TCGA is a comprehensive, freely accessible database that contains over 20,000 primary cancer and matched normal samples across 33 cancer types (https://www.cancer.gov/about-nci/organization/ccg/research/structural-genomics/tcga). It provides high dimensionality and quality data, including genomic, epigenomic, transcriptomic, and proteomic data. The details of the data include clinical, gene expression, copy number variation, mutation, methylation and protein (19).

GTEx (https://www.gtexportal.org/) contains genotype data of nearly 17,382 RNA-seq samples across 54 tissue sites (from 948 postmortem donors) and 2 cell lines. Full gene expression datasets are available for free. This portal established a comprehensive catalog of genetic variants that affect gene expression across multiple tissues, for the research community, to evaluate tissue-specific gene expression and regulation in numerous different tissues.

Oncomine (https://www.oncomine.org) is a public database that can easily obtain microarray data across different cancers. The mRNA levels of CDC45 were compared between cancer and control tissues using the following threshold: P<0.05; fold change, >2; gene rank, Top 10%.

TISIDB (http://cis.hku.hk/TISIDB/index.php) is an online website that focuses on the interactions between cancer and the immune system and contains multiple heterogeneous data types (20). The association between CDC45 expression and the grade or stage of different cancers from TCGA was investigated. In addition, it was also used to analyze the interaction between CDC45 levels in different immune subtypes, including C1 (wound healing), C2 (IFN-γ dominant), C3 (inflammatory), C4 (lymphocyte depleted), C5 (immunologically quiet), and C6 (TGF-β dominant).

TIMER (http://timer.cistrome.org/) is a comprehensive online website to explore tumor immunological, clinical and genomic features (21). Moreover, gene correlation analysis was also performed using TIMER across 32 cancers. The ‘Expression’ module was searched using the keyword ‘CDC45’ for CDC45 expression in different cancers and matched normal samples from TCGA. Similarly, the ‘Immune’ module was used to investigate the correlation between CDC45 and different immune cells, including CD8⁺ T cells, CD4⁺ T cells, neutrophils, myeloid dendritic cells, macrophages, and B cells, across all TCGA cancers. Moreover, the ‘Gene_-Corr’ module was adopted to determine the correlation between CDC45 and m⁶A-related genes.

HPA (https://www.proteinatlas.org/) has collected the entire human proteome and characterized it using immunohistochemistry and immunocytochemistry (22). It is also used to detect protein localization and expression in human tissues and cells. From this program, immunohistochemical images of different cancers and matched normal samples were downloaded.

The cBioPortal (https://www.cbioportal.org/) is an open source that contains gene data, including mutation and putative copy number alterations (CNVs) (23). The ‘TCGA Pan Cancer Atlas Studies’ in ‘Quick select’ was searched using the keyword ‘CDC45’ for gene mutations and CNVs across 32 cancers. Then, the summary results in the ‘Cancer Types Summary’ module were obtained. In addition, the details of mutated information of CDC45 are displayed in the diagram of protein structure or 3D graphics. Furthermore, the correlation between CDC45 mutation and clinical outcomes was explored using the ‘Comparison/Survival’ module.

GEPIA is an interactive website for exploring RNA-seq expression from TCGA and GTEx based on the normalized processing pipeline (http://gepia2.cancer-pku.cn/#index). It provides differential expression analysis, cancer type, pathological grade, survival analysis, similar gene detection and correlation analysis (24). By searching different modules, the correlation between CDC45 expression and cancer stage, OS and DFS across 33 cancers was analyzed using the following CDC45 cutoff value: High, 50%; low, 50%. Similar genes for GO and KEGG analyses were also downloaded.

The UALCAN (http://ualcan.path.uab.edu/index.html) database was used to analyze the DNA methylation of CDC45 between different cancer and matched normal samples from TCGA (25). Herein, we compared the methylation level of the CDC45 promoter region between cancer and matched control tissues and found that five cancer types exhibited significant differences, including bladder urothelial carcinoma (BLCA), breast invasive carcinoma (BRCA), testicular germ cell tumors (TGCTs), lung adenocarcinoma (LUAD) and head and neck squamous cell carcinoma (HNSC).

