The 2G allele of promoter region of Matrix metalloproteinase-1 as an essential pre-condition for the early onset of oral squamous cell carcinoma

Matrix metalloproteinase (MMP)-1 (Collagenase-1) is a major proteinase of the MMP family that specifically degrades type I collagen, which is a major component of the extracellular matrix (ECM), as well as other fibrillar collagens of types II, III, V and IX [1, 2]. MMP-3 (Stromelysine-1) is responsible for degradation of type IV collagen, which forms the basement membrane, and collagen V, IX, X [3]. MMP-3 also has a role in activation of proMMP-1 in tumor tissue into the active form of MMP-1 [4]. Gene expression data in our previous report demonstrated enhanced expression of MMP family genes in OSCC tissues, and suggested correlation of high expression levels of MMP-1 and MMP-3 with aggressive behavior, such as metastasis, and clinical prognosis [5]. Similar correlations have been reported in pharyngeal, colon and various other tumor types [6‐8]. Colocalization of MMP-1 and MMP-3 with destruction of ECM in the invasive front of cancer tissue suggests a direct role in cancer invasion [9].

It has also been shown that a function of MMPs affects susceptibility to different kinds of carcinoma. To date, polymorphisms of the promoter domain have been described in MMP-1, -3, -9, -12 to influence the expression level of the genes [10]. A 2G type of single nucleotide polymorphism (SNP) at -1607 bp site in the promoter domain of MMP-1 creates a sequence, 5'-GGA-3,' that is the core recognition sequence of the binding site for Ets family transcription factors. The 2G type promoter results in higher transcription activity of the MMP-1 gene than does a 1G type promoter [11]. A 5A type promoter at -1171 bp site of MMP-3 is also known to have a twofold higher transcriptional activity in vitro than that of a 6A type [12]. It has been documented that the 2G type SNP of MMP-1 confers increased susceptibility to colorectal [13], ovarian [14], lung [15], endometrial [16], renal cell [17] and head and neck [18] cancers; and the 5A type SNP of MMP-3 is associated with an increased susceptibility to breast cancer [19].

An in situ hybridization study revealed that MMP-1 expression in normal oral mucosa is controlled at a low level, whereas a remarkably elevated expression level is observed in cases of oral epithelial dysplasia, which becomes even higher in cases of OSCC [20]. It has also been documented that cases of oral epithelial dysplasia showing high expression level of MMP-1 developed into OSCC at higher frequency than cases with low expression of MMP-1 [21]. Taking all the observations noted above into consideration, it seems likely that MMPs may serve as key factors in all stages of the OSCC progression from carcinogenesis in the early precancerous condition to the advanced invasive and metastatic phases.

The purpose of this study is to document the effect of genomic polymorphisms of MMP genes in the development of OSCC. We compared genotype distribution in the promoter domains of the functional SNPs that influence the transcriptional activity of MMP-1 and MMP-3 between OSCC patients and healthy control groups. Multivariate analysis was effectively used to assess correlations among parameters such as OSCC, the genotypes, age and sex. In this report, we describe the role of functional SNP of the MMP-1 gene in susceptibility to OSCC and, as a remarkable finding; we discuss the crucial impact of the MMP-1 2G allele in the development of OSCC in younger individuals.

In this study, we examined SNPs in the promoter regions of MMP-1 and MMP-3 genes in relation to OSCC risk with case-control analyses. Significant differences, i.e., high frequency of 2G/2G genotype and decreased frequency of 1G/1G genotype, were observed in the MMP-1 genotype distribution in OSCC cases (p = 0.034, Table 1). Similarly, in comparisons of allele frequency a significant correlation was detected between onset of OSCC and genotypes that carry the MMP-1 promoter 2G allele in homo- or hetero-form (p = 0.016, OR = 2.39, 95%CI = 1.21–4.72). MMP-1 with the 2G type promoter caused a higher expression level of MMP-1 in tissues [11]. The biological mechanism of the elevated risk of OSCC as related to the 2G allele of MMP-1 promoter includes genetic instability, which is accompanied by the cell cycle activation caused by a sequence of events led by increased MMP-1 enzyme activity, i.e. activation of tissue disruption, reconstruction and resultant release of ECM binding growth factors. The relationship between MMP-1 promoter polymorphism and risk of OSCC has been investigated by Lin et al. in a group of 121 cases [22] and by Cao et al. in a group of 96 cases [23] and by Vairaktaris et al. in a group of 156 cases [24]. Although both Lin et al. and Cao et al.reported an elevation of OSCC risk correlated with the MMP-1 2G allele, further studies were needed to reach a conclusion mainly because the numbers of cases in the previous studies were relatively small, and there were differences between these reports in background analyses on correlation with environmental risk factors, such as smoking and areca use. In this study, significant correlations were found between the presence of the MMP-1 promoter 2G allele and increased OSCC risk among 170 Japanese OSCC patients. These results are consistent with existing reports on malignant neoplasms in general. It is notable that a new finding on the molecular function of MMP-1 in the onset of OSCC was obtained from the characteristic pattern of the MMP-1 promoter genotype distribution in younger patients.

