gga-miRNA-18b-3p Inhibits Intramuscular Adipocytes Differentiation in Chicken by Targeting the ACOT13 Gene

Intramuscular fat (IMF) is the most important evaluating indicator of chicken meat quality, the content of which is positively correlated with tenderness, flavor, and succulence of the meat. Chicken IMF deposition process is regulated by many factors, including genetic, nutrition, and environment. Although large number of omics’ studies focused on the IMF deposition process, the molecular mechanism of chicken IMF deposition is still poorly understood. In order to study the role of miRNAs in chicken intramuscular adipogenesis, the intramuscular adipocyte differentiation model (IMF-preadipocytes and IMF-adipocytes) was established and subject to miRNA-Seq. A total of 117 differentially expressed miRNAs between two groups were obtained. Target genes prediction and functional enrichment analysis revealed that eight pathways involved in lipid metabolism related processes, such as fatty acid metabolism and fatty acid elongation. Meanwhile a putative miRNA, gga-miR-18b-3p, was identified be served a function in the intramuscular adipocyte differentiation. Luciferase assay suggested that the gga-miR-18b-3p targeted to the 3′UTR of ACOT13. Subsequent functional experiments demonstrated that gga-miR-18b-3p acted as an inhibitor of intramuscular adipocyte differentiation by targeting ACOT13. Our findings laid a new theoretical foundation for the study of lipid metabolism, and also provided a potential target to improve the meat quality in the poultry industry.


Introduction
Meat products are essential for humans to maintain a balanced diet. Recently, great progress has been achieved in meat quantity by genetic selection for growth rate and meat yield. However, the strong selection led to the decline in the meat quality, which presented as larger fiber diameters, higher proportion of glycolytic fibers, and lower intramuscular fat (IMF) [1,2]. Among them, IMF was considered to be the main factor affecting meat quality. It has been demonstrated that IMF content, the amount of fat within muscles, was positively correlated with flavor, juiciness, and tenderness [3]. Gushi chicken as a Chinese local chicken breed with high meat quality, has been favored by most consumers for its delicious and unique-flavored meat. Numerous studies have shown that intramuscular fat not only correlated with muscle flavor, but also related to many other meat quality traits (such as muscle pH, hydraulics, tenderness, etc.) [2][3][4][5][6]. Our previous studies have found that there were significant differences in the density and size of lipid droplets in the breast muscle between the juvenile and

Small RNAs Library Construction and Deep Sequencing
Total RNA of IMF-preadipocytes and IMF-adipocytes were prepared using mirVana™ miRNA Isolation Kit (Austin TX, US). After quality checking,~2 µg total RNA for each sample was used to construct the small RNA library according to the protocol (Illumina Small RNA Sample Prep Kit). A total of 4 libraries (IMF-preadipocytes and IMF-adipocytes, each with 2 replicates) were sequenced with Illumina Genome Analyzer (Illumina, San Diego, CA, USA). Meanwhile, cDNA was also prepared using PrimeScript™ RT reagent kit with gDNA Eraser (Takara, Dalian, China) and stored at −20 • C for the following wet-lab experiment.

Identification and Prediction of miRNAs
After a series of quality control, clean reads for each sample were mapped to Gallus gallus 5.0 using Bowtie (http://bowtie.cbcb.umd.edu) (The University of Maryland, Maryland, MD, America) [27], and then aligned with the known miRNAs in miRbase [28] with allowing less than one mismatch to identify known miRNAs. The identified miRNAs were classified into families based on their sequence similarities. Unmatched reads remained were further processed to predict novel miRNAs using miRCat [29]. RegRNA 2.0 was applied to predict the structures of novel miRNAs precursors, hairpins, secondary structures, and minimum free energy [30].

Prediction of miRNA Target Genes and GO and Pathway Enrichment Analysis
Differentially expressed miRNAs with p value ≤ 0.05 were identified and their target genes were predicted by TargetScan (http://www.targetscan.org/) and miRanda (microRNA.org). Correlation networks was visualized using Cytoscape software [31]. Gene ontology (GO) enrichment and Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis were performed to investigate the underlying functions and some vital pathways involved in lipid metabolism using DAVID (The Database for Annotation, Visualization, and Integrated Discovery) database (https://david.ncifcrf.gov/).

