When human blastocysts are cultured in vitro to derive embryonic stem cells, they undergo profound molecular changes that alter their identity.
Human embryonic stem cells (hESCs) offer exciting possibilities in the fields of regenerative medicine and disease modeling as they possess the potential to differentiate into all somatic lineages. Much progress has been made in understanding the molecular regulation of hESCs during in vitro propagation, but little is known about the derivation process itself. A study in this issue by O'Leary and colleagues1 tracks a handful of key molecular markers during the culture of >200 human blastocysts and provides evidence that pluripotent cells of the blastocyst inner cell mass (ICM) undergo substantial molecular and epigenetic changes during the transition to hESCs (Fig. 1).
In the pre-implantation epiblast of the mouse, Oct4+ ICM cells initiate expression of the pluripotency factor Nanog, a process that in female embryos coincides with reactivation of the inactive X chromosome, producing cells with two active X chromosomes (XaXa)2. This pluripotent state is transient: in the post-implantation epiblast, the Oct4+ cells reduce expression of Nanog and of other pluripotency regulatory transcription factors such as KLF2, 4, 5, and female cells inactivate one of the X chromosomes. By adding factors that block lineage commitment and maintain pluripotent cell growth, researchers have been able to artificially preserve two mouse pluripotent cell states: (i) a 'naive' state, similar to that of the ICM or ESCs, and (ii) a 'primed' state, similar to that of post-implantation epiblast stem cells (EpiSCs)2,3, where 'primed' indicates an increased propensity for early lineage commitment (based on features such as X-chromosome inactivation and expression of major histocompatibility class I molecules). Importantly, mouse ICM cells or ESCs can be coaxed to differentiate into EpiSC-like cells in vitro after expansion in conditions containing bFGF and Activin3,4.
Cultured naive ESCs and primed EpiSCs have been shown by various methods to closely resemble at the molecular level their in vivo counterparts, the pre- and post-implantation epiblast, respectively2. This claim has been reinforced by a recent study that used lineage-tracing experiments to show that murine ESC derivation under defined naive stem-cell growth conditions does not transiently activate the early germ cell differentiation marker Blimp1 during the derivation process5, thus refuting theories claiming that murine ESCs must originate from cells committed to the germ-cell lineage.
Comparing embryonic development in mice to that of humans is difficult given ethical restrictions on the culture of human embryos. As an alternative, scientists can study certain questions of human development using hESCs. A major puzzle in the field is that although hESCs are derived from blastocyst-stage embryos6, they are remarkably different from mouse naive ESCs and share many defining features with mouse primed EpiSCs2. These shared features include pluripotency maintenance through bFGF and ACTIVIN/NODAL signaling, differentiation upon ERK signaling inhibition, cell morphology, and X-chromosome inactivation in most female hESCs7. Considering that mouse pre-implantation ICM cells can adopt a primed state in vitro3 and that hESCs can be transiently and artificially rewired to adopt a naive state similar to that of mouse ESCs7, it is possible that the process of deriving hESCs proceeds from a naive state to a primed state.
To investigate this question, O'Leary and colleagues1 closely followed the molecular changes associated with the conversion of human blastocyst-derived ICM cells into hESCs. These experiments revealed a new, defined, transitional cellular state called the post-ICM intermediate (PICMI). Cells within PICMI structures are characterized by a transient molecular and morphological configuration that emerges from ICM cells 3–5 days after blastocyst plating. PICMI formation in vitro immediately precedes the formation of hESCs and predicts with an 80% success rate the subsequent derivation of stable hESC lines, regardless of the ICM isolation technique or embryo source used (Fig. 1).
