Unveiling DNA methylation: early diagnosis, risk assessment, and therapy for endometrial cancer.

Endocrine therapy, particularly high-dose progesterone, has been widely used in ER/progesterone receptor (PR)-expressing early-stage EC patients, but progesterone resistance still occurs in some cases. Moreover, the response rates are poor in advanced and recurrent women. Repeated use of progesterone may lead to loss of PR and eventually treatment resistance ( 108 ). In addition, some epigenetic mechanisms and signaling pathways are also involved in the formation of progesterone insensitivity. CpG islands exist in the promoter region and first exon of PR gene, where DNA methylation is abnormally active in EC, endometriosis, breast cancer and other hormone-related diseases. Studies have shown that aberrant DNA methylation brings about suppressed or even absent expression of PR gene in tumors ( 108 , 109 ). PRA and PRB, two isoforms of PR, play different roles in progesterone therapy. PRB works by activating gene transcription, while PRA interferes with the therapeutic effect via inhibiting PRB function. Interestingly, PRA and PRB are not methylated simultaneously, suggesting independent regulation of PRA and PRB methylation processes ( 110 , 111 ). In breast cancer, low PR levels were significantly associated with poor prognosis, and the methylation of PRA , rather than PRB , worsened tamoxifen treatment outcomes ( 112 ). More than a decade ago, it was demonstrated that DNMT inhibitors (DNMTi) combined with endocrine therapy were effective against constitutive-resistant breast carcinoma with high DNMT levels ( 113 ). In EC, PR gene methylation was related to congenital progesterone resistance ( 114 ). A significantly higher frequency of PR promoter methylation was observed in metastatic tumor than in primary lesion ( 115 ), posing a barrier to hormone therapy in advanced metastatic cases. Small molecule DNMTi sensitized poorly differentiated PRB-negative EC to progesterone therapy ( 116 ). Moreover, Jones et al. observed that HAND2 methylation levels were significantly higher in non-cancerous hyperplastic endometrium that did not respond to progesterone for 3 months compared with endometrial lesions that subsided after treatment, suggesting that progesterone responsiveness was reliant upon HAND2 expression ( 51 ). Chemotherapy is one of the main treatment strategies for post-surgery patients and those with advanced or recurrent EC. Commonly used chemotherapy drugs include paclitaxel, platinum, doxorubicin, and topotecan. Drug resistance, significantly reducing progression-free survival (PFS) and overall survival (OS), is related to gene mutation, microenvironment change, abnormal regulation of signaling pathways and other factors, covering variants in DNA methylation patterns ( 14 ). The checkpoint with forkhead-associated and ring finger (CHFR) protein acts as a cell cycle checkpoint component to delay the entry of cells into mitosis by diminishing cyclin dependent kinase 1 (CDK1) activity. When tubulin homeostasia is disrupted by agents such as paclitaxel, CHFR-positive cells stall in the G2 phase and even retreat from early mitosis, allowing them to escape, to some extent, cell death ( 117 ). The CHFR gene is often inactivated by methylation in cancer cells, and its methylation is associated with poor prognosis and increased sensitivity to paclitaxel in multiple cancer types, including ovarian and gastric cancers ( 117 , 118 ). In paclitaxel-sensitive EC cells, CHFR gene hypermethylation occurred more frequently, and restoring CHFR expression could reduce cell sensitivity to paclitaxel ( 119 , 120 ). Zhou et al. showed that phosphoglycerate kinase 1 (PGK1) mediated upregulation of DNMTs (DNMT1, DNMT3A and DNMT3B) through the HSP90/ERK pathway, leading to increased methylation levels and enhancing cisplatin resistance in EC ( 121 ) ( Figure 2 ). Fialkova et al. examined the changes of promoter DNA methylation in apoptosis-associated genes and observed the influences of BCL2L11 , CIDEB and GADD45A methylation in endometrial carcinogenesis, which may contribute to deregulation of mitochondrial apoptotic pathway and development of pro-apoptotic drug resistance ( 122 ). Underlying mechanisms of DNA methylation involved in therapy resistance of endometrial cancer. (A) Abnormal methylation of PR and HAND2 leads to downregulation of their protein expression, affecting the interaction between progesterone and PR, resulting in resistance of endometrial cells to progesterone. (B) The CHFR protein delays cell entry into mitosis by inhibiting CDK1, thereby weakening the cytotoxic effects of paclitaxel. Methylation of CHFR gene in EC cells leads to protein inactivation, heightening sensitivity to paclitaxel. (C) Elevated levels of PGK1 in EC induce ERK phosphorylation by directly binding to HSP90, causing upregulation of DNMTs expression, increasing methylation