Localized conditional induction of brain arteriovenous malformations in a mouse model of hereditary hemorrhagic telangiectasia

Abstract

Abstract Background Longitudinal mouse models of brain arteriovenous malformations (AVMs) are crucial for developing novel therapeutics and pathobiological mechanism discovery underlying brain AVM progression and rupture. The sustainability of existing mouse models is limited by ubiquitous Cre activation, which is associated with lethal hemorrhages resulting from AVM formation in visceral organs. To overcome this condition, we developed a novel experimental mouse model of hereditary hemorrhagic telangiectasia (HHT) with CreER-mediated specific, localized induction of brain AVMs. Methods Hydroxytamoxifen (4-OHT) was stereotactically delivered into the striatum, parietal cortex, or cerebellum of R26CreER; Alk12f/2f (Alk1-iKO) littermates. Mice were evaluated for vascular malformations with latex dye perfusion and 3D time-of-flight magnetic resonance angiography (MRA). Immunofluorescence and Prussian blue staining were performed for vascular lesion characterization. Results Our model produced two types of brain vascular malformations, including nidal AVMs (88%, 38/43) and arteriovenous fistulas (12%, 5/43), with an overall frequency of 73% (43/59). By performing stereotaxic injection of 4-OHT targeting different brain regions, Alk1-iKO mice developed vascular malformations in the striatum (73%, 22/30), in the parietal cortex (76%, 13/17), and in the cerebellum (67%, 8/12). Identical application of the stereotaxic injection protocol in reporter mice confirmed localized Cre activity near the injection site. The 4-week mortality was 3% (2/61). Seven mice were studied longitudinally for a mean (SD; range) duration of 7.2 (3; 2.3−9.5) months and demonstrated nidal stability on sequential MRA. The brain AVMs displayed microhemorrhages and diffuse immune cell invasion. Conclusions We present the first HHT mouse model of brain AVMs that produces localized AVMs in the brain. The mouse lesions closely resemble the human lesions for complex nidal angioarchitecture, arteriovenous shunts, microhemorrhages, and inflammation. The model’s longitudinal robustness is a powerful discovery resource to advance our pathomechanistic understanding of brain AVMs and identify novel therapeutic targets.