Identification of aberrantly methylated differentially expressed genes in prostate carcinoma using integrated bioinformatics

Prostate cancer (PCa) is the second most common malignant tumour in males, and it is the fifth leading cause of cancer mortality [1]. Due to prostate-specific antigen (PSA) screening and advanced biopsy techniques, early-stage PCa patients often show good prognosis after comprehensive treatment. However, PCa is a latent disease and can occur as an asymptomatic tumour in 20- to 30-year-old men [2]. The disease becomes symptomatic in the advanced stage, at which point fewer effective treatment options are available than at earlier stages [1]. Consequently, the overall survival (OS) of patients with advanced PCa is significantly diminished [3, 4]. Therefore, new specific biomarkers for early PCa detection are urgently needed.

In recent years, tumour epigenetic modifications, acknowledged as inherited modifications in gene expression, including DNA methylation, histone acetylation, and noncoding RNA-related modifications, have garnered significant research interest [5]. As the main epigenetic modification, DNA methylation has been extensively studied with respect to angiogenesis, apoptosis, cell cycle regulation, and DNA damage repair [6, 7]. Abnormal methylation, including hypomethylation of oncogenes and hypermethylation of tumour suppressor genes, is intimately involved in tumour pathogenesis and is significantly correlated with patient survival in many cancers [8, 9].

Genetic testing based on microarray and sequencing platforms has emerged as a promising and effective tool to screen significant genetic or epigenetic changes in carcinogenesis and to identify biomarkers useful in diagnosis and prognosis determination [10]. A number of differentially expressed genes (DEGs) and differentially methylated genes (DMGs) have been identified in PCa though microarray analysis [11, 12]. Although some studies have focused on specific genes with aberrant DNA hypermethylation or hypomethylation in PCa, an integrated analysis of gene expression, methylation, and signalling pathway interactions has not been performed.

In the present study, we assessed the interaction network of DEGs and DMGs along with interrelated signalling pathways in PCa by analysing gene expression microarray data (GSE29079), gene methylation microarray data (GSE76938), oncogenes, and tumour suppressor genes (TSGs) using bioinformatic tools; such an analysis has not been reported in previous research. We then used a dataset from the Cancer Genome Atlas (TCGA) as a validation cohort for our findings. We aimed to provide novel insights into the biological characteristics and pathways of DEGs/DMGs in PCa and to identify putative biomarkers important for the development and progression of PCa. Target genes such as AKT1, PRDM10, FASN, and FLNA may play essential roles in the diagnosis and treatment of PCa.

The initiation and development of PCa is a complex and multistage process regulated by genetic and epigenetic changes in pro-tumourigenic oncogenes and anti-tumourigenic TSGs. As with many other tumours, aberrant changes in epigenetic modifications, such as acetylation, phosphorylation, and in particular, DNA methylation, have been detected in PCa [16, 17]. Identifying novel biomarkers in PCa will contribute to improving the diagnosis, treatment, and prognostic assessment of PCa patients.

The GEO database is a free repository of microarray and next-generation sequencing analyses that was used to obtain expression profile (GSE29079) and methylation profile (GSE76938) datasets. R software can be used effectively for the analysis of genes in different groups of samples that are differentially expressed under various experimental conditions. In our study, 3 upregulated hypomethylated oncogenes and 1 downregulated hypermethylated TSG were identified. Functional enrichment analysis revealed that aberrant methylation affected certain pathways and hub genes. These results may offer novel insights into PCa pathogenesis.

DAVID analysis of upregulated hypomethylated genes demonstrated enrichment of the genes in biological processes such as regulation of RNA polymerase II-driven transcription, osteoblast differentiation, intracellular signal transduction, the Wnt signalling pathway, and actin cytoskeleton organization. This finding is consistent with previous studies that have shown that PCa cells stimulate the differentiation of pre-osteoplastic cells through regulators of bone metabolism, thereby facilitating prostate cancer metastasis to bones [18]. Additionally, the Wnt cascade can act as a master regulator by integrating signals from the PI3K/mTOR, MAPK, and AR pathways [19]. The MF category in GO analysis largely showed enrichments in protein and transcription regulatory region DNA binding, ligase activity, and GTPase activator and transcription coactivator activity. Previous studies have reported that GTPase activators regulate intercellular junctions and are disrupted during tumourigenesis [20]. KEGG pathway enrichment analysis suggested significant enrichment in AMPK signalling, cancer, and adherens junction pathways, consistent with the fact that the activated AMPK pathway is involved in the growth and survival of human PCa [21].

