human pluripotent stem cells (PSCs), either directly isolated from embryos or induced by exogenous factors, have not only opened up the possibility of deriving genetically normal cell types in large numbers, but they also offer a window into normally inaccessible human embryogenesis. The understanding and control of human PSC self-renewal and differentiation rely on a thorough molecular understanding of control of the stem cell state: how cells exit the pluripotent state, and how cells are specified to particular fates. Much of the understanding of PSC biology has come from studies in the mouse. However, it has recently become apparent that PSCs from mouse and human, previously thought of as equivalent types, are, in fact, quite different entities. Mouse embryonic stem cells (ESCs) can be grown on feeders with leukemia inhibitory factor or by restriction of differentiation and, thus, heterogeneity using chemical inhibitors (2i dual-inhibition conditions) (4). PSCs have also been derived from postimplantation mouse embryos; such cells are referred to as epiblast stem cells (EpiSCs). These cells share many of the characteristics of human ESCs (hESCs), including basic gene expression and growth factor requirements. However, there are some differences between human and mouse EpiSCs, such as the tendency to express markers of the primitive streak stage of development (6). To complicate the picture further, hESCs appear to be complex mixtures of cell states and seem to be more heterogeneous than mouse primed or naive ESCs (2a).
Much research has been conducted over the past decade into the mechanisms by which pluripotency is controlled. Significant insights into the regulation of PSCs have been obtained from the investigation of transcription factors and their associated networks. Most of these studies have treated the mouse as a paradigm for the human system, which is likely to be a misleading assumption. Three transcription factors, octamer-binding transcription factor 4 [OCT4, POU domain class 5 transcription factor 1 (POU5F1)], sex-determining region Y-box 2 (SOX2), and NANOG, have emerged as central players in the control of pluripotency (3). These studies, however, have tended to use mixtures of cells, and, given the apparent heterogeneity in human PSC (hPSC) cultures, further refinement of the gene regulatory networks is required.
The realization that splice variation may play a role in the control of the pluripotent phenotype in humans adds a further layer of complexity (1, 11). In the human system, OCT4 (POU5F1) can be expressed as different splice variants: OCT4A, OCT4B, and OCT4B1 (9). OCT4A is the direct ortholog of the mouse Oct4 transcript and is critical for self-renewal of hPSCs (9). The roles of OCT4B and OCT4B1 are much less well understood, despite their expression at the transcript level in hPSCs. Previous observations of subcellular location are suggestive of functional differences between OCT4A and OCT4B (2), but proper functional analysis is lacking. The presence of six transcribed pseudogenes makes primer design and interpretation of PCR results for OCT4 expression tricky. Thus previous studies of the function of OCT4 in hPSCs have not distinguished between the similar OCT4A and OCT4B and, thus, could not assess the role of the alternate splicing in hPSC self-renewal and differentiation. In light of the apparent complexity of the human system, researchers must begin the task of disentangling heterogeneity and splice variation to fully understand the control of the pluripotent state in humans.
The study by Tsai et al. (8) in this issue of American Journal of Physiology-Cell Physiology is a first step to come to grips with the role that alternate splicing of OCT4 may play in the maintenance of pluripotency. Tsai et al. focused on the role of OCT4A, as overexpression of OCT4B does not result in increased protein production, suggesting posttranslational reduction of OCT4B levels in hESCs. The ability to specifically overexpress or knock down OCT4A yielded subtle differences between this study and previous efforts that did not discriminate between OCT4 isoforms. Specifically, targeting OCT4A for knockdown resulted in reduction of the transcription factors SOX2 and NANOG and increases in the transcription factors paired box 6 (PAX6), neural cell adhesion molecule (NCAM), and fibroblast growth factor 5 (FGF5). Most previous studies have tended to report that the effect of OCT4 knockdown in humans is similar to the mouse paradigm, namely, trophectoderm and endoderm differentiation. The specific upregulation of expression of PAX6 and NCAM, markers of neuroectoderm lineages, in the study of Tsai et al. is novel. The induction of PAX6 and NCAM upon OCT4A knockdown suggests that the relative levels of OCT4 splice variants may play a role in lineage specification in the human system. However, differences in phenotypes observed by Tsai et al. and in studies using a nonspecific OCT4 knockdown may be influenced by the signaling environment, not just which splice variant is knocked down; for example, bone morphogenetic protein 4 concentration specifies ectodermal vs. extraembryonic lineages in cells with low OCT4 levels (10).
The increase in expression of FGF5 following OCT4A knockdown is interesting, as this classical marker of the mouse epiblast has proved to be very difficult to detect in differentiating hESC cultures. There may be many reasons for the lack of FGF5 detection in previous studies, but, at the very least, the results presented by Tsai et al. (8) suggest that the OCT4A knockdown cells may pass through an FGF5-positive “epiblast-like” stage during differentiation. In contrast, overexpression of OCT4A, which initially seemed to induce lineage genes, appeared to stabilize the self-renewing state, resulting in cells with higher single-cell replating ability (Fig. 1). This effect may be a function of the fact that the protein levels of OCT4A, SOX2, and NANOG did not appear to increase, even though a proportion of the OCT4A expression was being driven from a transgene. This stabilization of the self-renewing state, despite OCT4A overexpression, again suggests the presence of subtle posttranslational control. The ability of OCT4A-overexpressing cells to exist as a stable entity is interesting, as data from the mouse suggest that overexpression of Oct4 drives differentiation (5), and there is a narrow range of OCT4 expression that permits pluripotent differentiation (7). Whether the OCT4A-overexpressing cells are compromised in their ability to differentiate or whether in humans the system is able to compensate for increased OCT4A expression remains to be tested.
Overall, the results presented by Tsai et al. (8) and other studies (1, 11) point to an emerging role for differential splicing in the control of the decisions that PSCs make between self-renewal and differentiation. It seems likely that much more attention to exactly which splice variants are expressed and which variants are being manipulated in experimental settings is needed, if we are to fully understand the control mechanisms that operate in PSCs in humans. The observation that increased OCT4A can stabilize a more clonogenic state in humans (the poor single-cell plating efficiency being a major drawback to handling human compared with mouse PSCs) may enable researchers in the future to establish what extrinsic controls, such as growth factors or inhibitors, could be applied to cultures to keep them in a clonogenic state and, thus, minimize stresses during expansion.
No conflicts of interest, financial or otherwise, are declared by the author.
P.J.G. prepared the figure; P.J.G. drafted the manuscript; P.J.G. edited and revised the manuscript; P.J.G. approved the final version of the manuscript.
- Copyright © 2014 the American Physiological Society