SEARCH

SEARCH BY CITATION

Keywords:

  • hepatocellular carcinoma;
  • carcinogenesis;
  • chronic hepatitis c;
  • interleukin-6;
  • liver inflammation

Abstract

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Interleukin-6 (IL-6) may play a role in the pathogenesis of hepatocellular carcinoma (HCC). Recently, it was reported in mouse models that estrogen-mediated inhibition of IL-6 production explains the gender disparity in HCC. We conducted a retrospective cohort study to examine whether this hypothesis is applicable to human HCC. We enrolled 330 patients with chronic hepatitis C whose serum samples were collected between January 1994 and December 2002. Serum IL-6 concentrations were measured and patients were divided into three groups according to IL-6 levels: low, middle, and high. We evaluated the association between serum IL-6 levels and the risk of subsequent HCC development, including subgroup analysis on each gender. During the follow-up period (mean 9.0 yr), HCC developed in 126 patients. The incidence rates differed significantly among the three groups (p = 0.015), increasing in accordance with serum IL-6 levels. However, unexpectedly, this tendency was significant only in female patients. In a multivariate analysis, higher serum IL-6 level was an independent risk factor for HCC development in female patients, with a hazard ratio of 1.61. Although female patients showed a weak negative correlation between serum IL-6 levels and estradiol levels, the lower risk of HCC in female patients cannot be fully explained by estrogen-mediated inhibition of IL-6 production. In conclusion, higher serum IL-6 level was an independent risk factor for HCC development in female but not male chronic hepatitis C patients. Measurement of serum IL-6 levels may provide useful information for predicting future HCC development in female chronic hepatitis C patients. © 2009 UICC

Hepatocellular carcinoma (HCC) is the fifth most common cancer worldwide and chronic hepatitis C virus (HCV) infection is a major cause of HCC in the United States, Southern European countries and Japan.1 The host immune responses to HCV are often not strong enough to completely clear the infection, resulting in chronic stimulation of antigen-specific immune response.2 Hepatocyte damage is induced by the continued expression of cytokines and recruitment of activated inflammatory cells to the liver, which is followed by hepatocyte regeneration. This persistent cycle of necro-inflammation and hepatocyte regeneration is thought to provide a mitogenic and mutagenic environment leading to HCC development.3, 4

Interleukin-6 (IL-6) is a multifunctional cytokine largely responsible for the hepatic response to infections or systemic inflammation. Serum IL-6 levels are elevated in patients with chronic liver inflammation including alcoholic hepatitis, hepatitis B, HCV infections and steatohepatitis.5–7 Furthermore, serum IL-6 levels are reportedly higher in patients with HCC than in those without.8, 9 In chronic hepatitis, IL-6, produced mainly by activated Kupffer cells, intensifies local inflammatory responses and induces compensatory hepatocyte proliferation, facilitating malignant transformation of hepatocytes.10–12 Double transgenic mice that express both IL-6 and its soluble receptor under a liver-specific promoter develop hepatocellular hyperplasia and adenomas, mimicking precancerous lesions in humans.13 However, despite accumulating evidence suggesting that IL-6 plays an important role in hepatocarcinogenesis, it is not known how IL-6 contributes to HCC development in human chronic hepatitis.

HCC develops much more frequently in men than in women, with a male-to-female ratio of 2:1 to 4:1, which varies according to disease etiology and geographic region.1 This gender difference may be in part attributable to differences in exposure to lifestyle-related risk factors for HCC, such as alcohol consumption and cigarette smoking.14, 15 However, sex hormones and X chromosome-linked genetic factors may also be important, because males are more susceptible to HCC than females in experimental rat and mice models.16–18 In fact, male chronic hepatitis C patients have a higher risk for HCC than female patients even after adjusting for liver fibrosis and alcohol consumption.19, 20 Recently, Naugler et al.18 reported a molecular mechanism explaining the lower HCC susceptibility in females, using diethylnitrosamine (DEN)-induced murine HCC models. In their study, DEN administration caused a greater increase in serum IL-6 concentrations in males than in females. Genetic ablation of IL-6 abolished the gender differences in DEN-induced liver injury and HCC development. Furthermore, administration of estrogens to DEN-treated male mice decreased circulating IL-6 concentration and liver damage to the levels found in female mice. From these findings, they concluded that estrogen-mediated inhibition of IL-6 production by Kupffer cells explains the gender disparity in HCC development, focusing attention on the association between IL-6 and gender disparity in hepatocarcinogenesis.21–23

To examine whether this hypothesis is applicable to human HCC, we conducted this retrospective cohort study on the association between serum IL-6 levels and the risk of hepatocarcinogenesis in chronic hepatitis C patients, paying close attention to gender disparity in particular.

