Exploring the Protective Mechanism of Xuebijing Injection Against Sepsis

Introduction

Sepsis is a life-threatening condition caused by a dysregulated host response to infection or inflammation. In septic shock, hospital mortality exceeds 40%.^1–5 According to the Global Burden of Diseases report, sepsis caused 48.9 million cases and accounted for nearly 20% of global deaths in 2017.⁶ Timely recognition relies on an accurate definition, which has been revised over time, from the first consensus in 1991 to the latest international guidelines in 2021, in response to its rising incidence.²

Given this significant global health burden, improving survival is a critical objective. Xuebijing injection (XBJ), a traditional Chinese medicine formulation, has been approved by China’s State Food and Drug Administration as a Category II New Drug for sepsis treatment based on its demonstrated clinical efficacy.^7–9 XBJ contains five herbal components: Hong hua (Carthami Flos) from Carthamus tinctorius L., Chi shao (Paeoniae Radix Rubra) from Paeonia lactiflora Pall., Chuan xiong (Chuanxiong Rhizoma) from Ligusticum chuanxiong Hort., Dan shen (Salviae Miltiorrhizae Radix Et Rhizoma) from Salvia miltiorrhiza Bge., and Dang gui (Angelicae Sinensis Radix) from Angelica sinensis (Oliv.) Diels.¹⁰ Previous studies have demonstrated that XBJ possesses antioxidant, anti-inflammatory, and immune-modulating properties.^11–14 However, the bioactive components of XBJ and the mechanisms of its anti-sepsis effects remain to be elucidated or described in a convincing manner.

As a promising approach to understanding multiple-component drug systems such as TCM formulae, network pharmacology has become increasingly important for the development of drugs in recent years.¹⁵ Network pharmacology offers distinct advantages over traditional methods by comprehensively capturing the intricate interactions between biological macromolecules and chemical components, contributing to a more holistic understanding of the targets of interest.^15,16 However, there are also some limitations in network pharmacology-based analysis. For example, the conclusions of network pharmacology analysis may be preliminary due to the lack of reliable experimental validation. On the other hand, it is important to acknowledge that the traditional experimental validation methods are also inadequate for accurately presenting the mechanism by which a drug exerts its function. This is because the results from a single batch or research team are inherently subject to both subjective bias and random variability.

In light of the above considerations, this study utilized a network pharmacology approach, integrating data from extensive previous experimental research on XBJ, to systematically investigate the potential mechanisms underlying its anti-sepsis effects. Ultimately, we constructed a comprehensive regulatory network encompassing herbs, compounds, targets, and signaling pathways. The conclusions of this study are based on the screening and synthesis of hundreds of retrieved published papers on XBJ. Furthermore, we experimentally validated the key regulatory genes of XBJ using a septic mouse model. Therefore, the anti-sepsis mechanism of XBJ revealed in this study is both reliable and representative. This work contributes to a more accurate understanding of the pharmacological actions of XBJ and serves as a valuable reference for future in-depth pharmacological research.

Materials and Methods

Screening of the Compounds and Target Genes Associated with the Compounds in XBJ

The principal active ingredients of XBJ were identified through a review of some published papers,^17–19 and a total of 37 compounds, identified through UPLC-MS/MS, were subsequently utilized in the following analysis. To identify the target genes associated with the active ingredients, the Traditional Chinese Medicine Systems Pharmacology (TCMSP, https://tcmspw.com/tcmsp.php), the Integrative Pharmacology-based Research Platform of Traditional Chinese Medicine (TCMIP, http://www.tcmip.cn/TCMIP/index.php/Home/Login/login.html), and the Comparative Toxicogenomics Database (CTD, https://ctdbase.org/, cutoff values: interactions ≥ 8) were employed. To ascertain the source herbs for the 37 compounds in XBJ, TCMSP, TCMIP and the Natural Product Activity and Species Source Database (NPASS, https://bidd.group/NPASS/index.php) were utilized.

