Two experiments are described in which the interaction of prolactin, cortisol and growth hormone on hypoosmoregulation in the anadromous brown trout was studied. In experiment 1, fish at the postsmolt stage in freshwater (FW) were given four injections on alternate days of 5 μg cortisol (F)/g in combination with 0, 0.2, 1 or 2 μg ovine prolactin (oPRL)/g. Additional groups received 2 μg oPRL/g or saline as control. In experiment 2, FW parr were given four injections on alternate days with 5 μg F/g and 2 μg ovine growth hormone (oGH)/g in combination with 0, 0.2, 1 or 2 μg oPRL/g. Additional groups received 2 μg oPRL/g, 2 μg oGH/g, 5 μg F/g or saline as control. In both experiments, the fish were subjected to a 48 hr seawater (SW) challenge test 24 hr after the last injection. Muscle water content, plasma osmolality and ion levels, kidney and gill Na , K -ATPase activity were measured. In experiment 1, F-treated fish had better hypoosmoregulatory capacity than control fish as judged by a higher level of muscle water content and lower plasma osmolality after 48 hr in SW. All three doses of oPRL completely abolished this action of F. Gill Na ,K -ATPase activity was stimulated by F and unaffected by oPRL at any dose. In experiment 2, oPRL impaired, whereas F and oGH (injected individually or together) improved performance in the 48 hr SW test relative to control fish, judged by plasma osmolality and muscle water content. Ovine PRL inhibited the combined action of F and oGH in a dose-related manner but could not completely counteract the combined effect of these hormones. F and oGH had additive stimulatory effects on gill Na ,K -ATPase activity. This activation was counteracted by coinjection of oPRL in a dose-related manner but not below the level found in the F-treated fish. We conclude that exogenous oPRL inhibits the hypoosmoregulatory action of F. However, oPRL does not antagonize the F-dependent stimulation of gill Na ,K -ATPase activity. The observed antagonism between PRL and F in SW-acclimation may therefore occur on a target different from gill Na ,K -ATPase.
INTRODUCTION
In salmonids of the genera Salmo and Oncorhynchus, cortisol (F), growth hormone (GH), and prolactin (PRL) are believed to be key hormones in the development of seawater (SW)-tolerance during smoltification (e.g., Björnsson et al., 1989; Young et al., 1989). Both F and GH stimulate hypoosmoregulation by increasing gill chloride cell density and size and Na+,K+-ATPase activity in vivo (Richman and Zaugg, 1987; Madsen, 1990a,c; Bisbal and Specker, 1991; Almendras et al., 1993; Boeuf et al., 1994; Nonnotte and Boeuf, 1995). Gill Na+,K+-ATPase activity is stimulated by F in vitro but is unaffected by GH (McCormick et al., 1991a). Madsen and Bern (1993) showed, however, that insulin-like growth factor I (IGF-I) may be the mediator for GH's stimulatory action in vitro.
PRL, on the other hand, is important for hyperosmoregulation in FW teleosts and is generally thought to regulate branchial ion-exchange by decreasing epithelial permeability (review by Hirano, 1986). PRL also decreases the ability of salmonids to hypoosmoregulate (Hasegawa et al., 1986; Madsen and Bern, 1992; Boeuf et al., 1994), and increases (Salmo salar. Boeuf et al., 1994), decreases (Oncorhynchus mykiss: Madsen and Bern, 1992), or has no effect on (Salmo trutta: Madsen et al., 1995) gill Na+,K+-ATPase activity. Recently, gill Na+,K+-ATPase α-subunit mRNA-levels were shown to increase in S. trutta after treatment with F, recombinant salmon (rs) GH or recombinant bovine (rb) IGF- I, whereas salmon (s) PRL had no effect (Madsen et al., 1995). A few studies have emphasized the importance of endocrine interactions in the control of hypoosmoregulation in salmonids. Cortisol and the GH-IGF-I axis have additive effects on SW- tolerance and gill Na+,K+-ATPase activity in vivo (Björnsson et al., 1987; Madsen, 1990b,c; McCormick, 1996), whereas GH and PRL have antagonistic effects on the development of hypoosmoregulatory ability (Madsen and Bern, 1992; Boeuf et al., 1994). Interactions between PRL and F have not been described yet, and the aim of the present study was to investigate how oPRL affects the hypoosmoregulatory ability of S. trutta treated simultaneously with either F or F in combination with oGH.
