- Email: [email protected]
A single nucleotide polymorphism in the matrix metalloproteinase-1 and interleukin-8 gene promoter predicts poor prognosis in tongue cancer
Auris Nasus Larynx 35 (2008) 381–389 www.elsevier.com/locate/anl
A single nucleotide polymorphism in the matrix metalloproteinase-1 and interleukin-8 gene promoter predicts poor prognosis in tongue cancer Yoshinori Shimizu *, Satoru Kondo, Akiko Shirai, Mitsuru Furukawa, Tomokazu Yoshizaki Department of Otolaryngology, School of Medicine, Kanazawa University, Takaramachi 13-1, Ishikawa 920-8640, Japan Received 7 March 2007; accepted 10 December 2007 Available online 13 February 2008
Abstract Objective: Matrix metalloproteinase-1 (MMP-1) and interleukin-8 (IL-8) play an important role in cancer development and metastasis. There is a single nucleotide polymorphism (SNP) located in the promoter region of MMP-1 and IL-8 that regulates gene expression. MMP-1 1607 2G/2G and IL-8 251 A/A genotypes enhance transcriptional activity and may be associated with increased risk in malignant tumors. We therefore evaluated the impact of these SNPs in tongue squamous cell carcinoma (SCC). Methods: In this study, we genotyped 69 tongue SCC patients. The expression of MMP-1 and IL-8 in tongue SCC patients was analyzed by immunohistochemistry. Results: We found a significant difference in IL-8 A/A genotypes with nodal recurrence (P = 0.0068). An analysis of disease-free survival rates showed that the presence of both MMP-1 2G/2G and IL-8 A/A genotypes was associated with a particularly poor prognosis (P = 0.0032) and was an independent prognostic factor (P = 0.001). The expression of MMP-1 was significantly correlated with the frequency of MMP-1 2G/2G genotypes (P = 0.049). Conclusion: These results suggest that SNP in the promoter region of MMP-1 and IL-8 plays an important role in tumor progression and recurrence through its expression in tongue SCC. # 2007 Elsevier Ireland Ltd. All rights reserved. Keywords: Single nucleotide polymorphism; Matrix metalloproteinase-1; Interleukin-8; Oral cancer; Immunohistochemistry
1. Introduction Malignant tumors have the ability to invade normal tissue and spread to distant sites, giving rise to metastasis, the major factors in the morbidity and mortality of cancer. The extracellular matrix acts as a structural support network within tissues and as a barrier to cell migration. Matrix degradation is mediated by the concerted action of several proteinases, including matrix metalloproteinase (MMPs), a family of enzymes [1]. Several recent studies have suggested * Corresponding author at: Division of Otolaryngology, Graduate School of Medicine, Kanazawa University, 13-1 Takara-machi, Kanazawa 9200934, Japan. Tel.: +81 76 265 2413; fax: +81 76 234 4265. E-mail address: [email protected] (Y. Shimizu).
new functional roles for MMPs in supporting tumor growth, modulating the extracellular matrix, regulating the availability of growth factors, and facilitating angiogenesis [2–4]. Matrix metalloproteinase-1 (MMP-1) has been implicated in tumor invasion and metastasis [5] as its specific ability to degrade type-I collagen, which is the most abundant substrate in the tumor surrounding stroma. Many reports have shown a significant negative correlation between the expression of MMP-1 and survival in advanced cancers [6–8]. The presence of a MMP-1 1G/2G polymorphism in the promoter region of the MMP-1 gene at 1607 bp (GenBank AF023338 2767) affects the transcriptional level of MMP-1 in both normal fibroblasts and in melanoma cells [9]. The MMP-1 2G polymorphism is positively correlated with an increased risk for developing lung cancer, ovarian cancer,
0385-8146/$ – see front matter # 2007 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.anl.2007.12.002
382
Y. Shimizu et al. / Auris Nasus Larynx 35 (2008) 381–389
colorectal cancer, and head and neck cancer [10–13]. These data establish a potential role for the 2G polymorphism in cancer development. Angiogenesis, or neovascularization, is the formation of new blood vessels that occurs during embryogenesis, wound healing, and tumor progression. In addition to the fact that angiogenesis is an indispensable phenomenon for the growth of tumors, the close association of angiogenesis with tumor metastasis has been well documented [14,15]. In recent studies, interleukin-8 (IL-8) has been shown to be a potent angiogenic factor implicated in tumor growth, metastasis and poor prognosis [16–19]. Promoter regions of a number of cytokine genes contain polymorphisms that directly influence cytokine production [20]. In the IL-8 251 (GenBank AF385628 3470) A/T polymorphism, the A allele tended to be associated with increased IL-8 production [21]. The frequency of IL-8 T/T genotypes was less than that of the IL-8 A/T or A/A genotype in prostate cancer patients [22]. We have been working on the regulation of invasion and metastasis in head and neck cancers, and revealed the important role of both MMPs and angiogenic factors. It is generally accepted that overexpression of such factors predicts a poor outcome in patients with various malignant tumors; however, studies on the expression level are always accompanied by the problem of reproducibility of the determined data. Analysis of genetic events, such as genotyping, provides quite stable data; however, the results may not be as impressive as the data from expression profiles. From this viewpoint, we evaluated the risk and prognostic value of MMP-1 1607 1G/2G and IL-8 251 A/ T polymorphisms in tongue squamous cell carcinoma (SCC). Moreover, we examined the expression of MMP-1 and IL-8 proteins by immunohistochemical examination, and the relationship between immunohistochemical examination and genotype data was also evaluated.
2. Materials and methods 2.1. Study subjects This study consisted of 69 patients with tongue SCC and 91 healthy controls, and all subjects were ethnic Japanese. The healthy controls consisted of staff, medical students, and inpatients of Kanazawa University Hospital without a previous diagnosis of malignant neoplasm. The data of tongue SCC patients in this study were obtained from medical records. From April 1983 to January 2000, 99 tongue SCC patients without metastasis were diagnosed and treated at the Department of Otolaryngology, Kanazawa University Hospital and Toyama City Hospital in Japan. Among them, 69 patients from whom DNA samples were available were recruited in this study. Clinical status was determined with the 1997 UICC/AJCC staging system [23]. Sixty-five patients were treated with surgery. Therapeutic
neck dissection was performed on 17 patients who had clinically positive nodes. Selective neck dissection was performed on 23 patients with clinically negative nodes who had T2 or more advanced tumor status. Neck dissection was not performed on 23 patients with stage I disease. Postoperative radiotherapy was administered to the neck of patients with pathologically positive lymph node metastasis. Although the tumor-free surgical margin was confirmed from frozen sections of the primary tumor, four patients were shown to have positive margins by using permanent pathological specimens. These patients received postoperative irradiation. Four patients underwent chemotherapy with or without radiation. The median follow-up time was 46 months (range, 4–96 months). Disease-free survival time was calculated from the date of treatment until the time of recurrence, defined as disease recurrence at the same site or the detection of metastases, including recurrence in the neck lymph nodes. 2.2. Sample preparation Surgical specimens were fixed with 10% formalin, embedded in paraffin, and the primary tumors were examined by immunohistochemical analysis. Paraffinembedded specimens were retrieved from the surgical pathology files of the Pathology Section of Kanazawa University Hospital and Toyama City Hospital in Japan. DNA samples of tongue SCC were extracted from paraffin blocks and those of healthy individuals from whole blood samples. This study design was approved by the Ethics Committee of Kanazawa University. 2.3. DNA extraction DNA from whole blood or paraffin-embedded tissue was extracted with a MagExtractor System MFX-2000 (TOYOBO, Osaka, Japan); DNA from paraffin-embedded tissue was also extracted with a modified protocol [24]. In brief, 10 mm 10 mm paraffin sections were incubated with 1 ml xylene at 55 8C for 15 min. After centrifugation, the supernatant fluid was discarded. The pellet was washed twice with 1 ml of 1/1 xylene/100% ethanol and twice with 100% ethanol. Pellets were incubated with 180 ml buffer ATL and 20 ml proteinase K at 55 8C for 48 h with intermittent vortexing, after which the manufacturer’s protocol was followed. 2.4. Genetic analysis Genotyping was accomplished using PCR followed by melting curve analysis with specific fluorescent hybridization probes in a Light Cycler System (Roshe, Mannheim, Germany) [25,26]. To detect MMP-1 1607 (GenBank AF023338 2767) 1G/ 2G genotypes, a 230-bp fragment was amplified by primer extension reaction (PCR) using forward and reverse primers,
Y. Shimizu et al. / Auris Nasus Larynx 35 (2008) 381–389
