Gene transfer of drug resistance genes. Implications for cancer therapy.

Two general approaches to the gene therapy of cancer have been proposed: (1) strategies that use exogenous genes to modify cancer cells so that they are less malignant or more susceptible to host defenses or to killing by exogenous agents; and (2) approaches that modify host cells so that they are more effective in eliminating cancer cells or more resistant to agents that are used to treat cancer. In both cases, the development of vectors that encode in vivo selectable phenotypes, such as drug resistance, would be extremely valuable because of the inherent inefficiency of gene transfer and the potential of such vectors to protect normal tissues against toxic agents. To allow the selection of cells in vivo that have been transduced with vectors for gene therapy, we have utilized the human multidrug resistance (MDR1) gene. The product of this gene is a 170,000-dalton glycoprotein known as P-glycoprotein, which acts as an energy-dependent efflux pump for a great many cytotoxic anticancer drugs, including doxorubicin, daunorubicin, etoposide, teniposide, actinomycin D, and taxol. Vectors encoding an MDR1 cDNA are able to transduce many cell types, including bone marrow cells, with high efficiency to allow selection of drug resistance in vitro and in vivo in mouse models. Thus, it should be possible to protect the bone marrow of patients undergoing intensive chemotherapy by transduction of their bone marrow with MDR1 vectors. Furthermore, the ability to select for the presence of the MDR1 cDNA in vivo means that it can be used to introduce otherwise nonselectable genes into the bone marrow for therapy of cancer and other diseases.