Induction-dependent and lineage-dependent models for cell diversification are mutually exclusive.

The purpose of this brief review is to put into perspective just how little is known about the mechanisms that control the assembly of the differentiation program of any cell type. Any number of "trivial" changes in the microenvironment of a Friend erythroleukemic or of a neuroblastoma cell induces both covertly differentiated cells to reveal their lineage affiliations. Demethylating molecules, BudR, retinoic acid, cAMP, butyrate or other "inducing molecules" do not, however, transform the descendents of the neuroblastoma cell into a Hb- synthesizing cell or vice versa. For thousands of generations both of these immortialized lines transmit to their daughters their unique, lineage-dependent differentiation programs with great fidelity. The stability of the inherited transcription complex that is ultimately responsible for this covert differentiation program of these cell lines--or of normal precursor cells--is awesome. Clearly, with these immortalized cells as with normal chick blastodisc cells, the cell's microenvironment plays a major role in permitting or blocking the expression of the cell's inherited differentiation program. But the program itself must be generated by intracellular mechanisms and must be inherited; its assembly is not dependent upon inductive events initiated by exogenous molecules.