Clinical Cancer Research Molecular Diagnostics in Cancer Therapeutic Development: Fulfilling the Promise of Personalized Medicine Infection and Cancer: Biology, Therapeutics, and Prevention
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Cancer Research Clinical Cancer Research
Cancer Epidemiology Biomarkers & Prevention Molecular Cancer Therapeutics
Molecular Cancer Research Cancer Prevention Research
Cancer Prevention Journals Portal Cancer Reviews Online
Annual Meeting Education Book Meeting Abstracts Online
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Sun, T.
Right arrow Articles by Lin, D.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Sun, T.
Right arrow Articles by Lin, D.
Clinical Cancer Research Vol. 12, 7009-7017, December 1, 2006
© 2006 American Association for Cancer Research

Imaging, Diagnosis, Prognosis

Haplotypes in Matrix Metalloproteinase Gene Cluster on Chromosome 11q22 Contribute to the Risk of Lung Cancer Development and Progression

Tong Sun1, Yang Gao3, Wen Tan1, Sufang Ma3, Xuemei Zhang1, Yonggang Wang2, Qingrun Zhang3, Yongli Guo1, Dan Zhao1, Changqing Zeng3 and Dongxin Lin1

Authors' Affiliations: Departments of 1 Etiology and Carcinogenesis and 2 Thoracic Surgery, Cancer Institute and Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College; 3 Beijing Genomics Institute, Chinese Academy of Sciences, Beijing, China

Requests for reprints: Dongxin Lin, Department of Etiology and Carcinogenesis, Cancer Institute and Hospital, Chinese Academy of Medical Sciences, Beijing 100021, China. E-mail: dlin{at}public.bta.net.cn or Changqing Zeng, Beijing Genomics Institute, Chinese Academy of Sciences, Beijing 101300, China. E-mail: czeng{at}genomics.org.cn.


   Abstract
Top
 Abstract
Materials and Methods
 Results
 Discussion
 References
 
Purpose: Matrix metalloproteinases (MMP) play important roles in cancer development and single nucleotide polymorphisms (SNP) in some MMP genes were shown to confer susceptibility to certain cancers. This study examined the association between genotypes and haplotypes in the MMP1-MMP3-MMP12 gene cluster and risk of lung cancer development and metastasis.

Experimental Design: A two-stage investigation was conducted. First, 35 SNPs covering these genes were selected and validated in 190 patients and 190 controls. Twenty-two validated SNPs were then analyzed in an entire case-control panel consisting of 711 patients and 716 controls. Associations with the risk of lung cancer were estimated by logistic regression.

Results: The investigated MMP gene region could be partitioned into two major haplotype blocks. One common haplotype in the block composed of major part of MMP1 transcription region was significantly associated with increased risk for the development [odds ratio (OR), 1.35; 95% confidence interval (95% CI), 1.11-1.63; P = 0.01; permutated P = 0.134] and distant metastasis of lung cancer (ORs for stage IV versus stages I-III, 1.67; 95% CI, 1.12-2.50; P = 0.009; permutated P = 0.048) and the other showed a protective effect against metastasis (ORs for stage IV versus stages I-III, 0.22; 95% CI, 0.07-0.62; P = 0.001; permutated P = 0.011). Another common haplotype in the block across MMP3 was significantly associated with decreased risk for developing lung cancer (OR, 0.71; 95% CI, 0.59-0.86; P = 0.003; permutated P = 0.027).

Conclusions: The observed multiple cancer-associated genetic variants suggested that the MMP1-MMP3-MMP12 gene cluster plays important roles in lung cancer development and progression.

Discovery and application of biomarkers that incorporate with traditional cancer diagnosis, staging, and prognosis could largely help to improve early diagnosis and patient care (1). With the completion of human genome project, millions of single nucleotide polymorphisms (SNP) have been identified, which are thought to be attractive biomarkers in cancer risk assessment, screening, staging, or grading (2). However, the application of individual SNPs has been limited thus far because they are of low penetrance and the effect of risk alleles is relatively difficult to identify (2, 3). To date, most studies focus on "functional" SNPs, but the number of SNPs with clear function is limited. Therefore, how to incorporate SNPs in studies of cancer predisposition and prognosis and how to find out the true association are still challenging tasks (2–5). Recently, haplotype-based association study has been proposed as a powerful and comprehensive approach to identify causal genetic variation underlying complex diseases (6, 7).

Lung cancer is the leading cause of cancer-related death all over the world. In many countries, including China, the incidence and mortality rates of lung cancer have increased rapidly in recent years. Despite significant advances have been made in diagnosis and treatment in the last decades, the prognosis of lung cancer remains rather poor, with a 5-year overall survival rate <10% (8). Metastatic disease eliminates possibility of surgical cure of lung cancer. Unlike some other cancer, lung cancer lacks specific biomarkers for early detection and prognosis determination at diagnosis, although many efforts have been made to discover potential biomarkers for lung cancer risk assessment and clinical outcome prediction (9–11).

