Abstract
Background
We conducted a hospital-based study of 1110 SCCHN cases and 1129 controls to replicate the associations reported by a recent large European study between two potentially functional single nucleotide plymorphisms (SNPs) of the alcohol dehydrogenases genes, ADH1B R48H (rs1229984: G>A) and ADH7 A92G (rs1573496: C>G), and risk of squamous cell carcinoma of the head and neck (SCCHN).
Methods
Multivariate logistic regression was used to calculate adjusted odds ratios (OR) and 95% confidence intervals (95% CI). False-positive report probabilities were also calculated for significant findings.
Results
We found that the ADH7A92G GG and combined CG+GG genotypes were associated with a decreased risk of SCCHN (adjusted OR, 0.32; 95% CI, 0.13-0.82 for GG and adjusted OR, 0.74; 95% CI, 0.59-0.94 for CG+GG; FPRP, .098) compared with the CC genotype. This association was also evident in subgroups of older (> 57 years) subjects, males, former smokers, oral cancer, and N0 lymph node metastasis (P < .05 for all); however, such associations were not observed for the ADH1B R48H SNP.
Conclusion
Our results support the ADH7 A92G SNP as a marker for risk of SCCHN in non-Hispanic White populations.
Keywords: ADH7, genetic variant, genetic susceptibility, head and neck cancer, molecular epidemiology
Introduction
Head and neck cancer, which includes cancers that arise in the oral cavity, pharynx, and larynx, is the sixth most common type of cancer. It is estimated about 650 000 new cases and 350 000 new deaths worldwide every year.1 All head and neck cancers are chracteristic of squamous cell carcinoma (SCCHN).2 Although smoking and alcohol consumption are known as major risk factors for SCCHN, only few smokers and drinkers develop SCCHN, suggesting that there exists genetic susceptibility to this disease in the general population.
Acetaldehyde, as the first product of ethanol metabolism, is a main carcinogen involved in the etiology of alcohol-related cancer.3 Although studies have indicated that ethanol could be oxidized to acetaldehyde by several pathways in humans, alcohol dehydrogenase (ADH) is believe to play a major role in this metabolic process.4 There are seven isozymes of human ADH coded by different genes. Because of high similarities in structure and kinetic properties, these ADH isozymes are classified into five classes (i.e., Class I: ADH1A, ADH1B, ADH1C; Class II: ADH4; Class III: ADH5; Class IV: ADH7; and Class V: ADH6).5 There are several hundreds of genetic variants reported in these seven ADH genes,6 and cumulative evidence has shown that some of these genetic polymorphisms may play an important role in the etiology of SCCHN.
For example, studies in Asian populations have indicated that carriers of the GG genotype (formerly named ADH1B *1/*1) of ADH1B R48H (rs1229984: G>A) may have a higher risk, than carriers of other genotypes, of developing cancers of the upper aerodigestive tract (UADT) (including oral cavity, pharynx and esophagus cancer) in moderate or heavy drinkers compared with lighter drinkers.7-9 Moreover, studies in European ethnic origin populations have showed that the A allele of the ADH1B R48H may be associated with a decreased risk of UADT cancers regardless of drinking status.10 In addition, the GG genotype of ADH1C I350V (rs698: A>G) was also found to be associated with an increased risk of SCCHN by modifying the biologically-effective dose of alcohol in a US case-control study;11 however, a study in European populations suggested an opposite effect10. To date, most published studies of the associations between ADH polymorphisms and risk of SCCHN have dealt with ADH1B and ADH1C, but the results were not conclusive.12-14
All ADH genes are located in a cluster (5′- ADH7 - ADH1C - ADH1B - ADH1A - ADH6 - ADH4 - ADH5 - 3′) of approximately about 370 kb in the long arm of chromosome 4.15 Studies have already indentified that there is strong linkage disequilibrium (LD) among variants of the ADH1B and ADH1C genes.16, 17 Moreover, LD analysis for SNPs that cover this region from the HapMap also showed a strong LD among seven ADH genes.18 The strong LD over most of the region covering these ADH genes makes it difficult to determine the exact contributions of individual variants. To date, most published studies of associations between ADH genes and cancers, as well as other alcohol-related diseases, failed to address this issue, which may explain the inconsistent results in published studies.19 Since there is a high LD among SNPs in these ADH genes, it is ideal to systemically assess the associations between all genetic variants of these ADH genes and risk of SCCHN.
Recently, having analyzed the LD pattern of SNPs in all ADH genes, Hashibe and colleagues had selected six representative missense SNPs in ADH genes to assess the impact of genetic variants of ADH genes on risk of UADT cancers in a large European multicenter study. As a result, two significant SNPs, ADH1B R48H (rs1229984:G>A) and ADH7 A92G (rs1573496:C>G), were found to be associated with risk of the UADT cancers in 3,800 cases and 5,200 controls.18 Since SCCHN is part of the UADT cancers, we used the DNA samples from our ongoing SCCHN study to perform a replication study to test the hypothesis that these two functional SNPs, ADH1B R48H (rs1229984: G>A) and ADH7 A92G (rs1573496: C>G) as identified as important SNPs in UADT cancers, are associated with the risk of SCCHN in non-Hispanic white subjects in a Texas population.
Materials and Methods
Study Subjects
We used the subjects recruited from our ongoing molecular epidemiology of SCCHN, for which the recruitment of subjects has been described elsewhere.20, 21 In brief, all newly diagnosed, untreated SCCHN patients were histopathologically confirmed at The University of Texas M. D. Anderson Cancer Center between May 1995 and December 2007. Patients with second SCCHN primary tumors, primary tumors of the nasopharynx or sinonasal tract, primary tumors outside the UADT, cervical metastases of unknown origin, or any histopathological diagnosis other than SCCHN were excluded. All eligible cases were approached for participation, and the response rate was approximately 90%. As a result, this case-control study included 1110 subjects with primary tumors of the oral cavity (n = 327; 29.4%), pharynx (n = 609; 54.9%, including 566 oropharynx and 43 hypopharynx), or larynx (n = 174; 15.7%).
