Cancer Identiﬁcation of Pleiotropic Cancer Susceptibility ... · cancer risk, comprised of 989 unique variants associated with 27 unique cancer sites. We found 20 pleiotropic variant

Research Article

Identification of Pleiotropic Cancer SusceptibilityVariants from Genome-Wide Association StudiesReveals Functional CharacteristicsYi-Hsuan Wu1, Rebecca E. Graff1, Michael N. Passarelli2, Joshua D. Hoffman1,Elad Ziv3,4,5, Thomas J. Hoffmann1,3, and John S.Witte1,3,5,6

Abstract

Background: There exists compelling evidence that somegenetic variants are associated with the risk of multiple cancersites (i.e., pleiotropy). However, the biological mechanismsthrough which the pleiotropic variants operate are unclear.

Methods: We obtained all cancer risk associations from theNational Human Genome Research Institute-European Bioinfor-matics Institute GWASCatalog, and correlated cancer risk variantswere clustered into groups. Pleiotropic variant groups and geneswere functionally annotated. Associations of pleiotropic cancerrisk variants with noncancer traits were also obtained.

Results:We identified 1,431 associations between variants andcancer risk, comprised of 989 unique variants associated with 27unique cancer sites. We found 20 pleiotropic variant groups(2.1%) composed of 33 variants (3.3%), including novel pleio-tropic variants rs3777204 and rs56219066 located in the ELL2

gene. Relative to single-cancer risk variants, pleiotropic variantswere more likely to be in genes (89.0% vs. 65.3%, P ¼ 2.2 �10�16), and to have somewhat larger risk allele frequencies(median RAF ¼ 0.49 versus 0.39, P ¼ 0.046). The 27 genes towhich the pleiotropic variantsmappedwere suggestive for enrich-ment in response to radiation and hypoxia, alpha-linolenic acidmetabolism, cell cycle, and extension of telomeres. In addition,we observed that 8 of 33 pleiotropic cancer risk variants wereassociated with 16 traits other than cancer.

Conclusions: This study identified and functionally character-ized genetic variants showing pleiotropy for cancer risk.

Impact: Our findings suggest biological pathways common todifferent cancers andotherdiseases, andprovide abasis for the studyof genetic testing for multiple cancers and repurposing cancertreatments. Cancer Epidemiol Biomarkers Prev; 27(1); 75–85. �2017 AACR.

IntroductionAn emerging focus in cancer research is the discovery and

understanding of the shared genetic basis underlying the devel-opment of different cancer types. In the past 10 years, genome-wide association studies (GWAS) have identified hundreds ofgenetic variants associated with cancer risk (1–3), and several locihave been associated with multiple cancer sites. For example,variants at the 8q24 locus have been associated with prostate(4, 5), colorectal (6–8), bladder (9), breast (10), and ovariancancers (11), glioma (12), and chronic lymphocytic leukemia

(13, 14). The genes closest to this locus are FAM84B and MYC,both known cancer-related genes. As another example, the 5p15locus containing TERT and CLPTM1L is associated with multiplecancer types, including lung (15, 16), testicular (17), prostate(18–20), breast (21), colorectal (22) cancers, and glioma (12).

Pleiotropy refers to the phenomenon of a gene or geneticvariant affecting more than one phenotypic trait. Identifying andcharacterizing pleiotropic genes and variantsmay have importantclinical and pharmacologic implications (23, 24). For example, adrug used for one cancer type may be repurposed to treat anothercancer type if the therapeutic target is common to both cancers. Inaddition, genetic tests for pleiotropic variants may provide anefficient way to identify patients at high risk of multiple cancers.Understanding the functional mechanisms by which variantsexhibit pleiotropy is important toward prioritizing potential drugor genetic testing targets.

Recent studies have looked at whether genetic variants previ-ously associatedwith one cancer are associatedwith other cancers.Cancers studied in this way include endometrial (25), colorectal(26, 27), pancreatic (28), esophageal (29, 30), prostate (31), lung(32), ovarian (33), gastric (34), and estrogen receptor negative(ER-) breast cancers (35), and non-Hodgkin lymphoma (36).Cross-cancer GWAS analyses for two to five cancers have also beenconducted to identify pleiotropic variants (37–40). Previousworkhas also estimated the genetic correlation between pairs of cancersusing data from GWAS for multiple cancer sites (41, 42).

We build on this previous work by investigating pleiotropyacross all cancer results presented in the National Human

1Department of Epidemiology and Biostatistics, University of California SanFrancisco, San Francisco, California. 2Department of Epidemiology, GeiselSchool of Medicine, Dartmouth College, Hanover, New Hampshire. 3Institutefor Human Genetics, University of California San Francisco, San Francisco,California. 4Division of General Internal Medicine, Department of Medicine,University of California San Francisco, San Francisco, California. 5Helen DillerFamily Comprehensive Cancer Center, University of California San Francisco,San Francisco, California. 6Department of Urology, University of California SanFrancisco, San Francisco, California.

Note: Supplementary data for this article are available at Cancer Epidemiology,Biomarkers & Prevention Online (http://cebp.aacrjournals.org/).

Corresponding Author: John S. Witte, University of California San Francisco,1450 3rd St, San Francisco, CA 94158. Phone: 415-502-6882; Fax: 415-476-1356;E-mail: [email protected]

doi: 10.1158/1055-9965.EPI-17-0516

�2017 American Association for Cancer Research.

CancerEpidemiology,Biomarkers& Prevention

www.aacrjournals.org 75

on May 1, 2020. © 2018 American Association for Cancer Research. cebp.aacrjournals.org Downloaded from

Published OnlineFirst November 17, 2017; DOI: 10.1158/1055-9965.EPI-17-0516

http://crossmark.crossref.org/dialog/?doi=10.1158/1055-9965.EPI-17-0516&domain=pdf&date_stamp=2017-12-28

http://cebp.aacrjournals.org/

Genome Research Institute-European Bioinformatics Institute(NHGRI-EBI) GWAS Catalog (1–3). The GWAS Catalog providespublicly available, manually curated, and literature-derived SNP-trait associations with P values <10�5 from GWAS assessing atleast 100,000 SNPs.

Previous analyses of the GWAS Catalog data found substantialevidence of pleiotropy across various traits (43). However, thiswork did not fully investigate potential pleiotropy arising fromvariants in linkage disequilibrium (LD) with the associated var-iants, the functional implications of pleiotropic variants, or theancestral populations in which the variants were detected. In thisstudy, we addressed these limitations by evaluating variants in LDwith the reported variants, functionally characterizing the GWASreported variants, and incorporating ancestry information. Fur-thermore, we investigated the associations of the pleiotropiccancer risk variants with other diseases and traits.

Materials and MethodsDetermining associations with cancer risk in the GWAS catalog

Weaccessed all associations reported in theGWASCatalog as ofSeptember 27, 2016. These were mapped to Ensembl releaseversion 85 and contained associations published from March10, 2005 through October 30, 2015. For associations with anygiven trait, the data contained the most statistically significantvariant from each independent locus for each study. To performan initial screening for associations with cancer risk, we utilizedExperimental Factor Ontology (EFO) terms (release 2016-03-15;refs. 44, 45). The curated traits in the GWAS Catalog are mappedto EFO terms to facilitate cross-study comparisons. The initial setof associations we evaluated included traits mapped to the term"neoplasm", defined as benign or malignant tissue growth result-ing from uncontrolled cell proliferation (44). "Neoplasm" and itsdescendant terms include both cancers and benign tumors (Sup-plementary Fig. S1).

Our goal was to identify variants pleiotropic for cancer suscep-tibility, so we limited the associations included in our analysis tothose specific to the risk of individual cancer types andnot toothercancer outcomes. We excluded associations with curated traitscontaining any of the following terms: "survival," "recurrence,""prognosis," "level," "symptom," "toxicity," "mortality," "treat-ment," "response," "metastasis," "aggressiveness," and "interac-tion." We then manually reviewed each of the remaining associa-tions and excluded those reported for gene–gene interactions,noncancerous traits, andmixed cancer sites (i.e., combining lung,gastric and esophageal cancers). We also excluded associationsreported for haplotypes and for variants in the HLA region forwhich rs number, chromosomal position, and allele name wereunavailable.

Associations with the same cancer site but different histologicsubtypes were categorized as being from the same cancer site. Wethen calculated thenumber of variants associatedwith each cancersite, and the number of studies reporting associations for eachcancer site.

Estimating LD among cancer risk–associated variantsTo determine the ancestry of the discovery samples within

which associations were identified, we relied on data providedby the GWAS Catalog, which assigned one of 15 ancestral cate-gories to each association. We collapsed categories by the 1000Genomes Project's super populations [European (EUR), East

Asian (EAS), Ad Mixed American (AMR), African (AFR), andSouth Asian (SAS)]. The number of cancer risk associations foreach super population was calculated. For associations withoutancestry data reported in the Catalog, we reviewed the originalpublications to obtain the ancestry information.

