An Overview of Genetic Polymorphisms and Pancreatic Cancer Risk in Molecular Epidemiologic Studies.

0.18 MB·English

by  Lin, Yingsong

J Epidemiol 2011;21(1):2-12
doi:10.2188/jea.JE20100090
Review Article
An Overview of Genetic Polymorphisms and Pancreatic Cancer
Risk in Molecular Epidemiologic Studies
Yingsong Lin1, Kiyoko Yagyu1, Naoto Egawa2, Makoto Ueno3, Mitsuru Mori4, Haruhisa Nakao5,
HiroshiIshii6,KozueNakamura7,KenjiWakai8,SatoyoHosono9,AkikoTamakoshi1,andShogoKikuchi1
1DepartmentofPublicHealth,AichiMedicalUniversitySchoolofMedicine,Nagakute,Japan
2DepartmentofInternalMedicine,TokyoMetropolitanKomagomeHospital,Tokyo,Japan
3HepatobiliaryandPancreaticMedicalOncologyDivision,KanagawaCancerCenterHospital,Kanagawa,Japan
4DepartmentofPublicHealth,SapporoMedicalUniversitySchoolofMedicine,Sapporo,Japan
5DivisionofGastroenterology,DepartmentofInternalMedicine,AichiMedicalUniversitySchoolofMedicine,Nagakute,Japan
6HepatobiliaryandPancreaticSection,GastroenterologicalDivision,CancerInstituteHospital,Tokyo,Japan
7DepartmentofEpidemiologyandPreventiveMedicine,GifuUniversityGraduateSchoolofMedicine,Gifu,Japan
8DepartmentofPreventiveMedicine,NagoyaUniversityGraduateSchoolofMedicine,Nagoya,Japan
9DivisionofEpidemiologyandPrevention,AichiCancerCenterResearchInstitute,Nagoya,Japan
ReceivedJune9,2010;acceptedSeptember14,2010;releasedonlineNovember6,2010
ABSTRACT
Background: Although pancreatic cancer has been extensively studied, few risk factors have been identified, and
novalidatedbiomarkersorscreeningtoolsexistforearlydetectioninasymptomaticindividuals.Wepresentabroad
overviewofmolecularepidemiologicstudiesthathaveaddressedtherelationshipbetweenpancreaticcancerriskand
genetic polymorphisms in several candidate genes and suggest avenues for future research.
Methods: A comprehensive literature search was performed using the PubMed database.
Results: Overall, individual polymorphisms did not seem to confer great susceptibility to pancreatic cancer;
however, interactions of polymorphisms in carcinogen-metabolizing genes, DNA repair genes, and folate-
metabolizing genes with smoking, diet, and obesity were shown in some studies. The major problem with these
studies is that, due to small sample sizes, they lack sufficient statistical power to explore gene–gene or
gene–environment interactions. Another important challenge is that the measurement of environmental influence
needs to be improved to better define gene–environment interaction. It is noteworthy that 2 recent genome-wide
association studies of pancreatic cancer have reported that variants in ABO blood type and in 3 other chromosomal
regions are associated with risk for this cancer, thus providing new insight into pancreatic cancer etiology.
Conclusions: As is the case in other complex diseases, common, low-risk variants in different genes may act
collectively to confer susceptibility topancreatic cancer inindividuals with repeated environmental exposures, such
as smoking and red meat intake. Clarification of gene–gene and gene–environmental interaction is therefore
indispensableforfuturestudies.Toaddresstheseissues,arigorouslydesignedmolecularepidemiologicstudywitha
large sample is desirable.
Key words: pancreatic cancer; genetic polymorphisms; molecular epidemiology; gene–environment interaction
The etiology of sporadic pancreatic cancer is not well
INTRODUCTION
understood. However, mounting evidence suggests that
Pancreatic cancer is a major cause of cancer mortality in pancreatic carcinogenesis involves a complex interaction
developed countries.1 In Japan, approximately 25000 between genetic mutations, epigenetic alterations, and
Japanese men and women died from pancreatic cancer in environmental risk factors.4 Among these environmental risk
2007, making it the fifth leading cause of cancer death.2 factors, epidemiologic studies have identified only cigarette
Pancreatic cancer is a rapidly fatal disease, with mortality smoking and type II diabetes as clear risk factors for
almost identical to incidence. The all-stage 5-year survival pancreatic cancer.5,6 An association of pancreatic cancer
rate is less than 10%, the lowest among all cancer sites.3 with dietary habits remains unclear because of wide
Address for correspondence. Shogo Kikuchi, MD, PhD, Department of Public Health, Aichi Medical University School of Medicine, 21 Karimata, Yazako,
Nagakute-cho,Aichi480-1195,Japan(e-mail:[email protected]).
Copyright©2010bytheJapanEpidemiologicalAssociation
2
Lin Y,et al. 3
variationindietaryhabitsacrosspopulationsandthedifficulty increased risk for current smokers as compared with non-
of accurate diet measurement.7 smokers.5 Tobacco smoke contains a variety of carcinogens,
Due to the completion of human genome sequencing and ofwhich4-(methylnitrosamino)-1-(3-pyridyl)-1-butanoneand
rapid progress in sequencing techniques, an increasing its metabolite 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanol
number of studies are exploring the associations between are the major carcinogens involved in pancreatic carcino-
polymorphismsincandidategenesandpancreaticcancerrisk. genesis.15Whiletheroleofpolycyclicaromatichydrocarbons
In a search of the PubMed database using the keywords exposure and metabolism in pancreatic cancer needs further
“genetic polymorphism” plus “pancreatic cancer,” we found investigation, aromatic amine and heterocyclic amine have
217 publications that had been published in the last 10 years. been implicated in the pathogenesis of pancreatic cancer.16
Furthermore, a small but growing body of research has Metabolic activation of carcinogens forms DNA adducts,
addressed gene–environment interaction contributing to causing mutations in crucial genes, including RAS, MYC,
pancreatic cancer development.8–10 Of special importance TP53,andP16.15Theaccumulationofthesegeneticmutations
are 2 recent genome-wide association studies (GWAS), leads to uncontrolled cell growth and tumor development.
