For over 1,500 years, alchemists struggled to convert base metals into gold. But their search for the legendary 'philosopher's stone' that could make such a transformation resulted in failure and the slide of this pseudoscience into disrepute. Not so in cell biology, however. In the December issue of Nature Cell Biology, David Tosh and colleagues describe an equally startling transition — the conversion of pancreatic cells into hepatocytes, with no intervening cell division.

The interconversion of differentiated cells has been observed before. For example, hepatic foci appear in the rat pancreas after various experimental treatments, and Tosh and colleagues chose to study this process in vitro. They used a synthetic glucocorticoid called dexamethasone to convert a rat pancreatic cell line into hepatocyte foci.

Using an assay based on green-fluorescent protein, the authors conclude that this conversion occurs directly from an exocrine cell type to a hepatocyte — a process known as transdifferentiation. Although they cannot rule out the possibility that an intermediate stem cell is involved, this seems unlikely. Moreover, as the transition is not blocked by 5-bromo-2′-deoxyuridine (a thymine analogue that can be incorporated into DNA), the authors believe that it need not involve cell division.

What, then, is responsible for this transdifferentiation? To find out, Tosh and colleagues studied the expression of transcription factors associated with hepatic differentiation. One such factor, CCAAT-enhancer binding protein β (C/EBPβ), was not detectable in the pancreatic cell line but it could be induced after treatment of these cells with dexamethasone. And, as the figure shows, C/EBPβ could induce the pancreatic cells to transdifferentiate (C/EBPβ is labelled green, with the liver-cell marker glucose-6-phosphatase in red).

Tosh and co-workers confirmed many of their results using an in vitro system that more closely resembles the situation in vivo (pancreatic buds isolated from mouse embryos). Nonetheless, they point out that some findings need further clarification. For example, the authors do not claim that glucocorticoids are involved in the formation of hepatic foci in vivo; only that such treatment probably initiates the same molecular pathway. But in starting to identify some of the molecular elements that may be involved in this pathway, the authors may well have struck gold.