MATERIALS AND METHODS
Fish maintenance
Immature anadromous brown trout of mixed sexes were obtained from the Haarkjaer Hatchery, Denmark, where they had been reared in ponds under natural photoperiod and temperature. For Experiment 1, 70 postsmolts, age 1+, weight 62.2 ± 1.5 g (mean ± SEM), were obtained in June 1995; for Experiment 2, 90 parr, age 0+, weight 24.7 ± 0.3 g (mean ± SEM), were obtained in September 1995. The fish were brought to Odense University and acclimated in 2000-I flow-through freshwater tanks (FW, 15°C), and artificial photoperiod (12:12, L:D). They were fed 2% body weight commercial trout pellets once daily. After 14 days in Experiment 1, and one month in Experiment 2 the fish were randomly sorted into groups of ten fish (7 and 9 groups in Exp. 1 and 2, respectively) and transferred to 400-I tanks. A group of untreated fish was left in one tank. Two groups were pooled in each of the other tanks and tagged by adipose fin clipping. Food was withheld from this day onwards.
Experimental protocol and hormone treatment
Hormone treatments were initiated four days after the fish were divided into groups. Prior to intraperitoneal injection, the fish were slightly anesthetized in 0.05% phenoxyethanol. All injections were repeated on alternate days in FW for a total of four injections per fish. The fish were then transferred to dilute seawater (SW, 15°C) 24 hr after the last injection and sampled after 48 hr. The salinity was 27.3 ppt and 24.8 ppt in Exp. 1 and 2, respectively. The following vehicle and hormones were used: saline (0.9% NaCl, 0.05% NaOH, 0.4% bovine serum albumin), oPRL (NIADDK-oPRL-19, Baltimore, MD), F (Na-hydrocortisone hemisuccinate, Sigma, St. Louis, MO), and oGH (NIADDK-oGH-15). All hormones were dissolved in vehicle immediately prior to injection. The fish were injected according to body weight with one of the following combinations and doses of hormones. Experiment 1: 1) none (untreated group), 2) saline, 3) 2 μg oPRL/g, 4) 5 μg F/g, 5) 5 μg F + 0.2 μg oPRL/g, 6) 5 μg F + 1 μg oPRL/g, 7) 5 μg F + 2 μg oPRL/g. The injection volume was 2 μl/g. Experiment 2: 1) none (untreated group), 2) saline, 3) 2 μg oPRL/g, 4) 2 μg oGH/g, 5) 5 μg F/g, 6) 5 μg F + 2 μg oGH/g, 7) 5 μg F + 2 μg oGH + 0.2 μg oPRL/g, 8) 5 μg F + 2 μg oGH + 1 μg oPRL/g, 9) 5 μg F + 2 μg oGH + 2 μg oPRL/g. The injection volume was 3 μl/g.
Sampling and analysis
The fish was stunned by a blow to the head, and blood was drawn from the caudal vessels into a heparinized syringe and immediately centrifuged at 5000 × g for 3 min. The fish was then decapitated. One second gill arch and the posterior part of the body kidney were dissected and placed in sucrose-EDTA-imidazole buffer (SEI; 300 mM sucrose, 20 mM EDTA, 50 mM imidazole, pH 7.3). All samples were immediately frozen in liquid nitrogen and stored at −80°C until analyzed. A piece of paraxial muscle was dissected and weighed for determination of muscle water content (MWC).