which reported that a variant in ABO blood type and 3 Enzymes such as cytochrome P450S, glutathione-S-
other variants in chromosomal regions are associated with transferase (GST), N-acetyltransferase 1 (NAT1), and N-
pancreatic cancer risk.11,12 acetyltransferase (NAT2) are involved in the metabolic
Anestimated10%ofpancreaticcancercasesareassociated activation of carcinogens to DNA adducts and in detoxi-
with inherited predisposition, based on familial clustering.13 fication to other products.17 Human cytochrome P450 1A1
Several germline mutations have been linked to familial (CYP1A1) encodes aryl hydrocarbon hydroxylase, a phase I
pancreaticcancer.14The roleofgermline mutationsinseveral enzyme involved in the activation of tobacco-related
genes, such as INK4A, BRCA2, and LKB1, and their asso- carcinogens.17 The GSTs are a family of phase II iso-
ciationswithpancreaticcancerriskisbeyondthescopeofthis enzymes that are involved in phase II drug metabolism by
review. However, in this article we provide a broad overview conjugation of electrophilic substances with glutathione.18
of molecular epidemiologic studies that have investigated GST detoxifies a broad range of substances, including
the relationship between genetic polymorphisms in several carcinogens, environmental toxins, and drugs. Genetic
candidate genes and their interactions with environmental polymorphisms resulting in lack of enzyme activity due to
factors in conferring pancreatic cancer risk. We focus on homozygous deletion of the GSTM1 and GSTT1 genes have
polymorphisms in carcinogen-metabolizing genes, DNA been described.18 The frequencies of these deletions vary
repair genes, and folate-metabolizing genes because the across populations.19 Molecular epidemiologic studies have
functional importance of these genes has been elucidated shown increased risk for various cancers among individuals
and because most published studies have examined these with the NAT1 rapid acetylator or NAT2 slow acetylator
genes. Furthermore, on the basis of findings from GWAS genotypes, in the presence of known carcinogen exposure,
and biomarker, epidemiologic, and experimental studies, such as smoking or dietary exposure to heterocyclic amine.20
we identify additional genetic polymorphisms that require Although several molecular epidemiologic studies have
analysis due to their potentially important role in the etiology examined the associations between variants of the genes
of pancreatic cancer. encodingCYP,GST,andNATenzymesandpancreaticcancer
risk,mostoftheimportantfindingsregardingthemaineffects
of genetic variations or gene–environment interactions were
METHODS
reportedby2case–controlstudies:anongoinghospital-based
We performed a comprehensive literature search using case–control study at MD Anderson Cancer Center21 and a
the PubMed database. The keywords used were “genetic population-based case–control study in 6 areas of San
polymorphism” plus “pancreaticcancer.” Inaddition,we also Francisco Bay from 1994 to 20018 (Table 1). Both studies
cite the published literature addressing candidate gene poly- enrolled a relatively large number of case and control sub-
morphisms and their associations with other cancer types, as jects, which allowed for an analysis of gene–environment
well as findings from GWAS and experimental studies. interactions. The case–control study conducted at MD
Anderson Cancer Center revealed the following. (1) NAT1
rapid alleles were associated with 1.5-fold increased risk, and
RESULTS
the effect was more prominent among ever smokers and
Tobacco smoking and genetic polymorphisms in females.21 A significant synergistic effect of the CYP1A2*1F
carcinogen-metabolizing genes allele and NAT1 rapid alleles with respect to the risk for
Pancreatic cancer is a tobacco-induced cancer: epidemiologic pancreaticcancerwasalsodetected. (2)TherareNAT1*10or
studies have consistently shown that cigarette smoking NAT1*11-NAT2*6A diplotype may be an “at-risk” genetic
increases the risk for pancreatic cancer. A meta-analysis variant for pancreatic cancer.22 (3) The GSTP1*C variant
of 82 case–control and cohort studies reported a 1.8-fold conferred a possible protective effect against pancreatic
J Epidemiol 2011;21(1):2-12
4 Genetic Polymorphisms andPancreatic Cancer Risk
Table 1. Summaryoffindingsfromcase–controlstudiesofgeneticpolymorphismsincarcinogen-metabolizinggenesandtheir
interactionswithenvironmentalfactorsandpancreaticcancerrisk
Study No.of No.of Maineffects Gene–environment
Studyandyear Geneticpolymorphisms
population cases controls ofpolymorphisms interaction
Leeetal,1997 Koreans 45 53 P-450(1A1,2D6,and2E1) Noassociation Unreported
(Ref27)
BartschH,etal,1998 Whites 81 78 NAT1,NAT2,GSTM1, GSTM1andNAT1 Unreported
(Ref25) NAD(P)H:NQO1 enzymesassociatedwith
modestincreasein
susceptibilitytopancreatic
cancer
Liuetal,2000 Canadian 149 146 CYP1A1,GSTM1,GSTT1 Noassociation Notobserved
(Ref26)
DuellEJ,etal,2002 Whites 309 964 CYP1A1,GSTM1,GSTT1 Nosignificantmaineffects NeversmokerswithGSTT1-
(Ref8) presentgenotypevsheavy
smokerswithGSTT1-null
genotype:OR,3.2(95%CI,
1.3–8.1)formenand5.0
(1.8–14.5)forwomen
Lietal Non-Hispanic 365 379 P4501A1,NAT NAT1rapidalleles Interactionwithsmoking
(Ref21) whites associatedwith1.5-fold
increasedrisk
Jiaoetal,2007 Non-Hispanic 352 315 GSTM1,GSTT1,GSTP1 Nosignificantmaineffects GSTP1*Cvariantconferred
(Ref22) whites possibleprotectiveeffectin
oldersubjects
Jiaoetal,2007 Non-Hispanic 532 581 HaplotypeofNAT1andNAT2 RareNAT1*10orNAT*11- InteractionsbetweenNAT2
(Ref23) whites NAT2*6Adiplotypeasso- slowgenotypeandsmoking
ciatedwithincreasedrisk andhistoryofdiabetes
Suzukietal,2008 Non-Hispanic 755 636 P4501A2,SULT1A1,andNAT Nosignificantmaineffects Interactionsbetween
(Ref24) whites CYP1A2,NAT1,andheavy
smokinganddietary
mutagenintake
Abbreviations: NAT, N-acetyltransferase; GSTM, glutathione-S-transferase; CYP1A1, cytochrome P450 1A1; OR, odds ratio; CI, confidence
interval.
ResultsreportedbyLietal,Jiaoetal,andSuzukietalcamefromthesameresearchgroup.
cancer.23 (4) A significant interaction was noted between More studies are needed to examine genetic variations in
CYP1A2, NAT1, and heavy smoking and dietary mutagen tobacco-metabolizing genes and their associations with
intake.24 pancreatic cancer among different ethnic populations.
In the population-based case–control study carried out However, based on current, limited evidence from molecular
in 6 areas of San Francisco Bay, Duell et al examined epidemiologic studies, it seems unlikely that individual
polymorphisms in carcinogen-metabolizing genes, smoking, polymorphisms themselves confer major susceptibility to
and pancreatic cancer risk in whites and found that there was pancreatic cancer. Given the reported synergistic effects of
no significant increase in risk associated with any genotype smoking and certain carcinogen-metabolizing gene poly-
examined.8 However, the odds ratio (OR) was 5.0 (95% con- morphisms, it is essential to clarify gene–environment
fidence interval [CI], 1.8–14.5) for heavy smokers who had a interactions in future studies.
