CCNA2, MAD2L1, AURKA, and PTTG1 promote proliferation and migration in hepatocellular carcinoma Authors: Wang Jinli, Hu Shiping Published: 2026 ABSTRACT Liver hepatocellular carcinoma (LIHC) is a leading cause of cancer-related mortality worldwide, with limited therapeutic options and poor prognosis, particularly at advanced stages. Identifying key molecular drivers involved in LIHC progression is critical for improving diagnosis and treatment strategies. In this study, we performed an integrated bioinformatics analysis using four independent GEO datasets (GSE19665, GSE54236, GSE55092, and GSE84598) to identify differentially expressed genes (DEGs) associated with LIHC. A total of 147 common DEGs were identified, and protein–protein interaction network analysis revealed CCNA2, MAD2L1, AURKA, and PTTG1 as central hub genes. Functional enrichment analyses indicated that these genes are primarily involved in cell cycle regulation, mitosis, and chromosomal stability. Expression analyses using TCGA-based databases demonstrated significant upregulation of these hub genes at both mRNA and protein levels in LIHC tissues compared with normal liver tissues. Elevated expression was associated with advanced tumor stage, poor overall survival, immune cell infiltration, and altered drug sensitivity, highlighting their clinical relevance. Promoter methylation analysis suggested that epigenetic regulation may contribute to gene dysregulation, while genetic alterations were infrequent. Experimental validation using RT-qPCR confirmed overexpression of hub genes in LIHC cell lines. Furthermore, siRNA-mediated knockdown of these genes significantly suppressed proliferation, colony formation, and migration in HepG2 and Huh7 cells. Collectively, these findings identify CCNA2, MAD2L1, AURKA, and PTTG1 as key oncogenic drivers and potential biomarkers and therapeutic targets in LIHC.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12876-026-04845-y. Introduction Liver hepatocellular carcinoma (LIHC), the most common type of primary liver cancer, is a leading cause of cancer-related mortality worldwide, accounting for nearly 75–85% of all liver cancer cases [1, 2]. The global incidence of LIHC has been steadily rising, with over 800,000 new cases diagnosed annually [3, 4]. The disease is particularly prevalent in regions with high rates of chronic viral infections, such as sub-Saharan Africa and Southeast Asia, where hepatitis B virus (HBV) and hepatitis C virus (HCV) are endemic [5]. Chronic HBV and HCV infections are major risk factors for LIHC, contributing to liver cirrhosis and eventual tumorigenesis through persistent liver inflammation, fibrosis, and regenerative proliferation [6]. According to recent global estimates, approximately 254 million people are living with chronic HBV infection and 50 million with HCV worldwide, highlighting the substantial burden of viral hepatitis in these regions [7]. In addition to viral infections, lifestyle-related factors also play a significant role in the development of LIHC [8]. Excessive alcohol consumption is a well-established risk factor, leading to alcoholic liver disease and cirrhosis, which significantly increases the likelihood of hepatocellular carcinoma [9]. Current treatment options for LIHC include surgical resection, radiofrequency ablation, transarterial chemoembolization (TACE), and systemic therapies such as sorafenib and lenvatinib, which target angiogenesis and signaling pathways involved in tumor growth [10, 11]. However, these treatments offer limited efficacy, particularly for patients with advanced-stage disease, and the overall 5-year survival rate remains below 20% [12]. Over recent years, considerable efforts have been made to elucidate the molecular mechanisms driving LIHC. Multiple studies have identified potential biomarkers through comprehensive analyses of gene expression profiles, with some genes showing promise as diagnostic or therapeutic targets [13, 14]. For example, TP53 and CTNNB1 mutations, two of the most commonly altered genes in LIHC, have been linked to tumor initiation and progression, affecting cell cycle regulation, DNA repair, and apoptosis [15–17]. Studies have also identified aberrant activation of signaling pathways such as the Wnt/β-catenin, MAPK, and PI3K/AKT pathways, contributing to uncontrolled cell proliferation and metastasis [18]. Gene expression profiling has been widely utilized to uncover differentially expressed genes (DEGs) that are associated with disease progression [19, 20]. Notably, the analysis of the GSE14520 dataset revealed that genes like AFP, GPC3, and MDM4 are upregulated in LIHC tissues compared to normal controls, and their overexpression correlates with poor prognosis [21]. TCGA-LIHC data have further expanded the understanding of LIHC, highlighting genes like ALB, RPS20, and BIRC5 as important regulators of cell growth and survival [22, 23]. In our study, we aim to deepen the understanding of LIHC by using multiple publicly available datasets, including GSE19665, GSE54236, GSE55092, and GSE84598. These datasets provide comprehensive gene expression data that can be analyzed to uncover key differentially expressed genes (DEGs) and hub genes potentially involved in the development and progression of LIHC. In addition to bioinformatics analyses, we validate the identified genes through in vitro experiments, providing a more robust and accurate confirmation of their relevance in LIHC. Our ultimate goal is to identify novel biomarkers that can aid in early detection, diagnosis, and prognosis of LIHC patients. Furthermore, by pinpointing therapeutic targets, we aim to uncover new potential treatment options that could be explored in future clinical trials. The integration of bioinformatics tools allows us to systematically identify gene signatures associated with key biological processes and pathways in LIHC. Additionally, validating these results through in vitro experiments will help us assess the functional roles of these DEGs and hub genes in tumor cell proliferation, migration, and drug resistance, which are crucial aspects of LIHC pathogenesis. By combining computational predictions with experimental data, we aim to provide new insights into the molecular landscape of LIHC, which could contribute to the development of more effective therapeutic strategies for this aggressive form of liver cancer. This research not only advances our understanding of LIHC but also lays the groundwork for further studies aimed at clinical translation. This study provides valuable insights into the role of hub genes CCNA2, MAD2L1, AURKA, and PTTG1 in LIHC. Our findings revealed that these genes were significantly upregulated in LIHC tissues, with elevated expression associated with advanced disease stages and poorer overall survival. We demonstrated that these hub genes were involved in key cellular processes, such as cell cycle regulation, mitosis, and chromosomal stability, and their overexpression may contribute to tumor progression and metastasis. Furthermore, promoter methylation analysis highlighted the potential impact of epigenetic regulation on gene expression in LIHC, while genetic alterations were minimal. Drug sensitivity analysis suggested that these genes may influence chemotherapeutic resistance, pointing to their potential as predictive biomarkers for treatment response. The functional assays confirmed that hub genes play critical roles in regulating proliferation, colony formation, and migration in HepG2 cells, reinforcing their oncogenic potential in LIHC. Methodology Datasets retrieval, DEGs, and hub genes identification We retrieved four gene expression datasets (GSE19665, GSE54236, GSE55092, and GSE84598) from the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/) [24] for this study. The datasets were selected based on the following criteria: (i)inclusion of both LIHC tumor and corresponding normal liver tissue samples;(ii)availability of sufficient sample size to ensure statistical reliability;(iii)use of well-annotated human gene expression platforms; and.