Patient demographics and clinical characteristics

In this study, 106 subjects (66 males and 40 females), including 20 individuals in the HI group, 11 in the SL-HBV group, 20 in the CHB group, 18 in the LC group, 18 in the DN group, and 19 in the HCC group, were included; their mean age was 52.32 ± 12.82 years. There were significant differences in the levels of ALT, AST, AFP, TBIL, ALB, and GLO and in the A/G among the HI, SL-HBV, CHB, LC, DN, and HCC groups. There were significant differences in the HBV-DNA load, HBsAg level, HBeAg content, and AFU level among the CHB, LC, DN, and HCC groups. GP73 levels and Child‒Pugh stage were significantly different among the LC, DN, and HCC groups. The demographic and clinical characteristics of all patients are summarized in Table 1.

Table 1 Demographics and clinical characteristics of all subjectsReliability and accuracy analysis of test data

Due to the low content of some samples, we used twice-repeated fluorescence intensity (MFI) data to calculate and analyze cytokine concentrations via 5- and 3-parameter logistic methods. For samples with a lower content, the five-parameter analysis gave a result “less than the lower limit”, while the three-parameter analysis gave specific values. Statistical analysis revealed that the two algorithms were the same for the four chemokines with high expression levels (GRO-α, IL-8, IP-10, and MCP-1). Although there were many samples with IL-6 and IL-10 concentrations at the lower detection limits (1.69 and 2.45, respectively), the lower detection limits of the two experiments were the same, so the results were not significantly different. However, there were many samples with FGF-2, IFN-α2, IL-4, and VEGF-A concentrations at the lower detection limits, and the lower detection limits were significantly different, resulting in significant differences in the data from the two experiments. After combining the two experimental datasets, three- and five-parameter analyses were performed. The results for six cytokines/chemokines (IL-6, IL-10, GRO-α, IL-8, IP-10, and MCP-1) were consistent and reliable, while those for four cytokines (FGF-2, IFN-α2, IL-4, and VEGF-A) were different; moreover, the results of the three-parameter analysis were relatively more reliable.

Expression levels of the 10 cytokines/chemokines in the HI group and HBV infection groups

As shown in Table 2; Fig. 1, the serum expression levels of the 10 cytokines/chemokines were not significantly different between the HI group and the SL-HBV group. Except VEGF-A, the other 9 cytokines were significantly different (P < 0.05) among the four liver disease groups and between the liver disease groups and the HI/SL-HBV group. FGF-2 expression in the CHB group was significantly greater than that in the DN and HCC groups (P = 0.023; P = 0.013) (Fig. 1A). The expression of IFN-α2 in the CHB and HCC groups was significantly greater than that in the HI (P = 0.030; P = 0.050) and SL-HBV groups (P = 0.029; P = 0.045) (Fig. 1B). The expression of IL-4 in the CHB group was the highest, and there were significant differences between the CHB group and the HI, SL-HBV, DN and HCC groups (P = 0.000; P = 0.001; P = 0.000; P = 0.022) (Fig. 1C). In addition, the expression of IL-4 in the LC and HCC groups was greater than that in the HI group (P = 0.000; P = 0.040) (Fig. 1C). The expression of IL-6 in the LC group was the highest, followed by that in the CHB and HCC groups, which significantly differed from that in the HI (P = 0.000; P = 0.000; P = 0.011) and SL-HBV (P = 0.000; P = 0.002; P = 0.030) groups (Fig. 1D). In addition, IL-6 expression differed between the LC and DN groups (P = 0.010) (Fig. 1D). IL-10 expression significantly differed between the HI group and the CHB, DN and HCC groups (P = 0.016; P = 0.021; P = 0.000), and the expression of IL-10 was significantly increased in the HCC group (Fig. 1E). The GRO-α concentrations in the HI and SL-HBV groups were significantly greater than those in the CHB (P = 0.010; P = 0.020), LC (P = 0.000; P = 0.000), DN (P = 0.001; P = 0.002) and HCC (P = 0.000; P = 0.001) groups (Fig. 1F). The expression of IL-8 in the CHB group was the highest, followed by that in the LC group (Fig. 1G). There were significant differences in IL-8 expression between the CHB group and the HI, SL-HBV, DN and HCC groups (P = 0.001; P = 0.011; P = 0.000; P = 0.000); moreover, the LC group and the HI, DN and HCC groups were significantly different (P = 0.011; P = 0.000; P = 0.000) (Fig. 1G). IP-10 was highly expressed in the CHB, LC, DN and HCC groups, and there were significant differences between these four groups and the HI (P = 0.000; P = 0.000; P = 0.000; P = 0.000) and SL-HBV (P = 0.004; P = 0.005; P = 0.003; P = 0.007) groups (Fig. 1H). However, MCP-1 expression significantly differed between the CHB and SL-HBV groups (P = 0.019) (Fig. 1I). These data suggest that the serum immunological markers in the SL-HBV group are similar to those in the HI group, while the CHB, LC, DN and HCC groups have different serum immunological markers due to differences in HBV activity status or disease progression. The expression of cytokines/chemokines, which are in the active phase of HBV replication, was also greater in patients with CHB or LC.