According to the American Joint Committee on Cancer (AJCC),22 the regional lymph node involvement of SCCHN was defined as N0 to N3 as follows: N0, no regional node metastasis; N1, metastasis in a single ipsilateral lymph node, ≤3 cm in the greatest dimension; N2, metastasis in a single ipsilateral lymph node, >3 cm but <6 cm in the greatest dimension; or in multiple ipsilateral lymph nodes, none ≥6 cm in the greatest dimension; or in any bilateral or contralateral lymph node, <6 cm in the greatest dimension; N3, metastasis in any lymph node, ≥6 cm in the greatest dimension. The extent of the primary SCCHN was defined as T1 to T4 as follows: T1, tumor ≤2 cm at the greatest dimension; T2, tumor >2 cm but <4 cm in the greatest dimension; T3, tumor ≥4 cm in the greatest dimension; T4, tumor invading adjacent structures.
Self-reported cancer-free control subjects were recruited from M. D. Anderson Cancer Center visitors during the same period, who were genetically unrelated to the enrolled case subjects or each other and accompanied patients to the clinics but were not seeking medical care. We first surveyed potential control subjects at the clinics by using a short questionnaire to determine their willingness to participate in research studies and to obtain demographic information for frequency matching to the cases by age (± 5 years) and sex. Of the eligible controls, the response rate was approximately 90%.
Having given a written informed consent, each subject completed a questionnaire to provide additional information about demography and risk factors, such as age, sex, ethnicity, tobacco smoking and alcohol use. Subjects who had smoked less than 100 cigarettes in their lifetime were considered “never smokers;” and all others were considered “ever smokers.” For ever smokers, those who had quit smoking for more than one year before recruitment were considered “former smokers”, and the remaining “current smokers.” Subjects who had drunk alcoholic beverages at least once a week for more than one year were considered “ever drinkers” and all other were considered “nondrinkers.” Among ever drinkers, those who had quit drinking for more than one year before recruitment were considered “former drinkers”, and the others “current drinkers.” Each subject also provided a one-time sample of 30-ml blood for further biomarker assays. Considering genotype frequencies may vary between ethnic groups and few minority patients were recruited, we only included non-Hispanic whites in this analysis. The research protocol was approved by the M. D. Anderson Cancer Center institutional review board.
Genotyping
Genomic DNA were extracted from the buffy-coat fraction of the blood samples by using a blood DNA mini kit (Qiagen Inc., Valencia, CA) according to the manufacturer's instructions. Restriction fragment length polymorphism (RFLP) - polymerase chain reaction (PCR) was used to identify the genotypes of the ADH1B R48H (rs1229984: G>A) and ADH7 A92G (rs1573496: C>G) polymorphisms. The PCR mixture included approximately 20 ng of genomic DNA, 0.1 mM deoxynucleotide triphosphate, and 1× PCR buffer (50 mM KCl, 10 mM Tris HCl, and 0.1% Triton X-100); 1.5 mM MgCl2, 0.5 units of Taq polymerase (Denville Scientific Inc.; Metuchen, NJ); and 2 pmol of each primer. The genomic DNA samples were amplified with two primers (mismatch bases are underlined): 5′- AGAAACACAATTTCAGGAATTTGGGT -3′ (forward) and 5′- ACTAACCACGTGGTCATCTGCG -3′ (reverse) for ADH1B R48H and 5′- TTGTGGTTTGACACCTGCAT -3′ (forward) and 5′- ATTTTGGCCACAGGAATCTG -3′ (reverse) for ADH7 A92G. There were 132-bp and 163-bp PCR products for ADH1B R48H and ADH7 A92G, respectively. The final PCR reaction products were digested with restriction enzymes, HhaI for ADH1B R48H and BtsI for ADH7 A92G (New England Biolabs, Beverly, MA) and were separated in 3% and 2% gels, respectively. The PCR assays were conducted, and the results were evaluated without knowing the subjects' case and control status. More than 10% of the samples were randomly selected for repeated assays, and the results of both sets of analyses were 100% concordant.
Statistical Analysis
The chi-square tests were used to compare the differences in frequency distributions of demographic variables, other known risk factors, and alleles and genotypes of the ADH1B R48H and ADH7 A92G SNPs between the cases and controls. Additionally, the associations between the SNPs and SCCHN risk were estimated by computing odds ratio (ORs) and 95% confidence intervals (95% CI) with and without adjustment for age, sex, smoking and drinking status in unconditional multivariate logistic regression analyses. Stratified analysis was performed to explore possible interactions between each of SNPs and stratified variables using similar unconditional multivariate logistic regression models. The SAS ALLELE and HAPLOTYPE procedures were used to calculate LD and infer haplotype frequencies based on the observed genotypes.
We also calculated the false-positive report probability (FPRP) 23 that depends on the prior probability by which the SNP is associated with SCCHN, the power of the present study and the observed P value. We set 0.2 as a FPRP threshold, i.e., the probability of a false positive result is less than 20%. For all significant results in the present study, we assigned a prior probability of 0.01 to detect an OR of 1.56 (for a risk effect) or 0.67 (for a protective effect) for an association with genotypes and heplotypes of each SNP. Only significant results with a FPRP value less than 0.2 were considered a noteworthy association.
All statistical tests were two-sided, a P-value of 0.05 was considered significant, and all tests were performed by using the SAS software (version 9.13.; SAS Institute, Cary, NC).
