Many hepatic differentiation protocols involve EB formation from ES cells grown in static culture [25, 27]. Similar to embryos, EBs contain an endoderm, ectoderm, and mesoderm, which are important for the induction of complex differentiation . Several techniques, such as the hanging-drop and conical-tube methods , have been developed to facilitate EB generation. The stirred-tank bioreactor and perfused 3D bioreactor have been applied for the hepatic differentiation of ES cells [30, 31]. However, none of the current methods is sufficiently stable and controllable for the large-scale differentiation of ES cells for practical applications. Here, we used a rotating bioreactor to induce EB formation and hepatic differentiation. Cell culture in a rotating bioreactor has been shown to support the efficient, large-scale production of transplantable cells for bone-tissue engineering and of hematopoietic cells [32, 33]. To the best of our knowledge, although a few studies have examined the formation and multiple-tissue or tissue-specific differentiation of EBs in a RCCS [15, 34, 35], no one has reported the hepatic differentiation of mES cells via EB formation in a rotating bioreactor. In our previous study , the hepatic differentiation of mES cells within the 3D condition of a rotating bioreactor has been described, which is similar to the present study. A 3D-culture system consisted of biodegradable scaffolds, growth-factor-reduced Matrigel, and a rotating bioreactor in our previous study. Scaffolds and Matrigel provided a 3D environment for the cell growth, proliferation, and differentiation of mES cells. Tissue mass with functional hepatocytes was obtained from the rotating bioreactor for bioartificial livers and engineered liver tissue. In this study, however, the formation of EBs was dependent on mES cell self-assembly. A native 3D microenvironment with cell-cell and cell-matrix interactions was recapitulated in EBs. A large number of EBs with functional hepatocytes were prepared in the rotating bioreactor. This method is suitable to large-scale generation of hepatocytes derived from ES cells for cell transplantation and the development of bioartificial livers. Therefore, our data indicate that the rotating bioreactor is an effective tool to provide a native 3D microenvironment for the large-scale production of hepatocytes derived from ES cells.
Several studies demonstrated that the optimal speed for EB formation is 15 to 20 rpm in the rotating bioreactor [33, 36]. To enhance the efficiency of EB formation, we chose a slow rotational speed and a high cell concentration, resulting in large numbers of homogeneous EBs within 1 week. By culturing ES cells in suspension in this dynamic and mild environment, which is characterized by extremely low fluid shear stress and optimal nutrient and gas exchange, our protocol for EB formation not only avoids massive cellular damage but also controls the extent of EB agglomeration, thus reducing large central necrosis in EBs. More important, both histologic and morphologic observations revealed that this setup is more beneficial for the differentiation of ES cells into hepatocytes than standard conditions.
It has been previously shown that exogenous growth factors and hormones play an important role in the differentiation of ES cells into hepatocytes and in the expression of the hepatocyte phenotype . However, it is still unclear which growth factor, hormone, or combination of the two is critical for the hepatic differentiation of ES cells. Additionally, the optimal timing for the use of these factors and hormones to differentiate ES cells into hepatocytes is not known. In the current study, we found that the hepatic differentiation of ES cells in a rotating bioreactor was more efficient with than without the presence of HGF, FGF4, dexamethasone, DMSO, and ITS. More specifically, the early use of medium containing these factors resulted in the efficient differentiation of ES cells into hepatocytes, as determined by morphologic analysis (data not shown).
In this study, we investigated the expression of TTR, AFP, ALB, G6P, and TAT at the mRNA level. Among these genes, TTR is a marker of endodermal differentiation, AFP represents endodermal differentiation as well as the early differentiation of fetal hepatocytes , ALB appears in early fetal hepatocytes and reaches a maximal level in adult hepatocytes , G6P is predominantly expressed in hepatocytes in the late gestational and perinatal stages , and TAT is an excellent enzymatic marker for perinatal or postnatal hepatocyte-specific differentiation and represents an important function of the liver . In general, the expression of these specific genes in cells derived from EBs is characteristic of differentiated hepatocyte-like cells, although a proportion of these genes is also expressed in the yolk sac and other cells. The expression of such genes in EBs differentiating in the presence of growth factors parallels the expression observed in the embryonic liver in vivo. However, in our tests, the mRNAs of TTR, AFP, ALB, G6P, and TAT were expressed by day 7, and the expression of most of the genes, such as ALB, G6P, and TAT, was noticeable earlier than in previous reports [27, 41]. The level expression of these genes was higher in the EB-derived cells cultured in the bioreactor than the mES cells differentiated in the 2D culture. Because the expression of the ALB, AFP, and CK18 proteins was also detected in the day-7 EBs by WB analysis, and based on our immunostaining results, we believe that growth of ES cells in a rotating bioreactor supplemented with exogenous growth factors not only strongly promotes differentiation into hepatocyte-like cells but also enhances cell maturation.
The assay for ALB synthesis is well known as a relatively specific test for the presence and activity of hepatocytes during the hepatic differentiation of stem cells [23, 25]. We examined the ALB synthesis of mES cell-derived EBs to confirm the differentiation of the ES cells into functional hepatocytes. However, we failed to detect ALB directly at measurable levels during the rotating bioreactor culture period. As the probable reason for this issue was the relatively low cell concentration in the bioreactor, we removed the EBs from the bioreactor and seeded them in 12-well plates at a concentration of approximately 50 EBs/well. After incubation with fresh medium for 72 hours, these EBs secreted ALB into the culture medium. The ability of the EBs to synthesize ALB at days 14 and 21 was approximately sixfold to eightfold higher than at day 7, confirming the hepatic differentiation of the EBs.
Because normal hepatocytes possess many complex functions, in addition to our ALB analysis, we used other independent functional assays to verify further the hepatocyte-like attributes of the differentiated ES cells. Four functional tests were performed on both the EBs from the rotating culture and the EB-derived cells from the static culture more precisely to characterize the hepatic functions of the differentiated cells. Our data demonstrate that the EB-derived cells possessed cytochrome P450 activity, stored glycogen, and took up ICG and LDL, which were higher than that of cells derived from 2D culture. Together, all of these results strongly suggest that the cells differentiated in the rotating bioreactor, along with growth factors and hormones, were functional hepatocytes.