miR-29a modulates SCD expression and is regulated in response to a saturated fatty acids diet in juvenile GIFT (Oreochromis niloticus)

MicroRNAs (miRNAs) are small non-coding RNAs that regulate target gene expression by binding to the 3′untranslated region (3′UTR) of the target mRNA. MiRNAs regulate a large variety of genes, including those involved in liver biology and disease. Here, we report for the first time that miR-29a post-transcriptionally regulates stearoyl-CoA desaturase (SCD) by binding to its 3′UTR in genetically improved farmed tilapia (GIFT), Oreochromis niloticus, as shown by a 3′UTR luciferase reporter assay. miR-29a antagomir treatment in vivo resulted in significant up-regulation of SCD expression. We found that miR-29a expression was negatively correlated with SCD expression in GIFT liver. Inhibition of miR-29a led to a significant increase in SCD expression on day 60 induced by a saturated fatty acids diet, thereby increasing conversion of 16:0 and 18:0 to 16:1 and 18:1 and activating serum insulin, which would favor glucose and lipid uptake by the liver. These results indicate that miR-29a regulates SCD levels by binding to its 3′UTR and this interaction affects saturated fatty acids stress induction and insulin and lipid accumulation in serum. Our results suggest that miR-29a is critical in regulating lipid metabolism homeostasis in GIFT liver and this might provide a basis for understanding the biological processes and therapeutic intervention encountered in fatty liver. Jo ur na l o f E xp er im en ta l B io lo gy • A dv an ce a rt ic le INTRODUCTION Genetically improved farmed tilapia (GIFT), Oreochromis niloticus, is a very adaptable species in southern China, such as Hainan, Guangxi, Guangdong, Fujian(Qiang et al., 2012). Increasing the amount of carbohydrate or lipid in their diet might help to improve growth and reduce feed cost (Tan et al., 2009). However, there is growing concern about fatty liver disease in GIFT fed high fat or high carbohydrate diets (Xiong et al., 2014; Qiang et al., 2016). Excess accumulation of lipid droplets within hepatocytes results in hepatic steatosis, which may cause metabolic dysregulation, reduce growth performance, and impair both bone development and the oxidative response (Tan et al., 2009; Boren et al., 2013; Qiang et al., 2016). In particular, the accumulation of excess saturated fatty acids (SFAs) in liver of fish has been tied intimately to pathogenesis of hepatic lipid metabolism, such as accumulation of triglyceride (TG) and total cholesterol (TC), and oxidative stress (Hixson et al., 2016; Yu et al., 2012). In fish, the liver plays a critical role in lipid homeostasis, which governs lipid synthesis, catabolism, storage, and secretion (Zhang et al., 2014). Although the mechanism underlying SFA-induced metabolic dysregulation in fish liver is unclear, several studies in mammals have suggested that the effected fatty acid oxidation and insulin resistance resulted from the regulation of stearoyl-CoA desaturase (SCD) (Yokoyama et al., 2012; Miyazaki et al., 2009), insulin receptor (Yang et al., 2014a), and insulin receptor substrate (Yang et al., 2014b) modification and expression. SCD is an enzyme involved in the biosynthesis of monounsaturated fatty acids (MUFAs), which play an important role in the regulation of hepatic lipid metabolism (Dobrzyn and Ntambi 2004; Ntambi and Miyazaki 2004). SCD plays a key role in introducing the first double bond between carbons 9 and 10 of palmitoyl (16:0)-CoA and stearoyl (18:0)-CoA to form monounsaturated palmitoleic acid (16:1) and oleic acid (18:1), respectively (Heinemann and Ozols 2003). High expression of SCD1 were found in adipose and liver of mouse. Repressed SCD1 expression of mouse were hyperphagic, but they were lean and reduced obesity induced by high lipid diet (Miyazaki et al., 2009). Additionally, the tolerance of glucose and insulin in whole body were stimulated in lack of SCD1 mice (Miyazaki et al., 2009). MicroRNAs (miRNAs) are small non-coding, single-stranded RNAs (18–25 nucleotides) found in