Genetic Variation in Myeloperoxidase Modifies the Association of Serum α-Tocopherol with Aggressive Prostate Cancer among Current Smokers

Abstract

We investigated associations of serum α- and γ-tocopherols and their effect modification by polymorphisms in oxidative stress regulatory enzymes in relation to prostate cancer risk. In a nested case-control study in the Carotene and Retinol Efficacy Trial, prerandomized serum α- and γ-tocopherol were assayed among 684 men with incident prostate cancer [375 nonaggressive and 284 aggressive cancer (stage III/IV or Gleason score ≥7)] and 1441 controls. Manganese superoxide dismutase Ala-16Val (rs4880), glutathione peroxidase 1 Pro200Leu (rs1050450), catalase −262 C > T (rs1001179), and myeloperoxidase (MPO) G–463A (rs2333227) were genotyped. A multivariate-adjusted inverse association of serum α-tocopherol with total prostate cancer risk was observed in current smokers (OR = 0.62, 95% CI = 0.40–0.96, 4th vs. 1st quartiles). High (≥median) compared to low serum concentrations of α- and γ-tocopherol were inversely associated with aggressive prostate cancer in current smokers (OR = 0.50, 95% CI = 0.32–0.78 and OR = 0.64, 95% CI = 0.43–0.95, respectively). The association was stronger among those with MPO G/A+A/A genotypes. Among current smokers with low serum α-tocopherol concentrations, MPO G/A+A/A, the genotypes downregulating oxidative stress, were associated with an increased risk for aggressive prostate cancer (OR = 2.06, 95% CI = 1.22–3.46). Conversely, current smokers with these genotypes who had high α-tocopherol concentrations had a reduced risk for aggressive prostate cancer (OR = 0.34, 95% CI = 0.15–0.80; P-interaction = 0.001). In conclusion, among current smokers, both high serum α- and γ-tocopherol concentrations were associated with reduced risks of aggressive prostate cancer. The α-tocopherol–associated risks are modified by polymorphism in MPO G–463A.

Introduction

Oxidative stress occurs when antioxidant activities are insufficient to balance the formation of oxidative reactions in the body, resulting in a relative gain of prooxidant over antioxidant functions (1). Vitamin E is an important chain-breaking antioxidant that prevents free radical reactions and lipid peroxidation (2). It is this antioxidant function that may be particularly relevant in relation to the carcinogenesis of several cancers, including prostate (3). However, a major controversy over the association of vitamin E with prostate cancer risk prevails, because a beneficial effect of vitamin E supplements was found in a randomized trial of smokers but not in two other trials not specifically recruiting smokers (4–6). Whether the beneficial association exists only among smokers or other factors, such as genetic polymorphisms related to oxidative stress, play a role is inconclusive.

A hypothesized mechanism of oxidative stress in relation to prostate cancer involves four important enzymes: MnSOD⁸, GPX1, CAT, and MPO (7, 8). MnSOD, located in mitochondria, converts ROS and superoxide radicals to oxygen and H₂O₂. The increased level of H₂O₂ leads to oxidative stress in prostate tissue. Yet, CAT and GPX1 reduce H₂O₂ to H₂O, an action influenced by antioxidant status and, particularly for GPX1, selenium availability. Dietary intake of antioxidants, including selenium, helps facilitate the redox reaction, but oxidative stress, such as heavy smoking, may deter it. On the contrary, endogenous ROS can be produced by MPO as it converts H₂O₂ and chloride anion (Cl^–) into HOCl (9). Research has found that the expression of CAT and MnSOD decreases in prostatic adenocarcinoma tissue (10), and the genotype of MnSOD that confers a higher level of oxidative stress is associated with an increased risk of high-grade prostate cancer among smokers (11).

The primary objective of this study was to examine associations of serum α- and γ-tocopherol with prostate cancer risk in the CARET using a nested case-control study design. We further investigated whether the polymorphisms in four selected oxidative stress regulatory enzymes, MnSOD, GPX1, CAT, and MPO, modified these associations.

