Integrated Identification and Immunotherapy Response Analysis of the Prognostic Signature Associated With m6A, Cuproptosis‐Related, Ferroptosis‐Related lncRNA in Endometrial Cancer

Endometrial cancer (EC) stands as the most common gynecological malignancy impacting the female reproductive system on a global scale [1]. While early‐stage EC patients generally exhibit favorable prognoses following treatment [2, 3], the survival rates of those in advanced, metastatic, or recurrent stages significantly decrease [3, 4]. The mortality rate of EC is growing at a faster pace than its incidence rate, resulting in an estimated 76 000 female deaths globally each year [5]. The escalation in disease‐related mortality correlates with advanced stage, aggressive histology, and the occurrence of metastasis [6]. For patients with EC who are deemed suitable candidates for surgical intervention, standard treatment recommendations include total hysterectomy and bilateral salpingo‐oophorectomy, followed by lymph node staging [7]. However, late‐stage EC patients who have already received standard management often lack effective treatment options [8]. Despite progress and collaboration among various disciplines and institutions, the development and success of prognostic biomarkers remain limited [9]. Hence, it is imperative to identify predictive and prognostic features in high‐risk patients to further enhance the overall prognosis of individuals with EC.

Long noncoding RNAs (lncRNAs), characterized by their inability to encode proteins and lengths exceeding 200 nucleotides, are integral to a myriad of cellular functions [10]. Despite not encoding proteins directly, lncRNAs are considered to have essential functions in numerous cellular processes through various mechanisms [11]. Abnormal expression of lncRNAs in many tumors can result in tumor progression, metastasis, drug resistance, and other processes [12, 13]. Evidence suggests lncRNA features can effectively predict the prognosis of EC patients and further improve their survival rates [14, 15]. N6‐methyladenosine (m6A) is a prevalent RNA modification that widely exists in lncRNA, which participates in regulating the process of RNA encoding proteins [16, 17]. It exerts a significant influence on oncogenesis, growth, and metastasis. Studies have shown that m6A methylation can function as a biological indicator for early tumor diagnosis and prognosis evaluation, contributing to the improvement of patients' survival outcomes [18, 19, 20]. Furthermore, m6A modification has the capacity to regulate tumor microenvironment penetration and immune system activation, potentially impacting the efficacy of immunotherapy.

Ferroptosis and cuproptosis are both recently discovered regulated forms of programmed cell death. Ferroptosis is a cell death pathway caused by the accumulation of cytotoxic lipid peroxides regulated by iron dependency [21, 22]. The interplay of ferroptosis in tumorigenesis and resistance to systemic therapies is underscored by the role of specific proteins; for example, ferroptosis suppressor protein 1 (FSP1), which attenuates ferroptotic cell death by modulating levels of the lipid peroxidation‐associated molecule, coenzyme Q10 (CoQ10) [23]. Additionally, among all gynecological cancers, EC is particularly associated with metabolic abnormalities such as obesity, hyperglycemia, and hypertension [24]. Studies suggest that metabolic changes during obesity may render EC more sensitive to ferroptosis inducers [25].Copper, an indispensable micronutrient, is a catalyst and structural component essential for a plethora of biochemical reactions vital to life, encompassing energy metabolism, antioxidant defenses, and mitochondrial respiration [26]. However, an overabundance of copper can precipitate adverse outcomes. In various cancer types, including EC, heightened copper concentrations have been documented [27]. The oxidative stress engendered by copper may induce angiogenesis in EC, leading to the proliferation of new vasculature from a previously dormant state, thus fueling tumor progression [28]. Excess copper ions can also interact with lipid‐acylated components within the tricarboxylic acid (TCA) cycle, culminating in protein aggregation and the disintegration of iron–sulfur cluster proteins, which initiate proteotoxic stress and cell death [29]. This interaction triggers a cascade of toxic protein stress, culminating in cellular demise. This mechanism highlights the critical role of copper homeostasis in cellular metabolism and its potential impact on pathological states [26, 29]. Notably, targeting the regulation of ferroptosis‐related proteins has become an effective therapeutic approach to induce apoptosis in tumor cells, particularly for radioresistant and drug‐resistant malignant tumors [30, 31]. Adverse effects may occur if the copper content in the body passes the certain limit that body can tolerate. Elevated copper levels were observed in various types of malignancy, including EC [27]. Therefore, ferroptosis and cuproptosis related signatures may be the potential targets for the development of promising biomarkers.

