Functional Polymorphisms of Matrix Metalloproteinase-9 Are Associated with Risk of Occurrence and Metastasis of Lung Cancer

The extracellular microenvironment is a dynamic entity and provides regulatory signals on an intricate network of pathways that include cell adhesion, differentiation, division, and apoptosis. Therefore, cells with disruption of these pathways may acquire tumorigenic properties, such as loss of contact inhibition, aberrant cell division, and evasion of apoptosis (1). Matrix metalloproteinases (MMP), a family of structurally and functionally related zinc endopeptidases, are essential regulators of the microenvironment of the cell, through their control of extracellular proteolysis, and have been implicated in invasion and metastasis of tumor cells (2–5). Recent studies on the roles of MMPs revealed that these proteolytic enzymes are involved in the regulation of various cell behaviors, including cancer cell growth, differentiation, apoptosis, migration, invasion, and the regulation of tumor angiogenesis and immune surveillance (6); thus, MMPs play an important role in cancer development and aggression.

MMP-9, a member of the MMP family also known as gelatinase B or 92 kDa type IV collagenase, is the major structural component of basement membrane (7) but is highly expressed in a wide variety of human cancers, including lung cancer (8–13). Animal studies showed that MMP-9 overexpression contributes to cancer development and progression. For example, cancer cells were less capable of colonizing the lungs of MMP-9–deficient mice than the lungs of wild-type mice, and MMP-9 null mice develop fewer cancers than wild-type mice (14, 15). Notably, in two transgenic models of tumor progression, the K14-HPV16 skin cancer model (16) and the RIP1-Tag2 insulinoma model (17), cancer cell proliferation was decreased in tumors from MMP-9–deficient mice compared with wild-type mice, indicating that MMP-9 generated growth-promoting signals. Furthermore, studies documented that MMP-9 was also involved in other steps of cancer development, including decreasing cancer cell apoptotic potential (17), promoting angiogenesis (17–19), and regulating immune responses to cancer (16, 20–22). Therefore, a better understanding of the diverse role of MMP-9, particularly its contribution to early cancer development and tumor aggression, is needed.

Single-nucleotide polymorphisms (SNP) in the coding region, especially nonsynonymous SNPs, may influence the protein activity and therefore may be associated with cancer development and metastasis. We used the public SNP database (http://pga.gs.washington.edu) and literature searches (23, 24) to select nonsynonymous coding SNPs in the MMP-9 gene, and we identified three nonsynonymous SNPs, R279Q, P574R, and R668Q, with a rare allele frequency of >0.05. As shown in Fig. 1, R279Q (exon 6, base G to A) is located in the gelatinase specific fibronectin type II domains, which presumably enhance substrate binding (25, 26); P574R (exon 10, base C to G) and R668Q (exon 12, base G to A) are located in the hemopexin domain, which is thought to affect both substrate and inhibitor binding (27, 28). Therefore, these polymorphisms potentially alter protein structure of MMP-9 and may have some functional relevance and affect an individual's susceptibility to cancers.

The aim of the present case-control study was to investigate whether genetic variations of the MMP-9 gene may constitute markers for lung cancer risk and metastasis. We focused on these three polymorphisms (R279Q, P574R, and R688Q) that cause amino acid changes of the MMP-9 enzyme, and evaluated the associations between the genotypes and haplotypes and risk of lung cancer development and metastasis.

