SEARCH

SEARCH BY CITATION

Keywords:

  • macrophage migration inhibitory factor;
  • hepatocellular carcinoma;
  • plasma;
  • diagnosis;
  • prognosis

Abstract

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

We performed our study to determine whether plasma macrophage migration inhibitory factor (MIF) levels have diagnostic and prognostic value in hepatocellular carcinoma (HCC) patients. Enzyme-linked immunosorbent assay (ELISA) and immunohistochemistry were used to measure the expression of MIF in plasma and tissues, respectively. Plasma MIF levels were compared to HCC occurrence, clinicopathological features and outcomes. Cutpoints of plasma MIF levels for diagnosis and prognosis were, respectively, determined by receiver operating characteristic analysis and X-tile in corresponding training cohort, and then were confirmed in the validation cohort. The postoperative plasma MIF levels of HCC patients were detected in an independent cohort (80 HCC patients). As a result, MIF expression in situ was mainly observed in the cytoplasm of HCC cells. Intratumoral MIF expression was positively correlated with plasma MIF levels (r = 0.759, p < 0.001). Compared to serum α-fetoprotein (AFP), plasma MIF had a higher diagnostic value for discrimination of HCC from controls at 35.3 ng/ml. With determined cutpoints, plasma MIF levels demonstrated a significant association with overall survival (OS) and disease-free survival (DFS) of HCC patients even in patients with normal serum AFP levels and Tumor Node Metastasis (TNM) stage I. In addition, the plasma MIF levels were identified as an independent factor for OS [hazard ratio (HR) = 1.754; p = 0.012] and DFS (HR = 2.121; p < 0.001). Plasma MIF levels decreased markedly within 30 days after tumor resection (p < 0.001). Therefore, plasma MIF levels have potential as a diagnostic and prognostic factor for HCC.

Hepatocellular carcinoma (HCC) is the second and the third most common cause of cancer-associated death in China and worldwide, respectively.1, 2 Although surgical resection and liver transplantation provide valid approaches to treat HCC, the 5-year recurrence rate after curative resection remains high, up to 54.1–61.5%, of which 62.4–77.8% occurs within 2 years. This ultimately results in poor overall survival (OS).3 Presently, although several HCC-related biomarkers have been identified, no evidence has been obtained indicating that these markers accurately correlate with the prognosis of HCC patients. Therefore, the search continues for HCC-secreted proteins that can serve as biomarkers for predicting the prognosis of HCC.

Macrophage migration inhibitory factor (MIF) was originally identified as a lymphokine derived from activated T cells.4, 5 Recently, the roles of MIF in tumorigenesis, proliferation of tumor cells and tumor angiogenesis have attracted much more attention.6, 7 Bucala and Donnelly8 presumed that MIF was the key cytokine linking inflammation and cancer, because MIF promoted cell proliferation and oncogenesis by sustaining the activity of extracellular signal-regulated kinase (ERK) and ERK2 mitogen-activated protein (MAP) kinases and inhibiting P53-dependent apoptosis. Meyer-Siegler and Hudson9 reported that MIF mRNA levels were higher in prostatic adenocarcinoma than those in normal prostatic tissue. Ren et al.10 found that MIF mRNA was strongly expressed in HCC tissue, while weakly or not expressed in normal liver tissue. Taken together, these results show that MIF may act as an upstream regulator of growth factor-dependent tumor progression and angiogenesis.

Although a number of studies revealed that MIF is highly expressed in various tumor tissues, there are few studies on MIF in the peripheral blood. Therefore, we designed a retrospective study, which stringently followed the reporting recommendations for tumor marker prognostic studies established in 2005.11 The goal of our study was to evaluate the value of plasma MIF levels in the diagnosis and prognosis of HCC patients. To the best of our knowledge, this is the first study, which describes that elevated plasma MIF levels are associated with HCC development, tumor recurrence and survival of HCC patients after curative resection.

Material and Methods

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Patients

HCC patients (732) underwent surgeries between January 2004 and December 2004 in the Liver Cancer Institute, Zhongshan Hospital, Fudan University. Available preoperative plasma samples were accessed from 581 patients.