Vector Construction
The chicken ACOT13 coding region was cloned using primer ACOT13(CDS), isolated with EcolI and HindIII, then inserted into pcDNA3.1-EGFP to obtain the pcDNA3.1-ACOT13. Similarly, a fragment amplified from the GFP gene was also cloned into pcDNA3.1-EGFP as a negative control (NC). The 3 UTR region of ACOT13 gene fragments containing the predicted target site, including wild type (WT) and mutant type (MUT), were cloned into psiCHECK2 vector (isolated using XhoI and NotI (Takara, Dalian, China)). The recombinant vectors were named as ACOT13-3 UTR-WT and ACOT13-3 UTR-MUT, respectively. The primers are shown in Table 1. DF1 cell, a cell line of chicken embryo fibroblasts, was used to validate the miRNA target. Cells were seeded into 24-well plates. Co-transfection with 500 ng ACOT13-3 UTR-WT or ACOT13-3 UTR-MUT and 2 µL of gga-miR-18b-3p mimics or negative control were performed with Lipofectamine 2000 (Invitrogen). Then luciferase activities were measured using the Dual Luciferase Reporter Assay System (Promega, WI, USA) at 48 hours post-transfection. Assays were performed in triplicate.

Overexpression and Knockdown Assay in Preadipocytes
Primary preadipocytes were seeded into 6-well plates and transfected with gga-miR-18b-3p mimics or si-ACOT13 (small interfering RNA of ACOT13 gene, GenePharma, Shanghai, China), meanwhile their non-targeting sequences were used as negative control. Furthermore, a rescue experiment was conducted after ACOT13 gene knockdown in preadipocytes, that was to transfect the pcDNA3.1-ACOT13 into cells. Post-transfected cells were collected for analyzing the transfection efficiency, the cell differentiation, and gene expression at 48 hours post-transfection. 2.7.4. Quantitative PCR cDNA was obtained as mentioned above in Section 2.4. Specially, the cDNA synthesis and quantitative PCR (qPCR) for miRNAs were performed with Bulge-loopTM miRNA qPCR Starter Kit (Ribobio, Guangzhou, China). β-actin and U6 genes were used as reference genes for genes and miRNAs respectively. qPCR amplification was conducted as described previously [32,33]. All the primers used are listed in Table 1.

Statistical Analysis
Relative gene expression obtained from q-PCR data was calculated using the 2 −∆∆Ct method [34]. All of the data was presented as the mean ±SE (standard error) with at least three independent replicates and visualized using the "ggplot2" package in R (version 3.2.2, the University of Auckland, Auckland, New Zealand) and GraphPad Prism 7 software (San Diego, CA, USA). Significant differences in the data between groups were tested by Student's t-test at 5% level using SPSS 19.0 software (IBM, Chicago, IL, USA).

Phenotypic Variations between IMF-Preadipocytes and IMF-Adipocytes
To visualize the difference of adipogenesis between IMF-preadipocytes and IMF-adipocytes, the morphological characteristics of chicken IMF-preadipocytes and IMF-adipocytes are shown in Figure 1. At 0 day, IMF-preadipocytes we captured were shown in presence ( Figure 1A) or absence of Oil Red O staining ( Figure 1B). At 10 days after differentiation induction, IMF-preadipocytes were fully differentiated into IMF-adipocytes filled with lots of lipid droplets ( Figure 1C,D). Triacylglycerol content in the IMF-adipocytes was significantly higher than that in IMF-preadipocytes ( Figure 1E). To visualize the difference of adipogenesis between IMF-preadipocytes and IMF-adipocytes, the morphological characteristics of chicken IMF-preadipocytes and IMF-adipocytes are shown in Figure  1. At 0 day, IMF-preadipocytes we captured were shown in presence ( Figure 1A) or absence of Oil Red O staining ( Figure 1B). At 10 days after differentiation induction, IMF-preadipocytes were fully differentiated into IMF-adipocytes filled with lots of lipid droplets ( Figure 1C-D). Triacylglycerol content in the IMF-adipocytes was significantly higher than that in IMF-preadipocytes ( Figure 1E).

