## Abstract

Background: The p53 tumor suppressor protein is important in cell-cycle control, apoptosis, and DNA repair. Mutations in p53 have been associated with inherited cancer susceptibility. Because there is a difference in the risk of lung cancer among different ethnic groups, we examined associations between ethnicity and three polymorphisms in p53 (one exonic and two intronic) and haplotypes for the three loci and risk of lung cancer. We also examined the functionality of the p53 variants in apoptosis and DNA repair. Methods: In a case–control study, we frequency matched (by age, sex, and ethnicity) 635 lung cancer case patients and 635 control subjects. p53 genotypes and haplotypes at the three polymorphic sites were determined by restriction fragment length polymorphism analysis of lymphocyte DNA. Odds ratios (ORs) and 95% confidence intervals (CIs) for the association between genotype or haplotype and lung cancer risk were determined by logistic regression analysis. Apoptosis and DNA repair capacity were measured in 22 lymphoblastoid cell lines to determine the functional effects of the polymorphisms. All statistical tests were two-sided. Results: Genotype and haplotype frequency distributions were strongly dependent on ethnicity; variant allele frequencies were highest in African-Americans (29.1%) and lowest in Mexican-Americans (12.2%). Each of the three polymorphisms was associated with an increased risk of lung cancer among all ethnic groups. Moreover, for all three polymorphisms, increased variant allele copy number was associated with increased risk of lung cancer. Similarly, the variant haplotypes were also associated with an increased risk for lung cancer. Lymphoblastoid cell lines with all wild-type alleles at the three loci had statistically significantly higher apoptotic indices (13.66%, 95% CI = 8.61% to 18.71%) and DNA repair capacity (27.63%, 95% CI = 21.72% to 33.53%) than cell lines with at least one variant allele at all three loci (3.50%, 95% CI = 1.08% to 5.91%; and 17.48%, 95% CI = 7.99% to 26.96%, respectively). Conclusions: p53 polymorphisms may be associated with increased lung cancer risk and may affect p53 function.

Substantial evidence has accumulated to indicate that there is interindividual variation in susceptibility to lung cancer (1). This susceptibility is modulated, at least in part, by polymorphisms in genes encoding DNA repair proteins, cell-cycle control proteins, and metabolic enzymes. The gene for the tumor suppressor protein, p53, which is critical for DNA repair and cell-cycle control, is known to contain a variety of polymorphisms and mutations. Thirteen different polymorphisms have been detected in the human p53 gene, of which five were found in exons and eight in introns. The five coding region polymorphisms include a C → T transition in codon 21 (2), a G → A transition in codon 36 (3), a C → T transition in codon 47 (4), a G → C transversion in codon 72 (57), and an A → G transition in codon 213 (8). Of the five polymorphisms, those in codons 21, 36, and 213 are silent—that is, they do not result in a change in the encoded amino acid sequence. The polymorphism in codon 47 results in a proline to serine substitution, and the polymorphism in codon 72, which is in exon 4, results in an arginine to proline substitution. Whereas the codon 47 polymorphism is rare (4), the codon 72 polymorphism is common (7).

Mutation of p53 is considered an important genetic event in the development of lung cancer (9). Although the relationship between the polymorphism in codon 72 and lung cancer has been studied, the results have been inconsistent (7,1013). Three studies (7,10,14) found no association between the codon 72 p53 single-nucleotide polymorphisms and lung cancer. However, Fan et al. (13) found that, compared with the wild-type Arg/Arg genotype, the Pro/Pro and Pro/Arg genotypes were associated with an increased risk of lung adenocarcinoma in Caucasians. Furthermore, we previously reported that patients with the Pro/Pro genotype were diagnosed with lung cancer at an earlier age and had smoked fewer pack-years than patients with the Arg/Arg or Arg/Pro genotypes (10). Compared with the wild-type Arg/Arg genotype, risk estimates for the Pro/Pro genotype were statistically significantly high for patients younger than 53 years and for patients who reported smoking fewer than 30 pack-years (10).

The eight intronic p53 polymorphisms include a variable number of tandem repeat regions in intron 1 (15), a HaeIII restriction fragment length polymorphism (RFLP) in intron 1 (16), a G → C transversion in intron 2 (38 basepairs [bp] downstream of exon 2) (17), a 16-bp duplication in intron 3 (18,19), a G → A transition in intron 6 (20), a G → C transversion 37 bp upstream of the 5 end of exon 7 (21), an ApaI RFLP in intron 7 (22), and an A → T transversion in intron 10 (23). Although introns were originally believed to be nonfunctional because they do not code for proteins, it has been suggested that some of these sequences do indeed have relevance (24). Mutations in intron sequences may initiate aberrant pre-messenger RNA (mRNA) splicing, producing an mRNA that may be translated into a defective protein (21). Furthermore, intronic polymorphisms may influence coding-region mutations that increase the likelihood of a deleterious phenotype (25). Because introns have also been implicated in regulating gene expression and DNA–protein interactions (2628), mutations in intron sequences may affect these functions.

Similar to the studies regarding the polymorphism in codon 72, studies of the association between certain types of cancers and polymorphisms in introns 3 and 6 of p53 have yielded conflicting results (14,2933). Birgander et al. (33) found no association between the two polymorphisms and lung cancer risk. However, Biros et al. (14) reported a higher frequency of the intron 6 polymorphism in patients with lung cancer than in control subjects. Studies of the association between polymorphisms in introns 3 or 6 with other cancers also had inconsistent results (2932,34). For example, although Runnebaum et al. (29) noted an eightfold elevated risk for ovarian cancer in patients with the intron 3 variant genotype compared with patients with the intron 3 wild-type genotype, Lancaster et al. (31) and Campbell et al. (32) found no such association. Mavridou et al. (30) found a statistically significant difference in the distribution of the intron 6 polymorphism between healthy Caucasian control subjects and patients with ovarian cancer but found no difference between control subjects and patients with breast cancer. In contrast, in a small sample, Peller et al. (34) reported a statistically significant association between the intron 6 polymorphism and the risk of breast and colon cancers.

Several studies (33,3537) have estimated the haplotype frequencies for the polymorphisms in exon 4 and introns 3 and 6 and determined that the p53 haplotypes were associated with risk for lung (14,33), colorectal (35), and breast cancers (38). Haplotypes constitute the allele status at several linked loci on the same chromosome. Knowledge of the accumulated number of variants on the same chromosome may provide additional information about changes in p53 function because there could be additive or multiplicative changes in function. Furthermore, because haplotypes define multiple loci, they may be more precise and useful than genotypes in providing risk estimates for particular cancers.

The frequencies of the polymorphism tend to differ by ethnicity. Therefore, we carried out a case–control analysis to examine whether the polymorphisms in exon 4 and introns 3 and 6 of p53 and their haplotypes are associated with differences in the risk of lung cancer in Caucasians, African-Americans, and Mexican-Americans. We also examined the functional consequences (i.e., on several p53 activities) of the polymorphisms.

## Methods

### Study Population

Case patients were newly diagnosed, histologically confirmed, and previously untreated (by radiotherapy or chemotherapy) lung cancer patients consecutively recruited from The University of Texas M. D. Anderson Cancer Center in Houston, Texas. Approximately 22.6% of eligible case patients refused or were too ill to participate, leaving 635 case patients for the analyses. The control subjects were recruited from 20 area clinics of a large multispeciality health maintenance organization in the Houston metropolitan area. With a response rate of participation for the control subjects around 75%, 635 control subjects were ultimately recruited into the study. The control subjects visited the clinics for annual checkups. They had no prior history of cancer, with the exception of nonmelanoma skin cancer. There were no age, sex, or ethnic restrictions.

