Bioinformatics analysis implicates ferroptosis and key hub genes in the pathogenesis of endometriosis

Objective Endometriosis (EMs) is a chronic gynaecological condition characterised by the ectopic growth of endometrial tissue; however, its molecular mechanisms remain insufficiently understood. Ferroptosis, an iron-dependent form of regulated cell death, has been suggested as a potential contributor to its pathogenesis. This study aimed to identify differentially expressed ferroptosis-related genes (DE-FRGs) in EMs through bioinformatics analysis and to explore their underlying molecular mechanisms.

The gene expression datasets GSE7305 and GSE25628 were obtained from the GEO database. Ferroptosis-related genes (FRGs) were extracted from the FerrDb database. DE-FRGs were identified by intersecting differentially expressed genes (DEGs) with FRGs across the two datasets. Protein-protein interaction (PPI) networks were constructed using STRING and Cytoscape software, while core genes were identified through the cytoHubba plugin. Functional enrichment analysis was performed via the KEGG pathway.

A total of 11 DE-FRGs were identified, all of which demonstrated consistently upregulated expression in EMs tissues across both datasets. Four core genes – GATA6, KLF2, BGN and AEBP1 – were selected for further analysis owing to their significant enrichment. KEGG analysis indicated that these genes were particularly enriched in the ‘taurine and hypo-taurine metabolism’ and ‘muscle cell cytoskeleton’ pathways.

This study identified GATA6, KLF2, BGN, and AEBP1 as potential core genes associated with ferroptosis in EMs, highlighting their roles in metabolic and cytoskeletal pathways. These findings provide a novel perspective on the pathogenesis of EMs and suggest new therapeutic targets that warrant further experimental validation. PLAIN LANGUAGE SUMMARY Endometriosis is a prevalent condition characterised by the growth of tissue resembling the endometrium outside the uterus, resulting in pain and infertility. This study employed computational techniques to analyse genetic data and identified four pivotal genes that may affect a cell death process referred to as ‘ferroptosis’ in endometriosis. These genes are associated with metabolic and cellular structural alterations. Elucidating the functions of these genes will facilitate the development of novel therapeutic approaches for treating endometriosis in the future. GRAPHICAL ABSTRACT

Endometriosis (EMs) is a common chronic gynaecological disorder defined by the presence of endometrial-like tissue outside the uterine cavity, predominantly affecting the ovaries and peritoneum (Arrigo Fruscalzo et al. Citation2025). This condition often manifests as dysmenorrhoea, chronic pelvic pain, and infertility (Xie et al. Citation2025), significantly diminishing patients’ quality of life and work capacity. Epidemiological data suggest that EMs affects approximately 10% of women of reproductive age globally, with a notably higher prevalence of up to 35% among those experiencing infertility (Arrigo Fruscalzo et al. Citation2025). Furthermore, an increasing incidence has been reported in younger populations. Current diagnostic methods primarily utilise ultrasonography and serum CA-125 measurements, while treatment typically involves surgical intervention in conjunction with gonadotropin-releasing hormone agonists (GnRH-a) (Hussaini et al. Citation2024). Nonetheless, recurrence rates remain alarmingly high, underscoring the limitations of existing therapeutic strategies. Although EMs is histologically benign, it exhibits malignant-like characteristics such as invasive growth and recurrence. Its aetiology is multifactorial, encompassing genetic, immunological, endocrine, inflammatory, and environmental factors (Zhan and Wu Citation2025). Various hypotheses have been proposed to elucidate the origin of ectopic lesions, including retrograde menstruation, coelomic metaplasia, lymphatic and vascular dissemination, and stem cell involvement (Nakamura et al. Citation2024). Ferroptosis is a recently identified form of regulated cell death that is distinct from apoptosis and necrosis, characterised by the iron-dependent accumulation of lipid hydroperoxides (Zhou et al. Citation2023). This process arises from the dysregulation of cellular lipid metabolism, iron homeostasis, and glutathione-dependent antioxidant pathways (Lu et al. Citation2022). Emerging evidence suggests that ferroptosis significantly contributes to the pathogenesis and treatment of various gynaecological malignancies, including cervical, endometrial, and ovarian cancers (Tang and Chen Citation2024). Its role in endometriosis has increasingly attracted research attention. Notably, endometriotic lesions consistently demonstrate iron overload and oxidative stress (Yi et al. Citation2022), which closely align with the hallmark features of ferroptosis. Despite the growing body of evidence indicating a potential role for ferroptosis in EMs, our current understanding remains in its infancy. The systematic involvement of ferroptosis-related mechanisms in the pathogenesis of this condition has yet to be clarified. Previous research has demonstrated that endometrial cells in EMs develop an adaptive capacity to evade ferroptosis (Zubrzycka et al. Citation2020), a regulated form of cell death that could otherwise inhibit the survival, implantation, and establishment of ectopic lesions following retrograde menstruation (Ng et al. Citation2020). Nevertheless, the specific molecular mechanisms that facilitate the evasion of ferroptosis and its causal role in the initiation and progression of EMs are still inadequately understood. Moreover, the relationship between disrupted iron homeostasis and the subsequent pathophysiology of the disease necessitates further exploration. In light of these knowledge gaps, we propose the following hypothesis: Aberrant regulation of ferroptosis serves as a crucial mechanism in the pathogenesis of endometriosis, whereby endometrial cells acquire a survival advantage by evading ferroptosis, thus promoting the establishment and progression of ectopic lesions. This study aims to systematically investigate the role of ferroptosis-related pathways in endometriosis, elucidate their functional significance in disease development, and offer novel insights into the underlying mechanisms, ultimately striving to identify potential therapeutic targets for this challenging condition.

