Skip to main content
NCBI home page
BMC Pharmacology & Toxicology logoLink to BMC Pharmacology & Toxicology
. 2026 Jul 6;27:137. doi: 10.1186/s40360-026-01170-8

Prioritization of molecular signatures between BDE-209-relevant targets and ulcerative colitis: a network toxicology and bioinformatics analysis

Yue Zhang 1, Hai Zhang 1, Danli Hu 2, Hailong Zhang 3,4, Da Wang 1, Ziyang Hong 1, Si Su 1,✉, Ling Chen 1,✉
PMCID: PMC13628797  PMID: 42410645

Abstract

Background

Decabromodiphenyl ether (BDE-209) is a widely used flame retardant and persistent environmental contaminant. However, the overlap between predicted BDE-209-related targets and ulcerative colitis (UC)-associated molecular signatures has not been systematically evaluated.

Methods

Public UC transcriptomic datasets and predicted BDE-209-related targets were integrated for comparative bioinformatic analysis. Differential expression analysis, enrichment analysis, machine learning algorithms, protein–protein interaction network analysis, immune infiltration analysis, single-cell RNA sequencing analysis, and molecular docking were performed to prioritize candidate genes associated with BDE-209-related targets and UC-associated signatures.

Results

A total of 87 overlapping genes were identified between predicted BDE-209-related targets and UC-associated differentially expressed genes. Functional enrichment analysis showed that these genes were mainly enriched in inflammation- and immune-related pathways. Machine learning analysis and protein–protein interaction network analysis prioritized eight candidate hub genes (IL1B, MMP9, ICAM1, PPARG, VCAM1, AREG, FABP3, and ACSL5). Immune infiltration and single-cell transcriptomic analyses showed that these genes were mainly expressed in immune-related cell populations within UC datasets. Molecular docking analysis suggested potential structural compatibility between BDE-209 and the protein structures encoded by the identified hub genes.

Conclusions

This study identified shared molecular features between predicted BDE-209-related targets and UC-associated transcriptomic signatures. These findings provide a hypothesis-generating resource for future studies investigating the potential relevance of environmental contaminant-related target networks in UC-associated inflammatory contexts.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s40360-026-01170-8.

Keywords: BDE-209, Ulcerative colitis, Network toxicology, Inflammation signaling

Introduction

The commercial mixture decabromodiphenyl ether (c-decaBDE) is a mixture of polybrominated diphenyl ethers (PBDEs) used as a flame retardant, with BDE-209 as its main component [1]. It is widely used as an additive flame retardant in plastics, textiles, and electronic equipment [2]. Due to its strong hydrophobicity and lipophilicity, BDE-209 is highly resistant to degradation in the environment, making it the most prevalent PBDE, accounting for approximately 93% of total PBDEs [3–5]. Studies have shown that food is the primary route of human exposure to BDE-209, particularly high-fat foods such as vegetable oils, meat products, seafood, etc [6]. BDE-209 has been detected in human blood, adipose tissue, and breast milk [7]. Due to its toxic effects on human health and environmental persistence, BDE-209 has been widely recognized as a hazardous pollutant. Consequently, it has been classified as a persistent organic pollutant (POP) under the Stockholm Convention supervised by the United Nations Environment Programme [8].

According to the literature, BDE-209 can exert toxic effects on various biological systems through epigenetic regulation and the induction of oxidative stress [1]. Studies have shown that BDE-209 exhibits neurotoxicity, endocrine-disrupting effects, immunotoxicity, reproductive toxicity, and carcinogenic potential [9–13]. In the digestive system, the toxic effects of BDE-209 are primarily manifested as liver and intestinal injury. Studies indicate that BDE-209 may contribute to hepatotoxicity by disrupting normal liver metabolic functions through oxidative stress and inflammation, ultimately leading to liver dysfunction [14–16]. The intestine is a vital digestive organ, and it is often the first to be damaged when toxins pass through it. Current toxicological studies have shown that BDE-209 is associated with intestinal oxidative stress and inflammatory responses [17]. Multiple inflammatory signaling pathways may be involved in these processes, which are potentially linked to intestinal injury. Studies on zebrafish exposed to varying doses of BDE-209 showed that high exposure to BDE-209 disrupted intestinal barrier function, leading to increased oxidative stress and inflammatory responses in the gut [18]. Animal experiments have shown that BDE-209 is involved in structural and functional damage to the intestine by mediating oxidative stress and impairing the antioxidant defense system [19].

Ulcerative colitis (UC) is a chronic, relapsing, and nonspecific form of inflammatory bowel disease (IBD). An increasing number of studies suggest that environmental factors play an important role in the onset of UC [20]. Chronic inflammatory injury of the intestinal mucosa is considered a major feature of UC pathogenesis and disease progression. Studies have demonstrated that BDE-209 contributes to intestinal barrier disruption and necroptosis, as well as gut microbiota dysbiosis, all of which are critical processes in the initiation and progression of UC [21, 22]. Both BDE-209-associated intestinal injury and UC pathogenesis involve inflammatory responses, oxidative stress, immune-cell infiltration, and gut microbiota dysbiosis [23–25]. Several signaling pathways, including the NF-κB, TNF, and Toll-like receptor signaling pathways, have been implicated in these processes. Previous studies have suggested that UC and BDE-209-associated toxicological processes involve partially overlapping inflammatory and immune-related pathways. However, the relationship between predicted BDE-209-associated target profiles and UC-associated molecular signatures has not yet been systematically investigated. Therefore, we explored whether predicted BDE-209-associated targets share molecular features with UC-related transcriptomic signatures.

In our study, we integrated network toxicology with machine learning algorithms, immune infiltration analysis, single-cell RNA sequencing analysis, and molecular docking for comprehensive bioinformatic analysis. These approaches were used to prioritize overlapping genes and investigate their expression features in UC-related datasets. This study aimed to identify candidate molecular signatures shared between predicted BDE-209-associated targets and UC-related inflammatory transcriptomic profiles. The findings may provide a hypothesis-generating basis for future experimental studies. The procedure of our study is illustrated in Fig. 1.

Fig. 1.

Fig. 1

Open in a new tab

The flowchart illustrating the research design of this study

Methods

Management of the BDE-209 target

The potential target of BDE-209 associated with Homo sapiens comes from the Comparative Toxicology Genomics database (CTD) (https://ctdbase.org/) [26], PubChem (https://pubchem.ncbi.nlm.nih.gov/) database [27], and the SwissTargetPrediction (https://www.expasy.org/resources/swisstargetprediction) tool [28]. The candidate targets obtained from each database are presented in the Supplementary Tables 1–3 (Excel file). The integrated target set was subsequently used as the candidate BDE-209-related target library for downstream analyses.

Screen potential UC-related targets

To identify overlapping molecular signatures between predicted BDE-209-related targets and UC-associated transcriptomic profiles, the GSE87466 and GSE75214 datasets were obtained from the Gene Expression Omnibus (GEO) database (http://www.ncbi.nlm.nih.gov/geo/). The datasets contain mucosal biopsy samples from 184 patients with UC and 32 healthy controls, totaling 216 samples [29, 30]. All data analyses were performed using R version 4.4.0. The normalizeBetweenArrays function from the limma package (version 3.60.6) was used for data standardization. ComBat function from the sva package (version 3.52.0) was applied to remove batch effects between the two datasets. Efficacy of batch effect removal was evaluated by principal component analysis (PCA). The quality assessment of data standardization and batch-effect correction is presented in Fig. S1.

