Abstract
The aim of this study was to investigate the association between the risk for colorectal cancer and single nucleotide polymorphisms (SNP) of matrix metalloproteinase-2 (MMP2) -1306C/T, vascular endothelial growth factor (VEGF) 936C/T and hypoxia inducible factor-1α (HIF1A) 1772C/T. Patients and Methods: A total 50 colorectal cancer patients (46% women, mean age 68±11 years) were enrolled. Healthy controls without evidence of cancer history or family cancer predispositions were frequency-matched to the cases by sex and age (±5 years). Genotyping was performed by polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP) method and the genotype distribution and risk estimate were analyzed. The correlation between the genotypes and clinicopathological parameters (Dukes stage, phenotype, location, differentiation and size) among colorectal cancer patients were investigated. Results: There was a significant association between colorectal cancer and T allele-bearing genotype distribution of HIF1A 1772C/T polymorphism (Odds ratio, OR=3.63, 95% confidence interval, CI=1.08-12.18, p=0.03 for CT and TT genotypes relative to CC genotype). In addition, when stratified by age, the association remained in patients older than 60 years old (OR=13.60, 95% CI=1.63-113.24, p=0.01). However, there was no association between the genotypes of the MMP2, VEGF and HIF1A SNP and clinicopathological parameters of colorectal cancer. Conclusion: There is a significant association between the HIF1A 1772C/T SNP and the risk of developing colorectal cancer, especially in individuals older than 60 years.
Colorectal cancer is one of the major causes of mortality worldwide. It is the most common malignancy of the gastrointestinal tract and its occurrence is rapidly increasing in Korea (1). The adenoma-carcinoma sequence has been accepted as the main pathway for colorectal cancer development, involving numerous molecular events, such as the activation of oncogenes and the inactivation of suppressor genes (2). Recently, with the completion of the Human Genomic Project, single nucleotide polymorphisms (SNPs) have been the focus of biomedical research (3).
Polymorphisms are naturally occurring DNA sequence variations, which differ from gene mutations. They occur in the ‘normal’ healthy population and have a frequency of at least 1% (4). Approximately 90% of DNA polymorphisms are SNPs due to their single base substitutions, while others include insertion, deletion, minisatellite, and microsatellite polymorphisms (5). Although most polymorphisms are functionally neutral, some have effects on the regulation of gene expression or on the function of the coded protein. These functional polymorphisms could contribute to the difference between individuals according to the susceptibility and severity of diseases (5). Polymorphisms alone or in combination with environmental factors may affect the angiogenic pathway, and thereby the susceptibility and severity of cancer (4). SNPs in several genes have been suggested to be implicated in the pathogenesis of colorectal cancer (6).
MMPs function as proteolytic enzymes degrading the extracellular matrix and the basement membrane (5). The overexpression of MMPs is known to be associated with tumor invasion, metastasis, and a worse prognosis (7). Of all the MMPs, MMP2 (gelatinase A), after being processed into its active form as type IV collagenase, primarily hydrolyzes type IV collagen, which is the major structural component of the basement membrane. Several polymorphisms within the MMP2 promoter regions have been reported in cases with oncogenesis and tumor progression, especially in colorectal carcinogenesis (8). Recently, the SNP C →T transition at -1306 is reported to disrupt the Sp1-type promoter site (CCACC box), displaying a strikingly lower promoter activity and also reducing the transcriptional activity (9). Sp1 is a ubiquitously expressed transcription factor that binds to GC/GT-rich elements to regulate a variety of genes. The CCACC box has been shown to be essential for Sp1 binding and promoter function in several genes by invariably activating transcription (10, 11). The MMP2 −1306C/T polymorphism that abolishes Sp1 binding has the potential to affect the level and specificity of gene transcription. This has been demonstrated by in vitro experiments using the transient transfection method (12). Previous studies have demonstrated that the −1306CC genotype can double or quadruple the risk factors for maliganacies, such as lung, stomach, colorectal, oral cavity and breast cancer (9). But these findings are not always in accordance with other results and the association of MMP2 −1306C/T polymorphism in colorectal cancer is still in investigation.
