Development and molecular composition of the hepatic progenitor cell niche

End-stage liver diseases represent major health problems that are currently treated by liver transplantation. However, given the world-wide shortage of donor livers novel strategies are needed for therapeutic treatment. Adult stem cells have the ability to self-renew and differentiate into the more specialized cell types of a given organ and are found in tissues throughout the body. These cells, whose progeny are termed progenitor cells in human liver and oval cells in rodents, have the potential to treat patients through the generation of hepatic parenchymal cells, even from the patient's own tissue. Little is known regarding the nature of the hepatic progenitor cells. Though they are suggested to reside in the most distal part of the biliary tree, the canal of Hering, the lack of unique surface markers for these cells has hindered their isolation and characterization. Upon activation, they proliferate and form ductular structures, termed "ductular reactions", which radiate into the hepatic parenchyma. The ductular reactions contain activated progenitor cells that not only acquire a phenotype resembling that observed in developing liver but also display markers of differentiation shared with the cholangiocytic or hepatocytic lineages, the two parenchymal hepatic cell types. Interactions between the putative progenitor cells, the surrounding support cells and the extracellular matrix scaffold, all constituting the progenitor cell niche, are likely to be important for regulating progenitor cell activity and differentiation. Therefore, identifying novel progenitor cell markers and deciphering their microenvironment could facilitate clinical use. The aims of the present PhD thesis were to expand knowledge of the hepatic progenitor cell niche and characterize it both during development and in disease. Several animal models of hepatic injury are known to induce activation of the progenitor cells. In order to identify possible progenitor cell markers and niche components, we examined several genes upregulated in a global gene expression array conducted on one of these models, in which progenitor cells are activated. The protein expression patterns were evaluated in our collections of human embryonic and fetal livers, human liver diseases, and rodent hepatic injury models. When analyzing standard histological liver sections underlying connections and tissue architecture are not immediately evident. We therefore developed models for digitally reconstructing not only protein expression in serially cut tissue sections, but also vessels of the portal area. Article I constituted our earliest attempts to create 3D reconstructions of biological material. Human embryonic stem cell cultures were previously thought to consist of homogenously undifferentiated cells. The protocols for 3D reconstructions developed in this study demonstrated micro heterogeneity in expression of differentiation markers and provided the basis for later reconstructions of hepatic tissues. In article II we examined the expression patterns of chosen proteins seen upregulated in the gene array as well as classical hepatocytic and cholangiocytic markers in human liver disease and during prenatal development. Previous studies had indicated direct connections between activated progenitor cells apparently isolated in the parenchyma and the intrahepatic biliary tree. Our developed protocols for 3D reconstructions visually demonstrated direct connections between these entities. Analysis of protein expression in prenatal liver revealed the formation of the intrahepatic tree to occur through a special form of asymmetric tubulogenesis, only recently described in mice. In order to describe the composition of the hepatic progenitor cell niche and the localization of cell surface proteins in article III, the expression patterns of certain genes upregulated in the gene array analysis were analyzed in different models of rodent liver regeneration. We observed that the extracellular matrix molecules collagen 1a1, laminin, nidogen-1 and agrin embraced the biliary cells and sharply defined the hepatic progenitor cell niche, which was encircled by desmin positive support cells. In all injury models biliary cells expressed the cell surface proteins matriptase and HAI-1. However, in the so-called 2-AAF/PHx model of progenitor cell activation, a subpopulation of hepatic progenitor cells was positive for Dlk1. 3D reconstructions clarified that the Dlk1-subpopulation was entirely located in the portal area periphery, and connected to the bile ducts via HAI-1 positive biliary cells. The heterogeneous expression patterns of matriptase, HAI-1 and Dlk1 in this particular injury model indicate the presence of a cellular hierarchy containing possibly less differentiated Dlk1-positive hepatic progenitor cells. In conclusion, our studies characterized the hepatic progenitor cell niche in humans and rodents. We successfully developed protocols for digitally visualizing, not only hepatic, but virtually any tissue through two fundamentally distinct approaches. The identification of an asymmetric form of tubulogenesis in humans added new knowledge to the development of the intrahepatic biliary tree, and thereby the formation of the progenitor cell niche. The identification of heterogeneously expressed cell surface proteins and extracellular matrix components provided knowledge of the constituents defining the niche. These pieces of information are important for future isolation and characterization studies of biliary subpopulations and their differentiation abilities in vitro.