Integration of yeast phenomics and cancer pharmacogenomics to model precision medicine

Authors

Sean Santos, University of Alabama at Birmingham

Advisory Committee Chair

John L Hartman

Advisory Committee Members

Stephen Barnes

Mary-Ann Bjornsti

Keshav K Singh

Daniel L Smith

Tim M Townes

Document Type

Dissertation

Date of Award

2019

Degree Name by School

Doctor of Philosophy (PhD) Heersink School of Medicine

Abstract

Precision medicine aims to optimize disease treatment by considering the differential influence of functional genetic variation on phenotypic outcomes and therapeutic efficacy. However, current precision medicine paradigms lack consideration of the abundance of genetic interaction and ensuing complexity of phenotypes, thus thwarting resolution of gene-drug interaction at a systems level and ‘precise’ predictions for most patients. Yeast phenomics enables quantitative, high-resolution experimental modeling of gene-drug interaction phenotypes at a systems level by measuring growth curves for the ~6000 yeast knockout/knockdown library strains, which can guide a more global resolution of disease and treatment complexity at the organism level. We postulate that gene-drug interaction is evolutionarily conserved, and thus integration of yeast phenomic analysis with human pharmacogenomics data could help prioritize functional variants likely to influence patient therapeutic responses. We focused on cancer for precision medicine-based application of yeast phenomics due to availability of cancer pharmacogenomics data and the known relevance of yeast genetics. For doxorubicin, we addressed influence of the Warburg effect, termed aerobic glycolysis, by comparing gene-drug interaction in the context of aerobic glycolysis vs. respiration. Our analysis revealed greater cytotoxicity in the context of respiration, suggesting that the Warburg transition to glycolysis can reduce cancer vulnerabilities to chemotherapy. Respiratory-specific influence on doxorubicin included function in chromatin organization, protein folding and modification, and other metabolic processes, while relatively few genes had glycolytic-specific influence. We further used yeast phenomics in a humanized yeast model of deoxycytidine kinase (dCK) to resolve differential gene-drug interaction for gemcitabine and cytarabine. These structurally similar nucleoside analogs are activated by dCK, yet display remarkably different anti-tumor efficacy. We found greater influence on gemcitabine for autophagy, chromatin organization, and apoptotic-metabolism. Several conserved genes, but not enriched biological processes, exerted greater influence on cytarabine cytotoxicity. Correlation of yeast phenomic gene-drug interaction with gene expression and tumor sensitivity data from PharmacoDB for doxorubicin, gemcitabine, and cytarabine supported the concept that evolutionarily conserved gene-drug interaction could help to predict relative anti-cancer efficacy of cytotoxic chemotherapy based on cancer genomic profiles of individual patients’ disease tissue. Thus, we propose yeast phenomics as a methodology to advance systems-based precision medicine.

Recommended Citation

Santos, Sean, "Integration of yeast phenomics and cancer pharmacogenomics to model precision medicine" (2019). All ETDs from UAB. 2899.
https://digitalcommons.library.uab.edu/etd-collection/2899