Differential gene screening

Differential expression analysis was conducted using the limma package [31]. P values were adjusted for multiple testing using the Benjamini–Hochberg method to control the false discovery rate (FDR). Genes with an adjusted P value (FDR) < 0.05 and |log2FC| > 0.5 were considered differentially expressed genes (DEGs). Volcano plots were generated to visualize DEG distributions. PCA was conducted to assess the overall separation of gene expression profiles between UC and normal samples.

Pathway enrichment analysis of gene targets

To characterize the biological functions and pathways associated with the overlapping gene set, the intersection between candidate BDE-209-related targets and UC-related DEGs was identified. The clusterProfiler package (version 4.12.6) and org.Hs.eg.db package (version 3.19.1) were employed to perform Gene Ontology (GO) functional annotation and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis on the intersecting genes [32]. GO enrichment analysis included the Biological Process (BP), Cellular Component (CC), and Molecular Function (MF) categories. P values were adjusted using the Benjamini–Hochberg method, and pathways with FDR < 0.05 were considered significantly enriched [33].

Select core targets through machine learning

Two machine learning algorithms, Least Absolute Shrinkage and Selection Operator (LASSO) regression and Random Forest (RF), were applied for feature prioritization of candidate genes associated with overlapping BDE-209-related and UC-associated signatures. For LASSO regression, the cv.glmnet function from the glmnet package (version 4.1.10) was used for cross-validation with nfolds = 10, and the glmnet function was employed to train the model with lambda.1se selected as the optimal lambda parameter, resulting in the prioritization of 9 key genes [34]. For Random Forest, the randomForest function from the randomForest package (version 4.7.1.2) was used to train the model with ntree = 100. Candidate genes were ranked according to RF-derived variable importance scores, and the top-ranked genes were retained for subsequent hub-gene prioritization.

Construction of protein-protein interaction networks and screening of hub genes

To investigate interactions among the overlapping candidate genes, a protein–protein interaction (PPI) network was constructed using the STRING database (https://string-db.org/) [35], with the organism restricted to Homo sapiens. The interaction data from STRING were imported into Cytoscape (Version 3.10.2) for network visualization and topological analysis. Cytoscape was further used to construct the PPI network graph and evaluate the properties of network nodes and edges [36]. The cytoHubba plugin was used to rank candidate genes based on MCC (Maximum Clique Centrality) for hub-gene prioritization. To improve the transparency of hub-gene prioritization, Supplementary Table 4 (Word file) summarizes the evidence supporting the eight prioritized hub genes. Chemical evidence grading was assigned based on the availability of prior evidence linking a gene to BDE-209 exposure. Genes with direct experimental evidence of regulation by BDE-209 reported in the CTD and published studies were classified as high-evidence genes. Genes identified only through computational prediction or indirect inference without direct experimental validation were classified as low-evidence genes. In addition, the same analytical workflow was applied to an unrelated toxicant (PFOS) for comparative analysis to assess potential overlap in prioritized inflammatory signatures [37].

Internal and external validation of hub genes

To further evaluate the stability of the identified hub genes, differential expression analysis was performed independently in each dataset, and the results were intersected with the 8 hub genes of interest. An independent UC dataset, GSE179285, including 31 healthy control samples and 23 UC samples, was used for external validation [38]. Expression analysis of the 8 hub genes was conducted, and statistical significance was assessed using the Wilcoxon rank-sum test. The ggpubr package (version 0.6.2) was used for visualization of differential expression results (Fig. S2). Receiver operating characteristic (ROC) curve analysis was subsequently performed to evaluate the discriminative performance of the identified hub genes, and the area under the curve (AUC) was calculated.

Analysis of immune cell infiltration

UC is characterized by complex inflammatory and immune-related alterations [39]. To characterize immune-cell composition differences between UC and control samples, the CIBERSORT algorithm was used to calculate the estimated proportions of 22 immune-cell types [40]. The IOBR package (version 0.99.0) and CIBERSORT package (version 0.1.0) were employed for this analysis, with the deconvolution algorithm performed using LM22.txt as the reference signature matrix (perm = 1000) [41]. The estimated proportions of immune cell types in each sample are presented in Supplementary Table 5. In addition, the differences in immune cell infiltration among the expression groups of eight core genes (IL1B, MMP9, ICAM1, PPARG, VCAM1, AREG, FABP3, and ACSL5) were compared. To assess the robustness of these findings, we used the EPIC algorithm as a robustness check. The EPIC package (version 1.1.7) was employed to calculate the infiltration of 8 immune cell types, and correlation analysis between hub gene expression and immune cell infiltration levels was performed (Fig. S3).

ScRNA-seq data processing, clustering, and cell type labeling

The scRNA-seq dataset GSE231993 was obtained from a public database, comprising colon biopsies from 4 healthy controls and 4 UC patients [42]. Quality control filtering was conducted based on nCount, nFeature, and mitochondrial gene percentage (percent.mt), and cells with nCount > 40,000, nFeature > 6,000, or mitochondrial gene proportion > 25% were excluded. The filtered data were normalized, and PCA was performed with 20 principal components (PCs). The first 15 PCs were subsequently used for clustering and UMAP visualization. Clustering was conducted at a resolution of 0.5, followed by dimensionality reduction and visualization using uniform manifold approximation and projection (UMAP) analysis. Cell populations were annotated based on specific markers from the original study, including B cells, T cells, epithelial cells, smooth muscle cells, macrophages, endothelial cells, mast cells, and fibroblasts [43]. Batch-effect correction across samples was performed using the RunHarmony function. To avoid inter-sample heterogeneity, cells derived from UC patients and paired non-inflamed tissues from the same patients were selected for differential expression analysis in Supplementary Table 6 (Excel file). The FindMarkers() function was applied using the Wilcoxon rank-sum test with filtering criteria of |avg_log2FC| > 0.2 and raw P value < 0.05. Differential-expression analyses were performed for exploratory cellular-context characterization and were not intended as FDR-controlled gene-discovery analyses.

Molecular docking

The two-dimensional structure of the environmental pollutant BDE-209 was obtained from the PubChem database (https://pubchem.ncbi.nlm.nih.gov/; PubChem CID: 14410). Crystal structures of proteins encoded by the identified hub genes were retrieved from the Protein Data Bank (PDB)(https://www.rcsb.org/), according to the following criteria: (1) organism: Homo sapiens; (2) experimental method: X-ray diffraction; (3) resolution < 2.5 Å; and (4) availability of co-crystallized ligands. The positions of co-crystallized ligands were used to define the docking pockets. Molecular docking analysis was performed using the CB-Dock2 platform (https://cadd.labshare.cn/cb-dock2/) [44, 45]. Docking scores were used as approximate indicators of potential structural compatibility between BDE-209 and the selected protein structures for exploratory analysis, with lower docking scores indicating relatively more favorable predicted interactions. A summary of the analytical methods, software packages, and key implementation parameters used throughout the study is provided in Table 1.