Our goal was to identify the following two types of cancer riskvariants: (i) pleiotropic within ethnic group, and (ii) pleiotropicacross ethnic groups. To identify the first type, we estimatedpairwise LD among variants discovered in the same super pop-ulation using reference genotype data from the correspondingsuper population. To identify the second type, we estimatedpairwise LD among all variants, regardless of the discovery pop-ulation, using reference genotype data from each of the five 1000Genomes Project's (46) super populations individually.

We ensured that all rs numberswereupdated tobuild 142of theSingle Nucleotide PolymorphismDatabase (dbSNP; http://www.ncbi.nlm.nih.gov/projects/SNP/; ref. 47). For variants lacking rsnumbers in the original publications, we used chromosomalpositions and the UCSC Genome Browser (48) to obtain rsnumbers. LD was estimated with LDlink (http://analysistools.nci.nih.gov/LDlink/) (49), which uses genotype data from Phase3 of the 1000 Genomes Project and variant rs numbers indexedbased on dbSNP build 142. HaploReg v4.1 (http://www.broadinstitute.org/mammals/haploreg; refs. 50, 51) was used to eval-uate LD for variants that could not be assessed by LDlink.Wewereunable to calculate LD for variants that were monoallelic in agiven population and/or not in the 1000 Genomes data.

Identifying variants associated with the risk of multiple cancersites

First, to identify variants pleiotropic within the same ethnicgroup, we grouped variants based on LD estimated in each superpopulation. Second, to identify variants pleiotropic across ethnicgroups, we grouped all variants based on LD estimated in each ofthe five super populations; doing so yielded five different sets ofvariant groupings.

We took two steps to group cancer risk variants in high LD: (i)variant pairs with R2� 0.8 were clustered into variant groups, and(ii) variant groups sharing at least one variant were merged.Therefore, within each variant group, each variant was in LD withat least one other variant (e.g., Supplementary Fig. S2). A variantgroup was defined as pleiotropic if it was associated with the riskof more than one cancer site (P < 10�5). Single-cancer variantgroups were associated with the risk of only one cancer site. Insensitivity analyses, we explored variant groupings based ondifferent levels of LD (R2 of � 0.7 or � 0.6).

We calculated the median and interquartile range (IQR) of theassociation odds ratios and risk allele frequencies (RAF) forpleiotropic and single-cancer variants. As these were not normallydistributed, we mainly compared them using the Wilcoxon test.

Functional annotations of variants and genes pleiotropic forcancer risk

For variant-level functional annotation, we first usedHaploRegv4.1 (http://compbio.mit.edu/HaploReg; refs. 50, 51) to obtainall variants in strong LD (R2 � 0.8) with the pleiotropic riskvariants [based on the 1000 Genomes Phase 1 European (EUR)population]. The locations of and consequences on proteinsequences for these variants were determined using the EnsemblVariant Effect Predictor (VEP; ref. 52). We picked consequencetypes for each variant using two different options in VEP. The

Wu et al.

Cancer Epidemiol Biomarkers Prev; 27(1) January 2018 Cancer Epidemiology, Biomarkers & Prevention76



http://www.ncbi.nlm.nih.gov/projects/SNP/

http://www.ncbi.nlm.nih.gov/projects/SNP/

http://analysistools.nci.nih.gov/LDlink/

http://analysistools.nci.nih.gov/LDlink/

http://www.broadinstitute.org/mammals/haploreg

http://www.broadinstitute.org/mammals/haploreg

http://compbio.mit.edu/HaploReg


"–most_severe" option was used to select only the most severeconsequence (Supplementary Table S1). The "–pick" option wasused to select one or more consequences according to an orderedset of criteria (Supplementary Table S2). We grouped the con-sequences into three main categories: "gene variant," "intergenicvariant," and "regulatory region variant." In addition, we didvariant group-level functional annotation, and the "–most_severe" and "–pick" categories were used to select one conse-quence type per variant group.

Functional annotation was also performed on the gene level.Genic cancer risk variants and all other variants in strong LD(R2 � 0.8) were mapped to RefSeq genes using HaploReg v4.1(50, 51).We used theGene IDConversion Tool inDAVID (http://david.ncifcrf.gov/; refs. 53, 54) to convert RefSeq Accession toEntrez Gene ID. Overrepresentation of pleiotropic genes in bio-logical processes based on Gene Oncology (GO) was tested usingConsensusPathDB (55). Overrepresentation tests comparingpleiotropic genes inReactome (releaseDecember 7, 2016; refs. 56,57) pathways were conducted using the PANTHER Overrepre-sentation Test (release April 13, 2017; ref. 58).

Assessing associations between pleiotropic cancer risk variantsand other traits

As above, we used LDlink or HaploReg to obtain all variants instrong LD (R2 � 0.8) with the pleiotropic cancer risk variants. LDwas based on the 1000 Genomes Project super population thatreflected the discovery sample of the variant. These variants weresearched in the GWAS Catalog to identify associations with traitsother than cancer risk.

ResultsSummary of cancer risk associations in the GWAS Catalog

We evaluated the 28,643 associations with P < 10�5 publishedin the GWAS Catalog, and identified 1,711 that mapped to one of1,395 relevant EFO terms (i.e., "neoplasm" or its descendants;Fig. 1). Among the 1,711 associations, we excluded 171 that didnot address susceptibility (e.g., gene–gene and gene–environmentinteractions, survival, and aggressiveness). After manually review-

ing each of the remaining associations, we further excluded85 with noncancerous traits (e.g., cutaneous nevi, percent mam-mographic density), two associations with mixed cancer sites(both combining lung, gastric, and esophageal cancers; ref. 39),20 associations with haplotypes, and two associations missingrs number, chromosomal position, and name of HLA allele.Ultimately, 1,431 cancer risk associations with P < 10�5 (927with P < 5 � 10�8) formed by 989 variants were identified from227 studies.

The associations were grouped into 27 cancer sites (Table 1).The number of variants associated with prostate or breast cancerwas almost twice the number of variants associated with all otherindividual cancer sites, partially reflecting the larger number ofGWAS conducted for these two cancer sites. Other cancers withmore than 50 associated variants were leukemia, lymphoma, andcolorectal, pancreatic, skin, and lung cancers.

Cancer risk associations were discovered from 12 differentpopulations. We observed that 993 (69.4%) associationswere identified in an initial sample of Europeans, 250 (17.5%)were from East Asians, and 98 (6.8%) were from a samplecontaining European, East Asian, Hispanic, and African ancestries

Figure 1.

Flow diagram for obtaining all cancer risk associations. On the basis of the 1,395 EFO "neoplasm" terms, we identified 1,711 associations with "neoplasm." Followingexclusions, we obtained 1,431 cancer risk associations.

Table 1. Number of variants and GWAS for each of the 27 cancer sites

Cancer site

Number ofvariants(studies) Cancer site

Numberof variants(studies)

Prostate 166 (26) Cervix 17 (3)Breast 145 (34) Stomach 17 (5)Leukemia 95 (14) Brain 15 (8)Colon & rectum 81 (20) Nasopharynx 14 (4)Pancreas 59 (7) Liver 10 (5)Skin 55 (15) Neuroblastoma 10 (4)Lung 52 (20) Bone 9 (2)Lymphoma 52 (13) Endometrium 9 (2)Multiple myeloma 45 (4) Thyroid 8 (5)Esophagus 37 (7) Larynx 6 (1)Ovary 34 (9) Gallbladder 5 (1)Testis 27 (7) Salivary gland 5 (1)Urinary bladder 22 (9) Upper aerodigestive tracta 5 (1)Kidney 18 (5) All sites 989 (227)aUpper aerodigestive tract: oral cavity, pharynx, larynx, and esophagus.

Pleiotropic Cancer Risk Variants

www.aacrjournals.org Cancer Epidemiol Biomarkers Prev; 27(1) January 2018 77



http://david.ncifcrf.gov/

http://david.ncifcrf.gov/


(Supplementary Table S3). The variation in these percentagesreflects the differences in how many people from each of thesepopulations have been included in GWAS.

Genetic variants showing pleiotropy for cancer riskAmong the 939 variant groups obtained using R2 � 0.8 as a

threshold (Supplementary Table S3), 20 (2.1%) exhibited plei-otropy for cancer risk within the same ethnic group (Table 2). Weconfirmed that all grouped variants had high LD. In particular,within the 20 pleiotropic variant groups that we identified, thelowest LD between any two variants was R2 ¼ 0.735.

