Elevated GPX3 expression in endometriosis: implications for pathogenesis and immune cell infiltration

Three cohorts of ectopic endometrium and control samples were obtained from three GEO repositories (Fig.  1 A). After the removal of the batch effect, the integrated dataset was obtained and normalized, comprising 101 endometriosis samples and 87 control samples (Fig.  1 B). Differential analysis between the endometriosis and healthy groups identified 67 hub genes based on a cutoff criterion of adjusted p < 0.05 and |log2(fold change)|≥ 1, including 21 increased genes and 46 decreased genes. A volcano plot and heatmap were used to visualize the expression patterns of these DEGs (Fig.  1 C, D). Two machine learning algorithms were employed to identify signature genes from the disorderly DEGs in endometriosis and control samples. The LASSO analysis identified 19 signature genes (Fig.  1 E, F), while the SVM algorithm identified 24 signature genes (Fig.  1 G). The Venn diagram revealed that 12 hub genes were common to both algorithms (Fig.  1 H), and a heatmap illustrated the expression of these genes (Fig.  1 I). These selected hub genes were significantly differentially expressed between the endometriosis and control groups, indicating their potential importance in endometriosis (Figure S1 A–L). Additionally, the AUC of the ROC curve for each of these signature genes was displayed (Figure S1 M–X). Fig. 1 The performance of the signature genes in endometriosis. A – B The PCA cluster plot of the GSE25628 , GSE51981 , and GSE7305 datasets before ( A ) and after ( B ) sample correction. C The volcano plot illustrates the differential gene expression between ectopic endometrium and control endometrium. Red dots represent up-regulated genes, green dots represent down-regulated genes, and grey dots indicate genes with no significant differential expression. D The heatmap showed the up-regulated and down-regulated differentially expressed genes between the ectopic endometrium and control endometrium. E Least absolute shrinkage and selection operator (LASSO) logistic regression algorithm to screen diagnostic markers. Different colors represent different genes. F The cross-validation for tuning the parameter selection in the LASSO regression. G Support vector machine (SVM) algorithm to screen 24 signature genes as diagnostic markers. H The interaction of signature genes by the LASSO and SVM algorithms. I The heatmap depicts the expression levels of 12 signature genes The performance of the signature genes in endometriosis. A – B The PCA cluster plot of the GSE25628 , GSE51981 , and GSE7305 datasets before ( A ) and after ( B ) sample correction. C The volcano plot illustrates the differential gene expression between ectopic endometrium and control endometrium. Red dots represent up-regulated genes, green dots represent down-regulated genes, and grey dots indicate genes with no significant differential expression. D The heatmap showed the up-regulated and down-regulated differentially expressed genes between the ectopic endometrium and control endometrium. E Least absolute shrinkage and selection operator (LASSO) logistic regression algorithm to screen diagnostic markers. Different colors represent different genes. F The cross-validation for tuning the parameter selection in the LASSO regression. G Support vector machine (SVM) algorithm to screen 24 signature genes as diagnostic markers. H The interaction of signature genes by the LASSO and SVM algorithms. I The heatmap depicts the expression levels of 12 signature genes To further elucidate the characteristics of endometriosis, we performed the proteomic assay to detect significant proteins from eutopic endometrial tissues and ectopic endometrial tissues from 10 patients with endometriosis, alongside 5 control endometrial samples serving as controls, as depicted in Fig.  2 A. Volcano plots illustrated differentially expressed proteins, revealing minimal distinctions between control endometrial samples with endometriosis patients’ eutopic endometrial samples (Fig.  2 B). However, significant differential expression was observed between ectopic with eutopic endometrial samples (Fig.  2 C), as well as between ectopic and control endometrial tissues (Fig.  2 D). A Venn diagram indicated that 330 