N-(4-[2-(1,2,3,4-tetrahydro-6,7-dimethoxy-2-isoquinolinyl)ethyl]- phenyl)-9,10-dihydro-5-methoxy-9-oxo-4-acridine carboxamide (GF120918) has been selected from a chemical program aimed at identifying an optimized inhibitor of multidrug resistance (MDR). The potency of GF120918 is assessed by dose-dependent sensitization of CHRC5, OV1/DXR and MCF7/ADR cells to the cytotoxicity of doxorubicin and vincristine respectively: GF120918 fully reverses multidrug resistance at 0.05 to 0.1 microM and is half maximally active at 0.02 microM. The spectrum of drugs sensitized by GF120918 coincides with those having the classical MDR phenotype. In CHRC5 cells, 0.01-0.1 microM GF120918 enhances the uptake of [3H]daunorubicin and blocks the efflux from preloaded cells. It is also shown that GF120918 is still active several hours after being taken away from the culture medium showing that it is not, like verapamil, effluxed rapidly by P-glycoprotein. GF120918 effectively competes with [3H]azidopine for binding P-glycoprotein, pointing to this transport membrane protein as its likely site of action. After i.v. administration to mice, GF120918 penetrates thoroughly various organs that have a tissue level/blood level ratio above 10. It is eliminated from organs and blood with a half-time of approximately 2.7 h. It is well absorbed after p.o. administration. In mice implanted i.p. with the MDR P388/Dox tumor, a single i.v. or p.o. dose of GF120918 restores sensitivity of the tumor to a single i.p. dose (5 mg/kg) of doxorubicin administered 1 h later. A statistically significant effect is observed at 1 mg/kg GF120918 i.v. and maximal effect is reached at 5 mg/kg. Similarly, whereas neither drug alone is effective, GF120918 (10 mg/kg i.p.) associated with doxorubicin (5 mg/kg i.p.) inhibits the growth of the moderately MDR C26 tumor implanted s.c. as assessed by tumor size at day 19. GF120918 does not modify significantly the distribution or the elimination of doxorubicin in mice ruling out the possibility that the antitumor effects seen might be explained by pharmacokinetic interactions.