Z Gastroenterol 2015; 53 - A3_35
DOI: 10.1055/s-0035-1568055

Transcriptional regulatory networks governing stem cell differentiation into hepatocytes in vitro

P Godoy 1, W Schmidt-Heck 2, G Campos 1, A Widera 1, R Stoeber 1, T Weiss 3, A Nussler 4, G Damm 5, B Küppers-Munther 6, DC Hay 7, JG Hengstler 1
  • 1IfADo-Leibniz Research Centre for Working Environment and Human Factors at the Technical University Dortmund, Systems toxicology, Dortmund, Germany
  • 2Leibniz Institute for Natural Product Research and Infection Biology eV-Hans-Knöll Institute, Jena, Germany
  • 3University of Regensburg Hospital, Center for Liver Cell Research, Department of Pediatrics and Juvenile Medicine, Regensburg, Germany
  • 4Eberhard Karls University Tübingen, BG Trauma Center, Siegfried Weller Institut, Tübingen, Germany
  • 5Charité University Medicine Berlin, Department of General-, Visceral- and Transplantation Surgery, Berlin, Germany
  • 6Takara Bio Europe AB (former Cellartis AB), Gothenburg, Sweden
  • 7University of Edinburgh, MRC Centre for Regenerative Medicine, Edinburgh, United Kingdom

The differentiation of embryonic stem cells (ESC) to hepatocytes offers the possibility of unlimited supply of human hepatocytes for cell therapy and biomedical research. However, current protocols achieve only a partial hepatocyte differentiation. This is largely due to our poor understanding of the transcriptional regulatory networks (TRN) governing stem cell and hepatocyte plasticity.

Therefore, we performed a bioinformatics approach based on whole-genome expression analysis of human stem cells, hepatocyte-like cells (HLC) and primary hepatocytes (Hep) freshly after isolation (FH) and in different culture conditions. This allowed us to identify the TRN and biological motifs associated with gain (stem cell vs. HLC) and loss of hepatocyte phenotype (FH vs. cultivated Hep). Primary human (and mouse) hepatocytes from three independent donors were cultured in collagen monolayers or sandwich systems for up to 14 days. The gene networks were identified by applying the novel CellNet algorithm and by fuzzy-c means gene clustering and gene set enrichment analysis (GSEA). The most important genes identified by these approaches were validated by real time PCR.

Three different HLC models showed highly comparable phenotypes, with a partial gain of hepatocyte and concomitant fibroblast and colon gene networks. Interestingly, genes associated with failed acquisition of Hep features were largely correlated with genes repressed during primary hepatocyte cultivation. CellNet and GSEA identified novel transcription factors associated with the undesired phenotypes, including KLF5 and CDX2 for the “colon”, FOXQ11 and SOX11 for the “fibroblast” and HNF1, HNF4, CAR, FXR and PXR for the Hep failed gene clusters. In both human and mouse hepatocytes, we observed thousands of upregulated genes, albeit their fold change was larger in mouse hepatocytes. These included genes involved in mRNA processing, inflammation and cell migration. The loss of mature liver phenotype and inflammation motifs were strongest in mouse hepatocytes, particularly in monolayer culture, supporting the concept that 3D matrix sustains a differentiated phenotype more effectively than 2D systems. In spite of this difference, we could detect a significant set of genes that represents an interspecies motif of alterations in primary hepatocyte culture.

In conclusion, we present a blueprint for TRN involved in stem cell and hepatocyte plasticity, which will serve as basis for future interventions to improve the protocols for stem cell differentiation into hepatocytes.

Corresponding author: Godoy, P

E-Mail: godoy@ifado.de