Data acquisition The gene expression profile data for ectopic and normal endometrial tissues were retrieved from the Gene Expression Omnibus (GEO) database. (https://www.ncbi.nih.gov/geo/) by searching the keyword ‘endometriosis’. The inclusion criteria were as follows: (1) samples derived from human tissue (Homo sapiens); (2) samples containing both in situ endometrial and normal population endometrial microarrays; (3) all samples in the dataset sourced from the Affymetrix Human Genome U133 Array. The exclusion criteria included: (1) exclusion of adenomyosis; (2) samples derived from cell lines; (3) studies focused on non-coding RNA; (4) studies that failed to clearly distinguish between in situ/ectopic endometrium or controlled abnormal endometrium. Ferroptosis-related genes (FRGs) were sourced from the FerrDb database (https://www.zhounan.org/ferrdb/v3/pages/index.html), encompassing ferroptosis drivers, markers, inhibitors, and unclassified factors. Following data cleaning, which involved the elimination of duplicate entries, the correction of gene symbols, and the retention of genes with complete annotations, a distinct set of FRGs (n = 1673) was identified for further analysis. All data are publicly accessible. Identification of differentially expressed ferroptosis-related genes (DE-FRGs) Following the screening of the GEO database for datasets that met the specified criteria, patient information within these datasets was meticulously verified. Annotations were subsequently made for both the case group and the normal control group. The differential expression of genes (DEGs) between ectopic endometrium and normal endometrium in each dataset was analysed using GEO2R, the integrated differential analysis tool provided by GEO. The GEO2R tool conducts rigorous statistical analyses utilising the R/Bioconductor package ‘limma’ (linear model for microarray data). To minimise false-positive rates and ensure reliable outcomes, the threshold for identifying significantly differentially expressed genes (DEGs) was established at |log2 fold change (FC)| ≥ 2 and an adjusted P-value < 0.05. The adjusted P-value was computed employing the Benjamini & Hochberg (BH) method for the correction of multiple testing. A volcano plot was generated to visualise the gene expression profile, with upregulated genes denoted in red and downregulated genes in blue, where each point represents an individual gene. The x-axis indicates the log2 FC, reflecting the magnitude of expression change, while the y-axis represents the -log10(adjusted P- value), signifying the statistical significance. Subsequently, an online Venn diagram tool (https://bioinfogp.cnb.csic.es/tools/venny/index.html) was employed to identify the shared differentially expressed genes (DEGs) between the two datasets and the ferroptosis-related gene (FRG) dataset, referred to as DE-FRGs. A heatmap illustrating the expression patterns of the top-ranked differentially expressed genes (DEGs) across all samples was generated using the online platform for data analysis and visualisation available at https://www.bioinformatics.com.cn (last accessed on 10 Dec 2024) (Tang et al. Citation2023). Visualisation of DE-FRGs expression in datasets validation of hub gene expression in GSE datasets To visualise the expression patterns of DE-FRGs in EMs, we plotted box plots to compare their expression levels in normal endometrial and ectopic endometrial tissues, with analysis performed using the Mann-Whitney U test. The box plot analysis clearly demonstrated intergroup expression differences, where the median, interquartile range (IQR), and upper boundary represented the distribution of gene expression values in each group. These results confirmed that the selected DE-FRGs exhibited consistent differential expression between normal and ectopic endometrium, suggesting their potential involvement in the pathogenesis of EMs. Construction and module analysis of protein-protein interaction networks (PPI) The STRING (Search Tool for the Retrieval of Interaction Genes/Proteins) database (https://string-db.org/) was utilised with a predefined threshold of confidence > 0.4 to construct a protein-protein interaction (PPI) network encoding DE-FRGs, thereby gaining deeper insights and predicting the cellular functions and biological behaviours of the identified genes. Additionally, the PPI network was visualised using Cytoscape software (version 3.10.2). To identify hub genes, the protein-protein interaction network was analysed using the cytoHubba plugin (version 0.1) in Cytoscape. Genes were ranked by four topological algorithms: maximum central connectivity (MCC), maximum neighbourhood component (MNC), degree, and maximum neighbourhood component density (DMNC). The top ranked genes from each algorithm were intersected as final hub genes. Given the moderate size of the network and our focus on global centrality measures, we did not perform module detection prior to hub gene identification. Functional enrichment analysis of DE-FRGs The Kyoto Encyclopaedia of Genes and Genomes (KEGG) pathway analysis method was employed to identify major metabolic and signalling pathways associated with DE-FRGs. P-values <0.05 were considered statistically significant for KEGG enrichment. Ethical approval and consent Ethical approval and consent to participate were not required for this study. All data analysed were obtained from the GEO database (Accession: GSE7305 and GSE25628), which is a publicly available repository. The original studies contributing data to GEO have obtained necessary ethical approvals and informed consent from participants, as documented in their respective publications.