50 -GCA AGT GTT CTT TGG TCT CT-30 and 50 -CTT TCT GCG TCA AGA CTG-30 , respectively. The PCR profile to amplify a 230-bp product consisted of an initial melting step of 10 min at 95 8C, followed by 40 cycles of 10 s at 95 8C, 10 s at 60 8C, and 9 s at 72 8C. Following PCR, melting curve analysis was performed using a fluorescein-labeled mutation probe (50 -GAT AAG TCA TAT CTT TCT AAT TAT T-30 ) and Light Cycler Red 640-labeled anchor probe (50 -AAC TAC AAT TTC CTC ATC TAA GTG GCA TAA CAT GTG AGT TCA AGA GGG-30 ). For the IL-8 251 (GenBank AF385628 3470) A/T genotypes, the forward and reverse primers for PCR were 50 -CAT CCA TGA TCT TGT TC-30 and 50 -AAA TAC GGA GTA TGA CGA-30 , respectively. The PCR profile to amplify a 192-bp product consisted of an initial melting step of 10 min at 95 8C, followed by 40 cycles of 10 s at 95 8C, 15 s at 58 8C, and 8 s at 72 8C. Following PCR, melting curve analysis was performed using a fluorescein-labeled mutation probe (50 -GGT GAA TTA TCA AAT GTA TGC TTT TTT ATT TCT AGATAA C-30 ) and Light Cycler Red 640-labeled anchor probe (50 -TTA TCA TAG TAC CAT GTG ACA GAT ATA GTA CTA GAG TGG C-30 ). The genotypes of MMP-1 1607 1G/2G and IL-8 251 A/T revealed by Light Cycler System analysis were further confirmed with the ABI PRISM 310 genetic analyzer (Perkin-Elmer ABI, Foster City, CA). 2.5. Immunohistochemical analysis Immunohistochemical analysis was performed on paraffin-embedded 4-mm thick tissue sections as described previously [27]. The paraffin sections were dewaxed with xylene and rehydrated through graded concentrations of ethanol. Endogenous peroxidase was blocked with Endo-blocker (Biomeda, Foster City, CA) and nonspecific reactions were blocked with normal bovine serum (Wako, Osaka, Japan), then incubated with primary antibodies for 15 min at 45 8C. Primary antibodies against MMP-1 (41-1E5, 1:100 dilution) (Daiichi Fine Chemical, Takaoka, Japan) and IL-8 (6217, 1:20 dilution) (R&D Systems, Minneapolis, MN) were used. The sections were exposed to Universal Secondary Antibody (Research Genetics, Inc., Huntsville, AL). The sections were exposed to horseradish peroxidase-conjugated streptavidin, followed by diaminobenzidine as a chromogen. The nuclei were stained with methyl green. Skin tissue with epidermal ulcers was used as a positive control of MMP-1 protein expression and tonsil tissue for IL-8 protein expression. Negative controls were confirmed by staining with phosphatebuffered saline (PBS) instead of the primary antibody. 2.6. Evaluation of immunohistochemical staining for MMP-1 and IL-8 Stained samples of the primary tumors were reviewed as follows: using light microscopy, 2 areas that showed a high density of stained cells were selected in a 40 field, and then the number of stained cells of 200 cells was carefully
383
counted in the 2 areas in a 200 field without knowledge of the clinical data. The percentage of positive cell numbers obtained as an average of two counts was used as the expression score of MMP-1 and IL-8. For the purpose of correlation with SNPs, the expression scores of these proteins were arbitrarily divided into two categories: positive (MMP-1, positive cells 350% of tumor cells; IL-8, 330%) and negative (MMP-1, positive cells <50% of tumor cells; IL-8, <30%). 2.7. Statistical analysis Data were analyzed with a Macintosh personal computer (Apple Computer, Cupertino, CA) with Stat View software (Abacus Concepts, Berkeley, CA). Chi-square tests were used to compare the distribution of MMP-1 and IL-8 genotypes between cases and controls. Odds ratio and 95% confidence intervals were also calculated. Association of the MMP-1 and IL-8 genotype with clinicopathological parameters was evaluated by Fisher’s exact test. Statistical significance between the genotype and the survival rates was analyzed using Kaplan–Meier survival curves and the logrank test. The significance level was set at P < 0.05. Univariate and multivariate analyses were performed using the Cox proportional hazard regression model among the variables analyzed (age, gender, T, N, and M classifications, stage, and genotype). All variables other than age were dichotomized prior to data analysis, which was done to reduce the modeling flexibility and thereby yield a more robust final model. P-values <0.05 in the univariate models were included in multivariate analysis. Multivariate analysis was performed with backward elimination using a significance-level-to-stay of 0.05.
3. Results 3.1. Patient characteristics Patient characteristics are shown in Table 1. 3.2. Distribution of MMP-1 and IL-8 SNP Real-time PCR-assisted analysis of MMP-1 1G/2G and IL-8 A/T polymorphisms was carried out using the melting curves of the hybridization probes. Figs. 1 and 2A show the first differentiation of the fluorescence over time in MMP-1 and IL-8. The melting maximum peaks of MMP-1 1G/1G, 1G/2G, and 2G/2G genotypes were 48.3 8C, 48.3 8C and 54.4 8C, and 54.4 8C, respectively. The melting maximum peaks of the IL-8 A/A, A/T, and T/T genotypes were 62.0 8C, 62.0 8C and 64.6 8C, and 64.6 8C, respectively. The genotypes of MMP-1 and IL-8 were further confirmed by DNA sequencing. Figs. 1 and 2B show representative electropherograms of each genotype. A 10% random sample of the genotypes identified by the Light Cycler System was
384
Y. Shimizu et al. / Auris Nasus Larynx 35 (2008) 381–389
Table 1 Clinicopathological backgrounds of patients with tongue SCC
further evaluated by DNA sequencing, and 100% reproducibility was confirmed. Table 2 shows the distribution of MMP-1 1G/2G and IL-8 A/T SNP in tongue SCC patients and controls. The genotype distribution among controls was compatible with the Hardy– Weinberg equilibrium. No statistically significant difference in the frequency of MMP-1 or IL-8 genotypes was detected between patients and controls. 3.3. The relationship between genotypes and clinicopathological factors
The values are the numbers of patients. Tumor (T) and node (N); classification and stage were based on the International Union Against Cancer (UICC) Classification, 1997. SD: standard deviation.
The relationship between genotypes and the clinicopathological factors of 69 patients including age, sex, T classification, N classification, clinical stage, local recurrence, nodal recurrence, and distant metastatic recurrence were evaluated by Fisher’s exact test (Table 3). We found a significant difference in genotype distribution (IL-8 T/T or A/T vs. A/A) by nodal recurrence. Patients with the IL-8 A/A genotype had significantly more frequent nodal recurrence than those with the IL-8 T/T or A/T genotype. There was no significant association between the polymorphism and other clinicopathological factors.
Fig. 1. (A) Light Cycler System analysis of MMP-1 genotype. First differentiation of fluorescence over time. The melting maximum of the 2G-hybridization probe is 48.3 8C in the case of the 1G/1G genotype, 48.3 8C and 54.4 8C in the 1G/2G genotype, and 54.4 8C in the 2G/2G genotype. (B) Representative electropherogram of each type of polymorphism. Boxes indicate MMP-1 SNP, homozygous 1G/1G genotype with single G peak and clean sequence, heterozygous 1G/2G genotype with double G peak and uninterpretable sequence, homozygous 2G/2G genotype with double G peak and clean sequence.
Y. Shimizu et al. / Auris Nasus Larynx 35 (2008) 381–389
385
Fig. 2. (A) Light Cycler System analysis of IL-8 genotype. First differentiation of fluorescence over time. The melting maximum of the T-hybridization probe is 62.0 8C in the case of the A/A genotype, 62.0 8C and 64.6 8C in the A/T genotype, and 64.6 8C in the case of the T/T genotype. (B) Representative electropherogram of each type of polymorphism. Boxes indicate the IL-8 SNP, homozygous A/A genotype with single A peak, heterozygous A/T genotype with overlap A/T peak, homozygous T/T genotype with single T peak.
3.4. Immunohistochemical analysis The expression of MMP-1 and IL-8 was examined by immunohistochemical staining of the samples (Fig. 3). These signals were detected as dark brown staining. One of 69 paraffin sections was not available for this study because of insufficient residual tumor volume after DNA extraction. Thus, the immunohistochemical study of MMP-1 and IL-8 expression in the tumor was performed in the remaining 68 Table 2 Single nucleotide polymorphisms (SNP) in MMP1 and IL-8 promoter
P-values are based on the comparison of odds ratios in patients with 2G/2G vs. 1G/1G or 1G/2G genotypes for MMP-1 SNP, and those with A/A vs. T/T or A/T genotypes for IL-8 SNP. Odds ratio and 95% CI were calculated with 1G/1G or 1G/2G genotypes as a reference group in MMP-1 SNP, and T/T or A/T genotypes as a reference group in IL-8 SNP. CI: confidence interval.
cases. Table 4 shows the association of genotypes with the expression of MMP-1 and IL-8. In the MMP-1 stain, a high expression of MMP-1 was significantly associated with the MMP-1 2G/2G genotype. On the other hand, IL-8 expression had no significant correlation with the IL-8 genotype. The expression of MMP-1 and IL-8 had no significant relationship with survival (data not shown). 3.5. Association of MMP-1 and IL-8 genotypes with disease-free survival To evaluate the prognostic value of genotypes, the relationship between disease-free survival and MMP-1 and IL-8 genotypes was studied. Patients carrying neither the MMP-1 2G/2G nor the IL-8 A/A genotype had a significant trend toward decreased disease-free survival (Figs. 4 and 5); however, patients who carried both MMP-1 2G/2G and IL-8 A/A genotypes (MMP-1 + IL-8 high expression genotypes) had significantly shorter disease-free survival (Fig. 6). The short disease-free survival mainly reflected nodal recurrence. All four patients carrying both the MMP-1 2G/2G and IL-8 A/A genotypes had nodal recurrence, but did not have an advanced N stage (N0–1) at first diagnosis. Prognostic factors were analyzed by the Cox proportional hazard regression model among the variables (age, gender,
386
Y. Shimizu et al. / Auris Nasus Larynx 35 (2008) 381–389
Table 3 Relationship of MMP-1 and IL-8 genotypes with clinicopathological parameters
Table 4 Relationship of MMP-1 and IL-8 expression with SNP
*
Significant.
