Modern Pathology (2011) 24, 924–931

924 & 2011 USCAP, Inc. All rights reserved 0893-3952/11 $32.00

Androgen receptor expression in breast

cancer in relation to molecular phenotype:
results from the Nurses’ Health Study
Laura C Collins1, Kimberly S Cole1, Jonathan D Marotti2, Rong Hu3,4, Stuart J Schnitt1
and Rulla M Tamimi3,4
1
Department of Pathology, Beth Israel Deaconess Medical Center and Harvard Medical School, Boston,
MA, USA; 2Department of Pathology, Dartmouth-Hitchcock Medical Center, Lebanon, NH, USA; 3Channing
Laboratory, Department of Medicine, Brigham and Women’s Hospital and Harvard Medical School,
Boston, MA, USA and 4Department of Epidemiology, Harvard School of Public Health, Boston, MA, USA

Previous studies have demonstrated that androgen receptor is expressed in many breast cancers, but its
expression in relation to the various breast cancer subtypes as defined by molecular profiling has not been
studied in detail. We constructed tissue microarrays from 3093 breast cancers that developed in women
enrolled in the Nurses’ Health Study. Tissue microarray sections were immunostained for estrogen receptor
(ER), progesterone receptor (PR), human epidermal growth factor receptor 2 (HER2), cytokeratin 5/6, epidermal
growth factor receptor (EGFR) and androgen receptor (ER). Immunostain results were used to categorize each
cancer as luminal A or B, HER2 and basal like. The relationships between androgen receptor expression and
molecular subtype were analyzed. Overall, 77% of the invasive breast carcinomas were androgen receptor
positive. Among 2171 invasive cancers, 64% were luminal A, 15% luminal B, 6% HER2 and 11% basal like. The
frequency of androgen receptor expression varied significantly across the molecular phenotypes (Po0.0001).
In particular, androgen receptor expression was commonly observed in luminal A (91%) and B (68%) cancers,
but was less frequently seen in HER2 cancers (59%). Despite being defined by the absence of ER and PR
expression and being considered hormonally unresponsive, 32% of basal-like cancers expressed androgen
receptor. Among 246 cases of ductal carcinoma in situ, 86% were androgen receptor positive, but the frequency
of androgen receptor expression differed significantly across the molecular phenotypes (P ¼ 0.001), and high
nuclear grade lesions were less likely to be androgen receptor positive compared with lower-grade lesions.
Androgen receptor expression is most commonly seen in luminal A and B invasive breast cancers. However,
expression of androgen receptor is also seen in approximately one-third of basal-like cancers, providing further
evidence that basal-like cancers represent a heterogeneous group. Our findings raise the possibility that
targeting the androgen receptor pathway may represent a novel therapeutic approach to the management of
patients with basal-like cancers.
Modern Pathology (2011) 24, 924–931; doi:10.1038/modpathol.2011.54; published online 6 May 2011

Keywords: androgen receptor; breast cancer; molecular subtype

Recent gene expression profiling studies using subgroups that have been the most reproducibly
microarrays and unsupervised cluster analysis identified by these studies are luminal subtypes
have provided new insights into the classification A and B, both of which are estrogen receptor (ER)
of invasive breast cancers.1–4 The breast cancer positive and/or progesterone receptor (PR) positive;
the human epidermal growth factor receptor 2
(HER2) subtype; and the basal-like group.1–4 The
Correspondence: Dr LC Collins, MD, Division of Anatomic clinical relevance of these molecular subgroups is
Pathology, Beth Israel Deaconess Medical Center, 330 Brookline supported by studies that have demonstrated differ-
Avenue, Boston, MA 02215, USA.
E-mail: [email protected]
ences in outcome and response to various adjuvant
Received 4 November 2010; revised 29 November 2010; accepted and neoadjuvant therapies according to molecular
29 November 2010; published online 6 May 2011 subtype.3–6 Furthermore, molecular profiling has

www.modernpathology.org
 Androgen receptor expression in breast cancer
LC Collins et al 925