proteins are significantly associated with endometriosis (Fig.  2 E). Pathway analysis highlighted the enrichment of dysregulated proteins in cellular migration, vascularization, and other cellular signaling pathways significantly associated with endometriosis (Fig.  2 F). Fig. 2 Proteomic profiling reveals endometriosis features. A A schematic illustrating the workflow of proteomic identification. Ectopic and eutopic endometrial tissues were obtained from 10 patients with endometriosis and 5 healthy control subjects. B – D Volcano plots displaying proteins exhibiting dysregulated expression in endometriosis. Specifically, contrasts normal endometrium with eutopic endometrium of endometriosis patients ( B ); compares normal endometrium with ectopic endometrium of endometriosis patients ( C ); evaluates the differences between eutopic tissue of endometriosis and ectopic tissue of endometriosis ( D ). E Venn diagram delineating the disorganized protein expression in ectopic endometrium. Blue denotes proteins with differential expression between ectopic and eutopic endometrium in patients with endometriosis, while red indicates differential expression between normal endometrium and ectopic endometrium in patients with endometriosis. F Functional enrichment analysis conducted on proteins exhibiting dysregulated expression in ectopic endometrium tissue of patients with endometriosis Proteomic profiling reveals endometriosis features. A A schematic illustrating the workflow of proteomic identification. Ectopic and eutopic endometrial tissues were obtained from 10 patients with endometriosis and 5 healthy control subjects. B – D Volcano plots displaying proteins exhibiting dysregulated expression in endometriosis. Specifically, contrasts normal endometrium with eutopic endometrium of endometriosis patients ( B ); compares normal endometrium with ectopic endometrium of endometriosis patients ( C ); evaluates the differences between eutopic tissue of endometriosis and ectopic tissue of endometriosis ( D ). E Venn diagram delineating the disorganized protein expression in ectopic endometrium. Blue denotes proteins with differential expression between ectopic and eutopic endometrium in patients with endometriosis, while red indicates differential expression between normal endometrium and ectopic endometrium in patients with endometriosis. F Functional enrichment analysis conducted on proteins exhibiting dysregulated expression in ectopic endometrium tissue of patients with endometriosis To further elucidate the molecular biomarkers associated with endometriosis, we comprehensively analyzed the results of the GEO data analysis and proteomic data. The heatmap displays the levels of eight proteins detected through mass spectrometry analysis in individual samples (Fig.  3 A). Additionally, we assessed the diagnostic accuracy of each signature gene in predicting endometriosis within this proteome validation cohort. Subsequently, ROC analysis was performed on these eight genes, revealing prognostic significance for six of them. The AUC values from the ROC analysis were 0.63 for OLFM4, 0.8 for RRM2, 0.95 for ADIRF, 0.93 for APOBEC3B, 1.0 for GPX3, and 0.93 for LMNB1 (Fig.  3 B–G). Box plots further illustrate the levels of these 6 genes in the control samples compared to those from endometriotic patients (Figs.  3 H–M). Overall, the identified signature genes demonstrate high diagnostic efficiency for endometriosis. Fig. 3 Proteomics reveals the signature proteins in endometriosis. A The heatmap shows the intersection of signature gene expression identified by the GEO dataset and validation set in proteomic datasets. B – G Receiver operating characteristic (ROC) curves depict the diagnostic performance of the signature genes in the endometriosis proteome validation set. H – M The expression levels of signature genes within the proteomic validation set are presented. Data are shown as the median with interquartile range. *P < 0.01; **P < 0.01; ***P < 0.001; ns, nonsignificant Proteomics reveals the signature proteins in endometriosis. A The heatmap shows the intersection of signature gene expression identified by the GEO dataset and validation set in proteomic datasets. B – G Receiver operating characteristic (ROC) curves depict the diagnostic