RNU2-2-Related Neurodevelopmental Disorders: A Research Summary

At a glance

• Established Variants in RNU2-2, a small non-protein-coding gene on chromosome 11, cause a neurodevelopmental disorder with both dominant and recessive forms. [1][3][4][5]
• Established The dominant form is usually caused by new (de novo) variants at two recurrent positions, n.4G>A and n.35A>G. Mosaicism and rare cases of uncertain parental origin have been reported. [1][2][5]
• Established The recessive form is caused by two variants, one on each copy of the gene (biallelic). In some families one of the two is a new (de novo) variant; how often this happens varies by study. [3][4][5]
• Cohort finding Epilepsy was common in both forms. In Leitão et al.'s research study, 88% of dominant and 83% of recessive participants had epilepsy, and those with epilepsy met criteria for developmental and epileptic encephalopathy (DEE). [5]
• Established Exome sequencing and standard gene testing panels often do not include this non-coding gene. Genome sequencing has been the main route to diagnosis in published studies. Check coverage with your laboratory. [1][4][5]
• Research method A reduced ratio of U2-2 to U2-1 RNA in blood is a promising research biomarker for the recessive form. It is not yet a clinically validated test. [4]
• Hypothesis Explanations for why the two forms differ, and why blood shows few splicing changes in recessive cases, are still being studied. [3][4][5]

1. Gene and inheritance: what is established

The gene

• Established RNU2-2 is a single-copy gene on chromosome 11q12.3 (HGNC). It makes U2-2, a 191-nucleotide small nuclear RNA (snRNA) that forms part of the spliceosome complex, the machinery that removes introns from pre-messenger RNA. [1][5]
• Established It was previously annotated as a pseudogene (RNU2-2P). It was reclassified after studies showed it is expressed at levels similar to the main U2 genes. Older literature and databases may still use the old name. [1][2][5]
• Established The other main U2 gene, RNU2-1, is a multi-copy gene arranged in tandem repeats on chromosome 17q21.31. Copy-number estimates vary by study: Leitão et al. report 6 to more than 80 copies per chromosome [5], and Jackson et al. (2026) describe a tandem repeat array of 5 to 82 identical copies on chromosome 17 [4].
• OMIM: #621304 is the phenotype entry for developmental and epileptic encephalopathy 119 (DEE119), associated with RNU2-2. It is not a separate entry for the recessive form. *621238 is the RNU2-2 gene entry.

Dominant form

• Established Most reported dominant cases carry one of two recurrent variants, n.4G>A or n.35A>G. [1][2][5]
• Cohort finding Jackson et al. (2025) also reported other de novo candidate variants near the 5′ end (n.3C>A, n.7_8insA, n.35A>C). Unlike n.4G>A and n.35A>G, these are not established recurrent pathogenic variants; the authors note that only specific variants in this region may be pathogenic. [2]
• Established Where parents were tested, most of these variants arose de novo (new in the child). [1][2][5]
• Cohort finding Exceptions and uncertainties have been reported:
  • Leitão et al. describe one person with mosaic n.35A>G (variant present in 12% of blood cells, 20% of urine cells and 25% of cheek cells). [5]
  • In the same study, one n.4G>A case was confirmed only as "nonmaternal" (father not tested), and one of three n.35A>G cases was not confirmed de novo. [5]
  • Jackson et al. (2025) identified six other individuals in the 100,000 Genomes Project carrying one of their five discovery variants (n.3C>A, n.4G>A, n.7_8insA, n.35A>C, n.35A>G) that were not, or could not be proven to be, de novo: parental samples were unavailable in four, and there was one maternal and one paternal transmission. [2]
• Practical point: Confirm with parental testing, and consider whether a second variant could be present. De novo status alone does not establish the inheritance pattern.

Recessive form

• Established The recessive form requires a variant on both copies of RNU2-2 (biallelic): either the same variant twice (homozygous) or two different variants (compound heterozygous). [3][4][5]
• Established Recessive variants are spread across the whole gene, unlike the two hotspots of the dominant form. [3][4][5]
• Cohort finding How often one variant is de novo varies by cohort. Leitão et al. describe a de novo variant in trans with an inherited one as "often" seen [5]. In Greene et al. (2026), only 1 of 11 high-confidence (tier 1) pedigrees with both parents tested had a de novo variant, compared with 4 of 9 tier 2 pedigrees [3].
• Cohort finding Heterozygous parents of children with the recessive form were generally unaffected. Leitão et al. noted five exceptions where a carrier parent had epilepsy or intellectual disability. [5]
• Practical point: A single de novo variant does not rule out the recessive form. Look for a second variant on the other copy. Reproductive counseling should be provided by a genetics professional. Recurrence risk should be calculated from the family's actual variants, their phase and parental test results.

2. Clinical features reported in published cohorts

Both forms share developmental delay, intellectual disability and frequent epilepsy. All figures below are Cohort finding and describe the named study only.