Material and methods

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Patients

Between January 1994 and December 2002, a total of 1,428 HCV RNA-positive patients, excluding those with (or with a history of) HCC, visited the liver clinic of the Department of Gastroenterology at the University of Tokyo Hospital. Among these patients, we enrolled patients whose serum samples were collected with informed consent. Exclusion criteria were positivity for hepatitis B surface antigen, ongoing interferon therapy at the time of serum collection, infections in addition to HCV, and malignancies and autoimmune disorders known to alter IL-6 levels. Patients who visited only for consultation purpose were also excluded. We examined the history of interferon therapy and responses during the follow-up period. Patients who achieved a sustained virologic response, defined by undetectable HCV-RNA at least 24 weeks after the end of therapy, were also excluded. Furthermore, we excluded patients who developed HCC within 1 year after serum collection to rule out the possibility of occult HCC affecting serum IL-6 levels.24 Thus, 330 patients were enrolled, and the association between serum IL-6 levels at entry and the subsequent incidence of HCC was analyzed. All blood tests were performed at the time of serum collection. Control serum samples were collected from 80 age- and gender-matched healthy subjects in whom liver diseases were ruled out, recruited from the Center for Multiphasic Health Testing and Services, Mitsui Memorial Hospital. The study protocol conformed to the ethical guidelines of the 1975 Helsinki Declaration and was approved by the institutional review board.

Patient follow-up and diagnosis of hepatocellular carcinoma

Each patient was screened for HCC with ultrasonography at or immediately after the first visit and those in whom HCC was detected were excluded from this study. Afterward, patients were followed up at the outpatient clinic with blood tests including tumor markers and ultrasonography every 3–6 months. Contrast-enhanced computed tomography was performed when tumors were detected as possible HCC in ultrasonography and/or the serum α-fetoprotein (AFP) level showed an abnormal increase. HCC was diagnosed by dynamic computed tomography, considering hyperattenuation in the arterial phase with washout in the late phase as a definite sign of HCC. When diagnosis of HCC was ambiguous, ultrasound-guided tumor biopsy was performed and a pathologic diagnosis was made based on the Edmondson and Steiner criteria. Time to HCC occurrence was defined as the interval between the date of serum collection and the diagnosis of HCC. Patients were censored at the time of death without HCC development, the last visit when lost to follow-up, or the end of the study period. The last observation in this study was January 31, 2009.

Assay for IL-6 and estradiol

Serum samples were frozen at −70°C until assayed. IL-6 levels were measured by a commercially available enzyme-linked immunosorbent assay (ELISA) kit (R&D Systems, Minneapolis, MN) according to the manufacturer's instructions. Estradiol levels were measured by electrochemiluminescence immunoassay in a commercial laboratory (SRL, Tokyo, Japan). Detection limits were 0.1 pg/ml for IL-6 and 10 pg/ml for estradiol. Each sample was examined in duplicate and the average value was used for data analysis.

Statistical analysis

Data were expressed as the median and interquartile range (25th–75th percentiles) unless otherwise indicated. The categorical variables were compared by chi-square tests, whereas continuous variables were compared to the unpaired Student's t test (parametric) or Mann-Whitney U-test (nonparametric). Trends in accordance with serum IL-6 levels were assessed with the exact trend test (proportion) or the Jonckeere-Terpstra test (continuous variables). A partial correlation coefficient was calculated to remove the influence of possible confounders. A p value of less than 0.05 on a two-tailed test was considered significant. Cumulative HCC incidence was estimated using the Kaplan-Meier method, and the differences between groups were assessed with the log-rank test. In the analysis of risk factors for hepatocarcinogenesis, we tested the following variables obtained at the time of entry in univariate and multivariate Cox proportional hazard regression analysis: age, body mass index (BMI), heavy alcohol drinking (alcohol intake >50 g/day), serum albumin concentration, total bilirubin concentration, alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels, prothrombin activity, platelet count, AFP concentration and serum IL-6 concentration. Data processing and analysis were performed using S-PLUS 2000 (MathSoft, Seattle, WA) and SAS Software version 9.1 (SAS Institute, Cary, NC).

