Insulin regulates the expression of several metabolism-related genes in the liver and primary hepatocytes of rainbow trout (Oncorhynchus mykiss)

Carnivorous fish such as rainbow trout (Oncorhynchus mykissWalbaum) are characterized by a limited ability to use dietary carbohydrates efficiently (Wilson, 1994) and are thus considered to be `glucose intolerant'(Hemre et al., 2002; Moon, 2001). Oral or intravenous administration of glucose and a carbohydrate-rich diet result in persistent hyperglycemia in various fish species, including rainbow trout(Bergot, 1979; del sol Novoa et al., 2004; Legate et al., 2001; Palmer and Ryman, 1972; Parrizas et al., 1994b), and are also associated with an increase in insulin levels. The magnitude of the insulin response is very fish-species dependent and less than that in mammals(Furuichi and Yone, 1981). Insulin receptors are also present in major insulin-responsive tissues in fish, i.e. white muscle, liver and adipose tissues(Gutierrez et al., 1991; Navarro et al., 1999). Up-regulation of insulin binding and tyrosine kinase activity have been observed after insulin treatment and a carbohydrate-rich diet, respectively(Banos et al., 1998; Gutierrez et al., 1991; Parrizas et al., 1994a; Parrizas et al., 1994b). The persistent hyperglycemia in rainbow trout fed with carbohydrates despite the existence of insulin secretion and receptors suggests an insulin resistance state comparable to that observed in humans with type II diabetes.

Glucose uptake mediated by the facilitative insulin-regulated glucose transporter GLUT4 homologue in white muscle and fat tissues and by GLUT2 in the liver has been reported in trout(Capilla et al., 2002; Diaz et al., 2007a; Diaz et al., 2007b; Krasnov et al., 2001). Fish glucose transporters GLUT2 and GLUT4 are characterized by a lower affinity for glucose, which may at least partly explain the persistent post-prandial hyperglycemia observed in trout fed with carbohydrates(Capilla et al., 2004a; Krasnov et al., 2001). At the metabolic level, most of the key enzymes involved in carbohydrate metabolism have been described in fish (Cowey and Walton, 1989). For some of these, including glucokinase (liver),phosphofructokinase (liver and muscle) and pyruvate kinase (liver and muscle),their induction by dietary carbohydrates is similar to that described in mammalian systems (Fideu et al.,1983; Panserat et al.,2001a; Panserat et al.,2000b). However, findings concerning other metabolic mediators of glucose metabolism suggest the existence of atypical regulation after carbohydrate intake by trout, i.e. a lower capacity for glucose phosphorylation by hexokinases in the muscle of fish than in mammalian species, as confirmed by the poor role of exogenous glucose as a glycogenic substrate in trout muscle (Kam and Milligan, 2006; Kirchner et al., 2005; Wilson,1994). Moreover, in rainbow trout a carbohydrate-rich diet does not affect the activity or gene expression of key enzymes of gluconeogenesis such as glucose 6-phosphatase (G6Pase), fructose 1,6-bisphosphatase (FBPase)and phosphoenolpyruvate carboxykinase (PEPCK)(Panserat et al., 2001b; Panserat et al., 2000a; Panserat et al., 2001c; Tranulis et al., 1991). Such impaired post-prandial down regulation of gluconeogenesis in rainbow trout fed with carbohydrates is similar to that observed in human patients with type II diabetes. In healthy patients insulin lowers plasma glucose levels by inducing plasma glucose uptake into skeletal muscle and adipose tissues and by inhibiting endogenous hepatic production of glucose. In mammals, gluconeogenic genes are mainly under insulin control(Barthel and Schmoll, 2003b). Insulin inhibits expression of PEPCK and G6Pase enzymes at the transcriptional level (Barthel and Schmoll,2003a) through the activation of the protein kinase Akt (also known as protein kinase B) (Liao et al.,1998; Schmoll,2000), a critical node in the insulin signaling pathway(Taniguchi et al., 2006). Because of insulin resistance, patients with type II diabetes present fasting hyperglycemia that is correlated with an increased glucose production by the liver. This is illustrated by the higher level of expression of two key enzymes of hepatic gluconeogenesis, PEPCK and G6Pase. As observed in these diabetic patients, the absence of post-prandial inhibition of hepatic gluconeogenesis in rainbow trout probably contributes to persistent hyperglycemia. The reason for the persistent post-prandial endogenous production of glucose in rainbow trout therefore remains to be elucidated.