Study population. This study included 744 lung cancer patients and 747 cancer-free controls. All subjects were genetically unrelated ethnic Han Chinese and were from Nanjing City and surrounding regions. Patients were newly diagnosed incident lung cancer cases and were recruited between July 2002 and August 2004 at the Cancer Hospital of Jiangsu province (Nanjing) and the First Affiliated Hospital of Nanjing Medical University, Nanjing, China, without any restrictions. All histopathologically confirmed lung cancer cases in these two hospitals during the study period were recruited and the response rate of the patients was 91.8% (744 of 810). The exclusion criteria included previous cancer, other metastasized cancer, and previous radiotherapy or chemotherapy. Of the 744 patients, 494 (66.4%) underwent surgical resection and had detailed clinical stage classification and metastasic data on the basis of the postoperative histopathological examination (29). Cancer-free controls were randomly selected from 10,500 individuals who participated in a community-based screening program for noninfectious diseases conducted in Jiangsu province during the same time period as the cases were recruited, with a response rate of 90.8%. These control subjects had no history of cancer and were frequency matched to the cases in age (±5 years), sex, and residential area. Each subject was scheduled for a face-to-face interview after informed consent was obtained, and a structured questionnaire was administered by interviewers to collect information on demographic data and environmental exposures including tobacco smoking. After interview, ∼5 mL of venous blood sample were collected from each subject. The study was approved by the Institutional Review Board of Nanjing Medical University.

MMP-9 genotyping. Genomic DNA was extracted from the leukocyte pellet obtained from each blood sample buffy coat by centrifugation of 5 mL whole blood. The MMP-9 R279Q was detected using the PCR-RFLP assay. The primers used to detect this polymorphism were 5′-GAGAGATGGGATGAACTG-3′ (forward) and 5′-GTGGTGGAAATGTGGTGT-3′ (reverse), which generated a 439 bp fragment containing the R279Q site in a 20 μL PCR mixture consisted of ∼50 ng of genomic DNA, 12.5 pmol of each primer, 0.1 mmol/L each deoxynucleotide triphosphate, 1× PCR buffer (50 mmol/L KCl, 10 mmol/L Tris-HCl, and 0.1 % Triton X-100), 1.5 mmol/L MgCl₂, and 1.0 unit of Taq polymerase. The PCR product was then digested with the restriction enzyme MspI (New England BioLabs, Beverly, MA) and separated on a 3% agarose gel. The 279R allele has two restriction sites and produces three fragments of 187, 129, and 123 bp, and the 279Q allele has only one restriction site, resulting in two fragments of 252 and 187 bp (Fig. 2A).

The MMP-9 P574R and R668Q polymorphisms were detected using a primer-introduced restriction analysis-PCR assay (30). For the P574R polymorphism, the sense primer was introduced a mismatched T to replace G at −2 bp from the polymorphic site to create a BamHI restriction site. The primers used to detect this polymorphism were 5′-GCTGGACTCGGTCTTTGAGGATC-3′ (forward) and 5′-TTGAGCCTCCTTGACTGATGGG-3′ (reverse). Similarly, a mismatched T was introduced in the sense primer to replace C at −2 bp from the R668Q site and a TaqI restriction site was then created. The primers were 5′-ACACGCACGACGTCTTCCAGTATC-3 (forward) and 5′-GGGGCATTTGTTTCCATTTCCA-3′ (reverse). The 112 bp (P574R) and 138 bp (R668Q) PCR products were then digested by restriction enzymes of BamHI (for P574R) and TaqI (for R668Q; New England BioLabs), respectively. The rare (574R) allele produces a single 112 bp fragment and the common (574P) allele produces two fragments of 93 and 19 bp. Similarly, the rare (668Q) allele produces a single 138 bp fragment, whereas the common (668R) allele produces two fragments of 115 and 23 bp (Fig. 2B and C).

Genotyping was done without knowing the subjects' case and control status and the same number of cases and controls were assayed in each 96-well PCR plate with a positive control of a DNA sample with known heterozygous genotype. Two research assistants independently read the gel pictures and did the repeated assays if they did not reach a consensus on the tested genotype. In addition, 10% of the samples were randomly selected to perform the repeated assays, and the results were 100% concordant. To further confirm our genotyping results by the primer-introduced restriction analysis-PCR assay, PCR products of the P574R loci with different genotypes were randomly selected for direct sequencing using an automated sequencer (ABI model 377 genetic analysis; Perkin-Elmer Applied Biosystems, Foster City, CA).

