Zebrafish col4a1 loss-of-function models mirror key neurovascular and ocular features of COL4A1/A2 syndrome and enable human variants assessment in vivo

Pathogenic variants in COL4A1 and COL4A2, encoding type IV collagen α1 and α2 chains-core components of all basement membranes-cause a multisystem disorder with variable expressivity. Affected individuals commonly present with cerebral small vessel disease (cSVD), unmanageable intracerebral haemorrhage (ICH), drug-resistant epilepsy, microphthalmia, and congenital cataract. Severe phenotypes are often linked to glycine substitutions that disrupt α1/α2 heterotrimer assembly, though insertions, deletions, and haploinsufficiency seem to also be pathogenic. Limited insight into collagen IV α1 and α2 biology and how specific variants affect their functions-coupled with a lack of rapid in vivo assays for functional variants classification-hampers patient stratification and therapy development. Here, we established and characterized two complementary col4a1 knockdown (KD) models in zebrafish. Taking advantages of their transparency and rapid development we set-up in vivo assays for neurovascular and ocular phenotyping. Both models reproduced key features of human disease, including ventriculomegaly, vascular fragility with spontaneous and trauma-induced ICH, microphthalmia, and cataracts. Notably, expression of human wild-type COL4A1 partially rescued most of the observed defects, while pathogenic glycine-substitution variants failed to do so. These findings validate col4a1 KD in zebrafish as a robust in vivo model of some aspects of COL4A1/A2 syndrome, highlighting a conserved role of collagen IV α1 in neurovascular and ocular development. Our results also support haploinsufficiency as a contributing pathogenic mechanism, alongside dominant-negative effects. This work lays the foundation for the use of zebrafish to support rapid COL4A1 and COL4A2 variants pathogenicity assessment and mechanistic studies, with the potential to accelerate development of targeted therapies.