Ethics Statement

The study is in conformity with Norwegian laws and the Helsinki declaration, and all participants signed a document of informed consent. The study was approved by the Regional Committees for Medical and Health Research Ethics (REC Central).

Subjects

The third wave of the Nord-Trøndelag Health Study (HUNT3 Study) in Norway was carried out between 2006 and 2008. Participants in the present study attended a sub-study in HUNT3 designed to measure maximal oxygen uptake (VO_2max) in healthy adult subjects and was called the HUNT Fitness Study [18]. Participants in the Fitness Study reported to be free from heart- or lung disease (details previously described [18]). Other exclusion criteria were cancer, pregnancy, or any other medical contraindication or orthopaedic limitation not permitting a maximal exercise stress test.

From the eligible 4631 participants who had completed a successful VO_2max-test in the HUNT Fitness Study, 12 participants with high VO_2max and 12 participants with low VO_2max were selected for the initial screening study. All subjects were between 40 and 45 years old and were matched on gender, age (within 0.9 year between case and control), and equal levels of physical activity. In addition, the subjects were matched on the basis of having as similar as possible body mass index (BMI), blood pressure (BP), and serum levels of glucose, cholesterol and triglycerides. Thus, the only cardiovascular risk factor that was different between the cases and the controls was VO_2max.

To verify the findings in a validation cohort, 76 new participants from the HUNT Fitness Study were selected (38 with low VO_2max and 38 with high VO_2max) based on the same matching criterias as for the screening cohort.

Anthropometric Measurements

Weight and height were measured on a combined scale DS-102 (Arctic Heating AS, Nttery, Norway), and BMI was calculated as weight divided by height squared (kg/m²). An individualized protocol was applied to measure VO_2max [19]. Oxygen uptake kinetics were measured directly by a portable mixing chamber gas-analyzer (Cortex MetaMax II, Cortex, Germany) with the participants wearing a tight face mask (Hans Rudolph, Germany) connected to the MetaMax II. The system has formerly been found to be valid [20]. Heart rate was measured by radio telemetry (Polar S610i, Polar Electro Oy, Finland). From the warm-up pace, the load was regularly increased when oxygen uptake kinetics flattened. Along with a respiratory quotient of 1.05 or higher, a maximal test was considered achieved when the oxygen uptake did not increase more than 2 ml/kg/min at the highest effort or before the participant disembarked the treadmill [21]. VO_2max was measured as litres of oxygen per minute (l/min) and subsequently calculated as VO_2max relative to body mass (ml/kg/min) and VO_2max scaled (ml/kg^0.75/min).

Blood Sampling and Standard Biochemical Analyses

Blood sampling was performed before the start of the exercise test. Standard biochemical analyses were performed on fresh venous non-fasting blood samples at Levanger Hospital, Norway. Non-fasting glucose was analysed by Hexokinase/G-G-PDH methodology reagent kit 3L82-20/3L82-40 Glucose, high-density lipoprotein (HDL) cholesterol by the Accelerator selective detergent methodology reagent kit 3K33-20 Ultra HDL, triglycerides by Glycerol Phosphate Oxidase methodology reagent kit 7D74 Triglyceride, alanine aminotransferase (ALAT) by NADH (with P-5′-P) methodology reagent kit 8D36-30 Alanine aminotransferase activated, aspartate aminotransferase (ASAT) by NADH (with P-5′-P) methodology reagent kit 8D37-30 Aspartate aminotransferase activated, and C-reactive protein (CRP) was analysed by the Areoset CRP Vario kit (all analyses from Abbott Diagnostics, Illinois, US). ALAT measurements below the detection limit (LoD) were assigned a concentration of 9 u/l, ASAT measurements below LoD a concentration of 7 u/l, and CRP levels below LoD were recorded as 0 mg/l.