Of the 20 pleiotropic variant groups, 17 variant groups werefrom European populations and three variant groups were fromEast Asian populations. The 20 pleiotropic variant groups werecomposed of 33 (3.3%) variants, and the remaining 956 (96.7%)cancer risk variants were classified as single-cancer variants. Thepleiotropic variants are located in 27 genes such asMDM4, ELL2,MLLT10, BCL2, BRCA2, BABAM1 and ANKLE1. We observed thatthe cancer risk associations were similar for pleiotropic variants[median associationOR¼1.20; interquartile range (IQR)¼1.15–1.27] and for single-cancer risk variants (median associationOR¼ 1.23; IQR¼ 1.14–1.38; Wilcoxon test P¼ 0.15). However,the pleiotropic variants had slightly higher risk allele frequency(median RAF¼ 0.49; IQR¼ 0.30–0.54) than observed for single-cancer risk variants (median RAF ¼ 0.39; IQR ¼ 0.21–0.59;Wilcoxon test P ¼ 0.046; Supplementary Table S4).

In addition, we clustered variant groups according to differentLD thresholds (Supplementary Table S3). Using a threshold of R2

�0.7,we identified one additional variant group (21 total groups;41 variants) showing pleiotropy for cancer risk within the sameethnic group. Using a threshold of R2� 0.6, we identified yet onemore additional variant group (22 total groups; 48 variants).

Variants were also grouped regardless of the discovery samplesto identify those which were pleiotropic across ethnic groups.Using R2� 0.8 as the threshold, we identified 9 variant groups (18variants) pleiotropic for cancer risk across ethnic groups (Sup-plementary Tables S5 and S6). The lower theR2 threshold used forvariant grouping, the more pleiotropic variants we identified.Overall, approximately 2%–4% of variant groups (3%–7% ofvariants) were pleiotropic (Supplementary Fig. S3).

Functional characterizations of pleiotropic cancer risk variantsand genes

As variants reported in theGWASCatalogmaynot be causal butrather tag the true causal variants, we incorporated variants in LDwhenperforming functional annotations to try to capture asmuchinformation as possible. We identified 518 variants in strong LD(R2� 0.8) with the 33 pleiotropic cancer risk variants, and 18,069variants in strong LD with the 956 single-cancer variants. Weobserved that the most severe consequences were statisticallydifferent between pleiotropic and single-cancer variants (Fisherexact test P ¼ 2.2 � 10�16; Table 3). A higher percentage ofpleiotropic cancer risk variants were genic (89.0%) comparedwith single-cancer variants (65.3%). Within genes, most of thepleiotropic variants were in introns, 30 untranslated regions(UTR), or they changed exon sequence in a noncoding transcript.Pleiotropic cancer risk variantswere also less likely tobe intergenic(11.0%) compared with single-cancer variants (31.8%) andmorelikely to be upstreamof genes (7.4%vs. 5.3%). Interestingly, noneof the pleiotropic cancer risk variants were located in regulatory

regions such as transcription factor binding sites or other non-genic regions (0% vs. 2.9%). Selecting consequence accordingto an ordered set of criteria ("–pick" option) provided similarpercentages of gene and intergenic variants (SupplementaryTable S7). Likewise, annotations on variant-group levelshowed that pleiotropic variant groups tended to be in genesmore often than in intergenic regions (Supplementary TablesS8 and S9).

For the genic variants, the 460 pleiotropic ones mapped to 27genes, while the 11,755 single-cancer variants mapped to 612genes. Relative to single cancer genes, pleiotropic genes hadsuggestive enrichment (P < 0.008) in the following: response tostimuli such as light, radiation, oxygen and organic cyclic com-pounds, cell aging, and directing movement of a protein to aspecific location on a chromosome (Supplementary Table S10).Although not statistically significant after correction for multipletesting, the most overrepresented pathways for pleiotropic genesincluded alpha-linolenic acid (ALA) metabolism, cell cycle, andextension of telomeres (Supplementary Table S11).

Associations between pleiotropic cancer risk variants and traitsother than cancer

Detecting the associations between pleiotropic cancer riskvariants and noncancer traits has the potential to suggest sharedunderlying biology across different traits, which may reflect com-monunderlyingmechanisms (e.g., inflammation).We found that8 of 33 pleiotropic cancer risk variants that we identified wereassociatedwith 16 other complex diseases or traits investigated byGWAS (Fig. 2). Variants rs10936599 and rs12696304 located inMYNN and near TERC were associated with telomere length,celiac disease, and multiple sclerosis. Variant rs2736100 in TERTwas associated with telomere length, red blood cell count, andlung diseases. The CLPTM1L gene variants rs401681, rs31489,rs31490, and rs4975616 were associated with serum prostate-specific antigen (PSA) levels. Variant rs2294008 located in PSCAgene was associated with duodenal ulcers. Pleiotropic variantsrs174537 in MYRF and rs174549 in FADS1 were associated withmany lipid metabolism-related traits. Variants rs4430796 andrs8064454 in HNF1B were also associated with type 2 diabetesand PSA levels. The 672 variants in strong LD (R2 � 0.8) with the33 pleiotropic cancer risk variants were associated with an addi-tional 24 traits (Supplementary Table S12).

DiscussionThere is considerable interest in cancer pleiotropy as it may

highlight important molecular mechanisms and have implica-tions for drug development. Our analysis across 27 cancer sitesusing the publicly available NHGRI-EBI GWAS Catalog detectednumerous pleiotropic cancer risk variants and evaluated theirfunctional characteristics.

We uncovered some novel patterns of pleiotropy for knowncancer risk loci. Our study is the first to highlight that variants inMLLT10 at 10p12.31 are associated with both ovarian cancer andmeningioma. MLLT10 is known to encode a transcription factorinvolved in chromosomal rearrangements in leukemia (59).Nonetheless, there is currently nodirect evidence for howMLLT10is involved in developing ovarian cancer or meningioma otherthan GWAS. Another novel pattern of pleiotropy that we foundwas that the variant rs4245739 in MDM4 at 1q32.1 is associatedwith prostate cancer and ER-negative and triple-negative breast

Wu et al.





Table

2.The

20varian

tgroup

s(33varian

ts)pleiotropic

forcancer

risk

withinthesameethn

icgroup

Variant

group

Reg

ion

Can

cersites

Variant

rsnu

mber

LD(R

2)a

Can

cersite

asso

ciated

with

theva

rian

tEAb

EAFb

Pc

ORb

95%

confi

den

ceinterval

of

ORb

Anc

estryof

thedisco

very

sample

dMap

ped

gen

ee

Pub

Med

IDofthe

stud

y

rs424

5739

1q32

.1Breast,Prostate

rs424

5739

1Breast

C0.26

2E-12

1.14

[1.10

–1.18

]Europea

nMDM4

2353

5733

rs424

5739

1Prostate

C0.25

2E-11

0.91

[0.88–0

.95]

Europea

nMDM4

2353

5732

rs10936

599,

rs12638

862,

rs1269630

4

3q26

.2Colon&rectum

,Le

ukem

ia,M

ultiple

mye

loma,

Skin,

Urina

rybladder

rs12638

862

1(Ref)

Multiplemye

loma

A0.74

2E-06

1.37

[1.20–1.56]

Europea

n4.9

kb30

ofTE

RC

2350

2783

rs1269630

40.945

Skin

C0.73

3E-07

1.10

[NR]

Europea

n1.1

kb30

ofTE

RC

2623

7428

rs10936

599

0.928

Urina

rybladder

C0.76

5E-09

1.18

[1.11–1.23]

Europea

nMYNN

24163127

rs10936

599

0.928

Colon&rectum

C0.75

3E-08

1.08

[1.04–1.10

]Europea

nMYNN

20972

440

rs10936

599

0.928

Leuk

emia

C0.75

2E-09

1.26

[1.17

–1.35]

Europea

nMYNN

2429

2274

rs10936

599

0.928

Multiplemye

loma

C0.80

9E-14

1.26

[1.18

–1.33]

Europea

nMYNN

23955

597

rs10069690

5p15.33

Breast,Ovary

rs10069690

1Breastf

T0.32

5E-12

1.15

[1.11–1.20]

Europea

nTE

RT

2353

5733

rs10069690

1Ovary

T0.26

9E-09

1.14

[1.10

–1.19

]Europea

nTE

RT

2558

1431

rs27

36100

5p15.33

Brain,L

ung,T

estis

rs27

36100

1Brain

G0.49

2E-17

1.27

[1.19

–1.37]

Europea

nTE

RT

1957

836

7rs27

36100

1Lu

ngf

G0.50

2E-10

1.12

[1.08–1.16

]Europea

nTE

RT

19836

008

rs27

36100

1Testis

G0.51

8E-15

0.75

[0.67–

0.85]

Europea

nTE

RT

2054

3847

rs31489,

rs31490,

rs401681,

rs4975

616

5p15.33

Leuk

emia,L

ung,

Pan

crea

s,Skin,

Urina

rybladder

rs4975

616

1(Ref)

Lung

NR

NR

3E-09

1.15

[1.10

–1.20]

