Limitations of DNA histogram analysis by flow cytometry as a method of predicting chemosensitivity in a rat renal cancer model.

Analysis of DNA histograms obtained from rat renal cancer cells stained with propidium iodide and submitted to flow cytometry revealed a tumor cell population with prominent 2C and 4C peaks and with usually less than 10% each of S-phase or hyper-4C cells. The presence of an increased proportion of 4C cells was found to depend on the age and size of the tumor nodule. Sampling replicate portions of the same tumor or different tumors of the same size and age frequently revealed highly variable 4C:2C ratios. Treatment of animals bearing this tumor with a single i.p. injection of cyclophosphamide (CY), under conditions known to reduce the tumor burden by 80% within 1 week, or with 5-fluorouracil (FUra), which is ineffective against this tumor, in many cases did not yield changes in DNA histograms that permitted one to distinguish the effective drug. Either no marked difference in histogram shape occurred after therapy, or FUra induced more striking differences in cell cycle position than did CY. In tumor generations with greater than 15% S-phase cells, treatment with CY resulted in multiple effects on DNA histograms. These included detecting increased numbers of moribund cells (hypo-2C), a decrease in 4C cells, and an increase in hyper-4C cells. These changes did not occur with the ineffective agent FUra. The tumors grown in vitro show no evidence of replication of 4C cells. The DNA histograms of late-log-phase cultures have a major 2C peak and a minor S plus G2 hump. Since neither the untreated tumor in vivo nor that grown in vitro has a major hyper-4C cell population, it is probable that the tumor stem cells are chiefly 2C (diploid-hyperdiploid). Treatment in vitro of late-log-phase cultures with CY or FUra produces DNA histograms which permit identification of the effective agent. After CY, a major part of the cell population was hypo-2C (moribund) cells.