Alternative drug-response metric based on cellular response phenotypes

Abstract

Background and PurposeAssessment of drug response is typically performed by constructing a dose-response curve of viability and summarizing it to a representative value. However, this is limited by its dependency on the assay duration and lack of reflections regarding actual cellular response phenotypes. To resolve these limitations, we considered contribution of each response phenotype upon a drug treatment to the overall growth behavior. Experimental ApproachThe differential equation of phenotype population dynamics was solved analytically without numerical computation. By using the properly figured population dynamics, we explored how the conventional assessment method itself affects the assessment result of drug response, in the context of drug screening. Alternative phenotype metric was compared with the conventional metrics through evaluation of the publicly available drug response data. Key ResultsThe conventional assessment showed several limitations in the comparative analysis of drug response: a significant time-dependency, and ambiguities in assessment results based on a dose-response curve. Instead, the alternative phenotype metrics provide time-independent phenotype rates of change, that contain all the information of the drug response at a given dose, and better classification including the mechanism underlying growth inhibition. Conclusion and ImplicationsThe conventional dose-response curve is useful for a visual presentation of overall drug responses upon a certain molecular feature qualitatively. In contrast, the phenotype metric is better for assessing therapeutic effectiveness, and would improve preclinical pharmacogenomic analysis through its relevance to a response phenotype. Bullet point summaryO_ST_ABSWhat is already knownC_ST_ABSO_LIDrug response is investigated by constructing a dose-response curve in wide range. C_LIO_LIConventional assessment metrics of drug response lack reflections regarding actual cellular response phenotypes. C_LI What this study addsO_LILimitations of conventional assessments are due to time-dependency of dose-response curve and customary summarization. C_LIO_LIPhenotype metric evaluate a single dose-response that is time-independent and phenotype-relevant. C_LI Clinical significanceO_LIPhenotype metric would improve pharmacogenomic analysis with better classification and phenotype-relevance of drug response. C_LIO_LIImprovement in preclinical pharmacology would bring better translation and useful information in clinical studies. C_LI