The purpose of this study was to evaluate to what extent the ability of various chemosensitisers (CS) to reverse P-glycoprotein-associated multidrug resistance (MDR) is reduced when tested in physiological serum protein concentrations. Utilising drug sensitivity and accumulation assays, the CS were tested in medium containing 10% fetal bovine serum and in 100% horse or human serum. Two RPMI 8226 human myeloma sublines were used which express different levels of P-glycoprotein. The CS were tested at various concentrations, including clinically achievable blood levels. When using the CS at high doses, wide differences were observed in the extent CS activity was diminished by serum. Verapamil, cyclosporin A and quinine were not affected, quinidine and medroxyprogesterone acetate were moderately inhibited, and amiodarone and trifluoperazine were largely inactivated. When tile CS were used at concentrations achievable in humans, the activity of all agents except quinine was markedly reduced by serum. With respect to the extent to which CS activity was diminished by serum, good statistical correlation (r > 0.90, P < 0.001) was found between the use of cytotoxicity and drug accumulation assays, horse and human serum or cell lines with high and low levels of P-glycoprotein, respectively. These studies demonstrated that physiological serum protein concentrations can profoundly diminish the MDR reversing activity of particular CS. Some drugs, such as amiodarone and trifluoperazine, are largely inactivated by serum when used at a wide range of concentrations. Other agents, such as verapamil and cyclosporin A, are essentially unaffected when used at high doses but markedly inhibited at concentrations achievable in humans. These data suggest that in vitro studies of CS in medium containing low serum protein concentrations can result in misleading conclusions regarding the potential clinical activity of such agents. Copyright (C) 1996 Elsevier Science Ltd
