Functional Polymorphisms in the TERT Promoter Are Associated with Risk of Serous Epithelial Ovarian and Breast Cancers

Abstract

Genetic variation at the TERT-CLPTM1L locus at 5p15.33 is associated with susceptibility to several cancers, including epithelial ovarian cancer (EOC). We have carried out fine-mapping of this region in EOC which implicates an association with a single nucleotide polymorphism (SNP) within the TERT promoter. We demonstrate that the minor alleles at rs2736109, and at an additional TERT promoter SNP, rs2736108, are associated with decreased breast cancer risk, and that the combination of both SNPs substantially reduces TERT promoter activity.

Full text

Functional Polymorphisms in the
TERT
Promoter Are
Associated with Risk of Serous Epithelial Ovarian and
Breast Cancers
Jonathan Beesley
1
*
.
, Hilda A. Pickett
2,3.
, Sharon E. Johnatty
1
, Alison M. Dunning
4
, Xiaoqing Chen
1
, Jun
Li
1
, Kyriaki Michailidou
4
,YiLu
1
, David N. Rider
5
, Rachel T. Palmieri
6
, Michael D. Stutz
2,3
, Diether
Lambrechts
7,8
, Evelyn Despierre
9
, Sandrina Lambrechts
9
, Ignace Vergote
9
, Jenny Chang-Claude
10
,
Stefan Nickels
10
, Alina Vrieling
10
, Dieter Flesch-Janys
11
, Shan Wang-Gohrke
12
, Ursula Eilber
10
, Natalia
Bogdanova
13,14,15
, Natalia Antonenkova
14
, Ingo B. Runnebaum
16
, Thilo Do
¨rk
13
, Marc T. Goodman
17
,
Galina Lurie
17
, Lynne R. Wilkens
17
, Rayna K. Matsuno
17
, Lambertus A. Kiemeney
18,19,20
, Katja K. H.
Aben
18,20
, Tamara Marees
18
, Leon F. A. G. Massuger
21
, Brooke L. Fridley
5
, Robert A. Vierkant
5
, Elisa V.
Bandera
22
, Sara H. Olson
23
, Irene Orlow
23
, Lorna Rodriguez-Rodriguez
22
, Linda S. Cook
24
, Nhu D. Le
24
,
Angela Brooks-Wilson
24
, Linda E. Kelemen
24
, Ian Campbell
25
, Simon A. Gayther
26
, Susan J. Ramus
26
,
Aleksandra Gentry-Maharaj
27
, Usha Menon
27
, Shahana Ahmed
4
, Caroline Baynes
4
, Paul D. Pharoah
4
,
kConFab Investigators
25
, Australian Ovarian Cancer Study Group
1,25,28
, ABCTB Investigators, Kenneth
Muir
30
, Artitaya Lophatananon
30
, Arkom Chaiwerawattana
31
, Surapon Wiangnon
32
, Stuart Macgregor
1
,
Douglas F. Easton
4
, Roger R. Reddel
2,3
, Ellen L. Goode
5
, Georgia Chenevix-Trench
1
on behalf of the
Ovarian Cancer Association Consortium
1Division of Genetics and Population Health, Queensland Institute of Medical Research, Brisbane, Queensland, Australia, 2Cancer Research Unit, Children’s Medical
Research Institute, Westmead, New South Wales, Australia, 3Sydney Medical School, University of Sydney, Sydney, New South Wales, Australia, 4Department of Oncology,
University of Cambridge, Strangeways Research Laboratory, Cambridge, United Kingdom, 5Department of Health Sciences Research, Mayo Clinic College of Medicine,
Rochester, Minnesota, United States of America, 6Department of Community and Family Medicine, Duke University Medical Center, Durham, North Carolina, United
States of America, 7Vesalius Research Center, VIB, Leuven, Belgium, 8Vesalius Research Center, University of Leuven, Leuven, Belgium, 9Division of Gynaecologic
Oncology, Department of Obstetrics and Gynaecology, University Hospitals Leuven, University of Leuven, Leuven, Belgium, 10 Division of Cancer Epidemiology, German
Cancer Research Center (DKFZ), Heidelberg, Germany, 11 Institute for Medical Biometrics and Epidemiology, University Clinic Hamburg-Eppendorf, Hamburg, Germany,
12 Department of Obstetrics and Gynecology, University of Ulm, Ulm, Germany, 13 Clinics of Obstetrics and Gynaecology, Hannover Medical School, Hannover, Germany,
14 Byelorussian Institute for Oncology and Medical Radiology Aleksandrov N.N., Minsk, Belarus, 15 Clinics of Radiation Oncology, Hannover Medical School, Hannover,
Germany, 16 Clinics of Obstetrics and Gynaecology, Friedrich Schiller University, Jena, Germany, 17 Cancer Epidemiology Program, University of Hawaii Cancer Center,
Honolulu, Hawaii, United States of America, 18 Department of Epidemiology, Biostatistics and HTA, Radboud University Nijmegen Medical Centre, Nijmegen, The
Netherlands, 19 Department of Urology, Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands, 20 Comprehensive Cancer Center, The Netherlands,
