749 860
1.
Introduction
Telomeres are TTAGGG nucleotide repeats and a protein
complex at chromosome ends that play an essential role in
maintaining chromosomal stability. Due to the inability of
DNA polymerase to fully extend 3
0
DNA ends, telomeres
become gradually shorter with each cell division in the
absence of telomerase activity
[1]. Although in normal cells
critically short telomeres will trigger cellular senescence
and death, cancer cells can continue to divide despite
telomere shortening and the resultant genomic instability
[2]. Alternatively, upregulated telomerase activity leading
to increased telomere length may also promote tumorigen-
esis by conferring properties of immortal growth
[3] .Indeed,
recent studies suggest longer telomere length may be a risk
factor for select tumor types including melanoma, lung
cancer, chronic lymphocytic leukemia, glioma, and ovarian
cancer
[4 – 7].
As such, relative telomere length in peripheral blood
leukocytes has been evaluated in numerous population-
based studies as a suspected marker of cancer risk
[8]. Most
of these studies have characterized telomere length using
multiplex quantitative polymerase chain reaction (qPCR)
assays
[9] .Results of studies of leukocyte telomere length
and risk of renal cell carcinoma (RCC) have been inconsis-
tent. Two small hospital-based case-control studies
reported inverse associations between telomere length
and risk of RCC
[10,11] ,whereas no significant evidence of
an association was observed in a larger population-based
case-control study
[12]and two cohort-based investiga-
tions using prediagnostic samples
[13,14] .In contrast,
longer leukocyte telomere length has been associated with
reduced RCC survival
[15]. Telomerase activity is elevated in
renal tumors compared with adjacent normal renal tissue
and has been associated with clinicopathologic features of
advanced disease
[16,17].
These previous studies have several limitations. Leuko-
cyte telomere length measurements in case-control studies,
using postdiagnosis blood samples, may have been influ-
enced by effects of the disease. All studies measured
telomere length from a single time point, which may not
adequately reflect telomere length status in the etiologically
relevant time window, and were susceptible to confounding
from RCC risk factors that may be associated with telomere
length such as smoking
[13,18]and obesity
[19]. Further-
more, qPCR-based measurements of telomere length are
sensitive to preanalytic factors such as DNA source material
and extraction method
[12,20,21] .Nine common genetic variants have been identified in
genome-wide association studies (GWAS) that are associ-
ated with leukocyte telomere length at a level of genome-
wide significance (
p
<
5 10
8
)
[22 – 24] .Recent studies
have evaluated the relationship between these genetic
proxies of telomere length and risk of cancer and found
evidence suggesting longer genetically inferred telomere
length is associated with increased cancer risk
[4 – 7]. The
approach employed by these studies, Mendelian randomi-
zation, uses genetic variants associated with leukocyte
telomere length as genetic instruments to investigate the
relationship between leukocyte telomere length and RCC
risk. For resulting effect estimates to have a valid causal
interpretation, several conditions must hold: (1) the
telomere length associated variants must be associated
with telomere length in circulating leukocytes, (2) the
telomere length-associated variants should not be associ-
ated with other factors that are associated with telomere
length and RCC risk, and (3) the telomere length associated
variants can only influence RCC risk by their effect on
telomere length, that is they cannot have pleiotropic effects.
An advantage of this approach is that it is not susceptible to
the biases associated with measured telomere length as
described above. A recent investigation surveying several
chronic conditions suggested a marginal positive associa-
tion (
p
= 0.01) between genetically predicted telomere
length and RCC risk, although the sample size was smaller
(
N
= 2461 RCC cases)
[7].
In the present study, we evaluated RCC risk in relation to
individual telomere length-related genetic variants and an
aggregate genetic risk score (GRS) of telomere length-
associated genetic variants in a large sample of six RCC
GWAS datasets combined by meta-analysis to investigate a
potential etiologic relationship between telomere length
and RCC risk. We evaluated whether a genetic profile that is
associated with longer telomere length is associated with
risk of overall RCC and RCC subtypes, and investigated
potential modifiers of this relationship.
2.
Material and methods
The RCC GWAS meta-analysis included a total of 10 784 RCC cases and 20
406 controls of European ancestry from six independent scans
conducted at the International Agency for Cancer Research (IARC; two
scans totaling 5219 RCC cases and 8011 cancer-free controls; analyzed as
a combined dataset), the MD Anderson Cancer Center (893 RCC cases,
556 cancer-free controls), the US National Cancer Institute (NCI-1:
1311 RCC cases, 3424 cancer-free controls; NCI-2: 2417 RCC cases,
4391 cancer-free controls; analyzed separately), and the Institute of
Cancer Research (UK; 944 RCC cases, 4024 cancer-free controls)
[25]. Cases were restricted to adults diagnosed with RCC, de
fi
ned on
the basis of the International Classi
fi
cation of Disease for Oncology 2nd
and 3rd Edition topography code C64. Samples were genotyped on
commercially available Illumina Single Nucleotide Polymorphism (SNP)
microarrays (HumanHap 300, HumanHap 500, HumanHap 610,
HumanHap 660w, HumanHap 1.2 M, OmniExpress, Omni5 M) after
standard quality control metrics. High-quality genotypes were phased
and imputation was performed using either MaCH (IARC) or IMPUTE2
(UK, NCI1, NCI2, and UK) with 1000 Genomes Project (Phase 1, Version 3)
samples used as a reference panel for imputing missing genotypes.
Protocols for studies participating in each GWAS were reviewed by the
Institutional Review Boards of their respective institutions. All partici-
pants provided written informed consent. Further details on study
design and methods have been previously reported
[25] .For each study participant, genotypes were extracted for nine
previously identi
fi
ed common SNPs associated with telomere length in
circulating leukocytes (rs10936599, rs11125529, rs2736100, rs3027234,
rs6772228, rs755017, rs7675998, rs8105767, and rs9420907). Telomere
length-associated SNPs not directly genotyped were extracted from
imputed data for each scan (Supplementary Table 1)
[25] .Risk of RCC was evaluated in relation to each of the nine telomere
length-associated variants. Association testing was conducted separately
E U R O P E A N U R O L O GY 7 2 ( 2 0 17 ) 74 7
–
7 5 4
749