|Home | About | Journals | Submit | Contact Us | Français|
Previous studies did not discriminate wild-type from hemizygous genotypes of GSTM1 and GSTT1. In this study, we investigated wild-type, hemizygous deletion, and homozygous deletion genotypes of GSTM1 and GSTT1 and lung cancer risk.
We conducted a nested case-control study of 143 primary incident lung cancer cases and 447 matched to cancer-free controls Genotyped data were obtained using a real-time PCR-based assay. Conditional logistic regression models were used to estimate odds ratios (ORs) and 95% Confidence Intervals (CIs).
Compared to GSTM1 wild-type carriers, the relative odds of lung cancer increased from 1.49 (95% CI=0.66–3.40) to 1.80 (95% CI=0.81–4.02) for the hemizygous and homozygous deletion genotypes, respectively (p-trend=0.13). The strongest associations were seen among those who smoked <1 pack/day and had ≥deletion variant of GSTM1 (OR=3.25; 95% CI=0.93–11.34; p-trend=0.07) whereas the reverse was observed for smokers who smoked ≥1 pack per day (OR=0.80; 95% CI=0.24–2.67; p-interaction=0.08). No clear associations were observed for GSTT1 genotypes.
Risk of lung cancer increased as the number of deletion variants increased for GSTM1, though the associations were non-significant. Discriminating between the wild-type, hemizygous, and homozygous deletion GSTM1 genotypes permitted a more precise characterization of the associations between GSTM1 deletion variants and lung cancer.
Lung cancer is the leading cause of cancer death worldwide (1). Cigarette smoking is responsible for approximately 85% of lung cancer deaths (2) and constituents of cigarette smoke such as polycyclic aromatic hydrocarbons (PAHs) are known carcinogens (3). Hydrophobic PAH-derived electrophiles bind to DNA to form PAH-DNA adducts, which may initiate carcinogenesis (4–6). Genes belonging to the glutathione S-transferase (GST) family detoxify tobacco-related carcinogens such as PAHs before the carcinogens damage DNA (7). GSTM1 and GSTT1 have been identified to have a deletion genotype (homozygous deletion) that results in a total absence of enzymatic activity. Because of their functional role in detoxifying tobacco-related carcinogens, variants of GSTs, such as GSTM1 and GSTT1, may modulate lung cancer risk and contribute to individual susceptibility (7).
An extensive body of evidence has accumulated on the association between GSTM1 and GSTT1 and lung cancer risk comparing null-versus-present genotypes (8). Greater concentrations of PAH-DNA adducts have been found in lung tissues of GSTM1 homozygous deletion (“null”, or 0/0) smokers compared to smokers with GSTM1 present genotypes (wild-type or “+/+” plus hemizygous deletion or “ +/0” genotypes) (9). There is evidence of increased lung cancer risk among East Asians with GSTM1 homozygous deletion, but not for Caucasians (meta-analysis of 98 studies) (10). Until recently genotyping assays were unable to discriminate between wild-type and hemizygous deletion genotypes (11). Hence, most of the evidence on this topic to date has not distinctly accounted for hemizygous GSTM1 and GSTT1 genotypes. Consequently, previous investigations often compared the risks associated with the homozygous deletion genotype to a referent “present” genotype that was comprised of both wild-type and hemizygous deletion genotypes. Data suggest that distinguishing the wild-type from the hemizygous genotype is important, as studies using new real-time PCR-based assays that discriminate between the wild-type, hemizygous deletion, and homozygous deletion GSTM1 and GSTT1 genotypes show these genotypes to exhibit a high, intermediate, and absence of enzymatic activity, respectively (12–15). This raises the possibility that previously published results may tend to underestimate the true associations of GSTM1 and GSTT1 in relation to lung cancer risk because this misclassification would bias results toward the null (16). Support for this hypothesis has been observed for other malignancies (17–19). The present study was carried out to test whether the lung cancer risk varies according the number of functional alleles of GSTM1 and GSTT1 using a case-control study nested within a community-based cohort in Washington County, Maryland.
