Various cholestatic liver diseases are accompanied by a striking increase in the number of bile ductules. This so-called ductular reaction is thought to arise both from ductular metaplasia of hepatocytes and from proliferation of pre-existing bile ductules. Previous studies have shown that these reactive bile ductules differ from their normal counterpart in enzyme and immunohistochemical make-up. Using monoclonal antibodies directed to neuroendocrine markers and immunohistochemistry, we found that reactive bile ductules in cholestatic liver disease display neuroendocrine features. In all cases of primary biliary cirrhosis (PBC), primary sclerosing cholangitis (PSC), extrahepatic obstruction, and acute hepatitis A, reactive bile ductules expressed the neural cell adhesion molecule (N-CAM) and reacted with monoclonal antibody A2B5. Both N-CAM and the ganglioside, recognized by A2B5, are restricted to neuroendocrine cells and tissues. In all but four of these cases, the same bile ductules expressed chromogranin-A, present in the matrix of neuroendocrine granules. Furthermore, in three cases of longstanding cholestasis, scattered periportal hepatocytes expressed chromogranin-A but not N-CAM. Other neuroendocrine markers such as neuron-specific enolase, synaptophysin, or myelin-associated glycoprotein were lacking from both bile ductules and hepatocytes. The neuroendocrine phenotype of bile ductules and hepatocytes was confirmed on electronmicroscopy, demonstrating various numbers of dense-cored, neuroendocrine granules near the peripheral cell membrane in bile ductules as well as in cells intermediate between hepatocytes and bile ductular cells. In 10 cases of normal liver tissue without ductular reaction, bile ductules lacked neuroendocrine markers except in two cases in which very weak reactivity for chromogranin-A was observed. These findings illustrate the presence of a new, neuroendocrine cell type that emerges in the liver during cholestasis. Elucidation of the significance of the neuroendocrine substance(s) produced in the dense cored granules of reactive bile ductules awaits their isolation and characterization. We can speculate that this molecule plays an autocrine or paracrine regulatory role in the process of ductular metaplasia of hepatocytes or growth of bile ductules.