Thus, we put forward the hypothesis that the glucose intolerance observed in rainbow trout may be related to the impaired ability of insulin to regulate mRNA levels of hepatic target genes. In order to test this hypothesis, we performed intraperitoneal administration of bovine insulin (10 u kg^–1) to 48 h-fasted rainbow trout and analyzed the activation of Akt and the mRNA levels of genes involved in gluconeogenesis(PEPCK, G6Pase and FBPase). Since regulation of target gene expression by insulin may be more widely impaired in rainbow trout, the study was enlarged to include other hepatic genes identified as being under insulin control in mammals. We therefore also analyzed the insulin regulation of the mRNA expression of glucokinase (GK) and pyruvate kinase (PK), which catalyse the phosphorylation of glucose and the conversion of phosphoenolpyruvate to pyruvate, the first and the final steps in glycolysis, respectively(Towle et al., 1997). We also investigated the regulation of mRNAs encoding two proteins involved in fatty acid metabolism, i.e. fatty acid synthase (FAS), which promotes the synthesis of fatty acids, primarily palmitate, through the condensation of malonyl-CoA and acetyl-CoA (Smith et al.,2003), and carnitine palmitoyltransferase 1 (CPT1), which is a limiting enzyme of fatty acid β-oxidation responsible for the entry of long chain fatty acids into the mitochondria(Bartlett and Eaton,2004).

Despite the well-established and fully described potential and advantages of fish hepatocyte systems (Moon et al.,1985; Scholz et al.,1998; Segner,1998a), few studies have used cultured trout hepatocytes to analyze the regulation of intermediary metabolism(Alvarez et al., 2000). In order to rule out the possibility of an indirect effect of insulin and of possible interactions with other endocrine mediators on target gene expression in vivo, we used a primary cell culture of rainbow trout hepatocytes to analyze the transcriptional effects of insulin in this model.

In vivo and in vitro experiments were performed using sexually immature rainbow trout (250–300 g) obtained from the INRA experimental fish farm facilities of Donzacq, France. Fish were maintained in tanks kept in open circuits with 18°C well-aerated water. Fish were fed with a commercial diet (T-3P classic, Trouw, France). All experiments were carried out in accordance with legislation governing the ethical treatment of animals, and investigators were certified by the French Government to carry out animal experiments.

To study the regulation of hepatic gene expression by insulin in vivo, rainbow trout were placed in two separate tanks. Fish were kept unfed for 48 h before the day of the experiment in order to obtain fish with empty digestive tracts but without onset of body protein degradation. Fasting also allowed us to obtain basal plasma glucose levels and to limit individual variability. On the day of the experiment, fish received an intraperitoneal injection of bovine insulin (10 u kg^–1 body weight; Sigma, St Louis, MO, USA) as previously described(Salgado et al., 2004) or were sham treated with vehicle (saline). Four fish per tank were left untreated and unfed, and served as controls. At 2, 4 and 6 h after treatment, fish(N=8) were killed by cervical section. Blood was collected from the dorsal aorta using a syringe pre-treated with a solution containing 4% NaF and 4% K₂C₂O₄, and centrifuged, and plasma was kept frozen at –20°C until analysis. Liver samples were collected,snap frozen in liquid nitrogen then stored at –80°C prior to further mRNA and protein analyses.

