Anti-cancer drug characterisation using a human cell line panel representing defined types of drug resistance

Abstract

Differential drug response in a human cell line panel representing defined types of cytotoxic drug resistance was measured using the non-clonogenic fluorometric microculture cytotoxicity assay (FMCA). In total 37 drugs were analysed; eight topoisomerase II inhibitors, eight anti-metabolites, eight alkylating agents, eight tubulin-active agents and five compounds with other or unknown mechanisms of action, including one topoisomerase I inhibitor. Correlation analysis of log IC50 values obtained from the panel showed a high degree of similarity among the drugs with a similar mechanism of action. The mean percentage of mechanistically similar drugs included among the ten highest correlations, when each drug was compared with the remaining data set, was 100%, 92%, 88% and 52% for the topoisomerase II inhibitors, alkylators, tubulinactive agents and anti-metabolites respectively. Classification of drugs into the four categories representing different mechanisms of action using a probabilistic neural network (PNN) analysis resulted in 29 (91%) correct predictions. The results indicate the feasibility of using a limited number of cell lines for prediction of mechanism of action of anti-cancer drugs. The present approach may be well suited for initial classification and evaluation of novel anti-cancer drugs and as a potential tool to guide lead compound optimisation.