The study was conducted between March and May 2010 in the Caihe community of Hangzhou (Zhejiang, China). A total of 624 eligible Han Chinese participants (age, 40-65 years) were recruited. The Medical Ethics Committee of Sir Run Run Shaw Hospital affiliated to School of Medicine, Zhejiang University (Hangzhou, China) approved the present study. Written informed consent was obtained from the patients. Participants were 56.80±6.54 years and 51.25% were male. All patients completed a population-based cross-sectional survey and were assigned a number. Random numbers were generated electronically and 40 patients with MetS (56.78±6.69 years; 45% male) and 40 control subjects (56.83±6.00 years; 57.5% male) were selected. Baseline anthropometric and metabolic measures were recorded using standardized methods (19,20). Participants were interviewed face-to-face, completing a questionnaire regarding demographic data, life style, present and past illness, medical therapy and other health-associated information. All measurements were referred to as previously reported (19,20).

MetS was diagnosed according to criteria established by the Joint Committee for Developing Chinese Guidelines on Prevention and Treatment of Dyslipidemia in Adults (21). Individuals with ≥3 of the following abnormalities were considered to have MetS: Central obesity [waist circumference (WC), >90 cm for men and >85 cm for women]; hypertriglyceridemia (≥1.70 mmol/l); high density lipoprotein-cholesterol (HDL-c; <1.04 mmol/l); blood pressure (BP; ≥130/85 mmHg or ongoing treatment for hypertension); and hyperglycemia [fasting plasma glucose (FPG) ≥6.1 mmol/l or 2 h postprandial glucose ≥7.8 mmol/l]. Patients suffering from abnormal renal function or renal dysfunction at the time of recruitment were excluded. Patients fulfilling none of the above criteria were selected as healthy controls.

Baseline anthropometric and metabolic measures were collected from all patients. Anthropometric data included height, weight, WC, hip circumference, heart rate and BP. The body mass index (BMI) was calculated as weight (kg)/[height (m)]². Systolic (S) BP and diastolic BP were calculated as the mean of three measurements, using a mercury sphygmomanometer at 3-min intervals. The body fat (%) was assessed using a Tanita Body Composition Analyzer TBF-300 (Tanita Corporation, Tokyo, Japan). All subjects underwent abdominal magnetic resonance imaging (MRI) using a whole-body imaging system (SMT-100; Shimadzu Corporation, Kyoto, Japan) with TR-500 and TE-200 of spin-echo sequences (22). MRI scans were performed at the level of umbilicus between L4 and L5 with the subject in the supine position. Abdominal visceral fat area (VFA) and subcutaneous fat area (SFA) were calculated using Slice O'matic software (version 4.2; TomoVision, Magog, Canada). Following overnight fasting, blood and urine samples were collected; 5 ml blood were stored in collection tubes with EDTA as an anticoagulant.

FPG, 2 h postprandial glucose, triglyceride (TG), total cholesterol, aspartate aminotransferase, alanine aminotransferase, low-density lipoprotein-cholesterol (LDL-c) and HDL-c were measured using an auto-analyzer (Abbott Pharmaceutical Co. Ltd., Lake Bluff, IL, USA). Fasting serum insulin levels (FINS), 2h postprandial insulin levels were measured by using an insulin detection kit (cat. no. 2400218; Beijing North Institute Biological Technology, Beijing, China). Insulin sensitivity was assessed by homeostasis model assessment for insulin resistance (HOMA-IR) based on FPG and insulin measurements as follows: [FINS (mU/l) x FPG (mmol/l) /22.5(23). Glycosylated hemoglobin A1c (HbA1c) values were tested using ion-exchange high-performance liquid chromatography VARIANT II (D-100TM Hemoglobin Testing System; Bio-Rad Laboratories, Inc., Hercules, CA, USA) as described previously (24).

Genome-wide urinary miRNA profiles were detected using a microarray in whole urinary samples from 4 patients with MetS and 4 controls. Patients for the microarray analysis were selected at random out of the 40 patients previously identified per group. Comprehensive coverage was ensured using the TaqMan^® Array Human MicroRNA Cards B v3 (cat. no. 4444910; Thermo Fisher Scientific, Inc., Waltham, MA, USA), resulting in a total of 377 unique assays specific for human miRNAs. A total of 8 RNA samples were analyzed by Kangchen BioTech Co., Ltd. (Shanghai, China) according to manufacturers' instructions.