verified that expression of ER, PR and related genes Breast Cancer Case Confirmation
is the major determinant in the subclassification of
breast cancers. All women reporting incident diagnoses of cancer
ER is known to play an important role in were asked for permission to review their medical
endocrine-mediated tumor development and has records to confirm the diagnosis and to classify
been shown to influence breast cancer development cancers as in situ or invasive, by histologic type, size
and progression.7 Although ER and PR have been and presence or absence of metastases. To identify
studied extensively, relatively little is known about cases of cancer in nonrespondents who died, death
the role of androgens and androgen receptor in certificates for all deceased participants and medical
breast cancer. Previous studies have shown that records for the incident cancers were obtained.
androgen receptor is expressed in 60–85% of breast Following medical record review, 99% of self-
cancers, and in some cases it is more highly reported breast cancers were confirmed.
expressed than ER or PR.8–13 Additionally, epide-
miologic studies have found that high circulating Breast Cancer Tissue Block Collection
androgen levels are associated with an increased
risk of developing breast cancer, particularly among In 1993, we began collecting archived formalin-fixed
postmenopausal women.14–18 The biologic roles of paraffin-embedded breast cancer blocks for partici-
androgens in the breast are incompletely under- pants with primary incident breast cancers over 20
stood, as it is unclear whether the effects of years of follow-up (1976–1996). Cases who reported
androgens on breast cells are predominantly pro- a prevalent cancer including breast cancer at base-
liferative or antiproliferative. Recently, the effect of line were excluded from collection. Of the 5610
androgens on breast cancer cell lines and the breast cancers that were eligible for block collection,
potential role of the androgen receptor pathway in we were unable to obtain any pathology material for
breast cancer have been explored.19–21 Results 1858 cases. The primary reason was because they
suggest a possible antiproliferative effect of andro- had been destroyed by the hospital (45%). Because
gen receptor stimulation and pathway activation in the majority of hospitals archive tissue blocks for
breast cancer.19–21 only 5–10 years, we were more successful in
There has been recent interest in evaluating the obtaining more recent blocks. Because year of
expression of androgen receptor among the molecu- diagnosis and age at diagnosis are highly correlated
larly defined categories of invasive breast cancer, (Spearman’s correlation ¼ 0.49; Po0.0001), the tem-
particularly among the triple-negative (or basal like) poral effect on our collections is evident not only in
and the HER2 groups that are considered to be the differences in age at diagnosis, but also in the
hormone receptor negative.11,22,23 However, large frequency of premenopausal breast cancers when
population-based studies investigating expression comparing the women from whom we obtained
of androgen receptor in relation to molecular specimens with those for whom we did not.
phenotype or among women with ductal carcinoma However, these two groups of women were very
in situ are lacking. Therefore, the objective of this similar regarding a number of other breast cancer
study was to examine the expression of androgen risk factors and tumor characteristics (data pub-
receptor in relation to tumor stage, pathologic lished previously24). After taking into account the
features and molecular phenotype using a large, age and year of diagnosis, the participants whose
well-characterized population of women with breast tumors were included in the tissue microarrays were
cancer. very similar to those for whom we were unable to
obtain tissue blocks.
We obtained pathology material for 3752 partici-
pants. Of these, 390 specimens were hematoxylin
Materials and methods and eosin (H&E)-stained slides only and 45 tissue
Study Population blocks had to be returned to the lending hospital
before construction of the tissue microarrays and
Study design and population thus could not be included. H&E sections of the
The Nurses’ Health Study was initiated in 1976, corresponding 3317 paraffin-embedded tissue
when 121 700 US registered nurses aged 30–55 years blocks were reviewed by a single pathologist to
returned an initial questionnaire. The cohort confirm the cancer diagnosis, classify the cancer
has been followed by mailed questionnaires bien- according to histologic type and grade (Nottingham),
nially to update exposure information and ascertain and circle the area from which the cores for the
nonfatal incident diseases. Information on body tissue microarrays would be taken. Pathology review
mass index, reproductive history, age at menopause identified 420 tumor blocks as unusable for tissue
and postmenopausal hormone use as well as microarray construction. The majority of exclusions
diagnosis of cancer and other diseases are updated were because the block did not contain residual
every 2 years through questionnaires. The follow- tumor (60%) or there was insufficient tumor for the
up rate among this cohort was over 90% through tissue microarrays (26%). Tissue microarrays were
1996. constructed in the Dana Farber Harvard Cancer