MEXPRESS (https://mexpress.be) is an online website that allows researchers to explore DNA methylation from TCGA (26). Herein, we searched the DNA methylation of BLCA, BRCA, TGCT, LUAD and HNSC to verify the results obtained from UALCAN, and promoter region probes were obtained.

m⁶A-Atlas (http://180.208.58.66/m6A-Atlas/index.html) is a comprehensive online tool for revealing the m⁶A epitranscriptome (27). By selecting ‘Human gene’ and ‘CDC45’, detailed results, including m⁶A records, overall distribution pattern of m⁶A sites, and the distribution of all m⁶A sites and RNA binding proteins (RBPs) were obtained.

The REPIC (https://repicmod.uchicago.edu/repic) database is an online source that records numerous data retrieved from the Gene Expression Omnibus (GEO) and the Sequence Read Archive (SRA) of m⁶A-seq and MeRIP-seq using their unified pipeline (28). This database is used to query m⁶A modification sites by specific cell lines or tissue types. The m⁶A modification sites in different cell lines or tissues of CDC45 were explored.

The STRING database (http://string-db.org) is an online server that focuses on protein-protein interactions (PPIs) and functional protein networks (29). By querying ‘CDC45’ in this database and setting up parameters as follows: Minimum required interaction score [‘low confidence (0.150)’], meaning of network edges (‘evidence’), max number of interactors to show (‘no more than 50 interactors’ in 1st shell) and active interaction sources (‘experiments, database, co-expression’), the top 50 proteins that bind with CDC45 were obtained.

BioGRID (https://thebiogrid.org/), HIPPIE (http://cbdm-01.zdv.uni-mainz.de/~mschaefer/hippie/) and HitPredict (http://www.hitpredict.org/) databases were searched for PPI analysis.

LinkedOmics (http://linkedomics.org/login.php) is a publicly available portal that provides a unique platform for biologists and clinicians to access, analyze and compare cancer multi-omics data within and across tumor types.

To analyze TCGA and GTEx data, R (version 3.6.3) software (https://cran.r-project.org/bin/windows/base/old/3.6.3/) was used. With regard to certain cancers without control samples or those with highly limited normal tissues, the corresponding samples from GTEx were extracted as control groups, and the ‘ggplot2 package’ was used to visualize gene expression. The ‘ggradar’ and ‘ggplot2’ packages were used to reveal the MSI and neoantigens of immunity. The ‘timeROC’ package was used to compare the predictive accuracy of CDC45 and the risk score, and the log-rank test was also used to compare the survival difference between the two groups. The ‘clusterProfiller’ and ‘org.Hs.eg.db’ packages were used for GO and KEGG analyses. P<0.05 was considered to indicate a statistically significant difference.

Two human GBM cell lines, U251 and glioblastoma of unknown origin U87, were purchased from American Type Culture Collection. Cells were cultured in high-glucose Dulbecco's modified Eagle's medium (DMEM; Corning, Inc.) supplemented with 10% fetal bovine serum (FBS; Thermo Fisher Scientific, Inc.) at 37°C and 5% CO₂. These cell lines underwent Mycoplasma tests and were authenticated by Short Tandem Repeat (STR) analysis (Beijing Microread Genetics Co., Ltd).

The sequences of CDC45 siRNA (Hanbio Biotechnology Co., Ltd.) were as follows: Negative control (NC), 5′-UUCUCCGAACGUGUCACGU-3′ (sense) and 5′-ACGUGACACGUUCGGAGAA-3′ (antisense); siCDC45-1, 5′-GCGUGCAGACUUUCAGCAUTT-3′ (sense) and 5′-AUGCUGAAAGUCUGCACGCTT-3′ (antisense); and siCDC45-2, 5′-GCAAGACAAGAUCACUCAA-3′ (sense)and 5′-UUGAGUGAUCUUGUCUUGC-3′ (antisense).

U87 and U251 cells were seeded into 24- and 96-well plates and transfected 48 h later using Lipofectamine 3000 (Life Technologies Corporation; Thermo Fisher Scientific, Inc.) according to the manufacturer's protocol. Briefly, for 24-well plates, 1 µl siRNA (20 µM) was transfected into target cells using 1.5 µl Lipofectamine 3000 at room temperature. For 96-well plates, 0.2 µl siRNA (20 µM) was transfected into target cells using 0.3 µl Lipofectamine 3000 at room temperature. Following 48 h of incubation at 37°C and 5% CO₂, subsequent experiments were conducted.

