A genetic variant in the APE1/Ref-1 gene promoter -141T/G may modulate risk of glioblastoma in a Chinese Han population

The current study included 766 glioma patients and 824 healthy cancer-free controls. According to histological diagnosis, the glioma patients were stratified into three subgroups: 241 glioblastomas, 284 astrocytomas except for glioblastoma and 241 other gliomas. The subject enrollment commenced in October 2004 and is ongoing. All newly diagnosed glioma cases were inpatients for tumor resection at the Neurosurgery Department of Huashan Hospital of Fudan University, Shanghai, China. This hospital is the biggest center of neurosurgery in the east of China [30]. Histopathological diagnoses of potentially eligible cases were carried out on the basis of the biopsies or resected specimens in the Department of Neuropathology of the same hospital. Although there were no restrictions on age, gender and histology, patients who had previous cancers and/or previous radiotherapy or chemotherapy were excluded. Cancer-free control subjects were randomly recruited from annual checkup visitors (80%) and trauma outpatients (20%) at the same hospital during the similar time period and were frequency- matched to cases by gender, age (±5 years) and residence (urban or rural). The included controls self-reported no history of malignancies, genetic neurologic disorders and radiotherapy/chemotherapy. No evidence of demographic differences was detected between these trauma outpatients and the annual checkup subjects (data not shown). All participants were genetically unrelated ethnic Han Chinese people from Shanghai and its surrounding regions including Jiangsu, Zhejiang, and Anhui provinces.

Each eligible subject was face-to-face interviewed by a trained personnel who was blinded to the case-control status with a shortened structured questionnaire, which was adapted from the original version provided by Dr. Melissa Bondy for brain study in the Department of Epidemiology at M.D. Anderson Cancer center, to obtain data on demographic factors, smoking status, radiation exposure, family history of cancer and other factors. Each subject provided the informed consent and donated 3-5 ml of peripheral blood. As a result, DNA samples and questionnaires were available from 766 case and 824 control subjects representing an 85.7% and 79.2% of all eligible case and control subjects, respectively. The research protocol was approved by the Ethics Committee for Human Subject Research of Fudan University.

Genomic DNA was extracted from whole-blood samples using the Qiagen Blood Kit (Qiagen, Chatsworth, CA, USA), according to the manufacturer's instructions. After extraction, genomic DNA was diluted to a final concentration of 15-20 ng/μl for the genotyping assays. Polymorphism spanning fragments were amplified by the polymerase chain reaction (PCR) and performed genotyping with Sequenom MassARRAY iPLEX platform by use of allele-specific MALDI-TOF mass spectrometry assay [31]. Primers for amplification and extension reactions were designed using MassARRAY Assay Design Version 3.1 software (Sequenom, San Diego, CA). And SNP genotypes were obtained according to the iPLEX protocol provided by manufacturer. Genotyping quality was examined by a detailed QC procedure consisting of > 95% successful call rate, duplicate calling of genotypes, internal positive control samples and Hardy-Weinberg Equilibrium (HWE) testing.

Differences in demographic variables, smoking status, family history of cancer, and grouped allele and genotype frequencies between cases and controls were compared by the χ² test (df = 2). The Hardy-Weinberg Equilibrium was determined using the standard χ² test (df = 1) to compare the observed frequency with the expected frequency in controls. To estimate the association between the APE1/Ref-1 -141T/G variant and risk of gliomas, odds ratios and 95% confidence intervals (95% CI) were calculated using unconditional logistic regression with adjustment for age and gender. The homozygosity with the more frequent allele among controls was set as the reference group. A test of trend was calculated by treating the three genotypes (major allele homozygous, heterozygous and variant allele homozygous) as ordinal variables in the regression models. Furthermore, we performed stratified analyses for the APE1/Ref-1 -141T/G polymorphism by age (two groups: < 18 years and > 18 years), gender, smoking status, family history of cancer and histologic type. The SPSS 15.0 software (SPSS, Chicago, IL, USA) was used for all statistical analyses. All P values presented were two sided, and a level of P < 0.05 was considered statistically significant.

