Integrative multi-omics, machine learning, computational toxicology, and experimental validation reveal PFAS-associated immune–inflammatory signatures in chronic obstructive pulmonary disease
Abstract
BackgroundPer- and polyfluoroalkyl substances (PFAS) are persistent environmental contaminants increasingly implicated in respiratory disorders. However, the systems-level connections among PFAS-associated molecular perturbations, immune dysregulation, and chronic obstructive pulmonary disease (COPD) remain poorly understood. This study aimed to prioritize candidate links between PFAS-associated target information and COPD-related transcriptional and inflammatory signatures.MethodsPFAS-associated target information was integrated with bulk transcriptomic differential expression and weighted gene co-expression network analysis to identify overlapping candidate genes. Protein–protein interaction analysis and 113 machine-learning model combinations were used for candidate prioritization and cross-cohort classification evaluation. CIBERSORT and xCell provided computational immune-cell signatures, with Benjamini–Hochberg correction applied to group comparisons and gene–signature correlations; single-cell transcriptomics provided cell-type expression context. Candidate expression was examined in cigarette-smoke-extract-treated cells, clinical samples, and public human and murine datasets. Molecular docking and molecular dynamics simulations explored candidate PFAS–protein interactions.ResultsHMOX1, JUNB, CXCL12, and PTGS2 were prioritized as candidate genes overlapping PFAS-associated target information and COPD-related expression patterns. Enrichment and immune-deconvolution analyses suggested associations with inflammatory and immune-related signatures, while single-cell analysis described their distribution across annotated cell types. Exploratory plasma ELISA measurements supported higher levels of the four candidate proteins in participants with COPD, with additional expression support varying by gene and experimental model. Docking and molecular dynamics generated structural hypotheses for PFOA/PFOS–protein interactions, including HMOX1.ConclusionsThis integrative systems-immunology framework delineates immune–inflammatory signatures at the PFAS–COPD interface and identifies HMOX1, JUNB, CXCL12, and PTGS2 as convergent molecular nodes. The findings do not establish exposure causality but provide a multi-scale basis for future epidemiological and mechanistic studies of how persistent environmental pollutants may contribute to COPD-associated immune and inflammatory alterations.
Related
- Potential molecular links between triclosan exposure and osteoarthritis: An integrative systems toxicology study
- Genetic-multi-omics causal inference coupled with integrative in-vitro toxicology reveals immune-metabolite mediators for aortic aneurysm risk
- Integrative network toxicology and machine learning reveal the potential role of 5-hydroxymethylfurfural in exacerbating pulmonary fibrosis
- Environmental PFOA exposure and the risk of metabolic dysfunction-associated steatotic liver disease: An integrated computational toxicology and multi-omics study
- Molecular Mechanisms of 6PPD and 6PPD-Q Toxicity in Neurodegenerative Diseases: A Network Toxicology and Experimental Validation Study
BibTeX
@article{hu2026integrative,
title = {{Integrative multi-omics, machine learning, computational toxicology, and experimental validation reveal PFAS-associated immune–inflammatory signatures in chronic obstructive pulmonary disease}},
author = {Zhengyu Hu and Yuanlan Yang and Jialin Yu and Miaomiao Nian and Jieda Cui and Pixin Ran and Xia Cao and Hongxia Xin and Xiuyan Wang and Guilan Yang and Zhenzhen Gui and Zhanbing Ma and Juan Chen},
journal = {Frontiers in Immunology},
year = {2026},
month = nov,
doi = {10.3389/fimmu.2026.1965629},
url = {https://www.frontiersin.org/articles/10.3389/fimmu.2026.1965629},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Integrative multi-omics, machine learning, computational toxicology, and experimental validation reveal PFAS-associated immune–inflammatory signatures in chronic obstructive pulmonary disease
AU - Zhengyu Hu
AU - Yuanlan Yang
AU - Jialin Yu
AU - Miaomiao Nian
AU - Jieda Cui
AU - Pixin Ran
AU - Xia Cao
AU - Hongxia Xin
AU - Xiuyan Wang
AU - Guilan Yang
AU - Zhenzhen Gui
AU - Zhanbing Ma
AU - Juan Chen
PY - 2026
DA - 2026/11/06
JO - Frontiers in Immunology
DO - 10.3389/fimmu.2026.1965629
UR - https://www.frontiersin.org/articles/10.3389/fimmu.2026.1965629
AB - BackgroundPer- and polyfluoroalkyl substances (PFAS) are persistent environmental contaminants increasingly implicated in respiratory disorders. However, the systems-level connections among PFAS-associated molecular perturbations, immune dysregulation, and chronic obstructive pulmonary disease (COPD) remain poorly understood. This study aimed to prioritize candidate links between PFAS-associated target information and COPD-related transcriptional and inflammatory signatures.MethodsPFAS-associated target information was integrated with bulk transcriptomic differential expression and weighted gene co-expression network analysis to identify overlapping candidate genes. Protein–protein interaction analysis and 113 machine-learning model combinations were used for candidate prioritization and cross-cohort classification evaluation. CIBERSORT and xCell provided computational immune-cell signatures, with Benjamini–Hochberg correction applied to group comparisons and gene–signature correlations; single-cell transcriptomics provided cell-type expression context. Candidate expression was examined in cigarette-smoke-extract-treated cells, clinical samples, and public human and murine datasets. Molecular docking and molecular dynamics simulations explored candidate PFAS–protein interactions.ResultsHMOX1, JUNB, CXCL12, and PTGS2 were prioritized as candidate genes overlapping PFAS-associated target information and COPD-related expression patterns. Enrichment and immune-deconvolution analyses suggested associations with inflammatory and immune-related signatures, while single-cell analysis described their distribution across annotated cell types. Exploratory plasma ELISA measurements supported higher levels of the four candidate proteins in participants with COPD, with additional expression support varying by gene and experimental model. Docking and molecular dynamics generated structural hypotheses for PFOA/PFOS–protein interactions, including HMOX1.ConclusionsThis integrative systems-immunology framework delineates immune–inflammatory signatures at the PFAS–COPD interface and identifies HMOX1, JUNB, CXCL12, and PTGS2 as convergent molecular nodes. The findings do not establish exposure causality but provide a multi-scale basis for future epidemiological and mechanistic studies of how persistent environmental pollutants may contribute to COPD-associated immune and inflammatory alterations.
SN - 1664-3224
ER -