INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

DURING RECENT YEARS, the hormonal regulation of carbohydrate metabolism in fish, especially in liver, has received increasing attention. It has been demonstrated that the mobilization of glucose from the liver is stimulated by hormones such as catecholamines, cortisol, and glucagon (4, 5, 10, 25), in particular during stress conditions such as hypoxia (43). In contrast, little information is available on the regulation of lipid metabolism during hypoxia in teleost fish. Whereas in mammals, hypoxia results consistently in increases of free fatty acid (FFA) levels (23, 24), the effect of hypoxia on FFA levels in fish is not clear, and the available results are conflicting. In the hypoxia-tolerant goldfish, carp, and bream, plasma FFA levels were found to decrease, whereas FFA concentrations increased in trout, lamprey, and plaice (7, 13, 22). Recently, we demonstrated that a maximal exposure to severe hypoxia, under well-controlled conditions, resulted in a considerable decrease of plasma FFA levels in carp, while a small decline was found in rainbow trout (38).

It has been proposed that a decline of FFA levels would be a secondary effect of the increased availability of carbohydrates during hypoxia or catecholamine administration by enhancing the rate of fatty acid reesterification (7, 22, 32). Moreover, the increased availability of carbohydrates was assumed to result in a stimulation of insulin secretion, leading to inhibition of lipolysis and stimulation of lipogenesis (13, 17). However, several arguments may be proposed against this hypothesis. First, a decrease of plasma FFA levels is not consistently accompanied by hyperglycemia in several fish species (13). Second, elevated plasma glucose levels do not necessarily result in an increase of insulin levels in fish (18). Third, the uptake of blood glucose in fish is normally quite low because of the low activity of hexokinase (11, 21). Finally, during stress conditions in mammals, when fuel mobilization is favored (e.g., exercise or hypoxia), insulin levels are suppressed as a result of a catecholamine-mediated inhibition of the insulin secretion (20, 30). Thus a decrease of plasma FFA levels during hypoxia must be caused by a reduction of lipolysis. Species-specific differences in plasma FFA responses were observed when fish were injected with catecholamines: a decrease of plasma FFA in goldfish, carp, and bream (7, 13, 16, 17), and an increase of plasma FFA in rainbow trout, perch, and scorpion fish (13, 22). Because hypoxia is a potent stimulator of the release of catecholamines in fish, these hormones may play a pivotal role in the mobilization of FFA.

In a study (37) we showed that, although both catecholamines induce a significant hyperglycemia, stress-free infusion of epinephrine induces an increase of plasma FFA levels in carp whereas infusion of norepinephrine induces a significant decrease of plasma FFA concentrations. Mammalian adipose tissue and adipocytes always respond to catecholamines with lipolysis. The response of carp to norepinephrine, i.e., a decrease of blood FFA levels, was a new phenomenon. Considering the strong lipolytic character of these hormones in mammals, a reduction of lipolysis during hypoxia in fish is puzzling and strongly suggests a different mechanism of regulation. From a biological point of view, it is evident that suppression of lipolysis can have a protective effect on the animal under hypoxic conditions. Since the flux through the -oxidation is high during normoxic conditions, fatty acids and their metabolites accumulate rapidly during anaerobiosis in mammalian tissues (19) and will quickly start to disrupt biomembranes. Anoxia-induced membrane damage, however, does not occur in fish (37), which may be related to a better control of lipid mobilization.

On the basis of the mammalian literature, it is expected that -adrenergic stimulation results in a stimulation of lipolysis. However, suppression of FFA levels may be possible via the stimulation of ₂-adrenoceptors, which are mostly coupled to adenyl cyclase via the inhibitory G_i protein, resulting in a decrease of cAMP levels, thus decreasing triacylglycerol lipase activity (12, 34). However, at present it is not known whether ₂-adrenoceptors are expressed in fish lipid depot tissues. In this study we first tested the hypothesis that the decrease of plasma FFA in carp is mediated by the ₂-adrenoceptor. The results, however, indicate that, in addition to the ₂-adrenoceptor, the ₁-adrenoceptor is importantly involved. The first hypothesis was tested by selectively inhibiting and stimulating ₂-adrenoceptors using yohimbine and clonidine, respectively. Since the outcome did confirm the hypothesis only in part, while, in contrast, isoproterenol infusion did result in strongly reduced FFA levels, it was assumed that ₁- and/or ₂-adrenoceptors were involved. Therefore the selective ₁- and ₂-antagonists atenolol and ICI-118,551, respectively, were applied in combination with isoproterenol.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Chemicals. Isoproterenol HCl, yohimbine HCl, and clonidine HCl were obtained from Sigma (St. Louis, MO). Norepinephrine HCl was obtained from BDH (Poole, UK), and atenolol HCl and ICI-118,551 were kind gifts from Zeneca (Wilmslow, Cheshire, UK).