To further assess the probability of a spurious association due to multiple testing, we calculate the false-positive report probability (FPRP) for a SNP from an estimated OR and 95% CIs using the methodology described by Wacholder et al [32]. To assign a prior probability for the APE1/Ref-1 -141T/G variant, we considered that the genetic variant has been previously shown to influence expression of the APE1/Ref-1 gene [28, 29] and has been previously reported to be associated with lung cancer risk [28, 29]. In the light of this, prior probabilities of 25% (and 10%) were assigned. However, no previous studies have been reported on the association between the APE1/Ref-1 promoter polymorphism and glioma risk. Hence, we selected more conservative prior probabilities, 1% and 10%. In accordance with Wacholder et al [32], a standard FPRP cut-off of < 0.5 was selected with a cut-off of < 0.2 being considered more stringent.

The details of selected characteristics of 766 glioma cases ad 824 control subjects have been shown in Table 1. The differences in the distributions of age, gender and smoking status between cases and controls were not statistically significant (P = 0.353, 0.919, and 0.535, respectively). The mean age of cases (42.4 years) at diagnosis was slightly older than the reference age for controls (41.5 years). 59.1% cases and 59.9% controls were male. As shown in the previous study, the cases were more likely to report a history of cancer in first-degree relatives than controls (20.3% vs. 15.5%; P = 0.027). Among 766 glioma cases, 241 were glioblastoma, 284 astrocytomas except for glioblastoma and 241 other gliomas (Table 1).

Allele and genotype frequencies and associated ORs (95% CI) for glioma cases and controls are presented in Table 2. The variant allele G was 45.0% in controls and 42.8% in cases. The frequencies of the TT, TG and GG genotypes of APE1/Ref-1 rs1760944 were 29.0%, 52.0% and19.0% in controls, and 30.4%, 51.2% and 16.3% in cases, respectively. No significant differences were found in allele and genotype distributions of APE1Ref-1 rs1760944 T/G between glioma cases and controls (P = 0.220 and 0.410, respectively). In logistic regression analysis, neither the homozygous -141GG nor heterozygous TG genotype was associated with risk of gliomas in overall population. Further analysis showed that rs1760944 T/G genotypes were not associated with gender, age, smoking status, or family history of cancer (Table 3).

Next, we performed the stratified analysis for APE1/Ref-1 rs1760944 by glioma histology. As shown in Table 4, statistically significant differences were observed in allele and genotype distributions of rs1760944 T/G between glioblastoma patients and control subjects (P = 0.032 and 0.041, respectively). Overall, the variant G allele was associated with a decreased risk of glioblastoma compared with the T allele (OR = 0.80, 95% CI = 0.65-0.98, P = 0.032). Individuals with the homozygous -141GG genotype exhibited 46% reduced risk of glioblastoma (adjusted OR = 0.54, 95% CI = 0.34-0.87, P = 0.012), compared with the TT homozygote. According to the criteria suggested by Wacholder et al [32], this result remained robust given the appropriate prior probability of 25% (FPRP = 0.052) and 10% (FPRP = 0.140), but not with a prior probability of 1% (FPRP = 0.643). The P-associated with the trend test was 0.014 (P for trend = 0.014). However, no significant association of rs1760944 with risk of low-grade astrocytomas (astrocytomas except for glioblastoma) and other gliomas was found in this study. As the number of glioblastoma subjects was not large enough for stratification analysis, we did not further evaluate the association between the rs1760944 T/G and risk for glioblastoma by age, gender, smoking status and family history of cancer.

We also performed post-hoc power calculation based on the observed variant allele frequencies in both control and case groups, using post hoc power analyses of G*Power3 (http://​www.​psycho.​uni-duesseldorf.​de/​abteilungen/​aap/​gpower3). It showed that with a sample size 1052 (236 glioblastoma cases and 816 controls), the study had > 90% power in detecting associations of rs1760944 polymorphism with risk of glioblastoma, at a significance level of 0.05 (df = 2).