Upload
others
View
1
Download
0
Embed Size (px)
Citation preview
University of Groningen
Hormonal and molecular aspects of endometrioid endometrial cancerJongen, Vincentius Hubertus Willibrordus Maria
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite fromit. Please check the document version below.
Document VersionPublisher's PDF, also known as Version of record
Publication date:2008
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):Jongen, V. H. W. M. (2008). Hormonal and molecular aspects of endometrioid endometrial cancer. [S.l.]:[s.n.].
CopyrightOther than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of theauthor(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policyIf you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediatelyand investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons thenumber of authors shown on this cover page is limited to 10 maximum.
Download date: 13-10-2020
Hormonal and molecular aspects of
endometrioid endometrial cancer
V.H.W.M. Jongen
Hormonal and molecular aspects of endometrioid endometrial cancer
Vincent Jongen
STELLING EN
Behorend bij het proefschrift
Hormonal and molecular aspects of endometrioid endometrial cancer
van
V.H.W.M. Jongen
Groningen
12 november 2008
1. Aromatase is de grote gemene deler in de carcinogenese van het endometriumen mammacarcinoom (dit proefschrift).
2. De mate van ovariele stroma hyperplasie en de lokale productie van androgenen in het ovarium zijn aan elkaar gerelateerd in het endometrioid endometriumcarcinoom ( dit proefschrift ).
3. Gelijktijdige aanwezigheid van zowel androgenen als het enzym aromatase in het endometrium ondersteunt de hypothese dat lokale aromatase activiteit een rol speelt bij het ontstaan van endometrioid endometriumcarcinoom ( dit proefschrift).
4. Een zelfde relatie tussen aromatase en HER-2/neu bij endometrioid endometrium- en mammacarcinoom suggereert een vergelijkbare regulatie van aromatase expressie.
5. P53 positief endometrioid endometriumcarcinoom vertoont een biologisch gedrag dat lijkt op dat van non-endometrio'id endometriumcarcinoom.
6. Bepaling van de oestrogeen en progesteron receptor alpha is zinvol bij endometrioid endometriumcarcinoom ten behoeve van een meer op de individuele patient toegespitst behandelplan.
7. Voor de meeste vrouwen geldt dat hun eierstokken vrienden zijn, het is een bittere ervaring dat het opeens ook vijanden kunnen zijn.
8. Een ziekenhuis is een 24-uurs bedrijf en geen digitale papierwinkel.
9. Bezuiniging leidt niet alleen tot een afname van de thuiszorg maar tevens tot een toename hiervan (R. van Doom).
10. Een reeds gepromoveerde echtgenote is de beste prikkel om een proefschrift te voltooien, waarbij geldt: "twee promovendi op een kussen, daar komt niemand tussen".
11. In een goed huwelijk mag het best eens donderen. Als na elke donderdag maar weer een vrijdag komt (bron: onbekend).
C1..1,trJ!r. U
tvleGt',. he M Bi� Ii ,.lirck C
Grt.:11ingi:n G
L_ ...
Colofon
Publication of this thesis was generously supported by Abbott Laboratorie� Bayer Schering Pharma, GlaxoSmithKline, Merck Serono, Novartis Oncology, Organon Nederland BV ( onderdeel Schering-Plough Corporation), and Sanofi Pasteur MSD, and is gratefully acknowledged.
Hormonal and molecular aspects of endometrioid endometrial cancer Thesis Rijksuniversiteit Groningen
Cover: immunohistochemical staining for aromatase expression in endometrioid endometrial cancer, using monoclonal aromatase 677 antibodies.
© V.H.W.M. Jongen, 2008
ISBN: 978-90-367-3543-8
Printed by: Drukkerij Bari et, Ruin en, The Netherlands
RIJKSUNIVERSITEITGRONINGEN
Hormonal and molecular aspects of endometrioid endometrial cancer
Proefschrift
ter verkrijging van het doctoraat in de Medische Wetenschappen
aan de Rijksuniversiteit Groningen op gezag van de
Rector Magnificus, dr. F. Zwarts, in het openbaar te verdedigen op
woensdag 12 november 2008 om 14.45 uur
door
Vincentius Hubertus Willibrordus Maria Jongen
geboren op 23 mei 1963 te Valkenswaard
Cl!nttale Medische
Bibliotheek Groningen
-�-
u M C G
Promotores: Prof. Dr. A.G.J. van der Zee Prof. Dr. M.J. Heineman Prof. Dr. H. Hollema Prof. Dr. H.W. Nijman
Beoordelingscommissie: Prof. Dr. J.G. Aalders Prof. Dr. J.H.H.Thijssen Prof. Dr. P.H.B. Willemse
Opgedragen aan onze lieve Renske
Contents
Chapter 1 General outline of the thesis and aims of the studies.
Chapter 2 General introduction.
The postmenopausal ovary as an androgen-producinggland;
hypothesis on the etiology of endometrial cancer.
9
Maturitas 2002;43: 77-85. 15
Chapter 3 Aromatase in the context of breast and endometrial cancer.
A review.
Minerva Endocrinologica 2006;31:47-60.
Chapter 4 Ovarian stromal hyperplasia and ovarian vein steroid levels
in relation to endometrioidendometrial cancer.
35
British Journal of Obstetrics and Gynaecology 2003; 110:690-5. 67
Chapter 5 Is aromatase P450 involved in pathogenesis of endometrioid
endometrial cancer?
International Journal of Gynecological Cancer 2005; 15:529-36. 87
Chapter 6 Aromatase, COX-2, HER-2/neu and P53 as prognostic factors in
endometrioid endometrial cancer.
Accepted for publication in the International Journal of
Gynecological Cancer 107
Chapter 7 Expression of estrogen and progesterone receptors alpha and beta in
a large cohort of patients with endometrioid endometrial cancer.
Submitted. 131
Chapter 8 Summary and future perspectives. 153
Chapter 9 Samenvatting en overwegingen voor de toekomst. 165
Authors and affiliations 176
Dankwoord 179
List of publications 184
Curriculum vitae 187
General outline of the thesis and aims of the studies
9
10
This thesis concerns the expression and prognostic value of various hormones
and molecular markers playing a role in endometrioid endometrial cancer.
Especially we were interested in the enzyme aromatase, its expression and
(prognostic) role in endometrioid endometrial cancer. Endometrial cancer is the
most common genital tract malignancy in women. Two major histological types
can be distinguished; endometrioid endometrial cancer, and nonendometrioid
endometrial cancer including high-grade malignancies like serous papillary and
clear cell carcinoma. The most common form, involving more than 75% of all
cases of endometrial cancer is endometrioid carcinoma. Generally it is believed
that continuous estrogenic hyperstimulation of the postmenopausal endometrium
gives rise to endometrial tissue hyperplasia followed by atypical hyperplasia.
Eventually endometrioid endometrial cancer may develop. These estrogens
mainly originate from peripheral conversion in body fat of androgens,
originating from the adrenal glands and ovaries. The enzyme complex
cytochrome P450 aromatase, main subject of this thesis, plays a key role in the
(peripheral) conversion of androgens into estrogens.
However, it is possible that various hormonal pathways exist contributing to the
early development or promotion of endometrioid endometrial cancer. In this
context it cannot be excluded that increased production of androgens by
postmenopausal ovaries leads to increased availability of aromatizable
androgens, which may contribute to estrogen synthesis in the endometrial tissue
as well. Following the local conversion of ovarian androgens into estrogens, a
reaction catalyzed as well by the cytochrome P450 aromatase, estrogens can
function as a local mitogenic factor contributing to the development of
endometrial cancer. We consider the local availability of androgens and the local
activity of aromatase in endometrial tissue relevant for this process.
11
Shortly, in this thesis we will demonstrate that androgens are present in the local
hormonal environment of the uterus, and that aromatase expression is present in
the endometrial tissues. If so, this means that the local ingredients are available
to facilitate another pathway contributing to the development or promotion of
endometrioid endometrial cancer. In this view it may be possible that different
hormonal pathways exist, possibly next to or promoting each other, ultimately
leading to the development of endometrioid endometrial cancer.
Our studies will examme the following hypothesis, which will be further
elaborated in the Introduction (Chapter 2):
"The local availability of androgens in the utero-ovarian circulation, possibly
related to the amount of ovarian stromal hyperplasia, and the local availability of
aromatase activity in the endometrial tissue, support the hypothesis that ovarian
androgens contribute to the development of endometrioid endometrial cancer
through aromatase driven local conversion into estrogens."
Are there similarities between estrogen dependent breast and endometrial cancer?
The presence of aromatase expression appeared to be a relevant factor in the
development of breast cancer, and nowadays treatment with aromatase inhibitors
has found an essential place in breast cancer treatment. Breast cancer aspects
will be discussed in relation to endometrioid endometrial cancer, since both
disorders are partly estrogen dependent ( Chapter]).
To verify our hypothesis the following questions were asked:
1. Is the degree of ovarian stromal hyperplasia related to the presence of
endometrioid endometrial cancer? In a retrospective study, the degree of ovarian
stromal hyperplasia of patients with and without endometrioid endometrial cancer
was compared ( Chapter 4).
12
2. Are steroid levels in the utero-ovarian circulation related to the degree of
ovarian stromal hyperplasia, and to the presence or absence of endometrioid
endometrial cancer? In a prospective study steroid levels in the utero-ovarian
circulation of patients with and without endometrioid endometrial cancer were
compared to the degree of ovarian stromal hyperplasia ( Chapter 4).
3. Are androgen levels in the utero-ovarian circulation in endometrioid
endometrial cancer related to aromatase activity in endometrial and body fat
tissues? A prospective study was performed, in which aromatase activities in
endometrial tissue were related to the utero-ovarian venous steroid levels, using
tritium water release assays ( Chapter 5).
4. What is the expression and prognostic value of aromatase in endometrioid
endometrial cancer? Two retrospective studies were performed in archival paraffin
embedded tumour tissue, using tissue microarray analysis. We determined the
expression and prognostic value of the molecular markers aromatase in both
cancer and stromal cells, next to the prognostic value of COX-2, HER-2/neu and
P53 in a large cohort of endometrioid endometrial cancer patients. In a second
study using the same cohort of endometrial cancer patients, we assessed the
prognostic value of the estrogen and progesterone receptors alpha and beta in
endometrioid endometrial cancer. Expression of receptors and molecular
markers were determined by immunostaining and related to classical clinico
histopathological parameters, in addition to disease-free and overall survival.
( Chapter 6 and 7).
In Chapter 8 and 9 our findings will be summarized and discussed in relation to
each other.
13
14
2
The postmenopausal ovary as an androgen
producing gland; hypothesis on the etiology
of endometrial cancer
VHWM Jongen
AVSluijmer
MJHeineman
Maturitas 2002;43:77-85
15
Abstract
Postmenopausal estrogens originate from the peripheral conversion of androgens
which are produced by the adrenal glands and the ovaries. Estrogens are
considered to contribute to the neoplastic development of endometrium.
Hyperplasia of ovarian stroma is associated with an increased androgen production
by the ovaries and with the development of endometrial pathology. We
hypothesize that, in cases of endometrial pathology, an increased production of
aromatizable androgens by postmenopausal ovaries will lead to elevated
prehormone availability for estrogen formation in utero. Following the conversion
of ovarian androgens, a reaction catalyzed by the cytochrome P450 aromatase,
estrogens may function as a local mitogenic factor eventually leading to the
development of endometrial cancer. We consider the local availability of
androgens and the local activity of aromatase relevant for this process. If this
hypothesis proves to be right it may give rise to the introduction of aromatase
inhibitors in treatment strategies of hormone dependent endometrial malignancies.
16
Introduction
Some endometrial cancers are considered to be estrogen-dependent. After
menopause estrogen biosynthesis is mainly peripheral. In fat and other tissues
androgens from the adrenal glands and the ovaries are converted to oestrogens:
androstenedione and estrone are converted to testosterone and estradiol
respectively under the influence of 17B OH-dehydrogenase, and androstenedione
to estrone and testosterone to estradiol by the enzyme complex cytochrome P450
aromatase. Continuous hyperstimulation of the endometrium by ( endogenous and
exogenous) estrogens is supposed to be the causative mechanism eventually
leading to cancer or tumour progression. A positive correlation was reported
between postmenopausal estrogen levels and endometrial cancer.1-4 Obesity leads
to increased adrenal secretion of androgens and increased peripheral conversion
of androgens to estrogens in adipose cells.5•6 Possibly androgen production by the
postmenopausal ovary relates to estrogenic stimulation of endometrial tissues.1- 12
Apart from estrogens of peripheral origin local activation or failure of the
inhibition of aromatase activity in the endometrium may promote estrogenic
stimulation of the endometrium and eventually give rise to neoplastic development
by local conversion of androgens to estrogens.11•13•14 Indeed aromatase activity was
demonstrated in endometrial tumours, whereas its expression was not detected in
disease-free endometrium!1•15 Whether systemically delivered or locally produced,
elevated estrogen levels will promote the growth of steroid-responsive tissues:
speculatively these estrogens may result from an increased estrone conversion
from androstenedione in adipose tissues, or from an increased ovarian and adrenal
androgen secretion. In this article, we discuss the available evidence that a causal
relationship may exist between elevated ovarian androgen production after
menopause, the presence of ovarian stromal hyperplasia, aromatase activity and
the development of hormone-dependentendometrial cancer.
17
Postmenopausal ovarian function
During recent years the concept was developed of the postmenopausal ovary as a
viable androgen-producinggland.1-10,12,16-20 Apart from changes in morphology the
ovary also appears subject to changes in function, producing 40% respectively
20% of the total amount of testosterone and androstenedione in postmenopausal
years.7•2 1 Steroid measurements across the postmenopausal ovarian gland showed
that there exists a testosterone gradient across the postmenopausal ovary .10, 12, 19•22 In
order to study hormone production of the postmenopausal ovary techniques were
developed for the collection of utero-ovarian vein blood.23 As the amount of
ovarian stroma increased, the androgen content of the ovarian vein increased .17•18
Almost all studies in postmenopausal women comparing utero-ovarian vein and
peripheral vein levels of estrone, 17B-estradiol, androstenedione and testosterone,
demonstrate that the utero-ovarian androgen level is elevated when compared with
the peripheral serum concentration.8-1 3•17-19•2 1-25 Testosterone seems to be the more
important androgen secreted by the postmenopausal ovary. After oophorectomy
testosterone and androstenedione levels in postmenopausal women indeed
dropped significantly.21•25•26 Postmenopausal ovarian testosterone production
appears to be at least partly gonadotrophin (LH) dependent, since pituitary down
regulation with a GnRH-agonist resulted in a significant decrease of testosterone
levels.25•27 Elevated levels of luteinizing hormone (LH) were demonstrated in
postmenopausal women with endometrial cancer, which may possibly contribute
to increased ovarian androgen production.28 Receptors for LH and to a lesser
extent follicle stimulating hormone (FSH) are found in ovarian stromal cells.29
Ovarian androgen secretion after menopause could be stimulated by human
chorionic gonadotropin and LH. 30•31 Using an aromatase-specific monoclonal
antibody, the distribution of the aromatase cytochrome P450 enzyme was studied
in human ovaries. It was found that in premenopause ovarian follicles but not
18
stromal cells stained for the aromatase, whereas aromatase immunostaining was
present in three of seven postmenopausal ovaries. Rather conflicting results were
found recently in a small but interesting study in postmenopausal women with
complete adrenal insufficienci2 In the absence of adrenal steroids after
menopause, no circulating androgens were found, suggesting the climacteric ovary
not being a source of androgens at all. No hormonal changes were noted after
administration of human chorionic gonadotropin. Also no FSH and LH receptors
or aromatase P450 enzyme could be demonstrated in the postmenopausal ovaries.
A better understanding of postmenopausal ovarian endocrine function will help to
reveal more of the etiology of endometrial carcinoma.
After menopause the ovaries decrease in size, ovarian follicles disappear and the
ovary consists mainly of stromal cells. Histological distinction can be made
between postmenopausal ovaries with minimal stroma ( atrophic ovaries) and
ovaries with an increased volume of stroma (stromal hypeiplasia). It is suggested
that the extent of stromal hyperplasia correlates well with postmenopausal ovarian
steroid production, which would mean that any evaluation of the hormonal pattern
in postmenopausal women should also take into account the stromal characteristics
of the ovary.10•17• 18•33 Marked stromal hyperplasia cannot be considered a
premalignant condition as it may occur as well in benign as in malignant
conditions. We report a significant increase of ovarian stromal hyperplasia in case
of endometrioid endometrial cancer [submitted]. Contrary to our results Lucisano
found no correlation between the degree of ovarian stromal hyperplasia and
endometrioid endometrial cancer.17 Comparison with other studies is hampered by
the fact that the exact histological type of endometrial cancer most often is not
mentioned. Also widely varying percentages of endometrial cancer in these studies
prevent statements on odds ratio for developing endometrioid endometrial cancer
in the presence or absence of ovarian stromal hyperplasia.18•3 1
•34
•35
19
Endometrial cancer
Estrogens
Generally, two different clinicomorphologic forms of endometrial cancer have
been described: a low-grade malignancy related to unopposed estrogenic
stimulation, and a second, more virulent estrogen- independent form.36•37 This
second pathogenetic type, denominated nonendometrioid endometrial cancer,
includes poorly differentiated and high-grade malignancies like serous papillary
and clear cell carcinoma. Estrogen or progesterone receptors are rarely found in
these poorly differentiated, clear cell, and serous papillary uterine cancers, which
also do not respond to hormonal therapy !8•39 The former type is called endome
trioid endometrial carcinoma, and arises mainly in women with obesity,
hyperlipidemia, hypertension or diabetes and signs of hyperestrogenism:
anovulatory uterine bleeding, nulliparity, early age at menarche and late onset of
menopause, and a background of hyperplasia of the stroma of ovaries or
endometrium. Obesity increases the risk, the postulated mechanism being incre
ased conversion of androgens to estrogens in the adipose tissue. Sherman
demonstrated higher circulating estrogen levels in women with endometrioid
endometrial cancer than in women with serous carcinoma.40 Also, obesity was
associated with a decrease of sex hormone binding globulin production in the
liver and an increase of available free estradiol.16 The concept of continuous
estrogenic hyperstimulation of the postmenopausal endometrium was developed,
in which endometrioid endometrial cancer gradually develops from atypical
hyperplasia. In this view endometrial cancer is a disease of the proliferative effects
of estrogen on the endometrium, especially when unopposed by the differentiating
action of progesterone. Estrogenic stimulation drives cell proliferation and by the
increase of the number of cell divisions the risk of random DNA replication errors
(activation of oncogenes, inactivation of tumour-suppresion genes) increases.
Depending on the kind of genetic damage a malignant phenotype may arise with
20
progressive growth under the influence of estrogens. Thus, estrogens act as tumour
promotors but not as carcinogens. Initially anti-hormone therapy by e.g. progestins
is able to slow down the process of progression of cancer via different anti
estrogenic mechanisms, including a reduction in estrogen receptors and an
increase in metabolic inactivation of estrogens, which suggests that interruption of
continuous estrogenic stimulation may disturb the process of neoplastic
deterioration, until a 'point of no return' is reached.5'36
'41 Progestins also counteract
the development of hyperplasia by inducing cellular differentiation.
Postmenopausal women have low estradiol levels and may be sensitive to even
small increases in estrogen levels, which may enhance mitotic rate and thus
increase the risk of malignant degeneration.16 Under continuous estrogenic
stimulation nearly one quarter of atypical hyperplasias progress to well
differentiated endometrial carcinoma, indicating that atypical hyperplasia is a
precursor of endometrioid carcinoma.42 Endometrioid carcinoma is the most
common form of endometrial cancer, explaining for more than 75% of all cases.
These tumours, by definition, do not contain areas showing more than I 0% of
squamous, mucinous, or clear cell differentiation. Estrogen receptor generally
appears positive. Also, mixed endometrioid/serous carcinomas exist, which are
associated with endometrial hyperplasia in 46% of cases. Finally, 5-8% of serous
endometrial cancers are associated with endometrial hyperplasia.43 The
antiestrogen tamoxifen, used a.o. as an alternative endocrine treatment for
endometrial cancer, was associated with an increased risk of endometrial cancer
after menopause, so this strategy was left.44 It was suggested that tamoxifen
functions as an estrogen agonist in the estrogen-poor environment existing after
menopause, dependent on cumulative dose and duration of use.45
21
Androgens
Nagamani reported significantly higher ovarian venous levels of both
androstenedione and testosterone in thin women with endometrioid
adenocarcinoma of the endometrium compared to those in obese women with can
cer.46 These findings could imply that an increase in prehormone availability from
the ovaries plays a role in the development of endometrioid cancer in thinner
women, who are also known to be more prone to nonendometrioid cancer. An
interrelation may exist between postmenopausal ovarian morphology, local
androgen production and hyperplastic and malignant diseases of the endometrium.
Remarkably the testosterone levels in the utero-ovarian vein blood samples from
women with endometrial carcinoma were significantly elevated as compared with
the utero-ovarian vein levels from women without endometrial carcinoma, without
significant increases in the peripheral levels of these androgens.47·48 Two large
epidemiologic studies report an elevation of plasma androstenedione level in case
of endometrial cancer, which also may reflect increased precursor availability for
peripheral conversion.2·4 In vitro the ovarian stroma of postmenopausal women
with endometrial neoplasm released significantly more androstenedione,
testosterone and dehydroepiandrosterone than that of women without cancer.49
Stromal hyperplasia in the postmenopausal ovary is related with an increase of the
ovarian vein levels of androstenedione and testosterone?· 18·33 Increased proouction
of aromatizable androgens by the ovarian stroma may lead to increased pre
hormone availability for estrogen formation from peripheral conversion in fat
tissue. Alternatively, the demonstration of androgen receptors in the endometrium
by immuno-histochtmical techniques suggests a possible direct role of androgens
in causing endometrial cancer.50-52 However, it was suggested as well that intra
tumoral aromatase is able to convert circulating androgens derived from adrenal
cortex and/ or ovary to estrogens, and thus give rise to a growth promoting effect
on endometrium. In a in vivo mouse model androgen excess had little growth-
22
promoting effect on a well-differentiated hormone-responsive human endometrial
cancer line.53 Possibly low androgen receptor expression influenced the results.
Aromatase
Aromatase activity
In postmenopausal women approximately 2% of circulating androstenedione is
converted to estrone, which can be further converted to the biologically more
potent estradiol by 17B OH-dehydrogenase. Aromatase is the key enzyme for
estrogen production. The conversion of androgens to estrone is catalyzed by
aromatase cytochrome P450 (p450arom, product of the CYP19 gene, which is
involved in steroidgenesis and steroid metabolism). Aromatase is responsible for
binding the C 19 steroid substrate and catalyzing the series of reactions leading to
formation of the phenolic A ring characteristic of estrogens. The enzyme
aromatase is found in ovarian granulosa cells, adipose tissue, placental
syncytiotrophoblast, adipose and skin fibroblasts, vascular endothelial and aortic
smooth muscle cells, bone and brain. Conversion of androstenedione to estrone
increases as a function of age, obesity and aromatase activity in extraglandular
(adipose) tissue.54 These peripheral tissues, especially adipose cells, are dependent
for substrate circulating androgens from primarily the adrenal cortex and to a
lesser extent from the ovaries.
Estrogens may thus originate in extragonadal sites where they act locally in a
paracrine or intracrine way and may only partly escape into the circulation
without significantly affecting circulating levels.55 This means that the total
amount of estrogens produced from these sites may be small, but the local tissue
levels may be high with consequently significant biological influences, as the
estrogenic effects of locally produced steroids add to and amplify the effects of
systemic estrogens. Taking into account the above, it is hypothesized that, apart
23
from extraglandular aromatase activity in adipose tissue, aromatase activity in the
endometrium may play a role in malignant degeneration.1 1 • 13 Earlier research by
Tseng et al on testosterone aromatization in the human endometrium showed
estrogen production in these tissues.56 Increased aromatization activity might result
in increased estrogen concentrations in neoplastic endometrium. Intratumoral
estrogens derived from in situ aromatization may function as an autocrine growth
and mitogenic factor, regardless of serum concentration of estrogens. Indeed
aromatase activity was found in endometrial malignancy, but not in normal
endometrial tissue. 1 1•56-62 As endometrial carcinoma tissue consists of different cell
types ( epithelial, stromal, malignant cells), localization of aromatization sites in
endometrial cancer is of importance. Aromatase immunoactivity was found
mainly in stromal cells of endometrioid endometrial carcinoma, but not in normal
endometrium or tissue with ( atypical) hypeiplasia. Also mRNA hybridization
signals of aromatase accumulated in stromal cells but not in neoplastic or
(atypical) hyperplastic cells (without stromal invasion). The distribution of
aromatase mRNA correlated well with the immunohistochemical localization of
aromatase. It was concluded that intratumoral aromatase in endometrial malignan
cy is associated with stromal invasion and is expressed during the process of carci
noma-stromal interaction. The presence of aromatase in stromal or interstitial cells
suggests that locally produced estrogens may act on carcinoma cells in a paracrine
way. Watanabe et al reported no correlation between aromatase expression or acti
vity and steroid receptor status, clinical stage or histological grade. 15 Aromatase
activity was noted to increase proportionally with increasing concentrations of
insulin, suggesting that hyperinsulinemiamay predispose to endometrial neoplasia
by enhancing endogenous endometrial estrogen production.63 It has long been
recognized that conversion of androstenedione to estrone, and subsequently to
estradiol, increases with both age and body weight.64•65 These findings support the
alleged association between diabetes, obesity, aromatase activity and endometrial
cancer.
24
Gene expression
Regulation of aromatase ( over-)expression in endometrial cancer remains an
unsolved issue. Interaction between hormonal and genetic factors during the
development of precancerous lesions may eventually cause cancer, but it remains
still questionable which kinds of genetic predispositions favour malignant
degeneration The human CYP 19 gene contains a number of promotors that direct
tissue-specific aromatase expression. In adipose stromal cells this transcription is
regulated through the use of promotor 1.4 stimulated by glucocorticoids.66 In
endometrial cancer levels of aromatase transcripts were reported similar to those in
adipose tissue, whereas disease-free endometrium did not express aromatase.1 1•60
In small studies promotor II ( exon 1 d and 1 b) and 1.3 ( exon 1 c) seem to be
preferably used for aromatase expression in endometrial malignancy.1 1 • 1 4 Various
hormones, growth factors or cytokines may increase intracellular cAMP levels to
produce a signal transduction mechanism, responsible for stimulation of
transcription, in the end leading to local estrogen biosynthesis. Above observations
suggest that in endometrial cancer the normal mechanisms that inhibit aromatase
expression may be lost or, alternatively, that aromatase-stimulatingfactors or their
receptors become overexpressed. Women who may be at increased risk are those
who are not able to deal with endogenous or exogenous hyperestrogenism, due to
e.g. polymorphisms in estrogen metabolizing genes. Recently it was demonstrated
that CYP 19 polymorphism might be a genetic risk factor for the development of
endometrial cancer, as women carrying the A6 and/ or A 7 alleles of CYP 19 were
over-represented in patients with endometrial cancer. 67 Like CYP19 also CYPl 7
is involved in steroid production. Postmenopausal women who were homozygote
for the A2 allele of CYP 17 had higher estrogen and androgen levels.68
Mutation of the PTEN tumour suppression gene, which is found in 34-80% of
endometrioid endometrial malignancies and hyperplasia, has emerged as a primary
cause of cancer origin.69-7 1 A relationship between PTEN expression and steroid
25
hormone levels and thus hormonal endometrial risk factors is suspected: possibly
PTEN mutation enhances (pre-)malignant development as a promotor or initiator,
in this way genetically predisposing to the development of endometrial cancer.
PTEN immunostaining may be a marker for identification of premalignant lesions
that are likely to degenerate in a malignant fashion.69
Endometrial microsatellite instability, mutations m the ras oncogene and
replication error repair (RER) phenotype are found in 9 to 45% of endometrial
premalignancy and adenocarcinoma. Microsatellites are short repeating noncoding
units widely dispersed throughout the genome. The basis of microsatellite
instability is probably epigenetic silencing of DNA mismatch repair genes. It is
associated with neoplastic endometrium and may develop long before clinical
manifestation of malignant disease, but is rarely found in randomly sampled
histologicallynormal endometrial tissues.72•73
Possibly due to these genetic disorders or mutations clonal clusters are formed,
derived from common progenitor endometrial cells. In these proliferative
monoclonal expansions atypical endometrial hyperplasia may be histologically
diagnosed as a direct precursor of endometrial cancer.74 Whether aromatase
expression also plays a role in these hypothethical processes by promoting
protracted unopposed estrogen exposure remains questionable.
Aromatase inhibition in endometrial cancer
Although it is tempting to assume a causal relationship between ovarian stromal
hyperplasia, elevated androgen concentrations, and endometrial cancer, further
studies on aromatase activity in endometrial tissue are warranted. Treatment of
histocultures of endometrioid endometrial cancer with aromatase inhibitors
suggested depletion of in situ tumor estrogen, which might result in decreased cell
26
proliferation of tumor cells.75 The decreased cell proliferation due to aromatase
inhibitors in vitro appeared to be not directly related to intratumoral aromatase and
estrogen receptor status. Apart from their antagonistic oestrogenic action, anti
aromatase agents have no in vivo oestrogen agonistic properties. In contrast to
tamoxifen aromatase inhibitors do not induce disturbing adverse events like
endometrial hyper- or neoplasia.76 A better understanding of intratumoral
aromatase in estrogen dependent malignancies like endometrioid endometrial
carcinoma may give rise to the potential development of endocrine therapy by
aromatase inhibitors, as already occured in breast cancer.77 Already in 1984
objective responses to the aromatase inhibitor aminogluthethimide were obtained
in four out of 18 women with advanced endometrial cancer.78 In vivo aromatase
inhibitors were dose-dependently able to inhibit aromatase activity in human
uterine tumours.79 A small pilot study showed decrease of aromatase activity in
endometrial tumour tissue after administration of the aromatase inhibitor
letrozole.80 In a recent Phase II study the use of the oral nonsteroidal aromatase
inhibitor anastrazole was studied in endometrial cancer.8 1 In a group of 23 women
with advanced or recurrent endometrial malignancy, most of whom had poorly
differentiated tumours and negative estrogen receptors, only two partial
responses were noted. However, only 39% of enrolled patients were diagnosed
with endometrioid endometrial cancer, the type of malignancy where aromatase
inhibibition might be of therapeutical benefit. The possible benefit of aromatase
inhibitors in well-differentiated or hormone-receptor-positive cancer in the
adjuvant situation remains yet to be ascertained.
Acknowledgements
We thank JLH Evers, J Grond, H Hollema, FH De Jong, HJMM Mertens, JG
Santema and AGJ van der Zee for their stimulation and support.
27
References
1. Zeleniuch-Jacquotte A, Akhmedkhanov A, Kato I, Koenig KL, Shore RE, Kim MY,
Levitz M, et al. Postmenopausal endogenous oestrogens and risk of endometrial cancer: results of a prospective study. Br J Cancer 2001; 84(7): 975-8 1.
2. Austin H, Austin JM, Partridge EE, Hatch KD, Shingleton HM. Endometrial cancer,
obesity and body fat distribution. Cancer Res 199 1 ; 5 1 : 568-72. 3. Nyholm HCJ, Nielsen AL, Lyndrop J, Dreisler A, Hagen C. Plasma oestrogens in
postmenopausal women with endometrial cancer. Br J Obstet Gynaecol 1993; 100: 1 1 15-
19.
4. PotischmanN, Hoover RN, Brinton LA, Siiteri PK, Dorgan JF, Swanson CA, et al. Case
control study of endogenous steroid hormones and endometrial cancer. J Natl Cancer Inst 1996; 88 : 1 127-35 .
5. Akhmedkhanov A, Zeleniuch-Jacquotte A, Toniolo P. Role of exogenous and endogenous hormones in endometrialcancer. Ann N Y Acad Sci 2001 ; 943 : 296-3 15 .
6. Grodin JM, Siiteri PK, MacDonald PC. Source of estrogen production in postmenopausal
women. J Clin EndocrinolMetab 1973; 36: 207- 14 .
7. Adashi EY. The climacteric ovary as a functional gonadotropin-driven androgenproducing gland. Fertil Steril 1994; 62: 20-7.
8. Plouffe L. Ovaries, androgens and the menopause: practical applications. Semin Reprod
Endocrinol 1998; 16: 1 17-20.
9. Sluijmer AV.The postmenopausalovary. Functional and morpholologicalaspects. Thesis
Maastricht 1999 . ISBN 90-9012802-6.
10. Ala-Fossi SL, Maenpaa J, Aine R, Punnonen R. Ovarian testosterone secretion during perimenopause.Maturitas 1998 ; 29: 239-45 .
1 1 . Bulun SE, Economos K, Miller D, Simpson ER. CYP 19 ( aromatase cytochrome P450)
gene expression in human malignant endometrial tumors. J Clin Endocrinol Metab
1994;79: 183 1-4. 12. Longcope C. Endocrine function of the postmenopausal ovary. J Soc Gynecol Investig
2001; 8: S67-8. 13 . Bulun SE, Noble LS, TakayamaK, Michael MD, Agarwal V, Fisher C, et al. Endocrine
disorders associated with inappropriately high aromatase expression. J Steroid Biochem
Molec Biol 1997; 61 : 133-9.
28
14 . Sasano H, Harada N. Intratumoral aromatase in human breast, endometrial, and ovarian
malignancies.Endocr Rev 1998; 19(5): 593-607. 15 . WatanabeK, Sasano H, Harada N, Ozaki M, Niikura H, Sato S, YajimaA. Aromatase in
human endometrial carcinoma and hyperplasia. Immunohistochemical, in situ hybridization,and biochemical studies. Am J Pathol 1995; 146: 491-500.
16. Key TJA, Pike MC. The dose-effect relationship between 'unopposed' oestrogens and
endometrial mitotic rate: its central role in explaining and predicting endometrial cancer risk. Br J Cancer 1988; 57: 205- 12.
17. Lucisano A, Russo N, Acampora MG, Fabiano A, Fattibene M, Parlati E, Maniccia E, Dell"Acqua S. Ovarian and peripheral androgen levels in postmenopausal women:
correlations with ovarian histology.Maturitas 1986; 8: 57-65 .
18. Sluijmer AV, Heineman MJ, Koudstaal Joh, Theunissen PHMH, De Jong FH, Evers
JLH.Relationship between ovarian production of estrone, estradiol, testosterone and androstenedione and the ovarian degree of stromal hyperplasia in postmenopausal
women. Menopause 1998; 5 : 207- 10.
19. Longcope C, Hunter R, Franz C. Steroid secretion by the postmenopausal ovary. Am J
Obstet Gynecol 1980; 138 : 564-8.
20. Mattingly RF, Huang WY. Steroidogenesisof the menopausal and postmenopausal ovary. Am J Obstet Gynecol 1969; 103 : 679-93.
21 . Judd HL, Lucas WE, Yen SSC. Effect of oophorectomy on circulating testosterone and
androstenedione levels in patients with endometrial cancer. Am J Obstet Gynecol 1974;
1 18 : 793-8.
22. Judd HL, Judd GE, Lucas WE, Yen SSC. Endocrine function of the postmenopausal ovary: concentration of androgens and estrogens in ovarian and peripheral vein blood. J
Clin Endocrinol Metab 1974; 39 : 1020-4.
23 . Heineman MJ, Sluijmer AV, Evers JL. Utero-ovarian vem blood sampling in
postmenopausalwomen. Fertil Steril 1993 ; 60: 184-6.
24. Greenblatt RB, Colle ML, Mahesh VB. Ovarian and adrenal steroid production in the postmenopausalwoman. Obstet Gynecol 1976; 47: 383-7.
25. Sluijmer AV, Heineman MJ, De Jong FH, Evers JLH. Endocrine activity of the postmenopausal ovary: the effects of pituitary down-regulation and oophorectomy.J Clin
EndocrinolMetab 1995; 80: 2 163-7.
29
26. Judd HL, Lucas WE, Yen SSC. Serum 17 beta-estradiol and estrone levels in postmenopausal women with and without endometrial cancer. J Clin Endocrinol Metab 1976; 43 : 272-8.
27. Dowsett M, Cantwell B, Anshumala L, Jeffcoate SL, Harris AL. Suppression of
postmenopausal ovarian steroidogenesis with the luteinizing hormone-releasing hormone
agonist Goserilin. J Clin Endocrin Metabol 1988; 66: 672-7. 28. Nagamani M, Doherty MG, Smith ER, ChandrasekharY. Increased bioactive luteinizing
hormone levels in postmenopausal women with endometrial cancer. Am J Obstet Gynecol 1992; 167 : 1825-30.