deletion polymorphism in GSTT1, suggesting that inherited
deletion polymorphisms in GSTT1 increase susceptibility to DNArepairandpolymorphisms inDNArepairgenes
smoking-related pancreatic cancer. Another notable finding DNA repair plays a crucial role in cellular defense against
from this study was that the interaction was stronger mutations caused by carcinogens and endogenous mech-
in women than in men. Although these results require anisms. Four types of DNA repair systems—nucleotide
replication in other studies, the findings suggest that women excision repair, base excision repair (BER), mismatch repair,
with a GSTT1-null or GSTM1-null genotype may be more and recombination repair—have been identified so far.28
susceptible than men to the effects of DNA adducts. In Defects in these pathways may result in a predisposition to
additiontothe2studiesmentionedabove,other,smallstudies cancer.EachtypeofDNArepairisunderstoodinconsiderable
haveaddressedthisissue,althoughthefindingsaredifficultto detail. We focus on BER because the genetic polymorphisms
interpret due to the small sample size.25–27 in this repair pathway are the most extensively studied in
J Epidemiol 2011;21(1):2-12
Lin Y,et al. 5
molecular epidemiologic studies. DNA bases are particularly Because hundreds of genetic polymorphisms may be
susceptible tooxidationmediatedbyreactive oxygenspecies, involved in maintaining genomic integrity, additional studies
whichcanbeproducedasaconsequenceofionizingradiation with large sample sizes are needed to elucidate multiple
or environmental exposure to transition metals, chemical sequencevariantsinageneormultiplegeneswithinanentire
oxidants,andfreeradicals.Reactiveoxygenspecieshavebeen pathway.
linkedtotheinitiationandprogressionofcancer.29BERplays
an important role in preventing mutations associated with a Folate intake and polymorphisms in folate-
commonproductofoxidativedamagetoDNA,8-oxoguanine. metabolizing genes
X-Ray Repair Cross-Complementing Group 1 (XRCC1), Folate is a water-soluble B vitamin abundant in green leafy
located on 19q13.2, is a polymorphic BER gene that has vegetables, citrus fruit, legumes, and cereals. Substantial
been the most extensively examined in molecular epidemi- evidencefromepidemiologicandlaboratoryresearchsupports
ologic studies of the risk of various cancers.30 In the a role for folate in carcinogenesis.36,37 Epidemiologic studies
above-mentioned population-based case–control study con- have consistently shown an inverse association between
ducted in 6 areas of San Francisco Bay,31 a synergistic effect folate intake and pancreatic cancer risk. Based on a meta-
between the XRCC1 399Gln allele and tobacco smoking in analysis in which data from 4 cohort studies and 1
relation to pancreatic cancer risk was observed, although case–control study were analyzed, individuals with the
no significant associations were noted between XRCC1 highest folate intake had a 51% lower risk than those with
genotypes and pancreatic cancer risk. As compared with thelowestfolateintake.38Mechanisticstudieshaveelucidated
never-active smokers and passive smokers with the Arg/Arg 2 major underlying mechanisms that may be involved.
genotype, the age- and race-adjusted ORs for heavy smokers Folate deficiency may induce misincorporation of uracil
(≥41pack-years)withtheGln/GlnorArg/Glngenotypeswere into DNA, leading to chromosomal breaks and mutations.
7.0(95%CI,2.4–21)inwomenand2.4(1.1–5.0)inmen.The In addition, folate deficiency may cause aberrant DNA
interaction suggests that XRCC1 Arg399Gln and BER methylation, resulting in altered expression of critical proto-
capacity are important in susceptibility to smoking-induced oncogenes and tumor suppressor genes.39 Moreover,
pancreatic cancer. However, these findings need to be con- functional polymorphisms in folate-metabolizing genes may
firmed in other studies, as the number of study subjects was confer susceptibility to cancer. Among the several poly-
small intheanalysis exploringgene–environmentinteraction. morphisms in the folate metabolic pathway, polymorphisms
The 8-oxoguanine DNA glycosylase (OGG1) gene is in the 5-10 methylenetetrahydrofolate reductase (MTHFR)
another BER gene that removes oxidative DNA lesions.28 gene are the most extensively studied. A central enzyme
OGG1hasbeenassociatedwithalteredriskofhumancancers. in folate metabolism, MTHFR irreversibly converts 5,10-
In data from the hospital case–control study conducted at methylenetetrahydrofolate to 5-methyltetrahydrofolate, the
MD Anderson Center in the United States, Li et al noted predominant form of folate in systemic circulation. Thus,
significantly reduced overall survival in patients with MTHFR acts as a critical junction in folate metabolism by
the OGG1 C315G (rs1052133) GG homozygous variant directing folate metabolites toward the DNA methylation
genotype.32 Furthermore, they reported a weak interaction pathway and away from the DNA synthesis pathway. Two
of the OGG1 C315G CC/CG genotype with diabetes in common functional polymorphisms of the MTHFR gene,
pancreatic cancer. These findings suggest that the CC/CG C677T and A1298C, have been identified.39 Regarding
genotype, combined with environmental exposure, confers C677T, the TT genotype (variant type) has been shown
increased susceptibility to pancreatic cancer. to have 35% lower enzyme activity than the CC genotype
Li et al also examined associations of pancreatic cancer (wild type).39 As for A1298C, homozygotes (CC) have
with selected DNA repair polymorphisms in other types approximately 60% of normal MTHFR activity.
of DNA repair pathways, including XRCC2, XRCC3,33 Since 2005, five studies have reported an association
RAD54L, and RecQ1 in the recombination repair pathway,34 between the MTHFR C677T genotype and pancreatic cancer
and the xeroderma pigmentosum group D (XPD) in the risk, but the results were not consistent9,40–43 (Table 2). The
NER pathway.35 They found that variant alleles of XRCC2 TT genotype was associated with significantly increased risk
R188H and XRCC3 A17893G were associated with for pancreatic cancer in 2 case–control studies carried out
significantly reduced survival in pancreatic cancer patients in the United States9 and China40: the ORs were 2.14 (95%
and that XRCC2 Arg188His polymorphisms may be genetic CI, 1.14–4.01) and 5.12 (2.94–9.10), respectively. Notably, a
modifiers for smoking-related pancreatic cancer. significant interaction between TT genotype and smoking in
Overall, evidence from a small number of molecular pancreatic cancer risk was also observed. In these 2 studies,
epidemiologic studies supports a role for genetic variability heavy smokers with the TT genotype had an approximately
in DNA repair in the risk for pancreatic cancer. Due to small 7-fold increased risk as compared with nonsmokers with
sample sizes and heterogeneous study designs, however, the CC genotype.9,40 In contrast, 2 Japanese case–control
the results are inconclusive and require confirmation. studies reported no increased risk associated with the TT
J Epidemiol 2011;21(1):2-12
6 Genetic Polymorphisms andPancreatic Cancer Risk
Table 2. Summary of findings from case–control studies of genetic polymorphisms in folate-metabolizing genes and their
interactionswithenvironmentalfactorsandpancreaticcancerrisk
Study No.of No.of Genetic
Studyandyear Maineffectsofpolymorphismsa Gene–environmentinteractiona
population cases controls polymorphisms
Lietal,2005 Non-Hispanic 347 348 MTHFRC677T, SignificanteffectforC677T: HeavysmokerswithTTvs
(Ref9) USwhites A1298C CT,0.90(0.63–1.27),TT,2.14 neversmokerswithCC/CT:6.83
(1.14–4.01);noassociationfor (1.91–24.38)
A1298C HeavyalcoholdrinkerswithTT
vsnondrinkerswithCC/CT:4.23
(0.88–20.3)
Wangetal,2005 Chinese 163 337 MTHFRC677T, SignificanteffectforC677T:CT, HeavysmokerswithTTvs
(Ref40) A1298C,TS 2.60(1.61–4.29),TT,5.12 neversmokerswithCC/CT:6.69
(2.94–9.10);noassociationfor (3.39–13.63)
A1298C AlcoholdrinkerswithTT/CTvs
nondrinkerswithCC:4.39
(2.25–8.78)
Matsubayashietal, Americans 303 305 MTHFRC677T, NoassociationforC677T:CT, Nosignificantinteractionwith
2005(Ref41) A1298C 0.79(0.56–1.11),TT,1.10 smoking
(0.67–1.82)
Suzukietal,2008 Japanese 157 785 MTHFRC677T,MTR NoassociationforC677T:CT, Nosignificantinteractionwith
(Ref42) A2756G,TSvariable 0.98(0.65–1.47),TT,0.75 alcoholdrinking
numberoftandem (0.41–1.35)
repeat
Ohnamietal,2008 Japanese 198 182 MTHFRC677T,MTRR NoassociationforC677T,but Noassociation
(Ref43) (rs1801394,rs162049, MTRRpolymorphisms
rs10380) associatedwithincreasedrisk
Abbreviations:MTHFR,methylenetetrahydrofolatereductase;MTR,methioninesynthase;MTRR,methioninesynthasereductase;OR,oddsratio.