(iv)availability of normalized or raw data suitable for differential expression analysis. (i)inclusion of both LIHC tumor and corresponding normal liver tissue samples; inclusion of both LIHC tumor and corresponding normal liver tissue samples; (ii)availability of sufficient sample size to ensure statistical reliability; availability of sufficient sample size to ensure statistical reliability; (iii)use of well-annotated human gene expression platforms; and. use of well-annotated human gene expression platforms; and. (iv)availability of normalized or raw data suitable for differential expression analysis. availability of normalized or raw data suitable for differential expression analysis. These datasets collectively include transcriptomic profiles of 267 LIHC samples and 169 normal tissue samples (Table 1). Firstly, the analysis was conducted using the Limma package in R to DEGs from each dataset. The top 1000 DEGs from each dataset were selected for further analysis. To ensure comparability across datasets and to reduce potential noise arising from dataset-specific variability, the top 1000 DEGs ranked by adjusted p-value and fold change were selected from each dataset for downstream analysis. This approach is commonly used in integrative bioinformatics studies to focus on the most biologically significant and consistently dysregulated genes while maintaining a manageable gene set size for cross-dataset comparison. To identify common DEGs across the four datasets, a Venn diagram analysis was performed, highlighting the overlapping genes between the datasets. The genes common to all datasets were subjected to protein-protein interaction (PPI) network analysis using the STRING database (https://string-db.org/) [25] to explore functional interactions between the DEGs. A PPI network was constructed based on these interactions. Next, Cytoscape software (Version 3.10.2), along with the CytoHubba plugin, was used to analyze the PPI network and identify hub genes. The degree method was applied to rank the genes based on their connectivity, and the genes with the highest degree values were selected as hub genes. Finally, these hub genes were further investigated through in silico and in vitro experiments to understand their roles in LIHC. Table 1Overview of LIHC and normal samples from GEO datasets used in the studyDataset IDPlatformLIHC SamplesNormal SamplesTotal SamplesGSE19665GPL6480 Illumina HumanWG-6 v2.0242145GSE54236GPL570 Affymetrix Human Genome U133 Plus 2.0 Array8030110GSE55092GPL570 Affymetrix Human Genome U133 Plus 2.0 Array9181172GSE84598GPL10558 Illumina HumanHT-12 V4.0 expression beadchip7237109Total-267169436 Table 1Overview of LIHC and normal samples from GEO datasets used in the studyDataset IDPlatformLIHC SamplesNormal SamplesTotal SamplesGSE19665GPL6480 Illumina HumanWG-6 v2.0242145GSE54236GPL570 Affymetrix Human Genome U133 Plus 2.0 Array8030110GSE55092GPL570 Affymetrix Human Genome U133 Plus 2.0 Array9181172GSE84598GPL10558 Illumina HumanHT-12 V4.0 expression beadchip7237109Total-267169436 Overview of LIHC and normal samples from GEO datasets used in the study Gene enrichment analysis DAVID (https://david.ncifcrf.gov/) is a bioinformatics tool that provides comprehensive functional annotation for large gene or protein datasets [26]. It helps researchers interpret the biological significance of genes by offering insights into Gene Ontology (GO), biological pathways, and protein interactions. DAVID enables the identification of enriched functional terms, gene-disease associations, and pathway involvement, making it a valuable resource for understanding complex biological systems and gene functions. Herein, we utilized DAVID for the gene enrichment analysis of hub genes. Expression analysis of hub genes using TCGA datasets UALCAN (https://ualcan.path.uab.edu/) and OncoDB (https://oncodb.org/) are powerful online platforms for cancer research, providing access to comprehensive molecular data across various cancer types. UALCAN offers an easy-to-use interface for exploring transcriptomic data from TCGA, including gene expression analysis across tumor and normal samples, survival analysis, and clinicopathological correlations [27]. OncoDB, on the other hand, integrates multiple datasets to provide insights into protein-level expression, genetic mutations, and pathway alterations in cancer [28]. Together, these databases enable researchers to conduct in-depth analyses of mRNA and protein expression, aiding in biomarker discovery and therapeutic target identification in cancers like LIHC. In this study, UALCAN and OncoDB databases were used for the expression analysis of hub genes in LIHC samples. Promoter methylation and mutational analysis of hub genes Promoter methylation analysis of hub genes in LIHC was carried out using OncoDB (https://oncodb.org/) database [28]. While for mutational analysis of hub genes in LIHC, we utilized cBioPortal (https://www.cbioportal.org/) database. cBioPortal is an open-access, web-based platform that provides interactive exploration of multidimensional cancer genomics data [29]. It integrates datasets from prominent cancer research initiatives like The Cancer Genome Atlas (TCGA) and the International Cancer Genome Consortium (ICGC). Researchers can explore mutations, copy number alterations, gene expression, and clinical attributes across various cancer types. Prognostic analysis of hub genes KM plotter (https://kmplot.com/analysis/) is an online tool designed for survival analysis based on gene expression data across multiple cancer types [30]. It integrates data from public repositories, such as TCGA and GEO, allowing researchers to explore the prognostic impact of specific genes. By generating Kaplan-Meier survival curves, KM plotter evaluates the association between gene expression and overall survival, progression-free survival, or relapse-free survival. In this work, KM plotter tool was used for analyzing the prognostic values of hub genes across LIHC patients. Correlations of hub genes with immune cells and drug sensitivity in LIHC The Gene Set Cancer Analysis (GSCA, https://guolab.wchscu.cn/GSCA) database is an integrative bioinformatics platform designed to analyze cancer-related gene sets [31]. It combines multi-omics data from sources like TCGA and GTEx, enabling users to explore gene mutations, expression patterns, and pathway alterations in various cancers. GSCA supports the identification of key oncogenes, tumor suppressors, and drug targets, facilitating the discovery of potential biomarkers and therapeutic strategies across cancer types. We used the database in this work to explore correlations of hub genes with immune cells and drug sensitivity in LIHC. Cell lines and culture conditions LIHC cell lines, including HepG2 (HB-8065), Hep3B (HB-8064), and PLC/PRF/5 (CRL-8024) were obtained