FBN2 Variants and Spinal CSF Leaks
In one selected type 1b leak cohort, rare functional FBN2 variants were statistically enriched and supported by cell and mouse experiments, making fibrillin-2 a plausible contributor to susceptibility rather than a complete explanation or established basis for care.
> Research explainer: This briefing examines verified primary research published 54 days before the briefing date. It is not a same-day research update and does not provide medical advice.
Evidence
This research explainer concerns a study published on 2026-07-01. Investigators retrospectively identified 42 unrelated people with lateral, type 1b spontaneous spinal cerebrospinal fluid (CSF) leaks who were seen at one hospital between 2006 and 2019, then performed whole-exome sequencing. The cohort included 35 women (83%), seven men (17%), 38 White participants (90%), two Black participants (5%), and two Hispanic participants (5%). [pmid:42309087]
Nine of the 42 participants (21%) carried rare functional variants in FBN2, the gene encoding fibrillin-2. The investigators compared this finding with three independent sequencing cohorts. Rare functional FBN2 variants occurred in 177 of 2,244 unaffected adults (8%) in the Mendel discovery cohort, producing an odds ratio of 3.18 (95% CI 1.50–6.76; p=0.041). They occurred in 51 of 714 people (7%) in a Belgian whole-exome cohort, with an odds ratio of 3.55 (1.61–7.81; p=0.004), and in 45 of 913 people (5%) in a Belgian thoracic aortic aneurysm and dissection cohort, with an odds ratio of 5.26 (2.38–11.66; p=0.0003). [pmid:42309087]
The variants found in participants did not appear randomly distributed across fibrillin-2: they were enriched in transforming growth factor beta binding protein-like (TB) domains. To test a possible functional consequence, the team examined wild-type and mutant fibrillin-2 fragments for integrin-mediated binding to human dural fibroblasts. Two of three tested variants reduced fragment adhesion in vitro. [pmid:42309087]
The study also used CRISPR-Cas9 to generate three mouse models carrying equivalents of FBN2 variants observed in people with type 1b leaks: Fbn2 A1052T/+, Fbn2 D1581V/+, and Fbn2 M2387T/+. Under controlled leak induction by intrathecal infusion testing, mice carrying each of these variants showed predisposition to dural rupture. An established Marfan syndrome mouse model, Fbn1 C1039G/+, showed increased meningeal compliance in the reported comparison. [pmid:42309087]
Analysis — Genetic susceptibility signal
The convergence of enrichment, domain localization, fibroblast-binding experiments, and engineered mouse experiments makes FBN2 a biologically coherent candidate for susceptibility to type 1b spontaneous spinal CSF leaks. The human comparison supplies the association: rare functional variants were more common in the selected leak cohort than in each of the three control cohorts. The cellular experiment then links two tested variants to reduced adhesion of fibrillin-2 fragments to dural fibroblasts, while the mouse work links variant-equivalent genotypes to dural rupture when a leak is induced. Together, these results support the authors’ interpretation that disruption of adhesion to extracellular-matrix proteins may participate in pathophysiology. [pmid:42309087]
That interpretation should remain narrower than a universal genetic account of spontaneous spinal CSF leaks. The study specifically examined lateral type 1b leaks, not every imaging-defined leak type. It identified rare functional FBN2 variants in nine participants, so its result neither shows that all affected people carry such variants nor establishes their contribution to other leak categories. The TB-domain concentration is notable because it gives the association a protein-structure context, but it does not itself establish the effect of every variant. [pmid:42309087]
The authors state that their findings support integrating FBN2 genetic testing into clinical practice. In research terms, that is a study implication grounded in this selected cohort and its functional follow-up, not evidence here of improved diagnosis, treatment, outcomes, or testing performance. The mouse models and quantitative dural-integrity assay provide tools for future investigation; they do not demonstrate a pharmacological therapy. [pmid:42309087]
Limitations
The human evidence comes from a retrospective, single-hospital cohort of 42 people. Its demographic composition was predominantly women and White participants, which limits how confidently the observed frequencies can be generalized beyond this sample. [pmid:42309087]
The study combines association with supportive functional evidence, but association does not show that FBN2 explains all or most type 1b leaks. Only two of three variants tested in the fibroblast assay reduced adhesion, so those results cannot be assumed for every identified FBN2 variant. [pmid:42309087]
Finally, the mouse outcome was predisposition to dural rupture during controlled leak induction. It is not proof of spontaneous human CSF leaks, and it does not establish clinical benefit from genetic testing or any intervention. Claims about connective-tissue integrity, diagnostic strategy, and possible therapies should therefore be read as bounded mechanistic interpretations or study implications. [pmid:42309087]