Circulating miR-3659 may be a potential biomarker of dyslipidemia in patients with obesity

Background The present study attempted to identify potential key genes and miRNAs of dyslipidemia in obese, and to investigate the possible mechanisms associated with them. Methods The microarray data of GSE66676 were downloaded, including 67 obese samples from the Gene Expression Omnibus (GEO) database. The weighted gene co-expression network (WGCNA) analysis was performed using WGCNA package and grey60 module was considered as the highest correlation. Gene Ontology annotation and the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses for this module were performed by clusterProfiler and DOSE package. A protein–protein interaction (PPI) network was established using Cytoscape software, and significant modules were analyzed using molecular complex detection. Results Collagen type I alpha 1 chain gene (COL1A1) had the best significant meaning. After bioinformatic analysis, we identified four miRNAs (hsa-miR-3659, hsa-miR-4658, hsa-miR151a-5p and hsa-miR-151b) which can bind SNPs in 3′UTR in COL1A1. After validation with RT-qPCR, only two miRNAs (hsa-miR-3659 and hsa-miR151a-5p) had statistical significance. Conclusions The area of 0.806 for miR-3659 and 0.769 for miR-151a-5p under the ROC curve (AUC) may have good diagnostic value for dyslipidemia. Circulating miR-3659 may be a potential biomarker of dyslipidemia in patients with obesity. Electronic supplementary material The online version of this article (10.1186/s12967-019-1776-8) contains supplementary material, which is available to authorized users.


Background
Obesity, especially abdominal obesity, is the key reason to result in metabolic syndrome (MetS), which refers to insulin resistance, type 2 diabetes mellitus, hypertension, and dyslipidemia, and all above risk factors finally lead to cardiovascular disease [1,2]. A recent study showed that about 2.2 billion people were overweight or obese in 2015 [3]. As a complex and multifactorial disease, lots of environmental and genetic factors can result in this disorder [4,5].
MicroRNA (miRNA), a class of non-coding RNA molecules (~ 22 nucleotides), is short and highly conserved. When it dysregulated, lots of human diseases would be happened [6]. MiRNAs mediate post-transcriptional regulation of protein-coding genes by complementary binding to the 3′ untranslated region (3′UTR) and occasionally to the 5′UTR or coding regions of target mRNAs [7]. Previous study has shown that single nucleotide polymorphisms (SNPs) in the miRNA regulatory networks were a novel class of functional variants in the human genome. Genetic variants that potentially influence miRNA-mediated cellular function may be classified in two major categories: SNPs affecting miRNA biogenesis and SNPs in the miRNA targetome [8,9]. Early detection and treatment contribute to a promising effect. Since the discovery of circulating miRNAs in body fluids, an increasing number of studies have focused on their potential and non-invasive biomarkers, and therapeutic targets or tools for many diseases [10]. Martino et al. found that both miR-33a and miR-33b were early biomarkers for cholesterol levels in childhood [11]. Besides, Iacomino et al. had further found about the role of miRNAs in obesity and related metabolic abnormalities [12]. In this study, we performed the integrated bioinformatic methods to construct the co-expression network and mark significant miRNA. The outcomes may help us for further elucidating the innate character of dyslipidemia, and provide new insights to potential biomarkers and signaling pathways to treat dyslipidemia in obese.

Microarray data
Microarray data of GSE66676 [13] were downloaded from the National Center for Biotechnology Information (NCBI) Gene Expression Omnibus (GEO, http://www. ncbi.nlm.nih.gov/geo/) database [14]. GSE66676 contains 67 Liver wedge biopsy samples from obesity patients (mean age = 16.88, mean body mass index, BMI = 52.01). The CEL files were transformed into the expression value matrix using the Affy package in R [15], and the probe information was then transformed into the gene name using Bioconductor in R [16]. If one gene had more than one probe, the mean expression value of this gene was selected. The specific workflow is shown in Fig. 1.

