HMGR overexpression and interference affect the expression of steroidogenic genes and cholesterol content in bovine intramuscular adipocytes

Previously, we found that mevalonic acid stimulates 3-hydroxy-3-methyl-glutaryl-coenzyme A reductase (HMGR) expression in bovine intramuscular adipocytes to influence adipocyte differentiation. However, any direct links among HMGR, steroidogenic genes, and cholesterol content remain unclear. RNA-Seq was conducted to determine the differences between the gene expression profiles of bovine adipocytes containing different HMGR expression constructs. In total, 10,234 differentially expressed genes (DEGs) were found. Of these, 35 and 6 DEGs between the control and the overexpression groups were functionally related to lipid and energy metabolism, respectively. In addition, 43 and 8 DEGs between the control and the HMGR inhibition groups were related to lipid and energy metabolism, respectively. Several DEGs related to lipid and energy metabolism were also identified between the HMGR overexpression group and the HMGR interference group, and many DEGs were correlated positively or negatively with the overexpression or inhibition of HMGR. We also found that, following the activation or inhibition of the HMGR gene, AMP-activated protein kinase (AMPK) and sirtuin type 1 (SIRT1) had opposite expression patterns in bovine intramuscular adipocytes. Interestingly, the HMGR gene was downregulated when HMGR was overexpressed, and upregulated when HMGR was inhibited. Our findings establish a theoretical understanding of signaling pathways involved in cholesterol synthesis by elucidating the relationships between key genes.


Scientific RepoRtS
| (2020) 10:16606 | https://doi.org/10.1038/s41598-020-73626-8 www.nature.com/scientificreports/ supplementation can improve lipid disorders and decrease fat accumulation by regulating the SIRT1-autophagy pathway 12 . However, it remains unclear whether there is any direct link among HMGR, steroidogenic genes, and cholesterol content. Cholesterol content is a complex genetic trait. The genetics underlying the biochemical process and molecular background that determine cholesterol content are not yet fully understood, particularly with regard to local Chinese cattle breeds. Using RNA-Seq analysis to select differently expressed genes (DEGs) correlated with cholesterol levels, the present study aimed to explore the effects of HMGR overexpression and interference on the expression patterns of cholesterol-related genes, as well as on differential gene expression in bovine intramuscular adipose cells.

Results
Adipose cell morphology and lipid accumulation. We first evaluated morphological changes in induced adipogenic bovine intramuscular adipocytes, and then stained cells with Oil Red O to visualize lipid droplet accumulation after induction. As shown in Fig. 1A, B, induced adipocytes were large, round, and filled with fat droplets, compared to immature adipocytes. DEG detection. We sequenced RNA from the five experimental groups using the Illumina HiSeq platform, and generated an average of 6.58 Gb for each sample. As showed in Fig. 2, after mapping sequenced reads to the reference genome and reconstructing the transcripts, we obtained 13,264 novel transcripts from all samples. Of these, 9444 were previously unknown splicing events for known genes, 862 were novel coding transcripts without any known features, and the remaining 2958 were long noncoding RNAs.   www.nature.com/scientificreports/ We then analyzed differential gene expression between samples. The summary of DEGs is shown in Fig. 2. There were a total of 10,234 DEGs, and this number was higher when the HMGR gene was overexpressed or suppressed. There were 1554 DEGs found in the HMGR overexpression group versus the HMGR interference group, where genes were mostly upregulated. We also found that there were more upregulated genes within DEGs between all experimental groups (Supplementary table S1; Supplementary Figure S1). Pathway analysis of DEGs. We performed KEGG pathway classification and functional enrichment analyses for DEGs (Fig. 3). These DEGs were found to be mostly involved in signal transduction, cancers, infectious disease, global and overview maps, and the immune system. As shown in Fig. 3A, B, 35 DEGs after HMGR overexpression and 43 DEGs after HMGR inhibition, were functionally involved in lipid metabolism. Furthermore, 6 DEGs after HMGR overexpression and 8 DEGs after HMGR inhibition, were functionally involved in energy metabolism.
AMPK signaling. Changes in gene expression relevant to the AMPK signaling pathway following the overexpression and interference of the HMGR gene are shown in Fig. 4A, B. Interestingly, the HMGR gene was downregulated when HMGR was overexpressed, whereas the SIRT1 gene was upregulated. In contrast, the HMGR gene itself was upregulated following HMGR interference.

TC, LDL-C, and HDL-C content in adipose cells.
To understand the impact of HMGR overexpression on the amount of TC, LDL-C, and HDL-C, we examined the levels of these types of cholesterol in the five experimental groups. In contrast with other groups, the HMGR overexpression group had the highest amounts of TC, LDL-C, and HDL-C, particularly for TC and HDL-C (Fig. 5).
qRT-PCR of HMGR, AMPK, SIRT1, and LPL. To investigate the relationship between HMGR expression and cholesterol synthesis in bovine adipose cells, the expression levels of HMGR, AMPK, SIRT1, and LPL, which are four steroidogenesis-related genes, were assessed in adipose tissue using qRT-PCR (Fig. 6).
For HMGR, the overexpression group showed higher mRNA expression levels than those in other groups, with the HMGR interference group showing the lowest levels out of all groups. AMPK mRNA expression was the lowest when HMGR was overexpressed; even though AMPK protein expression levels were increased (as shown in Fig. 7). SIRT1 showed similar results as AMPK, whereas LPL expression levels were higher when HMGR was overexpressed and lower when HMGR was inhibited.
Western blotting of HMGR, AMPK, SIRT1, and LPL. To investigate changes in protein expression, we performed western blotting in the five experimental groups (Fig. 7). The results showed that HMGR protein expression in the HMGR overexpression group was higher than that in the other four groups. AMPK protein www.nature.com/scientificreports/ expression levels showed little differences between the groups. SIRT-1 and LPL protein expression levels were higher in the HMGR interference group and lower in the HMGR overexpression group.

