EZH2 expression in hepatocellular carcinoma and its relationship with circadian rhythm-related genes Authors: Xingyue Wang, Miaolong He, Limian Zhou, Wei Chen Published: 2025 ABSTRACT AbstractHepatocellular carcinoma (HCC) is the most prevalent liver malignancy, with increasing incidence and high mortality, primarily associated with hepatitis virus infection and cirrhosis. Herein, we explored the association between enhancer of zeste homolog 2 (EZH2) expression and HCC progression, prognosis, and circadian rhythm-related genes. Differentially expressed mRNAs in HCC were identified from The Cancer Genome Atlas and Gene Expression Omnibus databases and intersected with circadian rhythm-related genes. Core genes were screened using protein–protein interaction networks, least absolute shrinkage and selection operator regression, and Cox regression analysis. Functional enrichment and immune infiltration analyses were conducted using gene set variation analysis, single-sample gene set enrichment analysis, and MethSurv. Experimental validation included western blotting, small interfering RNA knockdown, reverse transcription quantitative polymerase chain reaction, and functional assays. EZH2 was overexpressed in HCC and was associated with advanced tumor stage, poor prognosis, Th2 and dendritic cell infiltration, and promoter hypermethylation. EZH2 knockdown inhibited proliferation, induced DNA damage, and downregulated CLOCK and CRY1 expression. A prognostic nomogram integrating EZH2 expression with clinicopathological parameters demonstrated strong predictive performance. Methylation analysis identified six CpG sites significantly associated with survival. In conclusion, EZH2 is a key regulator of HCC progression and potential prognostic biomarker and therapeutic target.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-025-26175-x.Keywords: EZH2, Biomarker, Circadian rhythm, Hepatocellular carcinoma, PrognosisSubject terms: Cancer, Genetics, Immunology, Molecular biology Introduction Hepatocellular carcinoma (HCC) is the most prevalent form of liver malignancy worldwide, with a significant increase in its incidence and mortality rates. The World Health Organization reported that liver cancer was responsible for nearly 830,000 fatalities in 2020, making it the third most common cause of cancer-related deaths globally1. The increasing prevalence of HCC is closely associated with risk factors, including chronic hepatitis B and C infections, liver cirrhosis, and metabolic disorders such as nonalcoholic fatty liver disease2. Current treatment strategies—including surgical resection, liver transplantation, and systemic therapies—often show limited efficacy, particularly in the advanced phases of the disease, underscoring the urgent need for novel therapeutic targets and prognostic biomarkers3. Recent studies have demonstrated a close association between circadian rhythm disruption and the initiation and progression of HCC. Circadian rhythm, an intrinsic and highly sophisticated timekeeping system within various organisms, regulates numerous physiological processes essential for life, including the cell cycle, metabolic mechanisms, and DNA repair4. Disruption of circadian rhythms has been implicated in metabolic disorders and tumor development and progression by influencing mechanisms such as cell proliferation, apoptosis, and DNA repair5. Domestic and international studies have revealed abnormal expression of multiple circadian rhythm-related genes (e.g., CLOCK, ARNTL, PER1, CRY1) in HCC and their correlation with patient prognosis6,7. For instance, the aberrant expression of the CLOCK gene has been linked to enhanced invasiveness and metastatic potential in HCC5, whereas reduced expression of PER1 is correlated with unfavorable outcomes in patients with HCC8. Recently, the significance of epigenetic regulation in HCC has attracted increasing attention. enhancer of zeste homolog 2 (EZH2), the central catalytic component of polycomb prepressive complex 2 (PRC2), plays a pivotal role in transcriptional silencing, cell cycle regulation, and the maintenance of stem cell pluripotency by catalyzing histone H3K27 trimethylation (H3K27me3)9. Elevated EZH2 expression is associated with poor prognosis in various cancers, including HCC, suggesting its potential as a prognostic biomarker10, the interaction between EZH2 expression and circadian rhythm-related genes has been emphasized, as disturbances in circadian rhythms can influence tumor development and patient outcomes11. The association between EZH2 and circadian genes, including CLOCK and BMAL1, could shed light on the molecular mechanisms that drive HCC progression and underscore the significance of circadian regulation in cancer biology. This study aimed to examine the relationship between the expression of genes associated with circadian rhythms and HCC, with particular emphasis on the central gene EZH2. This gene is associated with tumor development and may act as a prognostic indicator in individuals diagnosed with HCC. This study integrated large-scale genomic data with advanced statistical methodologies to shed light on the molecular processes that underlie HCC, revealing EZH2 as a key factor influencing patient prognosis. Accordingly, the findings may contribute to the development of targeted therapeutic approaches and enhance the accuracy of prognostic evaluations in patients with HCC. Results Screening for core genes between differentially expressed genes (DEG) and circadian rhythm-related genes Comparative analysis of HCC versus normal liver tissues revealed 2590 DEGs with 2036 genes upregulated and 554 downregulated genes, as illustrated by the volcano plot in Fig. 1a. We screened circadian rhythm-associated genes from the Molecular Signatures