In vivo regulation of rainbow trout lipolysis by catecholamines

Fatty acids play key physiological roles by acting as membrane components,oxidative fuels and metabolic signals. Therefore, the regulation of fatty acid supply, or lipolysis, affects many fundamental life processes(Clarke et al., 1997; Raclot and Oudart, 1999; Hulbert et al., 2007). In the intact organism, lipolysis is measured as the rate of appearance of glycerol(R_a glycerol) (Wolfe,1992), and it has been thoroughly investigated in mammals, mainly in humans (Klein et al., 1989; Wolfe et al., 1990; Klein et al., 1995; Quisth et al., 2005). Unfortunately, the information available for other vertebrates is restricted to two in vivo studies on birds(Bernard et al., 2002; Vaillancourt and Weber, 2007)and one on fish (Bernard et al.,1999). The lack of data for ectotherms is particularly unfortunate because an adequate supply of fatty acids is crucial to restructure membrane phospholipids during homeoviscous adaptation(Hazel and Williams, 1990) and to fuel endurance swimming in fish (Moyes and West, 1995).

The rate of lipolysis is modulated by several hormones, including catecholamines, insulin and glucagon(Coppack et al., 1994; Londos et al., 1999). In mammals, lipolysis is stimulated by norepinephrine (NE) as well as epinephrine(Epi) (Fain and Garcia-Sainz,1983; Nonogaki,2000), but these hormones have effects that are controversial in fish. Various species show different responses in vitro, and the effects of catecholamines on fish lipolysis have not been measured in vivo using tracer methods. Overall, however, NE appears to inhibit lipolysis in fish adipose tissue (Farkas,1967a, Farkas,1967b; Vianen et al.,2002), although not in all species(Farkas, 1967b). NE-induced inhibition of fish lipolysis also appears to be supported by in vivoobservations that the hormone decreases the level of plasma non-esterified fatty acids (NEFA) in many species(Farkas, 1967b; Minick and Chavin, 1973; Ince and Thorpe, 1975; Van Raaij et al., 1995), with a few exceptions (Leisbon et al.,1968; Plisetskaya,1980). However, to assume that the plasma concentration of NEFA reflects changes in lipolytic rate might lead to erroneous conclusions(Haman et al., 1997). Even for Epi, the response can be very different between fish and mammals. Some studies report in vivo inhibition of lipolysis by Epi for some fish species(Minick and Chavin, 1973; Ince and Thorpe, 1975),whereas others show no local effect on adipose tissue(Migliorini et al., 1992). The only information available for rainbow trout shows activation of lipolysis by both catecholamines in isolated hepatocytes(Van Heeswijk et al., 2006)but no effect of Epi on red muscle or on the plasma concentration of NEFA in vivo (Bilinski and Lau,1969; Perrier et al.,1972).

The goal of the present study was to investigate the effects of catecholamines on R_a glycerol in intact fish to obtain an integrated hormonal response rather than tissue- or cell-specific contributions to total fatty acid supply. Our aim was to quantify the effects of NE and Epi on the lipolytic rate of rainbow trout using in vivotracer kinetics. Based on the balance of evidence presently available, it was hypothesized that trout lipolysis would be inhibited by NE and stimulated by Epi. We also examined the relationship between R_a glycerol and glycerol concentration to determine whether the concentration of glycerol can be used as a reliable index of fish lipolysis.

Adult rainbow trout (Oncorhynchus mykiss Walbaum) were obtained from Linwood Acres Trout Farm (Campbellcroft, Ontario, Canada). They were kept in a 1300 litre flow-through holding tank in dechloraminated, well-oxygenated water at 13°C under a 12 h:12 h light:dark photoperiod. Animals were acclimatized to these conditions for at least one month before being subject to experimentation. The same water quality and photoperiod were used during the measurements. Trout were fed floating fish pellets (Martins Mills, Elmira,Ontario, Canada) until satiation five times a week and were randomly assigned to three groups: saline control (body mass 460±52g; N=4), Epi(485±35g; N=8) and NE (514±40g; N=8).

