Cancer Cell Biology
Downregulation of KiSS-1 expression is responsible for tumor invasion and worse prognosis in gastric carcinoma
Article first published online: 20 MAY 2004
Copyright © 2004 Wiley-Liss, Inc.
International Journal of Cancer
Volume 111, Issue 6, pages 868–872, 10 October 2004
How to Cite
Dhar, D. K., Naora, H., Kubota, H., Maruyama, R., Yoshimura, H., Tonomoto, Y., Tachibana, M., Ono, T., Otani, H. and Nagasue, N. (2004), Downregulation of KiSS-1 expression is responsible for tumor invasion and worse prognosis in gastric carcinoma. Int. J. Cancer, 111: 868–872. doi: 10.1002/ijc.20357
- Issue published online: 3 AUG 2004
- Article first published online: 20 MAY 2004
- Manuscript Accepted: 18 MAR 2004
- Manuscript Received: 5 JAN 2004
- gastric adenocarcinoma;
- RNase protection assay
KiSS-1 is a promising candidate tumor-suppressor gene and may play a key role in the metastatic cascade. The expression profile and the role of KiSS-1 in cancer progression are largely unknown in most of the cancers, including gastric cancer. In this study, KiSS-1 expression was evaluated by RNase protection assay and localization was done by in situ hybridization in 40 gastric cancers and their adjacent normal gastric mucosa. For comparison with clinicopathologic characteristics and patient prognosis, all patients were divided into 2 groups having high and low KiSS-1 expression by using the median as the cutoff value of KiSS-1 expression as determined by the RNase protection assay. Gastric cancers with low KiSS-1 had frequent venous invasion, distant metastasis and tumor recurrence. Accordingly, patients with low KiSS-1-expressing tumors had a significantly worse overall and disease-free survival. In multivariate analysis, KiSS-1 became the strongest independent prognostic factor among the conventional prognosticators for gastric cancer patients. Collectively, these findings suggest that KiSS-1 may play a crucial role in gastric cancer invasion and could be a useful target for therapeutic intervention. © 2004 Wiley-Liss, Inc.
Metastasis is the hallmark of malignancy and is an important decisive prognostic marker in most of the patients with malignant tumors. It is a complex multistep process and current theories embrace the notion of a concerted involvement of several proponent and inhibitory molecules in the metastatic process.1 To date, only a few such molecules are discovered and among them KiSS-1 is only recently discovered as a promising metastasis suppressor gene.2 The KiSS-1 cDNA was originally cloned by subtraction hybridization between a metastatic and a nonmetastatic melanoma cell line and has been proved to be competent in ameliorating the metastatic potentiality of several cancer cell lines in in vitro study.2, 3 Significant reduction in KiSS-1 or metastin expression in tumors with high metastatic potential has been reported.4, 5, 6 Moreover, reduced KiSS-1 expression became a strong prognostic marker in patients with urinary bladder cancer.4
More recently, it has been reported that kisspeptin or metastin, a product of the KiSS-1 gene, is thought to possess potent antimetastatic property and an orphan receptor for this protein has been detected.7, 8, 9, 10, 11, 12 Although the exact mechanism of the tumor suppressor activity of KiSS-1 is yet to be revealed, the KiSS-1 gene product has recently been shown to repress type IV collagenase (MMP-9) expression and a loss of KiSS-1 expression has been associated with loss of membrane E-cadherin expression.4, 10 Moreover, very recently it was shown that activation of a G-protein-coupled receptor (also known as AXOR12, GPR54, or hOT7T175) by KiSS-1 peptide changed cell morphology and actin filament reorganization in NIH3T3 cells.11 Suppression of cellular invasion and metastasis by transfection of KiSS-1 cDNA into metastatic cancer cell lines and suppressed metastasis in nude mice support a definite role of kiss-1 in the metastatic process.3, 7 These findings may open the possibility of future clinical application of this protein for prevention of cancer invasion and metastasis and thus may improve patient prognosis. These promising results provoked us to evaluate the expression and prognostic impact of KiSS-1 in gastric carcinoma, which is one of the most malignant tumors with frequent metastasis.