Established in 1989, the CLUE II cohort was named for its campaign slogan, “Give Us a Clue to Cancer and Heart Disease.” The details of the establishment of the cohort have been reported elsewhere (20, 21). Briefly, from May through October 1989, 32,897 residents of Washington County, Maryland agreed to participate in CLUE II. Of these individuals, 25,081 adults (>18 year old) provided a Washington County address and are covered by the county cancer registry. Starting in 1996, the CLUE II cohort members received periodic follow-up questionnaires.
At baseline, basic demographic characteristics, smoking status, and number of cigarette smoked per day were obtained using a brief questionnaire. At that time, participants also provided blood samples (20 ml) drawn into a 20 mL Vacutainer tube containing heparin and immediately refrigerated until centrifugation. Centrifugation usually took place within 6 hours and never exceeded twenty-four hours. Once centrifuged, aliquots of plasma, red blood cells, and buffy coats were separated and stored at-70° C in a specimen bank.
Ascertainment of lung cancer cases was achieved through linkage with the Washington County Cancer Registry, the Maryland State Cancer Registry, and death certificates. The Washington County Cancer Registry receives its data primarily from the county’s only general hospital, Washington County Hospital. CLUE II cohort members were also linked to the Maryland State Cancer Registry, which was established in 1993 and has a mandatory cancer reporting policy.
The present study is comprised of 143 incident lung cancer cases that occurred from 1990 to October, 2005 and 447 matched cancer-free controls with genotype data for at least one GSTM1 or GSTT1. With the exception of non-melanoma skin cancer or cervical carcinoma in situ, cases were first-time cancer diagnosed with primary lung cancer (International Classification of Diseases, Eighth Revision [ICD-8] for cases diagnosed before 1992 code 162; ICD-9 code 162 for cases diagnosed from 1992–2000, and ICD-10 codes C33-C34 for cases diagnosed from 2000-present).
For each case, we selected up to four controls. At least one control was matched to each case (n=17 case-control sets with one control) and the average number of controls per case was three. Eligibility criteria for control selection were: (1) completion of the baseline questionnaire; (2) no prior history of cancer except for non-melanoma skin cancer or cervical cancer in situ; (3) cancer-free and known to be alive at the time of case diagnosis.
Controls were individually matched to cases on the following variables: gender, age (±5 years), and smoking status (never, former, or current smokers). For former and current smokers, cases and controls were further matched on the number of cigarettes smoked per day as follows. Ever smokers were categorized into three smoking groups: ≤ 19, 20 to 29, and ≥ 30 cigarettes per day. For individuals who smoked <30 cigarettes per day, cases and controls were matched within ±5 cigarettes smoked per day. Those who smoked between 30–45 cigarettes per day were matched within ±10 and the heaviest smokers (≥45) were matched within ±20 cigarettes. The 1996 follow-up questionnaire was relevant to this study as it provided additional detail on pack-years of smoking on cases and controls. Among cases who provided information on the 1996 follow-up questionnaire about pack-years of smoking, we matched controls to within ±5 pack-years of cigarette smoking.
DNA was extracted from buffy coat, frozen at −70°C, using an alkaline lysis method (22). GSTM1 and GSTT1 genotyping was performed by Applied Biosystems (Foster City, California; www.appliedbiosystems.com) using TaqMan® Gene Copy Number Assays (PN4331182). The assay consisted of two primers and a FAM™-MGB probe in 20X formulation and measurements were made in real time. Primers and probes were designed from genomic sequence (hg18/Build 36) using Applied Biosystems proprietary software. For quality control, each assay was run as a duplex TaqMan real-time PCR reaction, one containing a FAM dye-based assay for the targeted gene and a VIC dye-based assay for the reference gene. An additional, a known sample was inserted randomly in each batch which was blinded to the technician performing the assay.
All assays were conducted in a 96-well plate (MicroAmp™ Optical 384-Well Reaction Plate). Each well contained FAM-labeled TaqMan probe for either GSTM1 or GSTT1 and VIC-labeled TaqMan probe for the reference gene. PCR was performed in a reaction mixture containing the following: 2x TaqMan® Universal PCR Master Mix; both GSTM1 primer and probe or GSTT1 primer and probe; DNase-free water; and genomic DNA sample. Real-time data analysis was performed using Sequence Detection Software v2.1 (Absolute Quantification and Copy Number Macro, www.allgenes.com). Relative quantity is determined by the Ct ((FAM Ct - VIC Ct)sample - (FAM Ct - VIC Ct)calibrator) method, where a reference sample or calibrator known to have two copies of the test sequence is used as the basis for comparative results. The gene copy number is two times the relative quantity (23).