Isolated liver cells were prepared from 4 day-fasted rainbow trout as previously described by Mommsen and colleagues(Mommsen et al., 1994). Fish were anesthetized by placing them in water containing 60 mg l^–1 aminobenzoic acid. After in situ perfusion using liver perfusion medium (1×; 17701-038, Invitrogen, Carlsbad, CA, USA),livers were excised, minced with a razor blade and immediately digested in a liver digest medium (17703-034, Invitrogen) for 20 min at 18°C. After filtration and centrifugation (120 g, 2 min), the resulting cell pellet was resuspended three successive times in modified Hanks' medium(136.9 mmol l^–1 NaCl, 5.4 mmol l^–1 KCl, 0.81 mmol l^–1 MgSO₄, 0.44 mmol l^–1KH₂PO₄, 0.33 mmol l^–1Na₂HPO₄, 5 mmol l^–1 NaHCO₃and 10 mmol l^–1 Hepes) supplemented with 1.5 mmol l^–1 CaCl₂ and 1.5% defatted bovine serum albumin(BSA; Sigma), and centrifuged (70 g, 2 min). Cells were finally taken up in modified Hanks' medium supplemented with 1.5 mmol l^–1 CaCl₂, 1% defatted BSA, 3 mmol l^–1 glucose, MEM essential amino acids (1×;Invitrogen), MEM non-essential amino acids (1×; Invitrogen) and antibiotic antimycotic solution (1×; Sigma). Cell viability (>98%)was assessed using the Trypan Blue exclusion method (0.04% in 0.15 mol l^–1 NaCl) and cells were counted using a hemocytometer. The hepatocyte cell suspension (CS) was plated in a six well Primaria culture dish(BD Biosciences, NJ, USA) at a density of 3×10⁶ cells per well and incubated at 18°C. The incubation medium was replaced every 24 h over the 72 h of primary cell culture. Microscopic examination ensured that hepatocytes progressively re-associated throughout culture to form two-dimensional aggregates, in agreement with earlier reports(Ferraris et al., 2002; Segner, 1998b). Cell viability and cytotoxicity were monitored throughout culture using CellTiter 96®aqueous one solution cell proliferation assay (Promega, Madison, WI, USA) and CytoTox 96® non-radioactive cytotoxicity assay (Promega) respectively,according to the manufacturer's recommendations.

For analysis of Akt phosphorylation, 48 h-cultured hepatocytes were stimulated using 4×10^–9 mol l^–1 bovine insulin (Sigma) for 5, 15, 30 and 60 min [4×10^–9 mol l^–1 insulin corresponds to the post-prandial level of insulin irrespective of dietary carbohydrate level(del sol Novoa et al., 2004)]. At the end of the stimulation period, cells were collected for protein extraction.

For gene expression analysis, 48 h-cultured cells were exposed for an additional 24 h to 4×10^–9 mol l^–1bovine insulin in the presence of 3 or 20 mmol l^–1 glucose,the mean glucose concentrations measured in fasted trout and trout fed the carbohydrate-rich diet, respectively(Bergot, 1979; del sol Novoa et al., 2004; Hemre et al., 2002; Panserat et al., 2001c). Osmotic pressures of insulin–glucose-supplemented cell culture media were verified before each treatment and calibrated to 300 mosmol l^–1 using a K7400 osmometer (Knauer, Berlin, Germany). At the end of the stimulation period, cells were collected for total RNA extraction and subsequent RT-PCR analysis.

Plasma glucose levels were measured with the Glucose RTU kit from BioMerieux (Marcy l'Etoile, France) according to the manufacturer's recommendations.

Total RNA samples were extracted from livers and hepatocytes using TRIzol reagent (Invitrogen) according to the manufacturer's recommendations and treated with DNAse to avoid any genomic DNA contamination. A 1 μg sample of the resulting total RNA was reverse transcribed into cDNA using the SuperScript III RnaseH^– reverse transcriptase kit(Invitrogen) and oligo dT primers (Promega) according to the manufacturer's instructions.

In order to confirm that the cultured cells might be considered as hepatocyte-like cells, expression of hepatocyte-specific genes such as albumin precursor (AP), tyrosine amino transferase (TAT), hepatic nuclear factor 4 (HNF4) and serotransferine precursor (STP)was examined by conventional RT-PCR. We also checked the mRNA expression of the glucose transporter (Glut2) and the insulin receptor(InsR) for further stimulation of these cells by glucose and/or insulin. Expression of these genes was studied in cell suspension as well as after 4, 24, 48 and 72 h of culture. For this purpose, cDNA samples were amplified by PCR using specific primers chosen from the Sigenae rainbow trout cDNA sequence public database(http://www.sigenae.org; Table 1). The PCR reaction was carried out using 2 μl of cDNA as template in a final reaction volume of 25μl containing 1.5 mmol l^–1 MgCl₂, 0.16μmol l^–1 of each primer, and 1 U of Taq polymerase(Promega). Samples were subjected to 35 amplification cycles of a standard PCR protocol (94°C for 20 s, specific primer hybridization temperature for 20 s, and 72°C for 20 s). The different PCR products were monitored by sequencing (GENOME Express, Meylan, France) to confirm the nature of the amplicon.