Diagnosis cohorts

For diagnosis analysis, 344 consecutive HCC patients and 390 controls [130 patients with hepatitis B-related liver cirrhosis, five patients with hepatitis C-related liver cirrhosis, 106 patients with liver benign lesions (focal nodular hyperplasia, cyst and hemangioma) and 149 healthy individuals] between January 2004 and May 2004 were included in the training cohort. In addition, 237 consecutive HCC patients and 257 controls (82 patients with hepatitis B-related liver cirrhosis, four patients with hepatitis C-related liver cirrhosis, 79 patients with liver benign lesions and 92 healthy individuals) between August 2004 and December 2004 were included in the validation cohort. The HCC patients and controls were well matched according to gender and age. The patients who were complicated by other primary malignancy and inflammatory disease were excluded from diagnosis analysis. All liver cirrhosis patients were initially diagnosed in the Department of Gastroenterology, Zhongshan Hospital, Fudan University. All healthy individuals underwent physical examination at the Zhongshan Hospital, Fudan University. To observe the change of plasma MIF levels after surgery, we enlisted 80 independent HCC patients, who recently underwent curative resection for HCC.

Prognosis cohorts

For prognosis analysis, 269 consecutive HCC patients who underwent curative resection between January 2004 and May 2004 were included in the training cohort. In addition, 173 consecutive HCC patients who underwent curative resection between August 2004 and December 2004 were included in the validation cohort. Patients who underwent palliative surgeries or who received prior intervention (such as trans-hepatic artery embolization, chemotherapy or radiotherapy) or who died in postoperative 2 months or who were complicated by other primary malignancies and inflammatory diseases during the course of our study were excluded from prognosis analysis. Tumor cell differentiation grade followed the Edmonson classification system. Tumor clinical staging followed the sixth edition Tumor Node Metastasis (TNM) classification standard revised by the International Union Against Cancer. Curative resection was defined as: (i) the complete resection of all tumor nodules and the cut surface being free from cancer on histological examination; (ii) no cancerous thrombus found in the portal vein (main trunk or two major branches), hepatic veins, or bile duct; (iii) no extrahepatic metastasis and (iv) negative serology and imaging studies at 2 months after operation.12, 13

Patients in the training and the validation cohorts were monitored after surgery until May 2010. The median follow-up time of training cohort and validation cohort were 47.8 months (range, 2.3–74.0 months) and 38.6 months (range, 2.7–67.3 months), respectively. Follow-up procedures were described in our previous study.12 During follow-up, serum α-fetoprotein (AFP) levels and ultrasound examinations were performed every 2–3 months after surgery, and computed tomography was performed if necessary. Patients with confirmed recurrence received further treatment. Second liver resection, radiofrequency ablation or percutaneous ethanol injection were suggested for local recurrence, while transcatheter arterial chemoembolization was suggested for multiple or diffused recurrence.

Samples

Preoperative plasma samples were collected from patients with HCC and benign lesions in the Liver Cancer Institute, Zhongshan Hospital, Fudan University. Plasma samples of liver cirrhosis were obtained at the first blood collection after hospitalization. Plasma samples of healthy individuals were obtained from those who underwent physical examination at the Zhongshan Hospital, Fudan University. To extensively investigate the correlation of plasma MIF levels of HCC patients with MIF expression in tumor tissue, we collected the preoperative plasma samples from a peripheral vein before intravenous injection in the operating room and corresponding tissue samples from the training cohort for prognosis analysis. The median time difference between plasma and tissue collection was 3.2 hr (range, 2.0–4.2 hr). In addition, corresponding pre- and postoperative (days 3, 7 and 30) plasma samples were also collected from 80 HCC patients in an independent cohort (totally 320 plasma samples). None of these patients showed evidence of recurrence at the time of the last blood collection. No postoperative complication (e.g., infection, bile leak or peritonitis) was found in all recruited patients. The study was approved by the Zhongshan Hospital Research Ethics Committee. All patients were recruited into our study after informed consent.

Enzyme-linked immunosorbent assay for MIF levels in plasma samples

Peripheral blood samples were collected, anticoagulated by ethylene diamine tetraacetie acid (EDTA) and then centrifuged at 4°C for 15 min (3000 rpm). The plasma was removed, aliquoted, and snap frozen at −70°C until used. MIF levels in plasma were measured by quantitative sandwich enzyme-linked immunosorbent assay (ELISA) kits (Quantikine, R&D Systems, Minneapolis, MN) according to the manufacturer's protocols as previously described.14 A subset of samples was reassayed five times in every ELISA plate for quality control.