Hardy-Weinberg Equilibrium (HWE) was assessed based on the distribution of the GSTM1 and GSTT1 genotypes among the controls using SIBPAIR version 0.99.0 (http://www.qimr.edu.au/davidD/davidd.html). Unless otherwise indicated, conditional logistic regression was used to estimate odds ratios (OR) and 95% confidence intervals (CI) for lung cancer risk comparing the GSTM1 and GSTT1 wild-type to the hemizygous deletion and the homozygous deletion genotypes, respectively.
All estimates were adjusted for age (continuous) and number of cigarettes smoked per day (continuous) to account for possible residual confounding that could persist after matching. Additional analyses were stratified by gender, above-versus-below the mean age of the controls, smoking status (never, former, and current), and above-versus-below 20 cigarettes smoked per day (median among the controls). The likelihood ratio test was used to test for dose-response trends across the three genotypes GST genotypes by fitting a single genotype variable as a single categorical variable. To assess how the results of analyses using the refined genotyping compared to the traditional dichotomous null-versus-present genotyping, analyses were also performed comparing the homozygous deletion genotype to a referent group comprised of the combined hemizygous deletion (+/0) plus wild-type genotypes (+/+). A two-tailed p-value of < 0.05 was considered to be statistically significant. All analyses were performed using STATA version 9.1
Cases and controls were matched on age, gender and smoking status and did not differ significantly with regard to other selected demographic characteristics (Table I). On average, the study population was 62 years of age at baseline. As expected, never smokers comprised only a small percentage (~8%) of lung cancer cases.
Among the controls, the frequencies observed for GSTM1 wild-type, hemizygous deletion, and homozygous deletion genotypes were 9%, 40%, and 50% respectively (Table II). For GSTT1, the distribution was 24%, 51%, and 20% for wild-type, hemizygous deletion, and homozygous deletion genotypes, respectively. These frequencies of GSTM1 and GSTT1 homozygous deletion genotype seen in the controls were similar to those previously observed for Caucasians (8). The tests for HWE showed no deviation (for GSTM1: p-value = 0.80 and for GSTT1: p-value=0.11).
Compared to those with the GSTM1 wild-type genotype, the risks of lung cancer were 1.49 (95% CI =0.66–3.40) and 1.80 (95% CI = 0.81–4.02) for those with the hemizygous deletion and homozygous deletion genotypes, respectively (p- trend = 0.13, Table III). For GSTT1, compared to the wild-type genotype, the risks of lung cancer were 1.17 (95% CI = 0.71–1.92) and 1.06 (95% CI= 0.58–1.95) for those with the hemizygous and homozygous deletion genotypes, respectively (p-= 0.83). None of these associations were statistically significant.
When stratified by smoking status, the odds ratios for lung cancer among those with the hemizygous and homozygous deletion GSTM1 genotypes ranged between 1.40 and 1.99 in both former and current smokers (Table III). Among those who smoked <=20 cigarettes per day equals to 1 pack or less per day, compared to the wild-type genotype the odds ratios were 3.06 (95% CI= 0.83–11.28; p-value= 0.09) for the hemizygous deletion genotype and 3.35 (95% CI= 0.94–11.86; p-value= 0.06) for the homozygous deletion genotype (p-trend= 0.12) (Table III). In contrast, the odds ratios were slightly in the protective direction for smokers who smoked > 1pack per day. The p-value for the test for interaction by smoking intensity was 0.07.
The smoking-stratified results for GSTT1 were opposite those seen for GSTM1. Specifically, no associations were seen in ever smokers who smoked ≤1 pack per day, whereas among heavier smokers the associations for the hemizygous deletion and homozygous deletion GSTT1 genotypes were in the direction of increased risk.
