The most frequent drawback of doxorubicin is the onset of drug resistance, due to the active efflux through P-glycoprotein (Pgp). Recently formulations of liposome-encapsulated doxorubicin have been approved for the treatment of tumors resistant to conventional anticancer drugs, but the molecular basis of their efficacy is not known. To clarify by which mechanisms the liposome-encapsulated doxorubicin is effective in drug-resistant cancer cells, we analyzed the effects of doxorubicin and doxorubicin-containing anionic liposomal nanoparticles ("Lipodox") on the drug-sensitive human colon cancer HT29 cells and on the drug-resistant HT29-dx cells. Interestingly, we did not detect any difference in drug accumulation and toxicity between free doxorubicin and Lipodox in HT29 cells, but Lipodox was significantly more effective than doxorubicin in HT29-dx cells, which are rich in Pgp. This effect was lost in HT29-dx cells silenced for Pgp and acquired by HT29 cells overexpressing Pgp. Lipodox was less extruded by Pgp than doxorubicin and inhibited the pump activity. This inhibition was due to a double effect: the liposome shell per se altered the composition of rafts in resistant cells and decreased the lipid raft-associated amount of Pgp, and the doxorubicin-loaded liposomes directly impaired transport and ATPase activity of Pgp. The efficacy of Lipodox was not increased by verapamil and cyclosporin A and was underwent interference by colchicine. Binding assays revealed that Lipodox competed with verapamil for binding Pgp and hampered the interaction of colchicine with this transporter. Site-directed mutagenesis experiments demonstrated that glycine 185 is a critical residue for the direct inhibitory effect of Lipodox on Pgp. Our work describes novel properties of liposomal doxorubicin, investigating the molecular bases that make this formulation an inhibitor of Pgp activity and a vehicle particularly indicated against drug-resistant tumors.