Delineation of a Human Mendelian Disorder of the DNA Demethylation Machinery: TET3 Deficiency

Abstract

Germline pathogenic variants in chromatin-modifying enzymes are a common cause of pediatric developmental disorders. These enzymes catalyze reactions that regulate epigenetic inheritance via histone post-translational modifications and DNA methylation. Cytosine methylation (5-methylcytosine [5mC]) of DNA is the quintessential epigenetic mark, yet no human Mendelian disorder of DNA demethylation has yet been delineated. Here, we describe in detail a Mendelian disorder caused by the disruption of DNA demethylation. TET3 is a methylcytosine dioxygenase that initiates DNA demethylation during early zygote formation, embryogenesis, and neuronal differentiation and is intolerant to haploinsufficiency in mice and humans. We identify and characterize 11 cases of human TET3 deficiency in eight families with the common phenotypic features of intellectual disability and/or global developmental delay; hypotonia; autistic traits; movement disorders; growth abnormalities; and facial dysmorphism. Mono-allelic frameshift and nonsense variants in TET3 occur throughout the coding region. Mono-allelic and bi-allelic missense variants localize to conserved residues; all but one such variant occur within the catalytic domain, and most display hypomorphic function in an assay of catalytic activity. TET3 deficiency and other Mendelian disorders of the epigenetic machinery show substantial phenotypic overlap, including features of intellectual disability and abnormal growth, underscoring shared disease mechanisms.

Keywords: 5-hydroxymethylcytosine; 5-methylcytosine; DNA methylation; TET; epigenetic; genetic; intellectual disability.

Figure 1

Craniofacial Features and Inheritance in Individuals with TET3 Deficiency (A) Images showing craniofacial characteristics of a subset of affected individuals. (B) Pedigrees illustrating inheritance patterns in each family; specific variants are listed. Abbreviations are as follows: Pat, paternal; Mat, maternal; y, years; and m, months.

Figure 2

TET3 Variants and Predicted Functional Consequences (A) Schematic depiction of TET3 showing domain structure with the catalytic dioxygenase domain in green (aa, amino acids 773–1776) and specific subdomains indicated as follows: the Cys-rich insert in yellow (aa 825–1012) and the double-stranded β helix domain in dark blue (DSBH; aa 1012–1159; aa 1636–1719). The DSBH domain is split in two by a low-complexity insert. The N-terminal CXXC DNA binding domain is shown in light blue (aa 46–102). Specific variants are annotated in orange for recessive alleles and purple for dominant alleles, and underlined variants occur within the last exon. (B) Alignment of missense variants in TET3 across multiple species, including hs, Homo sapiens; pt, Pan troglodytes; cf., Canis familiaris; fc, Felis catus; rn, Rattus norvegicus; and mm, Mus musculus, and among TET enzymes. (C) Crystal structure of TET2 (PDB: 5DEU) bound to DNA; TET3 mutations M2, M3, M4, and M5 are highlighted.

Figure 3

Cells Overexpressing TET3 Variants Have Decreased Amounts of 5hmC (A) Schematic depiction outlining the enzymatic activity assay for measuring TET3 catalytic activity. (B) Representative dot blot showing 5hmC amounts in HEK293 cells overexpressing wild-type or mutant HA-tagged TET3 constructs. (C) Representative western blots showing wild-type and mutant TET3 protein expression in HEK293 cells. β-tubulin was used as a loading control. (D) Quantification of the 5hmC signal relative to TET3 wild-type transfections. The dotted line indicates the wild-type signal. Error bars represent the standard error of the mean. Abbreviations are as follows: WT, wild type; and ΔCAT, catalytically inactive control (p.His1077Tyr/Asp1079Ala, also known as HxD²⁹).