Serum and cerebrospinal fluid neuroinflammatory biomarkers and trimethylamine N-oxide: associations with white matter lesion severity.

Introduction: This study aimed to investigate the association between a multi-panel biomarker profile across the blood-brain barrier and the severity of white matter lesions (WML) in 42 patients [mean age: 60.3 ± 12.1 years; 45.2% male], suggesting a potential inflammatory endotype rather than defining a distinct phenotype due to the cross-sectional design and limited sample size. We assessed the gut-brain axis role in WML pathogenesis through analysis of paired serum and cerebrospinal fluid samples. Methods: In this cross-sectional clinical study, 42 patients (aged 16-78 years; inclusion of younger subjects was based on radiological confirmation of WML unrelated to acute cerebrovascular events) with radiologically confirmed WML were stratified into four severity groups based on combined Fazekas global score (range 0-3, incorporating both periventricular and deep white matter hyperintensities). Levels of six biomarkers-interleukin-1β (IL-1β), matrix metalloproteinase-2 (MMP-2), tumor necrosis factor-α (TNF-α), trimethylamine N-oxide (TMAO), S100 calcium-binding protein β (S100β), and immunoglobulin G (IgG)-were simultaneously quantified in paired serum and cerebrospinal fluid (CSF) samples using ELISA. Statistical analyses included: (1) group comparisons using one-way ANOVA with LSD Results: Severity-dependent increases in IL-1β, TNF-α, and the gut-derived metabolite TMAO were observed in both serum and CSF. After adjustment for confounders, CSF TNF-α (adjusted OR = 1.95, 95% CI: 1.42-2.68, P<0.001) and serum MMP-2 (adjusted OR = 1.72, 95% CI: 1.28-2.31, P<0.001) demonstrated the strongest correlations with WML severity among all biomarkers, establishing them as primary candidate biomarkers. An inverse concentration pattern (elevated serum/decreased CSF) was observed for MMP-2, suggesting stage-specific roles in blood-brain barrier dynamics. CSF S100β levels increased with severity (r=0.487, P<0.001), consistent with parenchymal injury. Regarding IgG, while univariate analysis showed marginal significance in CSF (P = 0.045), FDR correction rendered this non-significant (q=0.12), and no significant correlation was found with Fazekas grade (P = 0.068), suggesting limited association with WML severity in this cohort. Conclusion: This study suggests a potential neuroinflammatory profile in WML, characterized by coupled peripheral and central inflammation associated with gut-derived TMAO, though causality cannot be inferred from this cross-sectional design. CSF TNF-α and serum MMP-2 are proposed as precision biomarkers for stratifying WML severity. These findings should be interpreted as a hypothesis-generating framework integrating dysregulated gut-brain axis signaling, central cytokine surges, and blood-brain barrier disruption into a unified pathogenic cascade, providing mechanistic insights for targeted anti-inflammatory strategies in cerebrovascular disease. However, the clinical applicability requires further validation given the invasive nature of CSF collection.