Annals of Oncology 19: 1910–1914, 2008
Published online 15 July 2008
Association of the polymorphism of the CAG repeat in
the mitochondrial DNA polymerase gamma gene
(POLG) with testicular germ-cell cancer
M. Blomberg Jensen1*, H. Leffers1, J. H. Petersen1,2, G. Daugaard3, N. E. Skakkebaek1&
E. Rajpert-De Meyts1*
1University Departments of Growth & Reproduction;2Biostatistics and;3Oncology, Rigshospitalet and Copenhagen University, Copenhagen,
Received 20 February 2008; revised 25 April 2008; accepted 4 June 2008
Background: A possible association between the polymorphic CAG repeat in the DNA polymerase gamma (POLG)
gene and the risk of testicular germ-cell tumours (TGCT) was investigated in this study. The hypothesis was prompted
by an earlier preliminary study proposing an association of the absence of the common 10-CAG-long POLG allele with
testicular cancer as well as previously reported in some European populations’ association with male subfertility, which
is a condition carrying an increased risk of TGCT.
Patients and methods: The number of CAG repeats in both POLG alleles was established in 243 patients with
TGCT and in 869 controls by the analysis of the genomic DNA fragment.
Results: A significantly higher proportion of men homozygous allele of other than the common 10 CAG repeats was
found among the patients with TGCT in comparison to the controls (4.9% versus 1.3%, respectively, P = 0.001). The
vast majority of the homozygous patients had a seminoma (11 of 12; 97%), despite that only about half (55%) of the
studied patients had this tumour type.
Conclusions: The findings indicate that the POLG polymorphism may be a contributing factor in the pathogenesis of
TGCT particularly in seminoma, but the mechanisms remain to be elucidated.
Key words: CAG repeat, mitochondria, POLG gene, seminoma, testicular dysgenesis syndrome, testicular
The incidence of testicular germ-cell tumours (TGCT) has
more than doubled in the last 30 years and has become the
most common malignant neoplasm among men aged 25–40
years . The vast majority of TGCTs, except infantile germ-
cell tumour and spermatocytic seminoma of the elderly men,
are derived from carcinoma in situ testis (CIS-T)  and that
includes both main types of overt TGCTs, seminoma (so-called
classical, to distinguish from spermatocytic seminoma) and
non-seminoma [3, 4].
In addition to the observed increase in TGCT, there has also
been a rise in congenital malformations of male genitalia, such
as cryptorchidism or hypospadias, and an apparent decline in
sperm counts [5, 6]. The latter—even though is first observed
in young adults—may in many cases be linked to prenatal
development, as recently shown in men exposed in utero to
maternal smoking . According to our hypothesis, TGCT and
the above-mentioned reproductive disorders are manifestations
of the testicular dysgenesis syndrome (TDS). We proposed
that the aetiology of TDS is linked to adverse influence of
environmental/lifestyle factors during early foetal development,
most probably combined with genetic predisposition [8, 9].
It has been shown in animal models that prenatal exposure
to hormones or endocrine disrupters may target specific
genes and lead to different forms of TDS, for example, in the
case of oestrogens that target Insl3 and lead to cryptorchidism
in rodents . The observed effects are, however, highly
variable, probably depending upon differences in genetic
background of the animals . Genetic predisposition is also
an obvious explanation for the ethnic differences in incidence
of TGCT in humans [12, 13]. We believe that multiple genes
may be responsible for predisposition to TGCT and other
forms of TDS, but the genes in question remain elusive.
One of the genes that have recently attracted our attention is
DNA polymerase gamma gene (POLG), which encodes the
catalytic subunit of DNA polymerase gamma (POLG), the key
nuclear enzyme responsible for replicating, elongation and
repair of the mitochondrial DNA (mtDNA). The POLG gene
*Correspondence to: Dr M. B. Jensen, Department of Growth and Reproduction,
Rigshospitalet, Blegdamsvej, Copenhagen 2200, Denmark. Tel: 35455017;
Fax: 35456054; E-mail: email@example.com
ª The Author 2008. Published by Oxford University Press on behalf of the European Society for Medical Oncology.