29. Kobayashi M, Nakano R, Shima K. Immunohistochemical localization of pituitary gonadotropins and estrogen in human postmenopausal ovaries. Acta Obstet Gynecol
Scand 1993 ; 72: 76-80. 30. Inkster SE, Brodie AMH. Expression of cytochrome P450 in premenopausal and
postmenopausal human ovaries: an immunocytochemical study. J Clin Endocrinol Metab 199 1 ; 73 : 7 17-26.
3 1. Dennefors BL, Janson PO, Knutson F, Hamberger L. Steroid production and resposiveness to gonadotropin in isolated stromal tissue of human post-menopausal
ovaries. Am J Obstet Gynecol 1980; 136: 997-1002.
32. Couzinet B, Meduri G, Lecce MG, Young J, Brailly S, Loosfelt H, et al. The postmenopausal ovary is not a major androgen-producing gland. J Clin Endocrinol
Metab 2001; 86: 5060-6. 33 . Longcope C. Editorial. The relationship of ovarian production of hormones and degree of
ovarian stromal hyperplasia in postmenopausal women. Menopause 1998; 5: 205 .
34. Boss JH, Scully RE, Wegner KH, Cohen RB. Structural variations in the adult ovary: clinical significance. Obstet Gynecol 1965 ; 25 : 747-64.
35. Woll E, Hertig AT, Smith GVS, Johnson LC. The ovary in endometrial carcinoma: with
notes on the morphological history of the aging ovary. Am J Obstet Gynecol 1948; 56:
617-33 .
36. Emons G, Fleckenstein G, Hinney B, Huschmand A, Heyl W. Hormonal interactions in endometrial cancer. Endocr Rel Cancer 2000; 7 : 227-42.
37. Sherman ME. Theories of endometrial carcinogenis: a multidisciplinary approach. Modem Pathol 2000; 13 : 295-308.
38. Carcangiu ML, Chambers JT, VoynickIM, Pirro M, Schwartz PE. Immunohistochemical evaluation of estrogen and progesterone receptor content in 183 patients with
30
endometrial carcinoma: part I: clinical and histological correlations. Am J Clin Pathol
1990; 94 : 247-54. 39. Umpierre SA, Burke TW, Tomos C, Ordonez N, Levenback C, Morris M.
Immunocytochemical analysis of uterine papillary serous carcinomas for estrogen and
progesteronereceptors. Int J Gynecol Pathol 1994; 1 3 : 127-30. 40. Sherman ME, Sturgeon LA, Brinton LA. Risk factors and hormone levels in patients
with serous and endometrioid uterine carcinomas. Mod Pathol 1997; 10 : 963-8.
4 1 . Henderson BE, Ross RK, Bernstein L. Estrogens as a cause of human cancer: the
Richard and Hinda Rosenthal Foundation Award Lecture. Cancer Res 1988; 48 : 246-53 . 42. Kurman RJ, Kaminski PF, Norris HJ. The behaviour of endometrial hyperplasia. A long
term study of"untreated"hyperplasiain 170 patients. Cancer 1985 ; 56: 403- 12. 43 . Carciangiu ML, Chambers JT. Uterine papillary serous carcinoma: a study of 108 cases
with emphasis on the prognostic significance of associated endometrioid carcinoma, absence of invasion, amd concomitant ovarian carcinoma. Gynecol Oncol 1992; 47:
298-305. 44. Moore TD, Phillips PH, Nerenstone SR, Cheson BD. Systemic treatment of advanced
and recurrent endometrial carcinoma: current status and future directions. J Clin Oncol 199 1 ; 9: 107 1 -88.
45. Mourits MJM, De Vries EGE, Willemse PHB, Ten Hoor KA, Hollema H, Van der Zee
AGJ. Tamoxifen treatment and gynecologic side effects: a review. Obstet Gynecol 200 I ; 97(5 Suppl 1 ) : 855-66.
46. Nagamani M, Hannigan EV, Dillard EA jr, Dinh TV. Ovarian steroid secretion in
postmenopausal women with and without endometrial cancer. J Clin Endocrinol Metab
1986; 62: 508- 12.
47. Bremond AG, Claustrat B, Rudigoz RC, Seffert P, Comiau J. Estradiol, androstenedione, testosterone, and dehydroepiandrosteronesulfate in the ovarian and peripheral blood of postmenopausal patients with and without endometrial cancer. Gynecol Oncol 1982; 14 :
1 19-24.
48. Mertens HJMM, Heineman MJ, De Jong FH, Evers JLH. Postmenopausal ovarian
function and endometrialcancer. Hum Reprod 14, Abstractbook; June 1999: 89-90.
49. Nagamani M, Stuart CA, Doherty MG. Increased steroid production by the ovarian stromal tissue of postmenopausal women with endometrial cancer. J Clin Endocrinol
Metab 1992; 74: 172-6.
3 1
50. Horie K, Takakura K, Imai K, Liao S, Mori T. Immunohistochemical localization of androgen receptor in the human endometrium, decidua, placenta and pathological conditions of the endometrium.Hum Reprod 1992; 7: 1461-6.
5 1. Mertens HJMM, Heineman MJ, Koudstaal J, Theunissen P, Evers JLH. Androgen
receptor content in human endometrium. Eur J Obstet Gynecol Reprod Biol 1996; 70: 1 1-3.
52. Hackenberg R, Schulz KD. Androgen receptor mediated growth control of breast cancer and endometrialcancer modulated by antiandrogen-and androgen-like steroids. J Steroid
Biochem Molec Biol 1996; 56: 1 13-7 .
53. Legro RS, Kunselman AR, Miller SA, Satyaswaroop PG. Role of androgens in the growth of endometrial carcinoma: an in vivo animal model. Am J Obstet Gynecol 2001; 184 : 303-8 .
54. Nelson LR, Bulun SE. Estrogen production and action. J Am Acad Dermatol 2001 ; 45 :
S 1 16-24. 55. Simpson ER, Davis SR. Minireview: aromatase and the regulation of estrogen
biosynthesis-somenew perspectives. Endocrinol 2001 ; 142: 45 89-94. 56. Tseng L, Mazella J, Mann WJ, Chumas J. Estrogen synthesis in normal and malignant
human endometrium. J Clin Endocrinol Metab 1982; 55 : 1029-3 1.
57. Yamamoto T, Kitawaki J, Urabe M, Honjo H, Tamura T, Noguchi T, et al. Estrogen productivity of endometrium and endometrial cancer tissue; influence of aromatase on
proliferation of endometrial cancer cells. J Steroid Biochem Molec Biol 1993; 44 : 463-8.
58. Sasano H, Kaga K, Sato S, Yajima A, Nagura H, Harada N. Aromatase P450 gene expression in endometrial carcinoma. Br J Cancer 1996; 74 : 154 1-4.
59. Yamaki J, Yamamoto T, Okada H. Aromatization of androstenedione by normal and neoplasticendometriumofthe uterus. J Steroid Biochem Molec Biol 1985; 22: 63-6.
60. Bulun SE, Mahendroo MS, Simpson ER. Polymerase chain reaction amplification fails to detect aromatase cytochrome P450 transcripts in normal human endometrium or
decidua. J Clin EndocrinolMetab 1993 ; 76: 1458-63.
61. Noble LS, S impson ER, Johns A, Bulun SE. Aromatase expression in endometriosis. J Clin EndocrinolMetab 1996; 8 1 : 174-9.
62. Kitawaki J, Noguchi T, Amatsu T, Maeda K, Tsukamoto K, Yamamoto T, et al. Expression of aromatase cytochrome P450 protein and messenger ribonucleic acid in
human endometriotic and adenomyotic tissues but not in normal endometrium. Biol
Reprod 1997; 57: 5 14-9.
32
63. Randolph JF, Kipersztok S, AyersJWT, Ansbacher R, Peegel H, Menon K.MJ. The effect
of insulin on aromatase activity in isolated human endometrial glands and stroma. Am J Obstet Gynecol 1987; 157: 1534-9.
64. Hemsell DL, Grodin JM, Brenner PF, Siiteri PK, MacDonald PC. Plasma precursors of
estrogen.II. Correlation of the extent of conversion of plasma androstenedione to estrone
with age. J Clin EndocrinolMetab 1974; 38 : 476-9. 65. MacDonald PC, Edman CD, Hemsell DL, Porter JC, Siiteri PK. Effect of obesity on
conversion of plasma androstenedione to estrone in postmenopausal women with and
without endometrialcancer. Am J Obstet Gynecol 1978; 130: 448-55. 66. Simpson ER, Mahendroo MS, Means GD, Kilgore MW, Hinshelwood MM, Graham
Lorence S, et al. Aromatase cytochrome P450, the enzyme responsible for estrogen
biosynthesis.Endocr Rev 1994; 15 : 342-55.
67. Berstein LM, Imyanitov EN, Suspitsin EN, Grigoriev MY, Sokolov EP, Togo A, et al.
CYPl 9 gene polymorphism in endometrial cancer patients J Cancer Res Clin Oncol
200 1 ; 127: 135-8. 68. Haiman CA, Hankinson SE, Colditz GA, Hunter DJ, De Vivo I. A polymorphism in CYP
17 and endometrialcancer risk. Cancer Res 2001; 61 : 3955-60. 69 . Mutter GI, Lin MC, Fitzgerald JT, Kum JB, Baak JPA, Lees JA, et al. Altered PTEN
expression is a diagnostic marker for the earliest endometrial precancers. J Natl Cancer
Inst 2000; 92: 924-30. 70. Mutter GL, Lin MC, Fitzgerald JT, Kum JB, Eng C. Changes in endometrial PTEN
expression throughout the human menstrual cycle. J Clin Endocrinol Metab 2000; 85 :
2334-8.
7 1 . Dahia PLM. PTEN, a unique tumor suppression gene. Endocr Rel Cancer 2000; 7: 1 15-29.
72. Faquin WC, Fitzgerald JT, Lin MC, Boyton KA, Muto MG, Mutter GL. Sporadic
microsatellite instability is specific to neoplastic and preneoplastic endometrial tissues.
Am J Clin Pathol 2000; 1 13 : 576-82. 73. MacDonald ND, Salvesen HB, Ryan A, Iversen OE, Akslen LA, Jacobs IJ. Frequency
and prognostic index of microsatellite instability in a large population based study of
endometrial carcinomas. Cancer Res 2000; 60: 1750-2. 74. Jovanovic AS, Boynton KA, Mutter GL. Uteri of women with endometrial carcinoma
contain a histopathological spectrum of monoclonal putative precancers, some with microsatelliteinstability.Cancer Res 1996; 56: 19 17-21 .
33
75. Sasano H, Sato S, Ito K, Yajima A, Nakamura J, Yoshihama M, et al. Effects of
aromatase inhibitors on the pathobiology of human breast, endometrial and ovarian carcinoma. Endocr Rel Cancer 1999; 6: 197-204.
76. Bonte J. Recent data about endometrial carcinoma: potential of anti-aromatase agents.
Eur J Gynaec Oncol 2001; 22: 96-8. 77. de Jong PC, van de Ven J, Nortier HW, Maitimu-Smeele I, Donker TH, Thijssen JH, et
al. Inhibition of breast cancer tissue aromatase activity and estrogen concentrations by the third-generationaromatase inhibitor vorozole. Cancer Res 1997; 57: 2 109- 1 1 .
78. Murray RLM, Pitt P. Treatment of advanced metastatic breast cancer, carcinoma of the
prostate and endometrial cancer with aminogluthethimidine.In: Aminoglutethimideas an aromatase inhibitor in the treatment of cancer 1984; 1 09- 122. Eds Nagel GA & Santen
RJ. Bern: Huber 79. Yamamoto T, Fukuoka M, Fujimoto Y, Kitawaki J, Nakakoshi M, YoshihamaM, et al.
Inhibitory effect of a new androstenedione derivate, 14a-hydroxy-4-androstene-3,6, l 7-trione ( 14a-OHAT) on aromatase activity of human uterine tumors. J Steroid Biochem Molec Biol 1990; 36: 5 17-21.
80. Bershtein LM, Maksimov SI, Gershfel'd ED, Gamaiunova VB, Meshkova IE,
Kovalenko IG, et al. Neoadjuvant use of the aromatase inhibitor letrozole in uterine cancer: endocrine and clinical effects. VoprOnkol 200 1 ; 47(5): 5 17-74.
8 1 . Rose PG, Brunetto VL, VanLe L, Bell J, Walker JL, Lee RB. A phase III trial of
Anastrozole in advanced recurrent or persistent endometrial carcinoma: a gynecologic
oncology group study. Gynecol Oncol 2000; 78: 212-6.
34
3
Aromatase in the context of breast and endometrial
cancer. A review
VHWM Jongen
H Hollema
AGJ van der Zee
MJ Heineman
Minerva Endocrinologica 2006;31 :47-60
35
Abstract
Generally, estrogens are considered to be involved in the neoplastic
transformation of endometrium. After the menopause these estrogens mainly
originate from conversion of adrenal androgens by aromatization in body fat.
However, in case of stromal hyperplasia of the ovaries, it can not be excluded
that production of aromatizable androgens by postmenopausal ovaries leads to
increased availability of androgen precursors for intratumoural estrogen
synthesis in the endometrial tissue as well. The local presence of androgens and
the local expression and activity of aromatase is considered important for this
steroidogenesis. In this review, we will discuss the available evidence that
androgens, produced in hyperplastic ovarian stroma or body fat tissues, play a
role in the development of endometrial cancer through conversion into
estrogens, a reaction catalyzed in the endometrium by the enzyme aromatase
cytochrome P450. As the presence of aromatase appeared to be a
pathophysiological factor in the formation of breast cancer, the latter will be
evaluated in relation to the development of endometrioid endometrial cancer as
well, since both disorders appear partly estrogen dependent. As treatment with
aromatase inhibitors appeared feasible in breast cancer, current knowledge of
comparable treatment modalities in hormone dependent endometrial cancer will
be reviewed.
36
Aromatase
The enzyme complex aromatase plays an indispensable role in the production of
estrogens. Step by step cholesterol is converted to estrogen by steroidal
hydroxylation (Figure 1 ). Finally estradiol is biosynthesized from androgens by
aromatase. The enzyme complex aromatase consists of the cytochrome P450
aromatase enzyme, a haeme group, and an NADPH-reductase flavoprotein. 1
Cytochrome P450 is the product of the CYP19 gene. The enzyme aromatase is
localized in the endoplasmic reticulum of cells found in tissues as ovaries,
adipose tissue, placental syncytiotrophoblast, breast, bone and brain. Androgens
are important C 19 steroid precursor substrates for estrogen biosynthesis in these
tissues. Aromatase binds the androgen steroid substrate and catalyzes a series of
three consecutive reactions which will finally lead to estrogen production. After
two hydroxylation steps, the third aromatization step is defined by an oxidation
reaction in which cytochrome P-450 converts C19 androgens into C18
estrogens.2 In postmenopausal women approximately 2% of circulating
androstenedione is converted to estrone, which can be further processed to the
more potent estradiol.
From a clinicopathological point of view, estrogens are considered to be the
most important factor promoting cell growth of endometrioid endometrial
carcinoma. The principle precursors of estrogens are circulating androgens,
derived from primarily the adrenal cortex and to a lesser extent from the ovaries.
Before menopause the granulosa cells in the ovaries are the main source of
estrogens, after menopause the adipose tissue. The role of the postmenopausal
ovary in the production of hormonal precursors can not be neglected, as these
prehormones can be processed by a.o. the aromatase complex. After menopause,
the ovary still produces 40% respectively 20% of the total body quantity of
testosterone and androstenedione, which can be converted into estrogens. 3 With
37
increasing age, body weight and aromatase activity in extraglandular (adipose)
tissue more peripheral conversion of androgens to estrogens is observed. 4 The
contribution of aromatase activity to estrogen formation must be considered
important, as aromatization is the rate-limiting factor. So it seems logical that
endocrinological treatment of estrogen levels in hormone dependent tumours
focusses on inhibition and reduction of aromatization processes. In breast cancer
reduction of estrogenic stimulation became the main pharmacological target,
leading to the discovery that treatment with aromatase inhibitors is quite
effective in depriving the tumour of estrogenic growth stimulation.
Cholesterol ----+ Dehydroepiandrosterone ----+
+
Androstenediole
Androstenedione ----+ (J ) Estrone
t (2) t (2)
Testosterone ----+ ( J l Estradiol
Fig 1 . Pathways of sex hormone metabolism and biosynthesis, influenced by a.o. ( 1 )
aromatase, and (2) l 7beta-hydroxysteroid dehydrogenase activity.
Breast cancer and hormone-dependent endometrial cancer; biological
similarities or not?
The incidence of breast and endometrial cancer is near 28% and 1 0% of all
cancers in women. 5 Invasive breast cancer and endometrioid endometrial
carcinoma are both preceded by premalignant disorders, characterized by
estrogen induced hyperproliferation.6'7 Both breast and endometrioid
endometrial cancer are partly considered as estrogen sensitive and in both
bioconversion of androgens to estrogens plays a role in tumour development.
Aromatase expression and activity are of importance in this process . Although
endogenous hormones are not an easily modifiable risk factor, the hormone
38
dependent evolution m both disorders and the biological alterations and
differences ask for comparison, as prognostic factors and targets for
endocrinological therapy may be shared.
Breast cancer
Cumulative doses of unapposed estrogens over time increase the risk of
developing breast cancer, so a lower age at menarche, a higher age at menopause
and a higher body mass index have been associated with an increasing risk of
developing breast cancer.6 Prolonged lactation diminishes the number of
menstrual cycles and subsequently breast cancer risk decreases, whereas alcohol
consumption may increase breast cancer rbisk by increasing plasma estrogen
levels. Postmenopausal obesity causes increased conversion of androstenedione
to estrone in adipose tissue, and is an accepted risk factor for breast cancer.
Two distinctive clinicopathological pathways have been described in the
histological pathophysiology of breast cancer.8 Ductal carcinoma, accounting for
85% of invasive breast cancers, develops from terminal duct units, which give
rise to hyperplastic unfolded lobules, followed by atypical ductal hyperplasia,
ductal carcinoma in situ and beyond. A second pathway involves the so-called
lobular neoplasia, where atypical lobular hyperplasia evolves to lobular
carcinoma in situ. Finally invasive breast cancer establishes.
Nearly all premalignant lesions are estrogen receptor positive. The percentage of
estrogen receptor positive cells in breast cancer tissue differs.9 In one-third of
breast cancers, growth can be stimulated by estrogens, while tumour regression
has been observed after estrogen deprivation. 10 The most important prognostic
factor for response to hormonal treatment is the presence of estrogen or
progesterone receptors in the malignant tissues; in case of receptor positivity 60
to 70% of women respond to hormonal therapy, so estrogen receptor status is
39
considered as a prognostic biomarker predicting hormonal therapy sensitivity .
Breast cancers that have no or low estrogen receptor density do not appear to
require estrogen for their continued growth. These tumours are unlikely to
respond to therapy with aromatase inhibitors, designed to reduce intratumoural
estrogen levels. About 50 to 80% of breast cancers appear estrogen receptor
positive, but breast cancers with a more malignant histological profile are mostly
estrogen receptor negative, so it can not be excluded that breast cancer can
become estrogen receptor negative and estrogen independent as the tumour
progresses. Progesterone receptors are probably induced by estrogens through
the estrogen receptor and were observed in 40-45% of breast cancers. 1 1
Breast cancer is mainly diagnosed after menopause. Due to a decrease of
postmenopausal ovarian estrogen production plasma estrogen levels drop
significantly after menopause. However, as the majority of breast cancers
depend on estrogens for their growth and appear estrogen receptor positive,
growth of breast cancers is expected to depend more on local estrogen levels
than on circulating plasma estrogen levels. No significant differences in
intratumoural estrogen levels exist between pre- and postmenopausal women
with breast cancer. 1 2 Estrogens are produced locally within the breast tumour
after conversion of androgens due to the activity of the aromatase enzyme
complex, which enables the tumour to maintain high estrogen levels
independently of menopausal state. 1 •1 3 Estrogen biosynthesis in the tumour or
surrounding tissues is thought to be of importance. About 63 to 72% of breast
tumours appear to show measurable aromatase activity, but aromatase activity
between tumours can vary tremendously.1 4 Remarkably tumour aromatase
activity was found to be an independent endocrinological factor, not correlated
with estrogen or progesterone receptor status. 1 5 However, the exact localisation
of aromatase activity within the breast is still not certain. Controversial results
are reported concerning localisation of aromatase activity in tumour or adipose
40
stromal cells.5•1 6 Aromatase activity was detected in higher amounts in the
tumour-bearing part of the breast, but in breast fat tissues as well. 1 7 Possibly the
fat surrounding tumour cells is important, since cancer cells are often
surrounded by adipose stromal cells and adipocytes. To some degree
intratumoural aromatase activity was found to have predictive value as a
prognostic factor, but no consistent relationship was found between aromatase
activity in breast tumours and known prognostic factors like age, menopausal
status, tumour size, nodal status, estrogen and progesterone receptor
concentrations, tumour differentiation grade and survival. 1 8-2 1
Estrogen deprivation as a therapeutic target can be achieved by blocking the
hormone receptor with a selective estrogen receptor modulator (SERM) like
tamoxifen, which is first-line therapy for advanced breast cancer in
postmenopausal women and is most effective in estrogen receptor positive
disease. Alternatively, before menopause estrogen concentrations can be
lowered by ovariectomy or by the use of GnRH-analogues, but aromatase
inhibitors can be used as well, aiming at the cytochrome P450 enzyme complex
aromatase. Aromatase inhibitors are effective as second line treatment after
developing resistance to tamoxifen, as adjuvant treatment or as first line therapy
in metastatic disease. A major advantage of recently developed aromatase
inhibitors seem their improved efficacy and tolerability, when compared to the
earlier aromatase inhibitors. In the near future an even more prominent and first
line role for aromatase inhibitors is expected in the treatment of breast cancer.
The exact mechanism by which the use of aromatase inhibitors leads to tumour
reduction is still unknown, possibly apoptosis plays a major role. Most probably
a combination of reduced cell proliferation, increasing coagulative and extensive
hyelinisation of the cancer and stromal cells in correlation with host reactions
such as tumour infiltration by lymphocytes and/or macrophages are of
importance. 1 9 Until now, it remains difficult to determine aromatase positivity,
41
as conclusive results are hampered by the lack of specificity of the aromatase
antibodies used, which off course is a major problem and applies for endometrial
cancer as well, as will be discussed later on.
Endometrial cancer, estrogens and androgens
Analogous to breast cancer, carcinogenesis of endometrioid endometrial cancer
has been related to excessive life-time exposure to estrogens. Hyperlipidemia,
hypertension, diabetes and obesity increase life-time risk, probably due to incre
ased conversion of androgens to estrogens in the adipose tissue. Ongoing
menopausal estrogen replacement therapy, sequential oral contraceptives,
chronic anovulation, anovulatory uterine bleeding, nulliparity, early age at
menarche and late onset of menopause increase the risk of developing
endometrioid endometrial cancer. This gives rise to the concept of continuous
estrogenic hyperstimulation of the postmenopausal endometrium, in which
endometrioid endometrial cancer gradually evolves from simple and atypical
hyperplasia. Diminished risks were observed in association with use of
combination-type oral contraceptives, and in case menopausal hormone
replacement therapy was combined with progestins, as progesterone response is
essential to control normal cell growth. With each pregnancy ( or lactation) the
life-time risk decreases, which is explained by the fact that no mitotic activity is
seen due to the continuously high placental production of progestagens.
Although both breast and endometrial cancer may develop under the influence
of estrogens, a major difference is evident: due to the endometrial capacity of
cyclic changes before menopause, progestins exert a protective action in
endometrial tissue. Conflicting data are present on the issue whether endogenous
progestins increase breast cancer risk, or not. 22•23 In the last two decades the
prevailing opinion has been that progesterone contributes to the development of
breast cancer, since exogenous progestins increase the breast cancer risk more
42
than estrogen alone when added to estrogen in hormone replacement therapy for
menopausal complaints.
On morphologic grounds endometrial cancer can be divided in two different
classes: endometrioid endometrial carcinoma related to unopposed estrogenic
stimulation, and a second estrogen independent form including poorly
differentiated and high-grade malignancies like serous papillary and clear cell
carcinoma, and mixed forms with endometrioid components. 24 Estrogen or
progesterone receptors are rarely present in these poorly differentiated cancers,
which also hardly respond to classic hormonal therapy. More than 75% of all
cases of endometrial cancers consist of estrogen and progesterone receptor
positive endometrioid endometrial cancers. As endometrioid endometrial cancer
has a relatively good prognosis, estrogen and progesterone receptor positivity
have been used as prognostic biomarkers. Especially progesterone receptor
positivity was associated with a more favourable outcome and disease-free
survival. 25 Endometrioid endometrial cancer gradually evolves from simple
hyperplasia to atypical hyperplasia and fullblown cancer. Under continuous
estrogenic stimulation nearly one quarter of atypical hyperplasias progress to
well-differentiated endometrial carcinoma. It is assumed that the increased
mitotic activity of the endometrium caused by estrogenic stimulation increases
the probability of the accumulation of random mutations, finally leading to
malignant transformation by activation of oncogenes, defective mismatch repair
genes and inactivation of tumour-suppressor genes. Continuous
hyperstimulation of the endometrium by ( endogenous and exogenous) estrogens
is supposed to be the causative mechanism eventually leading to endometrioid
endometrial cancer or its progression. However, androgens may play a role as
well. Several causative pathways may co-exist next to each other, or even
amplify each other, leading to the gradual development of endometrial cancer.
Since the ovary still produces considerable amounts of androgens after
43
menopause, these androgens may not only be processed to estrogens in
peripheral adipose tissues, but may also lead to elevated prehormone availability
for estrogen formation in utero. The hypothesis that ovarian androgens play a
role in endometrial carcinogenesis is supported by the fact that androgen levels
in the utero-ovarian veins from women with endometrial cancer were
significantly elevated as compared to women without malignant disease. 3•26-3 1
One may argue that ovarian androgens stimulate the development of endometrial
cancer through local aromatization into estrogens. In that case local aromatase
activity may give rise to high local estrogen concentrations which could result
in the stimulation of endometrial growth. Thus, in addition to extraglandular
aromatase activity through conversion of androgens into estrogens in adipose
tissue, aromatase activity in the endometrium itself may play a role in malignant
transformation. 32-34 The local availability of androgens and the presence of
aromatase activity within endometrial tissue, are the two essential conditions
supporting the hypothesis that estrogens may function as a mitogenic factor after
local conversion of ( ovarian) androgens. Thus, in this view some women may
develop endometrioid endometrial cancer through continuous hyperstimulation
of the endometrial tissue by estrogens systematically delivered from peripheral
conversion in body fat, while in other women malignant transformation is
enhanced by local conversion of ovarian androgens into estrogens, catalyzed by
(perhaps overexpression of) endometrial stromal aromatase cytochrome P450, or
both happen simultaneously. Possibly genetic risk factors like CYP19
polymorphisms increase the risk of developing endometrial cancer, as the normal
mechanisms that inhibit aromatase expression may be lost or aromatase
stimulating factors or their receptors become overexpressed, causing
hyperestrogenism:5 The human CYP 19 gene contains a number of promotors that
direct tissue-specific aromatase expression. Especially women carrying the A6
and/or A7 alleles of CYP19 were over-represented in patients with endometrial
cancer.36 Remarkably, evidence was found that a correlation exists between CYP
44
19 gene expression and the endometrial cancer types until now generally
considered as hormone independent.36-38
Ovarian stromal hyperplasia and androgens
A relation may exist between the production of ovarian androgens and the degree
of ovarian stromal hyperplasia.32•39 Using the criteria developed by Boss, defining
stromal hyperplasia of the ovary as nodular or diffuse proliferation of ovarian
stroma, higher degrees of ovarian stromal hyperplasia were found in endometrioid
endometrial cancer.40 Also, with increasing degrees of ovarian stromal hyperplasia
levels of ovarian vein androgens were higher.32•4 1
-43 So a causal relationship in the
origin of hormone dependent endometrial pathology may exist between ovarian
stromal hyperplasia, ovarian vein androgen levels and endometrioid endometrial
carcinoma, with the aromatase complex as the main linking enzyme.
Aromatase and endometrial cancer
The question arises what evidence is available relating aromatase directly to
endometrial cancer? Aromatase activity in endometrial cancer was first
demonstrated by Tseng.44 Aromatase activity was found not only in endometrial
cancer, but also in endometriosis and adenomyosis. 1 7•33
•45
-53 Aromatase
immunoreactivity was demonstrated m stromal cells of endometrioid
endometrial carcinoma. 45•46 It was demonstrated that intratumoral aromatase in
endometrial malignancy is associated with stromal invasion and is expressed
during the process of carcinoma-stroma interaction. Varying levels of aromatase
expression responsible for local estrogen synthesis in endometrial cancer may
contribute to malignant degeneration. Possibly, aromatase expression inhibiting
mechanisms may be lost during the process of neoplastic transformation or
aromatase-stimulating factors are activated. No correlation was found between
aromatase expression or activity and steroid status, clinical grade or histological
45
stage, although aromatase expression was more often reported in low grade, than
in high grade stromal endometrial sarcomas, and higher expression of aromatase
was more frequently associated with advanced disease or more poorly
differentiated endometrial cancer.33•48
•54 One study reported aromatase expression
in 65% of endometrial cancers. A correlation was found between aromatase and
cyclooxygenase-2 expression, but not with estrogen or progesteron receptors,
indicating that indeed intratumoural production of estrogens in endometrial
cancer may be an important mechanism in carcinogenesis. 55
The presence of aromatase in stromal or interstitial cells rather than in the
epithelial carcinoma cells suggests that locally produced estrogens may act on
carcinoma cells in a paracrine way .48 Thus, intratumoural estrogens originating
from in situ aromatization may function as an extra local carcinogenic factor in
an already existing malignancy. Since invasive carcinoma may induce aromatase
activity in stromal cells, local estrogen production in an existing endometrial
cancer could play a role in its progression. Even a small increase in the quantity
of local estrogen synthesis may be of significance for carcinogenesis or tumour
progression, as in the microenvironment of the tumour tissue estrogen receptors
are available within the malignant cells.47 It may be that the presence of
aromatase within the endometrial tissues creates a growth advantage
independently of circulating estrogen levels. If so, aromatase positivity may be
associated with a negative prognosis for survival chances. As mentioned before,
indeed evidence was found that a correlation exists between CYP 19 gene
expression (and aromatase expression) and the so-called hormone independent
endometrial cancer types, which are generally considered to have a worse
prognosis than well differentiated endometrioid endometrial cancer.36-38 Increased
aromatase expression in breast and endometrial cancer, leiomyomas and
endometriosis do not give rise to elevated estrogen levels in the peripheral
blood. The relatively high aromatase activity in leiomyoma-derived smooth
46
muscle cells and (adipose) fibroblasts surrounding breast tumours suggest the
presence of significant quantities of locally produced estrogens in these tissues. 1 7
In human endometrial cancer cell lines a high aromatase activity correlated with
a high DNA synthesis responsiveness to steroid hormones, both estradiol and
testosterone. 46•56 Aromatase immunoreactivity was found in 67% of stromal cells
of endometrioid endometrial carcinoma, but not in hyperplastic or normal
endometrium. 48 Because marked aromatase expression was detected at the sites
of stromal invasion a role for aromatase in invasion was suggested.
Anti-estrogens and arom.atase inhibitors in breast and endometrial cancer
Effects of anti-estrogens like tamoxifen and aromatase inhibitors are illustrative,
as both may influence breast and endometrial cancer tissue in a different way.
Aromatase inhibitors suppress endogenous estrogen production without directly
interacting with estrogen receptors, and have a different mode of action than the
anti-estrogen, tamoxifen, which blocks estrogen receptors but leaves
endogenous hormone levels unchanged. Tamoxifen is a selective estrogen
receptor modulator with tissue-specific effects on breast and endometrium.
Tamoxifen competes with estradiol for binding to the estrogen receptor and
changes the receptor in such a way that downstream gene transcription is
inhibited. Possibly, intratumoural estrogens increase estrogen receptor gene
expression. It has been observed that breast tumours can become estrogen
receptor negative and tamoxifen resistant in the course of disease progression
and treatment. 1 6 Also in atypical hyperplasia of the endometrium and
endometrial cancer estrogen and progesterone receptor density decrease with
increasing histological grade. 57•58 In the endometrium, tamoxifen effects depend
on estradiol concentration: only in postmenopausal women estrogen-agonistic
properties predominate and the risk of developing endometrial cancer
increases. 59 As aromatase seems to play a significant role both in the
47
/
development of breast cancer and endometrial cancer, treatment targets may be
shared. The ATAC trial is a large phase III trial comparing primary adjuvant
tamoxifen with the aromatase inhibitor anastrozole for early breast cancer in
postmenopausal women.60 In the anastrozole group significantly less cases of
endometrial cancer or endometrial abnormalities occured, as compared to the
rate for the general population, suggesting that anastrozole offers a protective
effect on the endometrium.6 1•62 Also, median endometrial thickness appeared to
increase in the tamoxifen group, whereas it remained unchanged in the
anastrozole group, which could mean a protective role for aromatase inhibitors. 63
One may add that in breast cancer patients increased endometrial thickness,
related to treatment with tamoxifen after menopause, reversed following
treatment with aromatase inhibitors. 64 These findings are consistent with the
lack of correlation between aromatase activity and expression of estrogen and
progesteron receptors in breast and endometrial cancer. 1 6• 1 8• 1 9•48•65
In conclusion, pathophysiologic and therapeutic similarities and differences exist
between breast and endometrioid endometrial cancer. Both partly find their
origin in excessive exposure to unapposed estrogens and develop from
premalignant hyperproliferative disorders. Progestins may increase breast cancer
risk, but protect the endometrial tissue. Anti-estrogens like tamoxifen have
tissue-specific effects with estrogen-agonistic properties on the postmenopausal
endometrium, whereas tamoxifen blocks estrogen receptors in breast cancer.
Intratumoural estrogen production under the influence of the enzyme complex
aromatase is an important factor for the promotion of breast cancer growth, and
may similarly be of importance in the pathogenesis of endometrial
cancer. 1 8•32•33•39 The presence of aromatase within the endometrium may appear
to form a negative prognostic factor independently of hormone receptor
positivity with future implications for the further development and introduction
48
of aromatase inhibitors in treatment strategies of advanced or recurrent
endometrial cancer.
Aromatase inhibitors in advanced or recurrent endometrial carcinoma
Late stage endometrial carcinoma
Of the female genital tract malignancies endometrial cancer is the most frequent,
with a lifetime risk of 1.2% for women. Annual incidence rates between 10 to
130 per 100.000 women are found.66'67 More than seventy-five percent of all
cases of endometrial carcinoma involves endometrioid endometrial carcinoma. 68
The mean age of women with endometrioid carcinoma is 59 years. The FIGO
stage is the surgical-pathological indicator of progression of the disease at the
time of diagnosis, and stage and age determine treatment. Staging includes
assessment of pelvic and paraaortic lymph nodes, adnexa, and peritoneal fluid
cytology, and requires a.o. histological evaluation of the tumour grade and depth
of invasion in the myometrium, but staging fails to assess individual risks of
local recurrence or distant metastases in the future. Most patients (7 5 to 80%) of
women present with a stage I or II malignancy with a total mortality of 10 to
25%. Creasman reports a 5-year survival of 9 1 % for stage Ia, 88% for stage lb
and 8 1 % for stage le. In stage II surival after 5 years drops to 67-77%, and to
52-60% in stage III. In case of distant metastases only 5% of affected women is
alive after 5 years. As most patients are diagnosed in early stages overall 5-year
survival was reported to be 73%.69 In a population-based Norwegian study, the
5- and 10-year survival for all FI GO-stages was 78% and 67%, respectively.70
For stage III patients these percentages were 39% and 34%, and for stage IV
patients these rates were only 27% and 23%. As patients often present with early
symptoms, such as vaginal bleeding, more than 80% of cases are diagnosed in
the early stages with coherent better prognosis. At diagnosis, 3% of women
present with a clinical stage IV, and about 4% have stage IV disease. Despite
49
treatment with surgery and radiotherapy, 50% of stage III tumours recur. Distant
metastases and local recurrence are major problems. Spread to the lungs occurs
in 36% of women. One must realize that these survival precentages may be
influenced by the fact that most patients present with endometrioid endometrial
cancer with a more favourable prognosis in the early, often curable stages.