Fish were fasted for 48 h before surgery and did not feed during recovery or measurements. Double catheterization was performed as described previously(Haman and Weber, 1996). The animals were anesthetized in a solution of 0.1 g l^–1ethyl-N-aminobenzoate sulfonic acid (MS-222) buffered with 0.2 g l^–1 sodium bicarbonate, and two PE-50 catheters (Intramedic,Clay-Adams, Sparks, MD, USA) were implanted in the dorsal aorta. Both catheters were filled with Cortland saline and sutured to the roof of the buccal cavity. After surgery, the animals were placed in opaque Plexiglas chambers (60×16×18 cm) supplied with the same quality water as the acclimatization tank at a rate of 5–6 l min^–1. This low rate was selected to maintain normoxic conditions without causing swimming. During recovery from surgery, the catheters were regularly checked for patency and flushed with Cortland saline containing sodium citrate as an anticoagulant(13 μmol ml^–1). Only animals with a hematocrit (packed cell volume) >20% after recovery from surgery were used in experiments. Glycerol kinetics were then measured in the same chambers 24 h after surgery.

The R_a glycerol, also commonly called the lipolytic rate, was measured by continuous infusion of 2-[³H]glycerol (37 GBqmmol^–1; Amersham, Buckinghamshire, UK), as described previously (Bernard et al.,1999). The infusate was prepared daily by drying 0.94MBq of the radioisotope under N₂ and resuspending in Cortland saline. This solution was infused for 3 h at 1 ml h^–1 through one of the catheters using a calibrated syringe pump (Harvard Apparatus, South Natick,MA, USA). Tracer infusion rates ranged between 375×10³ and 148×10⁴ d.p.m. kg^–1 min^–1. Because the infusate had a high specific activity, the total amount of glycerol administered (labeled and unlabeled) was <0.5 nmol kg^–1 min^–1. Isotopic steady state was reached in <1 h. Baseline glycerol kinetics were quantified between 60 and 90 min of infusion before investigating the effects of catecholamines. After quantifying baseline kinetics, isotope infusion was continued for 30 min while saline, Epi or NE were administered (see next section) and for an extra 60 min after catecholamine administration to monitor recovery from hormone treatment.

A second calibrated syringe pump was activated after 90 min of tracer infusion to administer saline (control), NE or Epi for 30 min at 1 ml h^–1. To avoid light-induced breakdown of catecholamines,hormone solutions were prepared under red, low-intensity light, and similar light conditions were used during administration. The rates of catecholamine administration were 0.45 nmol kg^–1 min^–1 for NE and 1.34 nmol kg^–1 min^–1 for Epi. These values were selected to achieve plasma concentrations of ∼170 nmol l^–1 (NE) and ∼500 nmol l^–1 (Epi) that mimic the hormone levels noted after exhaustive swimming(Butler et al., 1986). A previous study shows that this rate of epinephrine administration leads to plasma levels of 500 nmol l^–1 in trout(Weber and Shanghavi,2000).

Eight 0.3 ml blood samples were drawn in each experiment. Plasma was immediately separated by centrifugation (5000 g for 10 min)and stored at –80°C until subject to analyses. Each plasma sample was divided into two aliquots to measure glycerol concentration by spectrophotometry (Weber et al.,1993) and glycerol activity by scintillation counting (Beckman Coulter CS6500, Palo Alto, CA, USA). Infusion of 2-[³H]glycerol only leads to significant radioactivity in plasma H₂O, glucose and glycerol. The specific activity of glycerol was determined after drying plasma to eliminate labeled water and, after correction for glucose activity, using thin layer chromatography, as described previously(Bernard et al., 1999).

The rate of appearance of glycerol, or lipolytic rate, was calculated using the steady-state equation of Steele(Steele, 1959). This equation was used because it provides more accurate estimates of flux than the non-steady-state equation when, first, specific activity varies little over time and, second, when rapid equilibration takes place within the metabolite pool (caused by a rapid turnover rate relative to pool size)(Beylot et al., 1987; Wolfe et al., 1990). The effects of hormone administration over time were assessed by one-way analyses of variance with repeated measures (RM ANOVA). When significant changes were detected, the Holm-Sidak method was used to determine which means were different from baseline (Figs 1, 2, 3). Differences in mean glycerol concentrations and R_a glycerol between control and catecholamine-treated animals were evaluated with unpaired t-tests (Table 1). The values given are means ± s.e.m. A level of significance of P<0.05 was used in all tests.

The hematocrit was measured before and after treatment with saline(control), Epi or NE. The hematocrit did not change over time and averaged 24±2% (saline), 25±4% (Epi) and 24±3% (NE). Fig. 1 shows that administration of saline has no effect on plasma glycerol concentration,specific activity or R_a glycerol (P>0.05). These control experiments demonstrate that the isotopic and concentration steady state of glycerol can be maintained for 3 h.

