Embed Size (px)
Citation preview
sevier.com/locate/ygyno
Gynecologic Oncology 1
Two functionally relevant polymorphisms in the human progesterone
receptor gene (+331 G/A and progins) and the predisposition for
breast and/or ovarian cancer
Andrea Romano a,b,*,1, Patrick J. Lindsey c,d,2, Dagmar-C. Fischer a,b,1, Bert Delvoux a,b,1,
Aimee D.C. Paulussen c,d,2, Rob G. Janssen b,c,2, Dirk G. Kieback a,1
a Department of Obstetrics and Gynaecology, University Hospital of Maastricht, The Netherlandsb Research Institute Growth and Development (GROW), The Netherlands
c Department of Population Genetics, University and University Hospital of Maastricht, The Netherlandsd Nutrition and Toxicology Research Institute Maastricht (NUTRIM), Maastricht, The Netherlands
Received 6 June 2005
Available online 19 December 2005
Abstract
Objective. Two polymorphisms affecting either expression (+331 G/A) or transcriptional activity (progins) of the progesterone receptor have
been described. No clear correlation between either polymorphism and breast or ovarian cancer has been shown. Our objective is to clarify
whether the two progesterone receptor polymorphisms modify the risk for breast or ovarian cancer.
Methods. Healthy women and women suffering from either ovarian or breast cancer were enrolled in a case-control-based study to compare
the frequencies of women carrying either one, both or none of the two polymorphisms. Patient and control populations resided in the same region
of South Germany. PCR-RFLP analysis was used to determine the polymorphic alleles.
Results. Women diagnosed with ovarian cancer showed a not significant increased frequency of +331 A carriers and a significantly increased
frequency of progins carriers. Both polymorphisms appeared to be associated with a significantly increased risk for the disease in women below 51
years [OR: 4.1 (CI: 1.2–13.9) and 3.2 (CI: 1.1–9.1), respectively]. No association was detected between either of the two polymorphisms and
breast cancer. Among ovarian and breast cancer patients, the number of individuals carrying both rare polymorphic alleles was significantly higher
compared to healthy women.
Conclusions. Our findings support the hypothesis that low penetrant polymorphisms of progesterone receptor may modify the risk for ovarian
cancer. Our data do not allow drawing a clear conclusion on the risk for breast cancer.
D 2005 Elsevier Inc. All rights reserved.
Keywords: Progesterone receptor; +331 G/A promoter polymorphism; Progins; Ovarian cancer; Breast cancer
0090-8258/$ - see front matter D 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.ygyno.2005.10.040
* Corresponding author. Department of Obstetrics and Gynaecology, Uni-
versity Hospital of Maastricht, The Netherlands. Fax: +31 43 38 74765.
E-mail addresses: [email protected] (A. Romano),
[email protected] (P.J. Lindsey), [email protected]
(D.-C. Fischer), [email protected] (B. Delvoux),
[email protected] (A.D.C. Paulussen),
[email protected] (R.G. Janssen), [email protected]
(D.G. Kieback).1 Fax: +31 43 38 84573.2 Fax: +31 43 38 74765.
Introduction
Growth, differentiation, maturation and proliferation of the
female genital tract are highly dependent on steroid hormones,
which in turn exert their multiple effects after binding to their
cognate steroid hormone receptor. Progesterone is not only an
important regulator of ovulation but is also responsible for
pregnancy-associated proliferation and differentiation of breast
and endometrium. The progesterone receptor gene (PR) is
located on chromosome 11q22–23 and the encoded receptor
protein consists of several functional domains including an N-
terminal domain (NTD or A/B domain), a DNA-binding
domain (DBD) and a C-terminal ligand-binding domain
01 (2006) 287 – 295
www.el
A. Romano et al. / Gynecologic Oncology 101 (2006) 287–295288
(LBD). Whereas the NTD contains the ligand-independent
activation function, AF1 [1], steroid hormone induced tran-
scriptional activation is controlled by binding between the
natural agonist and the LBD, which harbors the ligand-
dependent activation function 2 (AF2). Two different promo-
ters give rise to two PR transcripts, subsequently translated into
the two receptor isoforms PR-A and PR-B ([1] and references
therein). Whereas PR-A consists of 764 amino acids (ca. 90
kDa), PR-B contains an additional 164 amino acids at the N-
terminus, resulting in a receptor-protein of about 114 kDa. PR-
B harbors a third trans-activation function (AF-3), which is
located inside the N-terminal part specific for PR-B ([1] and
references therein). Although both receptors bind to and
mediate progesterone activity, the responses of ligand-activated
PR-A and PR-B strongly depend on the cellular context.
Two PR polymorphic variants, which alter either function or
expression of PR, have been identified. In 2002, de Vivo et al.
described a G to A exchange at position +331 of the promoter
region [2]. This single nucleotide polymorphism (SNP) results in
the introduction of a TATA-box, which exclusively enhances
transcription of PR-B thereby increasing the ratio PR-B to PR-A.
