CUB and Sushi domains-containing CSMD proteins have recognized roles in neuron development. Prior genetic studies of the CSMD paralogs (CSMD1–3) linked these genes to cognition, attention-deficit hyperactivity disorder (ADHD), schizophrenia and brain structure by GWAS, and recent reports associated rare biallelic variants in CSMD paralogs with epilepsy and other neurodevelopmental phenotypes. Building on this context, the current study identifies multiple rare CSMD2 variants in individuals with neurodevelopmental disorders.
The authors report nine individuals from eight families who carry biallelic missense variants in CSMD2 identified in developmental delay/intellectual disability cohorts. They also describe two families with monoallelic CSMD2 frameshift variants, consistent with gnomAD metrics that support CSMD2 haploinsufficiency. Additionally, the manuscript references a previously described de novo truncating CSMD2 variant in an ASD proband and adds two newly reported individuals with de novo CSMD2 missense variants. The study further compares these findings to twelve de novo CSMD2 cases identified in published trio analyses.
Affected individuals in the reported cohorts presented with a range of neurodevelopmental phenotypes. Clinical features included developmental delay or intellectual disability, autism spectrum disorder (ASD), ADHD, and structural brain anomalies. The paper does not provide exhaustive individual-level phenotype tables in the summary text; details were reported in the source cohort descriptions.
The evidence presented supports multiple genetic mechanisms affecting CSMD2. Biallelic missense variants found in several affected individuals are consistent with an autosomal recessive route to disease. Separately, monoallelic frameshift variants segregating in two families and an existing report of a de novo truncating variant in an ASD proband are compatible with haploinsufficiency and dominant effects. The detection of de novo missense variants—both those newly reported here and those from prior trio studies—further underlines that CSMD2 variation can contribute to both autosomal recessive and dominant forms of neurodevelopmental disorder.
To assess broader population-level associations, the authors analyzed UK Biobank data. Volunteers carrying at least one rare CSMD2 variant showed nominally significant enrichment for ADHD, ASD and neuroticism. The source describes these findings as nominally significant; the summary does not report effect sizes, p-values or full statistical models, which were not provided in the abstract text.
In silico structural modelling and cellular immunofluorescence assays were used to explore consequences of CSMD2 missense variants. Mutated residues from both biallelic and de novo cases clustered in two regions of the protein. Functional assays showed that the mutant proteins exhibited impaired membrane localization and increased cytoplasmic aggregation compared with the wild-type protein. These cellular findings suggest altered subcellular trafficking or stability as a potential molecular mechanism.
The authors used zebrafish crispant larvae targeting the orthologous csmd2 to probe in vivo effects. csmd2-deficient larvae exhibited multiple neurodevelopmental and behavioral phenotypes: microcephaly, a smaller cerebellum, and reduced Purkinje cell size. Behavioral and connectivity phenotypes included decreased swimming velocity, diminished connectivity between the optic tecta, and abnormal peripheral neuronal branching. These phenotypes provide complementary functional evidence that loss or disruption of csmd2 impairs neurodevelopmental processes.
Taken together, human genetic data, cellular assays and zebrafish functional experiments support a role for rare, likely pathogenic CSMD2 variants in neurodevelopmental disorders. The evidence indicates that CSMD2 can be involved via biallelic missense changes (recessive), as well as via monoallelic loss-of-function or de novo missense/truncating variants (dominant/haploinsufficiency). The findings add to prior work implicating CSMD paralogs in neurogenesis and synapse formation and strengthen CSMD2 as a candidate gene for diverse neurodevelopmental phenotypes including ASD and ADHD.
The study declares no competing interests and reports that the research was approved by relevant institutional review boards (for example, CER-VD 2021-01400). The authors note that generated data are available upon reasonable request. Further details, including full phenotype descriptions, variant-level data, and statistical results, were not reported in the abstract and would require consulting the full manuscript or supplementary materials.