Database (MSigDB) to delve deeper into the circadian rhythm–HCC nexus. We used Venn diagrams to precisely isolate genes concurrently dysregulated in HCC and circadian control. Intersection analysis of DEGs with those associated with circadian rhythms revealed 40 differentially expressed circadian rhythm-related genes (DECRGs) (Fig. 1b). The 40 DECRGs were then subjected to Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) (GOKEGG) enrichment analyses, including molecular function (MF), components of cells (CC), and biological processes (BP) categories. The results indicated that BP closely tied to circadian rhythms, including rhythmic processes, circadian rhythms, and their regulation, were notably enriched (Fig. 1c). The CC analysis highlighted the enrichment of DECRGs at specific neuronal loci, enhancing our understanding of the HCC cell microenvironment. MF analysis underscored histone kinase activity as the primary function among the 40 overlapping genes, suggesting a pivotal role for epigenetic modifications in regulating the circadian rhythm and tumorigenesis in HCC (Fig. 1c). KEGG pathway enrichment analysis further revealed pivotal links to neuroactive ligand–receptor crosstalk, viral-induced carcinogenesis, and the AGE-RAGE signaling pathway in diabetic complications (Fig. 1c). These findings implicate multifactorial mechanisms in HCC progression, including neural, viral, and metabolic disruptions, and provide a molecular framework for its complex etiology. Several core circadian genes are particularly important, including CLOCK, ARNTL, PER1, CRY1, RORA, and NR1D16. Correlation analysis of selected upregulated (Fig. 1d) and downregulated (Fig. 1e) overlapping genes revealed associations with CLOCK and CRY1, among others. Subsequently, a gene–gene interaction network was constructed based on the 40 DECRGs to explore the central hub genes. Network analysis identified eight key hub genes regulating this complex network (Fig. 1f). Distributions of mRNA in HCC and identification of DEGs associated with circadian rhythm-related genes. (a) Volcano plot of 2588 DEG. The red and blue dots represent the upregulated and downregulated genes, respectively. (b) Venn diagram of circadian rhythm-related genes and DEGs of HCC patients. The blue circle represents HCC-related DEGs, and the pink circle represents circadian rhythm-related genes. (c) GO and KEGG enrichment analyses of 27 upregulated genes and 13 downregulated genes. (d) Spearman correlation analysis between 27 upregulated genes and circadian rhythm-related genes* (p < 0.05). (e) Spearman correlation analysis between the 13 downregulated genes and circadian rhythm-related genes * (p < 0.05). (f) Hub genes in 27 upregulated genes and 13 downregulated genes. These eight hub genes were further filtered using least absolute shrinkage and selection operator (LASSO) regression analysis, which identified two genes that emerged as the most significant predictors (Fig. 2a-b). Moreover, univariate and multivariate Cox regression analyses were performed to explore the association between the expression levels of the two genes and the overall survival (OS) duration of patients within the HCC cohort, which included clinical variables. (Fig. 2c, d). These results suggested that EZH2 is an independent risk factor for HCC. Prognostic LASSO regression and cox analysis of CHEK1 and EZH2 expression. (a) Prognostic LASSO coefficient screening for the eight hub genes. (b) Prognostic LASSO variable trajectories for the eight hub genes. (c) Univariate Cox analysis of CHEK1 and EZH2 expression and other clinicopathological variables. (d) Multivariate Cox analysis of CHEK1 and EZH2 expression and other clinicopathological variables. High expression of EZH2 in HCC Through a detailed examination of gene expression datasets sourced from The Cancer Genome Atlas (TCGA) database, we identified a significant increase in the expression levels of the EZH2 within HCC tissues compared with normal liver tissues (Figs. 3a, b). These findings offer preliminary evidence for the involvement of EZH2 in HCC progression. We calculated an area under the curve (AUC) of 0.977, indicating strong diagnostic potential for EZH2 in distinguishing HCC from normal tissues (Fig. 3c). Patients diagnosed with HCC were categorized into cohorts based on the levels of EZH2 expression, revealing significantly prolonged OS, disease-specific survival (DSS), and progression-free interval (PFI) in the low-EZH2 group (p < 0.001), with notably extended survival times (Fig. 3d–f). This finding support the potential of EZH2 as a prognostic biomarker, further elucidating the EZH2–clinical outcome correlation. In addition, we used the Gene Expression Omnibus (GEO) database and UALCAN to validate our findings. The results were consistent with those from TCGA database, showing that EZH2 expression was elevated in cancer tissues relative to normal tissues (Fig. 3g, j), with an AUC of 0.954 (Fig. 3i), and high expression of EZH2 was associated with poor prognosis (Fig. 3h). Immunohistochemical staining images of normal liver and tumor tissues were obtained from the Human Protein Atlas (HPA) database. The images indicated that the expression level of EZH2 in HCC tissues was significantly higher than that in normal tissues (Fig. 3k). Expression and prognostic value of EZH2 in HCC. (a–b) The expression of EZH2 in normal versus tumor tissues from TCGA cohort. (c) ROC curve assessing the diagnostic value of EZH2 in HCC patients (TCGA). (d–f) Kaplan–Meier survival curves comparing the high and low expression of EZH2 with regards to OS (D), DSS (E), and PFI (F) in TCGA. (g) Expression of EZH2 in normal and tumor tissues from the GEO dataset. (h) Kaplan–Meier survival curve comparing the high and low expression of EZH2 