New embryological perspectives on imprinting disease

Mammalian cells typically have two sets of chromosomes, one each inherited from the mother and the father. In most cases, both parental genes (alleles) have similar activities, but in an important minority (~0.5% of the total), the allele inherited from one parent is active relative to the corresponding allele from the other. These genes are said to be imprinted, and the balance of their expression is critical for embryonic viability and the avoidance of disease.

The chromatin marks that engender imprinted gene expression can take the comparatively stable form of genomic DNA methylation and have been known for some thirty years. Recently, it has been found that histones, which package genomic DNA, also carry imprints that seem only to last during preimplantation development; after this, expression from each parental allele becomes equivalent.

We recently documented both imprint types in the mouse by combining parent-discriminatory single embryo genome methylation and gene expression analyses. This work confirmed well-characterised imprints, but suggested that there exist multiple new imprinted regions that had not previously been reported, including 71 new imprinted genes called nBiX (for novel blastocyst-imprinted expressed). Mouse nBiX genes were marked by the modified histone (trimethylation of lysine 27 in histone 3, H3K27me3) and uniparentally expressed at the blastocyst stage before implantation, but equivalently expressed shortly thereafter: they are transiently imprinted.

nBiX genes are highly expressed in the developing brain and almost all have human counterparts that are associated with disease: many are involved in metabolic and membrane regulation, so their dysregulation manifests in diverse pathologies. Disruption of known imprinted genes is oncogenic, linking imprinted gene regulation and cancer, and there is increasing evidence that epigenetic dysregulation in early embryos results in adult disease: examples include nuclear transfer cloning (which produces obesity and other pathologies) and inter-generational epigenetic inheritance, in which parentally-acquired disease traits become heritable.

The present proposal seeks to test the unscrutinised hypothesis that transient disruption of imprinted (particularly nBiX) gene expression during the preimplantation window of uniparental expression causes disease. We will determine at high resolution which traits are dysregulated when nBiX expression levels are briefly disrupted during preimplantation development. This will be accomplished using embryological tools developed by us to increase or reduce nBiX transcript activity at specified stages in mouse preimplantation development.

From preliminary analysis, such transient disruption of nBiX expression impedes embryogenesis. We will extend this to a fuller analysis of the pathological consequences of nBiX disruption using molecular, cellular and physiological approaches, as it manifests at all stages pre- and post-implantation, peri-natal development and into adulthood. The mouse studies will be informed by in-depth analysis of parent-specific genome activity in human blastocysts, allowing us to map mouse nBiX dysregulation in disease onto human imprints.

There is currently no alternative experimental strategy to the one proposed reversibly to abrogate transient imprinted gene expression and reveal links to disease. The imprinting phenomenon is transient and subtle: H3K27me3 may be displaced long before disease is manifest, requiring new perspectives to study it. The work promises to reveal epigenetically-regulated mechanisms predisposing to cancer, and to establish a new and tractable embryonic model. More broadly it will show how the earliest events in embryonic development can effect epigenetic inheritance of disease traits, informing ART so that it may be adapted to minimise epigenetic contributions that predispose to disease.

We have combined micro-whole genome bisulphite sequencing of uniparental mouse embryos with RNA-seq of reciprocal crosses to identify 71 novel blastocyst-imprinted expressed (nBiX) genes with transient allelic expression asymmetry. Asymmetry is lost after implantation and mostly H3K27me3-dependent. The human orthologues of 85% of nBiX genes are associated with disease. However, the relationship between preimplantation imprinted gene expression and disease is poorly understood. We hypothesise that nBiX expression levels in preimplantation embryos are critical. This will be tested by reversibly altering nBiX transcript levels exclusively during preimplantation development, thereby isolating consequences due to early disruption from subsequent effects. nBiX transcript levels will be increased by coinjecting oocytes with sperm plus different concentrations of nBiX cRNA, and depleted by coinjecting oocytes with sperm plus either siRNA (RNAi) or morpholinos (blocking translation) that target nBiX transcripts. Outcomes will be assessed in a battery of tests to characterise molecular, cellular and developmental perturbations that occur during pre- and post-implantation development, and combining them with disease metrics (eg weight gain) post-term. This will relate specific transient preimplantation nBiX RNA perturbation with downstream phenotypic outcomes. We will determine parental expression bias in human blastocysts by ultra-deep RNA-seq with associated whole-exome sequencing. This will generate a unique human dataset, delineate imprinted human gene expression to reveal the degree of conservation to mouse imprinted genes, and inform the selection of nBiX targets. It will also reveal the relationship between early embryonic imprinted (nBiX) gene dysregulation and disease, illuminating regenerative processes (eg pluripotency induction), mechanisms of reprogramming in nuclear transfer cloning and epigenetic inheritance.