Total RNA was extracted from the cells using TRIzol (Takara Biotechnology Co., Ltd.) according to the manufacturer's protocol. cDNA was synthesized using the GoScript Reverse Transcription System (Promega Corporation) and 1 µg of RNA according to the manufacturer's protocol. qPCR was performed using a SYBR Green PCR kit (Bimake.com) according to the manufacturer's instructions on an Applied Biosystems QuantStudio instrument (Thermo Fisher Scientific, Inc.). Comparative quantification was performed using the 2^−ΔΔCq (30) method with GAPDH as the endogenous control, The PCR system was 20 µl, and the thermocycling conditions were as follows: 40 cycles were operated at 95°C for 3 min, 95°C for 15 sec and 60°C for 1 min. The dissolution curve program was as follows: 95°C for 15 sec, 60°C for 1 min, and 95°C for 1 sec. The primer sequences were as follows: CDC45 forward, 5′-TTCGTGTCCGATTTCCGCAAA-3′ and reverse, 5′-TGGAACCAGCGTATATTGCAC-3′; GAPDH forward, 5′-GGAGCGAGATCCCTCCAAAAT-3′ and reverse, 5′-GGCTGTTGTCATACTTCTCATGG-3′.

Following transfection, U87 and U251 cells were divided into 96-well plates at 3×10³ cells per well. At 24, 48, 72, 96, 120, and 144 h, CCK-8 solution was added (10 µl; Beyotime Institute of Biotechnology), and the cells were then incubated for 2 h at 37°C and 5% CO₂. The optical density (OD) at 450 nm (OD₄₅₀) was then assessed using a microplate reader (Biotek Instruments, Inc.).

U87 and U251 cells (5×10³/well) were inoculated into 24-well plates. Following 48 h of incubation at 37°C and 5% CO₂, an EdU In Vitro Kit (cat. no. C10310-1; Guangzhou RiboBio Co., Ltd.) was employed based on the manufacturer's protocol. Briefly, the EdU medium mixture (50 µM) was discarded, and 4% paraformaldehyde was added to fix cells at room temperature for 30 min. The cells were then washed with glycine (2 mg/ml) for 5 min on a shaker, and 0.2% Triton X-100 was added for 10 min. The cells were then washed twice with PBS, and click reaction buffer (Tris-HCl, pH 8.5, 100 mM; CuSO4, 1 mM; Apollo 550 fluorescent azide, 100 µM; and ascorbic acid, 100 mM) was added for 10–30 min while protecting from light. Subsequently, the cells were washed three times with 0.5% Triton X-100, and stained with Hoechst (5 µg/ml) for 30 min at room temperature. Finally, the cells were washed five times with 0.5% Triton X-100, and then 150 µl PBS was added. Images were obtained using a fluorescence microscope (magnification, ×4; cat. no. IX81; Olympus Corporation).

The cell cycle was measured using flow cytometry and a detection reagent kit (Nanjing KeyGen Biotech Co., Ltd.) according to the manufacturer's protocols. Briefly, 2×10⁶/well cells were seeded into 6-well plates and cultured for 48 h at 37°C and 5% CO₂. Next, cells were collected using trypsin (Beijing Solarbio Science & Technology Co., Ltd.), washed with PBS (centrifuged at 396 × g for 5 min at 4°C) three times, and fixed in 70% ice-cold ethanol at 4°C overnight. After washing with ice-cold PBS (centrifuged at 742 × g for 5 min at 4°C) twice, cells were stained with 500 µl propidium iodide containing RNase A buffer for 30 min at room temperature away from light. The percentage of cells was calculated and analyzed using a BD FACSAria flow cytometer (BD Biosciences), and the results were interpreted using FlowJo version 10 (BD Biosciences).