Immunohistochemical (IHC) staining

HCC tissue and peritumoral liver tissue samples within 1 cm of the tumor were fixed in 10% formalin and subsequently embedded in the paraffin. Serial 4-μm-thick sections were prepared from paraffin blocks and stained with hematoxylin and eosin. To eliminate the nonuniformity due to sequential staining, the slide staining was processed with Dako Autostainer Plus™ IHC staining system (Dako, Carpinteria, CA). Primary antibody was mouse antihuman monoclonal antibody against MIF (Abcam, Cambridge, UK; catalog numbers: ad55445; 1:500). Negative controls were treated identically but with the primary antibodies omitted. The density of positive staining was measured by computerized image system as described previously.15 Briefly, for the reading of antibody staining, a uniform setting was applied for all slides. Integrated optical density of all positive staining of MIF in each photograph was measured, and its ratio to total area of each photograph was calculated as the density of MIF by Image-Pro Plus v6.2 software (Media Cybernetics, Bethesda, MD).15

Statistical analyses

Statistical analyses were performed using the SPSS 13.0 software (SPSS, Chicago, IL). Continuous variables were expressed as median (interquartile range). Differences between two independent samples were tested with the Mann–Whitney U test (nonparametric). Correlation between two continuous variables was determined by the Spearman correlation coefficient (nonparametric). Differences between pre- and postoperative plasma MIF levels were tested with the Friedman M test (nonparametric).

Diagnosis analysis

In the training cohort, optimal cutpoint of plasma MIF for diagnosis was calculated with receiver operating characteristic (ROC) analysis, according to the highest area under curve (AUC) for discriminating HCC patients from controls (liver cirrhosis, benign lesions and healthy individuals). The concentration with the highest AUC evaluated by Youden's index (sensitivity + specificity − 1) was the optimal cutpoint to diagnosis. Then, the optimal cutpoint of plasma MIF was confirmed in the validation cohort for diagnosis. In our study, to compare with plasma MIF, we use the widely accepted serum AFP cutoff (20 μg/l) instead of finding an optimal cutpoint.

Prognosis analysis

To estimate the prognostic value of plasma MIF levels, we did not simply choose median or quartile as cutpoints due to the fact that they do not optimize for biological variability. Instead, we determined cutpoints in HCC patients from the training cohort using a statistical method called X-tile, Version 3.6.1.16–18 X-tile divided the training cohort for prognosis analysis into two subpopulations (low and high) by every cutpoint of plasma MIF levels. The association of each division with patient outcome was calculated by the standard log-rank test. The optimal cutpoints for prognosis were defined as the plasma MIF level with the highest log rank χ2 value. Then, these cutpoints were confirmed in corresponding validation cohort. The cumulative OS and disease-free survival (DFS) was computed by the Kaplan–Meier method and compared by the log-rank test. The relative prognostic importance of plasma MIF levels and other clinicopathological variables on OS and DFS was studied by multivariate analyses using a Cox regression model. Factors showing significance by univariate analysis were adopted when multivariate Cox proportional hazards analysis was performed as previously described.15

Statistical significance was taken as p < 0.05 (two tailed).

Results

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Expression of MIF in HCC tissue

MIF staining was mainly observed in intratumoral tissue, showing a predominant cytoplasmic staining in HCC cells (Figs. 1a and 1b). Peritumoral tissue with cirrhosis also showed weak positive MIF staining in hepatocyte cytoplasm (Fig. 1c). However, peritumoral tissue without cirrhosis showed weak or negative MIF staining (Fig. 1d). Intratumoral expression of MIF was significantly higher than that of peritumoral expression (median density, 0.083 vs. 0.007; p < 0.001) even in patients with liver cirrhosis (median density, 0.093 vs. 0.013; p < 0.001). In addition, intratumoral MIF expression was significantly correlated with large tumor size (p < 0.001), venous invasion (p = 0.014) and advanced disease stage (p = 0.001) (Supporting Information Table 1). Furthermore, we found that plasma MIF levels in HCC patients were positively correlated with corresponding intratumoral MIF expression [r = 0.759 (95% confidence interval, 0.701–0.806), p < 0.001; Fig. 2a].

