Hepatocellular carcinoma (HCC) is the most common primary malignant tumor of the liver. Various factors contribute to HCC being regarded as one of the major malignant diseases (Kew,2002). Among these are its high incidence in many of the most populous countries in the world, its often fulminant course, poor response to conservative treatment, high recurrence rate after resection and liver transplantation, and grave prognosis. Since the 1990s, HCC has been the second most common cause of death from cancer in China. The most important and reliable prognostic factor of HCC is the occurrence of portal vein tumor thrombus (PVTT) (Giannelli et al.,2002). It is also known that the survival of patients with HCC is influenced by the ability of cancer cells to invade surrounding tissue and vessels (Giannelli et al.,2002). However, more studies are needed to elucidate the mechanisms that dictate tumor cell motility in HCC. The KiSS-1 gene has been found to be a metastasis suppressor gene in human melanoma and breast carcinoma cells without affecting tumorigenicity (Lee et al.,1996). Although loss of KiSS-1 expression has been shown to be associated with progression and poor prognosis of various cancers including esophagus cancer (Ikeguchi et al.,2004), gastric cancer (Dhar et al.,2004), and breast cancer (Stark et al.,2005), the relationship between loss of KiSS-1 expression and invasion and metastasis of HCC has not been well-defined. In this study, we investigated the expression of KiSS-1 gene in HCC and studied its role in invasion and metastasis. We showed that the KiSS-1 gene was downregulated in HCC and that it plays a role in regulating transcription of MMP-9, a matrix metalloproteinase important for cell motility and metastasis.
KiSS-1 has been identified as a putative metastasis-suppressor gene in human melanomas and breast cancer cell lines. Although loss of KiSS-1 expression has been associated with progression and poor prognosis of various cancers, the exact role of KiSS-1 expression in HCC is not well-defined. Our study investigated KiSS-1 expression levels in HCC and its role in invasion and metastasis of human HCC. The expression levels of KiSS-1 and MMP-9 protein were determined by tissue microarray (TMA) serial sections, immunohistochemistry and semi-quantitative image analysis. All clinical and histological data obtained were subjected to statistical analysis. The expression of KiSS-1 protein in HCC and intrahepatic metastasis lesions was significantly lower (P < 0.01) when compared with non-tumor liver tissue and normal liver tissue. Multivariate analysis revealed a significant inverse correlation between KiSS-1 expression and ○1 TNM stage, (F = 7.113, P < 0.01) and ○2intrahepatic metastasis (t = 2.898, P < 0.01). Loss of KiSS-1 in intrahepatic metastasis versus primary carcinomas was statistically significant (P<0.01). We also found a negative correlation between KiSS-1 and MMP-9 expression in HCC (r = -0.506, P < 0.01). We conclude that loss of KiSS-1 during HCC metastasis, along with a concomitant upregulation of MMP-9 suggests a possible mechanism for cell motility and invasion during HCC metastasis, with KiSS-1 emerging as a possible therapeutic target during HCC metastasis. Anat Rec, 292:1128–1134, 2009. © 2009 Wiley-Liss, Inc.
HCC tissue with surrounding non-tumor liver tissue and intrahepatic metastasis lesions (Nakashima and Kojiro,2001) (30 specimens of satellite nodules and 7 specimens of PVTT) were obtained from 150 HCC patients who underwent surgical hepatectomy between June 2000 and January 2007. This group comprised 124 men and 26 women ranging in age from 27–77 years (average age 51 years). None of the patients received preoperative chemotherapy or radiation therapy. The diagnosis of HCC was confirmed in all cases by pathology. According to the TNM staging system recommended by the International Union Against Cancer (1997), 14 patients were in stage I, 45 in stage II, and 91 in stage III. Edmondson histopathological examination indicated that 38 HCC samples were moderately differentiated and 112 samples were poorly differentiated. Non-tumor liver tissue was obtained from regions distant from the tumors. Normal liver tissue was obtained from 16 patients with hepatic hemangioma to serve as a control. Specimens were fixed in a 40 g/L formaldehyde solution and embedded in paraffin.