T, N, M classifications, stage, and genotypes). Among them, T4, advanced node (N2 and N3), stage IV, and the presence of both the MMP-1 2G/2G and IL-8 A/A genotypes were significant hazards in univariate analysis (Table 5). As T4 and advanced node (N2 and N3) are involved in stage IV, multivariate analysis was performed with T4, N2–3 and the presence of both MMP-1 2G/2G and IL-8 A/A genotypes. As shown in Table 6, the presence of both the MMP-1 2G/2G and IL-8 A/A genotypes remained an independent prognostic factor (P = 0.001; hazard ratio of other genotypes = 6.51). The category of N2–3 also has independent prognostic significance (P = 0.0001; hazard ratio = 8.74).
4. Discussion *
Significant.
Information on the role of MMPs and angiogenic factors in invasion and metastasis has been accumulating in terms of neoplastic cells and their surrounding microenvironment [2– 4]. Recent studies indicate a strong link between increased
Fig. 3. Immunohistochemical study of MMP-1 (A and B) and IL-8 (C and D) in cancer specimens (3,30 -diaminobenzidine tetrahydrochloride and methyl green stain, original magnification 200). MMP-1 2G/2G genotype (A) shows high expression in comparison with the 1G/2G genotype (B). IL-8 A/A genotype (C) shows high expression in comparison with the T/T genotype (D), although IL-8 expression had no significant correlation with the IL-8 genotype.
Y. Shimizu et al. / Auris Nasus Larynx 35 (2008) 381–389
Fig. 4. Kaplan–Meier analysis of disease-free survival according to the MMP-1 genotype (1G/1G + 1G/2G vs. 2G/2G). Statistical analysis was performed by the log-rank test.
MMP-1 expression, the presence of the 2G allele, and a poor clinical outcome [9–12]. Moreover, the association of IL-8 expression, A/A genotype, with poor outcome has also become a popular paradigm in the clinical cancer research setting [22]; however, these studies did not show a link based on genotype protein expression and clinical outcome. In our study, there was no significant difference in the frequency of either the MMP-1 or IL-8 genotype between tongue SCC and the controls. Although the distribution of MMP-1 1G/2G SNP in tongue SCC patients and controls had no statistically significant difference in the frequency of MMP-1, the Pvalue was 0.077, very close to the 0.05 that we set as the significance level in this study. A larger sample study may show a significant increased risk for tongue cancer. On the other hand, we found a significant correlation between the presence of both the MMP-1 2G/2G and IL-8 A/A genotypes with disease-free survival. Despite the small sample size, these data suggest that MMP-1 and IL-8 SNP are not strongly associated with the carcinogenesis of tongue SCC but may contribute to the progression of tongue SCC via the upregulation of MMP-1 and IL-8 gene expression. This
Fig. 5. Kaplan–Meier analysis of disease-free survival according to the IL-8 genotype (T/T + A/T vs. A/A). Statistical analysis was performed by the logrank test.
387
Fig. 6. Kaplan–Meier analysis of disease-free survival according to the presence of both the MMP-1 2G/2G and IL-8 A/A genotype vs. the combination of the MMP-1 and IL-8 genotype. MMP-1 + IL-8 high expression genotype: the presence of both MMP-1 2G/2G and IL-8 A/A genotypes. MMP-1 + IL-8 low expression genotype: the other combination of MMP-1 and IL-8 genotypes. Statistical analysis was performed by the log-rank test.
suggests that the MMP-1 2G/2G or IL-8 A/A genotype could not predict the prognosis of patients with tongue SCC, but this combination of two genotypes could have a critical role in malignant formation by matrix degradation and angiogenesis. Many reports showed a significant negative correlation between the expression of MMP-1 and survival in advanced cancers [6–8]. Previously, we reported that the MMP-1 2G/2G genotype predicted a poor prognosis in nasopharyngeal carcinoma [28]. Angiogenesis has been used as a prognostic indicator in a variety of cancers and is believed to be controlled by angiogenic factors, including IL-8. Many previous studies have shown a significant relationship between IL-8 and clinicopathological variables, including prognosis in malignant neoplasms [16–19]. Head and neck SCC patients exhibiting high levels of IL-8 were reported to have clinically more aggressive disease manifested by a higher TNM stage, more frequent Table 5 Univariate Cox regression analyses of variables in patients with tongue SCC
Factors whose P-values were >0.05 in univariate models were not included in multivariate analysis. MMP-1 2G/2G + IL-8 A/A represents the presence of both MMP-1 2G/2G and IL-8 A/A genotypes. y.o.: years old. CI: confidence interval. *Significant.
388
Y. Shimizu et al. / Auris Nasus Larynx 35 (2008) 381–389
Table 6 Multivariate Cox regression analyses of variables in 69 patients with tongue squamous cell carcinoma
Multivariate analysis was performed with backward elimination using a significance-level-to-stay of 0.05. MMP-1 2G/2G + IL-8 A/A represents the presence of both MMP-1 2G/2G and IL-8 A/A genotypes. CI: confidence interval. *Significant.
recurrence, and shorter disease-free interval [29]. Therefore, IL-8 SNP affecting the transcriptional level was thought to induce the overexpression of its protein and poor prognosis in malignant tumors. Moreover, two high-affinity IL-8 receptors, IL-8 receptor A and IL-8 receptor B, have been identified [30], and are reported to be present in malignant tumors, including head and neck squamous cell carcinoma [31]. The expression of IL-8 receptor A and B on tumor cells is thought to promote the metastatic ability of tumor cells by IL-8. Recently, we reported that the expression of IL-8 receptor A in nasopharyngeal carcinoma significantly correlated with a shorter overall survival rate [32]. IL-8 receptor A is specific to IL-8, whereas several other cytokines also bind to IL-8 receptor B; therefore, this result suggests that IL-8 promotes tumor growth as well as metastasis through its receptors. Patients with IL-8 A/A genotypes showed significantly more frequent nodal recurrence compared to those with either IL-8 T/T or IL-8 A/T, suggesting that IL-8 promotes lymph node metastasis in patients with tongue SCC. Recent studies of lymph angiogenesis revealed the contribution of angiogenic factors to lymphogenesis as well as angiogenesis. Overexpression of IL-8 mRNA was highly associated with non-small-cell lung cancer in advanced stages, distant lymph node metastasis, high tumor microvessel count, short survival, and early relapse [33]. In human gastric carcinoma, the IL-8 level in the neoplasms [34] and drainage veins [35] correlated significantly with lymphatic invasion. Thus, patients with IL-8 A/A genotypes, who are prone to the induction of more IL-8 [21], should be carefully treated, especially for nodal disease. The relationship between nodal recurrence and the type of treatment, including neck dissection, radiotherapy and chemotherapy, were evaluated by Fisher’s exact test. We found no significant difference in the therapy by nodal recurrence. We analyzed the disease-free survival of each type of treatment. The factors among treatments were analyzed by the Cox proportional hazard regression model among the variables (neck dissection, radiotherapy, chemotherapy), none of which were significant hazards in multivariate analysis (P = 0.29, 0.78 and 0.1, respectively). Thus, we thought that the type of treatment did not affect nodal recurrence and disease-free survival in this study.
In our study, the high expression of MMP-1 protein in tongue SCC tissue correlated significantly with the MMP-1 2G/2G genotype. Rutter et al. [9] described using transient transfections in A2058 melanoma cells, in which the MMP-1 2G polymorphism led to a more than 20-fold increase of mRNA expression in melanoma cells, as well as 5–10-fold in normal fibroblasts. In ovarian cancers, tumors carrying 2G alleles have seven times higher expression levels of MMP-1 than tumors carrying no 2G alleles [11]. As MMP-1 protein levels mirror MMP-1 mRNA expression, the increase in MMP-1 transcription may increase protein expression. On the other hand, IL-8 expression had no significant correlation with IL-8 genotypes, which may reflect that the DNA sequence around 251 SNP is less influential than other cis-acting transcription regulatory elements in the promoter of IL-8. Alternatively, similar studies with larger sample sizes may result in a statistically significant correlation of IL-8 genotypes with IL-8 protein expression. Multivariate Cox regression analysis demonstrated that advanced nodal status (N2 and N3) and the presence of both the MMP-1 2G/2G and IL-8 A/A genotypes were independent prognostic factors. On the other hand, the expression of MMP-1 and IL-8 proteins had no significant relationship with survival. This suggests that the MMP-1 2G/2G and IL-8 A/A genotypes could predict the prognosis of patients with tongue SCC more precisely than protein expression. It is generally accepted that overexpression of MMP-1 and IL-8 predicts a poor outcome in patients with various malignant tumors [6–8,16–19] and there was no evidence of an increased expression of MMP-1 and IL-8 in this study; however, the quantitative study of gene expression, especially in immunohistochemical studies, is always associated with the problem of reproducibility and quantitative results. Furthermore, the expression of gene products is easily influenced by tumor-surrounding conditions compared with the DNA sequence. Therefore, the analysis of genetic events, such as genotyping, provides quite stable data, even though it may not be as impressive as the data from expression profiles. This drawback can be overcome by using a combination of several genotype analyses, as shown in this study. Patients with the IL-8 A/A genotype had more frequent nodal recurrence than those with either IL-8 T/T or IL-8 A/T in this study. In addition, we demonstrated that the presence of both the MMP-1 2G/2G and IL-8 A/A genotypes were independent prognostic factors and had a poor outcome in disease-free survival. The IL-8 A/A genotypes could be a risk factor for nodal recurrence; thus, patients with IL-8 A/A genotypes, especially with MMP-1 2G/2G, should be carefully treated for nodal disease, including selective neck dissection or the addition of adjuvant chemotherapy for T12N0 stage, and should be frequently followed for nodal recurrence. These therapeutic strategies could improve the prognosis of patients with IL-8 A/A genotypes without increasing morbidity in patients with IL-8 A/T or T/T genotypes.