The second polymorphism is referred to as ‘‘progins’’ [3,4] and
consists of a 320-bp PV/HS-1 Alu insertion in intron G3 and two
point mutations affecting exon 4 (V660L) and exon 5 (H770H)4.
These three aberrations have been found in complete linkage
disequilibrium in all populations studied [2,5,6] and the 660L
variant of both A and B isoforms showed increased stability and
trans-activation activity [7].
Functional polymorphisms in genes involved in proliferation
and cellular homeostasis may affect the risk for benign and
malignant disorders. Although ovarian and breast cancer show
familial clustering [8,9], high penetrant mutations in BRCA-1 and/
or BRCA-2 genes are present only in a minority of patients [8–
10]. Thus, it has been hypothesized that the development of either
cancer might be related to multiple low penetrant polymorphisms
rather than to single highly penetrant mutations [9,10].
Given the pivotal role played by progesterone and its
receptor in breast and ovary, we evaluated the frequencies of
the promoter SNP (+331 G/A) and the progins haplotype in
unrelated groups of healthy women and women diagnosed with
either ovarian or breast cancer of Caucasian origin and residing
in the same region of southern Germany.
Materials and methods
Study populations
Samples from two groups of female patients were enrolled in this study.
Patients were diagnosed between 1980 and 1999 with either epithelial ovarian
or breast cancer. Samples of epithelial ovarian cancer (OC, n = 75) originated
from the Department of Obstetrics and Gynaecology, Ulm University Medical
Center and samples from breast cancer patients (BC, n = 569) originated from
3 Accession number NCBI Z49816.4 Accession number NCBI AF016381. SNP-1: Reference SNP Cluster
Report (ref SNP ID): rs1042838, consisting of a G to T substitution in exon
4 at position 3432, counting reference NCBI RefSeq NM_000926.2, causing a
substitution of a valin into a leucin in codon 660; SNP-2. ref SNP ID:
rs1042839, a silent C to T substitution in exon 5 at position 3764.
the Departments of Obstetrics and Gynaecology at either Ulm University
Medical Center or Freiburg University Medical Center. Specific cancer
characteristics such as stage of the disease, grading, metastasis or histological
subtype were not used as a criterion for the exclusion or inclusion of samples.
For the control population, DNA isolated from peripheral blood lymphocytes of
healthy women (n = 484) was provided anonymously by the blood bank of
Ulm. Both patients and anonymous healthy volunteers consented to the
utilization of these specimens for research purposes in advance. Age and
genotype distributions for each group of women and each cohort selected for
subsequent association analyses are given in Tables 1A–C.
DNA extraction and genotype analysis
Genomic DNAwas isolated from peripheral blood lymphocytes or paraffin-
embedded tissue and submitted to PCR as described previously [5,11]. In total,
484 controls, 75 samples from ovarian cancer patients and 569 samples from
breast cancer patients were available. Due to occasionally poor quality of the
DNA, 379 controls, 52 OC and 535 BC were successfully genotyped for the
+331 G/A polymorphism, 443 controls, 67 OC and 546 BC were genotyped for
the progins polymorphism and 361 controls, 50 OC and 505 BC samples were
successfully genotyped for both polymorphisms.
The +331 G/A promoter polymorphism and the two polymorphisms
characteristic for progins were detected by PCR-RFLP as previously described
[12,13]. The presence or absence of the 320 bp HS-Alu insertion was judged
directly from the size of the amplified fragment [5]. Two independent
investigators masked to case-controls performed independent PCR-RFLP
analyses and assigned genotypes. The genotype of samples that were hetero/
homozygous for +331 A or progins was confirmed by an extra PCR-RFLP
analysis. We adopted this strategy to reduce the potential for false-positive
results. All PCR reactions (20 Al final volume) consisted of 50 ng genomic DNA,
500 nM of each primer, 0.05 mMof each dNTP, 0.5 unit of Taq DNA polymerase
and 1� buffer with 2.5 mM MgCl2. The Taq polymerase (Qiagen, Hilden,
Germany) for PCR and the restriction endonucleases NlaIV (isoschizomer
BspLI, Fermentas), BrsI (isoschizomer BrsSI, Promega) and NlaIII (BioLabs)
were used according to the manufacturer’s recommendations. The digested PCR
products were resolved by electrophoresis on a 2.5% agarose gel, stained with
Gelstar (Cambrex, ME, USA) and visualized by UV illumination. Although
several groups showed that the three variants belonging to the progins
complex of aberrations (i.e. V660L, H770H and Alu insertion) are always
in complete linkage disequilibrium [2,5,6], we confirmed this in a subset of
150 samples before deciding that PCR-RFLP analysis of exon 4 (V660L)
would be sufficient to determine the progins status. The most common
allele (V660) is indicated as A1 whereas the progins rare allele (L660) is
indicated as A2.