Animals. Common carp (1.0-2.5 kg) were purchased from the Dutch Fisheries Organisation (OVB, The Netherlands). They were fed ad libitum with carp feed from Trouw (Putten, The Netherlands) and were kept in local running tap water in the laboratory for at least 2 mo. The fish were acclimated to a temperature of 20°C and a 14:10-h light-dark cycle. In total 51 experimental animals were used.

Preexperimental protocol. Experiments were performed in a recirculation system as described by Van Dijk et al. (35). Fish handling and surgery was performed as described by van Raaij et al. (38). The cannula was connected to a Y-piece just outside the fish and was split into two PE-50 tubes. One tube (45 cm) was used for blood sampling while the other tube (90 cm) was used for infusion. After cannulation, the fish were placed in the flow boxes and left for 2 days to recover from surgery before starting the experiments. The PE cannula was filled with a viscous solution of polyvinylpyrrolidone (PVP; Merck, Darmstadt, Germany; 1 g PVP/ml saline) containing 500 IU/ml heparin (Sigma). The heparin in the PVP solution prevents clot formation in the cannula, whereas the diffusion of heparin into the blood is negligible, thus preventing possible activation of endothelium-bound lipoprotein lipase.

Experiments. Experiments were started between 8:30 and 9:30 A.M. by taking two blood samples from each fish at t = 0.5 and 0 h (before infusion) to determine initial values. After the second blood sample, the 90-cm PE tube was connected to a 0.5-ml Hamilton syringe containing one of the five solutions: 1) carp Ringer, or carp Ringer containing 2) norepinephrine, 3) yohimbine (₂-antagonist) + norepinephrine, 4) clonidine (₂-agonist), 5) isoproterenol (-agonist). Because the volume of PE tubes was known, we were able to fill the cannula with the test solution just before entering the circulation (negligible dead space). The syringe was then placed in a microinfusion pump (Fine Mechanic Service Dept., Leiden, The Netherlands), which was set at a rate of 3.6 µl/min. The solutions were infused in the dorsal aorta at t = 0 h. Norepinephrine was infused over a period of 90 min at a rate of 2 µg · min¹ · kg¹ (n = 5). Yohimbine was infused as a bolus of 5 mg/kg, followed by the same norepinephrine infusion protocol. Isoproterenol was infused over 90 min at 1 µg · min¹ · kg¹ (n = 6). Clonidine was infused at three, stepwise-increasing levels, i.e., 0.01, 0.1, and 1 µg · min¹ · kg¹ (n = 6) each for 1 h. Control animals (n = 6) received an infusion of carp Ringer saline only.

Infusion quantities. The chosen infusion rates and concentrations of the different drugs were mainly based on the predicted end concentrations and on previously published data. The effective dose for norepinephrine in carp is taken from recent hypoxia studies: during normoxia norepinephrine levels were found to be 0.25 ± 0.07 ng/ml; this level rose to 15 ng/ml during moderate hypoxia and during anoxia to 49 ng/ml (38). This level corresponds to 0.5 × 10⁶ M. The applied infusion protocol for norepinephrine was similar to the one we used previously (37), i.e., 2 µg · min¹ · kg¹. Since the clearance of isoproterenol is ~10-fold lower than that of norepinephrine (preliminary observations), we infused isoproterenol at 1 µg · min¹ · kg¹, which at an extracellular volume of 8% would end up, after 90 min infusion, at ~10⁵ M. The yohimbine dose was based on mammalian literature: 0.2-3.0 mg/kg (2, 12, 40). We applied 5 mg/kg, assuming that this high concentration would block ₂-adrenoceptors virtually completely. To choose the appropriate concentration for clonidine was more difficult, because it is a partial ₂-agonist, and such agonists may turn into antagonists at higher concentrations. Clonidine is thus far only applied in mammals with an effective dose of 1-3 µg · min¹ · kg¹ (3, 8). Therefore we decided to apply three, stepwise-increasing concentrations: 0.01, 0.1, and 1.0 µg · min¹ · kg¹. The bolus infusion of isoproterenol was calculated to obtain an initial level of ~10⁶ M, which is enough to activate the -adrenoceptors. The doses of ICI-118,551 and atenolol chosen were reported previously to induce an effective blockade of mammalian vascular ₁- and ₂-adrenoceptors (28).

