Dynamic chromatin accessibility unveils a regulatory landscape in early embryogenesis in human

One of the most important events after fertilization is zygotic gene activation (ZGA), when the quiescent zygotic genes begin to be expressed and to contribute to the proper regulation of embryo development. After several rounds of divisions, the embryo enters the morula stage, in which cell lineages are specified and pluripotent cells are subsequently established in the inner cell mass (ICM) [1]. It has been suggested that epigenetic reprogramming plays a critical role in early embryogenesis and is required for ZGA and normal embryonic development [2–8].

Despite its importance, current research on gene regulation during early mammalian embryogenesis has been limited, mainly due to the technical challenges and the scarcity of materials. Over the last couple of years, however, there has been some breakthrough in mice. For example, using reduced representation bisulfite-sequencing, Alexander Meissner's group observed dramatic DNA demethylation upon fertilization and genome-wide hypomethylation at the zygote stage [2]. Only subtle global demethylation from zygote to the ICM stage was reported. In 2016, Yi Zhang's and Wei Xie's groups separately generated open-chromatin profiles in mice through modified protocols with DNase-seq and ATAC-seq [4, 5]. What they found is that the accessible chromatin regions in zygotes are low, but upon ZGA at the 2-cell stage, the number of open chromatin regions increases gradually and there is a big leap between the 4-cell and 8-cell stage. Further, Yi Zhang's group also observed that transcription factors NFYA and OCT4 contribute to DHS increase in ZGA and the 8-cell stage, respectively.

Figure 1.

Schematic of the establishment of accessible chromatin map and gene expression in human early embryogenesis. The number of DHSs gradually increases from 2-cell embryos to blastocyst. OCT4 plays a critical role in ZGA at the 8-cell stage. Evolutionary older genes tend to be expressed at earlier stages and younger genes tend to be expressed at later stages. Certain active human transposons, including HERV-Ks, HERV-Hs and SVAs, are more accessible in early embryos and lead to higher expression level during early embryogenesis

Although human and mice share a lot of similarities in embryogenesis, they do have some important differences. The major ZGA in human occurs at the 8-cell stage, while in mice it happens at the 2-cell stage [9]. It has also been reported that global DNA demethylation is almost complete at the 2-cell stage in human [3], but in mice, subtle global demethylation continues until the ICM stage [2]. Therefore, it is critical to study epigenomic regulation in human embryogenesis. Further, from the perspective of evolution, it is also interesting to identify the conserved and divergent events between the two species.

To fill in these gaps, using in vitro fertilized embryos, Dr. Jiang Liu's group from Beijing Institute of Genomics has now profiled chromatin accessibility for the first time in human early embryos at different timepoints, from the 2-cell stage through blastocyst [10]. The technique they used, termed liDNase-seq, is a low-input DNase-seq that requires as few as 30 cells and has been used previously to study mouse early embryogenesis [4]. They found that the chromatin accessibility map is progressively established, and the largest portion of DNase I hypersensitive sites (DHSs) detected at each stage is located within intergenic regions. To establish the relationship between enhancers and gene expression, they performed RNA-seq experiment in the same system. They observed that the majority of ZGA genes harbor promoter DHSs and that the number of promoter DHSs is consistent with the timing of ZGA. Genes with promoter DHSs have higher expression level compared to those without promoter DHSs, and this correlation becomes more pronounced from the 8-cell stage onward. Interestingly, they found that a small proportion of genes are primed at earlier stage and expressed at later stage.

Next, the authors performed Gene Ontology (GO) analyses and showed that the promoter DHSs at later stages are strongly associated with embryonic development. GO terms such as DNA and RNA processing are enriched for stages before morula, while pathways related with cell development and cell fate commitment begin to show enrichment in the morula stage. By comparing enriched transcription factor binding motifs before and at the time of ZGA, Jiang's group hypothesized that OCT4 is a crucial regulator of ZGA. Therefore, they performed siRNA knockdown of OCT4 in zygotes, which induced downregulation of 25% of ZGA genes at the 8-cell stage. Among them is SOX2, whose expression is reduced by 80%. To further study the role of SOX2, they knocked down SOX2 and found most of the genes downregulated by SOX2 knock-down (KD) embryos overlap with those in OCT4 KD embryos.

Previous work has shown that major global DNA demethylation is complete in human 2-cell stage embryos [3]. In this work, the authors showed that global DHS signals and DNA methylation levels are negatively correlated. Further, DHSs establishment at promoter regions is affected by CpG density: the lower the CpG density is, the higher the methylation level is and the later stage promoter region will be open.

To study whether the regulatory mechanisms in embryogenesis are conserved in evolution, the authors performed the same set of experiments (including liDNase-seq and RNA-Seq) at the comparable embryonic stages in mice. While most of the findings in human and mice are consistent, the authors did observe some key differences. As previously mentioned, the largest proportion of DHSs detected at each stage in human is within intergenic regions, while in mice the majority of DHSs localize at promoter regions at the 2-cell stage. siRNA knockdown of OCT4 in zygotes causes the downregulation of 25% ZGA genes in humans, but in mice, only 1.4% of ZGA genes are downregulated at the 4-cell stage, suggesting there are other important players involved in ZGA.

Transposons compose nearly half of the human genome, but only a small portion of them is active. In this work, Jiang's group further studied several active human transposons, including HERV-Ks, HERV-Hs, and SVAs. They showed that the chromatin of these transposons is in a more accessible state in early embryos, which thus leads to a higher expression level during early embryogenesis. Mutations caused by mobility of active transposons in early embryos may be passed to germline cells and offspring and may be beneficial in human evolution. Furthermore, evolutionarily older genes tend to establish DHSs and be expressed at earlier stages, as compared with younger genes. The pathways for younger genes are consistent with the time of embryonic development: for example, immune response is enriched in blastocyst embryos.

Briefly, this is the first time that the landscape of chromatin accessibility in human early embryos has been reported. This set of data will be invaluable for further investigations of the underlying mechanism of ZGA regulation and the possible role of transposons in human evolution. With the advent of new techniques, we expect to see more types of epigenomic data such as histone modifications and chromatin interaction data, which will no doubt deepen our understanding of the gene regulation events in embryogenesis.