Gene expression analyses were performed on samples from the livers of control and treated fish 6 h after intraperitoneal administration as well as from hepatocytes after 24 h of stimulation by insulin and glucose. Target gene expression levels were determined by real-time quantitative RT-PCR (qRT-PCR)using specific real-time PCR primers (Table 2). To avoid amplification of genomic DNA, when possible primer pairs included one intron-spanning oligonucleotide. Real-time RT-PCR was carried out on an iCycler iQ™ real-time PCR detection system (Bio-Rad,Hercules, CA, USA) using iQ™ SYBR® Green Supermix. mRNA levels of GK, PK, G6Pase isoforms 1 and 2, PEPCK, FBPase, CPT1isoforms A and B, and FAS were evaluated. Elongation factor 1α(EF1α) was employed as a non-regulated reference gene as previously used in rainbow trout(Gabillard et al., 2003; Kamangar et al., 2006). The absence of a change in EF1α gene expression was also observed in our in vivo and in vitro studies (data not shown). PCR was performed using 10 μl of the diluted cDNA mixed with 5 pmol of each primer in a final volume of 25μl. The PCR protocol was initiated at 95°C for 3 min for initial denaturation of the cDNA and hot-start iTaq™ DNA polymerase activation, and continued with a two-step amplification program (20 s at 95°C followed by 30 s at specific primer hybridization temperature) repeated 40 times. Melting curves were systematically monitored (temperature gradient at 0.5°C/10 s from 55 to 94°C) at the end of the last amplification cycle to confirm the specificity of the amplification reaction. The different PCR products were initially checked by sequencing to confirm the nature of the amplicon. Each PCR run included replicate samples (duplicate of reverse transcription and duplicate of PCR amplification) and negative controls (reverse transcriptase-free samples, RNA-free samples).

Frozen liver (300 mg) and hepatocytes were homogenized on ice with an Ultraturrax homogenizer (IMLAB sarl, Lille, France) in a buffer containing 150 mmol l^–1 NaCl, 10 mmol l^–1 Tris, 1 mmol l^–1 EGTA, 1 mmol l^–1 EDTA (pH 7.4), 100 mmol l^–1 sodium fluoride, 4 mmol l^–1 sodium pyrophosphate, 2 mmol l^–1 sodium orthovanadate, 1% Triton X-100, 0.50% NP40-IGEPAL and a protease inhibitor cocktail (Roche, Basel,Switzerland). Homogenates were centrifuged at 1000 g for 30 min at 4°C and supernatants were then centrifuged for 45 min at 150 000 g. The resulting supernatants were aliquoted and stored at–80°C. Protein concentrations were determined using the Bio-Rad protein assay kit. Liver and cell lysates (20 μg of protein) were subjected to SDS-PAGE and Western blotting using anti-phospho-Akt Ser473 and anti-Akt antibodies (Cell Signaling Technology, Ozyme, St Quentin-en-Yvelines, France),consecutively. Rabbit phospho-Akt Ser473 and Akt antibodies were directed against synthetic peptides corresponding to residues surrounding phosphorylated Ser473 of mouse Akt and the carboxyterminal sequence of mouse Akt, respectively. These antibodies have been shown to successfully cross-react with rainbow trout Akt protein(Seiliez et al., 2008). After washing, protein detection was performed by chemi luminescence using LumiGLO reagents (Cell Signaling Technology) and horseradish peroxidase-conjugated anti-IgG as the secondary antibody (Cell Signaling Technology).

The results of in vivo mRNA expression analyses are expressed as means ± s.e.m. (N=8) and were analyzed by one-way ANOVA. Results from plasma glucose levels (N=8) and mRNA expression in primary hepatocyte cell culture (N=6) are expressed as means ±s.e.m. and were analyzed by two-way ANOVA. Means were compared by Student–Newman–Keuls multiple comparison test. The level of significance was set at P<0.05.