Methods

CARET overview.

CARET was a multicenter, randomized, double-blind, placebo-controlled chemoprevention trial to test whether daily supplementation with 30 mg β-carotene and 25,000 IU (13.75 mg) retinyl palmitate would reduce the risk of lung cancer among 18,314 heavy smokers and asbestos-exposed workers. Other details about the design and primary results of CARET have been published (12). In short, the recruitment began in 1985. Prerandomized, baseline serum specimens were collected at the first CARET study center visit. These baseline serum samples of selected cases and controls were used for assaying tocopherols in this study. In 1995, whole blood was collected from a subset (68%) of active, consenting participants for DNA extraction and genotyping. The intervention was stopped in 1996; 94% of participants remained in active follow-up as of 2005. This study was restricted to male participants. The Institutional Review Board of the Fred Hutchinson Cancer Research Center and each of the five other participating institutions approved all procedures for the study, and participants provided written informed consent at recruitment and throughout the study. For the analyses in this study, additional institutional review was obtained from the Roswell Park Cancer Institute.

Case and control selection and endpoint ascertainment.

At each CARET annual visit, as well as in quarterly follow-up telephone calls, participants were asked to report if they had been diagnosed with any new cancers. All endpoints, including prostate cancer, were verified by the CARET Endpoints Committee. For this nested case-control study, participants with a previous report of prostate cancer were approached to request permission to obtain data on Gleason score and stage of disease at diagnosis through review of their medical records. For participants whose medical records were not available, stage and Gleason score were obtained from pathology reports in the Cancer Surveillance System of Western Washington, SEER registry. All medical records and SEER pathology records were reviewed and confirmed by one of the coauthors (G.G.). Case selection for this study was based on follow-up through 2003, by which a total of 778 prostate cancer cases had been confirmed. After excluding 50 men with prior cancer history, reported at the baseline visit, and 19 without specimens available for laboratory analyses, 709 cases were eligible for this study (13). Eligible controls were men who were free of both prostate cancer and lung cancer (the primary end point in CARET) and had available whole blood or extracted DNA. Cases and controls were frequency matched on age (5-y groups), race/ethnicity, and follow-up time to diagnosis of the matched cases. The case-control ratio was 1:4 for blacks, wherever achievable, and 1:2 for other races. As a result, a total of 724 cases and 1474 controls were selected (after reassigning 15 participants who were originally selected as controls and diagnosed subsequently with prostate cancer). Staging information and Gleason scores were available for 627 (87%) and 674 (93%) of the cases, respectively (13). Forty cases (6%) and 33 controls (2%) did not have serum tocopherol data due to insufficient serum or dietary data. Consequently, the current study was restricted to 684 cases and 1441 controls for main analyses. In addition, 219 cases (32%) and 86 controls (6%) did not have complete genotyping data. Participants without information on specific genotypes and prostate cancer disease status were excluded from the corresponding tocopherol/gene stratified analyses.

Serum tocopherols assay.

α-Tocopherol and γ-tocopherol were assayed by HPLC. Briefly, a hexane extract of serum was injected onto a 3-μm C-18 Spherisorb ODS-2 HPLC column and eluted with an isocratic solvent consisting of 73% acetonitrile, 12% tetrahydrofuran, 8% methanol, 7% water, 0.025% ammonium acetate, and 0.05% diethylamine (v:v) at a flow rate of 1.2 mL/min. α-Tocopherol and γ-tocopherol were detected at 292 nm. The CV for the pooled samples was <5% for both tocopherols.

Genotyping.