expression for OS in GEO. (i) ROC curve evaluating the diagnostic performance of EZH2 in HCC patients in (GEO). (j) EZH2 expression in normal versus tumor tissues using UALCAN. (k) Representative immunohistochemical staining for EZH2 in normal liver and HCC tissues from the HPA(used with permission). Correlation between EZH2 expression levels and clinicopathological characteristics of patients with HCC Employing TCGA database, we analyzed the correlation between EZH2 expression and diverse clinical parameters in patients with HCC (Table 1). We specifically examined the association between EZH2 expression and 10 key clinicopathological features of HCC, including patient demographics (age and sex), tumor stage (T, N, M), pathological stage, presence of residual tumor, histological grading, and OS events (Supplementary Fig. 2). Our findings revealed a statistically significant correlation between elevated EZH2 expression and advanced T stage (Supplementary Fig. 2d), higher pathological stage (Supplementary Fig. 2f), poorer histological grade (Supplementary Fig. 2H), and reduced OS (Supplementary Fig. 2i), all with p < 0.001. These findings indicate marked upregulation of EZH2 in HCC, which correlates with multiple adverse clinical parameters, reinforcing its potential utility as a prognostic marker. Table 1Relationship between EZH2 expression and clinicopathological features in TCGA database.VariableLow expression of EZH2High expression of EZH2p-valuen187187Gender, n (%)0.224 Female55 (14.7%)66 (17.6%) Male132 (35.3%)121 (32.4%)Age, n (%)0.043 ≤ 6079 (21.2%)98 (26.3%) > 60108 (29%)88 (23.6%)Pathologic T stage, n (%)0.088 T1&T2145 (39.1%)133 (35.8%) T3&T439 (10.5%)54 (14.6%)Pathologic N stage, n (%)0.670 N0123 (47.7%)131 (50.8%) N11 (0.4%)3 (1.2%)Pathologic M stage, n (%)0.583 M0130 (47.8%)138 (50.7%) M13 (1.1%)1 (0.4%)Histologic grade, n (%)0.000 G1&G2138 (37.4%)95 (25.7%) G3&G447 (12.7%)89 (24.1%)Pathologic stage, n (%)0.050 Stage I & Stage II138 (39.4%)122 (34.9%) Stage III & Stage IV37 (10.6%)53 (15.1%) Relationship between EZH2 expression and clinicopathological features in TCGA database. Construction and validation of a prognostic and diagnostic model We established prognostic models to improve clinical decision-making and advance personalized medicine by selecting risk factors derived from the Cox regression analysis. A nomogram was constructed based on key predictors of OS in HCC, including T stage, M stage, pathological stage, and EZH2 expression, to further validate the prognostic significance of EZH2 on survival rates at 1, 3, and 5 years. The cumulative score, derived by summing the scores assigned to each prognostic factor, was used to predict the OS of patients with HCC. The nomogram revealed a higher total score associated with a poorer prognosis (Fig. 4a). Furthermore, calibration curves were utilized to evaluate the alignment between the predicted probabilities from the nomogram at different time intervals and the actual survival probabilities, indicating that the model could provide an accurate estimation of survival probabilities for 1-, 3-, and 5-year survival rates (Fig. 4b–d). Receiver operating characteristic (ROC) curve analysis showed that the AUC values for time-dependent ROC at 1, 3, and 5 years were 0.760, 0.742, and 0.746, respectively. These results suggest that the prognostic model demonstrated a strong predictive capability (Fig. 4e–g). Decision curve analysis (DCA) is a statistical method used to evaluate the clinical utility of predictive models in practical applications by comparing the net benefit of different models across various threshold probabilities, assisting physicians and researchers. The DCA revealed that the model provided valuable guidance for clinical decision-making within this threshold range (Fig. 4h–j). Predictive performance of EZH2 for OS in HCC. (a) Nomogram predicting 1-, 3-, and 5-year OS based on EZH2 expression and clinical variables. (b–d) Calibration plots for 1-, 3-, and 5-year OS used to validate the nomogram model. (e–g) Time-dependent ROC curves for 1-, 3-, and 5-year OS. (h–j) DCA for 1-, 3-, and 5-year OS. Functional enrichment analysis of EZH2 in patients with HCC We analyzed genes co-expressed with EZH2 in the TCGA HCC dataset. A total of 8301 genes were identified based on the criteria of |Spearman’s correlation coefficient| > 0.3 and p < 0.05. The top 30 positively correlated genes are shown in Supplementary Fig. 3a. Functional annotations of these co-expressed genes were explored using GO terms and KEGG pathway analysis. Bubble plots illustrate the top three enriched categories of BP, CC, MF, and KEGG pathways. GO analysis revealed that these genes were predominantly involved in organelle fission, nuclear division, chromosome segregation, various catalytic activities, ATP-dependent mechanisms, and single-stranded DNA helicase activity (Supplementary Fig. 3b). The KEGG pathway analysis highlighted the significant roles of these genes in regulating the cell cycle, DNA replication, and Fanconi anemia pathways (Supplementary Fig. 3b). Moreover, gene set variation analysis (GSVA) was conducted to explore the potential impact of EZH2 expression on the development and progression of HCC. The findings revealed that genes in the high- and low-expression cohorts were predominantly enriched in an array of biological pathways such as E2F targets, G2M checkpoint, DNA repair, PI3K/AKT/MTOR signaling, mitotic spindle, spermatogenesis, and MYC targets v1 (Supplementary Fig. 4). These results suggest that genes related to the circadian rhythm may play a significant role in modulating the processes of DNA damage and repair, DNA replication, and the cell cycle. EZH2 expression is significantly correlated with immune infiltration Our investigation