Y. Shimizu et al. / Auris Nasus Larynx 35 (2008) 381–389
These genotypes can be confirmed in 2 days by the Light Cycler System using DNA from the whole blood of patients. Therefore, analysis of genotypes can provide quite stable data than data from expression profiles such as immunohistochemical studies, and is suitable for the treatment of patients with tongue SCC. Although it is not conclusive because of the rather small sample size, SNPs in the promoter region of MMP-1 and IL8 can be independent prognostic factors, and thus, the combination of this information will be beneficial to predict the outcome of patients with tongue SCC.
References [1] Brinckerhoff CE, Rutter JL, Benbow U. Interstitial collagenases as markers of tumor progression. Clin Cancer Res 2000;6:4823–30. [2] Basset P, Okada A, Chenard MP, Kannan R, Stoll I, Anglard P, et al. Matrix metalloproteinases as stromal effectors of human carcinoma progression: therapeutic implications. Matrix Biol 1997;15:535–41. [3] Nelson AR, Fingleton B, Rothenberg ML, Matrisian LM. Matrix metalloproteinases: biologic activity and clinical implications. J Clin Oncol 2000;18:1135–49. [4] Stetler-Stevenson WG. Matrix metalloproteinases in angiogenesis: a moving target for therapeutic intervention. J Clin Invest 1999;103: 1237–41. [5] Rigel DS, Friedman RJ, Kopf AW, Silverman MK. Factors influencing survival in melanoma. Dermatol Clin 1991;9:631–42. [6] Murray GI, Duncan ME, O’Neil P, Melvin WT, Fothergill JE. Matrix metalloproteinase-1 is associated with poor prognosis in colorectal cancer. Nat Med 1996;2:461–2. [7] Murray GI, Duncan ME, O’Neil P, McKay JA, Melvin WT, Fothergill JE. Matrix metalloproteinase-1 is associated with poor prognosis in oesophageal cancer. J Pathol 1998;185:256–61. [8] Ito T, Ito M, Shiozawa J, Naito S, Kanematsu T, Sekine I. Expression of the MMP-1 in human pancreatic carcinoma: relationship with prognostic factor. Mod Pathol 1999;12:669–74. [9] Rutter JL, Mitchell TI, Buttice` G, Meyers J, Gusella JF, Ozelius LJ, et al. A single nucleotide polymorphism in the matrix metalloproteinase-1 promoter creates an Ets binding site and augments transcription. Cancer Res 1998;58:5321–5. [10] Zhu Y, Spitz MR, Lei L, Mills GB, Wu X. A single nucleotide polymorphism in the matrix metalloproteinase-1 promoter enhances lung cancer susceptibility. Cancer Res 2001;61:7825–9. [11] Kanamori Y, Matsushima M, Minaguchi T, Kobayashi K, Sagae S, Kudo R, et al. Correlation between expression of the matrix metalloproteinase-1 gene in ovarian cancers and an insertion/deletion polymorphism in its promoter region. Cancer Res 1999;59:4225–7. [12] Ghilardi G, Biondi ML, Mangoni J, Leviti S, DeMonti M, Guagnellini E, et al. Matrix metalloproteinase-1 promoter polymorphism 1G/2G is correlated with colorectal cancer invasiveness. Clin Cancer Res 2001;7:2344–6. [13] Hashimoto T, Uchida K, Okayama N, Imate Y, Suehiro Y, Hamanaka Y, et al. Association of matrix metalloproteinase (MMP)-1 promoter polymorphism with head and neck squamous cell carcinoma. Cancer Lett 2004;211:19–24. [14] Weidner N, Semple JP, Welch WR, Folkman J. Tumor angiogenesis and metastasis-correlation in invasive breast carcinoma. N Engl J Med 1991;324:1–8. [15] Carmeliet P, Jain RK. Angiogenesis in cancer and other diseases. Nature 2000;407:249–57. [16] Ugurel S, Rappl G, Tilgen W, Reinhold U. Increased serum concentration of angiogenic factors in malignant melanoma patients correlates with tumor progression and survival. J Clin Oncol 2001;19:577–83.
389
[17] Ren Y, Poon RT, Tsui HT, Chen WH, Li Z, Lau C, et al. Interleukin-8 serum levels in patients with hepatocellular carcinoma: correlations with clinicopathological features and prognosis. Clin Cancer Res 2003;9:5996–6001. [18] Kassim SK, El-Salahy EM, Fayed ST, Helal SA, Helal T, Azzam EelD, et al. Vascular endothelial growth factor and interleukin-8 are associated with poor prognosis in epithelial ovarian cancer patients. Clin Biochem 2004;37:363–9. [19] Benoy IH, Salgado R, Van Dam P, Geboers K, Van Marck E, Scharpe´ S, et al. Increased serum interleukin-8 in patients with early and metastatic breast cancer correlates with early dissemination and survival. Clin Cancer Res 2004;10:7157–62. [20] Bidwell J, Keen L, Gallagher G, Kimberly R, Huizinga T, McDermott MF, et al. Cytokine gene polymorphism in human disease: on-line databases, supplement 1. Genes Immun 2001;2:61–70. [21] Hull J, Thomson A, Kwiatkowski D. Association of respiratory syncytial virus bronchiolitis with the interleukin 8 gene region in UK families. Thorax 2000;55:1023–7. [22] McCarron SL, Edwards S, Evans PR, Gibbs R, Dearnaley DP, Dowe A, et al. Influence of cytokine gene polymorphisms on the development of prostate cancer. Cancer Res 2002;62:3369–72. [23] Sobin LH, Wittekind C, editors. TNM classification of malignant tumors. 5th ed., New York, NY: Wiley-Liss; 1997. p. 4–8. [24] Wickham CL, Boyce M, Joyner MV, Sarsfield P, Wilkins BS, Jones DB, et al. Amplification of PCR products in excess of 600 base pairs using DNA extracted from decalcified, paraffin wax embedded bone marrow trephine biopsies. Mol Pathol 2000;53:19–23. [25] Wittwer CT, Ririe KM, Andrew RV, David DA, Gundry RA, Balis UJ. The LightCycler: a microvolume multisample fluorimeter with rapid temperature control. Biotechniques 1997;22:176–81. [26] Nojiri T, Morita H, Imai Y, Maemura K, Ohno M, Ogasawara K, et al. Genetic variations of matrix metalloproteinase-1 and -3 promoter regions and their associations with susceptibility to myocardial infarction in Japanese. Int J Cardiol 2003;92:181–6. [27] Yoshizaki T, Sato H, Maruyama Y, Murono S, Furukawa M, Park CS, et al. Increased expression of membrane type 1-matrix metalloproteinase in head and neck carcinoma. Cancer 1997;79: 139–44. [28] Kondo S, Wakisaka N, Schell MJ, Horikawa T, Sheen TS, Sato H, et al. Epstein-Barr virus latent membrane protein 1 induces the matrix metalloproteinase-1 promoter via an Ets binding site formed by a single nucleotide polymorphism: enhanced susceptibility to nasopharyngeal carcinoma. Int J Cancer 2005;115:368–76. [29] Eisma RJ, Spiro JD, Kreutzer DL. Role of angiogenic factors: coexpression of interleukin-8 and vascular endothelial growth factor in patients with head and neck squamous carcinoma. Laryngoscope 1999;109:687–93. [30] Holmes WE, Lee J, Kuang WJ, Rice GC, Wood WI. Structure and functional expression of a human interleukin-8 receptor. Science 1991;253:1278–80. [31] Richards BL, Eisma RJ, Spiro JD, Lindquist RL, Kreutzer DL. Coexpression of interleukin-8 receptors in head and neck squamous cell carcinoma. Am J Surg 1997;174:507–12. [32] Horikawa T, Kaizaki Y, Kato H, Furukawa M, Yoshizaki T. Expression of interleukin-8 receptor A predicts poor outcome in patients with nasopharyngeal carcinoma. Laryngoscope 2005;115:62–7. [33] Yuan A, Yang PC, Yu CJ, Chen WJ, Lin FY, Kuo SH, et al. Interleukin8 messenger ribonucleic acid expression correlates with tumor progression, tumor angiogenesis, patient survival, and timing of relapse in non-small-cell lung cancer. Am J Respir Crit Care Med 2000;162: 1957–63. [34] Kido S, Kitadai Y, Hattori N, Haruma K, Kido T, Ohta M, et al. Interleukin 8 and vascular endothelial growth factor—prognostic factors in human gastric carcinomas? Eur J Cancer 2001;37:1482–7. [35] Konno H, Ohta M, Baba M, Suzuki S, Nakamura S. The role of circulating IL-8 and VEGF protein in the progression of gastric cancer. Cancer Sci 2003;94:735–40.