aValuesareoddsratios(95%confidenceinterval).
Table 3. Associationsofgeneticpolymorphismsandpancreaticcancerriskthatrequireassessmentinfuturestudies
Potentialinteractionswith
Candidategenes Selectedpolymorphisms Circulatingbiomarker
environmentalfactors
VitaminDsignaling rs11574143 Sunexposure,diet Plasma25-hydroxyvitaminD
Melatoninreceptorsandclockgenes MTNR1B,rs10830963,rs11133373inCLOCK Diabetes Plasmaorurinarymelatonin
(6-sulfatoxymelatonin)
Insulin,IGFgene IGF1haplotypeandtheIGF2Ex4-233C>TTT Diabetes,obesity PlasmaorserumIGF
genotype
TGF-βsignaling TGFBR1*6A Diabetes PlasmaorserumTGF-β
Infection-relatedgenepolymorphisms COX-2polymorphisms N/A N/A
ABOgene rs505922 N/A N/A
Genesinchromosome13q22.1 Novelpolymorphismstobeidentified N/A N/A
Abbreviations:IGF,insulin-likegrowthfactor;TGF-β,transforminggrowthfactor-β;MTNR1B,melatoninreceptor1B;COX-2,cyclooxygenase-2.
N/A:notapplicable.
genotype.41,42 In one of these studies, the OR for pancreatic one study suggested a 1.8-fold increased risk in subjects with
cancer in individuals with the TT genotype was 0.75 the CC genotype.38
(0.41–1.35), but the association was statistically insig- Given the strong interaction of the MTHFR genotype
nificant.42 Due to the small number of studies and the wide with environmental factors such as smoking and alcohol
heterogeneity in results, a summary OR could not be consumption, it is important to unravel their complex
calculated in a meta-analysis of the MTHFR C677T relationships in additional large, adequately powered studies.
genotype and pancreatic cancer risk, as only 3 published Furthermore, because the balance between the use of
studies were included.38 Inadequate sample size and different methylenetetrahydrofolate for DNA synthesis rather than for
criteria for control selection might have contributed to the methionine synthesis might depend on the presence of the
inconsistent results reported so far. As for the A1298C 677T variant of MTHFR and nutritional folate status, studies
genotype, evidence is insufficient to draw a conclusion: 2 targeting populations with folate deficiency in developing
studiesshowednoimportanteffectsonpancreaticcancer,and countries may provide valuable information.
J Epidemiol 2011;21(1):2-12
Lin Y,et al. 7
Alcohol consumption and polymorphisms in lower risk of pancreatic cancer. However, the role of vitamin
alcohol-metabolizing enzymes D in the development of pancreatic cancer remains unclear
Although the majority of prospective cohort studies found due to the small number of studies.
no significant increase in the risk of pancreatic cancer High vitamin D exposure is hypothesized to decrease
with moderate to high levels of alcohol intake in a general cancer risk, possibly through genomic effects modulated by
population,7 some evidence suggests that excessive drinking thevitaminDreceptor,andbyautocrine/paracrinemetabolism
may increase risk in population subsets.44 ofthevitaminDreceptor’sligand,1α,25-(OH)2-vitaminD3.49
Ethanol is mainly metabolized to acetaldehyde by alcohol Recently, an increasing number of studies have examined
dehydrogenase enzymes and further oxidized to acetate by polymorphisms in vitamin D receptor and selected genes in
acetaldehyde dehydrogenase. Acetaldehyde has been shown the vitamin D pathway in relation to colorectal, breast, and
tohavecarcinogeniceffectsinexperimentalstudiesandisthe prostate cancer risk.48 However, there is currently no strong,
mainmechanismtoexplainalcohol-inducedcarcinogenesis.45 consistent epidemiologic evidence for a substantial influence
Variations in the production and/or oxidation of acetaldehyde of any single variant in vitamin D pathway genes on cancer
among individuals are caused by single-nucleotide poly- risk. The association of pancreatic cancer with serum vitamin
morphisms (SNPs) of ADH1B, ADH1CI, and ALDH2.45 In 25(OH)D levels and polymorphic variants in genes encoding
particular, people homozygous for ALDH2*2 display flush for enzymes that synthesize, carry, and degrade vitamin D is
syndrome, which is characterized by nausea, vomiting, and an important research subject for future studies.
facial flushing after ingestion of a small amount of alcohol.
Veryfewstudieshaveaddressedtheroleofpolymorphisms Circadian disruption, melatonin, and genetic
in alcohol-metabolizing enzymes and pancreatic cancer risk. variations in clock genes
A recent case–control study involving 160 pancreatic cancer To date, no studies have examined the association between
patients and 800 age- and sex-matched controls in Japan circadian disruption and pancreatic cancer risk in humans.
foundthatalcoholconsumptionwasassociatedwithincreased Experimental data, however, have shown that disruption of
risk in individuals with the ALDH2 Lys+ allele or ADH1B circadian rhythms in mice is associated with accelerated
His/His or ADH1C Arg/Arg genotypes, but not in those with growth of pancreatic cancer.52 Although the findings are not
theALDH2Glu/GlugenotypeorADH1BArgorADH1CGln entirely consistent, epidemiologic studies have indicated that
alleles.46 This suggests that the risk of pancreatic cancer is shift work is significantly associated with increased risks
associated with the combined effect of alcohol consumption of breast, colorectal, and prostate cancer.53 On the basis of
and certain polymorphisms in alcohol-metabolizing enzymes. considerable evidence from animal studies and limited
Because the metabolism of alcohol and acetaldehyde is evidence from epidemiologic studies, the working group of
strongly influenced by alcohol-metabolizing enzymes, future IARC concluded in 2007 that “shift-work that involves
molecular epidemiologic studies need to examine the effect circadian disruption is probably carcinogenic to humans.”54
of these polymorphisms on pancreatic cancer risk while The principal mechanism involves melatonin, a neuro-
accounting for alcohol consumption. hormone that regulates the circadian rhythm.55 Three recent
GWAS have shown that the common variant MTNR1B
Vitamin D and polymorphisms in vitamin D pathway (melatonin receptor 1B) is associated with insulin and
genes glucose concentrations.56–58 The melatonin receptor MT1 is
Humans get vitamin D mainly from exposure to sunlight highly expressed in pancreatic islet cells, and the expression
or their diet. Vitamin D is metabolized in the liver to of MTNR1B has been confirmed in both islets and sorted
25-hydroxyvitamin D [25(OH)D], which is further beta cells.59 Given the close relationship of hyperinsulinemia
metabolized in the kidneys by the enzyme 25-hydro- and diabetes with pancreatic cancer risk, it might prove
xivitamin D-1α-hydroxylase (CYP27B1) to its active form, interesting to examine the risk genotypes of MTNR1B and
1,25-dihydroxyvitamin D.47 Vitamin D receptor, a crucial their interactions with diabetes and other environmental
mediator of the cellular effects of vitamin D, is present in factors in pancreatic cancer development.