Statistical analysis. Differences in select demographic variables, smoking status, pack-years smoked, and frequencies of the MMP-9 genotypes, alleles, and haplotypes between the cases and the controls were evaluated by using the χ² test. The associations between MMP-9 variants and lung cancer risk were estimated by computing the odds ratios (OR) and 95% confidence intervals (95% CI) from both univariate and multivariate logistic regression analyses with adjustment for age, sex, and pack-years of smoking. We used the PHASE 2.0 program (31) to infer haplotypes based on the known MMP-9 genotypes. Stratified analyses were also done using age, sex, smoking status, and histologic types. Those who had smoked <100 cigarettes in their lifetime were defined as nonsmokers; otherwise, they were considered as smokers. Pack-years smoked [(cigarettes per day ÷ 20) × years smoked] were calculated to indicate the cumulative smoking dose. All statistical analyses were done with Statistical Analysis System software (v.8.0e; SAS Institute, Cary, NC).

Selected characteristics of the 744 lung cancer patients and 747 controls are summarized in Table 1. There were no significant differences in terms of distributions on age and sex between the cases and the controls (P = 0.49 and 0.96, respectively), suggesting that our frequency matching of the demographic characteristics was satisfactory. About 35.3% of the lung cancer cases smoked more than 30 pack-years, which was significantly higher than that of the controls (18.7%; P < 0.0001). Of the total 744 lung cancer cases, 281 (37.8%) were adenocarcinoma, 261 (35.1%) were squamous cell carcinoma, 48 (6.5%) were small cell carcinoma, and 154 (20.7%) were large-cell, mixed-cell, or undifferentiated carcinomas. Four-hundred ninety-four cases (66.4%) who had undergone surgery had detailed metastasis information and therefore were classified into two groups: 290 patients (58.7%) with metastases and 204 (41.3%) without.

The allele and genotype distributions of the MMP-9 R279Q, P574R, and R668Q polymorphisms in the cases and the controls are shown in Table 2. The observed genotype frequencies for these three polymorphisms were all in Hardy-Weinberg equilibrium in the controls (P = 0.70, 0.73, and 0.14, respectively). In the linkage disequilibrium analyses, we found that the R668Q locus was not in linkage disequilibrium with either the R279Q locus (R² = 0.03, D′ = 0.60) or the P574R locus (R² = 0.02, D′ = 0.58). However, the R279Q locus was in linkage disequilibrium with the P574R locus (R² = 0.66, D′ = 0.87). The frequencies of the 279R, 574P, and 668Q alleles were 0.70, 0.75, and 0.13 in the cases, and 0.68, 0.70, and 0.13 in the controls, respectively, and differences between the cases and the controls were statistically significant for the 574P allele (P = 0.01) but not for the 279R and 668Q alleles (P = 0.22 and 0.80, respectively). Compared with the 279QQ genotype, nonsignificantly elevated risks were associated with the 279RQ or 279RR genotypes [OR (95% CI) being 1.29 (0.89-1.89) and 1.36 (0.93-1.97), respectively], and the risks were more pronounced among the patients with metastasis [OR (95% CI) being 1.68 (0.97-2.92) and 1.79 (1.03-3.08), respectively; Table 2]. For the MMP-9 P574R polymorphism, the genotype frequencies were 8.6% (RR), 42.3% (PR), and 49.1% (PP) in controls and 5.5% (RR), 40.1% (PR), and 54.4% (PP) in the cases, and the difference was also statistically significant (P = 0.02). Logistic regression analysis revealed that subjects carrying the 574PR heterozygote had a 1.46-fold increased risk of lung cancer with borderline significance (95% CI, 0.94-2.26) and those carrying the 574PP homozygote had a 1.69-fold significantly elevated risk (95% CI, 1.10-2.60) compared with the 574RR genotype. When we examined the combined effects of these two MMP-9 variants and used 279R and 574P as the risk alleles, a significantly increased risk of lung cancer was associated with both the genotypes containing “1 to 2 risk alleles” (adjusted OR, 2.16; 95% CI, 1.30-3.59) and containing “>2 risk alleles” (adjusted OR, 2.44; 95% CI, 1.48-4.03). In addition, this significant association was more evident among the 290 lung cancer patients with metastasis [adjusted OR of 2.30 (95% CI, 1.09-4.85) for the subjects carrying 1 to 2 risk alleles, and adjusted OR of 2.82 (95% CI, 1.35-5.88) for those carrying >2 risk alleles], compared with those without any risk alleles (Table 2). However, no overall significant associations were observed between the R668Q polymorphism and lung cancer risk in this study population.