Statistical analyses
Hardy–Weinberg equilibrium for the two polymorphisms was assessed in
each cohort by comparing the observed genotype distribution with that
expected under Hardy–Weinberg assumption for the estimated allele frequen-
cy, and comparing the Pearson’s goodness-of-fit with a chi-square distribution
with one degree of freedom (Institute of Human Genetic, Munich, Germany:
http://www.ihg.gsf.de/). Because the frequencies of homozygous carriers of
both rare polymorphic alleles were very low, the frequencies of homo- and
heterozygous carriers of either the +331 A or the A2 allele were combined for
association analysis using a Fischer’s Exact test. Frequencies of individuals
carrying at the same time both polymorphisms were compared between healthy
controls and patients using Fisher’s Exact test.
Results
Age distribution, allele and genotype frequencies (+331 G/A,
A1/A2) in controls, OC- and BC-patients are given in Tables 1A
to C. +331 G and A1 are the common alleles, whereas +331 A
and A2 are the rare polymorphic alleles. Subsequently, age-
matched cohorts were analyzed for the presence of any
A. Romano et al. / Gynecologic Oncology 101 (2006) 287–295 289
association between disease and genotype. In order to compare
cohorts with similar mean ages, fixed age intervals of patients
and controls (19/23–50 years and 51–64 years) were used to
select cohorts. Hardy–Weinberg equilibrium was confirmed for
each polymorphism (+331 G vs. +331 A and A1 vs. A2,
respectively; Tables 1A to C) except for the group of women
being older than 51 years at the time breast cancer was diagnosed
(with respect to both polymorphisms; Table 1C, left column) and
for the group of women diagnosed with breast cancer and
regardless of the age at diagnosis (between 23 and 95 years, with
respect to +331 G/A only; Table 1C).
PR polymorphisms and ovarian cancer
The +331 A allele and predisposition to ovarian cancer
Although the percentage of women carrying the +331A allele
was higher among ovarian cancer patients than among healthy
controls (17.3% vs. 10.6%; Fig. 1a and Tables 1A and B), this
was not statistically significant. As epithelial ovarian cancer is
preferentially diagnosed in postmenopausal women (>50 years
of age), we used ‘‘age at diagnosis’’ as an additional criterion to
define subgroups of patients and controls. In women up to 50
years of age at time of diagnosis, the +331 A allele was
associated with an increased risk for OC (odds ratio, OR, 4.1;
confidence interval, CI, 1.2–13.9; P = 0.02; Fig. 1a). The
percentage of the +331 A carriers was not increased among
patients diagnosed with ovarian cancer between 51 and 64 years
of age compared to healthy controls (18.8% vs. 12.0%,
corresponding to an OR of 0.6 and CI of 0.2–2.3; P = 0.44).
The A2 allele and predisposition to ovarian cancer
The progins allele (A2) was detected more frequently in
ovarian cancer patients than in healthy controls (progins
carriers; 37.3% vs. 21.7%) corresponding to a significantly
increased risk for OC associated with A2 (OR 2.2; CI 1.2–3.7;
P < 0.01; Fig. 1b). When using ‘‘age at diagnosis’’ as an
additional criterion, the A2 allele was associated with an even
higher risk for OC (OR 3.2; CI 1.1–9.1; P = 0.02; Fig. 1b)
among women younger than 50 years of age. No increased risk
for OC was observed in women between 51 and 64 years of
age at diagnosis (OR 1.7; CI 0.7–3.9; P = 0.25).
PR polymorphisms and breast cancer
The +331 A allele and predisposition to breast cancer
The +331 A allele did not show any statistically significant
effect on the risk to develop BC (Fig. 2a). In the total BC
population, the presence of the +331 A allele was associated
with an OR of 1.1 and a CI of 0.7–1.6 (P = 0.82) compared to the
healthy controls. OR and CI were 1.3 (0.5–3.4; P = 0.56) and 0.5
(0.3–1.1; P = 0.10) in women younger than 50 (23–50) or older
than 51 (51–64) years of age at time of diagnosis, respectively.
However, the distribution of genotypes among women diag-
nosed with BC diverged significantly from the frequencies
expected under the Hardy–Weinberg assumption (P = 7.3 �10�10; Table 1C). This divergence was caused by the group of
women older than 51 years of age at the time of diagnosis (Table
1C; P = 1� 10�5). Women bearing at least one copy of the +331
A allele were less frequently seen (without statistical signifi-
cance) among patients diagnosed with breast cancer between 51
and 64 years of age than among healthy controls of the same age
(24 out of 216; 11.1% vs. 13 out of 69; 18.8%, respectively). In
contrast, homozygous carriers of the +331 A allele were more
frequent (without statistical significance) among BC patients
than in controls (5 out of 216; 2.3% vs. 1 out of 69; 1.4%; Tables
1A and C), which caused the observed deviation from the
Hardy–Weinberg equilibrium.