All subjects enrolled in the study were asked to sign an informed consent form, following the Institutional Review Board guidelines, before being interviewed. A detailed questionnaire was administered by trained interviewers to all study participants to gather demographic data and smoking history (smoking status, number of pack-years smoked, number of years smoked, and average number of cigarettes smoked per day). After the interview, 30 mL of blood was collected into heparinized tubes, which were coded with a unique study identification number and delivered to the laboratory for immediate DNA isolation. The control subjects were matched to the case patients by age, sex, ethnicity, and smoking status. Smoking status was divided into three categories: never, former, and current smokers. A never smoker was someone who had never smoked or had smoked fewer than 100 cigarettes in his or her lifetime. A former smoker was someone who reported a history of smoking but had stopped at least 1 year before being diagnosed with lung cancer (or 1 year before enrollment in the study, for control subjects). A current smoker was someone who currently smoked or who had stopped smoking less than 1 year before being diagnosed with lung cancer (or less than 1 year before enrollment in the study, for control subjects). Pack-years were defined as the average number of cigarettes smoked per year divided by 20 cigarettes and then multiplied by years smoked.

### Genotyping p53 Polymorphisms

Genomic DNA was isolated from peripheral blood lymphocytes by proteinase K digestion followed by isopropanol extraction and ethanol precipitation. Next, restriction fragment length polymorphism (RFLP)–polymerase chain reaction (PCR) was used to amplify p53 exon 4 codon 72, intron 3, and intron 6 fragments within which the polymorphisms fall. Each PCR reaction was performed in a 40-μL mixture containing 1 × PCR buffer (500 mM KCl, 100 mM Tris–HCl, and 1.0% Triton X-100; Promega, Madison, WI), 1.9 mM MgCl2, 0.25 mM dNTPs, 0.5 μM each primer, 50 ng template DNA, and 2 U Taq polymerase in storage buffer B (20 mM Tris–HCl, 100 mM KCl, 0.1 mM EDTA, 1 mM DTT (1,4-dithiothreitol), 50% glycerol, 0.5% Nonidet P-40, and 0.5% Tween 20; Promega). The PCR cycling conditions were 94 °C for 5 minutes, followed by 30 cycles at 94 °C for 30 seconds, 55 °C for 30 seconds, and 72 °C for 30 seconds, with a final extension step at 72 °C for 7 minutes.

For the exon 4 codon 72 polymorphism, the primers (5-CTGGTAAGGACAAGGGTTGG-3 and 5-ACTGACCGTGCAAGTCACAG-3) amplified a 396-bp DNA fragment. Next, the PCR product was digested with 10 U of BstUI (New England Biolabs, Beverly, MA). After an overnight digestion, the products were separated by gel electrophoresis (4% agarose gel for 30 minutes at 220 V) and visualized after staining with ethidium bromide. DNA from wild-type (WW) homozygotes produced 165-bp and 231-bp bands; DNA from wild-type/mutant (WM) heterozygotes produced 165-bp, 231-bp, and 396-bp bands; DNA from mutant (MM) homozygotes produced a 396-bp band.

For the intron 3 polymorphism, the primers (5-TGGGACTGACTTTCTGCTCTT-3 and 5-TCAAATCATCCATTGCTTGG-3) amplified 180-bp or 196-bp fragments. DNA from WW homozygotes produced the 180-bp band; DNA from WM heterozygotes produced both bands; DNA from MM homozygotes produced the 196-bp band.

For the intron 6 polymorphism, the primers (5-TGGCCATCTACAAGCAGTCA-3 and 5-TTGCACATCTCATGGGGTTA-3) amplified a 404-bp fragment. The PCR product was digested with 10 U of MspI (New England Biolabs). DNA from WW homozygotes produced 68-bp and 336-bp bands; DNA from WM heterozygotes produced 68-bp, 336-bp, and 404-bp bands; and DNA from MM homozygotes produced a 404-bp band.

The genotyping data were recorded by two independent readers who were blinded to the case–control status of the samples. Samples with ambiguous results, and 10% of randomly selected samples, were run in duplicate to ensure quality control.

### p53 Haplotype Determination

p53 haplotypes could be inferred directly from the genotyping results for individuals who were heterozygous at only one site or at no sites. For the other individuals, modified allele-specific PCR and RFLP–PCR methods described by Weston et al. (38) were used to determine the haplotypes.

For individuals who were heterozygous at only the intron 6 MspI restriction site (WW–WW–WM for exon 4–intron 3–intron 6 polymorphisms, respectively), PCR with the primer 5-CCCCCCTACTGCTCACCC-3 and the 5-end primer 5-TTCCTGAAAAACAACGTTCTGG-3 were used to amplify a 1618-bp fragment. Because the 3-end primer was designed to be complementary to the MspI restriction site wild-type allele, these primers amplify only the wild-type allele of intron 6.

To determine the haplotype relationship between intron 3 and intron 6, PCR with the primers 5-TTCCTGAAAAACAACGTTCTGG-3 and 5-TCAAATCATCCATTGCTTGG-3 were used to amplify 147-bp or 163-bp fragments from the wild-type intron 6 PCR product. The presence of a 147-bp band indicated that the two variant alleles (M-alleles) were on the same chromosome, whereas the presence of a 163-bp band indicated that the two M-alleles were on different chromosomes.

To determine the haplotype relationship between the exon 4 and intron 6 alleles, PCR with the primers used to genotype exon 4 was used to amplify the 396-bp or 165-bp and 231-bp fragments. The presence of a single 396-bp band indicated that the two M-alleles were on different chromosomes, whereas the presence of the 165-bp and 231-bp bands indicated that the M-alleles were on the same chromosome.

For individuals who were heterozygous at exon 4 and intron 3 but not intron 6 (WM–WM–WW and WM–WM–MM), PCR with the primers 5-ATGGGACTGACTTTCTGCTCTT-3 and 5-ACTGACCGTGCAAGTCACAG-3 were used to amplify 430-bp and 446-bp fragments. After amplification, the PCR product was digested overnight with 10 U of BstUI (New England Biolabs) and then separated on a 4% agarose gel by electrophoresis, following the same protocol as that for exon 4. The presence of 215-bp, 231-bp, and 430-bp bands indicated that the two M-alleles were on the same chromosome, whereas the presence of 199-bp, 231-bp, and 446-bp bands indicated that the two M-alleles were on different chromosomes.

Two independent readers who were blinded to the case–control status of the samples recorded the haplotying data. Samples with ambiguous results, and 10% of randomly selected samples, were run in duplicate to ensure quality control. We also compared the haplotype data with the genotype data for consistency.

### Lymphoblastoid Cell Lines

Twenty-two B-lymphoblastoid cell lines from healthy donors were provided by Dr. T. C. Hsu (M. D. Anderson Cancer Center, Houston) and Dr. G. E. Tomlinson (The University of Texas Southwestern Medical Center). These cell lines were not derived from any of the control subjects used in the genotype and haplotype analyses. The lymphoblastoid cell lines were genotyped at the three p53 loci.

