Myeloperoxidase and superoxide dismutase polymorphisms are associated with an increased risk of developing pancreatic adenocarcinoma

Authors

  • Paul Wheatley-Price MB, ChB,

    1. Department of Medical Oncology and Hematology, Department of Medicine, Princess Margaret Hospital, Toronto, Ontario, Canada
    Search for more papers by this author
  • Kofi Asomaning MD,

    1. Environmental and Occupational Medicine and Epidemiology, Department of Environmental Health, Harvard School of Public Health, Boston, Massachusetts
    Search for more papers by this author
  • Amy Reid BSc,

    1. Environmental and Occupational Medicine and Epidemiology, Department of Environmental Health, Harvard School of Public Health, Boston, Massachusetts
    Search for more papers by this author
  • Rihong Zhai MD, PhD,

    1. Environmental and Occupational Medicine and Epidemiology, Department of Environmental Health, Harvard School of Public Health, Boston, Massachusetts
    Search for more papers by this author
  • Li Su MSc,

    1. Environmental and Occupational Medicine and Epidemiology, Department of Environmental Health, Harvard School of Public Health, Boston, Massachusetts
    Search for more papers by this author
  • Wei Zhou MD, PhD,

    1. Environmental and Occupational Medicine and Epidemiology, Department of Environmental Health, Harvard School of Public Health, Boston, Massachusetts
    Search for more papers by this author
  • Andrew Zhu MD, PhD,

    1. Division of Medical Oncology and Hematology, Department of Medicine, Massachusetts General Hospital, Boston, Massachusetts
    Search for more papers by this author
  • David P. Ryan MD,

    1. Division of Medical Oncology and Hematology, Department of Medicine, Massachusetts General Hospital, Boston, Massachusetts
    Search for more papers by this author
  • David C. Christiani MD, MPH,

    1. Environmental and Occupational Medicine and Epidemiology, Department of Environmental Health, Harvard School of Public Health, Boston, Massachusetts
    2. Division of Medical Oncology and Hematology, Department of Medicine, Massachusetts General Hospital, Boston, Massachusetts
    Search for more papers by this author
  • Geoffrey Liu MD

    Corresponding author
    1. Department of Medical Oncology and Hematology, Department of Medicine, Princess Margaret Hospital, Toronto, Ontario, Canada
    2. Division of Medical Oncology and Hematology, Department of Medicine, Massachusetts General Hospital, Boston, Massachusetts
    3. Division of Applied Molecular Oncology, Department of Medical Biophysics, Ontario Cancer Institute, Toronto, Ontario, Canada
    • Department of Medical Oncology, Princess Margaret Hospital, 610 University Avenue, Toronto, Ontario, M5G 2M9
    Search for more papers by this author
    • Fax: (416) 946-6546


Abstract

BACKGROUND.

Pancreatic cancer risk has been linked to chronic pancreatitis and periodontitis, suggesting a role for inflammation in disease etiology. Myeloperoxidase (MPO) and superoxide dismutase (SOD2) are enzymes that regulate reactive oxygen species and contain recognized single nucleotide polymorphisms (SNPs) that confer altered enzyme activity.

METHODS.

One hundred twenty-two patients with pancreatic cancer and 331 age- and sex-matched controls were analyzed for polymorphisms of the MPO − guanine 463 adenine (−G463A) and the SOD2 alanine (Ala)-to-valine (Val) polymorphism at codon 16 (Ala16Val) genes. Cases and controls were analyzed for associations between these polymorphisms, adjusting for sex, age, history of alcohol use and smoking history.

RESULTS.

The variant A allele of MPOG463A was associated with a lower risk of pancreatic cancer (adjusted odds ratio [OR] for pancreatic cancer, 0.57; 95% confidence interval [95% CI], 0.4–0.9; P = .02). The SOD2 homozygous variant genotype (Val/Val) was associated with a greater risk of pancreatic cancer (adjusted OR, 1.96; 95% CI, 1.0–3.8; P = .04). Compared with individuals who carried both low-risk alleles (A/− and Ala/−), significantly more cases than controls carried both high-risk genotypes (G/G and Val/Val; adjusted OR, 4.31; 95% CI, 1.8–10; P = .001), or 1 high-risk genotype (adjusted OR, 1.96; 95% CI, 1.1–3.4; P = .01).

CONCLUSIONS.

Polymorphisms of the inflammatory pathway genes MPOG463A and SOD2Ala16Val are associated with elevated pancreatic cancer risk. Oxidative stress may play an important role in pancreatic cancer carcinogenesis. Cancer 2008. © 2008 American Cancer Society.

