Rare Nonsense Mediated Decay-Escaping Stop Gained Variants

 

This is a blog post about rare Nonsense Mediated Decay-escaping stop gained variants.

A stop-gained variant, also known as a nonsense mutation, is a type of genetic change where a single nucleotide substitution introduces a premature termination codon (PTC) within the coding sequence of a gene. This leads to early termination of translation, producing a truncated protein. Depending on the location of the PTC and the structure of the affected transcript, the variant may trigger nonsense-mediated decay (NMD) or escape it, making the biological impact isoform-specific and context-dependent.

Nonsense-mediated decay–escaping (NMD-escaping) describes a situation where a transcript containing a premature termination codon (PTC) avoids degradation by the cell’s NMD pathway and remains stable enough to be translated. While NMD typically prevents the production of truncated proteins by degrading transcripts with early stop codons, escape occurs when the PTC is located in the last exon or within ~50–55 nucleotides upstream of the final exon-exon junction. Additional transcript-specific features—such as exon structure, 3′ UTR length, and splicing—can also influence escape. The resulting protein may be benign, harmful, or functionally neutral, depending on domain context and biological role, making isoform-aware interpretation essential.

I was checking out Expert Panel Pathogenic Stop Gained Variants in connection to neurodevelopmental disorders. I killed at those in connection to Rett and Angelman-like disorders. I found one that was quite intriguing and perplexing because overall in-silico predictions showed strong signs of it being benign. 

Pathogenic Rett and Angelman-like Disorders Variant Curation Expert Panel

NM_001323289.2(CDKL5):c.2842C>T (p.Arg948Ter)

rs1555955296

chrX:g.18628716C>T  

Rett and Angelman-like Disorders VCEP

The p.Arg948Ter variant in CDKL5 is absent from gnomAD (PM2_Supporting). The p.Arg948Ter variant in CDKL5 has been reported as a de novo occurrence (biological parentage confirmed) in at least 2 individuals with features of CDKL5-associated disorder (PMID 32366967; internal database) (PS2_Very Strong). The p.Arg948Ter variant in CDKL5 is predicted to cause a premature stop codon that leads to a truncated or absent protein in a gene where loss-of-function is an established mechanism. While loss-of-function variants are commonly observed in affected individuals in this gene, there is a paucity of these variants in this region of the gene to date (PVS1). In summary, the p.Arg948Ter variant in CDKL5 is classified as pathogenic for CDKL5 disorder based on the ACMG/AMP criteria (PM2_Supporting, PS2_Very Strong, PVS1).

Met criteria codes

PS2_Very Strong - The p.Arg948Ter variant in CDKL5 has been reported as a de novo occurrence (biological parentage confirmed) in at least 2 individuals with features of CDKL5-associated disorder (PMID 32366967; internal database)

PM2_Supporting - The p.Arg948Ter variant in CDKL5 is absent from gnomAD

PVS1 - The p.Arg948Ter variant in CDKL5 is predicted to cause a premature stop codon that leads to a truncated or absent protein in a gene where loss-of-function is an established mechanism. While loss-of-function variants are commonly observed in affected individuals in this gene, there is a paucity of these variants in this region of the gene to date (PVS1)

https://erepo.genome.network/evrepo/ui/classification/6f4ef230-8fe8-4a41-9161-47767f6eabf7?version=1.0

The following is from submission in ClinVar

Comment:

The p.Arg948X variant in CDKL5 has been reported in the hemizygous state in a young male with infantile spasms, atonic seizures, developmental delays, and regression (Schoch et al, 2020) and in the heterozygous state in a young female with infantile spasms, epilepsy, and developmental delay (Broad Institute Rare Genomes Project). In both individuals the variant was found to be de novo through trio whole genome sequence analysis. This variant has also been reported by a clinical laboratory in ClinVar (Variation ID 482247) and is absent from large population studies. This variant is present in the NM_001323289.2 transcript, which is confirmed to be expressed in the brain in both humans and mice (Hector RD et al 2016). This nonsense variant leads to a premature termination codon at position 948. This alteration occurs within the last exon and is, therefore, likely to escape nonsense mediated decay (NMD) and result in a truncated protein. However, identification of other de novo truncating variants in this region suggest the region is critical for protein function. In summary, this variant meets criteria to be classified as pathogenic for X-linked dominant early infantile epileptic encephalopathy based on case observations, de novo occurrence, absence from large population studies, and predicted impact on protein. ACMG/AMP Criteria applied: PS2, PVS1_Strong, PM2_Supporting. (less)

Early infantile epileptic encephalopathy with suppression bursts

Affected status: unknown

Allele origin: germline

Laboratory for Molecular Medicine, Mass General Brigham Personalized Medicine

Accession: SCV004848397.1

First in ClinVar: Apr 20, 2024

Last updated: Apr 20, 2024

https://www.ncbi.nlm.nih.gov/clinvar/variation/489299/?oq=((482247[AlleleID]))&m=NM_001323289.2(CDKL5):c.2842C%3ET%20(p.Arg948Ter)

I checked the in-silico predictions at Varsome website.   The overall in-silico predictions show it as being benign. 

BayesDel addAF - Benign Strong

BayesDelnoAF - Benign Moderate. 