Results

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Patient profiles

A total of 330 patients, 150 men and 180 women, with a median age of 60 years (interquartile range, 52–65) were included in the study. Detailed demographic data are shown in Table I. Baseline characteristics were largely comparable between male and female patients. However, the proportion of patients with heavy alcohol consumption was much higher in male (29.3%) than in female patients (2.2%). The difference in total bilirubin concentration (median [interquartile range], 0.7 [0.6–0.9] vs. 0.6 [0.5–0.8] mg/dl) was considered to be physiological.

Table 1. Baseline Characteristics
VariableAll (n = 330)Male (n = 150)Female (n = 180)p
  • a

    Expressed as median (25th–75th percentiles).

Age, yra60 (52–65)60 (51–66)60 (53–65)0.96
Platelet count, ×103/μla145 (10.5–18.7)145 (10.5–18.1)145 (10.4–19.3)0.44
Total bilirubin, mg/dla0.7 (0.5–0.9)0.7 (0.6–0.9)0.6 (0.5–0.8)0.003
Serum albumin level, g/dla4.0 (3.8–4.2)4.0 (3.7–4.2)4.0 (3.8–4.2)0.3
AST, IU/la53 (36–81)55 (42–80)52.0 (32.0–83.0)0.83
ALT, IU/la59 (33–95)65 (47–100)50.5 (30.0–92.5)0.11
AFP, ng/mla5.0 (3.0–11.8)6.0 (3.0–12.0)5.0 (3.0–10.5)0.22
Prothrombin time activity (%)a85.3 (74.3–98.2)85.5 (72.7–96.7)85.3 (74.4–99.2)0.57
Drinking >50 g/day, n (%)48 (14.5)44 (29.3)4 (2.2)<0.0001
BMI (kg/m2)a22.5 (20.4–24.6)22.8 (20.9–24.6)22.2 (20.2–24.7)0.4
Patients who received IFN, n (%)49 (14.8)21 (13.1)28 (15.5)0.81

Serum IL-6 levels

Chronic hepatitis C patients had significantly higher serum IL-6 levels than healthy controls (7.1 [2.1–42.5] vs. 0.2 [0.1–0.6] pg/ml, p < 0.0001) and the difference was similar in male and female patients (p = 0.83; Fig. 1). In healthy controls, men tended to have higher serum IL-6 levels than women, although the difference was not statistically significant (p = 0.055).

thumbnail image

Figure 1. Serum IL-6 levels in healthy subjects and chronic hepatitis C patients. The box plot indicates the median (horizontal line), interquartile range (the box itself), 10th and 90th percentiles (the whiskers) and outliers outside the 10th and 90th percentiles (gray dots). *p < 0.01.

Download figure to PowerPoint

Incidence of HCC stratified according to serum IL-6 levels

The mean follow-up period was 9.0 years, constituting the total observation of 2,981 person-years. During the follow-up period, 36 (10.9%) patients were lost to follow-up evaluation. By the end of the follow-up period, HCC developed in 126 patients (4.2% per 1 person-year). Patients were divided into three groups according to serum IL-6 levels: low (serum IL-6 concentration ≤5 pg/ml; 74 males and 72 females), middle (5–50 pg/ml; 42 males and 65 females) and high (>50 pg/ml; 34 males and 43 females). Cumulative incidence rates at 5 and 10 years were 19.9 and 29.7% (3.1% per 1 person-year) in the low IL-6 group; 27.1 and 39.7% (4.7% per 1 person-year) in the middle IL-6 group; and 29.9 and 45.9% (5.9% per 1 person-year) in the high IL-6 group, respectively (Fig. 2). The incidence rates differed significantly among the three groups (p = 0.015, log-rank test), increasing in accordance with serum IL-6 levels.

thumbnail image

Figure 2. Cumulative incidence of HCC development stratified according to serum IL-6 levels; low (≤5 pg/ml), middle (>5-50 pg/ml) and high (>50 pg/ml).