from the American Type Culture Collection (ATCC, USA), while Huh7, JHH-1, JHH-7, Li-7, and HLF were obtained from the JCRB Cell Bank (Japan). The SNU series cell lines (SNU-182, SNU-387, SNU-398, SNU-423, SNU-449, and SNU-475) were obtained from the Korean Cell Line Bank (KCLB). Mahlavu cells were obtained from DSMZ (Germany). Normal liver models included THLE-2 (CRL-2706) and THLE-3 (CRL-11233) obtained from ATCC, primary human hepatocytes (PHHs) purchased from Lonza (Switzerland) and Thermo Fisher Scientific (USA), NKNT-3 obtained from RIKEN (Japan), WRL-68 (CL-48) from ATCC, and induced pluripotent stem cell (iPSC)-derived hepatocytes from Thermo Fisher Scientific. HCC cell lines were cultured in Dulbecco’s Modified Eagle Medium (DMEM, Gibco) supplemented with 10% fetal bovine serum (FBS, Gibco) and 1% penicillin–streptomycin (Gibco). THLE-2 and THLE-3 cells were cultured in Bronchial Epithelial Cell Growth Medium (BEGM, Lonza) supplemented with the recommended growth factors according to the manufacturer’s instructions. Primary human hepatocytes and iPSC-derived hepatocytes were maintained in hepatocyte culture medium (HCM, Lonza or Thermo Fisher Scientific) supplemented with hepatocyte growth factors. NKNT-3 and WRL-68 cells were cultured in RPMI-1640 medium (Gibco) supplemented with 10% FBS and 1% penicillin–streptomycin. All cells were maintained at 37 °C in a humidified incubator with 5% CO₂. Cell lines were authenticated by short tandem repeat (STR) profiling and routinely tested for mycoplasma contamination prior to use. RNA extraction Total RNA was extracted from the LIHC and normal liver cell lines using Thermo Fisher Scientific TRIzol™ Reagent (Catalog No. 15596026). Cells were harvested at 80–90% confluence, lysed in TRIzol reagent, and incubated for 5 min at room temperature. Chloroform was added to the lysates, followed by vigorous shaking and centrifugation at 12,000 × g for 15 min at 4 °C. The aqueous phase was collected, and RNA was precipitated with isopropanol. The concentration and purity of RNA were determined using a NanoDrop™ 2000 spectrophotometer (Thermo Fisher). cDNA synthesis cDNA was synthesized using the Thermo Fisher Scientific High-Capacity cDNA Reverse Transcription Kit (Catalog No. 4368814). Real-Time Quantitative PCR (RT-qPCR) The expression levels of hub genes CCNA2, MAD2L1, AURKA, and PTTG1 were quantified using Thermo Fisher Scientific TaqMan™ Gene Expression Assays and the TaqMan™ Universal PCR Master Mix (Catalog No. 4304437). The specific TaqMan assays for each gene were as follows: CCNA2: Assay ID Hs00996788_m1, MAD2L1: Assay ID Hs00980633_m1, AURKA: Assay ID Hs01582072_m1, and PTTG1: Assay ID Hs00153318_m1 the RT-qPCR reactions were performed in triplicate in a 96-well format using the Applied Biosystems™ 7500 Real-Time PCR System (Thermo Fisher). GAPDH (Assay ID Hs02758991_g1) was used as the internal control to normalize gene expression levels. The relative gene expression was calculated using the 2^−ΔΔCt method. The expression of CCNA2, MAD2L1, AURKA, and PTTG1 was compared between the LIHC and normal liver cell lines. Knockdown of PTTG1, CCNA2, MAD2L1, and AURKA genes HepG2 cells were transfected with small interfering RNAs (siRNAs) targeting PTTG1, CCNA2, MAD2L1, and AURKA (Thermo Fisher Scientific) using Lipofectamine™ RNAiMAX (Thermo Fisher Scientific), according to the manufacturer’s instructions. A non-targeting siRNA was used as a negative control. Cells were harvested 48 h after transfection to evaluate knockdown efficiency. For Western blot analysis, a protocol given in a previous study was followed [32]. Cell proliferation assay Cell proliferation was assessed using the Cell Counting Kit-8 (CCK-8; Dojindo Laboratories). Cells transfected with siRNAs or negative control siRNA were seeded in 96-well plates (5,000 cells/well). At 24-, 48-, and 72-hours post-seeding, 10 µL of CCK-8 reagent was added to each well and incubated for 2 h at 37 °C. Absorbance was measured at 450 nm using a microplate reader (BioTek). Colony formation assay For colony formation, cells transfected with siRNAs were seeded in 6-well plates at a density of 500 cells/well. Cells were cultured for 10–14 days, fixed with 4% paraformaldehyde, and stained with 0.5% crystal violet. Colonies containing > 15 cells were counted manually. Images were acquired using a light microscope. Wound healing assay Cells transfected with siRNAs were grown to confluency in 6-well plates. A scratch was introduced using a sterile 200 µL pipette tip, and cells were washed with PBS to remove debris. Fresh serum-free medium was added, and images of the wound were taken at 0 and 24 h using an inverted microscope. Wound closure was quantified by measuring the wound area using ImageJ software. Statistics DEGs were identified using the Limma package in R, with an adjusted p-value < 0.05 and |log2FC| > 1. Limma applies empirical Bayes moderated two-tailed t-tests, which assume approximate normal distribution of expression values after data normalization. Venn diagram analysis was applied to identify overlapping DEGs across datasets. Functional enrichment analysis of hub genes was performed using DAVID with a significance threshold of p < 0.05 (two-tailed). Survival analysis was conducted using the KM plotter tool, where statistical significance was determined using the two-tailed log-rank test. Correlation with immune infiltration and drug sensitivity was explored using the Pearson correlation method, which assumes linear relationships and approximate normal distribution of variables. Given the large sample sizes in TCGA-based datasets, the normality assumption was considered acceptable based on the central limit theorem. For experimental validation (RT-qPCR), comparisons between two groups were performed using a two-tailed Student’s t-test, assuming normal distribution of data. All statistical tests were considered significant at p < 0.05. Results Identification of DEGs and hub genes Four gene expression profiles of LIHC patients (GSE54236, GSE55092, GSE19665, and GSE84598) were analyzed by Limma package to obtain DEGs. Top 1000 DEGs from each dataset were obtained for further analysis. Next, Venn diagram analysis was conducted to identify common DEGs between these datasets. Figure 1A presents the results of Venn diagram analysis showcasing the overlap of DEGs from GSE54236, GSE55092, GSE19665, and GSE84598 datasets. The numbers inside the overlapping areas indicate the count of shared DEGs between the datasets, with the central region showing 147 DEGs that are common across all four datasets (Fig. 1A). A PPI network of the 147 DEGs was constructed using STRING database (Fig. 1B). Then, hub genes in the constructed PPI network were denoted using CytoHubba application of the Cytoscape software. Based on the degree method, PTTG1, CCNA2, AURKA, and MAD2L1 genes were emerged as hub genes (Fig. 1C-D). These hub genes are connected to each other, suggesting they may have critical roles in regulating disease progression. Fig. 1Identification of overlapping Differentially Expressed Genes (DEGs) and key hub genes in Liver Hepatocellular Carcinoma (LIHC). A Venn diagram illustrating the overlapping DEGs identified across four Gene Expression Omnibus (GEO) datasets (GSE19665, GSE54236, GSE55092, and GSE84598). B Protein-protein interaction (PPI) network of the 147 common DEGs generated using STRING database and visualized in Cytoscape. C PPI network highlighting