Matrix metalloproteinases (MMP), a family of proteases degrading extracellular matrix and basement membrane barriers, are not only involved in multiple steps of cancer development but also play important roles in cancer metastasis (12). Molecular epidemiologic studies have shown associations between genetic polymorphisms in MMPs and cancer susceptibility or prognosis, including cancers of the lung (13–17), esophagus (18, 19), colorectum (20–22), and cervix (23, 24). However, these association studies were limited to few SNPs or constructed haplotypes from two or three polymorphic sites. A cluster of eight MMP genes, including MMP20, MMP27, MMP8, MMP10, MMP1, MMP3, MMP12, and MMP13, is defined on chromosome 11q 22.3 (25). The aberration of this chromosome region has been associated with risk of primary lung cancer and its metastatic disease (26, 27). Specifically, the expression or activation of MMP1, MMP3, and MMP12 in this region showed significant correlation with advanced stages of lung cancer and poor patient survival (28–31). MMP1 and MMP12 also play critical role in smoke-induced lung injury (32, 33). In addition, functional SNPs in MMP1 and MMP3 have been associated with susceptibility to lung cancer (13, 14, 16). These findings warrant more powerful and comprehensive studies to explore the relationship between genetic variations in MMPs and lung cancer.

In this study, we examined linkage equilibrium (LD) and haplotype structure of the genomic region across MMP1-MMP3-MMP12 loci on chromosome 11q22 and assessed the roles of genotypes and haplotypes in risk for the development and metastasis of lung cancer.


   Materials and Methods
Top
 Abstract
 Materials and Methods
Results
 Discussion
 References
 
Subject selection. This study consisted of 711 primary lung cancer patients and 716 controls and all subjects were ethnical Han Chinese. The subject characteristics have been described previously (15, 17). Briefly, eligible patients were consecutively recruited between January 1997 and November 2001 at the Cancer Hospital, Chinese Academy of Medical Sciences (Beijing, China). Because we examined the relationship between MMP polymorphisms and lung cancer staging, surgically resectable patients (91%) were mostly recruited, which allowed us to obtain accurate tumor-node-metastasis data and histologic classification at the time of diagnosis. The response rate for patients was 93%. The exclusion criteria included previous cancer, metastasized cancer from other organs, and previous radiotherapy or chemotherapy. The pathologic stages were determined according to American Joint Committee on Cancer staging system (34) and tumor grade was classified into low (well differentiated), intermediate (moderately differentiated), or high grade (poorly differentiated) according to the WHO grade classification (35). Controls were cancer-free individuals selected from a nutritional survey consisted of 2,500 individuals, which was conducted in the same region during the same period as patients were collected. The selection criteria included no individual history of cancer and frequency matching to patients on sex and age (±5 years). The participation response rate for controls was 83%. Of the 770 patients and 777 controls who participated in the previous study (15), only 711 patients and 716 controls were successfully genotyped in this study because DNA samples of the rest of the subjects were no longer available. At recruitment, written informed consent was obtained from each subject and each participant was then interviewed to collect detailed information on demographic characteristics and lifetime history of tobacco use. Subjects were considered current smokers or ex-smokers if they smoked up to 1 year before the date of cancer diagnosis or if they smoked up to 1 year before the date of the interview for control subjects, otherwise were defined as nonsmokers. Information was collected on the number of cigarettes smoked daily, the age at which the subjects started smoking, and the age at which ex-smokers stopped smoking. Because only 24 patients and 38 controls were ex-smokers, they were combined with current smokers for analysis. This study was approved by the Institutional Review Board of the Chinese Academy of Medical Sciences Cancer Institute.

Selection of candidate SNPs. SNPs across the 84.5-kb region spanning MMP1-MMP3-MMP12 loci on chromosome 11q22, from 0.1 kb upstream of MMP12 transcriptional region to 7.6 kb downstream of MMP1 transcriptional region, were surveyed in SNP database4 and Celera Discovery System5. We also referred to International HapMap Project (reference) for genotyped SNPs in Han Chinese population6. SNPs with a minor allele frequency ≥5% were selected and those located in coding regions were included as many as possible. We selected SNPs every 1 to 3 kb across these gene loci to ensure a high density of markers and to provide adequate characterization of haplotype diversity within previously defined LD blocks. SNPs were selected in an iterative manner until reaching four to seven common SNPs (frequency, ≥5%) per LD block. The distance between adjacent markers was <5 kb, except for the MMP12 locus where the last two intervals between markers were 9.2 and 11.9 kb, respectively. Specifically, we selected 10 SNPs in the MMP1 locus (8.2-kb transcript region), 9 SNPs in the MMP3 locus (7.8-kb transcript region), and 2 SNPs in the MMP12 locus (12.3-kb transcript region). In addition, 14 SNPs located beyond the boundaries of these gene loci were also selected, totaling up to 35 SNPs as candidate markers. These SNP markers cover an 84.5-kb region in chromosome 11q22 with the average resolution of one SNP per 2.4 kb.

SNP analysis and validation. SNPs were typed by the MassARRAY system (Sequenom, San Diego, CA) as described (36). To ensure the typing quality, multiple positive and negative samples were incorporated into every genotyping plate and genotyping data of all duplicate samples were consistent. The laboratory persons were blinded to sample arrangement during the process. In our two-stage study design, the first stage was to validate above described SNPs in our study population. All the selected 35 SNPs were first typed among 190 controls and 190 patients. Consequently, we removed 10 SNPs that either were monomorphic or had the minor allele frequency <2% in our study population and 3 SNPs where the successful rate for genotyping was <90% in these 380 DNA samples. Therefore, a total of 22 SNPs with the minor allele frequency ≥2% and call rate >90% were used for the second-stage study. The observed genotype distributions of these 22 SNPs did not differ from those expected from Hardy-Weinberg equilibrium in controls.

