Blood-brain barrier disruption and microglia-associated synaptic pruning contribute to cognitive decline after electroconvulsive therapy.

BACKGROUND

Modified electroconvulsive therapy (MECT) is a highly effective intervention for major depressive disorder (MDD), yet its broader clinical application is constrained by poorly understood cognitive side effects. We hypothesized that MECT disrupts hippocampal blood-brain barrier (BBB) integrity, permitting extravasation of peripheral factors that activate microglia, drive synaptic phagocytosis, and precipitate cognitive decline.

METHODS

We employed a multi-modal translational approach integrating: (1) clinical neuroimaging (rs-fMRI and DCE-MRI) and cognitive assessments in 41 MDD patients undergoing MECT (including 9 with DCE-MRI); (2) a chronic restraint stress (CRS) mouse model subjected to electroconvulsive shock (ECS), with behavioral, molecular, and transcriptomic analyses; and (3) causal validation via microglial depletion (PLX5622) and targeted BBB repair (Axitinib). In vitro, HT22 hippocampal neurons were stimulated with soluble inflammatory mediators released by primed microglia to assess direct synaptic injury.

RESULTS

MECT significantly impaired immediate and delayed associative memory and reduced left hippocampal ALFF in MDD patients; memory decline correlated with ALFF reduction. In a paired subgroup of patients, DCE-MRI revealed increased hippocampal Ktrans after MECT, providing preliminary clinical imaging support for treatment-associated BBB hyperpermeability. In CRS mice, ECS downregulated tight junction proteins (Claudin-5, ZO-1, Occludin), induced Evans blue extravasation, and promoted hippocampal deposition of fibrinogen and complement C3. ECS also triggered microglial activation, enhanced PSD95 synaptic phagocytosis, elevated IL-6/IL-1β, and reduced synaptic proteins. Critically, both PLX5622-mediated microglial depletion and Axitinib-mediated BBB repair reversed these pathological changes and rescued cognitive function, establishing a dual causal validation system. Axitinib failed to suppress LPS-induced microglial activation, suggesting its protective effects were mediated via BBB restoration rather than direct immunomodulation.

CONCLUSIONS

In conclusion, this study identifies a hippocampus-centered neurovascular-immune mechanism underlying MECT/ECS-related cognitive impairment. Our findings demonstrate that BBB disruption is an early upstream event that permits peripheral factors to enter the brain, activate microglia, and drive synaptic injury and cognitive decline. These results establish a novel mechanistic framework for the cognitive side effects of MECT and highlight BBB protection and microglial modulation as potential strategies for cognitive protection.