Results
Characteristics of study subjects
The characteristics of the study population are shown in Table 1. Both cases and controls had similar frequency distributions of age and sex (P = .482 and P = .708, respectively), suggesting the frequency matching was adequate. However, there were more current smokers and drinkers in cases than in controls (P < .001 for both), which were controlled for in later multivariate logistic regression analysis.
Table 1.
Frequency distributions of selected variables in SCCHN patients and cancer-free controls
| Variable | No (%) | P* | |
|---|---|---|---|
| Cases (n = 1110) | Controls (n = 1129) | ||
| Mean age (± SD) | 57.2 (±11.1) | 56.8 (±11.0) | .414 |
| Age | |||
| ≤ 50 | 302 (27.2) | 317 (28.1) | .482 |
| 51 to 57 | 290 (26.1) | 270 (23.9) | |
| > 57 | 518 (46.7) | 542 (48.0) | |
| Gender | |||
| Female | 273 (24.6) | 270 (23.9) | .708 |
| Male | 837 (75.4) | 859 (76.1) | |
| Smoking status | |||
| Never | 308 (27.8) | 552 (48.9) | <.001 |
| Former | 380 (34.2) | 412 (36.5) | |
| Current | 422 (38.0) | 165 (14.6) | |
| Alcohol use | |||
| Never | 304 (27.4) | 494 (43.8) | <.001 |
| Former | 242 (21.8) | 182 (16.1) | |
| Current | 564 (50.8) | 453 (40.1) | |
| Primary tumor† | |||
| T1 | 282 (25.4) | ||
| T2 | 401 (36.1) | ||
| T3 | 222 (20.0) | ||
| T4 | 204 (18.5) | ||
| Lymph node metastasis† | |||
| N0 | 404 (36.4) | ||
| N1 | 160 (14.4) | ||
| N2 | 502 (45.3) | ||
| N3 | 43 (3.9) | ||
| Tumor site | |||
| Oral cavity | 327 (29.4) | ||
| Pharynx | 609 (54.9) | ||
| Larynx | 174 (15.7) | ||
Two-sided χ2 test for differences in the distributions between the cases and controls.
One case has missed information for primary tumor, lymph node metastasis.
Association between ADH1B and ADH7 genotypes and SCCHN risk
The genotype and allele distributions of the two selected SNPs in the cases and controls are summarized in Table 2. The observed genotype frequencies of ADH1B R48H in controls were in Hardy-Weinberg equilibrium (P = .108), but that of ADH7 A92G were not (P = .025). There was no statistically significant difference in the frequencies of both ADH1B R48H genotypes and alleles between cases and controls (P = .374 and .689, respectively). However, for ADH7 A92G, the cases had a significant lower frequencies of CG and GG genotypes (P < .001) and G allele (P < .001), compared with the controls. With adjustment for age, sex, smoking and drinking status, logistic regression analysis suggested a significantly lower risk of SCCHN associated with the ADH7 GG (adjusted OR, 0.32; 95% CI, 0.13-0.82) and combined CG+GG genotypes (adjusted OR, 0.74; 95% CI, 0.59-0.94). However, these associations were not observed for ADH1B R48H.
Table 2.
Genotype frequencies of the ADH1B and ADH7 polymorphisms among SCCHN cases and control subjects and their associations with risk of SCCHN
| Variables | No. (%) | P* | OR (95% CI) | ||||
|---|---|---|---|---|---|---|---|
| Cases (n=1110) | Controls (n=1129) | Crude | Adjusted † | ||||
| ADH1B R48H (rs1229984: G>A) | |||||||
| GG | 1059 | 95.4 | 1075 | 95.2 | .374 | 1.00 | 1.00 |
| AG | 51 | 4.6 | 52 | 4.6 | 1.00 (0.67-1.48) | 1.33 (0.88-2.02) | |
| AA | 0 | 0.0 | 2 | 0.2 | NA | NA | |
| AG and AA | 51 | 4.6 | 54 | 4.8 | .823 | 0.96 (0.65-1.42) | 1.29 (0.86-1.94) |
| A allele frequency | 0.023 | 0.025 | .689 | ||||
| ADH7 A92G (rs1573496: C>G) | |||||||
| CC | 948 | 85.4 | 902 | 79.9 | .000 | 1.00 | 1.00 |
| CG | 156 | 14.1 | 206 | 18.2 | 0.72 (0.57-0.90) | 0.79 (0.62-1.00) | |
| GG | 6 | 0.5 | 21 | 1.9 | 0.27 (0.11-0.68) | 0.32 (0.13-0.82) | |
| CG and GG | 162 | 14.6 | 227 | 20.1 | .000 | 0.68 (0.54-0.85) | 0.74 (0.59-0.94) |
| G allele frequency | 0.076 | 0.110 | <.001 | ||||
Two-side χ2 test for differences in the frequency distributions of genotypes, combined genotypes, or alleles between cases and controls.
Adjusted for age, sex, smoking and drinking status.