key hub genes (red nodes), identified through degree method using the CytoHubba plugin in Cytoscape. D Simplified interaction network of the top four hub genes (PTTG1, CCNA2, AURKA, and MAD2L1), suggesting strong interconnectivity and potential regulatory roles in LIHC progression Identification of overlapping Differentially Expressed Genes (DEGs) and key hub genes in Liver Hepatocellular Carcinoma (LIHC). A Venn diagram illustrating the overlapping DEGs identified across four Gene Expression Omnibus (GEO) datasets (GSE19665, GSE54236, GSE55092, and GSE84598). B Protein-protein interaction (PPI) network of the 147 common DEGs generated using STRING database and visualized in Cytoscape. C PPI network highlighting key hub genes (red nodes), identified through degree method using the CytoHubba plugin in Cytoscape. D Simplified interaction network of the top four hub genes (PTTG1, CCNA2, AURKA, and MAD2L1), suggesting strong interconnectivity and potential regulatory roles in LIHC progression Gene enrichment analysis of hub genes Gene enrichment analysis of the hub genes was conducted using DAVID tool. Figure 2 represents the gene enrichment analysis of hub genes in various biological processes, molecular functions, and pathways. Figure 2A shows the cellular component enrichment, where the hub genes are significantly enriched in components such as the “germinal vesicle,” “mitotic checkpoint complex,” and “Cyclin A2-CDK2 complex.” These components are crucial for cell division and regulation, particularly in the mitotic process. The fold enrichment values, represented on the x-axis, indicate a strong over-representation of these components, with higher -log10(FDR) values signifying greater significance. Figure 2B illustrates the molecular function enrichment of the hub genes, with significant enrichment observed in activities such as “histone serine kinase activity,” “protein kinase binding,” and “peptidase inhibitor activity.” These functions are critical for regulating gene expression, protein modification, and signaling pathways, suggesting that the hub genes play essential roles in cellular signaling and protein regulation. Figure 2C focuses on the biological process enrichment, showing that the hub genes are highly enriched in processes like “histone phosphorylation,” “cell cycle,” “mitotic nuclear division,” and “negative regulation of chromosome segregation.” These processes are closely related to the regulation of the cell cycle, particularly during mitosis, and are essential for maintaining genomic stability. Finally, Fig. 2D presents pathway enrichment analysis, revealing that the hub genes are involved in pathways such as “progesterone-mediated oocyte maturation,” “cell cycle,” “acute myeloid leukemia,” and “viral carcinogenesis (Fig. 3).” The involvement of these pathways indicates that the hub genes may be pivotal in cancer progression and cell cycle dysregulation, making them potential targets for therapeutic intervention in diseases like cancer. The size of the dots corresponds to the number of genes involved, and the color gradient indicates the significance of the enrichment, with higher values representing more significant findings. Fig. 2Functional enrichment analysis of hub genes in Liver Hepatocellular Carcinoma (LIHC) using DAVID (https://david.ncifcrf.gov/). A Gene Ontology (GO) cellular component enrichment analysis of hub genes. B GO molecular function enrichment analysis of hub genes. C GO biological process enrichment analysis of hub genes. D Pathway enrichment analysis of hub genes. The fold enrichment values are presented on the x-axis, and the number of enriched genes is denoted by the dot size, with color representing the statistical significance (log10 FDR). P-value < 0.05 Functional enrichment analysis of hub genes in Liver Hepatocellular Carcinoma (LIHC) using DAVID (https://david.ncifcrf.gov/). A Gene Ontology (GO) cellular component enrichment analysis of hub genes. B GO molecular function enrichment analysis of hub genes. C GO biological process enrichment analysis of hub genes. D Pathway enrichment analysis of hub genes. The fold enrichment values are presented on the x-axis, and the number of enriched genes is denoted by the dot size, with color representing the statistical significance (log10 FDR). P-value < 0.05 Fig. 3Top three highly enriched pathways with hub genes in Liver Hepatocellular Carcinoma (LIHC). A The progesterone-mediated oocyte maturation pathway, highlighting the key roles of hub genes in regulating processes such as G2/M transition, MAPK signaling, and cell cycle progression. Genes highlighted in red are significantly enriched and play a critical role in this pathway. B The cell cycle pathway, demonstrating significant involvement of hub genes in the regulation of cell cycle checkpoints, DNA replication, and mitosis. Red-highlighted genes represent those enriched in the pathway, underscoring their role in hepatocellular carcinoma progression. C The oocyte meiosis pathway, showing the involvement of hub genes in processes related to meiotic progression and chromosomal segregation. Red-highlighted genes indicate significant enrichment within this pathway Top three highly enriched pathways with hub genes in Liver Hepatocellular Carcinoma (LIHC). A The progesterone-mediated oocyte maturation pathway, highlighting the key roles of hub genes in regulating processes such as G2/M transition, MAPK signaling, and cell cycle progression. Genes highlighted in red are significantly enriched and play a critical role in this pathway. B The cell cycle pathway, demonstrating significant involvement of hub genes in the regulation of cell cycle checkpoints, DNA replication, and mitosis. Red-highlighted genes represent those enriched in the pathway, underscoring their role in hepatocellular carcinoma progression. C The oocyte meiosis pathway, showing the involvement of hub genes in processes related to meiotic progression and chromosomal segregation. Red-highlighted genes indicate significant enrichment within this pathway Expression analysis of hub genes The expression analysis of hub genes (CCNA2, MAD2L1, AURKA, and PTTG1) at both mRNA and protein levels was conducted using databases, including UALCAN and OncoDB. Figure 4A highlights the mRNA expression of these hub genes in LIHC samples using the UALCAN database. Across all genes, a significant (p-value < 0.05) upregulation in LIHC samples compared to normal tissues was observed, as indicated by the higher transcript per million (TPM) values in tumors. Specifically, CCNA2 shows a pronounced increase in tumor tissues, emphasizing its role in cell cycle regulation and proliferation (Fig. 4A). Similarly, MAD2L1 was elevated in LIHC tumors, reflecting its involvement in mitotic checkpoint control (Fig. 4A), while AURKA was notably upregulated in LIHC samples, suggesting its role in mitosis and chromosomal stability (Fig. 4A). PTTG1 also displays high expression in tumors, reinforcing its association with tumorigenesis despite lacking protein-coding capacity (Fig. 4A). Figure 4B confirms these findings using the OncoDB database, where the same trend of significantly (p-value < 0.05) higher expression of hub genes in LIHC tumor samples compared to normal tissues was evident. Figure 4C focuses on protein expression analysis of CCNA2, MAD2L1, and AURKA in LIHC samples using UALCAN. Since PTTG1 was unavailable on the UALCAN, it was excluded from protein analysis. The protein