Unit: Genetics of Differentiation - URA 2578 CNRS
Director: Mary C. Weiss
The unit studies the mechanisms implicated in the regulation of expression of hepatic genes and in the roles of the transcription factors that are necessary for the establishment and maintenance of hepatic differentiation. In addition, we have established new experimental models of liver cell differentiation, and are attempting to identify transcripts that are specifically induced when hepatocytes are infected by sporozoites of Plasmodium berghei.
Role and regulation of HNF4α (N. Briançon, A. Bailly, C. Rachez)
At present, our efforts are concentrated on the study of HNF4α, a transcription factor of the nuclear receptor family which has been identified as a key factor for execution of the liver differentiation program. The HNF4α gene has two alternative promoters whose use is regulated during the course of liver development. (Other isoforms, not discussed here, derive from internal splicing events.) The distal P2 promoter directs the expression of the HNF4α7 isoform, and it is used during embryogenesis in the liver, while the HNF4α1 transcript of the proximal P1 promoter is always the majority form in the liver, and the quasi-exclusive one after birth.
To understand the basis for the differences in transcriptional activity of the HNF4α1 and HNF4α7 isoforms, which differ only in their N-terminal amino acids, the two transactivation domains known for nuclear receptors were analysed using tests of in vitro interaction as well as co-transfection assays. It was found that the region corresponding to the AF-1 of HNF4α7 is devoid of activity while that of HNF4α1 showed the anticipated activity, that was increased by interaction with GRIP and CBP. For both isoforms the AF-2 domain was able to mediate interactions with GRIP-1, p300 and the corepressor SMRT. However, the repression imposed by SMRT was less robust for HNF4α7 than for HNF4α1. Finally, both isoforms, in association with SMRT, are able to recuite HDAC1 and 4, and when the three molecules are associated, a larger quantity of HNF4α1, but not of HNF4α7, is fixed. Moreover, in a test of transfection in the presence of an inhibitor of HDACs, a dramatic increase of reporter gene activity was observed for HNF4α 1 but not HNF4α7. These results imply that HNF4α1, the isoform found in adult liver, is more subject to regulation by interaction with coactivators and corepressors than the HNF4α7 isoform that is present mainly in embryonic and fetal liver.
To analyse the roles of the two isoforms of HNF4, α 1 and α7, we have used directed mutagenesis in the mouse to create animals that express only one of the two isoforms (as well as its internal splicing derived isoforms) , but under control of both of the respective promoters. In this context, the total amount of HNF4α protein should not vary in the knock-in geneotypes. Mice that produce α1-only and α7-only are viable and reproduce normally, indicating that the two isoforms show functional redundancy. Nevertheless, subtle differences are observed in the adults : in particular, lipid metabolism is affected in the α7-only animals. Furthermore, it has been possible to identify target genes of HNF1α which require a functional AFl for activity, including ApoAIV and Constitutive Androstane Receptor (CAR). These geneticaly modified animls permit a direct test of the role of the AF1 domain of this member of the nuclear receptor family, for only HNF4α1 possesses a functional AF1.
In order to analyze, on a molecular basis, the different transcriptional activities of HNF4α1 and HNF4α7, we have initiated a proteomic approach to isolate and characterize proteins that interact differentially with each isoform of HNF4α, and that may be linked to their different transcriptional activities. This approach reproduces a strategy previously employed in the identification and characterization of novel coregulators of the retinoic acid receptor RARα. The work performed on the coregulators of RARα, initiated in a different research unit, is now being finalized in the laboratory.
Analysis of the P2 promoter has revealed that it is regulated primarily by binding sites for HNF1 and HNF6/Oc2. Transgenic knock-in mice that express the P2 directed HNF4α7 isoform under the control of both promoters show strong up-regulation of the P2 promoter in the adult liver, indicating that the HNF4α1 isoform is a negative regulator of P2 through indirect binding to DNA. These results explain the surprising finding that the major hypersensitive site of the P2 promoter remains present in the adult liver, even though the promoter is hardly expressed at this stage.
Bipotential hepatic cell lines (H. Strick-Marchand, C. Deschatrette, T. Imaizumi Scherrer, D. Faust, G. Hayhurst, G. Yeoh)
We have continued to study bipotential cell lines obtained from embyronic liver of the mouse. While the first lines to be studied here were obtained from transgenic mice expressing an activiated form of human c-Met in the liver (MMH or Mouse Murine Hepatocyte lines), our recent attempts to obtain them from non-transgenic mice have been successful. Thus, we now have bipotential liver lines designated BMEL (Bipotential Mouse Embryonic Liver) from a large variety of mouse lines. In addition, we have improved the methods for obtaining their differentiation into cholangiocytes and hepatocytes. Hepatocyte functions such as apolipoproteins, albumin, AFP, alcohol dehydrogenase and aldolase B are induced upon culture for a few days in the form of aggregates. When cultured in Matrigel, not only are cholangiocyte/oval cell markers such as CD34, c-kit, integrin b4, connexins and GGTIV induced, but morphogenesis into bile duct units is observed. Thanks to an NIH Genome Anatomy Project grant, in collaboration with the laboratory of Gretchen Darlington (www.scgap.org), Affymetrix array analysis has been carried out using these cell lines under basal and differentiation conditions.
At present, we are analysing liver cell lines obtained from adult wild type mice or mouse embryos carrying targeted mutations in transcription factors important for liver dvelopment. These studies will permit us to clarify the roles of these factors in the maintenance of the bipotential phenotype and in the execution of the cholangiocyte and hepatoctyte programs.
Finally, the BMEL cell lines have been tested for their capacity to participate in liver regeneration, using the transgenic ALB-uPA mouse where excess quantities of the toxic uPA cause hepatocyte degeneration. In collaboration with Dina Kremsdorf and Serban Morosan (INSERM unit 370 at Hopital Necker) and Pierre Charneau (IP), BMEL-GFP cells have been injected into the spleen of young ALB-uPA/SCID mice where they form multiple islands of proliferating hepatocytes and bile ducts. Cells within the BMEL islands express the donor derived H2Kk while the H2 antigen of the host, H2Dd is absent, proving that the BMEL cells have not fused with host hepatocytes.Moreover, the BMEL-GFP hepatocyte islands and bile ducts show appropriate expression respectively of hepatocyte (HNF4a, HNFa glutamine synthetase, carbamoyl phosphate synthetase) and bile duct markers (cytokeratins 7 and 19, HNF1b). Furthermore, the hepatocyte enzymes show regulated zonal expression within the hepatic acini. These results demonstrate that the bipotentiality of BMEL cells extends to the liver in vivo, where the injected cells behave exactly as their normal adult tissue cell neighbors, responding to in vivo signals for growth and differentiation. In addition, similar cell lines have been isolated from healthy livers of C57BL6 adult mice: the clonal lines are also bipotential in vitro and in vivo.
Keywords: Hepatic differentiation, liver-enriched transcription factor, HNF4a isoforms, co-regulator, bipotential cells, liver repopulation