thumbnail image

Figure 1. Representative intratumoral [(a) and (b)] and peritumoral [(c) and (d)] immunohistochemical staining of MIF. Patient 18 [(a) and (c)] showed a strong positive staining in intratumoral tissue and a weak positive staining in peritumoral tissue with liver cirrhosis. Patient 57 [(b) and (d)] showed an intermediate positive staining in intratumoral tissue and a negative staining in peritumoral tissue without cirrhosis. Magnification: 200×. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Download figure to PowerPoint

thumbnail image

Figure 2. Correlation of plasma MIF levels with intratumoral MIF expressions in HCC tissue was shown in (a) determined by the Spearman correlation coefficient. Box and whisker plots (b) showed difference of plasma MIF level in patients with HCC, liver cirrhosis, benign lesion and healthy individuals calculated by Mann–Whitney U test. There is no difference of plasma MIF level between patients with benign lesions and healthy individuals. Box, the range of the middle 50% of plasma MIF level; line inside box, median; cross inside box, mean; whiskers, 5th and 95th percentile; *p < 0.001 compared to HCC; **p < 0.001 compared to liver cirrhosis.

Download figure to PowerPoint

Diagnosis analysis

In the training cohort, plasma MIF levels were high in HCC patients (median, 85.7 ng/ml; interquartile range, 55.9–114.7 ng/ml), intermediate in liver cirrhosis patients (median, 24.9 ng/ml; interquartile range, 18.5–31.1 ng/ml) and low in benign lesion patients (median, 12.5 ng/ml; interquartile range, 9.5–16.8 ng/ml) or healthy individuals (median, 13.2 ng/ml; interquartile range, 9.6–17.3 ng/ml; Fig. 2b). By ROC analysis, we got the optimal cutpoint at 35.3 ng/ml of plasma MIF level which had the highest AUC (0.939; 95% confidence interval, 0.918–0.960) to discriminate HCC patients from controls (Supporting Information Fig. 1).

In the validation cohort, we verify the robustness of the diagnostic role of plasma MIF level with the cutpoint of 35.3 ng/ml. All HCC patients from controls (liver cirrhosis, benign lesions and healthy individuals), early stage HCC from controls, HCC from liver cirrhosis and small HCC from cirrhosis could be discriminated with similar positive predictive values and much higher negative predictive values compared to serum AFP (Table 1 and Supporting Information Table 2).

Table 1. The sensitivity, specificity, positive predictive value and negative predictive value of plasma MIF and serum AFP for diagnosing HCC in the validation cohort
inline image

Prognosis analysis

The demographic or clinicopathological characters of HCC patients in the training and validation cohorts are listed in Table 2. There was no significant difference between the two cohorts. We investigated the correlation of plasma MIF levels with clinicopathological parameters of HCC patients in the training cohort. We found that there was no significant correlation between plasma MIF levels and patients' gender, age, HBsAg status, hepatitis type B virus (HBV)-DNA titer, serum alanine aminotransferase (ALT) level, liver cirrhosis, Child-Pugh score, preoperative serum AFP levels, tumor number, tumor size, tumor capsule or Edmonson grade. In contrast, high plasma MIF levels were significantly associated with the presence of venous invasion and an advanced disease stage (Supporting Information Table 1). These results indicate that plasma MIF levels significantly correlated with the aggressive behavior of HCC.

Table 2. The basic information of HCC patients in two cohorts for prognosis analysis
inline image

Optimal plasma MIF concentration cutpoints were obtained at 90.8 ng/ml for OS and at 92.3 ng/ml for DFS. HCC patients with high plasma MIF levels had significantly worse prognosis than those with low plasma MIF levels (Figs. 3a and 3b).

thumbnail image

Figure 3. With the determined cutpoints, prognostic significance of plasma MIF level for overall survival and disease-free survival were assessed by Kaplan–Meier and log-rank tests in the training cohort [(a) and (b)] and confirmed in the validation cohort [(c) and (d)]. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Download figure to PowerPoint

With these cutpoints, we drew survival curves of HCC patients in the validation cohort and found that HCC patients with high plasma MIF levels had significantly worse prognosis than those with low plasma MIF levels. The 1-year, 3-year and 5-year OS rates of patients with high plasma MIF levels were 84.1%, 59.4% and 39.1%, respectively, which were much lower than those of patients with low plasma MIF levels (93.3%, 79.8% and 62.4%, respectively) [hazard ratio (HR), 1.843; p = 0.004] (Fig. 3c). Furthermore, the 1-year, 3-year and 5-year DFS rates of patients with high plasma MIF levels (65.4%, 37.9% and 26.2%, respectively) were significantly lower than those of patients with low MIF levels (86.3%, 68.7% and 55.2%, respectively) (HR, 2.385; p < 0.001) (Fig. 3d). Stratified analysis according to serum AFP levels and TNM stages were also conducted to confirm the prognostic value of plasma MIF levels (Supporting Information Fig. 2). As a result, even in patients with normal serum AFP levels and TNM stage I, plasma MIF levels still could predict prognosis of HCC after curative resection. Univariate analyses revealed that plasma MIF levels were an independent prognostic factor for OS (HR, 1.754; p = 0.012) and DFS (HR, 2.121; p < 0.001) (Table 3). The plasma MIF had the highest predictive value for death (1-year OS) and recurrence (1-year DFS) compared to other factors (Supporting Information Table 3).