levels, and enhancing resistance to cisplatin. PR, progesterone receptor; EC, endometrial cancer; CHFR, checkpoint with forkhead-associated and ring finger; CDK1, cyclin dependent kinase 1; PGK1, phosphoglycerate kinase 1; DNMTs, DNA methyltransferases. DNA methylation is reversible and dynamic. Under normal circumstances, DNA methylation and demethylation are in dynamic equilibrium, with extensive changes in DNA methylation patterns observed at all stages of tumors ( 123 ). This indicates that DNA demethylating agents may hold promise in reprogramming tumor cells back to normal state, representing a new strategy of cancer therapy. DNA demethylating compounds are generally referred to as DNMTi, which can be divided into two groups: nucleoside and non-nucleoside inhibitors ( 124 ). Nucleoside analogues have been developed for over 40 years since their DNA demethylation activity was first discovered in 1980 ( 125 ). They capture DNMTs by binding DNA for proteasomal degradation. As representative drugs, azacitidine and decitabine have been approved by the US Food and Drug Administration (FDA) against hematological malignancies, including myelodysplastic syndrome (MDS) and acute myeloid leukaemia (AML) ( 126 ). Decitabine, also known as 5-aza-2’-deoxycytidine, is the most common medicament used to induce DNA demethylation. Treatment of EC cells with decitabine resulted in the downregulation of DNMT3B and upregulation of MLH1 . This was accompanied by cell growth inhibition, cycle arrest and apoptosis, along with elevated E-cadherin and decreased Bcl-2 expression ( 127 , 128 ). Azacitidine appeared to target DNMTs more effectively than decitabine in MLH1 -hypermethylated mismatch repair (MMR) deficient ECs ( 129 ). Azacitidine could rescue secreted protein acidic and rich in cysteine (SPARC), an albumin-binding protein, from a hypermethylated state, thereby increasing albumin-bound paclitaxel accumulation at EC lesions ( 130 ). Guadecitabine (SGI-110) is a second-generation demethylation drug that prolongs plasma half-life and improves binding strength compared to azacitidine and decitabine due to its resistance to degradation by cytidine deaminase ( 131 ). Zebularine, a relatively new cytidine analogue, is more stable and has a longer half-life than azacitidine ( 132 ). Unfortunately, there is no research evidence available for these drugs in EC. Nucleoside analogues binding to DNA may induce mutagenic damage that results in unnecessary toxicity and side effects, therefore, non-nucleoside analogues are currently under extensive investigation. Unlike nucleoside analogues, non-nucleoside inhibitors do not mimic cytosine. Instead, they work by directly binding and inhibiting specific target proteins ( 132 ). In EC, RG108, suppressing DNMT3B and upregulating MLH1 , could inhibit tumor cell growth, block cell cycle, and induce apoptosis, seeming to be considered as a new candidate drug for EC treatment ( 133 ). However, other non-nucleoside inhibitors, such as SGI-1027, hydralazine, procainamide and EGCG ( 134 ), have not yet been developed in EC. The mechanism of reported DNA demethylating agents in EC is illustrated in Figure 3 . Mechanism illustration of demethylation drugs in endometrial cancer. Nucleoside analog decitabine and non-nucleoside inhibitor RG108 reduce MLH1 methylation levels by inhibiting DNMT3B, thereby inhibiting proliferation, inducing cell cycle arrest, and apoptosis in EC cells. Additionally, decitabine can suppress the EMT process. Another nucleoside drug, azacitidine, promotes the demethylation process of SPARC, an albumin-binding protein, which enhances the accumulation of albumin-bound paclitaxel in EC lesions, consequently increasing its anti-tumor efficacy. DNMT3B, DNA methyltransferase 3B; EMT, epithelial-mesenchymal transition; SPARC, secreted protein acidic and rich in cysteine; EC, endometrial cancer. At present, DNA demethylation reagents have been widely explored in hematological tumors and solid tumors, such as breast, liver, pancreatic and lung cancer, while studies in EC are limited to the preclinical stage and few in number. DNA demethylating agents still face great challenges in the treatment of EC, highlighting an urgent need to identify epigenetic drivers specific to EC and its subtypes. Treatment experience from other tumors suggests that not all patients benefit from demethylation monotherapy, underscoring the importance of actively seeking combination regiments. For instance, there are some superficial preclinical studies on the combination of DNMTi and histone deacetylase inhibitor (HDACi) in EC ( 128 , 135 , 136 ). Moreover, it is essential to note that demethylation agents are not region-specific and may reactivate oncogenes leading to therapy failure or even tumor progression. Therefore, novel demethylation reagents need to be developed that can, for example, bind to specific targets by linking unique sequences.