Download figure to PowerPoint

Incidence of HCC, gender-separated analysis

The mean follow-up period was 8.3 years for male patients and 9.6 years for female patients. A total of 19 (12.7%) male patients and 17 (9.4%) female patients were lost to follow-up (p = 0.38, Fisher's exact test). The cumulative incidence rates at 5 and 10 years in each gender were 32.7 and 45.2% (5.6% per 1 person-year) in male patients, and 17.8 and 29.7% (3.2% per 1 person-year) in female patients. The cumulative incidence rate of HCC was significantly higher in men than in women in the cohort (p = 0.0021, log-rank test). Gender disparity in hepatocarcinogenesis was investigated by further dividing patients of each gender according to serum IL-6 levels as described above. In male patients, cumulative incidence rates at 5 and 10 years were 28.4 and 41.0% (5.0% per 1 person-year) in the low IL-6 group, 40.5 and 50.4% (6.5% per 1 person-year) in the middle IL-6 group, and 29.4 and 47.5% (6.1% per 1 person-year) in the high IL-6 group, respectively (Fig. 3a). The difference among the groups was not significant (p = 0.54). In contrast, in female patients, cumulative incidence rates at 5 and 10 years were 11.1 and 18.2% (1.6% per 1 person-year) in the low IL-6 group, 16.9 and 32.7% (3.6% per 1 person-year) in the middle IL-6 group, and 30.2 and 44.2% in the high IL-6 group (5.7% per 1 person-year), respectively (Fig. 3b), revealing statistically significant difference (p = 0.0025), increasing in accordance with serum IL-6 levels.

thumbnail image

Figure 3. Cumulative incidence of HCC development stratified according to serum IL-6 levels; gender-separated analysis. (a) male patients and (b) female patients.

Download figure to PowerPoint

Risk analyses

We analyzed the risk factors for HCC development separately for each gender. In the univariate analyses, older age, lower serum albumin concentration, higher total bilirubin concentration, higher AST level, lower prothrombin time, lower platelet count and AFP levels greater than 20 ng/ml were significant risk factors for HCC in both male and female patients (Table II). Although heavy alcohol consumption was a significant risk factor only in male patients, higher BMI and higher serum IL-6 levels were significant risk factors only in female patients. Multivariate proportional hazard regression analyses revealed that older age, AFP levels greater than 20 ng/ml, lower prothrombin time and heavy alcohol consumption were independent risk factors in male patients, and older age, lower platelet count, AFP levels greater than 20 ng/ml, lower prothrombin time, higher BMI and higher serum IL-6 levels were independent risk factors in female patients (Table II). Thus, higher serum IL-6 level was significant only in female patients, with a hazard ratio of 1.61 per 1 log increase (95% confidence interval, 1.12–2.31, p = 0.01).