Nijmegen, The Netherlands, 21 Department of Gynaecology, Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands, 22 The Cancer Institute of New
Jersey/Robert Wood Johnson Medical School, New Brunswick, New Jersey, United States of America, 23 Department of Epidemiology & Biostatistics, Memorial Sloan
Kettering Cancer Center, New York, New York, United States of America, 24 Department of Population Health Research, Alberta Health Services-Cancer Care, Calgary,
Alberta, Canada, 25 Peter MacCallum Cancer Centre, East Melbourne, Victoria, Australia, 26 Department of Preventive Medicine, Keck School of Medicine, University of
Southern California, Los Angeles, California, United States of America, 27 Department of Gynaecological Oncology, University College London, EGA Institute for Women’s
Health, London, United Kingdom, 28 Department of Gynaecological Oncology and Westmead Institute for Cancer Research, University of Sydney at the Westmead
Millennium Institute, Westmead Hospital, Sydney, New South Wales, Australia, 29 Westmead Millennium Institute, Sydney Medical School, Westmead, The University of
Sydney, Sydney, New South Wales, Australia, 30 Health Sciences Research Institute, Warwick Medical School, University of Warwick, Coventry, United Kingdom,
31 Department of Academic Support, The National Cancer Institute of Thailand, Ministry of Public Health, Nonthaburi, Thailand, 32 Department of Pediatrics, Medical
School, Khon Kaen University, Khon Kaen, Thailand
Abstract
Genetic variation at the TERT-CLPTM1L locus at 5p15.33 is associated with susceptibility to several cancers, including
epithelial ovarian cancer (EOC). We have carried out fine-mapping of this region in EOC which implicates an association with
a single nucleotide polymorphism (SNP) within the TERT promoter. We demonstrate that the minor alleles at rs2736109, and
at an additional TERT promoter SNP, rs2736108, are associated with decreased breast cancer risk, and that the combination
of both SNPs substantially reduces TERT promoter activity.
PLoS ONE | www.plosone.org 1 September 2011 | Volume 6 | Issue 9 | e24987

Citation: Beesley J, Pickett HA, Johnatty SE, Dunning AM, Chen X, et al. (2011) Functional Polymorphisms in the TERT Promoter Are Associated with Risk of Serous
Epithelial Ovarian and Breast Cancers. PLoS ONE 6(9): e24987. doi:10.1371/journal.pone.0024987
Editor: Lin Zhang, University of Pennsylvania School of Medicine, United States of America
Received August 12, 2011; Accepted August 21, 2011; Published September 15, 2011
This is an open-access article distributed under the terms of the Creative Commons Public Domain declaration which stipulates that, once placed in the public
domain, this work may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose.
Funding: The Ovarian Cancer Association Consortium is supported by a grant from the Ovarian Cancer Research Fund. kConFaB is supported by the National
Breast Cancer Foundation, the National Health and Medical Research Council of Australia, and the Cancer Councils of Queensland, New South Wales, Western
Australia, South Australia, and Victoria. The Clinical Follow-up Study of kConFaB is funded by the National Health and Medical Research Council (NHMRC) of
Australia (145684 and 288704). AOCS was supported by the U.S. Army Medical Research and Materiel Command (DAMD17-01-1-0729), the National Health and
Medical Research Council of Australia (199600), and the Cancer Council Tasmania and Cancer Foundation of Western Australia. Tissues and samples were received
from the Australian Breast Cancer Tissue Bank which is supported by the National Health and Medical Research Council of Australia, The Cancer Institute NSW and
the National Breast Cancer Foundation. The tissues and samples are made available to researchers on a non-exclusive basis. The BEL study was supported by the
Nationaal Kankerplan – actie 29. The HAW study is supported by the U.S. National Institutes of Health (R01 CA58598, N01-CN-55424, N01-PC-67001). SEARCH was
supported by Cancer Research-UK grants [C1287/A10118, C490/A11021, C8197/A10123]. The MAYO study was supported by R01 CA122443 and P50 CA136939.