Polymorphisms for MnSOD Ala–16Val (rs 4880, T to C substitution), GPX1 Pro200Leu (rs 1050450, C to T substitution), CAT –262C > T (rs 1001179), and MPO G–463A (rs 2333227) were selected for genotyping, because their variants have been shown to influence the capacity for responding to oxidative stress. Genomic DNA was extracted with the use of QIAamp DNA blood Midi kits (Qiagen). Genotyping was performed with high-throughput matrix-assisted laser desorption/ionizing time-of-flight MS (Sequenom) by BioServe Biotechnologies. Procedures and primers for PCR were previously reported (9, 14). The interassay agreement were excellent among the 8% of randomly selected duplicates (k statistic: 0.95) with <1% assay failure rate. All polymorphisms in the controls were in Hardy-Weinberg equilibrium.

Other data collection.

Age, race/ethnicity, family history of prostate cancer, smoking history, height, and weight were collected by a self-administered questionnaire at baseline. Current smokers were defined as those who smoked any cigarettes in the past month. BMI in kg/m² was calculated from current weight and height. Alcohol consumption was assessed by averaging multiple self-administered FFQ (15, 16) completed prior to the prostate cancer diagnosis.

Statistical analyses.

The serum tocopherol concentrations were categorized into quartiles based on the distribution in controls. Unconditional logistic regression was used to estimate OR and 95% CI for associations of α- and γ-tocopherols (in separate models) with prostate cancer risk. Tests for linear trend across the tocopherol quartiles were based on the linear contrast of the regression coefficients corresponding to each quartile.

A covariate was included in multivariate models if a priori knowledge suggested that the variable was a confounder and its contribution to the model, evaluated by likelihood ratio tests, was significant. Using these criteria, the multivariate models were adjusted by age at enrollment (continuous), race (white, black, or others), randomization assignment (retinol plus β-carotene or placebo), family history of prostate cancer in first-degree relatives (yes or no), alcohol consumption (nondrinker, below median, or at or above median based on total alcohol amount in controls; the median was 10 g/d), smoking status (current or former/never), smoking pack-year (<40, 40 to <60, or ≥60), BMI (continuous), and serum cholesterol (continuous). We did not include intake of calcium and dairy products or serum lycopene in final models, because they neither contributed to the models nor changed OR estimates. All analyses were conducted first for all participants, followed by stratified analyses by participants’ smoking status (current smokers and former/never smokers) and disease status of the prostate cancer cases (nonaggressive and aggressive disease). Aggressive prostate cancer was defined as stage III or IV (extraprostatic extension or metastasis) tumors or with Gleason score ≥7 (16, 17). We also analyzed the risk for aggressive prostate cancer using Gleason score ≥8 (with the same tumor stages). The finding showed no major differences in OR, but the number of cases was not large enough to provide stable risk estimates. Thus, we presented findings limited to Gleason score ≥7 for aggressive prostate cancer.

To explore whether the main effects of serum tocopherols were modified by genotypes, variables were coded with a common reference group. Serum tocopherol concentrations were dichotomized at median values of the controls to maintain sufficient number of participants in each cell. The reference group for a given model was selected as men with tocopherol concentrations below the median and in the presence of the genotype that confers a higher risk of prostate cancer or other cancer outcomes. Participants with different genotypes were combined into one group if two genotypes have the same transcriptional activity. Thus, strata by genotypes in analyses were Val/Val+Val/Ala and Ala/Ala (reference) for MnSOD (7, 11, 17, 18), Pro/Pro and Pro/Leu+Leu/Leu (reference) for GPX1 (19), C/C and C/T+T/T (reference) for CAT (14, 20), and G/G (reference) and G/A+A/A for MPO (21, 22). A cross-product term of the dichotomized serum tocopherols and the genotypes was created and the likelihood ratio tests were used to examine the interaction.

