Myelosuppression is a major dose-limiting factor in cancer chemotherapy. Introduction of drug-resistance genes into bone marrow cells of cancer patients has been proposed to overcome this limitation. In theory, any gene whose expression protects cells against the toxic effects of chemotherapy should be useful in vivo for this purpose. Among such genes, human multidrug-resistance gene (MDR1) and O6-methylguanine DNA methyltransferase gene (MGMT) have been studied most extensively for this purpose, and clinical trials of drug-resistance gene therapy have been started in the US for cancer patients who undergo high-dose chemotherapy with autologous hematopoietic stem cell transplantation. In Japan, our clinical protocol of MDR1 gene therapy, "A clinical study of drug-resistance gene therapy to improve the efficacy and safety of chemotherapy against breast cancer", has been approved by our IRB and submitted to the government. To improve the efficacy and safety of this drug-resistance gene therapy, we have constructed a series of MDR1-bicistronic retrovirus vectors using a retrovirus backbone of Harvey murine sarcoma virus and internal ribosome entry site (IRES) from picornavirus to coexpress a second gene with the MDR1 gene. MDR1-MGMT bicistronic vectors can be used to protect bone marrow cells of cancer patients from combination chemotherapy with MDR1-related anticancer agents and nitrosoureas. In addition, MDR1-bicistronic retrovirus vectors can be designed to use the MDR1 gene as an in vivo selectable marker to enrich the transduced cells which express therapeutic genes, if disease is curable by the expression of a single-peptide gene in bone marrow cells or any types of peripheral blood cells.