We further stratified the observed associations of ADH1B R48H and ADH7 A92G genotypes with risk for SCCHN stratified by age, sex, smoking and drinking status, lymph node metastasis, primary tumor status and tumor site. As shown in Table 3, for ADH7 A92G, the decreased risk of SCCHN associated with the combined ADH7 CG+GG genotypes was significant for subgroups of older (>57 years) subjects (adjusted OR, 0.67; 95% CI, 0.48 -0.94), males (adjusted OR, 0.70; 0.54 – 0.91), former smokers (adjusted OR, 0.68; 0.46-0.99), N0 lymph node metastasis (adjusted OR, 0.64; 95% CI, 0.45-0.90), oral cavity (adjusted OR, 0.63; 95% CI, 0.43-0.91), marginal for T1 (adjusted OR, 0.69; 0.48-1.01) and T2 primary tumors (adjusted OR, 0.72; 0.52-1.00). For ADH1B R48H, although no overall risk was observed, a significantly increased risks of SCCHN were associated with the combined ADH1B AG+AA genotypes in subgroups of young (≤50 years) subjects (adjusted OR, 2.39; 95% CI, 1.12 -5.10) and N2 lymph node metastasis (adjusted OR, 1.62; 95% CI, 1.00-2.63). Because the findings in fewer subgroups were significant, the results for ADH1B R48H were likely by chance. Furthermore, there were no associations modulated by subgroups of alcohol use, although these two genes are involved in alcohol metabolism.
Table 3.
Stratification analysis of associations of the genotypes of the ADH1B and ADH7 with risk of SCCHN
| Stratified variables | No. (case/control) | ADH1B R48H rs1229984: G>A | P* | ADH7A92G rs1573496: C>G | P* | ||
|---|---|---|---|---|---|---|---|
| Percentage (case/control) | Adjusted OR (95% CI) † | Percentage (case/control) | Adjusted OR (95% CI) † | ||||
| GG | GG vs. AG + AA | CC | CC vs. CG + GG | ||||
| Age | |||||||
| ≤ 50 | 302/317 | 93.4/96.2 | 2.39 (1.12-5.10) | .025 | 85.5/80.4 | 0.74 (0.48-1.14) | .171 |
| 51 to 57 | 290/270 | 94.8/93.0 | 1.00 (0.47-2.10) | .996 | 84.5/80.7 | 0.90 (0.57-1.44) | .664 |
| ≥ 57 | 518/542 | 96.9/95.8 | 0.90 (0.45-1.79) | .757 | 85.7/79.2 | 0.67 (0.48-0.94) | .022 |
| Gender | |||||||
| Male | 837/859 | 95.3/95.2 | 1.26 (0.79-2.02) | .333 | 85.9/79.9 | 0.70 (0.54-0.91) | .008 |
| Female | 273/270 | 95.6/95.2 | 1.34 (0.57-3.14) | .501 | 83.9/80.0 | 0.90 (0.56-1.44) | .654 |
| Smoking status | |||||||
| Never | 308/552 | 91.6/94.4 | 1.59 (0.90-2.67) | .118 | 81.2/79.4 | 0.90 (0.63-1.28) | .547 |
| Former | 380/412 | 96.3/94.9 | 0.77 (0.38-1.54) | .453 | 85.8/79.9 | 0.68 (0.46-0.99) | .043 |
| Current | 422/165 | 97.4/98.8 | 2.14 (0.46-10.00) | .334 | 88.2/81.8 | 0.61 (0.36-1.01) | .055 |
| Alcohol use | |||||||
| Never | 304/494 | 91.8/93.3 | 1.39 (0.80-2.40) | .240 | 83.6/78.7 | 0.74 (0.51-1.07) | .110 |
| Former | 242/182 | 97.9/96.7 | 0.77 (0.22-2.68) | .682 | 85.1/76.9 | 0.67 (0.40-1.13) | .134 |
| Current | 564/453 | 96.3/96.7 | 1.35 (0.66-2.76) | .408 | 86.5/82.3 | 0.80 (0.56-1.14) | .219 |
| Primary tumor | |||||||
| T1 | 282/1129 | 95.0/95.2 | 1.24 (0.67-2.30) | .500 | 85.8/79.9 | 0.69 (0.48-1.01) | .055 |
| T2 | 401/1129 | 94.5/95.2 | 1.53 (0.90-2.61) | .113 | 85.8/79.9 | 0.72 (0.52-1.00) | .047 |
| T3 | 222/1129 | 96.4/95.2 | 1.05 (0.48-2.31) | .895 | 84.2/79.9 | 0.78 (0.52-1.17) | .226 |
| T4 | 204/1129 | 96.6/95.2 | 1.09 (0.47-2.55) | .828 | 85.3/79.9 | 0.76 (0.49-1.18) | .222 |
| Lymph node metastasis | |||||||
| N0 | 404/1129 | 96.5/95.2 | 0.93 (0.50-1.75) | .826 | 87.6/79.9 | 0.64 (0.45-0.90) | .010 |
| N1 | 160/1129 | 95.6/95.2 | 1.24 (0.54-2.86) | .611 | 86.9/79.9 | 0.63 (0.39-1.04) | .069 |
| N2 | 502/1129 | 94.0/95.2 | 1.62 (1.00-2.63) | .048 | 83.7/79.9 | 0.81 (0.60-1.07) | .140 |
| N3 | 43/1129 | 100.0/95.2 | -- | -- | 79.1/79.9 | 1.15 (0.54-2.47) | .720 |
| Tumor site | |||||||
| Oral cavity | 327/1129 | 94.5/95.2 | 1.48 (0.82-2.66) | .189 | 87.5/79.9 | 0.63 (0.43-0.91) | .014 |
| Pharynx ‡ | 609/1129 | 95.2/95.2 | 1.20 (0.74-1.94) | .466 | 83.9/79.9 | 0.80 (0.61-1.04) | .093 |
| Larynx | 174/1129 | 97.7/95.2 | 0.92 (0.31-2.73) | .877 | 86.8/79.9 | 0.68 (0.41-1.13) | .141 |
Two-side χ2 test obtained from multivariate logistic regression models with adjustment for age, sex, smoking and drinking status. The sum of subjects in some strata was less than the total number of subjects due to missing information.
Adjusted by age, sex, smoking and drinking status.
Included 565 oropharynx cases and 44 hypopharynx cases.
