FGF21 Mimics a Fasting-Induced Metabolic State and Increases Appetite in Zebrafish

Fibroblast growth factor 21 (FGF21) is a member of the FGF superfamily that acts in an endocrine manner. FGF21 is a key regulator of energy balance and metabolism in mammals, and has emerged as a therapeutic potential for treating obesity and diabetes. Here, we report that mRNAs encoding FGF21 and its receptors are widely distributed within the zebrafish tissues and are importantly modulated by fasting (decreased in brain and liver, and increased in gut). FGF21 stimulates food intake in zebrafish, likely in part by modulating brain npy/agrp and nucb2/nesfatin-1 and gut ghrelin and cck mRNA expression. In accordance with this orexigenic role, the expression of FGF21 and its receptors were observed to increase preprandially and decrease post-feeding in the foregut and/or liver. Finally, we found important evidence in favor of a role for FGF21 in regulating glucose and lipid metabolism in the zebrafish liver in a way that mimics a fasting metabolic state.

FGF21 increases food intake in zebrafish. Intraperitoneal (IP) administration of human recombinant FGF21 resulted in a significant increase in food intake at 2 h (all doses tested), 6 h (only 100 ng/g bw dose) and 24 h (only 1 ng/g bw dose) post-injection when compared to the control groups. Magnitude of feeding increase was about 60-80% in all cases. No significant differences in food intake were observed between saline and FGF21-injected fish at 1 h post-injection (Fig. 2a). Injection of all doses of FGF21 also caused a significant upregulation of neuropeptide y (npy) and agouti-related protein (agrp) mRNAs, and a significant reduction in nucleobindin 2a (nucb2a) and nucleobindin 2b (nucb2b) mRNAs in the zebrafish brain (Fig. 2b,c,f and g). Brain levels of cocaine-and amphetamine-regulated transcript (cart) mRNAs were slightly inducted (0.8-fold) by the peptide, but only at the highest dose (Fig. 2e). In peripheral tissues, IP injection of FGF21 downregulated the expression of cholecystokinin (cck) mRNAs in the foregut (Fig. 2i), and caused a significant increase in mRNAs encoding ghrelin in the foregut (Fig. 2h), and leptin a and leptin b in the liver (Fig. 2j,k).

Figure 1.
Distribution of mRNAs encoding FGF21 and its receptors in tissues involved in feeding and metabolism in zebrafish. Quantitative analysis of mRNA expression was performed by RT-qPCR considering β-actin as reference gene. Data are expressed as mean + SEM (n = 6), relative to the tissue with the lowest mRNA expression. fgf21, fibroblast growth factor 21; fgfr, fibroblast growth factor receptor.
The FGF21 system displays periprandial variations in expression in the zebrafish gut and liver. Figure 4 shows the periprandial mRNA expression profiles of fgf21 and its receptors in the zebrafish brain, foregut and liver. Major findings described a significant preprandial increase in the expression of fgf21, fgfr1a, fgfr2b and fgfr4 in the foregut, as demonstrated by significant higher levels of mRNAs at scheduled feeding time (0 h) compared to levels at −1 h and −3 h (Fig. 4g-i,l). Expression of fgf21 and fgfr2b decreased after feeding (+1 h and +3 h), although a similar drop was observed in unfed fish (Fig. 4g,i). Fgfr1a mRNA levels remained unaltered after feeding in the foregut of fish that received food; however, a significant drop was detected at +3 h compared to 0 h in those fish that skipped the scheduled feeding (Fig. 4h). No significant postprandial variations were observed in fgfr4 expression (Fig. 4l). As for fgfr2c and fgfr3, mRNA levels did not change preprandially in the foregut, and only a postprandial reduction in expression was observed for fgfr2c in both fed and unfed fish and for fgfr3 in unfed fish (Fig. 4j,k).
