Purpose: Folate deficiency and reduced DNA repair capacity are established risk factors for squamous cell carcinoma of the head and neck (SCCHN). We hypothesized that polymorphisms of the thymidylate synthase (TYMS) gene, which regulates a key enzyme in folate metabolism required for DNA synthesis and repair, are associated with SCCHN risk.
Experimental Design: In a hospital-based case-control study of 704 SCCHN cases and 1,085 controls, frequency matched by age, sex, and ethnicity, we genotyped the TSER (thymidylate synthase in the 5′-untranslated enhanced region) and TS3′UTR (thymidylate synthase in the 3′-untranslated region) polymorphisms.
Results: The TS3′UTR 0bp/0bp genotype was associated with a significantly decreased risk of SCCHN [adjusted odd ratio (OR) = 0.67, 95% confidence interval (CI) = 0.47–0.94] compared with the 6bp/6bp genotype, but the TSER polymorphism had no main effect on risk of SCCHN. When we evaluated the two polymorphisms together by the number of protective alleles (the TSER 3R and TS3′UTR 0bp alleles), we found that the combined genotypes with four protective alleles (the TSER 3R3R and TS3′UTR 0bp/0bp) was associated with significantly decreased SCCHN risk (OR = 0.60, 95% CI = 0.37–0.98). In addition, the TS3′UTR 0bp genotypes were associated in an allele dose-dependent manner with a decreased risk of overall stage IV oral cancer (OR = 0.84, 95% CI = 0.52–1.34 for the 6bp/0bp genotype and OR = 0.26, 95% CI = 0.08–0.87 for the 0bp/0bp genotype; Ptrend = 0.035).
Conclusion: The TSER and TS3′UTR polymorphisms are associated with SCCHN risk. The TSER 3R and TS3′UTR 0bp alleles seemed to jointly protect against SCCHN. In particular, the 0bp allele seemed to protect against oral cancer progression.
Squamous cell carcinoma of the head and neck (SCCHN), which includes cancers of the oral cavity, pharynx, and larynx, is the sixth most common cancer and the seventh leading cause of cancer-related death in the world, affecting >500,000 individuals each year worldwide (1) . In the United States, it was estimated that there would be 38,530 newly diagnosed SCCHN cases and 11,060 SCCHN-related deaths in 2004 (2) . Although tobacco smoking and alcohol use are the primary risk factors for SCCHN (3) , only a fraction of individuals exposed to tobacco or alcohol develops SCCHN, suggesting that individual susceptibility to exposure-related carcinogenesis varies.
Low dietary intake of fruits and vegetables has been implicated in SCCHN risk (4 , 5) . Folate, as one of the constituents in fruits and vegetables, may provide protection against SCCHN (5) . Several key enzymes, including thymidylate synthase (TYMS; EC18.104.22.168), are known to be involved in folate metabolism (6 , 7) . The TYMS gene, which is located at chromosome 18p11.32, catalyzes the conversion of dUMP to dTMP using the 5, 10-methylenetetrahydrofolate as a methyl donor (7) . This reaction is the de novo source of cellular thymidine, which is essential for the provision of nucleotides required for DNA synthesis and repair. It is also a primary target for 5-fluorouracil, the most common chemotherapy agent used to treat many cancers including SCCHN (8) . Thymidylate stress may cause chromosomal breakage and fragile sites induction (9 , 10) , which may ultimately modulate susceptibility to cancer.
The TYMS gene has a promoter enhancer region (TSER, thymidylate synthase in the 5′-untranslated enhancer region) polymorphism in the 5′-untranslated region, containing either double (2R) or triple (3R) tandem repeats of a 28 base-pair (bp) sequence (11) . In vitro and in vivo studies have revealed that the number of the tandem repeats can affect TYMS expression activity and protein translation efficiency (11, 12, 13) . TSER is considered an important functional polymorphism of the TYMS gene, because it modulates plasma folate and homocysteine levels (14 , 15) , suggesting that the TSER polymorphism may be a risk factor of cancer. Several studies have investigated the association between the TSER polymorphism and risk of cancer, but the results were mixed (16, 17, 18, 19) .
Recently, a novel polymorphism of the TYMS gene, a 6-bp deletion/insertion at bp 1494 in the 3′-untranslated region (TS3′UTR, thymidylate synthase in the 3′-untranslated region) was identified (6) and is thought to influence TYMS mRNA stability and tumor mRNA levels (6 , 20) . Several studies reported that the TS3′UTR polymorphism was not associated with risk of cancer (15 , 18 , 21) ; however, there seemed to be a significant combined effect of the TSER and TS3′UTR polymorphisms on risk of colorectal cancer (21) .
