Abstract 4254: The ratio of key metabolic transcripts is a predictive biomarker of breast cancer metastasis to the lung

Abstract

Abstract Understanding the rewired metabolism of organ-specific metastasis in breast cancer is an under-appreciated problem, with implications for the treatment and prevention of metastatic disease. Here, we used a systems biology approach and the MDA-MB 231 model to compare metabolic fluxes used by parental breast cancer cells and their brain- and lung-homing derivatives. We combined metabolomic profiling, flux measurements, spatial tissue mimetic systems, and mathematical modeling to dissect the cells’ metabolic rewiring, and identified changes specific for lung metastatic selection. We validated our results in other cell lines as well as patient data from project GENIE and the MBCP. Our results produced five biological insights: First, levels of mRNA and metabolic intermediates may anticorrelate with flux. Second, different lineages evolved from the same line can have distinct heritable metabolic fluxes. Metastatic lineages in our model display higher glycolytic flux and bioenergetics than parental cells, with lung-homing cells exhibiting by far the greatest glucose uptake and lactate production. Importantly, this occurs despite low levels of glycolytic intermediates in lung-homing cells. Third, this apparent paradox can be reconciled if feedback inhibition of the glycolysis pathway is prevented, a finding we modeled with flux-balance analysis and confirmed by 13C glucose tracing and measuring glycolytic enzymatic activities directly. Fourth, this distinct metabolic behavior in lung-homing cells is maintained by a high ratio of lactate dehydrogenase (LDH) to pyruvate dehydrogenase (PDH) gene expression, which also correlates with lung metastases in patients with breast cancer. Feature classification models trained on clinical characteristics alone were unable to predict tropism; however, the LDH/PDH ratio was a significant predictor for lung but not brain metastases, independent of other transcriptomic signatures, suggesting that this feature is a potential biomarker for lung metastasis. Fifth, this metabolic effect did not increase cellular growth rate, suggesting that lactate secretion may itself be a trait under selection in breast cancer lung metastasis. Follow up experiments and game theory-based models in a spatially structured tissue-mimetic system showed that lung-homing cells grow more favorably than brain-homing cells in nutrient-deprived gradients, demonstrating how divergent metabolisms of the lineages could lead to selection in different environments. Notably, lung metastases in a mouse model of pancreatic cancer had higher lactate production than other metastases, suggesting that this metabolic trait could be important for other cancer types. Together, our in vitro, in silico, and clinical data analyses highlight that metabolism—currency of all physiological processes—plays an essential role in the connection from gene to phenotype in metastatic disease. Citation Format: Deepti Mathur, Chen Liao, Wendy Lin, Alessandro La Ferlita, Salvatore Alaimo, Alfredo Ferro, Yi Zhong, Christine Iacobuzio-Donahue, Joao Xavier. The ratio of key metabolic transcripts is a predictive biomarker of breast cancer metastasis to the lung. [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 4254.