LD block determination and haplotype construction. LD between the 22 SNPs used in haplotype analysis was measured by pairwise D' statistic. The structure of LD block was examined using the method of Gabriel et al. (37), using the 80% confidence bounds of D' to define sites of historical recombination between SNPs. Block structure was assessed using SNPs with the minor allele frequency ≥5%. Haplotypes were constructed from genotype data in the full-size case-control panel within blocks by using an accelerated expectation-maximization algorithm method with Haploview 3.2 software (38). Briefly, this method creates highly accurate population frequency estimates of the phased haplotypes based on the maximum likelihood as determined from unphased input (39). Haplotypes in the two LD blocks were reconstructed by either all genotyped loci or haplotype-tagging SNPs (htSNP). The haplotype frequencies instead of individual haplotype phase among controls, total patients, patients with different tumor stages, or patients with other clinical features were estimated separately for comparison.

Selection of htSNPs. The htSNPs were selected with Haploview 3.2 software on a block-by-block basis using the method described by Carlson et al. (40) with the sample size inflation factor Rh2 ≥ 0.8. We also calculated the multivariate squared correlation, Rs2, an index exploiting the multivariate correlation between measured and unmeasured SNPs in a region of high LD. The Rs2 values for the two main haplotype block were 0.980 and 0.972, respectively, indicating that our selection of htSNPs provided a good prediction of other unmeasured SNPs and an optimal prediction of haplotypes in the cl0uster of the MMP genes.

Statistical analysis. Two-sided {chi}2 test was done to compare differences in allele frequencies of each locus between cases and controls. Logistic regression was used to analyze the association between a single locus and lung cancer risk, adjusted for sex, age, and smoking status. These statistical analyses were done with Statistical Analysis System software (version 8.0; SAS Institute, Cary, NC) and Haploview 3.2 software (38) separately. To assess significance, we did both Bonferroni correction (41) and permutation procedure (100 and 1,000 tests, respectively) to correct the P value of single-locus association results. The haplotype analyses were done when its frequency was >5% in both all-loci-constructed and htSNP-constructed haplotype estimates. We also used Haplo.stats (42, 43) to assess the relative effects of haplotypes and adjusted for sex, age, and smoking status.


   Results
Top
 Abstract
 Materials and Methods
 Results
Discussion
 References
 
Subject characteristics. As shown in Table 1 , there was no statistically significant difference in the distribution of sex and age between patients and controls. However, significantly more smokers were present among patients compared with controls (63.4% versus 52.0%; P < 0.001). In addition, patients had a higher value of pack-years than controls; 30.4% of smokers among patients smoked >41 pack-years compared with 19.9% among controls (P < 0.001). Among patients, 279 (39.3%) were classified as squamous cell carcinoma, 217 (30.5%) as adenocarcinoma, 29 (4.1%) as large cell carcinoma, 84 (11.8%) as small cell carcinoma, and 60 (8.5%) as other types, including bronchioalveolar carcinoma (n = 12) and undifferentiated cancer (n = 8). Forty-two patients (5.8%) had histologic type–unknown lung cancer. Of the 711 patients, 210 (29.5%) had stage I lung cancer, 157 (22.1%) had stage II lung cancer, 212 (29.8%) had stage III lung cancer, 71 (10.0%) had stage IV lung cancer, and 61 (8.6%) had stage-unknown disease. About tumor grade, 79 (11.1%) patients were classified into well-differentiated lung cancer (low grade) and 341 (48.0%) and 213 (29.9%) were classified into moderately differentiated (intermediate grade) or poorly differentiated (high grade) lung cancer, respectively, whereas data for the rest of the 78 patients (11.0%) were unavailable.


View this table:
[in this window]
[in a new window]

  		Table 1. Distributions of selected characteristics by case-control status

 
LD and haplotype structure of the MMP1-MMP3-MMP12 genomic region. We assembled a high-density SNP map across the MMP1-MMP3-MMP12 genomic region on 11q22 to determine LD block and haplotype structure (Fig. 1A ); 35 SNPs were selected using an iterative strategy (see Materials and Methods) and the average distance between SNPs across the 84.5-kb region was 2.4 kb. We determined this MMP gene cluster region to contain two LD blocks (Fig. 1B). Block1 (SNPs 1-4) covered 5.0 kb, spanning exons 5 to 10 and 0.2-kb 3'-untranscriptional region of MMP1; block 2 (SNPs 7-19) spanned 35.2 kb, encompassing the whole locus, 0.3 kb upstream, and 27.7 kb downstream region of MMP3. The distance between two blocks was 12.7 kb. There was a 31.0-kb region around the MMP12 locus that could not be included into the blocks. The LD plot for all samples is shown in Fig. 1B. We observed that within each block, the haplotype diversity was low in our study population. For block 1, only four common haplotypes were observed, which could be distinguished by three htSNPs and represented 99.9% subjects. Block 2 is much larger than block 1, but it also contained only four common haplotypes that represented 97.8% subjects and could be distinguished by three htSNPs.


Figure 1
View larger version (49K):
[in this window]
[in a new window]

  		Fig. 1. SNPs in the region of MMP1-MMP3-MMP12 gene cluster located in chromosome 11q22. A, MMP1-MMP3-MMP12 gene structure. White boxes, coding regions; black boxes, untranslated regions; arrows, transcription start sites. bullet, rs numbers denied in stage II study because the successful genotyping rate was <90% in stage I study; {blacktriangledown}, rs numbers that are htSNPs; *, rs numbers that are loci where the minor allele frequencies were <10%. B, diagram of block structure of MMP1-MMP3-MMP12 in chromosome 11q22 generated by using Haploview. LD plots were identified by strong LD. Depth of gray color, computed pairwise D'; deeper gray, higher D' value. The selected htSNPs and estimated haplotype frequencies in two major haplotype blocks of MMP1-MMP3-MMP12 gene cluster were shown. Marker numbers along with a tick beneath above haplotypes are htSNPs. The frequency of each haplotype within a block is to the right of the haplotype. The thickness of the lines connecting the haplotypes across blocks represents the relative frequency [i.e., high (thick) versus low (thin) with which a given haplotype is associated with the haplotype in the other block].

