Lipids are essential nutrients for fish. Dietary lipids at appropriate levels can provide energy and essential fatty acids that are beneficial for fish growth. Within certain limits, increasing dietary lipid levels can reduce the amount of protein required in the diet (Figueiredo-Silva et al., 2010; Zheng et al., 2010). However, a high-fat diet (HFD) can result in unwanted lipid accumulation in fish, resulting in fatty liver. Several studies have shown that fatty liver induced by a HFD is related to changes in blood biochemistry parameters, liver histology, and fatty acid composition. Such changes, which have been observed in blunt snout bream (Megalobrama amblycephala) (Lu et al., 2013a) and grass carp (Ctenopharyngodon idella) (Du et al., 2006), are indicators of abnormal metabolism and physiological status. Excess lipid intake has been shown to cause oxidative stress and affect the immune responses and disease resistance of marbled rockfish (Sebastiscus marmoratus) (Shi et al., 2013), turbot (Scophthalmus maximus) (Jia et al., 2017), and blunt snout bream (Adjoumani et al., 2017). Several studies have explored the mechanism of fatty liver formation in fish induced by a HFD (Lu et al., 2013b; Wang et al., 2015). The results of those studies suggested that hepatic steatosis induced by a HFD is strongly associated with abnormal lipid metabolism, including changes in lipid synthesis, uptake, and transport. For example, Lu et al. (2013b) showed that a HFD activated the activities of lipoprotein lipase and changed the mRNA expression levels of PPARα, encoding peroxisome proliferator-activated receptor-α, and PPARγ, encoding peroxisome proliferator-activated receptor-γ. Those changes resulted in the inhibition of lipid transport out of the liver and elevated lipid uptake, which contributed to fatty liver in blunt snout bream. However, few studies have focused on the physiology and molecular biology of fat deposition in tilapia.

With the rapid development of intensive fish farming (Malik et al., 2018), tilapia has become susceptible to hepatic steatosis when fed a HFD with an aim to use less protein (Ng and Romano, 2013; Huang et al., 2016, 2018). Several reports have shown that HFD-induced fatty liver is often accompanied by low growth and low immune function in farmed fish (Lu et al., 2013a; Ma et al., 2018). Therefore, it is particularly important to treat or alleviate HFD-induced fatty liver in tilapia. Genetically improved farmed tilapia (GIFT, Oreochromis niloticus) is bred from selected tilapia breeding stocks from Africa and Asia (Tendencia et al., 2006). Because of its excellent growth performance and disease resistance, GIFT is highly marketable and potentially has a broad market (Ng and Romano, 2013; Qiang et al., 2013). It is one of the most important commercial fish, and accounts for almost 75% of the total aquacultured tilapia in China at present. Thus, research to better understand the mechanism underlying the development of fatty liver in GIFT is particularly important from both biological and economic perspectives.

Recently, microRNAs (miRNAs) have emerged as key regulators of gene expression at the posttranscriptional level. These molecules are involved in multiple biological processes (Bartel, 2004). Thus, exploring the role of miRNAs can help us to better understand the underlying mechanisms that guide the physiological changes in living organisms. Most miRNAs in animals inhibit effective mRNA translation of target genes through imperfect complementary base-pairing with the 3′-untranslated region (3′-UTR) of their target mRNAs (Guo et al., 2010; Bizuayehu and Babiak, 2014). Many studies have revealed that miRNAs are important regulators of liver function in teleost fish (Her et al., 2011; Mennigen et al., 2014; Zhang et al., 2014; Wang et al., 2016). Particularly, miR-122 has been shown as a biomarker of liver injury (Florczyk et al., 2016) and play a role in hepatic lipid metabolism in grass carp (Wang et al., 2016), rainbow trout (Oncorhynchus mykiss) (Mennigen et al., 2014), and zebrafish (Danio rerio) (Her et al., 2011). Analyses of our deep sequencing data revealed that some miRNAs were differentially expressed in the liver between GIFT fed a HFD and those fed a normal-fat diet (NFD) for 60 days. This finding suggested that these miRNAs may play important roles in the GIFT liver in response to a high fat intake (Tao et al., 2017). In this study, we chose three clearly differentially expressed miRNAs (miR-29a, miR-145-5p, and miR-34a) and miR-122, a potential hepatic lipid metabolism-related miRNA in teleost fish (Her et al., 2011; Mennigen et al., 2014; Wang et al., 2016), to further explore their roles in hepatic lipid metabolism in GIFT.