performance of the signature genes in the endometriosis proteome validation set. H – M The expression levels of signature genes within the proteomic validation set are presented. Data are shown as the median with interquartile range. *P < 0.01; **P < 0.01; ***P < 0.001; ns, nonsignificant To further validate these candidate genes, RT-qPCR was performed to assess their expression in ectopic and eutopic endometrial tissues. GPX3 and ADIRF showed consistently significant overexpression in ectopic lesions compared with eutopic tissues (Fig.  4 A). To further confirm GPX3 expression at the protein level, western blotting was performed using normal endometrial tissues and ectopic endometrial tissues. GPX3 protein levels were markedly elevated in ectopic tissues (Fig.  4 B). We found that the expression of GPX3 in the ectopic endometrial tissue is much higher than that of the eutopic endometrial tissue or the normal endometrial tissue. Furthermore, we investigated whether GPX3 is involved in the progression of endometriosis. We constructed the interactome network of GPX3 proteins, which suggested its involvement in antioxidant activity (Fig.  4 C). We isolated primary endometriotic stromal cells and verified their identity by immunofluorescence staining. The cells exhibited strong positivity for the stromal marker Vimentin, whereas the epithelial marker E-cadherin and the macrophage marker CD68 were undetectable (Figure S2). GPX3 was silenced with siRNA in three ectopic endometrial primary cells (Fig.  4 D). Then ROS was detected, and the results showed that GPX3 knockdown significantly promoted ROS accumulation (Fig.  4 E). Similarly, we explored the impact of GPX3 on apoptosis in ectopic endometrial primary cells, revealing that GPX3 suppression did not affect apoptosis (Fig.  4 F). Interestingly, we found that silencing GPX3 significantly inhibited cell migration capacity (Fig.  4 G). These findings indicated the high expression of GPX3 in ectopic endometriotic tissues, its regulation of ROS levels, and the migration capability of endometrial cells. Fig. 4 High expression of GPX3 is associated with malignant progression of endometriosis. A The signature genes were identified by RT-qPCR. Specifically, GPX3 and ADIRF exhibited significant upregulation in ectopic tissue samples from individuals with endometriosis. Data are shown as the median with interquartile range. ***P < 0.001; ns, nonsignificant. B Western blot analysis was employed to assess the expression of GPX3 in normal endometrial tissue and ectopic endometrial tissue. The graphical representation of the statistical analysis demonstrates a significant increase in GPX3 expression in ectopic tissue. The full images of the blots are provided in Figure S3. *P < 0.05. Bars, SD. C The PPI network for GPX3 was constructed in GeneMANIA. The network diagram shows that GPX3 is mainly involved in antioxidant activity. D Western blotting validation confirmed the knockdown of GPX3 in ectopic endometrium primary cells derived from endometriosis tissue. The full images of the blots are provided in Figure S3. E Levels of ROS were quantified in ectopic endometrium primary cells following GPX3 knockdown. The statistical analysis depicted in a graph illustrates an elevation in ROS levels upon GPX3 knockdown. *P < 0.05. F Flow cytometry experiments showed that knocking down GPX3 did not affect the level of ectopic endometrium primary cell apoptosis. G The Transwell assay showed that the knockdown of GPX3 inhibited the migration of endometriosis primary cells. ****P < 0.001. Bars, SD High expression of GPX3 is associated with malignant progression of endometriosis. A The signature genes were identified by RT-qPCR. Specifically, GPX3 and ADIRF exhibited significant upregulation in ectopic tissue samples from individuals with endometriosis. Data are shown as the median with interquartile range. ***P < 0.001; ns, nonsignificant. B Western blot analysis was employed to assess the expression of GPX3 in normal endometrial tissue and ectopic endometrial tissue. The graphical representation of the statistical analysis demonstrates a significant increase in GPX3 expression in ectopic tissue. The full images of the blots are provided in Figure S3. *P < 0.05. Bars, SD. C The PPI network for GPX3 was constructed in