Proteomic Signatures in Cerebrospinal Fluid Reveal Biomarkers for Central Nervous System Infections

Abstract

Central nervous system (CNS) infections are life-threatening conditions that require rapid and accurate diagnosis to guide appropriate treatment. Conventional diagnostic methods such as cerebrospinal fluid (CSF) culture, Gram stain, and PCR often lack sensitivity, particularly after prior antibiotic exposure or delayed sampling and emerging next generation sequencing is currently time intensive. Proteomic analysis offers a promising complementary approach by enabling unbiased, high-throughput characterization of the CSF protein landscape which can identify biomarkers of CNS infection. Unlike pathogen-specific tests, proteomics captures the host’s immune and inflammatory responses, which may persist after pathogen clearance. Previous studies suggest that proteomic profiling can distinguish bacterial from viral infections, identify disease-specific protein signatures, and reveal biomarkers associated with blood–brain barrier dysfunction, neuronal injury, and immune activation. This study explores the feasibility of using CSF proteomics to differentiate bacterial, viral, and autoimmune CNS pathologies from noninfectious controls. Up to six female patients were included across four groups: confirmed bacterial CNS infection, confirmed viral CNS infection, autoimmune encephalitis, and absence of CNS infection. Restricting to female patients reduced potential sex-related variability in proteomic analysis. Demographic and clinical data were extracted from medical records. CSF samples were collected within 24 hours of completion of routine diagnostic testing, frozen, and subsequently analyzed by mass spectrometry. Bacterial pathogens included Haemophilus influenzae, Streptococcus pneumoniae, Enterococcus faecalis, Streptococcus agalactiae, and Enterobacter cloacae. Viral pathogens included HSV, parechovirus, HHV-6, and influenza. Differential protein expression was observed across all disease groups compared with controls. The bacterial CNS infection group demonstrated the greatest proteomic changes, with 214 upregulated and 240 downregulated proteins versus controls. Viral CNS infection showed 40 upregulated and 55 downregulated proteins, while autoimmune encephalitis exhibited 15 upregulated and 28 downregulated proteins. Between-group comparisons revealed distinct expression patterns, including 294 upregulated and 303 downregulated proteins in bacterial versus viral infection. Notably, S100 calcium-binding protein A9 (S100A9) emerged as a potential biomarker differentiating bacterial CNS infection from viral, autoimmune, and noninfectious conditions. This pilot study demonstrates the feasibility of using CSF proteomics to distinguish bacterial, viral, and autoimmune CNS pathologies from noninfectious states. Differential protein expression profiles highlight the diagnostic potential of host-response biomarkers, supporting proteomics as a valuable adjunct in the evaluation of CNS infections, particularly when conventional microbiologic testing is inconclusive. Furthermore, identification of dynamic protein signatures may enable development of prognostic and treatment-monitoring biomarkers, advancing precision management of CNS infections.