Differences in microRNA levels across metabo-endotypes reveal novel insights into asthma heterogeneity.

RATIONALE: We previously validated five clinically distinct asthma metabo-endotypes (mechanistically derived asthma subgroups). We hypothesize that metabo-endotype membership may be partially driven by differences in serum microRNA profiles and their influence on metabolite levels. OBJECTIVES: To determine whether serum miRNA levels can help understand the underlying drivers of metabolic dysregulation across metabo-endotypes. METHOD: We compared expression levels of serum microRNAs across 1121 children grouped into five asthma metabo-endotypes using ANCOVA. A LASSO model was leveraged to determine the most important miRNAs for discriminating metabo-endotype membership. Finally, multiple linear regression models and two-sample t-tests were employed to determine whether serum microRNA ~ plasma metabolite relationships differed between individuals within different metabo-endotypes. MEASUREMENTS AND MAIN RESULTS: Of 317 serum miRNAs, 132 (41.6%) demonstrated significantly different expression across metabo-endotypes (FDR < 0.05), with miR-143-3p showing the greatest variation (FDR p = 5.7 × 10− 19). Most differences were driven by metabo-endotypes 2 and 3, the most and least severe. Enrichment analysis of microRNAs’ predicted target genes revealed critical asthma pathways, including Th17 and Th1/Th2 cell differentiation. A model based on 17 miRNAs was able to discriminate membership of metabo-endotype 2 versus 3 (AUC:81%, CI: 73%-88%). There was some evidence that relationships between specific miRNAs and metabolites differed between individuals in metabo-endotypes 2 and 3, which may suggest differential posttranscriptional regulation of pathways including eicosanoid and arginine metabolism. CONCLUSIONS: The results provide some evidence to suggest differential miRNA regulated gene expression between biologically and clinically distinct asthma metabo-endotypes, with a potentially important role for miR-143-3p. Understanding these relationships may uncover novel therapeutic targets and guide more personalized treatment strategies.

Duke Scholars

Author Qingwen Chen