 
Associations between individual SNPs and lung cancer risk. The allelic frequencies of 22 second-stage SNPs in the three MMP gene loci among patients and controls are shown in Table 2 . Among the 22 SNPs, only 5 had allelic frequency that differed significantly between patients and controls. In the MMP1 locus, risk allelic frequencies of rs7125062, rs2075847, and rs470206 were higher in patients than in controls (P = 0.035, 0.037, and 0.038, respectively), with the odds ratio (OR) being 1.22 [95% confidence interval (95% CI), 1.00-1.51], 1.16 (95% CI, 0.86-1.57), or 1.21 (95% CI, 0.95-1.54), respectively. The allelic frequencies of both rs522616 and rs586701, which are located in the promoter region of MMP3, seemed to be significantly higher in patients than in controls (P = 0.013 and 0.003). Subjects carrying at least one risk allele of corresponding polymorphisms had an OR of 1.21 (95% CI, 0.99-1.51) or 1.57 (95% CI, 0.74-3.37). However, after Bonferroni correction or permutation test, all P values for these differences between patients and controls increased beyond the significant level of 0.05, despite that the rs586701 had a permutated P value of 0.051 after 1,000 tests.


View this table:
[in this window]
[in a new window]

  		Table 2. Allele frequencies of 22 second-stage SNPs in three MMP gene loci among 711 patients and 716 controls

 
Associations between haplotypes and lung cancer risk. We examined the difference in frequency distribution of all common haplotypes between patients and controls (Table 3 ) and found a significant haplotype effect of block 2 (P = 0.0008) but not block 1 (P = 0.080) on lung cancer risk (Table 3). Within block 2, which spans the whole transcript region of MMP3, we observed two haplotypes (2a and 2c) that differentially distributed in patients and controls. Haplotype 2a seemed to be more prevalent among patients than among controls (40% versus 35%; P = 0.010 and turned into P = 0.140 after 1,000 permutation tests), whereas haplotype 2c, differing from haplotype 2a only in the maker of rs522616, was significantly less prevalent among patients compared with controls (16% versus 23%; P = 0.000049 and turned into P = 0.004 after 1,000 permutation tests). Multivariate logistic regression analysis showed that haplotype 2c carriers were at a decreased risk of developing lung cancer compared with noncarriers (OR, 0.68; 95% CI, 0.56-0.83). In block 1, haplotype 1c had a frequency that was significantly different between patients and controls (21% versus 17%; P = 0.01 and turned into P = 0.134 after 1,000 permutation tests). Subjects with haplotype 1c had an increased risk of developing lung cancer (OR, 1.28; 95% CI, 1.06-1.56) compared with those without haplotype 1c. Because limited interblock recombination may result in long-range LD, we also evaluated whether there was long-range haplotype composed of a subset of the common block-specific haplotypes in block 1 and block 2 and the potential effect of the whole candidate region. Unfortunately, we did not find any common haplotypes that were more strongly associated with lung cancer, most likely due to a relatively high recombination rate between these two blocks because the D' value between them was 0.13.


View this table:
[in this window]
[in a new window]

  		Table 3. Haplotype frequencies of MMP1-MMP3-MMP12 loci in 711 patients and 716 controls

 
Associations between haplotypes and lung cancer disease status. We additionally evaluated whether there was an association between SNPs in this MMP gene cluster and lung cancer disease status at the time of diagnosis (Table 4 ). The increased risk for distant metastasis of lung cancer seemed to be associated with haplotype 1c in block 1. The estimated frequency of haplotype 1c among patients with distant metastasis of lung cancer (stage IV) was 30%, which was significantly higher than that among patients with stage III (23%), stage II (21%), or stage I (18%) lung cancer (P = 0.009; Ptrend = 0.042). Haplotype 1c carriers had an OR of 1.66 (95% CI, 1.11-2.48) for developing distant metastasis of lung cancer compared with noncarriers. In contrast with haplotype 1c, haplotype 1d showed a protective effect against lung cancer progression. Among stage IV patients, the estimated frequency of haplotype 1d was 3%, which was significantly lower than those among stage I to III patients (10-13%; P = 0.001). Patients carrying haplotype 1d had an OR of 0.22 (95% CI, 0.07-0.62) for developing distant metastasis of lung cancer compared with noncarriers. However, we did not observe any significant association between common haplotypes in block 2 and lung cancer disease status.


View this table:
[in this window]
[in a new window]

  		Table 4. Frequency distributions of constructed haplotypes in 711 patients with different disease stage at the time of diagnosis

 
Associations between tagging SNP-constructed haplotypes and lung cancer risk. To explore the minimum SNPs that might represent the cancer-related haplotypes, the associations between SNPs in the MMP genes and risk of lung cancer development and metastasis were investigated using htSNP-constructed haplotypes. We found that the results obtained by using haplotypes reconstructed with the three htSNPs in each LD block (Table 5 ) were very similar to those obtained by using haplotypes constructed with all second-stage SNPs (Tables 3 and 4). Specifically, the frequencies of the haplotypes constructed by htSNPs or all SNPs among patients, controls, and patients with different disease status were almost the same and the haplotypes 1c and 1d in block 1 and haplotype 2c in block 2 constructed with htSNPs remained to be significantly associated with lung cancer development or metastasis, respectively. These results indicate that the htSNPs represented well all SNPs and may serve as simple markers.