Table 3. Plasma MIF level is an independent prognostic factor for HCC patients after curative resection
inline image

Changes of plasma MIF levels after resection of tumor

To be a useful tumor marker, sensitive postoperative change, which represents the remission or recurrence of the disease, must be paralleled by specific changes in plasma MIF levels. In our study, we measured pre- and postoperative (days 3, 7 and 30) plasma MIF levels in an independent cohort of 80 HCC patients. We considered the plasma MIF levels of controls (patients with liver cirrhosis, benign lesions and healthy individuals) in the training cohort as the normal range (median, 15.5 ng/ml; range, 4.7–36.3 ng/ml). As shown in Supporting Information Figure 3, all HCC patients underwent a time-dependent decline of plasma MIF levels after tumor resections. Furthermore, all plasma MIF levels at postoperative day 30 (median, 18.7 ng/ml; range, 6.0–33.8 ng/ml) fell into the normal range and below the diagnostic cutpoint (35.3 ng/ml).

Discussion

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Previous studies about MIF have focused on inflammatory disease.19, 20 Accumulating evidences indicate that MIF might be the key cytokine linking inflammation and cancer.8 A malignant cycle of inflammation and renovation promotes oncogenesis. With regard to HCC, most lesions develop in a milieu of chronic inflammation produced by hepatitis B or C viruses as well as dysregulated cellular proliferation due to cirrhosis. In our study, we found a stepwise increase of plasma MIF levels in patients with benign lesions or healthy individuals, hepatitis-related liver cirrhosis and HCC. In agreement with our study, He et al.21 found that expression of MIF in both tissues and circulation was low in normal mucosa but substantially higher in gastritis, intestinal metaplasia and gastric cancer. Our results imply that circulative MIF levels are associated with the development of HCC and may be a potential diagnostic marker for HCC patients.

Our study showed that, with the cutpoint 35.3 ng/ml, plasma MIF could well diagnose HCC, even early stage and small HCC. HCC is difficult to diagnose by measuring serum AFP levels because 30–40% of all HCC patients were serum AFP negative (≤20 μg/l). So far, the AFP assay is not sensitive or specific enough to diagnose HCC. A more sensitive and specific biomarker should be pursued to improve the diagnosis of HCC. Our study shows that plasma MIF had higher diagnostic values, even in early stage HCC and small HCC, than serum AFP to discriminate HCC.

Expression of MIF in HCC tissue was found to be reversely proportional to prognosis of HCC patients.22 However, the correlation of MIF levels in peripheral blood with the prognosis of HCC was never investigated. Our study revealed that preoperative plasma MIF levels are related to venous invasion and advanced disease stage. We further investigated the association of plasma MIF levels with the prognosis of HCC and found that both OS and DFS of HCC patients with low MIF levels were better than that of the patients with high MIF levels, even when stratified by serum AFP levels and TNM stage.

Ren et al.23 showed that the serum MIF levels in patients with esophageal squamous cell carcinoma (ESCC) were not related to survival and explained that MIF acts as a local autocrine factor rather than a systemic mediator in ESCC, indicating that serum MIF alone was not a useful marker in clinical practice. To our knowledge, MIF is more likely an endocrine factor rather than a local autocrine factor.24 We also found a strong positive correlation between plasma MIF levels and intratumoral levels of MIF, mainly expressed by HCC cells. In agreement with our findings, He et al.21 showed a positive correlation between peripheral blood MIF levels in gastric cancer patients and MIF expression in gastric cancer tissue. Therefore, we postulate that MIF in peripheral blood is mainly produced by HCC cells and released into blood, but not just in an “over-spilled” phenomenon as explained by Ren et al.23 As opposed to the study of Ren et al., we measured MIF levels in plasma rather than serum. Compared to plasma, serum contains massive cytokines, perhaps including MIF, which are released from platelets when coagulation happens.25–27 Therefore, we also measured MIF levels in plasma and corresponding serum MIF concentration by ELISA in an independent study of 35 HCC patients and found that plasma MIF levels were significantly lower than their corresponding serum levels (data not shown). Thus, it is reasonable to consider that platelets release MIF during blood coagulation. We deem that measuring the amounts of free MIF in plasma is more likely to reflect biological function and patient outcome compared to the total amount in serum.