The A2 allele and predisposition to breast cancer
The frequency of A2 carriers was similar among controls
and BC patients (21.7% vs. 26.9%; Fig. 2b). In the total BC
population, the presence of the progins allele was associated
with an OR of 1.3 and a CI of 0.9–1.8 (P = 0.06) compared to
healthy controls. In women between 23 and 50 or 51 and 64
years at diagnosis, OR and CI were 1.4, 0.7–2.6 (P = 0.34) and
1.1, 0.6–2.0 (P = 0.77), respectively. A non-significant
increase in the number of women carrying two A2 alleles
was observed in patients diagnosed with breast cancer between
51 and 64 years of age (Tables 1A and C; P = 0.037).
Combination of +331 G/A and A1/A2 and ovarian and
breast cancer risk
Despite the low frequencies of the two rare polymorphic
alleles, we could analyze a sufficient number of samples to
compare healthy controls and patients with respect to two
possible combinations: women carrying no or either one of the
rare polymorphic alleles (i.e. either +331 A or A2; combination
A) and women being either homo- or heterozygous for both rare
polymorphic alleles (+331 A and A2; combination B). For this
analysis, data from 361 healthy controls, from 50 women
diagnosed with ovarian cancer and from 505 women diagnosed
with breast cancer were available (Table 2). Whereas 3 out of
361 healthy controls (0.8%) carried simultaneously the A2 and
+331 A alleles, this was detected in 3 out of 50 (6%) women
diagnosed with ovarian cancer and 16 out of 505 (3.2%) of
women diagnosed with breast cancer. Thus, women carrying
both rare alleles (+331 A and A2, with no respect to the number
of each allele in one woman) are at increased risk to develop
either ovarian or breast cancer (OR = 7.6, CI = 1.5–38.8, P <
0.01 and OR = 3.9, CI = 1.1–13.5, P = 0.02, respectively).
Discussion
Breast is the most frequent site for cancer onset in women
[14], and ovarian cancer has the highest mortality among
gynecological cancers [15]. Epidemiological and twin studies
have shown that the susceptibility to breast and ovarian cancer
is largely inherited [8,9] but only a subfraction of these patients
carry mutations in either the BRCA-1 or BRCA-2 gene [8,16].
Thus, other genetic factors may contribute to the risk for breast
and ovarian cancer [9]. Besides the estrogens, progesterone is
of major importance for differentiation and proliferation in
breast and ovarian tissues. The multiple actions of this
hormone require binding to the PR, which subsequently acts
A. Romano et al. / Gynecologic Oncology 101 (2006) 287–295290
as a nuclear transcription factor ([1] and references therein).
The presence of PR polymorphisms which alter either the
expression (+331 G/A) or the function (progins) of the
receptor, is likely to disturb not only the hormonal homeostasis
at the cellular level but also to modulate the risk for benign and
malignant gynecological disorders [2–4,7].
Although the frequencies of either polymorphisms in
women suffering from infertility, abortion, breast–ovarian or
endometrial cancer have been investigated by several groups
[2,6,12,13,17], results are still inconclusive. We have analyzed
the percentages of both polymorphic alleles and their combina-
tions among healthy women and women diagnosed with either
breast or ovarian cancer. All women enrolled in the present
study were of Caucasian origin and resided in the same region
of southern Germany.
Table 1
Overview of the age and genotype distribution among controls, patients and the co
A: Healthy controls
19 years–
Total (17 years–64 years) 19 years–
n 379 302
Mean age T SD 36.5 years T 11.9 years 33.0 years
+331 G/G 339 (89.4%) 275 (91.1%
+331 G/A 37 (9.8%) 25 (8.3%)
+331 A/A 3 (0.8%) 2 (0.7%)
All. frequency (+331 G) 0.94 0.95
p Hardy–Weinberg 0.057 0.1007
19 years–
19 years–
n 443 362
Mean age T SD 36.2 years T 12.0 years 32.7 years
A1/A1 347 (78.3%) 284 (78.5%
A1/A2 87 (19.6%) 71 (19.6%
A2/A2 9 (2.0%) 7 (1.9%)
All. frequency (A1) 0.88 0.88
p Hardy–Weinberg 0.2064 0.3079
B: OC patients
19 years–
Total (19 years–84 years) 19 years–
n 52 14
Mean age T SD 57.4 years T 11.5 years 44.7 years
+331 G/G 43 (82.7%) 10 (71.4%
+331 G/A 9 (17.3%) 4 (28.6%)
+331 A/A 0 0
All. frequency (+331 G) 0.91 0.86
p Hardy–Weinberg 0.4945 0.5329
19 years–
19 years–
n 67 15
Mean age T SD 57.5 years T 10.5 years 44.7 years
A1/A1 42 (62.7%) 7 (46.7%)
A1/A2 24 (35.8%) 8 (53.3%)
A2/A2 1 (1.5%) 0
All. frequency (A1) 0.81 0.73
p Hardy–Weinberg 0.2343 0.159
PR polymorphisms and ovarian cancer
Percentages of women carrying the +331 A or A2 allele
were higher (without statistical significance for +331 A and
with statistical significance for A2) among those diagnosed
with epithelial ovarian cancer (Fig. 1). Although epithelial
ovarian cancer preferentially occurs at postmenopausal age, 15
