Association of Polymorphisms in the Paraoxonase 1 Gene with Breast Cancer Incidence in the CPS-II Nutrition Cohort

Paraoxonase 1 (PON1) is found in serum associated with high-density lipoproteins (1). Although the natural substrate(s) of this enzyme are unknown, it is capable of the hydrolysis of a variety of xenobiotics, including organophosphates and aryl esters (2), that are common components of insecticides and nerve agents. Paraoxonase has been shown to metabolize lipid-soluble radicals (1) and contribute to the high-density lipoprotein–mediated prevention of low-density lipoprotein oxidation (3), suggesting that it plays an important role in the anti-inflammatory effect of high-density lipoproteins. Because oxidized low-density lipoproteins contribute to the development of atherosclerosis, studies have examined whether paraoxonase activity is associated with increased or decreased cardiovascular disease risk. Patients with cardiovascular disease have lower serum levels of paraoxonase than matched controls (reviewed in ref. 3). Additionally, a polymorphism in the PON1 gene known as Q192R, that changes the amino acid sequence and reduces enzymatic activity, was associated with significantly increased risk of heart disease in a meta-analysis of 43 studies (4). These findings suggest that lower paraoxonase activity may increase risk of cardiovascular disease.

Although oxidative stress and inflammation are believed to be important in carcinogenesis, there has been little investigation of changes in PON1 activity in cancer. Serum levels of paraoxonase were lower in patients with gastric or pancreatic cancer than in healthy controls in two small case-control studies (5, 6). Four studies of PON1 in cancer have focused on the Q192R single nucleotide polymorphism (SNP; ref. 7) or a second polymorphism, L55M (8). The L55M SNP also changes the amino acid sequence and reduces paraoxonase protein levels but was not associated with heart disease in the large meta-analysis mentioned above (4). Neither polymorphism was associated with colorectal cancer incidence (9), whereas the 192RR genotype was associated with significantly higher risk of non-Hodgkin's lymphoma [odds ratio (OR), 2.92; 95% confidence interval (95% CI), 1.49-5.72; ref. 10] in a moderately sized case-control study. A cohort study found no association between either polymorphism and prostate cancer (11), whereas a case-control study found significant positive associations for both the Q192R and L55M SNPs and prostate cancer (12). Cumulatively, these data do not indicate whether PON1 is likely to influence cancer development.

In this study, we investigated the association of the Q192R and L55M PON1 SNPs with breast cancer, the most common incident cancer in U.S. women (13). Because many of the risk factors for this cancer, including BRCA1 mutations and increased estrogen metabolism, are known to influence oxidative stress (14), it is reasonable to expect that PON1 may influence breast cancer development. Both of these polymorphisms are common and influence PON1 activity. The Q192R SNP results in a replacement of a glutamine with an arginine, whereas the L55M SNP changes a leucine to a methionine. The Q192R SNP affects the enzymatic activity directly, with the type and magnitude of the change observed dependent on the substrate used (7), whereas the L55M SNP alters levels of the PON1 protein (8).

Participants were drawn from the 21,965 women who provided a blood sample as part of the American Cancer Society Cancer Prevention II Nutrition Cohort. This is a prospective study of cancer incidence among ∼184,000 U.S. adults, ages 50 to 74 years, who completed a mailed questionnaire in 1992 that included questions on demographics, diet, and other lifestyle factors. The recruitment and characteristics of this cohort have been described previously (15). Follow-up questionnaires were sent to all living Nutrition Cohort members in 1997 and every 2 years afterwards to update exposure information and to ascertain newly diagnosed cases of cancer. Incident cases reported via questionnaire response were verified through medical records, linkage with state cancer registries, or death certificates. Between June 1998 and June 2001, blood samples were collected from of subset of Nutrition Cohort participants (21,965 women and 17,411 men) who lived in urban or suburban areas in 20 of the 21 states included in the larger cohort and were believed to be alive and cancer-free when invited to donate blood. Each blood sample was fractionated into serum, plasma, buffy coat, and RBCs and stored in liquid nitrogen vapor phase at −130°C until needed for analysis.

We identified 502 postmenopausal women who were diagnosed with breast cancer between 1992 and 2001 and had no history of any other cancer (except nonmelanoma skin cancer) from among the women who provided a blood sample. Of these, 361 women had invasive breast cancer (stage I or higher); the remainder had in situ cancer. An equal number of controls were matched to cases on age (±6 months), race/ethnicity (White, African American, Hispanic, Asian, or other/unknown), and date of blood collection (±6 months). Controls were selected from women who were cancer-free at the time of diagnosis of the matching case using risk-set sampling (16).

The questionnaire information on demographic characteristics, reproductive history, medication use, and personal and family history was collected at enrollment in 1992 to 1993. Thus, exposure information was collected before the cancer diagnosis for all cases.

Genotyping for the Q192R and L55M SNPs was done using Taqman at Applied Biosystems, Inc. (Foster City, CA). Overall success rate for genotyping was >98%, and the genotype distributions of the controls were all found to be in Hardy-Weinberg equilibrium (P > 0.05).

Conditional logistic regression was used to calculate ORs and 95% CIs for the association between the PON1 SNPs and breast cancer incidence. Other covariates that were considered for the analysis were oral contraceptive use, breast cyst, physical activity, hormone replacement therapy use, parity, education, and alcohol intake. None of these covariates were included in the logistic model because they did not alter the OR by >5%, either separately or in combination.

