A biomarker combination for sepsis diagnosis and application thereof

Abstract

This invention discloses a biomarker combination for sepsis diagnosis and its application, relating to the field of biomedical technology. This biomarker combination for sepsis diagnosis includes LTF, MMP9, MMP8, IL10, LCN2, and S100A12, which are significantly upregulated in sepsis patients, and the constructed nomogram model exhibits excellent diagnostic efficacy. Single-cell analysis shows that the biomarkers are mainly highly expressed in myeloid cells, especially neutrophils, indicating remodeling of cell communication networks in sepsis. Molecular docking and kinetic simulations confirm that the compound trimethoprim can stably bind to IL10, LCN2, LTF, and MMP9. This study provides novel biomarkers for the early diagnosis of sepsis and also provides a theoretical basis for therapeutic drugs targeting citrullination-related pathways.

Description

Technical Field

[0001] This invention relates to the field of biomedical technology, and in particular to a combination of biomarkers for the diagnosis of sepsis and their application. Background Technology

[0002] Sepsis is a severe systemic inflammatory response syndrome with a complex pathogenesis and high mortality rate, necessitating the discovery of diagnostically valuable biomarkers and potential therapeutic targets. However, early diagnosis of sepsis currently relies primarily on traditional biomarkers, which have limited sensitivity and specificity, failing to meet the precise clinical needs for early diagnosis and disease stratification. Therefore, screening novel biomarker combinations with high sensitivity and specificity, and deeply elucidating the immune regulatory mechanisms of sepsis, has become a key direction in sepsis diagnosis and treatment research.

[0003] Citrullination is a post-translational modification process in which arginine residues in proteins are converted into citrulline residues under the catalysis of the peptidylarginine deiminase (PAD) family. It is widely involved in protein structure regulation, inflammatory responses, and the regulation of immune cell function. There is an urgent need in this field to develop a comprehensive and systematic method for screening citrullination-related biomarkers in sepsis, and to identify novel combinations of biomarkers with clinical diagnostic value and therapeutic targeting potential. This would provide new technical solutions and theoretical basis for the early diagnosis, prognostic assessment, and targeted therapies for sepsis. Summary of the Invention

[0004] In view of this, the present invention addresses the deficiencies of the existing technology, and its main objective is to provide a combination of biomarkers for the diagnosis of sepsis and their applications. The present invention identifies a combination of biomarkers composed of six genes: LTF, MMP9, MMP8, IL10, LCN2, and S100A12. This combination is significantly upregulated in sepsis patients, and the diagnostic efficacy of the diagnostic model (AUC > 0.9) provides a novel and reliable molecular diagnostic tool for the early identification and risk stratification of sepsis.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: A combination of biomarkers for the diagnosis of sepsis, the combination of biomarkers comprising: LTF, MMP9, MMP8, IL10, LCN2, and S100A12.

[0006] As a preferred embodiment, the product includes a reagent kit.

[0007] As a preferred embodiment, the reagent comprises a substance that specifically binds to the biomarker combination LTF, MMP9, MMP8, IL10, LCN2 and S100A12.

[0008] As a preferred embodiment, the biomarker combination is upregulated in patients with sepsis, and the biomarker combination is used to construct a diagnostic prediction model.

[0009] As a preferred option, the diagnostic prediction model is a nomogram model, and the constructed diagnostic model has an AUC > 0.9.

[0010] The use of the aforementioned biomarker combination in the preparation of a drug for treating sepsis, wherein the drug is used to modulate the expression or activity of the biomarker combination.

[0011] As a preferred embodiment, the drug is used to target and regulate the biomarkers LTF, MMP8, MMP9, LCN2 and S100A12 enriched in neutrophils and to target and regulate the biomarker IL10 enriched in CD14++ monocytes / macrophages.

[0012] As a preferred option, the drug is trimethoprim.