Plasma osmolality was analyzed on a micro vapor pressure osmometer (Wescor 5500, Logan, UT). Plasma sodium was determined by flame photometry (Instrumentation Laboratory 243, Italy). Plasma total calcium and total magnesium were analyzed by atomic absorption spectrophotometry (Perkin Elmer 2380, Mountain View, CA) in samples diluted with 0.1% La2O3. Plasma chloride was determined by coulometric titration (Radiometer CMT 10, Copenhagen). MWC was determined as percent weight loss after drying at 105°C for 3 days. Gill and kidney Na+,K+-ATPase activities were analyzed at 25°C by the method of McCormick (1993) using a plate reader (Ceres 900UV; Bio-Tek, Winooski, VT). Protein content was measured by the method of Lowry et al. (1951) modified for plate reader.
Statistics
Statistical differences among groups were analyzed using SYSTAT 5.03 (Systat, 1991, Evanston, IL). When necessary, transformation of data was done to meet the assumption of homogeneity of variances (Bartlett test). Differences were then analyzed by one-way analysis of variance (ANOVA) followed by Tukey's Honestly Significant Difference Test (P ≤ 0.05). Correlation analyses were performed on plasma osmolality versus muscle water content and osmolality versus gill Na+,K+-ATPase activity.
RESULTS
Experiment 1
Two saline- and one oPRL-injected fish died during hormone treatment in FW. After 48 hr in 27.3 ppt SW, plasma [Na+], [Cl−] (Fig. 1A, B) and osmolality (Fig. 2A) were affected similarly by hormone treatment. MWC (Fig. 2B) was inversely related to changes in plasma osmolality (linear correlation analysis: r = 0.837, P < 0.001). F treatment improved SW- acclimation (decreased plasma [Cl−]), and oPRL treatment impaired SW-acclimation (increased plasma [Na+], [Cl−] and osmolality and decreased MWC). When coinjected with F, oPRL abolished the SW-adaptive effect of F. This effect was significant for plasma [Na+], [Cl−] and osmolality at the high dose of oPRL and for MWC at the low and medium doses of oPRL. Plasma total [Ca] and [Mg] (Fig. 1C, D) were unaffected by hormone treatment.
Gill Na+,K+-ATPase activity (Fig. 2C) was stimulated by F but unaffected by oPRL treatment at any dose. Kidney Na+,K+- ATPase activity was significantly lower in the untreated control group than in the saline-injected control group (Fig. 2D) but otherwise unaffected by experimental treatment.
Experiment 2
A few fish died during the experiment but there was no clear relation to treatment. After 48 hr in 27.3 ppt SW, the saline-injected control fish had a poorer hypoosmoregulatory ability than the untreated control group as judged from a higher plasma osmolality and a lower MWC (Fig. 3A, B). As in Exp. 1, plasma osmolality and [Na+] (Fig. 3C) were affected similarly by hormone treatment and MWC was negatively correlated with plasma osmolality (linear correlation analysis: r = 0.956, P < 0.001). As judged from plasma osmolality, [Na+] and MWC, oGH and F treatment improved, and oPRL treatment impaired the overall hypoosmoregulatory performance. Ovine PRL antagonized the SW-adaptive effect of oGH+F treatment in a dose-related manner, with 1 and 2 μg oPRL/g being equipotent. Neither dose of oPRL was able to completely abolish the hypoosmoregulatory effect of oGH+F treatment.
Plasma total [Ca] (Fig. 3D) showed a different response to hormone treatment from plasma [Na+] after 48 hr in SW. Treatment with F, oGH or oPRL alone did not alter plasma total [Ca]. The combined treatment of F and oGH, however, significantly decreased plasma total [Ca]. Ovine GH and F treatment significantly increased gill Na+,K+-ATPase activity (Fig. 3E), and when combined they had an additive stimulatory effect. Ovine PRL had no effect on gill Na+,K+-ATPase when injected alone but counteracted the effect of oGH+F in a doserelated manner with maximal effect at 1 μg/g. In all the combined treatment groups, oPRL was unable to depress enzyme activity completely to the level of the saline-injected control.
