Neuroinflammation Drives Perivascular Space Enlargement in Brain Small Vessel Disease

Cerebral small vessel disease (cSVD) is responsible for approximately 25% of all strokes and is the most common pathological substrate of vascular dementia. Among its MRI hallmarks, enlarged perivascular spaces (PVS) are thought to reflect impaired glymphatic drainage and fluid clearance around small brain vessels. Despite their clinical importance, the mechanisms driving PVS enlargement have remained poorly understood — with neuroinflammation, blood-brain barrier (BBB) dysfunction, and systemic inflammation all proposed but never rigorously tested together in humans using spatially resolved imaging.

This study recruited 54 symptomatic, sporadic cSVD patients (combining an initial observational cohort and baseline data from the MINERVA randomized trial) who underwent simultaneous PET-MRI on a 3T GE SIGNA scanner. Microglial activation was quantified using the radioligand 11C-PK11195, which binds the translocator protein (TSPO) upregulated during neuroinflammation. BBB permeability was measured concurrently using dynamic contrast-enhanced MRI (DCE-MRI) with a low-dose gadolinium contrast agent. PVS were quantified both by standardized visual rating (0–4 scale, separately for white matter and basal ganglia) and by automated volumetric segmentation. A key methodological innovation was the construction of concentric 'penumbra shells' around each segmented PVS, allowing the investigators to map 11C-PK11195 binding and BBB permeability as a function of distance from individual PVS. Systemic inflammation was assessed using a 93-protein Olink proximity extension assay panel covering cardiovascular disease, inflammatory, and endothelial biomarkers.

The spatial analysis revealed a clear neuroinflammatory signature around white matter PVS: tissue in the innermost shell (0–2 mm from PVS) showed significantly higher 11C-PK11195 binding than more distant shells (p < 0.001), with binding decreasing progressively as distance from the PVS increased. At the regional level, higher white matter PVS burden on the visual rating scale correlated with higher mean white matter 11C-PK11195 binding (Spearman ρ = 0.469, FDR-corrected p = 0.009), with a similar non-significant trend observed for PVS volume. Notably, these associations were specific to white matter PVS — no significant relationship was found between basal ganglia PVS burden and 11C-PK11195 binding in any analysis.

In striking contrast, no marker of PVS burden (visual rating or volume, in either white matter or basal ganglia) was associated with BBB permeability measured by DCE-MRI, either locally in PVS penumbrae or across the broader white matter region. Similarly, none of the 93 blood biomarkers of systemic inflammation showed significant correlation with any PVS measure after correction for multiple comparisons. This dissociation between CNS neuroinflammation and both BBB leakage and systemic inflammation suggests that the relationship between PVS and microglial activation represents a distinct, brain-intrinsic pathological process rather than a consequence of peripheral immune activation or vascular leak.

The findings carry meaningful clinical implications. They support neuroinflammation — specifically microglial activation — as a plausible mechanism in PVS enlargement and dysfunction, and by extension potentially in the broader glymphatic failure implicated in cSVD and dementia. The white matter specificity of the association (absent in basal ganglia PVS) is intriguing, potentially reflecting distinct pathophysiological mechanisms in these two locations. The authors note that the cross-sectional design precludes causal inference — it remains unknown whether microglial activation causes PVS enlargement, or whether enlarged PVS recruit inflammatory cells — and call for longitudinal and intervention studies, particularly trials of anti-inflammatory agents, to resolve this question.