All rights reserved. For permissions, please email: firstname.lastname@example.org
by guest on June 2, 2013
contains a polymorphic CAG repeat region [14, 15] with
a high frequency of repeat lengths of 10 codons (75%–80%),
which indicates that it is maintained by selection . A new
study supported the idea that the length of the POLG
microsatellite region, rather than its exact nucleotide or amino
acid sequence, is what is maintained in animals and humans
. The repeat located in the mature polypeptide is not found
in the orthologous genes in mouse and yeasts, but a shorter
repeat is found in African great apes. Deletion of the POLG
CAG repeat was shown not to affect enzymatic properties,
but modestly up-regulate expression [17, 18].
Absence of the 10-codon repeat on both alleles (named here
the homozygous variant) has been associated with male
subfertility in Finland and Denmark, where the frequency of
this polymorphism among fertile men is ?1%–1.5% , while
it was 2.8% in the Danish subfertile patients . Studies in
other countries, however, did not confirm this association
[20–23]. Despite that the expression of POLG was found in
human sperm , the previous studies taken together
indicated that the POLG polymorphism had no consistent
influence on semen concentration, sperm motility or
morphology, thus limiting its predictive clinical value in
fertility clinics . In our study, however, we observed one
case of TGCT among the few subfertile men homozygous for
the lack of the common 10-codon repeat . Moreover,
a small preliminary study of 49 polish patients reported that the
POLG polymorphism may be a genetic risk for TGCT .
Another reason for our interest in the CAG–POLG
polymorphism was a previously reported CAG repeat
instability, which was first observed in tumour genome and
also described in germline DNA of some families with
testicular cancer , although this was disputed in another
study . We hypothesized that—whether or not linked to
a general CAG repeat instability in the genome—CAG repeat
polymorphism in the POLG gene could perhaps subtly
impair its function.
To investigate whether or not there is a link between the
POLG polymorphism and TGCT, we analysed a large
group of patients with testicular germ-cell neoplasia and
well-characterized controls and report here the results
indicating that this polymorphism may indeed be associated
with an increased risk of testicular cancer.
materials and methods
subjects and clinical analysis
The study group included 243 Danish patients diagnosed with testicular
germ-cell neoplasia and referred to our hospital for oncological treatment,
semen banking or post-therapy fertility assessment. The average age of
patients at the time of the diagnosis was 33 years. Patients with testicular
tumours derived from other cell types (e.g. lymphoma or Leydig tumour)
were excluded. In the TGCT group, 55% of the patients had a seminoma,
while nearly 42% had a non-seminoma or mixed germ-cell tumours. The
remaining nine patients (3.7%) harboured pre-invasive CIS-T, without an
overt tumour. Diagnosis was on the basis of histological evaluation of the
orchidectomy preparation by pathologists.
The control group included 869 young healthy men, without any signs
of malignancy, 495 were young military conscripts and 374 proven fertile
men. All subjects were Danish. The same control groups were used in our
previous study of the POLG gene polymorphism in subfertile men .
A subset of TGCT patients (those referred for semen banking or
fertility treatment) and all control subjects underwent an andrological
examination including analysis of semen and reproductive hormones in
serum. The study was conducted in accordance with the law concerning
the protection of personal information and approved by the regional
Medical Research Ethics Committee.
DNA was isolated from peripheral blood samples using a kit (Roche
Diagnostics Gmbh, Mannheim, Germany). Two primers
(GGTCCCTGCACCAACCATGA and CTTGCCCGAAGATTTGCTCGT),
matching positions 267–286 and 533–553, respectively, of the POLG
transcript (ENSEMBL gene ENSG00000140521), were used to amplify
a 286-bp DNA fragment, using the Pfu DNA polymerase (Stratagene,
San Diego, CA). PCR was carried out in 30 ll of (final concentrations)
12 mM Tris–HCl, pH 8.3; 50 mM KCl; 1.9 mM MgCl2; 0.1% Triton X-100;
0.005% gelatine; 250 lM dNTP; 30 pmol of each primer and 2 U Pfu
polymerase. PCR conditions were 98?C for 5 min, 40 cycles of 98?C for 30 s,
63?C for 1 min, 72?C for 1:45 min and one cycle of 72?C for 5 min.
Eight per cent of the samples failed the amplification and were excluded.