Pancreatic cancer carries a dismal prognosis with a mortality rate close to the incidence rate.1, 2 In 2002, there were approximately 232,000 cases worldwide and 226,000 deaths from the disease.2 The majority of patients present with unresectable, locally advanced, or metastatic disease; and, even with optimal chemotherapy, the median survival is under 6 months.3 Chronic pancreatitis is an independent risk factor for developing pancreatic adenocarcinoma in both hereditary and sporadic disease.4–6 Other clinically recognized risk factors include cigarette smoking, certain ethnic groups (including African Americans and New Zealand Maori), and increasing age. Diabetes mellitus has been linked to pancreatic cancer risk, although it also can be an early manifestation of the disease. Although alcohol consumption is the major risk factor for developing chronic pancreatitis, its role as a risk factor for pancreatic cancer is poorly established.7–9 High-penetrant, low-prevalent, inherited mutations account for less than 10% of pancreatic cancers.7, 10

The inflammatory pathway genes myeloperoxidase (MPO), and superoxide dismutase (SOD2) are involved with regulation of inflammation through reactions with hydrogen peroxide (H2O2). The MPO enzyme is expressed abundantly in neutrophils, in which its antimicrobial function converts H2O2 to the bacterocidal and DNA-damaging hypochlorous acid. The MPO gene contains a common single nucleotide polymorphism (SNP) within the MPO gene promoter. This −guanine 463 adenine (−G463A) base transition has been identified at the SP1 binding site, where the variant A allele is associated with reduced messenger RNA (mRNA) expression, resulting in approximately 25 times less transcription activity than the G allele. This results in reduced binding of SP1 and, hence, lower inflammatory potential.11 MPO has been studied most commonly in relation to lung cancer, in which several early publications reported a reduction in the risk of lung cancer for carriers of the homozygous AA variant.12–15 More recent publications, generally with larger numbers, have suggested that there may be no greater risk after all.16–18

Conversely, SOD2 protects cells from reactive oxygen species (ROS) by converting superoxide radicals, which are produced by mitochondrial activity, into H2O2 and oxygen.19 A single thymine (T) to cytosine (C) base change in the mitochondrial targeting sequence (MTS) of the SOD2 gene results in a valine (Val) to alanine (Ala) amino acid change in the MTS.20 It has been demonstrated that Val-containing genotypes carry an increased risk of breast cancer, prostate cancer, lung cancer, and bladder cancer.21–24 However, in other studies, Ala-containing genotypes have shown an increased risk of breast cancer, prostate cancer, and ovarian cancer.20, 25, 26 The functional significance of this SNP remains unclear.27–29

In the current study, we postulated that chronic, low-level inflammation is important to the development of pancreatic cancer through the excess production of potentially carcinogenic ROS. The function of the MPOG463A polymorphism and the SOD2 Ala-to-Val polymorphism at codon 16 (Ala16Val) may have an impact on the capacity to regulate ROS, thereby affecting the risk of pancreatic carcinoma. We investigated this hypothesis by using a case–control design.

MATERIALS AND METHODS

Study Population

In total, 122 patients with histologically confirmed pancreatic adenocarcinoma were recruited prospectively at Massachusetts General Hospital between August 2003 and September 2005. All patients were aged >18 years and were diagnosed within 6 months before study entry. Between 2000 and 2005, other healthy controls also were recruited at Massachusetts General Hospital; and, for the current analysis, 331 of these individuals were frequency matched for age and sex. The controls were healthy nonblood-related family members (usually spouses) and friends of other cancer/surgical patients and were used as a shared set of controls for aerodigestive cancers. Individuals in the control group who had a prior diagnosis of cancer (except nonmelanoma skin cancer) were excluded from participation. After obtaining consent, a standardized questionnaire was administered, and whole blood was drawn from all participants.

Genotyping

DNA was extracted by using the Puregene DNA Isolation Kit (Gentra Systems, Minneapolis, Minn) and was stored at −70 °C until analysis. The MPOG463A and SOD2Ala16Val (G to A) polymorphisms were detected by using an allelic discrimination assay (ABI 7900HT Taqman assay; Applied Biosystems, Foster City, Calif). For MPOG463A, the primers used were TCTTG GGCTGGTAGTGC (forward) and GTATTTTTAGTAGAT ACAGGGTTTCA (reverse). The probes used were AGGCTGAGGCAGGTGGAT (FAM allele-specific probe) and TGAGGCGGGTGGATCACT (VIC allele-specific probe). Polymerase chain reaction conditions consisted of an initial reaction at 95 °C for 15 minutes, then 55 cycles at 95 °C for 30 seconds alternating with 60 °C for 60 seconds. For SOD2Ala16Val, the primers used were GCTGTGCTTTCTCGTCTTCAG (forward) and CTGCCTGGAGCCCAGATAC (reverse). The probes used were CCCAAAACCGGAGCC (FAM) and CCAAAG CCGGAGCC (VIC). Polymerase chain reaction conditions consisted of an initial reaction at 95°C for 15 minutes, then 50 cycles at 95°C for 30 seconds and at 60 °C for 60 seconds. The reaction volumes were 3 μL for both SNPs.