During the development of endometrial carcinoma, ferroptosis and cuproptosis may also be influenced by m6A modifications on lncRNAs. However, the association between m6A‐modified lncRNAs and ferroptosis, as well as cuproptosis, in the onset and progression of EC remains unexplored. In this study, based on bioinformatics approach, we identified lncRNAs co‐expressing with cuproptosis‐, ferroptosis‐, and m6A‐related lncRNAs from expression data of EC. By constructing the prognosis model in EC, we screened hub lncRNA signatures affecting prognosis of EC patients. Furthermore, the guiding value of m6A‐modified ferroptosis‐related lncRNA (mfrlncRNA) features was assessed in terms of prognosis, immune microenvironment, and drug sensitivity. Our finding would provide valuable reference for effective targeted therapy and prognostic assessment of EC.

Despite advancements in surgical techniques and chemotherapy regimens for EC, the prognosis for patients often remains poor, significantly impacting patient survival. Besides, a subset of women diagnosed with EC, particularly those of reproductive age, may suffer significant psychological impacts from the loss of fertility if standard treatments involving salpingo‐oophorectomy are uniformly applied [41]. In recent years, the safety of vitrification techniques for human oocytes and embryos has been widely validated [42]. The development and maturation of these technologies have propelled advancements in assisted reproductive technologies (ART), providing more options for EC patients who wish to preserve their fertility [43]. Nevertheless, the effectiveness of fertility‐sparing treatments can be influenced by the patient's physiological characteristics and the progression of the tumor [44, 45]. Therefore, developing optimized molecular prognostic and predictive models for EC patients is crucial. These models not only facilitate the rapid identification of prognostic markers and assessment of mortality risk, thereby aiding in clinical decision‐making, but they also allow for personalized treatment plans based on the molecular characteristics of early‐stage EC [46]. This approach can evaluate the feasibility of fertility preservation techniques during cancer treatment, offering hope and support to patients wishing to maintain their fertility. In addition, increasingly evidences indicated that the cuproptosis‐, ferroptosis‐, and m6A methylation‐related lncRNAs played the potential roles in predicting the prognosis of cancers [47]. Therefore, we utilize cuproptosis‐, ferroptosis‐, and m6A methylation‐related lncRNAs to construct prognostic models for EC patients, aiming to identify biomarkers that may influence the prognosis of EC patients. These models may not only help in the rapid identification of prognostic markers and assessment of mortality risk, providing a comprehensive risk evaluation and guiding tailored treatment strategies, but also facilitate the formulation of personalized treatment plans based on the molecular characteristics of early‐stage EC. This approach can evaluate the feasibility of fertility preservation techniques during cancer treatment, may offer hope and support to patients wishing to maintain their reproductive potential.

In our investigation, we extracted co‐expressed lncRNA expression data from the TCGA cohort to identify critical signatures through a series of analytical methods: univariate Cox regression, LASSO Cox regression, and stepwise Cox regression analyses. Although we attempted to utilize machine learning methods to construct predictive models, our comparative analysis revealed that these models did not outperform the LASSO Cox regression model. Research indicates that due to the complexity of real‐world data, the Cox proportional hazards regression model provides predictive accuracy comparable to that of machine learning models while offering advantages in variable interpretability [48]. Consequently, we opted to use the LASSO Cox regression model to develop the prognostic model for EC patients. This comprehensive approach allowed us to uncover relationships between identified genes, clinical characteristics, and potential immunotherapy effects. We successfully established a prognostic model based on five multifunctional lncRNAs (mfclncRNAs)—FZD10‐AS1, CDKN2B‐AS1, LINC01480, BX322234.1↗, and SENCR—and calculated the model's risk scores. These signatures have been implicated in EC, reinforcing the relevance of our findings. Previous literature has highlighted the prognostic significance of these lncRNAs in various cancers. For instance, Li et al. recognized FZD10‐AS1 as a prognostic marker in colorectal cancer through the construction of a competitive endogenous RNA network [49]. This study's observation that FZD10‐AS1 is downregulated in high‐risk populations mirrors our own findings, suggesting its broader applicability across cancer types. Similarly, LINC01480's abnormal expression has been linked to the progression of multiple tumors, including its role in activating the PI3K/AKT pathway, a critical route for cell proliferation and inhibition of apoptosis [9, 50, 51, 52, 53]. This pathway's potential as a therapeutic target for EC underscores the importance of our results [51, 54]. Moreover, BX322234.1↗ 's role in regulating genomic instability and its high expression in high‐risk groups point to its significance as a biomarker for aggressive tumor behavior [3, 55]. Therefore, further investigating the biological functions related to BX322234.1↗ would be of significant interest and importance. The oncogenic potential of CDKN2B‐AS1 across different cancers, including its regulatory effects on miR‐378b in lung cancer, further validates the multifaceted roles of these lncRNAs in oncogenesis [56, 57]. Lastly, SENCR's involvement in vascular development and its newly discovered contribution to cisplatin resistance in non‐small cell lung cancer highlight the expansive impact of lncRNAs on cancer biology and treatment resistance [58, 59, 60]. However, the role and underlying mechanisms of SENCR in malignant tumors remain largely unexplored [61]. Our research supports the need for further in‐depth investigations into this lncRNA as a potential factor influencing cancer susceptibility.