Earlier investigations showed a success rate ?90%, and our rate was slightly
higher after amplification and determining the CAG repeat length. DNA
fragments were purified from 1% agarose gels and analysed first by applying
a Cy5-labelled sequencing primer (CY5-CTGGATGTCCAATGGGTTGT,
positions 494–513) under standard PCR conditions, resulting in a DNA
fragment, which was run on an ALF express sequencer and analysed by
the Fragment Analyser software (Amersham-Pharmacia Biotech, Uppsala,
The reproducibility of the assay was tested in the beginning of our
first study . It included 1298 men and the first 300 were analysed
using two different methods: by direct sequencing and by fragment
length on the ALF express sequencer, with no difference in outcome. All
bands that differed from a normal CAG 10/10 band (putative x/x) were
Differences between allele frequencies were assessed using Fisher’s exact
test, which correctly takes into account the low number of polymorphic
homozygotes in both cancer patients and controls.
frequency and distribution of the POLG genotypes
All the frequencies of the alleles in the studied groups are
shown in Table 1. In concert with the previous report ,
homozygotes with 10 repeats on both alleles (10/10) were the
most common (76.6% of all studied subjects). The frequency of
the common POLG variant among the TGCT patients was
75.7%, which was close to that in the combined control
groups (76.6%), P = 0.80. The frequency of the absence of
the common 10 repeats on both alleles among the TGCT
patients was significantly higher compared with the controls
(4.9% versus 1.3%, respectively, P = 0.001).
CAG repeat lengths ranging from 4 to 12 triplets were
detected. This range and distribution are comparable to what
was shown earlier , where the majority of alleles that
deviate from 10 are close to 10 in length, with a small
tendency to be >10, the majority being 11. The proportion
of repeats <10 was comparable to the Finnish study  but
Annals of Oncology
Volume 19|No. 11| November 2008doi:10.1093/annonc/mdn407 | 1911
by guest on June 2, 2013
was much higher than that in the Italian study  (data not
phenotype versus genotype
The distribution of homozygous patients among tumour types
was highly non-random (P = 0.035). Nearly all homozygous
patients had a seminoma (11 of 12, 96.7%, Tables 1 and 2).
Except this finding there were no obvious phenotypic
differences between the TGCT patients with and without the
POLG polymorphism. Low serum inhibin B concentrations in
two patients homozygous for the POLG polymorphism indicate
that they probably had impaired spermatogenesis, but their
semen characteristics were unknown (Table 2). There was only
one known TGCT patient with severe oligozoospermia. There
was no significant trend or difference in the average sperm
concentration stratified on the three different haplotypes (data
not shown). As illustrated in Table 3, in patients with
heterozygosity we found no association of the numbers of CAG
repeats in the POLG gene with tumour histology, clinical
history or aggressiveness of TGCT.
In this study, we found a significantly higher frequency of the
homozygous absence of the common POLG haplotype among
patients with testis cancer (TGCT) compared with young
Danish men without TGCT and frequencies of this
polymorphism in the European populations . The
frequency of the homozygous POLG polymorphism (4.9%) was
even higher than that reported among infertile men,
a condition linked epidemiologically to testicular cancer .
We found a clear predominance of one tumour type
(seminoma) among the patients with the homozygous absence
of the common allele, but did not find any correlation of the
POLG polymorphism with other clinical parameters.
We are aware of only one previous study of POLG genotypes
in patients with testicular cancer; a very small preliminary study
which found a higher frequency of polymorphic genotypes;
26% versus 11% (13 heterozygotes and one homozygous
combined) in a group of 49 Polish TGCT patients versus 55
controls . That study, however, showed a large difference in
the frequency of the POLG haplotypes among the controls
Table 1. Distribution of the DNA polymerase gamma (POLG) alleles in patients with testis cancer [testicular germ-cell tumour (TGCT)] and controls
GroupNN and frequency (with 95% confidence intervals) of the three CAG repeat genotype patterns in both alleles
243 184 (75.7%, 70.0% to 80.7%)
6 (66.7%, 35.4% to 87.9%)
100 (75.2%, 67.2% to 81.8%)
78 (77.2%, 68.1% to 84.3%)
666 (76.6%, 73.7% to 79.3)
47 (19.3%, 14.9% to 24.8%)
3 (33.3%, 12.1% to 64.6%)
22 (16.5%, 11.2% to 23.8%)
22 (21.8, 14.8% to 30.8%)
192 (22.1%, 19.5% to 25.0%)
12 (4.9%, 2.8% to 8.4%)*
0 (0%, 0% to 29.9%)
11 (8.3%, 4.7% to 14.2%)**
1 (1.0%, 0.2% to 5.4%)
11 (1.3%, 0.7% to 2.3%)
The distribution of the wild-type 10/10, heterozygotes 10/x and the homozygous x/x variants among patients and controls. The patients with TGCT are
divided according to the histological pattern. CIS-T, carcinoma in situ testis.
