Predictive Power of Biomarkers of Oxidative Stress and ... - CiteSeerX

3 downloads 1897 Views 296KB Size Report
for predicting disease-free survival in patients with HCC and HCV. Conclusions: These .... formed using image analysis software (Scion Image;. Scion Corp.
Annals of Surgical Oncology 14(3):1182–1190

DOI: 10.1245/s10434-006-9049-1

Predictive Power of Biomarkers of Oxidative Stress and Inflammation in Patients with Hepatitis C Virus-Associated Hepatocellular Carcinoma Akira Maki,1,2 Hiroshi Kono,2 Mayetri Gupta,3 Masami Asakawa,2 Tetsuya Suzuki,2 Masanori Matsuda,2 Hideki Fujii,2 and Ivan Rusyn1

1

Department of Environmental Sciences and Engineering, School of Public Health, University of North Carolina at Chapel Hill, CB #7431, Chapel Hill, North Carolina 27599, USA 2 First Department of Surgery, Yamanashi University, Tamaho, Nakakoma, Yamanashi Prefecture, Japan 3 Department of Biostatistics, School of Public Health, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599, USA

Background: This study evaluated the relationship between inflammation, intra-hepatic oxidative stress, oxidative DNA damage and the progression of liver carcinogenesis in hepatitis C virus (HCV)-infected humans. Methods: Non-cancerous liver tissues were collected from 30 patients with an HCV-associated solitary hepatocellular carcinoma (HCC) who received curative tumor removal. After surgery, the patients were followed at monthly intervals at the outpatient clinic. Distribution of the inflammatory cells (CD68+), the number of 8-hydroxydeoxyguanosine (8-OHdG) DNA adducts and 4-hydroxynonenal (HNE) protein adducts and the expression of apurinic/ apyrimidinic endonuclease (APE) were determined by immunohistochemical analysis in serial liver sections from tumor-free parenchyma at the surgical margin around the tumor. Results: Significant positive correlations were observed between the number of CD68+ cells, the amount of HNE protein adducts, and the number of 8-OHdG adducts in liver tissue of patients with HCC and HCV. The cumulative disease-free survival was significantly shorter in patients with the highest percentage of 8-OHdG-positive hepatocytes. Using a Cox proportional hazard model, 8-OHdG, HNE and CD68 were determined to be good biomarkers for predicting disease-free survival in patients with HCC and HCV. Conclusions: These results support the hypothesis that HCV-induced inflammation causes oxidative DNA damage and promotes hepatocarcinogenesis which directly affects the clinical outcome. Since patients with greater intra-hepatic oxidative stress had a higher incidence of HCC recurrence, we suggest that oxidative stress biomarkers could potentially be used as a useful clinical diagnostic tool to predict the duration of disease-free survival in patients with HCV-associated HCC. Key Words: Hepatocellular carcinoma—Hepatitis C virus—Oxidative stress—Disease—free survival.

million deaths annually.1 While risk factors for HCC are well-established and include liver cirrhosis due to alcohol and chronic viral infection (hepatitis virus B and C), aflatoxin B1 exposure and a variety of metabolic liver diseases,2 the mechanisms that lead to this devastating disease are largely unknown. More than 90% of HCC in Japan is predominantly related to hepatitis C virus (HCV) infection.3 Epidemiological and experimental studies suggest that HCV-associ-

Hepatocellular carcinoma (HCC) is one of the worldÕs most common cancers, causing almost one

Received January 9, 2006; accepted April 26, 2006; published online December 31, 2006. Address correspondence and reprint requests to: Ivan Rusyn; E-mail: [email protected] Published by Springer Science+Business Media, Inc.  2006 The Society of Surgical Oncology, Inc.