MATERIAL AND METHODS
Patients and samples
Between 1991 and 1998, 40 patients who underwent curative resection for gastric adenocarcinoma and had sufficient frozen tissue sample for conducting molecular analysis of KiSS-1 expression were included in this study. Of these patients, 10 were female and 30 were male; their ages ranged from 32 to 94 (mean, 65.8). Frozen tumor tissue and normal mucosal tissue were retrieved from tissue bank and used for RNase protection assay (RPA). An experienced pathologist checked all hematoxylin and eosin-stained slides and documented the pathologic characteristics of tumors, including venous and lymphatic invasion. Tumor stage was defined according to the UICC TNM (1997) staging system. All patients provided consent for use of tissue material for clinical research, and the University Ethical Committee approved the research protocol. Follow-up was complete in all patients. Data obtained at regular follow-up visits at the outpatient department were stored in a database specially designed for gastric carcinoma patients. In case of any recurrence, the exact date of disease recurrence was collected from the referring physician or from the physician who attended the patient for the first time for the diagnosis of the disease recurrence.
mRNA isolation and multiprobe RNase protection assay
Total RNA was extracted using the TRI reagent (Molecular Research Center, Cincinnati, OH) following the manufacturer's instruction.
Probe synthesis by in vitro transcription.
A cocktail of linearized plasmids (at XbaI site for KiSS-1) containing the inserted cDNA fragment of KiSS-1 (a kind gift from Prof. Danny R. Welch, University of Alabama, Birmingham, AL) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH; Pharmingen, San Diego, CA) was used for probe construction. The T7 polymerase-directed and [32P] (Amersham, Tokyo, Japan)-labeled antisense RNA probe was synthesized by a commercially available kit (Pharmingen). The in vitro transcription reaction was terminated by adding RNase-free DNase, incubated at 37°C for 1 hr and the synthesized probe was extracted by phenol/chloroform:isoamyl alcohol (50:1). The probe cocktail was ethanol-precipitated and the pellet was dissolved in hybridization buffer; 1 μl of probe was taken and was quantified with a scintillation counter.
Hybridization and RNase protection assay and quantification of protected fragments.
The probe cocktail was diluted with hybridization buffer to optimize the concentration at 3.63 × 105 cpm/μl probe counts; 10 μg of total RNA pellet from each sample was dissolved in 8 μl hybridization buffer and subsequently 2 μl diluted probe was added in each sample. The following steps of RPA (Pharmingen) and quantification of protected fragments were done as previously described.13
KiSS-1 signals were normalized to the GAPDH signals. mRNA from human placenta specimen was used as a positive control for comparative purposes.
In situ hybridization
To localize KiSS-1 mRNA, in situ hybridization technique was employed. Serial sections of selected samples positive (n = 5) or negative (n = 5) for KiSS-1 by the RPA were used for in situ hybridization. Paraffin-embedded sections were used for in situ hybridization. Serial sections were used from each patient for sense and antisense probe. A digoxigenin-labeled sense and antisense RNA probe was transcribed by T3 and T7 RNA polymerase, respectively, with a DIG RNA labeling kit according to the manufacturer's instructions (Boehringer Mannheim, Mannheim, Germany). Hybridization and the immunohistochemical steps were done as described previously.14 Placental tissue served as a positive control. Evaluation of KiSS-1 was done blindly and independently by 2 of the authors (D.K.D., Y.T.) according to a comprehensive scoring formula. Briefly, the intensity was scored as follows: 1, weak expression; 2, moderate expression; and 3, strong expression. The extent was scored as: 1, < 33% of the tumor cells had positive staining; 2, 33–67% of the tumor cells positive staining; and 3, > 67% of the tumor cells had positive staining. The results obtained with the 2 scales of the staining score were multiplied against each other, yielding a single scale with scores of 1, 2, 3, 4, 6 and 9.