Association between ADH1B and ADH7 haplotypes and SCCHN risk
The LD analysis found that these two SNPs were in incomplete LD in this study population (D′ = .386, r2 = .031, P < .001), and four possible haplotypes were inferred based on the observed genotype data (Table 4). The global test for the haplotype distribution between the cases and the controls was statistically significant (P < .001). When the most common haplotype “GC” (i.e., ADH1B G + ADH7 C) was used as the reference, the haplotype “GG” (i.e., ADH1B G + ADH7 G) was associated with a significant protection against risk of SCCHN (adjusted OR, 0.74; 95% CI, 0.59-0.92), suggesting a major role of the ADH7 G allele, although the ADH1B G allele and ADH7 G allele had an opposite effect. In contrast, the haplotype “AC” (i.e., ADH1B A + ADH7 C) was associated with a significantly increased risk of SCCHN (adjusted OR, 1.71; 95% CI, 1.02 - 2.87), suggesting a major role of ADH7 C allele. Although the effect of combined haplotype “AG” did not reach significance (adjusted OR, 0.70; 95% CI, 0.36-1.34), the non-significant protective effect of this haplotype further supports a protective role of the ADH7 G allele.
Table 4.
Frequency of inferred haplotypes of ADH1B and ADH7 based on observed genotypes and their association with risk of squamous cell carcinoma of the head and neck
| Haplotype | No. chromosomes (%)*† | Crude OR (95% CI) |
Adjusted OR (95% CI) ‡ |
||
|---|---|---|---|---|---|
|
ADH1B R48H rs1229984 (G>A) |
ADH7A92G rs1573496 (C>G) |
Cases No. (%) |
Controls No. (%) |
||
| G | C | 2016 (90.8) | 1982 (87.78) | 1.00 | 1.00 |
| G | G | 153 (6.9) | 220 (9.74) | 0.68 (0.55-0.85) | 0.74 (0.59-0.92) |
| A | C | 36 (1.6) | 28 (1.24) | 1.26 (0.77-2.08) | 1.71 (1.02-2.87) |
| A | G | 15 (0.7) | 28 (1.24) | 0.53 (0.28-0.99) | 0.70 (0.36-1.34) |
Data presented as no. (%) of total chromosomes for cases (total chromosomes, N = 2,220) and controls (total chromosomes, N = 2,258).
Two-sided χ2 test for haplotype frequency distributions between the cases and controls was 16.93 with 3 degrees of freedom (p=.001).
Adjusted for age, sex, smoking and drinking status.
Finally, the FPRP values at different prior probability levels for all significant findings are summarized in Table 5. For the prior probability .01, assuming that the OR for specific genotype was 0.67 (if a protection) or 1.56 (if a risk), with a statistical power of .548 and .421, FPRP were .098 and .190, respectively, for the association of the combined ADH7 CG+GG genotypes with a reduced risk of SCCHN in all subjects and male subjects. The FPRP value for an association between the haplotype “GG” and risk of SCCHN was .103 with a power of .601. These three significant associations were considered noteworthy findings in the present study, because the probabilities of falsely positive were less than 20%. In contrast, greater FPRP values were observed for the other significant associations between ADH variants and SCCHN, suggesting some possible bias in the findings.
Table 5.
FPRP Values for Associations Between risk of SCCHN and Frequencies of Genotypes and Haplotypes of ADH1B and ADH7
| Genotype / Haplotype | Positive OR (95% CI*) |
Statistical power‡ | Prior probability | |||||
|---|---|---|---|---|---|---|---|---|
| P† | .25 | .1 | .01 | .001 | .0001 | |||
| ADH7 A92G rs1573496: C>G | ||||||||
| GG vs. CC | ||||||||
| All subjects | 0.27 (0.11-0.68) | .000 | .007 | .115 | .281 | .812 | .978 | .998 |
| GG+CG vs. CC | ||||||||
| All subjects | 0.68 (0.54-0.85) | .001 | .548 | .003 | .010 | .098 | .522 | .916 |
| AGE ≥ 57 | 0.63 (0.46-0.87) | .005 | .373 | .039 | .110 | .575 | .932 | .993 |
| Male | 0.65 (0.50-0.84) | .001 | .421 | .007 | .021 | .190 | .703 | .960 |
| Former smoking | 0.66 (0.45-0.96) | .027 | .461 | .151 | .349 | .855 | .983 | .998 |
| N0 | 0.56 (0.40-0.78) | .001 | .152 | .010 | .029 | .246 | .767 | .971 |
| Oral cavity | 0.57 (0.40-0.82) | .002 | .197 | .028 | .080 | .489 | .906 | .990 |
| Haplotype§ | ||||||||
| A-C vs. G-C | 0.53 (0.28-0.99) | .001 | .020 | .095 | .240 | .776 | .972 | .997 |
| G-G vs. G-C | 0.68 (0.55-0.85) | .001 | .601 | .003 | .010 | .103 | .538 | .921 |
The crude OR.
The omnibus chi-square test of the genotype and haplotype frequency distributions.
Statistical power was calculated using the number of observations in the subgroup and the OR and P values in this table.
Haplotypes constructed as the order of ADH1B R48H and ADH7 A92G
Discussion
In this hospital-based case-control study, we investigated the associations between two reportedly important SNPs, ADH1B R48H (rs1229984: G>A) and ADH7 A92G (rs1573496: C>G), and risk of SCCHN in a non-Hispanic white population. We found that an overall reduced risk of SCCHN was associated with the GG or the combined CG+GG genotypes of the ADH7 A92G SNP, and this reduced risk was also evident in subgroups of older and male subjects, former smokers, oral cavity cancer and tumors without metastasis. These results are consistent with that in the previously reported large European study of the UADT cancers.18 Although we did not find an overall association between ADH1B R48H SNP and risk of SCCHN, the “GG” haplotype of these two SNPs appeared to be associated with a significantly lower SCCHN risk (mostly the effect of the ADH7 G allele) but the “AC” haplotype was associated with a significantly higher SCCHN risk (mostly the effect of the ADH1B A allele), compared with the most common “GC” haplotype.