GeneMANIA. The network diagram shows that GPX3 is mainly involved in antioxidant activity. D Western blotting validation confirmed the knockdown of GPX3 in ectopic endometrium primary cells derived from endometriosis tissue. The full images of the blots are provided in Figure S3. E Levels of ROS were quantified in ectopic endometrium primary cells following GPX3 knockdown. The statistical analysis depicted in a graph illustrates an elevation in ROS levels upon GPX3 knockdown. *P < 0.05. F Flow cytometry experiments showed that knocking down GPX3 did not affect the level of ectopic endometrium primary cell apoptosis. G The Transwell assay showed that the knockdown of GPX3 inhibited the migration of endometriosis primary cells. ****P < 0.001. Bars, SD We evaluated the signaling pathways associated with GPX3 in endometriosis using GSEA analysis. The results revealed that the leading signaling pathways are shown in Fig.  5 A. The findings indicated a significant correlation between GPX3 and inflammation, including the Jak-stat signaling pathway, chemokine signaling, cytokine receptor interaction, and natural killer cell-mediated cytotoxicity. GPX3 is involved in the construction of tumor immune microenvironment [ 10 , 18 ]. Then, immune cell infiltration was evaluated to determine the association of GPX3 with immunological characteristics. Patients with elevated GPX3 expression, in contrast to those with lower levels, exhibit enhanced immune-related biological processes, including increased cytolytic activity, macrophage, and inflammation (Fig.  5 B). Multiple immunohistochemical staining further validated that ectopic endometrial tissue from patients with elevated GPX3 expression had a higher proportion of PD1-positive immune cells and CD68 + macrophages. In contrast, GPX3 expression had little effect on cell proliferation marker Ki67 (Fig.  5 C). Taken together, GPX3 expression in endometriosis cells is positively associated with the infiltration of immune cells. Fig. 5 GPX3 expression is correlated with immune cell infiltration in the ectopic endometrium tissue. A GSEA analysis revealed that high GPX3 expression was mainly enriched in JAK-STAT signaling and immune-related signaling pathways. B Higher GPX3 expression was associated with increased immune cell infiltration according to the CIBERSORT algorithm analysis. Data are shown as the median with interquartile range. *P < 0.05; **P < 0.01; ***P < 0.001. C Multiplex immunohistochemical staining showed that endometriosis patients with high GPX3 expression had a significantly higher proportion of PD1-positive T cells and CD68 + macrophages in ectopic endometrial tissues GPX3 expression is correlated with immune cell infiltration in the ectopic endometrium tissue. A GSEA analysis revealed that high GPX3 expression was mainly enriched in JAK-STAT signaling and immune-related signaling pathways. B Higher GPX3 expression was associated with increased immune cell infiltration according to the CIBERSORT algorithm analysis. Data are shown as the median with interquartile range. *P < 0.05; **P < 0.01; ***P < 0.001. C Multiplex immunohistochemical staining showed that endometriosis patients with high GPX3 expression had a significantly higher proportion of PD1-positive T cells and CD68 + macrophages in ectopic endometrial tissues

Three GEO datasets GSE25628 , GSE51981 , and GSE7305 were sourced from the Gene Expression Omnibus ( https://www.ncbi.nlm.nih.gov/geo/ ). The R software was utilized to analyze differentially expressed genes between endometriosis and control samples, with adjusted p-value < 0.05 and |log fold change|≥ 1. A volcano plot was created to visualize these DEGs, and a heatmap highlighted the most upregulated and downregulated DEGs. This investigation was approved by the Human Research Ethics Committee of the Xiangyang Central Hospital, and written informed consent was obtained from all participants. All participants were aged 20–49 years old and had not used hormonal drugs for at least 3 months prior to tissue collection. A total of 20 patients with ovarian endometriosis, confirmed by laparoscopy and meeting the American Society for Reproductive Medicine (ASRM) diagnostic criteria, were included. From each patient, both ectopic cyst wall tissue and eutopic endometrial tissue