Meta- and cross-species analyses of insulin resistance based on gene expression datasets in human white adipose tissues

Ample evidence indicates that insulin resistance (IR) is closely related to white adipose tissue (WAT), but the underlying mechanisms of IR pathogenesis are still unclear. Using 352 microarray datasets from seven independent studies, we identified a meta-signature which comprised of 1,413 genes. Our meta-signature was also enriched in overall WAT in in vitro and in vivo IR models. Only 12 core enrichment genes were consistently enriched across all IR models. Among the meta-signature, we identified a drug signature made up of 211 genes with expression levels that were co-regulated by thiazolidinediones and metformin using cross-species analysis. To confirm the clinical relevance of our drug signature, we found that the expression levels of 195 genes in the drug signature were significantly correlated with both homeostasis model assessment 2-IR score and body mass index. Finally, 18 genes from the drug signature were identified by protein-protein interaction network cluster. Four core enrichment genes were included in 18 genes and the expression levels of selected 8 genes were validated by quantitative PCR. These findings suggest that our signatures provide a robust set of genetic markers which can be used to provide a starting point for developing potential therapeutic targets in improving IR in WAT.

4

Jung J, Kim GW, Lee W, Mok C, Chung SH, Jang W