Data on plasma glucose levels measured 2, 4 and 6 h after intraperitoneal administration of insulin or saline solution are presented in Fig. 1. Administration of insulin gradually and significantly decreased plasma glucose levels to reach 1.78±0.18 mmol l^–1 6 h after intraperitoneal injection compared with saline-treated trout (4.49±0.19 mmol l^–1). A significant increase in plasma glucose levels was recorded 2 h after administration of saline solution but no further difference between control and vehicle-treated fish was observed 4 and 6 h after the intraperitoneal administration.

Since insulin signaling is mediated through Akt phosphorylation on threonine 308 and serine 473 in mammals, Western blot analysis was performed to study the phosphorylation of Akt on serine 473 compared with endogenous levels of total Akt in rainbow trout. As illustrated in Fig. 2A, insulin induced Akt phosphorylation whereas administration of saline solution did not. The phosphorylation of Akt was observed as early as 2 h after intraperitoneal administration of insulin and was maintained until at least 6 h after insulin administration.

As illustrated in Fig. 2B,insulin administration significantly decreased the expression of gluconeogenic genes, including both isoforms of G6Pase (2.1- and 2.6-fold decrease for G6Pase 1 and 2, respectively), PEPCK (2.0-fold)and FBPase (2.4-fold). Expression of both isoforms of CPT1also decreased about 3-fold under insulin treatment. Expression of FAS and PK mRNA was not affected by administration of insulin or saline solution. GK mRNA levels were also measured but were mostly undetectable in either saline or insulin-treated fish (data not shown).

Using end-point RT–PCR we monitored the expression of hepatocyte-specific markers (such as albumin precursor, serotransferin precursor and tyrosine aminotransferase), the hepatocyte nuclear factor 4(HNF4) required for hepatocyte differentiation, the hepatocyte-specific glucose transporter GLUT2, and the c form insulin receptor(Fig. 3). Fragments of the expected length were amplified for the albumin precursor, serotransferin precursor, tyrosine aminotransferase, HNF4 and GLUT2 genes in cell suspension and in cultured cells over 4 to 72 h. For the c form insulin receptor, the expected 413 bp amplification fragment was detected in cell suspension as well as during the full time course of incubation. However,we also detected the amplification of a larger fragment in cell cultures during the first 24 h. Sequencing of this larger fragment revealed that it corresponded to an alternatively spliced form (data not shown).

To analyze the effects of insulin on the phosphorylation of Akt in vitro, 48 h-incubated cells were treated with 4×10^–9 mol l^–1 insulin for 5, 15, 30 and 60 min. After Western blot analysis of phospho-Akt (serine 473), we observed that phosphorylation of Akt occurred as early as 5 min after insulin exposure, while the phosphorylated form of Akt was undetectable in untreated control cells (Fig. 4). Akt phosphorylation was detected until 60 min of incubation with insulin but the signal declined progressively.

As shown in Fig. 5, we observed that the mRNA level of GK was significantly enhanced by both insulin and glucose (average of 1.7- and 1.4-fold, respectively. PKand FAS gene expression were found to be regulated in the same way:insulin increased the expression of PK and FAS genes but the insulin stimulatory effect was enhanced in the presence of 20 mmol l^–1 glucose, as indicated by the significant statistical interaction between insulin and glucose. We found that insulin and glucose have opposite effects on the regulation of expression of both isoforms of G6Pase. While glucose significantly increased the expression of G6Pase, G6Pase mRNA levels declined in the presence of insulin. In the absence of insulin, the effect of glucose on G6Pase mRNA levels was more pronounced on the first isoform of this gene (3.7-fold increase) than on the second isoform (1.1-fold increase). The level of expression of the FBPase gene was not affected by the glucose concentrations tested in this experiment. We recorded a significant increase (1.6-fold) in FBPase mRNA level only when cells were incubated with insulin in the presence of 20 mmol l^–1 glucose. On investigating the regulation of expression of PEPCK (mitochondrial isoform), we found that the PEPCK mRNA level was slightly suppressed by glucose in the absence of insulin but strongly reduced by insulin (5.5- and 2-fold,respectively) when cells were incubated with 3 and 20 mmol l^–1 glucose. Expression of CPT1 (isoforms A and B)was almost undetectable in hepatocyte cultures.