Europea

n2.2kb

30ofCLP

TM1L

19654

303

rs401681g

0.849

Urina

rybladder

C0.54

4E-11

1.12

[1.08–1.16

]Europea

nCLP

TM1L

24163127

rs401681

0.849

Lung

C0.57

8E-09

1.15

[1.09–1.19

]Europea

nCLP

TM1L

18978

787

rs401681

0.849

Skin

C0.55

9E-13

1.21

[NR]

Europea

nCLP

TM1L

25855

136

rs31489

0.735

hLu

ngC

0.59

2E-10

1.12

[1.09–1.16

]Europea

nCLP

TM1L

19836

008

rs31490

0.849

Leuk

emia

A0.43

2E-07

1.18

[1.11–1.26]

Europea

nCLP

TM1L

2429

2274

rs31490

0.849

Pan

crea

sA

0.44

2E-11

1.20

[1.14

–1.27]

Europea

nCLP

TM1L

25086665

rs37

7720

4,

rs56

219066

5q15

Salivarygland

,Multiple

mye

loma

rs37

7720

4i

1(Ref)

Salivarygland

C0.29

1E-07

1.86

[1.48–2

.34]

Europea

nELL

2j25

823

930

rs56

219066

0.971

Multiplemye

loma

T0.71

1E-09

1.25

[1.16

–1.34]

Europea

nELL

2j26

0076

30rs24

94938

6p21.1

Lung

,Stomach

rs24

94938

1Lu

ngA

0.23

2E-06

1.15

[1.08–1.22]

EastAsian

LRFN2

231032

27rs24

94938

1Stomach

A0.23

5E-09

1.18

[1.12

–1.25]

EastAsian

LRFN2

231032

27rs22

859

47

7p15.3

Esopha

gus,L

ung,

Stomach

rs22

859

47

1Esopha

gus

A0.27

3E-06

1.14

[1.08–1.21]

EastAsian

DNAH11

231032

27rs22

859

47

1Lu

ngA

0.26

2E-08

1.17

[1.11–1.24]

EastAsian

DNAH11

231032

27rs22

859

47

1Stomach

A0.27

1E-06

1.14

[1.08–1.21]

EastAsian

DNAH11

231032

27rs1050

5477

,rs69832

67

8q24

.21

Colon&rectum

,Prostate

rs1050

5477

1(Ref)

Colon&rectum

T0.54

8E-13

1.20

[NR]

Europea

n20

kb50

ofPOU5F

1B24

7377

48

rs1050

5477

1(Ref)

Prostate

T0.49

9E-09

1.39

[1.28–1.50]

Europea

n20

kb50

ofPOU5F

1B24

740154

rs69832

67

0.916

Colon&rectum

fG

0.49

1E-14

1.27

[1.16

–1.39]

Europea

n15

kb50

ofPOU5F

1B1761828

4rs69832

67

0.916

Prostatef

G0.50

4E-15

1.34

[1.25–

1.43]

Europea

n15

kb50

ofPOU5F

1B24

7535

44

rs22

94008

8q24

.3Stomach,

Urina

rybladder

rs22

94008

1Urina

rybladder

T0.46

3E-15

1.13

[1.10

–1.16

]Europea

nPSC

A24

163127

rs22

94008

1Stomach

T0.47

2E-07

1.21

[NR]

Europea

nPSC

A26

098866

rs1101273

2,rs1243180

10p12.31

Brain,O

vary

rs1101273

21(Ref)

Brain

A0.32

2E-14

1.46

[1.32–

1.61]

Europea

nMLL

T10

2180454

7rs1243180

0.858

Ovary

A0.31

1E-09

1.10

[1.06–1.14

]Europea

nMLL

T10

2558

1431

rs17453

7,rs17454

911q12.2

Colon&rectum

,Laryn

xrs17453

71(Ref)

Colon&rectum

G0.59

9E-21

1.16

[1.12

–1.19

]EastAsian

MYRF

24836

286

rs17454

90.923

Larynx

A0.59

1E-20

1.37

[1.28–1.47]

EastAsian

FADS1

2519428

0rs73

5665

11q24

.1Le

ukem

ia,L

ympho

ma

rs73

5665

1Le

ukem

iaA

0.19

4E-39

1.62

[NR]

Europea

n35

kb50

ofGRAMD1B

2377

0605

rs73

5665

1Ly

mpho

ma

A0.21

4E-09

1.81

[1.50–2.20]

Europea

n35

kb50

ofGRAMD1B

20639

881

rs1157

1833

13q13.1

Breast,Lu

ngrs1157

1833

1Breast

T0.01

5E-08

1.26

[1.14

–1.39]

Europea

nBRCA2

2353

5729

rs1157

1833

1Lu

ngT

0.01

5E-20

2.47

[2.03–

3.00]

Europea

nBRCA2

2488034

2rs35

158985,

rs9929

218

16q22

.1Colon&rectum

,Skin

rs35

158985

1(Ref)

Skin

G0.30

3E-07

1.10

[NR]

Europea

nCDH1

2623

7428

rs9929

218

0.861

Colon&rectum

G0.29

1E-08

1.10

[1.06–1.12

]Europea

nCDH1

19011631

rs4430

796,

rs8064454

17q12

End

ometrium

,Prostate

rs4430

796

1(Ref)

End

ometrium

A0.52

7E-10

1.19

[1.12

–1.27]

Europea

nHNF1B

2149925

0rs4430

796

1(Ref)

Prostatef

A0.49

1E-11

1.22

[1.15

–1.30]

Europea

nHNF1B

176034

85

rs8064454

0.949

Prostate

C0.52

8E-29

1.24

[1.19

–1.29]

Europea

nHNF1B

25939

597

(Continue

donthefollowingpag

e)






Table

2.The

20varian

tgroup

s(33varian

ts)pleiotropic

forcancer

risk

withinthesameethn

icgroup

(Cont'd)

Variant

group

Reg

ion

Can

cersites

Variant

rsnu

mber

LD(R

2)a

Can

cersite

asso

ciated

with

theva

rian

tEAb

EAFb

Pc

ORb

95%

confi

den

ceinterval

of

ORb

Anc

estryof

thedisco

very

sample

dMap

ped

gen

ee

Pub

Med

IDofthe

stud

y

rs75

01939

,rs75

7210

17q12

Prostate,

Testis

rs75

7210

1(Ref)

Ovary

G0.37

8E-10

1.12

[1.08–1.17

]Europea

nHNF1B

2353

5730

rs75

01939

0.8

Prostatef

NR

NR

3E-18

NR

[NR]

Europea

nHNF1B

1976

7753

rs75

01939

0.8

Testis

C0.62

1E-09

1.28

[1.19

–1.39]

Europea

nHNF1B

25877

299

rs1774

956

1,rs49878

5518q21.33

Leuk

emia,L

ympho

ma

rs1774

956

11(Ref)

Lympho

ma

G0.91

8E-10

1.34

[1.22–

1.47]

Europea

n7.4kb

30ofBCL2

2527

9986

rs49878

550.951

Leuk

emia

G0.91

3E-12

1.47

[1.32–

1.61]

Europea

nBCL2

2377

0605

rs8170

19p13.11

Breast,Ovary

rs8170

1Breast

C0.48

4E-13

1.19

[1.14

–1.25]

Europea

nBABAM1

2354

4013

rs8170

1Ovary

NR

0.19

3E-14

1.19

[1.14

–1.25]

Europea

nBABAM1

2353

5730

rs23

639

5619p13.11

Breast,Ovary

rs23

639

561

Breast

NR

NR

2E-08

1.22

[1.14

–1.30]

Europea

nANKLE

124

3259

15rs23

639

561

Ovary

T0.43

1E-07

1.10

[1.06–1.15

]Europea

nANKLE

120

852

633

aLink

agediseq

uilib

rium

(LD).

bEffecta

llele(EA),effectallelefreq

uency(EAF),oddsratio(O

R),95%

confi

den

ceintervalofO

Rwereretrieve

dfromtheGWASCatalog.W

edescribed

theeffectalleleinRefSNPorien

tationas

dbSNPdid.W

ealso

man

ually

review

edtheoriginalpub

lications

anddbSNPto

fillintheinform

ationno

trep

orted

intheGWASCatalog.N

R:Inform

ationno

trep

orted

ineither

theGWASCatalogorthe

originalpub

lications

ofthe

stud

y,an

dno

tableto

be

retrieve

dfrom

dbSNP.

c Forthesamevarian

t-cancer

risk

associationrepea

tedly

iden

tified

bymultiple

stud

ies,theone

withthesm

allest

Pvaluean

d/oriden

tified

bythemost

recent

stud

ywas

reported

inthistable.

dClusteringofvarian

tsinto

varian

tgroup

sbased

onpairw

iseLD

was

perform

edwithinea

chofthe17

ethn

icgroup

sshownin

Tab

le2,

andonlythevarian

tgroup

sin

theEuropea

nan

dEastAsian

populationshowed

pleiotropic

effectsforcancer

risk.

eForvarian

tslocatedwithinagen

e,thegen

esymbolwas

reported

directly.

