In melanoma, patient-derived xenografting (PDX) into mice has permitted testing of the cancer stem cell (CSC) model, yielding conflicting results. However, none of these studies has comprehensively evaluated markers of melanocytic differentiation because of the paucity of such markers that enable separation of more and less differentiated cells for comparison.
We observe within pigmented melanomas extensive heterogeneity in pigmentation; although some cells contain melanin pigment, the canonical marker of melanocytic differentiation, the majority are non-pigmented (NP). We thus devised a novel flow cytometry method to separate melanoma cells according to melanin content. This method, which relies on the near-infrared emission of eu-melanin, allows separation of pigmented (P) and NP melanoma cells from the same tumor (or cell culture) for side-by-side assays.
Using this method, we found that NP cells were far more clonogenic than P cells in B16-F10 mouse melanoma cultures. Further, only NP B16 cells recapitulated the heterogeneous phenotypes of parental cultures. In pigmented PDX melanomas, NP cells exhibited enhanced clonogenicity compared to P cells in the same tumors, although variability existed among melanomas from different patients. This suggests that NP and P cells may be hierarchically related in melanoma cell lines and in some PDX melanomas, consistent with a CSC model, with NP cells primarily driving disease propagation.
Our data suggest that the CSC model may apply to some melanomas, despite contradictory studies of markers less directly related to melanocytic differentiation. However, the abundance of NP cells in most pigmented cell lines and pigmented PDX tumors indicates that the cellular hierarchies of these entities are likely to be flat, such that most melanoma cells would need to be eliminated in therapy. Moreover, it remains to be determined whether the molecular basis of pigment differences between melanoma cells is epigenetic, as required in the CSC model.