*P = 0.001 in comparison to the control group.
**P < 0.001 in comparison to the control group.
Table 2. Characteristics of the testicular germ-cell tumour (TGCT) patients homozygous for the DNA polymerase gamma (POLG) polymorphism
N POLG genotype Histology Stage and spread of diseaseRelapse Sperm concentration
Inhibin B serum
Advanced (good prognosis*)
Advanced (good prognosis*)
Note that it was not possible to gather information on all the patients because they were treated in different hospitals and some records were no longer
available. S, seminoma; NS, non-seminoma; EC, embryonal carcinoma; n.a., not available.
*According to International Germ Cell Cancer Collaborative Group classification.
**With a timespan of 3 years, he harboured in the contralat. testis a Seminoma and a focal EC in the frozen section, EC not found in paraffin after
Annals of Oncology
1912 | Blomberg Jensen et al.Volume 19|No. 11| November 2008
by guest on June 2, 2013
compared with the study of Eurasian populations that included
also 102 Poles; 19.6% of whom had the polymorphic alleles
. The inconsistency of the distribution of the genotypes
between the two studies raises the question of a possible
chance finding in the small study of Nowak et al. .
Nevertheless, our current results obtained in much larger
groups of patients and controls support the existence of a true
association of the POLG polymorphism with TGCT. In the
Danish population, however, the difference was only found
in the distribution of the homozygous variant genotype (with
the lack of 10 CAG repeats on both alleles) and there was no
difference in the frequency of heterozygotes.
The association of the POLG polymorphism with testicular
cancer is consistent with the previously reported link to male
subfertility, although—except for two studies [16, 24]—the
latter link was not confirmed by several studies mentioned
in the introduction, gathered recently by Westerweld et al. .
Also in the current study, no association was found between the
POLG polymorphism and any phenotypic parameter linked
to sperm quality, such as sperm motility or morphology.
Although fertility is impaired in the majority of patients
with TGCT , the sperm concentration is often retained
within the normal range.
An association of the POLG polymorphism with both
subfertility and testicular cancer in the Danish population
supports our hypothesis that impaired spermatogenesis and
TGCT are likely to be aetiologically linked in some patients. As
our hypothesis stipulates that these two diseases are parts of
TDS, the POLG polymorphism may be considered
a predisposing genetic factor, perhaps increasing sensitivity to
environmental factors such as endocrine disrupters. Higuchi
et al.  showed that depletion of the mtDNA determines
androgen dependence in prostate cancer cell lines, and this
could be a possible explanation of this polymorphism’s
effect on testicular germ cells. Still we do not know the effect
of this genetic ‘defect’, but a new study indicated that the length
of the POLG microsatellite region is maintained in all
investigated species, including humans .
Surprisingly, we found a significant (P = 0.035) association
with the absence of the common CAG repeat length on both
chromosomes and the histology of the tumour as 11 out of
12 tumours were seminomas. The one patient with a non-
seminoma (embryonal carcinoma) had the genotype 7/12
CAG repeats, which included the shortest repeat of all the
homozygous variants. Interestingly, we found one heterozygote
patient with 7/10 genotype and the tumour was a non-
seminoma as well. We did not find any correlation between the
repeat length and histology among all the heterozygotes,
where ?50% had seminomas, while the remaining half had
non-seminomas. We cannot explain the high frequency of
seminomas among the subjects with homozygous
polymorphism but we speculate that this genotype may, by
an unknown mechanism, promote the development of
seminomas, rather than non-seminomas. The progression of
CIS-T to a seminoma is thought to involve mainly increased
proliferation of tumour cells, which have essentially the
same phenotype as CIS-T cells, whereas the progression to
non-seminomas is associated with a profound
reprogramming with loss of many germ-cell-specific features
and up-regulation of the embryonic phenotype.