### Apoptosis Assay

Radiation exposure induces apoptosis in lymphoblastoid cells. We assessed whether there was any difference in the apoptotic response to radiation in the lymphoblastoid cells by p53 genotypes. Cells in the logarithmic phase of growth were plated at 1 × 105/mL in 10 mL medium in 25-cm2 flasks 24 hours before irradiation. Standardization experiments (39) were performed at a γ-radiation dose of 2.5 Gy from a 137Cs source at room temperature. All cells were processed together, with the exception that the control cells were not irradiated. Irradiated and control cells were harvested at 48 hours after γ-irradiation. All cells were pelleted by centrifugation, resuspended with 4% paraformaldehyde in phosphate-buffered saline (PBS), fixed for 1 hour, rinsed with PBS, permeabilized with 0.5% Triton X-100 for approximately 10 minutes, and rinsed with PBS twice more. The cells were next processed according to the manufacturer's recommended protocol for the detection of apoptosis with the use of the TUNEL (transferase uridyl nick end labeling assay; in situ cell death detection) assay kit (Boehringer Mannheim Corp., Indianapolis, IN). The difference in the percentage of TUNEL signal positive cells in irradiated versus control cells was recorded as the apoptotic index. Each cell line was tested three times.

### DNA Repair Assay

To assess DNA repair capacity (DRC) in the lymphoblastoid cell lines, the cells were transfected with a chemically damaged luciferase reporter gene plasmid construct. Because only the plasmids whose DNA damage was repaired would express the luciferase protein, the quantity of luciferase protein expression was used as a marker for the efficiency of DNA damage repair. Briefly, pGL-3 plasmid DNA (0.5 mg/mL) (Promega) was treated with benzo[a]pyrene diol epoxide (BPDE) (45 μM) (Midwest Research Institute, Kansas City, MO) and then incubated for 3 hours at room temperature. After incubation, the plasmid DNA was precipitated with 99% ethanol, washed with 75% ethanol, dissolved in 0.1 M Tris-EDTA buffer at a final concentration of 0.25 mg/mL, and stored in a –20 °C freezer until used to transfect the cells. Lymphoblastoid cell lines (2 × 105 cells/mL) were transfected with untreated or BPDE-treated pGL-3 plasmid DNA (0.5 μg) in triplicate by the DEAE-Dextran Transfection System (Promega). After 48 hours, luciferase activity was measured in a luminometer (Lumat LB 9507; Berthold Technologies GmbH, Bad Wildbad, Germany) using Luciferase Assay Systems (Promega), following the manufacturer's instructions. DRC was defined as the ratio of the luciferase activity of the cells transfected with BPDE-treated plasmids to that of cells transfected with untreated plasmids. For each cell line, the average of three experiments was recorded.

### Statistical Analysis

Pearson's χ2 test or Fisher's exact test (when the expected number in any cell was less than five) was used to compare the distribution of the p53 genotypes and haplotypes between case patients and control subjects. All analyses were stratified by ethnicity because it was a major confounding factor. Student's t test was used to compare the continuous independent variables, such as age and pack-years smoked, between case patients and control subjects. The overall ethnic-specific genotype frequencies in the control subjects were compared with the frequencies expected by the Hardy–Weinberg equilibrium by goodness-of-fit χ2. For sparse genotype data, we used permutation testing as described in Guo and Thompson (40) to assess the statistical significance of tests for the Hardy–Weinberg equilibrium. Linkage disequilibrium analysis was performed for all three ethnic groups. The disequilibrium parameters (D and D) and the χ2 values for statistical significance and their corresponding P values were calculated. The Lewontin (41) two-locus model and the Thomson and Baur (42) three-locus model were used to determine D. D is a measure of the difference between the frequency of a haplotype and the frequency that is expected under Hardy-Weinberg Equilibrium. D is a standardized version of D that ranges from 0 to 1. Genotype frequencies were compared between case patients and control subjects using a goodness-of-fit test. Odds ratios (ORs) and 95% confidence intervals (CIs) were calculated as a measure of association with cancer risk. Unconditional logistic regression analysis was used to adjust for the effect of covariates such as age, sex, and smoking status or pack-years smoked. For estimating the risk of developing lung cancer for case patients with combinations of genotypes and haplotypes, we restricted the analysis to Caucasians because too few Mexican-Americans and African-Americans were included in the study. The trend tests were conducted using a score test (43). Interaction terms were created between genotypes and then tested in a logistic regression model. These statistical analyses were performed using STATA software (44). The linkage disequilibrium analysis was performed using EH (EH-plus) software (45). StatExact (46) was used to calculate the P value by using a Monte-Carlo method to simulate 1 000 000 replicate tables having the same marginal proportions; Fisher's exact test was then used to compare the observed data to simulated data and thus generate a significance level. All statistical tests were two-sided.

## Results

### Distribution of Demographic Characteristics

We used a case–control approach to examine whether the exon 4 and introns 3 and 6 polymorphisms and the haplotypes in p53 were associated with risk of lung cancer. The distributions of demographic characteristics of the case patients and control subjects are summarized in Table 1. The case patients and control subjects were matched on age, sex, ethnicity, and smoking status. Although the case patients smoked more cigarettes per day than the control subjects (28.3, 95% CI = 27.1 to 29.5; and 27.9, 95% CI = 26.6 to 29.2, respectively), this difference was not statistically significant (P = .285). The only statistically significant difference between the case patients and control subjects was in the pack-years smoked (52.0, 95% CI = 49.1 to 54.9; and 47.3, 95% CI = 44.6 to 50.0, respectively; P = .005).

### Risk Estimates for p53 Polymorphisms

We estimated the risk of lung cancer associated with each p53 polymorphism by comparing the risk associated with each genotype with that associated with the WW genotype. We analyzed the mutant genotypes separately for Caucasians but combined them for Mexican-Americans and African-Americans because of the small sample sizes.

Compared with the wild-type alleles, only the intron 3 polymorphism was associated with a statistically significantly increased risk for lung cancer for Mexican-Americans (adjusted OR = 4.43; 95% CI = 1.40 to 14.00), although the other polymorphisms were also associated with elevated risks (adjusted OR for exon 4 polymorphism = 1.63; 95% CI = 0.65 to 4.05, and adjusted OR for intron 6 polymorphism = 1.79; 95% CI = 0.63 to 5.06) (data not shown). For African-Americans, the adjusted OR associated with the intron 3 polymorphism was 1.23 (95% CI = 0.55 to 2.72), the adjusted OR for intron 6 polymorphism was 1.33 (95% CI = 0.63 to 2.85), and the adjusted OR for exon 4 polymorphism was 0.93 (95% CI = 0.41 to 2.14) (data not shown).

We found no statistically significant differences between the genotype distribution and cancer histology within any ethnic group (P = .51 for intron 3, P = .12 for intron 6, and P = .13 for exon 4 for Caucasians; P = .40 for intron 3, P = .77 for intron 6, and P = .89 for exon 4 for Mexican-Americans; and P = .41 for intron 3, P = .79 for intron 6, and P = .63 for exon 4 for African-Americans) (data not shown).

### Risk Estimates for Combined Genotypes in Caucasians

We next tested for an interaction among the polymorphisms and found that there was no statistically significant interaction between intron 3 and intron 6 (P = .42). We did detect a negative interaction between intron 3 and exon 4 (P = .04) and a negative interaction between intron 6 and exon 4 (P = .03) (data not shown).

### Risk Estimates for p53 Haplotypes

To examine the linkage among the three loci (exon 4, intron 3, and intron 6), we performed linkage disequilibrium analysis. In Caucasians, the χ2 test of statistical significance for a three-locus disequilibrium gave a test statistic value of 97.5 (D = 0.62) for the case patients, 85.4 (D = 0.75) for the control subjects, and 186.3 (D = 0.52) for all subjects. The three-locus disequilibrium was statistically significant (P<.001).

For Mexican-Americans, the χ2 test for linkage disequilibrium was statistically significant for both case patients (χ2 = 8.46, P = .004, D = 0.54) and control subjects (χ2 = 4.67, P = .031, D = 1.0). For African-Americans, the χ2 test was significant for case patients (χ2 = 7.21, P = .007, D = 0.60) but not for control subjects (χ2 = 2.30, P = .129, (D = 0.00). Two loci's D values ranged between 0.53 and 1.0.