Unlike in mammals, expression of the gluconeogenic genes is poorly regulated by carbohydrate intake in rainbow trout(Panserat et al., 2001b; Panserat et al., 2000b),suggesting a potentially impaired ability of insulin to regulate the expression of gluconeogenic genes and, more widely, hepatic target genes in teleost carnivorous fish. However, combining both in vivo and in vitro experiments, we have clearly demonstrated here that insulin is able to regulate the expression of hepatic target genes in rainbow trout.

The in vivo study based on intraperitoneal administration of insulin confirmed the ability of exogenous and heterologous insulin to induce hypoglycemia in rainbow trout, as already shown(Albalat et al., 2006; Salgado et al., 2004). In vivo insulin administration led not only to the reduction of plasma glucose levels but also to changes in other metabolic and hormonal signals(fatty acids, glucagons, etc.) potentially involved in the regulation of hepatic gene expression. In mammals, glycolytic and lipogenic genes are reciprocally regulated by glucose and fatty acids in liver through molecular mechanisms involving the recently discovered glucose signaling transcription factor carbohydrate-responsive element binding protein (ChREBP) (for a review,see Dentin, 2006). Glucagon increases glucose output by dephosphorylating the CREB (cAMP-responsive element binding protein) coactivator TORC2 (transducer of regulated CREB activity), which is transported to the nucleus where it enhances CREB-dependent transcription of gluconeogenic genes such as PEPCK and G6Pase (Alan Cheng,2006). We therefore used primary cell culture of rainbow trout hepatocytes to eliminate the relative contributions of other confounding factors. The study with primary cultured hepatocytes was focused on the contribution of glucose and insulin in the regulation of expression of insulin target genes. Displaying both the expression of genes representative of liver gene expression and the expression of the first mediators of glucose and insulin signaling, this primary hepatocyte culture thus provides a good model for this purpose.

Insulin regulates the transcription of target genes by controlling Akt activation of transcription factors such as Forkhead transcription factors of the FoxO subfamily and transcription factors belonging to the family of sterol regulatory element binding proteins (SREBPs)(Foufelle and Ferre, 2002). Our findings clearly show that insulin was also able to induce Akt phosphorylation in the liver and primary hepatocyte cell culture from rainbow trout, confirming activation of the insulin signaling pathway after insulin treatment both in vivo and in vitro. These results are in agreement with the recent demonstration that insulin and insulin-like growth factors (IGFs) are able to activate the phosphorylation of Akt in zebra fish embryonic cells (Pozios et al.,2001) as well as in primary cell culture of muscle cells from two different fish species, i.e. rainbow trout and gilthead sea bream (Sparus aurata) (Castillo et al.,2006; Montserrat et al.,2007).