Table II. Risk Factors for HCC Development: Univariate and Multivariate Analysis
VariableUnivariate analysesMultivariate analyses
Hazard ratio (95% CI)pHazard ratio (95% CI)p
Male
 Age (per 1 year)1.106 (1.046–1.106)<0.00011.058 (1.023–1.094)0.0009
 Platelet count (per 103/μl)0.987 (0.982–0.992)<0.00010.961 (0.907–1.019)0.18
 Total bilirubin (per 1 mg/dl)1.92 (1.109–3.325)0.0190.886 (0.421–1.865)0.75
 Serum albumin (per 1 g/dl)0.228 (0.118–0.441)<0.00010.745 (0.365–1.521)0.41
 AST (per 1 IU/l)1.005 (1.001–1.008)0.00621.005 (0.992–1.018)0.43
 ALT (per 1 IU/l)1.002 (0.999–1.005)0.280.998 (0.988–1.009)0.72
 AFP >20 vs. ≤20 ng/ml4.484 (2.684–7.489)<0.00013.429 (1.953–6.022)<0.0001
 Prothrombin time activity (per 1%)0.942 (0.924–0.959)<0.00010.952 (0.931–0.974)<0.0001
 Drinking >50 vs. ≤50 g/day1.865 (1.152–3.022)0.01131.821 (1.062–3.121)0.02
 BMI (per 1 kg/m2)0.979 (0.912–1.05)0.550.993 (0.918–1.073)0.85
 Log IL-6 (per 1 log)1.2 (0.928–1.553)0.161.214 (0.901–1.636)0.2
Female
 Age (per 1 yr)1.12 (1.074–1.168)<0.00011.106 (1.052–1.162)<0.0001
 Platelet count (per 103/μl)0.977 (0.971–0.984)<0.00010.98 (0.98–0.99)0.003
 Total bilirubin (per 1 mg/dl)4.216 (2.163–8.218)<0.00010.435 (0.144–1.315)0.14
 Serum albumin (per 1 g/dl)0.158 (0.076–0.327)<0.00010.84 (0.321–2.199)0.72
 AST (per 1 IU/l)1.008 (1.004–1.012)0.00020.998 (0.986–1.011)0.76
 ALT (per 1 IU/l)1.004 (1.000–1.007)0.02590.999 (0.99–1.009)0.87
 AFP >20 vs. ≤20 ng/mL7.328 (4.301–12.485)<0.00013.741 (2.014–6.95)<0.0001
 Prothrombin time activity (per 1%)0.915 (0.896–0.935)<0.00010.958 (0.928–0.988)0.007
 Drinking >50 vs. ≤50 g/day0.482 (0.178–1.306)0.150.84 (0.321–2.199)0.72
 BMI (per 1 kg/m2)1.153 (1.083–1.227)<0.00011.106 (1.009–1.2132)0.03
 Log IL-6 (per 1 log)1.877 (1.354–2.601)0.00021.607 (1.116–2.314)0.01

Correlation between serum IL-6 levels and clinical parameters

We evaluated the correlation between serum IL-6 levels and other clinical factors to investigate the clinical relevance of serum IL-6 levels in chronic hepatitis C patients (Table III). In male patients, serum IL-6 levels were not significantly correlated with any clinical characteristics. In contrast, serum IL-6 levels were positively correlated with serum transaminase levels in female patients, suggesting ongoing inflammation, and negatively correlated with platelet count and prothrombin time, suggesting advanced fibrosis. Because inflammation and fibrosis are mutually related, each of them may act as a confounder to the other. Thus, we calculated the partial correlation coefficients between serum IL-6 and AST controlling for platelet counts, and that between IL-6 and platelet counts controlling for AST levels (Table IV). In female patients, serum AST levels were positively correlated with serum IL-6 levels (p = 0.01), whereas platelet counts were not (p = 0.67). No significant partial correlations were found in male patients. These findings suggest that serum IL-6 levels are closely associated with liver inflammation status in female chronic hepatitis C patients.

Table III. Correlations Between Serum IL-6 and Other Parameters
VariableSerum IL-6 level
Low (≤5 pg/mL)Middle (>5–50 pg/mL)High (>50 pg/mL)p
  • a

    Expressed as median (25th–75th percentiles).