CADD - 16.34 - low for a Pathogenic variant (especially for a Stop Gained variant). I use a minumum of 20. 

DANN -  0.80557297479159784 suggesting benign

I checked the variant using Variant Effect Predictor (VEP). It affects two transcripts with the following properties:

• Transcript 1: 18/18 exons

• cDNA position: 3092

• CDS position: 2842

• Protein position: 948

• Transcript 2: 19/19 exons

• cDNA position: 3215

• CDS position: 2965

• Protein position: 989

• Consequence: Stop-gained in the final coding exon of both transcripts

• NMD status: Escapes nonsense-mediated decay

This CDKL5 variant affects two structurally shortened transcripts, one including 18 of the gene’s 24 exons, and the other including 19. Both transcripts are incomplete even before the variant occurs. The stop-gained variant then truncates the remaining tail, compounding the structural loss. As a result, the expressed protein is malformed not only by premature termination, but by originating from an already incomplete blueprint.

The fact that this is a shortened, NMD-escaping transcript amplifies the biological impact. It expresses a malformed protein that wouldn’t exist if the variant triggered NMD—making it a potential driver of pathogenicity.

PVS1_Strong is justified not by generic LoF, but by domain-level disruption in an expressed isoform.

• PS2 (de novo) adds weight, but the biological plausibility hinges on transcript-aware reasoning is now confirmed.

• In-silico benignity is overridden by functional domain loss, which most predictors don’t model.

I have ultra rare NMD-escaping Stop Gained variants in NUAK1 and SGSM2 which are genes that have strong connections to the nervous system.  I am a neurodivergent with Dyslexia, Dyspraxia, ADHD.  I also have Ataxia which is a rare neurological condition that involves coordination problems like Dyspraxia does. 

I went over the variants in my blog post 

The Differences/Comorbidity Between Ataxia & Dyspraxia, My Testing, My Potential Causal Genetic Variants

https://neurodivergence.blogspot.com/2024/11/the-differencescomorbidity-between.html

NM_014840.3(NUAK1):c.1369A>T (p.Lys457Ter)

chr12:g.106067419T>A 

Stop Gained NMD-escaping variant

not found in any population

BayesDel addAF - Pathogenic Strong

BayesDelnoAF - Pathogenic Strong

CADD - 38 indicating highly deleterious

DANN -  0.9961 indicating Damaging

A NMD-escaping stop-gained variant in NUAK1 introduces a premature stop codon in the final coding exon of a transcript, allowing production of a truncated protein. NUAK1 encodes a serine/threonine kinase involved in cell survival and metabolic stress response. Truncation may impair signaling pathways relevant to hypoxia adaptation, tumor progression, or neural resilience.

https://reg.clinicalgenome.org/redmine/projects/registry/genboree_registry/by_caid?caid=CA386365698

I checked the variant using Variant Effect Predictor (VEP). It affects one transcript with the following properties:

• Transcript structure: 7/7 exons

• cDNA position: 1670

• CDS position: 1369

• Protein position: 457

• Consequence: Stop-gained in the final coding exon

• NMD status: Escapes nonsense-mediated decay

This NUAK1 variant affects a structurally shortened transcript that includes only 7 of the gene’s 12 exons. The transcript is incomplete even before the variant occurs. The stop-gained variant then truncates the remaining tail, compounding the structural loss. As a result, the expressed protein is malformed not only by premature termination, but by originating from an already incomplete blueprint.

NM_014853.3(SGSM2):c.3010C>T (p.Arg1004Ter)

chr17:g.2379146C>T

rs867998099 

Stop Gained NMD-Escaping variant

3 out of 1,614,054 (0.0001859%) GnomAD

2 out of 490,784 (0.0004%) AllofUs

2 out of 264,690 (0.0008%) TOPMED

2 out of 7234 (0.03%)  Korea4K

1 out of 1832 (0.05%)  Korea1K

1 out of 16,760 (00.006%) 8.3KJPN 

found in a total of 11 people 

BayesDel addAF - Pathogenic Strong

BayesDelnoAF - Pathogenic Strong

CADD - 40 indicated highly deleterious

DANN -  0.9978 suggesting Damaging

A NMD-escaping stop-gained variant in SGSM2 introduces a premature stop codon in the final coding exon of a transcript, allowing production of a truncated protein. SGSM2 encodes a modulator of small G protein signaling involved in intracellular trafficking and membrane dynamics. Truncation may alter vesicle transport or disrupt isoform-specific functions in neural or epithelial tissues.

https://reg.clinicalgenome.org/redmine/projects/registry/genboree_registry/by_caid?caid=CA286906028

I checked the variant using Variant Effect Predictor (VEP). It affects one transcript with the following properties:

• Transcript structure: 2/2 exons

• cDNA position: 152

• CDS position: 154

• Protein position: 52

• Consequence: Stop-gained in the final coding exon

• NMD status: Escapes nonsense-mediated decay

This SGSM2 variant affects a structurally shortened transcript that includes only 2 of the gene’s 24 exons. The transcript is incomplete even before the variant occurs. The stop-gained variant then truncates the remaining tail, compounding the structural loss. As a result, the expressed protein is malformed not only by premature termination, but by originating from an already incomplete blueprint.