mRNA-miRNA networks identify metabolic pathways associated to the anti-tumorigenic effect of thyroid hormone on preneoplastic nodules and hepatocellular carcinoma

Background: Thyroid hormones (THs) inhibit hepatocellular carcinoma (HCC) through different mechanisms. However, whether microRNAs play a role in the antitumorigenic effect of THs remains unknown. Methods: By next generation sequencing (NGS) we performed a comprehensive comparative miRNomic and transcriptomic analysis of rat hepatic preneoplastic lesions exposed or not to a short-term treatment with triiodothyronine (T3). The expression of the most deregulated miRs was also investigated in rat HCCs, and in human hepatoma cell lines, treated or not with T3. Results: Among miRs down-regulated in preneoplastic nodules following T3, co-expression networks revealed those targeting thyroid hormone receptor-β (Thrβ) and deiodinase1, and Oxidative Phosphorylation. On the other hand, miRs targeting members of the Nrf2 Oxidative Pathway, Glycolysis, Pentose Phosphate Pathway and Proline biosynthesis – all involved in the metabolic reprogramming displayed by preneoplastic lesions– were up-regulated. Notably, while the expression of most miRs deregulated in preneoplastic lesions was not altered in HCC or in hepatoma cells, miR-182, a miR known to target Dio1 and mitochondrial complexes, was down-deregulated by T3 treatment at all stages of hepatocarcinogenesis and in hepatocarcinoma cell lines. In support to the possible critical role of miR-182 in hepatocarcinogenesis, exogenous expression of this miR significantly impaired the inhibitory effect of T3 on the clonogenic growth capacity of human HCC cells. Conclusions: This work identified several miRNAs, so far never associated to T3. In addition, the precise definition of the miRNA-mRNA networks elicited by T3 treatment gained in this study may provide a better understanding of the key regulatory events underlying the inhibitory effect of T3 on HCC development. In this context, T3-induced down-regulation of miR-182 appears as a promising tool.