We also demonstrated that activation of the insulin signaling pathway was associated with the regulation of hepatic gene expression both in vivo and in vitro. For instance, G6Pase, PEPCK and FBPase, three key enzymes of endogenous glucose production(Granner and Pilkis, 1990),presented a significantly lowered gene expression level after insulin treatment. This inhibition was observed not only in vivo following intraperitoneal administration of insulin but also in vitro, at least for G6Pase and PEPCK. This suggests a direct effect of insulin on the transcription of these genes that probably involves the activation of FoxO1 and PGC1α transcription factors, as has been described in mammals(Puigserver et al., 2003). However, the effects of insulin on the regulation of gluconeogenic gene expression were studied in animals that presented plasma glucose levels(1.78±0.18 mmol l^–1) far below those normally measured in fasted animals (4.10±0.40 mmol l^–1) or even in rainbow trout fed with 28% carbohydrates (12.2 mmol l^–1)(del sol Novoa et al., 2004). In this context, we also analyzed the effects of insulin on target gene expression in primary hepatocytes in relation to glucose concentrations corresponding to plasma glucose levels of fasted and carbohydrate-fed trout,respectively. Our findings demonstrate that G6Pase and PEPCKgene expression are subject to regulation by glucose but in opposite ways. Glucose stimulated G6Pase gene expression in rainbow trout hepatocytes independently of insulin, as in diabetic rats in vivo(Massillon et al., 1996) and in vitro (Argaud et al.,1997). Although paradoxical, this may contribute to the maintenance of an appropriate level of G6P, a key regulator of glycogen synthesis and glycolysis, in the face of increased GK activity(Aiston et al., 1999). Glucose also inhibited PEPCK gene expression in rainbow trout hepatocytes independently of insulin, but the down-regulating effect of insulin was above that of glucose. The significant inhibition of PEPCK gene expression by glucose may be related to the stimulatory effect of glucose on GKgene expression. The inhibition of PEPCK gene expression by glucose in mammals is mediated through the activation of GK(Scott et al., 1998). However,instead of confirming the inhibition of FBPase gene expression by insulin in vitro we observed an up regulation of FBPase mRNA expression under insulin stimulation at high glucose concentrations, whereas insulin clearly decreased the expression of the endogenous FBPasegene in diabetic rats and rat primary hepatocyte cell culture(El-Maghrabi et al., 1991; El-Maghrabi et al., 1988). Further studies are needed to investigate the potential role of glucose and insulin in FBPase gene expression in the rainbow trout liver. Except for FBPase, our in vivo and in vitro findings suggest that, unlike what happens in the post-prandial state, insulin is clearly able to down regulate the expression of the gluconeogenic genes in rainbow trout as previously demonstrated in mammals.

Control of blood glucose levels is closely related to the efficiency of the glycolytic pathway and this, like gluconeogenesis, is under insulin control. Upon entering the hepatocyte, glucose is phosphorylated to G6P by glucokinase(Granner and Pilkis, 1990)before being stored as glycogen, transformed into fatty acids or catabolized. Insulin is a major regulator of GK gene expression in the mammalian liver (Iynedjian et al.,1989). In our study GK mRNA was undetectable in fasted rainbow trout, in accordance with previous findings in the same species(Panserat et al., 2001c) and in gilthead sea bream (Caseras et al.,2000). However, intraperitoneal administration of insulin failed to induce GK mRNA expression. In contrast, both glucose and insulin were able to increase GK gene expression in hepatocytes in vitro and their effects were additive. This represents the first demonstration that in rainbow trout GK gene expression may be directly regulated by glucose, as to our knowledge this does not occur in mammals. The potential control of GK gene expression by glucose in rainbow trout may provide an explanation for the absence of GK regulation in vivo following insulin administration to fasted rainbow trout. Indeed, plasma glucose levels were so drastically reduced 6 h after intraperitoneal administration of insulin (1.78±0.18 mmol l^–1) that the expected effect of insulin was counteracted by the inhibitory effect of the very low glucose level. These results are in agreement with the previous demonstration that GK gene expression was only enhanced when rainbow trout were fed with a single meal of glucose or a carbohydrate-rich diet (Panserat et al.,2000b; Panserat et al.,2001c) whereas insulin secretion was similar in trout fed with high or low carbohydrate diets (Capilla et al., 2004b; del sol Novoa et al., 2004). The regulation of the second glycolytic enzyme that we studied – PK, the last enzyme of glycolysis(Towle et al., 1997) –exhibited a different response under insulin stimulation in in vivoand in vitro studies. While insulin had no effect on PK gene expression in vivo, it enhanced PK mRNA accumulation in the primary hepatocytes. In vitro, it was noted that glucose was essential to stimulate the positive effect of insulin on PK gene expression. Since fish were hypoglycemic 6h after insulin administration,insulin was not able to stimulate PK gene expression. In fact, in mammals PK gene transcription is induced by insulin-stimulated glucose metabolism (Alam and Saggerson,1998; Towle,2005). For example, PK gene expression is stimulated by glucose in cultured hepatocytes expressing GK, independently of insulin (Alam and Saggerson,1998; Doiron et al.,1994). Finally, we demonstrated that insulin is involved in the control of gene expression of at least the first and last enzymes of glycolysis. However, in contrast to the gluconeogenic genes, it seems that glucose represents an important co-factor in the regulation of expression of GK and PK genes by insulin in the rainbow trout.