1182

1183

OXIDATIVE STRESS BIOMARKERS IN HCC PROGNOSIS

ated carcinogenesis in the liver is dependent upon prolonged viral infection, resulting in cell necrosis, liver cell renewal and chronic inflammation.4 One strong mechanistic link between chronic inflammation and cancer is through the increased production of free radicals at the site of inflammation and the resulting molecular changes, which include lipid peroxidation and oxidative DNA damage.5 Indeed, markers of DNA damage, such as 8-hydroxydeoxyguanosine (8-OHdG), and lipid peroxidation, such as 4-hydroxynonenal (HNE) and malondialdehyde (MDA), are commonly elevated in liver of patients with chronic HCV infection and correlate well with the degree of viral infection and inflammation, known risk factors for HCC.6 Improved surveillance and imaging techniques have led to earlier detection of HCC and an increased opportunity to treat patients; however, the prognosis for patients with HCC remains to be poor.7 The recurrence of HCC is high, and it has been suggested that the degree of liver inflammation is associated to the length of patient disease-free survival.8 While the results from several studies have suggested the value of markers of inflammation in predicting the survival of patients with HCC, including those who are HCVpositive,9 little is known about the potential predictive power of oxidative stress markers. In this study, we tested the hypothesis that the degree of oxidative DNA damage, inflammation and lipid peroxidation correlates with the prognosis of disease-free survival after resection of HCC in patients with HCV infection. Patients enrolled in our study who had undergone a surgical removal of the liver tumor were followed up with regular visits to an outpatient clinic. The duration of time before a new liver tumor was diagnosed by tumor markers and/or imaging examination, such as ultrasonography (USG), computed tomography (CT) or magnetic resonance imaging (MRI), i.e. disease-free survival, was recorded. One of our goals was to determine whether any of the available routine clinical markers retrieved from a patientÕs records or any of the markers of oxidative stress and inflammation that were measured in this study correlated with the length of disease-free survival and thus potentially assist in clinical prognosis. By evaluating the relationship between inflammation, intra-hepatic oxidative stress, oxidative DNA damage and the progression of carcinogenesis in HCV-infected human liver we were able to show that patients with greater intra-hepatic oxidative stress exhibited a much faster recurrence of HCC. We therefore suggest that the measurement of oxidative stress biomarkers in tumor-free liver

parenchyma removed at the surgical margin around the tumor can be a potentially useful clinical diagnostic tool to predict the length of disease-free survival in patients with HCV-associated HCC. PATIENTS AND METHODS Patients Thirty patients with solitary HCC diagnosed as having no vascular invasion were enrolled in this study and underwent complete removal of the tumor at the University of Yamanashi Hospital (Yamanashi Prefecture, Japan) between July 1994 and December 2001. All of the patients tested positive for HCV and negative for HBV (data not shown). The presence and identification of the hepatitis virus was determined by one or more of the following techniques: (1) presence of anti-HCV and anti-HBV reactive serum proteins, (2) reverse transcription-PCR for serum HCV-RNA or (3) branched DNA-HCV probe assay. Following the liver resection patients returned to the ambulatory care clinic for additional tests monthly. Serum a-fetoprotein levels were measured every month. In addition, USG and CT of the liver were performed every 3 and 6 months, respectively. Six patients who were HCV/HBV-free but were diagnosed with metastatic liver tumors that required liver resection of the tumor mass were enrolled as the control group. Informed consent was obtained from all subjects who participated in the study, and the study was approved by the Institutional Board on Ethics for Human Science at the University of Yamanashi. All available clinical data for the patients enrolled in this study are summarized in Table 1. Tissues Sections of tumor-free liver parenchyma from the surgical margin removed during tumor resection were collected at the time of the operation. Tissues were fixed in formalin (10%), dehydrated in absolute ethanol and embedded in paraffin. Serial sections (5 lm thick) were prepared at the University of Yamanashi and sent to the University of North Carolina for further immunohistochemical analysis. Immunohistochemical Detection of CD68, 8-OHdG, HNE and Apurinic/Apyrimidinic Endonuclease 1 (APE) Paraffin-embedded serial sections of liver tissues were stained immunohistochemically with antiAnn. Surg. Oncol. Vol. 14, No. 3, 2007

1184

A. MAKI ET AL.

TABLE 1. Clinical characteristics of patients Hepatocellular carcinoma (HCC) (n = 30) Characteristics

HCC overall

CD68 Low

CD68 High

HNE Low

HNE High

8-OHdG Low

8-OHdG High

Age (years)a Gender Males Females Virus infection

57.8 ± 8.7 25 (83.3%) 5 (16.7%)

13 (86.7%) 2 (13.3%)

12 (80.0%) 13 (86.7%) 12 (80.0%) 3 (20.0%) 2 (13.3%) 3 (20.0%) HBsAg ()), HCVAb (+)

13 (86.7%) 2 (13.3%)

12 (80.0%) 3 (20.0%)

Tumor size (cm)a TMN classification Stage I Stage II Tumor differentiation Well Moderate Poor N/Ab White blood cells (per ml)a Alanine aminotransferase (ALT) (IU/l)a Indocyanine green (%)a Alpha-fetoprotein (AFP) (ng/ml)a Fibrosis score Score I/II Score III/IV

3.5 ± 2.3

2.8 ± 1.4

4.5 ± 2.7

2.9 ± 1.7

4.4 ± 2.6

4.3 ± 1.7

3.1 ± 2.5

11 (42.3%) 15 (57.7%)

6 (46.2%) 7 (53.8%)

5 (38.4%) 8 (61.6%)

6 (46.2%) 7 (53.8%)

5 (38.4%) 8 (61.6%)

4 (30.7%) 9 (69.3%)

7 (53.8%) 6 (46.2%)

Metastatic liver tumors (MLT) (n = 6) 69.8±8.9 3 (50.0%) 3 (50.0%) HBsAg()), HCVAb ()) 8.0 ± 3.7 N/Ab