Modified Protocol for Preparing Paraffin Tissue Microarrays (TMA) (Zang et al.,2007)
Paraffin tissue microarrays were prepared using a modification of the TMA method of Kononen et al., (1998). Conventional paraffin tissue blocks were marked on the area of interest and the area was transferred to the recipient block using a holing needle, sampling needle and locating board. A total of 150 formalin-fixed, paraffin-embedded HCC tissues, the corresponding non-tumor liver tissues, 37 specimens of intrahepatic metastasis lesions, and 16 specimens of normal liver tissues were placed on the TMA. TMA serial sectioning was carried out in 4 μm using a microtome. Every 10th sections of tissue microarray has to be stained with H & E and made sure the pathology diagnosis are concurrent and matched to the adjacent serial sections.
Immunohistochemistry was performed using the EliVision™ Plus two-step system according to the manufacturer's instructions. The TMA serial slides were deparaffinized, rehydrated, and immersed in 3% hydrogen peroxide solution for 10 min. They were then boiled in citrate buffer (pH 6.0) for 5 min and cooled at room temperature for 1 hr. Blocking was carried out with 10% normal goat serum at 37°C for 30 min. The slides were then incubated in a 1:400 dilution of rabbit polyclonal anti-human KiSS-1 antibody (Santa Cruz, Biotechnology, USA) or a 1:350 dilution of mouse monoclonal anti-human MMP-9 antibody (Santa Cruz Biotechnology, USA) for 1 hr at 37°C, followed by three washes with PBS. The slides were incubated with polymerized HRP-anti mouse/rabbit lgG and visualized with 3-3′ diaminobenzidine (DAB). Counterstaining was carried out with hematoxylin. Paraffin sections of normal human placenta were used as a positive control. Negative control sections were incubated with preimmune serum, or with antiserum preabsorbed with a 100-fold excess of blocking peptide. PBS was substituted for primary antibody to serve as the blank control.
Microscopic Evaluation of KiSS-1 and MMP-9 Expression
Cells with a deposition of buffy-colored granules in the cytoplasm were scored as KiSS-1 or MMP-9 positive. Semi-quantitative image analysis was performed with Image-Pro Plus 4.5 system (Media Cybernetics, Inc. USA). Under a stereomicroscope, five prominently dense specimen areas with positive components were recorded in 400HF and the integral OD value (IOD) for each image was measured. The mean IOD of five images represented the level of KiSS-1 or MMP-9 protein expression for each specimen.
All values are presented as mean ± SD. The statistical significance in differences between two groups and among more than two groups was assessed by Student's t test and one-way ANOVA using SPSS 11.5 software. The influence of each variable on the intrahepatic metastasis was assessed by multivariate logistic regression analysis. Correlation between KiSS-1 and MMP-9 expression in HCC was investigated by Pearson correlation coefficient test. A P value of less than 0.05 was considered to be statistically significant.
Expression of KiSS-1 and MMP-9 in Human HCC and Intrahepatic Metastasis Lesions
Normal controls (Fig. 2A) and non-tumor liver tissues (Fig. 1B) showed high levels of KiSS-1 expression in the cytoplasm. There was moderate or no expression of KiSS-1 in HCC (Fig. 1C) or intrahepatic metastatic lesions (Fig. 1D). The IOD of KiSS-1 in HCC (2.616 ± 0.689) or intrahepatic metastasis lesions (1.822 ± 0.445) was significantly lower than that in non-tumor liver tissues (3.472 ± 0.395, P < 0.01) or in normal controls (3.650 ± 0.213, P < 0.01). The IOD of KiSS-1 in non-tumor liver tissues was not significantly different from that in normal controls (P > 0.05). Strong immunoreactivity of MMP-9 was seen in intrahepatic metastatic lesions (Fig. 2D). There was mild immunoreactivity of MMP-9 in normal controls (Fig. 2A), non-tumor liver tissues (Fig. 2B) and HCC without metastasis (Fig. 2C).