Summary of Sequencing Data
Four small RNA libraries were constructed as follows: IMF-Preadipocyte 2-1, IMF-Preadipocyte 2-2, IMF-Adipocyte 1, and IMF-Adipocyte 2, with two replicates for each treatment. And then were sequenced separately using an Illumina Genome Analyzer System. After discarding low-quality reads and masking adaptor sequences, clean reads with 21 to 24 nt in length were obtained, with the vast majority being 22 nt in length, which was consistent with the common size of miRNAs ( Figure  2A). After quality control and adaptor removal, a total of 19,178,226; 19,786,345; 20,109,174; and 19,530,973 clean reads were obtained from four libraries respectively. The reads count of each library were about 20 M with more than 90% of Q20 score. The percentage of the 22 nt reads of the total were 37.68%, 38%, 48.83% and 49.56% for the four libraries. The high-quality reads were subsequently annotated to different categories of RNA (identified miRbase, repeats-associated RNA, rRNA, tRNA, snRNA, snoRNA, etc) using databases such as miRBase (V19.0) [35] and Genbank (Gallus_gallus-5.0). The most abundant RNA species (based on read count) in the four libraries was classified as miRNAs, accounting for 71.58%, 56.53%, 72.41%, and 57.06% in the four libraries. This indicated that the deep sequencing data were highly enriched for mature miRNA sequences and well suitable for subsequent profiling analysis. The second most abundant category was rRNAs, accounting for 4.36%, 8.34%, 3.89%, and 14.02% in the four libraries, respectively. In addition, unknown RNAs also represented a high percentage (17.82%, 29.08%, 17.65%, and 24.13%, respectively). Finally, all reads were aligned against the chicken genome using the program miRCat [34].

Summary of Sequencing Data
Four small RNA libraries were constructed as follows: IMF-Preadipocyte 2-1, IMF-Preadipocyte 2-2, IMF-Adipocyte 1, and IMF-Adipocyte 2, with two replicates for each treatment. And then were sequenced separately using an Illumina Genome Analyzer System. After discarding low-quality reads and masking adaptor sequences, clean reads with 21 to 24 nt in length were obtained, with the vast majority being 22 nt in length, which was consistent with the common size of miRNAs ( Figure 2A). After quality control and adaptor removal, a total of 19,178,226; 19,786,345; 20,109,174; and 19,530,973 clean reads were obtained from four libraries respectively. The reads count of each library were about 20 M with more than 90% of Q20 score. The percentage of the 22 nt reads of the total were 37.68%, 38%, 48.83% and 49.56% for the four libraries. The high-quality reads were subsequently annotated to different categories of RNA (identified miRbase, repeats-associated RNA, rRNA, tRNA, snRNA, snoRNA, etc.) using databases such as miRBase (V19.0) [35] and Genbank (Gallus_gallus-5.0). The most abundant RNA species (based on read count) in the four libraries was classified as miRNAs, accounting for 71.58%, 56.53%, 72.41%, and 57.06% in the four libraries. This indicated that the deep sequencing data were highly enriched for mature miRNA sequences and well suitable for subsequent profiling analysis. The second most abundant category was rRNAs, accounting for 4.36%, 8.34%, 3.89%, and 14.02% in the four libraries, respectively. In addition, unknown RNAs also represented a high percentage (17.82%, 29.08%, 17.65%, and 24.13%, respectively). Finally, all reads were aligned against the chicken genome using the program miRCat [34].

Screening for Differentially Expressed miRNAs
To identify the major miRNAs involved in intramuscular fat deposition in chicken, differentially expressed miRNAs were screened between the IMF-preadipocytes and IMF-adipocytes groups. In total, 117 differential miRNAs were identified, including 91 known miRNAs and 26 novel miRNAs (Table S1). Of these, 95 were up-regulated and 22 were down-regulated miRNAs in IMF-adipocytes relative to IMF-preadipocytes ( Figure 2B,C). Furthermore, clustering analysis indicated that high similarity was shown within two replicates for each group ( Figure 2D

Screening for Differentially Expressed miRNAs
To identify the major miRNAs involved in intramuscular fat deposition in chicken, differentially expressed miRNAs were screened between the IMF-preadipocytes and IMF-adipocytes groups. In total, 117 differential miRNAs were identified, including 91 known miRNAs and 26 novel miRNAs (Table S1). Of these, 95 were up-regulated and 22 were down-regulated miRNAs in IMF-adipocytes relative to IMF-preadipocytes ( Figure 2B,C). Furthermore, clustering analysis indicated that high similarity was shown within two replicates for each group ( Figure 2D). To double check the results of miRNA-seq, nine differentially expressed miRNAs, including eight down-regulated and one up-regulated, were randomly selected for qPCR assay. Figure 2E showed the result of relative expression level from qPCR, which was consistent with the miRNA-Seq data. It indicated that identification and abundance estimation of miRNAs were reliable in this study.