Table 1.

Summary of analytical methods, software packages, and key parameters

Analysis type Software / packages (Version) Key parameters / implementation
DEG analysis limma (v3.60.6) adj.P.Val < 0.05, |log2FC| > 0.5
Enrichment analysis clusterProfiler (v4.12.6), org.Hs.eg.db (v3.19.1) enrichGO(ont="ALL”, pAdjustMethod="fdr”, pvalueCutoff = 0.05); enrichKEGG(organism=’hsa’, pvalueCutoff = 0.05)
Machine learning glmnet (v4.1.10), randomForest (v4.7.1.2) LASSO: cv.glmnet(family="binomial”, alpha = 1, nfolds = 10), λ: lambda.1se; RF: randomForest(ntree = 100)
PPI & Network STRING Database, Cytoscape (v3.10.2), cytoHubba Organism: Homo sapiens; Minimum required interaction score: 0.4; Hub Selection: Top 8 via MCC algorithm; Other Settings: Default.
Immune deconvolution CIBERSORT (v0.1.0), IOBR (v0.99.0), EPIC (v1.1.7) deconvo_tme(method="cibersort”, arrays = T, perm = 1000); EPIC(bulk=expr_mat, reference="TRef”)
scRNA-seq Seurat (v5.4.0) QC: percent.mt < 25, nFeature_RNA < 6000; Cluster: dims = 1:15, resolution = 0.5; FindMarkers(test.use="wilcox”)
Molecular docking CB-Dock2 (Website) Default pipeline via cadd.labshare.cn/cb-dock2/
Open in a new tab

Results

Screening of potential targets of BDE-209 associated with ulcerative colitis

In this study, candidate BDE-209-related targets were collected from the CTD, PubChem, and SwissTargetPrediction databases using the keyword “Decabromodiphenyl ether.” After merging and deduplication, a total of 417 candidate targets were obtained. The GSE87466 and GSE75214 datasets were downloaded from the GEO database. These datasets contained UC-related gene expression profiles generated using the GPL13158 and GPL6244 platforms, respectively. PCA was performed to visualize overall gene expression patterns between normal colonic tissues and UC samples. The first principal component (PC1) and second principal component (PC2) accounted for 17.3% and 10.6% of the total variance, respectively, suggesting partial transcriptomic separation between the two groups (Fig. 2A). Differential expression analysis identified 2,493 UC-related DEGs, including 1,058 downregulated genes and 1,435 upregulated genes (Fig. 2B). Intersection analysis between the candidate BDE-209-related targets and UC-related DEGs identified 87 overlapping genes for subsequent analyses (Fig. 2C).

Fig. 2.

Fig. 2

Open in a new tab

Identification of overlapping genes between predicted BDE-209-associated targets and UC-related DEGs. (A) PCA showing the distribution of gene expression profiles between UC and normal colonic tissues. (B) Volcano plot of DEGs identified from the GSE75214 and GSE87466 datasets. (C) Venn diagram showing the overlap between predicted BDE-209-associated targets and UC-related DEGs

Target analysis and pathway enrichment analysis

To further characterize the biological functions and pathways associated with the overlapping gene set, GO and KEGG enrichment analyses were performed using R software. The results are presented in the supplementary Tables 7, 8 (Excel file). During GO enrichment analysis, one gene (MUC5B) could not be mapped to the background universe used by the GO function, despite having a valid ENTREZ ID. Therefore, 86 genes were ultimately included in the GO enrichment output. GO enrichment analysis (Fig. 3A) showed that the candidate genes were mainly enriched in MF terms including DNA-binding transcription factor binding, cytokine receptor binding, and cytokine activity. In the CC category, enrichment was observed in the external side of the plasma membrane and apical part of cell. In the BP category, enriched terms included regulation of inflammatory response and response to molecule of bacterial origin. KEGG pathway analysis demonstrated enrichment in pathways including the Toll-like receptor signaling pathway, IL-17 signaling pathway, NF-κB signaling pathway, TNF signaling pathway, and Th17 cell differentiation (Fig. 3B). Overall, both GO and KEGG enrichment analyses consistently suggested that the overlapping gene set was primarily associated with inflammatory and immune-related signaling pathways relevant to UC-associated inflammatory signatures.

Fig. 3.

Fig. 3

Open in a new tab

GO and KEGG enrichment analyses of overlapping genes. (A) Bar plot showing the top five enriched GO terms in the BP, CC, and MF categories. The bar length represents the number of enriched genes, and enrichment significance was evaluated using adjusted P values presented as -log10(FDR). (B) KEGG pathway enrichment analysis showing the top 15 enriched pathways. Enrichment significance was evaluated using adjusted P values presented as -log10(FDR)

Machine learning analysis and acquisition of Hub genes

To further refine the overlapping gene set, LASSO regression and RF analyses were applied for feature prioritization of candidate genes associated with overlapping BDE-209-related and UC-associated signatures. The genes selected from machine learning are listed in Supplementary Tables 9 and 10 (Excel file). LASSO regression analysis prioritized 9 genes from the 87 overlapping genes (Fig. 4A, B). RF analysis results are shown in Fig. 4C-E, and genes were ranked according to RF-derived importance scores. The top 15-ranked genes were retained for subsequent analyses. A total of 19 prioritized genes were identified by integrating the results from the two machine-learning approaches (Fig. 5A). The heatmap demonstrated distinct expression patterns of the 19 prioritized genes between UC and control groups (Fig. 5B). Several inflammation-related genes, including IL1B, ICAM1, VCAM1, and MMP9, were upregulated in UC samples, whereas several metabolism-related genes, including PPARG, showed relatively lower expression. These prioritized genes were subsequently imported into the STRING database for protein–protein interaction analysis (Fig. 5C). Using the CytoHubba plugin in Cytoscape, eight prioritized hub genes were identified: IL1B, MMP9, ICAM1, PPARG, VCAM1, AREG, FABP3, and ACSL5 (Fig. 5D). To evaluate consistency across datasets, differential expression analysis was independently performed in each dataset, and overlap analysis was conducted with the eight hub genes. The UpSet plot demonstrated that all eight hub genes were consistently identified in both the GSE75214 and GSE87466 datasets (Fig. S4). To further assess whether the screening workflow preferentially captures broadly shared inflammatory genes, the same analytical pipeline was applied to an unrelated toxicant (PFOS) as a comparative control. Partial overlap was observed between the PFOS-derived genes and the BDE-209-related hub genes, with IL1B, MMP9, and PPARG identified in both analyses (Supplementary Table 11). Other hub genes identified in the BDE-209-related analysis were not observed in the PFOS comparison. External validation using the independent GSE179285 dataset (31 controls and 23 UC samples) showed that hub genes remained significantly differentially expressed between UC and control samples (Fig. S2). ROC curve analysis further demonstrated relatively good discriminatory performance of several hub genes between UC and control samples in the external validation cohort (Fig. 5E). These findings support the reproducibility of the prioritization strategy across independent UC datasets. To improve the transparency of the hub-gene screening strategy, a stepwise prioritization workflow illustrating how the candidate genes were reduced from 87 intersecting genes to the final eight hub genes is presented in Fig. 5F.

Fig. 4.