The ADH1B R48H AA genotype encodes an enzyme whose homodimers have about 40 times and more active than the enzymes encoded by the AG and GG genotype,4 but the AA genotype is common in Asian but rare in European populations,15, 24 as most of studies for the effects of ADH1B polymorphisms on cancer were in Asian populations. For example, esophageal squamous cell carcinoma risk increased as the numbers of ADH1B R48H G alleles increased in Taiwanese,7 and the GG genotype was associated with more than 5-fold increased SCCHN risk among moderate-to-heavy drinkers in Japanese men.25 Nevertheless, several later studies in European populations also identified the ADH1B GG genotype was associated with an increased risk of the UADT cancers with an interaction with alcohol consumption,10, 18 although there was no association between the GG genotype and laryngeal cancer risk, even taking into account alcohol consumption, in a German population.26
Our study did not observe an overall association between ADH1B R48H genotypes and risk of SCCHN, nor in different subgroups of alcohol use. The discrepancy could result from several factors. First, the ADH1B R48H A allele frequency was rare in our study (.023 for cases and .025 for controls). Second, the effect of ADH1B variants on SCCHN was striking only in the moderate and high levels of alcohol consumption,27 but there were more never drinkers in our study than in other studies.28 It is conceivable that the rare ADH1B R48H A allele with low alcohol consumption did not allow us to detect the potential effect of this SNP on SCCHN risk. However, the present study did validate the association between the more frequent ADH7 A92G SNP (the G allele frequency: .076 in cases and .110 in controls) and SCCHN risk in our study population, consistent with the result from a large European multicenter case-control.18 Studies have found that the ADH7 SNPs were associated with alcoholism,29 which appears to be independent from a causative factor in ADH1B.30 Nevertheless, we have not observed any associations that were modulated by the alcohol use status, although the effect of ADH7 A92G has been reportedly more striking in moderate and heavy alcohol drinkers.18 It is likely that the former and current drinkerss in our study may have included both light and heavy drinkers, which may weaken the association between ADH7 A92G and alcohol use.
The ADH7 gene is only expressed in the epithelial tissues of the upper gastrointestinal tract, down to the stomach, as well as in the eyes, whereas other ADH genes are mainly expressed in the liver,31, 32 and ADH7 has the highest maximal activity for ethanol among the ADHs.33 These suggest that ADH7 is probably the first of the ADHs in metabolizing the ingested alcohol before alcohol has been absorbed into the blood. A recent in vivo study also indentified that alleles of selected ADH7 SNPs influenced the early stages of alcohol metabolism 34. Since the upper gastrointestinal tract is the first organ to ingest ethanol, it would be expected that there is a causal association between variants in ADH7 and risk of SCCHN.
There are two main polarized haplotype blocks that have been identified in the ADH7 gene region, which include sequences flanking the 5′ and 3′ ends of the ADH7 transcription unit.35 The SNP A92G locus is located in the 5′ haplotype block of the ADH7 gene. Therefore, our finding suggests that the effect of this SNP on risk of SCCHN may also be due to other SNPs in high LD with it. Since LD between ADH7 and other ADH genes was low36, it is likely that there are other potential causal polymorphisms in the ADH7 gene, especially in the 5′ haplotype block of ADH7.
There are several limitations in our study. First, although sample size in our study was relatively large, it still not large enough to identify the possible effect of alcohol-using pattern on the association. Second, the present study is a hospital-based case-control study; the genotype frequency in the controls may not represent the true frequency in the general population because of potential selection bias. It may have led to the deviation from HWE in the controls by the genotype frequency of ADH7 A92G, which may weaken any possible interaction between variants in ADH and smoking and alcohol use. These limitations should be overcome in future large, well designed prospective studies. In summary, we have confirmed that the G allele of ADH7 A92G (rs1573496: C>G) SNP is associated with a reduced risk of SCCHN in non-Hispanic Whites. More studies with multi-ethnic groups on the role of ADH7 SNPs in the etiology of SCCHN are warranted.
Acknowledgments
We thank Margaret Lung and Kathryn Patterson for their assistance in recruiting the subjects and gathering the questionnaire information; Yawei Qiao, Zhaozheng Guo, Min Zhao for technical assistance; Jianzhong He and Kejin Xu for their laboratory assistance.
Grant support: National Institutes of Health grants R01 ES11740-07 (PI, Q.W.), R01 CA131274 -01 (PI, Q.W.), P50 CA97007-07 (PI, S.L.) and P30 CA16672 (M. D. Anderson Cancer Center).