VEGF is one of the most important activators of tumor angiogenesis, stimulating new blood vessel formation from existing vasculature, increasing the tumor oxygenation, and ultimately, leading to tumor growth (13). Cells under hypoxic conditions consistently produce pro-angiogenic factors, mostly VEGF which triggers multiple signaling pathways that result in endothelial cell survival, mitogenesis, migration, differentiation, and also increased vascular permeability and mobilization of endothelial progenitor cells (14). The overexpression of VEGF mRNA and its protein has been associated with tumor progression and the poor prognosis of various malignancies, including melanoma, ovarian carcinoma, prostate carcinoma, and colon cancer (15). Moreover, the overexpression of VEGF was clearly demonstrated in metastatic compared to non-metastatic colon cancer, and known to be directly correlated with increased vessel counts, the extent of neovascularization and the degree of proliferation (15). DNA sequence variations in the VEGF gene seem to affect the VEGF level and/or its activity, thereby causing differences between patients in lymphangiogenesis and lymphatic tumor spread. In particular, variations occurres in the 3’-untranslated region (UTR) are known to be correlated with key regulatory elements that are sensitive to hypoxia and, therefore, contribute to high variability in VEGF level among tissues (16). The SNP C →T transition especially at 936 within the 3’-UTR region of the VEGF gene has recently been associated with colorectal cancer (17).
HIF1A is a key transcription factor that regulates the cellular response to hypoxia (18). Under hypoxic conditions, degradation of HIF1A is suppressed and its intracellular accumulation activates VEGF, which induces the tumor growth via angiogenesis (19-20). An increased HIF1A level, therefore, plays a critical role in the ability of cancer cells to survive and metastasize in the hostile hypoxic environment. Recent studies have demonstrated the existence of SNPs in the human HIF1A gene that lead to amino acid substitutions within the oxygen-dependent degradation domain (ODD) (21). Mutations in this critical regulatory domain may lead to the overexpression of this protein and also the subsequent changes in the expression of downstream target genes, thus contributing to the development of cancer cell resistance. SNP in HIF1A, C →T transition at 1772 (which may result in an amino acid change from proline 582 to serine) was shown to have a higher transactivating capability in vitro compared to the wild-type allele (21), and the genotype distribution and functional significance of the HIF1A SNP in human malignancy, especially in colorectal cancer, is still under investigation.
In our previous study, we confirmed the up-regulated mRNA expression of MMP2, VEGF, and HIF1A in colorectal cancer tissues compared with the noncancerous tissue (22). Here, we investigated the risk of colorectal cancer and the clinical manifestations of the SNPs, including genes for MMP2 (C →T transition at −1306), VEGF (C →T transition especially at 936 within the 3’-UTR), and HIF1A (C →T transition at 1772) in 50 colorectal cancer patients and healthy controls in Korea.
Patients and Methods
Patients and samples. A total of 50 colorectal cancer patients (46% women, mean age of 68±11 years) were enrolled from January to December 2008 at Ewha Womans University Mokdong Hospital. Clinicopathological data were reviewed. Healthy controls without any evidence of cancer history or family cancer predisposition were frequency-matched to the cases by sex and age (±5 years). DNA of each sample was amplified by polymerase chain reaction for MMP2, VEGF, and HIF1A SNP. Genotyping was performed by polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP) method, and the genotype distribution and risk estimate were also analyzed. The tumor stage was described according to the Dukes staging system. The correlation between genotypes and clinicopathological parameters such as the tumor location, differentiation, size, lymphatic invasion, lymph node metastasis, and phenotype among colorectal cancer patients were investigated. This study was approved by the Institutional Review Board (IRB number ECT 216-1) of Ewha Womans University, and the written consent was obtained from each participant.
Genomic DNA extraction. Genomic DNA samples were prepared by proteinase K digestion and extracted with LaboPass™ Tissue Mini (COSMO GENETECH, Seoul, Korea) and Genomic Blood DNA Extraction Mini Kit (iNtRON Biotechnology, SungNam, Korea) from surgically resected cancer tissues and whole blood samples of controls respectively. The concentration of each genomic DNA was measured by the absorbance at 260 nm and samples were frozen at −20°C until further analysis.