In the liver, almost no preprandial variations in the levels of mRNAs encoding FGF21 or any of its receptors were detected. Nevertheless, we observed that mRNA expression of almost all of the genes studied decreased in the liver after a meal in fish that were fed at their scheduled feeding time, while remained high in food deprived fish ( Fig. 4m-r). The only exception to this profile was observed for fgfr3 mRNAs, whose levels decreased at +3 h both in fed and unfed fish (Fig. 4q). In the brain, no major periprandial changes were observed for any of the genes studied. Only a slight increase in fgf21, fgfr2b and fgfr4 mRNAs was detected at +3 h in fed fish ( Fig. 4a-f).  Results correspond to the mean + SEM of the results obtained in three different experiments (n = 3 in each experiment). Asterisks denote significant differences between control and treated groups assessed by t-test (*p < 0.05, **p < 0.01). (b-k) Expression of mRNAs encoding key appetite-regulating peptides in the zebrafish brain (b-g), gut (h-i) and liver (j-k) 2 h after intraperitoneal administration of saline alone (control) or containing 1, 10 or 100 ng/g bw of FGF21. mRNA expression was quantified by RT-qPCR considering β-actin as reference gene. Data are expressed as mean + SEM (n = 6). Asterisks denote significant differences between control and treated groups (*p < 0.05, **p < 0.01, ***p < 0.001). agrp, agouti-related protein; cart, cocaine-and amphetamine-regulated transcript; cck, cholecystokinin; FGF21, fibroblast growth factor 21; grl, ghrelin; npy, neuropeptide Y; nucb2, nucleobindin 2; pomc, proopiomelanocortin.
(Glut2, glucose transporter 2, and Sglt1, sodium-glucose cotransporter 1) in the zebrafish liver are shown in Fig. 5. IP administration of FGF21 did not modulate the mRNA levels of glut2 at 2 h post-injection (Fig. 5d). However, it caused a significant reduction in the hepatic sglt1 mRNA expression at doses of 10 and 100 ng/g bw (Fig. 5g). FGF21-induced changes in the expression of Glut2 and Sglt1 were also studied in vitro using ZFL cells. Prior to this, the location of the two transporters within such cells was described using immunocytochemistry. Both Glut2 and Sglt1 were abundant in the ZFL cells and located within the cytoplasm of the cells (Fig. 5a,b). Levels of glut2 and sglt1 mRNAs were upregulated in ZFL cells exposed to different concentrations of FGF21 (0.1, 1 and 10 nM) during 1 and 6 h (Fig. 5e,h). Likewise, the exposure of cells to 1 nM and 10 nM FGF21 for 1 h increased Glut2 and Sglt1 protein levels, respectively (Fig. 5f,i).
Expression of genes involved in glucose metabolism was also modulated by treatment with FGF21, as shown in Fig. 6. When administered intraperitoneally, all doses of FGF21 tested (1, 10 and 100 ng/g bw) caused an induction in the levels of glucokinase (gck), glucose 6-phosphatase b (g6pcb) and glycogen phosphorylase mRNAs, while 10 and 100 ng/g bw FGF21 (but not 1 ng/g) upregulated the mRNA expression of phosphoenolpyruvate carboxykinase 2 (pck2) and fructose 1,6-bisphosphatase 1a (fbp1a) at 2 h post-injection (Fig. 6a,k,q,m,s). We also observed an increase in phosphofructokinase b (pfklb) mRNAs upon administration of 100 ng/g bw FGF21 (Fig. 6e). No changes in the levels of phosphofructokinase a (pfkla), pyruvate kinase (pklr), phosphoenolpyruvate carboxykinase 1 (pck1) and fructose 1,6-bisphosphatase 1b (fbp1b) mRNAs were observed following IP treatment with FGF21 ( Fig. 6c,g,i,o). In vitro exposure of ZFL cells to FGF21 during 1 and 6 h resulted in a concentration-dependent induction in the expression of gck, pfkla, pfklb, pklr, pck1, pck2, fbp1a, fbp1b and g6pcb (Fig. 6b,d,f,h,j,l,n,p,r) mRNAs. Glycogen phosphorylase mRNAs were also upregulated by the exposure to 10 nM FGF21 during 1 h, but was found downregulated at 6 h (Fig. 6t). Figure 7 shows the FGF21-induced changes in the expression of genes involved in lipid metabolism in zebrafish liver. IP administration of 1, 10 and 100 ng/g bw FGF21 was observed to significantly upregulate the