Total RNA was extracted from urine using miRNeasy kit (Qiagen Inc.) and purified using the Urine RNA Purification kit (Abnova, Taipei, Taiwan) according to manufacturers' instructions. For each reaction, 0.5-1 µg of total RNA isolated from 5-10 ml of whole urine samples was used. MiDETECT A Track^TM miRNA RT-PCR Start kit (R10048.3; Guangzhou RiboBio Co., Ltd., Guangzhou, China) was used for validating miR-29a-3p and U6 expression via stem-loop-based reverse transcription-quantitative polymerase chain reaction (RT-qPCR) following the manufacturers' instructions and the following thermocycling conditions: 95˚C for 5 min followed by 35 cycles of 95˚C for 10 sec, 60˚C for 20 sec and 70˚C for 10 sec. miRNA expression was normalized to U6 using the 2^-ΔΔCq method (25). Primer details are presented in Table I.

293 cells were selected for the transfection experiments. Cells were purchased from the American Type Culture Collection (Manassas, VA, USA) and cultured in Dulbecco's modified Eagle's medium (DMEM; 11965-092; Thermo Fisher Scientific, Inc.) containing 10% fetal bovine serum (FBS; Bio-Rad Laboratories, Inc.) and 100 IU/ml penicillin/streptomycin (Sigma-Aldrich; Merck KGaA, Darmstadt, Germany) at 37˚C with 5% CO₂. Cholesterol-conjugated 2'-O-methyl-modified mimics agomir-29a-3p (cat. no. miR40000210), agomir-NC (cat. no. miR04101), antagomir-29a-3p (cat. no. miR30000256) and antagomir-NC (cat. no. miR03101) were synthesized by Guangzhou RiboBio Co., Ltd. and used in transfection experiments. 293 cells were grown to 30-40% confluence and agomir-29a-3p (100 nM) or antagomir-29a-3p (200 nM) or respective controls were transiently transfected using riboFECTTM CP transfection kit (cat. no. R10035.3; RiboBio Co., Ltd, Guangzhou, China) according to the manufacturer's instructions (26). Cells were harvested at 48 h following treatment and RT-qPCR analysis was performed to assess transfection efficiency.

The homo sapiens insulin-like growth factor 1 (IGF1) 3'-untranslated region (3'UTR) was amplified from 293 cells and cloned into the XhoI/HindIII sites of the pGL4-basic vector (Promega Corporation, Madison, WI, USA). Cloning products were verified by sequencing and primer sequences were as follows: Forward, 5'-CCGCTCGAGACC ATCTCATGCTCTGTGGC-3' and reverse, 5'-CCCAAG CTTCTTTGCAAGGGAGGGGCATA-3'. 293 cells (5x10⁴) were seeded in 24-well plates 24 h prior to transfection. IGF1 plasmid (500 ng) and internal control Renilla luciferase plasmid (200 ng) were co-transfected using Lipofectamine 2000 (Invitrogen; Thermo Fisher Scientific, Inc.). Following 6-8 h, 100 nM agomir-29a-3p and agomir-NC were transiently transfected into the cells using the riboFECTTM CP transfection kit. Cells were harvested at 48 h post-transfection and assayed for luciferase activity using the Dual-Luciferase Reporter Assay System (Promega Corporation). Activities were normalized to Renilla and all experiments were repeated four times.

Total RNA was isolated from 293 cells using TRIzol reagent (Invitrogen; Thermo Fisher Scientific, Inc.) and RT was performed using a reverse transcription reagent kit (Takara Bio, Inc., Otsu, Japan) to generate the cDNA and samples were incubated for 30 min at 16˚C, followed by 30 min at 42˚C and 5 min at 85˚C. qPCR was performed using a PCR kit containing SYBR Green (Takara Bio, Inc.) and a 7500 real-time PCR detection system (Applied Biosystems; Thermo Fisher Scientific, Inc.). Thermocycling conditions were as follows: 95˚C for 30 sec followed by 40 cycles of 95˚C for 5 sec, 60˚C for 30 sec and 70˚C for 10 sec. IGF1 (forward, 5'-GGAGGCTGGAGATGTATTGC-3'and reverse, 5'-ACTTGCTTCTGTCCCCTCCT-3') expression was normalized to GAPDH (forward, 5'-AGCAGTCCCGTACAC TGGCAAAC-3'and reverse, 5'-TCTGTGGTGATGTAAAT GTCCTCT-3') and samples were quantified as described above. Measurements were conducted in triplicate.