Target Gene Prediction, Functional Enrichment Analysis, and Candidate Genes Identification
To better understand the potential miRNAs responsible for adipogenesis in chicken, target prediction of differentially expressed miRNAs and functional enrichment analysis were conducted. A total of 4005 target genes were identified by 117 differentially expressed miRNAs. All target genes were subjected to enrichment analysis of functions and signaling pathways. Biological processes significantly enriched in adipogenesis-related terms, including regulation of cell differentiation (GO:0045595), cellular lipid metabolic process (GO:0044255) and lipid metabolic process (GO:0006629) (Table S2). Furthermore, many lipid-related pathways were also enriched by target genes, such as fatty acid metabolism (gga01212), fatty acid elongation (gga00062), citrate cycle (TCA cycle) (gga00020) and fatty acid degradation (gga00071) ( Figure 3A). Eight pathways related to lipid metabolism were focuse ( Figure 3B and Table 2). Of these, "fatty acid elongation" was significantly enriched by the ELOVL gene family, HACD gene family, ACOT7, etc. Lots of published data has proven that the ELOVL and HACD gene families catalyzed the reactions of the long-chain fatty acids elongation cycle [35][36][37]. Additionally, it has already been reported that the ACOT gene family, such as ACOT7 and ACOT13, played a key role in the regulation of hepatic lipid metabolism [38,39].

Expression of gga-miR-18b-3p and ACOT13 in Preadipocytes and Adipocytes
To illustrate the homology of miR-18b-3p in various species, cluster analysis was performed, which indicated that chicken miR-18b-3p was highly identical with alpaca and Capra hircas ( Figure 4A). Besides, the seed region of the mature miR-18b-3p sequence was highly conserved in mouse, reptile, chicken, goat, etc. ( Figure 4B). A potential binding site for miR-18b-3p was identified in the 3 UTR region of the ACOT13 gene ( Figure 4C). Furthermore, the expression pattern of gga-miR-18b-3p and its putative target ACOT13 were detected between pre-differentiation (0 day) and post-differentiation (10 days). As shown in Figure 4D, oppose expression pattern of miR-18b-3p and ACOT13 gene was observed at 10 days post-differentiation ( Figure 4D). It indicated that there was a negative correlation between the expression of miR-18b-3p and the ACOT13 gene.

miR-18b-3p Functioned as an Inhibitor during Preadipocytes Differentiation by Targeting the ACOT13 Gene
To investigate the regulatory role of gga-miR-18b-3p on preadipocyte differentiation, an over-expression experiment was conducted by transfecting gga-miR-18b-3p mimics into chicken preadipocytes. Compared with negative control, the expression level of gga-miR-18b-3p was significantly increasing in the overexpression group ( Figure 5A). Expectedly, the expression levels of ACOT13, PPARG, and FABP4 were significantly decreased in the gga-miR-18b-3p overexpressing group relative to negative control. The differentiation of preadipocytes was determined using Oil Red O staining and intracellular triglycerides concentration. Evidence from Oil Red O staining suggested that the number of lipid droplets was decreased in the overexpression group compared with the control group ( Figure 5B). Furthermore, the concentration of intracellular triglyceride in the gga-miR-18b-3p overexpressing group was significantly lower than that of the negative control group ( Figure 5C). All of the above results demonstrated that gga-miR-18b-3p functioned as an inhibitor for preadipocytes differentiation.