[0013] Compared with existing technologies, this invention has significant advantages and beneficial effects. Specifically, as shown in the above technical solution, this invention identifies a biomarker combination consisting of six genes: LTF, MMP9, MMP8, IL10, LCN2, and S100A12. This combination is significantly upregulated in sepsis patients. The diagnostic model constructed through multi-omics integrated analysis and machine learning algorithms exhibits excellent diagnostic efficacy (AUC > 0.9) on both the training and external validation sets, providing a novel and reliable molecular diagnostic tool for the early identification and risk stratification of sepsis. Furthermore, this biomarker combination can be further used to prepare sepsis treatment drugs, exerting its effects by regulating the expression or activity of the aforementioned biomarkers. Specifically, it can target LTF, MMP8, MMP9, LCN2, and S100A12 enriched in neutrophils, and IL10 enriched in CD14++ monocytes / macrophages. Furthermore, trimethoprim can stably bind to IL10, LCN2, LTF, and MMP9, providing a candidate drug basis for sepsis treatment strategies targeting citrullination-related pathways and NETosis. This invention overcomes the limitations of traditional single-marker diagnostic sensitivity and specificity, constructing a systematic framework from marker screening and cell origin localization to immune network analysis, and has promising prospects for clinical translational applications.

[0014] To more clearly illustrate the structural features and effects of the present invention, a detailed description is provided below in conjunction with the accompanying drawings and specific embodiments. Attached Figure Description

[0015] Figure 1 This is a schematic diagram illustrating the construction and functional enrichment analysis of sepsis-related differentially expressed genes and WGCNA co-expression network of the present invention; wherein, Figure 1 (A) Volcano plot of differentially expressed genes between the sepsis group and the normal control group in the GSE185263 training set. Red indicates significantly upregulated genes, blue indicates significantly downregulated genes, and gray indicates genes with no significant difference. Representative differentially expressed genes are marked in the figure. (B) Cluster heatmap of the top 20 differentially expressed genes, showing a clear separation in transcriptional expression patterns between sepsis samples and normal control samples. (CD) WGCNA soft threshold screening results, including scale-free topological fit index and average connectivity changes, used to determine parameters for subsequent co-expression network construction. (E) Gene hierarchical clustering tree and module colors obtained by dynamic pruning, showing that genes with similar expression patterns are divided into different co-expression modules. (F) Module-trait correlation heatmap, used to assess the correlation between each gene module and sepsis and normal phenotypes, where the MEmagenta module shows a strong positive correlation with the sepsis phenotype. (G) Intersection analysis of differentially expressed genes, WGCNA-Sepsis module genes, and citrullination-related genes, resulting in 35 candidate genes. (H) GO functional enrichment analysis of 35 candidate genes, mainly involving biological processes and functions such as inflammatory response regulation, pathogen defense, neutrophil granule and RAGE receptor binding. (I) KEGG pathway enrichment analysis of 35 candidate genes, mainly enriched in cytokine-cytokine receptor interactions, JAK-STAT, TNF and IL-17 and other inflammation and immune-related signaling pathways.

[0016] Figure 2 This is a schematic diagram illustrating the screening and validation of key sepsis genes and candidate diagnostic biomarkers according to the present invention; wherein, Figure 2 (A) A PPI network was constructed based on 35 candidate genes, and four topology algorithms (Degree, EPC, MCC, and MNC) from the CytoHubba plugin were used to screen key nodes. After intersection analysis of the results of different algorithms, nine stable hub genes were obtained. (B) A diagnostic model was constructed based on the six feature genes finally selected. ROC curve analysis was performed on the training set and two external validation sets. The results showed that the model had high discrimination ability in different datasets. (C) Expression validation of the six feature genes IL10, LCN2, LTF, MMP8, MMP9, and S100A12 in the GSE185263 training set, GSE134347 validation set, and GSE154918 validation set. Box plots showed that these genes were significantly higher in the sepsis group than in the normal control group, suggesting that they have relatively stable cross-cohort expression differences and potential diagnostic value.