According to the latest cancer statistics spanning 185 countries in 2022, cervical cancer ranked fourth among all women’s cancers with an incidence rate of 3.3%, while endometrial cancer (EC) ranked sixth at 2.1%, making them prominent gynecological tumors globally. In 2022, there were 420,242 newly diagnosed cases of EC and 97,704 deaths ( 1 ). EC predominantly affects postmenopausal women, yet 2% to 14% of cases occur in patients with reproductive age, with a notable 7.8% affecting individuals younger than 40 years old ( 2 , 3 ). Although most patients are diagnosed at an early stage (I-II) confined to the uterus, 10-20% of them experience relapse, with poorer prognosis observed in advanced and recurrent patients ( 4 ). Traditional Bokhman’s dualistic model divides EC into type I and type II. Type I tumors, such as endometrioid adenocarcinoma, are generally associated with elevated estrogen levels, accounting for more than 85% of EC cases, and typically exhibit a more favorable prognosis. Conversely, type II tumors, representing about 10% of ECs, are hormone receptor-negative and include serous and clear cell carcinoma, characterized by regrettable outcomes and recurrence rates exceeding 50% ( 5 , 6 ). Advances in molecular biology and sequencing technology have led to a more nuanced understanding of the molecular mechanisms underlying EC. In 2013, The Cancer Genome Atlas (TCGA) database classified EC into ultramutated/ POLE mut, hypermutated/MMRd, copy number-high/p53abn and copy number-low/no specific molecular profile (NSMP) types ( 7 , 8 ). Among these, POLE mut and p53abn types show the best and worst prognoses, respectively. TCGA molecular typing provides key information for clinicians to evaluate disease prognosis and formulate individualized treatment, holding promising application prospects. An updated staging system ( 9 ) for EC was published by International Federation of Gynecology and Obstetrics (FIGO) in 2023, which adopted molecular classification as one of the staging criteria. Further understanding of the molecular characteristics and regulatory mechanisms of EC is imperative for clinical risk stratification and treatment decisions. Gene expression is regulated through diverse mechanisms, including gene copy number variations, point mutations, and epigenetic modifications, each playing a critical role in cellular function and tumorigenesis ( 6 ). In recent years, increasing attention has been directed towards the pivotal role of epigenetic processes in cancer. DNA methylation is a type of epigenetic modification, and its abnormal alteration stands as a hallmark of human cancer occurrence and development ( 10 ). In eukaryotes, the fifth carbon atom of cytosine in the cytidine-phosphateguanosine (CpG) of the genome covalently binds to active methyl group (-CH3) to form 5-methylcytosine (5mC) under the catalysis of DNA methyltransferases (DNMTs), causing epigenetic changes that regulate gene expression ( 5 ). CpG is highly aggregated into CpG islands, 70% of which are present in gene promoter regions. Under normal circumstances, about 70% to 80% of CpG outside the CpG islands is methylated, while the vast majority inside CpG islands is hypomethylated or unmethylated ( 11 ). DNA methylation process is mediated by DNMTs, which consist of maintenance methyltransferase (DNMT1), and de novo DNA methyltransferases (DNMT3A, DNMT3B, and DNMT3L). DNMT1 is required to maintain DNA methylation patterns during cell activity and silence tumor suppressor genes aberrantly in tumor cells. Importantly, DNMT1 is the key enzyme that ensures the smooth transmission of epigenetic marks to the next generation ( 12 ). DNMT3A/B are thought to play an important role in shaping the epigenetic landscape of developing individuals, enabling the establishment of new DNA methylation patterns based on environmental factors during embryonic development ( 13 ). DNMT3L, a novel regulatory protein for de novo methylation, is involved in mediating the activity of DNMT3A/B ( 14 ). Dysregulation of DNMTs expression and function has been observed in a variety of diseases, including tumors. Researches have revealed distinct methylation states at different stages of menstruation cycle. For example, during the proliferative phase, stromal and glandular epithelial cells exhibit higher levels of cytosine methylation, which notably decline in adenocytes during secretory phase ( 15 ). These fluctuations may be attributed to the changing steroid hormone levels regulating DNMTs expression and function throughout the menstrual cycle, although the precise molecular mechanisms remain unclear. Compared to normal cells, cancer genomes are characterized by gene-specific hypermethylation of CpG islands and global hypomethylation ( 16 ). The methylation of CpG island promoters typically represses gene transcription by inhibiting promoter activity, whereas hypomethylated regions may enhance transcription by facilitating physical interactions between regulatory elements and gene promoters ( 17 , 18 ). Methylation status in EC is influenced by race ( 19 ), with black women displaying lower and more variable DNA methylation levels than white patients. Additionally, methylation patterns correlate with age, aging, body mass index, physical activity, and histological subtype to some extent ( 20 – 24 ). As early as over 20 years ago, studies indicated the hypermethylation of MLH1 and PTEN , leading to tumorigenesis and advanced stage in EC ( 25 , 26 ). Subsequently, the methylation status of tumor suppressor genes like RASSF1A , APC , p16, E-cadherin , CDH13 , ESR1 , and PRs have been documented ( 6 , 11 ). However, there are relatively few reports of hypomethylation of oncogenes, including PARP1 , BMP , CTCFL , PAX2 , NCAPH , MCM , and CASP8 . These differentially methylated regions and genes are implicated in many crucial cancer-related biology processes, encompassing cell differentiation, adhesion, invasion, apoptosis, cell cycle control, DNA mismatch repair, and epithelial-mesenchymal transition (EMT) process, cAMP signaling, Wnt signaling, and fibroblast growth factor signaling pathway ( 27 – 32 ). While foundational studies provided valuable insights into the molecular underpinnings of EC, their scope was constrained by methodological limitations and insufficient technological advancements. With the advent of high-throughput technologies, such as whole-genome bisulfite sequencing (WGBS), DNA methylation research has been further expanded. In addition, an increasing number of molecular methylation patterns in EC are being extensively investigated. For instance, novel hypermethylated genes, including ZSCAN12 and GYPC , have been identified as potential diagnostic markers with improved sensitivity and specificity ( 33 ). Apart from DNA methylation, other epigenetic mechanisms, including non-coding RNAs, histone modifications, and chromatin remodeling, also contribute to the pathophysiology of EC. Long non-coding RNAs (lncRNAs) can directly interact with chromatin by forming complexes with DNA, thereby influencing the binding of transcription factor and regulating gene expression ( 34 ). MicroRNAs (miRNAs) can pair complementarily with target gene sequences, mediating post-transcriptional suppression of gene expression ( 35 ). They play a critical role in the initiation and progression of EC. Histone lactylation promoted the malignant biological behavior of EC cells ( 36 ). Histone 3 lysine 27 (H3K27) methylation dysregulation may be an underlying cause of dedifferentiated EC ( 37 ), and the histone methyltransferase SMYD3 was highly expressed in EC ( 38 ). Chromatin remodeling genes (CRGs), such as ARID1A , CTCF , and KMT2D , were frequently altered in EC and associated with an increased likelihood of lymphovascular and myometrial invasion ( 39 ). Additionally, SP-1, ZEB1, and other transcription factors have been widely recognized for their roles in driving the progression of EC ( 40 , 41 ). In this review, we summarize the application of DNA methylation in early diagnosis, risk assessment, and therapy of EC, focusing on improving EC diagnostics, treatments, and management strategies.

With the rapid advancement in tumor diagnosis and treatment, the significance of DNA methylation cannot be ignored. In this review, firstly, we comprehensively reviewed the biomarkers based on DNA methylation in EC, encompassing valuable research findings published over the past two years. These findings serve as a swift reference for researchers investigating specific genes and contribute to the development of mature methylation detection kits. Secondly, we delineated the multifaceted role of DNA methylation in risk assessment of EC, spanning carcinogenesis risk, feasibility of fertility preservation, and overall prognosis, tailoring personalized treatment plans for patients. For accurate interpretation of DNA methylation data in EC diagnosis and prognosis, large-scale multicenter studies conducted in a standardized manner are imperative for validation. Furthermore, we shed light on the current landscape of new drug development centered on DNA methylation. In contrast to hematological tumors, research on DNA demethylating agents in EC remains limited. Nonetheless, the existing theoretical framework and research outcomes tentatively hint at the feasibility of demethylating agents in EC treatment, necessitating further preclinical investigations to assess their efficacy and safety. The methylation regulation mechanism is intricate, and existing research mostly focuses on DNMTi. However, attention should also be directed towards methylation-binding proteins and demethylases as prospective targets for demethylation. Additionally, methylation may be involved in the occurrence of chemotherapy resistance, underscoring the importance of actively exploring combination therapy regimens to enhance efficacy and mitigate adverse events. Admittedly, several limitations must be acknowledged to better interpret the findings of this review. One major challenge lies in the lack of consistency in data reporting across the cited studies. Key factors such as racial background, age, menopausal status, and tumor stage, which significantly influence DNA methylation patterns in EC, were inconsistently documented across studies, making it difficult to conduct comprehensive analyses or emphasize these variables. Moreover, methodological variability in DNA methylation detection platforms and analytical techniques, combined with the inclusion of some studies with small sample sizes, further complicates the comparability of results, limiting the generalizability of the synthesized findings. These limitations underscore the urgent need for more rigorous, standardized, and population-diverse studies to validate the role of DNA methylation in the diagnosis, risk assessment, and treatment of EC. Overall, DNA methylation holds promising research prospects and expansive opportunities for exploration in EC, offering valuable insights for early diagnosis, risk assessment, and treatment. Nonetheless, overcoming the current limitations is essential for driving the clinical application of DNA methylation forward.