In both experiments, plasma osmolality and gill Na+,K+- ATPase activity were nonlinear negatively correlated (Figs. 4A, B) and could be described by the following equations: Exp. 1: Osmolality = 356.3 + 553.7 * e−1.6*activity (r = 0.636, P < 0.001). Exp. 2: Osmolality = 322.3 + 853.2 * e−1.3*activity (r = 0.793, P < 0.001).
DISCUSSION
Several studies have shown that treatment with either F or GH improves SW-tolerance and hypoosmoregulation in salmonid teleosts. These effects were also evident in both of the present experiments (plasma [Na+], [Cl−] and osmolality and MWC; Figs. 1–3), and may be causally related to stimulated levels of gill Na+,K+-ATPase activity after F and GH treatment (Figs. 2C, 3E; Richman and Zaugg, 1987; Madsen, 1990a,c; Bisbal and Specker, 1991; Almendras et al., 1993; Boeuf et al., 1994; Nonnotte and Boeuf, 1995). This is supported by the nonlinear negative correlation between plasma osmolality and gill Na+,K+-ATPase activity observed in the present experiments (Fig. 4A, B). The correlation analyses further suggest that there may be a gill Na+,K+-ATPase threshold level (of about 2.5 μmol ADP/mg protein/hr) below which the brown trout cannot balance the influx of salt after SW-transfer with increased salt excretion. Treatment with oPRL impairs the ability of salmonids to hypoosmoregulate (Hasegawa et al., 1986; Figs. 1–3) although in some studies these effects were less significant (Madsen and Bern, 1992; Boeuf et al., 1994).
The use of ovine GH and PRL is justified by the fact that these hormones are known to bind specifically to their respective teleost receptors, although with lower affinity than the homologous hormones (oPRL: Auperin et al., 1994; Prunet and Auperin, 1994; oGH: Yao et al., 1991). Several studies have shown that the mammalian hormones induce qualitatively similar osmoregulatory effects as the teleost hormones. The receptor's lower affinity for the heterologous hormone is, however, reflected in the reduced potency compared to the native hormones (oPRL: Hasegawa et al., 1986; oGH: Bouef et al., 1994).
None of the present hormone treatments had a significant effect on kidney Na+,K+-ATPase activity when compared to the saline-injected control. This agrees with Madsen et al. (1995), who found no effect of F, rsGH, sPRL or rblGF-I on kidney Na+,K+-ATPase activity and its α-subunit mRNA expression in S. trutta. Similarly, kidney Na+,K+-ATPase activity and mRNA levels were unaffected during SW-acclimation (O. mykiss: Jürss et al., 1985; S. salar. McCormick et al., 1989; S. trutta: Madsen et al., 1995). Kidney Na+,K+-ATPase decreases during SW-acclimation in other teleosts (F. heteroclitus: Epstein et al., 1969; Chelon labrosus: Gallis et al., 1979) and the activity is stimulated by PRL and F in F. heteroclitus (Pickford et al., 1970a,b) The adaptive role and the endocrine control of changes in kidney Na+,K+-ATPase activity in salmonids remain uncertain. Plasma total [Ca] (Figs. 1C and 3D) and [Mg] (Fig. 1D) showed little response to the experimental treatment. F, oGH or oPRL treatment had no effect on plasma total [Ca] after 48 hr in SW. This is