Hepatocyte differentiation potential of mesenchymal cell lineages for liver regenerative medicine

Human mesenchymal stem cells (MSCs) are being largely studied for their differentiation potential and immunological properties. In the present study, we evaluated the ability to reliably differentiate mesenchymal lineages into hepatocyte-like cells both in vitro and in vivo. For this purpose, we handled several tissue sources and compared typical MSCs from bone marrow (BM) or umbilical cord, to liver-derived mesenchymal-like cells and to fibroblasts. We observed that hepatocyte differentiation of BM-MSCs was incomplete and variable with elective expression of some specific markers. These mesenchymal-derived hepatocyte-like cells (MDHLCs) were also chimerical in their phenotype as they expressed mesenchymal markers while these were down-regulated. We therefore designed differentiation cocktails with an aim to improve MDHLC phenotype and some unexpected results were obtained with LIF cytokine whose action on stem cells for hepatocyte differentiation was not documented. Nevertheless, we observed a limitation in the acquisition yield of hepatic features. Furthermore, the hepatocytelike phenotype of MDHLCs completely disappeared when the cells were incubated into growth medium. However, we showed that hepatic functionality of these cells, as urea secretion and gluconeogenesis, could be increased under specific conditions, suggesting the potential to improve MDHLC phenotype. In vivo, MSCs were able to express hepatic markers into SCID-mice livers while their chimerical phenotype remained. In contrast, MDHLCs down-regulated their hybrid phenotype after transplantation suggesting a beneficial influence of in vitro differentiation step. MSCs were also able to engraft and even partially differentiate into wild-type mice which was a strong argument for their low immunogenicity. Surprisingly, fibroblasts showed highly similar potential than MSCs to differentiate into hepatocyte-like cells both in vitro an in vivo and these results underlined the difficulty to accurately distinguish between both cell types using current techniques. Umbilical cord-derived stem cells (UCMSCs) and adult-derived human liver stem cells (ADHLSCs) were different in nature and displayed a native hybrid phenotype while their differentiation allowed high levels of hepatocyte-like feature acquisition. Together all these data suggest the current possibility to engineer mesenchymal-derived hepatocyte-like cells owning specific features acquisition while remaining limited in their commitment. This highlights the need for further investigations to evidence the usefulness of these mesenchymal lineages for liver cell therapy.