Many cytokine levels in the peripheral blood have been shown to decrease after tumor ablation.27 Postoperative change of plasma MIF levels may provide additional predictive value on tumor recurrence and prognosis. We compared preoperative and postoperative plasma MIF levels in an independent cohort of 80 HCC patients and found that most patients underwent a slight decline of plasma MIF levels on postoperative day 3, an intermediate decline on postoperative day 7, and an intense decline on postoperative day 30 (Supporting Information Fig. 3). Gando et al.28 suggested that MIF production occurs in patients undergoing hepatic resection and that surgical stress may play an important role in elevated circulative MIF levels. Liver injury and inflammatory response, which inevitably happens in liver resection, are always correlated with elevated MIF secretion.29 It is reasonable to propose that the postoperative plasma MIF levels only slightly decreased at the early stage due to the upregulation effects from surgical stress, as well as the liver inflammatory response. Subsequently, the marked decrease of plasma MIF levels at postoperative day 30 was ascribed to tumor ablation. Therefore, we suggest that measurement of postoperative plasma MIF levels should be performed after postoperative day 30 in future studies. Due to technical reasons, the sample size and follow-up time were not adequate to determine the relationship between postoperative plasma MIF levels and prognosis of HCC patients. Further prospective multicenter studies are imperatively needed. Nonetheless, we confirmed that elevated plasma MIF levels are likely derived from HCC cells in situ as most patients have markedly reduced levels after surgical resection of tumors.

Akbar et al.30 detected the serum MIF levels in patients with a series of liver disease and healthy controls and found that serum MIF levels were higher in HCC patients, liver cirrhosis and chronic hepatitis patients compared to normal controls. In addition, they detected the MIF expression in HCC tissue and peripheral blood mononuclear cell of patients with HCC and normal controls to speculate the resources of serum MIF. Compared to their study, our study has some superiority. Firstly, we did not simply confirm Akbar et al.'s results but analyzed the correlations of plasma MIF levels with clinicopathological characteristics of HCC patients and evaluated the diagnostic and prognostic value of plasma MIF levels in HCC patients. More importantly, we compared plasma MIF with the widely accepted HCC biomarker serum AFP and found the former have more powerful diagnostic and prognostic value. Secondly, in Akbar et al.'s study, the serum MIF levels in HCC and liver cirrhosis patients were 25.6 ± 15.3 and 18.9 ± 10.7 ng/ml, respectively, which are much lower than our data. The discrepancy is explainable. The liver disease background in Akbar et al.'s study mainly consists of hepatitis type C virus (HCV) infection [86.4% (57/66) in HCC patients and 57.7% (15/27) in liver cirrhosis patients]. This was different from our study, which largely consisted of HBV infection [88.5% (238/269) in HCC patients and 96.3% (130/135) in liver cirrhosis patients]. As a consequence, the disparity of serum MIF levels between 66 HCC patients and 26 liver cirrhosis patients was not enough to discriminate HCC from liver cirrhosis patients in Akbar et al.'s study. However, our study used a larger sample size cohort and clarified the powerful diagnostic value of plasma MIF. It also further confirmed the advantage of plasma over serum for HCC diagnosis. Lastly, our study collected more direct evidence to demonstrate the resource of peripheral MIF. The findings that intratumoral MIF expression significantly correlated with plasma MIF levels, which decreased to the normal range after tumor ablation, strongly suggest that plasma MIF is mainly derived from HCC cells.

Even so, there are still some limitations in our study. Firstly, diagnosis of early stage HCC could be missed in some of cirrhosis patients. That will partly led to the false-negative of HCC diagnoses and the bias of our study. Secondly, the cutoff of 20 μg/l was routinely used to perform serum AFP test in China. In our study, we compared this fixed cutoff of serum AFP with plasma MIF without selecting an optimal cutpoint of serum AFP. This could also bias our comparison toward favoring plasma MIF, as the AFP cutoff of 20 μg/l may not be optimal in China. Thirdly, our study was a retrospective single-centric study based on the hospital population. Although with robust statistical evidence, further multicentric prospective studies based on community or high HCC risk populations are still needed to enhance the generalizability and clinical applicability of our findings. Fourthly, the long-term follow-up should be performed in postoperative observation to further confirm the value of plasma MIF levels as an indicator for tumor remission. Finally, as most liver disease backgrounds of HCC and cirrhosis are hepatitis B virus infections in China, further studies should be performed in HCV infection patients to elevate their respresentativeness.

