Elucidation of the mechanism enabling tumor selective PMT in vivo with appropriate glucuronyl-spacer-doxorubicin prodrugs, such as HMR 1826, is important for the design of clinical studies, as well as for the development of more selective drugs. Enzyme histochemistry, immunohistochemistry, and the terminal deoxytransferase technique were applied using human cryopreserved cancer tissues, normal human, monkey, and mouse tissues, and human tumor xenografts to examine mechanisms underlying the selectivity of successful PMT with HMR 1826. It could unambiguously be shown by enzyme histochemistry that necrotic areas in human cancers are the sites in which lysosomal beta-glucuronidase is liberated extracellularly in high local concentrations. The cells responsible for the liberation of the enzyme are mainly acute and chronic inflammatory cells, as shown by IHC. Furthermore, it could be demonstrated that beta-glucuronidase liberated in necrotic areas of tumors can activate HMR 1826, resulting in increased doxorubicin deposition in human tumor xenografts or in human lung cancers subjected to extracorporal perfusion, compared to chemotherapy with doxorubicin. Additionally, the doxorubicin load to normal tissues was significantly reduced compared to chemotherapy with doxorubicin. Surprisingly, the increased doxorubicin deposition in tumors also resulted in strong antitumor effects also in cancers resistant to maximum tolerated doses of systemic doxorubicin. Finally, toxicity studies in mice and monkeys revealed an excellent tolerability of HMR 1826, up to a dose of 3 g/m2 (monkeys). These data suggest that HMR 1826 is a promising candidate for clinical development.