A single nucleotide polymorphism in the matrix metalloproteinase-1 and interleukin-8 gene promoter predicts poor prognosis in tongue cancer Yoshinori Shimizu *, Satoru Kondo, Akiko Shirai, Mitsuru Furukawa, Tomokazu Yoshizaki Department of Otolaryngology, School of Medicine, Kanazawa University, Takaramachi 13-1, Ishikawa 920-8640, Japan Received 7 March 2007; accepted 10 December 2007 Available online 13 February 2008
Abstract Objective: Matrix metalloproteinase-1 (MMP-1) and interleukin-8 (IL-8) play an important role in cancer development and metastasis. There is a single nucleotide polymorphism (SNP) located in the promoter region of MMP-1 and IL-8 that regulates gene expression. MMP-1 1607 2G/2G and IL-8 251 A/A genotypes enhance transcriptional activity and may be associated with increased risk in malignant tumors. We therefore evaluated the impact of these SNPs in tongue squamous cell carcinoma (SCC). Methods: In this study, we genotyped 69 tongue SCC patients. The expression of MMP-1 and IL-8 in tongue SCC patients was analyzed by immunohistochemistry. Results: We found a significant difference in IL-8 A/A genotypes with nodal recurrence (P = 0.0068). An analysis of disease-free survival rates showed that the presence of both MMP-1 2G/2G and IL-8 A/A genotypes was associated with a particularly poor prognosis (P = 0.0032) and was an independent prognostic factor (P = 0.001). The expression of MMP-1 was significantly correlated with the frequency of MMP-1 2G/2G genotypes (P = 0.049). Conclusion: These results suggest that SNP in the promoter region of MMP-1 and IL-8 plays an important role in tumor progression and recurrence through its expression in tongue SCC. # 2007 Elsevier Ireland Ltd. All rights reserved. Keywords: Single nucleotide polymorphism; Matrix metalloproteinase-1; Interleukin-8; Oral cancer; Immunohistochemistry
1. Introduction Malignant tumors have the ability to invade normal tissue and spread to distant sites, giving rise to metastasis, the major factors in the morbidity and mortality of cancer. The extracellular matrix acts as a structural support network within tissues and as a barrier to cell migration. Matrix degradation is mediated by the concerted action of several proteinases, including matrix metalloproteinase (MMPs), a family of enzymes [1]. Several recent studies have suggested * Corresponding author at: Division of Otolaryngology, Graduate School of Medicine, Kanazawa University, 13-1 Takara-machi, Kanazawa 9200934, Japan. Tel.: +81 76 265 2413; fax: +81 76 234 4265. E-mail address: [email protected] (Y. Shimizu).
new functional roles for MMPs in supporting tumor growth, modulating the extracellular matrix, regulating the availability of growth factors, and facilitating angiogenesis [2–4]. Matrix metalloproteinase-1 (MMP-1) has been implicated in tumor invasion and metastasis [5] as its specific ability to degrade type-I collagen, which is the most abundant substrate in the tumor surrounding stroma. Many reports have shown a significant negative correlation between the expression of MMP-1 and survival in advanced cancers [6–8]. The presence of a MMP-1 1G/2G polymorphism in the promoter region of the MMP-1 gene at 1607 bp (GenBank AF023338 2767) affects the transcriptional level of MMP-1 in both normal fibroblasts and in melanoma cells [9]. The MMP-1 2G polymorphism is positively correlated with an increased risk for developing lung cancer, ovarian cancer,
0385-8146/$ – see front matter # 2007 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.anl.2007.12.002
382
Y. Shimizu et al. / Auris Nasus Larynx 35 (2008) 381–389
colorectal cancer, and head and neck cancer [10–13]. These data establish a potential role for the 2G polymorphism in cancer development. Angiogenesis, or neovascularization, is the formation of new blood vessels that occurs during embryogenesis, wound healing, and tumor progression. In addition to the fact that angiogenesis is an indispensable phenomenon for the growth of tumors, the close association of angiogenesis with tumor metastasis has been well documented [14,15]. In recent studies, interleukin-8 (IL-8) has been shown to be a potent angiogenic factor implicated in tumor growth, metastasis and poor prognosis [16–19]. Promoter regions of a number of cytokine genes contain polymorphisms that directly influence cytokine production [20]. In the IL-8 251 (GenBank AF385628 3470) A/T polymorphism, the A allele tended to be associated with increased IL-8 production [21]. The frequency of IL-8 T/T genotypes was less than that of the IL-8 A/T or A/A genotype in prostate cancer patients [22]. We have been working on the regulation of invasion and metastasis in head and neck cancers, and revealed the important role of both MMPs and angiogenic factors. It is generally accepted that overexpression of such factors predicts a poor outcome in patients with various malignant tumors; however, studies on the expression level are always accompanied by the problem of reproducibility of the determined data. Analysis of genetic events, such as genotyping, provides quite stable data; however, the results may not be as impressive as the data from expression profiles. From this viewpoint, we evaluated the risk and prognostic value of MMP-1 1607 1G/2G and IL-8 251 A/ T polymorphisms in tongue squamous cell carcinoma (SCC). Moreover, we examined the expression of MMP-1 and IL-8 proteins by immunohistochemical examination, and the relationship between immunohistochemical examination and genotype data was also evaluated.
2. Materials and methods 2.1. Study subjects This study consisted of 69 patients with tongue SCC and 91 healthy controls, and all subjects were ethnic Japanese. The healthy controls consisted of staff, medical students, and inpatients of Kanazawa University Hospital without a previous diagnosis of malignant neoplasm. The data of tongue SCC patients in this study were obtained from medical records. From April 1983 to January 2000, 99 tongue SCC patients without metastasis were diagnosed and treated at the Department of Otolaryngology, Kanazawa University Hospital and Toyama City Hospital in Japan. Among them, 69 patients from whom DNA samples were available were recruited in this study. Clinical status was determined with the 1997 UICC/AJCC staging system [23]. Sixty-five patients were treated with surgery. Therapeutic
neck dissection was performed on 17 patients who had clinically positive nodes. Selective neck dissection was performed on 23 patients with clinically negative nodes who had T2 or more advanced tumor status. Neck dissection was not performed on 23 patients with stage I disease. Postoperative radiotherapy was administered to the neck of patients with pathologically positive lymph node metastasis. Although the tumor-free surgical margin was confirmed from frozen sections of the primary tumor, four patients were shown to have positive margins by using permanent pathological specimens. These patients received postoperative irradiation. Four patients underwent chemotherapy with or without radiation. The median follow-up time was 46 months (range, 4–96 months). Disease-free survival time was calculated from the date of treatment until the time of recurrence, defined as disease recurrence at the same site or the detection of metastases, including recurrence in the neck lymph nodes. 2.2. Sample preparation Surgical specimens were fixed with 10% formalin, embedded in paraffin, and the primary tumors were examined by immunohistochemical analysis. Paraffinembedded specimens were retrieved from the surgical pathology files of the Pathology Section of Kanazawa University Hospital and Toyama City Hospital in Japan. DNA samples of tongue SCC were extracted from paraffin blocks and those of healthy individuals from whole blood samples. This study design was approved by the Ethics Committee of Kanazawa University. 2.3. DNA extraction DNA from whole blood or paraffin-embedded tissue was extracted with a MagExtractor System MFX-2000 (TOYOBO, Osaka, Japan); DNA from paraffin-embedded tissue was also extracted with a modified protocol [24]. In brief, 10 mm 10 mm paraffin sections were incubated with 1 ml xylene at 55 8C for 15 min. After centrifugation, the supernatant fluid was discarded. The pellet was washed twice with 1 ml of 1/1 xylene/100% ethanol and twice with 100% ethanol. Pellets were incubated with 180 ml buffer ATL and 20 ml proteinase K at 55 8C for 48 h with intermittent vortexing, after which the manufacturer’s protocol was followed. 2.4. Genetic analysis Genotyping was accomplished using PCR followed by melting curve analysis with specific fluorescent hybridization probes in a Light Cycler System (Roshe, Mannheim, Germany) [25,26]. To detect MMP-1 1607 (GenBank AF023338 2767) 1G/ 2G genotypes, a 230-bp fragment was amplified by primer extension reaction (PCR) using forward and reverse primers,
Y. Shimizu et al. / Auris Nasus Larynx 35 (2008) 381–389