## Discussion

We evaluated the frequency distributions of p53 exon 4, intron 3, and intron 6 polymorphisms in a lung cancer case–control study. When participants were stratified according to ethnicity, the p53 polymorphisms were associated with an elevated lung cancer risk for all three ethnic groups (Caucasians, Mexican-Americans, and African-Americans), with the exception of exon 4 in African-Americans. The three polymorphisms were strongly linked, and there were gene–dose effects for all three polymorphisms among Caucasians. We observed that the haplotypes with the M-allele, consistent with these findings, were associated with an increased risk for lung cancer. The association may be attributable to changes in p53 function because lymphoblastoid cell lines with all wild-type alleles exhibited the highest mean apoptotic index and DNA repair capacity, whereas those with at least one mutant allele at all three polymorphic sites had the lowest indices.

Although it is unclear which polymorphism is responsible for the changes in p53 function in the lymphoblastoid lines, the exon 4 polymorphism is the most likely candidate. Thomas et al. (47) reported that the p53 exon 4 mutant polymorphism was less efficient in suppressing cell transformation and slower in inducing apoptosis than was the p53 exon 4 wild-type protein. Likewise, Marin et al. (48) demonstrated that the p53 exon 4 wild-type protein was more efficient than the p53 exon 4 mutant protein at binding and inactivating p73, a tumor suppressor protein responsible for apoptosis. Although the effect of the polymorphism may be to inhibit p53 function in vitro, whether this effect also occurs in vivo remains to be determined. Brooks et al. (49) noted that p53 loss of heterozygosity in tumor tissue was more common in individuals who were heterozygous at codon 72 than in those who were homozygous at codon 72. In our study, in agreement with others (40,41), we noted that the exon 4 polymorphism may have a functional effect. We found that the exon 4 polymorphism alone was a statistically significant risk factor for lung cancer in Caucasians; however, in conjunction with introns 3 and 6 variants, the exon 4 polymorphism was associated with lower risk. Likewise, Birgander et al. (33) suggested that the exon 4 variant, in conjunction with an intron 3 variant, might exert a protective effect rather than a detrimental effect for lung and colorectal cancers. Clearly, more studies are needed to assess the relationship between the exon 4 polymorphism and p53 function.

Our data also suggest that the intron 3 and intron 6 M-alleles might exert a functional effect. We noted that as the copy number of the M-allele in introns 3 and 6 increased, the level of the apoptotic index decreased. In addition, cell lines with at least one variant allele at all three polymorphic sites had a statistically significantly reduced DNA repair capacity compared with cell lines with all wild-type alleles. Reduced DNA repair capacity and apoptotic responses have been linked to an increased risk of lung cancer (50,51). Avigad et al. (28) reported a germline variant in p53 intron 6 (a G-to-A transition 39 bp upstream of exon 7) in six pediatric patients with diverse tumors, all of which were part of the Li–Fraumeni syndrome. Lehman et al. (52) reported a G-to-C base change at nucleotide 13 964 in intron 6 of p53 in patients with hereditary breast cancer but not sporadic breast cancer. The patients with hereditary breast cancer with this mutation had strong family histories of multiple types of cancers and high levels of mutant p53 protein. Two immortalized lymphoblastoid cell lines derived from these patients had prolonged in vitro survival and decreased chemotherapy-induced apoptosis after treatment with cisplatinum, suggesting that the mutation in the p53 gene changes p53 function (52).

We found that the WW–WW–WW genotype was the most prevalent genotype in lung cancer case patients and control subjects but was more common in control subjects than in case patients. A similar observation has been documented for colorectal cancer and breast cancer (35,37). The haplotype distribution in our study is similar to that found by Weston et al. (38), who suggested that p53 minor haplotypes are associated with risk of breast cancer in Caucasians. Similarly, we found that the minor haplotypes were statistically significantly associated with an elevated risk of lung cancer, with the highest risk found in individuals with the W–M–M haplotype. In addition, we and Sjalander et al. (53) found that the three polymorphisms are in strong linkage disequilibrium. Wang-Gohrke et al. (54) reported a strong linkage disequilibrium between intron 3 and intron 6 polymorphisms in healthy control subjects and ovarian cancer patients, and Birgander et al. (33) showed a strong linkage disequilibrium between the exon 4 and intron 3 variant alleles. Although we had originally hypothesized that the inclusion of haplotype information might improve the fit of our risk models, we did not observe any substantial improvement in models that included haplotypes over models with alleles at each locus. Because of the strong linkage disequilibria among the markers, larger sample sizes may resolve whether any relative improvement in modeling might be obtained by haplotyping. However, although there may be some rare genotype combinations for which further studies with haplotypes may be warranted, for the majority of subjects it appears that genotyping the different loci provides adequate information for risk model-building.

It is evident from our data that the variant allele and haplotype frequencies were strongly dependent on ethnicity. An expanding body of literature demonstrates that ethnic differences in cancer incidence may be partly caused by host genetic factors, particularly genetic polymorphisms (5456). Lung cancer incidence rates are highest in African-Americans and lowest in Mexican-Americans. The difference is not explained solely by patterns of cigarette smoking (56,57). The difference may be related, in part, to the prevalence of the MM genotype for exon 4; 21.6% in African-American control subjects and only 7.7% in Mexican-American control subjects, similar to that previously reported (10). Among the control subjects in this study, the frequencies of the variant allele were highest in African-Americans, intermediate in Caucasians, and lowest in Mexican-Americans. There was a similar trend in haplotype frequencies. If the M-allele variant is associated with a risk of lung cancer, the lower frequencies in Mexican-Americans may partly explain their apparent lower rates of lung cancer.

In summary, we present data suggesting that p53 polymorphisms are associated with the risk of lung cancer, possibly by affecting p53 functional activity. Further studies are needed to determine the mechanism by which exon 4 and introns 3 and 6 variants modify cancer risk.

Table 1.

Distribution of select characteristics among case patients and control subjects

Variables Case patients n = 635 Control subjects n = 635 P value*
*P values derived from the Pearson χ2 test for categorical variables (sex, ethnicity, and smoking status) and the Student t test for continuous variables (age, no. cigarettes/day, and pack-years). All P values are two-sided.
†Never smokers were defined as people who had never smoked or had smoked fewer than 100 cigarettes in his or her lifetime. Former smokers reported a history of smoking but had stopped at least 1 year before being diagnosed with lung cancer (or 1 year before enrollment in the study, for control subjects). Current smokers were currently smoking or had stopped smoking less than 1 year before being diagnosed with lung cancer (or less than 1 year before enrollment in the study, for control subjects).
‡For ever smokers only.
Sex (%)
Male 337 (53.1) 354 (55.7)
Female 298 (46.9) 281 (44.3) .338
Ethnicity (%)
Caucasians 517 (81.4) 544 (85.7)
Mexican-Americans 54 (8.5) 40 (6.3)
African-Americans 64 (10.1) 51 (8.0) .120
Smoking Status† (%)
Never 63 (10.4) 63 (10.4)
Former 312 (51.6) 312 (51.6)
Current 230 (38.0) 230 (38.0)
Age Mean, years (SD) 61.4 (9.7) 60.6 (9.8) .092
No. Cigarettes/Day (SD)‡ 28.3 (14.5) 27.9 (15.3) .285
Pack-years (SD)‡ 52.0 (33.7) 47.3 (32.1) .005
Variables Case patients n = 635 Control subjects n = 635 P value*
*P values derived from the Pearson χ2 test for categorical variables (sex, ethnicity, and smoking status) and the Student t test for continuous variables (age, no. cigarettes/day, and pack-years). All P values are two-sided.
†Never smokers were defined as people who had never smoked or had smoked fewer than 100 cigarettes in his or her lifetime. Former smokers reported a history of smoking but had stopped at least 1 year before being diagnosed with lung cancer (or 1 year before enrollment in the study, for control subjects). Current smokers were currently smoking or had stopped smoking less than 1 year before being diagnosed with lung cancer (or less than 1 year before enrollment in the study, for control subjects).
‡For ever smokers only.
Sex (%)
Male 337 (53.1) 354 (55.7)
Female 298 (46.9) 281 (44.3) .338
Ethnicity (%)
Caucasians 517 (81.4) 544 (85.7)
Mexican-Americans 54 (8.5) 40 (6.3)
African-Americans 64 (10.1) 51 (8.0) .120
Smoking Status† (%)
Never 63 (10.4) 63 (10.4)
Former 312 (51.6) 312 (51.6)
Current 230 (38.0) 230 (38.0)
Age Mean, years (SD) 61.4 (9.7) 60.6 (9.8) .092
No. Cigarettes/Day (SD)‡ 28.3 (14.5) 27.9 (15.3) .285
Pack-years (SD)‡ 52.0 (33.7) 47.3 (32.1) .005
Table 2.