Because folate is involved in nucleotide synthesis (a process required for DNA repair), low folate intake is associated with suboptimal DNA repair capacity (22) , and suboptimal DNA repair capacity is a risk factor for SCCHN (23) , we hypothesized that the TSER and TS3′UTR polymorphisms are associated with SCCHN risk. We tested this hypothesis in an ongoing, hospital-based, case-control study of SCCHN.
MATERIALS AND METHODS
The subject recruitment for the ongoing SCCHN study has been described previously (24) . Briefly, 704 newly diagnosed SCCHN cases and 1,085 cancer-free controls were recruited into the study between May 1995 and September 2003. All of the cases and 663 of the controls were recruited from The University of Texas M. D. Anderson Cancer Center in Houston, Texas. These controls were recruited from among biologically unrelated visitors who were accompanying the cases to the hospital. There were 422 additional controls who were recruited from a multispecialty physician practice, the Kelsey-Seybold Clinic, which has multiple clinics throughout the Houston, Texas, metropolitan area. These additional controls were older men who were former and current smokers, which we needed for frequency matching to our cases. Because genotype frequencies can vary among ethnic groups and few minority patients were recruited, only non-Hispanic white patients and controls were included in this analysis.
Approximately 95% of eligible patients who were contacted chose to participate. The 704 SCCHN patients with primary tumors included in the analysis had cancer of the oral cavity (n = 214; 30.4%), pharynx (n = 355; 50.4%), or larynx (n = 135; 19.2%). Patients with second SCCHN primary tumors, primary tumors of the nasopharynx or sinonasal tract, primary tumors outside the upper aerodigestive tract, cervical metastases of unknown origin, or histopathologic diagnoses other than squamous cell carcinoma were excluded. The overall stage of SCCHN was defined according to the American Joint Committee for Cancer Staging and End-Results reporting, 1992: overall stage I (T1N0M0), overall stage II (T2N0M0), overall stage III (T3N0M0 or T1–3N1M0), and overall stage IV (T4N0–1M0 or any TN2–3M0 or any T and N, M1). T, N, and M represent tumor stage, nodal status, and distant metastases, respectively. M was M0 for all cases in this study.
We first surveyed the potential control subjects by using a short questionnaire to determine their willingness to participate in research studies and to obtain information about their smoking behavior and demographic factors. Of the respondents we contacted for recruitment, the response rate was >80%. We interviewed each eligible subject to obtain data on age, sex, smoking status, and alcohol use. We frequency matched the controls to the cases by age (±5 years), sex, and smoking status. Those subjects who had smoked >100 cigarettes in their lifetimes were defined as ever smokers. Ever smokers who had quit smoking >1 year previously were defined as former smokers and the other smokers as current smokers. After the subjects signed informed consent forms, each subject donated 30 mL of blood, which was collected into a heparinized tube. A leukocyte cell pellet obtained from a buffy coat was used for DNA extraction with a DNA blood mini kit (Qiagen, Valencia, CA) according to the manufacturer’s instructions. The research protocol was approved by The University of Texas M. D. Anderson Cancer Center and Kelsey-Seybold Clinic institutional review boards.
PCR and PCR-based restriction fragment length polymorphism assays were used to identify the TSER and TS3′UTR polymorphisms, respectively. The primers of the TSER polymorphism were 5′-GTGGCTCCTGCGTTTCCCCC-3′ (forward) and 5′-GGCTCCGAGCCGGCCACAGGCATGGCGCGG-3′ (reverse; ref. 12 ), which generated 243-bp (triple repeats = 3R) and 215-bp (double repeats = 2R) fragments. The primers of the TS3′UTR polymorphism were 5′-CAAATCTGAGGGAGCTGAGT-3′ (forward) and 5′-CAGATAAGTGGCAGTACAGA-3′ (reverse; ref. 6 ), which generated 152-bp fragment for 6-bp deletion (i.e., 0bp) or 158-bp for 6-bp insertion (i.e., 6bp) of the TS3′UTR polymorphism. The fragments of the TSER polymorphism were amplified in 20 μL of reaction mixture containing approximately 50 ng of genomic DNA template, 12.5 pmol of each primer, 10% DMSO, 0.1 mmol/L of each deoxynucleoside triphosphate, 1 × PCR buffer (50 mmol/L KCl, 10 mmol/L Tris-HCl, and 0.1% Triton X-100), 1.5 mmol/L MgCl2, and 1.5 units of Taq polymerase (Sigma-Aldrich Corp., St. Louis, MO). The fragments of the TS3′UTR polymorphism were amplified in the same cycling condition as that of the TSER polymorphism but without 10% dimethyl sulfoxide. The PCR amplification parameters of these two polymorphisms were as follows: 5-minute denaturation cycle at 95°C; 35 cycles of 95°C for 30 seconds, 63°C for 45 seconds, and 72°C for 1 minute; and a final extension at 72°C for 10 minutes.