levels of CCNA2, MAD2L1, and AURKA were significantly (p-value < 0.05) higher in primary tumor samples compared to normal tissues, reinforcing the trends seen at the mRNA level (Fig. 4C). These findings suggest that the dysregulated expression of these genes at both the transcript and protein levels contributes to the pathogenesis of LIHC. Fig. 4Expression validation of hub genes in Liver Hepatocellular Carcinoma (LIHC) tissue samples compared to normal tissues. A UALCAN-based (https://ualcan.path.uab.edu/) boxplots showing the mRNA expression levels (transcript per million) of CCNA2, MAD2L1, AURKA, and PTTG1 in normal (n = 50) and LIHC (n = 371) samples. B OncoDB-based (https://oncodb.org/) mRNA expression analysis of CCNA2, MAD2L1, AURKA, and PTTG1 in normal (n = 50) and LIHC (n = 371) samples. C UALCAN-based protein expression levels (Z-value) of CCNA2, MAD2L1, and AURKA in clear cell renal cell carcinoma (RCC) from CPTAC samples. Due to limited availability of CPTAC-based LIHC proteomic data in UALCAN, RCC data were used as a reference to assess protein-level expression trends of hub genes across cancer types. P*-value < 0.05 Expression validation of hub genes in Liver Hepatocellular Carcinoma (LIHC) tissue samples compared to normal tissues. A UALCAN-based (https://ualcan.path.uab.edu/) boxplots showing the mRNA expression levels (transcript per million) of CCNA2, MAD2L1, AURKA, and PTTG1 in normal (n = 50) and LIHC (n = 371) samples. B OncoDB-based (https://oncodb.org/) mRNA expression analysis of CCNA2, MAD2L1, AURKA, and PTTG1 in normal (n = 50) and LIHC (n = 371) samples. C UALCAN-based protein expression levels (Z-value) of CCNA2, MAD2L1, and AURKA in clear cell renal cell carcinoma (RCC) from CPTAC samples. Due to limited availability of CPTAC-based LIHC proteomic data in UALCAN, RCC data were used as a reference to assess protein-level expression trends of hub genes across cancer types. P*-value < 0.05 Expression of hub genes in different clinical variables The expression of hub genes (CCNA2, MAD2L1, AURKA, and PTTG1) across different clinical variables (such as cancer stage, race, gender, and age) in LIHC patients was analyzed via UALCAN. CCNA2 expression was significantly (p-value < 0.05) elevated in LIHC tissues across all cancer stages, with the highest expression in stage 3, compared to normal tissues. Racial stratification revealed higher expression in Caucasian and African-American patients, while Asian patients had slightly lower levels. Gender-wise, males exhibited significantly higher CCNA2 expression than females (Fig. 5A). The expression was highest in the 41–60 age group, followed by the 61–80 age group, with both having significantly higher expression than younger (21–40) and older (81–100) groups (Fig. 5A). Similarly, MAD2L1 was significantly (p-value < 0.05) up-regulated across all cancer stages, with the peak in stage 3. Expression by race showed higher levels in Caucasian and African-American patients compared to normal controls, with Asian patients having comparatively lower levels (Fig. 5B. Males had higher MAD2L1 expression than females, and age analysis mirrored that of CCNA2, with the 41–60 age group showing the highest expression (Fig. 5B). AURKA expression followed a similar pattern, being significantly (p-value < 0.05) up-regulated across all cancer stages, with stage 3 showing the highest levels. Racially, Caucasians and African-Americans showed elevated AURKA expression, while Asian patients had lower levels (Fig. 5C). Male patients displayed higher AURKA expression compared to females, and again, the 41–60 age group exhibited the highest expression, with lower levels in the younger and older age groups (Fig. 5C). Lastly, PTTG1 expression was significantly (p-value < 0.05) elevated across all cancer stages, particularly in stage 3. Race-based analysis revealed higher PTTG1 expression in Caucasian and African-American patients, with Asian patients showing lower expression (p-value < 0.05). Males had higher PTTG1 expression compared to females, and the highest expression was observed in the 41–60 age group, with significantly lower levels in both younger and older groups (p-value < 0.05). Overall, these findings highlight the elevated expression of these hub genes in LIHC, particularly in certain demographic and clinical subgroups, suggesting their potential relevance in the progression and prognosis of the disease. Fig. 5Expression of hub genes in Liver Hepatocellular Carcinoma (LIHC) across different cancer stages, races, genders, and age groups. A UALCAN-based (https://ualcan.path.uab.edu/) CCNA2 expression analyzed by cancer stages (Normal, Stage 1, Stage 2, Stage 3, Stage 4), race (Caucasian, African-American, Asian), gender (Male, Female), and age groups (21–40, 41–60, 61–80, 81–100 years). B MAD2L1 expression analyzed by the same variables. C AURKA expression analyzed by the same variables. D PTTG1 expression analyzed by the same variables. P*-value < 0.05 Expression of hub genes in Liver Hepatocellular Carcinoma (LIHC) across different cancer stages, races, genders, and age groups. A UALCAN-based (https://ualcan.path.uab.edu/) CCNA2 expression analyzed by cancer stages (Normal, Stage 1, Stage 2, Stage 3, Stage 4), race (Caucasian, African-American, Asian), gender (Male, Female), and age groups (21–40, 41–60, 61–80, 81–100 years). B MAD2L1 expression analyzed by the same variables. C AURKA expression analyzed by the same variables. D PTTG1 expression analyzed by the same variables. P*-value < 0.05 Promoter methylation, genetic alteration, and survival analysis of hub genes Next, a comprehensive promoter methylation, mutational, and survival analyses of hub genes (CCNA2, MAD2L1, AURKA, and PTTG1) were conducted using OncoDB, cBioPortal, and KM plotter tools. Figure 6A shows the promoter methylation analysis of these genes via OncoDB, where promoter regions exhibit hypermethylation in LIHC compared to normal tissues, which may influence gene expression regulation (Fig. 6A). Figure 6B presents a mutational analysis of hub genes using cBioPortal, showing that alterations were found in only 1.37% of 364 LIHC samples. Among these, AURKA showed the highest alteration rate at 1%, primarily through missense mutations and multi-hit events, while the other genes (CCNA2, MAD2L1, and PTTG1) had no detected mutations. Figure 6C provides details on variant classification, with the majority being single nucleotide polymorphisms (SNPs) leading to missense mutations, and the most common nucleotide changes observed were C > T transitions. Lastly, Fig. 6D shows survival analysis using the KM plotter tool, highlighting the significant (p-value < 0.05) impact of these hub genes on LIHC patient prognosis. High expression of CCNA2, MAD2L1, AURKA, and PTTG1 was associated with poorer overall survival (Fig. 6D), as indicated by hazard ratios (HR) ranging from 1.98 to 2.29 and highly significant p-values. This suggests that these genes may serve as potential prognostic biomarkers in LIHC. Fig. 6Promoter methylation level, genomic alterations, and survival impact of hub genes in Liver Hepatocellular Carcinoma (LIHC). A OncoDB-based (https://oncodb.org/) DNA methylation profiles of CCNA2, MAD2L1, AURKA, and PTTG1 in LIHC primary tumor samples (n = 371) compared to normal tissues (n = 41) from The Cancer Genome Atlas (TCGA) dataset. B cBioPortal-based (https://www.cbioportal.org/) Oncoprint summarizing genetic alterations (missense