Results

Cases had slightly lower median concentrations of serum α-tocopherol than controls (Table 1). Serum α- or γ-tocopherol concentrations were not associated with prostate cancer risk in the overall study population (Table 2). Nevertheless, among current smokers, there was an inverse association between serum α-tocopherol concentrations and prostate cancer risk (P-trend = 0.008). Current smokers in the 4th quartile of α-tocopherol had a 38% decreased risk of prostate cancer compared to those in the first quartile. In the analysis stratified by disease status, we observed an inverse trend between serum α-tocopherol concentrations and aggressive prostate cancer risk (P-trend = 0.039). Participants with serum γ-tocopherol concentrations in the 4th quartile had a nonsignificant, 31% reduced risk of aggressive prostate cancer (P = 0.060).

TABLE 1.

Characteristics of prostate cancer cases and controls in the Carotene and Retinol Efficacy Trial (CARET)¹

Characteristics	Cases (n = 684)	Controls (n = 1441)
Age, y
Baseline	60.6 ± 5.7	60.3 ± 5.8
Diagnosis	67.2 ± 6.0	N/A2
Race/ethnicity, n (%)	618 (90.4)	1269 (88.0)
White	42 (6.1)	122 (8.5)
African American	24 (3.5)	50 (3.5)
Other
Family history of prostate cancer, n (%)
Yes	44 (6.4)	47 (3.3)
Smoking status, n (%)
Current	352 (51.5)	761 (52.8)
Never3/former	332 (48.5)	680 (47.2)
Amount smoked, pack-years
Current smokers	52.2 ± 20.3	51.2 ± 21.3
Former smokers	46.5 ± 27.2	45.7 ± 27.2
BMI, kg/m2	28.3 ± 4.3	28.1 ± 4.4
Serum tocopherol, μmol/L
α-Tocopherol	27.9 (23.0–36.5)	28.8 (23.0–36.9)
γ-Tocopherol	6.0 (4.1–8.2)	6.0 (4.3–8.4)
Gleason score, n (%)
<7	368 (53.8)	N/A
≥7	269 (39.3)
Unknown	47 (6.9)
Stage, n (%)
0–I	171 (25.0)	N/A
II	292 (42.7)
III	26 (3.8)
IV	25 (3.7)
Unknown	170 (24.8)
Year of diagnosis,4 n (%)
1986–1993	152 (22.2)	N/A
1994–2005	532 (77.8)

¹Values are mean ± SD, median (IQR), or (%).

²N/A, not applicable.

³Represents <2%. They were recruited in the CARET because of their occupational asbestos exposure.

⁴1994 approximates the advent of the PSA era.

TABLE 2.

OR and 95% CI for prostate cancer risk according to quartiles of serum α- and γ-tocopherol, overall and stratified by smoking status, and disease status in 684 cases and 1441 controls in the Carotene and Retinol Efficacy Trial (CARET)

