Comprehensive analysis of mRNA-miRNA network in preneoplastic hepatocytes undergoing regression following treatment with thyroid hormone.

Background: Hepatocellular carcinoma (HCC) is the third leading cause of cancer-related deaths worldwide. Despite the approval of new molecular targeted therapies, benefits from current therapies remain unsatisfactory. Thus, it is mandatory to find new and effective therapeutic treatment strategies to treat HCC. Recent studies revealed a status of severe local hypothyroidism in rat hepatic preneoplastic lesions, and in human HCCs, suggesting that this condition may represent a favorable event for HCC development. Accordingly, a short treatment with thyroid hormone (T3) caused the regression of pre-neoplastic lesions in a rat model of hepatocarcinogenesis. Thyroid hormones (THs) inhibit HCC through different mechanisms. To investigate whether microRNAs (miRs) play a role in the antitumorigenic effect of THs, by Next Generation Sequencing (NGS) we performed a comprehensive comparative miRNomic and transcriptomic analysis of hepatic preneoplastic lesions of rats subjected to the Resistant-Hepatocyte (RH) protocol and then exposed or not to a four-day treatment with triiodothyronine (T3). The most deregulated miRs were also analysed in HCCs generated by the same protocol, with or without T3. Analysis of the transcriptomics data can be tricky, and often requires substantial expertise in bioinformatics. Prior to the final analysis, the raw data needs to be processed through number of steps, thus resulting in a gene or miRNA expression quantification. Bioinformatic workflows are generally employed to automate these steps. A number of workflows are available to help the researchers preprocess their data. However, many existing workflows have their own approach towards the analysis, or they are limited to specific studies. Here, we have implemented two automated, scalable, and reproducible Snakemake based pipelines called RIDE and SRIDE for performing transcriptome and microRNA analysis respectively. Result: Pathway analysis in preneoplastic nodules showed that Oxidative Phosphorylation and NRF2-mediated Oxidative Stress Response were the most deregulated pathways. These pathways were resulted to be reversed in preneoplastic nodules exposed to T3. Gene expression analysis showed that T3 also inhibited proline biosynthesis, which was up-regulated in preneoplastic lesions. With regard to microRNA, most downregulated miRNA in preneoplastic lesions have already been described to act as oncosuppressors, whereas, some of the differentially up-regulated miRNAs have been shown to contribute substantially in carcinogenesis process. miRNA’s down-regulated in preneoplastic nodules following T3 treatment targeted Thyroid hormone receptor- (THR), deiodinase, and oxidative phosphorylation pathways, while on the contrary, most upregulated miRNAs were those of targeting Nrf2 Oxidative Pathway, Glycolysis, Pentose Phosphate Pathway and Proline biosynthesis – all involved in metabolic reprogramming. Notably, while T3 exerted similar effects on gene expression in both preneoplastic nodules and HCCs, no change in the expression of these miRNAs regulated in preneoplastic lesions was found at the late stages. This study investigated the miRNA-mRNA networks elicited by T3 treatment in liver preneoplastic lesions. The identification of several miRNAs, so far never associated to T3, may improve our understanding of the key regulator events underlying the early stages of HCC development and help to design therapeutic strategies.