Study design The overall workflow of this study is illustrated in . Briefly, gene expression profiles were obtained from the GEO database (GSE7305 and GSE25628), and ferroptosis-related gene sets were retrieved from the FerrDb database. DEGs were identified through differential expression analysis, followed by intersection with FRGs to obtain DE-FRGs. A protein-protein interaction network was then constructed using the STRING database, and core genes were identified using Cytoscape software. Finally, KEGG pathway enrichment analysis was performed to explore the potential biological pathways associated with the core genes. A comprehensive summary of the analytical workflow and key findings is presented in . Data screening This study selected two micro array datasets, GSE7305 and GSE25628, that met the criteria. Patients with EMs confirmed by pathology were defined as the case group, while normal endometrium from non-endometriosis patients and the in-situ endometrium from patients with confirmed EMs were defined as the normal control group. GSE25628 included 8 cases and 6 normal samples, whereas GSE7305 included 10 cases and 10 normal samples. Identification of DE-FRGs Differential analysis of the dataset GSE25628 identified 3,901 differentially expressed genes, including 1,194 down-regulated genes and 2,707 up-regulated genes. The GSE7305 dataset contains 430 differentially expressed genes, with 222 downregulated and 208 upregulated genes. The intersection of the differential expression genes (DEGs) from the two datasets and the ferroptosis dataset yielded 11 shared DE-FRGs, which were visualised using a Venn diagram (as shown in the ). To visually characterise the expression patterns of the 11 DE-FRGs identified in our study, we generated heatmaps based on their normalised expression levels in the GSE7305 and GSE25628 datasets. As shown (), the 11 genes exhibited distinct expression profiles between EMs and normal endometrial tissues. Unsupervised hierarchical clustering of samples revealed two major clusters that corresponded closely to the disease and control groups, indicating that the combined expression signature of these DE-FRGs effectively distinguishes EMs tissues from normal tissues. Consistent with our previous validation results, the majority of DE-FRGs – including Adipocyte enhancer-binding protein 1 (AEBP1), Biglycan (BGN), Fatty acid binding protein 4 (FABP4), Flavin containing dimethylaniline monoxygenase 1 (FMO1), Frizzled class receptor 7 (FZD7), Growth arrest specific 1 (GAS1), GATA binding protein 6 (GATA6), Krüppel-like factor 2 (KLF2), PDZ and LIM domain 3 (PDLIM3), and Regulator of G protein signalling 2 (RGS2) – showed markedly higher expression in EMs samples across both datasets, whereas NTRK2 displayed no consistent difference between the two groups. The clear segregation of sample groups based on DE-FRG expression further supports the potential involvement of these genes in EMs pathogenesis and highlights their value as candidate diagnostic biomarkers. Validation of the expression of hub genes in the GSE datasets To validate the expression patterns of the 11 identified DE-FRGs, we analysed two independent microarray datasets, GSE7305 and GSE25628. As shown in , the majority of these genes exhibited significantly higher expression levels in EMs tissues compared to normal controls. Specifically, AEBP1, FABP4, FMO1, FZD7, GAS1, GATA6, KLF2, PDLIM3, and RGS2 showed highly significant upregulation (P < 0.001 or P = 0.0001–0.0002), while BGN also demonstrated significantly increased expression (P = 0.0044). In contrast, NTRK2 did not show a statistically significant difference between EMs and normal tissues (P = 0.7155), suggesting it may not be a key player in this context or that its expression is more variable. The consistent upregulation of the remaining 10 DE-FRGs across two independent datasets strengthens the reliability of our initial screening and supports their potential involvement in the pathogenesis of EMs. These findings warrant further experimental validation to elucidate the functional roles of these genes in ferroptosis-mediated disease progression. Construction and module analysis of PPI network To explore the functional interactions among the differentially expressed ferroptosis-related genes, we constructed a protein-protein interaction (PPI) network using the STRING database. As illustrated in , the network revealed complex interconnections among the DE-FRGs, suggesting potential synergistic roles in ferroptosis regulation. To identify the most significant functional clusters, we performed module analysis using the cytoHubba plugin in Cytoscape. The top-ranked module, consisting of four hub genes. These hub genes include GATA binding protein 6 (GATA6), Krüppel-like factor 2 (KLF2), Biglycan (BGN), and AE binding protein 1 (AEBP1). Notably, all four genes exhibited high node degrees and were centrally positioned within the network, indicating that they may serve as key regulators of ferroptosis in the context of EMs. Their identification provides a focused set of candidates for further functional validation and mechanistic studies. DE-FRGs functional enrichment analysis To gain insights into the biological functions of the differentially expressed ferroptosis-related genes (DE-FRGs), we performed pathway enrichment analysis using the KEGG database. As illustrated in , the DE-FRGs were significantly enriched in two pathways: ‘Taurine and hypo taurine metabolism’ and ‘Cytoskeleton in muscle cells’. Both pathways exhibited statistical significance (p < 0.05), with enrichment scores expressed as –log10(p-value) of 1.72 and 1.80, respectively. These pathways are known to be involved in the regulation of oxidative stress and cytoskeletal dynamics – processes closely linked to ferroptosis. The enrichment of DE-FRGs in these pathways suggests that ferroptosis may exert its effects in the context of EMs partly through modulating taurine metabolism and cytoskeletal organisation.

