Tumor cell heterogeneity: Divided-colony assay for measuring drug response

In vitro tests for predicting the response of tumors to chemotherapeutic agents might be improved if they were modified to take into account tumor-cell heterogeneity. We have studied the heterogeneity of cellular growth rate and drug response in mouse fibroblast NIH 3T3 cells and in NIH 3T3 cells transformed with the human HRAS gene (homologue of the Harvey sarcoma virus oncogene v-Ha-ras) from the EJ human bladder carcinoma cell line. Growth-rate heterogeneity was detected as a broad distribution of numbers of cells per colony. In spite of this heterogeneity, secondary colonies have numbers of cells per colony that resemble that of the primary colony from which they were derived. The variance between unrelated secondary colonies is increased by HRAS(EJ). Colony-size measurements are reliable because primary colonies divided in half formed two groups of secondary colonies (on two separate plates) that have indistinguishable mean colony sizes. Based on these observations, a divided-colony procedure was devised to detect the drug response of heterogeneous cell populations. Primary colonies are divided into two groups of cells, one of which is treated with a drug and the other is left untreated as a control. The size distribution of treated secondary colonies is then compared to that of the untreated control and to that of the primary colony from which it was derived. The divided-colony procedure is proposed as a modification of the human-tumor-cloning system to increase the sensitivity and reliability of in vitro procedures used to determine the drug response of heterogeneous tumor-cell populations.