Developmental and epileptic encephalopathy (DEE) applies to both forms

• Cohort finding In Leitão et al., epilepsy occurred in 88% of dominant (28 of 32) and 83% of recessive (60 of 72) participants, and "patients with epilepsy met the criteria for developmental and epileptic encephalopathy." [5]
• Jackson et al. (2026) describe the recessive form as a severe DEE, which matches their title. [4] This should not be read as meaning DEE is limited to the recessive form.

Features by study

Hyperventilation

• Cohort finding Greene et al. (2025) recorded hyperventilation in 3 of the 9 initially HPO-coded cases, each coded by a different clinician, and noted it as unusual. [1]
• The papers reviewed here do not establish that hyperventilation triggers seizures.

Brain MRI

• Cohort finding MRI findings vary. Greene et al. (2026) describe scans as often normal early in life, with cerebral or cerebellar atrophy or white matter changes seen later in some. They give no counts or scan timing, so this does not show that an individual's scan will change over time. [3]
• Cohort finding Jackson et al. (2026) reported abnormal MRI in 46% (13 of 28), most often cerebral atrophy (9 of 28). [4] Separately, Leitão et al. (Table 1) reported MRI abnormalities in 9 of 28 monoallelic (32%) and 38 of 63 biallelic (60%) participants with available MRI data. [5]
• The word "progressive" is not used for MRI here, because the papers do not report enough repeat scans to show progression across patients. Greene et al. (2025) did describe the overall dominant phenotype as progressive. [1]

Mortality

• Cohort finding Leitão et al. reported eight deaths (ages 3 to 19 years, median 15), all in participants with biallelic variants. Causes were respiratory, infectious or epilepsy-related complications. [5]
• Cohort finding Greene et al. (2026) reported one death in their recessive dataset, related to prolonged seizures and respiratory compromise. [3]
• These figures come from each study's own dataset. The cohorts may overlap, so the numbers should not be added together.

3. Genetic testing considerations

These are research-informed considerations drawn from the published studies. They are not a formal consensus recommendation. Testing decisions should be made with the reporting laboratory.

Which test

• Established Exome sequencing and standard epilepsy or neurodevelopmental panels often do not include RNU2-2, because it is a non-coding gene. Whether a given test covers it depends on the laboratory's design and analysis. Ask the laboratory directly. [1][4][5]
• Established Genome sequencing (WGS) has been an important method in both research and diagnosis of RNU2-2 disorders. Confirm that the laboratory's pipeline covers and reports this gene, and that the analysis is clinically validated. [1][4][5]
• A person with a previous negative exome or panel may still be worth re-evaluating for RNU2-2, depending on what the earlier test covered.

Technical considerations for laboratories

• Cohort finding Confirm that two variants are on opposite copies of the gene (in trans) using an appropriate validated method. Jackson et al. (2026) manually reviewed reads for 190 individuals with two different rare variants and found 15 phase-switch errors in statistically phased data: 7 truly in trans had been called in cis, and 8 truly in cis had been called in trans. [4]
• Cohort finding The RNU2-1 repeat region is not annotated in the older GRCh37 reference genome. Jackson et al. (2026) reported that RNU2-1 reads may be mis-mapped as low-quality RNU2-2 reads, which lowers the alternate allele fraction of true RNU2-2 variants and increases false-negative calls. This does not mean every GRCh37 result is unreliable, but laboratories may wish to review how this region is handled. [4]

Variant interpretation

• Cohort finding Leitão et al. classified RNU2-2 variants using ACMG criteria in their study. [5] Classification of any individual variant should come from the reporting laboratory.
• Cohort finding Jackson et al. (2026) found that compound heterozygous genotypes with at least one variant in n.1 to n.67 (the constrained 5′ region) were strongly enriched in people with unsolved neurodevelopmental disorders compared with controls. They found no significant difference for homozygous variants in this region. [4]
• Hypothesis Variants predicted to break base pairing in the gene's stem-loop structures may be more likely to be pathogenic in the recessive form. This is a structural prediction, not proof for any single variant. [3][5]
• Because snRNA genes have high mutation rates, de novo status and absence from population databases are weaker evidence of pathogenicity than usual, and variants should be interpreted with caution. [2]

Reproductive implications

• Established When a recessive diagnosis is confirmed, carrier testing and reproductive counseling are relevant for the family. Refer to a genetics professional.

4. Research findings and methods (not routine clinical tests)

The approaches below were used in research studies. None should be treated as standard clinical workup unless a laboratory offers a clinically validated version.

Why blood may show few splicing changes in the recessive form

• Cohort finding In Greene et al. (2026), U2-1 levels rose in a small group of 5 bona fide recessive cases with low U2-2, keeping total U2 at around normal levels in blood. [3] Jackson et al. (2026) found that 7 of 8 individuals with unusually low U2-2 also had unusually high U2-1. [4]
• Hypothesis This compensation by U2-1 may explain why little splicing change is seen in blood. It has not been shown directly, and its role in brain or other relevant tissues is unknown. Leitão et al. note that U2-2 is relatively more highly expressed in brain than in blood. [5]

5. Hypotheses, predictions and open questions

These ideas are useful for research but are not settled. They should not be used as clinical rules.