The correlation between KiSS-1 expression and various clinicopathologic parameters was determined by chi-square with Yate's correction or Mann-Whitney U-test as appropriate. The survival rate was calculated by the Kaplan-Meier method and statistical significance between groups was determined by the log-rank test. Independent variables predicting survival were evaluated by multiple stepwise regression analysis using the Cox model. The Statview 4.5J (Abacus Concepts, Berkeley, CA) software was used for data analysis.
In situ hybridization
KiSS-1 had a very strong expression in positive control placental tissue. Strong and specific signals were noticed inside the placental syncytiotrophoblasts by the in situ hybridization technique (Fig. 1a). Compared to the strong expression in the placental tissue, gastric normal mucosa and tumors had a relatively weak expression of KiSS-1 (Fig. 1b and c). Only a faint background color was noticed in serial sections treated with the sense probe (Fig. 1d). Homogeneous expression of KiSS-1 was noticed in gastric mucosa sparing the proliferative zone of the mucosa. Most of the superficial foveolar epithelial cells and deep glandular cells were positive for KiSS-1 (Fig. 1b). Intestinal metaplasia adjacent to the tumor was positive for KiSS-1 (Fig. 1f), whereas KiSS-1 expression in tumors was weak and heterogeneous (Fig. 1e). Occasionally, infiltrating mononuclear cells also had faint signal for KiSS-1 (Fig. 1c). Positive samples by the RNase protection assay had distinct KiSS-1 expression by the in situ hybridization and negative samples by the RNase protection assay had only faint background color without any distinct signal for KiSS-1. All 5 positive tumors by RPA had an in situ hybridization score of > 4 and out of 5 negative tumors 4 had a score of < 2 and only 1 had a score of 3.
RNase protection assay
As it was noticed in in situ hybridization, KiSS-1 expression was much stronger in the placenta than in gastric samples (40:1; Fig. 1g). KiSS-1 expression was downregulated more in gastric tumors than in adjacent gastric mucosa and a marginal statistical significance was achieved only by the nonparametric statistical test (23.1 ± 7.0 vs. 18.3 ± 4.7, mean ± SE; 2-tailed t-test p = 0.3883; Wilcoxon matched pairs signed ranked test, p = 0.0452).
Protected bands were quantified by BAS2000 II (Fujix Basstation, version 1.31; Fuji Film, Tokyo, Japan) and were normalized to that of GAPDH signals. Tumors were divided into high and low KiSS-1 expression groups by using the median value as the cutoff value. Correlations between KiSS-1 expression and the clinicopathologic characteristics are shown in Table I. Among the several pathologic parameters, only venous invasion (p = 0.005) had a significant correlation with KiSS-1 expression. Downregulation of KiSS-1 expression was associated with increased venous invasion. Higher tendencies of lymph vessel invasion and lymph node metastasis were noticed in lower KiSS-1 expression group than in higher KiSS-1 expression group; however, neither reached statistically significant value (p = 0.056 and p = 0.088, respectively). Among the other parameters, only the tumor recurrence rate (p = 0.004) and distant metastasis (p = 0.038) had significant correlation with KiSS-1 expression.
|High KiSS-1 (n = 20)||Low KiSS-1 (n = 20)||p-value1|
|Tumor size (cm)||7.1 ± 4.6||6.7 ± 3.1||0.749|
|Lymph node metastasis|
Both the overall survival (OS) and the disease-free survival (DFS) were stratified into significant prognostic groups by KiSS-1 expression (Fig. 2). Patients with lower KiSS-1 expression had significantly higher probability of disease recurrence (p = 0.000, log-rank test) and shorter overall survival (p = 0.001, log-rank test). To evaluate the strength of the individual parameters including the conventional prognostic markers and KiSS-1 expression, a stepwise multivariate analysis was carried out. In both OS and DFS, KiSS-1 became the strongest [OS, risk rate (RR) = 4.89, p = 0.009; DFS, RR = 11.869, p = 0.005] independent prognostic marker among the other independent prognostic markers (T-stage: OS, RR = 2.029, p = 0.014, DFS, RR = 4.148, p = 0.0021; N-stage: DFS, RR = 1.812, p = 0.035; sex: OS, RR = 3.935, p = 0.040) in this cohort of gastric carcinoma patients.