In conclusion, this investigation correlated plasma MIF levels with tumor development, occurrence, invasive phenotypes, prognosis and ablation of tumor focus. After validating the diagnostic and prognostic value of plasma MIF levels in an independent cohort, we suggest that plasma MIF levels have the potential to be a useful diagnostic and prognostic marker for HCC.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

We thank Wei-De Zhang for help in collection and primary analysis of clinicopathological data.

References

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information
  • 1
    Li L, Zhang S, Lu F. Research on characteristics of mortality spectrum and type composition of malignant tumors in China. Chin J Oncol 1997: 19: 3238.
  • 2
    Parkin DM, Bray F, Ferlay J, Pisani P. Global cancer statistics, 2002. CA Cancer J Clin 2005; 55: 74108.
  • 3
    Tang ZY. Hepatocellular carcinoma surgery—review of the past and prospects for the 21st century. J Surg Oncol 2005; 91: 956.
  • 4
    David JR. Delayed hypersensitivity in vitro: its mediation by cell-free substances formed by lymphoid cell-antigen interaction. Proc Natl Acad Sci USA 1966; 56: 727.
  • 5
    Bloom BR, Bennett B. Mechanism of a reaction in vitro associated with delayed-type hypersensitivity. Science 1966; 153: 802.
  • 6
    Chesney J, Metz C, Bacher M, Peng T, Meinhardt A, Bucala R. An essential role for macrophage migration inhibitory factor (MIF) in angiogenesis and the growth of a murine lymphoma. Mol Med 1999; 5: 18191.
  • 7
    Wilson JM, Coletta PL, Cuthbert RJ, Scott N, MacLennan K, Hawcroft G, Leng L, Lubetsky JB, Jin KK, Lolis E, Medina F, Brieva JA, et al. Macrophage migration inhibitory factor promotes intestinal tumorigenesis. Gastroenterology 2005; 129: 1485503.
  • 8
    Bucala R, Donnelly SC. Macrophage migration inhibitory factor: a probable link between inflammation and cancer. Immunity 2007; 26: 2815.
  • 9
    Meyer-Siegler K, Hudson PB. Enhanced expression of macrophage migration inhibitory factor in prostatic adenocarcinoma metastases. Urology 1996; 48: 44852.
  • 10
    Ren Y, Tsui HT, Poon RT, Ng IO, Li Z, Chen Y, Jiang G, Lau C, Yu WC, Bacher M, Fan ST. Macrophage migration inhibitory factor: roles in regulating tumor cell migration and expression of angiogenic factors in hepatocellular carcinoma. Int J Cancer 2003; 107: 229.
  • 11
    McShane LM, Altman DG, Sauerbrei W, Taube SE, Gion M, Clark GM. Reporting recommendations for tumor marker prognostic studies (REMARK). J Natl Cancer Inst 2005; 97: 11804.
  • 12
    Sun HC, Zhang W, Qin LX, Zhang BH, Ye QH, Wang L, Ren N, Zhuang PY, Zhu XD, Fan J, Tang ZY. Positive serum hepatitis B e antigen is associated with higher risk of early recurrence and poorer survival in patients after curative resection of hepatitis B-related hepatocellular carcinoma. J Hepatol 2007; 47: 68490.
  • 13
    Poon RT, Ng IO, Lau C, Yu WC, Yang ZF, Fan ST, Wong J. Tumor microvessel density as a predictor of recurrence after resection of hepatocellular carcinoma: a prospective study. J Clin Oncol 2002; 20: 177585.
  • 14
    Ren Y, Poon RT, Tsui HT, Chen WH, Li Z, Lau C, Yu WC, Fan ST. Interleukin-8 serum levels in patients with hepatocellular carcinoma: correlations with clinicopathological features and prognosis. Clin Cancer Res 2003; 9: 59966001.
  • 15
    Zhu XD, Zhang JB, Zhuang PY, Zhu HG, Zhang W, Xiong YQ, Wu WZ, Wang L, Tang ZY, Sun HC. High expression of macrophage colony-stimulating factor in peritumoral liver tissue is associated with poor survival after curative resection of hepatocellular carcinoma. J Clin Oncol 2008; 26: 270716.
  • 16