Fig. 4

Open in a new tab

Machine learning–based prioritization of overlapping genes. (A) LASSO coefficient profiles of overlapping genes. (B) Cross-validation plot for optimal lambda selection in the LASSO model. (C) Random forest error rate plot. (D) MeanDecreaseGini distribution of candidate genes in the RF model. (E) Feature importance ranking of the top 15 genes identified by RF analysis

Fig. 5.

Fig. 5

Open in a new tab

Identification and prioritization of hub genes. (A) Venn diagram showing the overlapping genes identified by LASSO regression and RF analyses. (B) Heatmap showing the expression patterns of the 19 prioritized genes between UC and normal control samples. (C) PPI network of the 19 prioritized genes. (D) Hub-gene network constructed from the top-ranked genes identified by topological analysis of the PPI network. (E) ROC curves evaluating the discriminatory performance of prioritized hub genes between UC and control samples. (F) Stepwise hub-gene prioritization workflow. The workflow illustrates the sequential filtering and prioritization process used to identify the final hub genes from the initial intersecting gene set

Analysis of immune cell infiltration

To further characterize the immune microenvironment in UC, immune infiltration analysis was performed using the CIBERSORT algorithm in the GSE87466 and GSE75214 datasets. Potential associations between hub genes and immune-cell populations were subsequently evaluated. The stacked bar plot showed altered proportions of 22 immune cell subsets across UC and normal colon tissue samples, including macrophage subtypes (M0, M1, and M2 macrophages) (Fig. 6A). These findings suggest a potential association between macrophage-related innate immune responses and the UC inflammatory microenvironment. In addition, significant differences in the relative proportions of several immune-cell populations were observed between the control and disease groups. These included M0, M1, and M2 macrophages, plasma cells, and neutrophils (Fig. 6B). Correlation analysis further revealed distinct relationships among immune-cell subsets within the UC immune microenvironment (Fig. 6C). Finally, correlation analysis between hub genes and immune-cell infiltration levels showed that VCAM1, MMP9, IL1B, and ICAM1 were correlated with multiple infiltrating immune-cell populations (Fig. 6D). These findings suggest potential relationships between these hub genes and immune-related features in UC tissues.

Fig. 6.

Fig. 6

Open in a new tab

Immune infiltration analysis in UC and normal colon tissues. (A) Stacked bar plot showing the relative proportions of 22 immune cell types estimated by the CIBERSORT algorithm in UC and normal colon tissues from the GSE87466 (UC n = 87; normal n = 21) and GSE75214 (UC n = 97; normal n = 11) datasets. Each column represents an individual sample, and each color represents an immune cell subtype. (B) Comparison of inferred immune cell infiltration levels between UC and normal groups. (C) Spearman correlation heatmap of the 22 immune cell subsets. (D) Correlation analysis between eight hub genes and immune cell infiltration levels was calculated using Spearman correlation. Correlation coefficients range from − 1 to 1, representing negative to positive associations

ScRNA-seq data processing and analysis

Single-cell RNA sequencing (scRNA-seq) provides cell-specific transcriptomic information for exploring the cellular landscape of UC tissues. scRNA-seq data were obtained from colon tissues of healthy individuals and UC patients from the GEO database. After quality control, a total of 47,787 high-quality cells were retained, including 17,148 cells from healthy controls, 15,329 cells from inflamed UC tissues, and 15,400 cells from non-inflamed tissues of UC patients. Non-inflamed tissues from UC patients refer to colonic regions without macroscopic or histological signs of inflammation sampled from the same patients as the inflamed tissues. After quality control and clustering analysis, a total of 18 cell clusters were identified using the Seurat workflow and visualized by UMAP analysis (Fig. 7A). Based on canonical marker-gene expression patterns, these clusters were annotated into major cell types, including B cells, T cells, epithelial cells, smooth muscle cells, macrophages, endothelial cells, mast cells, and fibroblasts (Fig. 7B). The expression patterns of representative marker genes were further examined. Canonical markers were predominantly enriched in their corresponding cell populations, supporting the reliability of the cell-type annotations (Fig. 7C). Further analysis of differential expression patterns between UC and control groups demonstrated that several prioritized hub genes exhibited distinct cell-type-associated expression patterns across major immune and stromal cell populations (Fig. 7D). In particular, MMP9 and IL1B were relatively increased in macrophages and endothelial cells in UC samples, whereas PPARG displayed comparatively lower expression levels in several inflammatory-associated cell populations. These findings provide supportive evidence regarding the cellular distribution of hub genes and inflammatory-related transcriptional changes in UC tissues.

Fig. 7.

Fig. 7

Open in a new tab

Single-cell transcriptomic characterization. (A) UMAP visualization showing unsupervised clustering of single cells. Each dot represents an individual cell, and different colors indicate distinct Seurat clusters. A total of 18 clusters were identified. (B) Cell-type annotation of the identified clusters based on canonical marker gene expression patterns. (C) Expression patterns of representative marker genes across annotated cell types. Dot size represents the proportion of expressing cells, and color intensity indicates the average expression level. (D) Distribution of differentially expressed genes between UC and control groups across major cell types. The y-axis represents log2 fold change (UC vs. control), and hub genes are highlighted

Molecular docking results

To further explore the potential structural compatibility between BDE-209 and the prioritized hub-gene-related proteins, molecular docking analysis was performed using the CB-Dock2 platform. Predicted docking conformations and putative binding cavities of BDE-209 with ACSL5, AREG, FABP3, ICAM1, IL1B, MMP9, PPARG, and VCAM1 are shown in Fig. 8. The docking results suggested that BDE-209 could be accommodated within the predicted binding pockets of multiple hub-gene-related proteins, with vina scores ranging from − 5.2 to -7.8 kcal/mol. Among these proteins, FABP3 exhibited the lowest docking score, followed by MMP9, ICAM1, VCAM1, PPARG, and ACSL5. Several hydrogen-bond and hydrophobic interaction patterns were observed within the predicted docking cavities. Overall, these findings provide supportive exploratory evidence for possible interactions between BDE-209 and BDE-209-UC–related targets.

Fig. 8.

Fig. 8

Open in a new tab

Exploratory molecular docking analysis between BDE-209 and prioritized 8 hub-gene-related proteins

Discussion

In this study, we integrated the CTD database, PubChem database, SwissTargetPrediction tool, and GEO datasets to identify overlapping candidate genes between predicted BDE-209-related targets and UC-associated transcriptomic signatures. Differential expression analysis, GO and KEGG enrichment analyses, machine learning approaches (LASSO and random forest), immune infiltration analysis, and single-cell RNA sequencing were subsequently performed using R software. After hub-gene prioritization, molecular docking analysis was performed using the PubChem database, PDB database, and CB-Dock2 platform. This analysis was conducted to explore the structural plausibility of interactions between BDE-209 and the encoded target proteins. Overall, our study prioritized several candidate targets and pathways that may be relevant to the molecular characteristics of UC within the context of predicted BDE-209-related targets. In addition, external validation across independent UC datasets showed that several prioritized hub genes retained relatively stable discriminatory performance, partially supporting the robustness of the screening strategy. Our findings provide exploratory and hypothesis-generating evidence for potential associations between BDE-209-related targets and UC-associated inflammatory and metabolic alterations.