Abbreviations
- ADH
alcohol dehydrogenase
- SNP
single nucleotide polymorphism
- SCCHN
squamous cell carcinoma of the head and neck
- OR
odds ratio
- CI
confidence interval
References
- 1.Parkin DM, Bray F, Ferlay J, Pisani P. Global cancer statistics, 2002. CA Cancer J Clin. 2005;55:74–108. doi: 10.3322/canjclin.55.2.74. [DOI] [PubMed] [Google Scholar]
- 2.Sturgis EM, Wei Q, Spitz MR. Descriptive epidemiology and risk factors for head and neck cancer. Semin Oncol. 2004;31:726–33. doi: 10.1053/j.seminoncol.2004.09.013. [DOI] [PubMed] [Google Scholar]
- 3.Seitz HK, Stickel F. Molecular mechanisms of alcohol-mediated carcinogenesis. Nat Rev Cancer. 2007;7:599–612. doi: 10.1038/nrc2191. [DOI] [PubMed] [Google Scholar]
- 4.Seitz HK, Becker P. Alcohol metabolism and cancer risk. Alcohol Res Health. 2007;30:38–41. 44–7. [PMC free article] [PubMed] [Google Scholar]
- 5.Holmes RS. Alcohol dehydrogenases: a family of isozymes with differential functions. Alcohol Alcohol Suppl. 1994;2:127–30. [PubMed] [Google Scholar]
- 6.Li H, Gu S, Cai X, Speed WC, Pakstis AJ, Golub EI, et al. Ethnic related selection for an ADH Class I variant within East Asia. PLoS ONE. 2008;3:e1881. doi: 10.1371/journal.pone.0001881. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Lee CH, Lee JM, Wu DC, Goan YG, Chou SH, Wu IC, et al. Carcinogenetic impact of ADH1B and ALDH2 genes on squamous cell carcinoma risk of the esophagus with regard to the consumption of alcohol, tobacco and betel quid. Int J Cancer. 2008;122:1347–56. doi: 10.1002/ijc.23264. [DOI] [PubMed] [Google Scholar]
- 8.Hiraki A, Matsuo K, Wakai K, Suzuki T, Hasegawa Y, Tajima K. Gene-gene and gene-environment interactions between alcohol drinking habit and polymorphisms in alcohol-metabolizing enzyme genes and the risk of head and neck cancer in Japan. Cancer Sci. 2007;98:1087–91. doi: 10.1111/j.1349-7006.2007.00505.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Yokoyama A, Kato H, Yokoyama T, Tsujinaka T, Muto M, Omori T, et al. Genetic polymorphisms of alcohol and aldehyde dehydrogenases and glutathione S-transferase M1 and drinking, smoking, and diet in Japanese men with esophageal squamous cell carcinoma. Carcinogenesis. 2002;23:1851–9. doi: 10.1093/carcin/23.11.1851. [DOI] [PubMed] [Google Scholar]
- 10.Hashibe M, Boffetta P, Zaridze D, Shangina O, Szeszenia-Dabrowska N, Mates D, et al. Evidence for an important role of alcohol- and aldehyde-metabolizing genes in cancers of the upper aerodigestive tract. Cancer Epidemiol Biomarkers Prev. 2006;15:696–703. doi: 10.1158/1055-9965.EPI-05-0710. [DOI] [PubMed] [Google Scholar]
- 11.Peters ES, McClean MD, Liu M, Eisen EA, Mueller N, Kelsey KT. The ADH1C polymorphism modifies the risk of squamous cell carcinoma of the head and neck associated with alcohol and tobacco use. Cancer Epidemiol Biomarkers Prev. 2005;14:476–82. doi: 10.1158/1055-9965.EPI-04-0431. [DOI] [PubMed] [Google Scholar]
- 12.Brennan P, Lewis S, Hashibe M, Bell DA, Boffetta P, Bouchardy C, et al. Pooled analysis of alcohol dehydrogenase genotypes and head and neck cancer: a HuGE review. Am J Epidemiol. 2004;159:1–16. doi: 10.1093/aje/kwh003. [DOI] [PubMed] [Google Scholar]
- 13.Wang D, Ritchie JM, Smith EM, Zhang Z, Turek LP, Haugen TH. Alcohol dehydrogenase 3 and risk of squamous cell carcinomas of the head and neck. Cancer Epidemiol Biomarkers Prev. 2005;14:626–32. doi: 10.1158/1055-9965.EPI-04-0343. [DOI] [PubMed] [Google Scholar]
- 14.Homann N, Stickel F, Konig IR, Jacobs A, Junghanns K, Benesova M, et al. Alcohol dehydrogenase 1C*1 allele is a genetic marker for alcohol-associated cancer in heavy drinkers. Int J Cancer. 2006;118:1998–2002. doi: 10.1002/ijc.21583. [DOI] [PubMed] [Google Scholar]
- 15.Osier MV, Pakstis AJ, Soodyall H, Comas D, Goldman D, Odunsi A, et al. A global perspective on genetic variation at the ADH genes reveals unusual patterns of linkage disequilibrium and diversity. Am J Hum Genet. 2002;71:84–99. doi: 10.1086/341290. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Chen CC, Lu RB, Chen YC, Wang MF, Chang YC, Li TK, et al. Interaction between the functional polymorphisms of the alcohol-metabolism genes in protection against alcoholism. Am J Hum Genet. 1999;65:795–807. doi: 10.1086/302540. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Osier M, Pakstis AJ, Kidd JR, Lee JF, Yin SJ, Ko HC, et al. Linkage disequilibrium at the ADH2 and ADH3 loci and risk of alcoholism. Am J Hum Genet. 1999;64:1147–57. doi: 10.1086/302317. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Hashibe M, McKay JD, Curado MP, Oliveira JC, Koifman S, Koifman R, et al. Multiple ADH genes are associated with upper aerodigestive cancers. Nat Genet. 2008;40:707–9. doi: 10.1038/ng.151. [DOI] [PubMed] [Google Scholar]