The MARIE study was supported by the Deutsche Krebshilfe e.V., grant number 70-2892-BR I, the Hamburg Cancer Society, the German Cancer Research Center
(DKFZ) and the DNA extraction and genotype work in part by the Federal Ministry of Education and Research (BMBF) Germany grant 01KH0402. The GESBC
epidemiological study was supported by the Deutsche Krebshilfe e. V. [70492] and GESBC genotyping in part by the state of Baden-Wu
¨rttemberg through the
Medical Faculty of the University of Ulm [P.685]. The OVAL-BC Study is supported by the British Columbia Workers Compensation Board and the Canadian
Institutes for Health Research. The NJO study was funded by NIH-K07 CA095666, R01CA83918, and The Cancer Institute of New Jersey. The UKO study is
supported by funding from Cancer Research UK, the Eve Appeal, and the OAK Foundation; some of this work was undertaken at UCLH/UCL which received some
funding from the Department of Health’s NIHR Biomedical Research Centre funding scheme. ACP study is funded by The Breast Cancer Research Trust, UK. JB,
HAP, SEJ, MDS and RRR are supported by NHMRC project grant 1012023. DFE is a Principal Research Fellow of Cancer Research UK. SM is supported by an NHMRC
career development award. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: jonathan.beesley@qimr.edu.au
.These authors contributed equally to this work.
Introduction
Genome-wide association studies (GWAS) have identified more
than 140 cancer susceptibility loci for 17 different cancers (www.
genome.gov/gwastudies), including a locus at 5p15.33 which has
been implicated in susceptibility to melanoma [1,2], glioma [3],
lung [4,5], pancreas [6], prostate [2], testicular [7], and bladder
cancers [8]. This locus harbours TERT, encoding the reverse
transcriptase component of telomerase, and cleft lip and palate
transmembrane protein 1-like (CLPTM1L). GWAS of ovarian and
breast cancer have not detected an association with this locus to
date, which may be due to poor tagging of this region on the chips
employed, or the lack of statistical power to detect associations.
Using a candidate gene approach we previously reported evidence
of an association between an intronic SNP in TERT (rs7726159)
and EOC risk, particularly of the serous histological subtype, using
a per-allele model [9]. In this study we carried out fine-mapping,
followed by functional analyses of associated SNPs identified
within the promoter region of TERT. We directly demonstrate
that the presence of a common haplotype, which is associated with
decreased cancer risk, substantially reduces TERT promoter
activity. The vast majority of cancers depend on expression of
telomerase, which requires substantial upregulation of TERT
expression, for their continued proliferation, strongly implicating
TERT as the predominant gene involved in this association.
Results and Discussion
To further clarify the association previously reported with
TERT, we employed a fine-mapping strategy in nine case-control
studies from the Ovarian Cancer Association Consortium (OCAC)
(Table S1). SNPs in a region spanning 6250 kb across the TERT,
CLPT1ML,SLC6A18 and SLC6A19 genes that were correlated
(0.2#r
2
#0.99) with rs7726159, rs11133719, rs2735940 and
rs2736100 [9] were selected from the 1000 Genomes low coverage
pilot release of April 2009. The SNPs implicated by our previous
study of EOC [9] (rs11133719, rs7726159, rs2736100 and
rs2735940) or by cancer GWAS (rs2736100) were also included
in the panel. We genotyped 36 SNPs by iPLEX (Sequenom Inc.)
in 2,130 invasive EOC cases and 3,975 controls, all of Caucasian
ancestry. After excluding monomorphic loci (n = 6) and two SNPs
that failed OCAC’s quality control criteria [10], 28 SNPs were
analysed for association with risk of EOC. We used single marker
and stepwise logistic regression models, adjusted for study and age
(at interview for controls and at diagnosis for cases) with a
threshold of P#0.05 for addition (forward stepwise) or removal
(backward stepwise) of SNPs (Table S2). We found that
rs2736109 showed the strongest association with serous EOC
(adjusted OR
per-allele
0.86 (0.77–0.96), P= 0.005) (Table S3), but
not with invasive EOC risk overall (adjusted OR
per-allele
0.96
(0.89–1.05), P= 0.38) (data not shown). Likelihood ratio tests
comparing logistic regression models with and without a genotype-
by-study interaction term revealed no significant study heteroge-
neity (P= 0.4). rs2736109 is in a region of low linkage
disequilibrium that encompasses the 59end of the TERT gene
and the TERT promoter (Figure 1). This region also contains the
SNPs, rs2736108 and rs2853669, which have pairwise correlations
(r
2
) with each other, and with rs2736109 of greater than 0.6. It has
previously been reported that rs2853669 is associated with breast
cancer risk [11].
