Transcriptional heterogeneity and cell cycle regulation as central determinants of Primitive Endoderm priming

Curation statements for this article:
  • Curated by eLife

    eLife logo

    Evaluation Summary:

    This study uses media-based conversion of stem cell cultures towards to investigate how cell cycle regulation affects the transition of cell populations between pluripotent and differentiated states. Through a detailed analysis of cell cycle properties in different primed subpopulations, under a range of growth conditions, the authors propose that both the maintenance of pluripotency as well as the conversion towards a more differentiated state is influenced by selective shortening of the cell cycle in different primed subpopulations. By using new reporter systems and long-term imaging, this study thus sheds new light on the old question of whether extracellular signals control differentiation in cell populations through selection or induction.

    (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. The reviewers remained anonymous to the authors.)

This article has been Reviewed by the following groups

Read the full article

Abstract

During embryonic development cells acquire identity as they proliferate, implying that an intrinsic facet of cell fate choice requires coupling lineage decisions to cell division. How is the cell cycle regulated to promote or suppress heterogeneity and differentiation? We explore this question combining time lapse imaging with single-cell RNA-seq in the contexts of self-renewal, priming, and differentiation of mouse embryonic stem cells (ESCs) towards the Primitive Endoderm (PrE) lineage. Since ESCs are derived from the inner cell mass (ICM) of the mammalian blastocyst, ESCs in standard culture conditions are transcriptionally heterogeneous containing dynamically interconverting subfractions primed for either of the two ICM lineages, Epiblast and PrE. Here, we find that differential regulation of cell cycle can tip the balance between these primed populations, such that naïve ESC culture promotes Epiblast-like expansion and PrE differentiation stimulates the selective survival and proliferation of PrE-primed cells. In endoderm differentiation, this change is accompanied by a counter-intuitive increase in G1 length, also observed in vivo . While fibroblast growth factor/extracellular signal-regulated kinase (FGF/ERK) signalling is a key regulator of ESC differentiation and PrE specification, we find it is not just responsible for ESCs heterogeneity, but also the inheritance of similar cell cycles between sisters and cousins. Taken together, our results indicate a tight relationship between transcriptional heterogeneity and cell cycle regulation in lineage specification, with primed cell populations providing a pool of flexible cell types that can be expanded in a lineage-specific fashion while allowing plasticity during early determination.

Article activity feed

  1. Evaluation Summary:

    This study uses media-based conversion of stem cell cultures towards to investigate how cell cycle regulation affects the transition of cell populations between pluripotent and differentiated states. Through a detailed analysis of cell cycle properties in different primed subpopulations, under a range of growth conditions, the authors propose that both the maintenance of pluripotency as well as the conversion towards a more differentiated state is influenced by selective shortening of the cell cycle in different primed subpopulations. By using new reporter systems and long-term imaging, this study thus sheds new light on the old question of whether extracellular signals control differentiation in cell populations through selection or induction.

    (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. The reviewers remained anonymous to the authors.)

  2. Reviewer #1 (Public Review):

    In this manuscript the authors set out to characterise the process of differentiation of inner cell mass cells within the mouse blastocyst into either epiblast or primitive endoderm, which is a binary fate choice, using various models. To this end, they made use of well-established reporter cell lines previously generated in their lab as well as a widely used fluorescent system (FUCCI) that allows stages in the cell cycle to be visualised and sorted. The experimental output was compared with computational models and published data generated from mouse embryos during the process of primitive endoderm and epiblast segregation. Their data uncovered interesting mechanistic insight into the dynamics of the cell cycle and how these correlate with lineage choice and amplification. The methods have been carefully considered and validated in previous work by the group and the analysis is thorough. The single cell profiling is particularly well presented, and backed up by immunofluorescence data using well-characterised lineage reporters with appropriate statistical analysis. Probably the most interesting finding, which the authors identify as unexpected, is the considerable lengthening of the G1 part of the cell cycle in cells differentiating into PrE, but coinciding with a reduction in overall cell cycle length. Also, cell cycle length from mother to daughter cells in all conditions appears not to be inherited, yet sister cells, and to a lesser extent, cousins, appear to retain similar cell cycle dynamics. This feature is attributed to differential levels of FGF, suggested by the use of PD03 or PD17 as downstream inhibitors. Not surprisingly, levels of the PrE-associated factor Hex could predict the likelihood of differentiation to PrE, but also higher levels of Hex correlated with a shorter cell cycle. Also, blocking MEK/ERK signalling increased cell cycle duration as well as reducing differentiation to PrE in the culture conditions designed to promote differentiation to epiblast. The aims of the paper appear to be achieved and the results adequately support the authors' conclusions. A similar system to the one established here could be envisaged for downstream developmental processes, such as those involving binary decisions for specific tissue formation in organogenesis, but would require the generation and validation of different reporter cell lines.