Interestingly, Vairaktaris et al. reported an increased risk of oral cancer with MMP-1 1G/2G polymorphism [24]. The discrepancy of the results of the Chinese studies and our report on one hand and the European (German and Greek) study on the other hand may be explained by the diverse ethnic background of the different studied populations. Another point of difference is the fact that their patient sample included patients with positive family history of thrombophilia.

The observed bias of the average age by the MMP-1 promoter genotype among the 170 OSCC cases of this study suggested an influence of the MMP-1 promoter 2G allele on the age of OSCC onset. The scatter plot of age on the MMP-1 promoter genotypes also revealed disappearance of the 1G/1G-genotype distribution among the OSCC cases under 45 years old. In agreement with this observation, the ROC analysis suggested that the age of 45 years is the borderline age above which a change occurs in the incidence of OSCC. Although the specific reasons for these results have not been clarified, they suggest that some kind of biological conditions related to carcinogenesis have initiated around the age of 45. Carcinogenesis is caused by synergistic effects of various factors [25, 26]. The multivariate logistic regression analysis showed that aging and presence of the MMP-1 2G allele (i.e., 2G/2G or 1G/2G) are independently involved in onset of OSCC. In fact, it was clarified that the significance of the presence of the MMP-1 2G allele was lower in the subject group 45 years old or older (p = 0.238, OR = 1.65, 95% CI = 0.75–3.55), and that the impact of the allele was highly significant in the younger group, which was under 45 years old (p = 0.014, OR = 9.67, 95% CI = 1.20–78.15) (data not shown).

As concrete causes of OSCC, mainly tobacco smoking and alcohol usage and, in some cases, viral infections have attracted attention as environmental causative factors [25, 27]. It has been suggested that tobacco smoking and alcohol usage alone may not explain the mechanism of the entire early onset of OSCC, because the duration of exposure to these risk factors in young OSCC patients is shorter than in the older group, and also because there are some cases with stable incidence that have no known risk factors [28]. The association with smoking and habitual drinking has been supported as the risk factors of OSCC, which generally shows a higher incidence in older age, but other investigators believe that the mechanism of the early onset OSCC is fundamentally different from that of elderly onset OSCC [25]. Cao et al. considered the relationship with smoking as a behavioral risk factor and the MMP-1 2G allele as a genetic risk factor in the OSCC group [23]. In lung cancer, it has been similarly shown that a genetic factor of the MMP-1 promoter 2G allele increases the risk of cancer that occurs in a tobacco-usage-dependent manner only in cases with a history of tobacco use. These reports support the role of MMP-1 promoter polymorphism as an endogenous background factor, in contrast to exogenous environmental factors [15]. However, no evidence to support a notion of synergistic interactions between risk factors, such as habitual smoking and drinking and MMP-1 promoter polymorphism in OSCC cases, was shown in this study. Further investigations with regard to frequency and duration of exposure are required to consider interactions between environmental factors and multiple genetic background factors.

Examples are known of malignant neoplasms caused by genetic factors, for example, mutations in RB in Retinoblastoma [29], p53 in Li-Fraumeni Syndrome [30], and APC in familial adenomatous polyposis [31]. There are also reports of mutations in BRCA1 and BRCA2, which affect familial breast cancer occurrence, though their penetrance is low [32]. These are neoplasms that are all caused by loss-of-function mutations in molecules functionally classified as tumor-suppressor genes, and they are characterized by familial, juvenile or young onset and may be multicentric or bilateral. Although incidents are rare in the population, they are examples of germline mutations that directly involve heterofamilial tumorigenesis.