Ferroptosis is a recently identified form of iron-dependent programmed cell death that is crucial in various pathological processes, including tumours and neurodegenerative diseases (Liang et al. Citation2025). Nonetheless, its precise role in EMs remains inadequately elucidated. This study utilised bioinformatics approaches to identify potential pathways and key genes associated with the development of EMs, thereby offering a novel perspective on the molecular mechanisms underlying this condition and highlighting the potential role of ferroptosis in its pathogenesis. Previous research has established the involvement of ferroptosis in the pathogenesis and progression of EMs. By integrating data from the FerrDb database, we identified 11 differentially expressed ferroptosis-related genes (DE-FRGs) between EMs tissues and normal endometrial tissues, utilising the GSE25628 and GSE7305 datasets. This approach elucidated the molecular characteristics of these genes within the context of EMs. Using the Cytoscape and cytoHubba plugins, we identified four key hub genes: GATA6, KLF2, BGN, and AEBP1. GATA6 is a pivotal hub gene identified in this study, belonging to the GATA transcription factor family. Its involvement in EMs has previously been associated with the establishment of progesterone resistance, a hallmark of the disease, as evidenced by low methylation and elevated expression in ectopic stromal cells (Dyson et al. Citation2014). Furthermore, GATA6 has been demonstrated to promote pyroptosis via the PI3K/AKT pathway, thereby facilitating lesion growth (Zhao et al. Citation2022). However, its direct role in ferroptosis within the context of EMs remains to be investigated. Drawing on evidence from other disease models, where GATA6 was shown to inhibit neuronal ferroptosis through the miR-193b/ATG7 axis (Fan et al. Citation2023), we propose a context-dependent mechanistic hypothesis regarding its function in EMs. We hypothesise that GATA6 may exert a dual, or even opposing, regulatory effect on ferroptosis, contingent upon the specific cellular microenvironment. For instance, in ectopic endometrial stromal cells, where the PI3K/AKT pathway is hyperactivated, GATA6 may promote cell survival by suppressing ferroptosis. This could occur through the direct upregulation of antioxidant systems, such as the SLC7A11-GPX4 axis, or by activating the NRF2-mediated transcriptional programme, a master regulator of anti-ferroptotic defences. Conversely, within the highly inflammatory environment of EMs lesions, GATA6 may indirectly sensitise cells to ferroptosis by upregulating the expression of pro-inflammatory factors, such as IL-1β, through the NLRP3 inflammasome, which can facilitate lipid peroxidation. Future investigations employing gain- and loss-of-function strategies in EMs cell lines, in conjunction with specific ferroptosis modulators, including ferrostatin-1 and RSL3, are crucial for elucidating these context-dependent mechanisms and determining whether GATA6’s primary role in EMs is to promote or to protect against ferroptotic cell death. KLF2, a transcription factor that plays a critical role in inflammation, oxidative stress, and angiogenesis, has previously been implicated in EMs through its negative regulation of mesenchymal–epithelial transition and adhesion in endometrial stromal cells via AKT phosphorylation (Yin et al. Citation2025). However, its direct involvement in ferroptosis within the context of EMs remains unexamined. Emerging evidence from other disease models indicates a complex, context-dependent role for KLF2 in the regulation of ferroptosis. In osteomyelitis and osteoarthritis, KLF2 demonstrates a protective effect by transcriptionally activating GPX4, engaging the SIRT1 signalling pathway, and maintaining mitochondrial integrity, thereby inhibiting ferroptosis (Shi et al. Citation2024, Sun et al. Citation2025). Conversely, in prostate and colon cancers, KLF2 functions as a tumour suppressor by promoting ferroptosis through the inhibition of the PI3K/AKT pathway or by repressing the SLC7A11/GPX4 axis (Zhang et al. Citation2024, Li et al. Citation2023). We propose that the divergent roles of KLF2 in ferroptosis can be reconciled by examining the distinct cellular states and microenvironments associated with EMs. Within EMs, ectopic endometrial stromal cells display hyperactivated PI3K/AKT signalling and altered cell adhesion properties, characteristics that may redirect KLF2 towards a pro-ferroptotic function. We hypothesise that, in the context of EMs, KLF2 may facilitate ferroptosis by antagonising the PI3K/AKT pathway, a crucial survival signal, thereby sensitising ectopic cells to lipid peroxidation. This pro-ferroptotic effect could be further enhanced by KLF2-mediated transcriptional repression of SLC7A11 or GPX4, as observed in certain malignancies. Alternatively, considering the chronic inflammatory environment of EM lesions, KLF2 may initially seek to exert its anti-ferroptotic, cytoprotective role to uphold cellular homeostasis; however, this protective mechanism may be overwhelmed by persistent oxidative stress and pro-inflammatory signals, ultimately shifting the balance towards ferroptotic cell death. To test these hypotheses, systematic gain- and loss-of-function experiments in EM cell models will be necessary, alongside the use of specific ferroptosis modulators and targeted manipulation of the PI3K/AKT and SIRT1 pathways. Furthermore, given