Table 2 Cytokine/chemokine levels of all subjects (pg/ml)Fig. 1

Expression of nine cytokines/chemokines in the HI, SL-HBV, CHB, LC, DN and HCC groups. The nine cytokines/chemokines included FGF-2, IFN-α2, IL-4, IL-6, IL-10, GRO-α, IL-8, IP-10, and MCP-1. (* P < 0.05; ** P < 0.01; *** P < 0.001)

Expression levels of the 10 cytokines/chemokines in HCC patients during different periods after surgical intervention

As shown in Table 3; Fig. 2, the results indicated that the serum expression levels of the 10 cytokines/chemokines were not significantly different between the preoperative group and the two-day postoperative group. However, compared with those in the other two groups, the expression levels of IL-4, VEGF-A and IL-8 tended to increase, and there was a significant difference in the 3-month postoperative group (Fig. 2B, D, F). The expression levels of IL-10 and GRO-α showed a downward trend, and there was a significant difference between the 3-month postoperative group and the other two groups (Fig. 2C, E). The expression of IP-10 decreased in the 2-day postoperative group but increased in the preoperative and 3-month postoperative groups (Fig. 2G).

Table 3 Cytokine/chemokine levels in preoperative and postoperative HCC periods (pg/ml)Fig. 2

Expression of cytokines/chemokines in the CHB group, LC group, and before and after HCC surgery. * P < 0.05; ** P < 0.01; *** P < 0.001)

Almost all HBV-related HCCs in China evolve from LC, and the patients with HCC in this study also had a background of LC. Therefore, the differences among the preoperative HCC group, 2-day postoperative group, 3-month postoperative CHB group and LC group were further analyzed. In the preoperative HCC group, the FGF-2, IL-4, and IL-8 levels were significantly lower than those in the CHB group (Fig. 2A, B, F); the IL-8 expression was also significantly lower than that in the LC group (Fig. 2F); and the IL-10 expression was significantly greater than that in the LC group (Fig. 2C). In the 2-day postoperative group, the IL-4, IL-8, and IP-10 levels were significantly lower than those in the CHB group (Fig. 2B, F, G); the IL-8 and IP-10 levels were also significantly lower than those in the LC group (Fig. 2F, G); and the GRO-α expression was significantly greater than that in the LC group (Fig. 2E). For the 3-month postoperative group, only the GRO-α level was lower than that in the CHB group (Fig. 2E). The above data indicate that the specific cytokine changes observed in the preoperative (tumor-bearing) group might be closely related to HCC, and the markedly elevated IL-10 and markedly decreased IL-8 levels in at least some HCC patients might be related to HCC progression. The immunological status of the 3-month postoperative group who had not relapsed after resection was similar to that of the CHB and LC patients, mainly in terms of the inflammatory phenotype, as they no longer exhibited a phenotype caused by cancer cells.

Correlation analysis of the 10 cytokines/chemokines in HCC patients during different periods after surgical intervention

Studies have shown complex interactions and significant differences in cytokine/chemokine levels at different periods of HCC treatment relative to the time of surgery. Most of the cytokines/chemokines with significant correlation differences were positively correlated both before HCC and after HCC (Fig. 3), implying that the increases in these cytokines/chemokines were mainly accompanied by increases in the levels of other inflammatory molecules. Notably, IL-6 was the marker most closely linked to the preoperative HCC group, and IL-10, IL-8 and MCP-1 contributed the most to this increase (r = 0.767, P = 0.000; r = 0.720, P = 0.001; r = 0.612, P = 0.005; respectively); there was also a positive correlation between IL-10 and IL-8 (r = 0.473, P = 0.041). In addition, FGF-2 was positively correlated with IL-4 (r = 0.515, P = 0.024). At 2 days after HCC surgery, IL-10 expression decreased sharply, and IL-6 and IL-10 were strongly positively correlated (r = 0.731, P = 0.000). FGF-2 was positively correlated with VEGF-A and GRO-α (r = 0.479, P = 0.038; r = 0.485, P = 0.035; respectively); IFN-α2 was positively correlated with IP-10 (r = 0.525, P = 0.021). However, the increase in IL-8 expression was closely related to the increase in IL-6 and MCP-1 at 3 months after HCC surgery (r = 0.617, P = 0.005; r = 0.653, P = 0.002; respectively), and the increase in IL-6 and MCP-1 was strongly correlated (r = 0.733, P = 0.000). However, IP-10 was negatively correlated with IL-4 (r = -0.462, P = 0. 047).