Modern Pathology (2011) 24, 924–931

 Androgen receptor expression in breast cancer
926 LC Collins et al

Center Tissue Microarray Core Facility (Boston, DAB as the chromogen-substrate. Tissue sections
MA). From each breast cancer, three 0.6-mm cores from samples of prostate carcinoma were used as
were obtained and inserted into the recipient tissue both positive and negative controls and were
microarray blocks. In total, 23 tissue microarray included in all staining runs.
blocks were constructed from 3093 cancers and Immunostained tissue microarray slides were
positive lymph nodes from 2897 participants. We evaluated for ER, PR and androgen receptor expres-
excluded from the current analysis participants with sion, HER2 protein overexpression, and expression
positive lymph nodes only (n ¼ 25), rare tumor types of CK5/6 and EGFR in each core. Tumor cells that
including malignant phyllodes tumors, neuroendo- showed nuclear staining for ER, PR or androgen
crine carcinoma and angiosarcoma (n ¼ 10), lobular receptor were considered ER positive, PR positive or
carcinoma in situ (n ¼ 31), in situ carcinomas with androgen receptor positive, respectively, whereas all
both ductal and lobular features (n ¼ 13) and miss- ER-negative, PR-negative and androgen receptor-
ing information on androgen receptor status negative cases showed complete absence of tumor
(n ¼ 287). Among 2258 invasive breast cancers and cell staining in all three tissue cores. Of note, low
273 women with ductal carcinoma in situ with positive ER, PR or androgen receptor (1–10% of
androgen receptor data, 2171 and 246 women also tumor cell nuclei staining) and positive ER, PR or
had complete information on immunophenotype, androgen receptor (410% of tumor cell nuclei
respectively. staining) were collapsed into a single ER, PR or
androgen receptor ‘positive’ category for the pur-
Immunohistochemical Analysis poses of this analysis. Tumor cells were considered
positive for HER2 protein overexpression when
We performed immunohistochemical staining for 410% of the cells showed moderate or strong
ER-a, PR, HER2, cytokeratin (CK)5/6, and epidermal membrane staining (2 þ and 3 þ ). The results of
growth factor receptor (EGFR) on 5-mm paraffin analyses in which HER2 positivity was defined as 3
sections cut from the tissue microarray blocks. þ were very similar to those with a definition of 2 þ
Immunostains for each marker were performed in a and 3 þ .24 Cases were considered basal CK positive
single staining run on a Dako Autostainer (Dako or EGFR positive if any cytoplasmic and/or mem-
Corporation, Carpinteria, CA, USA). These particu- branous staining was detected in the tumor cells,
lar biomarkers were selected for analysis because even if focal. These latter criteria are similar to those
they have been commonly used as a surrogate to previously used for scoring these markers in
classify invasive breast cancers according to their invasive basal-like cancers.4,25,26
molecular phenotypes.4,25–28 The sources and dilu-
tions of the primary antibodies used in this study are Classification of Molecular Phenotype
listed in Table 1. The immunostaining protocols for
ER, PR, HER2, CK5/6 and EGFR have been pre- Immunostained tissue microarray sections were
viously described in detail.24 Immunostaining for reviewed under a microscope and visually scored
androgen receptor was performed on tissue sections for each individual tissue core as described above.
following deparaffinization in two 5-min changes of We classified a case as positive if there was staining
xylene and rehydration through graded alcohols to in any of the three cores from that case and negative
distilled water. After blocking endogenous perox- if there was no immunostaining present. Cases that
idase activity, sections were subjected to heat- were ER positive and/or PR positive, HER2 negative
induced epitope retrieval by heating in a vegetable and histologic grade 1 or 2 were classified as
steamer in citrate buffer (pH 6.1) for 20 min. luminal A cancers; cases that were ER positive
Following heat-induced epitope retrieval, the pri- and/or PR positive and HER2 positive or ER positive
mary monoclonal antibody AR441 (DAKO) was and/or PR positive and HER2 negative but histologic
applied to the sections in a dilution of 1:200 and grade 3 were classified as luminal B cancers; cases
the slides were incubated at room temperature for that were ER negative, PR negative and HER2
30 min followed by incubation with the biotinylated positive were classified as HER2 type; and cases
universal secondary antibody and the avidin-biotin that were negative for ER, PR and HER2 and positive
complex. Visualization was performed with liquid for CK5/6 and/or EGFR were categorized as basal

Table 1 Antibodies and dilutions used

Antibody to Clone Manufacturer Dilution

Androgen receptor AR441 Dako 1:200

Estrogen receptor 1D5 Dako 1:200
Progesterone receptor PgR 636 Dako 1:50
Human epidermal growth factor receptor 2 A0485 Dako 1:400
Cytokeratin 5/6 D5/16B4 Dako 1:50
Epidermal growth factor receptor 2–18C9 Dako Prediluted (pharmDX kit)

