Juvenile polyposis syndrome (JPS; Online Mendelian Inheritance in Man2 174900) is a rare Mendelian disorder in which individuals have typical hamartomatous polyps within the gastrointestinal tract. The stromal element of the polyps has classically been thought to be the proliferative component, although epithelial malignancies (largely gastrointestinal cancers) occur more frequently than expected in JPS patients. Germ-line mutations in SMAD4 (DPC4) account for about a third of JPS cases. It has been postulated that the apparent paradox of a stromal lesion predisposing to epithelial malignancy can be resolved by the "landscaper" effect: an abnormal stromal environment affects the development of adjacent epithelial cells, and the resulting regeneration of damaged epithelium leads to an increased risk of cancer. We have found allele loss at the SMAD4 locus on 18q in polyps from JPS individuals with a germ-line SMAD4 mutation, showing that SMAD4 is acting as a tumor suppressor gene in JPS polyps, as it does in sporadic cancers of the gastrointestinal tract. Interphase fluorescence in situ hybridization showed deletion of one copy of SMAD4 in the epithelial component of JPS polyps, but not in the inflammatory infiltrate. Fluorescence in situ hybridization also suggested that a single copy of SMAD4 was present in stromal fibroblasts of JPS polyps. Thus, biallelic inactivation of SMAD4 occurs in both the epithelium and some of the stromal cells in these lesions, suggesting a common clonal origin. Epithelial malignancies almost certainly develop in juvenile polyposis through direct malignant progression of the epithelial component of the hamartomas. SMAD4/DPC4 probably acts as a "gatekeeper" tumor suppressor in juvenile polyps, and there is no need to invoke a "landscaper hypothesis."