explored the relationship between EZH2 expression and immune cell infiltration in tumors using both single-sample gene enrichment analysis (ssGSEA) and CIBERSORT algorithms. The analysis highlighted the three most significant correlations between EZH2 and Th2 cells (r = 0.662, p < 0.01), Th17 cells (r = − 0.417, p < 0.01), and dendritic cells (DCs) (r = − 0.404, p < 0.01), as determined by the ssGSEA algorithm (Supplementary Fig. 5a). Concurrently, using the CIBERSORT algorithm, our findings revealed that the three most notable associations were between EZH2 and follicular helper T cells (r = 0.269, p < 0.01), M0 macrophages (r = 0.239, p < 0.01), and activated CD4+ memory T cells (r = 0.235, p < 0.01) (Supplementary Fig. 5b). We further assessed immune cell enrichment scores in high- and low- EZH2 expression groups using both ssGSEA and CIBERSORT. According to the ssGSEA method, the high-expression group displayed increased enrichment scores in four categories of immune cells, whereas the low-expression group demonstrated higher scores in eight categories of immune cells (Supplementary Fig. 5c). In contrast, the CIBERSORT results showed that the high-expression group had elevated enrichment scores across five immune cell types, whereas the low-expression group had increased scores in two immune cell types (Supplementary Fig. 5d). EZH2 gene mutation and methylation Given that genetic mutations are a primary factor contributing to the onset of cancer12, we examined EZH2 mutations and copy number variations (CNVs) using two datasets on HCC from cBioPortal: INSERM, Nat Genet 2015 (n = 243) and TCGA, Firehose Legacy, (n = 379). Our findings indicated that EZH2 exhibited missense mutations, truncating mutations, amplifications, and deep deletions at a rate of 2.9% (Fig. 5a, b). Irregular DNA methylation plays a crucial role in the initial stages of HCC progression13,14. Nevertheless, our investigation determined that mutations in EZH2 did not correlate with variations in OS (p = 0.803) or DFS (p = 0.409) (Fig. 5c, d). Further assessments using the UALCAN database indicated that the elevated DNA methylation levels of EZH2 in the promoter region of HCC tissues compared with normal liver tissues were statistically significant (p < 0.01) (Fig. 5e). Furthermore, utilizing the MethSurv database facilitated a more nuanced exploration of the DNA methylation levels of EZH2 and the prognostic significance of its associated CpG islands. The outcomes revealed that most CpG sites were hypomethylated (Fig. 5f). Importantly, six CpG sites displayed methylation levels associated with HCC prognosis (Fig. 5g). In particular, hypermethylation of cg18494018, cg06816161, and cg13751213 correlated with a worse prognosis (Fig. 5h–k). Mutations and DNA methylation levels of EZH2 and their associations with HCC prognosis. (a, b) OncoPrint plots showing EZH2 mutation profiles from the cBioPortal database. (c, d) Kaplan-Meier survival curves evaluating the association between EZH2 mutation status and OS (C) and DFS (D) in HCC. (e) EZH2 methylation levels in HCC compared to normal tissue, retrieved form the UALCAN database. (f) Correlation between EZH2 mRNA expression level and methylation level form the MethSurv database. (g) Association between EZH2 methylation levels and HCC prognosis. (h–k) Kaplan–Meier survival curves showing the prognostic impact of methylation at specific EZH2 CpG sites: (h) cg26118713, (i) cg18494018, (j) cg06816161, (k) cg13751213. Correlation of EZH2 expression with circadian rhythm in HCC Functional enrichment analysis revealed that EZH2-associated genes were predominantly implicated in crucial cellular processes, including cell cycle regulation, DNA replication, the Fanconi anemia repair pathway, and spliceosome function (Supplementary Fig. 3b). Notably, circadian clock genes such as PER2 have been shown to directly or indirectly modulate pivotal cell cycle regulators, including p53, p21, Cyclin D1, CDK1, and Cyclin B1, suggesting a potential interplay with circadian rhythm control mechanisms15,16. Figure 6a and b illustrate a significant positive correlation between EZH2 expression and CLOCK (r = 0.316, p < 0.001) and CRY1 (r = 0.440, p < 0.001) expression, as depicted in the scatter plot. In our study, we compared the expression levels of EZH2 in the normal liver cell line MIHA with those in the HCC cell lines Huh7 and JHH7. The results indicated that the expression levels of EZH2 in HCC cell lines were significantly higher than those observed in the normal liver cell line MIHA (Fig. 6c, d). We knocked down EZH2 in JHH7 and Huh7 cells using two specific siRNAs to validate this correlation, and the effectiveness of the knockdown was confirmed by quantitative polymerase chain reaction (PCR) and western blot analysis. Experimental results indicate that the knockdown of EZH2 leads to the decreased expression of CLOCK and CRY1 (Fig. 6e–h). Effect of EZH2 knockdown on the expression of circadian rhythm-related genes in HCC cells. JHH7 and Huh7 cells were transfected with control or EZH2-targeting siRNAs. (a, b) Scatter plots were plotted showing the correlations between EZH2 and expression of circadian rhythm-related genes, including CLOCK, and CRY1. (c) Protein expression of EZH2 in MIHA, Huh7, and JHH7. (d) mRNA expression of EZH2 in MIHA, Huh7, and JHH7. (e, f) Protein levels of EZH2, CLOCK, and CRY1 as measured by western blotting. (g, h) mRNA levels of EZH2, CLOCK, and CRY1 assessed by real-time quantitative PCR. p < 0.05, p < 0.01, p < 0.001. EZH2 knockdown inhibits HCC cell proliferation and induces DNA damage To investigate the effect of EZH2 on cell proliferation, we used two specific siRNAs to knock down EZH2 expression in Huh7 and JHH7 cell lines. Cell Counting Kit-8 (CCK-8) assay results