Screening of Sepsis Target Genes

The Therapeutic Target Database (TTD, https://db.idrblab.net/ttd/), the GeneCards database (https://www.genecards.org/), TCMIP, CTD, and TCMSP were utilized to collect disease-associated targets with “sepsis” as the keyword. It should be noted that the genes related to “serine protease unspecific” and “cytokine receptor unspecific” in the research results of the TTD were obtained by searching UniProt (https://www.uniprot.org/) based on the keywords “serine protease” and “cytokine receptor”. Ultimately, a total of 390 genes were identified as sepsis-associated target genes.

Screening and Analysis of RNA-Sequencing Data

To obtain transcriptome sequencing data related to the effect of XBJ on sepsis intervention, searches were conducted in the BioProject (https://www.ncbi.nlm.nih.gov/bioproject/) and GEO (https://www.ncbi.nlm.nih.gov/geo/) databases using the search terms “Xuebijing OR XBJ”. Only one study data set (Accession: PRJNA1012654) was retrieved in BioProject database that was consistent with the search objectives.

The available study utilized the cecal ligation and puncture (CLP) model to induce sepsis in mice, and the heart tissues from XBJ injection-treated and untreated mice were collected for subsequent RNA-sequencing (RNA-seq) at 24h following CLP surgery.²⁰ This RNA-seq data was utilized for further analysis in this study, the R package “DESeq2” was employed for normalization and the identification of differentially expressed genes (DEGs). A false discovery rate (FDR) of less than 0.05 was considered the screening criterion.

Screening of XBJ Injection-Related Experimental Findings

A search of the PubMed database (https://pubmed.ncbi.nlm.nih.gov/) was conducted using the search terms “((Xuebijing) OR (XBJ)) AND ((Sepsis) OR (anti-inflammation) OR (antioxidant))”, and a total of 143 relevant papers were retrieved (retrieved on May 25, 2024). After removing reviews and other irrelevant or non-experimental papers (such as meta-analysis), 27 relevant papers were obtained. The regulatory genes of XBJ in these filtered papers were summarized and subjected to further analysis.

Mice

Female C57BL/6 mice (8–10 weeks old) were obtained from Beijing Vital River Laboratories (specific pathogen-free). All animals were housed at 22 ± 1 °C under a 12-h light/12-h dark cycle. The animal experiment protocols were approved by the Institutional Animal Care and Use Committee (IACUC) of the 927th Hospital of the People’s Liberation Army of China (927-KYLL-2025001), and were conducted in accordance with the Regulations for the Care and Use of Laboratory Animals and the Guideline for Ethical Review of Animal Welfare (China, GB/T 35892-2018).

Sepsis Murine Model

The murine sepsis model was established using CLP according to a previously described method.²¹ To robustly evaluate survival differences, 36 mice subjected to CLP were randomly allocated into three treatment groups (n = 12 per group): a PBS group, a Hydroxysafflor Yellow A group, and an XBJ group. The mice received an intraperitoneal injection of 0.6 mL containing PBS, Hydroxysafflor Yellow A (MeilunBio, 22.5 mg/kg), or XBJ (2304031, Tianjin Hongri Pharmaceutical Co., Ltd., 30mL/Kg) at 0.5 hours before surgery and at 12 hours after surgery, and the number of surviving animals was recorded every 6 hours. During the CLP surgery, all mice received an intraperitoneal injection of sodium pentobarbital solution at a dose of 100 mg/kg body weight.

At 24 hours post-surgery, three surviving mice were randomly selected from each group. Peripheral blood was collected from these mice for the isolation of peripheral blood mononuclear cells (PBMCs). At the experimental endpoint (24 hours after CLP surgery), all of the remaining surviving mice were euthanized by cervical dislocation. The isolated PBMCs were then entrusted to Zhejiang Tianke High-Tech Development Co., Ltd. for transcriptome sequencing. The FPKM values of the interested genes were visualized using a clustering heatmap generated with R software. Plasma samples from mice were used to measure SOD activity and MDA levels, using assay kits purchased from Nanjing Jiancheng Biotechnology Co., Ltd.

