A Trim28-ERV axis drives phenotypic variation in obesity

Modified Project Summary/Abstract Section Estimates from twin and inbred animal models suggest that up to 50% of metabolic disease heterogeneity is due to something other than genetics or the environment. We previously discovered that Trim28 is a master regulator of non-genetic, non-environmental phenotypic heterogeneity; demonstrated that Trim28 buffers against a unique (bistable) epigenetic obesity in isogenic Trim28+/D9 mice; and identified a comparable TRIM28-associated mRNA expression signature in human obesity. Individuals at the obese, epigenetically “ON” end of the spectrum are characterized by dysregulation of Trim28-sensitive loci, including endogenous retroviruses (ERVs). Trim28-normally interacts with KZFPs to keep ERVs in a stable, fully silenced state throughout development. However, a subset of ERVs partially escape this silencing and become variably methylated (VM-ERVs), which translates into variable levels of repression/activation, subsequent variability in neighboring gene expression, and ultimately phenotypic variation. Our preliminary data indicate that VM-ERVs have sequence content allowing them to be bound by Zinc-Finger-CxxC (ZF-CxxC) domain containing proteins, thereby protecting these loci from methylation. We hypothesize that activation vs silencing at VM-ERV loci is controlled by a competition between KZFP/Trim28 and ZF-CxxC binding, and that TRIM28-sensitive VM-ERV expression is a significant driver of unexplained phenotypic variation and metabolic disease heterogeneity. We will test these hypotheses through three independent Aims. In Aim 1, we will define the epigenomic characteristics of ERVs in hepatocytes and adipocytes to identify the epigenomic and genomic features that define VM-ERVs. In Aim 2, we will profile binding of specific ZF-CxxC proteins (CFP1 and TET1) to prove that these proteins gain a competitive binding advantage as Trim28 levels are reduced and demonstrate that ZF-CxxC – KZFP/Trim28 competition triggers epigenetic variability at VM-ERVs. In Aim 3, we will determine if VM-ERV activation is coupled to metabolic disease heterogeneity. We expect this project will identify Trim28/ZF-CxxC competition as a novel mechanism governing the variable silencing at ERVs and provide the research community with a comprehensive list of target proteins and pathways to study and begin understanding how VM-ERV dysregulation triggers and/or influences complex disease etiology.