a variety of cell types, including pancreatic beta cells.48 In addition to melatonin, circadian rhythms are
Experimental evidence shows that vitamin D receptor controlled and maintained by several circadian genes via
interacts with other cell-signaling pathways to influence transcription–translation feedback loops that include positive
cancer development.49 Ecologic studies have linked sun activators, such as Clock, neuronal PAS domain protein 2
exposure to lower pancreatic cancer mortality.50 Individuals (NPAS2), cryptochrome 1 (CRY1) and CRY2, and period 1
with higher circulating 25(OH)D levels have been found (PER1), PER2, and PER3.60 To test the hypothesis that
to have decreased risks of breast, colorectal, and prostate genetic variations in these genes may confer susceptibility
cancerinnumerousprospectivestudies.51Giventhecollective to prostate cancer, Zhu et al genotyped a total of 41 tagging
evidence from epidemiologic and experimental studies, it is and amino acid-altering SNPs in 10 circadian genes in a
plausible that high vitaminDlevelsmay be associated with a population-based case–control study of white men and found
J Epidemiol 2011;21(1):2-12
8 Genetic Polymorphisms andPancreatic Cancer Risk
that NPAS2 showed the most robust association with prostate TGFBR1*6A is emerging as a high frequency, low-
cancer risk.61 No studies, however, have examined genetic penetrance tumor susceptibility allele that confers sus-
polymorphisms in clock genes and pancreatic cancer risk. ceptibility to breast, ovarian, and colorectal cancer. A meta-
Because of the important role of melatonin and circadian analysisof7case–controlstudiesofTGFBR1*6Aandvarious
genes in maintaining circadian rhythm, future studies may cancer types combinedshowedthat TGFBR1*6A carriers had
address genetic variations in these genes and the risk of a26%increasedrisk.72TheroleofTGFBR1*6Ainpancreatic
pancreatic cancer. cancer remains unclear and is a subject of future study.
Insulin and insulin-like growth factor gene poly- Inflammation and infection-related gene poly-
morphisms morphisms
Obesity and type II diabetes are well established risk factors Inflammation has been implicated in pancreatic carcino-
for pancreatic cancer, especially in developed countries. genesis. Cyclooxygenase-2 (COX2) is a key enzyme in-
Elevated levels of insulin and insulin-like growth factors volvedinbiologicprocessesincludinginflammation,immune
(IGFs), such as IGF-I, are important mechanisms underlying function, and cell proliferation.73 The overexpression of this
the association between obesity, diabetes, and pancreatic enzyme has been shown in pancreatic cancer.74 However,
cancer.62 Insulin, IGF-1, and the insulin receptor-related there have been few studies addressing inflammation-related
receptor can form functional hybrids.63 IGF1 and IGF1 genetic polymorphisms and pancreatic cancer risk. Zhao et al
receptors are highly expressed in pancreatic cancer cells, showed that functional COX-2 polymorphisms are associated
and IGF2 imprinting is disrupted in many tumors.64 with susceptibility to pancreatic cancer.75 In another
Despite strong experimental evidence indicating that case–control study, 3 infection-related polymorphisms (TNF-
IGFs play an important role in carcinogenesis—including A, RANTE5, and CCR5) were examined, but no significant
the regulation of cell proliferation, differentiation, and effects were found.76
apoptosis—the results of epidemiologic studies examining Helicobacterpyloriinfectioninduceschronicinflammation
IGFs in relation to cancer risk are less persuasive. Using data and has been established as a risk factor for gastric cancer.
from a nested case–control study in the Japan Collaborative It remains controversial, however, whether H. pylori
Cohort (JACC) Study, we found a positive association infection plays a role in pancreatic cancer development.77
between baseline IGF-1 levels and the risk of pancreatic Epidemiologic studies examining this issue have produced
cancer mortality in apparently healthy Japanese.65 However, mixed results. The hypothesis has been proposed that
no significant associations were observed in other studies.66 polymorphisms in genes involved in inflammatory response,
Only 1 study has examined the association between genetic suchasIL-1A,IL-1B,IL-6,IL-8,mayhelpexplainwhyonlya
polymorphisms in IGF genes and pancreatic cancer risk.67 subset of individuals infected with H. pylori develops gastric
Of 6 SNPs of IGF1 and IGF2 that were examined in a cancer. Similarly, it is important to comprehensively analyze
case–control study by Suzuki et al, the IGF1 haplotype and the effects of these polymorphisms on pancreatic cancer risk.
the IGF2 Ex4 -233 C>T TT genotype were significantly
associated with decreased risk of pancreatic cancer, which SNPs indentified by genome-wide association
suggests that polymorphic variants of the IGF genes may studies
serve as a susceptibility factor for pancreatic cancer. Future Genome-wide association studies (GWAS) have been proven
studies are warranted to explore polymorphisms in IGF gene to be a valuable tool for identifying common alleles that
pathways and their interaction with obesity and physical influence disease risk.78 Fast-evolvingsequencingtechnology
activity in pancreatic cancer risk. allows researchers toscan across thegenomein a large set of
cases and controls to identify new associations that link
Transforming growth factor-β (TGF-β) and poly- certain regions to disease risk.
morphisms in the TGF-β pathway The first GWAS for pancreatic cancer, published in Nature
TGF-β regulates tumor initiation, progression, and metastasis GeneticsinAugust2009,identifiedcommonriskvariantsthat
via its signaling pathway involving membrane receptors and map to the first intron of the ABO gene on chromosome
SMAD transcription factors.68 The dual role of TGF-β in 9q34.2 (SNP rs505922).11 This finding implies that people
cancer, both as a tumor suppressor and tumor promoter, has with blood group O may have a lower risk than those with
been well defined.69 Several lines of evidence demonstrate groups A or B.
that pancreatic cancer is clearly linked to TGF-β.70 In Base on DNA collected from nearly 4000 patients in 13
particular, SMAD4, a component of the TGF-β pathway, is different studies, a second GWAS for pancreatic cancer has,
mutated in approximately 50% of pancreatic cancers.71 for the first time, identified pancreatic cancer susceptibility
Because of the presence of plausible mechanisms, a study locion 3chromosomes—13q22.1, 1q32.1,and 5p15.33.12Of
of the associations of polymorphisms in TGF-β pathway the 3 regions, the locus on 13q22.1 appears to be specific for
with pancreatic cancer risk should prove interesting. pancreatic cancer.