miR-18b-3p Functioned as an Inhibitor during Preadipocytes Differentiation by Targeting the ACOT13 Gene
To investigate the regulatory role of gga-miR-18b-3p on preadipocyte differentiation, an overexpression experiment was conducted by transfecting gga-miR-18b-3p mimics into chicken preadipocytes. Compared with negative control, the expression level of gga-miR-18b-3p was significantly increasing in the overexpression group ( Figure 5A). Expectedly, the expression levels of ACOT13, PPARG, and FABP4 were significantly decreased in the gga-miR-18b-3p overexpressing group relative to negative control. The differentiation of preadipocytes was determined using Oil Red O staining and intracellular triglycerides concentration. Evidence from Oil Red O staining suggested that the number of lipid droplets was decreased in the overexpression group compared with the control group ( Figure 5B). Furthermore, the concentration of intracellular triglyceride in the gga-miR-18b-3p overexpressing group was significantly lower than that of the negative control group ( Figure 5C). All of the above results demonstrated that gga-miR-18b-3p functioned as an inhibitor for preadipocytes differentiation.
In order to confirm the gga-miR-18b-3p target on ACOT13, a plasmid-containing wild-type or mutant target site of the 3′UTR region of the ACOT13 gene was transfected into DF1 cells together with gga-miR-18b-3p or negative control mimics. gga-miR-18b-3p mimics significantly decreased the luciferase activity of ACOT13-3′UTR-WT, but it did not affect the activity of ACOT13-3′UTR-MUT ( Figure 5D). This result indicated that gga-miR-18b-3p could directly bind on the 3′UTR region of the ACOT13 gene. Taken together, these results suggested that miR-18b-3p inhibited the preadipocytes differentiation by down-regulating the ACOT13 gene. In order to confirm the gga-miR-18b-3p target on ACOT13, a plasmid-containing wild-type or mutant target site of the 3 UTR region of the ACOT13 gene was transfected into DF1 cells together with gga-miR-18b-3p or negative control mimics. gga-miR-18b-3p mimics significantly decreased the luciferase activity of ACOT13-3 UTR-WT, but it did not affect the activity of ACOT13-3 UTR-MUT ( Figure 5D). This result indicated that gga-miR-18b-3p could directly bind on the 3 UTR region of the ACOT13 gene. Taken together, these results suggested that miR-18b-3p inhibited the preadipocytes differentiation by down-regulating the ACOT13 gene.

ACOT13 Promoted Preadipocyte Differentiation by Fatty Acid β-Oxidation and Lipogenic Transcription
To elucidate the function of ACOT13 on chicken preadipocyte differentiation, overexpression, knockdown and rescue experiment of the ACOT13 gene in chicken intramuscular preadipocytes were designed. After transfection for 48 hours, expression of marker genes that involved in fatty acid β-oxidation and adipogenesis were investigated. Meanwhile, the content of intracellular triglyceride was measured.
First, overexpression of the ACOT13 gene in chicken intramuscular preadipocytes. Compared with the negative control group, the expression level of ACOT13 significantly increased ( Figure 6A). The marker genes related to fatty acid β-oxidation, including ACOX1, ACOX3, and ATGL, were significantly higher in the ACOT13 overexpressing group relative to the control group ( Figure 6B). Meanwhile, marker genes involved in adipogenesis, including PPARG and FABP4, also presented a higher expression level in the ACOT13 overexpressing group relative to the control ( Figure 6B). Results from Oil Red O staining and content of intracellular triglyceride suggested that the accumulation of lipid droplets significantly increased in the ACOT13 overexpressing group compared with the control (Figure 6C,D). Second, knockdown of ACOT13 in chicken intramuscular preadipocytes. As we expected, the expression level of ACOT13 and marker genes involved in fatty acid β-oxidation (ACOX3 and ATGL) and adipogenesis (PPARA, PPARD, CEBPA, and FABP4) were significantly lower in ACOT13 knockdown group than that in the control group ( Figure 7A,B). Furthermore, evidence from Oil Red O staining and intracellular triglyceride assay demonstrated that lipid droplet accumulation and Second, knockdown of ACOT13 in chicken intramuscular preadipocytes. As we expected, the expression level of ACOT13 and marker genes involved in fatty acid β-oxidation (ACOX3 and ATGL) and adipogenesis (PPARA, PPARD, CEBPA, and FABP4) were significantly lower in ACOT13 knockdown group than that in the control group ( Figure 7A,B). Furthermore, evidence from Oil Red O staining and intracellular triglyceride assay demonstrated that lipid droplet accumulation and intracellular triglyceride content significantly declined in the ACOT13 knockdown group (Figure 7C,D). Third, co-transfection of gga-miR-18b-3p mimicswith ACOT13 overexpression plasmid was performed to further verify the effects of gga-miR-18b-3p and ACOT13 on preadipocyte differentiation. The treatment of negative control mimics together with the negative control plasmid was also conducted. The results showed that there was no significance in both lipid droplet accumulation and intracellular triglyceride concentration between the two treatments ( Figure 8A,B). Third, co-transfection of gga-miR-18b-3p mimicswith ACOT13 overexpression plasmid was performed to further verify the effects of gga-miR-18b-3p and ACOT13 on preadipocyte differentiation. The treatment of negative control mimics together with the negative control plasmid was also conducted. The results showed that there was no significance in both lipid droplet accumulation and intracellular triglyceride concentration between the two treatments ( Figure 8A Taken together, overexpression of ACOT13 promoted preadipocyte differentiation, fatty acid βoxidation, and adipogenesis. Conversely, knockdown of ACOT13 would inhibit these processes. Moreover, the promoting effect of ACOT13 on preadipocyte differentiation, fatty acid β-oxidation and adipogenesis can be eased by the inhibitory effect of gga-miR-18b-3p. All the results demonstrated that gga-miR-18b-3p targeting ACOT13 inhibited intramuscular adipocytes differentiation.

