Targeted depletion of uterine glandular Foxa2 induces embryonic diapause in mice

  1. Mitsunori Matsuo
  2. Jia Yuan
  3. Yeon Sun Kim
  4. Amanda Dewar
  5. Hidetoshi Fujita
  6. Sudhansu K Dey
  7. Xiaofei Sun

  • Curated by eLife

    eLife logo

    Evaluation Summary:

    This manuscript reports the complex interactions that take place in the uterus between the endometrium and the blastocyst during and after embryonic diapause, a period of suspended animation that occurs in some mammals including the mouse, the model used here. The authors showed that one gene, Foxa2, interacts with two other genes, Msx1 and LIF, to control the success and duration of diapause. This will be of broad interest to researchers in the field of developmental biology and reproduction. It is a carefully done study, providing new information on the complex process that is diapause in which an embryo goes into suspended animation until it receives appropriate signaling from the uterine endometrial secretions to reactivate.

    (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. Reviewer #2 agreed to share their name with the authors.)

Reviewed by

Read the full article See related articles

Listed in

Log in to save this article

Abstract

Embryonic diapause is a reproductive strategy in which embryo development and growth is temporarily arrested within the uterus to ensure the survival of neonates and mothers during unfavorable conditions. Pregnancy is reinitiated when conditions become favorable for neonatal survival. The mechanism of how the uterus enters diapause in various species remains unclear. Mice with uterine depletion of Foxa2, a transcription factor, are infertile. In this study, we show that dormant blastocysts are recovered from these mice on day 8 of pregnancy with persistent expression of uterine Msx1 , a gene critical to maintaining the uterine quiescent state, suggesting that these mice enter embryonic diapause. Leukemia inhibitory factor (LIF) can resume implantation in these mice. Although estrogen is critical for implantation in progesterone-primed uterus, our current model reveals that FOXA2-independent estrogenic effects are detrimental to sustaining uterine quiescence. Interestingly, progesterone and anti-estrogen can prolong uterine quiescence in the absence of FOXA2. Although we find that Msx1 expression persists in the uterus deficient in Foxa2, the complex relationship of FOXA2 with Msx genes and estrogen receptors remains to be explored.

Article activity feed

  1. Version published to 10.7554/elife.78277 on eLife
  2. eLife

    Author Response

    Reviewer #2 (Public Review):

    The authors have used every possible combination and permutation of treatments at different stages of diapause and post diapause development in the mouse and used conditional gene knockouts at different stages to tease out the interactions of Foxa2 with Msx1 and LIF in the reactivation and implantation process in mice. The authors extend diapause further after treatments with progesterone and an estrogen-degrading chemical to show that this will prolong diapause in the presence of Msx1. Overall this study advances our knowledge of the cross-talk between uterine endometrium and the blastocyst during and after the remarkable phenomenon that is diapause.

    Strengths

    Demonstrating that Msx1 is critical to maintaining diapause, and that diapause is maintained in Foxa2 deficient mice have clarified their interactions. It is interesting that LIF triggers implantation on day 8 but cannot support the pregnancy to full term. Suppression of the estrogen effects by progesterone or fulvestrant increases the duration of diapause. Demonstrating that Foxa2 induces diapause via interactions with MSX1 shows Foxa2 plays such an important role in the control of diapause and adds another 'cog' to the complex wheel of its control.

    Weaknesses

    There is an assumption that everyone will understand the various manipulations that are done in this study - some effort needs to be made to clarify each experimental stage. How long are the embryos viable after the extension of the diapause by the various manipulations.

    The very positive review by a well-known expert in the field of diapause is reassuring, and we agree with her suggestions to improve the quality of the manuscript. As recommended, we now provide a scheme to summarize our findings to illustrate the length of embryo dormancy (see Fig. 7).