Since other loci confer susceptibility to both EOC and breast
cancer [12,13] we investigated associations of TERT SNPs with
breast cancer risk in various data sets. First, we genotyped
rs2736109 by iPLEX in 4,277 invasive breast cancer cases and
7,000 controls from Australian, German and Thai studies from the
Breast Cancer Association Consortium (BCAC) (Table S4). In the
combined analysis of the Australian and German studies, there
was no association with invasive breast cancer risk overall
(adjusted OR
per-allele
0.95 (0.90–1.01), P= 0.10), but there was
an association in cases $50 years at diagnosis that approached
significance (adjusted OR
per-allele
0.94 (0.88–1.00), P= 0.049). We
also found a stronger association for ER-negative tumours
(adjusted OR
per-allele
0.88 (0.80–0.98), P= 0.022; for ER-positive
tumours adjusted OR
per-allele
0.98 (0.92–1.05), P= 0.562) (Tables
S5, S6, and S7). We found the strongest evidence of association
among ER-negative cases over the age of 50 (n = 636) (adjusted
OR
per-allele
0.84 (0.75–0.95), P= 0.005). We did not find any
TERT SNPs and Ovarian and Breast Cancer Risk
PLoS ONE | www.plosone.org 2 September 2011 | Volume 6 | Issue 9 | e24987

association with breast cancer risk in a Thai study (n = 327 cases)
genotyped for rs2736109.
Next, we analysed genotypes imputed using MaCH for
rs2736109 (r
2
= 0.45) from a breast cancer GWAS of 3,931 cases
and 3,622 controls from the United Kingdom, from which neither
age nor ER status were available [14]. We observed a significant
association between rs2736109 and overall breast cancer risk in
this population (OR
per-allele
0.91 (0.83–0.99), P= 0.037), and from
a weighted meta-analysis of imputed and genotyped data from all
studies (OR
per-allele
0.94 (0.89–0.98), P= 0.011).
In an additional replication sample set (SEARCH) we
genotyped a correlated SNP, rs2736108 (r
2
= 0.96, 1000 Genomes
Project, Dec 2009, r
2
= 0.63 based on 345 Australian controls) in
6,788 cases and 6,426 controls because rs2736109 was not
amenable to genotyping by TaqMan (the genotyping platform
used by this group). We observed a significant association with
breast cancer risk overall (adjusted OR
per-allele
0.92 (0.87–0.97),
P= 0.003), which was restricted to the subset of cases diagnosed at
age 50 or older (adjusted OR
per-allele
0.91 (0.85–0.97), P= 0.002)
(Table S6). Estimates according to ER status showed a significant
association between the rs2736108 genotypes and ER-positive
tumours (adjusted OR
per-allele
0.93 (0.88–0.99), P= 0.031); there
was no significant association in ER-negative tumours but the
estimated OR was similar (OR
per-allele
0.95 (0.85–1.06)) and the
sample size much smaller. Comparison of models for both
rs2736109 and rs2736108 with and without genotype by age
group interaction terms showed no evidence of a statistical
interaction on the multiplicative scale (P
interaction
$0.25).
The SNPs of interest (rs2736108, rs2736109 and rs2853669) lie
within the upstream promoter region of TERT. To determine the
functional significance of these sites, we generated combinations of
these variants in a luciferase reporter construct containing 3.9 kb
of the TERT promoter [15]. Relative promoter activity was
determined in an EOC cell line (27/87), a breast adenocarcinoma
cell line (MDA-MB-468), and in post-selection normal breast
epithelial cells (Bre16) (Figure 2). Introduction of a mutation into
the human estrogen-responsive element in the TERT promoter
(TERT-ERE) [16] was used as a positive control and confirmed
diminished reporter activity. In all three cell types, luciferase
activity was substantially reduced for the construct carrying both
minor (A) alleles at rs2736108 and rs2736109, but remained
unaltered for those with the individual minor allele at either SNP.
We observed no change in expression for the minor allele at
rs2853669.
Our analysis of Australian controls estimated a frequency of 32%
for the A-A haplotype at rs2736108 and rs2736109, suggesting that
this relatively common promoter haplotype may lower the risk of
ovarian and breast cancer through decreasing TERT expression. This
finding provides no support for the hypothesis that decreased
telomerase activity predisposes to genomic instability and consequent
oncogenic progression but instead our data imply the opposite,
namely, that decreased TERT expression confers decreased cancer risk.
Figure 1.
