Clinicopathological significance of S100A4 expression in human hepatocellular carcinoma

Abstract

Objective

Prognostic markers for hepatocellular carcinoma (HCC) could help in the clinical management and understanding of its poor prognosis. S100 calcium binding protein A4 (S100A4) is directly involved in tumour metastasis. This study evaluated S100A4 gene expression in human HCC, to identify its role in tumour progression.

Methods

In this retrospective study, S100A4 protein levels in HCC samples from Chinese patients were evaluated using immunohistochemistry and compared with paired normal tumour-adjacent tissue samples (controls). All patients were evaluated for HCC recurrence.

Results

S100A4 levels were significantly increased in HCC samples compared with controls (n = 72, each sample type). Moderately or poorly differentiated tumours had significantly higher levels of S100A4 protein than well-differentiated tumours, and those with tumour vascular invasion showed significantly higher S100A4 levels than those without invasion. Risk of recurrence increased in patients positive for S100A4, and S100A4 positivity correlated with a shorter overall survival time.

Conclusion

This pilot study suggests S100A4 as a likely tumour marker for HCC that correlates with tumour differentiation, invasion, recurrence and overall survival. S100A4 could be a useful marker of tumour aggressiveness and prognosis.

Introduction

Tumours transition from the benign to the malignant state when their cells develop the ability to cross tissue barriers and invade surrounding tissue. ¹ Crossing basement membranes is considered to be the critical event in initiating metastasis; it involves attachment of tumour cells to the extracellular matrix, local proteolysis, and cellular migration. ² Upregulation of S100 calcium binding protein A4 (S100A4, a member of the S100 family of calcium-binding proteins) can result in a phenotype characterized by metastatic cancer.^3,4 Increased expression of the S100A4 gene has been shown in multiple cancer types.^5–7

The potential for use of S100A4 as a prognostic marker in cancer has been widely reported. One study in 349 breast cancer patients showed S100A4 to be the most significant predictor of patient survival, even when compared with well-established markers of disease progression. ⁸ A high survival rate was observed among patients with S100A4-negative tumours (80%) at 19 years' follow-up, compared with those with S100A4-positive tumours (11%). Two separate cohort studies confirmed this improved survival advantage, although survival outcomes in those studies were less pronounced.^9,10 The S100A4 gene is also overexpressed in gastric carcinoma, and its expression is significantly related to tumour size and lymph node metastasis. ¹¹ An association between S100A4 protein and patient survival has been demonstrated in a variety of tumour types including ovarian, pancreatic, bladder, nonsmall-cell lung and gallbladder cancer, oesophageal squamous cell carcinoma and malignant melanoma.^12,13

Hepatocellular carcinoma is one of the most common and aggressive tumours; it is ranked as the fifth most frequent cancer, but the third leading cause of cancer death worldwide. ¹⁴ The prognosis for HCC is extremely poor and is mainly attributed to a high rate of intrahepatic metastasis after treatment. ¹⁵ Despite advances in molecular technology that allow the identification of novel targets involved in HCC metastasis, the precise mechanisms involved in this process remain elusive.

The present retrospective study investigated the presence of S100A4 protein in HCC tumour samples from a comparatively large group of patients, with sufficient follow-up to assess whether its presence at the time of diagnosis was significantly associated with disease progression.

Patients and methods

Patients

This retrospective study included surgically resected HCC tumour tissue samples and paired normal tumour-adjacent tissue samples (controls) from consecutive Chinese patients who underwent hepatectomy for primary HCC, between January 2002 and July 2007. Data were collected from the archives of the Department of Pathology, The Affiliated Hospital Medical College, Qingdao University, Qingdao, China. Patients' tumours were classified according to the International Union Against Cancer tumour–node–metastasis classification system ¹⁶ and World Health Organization criteria. ¹⁷ Data collected from hospital archives included tumour size, stage and grade, extent of vascular invasion and α-fetoprotein levels. Tumour recurrence during patient follow-up was also recorded.

The study was approved by the Ethics Committee of The Affiliated Hospital Medical College, Qingdao University. Written informed consent was obtained from each patient before tissue acquisition.

Histopathology and immunohistochemical analysis

Tumour and control tissue samples were fixed in 10% neutral-buffered formalin, and subjected to routine paraffin-wax embedding and sectioning (4 -µm thick slices) using silanized slides. Each diagnosis was based on histopathological examination of Mayer's haematoxylin and eosin-stained, paraffin-wax embedded, sections.

To evaluate the level of S100A4 protein, sections were deparaffinized and rehydrated by passing through xylene three times for 5 min, followed by a graded series of ethanol (100%, 80%, 70% and 50% once each for 5 min) and distilled water for 10 min. Endogenous peroxidase activity was blocked using 3% hydrogen peroxide in methanol for 30 min at room temperature, between the first ethanol rehydration steps. Sections were washed in 0.5 mol/l Tris-buffered saline (TBS; pH 7.6) three times for 10 min after each of the following stages. Sections were blocked with 2% normal rabbit serum in 0.5 mol/l TBS for 20 min, then incubated with mouse antihuman S100A4 monoclonal antibodies (1 : 100 dilution; Santa Cruz Biotechnology, Santa Cruz, CA, USA) overnight at 4°C. Sections were then incubated with a biotinylated rabbit antimouse secondary antibody (1 : 50 dilution; DAKO, Hamburg, Germany) for 40 min at room temperature, followed by the avidin-biotinylated peroxidase complex for an additional 40 min at room temperature. After washing with distilled water for 10 min, antibody binding was visualized using 3,3′-diaminobenzidine according to the manufacturer's instructions, counterstained with Mayer's haematoxylin and eosin and mounted for light microscopy. For each tumour and control tissue sample, a corresponding section was incubated in TBS without the primary antibody, as a control for nonspecific binding. Breast cancer cases with known positive immunostaining for S100A4 protein were included as positive controls.