	Quartile of serum tocopherols
Biomarker/subgroup	First	Second	Third	Fourth	P-trend
α-Tocopherol (range), μmol/L	8.1 to <23.0	23.0 to <28.8	28.8 to <37.1	37.1–115.4
All, n cases/n controls	171/359	199/360	151/362	163/360
OR (95% CI)1	1.00 (ref)	1.08 (0.83–1.41)	0.80 (0.60–1.07)	0.82 (0.61–1.10)	0.06
Smoking status
Former smokers, n cases/n controls	70/132	80/181	73/179	109/188
OR (95% CI)2	1.00 (ref)	0.82 (0.54–1.22)	0.74 (0.48–1.13)	0.97 (0.64–1.48)	0.78
Current smokers, n cases/n controls	101/227	119/179	78/183	54/172
OR (95% CI)2	1.00 (ref)	1.38 (0.97–1.95)	0.87 (0.58–1.29)	0.62 (0.40–0.96)*	0.008
Disease status
Nonaggressive disease, n cases/n controls	93/359	104/360	89/362	89/360
OR (95% CI)1	1.00 (ref)	1.08 (0.78–1.51)	0.89 (0.63–1.27)	0.87 (0.60–1.25)	0.30
Aggressive disease, n cases/n controls	72/359	89/360	58/362	65/360
OR (95% CI)1	1.00 (ref)	1.12 (0.78–1.61)	0.69 (0.46–1.05)	0.73 (0.47–1.11)	0.039
γ-Tocopherol (range), μmol/L	0.43 to <4.3	4.3 to <6.0	6.0 to <8.4	8.4–35.5
All, n	186/359	166/360	173/358	159/364
OR (95% CI)1	1.00 (ref)	0.90 (0.69–1.16)	0.94 (0.72–1.21)	0.83 (0.63–1.09)	0.24
Smoking status
Former smokers, n cases/n controls	90/160	71/173	80/157	91/190
OR (95% CI)2	1.00 (ref)	0.74 (0.50–1.09)	0.92 (0.63–1.34)	0.86 (0.58–1.25)	0.68
Current smokers, n cases/n controls	96/199	95/187	93/201	68/174
OR (95% CI)2	1.00 (ref)	1.04 (0.73–1.48)	0.95 (0.67–1.36)	0.78 (0.53–1.16)	0.19
Disease status
Nonaggressive disease, n cases/n controls	91/359	98/360	94/358	92/364
OR (95% CI)1	1.00 (ref)	1.08 (0.78–1.50)	1.05 (0.76–1.47)	1.01 (0.72–1.42)	0.99
Aggressive disease, n cases/n controls	86/359	63/360	71/358	64/364
OR (95% CI)1	1.00 (ref)	0.74 (0.51–1.06)	0.82 (0.57–1.17)	0.69 (0.47–1.01)	0.11

¹Multivariate adjusting age at enrollment (continuous), race (white, black, others), random assignment (intervention, placebo), family history of prostate cancer in first-degree relatives (yes, no), alcohol consumption (nondrinker, <median, ≥median), smoking status (current, former/never), smoking pack-year (<40, 40 to <60, ≥60), BMI (continuous), and serum cholesterol (continuous).

²Multivariate adjusting variables listed in footnote 1 except for smoking status. * < 0.05.

We next conducted analyses stratified by polymorphisms in the oxidative stress regulatory genes (Table 3). We restricted the analyses to current smokers and estimated risk for aggressive prostate cancer, because the main effects were stronger in smokers or for aggressive tumors (Table 2). In this subgroup, high serum α-tocopherol was associated with a 50% reduced aggressive prostate cancer risk. Genotypes of MnSOD, GPX1, and CAT did not modify the association. However, a significant interaction was observed between the MPO genotype and serum α-tocopherol. MPO G/A and A/A genotypes were associated with a 2-fold increased prostate cancer risk compared to G/G genotypes when serum α-tocopherol concentrations were low. Conversely, men with high serum α-tocopherol concentrations and MPO G/A+A/A genotypes had a 66% reduced risk, compared to those low serum α-tocopherol concentrations and G/G genotypes (P-interaction = 0.001). We conduced a sensitivity analysis including both current and former/never smokers and thus a larger number of aggressive prostate cancer cases (n = 212; 29 in the MPO G/A+A/A and high serum α-tocopherol group) were used. In this sensitivity analysis, the OR of aggressive prostate cancer among men with MPO G/A+A/A and high serum α-tocopherol concentrations compared to those with MPO G/G and low serum α-tocopherol concentrations remained significant (OR = 0.53, 95% CI = 0.33–0.87; P-interaction = 0.11; data not shown). The interaction for the total prostate cancer in the entire population was not significant.

TABLE 3.