The liver also plays a central role in fatty acid metabolism, and interactions between glucose and lipid metabolism are well characterized(Weickert and Pfeiffer, 2006). Lipogenesis in fish mainly occurs in the liver, which is also a site of lipid catabolism – fatty-acid β-oxidation(Gutieres et al., 2003). Both mechanisms are known to be under insulin control in mammals. We focused this study on FAS and CPT1, which are involved in lipid synthesis and degradation,respectively, and have been identified as insulin targets in mammals. Insulin regulation of FAS transcription in mammals is mediated by the PI3-kinase/Akt signaling pathway (Sul et al., 2000). In the present study, insulin regulation of rainbow trout FAS gene expression was in every way similar to that observed for the PK gene. As reported in mammals(Doiron et al., 1994; Sul and Wang, 1998), our results showed that glucose was essential for insulin stimulation of FAS gene expression. On the other hand, we demonstrated that CPT1 gene expression was down regulated in the rainbow trout liver. This down regulation was only shown in vivo following administration of insulin since expression of the two isoforms of CPT1 was undetectable in rainbow trout hepatocytes. Chatelain and colleagues demonstrated that long chain fatty acids markedly increased CPT1 mRNA level in primary culture of fetal rat hepatocytes(Chatelain et al., 1996). This accumulation resulted from two mechanisms: stimulation of gene transcription and stabilization of the mRNA. Thus the undetectable expression of CPT1 might be related to the total absence of fatty acids in our culture medium. However, our in vivo results were consistent with a previous study showing that insulin decreased the CPT1 mRNA level in rat H4IIE hepatoma cells (Park et al.,1995). Finally, the regulation of FAS and CPT1gene expression by insulin and glucose in fish was on the whole consistent with what happens in mammals.

Based on earlier findings showing the absence of post-prandial inhibition of gluconeogenic gene expression (Panserat et al., 2001b; Panserat et al., 2000b), and since insulin represents the main regulator of the expression of these genes, we investigated whether the ability of insulin to control gluconeogenic gene expression was impaired in fish. As we thought that the transcriptional effect of insulin may be affected in its entirety, we enlarged the study to other hepatic insulin target genes. By combining in vivo and in vitro approaches, we clearly demonstrated that in the rainbow trout insulin possesses the intrinsic ability to activate its signaling pathway and regulate expression of hepatic target genes, including gluconeogenic genes. The absence of post-prandial regulation of hepatic gluconeogenic gene expression cannot therefore be directly attributed to a fault in insulin signaling per se. However, the mechanisms involved in the regulation of metabolism often depend on the cross-talk between nutritional and hormonal signals. This study revealed the importance that other factors, including nutrients such as glucose, may have in the insulin regulation of target gene expression. In mammals, excessive dietary proteins and amino acids have detrimental effects on glucose homeostasis by promoting insulin resistance and increasing gluconeogenesis(Tremblay et al., 2005). One particular feature of the rainbow trout diet is that total protein content may exceed 45% of the dry matter. This consistently high dietary amino acid intake can thus have undesirable effects on insulin sensitivity, particularly on insulin-regulated gene expression. This could explain the absence of post-prandial down regulation of the expression of insulin target genes such as G6Pase and PEPCK and the restoration of their inhibition by reducing dietary protein levels(Kirchner et al., 2003). Further experiments are needed to examine this hypothesis and to analyze the effects of excessive amino acids on the regulation of the hepatic target gene by insulin.

We thank F. Terrier, F. Sandres and Y. Hontang for their assistance at the experimental fish farm, and V. Gentil, M. Saint Laurent and C. Delcambre for technical assistance. This study was supported by the European Commission`Perspectives of Plant Protein use in Aquaculture' (PEPPA) FAIR-no. QLK5-2000-30068, `Sustainable Aquafeeds to Maximise the Health Benefits of Farmed Fish for Consumers' (AQUAMAX no. 016249-2) and the Aquitaine region(no. 20061203001 and CCRRDT-2002-0308002C). The European Commission provided a fellowship to M.L. (AQUAMAX no. 016249-2).