TERT-CLPTM1L
locus SNPs genotyped in epithelial ovarian cancer cases and controls. Panel Adepicts the pattern of linkage
disequilibrium using data from the HapMap phase II+III CEU population (2010), where white represents r
2
= 0, and black r
2
= 1. The plot in B
represents 2log
10
pagainst the chromosomal position for genotyped SNPs in the ovarian cancer fine mapping panel. The purple diamond represents
the SNP (rs2736109) with the strongest observed association. The point colours represent the strength of LD according to 1000 Genomes Data.
doi:10.1371/journal.pone.0024987.g001
TERT SNPs and Ovarian and Breast Cancer Risk
PLoS ONE | www.plosone.org 3 September 2011 | Volume 6 | Issue 9 | e24987

We also examined tumour expression and germline variants at
the TERT-CLPTM1L locus using data from The Cancer Genome
Atlas (TCGA) ovarian serous cystadenocarcinoma set. mRNA
expression profiling was available for 574 tumour samples using
Affymetrix U133A platform This assay has one probe for TERT,
but none for CLPTM1L.TERT was expressed at low levels, as
expected, but this does not necessarily reflect low levels of
telomerase protein expression or activity [17]. 508 normal DNA
samples were genotyped using Illumina 1 M array, and the SNP
rs2736108 was typed on this array. We also imputed the TCGA
samples with reference to the 1000 Genome data (June 2010
release); rs2853669 was successfully imputed, but rs2736109 failed
quality control. There was no evidence suggesting that the
genotyped SNP, rs2736108, or the imputed SNP, rs2853669,
are associated with expression of TERT. However one SNP,
rs2735845, in partial linkage disequilibrium with the typed SNP,
rs2736108 (r
2
= 0.306), was associated with significantly altered
transcript abundance of TERT (P=4610
24
and P= 0.0029 after
correction for multiple testing). Gene copy number aberration was
found in TERT; it was amplified in ,20% of tumour samples.
This SNP showed robust association after adjusting for the copy
number variation (P= 1.8610
23
and P= 0.015 after correction for
multiple testing), and it explained 1.85% of variance of TERT
transcription in ovarian cancer tumours.
A recently published meta-analysis identified a significantly
decreased risk of breast cancer with a TERT SNP, rs2853669 (OR
0?76 (0?64–0?91), P=0?002) [11]. This SNP lies in the same
linkage disequilibrium block as rs2736108 and rs2736109 (r
2
.0.6
with rs2736108); therefore, this meta-analysis provides additional
support for an association between TERT promoter SNPs and
breast cancer risk. However, our functional analyses, which were
carried out in normal breast epithelial cells and a breast
adenocarcinoma cell line did not indicate any change to TERT
expression with this individual SNP.
We computed a gene-based test [18] of association at TERT
and CLPTM1L which yielded evidence in GWAS data available
from dbGAP (http://www.ncbi.nlm.nih.gov/gap) for association
with risk of lung, prostate and pancreatic cancer, but not breast
cancer overall. Combining data for all cancers in a cross-cancer
meta-analysis, revealed a genome-wide significant gene-based
P= 4.1610
27
for CLPTM1L (P= 0.008 after correction for 19,000
genes). A similar result was obtained for TERT (all cancer
P= 7.7610
25
). This gene-based test includes all SNPs within
50 kb of the start/stop site of each gene. Most of the associated
SNPs lie in the interval between TERT and CLPTM1L but with
slightly more evidence for SNPs near to CLPTM1L, leading to a
slightly higher gene-based Pvalue for CLPTM1L. These results are
based on marker data from Illumina GWAS arrays and hence the
exact location of the maximum association test statistic is
dependent to some degree upon the arbitrary set of SNPs that
are on the arrays. Clearly, additional analysis of the entire TERT-
Figure 2.
TERT
promoter activity. Luciferase reporter assays
following transient transfection with pGL2-control (SV40 promoter
and enhancer), pGL2-basic (lacks promoter and enhancer), transfection
control (blank), and the TERT reporter vectors TERT WT (3.9 kb of TERT
promoter), the positive control TERT-ERE (containing a mutation in the
estrogen-responsive element), and the minor alleles of rs2736108 (A
allele), rs2736109 (A allele), rs2736108/9 (A alleles at both sites),
rs2853669 (C allele) in (A) an EOC cell line (27/87), (B) a breast
adenocarcinoma cell line (MDA-MB-468) and (C) a normal breast
epithelial cell strain (Bre16). Error bars represent standard deviation
between three separate experiments. * represents statistical signifi-
cance using one-way ANOVA with post hoc Dunnett’s tests.
doi:10.1371/journal.pone.0024987.g002
TERT SNPs and Ovarian and Breast Cancer Risk
PLoS ONE | www.plosone.org 4 September 2011 | Volume 6 | Issue 9 | e24987

CLPM1L locus is warranted in breast, and other, cancers to
validate the associations we have identified and fine map putative
causal variants if our findings are confirmed.