Downregulated hypermethylated genes in PCa were enriched for positive regulation of MAPK cascade and phosphatidylinositol 3-kinase signalling, muscle contraction, ageing, and signal transduction in the BP category. The MF category in GO analysis indicated involvement in actin binding, protein binding, PDZ domain binding, glycosaminoglycan binding and transmembrane receptor protein tyrosine kinase activity. Previous studies have shown that actin binding proteins can serve as key suppressors of cell migration and micrometastatic dissemination in PCa [22]. Additionally, tyrosine receptor kinase is an essential regulator of PCa proliferation and tumour growth [23]. KEGG pathway analysis revealed enrichment in ARVC, focal adhesion, HCM, dilated cardiomyopathy, and PI3K-AKT signalling pathways. The role of focal adhesion kinase (FAK) signalling in tumourigenesis and tumour progression has been extensively researched and has led to the development of FAK tyrosine kinase inhibitors as potential anticancer drugs [20]. Activation of the PI3K/AKT pathway has also been shown to play a major role in the aggressive nature of many prostate cancers [24]. Understanding the biological processes and signalling pathways in which aberrantly methylated DEGs are involved can help illuminate the pathogenesis of PCa and identify new therapeutic targets.

In the PPI network generated with Cytoscape, significantly more interactions than expected were observed for the aberrantly methylated DEGs. Importantly, a number of upregulated hypomethylated genes appeared to be involved in the tumourigenesis and tumour progression of PCa. We visualized the networks in Cytoscape, identified hub genes using cytoHubba, and validated the identified oncogenes AKT1, PRDM10, and FASN using the TCGA PRAD patient dataset. The serine/threonine kinase Akt, with 3 isoforms (Akt1, Akt2, and Akt3), plays a critical role in regulating diverse cellular functions, including cell growth, proliferation, survival, transcription, and protein synthesis [25, 26]. Akt1 expression is frequently elevated in breast and prostate cancers [27, 28]. Additionally, Akt1 appears to be robustly involved in the tumourigenesis and invasion of cancer cells [29]. Studies have shown that Akt1 is almost completely hypomethylated in bladder cancer tissues [30]. However, the methylation pattern for Akt1 in PCa tissue has not been defined. The rate of Akt1 mutation in PCa was 9%; we hypothesize that the mutations cause the aberrant methylation and/or upregulation of Akt1. Akt1 is the target of the antitumour drug arsenic trioxide, which is currently used for patients with acute promyelocytic leukaemia [31]. Therefore, it is possible that Akt1 may serve as a potential drug target in other tumour types, including PCa.

PRDM10 is a poorly studied member of the PRDM family. It lacks enzymatic activity and is believed to function as a transcriptional cofactor by recruiting histone-modifying enzymes to target promoters [32]. PRDM10 may serve as a transcriptional regulator for normal tissue differentiation and play important roles in promoting tumour development [33, 34]. In PCa, PRDM10 has been found to be altered in approximately 7% of cases, indicating a similar mechanism may occur in PCa tumourigenesis. It has been reported that breast cancer cells show upregulated expression of PRDM10 associated with hypomethylation of the PRDM10 gene, suggesting the involvement of this gene in the proliferation and invasion of breast cancer cells [35]. Similarly, we found that hypomethylation of PRDM10 in PCa led to high expression in PCa samples, which may affect antitumour activity during tumour development. The Fatty Acid Synthase (FASN) protein-coding gene exhibits high expression levels in tumours, including PCa [36]. This gene has been shown to play critical roles in cancer progression and aggressiveness and to be highly associated with poor prognosis, high risk of disease recurrence, and drug resistance [37, 38]. In our study, we found that hypomethylation of FASN led to high expression of FASN in PCa, indicating that FASN may function as a promoter of PCa tumourigenesis. FASN was altered in approximately 9% of PCa patients; we hypothesize that the mutations may drive the high expression of FASN in PCa. In addition, the FASN-targeting drug cerulenin can dose-dependently decrease HER2/neu protein levels in breast cancer cells (from a 14% decrease at 1.25 mg/L to a 78% decrease at 10 mg/L), and it has been suggested as a possible anticancer treatment [39]. In this study, we identified 3 oncogenes that were highly expressed in PCa samples compared to normal tissues, which suggested that these genes may play vital roles in PCa occurrence. Aberrant methylation of these 3 oncogenes may lead to their upregulated expression in PCa.