Construction of the Protein-Protein and Herb-Ingredient-Target Network

Protein-protein interactions (PPI) were predicted using the STRING database (https://string-db.org/cgi/input.pl), and herb-ingredient and ingredient-target correlation data were obtained based on the results of the aforementioned search. The construction and the hub factor screening of the protein-protein and herb-ingredient-target network were built using Cytoscape 3.8.2, and the hub factors were identified using cytoHubba’s maximal clique centrality (MCC) algorithm.

Molecular Docking

The screened ingredients and the disease-associated target genes were selected for molecular docking. The Protein Data Bank (PDB) database (https://www.rcsb.org/) was utilized for the retrieval of the pertinent protein 3D structures in the form of PDB files. The PubChem database was utilized for the retrieval of SDF files pertaining to the selected components. The AutoDock Vina 1.1.2 software was employed for the subsequent molecular docking, with the resulting binding energies presented as the docking outcomes for the proteins and their corresponding ligands. The 3D structure of the protein-ligand complex was then visualized using the PyMOL 2.5.0 software.

Visualization and Statistical Analysis

As described in our previously published papers,^22,23 additional visualization results appearing in this study were analyzed and plotted using the relevant packages in R 4.1.0 software. These included Gene Ontology (GO) enrichment analysis, Disease Ontology (DO) enrichment analysis, Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis, principal component analysis (PCA), chord diagram, heat map, Venn diagrams and volcano map. Furthermore, bar charts were generated using GraphPad Prism (version 8.4.3), and the Sankey diagram was created with the online tool Flourish (https://app.flourish.studio/projects). The Log rank test and Cox proportional hazards regression model were used to quantitatively evaluate the survival benefit in mice using R. After confirming normality and homogeneity of variance, a two-tailed t-test was used for statistical analysis of the two groups.

Results

Characterization of Compounds and Target Genes Associated with These Compounds in XBJ

Firstly, based on a review of some published papers,^17–19 37 compounds present in XBJ were selected for further investigation. A database search for these compounds yielded 668 potential target genes. The relationships between the herbal constituents, their compounds, and the number of corresponding target genes were visualized in Figure 1. It should be noted that three compounds (Naringenin, Taxifolin, and Matrine) whose source herbs could not be confirmed were excluded from this figure. The number of target genes for each compound, along with its proportion relative to the total, is shown in Figure 2A. The results indicated that Quercetin had the highest number of target genes (326), whereas no target genes were retrieved for Benzoylpaeoniflorin, Danshensu, Hydroxysafflor Yellow A, or Galloylpaeoniflorin.

Figure 1 A visual presentation of the herb-compound-number of target genes for XBJ by the chord diagram.

Figure 2 Characterization of XBJ-regulated genes and compounds in XBJ. (A) The presentation of the number of target genes and the ratio of the gene count to the total number of XBJ-regulated genes. (B) GO enrichment analysis of all the XBJ-regulated genes. (C) DO enrichment analysis of all the XBJ-regulated genes.

Subsequently, we performed GO and Disease Ontology (DO) enrichment analyses on the 668 target genes. The GO analysis revealed significant enrichment in biological processes such as the response to oxidative stress, response to extracellular stimulus, and response to oxygen levels (Figure 2B). The DO analysis indicated association with diseases including chronic obstructive pulmonary disease, ischemia, and urinary system diseases (Figure 2C). These findings are consistent with the reported pharmacological effects of XBJ, such as antioxidant and anti-inflammatory activities. Moreover, the enriched disease terms are strongly linked to the pathogenesis of sepsis; for instance, global tissue hypoxia (a cause of elevated lactate) and acute kidney injury are established clinical hallmarks of sepsis.^24,25

