miRNAome-metabolome wide association study reveals effects of miRNA regulation in eosinophilia and airflow obstruction in childhood asthma

Abstract

Background: There are important inter-relationships between miRNAs and metabolites: alterations in miRNA expression can be induced by various metabolic stimuli, and miRNAs play a regulatory role in numerous cellular processes, impacting metabolism. While both specific miRNAs and metabolites have been identified for their role in childhood asthma, there has been no global assessment of the combined effect of miRNAs and the metabolome in childhood asthma.

Methods: We performed miRNAome-metabolome-wide association studies ('miR-metabo-WAS') in two childhood cohorts of asthma to evaluate the contemporaneous and persistent miRNA-metabolite associations: 1) Genetic Epidemiology of Asthma in Costa Rica Study (GACRS) (N = 1121); 2) the Childhood Asthma Management Program (CAMP) (N_Baseline = 312 and N_{End of trial} = 454). We conducted a meta-analysis of the two cohorts to identify common contemporaneous associations between CAMP and GACRS (false-discovery rate (FDR) = 0.05). We assessed persistent miRNA-metabolome associations using baseline miRNAs and metabolomic profiling in CAMP at the end of the trial. The relation between miRNAs, metabolites and clinical phenotypes, including airway hyper-responsiveness (AHR), peripheral blood eosinophilia, and airflow obstruction, were then assessed via. Mediation analysis with 1000 bootstraps at an FDR significance level of 0.05.

Findings: The meta-analysis yielded a total of 369 significant contemporaneous associations, involving 133 miRNAs and 60 metabolites. We identified 13 central hub metabolites (taurine, 12,13-diHOME, sebacate, 9-cis-retinoic acid, azelate, asparagine, C5:1 carnitine, cortisol, 3-methyladipate, inosine, NMMA, glycine, and Pyroglutamic acid) and four hub miRNAs (hsa-miR-186-5p, hsa-miR-143-3p, hsa-miR-192-5p, and hsa-miR-223-3p). Nine of these associations, between eight miRNAs and eight metabolites, were persistent in CAMP from baseline to the end of trial. Finally, five central hub metabolites (9-cis-retinoic acid, taurine, sebacate, azelate, and 12,13-diHOME) were identified as primary mediators in over 100 significant indirect miRNA-metabolite associations, with a collective influence on peripheral blood eosinophilia, AHR, and airflow obstruction.

Interpretation: The robust association between miRNAs and metabolites, along with the substantial indirect impact of miRNAs via 5 hub metabolites on multiple clinical asthma metrics, suggests important integrated effects of miRNAs and metabolites on asthma. These findings imply that the indirect regulation of metabolism and cellular functions by miRNA influences Th2 inflammation, AHR, and airflow obstruction in childhood asthma.

Funding: Molecular data for CAMP and GACRS via the Trans-Omics in Precision Medicine (TOPMed) program was supported by the National Heart, Lung, and Blood Institute (NHLBI).

Keywords: Airflow obstruction; Airway hyper-responsiveness; Eosinophilia; Metabolite; microRNA.

Fig. 1

Study design and workflow. The study includes two independent cohorts of children with asthma: GACRS and CAMP. Samples of plasma and serum were gathered from participants in both cohorts. MiRNA sequencing was conducted on the serum samples from both cohorts. The GACRS cohort underwent metabolomic profiling using plasma samples, while the CAMP cohort had baseline serum samples and end-of-trial plasma samples analysed for metabolomic profiling. The miRNA sequencing data and metabolome profiles from both cohorts were harmonized and employed a shared set of miRNA and metabolite data for subsequent analysis. The workflow includes global miR-metabolite wide association analyses separately within each cohort, and pooling effect sizes using fixed-effect meta-analysis. Subsequently, causal inference modelling was employed via mediation analysis to integrate the miR-metabo-WAS findings with asthma outcomes.

Fig. 2

Contemporaneous miRNAome-metabolome-wide association (miR-metabo-WAS) analysis in two childhood asthma cohorts. Edges: Red are positive beta-estimates; Blue are negative beta-estimates, and weighted according to beta-estimate magnitude. Metabolites are grouped according to sub-pathways. The dotted encircled hub metabolites were also found in prospective miR-metabo-WAS.

Fig. 3

Persistent association between miRNA and metabolite at TrialEnd timepoint in CAMP cohort. Nine miRNA-metabolite pairs, meeting the FDR cut-off in meta-analysis (combining GACRS and CAMP Baseline miR-metabo-WAS), displayed consistent effects at a 5% p-value significance level in the CAMP TrialEnd point.

Fig. 4

A network view depicting hub metabolites, their associated miRNAs, and the target genes within the corresponding metabolic pathway. In mediation analysis, five metabolites and their associated miRNAs (N = 59) showed a significant mediation effect on all three asthma outcomes—peripheral blood eosinophilia, AHR, and airflow obstruction—at a 5% FDR. The network view also integrates the genes targeted by the associated miRNAs within the metabolomic pathways of the respective metabolites.