of the samples investigated for either polymorphism originated
from patients being between 19 and 50 years of age at
diagnosis, i.e. the cancer had developed already at premeno-
pausal age. In fact, in this subgroup of patients, percentages of
both rare polymorphic alleles were significantly higher than
among healthy controls. However, one should keep in mind,
that the number of patients diagnosed with OC below the age
of 50 years (15 women analyzed for progins and 14 for +331
horts selected for further analyses
64 years; n = 371 and 23 years–64 years; n = 332
50 years 23 years–50 years 51 years–64 years
263 69
T 8.9 years 35.3 years T 7.8 years 54.6 years T 3.0 years
) 239 (90.9%) 56 (81.2%)
22 (8.4%) 12 (17.4%)
2 (0.8%) 1 (1.4%)
0.95 0.9
0.0748 0.7019
64 years; n = 435 and 23 years–64 years; n = 368
50 years 23 years–50 years 51 years–64 years
295 73
T 9.3 years 35.4 years T 8.1 years 54.5 years T 3.0 years
) 236 (80.0%) 55 (75.3%)
) 52 (17.6%) 16 (21.9%)
7 (2.4%) 2 (2.7%)
0.89 0.86
0.0525 0.5327
64 years; n = 39
50 years 51 years–64 years
25
T 7.9 years 62.0 years T 8.8 years
) 22 (88.0%)
3 (12.0%)
0
0.94
0.7497
64 years; n = 52
50 years 51 years–64 years
37
T 7.8 years 57.2 years T 4.0 years
24 (64.9%)
13 (35.1%)
0
0.82
0.194863
C: BC patients
23 years–64 years; n = 267
Total (23 years–95 years) 23 years–50 years 51 years–64 years
n 535 51 216
Mean age T SD 64.8 years T 11.7 years 41.2 years T 5.7 years 58.9 years T 3.4 years
+331 G/G 476 (89.0%) 45 (88.2%) 192 (88.9%)
+331 G/A 48 (9.0%) 6 (11.8%) 19 (8.8%)
+331 A/A 11 (2.1%) 0 5 (2.3%)
All. frequency (+331 G) 0.93 0.94 0.93
p Hardy–Weinberg 7.3 � 10�10 0.6554 0.00001
23 years–64 years; n = 279
23 years–50 years 51 years–64 years
n 546 63 216
Mean age T SD 64.2 years T 12.1 years 41.4 years T 6.0 years 58.8 years T 3.5 years
A1/A1 399 (73.1%) 47 (74.6%) 159 (73.6%)
A1/A2 133 (24.4%) 15 (23.8%) 48 (22.2%)
A2/A2 14 (2.6%) 1 (1.6%) 9 (4.2%)
All. frequency (A1) 0.85 0.87 0.85
p Hardy–Weinberg 0.468 0.8741 0.0375
n: number of samples successfully genotyped and thus included in further analyses. SD = standard deviation. All. freq.: frequency of the most common allele. p
Hardy–Weinberg: Pearson’s goodness of fit chi-square for the Hardy–Weinberg assumption.
Table 1 (continued)
A. Romano et al. / Gynecologic Oncology 101 (2006) 287–295 291
G/A) is small. Furthermore, since age is an important risk
factor for ovarian cancer, the difference of mean age between
the subgroups analyzed (cases vs. controls; Tables 1A and B)
represents an extra limitation of the present study. Thus, these
results need to be confirmed within a larger study population.
Subsequently, we assessed the association of combined
+331 G/A and A1/A2 genotypes with ovarian cancer risk. The
combination of both rare polymorphic alleles in the same
individual (+331 A plus A2 regardless of homo- or heterozy-
gosity) increased the risk for the disease (Table 2) compared to
women carrying none or either one of the rare polymorphic
alleles (regardless of homo- or heterozygosity).
One recent study conducted on a population from North
Carolina and one from Australia [21] reported that the +331 A
allele decreased the risk for OC, and the reduction was more
pronounced for some subtypes of cancer. However, two larger
case-control studies in Caucasian women residing in New
Hampshire and eastern Massachusetts [22] and on women from
different ethnicity in the US [23] did not confirm these results.
The association between OC and progins (A2) has been
studied by several authors in Caucasian populations from
different origins, and contrasting results have been published
(all references have been recently reviewed by Modugno
[16]). Nevertheless, several authors suggested that the progins
allele (A2) increases the risk for ovarian cancer in homozy-
gous status [24,25]. Other groups presented evidence that
additional external (use of oral contraception) and/or internal
(presence of a mutated BRCA-1 and/or BRCA-2 gene)
effectors have a combined effect on OC risk [26,27]. The
same study assessing the risk for OC and +331 G/A
polymorphisms on Caucasian women [22] reported a reduced
risk for OC in A2 carriers, and this association was more
pronounced for some ovarian cancer subtypes. A second very
recent study conducted in the US considers several SNP
haplotypes in the PR gene [23], including one haplotype that
was identified by the A2 allele. An increased risk for OC in
homozygous carriers of the haplotype identified by the A2
allele was found. Nevertheless, although very complex and
exhaustive, this haplotype-based study did not clarify whether
the genetic risk for OC is associated with the haplotype
identified by the A2 allele or with a second larger haplotype,
which contained the A2 haplotype in many, but not in all
samples analyzed.