plants, animals, and some viruses that function in RNA silencing and post-transcriptional regulation of gene expression by binding to the 3′untranslated region (3′UTR) of their target mRNAs (Bazzini et al., 2012; Pillai et al., 2007). MiRNAs play extremely important roles in regulating gene expression and have been reported to be instrumental in mediating liver biology and Jo ur na l o f E xp er im en ta l B io lo gy • A dv an ce a rt ic le disease in mammals (Bandiera et al., 2015; Bhatia et al., 2014). The miR-29 family is one of the best-known miRNA families, which has been associated with intermediate lipid metabolites such as ceramide and diacylglycerol (Yang et al., 2014; Mattis et al., 2015). In C57BL/6 mice fed SFA palmitate and a high-fat diet, miR-29a was induced and down-regulated the mRNA expression level of insulin receptor substrate-1 (IRS-1) to regulate insulin signaling in myocytes (Yang et al., 2014). In addition, inhibition of adipogenic gene expression and impaired accumulation of lipid in mice hepatocytes were induced by miR-29a regulation, which its plays a key role in energy metabolism of high fat diet (Yang et al., 2014). However, the detailed mechanism of hepatic steatosis in GIFT by the miR-29a regulatory remains unclear. We conducted this study to evaluate the involvement of miR-29a in GIFT lipid metabolism. We identified for the first time the SCD 3′UTR as a putative target of miR-29a through RNA-seq screening. Expression analysis by quantitative real-time RT-PCR (qRT-PCR) further revealed the miR-29a-mediated control of SCD in GIFT fed SFA diets. The present work suggests a novel mechanism where miR-29a is related to the development of hepatic steatosis in GIFT, and provides an improved understanding of lipid metabolism and the mechanisms involved in fatty liver disease for possible therapeutic intervention. Jo ur na l o f E xp er im en ta l B io lo gy • A dv an ce a rt ic le MATERIAL AND METHODS Ethics statement The study protocols were approved by the Freshwater Fisheries Research Center at the Chinese Academy of Fishery Sciences (Wuxi, China). The fish in well-aerated water were anesthetized by injecting 0.01% tricaine methanesulfonate (Sigma, USA) before sampling, and the livers were extracted based on the Guide for the Care and Use of Laboratory Animals in China. Small RNA library construction and verification Healthy GIFT from the Yixing tilapia farm of the Freshwater Fisheries Research Center, Chinese Academy of Fishery Sciences were selected as the experimental fish. Twenty-five GIFT were fed high fat diets (17% fat diet) as the stress group (HFD), and another 25 fish was fed normal fat diets (6.8% fat diet) as the normal group (NFD) (Table S1). Feed was offered to apparent visual satiety two time per day (8:00 and 15:00 h). Continuous aeration was applied during the 30-day experiment. Feces were removed daily using a siphon, and about 20% of the water was replaced every 3 days, with a temperature difference <0.3C. Dissolved oxygen was maintained at ≥6 mg·L, and pH was 7.4–7.8. Ammonia-N and nitrite concentrations were both <0.01 mg·L, and the photoperiod was 12 h:12 h (light:dark). At the end of the trial, liver tissues from nine fish (NFD group or HFD group) were collected separately and immediately stored in liquid nitrogen at −80C for transcriptome sequencing and expression analysis. The samples, which contained equal amounts of RNA extracted from the liver tissues from the HFD and NFD groups, were mixed and pooled to construct the miRNA libraries by high-throughput sequencing on a HiSeq 2000 system (Illumina, San Diego, USA). Low quality reads and reads contaminated with adapter sequences were removed before analysis (Xu et al., 2015). The filtered sequences (18–30 nt) were aligned to the zebrafish (Danio rerio) genome sequence using the short oligonucleotide alignment program (SOAP) (http://soap.genomics.org.cn) with a tolerance of one mismatch. To determine the differential expression levels of the miRNAs between the HFD and NFD libraries, the sequencing data were normalized (Zhu et al., 