Back

Systematic analysis of loss-of-function variants across MODY genes demonstrates gene- and NMD-specific effects and identifies NMD-escape INS variants as a novel cause of MODY

Laver, T. W.; Sriram, A.; Wakeling, M. N.; Siklar, Z.; Kalyon, O.; Hattersley, A.; Weedon, M. N.; Flanagan, S.; De Franco, E.; Colclough, K.; Patel, K. A.

2025-09-08 endocrinology
10.1101/2025.09.07.25334427 medRxiv
Show abstract

Aims/hypothesisAccurate interpretation of loss-of-function (LOF) variants in MODY genes is essential for diagnosis but remains challenging, particularly for variants predicted to escape nonsense-mediated decay (NMD). We aimed to systematically evaluate the pathogenicity of LOF variants, stratified by NMD-triggering and NMD-escape status, across all known MODY genes. MethodsWe analysed ultra-rare LOF variants (minor allele frequency <1 in 10,000) in 5171 individuals of European ancestry with suspected MODY, compared with 155,501 population-based controls from UK Biobank. LOF variants in ABCC8, GCK, HNF1A, HNF4A, HNF1B, INS, KCNJ11, NEUROD1, PDX1, and RFX6 were classified as NMD-triggering or NMD-escape. We tested for gene-level enrichment in cases versus controls. For novel associations, we performed replication in additional MODY cases, assessed familial co-segregation, and undertook in-silico protein modelling. ResultsLOF variants were significantly enriched in all MODY genes except ABCC8 and KCNJ11. Both NMD-triggering and NMD-escape variants were enriched in GCK, HNF1A, and HNF4A, consistent with haploinsufficiency (all P <10-3). HNF1B and RFX6 showed enrichment only for NMD-triggering variants, while NEUROD1 and PDX1 were enriched for only NMD-escape variants. A novel finding was significant enrichment of only NMD-escape LOF variants in INS (OR = 181, P < 10-5). Including replication, we identified eight families with 17 affected individuals carrying INS variants. These variants co-segregated with diabetes (LOD score = 3), included one de novo case, and were absent from >800,000 population controls. Individuals presented with diabetes at a median age of 19 years, had median BMI of 22.9 kg/m2, were negative for islet autoantibodies, and had low type 1 diabetes genetic risk scores. Compared with INS missense MODY, diagnosis occurred [~]10 years later. Protein modelling suggested that INS NMD-escape variants produce aberrant proinsulin molecules with unpaired B chain cysteines, leading to milder misfolding. Conclusions/interpretationThe pathogenicity of LOF variants in MODY genes depends on gene context and NMD status. Heterozygous NMD-escape LOF variants in INS are a novel cause of MODY. These findings provide systematic gene-level evidence to inform variant interpretation guidelines and improve the accuracy of MODY diagnosis in clinical practice. Research in ContextO_ST_ABSWhat is already known about this subject?C_ST_ABSO_LIInterpretation of loss-of-function (LOF) variants is complex and strongly influenced by predicted NMD status. C_LIO_LICurrent evidence for LOF variants in MODY genes is mainly based on case reports. C_LIO_LISystematic genetic evidence is needed to improve diagnosis and to inform gene-specific variant interpretation guidelines. C_LI What is the key question?O_LIDo NMD-triggering and NMD-escape LOF variants contribute to MODY across all known MODY genes? C_LI What are the new findings?O_LIHeterozygous NMD-escape LOF variants in INS are a novel cause of MODY. C_LIO_LIHeterozygous LOF variants in ABCC8 and KCNJ11 are not enriched in MODY. C_LIO_LIEnrichment of NMD-triggering and NMD-escape LOF variants differs across MODY genes, clarifying underlying mechanisms and supporting gene-specific interpretation. C_LI How might this impact on clinical practice in the foreseeable future?O_LIThese results improve the accuracy of MODY diagnosis by providing robust evidence for variant interpretation and supporting gene-specific clinical guidelines. C_LI

Matching journals

The top 4 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.