Untargeted sequencing of circulating microRNAs in a healthy and diseased older population

Abstract

We performed untargeted profiling of circulating microRNAs (miRNAs) in a well characterized cohort of older adults to verify associations of health and disease-related biomarkers with systemic miRNA expression. Differential expression analysis revealed 30 miRNAs that significantly differed between healthy active, healthy sedentary and sedentary cardiovascular risk patients. Increased expression of miRNAs miR-193b-5p, miR-122-5p, miR-885-3p, miR-193a-5p, miR-34a-5p, miR-505-3p, miR-194-5p, miR-27b-3p, miR-885-5p, miR-23b-5b, miR-365a-3p, miR-365b-3p, miR-22-5p was associated with a higher metabolic risk profile, unfavourable macro- and microvascular health, lower physical activity (PA) as well as cardiorespiratory fitness (CRF) levels. Increased expression of miR-342-3p, miR-1-3p, miR-92b-5p, miR-454-3p, miR-190a-5p and miR-375-3p was associated with a lower metabolic risk profile, favourable macro- and microvascular health as well as higher PA and CRF. Of note, the first two principal components explained as much as 20% and 11% of the data variance. miRNAs and their potential target genes appear to mediate disease- and health-related physiological and pathophysiological adaptations that need to be validated and supported by further downstream analysis in future studies.Clinical Trial Registration: ClinicalTrials.gov: NCT02796976 ( https://ichgcp.net/clinical-trials-registry/NCT02796976 ).

Conflict of interest statement

The authors declare no competing interests.

© 2022. The Author(s).

Figures

Figure 1

Summary of analysis of EXAMIN AGE data. (A). Distribution of the number of individual miRNAs detected in specific numbers of samples and number and percentage of mature miRs identified in the indicated proportions of samples. (B). Numbers of miRNAs with specific values of median expression across samples (miRNAs with zero median expression are not shown). (C). Outline of the comparison of miRNA expression between groups of patients. DEA differential expression analysis, HA healthy active, HS healthy sedentary, SR sedentary at increased CV risk, FC fold change, FDR false discovery rate.

Figure 2

miRNA expression changes between conditions. (A) ‘Volcano plots’ showing the fold-change versus statistical significance measure for pairs of sample groups. (B) Mean versus fold-change expression of individual miRNAs in pairs of sample groups. HA healthy active, HS healthy sedentary, SR sedentary at increased CV risk, FDR false discovery rates, CPM count-per-million.

Figure 3

Relationship between miRNA expression levels and values of various physiological parameters. (A) Correlation coefficients have been calculated between the miRNA expression levels and the values of the indicated physiological parameters. (B) Fold-changes of the corresponding miRNAs among the indicated pairs of samples. (C) Median expression level of these miRNAs across samples (in counts per million, CPM). Fat fat mass, BMI body mass index, hipcirc hip circumference, hscrp high sensitive C-reactive protein, waistcirc waist circumference, sys_24h systolic blood pressure based on 24 h monitoring, dia_24h diastolic blood pressure based on 24 h monitoring, pwv pulse wave velocity, ldl low-density lipoprotein, tnf_alpha tumor necrosis factor alpha, CRVE central retinal venular equivalent, pwv_24h pulse wave velocity based on 24 h monitoring, CRAE central retinal arteriolar equivalent, ACC accelerometer, active activity counts, iop intraocular pressure, IL interleukin, AVR arteriolar-to-venular diameter ratio, hdl high-density lipoprotein, VO2peak peak oxygen uptake, HA healthy active individuals, HS healthy sedentary individuals, SR sedentary individuals with at least 2 cardiovascular risk factors.

Figure 4

Principal component analysis of the combined miRNA and physiological parameter data. (A) Projection of the data on the first two principal components, which explain ~ 20% and 11% of the variance. PC1 imperfectly separates the healthy active individuals from sedentary at risk. (B) PCA loadings of the first 30 variables contributing most to explaining the variance in the data (as indicated on the colour scale).