J Epidemiol 2011;21(1):2-12
Lin Y,et al. 9
Findings from the 2 GWAS of pancreatic cancer have susceptibilityvariantswillbediscoveredfromGWAS,andthe
provided new insight into pancreatic cancer etiology; interactions of these variants with environmental factors will
however, they need to be replicated in other large studies. be more frequently confirmed in molecular epidemiologic
Furthermore, follow-up studies after the GWAS must address studies. Since we have yet to discover rare variants that
whether SNPs discovered by GWAS represent functional greatlyincreasetheriskofpancreaticcancer,perhaps,asisthe
variants or simply tag true variants located in the same case with other complex diseases, common low-risk variants
haplotype. in different genes act collectively to confer susceptibility
to pancreatic cancer in individuals who have repeated
Mitochondrial genetic polymorphisms environmental exposures, such as smoking and intake of red
Mitochondria play an important role in cellular energy meta- meat.Arecentstudyprovidedcriticalevidencetosupportthis
bolism, free radical generation, and apoptosis. Mitochondrial notion by demonstrating that pancreatic cancer results from
DNA mutations can initiate a cascade of events leading to genetic alterations of a large number of genes that function
persistent oxidative stress, a condition that probably favors through 12 pathways and processes,79 including TGF-β
tumor development.80 Although previous studies have signaling and DNA damage control, which were discussed
examined the association between mitochondrial genetic in this review.
polymorphisms and pancreatic cancer,81–83 the results were What is the future direction for research on the etiology
inconsistent. While a mitochondrial SNP in the 16519 of pancreatic cancer? First, we believe that unraveling the
mitochondrial DNA nucleotide was found to be associated functional SNP variants in a number of identified gene
with worse prognosis,81 this positive association was not pathways, combined with novel variants identified in GWAS,
replicated in a recent study involving 990 pancreatic cancer is essential in deepening our understanding of pancreatic
patients.82 A recent large case–control study comprised 955 cancer risk. To achieve this goal, the complex gene–gene
participants with primary pancreatic adenocarcinoma and and gene–environmental interactions must be clarified in a
1102 control subjects and examined 24 mitochondrial SNPs rigorously designed molecular epidemiologic study with a
and11commonhaplogroups,noneofwhichwassignificantly large sample size. Second, in addition to SNPs, there is
associated with pancreatic cancer risk.83 Their results did not increasing recognition of the role of genetic variations—such
supportthesignificant involvementofmitochondrialSNPs or asDNAcopynumbervariationsandvariable-numbertandem
haplogroupsinthedevelopmentofpancreaticcancer.Because repeats—in cancer predisposition.84 High-resolution SNP
oftheimportantroleofmitochondrialDNAincancer,further arrays have made it possible to identify copy number
investigations of mitochondrial genetic variations are variations. Moreover, there have been studies linking copy
necessary to provide insight into the etiology of pancreatic number variations and variable-number tandem repeats to
cancer. pancreaticcancerrisk.85,86Elucidatingtheseassociationsisan
important goal for future research.
DISCUSSION
ACKNOWLEDGMENTS
Molecular epidemiologic studies examining the associations
between polymorphisms in several gene pathways and This work was supported by Grant-in-Aid for Cancer
pancreatic cancer risk have produced mixed results. Overall, Research (21Shi-11-1) from the Ministry of Health, Labour
individual polymorphisms did not seem to confer marked and Welfare, Japan.
susceptibility; however, some studies implicated interactions Conflicts of interest: None declared.
of polymorphisms in carcinogen-metabolizing genes, DNA
repair genes, and folate-metabolizing genes with smoking,
REFERENCES
diet, and obesity. The principal weakness of these studies is
small sample size; thus, it is difficult to detect statistically 1. ParkinDM, BrayF, FerlayJ,Pisani P. Globalcancer statistics,
significant gene–gene or gene–environment interactions. 2002.CA Cancer JClin. 2005;55:74–108.
Because of this, no functional variants reported so far have 2. FoundationforPromotionofCancerResearch.CancerStatistics
inJapan-2008.
been used to predict pancreatic cancer risk in the clinical
3. Matsuno S, Egawa S, Fukuyama S, Motoi F, Sunamura M,
setting. Another critical challenge is that the measurement of
Isaji S, et al. Pancreatic cancer registry in Japan: 20 years of
environmental influence in epidemiologic studies must be
experience. Pancreas.2004;28:219–30.
improved to better define gene–environment interaction.
4. Bardeesy N, DePinho RA. Pancreatic cancer biology and
Two recent GWAS of pancreatic cancer have provided genetics.Nat RevCancer.2002;2:897–909.
intriguing results that need to be confirmed in additional
5. Iodice S, Gandini S, Maisonneuve P, Lowenfels AB. Tobacco
studies. With the decreasing cost of genotype sequencing, we and the risk of pancreatic cancer: a review and meta-analysis.
expect that future GWAS will unravel causal variants with LangenbecksArch Surg. 2008;393:535–45.
significant effects on pancreatic cancer. Hopefully, disease 6. Huxley R, Ansary-Moghaddam A, Berrington de González A,
J Epidemiol 2011;21(1):2-12
10 Genetic Polymorphisms andPancreatic Cancer Risk
BarziF,WoodwardM.Type-IIdiabetesandpancreaticcancer:a 2007;16:2379–86.
meta-analysis of 36studies. BrJ Cancer.2005;92:2076–83. 23. Jiao L, Bondy ML, Hassan MM, Chang DZ, Abbruzzese JL,
7. MichaudDS.Epidemiologyofpancreaticcancer.MinervaChir. EvansDB,etal.GlutathioneS-transferasegenepolymorphisms
2004;59:99–111. and risk and survival of pancreatic cancer. Cancer.
8. DuellEJ,HollyEA,BracciPM,LiuM,WienckeJK,KelseyKT. 2007;109:840–8.
A population-based case–control study of polymorphisms in 24. Suzuki H, Morris JS, Li Y, Doll MA, Hein DW, Liu J, et al.
carcinogen-metabolizing genes, smoking, and pancreatic Interaction of the cytochrome P4501A2, SULT1A1 and NAT
adenocarcinoma risk. J Natl Cancer Inst.2002;94:297–306. genepolymorphismwithsmokinganddietarymutagenintakein
9. LiD,AhmedM,LiY,JiaoL,ChouTH,WolffRA,etal.5,10- modification of the risk of pancreatic cancer. Carcinogenesis.
Methylenetetrahydrofolatereductasepolymorphismsandtherisk 2008;29:1184–91.
of pancreatic cancer. Cancer Epidemiol Biomarkers Prev. 25. Bartsch H, Nair U, Risch A, Rojas M, Wikman H, Alexandrov
2005;14:1470–6. K. Genetic polymorphisms of CYP genes, alone or in
10. McWilliams RR, Bamlet WR, Cunningham JM, Goode EL, combination, as a risk modifier of tobacco-related cancers.
de Andrade M, Boardman LA, et al. Polymorphisms in DNA Cancer Epidemiol Biomarkers Prev. 2000;9:3–28.
repair genes, smoking, and pancreatic adenocarcinoma risk. 26. Liu G, Ghadirian P, Vesprini D, Hamel N, Paradis AJ, Lal G,
Cancer Res. 2008;68:4928–35. etal.PolymorphismsinGSTM1,GSTT1,andCYP1A1andrisk
11. Amundadottir L, Kraft P, Stolzenberg-Solomon RZ, Fuchs CS, of pancreatic adenocarcinoma. Br JCancer. 2000;82:1646–9.
Petersen GM, Arslan AA, et al. Genome-wide association 27. Lee HC, Yoon YB, Kim CY. Association between genetic
study identifies variants in the ABO locus associated with polymorphismsofthecytochromesP-450(1A1,2D6,and2E1)
susceptibility to pancreatic cancer. NatGenet.2009;41:986–90. andthesusceptibilitytopancreaticcancer.KoreanJInternMed.