the established relationship between KLF2, AKT phosphorylation, and cell adhesion, future studies should investigate how the interplay between ferroptosis and epithelial–mesenchymal transition dynamics contributes to the establishment and persistence of ectopic lesions. BGN encodes a small leucine-rich proteoglycan that serves as a crucial component of the extracellular matrix (ECM). Recent evidence has identified BGN as a negative regulator of ferroptosis across various malignancies. In gastric cancer, the long non-coding RNA ZEB1-AS1 enhances BGN expression by sponging miR-429, thereby inhibiting ferroptosis induced by erastin and RSL3, as demonstrated by decreased levels of Fe2+, malondialdehyde, and lipid reactive oxygen species (Wu et al. Citation2023). n trastuzumab-resistant HER2-positive breast cancer, the circular RNA circ-BGN directly interacts with OTUB1 and SLC7A11, promoting OTUB1-mediated deubiquitination of SLC7A11 and stabilising the protein, which ultimately inhibits ferroptosis; notably, the knockdown of circ-BGN restores sensitivity to ferroptosis (Liu et al. Citation2023). Collectively, these studies position BGN as a significant suppressor of ferroptosis, primarily through the post-translational stabilisation of SLC7A11 or indirect upregulation via competing endogenous RNA networks. Although direct evidence for BGN-mediated regulation of ferroptosis in endometriotic lesions is currently lacking, we propose a mechanistic hypothesis based on the distinctive pathophysiological characteristics of EMs. Endometriotic tissues are marked by extensive extracellular matrix (ECM) remodelling, chronic inflammation, and progesterone resistance – conditions that frequently coincide with enhanced cellular survival mechanisms. We hypothesise that BGN may play a significant anti-ferroptotic role in ectopic endometrial cells. Specifically, we suggest that BGN, likely upregulated by local inflammatory cytokines or hypoxia-inducible factors within the EM microenvironment, promotes cell survival by stabilising the SLC7A11 protein. This stabilisation may occur through direct interaction with deubiquitinating enzymes such as OTUB1, akin to the mechanism observed in breast cancer, thereby sustaining glutathione biosynthesis and counteracting lipid peroxidation. Furthermore, the upregulation of BGN in EMs may be driven by dysregulated competing endogenous RNA networks involving specific microRNAs (e.g. miR-429 or miR-193b), which are known to be aberrantly expressed in EMs. This model positions BGN as a critical node linking ECM signalling, inflammatory stress, and resistance to ferroptosis. To test this hypothesis, functional experiments in EM cell lines and primary ectopic stromal cells will be necessary, including BGN knockdown and assessment of SLC7A11 ubiquitination status. Additionally, the upstream regulators of BGN in EMs require investigation to fully elucidate the molecular circuitry that confers ferroptosis resistance in ectopic lesions. AEBP1, a transcriptional repressor, has recently been identified as a central regulator of ferroptosis across various disease models, demonstrating context-dependent functional diversity. In glioblastoma, the toad venom derivative arenobufagin induces ferroptosis by upregulating miR-149-5p, which subsequently suppresses AEBP1 expression, leading to a reduction in reactive oxygen species, Fe2+, and malondialdehyde accumulation (Hu et al. Citation2025). In contrast, during brain ischemia–reperfusion injury, AEBP1 promotes neuronal ferroptosis by directly repressing PRKCA transcription, thereby inhibiting the PI3K–Akt survival pathway (Zhang et al. Citation2024). In oral cancer, silencing AEBP1 enhances sulfasalazine-induced ferroptosis, potentially through the activation of the JNK/p38/ERK pathway, and suppresses tumour growth in vivo (study on cisplatin-resistant cells). In lung adenocarcinoma, the transcription factor YY1 directly activates AEBP1 transcription, while AEBP1 silencing inhibits proliferation, migration, and angiogenesis, concurrently inducing ferroptosis (Zhou et al. Citation2022). Moreover, in a chronic unpredictable mild stress model and LPS-treated microglia, AEBP1 promotes ferroptosis and M1 polarisation by binding to the DDR2 promoter and activating the STAT3/P53 signalling axis (Zou et al. Citation2025). Collectively, these findings establish AEBP1 as a multifunctional regulator of ferroptosis, capable of both suppressing and promoting this process depending on the cellular context and upstream signalling. Although direct evidence for AEBP1-mediated regulation of ferroptosis in EMs is currently lacking, we propose a mechanistic hypothesis based on the distinctive pathophysiological characteristics of EMs. We suggest that AEBP1 primarily functions as a suppressor of ferroptosis in ectopic endometrial cells, thereby providing a survival advantage within the inhospitable peritoneal microenvironment. This pro-survival role may be realised through at least three convergent mechanisms. First, in light of the chronic inflammatory environment associated with EM lesions, AEBP1 may be upregulated by local inflammatory cytokines or hypoxia-inducible factors, subsequently promoting resistance to ferroptosis by repressing PRKCA and attenuating PI3K–Akt signalling, a pathway typically linked to cell survival. Second, AEBP1 may directly activate the DDR2/STAT3/P53 axis, as observed in neuroinflammatory