Genotyping of SNP MMP2, VEGF, and HIF1A by PCR-RFLP. Genotypes were determined by the PCR-RFLP method. The PCR reaction was performed in a 20 μL sample volume containing 100 ng of DNA template, 2 μL of 10× PCR buffer, 0.5 U of Taq DNA polymerase (Takara Bio Inc., Shiga, Japan), 200 μM of each dNTPs, and 0.5 μM of each primer. All PCR reactions were performed by a thermal cycler GeneAmp PCR system 9600 (Perkin-Elmer Corp., Norwalk, CT, USA). The PCR products were separated by running on 3% agarose gel stained with ethidium bromide and detected by UV transillumination. The GelDoc 2000 system (BioRad, Hercules, CA, USA) was used to detect each band.
Genotyping of MMP2 polymorphism. A 188 base pairs (bps) fragment was amplified from genomic DNA using the following primers: 5’-CTT CCT AGG CTG GTC CTT ACT GA-3(forward) and 5’-CTG AGA CCT GAA GAG CTA AAG AGC T-3(reverse). The PCR cycling conditions were 5 minutes at 94°C followed by 35 cycles of 45 seconds at 94°C, 45 seconds at 58°C and 45 seconds at 72°C, with a final step at 72°C for 5 minutes to allow the complete extension of all PCR fragments. The PCR products were digested at 37°C overnight with 5 U of restriction enzyme, Xsp I (Takara Bio Inc.).
On electrophoresis of PCR products, CC homozygotes showed a single band at 188 bp, and CT heterozygotes showed bands at 188, 162, and 26 bp, while TT homozygotes showed double bands at 162 and 26 bp (Figure 1).
Genotyping of VEGF polymorphism. The PCR primers used to detect the VEGF 936C/T polymorphism were 5’-AAG GAA GAG GAG ACT CTG CGC AGA GC-3(forward) and 5’-TAA ATG TAT GTA TGT GGG TGG GTG TGT CTA CAG-3(reverse). Thermal conditions included a denaturation step at 94° for 5 minutes, followed by 35 cycles at 94°C for 40 second, annealing at 64°C for 1 minute, and extension at 72° for 40 seconds with a final extension step at 72° for 5 minutes. The PCR product was digested overnight with 5 U of restriction enzyme, N1a III (New England Biolabs, Ipswich, MA, USA). On electrophoresis of PCR products, CC homozygotes showed a single band at 208 bp, CT heterozygotes showed bands at 208, 122, and 86 bp, while TT homozygotes showed double bands at 86 bp and 122 bp (Figure 2).
Genotyping of HIF1A polymorphism. The primer sequences to detect the 1772C/T polymorphism were 5’-TGT GGC CAT TGT AAA AAC TCA-3(forward) and 5’-CTT GCG GAA CTG CCT TCT AA-3(reverse). The PCR temperature profile was 1 cycle of 94°C for 5 minute, followed by 35 cycles of 94°C for 1 minute, 55°C for 1 min, and 1 min at 72°C with the final extension step at 72°C for 5 min. The 147 bp PCR fragment was incubated overnight with 10 U of restriction enzyme, Bsl I (New England Biolabs), at 37°C. On electrophoresis of PCR products, CC homozygotes showed double bands at 131 and 19 bp, CT heterozygotes showed bands at 147, 131, and 19 bp, while TT homozygotes showed a single band at 147 bp only (Figure 3). The details of primer sequence and reaction conditions used for RFLP typing are summarized in Table I.
Statistical analysis. All data were analyzed using SPSS version 13.0 for Windows (SPSS Inc, Chicago, IL, USA). Independent t-test was used for continuous variables such as age, and the Pearson chi-square test was for genotype distribution and risk estimate for MMP2, VEGF, and HIF1A SNP. Odds ratio (OR) and 95% confidence intervals (95% CI) were calculated by logistic regression analysis. Clinicopathological parameters such as the tumor location, differentiation, size, and phenotype among colorectal cancer patients were investigated using the Pearson Chi-square test or the Fischer's exact test according to the genotype. P-values less than 0.05 were regarded as statistically significant.