Malen = 74n = 42n = 34 
 Age, yra59.5 (48–66)61 (52–67)60 (46–65)0.91
 Platelet count, ×103/μla152 (115–193)121 (94–166)151 (106–188)0.27
 Total bilirubin, mg/dla0.7 (0.6–0.9)0.7 (0.6–1.0)0.7 (0.6–0.9)0.87
 Serum albumin, g/dla4.0 (3.8–4.3)4.0 (3.7–4.2)3.9 (3.6–4.2)0.15
 AST, IU/la57.5 (42–80)54.0 (42–83)52.0 (41.0–65.0)0.57
 ALT, IU/la73 (39–108)59 (44–82)62.5 (49–88)0.22
 AFP, ng/mLa6.0 (3.0–15.0)5.0 (3.0–10.0)6.0 (4.0–10.0)0.27
 Prothrombin time activity (%)a85.5 (76.5–100)86.1 (71.4–96.7)84.6 (69.6–97.6)0.3
 Drinking >50 g/day, n (%)23 (31.1)10 (23.8)11 (32.3)0.95
 BMI (kg/m2)a22.8 (21.3–24.9)21.3 (20.2–24.3)23.5 (22.3–25.3)0.87
 Patients who received IFN, n (%)9 (12.2)5 (12.0)7 (20.1)0.3
Femalen = 72n = 65n = 43 
 Age, yearsa59 (51.5–64)61 (53–64)61 (56–66.7)0.13
 Platelet count, ×103/μla158 (113–202)148 (104–203)127 (90–172)0.05
 Total bilirubin, mg/dla0.6 (0.5–0.85)0.7 (0.5–0.8)0.7 (0.5–0.9)0.61
 Serum albumin, g/dla4.0 (3.8–4.2)4.0 (3.8–4.2)4.0 (3.7–4.1)0.22
 AST, IU/la43.5 (26.5–74.5)54.0 (33.5–81.2)65.0 (43.8–90.0)0.004
 ALT, IU/la41.5 (23.5–88.0)49.0 (29.5–93.5)64.0 (42.8–95.8)0.01
 AFP level, ng/mla4.8 (3.0–7.5)5.0 (3.0–11.3)6.0 (4.0–16.5)0.08
 Prothrombin time activity (%)a90 (78.0–100)87.9 (75.4–100)78.0 (72.5–92.3)0.02
 Drinking >50 g/day, n (%)3 (4.1)1 (1.5)0 (0)0.2
 BMI (kg/m2)a22.1 (19.9–25.0)22.4 (20.4–24.3)22.2 (20.4–25.9)0.73
 Patients who received IFN, n (%)12 (16.6)11 (16.9)5 (11.6)0.6
Table IV. Partial Correlation Coefficients Between Serum IL-6 Level and Ast Level or Platelet Count
Variableρp
Male
 Platelet count−0.130.11
 AST level−0.010.85
Female
 Platelet count−0.030.67
 AST level0.190.01

Influence of estrogen on serum IL-6 levels and HCC development in female patients

To investigate the influence of estrogen on serum IL-6 levels, we measured serum levels of estradiol, the major female reproductive hormone, in female patients. Serum estradiol levels were relatively low (12.5 [undetectable-20] pg/ml) in this study population, probably because the majority of female patients had already reached menopause. There was a trend of negative correlation between serum IL-6 levels and serum estradiol levels, although not statistically significant (Spearman's ρ = −0.135, p= 0.07). In the univariate analysis, lower serum estradiol level (≤20 pg/ml) was a significant risk factor for HCC development in female patients, with a hazard ratio of 2.505 (95% CI, 1.134–5.532, p = 0.023). However, serum estradiol level did not retain statistical significance in the multivariate analysis including serum IL-6 (serum estradiol ≤ 20 pg/ml; HR 0.815, 95% CI 0.316–2.102, p = 0.67, serum IL-6; HR 1.606, 95% CI 1.117–2.31 per 1 log increase, p = 0.01). Serum IL-6 level was a significant risk factor for HCC development (HR 1.635, 95% CI 1.158–2.307 per 1 log increase, p = 0.005) also in subgroup analysis among female patients with lower serum estradiol levels (≤20 pg/ml).

Discussion

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Accumulating evidence indicates that sustained inflammatory reaction is the major contributing factor to HCC development in chronic hepatitis C patients.3, 25 Various inflammatory cytokines, including tumor necrosis factor (TNF)-α, IL-1α, IL-1β, IL-6 and IL-8, have been implicated in chronic liver inflammation, among which IL-6 is thought to be one of the most important.2, 25, 26 In this study, we found that serum IL-6 level was significantly higher in chronic hepatitis C patients than in healthy controls, and that female, but not male, patients with higher serum IL-6 levels were at an increased risk of HCC development.

