Background: Low selenium levels are associated with an increased incidence and advanced stage of thyroid cancers (THCAs). In response to changes in selenium levels, a hierarchy of selenoprotein biosynthesis allows tissue-specific fine-tuning of the 25 selenoproteins. To determine the role of individual selenoproteins on thyroid carcinogenesis, we carried out a multiomic data mining study. Methods: The expression levels of individual selenoproteins and their correlations with prognosis in THCAs were analyzed using Oncomine, GEPIA, and Kaplan-Meier plotter platforms. Co-expression analyses using the cBioportal database were carried out to identify genes that are correlated with selenoproteins. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichments were performed for genes correlated with selenoproteins that were identified as clinically significant. Results and Discussion: DIO1, GPX3, SELENOO, SELENOP, SELENOS, and SELENOV were significantly downregulated in THCAs and were associated with poor prognoses. Biological processes including negative regulation of growth and angiogenesis were enriched in DIO1-positively and DIO1-negatively correlated genes, respectively. Many biological processes including negative regulation of growth and MAPK cascade were enriched in GPX3-positively and GPX3-negatively correlated genes, respectively. The antitumor effects of SELENOS might be attributed to their protection against endoplasmic reticulum (ER) stress. SELENOO was revealed to be correlated with ER stress, mitochondrial translation, and telomere maintenance. Biological processes of SELENOV-correlated genes were enriched in redox processes and ER calcium ion homeostasis. Moreover, cell adhesion and angiogenesis were also shown to be negatively regulated by SELENOV, providing an antimetastatic effect similar as DIO1. Conclusion: This study explored the distinct roles of the 25 selenoproteins in THCA pathogenesis, providing potential oncosuppressing effects of 6 selenoproteins.
Keywords: biological processes; multiomic data mining study; pathogenesis; selenoproteins; thyroid cancer.