[0017] Figure 3 This is a schematic diagram of the sepsis risk nomogram model, molecular docking, and GSEA functional analysis of the present invention; wherein, Figure 3(A) A nomogram model of sepsis risk constructed based on LTF, MMP9, MMP8, IL10, LCN2, and S100A12. Each gene is assigned a score according to its expression level, and the total score is used to estimate the risk of sepsis. (B) ROC curve, decision curve, and calibration curve of the nomogram model. The ROC curve shows that the model has good diagnostic efficacy; the decision curve suggests that the model has a clinical net benefit within a certain threshold probability range; the calibration curve shows that there is good consistency between the predicted probability and the actual outcome. (C) Molecular docking results of the candidate drug trimethoprim with LTF, showing its binding conformation and key amino acid interactions in the LTF protein binding pocket, suggesting that the two have potential binding stability. (D) Single-gene GSEA enrichment analysis of six characteristic genes showed that the LTF, MMP9, MMP8, IL10, LCN2 and S100A12-related pathways are mainly involved in interferon response, inflammation activation, cellular stress, tissue remodeling and immune regulation, providing functional support for the relationship between these markers and sepsis-related inflammatory and immune changes.

[0018] Figure 4 This is a schematic diagram of Western blot verification of PAD4-CitH3-NETs axis-related protein expression in the kidney tissue of CLP mice according to the present invention; wherein, Figure 4 In the diagram: the three columns on the left represent the control group, and the three columns on the right represent the CLP model group. The detection indicators included Cit Histone H3, Histone H3, Ly-6G, MPO, and PAD4, with β-actin used as an internal reference. Detailed Implementation

[0019] The present invention is as follows Figure 1 As shown in Figure 4, a combination of biomarkers for the diagnosis of sepsis includes: LTF, MMP9, MMP8, IL10, LCN2, and S100A12.

[0020] This product includes a reagent kit.

[0021] The reagent includes substances that specifically bind to the biomarker combination LTF, MMP9, MMP8, IL10, LCN2, and S100A12.

[0022] This combination of biomarkers was upregulated in patients with sepsis and was used to construct a diagnostic prediction model.

[0023] The diagnostic prediction model is a nomogram model, and the constructed diagnostic model has an AUC > 0.9.

[0024] An application of the biomarker combination in the preparation of a drug for treating sepsis, wherein the drug is used to modulate the expression or activity of the biomarker combination.

[0025] This drug is used to target and regulate the biomarkers LTF, MMP8, MMP9, LCN2 and S100A12 enriched in neutrophils and to target and regulate the biomarker IL10 enriched in CD14++ monocytes / macrophages.

[0026] The drug is trimethoprim.

[0027] Example: A combination of biomarkers for the diagnosis of sepsis and its application Materials and Methods: Data sources and differential gene screening: Download the sepsis-related transcriptome dataset from the NCBI GEO public database: training set GSE185263 (sepsis=348, healthy=44), validation set GSE154918 (sepsis=24, healthy=40), validation set GSE134347 (sepsis=156, healthy=83), and single-cell dataset GSE167363 (sepsis=10, healthy=2). Also obtain the corresponding platform annotation files for probe ID conversion.

[0028] Citrullination-related genes were collected from the GeneCards database using the search terms "Citrullination", "PAD2", and "PAD4". After deduplication, a total of 1689 genes were obtained. In the training set GSE185263, DESeq2 (Version 1.38.3, https: / / github.com / thelovelab / DESeq2) was used to identify differentially expressed genes (DEGs) between sepsis patients and controls. The selection criteria were corrected p-value < 0.05 and |log2FC| > 2. ggplot2 (version 3.5.2, https: / / ggplot2.tidyverse.org / ) and pheatmap (version 1.0.13, https: / / www.rdocumentation.org / packages / pheatmap / versions / 1.0.13 / topics / pheatmap) were used for image plotting. The DEG results were presented as volcano plots and heatmaps. The "WGCNA" R... The package constructed a gene co-expression network from normal and diseased samples in the training set, identifying key module genes significantly associated with the disease (P<0.05), named WGCNA-Sepsis. The "Venn Diagram" R package was used to construct Venn diagrams for DEGs, WGCNA-Sepsis, and citrullination-related genes. Intersections were used to obtain candidate genes, which were then displayed using Venn diagrams. The "clusterProfiler" package (Version 4.16.0, https: / / www.bioconductor.org / packages / release / bioc / html / clusterProfiler.html) was used to perform GO and KEGG pathway enrichment analysis on the candidate genes, with p.adjust < 0.05 and Count>1 used as the criteria for significant enrichment.

