Treatment-induced changes in sensitivity in a multiclonal human tumor mixture model in vitro.

An in vitro model has been devised so that mixtures of human tumor cells can be grown together for studies related to drug-induced or -selected changes in sensitivity. In the studies reported here, two human astrocytoma clones, one sensitive and one resistant to 1-(2-chloroethyl)-3-(4-methylcyclohexyl)-1-nitrosourea (MeCCNU), were carefully matched for doubling times, cell cycle phase distributions, and colony-forming efficiencies. The clones were mixed and grown together, and after only three weekly treatments with MeCCNU (10 micrograms/ml for 1 h each week) the sensitive cells in the mixture were killed, leaving behind a population that was almost 100% resistant to further exposures to MeCCNU. The loss of the sensitive cells from the mixture each week was easily detected by visual observation of flow microfluorometry histograms since the clones had different DNA indices. Repeated weekly exposures of the unmixed resistant clone (AST 1-1) to MeCCNU caused very little accumulated cell kill. Similar exposures of the unmixed sensitive clone (AST 3-4) produced a linear decrease in survival over the first three weekly treatments with 10 micrograms MeCCNU/ml, but after that time these cells became progressively more resistant to MeCCNU. It is unlikely that the change to resistance in the AST 3-4 clone occurred because of contamination with the resistant AST 1-1 cells, because their DNA index remained stable. These data show that repeated treatments with a single agent can cause a tumor cell population to become more resistant. It remains to be tested whether this resistance was the result of cellular interactions, drug-induced changes in sensitivity, or selection for resistant cells already present in the populations. This mixture model may be useful in studies on how cellular interactions influence growth and drug sensitivity in tumor and normal cell populations.