Identification of 5' untranslated region variants in genes involved in neurodevelopmental disorders.

The importance of 5'-untranslated region (5'-UTR) variants in genetic diseases has become increasingly recognized. However, systematic frameworks for interpreting their pathogenic mechanisms remain underdeveloped. We performed genome sequencing (GS) or reanalyzed exome sequencing (ES) data from patients with neurodevelopmental disorders in whom no pathogenic variants had previously been identified, and searched for variants affecting upstream open reading frames (uORFs) in the 5'-UTR using UTRannotator, a tool for annotating 5'-UTR variants. We identified one patient with a maternally inherited single nucleotide duplication upstream of ATRX (c.-138dup), which is predicted to result in the formation of an out-of-frame uORF overlapping the coding sequence (CDS). The patient exhibited the core features of ATRX-related disorders. RNA sequencing of urine-derived cells (UDCs) revealed reduced ATRX expression in the patient. Luciferase reporter assays demonstrated that wild-type and mutant ATRX 5'-UTR sequences conferred significantly increased and decreased luciferase activity compared with the parental pGL3-promoter vector, respectively, suggesting that the c.-138dup variant may disrupt an enhancer-like regulatory element and impair translation. We also identified another patient with a de novo single nucleotide variant upstream of POU3F3 (c.-303C>A), which introduces a novel uORF overlapping the CDS in-frame. This patient showed phenotypes consistent with POU3F3-related disorder. Although immunoblotting using UDCs revealed no elongated POU3F3 proteins, the luciferase assay showed reduced activity with mutant 5'-UTR compared to the wild-type. Our study demonstrates that integrating GS or ES with UTRannotator is useful for identifying candidate 5'-UTR variants; however, the potential impact of predicted non-coding variants still requires careful experimental evaluation.