[0029] PPI Network and Machine Learning Model Construction: The obtained candidate genes were imported into the STRING database (https: / / string-db.org / ), with the species set as human (Homo sapiens) and an interaction score ≥0.4, to construct a protein-protein interaction (PPI) network. The PPI relationship network obtained from the STRING database was imported into Cytoscape software (Version 3.9.1, https: / / cytoscape.org), and four topology analysis algorithms (MCC, MNC, Degree, and EPC) from the CytoHubba plugin were used to predict and explore the top 20 important genes in the PPI network. The genes at the intersection of the four algorithms were identified as key genes. Twelve machine learning algorithms were employed: Lasso, Ridge, Stepglm, XGBoost, Random Forest (RF), Elastic Net (Enet), Partial Least Squares Regression for Generalized Linear Models (plsRglm), Generalized Boosted Regression Modeling (GBM), NaiveBayes, Linear Discriminant Analysis (LDA), Generalized Linear Model Boosting (glmBoost), and Support Vector Machine (SVM) to construct diagnostic models. 113 combinations of these algorithms were analyzed using the training dataset (GSE185263), and variable selection and model development were performed within a ten-fold cross-validation framework. The performance of the models was then validated using the analysis sets (GSE154918 and GSE134347) as external test datasets. The model with the highest average C-index in both the training and analysis sets was defined as the optimal model, and the genes selected from the optimal model were defined as feature genes. Then, diagnostic efficacy was evaluated using the pROC package (Version 1.18.5, https: / / www.rdocumentation.org / packages / pROC / versions / 1.18.5), with ROC curves plotted for the characteristic genes on both the training and validation sets. To assess the ability of the characteristic genes to distinguish between disease and control samples, Wilcoxon analysis was performed on the sepsis training and validation sets to analyze the expression of the characteristic genes in the sepsis and control groups. Finally, genes showing significant differences in expression levels (corrected P < 0.05) and consistent expression trends in both the sepsis training and validation sets were selected as biomarkers.To evaluate the predictive value of all biomarkers for sepsis, nomograms were constructed using the R package "rms" (Version 6.7-1, https: / / www.rdocumentation.org / packages / rms) based on the biomarkers in the sepsis training set. Calibration curve analysis was performed using the R package "rms", and DCA decision curves were plotted using the R package "ggDCA" (Version 1.2, https: / / github.com / yikeshu0611 / ggDCA) to evaluate the accuracy of the model's predictions. To assess the expression of characteristic genes in the sepsis group and the control group, Wilcoxon tests were performed on the training set (GSE185263) and the validation set (GSE154918, GSE134347). Genes with significantly different and consistent expression levels were ultimately selected as biomarkers.

[0030] Nodal plot construction and calibration and single-gene GSEA analysis: To evaluate the predictive value of the selected biomarkers for sepsis, nomograms were constructed using the R package "rms" on the training set (GSE185263). Each biomarker was scored based on its expression level, and the scores of all biomarkers were summed to obtain a total score. The incidence of sepsis was inferred from the total score. ROC curves were used to measure model performance, and DCA decision curves were used to evaluate model accuracy.