PR polymorphisms and breast cancer
When the frequencies of +331 A and A2 carriers were
analyzed among women diagnosed with BC (Fig. 2), no
association with the disease was found. Nevertheless, the
genotype distributions of +331 G/+331 A deviated significant-
ly (P < 0.001) from the Hardy–Weinberg equilibrium among
patients diagnosed with BC and this deviation seemed to be
restricted to women above 51 years of age at diagnosis (Table
1C for women between 51 and 64 years. The same was
observed in the group of 484 women between 51 and 95 years
of age, data not shown). This divergence was caused by a
higher frequency (without statistical significance) of women
harboring two copies of the +331 A allele and being diagnosed
with breast cancer after menopause compared to controls
(Tables 1A and C). Due to the low prevalence of women
carrying two +331 A alleles, no association analysis with BC
could be performed. Furthermore, a much larger study
population is required to investigate any possible recessive
behavior of the +331 A allele on the risk for BC. A slight
deviation of the A1/A2 allelic frequencies from the Hardy–
Weinberg equilibrium was also observed in women diagnosed
with breast cancer between 51 and 64 years of age (Table 1C),
but this divergence was lost if all women diagnosed with BC
Fig. 1. Association between +331 G/A (a) and progins (b) polymorphisms and OC. Percentages of women carrying either rare allele (+331 A or A2) are given among
controls ( ) and patients (h). * indicates a P value < 0.05; ** indicates a P value < 0.01. OR: odds ratio, CI = 95% confidence interval and p: Pearson’s goodness of
fit chi-square for the test for association (for the presence of the allele).
A. Romano et al. / Gynecologic Oncology 101 (2006) 287–295292
above the age of 51 (i.e. 51–95 years) were considered (not
shown). The combination of the +331 A and the A2 alleles in
one individual (regardless of homo- or heterozygosity)
significantly increased the risk for BC (Table 2). Although
the BC population was rather large, rare genotype combina-
tions, such as homozygosity for either one or both rare
polymorphic alleles, could not be studied. Furthermore, our
results require cautious interpretation, as the subgroups differ
with respect to the mean age.
The results of previous studies assessing BC risk and +331
G/A or A1/A2 polymorphisms in Caucasian populations are
inconclusive. De Vivo and co-workers concluded that post-
menopausal women carrying the +331 A allele had an
increased risk to develop BC [28]. It was not clear from this
study whether the specific cohort comprising postmenopausal
women was in Hardy–Weinberg equilibrium. A nested case-
control study on postmenopausal women [29] failed to confirm
this association.
Different Caucasian populations have been assessed to
investigate the effect of progins: one study describes a
protective action of the progins (A2) allele compared to the
more common A1 allele in women below 50 years of age [5],
other authors reported no association [18–20,30–32], or allelic
imbalance [33]. A recent meta-analysis considering all studies
mentioned above [10] hypothesized a protective effect of the
A2 allele against BC. The haplotype-based study, conducted by
Pearce and co-workers [23], confirmed this conclusion.
Conclusive remarks and recommendations
Although several studies have focussed on the progins
polymorphic allele and its association with ovarian and breast
Fig. 2. Association between +331 G/A (a) and progins (b) polymorphisms and BC. Percentages of women carrying either rare allele (+331 A or A2) are given among
controls ( ) and patients (h). OR: odds ratio, CI = 95% confidence interval and p: Pearson’s goodness of fit chi-square for the test for association (for the presence of
the allele).
A. Romano et al. / Gynecologic Oncology 101 (2006) 287–295 293
cancers, this is the first study assessing the association between
either the +331 G/A polymorphism or the combinations of
+331 G/A and progins with breast or ovarian cancer in a
Caucasian European population.
Published results describing frequencies of progins among
healthy women and patients are inconclusive (for a compre-
hensive review on OC association studies, refer to Modugno
Table 2
Combination of +331 G/A and A1/A2 and ovarian and breast cancer risk
Total n Combination A carrying one
polymorphism (progins or +331A)
Com
pol
Controls 361 358 (99.2%) 3
OC patients 50 47 (94.0%) 3
BC patients 505 489 (96.7%) 16
OR: odds ratio; CI: confidence interval; P: P value.