UC is a chronic, non-specific intestinal inflammatory disease with a complex pathogenesis involving genetic susceptibility, environmental factors, and immune dysregulation [20]. The etiology of UC remains elusive, but UC appears to persist in genetically susceptible individuals through an impaired immune response. This aberrant immune response is associated with dysregulation of innate and adaptive immune responses. In UC tissues, the lamina propria contains immune cell populations such as macrophages and dendritic cells, which can interact with adaptive immune cells and participate in inflammatory responses [46]. Activated macrophages have been reported to produce pro-inflammatory mediators, including TNF-α and IL1B, etc [47]. This is consistent with the results of our single-cell analysis on the differential expression of hub genes in various cells clustered in UC tissues.

Macrophages exhibit substantial functional heterogeneity in intestinal immunity. Previous studies have suggested that M1-like macrophages are associated with pro-inflammatory responses, whereas M2-like macrophages may participate in tissue repair and immune regulation [48]. Among the prioritized hub genes identified in this study, Amphiregulin (AREG) attracted particular attention because it has been implicated in epithelial repair and inflammatory regulation. AREG, secreted by macrophages, is a member of the epidermal growth factor (EGF) family and has been reported to participate in tissue homeostasis and mucosal repair following intestinal injury [49]. In addition, clinical and experimental studies have suggested that EGF family-related signaling may contribute to mucosal healing in UC [50, 51]. In the present study, AREG was identified as one of the prioritized hub genes and showed potential biological relevance in UC-related datasets. These observations are consistent with findings from previous studies. Together, they provide indirect support for a possible association between AREG-related signaling and inflammatory or tissue-repair processes in UC. However, our study does not provide direct mechanistic evidence regarding the role of AREG in mediating the effects of BDE-209 on UC pathogenesis.

Acyl-CoA Synthetase Long-Chain Family Member 5 (ACSL5) is a member of the intestinal acyl-CoA synthetase family that has been reported to participate in lipid metabolism and gut microbiota regulation through the hypoxia-responsive PPARGC1A/BAMBI/ACSL5 axis [52]. Previous studies on the acyl-CoA synthetase family have mainly focused on ACSL4, which is highly expressed in UC. ACSL4 may be associated with arachidonic acid metabolism and ferroptosis-related inflammatory responses in M2 macrophages [53]. In addition, experimental studies have suggested a potential association between increased ACSL5 expression and enhanced M1 macrophage infiltration under inflammatory conditions. These findings may reflect a possible link between lipid metabolism and intestinal immune responses [54]. We speculate that BDE-209 exposure may be associated with altered ACSL4 and ACSL5 expression, which could potentially influence macrophage-related inflammatory processes and intestinal repair in UC. Additionally, Fatty Acid Binding Protein 3 (FABP3) is an intestinal fatty acid-binding enzyme that also acts as an antimicrobial protein to regulate the gut microbiota [55]. Studies have shown that FABP3 and ACSL5 are involved in fatty acid metabolism within intestinal epithelial cells [56]. Dysregulation of their functions may disrupt microbial metabolic processes in the gut, potentially leading to intestinal immune dysregulation. Therefore, our findings provide indirect support for a possible association between lipid metabolic dysregulation and UC-related inflammatory alterations. However, the present study does not provide direct mechanistic evidence demonstrating that BDE-209 regulates ACSL5, FABP3, or macrophage-associated pathways in UC. Further experimental validation is therefore required.

Immune cell trafficking is a critical component of intestinal injury. Adhesion molecules such as Vascular Cell Adhesion Molecule 1 (VCAM1) and Intercellular Adhesion Molecule 1 (ICAM1) are expressed by intestinal tissues, vascular endothelial cells, and immune cells. Previous studies have reported that these molecules participate in immune-cell recruitment and inflammatory responses in UC [57]. Previous studies have shown that TNF-α may induce the expression of VCAM1 and ICAM1, thereby promoting immune-cell adhesion and infiltration into intestinal mucosal tissues [58]. These findings suggest that VCAM1- and ICAM1-related signaling may contribute to immune-mediated intestinal injury in UC. However, the mechanism by which BDE-209 affects the expression of VCAM1 and ICAM1 remains to be further investigated. Our study identified VCAM1 and ICAM1 as potential hub genes among the overlapping molecular signatures between predicted BDE-209-related targets and UC-associated transcriptomic alterations.

Oxidative stress has been recognized as an important contributor to the pathogenesis of IBD. In patients with UC, oxidative stress is closely associated with intestinal mucosa damage, thereby exacerbating the inflammatory response [59]. Previous studies evaluating oxidative stress biomarkers in UC patients have further supported the involvement of oxidative stress in disease progression [60]. Elevated oxidative stress levels are associated with the exacerbation of intestinal inflammation. Peroxisome Proliferator-Activated Receptor Gamma (PPARG) participates in the regulation of oxidative stress by maintaining intracellular redox homeostasis [61]. In experimental colitis models, activation of PPARG improved intestinal barrier integrity and alleviated intestinal inflammation [62]. PPARG expression is significantly reduced in DSS-induced UC animal models. By promoting the activation of M2 macrophages, PPARG alleviates intestinal injury in DSS-induced UC mice [22]. These findings provide a possible biological context linking oxidative stress-related pathways with predicted BDE-209-related targets in UC. However, the present study does not provide direct evidence that BDE-209 mediates UC progression through PPARG-dependent oxidative stress mechanisms, and further experimental validation is required.

Interleukin 1 Beta (IL1B) and Matrix Metallopeptidase 9 (MMP9) are well-established inflammation-related genes that play important roles in UC-associated inflammatory responses. IL1B is a key pro-inflammatory cytokine that can amplify local inflammatory responses by activating multiple inflammatory signaling pathways, including NF-κB and IL-17 pathways. Previous studies have demonstrated that IL1B is significantly upregulated in the intestinal tissues of patients with UC and is closely associated with disease activity and inflammatory severity [63]. Meanwhile, MMP9, an important member of the matrix metalloproteinase family, is primarily involved in extracellular matrix degradation, tissue remodeling, and inflammatory-cell migration. Previous studies have shown that MMP9 is frequently highly expressed in active UC tissues and is associated with intestinal mucosal barrier disruption, ulcer formation, and disease severity [64]. Taken together, the abnormal expression patterns of IL1B and MMP9 further support the potential association between the identified candidate hub genes and the UC inflammatory microenvironment.

In summary, our study prioritized candidate toxicant-related targets potentially linking BDE-209 with UC-associated molecular features. Eight hub genes were identified through complementary screening strategies integrating machine learning algorithms, protein-protein interaction analysis, and differential expression characteristics. These findings suggest potential associations between BDE-209-related target networks and UC-related inflammatory and immune-regulatory processes. KEGG enrichment analysis indicated that the identified genes were enriched in inflammation- and immune-related pathways, including the IL-17, NF-κB, TNF, Toll-like receptor, and Th17 cell differentiation pathways. GO enrichment analysis further suggested potential involvement in responses to bacterial molecules and regulation of oxidoreductase activity. Immune infiltration analysis revealed associations between hub genes and multiple immune-cell populations, suggesting potential immune-regulatory relevance in UC. Single-cell RNA-seq analysis further provided supportive evidence regarding the cellular localization of hub genes, particularly within macrophage-related cell populations. In addition, molecular docking analysis provided exploratory structural evidence suggesting potential compatibility between BDE-209 and the protein structures encoded by the identified hub genes. However, these findings should not be interpreted as evidence of direct binding or functional regulation.