The failure to date to identify an association between the
TERT-CLPTM1L locus and risk of EOC or breast cancer by
GWAS may be explained by the pattern of linkage disequilibrium
of the relevant SNPs: rs2736100, the tagSNP most commonly
identified by GWAS of other cancers is poorly correlated with
rs2736108 and rs2736109 (r
2
= 0.141 and 0.105, respectively), and
neither of these SNPs are on the Illumina 300 K, 610 K or 650 K
arrays used for most cancer GWAS.
In summary, we have demonstrated a direct association
between functional SNPs in the TERT promoter, which confer
decreased risk of ovarian and breast cancer, and reduced TERT
promoter activity. Decreased levels of TERT result in progressive
telomere shortening and the onset of cellular senescence, which
ultimately acts to suppress tumorigenesis. The association of
hypomorphic sequence variants in the TERT promoter with
decreased risk of cancer implicates downregulation of telomerase
and telomere shortening as an intrinsic tumour suppressive
mechanism. It is also possible that TERT variants associated with
elevated cancer risk may alter the stringency with which TERT is
regulated, potentially facilitating TERT activation and conse-
quently providing a tumorigenic advantage. Potential non-
canonical roles of TERT in cell signalling pathways may also
underlie cancer risk [19]. Our results add functional insight into
the increasingly important role of TERT as a cancer risk factor and
demonstrate the need for further mechanistic analysis of this multi-
cancer susceptibility locus.
Materials and Methods
Ethics Statement
Approval for this study was obtained from The Queensland
Institute of Medical Research Human Research Ethics Commit-
tee. All studies were approved by the review boards and ethics
committees of their respective institutions, and all participants
provided written informed consent.
Genotyping
iPLEX genotyping was carried out using MALDI-TOF
spectroscopy utilising Sequenom’s MassARRAY platform and
iPLEX GOLD chemistry. 10 ng of genomic DNA was used as
template, to which a PCR mix containing Qiagen Hot-StarTaq
was added. Shrimp alkaline phosphatase and primer extension
steps were carried out using Sequenom’s protocol and reagents.
Primers were obtained from Integrated DNA Technologies (Ohio
USA). Assays were designed with MassARRAY Assay Design
version 3.1(Sequenom). Raw genotype data were visualised and
processed with MassARRAY Typer software version 3.4. TaqMan
genotyping (SEARCH) was carried out as previously described
[20]. Strict quality control criteria were adhered to as part of the
Ovarian Cancer Association Consortium’s guidelines [10].
Statistical methods
We used single marker and stepwise logistic regression models to
screen 28 SNPs in non-Hispanic white ovarian cancer cases
(n = 2,130) and controls (n = 3,975) from nine OCAC studies
(Supplementary Table 1). Genotype data for all the previously
reported TERT SNPs [9] has been excluded from the current
analysis. Stepwise models were adjusted for study and age (at
interview for controls and at diagnosis for cases), with a threshold
of P#0.05 for addition (forward stepwise) or removal (backward
stepwise) of SNPs. All single marker risk estimates were obtained
from unconditional logistic regression models adjusted for age
(where available) and additionally for study where data was pooled
across multiple studies. Assuming a log additive model of
inheritance, the per-allele odds ratios (ORs) and their 95%
confidence intervals (CIs) associated for selected SNPs were
estimated by fitting the number of rare alleles carried as a
continuous covariate. Ovarian cancer risk associated with SNP
genotypes were obtained for all invasive cases as well as a subset of
serous cases. Breast cancer risk estimates were obtained for
invasive cases and by estrogen receptor (ER) status. Separate
comparisons were made for cases diagnosed before 50 vs. $50
years of age to explore effect modification by advancing age of
diagnosis. Summary estimates from pooled analyses using
genotyped and imputed SNP data were obtained from weighted
meta-analysis of study-specific parameter estimates (bcoefficients
and Standard Error). The minor allele frequency (MAF) for each
SNP was estimated from the control population for each study.
Study heterogeneity and risk differences associated with age
groups (,50 vs. $50) were assessed using the likelihood ratio test
to compare logistic regression models with and without a
multiplicative interaction term. All tests for association were two-
tailed, statistical significance was assessed at P#0.05, and were
performed in STATA SE v.11 (StataCorp, USA), and SAS v. 9.1.