2010) and the fold-change was determined as: fold-change=log 2(HFD/NFD). The data from the two libraries were statistically analyzed based on the Audic and Claverie method (Audic and Claverie 1997). The differential expression levels of miRNA were verified by qRT-PCR. Jo ur na l o f E xp er im en ta l B io lo gy • A dv an ce a rt ic le MiR-29a targets prediction and tissue distribution in GIFT We used the Nile tilapia transcriptome data (https://www.ncbi.nlm.nih.gov/genome/?term= Oreochromis+niloticus) and the miRanda v3.01 toolbox to predict target genes that contained a single candidate site in the 3′UTR for the miRNAs. Targets identified using the default parameters and cutoffs (Score S ≥140; Energy E ≤−7.0 Kcal/mol) (Betel et al., 2010) and related to lipid metabolism were selected as promising candidate genes. Six tissues (muscle, liver, kidney, spleen, blood, and gut) from healthy GIFT were selected and the miRNA tissue distribution was determined by qRT-PCR. RNA preparation and qRT-PCR MiRNAs were extracted using an amiRNeasy kit (Takara, Dalian, China) according to the manufacturer's protocol and a Mir-XTM miRNA First-Strand Synthesis kit (Takara, Dalian, China) was used to synthesize the first-strand cDNA. The 10.0 μL reverse transcription (RT) reaction mixture contained 5.0 μL 2× mRQ Buffer, 3.75 μL RNA sample (0.25–8 μg), and 1.25 μL mRQ Enzyme. The reactions were incubated at 37C for 60 min, 85C for 5 min, and held at 4C. The qRT-PCRs were performed using an miRNA SYBR Green qRT-PCR Kit (Takara, Dalian, China) with the provided miRNA reference gene (U6). The 25 μL PCR contained 2.0 μL of the RT product (template), 12.5 μL 2× SYBR Advantage Premix, 9 μL ddH2O, 0.5 μL 50× ROX Dye, 0.5 μL miRNA-specific primer (10 μM), and 0.5 μL mRQ 3′primer. The default thermal profile used for PCR amplification consisted of 95C for 10 s, followed by 40 cycles of 95C for 5 s and 60C for 20s, with a final dissociation curve at 95°C for 60 s, 55°C for 30s, and 95°C for 30s. Dissociation curve analysis of amplified products was performed at the end of each PCR reaction to confirm that only one PCR product was amplified and detected. For each cDNA, three well replicates were used. The threshold cycle (Ct) value was determined using the automatic setting on the ABI 7900HT Fast Real-Time PCR system (Applied Biosystems, USA). Ct values determined for each sample were normalized against the values for U6. The relative fold change in expression to U6 was calculated by 2 method (Livak and Schmittgen, 2001), and the values related to the control group represented the n-fold difference. In order to detect the presence of non-specific amplifications, control reactions without template were included for each primer set. The miR-29a-specific primer (5′-GCACCATTTGAAATCGGTTAG-3′) was synthesized by Genewiz, Inc. (Genewiz, Suzhou, China). Jo ur na l o f E xp er im en ta l B io lo gy • A dv an ce a rt ic le Total RNA was isolated with TRIzol Reagent (Invitrogen, USA). PrimeScriptTM RT Master Mix (Takara, Dalian, China) was used for the RT reaction of the miRNA target genes. The 10.0 μL RT reaction contained 2.0 μL 5× PrimeScript RT Master Mix, RNA sample (≤500ng), and RNase Free dH2O up to 10 μL. The reactions were incubated at 37C for 15 min, 85C for 5 s, and held at 4C. The qRT-PCRs were analyzed using SYBR Premix Ex Taq kits. The 20 μL PCR included 2.0 μL of the RT product, 10.0 μL 2× SYBR Premix Ex Taq II, 0.8 μL each of PCR forward and reverse primers (10 μM), 0.4 μL 50× ROX Dye, and 6 μL ddH2O. The reactions were incubated at 95C for 30 s followed by 40 cycles of 95C for 5 s and 60C for 30s. The 18S rRNA transcript level was taken as a reference to calculate the relative expression level of each target gene. The mRNA primers for SCD were F: 5′-ACAAGCTCTCCGTGCTGGTCAT-3′, R: 3′-GCAGAGTTGGGACGAAGTAGGC-5′ and for 18S rRNA they were F: 5′-GGCCGTTCTTAGTTGGTGGA-3′ and R: 5′-TTGCTCAATCTCGTGTGGCT-3′). The primers were synthesized by Shanghai GeneCore Bio Technologies Co., Ltd. (Shanghai, China). The mRNA expression levels were analyzed using the 2 method (Livak and Schmittgen, 2001), and the values related to the control group represented the n-fold difference. The mRNA expression levels were quantified using an ABI 7900HT Fast Real-Time PCR System (Applied Biosystems, USA) and compared using Relative Quantification (RQ) manager software. SCD 3′UTR luciferase reporter assay HEK 293T (human embryonic kidney) cells were obtained from the Shanghai Bioleaf Biotech Co., Ltd (Shanghai, China). The cells were maintained in Dulbecco’s modified Eagle’s medium with 10% fetal bovine serum, 100 U.ml penicillin, 100 μg.ml streptomycin and 250 ng.ml –1 amphotericin B, and maintained at 37°C in a humidified 5% CO2 incubator. The reagents were purchased from American Sigma-Aldrich Co. LLC (Sigma, USA). The SCD 3′UTR-luciferase reporter vector was constructed for analyzing the potential miRNA target sites and the effect of miR-29a on its activity was evaluated in the HEK293T cell line. The full-length 3′UTR from SCD was chemically synthesized and inserted into downstream of the luciferase gene in the pGL3-control vector (Promega, USA). To construct the pGL3-SCD mutant, six base pairs (5′-GTGCTA-3′) in the SCD 3′UTR region were deleted and six new base pairs (5'-AGCTAC-3′) were inserted. Synthetic miRNA mimics were synthesized by RiboBio Bioscience Co., Ltd. (Guangzhou, China) as RNA duplexes designed from the miR-29a sequence (5′-CUAGCACCAUUUGAAAUCGGUUA-3′) in the miRBase database. A scrambled miRNA Jo ur na l o f E xp er im en ta l B io lo gy • A dv an ce a rt ic le mimic (5′-UUUGUACUACACAAAAGUACUG-3′) with no homology to the tilapia genome was used as the negative control. Twenty-four hours prior to transfection, HEK 293T cells were plated at 1.0×10 cells per well in 12-well dishes. Cells were transfected with 25 ng firefly luciferase reporter vector containing either the wild-type or mutant 3′UTR constructs, with or without 50 nM miR-29a mimic or negative control. and 5 ng Renilla luciferase control vector (pRL-TK-Promega, USA), using Dharma FECT Duo (Thermo Scientific Dharmacon). Firefly luciferase activity was normalized to Renilla luciferase activity for each transfected well. Then, 36 h after transfection, the cells were washed with ice-cold PBS and centrifuged at 800 × g at 4°C for 5 min to harvest the cells. A liquid scintillation counter (Hitachi, Japan) was used to detect luciferase activity with a standard dual-luciferase reporter system according to the manufacturer’s instructions (Krek et al., 2005). The assay kits were purchased from Shanghai Lengton Bioscience Co., Ltd. (Shanghai, China). Five biological replications were performed for each treatment. Functional analysis of miR-29a in vivo A chemically modified antisense oligonucleotide (miR-29a antagomir: 5′-UAACCGAUUUCAAAUGGUGCUAG-3′) was synthesized to regulate miR-29a expression (Ribobio, Guangzhou, China). The 3′end of the oligonucleotide was conjugated to cholesterol, and all the bases were 2′-OMe modified. The antagomir oligonucleotide was deprotected, desalted, and purified by high-performance liquid chromatography. To analyze the silencing miR-29a expression levels, 270 juvenile GIFT weighing about 4.5 g were distributed randomly to nine 600-L tanks each containing 30 fish. The fish were tail-vein injected with miR-29a antagomir, negative antagomir (four mismatch mutations in each miRNA sequence), or the same volume of PBS at a dose of 50 mg kg body weight every three days for 21d. The PBS-treated fish were taken as the control (Yan et al., 2013a; b). Livers were sampled from three fish from each tank at the following time periods: 0, 7, 14, and 21 days. The liver tissues were collected, immediately frozen in liquid nitrogen, and stored at −80°C until used for qRT-PCR. At the end of the trial, another three fish were selected randomly from each tank. Liver samples were obtained and store for analyses of biochemistry and enzyme activities. After collecting the samples, the body of the 10 fish per tank were measured as: Jo ur na l o f E xp er im en ta l B io lo gy • A dv an ce a rt ic le Weight gain (WG) = (W2−W1)/W1 