In general, all patients undergo a hysterectomy and bilateral salpingo
oophorectomy when feasible, supplemented by surgical staging and pelvic
radiotherapy is administered to patients with poor prognostic factors that
increase risk of recurrence. However, local recurrence was noted in 4% of
patients even if adequate radiotherapy was administered. 71 Adjuvant hormonal or
cytotoxic chemotherapy and palliative radiotherapy hardly contribute to cure
rates, and significant side effects are noted. One may conclude, that in case
surgery in combination with radiotherapy fails to prevent recurrence, or if the
patient presents with a stage III or IV disease, prognosis remains poor in spite of
adjuvant therapy. Introduction of additional hormonal treatment possibilities
would be welcomed.
Palliative honnonal treatment
For recurrent or advanced hormone dependent cancer treatment is aimed at
slowing down tumour progression as long as possible, or to reduce symptoms of
local metastases. Endocrinological treatment aims at disease stabilisation and
preservation of quality of life, in the knowledge that disease progression is
unavoidable in the end. Particularly for women with well differentiated and/or
progesteron receptor positive endometrial cancers high-dose progestins like
megestrol acetate have been used with limited success. 72 Objective temporary
reponse rates of 10 to 25% were reported in advanced or recurrent disease. 73-75 In
spite of this relative success, one may also conclude that the therapy proved
ineffective in a high percentage of patients. Unfortunately, poor differentiation
50
and progression of endometrial cancers are associated with the loss of
progesterone receptor expression, which partly explaines the limited success of
progestin therapy in these patients. 76 Progesterone receptor positive tumours
show a six times higher response to hormonal treatment, when compared to
receptor negative tumours.25 The anti-estrogen tamoxifen, used a.o. as an
endocrine treatment for advanced endometrial cancer with objective responses
between O and 53%, was associated with an increased risk of endometrial
cancer, so this strategy was abandoned. 59•73 Growth inhibition of advanced
endometrial cancer can theoretically be achieved by treating patients with GnRH
analogs, but disappointing response rates between O and 35% were found.75•77
The potential of aromatase inhibitors
At present two types of aromatase inhibitors are prescribed to patients. Type I
steroidal aromatase inhibitors act as substrate analogues by binding to the
androgen substrate binding site, causing an increase of inactivated aromatase
enzymes. 1 8•78 The non-steroidal type II inhibitors bind to the haem group of the
aromatase enzyme complex and prevent NADPH utilisation, which obstructs
hydroxylation, and induces a decrease of estrogens. The first and most well
known inhibitor was the type II inhibitor aminoglutethimide.
Aminoglutethimide had several disadvantages like side-effects and a limited
specificity. New type I and II aromatase inhibitors were developed with more
binding affinity towards aromatase. These inhibitors like arimidex, vorozole and
letrozole appeared very powerfull in reducing estrogen levels. 18•79-83 Also, as
these new aromatase inhibitors showed less side-effects this was a major
improvement in the treatment of breast cancer. 18 Nowadays, these second
generation aromatase inhibitors are frequently used in the treatment of estrogen
dependent diseases such as breast cancer and also even in endometriosis. 84
Aromatase is overexpressed in endometriosis and reduction of aromatase
5 1
activity in endometriosis most likely diminishes estradiol production with
subsequent improvement of clinical symptoms. 85
The adjuvant use of aromatase inhibitors may offer new treatment approaches in
advanced or recurrent endometrial cancer with poor prognosis. The potent and
specific characteristics of the latest aromatase inhibitors could improve response
rates by suppressing estrogen levels in endometrial cancer with minimal side
effects. Theoretically, aromatase blockers may inhibit both peripheral and
endometrial stromal aromatization, as the aromatase enzyme converting
androgens into estrogens appears present in all tissues. There exists some
controversy about the effects of aromatase inhibitors on breast cancer growth, as
some authors report local effects in and around the tumour, while others claim
suppression of total body aromatase activity, shown by a reduction of serum
estrogen levels. 1 8 The same controversy may exist in endometrial cancer, as both
mechanisms may apply for endometrial cancer as well. Aromatase inhibitors
may suppress estrogen levels more than by e.g. surgical adrenalectomy or
hypophysectomy alone. 86 Although the exact pathophysiology is unknown, most
likely tumour cell apoptosis is increased due to estrogen deprivation. Aromatase
inhibitors do not influence the estrogen receptor status of tumour cells in human
endometrial carcinomas, contrary to the estrogen agonistic properties of
tamoxifen. 1 9 Even endometrial thickness is reduced in patients treated with
aromatase inhibitors, concluding that aromatase inhibitors do not induce
endometrial hyperplasia as a side-effect in contrast to tamoxifen.60-63
•87 In vitro
it was shown that aromatase inhibitors reduce proliferation and increase
apoptosis m endometrial cancer. 1 9•88 Sasano detected aromatase
immunoreactivity in stromal cells of six out of 14 cases of endometrial
carcinoma. 1 9 The addition of testosterone as a substrate for aromatase, resulted
in an increase of aromatase activity in two cases. Addition of an aromatase
inhibitor, NKSO 1, blocked the [3H]Thymidine incorporation stimulated by
testosterone in four out of 15 endometrioid carcinomas. Some studies reported
52
objective responses and a decrease of aromatase activity after administration of
aromatase inhibitors to women with advanced endometrial cancer. Already in
1 984 objective responses to the aromatase inhibitor aminoglutethimide were
found in four out of 1 8 women with metastatic endometrial carcinoma. 89 A small
study showed decrease of aromatase activity after administration of the
aromatase inhibitor letrozole.90 One study reported very limited response rates to
aromatase inhibition in a phase II trial in patients with advanced or recurrent
endometrial cancer. Only two partial responses were found in 23 women with
advanced or recurrent endometrial malignancy, of whom most cancers were
poorly differentiated or presented an agressive histology ( clear cell carcinoma or
papillary serous carcinoma). Only 39% of these patients were diagnosed with
(hormone-dependent) endometrioid endometrial cancer, and women with poorly
differentiated tumours did not respond.9 1 One may argue that patient selection
and short duration of treatment (four weeks) influenced results negatively. In
breast cancer the tumours with no or low estrogen receptors do not require
estrogen for their continued growth.92 These tumours are unlikely to respond to
therapy designed to block estrogen synthesis. Women with estrogen receptor
negative breast tumours are not eligible for endocrine treatment. This may
possibly apply for some endometrial cancer as well. Poorly differentiated and
agressive tumours like serous papillary and clear cell carcinoma rarely show
estrogen or progesterone receptors and hardly respond to endocrinological
therapy like progestins.46 Unfortunately most women to be considered for
treatment with adjuvant therapy are those with advanced or recurrent aggressive
tumours, who simultaneously are not expected to respond to treatment with
endocrinological drugs, as they are most often hormone receptor negative. On
the other hand, no persistent correlation was found between steroid receptor
status and aromatase expression or activity, in accordance with the absence of
correlation between aromatase activity and steroid receptor expression in breast
carcinoma.48•65
•93 So it remains doubtfull whether estrogen or progesterone
53
receptor positivity should be used as a suitable criterion justifying treatment
with aromatase inhibitors. This would mean that aromatase inhibitors possibly
can be used also in estrogen and progesterone negative tumours with a suspected
poor prognosis. As a correlation was found between CYP 19 gene expression and
hormone independent endometrial cancer, aromatase positivity may be an
important marker relating to tumour growth due to hormonal stimulation from
locallly produced estrogens, with prognostic and therapeutic consequences
independently of estrogen receptor status.36-38 In a recent phase II trial the
aromatase inhibitor letrozole was administered for 12 weeks to patients with
recurrent or advanced endometrial cancer, of whom 86% appeared positive for
estrogen or progesteron receptor.94 In 28 women overall response was 34%. One
complete and two partial responses were noted, whereas 1 1 patients had a stable
disease for a median duration of 6. 7 months, suggesting that a proportion of the
population could benefit. A small report describes two women with endometrial
stromal sarcoma. Both patients developed pulmonar metastases, respectively
three and six years after surgery. Both patients achieved a complete response
after treatment with the aromatase inhibitor aminoglutethimide.95 Successful
response on treatment with letrozole was reported in advanced low-grade
endometrial stromal sarcoma, corresponding with the observation that aromatase
expression was more often reported in low grade, than in high grade stromal
endometrial sarcomas, and that higher expression of aromatase was more
frequently associated with advanced disease and poorer differentiation.33•54
•96
Two premenopausal obese women with well-differentiated endometrial cancer
were successfully treated with progesterone and anastrozole, with the purpose of
maintaining fertility.97 Recently Berstein et al found a decrease in aromatase
activity in endometrial cancer tissue after two weeks of treatment with letrozole
and exemestane. 37
54
Possibly the classification of endometrial cancers in two distinctive classes on
morphologic features, endometrioid endometrial carcinoma related to
unopposed estrogenic stimulation, and a second estrogen independent form
including poorly differentiated and high-grade malignancies like serous
papillary and clear cell carcinoma, should be subdivided in a new model
according to new endocrinological criteria; aromatase positive and aromatase
negative endometrial cancers may exist among both classes of endometrial
cancer, irrespective of body weight, or estrogenic or hormone receptor status,
but related more to genetic and molecular predispositions like CYP 19
polymorphisms. Several pathways on molecular levels may overlap the origin
and evolution of endometrial carcinogenesis and influence each other.
In conclusion, only limited evidence is available concerning the use of
aromatase inhibitors in advanced endometrial cancer. But what about the use of
aromatase inhibitors in the early treatment of endometrial cancer? The AT AC
trial is a large phase III trial comparing primary adjuvant tamoxifen with the
aromatase inhibitor anastrozole for early breast cancer in postmenopausal
women. 60•98 In the anastrozole group less cases of endometrial abnormalities or
cancer occured during two years treatment with anastrozole compared with
tamoxifen, suggesting that anastrozole offers some protective effect on the
endometrial tissue. 61 Also, median endometrial thickness appeared to increase
in the tamoxifen group, whereas it remained unchanged in the anastrozole
group, suggesting a protective role for aromatase inhibitors. 63 Another study
found that after treatment with aromatase inhibitors endometrial abnormalities
due to treatment with tamoxifen diminished in breast canbcer patients. 64 Thus,
one may consider to use aromatase inhibitors in the treatment of complex
atypical endometrial hyperplasia in order to prevent progression of atypia.
However, conservative treatment of atypical hyperplasia can only be achieved
under adequate hysteroscopic monitoring, and one should realize that almost
55
20% of atypical hyperplasias show a carcinoma m their hysterectomy
specimens. 74
As in breast cancer a positive but inconsistent correlation was found between
intratumoural aromatase activity and response to treatment with aromatase
inhibitors, reliable methods must be developed for detection of aromatase
positivity in the individual patient, in order to select those patients who may
benefit from aromatase inhibition. 99- 1 0 1 Taking into account the need for
treatment of patients with advanced or recurrent endometrial disease who are
often elderly and who present with some comorbidity, and the present lack of
long-lasting responses in the adjuvant treatment of endometrial cancer, the
usefulness of treatment with aromatase inhibitors in advanced endometrial
cancer ought to be evaluated. If so, a simple and quantifiable test for
measurement of aromatase activity in formalin-fixed or paraffin-embedded
tissue would be needed for prediction of possible responses to aromatase
inhibitors.54•99 In breast cancer expression of intratumoural aromatase was found
to show to have prognostic value concerning the chances of successful treatment
with aromatase inhibitors.100,101 Intratumoural aromatase expression was related
to poor survival in endometrioid endometrial cancer, but confirmation from
other studies is still indicated. 1 02 A reliable simple test for measurement of
aromatase activity in archival materials or formalin-fixed and paraffin-embedded
tissue would be welcomed for prediction of sensitivity to treatment with aromatase
inhibitors. Conflicting results were found in previous studies using
immunohistochemistry, and the most important problem appeared the fact that
specificity of antibodies could not be characterized well enough.99 New
monoclonal antibodies for immunohistochemistry, such as the 677 clones, seem
quite useful for detection of aromatase activity not only in breast cancer, but also
in endometrial carcinoma.99•103 The extent of aromatase positivity in these
tumours may offer new endocrinological treatment modalities in the
56
management of endometrial carcinoma and new insights m individual
prognostic factors and chances of disease-free survival.
57
References
1. Reed MJ, Owen AM, Lai LC, et al. In situ oestrone synthesis in normal breast and breast tumour tissues: effect of treatment with 4-hydroxyandrostenedione. Int J Cancer
1989; 44 : 233-7. 2. Korzekwa KR, Trager WF, Mancewicz J, Osawa Y. Studies on the mechanism of
aromatase and other cytochrome P450 mediated deformylation reactions. J Steroid
Biochem Molec Biol 1993; 44 : 367-73 . 3 . Adashi EY. The climacteric ovary as a functional gonadotropin-driven androgen
producing gland. Fertil Steril 1994, 62, 20-7.
4 . Nelson LR, Bulun SE. Estrogen production and action. J Am Acad Dermatol 2001; 45 : S 1 16-24.
5. Brueggemeier RW. Aromatase, aromatase inhibitors, and breast cancer. Am J of
Therapeutics 2001; 8 : 333-44. 6. Henderson BE, Feigelson HS. Hormonal carcinogenesis. Carcinogenesis 2000; 21 :
427-33 . 7. Kurman RJ, Kaminski PF, Norris HJ. The behaviour of endometrial hyperplasia. A long
term study ofuntreated 'hyperplasia ' in 170 patients. Cancer 1985; 56: 403- 12. 8. Wellings SR, Jensen HM, Marcum RG. An atlas of subgross pathology of the human
breast with special reference to possible precancerous lesions. J Natl Cancer Inst 1975;
55 : 23 1 . 9 . Berardo M, Hilsenbeck SG, Allred DC. Histological grading of non-invasive breast
cancer and its relationship to biological features. Lab Invest 1996; 74 : 68. 10. Santen RJ, Manni A, Harvey H, Redmond C. Endocrine treatment of breast cancer in
women. Endocr Rev 1990; 1 1 : 1-45. 1 1. Osborne CK. Steroid hormone receptors in breast cancer management . Breast Cancer
Res Treat 1998; 5 1 : 227-38.
12. Blankenstein MA, van de Ven J, Maitimu-Smeele I, Danker GH, de Jong PC,
Daroszewski J, Szymczak J, Milewicz A, Thijssen JHH. Intratumoral levels of
estrogens in breast cancer. J Steroid Biochem Molec Biol 1999; 69 : 293-7. 13. van Landeghem AAJ, Poortman J, Nabuurs M, Thijssen JH. Endogenous
concentration and subcellular distribution of estrogens in normal and malignant human
breast tissue. Cancer Res 1985 ; 45 : 2900-6.
58
14. Miller WR, Mullen P, Sourdaine P, Watson C, Dixon JM, Telford J. Regulation of aromatase activity within the breast. J Steroid Biochem Malec Biol 1997; 61 : 193-202.
15 . Lipton A, Santner SJ, Santen RJ, Harvey HA, Feil PD, White-Hershey D,
Bartholomew MJ, Antle CE. Aromatase activity in primary and metastatic human
breast cancer. Cancer 1987; 59 : 779-82.
16. van de Ven J, Danker GH, Sprong M, Blankenstein MA, Thijssen JHH. Effect of
tibolone (Org OD14) and its metabolites on aromatase and estrone sulfatase activity in human breast adipose stromal cells and in MCF-7 and T47D breast cancer cells. J
Steroid Biochem Malec Biol 2002; 8 1 : 237-42. 17. Bulun SE, Noble LS, Takayama K, Michael MD, Agarwal V, Fisher C, et al.
Endocrine disorders associated with inappropriately high aromatase expression. J
Steroid Biochem Malec Biol 1997; 61 : 133-9 . 18 . De Jong P Chr, Blankenstein MA, van de Ven J, Nortier, JWR, Blijham GH, Thijssen
JHH. The importance of local aromatase activity in hormone dependent breast cancer: a review. The Breast 2001; 10: 9 1-9.
19. Sasano H, Sato S, Ito K, Yajima A, Nakamura J, Yoshihama M, Ariga K, Anderson TJ, Miller WR. Effects of aromatase inhibitors on the pathobiology of human breast,
endometrial and ovarian carcinoma. Endocr Rel Cancer 1999; 6: 197-204.
20. Sasano H, Harada N. Intratumoral aromatase in human breast, endometrial and
ovarian malignancies. Endocr Rev 1998; 19 : 593-607.
2 1. Miller WR, Anderson TJ, Jack WJ. Relationship between tumour aromatase activity, tumour characteristics and response to therapy. J Steroid Biochem Malec Biol 1990;
37: 1055-9 .
22. Cline JM, Soderqvist G, von Schoultz E, Skoog L, von Schoultz B. Effects of
hormone replacement therapy on the mammary gland of surgically postmenopausal
macaques. Am J Obstet Gynecol I 996; 174 : 93- 100.
23 . Campagnoli C, Abba C, Ambroggio S, Peris C. Pregnancy, progesterone and progestins in relation to breast cancer risk. J Steroid BiochemMolec Biol 2005; 97: 44 1-50.
24. Bokhman JV. Two pathogenetic types of endometrial carcinoma. Gynecol Oneal 1983; 15 : 10-7.
25. Pijnenborg JMA, Romano A, Dam-de Veen GC, Dunselman GAJ, Fischer DG,
Groothuis PG, et al. Aberrations in the progesterone receptor gene and the risk of
recurrent endometrial carcinoma. J Pathol 2005; 205 : 597-605.
59
26. Sluijmer AV.The postmenopausalovary. Functional and morpholologicalaspects. Thesis Maastricht 1999. ISBN 90-9012802-6.
27. Yamamoto T, Kitawaki J, Urabe M, Honjo H, Tamura T, Noguchi T, et al. Estrogen productivity of endometrium and endometrial cancer tissue; influence of aromatase on
proliferation of endometrial cancer cells. J Steroid Biochem Malec Biol 1993 ; 44 : 463-8 .
28. Mertens HJMM, Heineman MJ, de Jong FH, et al. Postmenopausal ovarian function and endometrialcancer.Hum Reprod 14, Abstractbook,Uune 1999): 89-90.
29. Heineman MJ, Sluijmer AV, Kampschoer PHNM, et al. Utero-ovarian vein blood smapling in postmenopausal women with and without endometrial cancer. Fertil Steril
1995; Suppl: S92 (Abstract O-187).
30. Nagamani M, Stuart CA, Doherty MG. Increased steroid production by the ovarian stromal tissue of postmenopausal women with endometrial cancer. J Clin Endocrinol Metab 1992; 74 : 172-6.
3 1 . Nagamani M, Hannigan EV, Dillard EA jr, Dinh TV. Ovarian steroid secretion in postmenopausal women with and without endometrial cancer. J Clin Endocrinol Metab 1986; 62: 508- 12.
32. Jongen VHWM, Hollema H, Santema JG, van der Zee AGJ, Heineman MJ. Endometrioidendometrial cancer, ovarian stromal hyperplasia and steroid production. Br J Obstet Gynaecol 2003; 1 10: 690-5 .
33. Jongen VHWM, Thijssen JHH, Hollema H, Donker GH, Santema JG, van der Zee AGJ, Heineman MJ. Is aromatase P450 involved in pathogenesis of endometrioid endometrial
cancer? Int J Gynecol Cancer2005; 15 : 527-34. 34 . Longcope C. Endocrine function of the postmenopausal ovary. J Soc Gynecol Investig
2001; 8: S67-8.
35 . Paynter RA, Hankinson SE, Colditz GA, Kraft P, Hunter DJ, De Vivo I. CYP19
( aromatase) haplotypes and endometrial cancer risk. Int J Cancer 2005; 1 16: 267-74.
36. Berstein LM, Imyanitov EN, Suspitsin EN, Grigoriev MY, Sokolov EP, Togo A, et al. CYPl 9 gene polymorphism in endometrial cancer patients. J Cancer Res Clin Oneal
2001; 127: 135-8. 37. Berstein L, Kovalevskij A, Zimarina T, Maximov S, Gershfeld E, Vasilyev D, et al.
Aromatase and comparative response to its inhibitors in two types of endometrial cancer. J Steroid Biochem Malec Biol 2005; 95 : 7 1-4.
60
38. Berstein L, Zimarina T, Kovalevskij A, Maximov S, VolkovO, Thijssen JH. CYP 19 gene expression and aromatase activity in endometrial cancer tissue: importance of the type of
disease. Neoplasma2005; 52: 1 15-8 .
39. Jongen VHWM, Sluijmer AV, Heineman MJ. The postmenopausal ovary as an androgen-producing gland; hypothesis on the etiology of endometrial cancer. Maturitas 2002; 43 : 77-85 .
40. Boss JH, Scully RE, Wagner KH, et al. Structural variations in the adult ovary: clinical
significance. Obstet Gynecol 1965; 25 : 747-64. 4 1. Sluijmer A V,Heineman MJ, Koudstaal J, et al. Relationship between ovarian production
of estrone, estradiol, testosteron, and androstenedione and the ovarian degree of stromal
hyperplasia in postmenopausalwomen. Menopause 1998; 5: 207-10.
42. Lucisano A, Russo N, Acampora MG, et al. Ovarian and peripheral androgen and estrogen levels in post-menopausal women: correlations with ovarian histology.
Maturitas 1986; 8: 57-65.
43. Ala-Fossi SL, Maenpaa J, Aine R, et al. Ovarian testosterone secretion during
perimenopause.Maturitas 1998; 29 : 239-45.
44 . Tseng L, Mazella J, Mann WJ, Chumas J. Estrogen synthesis in normal and malignant
human endometrium. J Clin Endocrinol Metab 1982; 55 : 1029-3 1 . 45 . Sasano H, Kaga K, Sato S , Yaj ima A, Nagura H, Harada N. Aromatase P450 gene
expression in endometrial carcinoma. Br J Cancer 1996; 74 : 154 1-4.
46. Yamamoto T, Kitawaki J, Urabe M, Honjo H, Tamura T, Noguchi T, et al. Estrogen
productivity of endometrium and endometrial cancer tissue; influence of aromatase on
proliferation of endometrial cancer cells. J Steroid Biochem Molec Biol 1993; 44 : 463-8.
47. Bulun SE, Economos K, Miller D, Simpson ER. CYP 19 (aromatase cytochrome P450) gene expression in human malignant endometrial tumors. J Clin Endocrinol
Metab 1994; 79 : 183 1-4.
48. WatanabeK, Sasano H, Harada N, Ozaki M, Niikura H, Sato S, Yaj imaA. Aromatase in human endometrial carcinoma and hyperplasia. Immunohistochemical, in situ
hybridization, and biochemical studies . Am J Pathol 1995; 146: 49 1-500.
49. Bulun SE, Mahendroo MS, Simpson ER. Polymerase chain reaction amplification fails to detect aromatase cytochrome P450 transcripts in normal human endometrium
or decidua. J Clin Endocrinol Metab 1993; 76: 1458-63.
50. Noble LS, Simpson ER, Johns A, Bulun SE. Aromatase expression in endometriosis. J
Clin Endocrinol Metab 1996; 8 1 : 174-9.
6 1
5 1. Kitawaki J, Noguchi T, Amatsu T, Maeda K, Tsukamoto K, Yamamoto T, et al. Expression of aromatase cytochrome P450 protein and messenger ribonucleic acid in human endometriotic and adenomyotic tissues but not in normal endometrium. Biol
Reprod 1997; 57: 5 14-9.
52. Prefontaine M, Shih C, Pan CC, Bhavnani BR. Applicability of the product isolation and radiometric aromatase assays for the measurement of low levels of aromatase: lack of aromatase activity in the human endometrium. J Endocrinol 1990; 127: 539-
5 1 .
53 . Baxendale PM, Reed MJ, James VHT. Inability of human endometrium or
myometrium to aromatize androstenedione. J Steroid Biochem Molec Biol 198 1; 14 :
305-6.
54. Reich 0, Regauer S. Aromatase expression in low-grade endometrial stromal sarcoma; an immunohistochemical study. Mod Pathol 2004; 17( 1) : 104-8.
55 . Fowler JM, Ramirez N, Cohn DE, Kelbick N, Pavelka J, Ben-Shachar I, et al. Correlation of cyclooxygenase-2 (COX-2) and aromatase in human endometrial cancer: tissue microarrayanalysis. Am J Obstet Gynecol 2005 : 192: 1262-7 1 .
56. Tada A, Sasaki H, Nakamura J, Yoshihama M, Terashima Y . Aromatase activity and the effect of estradiol and testosterone on DNA synthesis in endometrial carcinoma
cell lines. J Steroid Biochem Molec Biol 1993; 44 : 66 1 -6.
57. Mertens HJMM. Observational studies of human endometrium and endometrial cancer. The role of steroids and steroid hormone receptors in tissue proliferation, differentiation
and malignant transformation. PhD Thesis 2002 University of Maastricht, The
Netherlands.
58. McCarthy KS, Barton TK, Fetter BF, Creasman WT. Correlation of estrogen and progesterone receptors with histological differentiation in endometrial adenocarcinoma. Am J Pathol 1979; 96: 17 1-83 .
59 . Mourits MJE, Ten Hoor KA, van der Zee AGJ, Willemse PHB, de Vries EGE, Hollema H. The effects of Tamoxifen on proliferation and steroid receptor expression
in postmenopausal endometrium. J Clin Pathol 2002; 55 : 5 14-9.
60. The AT AC Trialists' Group. Anastrazole alone or in combination with tamoxifen versus tamoxifen alone for adjuvant treatment of postmenopausal woman with early breast cancer: first results of the ATAC randomized trial. Lancet 2002; 359: 2 13 1-9.
62
61 . Duffy SRG, Greenwood M. The endometrial data from the ATAC (Arimidex,
Tamoxifen, Alone or in Combination) trial indicates a protective effect of anastrozole
(arimidex) upon the endometrium. Breast Cancer Res Treat 2003; 83 : abs 134 . 62. Duffy S, Jackson TL, Lansdown M, Philips K, WellsM, Pollard S, Clack G, Coibon M,
Bianco AR. The AT AC (' Arimidex', Tamoxifen, Alone or in Combination) adjuvant
breast cancer trial; first results of the endometrial sub-protocol following 2 years of
treatment. Hum Reprod 2006; 21 (2): 545-53 .
63 . Duffy SRG, Jackson T. Fewer endometrial abnormalities with anastrozole than tamoxifen: Endometrial subprotocol results from the ATAC (Arimidex, Tamoxifen, Alone or in Combination) early breast cancer (EBC) trial in postmenopausal (PM)
patients (pts). Ann Oncol 2002; 13 : abs 139.
64. Morales L, Timmerman D, Neven P, Konstantinovic ML, Carbonez A, van Huff el S, et al. Third generation aromatase inhibitors may prevent endometrial growth and reverse
tamoxifen-induceduterine changes in postmenopausal breast cancer patients. Ann Oncol 2005; 16: 70-4.
65. Sasano H, Nagura H, HaradaN, Goukon Y,Kimura M. Immunolocalizationofaromatase
and other steroidogenicenzyme in human breast disorders. Hum Pathol 1994; 25 : 530-5 . 66. Lalloo F, Evans G. Molecular genetics and endometrial cancer. Best practice &
research. Clin Obstet & Gynaecol 2001 ; 15 : 355-63 .
67. Visser 0, Coebergh JWW, Schouten LJ, van Dijk JAAM. Incidence of cancer in the Netherlands 1995 . Netherlands Cancer Registry, Utrecht, pp. 43.
68. Blaustein. Pathology of the female genital tract. Springer-Verlag New York Inc. 1994,
4th ed.
69. Creasman WT. Endometrial cancer: incidence, prognostic factors, diagnosis and
treatment. Semin Oncol 1997; 24: SI 140- SI 150. 70. Abeler VM, Kjorstad KE, Berle E. Carcinoma of the endometrium in Norway: a
histopathological and prognostic survey of a total population. Int J Gynecol Cancer 1992; 2: 9-22.
7 1 . Creutzberg CL, van Putten WLJ, Koper PCM, Lybeert MLM, Jobs en JJ, et al. Surgery and postoperative radiotherapy versus surgery alone for patients with stage- I endometrial
carcinoma: multcenterrandomised trial. Lancet 2000; 355 : 1404- 1 1 .
72. Kaupilla A. Oestrogen and progestin receptors as prognostic indicators in endometrial
cancer. A review of the literature. Acta Oncol 1989; 28: 561-6.
63
73 . Moore TD, Phillips PH, Nerenstone SR, Cheson BD. Systemic treatment of advanced
and recurrent endometrial carcinoma: current status and future directions. J Clin Oncol 199 1 ; 9: 107 1 -88.
74 . 74.Quinn MA. Hormonal treatment of endometrial cancer. Hematology/Oncology
Clinic ofNorth America 1999; 1 3 : 163-87. 75 . Gallagher CJ, Oliver RTD, Oram DH, Fowler CG, Blake PR, Mantell BS. A new
treatment for endometrial cancer with gonadotropin releasing hormone analogue. Br J
Obstet Gynaecol 199 1 ; 98 : 1037-4 1 . 76. Fujimoto J, lchigo S , Hirose R, Sakaguchi H, Tamaya T. Clinical implication of
expression of progesterone receptor form A and B mRNAs in secondary spreading of
gynecologic cancers. J Steroid Biochem Molec Biol 1997; 62: 449-54.
77. Covens A, Thomas G, Shaw P, Ackerman I, Osborne R, Lukka H. A phase II study of
leuprolide in advanced/recurrent endometrial cancer. Gynecol Oncol 1997; 64: 126-9 . 78. Brodie AM, Njar VC. Aromatase and breast cancer. Semin Oncol 1996; 23 : 10-20.
79. Demers LM, Lipton A, Harvey HA, Kambic KB, Grossberg H, Brady C, Santen RJ. The effcacy of CGS 20267 in suppressing estrogen biosynthesis in patients with
advanced stage breast cancer. J Steroid Biochem Molec Biol 1993; 44: 687-9 1 .
80. Iveson TJ, Smith IE, Ahern J, Smithers DA, Trunet PF, Dowsett M. Phase I study of the oral nonsteroidal aromatase inhibitor CGS 20267 in postmenopausal patients with
advanced breast cancer. Cancer Res 1993 ; 53 : 166-70. 8 1 . Plourde PV, Dryoff M, Dukes M. Arimidex: a potent and selective fourth-generation
aromatase inhibitor. Breast Cancer Res Treat 1994; 30: 1 03- 1 1 . 82. Johnston SRD, Smith IE, Doody D, Jacobs S, Robershaw H, Dowsett M. Clinical and
endocrine effects of the oral aromatase inhibitor Vorozole in postmenopausal patients
with advanced breast cancer. Cancer Res 1994; 54 : 5875-8 1 . 83 . Gosse PE, Gwyn KMEH. Current perspectives on aromatase inhibitors in breast
cancer. J Clin Oncol 1994; 12: 2460-70. 84. Bulun SE, Zeitoun KM, Takayama K, Simpson ER, Sasano H. Aromatase as a
therapeutic target in endometriosis. Trends Endocrinol Metab 2000; 1 1 : 22-7. 85 . Bulun SE, Yang S, Fang Z, Gurates B, Tamura M, Zhou J, et al. Role of aromatase in
endometrial disease. J Steroid Biochem Molec Biol 200 1 ; 79 : 19-25. 86. 86.Samojlik E, Santen RJ, Worgul TJ. Suppression of residual oestrogen production
with aminoglutethimide in women following surgical hypophysectomy or adrenalectomy. Clin Endocrinol 1984; 20: 43-5 1 .
64
87. Matsuda A, Higucha K, Karasawa M, Yoneyama S, Deguchi J, Miyamoto M. Fourteen
day oral combination dose toxicity study of CGS 16949A with 5-Fluororouracil or
tamoxifen in rats. J Toxicol Sciences 1997; 22: 1-5. 88. Morsi H, Leers M, Nap M, et al. Apoptosis and anti-apoptosis in oestrogen-receptor
negative endometrial cancer cells in response to anastrazole, 4-hydroxytamoxifen and
medroxyprogesterone acetate. Eur J Cancer 2000; 36: 1 12-3. 89. Murray RLM, Pitt P. Treatment of advanced metastatic breast cancer, carcinoma of the
prostate and endometrial cancer with aminogluthethimidine. In: Nagel GA & Santen
RJ, ed.: Aminogluthethimidine as an aromatase inhibitor in the treatment of cancer.
Bern: Huber, 1984 : 109-22. 90. Berstein LM, Maksimov SI, Gershfel' d ED, Gamaiunova VB, Meshkova IE,
Kovalenko IG et al. Neoadjuvant use of the aromatase inhibitor letrozole in uterine
cancer: endocrine and clinical effects. Vopr Onkol 2001; 47(5): 5 17-74.
9 1. Rose PG, Brunetto VL, VanLe L, Bell J , Walker JL, Lee RB. A phase II trial of anastrozole in advanced recurrent or persistent endometrial carcinoma: a gynecologic oncology group study. Gynecol Oncol 2000; 78: 212-6.
92. Miller WR. Prediction of estrogen sensitivity/dependence. Estrogen and Breast
Cancer. Austin, texas: RG Landes Co, 1996; 15 1-69. 93. Ragaz J, Brodie a. Significant co-expression of Cox-2 and aromatase, with
implications on pathogenesis of human breast cancer, relevant to therapy with combination of aromatase and Cox2 inhibitors: long-term follow up of the British
Columbia Tissue Microarray study. Abstractbook Aromatase 2004, page 1 1. 94. Ma BBY, Oza A, Eisenhauer E, Stanimir G, Carey M, Chapman W, et al. The activity
of letrozole in patients with advanced or recurrent endometrial cancer and correlation
with biological markers - a study of the National cancer Institute of Canada Clinical
Trials Group. Int J Gynecol Cancer 2004; 14 : 650-8. 95. Spano JP, Soria JC, Kambouchner M, Piperno-Neuman S, Morin F, Morere JF, et al.
Long-term survival of patients given hormonal therapy for metastatic endometrial
stromal sarcoma. Med Oncol 2003 ; 20(1) : 87-93 . 96. Leunen M, Breugelmans M, De Sutter P, Bourgain C, Amy JJ. Low-grade endometrial
stromal sarcoma treated with the aromatase inhibitor letrozole. Gynecol Oncol 2004; 98 : 173-4.
97. Burnett AF, Bahador A, Amezcua C. Anastrozole, an aromatase inhibitor, and medroxyprogesterone acetate therapy in premenopausal obese women with endometrial
65
cancer: a report of two cases successfully treated without hysterectomy. Oynecol Oncol
2004; 94 : 832-4. 98. The ATAC (Arimidex, Tamoxifen Alone or in combination) Trialists 'Oroup.
Anastrazole alone or in combination with tamoxifen versus tamoxifen alone for adjuvant treatment of postmenopausal woman with early breast cancer: result of the
ATAC Trial efficacy and safety update analysis. Cancer 2003 ; 98: 1802-10.
99. Sasano H, Edwards DP, Anderson TJ, Silverberg SO, Evans DB, Santen RJ, et al. Validation of new aromatase monoclonal antibodies for immunohistochemistry:
progress report. J Steroid Biochem Molec Biol 2003 ; 86: 239-44.
100. Bezwoda WR, Mansoor R, Correlation of breast tumor aromatase activity and
response to aromatase inhibition with aminoglutethimide. Oncol 1987; 44 : 345-9. 101. Miller WR, Hawkins RA, Mullen P, Sourdaine P, Telford J. Aromatase inhibition:
determinants of response and resistance. Endocri Rel Cancer 1995; 2: 73-85.
102. Segawa T, Shozu M, Murakami K, Kasai T, Shinohara K, Nomura K, et al. Aromatase expression in stromal cells of endometrioid endometrial cancer correlates
with poor survival. Clin Cancer Res 2005 : 1 1 : 2188-94.
103. Sasano H, Anderson TJ, Silverberg SO, Santen RJ, Conway M, Edwards DP, et al. The validation of new aromatase monoclonal antibodies for immunohistochemistry--a correlation with biochemical activities in 46 cases of breast cancer. J Steroid Biochem Molec Biol 2005; 95 : 35-9.
