Cholangiocytes, the epithelial cells that line intrahepatic bile ducts, participate in bile secretion via basal and agonist-stimulated transport of solutes and water. On the basis of subtle structural differences between cholangiocytes lining small vs. large bile ducts, as well as known phenotypic variations among transporting epithelia in other organs, we demonstrated that cholangiocytes are functionally heterogeneous along the intrahepatic biliary tree of normal rats. In studies reported here, we confirm and extend the concept of functional heterogeneity of cholangiocytes by employing the bile duct-ligated (BDL) rat model of cholestasis associated with selective cholangiocyte proliferation. Using novel isolation and separatory techniques, we prepared subpopulations of pure small, medium, and large cholangiocytes from BDL rats and compared them with regard to gene expression and basal or agonist-responsive transport activities. Although transcripts for gamma-glutamyl transpeptidase and cytokeratin 19, two cholangiocyte-specific proteins, and glyceraldehyde-3-phosphate dehydrogenase, a housekeeping gene, were in all three subpopulations, genes for several proteins involved in solute transport [Cl-/HCO3- exchanger, cystic fibrosis transmembrane conductance regulator (CFTR), and secretin receptor] were expressed only in medium and large cholangiocytes. Consistent with these findings, secretin increased intracellular levels of adenosine 3',5'-cyclic monophosphate (cAMP) and 36Cl- efflux rates in medium and large cholangiocytes but not in small cholangiocytes. Also, forskolin/8-(4-chlorophenylthio)-cAMP stimulated 36Cl- efflux rates only in medium and large cholangiocytes, consistent with selective functional expression of CFTR in these subpopulations. These results support the molecular and functional heterogeneity of cholangiocytes within the intrahepatic biliary ductal system and are consistent with the notion that hormone-regulated transport of solutes after BDL occurs principally in medium and large cholangiocytes in a fashion similar to that observed in normal rat liver.

Secretion stimulates ductular bile secretion by binding to receptors on intrahepatic bile duct epithelial cells (i.e., cholangiocytes). In the rat, this choleretic effect increases after bile duct ligation (BDL). Although cholangiocyte proliferation induced by BDL contributes to secretin-induced hypercholeresis, the mechanisms modulating these alterations in secretin-induced ductular bile secretion are obscure. Thus we studied the expression of secretin receptor mRNA (SR-mRNA) in purified liver cells from normal and BDL rats. Northern blot analysis and RNase protection assays with mRNA from purified liver cells demonstrated SR-mRNA only in cholangiocytes; moreover, SR gene expression showed a seven- to ninefold increase in individual cholangiocytes from BDL rats compared with controls. This increase in SR-mRNA expression was related to a similar increase in the rate of transcription of SR-mRNA in cholangiocytes from BDL rats. Thus our studies indicate that 1) SR-mRNA is detected in liver only in cholangiocytes; 2) BDL causes an increase in SR-mRNA in individual cholangiocytes; and 3) the increase in SR-mRNA after BDL is partly related to an increase in the rate of transcription of SR-mRNA by cholangiocytes after BDL. Our data suggest that upregulation of the SR gene may contribute to secretin-induced hypercholeresis.

The development of new and refined separation techniques--including FACS, FFE, CFE and isopyknic gradients--has had a profound impact on the ability of investigators to isolate specific cell types from the liver. Although some of these techniques, such as FFE, may be of limited preparative value, they are nonetheless important analytical tools that detect subtle differences among cell subpopulations. The isolation of highly purified preparations of liver cells in large yields requires the use of more conventional purification methods such as CFE and isopyknic centrifugation. Immunological approaches represent a key development for the isolation of specific liver cell types, especially when they are used in combination with other techniques. Excellent, reliable and relatively simple techniques now exist to isolate highly purified preparations of hepatocytes, cholangiocytes, KCs, SCs, FSC, myofibroblasts and pit cells. Additional work is necessary to refine techniques for the isolation of dendritic cells and lymphocytes.

We have previously demonstrated that tamoxifen inhibits the growth of human cholangiocarcinoma cells in culture and inhibits tumor growth when cells are injected into nude mice. However, the mechanism of action of tamoxifen remains unknown. Here we demonstrate that tamoxifen and trifluoperazine, both potent calmodulin antagonists, induce apoptosis in vitro, probably acting via the Fas system, in human cholangiocarcinoma cells. Human cholangiocarcinoma cell lines heterogeneously express Fas antigen on their surface. Fas-negative and Fas-positive surface-expressing cells were isolated, cloned, and cultured. Fas antibody, tamoxifen, and trifluoperazine induced dose-dependent apoptosis only in Fas-positive cells; Fas-negative cells were unaffected. Furthermore, apoptosis induced by tamoxifen in Fas-positive cells was blocked by an inhibitory Fas antibody. Tamoxifen was not acting through an anti-estrogenic mechanism, because neither Fas-negative nor Fas-positive cells expressed estrogen receptors and the pure anti-estrogen compound, ICI 182780, did not induce apoptosis in either cell line. Fas-negative cells, but not Fas-positive cells, were able to produce tumors when subcutaneously injected into nude mice. These findings suggest Fas may be a candidate oncogene involved in the pathogenesis of cholangiocarcinoma. Furthermore, the similarity between the pro-apoptotic effects of tamoxifen and trifluoperazine support an underlying molecular mechanism for Fas-mediated apoptosis that involves calmodulin.

No abstract