Specific growth rate (%·d)(SGR) = [(ln W2−ln W1)/(t2−t1)]×100 where W1 and W2 are body weight (g) at the start (t1) and end (t2) of the experimental period. Saturated fatty acid-regulated trial The feeding trial was performed in a water recirculating system comprising six plastic tanks (800 L each) maintained at 290.3C. GIFT were fed either a control diet (5% fish oil) or SFAs diet (SFA, 5% coconut oil) (Table S2). About 600 L of treated water (aerated for three consecutive days) was added to each of the six plastic tanks, and 30 fish per tank were fed each experimental treatment (each treatment had three replicates). At the beginning of this experiment 30 fish were taken in triplicate from a common tank to record initial weights. Wet weight was recorded using an electronic digital balance (0.01 g), and the initial mean weight of each juvenile was 2.780.11g. Results of a multivariate analysis of variance showed no differences in weight between fish in the different treatments and replicate treatments (P>0.05). Rearing management was the same as that used in the experiment of high fat diets. Feeding was stopped 24 h prior to collecting the samples. Three fish were collected from each tank on days 20, 40, and 60 during the experiment, and liver samples were obtained, frozen immediately in liquid nitrogen, and stored at −80°C until mRNA levels and fatty acid compositions were measured. At the end of the trial, six fish were selected randomly from each tank. Blood samples were collected from the caudal vein of three fish from each tank. All the blood samples were kept at 4C for 2 h and then centrifuged at 4C and 3500 × g for 10 min to collect the serum, which was stored at −80C for later use. Liver samples were obtained from the remaining three fish, immediately frozen in liquid nitrogen, and stored at −80°C for later biochemistry analyses. Finally, liver tissue blocks of six fish from two treatments (one fish from each of six tanks) were washed with physiological saline and fixed with Bouin's solution and 2.5% glutaraldehyde solution for 24 h in order to prepare paraffin sections, which were examined using an optical microscope. After collecting the samples, the body and liver weights of the remaining fish were measured as: WG, SGR and hepatopancreas somatic index (%)(HSI) = (liver weight×100)/W2. Serum and hepatic biochemical analyses The levels of serum glucose, triglyceride (TG), and cholesterol (TC) were measured using a Roche Cobas C311 automatic biochemical analyzer (Roche, Basel, Switzerland). Serum alanine aminotransferase (ALT) and aspartate aminotransferase (AST), and hepatic TG, TC, and glycogen Jo ur na l o f E xp er im en ta l B io lo gy • A dv an ce a rt ic le were determined by enzyme-linked immunosorbent assay (ELISA) using test kits. All assay kits used were from Shanghai Lengton Bioscience, Shanghai, China. Serum insulin levels were measured by radioimmunoassay using bonito insulin as the standard and rabbit anti-bonito insulin as antiserum, according to the method described by Gutiérrez et al. (1984). The minimum detection limit was 0.15 mIU.L, with intraand inter-assay coefficients of variation of 4.4% and 10.3%, respectively (n=9). Fatty acid analysis Total lipids were extracted from liver tissue and fatty acid analyses were according to the methods described previously (He et al., 2015). Statistical analysis Statistical analysis of data was done by two-way ANOVA using the SPSS 17.0 program (SPSS Inc., Chicago, IL, USA). Each value represents nine replicates. Data were tested for normality and homogeneity of variances, and transformed to logarithm when necessary. When the effects of sampling times and/or type of lipid sources were significant, the two factors were analyzed separately by one-way ANOVA. A P-value <0.05 was considered significant. Significant differences (P <0.05) in different treatments at each sampling point were calculated with Duncan’s multiple range tests. Significant differences (P <0.05) between values obtained at the different sampling times were calculated with paired-samples t tests.