By contrast, the MMP-1 promoter 2G allele is a genetic polymorphism that exists at the high allele frequency of 80% to 90% in the general population. It is possible to call this a hereditary trait that is shared among races or humans in general. Other genetic factors that are classified into similar polymorphism as MMP-1 have also been reported, such as Cyclin D1 (CCND1) in head and neck squamous cell carcinoma (HNSCC) and colorectal cancer [33, 34], xeroderma pigmentosum complementary group D (XPD), DNA damage binding protein 2 (DDB2), and MMP-9 in lung cancer [35‐37]. Similar to the effects of mutations in tumor suppressor genes, diversity in gene expression level and molecular structure caused by genetic polymorphisms affect monitoring and repair mechanism of DNA replication, as well as control of the cell cycle, ultimately resulting in genetic instability. The common mechanism of carcinogenesis caused by those genetic factors is an elevated carcinogenic risk due to this genetic instability.

As examples of genetic polymorphisms that affect onset age of cancers other than that examined in this study, some reports discuss MMP-1 and DDB2 in lung cancer [15, 36], and CCND1 in HNSCC and colorectal cancer [33, 34]. These genetic factors, on the other hand, are present in patients with cancer as well as in the general population at certain frequencies. Therefore, it is hard to conclude that these genetic factors are the primary factors in early onset carcinogenesis, unlike germline single-gene mutations in Rb and APC that induce juvenile or early onset of cancers with high penetrance. Aging brings exposure to various carcinogenic factors and is associated with inevitable accumulation of genetic and epigenetic modifications of genes. These accumulations interact synergistically with biological background factors on the host side, and consequently cancer may develop [26, 38]. It seems that inherited genetic factors play a larger role in development of early onset cancer that has undergone a relatively short duration of exposure to carcinogen through environment and lifestyle habits. Therefore, it is conceivable that the impact of these genetic factors would appear more directly in younger than in older cases, although we still expect a synergistic involvement of various environmental and genetic factors. Nevertheless, because of the disappearance ofthe MMP-1 1G allele distribution observed in the OSCC cases with a clear boundary under the age of 45 years, the MMP-1 promoter 2G allele should be recognized as an essential genetic precondition for the development of early onset OSCC.

It is said that an onset age of 45 years and younger accounts for about 6% of OSCC cases [27]. It should be noted that in this study we found that the tongue was the primary site in 25 cases (86.2%) of 29 early onset OSCC cases in individuals under 45 years old. This primary site was obviously different from that in patients 45 years old and older in whom we noted 57 cases of tongue cancer (40.4%) out of 141 OSCC cases. Similar tendencies have been reported in previous studies [39, 40]. These findings suggest the possibility that each primary tumor sites in the oral cavity may be associated with differences in cancer susceptibility, specific carcinogenic stimulation, or genetic background. However, no environmental or genetic factors specific to a particular site of the oral mucosa have been identified for OSCC. It is conceivable that it would be important to carry out further investigations on genetic and environmental risk factors for each anatomical site of the oral regions, though we are not able to discuss this aspect of OSCC because the number of cases was inadequate in this study.

The number of cases in this study was relatively small; therefore, the conclusions may still be contingent upon confirmation in a larger study. To evaluate the conclusive impact of MMP-1 on OSCC risk, a sufficiently large number of cases or consistency among results from multiple age adjusted studies would be needed. Furthermore, although the MMP-1 promoter SNP-based determination of OSCC susceptibility has the potential to become a useful clinical tool, it is necessary to add a greater diversity of information for practical use. In the future, improvements in diagnosis of susceptibility and clinical application could become possible by incorporating multiple genetic and environmental risk factors into the prediction model for cancer occurrence. Information about the risk of various types of malignant neoplasm as well as OSCC is useful for prevention and early detection of cancer or for monitoring of postoperative recurrence. A diagnostic system for evaluating cancer risk is vital not only for improvement in treatment techniques but also as a theoretical foundation to build strategies for cancer prevention.