For quality control, positive and negative controls and blinded duplicate samples were run. Alternative genotyping approaches were used as required to verify technical reliability and accuracy. Blinded repeat samples were run in 10% of all samples. A second scientist checked all laboratory interpretations independently.

Statistical Analysis

We analyzed pancreatic cancer cases and controls with information on age, sex, smoking status (nonsmokers, exsmokers, and current smokers), pack-years of smoking, and alcohol history (nondrinkers, which we defined as <2 alcoholic drinks per year, or drinkers). Univariate and multivariate analyses of the data were performed, and significant differences in the distribution of population characteristics between cases and controls were tested by using the Fisher exact test and the Student t test. Hardy-Weinberg equilibrium was tested on the controls by using the chi-square test. When the frequency of the homozygous variant genotype was very low, heterozygous and homozygous variant genotypes were combined. Analyses of associations between the genetic polymorphisms and the risk of pancreatic cancer were based on logistic regression models. When appropriate, crude and adjusted odds ratios (ORs) and 95% confidence intervals (95% CIs) for the risk of pancreatic cancer were calculated. For adjusted analyses, we included age, sex, and pack-year smoking history. Because alcohol data were unavailable for a proportion of controls, we performed sensitivity analyses on alcohol status. All statistical analyses were undertaken by using the SAS statistical packages (version 9.1; SAS Institute, Cary, NC).

RESULTS

The demographic profile and genotype frequencies of the 122 cases and 331 controls are shown in Table 1. The 2 groups were balanced well by age, sex, and smoking status. Because the selected group of controls came from a shared pool of controls that were recruited for aerodigestive cancers, the controls from 2000 to 2001 did not have complete alcohol data. Thus, alcohol consumption data were not available in 112 controls. All genotypes in the control population were in Hardy-Weinberg equilibrium. In the control group, the MPO minor allele frequency was 29%, and the SOD2 minor allele frequency was 43%. These rates were similar to those reported in other Caucasian control groups (MPOA allele, 21%–30%; SOD2Val allele, 44%–51%).24, 30, 31 Increasing age, alcohol consumption, and cigarette smoking were associated independently with an increased risk of developing pancreatic cancer.

Table 1. Demographics and Genotypes in Cases and Controls
VariableNo. (%)P*
Cases (n = 122)Controls (n = 331)
  • MPO indicates myeloperoxidase; G, guanine; A, adenine; SOD2, superoxide dismutase; Ala, alanine; Val, valine.

  • *

    Determined using the chi-square test unless stated.

  • Determined using the Student t test.

  • A total of 112 controls had no available data regarding alcohol history.

  • §

    Three-way analysis (G/G vs G/G vs A/A).

  • Two-way analysis (G/G vs A−–).

  • Three-way analysis (Val/Val vs Ala/Ala vs Ala/Val).

  • #

    Two-way analysis (Val/Val vs Ala−–).

Median age [range], y64 [33–90]64 [37–82].74
Sex
 Men65 (53)162 (49).41
 Women57 (47)169 (51) 
Smoking history
 Never-smoker47 (38)119 (36) 
 Exsmoker59 (48)180 (54) 
 Current smoker16 (13)32 (10).41
 Median pack-y smoking in current and exsmokers [range]27 [0.2–166]26 [0.1–156].04
Alcohol history
 No alcohol (≤2 drinks per y)20 (16)57 (26) 
 Drinks alcohol (>2 drinks per y)102 (84)162 (74).04
MPO
 G/G79 (65)170 (51).01§
 G/A37 (30)132 (40).03
 A/A6 (5)29 (9) 
SOD2
 Ala/Ala31 (25)105 (32)11
 Ala/Val58 (48)165 (50).04#
 Val/Val33 (27)61 (18) 

The MPOG463A Polymorphism

The wild-type genotype (G/G) was more common in patients with pancreatic cancer than in controls (65% vs 51%; P = .03). The crude and adjusted ORs for developing pancreatic cancer by genotype are displayed in Table 2. Individuals who had the heterozygous G/A genotype were associated with a statistically significant reduced risk of developing pancreatic cancer with an adjusted OR of 0.56 (95% CI, 0.3–0.9; P = .02). The OR for the homozygous variant genotype (A/A) was of a similar magnitude and direction at 0.62 (95% CI, 0.2–1.8; P = .37), although there were only 6 cases and 29 controls with the A/A variant. Therefore, we combined the A/A and G/A groups for all other analyses. The adjusted OR for individuals with at least 1 variant A allele was 0.57 (95% CI, 0.4–0.9; P = .02).