models, which could exacerbate inflammatory responses and M1 macrophage polarisation. This activation may sensitise ectopic cells to inflammatory stress while paradoxically shielding them from ferroptotic cell death through yet-to-be-defined compensatory mechanisms. Third, AEBP1 expression in EMs may be driven by upstream transcription factors such as YY1, which is known to be dysregulated in EMs, thereby establishing a positive feedback loop that sustains AEBP1-mediated resistance to ferroptosis. Furthermore, considering the established role of AEBP1 in extracellular matrix remodelling and fibrosis, we propose that AEBP1 may connect ferroptosis resistance to the fibrotic progression of EM lesions, thus contributing to both the establishment and persistence of these lesions. Functional enrichment analysis identified two significantly enriched pathways: ‘taurine and hypotaurine metabolism’ and ‘cytoskeleton in muscle cells’. The enrichment of the taurine and hypotaurine metabolism pathway is particularly noteworthy in the context of EMs. Taurine, a cysteine-derived amino sulphonic acid, serves as a vital regulator of oxidative stress, osmotic balance, and membrane integrity (Liu et al. Citation2024). It effectively scavenges reactive oxygen species, and its metabolic disruption is known to sensitise cells to oxidative injury (Heidari et al. Citation2013). Considering that ferroptosis is driven by lipid peroxidation (Wang et al. Citation2025), we hypothesise that dysregulation of taurine metabolism in EMs may directly undermine the antioxidant capacity of ectopic endometrial cells, thereby reducing the threshold for ferroptotic cell death. This mechanism could influence the survival or clearance of ectopic lesions, contingent upon the balance between pro-oxidant and antioxidant forces within the local microenvironment. Consequently, the enrichment of this pathway suggests a potential metabolic vulnerability that could be therapeutically targeted to selectively induce ferroptosis in ectopic tissues. The second enriched pathway, ‘cytoskeleton in muscle cells’, includes essential cytoskeletal components, such as actin-binding proteins. The involvement of the hub gene BGN in the structure of the extracellular matrix (ECM) further highlights the significance of cytoskeletal and ECM dynamics in the pathogenesis of EMs (He et al. Arendt et al. Citation2025). Processes critical to lesion formation – including adhesion, invasion, and epithelial–mesenchymal transition – are fundamentally reliant on cytoskeletal remodelling (Wang et al. Citation2024). We hypothesise that the DE-FRGs identified in this study may regulate the migratory and invasive capabilities of ectopic cells not only through canonical ferroptosis pathways but also by modulating cytoskeletal integrity. For example, BGN, as a key ECM component, may affect cytoskeletal organisation via integrin-mediated signalling, thereby connecting structural support to ferroptosis sensitivity. Although the number of genes enriched in these pathways is limited, their significant convergence suggests potential mechanistic crosstalk between the regulation of ferroptosis and the key pathological features of EMs. These findings establish a basis for functional validation studies aimed at elucidating how metabolic and structural pathways integrate to govern ferroptosis in ectopic lesions and at investigating their potential as therapeutic targets. When interpreting KEGG enrichment results, caution is essential. Firstly, the issue of false positives in enrichment analysis presents a common challenge in bioinformatics research. The accuracy of such analyses is heavily reliant on the quality of the input gene list and the chosen background database. Although this study implemented stringent threshold screening for differential genes and utilised corrected P-values, the number of differentially expressed ferroptosis-related genes (DE-FRGs) ultimately included in the enrichment analysis was limited to only 11. The use of a smaller gene set for enrichment analysis may render the results highly sensitive to variations among individual genes, thereby increasing the likelihood of incidental discovery of significant pathways, which constitutes false positives. Consequently, our discussion centres on pathways that exhibit significant P-values and demonstrate potential biological relevance as documented in the existing literature. Secondly, this study identified only the aforementioned two pathways as significantly enriched, while other pathways commonly reported in the pathogenesis of endometriosis (EM), such as the ‘HIF-1 signalling pathway’, ‘PI3K-Akt signalling pathway’, or ‘oestrogen signalling pathway’, were not enriched. This phenomenon may be elucidated by several factors. Primarily, the specificity of the analytical focus is noteworthy. This study concentrated on ‘ferroptosis-related genes’, and the enrichment analysis highlighted the pathways most likely involved with these specific genes (DE-FRGs), rather than encompassing all pathways associated with EM. This suggests that ferroptosis-related genes may contribute to EM through a relatively specific mechanism, such as influencing redox balance and the cytoskeleton, rather than directly engaging with classical hormonal or inflammatory pathways. Secondly, there is insufficient statistical power. Due to the limited number of input genes, even if DE-FRGs include members of certain classical