In our previous study, miRanda software¹ was used to predict potential target genes with miRNA target sites in their 3′-UTRs based on Nile tilapia whole genome sequence data². The miRanda parameters and cutoffs were as follows: score ≥140 and free energy ≤-7 kcal/mol (Betel et al., 2010). Gene Ontology (GO)³ was used to detect the functions of the miRNAs’ target genes. The results of those analyses indicated that SCD (encoding stearoyl-coenzyme A desaturase), with the relatively high score (153) and low free energy (-25.0 kcal/mol) among target genes, may be a target of miR-122. The SCD gene also had a relatively high score (153) and low free energy (-16.07 kcal/mol) among potential target genes of miR-29a. The gene ELOVL6, which encodes elongation of the very long chain fatty acid protein 6, had a relatively high score (162) and low free energy (-23.17 kcal/mol) among target genes of miR-145-5p. The SRD5A2 gene, which encodes steroid 5 alpha-reductase 2, had a relatively high score (167) and low free energy (-26.47 kcal/mol) among potential target genes of miR-34a (Tao et al., 2017).

Stearoyl-coenzyme A desaturase (SCD) is a rate-limiting enzyme in monounsaturated fatty acids (MUFA) synthesis, and catalyzes the conversion of palmitic acid (C16:0) and stearic acid (C18:0) into palmitoleic acid (C16:1) and oleic acid (C18:1), respectively (Heinemann and Ozols, 2003). ELOVL6 plays a crucial role in elongating saturated fatty acids (SFA) and MUFA with 12, 14, and 16 carbons to form 18-carbon fatty acids (Moon et al., 2001). SRD5A2 was shown to suppress lipogenesis by inhibiting the effects of cortisol (Nasiri et al., 2015). Because these three enzymes are important regulators in lipid metabolism, we selected their encoding genes as potential miRNA targets that warranted further analysis.

The principal goal of this research was to characterize the potential mechanisms of GIFT fatty liver formation by investigating changes in physiological indexes, miRNA expression levels, and transcript levels of potential lipid metabolism-related target genes in the liver of GIFT in response to a HFD. Our results suggest that further research on techniques to attenuate hepatic steatosis induced by HFD in GIFT is warranted.

China is the world’s largest producer of aquacultured tilapia (Malik et al., 2018). However, tilapia is susceptible to hepatic steatosis as a result of dietary imbalances in modern culture systems (Ng and Romano, 2013; Huang et al., 2016, 2018). The mechanism underlying the development of fatty liver in tilapia is still not fully understood. Our results confirmed that a HFD (18.5% lipids) quickly and effectively induced the formation of fatty liver in GIFT juveniles. Our results also suggested that the dysregulation of hepatic miRNAs (miR-122, miR-29a, miR-145-5p, and miR-34a) may disturb lipid homeostasis in GIFT juveniles by affecting the expression of their target genes in lipid metabolism.