Questionnaire-based Information

Physical activity was registered based on the responses to a self-administered questionnaire [22]. The questionnaires included three questions: Question 1: “How frequently do you exercise?”, with the response options “Never” (0), “Less than once a week” (0.5), “Once a week” (1), “2–3 times per week” (2.5) and “Almost every day” (5). Question 2: “If you exercise as frequently as once or more times a week: How hard do you push yourself?” with the response options: “I take it easy without breaking a sweat or losing my breath” (1), “I push myself so hard that I lose my breath and break into sweat” (2) and “I push myself to near exhaustion” (3). Question 3: “How long does each session last?”, with the response options: “Less than 15 minutes” (0.1), “16–30 minutes” (0.38), “30 minutes to 1 hour” (0.75) and “More than 1 hour” (1.0). Each participant’s response to the above three questions (i.e. numbers in parentheses) were multiplied to calculate a physical activity index score (Kurtze score). As the second and third question only addressed people who exercised at least once a week, both “Never” and “Less than once a week” yielded an index score of zero. Participants with a zero score were categorized as inactive.

The Finnish Type 2 Diabetes Risk Score (FINDRISC) was calculated for all the participants based on lifestyle variables, family history and BMI. FINDRISC predicts the risk of being diagnosed with type 2 diabetes in the next ten years and requires no laboratory measurements [23]. We also calculated the Framingham Risk Score, which estimates the 10-year risk of having a myocardial infarction, and the NORRISK Score, which is optimized for the Norwegian population and estimates the 10-year risk of cardiovascular mortality [24], [25].

miR Isolation

miRs were extracted from serum using the miRNeasy® Mini Kit (Qiagen, US). To control isolation efficiency, two syntetic spike-ins were added to the samples (UniSp2 and UniSp4). Briefly, 750 µl of a Qiazol mixture containing 1.25 µg/ml of MS2 bacteriophage RNA (RNA carrier not containing miRs) was added to 200 µl serum. The MS2 bacteriophage was added to increase the extraction of miRs. miRs were then extractet using chloroform, ethanol and spin columns. miRs were eluted in 50 µl of RNase-free water and stored in a −80°C freezer prior to analysis.

Real-time Quantitative Polymerase Chain Reaction (RT-qPCR)

Fifteen µl RNA was reversely transcribed in 75µl reactions using the miRCURY LNA™ Universal RT microRNA PCR, Polyadenylation and cDNA synthesis kit (Exiqon, Denmark). cDNA was diluted 50× and assayed in 10 µl PCR reactions according to the protocol for miRCURY LNA™ Universal RT microRNA PCR. For the screening approach, all miRs were assayed once by RT-qPCR on the microRNA Ready-to-Use PCR, Human panel I and panel II. For the validation approach, selected candidate miR were analysed by the custom-made “Pick-and-mix” RT-qPCR system (Exiqon, Denmark). Negative controls excluding template from the reverse transcription reaction was included and profiled like the samples. The amplification was performed in a LightCycler® 480 RT-qPCR System (Roche, Switzerland) in 384 well plates. The amplification curves were analysed using the Roche LC software, both for determination of Cp (by the 2nd derivative method) and for melting curve analysis.

miR Data Analysis

Screening of 720 miRs was performed in serum samples and the amplification efficiency was calculated using algorithms similar to the LinReg software. All assays were inspected for distinct melting curves and the Tm was checked to be within known specifications of the assay. Individual miRs had to be detected with Cp<37 and 5 Cp’s less than the negative control (blank) to be included in the data analysis. Data that did not pass these criteria were omitted from further analysis. Using SLqPCR on the screening cohort, the best normalizer was found to be the average of assays detected in all samples (global mean) and therefore data were normalized to the global mean (average – assay Cp). For the validation cohort, we normalized data to the level of miR-425, as miR-425 levels were expressed close to the global mean in the screening cohort and miR-425 previously has been found valid as a housekeeping miR [26].

miR Target Prediction

The web-based computational tool DIANA-mirPath was used to identify molecular pathways potentially altered by the coordinated change in expression of the three fitness-miRs [27]. DIANA mirPath combines the prediction tool Targetscan 5.1 and the pathway tool KEGG (Kyoto Encyclopedia of Genes and Genomes). DIANA-mirPath was set to use TargetScan Human 5.1 to identify possible mRNA targets. The level of significance was set at p<0.05.

