Gene Set Enrichment Analysis of Selenium-Deficient and High-Selenium Rat Liver Transcript Expression and Comparison...

Better biomarkers of selenium (Se) status and a better understanding of toxic Se biochemistry are needed to set safe dietary upper limits. In previous studies, differential expression (DE) of individual liver transcripts in rats and turkeys failed to identify a single transcript that was consistently and significantly (q < 0.05) altered by high Se.

To evaluate the effect of Se status on rat liver transcript expression data at the level of gene sets, and to compare transcript expression in rats with that in turkeys to identify common regulated transcripts.

Gene set enrichment analysis (GSEA) was conducted on liver from weanling rats fed an Se-deficient basal diet (0.005 μg Se/g) supplemented with 0, 0.24 (Se-adequate), 2, or 5 μg Se/g diet as selenite for 28 d. In addition, transcript expression was compared with liver expression in turkeys fed 0, 0.4, 2, or 5 μg Se/g diet as selenite.

Se deficiency significantly downregulated the rat selenoprotein gene set but also upregulated gene sets for a variety of pathways, processes, and disease states. GSEA of 2 compared with 0.24 μg Se/g found no significantly up- or downregulated gene sets, showing that 2 μg Se/g is not particularly toxic to the rat. GSEA analysis of 5 compared with 0.24 μg Se/g transcripts, however, found 27 significantly upregulated gene sets for a wide variety of conditions. Cross-species GSEA comparison of transcript expression, however, identified no common gene sets significantly and consistently regulated by high Se in rats and turkeys. In addition, comparison of individual marginally significant (unadjusted P < 0.05) DE transcripts between rats and turkeys also failed to find common transcripts.

The dramatic increase in significant liver transcript DE and GSEA gene sets in rats fed 5 compared with 2 μg Se/g clearly appears to be a biomarker for Se toxicity, albeit not Se-specific. These analyses, however, failed to identify specific transcripts or pathways, biological states, or processes that were directly linked with high Se status, strongly indicating that adaptation to high Se lies outside transcriptional regulation.