View this table:
[in this window]
[in a new window]

  		Table 5. Frequency distributions of tagging SNP-constructed haplotypes in 711 patients and 716 controls and different disease stages

 
Additional analysis with stratification of lung cancer subtypes using both all SNPs and htSNP-constructed haplotypes revealed that the major histologic types did not exhibit heterogeneity in their relation to the genes studied (data not shown). We also examined the associations of these SNPs and haplotypes with lung cancer grade but the results were negative (data not shown).


   Discussion
Top
 Abstract
 Materials and Methods
 Results
 Discussion
References
 
In this study, we selected 35 SNPs in the region across MMP1-MMP3-MMP12 genes on chromosome 11q22.3 and examined their association with the development and disease status of lung cancer in a Chinese population. Two haplotypes (1c and 1d) in the LD block composed of major part of MMP1 were identified to be significantly associated with risk of lung cancer development and advanced disease status. We also found a haplotype (2c) in the LD block across MMP3 that was strongly associated with decreased risk of lung cancer. We showed that as few as three htSNPs in each LD block could well represent all SNPs to construct haplotypes that link to the disease. To our knowledge, this is the first fine-mapping association study to investigate the effect of inherited variations in this genomic region on lung cancer development and progression.

Previous studies showed that MMP1 –1607 1G>2G (rs1799750) and MMP3 –1171 5A>6A (rs3025058) SNPs are associated with risk of developing lung cancer (13, 14, 16). However, the results are conflicting about which allele is risk allele. Zhu et al. (13) reported an increased risk associated with the MMP1 2G allele, but Su et al. (16) reported an overall nonassociation for this SNP despite a positive association was observed in stratified analyses. In contrast, Fang et al. (14) showed a null association with lung cancer risk in a Chinese population. Conflicting results also exist for the MMP3 –1171 5A>6A polymorphism. Su et al. (44) reported recently a case-control study using MMP1 –1607 1G>2G, MMP3 –1171 5A>6A, and MMP12 –82A>G (rs2276109) and 1082A>G (rs652438) SNPs as markers and observed an increased risk among never smokers related to the MMP3 6A/6A genotype and associated haplotypes, which is inconsistent with previous study showing the 5A/5A genotype as risk genotype (14). These conflicting results suggest that the selected SNPs in the previous studies may not be efficient markers, and gene-based approach instead of individual SNP-based approach should be used because SNPs having opposite function may concomitantly exist in a gene. By using the strategy of LD-tagging SNPs, we found that it was specific haplotypes but not an individual SNP in these MMPs that play a significant role in lung cancer. In addition, we identified a haplotype profile for lung cancer risk in Chinese that is different from previous study in Caucasians (16, 44). Although we did not analyze the "functional" MMP1 –1607 1G>2G and MMP3 –1171 5A>6A SNPs, we found that the MMP1 –1607 SNP was not located in haplotype block 1. Furthermore, we found that the MMP1 –1607 and MMP3 –1612 SNPs were in weak LD and were not in the same LD block. Therefore, the haplotypes constructed in previous studies (14, 16) based on the selected functional SNPs may introduce inherited errors in the analysis and thus may not accurately reflect the association between the true causal genetic variants and the disease. These observations warrant further studies to find out the true functional variants.

Another interesting result in the present study is that we found haplotype 1c, which is composed of unique MMP1 variants, was not only associated with increased risk of developing lung cancer but also strongly associated with distant metastasis of the cancer. This observation is biologically plausible because overexpression of MMP1, which is capable of degrading interstitial type collagens resulting in expanding growth and migration of cancer cells (12, 45), has been correlated with tumor invasion and metastasis (28, 46–50). In addition, the MMP1 –1607 1G>2G SNP has been associated with the invasion, metastasis, and poor prognosis of many other types of cancer (20, 22–24, 51, 52). For lung cancer, Fang et al. (14) reported recently that the MMP3 –1171 5A/5A genotype and the MMP1 1G-MMP3 5A haplotype were associated with increased risk of lymphatic metastasis of lung cancer. However, we found that the haplotype in the block across MMP1 but not MMP3 was associated with metastasis of the cancer. The distance between MMP1 –1607 and MMP3 –1171 sites is ~45.5 kb and, due to a weak LD, these two SNPs are not located in the same LD block. On the other hand, the risk variant (haplotype 1c) we identified is unlikely to be in LD with the MMP1 –1607 SNP because this polymorphism is not located in haplotype block 1 and LD between block 1 and MMP1 –1607 site is low. These results suggested that there may be functional variant(s) other than –1607 1G>2G in MMP1 that determine lung cancer aggression.

We have implemented an efficient stepwise approach for searching lung cancer alleles in candidate chromosomal region in this study. This two-stage approach differs from htSNP approach described in other articles (53, 54). The major drawback of the htSNP approach is that it is difficult to explore SNPs outside LD blocks and to pinpoint the true causal locus (2, 6, 37). In addition, the htSNP approach may also miss moderately rare SNPs and haplotypes. Although it is more expensive than htSNP approach, our approach holds two noticeable advantages. First, with our approach, SNPs in the region between LD blocks would not be left out. It has been shown that such region is abundant in human genome and might encompass many essential genetic loci (37, 55). In this study, we observed a single locus (rs586701) located beyond the two major blocks that was associated with lung cancer risk, suggesting a need of more subtle analysis of the surround region. Second, with our approach, high-density markers would apply much finer map to pinpoint the disease-causal variants after association analysis, which is important for seeking minimum number of SNPs as genetic markers for practical applications, such as clinical genotyping. In this study, we showed that as few as three htSNPs in each LD block could well represent all SNPs to construct haplotypes that link to the disease.