4 (14.8%) 2 (14.3%) 2 (15.2%) 3 (21.4%) 1 (7.7%) 1 (7.7%) 3 (21.4%) N/Ab 19 (70.3%) 10 (71.5%) 9 (69.2%) 8 (57.1%) 11 (84.6%) 10 (66.7%) 9 (64.3%) 2 (7.4%) 1 (7.1%) 1 (7.6%) 1 (7.1%) 1 (7.7%) 1 (7.7%) 1 (7.1%) 2 (7.4%) 1 (7.1%) 1 (7.6%) 2 (14.2%) 0 1 (7.7%) 1 (7.1%) 4378 ± 1169 4457 ± 1267 4293 ± 1099 4457 ± 1302 4294 ± 1055 4272 ± 1093 4477 ± 1269 6922 ± 3356 57.5 ± 30.5 62.7 ± 35.1 52.3 ± 25.2 61.7 ± 35.9 53.3 ± 24.5 58.6 ± 26.2 56.4 ± 35.1 c

d

e

19.3± 10.2 15.0 ± 6.7 24.4 ±11.7 14.9 ± 6.9 4.6 ± 11.3 18.2 ± 6.8 20.4 ± 13.0 541 ± 1177 371 ± 1129 712 ± 1787 369 ± 1130 714 ± 1786 713 ± 1787 370 ± 1129 14 (56.0%) 11 (44.0%)

8 (66.7%) 4 (33.4%)

6 (50.0%) 6 (50.0%)

8 (66.7%) 4 (33.4%)

6 (46.2%) 7 (53.8%)

7 (53.8%) 6 (46.2%)

7 (63.6%) 4 (36.4%)

21.4 ± 7.6 7.4 ± 3.7 N/Ab N/Ab

a

The data shown are the mean ± standard deviation. N/A, not available. Significantly different from MLT group (P = .018). d Significantly different from ‘‘CD68 Low’’ group (P = .023). e Significantly different from ‘‘4-hydroxynonenal (HNE) Low’’ group (P = .016). b c

CD68 (PG-M1, DAKO, Kyoto, Japan), anti8OHdG (N45.1; Jaica, Shizuoka, Japan), anti-HNE (HNE11S; Alpha Diagnostics, San Antonio, Tex.) or anti-APE1/Ref-1 (NB100-116; Nubous Biologicals, Littleton, Colo.) antibodies. Immunohistochemical staining was performed on paraffin-embedded sections. Briefly, after deparaffinization and rehydration, the antigen retrieval procedure was applied (for CD68 and 8-OHdG staining) by heating the slides in .1 M citrate buffer (pH 6.0) for 10 min. The sections were then incubated first with .3% H2O2 in distilled water for 5 min to block endogenous peroxidase and then incubated with one of the following antibodies: monoclonal mouse antihuman CD68 for 30 min, HNE (diluted to 1:200) for 10 min, 8-OHdG (10 mg/ml) at 4C overnight or APE (diluted to 1:200) for 60 min. Biotinylated secondary antibody conjugated with avidin-biotinhorseradish peroxidase (Dako Envison kit/HRP; Dako, Kyoto, Japan) and 3¢,3¢-diaminobenzidine tetrahydrochloride were used for standard avidinbiotinylated peroxidase detection. Cells that stained Ann. Surg. Oncol. Vol. 14, No. 3, 2007

positive for CD68 and APE were counted in ten different random fields at a magnification of 400·. The hepatocytes that stained positively for 8-OHdG were counted (at least 1000 cells in each slide), and the percentage of positive cells was calculated. Quantitation of HNE-protein adducts was performed using image analysis software (Scion Image; Scion Corp., Frederick, M.D.) by evaluating five different fields (magnification: 100·) for areas positively stained for HNE-protein adducts and expressed as a percentage of total area. Statistics The STATVIEW software (ver. 5.0, SAS institute Cary, N.C) was used for statistical analysis. For the univariate analysis, two-sample t-test, linear regression or Logrank test statistical procedures were used to assess which endpoints could be used for predicting the prognosis of HCC patients after surgery. For the multivariate analysis, the Cox proportional hazard model was used to calculate hazard ratio and the

OXIDATIVE STRESS BIOMARKERS IN HCC PROGNOSIS

1185

P value of each parameter. A P value of less than .05 was considered to be significant.