The restriction enzyme DraI (New England BioLabs, Inc., Beverly, MA) was used to distinguish the TS3′UTR polymorphism, in which the presence of the 6-bp insertion creates a DraI restriction site, and the expected fragment sizes were 88 bp and 70 bp (Fig. 1B)⇓ . The 158-bp band in the heterozygous TS3′UTR 6bp/0bp is caused by an undigested wild-type fragment (6) . The PCR-amplified fragments of the TSER polymorphism and the digestion products of the TS3′UTR polymorphism were separated on 3% NuSieve 3:1 agarose gel (Cambrex Bio Science Rockland, Inc., Rockland, ME; Fig. 1A and B⇓ ). More than 10% of the samples were randomly selected for repeated assays, and the results were 100% concordant.
We used the χ2 test to evaluate differences between the cases and controls in the frequency distributions of selected demographic variables, smoking status, alcohol use, and each allele and genotype of the TSER and TS3′UTR polymorphisms. Univariate and multivariate logistic regression analyses were used to obtain the crude and adjusted odds ratios (ORs) and 95% confidence intervals (CIs). The multivariate adjustment included age, sex, smoking status, and alcohol use. The genotype data were further stratified by subgroups of age, sex, smoking status, and alcohol use. Considering potential interactions between the TSER and TS3′UTR polymorphisms on risk of SCCHN, we evaluated any potential association between SCCHN risk and the combined genotypes of these two polymorphisms. We determined all tests of statistical significance two sided at the level of P < 0.05 by using SAS software, version 8.2 (SAS Institute, Inc., Cary, NC).
Characteristics of the Study Population.
The frequency distributions of selected characteristics of the cases and controls are presented in Table 1⇓ . The mean age was 57 years (±11.8; range, 18–90 years) for the cases and 56.7 years (±11.5; range, 20–87 years) for the controls, and the difference was not statistically significant (P = 0.122). The distribution of sex was also similar between the cases and controls (P = 0.26). However, the frequency matching on smoking status was imperfect: i.e., there were more current smokers (34.7%) and current drinkers (51.3%) among the cases than among the controls (26.3 and 44.4%, respectively), and these differences were statistically significant (P < 0.001 for both tobacco and alcohol use; Table 1⇓ ). Therefore, these variables were further adjusted for in the logistic regression analysis.
Genotype Distributions of TYMS Polymorphisms among the Cases and Controls.
The genotype and allele frequencies of TSER and TS3′UTR polymorphisms are summarized in Table 2⇓ . Because the controls were selected from different sources, the distributions of the genotypes were compared between the cases and three groups of controls: the M. D. Anderson Cancer Center, the Kelsey-Seybold Clinic, and the two groups combined. We did not find any differences in the genotype distributions between the M. D. Anderson Cancer Center and the Kelsey-Seybold Clinic control groups (P = 0.883 for the TSER and P = 0.395 for TS3′UTR; data not shown). Nevertheless, we present the results from the analyses of the combined controls as well as the M. D. Anderson Cancer Center and the Kelsey-Seybold Clinic controls separately.
In this study, we identified three cases (0.4%) and four controls (0.4%) with the TSER 4R allele (data not shown, Fig. 1A⇓ ). But because they were rare, these seven subjects with the 4R allele were not included in the final analyses so that our results would be comparable with those of previously published reports (18) .
As shown in Table 2⇓ , the genotype frequencies of the TSER polymorphism were 26.1, 53.2, and 20.7% for 3R3R, 2R3R, and 2R2R, respectively, among the cases and 28.8, 48.5, and 22.7%, respectively, among the controls. This difference was not statistically significant (P = 0.158). The difference was not significant for M. D. Anderson controls (P = 0.256) or for Kelsey-Seybold controls (P = 0.258). The genotype frequencies of the TS3′UTR polymorphism were 48.1, 44.2, and 7.7% for the 6bp/6bp, 6bp/0bp, and 0bp/0bp, respectively, among the cases and 47.7, 41.1, and 11.2%, respectively, among the controls, and the difference was statistically significant (P = 0.039). However, a similar difference was observed for the Kelsey-Seybold controls (P = 0.023) but not for the M. D. Anderson controls (P = 0.15). This may be because the Kelsey-Seybold controls were selected based on the need for older male smokers for our frequency matching. For part of the study design, it was appropriate to combine the Kelsey-Seybold controls with the M. D. Anderson controls to reduce possible selection bias as well as to increase statistical power for further comparisons with the cases.