were collected. Among these samples, tissues from 10 patients were used for mass spectrometry analysis, while tissues from the remaining 10 patients were used for experimental validation. Additionally, five normal samples were obtained from women undergoing dilatation and curettage for benign gynecological conditions during the proliferative or secretory phase, with no visible evidence of endometriosis. Control participants were comparable in age and reproductive status to the endometriosis group. For primary cell isolation, the surgically separated tissue was lysed with 0.2% type IV collagenase in a 37 ℃ shaker for 2 h, and the digested tissue mixture was collected and filtered with a 75 μm screen, centrifuged at 500 ×  g for 10 min at 4 ℃. The supernatant was then carefully removed, PBS was added to resuspend, and centrifuged at 4 ℃ and 500 ×  g for 10 min. Then, PBS was removed, RPMI 1640 culture medium (Gibco, Invitrogen), containing 10% fetal bovine serum (Pricella, Wuhan) was added, and the cells were transferred to a cell culture dish. Primary cells of ectopic endometrial were grown in RPMI 1640 medium (Gibco, Invitrogen), enhanced with 10% fetal bovine serum (Pricella, Wuhan), and 1% penicillin/streptomycin (Gibco, Invitrogen). Lipofectamine 3000 (Gibco, Invitrogen) was utilized for the transfection experiment, following the instructions provided by the manufacture. The GPX3 siRNA sequence is 5′-GTGGAGGCTTTGTCCCTAATT-3′. Two machine learning algorithms were applied to screen hub genes, the support vector machine (SVM) model and the least absolute shrinkage and selection operator (LASSO). The overlapping genes identified by these two machine-learning algorithms were considered the hub genes for endometriosis. The diagnostic efficiency of these hub genes was evaluated using the area under the curve (AUC) of the receiver operating characteristic (ROC) curve. Proteome identification was carried out using Data-Independent Acquisition (DIA) MS, following the procedure we previously outlined [ 15 ]. Briefly, samples were dissolved in a 2% SDS lysis buffer and subjected to tryptic digestion using the FASP method. The resulting tryptic peptides were then reconstituted in 0.1% formic acid. These peptides were subsequently analyzed with an Orbitrap Fusion Lumos mass spectrometer (Thermo). Quantification was performed using Proteome Discoverer (Thermo Fisher Scientific) and Spectronaut (Biognosys) software. GSEA was conducted across various subgroups based on GPX3 expression levels. The c2.cp.kegg.v7.5.symbols.gmt gene set was used for the reference. Significant KEGG pathways were identified and presented according to the Net Enrichment Score (NES), gene ratio, and p-value. Gene sets with |NES|> 1 and a nominal adjusted p-value of < 0.05 were deemed significantly enriched. Western blot analysis was performed following the procedure previously described in our study [ 16 ]. Briefly, cells were lysed using Cell Lysis Buffer (Cell Signaling Technology, Boston, MA), and protein quantification was performed with a BCA Protein Assay Kit (Fdbio Science, Hangzhou). Equal amounts of protein (30 μg per sample) were mixed with an equal volume of loading buffer, boiled at 100 °C for 10 min, and separated via SDS-PAGE. The proteins were subsequently transferred onto a 0.22 μm polyvinylidene difluoride (PVDF) membrane (Millipore Sigma). After blocking with 5% non-fat milk at room temperature for one hour, the membrane was incubated overnight at 4 °C with primary antibodies, including GPX3 (Proteintech, Wuhan, dilution 1:1000, 13947-1-A, RRID: AB_3085426), Tubulin (Cell Signaling Technology, USA, dilution 1:1000, #2144, RRID: AB_2210548). After washing three times with TBST, membranes were incubated at room temperature for 1 h with HRP-linked goat anti-rabbit or anti-mouse IgG secondary antibodies (Servicebio, Wuhan, dilution 1:2000, RRID: AB_2892100, RRID: AB_2910572). Following three washes with TBST, chemiluminescent signals were developed using the SuperFemto ECL Chemiluminescence Kit (Vazyme Biotech Co., Ltd., China, E423). Target protein bands were visualized using a luminescence imaging system (Bio-Rad). Cell apoptosis assay was using the Annexin V-FITC/PI Apoptosis Detection Kit (KeygenBioTECH, KGA1102-50). Cells were gently collected and