    Reviewer #3 (Public Review):

    Matsuo et al. have authored a manuscript describing the effects of depletion of the forkhead box gene, Foxa2, on embryogenesis and gestation in the mouse. The effects of this treatment are the induction of the diapause arrest in the development of the embryo and consequent dormancy. The manuscript is wellprepared, and the figures, for the most part, are didactic and interpretable. Although the conclusions are interesting, the principal weaknesses of the manuscript are the lack of novelty and the perceived absence of some controls and follow-up experiments.

    Controls and Follow-ups:

    1. The Cre/lox system depletes rather than deletes genes. Although in situ data are presented, these are not judged to be quantitative. The usual qPCR analysis of tissues could have established the quantity of depletion. Stupid but can be done. This is important because the frequency of implantation sites in both Cre/lox models (lines 111-113) may be attributable to the residual expression of Foxa2.

    The Foxa2f/fPgrCre/+ and Foxa2f/fLtfCre/+ mouse models used in the current study have been used in the previous studies (refs 7 and 8 in the manuscript). The deletion efficiency of Foxa2 in Foxa2f/fPgrCre/+ mice was examined by RT-PCR and IHC (figure 2 in ref 7); while the deletion efficiency in Foxa2f/fLtfCre/+ mice was examined by IHC (figure S1 in ref 8). The deletion efficiency has been proven by hundreds of publications since the generations of Pgr-cre in 2005 and Ltf-cre mice in 2014.

    Although these mouse lines have been used before, we confirmed the deletion of Foxa2 at the beginning of our study at protein levels (fig 1c) and RNA levels (fig 1d). We understand that the reviewer is trying to link the observation that some of the knockout animals still carried implantation sites on day 8 of pregnancy with the possibility that the deletion of Foxa2 is not complete. However, it is not uncommon to observe such phenotypes that are not fully penetrant even in systemic knockout mouse models. Nonetheless, we now provide real time PCR results of uterine Foxa2 on day 4 of pregnancy in all mouse models used in the current manuscript in the new supplemental figure 1.

    1. The most novel and salient finding of the present study is that the depletion of Foxa2 results in embryos that are in a state that "morphologically resembled dormant blastocysts". A useful experiment would have been to transplant these embryos to normal recipients or to culture them in vitro to determine whether they were capable of reactivation from the dormant state.

    Whether dormant embryos in Foxa2f/fPgrCre/+ and Foxa2f/fLtfCre/+ uteri can be reactivated is the main question we studied. The results in figures 4-6 address this question. The blastocysts in Foxa2f/fPgrCre/+ and Foxa2f/fLtfCre/+ uteri can be activated on day 4 as shown in figure 4b. Without any support, blastocysts in Foxa2f/fLtfCre/+ uteri still can be reactivated on day 8 (figure 4b). In the following experiments and results shown in figures 5 and 6, we tried to improve the uterine environment by supplementing progesterone and estrogen. Dormant embryos are successfully re-activated by a LIF injection and the pregnancies proceeded to full terms.

    This reviewer suggests using normal recipients to test the reactivation of dormant embryos. Given dormant embryos can be reactivated in a knockout uterine environment, embryo transfer experiments using normal recipients are an addition measure to test the integrity of embryonic dormancy. The embryo transfer experiments may be futile attempt in our studies because of the following reasons.

    The numbers of mated mutant females that yield blastocysts are relatively meager and so are the numbers of blastocysts recovery, especially from diapausing donors. It is well known that implantation rates after blastocyst transfer are compromised due the surgical trauma and anesthesia. Therefore, the results from these experiments may not provide meaningful information.

