Characterizing CSF inflammatory proteomics in pediatric post-hemorrhagic hydrocephalus and Anti-NMDAR encephalitis

Background
Neuroinflammation is a key contributor to pathology in many central nervous system (CNS) diseases. While cerebrospinal fluid (CSF) proteomic studies in adult neurologic conditions have identified insightful inflammatory signatures, fewer studies have been conducted in pediatric diseases. Moreover, past studies primarily employed proteomic approaches that are less suited to detect novel, low-abundance inflammatory mediators that may be critical in pediatric CNS pathophysiology. To address this gap, we applied high-sensitivity, multi-targeted proteomic profiling to characterize the neuroinflammatory signatures across three distinct pediatric neurologic diseases: post-hemorrhagic hydrocephalus (PHH), N-methyl-D-aspartate receptor encephalitis (NMDARE), and brain tumor-associated hydrocephalus.

Methods
CSF samples from controls (n = 5) and patients with PHH (n = 9), NMDARE (n = 5), and brain tumor-associated hydrocephalus (n = 10) were obtained from a pediatric CSF biorepository. After proteomic profiling using the Olink Explore platform, 641 inflammation-related proteins were retained for analysis. Differentially abundant proteins (DAPs) were identified using limma with false discovery rate (FDR) correction (FDR < 0.05, |log₂FC| >1). Pathway enrichment of DAPs was performed with Reactome via Enrichr, and protein–protein interaction networks were constructed using STRING to identify functional modules and key hub proteins.

Results
Principal component and hierarchical clustering analyses revealed separation of PHH and brain tumor samples from controls, while NMDARE partially overlapped. CSF in PHH contained 532 DAPs, with pathway enrichment analysis identifying alternative complement activation, coagulation, and platelet degranulation pathways as top hits. Conversely, CSF in NMDARE showed only 65 DAPs, with the top pathways involving IL-10 and IL-18 signaling, and the top 3 DAPs involving humoral immunity (IGLC2, MZB1, CD79B). DAPs did not meet statistical significance in brain tumor patients. Longitudinal analysis of serial collections from PHH patients suggested a persistence of coagulation- and complement-related neuroinflammation over time. NELL2 emerged as a consistently downregulated protein in PHH for weeks after the initial hemorrhage.

Conclusions
PHH and NMDARE revealed distinct neuroinflammatory proteomic signatures compared to our control samples. PHH was marked by a broad increase in detection of the majority of inflammation-related proteins, with highest representation among the alternative complement and coagulation-related pathways. The persistent detection of these proteins for weeks after the initial hemorrhage may be indicative of chronic neuroinflammation, even at the time of permanent CSF diversion. Conversely, NMDARE induced a narrower lymphocyte-driven profile, more consistent with an antibody-mediated autoimmune disease. Furthermore, suppression of NELL2 and up-regulation of immunoglobulin-related markers (IGLC2, MZB1, CD79B) are potential candidates for biomarkers in PHH and NMDARE, respectively.