66
4
Ovarian stromal hyperplasia and ovarian vein steroid
levels in relation to endometrioid endometrial cancer
VHWM Jongen
H Hollema
AGJ van der Zee
JG Santema
MJHeineman
British Journal of Obstetrics and Gynaecology2003; 1 10:690-95
67
Abstract
In a retrospective and prospective study, respectively, the relationship was studied
between the presence of endometrioid endometrial cancer, the degree of ovarian
stromal hyperplasia and ovarian steroid production in postmenopausal women . . In
1 12 women with endometrioid endometrial cancer, 47 women with a benign
gynaecological condition and ten women with non-endometrioid endometrial
cancer, the degree of ovarian stromal hyperplasia was scored retrospectively on a
semi-quantitative scale (atrophy, slight, marked). All women were postmenopausal
and had undergone a hysterectomy with bilateral salpingo-oophorectomy.
Prospectively also blood sampling from the ovarian veins was performed in 60
women. Steroid levels ( estrone, estradiol, androstenedione, testosterone) were
determined and related to the degree of ovarian stromal hyperplasia and the
presence (n=52) or absence (n=8) of endometrioid endometrial cancer. In the
retrospective study the degree of ovarian stromal hyperplasia was higher in the
presence of endometrioid endometrial cancer (p=0.000). The prospective study
showed that an increasing degree of ovarian stromal hyperplasia was related to
higher ovarian levels of both testosterone and androstenedione (p<0.05 and
p<0.005 respectively), but not to estrone or estradiol. Mean ovarian vein levels of
both testosterone and androstenedione were higher in patients with endometrial
cancer than in patients with benign conditions, without reaching statistical
significance. In conclusion, higher degrees of ovarian stromal hyperplasia were
found in endometrioid endometrial cancer and with increasing degrees of ovarian
stromal hyperplasia levels of ovarian vein androgens were higher. A causal
relationship in the origin of hormone-dependentendometrial pathology may exist
between ovarian stromal hyperplasia, ovarian vein androgen levels and
endometrioid endometrial carcinoma.
68
Introduction
Endometrioid endometrial cancer is generally regarded as an estrogen-related
malignancy gradually developing from endometrial atypical hyperplasia towards
endometrial cancer. Estrogens have been demonstrated to act as promoting factors,
especially in the relative absence of progestagens.1 •2 Obese women are at risk for
developing endometrioid endometrial carcinoma due to the increased capacity in
adipose tissue to convert androstenedione to estrone and testosterone to estradiol
under the influence of the enzyme complex cytochrome P450 aromatase. The
postmenopausal ovary is a viable androgen-producing gland, producing 40% and
20% of the total body amount of testosteron and androstenedione, respectively; It
has been suggested that ovarian androgens, apart from being precursors for
peripheral conversion to estrogens in adipose tissue, could also play an indirect
growth promoting role in the development of endometrial pathology, as the
enzyme aromatase cytochrome P450 may be able to convert circulating androgens
to estrogens in the endometrium itself.4•7 The clinical significance of ovarian
stromal hyperplasia in the postmenopausal ovary is controversial. In smaller
studies it has been suggested that the extent of ovarian stromal hypeiplasia
correlates with postmenopausal ovarian steroid production. 8· 1 1 Thus, a causal
relationship between ovarian stromal hypeiplasia, ovarian steroid production and
the presence of endometrioid endometrial cancer may exist and the aim of the
present study was to explore this relationship.
69
Methods
Retrospective study
A retrospective study was performed in all 323 postmenopausal women, who
underwent a hysterectomy with bilateral sapingo-oophorectomy at the University
Hospital of Groningen between 1990 and 2001. The clinical data were extracted
from patients records (Table 1). Excluded were 126 women diagnosed with
ovarian metastases, cervical or ovarian cancer. Also not included were 24 women,
whose histological samples appeared absent or could not reliably be ascertained.
Women with atypical endometrial hyperplasia were excluded as a limited number
of four cases were found. A subdivision was made between 112 women with
endometrioid endometrial cancer, 47 women with benign disorders and a mixed
group of ten women with non-estrogen related endometrial cancer, like clearcell
carcinoma, serous papillary cancer, mixed mullerian tumour or uterine sarcoma.
The morphological characteristics of the ovaries of the included patients were
classified according to the degree of stromal hyperplasia using the parameters
originally described by Boss et al: an ovary with a cortex of less than 1 millimeter
(mm) was determined to be atrophic (class I), a cortex wider than 1 mm was
defined as slight stromal hyperplasia ( class II), an ovarian cortex wider than 1 mm
with apparent cellular cortical stroma present in the medulla was called moderate
or marked stromal hyperplasia ( class III).8-1 0
,1 3 The ovaries were judged without
being familiair with the patients' diagnosis. A representative part of the samples
(3 0%) was judged in almost full agreement by three investigators (VJ, HH, JS) to
rule out inter-observer variability. In the few cases of discordant histological
diagnoses full consensus was obtained after reassessment. The histological ovarian
tissue samples of the included patients were scored according to the degree of
ovarian stromal hyperplasia and related to the presence or absence of endometrial
cancer.
70
Table 1. Clinical data of 112 women with endometrioid endometrial cancer, 10
women with non-endometrioid endometrial cancer and 47 women with a benign
gynaecological disorder.
Age (years)
Year after menopause
Body weight (kg)
Hip-waist ratio (n=57)
Endometrioid
endometrial cancer (n= l 12)
66.2 ( 10.2)
14.7 ( 10.9)
8 1 .6 ( 19 .3)
1 .08 (0. 14)
Benign disorder Non-endometrioid
endometrial cancer (n=47) (n= l O)
63 .8 ( 10.3) 64.5 (9 .9)
12.7 ( 1 1 .6) 13.3 ( 12.7)
74.8 ( 12.7) 77.0 ( 12.6)
1.09 (0. 14)
Data represent mean values, standard deviation is shown in parentheses.
No significant differences were noted between subgoups.
Prospective study
In addition, a prospective study was performed between 1998 and 2001. Patients
suspected of endometrial cancer or a benign condition, who were scheduled to
undergo an abdominal hysterectomy with bilateral salpingo-oophorectomy in the
University Hospital of Groningen or the Medical Centre of Leeuwarden, were
invited to participate. Twenty of 94 eligible women were not included in the study
due to logistical problems. Fourteen women refused to participate in the study or
were not mentally fit enough to consider participation. All remaining 60 women
were included and fulfilled the same eligibility criteria as the women in the
retrospective part of the study. They were also part of the 169 women in the
retrospective study. None of the patients had received hormone replacement
therapy during the preceding 6 months and all had been postmenopausal for at
7 1
least one year. Approval of the Hospital Ethics committee was obtained. Tissue
ovarian samples obtained during surgery were fixed in formalin and repmentative
transverse sections were processed by routine methods including staining by
haematoxylin and eosin. Morphological characteristics of the ovaries of the
included patients were classified as described in the retrospective study. When
informed consent was received, peripheral blood samples and intra-operative
samples of blood from the utero-ovarian veins were collected as described
previously.12 When blood samples were collected, the samples were centrifuged
after clotting. The acquired serum was stored at -20° C until processed. Using non
commercial radio-immunoassays (RIA) serum levels of estrone, estradiol,
androstenedione and testosterone were measured. 1 4- 1 6 For the estron assay intra
and interassay coefficients of variation were less than 1 1 % and 15%, and for the
estradiol assay less than 5% and 14%, respectively. Intra- and interassay
coefficients of variation for the androstenedione assay were less than 7 .3% and
21 %, and for the testosterone assay less than 7 . 1 % and 19%. As sampling was
performed from left and right utero-ovarian veins, the average hormone levels
were calculated taking both sides together. Left and right vein levels are typically
similar.12
Hip-waist ratio and body weight were measured in 57 women. The hip-waist ratio
was defined as the hip circumference at the point of maximum protuberance of the
buttocks divided by the cirumference at the umbilical level, measured with a steel
tape while the subject was standing erect.
72
Statistics
For comparison of results in both studies data analysis was performed using the
SPSS 10.0 software package (SPSS Inc., Chicago, IL). Data are presented with
mean values and standard deviation or median and inter-quartile ranges, whenever
appropriate. Statistical comparison was made between the degree of ovarian
hyperplasia, the presence or absence of malignancy, and serum and ovarian vein
levels of estrogens and androgens. Differences in scores for hyperplasia were
evaluated by the Test of Yates and Cochran. Atrophy, slight and marked
hyperplasia of the ovaries were scored as 1,2 and 3. The Mann Whitney Test and
Kruskal Wallis Test were used for comparison of continuous variables. These non
parametric tests were chosen, as the values were not expected to be normally
distributed. Chi-square analysis was used for comparison of frequencies. Linear
multiple regression analysis was performed for calculating statistical correlation
between the presence of endometrial cancer and ovarian vein steroid levels. P
values lower than 0.05 were considered to reflect statistical significance.
Results
Clinical data of patients involved in the study are shown in Table 1 and 2. The
whole retrospective study group consisted of 169 women; 1 12 women had
endometrioid endometrial cancer, 4 7 women had a benign gynaecological
condition (Table 2) and ten women had non-endometrioid endometrial cancer.
The three different patient categories were comparable with respect to age, age
after menopause and body weight. In the prospective study blood sampling was
performed in 60 postmenopausal women, of whom 52 had endometrioid
endometrial cancer and eight had a benign condition.
73
Table 2. Clinical or pathological diagnosis of 4 7 women with benign
gynaecological conditions.
Clinical or pathological diagnosis
Uterine leiomyoma Adenomyosis Reactive changes Abdominal hysterectomywith normal findings Genital prolapse with normal findings Benign glandulocystic endometrial polyp G landulocysticendometrial hyperplasia Complex hyperplasia(no atypia)
Number ofwomen (n=47)
12 1 3
12 4
5 5 5
Table 3. Degree of ovarian stromal hyperplasia in relation to endometrioid
endometrial cancer (n= 1 12), benign gynaecological disorders (n=47) and
nonendometrioidendometrial cancer (n= l 0).
Stromal atrophy Slight stromal Moderate/ marked Mean Hyperplasia Hyperplasia Score
(score= l ) (score=2) (score=3) Endometrioid endometrial 10 (8.9 %) 5 1 (45 .5%) 5 1 (45 .5%) 2.4 carcinoma(n= l 12Y
Benign disorder 14 (29.8%) 25 (53.2%) 8 ( 17 .0%) 1.9 (n=47)1
Non-endometrioid endometrial 4 (40.0%) 5 (50%) 1 ( 10.0%) carcinoma ( n= 10)
95% C.I 0.24-0.74; p=0.000
74
Ovarian stromal hyperplasia and steroid levels
In the retrospective study analysis of the association between endometrioid
endometrial cancer and degree of ovarian stromal hyperplasia is demonstrated in
Table 3. In 47 women with benign conditions a 17.0% prevalence of marked
ovarian stromal hyperplasia was observed, as compared to an incidence of 45 .5%
in 112 women diagnosed with endometrioid endometrial cancer (p=0.000). In the
group of ten women with non-endometrioid endometrial cancer only one case of
marked ovarian stromal hyperplasia was diagnosed.
In the prospective part of the research, an increasing degree of ovarian stromal
hyperplasia was related to higher ovarian levels of both testosterone and
androstenedione (Table 4; p<0.05 and p<0.005 respectively). When benign
disorders were excluded and mean ovarian vein levels were related to the degree
of ovarian stromal hyperplasia, no other relations were found. Median ovarian vein
levels of both testosterone and androstenedione were higher in patients with
endometrioid endometrial cancer as compared with patients with a benign
condition, without reaching statistical significance (Table 5). When comparing
women with endometrial cancer and atrophic stromal ovarian hyperplasia to
women with endometrial cancer and marked stromal ovarian hyperplasia, higher
ovarian levels of both testostosterone and androstenedione were found in the
second group (p<0.05), suggesting that the levels of androgens were related more
to the degree of ovarian stromal hyperplasia than to the presence of endometrial
cancer. By performing backward linear multiple regression analysis we confirmed
that the presence or absence of endometrial cancer did not influence serum values
of estrogens and androgens. The ovarian vein levels of estrone and estradiol were
not correlated to the degree of ovarian stromal hyperplasia. No statistical
difference was found between hormone levels of peripheral blood samples and
degree of ovarian stromal hyperplasia and the presence or absence of endometrial
cancer.
75
Table 4. Median ovarian vein and peripheral serum levels ( + inter-quartile
ranges) of Testosterone (T), Androstenedione (A), Estrone (El ) and Estradiol
(E2) in 60 postmenopausal women and the relationship with the degree of
ovarian stromal hyperplasia.
Ovarian stromal
atroEhY (n=5) T ovarian vein
(nmol/L) 3 . 1 (2. 1 :4.9) 1
T peripheral
(nmol/L) 1 .7 ( 1 .6:2.2)
A ovarian vein
(nmol/L) 3.7 (2.9 :4 . 1)2
A peripheral
(nmol/L) 3.0 (2.5:4.6)
E2 ovarian vein
(nmol/L) 0. 1 (0. 1 :0. 1)
E2 peripheral
(nmol/L) 0. 1 (0. 1 :0.2)
E 1 ovarian vein
(nmol/L) 0. 1 (0.0:0. 1)
E 1 peripheral
(nmol/L) 0.0 (0.0:0.0)
I p < 0.05
2 p < 0.005
Slight stromal
hype�lasia (n=30)
3.8 (2.9 :7.3) 1
2.0 ( 1 .8 :2.4)
6.3 (4 .8:9.3)2
3 .5 (2.5 :4 .7)
0. 1 (0. 1 :0.3)
0. 1 (0.0:0.2)
0. 1 (0.0:0.2)
0.0 (0.0:0. 1)
Marked stromal
Hype!:£lasia (n=25)
7.0 (3 .3 : 18.4) 1
2. 1 ( 1 .5 :2.6)
14 .0 (8.9 :29.3)2
3 .6 (2.4 :5 .3)
0. 1 (0.0:0. 1)
0. 1 (0.0:0.2)
0. 1 (0.0:0.2)
0. 1 (0.0:0. 1)
76
Table 5. Median ovarian vein and peripheral serum levels (+ inter-quartile
ranges) of Testosterone (T), Androstenedione (A), Estrone (E 1) and Estradiol
(E2) in 60 women with endometrioid endometrial cancer or a benign
gynaecological condition.
Endometrial Benign Cancer disorders p-value (n=52) (n=8)
T ovarian vein 5.0 (2.8: 10.0) 4.7 (3 .5 : 16.7) 0.5 (nmol/L)
T peripheral 2. 1 ( 1 .6:2.6) 1 .9 (1 .8:2. 1) 0.4 (nmol/L)
A ovarian vein 9 .0 (4.8 : 15 .4) 8. 1 (4.4 :8.5) 0.3 (nmol/L)
A peripheral 3 .4 (2.5 :4 .7) 3 .6 (2.5 :5. 1) 1.0 (nmol/L) E2 ovarian vein 0. 1 (0.0:0.2) 0. 1 (0. 1 :5.8) 0.4 (nmol/L)
E2 peripheral 0. 1 (0.0:0.2) 0. 1 (0.0:0.5) 0.4 (nmol/L)
E 1 ovarian vein 0. 1 (0.0:0.2) 0. 1 (0.0:0.2) 0.4 (nmol/L)
E l peripheral 0. 1 (0.0:0.2) 0.0 (0.0:0. 1) 0.4 (nmol/L)
N.S.
One woman with endometrioid adenocarcinoma was later excluded from the
prospective study as she was also diagnosed with a well differentiated
endometrioid adenocarcinoma in the left ovary, possibly due to metastatic disease,
whereas a second primary tumour could not be excluded. In her case a high
ovarian vein androstenedione level (average 85.1 nmol/L; left side 90.9 and right
side 79.3 nmol/L) and elevated testosterone level (average 18.4 nmol/L; left side
21.5 and right side 15.2 nmol/L) were found, possibly due to high androgen
77
production from local stromal reaction. Remarkably androgens from the ovary
without malignancy were elevated as well. In three women with endometrioid
endometrial cancer and marked stromal hyperplasia an extremely high mean
ovarian vein androstenedione level of 1 15 nmol/L and a mean testosterone level
of 1 10 nmol/L were found, but pathological reexamination showed no sign of
micrometastasis in these ovaries.
Bodyweight
Degree of ovarian stromal hyperplasia was related to body weight (p<0.005)
(Figure 1). In 57 women hip-waist ratio had been measured prospectively, but no
correlation was found between degree of hyperplasia, estrogen or androgen levels
and upper or lower body obesity. Comparison between degree of stromal
hyperplasia and age or age after menopause or histological grade of endometrioid
endometrial carcinoma did not appear statistically significant.
u � If) 0)
1 00
90
80
:C 70
>, u 0
.0 C (I] Q)
� 60
Fig 1 . Relation between body weight, ovarian
stromal hyperplasia and endometrial cancer
I I
atrophy slight
Degree of ovarian stromal hyperplasia
I 0 I I Benign disorder
□
I Endometrlal cancer
□ I
marked
78
Discussion
After menopause, simultaneously with the decreasing number of follicles the
ovarian cortex becomes thin and gyrated, creating an irregular surface.1 7•
1 8 Ovaries
of postmenopausal women are composed mainly of stromal cells. Although the
ovarian volume decreases after menopause, the postmenopausal ovary is generally
regarded as a gland producing significant amounts of androgens.3• 1 0• 1 1 •1 9•20 A
postmenopausal ovary may show cortical and stromal atrophy, but also stromal
hyperplasia can be seen. Stromal hyperplasia is characterized by nodular or diffuse
proliferation of ovarian stroma. Varying degrees may be present. Until the criteria
developed by Boss were introduced, some regarded the amount of stromal
collagen or the presence of cortical stroma as features of stromal hyperplasia,
whilst others made that classification only in case nodules of cortical stroma of
one mm. in diameter or more were present. Boss himself found no consistent
correlation between degree of stromal proliferation and ovarian size.1 3 Marked
stromal hyperplasia appeared sometimes evident in a normal sized ovary, as the
amount of stroma is only one factor determining the size of the ovary, concluding
that the volume perse is not a reliable indicator of the content of the ovary. Several
studies showed that the extent of ovarian stromal hyperplasia is associated to some
degree with the degree of postmenopausal ovarian steroid production.8-10 Any
evaluation of the hormonal patterns in postmenopausal women with regard to the
development of hormonal dependent endometrial cancer should also take into
account the stromal characteristics of the ovary, as ovarian morphology may
reflect to the endocrine status and endometrial disease.1 1
We compared the degree of stromal hyperplasia in relation to the presence or
absence of endometrial cancer and report a higher frequency of ovarian stromal
hyperplasia in case of endometrioid endometrial cancer. Our study supports the
hypothesis that a relationship exists between endometrioid endometrial cancer and
79
ovarian stromal proliferation. Contrary to our results, Lucisano et al found no
correlation between the degree of stromal hyperplasia and the presence of
endometrial cancer, but this lack of relation might be explained by the relative
small number of patients in their study.9 In case of endometrial carcinoma we
found a 45.5% incidence of marked stromal hyperplasia in ovaries, which is
higher than Lucisano, who reported an incidence of 19% in 21 women with
adenocarcinoma of the endometrium. Other studies report incidences of 1 1 %,
28% and 4 7% in patients with and without endometrial cancer mixed together, but
comparison is hampered by widely varying percentages of endometrial malig
nancy in these studies.8• 1 3•2 1 Decades ago, without using the criteria developed by
Boss in 1965, marked ovarian hyperplasia was found in 84% of endometrial
malignancies, compared to only 44% in autopsy controls.1 3•2 1 Boss himself
mentions marked ovarian stromal hyperplasia in 8 1 % of 16 patients with
endometrial cancer.1 3 Except the study by Lucisano, all of the studies mentioned
failed to record the exact histological type of carcinoma in case of endometrial
pathology? The present study is the first one analysing only women with
endometrioid endometrium carcinoma separately. Endometrioid cancer is
supposed to be related to estrogenic stimulation and accounts for 75% of all cases
of endometrial cancer. Taking into account the assumption that increased androgen
production from ovarian hyperplasia may be related to endometrial pathology,
only hormone-dependent endometrioid endometrial malignancies should be
included in comparative studies. This conclusion is supported by our findings that
androgen levels from the ovarian veins were higher with an increasing degree of
ovarian stromal hyperplasia, and that the degree of ovarian stromal hyperplasia
was higher in the presence of endometrioid endometrial cancer. Analysis of
women with non-endometrioid endometrial cancer showed only 10% marked
stromal hyperplasia, suggesting that no increase of ovarian stromal hyperplasia is
found in hormone-independent uterine cancer. However, it should be noted that
this subgroup is rather small to draw firm conclusions.
80
Endometrioid endometrial cancer arises more often in women with obesity. In our
study mean body weight was higher in case of malignancy. Taking into account
the reported correlation between ovarian stromal hyperplasia and endometrioid
endometrial cancer, it is not surprising that in the present study a correlation was
found between ovarian stromal hyperplasia and body weight, although a causal
relationship cannot be presumed.
Whether increased androgen production by the hyperplastic ovary is a causative
mechanism leading to endometrial cancer remains to be ascertained. In 1 8
postmenopausal women Sluijmer et al demonstrated a correlation between the
degree of ovarian stromal hyperplasia and the production of androgens, both
androstenedione and testosterone.8 Other studies showed that ovaries with marked
stromal hypeiplasia secreted significant amounts of androgens, and that the degree
of hyperplasia correlated to some degree with the amount of androgen secreted.9'10
Androgen levels in the utero-ovarian vein blood samples from women with endo
metrial cancer were significantly elevated as compared with the utero-ovarian vein
levels from women without endometrial carcinoma.22-24 Like androgens from the
adrenal glands also androgens produced by the ovaries can be converted to
estrogens by aromatase cytochrome P450. Increased production of aromatizable
androgens by the ovarian stroma may lead to increased precursor availability for
estrogen production. It is an intriguing question whether, apart from extraglandular
aromatase activity in adipose tissue, aromatase activity in the endometrium plays a
role in malignant degeneration.6'7 Aromatase (over-) expression in endometrium is
hypothesized to be a possible mechanism linking ovarian stromal hyperplasia and
increased steroid production to the development of uterine cancer. In endometrial
cancer indeed aromatase activity has been reported, whereas disease-free
endometrium did not express aromatase.25-29 These findings may support the
alleged correlation between steroid secretion from the ovarian stromal hyperplasia
8 1
and endometrial disease.30 In their excellent work from the forties of the last
century Smith, Woll and Hertig were the first to focuss on this relationship.2 1 •3 1
Many decades later it still remains questionable whether ovarian stromal
hyperplasia simply reveals a local stromal reaction in the presence of endometrial
cancer or reflects an ovarian predisposition favouring future malignant
degeneration due to its endocrine properties. Therefore future research should aim
at the relationship of ovarian stromal hyperplasia, ovarian steroid production and
aromatase activity in the endometrium. Under the influence of the increased
steroid levels from the hyperplastic ovary clonal cellular clusters may arise in the
endometrium with progressive growth resulting in an increasing risk of DNA
replication errors, especially in case of specific genetic susceptibility, such as e.g.
CYP 19 polymorphisms, that have been associated with increased steroid
concentrations and might be related to a.o. aromatase activity.32 Eventually
endometrial cancer may develop, with ovarian stromal hyperplasia as another
concomitant hormonal risk factor promoting endometrial carcinogenesis.
Acknowledgements
The authors are grateful to Mrs. KA Ten Hoar and J Koerts for laboratory
assistance and to Dr. WJ Sluiter for his contribution in statistical analysis.
82
References
1. Emons G, Fleckenstein G, Hinney B, Huschmand A, Heyl W. Hormonal interactions in
endometrialcancer. EndocrRel Cancer2000; 7: 227-42. 2. Akhmedkhanov A, Zeleniuch-Jacquotte A, Toniolo P. Role of exogenous and
endogenous hormones in endometrialcancer. Ann NY Acad Sci 2001; 943 ; 296-3 15.
3. Adashi EY. The climacteric ovary as a functional gonadotropin-driven androgen
producing gland. Fertil Steril 1994; 62: 20-7 . 4 . Sasano H, Harada N. Intratumoral aromatase in human breast, endometrial and ovarian
malignancies.EndocrineRev 1998; 19 : 593-607. 5. WatanabeK, Sasano H, Harada N, et al. Aromatase in human endometrial carcinoma and
hyperplasia. Immunochemical, in situ hybridization, and biochemical studies. Am J
Pathol 1995; 146: 49 1-500. 6. Bulun SE, Noble LS, Takayama K, Michael MD, Agarwal V, Fisher C, et al. Endocrine
disorders associated with inappropriately high aromatase expression. J Steroid Biochem
Molec Biol 1997; 61 : 133-9.
7. Bulun SE, Economos K, Miller D, Simpson ER. CYP19 (aromatase cytochrome P450)
gene expression in human malignant endometrial tumors. J Clin Endocrinol Metab 1994;
79 : 183 1-4.
8. Sluijmer A V,Heineman MJ, Koudstaal J, et al. Relationship between ovarian production of estrone, estradiol, testosteron, and androstenedione and the ovarian degree of stromal
hyperplasiain postmenopausalwomen. Menopause 1998; 5: 207- 10. 9. Lucisano A, Russo N, Acampora MG, et al. Ovarian and peripheral androgen and
estrogen levels in post-menopausal women: correlations with ovarian histology.
Maturitas 1986; 8: 57-65.
10. Ala-Fossi SL, Maenpaa J, Aine R, et al. Ovarian testosterone secretion during
perimenopause.Maturitas 1998; 29 : 239-45 .
1 1 . Longcope C. Editorial. The relationship of ovarian production of hormones and the
degeree of ovarian stromal hyperplasia in postmenopausal women. Menopause 1998; 5 : 205-6.
12. Heineman MJ, Sluijmer AV, Evers JLH. Utero-ovarian vein blood sampling in
postmenopausalwomen. Fertil Steril 1993; 60: 184-6.
13 . Boss Jh, Scully RE, Wagner KH, et al. Structural variations in the adult ovary: clinical
significance. Obstet Gynecol 1965; 25 : 747-64.
83
14. Jurjens H, Pratt JJ, Woldring MG. Radioimmunoassay of plasma estradiol without extraction and chromatography.J Clin EndocrinolMetab 1975; 40: 19-25 .
15 . Pratt JJ, Wiegman T, Lappohn RE, Woldring MG. Estimation of plasma testostosteron without extraction and chromatography.Clin Chim Acta 1978; 84: 329-37 .
16. Pratt JJ. Steroid immunoassay in clinical chemistry. Thesis Groningen 198 1 . 17. Flaws JA, Rhodes JC, Langen berg P, et al. Ovarian volume and menopausal status.
Menopause2000; 7: 53-61 . 18 . Goswamy RK, Campbell S, Royston JP, et al. Ovarian size in postmenopausal women.
Br J Obstet Gynaecol 1988; 95 : 795-801 . 19. Sluijmer AV.The postmenopausal ovary. Functional and morphological aspects. Thesis
Maastricht 1999. ISBN 90-9012802-6. 20. Dennefors BL, Janson PO, Knutson F, et al. Steroid production and responsiveness to
gonadotropin in isolated stromal tissue of human post-menopausal ovaries. Am J Obstet Gynecol 1980; 136: 997- 1002.
21 . Woll E, Hertig AT, Smith GVS, et al. The ovary in endometrial carcinoma: with notes on the morphological history of the aging ovary. Am J Obstet Gynecol 1948; 56: 617-33.
22. Mertens HJMM, Heineman MJ, de Jong FH, et al . Postmenopausal ovarian function and endometrial cancer. Hum Reprod 14, Abstractbook, Uune 1999): 89-90.
23 . Heineman MJ, Sluijmer AV, Kampscoer PHNM, et al. Utero-ovarian vein blood smapling in postmenopausal women with and without endometrial cancer. Fertil Steril
1995; Suppl: S92 (Abstract O- 187).
24. Nagamani M, Stuart CA, Doherty MG. Increased steroid production by the ovarian stromal tissue of postmenopausal women with endometrial cancer. J Clin Endocrinol Metab 1992; 74: 172-6.
25. Bulun SE, MahendrooMS, Simpson ER. Polymerasechain reaction amplificationfails to
detect aromatase cytochrome P450 transcripts in normal human endometrium or decidua.
J Clin EndocrinolMetab 1993 ; 76: 1458-63 . 26. Tseng L, Mazella J, Mann WJ, et al. Estrogen synthesis in normal and malignant humen
endometrium. J Clin EndocrinolMetab 1982; 55 : 1029-3 1. 27. Yamaki J, Yamamoto T, Okada H. Aromatization of androstenedione by normal and
neoplasticendometriumofthe uterus. J Steroid Biochem 1985 ; 22: 63-6. 28. Yamamoto T, Kitawaki J, Urabe M, et al. Estrogen productivity of endometrium and
endometrial cancer cells. J Steroid BiochemMol Biol 1993 ; 44: 463-8 .
84
29. Sasano H, Kaga K, Sato S, et al. Aromatase P450 gene expression in endometrial
carcinoma. Br J Cancer 1996; 74 : 154 1-4 . 30. Longcope C. Editorial. The relationship of ovarian production of hormones and degree of
ovarian stromal hyperplasia in postmenopausalwomen. Menopause 1998; 5: 205-6. 3 1. Smith GVS. Carcinoma of the endometrium: a review with results of treatment trough
1935 . N Engl J Med 194 1 ; 225 : 608- 15 . 32. Berstein LM, Imyanitov EN, Suspitsin EN, Grigoriev MY, Sokolov EP, Togo A, et al.
CYP 19 gene polymorphism in endometrial cancer patients. J Cancer Res Clin Oncol
2001; 127: 135-8.
85
86
5
Is aromatase cytochrome P450 involved in the
pathogenesis of endometrioid endometrial cancer?
VHWM Jongen
JHH Thij ssen
H Hollema
GH Donker
JG Santema
AGJ van der Zee
MJ Heineman
International Journal of Gynecological Cancer 2005; 15:529-36
87
Abstract
Prospectively the relationship between androgen levels in the utero-ovarian
circulation, aromatase activity in endometrial and body fat tissue, and the presence
or absence of endometrioid endometrial cancer was studied in postmenopausal
women. In 43 women with endometrioid endometrial cancer and 8 women with a
benign gynaecological condition a hysterectomy with bilateral salpingo
oophorectomy was performed. Using tritium water release assays, aromatase
activities in endometrial and body fat tissue were determined and related to the
steroid levels from the peripheral and the utero-ovarian venous circulation
( estradiol, androstenedione, testosterone) and to the presence or absence of
endometrial cancer. Significant aromatase activity was found in both benign and
malignant endometrial tissue samples. Aromatase activity in samples of
endometrial tissue and in samples of body fat did not correlate with steroid levels
in peripheral or utero-ovarian venous blood. Aromatase activity in samples of
benign or malignant endometrium did not differ. Remarkably, in four women with
mainly poorly differentiated endometrial cancer very high aromatase activity was
found in endometrial tissue. It is likely that multiple pathogenetic pathways exist
which eventually lead to the formation of endometrioid endometrial cancer. The
local availability of androgens and the finding that aromatase activity is present in
both endometrial cancer and benign endometrial tissue, support the hypothesis that
aromatase activity in the endometrium may play a role in malignant
transformation, by converting androgens into mitogenic estrogens in the
endometrial tissue.
88
Introduction
The enzyme complex aromatase consists of the cytochrome P450 aromatase
enzyme, its associated haeme group, and a NADPH-reductase flavoprotein. 1
Aromatase is responsible for the binding of the C 19 steroid substrate
( androstenedione and testosterone) and catalyzes the series of reactions which
will eventually lead to estrogen production. The conversion of androgens to
estrogens involves three subsequent steps. The third step leads to the
aromatization of the A-ring, giving the characteristic phenolic ring of estrogens.2
Aromatization inside the ovaries leads to the synthesis of estrogens during the
menstrual cycle, after menopause the so-called peripheral conversion is
responsible for the production of estrogens. The overall peripheral conversion of
androgens to estrogens increases as a function of age, obesity and aromatase
activity in extraglandular (adipose) tissue. 3 The contribution of aromatase to
estrogen formation must be considered extremely important, as aromatization is
the essential and rate-limiting step. The principle substrates of estrogens are
circulating androgens, derived primarily from the adrenal cortex and to a lesser
extent from the ovaries. Estrogens are considered to be the main promoting
factor of endometrioid endometrial carcinoma. It can not be excluded that
androgens may play a role as well. After menopause the ovary still produces
40% respectively 20% of the total amount of testosterone and
androstenedione.4-7 Androgens, produced by the ovaries, are not only converted
to estrogens in peripheral adipose tissues, but also give rise to elevated
prehormone availability for estrogen formation in utero. In several studies it has
been demonstrated that the androgen levels in the utero-ovarian veins from
women with endometrial cancer were significantly elevated as compared to
women without malignant disease. 7- 1 0 In a recent study these findings were not
confirmed. 1 1 Ovarian androgens may enhance the development of endometrial
cancer through local aromatization into estrogens. Local aromatase activity may
89
even give rise to very high local estrogen concentrations which could result in
the stimulation of endometrial growth. Thus, in addition to extraglandular
aromatase activity by converting androgens into estrogens in adipose tissue,
aromatase activity in the endometrium may play a role in malignant
transformation. 12 Aim of the present study is to explore the possible relationship
between androgen levels in the utero-ovarian circulation, aromatase activity in
the endometrial tissue and the presence of endometrioid endometrial cancer.
Patients and methods
A prospective study was performed between 1998 and 2001. Patients suspected of
endometrial cancer or a benign condition, who were scheduled to undergo an
abdominal hysterectomy with bilateral salpingo-oophorectomy in the University
Hospital of Groningen or the Medical Centre of Leeuwarden, were invited to
participate, and were included after informed consent. Women with ovarian
metastases, cervical or ovarian cancer were not included. None of the patients had
received hormone replacanent therapy during the preceding 6 months and all had
been postmenopausal for at least one year. Approval of the Hospital Ethics
committee was obtained. Patients were part of a study described in a previous
paper.1 1
During surgery samples of blood from the utero-ovarian and peripheral veins were
collected, as described previously.13 When blood samples were collected, the
samples were centrifuged after clotting. The acquired serum was stored at -20° C
until processed in order to determine steroid levels. During or directly after the
procedure tissue specimens were obtained from both the subcutaneous abdominal
fat, and the endo- and myometrium. In case of endometrial cancer a
representative piece of endometrial tissue was chosen, directly next to or on the
border of the suspected endometrial cancer tissue. The acquired specimens
90
( approximately one cm3) were frozen at -20° C as well, until aromatase assays
could be carried out.
Steroid measurements
Using non-commercial radio-immunoassays (RIA) serum levels of estrone,
estradiol, androstenedione and testosterone were measured. 14-1 6 Intra- and
interassay coefficients of variation were less than 11 % and 15% for the estrone
assay and for the estradiol assay less than 5% and 14%, respectively. Intra- and
interassay coefficients of variation for the androstenedione assay were less than
7 . 3 % and 21 %, and for the testosterone assay less than 7. 1 % and 19%. As
sampling was performed from both ovarian sides, the average hormone levels
were calculated taking left and right utero-ovarian vein together, as they are
typically similar.13
Aromatase assays
[ l �-3H]-androst-4-ene-3, l 7-dione (spec.act. 25 Ci/mmol) was provided by NEN
(Boston MA, USA); Exemestane by Pharmacia & Upjohn (Milan, Italy); NAD
by Sigma (St.Louis, Missouri ,USA); NADP and glucose-6- phosphate by Roche
(Mannheim,Germany); ATP and glucose-6- phosphate dehydrogenase by ICN
(Ohio,USA); charcoal by Merck (Darmstadt, Germany); Dextran T by
Amersham Pharmacia (Uppsala, Sweden); diethylether by Rathbum
(Walkerbum, Scotland); scintillation fluid Ultima Gold by Packard (Meriden
CT, USA); Coomassie brilliant blue reagent by Biorad (Hercules CA, USA) and
human serum albumin by Kabi Vitrum (Stockholm,Sweden). Buffer A: 0.01 M
phosphate buffer pH 7.4 containing Trasylol 500 KIE/ml (Bayer, Leverkusen,
Germany). Buffer B: buffer A + co-factors: NAD (10 mM), glucose-6-
phosphate dehydrogenase (2 U/ml) and 2 mM each of ATP,NADP and glucose-
6- phosphate. Charcoal: 2.5% Dextran T70, 5% charcoal in aqua dest. (w/v).
91
Aromatase activity assay
Measurement of aromatase enzyme activity was based on the tritium water
release assay as described previously. 1 7 In brief: tissue samples (0.4-2 gram)
were minced on ice and pulverized at - 1 96° C using a Micro-Dismembrator
(Braun, Melsungen, Germany) and stored at -80° C. Just prior to assay the
powder was suspended in 2-3 ml ice-cold buffer A, centrifuged at 4° C at 800 g
for 5 min, the fatty layer was removed and 1 ml of the supernatant was used.
