Patterns of tumor colony development over time in soft-agar culture.

Human tumors were cultured by the two-layer soft-agar technique and the time course of tumor colony development was evaluated during periods of up to 6 weeks in culture. All colony counting was performed with an automated tumor colony counter (Omnicon; Bausch and Lomb, Inc, Rochester, NY, USA). This instrument provided colony counts per culture plate in six size categories from greater than 60 microns diameter colonies to greater than 149 microns diameter colonies. Six to 24 culture plates were used for each "growth curve", generally 24. Control (non-drug-treated) cultures were obtained from 117 tumors, of which 25 also provided enough cells to allow evaluation of the time course of colony development after exposure to cytostatic agents. The development of colonies in non-drug-treated plates usually demonstrated a lag phase, a logarithmic growth phase to maximum colony development and a subsequent deterioration of colonies. In spite of clumps seeded into the agar, real colony growth could be recognized by frequent colony counting of culture dishes, although the temporal patterns of growth were sometimes different if pure single-cell suspensions were compared with suspensions containing clumps from the same tumor. Drug pre-incubation caused changes in the temporal pattern of colony growth as well as in the total number of colonies. Some cultures showed drug sensitivity when evaluated at certain time points while evaluation at later time points showed only borderline drug effect or none at all. The potential utility of tumor colony growth curves in the clinical applications of tumor colony cultures is discussed.