Results
Patient demographics. The baseline characteristics of colorectal cancer patients and the control group are shown in Table II. Of the 50 patients with colorectal cancer at the time of diagnosis, the majority were found to have tumor located at the colon (n=36, 72.0%). The size of the tumor mass was found to be larger than 4 cm in the majority of patients (n=37, 74.0%). Ulcerative tumor was found in most patients (n=41, 82.0%) and most tumors were moderately differentiated (n=43, 91.4%).
MMP2 −1306 C/T polymorphism. The genotype distribution and risk estimate among the patient and control groups were analyzed (Table III). No significant association was found between the risk of developing colorectal cancer and genotype distribution of MMP2 −1306 C/T polymorphism (OR=0.675, 95% CI=0.246-1.854, p=0.446). Nor was any significant association observed in the genotype distribution and risk estimate when stratified by gender (male, OR=1.307, 95% CI=0.310-5.509, p=0.715; female, OR=1.676, 95% CI=0.403-6.966, p=0.447) (Table IV) or age (>60 years old, OR=0.517, 95% CI=0.165-1.625, p=0.259; ≤60 years old, OR=2.154, 95% CI=0.174-25.672, p=0.550) (Table V). Finally, no association was found between the genotype distribution of the MMP2 polymorphism and clinicopathological parameters (localization, differentiation, size, lymph node invasion, lymph node metastasis, ulcerative/polypoid phenotype and Dukes' stage) of colorectal cancer (Table VI).
VEGF 936 C/T polymorphism. No significant association was found between the risk of developing colorectal cancer and genotype distribution of VEGF 936 C/T polymorphism (OR=1.179, 95% CI=0.532-2.610, p=0.685). No significant association was observed in genotype distribution and risk estimate when stratified by gender (male, OR=1.169, 95% CI=0.391-3.494, p=0.780; female, OR=1.192, 95% CI=0.373-3.807, p=0.767) (Table IV) or age (>60 years old, OR=1.263, 95% CI=0.489-3.261, p=0.629; ≤60 years old, OR=1.000, 95% CI=0.232-4.310, p=1.000) (Table V). Finally, no association was found between the genotype distribution of the VEGF polymorphism and clinicopathological parameters of colorectal cancer (Table VII).
HIF1A 1772 C/T polymorphism. There was a significant association found in the risk of developing colorectal cancer and genotype distribution of HIF1A 1772 C/T polymorphism (OR=3.632, 95% CI=1.082-12.183, p=0.037) for CT and TT genotypes relative to CC genotype. Moreover, when stratified by age, significant correlation was found in patients older than 60 years (OR=13.600, 95% CI=1.633-113.246, p=0.016) for C/T and T/T genotypes relative to CC genotype (Table V). However, no significant association was observed in the genotype distribution and risk estimate when stratified by gender (male, OR=3.250, 95% CI=0.361-33.409, p=0.321; female, OR=4.286, 95% CI=0.981-18.721, p=0.053) (Table IV). Finally, no association was found in the genotype distribution of the HIF1A polymorphism with clinicopathological parameters of colorectal cancer (Table VIII).
Discussion
The role of genetic polymorphisms which function as the important factors determining endogenous causes of cancer, especially in the risk of colorectal cancer, has attracted increasing interest due to advances in DNA analyzing technologies and also the knowledge of the human genome. To date, however, very little about this matter is proven and known.
MMPs are zinc metalloproteases that degrade collagens of the extracellular matrix that are important for tissue remodeling and repair, during the development and inflammation. MMPs are also involved in controlling cell cycle checkpoints, genomic instability, and cell adhesion (23). Excessive or inappropriate expression of MMPs may contribute to the pathogenesis of cancer by facilitating tissue degradation. Currently, more than 20 MMPs have been identified, which can be categorized by their substrate specificity (7, 23). Accumulating evidence has shown that MMP promoter SNPs affecting the gene transcription are associated with an enhanced susceptibility to developing malignant diseases, including colorectal cancer, significantly affecting survival and prognosis.