    Camp RL, Dolled-Filhart M, Rimm DL. X-tile: a new bio-informatics tool for biomarker assessment and outcome-based cut-point optimization. Clin Cancer Res 2004; 10: 72529.
  • 17
    McCabe A, Dolled-Filhart M, Camp RL, Rimm DL. Automated quantitative analysis (AQUA) of in situ protein expression, antibody concentration, and prognosis. J Natl Cancer Inst 2005; 97: 180815.
  • 18
    Dolled-Filhart M, McCabe A, Giltnane J, Cregger M, Camp RL, Rimm DL. Quantitative in situ analysis of beta-catenin expression in breast cancer shows decreased expression is associated with poor outcome. Cancer Res 2006; 66: 548794.
  • 19
    Leech M, Metz C, Bucala R, Morand EF. Regulation of macrophage migration inhibitory factor by endogenous glucocorticoids in rat adjuvant-induced arthritis. Arthritis Rheum 2000; 43: 82733.
  • 20
    Mizue Y, Ghani S, Leng L, McDonald C, Kong P, Baugh J, Lane SJ, Craft J, Nishihira J, Donnelly SC, Zhu Z, Bucala R. Role for macrophage migration inhibitory factor in asthma. Proc Natl Acad Sci USA 2005; 102: 144105.
  • 21
    He XX, Yang J, Ding YW, Liu W, Shen QY, Xia HH. Increased epithelial and serum expression of macrophage migration inhibitory factor (MIF) in gastric cancer: potential role of MIF in gastric carcinogenesis. Gut 2006; 55: 797802.
  • 22
    Hira E, Ono T, Dhar DK, El-Assal ON, Hishikawa Y, Yamanoi A, Nagasue N. Overexpression of macrophage migration inhibitory factor induces angiogenesis and deteriorates prognosis after radical resection for hepatocellular carcinoma. Cancer 2005; 103: 58898.
  • 23
    Ren Y, Law S, Huang X, Lee PY, Bacher M, Srivastava G, Wong J. Macrophage migration inhibitory factor stimulates angiogenic factor expression and correlates with differentiation and lymph node status in patients with esophageal squamous cell carcinoma. Ann Surg 2005; 242: 5563.
  • 24
    Calandra T, Bernhagen J, Metz CN, Spiegel LA, Bacher M, Donnelly T, Cerami A, Bucala R. MIF as a glucocorticoid-induced modulator of cytokine production. Nature 1995; 377: 6871.
  • 25
    Wu CY, Wu MS, Chiang EP, Chen YJ, Chen CJ, Chi NH, Shih YT, Chen GH, Lin JT. Plasma matrix metalloproteinase-9 level is better than serum matrix metalloproteinase-9 level to predict gastric cancer evolution. Clin Cancer Res 2007; 13: 205460.
  • 26
    Poon RT, Lau CP, Cheung ST, Yu WC, Fan ST. Quantitative correlation of serum levels and tumor expression of vascular endothelial growth factor in patients with hepatocellular carcinoma. Cancer Res 2003; 63: 31216.
  • 27
    Poon RT, Fan ST, Wong J. Clinical implications of circulating angiogenic factors in cancer patients. J Clin Oncol 2001; 19: 120725.
  • 28
    Gando S, Nishihira J, Kobayashi S, Morimoto Y, Matsushita M, Kemmotsu O. Systemic macrophage migration inhibitory factor release following hepatic resection. Surg Today 2001; 31: 6059.
  • 29
    Bourdi M, Reilly TP, Elkahloun AG, George JW, Pohl LR. Macrophage migration inhibitory factor in drug-induced liver injury: a role in susceptibility and stress responsiveness. Biochem Biophys Res Commun 2002; 294: 22530.
  • 30
    Akbar SM, Abe M, Murakami H, Tanimoto K, Kumagi T, Yamashita Y, Michitaka K, Horiike N, Onji M. Macrophage migration inhibitory factor in hepatocellular carcinoma and liver cirrhosis; relevance to pathogenesis. Cancer Lett 2001; 171: 12532.

Supporting Information

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
IJC_25918_sm_suppfig1.tif900KSupporting Figure 1
IJC_25918_sm_suppfig2.tif1634KSupporting Figure 2
IJC_25918_sm_suppfig3.tif15476KSupporting Figure 3
IJC_25918_sm_supptables.doc96KSupporting Tables

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.