Fig. 6. The effect of EDNRB signaling on melanocyte development both cell-autonomously and cell non-autonomously. Both melanoblasts and non-melanogenic NT or neural crest cells express Ednrb. As the addition of Ednrb^+/+ NTs can induce tyrosinase expression in Ednrb-deficient melanoblasts, we postulate that in response to EDN3, these Ednrb^+/+ NTs or their neural crest derivatives produce one or several factors that indirectly help melanoblasts to differentiate. As these rescuing Ednrb^+/+ NTs need not be capable of generating melanocytes themselves but must be capable of providing KITL, the helper cell type(s) can be non-melanocytic but must provide KITL as a major helper factor. It is likely that KITL acts directly on melanoblasts even though an indirect action involving yet other cell types cannot be excluded. Nevertheless, even though the cell non-autonomous action of Ednrb can rescue Ednrb mutant cells to the tyrosinase-positive stage, it is insufficient to induce pigmentation. Pigmentation can be seen, however, if the rescued cells are given TPA which induces PKC as does EDNRB signaling. Thus, terminal differentiation to mature, pigmented melanocytes probably depends on an additional, cell-autonomous EDNRB signaling step. Note that the above scenario does not preclude the possibility that in wild-type melanoblasts, Ednrb can also act cell-autonomously throughout development.