Distribution of p53 exon 4, intron 3, and intron 6 genotypes among control subjects by ethnic group

Genotypes*
Site/Ethnicity WW (%) WM (%) MM (%) P value M-allele frequency HWE P value†
*WW = wild-type homozygous; WM = wild-type/mutant heterozygous; MM = mutant homozygous. Individuals with missing genotyping data were not included in the analysis. There were five control subjects missing exon 4 genotype data, two control subjects missing intron 3 genotype data, and three control subjects missing intron 6 genotype data.
†Hardy–Weinberg Equilibrium (38)
Exon 4
Caucasians 352 (65.2) 156 (28.9) 32 (5.9)  0.20 .016
Mexican-Americans 26 (66.7) 10 (25.6) 3 (7.7)  0.21 .170
African-Americans 15 (29.4) 25 (49.0) 11 (21.6) <.001 0.46 .999
Intron 3
Caucasians 440 (81.2) 92 (17.0) 10 (1.9)  0.10 .061
Mexican-Americans 35 (87.5) 5 (12.5) 0 (0.0)  0.06 .999
African-Americans 35 (68.6) 14 (27.5) 2 (3.9) .160 0.18 .637
Intron 6
Caucasians 423 (78.2) 107 (19.8) 11 (2.0)  0.12 .215
Mexican-Americans 32 (80.0) 8 (20.0) 0 (0.0)  0.10 .999
African-Americans 30 (58.8) 18 (35.3) 3 (5.9) .019 0.24 .999
Genotypes*
Site/Ethnicity WW (%) WM (%) MM (%) P value M-allele frequency HWE P value†
*WW = wild-type homozygous; WM = wild-type/mutant heterozygous; MM = mutant homozygous. Individuals with missing genotyping data were not included in the analysis. There were five control subjects missing exon 4 genotype data, two control subjects missing intron 3 genotype data, and three control subjects missing intron 6 genotype data.
†Hardy–Weinberg Equilibrium (38)
Exon 4
Caucasians 352 (65.2) 156 (28.9) 32 (5.9)  0.20 .016
Mexican-Americans 26 (66.7) 10 (25.6) 3 (7.7)  0.21 .170
African-Americans 15 (29.4) 25 (49.0) 11 (21.6) <.001 0.46 .999
Intron 3
Caucasians 440 (81.2) 92 (17.0) 10 (1.9)  0.10 .061
Mexican-Americans 35 (87.5) 5 (12.5) 0 (0.0)  0.06 .999
African-Americans 35 (68.6) 14 (27.5) 2 (3.9) .160 0.18 .637
Intron 6
Caucasians 423 (78.2) 107 (19.8) 11 (2.0)  0.12 .215
Mexican-Americans 32 (80.0) 8 (20.0) 0 (0.0)  0.10 .999
African-Americans 30 (58.8) 18 (35.3) 3 (5.9) .019 0.24 .999
Table 3.

Distribution of p53 exon 4, intron 3, and intron 6 haplotypes among control subjects by ethnic group

Haplotypes* (exon 4-intron 3-intron 6) Caucasians n† (%) Mexican-Americans n† (%) African-Americans n† (%) P value‡
*W = wild-type; M = variant.
†n is the number of alleles. Because each individual has two alleles, the total number of alleles will be twice that of the total number of individuals. Individuals with missing genotyping data were not included in the analyses.
‡Monto–Carlo method and Fisher exact test were used to calculate the P value.
W–W–W 809 (75.1) 60 (76.0) 51 (50.0)
M–W–W 110 (10.2) 11 (13.9) 26 (25.5)
W–M–W 19 (1.8) 0 (0.0) 0 (0.0)
W–W–M 21 (2.0) 2 (2.5) 1 (1.0)
M–M–W 12 (1.1) 0 (0.0) 1 (1.0)
M–W–M 26 (2.4) 1 (1.3) 6 (5.9)
W–M–M 11 (1.0) 1 (1.3) 3 (2.9)
M–M–M 70 (6.5) 4 (5.1) 14 (13.7) <.001
M allele frequency 458 (14.2) 29 (12.2) 89 (29.1)
Haplotypes* (exon 4-intron 3-intron 6) Caucasians n† (%) Mexican-Americans n† (%) African-Americans n† (%) P value‡
*W = wild-type; M = variant.
†n is the number of alleles. Because each individual has two alleles, the total number of alleles will be twice that of the total number of individuals. Individuals with missing genotyping data were not included in the analyses.
‡Monto–Carlo method and Fisher exact test were used to calculate the P value.
W–W–W 809 (75.1) 60 (76.0) 51 (50.0)
M–W–W 110 (10.2) 11 (13.9) 26 (25.5)
W–M–W 19 (1.8) 0 (0.0) 0 (0.0)
W–W–M 21 (2.0) 2 (2.5) 1 (1.0)
M–M–W 12 (1.1) 0 (0.0) 1 (1.0)
M–W–M 26 (2.4) 1 (1.3) 6 (5.9)
W–M–M 11 (1.0) 1 (1.3) 3 (2.9)
M–M–M 70 (6.5) 4 (5.1) 14 (13.7) <.001
M allele frequency 458 (14.2) 29 (12.2) 89 (29.1)
Table 4.

Risk estimates for p53 polymorphisms in Caucasian case patients and control subjects

