Accumulation of genetic and epigenetic changes alters regulation of a web of interconnected genes including microRNAs (miRNAs), which confer hallmark capabilities and characteristic cancer features. In this study, the miRNA and messenger RNA expression profiles of 126 non-small cell lung cancer specimens were analyzed, with special attention given to the diversity of lung adenocarcinomas. Of those, 76 adenocarcinomas were classified into two major subtypes, developing lung-like and adult lung-like, based on their distinctive miRNA expression profiles resembling those of either developing or adult lungs, respectively. A systems biology-based approach using a Bayesian network and non-parametric regression was employed to estimate the gene regulatory circuitry functioning in patient tumors in order to identify subnetworks enriched for genes with differential expression between the two major subtypes. miR-30d and miR-195, identified as hub genes in such subnetworks, had lower levels of expression in the developing lung-like subtype, whereas introduction of miR-30d or miR-195 into the lung cancer cell lines evoked shifts of messenger RNA expression profiles toward the adult lung-like subtype. Conversely, the influence of miR-30d and miR-195 was significantly different between the developing lung-like and adult lung-like subtypes in our analysis of the patient data set. In addition, RRM2, a child gene of the miR-30d-centered subnetwork, was found to be a direct target of miR-30d. Together, our findings reveal the existence of two miRNA expression profile-defined lung adenocarcinoma subtypes with distinctive clinicopathologic features and also suggest the usefulness of a systems biology-based approach to gain insight into the altered regulatory circuitry involved in cancer development.
!!!All Science Journal Classification (ASJC) codes