Multidrug-resistant (MDR) cell lines often have a compound phenotype, combining reduced drug accumulation with a decrease in topoisomerase II. We have analysed alterations in topoisomerase II in MDR derivatives of the human lung cancer cell line SW-1573. Selection with doxorubicin frequently resulted in reduced topo II alpha mRNA and protein levels, whereas clones selected with vincristine showed normal levels of topo II alpha. No alterations of topo II beta levels were detected. To determine the contribution of topo II alterations to drug resistance, topo II activity was analysed by the determination of DNA breaks induced by the topo II-inhibiting drug 4'-(9-acridinylamino)methane-sulphon-m-anisidide (m-AMSA) in living cells, as m-AMSA is not affected by the drug efflux mechanism in the SW-1573 cells. The number of m-AMSA-induced DNA breaks correlated well (r = 0.96) with in vitro m-AMSA sensitivity. Drug sensitivity, however, did not always correlate with reduced topo II mRNA or protein levels. In one of the five doxorubicin-selected clones m-AMSA resistance and a reduction in m-AMSA-induced DNA breaks were found in the absence of reduced topo II protein levels. Therefore, we assume that post-translational modifications of topo II also contribute to drug resistance in SW-1573 cells. These results suggest that methods that detect quantitative as well as qualitative alterations of topo II should be used to predict the responsiveness of tumours to cytotoxic agents. The assay we used, which measures DNA breaks as an end point of topo II activity, could be a good candidate.