Fig. 3

Correlation analysis of cytokines/chemokines in the preoperative and postoperative HCC periods. Correlation analysis of FGF-2, IFN-α2, IL-4, IL-10, VEGF-A, GRO-α, IL-8, IP-10, and MCP-1 in the HCC preoperative, HCC two-day postoperative and HCC three-month postoperative groups (A). Circos plots were used to illustrate the correlation networks of the 10 cytokines/chemokines in the preoperative and postoperative HCC periods, and only significant correlations (P < 0.05) are shown (B). (* P < 0.05; ** P < 0.01; *** P < 0.001)

Correlations between the 10 cytokines/chemokines and clinical data

In the CHB group, FGF-2 was positively correlated with TBIL (r = 0.451, P = 0.046); IL-10 was positively correlated with ALT, AST and TBIL (r = 0.640, P = 0.002; r = 0.647, P = 0.002; r = 0.612, P = 0.004; respectively) (Fig. 4A); and IP-10 was positively correlated with ALT and AST (r = 0.494, P = 0.027; r = 0.547, P = 0.013; respectively) but negatively correlated with HBsAg (r = -0.460, P = 0.041) (Fig. 4A).

Fig. 4

Correlations between the 10 cytokines/chemokines and clinical data. (* P < 0.05; ** P < 0.01; *** P < 0.001)

In the LC group, the chemokine GRO-α had the closest correlation with the clinical data. GRO-α was positively correlated with GLO, ALP, ALT, AST, HBV-DNA, HBsAg and HBeAg (r = 0.775, P = 0.000; r = 0.577, P = 0.012; r = 0.488, P = 0.040; r = 0.593, P = 0.009; r = 0.564, P = 0.015; r = 0.730, P = 0.001; and r = 0.719, P = 0.001, respectively) but negatively correlated with A/G (r = -0.722, P = 0.001) (Fig. 4B). In addition, IL-8 was positively correlated with GLO, HBV-DNA and HBeAg (r = 0.624, P = 0.006; r = 0.779, P = 0.000; and r = 0.581, P = 0.011, respectively) but negatively correlated with ALB and A/G (r = -0.631, P = 0.005; and r = -0.740, P = 0.000, respectively) (Fig. 4B). MCP-1 was positively correlated with AFP, AFU and GLO (r = 0.474, P = 0.047; r = 0.517, P = 0.028; and r = 0.595, P = 0.009, respectively) but negatively correlated with A/G (r = -0.489, P = 0.040) (Fig. 4B). IL-4 was positively correlated with AST and TBIL (r = 0.539, P = 0.009; r = 0.480, P = 0.044; respectively), while IL-6 was positively correlated with GLO (r = 0.688, P = 0.002) but negatively correlated with A/G (r = -0.591, P = 0.010) (Fig. 4B). VEGF-A was positively correlated with ALT and AST (r = 0.603, P = 0.008; r = 0.530, P = 0.024, respectively) (Fig. 4B).

In the DN group, FGF-2 was positively correlated with GLO (r = 0.575, P = 0.012) but negatively correlated with A/G (r = -0.481, P = 0.043) (Fig. 4C). IL-6 was positively correlated with AFU (r = 0.521, P = 0.027) (Fig. 4C). IL-8 levels were positively correlated with Child‒Pugh stage (r = 0.565, P = 0.015) but negatively correlated with A/G (r = -0.569, P = 0.014) (Fig. 4C).

In the HCC group, FGF-2 was negatively correlated with GP73 and Child‒Pugh stage (r = -0.542, P = 0.017; r = -0.546, P = 0.016, respectively) (Fig. 4D). IFN-α2 was positively correlated with HBV-DNA and HBeAg (r = 0.470, P = 0.042; and r = 0.634, P = 0.004, respectively) (Fig. 4D). IL-4 was positively correlated with AFU and GLO (r = 0.627, P = 0.004; r = 0.555, P = 0.014, respectively) but negatively correlated with A/G (r = -0.479, P = 0.038) (Fig. 4D). IL-6 was positively correlated with ALT (r = 0.523, P = 0.022). IL-10 was positively correlated with ALT (r = 0.502, P = 0.029) (Fig. 4D). GRO-α was positively correlated with AFP (r = 0.511, P = 0.026) but negatively correlated with ALT and TBIL (r = -0.538, P = 0.018; and r = -0.458, P = 0.049, respectively) (Fig. 4D). IL-8 was positively correlated with AST (r = 0.474, P = 0.040) (Fig. 4D).