Abstract
Neurobrucellosis is the most morbid form in brucellosis disease. Metabolomics is an emerging method which intends to explore the global alterations of various metabolites in samples. We aimed to identify metabolites in cerebrospinal fluid (CSF) as biomarkers that were potentially unique for neurobrucellosis. CSF samples from 25 neurobrucellosis patients and 25 normal controls (uninfected patients with hydrocephalus) were collected for metabolite detection using liquid chromatography-mass spectrometry (LC-MS) approach. Inflammatory cytokines in CSF were measured with Enzyme-linked immunosorbent assay (ELISA). The base peak chromatogram in CSF samples showed that small-molecule metabolites were well separated. Principal Component Analysis (PCA) analysis exhibited the examined samples were arranged in two main clusters in accordance with their group. Projection to Latent Structures Discriminant Analysis (PLS-DA) revealed there was a noticeable separation between neurobrucellosis and normal groups. Orthogonal Partial Least-Squares-Discriminant Analysis (OPLS-DA) could responsibly illuminate the differences between neurobrucellosis and normal controls. Neurobrucellosis showed a total of 155 differentiated metabolites. Prominent potential biomarkers including 30 metabolites were then selected out, regarded as more capable of distinguishing neurobrucellosis. TNF-α and IL-6 in CSF were remarkably increased in neurobrucellosis. We presented the heatmaps and correlation analyses among the identified 30 potential biomarkers. In conclusion, this study showed that CSF metabolomics based on LC-MS could distinguish neurobrucellosis patients from normal controls. Our data offered perspectives for diagnosis and treatment for neurobrucellosis.