Screening of co-pathogenic genes of non-alcoholic fatty liver disease and hepatocellular carcinoma

BackgroundNon-alcoholic fatty liver disease (NAFLD) is a risk factor for hepatocellular carcinoma (HCC). However, its carcinogenic mechanism is still unclear, looking for both diseases' transcriptome levels, the same changes as we are looking for NAFLD may provide a potential mechanism of action of HCC. Thus, our study aimed to discover the coexisting pathogenic genes of NAFLD and HCC. MethodsWe performed a variance analysis with public data for both diseases. At the same time, weighted gene correlation network analysis (WGCNA) was used to find highly correlated gene modules in both diseases. The darkturquoise gene module was found to be highly correlated with both diseases. Based on the diagnosis related module genes and the differential genes of the two diseases, we constructed diagnostic and prognostic models by logistic regression, univariate Cox regression, and LASSO regression. Public datasets verified the results. Meanwhile, we built a competing endogenous RNA (ceRNA) network based on the model genes and explored the related pathways and immune correlation involved in the two diseases by using Gene Ontology, Kyoto Encyclopedia of Genes and Genomes, and gene set enrichment analyses. Immunohistochemistry was used to verify the different expression of ABCC5 and TUBG1 among the normal liver, NAFLD, and HCC tissues. Sodium palmitate/sodium oleate was used to establish high-fat cell models, and Real Time Quantitative Polymerase Chain Reaction (RT-qPCR) was used to verify the messenger RNA (mRNA) expression of ABCC5 in lipidization cells. ResultsA total of 26 upregulated genes and 87 downregulated genes were found using limma package identification analysis. According to WGCNA, the darkturquoise gene module was highly correlated with the prognosis of both diseases. The coexisting genes acquired by the two groups were only three central genes, that is, ABCC5, DHODH and TUBG1. The results indicated that the diagnostic and prognostic models constructed by ABCC5 and TUBG1 genes had high accuracy in both diseases. The results of immunohistochemistry showed that ABCC5 and TUBG1 were significantly overexpressed in NAFLD and HCC tissues compared with normal liver tissues. The Oil Red O staining and triglyceride identified the successful construction of HepG2 and LO2 high-fat models using PA/OA. The results of RT-qPCR showed that the lipidization of LO2 and HepG2 increased the mRNA expression of ABCC5. ConclusionsThe gene model constructed by ABCC5 and TUBG1 has hi