Role of pericytes and endothelial cells in small vessel disease: from animal models to human disease

Abstract

Cerebral small vessel disease (SVD) is a prevalent disorder of the brain’s perforating microvessels and the underlying cause of a quarter of all ischaemic strokes and most haemorrhagic strokes and the commonest cause of vascular dementia. Despite its increasing incidence and importance in the public health domain, the underlying pathophysiology of SVD remains poorly understood and there are currently no available treatments beyond secondary strokeprevention.

An abundance of recently published evidence suggests the involvement of capillary-enwrapping pericyte and brain microvessel endothelial cell dysfunction in SVD pathogenesis, although significant controversy remains, and translational research is required to reconcile what we know about disease mechanisms on the basis of preclinical studies and human studies.

This thesis begins with an introduction to brain pericytes, endothelial cells, and an in-depth exploration into how they potentially contribute to SVD mechanisms. Subsequently, through a scoping review featuring 49 eligible publications, I provide a holistic overview of the existing literature on the molecular mechanisms underlying endothelial dysfunction in SVD, including discussions around gaps in literature and potential sources of inconsistencies amongst study results.

There are discrepancies in the literature regarding pericyte loss and endothelial activation in ageing and in the context of SVD, and the nature of the relationship between endothelial and pericyte function remains to be elucidated. Seeking to address some of these inconsistencies, I quantified pericyte coverage (%) and marker of endothelial activation, vascular cell adhesion molecule-1 (VCAM-1) expression across 9 brain regions in young (n=8, 3 months-old) and ageing (n=8, 12 months-old) mice. We found a 7.5% reduction in global pericyte coverage and more than a two-fold increase in VCAM-1 coverage in aged animals compared to their younger counterparts, after adjusting for sex and brain region using linear mixed effects models. We also demonstrated a potential inverse relationship between VCAM-1 capillary coverage (%) and pericyte signal intensity, wherein a 10% decrease in pericyte signal was associated with a 0.20% increase in VCAM-1 capillary coverage.

To corroborate the murine findings within a human SVD context, I used postmortem human brain tissue from 5 healthy control cases (mean age: 33 years (SD: 9.3)), as well as 3 mild (mean age: 66 (SD: 20)), 3 moderate (mean age: 80 (SD:6.2)), and 5 severe (mean age: 83 (SD:8.8)) SVD cases. We report an estimated 83-86% reduction in pericyte coverage and a 2-to-3-fold increase in VCAM-1 amongst those with moderate-to-severe SVD compared to healthy controls in linear mixed effects models after adjusting for potential covariates, including vascular risk factors. We also demonstrate a significant inverse relationship between VCAM-1 and pericyte signal intensity, wherein every 10% increase in pericyte signal was associated with a 0.80% decrease in VCAM-1 signal.

I then aimed to determine whether these dysfunctions could be detected in vivo and if they were associated with distinct manifestations of SVD burden and progression. In these exploratory analyses, I examined the relationship between a panel of biomarkers of vascular, endothelial, and pericyte function and SVD structural and vascular function neuroimaging measures and cognitive outcomes in 181 patients with mild, non-disabling ischaemic stroke, enriched for SVD features (mean age: 65 years (SD: 11), 31% female). Outcomes were measured at baseline and throughout 1 year and included several MRI markers of SVD, blood-brain barrier (BBB) permeability, cerebrovascular reactivity (CVR), and cognition. After adjusting for relevant covariates, we identified a novel association between higher circulating levels of Vascular Endothelial Growth Factor (VEGF) and incident lesions visible on diffusion-weighted imaging appearing during the first year post-stroke. We also detected associations between higher Intercellular Adhesion Molecule-1 (ICAM-1) and lower CVR, as well as elevated concentrations of P-Selectin and a higher risk of cognitive impairment at 1-year follow-up.

Finally, I report results from a longitudinal observational cohort study, the Mild Stroke Study 2, where we followed-up a cohort of patients with mild ischaemic stroke, enriched for SVD features, 8.5-years post-stroke (n=243, mean age: 67 years (SD:12), 42% female). During the 7-9 years of follow-up, despite the lacunar strokes being slightly younger at baseline, a similar proportion of cortical and lacunar stroke participants were diagnosed with dementia (11%), at least one recurrent stroke (32%), or died (26%). Cortical stroke participants exhibited higher mortality rates, particularly from cardiovascular causes, compared to those with lacunar stroke. Cardiac conditions, such as ischaemic heart disease, arrhythmias, and heart failure, were also more prevalent in cortical stroke participants. Conversely, hypertension, hypercholesterolaemia, and diabetes were more frequent amongst lacunar stroke participants. These profiles coincide with previous research results and likely reflect the differing vascular pathologies underlying the two stroke subtypes: large artery atherothrombotic disease associated with cortical strokes and small vessel disease characteristic of lacunar strokes. Importantly, 45% of patients reported struggling with everyday activities. Half of lacunar stroke patients and a third of cortical stroke patients reported experiencing anxiety and/or depression, and over half of lacunar stroke patients reported experiencing subjective cognitive difficulties or concerns. Results from this study highlight the far-reaching and life-altering impacts on patient quality of life that persist for long after a mild stroke.

This translational work reinforces the view that pericyte and endothelial cell dysfunction play key roles in ageing and SVD mechanisms and provides further evidence supporting the modulating influence of endothelium-pericyte crosstalk on disease pathophysiology. A major challenge impending the advancement of therapeutics and the mechanistic understanding of SVD arises from the absence of elucidative and specific blood biomarkers. If validated in larger cohorts, these findings may offer a particularly exciting prospect to link manifestations of disease to otherwise unseen biological mechanisms and could help bridge the gap between what we know of SVD from preclinical and clinical investigations.