A comparison of the molecular and epigenetic status of PICMI cells, ICM cells and hESCs showed that PICMI cells maintain the expression levels of core pluripotency factors (OCT4 and NANOG) seen in ICM cells, but also express bFGF and NODAL/TGFb signaling regulators (NODAL, LEFTY1 and LEFTY2, CRIPTO, for example) at high levels not detected in ICM cells. In addition, PICMI cells upregulate various early and late mouse epiblast markers, including FOXD3 and FOXA2. Notably, however, PICMI cells and hESCs retain expression of ICM markers, including REX1, NANOG and KLF4 (ref. 2), indicating that PICMI cells and hESCs differ from murine EpiSCs. Finally, when O'Leary and colleagues1 tested signaling pathway activity in ICM and PICMI cells, they found that only PICMI cells rely on epiblast NODAL signaling, with no indication of canonical Wnt signaling activation, which is known to promote the ground state of pluripotency. These findings may indicate an adaptive reconfiguration of signaling pathway activity during the hESC derivation process.
O'Leary and colleagues1 also examined X-chromosome status during hESC derivation, as in mouse development the pluripotent state is XaXa whereas the post-implantation epiblast has inactivated an X chromosome (XaXi). A previous study had concluded that X-chromosome dynamics in human blastocysts are similar to those of the mouse based on the finding that hESC derivation in 5% oxygen conditions (which is close to physiological oxygen levels during development) preserves the preexisting XaXa state indefinitely and prevents oxidative stress–induced precocious inactivation of the X chromosome8. O'Leary and colleagues1 stained Oct4+ cells in female PICMIs and hESCs for the repressive H3K27me3 mark, which indicates a silenced X chromosome, and reached a different conclusion. They found that accumulation of the H3K27me3 mark3 can be readily detected in human female ICM cells and PICMI structures. Neither ICM nor PICMI cells grown under 5% oxygen conditions showed indications of two active X chromosomes (Fig. 1). Rather, the authors noticed that some of the lines stochastically became XaXa only after extended expansion in vitro. After 7 days of differentiation, such hESC lines had inactivated an X chromosome, indicating that they retained the potential for X-chromosome inactivation.
The conflicting data among independent studies1,8,9 analyzing X-chromosome inactivation dynamics underscore the complexity of potential cross-species differences in X-chromosome regulation in ESCs and in early development. They further highlight the limits of our knowledge of early developmental epigenetic dynamics in humans and possibly other primates. In another recent example of developmental species differences, monkey blastomeres, but not ICM cells, were shown to generate chimeric monkeys through embryo aggregation10, whereas in rodents, both blastomeres and ICM cells have the unrestricted developmental potency to contribute to chimeric animals. Future studies focusing on in vivo early developmental dynamics in different organisms are likely to help resolve some of these important issues.
The insights of O'Leary and colleagues1 into the origin and identity of human pluripotent cells show clearly that, when it comes to closely observing explanted peri-implantation human embryos in bFGF/TGFb-stimulating conditions and awaiting the formation of ESCs after extended culture, what you initially see is not what you eventually get (Fig. 1). It is possible that PICMI formation corresponds to the activation of a molecular switch that masks the as-yet-undefined original identity of human ICM-derived cells and that artificially induces these cells to convert into what researchers currently and conventionally designate as hESCs. The findings of O'Leary and colleagues1 also suggest that human ESCs represent either a later and more primed potency stage in development compared with ICM-derived pluripotent stem cells or, alternatively, an artificial in vitro state that has no authentic in vivo counterpart.
O'Leary, B. Nat. Biotechnol. 30, 278–282 (2012).
Nichols, J. & Smith, A. Cell Stem Cell 4, 487–492 (2009).
Hanna, J. et al. Cell Stem Cell 4, 513–524 (2009).
Guo, G. et al. Development 136, 1063–1069 (2009).
Chu, L.F., Surani, M.A., Jaenisch, R. & Zwaka, T.P. Curr. Biol. 21, 1759–1765 (2011).
Thomson, J.A. et al. Science 282, 1145–1147 (1998).
Hanna, J. et al. Proc. Natl. Acad. Sci. USA 107, 9222–9227 (2010).
Lengner, C.J. et al. Cell 141, 872–883 (2010).
Okamoto, I. et al. Nature 472, 370–374 (2011).
Tachibana, M. et al. Cell 148, 285–295 (2012).
The authors declare no competing financial interests.
About this article
Cite this article
Pribluda, A., Hanna, J. Tracing the genesis of human embryonic stem cells. Nat Biotechnol 30, 247–249 (2012). https://doi.org/10.1038/nbt.2139