Fig. 2 shows the effects of administration of NE on the concentration of plasma glycerol, specific activity and R_a glycerol. Baseline values for glycerol concentration and R_a glycerol averaged 0.20±0.05μmol ml^–1 and 4.4±0.6 μmol kg^–1min^–1, respectively. The glycerol concentration and R_a glycerol were both decreased by norepinephrine(P<0.05). The concentration reached a minimal value of 0.10±0.03 μmol ml^–1 and R_aglycerol a minimal value of 2.3±0.7 μmol kg^–1min^–1 after 30 min of NE administration, and both returned to baseline during recovery (P>0.05).

Fig. 3 shows the effects of administration of Epi on the concentration of plasma glycerol, specific activity and R_a glycerol. Baseline values for glycerol concentration and R_a glycerol averaged 0.22±0.05μmol ml^–1 and 4.4±0.5 μmol kg^–1min^–1, respectively. The glycerol concentration and R_a glycerol were both stimulated by Epi(P<0.001). The concentration reached a maximal value of 0.55±0.07 μmol ml^–1 and R_aglycerol a maximal value of 8.2±0.8 μmol kg^–1min^–1 after 20 min of Epi administration. The concentration and R_a glycerol stayed elevated until the end of Epi administration (P<0.001), but they returned to baseline during recovery (P>0.05). Table 1 summarizes the mean values for plasma glycerol concentration and R_a glycerol during the final 10 min of administration of saline, Epi and NE. Both parameters were decreased by NE (P<0.05)and increased by Epi (P<0.001).

Fig. 4 plots the plasma glycerol concentration versus R_a glycerol (or lipolytic rate) during the final 10 min of administration of saline (stars), Epi (filled circles) and NE (unfilled circles). Separate linear regressions were fitted for the three groups. For Epi, the relationship had a positive slope different from 0 (r²=0.48, P<0.05). The concentration and rate of appearance were not significantly related for saline(r²=0.03, slope not different from 0, P=0.905)and for NE (r²=0.05, slope not different from 0, P=0.830).

Fig. 5 summarizes the ratios between R_a glycerol and oxygen consumption for vertebrate species measured to date at rest or during exercise. This ratio is an index of lipolytic rate standardized for differences in metabolic rate. All values are very similar, ranging from 0.009–0.031, except for resting fish, which show an extreme ratio of 0.115.

The present study investigates the kinetics of glycerol in rainbow trout and provides the first in vivo measurements of the effects of catecholamines on the lipolytic rate of an ectotherm. It shows that NE inhibits lipolysis in trout, instead of stimulating it as it does in mammals. Epi, by contrast, has the same activating effect in both groups of vertebrates. The changes in plasma glycerol concentration are weakly correlated with R_a glycerol (Epi stimulation) or not correlated at all (NE inhibition), making glycerol concentration a very poor predictor of lipolysis. Comparing all in vivo lipolytic rates measured to date reveals that trout maintain a disproportionately high resting R_a glycerol for their low metabolic rate compared with that of endotherms. No demonstrated functional explanation for maintaining elevated lipolytic rates in the resting state is available. However, we suggest that ectotherms might always need high rates of fatty acid supply for rapid remodeling of membrane phospholipids while coping with changes in body temperature.

R_a glycerol is recognized as the best estimate of lipolysis, and it has been widely used to quantify whole-body lipolytic rates in vertebrates, including mammals(Issekutz et al., 1975; Wolfe, 1992; Weber et al., 1993; McClelland et al., 2001; Henderson et al., 2007; Houser et al., 2007), birds(Bernard et al., 2002; Bernard et al., 2003; Vaillancourt and Weber, 2007)and fish (Bernard et al.,1999). Using R_a glycerol for this purpose is recommended because the glycerol produced by lipolysis must obligatorily appear in the circulation, in situ recycling being prevented by the negligible activities of glycerol kinase in the adipose tissue and muscle of vertebrates (Newsholme and Taylor,1969; Brooks et al.,1982; Reshef et al.,2003). By contrast, the rate of appearance of NEFA(R_a NEFA) cannot be used to estimate the lipolytic rate because a significant fraction of the fatty acids produced by lipolysis is reesterified in situ. This important difference between glycerol and NEFA is illustrated by the fact that typical ratios of R_aNEFA/R_a glycerol measured in vivo or in vitro are well below 3.0 (Bernard et al., 1999; Henderson et al.,2007). The other reason why R_a glycerol adequately reflects lipolysis is because this pathway is the only significant source of glycerol (Issekutz et al.,1975; Wolfe, 1992; Houser et al., 2007). It could be argued that this might not be true for fish because some species can use an alternative pathway to produce glycerol as an antifreeze from glucose and amino acids (Raymond, 1992; Raymond and Driedzic, 1997; Driedzic et al., 2006). However, this particular adaptation for cold tolerance depends on high activities of hepatic glycerol-3-phosphatase, an enzyme present only in unique species such as rainbow smelt (Osmerus mordax) but virtually absent in species lacking freeze tolerance such as trout and other members of the smelt family (Treberg et al.,2002; Lewis et al.,2004; Walter et al.,2006). In trout, there is no evidence that glycerol released in the circulation could be of nonlipolytic origin.