To illustrate the potential value of the refined classification of GST genotypes that explicitly accounts for hemizygotes, we reanalyzed our data using the traditional null-versus-present comparison in which the referent group included the hemizygous deletion GSTM1 genotype. When the data were re-classified to conform to the traditional null-versus-present genotype, the risks associated with the GSTM1 genotype were always attenuated toward the null compared to the more refined classifications we presented in our primary analyses (Null vs present comparison, Table III). This was particularly true among lighter smokers among whom the odds ratios diminished from 3.35 to 1.31 for the GSTM1 homozygous deletion variant.
GSTM1 and GSTT1 have been thoroughly investigated in relation to lung cancer risk because of their critical role in inactivating tobacco-related carcinogens (24). The value of the present study was the use of a real-time PCR-based assay for genotyping that distinguishes between the three distinct genotypes of wild-type, hemizygous deletion, and homozygous deletion, whereas most previous studies could only distinguished the homozygous deletion, “null” genotype, versus the “present” genotype consisting of the wild-type and hemizygous deletion genotypes. Based on the GSTM1 null-versus-present comparisons used predominantly in the past, overall weak or no associations have been observed (25–33). Notably, when we reanalyzed our GSTM1 data using this traditional genotyping classification, the associations were attenuated. This is consistent with misclassification of those lacking a functional allele. The results were not statistically significant, but the overall pattern of associations suggested the risk of lung cancer increased as the copy number of GSTM1 deletion variants increased from zero (wild-type) to one (hemizygous deletion) to two (homozygous deletion). These results are consistent with previous observation that hemizygous and wild-type genotypes are associated with function difference in enzymatic activity (15).
The only previous study we are aware of to report on the associations between GSTM1 and GSTT1 genotypes and lung cancer risk that classified hemizygous deletion separately from wild-type observed no difference between the hemizygous and homozygous null genotypes of GSTM1 and GSTT1 when compared to the referent wild-type genotype (34). Our results are in agreement with the previous lack of associations observed for GSTT1, and thus suggest caution in interpreting our results for GSTM1. Compared to the previous study, the careful matching of cases and controls on smoking exposures is a notable strength of our study design. This is primarily because cigarette smoking is the predominant cause of lung cancer such that even the residual effects of smoking could overwhelm smaller risks associated with genetic factors.
Although not statistically significant, our smoking-stratified results suggested that lung cancer risk associated with GSTM1 deletions was more pronounced among smokers who smoked ≤ 1 pack of cigarettes per day, whereas the converse was observed for heavy smokers. The lack of association among heavier smokers could possibly be due to the carcinogen exposure levels being so high that they overwhelm the effects of the GSTM1 deletion genotype. Our findings of increased risk for light smokers are consistent with other findings from case-control studies on GSTM1-smoking interaction in relation to lung cancer (35–37), suggesting that GSTM1 may have a more discernable influence in the milieu of lower carcinogenic exposure. The findings of our study also may be due to chance.
The present study benefits from the prospective study design that avoids methodological limitations of retrospective studies such as selection and recall bias. In this setting, the relevance of recall bias primarily pertains to the measurement of cigarette smoking. Blood samples were collected at baseline and before onset of lung cancer diagnosis. This minimizes the possibility of survival bias that might be attributable to GSTM1 variants (38, 39).
An important limitation of this study is that it lacked adequate statistical precision to detect statistical significance that may be associated with GST genotypes. We focused on only a few genes, but a pathway-based approach, in which the concerted risk of multiple at-risk variants of metabolic genes are investigated, would provide a more comprehensive characterization of variant genes and lung carcinogenesis (40). For example, a deficiency in one or two genes may be compensated by other genes within the GST pathway (40, 41).
In summary, the results of this nested case-control study provide limited support for the hypothesis that accounting for the hemizygous deletion GSTM1 genotype strengthens the association between GSTM1 genotype and lung cancer risk compared to previous evidence based on present-versus-null genotype comparisons. The results were not statistically significant, but these hypothesis-generating findings imply additional investigations with genotyping for hemizygote status will enhance the resolution of our understanding of this question.
This work was supported by the Institutional Research Epidemiology Fellowship (T32CA009314), National Institute of Aging (5U01AG018033) and National Cancer Institute (CA105069). The authors express their appreciation to the participants of the CLUE II cohort and thank the staff at the George W. Comstock Center for Public Health Research and Prevention for their dedication and contributions to the study.