RESULTS Markers of Oxidative Stress and Inflammation Accumulate in the liver of Patients with HCVAssociated HCC Hepatocellular carcinoma and viral hepatitis are known to produce a state of chronic oxidative stress in the liver that is thought to be associated with chronic inflammation. This study investigated whether mechanism-based molecular markers of oxidative stress and inflammatory response are useful endpoints that can be easily evaluated in specimens collected during liver tumor resection in (1) patients with solitary tumors in the liver that were a result of metastasis from other organs and (2) subjects manifesting HCC in combination with chronic HCV infection and the absence of known tumors at other sites (Fig. 1). With the exception of the indocyanine green (ICG) clearance test result, there was no significant difference in the clinical characteristics of these two groups of patients in this study when traditional clinical parameters were considered (Table 1). Consequently, additional experiments were performed using liver parenchymal tissue that appeared to be tumor-free but which was excised during tumor resection as a ‘‘surgical margin’’. We assumed that these samples of the margin tissue were an accurate representation of the remaining ‘‘normal’’ liver parenchyma in each patient. Immunohistochemical detection of 8-OHdG, a marker of oxidative DNA damage (Fig. 2A), HNE, a marker of lipid peroxidation (Fig. 2B) and CD68, a marker of inflammatory cells such as macrophages and neutrophils (Fig. 2C), was performed as described in the Patients and Methods section. Positive staining for CD68, HNE and 8-OHdG was detected in both the patients with HCC and patients with metastatic liver tumors (MLT); however, the number of positively stained cells or the area of the positive staining was consistently significantly higher (CD68: P < .001, HNE: P < .005, 8-OHdG: P < .01) in patients with HCC (n = 30) than in patients with metastatic liver tumors (n = 6). We next evaluated the relationship between markers of inflammation, intra-hepatic oxidative stress and oxidative DNA damage in patients with HCV-associated HCC (Fig. 2D, E). Interestingly, a strong positive correlation was observed between these three measures of

FIG. 1. Working hypothesis. We hypothesized that mechanismbased molecular markers of oxidative stress and inflammatory response would be useful for predicting disease-free survival after liver tumor resection in subjects manifesting hepatocellular carcinoma (HCC) in combination with chronic hepatitis C virus (HCV) infection and absence of known tumors at other sites.

the pathophysiological processes occurring in the liver of these patients. When patients with HCV-associated HCC were sub-divided into two groups (‘‘low’’ and ‘‘high’’) based on the median value of each of the three markers evaluated in this study (CD68, HNE and 8OHdG), 12 out of 15 patients were classified as ‘‘high’’ for all three markers and two were classified as ‘‘high’’ for two out of the three (data not shown). Furthermore, when clinical characteristics were compared between ‘‘low’’ and ‘‘high’’ patients for each marker, no significant differences were observed (Table 1), with the exception of the results of the ICG retention test (n = 24, P = .018). The latter was also significantly higher in patients who exhibited a greater inflammatory response (as determined by the number of CD68-positive cells in the liver; n = 11, P = .023) but lower in patients with high lipid peroxidation marker accumulation (HNE-positive staining; n = 13, P=.016). While several studies have linked chronic liver inflammation with an increase in organ dysfunction as measured by ICG clearance,10,11 the significance of the observed negative association between lipid peroxidation and organic anion transport in liver is uncertain. Markers of Oxidative Stress and Inflammation Improve Prediction for the Length of Disease-Free Survival in Patients with HCV-Associated HCC The univariate analysis of the prognostic power for various measurements taken in this study showed that none of the routinely collected indices of tumor progression were able to achieve statistical Ann. Surg. Oncol. Vol. 14, No. 3, 2007

1186

A. MAKI ET AL.

two groups (‘‘high’’ vs. ‘‘low’’) based on the median value for each of the markers (dashed and dotted lines, respectively), there was a dramatic significant difference (see Table 2 for group size and P values) in the clinical outcome between the groups. Importantly, the disease-free survival was significantly shorter in patients with strong staining for CD68, HNE and 8-OHdG, while there was no difference between these groups when traditional clinical parameters were compared (Table 1). Expression of DNA Repair Protein APE Is Elevated in HCV-Associated HCC but Does Not Correlate with Disease-Free Survival

FIG. 2. Immunohistochemical analysis for 8-hydroxydeoxyguanosine (8-OHdG), 4-hydroxynonenal (HNE) and CD68-positive cells in non-cancerous liver tissue from patients with HCV-associated HCC or metastatic liver tumors (MLT). Immunohistochemical staining for 8-OHdG (A), HNE (B) and CD68 (C) was performed in non-cancerous liver tissue collected from patients as described in detail in the Patients and Methods section. The boxes and error bars indicate the 25–75th percentile and the 10–90th percentile, respectively. Statistical significance is indicated with asterisks (*P < .01; **P < .005; and ***P < .001) as compared to the data in MLT patients. Panels D and E show correlations between markers assessed in this study in each of the patients with HCV-associated HCC.