expression of ATP citrate lyase (acly), carnitine palmitoyltransferase 1a (cpt1a), 3-hydroxy-3-methylglutaryl-CoA lyase (hmgcl) and acetyl-CoA acetyltransferase 1 (acat1) mRNAs at 2 h post-injection ( Fig. 7a,g,o,q). Levels of acetyl-CoA carboxylase (acaca) and 3-hydroxyacyl CoA dehydrogenase (hadh) mRNAs were also increased by FGF21, but at doses of 10 and 100 ng/g (not 1 ng/g) (Fig. 7c,m). Injection of FGF21 did not alter the expression of fatty acid synthase (fasn), acyl-CoA dehydrogenase medium-chain (acadm) and enoyl-CoA hydratase short-chain 1 (echs1) at the time tested (Fig. 7e,i,k). In vitro, exposure of ZFL cells to FGF21 dose-dependently induced levels of acaca, fasn, cpt1a, hmgcl and acat1 mRNAs at 1 and 6 h, while reduced the expression of hadh at 1 h (Fig. 7d,f,h,p,r). Acly mRNA expression in ZFL cells were significantly reduced by treatment with FGF21 at 1 h, but increased at 6 h (Fig. 7b). No significant FGF21-induced changes in mRNA expression were observed for acadm and echs1 in vitro (Fig. 7j,l). . Effects of 7-day fasting on the mRNA expression of fgf21 and its receptors in the zebrafish brain (a-f), foregut (g-l) and liver (m-r). Data obtained by RT-qPCR are expressed as mean + SEM (n = 6). Asterisks denote significant differences between control and treated groups (*p < 0.05, **p < 0.01, ***p < 0.001). fgf21, fibroblast growth factor 21; fgfr, fibroblast growth factor receptor. Periprandial changes in the mRNA expression of fgf21 and its receptors in the zebrafish brain (a-f), foregut (g-l) and liver (m-r). Samples were collected before scheduled feeding time (−3 h and −1 h), at feeding time (0 h) and after scheduled feeding time (+1 h and +3 h) in both fed and unfed fish. Data are expressed as mean ± SEM (n = 6) relative to the lowest average expression. Arrows denote feeding time. Different letters indicate significant differences (p < 0.05) among the different time points in fed (black dots) or unfed (white dots) groups, while asterisks indicate significant differences (*p < 0.05, **p < 0.01, ***p < 0.001) between groups at the same time point. fgf21, fibroblast growth factor 21; fgfr, fibroblast growth factor receptor. FGF21 affects the expression of genes regulating mitochondrial activity and metabolic transcription factors. Expression of ATP synthase F1 subunit beta (atp5f1b) mRNAs was observed to be upregulated by FGF21 both in vivo at 2 h post-IP injection, and in vitro 1 and 6 h after exposure of ZFL cells to the peptide (Fig. 7s,t). Expression of the transcription factor pparα and coactivator ppargc1α was unaltered by IP injection of FGF21 (Fig. 7u,w). However, in vitro exposure of ZFL cells to FGF21 resulted in a significant induction in ppargc1α and pparα mRNAs at 1 h, while a significant reduction was observed at 6 h ( Fig. 7v,x).

Discussion
This research considered the nutritional regulation of the FGF21 system in zebrafish, and determined the putative role of this peptide in glucose and lipid metabolism for the first time in a non-mammal. First, we reported that fgf21 mRNAs are present in zebrafish tissues involved in feeding and metabolism (brain, gut and liver), with considerably higher levels in the liver. Such abundant expression of FGF21 in the liver agrees with reports on tissue distribution in mammals 2,28 , but is not in agreement with the study by Wang and co-workers showing that fgf21 mRNAs in the Asian seabass are exclusively expressed in the kidney and intestine 27 . As it will also be pointed out later, it seems that the Asian sea bass gene encoding FGF21 might have evolved differently compared to mammals (and potentially other fish species, including the zebrafish) and might exert different physiological functions in this species. We also observed a tissue-specific presence of mRNAs encoding FGF21 receptors in the zebrafish brain, gut and liver. Such differential expression of FGF21 receptor subtypes points to tissue specificity in mediating the different physiological actions that FGF21 might be exerting in zebrafish.