Effect modification of the association between the SNPs and breast cancer was examined using the likelihood ratio test for tobacco smoking (never versus ever), nonsteroidal anti-inflammatory drug (NSAID) use (nonusers versus current users), regular NSAID use (nonusers versus users of ≥15 times a month), vitamin C use (below versus above median), vitamin E use (below versus above median), diabetes (no/yes), alcohol use (none versus any), dietary fat intake (below versus above median), and statin use (none versus any). Significance was determined by calculating P for the difference in the trends among genotypes for each strata. All reported Ps were two sided, and Ps ≤ 0.05 were defined as being statistically significant.

The associations between the Q192R and L55M polymorphisms and breast cancer incidence are shown in Table 1. No significant association was found for the Q192R SNP for either all cases or invasive cases of breast cancer. In contrast, L55M was associated with a significantly higher risk of breast cancer. Among all cases (Table 1A), the OR increased from 1.23 (95% CI, 0.92-1.65) for the heterozygous LM genotype to 1.58 (95% CI, 1.05-2.37) for the homozygous MM genotype (P_trend = 0.03). This association was somewhat stronger when only invasive cases were analyzed (OR, 1.85; 95% CI, 1.14-3.00 for MM; Table 1B). As with all cases, the magnitude of the association was greater with two copies of the variant allele (P_trend = 0.01), suggesting that it functioned in a codominant manner. None of the combinations of genotypes were significantly associated with breast cancer incidence (Table 1C).

PON1 activity and expression are potentially regulated by a number of factors that influence oxidative stress, including environmental exposures, dietary factors, medications (NSAIDS, statins, and fibrates), and some health conditions (diabetes; reviewed in refs. 17, 18). NSAID use is common among the women of the CPS-II Nutrition Cohort (55% reported being current users at enrollment; ref. 19) as well as those included in this nested case-control study (59% current users). NSAID use is not significantly associated with altered risk of breast cancer within the CPS-II Nutrition Cohort (19) or this nested case-control population [OR, 1.11; 95% CI; 0.83-1.49 for current use and OR, 1.24; 95% CI, 0.80-1.92 for regular use (defined as ≥15 times per month)]. However, we found that the association of the L55M SNP with breast cancer was significantly modified by NSAID use (Table 2). The risk associated with the M allele was limited to women who took NSAIDs (P_interaction = 0.04), regardless of whether current use (any) or regular use (≥15 times per month) was analyzed. Risk was somewhat higher for regular users (OR, 3.24; 95% CI, 1.17-9.00 for MM) than for women who reported any use (for MM: OR, 2.46; 95% CI, 1.39-4.36 for MM). Similar results were seen when only invasive cases were considered (data not shown). The significant level of these interactions was reduced (P_interaction = 0.32) for both current and regular NSAID use after Bonferroni correction for multiple testing.

NSAID use did not modify the association of the Q192R SNP with breast cancer incidence. None of the other factors tested, which included smoking, vitamin C use, vitamin E use, dietary fat intake, statin use, or diabetes, modified the association of either the L55M SNP or Q192R SNP with breast cancer.

The L55M SNP, which results in an amino acid change in the third exon of the PON1 gene, lowers paraoxonase activity by decreasing the amount of this enzyme present in blood (8, 20). We found that women with the MM genotype for this polymorphism had a 57% higher incidence of breast cancer and an 85% higher incidence of invasive breast cancer. In both cases, the M allele behaved in a codominant manner. This is consistent with measurements of blood activity levels in which individuals with the LM genotype were found to have paraoxonase activity levels between those of LL and MM individuals (20). Furthermore, our finding that risk of breast cancer was elevated even among women who had only one copy of the M allele suggests that even small changes in paraoxonase activity may be important.

Aspirin use has been found to be associated with increased serum levels of paraoxonase activity (21). The actual concentration of the PON1 protein was also increased (21), suggesting that aspirin influenced activity by altering the levels of this enzyme. We found that any use of NSAID significantly modified the association between the L55M SNP and breast cancer incidence. The association was stronger and limited to women who took NSAIDs. The effect modification was greater when only women who took NSAIDs regularly were analyzed, suggesting a dose dependence of the drug effect. However, these interactions were no longer statistically significant after Bonferroni correction for multiple testing. If real, these results could reflect a greater difference in the serum levels of PON1 between LL genotype carriers and LM and MM genotype carriers among women who use NSAIDs. The relationship of the L55M polymorphism to NSAID-induced changes in serum paraoxonase levels has not been investigated.

The Q192R SNP was not associated with breast cancer incidence in this study nor was any of combination of genotypes of the Q192R and L55M SNPs. Although somewhat different than the L55M SNP in that it affects PON1 enzymatic activity directly, this polymorphism also decreases the high-density lipoprotein–mediated protection of low-density lipoprotein oxidation. Therefore, if this plays an important role in breast cancer, the Q192R SNP might also be expected to be associated with increased rates of this disease. However, there is a moderate degree of linkage disequilibrium between two SNPs (D′ = 0.725), and the rare allele of one is commonly found with the common allele of the other (22). Thus, the rare occurrence of the 192R and 55M alleles (r² = 0.087) largely precludes both SNPs being associated with disease in the same population.

Strengths of this study are the prospective design, which limits recall bias in the questionnaire information, and availability of self-reported data on a large number of potential interacting factors. The primary limitation of this analysis is the sample size, which limits the statistical power to examine interactions. Our findings suggest that postmenopausal women with the 55MM PON1 genotype may have increased risk for breast cancer, and that NSAID use may interact to increase this risk. However, because this is the first study of PON1 in breast cancer, replication is needed in larger studies.