Construction of weighted gene co-expression network
The weighted gene co-expression network (WGCNA) is a widely used systems biology method, which is used to construct a scale-free network from gene expression data [17]. An appropriate soft threshold power (soft power = 6) was selected in accordance with standard scale-free networks, with which adjacencies between all differential genes were calculated by a power function [18]. The rest of the analysis strategy can refer to our previous research [19].

Finding module of interest and functional annotation
The correlation between modules and clinical features was evaluated by Pearson's correlation tests to search biologically meaningful modules. The module and clinical feature, which exhibited the highest correlation, were selected as module of interest and clinical feature to be studied. All genes of module of interest were analyzed by Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway by clusterProfiler and DOSE package in R [20]. A P value < 0.01 and false discovery rate (FDR) < 0.05 were set as the cutoff criteria.

Hub gene analysis
The module membership (MM) was defined as the correlation of gene expression profile with module eigengenes (Mes). And the gene significance (GS) measure was defined as (the absolute value of ) the correlation between gene and external traits. Genes with highest MM and highest GS in modules of interest were natural candidates for further research [21]. Thus, the intramodular hub genes were chosen by external traits based GS > 0.2 and MM > 0.6 with a threshold of P-value < 0.05 [17]. The gene-gene interaction network was constructed and visualized using Cytoscape software package [22] and molecular complex detection (MCODE) [23] was used to analyze the most notable clustering module. MCODE score > 6 was a threshold for next analysis.

Subjects
All of the two groups of study population were obese, including 424 unrelated participants of normals (128   [24]. The participants' age ranged from 18 to 80 years with a mean age of 55.31 ± 10.52 years in normal and 55.87 ± 11.13 years in dyslipidemic groups; respectively. The gender ratio and age distribution were matched between the two groups. All participants were essentially healthy with no history of coronary artery disease, stroke, diabetes, hyper-or hypo-thyroids, and chronic renal disease. They were free from medications known to affect lipid profiles.

Epidemiological survey
The epidemiological survey was carried out using internationally standardized method, following a common protocol [25]. Cigarette smoking status was categorized into groups of cigarettes smoker and non-smoker. Alcohol consumption was categorized into groups of alcohol drinker and non-drinker [26]. Several parameters such as blood pressure, height, weight and waist circumference (WC) were measured, while BMI (kg/m 2 ) was calculated.

Biochemical measurements
Venous blood samples were obtained from all subjects after at least 12 h of fasting. The levels of serum total cholesterol (TC), triglyceride (TG), high-density lipoprotein cholesterol (HDL-C), and low-density lipoprotein cholesterol (LDL-C) in samples were determined by enzymatic methods with commercially available kits, Tcho-1, TG-LH. Cholestest HDL, and Cholestest LDL, respectively. Serum apolipoprotein (Apo) A1 and ApoB levels were detected by the immunoturbidimetric immunoassay. All determinations were performed with an autoanalyzer in the Clinical Science Experiment Center of the First Affiliated Hospital, Guangxi Medical University [27].

Diagnostic criteria
The normal values of serum TC, TG, HDL-C, LDL-C, ApoA1, ApoB levels and the ApoA1/ApoB ratio in our Clinical Science Experiment Center were 3.10-5.17, 0.56-1.70, 0.91-1.81, 2.70-3.20 mmol/L, 1.00-1.78, 0.63-1.14 g/L, and 1.00-2.50; respectively. Hypertension was diagnosed according to the criteria from the 1999 World Health Organization-International Society of Hypertension Guidelines for the management of hypertension [28]. The diagnostic criteria of overweight and obesity were according to the Cooperative Meta-analysis Group of China Obesity Task Force. Normal weight, overweight and obesity were defined as a BMI < 24, 24-28 and > 28 kg/m 2 , respectively [29]. Dyslipidemia was defined according to World Health Organization criteria: TG ≥ 1.7 mmol/L and HDL-C < 0.9 mmol/L for men or < 1.0 mmol/L for women. Diabetes was defined as a fasting plasma glucose ≥ 7.0 mmol/L or 2 h postprandial plasma glucose ≥ 11.1 mmol/L or as having been previously diagnosed with diabetes and receiving therapy [30].