  3. Reviewer #2 (Public Review):

    In this paper, the authors show that the maintenance of pluripotent mouse stem cell cultures and their conversion towards primitive endoderm relies on selective effects of specific culture media that act on the survival and cell cycle properties of different primed subpopulations. They further demonstrate that FGF/ERK signaling underlies correlations between cell cycle length in daughter cells, and identify characteristic, lineage-specific differences in relative G1 length as cells differentiate along the endodermal lineage that is recapitulated in vivo. The study is based on a technically challenging combination of long-term time-lapse imaging with reporters for cell cycle and lineage priming and delivers new insight into the old question of whether extracellular signals regulate differentiation in cell populations by selection or by induction.

    A central conclusion of the study, that media conditions control differentiation through cell cycle regulation, is based on the analysis of time-lapse imaging of cells in PrE differentiation and pluripotency conditions. Even though the authors acknowledge that cell death contributes to population composition, the manuscript mainly talks about changes to the cell cycle. This focus on the cell cycle does not do the data justice, especially in the context of PrE differentiation: The time-lapse movie shows that there is massive cell death from 24 h onwards, to a degree that there is no net growth of the population but rather a decline in cell numbers over the course of the experiment. This impression is supported by the lineage trees, where the NEDiff cells appear to selectively die, rather than being outcompeted by PrE cells. Thus, while cell cycle regulation clearly contributes to differentiation at the population level, it remains an open question how important this effect is in the chosen differentiation paradigm, compared to selective effects that act through cell survival.

  4. Reviewer #3 (Public Review):

    The manuscript of Birckman and colleagues tackles the link between lineage priming, lineage specification, and cell cycle in the ESCs culture. This is an interesting piece of work, with several noteworthy findings, that elegantly explain how lineage priming can be efficiently achieved during the changing cultural conditions. There are several interesting points raised by the authors, relating to lineage priming, cell specification, and cell cycle, that can be presented to the scientific community. Namely:

    • Differential regulation of the cell cycle can tip the balance between populations of cells primed to different cell fate choices (here PrE and Epi).

    • Different culture conditions favour acceleration/stimulation of the cell cycle of different cell populations.

    • Only a small population of cells from the original culture enters a differentiation process which is followed by selected expansion and/or survival of their progeny.

    • In the case of endodermal type specification (towards PrE), a shortening of the cell cycle is accompanied by the proportional relative increase of G1 phase length.

    • FGF activity is responsible for cell cycle synchronisation, required for the inheritance of similar cell cycles between sisters and cousins

    Unfortunately, in the current version of the manuscript, the authors try to create the impression that the relationship between cell cycle, heterogeneity and cell fate found in ESCs can be directly translated to the in vivo system. It is not clear, however, how easily and reliably the information about the cell cycle in ESCs can be translated to an in vivo setting. The timeline of PrE vs Epi specification in vivo and in vitro are completely different. In embryos, PrE is specified within 24h, whereas with in vitro it takes 6 days. I cannot see how these two timelines - and also different cell cycle lengths - can be reliably compared.

  5. Author Response

    Reviewer 1

    In this manuscript the authors set out to characterise the process of differentiation of inner cell mass cells within the mouse blastocyst into either epiblast or primitive endoderm, which is a binary fate choice, using various models. To this end, they made use of well-established reporter cell lines previously generated in their lab as well as a widely used fluorescent system (FUCCI) that allows stages in the cell cycle to be visualised and sorted. The experimental output was compared with computational models and published data generated from mouse embryos during the process of primitive endoderm and epiblast segregation. Their data uncovered interesting mechanistic insight into the dynamics of the cell cycle and how these correlate with lineage choice and amplification. The methods have been carefully considered and validated in previous work by the group and the analysis is thorough. The single cell profiling is particularly well presented, and backed up by immunofluorescence data using well-characterised lineage reporters with appropriate statistical analysis. Probably the most interesting finding, which the authors identify as unexpected, is the considerable lengthening of the G1 part of the cell cycle in cells differentiating into PrE, but coinciding with a reduction in overall cell cycle length. Also, cell cycle length from mother to daughter cells in all conditions appears not to be inherited, yet sister cells, and to a lesser extent, cousins, appear to retain similar cell cycle dynamics. This feature is attributed to differential levels of FGF, suggested by the use of PD03 or PD17 as downstream inhibitors. Not surprisingly, levels of the PrE-associated factor Hex could predict the likelihood of differentiation to PrE, but also higher levels of Hex correlated with a shorter cell cycle. Also, blocking MEK/ERK signalling increased cell cycle duration as well as reducing differentiation to PrE in the culture conditions designed to promote differentiation to epiblast. The aims of the paper appear to be achieved and the results adequately support the authors' conclusions. A similar system to the one established here could be envisaged for downstream developmental processes, such as those involving binary decisions for specific tissue formation in organogenesis, but would require the generation and validation of different reporter cell lines.

    We thank this reviewer for their support of our manuscript.