12. PetersenGM,AmundadottirL,FuchsCS,KraftP,Stolzenberg- 1997;12:128–36.
Solomon RZ, Jacobs KB, et al. A genome-wide association 28. David SS, O’Shea VL, Kundu S. Base-excision repair of
study identifies pancreatic cancer susceptibility loci on chro- oxidative DNAdamage. Nature. 2007;447:941–50.
mosomes13q22.1,1q32.1,5p15.33.NatGenet.2010;42:224–8. 29. Klaunig JE, Kamendulis LM. The role of oxidative stress in
13. Schenk M, Schwartz AG, O’Neal E, Kinnard M, Greenson JK, carcinogenesis.AnnuRevPharmacolToxicol.2004;44:239–67.
FryzekJP,etal.Familialriskofpancreaticcancer.JNatlCancer 30. HungRJ,HallJ,BrennanP,BoffettaP.Geneticpolymorphisms
Inst.2001;93:640–4. in the base excision repair pathway and cancer risk: a HuGE
14. Hezel AF, Kimmelman AC, Stanger BZ, Bardeesy N, Depinho review. AmJ Epidemiol.2005;162:925–42.
RA.Geneticsandbiologyofpancreaticductaladenocarcinoma. 31. Duell EJ, Holly EA, Bracci PM, Wiencke JK, Kelsey KT. A
Genes Dev.2006;20:1218–49. population-based study of Arg399Gln polymorphism in X-Ray
15. Hecht SS. Tobacco carcinngens, their biomarkers and tobacco- Repair Cross-Complementing Group1 (XRCC1) and risk of
related cancer. Nat RevCancer.2003;3:733–44. pancreatic adenocarcinoma. CancerRes. 2002;62:4630–6.
16. Thompson PA, Seyedi F, Lang NP, MacLeod SL, Wogan GN, 32. Li D, Li Y, Jiao L, Chang DZ, Beinart G, Wolff RA, et al.
Anderson KE, et al. Comparison of DNA adduct levels Effects of base excision repair gene polymorphisms on
associated with exogenous and endogenous exposures in pancreatic cancer survival. Int JCancer. 2007;120:1748–54.
human pancreas in relation to metabolic genotype. Mutat Res. 33. Jiao L, Hassan MM, Bondy ML, Wolff RA, Evans DB,
1999;424:263–74. Abbruzzese JL, et al. XRCC2 and XRCC3 gene poly-
17. Nebert DW,DaltonTP.Therole of cytochrome P450enzymes morphisms and risk of pancreatic cancer. Am J Gastroenterol.
in endogenous signaling pathways and environmental 2008;103:360–7.
carcinogenesis. Nat RevCancer.2006;6:947–60. 34. LiD,FrazierM,EvansDB,HessKR,CraneCH,JiaoL,etal.
18. Rebbeck TR. Molecular epidemiology of the human gluta- Single nucleotide polymorphisms of RecQ1, RAD54L, and
thione S-transferase genotypes GSTM1 and GSTT1 in ATM genes are associated with reduced survival of pancreatic
cancer susceptibility. Cancer Epidemiol Biomarkers Prev. cancer. JClin Oncol. 2006;24:1720–8.
1997;6:733–43. 35. JiaoL,HassanMM,BondyML,AbbruzzeseJL,EvansDB,Li
19. Geisler SA, Olshan AF. GSTM1, GSTT1, and the risk of D. The XPD Asp312Asn and Lys751Gln polymorphisms,
squamous cell carcinoma Lin 30 of the head and neck: a mini- corresponding haplotype, and pancreatic cancer risk. Cancer
HuGE review. AmJ Epidemiol.2001;154:95–105. Lett.2007;245:61–8.
20. HeinDW,DollMA,FretlandAJ,LeffMA,WebbSJ,XiaoGH, 36. Kim YI. Folate and carcinogenesis: evidence, mechanisms, and
et al. Molecular genetics and epidemiology of the NAT1 implications.J NutrBiochem. 1999;10:66–88.
and NAT2 acetylation polymorphisms. Cancer Epidemiol 37. Larsson SC, Håkansson N, Giovannucci E, Wolk A. Folate
Biomarkers Prev.2000;9:29–42. intake and pancreatic cancer incidence: a prospective study of
21. Li D, Jiao L, Li Y, Doll MA, Hein DW, Bondy ML, et al. Swedishwomen andmen.J Natl Cancer Inst.2006;98:407–13.
PolymorphismsofcytochomeP4501A2andN-acetyltransferase 38. Larsson SC, Giovannucci E, Wolk A. Folate intake, MTHFR
genes, smoking and risk of pancreatic cancer. Carcinogenesis. polymorphisms, and risk of esophageal, gastric, and pancreatic
2006;27:103–11. cancer:ameta-analysis. Gastroenterology. 2006;131:1271–83.
22. Jiao L, Doll MA, Hein DW, Bondy ML, Hassan MM, Hixson 39. LucockM,YatesZ.Folicacid—vitaminandpanaceaorgenetic
JE, et al. Haplotype of N-Acetyltransferase 1 and 2 and risk timebomb?Nat RevGenet.2005;6:235–40.
of pancreatic cancer. Cancer Epidemiol Biomarkers Prev. 40. WangL,MiaoX,TanW,LuX,ZhaoP,ZhaoX,etal.Genetic
J Epidemiol 2011;21(1):2-12
Lin Y,et al. 11
polymorphisms in methylenetetrahydrofolate reductase and 58. ProkopenkoI,LangenbergC,FlorezJC,SaxenaR,SoranzoN,
thymidylate synthase and risk of pancreatic cancer. Clin Thorleifsson G, et al. Variants in MTNR1B influence fasting
Gastroenterol Hepatol. 2005;3:743–51. glucose levels.NatGenet. 2009;41:77–81.
41. Matsubayashi H, Skinner HG, Iacobuzio-Donahue C, Abe T, 59. MühlbauerE,PeschkeE.Evidencefortheexpressionofboththe
Sato N, Riall TS, et al. Pancreaticobiliary cancer with defi- MT-1 and in addition, the MT-2-melatonin receptor, in the rat
cient methylenetetrahydrofolate reductase genotypes. Clin pancreas, isletandbeta-cell. JPineal Res. 2007;42:105–6.