Characterization of Sepsis Target Genes

After analyzing the XBJ-regulated genes, we performed functional enrichment analysis on the 390 collected sepsis-related target genes. GO enrichment analysis revealed significant involvement in biological processes including the cytokine mediated signaling pathway, leukocyte-mediated immunity, and response to chemokines (Figure 3A). Similarly, DO enrichment analysis showed association with diseases such as chronic obstructive pulmonary disease, myocardial infarction, and bacterial infectious diseases (Figure 3B). Further analysis identified 63 overlapping genes between the 390 sepsis-related targets and the XBJ-regulated gene set (Figure 3C). We then predicted PPI for the sepsis-related genes and constructed a network of the top 60 hub genes using Cytoscape (Figure 3D). Notably, 29 of these hub genes (nearly 50%) were among the 63 overlapping genes, suggesting that approximately half of the pivotal regulatory genes in sepsis are also targeted by XBJ.

Figure 3 Characterization of sepsis target genes. (A) GO enrichment analysis of all the sepsis target genes. (B) DO enrichment analysis of all the sepsis target genes. (C) Venn diagram between the XBJ-regulated genes and the sepsis target genes. (D) PPI network among the top 60 hub sepsis genes. The 29 genes filled in blue at the top of the picture are also XBJ-regulated sepsis genes.

Analysis of XBJ-Regulated Genes Based on Publicly Available RNA-Sequencing Data

To elucidate the expression trends of XBJ-regulated genes, we reanalyzed RNA-sequencing data from a published study on septic mice treated with XBJ. The analysis revealed that XBJ treatment resulted in 3016 upregulated and 2152 downregulated DEGs in septic mice (Figure 4A). PCA based on these DEGs showed a clear separation between the XBJ-treated and CLP (untreated) groups, indicating a substantial difference in their global gene expression profiles (Figure 4B).

Figure 4 Characterization of XBJ-regulated genes from available RNA-sequencing data. (A) Number of DEGs between XBJ and CLP groups. (B) Principal component analysis showing that there was an obvious difference in the gene expression pattern between CLP and XBJ group. Ellipses show the distribution of the groups at 68% confidence levels. (C) Expression visualization of the 20 XBJ-regulated sepsis DEGs. The data were normalized for plotting. (D) KEGG enrichment analysis of the down-regulated DEGs in XBJ group. (E) DEGs are represented by volcano plots for XBJ vs. CLP. up: up regulation; down: down regulation; ns, no significance. The expression characteristics of 10 XBJ-regulated hub sepsis DEGs were shown in detail.

Furthermore, we identified 20 DEGs that overlapped with the 63 intersection genes from our prior analysis. A heatmap visualization of these 20 genes showed a downregulation trend in several pro-inflammatory genes, such as Il6 and Cxcl2, in the XBJ-treated group (Figure 4C). Based on this finding, we performed KEGG pathway enrichment analysis on the downregulated DEGs. The results revealed significant enrichment in pathways including cytokine-cytokine receptor interaction, JAK-STAT signaling, and FoxO signaling (Figure 4D), suggesting that XBJ may play an important role in mitigating inflammation and oxidative stress. This conclusion is consistent with the aforementioned findings for the XBJ-regulated gene set.

Further analysis showed that 10 DEGs were among the 29 previously identified XBJ-regulated hub genes for sepsis, most of which (7 out of 10) were downregulated. Volcano plots were used to display the distribution of these hub genes among all expressed genes and to detail their expression patterns (Figure 4E).

Network Pharmacology Analysis of XBJ-Regulated Sepsis Genes

To investigate how the constituents of XBJ regulate sepsis-associated genes, we constructed an herb-compound-target gene network using Cytoscape (Figure 5) and identified the 12 key hub nodes within it. As shown in Figure 6A, these hubs comprise one compound (Quercetin) and 11 genes, which are listed in descending order of their ranking scores: IL6, TNF, IL1B, ICAM1, JUN, ALB, NFKB1, CASP3, NFKBIA, PTGS2, and MAPK3. These results suggest that these 12 hub factors (one compound and 11 genes) are potentially critical to the mechanism by which XBJ alleviate sepsis.

Figure 5 The herb-compound-target sepsis gene network of XBJ.