significance in predicting the length of disease-free survival in patients with HCV-associated HCC (Table 2). Interestingly, all three markers of oxidative stress and inflammation were significant (see Table 2 for group size and P values) in this analysis with the number of 8-OHdG-positive hepatocytes in the tumor-free surrounding parenchyma collected at the time of tumor removal being the most important. Figure 3 shows that when disease-free survival was plotted using the Kaplan-Meier method for all patients (solid lines) or by sub-dividing them into Ann. Surg. Oncol. Vol. 14, No. 3, 2007

Chronic oxidative stress in liver not only increases DNA damage but also can lead to persistent upregulation of DNA repair genes.12 It has been hypothesized that such an adaptive response may in fact be contributing to carcinogenesis.13 Interestingly, in patients with ulcerative colitis (UC) the level of DNA repair protein APE has been found to be significantly higher in UC colon epithelium undergoing elevated inflammation and to be positively correlated with microsatellite instability.14 In the present study, immunohistochemical staining for APE was performed on non-cancerous liver tissues from patients with liver tumors, as described in the Patients and Methods section. Although positive staining for APE was detected in both groups of patients, the area of positive staining was significantly (n = 30, P = .037) greater in patients with HCV-associated HCC than in patients with metastatic liver tumors (Fig. 4A). APE protein level in livers of patients with HCC correlated with neither the degree of oxidative DNA damage (as measured by 8-OHdG-positive hepatocytes, Fig. 4B), lipid peroxidation or the number of inflammatory cells (data not shown); nor did it show statistically significant predicting power in the univariate analysis of disease-free survival (Table 2). To further determine if an imbalance between oxidative DNA damage (8-OHdG) and DNA repair (APE) exists in patients with HCV-associated HCC and has an adverse effect on disease-free survival, the patients were categorized into five groups using the regression tree-based classification method15 to maximize the difference between the groups (Fig. 5). While the resulting chi-square statistic has a value of 20.4 (4 df, with a P value = .000421), indicating that the groups are significantly different, no significant improvement in predictive power was observed when APE was included in the analysis (dashed and dashed-and-dotted lines).

1187

OXIDATIVE STRESS BIOMARKERS IN HCC PROGNOSIS

TABLE 2. Univariate analysis of prognostic markers associated with disease-free survival in patients with liver tumors Marker

Distribution

n

P value

Tumor size (cm) TMN classification (Stage I/II) Tumor differentiation (well/moderate/poor) Serum ALT (IU/l) Serum AFP (ng/ml) ICGR15 (%) CD68 (number of positive cells/400· field) HNE (% positive staining area) 8-Hydroxydeoxyguanosine (8-OHdG) (% positive hepatocytes/1000+ cells) Apurinic/apyrimidinic endonuclease (APE) (number of positive cells/400· field)

3.5 ± 2.3 11/15 4/19/2 57.5 ± 30.1 541 ± 1177 19.3 ± 10.2 57.47 ± 14.75 29.27 ± 8.06 58.1 ± 12.27

30 26 25 30 30 24 30 30 30

.3208 .9823 .8582 .2720 .8623 .6954 .0134 .0366 .0084

172.73 ± 103.69

30

.3982

Patients with HCV-associated HCC were divided into two groups based on the median of the distribution of each variable [total number of patients (n) in each group is shown]. Data shown are the mean ± SD for each variable (with the exception of TMN classification and tumor differentiation), or the number of patients in each group for tumor stage and differentiation. For each variable a cumulative survival rate for patients in each of the two groups was then calculated using the Kaplan-Meier method, and the P value for the difference between patients between groups was calculated using the Mantel-Cox logrank test. Bold highlights statistically significant data.

Multivariate Analysis of Disease-Free Survival Outcome in Patients with HCV-Associated HCC Shows the Utility of Additional Novel Molecular Markers of Oxidative Stress and Inflammation In general, the size and stage of the tumor, serum alanine aminotransferase (ALT) level and serum alpha-fetoprotein (AFP) concentration are thought to be useful in assessing the potential for a relapse of HCC after complete tumor resection. In our study, the univariate analysis failed to show that such clinical parameters alone can serve as significant predictors for the recurrence of the disease. To evaluate the effect of novel molecular markers of oxidative stress and inflammation on the predictive power of traditional clinical parameters, we included the percentage of 8OHdG-positive hepatocytes, size of the tumor, stage of the tumor, ALT and AFP all together in the multivariate analysis. The result from the Cox proportional hazard analysis showed that the percentage of 8-OHdG-positive hepatocytes, the size and stage of the tumor, serum ALT level and serum AFP concentration appear to be independent risk factors for the relapse of HCC (Table 3). Although the hazard ratio of the percentage of 8-OHdG-positive hepatocytes was 1.174 (95% C.I. 1.089–1.266), the association with the disease outcome was the strongest among all other endpoints (P < .0001), suggesting that the percentage of 8-OHdG-positive cells in non-cancerous liver parenchyma is a good predictor of a relapse of HCC. Since the CD68 and HNE markers were found to be co-linear with 8-OHdG, it is not surprising that they also were found to be strong predictors of disease-free survival in this patient cohort. When either of these two markers was included into the multi-