centrifuged to remove the culture medium. The cell pellet was resuspended in Annexin V-FITC staining buffer, followed by the addition of propidium iodide (PI) staining solution with gentle mixing. The samples were incubated at room temperature for 20 min in the dark. Flow cytometry was then performed to detect Annexin V–FITC and PI fluorescence. ROS levels were measured using the Reactive Oxygen Species Assay Kit (S0033S, Beyotime). After removing the culture medium, cells were washed twice with PBS. The DCFH-DA working solution prepared in PBS was added, and cells were incubated at 37 °C for 20 min. Subsequently, cells were washed three times with PBS to remove excess DCFH-DA. Cells were gently collected, and ROS levels were analyzed by flow cytometry. Cells were fixed with 4% paraformaldehyde for 30 min at room temperature and permeabilized with 0.05% Triton X-100 for 10 min. After permeabilization, cells were blocked with 5% bovine serum albumin (BSA) for 1 h at room temperature. Subsequently, cells were incubated with the appropriate primary antibodies overnight at 4 °C, followed by three washes with TBST. The following primary antibodies were used: Vimentin (Abclonal, Wuhan, A19607, RRID: AB_2862696), E-cadherin (Abclonal, Wuhan, A20798, RRID: AB_3107194), and CD68 (Abclonal, Wuhan, A23205, RRID: AB_3094597). Fluorescent secondary antibodies were applied for one hour at room temperature in the dark. Nuclei were counterstained with DAPI (Keygen BioTECH, KGA1523-25), and fluorescence images were captured via confocal microscopy. Total RNA was extracted from ectopic and eutopic endometrial tissue samples using TRIzol reagent (Yeasen Biotechnology, 19201ES60) according to the manufacturer’s protocol. RNA concentration was assessed using a NanoDrop spectrophotometer (Thermo Fisher Scientific). For cDNA synthesis, 1 µg of total RNA was reverse-transcribed using a cDNA synthesis kit (Yeasen Biotechnology, 11141ES60) according to the manufacturer’s instructions. Quantitative PCR (qPCR) was then performed with SYBR Green PCR Master Mix (Yeasen Biotechnology, 11202ES08), using 0.5 µL of the synthesized cDNA per reaction. β-actin served as the internal control for mRNA quantification, and relative gene expression levels were calculated using the 2 ⁻ΔΔCt method. Primer sequences are listed in Table  1 . Table 1 Primer sequence Target gene Sequence (5′–3′) ADIRF Forward: CACCCAGGAAACCATCGACA Reverse: CTAGATGGCGCCTGGAAGG GPX3 Forward: ATGGGCAATCCCCAGATGGAC Reverse: AGCTGGCCACGTTGACAAA LMNB1 Forward: CGCGTGCGTGTCTATGCTAA Reverse: GGCTTCCAACTGGGCAATCT OLFM4 Forward: AAATGCTCGAGAGTTGCGGA Reverse: CACAGCAATCGTGTTGGTGG RRM2 Forward: AAGAAGAAGGCAGACTGGGC Reverse: CCAGGCATCAGTCCTCGTTT APOBEC3B Forward: ACCGCACGCTAAAGGAGATT Reverse: CGGGTCCAACTCGTTGCATA β-actin Forward: ACGTGGACATCCGCAAAG Reverse: GACTCGTCATACTCCTGCTTG Primer sequence Primary cells of endometriosis underwent transfection with the specified treatment and were cultured for another 48 h. Afterward, cells were collected, and 1 × 10 5 cells suspended in 100 µL of serum-free medium were seeded into the upper chamber of a Transwell ® insert (Corning). The lower chamber contained 600 µL of RPMI 1640 medium supplemented with 10% FBS. Following a 12-h culture, the cells were fixed using methanol and stained with crystal violet. The cells remaining in the upper chamber were carefully removed with a cotton swab. Random five fields were then selected to count the migrated cells. All experiments were conducted in triplicate. The paraffin-embedded tissue was deparaffinization using the dewaxing solution, followed by dehydration using a series of ethanol concentrations. Then, antigen retrieval was carried out with citrate buffer (pH 6.0) and subjected to high-pressure boiling for 2 min, and then allowed to cool to room temperature. Endogenous peroxidase activity was quenched with hydrogen peroxide, followed by blocking with 3% BSA for 30 min. The slides were left to incubate with primary antibodies at 4 °C overnight. Afterward, they were washed three times with PBS for 5 min each and then treated with secondary antibodies for 50 min. Tyramide signal amplification was applied for 10 min at room temperature, followed by three 5-min washes with PBS. To remove nonspecifically bound antibodies, the slides