We identified FLNA as a TSG that was hypermethylated and downregulated in PCa. It has been reported that FLNA is required for the regulation of cell migration and invasion [40]. However, emerging evidence suggests that it may also be involved in different tumourigenic processes, such as DNA damage and angiogenesis [41, 42]. Downregulation of this gene has been observed in a wide spectrum of human malignancies, including gastric cancers and renal cancers. FLNA has also been shown to be significantly correlated with lymph node metastasis, disease stage, histological grade, and poor OS through promotion of the degradation of MMP-9 [43, 44]. The methylation patterns of FLNA in PCa have not been previously described. In our study, we found that FLNA was hypermethylated and downregulated in PCa, suggesting that aberrant methylation of FLNA in PCa may lead to the deregulation of this TSG, thereby impacting tumour development.

Core module analysis of the PPI network for the upregulated hypomethylated genes suggested that ubiquitin-mediated proteolysis, GABAergic synapses, the cell cycle, endocytosis, purine metabolism, focal adhesion, and biosynthesis of amino acids may be involved in PCa progression. Studies have shown that ubiquitin-mediated proteolysis, endocytosis, and focal adhesion pathways can impact PCa development and progression in different ways [45‐47]. In addition, in the GABAergic synapse pathway, GABA induces GRP secretion via GABBR1 in neuroendocrine-like cells, which is involved in PCa progression [48]. The cell cycle is a vital cellular process involving DNA replication and translation, and it tends to be deregulated in cancer [49]. However, the roles of purine metabolism and biosynthesis of amino acids in PCa, as well as the impact of aberrant methylation is unknown. In our study, the identified hypermethylated downregulated genes were enriched in cancer signalling and Rap1 signalling pathways. Cancer signalling pathways are fundamental for cancer development and sustained carcinogenesis. In addition, aberrant Rap1 activation leads to tumour progression, and it may be induced by cytokines such as galanin [50]. The specific manner in which aberrant methylation affects the functional roles of these pathways in PCa development and progression needs to be investigated in the future.

There were several limitations of the present study. First, the study focused on upregulated hypomethylated and downregulated hypermethylated genes. However, contra-regulated genes were not included; these need to be considered in the future. Second, validation of the aberrantly methylated genes was carried out with TCGA data using in silico approaches. Biological experiments will be necessary to validate these findings in the future. Third, our study was limited to only 2 datasets. Therefore, larger sample sizes are needed to validate the findings of this study. Lastly, more experiments, such as qRT-PCR experiments comparing expression in PCa tissues and normal tissues, should be conducted to confirm the target genes. We have collected PCa tissues and normal tissues, and the results of further analyses will be presented in the future.

In summary, our results identified a series of aberrantly methylated differentially expressed oncogenes and TSGs and their associated pathways in PCa using integrated bioinformatic analysis of gene expression and gene methylation microarray datasets. These results may contribute to a more comprehensive understanding of the molecular mechanisms underlying the occurrence and development of PCa. The 4 hub genes found, namely, AKT1, PRDM10, FASN, and FLNA, were validated using the TCGA PRAD patient dataset. These genes, when aberrantly methylated, may serve as putative biomarkers for the precise diagnosis and treatment of PCa in the future. In comparison to other studies that have focused on individual datasets, our study analysed multiple datasets to produce more robust results regarding gene expression changes and gene modifications that are important in the development and progression of PCa. Future studies will be aimed at validating the functional significance of the identified hub genes in PCa.