Barrett's esophagus is a complex metaplasia which can occur as a pure gastric or intestinal histopathological phenotype but more commonly occurs as a mosaic comprising gastric and intestinal forms. Intestinal metaplasia is typified by the presence of goblet cells seen on light microscopy. Some studies indicate that the chances of identifying goblet cells are related to the length of Barrett's, location of the biopsies (more goblet cells in proximal esophagus), number of biopsies, patient age and gender, being more common in men [1]. Most interest has focussed on those patients with an endoscopically visible columnar lined epithelium containing intestinal metaplasia since this is the form most clearly associated with malignancy with an annual progression rate estimated to be in the region of 0.3-0.4% per annum [2,3]. However, recent studies comparing goblet and non-goblet containing metaplastic epithelium have shown that both types demonstrate the same frequency of DNA content abnormalities [4,5], suggesting that the gastric form is not entirely benign.The molecular and cell biological process by which Barrett's esophagus arises is poorly understood. The emerging theories include trans-differentiation from a mature esophageal keratinocyte or altered differentiation of an esophageal or submucosal stem cell [6]. More recently it has been suggested that Barrett's may occur from expansion of an embryological remnant [7]. Whichever theory is correct, all appear to depend on the activation of transcription factors to alter the cell fate. To date most effort has focussed on the homeobox genes Cdx1 and 2 which, in contrast to the dramatic effect observed when ectopically expressed in the murine stomach [8], fail to induce intestinalisation when over-expressed in the esophagus. Other transcription factors examined have included Hedgehog [9] and another critical signalling molecule in embryonic development called BMP4 [10], which have generally failed to induce phenotypic intestinalisation and in particular goblet cells.The aim of this paper was therefore to explore whether the helix-loop-helix transcription factor Math1/Atoh1, which is required for the differentiation of the three secretory cell lineages in the intestine (enteroendocrine, Paneth and goblet cells), could induce intestinalisation of an esophageal cell line. In order to explore this Kong et al. [11] first confirmed earlier reports that this gene is over-expressed in Barrett's esophageal tissues [12]. RNA expression levels in endoscopic biopsies were found to be at a highly variable level (from 10-to 1,000-fold), but the levels were always in excess of that found in paired normal esophageal biopsies. It is unfortunate that no matching histology was available for these Barrett's samples, since as noted in the discussion, it is likely that this variation in RNA expression levels correlated with the goblet cell numbers which are known to be heterogeneous in this epithelium. This is a methodological point worth noting since many studies in the Barrett's field are...