Mônica Bragança Sousa¹, Giuseppe Gianini Figueiredo Leite¹, Jackelinne Yuka Hayashi¹, Alexandre Keiji Tashima¹, Reinaldo Salomão¹

Introduction: Sepsis is a major cause of death in intensive care units. In 2017, an estimated 48.9 million cases and 11 million related deaths occurred worldwide; in Brazil, lethality can reach 55%. Neutrophils, the first line of defense against pathogens, rely on glycolysis for their effector functions and, when dysfunctional, are associated with worse outcomes.

Methods: ICU patients diagnosed with sepsis within the first 48 hours after organ dysfunction or shock were included. PMNs were isolated by Ficoll-Paque Plus density gradient and stored in liquid nitrogen. A total of 45 patients were selected: 21 with community-acquired (CA) of whom 9 died and 12 survived, and among the remaining 24 patients with hospital acquired (HA), 12 patients died and 12 survived., along with 10 healthy controls. For proteomics, PMNs were lysed, proteins digested, and peptides labeled with TMT for nanoLC-MS/MS analysis on an Orbitrap Fusion Lumos. Normalized TMT reporter abundances were log₂-transformed in R and compared using empirical Bayes moderated t-tests with Benjamini-Hochberg correction. Proteins with p < 0.05 and |log₂ fold change| > 0.26 were considered differentially abundant. Untargeted Protein Set Enrichment Analysis (PSEA) was performed using Reactome and MSigDB Hallmark gene sets.

Discution: We identified 813 proteins in purified PMNs with distinct abundance profiles between septic patients and healthy controls. Of these, 333 were differentially abundant (281 upregulated, 52 downregulated; |log₂ FC| > 0.26; p ≤ 0.05), indicating substantial alterations in cellular function during sepsis. Comparing protein abundance by outcome (survivors vs. non-survivors) and infection source (community-acquired [CA] vs. hospital-acquired [HA]), only 13 proteins differed, suggesting significant changes are restricted to specific clinical subgroups. CA patients were older (73 ± 12 vs. 60 ± 18 years, p = 0.003) with higher respiratory dysfunction (38% vs. 8.3%) but lower septic shock incidence (38% vs. 79%, p = 0.007), with only 4 proteins differing significantly, indicating partial influence of clinical factors. Stratified analyses revealed 49 differentially abundant proteins in CA survivors vs. non-survivors and 30 in HA survivors vs. non-survivors. Untargeted protein set enrichment analysis (PSEA) using Reactome and Hallmark gene sets showed that CA non-survivors had suppressed core metabolic pathways (RNA/protein metabolism, mitochondrial respiration, central carbon metabolism) alongside upregulated inflammation, platelet activation, programmed cell death, glycolysis, and hypoxia, suggesting metabolic dysfunction and hyperinflammation drive fatal outcomes. HA non-survivors exhibited a similar but attenuated profile, with modest reductions in oxidative phosphorylation and lipid metabolism and enrichment in glycolysis, hypoxia, and apoptosis, indicating severity varies by infection source. These findings demonstrate that the most pronounced proteomic alterations occur in clinically homogeneous subgroups, providing insights into molecular mechanisms underlying sepsis mortality.

Conclusion: The most pronounced proteomic changes in neutrophils of patients occur in clinically homogeneous subgroups, particularly when considering outcome and infection source. Non-survivors show suppression of core metabolic pathways combined with inflammation, programmed cell death, and stress responses, suggesting that metabolic dysfunction and immune hyperactivation are linked to fatal sepsis outcomes.

Agradecimentos: We thank the São Paulo Research Foundation and the National Council for Scientific and Technological Development (CNPq) for financial support.