Analysis of the expression and prognostic significance of DDK complex in Hepatocarcinoma

Both the gene expression profiles of HCC patients and the corresponding clinical information were downloaded from the public TCGA HCC database (https:// portal.gdc.cancer.gov/repository) [14], in which 50 paired normal liver tissues and 374 HCC tissues were included. The mRNA expression of DDK complex members was downloaded from three GEO databases (GSE25097, GSE54236, GSE64041) (https://​www.​ncbi.​nlm.​nih.​gov/​geo/​) [15‐17]. All these public data were processed (downloaded, sorted and de-duplicated) by R studio software (R version 3.6.3, Boston, MA, USA). The normalized expression of DDK members was analyzed by GraphPad Prism 9.0 (GraphPad, San Diego, CA, USA).

The survival analysis based on DDC complex expression and clinical information of HCC patients in TCGA HCC dataset was accomplished by online Web-Based Survival Analysis Tool (http://​kmplot.​com/​analysis/​). The best cutoff used to define low and high expression for each analyzed gene is generated as described [18]. Briefly, to find the best cutoff, they iterate over the input variable values from the lower quartile to the upper quartile and compute the Cox regression for each setting. The most significant cut-off value is used as the best cutoff to separate the input data into two groups.

GSEA as a computational method is frequently used to determine whether an a priori defined set of genes shows statistically significant, concordant differences between two biological states. We firstly sorted out the genes that were negatively or positively correlated with our target genes by pearson correlation analysis based on TCGA HCC expression files, and GSEA was then used to identify the potential biological functions of target genes. In this study, GSEA was automatically carried out by the online website tool LinkedOmics (http://​www.​linkedomics.​org/​login.​php) [19]. Genes highly correlated with DBF4 and CDC7 (P < 0.05) were subjected to further analysis (Enrichment analysis, KEGG pathway; Rank criteria, P value; Minimum number of genes, 3; Simulations, 1000). Detailly, 1000 gene set permutations were performed for each analysis. The normalized enrichment score (NES) was calculated and the significantly enriched gene sets were screened. Then, the enrichment pathways with a normal p-value < 0.05 and a false discovery rate (FDR) < 0.25 were selected.

A cohort of 110 patients with confirmed HCC underwent tumor resection at West China Hospital of Sichuan University between 2016 and 2017. OS was defined as the time between initial surgery and death. The paraffin-embedded tumor samples were made into tissue microarray (TMA) cores (1 mm diameter). Immunohistochemistry (IHC) staining was then performed on the TMA slides with CDC7 antibody (ab229187, Abcam, Cambridge, UK), and the results were interpreted by three pathologists blinded to clinical information. The expression level of CDC7 was scored according to the signal intensity and distribution as previously described [20]. Briefly, five 400 × -magnified areas were examined and assigned according to the following categories: 0, < 5%; 1, 5–25%; 2, 25–75%; 3, > 75%. The staining intensity was scored as follows: 1, weak; 2, moderate; and 3, intense. Tumor tissues with an IHC score of 0–3 were designated as low expression, and those with scores of 4–9 were designated as high expression. Classic core clinical features, such as age, AFP level, HBsAg, TNM stage, and tumor size, etc., were adopted to analyze the correlation of CDC7 with HCC. Approval for this study was granted by the Ethics Committee of the West China Hospital, Sichuan University, and each patient provided written informed consent.

Fresh frozen tissues from HCC patients were collected and lysed with RIPA buffer according to the manufacturer’s instructions (Beyotime Biotechnology, Shanghai, China). For each lane, 50 μg protein was loaded for detecting DBF4 and CDC7 on different gels. Proteins were separated with SDS-PAGE and transferred to 0.45 μm PVDF membranes. Membranes were then blocked with PBS containing 0.05% tween and 5% non-fat milk. DBF4 antibody (ab124707, Abcam, Cambridge, UK) and CDC7 antibody (ab229187, Abcam, Cambridge, UK) were used to detect the protein level of DBF4 and CDC7, and GAPDH (ET1601-4, HuaBio, Hangzhou, China) was used as an internal control.