The results from both studies are consistent because primary reesterification was estimated at 68% of all fatty acids released by lipolysis(Bernard et al., 1999), and total reesterification estimated at 87% (present study), suggesting that the difference (19%) is explained by secondary reesterification.

At the whole-organism level, the present study shows that NE (inhibition)and Epi (activation) have opposite effects on the lipolytic rate of rainbow trout (Figs 2 and 3). Therefore, the regulation of lipolysis by catecholamines might play a particularly important role during hypoxia, an environmental stress causing the strong release of NE(Montpetit and Perry, 1998). This hormone probably causes the decreases in plasma NEFA concentration(Van Raaij et al., 1996) and NEFA turnover rate (Haman et al.,1997) previously observed when trout are exposed to hypoxic conditions. The inhibition of lipolysis probably is also responsible for the decreases in plasma NEFA concentration reported after administration of NE in bream Abramis brama and pike-perch Sander lucioperca(Farkas, 1967b), as well as in goldfish Carassius auratus(Minick and Chavin, 1973),pike Esox lucius (Ince and Thorpe, 1975) and carp Cyprinus carpio(Van Raaij et al., 1995). These in vivo effects are consistent with the inhibition of lipolysis demonstrated in vitro for adipose tissue in carp, pike-perch(Farkas, 1967b) and tilapia Oreochromis mossambicus (Vianen et al., 2002). Administration of Epi causes the stimulation of lipolysis (Fig. 3), and this result is consistent with previous studies in fish showing Epi-induced increases in plasma NEFA concentration [scorpion fish Scorpaena porcus (Leisbon et al.,1968), eel Anguilla anguilla(Larsson, 1973), lamprey Petromyzon marinus (Plisetskaya,1980), plaice Pleuronectes platessa(White and Fletcher, 1989) and carp (Van Raaij et al.,1995)], although no such increase could be demonstrated in trout Oncorhynchus mykiss (Perrier et al., 1972).

In mammals, NE and Epi increase fatty acid mobilization from adipose tissue and muscle by activating hormone-sensitive lipase (HSL). The presence of one or more lipases also activated by hormones, homologous or not to mammalian HSL, has been reported in the liver and adipose tissue of rainbow trout(Albalat et al., 2005; Van Heeswijk et al., 2006). In trout hepatocytes, both catecholamines activate lipolysis(Van Heeswijk et al., 2006),but no Epi-induced stimulation of lipolysis was detected in red muscle slices from trout (Bilinski and Lau,1969). Taken together, these in vitro studies suggest that the increase in lipolytic rate observed in the present study after in vivo administration of Epi might be caused by the activation of HSL-like lipases in liver and adipose tissue but not in red muscle. Furthermore, the potential implication of other lipases such as lipoprotein lipase (LPL) or other circulating hormones acting indirectly, such as glucagon, insulin or cortisol, cannot be eliminated (Zechner,1997; Merkel et al.,2002; An et al.,2005). In trout, a homolog of mammalian LPL sensitive to insulin has also been characterized in several tissues(Lindberg and Olivecrona,2002; Albalat et al.,2006; Magnoni and Weber,2007). Therefore, further studies will be needed to distinguish the effects of HSL-like lipases and LPL and to characterize the potential interactions of catecholamines with other hormones. Even though NE causes the stimulation of mammalian lipolysis in vivo, it also triggers local inhibition in adipocytes (Holm,2003). The same mechanism that causes inhibition in mammalian adipocytes might also be responsible for the in vivo lipolytic response reported here in trout (Fig. 2).