Table 2. Crude and Adjusted Odds Ratios for Developing Pancreatic Cancer by Genotype
VariableCrude (122 Cases and 331 Controls)Adjusted (122 Cases and 202 Controls)*Adjusted (122 Cases and 312 Controls)
OR [95% CI]POR [95% CI]POR [95% CI]P
  • OR indicates odds ratio; 95% CI, 95% confidence interval; MPO, myeloperoxidase; G, guanine; A, adenine; SOD2, superoxide dismutase; Val, valine; Ala, alanine.

  • *

    Adjusted for age, sex, and smoking and alcohol history (the smaller cohort was because of missing data regarding alcohol history for 112 patients in the control group).

  • Adjusted for age, sex, and smoking history.

MPO
 A/A vs G/G0.44 [0.2–1.1].080.62 [0.2–1.8].370.43 [0.2–1.1].07
 G/A vs G/G0.60 [0.4–0.9].030.56 [0.3–0.9].020.59 [0.4–0.9].03
 A/− vs G/G0.57 [0.4–0.9].010.57 [0.4–0.9].020.56 [0.4–0.9].01
SOD2
 Val/Val vs Ala/Ala1.83 [1.0–3.3].041.96 [1.0–3.8].051.96 [1.1–3.6].03
 Ala/Val vs Ala/Ala1.19 [0.7–2.0].490.94 [0.5–1.6].831.13 [0.7–1.9].63
 Val/Val vs Ala/−1.64 [1.0–2.7].052.05 [1.2–3.6].011.81 [1.1–3.0].02

The SOD2Ala16Val Polymorphism

The homozygous variant genotype (Val/Val) was more common in cases than in controls (27% vs 18%; P = .04). The adjusted OR among or patients who had the Val/Val genotype for developing pancreatic cancer was 1.96 (95% CI, 1.0–3.8; P = .05) compared with patients w ho had the wild type genotype (Ala/Ala), whereas the heterozygous genotype (Ala/Val) was not associated with an increased risk (OR, 0.94; 95%CI, 0.5–1.6; P = .83) (Table 2).

Combined Risk of the MPOG463A and SOD2Ala16Val Polymorphism

Because both the MPO G/G and the SOD2 Val/Val genotypes were associated independently with elevated pancreatic cancer risk, we also evaluated the combined risk of these SNPs. Sixteen percent of cases (n = 19) had both high-risk genotypes compared with only 8% of controls (n = 26). When the analysis was adjusted for age, sex, smoking, and alcohol, the OR for developing pancreatic cancer was 4.29 (95% CI, 1.8–10; P = .0007) for those with both high-risk genotypes. Sixty-one percent of cases (n = 74) had 1 high-risk genotype compared with 53% of controls (n = 174), with an OR for developing pancreatic cancer of 1.96 (95% CI, 1.1–3.4; P = .01). The data are presented in Table 3, in which individuals with 1 high-risk genotype are separated into those with high-risk MPOG/G and low-risk SOD2Ala/− and those with low-risk MPOA/− and high-risk SOD2Val/Val. The results from our sensitivity analysis indicted that, when alcohol was excluded from the multivariate analyses, thereby increasing the control sample size, both high-risk polymorphisms retained statistical significance as risk factors for developing pancreatic cancer (Table 3).

Table 3. Crude and Adjusted Odds Ratios of Developing Pancreatic Cancer According to Combinations of High-risk Polymorphisms
MPOSOD2Total No. (%)CrudeAdjusted*Adjusted
CasesControlsOR [95% CI]POR [95% CI]POR [95% CI]P
  • MPO indicates myeloperoxidase; SOD2, superoxide dismutase; OR, odds ratio; 95% CI, 95% confidence interval; G, guanine; Val, valine; A, adenine; Ala, alanine.

  • *

    Adjusted for age, sex, and smoking and alcohol history (the smaller cohort was because of missing data regarding alcohol history for 112 patients in the control group).

  • Adjusted for age, sex, and smoking history.