    Furthermore, during the pandemic our mouse colonies were drastically reduced, and we are still recovering from this downturn during this “New Normal”. Notably, pregnancy rate fluctuates throughout the year even if mice are housed in a controlled environment, and pregnancy rate is often relatively poor in mutant mice which of course depend on the genetic background and diets (DOI: 10.1126/scisignal.aam9011). Most importantly, viability of diapausing embryos is amply evident from our experiments (Figs. 4-6)

    1. Figure 3C indicates that embryos recovered on Day 8 had an extensive proliferation of ICM cells, but not trophoblast. Previous studies have explored the progression of entry and exit from diapause in the mouse (DOI: 10.1093/biolre/ioz017) showing that reactivation of the embryo from diapause commences in the ICM and then proceeds to the trophoblast. It therefore may be possible that proliferation in the trophoblast is not suspended, rather than the recovered blastocyst has resumed development and that mitotic activity has not yet reached the trophoblast.

    It is common to see KI67 expression in the ICM of dormant embryos. Figure 4D from the paper quoted by this reviewer presents Ki67 staining on embryos undergoing diapause at different stages. In our study, we showed Ki67 staining on dormant embryos collected on day 8, which equals D7.5 in their figure. Our data in figure 3C is consistent with observation shown. Without LIF, embryos remain dormant in Foxa2f/fPgrCre/+ and Foxa2f/fLtfCre/+ uteri.

    1. In Figure 4B, neither the Ltf nor the Pgr Cre treated uteri appear normal on Day 8. This is not consistent with the conclusion in lines 170 et seq. of the manuscript. It is difficult to discern normality from Figure 4C, but it is clear that the PgrCre-lox uterus does not conform to the controls. It is later noted that there is edema in the uteri at this time in the Day 8-treated PgrCre/lox mice (lines 217-218).

    We have clarified our description.

    Lines 173-176: Notably, implantation sites with a normal appearance were observed in Foxa2f/fLtfCre/+ uteri when LIF was given on day 8 of pregnancy (Figure 4b), albeit Foxa2f/fPgrCre/+ uteri with edema have only faint blue bands. Histology of implantation sites confirmed this observation.

    In line 217, we stated that “the uterine edema in Foxa2f/fPgrCre/+ females two days after LIF injection on day 8…”. Figure 4B showed that Foxa2f/fPgrCre/+ uteri with edema have some very faint blue bands suggesting implantation-like reaction. But we do not think they are real implantation, which is confirmed by figures 4c and e.

    1. In Figure 6B, the implantation sites appear substantially smaller in mice of both mutant genotypes. Supplemental Figure 4 suggests that this is not the case. It is unclear whether the samples chosen for figures are representative of the uteri and whether variation in the size of implantation sites was observed.

    In figure 6B, the Foxa2f/f uteri samples were collected on day 10 of pregnancy, which is same as when Foxa2f/fPgrCre/+ and Foxa2f/fLtfCre/+ tissues were collected. Since embryos implanted in Foxa2f/f uteri on day 4 night but in Foxa2f/fPgrCre/+ and Foxa2f/fLtfCre/+ uteri on day 8 after LIF injections, the implantation sites are bigger in Foxa2f/f uteri. However, in supplemental figure 4 the implantation sites were collected from Foxa2f/f females on day 6 of pregnancy, which show similar size as compared to implantation sites collected from Foxa2f/fPgrCre/+ and Foxa2f/fLtfCre/+ females 2 days after LIF injection.

  3. eLife

    Evaluation Summary:

    This manuscript reports the complex interactions that take place in the uterus between the endometrium and the blastocyst during and after embryonic diapause, a period of suspended animation that occurs in some mammals including the mouse, the model used here. The authors showed that one gene, Foxa2, interacts with two other genes, Msx1 and LIF, to control the success and duration of diapause. This will be of broad interest to researchers in the field of developmental biology and reproduction. It is a carefully done study, providing new information on the complex process that is diapause in which an embryo goes into suspended animation until it receives appropriate signaling from the uterine endometrial secretions to reactivate.

    (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. Reviewer #2 agreed to share their name with the authors.)