The presence of mRNAs encoding FGF21 and some of its receptors in the zebrafish brain and gut, key tissues involved in appetite regulation, suggests the involvement of FGF21 in feed intake control. Indeed, our results showed that human recombinant FGF21 increases short-term food intake in the zebrafish in a time-and dose-dependent manner when administered intraperitoneally. We used human peptide because, given the large length of FGF21, the synthesis of the corresponding fish peptide was prohibitive due to excessively high costs. While this is a limitation of the study, human recombinant FGF21 has been previously used to study feeding behavior in teleosts 27 . In addition, zebrafish FGF21 sequence maintains a conserved cysteine residue at position 122 of human FGF21 protein. Both zebrafish and human FGF21 sequences share several residues at the C-and N-terminus (Supplementary Figure 2), which in mammals have been demonstrated to be critical for the interaction with the co-receptor b-Klotho and FGF receptors, respectively 29,30 . Additional studies are required to elucidate the structural characteristics of FGF21 and its receptor interactions, as well as mechanism of action in zebrafish. The observation of FGF21 being an orexigen in zebrafish is concordant with previous observations in rats intracerebroventricularly injected with FGF21 20 , although other studies reported no difference in food intake upon infusion of FGF21 in rats or in FGF21-KO mice 23 . However, it is opposite to the anorexigenic role reported for FGF21 in the Asian sea bass 27 , reinforcing the hypothesis of the FGF21 gene evolving independently in this species. The orexigenic role of FGF21 in zebrafish is likely mediated by the upregulation of orexigens NPY and AgRP in the brain and ghrelin in the foregut, and the downregulation of the anorexigens NUCB2/nesfatin-1 in the brain and CCK in the foregut. In fact, this is supported by our qPCR results, but further studies are needed to confirm this. Recinella and coworkers 20 described that NPY and AgRP gene expression is increased, while expression of POMC and CART is decreased, in response to FGF21 in rats. Furthermore, FGF21-KO mice had higher Figure 8. Effects of FGF21 on the activity of key enzymes involved in glucose and lipid metabolism in ZFL cells. Cells were incubated with culture media alone (control) or containing 1 or 10 nM FGF21 during 1 h. Data are shown as mean + SEM (n = 8). Asterisks denote significant differences between control and treated groups (*p < 0.05, **p < 0.01, ***p < 0.001). Acly, ATP citrate lyase; Cpt1a, carnitine palmitoyltransferase 1a; Fas, fatty acid synthase; FGF21, fibroblast growth factor 21; Gck, glucokinase; Hoad, 3-hydroxyacyl CoA dehydrogenase; Pepck, phosphoenolpyruvate carboxykinase; Pk, pyruvate kinase.

Scientific RepoRtS |
(2020) 10:6993 | https://doi.org/10.1038/s41598-020-63726-w www.nature.com/scientificreports www.nature.com/scientificreports/ Figure 9. Schematic representation of the metabolic pathways proposed to be modulated by FGF21 in the zebrafish hepatic cells. Diagram shows the main pathways studied, and the transporters and enzymes whose mRNA expression (and activity for some of them) was shown to be upregulated (+), downregulated (−) or unaltered (=) by FGF21 treatment in vivo and/or in vitro. Based on the results of the present study, we hypothesize that FGF21 leads to an increase in the amount of glucose in the hepatic cells by enhancing its entrance into the cells, gluconeogenesis and glycogen degradation. Interestingly, we also observed evidence in favor of FGF21 stimulating not only gluconeogenesis but also glycolysis. We hypothesize that the final product of glycolysis, pyruvate, after being converted into citrate, might be released into the cytoplasm and be used as a substrate for fatty acid biosynthesis. Our results do not support that the increase in fatty acid synthesis is related to its β-oxidation, but instead it might be related to the synthesis of phospholipids for maintaining cell membrane structure. In addition, we propose that FGF21 might stimulate ketogenesis. ACAT, Acetyl-CoA acyltransferase; ACC, Acetyl-CoA carboxylase; ACLY, ATP citrate lyase; CPT1a, Carnitine palmitoyltransferase 1a; ECH, Enoyl-CoA hydratase; ER, endoplasmic reticulum; FAS, Fatty acid synthase; FBPase, Fructose 1,6-bisphosphatase; G6Pase, Glucose 6-phosphatase; GCK, Glucokinase; GLUT2, Glucose transporter 2; HMGCL, 3-hydroxy-3-methylglutaryl-CoA lyase; HOAD, 3-hydroxyacyl-CoA dehydrogenase; MCAD, Medium-chain acyl-coenzyme A dehydrogenase; PEPCK, Phosphoenolpyruvate carboxykinase; PFK, Phosphofructokinase; PK, Pyruvate kinase; SGLT1, sodium-glucose cotransporter 1; TCA cycle, tricarboxylic acid cycle (Krebs cycle).