G/GVal/Val19 (16)26 (8)3.17 [1.6–6.5].0024.31 [1.9–10].0013.33 [1.6–6.9].001
A/−Val/Val14 (11)33 (10)1.84 [0.9–3.9].112.28 [1.0–5.3].052.29 [1.1–4.9].034
G/GAla/−60 (49)141 (43)1.85 [1.1–3.1].021.89 [1.1–3.3].011.99 [1.2–3.3].009
A/−Ala/−29 (24)126 (39)1 [Reference] 1 [Reference] 1 [Reference] 

DISCUSSION

In this article, we report the first association to our knowledge between pancreatic adenocarcinoma risk and polymorphisms of the inflammatory pathway genes MPOG463A and SOD2Ala16Val. Our study demonstrated that patients with the wild-type MPOG/G variant had an increased risk of pancreatic cancer. Patients who were homozygous for the variant allele of SOD2Ala16Val also had an increased risk of pancreatic cancer compared with controls. The risk was additive when both polymorphisms were assessed together.

Oxidative stress is caused when the levels of superoxide radicals and ROS exceed regulatory antioxidant mechanisms.32 Malignant transformation may be accompanied by either reduced antioxidant activity or increased levels of ROS.33, 34 Reduced antioxidant gene expression, including that from SOD2, has been described in chronic pancreatitis and pancreatic cancer compared with the normal pancreas in controls.35

Our results are in line with several studies in nonsmall cell lung cancer, which demonstrated that patients who were homozygous for the variant MPO A/A genotype had a reduced risk of nonsmall cell lung cancer compared with controls.12–15 Furthermore, in preclinical models of pancreatic carcinoma cell lines, reduced SOD2 protein levels were correlated with increased rates of tumor cell proliferation.36 A small study of 24 cases and 23 controls that examined polymorphisms of the SOD2Ala16Val gene and pancreatic risk demonstrated no association.37 However, it has been demonstrated previously that Val-containing genotypes carry an increased risk of breast, lung, and bladder cancer.22–24 Duell et al. suggested a link between pancreatic cancer and the inflammatory gene polymorphisms tumor necrosis factor (TNF) and regulated upon activation in normal T cells, expressed, and secreted (RANTES) in a subgroup of patients with a history of pancreatitis but identified no overall association.38 Because pancreatitis may be a symptom of pancreatic cancer as well as a risk factor, we evaluated pancreatitis that first occurred >3 years before diagnosis. Less than 5% of patients had a remote history of pancreatitis, which made this subgroup too small to analyze separately.

Evaluation of other high-prevalence, low-penetrant genetic polymorphisms in pancreatic cancer has been sparse.38–42 Polymorphisms in the methylene tetrahydrofolate reductase (MTHFR) gene, which regulates folate metabolism, are associated with an increased risk of pancreatic cancer.40–42 Polymorphisms in the DNA repair genes O-6-methylguanine-DNA methyltransferase (MGMT), apurinic/apyrimidinic endonuclease class I (APE1), and x-ray repair cross-complementing protein 1 (XRCC1) have not demonstrated a significant interaction with smoking on the risk of developing pancreatic cancer.39

Recently, there have been several association studies linking sporadic pancreatic cancer risk with tooth loss and periodontitis, a condition that is characterized by chronic, low-level inflammation.43–45 This association remained significant even after accounting for cigarette smoking.44 During periodontitis, MPO and SOD2 are elevated in the gingival crevicular fluid.46, 47 The MPOG/G genotype also has a reported association with an increased risk of coronary artery disease.48 These associations raise the possibility that inflammatory processes may be important etiologically in pancreatic cancer, as reported in several other conditions.49–51

Our results are tempered because of modest sample sizes. In addition, alcohol data were missing in approximately 33% of controls, although sensitivity analyses consistently demonstrated statistical significance. We were not able to obtain diabetes data for most of the controls, although the association between diabetes and pancreatic cancer risk remains controversial. We had not collected data on periodontitis in either cases or controls, because this association had not been reported at the time of recruitment. These results will require validation in larger studies in which polymorphisms over the entire inflammatory pathway can be studied and correlated with serum inflammatory markers. In conclusion, our findings support the hypothesis that polymorphisms of MPOG463A and SOD2Ala16Val confer a reduced level of antioxidant activity may increase the risk of pancreatic carcinogenesis.

Acknowledgements

We thank the following: Rebecca Heist, Matthew Kulke, Penny Bradbury, Lawrence Blaszkowsky, Eunice Kwak, Peggy Suen, Andrea Shafer, Lucy Ann Principe, Salvatore Mucci, Martin Fruh, and Frances Shepherd.

Ancillary