DEHP Exposure Points to Immune and Epithelial Drivers in Endometriosis

Endometriosis is increasingly being read not only as an estrogen-dependent pelvic disorder, but also as a disease of persistent inflammation, tissue invasion, and immune miscommunication. In this Frontiers in Immunology study, integrated transcriptomic, toxicogenomic, and single-cell analyses traced a potential molecular path from DEHP exposure to the altered immune microenvironment seen in endometriosis. The signal converged on three candidate hub genes — CAPN2, CHRM3, and PRKCQ — with epithelial cells emerging as the main cellular compartment carrying the aberrant expression pattern.

The biological logic is plausible. DEHP is a widely used plasticizer with chronic human exposure through food, skin contact, and inhalation, and the ovary and endometrium are both vulnerable to toxicants that perturb steroid signaling, oxidative stress, and cellular survival pathways. In endometriosis, a lesion can persist only if ectopic endometrial cells survive, attach, invade, and evade immune clearance. A chemical exposure capable of nudging those processes in the wrong direction deserves close scrutiny.

One sentence captures the study’s central point: DEHP did not appear as a passive background exposure, but as a candidate molecular modifier of disease biology.

Cross-Platform Gene Filtering Narrowed the Signal

The analysis began with public bulk transcriptomic datasets, GSE7305 as the discovery cohort and GSE23339 as external validation. The discovery set included 10 endometriosis samples and 10 controls; the validation set included 10 endometriosis samples and 9 normal endometrial samples. Those datasets were paired with single-cell RNA sequencing data derived mainly from 7 endometriosis tissues and 3 control tissues, allowing the investigators to move from bulk association toward cell-specific context.

Differential expression analysis and weighted gene co-expression network analysis were used to identify endometriosis-associated genes, then those genes were intersected with predicted DEHP toxicity targets. Machine-learning feature selection, using LASSO regression and support vector machine recursive feature elimination, further reduced the list. From this layered approach, six candidate genes linked both endometriosis and DEHP exposure. Three rose to the top as candidate hub genes: CAPN2, CHRM3, and PRKCQ.

The workflow matters because it reduces the chance that a single noisy dataset drives the entire result. By forcing agreement across differential expression, co-expression structure, toxicology prediction, and machine-learning ranking, the study searched for convergent biology rather than isolated statistical signals.

Immune Microenvironment Changes Followed CHRM3 Expression

Pathway analysis placed the candidate genes inside immune-related and cell cycle-associated programs. That is important for endometriosis, where local survival of ectopic tissue depends on both proliferative capacity and a permissive inflammatory niche. Among the candidate genes, CHRM3 stood out in immune infiltration analysis. Its expression correlated negatively with resting natural killer cells and positively with M2 macrophages, suggesting a shift toward an immune environment more likely to tolerate lesion persistence than to eliminate abnormal tissue.

This pattern has clinical resonance. Natural killer cells help police aberrant cells, while M2 macrophages are often associated with tissue repair, immune dampening, and pro-survival signaling. If CHRM3 tracks with fewer resting natural killer cells and more M2 macrophages, the endometriotic environment may be leaning away from immune surveillance and toward lesion support. In a chronic disease where escape from immune clearance is central, such a skew is biologically coherent.

The study also used molecular docking and molecular dynamics simulations to explore potential interactions between DEHP and the proteins encoded by the candidate hub genes. Those analyses are exploratory, not proof of causality, but they reinforce the possibility that DEHP can intersect directly with molecular machinery relevant to lesion biology.

Epithelial Cells Carried the Strongest Aberrant Signature

Single-cell analysis added a layer that bulk RNA sequencing cannot provide. Epithelial cells emerged as the primary population showing abnormal candidate hub gene expression and immune interaction signatures. That finding is particularly relevant because endometrial epithelial cells are not inert structural elements; they participate in adhesion, invasion, cytokine signaling, and the construction of a tissue niche that supports endometriotic implants.

The epithelial focus also fits the disease phenotype. Ectopic endometrial lesions depend on epithelial programs that facilitate attachment and persistence outside the uterine cavity, and environmental chemicals may exert disproportionate effects on those programs. The single-cell data therefore give tissue-level meaning to the candidate genes: the signal is not diffuse, but concentrated in a compartment with direct responsibility for lesion behavior.

Tissue Validation and Cell Experiments Supported the Bioinformatic Pattern

Experimental validation strengthened the computational findings. In ectopic versus eutopic endometrial tissue, CAPN2 and CHRM3 were increased, while PRKCQ was decreased. The directionality matters because it suggests a coordinated remodeling rather than random fluctuation. CAPN2, a calpain family protease, is consistent with altered cytoskeletal dynamics and cellular remodeling; CHRM3, a muscarinic receptor, can plausibly influence signaling pathways tied to proliferation and microenvironmental response; PRKCQ, a protein kinase involved in immune signaling, points toward altered immune communication.

The in vitro DEHP exposure experiments extended those observations. Endometrial epithelial cells exposed to DEHP showed changes in CAPN2, CHRM3, and PRKCQ expression at both mRNA and protein levels. In practical terms, this gives the environmental exposure a laboratory footprint: DEHP was able to move the same molecular targets identified in patient tissue and computational screening.

One result deserves emphasis on its own. DEHP altered candidate hub gene expression in endometrial epithelial cells at both transcript and protein levels.

What This Means for Environmental Risk and Endometriosis Biology

The study does not establish that DEHP causes endometriosis in humans, and it does not define a clinical threshold of exposure. It does, however, offer a coherent molecular framework linking a common environmental plasticizer with immune dysregulation and epithelial dysfunction in the disease. That framework is valuable because endometriosis remains a disorder with substantial diagnostic delay and limited disease-modifying therapies. If environmental toxicants reinforce lesion persistence through immune and epithelial pathways, then mechanistic targets like CAPN2, CHRM3, and PRKCQ may become useful markers for future validation.

The strongest contribution here is conceptual: endometriosis can be studied as an interface disease, where endocrine signaling, immunity, epithelial biology, and environmental chemistry intersect. The present data place DEHP inside that interface. The next step is not assumption, but replication in larger clinical cohorts and functional experiments that can separate association from mechanism.

Study Constraints Temper the Inference

This was a multi-layered bioinformatic and experimental study, but its datasets were still modest in size, and several analytic steps relied on public repositories rather than prospectively collected clinical samples. The enrichment and docking analyses are hypothesis-generating, not definitive evidence of binding or disease causation. The single-cell component strengthens tissue localization, yet it does not resolve whether the observed molecular shifts initiate disease, amplify established lesions, or both.

Even with those limits, the signal is difficult to dismiss. DEHP tracked with a disease-associated immune pattern, epithelial involvement, and reproducible shifts in candidate hub genes across tissue and cell models. For a condition as biologically complex as endometriosis, that combination is enough to justify deeper mechanistic work rather than casual dismissal.


Reporting by Dr. A. Nikhil Krishnan, MBBS for Medicine World Foundation, with additional data provided by Frontiers in Immunology.
PRIMARY CLINICAL EVIDENCE & REGISTRY DATA