The direct effect of continuous exposure to prostaglandins on the cloning efficiency and proliferative capacity of human malignant melanoma colony-forming cells in soft agar was evaluated. Prostaglandin A1 (PGA1) and prostaglandin E1 (PGE1) effected a dose-dependent inhibition of colony formation and proliferative capacity. PGA1 at a concentration of 5 microgram/ml reduced colony formation of cells from human melanoma cell strains C8054, C8130, and C822 by at least 85%. PGA1 also inhibited colony formation of cells obtained directly from biopsies of melanoma tissues from eight patients by greater than 70% at a concentration of 5 microgram/ml. A steep dose-response curve was evident by the little effect of PGA1 on colony formation at a concentration of 0.5 microgram/ml. The mean 50% inhibitory doses for PGA1 and PGE1 were 1.25 and 4.25 microgram/ml, respectively. Prostaglandin A2 was much less effective than PGA1 in inhibiting melanoma colony formation. The related prostaglandins (prostaglandin B1, prostaglandin F1 alpha, and prostaglandin E2 alpha) had little or no effect on colony formation. Overall, these results suggested that the presence of a carbonyl group at position 9 of the cyclopentane ring may be required for inhibitory activity as prostaglandins of the A and E series inhibited human melanoma cell growth. PGA1 and PGE1 did not effect a rise in cyclic adenosine 3':5'-monophosphate levels in C8054 and C8130 cells. However, while alpha-melanocyte-stimulating hormone and prostaglandin F2 alpha did generate a rise in adenosine 3':5'-monophosphate levels in C8054 cells, these hormones had no effect on colony formation. These results are consistent with the notion that the PGA1 and PGE1 inhibition of melanoma colony-forming cells occurs via a noncyclic nucleotide mechanism.

An acid-sensitive growth factor for human melanoma has been partially purified from human platelets. TSK gel filtration HPLC provides a molecular weight estimation of 60,000 daltons. This factor is not only mitogenic for human melanoma, but also stimulates 3H-thymidine uptake and increases the number of bovine aortic endothelial cells, while fibroblasts are nonresponsive. Radioiodination of active HPLC fraction has been accomplished. The human melanoma cell line, M1RW5 demonstrates specific, time-dependent binding of the labeled protein. These studies suggest that the growth of human melanoma may in part be regulated by a newly described growth factor present in human platelets.

A human platelet sonicate was evaluated for its effects on the growth of human metastatic melanoma colony-forming cells in soft agar from cells in culture and from biopsies. The addition of platelet sonicate increased both cloning efficiency and proliferative capacity in that more and larger colonies were formed. In more detailed studies under growth-limiting conditions, melanoma cellular responses to known growth factors were compared to the activity found in the platelet sonicate. None of the growth factors tested either alone or in combination, including platelet-derived growth factor, epidermal growth factor, alpha-type transforming growth factor, and beta-type transforming growth factor, were capable of inducing melanoma colony formation to the 12-fold stimulation observed with the platelet sonicate. Treatment of platelet sonicate with dithiothreitol, trypsin, or acid resulted in loss of activity for human melanoma. Our results suggest that human platelets contain an acid-sensitive protein which can support the expression of the transformed phenotype of human melanoma, and this factor is distinct from acid-stable activities previously characterized from human platelets.

The radiation response of murine and human melanoma cells assayed in bilayer soft agar and monolayer was examined. Cells from the murine melanoma Cloudman S91 CCL 53.1 cell line and three human melanoma cell strains (C8146C, C8161, and R83-4) developed in our laboratory were irradiated by single dose X-rays and plated either in agar or on plastic. D0 values were the same within 95% confidence intervals for cells from the human melanoma cell strains C8146C, C8161, and R83-4 but were dissimilar for the murine cell line CCL 53.1 Dq values were different for all cells studied. The shape of the survival curve for all four melanomas was not identical for cells assayed in soft agar versus cells grown on plastic. This would indicate that apparent radiosensitivity was influenced by the method of assay although there were no apparent consistent differences between the curves generated by monolayer or bilayer soft agar assays.