The objective of this study was to use the established xenograft model of human melanoma (C8161.9) to test a pharmacological approach to the effect of the metastasis suppressor KISS1. A KISS1 analog was used to inhibit the metastatic development of C8161.9 cells in nude mice. Further experiments were performed to test the validity of the C8161.9 model and test the connection between KISS1 expression and loss of metastatic potential. New clones of C8161.9 cells were obtained, with or without KISS1 expression, and were tested for metastasis formation. The absence of benefit in survival with the KISS1 analog compared with PBS prompted us to revisit the C8161.9 model. We found that the cells expressing KISS1, used in the previous study and obtained by transfection and single-cell cloning, were defective for both formation of orthotopic tumors and metastases. In mixing experiments, these cells could not suppress orthotopic tumor growth of KISS1-negative C8161.9 cells, suggesting that the suppression of metastasis by C8161.9-KISS1 cells may be intrinsic to the selected clone rather than related to KISS1 expression. Isolation of clones from parental C8161.9 cells in soft agar yielded cell populations that phenotypically and genotypically mimicked the KISS1-positive clone. In addition, new clones expressing KISS1 did not show any decrease in metastatic growth. These data demonstrate the heterogeneity of cell types in the C8161.9 cell line and the high risk of artifact linked to single-cell selection. A different xenograft model will be necessary to evaluate the use of KISS1 analogs as antimetastatic therapy.