As the International Hapmap Project advances, more and more information on allele frequencies and LD status of SNPs is available, which benefits the hypothesis-driven association studies for SNP selection in candidate genes or candidate genomic region (56). Although genotyping data for chromosome 11 have been completed now, we could not obtain whole data from Chinese population for SNP selection in the time when this study was launched. SNPs directly chosen from public databases tend to be fallible because a large number of SNPs in the databases are not validated and often display frequencies that are dependent on ethnicity. In the present study, 28.6% (10 of 35) of initially chosen SNPs from the SNP database were verified to be monomorphic or <2% in our study population. Therefore, even after the completion of the Hapmap Project, two-stage study design would still be necessary because limited population sizes have been genotyped in the project (7, 56).

In this study, we used two correction approaches to avoid false-positive results in multiple tests. We did Bonferroni approach by multiplying the P value obtained in logistic regression by 22 (for each of the 22 SNPs) when analyzing single-locus association; meanwhile, all association results underwent 1,000 permutation tests. Having relatively large sample size, homogeneous study population, frequency-matched case-control design, solid and reproducible genotyping techniques, and small P values, our results are unlikely to be false positive.


   Footnotes
 
Grant support: State Key Basic Research Program grant 2004CB518701 (D. Lin) and 2002CB512902 (W. Tan), National "863" High Technology Projects grant 2002BA711A06 (D. Lin) and 2002AA232031 (C. Zeng), the Program for New Century Excellent Talents in University (W. Tan), and the Hundred Talents Program of Chinese Academy of Sciences (C. Zeng).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Note: T. Sun, Y. Gao, and W. Tan contributed equally to this work.

4 http://www.ncbi.nlm.nih.gov/SNP. Back

5 http://www.allsnps.com/snpbrower. Back

6 http://www.hapmap.org. Back

Received 2/27/06; revised 8/12/06; accepted 9/14/06.