OR and 95% CI of serum tocopherols for aggressive prostate cancer risk among current smokers according to genotypes of oxidative-stress regulatory genes¹²

	α-Tocopherol				γ-Tocopherol
	Low (<median)		High (≥median)		Low (<median)		High (≥median)
Genotypes of selected genes	n cases/n controls	OR (95% CI)	n cases/n controls	OR (95% CI)	n cases/n controls	OR (95% CI)	n cases/n controls	OR (95% CI)
All participants	98/406	1.00 (ref)	52/355	0.50 (0.32–0.78)*	84/386	1.00 (ref)	66/375	0.64 (0.43–0.95)*
MnSOD (rs 4880)
Ala/Ala3	17/94	1.00 (ref)	8/93	0.39 (0.16–1.00)	13/95	1.00 (ref)	12/92	0.88 (0.37–2.07)
Ala/Val+Val/Val	50/287	1.01 (0.55–1.84)	26/240	0.44 (0.22–0.91)*	44/265	1.17 (0.60–2.29)	32/262	0.82 (0.40–1.66)
P-interaction				0.83				0.65
GPX1 (rs 1050450)
Pro/Leu+Leu/Leu3	37/174	1.00 (ref)	16/168	0.38 (0.19–0.74)*	31/170	1.00 (ref)	22/172	0.64 (0.35–1.17)
Pro/Pro	31/205	0.74 (0.44–1.26)	18/166	0.42 (0.22–0.80)*	25/189	0.69 (0.39–1.22)	24/182	0.70 (0.39–1.27)
P-interaction				0.37				0.28
CAT (rs 1001179)
C/T+T/T3	20/146	1.00 (ref)	13/130	0.64 (0.29–1.38)	15/124	1.00 (ref)	18/152	1.02 (0.48–2.15)
C/C	50/244	1.77 (1.00–3.13)	19/206	0.58 (0.29–1.19)	40/246	1.54 (0.81–2.95)	29/204	1.27 (0.64–2.52)
P-interaction				0.17				0.64
MPO (rs 2333227)
G/G3	33/250	1.00 (ref)	28/206	0.82 (0.45–1.47)	33/232	1.00 (ref)	28/224	0.84 (0.48–1.47)
G/A+A/A	39/147	2.06 (1.22–3.46)*	8/145	0.34 (0.15–0.80)*	27/149	1.31 (0.75–2.28)	20/143	1.00 (0.54–1.85)
P-interaction				0.001				0.83

¹Multivariate adjusting age at enrollment (continuous), race (white, black, others), random assignment (intervention, placebo), family history of prostate cancer in first-degree relatives (yes, no), alcohol consumption (nondrinker, <median, ≥median), smoking status (current, former/never), smoking pack-year (<40, 40 to <60, ≥60), BMI (continuous), and serum cholesterol (continuous).

²CAT, catalase; GPX1, glutathione peroxidase 1; MnSOD, manganese superoxide dismutase; MPO, myeloperoxidase.

³“High-risk” genotypes, i.e. genotypes that result in higher oxidative stress levels due to the substitutions and transcriptional activities. See text for references. * < 0.05.

Among current smokers, high serum γ-tocopherol concentrations were associated with a 36% reduced risk of aggressive prostate cancer. The patterns of risks for serum γ-tocopherol stratified by the genotypes were largely similar as those for serum α-tocopherol, but the strength of associations was not as strong and none of the interaction tests were significant.

Discussion

In this nested case-control study in the CARET, higher serum α- and γ-tocopherol concentrations were associated with reduced risks of aggressive prostate cancer among current smokers. Our data extend prior literature by suggesting that the association is different among men with genetic variation in MPO G–463A. MPO G/A and A/A genotypes were associated with a nearly 2-fold increased risk of aggressive prostate cancer among current smokers with low serum α-tocopherol concentrations. However, among those with the same genotypes, high serum α-tocopherol concentrations were associated with a 66% reduced risk. To our knowledge, this is the first study investigating the interaction between serum tocopherols and polymorphisms in MPO for prostate cancer. These MPO G/A and A/A allelic variations were present in 40% of our study population. Therefore, our observations are important from a public health perspective, because these genotypes are not rare in the general population (21, 23, 24).