Scientific RepoRtS |
(2020) 10:6993 | https://doi.org/10.1038/s41598-020-63726-w www.nature.com/scientificreports www.nature.com/scientificreports/ levels of Pomc while lower levels of Agrp mRNAs in the hypothalamus compared to wild-type mice 31 . To exert actions in the zebrafish brain, it is likely that FGF21 crosses the blood-brain barrier, as it has been demonstrated in mammals 32 . In accordance with the orexigenic role reported in this study for FGF21 in zebrafish, we observed that mRNAs encoding FGF21 and its receptors in the foregut rise preprandially and decrease after a meal. Likewise, although no preprandial variations were detected, mRNAs encoding all components of the FGF21 system were reduced by feeding in the liver. This is the first report describing the periprandial variations in the FGF21 system in any species. Besides the periprandial profiles, the increase in the foregut expression of fgf21 and almost all of its receptors in response to fasting observed in this study is also a clear signal of an appetite enhancer. However, such a response was not observed in the brain and liver. Instead, our results demonstrated that a 7-day fasting significantly reduces the expression of FGF21 system mRNAs in zebrafish tissues. These observations contradicted with that of studies in mammals 12,33-35 , and might not be related to the role of FGF21 on zebrafish food intake, but with a different physiological action considering the multifunctional nature of the peptide suggested by its wide tissue distribution in this species. Differences in study methods, including the species used, time of sampling and the regions of tissues selected will all contribute to the difficulties in direct comparisons between this research and other studies. Further studies are needed to elucidate the physiological meaning of the fasting-induced downregulation of the FGF21 system in the zebrafish brain and liver. In addition, future research should consider studying long-term effects of FGF21 on feeding and body weight, as well as muscle and fat mass in fish.
Given the great abundance of the FGF21 system in the zebrafish liver, we hypothesized that FGF21 might also have an important role in glucose and lipid metabolism in zebrafish. Such a role for FGF21 has been reported in mammals (see reviews 10,36 ), but not in fish. Our results demonstrated that FGF21 regulates genes and enzymes involved in glucose and lipid metabolism in the zebrafish liver (Fig. 9), in a way that metabolic changes expected by results observed would mimic the metabolic effects of fasting, as in mammals 13 . Various observations support this major finding. First, FGF21 was found to upregulate the mRNA and protein levels of glucose transporters Glut2 and Sglt1 in vitro, which might suggest an increase in the amount of glucose entrance into hepatocytes. The reduced expression of sglt1 after FGF21 IP administration does not match the in vitro observations. This might be related to the fact that Sglt1 is an active transporter, and it might not be beneficial to use energy to increase the rate of glucose entrance in an in vivo situation. The suggested increase in the amount of glucose in the hepatic cells in response to FGF21 is also supported by the increased activity and/or expression of mRNAs encoding the gluconeogenic enzymes Pepck, fructose 1,6-bisphosphatase (Fbpase) and glucose 6-phosphatase (G6pase), and the increased mRNA levels of glycogen phosphorylase (implicated in glycogen degradation). In mammals, a FGF21-induced increase in gluconeogenesis and in the mRNA expression of G6pc and Pck was also described 13,37 . Previous reports have described that the expression of hepatic gluconeogenic enzymes 38 and glycogenolysis 39 are induced by leptin in fish. The increase in leptin a and leptin b mRNAs in the zebrafish liver after IP administration of FGF21 observed in this study suggests that the putative stimulation of gluconeogenesis and glycogenolysis by FGF21 in zebrafish might be occurring by the mediation of this hormone. Interestingly, our results suggest that not only gluconeogenic but also glycolytic pathways appear to be activated by FGF21 in the zebrafish liver, as suggested by the increased expression and activity of the glycolytic enzymes Gck, PfkA, PfkB and Pk. While these two processes are normally regulated so that they do not occur at the same time, previous reports have described the simultaneous stimulation of gluconeogenesis and glycolysis in fish 40 . In the context of the present study, it is possible that the high amount of glucose that might enter into the hepatic cells or is being synthesized in response to FGF21 is at the same time hydrolyzed to produce pyruvate. Resulting pyruvate can enter into the Krebs cycle and be used for obtaining energy. Indeed, FGF21 was previously observed to increase the Krebs cycle flux in mice 13 . A tendency towards increasing the gain of energy in response to FGF21 is also supported by the FGF21-induced upregulation of mRNAs encoding ATP5f1b, a subunit of mitochondrial ATP synthase.
