The large-scale production of differentiated cells suitable for clinical transplantation is a fundamental objective of regenerative medicine research. The replacement of damaged cells using in vitro differentiated functional cardiomyocytes has received more attention in the last decade with clinical trials on the horizon. Despite many developments in the stem cell field, we still lack the ability to produce adequate number of suitable differentiated cells that can meet clinical requirements. We conducted this study to evaluate the tumorgenicity of bioreactor derived cardiomyocytes, since we have observed that ESCs differentiated in stirred suspension culture maintain their pluripotency compare to those differentiated in static culture . In this study we observed that by day 12 following the removal of LIF in suspension, ESCs underwent efficient differentiation into cardiomyocytes in the presence of ascorbic acid and DMSO. Expression of the early mesodermal marker (PCAM) and cardiomyocyte specific markers (ALCAM (CD166) , ANF , α-MHC and β-MHC ) clearly confirmed the differentiation of ESCs into cardiomyocyte in stirred suspension culture. Similarly, in static culture, cells efficiently made EBs three days after the removal of LIF and the derived-cardiac bodies showed characteristics of cardiomyocytes including expression of the molecular marker α-MHC. It has previously been shown that both AA and DMSO enhance differentiation of mouse ESCs into cardiac myocytes. It has been suggested that AA induces permissive changes enabling cardiomyocyte differentiation, while DMSO has been shown to activate essential cardiogenic transcription factors, such as GATA-4 and Nkx-2.5. However, the exact related mechanism for triggering these genes is still not very well known [21, 22].
Ultra-structural studies by electron microscopy revealed the specific subcellular sarcomeric organization of cardiomyocytes, such as Z-banding. H- and A-bands, and intercalated discs including desmosome and gap junction were also obvious in derived cells. The chronotropical responses of cardiomyocytes confirmed the existence of specialized Ca+2 channels as well as α1 and β1-adrenergic receptors.
Importantly, no tumors were generated after cardiomyocyte lineage selection used to eliminate cells not expressing the α-MHC-neor gene. Based on these experiments, we assume that functional cardiomyocytes had been produced that didn't pose a risk of tumor formation in vivo. However, it should be emphasized that even in the presence of drug selection, ESCs differentiated in stirred suspension culture, still maintained the expression of pluripotency markers. This trend was even more apparent in cultures not undergoing drug selection pressure, where a sub-population expressed both Oct4 and α-MHC simultaneously in the same cell. It is important to remember, however, that drug selection did negate tumor formation signifying that the expression of pluripotent genes is either: (i) not maintained once the cells are taken out of expansion, or (ii) not sufficient to reprogram a cell that has undergone terminal differentiation.
Furthermore, the results presented here show that even after differentiation toward cardiomyocytes, a sub-population of cells (54%) still express Oct4 demonstrating some link between Oct4 gene expression and the environment. This observation is further confirmed by the difference observed in teratoma formation between static and suspension. The suspension bioreactor differentiations retained cell pluripotency to a degree that ESCs formed teratomas representative of all three germ layers. In contrast, cells from static cultures only formed an unstructured cell population: although not teratomas, tumor masses were still generated. While lineage selection using transgenic constructs is a useful research tool, this approach cannot be applied clinically. Hence, our observation that bioreactor-derived cardiomyocytes maintain hallmarks of pluripotency has significant implications. Before the large scale production of cardiomyocytes in suspension bioreactors will be possible, it will be necessary to eliminate bioreactor induced pluripotency.
We have recently evaluated the application of suspension bioreactor culture for the generation of cartilage and bone tissue . Unlike the static culture environment, bioreactor-differentiated aggregates caused teratoma formation when implanted subcutaneously into SCID mice, which implicated the existence of pluripotent cells in the bioreactor even after 30 days of suspension culture. Upon closer analysis of cells within aggregates, we discovered that cells on the ridges of the aggregate expressed the greatest amount of Oct4 in locations, which would be exposed to the greatest amount of laminar fluid flow. This result suggested that the maintenance of pluripotency within bioreactor differentiation cultures was the result of fluid shear stress. In the current study, much like all our other bioreactor studies, we apply an agitation rate of 100 rpm (shear stress of 6.1 dyne/cm2) in order to optimize mass transfer and avoid necrosis, which otherwise occurs at lower velocities .
We believe that liquid shear stress in the stirred suspension bioreactor plays an important mechanistic role in bioreactor induced pluripotency. Shear stress can modulate gene expression through mechanotransduction, where physical signals are sensed at the cell periphery, transduced into biochemical signals within the cell, ultimately resulting in cell responses, including changes in gene expression [23, 24]. Previous studies in other cells have demonstrated that shear stress can induce the nuclear translocation of ß-catenin into the nucleus [25, 26]. As ß-catenin is an important regulator of pluripotency , we are interested in the role of the non-canonical Wnt signaling pathway in this induced pluripotency process. Recently, using a LEF/TCF-GFP (Lymphoid enhancer factor/T cell factor) reporter system, we have confirmed that ß-catenin nuclear occupancy is considerably increased over controls when cells are exposed to 6.1 dynes/cm2 shear stress compared to lower levels of shear generated by lower velocities of stirring (unpublished data). Following nuclear translocation, β-catenin forms a complex with LEF/TCF transcription factors. This complex interacts with the specific sequence in the promoter region to activate transcription of certain genes.
Recently, Saha et al. has shown that biaxial cyclic strain above a certain threshold inhibits human ESC differentiation and enhances their self-renewal without selecting against growth or survival of differentiated or undifferentiated cells . This group later suggested that strain may induce autocrine or paracrine signaling through TGF-β(Transforming growth factor-beta) superfamily ligands in human ESCs since the TGF-β superfamily activation of Smad2/3 was necessary for suppression of spontaneous differentiation under strain .
Despite the fact that stirred suspension cultures are very useful for the generation of a large number of undifferentiated cells, we have found that the addition of medium enhancers is not adequate to force complete differentiation of the population in suspension bioreactors. By elucidating the exact mechanism(s) by which liquid shear stress may contribute to promoting pluripotency and preventing differentiation, we will be able to create an efficient environment for both the production of large quantities of pluripotent stem cells, and their differentiated progeny. This is an important objective for human regenerative medicine, as lineage selection using transgenes will not be possible.