mutations and multi-hit mutations) of CCNA2, MAD2L1, AURKA, and PTTG1 across LIHC patients (n = 364). C Analysis of hub gene variant classification, variant types, and single nucleotide variant (SNV) classes in LIHC. D KM plotter-based (https://kmplot.com/analysis/) Kaplan-Meier survival curves for LIHC patients stratified by the expression levels of CCNA2, MAD2L1, AURKA, and PTTG1. P-value < 0.05 Promoter methylation level, genomic alterations, and survival impact of hub genes in Liver Hepatocellular Carcinoma (LIHC). A OncoDB-based (https://oncodb.org/) DNA methylation profiles of CCNA2, MAD2L1, AURKA, and PTTG1 in LIHC primary tumor samples (n = 371) compared to normal tissues (n = 41) from The Cancer Genome Atlas (TCGA) dataset. B cBioPortal-based (https://www.cbioportal.org/) Oncoprint summarizing genetic alterations (missense mutations and multi-hit mutations) of CCNA2, MAD2L1, AURKA, and PTTG1 across LIHC patients (n = 364). C Analysis of hub gene variant classification, variant types, and single nucleotide variant (SNV) classes in LIHC. D KM plotter-based (https://kmplot.com/analysis/) Kaplan-Meier survival curves for LIHC patients stratified by the expression levels of CCNA2, MAD2L1, AURKA, and PTTG1. P-value < 0.05 Correlation of hub genes expression with immune infiltration and drug sensitivity in LIHC Correlation of hub genes (PTTG1, CCNA2, MAD2L1, and AURKA) expression with immune infiltration and drug sensitivity in LIHC were explored using the GSCA database. Figure 7A shows that PTTG1, CCNA2, MAD2L1, and AURKA expression positively correlates with infiltration levels of various immune cells, such as Tregs, B cells, and CD8 + T cells, with strong associations observed for Treg and exhausted CD4 + T cells (Fig. 7A). Conversely, a negative correlation was observed with M2 macrophages and monocytes, indicating potential immune modulation by these genes in LIHC (Fig. 7A). Furthermore, Fig. 7B demonstrates the correlation between gene expression and drug sensitivity from the GDSC dataset. AURKA expression shows strong positive correlations with sensitivity to several drugs, including Nutlin-3a, VX-11e, and BAY-61-3606, indicating that higher expression of AURKA may enhance the sensitivity of LIHC cells to these treatments (Fig. 7B). PTTG1 also showed positive correlations with sensitivity to these drugs, while CCNA2 and MAD2L1 displayed weaker associations (Fig. 7B). These results suggest that these hub genes not only influence immune infiltration in LIHC but may also play a role in determining the sensitivity of cancer cells to specific chemotherapeutic agents, potentially guiding personalized treatment strategies. Fig. 7Correlation analysis of hub gene expression with immune infiltrates in Liver Hepatocellular Carcinoma (LIHC) and drug sensitivity in Genomics of Drug Sensitivity in Cancer (GDSC). A GSCA-based (https://guolab.wchscu.cn/GSCA) heatmap shows the correlation between the expression of four hub genes (PTTG1, CCNA2, MAD2L1, and AURKA) and various immune cell infiltrates in LIHC. B GSCA-based heatmap shows the correlation between gene expression of hub genes (PTTG1, CCNA2, MAD2L1, and AURKA) and drug sensitivity from GDSC database. P-value < 0.05 Correlation analysis of hub gene expression with immune infiltrates in Liver Hepatocellular Carcinoma (LIHC) and drug sensitivity in Genomics of Drug Sensitivity in Cancer (GDSC). A GSCA-based (https://guolab.wchscu.cn/GSCA) heatmap shows the correlation between the expression of four hub genes (PTTG1, CCNA2, MAD2L1, and AURKA) and various immune cell infiltrates in LIHC. B GSCA-based heatmap shows the correlation between gene expression of hub genes (PTTG1, CCNA2, MAD2L1, and AURKA) and drug sensitivity from GDSC database. P-value < 0.05 Expression analysis of hub genes across LIHC cell lines via RT-qPCR Finally, the expression of hub genes across LIHC and normal control cell lines was validated using the RT-qPCR technique. In Fig. 8A, the box plots display significantly (p-value < 0.05) higher expression levels of all four genes in the LIHC group (n = 15) compared to the normal control group (n = 7). For CCNA2, the expression in LIHC cells is notably elevated compared to normal cells. Similarly, MAD2L1 shows a marked increase in expression in LIHC cells. Both AURKA and PTTG1 follow the same trend, with significant upregulation in LIHC cells, highlighting the potential roles of these genes in LIHC tumorigenesis (Fig. 8A). Figure 8B features ROC (Receiver Operating Characteristic) curves for each gene, which assess their diagnostic performance in distinguishing LIHC from normal controls. The AUC (Area Under the Curve) values are reported as follows: CCNA2 (0.98), MAD2L1 (0.96), AURKA (0.90), and PTTG1 (0.91), indicating excellent diagnostic accuracy, particularly for CCNA2 and MAD2L1, both of which demonstrate near-perfect discrimination between LIHC and normal tissues (Fig. 8B). The high AUC values for all four genes suggest that they could serve as valuable biomarkers for the detection and diagnosis of LIHC. Fig. 8Expression of Hub genes in Liver Hepatocellular Carcinoma (LIHC) cell lines and diagnostic performance analysis. A RT-qPCR-based box plots depicting the expression levels of hub genes (CCNA2, MAD2L1, AURKA, PTTG1) in LIHC cell lines (n = 15) compared to normal cell lines (n = 7). B Receiver Operating Characteristic (ROC) curves for each of the four hub genes (CCNA2, MAD2L1, AURKA, PTTG1). P*-value < 0.05 Expression of Hub genes in Liver Hepatocellular Carcinoma (LIHC) cell lines and diagnostic performance analysis. A RT-qPCR-based box plots depicting the expression levels of hub genes (CCNA2, MAD2L1, AURKA, PTTG1) in LIHC cell lines (n = 15) compared to normal cell lines (n = 7). B Receiver Operating Characteristic (ROC) curves for each of the four hub genes (CCNA2, MAD2L1, AURKA, PTTG1). P*-value < 0.05 Silencing of PTTG1, CCNA2, MAD2L1, and AURKA suppresses proliferation, colony formation, and migration in HepG2 and Huh7 cells To further investigate the functional roles of PTTG1, CCNA2, MAD2L1, and AURKA in hepatocellular carcinoma, we performed siRNA-mediated knockdown experiments in HepG2 and Huh7 cells. Efficient silencing of all four genes was first confirmed at both the mRNA and protein levels. RT-qPCR analysis demonstrated significant (p-value < 0.001) reductions in the transcript levels of PTTG1, CCNA2, MAD2L1, and AURKA following siRNA transfection in HepG2 cells (Fig. 9A) and Huh7 cells (Fig. 10A). Consistently, Western blot analysis showed markedly decreased protein expression of these genes compared with control cells (Figs. 9B and 10B, and supplementary data Fig. 1), with GAPDH serving as the loading control. Next, we assessed the effects of gene silencing on cell proliferation using the CCK-8 assay. As shown in Figs. 9C and 10C, knockdown of PTTG1, CCNA2, MAD2L1, or AURKA significantly (p-value < 0.001) reduced cell viability in both HepG2 and Huh7 cells compared with the control group, indicating a pronounced inhibitory effect on cell proliferation. Colony formation assays further revealed that suppression of these genes substantially impaired the clonogenic ability of hepatocellular carcinoma cells. Both HepG2 and Huh7 cells transfected with siRNAs targeting PTTG1, CCNA2, MAD2L1, or AURKA formed significantly (p-value < 0.001) fewer and smaller colonies than control cells (Figs. 9D–E and 10D–E), confirming their essential roles in long-term proliferative capacity. In