Applying Differential Gene Set Enrichment Analysis to Elucidate Metabolic Vulnerabilities in Tumor Cells Under Oxidative Stress

Introduction

In MCF-10A human mammary epithelial cells, expression of a constitutively active mutant AKT kinase (mAKT) renders cells sensitive to disruption of glycolysis. Specifically, when mAKT-MCF-10A cells are switched from normal glucose to galactose, mAKT-MCF-10A cells die for ~3 passages (15 days).

Differential Gene Set Enrichment Analysis (DGSEA) is a novel bioinformatic technique developed by the Graham lab that measures how two metabolic pathways are altered relative to each other. Because metabolic adaptations often involve tradeoffs between two different metabolic pathways, DGSEA may lead to novel insights compared to existing GSEA methods. Here, DGSEA can be used to analyze how metabolic pathways are regulated in MCF-10A cells expressing different oncogenes when cells are cultured in oxidative culture.

Methods

We used previously collected metabolomic data from MCF-10A cells expressing red fluorescent protein (negative control) or activated AKT (mAKT) cultured in glucose and short-term galactose for 24h. We ranked intracellular metabolite pool sizes by log₂ fold change between the short-term galactose cells and long-term glucose cells. Next, we used DGSEA to identify pairs of metabolic pathways that are differentially regulated. Large changes in pairs of metabolic pathways were indicated by large, normalized enrichment scores (NES). The p-values and false discovery rates (FDR) were estimated by permutation analysis. The criteria used to determine significant metabolic pathway pairs was a p-value < 0.05 and a false discovery rate < 0.25.

Results and Conclusions

Significant metabolic pathways pairs comparing mAKT-MCF-10A cells placed in short term galactose culture to those placed in glucose culture were analyzed using DGSEA. Four significant pathway pairs were found: pyrimidine and nicotinate/nicotinamide, pyrimidine and butanoate, pyrimidine and glycolysis/gluconeogenesis, and lastly pyrimidine and glutathione. For these pathways, pyrimidine metabolism was the downregulated pathway when mAKT cells were placed in galactose culture. These other paired pathways present potential tradeoff checkpoints that the cell may elect to prefer at the expense of pyrimidine nucleotide metabolism in times of oxidative stress.

Discussion

Our finding that pyrimidine is downregulated when tumor cells are under oxidative stress induced by being placed in galactose culture corroborates findings within the field of metabolomics suggesting that cancer cells rely on increased pyrimidine and nucleotide synthesis pathways to fuel tumor growth. The findings of this project suggest a possible mechanism that oxidative stress induced by the galactose culture led mAKT cells to downregulate pyrimidine synthesis.