The functional enrichment analysis delineated the pivotal roles of mfclncRNAs, predominantly associated with tumor progression and immune responses. Notably, within the GO terms, endopeptidase activity emerged as a crucial enzymatic function, facilitating protein synthesis, folding, and protection by cleaving peptide bonds in proteins. This activity plays a critical role in thwarting tumor progression by maintaining protein homeostasis [62]. Additionally, the analysis underlined biological processes mainly associated with inhibiting protease and hydrolase activities. According to KEGG pathway insights, the Wnt signaling pathway stands out for its integral role in both embryonic development and the emergence of tumors [63, 64].

To corroborate the validity of the risk score within our prognostic model, K‐M curves were generated, revealing significant survival discrepancies between the high‐risk and low‐risk groups. Further analytical rigor was provided by plotting the C‐index and its time‐dependent variant, both of which underscored the prognostic model's risk scores as a potent individual prognostic factor, offering distinct advantages over traditional clinical characteristics. To extend the validation process, a testing set was employed to craft a predictive nomogram alongside corresponding calibration curves. These tools not only affirmed the risk score's predictive capacity for the prognosis of EC patients but also facilitated a deeper investigation into the prognostic model's efficacy.

In exploring the gene mutation profiles across different risk groups, we found that PTEN has a lower mutation rate in the low‐risk group with better prognosis. Conversely, ARID1A and TP53 exhibit higher mutation rates in the high‐risk group with poorer prognosis. This finding corroborates the results reported by Gullo et al. [65]. In our endeavor to elucidate the functional impact of immune cell infiltration on EC, we delved into the association between prognostic lncRNA signatures and the extent of immune cell penetration within the tumor microenvironment. Our findings reveal that EC patients characterized by elevated ES, enhanced IS, and diminished TP are correlated with improved survival rates and more favorable prognostic indicators. The tumor microenvironment of EC is a complex network composed of a variety of cell types, including endothelial cells, fibroblasts, myofibroblasts, alongside key immune and inflammatory cells. Consistent with prior research, our analysis substantiates that the presence of tumor‐infiltrating immune cells plays a significant role in determining the prognosis of EC, underscoring the critical interplay between immune infiltration and cancer progression [66, 67]. Our findings excitingly reveal that, through risk stratification, patients categorized as high risk exhibit a pronounced infiltration of CD8⁺ cells. As crucial entities for eradicating cancer cells, CD8⁺ cytotoxic T lymphocytes are instrumental in both the development and assessment of immunotherapeutic strategies for combating cancer. This observation underscores the vital role these cells play in influencing patient outcomes. Conversely, individuals identified as low risk display more robust immune responses, correlating with superior prognostic results. Further examination of the relationship between risk scores and immune functionalities uncovered that high‐risk patients exhibit reduced activities in Type II IFN response, T‐cell co‐stimulation, T‐cell co‐inhibition, cytolytic activity, and antigen‐presenting cell (APC) co‐stimulation, whereas Type I IFN response remains notably active. The synergistic modulation of Type I and Type II responses plays a critical role in activating pro‐inflammatory pathways, while also initiating inhibitory feedback mechanisms that challenge cancer control efforts [68]. Particularly, the Type I IFN response, known for regulating both self‐immunity and antitumor activities, may under certain conditions impair self‐immune reactions or activate feedback loops that limit effective immune responses [69, 70]. The downregulation of Type II IFN response suggests a potential mechanism for immune evasion, leading to less favorable outcomes for high‐risk EC patients [8]. Moreover, our research aligns with studies on various cancer types, demonstrating that TMB is a predictive marker for the efficacy of immune checkpoint inhibitors [71, 72]. Significantly, EC patients with elevated TMB levels were associated with better prognoses, with the low‐risk group displaying higher TMB values. Analysis of TIDE scores between the high‐risk and low‐risk groups indicated a greater likelihood of immune escape phenomena in the former. Collectively, these findings highlight that the low‐risk group benefits from increased antitumor immune activities, offering insight into the prognostic significance of the identified features.