Forintergen

icvarian

ts,d

istancean

ddirectionto

theclosest

gen

ewerereported

.f Statisticallysignificant

associationwas

also

observed

inother

populations:rs10069690an

dbreastcancer

inEuropea

nan

dAfrican

American

/Afro-Caribbean

(Pub

Med

ID:220

3755

3);rs273

6100an

dlung

cancer

inEast

Asian

(Pub

Med

ID:21725

308);rs69832

67an

dco

lorectalcancer

inEastA

sian

(Pub

Med

ID:24836

286);rs69832

67an

dprostatecancer

inasampleofE

uropea

n,African

American

/Afro-Caribbea

n,EastA

sian

,Hispan

ic/Latin

American

(Pub

Med

ID:26034

056

);rs4430

796an

dprostatecancer

inEastAsian

(Pub

Med

ID:264434

49);an

drs75

01939

andprostatecancer

inEastAsian

(Pub

Med

ID:20676

098),an

dinasampleofE

uropea

n,African

American

/Afro-Caribbean

,EastAsian

,Hispan

ic/Latin

American

(Pub

Med

ID:2

6034

056

).gVariant

rs401681h

asbee

nassociated

withpan

crea

ticcancer

inasampleofE

uropea

n,African

unspecified

,Asian

unspecified

,Hispan

ic/LatinAmerican

(Pub

Med

ID:26098869)an

dan

other

sampleofE

uropea

nan

dEast

Asian

(Pub

Med

ID:2

0101243).

hVariant

rs31489isin

highLD

withrs31490(R

2¼

0.874

)an

drs401681(R

2¼

0.814).

i Variant

rs37

7720

4was

shownas

exm22

659

79in

theoriginal

pub

lication.

j ELL2isapleiotropic

gen

ethat

isno

veltothisstud

y.

Wu et al.





cancer.MDM4 encodes a repressor that binds and inactivates p53and is considered important in cancer development (60, 61).Interestingly, we observed that the association of the C allele ofrs4245739 is in the opposite direction for prostate (OR ¼ 0.91;

95%CI: 0.88–0.95; ref. 62) andbreast cancer (OR¼1.14; 95%CI,1.10–1.18; ref. 63). On the basis of the Genotype-Tissue Expres-sion (GTEx) Project (64), rs4245739 is not associated withexpression in prostate or breast tissue, but is correlated with

Table 3. Comparison of the variant consequences between the 518 pleiotropic variants and the 18,069 single-cancer variants (P value of Fisher's exact test¼ 2.2�10�16) using the "–most_severe" option

Variant consequencea ImpactbNumber of pleiotropicvariants (%)c

Number of single-cancervariants (%)c

Gene variant Total 460 (89.0) Total 11,755 (65.3)Intron variant Modifier 369 (71.4) 10,703 (59.4)3 prime UTR variant Modifier 29 (5.6) 277 (1.5)Non coding transcript exon variant Modifier 26 (5.0) 454 (2.5)5 prime UTR variant Modifier 11 (2.1) 85 (0.5)Synonymous variant Low 11 (2.1) 94 (0.5)Missense variant Moderate 8 (1.5) 102 (0.6)Splice region variant Low 4 (0.8) 26 (0.1)Stop gained High 2 (0.4) 3 (0.0)Frameshift variant High 0 (0.0) 4 (0.0)Splice donor variant High 0 (0.0) 3 (0.0)Splice acceptor variant High 0 (0.0) 2 (0.0)Inframe insertion Moderate 0 (0.0) 1 (0.0)Start lost High 0 (0.0) 1 (0.0)

Intergenic variant Total 57 (11.0) Total 5,737 (31.8)Upstream gene variant Modifier 38 (7.4) 954 (5.3)Downstream gene variant Modifier 11 (2.1) 731 (4.1)(Other intergenic variant) - 8 (1.5) 4,052 (22.5)

Regulatory region variant Total 0 (0.0) Total 523 (2.9)TF binding site variant Modifier 0 (0.0) 9 (0.0)(Other regulatory region variant) - 0 (0.0) 514 (2.9)

aVariant consequence was predicted using the Ensembl Variant Effect Predictor (VEP).bImpact was defined by Ensembl to classify the severity of the variant consequence, with four categories: high, moderate, low, and modifier.cVariant consequence annotations for one pleiotropic variant (rs35464379) and 54 single-cancer variants were not available.

Figure 2.

Associationsof pleiotropic cancer risk variantswith other complex diseases and traits studiedbyGWAS.We identified8out of 33pleiotropic variantswere associatedwith other 16 distinct traits. The arrow indicates the direction of the association. NR: Information not reported in the GWAS Catalog.






PIK3C2B expression in testis. PIK3C2B encodes a PI3K that plays arole in many oncogenic pathways.

Our analysis of the GWAS Catalog also identified the novelpleiotropic gene ELL2. The variant rs3777204 (exm2265979 inthe original publication) is associated with salivary gland carci-noma and rs56219066 is associated with multiple myeloma inEuropean populations (R2 ¼ 0.971 between the two variants).ELL2 encodes an elongation factor for RNA polymerase II, animportant component of the super elongation complex (SEC;ref. 65) that regulates the transcriptional elongation checkpointcontrol (TECC) stage of transcription. Dysregulation is related tocarcinogenesis (66). We also found that variant rs56219066 isassociated with a reduction of IgA and IgG levels, which couldaffect the pre-mRNA processing andmalignant transformation inmultiple myeloma (67).

We also replicated a number of previously known pleiotropicvariants and loci. Variants in the TERC-MYNN region at 3q26.2and TERT-CLPTM1L region at 5p15.33 are pleiotropic for manycancers (68, 69). The telomerase reverse transcriptase (TERT;ref. 70) and its integral RNA template (TERC; ref. 71) are twosubunits of the telomerase ribonucleoprotein complex thatmaintain telomere length. Variant rs2736100 in TERT is asso-ciated with lung and testicular cancers and glioma in Europeanpopulations, but only the lung cancer association in East Asianpopulations was identified. To our knowledge, only one can-didate gene study found an association between rs2736100 andglioma risk in East Asians (P ¼ 3.69 � 10�4; ref. 72). Theassociation between rs2736100 and testicular cancer is in theopposite direction for lung cancer and glioma. It has beenspeculated that rs2736100-T mediates the recruitment of sexdetermining region Y transcription factor to TERT, which mightincrease the telomerase in germ cells, leading to a potentialincreased risk of testicular cancer (73). As another example,BRCA2 at 13q13.1 is a well-known pleiotropic gene (74–76).Variants in the BABAM1-ANKLE1 region at 19p13.11 are asso-ciated with breast and ovarian cancer. The BABAM1 geneencodes a component of the BRCA1 complex and BRCA1activates DNA repair in double-strand breaks (DSB) in coop-eration with BRCA2 (77); defective repair in DSB can lead totumorigenesis (78). As mutations in BRCA1 affect both breastand ovarian cancer risk, the association of BABAM1 with bothof these cancers is consistent with its known interaction withBRCA1. Finally, the 8q24 region is associated with diversecancers, and the HNF1B gene variants at 17q12 are specificallyassociated with hormone-related cancers, including prostate,endometrial, ovarian and testicular cancers.

Previouswork evaluating the associations in theGWASCatalogthrough 2011 estimated that 4.8% of SNPs associated with cancerexhibit pleiotropy (43). There are several possible reasons whywedetected a lower level of pleiotropy [2.1% variant groups (3.3%variants)] within ethnic groups. First, the previous study includedSNPs pleiotropic across ethnic groups. Second, we estimated LDusing the populations in which the variants were discovered,while the previous study used the HapMap CEU population tocalculate LD for all variants. Finally, we focused specifically oncancer risk instead of all cancer outcomes. The evaluation ofpleiotropy for treatment response is certainly compelling, but itis also very complicated. One must decide how to considertreatment type, efficacy, toxicity, and time to response, amongother dimensions. To maintain focus and comparability withprevious research, we elected to only evaluate here cancer sus-

ceptibility. Future work exploring pleiotropy specific to the treat-ment of cancer can be performed.

Spurious cross–phenotype associations can occur when indi-viduals suffering from one cancer are more likely to receivediagnostic evaluation and detection of other cancers (ascertain-ment bias; refs. 23, 79). In our study, this potential bias wasminimized because the associations studied generally camefrom distinct GWAS in which few subjects had multiple cancers.Nevertheless, the frequency of pleiotropy that we observeddepended upon the associations collected in the GWAS Cata-log, and the traits that we selected to study were not a randomsample of all traits. In addition, the frequency of pleiotropy islarger than we estimated because many common and rarevariants associated with cancer risk have yet to be found. Thatsaid, utilizing the EFO terms that map the traits increased thesensitivity of detecting cancer risk associations, while manuallyreviewing each of the identified associations avoided includingnoncancer risk outcomes.

