Pericyte Loss and Inflamed Blood Vessel States Revealed in Brain Small Vessel Disease

Cerebral small vessel disease, the quiet saboteur of the brain’s tiniest blood vessels, is a leading cause of stroke and the principal vascular contributor to cognitive impairment and dementia. Yet despite its enormous clinical footprint, the cellular mechanisms that drive microvascular dysfunction in human disease have remained stubbornly unclear. A new study published in Acta Neuropathologica has now mapped the human cortical microvasculature at single-vessel resolution, revealing that the two most celebrated suspects in blood–brain barrier breakdown — pericyte loss and endothelial activation — do not march in lockstep, but instead define distinct, partially dissociated vascular states.

The research team, led by investigators at the University of Edinburgh and the UK Dementia Research Institute, analysed post-mortem brain tissue from 20 cases spanning the full spectrum of cerebral small vessel disease severity, from healthy controls to severe disease. The tissue, drawn from the Lothian Birth Cohort, the LINCHPIN study, and the Edinburgh Sudden Death Brain Bank, came from Brodmann area 4, the primary motor cortex, an understudied region in this disease. Using multiplex immunohistochemistry combined with spectral unmixing and automated image analysis, the researchers dissected 11,409 individual microvascular fragments, tracking three molecular markers simultaneously: PECAM-1 to identify endothelial cells, PDGFRβ to label pericytes, and VCAM-1 as a readout of endothelial activation.

The technical achievement underlying the study is considerable. Aged human brain tissue is notorious for autofluorescence, particularly from lipofuscin pigment, which confounds conventional immunofluorescence. The team sidestepped this problem with chromogenic multiplex staining imaged across 20 wavelengths on a liquid–crystal tuneable band-pass filter system, then computationally unmixed the overlapping chromogen signals into separate pseudo-fluorescent channels. A MATLAB-based segmentation pipeline called Morph-O-Matic then extracted each vessel’s length, diameter, area, and marker signal intensities, with extreme values filtered out to remove segmentation artefacts. The result was a quantitative atlas of capillaries averaging 9.1 micrometres in diameter, the most abundant vascular bed in brain tissue.

The first surprise came from vascular architecture. In healthy controls, white matter vessels were 51 percent less dense and 14 percent wider than their grey matter counterparts, underscoring the inherent vulnerability of white matter. As disease severity increased, vessel diameter remained essentially stable — but microvascular density in white matter doubled in severe cases compared with controls, and correlated strongly with age at death. This finding runs counter to previous reports of age-related capillary rarefaction. The authors suggest the increase may represent compensatory remodelling in response to hypoxia or hypoperfusion in deeper brain regions, noting that their sampling focused on superficial cortical layers and adjacent white matter, where cSVD-related remodelling is typically less pronounced.

The molecular story was more dramatic. PDGFRβ signal, a proxy for pericyte coverage, plummeted with increasing disease severity — dropping 94 percent in grey matter and 87 percent in white matter in the most severely affected cases relative to controls. The decline correlated robustly with both disease severity and age, and white matter pericyte signal was already 52 percent lower than grey matter even in healthy brains. In parallel, the endothelial markers climbed: PECAM-1 rose by 41 to 67 percent and VCAM-1 more than doubled in severe disease, hallmarks of a shift toward a pro-inflammatory, pro-thrombotic endothelial phenotype that promotes leukocyte adhesion and blood–brain barrier disruption.

But here lies the study’s central twist: when the researchers looked for a direct relationship between pericyte loss and endothelial activation, they found none. There was no clear correlation between falling PDGFRβ and rising VCAM-1 or PECAM-1. Instead, the data displayed striking heterogeneity, particularly in the intermediate severity groups, where some vessels retained robust pericyte coverage while others showed marked depletion. This heterogeneity, invisible to the bulk and region-averaged analyses that have dominated earlier neuropathological studies, may help explain why the literature has long been split between reports of pericyte loss and reports of preserved pericyte coverage in dementia.

To bring order to this complexity, the team applied unsupervised Gaussian mixture clustering to the three marker signals. The analysis resolved three distinct vascular states. Cluster 1 combined high pericyte coverage with low endothelial activation and contained all the healthy controls, plus a few mild and moderate cases. Cluster 2 showed reduced pericyte signal without endothelial activation, grouping intermediate cases. Cluster 3 paired pericyte loss with the highest levels of endothelial activation and captured all severe cases plus one moderate case. Notably, the marker-based clusters separated the cases more cleanly than the neuropathological diagnosis alone, as judged by PERMANOVA statistics, yet the authors are careful to stress that the clusters are descriptive molecular phenotypes, not proven sequential stages of disease.

Complementing the post-mortem work, the researchers examined circulating PDGF-BB and VCAM-1 in a previously published cohort of lacunar stroke patients. Patients whose blood showed high VCAM-1 with low PDGF-BB were significantly older than those in other biomarker quadrants, and their white matter damage scores, while highest, were no longer significant after adjusting for age. These exploratory data provide clinical context linking vascular marker profiles to ageing, but the authors emphasise they do not establish any temporal relationship between pericyte and endothelial changes.

The study also caught microglia in the act of change. In severe disease, Iba1-positive microglial profiles appeared more amoeboid, with segmented particle counts 46 percent lower and median particle areas up to 60 percent larger in grey matter compared with milder cases — a pattern consistent with reduced ramification and microglial activation. This supports the idea that advanced small vessel disease engages the wider neurogliovascular unit, though the thin 4-micrometre sections used precluded detailed reconstruction of microglial morphology or direct assessment of microglia–vessel interactions.

The authors are candid about the limitations. The cohort of 20 cases is small, the controls were not age-matched — a confound that partially entangles disease with ageing — and only five participants were female. Two severe cases also carried mild cerebral amyloid angiopathy, a potential confounder that could not be assessed separately. The anatomical scope was restricted to one cortical region and superficial white matter, whereas small vessel disease pathology is often most pronounced in deep white matter. And because the measurements are post-mortem marker expression, they do not directly measure blood–brain barrier permeability, tight junction integrity, or vascular physiology. Larger, longitudinal studies incorporating leakage markers and functional systems will be needed to determine whether the three vascular states are mechanistically connected or occur in a defined sequence.

Even with those caveats, the framework emerging from this work is a significant step forward. By showing that pericyte alterations and endothelial activation are partially dissociated processes in human disease, the study challenges the simple narrative in which pericyte loss automatically triggers endothelial inflammation. Instead, the brain’s smallest vessels appear to occupy distinct molecular states — intact and quiescent, pericyte-depleted but calm, or pericyte-depleted and inflamed — that may respond differently to the disease-modifying therapies that cerebral small vessel disease so urgently awaits. For a condition with no approved treatments that alter its course, defining these microvascular phenotypes is the essential first step toward targeting them.

Subject of Research: Pericyte loss and endothelial activation states in the human cerebral microvasculature in small vessel disease

Article Title: Pericyte loss and endothelial activation define distinct vascular states in cerebral small vessel disease

Article References: Pericyte loss and endothelial activation define distinct vascular states in cerebral small vessel disease. (n.d.). https://doi.org/10.1007/s00401-026-03102-6

Image Credits: AI Generated

DOI: 10.1007/s00401-026-03102-6

Keywords: cerebral small vessel disease, pericytes, endothelial activation, blood–brain barrier, PDGFRβ, VCAM-1, PECAM-1, microvasculature, white matter, dementia, multiplex immunohistochemistry, ageing