Figure 6 Comprehensive analysis of the XBJ-regulated sepsis genes. (A) Compound-target gene network of the top 12 hub factors screened from herb-compound-target sepsis gene network of XBJ. (B) Characterization of 17 XBJ-regulated genes or proteins from 27 published studies. up: up regulation; down: down regulation; WB: Western Blotting; IF: Immunofluorescence. (C) Schematic timeline of CLP surgery, drug treatment, survival observation, and sampling in mice. (D) Survival curves of CLP-induced septic mice following treatment with the indicated drugs. Survival differences among groups were assessed by the log‑rank test. Hazard ratios (HR) with 95% confidence intervals (CI) and P values were calculated using Cox proportional hazards regression with the PBS group as the reference. (E) Relative mRNA expression levels of key genes in PBMCs regulated by XBJ. Abbreviation: HSYA, Hydroxysafflor Yellow A.

Summary of XBJ-Regulated Genes from Published Studies

In order to gain a more precise understanding of the key genes involved in the treatment of sepsis with XBJ, we conducted a comprehensive review of the PubMed database for studies reporting XBJ-regulated genes, using the keywords “sepsis”, “anti-inflammatory” and “antioxidant”. To ensure the accuracy of our conclusions, we applied strict inclusion criteria: genes reported in only a single paper without validation by multiple detection methods (excluding sequencing) were omitted, as were genes with inconsistent expression trends across different studies. For example, IL10 exhibited contrasting expression patterns in the literature.^26–30 Following this screening, we identified a final set of 17 XBJ-regulated genes/proteins from 27 publications.^17,26–51

As shown in Figure 6B, IL6, TNF, and IL1B were the most frequently reported genes, and their downregulation by XBJ was consistently demonstrated by methods such as ELISA and PCR. Notably, these three genes also ranked as the top hub factors in our network pharmacology analysis of XBJ-regulated sepsis genes and compounds (Figure 6A). Moreover, the literature consistently indicates that XBJ-regulated genes are predominantly inflammation-related and generally exhibit a downregulation trend.

XBJ Improved the Survival Rate of CLP Mice

Based on the above analyses, we were encouraged to further evaluate the genuine protective effects of XBJ using the CLP mouse model, which is well-established for its ability to closely mimic human sepsis. Furthermore, since Hydroxysafflor Yellow A is the second most abundant component in XBJ, only after paeoniflorin,^18,19 and yet functional studies on this compound remain scarce—with its regulatory genes even failing to be identified in current databases—we concurrently assessed the protective effects of both XBJ and Hydroxysafflor Yellow A in CLP mice (Figure 6C). As shown in Figure 6D, although the improvement in survival rate of mice treated with XBJ and Hydroxysafflor Yellow A did not reach statistical significance compared to the PBS group (possibly due to the limited sample size or insufficient observation duration), the XBJ group exhibited a clear trend toward prolonged survival (HR = 0.361, P = 0.097) and demonstrated the strongest protective effect, with a survival rate at the endpoint twice that of the PBS group.

To further verify the accuracy of the previous analysis—which predicted expression trends of XBJ-regulated genes via network pharmacology and literature review—and to explore the potential role of Hydroxysafflor Yellow A in sepsis intervention, we also performed transcriptome sequencing on peripheral blood mononuclear cells (PBMCs) from CLP mice at 24-hour post-treatment. Key sepsis-related genes identified as being regulated by XBJ were visualized using a heatmap (Figure 6E). The results indicated that the regulatory effects of XBJ on gene expression observed in this study were largely consistent with the trends predicted previously, showing a concordance rate of nearly 80%, with only a few exceptions such as the Hspa5 and Casp9 genes. This confirms that the integrated approach of network pharmacology and literature data aggregation possesses robust reliability for identifying XBJ-regulated genes and predicting their expression trends.

Additionally, it is worth noting that the results in Figure 6E suggest a relatively high consistency between XBJ and Hydroxysafflor Yellow A in regulating sepsis-related genes. However, the suppressive effect of Hydroxysafflor Yellow A on the expression of genes such as Cxcl2, Tlr4, and Hmgb1 was weaker than that of XBJ.