    Reviewer 2

    In this paper, the authors show that the maintenance of pluripotent mouse stem cell cultures and their conversion towards primitive endoderm relies on selective effects of specific culture media that act on the survival and cell cycle properties of different primed subpopulations. They further demonstrate that FGF/ERK signaling underlies correlations between cell cycle length in daughter cells, and identify characteristic, lineage-specific differences in relative G1 length as cells differentiate along the endodermal lineage that is recapitulated in vivo. The study is based on a technically challenging combination of long-term time-lapse imaging with reporters for cell cycle and lineage priming and delivers new insight into the old question of whether extracellular signals regulate differentiation in cell populations by selection or by induction.

    A central conclusion of the study, that media conditions control differentiation through cell cycle regulation, is based on the analysis of time-lapse imaging of cells in PrE differentiation and pluripotency conditions. Even though the authors acknowledge that cell death contributes to population composition, the manuscript mainly talks about changes to the cell cycle. This focus on the cell cycle does not do the data justice, especially in the context of PrE differentiation: The time-lapse movie shows that there is massive cell death from 24 h onwards, to a degree that there is no net growth of the population but rather a decline in cell numbers over the course of the experiment. This impression is supported by the lineage trees, where the NEDiff cells appear to selectively die, rather than being outcompeted by PrE cells. Thus, while cell cycle regulation clearly contributes to differentiation at the population level, it remains an open question how important this effect is in the chosen differentiation paradigm, compared to selective effects that act through cell survival.

    • We acknowledge that there is a large amount of cell death in the beginning of differentiation and believe this could be a response to changes in media and cytokines. This is also observed in other differentiation and reprogramming protocols (Ying and Smith 2003; Hayashi et al. 2011; Yasunaga et al. 2005; Argelaguet et al. 2019; Rugg-Gunn 2022).

    • While our original analyses were focused on cell cycle, we did not mean to imply that survival is irrelevant to selecting the correct populations at the end of differentiation. Although there is clearly an increase in NEDiff cell death, these rates fluctuate during differentiation (Table S2) and it is difficult to make any hard conclusions about the timing of selection. In addition, we find no signature for apoptosis in all scRNA-seq clusters associated to early NEDiff. However, we thank the reviewers for raising this issue and we address the issue of survival more extensively in the revised manuscript (lines140-142, and 342-347, in addition to a new modelling section described below).

    • To further explore when relative survival is most important for effective differentiation, we have reformulated our model. Our new modelling results show that changes in the survival rates have stronger impact shifting the cell proportions at later stages of differentiation. Phase diagrams in Figure 3- Supplement 1 show that at day 5 reaching an appropriate proportion of PrE requires less change in the survival rates than in doubling times, suggesting that survival can have greater impact on the ratio than cell division. However, at day 3, the modelling suggests the opposite, consistent with a first wave of a cell cycle regulation and PrE fastest division time (Fig. 2D). We have discussed this in the Results section, lines 162-168.

    • We have added survival alongside proliferation to the abstract.

    Reviewer 3

    The manuscript of Birckman and colleagues tackles the link between lineage priming, lineage specification, and cell cycle in the ESCs culture. This is an interesting piece of work, with several noteworthy findings, that elegantly explain how lineage priming can be efficiently achieved during the changing cultural conditions. There are several interesting points raised by the authors, relating to lineage priming, cell specification, and cell cycle, that can be presented to the scientific community. Namely:

    • Differential regulation of the cell cycle can tip the balance between populations of cells primed to different cell fate choices (here PrE and Epi).

    • Different culture conditions favour acceleration/stimulation of the cell cycle of different cell populations.

    • Only a small population of cells from the original culture enters a differentiation process which is followed by selected expansion and/or survival of their progeny.

    • In the case of endodermal type specification (towards PrE), a shortening of the cell cycle is accompanied by the proportional relative increase of G1 phase length.

    • FGF activity is responsible for cell cycle synchronisation, required for the inheritance of similar cell cycles between sisters and cousins

    Unfortunately, in the current version of the manuscript, the authors try to create the impression that the relationship between cell cycle, heterogeneity and cell fate found in ESCs can be directly translated to the in vivo system. It is not clear, however, how easily and reliably the information about the cell cycle in ESCs can be translated to an in vivo setting. The timeline of PrE vs Epi specification in vivo and in vitro are completely different. In embryos, PrE is specified within 24h, whereas with in vitro it takes 6 days. I cannot see how these two timelines - and also different cell cycle lengths - can be reliably compared.

    • We were aware of the difficulties inherent in this comparison and apologize if our statements were not sufficiently conditional. We have now added caveats to the results section (line 328) and a revised discussion that explicitly discusses the difference in time lines (lines 404-409).

    • On line 111, we added a clarification that our in vitro model is a tool for deconstructing “transcriptional signatures” as this is precisely what we measure both in vivo and in vitro.

    • On line 322 we clarified that even though the time frames are different, we found a similar G1 signature trend in our in vitro dataset than the one found in the in vivo published dataset.

    • On line 337 we amended the sentence to clarify that some aspects of cell cycle regulation found in vitro could be also exploited in vivo.

    • On line 414 we have added a caveat line explaining that the use of G1 to drive increasing levels of commitment is a facet of cell fate choice both in vivo and in vitro, even though the time scales are different.