Researchers at Chiba University, with colleagues at the Kazusa DNA Research Institute and the National Institutes for Quantum Science and Technology, report a zebrafish carrying the Notch3 p.C680S mutation that develops age-related features of CADASIL, the inherited small vessel disease caused by NOTCH3 mutations. The paper, by Tohgo Kanoh and colleagues and led by Motoyuki Itoh, was published in Acta Neuropathologica Communications on 3 June 2026. This is animal work. According to the paper's abstract, the mutant fish showed cerebrovascular dysfunction, atrophy of the telencephalon, cognitive impairment and deposits of granular osmiophilic material, the granular deposit around small vessels that pathologists treat as a hallmark of CADASIL in people. In Chiba University's own announcement of the work, cerebral blood flow declined progressively from 11 months of age, learning was significantly worse by 14 months and working memory was impaired by 17 months.
The reason the finding is being noticed is the matrix protein, not the fish. Combined transcriptomic and proteomic analysis found significant downregulation of type IV collagen related molecules, including serpinh1b, col4a1 and col4a2, and immunostaining confirmed less type IV collagen around cerebral blood vessels. Type IV collagen is a core component of the basement membrane that sleeves small vessels. Most CADASIL research has centred on the NOTCH3 extracellular domain and its aggregation; the authors write that their results raise the possibility that reduced type IV collagen contributes to disease progression. Structural modelling with AlphaFold 3 placed p.C680S in what the authors call a nonconventional class of NOTCH3 mutation, and they frame the collagen link as most relevant to that class, which is a narrower claim than the field's headline reading of it.
The context is a disease with almost nothing on the shelf. CADASIL is the most common single gene small vessel disease and has no disease modifying treatment. In March 2026, a randomised, double blind, placebo controlled trial in eClinicalMedicine reported that a tocotrienol rich vitamin E complex given to 51 people with genetically confirmed CADASIL for 24 months did not reduce clinical progression, with a Bayesian analysis indicating futility. Existing mouse models, as the Kanoh paper notes, have not reproduced the full range of human pathology, which is the gap a cheap, transparent, fast breeding vertebrate is being offered to fill.
The distance to people remains large. What the fish reproduced were vascular dysfunction, telencephalic atrophy, cognitive impairment and GOM deposits; the abstract does not report lacunar infarcts or the leukoencephalopathy that give CADASIL the second half of its name, and the zebrafish brain has no structure equivalent to human cortical white matter. The collagen result is an association observed inside one mutant line: nothing in the study shows that restoring type IV collagen changes the course of the disease, in fish or in anyone else. No drug, supplement or dietary change follows from it.