were treated in a microwave for 10 min. All slides were then stained with the following antibodies: CD68 (Servicebio, GB113150 , RRID: AB_2924885) visualized with Opal440 TSA; Ki67 (Servicebio, GB151499 , RRID: AB_3677419) visualized with Opal647 TSA; and PD1 (Servicebio, GB12338, RRID: AB_3675596) visualized with Opal546 TSA. DAPI incubation for 10 min, then rinsed three times with PBS. Following, the slides were sealed with an anti-fade mounting medium to prevent fluorescence quenching. Finally, the sections were imaged and analyzed using 3DHISTECH slide scanners (Pannoramic MIDI). All statistical analyses in this study were conducted using GraphPad Prism 8.0 and R software (Version 4.1.2). For normally distributed data, comparisons between two groups were performed using the independent t-test. Data were obtained from at least three independent experiments and are presented as the mean ± standard deviation (SD). For non-normally distributed variables and clinical data, the Wilcoxon rank-sum test or Chi-square test was applied as appropriate. Data are presented as the median with interquartile range [ 17 ]. Receiver operating characteristic (ROC) curve analysis was used to assess diagnostic performance. A p-value < 0.05 was considered statistically significant for all analyses.

In this study, we performed an integrative analysis of both GEO and proteomic data, revealing GPX3 as a potential biomarker that can promote the progression of endometriosis. We observed markedly elevated GPX3 expression in the ectopic tissue from endometriosis patients. Functional experiments demonstrated that GPX3 knockdown reduced ROS levels and impaired the migratory capacity of ectopic endometrial cells, without significantly affecting apoptosis. Nevertheless, additional investigation is needed to clarify the exact mechanisms through which GPX3 influences the development of endometriosis. ROS are metabolic products that act as signaling molecules regulating gene expression and cell proliferation [ 19 ]. In endometriosis, iron overload from retrograde menstruation generates highly reactive free radicals via the Fenton reaction, causing peritoneal damage and inflammation and promoting lesion formation [ 20 ]. ROS levels are elevated in ectopic endometrial tissues, enhancing cell proliferation and migration and inducing lipid peroxidation, DNA damage, and inflammatory mediator expression, thereby establishing sustained oxidative stress. Under this pressure, increased superoxide triggers compensatory upregulation of antioxidant enzymes such as SOD and GPXs [ 21 ], which partially protect ectopic cells from ROS-induced apoptosis. However, this enhanced antioxidant response also grants lesion cells a survival advantage in the iron-rich, hypoxic, and inflammatory microenvironment, enabling continued growth. Thus, excessive ROS drives lesion initiation and progression, while antioxidant systems like GPXs, despite reducing oxidative injury, simultaneously support the persistence of endometriotic cells under oxidative stress [ 22 ]. Although the precise pathogenesis remains ambiguous, the current study revealed that immune-related mechanisms of dysregulation play a key role in the onset and progression of endometriosis [ 23 – 25 ]. In general, endometriotic lesions are composed of epithelial, stromal, endothelial, glandular, and immune cell components. Compared to normal endometrium, the immune-inflammatory characteristics of endometriotic lesions are altered [ 25 ]. Endometriosis cells were able to escape peritoneal cavity immune surveillance and promote the growth of endometriosis lesions [ 26 ]. Endometriosis is also associated with peritoneal inflammation, manifested by increased macrophages, inflammatory cytokines, growth factors, and proangiogenic factors in peritoneal fluid [ 27 ]. Interestingly, through bioinformatic analysis, we found a relationship between GPX3 and the infiltration of immune cells in the ectopic endometrial tissue of endometriosis. We found that the correlation between increased GPX3 levels and the presence of macrophages and PD1-positive immune cells. It has been reported that women with endometriosis exhibit a significant increase in macrophage concentration in peritoneal fluid [ 6 , 27 ]. Macrophages play a role in identifying