Bioinformatic analysis of miRNAs binding to SNP and linkage disequilibrium analysis
Bioinformatic software (http://bioin fo.life.hust.edu.cn/ miRNA SNP2/) was used to detect the candidate SNPs which could affect COL1A1 regulation via miRNAs [31].

RNA isolation
Fasting blood samples (5 mL

Statistical analysis
All statistical analyses were performed using the statistical software package SPSS 21.0 (SPSS Inc. Chicago, IL, USA) and R software (version 3.3.3). A Chi square analysis was used to evaluate the difference of the rate between the groups. Continuous data were presented as mean ± SD. For those, that are normally distributed, whereas the medians and interquartile ranges for TG, which is not normally distributed. Comparisons between groups for continuous data were made using Mann-Whitney nonparametric tests. The heart-map of correlation models and Bioinformatic analysis was measured by R software (version 3.3.3). The receiver operating characteristic (ROC) curve analysis was performed with plasma miR-3659 and miR-151a-5p to distinguish between dyslipidemia and control groups. The AUC was estimated to assess the diagnostic performance of miR-3659 and miR-151a-5p. All tests were two-sided, and P < 0.05 was considered statistically significant.

Data preprocessing
When the GSE66676 was analyzed, we can get 54560 expression probes separately from each gene expression profile. After data preprocessing, the expression matrices of 19938 genes were obtained from the 67 samples. All of the genes and the samples' phenotype were shown in Additional file 1: Tables S1 and S2.

Weighted gene co-expression networks
We selected soft-threshold β = 6 to construct gene modules using the WGCNA package (Additional file 2: Figure S1). After determining the soft threshold, all of genes were used to construct weighted gene co-expression networks. Then, we calculated the correlation matrix and adjacency matrix of the gene expression profile of the four lipid-profile groups, and then transformed them into a topological overlap matrix (TOM), and obtained a system clustering tree of genes on the basis of gene-gene non-ω similarity. Together with the TOM, we performed the hierarchical average linkage clustering method to identify the gene modules of each gene network (deepsplit = 2, cut height = 0.25). Six gene modules were recognized by the dynamic tree cut (Fig. 2).

Finding module of interest and functional annotation
It is a hugely valued biological significance to find out modules most significantly associated with clinical features. The highest association in the Module-feature relationship was found in grey60 module and TG (r 2 = 0.98, P = 7E-04), which was selected as module of interest and clinical feature to be studied in subsequent analyses (Fig. 3). The other modules without enough relationship or statistical significance for further consideration. In order to explore biological relevance of grey60 module, 139 genes which can be found in this module (Additional file 1: Table S3) were respective subjected to Gene Ontology (GO) functional and KEGG pathway enrichment analyses by R clusterProfiler package [20]. Biological processes, cell component, molecular function and KEGG pathway analysis of grey60 module were shown in Fig. 4. All of the databases were shown in Additional file 1: Tables S4 and S5.

Protein-protein interaction (PPI) network construction and identify hub genes
When the STRING database [33] was analyzed, a total of 64 nodes and 179 protein pairs were got with a combined weight score > 0.25 in grey60 module (Fig. 5).
After analysis in sub-module, only two modules with score > 6 were detected by MCODE. As shown in triangle cluster, COL1A1 had the highest score (Degree = 34, MCODE = 10.25). We hypothesized that COL1A1 as the hub gene was closely relevant to dyslipidemia occurs.

Demographic and biochemical characteristics
Demographic, epidemiological and clinical characteristics of the 857 analyzed study subjects are summarized in Table 1. All of the subjects were obese. The levels of weight, WC, BMI, systolic blood pressure (SBP), diastolic blood pressure (DBP), pulse pressure (PP), serum glucose, TC, TG, LDL-C and the percentages of diabetes and hypertension were higher, whereas the HDL-C levels were lower in dyslipidemia group as compared with normals.