50 -GCA AGT GTT CTT TGG TCT CT-30 and 50 -CTT TCT GCG TCA AGA CTG-30 , respectively. The PCR profile to amplify a 230-bp product consisted of an initial melting step of 10 min at 95 8C, followed by 40 cycles of 10 s at 95 8C, 10 s at 60 8C, and 9 s at 72 8C. Following PCR, melting curve analysis was performed using a fluorescein-labeled mutation probe (50 -GAT AAG TCA TAT CTT TCT AAT TAT T-30 ) and Light Cycler Red 640-labeled anchor probe (50 -AAC TAC AAT TTC CTC ATC TAA GTG GCA TAA CAT GTG AGT TCA AGA GGG-30 ). For the IL-8 251 (GenBank AF385628 3470) A/T genotypes, the forward and reverse primers for PCR were 50 -CAT CCA TGA TCT TGT TC-30 and 50 -AAA TAC GGA GTA TGA CGA-30 , respectively. The PCR profile to amplify a 192-bp product consisted of an initial melting step of 10 min at 95 8C, followed by 40 cycles of 10 s at 95 8C, 15 s at 58 8C, and 8 s at 72 8C. Following PCR, melting curve analysis was performed using a fluorescein-labeled mutation probe (50 -GGT GAA TTA TCA AAT GTA TGC TTT TTT ATT TCT AGATAA C-30 ) and Light Cycler Red 640-labeled anchor probe (50 -TTA TCA TAG TAC CAT GTG ACA GAT ATA GTA CTA GAG TGG C-30 ). The genotypes of MMP-1 1607 1G/2G and IL-8 251 A/T revealed by Light Cycler System analysis were further confirmed with the ABI PRISM 310 genetic analyzer (Perkin-Elmer ABI, Foster City, CA). 2.5. Immunohistochemical analysis Immunohistochemical analysis was performed on paraffin-embedded 4-mm thick tissue sections as described previously [27]. The paraffin sections were dewaxed with xylene and rehydrated through graded concentrations of ethanol. Endogenous peroxidase was blocked with Endo-blocker (Biomeda, Foster City, CA) and nonspecific reactions were blocked with normal bovine serum (Wako, Osaka, Japan), then incubated with primary antibodies for 15 min at 45 8C. Primary antibodies against MMP-1 (41-1E5, 1:100 dilution) (Daiichi Fine Chemical, Takaoka, Japan) and IL-8 (6217, 1:20 dilution) (R&D Systems, Minneapolis, MN) were used. The sections were exposed to Universal Secondary Antibody (Research Genetics, Inc., Huntsville, AL). The sections were exposed to horseradish peroxidase-conjugated streptavidin, followed by diaminobenzidine as a chromogen. The nuclei were stained with methyl green. Skin tissue with epidermal ulcers was used as a positive control of MMP-1 protein expression and tonsil tissue for IL-8 protein expression. Negative controls were confirmed by staining with phosphatebuffered saline (PBS) instead of the primary antibody. 2.6. Evaluation of immunohistochemical staining for MMP-1 and IL-8 Stained samples of the primary tumors were reviewed as follows: using light microscopy, 2 areas that showed a high density of stained cells were selected in a 40 field, and then the number of stained cells of 200 cells was carefully
383
counted in the 2 areas in a 200 field without knowledge of the clinical data. The percentage of positive cell numbers obtained as an average of two counts was used as the expression score of MMP-1 and IL-8. For the purpose of correlation with SNPs, the expression scores of these proteins were arbitrarily divided into two categories: positive (MMP-1, positive cells 350% of tumor cells; IL-8, 330%) and negative (MMP-1, positive cells <50% of tumor cells; IL-8, <30%). 2.7. Statistical analysis Data were analyzed with a Macintosh personal computer (Apple Computer, Cupertino, CA) with Stat View software (Abacus Concepts, Berkeley, CA). Chi-square tests were used to compare the distribution of MMP-1 and IL-8 genotypes between cases and controls. Odds ratio and 95% confidence intervals were also calculated. Association of the MMP-1 and IL-8 genotype with clinicopathological parameters was evaluated by Fisher’s exact test. Statistical significance between the genotype and the survival rates was analyzed using Kaplan–Meier survival curves and the logrank test. The significance level was set at P < 0.05. Univariate and multivariate analyses were performed using the Cox proportional hazard regression model among the variables analyzed (age, gender, T, N, and M classifications, stage, and genotype). All variables other than age were dichotomized prior to data analysis, which was done to reduce the modeling flexibility and thereby yield a more robust final model. P-values <0.05 in the univariate models were included in multivariate analysis. Multivariate analysis was performed with backward elimination using a significance-level-to-stay of 0.05.
3. Results 3.1. Patient characteristics Patient characteristics are shown in Table 1. 3.2. Distribution of MMP-1 and IL-8 SNP Real-time PCR-assisted analysis of MMP-1 1G/2G and IL-8 A/T polymorphisms was carried out using the melting curves of the hybridization probes. Figs. 1 and 2A show the first differentiation of the fluorescence over time in MMP-1 and IL-8. The melting maximum peaks of MMP-1 1G/1G, 1G/2G, and 2G/2G genotypes were 48.3 8C, 48.3 8C and 54.4 8C, and 54.4 8C, respectively. The melting maximum peaks of the IL-8 A/A, A/T, and T/T genotypes were 62.0 8C, 62.0 8C and 64.6 8C, and 64.6 8C, respectively. The genotypes of MMP-1 and IL-8 were further confirmed by DNA sequencing. Figs. 1 and 2B show representative electropherograms of each genotype. A 10% random sample of the genotypes identified by the Light Cycler System was
384
Y. Shimizu et al. / Auris Nasus Larynx 35 (2008) 381–389
Table 1 Clinicopathological backgrounds of patients with tongue SCC
further evaluated by DNA sequencing, and 100% reproducibility was confirmed. Table 2 shows the distribution of MMP-1 1G/2G and IL-8 A/T SNP in tongue SCC patients and controls. The genotype distribution among controls was compatible with the Hardy– Weinberg equilibrium. No statistically significant difference in the frequency of MMP-1 or IL-8 genotypes was detected between patients and controls. 3.3. The relationship between genotypes and clinicopathological factors
The values are the numbers of patients. Tumor (T) and node (N); classification and stage were based on the International Union Against Cancer (UICC) Classification, 1997. SD: standard deviation.
The relationship between genotypes and the clinicopathological factors of 69 patients including age, sex, T classification, N classification, clinical stage, local recurrence, nodal recurrence, and distant metastatic recurrence were evaluated by Fisher’s exact test (Table 3). We found a significant difference in genotype distribution (IL-8 T/T or A/T vs. A/A) by nodal recurrence. Patients with the IL-8 A/A genotype had significantly more frequent nodal recurrence than those with the IL-8 T/T or A/T genotype. There was no significant association between the polymorphism and other clinicopathological factors.
Fig. 1. (A) Light Cycler System analysis of MMP-1 genotype. First differentiation of fluorescence over time. The melting maximum of the 2G-hybridization probe is 48.3 8C in the case of the 1G/1G genotype, 48.3 8C and 54.4 8C in the 1G/2G genotype, and 54.4 8C in the 2G/2G genotype. (B) Representative electropherogram of each type of polymorphism. Boxes indicate MMP-1 SNP, homozygous 1G/1G genotype with single G peak and clean sequence, heterozygous 1G/2G genotype with double G peak and uninterpretable sequence, homozygous 2G/2G genotype with double G peak and clean sequence.
Y. Shimizu et al. / Auris Nasus Larynx 35 (2008) 381–389
385
Fig. 2. (A) Light Cycler System analysis of IL-8 genotype. First differentiation of fluorescence over time. The melting maximum of the T-hybridization probe is 62.0 8C in the case of the A/A genotype, 62.0 8C and 64.6 8C in the A/T genotype, and 64.6 8C in the case of the T/T genotype. (B) Representative electropherogram of each type of polymorphism. Boxes indicate the IL-8 SNP, homozygous A/A genotype with single A peak, heterozygous A/T genotype with overlap A/T peak, homozygous T/T genotype with single T peak.
3.4. Immunohistochemical analysis The expression of MMP-1 and IL-8 was examined by immunohistochemical staining of the samples (Fig. 3). These signals were detected as dark brown staining. One of 69 paraffin sections was not available for this study because of insufficient residual tumor volume after DNA extraction. Thus, the immunohistochemical study of MMP-1 and IL-8 expression in the tumor was performed in the remaining 68 Table 2 Single nucleotide polymorphisms (SNP) in MMP1 and IL-8 promoter
P-values are based on the comparison of odds ratios in patients with 2G/2G vs. 1G/1G or 1G/2G genotypes for MMP-1 SNP, and those with A/A vs. T/T or A/T genotypes for IL-8 SNP. Odds ratio and 95% CI were calculated with 1G/1G or 1G/2G genotypes as a reference group in MMP-1 SNP, and T/T or A/T genotypes as a reference group in IL-8 SNP. CI: confidence interval.
cases. Table 4 shows the association of genotypes with the expression of MMP-1 and IL-8. In the MMP-1 stain, a high expression of MMP-1 was significantly associated with the MMP-1 2G/2G genotype. On the other hand, IL-8 expression had no significant correlation with the IL-8 genotype. The expression of MMP-1 and IL-8 had no significant relationship with survival (data not shown). 3.5. Association of MMP-1 and IL-8 genotypes with disease-free survival To evaluate the prognostic value of genotypes, the relationship between disease-free survival and MMP-1 and IL-8 genotypes was studied. Patients carrying neither the MMP-1 2G/2G nor the IL-8 A/A genotype had a significant trend toward decreased disease-free survival (Figs. 4 and 5); however, patients who carried both MMP-1 2G/2G and IL-8 A/A genotypes (MMP-1 + IL-8 high expression genotypes) had significantly shorter disease-free survival (Fig. 6). The short disease-free survival mainly reflected nodal recurrence. All four patients carrying both the MMP-1 2G/2G and IL-8 A/A genotypes had nodal recurrence, but did not have an advanced N stage (N0–1) at first diagnosis. Prognostic factors were analyzed by the Cox proportional hazard regression model among the variables (age, gender,
386
Y. Shimizu et al. / Auris Nasus Larynx 35 (2008) 381–389
Table 3 Relationship of MMP-1 and IL-8 genotypes with clinicopathological parameters
Table 4 Relationship of MMP-1 and IL-8 expression with SNP
*
Significant.
