DISTINCT ULTRASTRUCTURAL CHARACTERISTICS OF CAPILLARY BASAL LAMINA IN HUMAN AND MOUSE BRAINS: RELEVANCE TO MURINE MODELS OF NEUROVASCULAR DISEASE

Abstract

Recent studies suggest that capillary dysfunction may be a source of neurodegeneration, stroke and cognitive decline (J Cereb Blood Flow Metab, 2016, 36:302; Neurobiol Aging, 2017, 50:107). Most neurovascular studies used mice to model human diseases. The morphologic changes in the capillary wall are less-well characterized in humans, especially in early or presymptomatic phases and in early onset diseases. Here we describe major differences in ultrastructure of capillary basal lamina (CBL) in humans (normal and diseased, <50 years) compared with mouse (young to middle age). Human autopsy brains without or with neurodegeneration, including Parkinson's disease, frontotemporal degeneration and amyotrophic lateral sclerosis, and mouse models of tauopathy were evaluated. Brain tissues were fixed in formaldehyde and/or glutaraldehyde and processed for electron microscopy (EM) or post-embedding immunogold EM. Antibodies to types I, III and IV collagen and to fibronectin were used. Mouse CBL is a thin and uniformly amorphous structure that is labeled with anti-type IV collagen in both normal mice and models of neurodegenerative disease at all ages. In contrast, the CBL in normal and diseased human brains has the amorphous BL, but also banded collagens. In addition, the BL frequently shows segmental splitting between endothelial and astrocytic sides of the BL. The split creates an expanded space that contains banded fibrils of type I and III collagen, but not type IV collagen. The expansion and splitting of BL increase with age and in diseased brains. The CBL in diseased human brains often extends into the pericapillary space. Anti-fibronectin co-localizes with collagens. These ultrastructural features are rarely observed in mice, but are detected in the youngest human brains studied. Human CBL has age- and disease-related alternations including split BL and accumulation of banded type I and III collagen fibrils associated with fibronectin. This BL “fibrosis” is rarely observed in mice brains. If these structural BL alterations are associated with capillary dysfunction, the lack of these changes in mice warrants caution about use of mouse models for neurovascular studies.