The MMP2 promoter contains sequences for the binding of AP-2, p53, Sp1, and Sp3 (11). Price et al. (12) identified the -1306C/T polymorphism in the MMP2 promoter and showed a strikingly lower promoter activity with the T allele via in vitro transient transfection experiment. The resulting base transition occurs in the CCACC box of the Sp1 binding site and eliminates promoter activity. It is likely that the - 1306CC genotype may be associated with a high transcription level and enzyme activity of MMP-2, and eventually, it might affect individual susceptibilities to neoplasms. Xu et al. (9) found some evidence that individuals with CC genotypes had a higher risk (OR=1.959, 95% CI=1.06-3.64) of developing colorectal cancer compared with those with CT or TT genotypes, and that patients with CC genotype more often had tumors with extended invasion into the serosa or adventitia layer. Langers et al. (24) found significant association between MMP2 - 1306C/T SNP, tumor stage, and the survival of the patient. Hettiaratchi et al. (25) and Elander et al. (26), however, showed no significant association between the highly active C allele of MMP2 -1306 C/T polymorphism and clinicopatholgical parameters or susceptibility of colorectal cancer. In our study, there was no significant association between genotype distribution of the MMP2 -1306 C/T polymorphism and colorectal cancer. Moreover, no significant difference was noted in MMP2 -1306 C/T polymorphism and clinicopathological parameters.
Evidence from preclinical and clinical studies have shown that VEGF functions as a predominant angiogenic factor in human colorectal cancer, which is associated with the metastases and poor prognosis (13). VEGF gene, which contains eight exons and seven introns, is known to be located on the chromosomal subband 6p21.3. At least 30 SNPs in this gene region have been described in published studies (27, 28). In particular, the 936C/T SNP in the 3’-UTR is of particular note. The 3’-UTR of the VEGF gene has been proven to increase the stability of mRNA and to be associated with the hypoxic induction of VEGF (29-32). Recently, genes designed as Hu family have been identified and their products have been shown to bind the AU-rich element of 3’-UTR of several genes, including the VEGF mRNA (29, 33-35). It is suggested that the proteins of the Hu family change the VEGF mRNA conformation so that the mRNA is not attacked by RNAase. Alterations in a few nucleotides in the 3’-UTR such as SNPs have been shown to be associated with the deregulation of affected genes (36). Therefore, it is suggested that nucleotide polymorphism in the 3’-UTR may alter the mRNA conformational integrity, resulting in genetic variation of VEGF gene expression.
Several studies have reported the significant association of the 936T allele of VEGF with the risk of developing cancer such as oral cancer, gastric cancer and colon cancer (37), whereas in other studies on breast and lung cancer (38), protective effects of the variant were found. Concerning colorectal cancer, Yamamori et al. (28) comparatively analyzed different VEGF gene polymorphisms located in the regions of the promoter, the 5’-and the 3’-UTR from the tumor tissues of 18 Japanese patients, and their findings indicate an association between reduced risk of colorectal cancer with certain SNPs in the VEGF gene. Bae et al. (17) found a significant gender difference by investigating the 936T allele-bearing genotype with the risk of colon cancer. Recently, some investigators found that T allele-bearing genotypes in VEGF 936C/T SNP seem to be related to a lower overall survival, and that VEGF polymorphism could be an independent prognostic marker and genetic determinent for colorectal patients (17, 27, 39). However, Hofmann et al. (40) have recently reported that there was no association between the 936T allele and the risk of colon cancer, and also, no correlation was found between genotype of 936C/T SNP and tumor characteristics such as size, histological grading, positive regional lymph node metastases or tumor stage. They also found that common SNPs including 936 C/T polymorphism in VEGF were not associated with an individual's susceptibility to CRC. In the present study, we did not find any association with either the genotype distribution or the other clinicopathological parameters. A reason for these inconsistant findings could be a linkage disequilibrium with other, perhaps not yet discovered, functional SNPs in the VEGF gene or unknown SNPs in angiogenesis-linked signal pathways (37).