Analytical procedures. Blood samples (450 µl) were taken with ice-cooled, heparin-flushed, gastight precision syringes (Hamilton) and were directly placed on ice. Of the whole blood, 50 µl were used for hemoglobin (Hb) and hematocrit (Hct) measurements; the remainder was centrifuged for 5 min at 10,000 g (Eppendorf 5415), and plasma was separated directly. Aliquots of untreated plasma were stored at 80°C until used for analysis of FFA (50 µl). For the measurement of lactate and glucose, 100 µl of plasma were diluted to 20% with trichloro-acetic acid solution (6% vol/vol) to precipitate plasma proteins. The supernatant was neutralized with 1 M K₃PO₄ and stored at 20°C for up to 3 days until analysis. For the assay of catecholamines, 100 µl of plasma were mixed immediately with 15 µl of a preservative solution (22 mg EDTA/ml) and stored at 80°C until analysis.

Presentation and statistics. The results of the 1.5-h infusion experiments were normalized to the mean of the first two sample points at t = 0.5 and 0 h (control=100%) and are presented as means ± SE. The results of the bolus infusion experiments are presented as means ± SE. Statistical significance between experimental data points and initial values and/or control (saline-infused group) values was defined at P < 0.05. Statistical analysis was performed with Sigmastat using the Student's two-tailed t-test. Since the data in Table 2 are not normally distributed, statistical differences were analyzed with the nonparametric Friedman test (Friedman repeated-measures analysis of variance on ranks), followed by the multiple-comparisons Dunnett's test; comparisons were made with respect to the initial value (t = 0.5 h). Comparisons between the different series and the saline controls were based on the Mann-Whitney rank sum test.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Data on Hb, Hct, and mean cellular hemoglobin concentration (MCHC) are not presented because no changes were observed. Hct values were ~30 and declined by maximally 10% during the experiments because of blood sampling. The MCHC values stayed constant at 18 throughout the experiments.

The plasma levels of lactate and epinephrine hardly changed during the different protocols. Therefore a limited number of data are presented in Table 1: 4 of 10 sampling points (t = 0.5, 1.5, 5.5, and 48 h). Under all conditions the lactate levels remained below 0.6 mmol/l, indicating the absence of struggling and anaerobic metabolism. The epinephrine levels at the reference points (first 2 samples in each series) and in the control series were extremely low, quite often below the detection level (10 pg/ml), also indicating low-stress conditions.

All norepinephrine levels measured in the nine series are presented in Table 2. The norepinephrine concentrations remained rather low during and after the different infusions except for the infusion experiments (norepinephrine and yohimbine + norepinephrine). Over the period t = 0.5 to 3.5 h all levels were significantly increased. With the norepinephrine infusion the highest value occurred at t = 3.5 h: 297 ng/ml, which is about a 100-fold increase compared with the saline controls. The combination yohimbine + norepinephrine resulted in even much higher values: the highest level of 3,545 ng/ml was found at t = 1.5 h, which is roughly a 1,000-fold increase compared with the saline controls.

The initial values of the FFA and glucose levels during the different protocols are presented in Table 3. The levels are close to previously described levels for carp (37).

The relative changes in plasma FFA and glucose content due to infusion of the different compounds are shown in Figs. 1-5. Figure 1 shows the FFA and glucose values of the control (saline) series; the data stay constant throughout 48 h and can thus be used for reference. Infusion of norepinephrine (Fig. 2A) results in a transient increase of glucose and was accompanied by a transient decrease of FFA. The glucose increase was significant at the end of the 1.5-h infusion period and stayed elevated until t = 9.5 h. The glucose recovery was most prominent between t = 3.5 and 5.5 h. In contrast to glucose the FFA decline was already significantly lower at 15 min after the start of the norepinephrine infusion. A 50% decline was reached after 1.5 h of infusion, after which a very slow recovery occurred; at t = 9.5 h the plasma FFA level was still 25% below the control level. At 24 and 48 h the data were not different from the controls.

View larger version (15K): [in this window] [in a new window]   Fig. 1.   Effect of arterially infused saline on the plasma concentrations of glucose () and free fatty acid (FFA; ) in common carp. Values are normalized with respect to the mean initial concentration (t = 0.5 and 0 h) and are expressed as means ± SE. No significant changes were observed during the experimental protocol.