The observed changes in R_a glycerol caused by Epi correlated weakly with changes in plasma glycerol concentration, but no such correlation was apparent for NE (Fig. 4). Because Epi causes a much stronger stimulation of cardiac output than NE, the relationship between flux and concentration could be influenced by differential effects of catecholamines on the cardiovascular system (Gannon and Brunstock,1969; Wood and Shelton,1980; Randall and Perry,1992). Previous studies have shown that plasma metabolite fluxes can be increased in two ways: first, by augmenting concentration in plasma(mass-action effect) or, second, by increasing blood flow (perfusion effect)(Weber et al., 1987). Therefore, the stimulating effect of Epi on cardiac output might have amplified the slope of the relationship between the glycerol concentration and R_a glycerol. Interestingly, the stimulating effect of NE on cardiac output appears to have been sufficient to offset the negative mass-action effect of a decrease in plasma glycerol concentration (thereby eliminating a potential relationship between flux and concentration in the NE experiments). A detailed analysis of trout glucose and NEFA metabolism also showed that the plasma concentration of these metabolites does not reflect changes in their flux (Haman et al.,1997). The present study reinforces the idea that changes in metabolite concentration cannot be used to speculate on possible changes in metabolite fluxes unless a clear relationship between the two parameters has been formally established under specified conditions.

In previous in vivo studies, the lipolytic rate has been measured almost exclusively in warm-blooded animals. The rates reported here for trout[2–8 μmol kg^–1 min^–1; Figs 1, 2, 3 and Bernard et al.(Bernard et al., 1999)] fit within the endotherm range (2–72 μmol kg^–1min^–1), but relevant comparisons can only be made if differences in metabolic rate are taken into account. Therefore, the ratios between lipolytic rate and metabolic rate were plotted in Fig. 5 for all vertebrates measured to date. In endotherms, the ratio varies between 0.01 and 0.03, and it does not vary much between rest and exercise because lipolytic rate is stimulated in parallel with metabolic rate. This analysis reveals that swimming trout fall within the normal vertebrate range (0.03) but that resting trout have a drastically different ratio of 0.10, more than six times the average (0.02). What is the physiological significance of this unexpected observation? We have calculated the theoretical metabolic rate needed to oxidize all fatty acids supplied by lipolysis at rest (4.6 μmol glycerol kg^–1 min^–1 or ∼14 μmol fatty acid kg^–1 min^–1). Assuming that energy metabolism is only supported by lipids, this theoretical metabolic rate would be 360μmol O₂ kg^–1 min^–1 (if oleate is considered as an average fatty acid, requiring 26 molecules of O₂ for oxidation, and if the contribution of glycerol oxidation is ignored). Because the real metabolic rate of a resting trout is only 46μmol O₂ kg^–1 min^–1(Burgetz et al., 1998), we calculate that 13% of their lipolytic rate can fuel the resting energy metabolism entirely. Therefore, 87% of the fatty acids released by trout lipolysis have to undergo reesterification: a much higher proportion than in endotherms (<70%; see Fig. 5for references). Such high lipolytic rates (present study) and triacylglycerol turnover rates (Magnoni et al.,2008) demonstrate that fatty acid mobilization and reesterification are particularly active in trout. We propose that this rapid cycling of fatty acids might be crucial for restructuring membrane phospholipids and, therefore, might be necessary in all ectotherms for adequate homeoviscous adaptation.

Our results show for the first time that NE inhibits whole-organism lipolysis in trout, rather than stimulating it as in mammals. Our data support previous in vitro studies suggesting that fish and mammals regulate lipid mobilization differently. NE-induced inhibition of fish lipolysis might play a particularly important role during aquatic hypoxia, when survival often depends on regulated metabolic depression. In trout, the changes in lipolytic rate are not well reflected by changes in the concentration of plasma glycerol. Therefore, glycerol concentration cannot be used as an index of glycerol flux (R_a glycerol), a recommendation previously made for other key intermediates of energy metabolism(Haman et al., 1997). Finally,the present study demonstrates that resting trout maintain a disproportionately high lipolytic rate because only 13% of the fatty acids supplied by lipolysis are sufficient to support total energy expenditure. Therefore, most mobilized fatty acids must undergo reesterification (87%). Such cycling implies rapid exchange between fatty acid pools, a metabolic adaptation improving the capacity for remodeling membrane phospholipids with fatty acids of different chain length and degree of saturation. Determining whether such an adaptation is ubiquitous among ectotherms strikes us as a fascinating avenue for future research.

We thank Bill Fletcher for his invaluable help in animal care. This work was supported by an NSERC discovery grant to J.M.W.