Based on the known biological functions of the hub genes, oxidative stress, inflammatory responses, and immune dysregulation may collectively contribute to the association between BDE-209 exposure and UC-related pathological processes. However, the present study remains exploratory and hypothesis-generating in nature, and all findings should be interpreted cautiously. Several limitations of the present study should be acknowledged. First, the current workflow may preferentially identify canonical inflammatory genes broadly associated with UC rather than toxicant-specific molecular signatures. Although comparison with PFOS suggested that some prioritized inflammatory signatures were not completely overlapping, the present analyses do not establish toxicant-specific molecular mechanisms for BDE-209. Therefore, these findings should be interpreted as exploratory associations derived from overlapping target-space analyses rather than evidence of toxicant-specific pathogenic mechanisms. Second, the immune infiltration analysis was based on computational deconvolution of bulk transcriptomic data and therefore reflected inferred immune-cell proportions rather than direct cellular measurements. The findings should be interpreted as correlation-based inferences rather than definitive mechanistic evidence. In addition, although scRNA-seq analysis provided supportive information regarding the cellular localization of hub genes, dedicated doublet-detection procedures were not performed during preprocessing. Thus, the single-cell findings should be considered supportive rather than definitive evidence. Third, the molecular docking analyses were exploratory and do not constitute direct evidence of molecular binding or functional regulation. Finally, because the transcriptomic datasets used in this study did not directly measure BDE-209 exposure, no causal relationship can be established between BDE-209 and UC pathogenesis within the scope of the present bioinformatics analyses. Further experimental and functional studies are required to validate these observations and clarify the underlying biological mechanisms.

Conclusion

In conclusion, this study integrated network toxicology and comprehensive bioinformatics analyses to identify candidate BDE-209-related targets potentially associated with UC-related inflammatory signatures. A systematic screening strategy integrating the CTD, PubChem, SwissTargetPrediction, and GEO databases identified 87 overlapping candidate genes. Among these candidates, eight hub genes (IL1B, MMP9, ICAM1, PPARG, VCAM1, AREG, FABP3, and ACSL5) were prioritized through complementary machine learning and protein–protein interaction analyses. Immune infiltration and single-cell transcriptomic analyses further provided supportive information regarding the potential immune-cell associations and cellular localization patterns of these hub genes in UC. Functional enrichment analyses suggested that these candidate targets were mainly associated with inflammation- and immune-related pathways. These pathways included the IL-17, NF-κB, TNF, Toll-like receptor, and Th17 cell differentiation pathways, as well as biological processes related to bacterial response and oxidoreductase activity. These findings suggest potential associations between BDE-209-related candidate targets and UC-associated inflammatory processes. Overall, the present study provides exploratory bioinformatics evidence and a systematic analytical framework for investigating potential links between environmental pollutants and inflammatory diseases. However, the findings are based on computational prediction and overlapping target-space analyses, and therefore should be interpreted cautiously. Further experimental, functional, and mechanistic studies are required to validate the biological relevance of these observations and clarify the precise relationship between BDE-209 exposure and UC pathogenesis.

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary Material 1 (262.9KB, pdf)
Supplementary Material 2 (199.7KB, pdf)
Supplementary Material 3 (173KB, pdf)
Supplementary Material 4 (104KB, pdf)
Supplementary Material 5 (1.2KB, txt)
Supplementary Material 6 (12.9KB, docx)
Supplementary Material 7 (57.7KB, xlsx)
Supplementary Material 8 (24.9KB, xlsx)
Supplementary Material 9 (14.9KB, xlsx)
Supplementary Material 10 (16.6KB, docx)
Supplementary Material 11 (228.5KB, xlsx)
Supplementary Material 12 (11.5KB, xlsx)
Supplementary Material 13 (145.1KB, xlsx)
Supplementary Material 14 (47KB, xlsx)
Supplementary Material 15 (9.7KB, xlsx)
Supplementary Material 16 (9.5KB, xlsx)
Supplementary Material 17 (256KB, xlsx)

Acknowledgements

We would like to express our gratitude to BioRender (BioRender.com) for the materials provided.

Author contributions

Y.Z.: Conceptualization, Methodology, Data curation, Writing – Original Draft. H.Z.: Formal analysis, Methodology, Visualization. D.H.: Software, Methodology, Data curation. H.Z.: Formal analysis, Software. D.W.: Data curation, Conceptualization. Z.Y.: Investigation. S.S: Writing – review & editing, Supervision. L.C.: Writing – review & editing, Project administration, Supervision.

Funding

This work was supported by the National Natural Science Foundation of China (NSFC-82173906).

Data availability

The datasets analyzed in this study are publicly available from the Gene Expression Omnibus (GEO) database under accession numbers GSE87466, GSE75214, GSE179285, and GSE231993. BDE-209-related target genes were obtained from the Comparative Toxicogenomics Database (CTD), PubChem, and SwissTargetPrediction databases. The scripts used for data processing and analysis are available in the GitHub repository at: https://github.com/hlzbioinfor/bio_ana_UC.git.

Code availability

The code that supports the findings of this study is openly available in GitHub at https://github.com/hlzbioinfor/bio_ana_UC.git. The repository includes the major scripts, workflow order, and package information required to reproduce the primary analyses.

Declarations

Ethical approval

Not applicable.

Consent for publication

Not applicable.

Consent to participate

Not applicable.

Consent to publish

Not applicable.

Competing interests

The authors declare no competing interests.

Footnotes

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Contributor Information

Si Su, Email: 552533925@qq.com.

Ling Chen, Email: chendoc@126.com.