Modern Pathology (2011) 24, 924–931

 Androgen receptor expression in breast cancer
LC Collins et al 927

like. Cases that lacked expression of all five markers androgen receptor expression was present in 91%
were considered ‘unclassified’. of luminal A cancers, 68% of luminal B cancers and
59% of HER2-type cancers. Of note, despite their
being defined by the absence of ER and PR
Statistical Analysis expression and being considered hormonally un-
responsive, 32% of basal-like cancers showed
The w2 tests were used to evaluate the independence expression of androgen receptor. Also, 46% of the
of selected variables under the null hypothesis. All unclassified cases expressed androgen receptor.
statistical tests were two sided and P-values o0.05 There were 246 cases of ductal carcinoma in situ
were considered statistically significant. with complete immunophenotypic data: 50% were
luminal A, 28% luminal B, 13% HER2, 7% basal
Results like and 2% unclassified. Overall, 86% of ductal
carcinoma in situ cases were androgen receptor
The population for this analysis consisted of breast positive, but the frequency of androgen receptor
cancers that developed in women in the Nurses’ expression differed significantly across the molecu-
Health Study after the baseline questionnaire lar phenotypes (P ¼ 0.001). Androgen receptor ex-
(1976) through the 1996 follow-up cycle that could pression was commonly observed in ductal
be classified into one of the four molecular pheno- carcinoma in situ with luminal A and luminal B
types and that had evaluable androgen receptor- phenotypes (93 and 84% of cases, respectively), but
stained tissue microarray cores (2171 invasive was less frequently seen in the HER2 subtype (78%
cancers; 246 DCIS). Based on immunostaining data of cases). Again, despite their being defined by
for the five markers used (ER, PR, HER2, EGFR and absence of ER and PR expression, 71% of basal-like
CK5/6), 1380 invasive tumors were classified as ductal carcinoma in situ showed expression of
luminal A (64%); 326 were luminal B (15%); 126 androgen receptor.
were HER2 (6%); and 237 were basal like (11%). The frequency of androgen receptor according to
There were also 102 invasive tumors that were histologic type of invasive cancer is shown in
considered unclassifiable (ER–/PR–/HER2–/EGFR–/ Figure 1. Androgen receptor expression was seen
CK5/6–) for which androgen receptor staining was in most types, including 71.0% of invasive ductal
available (Table 2). carcinomas, 96.3% of invasive lobular carcinomas,
Overall, 77% of the invasive cancers were andro- 80.5% of mucinous carcinomas and 100% of tubular
gen receptor positive. The frequency of androgen carcinomas. Overall, androgen receptor-positive
receptor expression varied significantly across the tumors at presentation were smaller (Po0.0001),
molecular phenotypes (Po0.0001). In particular, more often node negative (P ¼ 0.003), lower grade

Table 2 Distribution of androgen receptor expression in relation to other biomarkers and molecular phenotype

All invasive tumors Invasive ductal, NOS Ductal carcinoma in situ

N (%) Androgen Androgen Androgen Androgen Androgen Androgen P-valuea

receptor+ receptor receptor+ receptor receptor+ receptor

ER
+ 1476 (86.8) 225 (13.2) 916 (83.4) 183 (16.7) 166 (89.3) 20 (10.8) 0.04
 225 (44.0) 287 (56.1) 166 (39.1) 259 (60.9) 45 (72.6) 17 (27.4) o0.0001
PR
+ 1243 (86.9) 187 (13.1) 792 (83.9) 152 (16.1) 138 (87.9) 19 (12.1) 0.2
 475 (59.3) 326 (40.7) 297 (50.5) 291(49.5) 83 (81.4) 19 (18.6) o0.0001
HER2
+ (2+/3+) 154 (64.7) 84 (35.3) 117 (60.6) 76 (39.4) 56 (80.0) 14 (20.0) 0.003
 1551 (78.5) 425 (21.5) 968 (72.6) 365 (27.4) 166 (87.4) 24 (12.6) o0.0001

EGFR
+ 209 (50.4) 206 (49.6) 155 (45.2) 188 (54.8) 43 (74.1) 15 (25.9) o0.0001
 1486 (83.1) 302 (16.9) 929 (78.9) 249 (21.1) 170 (88.1) 23 (11.9) 0.003
CK5/6
+ 44 (35.8) 79 (64.2) 34 (32.1) 72 (67.9) 20 (87.0) 3 (13.0) o0.0001
 1663 (79.2) 438 (20.9) 1055 (73.9) 373 (26.1) 196 (84.9) 35 (15.2) 0.0003
Luminal A 1256 (91.0) 124 (9.0) 739 (88.8) 93(11.2) 115 (92.7) 9 (7.3) 0.19
Luminal B 220 (67.5) 106 (32.5) 187 (65.9) 97 (34.2) 57 (83.8) 11 (16.2) 0.004
HER2 type 74 (58.7) 52 (41.3) 57 (54.3) 48 (45.7) 25 (78.1) 7 (21.9) 0.02
Basal like 75 (31.7) 162 (68.4) 67 (31.0) 149 (69.0) 12 (70.6) 5 (29.4) 0.0009
Unclassified 47 (46.1) 55 (53.9) 25 (34.7) 47 (65.3) 2 (40.0) 3 (60.0) 0.81

a
P-value is comparing the androgen receptor status between invasive ductal, NOS and ductal carcinoma in situ within each receptor status group.