In the Krebs cycle, pyruvate is first converted into citrate, which apart from being used in that cycle, could also be released into the cytoplasm through a specific mitochondrial carrier and be used as a substrate for fatty acid biosynthesis 41 . Our results point towards the possibility that FGF21 increases lipogenesis in the zebrafish liver, as suggested by the increased expression and/or activity of the enzymes Acly (catalyzes conversion of citrate into acetyl-CoA), Acc (catalyzes conversion of acetyl-coA into malonyl-CoA) and Fas (catalyzes conversion of malonyl-CoA into palmitate). Palmitate formed by the action of Fas can be either β-oxidized in order to obtain more energy or incorporated into phospholipids. In the present study, the unaltered expression of acadm and echs1 in vivo and in vitro, and the reduced expression (in vitro) and activity of Hoad (all enzymes involved in β-oxidation) in response to FGF21, seem to indicate that β-oxidation is not being upregulated by this peptide in the zebrafish liver. Instead, palmitate might be used for synthesizing phospholipids for maintaining cell membrane structure. In mammals, it has been described that about 60-70% of palmitate in the liver and most other tissues is incorporated into phospholipids, and only 20-30% is β-oxidized 42 . The observations that FGF21 seems to promote lipogenesis and does not stimulate β-oxidation in the zebrafish liver are opposite to that reported for mammals 11,13,14 . Additionally, the latter does not agree with the increased expression and activity of Cpt1a, which allows the entrance of fatty acids into the mitochondria. This, however, could be related to an increase in ketogenesis by FGF21, which indeed is suggested by the increased expression of hmgcl and acat1 in response to treatment with the peptide. Ketogenesis was reported to be induced by FGF21 in the mammalian liver 11,12 . This might indicate a need to produce extra energy when glycogen stores are depleted and glucose levels are low, a state that occurs during prolonged periods of fasting. The increase in Acly expression and activity supports the hypothesis that FGF21 might be stimulating ketogenesis in the zebrafish liver, given that the acetyl-CoA produced in the reaction catalyzed Acly can be used for synthesizing ketone bodies. A limitation of the present research is that it studied only the gene expression and activity of key metabolic enzymes, but not the cellular metabolism. While the observed effects of FGF21 on the metabolic machinery point to an alteration of the metabolic processes discussed, further studies on metabolite levels should be needed to confirm such a role for FGF21 in fish. (2020) 10:6993 | https://doi.org/10.1038/s41598-020-63726-w www.nature.com/scientificreports www.nature.com/scientificreports/ In addition, the liver samples from in vivo studies are modulated by an array of hormones and metabolites that represent a multiple, redundant milieu. Meanwhile, the results obtained from ZFL cells reflect a direct action of FGF21, without the influence of the in vivo milieu. Therefore, direct comparison of results obtained using these two models has its limitations.