Substrate 1 µCi [ 1 �-3H]-androstenedione ( 15 nM) was added in 1 ml of buffer
B. In order to obtain blanks for each of the individual samples identical parallel
incubations were done to which the aromatase inhibitor exemestane (final
concentration 10-5 M) had been added. Aspecific influences have been
minimised because all individual samples have been corrected for by the use of
an identical sample to which exemestane had been added. Incubations were done
1 6 hours at 3 7 °C with continous shaking and finished by placing the tubes on
ice. To remove tritiated steroids, supernatants were extracted 6 times with 4 ml
diethylether; subsequently 0.5 ml suspension of dextran coated charcoal was
added and incubated for 20 minutes on ice. Samples were centrifugated at 4800g
at 4°C for 30 minutes. Two ml of the supernatant was added to 1 0 ml Ultima
Gold scintillation fluid and assayed for 3H radioactivity in a W allac model 14 14
liquid scintillation counter.
Protein assay
Protein content of the supernatants was measured according to the method of
Bradford, using a two wavelenght OD-ratio modification to improve linearity.5
µl sample was added to 200 µl Coomassie brilliant blue reagent. The OD was
measured at 595 nm and 450 nm in a Biorad microplate reader. Human serum
albumin was used as a standard at concentrations ranging from 0-700 µg/ml.
The sensitivity of the aromatase assay is approximately 0. 3 fmol/mg. Results of
aromatase levels below 0. 3 fmol/mg were considered to be 0.3 fmol/mg.
92
Statistics
Data analysis was performed using SPSS I 0.0 software package (SPSS Inc.,
Chicago, IL). Data are presented with mean values and standard deviation or
median and inter-quartile ranges, whenever appropriate. Statistical comparison
was made between aromatase concentrations in endometrial tissue and body fat,
the presence or absence of malignancy, and ovarian vein levels of androgens and
estrogens. The Mann-Whitney-U Test was used for comparison of continuous
variables. Using Spearman rank-correlation, aromatase levels were compared to
utero-ovarian steroids. P values lower than 0.05 were considered to reflect
statistical significance.
Results
Inclusion
In the study 60 postmenopausal women were included, of whom 52 had
endometrioid endometrial cancer and eight had a benign condition. Atypical
hyperplasia of the endometrium (N=2) was regarded as a benign condition.
Clinical data of patients involved in the study are shown in Table I . Of 43 women
with endometrioid endometrial cancer 21 women had a grade I malignancy, 13
women a grade II maligancy, and 9 women were reported with a grade III
malignancy. Five women were reported with stage la endometrial cancer, 16
women with a stage l b, and 16 women with a stage l e malignancy.Stage 2a and
stage 2b were found in 2 respectively 3 women. Finally, one woman had stage 3
endometrial cancer. No correlation could be detected between histopathological
stage, grade and aromatase activity or steroid hormone levels. No patients had
diseases of liver, kidney or of the endocrine system, apart from one woman with a
(grade I) stage I a malignancy. She had suffered from hyperadrenocorticism, but
all steroid hormone levels were below median levels. Nine patients with
93
endometrial cancer were excluded later, because not enough endometrial or fat
tissue was available for determination of aromatase levels, so aromatase levels
could be compared in 43 patients with endometrial cancer. In Table 2 we describe
clinical or pathological diagnoses or reasons for hysterectomy with bilateral
salpingo-oophorectomy in 8 women with a benign condition. Unfortunately,
only 8 women with a benign condition could be included in the study. The size
of the control group was limited by the fact that an abdominal hysterectomy
with bilateral salpingo-oophorectomy is rarely performed in case no malignancy
is present.
Table 1 . Clinical data of 51 postmenopausal women undergoing an abdominal
hysterectomy with bilateral salpingo-oophorectomy.
Endometrioid Benign Endometrial cancer Disorder
(n=43) (n=8)
Age (years) 63 .9 (9.8) 62. 1 (8.6) Years after menopause 12.5 ( 10.9) 1 1 .9 (7.6) Body weight 78. 1 (14.8) 69. l ( 1 3 .6) Grade I Endometrioid endometrial cancer 21 Grade II Endometrioid endometrial cancer 1 3 Grade II Endometrioid endometrial cancer 9 Stage 1 malignancy 37 Stage 2 malignancy 5 Stage 3 malignancy 1
Data represent mean values, standard deviation is shown in parentheses
94
Table 2. Clinical or pathological diagnosis/ reason for hysterectomy with
bilateral salpingo-oophorectomy in 8 women with a benign condition.
Diagnosis or reason for hysterectomy n Recurrent postmenopausal bleeding leading to diagnostic
hysterectomy 3
Benign endometrial polyp 2
Atypical hyperplasia 1
Adenomyosis and atypical hyperplasia 1
Moderate cervical dysplasia after menopause 1
Table 3 . Median utero-ovarian vein and peripheral serum levels (+ inter
quartile ranges) of Testosterone (T), Androstenedione (A), Estrone (E 1) and
Estradiol (E2) in 60 women with endometrioid endometrial cancer or a benign
gynaecological condition.
Benign Endometrioid p-value
disorders endometrial
(n=8) cancer (n=52) T ovarian vein 4.7 (3.5: 16.7) 5.0 (2.8 : 10.0) 0.5 T peripheral 1 .9 ( 1 .8 :2. 1) 2. 1 ( 1 .6:2.6) 0.4 A ovarian vein 8. 1 ( 4.4:8.5) 9.0 (4.8: 15.4) 0.3 A peripheral 3.6 (2.5:5. 1 ) 3.4 (2.5:4.7) 1 .0 E 1 ovarian vein 0. 1 (0.0:0.2) 0. 1 (0.0:0.2) 0.4 El peripheral 0. 1 (0.0:0.2) 0. 1 (0.0:0. 1 ) 0.4 E2 ovarian vein 0. 1 (0. 1 :5.8) 0 . 1 (0.0:0.2) 0.4 E2 peripheral 0. 1 (0.0:0.5) 0. 1 (0.0:0.2) 0.4 N.S.
95
Median utero-ovarian vein levels of both testosterone and androstenedione were
slightly higher in patients with endometrioid endometrial cancer as compared with
patients with a benign condition, without reaching statistical significance (Table
3 ). 1 1 When comparing activities of aromatase in endometrial and body fat tissues to
the presence or absence of endometrial cancer, no significant differences could be
noticed, although aromatase activities in benign disorders were generally higher.
Median activities of aromatase were higher in body fat tissues, both in the
presence and in the absence of endometrioid endometrial cancer (Table 4 ). Steroid
levels of androgens and estrogens from the peripherous and utero-ovarian veins
were compared to activities of aromatase in endometrial tissue and abdominal
subcutaneous fat (Table 5). No correlations could be found, although testosterone
levels in the utero-ovarian veins were related to the activity of aromatase in body
fat (p<0.05), but we do not have an explanation for that relation. In Figure 1
aromatase activities in endometrial tissue and body fat are shown in the presence
or absence of endometrioid endometrial cancer, respectively. It is shown that most
patients with endometrioid endometrial cancer had low aromatase activities in the
endometrial tissue samples; e.g. 20 times a low level of less than 0.3 fmol/mg
(limit of detection) was measured.
96
Table 4. Median aromatase activity in endometrial tissue and subcutaneous
body fat compared to the presence or absence of endometrioid endometrial
cancer.
Endometrioid Benign disorder Mann-Whitney-CT Test endometrial cancer ( p-value)
(n=43) (n=8)
Aromatase in endometrial 0.3 1 . 1 0.2 tissue (fmol/mg)
Aromatase in body fat 1 .6 2.7 1.0 (fmol/mg)
N.S.
Table 5. Comparison of aromatase activity in endometrial tissue and body fat to
utero-ovarian vein and peripheral levels of androstenedione (A), Testosterone
(T) and estradiol (E2).
Aromatase in endometrial tissue Aromatase in abdominal fat R p R p
A ovarian vein 0.019 0.895 0. 105 0.483
A peripheral -0.042 0.0773 -0. 138 0.355 T ovarian vein 0.050 0.0729 0.3 17 0.034* T peripheral 0.052 0.0721 -0.079 0.598
E2 ovarian vein 0. 106 0.503 0. 1 12 0.497 E2 peripheral 0.08 1 0.584 0. 197 0. 197
* Speannan-rank correlation; p<O. 05 Steroid hormone levels and aromatase
activity
97
Figure 1 . Aromatase activity (fmol/mg) in endometrial and body fat tissues of
patients with endometrioid endometrial cancer (n=43) and of patients with
benign disorders (n=8).
C ·a3
bO E
cE 0 (J}
C'd ... E
a
Aromatase levels endometrial and fatty tissues so ----------
.. ., •
1 0 ::-.,
.. y
'" .,, ... • .., I y ' •
� ' yly ""
t ... ... , ... 1 --- 'f' - ., • ...... - ., - 'flW • • • - ., • """" •
0.2 I I I I
I endometrium I I fatty tissue I
High aromatase activity
T malignant
• benign
Remarkably, whereas in all but four patients aromatase activities in the
endometrial cancer tissues were below 3 fmol/mg, activity of aromatase in those
four patients appeared extremely high ( 1 5.2, 17.5, 20. 1 and 2 1 .2 fmol/mg,
respectively) (Figure 1 ). No correlation could be found between aromatase
activities of the four patients with high aromatase concentrations in the
98
endometrial cancer tissue and utero-ovarian or peripheral steroid levels. None of
these four patients had a history of other malignancies like breast cancer.
Remarkably, three of these four women had poorly differentiated endometrioid
endometrial carcinoma, once with serous papillary components.
Discussion
Aromatase activity in endometrial cancer was first demonstrated by Tseng in the
early eighties. 18 In later years aromatase activity was found in endometrial
tumours, whereas its expression was not detected in disease-free endometrium.
Aromatase mRNA expression was demonstrable in several benign disease states
of endometrium, namely endometriosis and adenomyosis. 19-29 Very recently
evidence was presented showing aromatase activity in benign myometrial
disease.30 Watanabe et al found aromatase immunoactivity in 67% of stromal
cells of endometrioid endometrial carcinoma. 20 The presence of in situ
production of estrogens in the stromal cells of endometrial carcinoma was
proven by immunohistochemistry, in situ hybridisation and biochemical
methods. In stromal cells of endometrial cancer tissue, but not in neoplastic or
hyperplastic cells mRNA hybridization signals of aromatase accumulated and
the distribution of aromatase mRNA correlated well with the immunohistoche
mical localization of aromatase. It was concluded that intratumoral aromatase in
endometrial malignancy is associated with stromal invasion and is expressed
during the process of carcinoma-stromal interaction. The presence of aromatase
in stromal or interstitial cells suggests that locally produced estrogens may act
on carcinoma cells in a paracrine way. Thus, intratumoral estrogens derived
from in situ aromatization may function as a local mitogenic factor. Contrary to
our results, Watanabe found no aromatase activity in normal or hyperplastic
endometrial tissue. Like in our study, no correlation was found between
aromatase expression, activity or mRNA transcript levels and
99
clinicopathological factors such
Immunohistochemically Sasano
as clinical or histological stage. 20•27
confirmed the presence of aromatase
immunoreactivity in the stromal cells of endometrioid endometrial carcinoma.22
Contrary to other reports, we conclude that considerable activities of the enzyme
complex aromatase are present not only in endometrial cancer, but also in
normal endometrial tissue.20,22
-29 We found even higher median activities of
aromatase in endometrial tissue of benign origin than in endometrial cancer
tissue (Table 4). Among the women with endometrial cancer, we identified four
women with very high aromatase activities (Figure 1) . Remarkably, three of
these patients had poorly differentiated endometrial cancer, once with serous
papillary components. It is well known that estrogen and progesteron receptors
are less prevalent in poorly differentiated tumours. The role of androgen
receptors and androgen sensitivity of the endometrium needs to be studied more
extensively. In a recent study Mertens et al found an increase of androgen
receptor content with increasing histological grade of endometrial cancer.8 We
report a correlation between testosterone levels from the utero-ovarian veins and
the activity of aromatase in body fat (Table 5; p<0.05), but we are unable to
explain this correlation. Although we can not report an increase of aromatase
activity in the endometrial cancer tissue, our observations still fit well within
the hypothesis that multiple pathogenetic pathways may exist eventually leading
to the formation of endometrioid endometrial cancer. The local availability of
androgens and the aromatase activity present within endometrial tissue, are the
two essential conditions supporting the hypothesis that estrogens may function
as a mitogenic factor after local conversion of ( ovarian) androgens. Thus, in this
view some women may develop endometrioid endometrial cancer through
continuous hyperstimulation of the endometrial tissue by estrogens
systematically delivered from peripheral conversion, while in other women
malignant transformation is enhanced by local conversion of ovarian androgens
into estrogens, catalyzed by ( overexpression of) endometrial stromal aromatase.
1 00
Indeed, data from endometrial carcinoma cell lines suggest that some but not all
endometrial carcinomas may possess a heterogenous aromatase-dependent
growth stimulating system.2 1•3 1
•32 Possibly the normal mechanisms that inhibit
aromatase expression may be lost or aromatase-stimulating factors become
overexpressed. 27 Bokhman described two different pathogenetic types of
endometrial carcinoma. 33 The first pathogenetic type arises in women with
obesity, hyperlipidemia and signs of hyperestrogenism. In these women mainly
highly and moderately differentiated tumours are found. In women with the
second type of endometrial cancer poorly differentiated tumours are found,
including high grade malignancies like serous papillary or clear cell carcinoma.
Women with this second more virulent estrogen-independent form tend to be
less obese. Our observations partly correspond with this hypothesis. In the four
patients with high aromatase activity of the endometrial tissue mainly poorly
differentiated endometrial carcinoma was found, once with serous papillary
components. Remarkably, these patients indeed had a lower body mass index
with a mean value of 25.4 kg/m2, compared to a mean value of 28.8 kg/m2 in the
remaining 39 patients with endometrial cancer. Possibly a further subdivision
should be made within the first pathogenetic type between obese women with
signs of hyperestrogenism, and less obese women with generally poorly
differentiated endometrioid endometrial carcinoma,who appear to be aromatase
dependent.
Nowadays, the enzyme aromatase has become the most important target enzyme
in the endocrine treatment of advanced breast cancer. Generally, in about 70%
of breast tumours aromatase activity is found above the limit of detection. The
tumour-bearing quadrant of the breast was reported to have the highest
aromatase activity, suggesting that local estrogen production in the breast
increases the risk of tumour development, or that tumour tissue can stimulate
aromatase activity in the surrounding tissue. 34 It seems obvious to search for
10 1
similarities between breast and endometrial cancer in their pathogenesis and
endocrine treatment, as both are mainly hormone-dependant tumours. However,
as endometrial cancer in the early stages must be considered as an often curable
disease using surgery and radiotherapy, the potential use of aromatase inhibitors
may be of benefit mainly to only a minority of women with advanced disease.
Although in our study most women had detectable aromatase levels, we found
high aromatase levels in four of 43 women with endometrioid endometrial
cancer. Possibly these four women may be the ones with aromatase-dependent
endometrial cancer. It still remains unproven whether the adjuvant use of
aromatase inhibitors may improve prognosis of patients with endometrial
cancer, as in breast cancer. Theoretically, aromatase inhibition may influence
both peripheral and endometrial stromal aromatization. Some small studies
reported objective responses and a decrease of aromatase activity after
administration of aromatase inhibitors to women with advanced endometrial
cancer.35•36 One study reported very limited response to aromatase inhibition
therapy in patients with advanced or recurrent endometrial cancer. Only a
minority of these patients was diagnosed with (hormone-dependent)
endometrioid endometrial cancer, and women with poorly differentiated tumours
did not respond. 37
In conclusion, the local availability of androgens in the utero-ovarian
circulation, and the presence of aromatase activity in endometrium, support the
hypothesis that ovarian androgens may play a role in the development of
endometrioid endometrial cancer through conversion into estrogens.
Acknowledgements
The authors are grateful to Dr. WJ Sluiter for his contribution in statistical
analysis.
102
References
1. Simpson ER, Mahendroo MS, Means GD, Kilgore MW, Hinshelwood MM, Graham
Lorence S, et al. Aromatase cytochrome P450, the enzyme responsible for estrogen
biosynthesis. Endocr Rev 1994; 15 : 342-55. 2. Korzekwa KR, Trager WF, Mancewicz J, Osawa Y. Studies on the mechanism of
aromatase and other cytochrome P450 mediated deformylation reactions. J Steroid
BiochemMolec Biol 1993; 44 : 367-73.
3. Nelson LR, Bulun SE. Estrogen production and action. J Am Acad Dermatol 2001; 45 : S I 16-24.
4. Ala-Fossi SL, Maenpaa J, Aine R, Punnonen R. Ovarian testosterone secretion during
perimenopause. Maturitas 1998; 29 : 239-45. 5. Adashi EY. The climacteric ovary as a functional gonadotropin-driven androgen
producing gland. Fertil Steril 1994; 62: 20-7.
6. Longcope C. Editorial. The relationship of ovarian production of hormones and the degeree of ovarian stromal hyperplasia in postmenopausal women. Menopause 1998; 5 :
205-6.
7. Sluijmer AV. The postmenopausal ovary. Functional and morphological aspects. PhD
Thesis Maastricht, The Netherlands: University of Maastricht, 1999.
8. Mertens HJMM. Observational studies of human endometrium and endometrial cancer. The role of steroids and steroid hormone receptors in tissue proliferation, differentiation
and malignant transformation. PhD Thesis Maastricht, The Netherlands, University of
Maastricht2002.
9. Bremond AG, Claustrat B, Rudigoz RC, Seffert P, Corniau J. Estradiol, androstenedione,
testosterone, and dehydroepiandrosteronesulfate in the ovarian and peripheral blood of
postmenopausal patients with and without endometrial cancer. Gynecol Oncol 1982; 14 :
1 19-24.
10. Nagamani M, Stuart CA, Doherty MG. Increased steroid production by the ovarian stromal tissue of postmenopausal women with endometrial cancer. J Clin Endocrinol
Metab 1992; 74 : 172-6.
I 1. Jongen VHWM, Hollema H, van der Zee AGJ, Santema JG, Heineman MJ. Ovarian stromal hyperplasia and ovarian vein steroid levels in relation to endometrioid
endometrial cancer. Br J Obstet Gynaecol 2003 ; 1 10: 690-5.
103
12. Jongen VHWM, Sluijmer AV, Heineman MJ. The postmenopausal ovary as an androgen-producing gland; hypothesis on the etiology of endometrial cancer. Maturitas 2002; 43 : 77-85.
13. Heineman MJ, Sluijmer AV, Evers JLH. Utero-ovarian vein blood sampling in
postmenopausal women. Fertil Steril 1993; 60: 184-6.
14 . Jurjens H, Pratt JJ, Woldring MG. Radioimmunoassay of plasma estradiol without
extraction and chromatography. J Clin Endocrinol Metab 1975; 40: 19-25 . 15. Pratt JJ, Wiegman T, Lappohn RE, Woldring MG. Estimation of plasma testostosteron
without extraction and chromatography.Clin Chim Acta 1978; 84: 329-37.
16. Pratt JJ. Steroid immunoassay in clinical chemistry. PhD Thesis Groningen, The Netherlands, University of Groningen 1981 .
17. vd Ven J, Donker GH, Sprong M, Blankenstein MA, Thijssen JHH. Effect of tibolone (Org OD14) and its metabolites on aromatase and estrone sulphatase activity in human breast adipose stromal cells and in MCF-7 and T47D breast cancer cells. J Steroid Biochem Mol Biol 2002; 8 1 : 237-47.
18. Tseng L, Mazella J, Mann WJ, Chumas J. Estrogen synthesis in normal and malignant
human endometrium. J Clin Endocrinol Metab 1982; 55 : 1029-3 1. 19. Bulun SE, Noble LS, TakayamaK, Michael MD, Agarwal V, Fisher C, et al . Endocrine
disorders associated with inappropriately high aromatase expression. J Steroid Biochem Molec Biol 1997; 61 : 133-9.
20. WatanabeK, Sasano H, Harada N, Ozaki M, Niikura H, Sato S, YajimaA. Aromatase in human endometrial carcinoma and hyperplasia. Immunohistochemical, in situ
hybridization, and biochemical studies. Am J Pathol 1995; 146: 49 1-500.
21. Yamamoto T, Kitawaki J, Urabe M, Honjo H, Tamura T, Noguchi T, et al. Estrogen productivity of endometrium and endometrial cancer tissue; influence of aromatase on proliferation of endometrial cancer cells. J Steroid Biochem Molec Biol 1993; 44: 463-8.
22. Sasano H, Kaga K, Sato S, Yajima A, Nagura H, Harada N. Aromatase P450 gene
expression in endometrial carcinoma. Br J Cancer 1996; 74 : 1541-4.
23 . Yamaki J, Yamamoto T, Okada H. Aromatization of androstenedione by normal and neoplasticendometriumofthe uterus. J Steroid Biochem Molec Biol 1985; 22: 63-6.
24. Bulun SE, Mahendroo MS, Simpson ER. Polymerase chain reaction amplification fails
to detect aromatase cytochrome P450 transcripts in normal human endometrium or
decidua. J Clin EndocrinolMetab 1993; 76: 1458-63 .
104
25. Noble LS, Simpson ER, Johns A, Bulun SE. Aromatase expression in endometriosis. J Clin EndocrinolMetab 1996; 8 1 : 174-9.
26. Kitawak.i J, Noguchi T, Amatsu T, Maeda K, Tsukamoto K, Yamamoto T, et al.
Expression of aromatase cytochrome P450 protein and messenger ribonucleic acid in
human endometriotic and adenomyotic tissues but not in normal endometrium. Biol
Reprod 1997; 57: 5 14-9.
27. Bulun SE, Economos K, Miller D, Simpson ER. CYP 19 (aromatase cytochrome P450) gene expression in human malignant endometrial tumors. J Clin Endocrinol Metab
1994; 79 : 183 1-4. 28. Prefontaine M, Shih C, Pan CC, Bhavnani BR. Applicability of the product isolation
and radiometric aromatase assays for the measurement of low levels of aromatase: lack of aromatase activity in the human endometrium. J Endocrinol 1990; 127 : 539-
51 .
29. Baxendale PM, Reed MJ, James VHT. Inability of human endometrium or
myometrium to aromatize androstenedione. J Steroid Biochem Molec Biol 198 1 ; 14 :
305-6.
30. vd Ven J, Donker TH, Blankenstein MA, Thijssen JH. Differential effect of
gonadotropin-releasing hormone analogue treatment on estrogen levels and sulfatase activity in uterine leiomyoma and myometrium. Fertil Steril 2002; 77: 1227-32.
3 1 . Tada A, Sasaki H, Nakamura J, Yoshihama M, Terashima Y. Aromatase activity and
the effect of estradiol and testosterone on DNA synthesis in endometrial carcinoma
cell lines. J Steroid Biochem Molec Biol 1993; 44 : 661-6.
32. Sasano H, Sato S, Ito K, Yajima A, Nakamura J, Yoshihama M, et al. Effects of aromatase inhibitors on the pathobiology of human breast, endometrial and ovarian
carcinoma. Endocr Rel Cancer 1999; 6: 197-204.
33 . Bokhman JV. Two pathogenetic types of endometrial carcinoma. Gynecol Oncol 1983;
15 : 10-7.
34. van de Ven J. Local estrogen levels and estrogen biosynthesis in breast cancer and uterine leiomyomas: effect of endocrine therapies. PhD Thesis Utrecht, The
Netherlands, University of Utrecht, 2002.
35. Murray RLM, Pitt P. Treatment of advanced metastatic breast cancer, carcinoma of the
prostate and endometrial cancer with aminogluthethimidine. In: Nagel GA & Santen
RJ, ed. : Aminogluthethimidine as an aromatase inhibitor in the treatment of cancer.
Bern: Huber, 1984 : 109-22.
105
36. Bershtein LM, Maksimov SI, Gershfel 'd ED, Gamaiunova VB, Mesh.kova IE, Kovalenko IG et al. Neoadjuvant use of the aromatase inhibitor letrozole in uterine cancer: endocrine and clinical effects. Vopr Onkol 2001 ; 47(5): 5 1 7-74.
37. Rose PG, Brunetto VL, VanLe L, Bell J, Walker JL, Lee RB. A phase III trial of
anastrozole in advanced recurrent or persistent endometrial carcinoma: a gynecologic oncology group study. Gynecol Oncol 2000; 78 : 212-6.
106
6
Aromatase, COX-2, HER-2/neu and P53 as
prognostic factors in endometrioid endometrial
cancer
VHWM Jongen
JM Briet
RA de Jong
E Joppe
KA ten Hoor
HM Boezen
DB Evans
H Hollema
AGJ van der Zee
HW Nijman
Accepted for publication in the International Journal of Gynecological Cancer
107
Abstract
The prognostic value of aromatase, COX-2, HER-2/neu and p53 expression was
determined in endometrioid endometrial cancer. Tissue microarrays were
constructed comprising samples from 315 endometrioid endometrial cancer
patients. Expression of aromatase, COX-2, HER-2/neu and p53 was determined
by immunostaining and related to classical clinico-histopathological parameters,
in addition to recurrence of disease and survival. Median follow-up time for all
patients was 5.0 years. Patients were classified as FIGO stage I (59.0%), stage II
( 17. 1%), stage III ( 19.4%) and stage IV (4. 1%). Sixty-five patients (20.6%)
developed recurrent disease and 38 ( 12. 1 %) died because of endometrial cancer.
Aromatase, COX-2, HER-2/neu and p53 expression was observed in 133
(42.2%), 107 (34.0%), 17 (5.4%), and 21 (6.7%) tumour cases, respectively.
Aromatase expression in tumour cells was related to aromatase expression in
stromal cells (p<0.0001) and to HER-2/neu expression in tumour cells
(p=0.0 19). Aromatase expression in both tumour as well as stromal cells was
related to low stage of disease (p=0.02 and p=0.00 1, respectively), while
aromatase expression in stromal cells was also related to low tumour grade
(p=0.021). P53 expression was related to high stage and high grade (p=0.006
and p<0.0001, respectively). In multivariate analysis p53 overexpression was
independently related to death due to the disease (p=0.043, OR 3.0, 95% CI 1.0-
8.7). For COX-2, HER-2/neu and aromatase no relation with any other
histopathological parameter or survival was found. In conclusion, aromatase
expression in tumour and stromal cells and p53 expression in tumour cells are
related to tumour grade and stage of disease, while p53 is an independent
prognostic factor in endometrioid endometrial cancer. The possible interaction
of the studied parameters in the development of endometrial cancer is discussed.
108
Introduction
Over the past few years molecular markers such as aromatase, cyclooxygenase-2
(COX-2) and HER-2/neu have been identified as important cell biological
determinants, contributing to breast cancer development and simultaneously
becoming targets of new treatment modalities in breast cancer.1 •2 Studies in
breast cancer showed strong correlations between aromatase, COX-2 and HER-
2/neu enzyme systems, while interactions between these biomarkers were found
to play a role in the production of estrogens. 3•4 Comparable to breast cancer, also
in endometrial tissues aromatase activity may contribute to the development of
endometrioid endometrial cancer by local conversion of androgens into
estrogens, thereby inducing high local estrogen concentrations. 5•6 If indeed
activity of aromatase, COX-2 and HER-2/neu are involved in the development
of endometrioid endometrial cancer, a possible therapeutic role may be
considered for inhibitors of either aromatase, COX-2 or HER-2/neu in treatment
strategies of advanced or recurrent endometrial cancer.7 Likewise, strong
predictive markers could be used to further individualize treatment modalities,
next to classical clinico-histopathological factors.
Overexpression of p53 is another wellknown, cell biological prognostic factor
in endometrial cancer.8- 12 Data on a possible relation between p53 expression and
expression of HER-2/neu, aromatase or COX-2 are scarce and inconclusive in
endometrial cancer.13• 1 4 The aim of our study was to relate the expression of
aromatase, COX-2 and HER-2/neu to p53 expression, classical clinico
histopathological characteristics and to disease-free and overall survival m a
large, well-documented series of endometrioid endometrial cancer patients.
109
Patients and Methods
Study population
All patients were consecutively treated at the department of Obstetrics and
Gynecology of the University Medical Center of Groningen, the Netherlands,
between 1984 and 2004. Data of 411 patients with endometrial cancer were
prospectively collected based on hospital records including surgical notes and
pathology reports. Corresponding histological material, suitable for tissue
microarray (TMA) construction, could be obtained from 367 out of 411 patients.
In the remaining 44 patients paraffin-embedded tumour tissue was not available
for TMA construction because of either lack of cancer tissue, or insufficient
tissue quality. Finally 315 out of 367 patients had an endometrioid tumour type
and were selected for for the present study. From 1988 pelvic lymphadenectomy
was performed in case of a grade 3 endometrioid endometrial tumour and/ or
cervical involvement. Para-aortic lymphadenectomy was done in case of suspect
para-aortic lymph nodes at surgery, or proven positive pelvic nodes. Extensive
lymph node dissection was omitted in case of severe co-morbidity. Clinico
pathological characteristics and follow-up data were noted, using medical
records. Follow-up surveillance consisted of a clinical history, physical
examination and a yearly PAP smear of the vaginal vault with an increasing
interval until 10 years after treatment. Patient characteristics have been
described previously in a study from our group on lymph vascular space
involvement in endometrioid endometrial cancer. 1 5
Design and construction of the tissue microarray
The TMA was constructed, as described previously, using paraffin-embedded
tumour tissue of endometrial cancer patients. 16-19 In short, morphologically
representative areas of tumour were marked on hematoxylin- and eosin- stained
sections. Three 0.6 mm core biopsies of representative areas were taken from the
1 10
individual paraffin embedded tumour blocks. The core biopsies were then
assembled on a recipient paraffin block at pre-defined array locations, using a
precision instrument (Beecher Instruments, Silver Springs, Maryland).
Immunostaining
Four µm sections were cut and mounted on APES (3-aminopropyltriethoxylane;
Sigma-Aldrich, Diesenhofen, Germany) -coated glass slides. Sections were
dewaxed in xylene and rehydrated through graded concentrations of ethanol
( 100%, 90% and 70%) to distilled water. Antigen retrieval was performed when
needed by microwave method. For COX-2 antigen retrieval was performed for
15 minutes at 95 to 100° C in a microwave in lmM EDTA buffer (pH 8) and for
P53 1 OmM TRIS/EDT A (pH 9 .0) was used. Endogenous peroxidase activity was
blocked for 30 minutes with 0.3% H2O2. For aromatase endogenous avidin and
biotin activity was blocked for 15 minutes using a blocking kit (Vector
Laboratories, Burlingame, USA). Slides for aromatase staining were
preincubated with 1 % normal rabbit serum in Phosphate Buffered Saline (PBS)
pH 7.2 containing 1 % BSA. Slides for COX-2 and p53 were incubated with the
primary antibody for 60 minutes at room temperature and aromatase slides were
incubated overnight at 4 ° C. All dilutions were made in Phosphate Buffered
Saline (PBS) pH 7.2 containing 1 % BSA. For HER-2/neu a Ventana autostainer
was used (monoclonal anti rabbit 4B5 as primary antibody). As a primary
antibody for COX-2 we used anti mouse clone 33 (BD Biosciences) in a dilution
of 1 :200. For COX-2 rabbit anti mouse/HRP (DAKO) was used as secondary
antibody, followed by goat anti rabbit/HRP (DAKO), both 1 : 100 for 30 minutes
at room temperature. For p53 biomarker analysis anti mouse antibodies DO-7
(DAKO) were used as primary antibodies in a dilution of 1 :2000. Also for p53 a
goat anti-mouse-/rabbit secondary antibody conjugated with a peroxidase
labelled polymer (Dako En Vision+ system) was used. For determination of
aromatase expression we used antibody 677 (Novartis), as primary antibodies in
1 1 1
a dilution of 1 : 1 00.20•21 Biotinylated rabbit anti mouse (DAKO) was used as a
secondary antibody ( 1 :300 for 30 minutes at room temperature) for aromatase
after which Streptavidine/HRP (DAKO) 1 :300 was applied for 30 minutes at
room temperature. For all slides the antigen-antibody reaction was visualized
with 3 ,3 '-diaminobenzidine (DAB substrate kit, Vector Laboratories,
Burlingame, U.S.A.) for 1 0 minutes, sections were then counterstained with
Mayer's haematoxylin and the slides were dehydrated in graded ethanol, dried
and cover slipped. Negative controls were obtained by omission of the primary
antibody. The TMA included breast and colon cancer as well as placental tissue
of known biomarker expression for use as positive control.
Figure 1 . Positive immunohistochemical staining for aromatase expression in
tumour cells.
1 12
Analysis of immunostaining on TMA
Immunostaining was semi-quantitatively scored based on the intensity of
nuclear or cytoplasmic staining. Intensity of immunostaining for receptor
expression was scored as negative (0), weak ( 1+), positive (2+) or strongly
positive (3+) staining. For aromatase both staining of stromal as well as tumour
cells was scored (Figure 1 ). In accordance to literature, weak to positive
cytoplasmic staining in stromal or tumour cells was considered as aromatase
expression positivie. 1 7•2° COX-2 was considered positive if more than 1 0% of the
tumour cells showed positive cytoplasmatic staining. 1 7 Immunostaining for DO-
7, an antibody which detects both wild-type and mutant p53 protein, was scored
according to the intensity of nuclear staining and to the percentage of positively
stained tumour cells. P53 was considered positive ( overexpression) in case a
substantial percentage of the nuclei (more than 50%) stained positive as
described by others.8 HER-2/neu expression was scored as recommended by the
manufacturer's scoring guidelines. HER-2/neu staining was considered positive
if there was moderate to strong complete membrane staining in more than 10%
of tumour cells, comparable to a recently published study by Morrison et al. 22
Immunostaining was scored by two independent observers (KH and VJ).
Observers had no prior knowledge of clinicopathological information. Cases
with a discrepancy in score between the two observers were re-examined with a
gynaecological pathologist (HH) until full agreement was reached. Minimally
two cores with representative tumour had to be present on the TMA for a sample
to be entered into analysis. The mean overall core loss was 6 .6% ( 6 .3-7 .6% ) .
Statistics
All analyses were performed using SPSS 14.0. Differences between groups were
tested using the X2 or the Fisher's exact test. Kaplan-Meier survival analysis and
Cox regression analysis were performed to study the role of aromatase, COX-2,
1 1 3
Table 1 . Histological and clinical characteristics of included patients with
endometrioid endometrial cancer (n=315).
Patient age, years Median 64.7 Range 32-89
Tumour Stage I 186 59.0 II 54 17. 1 III 6 1 19.4 IV 13 4. 1 Missing 1 0.3
Histological grade Grade 1 162 5 1.4 Grade 2 96 30.5 Grade 3 56 17.8 Undifferentiated 1 0.3
Myometrial invasion < 50% 185 58.7 > 50% 126 40.0 Missing 4 1.3
L VSI Negative 222 70.5 Positive 73 23 .2 Missing 20 6.3
Recurrence No recurrence 250 79.4 Local 26 8.3 Regional 8 2.5 Distant 3 1 9.8
Pelvic lymph nodes Negative 97 30.8 Positive 29 9.2 No sampling 189 60.0
Paraaortic lymph nodes Negative 45 14.3 Positive 12 3 .8 No sam lin 258 8 1.9
LVSI: Lymph Vascular Space Involvement
1 14
HER-2/neu, and p53 in relation to recurrence (cq disease free survival) and
death ( survival) as end points. Multiple logistic regression analyses were
performed with COX-2, HER-2/neu, aromatase and p53 as the dependent
variables and known risk factors lymph vascular space invasion, myometrial
invasion, grade and stage of disease as independent variables entered
simultaneously into the model. Patients were censored at date of last contact or
death in case of no relapse. P-values < 0.05 were considered statistically
significant (tested 2-sided).
Results
In Table 1 the clinico-pathological characteristics of 315 patients with
endometrioid endometrial cancer are shown. The median follow-up time for all
patients was 4.97 years (range 0.0-21.5). Follow-up data were completed until
September 2006. Median age at time of diagnosis was 64.7 years (range 32.0-
89.0). Patients were classified as FIGO stage I (59.0%), stage II ( 17.1 %), stage
III ( 19 .4 % ) and stage IV ( 4 . 1 % ) . Hysterectomy and bilateral salpingo
oophorectomy was performed in 56.2% of the patients, and more extended
surgery in 41.9% of the patients. Adjuvant post-operative radiotherapy was
given to 168 patients (53.4%). In 65 patients (20.6%) recurrent disease was
diagnosed. Local relapse at the vaginal vault occurred 26 times, while regional
and distant metastases were diagnosed 8 and 31 times, respectively. The median
time to relapse was 1 . 7 years (range: 0.1-7 .6). During follow-up 8 1 out of 315
patients died (26.0%), 38 patients ( 12. 1 %) died from endometrial cancer.