How the variants may cause disease

• Hypothesis Dominant variants at n.4 and n.35 are predicted to disrupt how U2-2 interacts with other molecules, such as U6 RNA or the intron branch site. [1][5]
• Hypothesis Most recessive variants are predicted to disrupt the internal folding of U2-2 (its stem-loops), making the RNA less stable. Greene et al. (2026) state that most variants were "predicted to disrupt intramolecular interactions." [3]
• Hypothesis Leitão et al. propose a "gradient-of-impact" model linking dominant and recessive effects. [5]
• Hypothesis Jackson et al. (2026) propose a "Goldilocks" model: two strongly damaging variants may not be compatible with life, two mild ones may cause no obvious disease, and the recessive disorder may occur in between. The whole model is a proposal needing further evidence. [4]

Parent of origin of new variants

• Cohort finding For the dominant form, Greene et al. (2025) found maternal origin in all 10 families where it could be determined. [1] In Leitão et al., the 10 n.4G>A and n.35A>G variants in which parental origin could be determined, 8 and 2 respectively, were all maternal. This applies to those two recurrent variants only. [5]
• Cohort finding For de novo variants contributing to the recessive form, Greene et al. (2026) found 5 of 7 were paternal. [3] Across all 26 phaseable de novo variants in Leitão et al. (dominant and recessive combined), 16 were maternal and 10 paternal. [5]
• The numbers are small and the reason is unknown. This is a research observation, not a diagnostic clue.

Why snRNA genes gain new variants often

• Cohort finding Jackson et al. (2025) found an excess of de novo variants in genomic regions that form R-loops (DNA-RNA hybrids). [2]
• Hypothesis R-loops and transcription-related processes are proposed explanations for high mutation rates at snRNA genes, including RNU2-2. This is not an established cause. [2]

How common is it?

Each estimate below uses a different cohort, case definition and comparison. They should not be combined into a single ratio.

References

Five peer-reviewed papers in Nature Genetics (two from 2025, three from 30 March 2026) plus one peer-reviewed paper in Epilepsia (2026, previously a medRxiv preprint). Two papers have publisher corrections, linked in their entries.

1. Greene D, et al. Mutations in the small nuclear RNA gene RNU2-2 cause a severe neurodevelopmental disorder with prominent epilepsy. Nature Genetics 57, 1367–1373 (2025). First description of the dominant form in 25 individuals; reports recurrent de novo n.4G>A and n.35A>G variants and maternal origin in phaseable families.
2. Jackson A, et al. Analysis of R-loop forming regions identifies RNU2-2 and RNU5B-1 as neurodevelopmental disorder genes. Nature Genetics 57, 1362–1366 (2025). Independent identification of dominant RNU2-2 variants through an analysis of de novo variants in R-loop regions. See also the publisher correction (24 June 2025): https://doi.org/10.1038/s41588-025-02274-3.
3. Greene D, et al. Biallelic variants in RNU2-2 cause the most prevalent known recessive neurodevelopmental disorder. Nature Genetics 58, 774–781 (2026). DOI: 10.1038/s41588-026-02539-5. Statistical genetic evidence for the recessive form; reports reduced U2-2 expression in blood and paternal origin for most phaseable de novo alleles in recessive cases. Uses the name ReNU2 syndrome. See also the publisher correction (23 April 2026): https://doi.org/10.1038/s41588-026-02605-y.
4. Jackson A, et al. Biallelic variants in RNU2-2 cause a remarkably frequent developmental and epileptic encephalopathy. Nature Genetics 58, 798–809 (2026). DOI: 10.1038/s41588-026-02551-9. Reports 84 individuals with neurodevelopmental disorders from 67 families with candidate biallelic variants, across discovery and validation cohorts; describes the U2-2:U2-1 ratio as a potential biomarker; documents phasing and GRCh37 alignment issues; proposes the "Goldilocks" model.
5. Leitão E, et al. Systematic analysis of snRNA genes reveals frequent RNU2-2 variants in dominant and recessive developmental and epileptic encephalopathies. Nature Genetics 58, 782–797 (2026). DOI: 10.1038/s41588-026-02547-5. 141 individuals from 122 families, largely from the French PFMG cohort; compares dominant and recessive features; reports a mosaic case, eight deaths in biallelic participants, and exploratory methylation data.
6. Henry OJ, Pekkola Pacheco N, Duba I, et al. Phenotypic and transcriptomic characterization of biallelic RNU2-2 developmental and epileptic encephalopathy. Epilepsia, published online 1 September 2026. DOI 10.1002/epi.70473; PMID 42678714. Peer reviewed; previously a medRxiv preprint. DOI 10.64898/2026.02.19.26345867. 14 individuals from 9 families; reports aberrant splicing in fibroblasts that was not detectable in blood, and proposes fibroblast RNA sequencing for future variant validation.