Concerning the mechanisms and the consequences of this
polymorphism on the enzymatic function of POLG, there is no
obvious explanation. Perhaps, the polymorphism affects the
function of POLG, thus increasing probability of errors in
the replication or proofreading of the mtDNA. We know that
point mutations in the coding regions of the POLG gene
accelerate accumulation of mutations, deletions and frame
shifts in the mitochondrial genome, and deletions in mtDNA
have been associated with male infertility [18, 32]. However,
any explanation remains speculative until an independent
In conclusion, we found an association between the
POLG gene polymorphism and TGCT, especially classical
seminoma. The POLG polymorphism can be a contributing
genetic risk factor for testicular cancer but the possible
mechanisms remain elusive and require further studies.
Danish Cancer Society; Svend Andersen Foundation; Danish
Medical Research Council.
The authors thank Drs C. E. Hoei-Hansen, N. Jørgensen, E.
Carlsen, L. Frydelund-Larsen and T. K. Jensen for help with
collecting blood samples and clinical information, and L. G.
Pedersen for expert technical assistance.
1. Huyghe E, Matsuda T, Thonneau P. Increasing incidence of testicular cancer
worldwide: a review. J Urol 2003; 170: 5–11.
Table 3. Clinical characteristics of the testicular germ-cell tumour (TGCT) patients heterozygous for DNA polymerase gamma variants
Stage IAdvancedUnknown stage Relapseda
All the included patients have one allele with 10 CAG, and according to the repeat length on the other allele are divided in two groups >10 or <10. This allows
to evaluate if a low or a high repeat length contributes to a different clinical outcome, but no significant differences were found. S, seminoma; NS, non-
aPatients with inadequate follow-up information are not included.
Annals of Oncology
Volume 19|No. 11| November 2008doi:10.1093/annonc/mdn407 | 1913
by guest on June 2, 2013
2. Skakkebaek NE. Possible carcinoma-in-situ of the testis. Lancet 1972; 2: Download full-text
3. Skakkebaek NE, Berthelsen JG, Giwercman A, Muller J. Carcinoma-in-situ of the
testis: possible origin from gonocytes and precursor of all types of germ cell
tumours except spermatocytoma. Int J Androl 1987; 10: 19–28.
4. Oosterhuis JW, Looijenga LH. Testicular germ-cell tumours in a broader
perspective. Nat Rev Cancer 2005; 5: 210–222.
5. Carlsen E, Giwercman A, Keiding N, Skakkebaek NE. Evidence for decreasing
quality of semen during past 50 years. BMJ 1992; 305: 609–613.
6. Boisen KA, Kaleva M, Main KM et al. Difference in prevalence of congenital
cryptorchidism in infants between two Nordic countries. Lancet 2004; 363:
7. Jensen TK, Jørgensen N, Punab M et al. Association of in utero exposure to
maternal smoking with reduced semen quality and testis size in adulthood:
a cross-sectional study of 1,770 young men from the general population in five
European countries. Am J Epidemiol 2004; 159: 49–58.
8. Skakkebaek NE, Rajpert-De Meyts E, Main KM. Testicular dysgenesis syndrome:
an increasingly common developmental disorder with environmental aspects.
Hum Reprod 2001; 16: 972–978.
9. Moller H, Skakkebaek NE. Risk of testicular cancer in subfertile men: case-
control study. BMJ 1999; 318: 559–562.
10. Emmen JM, McLuskey A, Adham IM et al. Involvement of insulin-like factor 3
(Insl3) in diethylstilbestrol-induced cryptorchidism. Endocrinology 2000; 141:
11. Skakkebaek NE, Rajpert-De Meyts E, Jorgensen N et al. Germ cell cancer and
disorders of spermatogenesis: an environmental connection? APMIS 1998; 106:
12. McGlynn KA, Devesa SS, Sigurdson AJ et al. Trends in the incidence of testicular
germ cell tumors in the United States. Cancer 2003; 97: 63–70.