in agreement with Madsen and Bern (1992) who found no changes in plasma total [Ca] in either oGH- or oPRL-injected S. trutta transferred to 25 ppt SW. In contrast, implants of oGH decreased plasma total [Ca] after transfer of S. trutta to 35 ppt SW (Almendras et al., 1993). Bolton et al. (1987) found a hypocalcemic effect of oGH after transfer of O. mykiss to 80% SW, whereas PRL had no effect on plasma total [Ca]. In the present Exp. 2, combined F and oGH had a hypocalcemic effect, which is at variance with Madsen and Korsgaard (1991) who found no effect of coinjecting F and oGH in S. trutta on plasma total [Ca] after 2 days in 28.5 ppt SW. Calcium regulation is a complex process in fishes that may be controlled at various levels and by various hormones. GH, PRL and F have been reported as calcitropic hormones in O. mykiss and Oreochromis mossambicus (Flik and Perry, 1989; Flik et al., 1993, 1994) and may affect either active or passive steps of transbranchial transport. The variable responses may therefore be due to differences in experimental protocols such as salinity, hardness of water, nutritive status and stress. The mode of action of PRL's anti-SW-adaptive effect is unknown. Madsen and Bern (1992) proposed that one possible mechanism may occur via inhibition of gill Na+,K+-ATPase activity. In the present study, oPRL treatment did not inhibit basal levels of gill Na+,K+- ATPase activity (Figs. 2C and 3E), yet PRL had clear hyperosmotic effects in both experiments (Figs. 1–3). One reason for the lack of effect of oPRL when injected alone may be that the gill Na+,K+-ATPase activity is at near-basal level during the fall in FW-acclimated brown trout, and thus may not be further decreased by hormonal manipulation. Madsen et al. (1995) also found a similar lack of effect of sPRL on both gill enzyme activity and mRNA expression in FW brown trout. Conversely, when gill Na+,K+-ATPase activity was increased by long-term SW acclimation in 0+ brown trout, a single injection of 0.2 μg oPRL/g significantly inhibited enzyme activity after 3 days (M. Seidelin, 1996, unpublished). Madsen and Bern (1992) found that GH-stimulated gill Na+,K+-ATPase activity could be depressed to control fish levels by simultaneous PRL treatment in O. mykiss, and a similar antagonism has been found in S. trutta (M. Seidelin, A. Heilskov and S.S. Madsen, 1996, unpublished). Thus it appears that gill Na+,K+-ATPase activity needs to be at a stimulated level in order for PRL to have an inhibitory action in vivo.
When S. trutta were treated with oPRL in combination with F, it antagonized the SW-adaptive effect of F on plasma osmolality, [Na+], [Cl−] and MWC (Figs. 1A, B and 2A, B). However, oPRL did not affect the stimulated gill Na+,K+- ATPase activity of F-treated fish (Fig. 2C). In this situation, an alternative route for the anti-SW adaptive effect of PRL must be sought. Other studies have shown that PRL may decrease branchial ion-permeability (Hirano, 1986), or impair in situ chloride cell function, as shown for opercular membrane chloride cells in O. mossambicus (Foskett et al., 1982; Herndon et al., 1991).