Modern Pathology (2011) 24, 924–931

 Androgen receptor expression in breast cancer
928 LC Collins et al

Table 4 Androgen receptor status among ductal carcinoma in

situ cases according to grade

Androgen Androgen
receptor+ receptor

Nuclear gradea 0.0004

Low 27 (11.6) 1 (2.4)
Intermediate 126 (54.3) 13 (31.7)
High 79 (34.1) 27 (65.9)

a
From pathology review.

positive. The frequency of androgen receptor

expression among breast cancers in this series is
similar to that reported in previous smaller studies
that also used immunohistochemical methods for
Figure 1 Proportion of cases with androgen receptor staining
among different histologic types of invasive breast cancer.
androgen receptor detection (range 74.8–80%).8–13
As reported in previous studies, we found that the
majority of ER-positive tumors also express andro-
Table 3 Tumor characteristics according to androgen receptor gen receptor. Of interest, we also found that
status among women with invasive breast cancer, Nurses’ Health androgen receptor expression was frequent even in
Study (1976–1996)
molecular subtypes of invasive cancer that are ER
Characteristic Androgen Androgen negative. Specifically, we found androgen receptor
receptor+ receptor expression in 59% of cancers in the HER2 group,
N P-value 32% of those in the basal-like group and in 46% of
unclassified carcinomas (those that were ER, PR,
Tumor sizea o0.0001 HER2, CK5/6 and EGFR negative). Results from a
0.1–1.0 cm 401 (24.6) 79 (15.8) recent publication by Niemeier et al11 found a
1.1–2.0 cm 637 (39.1) 183 (36.5)
r2.0 cm 7 (0.4) 3 (0.6) similarly high proportion of HER2-type breast
2.1–4.0 cm 430 (26.4) 163 (32.5) cancers to coexpress androgen receptor (5/8 cases,
4.1+ cm 156 (9.6) 73 (14.6) 63%), although it should be noted that these authors
utilized a different classification system for the
Lymph node statusa 0.003
No nodes 1153 (66.5) 301 (57.4)
luminal B group than we did and that resulted in
1–3 Nodes 324 (18.7) 115 (22.0) all HER2-positive cancers being classified as HER2
4–9 Nodes 132 (7.6) 54 (10.3) type. In that study, only 10% (3/30) of triple-
10+ Nodes 81 (4.7) 37 (7.1) negative breast cancer cases coexpressed androgen
Metastatic at 44 (2.5) 17 (3.2) receptor, compared with 32% (75/237) in this study.
diagnosis
The effects of androgens on the breast are still
Stage at diagnosisa 0.001 incompletely understood. Evidence suggests that
I/II 1349 (82.3) 383 (75.7) estrogens and androgens have opposing effects and
III/IV 290 (17.7) 123 (24.3) that androgens may play a protective role with
Histologic gradeb o0.0001
regard to breast tumor development.29 In support of
1 426 (24.8) 27 (5.2) this argument, we, like others, found that androgen
2 1034 (60.3) 201 (39.0) receptor-negative tumors were more likely to be
3 255 (14.9) 288 (55.8) larger in size, higher grade, and have more
a
extensive lymph node involvement.11–13,22,30 More-
From medical/pathology record. over, androgen receptor expression has been shown
b
From pathology review. to be an independent prognostic factor for better
outcome, even among women with ER-positive
(Po0.0001) and lower stage than androgen receptor- breast cancers.30,31
negative tumors (P ¼ 0.001; Table 3). Among ductal Although androgens and androgen receptor and
carcinoma in situ cases, high nuclear grade lesions the association with breast cancer have been studied
were less likely to be androgen receptor positive previously, their potential role in the development
compared with low- or intermediate-grade lesions of breast cancer remains inconclusive.14–18,32,33 Han-
(Table 4). ley et al34 investigated the potential role of androgen
receptor in relation to breast tumor progression and
showed that 93% of 43 high-grade ductal carcinoma
Discussion in situ cases expressed androgen receptor, whereas
only 55% of 44 high-grade invasive ductal carcino-
In this large population-based study, 77% of mas showed androgen receptor expression.
invasive breast cancers were androgen receptor The authors argue that this may suggest a role for

