Effects of O6-methylguanine-DNA methyltransferase (MGMT) polymorphisms on cancer: a meta-analysis.

Yu Zhong,Yongsheng Huang,Yan Huang,Tianbao Zhang,Chengying Ma,Shuyong Zhang,Weiwei Fan,Hongyan Chen,Ji Qian,Daru Lu
DOI: https://doi.org/10.1093/mutage/gep050
2010-01-01
Mutagenesis
Abstract:O-6-methylguanine-DNA methyltransferase is one of the rare proteins to directly remove alkylating agents in the human DNA direct reversal repair pathway. Its two common single-nucleotide polymorphisms, Leu84Phe and Ile143Val, had previously been identified to contribute to susceptibility of cancer. However, there are conflicting results in studies on the association of the two polymorphisms with cancer. Therefore, we conducted a meta-analysis to clarify the paradox with a large collected sample (13 069 cancer patients and 20 290 controls). We found significant association between the T allele (84Phe) and cancer risk, under the recessive genetic model [P = 0.023, odds ratio (OR) = 1.251, 95% confidence interval (CI) 1.031-1.517, P-heterogeneity = 0.270], TT versus CC comparison (P = 0.035, OR = 1.239, 95% CI 1.015-1.511, P-heterogeneity = 0.225) and TT versus CT comparison (P = 0.007, OR = 1.292, 95% CI 1.071-1.559, P-heterogeneity = 0.374), using the random-effect model. In the ethnicity subgroup analysis, a significant association with cancer among Caucasians was found under the recessive genetic model, homozygote comparison and TT versus TC comparison. In the tumour sites subgroup analysis, only the protective effects of Leu84Phe polymorphism were found in colorectal cancer, under CT versus CC comparison. No significant association between the G allele of Ile143Val and cancer risk was found. The G allele showed an increased lung cancer risk under the dominant genetic model and AG versus AA comparison in all Hardy-Weinberg equilibrium subjects, only when the fixed-effect model was used. However, it was insignificant in the random-effect model.
What problem does this paper attempt to address?