Tests for study heterogeneity and age group interaction tests were
implemented in the R project for Statistical Computing (http://
www.r-project.org/).
VEGAS applied to dbGAP data
To evaluate evidence for association at TERT and CLPTM1L
with various cancers, we applied the gene-based test implemented
in VEGAS (all SNPs in gene test) [18] to data from dbGAP. In
brief, we selected cancer cases and controls from dbGAP that were
genotyped on ,550,000 SNPs (Illumina 610 quad or Illumina
HumanHap550 arrays). Studies were CGEMS breast (1145 cases,
1142 controls), CGEMS pancreatic cancer (2328 cases, 2351
controls), GENEVA lung cancer (EAGLE and PLCO combined
2748 cases, 2840 controls) and CGEMS prostate (1145 cases, 1054
controls). For full details see dbGAP website (http://www.ncbi.
nlm.nih.gov/gap). In each case, following standard quality control,
the genomic control lambda was as would be expected if cases and
controls were well matched; breast, pancreatic, lung, prostate
lambda values 1.01, 1.01, 1.03, 1.03, respectively. The VEGAS
gene based Pvalues from each of the four dbGAP studies were
combined in a meta-analysis using Fisher’s method for combining
Pvalues.
Luciferase assays
Variants were introduced into pGL3-hTERT-3915 [15] by
site-directed mutagenesis (Agilent Technologies). The A-A vector
was generated by introducing the variants sequentially. Cells were
transfected using siPORT NeoFX Transfection Agent (Ambion),
according to the manufacturer’s instructions, and harvested after
48 h. Cells were washed with phosphate buffered saline (PBS)
and lysed with 200 mL/well lysis buffer (25 mM tris pH 7.8,
2 mM EDTA, 10% glycerol, 1% Triton X-100, 0.2 mM DTT).
Luciferase activity was assayed in triplicate for each transfection
using 20 mL lysate and 50 mL reconstituted luciferase assay
reagent (Promega). Luminescence was measured immediately for
4 s in each well with a Wallac Victor
3
1420 multilabel counter.
The experiment was repeated three times and the results
averaged. Data were analysed by one-way ANOVA with post
hoc Dunnett’s tests in GraphPad Prism version 5.03 (GraphPad
Software).
TERT SNPs and Ovarian and Breast Cancer Risk
PLoS ONE | www.plosone.org 5 September 2011 | Volume 6 | Issue 9 | e24987

Supporting Information
Table S1 Participating EOC case-control studies.
(DOC)
Table S2 Per allele OR for all SNPs in EOC.
(DOC)
Table S3 The association of rs2736109 with EOC, by study.
(DOC)
Table S4 Participating invasive breast case-control studies.
(DOC)
Table S5 The association of rs2736109/rs2736108 with risk of
invasive breast cancer, by study.
(DOC)
Table S6 Breast cancer risk by age.
(DOC)
Table S7 Breast cancer risk by ER status.
(DOC)
Acknowledgments
We are grateful to the family and friends of Kathryn Sladek Smith for their
generous support of Ovarian Cancer Association Consortium through their
donations to the Ovarian Cancer Research Fund. MARIE thanks Tracy
Slanger and Elke Mutschelknauss for their valuable contributions, and S.
Behrens, R. Birr, W. Busch, U. Eilber, B. Kaspereit, N. Knese, K. Smit, for
their excellent technical assistance. GESBC thank Ursula Eilber and Tanya
Koehler for competent technical assistance. HJO and HMO acknowledge
the support of Matthias Du¨rst, Peter Schu¨ rmann and Peter Hillemanns.
UKO thanks Ian Jacobs, Eva Wozniak, Andy Ryan, Jeremy Ford and
Nayala Balogun for their contribution to the study.
Author Contributions
Conceived and designed the experiments: GC-T RRR ELG DNR.
Performed the experiments: JB HAP XC MDS SA CB. Analyzed the data:
JB HAP SEJ JL KM SM YL. Contributed reagents/materials/analysis
tools: RTP DL ED SL IV JC-C SN AV DF-J MTG GL LRW RKM NB
NA IBR TD BLF RAV ELG EVB SHO IO LR-R LEK NDL LSC AB-W
IC SAG SJR AG-M UM PDP AMD SW-G UE KM AL AC SW DFE
RRR GC-T LAK KKHA TM LFAGM. Wrote the paper: JB HAP GC-T.
References
1. Baird DM (2010) Variation at the TERT locus and predisposition for cancer.
Expert Rev Mol Med 12: e16.
