Differential protein expression and enriched pathways in pediatric sepsis: identification of novel brain-associated biomarkers revealed through proteomic profiling

Background: Sepsis, defined by confirmed or suspected infection with systemic inflammatory response syndrome, requires robust biomarker identification. Proteomics enables protein quantification and expression analysis across disease states. This study investigated differential protein expression patterns, particularly brain-associated proteins, between sepsis patients and healthy controls, while evaluating temporal variations and relevant molecular pathways. Methods: Study participants were prospectively enrolled based on established pediatric sepsis criteria with clinical and blood samples collected. Plasma protein concentrations were quantified using Nucleic Acid Linked Immuno-Sandwich Assay methodology. Statistical analyses incorporated conventional statistics, bioinformatics and machine learning approaches. Results: The study cohorts comprised 23 age- and sex-matched participants: pediatric sepsis patients (median 11 years, IQR 9.5–14) and healthy controls (median 11 years, IQR 7.8–13; P = 0.809). Analyses revealed 59 differentially expressed proteins (DEPs) on Pediatric Intensive Care Unit Day 1 (PICU D1). Random Forest Classification (RFC) with Boruta feature selection identified 29 proteins that facilitated distinct group stratification. Comparison between PICU D1 and D3 samples yielded 34 DEPs, with RFC and Boruta feature selection isolating 9 discriminatory proteins. Multiple proteins were correlated with PELOD-2 scores and mortality (P < 0.05). Novel brain-associated proteins demonstrated significant differential expression patterns between PICU D1 and healthy controls, and over 3 days of PICU stay in sepsis patients. PICU D1 samples demonstrated significant pathway upregulation when compared to healthy controls, including “Signaling by Interleukins”, “Cytokine Signaling in Immune system”, and “Interleukin-10 signaling”. By PICU D3, pathways associated with “Generic Transcription Pathway”, “RNA Polymerase II Transcription”, and “Gene expression (Transcription)” exhibited significant downregulation. Protein-protein interaction network analysis revealed TNF and IL1B as critical bridging proteins linking inflammatory and neurological processes. Disease enrichment analysis demonstrated significant over-representation of respiratory pathology-associated genes, with respiratory failure and adult respiratory distress syndrome as the most enriched categories. Conclusions: Our investigation revealed distinct proteomic signatures in inflammatory and transcriptional pathways, including brain-associated processes, that differentiated pediatric sepsis patients from healthy control participants and exhibited temporal dynamics. The identification of TNF and IL1B as bridging proteins between systemic inflammation and neurological processes, combined with respiratory-centric disease enrichment patterns, provides mechanistic insights into sepsis pathophysiology. These alterations may provide insight into the mechanisms underlying sepsis-associated encephalopathy and lingering cognitive impairment in sepsis survivors, warranting further investigation in future studies. The identified molecular signatures present potential diagnostic and prognostic biomarkers for pediatric sepsis management.