Relationship Between KiSS-1 Expression and Clinicopathologic Parameters
Table 1 shows the relationship between KiSS-1 expression and selected clinicopathologic parameters. KiSS-1 expression did not vary significantly with age, gender, tumor size, histological type or peplos invasion. However, we found a significant inverse correlation between KiSS-1 expression and ○1 TNM stage, (F = 7.113, P < 0.01) and○2 intrahepatic metastasis (t = 2.898, P < 0.01). Multivariate analysis using Logistic regression employed the variables of tumor size, histological grade, peplos invasion and KiSS-1 expression against intrahepatic metastasis. The results showed that intrahepatic metastasis had an inverse correlation with KiSS-1 expression (X2 = 4.471, P < 0.05, Table 2). When thirty seven specimens of intrahepatic metastasis were compared to thirty seven corresponding primary carcinomas, we found that the frequency of KiSS-1 loss in intrahepatic metastasis versus primary carcinomas was statistically significant (P < 0.01, Table 1).
|≤45 years||124||2.617 ± 0.718|
|>45 years||26||2.607 ± 0.555||t = 0.069||0.945|
|Male||49||2.600 ± 0.754|
|Female||101||2.624 ± 0.662||t = 0.213||0.831|
|≤3 cm||25||2.585 ± 0.709|
|>3 cm||125||2.767 ± 0.579||t = 1.205||0.230|
|Moderate||38||2.723 ± 0.729|
|Poor||112||2.579 ± 0.677||t = 1.112||0.268|
|Yes||122||2.580 ± 0.681|
|No||28||2.765 ± 0.723||t = 1.277||0.204|
|I||14||2.949 ± 0.630|
|II||45||2.841 ± 0.594||F = 7.113||0.001a|
|III||91||2.453 ± 0.700|
|Yes||111||2.884 ± 0.625|
|No||39||2.521 ± 0.690||t = 2.898||0.004a|
|P and Mb|
|Primary lesion||37||2.344 ± 0.623|
|Metastatic lesion||37||1.822 ± 0.445||t = 4.146||0.000a|
|Factors||Wald X2||95% CIa||P|
|≤3 cm vs. >3 cm||0.826||0.573–4.580||0.364|
|Yes vs. no||4.038||1.858–14.774||0.002b|
|Moderate vs. poor||9.806||0.207–1.453||0.226|
|Loss vs. preserved||4.471||0.177–0.936||0.034c|
Correlation Between KiSS-1 and MMP-9 in HCC
KiSS-1 and MMP-9 expression in HCC TMA serial sections stained with immunohistochemistry were analyzed by Pearson correlation coefficient test. The expression of MMP-9 showed a significant and negative correlation with KiSS-1 expression (r = −0.506, P < 0.01; Fig. 3). MMP-9 expression was significantly lower in KiSS-1 positive cells than in KiSS-1 negative cells.
Highly metastatic human melanoma cells were transfected with full-length human KiSS-1 cDNA (Lee et al.,1996). Athymic nude mice, injected in the tail vein with these KiSS-1 transfected cells showed a significant suppression in the number of lung metastases without affecting tumorigenicity, suggesting that KiSS-1 is a novel metastasis suppressor. KiSS-1 is known to encode a 145-amino acid residue precursor peptide that is processed to a 54-amino acid peptide with C-terminal amidation. The final peptide is called metastin and is a ligand for a G-coupled orphan receptor known as OT7T175/AXOR12/GPR54 (Muir et al.,2001; Ohtaki et al.,2001). Expression of KiSS-1 has been shown to be highest in human placenta and testis, followed by the pituitary and spinal cord, with moderate levels in the liver, pancreas, and intestines (Kotani et al.,2001; Harms et al.,2003).