addition, wound healing assays were performed to evaluate the impact of gene knockdown on cell migration. At 24 h after scratch induction, HepG2 and Huh7 cells with silenced PTTG1, CCNA2, MAD2L1, or AURKA exhibited markedly (p-value < 0.001) reduced wound closure compared with control cells (Figs. 9F–G and 10F–G), indicating impaired migratory ability. Fig. 9 Silencing of PTTG1, CCNA2, MAD2L1, and AURKA suppresses proliferation, colony formation, and migration in HepG2 cells. A RT-qPCR analysis showing efficient knockdown of PTTG1, CCNA2, MAD2L1, and AURKA mRNA levels in HepG2 cells following siRNA transfection. B Western blot analysis confirming reduced protein expression of PTTG1, CCNA2, MAD2L1, and AURKA, with GAPDH used as a loading control. C Cell proliferation assessed by the CCK-8 assay, showing significantly decreased viability in cells transfected with siRNAs targeting PTTG1, CCNA2, MAD2L1, or AURKA compared with control cells. D Representative images of colony formation assays in HepG2 cells after gene silencing. E Quantification of colony numbers, demonstrating a significant reduction in clonogenic capacity following knockdown of each gene. F Representative images of wound healing assays at 0 and 24 h. G Quantification of wound closure percentages at 24 h, indicating impaired migratory ability in siRNA-treated cells. Data are presented as mean ± SD. ***P < 0.001 versus control Silencing of PTTG1, CCNA2, MAD2L1, and AURKA suppresses proliferation, colony formation, and migration in HepG2 cells. A RT-qPCR analysis showing efficient knockdown of PTTG1, CCNA2, MAD2L1, and AURKA mRNA levels in HepG2 cells following siRNA transfection. B Western blot analysis confirming reduced protein expression of PTTG1, CCNA2, MAD2L1, and AURKA, with GAPDH used as a loading control. C Cell proliferation assessed by the CCK-8 assay, showing significantly decreased viability in cells transfected with siRNAs targeting PTTG1, CCNA2, MAD2L1, or AURKA compared with control cells. D Representative images of colony formation assays in HepG2 cells after gene silencing. E Quantification of colony numbers, demonstrating a significant reduction in clonogenic capacity following knockdown of each gene. F Representative images of wound healing assays at 0 and 24 h. G Quantification of wound closure percentages at 24 h, indicating impaired migratory ability in siRNA-treated cells. Data are presented as mean ± SD. ***P < 0.001 versus control Fig. 10Silencing of PTTG1, CCNA2, MAD2L1, and AURKA suppresses proliferation, colony formation, and migration in Huh7 cells. A RT-qPCR analysis confirming effective knockdown of PTTG1, CCNA2, MAD2L1, and AURKA mRNA expression in Huh7 cells. B Western blot analysis showing decreased protein levels of PTTG1, CCNA2, MAD2L1, and AURKA after siRNA transfection, with GAPDH as a loading control. C CCK-8 assay demonstrating significantly reduced cell proliferation in Huh7 cells following knockdown of each gene. D Representative images of colony formation assays in Huh7 cells. E Quantification of colony numbers, showing marked suppression of colony formation upon silencing of PTTG1, CCNA2, MAD2L1, or AURKA. F Representative wound healing images at 0 and 24 h. G Quantitative analysis of wound closure percentages at 24 h, indicating significantly impaired migratory capacity in siRNA-treated cells. Data are presented as mean ± SD. ***P < 0.001 versus control Silencing of PTTG1, CCNA2, MAD2L1, and AURKA suppresses proliferation, colony formation, and migration in Huh7 cells. A RT-qPCR analysis confirming effective knockdown of PTTG1, CCNA2, MAD2L1, and AURKA mRNA expression in Huh7 cells. B Western blot analysis showing decreased protein levels of PTTG1, CCNA2, MAD2L1, and AURKA after siRNA transfection, with GAPDH as a loading control. C CCK-8 assay demonstrating significantly reduced cell proliferation in Huh7 cells following knockdown of each gene. D Representative images of colony formation assays in Huh7 cells. E Quantification of colony numbers, showing marked suppression of colony formation upon silencing of PTTG1, CCNA2, MAD2L1, or AURKA. F Representative wound healing images at 0 and 24 h. G Quantitative analysis of wound closure percentages at 24 h, indicating significantly impaired migratory capacity in siRNA-treated cells. Data are presented as mean ± SD. ***P < 0.001 versus control Discussion LIHC is the most common type of primary liver cancer, accounting for approximately 80–90% of all liver malignancies [33, 34]. Globally, it is a major cause of cancer-related death, with a high prevalence in regions where hepatitis B and C infections are endemic [35, 36]. Despite advances in therapeutic options, the prognosis for LIHC remains poor, primarily due to late-stage diagnosis and resistance to conventional therapies [37]. The molecular mechanisms driving LIHC progression are complex, involving numerous genetic, epigenetic, and environmental factors [38, 39]. A better understanding of the molecular pathways and key regulatory genes involved in LIHC could enhance early detection and lead to more effective therapeutic strategies. In the present study, we performed an integrated bioinformatics and experimental analysis to identify key DEGs in LIHC. By analyzing four independent GEO datasets, we identified 147 common DEGs and further screened PTTG1, CCNA2, AURKA, and MAD2L1 as hub genes through protein–protein interaction (PPI) network analysis. The consistent identification of these genes across multiple datasets highlights their potential robustness and relevance in LIHC pathogenesis. These hub genes are functionally linked to cell cycle regulation, mitotic progression, and chromosomal stability, suggesting that their dysregulation may contribute to uncontrolled proliferation and genomic instability—two hallmarks of cancer. The significant upregulation of these genes in LIHC tissues, particularly in advanced stages, further supports their role in tumor progression rather than initiation. CCNA2 is known to regulate cell cycle progression, particularly during the S and G2/M phases, and its overexpression has been linked to numerous malignancies, including breast cancer, colorectal cancer, and now lung cancer [40, 41]. We observed that CCNA2 expression was particularly elevated in stage 3 LIHC, further emphasizing its role in advanced tumor proliferation and progression. Interestingly, our results revealed strong diagnostic potential for CCNA2 with an AUC of 0.98, one of the highest among the hub genes, emphasizes its utility as a biomarker for early diagnosis in LIHC. MAD2L1 plays a critical role in the mitotic spindle checkpoint, preventing premature anaphase entry [42]. Its dysregulation is associated with chromosomal instability, a hallmark of cancer [43]. In our study, MAD2L1 was significantly upregulated in LIHC and exhibited high diagnostic accuracy (AUC = 0.96). This is consistent with a previous report, which also identified MAD2L1 overexpression in hepatocellular carcinoma tissues, linking it to poor prognosis. Our study adds to this body of knowledge by demonstrating MAD2L1’s potential as both a diagnostic and prognostic marker, as high expression correlated with advanced disease stages and poorer overall survival. AURKA is a serine/threonine kinase critical for proper chromosome segregation and spindle assembly during mitosis [44, 45]. Dysregulation of AURKA has been implicated in tumorigenesis across various cancers, including breast