View larger version (21K): [in this window] [in a new window]   Fig. 2.   Effect of arterially infused norepinephrine (A) and norepinephrine in combination with yohimbine (B) on the plasma concentrations of glucose () and FFA () in common carp. Yohimbine was infused as a bolus (5 mg/kg) followed by the same norepinephrine infusion protocol as presented in A. Values are normalized with respect to the mean initial concentration (t = 0.5 and 0 h) and are expressed as means ± SE. Significant differences: * P < 0.05 vs. initial value,  P < 0.05 vs. control group,  P < 0.05 vs. initial value and control group.

View larger version (18K): [in this window] [in a new window]   Fig. 3.   Effect of arterially infused clonidine on the plasma concentrations of glucose () and FFA () in common carp. Values are normalized with respect to the mean initial concentration (t = 0.5 and 0 h) and are expressed as means ± SE. Significant differences: * P < 0.05 vs. initial value,  P < 0.05 vs. initial value and control group.

View larger version (17K): [in this window] [in a new window]   Fig. 4.   Effect of arterially infused isoproterenol on the plasma concentrations of glucose () and FFA () in common carp. Values are normalized with respect to the mean initial concentration (t = 0.5 and 0 h) and are expressed as means ± SE. Significant differences: * P < 0.05 vs. initial value,  P < 0.05 vs. initial value and control group.

View larger version (23K): [in this window] [in a new window]   Fig. 5.   Effect of an arterially injected bolus infusion of saline (), isoproterenol alone (19.8 µg/kg; ), and isoproterenol preceded by a bolus infusion of atenolol (213 µg/kg; ) or ICI-118,551 (250 µg/kg; ) on the plasma concentrations of glucose (A) and FFA (B). Values are expressed as means ± SE. Significant differences: * P < 0.05 vs. initial value;  P < 0.05 vs. initial value and control group; 3 P < 0.05, all 3 isoproterenol treatments vs. initial value and control group; # P < 0.05, atenolol or ICI-118,551 treatment vs. isoproterenol alone.

The same experiment was repeated for the combination of norepinephrine with the ₂-antagonist yohimbine (Fig. 2B). The glucose data were very similar to norepinephrine alone (Fig. 2A), with the same first significant point after 1.5 h infusion, and with the rapid recovery between 3.5 and 5.5 h. Remarkably different, however, was the FFA pattern: the FFA decline in the presence of yohimbine was clearly retarded, reaching the first significant point at 2.5 h; thereafter the data were similar to those without yohimbine.

The effect of infusion of the ₂-agonist clonidine is shown in Fig. 3. The concentration was increased stepwise (0.01, 0.1, and 1.0 µg · min¹ · kg¹), with each concentration being infused during 1 h. The glucose level did not significantly change except at t = 9.0 h, where the concentration was reduced by 50%. The plasma FFA levels were decreased to some extent at 0.75 and 1.75 h, the other points particularly after 3.25 h, and later showed a large variance with values around 30-40% above control.

Infusion with the nonselective -adrenoceptor agonist isoproterenol (Fig. 4) showed opposite effects for plasma glucose and FFA levels. A strong and highly significant increase during the infusion was observed for glucose. After the infusion the levels stayed high at ~300% of the control level and declined to control values between 3.5 and 9.5 h. The FFA response was also very clear but inverse. A decline of 40% below control occurred during the infusion; the levels stayed low until 3.5 h and slowly recovered to control values at 9.5 h.

To distinguish between ₁- and ₂-adrenoceptor effects, combinations of isoproterenol with selective ₁- and ₂-antagonists were tested, which were injected as bolus infusions. At several time points before and after the bolus injection(s) (t = 1.25, 0.25, 1.5, and 48 h), the isoproterenol concentration in the plasma was measured. Injection of isoproterenol caused a strong but transient increase in concentration. After 15 min the values were 1,000-6,000 ng/ml (5-25 × 10⁶ M) followed by a decrease to values of 100-600 ng/ml 1.5 h postinjection. After 48 h no differences with the controls were observed anymore. In Fig. 5 the effects on glucose and FFA levels are shown. The plasma glucose levels (Fig. 5A) increased rapidly after the infusion with isoproterenol. The highest level was reached at t = 1.5 h; the plasma concentrations stayed elevated until t = 5.5 h. After infusion with isoproterenol + atenolol and isoproterenol + ICI-118,551, the glucose levels increased significantly above the control values in a pattern similar to isoproterenol alone. The infusion with each of the antagonists, however, resulted in significantly lower glucose levels at t = 2.5, 3.5, and 5.5 h. This indicates that both ₁- and ₂-adrenoceptors are involved in glucose mobilization.