Discussion
Adipose tissue is the major depot for energy and cholesterol storage, with most of the intracellular cholesterol distributed in the form of lipid droplets 13,14 . Cholesterol homeostasis may have a role in the regulation of adipocyte size and function 15,16 . In addition, cellular cholesterol homeostasis is involved in various diseases of nongenetic origin, such as atherosclerosis, leading to lipid accumulation in target organs 15,17 .
In the present study, DEG and KEGG pathway analyses were performed to explore the regulatory network underlying bovine adipose cells following HMGR overexpression and interference. As expected, several wellknown pathways 18 related to lipid metabolism were identified, particularly the AMPK signaling pathway. In  www.nature.com/scientificreports/ AMPK signaling, the HMGR, AMPK, and SITR1 genes play a functional role in lipid metabolism or cholesterol biosynthesis.
In terms of cholesterol levels, we found that TC, LDL-C, and HDL-C levels were the highest when the HMGR gene was overexpressed; however, their levels were not majorly decreased when HMGR was inhibited. When we investigated the protein expression of HMGR, AMPK, SIRT-1, and LPL, we found that HMGR protein levels were higher in the HMGR overexpression group compared to the control group. However, SIRT-1 protein levels were lower in the HMGR overexpression group compared to those in the HMGR interference group. Therefore, the results obtained for protein expression were consistent with those obtained for gene expression.
Given that HMGR is a rate-limiting enzyme in cholesterol biosynthesis 10 , we generated an adipocyte gene expression dataset using RNA-Seq, and examined the relationship between steroidogenic genes and cholesterol levels by overexpressing or inhibiting the HMGR gene. Transcriptional and pathway analysis results showed that the overexpression of HMGR was correlated with the downregulation of AMPK and SIRT1 gene expression.
AMPK is a crucial energy sensor that maintains energy homeostasis. It is also a major regulator of the lipid metabolism through the phosphorylation and inactivation of numerous metabolic enzymes, including HMGR 19,20 . In adipose tissue, the chronic activation of AMPK led to the downregulation of HMGR expression in a new transgenic mouse model 21 . The 5-aminoimidazole-4-carboxamide (AICA) ribonucleotide was shown to induce the activation of AMPK and directly inhibit the expression of HMGR, one of its target genes 19 . In the present study, HMGR mRNA expression was reduced when it was overexpressed in bovine adipose cells. In contrast, HMGR expression was increased when the HMGR gene was inhibited. These results are important, given the key role of the HMGR gene in the process of cholesterol synthesis.
SIRT1 is a longevity-associated deacetylase enzyme that modulates metabolic homeostasis in response to cellular energy. AMPK and SIRT1 are related proteins and share common target pathways 22 . The present study showed that the expression of SIRT1 decreased when the HMGR gene was overexpressed. However, SIRT1 was upregulated when HMGR expression was modified. These results imply that the HMGR and SIRT1 genes have contrasting roles in the cholesterol synthesis pathway.
The LPL gene encodes a rate-limiting enzyme that has a key role in the hydrolysis of triglycerides. LPL deficiency and dysfunction is associated with many disorders of lipoprotein metabolism 23,24 . The results of the present study showed that the patterns of LPL gene expression were consistent with those of HMGR expression.  siRNA transfection. We selected four pairs of HMGR siRNAs (siRNA1, siRNA2, siRNA3, and siRNA4) for the transfection experiment, the fluorescence experiment, and the quantitative experiment to detect the transfection efficiency. The transfection efficiencies of siRNA1, siRNA2, siRNA3 and siRNA4 were found to be 46.8%, 63.8%, 46.8% and 19.1, respectively. Therefore, we selected the siRNA2 for further experiments, and the sequences of its primers as follows: sense, 5′-GUU CUA ACU CAC AGG AUG ATT-3′; and antisense, 5′-UCA UCC UGU GAG UUA GAA CTT-3′. The siRNA2 oligonucleotides were designed by Beijing Dingguo Changsheng Biotechnology Co. Ltd.. The siRNA2 inhibition group and siRNA2 negative control group were transfected for 48 h using the lipofectamine transfection reagent (Invitrogen), according to the manufacturer's protocol.
Detection of DEGs. The generation and processing of raw RNA-seq data were conducted by the Beijing Genomics Institute using the Illumina HiSeq platform. We detected DEGs using the PossionDis algorithm, which is based on the Poisson distribution, as described by Audic et al. 26 . The parameters to determine significant changes in gene expression were as follows: fold change ≥ 2 and false discovery rate (FDR) ≤ 0.001.

Pathway analysis of DEGs. DEGs were annotated according to the Kyoto Encyclopedia of Genes and
Genomes (KEGG) database. We performed pathway enrichment analysis using phyper, a function in R based on hypergeometric distribution 27,28 . We then calculated the FDR for each p-value, and considered an FDR < 0.001 to represent significant enrichment.
Statistical analysis. Differences in gene expression between the experimental groups were calculated by analysis of variance (ANOVA), followed by the Bonferroni test. Results were considered statistically significant at P < 0.05.

Data availability
The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request.