1 15
Association of aromatase, COX-2, HER-2/neu, and P53 with clinico
h istopathological factors.
Aromatase expression in tumour cells was observed in 133 (42.2%) tumours,
while 89 (28.3%) tumours showed aromatase expression in stromal cells. COX-
2, HER-2/neu and p53 expression was observed in 107 (34.0%), 17 (5.4%) and
2 1 (6.7%) patients, respectively. (Table 2). Expression of aromatase in tumour
cells was related to expression in stromal cells (p<0.0001 ). Aromatase
express10n in both tumour and stromal cells was related to HER-2/neu
expression (p=0.0 19 and p=0.02 1 respectively). Aromatase expression in tumour
and stromal cells was also related to low stage of disease (p=0.02 and p=0.00 1
respectively), while aromatase expression in stromal cells correlated with low
tumour grade (p=0.021 ) (Table 3). P53 expression was related to high stage of
disease and tumour grade (p=0.006 and p<0.000 1 , respectively) (Table 3 and 4).
In a multivariate (logistic regression) analysis absence of lymph vascular space
invasion was related to COX-2 expression (p=0.029, OR 0.46, 95% CI 0.2-0.9)
(Table 5). No relation was found between aromatase expression and disease free
or overall survival. In univariate analysis risk for death of disease was increased
in the presence of p53 (p=0.047) (Figure 2). P53 also had an independent
prognostic value in multivariate analysis with all prognostic factors for death
due to disease (p=0.043, OR 3.0, 95% CI 1 . 0-8.7) (Table 6). In a multivariate
analysis tumour grade was the strongest predictor for recurrence (p=0.003, OR
3. 6, 95% CI 1 . 5-8.3) (Table 7).
1 16
Table 2. Co-expression of aromatase, HER-2/neu, COX-2 and P53 in 315
patients with endometrioid endometrial cancer.
Aromatase Aromatase HER-2/ neu COX-2 cancer cells in stroma (n = 1 33) (n = 89) (n = 1 7) (n = 1 07)
Aromatase in cancer Positive X 83 (93.3%) 12 (70.6%) 87 (81 .3%) cells
Negative X 6 (6 .7%) 5 (29.4%) 20 ( 1 8.7%) Aromatase in stroma Positive 83 (62.4%) X 9 (52.9%) 24 (22.4%)
Negative 50 (37.6%) X 8 (47. 1 %) 83 (77.6%)
HER-2/ neu Positive 1 2 (9.0%) 9 ( 1 0. 1 %) X 1 1 ( 1 0.3%)
Negative 1 1 1 (91 .0%) 80 (89.9%) X 96 (89 .7%)
COX-2 Positive 43 (32.3%) 24 (27 .0%) 1 1 (64.7%) X
Negative 90 (67.7%) 65 (73.0%) 6 (35.3%) X
P53 Positive 9 (6.8%) 6 (6.7%) 4 (23.5%) 1 1 ( 1 0.3%)
Negative 1 24 (93.2%) 83 (93.3%) 1 3 (76 .5%) 96 (89.7%)
1 17
P53
(n=21 )
9 (42 .9%)
1 2 (57. 1 %)
6 (28.6%)
15 (71 .4%)
4 ( 1 9 .0%)
1 7 (81 .0%)
1 1 (52 .4%)
1 0 (47.6%)
X
X
Table 3 . Chi square analysis of receptor expression in relation to known
prognostic factors in patients with endometrioid endometrial cancer.
StaQe Grade Stage Stage Grade Grade 3 and I and I I I l l and IV 1 and 2 undifferentiated
Aromatase Neg in cancer n= 1 62 1 1 5 47 1 1 4 1 9 cells Pos
n= 1 33 1 09 23 1 26 36
p value 0.020 NS
Aromatase Neg in stroma n=206 146 60 1 61 45
Pos n=89 78 1 0 80 9
p value 0.001 0 .021 HER-2/neu Neg
n=274 2 10 64 220 54 Pos n= 1 7 1 2 5 1 5 2
p value NS NS
COX-2 Neg n= 1 87 140 47 1 46 41 Pos n=1 07 84 23 92 1 5
p value NS NS
P53 Neg n=274 2 1 3 60 229 45
Pos n=2 1 1 0 1 1 1 0 1 1
p value 0.006 <0.0001
P derived from 2 sided testing using Chi-square test Abbreviations Neg = negative
Pos = positive
LVSI LVSI Neaative
1 1 4
94
142
66
1 96
1 0
1 23
85
1 93
1 4
LVSI = Lymph Vascular Space Involvement
LVSI Positive
41
29
NS
53
1 7
NS
67
3
NS
55
1 6
0.006
66
6
NS
1 1 8
Table 4. The independent effect of stage of disease, tumour grade, myometrial
invasion and L VSI on the presence of p53 expression in endometrioid
endometrial cancer (logistic regression analysis).
95.0% Cl for OR EXP(B) P value
Lower Upper
Stage of 3 . 1 65 1 . 1 2 1 8.938 0.030 d isease
Tumour grade 4.797 1 .666 1 3.8 14 0.004
Myometrial invasion 0.732 0.250 2. 1 49 0.570
LVSI 0.51 7 0. 1 47 1 .8 1 5 0.303
LVSI = Lymph Vascular Space Involvement
Table 5. The independent effect of stage of disease, tumour grade, myometrial
invasion and LVSI on the presence of COX-2 expression in endometrioid
endometrial cancer (logistic regression analysis).
95.0% Cl for OR EXP(B) P value
Lower Upper
Stage of d isease 1 .201 0.633 2.279 0 .575
Tumour grade 0.787 0.387 1 .602 0.509
Myometrial invasion 0.822 0.473 1 .429 0.488
LVSI 0.461 0.231 0.923 0.029
LVSI = Lymph Vascular Space Involvement
1 1 9
Table 6. The risk of death due to endometrioid endometrial cancer according to
the presence of aromatase in cancer and stromal cells, HER-2/neu, COX-2, p53,
stage of disease, tumour grade, myometrial invasion and L VSI ( cox regression
analysis).
I -----OR�-
95.
�:�(�\
for
� Lower Upper
Aromatase in cancer cells
Aromatase in stroma
I HER-2/neu
COX-2
P53
Stage of disease
Tumour grade
Myometrial invasion
I LVSI
3 .702
0.584
0.660
0.474
3.01 0
1 .3 1 7
1 . 1 85
2. 1 1 4
1 .264
0.941 14 .562
0. 1 48
0. 1 02
0. 1 77
1 .037
0.583
0.435
0.866
0.5 1 9
2.300
4.278
1 .264
8.738
2.973
3.229
5. 1 56
3.080
LVSI = Lymph Vascular Space Involvement
P value
0.061
0.442
0.663
0 . 1 36
0.043
0.508
0.74 1
0 . 1 00
0.605
120
Table 7. The risk of relapse of endometrioid endometrial cancer according to the
presence of aromatase in cancer and stromal cells, HER-2/neu, COX-2, p53,
stage of disease, tumour grade, myometrial invasion and L VSI ( cox regression
analysis).
--95.0% Cl for
OR Exp(B) P value Lower Upper
Aromatase in 0.895 0.432 1 .856 0.766 cancer cells
Aromatase in 0.651 0.251 1 .688 0.377 stroma
HER-2/neu 1 .751 0.234 1 3.083 0 .585
COX-2 0.529 0.269 1 .037 0.064
P53 1 .086 0.359 3.282 0 .884
Stage of disease 0.854 0.487 1 .498 0.581
Tumour grade 3 .569 1 .535 8.299 0.003
Myometrial invasion 1 . 1 95 0.664 2. 1 50 0.553
LVSI 0.654 0.337 1 .272 0.21 1
LVSI=Lymph Vascular Space Involvement
12 1
Figure 2. Kaplan-Meier survival analysis concerning p53 expression and
cumulative disease-specific survival (p=0.047).
Survival Funct ions
0,0 2 ,0 4 P s p 8,0
follow-up in years
1 0,0
p53 e xp re ss ion -- negative
-- positive
122
Discussion
In this study in a large cohort of endometrial cancers aromatase expression was
frequently found both in tumour as well as in stromal cells. We found aromatase
in tumour and stromal cells to be related to low tumour grade and low stage of
disease. Also, aromatase expression in tumour cells was related to expression of
HER-2/neu. The aromatase enzyme is a product of the CYP 19 gene and is
capable of converting C 19 steroids (androgens), produced by adrenal glands
and ovaries, into C 18 steroids (estrogens). Conversion of circulating androgens
into estrogens may take place not only in peripheral adipose tissues, but in
endometrial tissue as well.6•7•17•23•24 In this way, aromatase activity may be
involved in the development of endometrial cancer.7•25-28 Our results suggest that
aromatase activity in tumour cells plays a role especially in lower stage and
grade of endometrial cancer, as the aromatase expression in tumour cells appears
to be lost with increasing stage and/ or dedifferentiation. As the presence or
absence of aromatase expression did not contribute to disease outcome, no
arguments could be found for adjuvant treatment with aromatase inhibitors in
the palliative setting. We found aromatase expression in cancer cells related to
aromatase expression in stromal cells. Possibly aromatase expression in the
nests of cancer cells is responsible for local conversion of androgens in
estrogens. Aromatase expression in the surrounding stroma may represent tissue
reaction, as aromatase expression was found far less frequently in stromal cells
than in tumour cells and expression in stromal cells was not related to expression
in cancer cells. In contrast, Watanabe localized aromatase expression in stromal
cells and suggested a paracrine growth mechanism through in situ production of
estrogens in stromal cells. 5 This different localization might be attributable to
differences in antibodies and techniques used. The aromatase antibody 677, that
was used for the first time in endometrial cancer in our study may be more
reliable, as its specificity was recently established in an international
123
collaborative effort, that had the specific purpose of intratumoural detection of
aromatase in archival or paraffin-embedded tissue.20•2 1
In the literature HER-2/neu overexpression has been correlated with non
endometrioid endometrial cancer, advanced stage, high grade, myometrial
invasion and recurrent disease and was suggested to be used as a prognostic
indicator for hormone-independent growth.29-3 1 Interestingly, in our patient group
with only endometrioid endometrial cancer a correlation was found between
aromatase and HER-2/neu expression. Also in breast cancer HER-2/neu and
aromatase overexpression were reported to be related, with HER-2/neu as the
primary molecular determinant of increased aromatase activity and COX-2 as
the potential functional intermediate. 3•4 Increased prostaglandin synthesis from
COX-2 expression prompts formation of aromatase and may play a role in the
initial development of endometrial cancer.1•
1 3•32
-34 This may lead to the interesting
possibility that the regulation of aromatase expression m breast and
endometrioid endometrial cancer follows similar pathways, as both appear
examples of aromatase expressing estrogen dependent tumours.
We found COX-2 expression in 36% of endometrioid tumour cases and others
have shown positive immunostaining from 39 to 69% in endometrial
cancer.1 3•32
•35
-37 COX-2 did not appear to be a prognostic marker in our large
cohort study. In other studies, conflicting results were found on the relation
between COX-2 expression and survival and clinical variables such as stage and
histological type.4•
1 3•
1 7•32
•36 A strong relation in breast cancer was found between
COX-2 and CYP19 gene expression, of which the aromatase enzyme is the
product. 38•39 In contrast to a previous report, we did not find a correlation with
aromatase, although a tendency existed for aromatase expression in stromal cells
to be related to COX-2 expression (p=0.074). 1 7
P53 is a tumour-suppressor protein which causes cell cycle arrest in case of
DNA damage, to allow DNA repair or induction of apoptosis in case of
substantial damage. In malignancies p53 is often overexpressed due to mutations
124
and thereby uncapable of performing DNA repair. In our study in endometrioid
endometrial cancer p53 overexpression was independently related to a high
stage of disease, grade and to poor prognosis. Previous studies found p53
overexpression especially associated with non-endometrioid endometrial
cancer.29•40 Although in our study p53 over-expression was a relatively rare
observation (7. 1%), it appeared to be a strong prognostic marker.
We conclude that the local presence of aromatase in early stage and low grade
endometrioid endometrial cancer supports the hypothesis that aromatase may
contribute to the promotion of endometrial cancer. As aromatase expression in
both tumour and stromal cells was related to HER-2/neu expression, the
regulation of aromatase expression in breast and endometrioid endometrial
cancer may follow similar pathways. Also, in endometrioid endometrial cancer,
p53 is an independent prognostic marker for disease-specific survival.
Acknowledgements
We acknowledge the support of Novartis Pharma AG, Basel, Switzerland, for
their help in disposing the 677 aromatase antibodies.
HW Nijman is supported by the Dutch Cancer Society (Grant 2002-2768).
125
References
I . Brueggemeier RW, Diaz-Cruz ES. Relationship between aromatase and cyclooxygenases in breast cancer: potential for new therapeutic approaches. Minerva
Endocrinol 2006; 3 1 : 13-26.
2. Ristimaki A, Sivula A, Lundin J, Lundin M, Salminen T, Haglund C, et al. Prognostic significance of elevated cyclooxygenase-2 expression in breast cancer. Cancer Res 2002; 62: 632-5.
3 . Subbaramaiah K, Norton L, Gerald W, Dannenberg AJ. Cyclooxygenase-2 is
overexpressed in HER-2/neu-positive breast cancer: evidence for involvment of AP- I and PEA3. J Biol Chem 2002; 277: 18649-57.
4 . Subbaramaiah K, Howe LR, Port ER, Brogi E, Fishman J, Liu CH, et al. HER-2/neu status is a determinant of mammary aromatase activity in vivo : evidence for a cyclooxygenase-2 dependent mechanism. Cancer Res 2006; 66: 5504- 1 1 .
5. Watanabe K, Sasano H, Harada N, Ozaki M, Niikura H, Sato S, et al. Aromatase in human endometrial carcinoma and hyperplasia. Am J Pathol 1995 ; 146: 49 1 -500.
6. Jongen VHWM, Thijssen JHH, Hollema H, Danker GH, Santema JG, van der Zee AGJ, et al. Is aromatase cytochrome P450 involved in the pathogenesis of endometrioid endometrial cancer? Int J Gynecol Cancer 2005 ; 15 : 529-36.
7. Jongen VHWM, Hollema H, van der Zee AGJ, Heineman MJ. Aromatase in the
context of breast and endometrial cancer. A review. Minerva Endocrinologica 2006;
3 1 : 47-60.
8. Alkushi A, Lim P, Coldman A, Huntsman D, Miller D, Gilks CB. Interpretation of P53 immunoreactivity in endometrial carcinoma: establishing a clinically relevant cut-off
level. Int J Pathol 2004; 23 : 409- 1 0. 9. Sung CJ, Zheng Y, Quddus MR, Kang X, Zhang ZF, Lauchlan, et al. P53 as a
significant prognostic marker in endometrial carcinoma. Int J Gynecol Cancer 2000; 10: 1 19-27.
1 0. Hachiguda T, Fukuda K, Uchiyama M, Matsao N, Isawaka T, Sugimori H. Immunohistochemical study of P53 expression in endometrial carcinomas: correlation
with markers of proliferating cells and clinicopathological features. Int J Gynecol
Cancer 1993; 3: 363 .
126
1 1 . Powell B, Soong R, Grieu F, Knowles S, Hammond I, Iacopetta B. P53 protein
overexpression is a prognostic factor of poor survival in stage I endometrial carcinoma. Int J Oncol 1999; 14 : 175-9.
12. Soong R, Knowles S, Williams KE, Hammond IO, Wysocki SJ, Iacopetta BJ. Overexpression of P53 protein is an independent prognostic factor in human
endometrial carcinoma. Br J Cancer 1996; 74: 562-7. 13 . Jeon YT, Kang SB, Hang DH, Yoo KY, Park IA, Bang YJ, et al. Cyclooxygenase-2
and P53 expressions in endometrial cancer. Cancer Epidem Biomarkers & Prevention
2004; 13 : 1538-42.
14. Ferrandina G, Ranalletti FO, Gallota V, Martinelli E, Zannoni OF, Gessi M, et al. Expression of cyclooxygenase-2 (COX-2), receptors for estogen (ER), and
progesterone (PR), P53, ki67, and neu protein in endometrial cancer. Gynecol Oncol
2005; 98: 383-9 . 15. Briet JM, Hollema H, Reesink N, Aalders JG, Mourits MJE, ten Hoor KA, et al.
Lymphvascular space involvment: an independent prognostic factor in endometrial
cancer. Gynecol Oncol 2005; 96: 799-804.
16. Kallioniemi OP, Wagner U, Kononen J, Sauter G. Tissue microarray technology for
high-throughput molecular profiling of cancer. Hum Mol Genet 2001; 10: 657-62. 17. Fowler JM, Ramirez N, Cohn DE, Kelbick N, Pavlka J, Ben-Shachar I, et al.
Correlation of cyclooxygenase-2 (COX-2) and aromatase expression in human
endometrial cancer: Tissue microarray analysis. Am J Obstet Gynecol 2005 ; 192: 1262-7 1.
18. Hoos A, Cordon-Cardo C. Tissue microarray proofiling of cancer specimens and cell
lines: opportunities and limitations. Lab Invest 2001; 8 1 : 133 1-8.
19. Hendriks Y, Franken P, Diersen JW, de Leeuw W, Wijnen J, Dreef E, et al. Conventional and tissue microarray immunohistochemical expression analysis of
mismatch repair in hereditary colorectal tumors. Am J Pathol 2003; 162: 469-77.
20. Sasano H, Edwards DP, Anderson TJ, Silverberg SO, Evans DB, Santen RJ, et al.
Validation of new aromatase monoclonal antibodies for immunohistochemistry:
progress report. J Steroid Biochem Molec Biol 2003; 86: 239-44.
21 . Sasano H, Anderson TJ, Silverberg SO, Santen RJ, Conway M, Edwards DP, et al. The
validation of new aromatase monoclonal antibodies for immunohistochemistry - A
correlation with biochemical activities in 46 cases of breast cancer. J Steroid Biochem Molec Biol 2005; 95 : 35-9.
127
22. Morrison C, Zanagnolo V, Ramirez N, Cohn D, Kelbick N, Copeland L, et al. HER-2 is an independent prognostic factor in endometrial cancer : association with outcome in a large cohort of surgically staged patients. J Clin Oncol 2006: 24 : 2376-85.
23. Ala-Fossi SL, Maenpaa J, Aine R, Punnonen R. Ovarian testosterone secretion during perimenopause. Maturitas 1998; 29 : 239-45.
24. Jongen VHWM, Sluijmer AV, Heineman MJ. The postmenopausal ovary as an androgen-producing gland; hypothesis on the etiology of endometrial cancer.
Maturitas 2002; 43 : 77-85 .
25. Sasano H, Kaga K, Sato S, Yajima A, Nagura H, Harada N. Aromatase P450 gene
expression in endometrial carcinoma. Br J Cancer 1996; 74 : 154 1 -4.
26. Yamamoto T, Kitawaki J, Urabe M, Honjo H, Tamura T, Noguchi T, et al. Estrogen productivity of endometrium and endometrial cancer tissue; influence of aromatase on
proliferation of endometrial cancer cells. J Steroid Biochem Molec Biol 1993 ; 44 : 463-8.
27. Pathirage N, Di Nezza LA, Salmonsen LA, Jobling T, Simpson ER, Clyne CD. Expression of aromatase, estrogen receptors, and their coactivators in patients with
endometrial cancer. Fertil Steril 2006; 86: 469-72.
28. Segawa T, Shozu M, Murakami K, Kasai T, Shinohara K, Nomura K, et al. Aromatase expression in stromal cells of endometrioid endometrial cancer correlates with poor
survival. Clin Cancer Res 2005 : 1 1 : 2 188-94. 29 . Coronado PJ, Vidart JA, Lopez-Asenjo JA, Fasero M, Furio-bacete V, Magrina J, et al.
P53 overexpression predicts endometrial carcinoma recurrence better than HER-2/neu overexpression. Eur J Obstet Gynecol Reprod Biol 2001; 98 : 103-8.
30. Rolitsky CD, Theil KS, McGaughy VR, Copeland LJ, Niemann TH. HER-2/neu amplification and overexpression in endometrial carcinoma. Int J Gynecol Pathol
1999; 18: 138-43.
3 1. Mariani A, Sebo TJ, Katzmann JA, Riehle DL, Dowdy SC, Keeney GL, et al. HER-2/neu overexpression and hormone dependency in endometrial cancer: analysis of
cohort and review of literature. Anticancer Res 2005; 25 : 2921-7.
32. Fujiwaki R, Iida K, Kanasaki H, Ozaki T, Hata K, Miyazaki K. Cyclooxygenase-2 expression in endometrial cancer: correlation with microvessel count and expression of vascular endothelial growth factor and thymidine phosphorylase. Hum Pathol 2002;
33 : 213-9 .
128
33. Richards JA, Brueggemeier RW. Prostaglandin E2 regulates aromatase activity and
expression in human adipose stromal cells via two distinct receptor subtypes. J Clin
Endocrinol Metab 2003; 88: 28 10-6.
34. Richards JA, Petrel TA, Brueggemeier RW. Signaling pathways regulating aromatase and cyclooxygenases in normal and malignant breast cells. J Steroid Biochem Malec Biol 2002; 80: 203- 12.
35. Orejuela F, Ramondetta LM, Smith J, Brown J, Lemos LB, Li Y, et al. Estrogen and
progesterone receptors and cyclooxygenbase-2 expression in endometrial cancer,
endometrial hyperplasia, and normal endometrium. Gynecol Oneal 2005; 97: 483-8. 36. Ferrandina G, Legge F, Ranelleti F, Zannoni GF, Maggiano N, Evangelista E, et al.
Cyclooxygenase-2 expression in endometrial carcinoma. Correlation with
clinicopathologic parameters and clinical outcome. Cancer 2002; 95 : 801-7.
37. Landen Jr. CR, Mathur SP, Richardson MS, Creasman WT. Expression of cyclooxygenase-2 in cervical, endometrial, and ovarian malignancies. Am J Obstet
Gynecol 2003 ; 188: 1 174-6.
38. Brodie AM, Lu Q, Long BJ, Fulton A, Chen T, Macpherson N, et al. Aromatase and COX-2 expression in human breast cancers. J Steroid Biochem Moecl Biol 2001; 79 :
4 1-7.
39. Brueggemeier RW, Quinn AL, Parrett ML, Joarder FS, Harris RE, Robertson FM. Correlation of aromatase and cyclooxygenase gene expression in human breast cancer
specimens. Cancer Lett 1999; 140: 27-35 .
40. Macwhinnie N, Managhan H. The use of P53, PTEN and C-erb-2 to differentiate uterine serous papillary carcinoma from endometrioid endometrial carcinoma. In J
Gynecol Cancer 2004; 14 : 938-46.
129
130
7
Expression of estrogen and progesterone receptors
alpha and beta in a large cohort of patients with
endometrioid endometrial cancer
VHWM Jongen
JM Briet
RA de Jong
KA ten Hoor
HM Boezen
AGJ van der Zee
HW Nijman
H Hollema
Submitted
1 3 1
Abstract
The prognostic impact of estrogen receptor (ER) and progesterone receptor (PR)
alpha and beta was determined in endometrioid endometrial cancer. Tissue
microarrays were constructed from 315 endometrioid endometrial cancer
patients. Receptor expression was assessed by immunostaining, and their semi
quantitatively determined expression levels were correlated to classical clinico
histopathological parameters in addition to disease free and disease specific
survival. Patients were classified as FIGO stage I (59.0%), stage II (17.1%),
stage III (19.4%) and stage IV (4.1 %). Sixty-five patients (20.6%) developed
recurrent disease and 38 (12.1 %) died due to endometrial cancer. In univariate
analysis, expression of ER alpha was related to early stage endometrial cancer
(p=0.020), while expression of ER alpha, PR alpha and PR beta was associated
with lower grade tumours (p<0.0001, p<0.001 and p=0.001 respectively). A
ratio of ER alpha / ER beta < l was related to a shorter disease free survival
(p=0.027), while the ratio of PR alpha / PR beta <l both was associated with a
shorter disease free survival as well as a shorter overall survival (p=0.044 and
p=0.005, respectively). In early stage disease, using multivariate analysis,
absence of ER alpha was independently related to death of disease (p=0.017, OR
7.28, 95% CI 1.42-37.25), while absence of PR alpha (p=0.015, OR 4.2, 95% CI
1.32-13.33) appeared to be an independent prognostic factor for relapse of
disease. We conclude that in early stage endometrioid endometrial cancer
absence of PR alpha is an independent prognostic factor for disease-free
survival, while patients with ER alpha positive tumours have a better overall
survival.
1 32
Introduction
Endometrial cancer is the most frequently found genital tract malignancy in
women, and consists of two major histological types, endometrioid endometrial
cancer, and nonendometrioid endometrial cancer including high-risk malignancies
such as serous papillary and clear cell carcinoma. Endometrioid carcinoma is the
most common form, accountable for more than 7 5% of all cases of endometrial
cancer.1 Surgical treatment of endometrial cancer consists of total abdominal
hysterectomy with bilateral salpingo-oophorectomy next to pelvic and para
aortic lymphadenectomy in high risk patients. Patients with stage I or II
endometrioid endometrial cancer and poor prognostic factors such as deep
myometrial invasion and / or poor differentiation grade are treated additionally
with pelvic radiotherapy, which reduces risk of local recurrence without
prolonging overall survival.2•3 Recently, it was suggested that patients with high
risk early stage endometrial cancer might benefit from adjuvant chemotherapy.4
The question remains how to identify patients, who might benefit from adjuvant
treatment modalities. While some studies report an independent prognostic value
for estrogen receptor (ER), and progesterone receptor (PR) expression in
endometrioid endometrial cancer, others could not confirm this.5-13 Also, clinical
data in relation to the prevalence of steroid receptor isoforms ER and PR alpha
and beta are scarce. 7-9,11 The present study aimed to assess the prognostic value
of ER alpha and beta, and PR alpha and beta in a large cohort of endometrioid
endometrial cancer patients. We were particularly interested in the potency of
steroidmarkers for prediction of prognosis, as this may lead to a more patient
specific risk profile and treatment. 6•14
133
Patients and Methods
Study population
All patients were consecutively treated at the department of Obstetrics and
Gynaecology of the University Medical Center of Groningen, the Netherlands,
between 1 984 and 2004. Data of 41 1 patients with endometrial cancer were
prospectively collected based on hospital records including surgical notes and
pathology reports. Corresponding histological material, suitable for tissue
microarray (TMA) construction, could be obtained from 367 out of 4 1 1 patients.
In the remaining 44 patients paraffin-embedded tumour tissue was not available
for TMA construction because of either lack of cancer tissue, or insufficient
tissue quality. Finally 3 15 out of 367 patients had an endometrioid tumour type
and were selected for the present study. Clinico-pathologic and follow-up data of
the patients have been described previously. 1 5 From 1 988 pelvic
lymphadenectomy was performed in case of a grade 3 endometrioid endometrial
tumour or cervical involvement. Para-aortic lymphadenectomy was done in case
of suspect para-aortic lymph nodes at surgery, or proven positive pelvic nodes.
Extensive lymph node dissection was omitted in case of severe co-morbidity.
Clinico-pathological characteristics and follow-up data were noted, using
medical records. Follow-up surveillance consisted of a clinical history, physical
examination and a yearly PAP smear of the vaginal vault with an increasing
interval until 10 years after treatment.
Design and construction of the tissue microarray
The TMA was constructed, as described previously, using paraffin-embedded
tumour tissue of endometrial cancer patients. 16-1 9 Three representative cores of
0.6 mm diameter were taken from the tumour tissue block and placed in a
recipient paraffin block on pre-defined array locations, using a precision
instrument (Beecher Instruments, Silver Spring, Maryland). For each TMA
1 34
block control cores with known biomarker expression were incorporated in the
arrays, together with endometrial cancer tissue originating from the same donor
blocks for use as internal positive control for intra-run variability. Controls
included tissues of breast cancer and benign endometrium, gut, and placenta.
Immunostaining
Four µm sections were cut from the array blocks and mounted on APES (3-
aminopropyltriethoxylane; Sigma-Aldrich, Diesenhofen, Germany) -coated
glass slides. Sections were dewaxed in xylene and rehydrated through graded
concentrations of ethanol (100%, 90% and 70%) to distilled water. Antigen
retrieval was performed by either autoclave or microwave method when needed.
Antigen retrieval by autoclave treatment 3 times 5 minutes at 1 l 5°C was used
for PR beta in Blocking reagent (Boehringer Mannheim) (2% block + 0.2% SDS
in maleic acid, pH 6.0) and for ER alpha and ER beta in 1 OmM citrate buffer
(pH 6.0). For PR alpha antigen retrieval was performed for 15 minutes at 95 to
100°C in a microwave 1 mM EDT A buffer (pH 8). Endogenous peroxidase
activity was blocked by incubating the slides for 30 minutes with 0.3% H2O2.
ER alpha slides were preincubated with 1 :5 diluted normal goat serum for 15
minutes at room temperature. For ER beta endogenous avidin and biotin activity
was blocked for 15 minutes using a blocking kit (Vector Laboratories,
Burlingame, USA). Slides were then incubated with the primary antibody for 60
minutes at room temperature. For ER alpha 0.05M TRI S Buffered Saline (TBS)
pH 7.2 containing 1 % BSA was used for all dilutions. For ER beta the primary
antibody was diluted in 0.05M TRIS/HCL/1 % BSA, all other dilutions were
made in PBS pH 7 .2 containing 1 % BSA. The primary antibodies and dilutions
used are mentioned in Table 1. For ER alpha goat anti mouse IgG 1/HRP 1 :40
and for PR alpha and PR beta rabbit anti mouse/HRP (DAKO, Glostrup,
Denmark) 1: 100 was used as secondary antibody, followed by goat anti
rabbit/HRP (DAKO), 1: 100 for 30 minutes at room temperature. Biotinylated
135
rabbit anti mouse (DAKO) was used as a secondary antibody (1 :300 for 30
minutes at room temperature) for ER beta after which Streptavidine/HRP
(DAKO), 1 :300 was applied for 30 minutes at room temperature. For all slides
the antigen-antibody reaction was visualised with 3,3 '-diaminobenzidine (DAB
substrate kit, Vector Laboratories) and sections were counterstained with
Mayer's haematoxylin. The TMA included breast and colon cancer as well as
placental tissue of known biomarker expression for use as positive control.
Negative controls were obtained by omission of the primary antibody.
Table 1 . Primary antibodies and dilutions, used for receptor analysis.
Antigen Antibody Clone Company Dilution
Estrogen receptor alpha Monoclonal mouse 6F l 1 Serotec 1 :20
Estrogen receptor beta Monoclonal mouse ppg5/1 0 Serotec 1 :20
Progesterone receptor Monoclonal mouse HPRa7 Neomarkers 1 :50
alpha
Progesterone receptor Monoclonal mouse HPRa2 Neomarkers 1 :50
beta
Analysis of immunostaining on TMA
The receptor immunoreactivity for ER and PR was semi-quantitatively scored
according to the intensity and percentage of tumour cells showing positive
nuclear staining, as described previously.8•20 Intensity of immunostaining for
receptor expression was scored as negative (0), weak ( l+), positive (2+) or
strongly positive (3+ ). ER and PR staining was considered positive if nuclear
staining was present in more than 10% of the tumour cells (Figure 1 a and 1 b
respectively).
1 36
Figure la . Positive immunohistochemical staining for expression of
progesterone receptor alpha.
Figure 1 b . Positive immunohistochemical staining for expression of estrogen
receptor alpha.
137
lmmunostaining was scored by two independent observers (KH and VJ).
Observers had no prior knowledge of clinicopathological information. Cases
with a discrepancy in score between the two observers were re-examined with a
gynaecological pathologist (HH) until consensus was reached. Minimally two
cores with representative tumour had to be present on the TMA for a sample to
be entered into analysis.The mean overall core loss was 5.7% (4.8-6.3%).
Statistics
All analyses were performed using SPSS 14.0. Differences between groups were
tested using the X2 or the Fisher's exact test if appropriate. The ratio of ER alpha
and beta, and the ratio of PR alpha and beta were calculated using the semi
quantitative numerical scores of the staining intensities, and classified as a ratio
> 1 or < 1, as described previously .8 Survival curves were constructed according
to Kaplan-Meier. Logistic regression analyses were performed with disease free
and disease specific survival as dependent variables, and ER alpha and PR alpha
as the independent variables, with adjustment for known risk factors as lymph
vascular space invasion, myometrial invasion, grade and stage of disease entered
simultaneously into the model. Cox regression analysis was used for time to
event analyses, with recurrence or death of disease as end point. Patients were
censored at date of last contact or death in case of no relapse. P-values < 0.05
were considered statistically significant (tested 2-sided).
Results
In Table 2 the clinico-pathological characteristics of 315 patients with
endometrioid endometrial cancer are shown. The median follow-up time was
4.97 years (range 0.0 - 21.5). Follow-up data were completed until September
2006. Median age at time of diagnosis was 64. 7 years (range 32.0-89.0).
Patients were classified as FIGO stage I (59.0%), stage II ( 17. 1 %), stage III
138
( 19.4%) and stage IV (4. 1%). Hysterectomy and bilateral salpingo
oophorectomy was performed in 56.2% of the patients, and more extended
surgery in 4 1 .9% of the patients. Adjuvant post-operative radiotherapy was
given to 168 patients (53.3%). In 65 patients (20.6%) recurrent disease was
diagnosed. Local relapse at the vaginal vault occured in 26 patients, while in 8
and 31 patients regional and distant metastases were diagnosed, respectively.
The median time to relapse was 1 . 7 years (range: 0. 1 - 7 .6). During follow-up
82 out of 3 15 patients died (26.0%), 38 patients ( 12. 1 %) died from endometrial
cancer.
Association of ER alpha and beta, PR alpha and beta with h istopathological
factors
ER alpha and beta expression was observed in 185 (58.7%) and 23 1 (73.3%)
tumours respectively, in addition to PR alpha and beta expression in 1 16 (36.8%)
and 12 1 (38.4%) tumours respectively. Co-expression of ER and PR alpha and
beta in patients with endometrioid endometrial cancer is depicted in Table 3. In
87% of ER alpha positive tumours expression of ER beta was seen, while ER
alpha expression was found in 70% of ER beta positive tumours. Only half of
ER alpha and beta positive tumours expressed PR alpha and beta. Positive PR
alpha and beta expression were both combined with expression of ER alpha and
beta in 86% and 88%, respectively. PR alpha was seen together with PR beta
expression in 86%, while PR beta expression went together with PR alpha
expression in 83% of cases. In Table 4 receptor expression is shown in relation
to traditional prognostic factors such as stage, grade and lymphvascular space
invasion. ER alpha was found more often in early stage endometrial cancer
(univariate analysis p=0.020). Next, expression of ER alpha, PR alpha and PR
beta were related to low tumour grade (p<0.000 1 , p<0.001 and p=0.001
respectively). In the presence of lymph vascular space involvement the tumours
were more likely to have no ER alpha expression (p=0.0 19, OR 2.2, 95% CI
139
1 . 14-4.26) . High grade tumours were more likely to be negative for ER alpha
and PR alpha expression (p<0.005, OR 6.0, 95% CI 2.94- 12.31 and p<0.005,
OR 8.4, 95% CI 3.08-22.84 respectively) (Table 5 and 6).
Table 2 . Histological and clinical characteristics of patients with endometrioid
endometrial cancer (n=315).
Patient age, years Median 64.7 65.4 Range 32-89 35-86
Tumour Stage I 1 86 59.0 1 86 77.5 II 54 1 7. 1 54 22.5 III 6 1 1 9.4 IV 1 3 4. 1 Missing 1 0.3
Histological grade Grade 1 1 62 5 1 .4 1 39 57.9 Grade 2 96 30.5 7 1 29.6 Grade 3 56 1 7.8 29 12 . 1 Undifferentiated 1 0.3 1 0.4
Myometrial invasion < 50% 1 85 58.7 1 60 66.7 > 50% 126 40.0 78 32.5 Missing 4 1 .3 2 0.8
L VSI Negative 222 70.5 1 88 78.3 Positive 73 23.2 38 1 5 .8 Missing 20 6.3 14 5.8
Recurrence No recurrence 250 79.4 204 85.0 Local 26 8.3 20 8.3 Regional 8 2.5 6 2.5 Distant 3 1 9.8 1 0 4.2
Pelvic lymph nodes Negative 97 30.8 78 32.5 Positive 29 9.2 No sampling 1 89 60.0 1 62 67.5
Paraaortic lymph nodes Negative 45 14.3 27 1 1 .3 Positive 1 2 3.8 No sam ling 258 8 1 .9 2 1 3 88.8
LVSI: Lymph Vascular Space Involvement
140
Table 3. Co-expression of the receptors m patients with endometrioid endometrial cancer.