Whole cell extracts were obtained by lysing cells in buffer (50 mM Tris pH 7.4, 150 mM NaCl, 1 mM EDTA, 10 mM sodium pyrophosphate, 0.5% Triton X-100, 1 mM sodium vanadate and 25 mM sodium fluoride) containing protease inhibitors (5 µg/ml phenylmethylsulfonyl fluoride, 0.5 µg/ml leupeptin, 0.7 µg/ml pepstatin and 0.5 µg/ml aprotinin). Protein concentrations were determined using a bicinchoninic acid protein assay. Equal amounts of protein (50 µg) were denatured by boiling, separated using 10% SDS-PAGE gels and transferred to polyvinylidene membranes. Membranes were blocked with 5% non-fat milk for 1 h at room temperature. Membranes were then incubated with primary antibodies, including polyclonal rabbit anti-IGF1 (cat. no. ab40657; 1:1,000; Abcam, Cambridge, UK) and monoclonal mouse anti-β-actin (cat. no. A5316; 1:5,000; Sigma-Aldrich; Merck KGaA) at 4˚C overnight followed by incubation with horseradish peroxidase-conjugated goat anti-rabbit/mouse secondary IgG antibodies (cat. nos. ab6721 and ab6789; 1:10,000; Abcam) at room temperature for 1 h. Immunoreactive proteins were detected using a chemiluminescent assay kit (EMD Millipore, Billerica, MA, USA). Quantity One software (v4.6.2; Bio-Rad Laboratories, Inc.) was used to quantify expression levels and normalize to β-actin.

To search for the enrichment of specific target genes associated with insulin signaling and the glucose and lipid metabolism, the Database for Annotation, Visualization and Integrated Discovery (https://david.ncifcrf.gov) were used. To predict target genes of miR-29a-3p, bioinformatics databases miRanda (http://www.microrna.org), mirBase (http://www.mirbase.org) and Targetscan (http://www.targetscan.org) were used.

For the patient study, all the continuous variables were tested for normal distribution and normally distributed variables are expressed as the mean ± standard deviation and variables with a skewed distribution are presented as the median with 25-75% interquartile range. Categorical variables were presented as frequencies and percentages. Differences in baseline characteristics were analyzed by two-sided Student's t-test for continuous variables and chi-square test for categorical variables. Correlation between urinary miR-29a-3p and metabolic parameters was determined using a Spearman correlation analysis. MetS incidence at different levels of urinary miR-29a-3p was analyzed by chi-square test. Logistic regression analysis and multiple stepwise regression analysis were applied in the analysis of the adjusted variables. A receiver operating characteristic (ROC) analysis was conducted and areas under the curve and optimal sensitivity and specificity levels were calculated. SPSS 17.0 (SPSS, Inc., Chicago, IL, USA) was used for statistical analyses. Cell experiments were evaluated using independent two-sided Student's t-test and one-way analysis of variance followed by Tukey's test for pairwise or multiple group comparisons, respectively. Data are presented as the mean ± standard deviation, representative of three replicates. P<0.05 was considered to indicate a statistically significant difference.

Anthropometric and metabolic characteristics of the study population at baseline are presented in Table II. Participants with MetS exhibited increased cardiovascular risk factors compared with the controls, including significantly higher BMI, waist-to-hip ratio, body fat, FINS levels, HOMA-IR, HbA1c, SFA, VFA, and MetS indicators, including significantly higher WC, blood glucose, TG, BP and significantly lower HDL-c (P<0.05).