The lack of effect of oPRL on the F-stimulated level of gill Na+,K+-ATPase activity, led us to investigate the ability of oPRL to antagonize the combined effect of F and oGH on the development of hypoosmoregulatory ability (see Madsen 1990b,c). In Experiment 2, fish treated with either oGH, F or oGH+F in FW performed well in the 48 hr SW challenge test (Fig. 3). Plasma osmolality and MWC in these groups were comparable to the levels measured in S. trutta acclimated to FW (Madsen, 1990c) or to 25 ppt SW for 50 days (Madsen et al., 1995). Ovine GH and F have clear additive effects on gill Na+,K+-ATPase activity (Fig. 3E; Björnsson et al., 1987; Madsen, 1990b,c). In the present study we did not find additive effects on plasma osmolality, [Na+] or MWC (Fig. 3A, B, C), as reported by Madsen (1990b,c). One likely explanation is that the salinity used in the SW-challenge test may have been too low to reveal such effects. Recent experiments have pursued some possible mechanism for the interaction of F and GH. F treatment of salmonids generally reduces the cytosolic and increases the nuclear fraction of the gill corticosteroid receptor (CR) resulting in a general decrease in the total amount of CRs (e.g., Weisbart et al., 1987; Shrimpton and Randall, 1994). GH treatment increases CRs in the gill (Shrimpton et al., 1995) and may thereby increase the responsiveness of the gill to F. Exogenous GH has no direct effect alone or together with F on gill Na+,K+-ATPase activity (McCormick et al., 1991a), and there is evidence that the interaction between F and GH at the level of gill Na+,K+- ATPase may be mediated by IGF-I. GH increases the hepatic and branchial production of IGF-I (Sakamoto and Hirano, 1993) and thus increases IGF-I levels in plasma (Moriyama, 1995). IGF-I stimulates hypoosmoregulatory ability (O. mykiss: McCormick et al., 1991b) as well as gill Na+,K+-ATPase activity and mRNA levels in vivo (S. trutta: Madsen et al., 1995). Furthermore, IGF-I has been shown to increase gill Na+,K+- ATPase activity in vitro when the gill epithelium had been preexposed to endogenous or exogenous GH in vivo (Madsen and Bern, 1993). Recently, it has been shown that IGF-I may increase the responsiveness of gill Na+,K+-ATPase activity to F in S. salar (McCormick, 1996).
As shown in Fig. 3, oPRL did not abolish the hypoosmoregulatory ability (plasma osmolality, [Na+] and MWC) of the oGH+F-injected fish to the level of the saline-injected controls. Nor was gill Na+,K+-ATPase activity depressed below the level of the F-injected fish, which suggests that oPRL only counteracts the oGH-induced stimulation of gill Na+,K+-ATPase activity. This is supported in Exp. 1, where oPRL was unable to inhibit F-stimulated Na+,K+- ATPase activity. The molecular mechanism by which PRL antagonizes the effect of GH on gill Na+,K+-ATPase activity is not known. Competition between oPRL and oGH at the receptor level appears unlikely, since displacement studies show that the teleost GH-receptor has less than 1% of the affinity for oPRL that it has for GH (Gray et al., 1990; Sakamoto and Hirano, 1991; Yao et al., 1991). On the other hand, based on the similarity of their receptors, GH and PRL may interact at the postreceptor signalling level (review by Roupas and Herington, 1994). PRL may also alter the amount of IGF-I binding proteins (as has been shown in vitro in Morone saxatilis by Fukazawa et al., 1995), thus possibly modifying the half-life of IGF-I in the plasma as well as changing the target orientation of IGF-I (review by Siharath and Bern, 1993).
Our finding that PRL is unable to inhibit the F-stimulated gill Na+,K+-ATPase activity may have relevance during the development of euryhaline capacity at smoltification. Typically, the plasma PRL-level decreases concomitantly with an increase in plasma F and GH just prior to development of maximum hypoosmoregulatory ability (e.g., Björnsson et al., 1989; Young et al., 1989). The decrease in plasma PRL may release the constraints on GH action at the level of the gill, which may be a prerequisite for increasing gill Na+,K+-ATPase activity to its maximum level. GH and F may have to act in concert to achieve this goal (Björnsson et al., 1987; Madsen, 1990b,c; Experiment 2, this study).
In conclusion, in S. trutta oPRL inhibits the hypoosmoregulatory action of exogenously administered F. Ovine PRL does not affect the increase in gill Na+,K+-ATPase activity brought about by F, and the gill Na+,K+-ATPase enzyme may therefore not be the target for the antagonistic interaction of PRL and F in salmonid hypoosmoregulation.
Acknowledgments
This study was supported by a grant from the Danish Research Council (11-9659-1). The National Hormone and Pituitary Program, the National Institute of Diabetes and Digestive and Kidney Diseases, the National Institute of Child Health and Human Development and the U.S. Department of Agriculture are acknowledged for the gift of pituitary hormones used in this study.