X chromosome inactivation in human pluripotent stem cells as a model for human development: back to the drawing board?

BACKGROUND: Human pluripotent stem cells (hPSC), both embryonic and induced (hESC and hiPSC), are regarded as a valuable in vitro model for early human development. In order to fulfil this promise, it is important that these cells mimic as closely as possible the in vivo molecular events, both at the genetic and epigenetic level. One of the most important epigenetic events during early human development is X chromosome inactivation (XCI), the transcriptional silencing of one of the two X chromosomes in female cells. XCI is important for proper development and aberrant XCI has been linked to several pathologies. Recently, novel data obtained using high throughput single-cell technology during human preimplantation development have suggested that the XCI mechanism is substantially different from XCI in mouse. It has also been suggested that hPSC show higher complexity in XCI than the mouse. Here we compare the available recent data to understand whether XCI during human preimplantation can be properly recapitulated using hPSC.

OBJECTIVE AND RATIONALE: We will summarize what is known on the timing and mechanisms of XCI during human preimplantation development. We will compare this to the XCI patterns that are observed during hPSC derivation, culture and differentiation, and comment on the cause of the aberrant XCI patterns observed in hPSC. Finally, we will discuss the implications of the aberrant XCI patterns on the applicability of hPSC as an in vitro model for human development and as cell source for regenerative medicine.

SEARCH METHODS: Combinations of the following keywords were applied as search criteria in the PubMed database: X chromosome inactivation, preimplantation development, embryonic stem cells, induced pluripotent stem cells, primordial germ cells, differentiation.

OUTCOMES: Recent single-cell RNASeq data have shed new light on the XCI process during human preimplantation development. These indicate a gradual inactivation on both XX chromosomes, starting from Day 4 of development and followed by a random choice to inactivate one of them, instead of the mechanism in mice where imprinted XCI is followed by random XCI. We have put these new findings in perspective using previous data obtained in human (and mouse) embryos. In addition, there is an ongoing discussion whether or not hPSC lines show X chromosome reactivation upon derivation, mimicking the earliest embryonic cells, and the XCI states observed during culture of hPSC are highly variable. Recent studies have shown that hPSC rapidly progress to highly aberrant XCI patterns and that this process is probably driven by suboptimal culture conditions. Importantly, these aberrant XCI states seem to be inherited by the differentiated hPSC-progeny.

WIDER IMPLICATIONS: The aberrant XCI states (and epigenetic instability) observed in hPSC throw a shadow on their applicability as an in vitro model for development and disease modelling. Moreover, as the aberrant XCI states observed in hPSC seem to shift to a more malignant phenotype, this may also have important consequences for the safety aspect of using hPSC in the clinic.