Participants’ smoking status, as well as characteristics of the prostate tumor, may be important determinants of the association of serum α-tocopherol with prostate cancer risk. The inverse association of serum α-tocopherol with prostate cancer risk is more evident among smokers than nonsmokers (25). Observational data from the ATBC Cancer Prevention Trial, Physicians’ Health Study, U.S. Health Professionals’ Follow-up Study, and a Swiss study suggested inverse associations in smokers and former smokers (26–29), but other studies analyzing smokers and nonsmokers together did not (30–33). In randomized controlled trials, compared to placebos, vitamin E supplementation reduced prostate cancer risk in ATBC (4), but not in the Selenium and Vitamin E Cancer Prevention Trial or Physicians’ Health Study II (5, 6), two studies recruiting participants with low smoking prevalence. A possible explanation for these observations is that the requirement for vitamin E and other antioxidants is known to be higher in smokers than in nonsmokers (34). In addition, the majority of studies found inverse associations of α-tocopherol with high-grade prostate cancer and mortality from prostate cancer, but not with low-grade disease (4, 28, 29, 35). In the Selenium and Vitamin E Cancer Prevention Trial, the majority of the prostate tumors ascertained in the trial was localized in part due to the exclusion of a group with high prostate-specific antigen values and widespread use of the screening (36). The reason why antiprostate tumor effects of tocopherols are more relevant to high-grade tumors needs further study. Future investigations should consider the demand of antioxidation and tumor characteristics in study populations.

Literature on whether serum γ-tocopherol is associated with prostate cancer risk is less consistent, although it is biologically plausible. Our findings are consistent with those from ATBC, which showed a reduced risk of prostate cancer in current smokers with higher serum γ-tocopherol concentrations overall, and the association was stronger in the α-tocopherol supplementation arm (26). Studies including both smokers and nonsmokers largely did not find the association (28, 31, 32), except one (30). Whether the characteristics of study participants or phenotypes of prostate tumor studied affect the findings is unclear. With respect to biological mechanisms, γ-tocopherol can quench nitrogen oxides, a ROS that reacts with unsaturated fatty acid and causes inflammation that contributes to prostate carcinogenesis (37, 38). An inverse association of γ-tocopherol with prostate cancer risk can be reasonably hypothesized in smokers, because cigarette smoking is an important external source of nitrogen oxides (39).

In our study population, the polymorphisms in MPO may modify aggressive prostate cancer risk. MPO functions in neutrophils, monocytes, and some macrophages, triggered by inflammation and immune responses (40). HOCl, a secondary free radical produced by reactions involving MPO, is associated with lipid peroxidation and mitochondria DNA damage, which enhances aggressive phenotypes of prostate cancer (3). HOCl can also result in chlorination of DNA base, such as guanine, leading the formation of 8-hydroxyguanine, a known biomarker of DNA base lesion in malignant prostate tissue (41). This oxidation process involving MPO is intensified when nicotine, a major compound of cigarette smoking, presents (42). Compared to the G allele, the MPO A variant confers a 25 times lower transcriptional activity, which downregulates oxidative stress (22). In CARET, MPO AA was associated with a 60% lower risk of aggressive prostate cancer (OR = 0.4, 95% CI = 0.2–0.9) compared to MPO GG+GA (16). In our current analysis, the A allele was associated with an increased risk of aggressive prostate cancer among current smokers with low serum α-tocopherol concentrations. This is in parallel with a previous analysis in CARET, indicating that high iron intake, creating a prooxidant environment, is associated with an increased prostate cancer risk in men with MPO G/A+A/A genotypes (OR = 1.6, 95% CI = 1.0–2.4) (16). A study investigating breast cancer also suggested that the MPO A allele might be associated with an increased risk among women with low fruit and vegetable consumption (OR = 2.09, 95% CI = 0.73–5.95, A/A vs. G/G) (23). On the other hand, our study observed a substantial (66%) reduction in aggressive prostate cancer risk when α-tocopherol concentrations were high in current smokers with MPO G/A and A/A genotypes. The pattern of risk reduction was consistent when both current and former smokers were included, as shown in the sensitivity analysis. A population-based study also showed a decreased breast cancer risk among those with high fruit and vegetable consumption (OR = 0.75, 95% CI = 0.58–0.97, A/A+G/A vs. G/G) (21). The Nurses’ Health Study found a similar result showing that women with MPO A allele plus high plasma carotenoid concentrations had a decreased risk of breast cancer (OR = 0.44, 95% CI = 0.18–1.09) compared to those with MPO G allele plus low carotenoid concentrations. However, the association of MPO genotypes and plasma α-tocopherol with breast cancer was less clear (the study genotyped MPO T–764C, but the genotypes have 100% concordance with G–463A) (24). These observations suggest that whether MPO genotypes are associated with reduced cancer risks, such as breast cancer and aggressive prostate cancer, may depend on the individual’s oxidative stress status.