Binding Interaction Between the Screened Ingredients and the Target Genes

Given that IL6, TNF, and IL1B are key target genes regulated by XBJ, and Quercetin is predicted as a hub compound in the XBJ-regulated sepsis network, we performed molecular docking to evaluate its binding affinity with these proteins. The results demonstrated robust binding interactions between Quercetin and all three targets (Figure 7A–C). Furthermore, among the 12 hub factors identified in our network pharmacology analysis, CASP3, MAPK3, and JUN are known to regulate the expression or activation of multiple downstream genes. Therefore, we also assessed the binding of Quercetin to these proteins. Similarly, strong binding interactions were observed (Figure 7D–F). In addition, we evaluated the binding of Hydroxysafflor Yellow A to IL-6, TNF, and IL-1B to explore its potential role, even though it did not suppress IL-1B expression in our previous experiments. The results indicated strong binding interactions with all three proteins (Figure 7G–I), suggesting that they may still be potential regulatory targets of Hydroxysafflor Yellow A.

Figure 7 Molecular docking of candidate compounds with key target proteins. (A–F) Molecular docking diagram of Quercetin with IL6 (A), TNF (B), IL1B (C), CASP3 (D), MAPK3 (E), or JUN (F). Quercetin is marked in red and the amino acids interacting with Quercetin are marked in purple. (G–I) Molecular docking diagram of Hydroxysafflor Yellow A with IL6 (G), TNF (H), or IL1B (I). Hydroxysafflor Yellow A is marked in red and the amino acids interacting with Hydroxysafflor Yellow A are marked in purple. For reference, known inhibitors of TNF (ZINC09609430) and CASP3 (MSI) demonstrate binding energies of −9.2 kcal/mol and −8.21 kcal/mol, respectively.52,53

Anti-Inflammatory and Antioxidant Properties Mediate the Therapeutic Effects of XBJ Against Sepsis

To systematically elucidate the target genes and signaling pathways regulated by XBJ, we compiled a gene set, the complete list of which is visually presented in the heatmap of Figure 6E. This set was composed of two subsets: 1) XBJ-regulated genes collected from the literature (gene names are shown in Figure 6B), and 2) the intersection between RNA-sequencing DEGs and the 29 sepsis-related hub genes, which yielded 10 genes (gene names are shown in Figure 4E). We employed this integrated gene set for GO and KEGG enrichment analysis, which revealed significant enrichment in pathways related to immune responses and oxidative stress. Since the above-listed genes do not directly demonstrate the antioxidant effects of XBJ, we further measured the plasma SOD activity and MDA levels in XBJ-treated septic mice. The significantly higher SOD activity (Figure 8A) and lower MDA levels (Figure 8B) observed in the XBJ-treated group confirmed the antioxidant capacity of XBJ.

Figure 8 Analysis of the antioxidant capacity of XBJ and construction of the XBJ regulatory network. (A) Detection of plasma SOD activity in septic mice after intervention with PBS or XBJ. (B) Detection of plasma MDA levels in septic mice after intervention with PBS or XBJ. (C) Herb-compound-target gene-signaling pathway network of XBJ. XXX: unspecified herbs (the originating herbs of Taxifolin, Matrine, and Naringenin are denoted as XXX for visual representation). *p < 0.05; **p < 0.01.

To illustrate the relationships among the herbal constituents of XBJ, their compounds, target genes, and signaling pathways, we constructed a Sankey diagram (Figure 8C). Of note, the source herbs for three compounds (Taxifolin, Matrine, and Naringenin) could not be definitively determined; for display purposes, these are labeled as “XXX”. Furthermore, only genes present in at least four pathways are displayed. The results suggest that at least 18 compounds in XBJ possess potential anti-inflammatory and antioxidant activities, and the collective action of these multi-component agents may underpin the therapeutic benefits of XBJ in sepsis.