pathways, the inadequate gene coverage may hinder them from achieving the threshold for statistically significant enrichment. Consequently, the inability to enrich classical pathways does not diminish their significance in the overall pathogenesis of EMs; rather, it suggests that these pathways are not the primary focus of enrichment within the DE-FRGs list of this study. This study has several limitations. Firstly, the bioinformatics analysis relied on a restricted public dataset with a relatively small sample size, potentially introducing selection bias. Secondly, the findings were analysed solely at the transcriptomic level and lack validation through protein-level and functional experiments. The specific mechanisms of the identified hub genes in endometriosis, particularly their role in influencing disease progression via the regulation of ferroptosis, necessitate further investigation through in vitro cell experiments and in vivo animal models. Lastly, this study primarily concentrated on differentially expressed genes and did not comprehensively incorporate information regarding epigenetic regulation and post-translational modifications. In summary, this study identified four core ferroptosis-related hub genes – GATA6, KLF2, BGN, and AEBP1 – that are significantly upregulated in EMs tissues, as determined through integrative bioinformatics analysis. Functional enrichment analysis indicated that the pathways of ‘taurine and hypotaurine metabolism’ and ‘cytoskeleton in muscle cells’ may serve as crucial mechanistic links between the regulation of ferroptosis and the pathogenesis of EMs. Based on these findings, we propose that these hub genes may modulate ferroptosis sensitivity in ectopic endometrial cells through context-dependent mechanisms, which involve the modulation of antioxidant capacity (GATA6, KLF2), the stabilisation of SLC7A11 (BGN), and the transcriptional reprogramming of survival pathways (AEBP1). While these mechanistic hypotheses necessitate experimental validation, they establish a theoretical framework that integrates dysregulation of iron metabolism, metabolic adaptation, and cytoskeletal dynamics with the hallmark features of EMs, including progesterone resistance, chronic inflammation, and enhanced cell survival. Despite the inherent limitations associated with bioinformatics analyses – such as the potential for false positives, reliance on publicly available datasets with relatively small sample sizes, and the lack of functional validation – these findings provide a novel theoretical foundation for understanding the molecular interactions between ferroptosis and the pathogenesis of EMs. The identified hub genes and enriched pathways present prioritised candidates for subsequent experimental investigation. Future studies should validate the expression patterns of these hub genes in larger, prospectively collected cohorts of EMs tissue using quantitative real-time PCR and immunohistochemistry. Functional gain- and loss-of-function experiments in primary ectopic endometrial cells and patient-derived organoid models, in conjunction with ferroptosis-specific modulators, are vital for elucidating the precise molecular mechanisms through which GATA6, KLF2, BGN, and AEBP1 regulate ferroptosis, proliferation, migration, and invasion in ectopic lesions. Furthermore, assessing the diagnostic utility of these genes in minimally invasive samples, such as peripheral blood or peritoneal fluid, alongside investigating their therapeutic potential via preclinical targeting strategies, will be crucial for clinical translation. By constructing a mechanistic framework that links susceptibility to ferroptosis with key pathological processes in endometriotic lesions, this study significantly enhances the molecular understanding of the disease and establishes a foundation for the development of diagnostic biomarkers and targeted therapeutic strategies aimed at overcoming treatment resistance and minimising recurrence. Supplemental material Highlights.docx Download MS Word (13.7 KB)Highlights.docxAcknowledgments Each author has made substantial contributions to the conception or design of the work, or the acquisition, analysis, or interpretation of data; has drafted or critically revised the manuscript for important intellectual content; has given final approval of the version to be published; and agrees to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. Disclosure statement No potential conflict of interest was reported by the author(s). Data availability statement The data that support the findings of this study are openly available in public repositories. The Gene Expression Omnibus (GEO) datasets GSE25628 and GSE7305 can be accessed via the following URLs: GSE25628: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE25628 GSE7305: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE7305. Additionally, the Ferroptosis Database (FerrDb V3) is available at https://www.zhounan.org/ferrdb/v3/pages/index.html. All data were derived from the National Centre for Biotechnology Information (NCBI) GEO database and the FerrDb V3 repository, which are in the public domain. Additional information Funding