Statistical Analysis

PASW Statistics 17.0 (IBM, US) was used for traditional statistical analyses. All statistical tests were two-sided, and p-values below 0.05 were considered statistically significant. Kolmogorov-Smirnov test was used to test for normality. One-Way ANOVA was used to compare variables between the high and the low VO_2max groups. Pearson’s correlation was used to study associations between normally distributed variables. Linear regression analysis was used to predict the contribution of miRs to the level of VO_2max. Logistic regression was uset to determine the odds ratios for different predictors and outcomes.

Correlations between miRs and Fitness Related Variables

A significant inverse correlation was found between VO_2max and miR-210 (n = 100, r = −0.35, p<0.001). Despite the observed group differences in miR-222 and miR-21, only a weak correlation was found between miR-21 and VO_2max (n = 100, r = −0.20, p<0.05) and no correlation was found between miR-222 and VO_2max. No correlations were found between miR-210, miR-21 and miR-222 and exercise habits like frequency, intensity, and duration of self-reported regular physical activity, nor with self-reported time since last bout of exercise in the validation cohort.

Linear regression analysis showed that the serum level of miR-210 explained 12% of the variation in VO_2max (r² = 0.12). When adding the serum levels of miR-21 and miR-222 to the analysis, these three miRs in combination explained 15% of the variation in VO_2max. However, the best prediction model for VO_2max was obtained by using miR-210 alone. miR-210 levels in the lowest quartile increased the odds ratio by 10.4 times for having a high VO_2max (>120 ml/kg^0.75/min).

Correlations between miRs and Variables Related to CVD Risk

We observed significant correlations between the FINDRISC Score and miR-21 (n = 100, r = 0.37, p<0.001) and miR-222 (n = 100, r = 0.34, p<0.001). Assessing correlations with other CVD risk factors, we found a positive correlation between miR-21 and CRP (n = 51, r = 0.35, p<0.05) and between miR-210 and serum ASAT (n = 16, r = 0.78, p<0.0005). let-7d (n = 34 r = 0.48, p<0.005), and miR-103 (n = 62 r = 0.31, p<0.05), were also positively correlated with age in the validation cohort. In contrast, we found no correlations between the most common risk factors for CVD (BP, cholesterol, smoking habit, or obesity) and miR-21, miR-210 and miR-222 levels. Moreover, no correlations were found between the Framingham Risk Score or the NORRISK Score and miR-21, miR-210 and miR-222 levels.

Targeted Pathways by Fitness-miRs

Bioinformatics may provide insight into signaling pathways that are targeted by different miRs [27]. Using bioinformatics, miR-21, miR-210 and miR-222 were predicted to target 611 genes. The KEGG pathway database recognized 138 of those genes. The signaling pathways that, based on prediction, were found to significantly segregate subjects with high and low VO_2max were the MAPK pathway, TGF-β pathway, B-cell receptor pathway, Wnt pathway, T-cell receptor pathway, mTOR pathway, p53 pathway, ErbB pathway, and VEGF pathway (p<0.05).

Correlations between miRs and Variables Related to Fitness and CVD Risk

In this study we have matched the contrasting groups on physical activity level (based on the Kurtze score [22]), to avoid finding typical exercise induced miRs. Previous studies have shown that circulating levels of miR-21 and miR-222 are regulated by acute and long-term exercise training [16]. However, circulating levels of miR-210 have previously been reported to be unaffected both by acute and long-term exercise training [16]. In our study, no correlations were found between miR-210, miR-21 and miR-222 and exercise habits like frequency, intensity, or duration of self-reported regular exercise training. Additionally, no correlations were found between the three miRs and the time since the last exercise bout. Based on this, we can probably rule out the possibility that acute effects of exercise are responsible for our findings.

Surprisingly, no correlations were found between miR-21, miR-210 and miR-222 levels and important CVD risk factors such as BP, cholesterol, smoking habit, and obesity. However, miR-21 and miR-222 correlated with the FINDRISC score, which predicts the risk of being diagnosed with type 2-diabetes in the next ten years. As mentioned above, this finding strengthens our hypothesis that fitness biomarkers also have potential as markers of future cardiovascular health independent of traditional biomarkers.