variate analysis in place of 8-OHdG, they were found to be the only significant predictors with HNE having a hazard ratio of 1.243 (95% C.I. 1.106–1.397, P < .0003) and CD68 with a hazard ratio of 1.072 (95% C.I. 1.027–1.119, P < .0015). DISCUSSION The hepatitis C virus is a significant cause of disease-associated morbidity and mortality worldwide.16 Chronic viral infection of the liver leads to hepatocellular necrosis, inflammation and liver regeneration, all of which are associated with the infiltration of immune cells that produce reactive oxygen and nitrogen species.17 In addition, HCV core protein overexpression in mice has been shown to contribute to hepatic oxidative stress through mitochondrial dysfunction.18,19 Oxygen free radical-induced DNA damage as well as the adaptive imbalance in DNA repair lead to the accumulation of missense mutations in cancer-related genes and microsatellite instability, and have been strongly implicated as primary causes of human cancers that are a result of chronic inflammation.5,15 The disease-free survival rate is a common clinically useful parameter that determines the outcome of the treatment of HCC. After tumor resection, which is the choice of treatment for HCC in non-cirrhotic patients, the tumor recurrence rate exceeds 70% at the 5-year follow-up.20–22 Many pathological features, such as tumor size, number of tumors in the liver, capsule state, cell differentiation, venous invasion and the extent of intra-hepatic spreading, are commonly used in clinical diagnosis as predictive risk Ann. Surg. Oncol. Vol. 14, No. 3, 2007

1188

A. MAKI ET AL.

FIG. 3. Molecular markers of oxidative stress and inflammation in non-cancerous liver tissue improve the prediction of disease-free survival in patients with HCV-associated HCC. Disease-free survival was plotted using Kaplan-Meier method for all patients with HCV-associated HCC (n = 30, solid lines), or after patients were divided into two groups (‘‘low’’ and ‘‘high’’; dotted and dashed lines, respectively; n = 15) according to the median value of measurements for 8-OHdG (A), HNE (B) and CD68 (C).

factors for HCC recurrence and health prognosis in patients. Overall, the most powerful predictors of recurrence are the presence of microvascular invasion and additional tumor sites.20–22 The recurrence of HCC is thought to occur primarily due to dissemination and the de novo formation of tumors.23 The latter is thought to be due to the same predisposition factors as those that lead to the formation of a primary tumor. Thus, an assessment of the state of the tumor-free liver parenchyma Ann. Surg. Oncol. Vol. 14, No. 3, 2007

FIG. 4. Expression of apurinic/apyrimidinic endonuclease 1 (APE) in non-cancerous liver tissue is higher in patients with HCV-associated HCC than in those with MLT. Immunohistochemical staining for APE was performed in non-cancerous liver tissue collected from patients as described in detail in the Patients and Methods section. A, The data are expressed as a number of APEpositive cells/400· field. The boxes and error bars indicate the 25– 75th percentile, and 10–90th percentile, respectively. Statistical significance is indicated with asterisk (*P = .0367) as compared to the data in MLT patients. Panel B shows the correlation between APE and 8-OHdG markers in each of the patients with HCVassociated HCC.

at the time of resection of the primary carcinoma may yield important clues regarding the relative importance of the factors that promote carcinogenesis in liver, such as oxidative stress to DNA. The present study examined a series of molecular markers of inflammation, oxidative stress and DNA damage in non-cancerous liver tissue from patients undergoing tumor resection with the objective of gaining an understanding of the extent of the molecular changes in remaining liver parenchyma and their potential value in predicting the outcome of the treatment by surgical resection. The molecular profiling of HCC and mechanismbased markers are expected to refine the assessment

OXIDATIVE STRESS BIOMARKERS IN HCC PROGNOSIS

FIG. 5. Concomitant evaluation of DNA repair and DNA damage does not improve the prognosis of disease-free survival in patients with HCV-associated HCC. Patients were categorized into five groups using the regression tree-based classification method as detailed in the Results.

TABLE 3. Multivariate analysis of prognostic markers associated with disease-free survival in patients with Hepatitis C virus (HCV)-associated HCC Marker 8-OHdG Size of tumor Stage of tumor ALT AFP

Hazard ratio 1.174 .714 3.723 .993 1.000

95% C.I.