[0031] Discussion of Results: Co-expression network and functional enrichment analysis of sepsis-related differentially expressed genes and WGCNA A total of 582 upregulated genes and 25 downregulated genes were obtained from the training set GSE185263. Based on the differentially expressed genes, a volcano plot was drawn for the top 20 genes. Figure 1 (A) and heat map ( Figure 1 (B). WGCNA analysis was performed on differentially expressed genes, with a threshold of 8, such as... Figure 1 C, Figure 1 In the middle D; based on clustering and dynamic pruning methods, highly correlated genes are grouped into the same module. The modules are then clustered, and modules with correlation coefficients > 0.75 are merged to obtain 9 modules, such as... Figure 1 In E, the correlation between the feature vector genes and phenotypes of each module is calculated, such as... Figure 1 In the middle F., the intersection of DEGs, WGCNA-Sepsis, and citrullination-related genes yielded 35 candidate genes, such as Figure 1 G. GO enrichment analysis was performed on candidate genes, such as... Figure 1In the H group, a total of 275 pathways were enriched, including 245 BP (Biological Process) pathways, 10 CC (Cellular Component) pathways, and 20 MF (Molecular Function) pathways. KEGG enrichment analysis was performed on candidate genes, such as... Figure 1 In the middle I, a total of 7 pathways were enriched, which, in descending order of gene number, are: Inflammatory bowel disease, JAK-STAT signaling pathway, Cytokine-cytokine receptor interaction, IL-17 signaling pathway, Amoebiasis, TNF signaling pathway, and Fructose and mannose metabolism.

[0032] Screening and validation of key gene biomarkers for sepsis: PPI network, intersection analysis, ROC assessment and expression validation To investigate the interactions between proteins encoded by 35 candidate genes, a PPI network of the candidate genes was constructed using the STRING website. Subsequently, four topology algorithms (Degree, EPC, MCC, and MNC) from the CytoHubba plugin were used to predict and explore the top 10 important genes in the PPI network. The four algorithms yielded nine key intersection genes, such as... Figure 2 The nine key genes identified are S100A12, IL10, MMP9, MMP8, LCN2, LTF, ARG1, RETN, and S100A8. For these nine selected genes, 12 machine learning algorithms were used in combination analysis on the training set (GSE185263). Variable selection and model development were performed within a ten-fold cross-validation framework, and the model performance was externally validated using validation sets (GSE154918 and GSE13447). The Stepglm[backward] + Ridge model was ultimately selected as the optimal model. Figure 2 In the B dataset, the average AUC value was 0.981. Subsequently, six genes upregulated in the disease were selected as characteristic genes: LTF, MMP9, MMP8, IL10, LCN2, and S100A12. ROC curves of the optimal model were plotted on both the training and validation sets. The AUC values ​​were all greater than 0.9, indicating good diagnostic performance on both sets. The expression of these six genes in sepsis and the control group was analyzed using the Wilcoxon test on three independent sepsis datasets. Figure 2 As shown in C.

[0033] Sepsis Risk Nodal Plot Model and Hub Gene Function Validation: Molecular Docking and GSEA Analysis To analyze the diagnostic efficacy of six biomarkers for sepsis, this study constructed a nomogram to visualize the predictive ability of each gene for sepsis, such as... Figure 3 As shown in Figure A, low expression of LTF and MMP9, and high expression of MMP8, IL10, LCN2, and S100A12 genes are risk factors for sepsis. The AUC of the ROC curve for the sepsis training set is ≥0.9. Figure 3 The result in B indicates that the nomogram model has excellent diagnostic performance. Decision curve results show that the threshold probability range of 0.2-0.8 is clinically applicable, and the calibration curve shows good consistency between predicted and actual probabilities. Drug prediction was performed on six biomarkers, and multiple biomarkers were screened to obtain trimethoprim, a drug targeting IL-10, LCN2, MMP9, and LTF genes, which underwent molecular docking. Figure 3 Among the C-type complexes, LTF exhibits the best binding affinity to trimethoprim. The LTF-trimethoprim complex system was evaluated using root mean square deviation (RMSD). It reached and maintained structural equilibrium in a 100 ns simulation, demonstrating both good stability and reasonable dynamic behavior, providing a reliable foundation for subsequent functional verification and binding mechanism analysis.