Although cause-and-effect relationships between IL-6 and hepatocarcinogenesis could not be evaluated in this study, IL-6 may enhance local inflammatory response and induce compensatory hepatocyte proliferation in chronic hepatitis patients.12 On IL-6-binding, the IL-6 receptor, which is highly expressed in hepatocytes, activates two signaling pathways, Janus activated kinase-signal transducer and activator of transcription, (JAK-STAT) and mitogen-activated protein kinase (MAPK), important in the regulation of cell survival and proliferation.27 The JAK-STAT pathway is enhanced in HCV-infected liver and HCC, and suppressors of this pathway such as SOCS3 are downregulated in HCC.28 Furthermore, DEN-induced hepatocarcinogenesis is reduced in IL-6 knockout mice,18 suggesting that IL-6 signaling plays an important role in hepatocarcinogenesis. Indeed, in the current study, the risk of HCC development was more strongly associated with serum IL-6 levels than with serum transaminase levels in female patients, suggesting a direct relationship between IL-6 and hepatocarcinogenesis in humans.

According to the hypothesis that estrogen-mediated inhibition of IL-6 production is the cause of gender disparity in hepatocarcinogenesis,18 serum IL-6 levels should be lower in female than in male hepatitis C patients. However, we found no significant difference in serum IL-6 levels between them. One possible explanation for this discrepancy is that the inhibitory effect of estrogen on IL-6 production may have been diminished in most female patients because the female cohort was mainly post-menopausal with low estrasdiol to begin with. About 80% of them were over 50 years, which is the mean age of menopause in Japanese women.29 Perhaps, when estradiol drops below a certain threshold, it is no longer the dominant factor affecting IL-6 expression. Nonetheless, significant difference in HCC incidence rates between male and female patients remains to be found in this study. This implies that the gender disparity in hepatocarcinogenesis in human cannot be attributed solely to the difference in IL-6 levels. The mechanism of gender disparity remains to be further investigated.

In female patients, serum IL-6 levels showed only weak negative correlation with serum estradiol levels. Serum IL-6 level is not strongly dependent on serum estradiol level in female patients with chronic hepatitis C, although we do not know whether this can be applicable to younger female patients. In multivariate analysis, not serum estradiol level but serum IL-6 level retained significance as an independent risk factor for HCC development. Additionally, even in the subgroup analysis limited to the patients with lower serum estradiol levels, serum IL-6 level was a significant risk factor for HCC development. These results indicate that serum IL-6 level in female patients with chronic hepatitis C represents certain pathological process other than serum estradiol level that may be directly engaged in hepatocarcinogenesis.

In this study, serum IL-6 level was not correlated with serum AST level nor associated with the risk of HCC development in male patients. The reason for this is not clear but androgenic hormones and other male-specific factors may be more important in hepatocarcinogenesis in men. In an animal model, DEN-induced hepatocarcinogenesis is significantly reduced with liver-specific androgen receptor knockout.30 Furthermore, androgens reportedly downregulate IL-6 production in human macrophages,31 making the interpretation of serum IL-6 levels more complicated. IL-6 production is less affected by sex hormones in postmenopausal women, and serum IL-6 levels may reflect liver inflammation status more accurately in female patients. However, even in males there was a positive correlation, albeit short of statistical significance, between serum IL-6 and HCC incidence. If the sample size is much larger, this correlation may be of statistical significance.

Recently, an association between serum IL-6 level and HCC development in chronic hepatitis B patients was reported from a nested case–control study of 37 HCC patients and 37 controls.32 Serum IL-6 level was correlated with serum transaminase level and associated with HCC development. However, in contrast to the present study, these findings were found in a male-dominant (89%) study population, and it is unclear whether this is due to differences in the etiology of hepatitis.

In summary, this study has shown that a higher serum IL-6 level is an independent risk factor for HCC development in female chronic hepatitis C patients. This finding supports the important role of IL-6 signaling in hepatocarcinogenesis. Although female patients showed a weak negative correlation between serum IL-6 levels and estradiol levels, the lower risk of HCC in female patients cannot be fully explained by estrogen-mediated inhibition of IL-6 production. The measurement of serum IL-6 levels may provide useful information on the risk of future HCC development in female hepatitis C patients.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

S.M. and M.O. were supported by Grants-in-Aid from the Ministry of Education, Culture, Sports, Science, and Technology, Japan (#00130036).

References

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References