In the present study, the TSER 2R and TS3′UTR 0bp allele frequencies were 0.473 and 0.298, respectively, among the 704 cases and 0.469 and 0.318, respectively, among the 1,085 controls; these differences were not statistically significant (P = 0.84 for the 2R; P = 0.318 for the 0bp allele; Table 2⇓ ). The genotype distributions of the TSER and TS3′UTR polymorphisms among subjects in the combined control group were in agreement with the Hardy-Weinberg equilibrium (χ2 test: P = 0.379 for the TSER, P = 0.085 for the TS3′UTR).
Associations and the Stratification Analysis of TYMS Polymorphisms and SCCHN Risk.
When we used the TSER 3R3R genotype as the reference, the 2R3R genotype was associated with a borderline increased risk of SCCHN [adjusted odds ratio (OR) = 1.23, 95% confidence interval (CI) = 0.98–1.55]; however, there was no significant association between the 2R2R genotype and SCCHN risk (OR = 1.01, 95% CI = 0.77–1.33). When we used the TS3′UTR 6bp/6bp genotype as the reference, the 0bp/0bp, but not the 6bp/0bp genotype, was associated with a statistically significantly decreased risk of SCCHN (adjusted OR = 0.67, 95% CI = 0.47–0.94 for the 0bp/0bp; OR = 1.07, 95% CI = 0.87–1.31 for the 6bp/0bp; Table 2⇓ ). These risk estimates were very close to those obtained from comparisons with either the Kelsey-Seybold or the M. D. Anderson control groups, but the combined controls provided a much greater study power, which allowed for further stratification analyses (Table 3)⇓ ⇓ .
In the stratification analysis for the TSER polymorphism, a significantly increased risk of SCCHN was associated with the 2R3R genotype and was confined to never smokers (adjusted OR = 1.59, 95% CI = 1.03–2.46) compared with the 3R3R genotype. For the TS3′UTR polymorphism, a significantly decreased risk of SCCHN was associated with the 0bp/0bp genotype and was confined to older subjects (>55 years; adjusted OR = 0.52, 95% CI = 0.32–0.84) and never smokers (OR = 0.47, 95% CI = 0.22–0.99), compared with the 6bp/6bp genotype. When stratified by tumor site, only the TSER 2R3R genotype was associated with a significantly increased risk of squamous cell carcinoma of the pharynx (adjusted OR = 1.34, 95% CI = 1.00–1.80), compared with the 3R3R genotype. For the TS3′UTR polymorphism, although there were more 6bp/6bp genotypes among subjects with oral cancer (53.7%) than among subjects with pharyngeal and laryngeal cancer (46.2 and 44.4%, respectively), there was no significant association between the TS3′UTR polymorphism and SCCHN risk for these three tumor sites (Table 3)⇓ ⇓ . These findings are limited because of the reduced number of observations in each subgroup analyzed.
Combined Analysis of Association between the Two TYMS Polymorphisms and SCCHN Risk.