SNP C →T transition at 1772 of HIF1A was shown to cause amino acid substitutions within or near the N-terminal transactivation domain (N-TAD) within the ODD of HIF1A (41) (Figure 4). Tanimoto et al. (21) showed the elevated transactivation capacity of variant forms of HIF1A, by assessing the transcription activity in co-transfection experiment with a hypoxia-response element (HRE)-driven luciferase reporter gene in COS7 cells, which implied a role of HIF1A polymorphisms in generating individually different tumor progression potential. They demonstrated that the presence of these polymorphic variants seems to cause a significantly higher transcriptional activity than the activity of the wild-type (21), and that these polymorphic variants were associated with higher microvessel density and higher disease stage in patients with head and neck squamous cell carcinoma (21). A significant difference in genotype distribution of the 1772C/T polymorphism between patients and the control group was seen in androgen-independent prostate cancer (42). Konac et al. (43) showed a statistically significant increase in the percentage of carriers of the T allele (CT, TT and CT+TT genotypes) in cervical and endometrial cancer, which indicates that carrying a T allele would be a risk factor in the development of those cancers. The polymorphic allele of 1772C/T polymorphism was also shown to be associated with increased risk of developing the breast cancer (44), but the association was not seen in some cases (45, 46).
To date, only few SNP studies concerning colorectal cancer have been published. In 2004, Kuwai et al. (47) claimed that HIF1A 1772C/T polymorphism is not involved in progression or metastasis of colorectal carcinoma. However, two years later, Fransen et al. (48) showed a significantly higher risk for the development of colorectal cancers with more severe ulcerative growth pattern in heterozygous (CT) patients, compared to the patients who were homozygous (CC) wild-type. In our results, we did not find any association between HIF1A polymorphism and the phenotype (polypoid vs. ulcerative) of tumor (p=0.465). But we did find a higher risk for the development of colorectal cancer in T-allele bearing genotypes of the HIF1A gene and this was much more evident in patients of age older than 60 years. Previous studies have suggested that molecular and pathophysiological changes occurring throughout life progressively modify molecular homeostasis of colonic epithelial cells and lead to neoplasia (49). DNA damage is certainly increased in older rodents, suggesting frequent stochastic cellular insults (50-52). Aging was also proving to increase epithelial proliferation in rodent (53) and human (54) colon. In spite of several studies on SNPs, the mechanism which results in the replacement with serine (T-allele) at amino acid 582, and therefore affect HIF1A structure and function, is unclear in the elderly group. Although any further attempt to explain the mechanism behind this interaction would be imprudent, these findings clearly warrant further investigations into the role of the HIF1A 1772C/T polymorphism, particularly with advancing age.
During the last few years, in spite of previous attempts on clarifying the association of SNP with colorectal cancer in order to determine the higher risk groups, its prognostic effect, and eventually help early diagnosis and cure, clinical relevance has not yet been shown. To our best knowledge, this is the first study to investigate the genotype distribution, clinicopathological impact, and susceptibility to colorectal cancer of three genes, MMP2, VEGF and HIF1A SNPs in the same study population. Moreover, we demonstrated significant association between 1772C/T polymorphism of HIF1A gene and colorectal cancer in this Korean population. Given the ethnically homogenous background of Korean patients, any potential confounding effect due to ethnicity is likely to be small in the present study.
In conclusion, our findings indicate that the HIF1A C/T polymorphism may be a useful indicator of the susceptibility to colorectal cancer and that this may be useful in detecting early carcinogenesis and progression of colorectal cancer. Furthermore, larger scaled genetic studies of MMP2, VEGF and HIF1A SNPs are needed for their functional relevance to be elucidated.
- Received November 26, 2010.
- Revision received January 11, 2011.
- Accepted January 12, 2011.
- Copyright© 2011 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved