Scholarlink

Background <p>As a prevalent plasticizer in industrial manufacturing and daily-use products, di(2-ethylhexyl) phthalate (DEHP) has raised growing concerns regarding its safety and potential role in disease pathogenesis. Although direct epidemiological evidence linking DEHP exposure to Sjögren’s syndrome (SS) remains limited, DEHP has been reported to exert immunomodulatory and endocrine-disrupting effects that are highly relevant to the core pathological mechanisms of SS. Therefore, this study aimed to systematically explore the potential links between DEHP and SS using an integrated strategy combining toxicity prediction, network toxicology, machine learning, molecular docking, single-cell atlas interrogation, and in vitro validation.</p> Methods <p>The toxicity profile of DEHP was predicted using two complementary online platforms (ProTox 3.0 and ADMETlab 2.0). SS-related transcriptomic datasets were obtained from GEO, and potential DEHP targets were collected from CTD, SwissTargetPrediction, and SEA. Network toxicology analyses (intersection screening, PPI construction, and GO/KEGG enrichment) were performed to identify key genes and pathways. Machine learning models were trained to classify SS versus controls using the DEHP–SS intersect genes, and SHAP was applied to interpret key predictors. Molecular docking (AutoDock Vina) was conducted to estimate the binding propensity between DEHP and prioritized proteins. A single-cell atlas of mouse submandibular gland tissue (PanglaoDB; Mus musculus; SRA693675:SRS3206192) was queried to localize gene expression. Finally, in vitro experiments were performed in human submandibular gland epithelial cells (HSG) to validate STAT1 expression changes following DEHP exposure.</p> Results <p>In silico toxicity predictions suggested that DEHP may exhibit immunotoxic potential among multiple predicted toxicity endpoints. Network toxicology analysis identified 27 DEHP–SS intersect targets, and the PPI network highlighted STAT1, CCL2, CXCL10, MYD88, and IRF7 as highly connected hub genes. Across 113 machine learning models, Ridge/Elastic Net–based models demonstrated stable classification performance, and SHAP analysis prioritized five core predictors (STAT1, PHGDH, ISG15, CXCL10, and CCL2). Molecular docking suggested favorable binding energies ( ≤ − 5.0&#xa0;kcal/mol) between DEHP and these five proteins. Single-cell atlas interrogation indicated that STAT1 is expressed in salivary mucous cells. In vitro experiments further demonstrated that DEHP exposure increases STAT1 mRNA and protein expression in HSG cells in a dose-dependent manner. Collectively, these results provide mechanistically plausible—yet non-causal—evidence linking DEHP to SS-relevant immune signatures.</p> Conclusion <p>This study identifies an overlapping immune-response signature between predicted DEHP-associated targets and SS-related transcriptomic features, highlighting STAT1 as a prioritized node within this intersection. In vitro experiments demonstrate that DEHP exposure upregulates STAT1 expression in HSG cells. These findings provide hypothesis-generating evidence that DEHP may modulate interferon-related immune pathways relevant to SS, while not establishing causality.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Network toxicology integrated with machine learning and SHAP analysis identifies overlapping immune signatures between Di(2-ethylhexyl) phthalate (DEHP) and Sjögren’s syndrome

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年3月7日

查看全文
  • Cheng Lili,
  • Tang Zhongfu,
  • Li Ming,
  • Chuanbing Huang

摘要

Background

As a prevalent plasticizer in industrial manufacturing and daily-use products, di(2-ethylhexyl) phthalate (DEHP) has raised growing concerns regarding its safety and potential role in disease pathogenesis. Although direct epidemiological evidence linking DEHP exposure to Sjögren’s syndrome (SS) remains limited, DEHP has been reported to exert immunomodulatory and endocrine-disrupting effects that are highly relevant to the core pathological mechanisms of SS. Therefore, this study aimed to systematically explore the potential links between DEHP and SS using an integrated strategy combining toxicity prediction, network toxicology, machine learning, molecular docking, single-cell atlas interrogation, and in vitro validation.

Methods

The toxicity profile of DEHP was predicted using two complementary online platforms (ProTox 3.0 and ADMETlab 2.0). SS-related transcriptomic datasets were obtained from GEO, and potential DEHP targets were collected from CTD, SwissTargetPrediction, and SEA. Network toxicology analyses (intersection screening, PPI construction, and GO/KEGG enrichment) were performed to identify key genes and pathways. Machine learning models were trained to classify SS versus controls using the DEHP–SS intersect genes, and SHAP was applied to interpret key predictors. Molecular docking (AutoDock Vina) was conducted to estimate the binding propensity between DEHP and prioritized proteins. A single-cell atlas of mouse submandibular gland tissue (PanglaoDB; Mus musculus; SRA693675:SRS3206192) was queried to localize gene expression. Finally, in vitro experiments were performed in human submandibular gland epithelial cells (HSG) to validate STAT1 expression changes following DEHP exposure.

Results

In silico toxicity predictions suggested that DEHP may exhibit immunotoxic potential among multiple predicted toxicity endpoints. Network toxicology analysis identified 27 DEHP–SS intersect targets, and the PPI network highlighted STAT1, CCL2, CXCL10, MYD88, and IRF7 as highly connected hub genes. Across 113 machine learning models, Ridge/Elastic Net–based models demonstrated stable classification performance, and SHAP analysis prioritized five core predictors (STAT1, PHGDH, ISG15, CXCL10, and CCL2). Molecular docking suggested favorable binding energies ( ≤ − 5.0 kcal/mol) between DEHP and these five proteins. Single-cell atlas interrogation indicated that STAT1 is expressed in salivary mucous cells. In vitro experiments further demonstrated that DEHP exposure increases STAT1 mRNA and protein expression in HSG cells in a dose-dependent manner. Collectively, these results provide mechanistically plausible—yet non-causal—evidence linking DEHP to SS-relevant immune signatures.

Conclusion

This study identifies an overlapping immune-response signature between predicted DEHP-associated targets and SS-related transcriptomic features, highlighting STAT1 as a prioritized node within this intersection. In vitro experiments demonstrate that DEHP exposure upregulates STAT1 expression in HSG cells. These findings provide hypothesis-generating evidence that DEHP may modulate interferon-related immune pathways relevant to SS, while not establishing causality.