The experiments were repeated three times independently. Results were analyzed using GraphPad software 8.0 (GraphPad Software, Inc.), and all quantitative data are presented as the means ± SDs. Unpaired Student's t-test was performed to compare differences between two different groups with parametric variables, while one-way ANOVA was used to analyze the significance of multiple group comparisons. Tukey's post hoc test was performed following ANOVA. Spearman's rank correlation coefficient was performed to analyze the strength and direction of association between two ranked variables. The Pearson correlation coefficient was used to assess a linear correlation. All statistical analyses were two-sided, and P<0.05 was considered to indicate a statistically significant difference. Variance was similar between the groups being statistically compared.

To analyze CDC45 expression in different cancers, the Oncomine database, was first searched. Compared with noncancerous samples, levels of CDC45 were markedly elevated in various cancers (Fig. 1A). Next, the relative CDC45 expression across different cancers of TCGA were explored using TIMER 2.0. CDC45 expression was higher in BLCA, BRCA, cervical squamous cell carcinoma and endocervical adenocarcinoma (CESC), cholangiocarcinoma (CHOL), colon adenocarcinoma (COAD), esophageal carcinoma (ESCA), glioblastoma multiforme (GBM), HNSC, kidney chromophobe (KICH), kidney renal clear cell carcinoma (KIRC), kidney renal papillary cell carcinoma (KIRP), liver hepatocellular carcinoma (LIHC), LUAD, lung squamous cell carcinoma (LUSC), pancreatic adenocarcinoma (PAAD), pheochromocytoma and paraganglioma (PCPG), prostate adenocarcinoma (PRAD), rectum adenocarcinoma (READ), stomach adenocarcinoma (STAD), thyroid carcinoma (THCA) and uterine corpus endometrial carcinoma (UCEC) (Fig. 1B) than in matched control samples. With regard to certain cancers without control samples or those with highly limited normal tissues, including adrenocortical carcinoma (ACC), lymphoid neoplasm diffuse large B-cell lymphoma (DLBC), acute myeloid leukemia (LAML), brain lower grade glioma (LGG), ovarian serous cystadenocarcinoma (OV), sarcoma (SARC), skin cutaneous melanoma (SKCM), TGCT, thymoma (THYM), and uterine carcinosarcoma (UCS), GTEx data were utilized as control groups, allowing the further investigation of CDC45 expression between various cancers and their matched control groups. The results revealed that CDC45 was significantly increased in almost all cancers, except for LAML, SARC, mesothelioma (MESO) and uveal melanoma (UVM) (Fig. 1C). The association between CDC45 expression and different cancer stages was further explored in each cancer, and it was determined that CDC45 expression was positively associated with ACC, KICH, KIRC, KIRP, LIHC, LUAD and UCEC, while COAD, OV and READ exhibited negative associations (Fig. 1D). In addition, association between CDC45 expression and different pathological grades of each cancer was evaluated, and the results demonstrated that levels of CDC45 were positively associated with CESC, HNSC, KIRC, LGG, LIHC PAAD and UCEC (Fig. 1E).

As proteins are the substances that execute biological processes, the protein levels of CDC45 across different cancers were next analyzed. Based on HPA, it was revealed that CDC45 protein levels were increased in most cancer tissues compared with matched control samples (Fig. S1).

Given that CDC45 is highly expressed in various cancers, the survival association between CDC45 and different cancers was next analyzed using the TCGA project. The results indicated that high levels of CDC45 were negatively associated with the overall survival in patients with ACC, KICH, KIRC, KIRP, LAML, LGG, LIHC, LUAD, MESO, PAAD, SARC, SKCM and UVM, while CESC showed a positive correlation with OS (Fig. 2). Finally, to further explore the association between CDC45 expression and the predictive ability of OS in different cancers, the raw counts of RNA-sequencing data and clinical information of different cancers from TCGA were downloaded, and time ROC analysis was also performed to compare the predictive accuracy of each gene and the risk score. The area under the ROC curve (AUC) of each cancer in the 1-, 3-, and 5-year survival rates conveyed a high predictive value, especially in ACC (1-year 0.834, 3-year 0.954, and 5-year 0.847), MESO (1-year 0.800, 3-year 0.805, and 5-year 0.856), KICH (1-year 0.983, 3-year 0.758, and 5-year 0.844) and KIRP (1-year 0.739, 3-year 0.769, and 5-year 0.661). These results indicated that CDC45 expression may be a biomarker for the prognosis of these cancers (Fig. 3).