Loss of KiSS-1 and/or hOT7T175 gene expression was found to be a significant predictor and potential biomarker of lymph node metastasis in esophageal squamous cell carcinoma (ESCC) (Ikeguchi et al.,2004). Gastric cancers with low KiSS-1 had frequent venous invasion, distant metastasis and tumor recurrence. Patients with low KiSS-1 expressing tumors had a significantly worse prognosis than patients with high KiSS-1 expressing tumors (Dhar et al.,2004). These findings suggest that KiSS-1 may play a crucial role in gastric cancer invasion and could be a useful target for therapeutic intervention. Similarly, a significantly reduced expression of KiSS-1 gene in breast cancer patients was found to lead to increased brain metastasis (Stark et al.,2005). However, the relationship between loss of KiSS-1 expression and invasion and metastasis of HCC has not been completely elucidated. In contrast to these studies, Ikeguchi et al., (2003) report the KiSS-1 expression did not vary significantly between non-cancerous cirrhotic livers and carcinomas and that overexpression of KiSS-1 was frequently correlated with HCC progression. Additionally, Schmid et al., (2007) studied the expression of KiSS-1 gene in HCC and its role in invasion, metastasis and prognosis of human HCC by immunohistochemistry with results that contradict ours. One possible explanation for these two studies differing significantly from other previous work and from our present study could be the existence of an additional level of KiSS-1 regulation at the translational level or existence of different cleavage variants of the KiSS-1 protein. In contrast to these two studies, Hou et al., (2007) reported that the levels of KiSS-1 mRNA and protein were significantly lower in patients with PVTT and patients with HCC and PVTT than in patients with HCC without PVTT. They found a negative correlation of KiSS-1 with MMP-9 expression in HCC and speculated that the KiSS-1 gene might suppress the formation of PVTT by suppressing MMP-9 synthesis. Our findings are in agreement with Hou et al., (2007). Furthermore, we demonstrated by multivariate analysis that loss of KiSS-1 expression in HCC was associated with intrahepatic metastasis in these patients. The KiSS-1 protein investigated in our study was identified by a rabbit polyclonal anti-human KiSS-1 antibody. This was raised against a recombinant protein corresponding to amino acids 1–145 representing full length KiSS-1 of human origin, consisting of kisspeptin-54 and other smaller peptides with stronger receptor binding activity which correlates to the biological behaviors of tumors.
Mechanisms that modulate invasion and metastasis in HCC are not clear. Invasion and metastasis of tumor cells is a multi-step process in which increased cell motility is accompanied by uncontrolled degradation of the basement membrane and components of the extracellular matrix. MMP-9 (gelatinase B, 92-kDa), contains fibronectin-like domains for collagen binding and is capable of degrading type I, IV, V, VII, and XI collagens and laminin. This proteolytic ability suggests that MMP-9 ultimately regulates cell migration, tumor growth, and angiogenesis (Freije et al.,2003; Ishii et al.,2003). Yan et al., (2001) reported that high KiSS-1 expression correlated with reduced binding of NF-kB p50/p65 to the MMP-9 promoter and diminished expression of MMP-9. Our study confirms that the expression of KiSS-1 was significantly and negatively correlated with MMP-9 expression. We propose that the KiSS-1 gene may suppress cancer metastases by regulating cell-matrix adhesion via altered MMP-9 levels.
In summary, we show that downregulation of KiSS-1 in HCC plays a key role in tumor invasion, particularly intrahepatic metastasis, as well as in the incidence of distant metastasis. We propose that KiSS-1 expression could be a useful prognostic marker in HCC patients to predict onset of metastasis and design appropriate therapeutic strategies. Further studies of KiSS-1 in HCC cells would give us valuable information on the exact mechanism by which cell motility and metastasis are enhanced in these tumors. The involvement of MMP-9 in this model makes this a particularly significant finding in light of the fact that elevated MMP-9 expression is associated with increased metastatic potential in many different cancers.