In order to investigate the effect of environmentally determined conditions on the cytotoxicity of anticancer treatments, Hoechst 33342 dye selected tumor subpopulations were separated after in vivo treatment and plated for single cell colony survival. The 10% brightest cells were assayed as putative normally oxygenated cells and the 20% dimmest as putative hypoxic cells. At single therapeutic doses, cyclophosphamide treatment resulted in the largest differential killing between bright and dim cells (6.3-fold bright > dim); 1,3-bis(2-chloroethyl)-1-nitrosourea was 3.2-fold more cytotoxic toward bright cells and carboplatin was 2.4-fold more toxic toward bright cells. Both radiation (10 Gy) and melphalan were 2.2-fold more toxic to bright cells, while cis-diamminedichloroplatinum(II) was 1.8-fold, thiotepa was 1.2-fold and procarbazine was 1.3-fold more toxic to bright cells. Actinomycin D was 3.4-fold more toxic to bright cells. Adriamycin was 2.2-fold, vincristine was 2.1-fold, and etoposide was 1.6-fold more toxic to bright cells. Bleomycin and 5-fluorouracil were also tested and were 1.5- and 2.3-fold more toxic to bright cells respectively. Only four treatments were more toxic to dim cells: mitomycin C (3.5-fold), misonidazole (1.5-fold), etanidazole (3.5-fold), and 43°C, 30 min local hyperthermia (2.6-fold). In an attempt to shift the pattern of dim cell sparing, Fluosol-DA plus carbogen (95% O2/5% CO2) breathing was added to treatment with radiation (10 Gy), melphalan, cis-diamminedichloroplatinum(II), and etoposide. Although each of these treatments became significantly more toxic with the addition of Fluosol-DA/carbogen, only with melphalan did the combination overcome the sparing of dim cells. These results indicate that cells located distally from the tumor vasculature are significantly less affected by most anticancer drugs and suggest that successful therapeutic strategies against solid tumors will involve greater use of the few treatments which are more toxic toward this tumor subpopulation.
This work was supported by National Cancer Institute Grant PO1 CA19589 and a grant from Bristol-Myers, Inc., Wallingford, CT.