Expression level of four miRNAs between the two groups
As compared with those of healthy controls, the relative expression levels of circulating miR-3659 and miR-151a-5p in dyslipidemic patients were significantly increased (P < 0.05; Fig. 6).

ROC curve for dyslipidemia
We performed a ROC analysis to determine the predictive values of miR-3659 and miR-151a-5p for dyslipidemia. The AUCs of miR-3659 and miR-151a-5p were 0.806 (95% CI 0.769-0.844; P < 0.001) and 0.769 (95% CI 0.729-0.808; P < 0.001), respectively. This indicated that Edge stands for the interaction between two genes. A degree was used for describing the importance of protein nodes in the network, red shows a high degree and blue presents a low degree. The significant two modules identified from the protein-protein interaction network shown with triangle (cluster 1) and diamond (cluster 2) using the molecular complex detection method with a score of > 6.0 the diagnostic performance of the miR-3659 was superior to miR-151a-5p (P = 0.02026; Fig. 7).

Discussion
Several recent reports showed that age, gender, smoking, obesity, dyslipidemia, lack of exercise, hypertension and diabetes mellitus are established risk factors for cardiovascular disease [34,35]. With the remarkable improvement of social living standard, obesity has turned into a worldwide epidemic [36]. As a complex and multifactorial disease, lots of environmental and genetic factors can result in obesity and dyslipidemia [37,38]. In the current study via WGCNA analysis, we have identified that COL1A1 may modify serum lipid levels in obese patients. Type 1 collagen is the main structural protein of bone and is encoded by two genes: COL1A1 and COL1A2. The COL1A1 is located in chromosome 17, region 17q21-22, and presents 51 exons. One of the most extensively studied polymorphisms is the so-called Sp1, which consists of the substitution of a guanine (G) by a thymine (T) in the first base of the first intron of the gene, which affects the binding site of the transcription factor Sp1 and thereby, the regulation of the gene transcription [39]. Besides this, a recent research found that COL1A1 polymorphisms could modify blood lipid levels, especially TG [40]. Previous studies also showed that the typical dyslipidemia of obesity consists of increased TG and free fatty acid, decreased HDL-C with HDL dysfunction and normal or slightly increased LDL-C with increased small dense LDL [41]. Therefore, in obese, when the COL1A1 expression changed, it may lead to serum TG increased and cause dyslipidemia.
MiRNAs, the endogenous and small noncoding RNAs, were found in lots of cell types and tissues especially the adipose tissue [42]. Besides, miRNAs may play an important role in the regulation of physiological and metabolic processes just as obesity, dyslipidemia, diabetes, aging, and others [43,44]. Given their role in regulating transcriptional networks, miRNAs in adipose tissue might offer attractive biomarkers of adiposity as well as potential therapeutic targets for treating metabolic disorders [45]. Lots of evidences demonstrated that SNPs localized at miRNA binding sites (miRSNPs) could affect the binding of miRNAs to the target genes and in turn result in reduction or increase in translation of the target mRNA and altered susceptibility to disease [46]. In the current study, we used bioinformatic software to identified four miRNAs which can bind SNPs in 3′UTR in COL1A1, but only two miRNAs had statistical significance. After ROC curve analysis, we showed that miR-3659 in obese might be a potential biomarker for dyslipidemia diagnosis.
This study had several limitations. First, it did not report serum hormone levels which had potential impact on obesity and dyslipidemia. Second, it is a cross-sectional study that suggested hypotheses but failed to describe the relationship between the putative cause and effect. Third, it is a single-center study involving a small number of patients, and large-scale multicenter studies are necessary to verify our findings. Finally, the Fig. 6 Binding of miRNA to COL1A1 SNPs minor alleles and the relative expression level. On top of the figure shown bioinformatic analysis of potential miRNAs binding to COL1A1 SNPs polymorphisms and below was the relative expression level of the four miRNAs between two groups