showed that cells with EZH2 knockdown exhibited a significantly slower proliferation rate compared with control cells (Fig. 7a, b). This observation suggests that the EZH2 gene may promote cell proliferation (Fig. 7a, b). In addition, the comet assay results showed that cells with EZH2 knockdown had longer comet tails than control cells (Fig. 7c). Our findings indicate that EZH2 knockdown causes DNA damage and reduces DNA repair capacity. EZH2 knockdown inhibits HCC cell proliferation and induces DNA damage. Huh7 and JHH7 cells were transfected with siRNA targeting EZH2. (a, b) Cell proliferation was assessed using the CCK-8 assay. (c) DNA damage was evaluated using the comet assays. p < 0.05, p < 0.01, p < 0.001. Discussion HCC continues to be a predominant contributor to cancer-related deaths globally, highlighting the urgent need to discover reliable biomarkers that facilitate early detection and prognosis. This study highlighted the significant overexpression of EZH2 in HCC tissues and its strong correlation with patient outcomes, underscoring its potential as a critical prognostic biomarker. Unlike previous studies that primarily focused on individual gene expression and exosomal micro-RNAs17,18, our research integrated circadian rhythm-related genes, providing a novel perspective on the multifaceted role of EZH2 in HCC progression. Circadian rhythms influence the initiation and progression of HCC by regulating multiple processes, including the cell cycle, metabolism, DNA repair, immune microenvironment, epigenetic modifications, oxidative stress, and tumor microenvironment. Dysregulation of circadian rhythm genes—such as CLOCK, BMAL1, PER, and CRY—may result in abnormal proliferation, metabolic reprogramming, genomic instability, immune evasion, and microenvironmental remodeling in HCC cells, thereby promoting tumor progression19–23. To our knowledge, this study is the first to elucidate the regulatory relationship between EZH2 and circadian rhythm-related genes (CLOCK and CRY1) in HCC. Functional enrichment analysis revealed that genes associated with EZH2 were markedly enriched in pathways related to the cell cycle, DNA replication, and Fanconi anemia repair, all of which strongly connect with regulating circadian rhythms. Experimental validation demonstrated that knockdown of EZH2 resulted in downregulation of CLOCK and CRY1, suggesting that EZH2 may influence HCC progression by modulating these core circadian rhythm genes. Both CLOCK and CRY1 play critical roles in maintaining the cell cycle and DNA repair5,8, and their dysregulation may contribute to genomic instability and tumorigenesis. Therefore, EZH2 may indirectly affect the proliferative capacity and DNA damage repair mechanisms of HCC cells by regulating the circadian rhythm-related genes. EZH2 is a core component of the PRC2 complex. It promotes the proliferation, invasion, and epithelial-mesenchymal transition (EMT) of HCC by catalyzing the trimethylation of histone H3 at lysine 27 (H3K27me3) to silence tumor suppressor genes, such as CDKN2A and PTEN10. EZH2 exhibits elevated expression levels in HCC, which is notably linked to advanced disease stages, vascular invasion, and unfavorable prognostic outcomes10, and regulates the HOXA-AS2/miRNA-ceRNA network to enhance pro-tumorigenic signaling and suppress anti-tumor immunity by reducing CD8+ T cell infiltration10. Our analysis revealed that elevated EZH2 expression was correlated with advanced tumor stages, poor histological grades, and reduced OS rates among patients with HCC. Furthermore, a prognostic model incorporating EZH2 expression and clinical parameters demonstrated predictive power for patient survival. These findings reinforce the necessity of EZH2 as an HCC biomarker and suggest its involvement in intricate regulatory networks influencing tumor behavior and patient prognosis. These results warrant further investigation into the mechanistic role of EZH2 in HCC and its potential as a therapeutic target. This study revealed a significant negative correlation between EZH2 expression and OS in patients with HCC, suggesting that EZH2 may play a critical role in the progression of HCC. This finding aligns with previous research, such as a study utilizing TCGA data, which reported that high EZH2 expression was associated with poor prognosis in patients with HCC (n = 243) and was correlated with advanced tumor stages and increased recurrence rates24. Another investigation involving 66 patients with HCC demonstrated that elevated EZH2 levels were associated with unfavorable clinical outcomes, reinforcing the notion that EZH2 is an independent prognostic marker for HCC25. These studies support the hypothesis that EZH2 contributes to HCC malignancy through mechanisms such as epigenetic regulation and tumor microenvironment modulation. However, some studies have reported conflicting results. For instance, a study analyzing 379 HCC samples found no significant association between EZH2 mutations and OS, suggesting that genetic alterations in EZH2 may not directly influence patient survival10. This discrepancy may stem from differences in sample sizes, methodologies, or specific patient populations studied, highlighting the complexity of EZH2’s role in HCC. The developed nomogram model, which incorporated EZH2 expression levels and clinical staging, demonstrated significant clinical utility. Overall, consistent findings across multiple studies underscore the potential of EZH2 as a therapeutic target and prognostic biomarker in HCC while emphasizing the need for further research to clarify its mechanistic pathways and interactions within the tumor microenvironment. This study identified various mutation types of the EZH2 gene in HCC, including missense mutations, truncating