Among the 1,085 controls, the TSER and TS3′UTR polymorphisms were in linkage disequilibrium (P < 0.001). Although all genotype combinations were observed, 32.3% of subjects with the TSER 3R3R genotype had the TS3′UTR 6bp/6bp genotype and 25.6% had the TS3′UTR 0bp/0bp genotype; however, the TSER 3R3R genotype was not as associated with SCCHN risk as the TS3′UTR 6bp/6bp genotype was, suggesting incomplete disequilibrium between these two polymorphisms. On the basis of the number of protective alleles of the TSER and TS3′UTR polymorphism (the TSER 3R and TS3′UTR 0bp alleles), we dichotomized the individuals into five genotype groups: (a) the TSER 2R2R and TR3′UTR 6bp/6bp (0 protective allele of either gene); (b) the TSER 2R3R and TS3′UTR 6bp/6bp or TSER 2R2R and TS3′UTR 6bp/0bp (only 1 protective allele); (c) TSER 3R3R and TS3′UTR 6bp/6bp or TSER 2R3R and TS3′UTR 6bp/0bp or TSER 2R2R and TS3′UTR 0bp/0bp (2 protective alleles); (d) TSER 3R3R and TS3′UTR 6bp/0bp or TSER 2R3R and TS3′UTR 0bp/0bp (3 protective alleles); and (e) TSER 3R3R and TS3′UTR 0bp/0bp (4 protective alleles). As shown in Table 4⇓ , the frequencies of the combined genotypes were 15.5, 29.7, 33.5, 17.0, and 4.3% for the groups with zero, one, two, three, and four protective alleles, respectively, among the cases and 15.9, 28.6, 32.6, 9.4, 15.5, and 7.4%, respectively, among the controls. However, the difference was not statistically significant (P = 0.102). Compared with the group with zero protective alleles (the 2R2R/6bp/6bp genotype), the group with four protective alleles (the 3R3R/0bp/0bp genotype) had a significantly decreased risk of SCCHN (adjusted OR = 0.60, 95% CI = 0.37–0.98) as assessed in the multivariate logistic regression analysis, but the trend of decreasing risk was not statistically significant (P = 0.343; Table 4⇓ ).
TSER and TS3′UTR Polymorphisms and Progression of SCCHN.
Because the TSER 2R3R and TS3′UTR 0bp/0bp genotypes were distributed differently by tumor site, we hypothesized that these genotypes may also be distributed differently by overall tumor stage (I to IV as described in “Materials and Methods”). Each stage level was related to a different tumor stage (T1–4), nodal status (N0–3), and distant metastases (M0–1) rating. There was no significant association between these two polymorphisms and overall stage of the pharyngeal and laryngeal cancers (because few subjects had these polymorphisms; data not shown). However, when we used the TS3′UTR 6bp/6bp as the reference, we found that the TS3′UTR variant genotypes (6bp/0bp and 0bp/0bp) were associated with a decreased risk of overall stage IV oral cancer (adjusted OR = 0.84, 95% CI = 0.52–1.34 for 6bp/0bp and adjusted OR = 0.26, 95% CI = 0.08–0.87 for 0bp/0bp) in an allele dose-dependent manner (Ptrend = 0.035). However, the trend of decreasing risk was not evident for the cases with an overall stage of I to III oral cavity cancer or for the TSER polymorphism (Table 5)⇓ .
In this study, we found that the TS3′UTR 0bp/0bp genotype was associated with a significantly decreased risk of SCCHN compared with the 6bp/6bp genotype, but the TSER polymorphism had no main effect on the risk of SCCHN. When we evaluated the two polymorphisms together by the number of protective alleles (the TSER 3R and TS3′UTR 0bp alleles), we found that the combined genotypes with four protective alleles (the TSER 3R3R and TS3′UTR 0bp/0bp) were associated with significantly decreased SCCHN risk. In addition, the TS3′UTR 0bp genotypes were associated in an allele dose-dependent manner with a decreased risk of overall stage IV oral cancer. These findings suggest that the TSER and TS3′UTR polymorphisms are associated with SCCHN risk. The TSER 3R and TS3′UTR 0bp alleles seemed to jointly protect against SCCHN. In particular, the 0bp allele seemed to protect against oral cancer progression.
Recent studies on the association between the TSER polymorphism and cancer risk found that individuals with the TSER 2R3R and 3R3R genotypes were protected against adult acute lymphocytic leukemia compared with those with the 2R2R genotype (17) . Similarly, individuals with at least one TSER 2R allele had a significantly increased risk of malignant lymphoma compared with those without the 2R allele (16) . However, other studies did not find a significant association between the TSER polymorphism and risk of colorectal adenomas (18 , 19) . In the present study, although the 2R3R genotype was associated with a borderline increased risk of SCCHN and with a significantly increased risk of pharyngeal cancer, the lack of an overall association between the TSER polymorphism and the risk of SCCHN diminishes the importance of any findings for this subgroup without supportive biological mechanisms.
The 3′-untranslated region of a gene is generally not translated into proteins, but it is thought to play an important role in maintaining mRNA stability (6 , 20) , which may indirectly affect protein expression. Several studies found that the TS3′UTR polymorphism was not associated with cancer risk (15 , 18 , 21) . In the present study, we found that the subjects with the TS3′UTR 0bp/0bp genotype had a 33% decreased risk for SCCHN than did those with the 6bp/6bp genotype, suggesting that the TS3′UTR polymorphism had a main effect on SCCHN risk. This decreased risk was more pronounced among older never smokers, which is not consistent with data on other cancers published previously (15 , 18 , 21) . Our data suggest that the TS3′UTR polymorphism plays an important role in the etiology of SCCHN. Because this is the first such report on SCCHN, larger and preferably prospective studies are needed to verify these findings.