13. Shah MN, Devesa SS, Zhu K, McGlynn KA. Trends in testicular germ cell tumours
by ethnic group in the United States. Int J Androl 2007; 30: 206–213.
14. Bolden A, Noy GP, Weissbach A. DNA polymerase of mitochondria is a gamma-
polymerase. J Biol Chem 1977; 252: 3351–3356.
15. Ropp PA, Copeland WC. Cloning and characterization of the human
mitochondrial DNA polymerase, DNA polymerase gamma. Genomics 1996; 36:
16. Rovio AT, Marchington DR, Donat S et al. Mutations at the mitochondrial DNA
polymerase (POLG) locus associated with male infertility. Nat Genet 2001; 29:
17. Rovio AT, Abel J, Ahola AL et al. A prevalent POLG CAG microsatellite
length allele in humans and African great apes. Mamm Genome 2004; 15:
18. Spelbrink JN, Toivonen JM, Hakkaart GA et al. In vivo functional analysis of the
human mitochondrial DNA polymerase POLG expressed in cultured human cells.
J Biol Chem 2000; 275: 24818–24828.
19. Jensen M, Leffers H, Petersen JH et al. Frequent polymorphism of the
mitochondrial DNA polymerase gamma gene (POLG) in patients with normal
spermiograms and unexplained subfertility. Hum Reprod 2004; 19: 65–70.
20. Aknin-Seifer IE, Touraine RL, Lejeune H et al. Is the CAG repeat of mitochondrial
DNA polymerase gamma (POLG) associated with male infertility? A multi-centre
French study. Hum Reprod 2005; 20: 736–740.
21. Krausz C, Guarducci E, Becherini L et al. The clinical significance of the POLG
gene polymorphism in male infertility. J Clin Endocrinol Metab 2004; 89:
22. Brusco A, Michielotto C, Gatta V et al. The polymorphic polyglutamine repeat in
the mitochondrial DNA polymerase gamma gene is not associated with
oligozoospermia. J Endocrinol Invest 2006; 29: 1–4.
23. Harris TP, Gomas KP, Weir F et al. Molecular analysis of polymerase gamma
gene and mitochondrial polymorphism in fertile and subfertile men. Int J Androl
2006; 29: 421–433.
24. Amaral A, Ramalho-Santos J, St John JC. The expression of polymerase gamma
and mitochondrial transcription factor A and the regulation of mitochondrial DNA
content in mature human sperm. Hum Reprod 2007; 22: 1585–1596.
25. Tu ¨ttelmann F, Rajpert-De Meyts E et al. Gene polymorphisms and male
infertility–a meta-analysis and literature review. Reprod Biomed Online 2007;
26. Nowak R, Zub R, Skoneczna I et al. CAG repeat polymorphism in the DNA
polymerase gamma gene in a Polish population: an association with testicular
cancer risk. Ann Oncol 2005; 16: 1211–1212.
27. King BL, Peng HQ, Goss P et al. Repeat expansion detection analysis of (CAG)n
tracts in tumor cell lines, testicular tumors, and testicular cancer families. Cancer
Res 1997; 57: 209–214.
28. Ono Y, De Meyts ER, Guellaen G, Bulle F. Sporadic testicular germ cell cancers
do not exhibit specific alteration in CAG/CTG repeats containing genes expressed
in human testis. Oncogene 2001; 20: 5548–5553.
29. Malyarchuk BA, Papuga M et al. Low variability of the POLG (CAG)n repeat in
north Eurasian populations. Hum Biol 2005; 77: 355–365.
30. Westerweld G, Kaaij-Visser L, Tanck M et al. CAG repeat length variation in the
polymerase gamma (POLG) gene: effect on semen quality. Mol Hum Reprod
2008; 14: 245–9.
31. Higuchi M, Kudo T et al. Mitochondrial DNA determines androgen dependence in
prostate cancer cell lines. Oncogene 2006; 25: 1427–1445.
32. May-Panloup P, Chre ´tien M-F, Svagner F et al. Increased sperm mitochondrial
DNA content in male infertility. Hum Reprod 2003; 18: 550–556.
Annals of Oncology
1914 | Blomberg Jensen et al.Volume 19|No. 11| November 2008
by guest on June 2, 2013