foreign and damaged cells in the abdominal cavity, and dysfunctional peritoneal macrophages in endometriosis patients may secrete growth factors and cytokines, potentially facilitating the survival of ectopic endometrial cells [ 27 ]. Similarly, alterations in the production of cytokines by helper T lymphocytes may contribute to changes in the viscosity of the peritoneal fluid, creating a favourable environment for the proliferation of ectopic endometrial tissue [ 28 , 29 ]. In our study, greater macrophage infiltration in patients with ectopic endometriosis who have high GPX3 expression suggests a potential role of GPX3 as an immunomodulatory factor. It should be noted that the present study has several limitations. First, given the relatively small sample size, the diagnostic accuracy of GPX3 may have been overestimated, which represents an important limitation. Therefore, validation in larger and independent cohorts will be essential to confirm the robustness and generalizability of these biomarkers. Second, considering the highly complex nature of ROS regulation in endometriosis, GPX3 may predominantly modulate localized rather than global oxidative stress within lesions, a possibility that requires further clarification. In addition, we did not directly measure ROS levels in ectopic tissues, nor did we quantify the proportions of CD68⁺ macrophages and PD-1⁺ T cells in animal models or human samples. These gaps limit our ability to fully establish a causal relationship between GPX3-mediated ROS modulation and immune evasion in vivo. Future studies will therefore focus on systematically validating the role of GPX3 in immune regulation and lesion progression using an endometriosis animal model, as well as performing in vivo quantification of immune-cell subsets.

In conclusion, GPX3 may serve as a potential biomarker for diagnosing and treating endometriosis, and we demonstrated that knockdown of GPX3 increased ROS levels and suppressed the migratory ability of endometrial cells. Moreover, we found that GPX3 expression levels were positively correlated with immune cell infiltration, including macrophages and PD1-positive immune cells, in endometriotic lesions.

Endometriosis is a localized chronic inflammatory condition in gynecology, marked by the presence of functional endometrial tissue, including stroma and glands, outside the myometrium and the uterine cavity [ 1 , 2 ]. It predominantly affects women of reproductive age, with an incidence ranging from 10 to 15% [ 3 ]. In recent years, the incidence of endometriosis has risen significantly, yet the underlying mechanisms of its development are still not fully elucidated [ 3 ]. Increasing evidence suggests that ectopic endometrial lesions can trigger oxidative stress responses, leading to local inflammatory reactions and alterations in the peritoneal microenvironment [ 3 – 5 ]. This can result in neutrophilic inflammatory infiltration and induce the production of pain-inducing factors such as macrophages, enhance the activity of pain-related proteases, and promote lipid peroxidation reactions, thus causing pelvic pain [ 5 , 6 ]. Excessive reactive oxygen species (ROS) release further regulates protein activity and gene expression, causing cellular damage and functional changes [ 6 , 7 ]. The glutathione peroxidase (GPX) family is fundamental in managing the redox balance [ 8 , 9 ]. Notably, glutathione peroxidase 3 (GPX3) is the only secretory member of the GPX family, playing a key role in protecting cells from ROS damage in the microenvironment [ 10 – 12 ]. Accumulating studies have demonstrated that GPX3 can function either as a tumor suppressor or promoter depending on the biological context [ 13 , 14 ]. However, its function in endometriosis has not been reported. In this study, we integrated public endometriosis-related datasets ( GSE7305 , GSE51981 , and GSE25628 ) from the Gene Expression Omnibus (GEO) with proteomic profiling of eutopic, ectopic, and normal endometrial tissues to identify key molecules involved in endometriosis. Given that GPX3 has not been previously reported in the context of endometriosis, the present study aimed to characterize its expression pattern and explore its potential functional relevance.

Supplementary Material 1. Supplementary Material 1.