Because the MMP-9 R279Q and P574R polymorphisms were in linkage disequilibrium, we then combined these two loci and did the haplotype analyses using the PHASE 2.0 program. There were a total of four possible haplotypes derived from their genotypes, and the haplotype distribution between the cases and the controls was statistically different (P = 0.0001; Table 3). Compared with the haplotype 279Q-574R, the other three haplotypes were associated with a significant increased risk of lung cancer [OR, 1.26 (95% CI, 1.06-1.50) for 279R-574P; OR, 1.93 (95% CI, 1.41-2.65) for 279Q-574P; and OR, 1.71 (95% CI, 1.11-2.64) for 279R-574R]. Furthermore, the elevated risks associated with the 279R-574P and 279R-574R haplotypes were more evident among the patients with metastasis [OR of 1.35 (95% CI, 1.07-1.71) for 279R-574P, and OR of 2.02 (95% CI, 1.17-3.49) for 279R-574R], which mainly resulted from the effect of the 279R allele.

The dichotomized genotypes of combined MMP-9 R279Q and P574R were further examined for subgroups by selective variables. However, there were no significant differences in terms of the associations between combined MMP-9 R279Q and P574R genotypes and lung cancer risk for individuals stratified by age, gender, smoking status, and histologic types (data not shown).

In this case-control study, we investigated the associations of three nonsynonymous SNPs of the MMP-9 gene with risk of lung cancer in a population in southeast China. We found that the MMP-9 574PR/PP genotypes were associated with a significantly increased risk of lung cancer in a dose-response manner compared with the 574RR genotype. We also observed a significant elevated risk of lung cancer metastasis associated with the MMP-9 279RR genotype in an allele-dose manner. To the best of our knowledge, this is the first large molecular epidemiologic study that has investigated the association of MMP-9 variants and risk of cancer development and progression.

Studies in genetically modified animals showed that the high level of constitutive expression of MMP-9 was associated with susceptibility to tumor formation (6) and tumor progression (16). Genetic alterations that may alter the functions of the MMP-9 gene are thought to contribute to cancer susceptibility and metastatic potential. The MMP-9 P574R and R668Q variants are located in the hemopexin-like domain. Removal of this domain in the collagenases eliminates their characteristic capability to cleave triple-helical collagen, but does not significantly affect hydrolytic activity toward gelatin, casein, or synthetic substrates (28). Furthermore, the hemopexin-like domains affecting inhibitor binding and the surface patch of positively charged residues are noteworthy, which might be involved in binding to the negatively charged COOH-terminal tail of tissue inhibitor of metalloproteinase (32). This specific tissue inhibitor of metalloproteinase tail and hemopexin-like domain interaction is important for formation of the MMP-tissue inhibitor of metalloproteinase progelatinase complex implicated in progelatinase activation (33, 34). Although the reported amino acid substitutions in these functional domains have not been fully understood, it is possible that the substitution of a hydrophobic 574P amino acid (proline) to a positive charged 574R (arginine) may influence the inhibitor binding and progelatinase activation (35), which is consistent with our observation of an association between this P574R polymorphism and lung cancer risk.

