Human pluripotent stem cells (hPSC), including human embryonic stem cells (hESC) and human induced plutipotent stem cells (hiPSC), have great potential for therapy, disease modeling, drug development, and for recapitulating human liver development. Liver regenerative medicine is a rich field with many approaches towards organ replacement (liver transplantation) and functional support (bioartificial reactor) approaches. Our approach is to focus on liver development and organogenesis in order to solve problems in liver regenerative medicine. The liver arises from the liver diverticulum (LD) which arises in mouse development (E9.5) and human development (d26). The LD is formed under low oxygen conditions and containins primitive liver cells (definitive endoderm (DE) and hepatic endoderm (HE)) within an epithelial sheet/tube encased by endothelial cells. The LD is surrounded by the septum transversum mesenchyme (STM) containing collagen-rich mesenchyme and a high density of mesenchymal cells, and receives cues from cardiac mesoderm Fibroblast growth factor 2 (FGF2), and STM (Bone morphogenetic protein 4 (BMP4)). We first determined the appropriate LD dimensions with an online mouse database. The LD is an elliptical structure, and we measured a long diameter of 180 µm, a short diameter of 90 µm, and a circumference of 810 µm. The columnar pseudostratified epithelium has an appreciable thickness of 70 µm and is encased in a single layer of endothelial cells. We hypothesized that hPSC-derived DE and HE retain the potential for three-dimensional (3D) tissue formation. Thus they can be used in an organoid format to model the LD. hPSC were differentiated into DE and HE respectively using the commercially available STEM Diff Kit media (Stem Cell Technologies) under 5% O2
on matrigel (MG)-coated tissue culture plastic for 4 days. Currently, this kit has not been used under low oxygen conditions that prevail during LD formation. Gene expression analysis demonstrated high expression of Forkhead box A2 (Foxa2) and (Sex determining region Y)- box 17 (Sox 17) on day 4 in the DE-like population. To model HE formation, day 4 DE was cultivated for an additional 5 days, in a serum free media (SFM)-containing bone morphogenetic protein 4 (BMP4) and fibroblast growth factor 2 (FGF2), based on stage-specific, in vivo liver development. Phase contrast microscopy of day 7 HE
demonstrated cuboidal morphology.
To engineer organoids that eventually resemble the LD, we first generated spheroids containing either
DE (day 4) or HE (day 7). Day 4 DE cells were harvested using accutase (enzymatic dissociation) and were seeded (20,000 cells) into one well of a 384-well round bottom, low attachment plate containing SFD ( IMDM, 20% knockout serum replacement (KSR), N2, B27, activin (50 ng/mL), FGF2, Glutamine, Ascorbic acid, Thioglycerol, BSA). After 24 hours, cells compacted to form distinct spheroids 500 µm in diameter, and maintained Foxa2 and Sox 17 compared to day 4 DE. In addition, day 7 HE (cultivated for 3 days in SFM withour KO serum, with FGF and BMP4) were harvested and seeded into spheroids as above for 5 days with SFD, as above, BMP4. The spheroid from both DE and HE remained stable with minor proliferation over 5 days. Gene expression analysis of DE and HE organoids demonstrated maintained expression of Foxa2 and near complete downregulation of Sox 17 as expected in DE to HE forward cell differentiation. Overall, we describe a novel approach and two new systems containing hPSC-derived spheroids for generating hepatic organoids. Future work will be to further model the LD by the addition of endothelial cells and mesenchyme with assessment of markers and phenotypes associated with liver development.