P value

1.089–1.266 .531–.961 1.061–13.064 .974–1.012 .999–1.000

< .0001 .0261 .0401 .4481 .0252

of recurrence and disease outcome. Many gene expression profiling studies have been performed to date in attempts to correlate changes in mRNA levels with disease-free survival in patients with HCVassociated HCC.24–26; however, none of the proposed markers have yet to gain wide-spread acceptance or to become routine in clinical practice.27 Still, while it was suggested that marked inflammatory cell infiltration in the tumor could provide a better prognosis of disease-free survival,22 subsequent microarray studies have confirmed that an expression signature characteristic of activated lymphocytes infiltrating the cirrhotic liver and activated liver macrophages can be detected in tumors of patients with HCV.26 Our study shows that immunohistochemical detection of CD68-positive inflammatory cells in non-tumor parenchyma may be a useful predictor of disease-free survival. We suggest that the strong

1189

correlation between CD68-positive inflammatory cells and both markers of oxidative stress (HNE and 8-OHdG) as well as shorter disease-free survival observed in this study in patients with HCC supports the theory that CD68-positive inflammatory cells in the liver may be a key source for reactive intermediates that exacerbate liver injury and lead to cancer. Thus, chronic inflammation in the liver should be one of the prime targets for therapeutic intervention to prevent relapse of HCC. Based on these observations, we have initiated a prospective human study that is evaluating the effect of anti-inflammatory therapy on disease-free survival in patients with HCV-associated HCC who underwent surgical removal of liver tumors. While there are numerous reports of oxidative stress to DNA in hepatocellular carcinoma,28,29 little is known about the prognostic value of oxidative DNA damage. One recent study found that a high 8-OHdG labeling index in non-cancerous liver tissue may be a potentially useful prognostic marker for the recurrence of HCC, while a multivariate analysis revealed that tumor grade and stage were not found to be significant predictors.9 Our study confirms this previous report in a different patient cohort and shows that the 8-OHdG labeling index is indeed a very good prognostic marker. Furthermore, we show here that accumulation of lipid peroxidation-modified proteins (HNE staining) in the liver is also a good independent predictor of disease-free survival. These findings strengthen the usefulness of these novel predictors. It was reported recently that adaptive overexpression of base excision repair enzymes, the primary pathway that repairs oxidative DNA lesions, was observed in another tumor that depends on chronic inflammation, ulcerative colitis.14 This study, along with others,13,30 presents convincing evidence that an imbalanced induction of DNA repair results in mutations and contributes to the progression of carcinogenesis. In our study, the protein expression of APE, a critical enzyme in the base excision DNA repair pathway, was increased in livers of HCV-infected patients with HCC. However, the level of APE expression did not correlate with disease-free survival. Furthermore, we did not detect a statistical relationship with shortened disease-free survival after resection of the primary tumor even in patients with strong discordant expression of APE and oxidative stress, DNA damage and inflammatory cell infiltration. The reasons for this lack of correlation between repair, DNA damage and recurrence of liver Ann. Surg. Oncol. Vol. 14, No. 3, 2007

1190

A. MAKI ET AL.

tumors are not clear and may be a factor of the small size of the patient population in this study. In conclusion, the molecular markers selected in this study based on the pathogenesis of the disease strongly correlate with each other and are able to serve as independent strong predictors of disease-free survival in patients with HCV-associated HCC. Furthermore, we argue that a panel of mechanismbased molecular markers, rather than a single measure of oxidative stress and/or inflammatory response, should be used to increase precision and improve confidence in the post-operative prognosis. ACKNOWLEDGMENTS This research was supported, in part, by grants from the National Institute of Environmental Health Sciences (P30 ES10126, R01 ES12686, U19 ES11391, and K22 ES11660). REFERENCES 1. Inoue H, Seitz HK. Viruses and alcohol in the pathogenesis of primary hepatic carcinoma. Eur J Cancer Prev 2001; 10:107–110. 2. Kountouras J, Lygidakis NJ. New epidemiological data on liver oncogenesis. Hepatogastroenterology 2000; 47:855–861. 3. Higuchi M, Tanaka E, Kiyosawa K. Epidemiology and clinical aspects on hepatitis C. Jpn J Infect Dis 2002; 55:69–77. 4. Idilman R, De Maria N, Colantoni A, Van Thiel DH. Pathogenesis of hepatitis B and C-induced hepatocellular carcinoma. J Viral Hepat 1998; 5:285–299. 5. Hussain SP, Hofseth LJ, Harris CC. Radical causes of cancer. Nat Rev Cancer 2003; 3:276–285. 6. De Maria N, Colantoni A, Fagiuoli S, et al. Association between reactive oxygen species, disease activity in chronic hepatitis C. Free Radic Biol Med 1996; 21:291–295. 7. Cahill BA, Braccia D. Current treatment for hepatocellular carcinoma. Clin J Oncol Nurs 2004; 8:393–399. 8. Matsumoto K, Yoshimoto J, Sugo H, Kojima K, Futagawa S, Matsumoto T. Relationship between the histological degrees of hepatitis and the postoperative recurrence of hepatocellular carcinoma in patients with hepatitis C. Hepatol Res 2002; 23:196–201. 9. Matsumoto K, Satoh Y, Sugo H, et al. Immunohistochemical study of the relationship between 8-hydroxy-2Õ-deoxyguanosine levels in noncancerous region and postoperative recurrence of hepatocellular carcinoma in remnant liver. Hepatol Res 2003; 25:435–441. 10. Grune S, Michl M, Schinharl D, et al. Rapid effects of lipopolysaccharides on indocyanine green clearance in rat liver. Eur J Gastroenterol Hepatol 2000; 12:679–685. 11. Poeze M, Ramsay G, Buurman WA, Greve JW, Dentener M, Takala J. Increased hepatosplanchnic inflammation precedes the development of organ dysfunction after elective high-risk surgery. Shock 2002; 17:451–458.