Next, CDC45 alterations in various cancers from the TCGA database were obtained. As revealed in Fig. S2A, the most frequent genetic alteration was ‘mutation’ in UCS (7.02%), which was entirely composed of the ‘amplification’ type, and the same statistical trends occurred in TGCT (1.34%), MESO (1.15%), PCPG (0.56%) and THCA (0.4%). The details of ‘mutation’ were also explored using a diagram (Fig. S2B), including sites, types and case numbers. Missense was the most common mutation type of CDC45 mutation (83/101), followed by truncation (10/101), splice (6/101), in frame (1/101) and SV/fusion (1/101). R175 and X163 had 3 mutations, and the mutation type of R175 is presented in the 3D structure of the CDC45 protein (Fig. S2C).CDC45 and DNA methylation in various cancers. Subsequently, the DNA promoter methylation levels between different cancers and matched control tissues were investigated using UALCAN. It was observed that the beta values were higher in normal samples than in cancer tissues in BLCA, BRCA, TGCT, LUAD and HNSC (Fig. 4; images on the left). In addition, potential probes of DNA methylation across different cancers were searched using MEXPRESS. As revealed in Fig. 4 (images on the right), all promoter probes were negatively correlated with CDC45 DNA methylation, indicating that DNA promoter methylation is reduced in BLCA, BRCA, TGCT, LUAD and HNSC. These results are consistent with the data obtained from UALCAN, suggesting that the promoter methylation level of CDC45 may participate in the processes of tumorigenesis in BLCA, BRCA, TGCT, LUAD and HNSC.

In recent years, immunotherapy has shown great potential in various cancers. It was next examined whether there were associations between cancers and immunology with respect to CDC45. First, the association between CDC45 and different immune cells, including CD8⁺ T cells, CD4⁺ T cells, neutrophils, myeloid dendritic cells, macrophages, and B cells, were analyzed across all TCGA cancers in the TIMER2.0 database. The data revealed that CDC45 was almost positively associated with these immune cells in KIRC, LIHC, THCA and THYM but negatively associated with LUAD, LUSC, and GBM (Fig. 5A). Considering the important role of tumor MSI, which establishes a significant association with the sensitivity to immune checkpoint inhibitors, Fig. 5B revealed that CDC45 was positively associated with MSI in 11 cancers, including HNSC (P=0.017), THCA (P=0.0071), KIRC (P=6.7e-06), STAD (P=0.0013), COAD (P=2.3e-12), BRCA (P=0.012), SARC (P=2.8e-07), LIHC (P=0.01), UCEC (P=4.3e-0.5), PRAD (P=6.3e-06) and LUAD (P=0.0098). Subsequently, the association between CDC45 expression and neoantigens in different cancers was explored, and the data (Fig. 5C) illustrated that CDC45 expression was positively associated with neoantigens in LUAD (P=0.0045), BRCA (P=3.5e-13), UCEC (P=0.00023), STAD (P=0.0011), BLCA (P=0.011), and PRAD (P=9.1e-05). In addition, the association between CDC45 expression and Estimation of stromal and immune cells in malignant tumor tissues using expression data (ESTIMATE) and checkpoints was analyzed (Fig. S3). Moreover, the different immune subtypes were also analyzed across various cancers using TISIDB, including C1 (wound healing), C2 (IFN-γ dominant), C3 (inflammatory), C4 (lymphocyte depleted), C5 (immunologically quiet) and C6 (TGF-β dominant). A significant difference in CDC45 expression was observed in ACC, BLCA, BRCA, COAD, ESCA, KIRC, KIRP, LGG, LIHC, LUAD, LUSC, MESO, OV, PAAD, PRAD, SARC, STAD, TGCT, THCA, UCEC, and UCS. Taking COAD as an example, the C2 subtype exhibited the highest CDC45 expression while the C6 subtype exhibited low CDC45 levels (Fig. 5D). These results indicated that CDC45 expression is associated with the immunology of various tumors.