Discussion
In recent years, understanding the mechanisms underlying the process of adipocyte differentiation has vastly expanded. Several potential miRNAs in our miRNA-seq data have been reported to be associated with cell differentiation, such as miR-221, miR-146, miR-30-5p, miR-301, and miR-148 [38]. Some of them played pivotal roles in the differentiation of adipocytes [39,40], which was consistent with our expected results, that we could obtain the differentially expressed miRNAs related to adipocyte differentiation.
KEGG pathway analysis was performed to investigate the biological interpretation of target genes for differentially expressed miRNAs. Three important pathways involved in fatty acid metabolism were excessively enriched, including the PPAR signaling pathway (gga03320), fatty acid metabolism (gga01212), and fatty acid degradation (gga00071). The PPAR signaling pathway is a very important intermediate metabolic pathway. Fatty acids are able to activate the expression of genes via PPARs [41]. PPARs are transcription factors that are ligand-activated, and they belong to the superfamily of nuclear hormone receptors. PPARα, PPARβ/δ, and PPARγ are included in PPARs. These three transcription factors act on their target genes and participate in lipid metabolism and adipocyte differentiation. Recent research found that PPARs had an influence on adipocyte differentiation [42,43] and fatty acid oxidation [44,45]. The marker gene, FABP [46], has been proven to have the same function in the PPARs signaling pathway. In the pathway of fatty acid metabolism and fatty acid degradation, ACOX1 [47] is also involved in fatty acid oxidation, while FASN [48] and ELOVL5 [49] are related to fatty acid biosynthesize.
Recent studies have found that miR-18b was considered as important marker of cell proliferation and cell adhesion. Murakami Y et al. found that the expression level of miR-18b was significantly associated with histological differentiation. Poorly differentiated hepatocellular carcinoma (HCC) Taken together, overexpression of ACOT13 promoted preadipocyte differentiation, fatty acid β-oxidation, and adipogenesis. Conversely, knockdown of ACOT13 would inhibit these processes. Moreover, the promoting effect of ACOT13 on preadipocyte differentiation, fatty acid β-oxidation and adipogenesis can be eased by the inhibitory effect of gga-miR-18b-3p. All the results demonstrated that gga-miR-18b-3p targeting ACOT13 inhibited intramuscular adipocytes differentiation.