In Fig. 5B the effects of the different infusion protocols on the plasma FFA levels are shown. Infusion with a bolus of isoproterenol showed a small but nonsignificant decline of the FFA levels. The combination of isoproterenol + ICI-118,551 showed a fast decline of the plasma FFA level with a significant point at t = 0.5 h, recovering to control values at t = 3.5 h. Infusion with isoproterenol + atenolol, however, showed a completely opposite effect: at three time points the values were significantly higher than the initial levels. These observations indicate that the plasma FFA levels are governed by inhibitory ₁- and stimulatory ₂-adrenoceptors.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

-Adrenoceptors. The results shown in Fig. 2 confirm our earlier findings that norepinephrine increases plasma glucose and decreases plasma FFA in carp (37), the latter being in sharp contrast to the mammalian response. The FFA decline observed after norepinephrine infusion (Fig. 2A) is similar to the one described in our former study: a significant fall already at 15 min, declining further to 50% at the end of the infusion period, followed by a very slow recovery. Since we found in carp that epinephrine increased the plasma FFA level (37), it was hypothesized that the antagonistic effect of both catecholamines may be mediated by their differential effects on ₂- and -adrenoceptors. ₂-Adrenoceptors act through G_i proteins, which inhibit adenyl cyclase. On the other hand, -adrenoceptors stimulate adenyl cyclase via G_s proteins. Recently it has been suggested that the ratio of ₂- to -adrenoceptors determines the lipolytic capacity in adipose tissues (34). Thus the ratio of activated ₂- to -adrenoceptors could determine the lipolytic rate in the adipose tissue, which could be controlled by the ratio of norepinephrine to epinephrine.

-Adrenoceptors. The involvement of -adrenoceptors in lipolysis was tested by the infusion of isoproterenol. This general -adrenoceptor agonist has been applied in many studies including fish. The effect of isoproterenol infusion is presented in Fig. 4 and shows an increase of blood glucose and a decrease of plasma FFA. The first effect, i.e., glucose release from the liver by -adrenoceptor stimulation, was expected. The metabolic effect of stimulation of -adrenoceptors on fish hepatocytes has been studied extensively, although these studies were almost exclusively aimed at carbohydrate metabolism (6, 27). Both epinephrine and norepinephrine stimulate predominantly glycogenolysis, and at a lower rate also gluconeogenesis (10, 41). The adrenoceptors responsible for glucose release in trout liver are, according to McKinley and Hazel (14), of the ₂-type, not the - or ₁-type. Also Reid et al. (27) showed that -adrenoceptors in trout hepatocytes are of the ₂-type. However, the results shown in Fig. 5A suggest the presence of two different -adrenoceptors on the liver cells. Atenolol and ICI-118,551 are selective ₁- and ₂-adrenoceptor antagonists, respectively, whereas isoproterenol is a nonselective -adrenoceptor agonist. In Fig. 5A the conditions isoproterenol + atenolol and isoproterenol + ICI-118,551 both reveal a significantly lower glucose response than the infusion with isoproterenol alone, which shows that ₁- as well as ₂-adrenoceptors are involved in glucose release. The experiments of McKinley and Hazel (14) and Reid et al. (27) were carried out with trout, whereas our results were with carp. Thus there may possibly be species-related differences. Also our experiments are whole animal responses; therefore other tissues may be involved as well. To prove that the glucose response of carp is mediated by both ₁- and ₂-adrenoceptors on the liver, appropriate in vitro experiments should be carried out with hepatocytes.

Perspectives

The regulation of lipolysis in fishes by catecholamines is an intriguing problem. The involvement of -adrenoceptors in inhibition of lipolysis as described in this study is a novel phenomenon and suggests a mechanism different from the well-described mammalian system. For analysis of this mechanism it is necessary to carry out in vitro experiments with adipose tissues or adipocytes. Ongoing experiments with isolated adipocytes from different fish species aim to characterize the receptor(s) and transduction mechanisms involved. The inhibition of lipolysis by norepinephrine in some fish species is likely a protection mechanism against accumulation of amphiphiles. This accumulation is responsible for tissue damage in mammals after ischemic and hypoxic insults. The transition from water to air breathing during evolution may be connected to a changed function of norepinephrine. Comparative studies are planned to test this hypothesis.