ER alpha ER beta PR alpha PR beta positive positive positive positive (n = 1 85) (n = 231 ) (n = 1 1 6) (n = 1 21 )
ER alpha X 1 6 1 (69.7%) 1 00 (86.2%) 1 04 (86 .0%)
ER beta 1 6 1 (87.0%) X 1 02 (87.9%) 1 06 (87.6%)
PR alpha 1 00 (54. 1 %) 1 02 (44.2%) X 1 00 (82.6%)
PR beta 1 04 (56.2%) 1 06 (45.9%) 1 00 (86.2%) X
Table 4. Chi square analysis of receptor expression in relation to known prognostic factors in patients with endometrioid endometrial cancer.
Neg
ER alpha n= l l 2 Pos
n= l 85 p value Neg
ER beta n=66 Pos
n=23 1 p value Neg
PR alpha n= 1 84 Pos
n= l 1 6 p value Neg
PR beta n= l 77 Pos
n= l 2 1 p value
Neg = negative Pos = positive
Stage
Stage I
and I I
76
148
56
1 69
1 3 8
90
136
90
Grade LVSI
Stage III Grade l Grade 3 and LVSI LVSI
and IV and 2 undifferentiated Negative Positive
35 69 42 64 42
36 1 7 1 14 144 30 0.020 <0.000 1 <0.0001
1 0 5 8 8 49 12
61 1 84 47 1 59 60 NS NS NS
45 1 32 5 1 126 52
25 1 1 1 5 85 20 NS <0.001 NS
4 1 1 33 44 12 1 49
30 1 09 1 2 89 23 NS 0.001 NS
LVSI = Lymph Vascular Space Involvement
141
Table 5. The independent effect of stage of disease, tumour grade, myometrial
invasion and LVSI on the presence of ER alpha expression in endometrioid
endometrial cancer (logistic regression analysis) .
95.0% C . l .for EXP(B)
OR Lower Upper P value Stage of disease 0.889 0.454 1 .743 0.733 Tumour Qrade 6.01 5 2.939 1 2 .309 0.000 Myometrial invasion 0 .943 0 .523 1 .698 0 .844
LVSI 2 .206 1 . 1 42 4.262 0 .01 9
LVSI = Lymph Vascular Space Involvement
Table 6. The independent effect of stage of disease, tumour grade, myometrial
invasion and L VSI on the presence of PR alpha expression in endometrioid
endometrial cancer (logistic regression analysis) .
95.0% C. l .for EXP(B)
OR Lower Upper P value Tumour grade 8.384 3.078 22.839 0.000 Stage of disease 0.700 0.357 1 .372 0.299 Myometrial invasion 0.867 0 .493 1 .526 0.621
LVSI 1 .297 0.654 2 .573 0.457
LVSI = Lymph Vascular Space Involvement
1 42
Figure 2a. Kaplan-Meier survival analysis concerning ER alpha expression and
cumulative disease-specific survival (p=0.0053) . .
1 , 0
0 ,8
.;!!: 0 ,6
::s tJ)
a 0 ,4
0 ,2
0 ,0
0 ,0 2 ] 4D 6] 8 D
Follow-up in years
1 0 ,0
ER a lfa exp ression ERA expression
NO ERA expression
143
Prognostic factors in endometrioid endometrial cancer
Univariate analysis using Kaplan-Meier survival analysis indicates that ER
alpha positive tumours had a better disease-specific survival and presence of PR
alpha reduces risk of recurrence (Figure 2a and 2b) (p=0.0053 and p=0.0335
respectively). Using cox regression analysis, high tumour grade (p=0.0 14, OR
2.4, 95% CI 1 . 19-4. 72) and absence of PR alpha expression (p=0.015, OR 4.2,
95% CI 1 .32- 13. 33) were independent prognostic factors for relapse of
endometrioid endometrial cancer (Table 7). Stage of disease was the strongest
independent predictor for survival (p=0.0 14, OR 2. 7, 95% CI 1 . 23-5.88) (Table
8).
Figure 2b. Kaplan-Meier survival analysis concerning PR alpha expression and
cumulative recurrence (p=0.0335).
1 ,0
0,8
C Q) t 0,6 :J
! E :J 0,4
0,2
0,0
0,0 1 ,0 2 , 0 3,0 4,0 5,0
timetill recurrencein years
PRalfa expession
NO PRA expesson
PRA
expesson
144
Table 7. The risk of relapse of endometrioid endometrial cancer according to the
presence of ER and PR alpha and beta, stage of disease, tumour grade,
myometrial invasion and L VSI ( cox regression analysis).
95.0% Cl for OR Exp(B) P value
Lower Upper
ER alpha 0.855 0.376 1 .943 0.708
ER beta 1 .076 0.499 2 .322 0.852
PR alpha 4. 1 98 1 .322 1 3 .333 0.01 5
PR beta 0.542 0. 1 82 1 .6 12 0.271 Stage of disease 0.859 0.483 1 .528 0.605 Tumour grade 2 .372 1 . 1 93 4.720 0.0 14 Myometrial invasion 1 .438 0.81 0 2 .553 0.2 1 5
LVSI 0 .91 4 0.489 1 .707 0.778
LVSI = Lymph Vascular Space Involvement
145
Table 8. The risk of death due to endometrioid endometrial cancer according to
the presence of ER and PR alpha and beta, stage of disease, tumour grade,
myometrial invasion and L VSI ( cox regression analysis).
95.0% Cl for OR Exp(B) P value
Lower Upper
ER alpha 2.079 0.670 6 .453 0.205
ER beta 1 .482 0.401 5.475 0.555
PR alpha 0.967 0. 1 75 5.362 0.970
PR beta 0.789 0. 1 87 3 .326 0.746 Stage of disease 2.685 1 .226 5.880 0.014 Tumour g rade 0.961 0.403 2.291 0.982 Myometrial i nvasion 2.009 0.851 4.747 0 . 1 1 2
LVS I 1 .093 0.470 2.544 0.836
LVSI = Lymph Vascular Space Involvement
Relevance of ER alpha and ER beta, PR alpha and PR beta in early stage
endometrial cancer
In patients with stage I and II (n=240) ER alpha expression was observed in 148
(6 1.7%) tumours, while ER beta expression was found in 169 cases (70.4%).
Both, PR alpha and beta expression was demonstrated in 90 patients (3 7 .5% ). In
patients with early stage disease local relapse was diagnosed in 20 patients, and
regional and distant metastases were diagnosed in 6 and 10 patients, respectively
(Table 2). In this early stage group, 14 patients (5.8%) died from endometrial
146
cancer. In a multivariate analysis absence of ER alpha was independently related
to death of disease (p=0.017, OR 7.28, 95% CI 1.42-37.25).
Prognostic value of ER alpha I beta and PR alpha I beta ratios
The ratio of ER alpha / ER beta <I was univariately related to a shorter disease
free survival, using Kaplan-Meier analysis (p=0.027). Patients with a ratio of
PR alpha / PR beta < I more often had a shorter disease free survival (univariate
analysis p=0.044) or died from endometrial cancer (p=0.005).
Discussion
We found ER alpha expression to be decreased in poorly differentiated tumours
and high stage of disease, and to be correlated with survival. Interestingly, in a
subgroup analysis for early stage of disease, absence of ER alpha appeared an
independent prognostic factor for death of disease. Our results suggest that
adjuvant or more extensive therapy may be considered in ER alpha negative
patients with early stage of disease, as these patients can be classified as being at
risk for death of disease. Recently adjuvant chemotherapy was suggested to
improve outcome in patients with high-risk early stage endometrial cancer.4
Immunohistological analysis of ER has been used previously to predict overall
prognosis of patients with endometrial cancer, but data on ER alpha and beta are
scarce. 7•8•2 1 •22 ER alpha consists of a single polypeptide chain of 595 amino
acids, and ER beta of 485 amino acids. ER beta has a similar binding affinity for
estrogens as ER alpha, but selective modes of estrogenic action in different types
of tissue may exist. 23 In contrast to our results, it has been suggested that ER beta
is important in the progession of myometrial invasion. 8•2 1 •24
Estrogen dependent carcinogenesis may depend on the ratios of the two ER
subtypes, as disruption of the balance between the isoforms ER alpha and beta is
likely related to different modes of estrogenic action. In literature ER alpha / ER
147
beta ratios decreased from normal to malignant endometrial tissue. 7•8•22•24 We
observed that the ratio of ER alpha / ER beta <1 appeared to be a predictor of a
shorter disease free survival, which was not found in a small study, using the
same semi-quantitative ratio.8 So status of both ER alpha and beta may be a
prognostic factor in endometrial cancer, as disturbances of the synchronized
expression of ER alpha and beta may play a role in the alteration of estrogen
action. As we did not demonstrate any correlation between ER beta and clinico
histopathological characteristics, determination of the expression of ER beta
seems only useful for prognostic purposes if analysed together with ER alpha
expression as a ratio.25
Several studies reported a correlation between PR expression and disease-free
survival, but data on the relative expression of PR alpha and beta, as reported in
this study, and their involvement in the pathogenesis of endometrial cancer are
scarce.9•11•2 1 We found the absence of PR alpha expression to be a powerful
independent prognostic factor for relapse of disease in endometrioid endometrial
cancer. This is contradicted by a recent study reporting decrease of PR beta to
have independent prognostic value, but the power of that study may be biased by
heterogeneous tumour inclusion as only 73% of analysed tumour cases
concerned endometrioid adenocarcinoma. 25 The availability of PR is an essential
factor determining progesterone response. Progesterone decreases the risk of
developing estrogen-dependent endometrial cancer by controlling cell
proliferation, as it is the natural antagonist of estrogenic hormones. PR is
expressed by a single gene encoding two proteins (PR alpha and PR beta). PR
beta consists of 933 amino acids, and PR alpha is a truncated form of PR beta
lacking 164 amino acids from the N-terminus of PR beta.9 PR alpha was found
to inhibit PR beta function and ER activation, while PR beta is involved in
regulation of endometrial responsiveness. 1 1 It was hypothesized that decreased
PR levels observed in endometrial cancer in fact arise from the loss of one PR
148
isoform, PR alpha or PR beta, as express10n of a single PR isoform was
associated with higher grade of tumour.9•1 1
As we found both PR alpha and beta to be related to low grade of tumour, the
loss of PR alpha and beta possibly contributes to tumour progression. We found
the ratio of PR alpha / PR beta <1 to be related to relapse and death of disease.
These results need to be confirmed by others to determine the strength of the PR
alpha / PR beta ratio as a prognostic factor. So the cellular ratio of PR alpha / PR
beta possibly determines progesterone action and may be involved in the
regulation of endometrial reaction.
In conclusion, together with histological grade, the absence of PR alpha is an
independent prognostic factor for recurrence of disease, while patients with ER
alpha positive tumours have a better survival, especially in the early stages. The
data of our large study population support the general idea of the prognostic
significance of estrogen and progesterone receptors, in addition to the classical
clinico-histopathological factors. These findings suggest that the
immunohistochemical evaluation of steroid receptors contributes to predicting
the biological behaviour of the tumour, next to the classical parameters. This
could lead to a more patient specific risk profile and therefore improve patient
tailored treatment.
Acknowledgements
HW Nijman is supported by the Dutch Cancer Society (Grant 2002-2768).
149
References
1 . Sherman ME. Theories of endometrial carcinogenis: a multidisciplinary approach.
Modem Pathol2000; 1 3 :295-308 . 2. Creutzberg CL, van Putten WL, Koper PC, et al. Surgery and postoperative
radiotherapy versus surgery alone for patients with stage- I endometrial carcinoma:
multicentre randomised trial. PORTEC Study Group. Postoperative radiation therapy
in endometrial carcinoma. Lancet 2000;355 : 1404- 1 1 . 3 . Keys HM, Roberts JA, Brunetto VL. A phase III trial of surgery with or without
adjunctive external pelvic radiation therapy in intermediate risk endometrial
carcinoma: a Gynecologic Oncology Group Study. Gynecol Oncol 2004;92:744-5 1. 4. Kodama J, Seki N, Hiramatsu Y. Chemotherapy for high-risk early-stage endometrial
cancer. Curr Opin Obstet Gynecol. 2007; 19(1) :42-7.
5. Gehrig PA, Van Le L, Olatidoye B, Geradts. Estrogen receptor status, determined by immunohistochemistry, as a predictor of the recurrence of stage I endometrial carcinoma. J Cancer 1999;86:2083-9.
6. Jeon YT, Part IA, Kim YB, Kim JW, Park NH, Kang SB. Steroid receptor expressions in endometrial cancer: clinical significance and epidemiological implication. Cancer Lett 2006;239: 198-204.
7. Fujimoto J, Sakaguchi H, Aoki I, Toyoki H, Tamaya T. Clinical implications of the
expression of estrogen-alpha and -beta in primary and metastatic lesions of uterine endometrial cancers. Oncol 2002;62:269-77.
8 . Mylonas I, Jeschke U, Shabani N, Kuhn C, Kriegel S, Kupka MS, et al. Normal and
malignant human endometrium express immunohistochemically estrogen receptor alpha (ER-alpha), estrogen receptor beta (ER-beta) and progestrone receptor (PR).
Anticancer Res 2005;25 : 1679-86. 9. Miyamoto T, Watanabe J, Hata H, Jobo T, Kawaguchi M, Hattori M, et al.
Significance of progesterone receptor-A and -B expressions in endometrial
adenocarcinoma. J Steroid Biochem Molec Biol 2004;92: 1 1 1 -8. 10. Hanekamp EE, Gielen SCJP, Smid-Koopman E, Kuhne LCM, de Ruiter PE, Chadha
Ajwani S, et al. Consequences of loss of progesterone receptor expression in
development of invasive endometrial cancer. Clin Cancer Res 2003;9 :4 190-9.
150
1 1. Arnett-Mansfield RL, deFazio A, Wain GV, Jaworski RC, Byth K, Mote PA, Clarke CL. Relative expression of progesterone receptors A and B in endometrioid cancers of
the endometrium. Cancer Res 2001;61 :4576-82. 12. Hirai M, Hirano M, Oosaki T, Hayashi Y, Yoshihara T, Matsuzaki 0. Pognostic factors
relating to survival in uterine endometrioid carcinoma. Int J Gynaecol Obstet
1999;66: 155-62. 13 . Steiner E, Eicher 0, Sagemuller J, Schmidt M, Pilch H, Tanner B, et al. Multivariate
independent prognostic factors in endometrial carcinoma: a clinicopathologic study in
18 1 patients : 10 years experience at the Department of Obstetrics and Gynecology of
the Mainz University. Int J Gynecologic Cancer 2003 ; 13 : 197-203 .
14. Oreskovic S, Babic D, Kalafatic D, Barisic D, Beketic-Oreskovic L. A significance of
irnmunohistochemical determination of steroid receptors, cell proliferation factor Ki-
67 and protein p53 in endometrial carcinoma. Gynecol Oncol 2004;93 :34-40.
15. Briet JM, Hollema H, Reesink N, Aalders JG, Maurits MJE, ten Hoor KA, et al.
Lymphvascular space involvment: an independent prognostic factor in endometrial cancer. Gyn Oncol 2005;96:799-804.
16. Hoos A, Cordon-Cardo C. Tissue microarray proofiling of cancer specimens and celllines: opportunities and limitations. Lab Invest 2001 ;8 1 : 133 1-8.
17. Hendriks Y, Franken P, Diersen JW, de Leeuw W, Wijnen J, Dreef E, et al.
Conventional and tissue microarray irnmunohistochemical expression analysis of
mismatch repair in hereditary colorectal tumors. Am J Pathol 2003 ; 162:469-77.
18. Kallioniemi OP, Wagner U, Kononen J, Sauter G. Tissue microarray technology for
high-throughput molecular profiling of cancer. Hum Mol Genet 2001; 10:657-62. 19. Nijhuis ER, Nijman HW, Oien KA, Bell A, ten Hoor KA, Reesink-PetersN, et al. Loss
of MSH2 protein expression is a risk factor in early stage cervical cancer. J Clin Pathol 2007;60:824-30.
20. Leitao MM, Soslow RA, Nonaka D, Olshen AB, Aghajanian C, Sabbatini P, et al.
Tissue microarray immunohistochemical expression of estrogen, progesterone, and
androgen receptors in uterine leiomyomata and leiomyosarcoma. Cancer 2004; 101 : 1455-62.
21 . Utsunomiya H, Suzuki T, Harada N, Ito K, Matsuzaki S. Analysis of estrogen receptor alpha and beta in endometrial carcinomas: correlation with ERbeta and
clinicopathologic findings in 45 cases. Int J Gynecol Pathol 2000; 19 :335-4 1.
15 1
22. Hu K, Zhong G, He F. Expression of estrogen receptors ER alpha and ER beta in endometrial hyperplasia and adenocarcinoma. In J Gynecol Cancer 2005 ; 15 :537-4 1 .
23 . Kuiper GG, Carlsson B, Grandien K, Enmark E, Haggblad J, Nilsson S, et al.
Comparison of the ligand binding specificity and transcript tissue distribution of estrogen receptors and . Endocrinol 1997; 138 :863-70.
24. Takama F, Kanuma T, Wang D, Kagami I, Mizunuma H. Oestrogen receptor beta
expression and depth of myometrial invasion in human endometrial cancer. Br J Cancer 200 I ;84:545-9.
25 . Shabani N, Kuhn C, Kunze S, Schulze S, Mayr D, Dian D, et al. Prognostic
significance of oestrogen receptor alpha (ERalpha) and beta (ERbeta), progesterone receptor A (PR-A) and B (PR-B) in endometrial carcinomas. Eur J Cancer 2007 ;43 :2434-44.
152
Summary and future perspectives
153
154
Summary and future perspectives
As outlined in Chapter 1 the aims of the studies described in the first part of this
thesis was to explore whether endogenously ( ovary) produced androgens after
conversion to estrogens by aromatase activity of the endometrium may contribute
to endometrial carcinogenesis. This hypothesis was examined (i) by evaluation of
the relation between androgens present in the utero-ovarian circulation and the
degree of ovarian stromal hyperplasia, and (ii) by demonstration of aromatase
expression in endometrial tissue.
In the second part of the thesis, we assessed the expression and prognostic value of
(i) estrogen and progesterone receptor alpha and beta, next to (ii) aromatase, COX-
2, HER-2/neu and p53 in a large cohort of endometrioid endometrial cancer
patients. Using tissue microarray analysis, the expression of the hormonal and
molecular markers was related to classical histopathological factors and disease
free and overall survival.
In Chapter 2 an introduction is given on the hormonal interactions contributing to
the development of endometrioid endometrial carcinoma. After the menopause
estrogens mainly originate from peripheral conversion of androgens by
aromatization in body fat. Hyperplasia of ovarian stroma may be associated with
an increased androgen production by the ovaries. In case of stromal hyperplasia
of the ovaries, increased production of aromatizable androgens by
postmenopausal ovaries may contribute to increased availability of androgen
precursors for intratumoural estrogen synthesis in the endometrial tissue as well.
Following the aromatase driven conversion of ovarian androgens, estrogens may
contribute in situ to the development of endometrial cancer. The local availability
of androgens and the local activity of aromatase in endometrial tissue are
considered relevant for this process.
155
Chapter 3 reviews available evidence that androgens, produced in hyperplastic
ovarian stroma or body fat tissues, play a role in the development of endometrial
cancer through conversion into estrogens, a reaction catalyzed in the
endometrium by the enzyme aromatase cytochrome P450. As the presence of
aromatase appeared to be a pathophysiological factor in the development of
hormone dependent breast cancer, the latter is evaluated in endometrioid
endometrial cancer as well. As treatment with aromatase inhibitors appeared
feasible in breast cancer, current knowledge of comparable treatment modalities
in hormone dependent endometrial cancer is also reviewed. Treatment of
endometrial cancer with aromatase inhibitors appears to be useful to some
extent, but available evidence in the literature is limited, especially in advanced
and recurrent endometrial cancer, where new treatment modalities are urgently
needed.
In chapter 4 two separate studies report on the relationship between the presence
of endometrioid endometrial cancer, the degree of ovarian stromal hyperplasia and
ovarian steroid production in postmenopausal women. In 1 1 2 patients with
endometrioid endometrial cancer, 4 7 patients with a benign gynaecological
condition and 10 patients with non-endometrioidendometrial cancer, the degree of
ovarian stromal hyperplasia was scored retrospectively on a semi-quantitative
scale (atrophy, slight, marked) . All patients were postmenopausal and had
undergone a hysterectomy with bilateral salpingo-oophorectomy. Prospectively
also blood sampling from the ovarian veins was performed in 60 patients. Steroid
levels ( estrone, estradiol, androstenedione, testosterone) were determined and
related to the degree of ovarian stromal hyperplasia and the presence (n=52) or
absence (n=8) of endometrioid endometrial cancer. In the retrospective study the
degree of ovarian stromal hyperplasia was higher in the presence of endometrioid
endometrial cancer (p=0.000). Our prospective study showed that an increasing
1 56
degree of ovarian stromal hyperplasia was related to higher ovarian levels of both
testosterone and androstenedione (p<0.05 and p<0.005 respectively), but not to
estrone or estradiol. Mean ovarian vein levels of both testosterone and
androstenedione were higher in patients with endometrial cancer than in patients
with benign conditions, without reaching statistical significance. In conclusion,
higher degrees of ovarian stromal hyperplasia were found in endometrioid
endometrial cancer and with increasing degrees of ovarian stromal hyperplasia
levels of ovarian vein androgens were higher. A causal relationship in the origin
of hormone-dependent endometrial pathology may exist between ovarian stromal
hyperplasia, ovarian vein androgen levels and endometrioid endometrial
carcinoma.
In chapter 5 prospectively the relationship between androgen levels in the utero
ovarian circulation, aromatase activity in endometrial and body fat tissue, and the
presence or absence of endometrioid endometrial cancer was assessed in a
prospective study. In 43 patients with endometrioid endometrial cancer and 8
patients with a benign gynaecological condition after menopause a hysterectomy
with bilateral salpingo-oophorectomywas performed. Using tritium water release
assays, aromatase activity in endometrial and body fat tissue was determined and
related to the steroid levels from the peripheral and the utero-ovarian venous
circulation ( estradiol, androstenedione, testosterone) and to the presence or
absence of endometrial cancer. Significant aromatase activity was found in both
benign and malignant endometrial tissue samples. Aromatase activity in samples
of endometrial tissue and in samples of body fat did not correlate with steroid
levels in peripheral or utero-ovarian venous blood. Aromatase activity in samples
of benign or malignant endometrium did not differ. Remarkably, in four patients
with mainly poorly differentiated endometrial cancer very high aromatase activity
was found in endometrial tissue. It is likely that multiple pathogenetic pathways
exist which eventually lead to the formation of endometrioid endometrial cancer.
157
The local availability of androgens and the finding that aromatase activity is
present in both endometrial cancer and benign endometrial tissue, support the
hypothesis that aromatase activity in the endometrium may play a role in
malignant transformation, by converting androgens into mitogenic estrogens in the
endometrial tissue.
In chapter 6 the prognostic value of aromatase, COX-2, HER-2/neu and p53
expression was determined in endometrioid endometrial cancer. Tissue
microarrays were constructed comprising samples from 315 endometrioid
endometrial cancer patients. Expression of aromatase, COX-2, HER-2/neu and
p53 was determined by immunostaining and related to classical clinico
histopathological parameters, in addition to recurrence of disease and survival.
Median follow-up time for all patients was 5.0 years. Patients were classified as
FIGO stage I (59.0%), stage II ( 17. 1%), stage III ( 19.4%) and stage IV (4.1%).
Sixty-five patients (20.6%) developed recurrent disease and 38 ( 12. 1 %) died
because of endometrial cancer. Aromatase, COX-2, HER-2/neu and p53
expression was observed in 133 (42.2%), 107 (34.0%), 17 (5.4%), and 21
(6.7%) tumour cases, respectively. Aromatase expression in tumour cells was
related to aromatase expression in stromal cells (p<0.0001) and to HER-2/neu
expression in tumour cells (p=0.0 19). Aromatase expression in both tumour as
well as stromal cells was related to low stage of disease (p=0.02 and p=0.00 1,
respectively), while aromatase expression in stromal cells was also related to
low tumour grade (p=0.021). P53 expression was related to high stage and high
grade (p=0.006 and p<0.000 1, respectively). In multivariate analysis p53
overexpression was independently related to death due to the disease (p=0.043,
OR 3.0, 95% CI 1.0-8.7). Although rarely found, presence of p53 expression
leads to a worse prognosis of endometrioid endometrial cancer. From literature it
is known that p53 is more frequently found in non-endometrioid tumours. As
patients with non-endometrioid endometrial cancer generally have a worse
158
prognosis than patients with endometrioid tumours, it seems that p53 positive
endometrioid tumours have a biologic behaviour more resembling non
endometrioid tumours. For COX-2, HER-2/neu and aromatase no relation with
any other histopathological parameter or survival was found. In conclusion,
aromatase expression in tumour and stromal cells and p53 expression in tumour
cells are related to tumour grade and stage of disease, while p53 is an
independent adverse prognostic factor in endometrioid endometrial cancer.
The frequent presence of aromatase in endometrioid endometrial cancer ( 42% of
tumour cases ! ) supports the hypothesis that aromatase activity in the endometrial
tissue may play a role in the development of endometrioid endometrial cancer.
With regard to the aims of the study, it must be noted that aromatase expression
was found predominantly in lower stage and lower grade of disease. Possibly its
expression therefore disappears with increasing stage and dedifferentiation. In
breast cancer a relationship was found between the COX-2, HER-2/neu and
aromatase enzyme systems, as interaction between these biomarkers was found
to play a role in the production of estrogens. In breast cancer HER-2/neu and
aromatase overexpression were related, with HER-2/neu as the primary
molecular determinant of increased aromatase activity. COX-2 is the functional
intermediate with prostaglandin E2, which regulates aromatase gene expression,
as its product. We found a correlation between HER-2/neu and aromatase, which
may suggest a similar correlation between these enzyme systems in endometrial
cancer. Our studies suggest a role for aromatase in the early stages and grades of
disease. It can not be excluded that, apart from local effects on the endometrial
tissue, a positive effect of the use of aromatase inhibitors on peripheral fat tissue
exists. Theoretically this estrogen decreasing effect in body fat will not
contribute much to a better prognosis of the higher stages and grades, as
estrogen receptor expression decreases in the higher stages and grades of disease
( chapter 7). So possibly some women develop endometrioid endometrial cancer
1 59
through continuous hyperstimulation of the endometrial tissue by estrogens
produced by peripheral conversion in body fat, while in other women malignant
transformation is enhanced by local conversion of ovarian androgens into
estrogens, catalysed by (perhaps overexpression of) endometrial stromal
aromatase cytochrome P450, or both phenomena may happen simultaneously.
In chapter 7 the prognostic impact of estrogen receptor (ER) and progesterone
receptor (PR) alpha and beta was determined in endometrioid endometrial
cancer. Tissue microarrays were constructed from 315 endometrioid endometrial
cancer patients. ER alpha and beta and PR alpha and beta were semi
quantitatively assessed by immunostaining, and together with their alpha / beta
ratios related to classical clinico-histopathological parameters in addition to
disease free and disease specific survival. Patients were classified as FIGO stage
I (59.0%), stage II (17.1%), stage III (19.4%) and stage IV (4.1%). Sixty-five
patients (20.6%) developed recurrent disease and 38 (12.1 %) died due to
endometrial cancer. In univariate analysis, expression of ER alpha was related to
early stage endometrial cancer (p=0.020), while expression of ER alpha, PR
alpha and PR beta was associated with lower grade tumours (p<0.0001, p<0.001
and p=0.001 respectively). A ratio of ER alpha / ER beta <1 was related to a
shorter disease free survival (p=0.027), while the ratio of PR alpha / PR beta <1
both was associated with a shorter disease free survival as well as a shorter
overall survival (p=0.044 and p=0.005, respectively). These findings illustrate
that the relative distribution of these biomarkers is of importance in disease
progress and prognosis. In early stage disease absence of ER alpha was
independently related to death of disease (p=0.017, OR 7.28, 95% CI 1.42-
37.25), while absence of PR alpha (p=0.015, OR 4.2, 95% CI 1.32-13.33)
appeared to be an independent prognostic factor for relapse of disease. Although
aromatase expression ( chapter 6) was not part of the study described in chapter
7, we also assessed relations between expression of aromatase and ER and PR
160
subtypes. Although expression of aromatase and ER alpha were both related to
early stage endometrial cancer, no relations between the aromatase and ER or
PR expression could be found. We conclude that in early stage endometrioid
endometrial cancer absence of PR alpha is an independent prognostic factor for
disease-free survival, while patients with ER alpha positive tumours have a
better overall survival. The data of our large study population support the
significance of estrogen and progesterone receptors for prognostic purposes, in
addition to the classical clinico-histopathological factors, and this may lead to a
more patient specific risk profile and treatment.
Future perspectives
Future research should be aimed at the molecular basis of endometrial cancer
carcinogenesis, and the exact role of aromatase. Our studies can not be
considered as true evidence for a causative role for aromatase in endometrial
cancer carcinogenesis, as we only describe statistical associations and relations.
Further evidence for the role of aromatase needs to be derived from in vitro
studies with human endometrial tissue cells. In normal, hyperplastic and
malignant endometrial cells more extensively the influences of addition or
deprivation of androgens and aromatase should be explored. In endometrial
cancer cell lines the levels of aromatase activity in relation to the effects of
steroids on DNA synthesis deserve further in depth evaluation.
Although in the treatment of breast cancer a revolution took place, when
aromatase inhibitors became available for depriving the tumour of estrogenic
growth stimulation ( see chapter 3), it is not expected that this will happen in a
same manner with endometrial cancer. Contrary to breast cancer, prognosis of
endometrioid endometrial cancer generally remains good with effective surgical
and radiotherapeutical treatment possibilities. Most likely only small numbers of
patients may benefit from aromatase inhibitors. A palliative role for aromatase
161
inhibitors especially in stage IV patients or metastatic disease should be
considered. Unfortunately until now studies did not show therapy with
aromatase inhibitors to be very effective in endometrial cancer ( chapter 3).
Future studies on the use of aromatase inhibitors in recurrent or advanced
endometrial cancer should relate objective disease response to potential
predictive markers, as described in chapter 6 and 7, including aromatase. Until
now only one clinical phase II trial on the efficacy of aromatase inhibitors
reported disappointing disease response in endometrial cancers with positive
expression of ER, PR and HER-2/neu. 1 So the potential role of aromatase
inhibitors in steroid receptor positive endometrial cancer is not yet resolved, as
theoretically estrogen receptor positive tumours can be expected to respond to
therapy. As aromatase is known to convert androgens into estrogens, it is
unlikely that the enzyme plays a significant role in hormone independent cancer
development, so research on the possible role of aromatase expression in non
endometrioid endometrial cancer does not seem useful. Indeed disappointing
response rates to aromatase inhibition were reported in a phase II trial in
patients with advanced or recurrent endometrial cancer. Only two partial
responses were found in 23 women with advanced or recurrent endometrial
malignancy, of whom most cancers were poorly differentiated or presented an
agressive histology ( clear cell carcinoma or papillary serous carcinoma). Only
39% of these patients were diagnosed with endometrioid endometrial cancer,
and women with poorly differentiated tumours did not respond. 2 One may argue
that patient selection and short duration of treatment influenced results
negatively. As aromatase may play a role in the origin of endometrial cancer,
future research should perhaps focus on the possible use of aromatase inhibitors
as prophylactic medical treatment of (atypical) endometrial hyperplasia, with the
aim of preventing a prophylactic hysterectomy eg. in young patients with
endometrial cancer or atypical hyperplasia who want to preserve fertility.
Another challenge may be found in evaluation of the endometrium in
1 62
postmenopausal breast cancer who are using aromatase inhibitors. Several
studies reported a potential protective effect of aromatase inhibitors on the
endometrium in patients previously treated with tamoxifen, but data are not
conclusive ( chapter 3). 3-5 As aromatase inhibitors are prescribed more frequently
in postmenopausal breast cancer patients according to new treatment protocols,
more research has to be done on the implications for the endometrial tissue.
Meanwhile the endometrial thickness of postmenopausal breast cancer patients,
who switched from treatment with tamoxifen to aromatase inhibitors needs to be
monitored carefully, until possible (late) effects of replacement of tamoxifen by
aromatase inhibitors are better defined.
References
1 . Ma BBY, Oza A, Eisenhauer E, Stanimir G, Carey M, Chapman W, et al. The activity
of letrozole in patients with advanced or recurrent endometrial cancer and correlation
with biological markers - a study of the National cancer Institute of Canada Clinical
Trials Group. Int J Gynecol Cancer 2004; 14 : 650-8.
2. Rose PG, Brunetto VL, VanLe L, Bell J, Walker JL, Lee RB. A phase II trial of
anastrozole in advanced recurrent or persistent endometrial carcinoma: a gynecologic
oncology group study. Gynecol Oncol 2000; 78 : 212-6.
3. Cohen I. Aromatase inhibitors and the endometrium. Maturitas (2008),
doi : 10. 1016/j.maturitas.2008.03 .001 .
4. Duffy SRG, Greenwood M. The endometrial data from the ATAC (Arimidex,
Tamoxifen, Alone or in Combination) trial indicates a protective effect of anastrozole
(arimidex) upon the endometrium. Breast Cancer Res Treat 2003; 83 : abs 134.
5. Duffy S, Jackson TL, Lansdown M, Philips K, Wells M, Pollard S, Clack G, Coibon M, Bianco AR. The ATAC ('Arimidex', Tamoxifen, Alone or in Combination)
adjuvant breast cancer trial; first results of the endometrial sub-protocol following 2
years of treatment. Hum Reprod 2006; 21(2): 545-53.
163
164
Samenvatting en overwegingen voor de toekomst
165
Samenvatting en overwegingen voor de toekom.st
Zoals uiteengezet in hoofdstuk 1 was bet doel van de in bet eerste deel van bet
proefschrift beschreven studies om te onderzoeken of door bet ovarium
geproduceerde androgenen bijdragen aan de carcinogenese van bet endometrioid
endomeriumcarcinoom, nadat deze androgenen tot oestrogenen zijn omgezet via
bet enzym aromatase. Hiertoe werd onderzocht of (i) in de utero-ovariele circulatie
aanwezige androgenen gerelateerd zijn aan de mate van ovariele stroma
hyperplasie, en of (ii) expressie van aromatase aangetoond kan worden in
endometriumweefsel. Daamaast onderzochten we in bet tweede deel van bet
proefschrift de expressie en prognostische waarde van (i) oestrogeen en
progesteron receptor alpha en beta, naast (ii) aromatase, COX-2, HER-2/neu en
p53 in een groot cohort van patienten met endometrioid endometriumcarcinoom.
Gebruikmakend van tissue microarray analyse werd de expressie van de
hormonale en moleculaire markers gerelateerd aan klassieke histopathologische
factoren en aan ziektevrije en totale overleving.