Metastasis remains an important clinical problem for most of the malignancies, including gastric cancer. The prognosis of gastric cancer patients presenting with regional or distant metastasis remains dismal.15, 16 Although significant progress has been made in exploitation of the molecular and cellular biologic markers of cancer invasion and metastasis, only little is revealed in gastric carcinoma.
To the best of our knowledge, the present study is the first of its kind in gastric cancer in which we showed that downregulation of KiSS-1 expression produced frequent tumor invasion and it became a strong prognostic determinant in gastric carcinoma patients. Although KiSS-1 expression level was lower in gastric cancers when compared with normal gastric mucosa, there was a trend of statistical significance between the groups. This might be explained by the frequent metaplastic changes in gastric mucosa adjacent to the tumors and might have affected the KiSS-1 expression in gastric mucosa. In thyroid cancer, there was no significant difference in metastin (KiSS-1 product) expression between the normal and the malignant thyroid tissue.6 To date, only a few studies reported KiSS-1 expression profile in clinical materials, including those in melanoma and urinary bladder cancer.4, 5 In melanoma, 44% of examined tumors had low KiSS-1 expression. The exact mechanism of downregulation of KiSS-1 in tumors is yet to be elucidated. KiSS-1 is mapped to chromosome 1q and a loss of heterozygosity (LOH) of the long arm of chromosome 6 was reported to be associated with loss of KiSS-1 expression and progression of melanoma.5 Interestingly, high frequencies of allele losses on both chromosome 1q and 6q were noticed in advanced gastric carcinomas, implicating a close association between these LOH and downregulation of KiSS-1.17, 18, 19 Other possibilities such as homozygous deletion, promoter methylation, or mutation of KiSS-1 gene might be responsible for downregulation of this gene in tumors and needs further studies for clarification.
Invasion by tumor cells into the newly grown blood vessels is a crucial step of tumor metastasis. In this study, tumors having low KiSS-1 expression had frequent vascular invasion. Sanchez-Carbayo et al.4 reported a similar correlation between low expression of KiSS-1 and frequent vascular invasion and increased distant metastasis rate in urinary bladder carcinoma and they concluded that KiSS-1 plays a tumor suppressor role in bladder cancer and loss of expression is associated with tumor progression. We concur with them and our findings showed that downregulation of KiSS-1 expression produced tumor invasion and metastasis as well as increased the rate of disease recurrence in gastric carcinoma.
As reported in patients with urinary bladder cancer, KiSS-1 expression became a significant independent prognostic marker for gastric carcinomas. In this study, KiSS-1 expression became a significant predictor of patient survival in terms of overall and disease-free survival. Moreover, among the conventional prognostic predictors of gastric carcinoma, KiSS-1 remained the strongest independent predictor of patient survival. A probable survival advantage for KiSS-1 expression could be attributable to its tumor suppressor role through the prevention of cellular invasion, metastasis and tumor recurrence. Moreover, KiSS-1 may have an antiproliferative effect on cancer cells and this could be responsible for decreased tumor growth and increased patient survival. Indeed, Stafford et al.11 very recently described that KiSS-1 peptide inhibited cellular proliferation through intracellular Ca++ release and activation of protein kinase C. An increase in intracellular Ca++ was found to inhibit cell proliferation and induce cell differentiation and apoptosis in cancer cells.
In conclusion, the results of our study indicate that KiSS-1 downregulation in gastric carcinomas plays a key role in tumor invasion, particularly venous invasion, as well as the incidence of distant metastasis. Also, KiSS-1 expression could be a useful prognostic marker in gastric carcinoma patients and could be used to screen patients for more aggressive therapy. However, before attaining such a final decision, further studies in a large number of gastric carcinoma patients should be carried out.
The authors thank Prof. Danny R. Welch (University of Alabama, Birmingham, AL) for providing the KiSS-1 plasmid for his thoughtful discussion of this article.