VEGF-induced angiogenesis is an early event of EA development. Polymorphisms that can alter VEGF expression and protein production may contribute to the risk of EA. We tested this hypothesis by evaluating the relationship between three functional polymorphisms of VEGF gene and the risk of EA. We found that the +936C/T polymorphism and specific VEGF haplotypes were significantly associated with higher risk of EA. In addition, we found a significant interaction between smoking and −460C/T polymorphism and a greater risk of developing EA.
polymorphism has been studied in relation to various cancers with disparate results. It has been reported to be associated with decreased risk of breast cancer and small cell lung cancer (22
), but another lung cancer study found no relationship (29
). In separate studies, this polymorphism has been studied in several cancers with mixed results (28
). We know of no previous studies investigating the role of +936C/T
polymorphism in EA risk. In the present study, we found that the variant T
allele of the +936C/T
polymorphism was significantly associated with increased risk of EA. Consistent with this result, haplotype CGT
that contains the +936T
allele showed a similar association with the risk of EA. The observed +936CT+TT
–EA associations in our study were consistent with reports of the function of +936CT+TT
in other cancers that are characterized histologically by adenocarcinomas, such as colorectal, breast and gastric cancers (30
The strong interaction effects of the −460C/T
polymorphism and smoking suggest that this polymorphism is not an independent risk factor, but probably an effect modifier that acts synergistically with smoking on EA risk. Previous studies that explored the relationship between VEGF −460C/T
genotypes and cancer risk found inconsistent results. The −460C/T
polymorphism was found to be associated with prostate cancer risk (34
), but no associations were observed in malignant melanoma, lung and renal cell cancers (29
). The reasons for these discrepancies are not completely understood, but if the effect of this polymorphism is only through an interaction with smoking, then main effects may not be observed depending on the distribution of smoking variables across cases and controls. Further, the relevant interactions with smoking may be different as the magnitudes of relationship between cigarette smoking and various cancer risks are different.
Mechanisms for the interaction between smoking and the −460C/T
polymorphism could be at the level of gene expression or protein activity through angiogenesis pathways (37
). The −460C/T
polymorphisms have been associated with variation in VEGF
expression and protein production (20
). Smoking can stimulate both angiogenesis and VEGF
expression, which exacerbates the procarcinogenic effect of angiogenesis (37
). Additionally, it has been shown that smoke-induced VEGF
expression and release were mediated by other pathways, such as the epidermal growth factor receptor–extracellular signal-regulated kinase, 5-lipoxygenase, oxidative stress, inflammatory responses and beta-adrenoceptors pathways (38
). It is possible that cigarette smoke and VEGF activate multiple effects on EA carcinogenesis.
We found no overall association of +405C/G
or its interaction with smoking on EA risk. There are conflicting reports in the literature on the exact function of the +405G/C
polymorphism. The variant C
allele of the +405G/C
polymorphism has been associated with lower VEGF production, and the presence of a C
allele may disrupt the myeloid zinc finger 1
transcription factor-binding site (20
). However, some groups reported that higher VEGF level or no association with the +405C/C
). Any association between +405C/G
and EA may be through its linkage disequilibrium with −460C/T
We recognize several limitations to our study. Firstly, we only evaluated three functional polymorphisms of the VEGF gene. Whether other genetic variants in the VEGF gene would also be associated with EA risk requires further investigation. Secondly, although we adjusted for various confounding variables in our analysis, diet, environmental and occupational exposure data were not included in our logistic regression models because of incomplete or missing information. However, given the consistent results of individual genotypes, gene–environment interaction effects and haplotype analysis, these potential confounders would probably not substantially change the results. Thirdly, the significant associations between VEGF polymorphisms and EA risk were observed in only one study population. These findings should be validated in independent studies, as well as expanded to cover more completely the entire angiogenic pathway. Finally, multiple comparisons may have led to spuriously significant results. Even if we corrected the level of significance for the number of main and gene–smoking interaction analyses (12 in total) using the Bonferroni method, the interaction between −460C/T and smoking status would still be significant.
In summary, this is the first report to our knowledge of the associations between VEGF genotypes, haplotypes and gene–smoking interaction with the risk of EA. Our findings emphasize the importance of considering joint effects of genetic polymorphisms and environmental factors when investigating the risks of EA carcinogenesis. Confirmatory studies and in vitro functional assays are needed to confirm the results and identify the mechanisms underlining the associations.