Search

BackgroundCerebral cavernous malformations (CCMs) are vascular malformations mostly located within the central nervous system. Most deleterious variants are loss of function mutations in one of the three CCMgenes. These genes code for proteins that form a ternary cytosolic complex with CCM2 as a hub. Very few CCM2missense variants have been shown to be deleterious by modifying the ternary CCM complex stability.ObjectivesTo investigate the causality of novel missense CCM2variants detected in patients with CCM.MethodsThe three CCM genes were screened in 984 patients referred for CCMmolecular screening. Interaction between CCM1 and CCM2 proteins was tested using co-immunoprecipitation experiments for the CCM2missense variants located in the phosphotyrosine binding (PTB) domain.Results11 distinct CCM2rare missense variants were found. Six variants predicted to be damaging were located in the PTB domain, four of them were novel. When co-transfected with CCM1 in HEK293T cells, a loss of interaction between CCM1 and CCM2 was observed for all six variants.ConclusionWe showed, using co-immunoprecipitation experiments, that CCM2 missense variants located in the PTB domain were actually damaging by preventing the normal interaction between CCM1 and CCM2. These data are important for diagnosis and genetic counselling, which are challenging in patients harbouring such variants.

Heterozygous COL2A1 variants cause a wide spectrum of skeletal dysplasia termed type II collagenopathies. We assessed the impact of this gene in our French series. A decision tree was applied to select 136 probands (71 Stickler cases, 21 Spondyloepiphyseal dysplasia congenita cases, 11 Kniest dysplasia cases, and 34 other dysplasia cases) before molecular diagnosis by Sanger sequencing. We identified 66 different variants among the 71 positive patients. Among those patients, 18 belonged to multiplex families and 53 were sporadic. Most variants (38/44, 86%) were located in the triple helical domain of the collagen chain and glycine substitutions were mainly observed in severe phenotypes, whereas arginine to cysteine changes were more often encountered in moderate phenotypes. This series of skeletal dysplasia is one of the largest reported so far, adding 44 novel variants (15%) to published data. We have confirmed that about half of our Stickler patients (46%) carried a COL2A1 variant, and that the molecular spectrum was different across the phenotypes. To further address the question of genotype–phenotype correlation, we plan to screen our patients for other candidate genes using a targeted next-generation sequencing approach.