Gastroenterol Hepatol. 2005;3:752–60. 60. Fu L, Lee CC. The circadian clock: pacemaker and tumour
42. SuzukiT,MatsuoK,SawakiA,MizunoN,HirakiA,KawaseT, suppressor.Nat RevCancer.2003;3:350–61.
etal.Alcoholdrinkingandone-carbonmetabolism-relatedgene 61. Zhu Y, Stevens RG, Hoffman AE, Fitzgerald LM, Kwon EM,
polymporphisms on pancreatic cancer risk. Cancer Epidemiol Ostrander EA, et al. Testing the circadian gene hypothesis in
Biomarkers Prev.2008;17:2742–7. prostate cancer: a population-based case–control study. Cancer
43. OhnamiS,SatoY,YoshimuraK,OhnamiS,SakamotoH,Aoki Res. 2009;69:9315–22.
K, et al. His595Tyr polymorphisms in the methionine synthase 62. McCarty MF. Insulin secretion as a determinant of pancreatic
reductase(MTRR)geneisassociatedwithpancreaticcancerrisk. cancerrisk.MedHypotheses. 2001;57:146–50.
Gastroenterology. 2008;135:477–88. 63. Saltiel AR, Kahn CR. Insulin signaling and the regulation of
44. SilvermanDT,BrownLM,HooverRN,SchiffmanM,Lillemoe glucose andlipidmetabolism. Nature. 2001;414:799–806.
KD, Schoenberg JB, et al. Alcohol and pancreatic cancer 64. Bergmann U, Funatomi H, Yokoyama M, Beger HG, Korc M.
in black and whites in the United States. Cancer Res. Insulin-likegrowthfactorIoverexpressioninhumanpancreatic
1995;55:4899–905. cancer: evidence for autocrine and paracrine roles. Cancer Res.
45. SeitzHK,StickelF.Molecularmechanismsofalcohol-mediated 1995;55:2007–11.
carcinogenesis. Nat RevCancer.2007;7:599–612. 65. LinY,TamakoshiA,KikuchiS,YagyuK,ObataY,IshibashiT,
46. Kanda J, Matsuo K, Suzuki T, Kawase T, Hiraki A, Watanabe et al. Serum insulin-like growth factor-I, insulin-like growth
M, et al. Impact of alcohol consumption with polymorphisms factorbindingprotein-3,andtheriskofpancreaticcancerdeath.
in alcohol-metabolizing enzymes on pancreatic cancer risk in IntJ Cancer.2004;110:584–8.
Japanese. Cancer Sci. 2008;100:296–302. 66. WolpinBM,MichaudDS,GiovannucciEL,SchernhammerES,
47. Holick MF. Vitamin D deficiency. N Engl J Med. Stampfer MJ, Manson JE, et al. Circulating insulin-like growth
2007;357:266–81. factorsaxisandtheriskofpancreaticcancerinfourprospective
48. Raimondi S, Johansson H, Maisonneuve P, Gandini S. Review studies. BrJ Cancer.2007;97:98–104.
and meta-analysis on vitamin D receptor polymorphisms and 67. Suzuki H, Li Y, Dong X, Hassan MM, Abbruzzese JL, Li D.
cancer risk.Carcinogenesis. 2009;30:1170–80. Effectofinsulin-likegrowthfactorgenepolymorphismsaloneor
49. Campbell FC, Xu H, El-Tanani M, Crowe P, Bingham V. The in interaction with diabetes on the risk of pancreatic cancer.
Yin and Yang of vitamin D receptor (VDR) signaling in Cancer Epidemiol Biomarkers Prev. 2008;17:3467–73.
neoplastic progression: operational networks and tissue-specific 68. Blobe GC, Schiemann WP, Lodish HF. Role of transforming
growth control.BiochemPharmacol. 2010;79:1–9. growth factor β in human disease. N Engl J Med.
50. Stolzenberg-SolomonRZ.VitaminDandpancreaticcancer.Ann 2000;342:1350–8.
Epidemiol. 2009;19:89–95. 69. MassaguéJ.TGFβin cancer. Cell. 2008;134:215–30.
51. Garland CF, Gorham ED, Mohr SB, Garland FC. Vitamin D 70. Truty MJ, Urrutia R. Basics of TGF-β and pancreatic cancer.
for cancer prevention: globe perspective. Ann Epidemiol. Pancreatology. 2007;7:423–35.
2009;19:468–83. 71. Tascilar M, Skinner HG, Rosty C, Sohn T, Wilentz RE,
52. Filipski E, King VM, Li X, Granda TG, Mormont MC, Liu X, Offerhaus GJ, et al. The SMAD4 protein and prognosis
et al. Host circadian clock as a control point in tumor of pancreatic ductal adenocarcinoma. Clin Cancer Res.
progression. JNatl CancerInst. 2002;94:690–7. 2001;7:4115–21.
53. DavisS,MirickDK.Circadiandisruption,shiftworkandriskof 72. Kaklamani VG, Hou N, Bian Y, Reich J, Offit K, Michel LS,
cancer:asummaryoftheevidenceandstudiesinSeattle.Cancer et al. TGFBR1*6A and cancer risk: a meta-analysis of seven
Causes Control.2006;17:539–45. case–control studies. J Clin Oncol. 2003;21:3236–43.
54. IARC. IARC monographs on the evaluation of carcinogenic 73. WilliamsCS,MannM,DuBoisRN.Theroleofcyclooxygenase
risks to humans; volume 98, Shift work, painting and fire- in inflammation, cancer, and development. Oncogene.
fighting.Lyon: International Agency forResearch onCancer. 1999;18:7908–16.
55. Brzezinski A. Melatonin in humans. N Engl J Med. 74. Juuti A, Louhimo J, Nordling S, Ristimäki A, Haglund C.
1997;336:186–95. Cyclooxygenase-2 expression correlates with poor prognosis in
56. Bouatia-NajiN,BonnefondA,Cavalcanti-ProençaC,SparsøT, pancreatic cancer. J Clin Pathol.2006;59:382–6.
Holmkvist J, Marchand M, et al. A variant near MTNR1B is 75. Zhao D, Xu D, Zhang X, Wang L, Tan W, Guo Y, et al.
associatedwithincreasedfastingplasmaglucoselevelsandtype Interaction of cyclooxygenase-2 variants and smoking in
2 diabetes risk.NatGenet. 2009;41:89–94. pancreatic cancer: a possible role of nucleophosmin.
57. Lyssenko V, Nagorny CL, Erdos MR, Wierup N, Jonsson A, Gastroenterology. 2009;136:1659–68.
Spégel P, et al. Common variant in MTNR1B associated with 76. Duell EJ, Casella DP, Burk RD, Kelsey KT, Holly EA.
increased risk of type 2 diabetes and impaired early insulin Inflammation, genetic polymorphisms in proinflammatory
secretion. NatGenet.2009;41:82–8. genes TNF-A, RANTES, and CCR5, and risk of pancreatic
J Epidemiol 2011;21(1):2-12