[16]) and variations exist in the observed allelic frequencies
between different studies. Also very recent and complex
studies in which corrections for many possible confounders
were introduced led to contrasting conclusions [21–23]. In
some cases, the size and the ethnicity of the populations studied
can explain the differences in the allelic frequency of controls
and the non-conclusive results obtained with respect to any
bination B carrying both
ymorphisms (progins and +331A)
OR CI P
(0.8%)
(6.0%) 7.6 1.5–38.8 <0.01
(3.2%) 3.9 1.1–13.5 0.02
A. Romano et al. / Gynecologic Oncology 101 (2006) 287–295294
association with cancer risk (and other disturbances). Further-
more, one should keep in mind that the technique used for
genotyping might be error prone as well, especially when DNA
extracted from paraffin-embedded tissues has to be used for
genotyping. As the progins status can be judged from
investigation of one aberration, PCR analysis of intron G
appears to be the most convenient technique. The Alu insertion
increases the size of the amplicon by 320 bp, and this
difference is clearly detectable after agarose gel electrophore-
sis. However, the amplification of such a fragment is hard to
achieve with DNA originating from paraffin-embedded tissue.
Therefore, we developed and validated a combination of
nested-PCR and RFLP analysis of exon 4 and recommend
this technology for further investigations on the progins
polymorphism.
Although the age-adjusted group of ovarian cancer patients
was rather small, our data suggest that the presence of either the
+331 A or the A2 allele is associated with an increased risk for
epithelial ovarian cancer in women younger than 50 years of
age. To further investigate the effect of the +331 A on the risk
for breast cancer in postmenopausal women, a larger popula-
tion-based study is necessary.
Acknowledgments
We would like to thank K. Korner from the Ulm University
Medical Center and L. Walz, B. Seidl and A. Ohm from the
Freiburg University Medical Center from providing tissue
material and helping in genomic DNA extraction.
References
[1] Li X, O’Malley BW. Unfolding the action of progesterone receptors.
J Biol Chem 2003;278(41):39261–4.
[2] De Vivo I, Huggins GS, Hankinson SE, Lescault PJ, Boezen M, Colditz
GA, et al. A functional polymorphism in the promoter of the progesterone
receptor gene associated with endometrial cancer risk. Proc Natl Acad Sci
U S A 2002;99(19):12263–8.
[3] Rowe SM, Coughlan SJ, McKenna NJ, Garrett E, Kieback DG, Carney
DN, et al. Ovarian carcinoma-associated TaqI restriction fragment length
polymorphism in intron G of the progesterone receptor gene is due to an
Alu sequence insertion. Cancer Res 1995;55(13):2743–5.
[4] McKenna NJ, Kieback DG, Carney DN, Fanning M, McLinden J, Headon
DR. A germline TaqI restriction fragment length polymorphism in the
progesterone receptor gene in ovarian carcinoma. Br J Cancer 1995;
71(3):451–5.
[5] Wang-Gohrke S, Chang-Claude J, Becher H, Kieback DG, Runnebaum
IB. Progesterone receptor gene polymorphism is associated with
decreased risk for breast cancer by age 50. Cancer Res 2000;60(9):
2348–50.
[6] Schweikert A, Rau T, Berkholz A, Allera A, Daufeldt S, Wildt L.
Association of progesterone receptor polymorphism with recurrent abor-
tions. Eur J Obstet Gynecol Reprod Biol 2004;113(1):67–72.
[7] Agoulnik IU, Tong X-W, Fischer D-C, Korner K, Atkinson NE, Edwards
DP, et al. A germline variation in the progesterone receptor gene increases
transcriptional activity and may modify ovarian cancer risk. J Clin
Endocrin Metabol 2004;89(12):6340–7.
[8] Sutcliffe S, Pharoah PD, Easton DF, Ponder BA. Ovarian and breast
cancer risks to women in families with two or more cases of ovarian
cancer. Int J Cancer 2000;87(1):110–7.
[9] Antoniou AC, Pharoah PD, McMullan G, Day NE, Ponder BA, Easton D.
Evidence for further breast cancer susceptibility genes in addition to
BRCA1 and BRCA2 in a population-based study. Genet Epidemiol
2001;21(1):1–18.
[10] Dunning AM, Healey CS, Pharoah PD, Teare MD, Ponder BA, Easton
DF. A systematic review of genetic polymorphisms and breast cancer risk.
Cancer Epidemiol, Biomarkers Prev 1999;8(10):843–54.
[11] Wright D, Manos M. Sample preparation from paraffin-embedded
tissues. In: Innis M, Gelfand D, Sninsky J, White T, editors. PCR
protocols—A guide to methods and applications. San Diego’ Academic
Press; 1990. p. 153–8.
[12] Pijnenborg JMA, Romano A, Dam-de Veen GC, Dunselman GAJ,
Fischer D-C, Groothuis PG, et al. Aberrations in the progesterone
receptor gene and the risk of recurrent endometrial carcinoma. J Pathol
2005;205:597–605.
[13] Pisarska MD, Carson SA, Casson PR, Tong X, Buster JE, Kieback DG. A
mutated progesterone receptor allele is more prevalent in unexplained
infertility. Fertil Steril 2003;80(3):651–3.
[14] Parkin DM, Pisani P, Ferlay J. Global cancer statistics. CA Cancer J Clin
1999;49(1):33–64.