[0034] Western blot validation of PAD4-CitH3-NETs axis-related proteins in CLP mouse model To further verify whether citrullination and neutrophil-related inflammatory activation exist in the sepsis animal model, this study used Western blot to detect changes in the protein expression of Cit Histone H3, Histone H3, Ly-6G, MPO, and PAD4 in the kidney tissues of mice in the CLP model group and the control group. Results Figure 4The three columns on the left represent the control group, and the three columns on the right represent the CLP model group. Compared with the control group, the Cit Histone H3 band in the model group was significantly enhanced overall, while the change in the total Histone H3 band was relatively small, suggesting an increase in histone H3 citrullination levels in renal tissue under CLP-induced sepsis. Simultaneously, PAD4 protein expression in the model group showed an increased trend compared to the control group, indicating that the key enzyme catalyzing histone citrullination was activated. Ly-6G and MPO, as indicators related to neutrophil infiltration and activation, also showed strong band signals in the model group, suggesting enhanced neutrophil recruitment and inflammatory response in renal tissue after CLP. Considering the trends in Cit Histone H3, PAD4, Ly-6G, and MPO, these results support the existence of PAD4-mediated enhanced histone citrullination in the CLP model, accompanied by an increase in neutrophil-related NET formation or activation. The results of this animal experiment further supplement the previous bioinformatics screening results at the in vivo level, suggesting that changes in citrullination-related molecules are involved in sepsis-related inflammatory damage to kidney tissue, and providing experimental evidence for further in-depth mechanistic studies on the relationship between the PAD4-CitH3-NETs axis and candidate biomarkers.

[0035] Overview of Citrullinated NETs Axis and Core Biomarkers Sepsis is a severe and rapidly progressing disease caused by immune dysregulation, characterized by abnormal neutrophil and macrophage function. Extraneutrophil traps (NETs), a key mechanism in the body's defense against disease, exhibit excessive formation (NETosis), leading to uncontrolled inflammatory responses, tissue damage, and organ failure. NETs play a dual role, providing antibacterial protection and driving inflammation. Citrullinated histone H3 (CitH3), a key component released during NETosis, has become a crucial mediator in sepsis and is currently a promising biomarker for the early diagnosis of sepsis and other inflammatory conditions. This study, through differential gene expression analysis and WCGNA analysis on multiple datasets, identified six sepsis biomarkers associated with citrullination: LTF, MMP8, MMP9, IL10, LCN2, and S100A12. These genes were significantly upregulated in both the training and validation sets. The constructed diagnostic model showed an AUC > 0.9, and the nomogram demonstrated good clinical applicability, providing a new tool for the early diagnosis of sepsis. Furthermore, four distinct subtypes were identified through single-cell analysis, key cell localization, and heterogeneity analysis: CD14++ Monocytes / Macrophages, Neutrophils, CD14+ Monocytes / Macrophages, and DCs. Temporal dynamics of gene expression experienced by each cell during changes in the state of different subtypes were investigated using pseudo-temporal analysis. The study also analyzed the crucial roles of citrullination-related genes in sepsis progression.

[0036] Research and Clinical Significance of Six Citrullinated Biomarkers The study identified six different citrulline-related sepsis biomarkers, with varying expression across different phenotypes. S100A12, a calcium-binding protein in the S100 family with core inflammatory regulatory functions, is a neutrophil-specific marker, significantly elevated in sepsis-induced myocardial dysfunction (SIMD) and ARDS. Furthermore, studies have reported that S100A12's diagnostic ability in sepsis is superior to that of traditional inflammatory markers. LTF, an iron-rich glycoprotein, is a major component of neutrophils and participates in the body's antimicrobial defense and inflammatory regulation. MMP8 and MMP9 both belong to matrix metalloproteinases (MMPs) and are released by neutrophils; IL-10 is one of the key anti-inflammatory cytokines in the sepsis immune regulatory network.

