Abstract

Multidrug resistance (MDR) and more specifically the expression of P-glycoprotein (Pgp) have been studied extensively in vitro. Unfortunately, it appears that the predictive value of MDR recognized in vitro is mostly an incorrect measure to determine the responsiveness of a particular tumour in the clinic. This misunderstood or overvalued role of MDR might explain the failure of strategies to reverse Pgp function by the use of modulators in solid tumours. To obtain more insight in in vivo drug resistance we investigated a panel of 15 human ovarian cancer xenografts consisting of the most common histological subtypes known in ovarian cancer patients. The response rate to cisplatin, cyclophosphamide and doxorubicin in the xenografts resembled the results of phase II trials with these agents in ovarian cancer patients. This resemblance justifies drug resistance studies in this experimental in vivo human tumour system. We determined the expression levels of MDR 1, MRP 1, LRP and topoisomerase IIalpha mRNA by the RNase protection assay and the presence of MRP1 and LRP proteins by immunohistochemistry. The S-phase fraction was investigated as a separate parameter by flow cytometry. In none of the 15 ovarian cancer xenografts was MDR 1 expression detectable. The expression levels of MRP 1 and LRP were low to moderate and resembled the presence of the MRP1 and LRP proteins. There was a weak, inverse relationship between the expression levels of LRP and sensitivity to cisplatin and cyclophosphamide (r = -0.44 and -0.45), but not to doxorubicin. The levels of topoisomerase IIalpha varied among the xenografts (0.73-2.66) and failed to correlate with doxorubicin resistance (r = 0.14). The S-phase fraction, however, showed a relation with the sensitivity to cisplatin (r = 0.66). Among the determinants studied in ovarian cancer in vivo, LRP mRNA and the S-phase fraction were the best predictive factors for drug response and most specifically for the activity of cisplatin.

Copyright 2000 Cancer Research Campaign.