The last aim of this research was to evaluate whether Ppargc1α and Pparα mediate the actions of FGF21 in the zebrafish liver. PPARGC1α and PPARα are key transcriptional regulators of energy homeostasis. Thus, the induction of PPARGC1α allows this protein to coactivate several transcription factors, including PPARα, which in turn regulates the expression of key genes and enzymes involved in gluconeogenesis, fatty acid oxidation, and other metabolic processes 43 . In mammals, FGF21 has been identified as a downstream of PPARα activation 11,12,35,44 , but also as a regulator of PPARGC1α 13,31,37,45 , suggesting that the mechanism of action for its physiological effects is mediated at least by this coactivator. The adipokine adiponectin has also been reported to mediate the metabolic effects of FGF21 in the mice liver 46 . In the present study, we reported that FGF21 modulates the expression of ppargc1α in vitro, suggesting a similar mechanism of action for FGF21 in fish than in mammals. However, the lack of FGF21-induced effects on ppargc1α mRNAs in vivo needs this hypothesis to be further confirmed. We also reported that Pparα might be a mediator of FGF21 actions in zebrafish, as its mRNA levels resulted modulated in ZFL cells treated by the peptide.
In summary, this study characterized FGF21 as an important orexigen in zebrafish, and reported that the expression of mRNAs encoding the peptide as well as its receptors is importantly modulated by feeding and food deprivation in central and peripheral tissues. We also showed for the first time in fish that FGF21 exerts an important role in regulating glucose and lipid metabolism in the zebrafish liver. Specifically, we observed that FGF21 appears to stimulate a fasting metabolic state, as it has been reported in mammals. Results presented here add significant new information to our growing knowledge on naturally occurring regulators of vertebrate physiology. To study the putative regulation of the FGF21 system by the composition of the diet, whether metabolic roles of FGF21 differ in the fed and fasted state, and a deeper characterization of the mechanisms underlying the physiological actions of the peptide, are important new directions to consider for future research. , supplemented with 0.15 g/L sodium bicarbonate, 15 mM HEPES, 0.01 mg/mL bovine insulin, 50 ng/ mL mouse epidermal growth factor (EGF), 5% heat-inactivated fetal bovine serum and 0.5% trout serum. At 80% confluency, ZFL cells were seeded at 5 × 10 5 cells/well in 24-well plates or 1 × 10 6 cells/well in 6-well plates, and the studies were performed when cells were 80-90% confluent (typically 48-72 h after seeding). An additional batch of cells was seeded in chamber slides for immunocytochemistry.

Methods
Reagents. The human recombinant FGF21 protein (amino acids 29 to 209) was obtained from Abcam (Catalog # ab54141; Toronto, ON, Canada). Recombinant human FGF21 was previously used in studies using fish models and was reported to exert biological effects 27 . For in vivo studies, peptide was prepared in sterile saline (0.9% NaCl) at a concentration of 1, 10 or 100 ng/10 µL saline. For in vitro studies, FGF21 was diluted in ZFL complete culture media at concentrations of 0.1 nM, 1 nM and 10 nM.
Experimental designs. Distribution of the FGF21 system mRNAs in zebrafish tissues involved in feeding and metabolism. Six zebrafish were anesthetized using tricaine methanesulfonate (MS-222; Syndel Laboratories, Nanaimo, BC, Canada) and sacrificed by decapitation. Samples of brain (without the hypothalamus), hypothalamus, foregut (intestinal bulb and anterior portion of the intestine), hindgut (posterior portion of the intestine) and liver (n = 6) were collected, immediately frozen in liquid nitrogen and stored at −80 °C until quantification of gene expression (see Real-time quantitative PCR section).
Effects of FGF21 on food intake. A total number of 12 tanks (3 zebrafish/tank) were set up, where fish were maintained as described earlier. Following a 3-week acclimation period, fish were divided into four experimental groups (3 tanks/group): i) Control, ii) 1 ng/g bw FGF21, iii) 10 ng/g bw FGF21, and iv) 100 ng/g bw FGF21. Food intake was registered daily during one week before treatment to evaluate basal levels of food intake. For this, a pre-weighed amount of food was offered to fish in each tank. After feeding for 30 min, the uneaten food was collected, dried for 24 h and weighed. The amount of food consumed by all fish in each tank was calculated by subtracting the dry weight of the amount of food retrieved from the tank after 30 min of feeding from the dry weight of the total amount of food originally provided. On the day of experiment, fish were lightly anesthetized with MS-222 at the scheduled feeding time, weighed and IP injected with either saline alone (control group) or containing 1, 10 or 100 ng/g of FGF21. Immediately after the injections, fish were allowed to recover (5 min) and (2020) 10:6993 | https://doi.org/10.1038/s41598-020-63726-w www.nature.com/scientificreports www.nature.com/scientificreports/ were offered a pre-weighed quantity of food. Uneaten food was recovered 1 h post-injections, dried and weighed in order to calculate the amount of food intake. Fish were again fed a pre-weighed amount of food at 2, 6 and 24 h post-injections, and the amount of food ingested at each time was quantified as described above. The experiment was repeated three times, and results shown correspond to the mean of three experiments.