T, N, M classifications, stage, and genotypes). Among them, T4, advanced node (N2 and N3), stage IV, and the presence of both the MMP-1 2G/2G and IL-8 A/A genotypes were significant hazards in univariate analysis (Table 5). As T4 and advanced node (N2 and N3) are involved in stage IV, multivariate analysis was performed with T4, N2–3 and the presence of both MMP-1 2G/2G and IL-8 A/A genotypes. As shown in Table 6, the presence of both the MMP-1 2G/2G and IL-8 A/A genotypes remained an independent prognostic factor (P = 0.001; hazard ratio of other genotypes = 6.51). The category of N2–3 also has independent prognostic significance (P = 0.0001; hazard ratio = 8.74).
4. Discussion *
Significant.
Information on the role of MMPs and angiogenic factors in invasion and metastasis has been accumulating in terms of neoplastic cells and their surrounding microenvironment [2– 4]. Recent studies indicate a strong link between increased
Fig. 3. Immunohistochemical study of MMP-1 (A and B) and IL-8 (C and D) in cancer specimens (3,30 -diaminobenzidine tetrahydrochloride and methyl green stain, original magnification 200). MMP-1 2G/2G genotype (A) shows high expression in comparison with the 1G/2G genotype (B). IL-8 A/A genotype (C) shows high expression in comparison with the T/T genotype (D), although IL-8 expression had no significant correlation with the IL-8 genotype.
Y. Shimizu et al. / Auris Nasus Larynx 35 (2008) 381–389
Fig. 4. Kaplan–Meier analysis of disease-free survival according to the MMP-1 genotype (1G/1G + 1G/2G vs. 2G/2G). Statistical analysis was performed by the log-rank test.
MMP-1 expression, the presence of the 2G allele, and a poor clinical outcome [9–12]. Moreover, the association of IL-8 expression, A/A genotype, with poor outcome has also become a popular paradigm in the clinical cancer research setting [22]; however, these studies did not show a link based on genotype protein expression and clinical outcome. In our study, there was no significant difference in the frequency of either the MMP-1 or IL-8 genotype between tongue SCC and the controls. Although the distribution of MMP-1 1G/2G SNP in tongue SCC patients and controls had no statistically significant difference in the frequency of MMP-1, the Pvalue was 0.077, very close to the 0.05 that we set as the significance level in this study. A larger sample study may show a significant increased risk for tongue cancer. On the other hand, we found a significant correlation between the presence of both the MMP-1 2G/2G and IL-8 A/A genotypes with disease-free survival. Despite the small sample size, these data suggest that MMP-1 and IL-8 SNP are not strongly associated with the carcinogenesis of tongue SCC but may contribute to the progression of tongue SCC via the upregulation of MMP-1 and IL-8 gene expression. This
Fig. 5. Kaplan–Meier analysis of disease-free survival according to the IL-8 genotype (T/T + A/T vs. A/A). Statistical analysis was performed by the logrank test.
387
Fig. 6. Kaplan–Meier analysis of disease-free survival according to the presence of both the MMP-1 2G/2G and IL-8 A/A genotype vs. the combination of the MMP-1 and IL-8 genotype. MMP-1 + IL-8 high expression genotype: the presence of both MMP-1 2G/2G and IL-8 A/A genotypes. MMP-1 + IL-8 low expression genotype: the other combination of MMP-1 and IL-8 genotypes. Statistical analysis was performed by the log-rank test.
suggests that the MMP-1 2G/2G or IL-8 A/A genotype could not predict the prognosis of patients with tongue SCC, but this combination of two genotypes could have a critical role in malignant formation by matrix degradation and angiogenesis. Many reports showed a significant negative correlation between the expression of MMP-1 and survival in advanced cancers [6–8]. Previously, we reported that the MMP-1 2G/2G genotype predicted a poor prognosis in nasopharyngeal carcinoma [28]. Angiogenesis has been used as a prognostic indicator in a variety of cancers and is believed to be controlled by angiogenic factors, including IL-8. Many previous studies have shown a significant relationship between IL-8 and clinicopathological variables, including prognosis in malignant neoplasms [16–19]. Head and neck SCC patients exhibiting high levels of IL-8 were reported to have clinically more aggressive disease manifested by a higher TNM stage, more frequent Table 5 Univariate Cox regression analyses of variables in patients with tongue SCC
Factors whose P-values were >0.05 in univariate models were not included in multivariate analysis. MMP-1 2G/2G + IL-8 A/A represents the presence of both MMP-1 2G/2G and IL-8 A/A genotypes. y.o.: years old. CI: confidence interval. *Significant.
388
Y. Shimizu et al. / Auris Nasus Larynx 35 (2008) 381–389
Table 6 Multivariate Cox regression analyses of variables in 69 patients with tongue squamous cell carcinoma
Multivariate analysis was performed with backward elimination using a significance-level-to-stay of 0.05. MMP-1 2G/2G + IL-8 A/A represents the presence of both MMP-1 2G/2G and IL-8 A/A genotypes. CI: confidence interval. *Significant.
recurrence, and shorter disease-free interval [29]. Therefore, IL-8 SNP affecting the transcriptional level was thought to induce the overexpression of its protein and poor prognosis in malignant tumors. Moreover, two high-affinity IL-8 receptors, IL-8 receptor A and IL-8 receptor B, have been identified [30], and are reported to be present in malignant tumors, including head and neck squamous cell carcinoma [31]. The expression of IL-8 receptor A and B on tumor cells is thought to promote the metastatic ability of tumor cells by IL-8. Recently, we reported that the expression of IL-8 receptor A in nasopharyngeal carcinoma significantly correlated with a shorter overall survival rate [32]. IL-8 receptor A is specific to IL-8, whereas several other cytokines also bind to IL-8 receptor B; therefore, this result suggests that IL-8 promotes tumor growth as well as metastasis through its receptors. Patients with IL-8 A/A genotypes showed significantly more frequent nodal recurrence compared to those with either IL-8 T/T or IL-8 A/T, suggesting that IL-8 promotes lymph node metastasis in patients with tongue SCC. Recent studies of lymph angiogenesis revealed the contribution of angiogenic factors to lymphogenesis as well as angiogenesis. Overexpression of IL-8 mRNA was highly associated with non-small-cell lung cancer in advanced stages, distant lymph node metastasis, high tumor microvessel count, short survival, and early relapse [33]. In human gastric carcinoma, the IL-8 level in the neoplasms [34] and drainage veins [35] correlated significantly with lymphatic invasion. Thus, patients with IL-8 A/A genotypes, who are prone to the induction of more IL-8 [21], should be carefully treated, especially for nodal disease. The relationship between nodal recurrence and the type of treatment, including neck dissection, radiotherapy and chemotherapy, were evaluated by Fisher’s exact test. We found no significant difference in the therapy by nodal recurrence. We analyzed the disease-free survival of each type of treatment. The factors among treatments were analyzed by the Cox proportional hazard regression model among the variables (neck dissection, radiotherapy, chemotherapy), none of which were significant hazards in multivariate analysis (P = 0.29, 0.78 and 0.1, respectively). Thus, we thought that the type of treatment did not affect nodal recurrence and disease-free survival in this study.
In our study, the high expression of MMP-1 protein in tongue SCC tissue correlated significantly with the MMP-1 2G/2G genotype. Rutter et al. [9] described using transient transfections in A2058 melanoma cells, in which the MMP-1 2G polymorphism led to a more than 20-fold increase of mRNA expression in melanoma cells, as well as 5–10-fold in normal fibroblasts. In ovarian cancers, tumors carrying 2G alleles have seven times higher expression levels of MMP-1 than tumors carrying no 2G alleles [11]. As MMP-1 protein levels mirror MMP-1 mRNA expression, the increase in MMP-1 transcription may increase protein expression. On the other hand, IL-8 expression had no significant correlation with IL-8 genotypes, which may reflect that the DNA sequence around 251 SNP is less influential than other cis-acting transcription regulatory elements in the promoter of IL-8. Alternatively, similar studies with larger sample sizes may result in a statistically significant correlation of IL-8 genotypes with IL-8 protein expression. Multivariate Cox regression analysis demonstrated that advanced nodal status (N2 and N3) and the presence of both the MMP-1 2G/2G and IL-8 A/A genotypes were independent prognostic factors. On the other hand, the expression of MMP-1 and IL-8 proteins had no significant relationship with survival. This suggests that the MMP-1 2G/2G and IL-8 A/A genotypes could predict the prognosis of patients with tongue SCC more precisely than protein expression. It is generally accepted that overexpression of MMP-1 and IL-8 predicts a poor outcome in patients with various malignant tumors [6–8,16–19] and there was no evidence of an increased expression of MMP-1 and IL-8 in this study; however, the quantitative study of gene expression, especially in immunohistochemical studies, is always associated with the problem of reproducibility and quantitative results. Furthermore, the expression of gene products is easily influenced by tumor-surrounding conditions compared with the DNA sequence. Therefore, the analysis of genetic events, such as genotyping, provides quite stable data, even though it may not be as impressive as the data from expression profiles. This drawback can be overcome by using a combination of several genotype analyses, as shown in this study. Patients with the IL-8 A/A genotype had more frequent nodal recurrence than those with either IL-8 T/T or IL-8 A/T in this study. In addition, we demonstrated that the presence of both the MMP-1 2G/2G and IL-8 A/A genotypes were independent prognostic factors and had a poor outcome in disease-free survival. The IL-8 A/A genotypes could be a risk factor for nodal recurrence; thus, patients with IL-8 A/A genotypes, especially with MMP-1 2G/2G, should be carefully treated for nodal disease, including selective neck dissection or the addition of adjuvant chemotherapy for T12N0 stage, and should be frequently followed for nodal recurrence. These therapeutic strategies could improve the prognosis of patients with IL-8 A/A genotypes without increasing morbidity in patients with IL-8 A/T or T/T genotypes.