2. Rafnar T, Sulem P, Stacey SN, Geller F, Gudmundsson J, et al. (2009) Sequence
variants at the TERT-CLPTM1L locus associate with many cancer types. Nat
Genet 41: 221–227.
3. Shete S, Hosking FJ, Robertson LB, Dobbins SE, Sanson M, et al. (2009)
Genome-wide association study identifies five susceptibility loci for glioma. Nat
Genet 41: 899–904.
4. Landi MT, Chatterjee N, Yu K, Goldin LR, Golds tein AM, et al. (2009) A
genome-wide association study of lung cancer identifies a region of chromosome
5p15 associated with risk for adenocarcinoma. Am J Hum Genet 85: 679–691.
5. McKay JD, Hung RJ, Gaborieau V, Boffetta P, Chabrier A, et al. (2008) Lung
cancer susceptibility locus at 5p15.33. Nat Genet 40: 1404–1406.
6. Petersen GM, Amundado ttir L, Fuchs CS, Kraft P, Stolzenberg-Solomon RZ,
et al. (2010) A genome-wide association study identifies pancreatic cancer
susceptibility loci on chromosomes 13q22.1, 1q32.1 and 5p15.33. Nat Genet 42:
224–228.
7. Turnbull C, Rapley EA, Seal S, Pernet D, Renwick A, et al. (2010) Variants near
DMRT1, TERT and ATF7IP are associated with testicular germ cell cancer.
Nat Genet 42: 604–607.
8. Rothman N, Garcia-Closas M, Chatterjee N, Malats N, Wu X, et al. (2010) A
multi-stage genome-wide association study of bladder cancer identifies multiple
susceptibility loci. Nat Genet 42: 978–984.
9. Johnatty SE, Beesley J, Chen X, Macgregor S, Duffy DL, et al. (2010)
Evaluation of candidate stromal epithelial cross-talk genes identifies association
between risk of serous ovarian cancer and TERT, a cancer susceptibility ‘‘hot-
spot’’. PLoS Genet 6: e1001016.
10. Ramus SJ, Vierkant RA, Johnatty SE, Pike MC, Van Den Berg DJ, et al. (2008)
Consortium analysis of 7 candidate SNPs for ovarian cancer. Int J Cancer 123:
380–388.
11. Zhang B, Beeghly-Fadiel A, Long J, Zheng W (2011) Genetic variants associated
with breast-cancer risk: comprehensive research synopsis, meta-analysis, and
epidemiological evidence. Lancet Oncol 12: 477–488.
12. Antoniou AC, Wang X, Frederick sen ZS, McGuffog L, Tarrell R, et al. (2010) A
locus on 19p13 modifies risk of breast cancer in BRCA1 mutation carriers and is
associated with hormone receptor-negative breast cancer in the general
population. Nat Genet 42: 885–892.
13. Bolton KL, Tyrer J, Song H, Ramus SJ, Notaridou M, et al. (2010) Common
variants at 19p13 are associated with susceptibility to ovarian cancer. Nat Genet
42: 880–884.
14. Turnbull C, Ahmed S, Morrison J, Pernet D, Renwick A, et al. (2010) Genome-
wide association study identifies five new breast cancer susceptibility loci. Nat
Genet 42: 504–507.
15. Chen YJ, Campbell HG, Wiles AK, Eccles MR, Reddel RR, et al. (2008) PAX8
regulates telomerase reverse transcriptase and telomerase RNA component in
glioma. Cancer Res 68: 5724–5732.
16. Kyo S, Takakura M, Kanaya T, Zhuo W, Fujimoto K, et al. (1999) Estrogen
activates telomerase. Cancer Res 59: 5917–5921.
17. Yi X, Shay JW, Wright WE (2001) Quantitation of telomerase components and
hTERT mRNA splicing patterns in immortal human cells. Nucleic Acids Res
29: 4818–4825.
18. Liu JZ, McRae AF, Nyholt DR, Medland SE, Wray NR, et al. (2010) A versatile
gene-based test for genome-wide association studies. Am J Hum Genet 87:
139–145.
19. Choi J, Southworth LK, Sarin KY, Venteicher AS, Ma W, et al. (2008) TERT
promotes epithelial proliferation through transcriptional control of a Myc- and
Wnt-related developmental program. PLoS Genet 4: e10.
20. Dunning AM, Healey CS, Baynes C, Maia AT, Scollen S, et al. (2009)
Association of ESR1 gene tagging SNPs with breast cancer risk. Hum Mol
Genet 18: 1131–1139.
TERT SNPs and Ovarian and Breast Cancer Risk
PLoS ONE | www.plosone.org 6 September 2011 | Volume 6 | Issue 9 | e24987