mutations, and amplifications. However, the mutation frequency was low (2.9%) and showed no significant correlation with patient survival. DNA methylation levels in the EZH2 promoter region were significantly elevated in HCC tissues, and the methylation status of certain CpG sites was associated with patient prognosis. These findings suggest that EZH2 expression in HCC is primarily regulated by epigenetic mechanisms rather than genetic mutations. Future research could further explore the regulatory mechanisms underlying EZH2 methylation in HCC and evaluate its potential as a therapeutic target. In particular, efforts should focus on identifying key transcription factors that regulate the expression of EZH2 in specific cellular subtypes. Understanding the functional roles of these transcription factors may reveal novel therapeutic targets. Investigating how these transcription factors modulate EZH2 activity and associated cellular pathways is crucial for developing targeted strategies to overcome treatment resistance in HCC. However, this study has several limitations that must be acknowledged. First, a significant limitation of this study is the lack of a specific in vivo model to validate the functional role of EZH2 in the progression of HCC. Moreover, the small sample size may limit the generalizability of our results, and the lack of clinical validation emphasizes the need for larger multicenter studies. Furthermore, the potential batch effects arising from the use of multiple datasets could introduce variability into the expression analysis, complicating the interpretation of the results. In conclusion, our investigation revealed that increased EZH2 expression correlates with greater malignancy and dismal prognosis in patients with HCC, underscoring its potential as a prognostic biomarker and therapeutic target. The machine learning approaches employed in this study provided a robust framework for identifying key molecular players in HCC, laying the groundwork for future studies aimed at to elucidating the underlying mechanisms of EZH2 in tumor progression. Ultimately, these findings may contribute to developing novel therapeutic strategies that could improve patient outcomes in HCC, emphasizing the importance of further investigations into the role of EZH2 in liver cancer biology. Materials and methods All data were obtained from publicly available databases (TCGA, GEO) and complied with relevant ethical guidelines. Screening for core genes between DEGs and circadian rhythm-related genes RNA sequencing data and relevant clinical details for patients diagnosed with HCC were obtained from TCGA (https://cancergenome.nih.gov/) and GEO (https://www.ncbi.nlm.nih.gov/) databases. Genes associated with circadian rhythms were sourced from the MSigDB (https://www.gsea-msigdb.org/gsea/msigdb/), as presented in Supplementary Tables 126,27. Differential expression profiles (HTSeq-Counts) between the HCC cohort and the control group were analyzed to identify DEGs using the Wilcoxon rank-sum test in R, specifically with the DESeq2 (version 1.36.0) and edgeR (version 3.38.2) packages. Threshold values for identifying DEGs were established as |log2Fold Change| > 1.5 and an adjusted p-value of < 0.0128,29. The “ggplot2” package was used to generate mRNA volcano plots. Venn overlap analysis was used to illustrate the relationship between DEGs and circadian rhythm-associated genes. The Search Tool for the Retrieval of Interacting Genes (STRING; http://string-db.org; version 10.0) was used to predict protein–protein interactions network for EZH2 co-expressed genes and analyze functional protein associations30. Statistical significance was set at a combined score exceeding 0.4, which was deemed statistically significant for interactions. We identified EZH2 as the central gene using Cystoscape (version 3.10.0). In addition, this approach involved a preliminary fitting process in which gene expression data were analyzed alongside survival duration using the LASSO regression technique. This was followed by a rigorous cross-validation procedure using the “cv.glmnet” function to ensure the model’s reliability and accuracy. The “coef” function was then used to extract and compute the weights of the selected genes, providing a clearer understanding of their contribution to the analysis. Finally, patients with complete clinical data from the TCGA database were selected for univariate and multivariate Cox regression analyses to identify candidate genes that serve as independent prognostic risk factors. Correlation of gene expression levels with circadian rhythm-associated genes in HCC The relationship between gene expression levels and circadian rhythm-associated genes such as CLOCK, ARNTL, PER1, CRY1, RORA, and NR1D1 was examined using the R statistical programming language. Correlation analysis between these genes was performed using an appropriate R package. Expression and prognostic significance of EZH2 in HCC The TCGA and GEO databases were used to assess the relationship between EZH2 mRNA expression and survival outcomes in patients with HCC and varying clinical characteristics. The UALCAN database was used to evaluate the association between EZH2 expression and patient survival. The R package “survival” (version 3.6) was used to generate OS, DSS, and PFI survival plots based on EZH2 expression in patients from TCGA database, as well as OS plots from the GEO database. The patients were divided into high- and low-expression groups based on the median expression level of EZH2. Immunohistochemical staining Immunohistochemical staining images of normal liver and HCC tissues were obtained from the HPA database (https://www.proteinatlas.org/) to evaluate the varying expression levels of EZH2 in healthy and tumor tissues31. Gene