Y. Shimizu et al. / Auris Nasus Larynx 35 (2008) 381–389
These genotypes can be confirmed in 2 days by the Light Cycler System using DNA from the whole blood of patients. Therefore, analysis of genotypes can provide quite stable data than data from expression profiles such as immunohistochemical studies, and is suitable for the treatment of patients with tongue SCC. Although it is not conclusive because of the rather small sample size, SNPs in the promoter region of MMP-1 and IL8 can be independent prognostic factors, and thus, the combination of this information will be beneficial to predict the outcome of patients with tongue SCC.
References [1] Brinckerhoff CE, Rutter JL, Benbow U. Interstitial collagenases as markers of tumor progression. Clin Cancer Res 2000;6:4823–30. [2] Basset P, Okada A, Chenard MP, Kannan R, Stoll I, Anglard P, et al. Matrix metalloproteinases as stromal effectors of human carcinoma progression: therapeutic implications. Matrix Biol 1997;15:535–41. [3] Nelson AR, Fingleton B, Rothenberg ML, Matrisian LM. Matrix metalloproteinases: biologic activity and clinical implications. J Clin Oncol 2000;18:1135–49. [4] Stetler-Stevenson WG. Matrix metalloproteinases in angiogenesis: a moving target for therapeutic intervention. J Clin Invest 1999;103: 1237–41. [5] Rigel DS, Friedman RJ, Kopf AW, Silverman MK. Factors influencing survival in melanoma. Dermatol Clin 1991;9:631–42. [6] Murray GI, Duncan ME, O’Neil P, Melvin WT, Fothergill JE. Matrix metalloproteinase-1 is associated with poor prognosis in colorectal cancer. Nat Med 1996;2:461–2. [7] Murray GI, Duncan ME, O’Neil P, McKay JA, Melvin WT, Fothergill JE. Matrix metalloproteinase-1 is associated with poor prognosis in oesophageal cancer. J Pathol 1998;185:256–61. [8] Ito T, Ito M, Shiozawa J, Naito S, Kanematsu T, Sekine I. Expression of the MMP-1 in human pancreatic carcinoma: relationship with prognostic factor. Mod Pathol 1999;12:669–74. [9] Rutter JL, Mitchell TI, Buttice` G, Meyers J, Gusella JF, Ozelius LJ, et al. A single nucleotide polymorphism in the matrix metalloproteinase-1 promoter creates an Ets binding site and augments transcription. Cancer Res 1998;58:5321–5. [10] Zhu Y, Spitz MR, Lei L, Mills GB, Wu X. A single nucleotide polymorphism in the matrix metalloproteinase-1 promoter enhances lung cancer susceptibility. Cancer Res 2001;61:7825–9. [11] Kanamori Y, Matsushima M, Minaguchi T, Kobayashi K, Sagae S, Kudo R, et al. Correlation between expression of the matrix metalloproteinase-1 gene in ovarian cancers and an insertion/deletion polymorphism in its promoter region. Cancer Res 1999;59:4225–7. [12] Ghilardi G, Biondi ML, Mangoni J, Leviti S, DeMonti M, Guagnellini E, et al. Matrix metalloproteinase-1 promoter polymorphism 1G/2G is correlated with colorectal cancer invasiveness. Clin Cancer Res 2001;7:2344–6. [13] Hashimoto T, Uchida K, Okayama N, Imate Y, Suehiro Y, Hamanaka Y, et al. Association of matrix metalloproteinase (MMP)-1 promoter polymorphism with head and neck squamous cell carcinoma. Cancer Lett 2004;211:19–24. [14] Weidner N, Semple JP, Welch WR, Folkman J. Tumor angiogenesis and metastasis-correlation in invasive breast carcinoma. N Engl J Med 1991;324:1–8. [15] Carmeliet P, Jain RK. Angiogenesis in cancer and other diseases. Nature 2000;407:249–57. [16] Ugurel S, Rappl G, Tilgen W, Reinhold U. Increased serum concentration of angiogenic factors in malignant melanoma patients correlates with tumor progression and survival. J Clin Oncol 2001;19:577–83.
389
[17] Ren Y, Poon RT, Tsui HT, Chen WH, Li Z, Lau C, et al. Interleukin-8 serum levels in patients with hepatocellular carcinoma: correlations with clinicopathological features and prognosis. Clin Cancer Res 2003;9:5996–6001. [18] Kassim SK, El-Salahy EM, Fayed ST, Helal SA, Helal T, Azzam EelD, et al. Vascular endothelial growth factor and interleukin-8 are associated with poor prognosis in epithelial ovarian cancer patients. Clin Biochem 2004;37:363–9. [19] Benoy IH, Salgado R, Van Dam P, Geboers K, Van Marck E, Scharpe´ S, et al. Increased serum interleukin-8 in patients with early and metastatic breast cancer correlates with early dissemination and survival. Clin Cancer Res 2004;10:7157–62. [20] Bidwell J, Keen L, Gallagher G, Kimberly R, Huizinga T, McDermott MF, et al. Cytokine gene polymorphism in human disease: on-line databases, supplement 1. Genes Immun 2001;2:61–70. [21] Hull J, Thomson A, Kwiatkowski D. Association of respiratory syncytial virus bronchiolitis with the interleukin 8 gene region in UK families. Thorax 2000;55:1023–7. [22] McCarron SL, Edwards S, Evans PR, Gibbs R, Dearnaley DP, Dowe A, et al. Influence of cytokine gene polymorphisms on the development of prostate cancer. Cancer Res 2002;62:3369–72. [23] Sobin LH, Wittekind C, editors. TNM classification of malignant tumors. 5th ed., New York, NY: Wiley-Liss; 1997. p. 4–8. [24] Wickham CL, Boyce M, Joyner MV, Sarsfield P, Wilkins BS, Jones DB, et al. Amplification of PCR products in excess of 600 base pairs using DNA extracted from decalcified, paraffin wax embedded bone marrow trephine biopsies. Mol Pathol 2000;53:19–23. [25] Wittwer CT, Ririe KM, Andrew RV, David DA, Gundry RA, Balis UJ. The LightCycler: a microvolume multisample fluorimeter with rapid temperature control. Biotechniques 1997;22:176–81. [26] Nojiri T, Morita H, Imai Y, Maemura K, Ohno M, Ogasawara K, et al. Genetic variations of matrix metalloproteinase-1 and -3 promoter regions and their associations with susceptibility to myocardial infarction in Japanese. Int J Cardiol 2003;92:181–6. [27] Yoshizaki T, Sato H, Maruyama Y, Murono S, Furukawa M, Park CS, et al. Increased expression of membrane type 1-matrix metalloproteinase in head and neck carcinoma. Cancer 1997;79: 139–44. [28] Kondo S, Wakisaka N, Schell MJ, Horikawa T, Sheen TS, Sato H, et al. Epstein-Barr virus latent membrane protein 1 induces the matrix metalloproteinase-1 promoter via an Ets binding site formed by a single nucleotide polymorphism: enhanced susceptibility to nasopharyngeal carcinoma. Int J Cancer 2005;115:368–76. [29] Eisma RJ, Spiro JD, Kreutzer DL. Role of angiogenic factors: coexpression of interleukin-8 and vascular endothelial growth factor in patients with head and neck squamous carcinoma. Laryngoscope 1999;109:687–93. [30] Holmes WE, Lee J, Kuang WJ, Rice GC, Wood WI. Structure and functional expression of a human interleukin-8 receptor. Science 1991;253:1278–80. [31] Richards BL, Eisma RJ, Spiro JD, Lindquist RL, Kreutzer DL. Coexpression of interleukin-8 receptors in head and neck squamous cell carcinoma. Am J Surg 1997;174:507–12. [32] Horikawa T, Kaizaki Y, Kato H, Furukawa M, Yoshizaki T. Expression of interleukin-8 receptor A predicts poor outcome in patients with nasopharyngeal carcinoma. Laryngoscope 2005;115:62–7. [33] Yuan A, Yang PC, Yu CJ, Chen WJ, Lin FY, Kuo SH, et al. Interleukin8 messenger ribonucleic acid expression correlates with tumor progression, tumor angiogenesis, patient survival, and timing of relapse in non-small-cell lung cancer. Am J Respir Crit Care Med 2000;162: 1957–63. [34] Kido S, Kitadai Y, Hattori N, Haruma K, Kido T, Ohta M, et al. Interleukin 8 and vascular endothelial growth factor—prognostic factors in human gastric carcinomas? Eur J Cancer 2001;37:1482–7. [35] Konno H, Ohta M, Baba M, Suzuki S, Nakamura S. The role of circulating IL-8 and VEGF protein in the progression of gastric cancer. Cancer Sci 2003;94:735–40.