Target Genes and Pathways

miR-210 is considered the hypoxia-miR [28]. Thus, the high circulating levels of miR-210 in subjects with low VO_2max, may suggest that hypoxia-induced pathways are activated. In line with this, the mirPath prediction tool identified the VEGF pathway, which is another important hypoxia pathway, to be differentially activated in the high and low VO_2max group. Interestingly, in endothelial cells miR-210 and the VEGF pathway seem to interact in the formation of new capillaries [30]. However, whether subjects with low aerobic fitness have increased activity in hypoxia- and angiogenesis pathways to compensate for the low oxygen uptake, will need confirmation in additional clinical and experimental studies.

The pathway analysis also predicted that miR-21, miR-210 and miR-222 inhibit the expression of several key proteins in more general pathways associated with growth and development (MAPK, TGF-β, mTOR, ErbB and the Wnt pathways), the immune system (T and B cell signaling) and stress (p53 pathway). Potentially induced pathways in the vessel wall and the heart of subjects with low VO_2max are illustrated in Figure 3. The implications of this for the pathophysiology of subjects with low VO_2max is not clear and will need more detailed studies of specific pathways. We also acknowledge that currently only limited information is available regarding the origin of circulating miRs. Hence, extrapolating from miRs in the peripheral blood to functional aspects in specific organs should be done with caution. Still, the association to hypoxia- and angiogenesis-related pathways seems interesting and should be further explored.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 3. Potentially increased microRNAs and associated pathways in subjects with low maximal oxygen uptake (VO2max).

The connections between the biological processes, microRNAs and the regulated pathways in A) the vessel wall and (B) the heart are based on the current knowledge in the literature. Red boxes illustrate the microRNAs that were increased in patients with low VO2max and the yellow boxes depict the associated pathways as identified by miR-Path analyses. Upward arrows inside the boxes illustrate up-regulation and downward arrows illustrate down-regulation. Blue dashed lines indicate the putative release of the microRNAs to the circulation. Pointed arrows illustrate activation and flat arrows illustrate inhibition. c-miR: Circulating microRNA, VEGF: Vascular endothelial growth factor, VO2max: Maximal oxygen uptake.

https://doi.org/10.1371/journal.pone.0057496.g003

To our knowledge only one previous study has described that circulating miRs may be associated with fitness measured as VO_2max [16]. This paper has reported a positive correlation between the circulating levels of miR-146a and VO_2max [16]. No such correlation was found in this study. However, this discrepancy may be explained by the differences in study-population as Baggish et al studied 10 male athletic subjects [16].

Limitations

One limitation of this study is that the analyzed blood samples were collected from non-fasting individuals. However, the time since last meal was similar in both the high and low VO_2max group (3.1 hours). We therefore believe that food intake did not influence the miR results. Another limitation of the study is that 8 of the 100 participants reported to have performed exercise training the same day as blood sampling, spanning from 1 hour to 11 hours before the samples were collected. Baggish et al have previously shown that exercise training <24 hours before blood sampling can influence the levels of circulating miRs such as miR-222 and miR-21 independent of VO_2max. However, removing these 8 participants from the statistical tests had no influence on the results. Furthermore, correlation analysis between miR-levels (miR-21, miR-210, miR-222) and self-reported time since last bout of exercise showed no correlations.

Conclusion

The results from this study suggest a potential value of measuring circulating miRNAs as biomarkers of aerobic fitness. The results indicated that miR-210 was the best marker of VO_2max-level, based on increment in the low VO_2max group in both the screening and validation cohort, and the significant correlation with VO_2max-level. The lack of association between miR-210, miR-21, miR-222, and other fitness related variables indicates that these three miRs may have potential as new independent biomarkers of fitness level. Furthermore, the lack of association between miR-210, miR-21, miR-222 and traditional CVD risk factors suggest a potential of these miRs as early biomarkers of future cardiovascular health, and should be further studied. Potential bias involve that blood samples were collected from non-fasting individuals, and that 8 of the participants reported to have performed exercise training within 24 hours before sampling.