Genotypes Case patients No. (%)* Control subjects No. (%)* P value† Univariate OR (95% CI)‡ Adjusted OR (95% CI)‡
*Individuals with missing genotyping data were not included in the analysis.
†Pearson χ2 test was used to calculate the P value.
‡ORs were adjusted by age, sex, and smoking status. OR = odds ratio; CI = confidence interval.
Caucasians
Exon 4
WW 299 (58.0) 352 (65.2)  Referent Referent
WM 177 (34.3) 156 (28.9)  1.34 (1.03 to 1.74) 1.37 (1.01 to 1.86)
MM 40 (7.8) 32 (5.9) .050 1.47 (0.90 to 2.40) 1.67 (0.94 to 2.98)
Ptrend = .018 Ptrend = .013
Intron 3
WW 377 (73.1) 440 (81.2)  Referent Referent
WM 125 (24.2) 92 (17.0)  1.59 (1.17 to 2.15) 1.77 (1.24 to 2.52)
MM 14 (2.7) 10 (1.9) .007 1.63 (0.72 to 3.72) 2.37 (0.76 to 7.39)
Ptrend<.001 Ptrend<.001
Intron 6
WW 347 (67.3) 423 (78.2)  Referent Referent
WM 155 (30.0) 107 (19.8)  1.77 (1.33 to 2.35) 1.95 (1.40 to 2.70)
MM 14 (2.7) 11 (2.0) <.001 1.55 (0.70 to 3.46) 2.45 (0.77 to 7.78)
Ptrend<.001 Ptrend<.001
Genotypes Case patients No. (%)* Control subjects No. (%)* P value† Univariate OR (95% CI)‡ Adjusted OR (95% CI)‡
*Individuals with missing genotyping data were not included in the analysis.
†Pearson χ2 test was used to calculate the P value.
‡ORs were adjusted by age, sex, and smoking status. OR = odds ratio; CI = confidence interval.
Caucasians
Exon 4
WW 299 (58.0) 352 (65.2)  Referent Referent
WM 177 (34.3) 156 (28.9)  1.34 (1.03 to 1.74) 1.37 (1.01 to 1.86)
MM 40 (7.8) 32 (5.9) .050 1.47 (0.90 to 2.40) 1.67 (0.94 to 2.98)
Ptrend = .018 Ptrend = .013
Intron 3
WW 377 (73.1) 440 (81.2)  Referent Referent
WM 125 (24.2) 92 (17.0)  1.59 (1.17 to 2.15) 1.77 (1.24 to 2.52)
MM 14 (2.7) 10 (1.9) .007 1.63 (0.72 to 3.72) 2.37 (0.76 to 7.39)
Ptrend<.001 Ptrend<.001
Intron 6
WW 347 (67.3) 423 (78.2)  Referent Referent
WM 155 (30.0) 107 (19.8)  1.77 (1.33 to 2.35) 1.95 (1.40 to 2.70)
MM 14 (2.7) 11 (2.0) <.001 1.55 (0.70 to 3.46) 2.45 (0.77 to 7.78)
Ptrend<.001 Ptrend<.001
Table 5.

Risk estimates for combinations of p53 genotypes in Caucasian case patients and control subjects*

Genotypes
Exon Intron 3 Intron 6 Case patients No. (%)† Control subjects No. (%)† Univariate OR (95% CI) Adjusted OR (95% CI)‡
*OR = odds ratio; CI = confidence interval; WW = wild-type homozygous; WM = wild-type/mutant heterozygous; MM = mutant homozygous; WM + MM = individuals with WM or MM genotypes.
†Individuals with missing genotyping data were not included in the analyses.
‡ORs were adjusted by age, sex and smoking status.
WW WW WW 226 (42.4) 307 (57.6) Referent Referent
WM + MM WW WW 95 (52.5) 86 (47.5) 1.51 (1.07 to 2.11) 1.50 (1.07 to 2.11)
WW WM + MM WW 20 (57.1) 15 (42.9) 1.81 (0.91 to 3.62) 1.78 (0.89 to 3.56)
WW WW WM + MM 30 (62.5) 18 (37.5) 2.26 (1.23 to 4.16) 2.28 (1.24 to 4.21)
WM + MM WM + MM WW 5 (29.4) 12 (70.6) 0.57 (0.20 to 1.63) 0.56 (0.20 to 1.63)
WM + MM WW WM + MM 25 (51.0) 24 (49.0) 1.42 (0.79 to 2.54) 1.36 (0.76 to 2.46)
WW WM + MM WM + MM 22 (66.7) 11 (33.3) 2.72 (1.29 to 5.72) 2.74 (1.30 to 5.79)
WM + MM WM + MM WM + MM 92 (59.0) 64 (41.0) 1.95 (1.36 to 2.81) 1.96 (1.36 to 2.81)
Genotypes
Exon Intron 3 Intron 6 Case patients No. (%)† Control subjects No. (%)† Univariate OR (95% CI) Adjusted OR (95% CI)‡
*OR = odds ratio; CI = confidence interval; WW = wild-type homozygous; WM = wild-type/mutant heterozygous; MM = mutant homozygous; WM + MM = individuals with WM or MM genotypes.
†Individuals with missing genotyping data were not included in the analyses.
‡ORs were adjusted by age, sex and smoking status.
WW WW WW 226 (42.4) 307 (57.6) Referent Referent
WM + MM WW WW 95 (52.5) 86 (47.5) 1.51 (1.07 to 2.11) 1.50 (1.07 to 2.11)
WW WM + MM WW 20 (57.1) 15 (42.9) 1.81 (0.91 to 3.62) 1.78 (0.89 to 3.56)
WW WW WM + MM 30 (62.5) 18 (37.5) 2.26 (1.23 to 4.16) 2.28 (1.24 to 4.21)
WM + MM WM + MM WW 5 (29.4) 12 (70.6) 0.57 (0.20 to 1.63) 0.56 (0.20 to 1.63)
WM + MM WW WM + MM 25 (51.0) 24 (49.0) 1.42 (0.79 to 2.54) 1.36 (0.76 to 2.46)
WW WM + MM WM + MM 22 (66.7) 11 (33.3) 2.72 (1.29 to 5.72) 2.74 (1.30 to 5.79)
WM + MM WM + MM WM + MM 92 (59.0) 64 (41.0) 1.95 (1.36 to 2.81) 1.96 (1.36 to 2.81)
Table 6.

Risk estimates for extended p53 haplotypes in Caucasian case patients and control subjects*

Haplotype (exon 4-intron 3-intron 6) Case patients n (%)† Control subjects n (%)† Univariate OR (95% CI) Adjusted OR (95% CI)‡
*OR = odds ratio; CI = confidence interval; W = wild-type; M = mutant.
†n is the number of alleles. Because each individual has two alleles, the total number of alleles will be twice the total number of individuals. Individuals with missing haplotyping data were not included in the analyses.
‡ORs were adjusted by age, sex and smoking status.
W–W–W 695 (67.5) 809 (75.1) Referent Referent
M–W–W 123 (11.9) 110 (10.2) 1.30 (0.99 to 1.72) 1.30 (0.99 to 1.71)
W–M–W 24 (2.3) 19 (1.8) 1.47 (0.80 to 2.71) 1.45 (0.78 to 2.66)
W–W–M 32 (3.1) 21 (2.0) 1.77 (1.01 to 3.10) 1.78 (1.02 to 3.12)
M–M–W 5 (0.5) 12 (1.1) 0.49 (0.17 to 1.38) 0.49 (0.17 to 1.39)
M–W–M 27 (2.6) 26 (2.4) 1.21 (0.70 to 2.09) 1.17 (0.67 to 2.02)
W–M–M 22 (2.1) 11 (1.0) 2.33 (1.12 to 4.83) 2.33 (1.12 to 4.83)
M–M–M 102 (9.9) 70 (6.8) 1.70 (1.23 to 2.34) 1.70 (1.23 to 2.35)
W–W–W 695 (67.5) 809 (75.1) Referent Referent
Others 335 (32.5) 269 (25.0) 1.45 (1.20 to 1.75) 1.47 (1.21 to 1.79)
Haplotype (exon 4-intron 3-intron 6) Case patients n (%)† Control subjects n (%)† Univariate OR (95% CI) Adjusted OR (95% CI)‡
*OR = odds ratio; CI = confidence interval; W = wild-type; M = mutant.
†n is the number of alleles. Because each individual has two alleles, the total number of alleles will be twice the total number of individuals. Individuals with missing haplotyping data were not included in the analyses.
‡ORs were adjusted by age, sex and smoking status.
W–W–W 695 (67.5) 809 (75.1) Referent Referent
M–W–W 123 (11.9) 110 (10.2) 1.30 (0.99 to 1.72) 1.30 (0.99 to 1.71)
W–M–W 24 (2.3) 19 (1.8) 1.47 (0.80 to 2.71) 1.45 (0.78 to 2.66)
W–W–M 32 (3.1) 21 (2.0) 1.77 (1.01 to 3.10) 1.78 (1.02 to 3.12)
M–M–W 5 (0.5) 12 (1.1) 0.49 (0.17 to 1.38) 0.49 (0.17 to 1.39)
M–W–M 27 (2.6) 26 (2.4) 1.21 (0.70 to 2.09) 1.17 (0.67 to 2.02)
W–M–M 22 (2.1) 11 (1.0) 2.33 (1.12 to 4.83) 2.33 (1.12 to 4.83)
M–M–M 102 (9.9) 70 (6.8) 1.70 (1.23 to 2.34) 1.70 (1.23 to 2.35)
W–W–W 695 (67.5) 809 (75.1) Referent Referent
Others 335 (32.5) 269 (25.0) 1.45 (1.20 to 1.75) 1.47 (1.21 to 1.79)
Table 7.