Modern Pathology (2011) 24, 924–931

 Androgen receptor expression in breast cancer
LC Collins et al 929

androgen receptor in progression to invasion in growth inhibitory effects of DHEA-S on the andro-
high-grade breast carcinomas. A study by Meijnen gen receptor pathway, it appears to be necessary to
et al35 also evaluated the expression of androgen coadminister an aromatase inhibitor to prevent
receptor in 163 cases of ductal carcinoma in situ. In conversion of DHEA-S to estrogen and other hor-
that study, androgen receptor expression was seen mones by aromatase.19,38
in 36, 51 and 26% of cases of low-, intermediate- Our study has several potential limitations. First,
and high-grade ductal carcinoma in situ, a distribu- we were unable to obtain tissue blocks from all
tion similar to that seen in this study. Paradoxically, breast cancers arising in this cohort. Our success in
a higher cutoff of 10% nuclear staining was used to doing so was highly correlated with time between
define androgen receptor positivity in both previous diagnosis and initiation of our tissue block collec-
studies that would not account for the higher tion. After taking into account the effect of age and
proportion of positive cases in the Hanley series.34,35 year of diagnosis, the women for whom we were
Further work is needed to define a role for androgen able to obtain tumor specimens were very similar to
receptor in the transformation of in situ to invasive those for whom we were unable to obtain specimens
carcinoma. (for expanded discussion, see Tamimi et al24).
In this study, androgen receptor expression was Moreover, the frequency of androgen receptor
observed in each of the molecularly defined pheno- positivity among invasive tumors was very similar
types of invasive breast cancer, although the propor- to that observed in other populations, which
tion of cases showing androgen receptor expression suggests that samples included in this study are
was highest in the luminal subtypes. Furthermore, representative of the overall US population. Second,
approximately one-third of basal-like cancers, we utilized immunohistochemical markers as a
which are defined by the absence of ER and PR surrogate to classify breast cancers into the mole-
expression and are considered to be hormonally cular phenotypes defined by expression profiling.
unresponsive, showed expression of androgen Although the antibody panel that we used in this
receptor. This finding raises the possibility that study has been shown to be a reliable proxy for
targeting the androgen receptor pathway may repre- classification of invasive breast cancers categorized
sent a novel therapeutic approach to the manage- by gene expression,4,25–28 the correlation is not
ment of patients with androgen receptor-positive perfect and there will be some misclassification of
basal-like cancers.36 In addition, almost 60% of these phenotypes. The categories as defined by
invasive cancers in the HER2-type molecular group the immunohistochemical markers have been
(also characterized by lack of ER and PR expression) shown to be associated with prognostic markers and
showed expression of androgen receptor. Given survival consistent with what has been seen with
experimental data that have shown cross-talk classification based on RNA expression assays,
between androgen receptor and ERBB2 pathways suggesting that both methods are capturing
in ER-negative breast cancer cell lines,37 combining distinct subgroups.1,2,39–41 More recently, it has been
androgen receptor antagonists with drugs that shown that the distinction between luminal A and B
block the HER2 pathway may provide an additional tumors can be refined by adding the proliferation
therapeutic strategy for patients with ER-negative, marker Ki67 to ER, PR and HER2.42 As Ki67
HER2-positive breast cancers that express androgen data were not available for our cases, we used
receptor. histologic grade as a surrogate for proliferation
The role of androgen receptor-targeted therapy in rate, given the close correlation between prolifera-
ER-negative breast cancers has been explored in tion rate and histologic grade. Thus, our definitions
experimental models.19,20,38 For example, Garreau for luminal A and B are different from those
et al38 used dehydroepiandrosterone sulfate used in our previous studies, but more in keeping
(DHEA-S) and an aromatase inhibitor to treat an with the most recently proposed classification
ER/PR/androgen receptor-negative breast cancer cell scheme.42
line transfected with androgen receptor and were In summary, in this large population-based study,
able to demonstrate cell death with that treatment. androgen receptor expression was commonly seen
The results of that experiment and others like it19–21 in luminal A and B types of invasive breast cancer.
suggest that androgen receptor is responsible Furthermore, expression of androgen receptor was
for this effect and validates the concept that ER/ also seen in a subset of HER2-type and basal-like
PR-negative breast cancers may respond to appro- cancers that are considered to be hormone receptor-
priate hormonal therapy provided that they are negative breast cancers. The role of androgen
androgen receptor positive.19,38 Of interest, andro- receptor in the development and progression of
gen receptor antagonists, such as DHEA-S, have the invasive breast cancers merits further investigation.
opposite effect in ER-positive, androgen receptor- The potential for targeting the androgen signaling
positive cell lines in which proliferation is pathway in breast cancers defined by lack of
induced.20 Thus, it would appear that in the expression of ER and PR raises the possibility of
presence of both ER and androgen receptor, the new therapeutic options for some patients with
estrogenic effect of DHEA-S predominates.20 It tumors previously considered to be hormone
should be emphasized that to fully exploit the independent.

