Efficiency of proteomics combined with transcriptomics in identifying differentially expressed genes in sepsis-induced lung injury and independent sample validation

Objective To investigate biomarkers for sepsis-induced lung injury based on results of proteomics combined with transcriptomics analyses.

Methods A total of 70 patients with sepsis and 70 patients with sepsis-induced lung injury were included as research objects. These patients were divided into experimental group and validation group. The experimental group included 10 patients with sepsis and 10 patients with sepsis-induced lung injury. Proteomics was used to analyze differentially expressed proteins in plasma, and Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes(KEGG) enrichment analyses were performed. The dataset GSE10474 of sepsis-induced lung injury was downloaded from the Gene Expression Omnibus (GEO), and GEO2R online database was used to analyze differential transcriptomics data for sepsis-induced lung injury. A Venn diagram was used online to analyze common differentially expressed genes related to sepsis-induced lung injury in proteomics and transcriptomics. The validation group included 60 patients with sepsis (sepsis group) and 60 patients with sepsis-induced lung injury (sepsis-induced lung injury group). Enzyme-linked immunosorbent assay (ELISA) was used to detect and compare the differences in protein expression level in peripheral blood between the sepsis group and the sepsis-induced lung injury group. Receiver operating characteristic (ROC) curve was used to analyze the clinical value of differential protein expression level in distinguishing sepsis and sepsis-induced lung injury.

Results Proteomics results confirmed the presence of 239 significantly differentially expressed proteins in the plasma of patients with sepsis and sepsis-induced lung injury. Compared with patients with sepsis, there were 96 significantly upregulated proteins and 143 significantly downregulated proteins in patients with sepsis-induced lung injury. The results of GO enrichment analysis of differentially expressed proteins included cytoplasm, microtubule binding, adenosine triphosphate (ATP) binding, defense response to virus, and immune response. The results of KEGG enrichment analysis included metabolic pathways, interleukin-17 (IL-17) signaling pathway, and phosphatidylinositol-3-kinase-protein kinase B (PI3K-Akt) signaling pathway. In the GSE10474 dataset, compared with patients with sepsis, there were 77 significantly upregulated genes and 142 significantly downregulated genes in patients with sepsis-induced lung injury. The Venn diagram results showed that there were 6 common differentially expressed genes in proteomics and transcriptomics, namely BTNL8, FCGR2B, TAK1, KCNC1, TREM1, and SEC31A. Compared with the sepsis group, the levels of TAK1 and TREM1 proteins in the peripheral blood of the sepsis-induced lung injury group were significantly increased (P < 0.01). ROC curve showed that the areas under the curve (AUC) for the expression levels of TAK1 and TREM1 proteins in serum to distinguish sepsis and sepsis-induced lung injury were 0.925 and 0.785 respectively; when the cut-off value for TAK1 was 71.28 pg/mL, the sensitivity and specificity were 94.45% and 97.89% respectively; when the cut-off value for TREM1 was 58.22 mg/mL, the sensitivity and specificity were 83.43% and 82.19% respectively.

Conclusion Proteomics and transcriptomics results confirm that the activation of TAK1, TREM1 and multiple inflammatory signaling pathways may play important roles in the progression of sepsis-induced lung injury.