An original protocol for pancreatic cell differentiation generates liver-like cells from four human embryonic stem cell lines: a role for BMP and FGF pathways

Owing to islets donor shortage, only selected diabetes patients can benefit from transplantation therapy. Development of new beta cell sources is required, including differentiated human embryonic stem (hES) cells progenies. Recent data from Beatge's group showed efficient definitive endoderm induction and further differentiation toward pancreatic endocrine progenitors that yielded immature beta cells in vitro or functional ones in vivo. These findings open new perspectives for the setup of hES-derived beta cell grafts. However they were not reproduced by other investigators using different cell lines.
In this study, we applied the original protocol to four hES cell lines derived in our institute. Although they sequentially gave rise to definitive endoderm, primitive gut tube and posterior foregut equivalents as initially reported, none of them differentiated to Pdx1-positive pancreatic progenitors at proportions described for CyT203 line. Unexpectedly, they efficiently (40-60%) generated hepatoblast- (AFP+/Albumin+) then hepatocyte-like (AFP-/Albumin+) cells displaying urea production, albumin secretion and glycogen storage. This hepatic differentiation was negatively regulated by single or combined antagonism with BMP and FGF pathways using noggin and SU5402 respectively. But in these conditions, no significant increase in pancreas markers expression was noticed.
We conclude that certain features of the original protocol favor hepatic at the expense of pancreatic fate and that supplemented FGF10 contributes to these observations, recalling the versatile role of this pathway in early liver/pancreas development. Additional modifications are being tested in view of inhibiting hepatic and promoting pancreatic fate.