In our data, although the studied polymorphisms in MnSOD, GPX1, and CAT had no clear role in modifying the associations of serum α- and γ-tocopherol with aggressive prostate cancer, several observations may provide clues for future research. First, the amino acid coded by MnSOD T allele (valine) results in less activity in converting ROS into oxygen and H₂O₂, suggesting that Val is a high risk allele leading to an increased level of oxidative stress (43). However, epidemiological studies suggest that MnSOD C allele (alanine) seems to pose a greater prostate cancer risk than Val allele in men with low intake or serum concentrations of antioxidant micronutrients (7, 17, 18) and in smokers (11, 44). Our findings suggest that high serum α-tocopherol concentrations were associated with a reduced risk of aggressive prostate cancer among current smokers in either genotype. Second, although GPX1 Leu has lower transcriptional activities (19) and is associated with increased risks of lung and breast cancer (19, 45), in CARET, the α-tocopherol–associated prostate cancer risk was independent of the GPX1 polymorphism. This finding should be interpreted with caution, because the activity of GPX1 by genotypes is selenium dependent (46) and information on serum selenium is not available in this CARET population. Third, our data suggest that among current smokers with low serum α-tocopherol concentrations, CAT C/C was associated with an increased risk of aggressive prostate cancer, but the risk may be largely attenuated among those with high serum α-tocopherol. This finding is in parallel with previous research that CAT activities can be modified by fruit/vegetable consumption (20) and women with CAT C/C genotype and high fruit/vegetable intake have a lower risk of breast cancer (14). Because a relatively small number of aggressive prostate cancer cases in CARET might have limited our statistical power on finding these α-tocopherol-gene interactions, replications of our findings with a larger sample size are warranted.

A major strength of our study is its nested case-control design, which used serum measurements prior to prostate cancer diagnosis as the exposure. The grade and stage of prostate cancer were confirmed by both medical records and cancer registry files. Additionally, a large number of cases enabled us to estimate risks for both nonaggressive and aggressive prostate cancer. Nevertheless, there are limitations to this study. First, it may not be appropriate to generalize our study findings to other populations, because the CARET participants were heavy smokers and/or had occupational asbestos exposure. Characteristics related to oxidative stress levels and expression of ROS detoxifying enzymes in the study population may be different from those in other populations. Second, we were not able to control for other potential confounders such as diabetes and serum selenium, because CARET did not have the information. Third, potential effects from the long-term systematic or random variations of serum tocopherols may have biased our OR toward to null, because serum tocopherols were measured at one point in time. Nevertheless, a single measurement of serum tocopherols can represent long-term vitamin E intake to a modest degree (47). Fourth, we conducted several stratified analyses that may lead to an increase in a Type I error. However, because we specified our hypotheses and tests a priori and our findings were consistent with published literature and biological mechanisms, the probability of our significant findings being due to chance alone is low.

In conclusion, our data suggest that, among current smokers, both high serum α- and γ-tocopherol concentrations are associated with reduced risks of aggressive prostate cancer. High serum α-tocopherol concentrations may be particularly important among men with MPO G/A and A/A genotypes in reducing prostate cancer risk.