Urine samples from 4 patients with MetS and 4 controls were analyzed using a microarray evaluating a total of 377 unique assays specific to human miRNAs. Among the 377 screened miRNAs, 20 were identified as differentially expressed in the urine of patients with MetS compared with the controls (fold change >2; P<0.05; Fig. 1A). Based on the microarray analysis, miR-29a-3p was selected for further validation by RT-qPCR using a larger sample size (n=40/group). RT-qPCR analyses confirmed that urinary miR-29a-3p levels were significantly increased in patients with MetS compared with the controls (P<0.01; Fig. 1B). Furthermore, participants with MetS had significantly increased urinary miR-29a-3p levels compared with the control (0.51±0.17 vs. 0.34±0.20; P<0.01).

Associations between urinary miR-29a-3p levels and parameters associated to adiposity, insulin resistance, lipid profiles and hepatic enzymes were further assessed. Spearman correlation analyses of urinary miR-29a-3p with metabolic risks were performed. As presented in Table III, following an adjustment for gender, age, smoking and drinking, urinary miR-29a-3p levels were correlated with HOMA-IR (r=0.2713; P<0.001), FINS (r=0.643; P<0.001) and HDL-c (r=0.242; P<0.001). Additionally, urinary miR-29a-3p levels were correlated with BMI, SBP, FPG and body fat (p<0.05; Table III).

Furthermore, it was assessed whether a high level of urinary miR-29a-3p may be used to predict the incidence of further metabolism subgroups, including central obesity, hypertension, hyperglycemia, dyslipidemia, and low HDL-c in patients with MetS. Multiple stepwise regression analysis using urinary miR-29a-3p levels as the dependent variable following the adjustment for gender, age, smoking and drinking revealed that urinary miR-29a-3p levels were independently associated with FINS (β=0.561; P<0.001), HDL-c (β=0.242; P<0.001) and BMI (β=-0.141; P<0.05; Table IV).

To assess the accuracy and clinical utility of urinary miR-29a-3p levels in differentiating between patients with MetS and control subjects in the validation phase (n=40/group), a ROC curve analysis was performed using urinary miR-29a-3p levels. Using a logistic regression model, analysis indicated that urinary miR-29a-3p may be a valuable biomarker in patient with MetS compared with healthy controls, with an area under the curve of 0.776 (95% CI, 0.603-0.829). The cut-off value for urinary miR-29a-3p levels was 0.375 and the optimal sensitivity and specificity were 77.5 and 60.0%, respectively (Fig. 2).

To identify potential molecular targets of miR-29a-3p in patients with MetS, gene ontology and biological association analyses using the Database for Annotation, Visualization and Integrated Discovery Bioinformatics Resources were performed (27) and searched for enrichment of specific target genes associated with insulin signaling, as well as glucose and lipid metabolism. Following evaluation of three widely used bioinformatics databases (20), it was determined that IGF1 was a potential target of miR-29a-3p (Fig. 3A).

Subsequently, a dual-reporter constructs was generated by fusion of the luciferase coding sequence to the 3'UTR of IGF1. Luciferase activity of the reporter constructs containing the predicted miR-29a-3p binding sites was determined. The dual-reporter luciferase assay revealed that agomir-29a-3p treatment significantly repressed luciferase activity of IGF1 compared with the control (48.97% reduction; P<0.01; Fig. 3B) and agomir-NC treatment had no significant effect. The results suggested that IGF1 was a direct target of miR-29a-3p.

To investigate whether miR-29a-3p expression effects to IGF1 levels, 293 cells were treated with miR-29a-3p agomir and antagomir for upregulation and knockdown of miR-29a-3p, respectively. Agomir-29a-3p significantly upregulated miR-29a-3p levels compared with the agomir-NC treated cells (7.24-fold increase; P<0.001; Fig. 3C). Antagomir-29a-3p treatment significantly reduced miR-29a-3p levels compared with antagomir-NC treated cells (~70% reduction; P<0.01; Fig. 3C). mRNA expression of IGF1was assessed following miR-29a-3p agomirs and antagomirs treatment. Compared with the negative controls, IGF1 levels were significantly reduced by agomir-29a-3p (P<0.05) and significantly increased by antagomir-29a-3p treatment (P<0.01; Fig. 3C). Results were confirmed by western blot analysis, exhibiting that protein levels of IGF1 were significantly downregulated by agomir-29a-3p and upregulated by antagomir-29a-3p compared with the controls (P<0.05; Fig. 3D).