In hoofdstuk 2 wordt een introductie gegeven over de hormonale factoren en
interacties, die van belang zijn bij bet ontstaan van bet endometrioid
endometriumcarcinoom. Langdurige blootstelling van het postmenopauzale
endometrium aan oestrogenen is een van de factoren die van belang is bij bet
ontstaan van endometrioid endometriumcarcinoom. Toename van lichaamsvet en
toegenomen oestrogeenproductie in het vetweefsel bevordert de kans op
endometriumcarcinoom. In bet lichaamsvet worden androgenen door o.a. het
enzym aromatase omgezet in oestrogenen. Deze androgenen zijn oorspronkelijk
afkomstig van productie in de bijnierschors en postmenopauzale ovaria. Er lijkt
een relatie te bestaan tussen de mate van ovariele stroma hyperplasie
(schorsverdikking in de ovaria), productie van androgenen door de ovaria, en het
ontstaan van endometriumcarcinoom. Toegenomen productie van androgenen
1 66
door de ovaria zou via de utero-ovariele circulatie kunnen leiden tot een
toegenomen beschikbaarheid van androgenen in het endometrium. Indien het zo
blijkt te zijn dat in het endometrium het enzym aromatase aanwezig is, zou dit
kunnen leiden tot extra lokale omzetting van androgenen in oestrogenen, met
dientengevolge een toegenomen kans op het ontwikkelen van
endometriumcarcinoom. Wij beschouwen de aanwezigheid van androgenen en
aromatase activiteit in het endometrium als onmisbare voorwaarden voor het
mogelijk maken van dit proces. Wij tonen in het vervolg aan dat androgenen
aanwezig zijn in het hormonale milieu van het endometrium, in samenhang met
ovariele stroma hyperplasie, en dat aromatase inderdaad lokaal aanwezig is in
het endometrium. Dan zijn de noodzakelijke ingredienten aanwezig zijn om een
lokale route, leidend tot endometrioid endometriumcarcinoom, mogelijk te
maken. Dit is van belang voor de gedachtenvorming met betrekking tot het
ontstaan van endometriumcarcinoom. Ook heeft dit wellicht praktische
consequenties, daar een vergelijkbaar pathofysiologisch concept bij
mammacarcinoom aanwezig bleek, hetgeen leidde tot succesvolle behandeling
van hormoonafhankelijke borstkanker met aromataseremmers.
In hoofdstuk 3 wordt de stelling besproken dat androgenen, geproduceerd in
ovaria en lichaamsvet, bijdragen aan het ontstaan van endometrioid
endometriumcarcinoom doordat deze androgenen lokaal warden omgezet in
oestrogenen. Het enzym aromatase cytochrome P450 in het endometrium is
hiervoor verantwoordelijk. Daar aangetoond is dat het enzym aromatase van
belang is bij het ontstaan van mammacarcinoom, wordt in dit hoofstuk dieper
ingegaan op overeenkomsten en verschillen tussen mamma- en
endometriumcarcinoom. Omdat behandeling met aromataseremmers haalbaar en
zinvol bleek bij mammacarcinoom, is de "state of the art" van de
endocrinologische behandeling van endometriumcarcinoom geevalueerd. Tot op
zekere hoogte werden aromataseremmers met matig succes gebruikt bij de
167
behandeling van endometriumcarcinoom, echter het wetenschappelijke bewijs
hiervoor is beperkt.
In hoofstuk 4 worden twee studies beschreven. In een archiefstudie werd een
microscopische vergelijking gemaakt van ovariumweefsel van 169 patienten met
endometrioid endometriumcarcinoom, en benigne aandoeningen. De mate van
ovariele stroma hyperplasie bleek toegenomen m geval van
endometriumcarcinoom (p=0.000). In een tweede (prospectieve) studie werd de
relatie bestudeerd tussen hormoonproductie door de postmenopauzale ovaria en
endometrioid endometriumcarcinoom. Bij zestig postmenopauzale vrouwen
werden uterus en ovaria verwijderd wegens endometriumcarcinoom of een
benigne aandoening. Tijdens de procedure werd bloed afgenomen uit de utero
ovariele circulatie. Spiegels van androgenen en oestrogenen uit het afgenomen
bloed werden gerelateerd aan de mate van ovariele stroma hyperplasie, en aan
de aanwezigheid (52 maal) of afwezigheid (8 maal) van endometriumcarcinoom.
Het bleek dat bij toename van de mate van ovariele stroma hyperplasie ook de
spiegels van testosteron en androsteendion toenamen (p<0.05 and p<0.005),
maar niet de spiegels van oestrogenen. Ook bleken de spiegels van de
androgenen hoger bij de vrouwen met endometriumcarcinoom, dan bij de
vrouwen met een benigne aandoening, zonder statistische significantie te
bereiken. Wij concludeerden <lat bij patienten met endometrioid
endometriumcarcinoom de mate van ovariele stroma hyperplasie hoger was.
Ook vonden wij bij toename van de mate van ovariele stroma hyperplasie
hogere spiegels van androgenen in de utero-ovariele circulatie. Deze
bevindingen suggereren een verband tussen ovariele stroma hyperplasie,
ovariele androgeenproductie en endometrioid endometriumcarcinoom.
In hoofstuk 5 wordt een prospectieve studie gepresenteerd, waarin de relatie
wordt onderzocht tussen de androgeen- en oestrogeenspiegels in de utero-
168
ovariele circulatie, en aromatase expressie m lichaamsvet en
endometriumweefsel bij patienten met en zonder endometriumcarcinoom. Bij 5 1
postmenopauzale vrouwen werd een abdominale hysterectomie met bilaterale
adnexectomie verricht. Gebruikmakend van een tritium water release assay werd
de aromatase expressie bepaald in endometriumweefsel, en in lichaamsvet.
Aromatase expressie werd gerelateerd aan androgeen- en oestrogeenspiegels in
de utero-ovariele circulatie, en aan de aan- of afwezigheid van
endometriumcarcinoom. Significante aromatase expressie bleek aanwezig in
zowel benigne als maligne endometriumweefsel. De mate van aromatase
expressie in endometriumweefsel en/of lichaamsvet correleerde niet met
steroidspiegels in de utero-ovariele circulatie. Ook bleek de mate van aromatase
expressie niet significant te verschillen tussen benigne of maligne
endometriumweefsel. Toch ondersteunt de lokale beschikbaarheid van
androgenen in de utero-ovariele circulatie, en de waarneming dat aromatase
expressie in het endometrium aanwezig is, de volgende hypothese; aromatase
activiteit lokaal in het endometrium speelt een rol bij het ontstaan van
endometrioid endometriumcarcinoom, doordat aromatase ter plekke in staat is
androgenen om te zetten in oestrogenen.
In hoofdstuk 6 wordt de prognostische waarde onderzocht van de expressie van
de moleculaire markers COX-2, HER-2/neu, p53 en aromatase bij het
endometrioid endometriumcarcinoom. Met behulp van tissue microarray analyse
werd een cohort van 315 chirurgisch behandelde patienten met endometrioid
endometriumcarcinoom onderzocht. Tissue microarray analyse is een nieuwe
technologie, die het mogelijk maakt tumorweefsel van vele patienten tegelijk te
beoordelen onder de microscoop. lmmunohistochemische kleuringen werden
semi-kwantitatief gescoord, gebaseerd op de intensiteit van nucleaire kleuring.
Ten aanzien van de aromatase antilichamen, moet warden opgemerkt dat in de
literatuur de specificiteit van de antilichamen in het verleden vaak niet goed kon
169
warden gekarakteriseerd, met onbetrouwbare resultaten als gevolg. Daarom
gebruikten wij monoclonale 677 aromatase antilichamen, die speciaal
ontwm-pen en gevalideerd zijn voor routinematige bepaling van aromatase
expressie in archief-materiaal. Onze studie is de eerste die deze antilichamen
gebruikt bij endometriumweefsel. Expressie van de moleculaire markers werd
gerelateerd aan klassieke klinische en histopathologische karakteristieken, naast
ziektevrije en algehele overleving. De mediane follow-up voor alle patienten
bedroeg 5.0 jaar. De patienten werden geklassificeerd als FIGO stadium I
(59.0%), stadium II ( 1 7. 1%), stadium III ( 19.4%) en stadium IV (4. 1%).
Vijfenzestig patienten ontwikkelden een recidief (20.6%) en 38 ( 12. 1 %) stierven
ten gevolge van endometriumcarcinoom tijdens de follow-up periode. Expressie
van aromatase, COX-2, HER-2/neu en p53 werd waargenomen bij 1 33 (42.2%),
107 (34.0%), 1 7 (5.4%), en 2 1 (6.7%) van de tumoren. Aromatase expressie in
tumorcellen bleek gerelateerd aan aromatase expressie in stromacellen
(p<0.0001), en aan HER-2/neu expressie (p=0.0 19). Expressie van aromatase in
tumor en stromacellen was gerelateerd aan een lager stadium
endometriumcarcinoom (p=0.02 en p=0.00 1 ), terwijl aromatase expressie in
stromacellen ook gecorreleerd bleek aan een lagere tumorgraad (p=0.021 ). Wij
concludeerden dat p53, hoewel slechts aanwezig bij 7% van de patienten, een
krachtige onafhankelijke prognostische factor bleek. Expressie van p53 bleek
gerelateerd aan hogere graad en stadium (p=0.006 en p<0.0001 ), met een
toegenomen kans op overlijden ten gevolge van de aandoening (p=0.043, OR
3.0, 95% CI 1 .0-8.7). Uit de literatuur is bekend dat p53 vaker voorkomt bij het
(prognostisch ongunstiger) non-endometrioid endometriumcarcinoom. Daar
p53 positieve tumoren een slechte prognose bleken te hebben bij endometrioid
endometriumcarcinoom, lijkt hun biologische gedrag meer op dat van non
endometrioid endometriumcarcinoom. COX-2, HER-2/neu en aromatase bleken
geen prognostische betekenis te hebben. Wij concludeerden dat aromatase
expressie in tumor- en stromacellen, en expressie van p53 in tumorcellen
170
gerelateerd blijken aan de graad en het stadium van het endometrioid
endometriumcarcinoom, terwijl express1e van p53 een onathankelijke
prognostische voorspeller is.
De frequente aanwezigheid van aromatase expressie (42%!) ondersteunt de
hypothese dat aromatase activiteit een rol speelt bij het ontstaan van
endometrioid endometriumcarcinoom. Opvallend genoeg werd aromatase
expressie voomamelijk gevonden bij lager stadium en graad van het
endometriumcarcinoom. W ellicht verdwijnt de aromatase expessie bij progressie
van de maligniteit en toenemende dedifferentiatie. Bij mammacarcinoom werd
in de literatuur een relatie gevonden tussen COX-2, HER-2/neu en aromatase
expressie, daar interactie tussen deze biomarkers een rol speelt bij
oestrogeenproductie. Wij vonden ook een correlatie tussen HER-2/neu en
aromatase, hetgeen een vergelijkbare relatie zou kunnen suggereren bij
endometrioid endometriumcarcinoom. Aan de andere kant, moet er niet alleen
gekeken worden naar de effecten van aromataseremmers lokaal in het
endometrium, maar ook naar de effecten m het lichaamsvet. Bij
hormoonathankelijk mammacarcinoom worden aromataseremmers namelijk
ingezet juist mede vanwege hun effect in het perifere vetweefsel. Op
theoretische gronden zal dit oestrogeenverlagende effect van aromatase in
vetweefsel niet veel bijdragen aan een betere prognose, daar de expressie van
oestrogeenreceptoren m de hogere stadia en graden van het
endometriumcarcinoom juist afneemt.
In hoofdstuk 7 wordt de prognostische waarde onderzocht van de aanwezigheid
van de oestrogeen receptor (ER) en progesteron receptor (PR) alpha en beta, bij
het endometrioid endometriumcarcinoom. Endometriumweefsel van 315
chirurgisch behandelde patienten met endometrioid endometriumcarcinoom
werd onderzocht. Het doel van de studie was om de expressie van de genoemde
1 7 1
receptoren te relateren aan klinische karakteristieken, tumor recidief en
overlevingsduur. Expressie van ER alpha was gerelateerd aan de vroege stadia
endometriumcarcinoom (p=0.020), terwijl expressie van ER alpha (p<0.0001),
PR alpha (p<0.001) en PR beta (p=0.001) gerelateerd bleken aan lagere graad
endometriumcarcinoom. Afwezigheid van ER alpha was bij het vroege stadium
endometriumcarcinoom onafhankelijk gerelateerd aan sterftekans (p=0.017, OR
7.28, 95% CI 1.42-37.25). Dit suggereert dat adjuvante therapie overwogen zou
kunnen worden bij vroeg stadium ER alpha negatief endometriumcarcinoom.
Afwezigheid van PR alpha bleek een onafhankelijke voorspeller van ziekte
recidief (p=0.015, OR 4.2, 95% CI 1.32-13.33). Ook bleek de verhouding ER
alpha/ ER beta<l en PR alpha/ PR beta<l gerelateerd aan risico op ziekte
recidief en sterfte. Deze waamemingen illustreren dat de verhouding tussen deze
biomarkers van prognostische betekenis kan zijn. Hoewel aromatase geen deel
uitmaakte van de studie, beschreven in hoofdstuk 7, bleken er geen relaties te
bestaan tussen ER en PR expressie en aromatase. Wij concludeerden dat
expressie van ER en PR alpha en beta en hun ratio's nuttige parameters zijn om
inzicht te krijgen in de kans op ziektevrije periode en overleving bij
endometrioid endometriumcarcinoom.
Toekomstig onderzoek.
Verder onderzoek dient gericht te z1Jn op het verder ontsluieren van de
moleculaire basis van het ontstaan van endometriumcarcinoom, en de exacte rol
van aromatase hierbij, daar onze observaties vanwege het associatieve karakter
niet beschouwd kunnen worden als een bewijs voor een rol voor aromatase bij
de carcinogenese. Wellicht dient het bewijs voor de rol van aromatase bij het
ontstaan van endometriumcarcinoom gevonden te worden in in vitro studies: in
cellijnen met normaal, hyperplastisch en maligne endometrium dient de invloed
op DNA synthese bestudeerd te warden van toevoeging en deprivatie van
androgenen en aromatase. Hoewel zich bij de behandeling van
172
hormoonafhankelijk mammacarcmoom een ware revolutie voordeed met de
introductie van aromataseremmers, die de omzetting van androgenen in
oestrogenen succesvol bleken te remmen (hoofdstuk 3), is het niet
waarschijnlijk dat dit ook zal gebeuren bij het endometriumcarcinoom. De
prognose van het endometrioid endometriumcarcinoom is over het algemeen
goed te noemen, in tegenstelling tot mammacarcinoom, mits tijdig chirurgische
en/of radiotherapeutische behandeling wordt ingesteld. Een palliatieve rol voor
aromataseremmers bij gemetastaseerd of hoger stadium endometriumcarcinoom
kan wel overwogen warden. Daarom is het interessant onderzoek te doen naar
de hormonale interacties bij recidief endometriumcarcinoom, daar de patienten
met een recidief juist degenen zijn die zouden kunnen profiteren van een
endocrinologische behandeling met aromataseremmers. Objectieve respons van
de ziekte op behandeling in toekomstige studies dient dan gerelateerd te warden
aan mogelijke voorspellende markers, zoals besproken in hoofdstuk 6 en 7,
inclusief aromatase. Een fase II studie vermeldt teleurstellende resultaten bij
behandeling met aromataseremmers van patienten met gevorderd
endometriumcarcinoom en positieve expressie van ER, PR en HER-2/neu, maar
de aantallen patienten zijn klein. 1 De mogelijke rol van aromataseremmers bij
de behandeling van steroid receptor positief endometriumcarcinoom is dus nog
niet duidelijk, daar men theoretisch effect zou mogen verwachten van
behandeling met aromataseremmers bij ER positieve tumoren. Omdat aromatase
een rol speelt bij de omzetting van androgenen in oestrogenen lijkt onderzoek
naar de rol van aromatase expressie bij non-endometrioid
endometriumcarcinoom minder zinvol. Inderdaad werden in een fase II studie
teleurstellende resultaten gevonden van behandeling met aromataseremmers bij
patienten met gevorderd of gemetastaseerd endometriumcarcinoom. 2 Slechts 2
van de 23 patienten toonden gedeeltelijke verbetering. De meeste van de
pa ti en ten hadden slecht gedifferentieerd of non-endometrioid
endometriumcarcinoom. Men kan opmerken dat selectie van de patienten en
1 73
korte duur van de behandeling de resultaten wellicht negatief hebben beinvloed.
Daar aromatase een rol zou kunnen spelen bij het ontstaan van
endometriumcarcinoom zou toekomstig onderzoek zich kunnen richten op de
vraag of aromataseremmers profylactisch gebruikt zouden kunnen worden bij de
behandeling van ( atypische) endometriumhyperplasie, met het doel een
hysterectomie te voorkomen, met name bij jonge patienten die hun
vruchtbaarheid willen behouden. Een andere uitdaging kan gevonden worden in
evaluatie van het endometrium bij postmenopauzale patienten met
mammacarcinoom die worden behandeld met aromataseremmers. Verschillende
studies vinden namelijk een mogelijk beschermend effect van
aromataseremmers op het endometrium bij patienten die eerder werden
behandeld met tamoxifen. 3-5 De dikte van het endometrium dient bij deze
patienten zorgvuldig vervolgd te worden, totdat de (late) effecten van
vervanging van tamoxifen door aromataseremmers goed in kaart zijn gebracht.
References
1. Ma BBY, Oza A, Eisenhauer E, Stanimir G, Carey M, Chapman W, et al. The activity of letrozole in patients with advanced or recurrent endometrial cancer and correlation
with biological markers - a study of the National cancer Institute of Canada Clinical
Trials Group. Int J Gynecol Cancer 2004; 14 : 650-8. 2. Rose PG, Brunetto VL, VanLe L, Bell J, Walker JL, Lee RB. A phase II trial of
anastrozole in advanced recurrent or persistent endometrial carcinoma: a gynecologic
oncology group study. Gynecol Oncol 2000; 78: 2 12-6.
3. Cohen I. Aromatase inhibitors and the endometrium. Maturitas (2008),
doi: 10. 1016/j .maturitas.2008.03 .001.
4. Duffy SRO, Greenwood M. The endometrial data from the ATAC (Arimidex,
Tamoxifen, Alone or in Combination) trial indicates a protective effect of anastrozole
(arimidex) upon the endometrium. Breast Cancer Res Treat 2003 ; 83 : abs 134. 5. Duffy S, Jackson TL, Lansdown M, Philips K, Wells M, Pollard S, Clack G, Coibon
M, Bianco AR. The ATAC ('Arimidex', Tamoxifen, Alone or in Combination)
174
adjuvant breast cancer trial; first results of the endometrial sub-protocol following 2
years of treatment. Hum Reprod 2006; 21(2): 545-53 .
175
Authors and affiliations
From the Department of Gynaecologic Oncology, University Medical Center
Groningen, University of Groningen, The Netherlands
JM Briet
KA ten Hoor
RA de Jong
E Joppe
HW Nijman
AGJ van der Zee
From the Department of Obstetrics and Gynaecology, University Medical Center
Groningen, University of Groningen, the Netherlands
( Present affiliation: Academic Medical Centre of the University of Amsterdam,
Amsterdam, The Netherlands)
MJ Heineman.
From the Department of Pathology and Laboratory Medicine, University
Medical Center Groningen, University of Groningen, The Netherlands
H Hollema
From the Department of Epidemiology and Statistics, University Medical Center
Groningen, University of Groningen, The Netherlands
M Boezen
176
From the Department of Endocrinology, University Medical Center Utrecht, The
Netherlands
GH Donker
JHH Thij ssen
From the Department of Obstetrics and Gynaecology of the Medisch Centrum
Leeuwarden, Leeuwarden, The Netherlands
JG Santema
From the Department of Obstetrics and Gynaecology of the Wilhelmina
Ziekenhuis Assen, Assen, The Netherlands
AVSluijmer
From Novartis Institutes for BioMedical Research Basel, Novartis Pharma AG,
Oncology Research, K136-P65, CH-4002 Basel, Switzerland
DB Evans
177
178
Dankwoord
Promoveren doe je niet alleen, wordt vaak gezegd. Oat is zeker waar en dat heh
ik heel duidelijk ondervonden. Dit proefschrift was nooit tot een goed einde
gekomen zonder de prettige samenwerking met en stimulans van de
gynaecologische oncologie-groep in het UMCG. Met name dit hechte team hen
ik dank verschuldigd.
Hooggeleerde van der Zee. Beste Ate, als kritische stimulator van onderzoek en
"bijschaver" van artikels hen je een ongelooflijke kei. Je gaf me op bet juiste
moment de duw in de rug om de weg der tma' s in te slaan, zodat de tweede helft
van dit proefschrift geschreven kon worden. Oat duwtje had ik trouwens ook
nodig! Je gaf duidelijke en ongezouten kritiek op de stukken, maar het werd er
natuurlijk altijd beter van. Fijn dat je bet stokje van Maas Jan wilde ovememen,
naast je reeds zo voile agenda. Ik vind bet een groot voorrecht dat je mijn boekje
hebt willen begeleiden.
Hooggeleerde Heineman. Beste Maas Jan, je hebt een zeer grote bijdrage
geleverd aan bet tot stand komen van dit proefschrift. De wetenschappelijke
gedachtenvorming over bet onderwerp aromatase en de eerste stimulans om in
dit onderwerp te duiken komen van jou vandaan. Steeds weer wist je me
vriendelijk te motiveren door te gaan: "bet boekje moet wel af' ! De gesprekken
op je kamer heh ik altijd zeer gewaardeerd. Je bent een stimulerend voorbeeld.
Hooggeleerde Hollema. Beste Harry, geweldig hoe je je zeer grote vakkennis
weet te combineren met gevoel voor humor! De plezierige en laagdrempelige
manier waarop je me corrigeerde en nieuwe zaken bijbracht op bet mij
onbekende terrein der pathologie was van grote waarde ! Door een microscoop
179
ziet de wereld er inderdaad anders uit. Daamaast hielp jouw relativerende
houding als ik weer eens de moed liet zakken. Ik ben buitengewoon blij <lat ook
jij mijn promotor wilde zijn.
Hooggeleerde Nijman. Beste Hans, als endometriumonderzoeker werd jij toch
pas in de tweede helft van <lit geheel bij de studies betrokken. En direct bleek jij
"de man". Als directe begeleider op de werkvloer konden de TMA studies vlot
voortgang vinden, en gaf je in alle voortgangsgesprekken duidelijk en
stimulerend richting aan. Heel veel dank daarvoor !
De leescommissie, bestaande uit de hooggeleerden Prof. Dr. J.G. Aalders,
Prof. Dr. J.H.H.Thijssen en Prof. Dr. P.H.B. Willemse ben ik zeer erkentelijk
voor hun beoordeling en vanzelfsprekend ook voor hun goedkeuring van mijn
manuscript.
Hooggeleerde Aalders, mijn opleider in het UMCG en als gynaecologisch
oncoloog zelf gepromoveerd op endometriumcarcinoom. Hoewel woonachtig en
werkzaam in het buitenland, bleek u toch bereid zitting te willen nemen in de
beoordelingscommissie van <lit manuscript.
Hooggeleerde Thijssen. Ik ben zeer vereerd dat u als wandelende vraagbaak op
endocrinologisch gebied uw medewerking heeft gegeven aan een deel van de
"aromatase" studie. U vormde een brug tussen Groningen en Utrecht. Ik ben erg
verheugd dat u als deskundige op het gebied van aromatase ook als lid van de
beoordelingscommissie bij het manuscript betrokken bent.
Hooggeleerde Willemse. Als oncoloog draagt u bij aan een evenwichtige
samenstelling van de beoordelingscommissie, die behalve een gynaecologisch
oncoloog en een endocrinoloog natuurlijk ook een intemistisch oncoloog diende
1 80
te bevatten. Alle drie deze vakgebieden ZlJn van belang gezien het
multidisciplinaire karakter van de inhoud van het manuscript. Ik dank u
vriendelijk voor uw bereidwilligheid zitting te willen nemen in de
beoordelingscommissie en uw bijdrage hieraan.
Mw. Ten Hoor. Beste Klaske, zonder jouw onmisbare hulp en inspanning bij het
beoordelen van alle coupes en laboratoriumwerkzaamheden was het allemaal
niks geworden. Inmiddels had je in al die jaren al wel tien keer zelf kunnen
promoveren. De dinsdagmiddagen "scoren" van weefsels met een kopje thee
waren altijd erg gezellig. Klaske, je bent geweldig.
Justine, Renske, en Enja. Ook jullie waren fijne partners bij het beoordelen van
de tissue microarrays. Door jullie voelde ik me weer een "jonge onderzoeker".
Met name noem ik heel speciaal Justines onmisbare hulp en inzet bij de
statistiek ondanks de drukte van de opleiding, hetgeen een van de belangrijkste
schakels in de studie bleek en zeer werd gewaardeerd! Justine, buitengewoon
bedankt! Ik ben blij dat we elkaar op deze manier konden helpen. Ik ben ervan
overtuigd dat jullie vervolgopleiding ongetwijfeld succesvol zal zijn.
Mw Donker. Beste Truus, fijn dat je vanuit Utrecht zoveel hebt bijgedragen aan
de aromatase-bepalingen. Dank voor al je tijdsinvestering en belangstelling!
Zeergeleerde Boezen. Beste Marike, dank voor je waardevolle hulp, waardoor
de betrouwbaarheid van het geheel verder werd vergroot. Al werkende aan het
manuscript is mijn respect voor statistiek nog verder toegenomen!
Dear Dean, thanks for your help in disposing the 677 antibodies. Thanks to your
help and Novartis we were able to continue and complete our studies in
18 1
aromatase, using as reliable antibodies as possible. Thanks for this pleasant
collaboration! I hope our results will be of benefit for you as well.
Zeergeleerde Brouwer. Beste Willem, opleider van het eerste uur. Ook al hen je
niet direct bij dit boekje betrokken geweest, toch hoor je absoluut in dit
dankwoord thuis. Ik heh veel aan je te dank.en. Jij wist nog wel een tropenarts in
de bush van Tanzania, die belangstelling had voor de gynaecologie. Je haalde
me naar Leeuwarden en stimuleerde me steeds weer op de jou eigen wijze in de
zoektocht naar een opleidingsplaats. Je creatief enthousiasme zal me levenslang
bijblijven. Ik beschouw jou als de meester, en mij als je gezel. Ik zal je nooit
vergeten, en hen trots te kunnen zeggen "ik hen nog opgeleid door Brouwer".
Zeergeleerde Santema. Job, beste vriend, jij hebt letterlijk aan de wieg van dit
proefschrift gestaan. Wij tweeen weten wat met deze zinsnede bedoeld wordt.
Samen met Willem beschouw ik jou als mijn eerste opleider. Bedankt.
Maatschapsleden Leeuwarden en oncologisch gynaecologen te Groningen, en
collega arts-assistenten. Veel dank voor het indertijd steeds maar weer
includeren van patienten voor "dat onderzoek van Vincent". Angelique en Peet,
die geregeld vanuit Leeuwarden weefsel in een tank met vloeibare stikstof
meebrachten naar Groningen, bedankt. Elke keer was ik weer bevreesd dat de
vloeibare stikstof in de auto zou gaan lekken met dramatische gevolgen.
Gelukkkig ging het steeds weer goed!
Ook de vakgroep Pathologie te Leeuwarden en met name Dr Joris Grond dank
ik voor de vriendelijke bereidwilligheid steeds weer endometriumweefsel uit te
willen snijden en gebruik te mogen maken van de archieffaciliteiten.
Geke de Ruijter en Janny Abels wil ik hartelijk bedanken voor hun hulp in het
UMCG bij het plannen van data, versturen van manuscripten en nog veel meer.
182
Ook mijn beide zeer gewaardeerde paranimfen wil ik buitengewoon bedanken
voor hun hulp en inzet! Fijn dat jullie erbij waren!
Ik mag niet vergeten ook de patienten te bedanken, die bereid waren te
participeren in de prospectieve studies, beschreven in hoofdstuk 4 en 5, en bereid
bleken peroperatief weefsel en bloed af te staan ten bate van de wetenschap.
Maatschapsleden Meppel, beste George, Guus en Aster. Met je maatschapsleden
ben je inderdaad ook een beetje getrouwd, zoals velen reeds in hun boekje
opmerkten. Dank voor de goede samenwerking en harmonie de afgelopen jaren.
Laten we vooral doorgaan onze "afdeling" verder uit te bouwen tot een goede
gynaecologische kliniek.
Natuurlijk wil ik ook mijn ouders bedanken, die mij in staat stelden te komen
waar ik kwam, dankzij hun stimulatie vanaf mijn prille jeugd. Lieve mamma en
pappa, ook jullie mogen bier genoemd worden!
Zeergeleerde van Vuuren. Lieve, lieve Anke, dank voor alles. Ik kon natuurlijk
niet achterblijven na jouw promotie. Samen hebben we erg ftjne, maar ook
verdrietige tijden doorgemaakt. We beleefden avonturen in Tanzania en
Nederland. Onze band is in al die jaren alleen maar hechter geworden. Het is fijn
met jou en de kinderen op onze boerderij in Drenthe. Een reeds gepromoveerde
echtgenote blijkt toch de beste prikkel om een proefschrift te voltooien, waarbij
geldt: "twee promovendi op een kussen, daar komt niemand tussen". Dr. van
Vuuren, vanaf nu slapen er twee zeergeleerden op een kussen.
Lieve Marije, Karlijn, Hielke en Jolien, ik ben erg blij met jullie. Op zulke
kinderen kun je trots zijn! Nu kunnen jullie weer wat vaker op de computer.
183
List of publications
Jongen VHWM, van den Heuvel MM. An abdominal pregnancy. TropicalDoct<?r 1995 ; 25 : 124.
Jongen VHWM. Alternative external fixation for open fractures of the lower leg. TropicalDoctor 1995; 25 : 173-4.
Jongen VHWM. Ectopic pregnancy and culdo-abdomino-centesis.
Int J Gynecol Obstet 1996; 55 : 7 5-6.
Jongen VHWM, Koetsier DW. Pitfalls in the diagnosis of complete hydatidiform mole with
coexistent live fetus. Trop Med & Int Health 1997; 2: 289-90.
Jongen VHWM. Autotransfusion and ectopic pregnancy; an experience from Tanzania.
TropicalDoctor 1997; 27: 78-9.
Jongen VHWM, van Holten J, Ti ems J. Neonatal Borrelia Duttoni infection: a report of3 cases. TropicalDoctor 1997; 27 : 1 15-7.
Jongen VHWM, van RoosmalenJ. Complications of cervical cerclage in rural areas.
Int J Gynecol Obstet 1997; 57: 23-6.
Jongen VHWM, van Roosmalen J, Wetsteyn JCFM, et al. Tick-borne relapsing fever and
pregnancyoutcome in rural Tanzania.Acta Obstet Gynecol Scand 1997; 76: 834-8.
Jongen VHWM, Santema JG, Grond AJK, van VeelenH. Uterine amyloidosis in menopause.
Br J Obstet Gynaecol 1998; 105 : 362-4.
Jongen VHWM, Halfwerk MOC, Brouwer WK. Vaginal delivery after previous caesarean
section for failure of second stage oflabour. Br J Obstet Gynaecol 1998; 105 : 1079-8 1 .
Jongen VHWM. A case of dystocia in conjoined twins. Tropical Doctor 1999; 29 : 54-5 .
184
Renes WB, Jongen VHWM, van RoosmalenJ. De prijs van een cerclage.
Ned TijdschrObstet Gynaecol 1999; 1 12: 121-4.
Jongen VHWM, Brouwer WK. Comparison of the modified Pereyra procedure using permanent
suture material and Burch urethropexy; subjective results of a one to six years follow-up study.
Eur J Obstet Gynecol Reprod Biol 1999; 84: 7- 1 1 .
Jongen VHWM, Santema JG, Wallenburg HCS. Het primaire antifosfolipidesyndroom in de
zwangerschap.Ned TijdschrObstet Gynaecol2000; 1 13 : 79-82.
Jongen VHWM, Holm JP, Verschuuren EAM, van der Bij W. Vaginal delivery after lung
transplantation. Acta Obstet Gynecol Scand 2000; 79 : 1 132-3 .
Jongen VHWM, Collins J, Lubbers JA, van Selm GM. Unsuspected pregnancy at
hysterosalpingographyFertil Steril 2001; 76: 610- 1.
Jongen VHWM, Sluijmer AV, Heineman MJ. The postmenopausal ovary as an androgen
producing gland; hypothesis on the etiology of endometrial cancer. Maturitas 2002; 43 : 77-85.
Jongen VHWM, Hollema H, Santema JG, van der Zee AGJ, Heineman MJ. Ovarian stromal hyperplasia and ovarian vein steroid levels in relation to endometrioidendometrial cancer.
Br J Obstet Gynaecol 2003; 1 10: 690-5.
Jongen VHWM, Thijssen JHH, Hollema H, Donker GH, Santema JG, van der Zee AGJ, Heineman MJ. Is aromatase P450 involved in pathogenesis of endometrioid endometrial cancer?
Int l Gynecol Cancer 2005; 15 : 529-36.
Jongen VHWM, Hollema H, van der Zee AG, Heineman MJ. Aromatase in the context of breast and endometrial cancer. A review. Minerva Endocrinol 2006; 3 1 : 4 7-60.
Jongen VHWM, van RoosmalenJ. Een kraamvrouw met opisthotonus en strekkrampen. Ned Tijdschr voor Geneeskunde2007; 15 1(16):9 14.
185
Jongen VHWM, Briet JM, de Jong RA, ten Hoor KA, Boezen M, van der Zee AGJ, Nijman H, Hollema H. Expression of estrogen and progesterone receptors alpha and beta in a large cohort of patients with endometrioid endometrial cancer. Submitted.
Jongen VHWM, Briet JM, Joppe E, de Jong RA, ten Hoor KA, Boezen M, Evans DB, Hollema H, van der Zee AGJ, Nijman HW. Aromatase, COX-2, HER-2/neu and P53 as prognostic factors in endometrioid endometrial cancer. Submitted.
Abstracts
Deutsche Gesellschaft fur Gynekologie und Geburtshilfe/ Tropengynekologie; 3e Workshop Heidelberg Geburtshilfe unter einfachen Bedingungen, dee 1996: 3 .Rundbrief 13. 12. 1996:47. "Threatening threads of cervical cerclage: is a cervical cerclage appropriate in rural Africa?" Oral presentation.
22nd Annual Meeting te Amsterdam van de International Urogynecological Association, aug
1997. Int Urogynecol J 1997;8( 1): S3. "Comparison of the modified Pereyra procedure and Burch urethropexy". Oral presentation.
GynaecongresGroningen, mei 2003. NTOG 2003 ; 1 16:27 1-3 . "Postmenopauzaleovariumfunctieen de pathogenese van endometriumcarcinoom".
Oral presentation. Winnaaragio-prijsNVOG.
15 th International meeting van de European Society of Gynaecological Oncology (ESGO) te
Berlijn, oktober 20007. Abstract nr. 1804. "Estrogen and progesterone receptors alpha and beta are strong prognostic factors in endometrioid endometrial cancer". Oral presentation.
15 th International meeting van de European Society of Gynaecological Oncology (ESGO) te Berlijn, oktober 2007. Abstract nr. 1809. "The role ofaromatase, COX-2, HER-2/neu and P53
as prognostic factors in endometrioid endometrial canrer". Poster presentation.
186
Curriculum vitae
Vincent Jongen werd op 23 mei 1963 geboren als zoon van Maria Bender en Ir.
Hubertus Jongen. Na het behalen van het diploma Gymnasium beta aan het
Hertog Jan College te Valkenswaard en het vervullen van de militaire
dienstplicht bij het Garderegiment Jagers werd in 1982 gestart met de studie
Geneeskunde te Utrecht. In 1989 werd het artsexamen gehaald. In het kader van
de opleiding tot tropenarts werkte hij in het Ziekenhuis Overvecht te Utrecht en
het Ikazia Ziekenhuis te Rotterdam als arts-assistent chirurgie, inteme
geneeskunde/ cardiologie, gynaecologie/verloskunde en kindergeneeskunde. Van
1992 tot 1996 was hij als tropenarts via Memisa Medicus Mundi uitgezonden
naar Ndala Hospital, Tanzania. Na terugkeer in Nederland werkte hij 2 jaar in
het Medisch Centrum Leeuwarden als agnio gynaecologie, totdat aldaar in april
1998 met de opleiding tot gynaecoloog gestart kon warden (opleider Dr. W.A.
Brouwer). De opleiding gynaecologie werd daama voortgezet in het
Academisch Ziekenhuis Groningen (opleider Prof. Dr. J.G. Aalders) en de Isala
Klinieken te Zwolle (opleider Dr. J. van Eijck). Vanafjuli 2003 is hij werkzaam
binnen de maatschap gynaecologie van het Diaconessenhuis te Meppel (in
associatie met mw. A.B. Koster, G.J. Pricker en Dr. G.M. Vermeulen).
Hij is getrouwd met Dr. Anke van Vuuren. Samen kregen zij vijf kinderen,
namelijk Marije (1991), Karlijn (1993), Renske (t l997), Hielke (1998) en
Jolien (2000). In zijn vrije tijd houdt hij zich graag bezig met de paarden op de
boerderij.
187