ontology and Kyoto encyclopedia of genes and genomes enrichment analysis Correlation analyses were performed between EZH2 and other mRNAs in HCC using data from TCGA to understand the biological mechanisms and pathways related to EZH2. GOKEGG pathway enrichment analyses were performed on genes that demonstrated significant co-expression using the clusterProfiler package in R. Following this, a prognostic and diagnostic nomogram was constructed and validated. We constructed a nomogram incorporating independent prognostic variables identified through Cox regression analysis to develop a prognostic model for forecasting OS in HCC. These variables include the expression level of EZH2, T stage, M stage, and pathological stage. DCA and calibration curves were used to validate the nomogram’s precision and dependability. Furthermore, independent prognostic factors were established, and ROC curves were generated using the “timeROC” software package in R32. EZH2-associated gene enrichment analysis In examining cohorts distinguished by high- and low- EZH2 expression, the 30 genes that exhibited the most pronounced differential expression were further analyzed based on their adjusted p-values. Heatmaps were generated to illustrate the associations between these genes and EZH2 expression. Functional characterization of the DEGs was performed using GO and KEGG enrichment analyses. Gene set variation analysis The h.all.v2024.1. Hs.symbols.gmt gene set was obtained from MSigDB and used to perform GSVA on EZH2 expression in HCC samples from TCGA-Liver Hepatocellular Carcinoma (TCGA-LIHC) dataset. Functional enrichment differences were assessed between high- and low-risk cohorts. The threshold for GSVA screening was set at an adjusted p-value of < 0.05, using the Benjamini–Hochberg (BH) method for multiple testing correction. The “GSVA” package was used for the analysis, and results were visualized using the “ggplot2” package. Immune cell infiltration level and correlation analysis ssGSEA and CIBERSORT were used to assess the relationship between various immune cell types and the core gene33,34. Gene mutations and DNA methylation This study examined the mutations and CNVs linked to EZH2 using cBioPortal, emphasizing the correlation between EZH2 genetic alterations and HCC prognosis35. The methylation status of the EZH2 promoter was assessed using UALCAN36, and MethSurv was used to analyze the prognostic significance of the methylation levels of EZH237. Cell lines and culture MIHA, Huh7, and JHH7 cells were kindly provided by Dr. Wei Chen. All cell lines were maintained in Dulbecco’s Modified Eagle’s medium (Hyclone, Logan, UT, USA) supplemented with 10% fetal bovine serum (FBS; Hyclone) and 1% penicillin–streptomycin. The cultures were kept in a humidified incubator set at 37 °C with an atmosphere of 5% CO2. Each cell line was confirmed to be free of Mycoplasma contamination, and authentication was conducted through genetic profiling of polymorphic short tandem repeat (STR) loci for cultures that had been passaged for more than 6 months after their arrival at our laboratory. Western blotting Cell lysates were prepared in radioimmunoassay buffer (Solarbio, Beijing, China) supplemented with a protease inhibitor cocktail (Roche, Basel, Switzerland). An equivalent quantity of total protein (30 µg) was analyzed via 8% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and subsequently transferred onto a polyvinylidene fluoride (PVDF) membrane (Millipore, Danvers, MA, USA) by applying a current of 300 mA for 90 min. After a blocking step with 5% skimmed milk at 25 °C for 1 h, the membranes were subsequently incubated overnight at 4 °C with primary antibodies against EZH2 (#5246; 1:1000; Cell Signaling Technology, Danvers, MA, USA), CLOCK (#5157; 1:1000; Cell Signaling Technology, Danvers, MA, USA), CRY1 (13474-1-AP; 1:1000; Proteintech Group, Inc, Chicago, USA), and GAPDH (60004-1-Ig; 1:50000; Proteintech Group, Inc, Chicago, USA). Subsequently, the samples were incubated with secondary antibodies conjugated to horseradish peroxidase, and detection was performed using the ECL Plus reagent (GE Healthcare, Chicago, IL, USA). SiRNA transfection The sequences of the si_EZH2_ vectors were 5′-GTGCCCTTGTGTGATAGCACAA-3′ and 5′-GGCACTTTCATTGAAGAACTAA-3′. The sequence for the non-target control vector was 5′-TTCTCCGAACGTGTCACGT-3′. The siRNAs were procured from Gema Gene (Shanghai, China). Cells were transfected with si_EZH2_ employing Lipofectamine 2000 (Thermo Fisher Scientific Inc. Waltham, MA, USA) according to the manufacturer’s instructions. Real-time reverse transcription-polymerase chain reaction Total RNA was isolated using TRIzol reagent (Life Technologies, Carlsbad, CA, USA) and treated with RNase-free DNase I (Promega, Madison, WI, USA) for 30 min. Reverse transcription was performed using an M-MLV reverse transcription kit (Promega). SYBR Green real-time reverse transcription PCR was performed using the ABI PRISM 7300 Sequence Detection System (Life Technologies). The primer sequences employed were as follows: CLOCK: 5′-TCTCAGACCCTTCCTCAACA-3′ (forward) and 5′-TGACCTTCTTTGCACCATCTT-3′ (reverse); CRY1: 5′-CTGCGTCTACATCCTGGACC-3′ (forward) and 5′-GAAGCAAAAATCGCCACCTGT-3′ (reverse). Gene expression analysis was performed using the 2 −∆∆Ct method. The ΔCt value was calculated by subtracting the Ct (threshold cycle) of GAPDH from that of the target gene, whereas the ΔΔCt was calculated by taking the difference between the ΔCt value of the experimental sample and that of the control sample. Fold-change was expressed as 2 ^−ΔΔCt, with the relative expression level of the control group normalized to 1. Comet assay Cells were transfected with siRNA, and after incubation for 48 h, they were harvested for further