In addition, we evaluated the effects of immunotherapy. Cancer immunotherapy has demonstrated efficacy in treating various malignant tumors by either activating systemic immune responses or restoring immune defects caused by tumors [73]. Several case reports have shown the effectiveness of immune checkpoint inhibitors targeting receptors such as PD‐1, PD‐L1, and CTLA4 in late‐stage or metastatic EC patients [74, 75]. Based on pharmacological sensitivity analysis, our findings indicate that 5z‐7‐oxozeaenol, dabrafenib, trametinib, nutlin‐3a, and PD‐0332991 exhibit heightened sensitivity in patients categorized as high‐risk, whereas YM155 demonstrates increased sensitivity in those classified under the risk group. The variance in drug sensitivity is correlated with the molecular characteristics and genetic mutations specific to these patient groups. The loss of PTEN activates the PI3K/AKT signaling pathway, augmenting the dependency of cells on survivin. Under these circumstances, the application of the survivin inhibitor YM155 can more effectively suppress the proliferation of tumor cells [76]. Within cells harboring TP53 mutations that modulate the p53 pathway, the TAK1 inhibitor 5Z‐7‐oxozeaenol can restore apoptotic functions in a subset of cells by inhibiting the NF‐κB signaling pathway, which bolsters the antitumor potency of chemotherapy agents [77]. Loss of PTEN boosts PI3K/AKT signaling, potentially leading to resistance against BRAF and MEK inhibitors like dabrafenib and trametinib [78]. PD‐0332991, a CDK4/6 inhibitor, enhances therapeutic outcomes by disrupting cell cycle regulation in tumor cells, particularly those lacking TP53‐mediated apoptosis and repair [79]. Recent clinical trials of trametinib and dabrafenib for EC patients showed partial responses in three patients [80, 81]. Clinical investigations suggest that PD‐0332991, a selective cyclin‐dependent kinase 4/6 inhibitor, has shown promise in treating EC, suggesting its potential applicability in a broader range of malignancies [82]. In summary, these pharmacological discoveries enable us to integrate prognostic models to assess risk factors in diverse patient populations, thereby facilitating a more tailored and personalized approach to treatment. However, further clinical trials are essential to refine these strategies.

Our study demonstrates notable strengths, including the use of gene expression profiles and clinical data from two public platforms, enhancing our prognosis model's reliability and validity. The model's risk scores were validated internally, reinforcing their role as potential prognostic indicators. Despite unequal baselines between training and test sets, both validations supported the risk scores' predictive value. Additionally, we analyzed the biofunction and immune response of identified signatures, offering insights into potential EC therapies. However, our study is not without limitations. A notable challenge was the incomplete clinical data from TCGA for EC, particularly the absence of TNM staging information. This gap underscores the need for comprehensive clinical data to strengthen future research. Moreover, to solidify the prognostic utility of the identified mfclncRNAs in EC, further validation through corresponding cellular and animal experiments is essential. These steps are crucial for translating our findings into practical therapeutic strategies.

In our investigation, we pinpointed five pivotal lncRNAs in EC via the development of a prognostic model. These lncRNAs' associated risk scores were confirmed as effective predictors of prognostic outcomes in EC patients. Subsequent evaluations incorporating clinical features, functional enrichment studies, and assessments of immunotherapy responses underscored the significant promise of these lncRNA signatures as innovative therapeutic and diagnostic markers for EC.

Yongkang Qian: conceptualization, methodology, software, formal analysis, resources, writing – original draft. Hualing Chen: software, data curation, validation, project administration. Pengcheng Miao: methodology, software, validation. Rongji Ma: investigation, software, validation. Beier Lu: methodology, visualization. Chenhua Hu: project administration, software. Ru Fan: project administration, supervision. Biyun Xu: conceptualization, writing – review and editing. Bingwei Chen: conceptualization, supervision, writing – review and editing.

The authors declare no conflicts of interest.

We also thank the Big Data Computing Center of Southeast University for providing the facility support on the numerical calculations in this paper.

All data utilized in this research are accessible through the TCGA database, available at (https://portal.gdc.cancer.gov/↗).