Modern Pathology (2011) 24, 924–931

 Androgen receptor expression in breast cancer
930 LC Collins et al

Acknowledgements mas: potential relevance to new therapeutic strategies.

Cancer 2003;98:703–711.
This study was supported by GlaxoSmithKline 13 Ogawa Y, Hai E, Matsumoto K, et al. Androgen
(WE234 (EP140307)); Public Health Service Grants receptor expression in breast cancer: relationship with
CA087969, and SPORE in Breast Cancer CA089393, clinicopathological factors and biomarkers. Int J Clin
from the National Cancer Institute, National Insti- Oncol 2008;13:431–435.
tutes of Health, Department of Health and Human 14 Baglietto L, Severi G, English DR, et al. Circulating
steroid hormone levels and risk of breast cancer for
Services and Breast Cancer Research Fund. postmenopausal women. Cancer Epidemiol Biomar-
kers Prev 2010;19:492–502.
15 Kaaks R, Berrino F, Key T, et al. Serum sex steroids in
Disclosure/conflict of interest premenopausal women and breast cancer risk
The authors declare no conflict of interest. within the European Prospective Investigation into
Cancer and Nutrition (EPIC). J Natl Cancer Inst 2005;
97:755–765.
References 16 Lillie EO, Bernstein L, Ursin G. The role of androgens
and polymorphisms in the androgen receptor in the
1 Perou CM, Jeffrey SS, van de Rijn M, et al. Distinctive epidemiology of breast cancer. Breast Cancer Res
gene expression patterns in human mammary epithe- 2003;5:164–173.
lial cells and breast cancers. Proc Natl Acad Sci USA 17 Sieri S, Krogh V, Bolelli G, et al. Sex hormone levels,
1999;96:9212–9217. breast cancer risk, and cancer receptor status in
2 Sorlie T, Perou CM, Tibshirani R, et al. Gene expres- postmenopausal women: the ORDET cohort. Cancer
sion patterns of breast carcinomas distinguish tumor Epidemiol Biomarkers Prev 2009;18:169–176.
subclasses with clinical implications. Proc Natl Acad 18 Missmer SA, Eliassen AH, Barbieri RL, et al. Endo-
Sci USA 2001;98:10869–10874. genous estrogen, androgen, and progesterone concen-
3 Sorlie T, Tibshirani R, Parker J, et al. Repeated trations and breast cancer risk among postmenopausal
observation of breast tumor subtypes in independent women. J Natl Cancer Inst 2004;96:1856–1865.
gene expression data sets. Proc Natl Acad Sci USA 19 Hardin C, Pommier R, Calhoun K, et al. A new
2003;100:8418–8423. hormonal therapy for estrogen receptor-negative breast
4 Brenton JD, Carey LA, Ahmed AA, et al. Molecular cancer. World J Surg 2007;31:1041–1046.
classification and molecular forecasting of breast 20 Toth-Fejel S, Cheek J, Calhoun K, et al. Estrogen and
cancer: ready for clinical application? J Clin Oncol androgen receptors as comediators of breast cancer cell
2005;23:7350–7360. proliferation: providing a new therapeutic tool. Arch
5 Millar EK, Graham PH, O0 Toole SA, et al. Prediction of Surg 2004;139:50–54.
local recurrence, distant metastases, and death after 21 Macedo LF, Guo Z, Tilghman SL, et al. Role of
breast-conserving therapy in early-stage invasive breast androgens on MCF-7 breast cancer cell growth and
cancer using a five-biomarker panel. J Clin Oncol on the inhibitory effect of letrozole. Cancer Res
2009;27:4701–4708. 2006;66:7775–7782.
6 Hugh J, Hanson J, Cheang MC, et al. Breast cancer 22 Rakha EA, El-Sayed ME, Green AR, et al. Prognostic
subtypes and response to docetaxel in node-positive markers in triple-negative breast cancer. Cancer
breast cancer: use of an immunohistochemical 2007;109:25–32.
definition in the BCIRG 001 trial. J Clin Oncol 2009;27: 23 Farmer P, Bonnefoi H, Becette V, et al. Identification of
1168–1176. molecular apocrine breast tumours by microarray
7 Althuis MD, Fergenbaum JH, Garcia-Closas M, et al. analysis. Oncogene 2005;24:4660–4671.
Etiology of hormone receptor-defined breast cancer: a 24 Tamimi RM, Baer HJ, Marotti J, et al. Comparison of
systematic review of the literature. Cancer Epidemiol molecular phenotypes of ductal carcinoma in situ and
Biomarkers Prev 2004;13:1558–1568. invasive breast cancer. Breast Cancer Res 2008;10:R67.
8 Gonzalez LO, Corte MD, Vazquez J, et al. Androgen 25 Nielsen TO, Hsu FD, Jensen K, et al. Immunohisto-
receptor expression in breast cancer: relationship with chemical and clinical characterization of the basal-like
clinicopathological characteristics of the tumors, prog- subtype of invasive breast carcinoma. Clin Cancer Res
nosis, and expression of metalloproteases and their 2004;10:5367–5374.
inhibitors. BMC Cancer 2008;8:149. 26 Abd El-Rehim DM, Ball G, Pinder SE, et al. High-
9 Isola JJ. Immunohistochemical demonstration of an- throughput protein expression analysis using tissue
drogen receptor in breast cancer and its relationship to microarray technology of a large well-characterised
other prognostic factors. J Pathol 1993;170:31–35. series identifies biologically distinct classes of breast
10 Kuenen-Boumeester V, Van der Kwast TH, van Putten cancer confirming recent cDNA expression analyses.
WL, et al. Immunohistochemical determination of Int J Cancer 2005;116:340–350.
androgen receptors in relation to oestrogen and 27 Abd El-Rehim DM, Pinder SE, Paish CE, et al.
progesterone receptors in female breast cancer. Int J Expression of luminal and basal cytokeratins in human
Cancer 1992;52:581–584. breast carcinoma. J Pathol 2004;203:661–671.
11 Niemeier LA, Dabbs DJ, Beriwal S, et al. Androgen 28 van de Rijn M, Perou CM, Tibshirani R, et al.
receptor in breast cancer: expression in estrogen Expression of cytokeratins 17 and 5 identifies a group
receptor-positive tumors and in estrogen receptor- of breast carcinomas with poor clinical outcome. Am J
negative tumors with apocrine differentiation. Mod Pathol 2002;161:1991–1996.
Pathol 2010;23:205–212. 29 Labrie F, Luu-The V, Labrie C, et al. Endocrine and
12 Moinfar F, Okcu M, Tsybrovskyy O, et al. Androgen intracrine sources of androgens in women: inhibition
receptors frequently are expressed in breast carcino- of breast cancer and other roles of androgens and their