Discussion

This study sought to elucidate the mechanism underlying the anti-sepsis effects of XBJ. To this end, we integrated network pharmacology with transcriptome sequencing data and a comprehensive collection of experimental evidence from the published literature. Our analysis identified that XBJ, via at least 18 of its constituent compounds, modulates 21 genes (11 of the most critical genes are depicted in Figure 8C) to exert anti-inflammatory and antioxidant effects against sepsis. Notably, the majority of these target genes, including well-established sepsis-associated factors such as IL1B, TNF, IL6, and HMGB1,^54,55 were downregulated.

While some previous studies have employed network pharmacology methods to investigate the anti-sepsis properties of XBJ,^17,32 this study differs in two key aspects. Firstly, the XBJ-derived compounds included in our analysis were sourced exclusively from published studies, all of which were identified by UPLC-MS/MS,^17–19 thus ensuring the reliability of the data. Secondly, the sepsis-related target genes regulated by XBJ were identified through an extensive review of published literature, providing a comprehensive understanding of its mechanisms of action. Subsequently, a regulatory network of herbs-compounds-target genes-signaling pathways was constructed using network pharmacology methods. Unlike approaches that rely entirely on public databases, our strategy of using carefully curated experimental data enhanced the reliability and accuracy of the analytical results. Furthermore, the anti-sepsis effects of XBJ were validated in a CLP mouse model of sepsis. Although the CLP model does not fully represent human sepsis and there is a possibility of inconsistency or off-target effects when translating findings to humans, our results showed that XBJ significantly improved survival, with the survival rate in the XBJ-treated group being twice that of the PBS-treated group at the observation endpoint.

Nevertheless, this study has several limitations. Specifically, the effects of XBJ on immune cells were not investigated. Although previous studies have suggested that XBJ administration reduces neutrophil levels and inhibits M1 polarization of macrophages,^56,57 it has also been reported to facilitate the expansion of Tregs and normalize the pro-inflammatory Th17 population, thereby inhibiting inflammation and regulating the balance between Tregs and Th17 cells.⁴⁸ Further identification of the specific cell types regulated by the active compounds in XBJ, as well as the cellular sources of the genes regulated by XBJ, remains to be explored in future research. Additionally, we were unable to trace three compounds (Naringenin, Taxifolin, and Matrine) back to their specific herbal origins within XBJ due to limited information in the available databases. Furthermore, the target genes for four other compounds (Benzoylpaeoniflorin, Danshensu, Hydroxysafflor Yellow A, and Galloylpaeoniflorin) could not be confidently identified through database mining.

On a positive note, we confirmed the protective effect of Hydroxysafflor Yellow A (administered at a dose equivalent to its mass content in XBJ) using a CLP-induced murine sepsis model. Notably, we found that this compound modulated key sepsis-related genes in a manner similar to, albeit slightly weaker than, the full XBJ formulation (Figure 6E). Specifically, compared to XBJ, Hydroxysafflor Yellow A was less effective in suppressing key sepsis-related genes such as Cxcl2, Tlr4, and Hmgb1. In subsequent studies, we will further investigate the potential synergistic effects for sepsis treatment between this compound and other bioactive constituents identified in our network pharmacology analysis, such as Quercetin, Kaempferol, and Luteolin. Such research will be crucial for developing new anti-sepsis drugs or optimizing the XBJ formulation itself.

Conclusion

In summary, this study, employing network pharmacology and curated target genes from the literature, reveals that XBJ may exerts its anti-sepsis effects by regulating target genes via its constituent compounds (at least 18). The majority of these genes, including TNF and IL6, were down-regulated following XBJ treatment, primarily influencing anti-inflammatory and antioxidant pathways. Furthermore, the exploration of the anti-sepsis function of Hydroxysafflor Yellow A in this study provides crucial reference data and identifies directions for future research aimed at the optimization and enhancement of the XBJ formula. Additionally, the research strategy established here provides a valuable framework for elucidating the mechanisms of other TCM formulations or complex drug systems.