Apoptotic index and DNA repair capacity in cell lines with various p53 haplotype combinations

Apoptotic Index* DNA repair capacity*
Haplotype (exon 4–intron 3–intron 6) No. Mean 95% CI† P value Mean 95% CI P value
*The apoptotic index was determined from the percentage of transferase uridyl nick end labeling assay-positive lymphoblastoid cells in irradiated samples relative to those in unirradiated control samples. DNA repair capacity was determined from the amount of luciferase detected in cells transfected with a chemically damaged expression vector relative to that in cells transfected with a control expression vector.
†CI = confidence interval; W = wild-type.
WW–WW–WW 11 13.66 8.61 to 18.71  27.63 21.72 to 33.53
≥1 variant allele 11 4.29 2.23 to 6.35 <.001 20.45 13.11 to 27.80 .11
≥1 variant allele at each site 3.50 1.08 to 5.91 .002 17.48 7.99 to 26.96 .04
Apoptotic Index* DNA repair capacity*
Haplotype (exon 4–intron 3–intron 6) No. Mean 95% CI† P value Mean 95% CI P value
*The apoptotic index was determined from the percentage of transferase uridyl nick end labeling assay-positive lymphoblastoid cells in irradiated samples relative to those in unirradiated control samples. DNA repair capacity was determined from the amount of luciferase detected in cells transfected with a chemically damaged expression vector relative to that in cells transfected with a control expression vector.
†CI = confidence interval; W = wild-type.
WW–WW–WW 11 13.66 8.61 to 18.71  27.63 21.72 to 33.53
≥1 variant allele 11 4.29 2.23 to 6.35 <.001 20.45 13.11 to 27.80 .11
≥1 variant allele at each site 3.50 1.08 to 5.91 .002 17.48 7.99 to 26.96 .04
Supported by grants RO1 CA 55769, RO1 74880, U19 CA68437, SPORE CA 95008, and UO1 CA86390 from the National Cancer Institute, National Institutes of Health, Department of Health and Human Services. We thank Dr. Maureen E. Goode for her valuable assistance in editing this manuscript.

## References

1
Spitz MR, Wei Q, Li G, Wu X. Genetic susceptibility to tobacco carcinogenesis.
Cancer Invest

1999
;
17
:
661
–2.
2
Ahuja HG, Testa MP, Cline MJ. Variation in the protein coding region of the human p53 gene.
Oncogene

1990
;
5
:
1409
–10.
3
Felix CA, Brown DL, Mitsudomi T, Ikagaki N, Wong A, Wasserman R, et al. Polymorphism at codon 36 of the p53 gene.
Br J Cancer

1999
;
81
:
179
–83.
4
Felley-Bosco E, Weston A, Cawley HM, Bennett WP, Harris CC. Functional studies of a germ-line polymorphism at codon 47 within the p53 gene.
Am J Hum Genet

1993
;
53
:
752
–9.
5
Harris N, Brill E, Shohat O, Prokocimer M, Wolf D, Arai N, et al. Molecular basis for heterogeneity of the human p53 protein.
Mol Cell Biol

1986
;
6
:
4650
–6.
6
Matlashewski GJ, Tuck S, Pim D, Lamb P, Schneider J, Crawford LV. Primary structure polymorphism at amino acid residue 72 of human p53.
Mol Cell Biol

1987
;
7
:
961
–3.
7
Weston A, Perrin LS, Forrester K, Hoover RN, Trump BF, Harris CC, et al. Allelic frequency of a p53 polymorphism in human lung cancer.
Cancer Epidemiol Biomarkers Prev

1992
;
1
:
481
–3.
8
Carbone D, Chiba I, Mitsudomi T. Polymorphism at codon 213 within the p53 gene.
Oncogene

1991
;
6
:
1691
–2.
9
Chiba I, Takahashi T, Nau MM, D'Amico D, Curiel DT, Mitsudomi T, et al. Mutations in the p53 gene are frequent in primary, resected non-small cell lung cancer. Lung Cancer Study Group.
Oncogene

1990
;
5
:
1603
–10.
10
Jin X, Wu X, Roth JA, Amos CI, King TM, Branch C, et al. Higher lung cancer risk for younger African-Americans with the Pro/Pro p53 genotype.
Carcinogenesis

1995
;
16
:
2205
–8.
11
Murata M, Tagawa M, Kimura M, Kimura H, Watanabe S, Saisho H, et al. Analysis of a germ line polymorphism of the p53 gene in lung cancer patients; discrete results with smoking history.
Carcinogenesis

1996
;
17
:
261
–4.
12
Wang YC, Chen CY, Chen SK, Chang YY, Lin P. p53 codon 72 polymorphism in Taiwanese lung cancer patients: association with lung cancer susceptibility and prognosis.
Clin Cancer Res

1999
;
5
:
129
–34.
13
Fan R, Wu MT, Miller D, Wain JC, Kelsey KT, Wiencke JK, et al. The p53 codon 72 polymorphism and lung cancer risk.
Cancer Epidemiol Biomarkers Prev

2000
;
9
:
1037
–42.
14
Biros E, Kalina I, Kohut A, Stubna J, Salagovic J. Germ line polymorphisms of the tumor suppressor gene p53 and lung cancer.
Lung Cancer

2001
;
31
:
157
–62.
15
Hahn M, Serth J, Fislage R, Wolfes H, Allhoff E, Jonas V, et al. Polymerase chain reaction detection of a highly polymorphic VNTR segment in intron 1 of the human p53 gene.
Clin Chem

1993
;
39
:
549
–50.
16
Ito T, Seyama T, Hayashi T, Mizuno T, Iwamoto KS, Tsuyama N, et al. HaeIII polymorphism in intron 1 of the human p53 gene.
Hum Genet

1994
;
93
:
222
.
17
Pleasants LM, Hansen MF. Identification of a polymorphism in intron 2 of the p53 gene.
Hum Genet

1994
;
93
:
607
–8.
18
Runnebaum IB, Tong XW, Moebus V, Heilmann V, Kieback DG, Kreienberg R. Multiplex PCR screening detects small p53 deletions and insertions in human ovarian cancer cell lines.
Hum Genet

1994
;
93
:
620
–4.
19
Lazar V, Hazard F, Bertin F, Janin N, Bellet D, Bressac B. Simple sequence repeat polymorphism within the p53 gene.
Oncogene

1993
;
8
:
1703
–5.
20
Chumakov PM, Jenkins JR. BstNI/NciI polymorphism of the human p53 gene (TP53).
Nucleic Acids Res