[15] Scully SR. Pathology of ovarian carcinoma. In: Piver SM, editor. Ovarian
malignancies. Edinburg’ Dhurcill Livingstone; 1987. p. 72–95.
[16] Modugno F. Ovarian cancer and polymorphisms in the androgen and
progesterone receptor genes: a HuGE review. Am J Epidemiol 2004;
159(4):319–35.
[17] Wieser F, Schneeberger C, Tong D, Tempfer C, Huber JC, Wenzl R.
PROGINS receptor gene polymorphism is associated with endometriosis.
Fertil Steril 2002;77(2):302–9.
[18] Fabjani G, Tong D, Czerwenka K, Schuster E, Speiser P, Leodolter S,
et al. Human progesterone receptor gene polymorphism PROGINS and
risk for breast cancer in Austrian women. Breast Cancer Res Treat
2002;72(2):131–7.
[19] Donaldson CJ, Crapanzano JP, Watson JC, Levine EA, Batzer MA.
PROGINS Alu insertion and human genomic diversity. Mutat Res
2002;501(1–2):137–41.
[20] Spurdle AB, Hopper JL, Chen X, McCredie MR, Giles GG, Venter DJ,
et al. The progesterone receptor exon 4 Val660Leu G/T polymorphism
and risk of breast cancer in Australian women. Cancer Epidemiol,
Biomarkers Prev 2002;11(5):439–43.
[21] Berchuck A, Schildkraut JM, Wenham RM, Calingaert B, Ali S,
Henriott A, et al. Progesterone receptor promoter +331A polymor-
phism is associated with a reduced risk of endometrioid and clear cell
ovarian cancers. Cancer Epidemiol, Biomarkers Prev 2004;13(12):
2141–7.
[22] Terry KL, De Vivo I, Titus-Ernstoff L, Sluss PM, Cramer DW. Genetic
variation in the progesterone receptor gene and ovarian cancer risk. Am J
Epidemiol 2005;161(5):442–51.
[23] Pearce CL, Hirschhorn JN, Wu AH, Burtt NP, Stram DO, Young S, et
al. Clarifying the PROGINS allele association in ovarian and breast
cancer risk: a haplotype-based analysis. J Natl Cancer Inst 2005;97(1):
51–9.
[24] Tong D, Fabjani G, Heinze G, Obermair A, Leodolter S, Zeillinger R.
Analysis of the human progesterone receptor gene polymorphism
progins in Austrian ovarian carcinoma patients. Int J Cancer 2001;
95(6):394–7.
[25] Whittemore AS, McGuire V. Analysis of the human progesterone receptor
gene polymorphism PROGINS in Austrian ovarian cancer patients. Int J
Cancer 2002;101(2):202 [author reply 203].
[26] Lancaster JM, Wenham RM, Halabi S, Calingaert B, Marks JR,
Moorman PG, et al. No relationship between ovarian cancer risk and
progesterone receptor gene polymorphism in a population-based, case-
control study in North Carolina. Cancer Epidemiol, Biomarkers Prev
2003;12(3):226–7.
[27] Runnebaum IB, Wang-Gohrke S, Vesprini D, Kreienberg R, Lynch
R, Moslehi R, et al. Progesterone receptor variant increases ovarian
cancer risk in BRCA1 and BRCA2 mutation carriers who were
never exposed to oral contraceptives. Pharmacogenetics 2001;11(7):
635–8.
[28] De Vivo I, Hankinson SE, Colditz GA, Hunter DJ. A functional
A. Romano et al. / Gynecologic Oncology 101 (2006) 287–295 295
polymorphism in the progesterone receptor gene is associated with an
increase in breast cancer risk. Cancer Res 2003;63(17):5236–8.
[29] Feigelson HS, Rodriguez C, Jacobs EJ, Diver WR, Thun MJ, Calle EE.
No association between the progesterone receptor gene +331G/A
polymorphism and breast cancer. Cancer Epidemiol, Biomarkers Prev
2004;13(6):1084–5.
[30] De Vivo I, Hankinson SE, Colditz GA, Hunter DJ. The progesterone
receptor Val660->Leu polymorphism and breast cancer risk. Breast
Cancer Res 2004;6(6):R636–9.
[31] Lancaster JM, Berchuck A, Carney ME, Wiseman R, Taylor JA.
Progesterone receptor gene polymorphism and risk for breast and ovarian
cancer. Br J Cancer 1998;78(2):277.
[32] Manolitsas TP, Englefield P, Eccles DM, Campbell IG. No association of a
306-bp insertion polymorphism in the progesterone receptor gene with
ovarian and breast cancer. Br J Cancer 1997;75(9):1398–9.
[33] Garrett E, Rowe SM, Coughlan SJ, Horan R, McLinden J, Carney DN,
et al. Mandelian inheritance of a TaqI restriction fragment length
polymorphism due to an insertion in the human progesterone receptor
gene and its allelic imbalance in breast cancer. Cancer Res Ther Control
1995;4:217–22.