- Arendt, W., et al., 2025. Endometriosis and cytoskeletal remodeling: the functional role of actin-binding proteins. Cells, 14 (5), 360–360. - Dyson, M.T., et al., 2014. Genome-wide DNA methylation analysis predicts an epigenetic switch for GATA factor expression in endometriosis. PLoS Genetics, 10 (3), e1004158. - Fan, W., et al., 2023. GATA6 inhibits neuronal autophagy and ferroptosis in cerebral ischemia-reperfusion Injury through a miR-193b/ATG7 axis-dependent mechanism. Neurochemical Research, 48 (8), 2552–2567. - Fruscalzo, A., et al., 2025. Comprehensive approaches to endometriosis management. Journal of Clinical Medicine, 14 (12), 4324–4324. - Heidari, R., Babaei, H. and Eghbal, M.A., 2013. Cytoprotective effects of taurine against toxicity induced by isoniazid and hydrazine in isolated rat hepatocytes. Arhiv za Higijenu Rada i Toksikologiju, 64 (2), 15–24. - Hu, R., et al., 2025. Arenobufagin induces ferroptosis in glioblastoma cells via modulating the MiR‐149‐5p/AEBP1 axis. Journal of Applied Toxicology: JAT, 45 (4), 606–619. - Hussaini, A., et al., 2024. Management of endometriosis-related pain: Comparing the effectiveness of hormonal therapy, surgical interventions, and complementary therapies. Cureus, 16 (12), e75590. - Li, J., et al., 2023. KLF2 inhibits colorectal cancer progression and metastasis by inducing ferroptosis via the pi3k/akt signaling pathway. The Journal of Pathology. Clinical Research, 9 (5), 423–435. - Liang, L., et al., 2025. Integrative multi-omics analysis reveals the interaction mechanisms between gut microbiota metabolites and ferroptosis in rheumatoid arthritis. Frontiers in Immunology, 16, 1608262. - Liu, H., et al., 2024. Therapies for cirrhotic cardiomyopathy: current perspectives and future possibilities. International Journal of Molecular Sciences, 25 (11), 5849. - Liu, R., Zhou, Y. and Cao, Y., 2023. CircRNA and ferroptosis in human disease: insights for new treatments. Animal Models and Experimental Medicine, 6 (6), 508–517. - Lu, C., et al., 2022. Ferroptosis in intracerebral hemorrhage: a panoramic perspective of the metabolism, mechanism and theranostics. Aging and Disease, 13 (5), 1348–1364. - Nakamura, A., et al., 2024. mTOR inhibitors as potential therapeutics for endometriosis: a narrative review. Molecular Human Reproduction, 30 (12), gaae041. - Ng, S.-W., et al., 2020. Endometriosis: the role of iron overload and ferroptosis. Reproductive Sciences (Thousand Oaks, Calif.), 27 (7), 1383–1390. - Shi, J., et al., 2024. KLF2 inhibits ferroptosis and improves mitochondrial dysfunction in chondrocyte through SIRT1/GPX4 signaling to improve osteoarthritis. Drug Development Research, 85 (7), e70015. - Sun, J., et al., 2025. Potential Involvement of the KLF2–GPX4 Axis in Ferroptosis during S.aureus-induced Osteomyelitis. Human Genomics, 19 (1), 138. - Tang, D., et al., 2023. SRplot: a free online platform for data visualization and graphing. PloS One, 18 (11), e0294236–e0294236. - Tang, S. and Chen, L., 2024. The recent advancements of ferroptosis of gynecological cancer. Cancer Cell International, 24 (1), 351. - Wang, J., et al., 2025. Urinary betaine/creatinine ratio enhances diagnostic accuracy of PLA2R-Ab in idiopathic membranous nephropathy. BMC Nephrology, 27 (1), 35. - Wang, Z., et al., 2024. CADM2 participates in endometriosis development by influencing the epithelial-mesenchymal transition. Reproductive Sciences (Thousand Oaks, Calif.), 31 (10), 3049–3057. - Wu, C., et al., 2023. Long noncoding RNA ZEB1‐AS1 attenuates ferroptosis of gastric cancer cells through modulating miR‐429/BGN axis. Journal of Biochemical and Molecular Toxicology, 37 (8), e23381. - Xie, Z., et al., 2025. Identification of biomarkers for endometriosis based on summary-data-based Mendelian randomization and machine learning. Medicine, 104 (14), e41804. - Yi, Z., et al., 2022. Baicalein relieves ferroptosis-mediated phagocytosis inhibition of macrophages in ovarian endometriosis. Current Issues in Molecular Biology, 44 (12), 6189–6204. - Yin, S., et al., 2025. Excessive KLF2 impairs mesenchymal–epithelial transition efficiency and adhesion in endometrial stromal cells via AKT phosphorylation. Reproductive Biomedicine Online, 51 (6), 105123. - Zhan, J. and Wu, J., 2025. Immunomodulatory insights of monoterpene glycosides in endometriosis: immune infiltration and target pathways analysis. Hereditas, 162 (1), 1. - Zhang, C., et al., 2024. MiR-15b-3p weakens bicalutamide sensitivity in prostate cancer via targeting KLF2 to suppress ferroptosis. Journal of Cancer, 15 (8), 2306–2317. - Zhang, Y., Li, Y. and Liu, F., 2024. AEBP1 Silencing protects against cerebral ischemia/reperfusion injury by regulating neuron ferroptosis and microglia M2 polarization through PRKCA‐PI3K‐Akt axis. Drug Development Research, 85 (8), e70032. - Zhao, S., et al., 2022. IGF2BP2 promotes the progression of ovarian endometriosis by regulating m6A-modified MEIS2 and GATA6. The International Journal of Biochemistry & Cell Biology, 152 (106296), 106296–106296. - Zhou, J., et al., 2023. Ferroptosis contributes to hemolytic hyperbilirubinemia-induced brain damage in vivo and in vitro. Molecular Medicine Reports, 28 (6), 236. - Zhou, Q., et al., 2022. Inhibition of AEBP1 predisposes cisplatin-resistant oral cancer cells to Ferroptosis. BMC Oral Health, 22 (1), 478. - Zou, Q., et al., 2025. AEBP1 targets DDR2 to induce ferroptosis and microglial activation through the STAT3/P53 signaling pathway, exacerbating depression Risk. Archives of Biochemistry and Biophysics, 774, 110632. - Zubrzycka, A., et al., 2020. Genetic, epigenetic, and steroidogenic modulation mechanisms in endometriosis. Journal of Clinical Medicine, 9 (5), 1309.