In the present study, we established human HLCs derived from hADSCs according to a three-step protocol 15. Nrf2 was activated during the procedure, and knockdown of Nrf2 delayed the maturation and reduced specific functions of HLCs. Thus, Nrf2 might be a notable target for developing highly functional human HLCs.
Generation of hepatocytes derived from hADSCs could be a promising alternative source of human hepatocytes. Of the donor sources of hepatocytes proposed, including induced pluripotent stem cells, embryonic stem cells, and MSCs, MSCs may be best cell source because they can overcome problems associated with genetic damage, rejection, and ethics. In particular, ADSCs – a type of MSC derived from fat tissues – can be collected from patients with minimal invasiveness. We previously reported the effectiveness of a three-dimensional culture system for HLCs generated from hADSCs with an recombinant peptide µ-piece 26. However, mechanisms underlying the differentiation of hepatocytes from hADSCs are not well elucidated. Differentiation of definitive endoderm is an essential and critical early step to generate hepatocytes from hADSCs. Thus, efficient and reproducible production of definitive endoderm in differentiation cultures plays a key role in developing high quality and homogeneous hepatocytes.
We previously reported that Nrf2, a transcription factor that regulates cellular defenses against oxidative stress 28, protects against liver ischemia/reperfusion injury 29 and calcineurin inhibitor-induced nephrotoxicity 30 by inhibiting oxidative stress, inflammation, and apoptosis. In addition, Nrf2 maintains optimal levels of intracellular reactive oxygens species to regulate adipocyte differentiation 31. Moreover, Nrf2 overexpression in MSCs induce stem-cell marker expression and enhances osteoblastic differentiation by preventing apoptosis under oxidative stress 32. Considering this perspective, we anticipated a role for Nrf2 in regulating hADSC differentiation.
Wnt signaling is key for the differentiation of MSCs 33. Wnt/β-catenin signaling regulates foregut endoderm fate, proliferation, and morphogenesis 34. Previous studies revealed that Wnt signaling is required to specify definitive endoderm from hADSCs, and manipulations of Wnt signaling through use of GSK3 inhibitors have been applied to direct differentiation of definitive endoderm and hepatocytes 35. GSK3 inhibitors upregulate transcription factors involved in specification of definitive endoderm, namely GATA4, FOXA2, SOX17, and CXCR4 15. In this study, we used the GSK3 inhibitor CHIR99021 in Step 1 and successfully acquired homogenous hepatic progenitor cells. Notably, GSK3β is a negative regulator of Nrf2 27. Exposure of cells to inducers of antioxidant response elements (AREs), such as oxidative stress, leads to the dissociation of Nrf2 from its cytosolic inhibitor Keap1 in a process mediated by protein kinase C 27. After nuclear translocation, Nrf2 promotes the AREs of many cytoprotective genes in combination with other transcription factors, such as small musculoaponeurotic fibrosarcoma, activation transcription factor 4, and polyamine-modulated factor 1. Nuclear export of Nrf2 is regulated by the tyrosine kinase Fyn. Moreover, GSK3β acts upstream of Fyn kinase to control nuclear export and subsequent proteasomal degradation of Nrf2. Because a previous report showed an equal ability of CHIR99021 and activin A to induce hepatic progenitor cells 15, the role of Nrf2 seems to be not vital for endoderm differentiation. In line with this notion, application of the GSK3 inhibitor CHIR99021 did not significantly enhance Nrf2 nuclear translocation during Step 1 in the present study compared with activin A. However, knockdown of Nrf2 during Step 1 resulted in a longer time for HLCs to mature in the present study, implicating the importance of the Nrf2 pathway in promoting HLC differentiation. Moreover, rates of Nrf2 nuclear translocation were higher in the presence of the GSK3 inhibitor compared with activin A in Step 2. Thus, Nrf2 may play an important role in hepatoblast differentiation.
Another possible mechanism is FGFR4-GSK3β signaling, which promotes translocation of Nrf2 to the nucleus. FGFR4 receptors on the surface of hepatocellular carcinoma cells were activated by FGF19 through inactivation of GSK3β in the tumor microenvironment 36. Although it the role of the FGFR4-GSK3β-Nrf2 pathway under basal conditions is unknown, the homodimer of FGFR4 binds with FGF subfamily members including FGF4 or FGF19. We used FGF4 for hepatoblast differentiation in Step 2. FGF4 is the key cytokine required for differentiation of MSCs into a hepatic lineage 37. Nuclear translocation of Nrf2 mainly occurred during Step 2. Taken together, the FGFR4-GSK3β-Nrf2 pathway may play a key role in hepatoblast differentiation, although further investigation is needed.
As a limitation of the present study, we did not investigate other molecules in the Wnt signaling pathway or GFR4-GSK3β-Nrf2 pathway. Second, we did not investigate the effect of a specific Nrf2 inducer, such as dimethyl fumarate. Third, we did not perform transplantation of HLCs in this study to verify their actual function in vivo.
In conclusion, we demonstrated that Nrf2 was activated during differentiation of HLCs, especially during the hepatoblast differentiation step, whereas knockdown of Nrf2 delayed the maturation and impaired specific functions of HLCs. Our findings suggest that Nrf2 is a notable target for developing highly functional human HLCs.