analysis. Following this, comet assays were conducted according to established protocols38, and the resulting comet images were visualized using a fluorescence microscope (Leica DMI3000 B; Leica, Wetzlar, Germany). Image analysis was conducted using the Comet Assay Software Project (CASP; Wrocław, Poland) to ensure precise evaluation of DNA damage. Cell proliferation Cells were seeded in 96-well plates at a density of 2 × 103 cells/well, 24 h post-transfection. Cell proliferation was evaluated using the CCK-8 assay (MedChemExpress, Monmouth Junction, NJ, USA) following the manufacturer’s instructions. Absorbance was measured at 490 nm using a multifunctional microplate reader (Thermo Fisher Scientific, Waltham, MA, USA). Statistics analysis Data analysis and visualization were conducted using the R software (version 4.3.2). Box lines and scatter plots were used to assess the expression levels of the EZH2 gene in patients diagnosed with HCC. The threshold value of EZH2 expression was determined using the median approach for gene expression. Spearman’s correlation analysis was used to investigate the extent of the correlation between EZH2-associated genes and EZH2 expression. Univariate and multivariate Cox regression analyses were used to identify potential prognostic factors. ROC analysis, along with the commonly utilized method for binary evaluation, was conducted using the pROC package to evaluate the ability of EZH2 expression to distinguish HCC samples from their normal counterparts. The calculated AUC values, ranging from 0.1 to 0.5, reflected discriminative power between 50% and 100%. In all analyses, * indicates p < 0.05, which denotes statistical significance; ** indicates p < 0.01, signifying high statistical significance; and *** indicates p < 0.001, representing extremely high statistical significance. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Fig. 1. Flow chat of the study. Supplementary Fig. 2. Expression levels in tumor tissues across clinical subgroups in the TCGA database. (a) Age, (b) Gender, (c) N stage, (d) T stage, (e) M stage, (f) Pathologic stage, (g) Residual tumor status, (h) Histologic grade, (i) OS events. p < 0.05, p < 0.01, p < 0.001; ns: no signification. Supplementary Fig. 3. Functional clustering and interaction network analyses of EZH2-related genes. (a) Heatmap of the top 30 genes positively correlated with EZH2 expression in HCC. (b) Enrichment analyses for BP, CC, MF, and KEGG of EZH2 co-expressed genes. Supplementary Fig. 4. GSVA of differentially enriched pathways between the high and low EZH2 expression groups. (a) Heatmap of significantly enriched pathways. (b) Bar plot showing pathway enrichment scores. Supplementary Fig. 5. Correlation betweenEZH2 expression and immune cell infiltration. (a-b) Bubble plots showing correlations between EZH2 expression and immune cell abundance using the ssGSEA and CIBERSORT algorithms, respectively. (c-d) Immune infiltration levels of different immune cells in high and low EZH2 expression groups, analyzed using the ssGSEA and CIBERSORT algorithms, respectively Supplementary Table 1. Genes associated with circadian rhythms were sourced from the MSigDB. Acknowledgements The authors are grateful to HelixLife for their expert advice and guidance. Author contributions Conceptualization: X.W. and M.H.; Methodology; W.C.; Software: M.H.; Validation: W.C. and L.Z.; Formal analysis: X.W.; Investigation: W.C. and M.H.; Resources: X.W and L.Z.; Data curation: M.Z.; Writing—original draft: X.W.; Writing—review and editing: X.W.; Visualization: M.H.; Supervision: X.W.; Project administration: X.W. Funding Our project has no funding support. Data availability The data supporting the findings of this study are available from the corresponding author upon reasonable request. Public datasets are available from TCGA ( [https://cancergenome.nih.gov/](https:/cancergenome.nih.gov) ) and GEO (https://www.ncbi.nlm.nih.gov/geo/). Declarations Competing interests The authors declare no competing interests. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Xingyue Wang and Miaolong He contributed equally to this work. References Associated Data Supplementary Materials Supplementary Fig. 1. Flow chat of the study. Supplementary Fig. 2. Expression levels in tumor tissues across clinical subgroups in the TCGA database. (a) Age, (b) Gender, (c) N stage, (d) T stage, (e) M stage, (f) Pathologic stage, (g) Residual tumor status, (h) Histologic grade, (i) OS events. p < 0.05, p < 0.01, p < 0.001; ns: no signification. Supplementary Fig. 3. Functional clustering and interaction network analyses of EZH2-related genes. (a) Heatmap of the top 30 genes positively correlated with EZH2 expression in HCC. (b) Enrichment analyses for BP, CC, MF, and KEGG of EZH2 co-expressed genes. Supplementary Fig. 4. GSVA of differentially enriched pathways between the high and low EZH2 expression groups. (a) Heatmap of significantly enriched pathways. (b) Bar plot showing pathway enrichment scores. Supplementary Fig. 5. Correlation betweenEZH2 expression and immune cell infiltration. (a-b) Bubble plots showing correlations between EZH2 expression and immune cell abundance using the ssGSEA and CIBERSORT algorithms, respectively. (c-d) Immune infiltration levels of different immune cells in high and low EZH2 expression groups, analyzed using the ssGSEA and CIBERSORT algorithms, respectively Supplementary Table 1. Genes associated with circadian rhythms were sourced from the MSigDB. Data Availability Statement The data supporting the findings of this study are available from the corresponding author upon reasonable request. Public datasets are available from TCGA ( [https://cancergenome.nih.gov/](https:/cancergenome.nih.gov) ) and GEO (https://www.ncbi.nlm.nih.gov/geo/).