Epigenetic Inheritance from Oocytes


Phenotypes can change in response to environment. Evolution occurs when altered phenotypes are selected and transmitted to the offspring. Oocytes and sperm (gametes) are responsible for transmitting phenotypic information to the next generation. The gametes contain the genome and the epigenome, which is the instruction manual for the genome. Not only the genome, but also a part of the epigenome is transmitted to the next generation and contributes to the phenotypes. The epigenome that is inherited across generations is called the "transgenerational epigenome”. Until recently, DNA methylation had been the only known transgenerational epigenome in mammals. However, our recent studies have revealed that chemical modifications to histone proteins (histone modifications) can also be transmitted from the oocyte to the next generation. Since the epigenome, unlike the genome, can change depending on the environment, the discovery of transgenerational histone modification raises the new possibility that environment-induced changes of histone modifications in oocytes can affect the phenotypes of the next generation and potentially contribute to evolution. In our laboratory, we study (1) the mechanisms of epigenome establishment in oocytes and its transmission to the next generation, (2) the biological significance of maternal epigenome transmission, and (3) the mechanism of disease predisposition via oocyte epigenome.

Keywords: Transgenerational epigenetic inheritance, Histone, Oocyte, Preimplantation embryo, Placenta


Preimplantation embryos

A new life begins when a sperm enters an oocyte. The two originally different cells are now fused to form a single cell, called zygote, in which two nuclei co-exist (paternal pronucleus and maternal pronucleus). The zygote divides to form a 2-cell embryo and the subsequent cleavages give rise to 4-cell, morula (8-16 cells), and blastocyst (32-128 cells) embryos. This developmental process takes place in the oviduct and the uterus before implantation and therefore is called ‘pre-implantation development’. Preimplantation embryos can be cultured in vitro up to the blastocyst stage, and they can give rise to live animals after transplanted into surrogate mothers. Such 'ex vivo' permissive characteristics of preimplantation embryos allow researchers to easily manipulate gene function in a physiological context. Importantly, the epigenetic landscape is genome-widely reprogrammed in preimplantation embryos, providing an excellent model to study epigenetics.


The epigenetics field is getting bigger and bigger. Epigenetics is nowadays an infrastructure to study life sciences. There are various definitions of epigenetics, but my understanding of epigenetics is a gene regulatory mechanism beyond the layers of genomic sequence and transcription factors. The nature of epigenetics include chemical modifications on DNA and histones. Epigenetic modifications compartmentalizes chromatin into transcriptionally-permissive and -repressive states. Each cell type has its own epigenetic landscape to enable a cell-type specific gene expression program. See more information about Epigenetics elsewhere if needed.


Epigenetic reprogramming

The epigenetic landscapes of sperm and oocytes are very different from each other. Shortly after fertilization, the sperm-derived paternal chromatin and the oocyte-derived maternal chromatin show epigenetic modifications totally different from each other. During preimplantation development, the epigenetic asymmetry between the parental chromatins is gradually resolved. This resetting process, called epigenetic reprogramming, is believed to be important to prepare a white canvas to establish new epigenetic information in a cell type-specific manner during cellular differentiation at post-implantation. 


Reprogramming resistance & Genomic imprinting

While the epigenetic landscapes of gametes are genome-widely reprogrammed during preimplantation development, certain genomic loci escape the reprogramming wave. Such reprogramming resistance results in transmission of parental epigenetic information into the next generation. When the epigenetic modifications at reprogramming resistant loci are parent-of-origin specific, it causes imprinted expression. The representatives of these loci are imprinting control regions that are marked by oocyte- or sperm-specific DNA methylation.  


Maternal inheritance of histone modification

Since the first identification of imprinted genes in mammals in 1991, DNA methylation had been the only known mechanism governing genomic imprinting. Nevertheless, a few studies have demonstrated that several paternally-expressed imprinted genes (PEGs) maintain their imprinted expression even in the lack of oocyte DNA methylation, suggesting the existence of a DNA methylation-independent imprinting mechanism. In 2017, we discovered that oocyte H3K27me3 is transmitted into embryos and the oocyte-derived H3K27me3 controls all of the PEGs reported to be independent of maternal DNA methylation (Inoue et al., 2017 NatureInoue et al., 2017 Genes Dev). These PEGs include Xist, an X-linked long non-coding RNA critical for X chromosome inactivation (XCI). Loss of maternal H3K27me3 leads to bi-allelic expression of Xist and bi-allelic XCI in preimplantation embryos. Interestingly, while DNA methylation-mediated imprinting is generally maintained in both embryonic and extra-embryonic cell lineages, maternal H3K27me3-mediated imprinting is maintained only in extra-embryonic cells such as placentae. A mouse model in which H3K27me3-, but not DNA methylation-, mediated imprinting is disrupted shows developmental retardation at post-implantation, embryonic sublethality, and placental enlargement (Inoue et al., 2018 Genes Dev, Mei et al., 2021 Nat Genet), demonstrating the importance of H3K27me3-mediated imprinting in mouse development and placental size regulation. It is fascinating to study how maternal H3K27me3-mediated imprinting is regulated during development, what its functions are, and why it has to be histone, but not DNA methylation.