- 19.Edenberg HJ. The genetics of alcohol metabolism: role of alcohol dehydrogenase and aldehyde dehydrogenase variants. Alcohol Res Health. 2007;30:5–13. [PMC free article] [PubMed] [Google Scholar]
- 20.Liu Z, Calderon JI, Zhang Z, Sturgis EM, Spitz MR, Wei Q. Polymorphisms of vitamin D receptor gene protect against the risk of head and neck cancer. Pharmacogenet Genomics. 2005;15:159–65. doi: 10.1097/01213011-200503000-00004. [DOI] [PubMed] [Google Scholar]
- 21.Li D, Wang LE, Chang P, El-Naggar AK, Sturgis EM, Wei Q. In vitro benzo[a]pyrene diol epoxide-induced DNA adducts and risk of squamous cell carcinoma of head and neck. Cancer Res. 2007;67:5628–34. doi: 10.1158/0008-5472.CAN-07-0983. [DOI] [PubMed] [Google Scholar]
- 22.Lippincott JB, editor. Cancer AJCO. Manual for Staging of Cancer. 6th. Springer; Philadelphia: 2002. [Google Scholar]
- 23.Wacholder S, Chanock S, Garcia-Closas M, El Ghormli L, Rothman N. Assessing the probability that a positive report is false: an approach for molecular epidemiology studies. J Natl Cancer Inst. 2004;96:434–42. doi: 10.1093/jnci/djh075. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Li H, Mukherjee N, Soundararajan U, Tarnok Z, Barta C, Khaliq S, et al. Geographically separate increases in the frequency of the derived ADH1B*47His allele in eastern and western Asia. Am J Hum Genet. 2007;81:842–6. doi: 10.1086/521201. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Asakage T, Yokoyama A, Haneda T, Yamazaki M, Muto M, Yokoyama T, et al. Genetic polymorphisms of alcohol and aldehyde dehydrogenases, and drinking, smoking and diet in Japanese men with oral and pharyngeal squamous cell carcinoma. Carcinogenesis. 2007;28:865–74. doi: 10.1093/carcin/bgl206. [DOI] [PubMed] [Google Scholar]
- 26.Risch A, Ramroth H, Raedts V, Rajaee-Behbahani N, Schmezer P, Bartsch H, et al. Laryngeal cancer risk in Caucasians is associated with alcohol and tobacco consumption but not modified by genetic polymorphisms in class I alcohol dehydrogenases ADH1B and ADH1C, and glutathione-S-transferases GSTM1 and GSTT1. Pharmacogenetics. 2003;13:225–30. doi: 10.1097/00008571-200304000-00007. [DOI] [PubMed] [Google Scholar]
- 27.Druesne-Pecollo N, Tehard B, Mallet Y, Gerber M, Norat T, Hercberg S, et al. Alcohol and genetic polymorphisms: effect on risk of alcohol-related cancer. Lancet Oncol. 2009;10:173–80. doi: 10.1016/S1470-2045(09)70019-1. [DOI] [PubMed] [Google Scholar]
- 28.Hashibe M, Brennan P, Benhamou S, Castellsague X, Chen C, Curado MP, et al. Alcohol drinking in never users of tobacco, cigarette smoking in never drinkers, and the risk of head and neck cancer: pooled analysis in the International Head and Neck Cancer Epidemiology Consortium. J Natl Cancer Inst. 2007;99:777–89. doi: 10.1093/jnci/djk179. [DOI] [PubMed] [Google Scholar]
- 29.Osier MV, Lu RB, Pakstis AJ, Kidd JR, Huang SY, Kidd KK. Possible epistatic role of ADH7 in the protection against alcoholism. Am J Med Genet B Neuropsychiatr Genet. 2004;126B:19–22. doi: 10.1002/ajmg.b.20136. [DOI] [PubMed] [Google Scholar]
- 30.Han Y, Oota H, Osier MV, Pakstis AJ, Speed WC, Odunsi A, et al. Considerable haplotype diversity within the 23kb encompassing the ADH7 gene. Alcohol Clin Exp Res. 2005;29:2091–100. doi: 10.1097/01.alc.0000191769.92667.04. [DOI] [PubMed] [Google Scholar]
- 31.Engeland K, Maret W. Extrahepatic, differential expression of four classes of human alcohol dehydrogenase. Biochem Biophys Res Commun. 1993;193:47–53. doi: 10.1006/bbrc.1993.1588. [DOI] [PubMed] [Google Scholar]
- 32.Yin SJ, Chou FJ, Chao SF, Tsai SF, Liao CS, Wang SL, et al. Alcohol and aldehyde dehydrogenases in human esophagus: comparison with the stomach enzyme activities. Alcohol Clin Exp Res. 1993;17:376–81. doi: 10.1111/j.1530-0277.1993.tb00779.x. [DOI] [PubMed] [Google Scholar]
- 33.Pares X, Cederlund E, Moreno A, Saubi N, Hoog JO, Jornvall H. Class IV alcohol dehydrogenase (the gastric enzyme). Structural analysis of human sigma sigma-ADH reveals class IV to be variable and confirms the presence of a fifth mammalian alcohol dehydrogenase class. FEBS Lett. 1992;303:69–72. doi: 10.1016/0014-5793(92)80479-z. [DOI] [PubMed] [Google Scholar]
- 34.Birley AJ, James MR, Dickson PA, Montgomery GW, Heath AC, Martin NG, et al. Adh Snp Associations with Alcohol Metabolism in Vivo. Hum Mol Genet. 2009 doi: 10.1093/hmg/ddp060. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Birley AJ, James MR, Dickson PA, Montgomery GW, Heath AC, Whitfield JB, et al. Association of the gastric alcohol dehydrogenase gene ADH7 with variation in alcohol metabolism. Hum Mol Genet. 2008;17:179–89. doi: 10.1093/hmg/ddm295. [DOI] [PubMed] [Google Scholar]
- 36.Edenberg HJ, Xuei X, Chen HJ, Tian H, Wetherill LF, Dick DM, et al. Association of alcohol dehydrogenase genes with alcohol dependence: a comprehensive analysis. Hum Mol Genet. 2006;15:1539–49. doi: 10.1093/hmg/ddl073. [DOI] [PubMed] [Google Scholar]