The DSS, PFS, RFS and OS of DDK complex members based on their expression and clinical data were automatically carried out by Kaplan–Meier Plotter (http://​kmplot.​com/​analysis) [18]. All other statistical analyses were conducted by using the SPSS version 19.0 software program (SPSS Inc., Chicago, USA) or GraphPad Prism 7.0 (GraphPad, San Diego, CA, USA). The Kruskal–Wallis test, Wilcoxon rank test and logistic regression were used to analyze the relationship between DDK complex members and the clinical characteristics of HCC. Cox regression analysis was then used to compare the effects of CDC7 and clinical characteristics on OS and RFS. A P value of less than 0.05 was considered statistically significant.

To investigate the expression of DDK complex members in HCC patients, we firstly sorted out the expression profile of DBF4 and CDC7 in TCGA dataset, and we found both DBF4 and CDC7 were significantly highly expressed in HCC tumor tissues (Fig. 1B). Meanwhile, according to an online analysis platform (http://​guotosky.​vip:​13838/​GTBA/​), we found DBF4 and CDC7 are highly correlated in all TCGA tumor types, especially in HCC (R = 7.459, P < 2.2e-16) (Fig. 1C, D). Additionally, we explored the expression of DBF4 and CDC7 from the expression profiles from GSE25097, GSE54236 and GSE64041, in which both of them were higher expressed in HCC tumor tissues compared with adjacent normal tissues and minimum expressed in healthy normal livers (Fig. 1E). Finally, the immunoblotting assay results further confirmed the elevated expression of DBF4 and CDC7 in HCC tissues (Fig. 2A).

Furtherly, we analyzed the expression of DDK complex members in different clinical statuses based on TCGA HCC dataset. As shown, the expression of DBF4 and CDC7 was positively correlated with tumor grade and stage, and was much higher in advanced HCC patients with higher clinical stage and grade (Fig. 2B, C). And both DBF4 and CDC7 were independent risk factors for HCC patients assessed by multiCox proportional hazards regression analyses (Fig. 3A, B). Additionally, according to Kaplan–Meier Plotter, HCC patients with higher DBF4 or CDC7 expression had worse OS, RFS and DSS (Fig. 3C, D).

To further evaluate the prognostic value of DDK complex members in HCC patients, immunohistochemical staining on TMA was carried out. Unfortunately, we didn’t find a suitable DBF4 antibody for the IHC application. Therefore, only CDC7 was evaluated. Based on the IHC staining results, HCC patients were divided into two groups with high (n = 71) or low (n = 39) CDC7 expression, and representative images of staining were shown in Fig. 4A. The association of CDC7 expression and the clinical characteristics was presented in Table 1, and accordingly, the expression of CDC7 was highly correlated with tumor size (P = 0.011) and TNM stage (P = 0.008). Consistent with the findings in TCGA dataset, univariate and multivariate analysis revealed that the expression of CDC7 was closely associated with OS (Table 2), and significantly longer OS was found in patients with lower CDC7 expression (P = 0.009; Fig. 4B).

According to our results above, high expression of DDK complex may play a critical role in HCC development, while the underlying events remain unclear. As a critical regulator in DNA replication, we firstly analyzed the cell proliferation status in CDC7^High and CDC7^Low tumor samples. As demonstrated in Fig. 4C, more proliferating tumor cells were observed in CDC7^High tumor tissues. In order to further explore the possible correlated signal pathways through which DDK complex functions in the development of HCC, we firstly screened the genes that positively or negatively correlated with DBF4 or CDC7 in TCGA HCC data by pearson correlation analysis (Table S1 and S2), and then GSEA was performed on Linkedomics online website (http://​www.​linkedomics.​org/​) to identify the signaling pathways these genes belong to (Fig. 5A-D). According to the P value < 0.05, FDR < 0.05 and normalized enrichment score (NES), significant enrichment pathways were screened, and the top-ranked pathways were presented in Fig. 5B and D. The results demonstrated that genes were enriched in pathways facilitating tumor cell proliferation, especially in cell cycle and DNA replication (Fig. 5B, D and E–H), and the top-ranked associated genes were listed in Fig. 5I. These results suggested that highly expressed DDK complex members promotes DNA replication, thereby enhancing tumor cell proliferation, and that DDK complexes can promote HCC development in this way.