Effects of fasting on the expression of the FGF21 system. Fish were divided into two groups, control and experimental (n = 6/group). Following a 2-week acclimation period, fish of the experimental group were not provided food for 7 days, while the fish in the control group were fed daily. On day 7, control and fasted fish were sacrificed at 11:30 h, and samples of whole brain, foregut and liver were collected as above described. The expression of FGF21 system genes was quantified as described below (see Real-time quantitative PCR section).
Periprandial changes in the expression of the FGF21 system. Seven groups of fish (n = 6/group) were established and acclimated to tank conditions during 2 weeks. On the day of the experiment, fish from three aquaria (6 fish per sampling time) were sampled at −3 h, −1 h and 0 h before the scheduled feeding time. At the scheduled feeding time, two of the remaining four tanks were fed while food was withheld from the other two tanks. Fish from both fed and unfed tanks were sampled at +1 h and +3 h after scheduled feeding time. Samples of brain, foregut and liver were collected as above described until analysis of mRNA abundance (see Real-time quantitative PCR section).
In vivo effects of FGF21 on the expression of genes involved in appetite regulation and glucose and lipid metabolism. Fish were divided into four experimental groups (6 fish/group): i) Control, ii) 1 ng/g bw FGF21, iii) 10 ng/g bw FGF21, and iv) 100 ng/g bw FGF21. After a 2-week acclimation period, fish were anesthetized at the scheduled feeding time and IP injected with either saline alone (control group) or containing 1, 10 or 100 ng/g of FGF21. Subsequently, fish were allowed to recover and fed. Following 2 h, fish were anesthetized again and sacrificed by decapitation in order to collect samples of brain, foregut and liver. Samples were kept at −80 °C until mRNA quantification (see Real-time quantitative PCR section).
In vitro concentration-and time-dependent effects of FGF21 on the expression of genes involved in glucose and lipid metabolism. ZFL cells were seeded at 5 × 10 5 cells/well in 24-well plates and grown to confluency as described earlier. Then, culture media was replaced by 1 mL of fresh media alone (6 wells) or containing 0.1, 1 or 10 nM FGF21 (6 wells each), and plates were incubated for 1 h and 6 h. At the end of each culture time, media was removed and 500 µL of PureZOL TM RNA Isolation Reagent (Bio-Rad, Mississauga, ON, Canada) was added to each well. Cells were then scraped from the bottom of the wells, collected and frozen at −80 °C until total RNA was extracted (see Real-time quantitative PCR section). This experiment was repeated twice.
In vitro effects of FGF21 on glucose transporters levels and the activity of enzymes implicated in glucose and lipid metabolism. For this assay, we chose the concentrations and time in which FGF21 exerts the most significant inductions in mRNA expression. ZFL cells were seeded at 1 × 10 6 cells/well in 6-well plates and grown to confluency. Once 80-90% confluency was achieved, media was replaced by 1 mL of fresh media alone (4 wells for Western blot and 8 wells for assessment of enzymatic activity) or containing 1 or 10 nM FGF21 (4 wells each for Western blot and 8 wells each for assessment of enzymatic activity). After an incubation period of 1 h, media was removed and 300 µL of lysis buffer was added to each well. For Western blot analysis, lysis buffer consisted of T-PER Tissue Protein Extraction Reagent (Thermo Fisher Scientific, Waltham, MA, USA), while samples for enzymatic activity assessment were lysed in a 80 mM Trizma buffer (pH 7.6) containing 5 mM EDTA, 2.6 mM DTT and protease inhibitor cocktail (Thermo Fisher Scientific). After the addition of the buffer, cells were scraped, transferred to tubes and stored at −80 °C until further analysis (see Sections Western blot and Enzymatic activity assessment).