cancer and liver cancer [46, 47]. We found AURKA to be significantly overexpressed in LIHC compared to normal tissues, with an AUC of 0.90, indicating strong diagnostic potential. Moreover, the high expression of AURKA in our study was most notable in LIHC patients at stage 3, reinforcing its role in advanced tumor progression. PTTG, has emerged as an oncogene involved in cancer cell proliferation, invasion, and metastasis [48, 49]. In our study, PTTG1 was significantly overexpressed in LIHC tissues and demonstrated excellent diagnostic performance with an AUC of 0.91. While PTTG1 has been less studied in LIHC compared to the other hub genes, its overexpression has been linked to tumorigenesis in colorectal, breast, and lung cancers [50–53]. Our study provides evidence of PTTG1 overexpression in LIHC, highlighting its potential role in hepatocarcinogenesis. Interestingly, although our analysis showed promoter hypermethylation of CCNA2, MAD2L1, AURKA, and PTTG1 in LIHC, these genes were simultaneously overexpressed at the mRNA level. At first glance, this appears paradoxical, since promoter hypermethylation is classically associated with transcriptional repression. However, this relationship is not always linear in cancer. One possible explanation is that the methylation detected in bulk tumor data may occur at CpG sites that do not directly block transcription factor binding or core promoter activity. In addition, cancer-associated methylation changes are often region-specific, and hypermethylation in one promoter segment may coexist with hypomethylation at critical regulatory elements, enhancer activation, copy number gain, altered chromatin accessibility, or strong oncogenic transcription factor input, all of which may override the suppressive effect of promoter methylation. Another possibility is that the observed methylation pattern reflects tumor heterogeneity, where subpopulations of cells contribute differently to methylation and expression signals. Therefore, the coexistence of promoter hypermethylation and increased expression in our study suggests that epigenetic regulation of these hub genes in LIHC is complex and likely context-dependent rather than purely suppressive. Further locus-specific methylation mapping and functional assays will be necessary to clarify how DNA methylation influences the transcriptional regulation of these genes in LIHC. Notably, genetic alterations in these hub genes were infrequent, indicating that epigenetic dysregulation may play a more prominent role in their aberrant expression. Survival analysis further demonstrated that high expression of these genes is associated with poor overall survival, consistent with previous studies linking cell cycle-related genes to adverse outcomes in LIHC [54]. Our drug sensitivity analysis revealed significant correlations between the expression of hub genes and the response to several chemotherapeutic agents in LIHC. Notably, higher expression of AURKA and PTTG1 was associated with increased sensitivity to compounds such as Nutlin-3a, VX-11e, and BAY-61-3606. However, it is important to emphasize that these findings are based on correlation analyses derived from public pharmacogenomic datasets and do not establish a causal relationship between gene expression and drug response. Therefore, while these results suggest that hub gene expression may be linked to variations in drug sensitivity, they should be interpreted with caution. Further experimental validation, such as drug response assays following siRNA-mediated knockdown of these genes in LIHC cell lines, will be necessary to determine whether these genes directly influence chemotherapeutic sensitivity or resistance. Nonetheless, our findings provide a preliminary framework for exploring the potential role of these hub genes as predictive biomarkers for therapeutic response in LIHC. Furthermore, the findings of this study demonstrate that PTTG1, CCNA2, MAD2L1, and AURKA are critical regulators of proliferation, colony formation, and migration in HepG2 cells, emphasizing their oncogenic roles in LIHC. The significant suppression of cell proliferation following siRNA-mediated knockdown of PTTG1, CCNA2, MAD2L1, and AURKA, as evidenced by reduced metabolic activity in the CCK-8 assay, indicates their involvement in driving cellular growth. Similarly, the marked reduction in colony formation capacity suggests that these genes are essential for the long-term survival and clonogenic potential of LIHC cells. Furthermore, the impaired wound closure in migration assays highlights their role in promoting cell motility, a key process in tumor metastasis. These results align with previous reports implicating PTTG1, CCNA2, MAD2L1, and AURKA in various cancers, where their overexpression correlates with aggressive tumor behavior, poor prognosis, and resistance to therapy [55, 56]. While our study provides valuable insights into the role of hub genes in LIHC, several limitations should be acknowledged. First, the analysis is based on publicly available datasets, which may introduce biases due to variations in experimental conditions, patient populations, and sequencing technologies. This could affect the generalizability of the results. Second, while we identified correlations between hub gene expression and immune cell infiltration, as well as drug resistance, these findings are based on in silico analyses and require further experimental validation in vitro and in vivo to establish causality. Additionally, the functional mechanisms through which these genes contribute to drug resistance and immune modulation remain unclear and warrant further investigation. Conclusion In conclusion, this study emphasizes the significant role of hub genes—CCNA2, MAD2L1, AURKA, and PTTG1—in the pathogenesis of LIHC. By integrating bioinformatics analysis with experimental validation, we have identified these genes as critical regulators of key cellular processes such as cell cycle progression, mitosis, and chromosomal stability. Their upregulation in LIHC tissues, particularly in advanced stages of the disease, highlights their potential as biomarkers for both diagnosis and prognosis, offering insight into the molecular mechanisms underlying tumor progression and metastasis. The findings from promoter methylation and genetic alteration analyses point to the importance of epigenetic regulation in the dysregulation of these genes in LIHC, suggesting that epigenetic therapies might complement traditional treatments. Furthermore, the significant association between the expression of these genes and chemotherapeutic resistance suggests that targeting these genes or their associated pathways could provide novel therapeutic strategies to overcome resistance and improve patient outcomes. Functional assays demonstrated that hub genes are pivotal in regulating the proliferation, migration, and clonogenic potential of hepatocellular carcinoma cells, reinforcing their role as potential therapeutic targets. Despite the promising results, the study is limited by its reliance on publicly available datasets and the need for further in vitro and in vivo validation to fully establish causality and therapeutic relevance. Future research should aim to clarify the functional mechanisms of these genes in LIHC progression and explore their therapeutic potential through clinical and experimental models. Supplementary Information Supplementary Material 1. Supplementary Material 1.