How dysregulated collagen IV homeostasis causes cerebral microbleeds

Millions of older adults have tiny brain hemorrhages called cerebral microbleeds, which are strongly associated with dementia, cognitive decline, and stroke. However, their precise molecular mechanisms have remained unclear, largely because of the lack of suitable animal models that isolate this condition from other confounding pathologies. A groundbreaking study published today in the journal Brain helps fill this critical gap.

Researchers at Ajou University School of Medicine utilized CRISPR/Cas9 gene editing to selectively delete Col4a1, a gene encoding a vital structural protein in brain blood vessel walls, specifically in adult mice. A single intravenous injection of an engineered virus, AAV-BR1, delivered the gene-editing machinery directly to brain microvascular endothelial cells. Within three months, the mice developed dozens of microbleeds distributed across the cortex and hippocampus, remarkably mirroring the location and size seen on MRIs of elderly patients. This burden increased progressively over six months and could be precisely scaled by adjusting the initial viral dose.

Unlike existing models that conflate microbleeds with amyloid plaques or ischemic injury, this new platform generates a pure cerebral microbleed phenotype, allowing researchers to study its effects in isolation. Electron microscopy confirmed that the affected vessels had dramatically thinned basement membranes. As these microbleeds accumulated over time, the mice exhibited progressive memory impairment and motor deficits, closely mirroring clinical patient observations.

The research team also uncovered a distinctive neuroinflammatory pattern driving this decline. They observed that reactive astrocytes spread widely and diffusely beyond the individual lesion sites, while microglial activation remained strictly localized. This widespread astrocytic response suggests a novel mechanism where multiple small, scattered lesions collectively disrupt broader neural networks and impair overall brain function.

To bridge these experimental findings with human pathology, the researchers analyzed magnetic resonance imaging and genomic data from 836 participants in the BICWALZS biobank, a chronic cerebrovascular disease human tissue bank. They discovered that genetic variants in TIMP2, which regulates the enzyme that breaks down collagen IV, were significantly associated with microbleed susceptibility, increasing individual risk by 1.50 to 1.96 times. These human genomic data perfectly alignhe mouse model, implicating dysregulated collagen IV homeostasis as a conserved cross-species mechanism underlying sporadic cerebral microbleeds.

This is the first model to generate a purely cerebral microbleed phenotype through a targeted molecular intervention in the adult brain, said Byung Gon Kim, MD, PhD, professor of brain science and neurology at Ajou University and co-corresponding author of the study. The ability to precisely modulate microbleed burden offers an unprecedented platform for testing future therapies aimed at halting microbleed progression and preserving cognitive function in aging populations.