Circulating MicroRNA: Incident Asthma Prediction and Vitamin D Effect Modification

Abstract

Of children with recurrent wheezing in early childhood, approximately half go on to develop asthma. MicroRNAs have been described as excellent non-invasive biomarkers due to their prognostic utility. We hypothesized that circulating microRNAs can predict incident asthma and that that prediction might be modified by vitamin D. We selected 75 participants with recurrent wheezing at 3 years old from the Vitamin D Antenatal Asthma Reduction Trial (VDAART). Plasma samples were collected at age 3 and sequenced for small RNA-Seq. The read counts were normalized and filtered by depth and coverage. Logistic regression was employed to associate miRNAs at age 3 with asthma status at age 5. While the overall effect of miRNA on asthma occurrence was weak, we identified 38 miRNAs with a significant interaction effect with vitamin D and 32 miRNAs with a significant main effect in the high vitamin D treatment group in VDAART. We validated the VDAART results in Project Viva for both the main effect and interaction effect. Meta-analysis was performed on both cohorts to obtain the combined effect and a logistic regression model was used to predict incident asthma at age 7 in Project Viva. Of the 23 overlapped miRNAs in the stratified and interaction analysis above, 9 miRNAs were replicated in Project Viva with strong effect size and remained in the meta-analysis of the two populations. The target genes of the 9 miRNAs were enriched for asthma-related Kyoto Encyclopedia of Genes and Genomes (KEGG) signaling pathways. Using logistic regression, microRNA hsa-miR-574-5p had a good prognostic ability for incident asthma prognosis with an area under the receiver operating characteristic (AUROC) of 0.83. In conclusion, miRNAs appear to be good biomarkers of incident asthma, but only when vitamin D level is considered.