from the RBC after addition of catecholamine causes rapid formation of intracellular HCO₃⁻via the catalysed hydration of CO₂. Since a significant portion of this HCO₃⁻ must leave the RBC in exchange for Cl⁻, and CO₂ cannot be reintroduced into the cell as rapidly as it is being hydrated to HCO₃⁻, a major reduction of intracellular must ensue. Therefore, a portion of the CO₂ formed at the uncatalysed rate from HCO₃⁻ and H⁺ in the plasma is recycled through the RBC rather than diffusing into the water (or the gas phase in the in vitro assay). Both the reduction of the transmembrane HCO₃⁻ gradient and CO₂ recycling will cause a transient inhibition of net CO₂ excretion in vivo. The CO₂ recycling is associated with the phase of extracellular acid-base disequilibrium. Red blood cell HCO₃⁻ levels presumably remain elevated (since pHi remains elevated; see Fig. 6 of Wood and Perry, 1991) after the Na⁺/H⁺ exchanger is once again inactivated and can therefore contribute to a more prolonged inhibition of RBC HCO₃⁻ dehydration. Despite this, however, we suggest that normal CO₂ excretion rates are re-established in vivo owing to a rise in as CO₂ backs up in the system, thereby restoring the transbranchial flux.

Adrenergic inhibition of RBC CO₂ excretion, mediated by activation of Na⁺/H⁺ exchange, is sufficient to account for the transitory retention of CO₂ in fish blood after exhaustive exercise (Wood and Perry, 1985; Perry and Wood, 1989) and during intra-arterial catecholamine infusion (Perry and Vermette, 1987; Vermette and Perry, 1988). In the present study, this effect was demonstrated at realistic levels of catecholamines (10–1000 nmol I⁻¹) using physiological levels of HCO₃⁻ in blood appropriately acidified to mimic post-exercise acidosis. Previous studies in vivo (Steffensen et al. 1987; Playle et al. 1990) or in vitro (Tufts et al. 1988) were unable to demonstrate adrenergic inhibition of whole-animal CO₂ excretion and RBC CO₂ excretion, respectively. There are several possible reasons for the discrepancy. First, we now know that the inhibitory effect of catecholamines occurs only under acidotic conditions, and not when blood pH is normal (Fig. 5). Presumably, this reflects the potentiating effects of extracellular acidosis on adrenergic stimulation of the RBC Na⁺/H⁺ antiporter (see review by Nikinmaa and Tufts, 1989) in addition to possible allosteric modification of the Na⁺/H⁺ antiporter itself (e.g. Aronson et al. 1982). Second, the response is transient, reaching a maximum after approximately 5 min and disappearing after about 30min (Wood and Perry, 1991). Third, inhibition of HCO₃⁻ dehydration in vivo would cause a rise in of the blood and quickly re-establish CO₂ excretion at a new steady state (see above). Thus, inappropriate methodology and/or sampling times may have prevented detection of adrenergic responses in other studies. Fourth, the protocol of acutely raising the blood HCO₃⁻ concentration to 100 mmol l⁻¹ in the ‘boat’ technique (Tufts et al. 1988) may have obscured the proposed linkage between RBC CO₂ excretion and activation of Na⁺/H⁺ exchange in that study.

The functional significance of adrenergic inhibition of CO₂ excretion is unclear but may be related to the control of ventilation. For example, after exhausting exercise fish are motionless and blood oxygen status is unchanged so stimulation of neither proprioceptors nor oxygen chemoreceptors can account for the observed hyperventilation at this time. Recent studies (Heisler et al. 1988; R. Kinkead and S. F. Perry, unpublished results; Wood et al. 1990) have shown that relatively small changes in blood pH and/or have marked effects on ventilation in fish. Moreover, injection of bovine carbonic anhydrase into the circulation of rainbow trout not only reduced the rise in but also significantly depressed the postexercise hyperventilation (Perry and Wood, 1989). Thus, we suggest that adrenergic inhibition of RBC CO₂ excretion and the associated respiratory acidosis after vigorous exercise serve to increase gill ventilation in the absence of other appropriate ventilatory stimulants.

This research was supported by NSERC of Canada Operating and Equipment grants to S.F.P. and C.M.W. and an NSF grant (DCB-8608728) to P.J.W. We are grateful to Dr R. Motais for helpful discussions.