Recently, Rebbeck et al. (36) emphasized the importance of assessing functional relevance of genetic variants in association studies, and the lack of reproducibility of many association studies might reflect the involvement of genetic variants with no functional significance (37, 38). Several computational algorithms have been recommended to predict the effect of the amino acid substitutions of the genes on protein structures, expressions, and/or functions (36). Among them, PolyPhen was developed to identify functionally important SNPs by predicting whether an amino acid substitution is likely to be deleterious for the protein on the basis of three-dimensional structure and multiple alignments of homologous sequences (39). Previously, Zhu et al. (37) applied PolyPhen to 166 molecular epidemiologic studies to examine the correlation between position-specific independent counts (PSIC) score and the OR associated with a particular nonsynonymous SNP. The authors found a significant inverse correlation between the ORs and PSIC score difference, and between tolerance index and PSIC score difference (37), indicating that using an algorithm, such as PolyPhen, to assess the functional significance of nonsynonymous SNP is helpful for SNP selection and evaluation and therefore can optimize molecular epidemiologic association studies. In the present study, we also used PolyPhen to infer the functional relevance of the three MMP-9 SNPs and found that only P574R may possibly be damaging, which is consistent with the findings from the current molecular epidemiology study that the MMP-9 574P allele was associated with a significantly increased risk of lung cancer. Therefore, this finding is biologically plausible.

Our results also indicate that individuals carrying the 279R allele of the MMP-9 gene may be predisposed to metastasis of lung cancer. Although currently there is no experimental data on this R279Q polymorphism, the amino acid substitution site is located in the catalytic domain of the MMP-9 gene, particularly in the fibronectin type II domains that play important roles in substrate binding (25, 26). The possible mechanism of this finding may be that the conversion from the positively charged amino acid arginine to uncharged glutamine has a potential effect on the activity of this enzyme. Further studies are warranted to examine this possibility.

Historically, MMPs were considered to play an important role in cancer invasion and metastasis because of their capability of degrading extracellular matrix and basement membrane barriers (4, 5). However, studies have also shown that MMPs are also involved in multiple steps of cancer development (reviewed in ref. 6). Recently, more molecular epidemiologic studies investigated the associations between genetic polymorphisms of MMPs and cancer susceptibility and prognosis, including lung (40, 41), colorectal (42), esophageal (43), and cervical (44) cancers. In a case-control study of 243 non–small cell lung cancer patients and 350 control subjects from North China, Fang et al. (40) reported that a MMP-3 promoter polymorphism may modify susceptibility to non–small cell lung cancer, and the MMP-1 1G-MMP-3 5A haplotype may predicate the risk of lymphatic metastasis of lung cancer. However, in a large case-control study of 1,752 Caucasian lung cancer patients and 1,363 healthy controls, Su et al. (41) reported that genotypes containing the 2G allele of the MMP-1 were associated with increased risk of lung cancer in never-smokers and in male subjects. In addition, Zinzindohoue et al. (42) showed that colorectal cancer patients with the 2G/2G genotype of the MMP-1 had a significantly worse clinical outcome, such as cancer invasion, metastasis, and prognosis. For the MMP-9 polymorphisms, no published articles were available to date, either for its association with cancer occurrence or prognosis. It is biologically plausible that multiple functional SNPs of MMPs in the same pathway may work together in cancer development and progression. Therefore, further larger, well-designed studies with simultaneous genotyping of multiple polymorphisms of MMPs family are warranted to elucidate the combined roles of MMPs SNPs in carcinogenesis.

Like all other case-control studies, inherited biases may lead to spurious findings. Because our study was a hospital-based case-control study and the cases were from hospitals and the controls were from the surrounding community the hospitals served, the study subjects may not be representative of the general population. However, we believe that our results are unlikely to be attributable to selection bias because we used a relatively large number of incident cases. In addition, except for tobacco smoking, other factors, such as occupational exposure and certain dietary components, might interact with the MMP-9 genotype or act as potential confounders in the analysis. Unfortunately, information on these factors in our case-control study was not available. It would be interesting to investigate the interaction between the MMP-9 genotypes and these risk factors in future studies.

In conclusion, our study provides for the first time the evidence that genetic variants R279Q and P574R of the MMP-9 gene may influence the development and progression of lung cancer. Our results are consistent with data from in vitro studies and from animal experiments that MMP-9 plays an important role in cell migration and proliferation, an important process in cancer development and progression. Our finding is an important addition to previously published work on MMPs and cancer susceptibility but needs to be further validated in functional studies of these two genetic variants.