We estimated that 89.0% of pleiotropic variants were in genes,in contrast with only 65.3% of nonpleiotropic variants. Pleiotro-pic variants may be more likely to be within a gene than to affectgene regulation because gene regulation is highly tissue specific.Thus, for a variant to affect risk inmultiple tissues, it needs to affectthe function of a gene in a way that transcends tissue specific generegulation in regulatory elements.

The suggestion of overrepresentation of the pathways responseto radiation and hypoxia, ALA metabolism, cell cycle, and exten-sion of telomeres in pleiotropic genes implies carcinogenicmechanisms common to the development of different cancersites. The comparison group for these functional enrichmentanalyses was the single-cancer variants. Misclassification of trulypleiotropic variants as observed single-cancer variants is possibleif associations with other cancer sites have not yet been detected(e.g., due to limited power). Future GWAS or sequencing focusedon rarer variants, cancer subtypes, or epigenetic regulations mayhelp identify additional novel pleiotropic mechanisms.

Some of our findings of pleiotropy are in agreement withcurrent clinical practices. For example, we observed that twohighly correlated variants (R2¼ 0.95) located in or near the BCL2gene, rs17749561 and rs4987855, were associated with follicularlymphoma (FL) and chronic lymphocytic leukemia (CLL). TheBCL2 gene family encodes proteins that regulate cellular apopto-sis, and serve essential roles of balancing cell survival and celldeath (80, 81). The drug venetoclax (also known as ABT-199) is aBCL-2 inhibitor that binds to BCL-2 with high affinity andselectivity (82). It was approved for CLL (83) but also showsfavorable efficacy and safety in patients with FL (84), reflecting thepleiotropic role of BCL2 in these two blood cancers. Our findingsalso suggest potential clinical applications. For example, variantsin the novel pleiotropic gene ELL2 that we identified were asso-ciated with salivary gland carcinoma and multiple myeloma.Expression of ELL2 was downregulated by microRNAs miR-155(85) and miR-299 (86). With the extensive development ofmiRNA therapeutics (87), ELL2 exhibits a promising candidateto treat these two cancers. Furthermore, our discovery of variantsthat exhibit opposite effects on cancers (e.g., variants inMDM4 orin TERT) may help identify drugs that should not be explored forrepurposing across cancers.

Our approach was limited by the selection of variants forgenotyping arrays and the imputation reference panels in thepublished GWAS. Genotyping arrays may have preferentially

Wu et al.





included variants located in genes or previously associated withdiseases, giving such variants an increased chance of being pleio-tropic. To try to avoid this potential bias, we compared thepleiotropic variants to single-cancer variants, which were alsotagged by GWAS arrays. We were also limited in our ability todistinguish biological from spurious pleiotropy. It is possible thatvariants in high LD could be functional in different genes. Ourstudy also had potential bias in that some cancers havemore thanone ethnicity represented inGWASand these cancersmaybemoreor less likely to have shared genetic causes. The absolute risk ofcancer affects howmanyGWAS are performed and their power, sopleiotropic variants may appear to cluster for common cancers,which might also have more ethnicities involved. In addition, wewere only able to evaluate pleiotropy among loci with strongenough associations to be reported in the GWAS Catalog. As allsummary statistics from GWAS become more widely available,more extensive evaluations of pleiotropy can be undertaken.

Overall, identification of pleiotropic cancer risk variants andgenes has important implications. The biological functions andpathways overrepresented in pleiotropic genes may inform ourunderstanding of the underlying mechanisms shared by differentcancers. Genetic tests for pleiotropic variants could be developedto efficiently identify high-risk patients. Drugs developed for onecancer might be valuable for use in treating other cancers as well.

Such repurposing of anticancer therapies may offer promisingopportunities for improving cancer treatment.

Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.

Authors' ContributionsConception and design: Y.-H. Wu, J.S. WitteDevelopment of methodology: Y.-H. Wu, R.E. Graff, T.J. Hoffmann, J.S. WitteAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): Y.-H. Wu, J.D. Hoffman, J.S. WitteWriting, review, and/or revision of the manuscript: Y.-H. Wu, R.E. Graff,M.N. Passarelli, J.D. Hoffman, E. Ziv, T.J. Hoffmann, J.S. WitteStudy supervision: E. Ziv

AcknowledgmentsThis work supported by NIH grants CA127298, CA088164, CA112355, and

CA201358 (to R.E. Graff, M.N. Passarelli, J.D. Hoffman, and J.S. Witte), and theUCSFGoldberg-Benioff Program inCancer Translational Biology (to J.S.Witte).

The costs of publication of this articlewere defrayed inpart by the payment ofpage charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received June9, 2017; revised September 5, 2017; acceptedOctober 17, 2017;published OnlineFirst November 17, 2017.

References1. Hindorff LA, Sethupathy P, Junkins HA, Ramos EM, Mehta JP, Collins FS,

et al. Potential etiologic and functional implications of genome-wideassociation loci for human diseases and traits. Proc Natl Acad Sci USA2009;106:9362–7.

2. MacArthur J, Bowler E, Cerezo M, Gil L, Hall P, Hastings E, et al. The newNHGRI-EBI Catalog of published genome-wide association studies (GWASCatalog). Nucleic Acids Res 2017;45:D896–D901.

3. Welter D, MacArthur J, Morales J, Burdett T, Hall P, Junkins H, et al. TheNHGRI GWAS Catalog, a curated resource of SNP-trait associations.Nucleic Acids Res 2014;42:D1001–D6.

4. Thomas G, Jacobs KB, YeagerM, Kraft P,Wacholder S, Orr N, et al. Multipleloci identified in a genome-wide association study of prostate cancer. NatGenet 2008;40:310–5.

5. Yeager M,Orr N, Hayes RB, Jacobs KB, Kraft P,Wacholder S, et al. Genome-wide association study of prostate cancer identifies a second risk locus at8q24. Nat Genet 2007;39:645–9.

6. Tomlinson I, Webb E, Carvajal-Carmona L, Broderick P, Kemp Z, Spain S,et al. A genome-wide association scan of tag SNPs identifies a susceptibilityvariant for colorectal cancer at 8q24. 21. Nat Genet 2007;39:984–8.

7. Zanke BW, Greenwood CM, Rangrej J, Kustra R, Tenesa A, Farrington SM,et al. Genome-wide association scan identifies a colorectal cancer suscep-tibility locus on chromosome 8q24. Nat Genet 2007;39:989–94.

8. Tenesa A, Farrington SM, Prendergast JG, Porteous ME, Walker M, Haq N,et al. Genome-wide association scan identifies a colorectal cancer suscep-tibility locus on 11q23 and replicates risk loci at 8q24 and 18q21. NatGenet 2008;40:631–7.

9. Rothman N, Garcia-Closas M, Chatterjee N, Malats N, Wu X, Figueroa JD,et al. A multi-stage genome-wide association study of bladder canceridentifies multiple susceptibility loci. Nat Genet 2010;42:978–84.

10. Easton DF, Pooley KA, Dunning AM, Pharoah PD, Thompson D, BallingerDG, et al. Genome-wide association study identifies novel breast cancersusceptibility loci. Nature 2007;447:1087–93.

11. Goode EL, Chenevix-Trench G, Song H, Ramus SJ, Notaridou M, Law-renson K, et al. A genome-wide association study identifies suscepti-bility loci for ovarian cancer at 2q31 and 8q24. Nat Genet 2010;42:874–9.

12. Shete S, Hosking FJ, Robertson LB, Dobbins SE, SansonM,Malmer B, et al.Genome-wide association study identifies five susceptibility loci for glio-ma. Nat Genet 2009;41:899–904.

13. Crowther-Swanepoel D, Broderick P, Di Bernardo MC, Dobbins SE, TorresM,MansouriM, et al. Common variants at 2q37. 3, 8q24. 21, 15q21. 3 and16q24. 1 influence chronic lymphocytic leukemia risk. Nat Genet 2010;42:132–6.

14. Ghoussaini M, Song H, Koessler T, Al Olama AA, Kote-Jarai Z, Driver KE,et al. Multiple loci with different cancer specificities within the 8q24 genedesert. J Natl Cancer Inst 2008;100:962–6.

15. McKay JD, Hung RJ, Gaborieau V, Boffetta P, Chabrier A, Byrnes G, et al.Lung cancer susceptibility locus at 5p15. 33. Nat Genet 2008;40:1404–6.

16. Wang Y, Broderick P, Webb E, Wu X, Vijayakrishnan J, Matakidou A, et al.Common 5p15. 33 and 6p21. 33 variants influence lung cancer risk. NatGenet 2008;40:1407–9.

17. Turnbull C, Rapley EA, Seal S, Pernet D, Renwick A, Hughes D, et al.Variants near DMRT1, TERT and ATF7IP are associatedwith testicular germcell cancer. Nat Genet 2010;42:604–7.

