The Analyses Online and Characteristics of Patients.
The expression levels of CDC7 in different cancers in humans were showed in Fig. 1a and 1b. GSEA results in Fig. 1c and 1d showed that cell cycle pathway was closely correlated with high CDC7 expression, while fatty acid metabolism was closely correlated with low CDC7 expression.
A total of 277 patients from initial screening were included in our analyses. The mean age of patients was 59.06 years (SD = 12.94, range 16 - 85). Of the 277 patients,139 with stage I, 70 with stage II, 67 with stage III, 1 with stage IV, 98 (35.37%) patients were CDC7 positive. Based on the immune signature score, we divided the patients into three subtypes with high, medium and low immune scores. The cut-off points of immune scores were − 786.8 and − 427.1(X‐tile plots are shown in the Fig. S1). Totally, 28 (10.11%) patients were lower than or equal to − 786.8 (low immune scores subcohorts), 57 (20.58%) were between − 786.8 and − 427.1 (intermediate immune scores subcohorts), and 192 (69.31%) patients were greater than − 427.1 (high immune scores subcohorts). The median OS time was 20.81 months (range 0 ‐120.82 months) and the median DFS time was 13.08 months (range 0-120.82 months). A summary of differentially expressed CDC7 gene and clinical characteristics for immune scores subcohorts are shown in Table 1.
Table 1
Differentially expressed CDC7 gene and clinical characteristics for immune scores subcohorts in 277 patients with HCC.
| | Immune scores | | |
Characteristics | Total | ≤ -786.8 | -786.8 to -427.1 | > -427.1 | χ2 value | P |
Sample sizes | 277 | 28 | 57 | 192 | - | - |
Age (y) | | | | | 3.971 | 0.868 |
≤ 40 | 24 | 2 | 6 | 16 | | |
40–50 | 34 | 3 | 7 | 24 | | |
50–60 | 85 | 12 | 16 | 57 | | |
60–70 | 85 | 8 | 15 | 62 | | |
> 70 | 49 | 3 | 13 | 33 | | |
Sex | | | | | 1.311 | 0.519 |
Male | 183 | 17 | 35 | 131 | | |
Female | 94 | 11 | 22 | 61 | | |
Race | | | | | 0.718 | 0.698 |
White | 134 | 12 | 26 | 96 | | |
Asian | 143 | 16 | 31 | 96 | | |
Stage | | | | | 7.049 | 0.133 |
Stage I | 139 | 10 | 26 | 103 | | |
Stage II | 70 | 8 | 12 | 50 | | |
Stage III - IV | 68 | 10 | 19 | 39 | | |
CDC7 | | | | | 2.972 | 0.226 |
Negative | 179 | 14 | 37 | 128 | | |
Positive | 98 | 14 | 20 | 64 | | |
Univariate and Multivariate Analyses for DFS and OS.
As shown in Fig. 2 to Fig. 5, there were statistically significant differences in univariate analysis/multivariate analysis for DFS and OS among patients with CDC7 positive (hazard ratios [HR]:1.6, 95% confidence interval [CI]: 1.1–2.2, P = 0.013; HR:2.1, CI:1.4–3.3, P < 0.001; HR:1.62, CI:1.12–2.36, P = 0.011; HR: 2.18, CI:1.40–3.39, P < 0.001. Meantime, the analyses in Fig. 2 to Fig. 5 indicating the lower immune scores are, the worse the prognosis is for the patient with HCC. The results in Fig. S2 also confirmed this tendency.
High CDC7 Expression Impacts the Prognosis of Patients with HCC.
Kaplan–Meier survival curves in Fig. 1e and 1f demonstrated that high expression level of CDC7 in patients with HCC was significantly correlated with shorter DFS and OS. This was consistent with previously published results [22].
The prognostic nomogram that integrated all considered independent factors for DFS and OS were shown in Fig. 6. The C-index for DFS and OS predictions were 0.662 (95%CI, 0.688‐0.636) and 0.692(95% CI,0.66‐0.724), respectively. Figure 7 showed that ROC for nomograms for DFS (AUC of 3-year = 0.686, AUC of 5-year = 0.67) and nomograms for OS (AUC of 3-year = 0.76, AUC of 5-year = 0.728).
Correlation Analysis between CDC7 Expression and Immune Cells.
The scatterplots in Fig. 8 revealed the high CDC7 expression was significantly correlated with CD4 + T cells, macrophages and DCs in HCC. Macrophages are divided into M1/pro-inflammatory macrophages and M2/anti-inflammatory macrophages, which possess different cytokine profiles. M2 macrophages can promote HCC cells invasion and migration via miR-149-5p/MMP9 signaling in previous research [23]. And tumor-associated macrophages (TAMs) are also important components of the tumor microenvironment, which have been shown to make malignant tumor cells acquired cell migration and invasion capabilities [24, 25]. For each type of above cells in Fig. 9 we determined the number of three types biomarkers according to previous studies [26]. The partial correlation coefficient between CDC7 expression and biomarkers of TAMs was 0.227 (p < 0.01), 0.269 (p < 0.01) and 0.333 (p < 0.01), respectively. The partial correlation coefficient between CDC7 expression and biomarkers of M1 macrophages was 0.382 (p < 0.01), 0.02 (p > 0.05) and 0.201(p < 0.01), respectively. The partial correlation coefficient between CDC7 expression and biomarkers of M2 macrophages was 0.383 (p < 0.01), 0.539 (p < 0.01) and 0.459 (p < 0.01), respectively.
Correlation Analysis between CDC7 Expression and HCC Metastasis.
Some biomarkers, such as NUF2, ENO1 and FOXM1, have been verified to be associated with HCC cell invasion and metastatic ability [27–32]. The partial correlation coefficient between CDC7 expression and above biomarkers was 0.777 (p < 0.01), 0.438 (p < 0.01) and 0.796 (p < 0.01) in Fig. 10, respectively.