References

  • 1.Sun Y, et al. A critical review on BDE-209: Source, distribution, influencing factors, toxicity, and degradation. Environ Int. 2024;183:108410. [DOI] [PubMed] [Google Scholar]
  • 2.English K, et al. BDE-209 in the Australian Environment: Desktop review. J Hazard Mater. 2016;320:194–203. [DOI] [PubMed] [Google Scholar]
  • 3.Robrock KR, Korytár P, Alvarez-Cohen L. Pathways for the anaerobic microbial debromination of polybrominated diphenyl ethers. Environ Sci Technol. 2008;42(8):2845–52. [DOI] [PubMed] [Google Scholar]
  • 4.Wang Y, et al. Characterization of PBDEs and novel brominated flame retardants in seawater near a coastal mariculture area of the Bohai Sea, China. Sci Total Environ. 2017;580:1446–52. [DOI] [PubMed] [Google Scholar]
  • 5.Wang J, et al. Health risk assessment and development of human health ambient water quality criteria for PBDEs in China. Sci Total Environ. 2021;799:149353. [DOI] [PubMed] [Google Scholar]
  • 6.Jiang Y, et al. Polybrominated diphenyl ethers in the environment and human external and internal exposure in China: A review. Sci Total Environ. 2019;696:133902. [DOI] [PubMed] [Google Scholar]
  • 7.Darnerud PO, et al. Polybrominated diphenyl ethers: occurrence, dietary exposure, and toxicology. Environ Health Perspect. 2001;109(Suppl 1):49–68. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Sharkey M, et al. Phasing-out of legacy brominated flame retardants: The UNEP Stockholm Convention and other legislative action worldwide. Environ Int. 2020;144:106041. [DOI] [PubMed] [Google Scholar]
  • 9.Mariani A, et al. Decabrominated diphenyl ether and methylmercury impair fetal nervous system development in mice at documented human exposure levels. Dev Neurobiol. 2015;75(1):23–38. [DOI] [PubMed] [Google Scholar]
  • 10.Sarkar D, Chowdhury JP, Singh SK. Effect of polybrominated diphenyl ether (BDE-209) on testicular steroidogenesis and spermatogenesis through altered thyroid status in adult mice. Gen Comp Endocrinol. 2016;239:50–61. [DOI] [PubMed] [Google Scholar]
  • 11.Lv QY, et al. In vitro immune toxicity of polybrominated diphenyl ethers on murine peritoneal macrophages: apoptosis and immune cell dysfunction. Chemosphere. 2015;120:621–30. [DOI] [PubMed] [Google Scholar]
  • 12.Li X, et al. BDE-209 induces male reproductive toxicity via cell cycle arrest and apoptosis mediated by DNA damage response signaling pathways. Environ Pollut. 2019;255(Pt 1):113097. [DOI] [PubMed] [Google Scholar]
  • 13.Marchi M, et al. BDE-209 exposure in murine melanoma (B16-F1) cells modulates tumor malignancy and progression in vivo. Food Chem Toxicol. 2024;184:114350. [DOI] [PubMed] [Google Scholar]
  • 14.Sun Y, et al. Hepatotoxicity of decabromodiphenyl ethane (DBDPE) and decabromodiphenyl ether (BDE-209) in 28-day exposed Sprague-Dawley rats. Sci Total Environ. 2020;705:135783. [DOI] [PubMed] [Google Scholar]
  • 15.Che S, et al. Decabromodiphenyl ether initiates mitochondria-dependent apoptosis by disrupting calcium homeostasis in mice livers. Chemosphere. 2022;291(Pt 1):132767. [DOI] [PubMed] [Google Scholar]
  • 16.Yang Y, et al. Hepatotoxicity evaluation and possible mechanisms of decabrominated diphenyl ethers (BDE-209) in broilers: Oxidative stress, inflammatory, and transcriptomics. Ecotoxicol Environ Saf. 2023;264:115460. [DOI] [PubMed] [Google Scholar]
  • 17.Zhang Z, et al. SeNPs alleviates BDE-209-induced intestinal damage by affecting necroptosis, inflammation, intestinal barrier and intestinal flora in layer chickens. Ecotoxicol Environ Saf. 2023;262:115336. [DOI] [PubMed] [Google Scholar]
  • 18.Wang J, et al. BDE-209-induced genotoxicity, intestinal damage and intestinal microbiota dysbiosis in zebrafish (Danio Rerio). Sci Total Environ. 2023;905:167009. [DOI] [PubMed] [Google Scholar]
  • 19.Li S, et al. Decabromodiphenyl ether induces ROS-mediated intestinal toxicity through the Keap1-Nrf2 pathway. J Biochem Mol Toxicol. 2022;36(4):e22995. [DOI] [PubMed] [Google Scholar]
  • 20.Ordás I, et al. Ulcerative colitis. Lancet. 2012;380(9853):1606–19. [DOI] [PubMed] [Google Scholar]
  • 21.Shaoyong W, et al. BDE-209 caused gut toxicity through modulating the intestinal barrier, oxidative stress, autophagy, inflammation, and apoptosis in mice. Sci Total Environ. 2021;776:146018. [Google Scholar]
  • 22.Yuan J, et al. Regulatory effects of flavonoids luteolin on BDE-209-induced intestinal epithelial barrier damage in Caco-2 cell monolayer model. Food Chem Toxicol. 2021;150:112098. [DOI] [PubMed] [Google Scholar]
  • 23.Dai Y, et al. Xianglian Pill attenuates ulcerative colitis through TLR4/MyD88/NF-κB signaling pathway. J Ethnopharmacol. 2023;300:115690. [DOI] [PubMed] [Google Scholar]
  • 24.Ghorbaninejad M, et al. Contribution of NOTCH signaling pathway along with TNF-α in the intestinal inflammation of ulcerative colitis. Gastroenterol Hepatol Bed Bench. 2019;12(Suppl1):S80–6. [PMC free article] [PubMed] [Google Scholar]
  • 25.Wei YY, et al. Shaoyao decoction attenuates DSS-induced ulcerative colitis, macrophage and NLRP3 inflammasome activation through the MKP1/NF-κB pathway. Phytomedicine. 2021;92:153743. [DOI] [PubMed] [Google Scholar]
  • 26.Davis AP, et al. Comparative Toxicogenomics Database (CTD): update 2023. Nucleic Acids Res. 2023;51(D1):D1257–62. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Kim S, et al. PubChem in 2021: new data content and improved web interfaces. Nucleic Acids Res. 2021;49(D1):D1388–95. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Daina A, Michielin O, Zoete V. SwissTargetPrediction: updated data and new features for efficient prediction of protein targets of small molecules. Nucleic Acids Res. 2019;47(W1):W357–64. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Vancamelbeke M, et al. Genetic and Transcriptomic Bases of Intestinal Epithelial Barrier Dysfunction in Inflammatory Bowel Disease. Inflamm Bowel Dis. 2017;23(10):1718–29. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Li K, et al. Molecular Comparison of Adult and Pediatric Ulcerative Colitis Indicates Broad Similarity of Molecular Pathways in Disease Tissue. J Pediatr Gastroenterol Nutr. 2018;67(1):45–52. [DOI] [PubMed] [Google Scholar]
  • 31.Ritchie ME, et al. limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res. 2015;43(7):e47. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Gene Ontology Consortium. going forward. Nucleic Acids Res. 2015;43(Database issue):D1049–56. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Kanehisa M, et al. KEGG: new perspectives on genomes, pathways, diseases and drugs. Nucleic Acids Res. 2017;45(D1):D353–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Engebretsen S, Bohlin J. Statistical predictions with glmnet. Clin Epigenetics. 2019;11(1):123. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Szklarczyk D, et al. STRING v11: protein-protein association networks with increased coverage, supporting functional discovery in genome-wide experimental datasets. Nucleic Acids Res. 2019;47(D1):D607–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Shannon P, et al. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res. 2003;13(11):2498–504. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Wang F, et al. Mechanistic insights into PFOS-induced inflammatory bowel disease: a network toxicology and molecular docking study. Int J Surg. 2025;111(8):5150–61. [DOI] [PubMed] [Google Scholar]