All samples were examined by three independent investigators. As the experimental conditions for all tissues were identical, evaluation of the staining for S100A4 protein was performed as described previously. ⁸ Positive staining for S100A4 protein was observed as yellow or brown staining in the cytoplasm of cancerous cells of HCC tumours. Briefly, the percentage of carcinoma cells with cytoplasmic staining for S100A4 protein was recorded from two sections for each specimen, using 10 fields/section at ×100 magnification. Samples were categorized into one of two groups, based on the level of immunostaining: positive, ≥5% of cells stained; negative, < 5% of cells stained.

Statistical analyses

All statistical analyses were performed using SPSS® statistical software, version 13.0 (SPSS Inc., Chicago, IL, USA) for Windows®. Differences in the rate of positive immunostaining for S100A4 protein by clinicopathological features of HCC were analysed using Pearson's χ²-test. Survival times were counted (in months) from the date of partial hepatectomy and survival analysis was performed by the Kaplan–Meier method with log-rank analysis. A P-value < 0.05 was considered to be statistically significant.

Results

In total, surgically resected tumour tissue samples and paired normal tumour-adjacent tissue samples from 72 consecutive Chinese patients were included in this study: 36 females and 36 males, with a mean age of 57 years (range 11–76 years). The study population included 62 (86%) cases of hepatitis B virus (HBV)-induced HCC. The mean tumour diameter was 4.8 cm (range, 2.2–12.6 cm). In terms of tumour grading, 12 samples were well differentiated, 29 were moderately differentiated and 31 were poorly differentiated. There were 43 tumours with, and 29 tumours without, vascular invasion. The α-fetoprotein levels were ≤400 µmmol/l in 18 cases and >400 µmmol/l in 54 cases. The tumour diameter was ≥3 cm in 49 cases and <3 cm in 23 cases. Tumour recurrence was recorded in 37 cases.

Positive staining for S100A4 protein in the cytoplasm of cancerous cells was diffusely scattered over the HCC tumour area (data not shown). HCC samples positive for S100A4 protein showed stronger staining intensity than their paired normal tumour-adjacent tissue (control) samples: 52.8% (38/72) of HCC samples were positive for S100A4 protein, whereas no positive paired controls were observed (P < 0.001).

Positive staining for S100A4 protein was significantly higher in poorly differentiated 87.1% (27/31) HCC samples than the in well-differentiated 8.3% (one of 12) or moderately differentiated 34.5% (10/29) HCC samples combined (P < 0.05). HCC samples with vascular invasion had significantly higher levels of S100A4 protein (72.1% [31/43]) than those without invasion (24.1% [7/29]) (P < 0.05). Higher levels of S100A4 protein were observed in HCC patients who went on to experience disease recurrence (75.7% [28/37]) compared with those who did not experience recurrence (28.6% [10/35]) (P < 0.05). No significant association was observed between S100A4 gene expression and other clinicopathological parameters (such as α-fetoprotein level or tumour size). There was a trend towards increased positive S100A4 gene expression in grade III tumours when compared with grade I tumours,^16,17 but this difference was not statistically significant.

A significant association was observed between the positive expression of S100A4 in cancerous cells and decreased survival time (P < 0.01, log-rank test). No such association was seen in cases where S100A4 expression was only observed in stromal cells.

Discussion

Hepatocellular carcinoma, which is the most common liver malignancy that is observed in adults, is the third leading cause of cancer deaths worldwide. ¹⁴ Clinically, high tumour differentiation grade, recurrence and presence of vascular invasion are significant independent predictors of patient survival.^18,19 The molecular mechanisms that underlie the clinical behaviour of HCC are still not well understood, however.

In addition to animal studies, an association between S100A4 gene expression and metastasis has been observed in a number of human studies examining the usefulness of S100A4 as a prognostic marker for cancer. A retrospective study of 349 invasive human breast cancer specimens, over a follow-up period of 14–20 years, demonstrated a high correlation between S100A4 expression and death, suggesting that S100A4 was of prognostic importance.^8,9 S100A4 has also be used a prognostic marker in a number of human cancers.^11–13 Expression of S100A4 in many cancers illustrates its potential use as a marker for metastasis and disease progression.

The present study documented increasing S100A4 expression as HCC progressed from low- to high-grade tumours, as well as identifying a possible association between tumour differentiation, vascular invasion and recurrence. S100A4 expression was significantly correlated with invasion and with undifferentiated tumour cells, which may reflect the biologically aggressive nature of the tumour. Increased S100A4 expression had an unfavourable impact on overall survival, further emphasizing the role of this molecule in HCC carcinogenesis and its relationship with poor prognosis. Positive S100A4 expression was associated with tumour recurrence, suggesting that the protein may play a role in predicting a patient subgroup that displays unfavourable outcomes. Analysing samples to establish the level of S100A4 expression could potentially lead to more targeted therapeutic strategies, with more intensive treatment and stricter follow-up surveillance for HCC tumours showing positive S100A4 expression.

There were some limitations to the present study. First, as is common in China, the majority of patients in the study population (86%) had HBV-induced HCC. It would therefore be important to confirm our findings in patients with other liver diseases. In addition, the sample size (72 patients) was relatively small. Our work should therefore be viewed as a pilot study, to be followed with larger prospective and multiethnic studies to confirm the present findings. Despite these limitations, the present study findings suggest that S100A4 expression could be a candidate marker for predicting HCC behaviour and serve as a potential therapeutic target in this disease.