To further explore the potential roles of CDC45, it was hypothesized that CDC45 levels are related to m⁶A methylation, which is the most abundant modification in eukaryotic cells (19). According to the literature, the core genes of m⁶A methylation, including m⁶A methyltransferase ‘writers’ (CBLL1, KIAA1429, METTL14, METTL3, RBM15, RBM15B, and WTAP), ‘readers’ (HNRNPA2B1, HNRNPC, IGF2BP1, IGF2BP2, IGF2BP3, RBMX, YTHDC1, YTHDC2, YTHDF1, YTHDF2, YTHDF3, and ZNF217) and ‘erasers’ (ALKBL5 and FTO) (20–23) were summarized; hence, the correlation between CDC45 expression and those genes was analyzed. The results revealed that CDC45 expression was primarily positively correlated with ‘writers’ and ‘readers’ in various cancers but negatively correlated with the ‘reader’ in THYM (Fig. 6A and B). However, the correlation between CDC45 levels and ‘eraser’ varied across different cancers (Fig. 6C). The REPIC database was then searched, and the data showed that CDC45 may be regulated by m⁶A posttranslational modification (Fig. S4). These results were further demonstrated using the m⁶A-Atlas database, where a total of 9 studies elucidated the m⁶A positions of CDC45 in different cell lines and tissues (HeLa, HepG2, 293T, 293, ESC, and HCT116) using different techniques (Fig. 6D). The distribution of all the m⁶A sites of CDC45 located on chr 22 (GSE63753) (31) are as follows: 19468535, 19470283, 19484961, 19491689, 19492898, 19495817, 19496177, 19496202, 19502339, 19504159, 19508041, 19508071, respectively (Fig. 6E and F), and 27 RBPs were obtained, including HNRNPC, RBM15 and YTHDC1 (Fig. 6G), which may be immediately regulated through the m⁶A site. These results indicated that CDC45 participates in tumorigenesis via m⁶A posttranslational modification in tumorigenesis.

To further explore the biological significance of CDC45 in tumorigenesis in various cancers, gene co-expression analysis was first performed, and the results showed that CDC45 was significantly correlated with various cancers (Fig. S5). Next, CDC45-binding proteins were screened using the STRING database. Their interaction networks are shown in Fig. 7A, and a total of 50 interactive genes validated by experimental studies were obtained for further research. Another database, GEPIA2, was searched, and the top 100 genes co-expressed with CDC45 were obtained. Subsequently, the data of those genes was combined to perform the GO and KEGG analysis, and the identified biological process (BP) was linked to ‘DNA replication’, ‘cell cycle G1/S phase transition’, ‘G1/S transition of mitotic cell cycle’, ‘DNA-dependent DNA replication’ and ‘DNA replication initiation’ (Fig. 7B), while the cellular component (CC) showed that those genes were primarily enriched in ‘chromosomal region’, ‘chromosome telomeric region’, ‘replication fork’, ‘nuclear replication fork’ and ‘MCM complex’ (Fig. 7C). Finally, the molecular function (MF) of those genes included ‘catalytic activity acting on DNA’, ‘DNA replication origin binding’, ‘single-strand DNA binding’, ‘DNA helicase activity’ and ‘3’-5’ DNA helicase activity’ (Fig. 7D). The KEGG pathway analysis indicated that CDC45 was involved in the ‘cell cycle’, ‘DNA replication’, ‘nucleotide excision repair’, ‘Fanconi anemia pathway’ and ‘mismatch repair’ in the tumorigenesis of various cancers (Fig. 7E). These results were further supported by three other databases BioGRID, HIPPIE and HitPredict (Fig. S6).

Considering the biases of bioinformatics analysis, in vitro experiments were next performed to validate the biological functions of CDC45. Following transfection with siRNA, CDC45 expression in U87 and U251 cells was assessed by RT-qPCR (Fig. 8A). CCK-8 assays were then performed, and the results revealed that knockdown of CDC45 inhibited the proliferation of U87 and U251 cells (Fig. 8B). These results were further supported by the EdU proliferation assay, which revealed that EdU-positive cells were higher in the control group than in the CDC45-knockdown group (Fig. 8C). Finally, cell cycle analysis was also conducted. After knocking down CDC45, cell cycle progression was primarily arrested at the G1 phase (Fig. 8D). These results revealed that CDC45 represents an oncogene in GBM.