Recent progress(Nat Genet 2021)

To investigate how Polycomb repressive complex 2 (PRC2)-mediated H3K27me3 is established during oogenesis and how it is passed on to the next generation, we studied the role of PRC1-mediated histone H2A lysine 119 mono-ubiquitination (H2AK119ub1). Using low-input CUT&RUN, we investigated the dynamics of H2AK119ub1 during oogenesis and early embryogenesis, and found that broad H2AK119ub1 domains are established, along with H3K27me3, during oocyte growth. On the other hand, these modifications differently behave after fertilization, with H2AK119ub1 being established prior to H3K27me3 at developmental genes that are typical PRC targets. As far as we know, this is the first report to show that H2AK119ub1 guides H3K27me3 in vivo.

   To determine whether H2AK119ub1 is required for establishment of H3K27me3 during oogenesis, we generated oocytes lacking polycomb group ring finger 1 (PCGF1) and PCGF6, which are essential components of variant PRC1. Pcgf1/6 KO oocytes showed that H2AK119ub1 is globally reduced and H3K27me3 is not established at a subset of genes (about 2,000 genes) (Below Fig. A). Profiling of H3K27me3 distribution in maternal KO (matKO) embryos, obtained by fertilization of Pcgf1/6 KO oocytes with WT sperm, revealed that H3K27me3-lost genes continued to be devoid of H3K27me3 after fertilization, even in the presence of active PRC2 (Fig. A). After implantation, Pcgf1/6 maternal KO embryos showed bi-alleic expression of non-canonical imprinted genes (Fig. B), partial embryonic lethality, and placental enlargement in the survivors (Fig. C). Thus, if H3K27me3 is failed to be established in oocytes, it can no longer be reversed even if H3K27me3-modifying enzymes are present. The phenotype of placental enlargement can be interpreted as the mother suppressing the size of the placenta via histone modifications in the oocytes. This means that non-canonical imprinting supports the conflict hypothesis (one of the evolutionary theories of genomic imprinting). This paper enriched interesting aspects of Polycomb molecular biology and genomic imprinting.

Nat Genet 2021 summary

Current research directions

1. Mechanisms of maternal epigenome establishment.

2. Functions and regulatory mechanisms of maternal H3K27me3-mediated imprinting

3. Mechanisms underlying intergenerational inheritance of maternal metabolic disorders. 


Unique research tools

In addition to general molecular biology tools, we have unique research tools as listed below. The ability to integrate these allows us to conduct high-originality researches. 

1. Epigenome analysis technologies applicable to a small number of cells 

Low-input DNaseI-seq (Lu et al., 2016 Cell, Inoue et al., 2017 NatureDjekidel et al., 2018 Cell Rep)

Low-input ChIP-seq (Inoue et al., 2017 Genes Dev)

CUT&RUN (Inoue et al., 2018 Genes Dev, Mei et al., 2021 Nat Genet)

Low-input DNA methylome(Shen et al., 2014 Cell Stem Cell, Inoue et al., 2015 Cell Rep


2. Micromanipulation & Developmental engineering

mRNA/siRNA injection, pronuclear isolation, pronuclear transfer, chromosome transfer, somatic cell nuclear transfer, in vitro fertilization (IVF), in vitro maturation (IVM), in vitro growth (IVG), intracytoplasmic sperm injection (ICSI), embryo transplantation, dissection of E6.5 epiblast/extra-embryonic ectoderm/visceral endoderm, etc.

3. Maternal protein depletion system

siRNA injection into GV oocytes followed by IVM-IVF

Since most of biological events occurring in zygotes and early preimplantation embryos are regulated by maternally-stored proteins, it is essential to deplete them to understand their mechanisms and functions. One way is to generate an oocyte-specific KO model, which is time-consuming and thereby extremely low throughput. An alternative is RNA interference-mediated knockdown of interested genes. Although siRNA injection into MII-stage oocytes is commonly used due to its relative easiness, it often fails to deplete maternal proteins because these proteins have already been accumulated in MII oocytes. To achieve efficient knockdown of maternal proteins, we inject siRNA into GV-stage oocytes and then let the oocytes reach the MII-stage using in vitro maturation (IVM). This allows a 20-42 hrs time window for injected siRNA to act before fertilization. The usefulness of this approach has been proved by our past works in which maternally-stored proteins, Tet3 and Nyfa, were efficiently depleted (Inoue et al., 2012 Cell Res; Lu et al., 2016 Cell). The detailed protocol will be distributed upon request.


siRNA injection into growing oocytes followed by IVG-IVM-IVF

Nevertheless, it sometimes happens that the above approach fails to knockdown targeted maternal proteins when the proteins are highly stable. To overcome this issue, we inject siRNA into growing oocytes followed by in vitro growth (Inoue et al, 2014 Protocol Ex).  In this system, secondary follicles are collected from ovaries of 12 day-old mice, oocytes within the follicles are injected with siRNA, and then the follicles are cultured in vitro for 12 days until the oocytes reach the fully-grown GV-stage. Then, the oocytes are matured in vitro (IVM) and fertilized in vitro (IVF) to obtain zygotes. This system provides a long time window (~2 weeks) for injected siRNA to act. The injected siRNA can act for such a long period because the oocyte is a non-dividing cell and siRNA is not diluted by cell division. By using this method, we have succeeded to knockdown NPM2 and HIRA in mouse oocytes (Inoue and Aoki, 2010 FASEB J; Inoue and Zhang, 2014 NSMB).