Background

A study of RNU4-2 variant function identified a recessive neurodevelopmental disorder distinct from the previously known ReNU syndrome. The researchers used saturation genome editing (SGE) to build a map of variant function and linked it to patients’ genotypes and clinical features.

RNU4-2 is a non-protein-coding gene that produces U4 small nuclear RNA (snRNA), a component of the major spliceosome. The spliceosome carries out RNA splicing. Previously known variants associated with ReNU syndrome were concentrated in an 18-base region in the middle of the gene, but it was unclear whether variants in other regions could also cause neurodevelopmental disorders.

Key findings

The team assessed 539 RNU4-2 variants in the haploid human cell line HAP1-LIG4-KO. These included every single-nucleotide substitution across the 145-nucleotide transcript and the six nucleotides immediately downstream, as well as single-nucleotide insertions and deletions in a core region and selected control variants. The study did not examine every possible insertion and deletion.

The researchers compared variant frequencies four and 14 days after editing and calculated a functional score for each variant based on its effect on cell growth. This score is not a direct measurement of splicing efficiency. All 18 ReNU-associated variants included in the assessment significantly decreased in frequency in the cell population. In a single-nucleotide variant comparison, the area under the receiver operating characteristic (ROC) curve (AUC) for distinguishing ReNU-associated variants from population variants was 0.93 for the SGE score and 0.65 for the computational prediction tool CADD. These figures describe variant-classification performance within this study; they do not represent diagnostic accuracy in patients.

ReNU syndrome is a dominant disorder that can result from a pathogenic variant in one allele; de novo variants that are not inherited from either parent have been reported. In this study, the team identified 20 patients with neurodevelopmental disorders who carried variants that reduced function on both alleles outside the ReNU core region. Ten were homozygous for the same variant on both alleles, and 10 were compound heterozygous, with different variants on each allele. The group included siblings.

The patients’ variants were located in four regions that interact with other spliceosome components. Citing the clinical analysis in a companion paper, the researchers also reported features that distinguish these cases from ReNU syndrome, including white-matter abnormalities and cerebellar atrophy. However, functional scores alone could not distinguish dominant from recessive variants. The distinction of the new disorder was based on considering variant location, biallelic genotype, and patient phenotype together.

Meaning and outlook

This functional map provides experimental evidence to support clinical interpretation when an RNU4-2 variant is found in a patient. For the ReNU core region, the researchers calibrated the scores to allow strong evidence for pathogenic or benign classification. They could not calibrate variants associated with the recessive disorder outside this region to the same level because there were too few independently established pathogenic variants. The researchers recommend assigning PS3 supporting, a supporting level of evidence for pathogenicity, to significant reductions in function in this region.

The cell-growth-based assay does not directly reveal the specific mechanisms behind splicing changes and cannot capture all effects that may vary by cell type. A normal functional score alone cannot rule out every clinical effect. The findings may inform treatment development, but the study did not demonstrate the efficacy of a personalized treatment.

The methods, results, and limits of their application are described in the Nature paper.