Ann. Surg. Oncol. Vol. 14, No. 3, 2007

12. Rusyn I, Asakura S, Pachkowski B, et al. Expression of base excision DNA repair genes is a sensitive biomarker for in vivo detection of chemical-induced chronic oxidative stress: Identification of the molecular source of radicals responsible for DNA damage by peroxisome proliferators. Cancer Res 2004; 64:1050–1057. 13. Cairns J. The contribution of bacterial hypermutators to mutation in stationary phase. Genetics 2000; 156:923–926. 14. Hofseth LJ, Khan MA, Ambrose M, et al. The adaptive imbalance in base excision-repair enzymes generates microsatellite instability in chronic inflammation. J Clin Invest 2003; 112:1887–1894. 15. Hothorn T, Lausen B, Benner A, Radespiel-Troger M. Bagging survival trees. Stat Med 2004; 23:77–91. 16. Lai MM. Hepatitis C virus proteins: direct link to hepatic oxidative stress, steatosis, carcinogenesis and more. Gastroenterology 2002; 122:568–571. 17. Murakami T, Kim T, Nakamura H. Hepatitis, cirrhosis, and hepatoma. J Magn Reson Imaging 1998; 8:346–358. 18. Okuda M, Li K, Beard MR, et al. Mitochondrial injury, oxidative stress, and antioxidant gene expression are induced by hepatitis C virus core protein. Gastroenterology 2002; 122:366–375. 19. Korenaga M, Wang T, Li Y, et al. Hepatitis C Virus Core Protein Inhibits Mitochondrial Electron Transport and Increases Reactive Oxygen Species (ROS) Production. J Biol Chem 2005; 280:37481–37488. 20. Okada S, Shimada K, Yamamoto J, et al. Predictive factors for postoperative recurrence of hepatocellular carcinoma. Gastroenterology 1994; 106:1618–1624. 21. Minagawa M, Makuuchi M, Takayama T, Kokudo N. Selection criteria for repeat hepatectomy in patients with recurrent hepatocellular carcinoma. Ann Surg 2003; 238:703–710. 22. Adachi E, Maeda T, Matsumata T, et al. Risk factors for intrahepatic recurrence in human small hepatocellular carcinoma. Gastroenterology 1995; 108:768–775. 23. Chen YJ, Yeh SH, Chen JT, et al. Chromosomal changes and clonality relationship between primary and recurrent hepatocellular carcinoma. Gastroenterology 2000; 119:431– 440. 24. Iizuka N, Oka M, Yamada-Okabe H, et al. Oligonucleotide microarray for prediction of early intrahepatic recurrence of hepatocellular carcinoma after curative resection. Lancet 2003; 361:923–929. 25. Kurokawa Y, Matoba R, Takemasa I, et al. Molecular-based prediction of early recurrence in hepatocellular carcinoma. J Hepatol 2004; 41:284–291. 26. Smith MW, Yue ZN, Korth MJ, et al. Hepatitis C virus and liver disease: global transcriptional profiling and identification of potential markers. Hepatology 2003; 38:1458–1467. 27. Llovet JM, Wurmbach E. Gene expression profiles in hepatocellular carcinoma: not yet there. J Hepatol 2004; 41:336–339. 28. Schwarz KB, Kew M, Klein A, et al. Increased hepatic oxidative DNA damage in patients with hepatocellular carcinoma. Dig Dis Sci 2001; 46:2173–2178. 29. Wanibuchi H, Hori T, Meenakshi V, et al. Promotion of rat hepatocarcinogenesis by dimethylarsinic acid: association with elevated ornithine decarboxylase activity and formation of 8hydroxydeoxyguanosine in the liver. Jpn J Cancer Res 1997; 88:1149–1154. 30. Glassner BJ, Rasmussen LJ, Najarian MT, Posnick LM, Samson LD. Generation of a strong mutator phenotype in yeast by imbalanced base excision repair. Proc Natl Acad Sci USA 1998; 95:9997–10002.