18. Kote-Jarai Z, Saunders EJ, Leongamornlert DA, Tymrakiewicz M, Dadaev T,Jugurnauth-Little S, et al. Fine-mapping identifies multiple prostate cancerrisk loci at 5p15, one of which associates with TERT expression. HumMolGenet 2013;22:2520–8.

19. Kote-Jarai Z, Al Olama AA, Giles GG, Severi G, Schleutker J, Weischer M,et al. Seven prostate cancer susceptibility loci identified by a multi-stagegenome-wide association study. Nat Genet 2011;43:785–91.

20. Rafnar T, Sulem P, Stacey SN, Geller F, Gudmundsson J, Sigurdsson A, et al.Sequence variants at the TERT-CLPTM1L locus associate with many cancertypes. Nat Genet 2009;41:221–7.

21. Haiman CA, Chen GK, Vachon CM, Canzian F, Dunning A, Millikan RC,et al. A common variant at the TERT-CLPTM1L locus is associated withestrogen receptor-negative breast cancer. Nat Genet 2011;43:1210–4.

22. Peters U, Hutter CM, Hsu L, Schumacher FR, Conti DV, Carlson CS, et al.Meta-analysis of new genome-wide association studies of colorectal cancerrisk. Human Genet 2012;131:217–34.

23. Solovieff N, Cotsapas C, Lee PH, Purcell SM, Smoller JW. Pleiotropy incomplex traits: challenges and strategies. Nat Rev Genet 2013;14:483–95.

24. Manolio TA.Bringing genome-wide association findings into clinical use.Nat Rev Genet 2013;14:549–58.

25. Setiawan VW, Schumacher F, Prescott J, Haessler J, Malinowski J, Went-zensen N, et al. Cross-cancer pleiotropic analysis of endometrial cancer:PAGE and E2C2 consortia. Carcinogenesis 2014;35:2068–73.






26. Cheng I, Kocarnik JM, Dumitrescu L, Lindor NM, Chang-Claude J, AveryCL, et al. Pleiotropic effects of genetic risk variants for other cancers oncolorectal cancer risk: PAGE, GECCO and CCFR consortia. Gut 2014;63:800–7.

27. Kinnersley B, Migliorini G, Broderick P, Whiffin N, Dobbins S, Casey G,et al. The TERT variant rs2736100 is associated with colorectal cancer risk.Br J Cancer 2012;107:1001–8.

28. Pierce BL, Ahsan H. Genome-wide "pleiotropy scan" identifies HNF1Aregion as a novel pancreatic cancer susceptibility locus. Cancer Res2011;71:4352–8.

29. Lee E, Stram DO, Ek WE, Onstad LE, MacGregor S, Gharahkhani P, et al.Pleiotropic analysis of cancer risk loci on esophageal adenocarcinoma risk.Cancer Epidemiol Prev Biomarkers 2015;24:1801–3.

30. Chang J, Wei L, Miao X, Yu D, TanW, Zhang X, et al. Two novel variants on13q22. 1 are associated with risk of esophageal squamous cell carcinoma.Cancer Epidemiol Prev Biomarkers 2015;24:1774–80.

31. Panagiotou OA, Travis RC, Campa D, Berndt SI, Lindstrom S, Kraft P, et al.A genome-wide pleiotropy scan for prostate cancer risk. Eur Urol2015;67:649–57.

32. Park SL, Fesinmeyer MD, Timofeeva M, Caberto CP, Kocarnik JM, Han Y,et al. Pleiotropic associations of risk variants identified for other cancerswith lung cancer risk: the PAGE and TRICL consortia. J Natl Cancer Inst2014;106:dju061.

33. LiD-K,Han J, Liu J-B, JinG-F,Qu J-W,ZhuM, et al.Genetic variants at 6p21.1 and 7p15. 3 Identified by GWASs of multiple cancers and ovarian cancerrisk: a case-control study in Han Chinese women. Asian Pac J Cancer Prev2014;15:123–7.

34. Du J, Xu Y, Dai J, Ren C, Zhu C, Dai N, et al. Genetic variants at 5p15 areassociated with risk and early onset of gastric cancer in Chinese popula-tions. Carcinogenesis 2013;34:2539–42.

35. Campa D, Barrdahl M, Tsilidis KK, Severi G, Diver WR, Siddiq A, et al. Agenome-wide "pleiotropy scan" does not identify new susceptibility locifor estrogen receptor negative breast cancer. PloS One 2014;9:e85955.

36. Lim U, Kocarnik JM, Bush WS, Matise TC, Caberto C, Park SL, et al.Pleiotropy of cancer susceptibility variants on the risk of non-Hodgkinlymphoma: the PAGE consortium. PLoS One 2014;9:e89791.

37. Cheng TH, Thompson D, Painter J, O'Mara T, Gorman M, Martin L, et al.Meta-analysis of genome-wide association studies identifies commonsusceptibility polymorphisms for colorectal and endometrial cancer nearSH2B3 and TSHZ1. Sci Rep 2015;5:17369.

38. Fehringer G, Kraft P, Pharoah PD, Eeles RA, Chatterjee N, Schumacher FR,et al. Cross-cancer genome-wide analysis of lung, ovary, breast, prostate,and colorectal cancer reveals novel pleiotropic associations. Cancer Res2016;76:5103–14.

39. Jin G, Ma H, Wu C, Dai J, Zhang R, Shi Y, et al. Genetic variants at 6p21. 1and7p15. 3 are associatedwith risk ofmultiple cancers inHanChinese. AmJ Hum Genet 2012;91:928–34.

40. LawPJ, SudA,Mitchell JS,HenrionM,OrlandoG, LeniveO, et al. Genome-wide association analysis of chronic lymphocytic leukaemia, Hodgkinlymphoma and multiple myeloma identifies pleiotropic risk loci. Sci Rep2017;7:41071.

41. Sampson JN, Wheeler WA, Yeager M, Panagiotou O, Wang Z, Berndt SI,et al. Analysis of heritability and shared heritability based on genome-wide association studies for 13 cancer types. J Natl Cancer Inst 2015;107:djv279.

42. Lindstr€omS, FinucaneH, Bulik-SullivanB, Schumacher FR, AmosCI,HungRJ, et al. Quantifying the genetic correlation betweenmultiple cancer types.Cancer Epidemiol Prev Biomarkers 2017;26:1427–35.

43. Sivakumaran S, Agakov F, Theodoratou E, Prendergast JG, Zgaga L, Man-olio T, et al. Abundant pleiotropy in human complex diseases and traits.Am J Hum Genet 2011;89:607–18.

44. Malone J, Holloway E, Adamusiak T, Kapushesky M, Zheng J, KolesnikovN, et al. Modeling sample variables with an Experimental Factor Ontology.Bioinformatics 2010;26:1112–8.

45. Whetzel PL, Noy NF, ShahNH, Alexander PR, Nyulas C, Tudorache T, et al.BioPortal: enhanced functionality via newWeb services from the NationalCenter for Biomedical Ontology to access and use ontologies in softwareapplications. Nucleic Acids Res 2011;39:W541–W5.

46. 1000 Genomes Project Consortium. A global reference for human geneticvariation. Nature 2015;526:68.

47. Sherry ST, Ward M-H, Kholodov M, Baker J, Phan L, Smigielski EM, et al.dbSNP: the NCBI database of genetic variation. Nucleic Acids Res 2001;29:308–11.

48. KentWJ, SugnetCW, Furey TS, RoskinKM,Pringle TH,Zahler AM, et al. Thehuman genome browser at UCSC. Genome Res 2002;12:996–1006.

49. Machiela MJ, Chanock SJ. LDlink: a web-based application for exploringpopulation-specific haplotype structure and linking correlated alleles ofpossible functional variants. Bioinformatics 2015;31:3555–7.

50. Ward LD, Kellis M. HaploReg: a resource for exploring chromatin states,conservation, and regulatory motif alterations within sets of geneticallylinked variants. Nucleic Acids Res 2012;40:D930–D4.

51. Ward LD, Kellis M. HaploReg v4: systematic mining of putative causalvariants, cell types, regulators and target genes for human complex traitsand disease. Nucleic Acids Res 2016;44:D877–D81.

52. McLaren W, Gil L, Hunt SE, Riat HS, Ritchie GR, Thormann A, et al. Theensembl variant effect predictor. Genome Biol 2016;17:122.

53. Huang DW, Sherman BT, Lempicki RA. Systematic and integrative analysisof large gene lists using DAVID bioinformatics resources. Nat Protoc2009;4:44–57.

54. Huang DW, Sherman BT, Lempicki RA. Bioinformatics enrichment tools:paths toward the comprehensive functional analysis of large gene lists.Nucleic Acids Res 2009;37:1–13.

55. Kamburov A, Stelzl U, Lehrach H, Herwig R. The ConsensusPathDBinteraction database: 2013 update. Nucleic Acids Res 2013;41:D793–D800.