   References
Top
 Abstract
 Materials and Methods
 Results
 Discussion
 References
 

  1. Ludwig JA, Weinstein JN. Biomarkers in cancer staging, prognosis, and treatment selection. Nat Rev Cancer 2005;5:845–56.[CrossRef][Medline]
  2. Pharoah PD, Dunning AM, Ponder BA, Easton DF. Association studies for finding cancer-susceptibility genetic variants. Nat Rev Cancer 2004;4:850–60.[CrossRef][Medline]
  3. Neale BM, Sham PC. The future of association studies: gene-based analysis and replication. Am J Hum Genet 2004;75:353–62.[CrossRef][Medline]
  4. Thomas DC, Clayton DG. Betting odds and genetic associations. J Natl Cancer Inst 2004;96:421–3.[Free Full Text]
  5. de Bakker PI, Yelensky R, Pe'er I, Gabriel SB, Daly MJ, Altshuler D. Efficiency and power in genetic association studies. Nat Genet 2005;37:1217–23.[CrossRef][Medline]
  6. Johnson GC, Esposito L, Barratt BJ, et al. Haplotype tagging for the identification of common disease genes. Nat Genet 2001;29:233–7.[CrossRef][Medline]
  7. Bonnen PE, Wang PJ, Kimmel M, Chakraborty R, Nelson DL. Haplotype and linkage disequilibrium architecture for human cancer-associated genes. Genome Res 2002;12:1846–53.[Abstract/Free Full Text]
  8. Houwen L. State of the art: lung cancer in China. Ann Thorac Cardiovasc Surg 2003;9:147–8.[Medline]
  9. Zhang X, Miao X, Liang G, et al. Polymorphisms in DNA base excision repair genes ADPRT and XRCC1 and risk of lung cancer. Cancer Res 2005;65:722–6.[Abstract/Free Full Text]
  10. Gurubhagavatula S, Liu G, Park S, et al. XPD and XRCC1 genetic polymorphisms are prognostic factors in advanced non-small-cell lung cancer patients treated with platinum chemotherapy. J Clin Oncol 2004;22:2594–601.[Abstract/Free Full Text]
  11. Wu X, Zhao H, Amos CI, et al. p53 Genotypes and haplotypes associated with lung cancer susceptibility and ethnicity. J Natl Cancer Inst 2002;94:681–90.[Abstract/Free Full Text]
  12. Egeblad M, Werb Z. New functions for the matrix metalloproteinases in cancer progression. Nat Rev Cancer 2002;2:161–74.[Medline]
  13. 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.[Abstract/Free Full Text]
  14. Fang S, Jin X, Wang R, et al. Polymorphisms in the MMP1 and MMP3 promoter and non-small cell lung carcinoma in North China. Carcinogenesis 2005;26:481–6.[Abstract/Free Full Text]
  15. Zhou Y, Yu C, Miao X, et al. Functional haplotypes in the promoter of matrix metalloproteinase-2 and lung cancer susceptibility. Carcinogenesis 2005;26:1117–21.[Abstract/Free Full Text]
  16. Su L, Zhou W, Park S, et al. Matrix metalloproteinase-1 promoter polymorphism and lung cancer risk. Cancer Epidemiol Biomarkers Prev 2005;14:567–70.[Abstract/Free Full Text]
  17. Yu C, Pan K, Xing D, et al. Correlation between a single nucleotide polymorphism in the matrix metalloproteinase-2 promoter and risk of lung cancer. Cancer Res 2002;62:6430–3.[Abstract/Free Full Text]
  18. Zhang J, Jin X, Fang S, et al. The functional polymorphism in the matrix metalloproteinase-7 promoter increases susceptibility to esophageal squamous cell carcinoma, gastric cardiac adenocarcinoma and non-small cell lung carcinoma. Carcinogenesis 2005;26:1748–53.[Abstract/Free Full Text]
  19. Yu C, Zhou Y, Miao X, Xiong P, Tan W, Lin D. Functional haplotypes in the promoter of matrix metalloproteinase-2 predict risk of the occurrence and metastasis of esophageal cancer. Cancer Res 2004;64:7622–8.[Abstract/Free Full Text]
  20. Ghilardi G, Biondi ML, Mangoni J, et al. Matrix metalloproteinase-1 promoter polymorphism 1G/2G is correlated with colorectal cancer invasiveness. Clin Cancer Res 2001;7:2344–6.[Abstract/Free Full Text]
  21. Xu E, Lai M, Lu B, Xing X, Huang Q, Xia X. A single nucleotide polymorphism in the matrix metalloproteinase-2 promoter is associated with colorectal cancer. Biochem Biophys Res Commun 2004;324:999–1003.[CrossRef][Medline]
  22. Zinzindohoue F, Lecomte T, Ferraz JM, et al. Prognostic significance of MMP-1 and MMP-3 functional promoter polymorphisms in colorectal cancer. Clin Cancer Res 2005;11:594–9.[Abstract/Free Full Text]
  23. Nishioka Y, Sagae S, Nishikawa A, Ishioka S, Kudo R. A relationship between matrix metalloproteinase-1 (MMP-1) promoter polymorphism and cervical cancer progression. Cancer Lett 2003;200:49–55.[CrossRef][Medline]
  24. Lai HC, Chu CM, Lin YW, et al. Matrix metalloproteinase 1 gene polymorphism as a prognostic predictor of invasive cervical cancer. Gynecol Oncol 2005;96:314–9.[CrossRef][Medline]
  25. Pendas AM, Santamaria I, Alvarez MV, Pritchard M, Lopez-Otin C. Fine physical mapping of the human matrix metalloproteinase genes clustered on chromosome 11q22.3. Genomics 1996;37:266–8.[CrossRef][Medline]
  26. Berrieman HK, Ashman JN, Cowen ME, Greenman J, Lind MJ, Cawkwell L. Chromosomal analysis of non-small-cell lung cancer by multicolour fluorescent in situ hybridisation. Br J Cancer 2004;90:900–5.[CrossRef][Medline]
  27. Pan H, Califano J, Ponte JF, et al. Loss of heterozygosity patterns provide fingerprints for genetic heterogeneity in multistep cancer progression of tobacco smoke-induced non-small cell lung cancer. Cancer Res 2005;65:1664–9.[Abstract/Free Full Text]
  28. Michael M, Babic B, Khokha R, et al. Expression and prognostic significance of metalloproteinases and their tissue inhibitors in patients with small-cell lung cancer. J Clin Oncol 1999;17:1802–8.[Abstract/Free Full Text]
  29. Nawrocki B, Polette M, Marchand V, et al. Expression of matrix metalloproteinases and their inhibitors in human bronchopulmonary carcinomas: quantificative and morphological analyses. Int J Cancer 1997;72:556–64.[CrossRef][Medline]
  30. Cho NH, Hong KP, Hong SH, Kang S, Chung KY, Cho SH. MMP expression profiling in recurred stage IB lung cancer. Oncogene 2004;23:845–51.[CrossRef][Medline]
  31. Hofmann HS, Hansen G, Richter G, et al. Matrix metalloproteinase-12 expression correlates with local recurrence and metastatic disease in non-small cell lung cancer patients. Clin Cancer Res 2005;11:1086–92.[Abstract/Free Full Text]
  32. D'Armiento J, Dalal SS, Okada Y, Berg RA, Chada K. Collagenase expression in the lungs of transgenic mice causes pulmonary emphysema. Cell 1992;71:955–61.[CrossRef][Medline]
  33. Hautamaki RD, Kobayashi DK, Senior RM, Shapiro SD. Requirement for macrophage elastase for cigarette smoke-induced emphysema in mice. Science 1997;277:2002–4.[Abstract/Free Full Text]
  34. American Joint Committee on Cancer. AJCC cancer staging manual. 6th ed. New York (NY): Springer; 2002.
  35. Travis WD, Colby TV, Corrin B, Shimosato Y, Brambilla E. Histological typing of lung and pleural tumors. 3rd ed. In: Travis WD, editor. WHO international histological classification of tumors. Berlin: Springer; 1999.
  36. Buetow KH, Edmonson M, MacDonald R, et al. High-throughput development and characterization of a genomewide collection of gene-based single nucleotide polymorphism markers by chip-based matrix-assisted laser desorption/ionization time-of-flight mass spectrometry. Proc Natl Acad Sci U S A 2001;98:581–4.[Abstract/Free Full Text]
  37. Gabriel SB, Schaffner SF, Nguyen H, et al. The structure of haplotype blocks in the human genome. Science 2002;296:2225–9.[Abstract/Free Full Text]
  38. Barrett JC, Fry B, Maller J, Daly MJ. Haploview: analysis and visualization of LD and haplotype maps. Bioinformatics 2005;21:263–5.[Abstract/Free Full Text]
  39. Qin ZS, Niu T, Liu JS. Partition-ligation-expectation-maximization algorithm for haplotype inference with single-nucleotide polymorphisms. Am J Hum Genet 2002;71:1242–7.[CrossRef][Medline]
  40. Carlson CS, Eberle MA, Kruglyak L, Nickerson DA. Mapping complex disease loci in whole-genome association studies. Nature 2004;429:446–52.[CrossRef][Medline]
  41. Bland JM, Altman DG. Multiple significance tests: the Bonferroni method. Br Med J 1995;310:170.[Free Full Text]
  42. Schaid DJ, Rowland CM, Tines DE, Jacobson RM, Poland GA. Score tests for association between traits and haplotypes when linkage phase is ambiguous. Am J Hum Genet 2002;70:425–34.[CrossRef][Medline]
  43. Lake SL, Lyon H, Hantisira K, et al. Estimation and tests of haplotype-environment interaction when linkage phase is ambiguous. Hum Hered 2003;55:56–65.[CrossRef][Medline]
  44. Su L, Zhou W, Asomaning K, et al. Genotypes and haplotypes of matrix metalloproteinase 1, 3, and 12 genes and the risk of lung cancer. Carcinogenesis 2006;27:1024–9.[Abstract/Free Full Text]
  45. Brinckerhoff CE, Rutter JL, Benbow U. Interstitial collagenases as markers of tumor progression. Clin Cancer Res 2000;6:4823–30.[Abstract/Free Full Text]
  46. Inoue T, Yashiro M, Nishimura S, et al. Matrix metalloproteinase-1 expression is a prognostic factor for patients with advanced gastric cancer. Int J Mol Med 1999;4:73–7.[Medline]
  47. Baker EA, Bergin FG, Leaper DJ. Matrix metalloproteinases, their tissue inhibitors and colorectal cancer staging. Br J Surg 2000;87:1215–21.[CrossRef][Medline]
  48. Lin TS, Chiou SH, Wang LS, et al. Expression spectra of matrix metalloproteinases in metastasis non-small cell lung cancer. Oncol Rep 2004;12:717–23.[Medline]
  49. Sunami E, Tsuno N, Osada T, et al. MMP-1 is a prognostic marker for hematogenous metastasis of colorectal cancer. Oncologist 2000;5:108–14.[Abstract/Free Full Text]
  50. 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.[CrossRef][Medline]
  51. Przybylowska K, Zielinska J, Zadrozny M, et al. An association between the matrix metalloproteinase 1 promoter gene polymorphism and lymphnode metastasis in breast cancer. J Exp Clin Cancer Res 2004;23:121–5.[Medline]
  52. Przybylowska K, Kluczna A, Zadrozny M, et al. Polymorphisms of the promoter regions of matrix metalloproteinase genes MMP-1 and MMP-9 in breast cancer. Breast Cancer Res Treat 2006;95:65–72.[CrossRef][Medline]
  53. Sun J, Wiklund F, Zheng SL, et al. Sequence variants in Toll-like receptor gene cluster (TLR6-1-TLR10) and prostate cancer risk. J Natl Cancer Inst 2005;97:525–32.[Abstract/Free Full Text]
  54. Haiman CA, Stram DO, Pike MC, et al. A comprehensive haplotype analysis of CYP19 and breast cancer risk: the multiethnic cohort. Hum Mol Genet 2003;12:2679–92.[Abstract/Free Full Text]
  55. Phillips MS, Lawrence R, Sachidanandam R, et al. Chromosome-wide distribution of haplotype blocks and the role of recombination hot spots. Nat Genet 2003;33:382–7.[CrossRef][Medline]
  56. Schmidt C. Latest HapMap update aims to direct researchers to genetic basis of disease. J Natl Cancer Inst 2005;97:1638–40.[Free Full Text]