Discussion
In recent years, understanding the mechanisms underlying the process of adipocyte differentiation has vastly expanded. Several potential miRNAs in our miRNA-seq data have been reported to be associated with cell differentiation, such as miR-221, miR-146, miR-30-5p, miR-301, and miR-148 [38]. Some of them played pivotal roles in the differentiation of adipocytes [39,40], which was consistent with our expected results, that we could obtain the differentially expressed miRNAs related to adipocyte differentiation.
KEGG pathway analysis was performed to investigate the biological interpretation of target genes for differentially expressed miRNAs. Three important pathways involved in fatty acid metabolism were excessively enriched, including the PPAR signaling pathway (gga03320), fatty acid metabolism (gga01212), and fatty acid degradation (gga00071). The PPAR signaling pathway is a very important intermediate metabolic pathway. Fatty acids are able to activate the expression of genes via PPARs [41]. PPARs are transcription factors that are ligand-activated, and they belong to the superfamily of nuclear hormone receptors. PPARα, PPARβ/δ, and PPARγ are included in PPARs. These three transcription factors act on their target genes and participate in lipid metabolism and adipocyte differentiation. Recent research found that PPARs had an influence on adipocyte differentiation [42,43] and fatty acid oxidation [44,45]. The marker gene, FABP [46], has been proven to have the same function in the PPARs signaling pathway. In the pathway of fatty acid metabolism and fatty acid degradation, ACOX1 [47] is also involved in fatty acid oxidation, while FASN [48] and ELOVL5 [49] are related to fatty acid biosynthesize.
Recent studies have found that miR-18b was considered as important marker of cell proliferation and cell adhesion. Murakami Y et al. found that the expression level of miR-18b was significantly associated with histological differentiation. Poorly differentiated hepatocellular carcinoma (HCC) presented high miR-18b and low TNRC6B (trinucleotide repeat containing 6B) expression levels. The mechanism underlying was that miR-18b inhibited HCC differentiation via targeting the TNRC6B gene [50]. Namløs HM et al. also reported that the expression of miR-18b in osteoblast and normal bone was identical with the differentiation level of osteosarcoma. From osteoblasts to osteosarcoma, then to bone, the degree of cell differentiation gradually increased. The higher the degree of cell differentiation, the lower the expression of miR-18b [51]. Both above studies suggested that miR-18b might function as a crucial factor in cell differentiation. Therefore, we concluded that miR-18b participated in the cell differentiation process. At the same time, miR-18b-3p was also a differential expressing miRNA between IMF-preadipocytes and IMF-adipocytes. We inferred that miR-18b-3p should be involved in the process of IMF preadipocytes. Meanwhile ACOT13 was one of the differential genes in IMF-preadipocytes and IMF-adipocytes (unpublished data). Prediction of miRNA target genes suggested ACOT13 was the target of gga-miR-18b-3p.
Acyl-CoA thioesterase 13 (ACOT13), a member of the acyl-CoA thioesterase (ACOT) gene family, catalyzed the key steps in fatty acid biosynthesis, bioluminescence, and non-ribosomal peptide synthesis [52]. ACOT13 abundantly expressed in oxidative tissues (e.g., liver, adipose tissue) [53] and co-localized with mitochondria and long-chain fatty acyl-CoA substrates [54]. ACOT gene family could catalyze the hydrolysis of fatty acyl-CoAs to form free fatty acids (FFAs) and coenzyme A (CoASH). Indeed, the balance between FFAs and fatty acyl-CoAs within cells regulated fatty acid oxidation and controlled the synthesis of complex lipids [55,56]. Hung J Y et al. found that ACOT11 and ACOT13 knockdown-mediated growth inhibition could be rescued by the addition of fatty acids [57]. Nicholls previous studies have revealed that activated ACOT13 at the mitochondrial membrane could result in increased fatty acid oxidation [58]. All these findings showed that ACOT13 took part in lipid metabolism and fatty acid β-oxidation. In this study, results from overexpression and knockdown of the ACOT13 gene demonstrated that ACOT13 played an important role in fatty acid β-oxidation and adipogenesis, which were all consistent with previous studies.
Zhang et al. reported that miR-140-5p affected IMF deposition in chicken with the association with meat quality [4]. However, gga-miR-140-5p was not the top miRNA in our present study. The differentially expressed miRNAs in Zhang's study were detected between 20 and 55 weeks in Gushi chicken, so the difference might be not only caused by the different IMF deposition, but also by different stages between groups. As a short-chain non-coding small RNA, the expression of miRNAs in different tissues, and stages were different. In addition, miRNAs have coordinated biological functions by targeting genes, meanwhile, the target interactions were different in different tissues and stages [59,60]. Although miR-140-5p had potential effects on cell growth, migration, and death [61] in many species, such as pig [62], human, and chicken [63,64], many gene were involved in the process. Therefore, the direct effect between preadipocytes and adipocytes in our study might not be caused by miR-140-5p.
In our study, we identified a candidate miRNA, gga-miR-18b-3p, involved in the adipocyte differentiation based on the results of miRNA-seq. Bioinformatics analysis and functional assay suggested that gga-miR-18b-3p inhibited the differentiation of chicken intramuscular preadipocytes via targeting the ACOT13 gene.

Conclusions
In conclusion, our study identified that miRNAs were associated with chicken preadipocyte differentiation between IMF-preadipocytes and IMF-preadipocytes. This study demonstrated that gga-miR-18b-3p could inhibit the intramuscular adipocytes differentiation by negatively regulating ACOT13 (Figure 9). These findings may provide fundamental basis for improving poultry meat quality, and it will also be helpful for a deeper understanding of chicken IMF deposition.   Figure 9. Proposed model of gga-miR-18b-3p regulation on intramuscular adipocyte differentiation.

Conflicts of Interest:
The authors declare no conflict of interest.