56. Croft D, Mundo AF, Haw R, Milacic M,Weiser J, Wu G, et al. The Reactomepathway knowledgebase. Nucleic Acids Res 2014;42:D472–D7.

57. Fabregat A, Sidiropoulos K, Garapati P, Gillespie M, Hausmann K, Haw R,et al. The reactome pathway knowledgebase. Nucleic Acids Res 2016;44:D481–D7.

58. Thomas PD, Campbell MJ, Kejariwal A, Mi H, Karlak B, Daverman R, et al.PANTHER: a library of protein families and subfamilies indexed byfunction. Genome Res 2003;13:2129–41.

59. Chaplin T, Ayton P, Bernard OA, Saha V, Della Valle V, Hillion J, et al. Anovel class of zinc finger/leucine zipper genes identified from the molec-ular cloning of the t (10; 11) translocation in acute leukemia. Blood1995;85:1435–41.

60. Shvarts A, Steegenga W, Riteco N, Van Laar T, Dekker P, Bazuine M, et al.MDMX: a novel p53-binding protein with some functional properties ofMDM2. EMBO J 1996;15:5349.

61. Shvarts A, Bazuine M, Dekker P, Ramos YF, Steegenga WT, Merckx G, et al.Isolation and identification of the human homolog of a new p53-bindingprotein, Mdmx. Genomics 1997;43:34–42.

62. Eeles RA, Al Olama AA, Benlloch S, Saunders EJ, Leongamornlert DA,Tymrakiewicz M, et al. Identification of 23 new prostate cancer suscep-tibility loci using the iCOGS custom genotyping array. Nat Genet 2013;45:385–91.

63. Garcia-Closas M, Couch FJ, Lindstrom S, Michailidou K, Schmidt MK,Brook MN. Genome-wide association studies identify four ER negative–specific breast cancer risk loci. Nat Genet 2013;45:392.

64. Lonsdale J, Thomas J, Salvatore M, Phillips R, Lo E, Shad S, et al. Thegenotype-tissue expression (GTEx) project. Nat Genet 2013;45:580–5.

65. Liu M, Hsu J, Chan C, Li Z, Zhou Q. The ubiquitin ligase Siah1 controlsELL2 stability and formation of super elongation complexes to modulategene transcription. Mol Cell 2012;46:325–34.

66. Luo Z, Lin C, Shilatifard A. The super elongation complex (SEC) family intranscriptional control. Nat Rev Mol Cell Biol 2012;13:543–7.

67. Swaminathan B, Thorleifsson G, J€oud M, Ali M, Johnsson E, Ajore R, et al.Variants in ELL2 influencing immunoglobulin levels associate with mul-tiple myeloma. Nat Commun 2015;6:7213.

68. Campa D, Rizzato C, StolzenbergSolomon R, Pacetti P, Vodicka P, ClearySP, et al. TERT gene harbors multiple variants associated with pancreaticcancer susceptibility. Int J Cancer 2015;137:2175–83.

69. Fletcher O, Houlston RS. Architecture of inherited susceptibility to com-mon cancer. Nat Rev Cancer 2010;10:353–61.

70. Nakamura TM, Morin GB, Chapman KB, Weinrich SL, Andrews WH,Lingner J, et al. Telomerase catalytic subunit homologs from fission yeastand human. Science 1997;277:955–9.

71. Feng J, Funk WD, Wang S-S, Weinrich SL. The RNA component of humantelomerase. Science 1995;269:1236.

Wu et al.





72. Chen H, Chen Y, Zhao Y, Fan W, Zhou K, Liu Y, et al. Association ofsequence variants on chromosomes 20, 11, and 5 (20q13. 33, 11q23. 3,and 5p15. 33) with glioma susceptibility in a Chinese population. Am JEpidemiol 2011;173:915–22.

73. Kinnersley B, Migliorini G, Broderick P, Whiffin N, Dobbins S, Casey G,et al. The TERT variant rs2736100 is associated with colorectal cancer risk.Br J Cancer 2012;107:1001.

74. Cavanagh H, Rogers KM. The role of BRCA1 and BRCA2 mutations inprostate, pancreatic and stomach cancers. Hered Cancer Clin Pract 2015;13:16.

75. Antoniou A, Pharoah P, Narod S, Risch HA, Eyfjord JE, Hopper J,et al. Average risks of breast and ovarian cancer associated withBRCA1 or BRCA2 mutations detected in case series unselected forfamily history: a combined analysis of 22 studies. Am J Hum Genet2003;72:1117–30.

76. King M-C, Marks JH, Mandell JB. Breast and ovarian cancer risksdue to inherited mutations in BRCA1 and BRCA2. Science 2003;302:643–6.

77. Yoshida K, Miki Y. Role of BRCA1 and BRCA2 as regulators of DNA repair,transcription, and cell cycle in response to DNA damage. Cancer Sci2004;95:866–71.

78. Helleday T, Lo J, van Gent DC, Engelward BP. DNA double-strand breakrepair: from mechanistic understanding to cancer treatment. DNA Repair2007;6:923–35.

79. Smoller JW, Lunetta KL, Robins J. Implications of comorbidity and ascer-tainment bias for identifying disease genes. Am J Med Genet Part A2000;96:817–22.

80. Czabotar PE, Lessene G, Strasser A, Adams JM. Control of apoptosis by theBCL-2 protein family: implications for physiology and therapy. Nat RevMol Cell Biol 2014;15:49.

81. Shamas-Din A, Brahmbhatt H, Leber B, Andrews DW. BH3-only proteins:Orchestrators of apoptosis. Biochimica Et Biophysica Acta 2011;1813:508–20.

82. Souers AJ, Leverson JD, Boghaert ER, Ackler SL, Catron ND, Chen J, et al.ABT-199, a potent and selective BCL-2 inhibitor, achieves antitumoractivity while sparing platelets. Nat Med 2013;19:202–8.

83. Stilgenbauer S, Eichhorst B, Schetelig J, Coutre S, Seymour JF,Munir T, et al.Venetoclax in relapsed or refractory chronic lymphocytic leukaemia with17p deletion: a multicentre, open-label, phase 2 study. Lancet Oncol2016;17:768–78.

84. Gerecitano JF, Roberts AW, Seymour JF, Wierda WG, Kahl BS, Pagel JM,et al. A phase 1 study of venetoclax (ABT-199/GDC-0199) monotherapyin patients with relapsed/refractory non-Hodgkin lymphoma. Blood2015;126:254.

85. Bhattacharyya S, Balakathiresan NS, Dalgard C, Gutti U, Armistead D,Jozwik C, et al. Elevated miR-155 promotes inflammation in cystic fibrosisby driving hyperexpression of interleukin-8. J Biol Chem 2011;286:11604–15.

86. Huang Q, Zhang X-W, Ma Y-S, Lu G-X, Xie R-T, Yang H-Q, et al. Up-regulated microRNA-299 corrected with poor prognosis of glioblastomamultiforme patients by targeting ELL2. Jp J Clin Oncol 2017;47:590–6.

87. Rupaimoole R, Slack FJ. MicroRNA therapeutics: towards a new era for themanagement of cancer and other diseases. Nat Rev Drug Discov 2017;16:203–22.






2018;27:75-85. Published OnlineFirst November 17, 2017.Cancer Epidemiol Biomarkers Prev Yi-Hsuan Wu, Rebecca E. Graff, Michael N. Passarelli, et al. CharacteristicsGenome-Wide Association Studies Reveals Functional Identification of Pleiotropic Cancer Susceptibility Variants from

Updated version

10.1158/1055-9965.EPI-17-0516doi:

Access the most recent version of this article at:

Material

Supplementary

http://cebp.aacrjournals.org/content/suppl/2017/11/17/1055-9965.EPI-17-0516.DC1

Access the most recent supplemental material at:

Cited articles

http://cebp.aacrjournals.org/content/27/1/75.full#ref-list-1

This article cites 87 articles, 10 of which you can access for free at:

Citing articles

http://cebp.aacrjournals.org/content/27/1/75.full#related-urls

This article has been cited by 10 HighWire-hosted articles. Access the articles at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected]

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cebp.aacrjournals.org/content/27/1/75To request permission to re-use all or part of this article, use this link



http://cebp.aacrjournals.org/lookup/doi/10.1158/1055-9965.EPI-17-0516

http://cebp.aacrjournals.org/content/suppl/2017/11/17/1055-9965.EPI-17-0516.DC1

http://cebp.aacrjournals.org/content/27/1/75.full#ref-list-1

http://cebp.aacrjournals.org/content/27/1/75.full#related-urls

http://cebp.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://cebp.aacrjournals.org/content/27/1/75


Documents

Cancer Identiﬁcation of Pleiotropic Cancer Susceptibility ... · cancer risk, comprised of 989 unique variants associated with 27 unique cancer sites. We found 20 pleiotropic variant