This article has been cited by other articles:


Home page
 Cancer Epidemiol. Biomarkers Prev.Home page
W. Sauter, A. Rosenberger, L. Beckmann, S. Kropp, K. Mittelstrass, M. Timofeeva, G. Wolke, A. Steinwachs, D. Scheiner, E. Meese, et al.
Matrix Metalloproteinase 1 (MMP1) Is Associated with Early-Onset Lung Cancer
Cancer Epidemiol. Biomarkers Prev., May 1, 2008; 17(5): 1127 - 1135.
[Abstract] [Full Text] [PDF]

Home page
 Am. J. Pathol.Home page
S. Yamada, K.-Y. Wang, A. Tanimoto, J. Fan, S. Shimajiri, S. Kitajima, M. Morimoto, M. Tsutsui, T. Watanabe, K. Yasumoto, et al.
Matrix Metalloproteinase 12 Accelerates the Initiation of Atherosclerosis and Stimulates the Progression of Fatty Streaks to Fibrous Plaques in Transgenic Rabbits
Am. J. Pathol., May 1, 2008; 172(5): 1419 - 1429.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Sun, T.
Right arrow Articles by Lin, D.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Sun, T.
Right arrow Articles by Lin, D.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Cancer Research Clinical Cancer Research
Cancer Epidemiology Biomarkers & Prevention Molecular Cancer Therapeutics
Molecular Cancer Research Cancer Prevention Research
Cancer Prevention Journals Portal Cancer Reviews Online
Annual Meeting Education Book Meeting Abstracts Online