  4. eLife

    Reviewer #1 (Public Review):

    The authors report in this study that uterine deletion of Foxa2, a transcription factor, leads to embryonic diapause. The phenotypic characterization was well done and the data are of quality and convincing. However, the study remains highly descriptive due to a lack of molecular mechanisms. Consequently, the exact role of Foxa2 in LIF induction by E2 remains unknown.

  5. eLife

    Reviewer #2 (Public Review):

    The authors have used every possible combination and permutation of treatments at different stages of diapause and post diapause development in the mouse and used conditional gene knockouts at different stages to tease out the interactions of Foxa2 with Msx1 and LIF in the reactivation and implantation process in mice. The authors extend diapause further after treatments with progesterone and an estrogen-degrading chemical to show that this will prolong diapause in the presence of Msx1. Overall this study advances our knowledge of the cross-talk between uterine endometrium and the blastocyst during and after the remarkable phenomenon that is diapause.

    Strengths
    Demonstrating that Msx1 is critical to maintaining diapause, and that diapause is maintained in Foxa2 deficient mice have clarified their interactions. It is interesting that LIF triggers implantation on day 8 but cannot support the pregnancy to full term. Suppression of the estrogen effects by progesterone or fulvestrant increases the duration of diapause. Demonstrating that Foxa2 induces diapause via interactions with MSX1 shows Foxa2 plays such an important role in the control of diapause and adds another 'cog' to the complex wheel of its control.

    Weaknesses
    There is an assumption that everyone will understand the various manipulations that are done in this study - some effort needs to be made to clarify each experimental stage.
    How long are the embryos viable after the extension of the diapause by the various manipulations.

  6. eLife

    Reviewer #3 (Public Review):

    Matsuo et al. have authored a manuscript describing the effects of depletion of the forkhead box gene, Foxa2, on embryogenesis and gestation in the mouse. The effects of this treatment are the induction of the diapause arrest in the development of the embryo and consequent dormancy. The manuscript is well-prepared, and the figures, for the most part, are didactic and interpretable. Although the conclusions are interesting, the principal weaknesses of the manuscript are the lack of novelty and the perceived absence of some controls and follow-up experiments.

    Controls and Follow-ups:
    1. The Cre/lox system depletes rather than deletes genes. Although in situ data are presented, these are not judged to be quantitative. The usual qPCR analysis of tissues could have established the quantity of depletion. This is important because the frequency of implantation sites in both Cre/lox models (lines 111-113) may be attributable to the residual expression of Foxa2.
    2. The most novel and salient finding of the present study is that the depletion of Foxa2 results in embryos that are in a state that "morphologically resembled dormant blastocysts". A useful experiment would have been to transplant these embryos to normal recipients or to culture them in vitro to determine whether they were capable of reactivation from the dormant state.
    3. Figure 3C indicates that embryos recovered on Day 8 had an extensive proliferation of ICM cells, but not trophoblast. Previous studies have explored the progression of entry and exit from diapause in the mouse (DOI: 10.1093/biolre/ioz017) showing that reactivation of the embryo from diapause commences in the ICM and then proceeds to the trophoblast. It therefore may be possible that proliferation in the trophoblast is not suspended, rather than the recovered blastocyst has resumed development and that mitotic activity has not yet reached the trophoblast.
    4. In Figure 4B, neither the Ltf nor the Pgr Cre treated uteri appear normal on Day 8. This is not consistent with the conclusion in lines 170 et seq. of the manuscript. It is difficult to discern normality from Figure 4C, but it is clear that the PgrCre-lox uterus does not conform to the controls. It is later noted that there is edema in the uteri at this time in the Day 8-treated PgrCre/lox mice (lines 217-218).
    5. In Figure 6B, the implantation sites appear substantially smaller in mice of both mutant genotypes. Supplemental Figure 4 suggests that this is not the case. It is unclear whether the samples chosen for figures are representative of the uteri and whether variation in the size of implantation sites was observed.

  7. Version published to 10.1101/2022.03.15.484401v1 on bioRxiv