  • 38.Keir ME, et al. Regulation and Role of αE Integrin and Gut Homing Integrins in Migration and Retention of Intestinal Lymphocytes during Inflammatory Bowel Disease. J Immunol. 2021;207(9):2245–54. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Saez A, et al. Pathophysiology of inflammatory bowel disease: innate immune system. Int J Mol Sci.2023;24(2):1526. [DOI] [PMC free article] [PubMed]
  • 40.Newman AM, et al. Robust enumeration of cell subsets from tissue expression profiles. Nat Methods. 2015;12(5):453–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Ito K, Murphy D. Application of ggplot2 to Pharmacometric Graphics. CPT Pharmacometrics Syst Pharmacol. 2013;2(10):e79. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Du J, et al. Selective oxidative protection leads to tissue topological changes orchestrated by macrophage during ulcerative colitis. Nat Commun. 2023;14(1):3675. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Wei W, et al. Identifying polyamine related biomarkers in diagnosis and treatment of ulcerative colitis by integrating bulk and single-cell sequencing data. Sci Rep. 2024;14(1):18094. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Cheng F, et al. Triclosan exposure potentiates ischemic stroke risk: Multi-omics integration and molecular docking unveil neurotoxic mechanisms. Ecotoxicol Environ Saf. 2025;302:118551. [DOI] [PubMed] [Google Scholar]
  • 45.Wang Z, et al. Efficient evaluation of osteotoxicity and mechanisms of endocrine disrupting chemicals using network toxicology and molecular docking approaches: triclosan as a model compound. Ecotoxicol Environ Saf. 2025;293:118030. [DOI] [PubMed] [Google Scholar]
  • 46.Hart AL, et al. Characteristics of intestinal dendritic cells in inflammatory bowel diseases. Gastroenterology. 2005;129(1):50–65. [DOI] [PubMed] [Google Scholar]
  • 47.Grimm MC, et al. Direct evidence of monocyte recruitment to inflammatory bowel disease mucosa. J Gastroenterol Hepatol. 1995;10(4):387–95. [DOI] [PubMed] [Google Scholar]
  • 48.Fritsch SD, et al. Metabolic support by macrophages sustains colonic epithelial homeostasis. Cell Metab. 2023;35(11):1931. –1943.e8. [DOI] [PubMed] [Google Scholar]
  • 49.Ko JH, et al. Mesenchymal Stem and Stromal Cells Harness Macrophage-Derived Amphiregulin to Maintain Tissue Homeostasis. Cell Rep. 2020;30(11):3806–20.e6. [DOI] [PubMed] [Google Scholar]
  • 50.Farrell RJ. Epidermal growth factor for ulcerative colitis. N Engl J Med. 2003;349(4):395–7. [DOI] [PubMed] [Google Scholar]
  • 51.Babyatsky MW, Rossiter G, Podolsky DK. Expression of transforming growth factors alpha and beta in colonic mucosa in inflammatory bowel disease. Gastroenterology. 1996;110(4):975–84. [DOI] [PubMed] [Google Scholar]
  • 52.Su Y, et al. Effect of dietary supplementation with sanguinarine on meat quality and lipid metabolism of broilers. Poult Sci. 2024;103(8):103925. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 53.Ye Y, et al. The ERK-cPLA2-ACSL4 axis mediating M2 macrophages ferroptosis impedes mucosal healing in ulcerative colitis. Free Radic Biol Med. 2024;214:219–35. [DOI] [PubMed] [Google Scholar]
  • 54.Ding X, et al. ACSL5 mediates macrophage infiltration and lipid metabolism in erythrotelangiectasia rosacea via potential pathogenic mechanisms and therapeutic targets. Sci Rep. 2025;15(1):11929. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 55.Baru A, et al. Anti bacterial function of secreted human FABP3. Biochim Biophys Acta Gen Subj. 2023;1867(12):130472. [DOI] [PubMed] [Google Scholar]
  • 56.Wang Z, et al. A host enzyme reduces metabolic dysfunction-associated steatotic liver disease (MASLD) by inactivating intestinal lipopolysaccharide. Elife. 2025;13:RP100731. [DOI] [PMC free article] [PubMed]
  • 57.Verstockt B, et al. Sphingosine 1-phosphate modulation and immune cell trafficking in inflammatory bowel disease. Nat Rev Gastroenterol Hepatol. 2022;19(6):351–66. [DOI] [PubMed] [Google Scholar]
  • 58.Yan M, et al. Neuroprotectant androst-3β, 5α, 6β-triol suppresses TNF-α-induced endothelial adhesion molecules expression and neutrophil adhesion to endothelial cells by attenuation of CYLD-NF-κB pathway. Biochem Biophys Res Commun. 2017;483(2):892–6. [DOI] [PubMed] [Google Scholar]
  • 59.Muro P, et al. The emerging role of oxidative stress in inflammatory bowel disease. Front Endocrinol (Lausanne). 2024;15:1390351. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 60.Rana SV, et al. Role of oxidative stress & antioxidant defence in ulcerative colitis patients from north India. Indian J Med Res. 2014;139(4):568–71. [PMC free article] [PubMed] [Google Scholar]
  • 61.Dovinova I, et al. The role of Nrf2 and PPARgamma in the improvement of oxidative stress in hypertension and cardiovascular diseases. Physiol Res. 2020;69(Suppl 4):S541–53. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 62.Gu Y, et al. Dietary artemisinin boosts intestinal immunity and healthy in fat greenling (Hexagrammos otakii). Front Immunol. 2023;14:1198902. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 63.Xue JC, et al. Natural products modulate NLRP3 in ulcerative colitis. Front Pharmacol. 2023;14:1265825. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 64.Farkas K, et al. The diagnostic value of a new fecal marker, matrix metalloprotease-9, in different types of inflammatory bowel diseases. J Crohns Colitis. 2015;9(3):231–7. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Material 1 (262.9KB, pdf)
Supplementary Material 2 (199.7KB, pdf)
Supplementary Material 3 (173KB, pdf)
Supplementary Material 4 (104KB, pdf)
Supplementary Material 5 (1.2KB, txt)
Supplementary Material 6 (12.9KB, docx)
Supplementary Material 7 (57.7KB, xlsx)
Supplementary Material 8 (24.9KB, xlsx)
Supplementary Material 9 (14.9KB, xlsx)
Supplementary Material 10 (16.6KB, docx)
Supplementary Material 11 (228.5KB, xlsx)
Supplementary Material 12 (11.5KB, xlsx)
Supplementary Material 13 (145.1KB, xlsx)
Supplementary Material 14 (47KB, xlsx)
Supplementary Material 15 (9.7KB, xlsx)
Supplementary Material 16 (9.5KB, xlsx)
Supplementary Material 17 (256KB, xlsx)

Data Availability Statement

The datasets analyzed in this study are publicly available from the Gene Expression Omnibus (GEO) database under accession numbers GSE87466, GSE75214, GSE179285, and GSE231993. BDE-209-related target genes were obtained from the Comparative Toxicogenomics Database (CTD), PubChem, and SwissTargetPrediction databases. The scripts used for data processing and analysis are available in the GitHub repository at: https://github.com/hlzbioinfor/bio_ana_UC.git.

The code that supports the findings of this study is openly available in GitHub at https://github.com/hlzbioinfor/bio_ana_UC.git. The repository includes the major scripts, workflow order, and package information required to reproduce the primary analyses.

Articles from BMC Pharmacology & Toxicology are provided here courtesy of BMC

RESOURCES