1991
;
19
:
6969
.
21
Hillebrandt S, Streffer C, Demidchik EP, Biko J, Reiners C. Polymorphisms in the p53 gene in thyroid tumours and blood samples of children from areas in Belarus.
Mutat Res

1997
;
381
:
201
–7.
22
Prosser J, Condie A. Biallelic ApaI polymorphism of the human p53 gene (TP53).
Nucleic Acids Res

1991
;
19
:
4799
.
23
Buller RE, Skilling JS, Kaliszewski S, Niemann T, Anderson B. Absence of significant germ line p53 mutations in ovarian cancer patients.
Gynecol Oncol

1995
;
58
:
368
–74.
24
Mattick JS. Introns: evolution and function.
Curr Opin Genet Dev

1994
;
4
:
823
–31.
25
Malkinson AM, You M. The intronic structure of cancer-related genes regulates susceptibility to cancer.
Mol Carcinog

1994
;
10
:
61
–5.
26
Lozano G, Levine AJ. Tissue-specific expression of p53 in transgenic mice is regulated by intron sequences.
Mol Carcinog

1991
;
4
:
3
–9.
27
Shamsher M, Montano X. Analysis of intron 4 of the p53 gene in human cutaneous melanoma.
Gene

1996
;
176
:
259
–62.
28
Avigad S, Barel D, Blau O, Malka A, Zoldan M, Mor C, et al. A novel germ line p53 mutation in intron 6 in diverse childhood malignancies.
Oncogene

1997
;
14
:
1541
–5.
29
Runnebaum IB, Tong XW, Konig R, Zhao H, Korner K, Atkinson EN, et al. p53-based blood test for p53PIN3 and risk for sporadic ovarian cancer [letter].
Lancet

1995
;
345
:
994
.
30
Mavridou D, Gornall R, Campbell IG, Eccles DM. TP53 intron 6 polymorphism and the risk of ovarian and breast cancer.
Br J Cancer

1998
;
77
:
676
–7.
31
Lancaster JM, Brownlee HA, Wiseman RW, Taylor J. p53 polymorphism in ovarian and bladder cancer.
Lancet

1995
;
346
:
182
.
32
Campbell IG, Eccles DM, Dunn B, Davis M, Leake V. p53 polymorphism in ovarian and breast cancer.
Lancet

1996
;
347
:
393
–4.
33
Birgander R, Sjalander A, Rannug A, Alexandrie AK, Sundberg MI, Seidegard J, et al. P53 polymorphisms and haplotypes in lung cancer.
Carcinogenesis

1995
;
16
:
2233
–6.
34
Peller S, Kopilova Y, Slutzki S, Halevy A, Kvitko K, Rotter V. A novel polymorphism in intron 6 of the human p53 gene: a possible association with cancer predisposition and susceptibility.
DNA Cell Biol

1995
;
14
:
983
–90.
35
Sjalander A, Birgander R, Athlin L, Stenling R, Rutegard J, Beckman L, et al. P53 germ line haplotypes associated with increased risk for colorectal cancer.
Carcinogenesis

1995
;
16
:
1461
–4.
36
Birgander R, Sjalander A, Zhou Z, Fan C, Beckman L, Beckman G. p53 polymorphisms and haplotypes in nasopharyngeal cancer.
Hum Hered

1996
;
46
:
49
–54.
37
Sjalander A, Birgander R, Hallmans G, Cajander S, Lenner P, Athlin L, et al. p53 polymorphisms and haplotypes in breast cancer.
Carcinogenesis

1996
;
17
:
1313
–6.
38
Weston A, Pan CF, Ksieski HB, Wallenstein S, Berkowitz GS, Tartter PI, et al. p53 haplotype determination in breast cancer.
Cancer Epidemiol Biomarkers Prev

1997
;
6
:
105
–12.
39
Zhao H, Spitz MR, Tomlinson GE, Zhang H, Minna JD, Wu X. Gamma-radiation-induced G2 delay, apoptosis, and p53 response as potential susceptibility markers for lung cancer.
Cancer Res

2001
;
61
:
7819
–24.
40
Guo SW, Thompson EA. Performing the exact test of Hardy-Weinberg proportion for multiple alleles.
Biometrics

1992
;
48
:
361
–72.
41
Lewontin RC, Feldman MW. A general asymptotic property of two-locus selection models.
Theor Popul Biol

1988
;
34
:
177
–93.
42
Thomson G, Baur MP. Third order linkage disequilibrium.
Tissue Antigens

1984
;
24
:
250
–5.
43
Breslow NE, Day NE. Statistical methods in cancer research, Vol. 1—The analysis of case-control studies.
IARC Sci Publ

1980
;
32
:
5
–338.
44
STATA. Version 7. College Station (TX): STATA, 1999.
45
Terwillinger J, Ott J. Handbook of human genetic linkage. Baltimore (MD): Johns Hopkins University Press; 1994.
46
Statistical Software for Exact Nonparametric Inference, StatXact-4, User Manual, Version 4.0.1. Cytel Software Corporation, Cambridge, MA, 2000.
47
Thomas M, Kalita A, Labrecque S, Pim D, Banks L, Matlashewski G. Two polymorphic variants of wild-type p53 differ biochemically and biologically.
Mol Cell Biol

1999
;
19
:
1092
–100.
48
Marin MC, Jost CA, Brooks LA, Irwin MS, O'Nions J, Tidy JA, et al. A common polymorphism acts as an intragenic modifier of mutant p53 behaviour.
Nat Genet

2000
;
25
:
47
–54.
49
Brooks LA, Tidy JA, Gusterson B, Hiller L, O'Nions J, Gasco M, et al. Preferential retention of codon 72 arginine p53 is squamous cell carcinomas of the vulva occurs in cancers positive and negative for human papillomavirus.
Cancer Res

2000
;
60
:
6875
–7.
50
Wei Q, Cheng L, Amos CI, Wang LE, Guo Z, Hong WK, et al. Repair of tobacco carcinogen-induced DNA adducts and lung cancer risk: a molecular epidemiologic study.
J Natl Cancer Inst

2000
;
92
:
1764
–72.
51
Robels AI, Harris CC. P53 mediated apoptosis and genomic instability diseases.
Acta Oncol

2001
;
40
:
696
–701.
52
Lehman TA, Haffty BG, Carbone CJ, Bishop LR, Gumbs AA, Krishnan S, et al. Elevated frequency and functional activity of a specific germ-line p53 intron mutation in familial breast cancer.
Cancer Res

2000
;
60
:
1062
–9.
53
Sjalander A, Birgander R, Kivela A, Beckman G. p53 polymorphisms and haplotypes in different ethnic groups.
Hum Hered

1995
;
45
:
144
–9.
54
Wang-Gohrke S, Weikel W, Risch H, Vesprini D, Abrahamson J, Lerman C, et al. Intron variants of the p53 gene are associated with increased risk for ovarian cancer but not in carriers of BRCA1 or BRCA2 germline mutations.
Br J Cancer

1999
;
81
:
179
–83.
55
Perera FP. Environment and cancer: who are susceptible?
Science

1997
;
278
:
1068
–73.
56
Ries LA, Hankeys BF, Miller BA, Hartman A, Edwards BK. SEER Cancer Statistics Review 1973–1988. Bethesda (MD): National Institutes of Health, National Cancer Institute; 1991 Publication No. 91–2789.
57
U.S. Dept. of Health and Human Services. 1989 Vital and Health Statistics: Deaths of Hispanic Origin, 15 Reporting States, 1979–81, Series 20: Data from the National Vital Statistics System No. 18, DHHS publication no [PHS] 91–1855, National Center for Health and Human Services. Hyattsville (MD): National Center for Health Statistics.