Hepatic steatosis is strongly associated with abnormal lipid metabolism, which is mainly reflected as an imbalance between hepatic lipid inputs and outputs (Du et al., 2006; Lu et al., 2013b; Wang et al., 2015). When excessive lipids accumulate in the liver, the relevant pathways are activated and lipoproteins deliver lipids to the peripheral tissues (Lu et al., 2013b). Blood serves as an important metabolic transport system to regulate lipid metabolism. In this study, we found that serum TG, TC, LDL-C, and insulin levels were increased in GIFT fed a HFD, suggesting that endogenous metabolic transport was activated in these GIFT. An increased level of serum insulin would promote glucose uptake by the liver and consequently contribute to better glycemic control (Qiang et al., 2016). Similar results have been reported for blunt snout bream (Zhang et al., 2018). In fish, AST and ALT are two important transaminases that are mainly found in hepatocytes. Their activities in the serum may increase when the liver is damaged or inflamed (Sheikhzadeh et al., 2012). Together with the liver histological characteristics and hepatic lipid index assay results, the higher serum ALT activity observed in the HFD group suggested that excessive fat deposition in the liver not only altered hepatocytes, but also led to liver damage in GIFT.

Du et al. (2006) reported that dietary fatty acids strongly affect the fatty acid composition in fish tissues. In this study, we used soybean oil, which is rich in linoleic acid (C18:2n-6) and α-linolenic acid (C18:3n-3), as the oil source. This may have led to the high proportion of hepatic C18:2n-6 and C18:3n-3 in the HFD-fed GIFT. Linoleic and linolenic acids are considered as essential fatty acids (EFAs) for most animals, including fish (Holman, 1986). Fish can convert C18:2n-6 and C18:3n-3 into longer-chain, more unsaturated fatty acids (Tocher et al., 2001). In this study, the higher ratio of hepatic C20:2n-6 and C20:5n-3 in HFD-fed GIFT suggested that these fish may have been able to convert C18 EFAs into longer-chain n-3 and n-6 PUFA. We also found that the HFD led to an increase in the proportion of hepatic PUFA. A study on grass carp (Wang et al., 2016) suggested that fatty acid metabolic disorders may occur when fish are fed a HFD, because this leads to higher PUFA contents in the liver and possibly to the formation of fatty liver.

Excessive fat deposition in fish tissues can induce oxidative stress and impair fish health (Chaiyapechara et al., 2003; Zhang et al., 2018). The product of lipid peroxidation is MDA, and so the MDA level can indirectly reflect the degree of oxidative damage in fish tissues (Jiang et al., 2016). In fish, SOD, CAT, and GSH-Px are three important antioxidant enzymes; SOD can convert superoxide radicals into hydrogen peroxide, which can be further eliminated by GSH-Px and CAT (Jiang et al., 2016). In this study, hepatic GSH-Px and CAT activities were increased by day 20 in HFD group, indicating that short-term high-fat feeding (20 days) may activate the antioxidant system to remove excess oxygen free radicals. However, 60 days of high-fat rearing led to decreased hepatic SOD, CAT, and GSH-Px activities and an increased MDA content, consistent with the findings of studies on Chinese sucker (Myxocyprinus asiaticus) (Wang et al., 2010) and Japanese pufferfish (Takifugu rubripes) (Sun Y. et al., 2013). This suggested that excessive fat deposition in the liver may have led to a hepatic metabolic disorder, which attenuated hepatic antioxidant defenses and increased oxidative stress.

To investigate the possible mechanism of HFD-induced fatty liver in GIFT juveniles, we determined the expression levels of several miRNAs (miR-122, miR-29a, miR-145-5p, and miR-34a) and their predicted lipid metabolism-related target genes (SCD, ELOVL6, and SRD5A2) on days 20, 40, and 60. miR-122 is a liver-specific miRNA that was shown to account for almost 60% of total miRNAs in GIFT liver (Tao et al., 2017). The large amount of miR-122 in GIFT liver suggests that it plays a crucial role in regulating liver function. An increased expression level of miR-122 was linked with hepatic steatosis in zebra fish (Her et al., 2011). miR-29a is one of the mature forms of the miR-29 family and it is highly conserved in most species (Liston et al., 2012). Yang et al. (2014) reported that miR-29a was activated in C57BL/6 mice fed a HFD and the SFA palmitate, and that it functionally targeted the 3′-UTR of insulin receptor substrate-1 to regulate insulin signaling. We obtained a similar result, in that a HFD induced the expression of hepatic miR-122 and miR-29a in juvenile GIFT. SCD is a rate-limiting enzyme in the synthesis of MUFA from SFA (Heinemann and Ozols, 2003). An SCD1 deficiency in mouse led to hyperphagy, but the mice were lean and protected from HFD-induced obesity (Miyazaki et al., 2009). Ntambi (1999) reported that MUFA are more likely than SFA to combine with acyl-CoA cholesterol and triglycerides to produce TG and TC, respectively. In this study, SCD expression was suppressed in the HFD group. The results suggested that miR-122 together with miR-29a may be involved in reducing the synthesis of MUFA through attenuating the expression levels of SCD. This may contribute to the control of fat synthesis (especially TG and TC).

