During the period of the ERC starting grant, we set to provide functional insight into the dynamics and function of DNA methylation during the course of embryonic development. Results of our work have all been published in a freely accessible maner.
1. To create the first fully time-resolved model for mouse gastrulation, we generated single-cell transcriptome data from over 150 individually isolated and processed embryos (Mittnenzweig, Mayshar et al., Cell. 2021). Since each embryo independently progresses along the same trajectory, cells making up each embryo could be placed on a time continuum. This facilitated a unified flow model of mouse gastrulation. The main findings of this study include: (i) We established scRNA-seq as a precise method for timing embryo development. (ii) Our work provides a holistic model of development that considers continuous, parallel, and converged differentiation towards multiple lineages. (iii) The flows model provides answers to many open questions on fate acquisition in the embryo. For example, it explains the pluripotency gradient in the differentiating epiblast and the lack of a stable meso-endoderm progenitor state. (iv) We show that at the population level - development is not a series of “yes/no” decisions. Rather, our model emphasizes semi-stable progenitor states that continuously multi-furcate in a time-dependent manner. (v) The time resolved atlas is demonstrated to be extremely useful to characterize mutant phenotypes with time-matched controls. Similarly, this model can also serve to evaluate the fidelity of stem-cell-cased embryo models, as we demonstrated in Lau, Rubinstein et al. (Cell Stem Cell 2022).
2. Building on this model, we focused on studying the regulation of DNA methylation as a paradigm for the emergence of epigenetic programs during gastrulation. As genes of interest, we focused on the Tet family of DNA demethylation enzymes. Importantly, Tet-null embryos fail to gastrulate, limiting the possibility to analyze cell-autonomous gene function. By comparative analysis of normal development, whole-embryo knockouts, and chimeras, distinguishing direct cell-autonomous vs. indirect non-cell-autonomous effects we could work out the precise functions of the Tet machinery (Cheng et al., Cell. 2022).
3. Imprinted genes are widely considered dosage-sensitive and nearly all imprinting clusters contain an intricate interplay between paternally and maternally expressed genes. In Weinberg-Shukron et al. (Nat Communications, 2022), we provide a significant advance in understanding mechanisms controlling imprinted genes and their impact on embryonic development. Our data shows that correct parent-of-origin imprinting pattern is secondary to balanced gene dosage. More provocatively, it suggests that an epigenetic state is irrelevant as long as appropriate gene expression is maintained at imprinted loci.
4. To lay the foundation for using the rabbit as a comparative system for understanding genome and cell regulation, we established a time-resolved atlas of rabbit gastrulation. By directly aligning to our previous mouse model, we were able to identify conserved and diverged processes over a shared time axis. Thus, the time course of different cellular lineages alongside the dynamics of the key genes involved could be determined. Many trajectories were found to be highly conserved, whereas others, mainly of extraembryonic tissues, but also primordial germ cells, were found to be highly species-specific (Mayshar et al., Cell 2023).
5. Chromatin properties such as three dimensional organization and genome regulation are influenced significantly by histones and their post-translational modifications. By measuring histone turnover in steady state, we could measure the exchange of different histone variants and their functional association, such as with transcription, both in mouse stem cells and in animals (Dunjic et al., Nat Communications, 2023).
6. By generating a model that combines placental development alongside gastrulation we revealed previously uncharacterized early cell specification events. This further allowed us to probe the functional importance of extraembryonic gene perturbation, and determine the fidelity of placentation in ex-utero and synthetic embryo model systems. Using a variety of gene perturbation strategies we could refine the roles of the key signaling peptide BMP4 for germline specification, identifying that its time-dependent production from different sources is at the same time crucial for their initial differentiation, but also in restricting their relative numbers (Hadas et al., Nature 2024).