[0037] This study identified six core biomarkers related to citrullination in sepsis (LTF, MMP9, MMP8, IL10, LCN2, and S100A12) through multi-omics integrated analysis and bioinformatics methods. These biomarkers demonstrated stable diagnostic efficacy in cross-cohort validation, suggesting their potential as candidate indicators for early sepsis identification and risk stratification. Further analysis combining immune infiltration and single-cell transcriptomics revealed that these biomarkers primarily originate from the myeloid cell system: LTF / MMP8 / MMP9 / LCN2 / S100A12 are mainly enriched in neutrophils, reflecting enhanced effector programs such as degranulation, protease release, tissue infiltration, and inflammatory amplification; IL10 is mainly enriched in CD14++ monocytes / macrophages, indicating the initiation of anti-inflammatory negative feedback and immune braking programs, revealing a structural imbalance in the sepsis immune response where a "pro-inflammatory axis" and an "immunosuppressive axis" coexist. Cell communication and pseudo-time series results further indicate that myeloid cells exhibit continuous state evolution and branching pathways, suggesting that different immune states correspond to different disease stages and clinical phenotypes. Overall, this study constructed a systematic framework for sepsis and citrullinated / NETs-related immune imbalances from three levels: biomarker screening, cell origin localization, and immune network analysis. This provides a theoretical basis and candidate targets for subsequent protein-level and mechanism validation, the establishment of translatable combined diagnostic / stratification models, and the exploration of intervention strategies targeting myeloid effector programs and NET-related pathways.

[0038] The key design focus of this invention is the identification of a biomarker combination consisting of six genes: LTF, MMP9, MMP8, IL10, LCN2, and S100A12. This combination is significantly upregulated in sepsis patients. A diagnostic model constructed using multi-omics integrated analysis and machine learning algorithms exhibits excellent diagnostic efficacy (AUC > 0.9) on both the training and external validation sets, providing a novel and reliable molecular diagnostic tool for early identification and risk stratification of sepsis. Furthermore, this biomarker combination can be further used to prepare sepsis therapeutic drugs, exerting their effects by regulating the expression or activity of the aforementioned biomarkers. Specifically, it can target LTF, MMP8, MMP9, LCN2, and S100A12 enriched in neutrophils, and IL10 enriched in CD14++ monocytes / macrophages. In addition, trimethoprim can stably bind to IL10, LCN2, LTF, and MMP9, providing a candidate drug basis for sepsis treatment strategies targeting citrullination-related pathways and NETosis. This invention breaks through the limitations of traditional single biomarker diagnostic sensitivity and specificity, and constructs a systematic framework from biomarker screening, cell origin localization to immune network analysis, which has good prospects for clinical translational application.

[0039] The above description is merely a preferred embodiment of the present invention and does not constitute any limitation on the technical scope of the present invention. Therefore, any minor modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention shall still fall within the scope of the technical solution of the present invention.

Claims

1. A combination of biomarkers for the diagnosis of sepsis, characterized in that: The biomarker combination includes: LTF, MMP9, MMP8, IL10, LCN2, and S100A12.

2. The application of the reagent for detecting the expression level of the biomarker combination described in claim 1 in the preparation of sepsis diagnostic products, characterized in that: The product includes a reagent kit.

3. The application according to claim 2, characterized in that: The reagents include substances that specifically bind to the biomarker combination LTF, MMP9, MMP8, IL10, LCN2 and S100A12.

4. The application according to claim 2, characterized in that: The biomarker combination is upregulated in patients with sepsis, and the biomarker combination is used to construct a diagnostic prediction model.

5. The application according to claim 4, characterized in that: The diagnostic prediction model is a nomogram model, and the constructed diagnostic model has an AUC > 0.

9.

6. The use of the biomarker combination as described in claim 1 in the preparation of a drug for treating sepsis, characterized in that: The drug is used to regulate the expression or activity of the combination of biomarkers.

7. The application according to claim 6, characterized in that: The drug is used to target and regulate the biomarkers LTF, MMP8, MMP9, LCN2 and S100A12 enriched in neutrophils and to target and regulate the biomarker IL10 enriched in CD14++ monocytes / macrophages.

8. The application according to claim 6, characterized in that: The drug in question is trimethoprim.

Images