Wang et al. (2017) reported that miR-145 is a key regulator of lipogenesis in goat mammary cells. Overexpression of miR-145 activated the expression of fat synthesis-related genes at the transcript level, resulting in greater fat accumulation. The expression levels of hepatic miR-145 decreased in blunt snout bream fed a HFD (Zhang et al., 2014). However, in this study, miR-145-5p expression was increased in the liver of HFD-fed GIFT on day 60. We speculated that, different from other fish species, GIFT may have some other adaptation strategies to a HFD. ELOVL6, the only elongase involved in de novo lipogenesis, catalyzes the rate-limiting step in 18-C fatty acid synthesis (Moon et al., 2001). A deficiency of ELOVL6 was shown to alter the ratio of hepatic fatty acids in mice (Matsuzaka et al., 2007). Inhibition of this elongase could be a new therapeutic approach for ameliorating diabetes (Matsuzaka et al., 2007; Sun H. et al., 2013). Several studies have reported that fish can change their metabolic strategy (to decrease lipogenesis and increase β-oxidation of fatty acids) to adapt to a high fat intake (He A.Y. et al., 2015; Li et al., 2016). In this study, the proportion of C18:2n-6 was almost 3.5-fold higher in the HFD group than in the NFD. We speculated that in juvenile GIFT fed a HFD, to maintain fatty acid homeostasis, the levels of miR-122, miR-29a, and miR-145-5p increased and the transcript levels of SCD and ELOVL6 decreased to reduce the conversion of fatty acids into C18:2n-6.

In mammals, miR-34a is thought to be involved in non-alcoholic fatty liver disease (Ding et al., 2015). Moreover, miR-34a^-/- mice were shown to be susceptible to HFD-induced obesity (Lavery et al., 2016). Nasiri et al. (2015) reported that SRD5A2 is involved in regulating lipid homeostasis in human hepatocytes. Overexpression of SRD5A2 was shown to inhibit the effects of cortisol and thereby suppress lipogenesis. In this study, a HFD downregulated miR-34a expression but upregulated SRD5A2 in the liver of GIFT. This result suggested that the interaction between miR-34a and SRD5A2 mRNA may play an important role in lipid metabolism in GIFT fed a HFD.

The results of our study indicated that a HFD can cause excess fat deposition in the liver of GIFT juveniles. The excess fat deposition in the liver altered serum parameters, hepatic fatty acid composition, and antioxidant enzyme activities. This is the first report of differential expression patterns of four miRNAs (miR-122, miR-29a, miR-145-5p, and miR-34a) and their potential target genes (SCD, ELOVL6, and SRD5A2) in the liver of GIFT during 60 days of growth on a HFD. Our results suggested that these four miRNAs may be involved in lipid metabolism by post-transcriptionally regulating the expression levels of SCD, ELOVL6, and SRD5A2. Further research including western blotting and luciferase reporter assays is required to validate the functions of miR-122-SCD, miR-29a-SCD, miR-145-5p-ELOVL6, and miR-34a-SRD5A2 pairs in the lipid metabolism of GIFT.

G-JY and PX conceived and designed the experiments. Y-FT, JQ, J-WB, D-JC, and H-JZ performed the experiments. Y-FT and JQ analyzed the data and wrote the manuscript. All authors read and approved the final version of the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.