Although the TSER polymorphism was in linkage disequilibrium with the TS3′UTR polymorphism, it was not independently associated with risk for SCCHN. This was described previously as “imperfect” disequilibrium (18) . However, our data suggest that the TSER polymorphism may interact with the TS3′UTR polymorphism in the etiology of SCCHN, because when the TSER and TS3′UTR polymorphisms were combined by the number of protective alleles, the SCCHN risk was significantly decreased, especially for the combined genotypes with four protective alleles. Thus, our data suggest that these two TYMS polymorphisms may jointly provide protection against SCCHN. Because the interactions among folate-related genes, folate, and related dietary factors in tumorigenesis are complex (7 , 25) , this possible gene-gene interaction and the mechanisms under which these polymorphisms affect risk of SCCHN warrant additional investigations.
Recent studies have reported that the TS3′UTR polymorphism may affect TYMS mRNA stability and tumor mRNA level (20) , and the 6bp/6bp genotype significantly increases the tumor mRNA level compared with the 0bp/0bp genotype (6) . Several studies have also shown that overexpression of the TYMS protein is associated with resistance to 5-fluorouracil–based treatment and, thus, may lead to poor survival (12 , 26 , 27) . In the present study, we found that the TS3′UTR 0bp genotype was associated with a decreased risk of overall stage IV oral cavity cancer in a dose-dependent manner. Because the sample size in the subgroup was relatively small and therefore had limited statistical power, this finding needs to be investigated in future studies.
A primary shortcoming of this study was the lack of detailed data on the alcohol and dietary intake of subjects as well as their serum levels of folate and the precursors or metabolites of folate such as homocysteine, because the effect of variations in folate metabolism genes on cancer risk depends on alcohol and folate intake status (18) . Because our study was hospital based, there are inherent limitations in our study design that could have introduced selection bias, compared with population-based or cohort studies. However, the genotype distributions in our study population were similar to reported distributions of other studies. For instance, the frequencies of the TSER 3R3R, 2R3R, and 2R2R genotypes among our 1,085 Texas non-Hispanic white controls were 28.8, 48.5, and 22.7%, respectively, compared with 26, 52, and 21%, respectively, of 625 Minnesota population-based controls, of whom 97% were Caucasians (18) and 29, 48, and 23%, respectively, of 454 Physicians’ Health Study cohort controls, of whom 93% were Caucasians (15) . Also, the frequencies of the TS′UTR 6bp/6bp, 6bp/0bp, and 0bp/0bp genotypes among our 1,085 controls were 47.7, 41.1, and 11.2%, respectively, compared with 50, 40, and 10%, respectively, of the 625 Minnesota controls (18) , and 44, 42, and 14%, respectively, of the 454 Physicians’ Health Study controls (15) . Because the TYMS genotype frequency estimates from our hospital-based controls are between those of population-based and cohort-based controls, the selection bias in genotype distribution, if any, is unlikely to be substantial.
In conclusion, we found a significant association between the TS3′UTR polymorphism and SCCHN risk in a relatively large, hospital-based, case-control study. The protective alleles of these two polymorphisms (the TSER 3R and TS3′UTR 0bp alleles) seemed to jointly protect against SCCHN; in particular, the TS3′UTR 0bp allele may protect against overall stage IV oral cavity cancers. However, larger studies are needed to verify our findings.
We thank Margaret Lung, Peggy Schuber, and Leanel Fairly for assistance in recruiting the subjects; Li-E Wang and Zhensheng Liu for technical support; Jianzhong He, John I. Calderon, and Kejin Xu for laboratory assistance; Joanne Sider for manuscript preparation; and Katie Matias (Department of Scientific Publications) for scientific editing.
Grant support: NIH Grant R01 ES11740 (to Q. Wei) and in part by U01 CA86390 (to M. Spitz), CA 97007 (to M. Spitz and W. Hong), and P30 CA16672 (to The University of Texas M. D. Anderson Cancer Center)
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Requests for reprints: Qingyi Wei, Department of Epidemiology, Unit 189, The University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Boulevard, Houston, TX 77030. Phone: 713-792-3020; Fax: 713-792-0807; E-mail:
- Received May 10, 2004.
- Revision received July 6, 2004.
- Accepted August 30, 2004.