Abstract A21: Functional differences among melanoma cells separated according to pigment content

Improved understanding of cancer progression is essential for the development of rational treatment strategies. In tumors that follow a cancer stem cell (CSC) model, targeting only small populations of tumorigenic cells should be sufficient to control disease. In contrast, targeting only some cancer cells will be unsuccessful if every cell in a tumor can convert plastically into another or can acquire genetic changes that enhance tumorigenesis. In these tumors, every cell needs to be eliminated. In melanoma, development of patient-derived xenografting (PDX) of melanomas in immunocompromised mice has permitted testing of the CSC model, yielding conflicting results. Surprisingly, none of these studies has comprehensively evaluated markers of melanocytic differentiation. This is mainly due to the paucity of markers of melanocytic differentiation that enable separation of more and less differentiated cells for side-by-side studies of in vitro clonogenesis and in vivo tumorigenesis. We observe within overtly pigmented melanomas significant heterogeneity in pigmentation of constituent cells; although some cells contain melanin pigment, the canonical marker of melanocytic differentiation, many cells usually a majority are non-pigmented. We have thus devised a novel flow cytometry method to separate melanoma cells according to their content of melanin. This method, which relies on the near-infrared emission properties of eumelanin, allows separation of pigmented (P) and non-pigmented (NP) melanoma cells from the same tumor (or cell culture) for side-by-side clonogenesis assays. Using this method, we surprisingly found that in B16-F10 mouse melanoma cells, NP cells are far more clonogenic than P cells from the same culture. Further, only NP cells recapitulated the heterogeneous phenotypes of the parental culture. In pigmented PDX melanomas, NP cells exhibited enhanced clonogenicity compared to their P cell counterparts in the same tumor, although variability among PDX melanomas was notable. This suggests that NP and P cells may be hierarchically related in pigmented melanoma cell lines and in some pigmented PDX melanomas, consistent with a CSC model, with NP cells primarily driving cell line or tumor propagation. Our data suggest that the CSC model may in fact apply to melanoma, at least in some contexts, despite the contradicting data published so far in studies of markers largely unrelated specifically to melanocytic differentiation. However, the abundance of NP