Modern Pathology (2011) 24, 924–931

 Androgen receptor expression in breast cancer
LC Collins et al 931

precursor dehydroepiandrosterone. Endocr Rev 2003; 37 Naderi A, Hughes-Davies L. A functionally significant

24:152–182. cross-talk between androgen receptor and ErbB2 path-
30 Castellano I, Allia E, Accortanzo V, et al. Androgen ways in estrogen receptor negative breast cancer.
receptor expression is a significant prognostic factor in Neoplasia 2008;10:542–548.
estrogen receptor positive breast cancers. Breast Can- 38 Garreau JR, Muller P, Pommier R, et al. Transgenic
cer Res Treat 2010;124:607–617. introduction of androgen receptor into estrogen-recep-
31 Agoff SN, Swanson PE, Linden H, et al. Androgen tor-, progesterone-receptor-, and androgen-receptor-
receptor expression in estrogen receptor-negative breast negative breast cancer cells renders them responsive
cancer. Immunohistochemical, clinical, and prognostic to hormonal manipulation. Am J Surg 2006;191:
associations. Am J Clin Pathol 2003;120:725–731. 576–580.
32 Key T, Appleby P, Barnes I, et al. Endogenous sex 39 Nielsen TO, Parker JS, Leung S, et al. A comparison of
hormones and breast cancer in postmenopausal wo- PAM50 intrinsic subtyping with immunohistochemis-
men: reanalysis of nine prospective studies. J Natl try and clinical prognostic factors in tamoxifen-treated
Cancer Inst 2002;94:606–616. estrogen receptor-positive breast cancer. Clin Cancer
33 Wysowski DK, Comstock GW, Helsing KJ, et al. Sex Res 2010;16:5222–5232.
hormone levels in serum in relation to the develop- 40 Voduc KD, Cheang MC, Tyldesley S, et al. Breast
ment of breast cancer. Am J Epidemiol 1987;125: cancer subtypes and the risk of local and regional
791–799. relapse. J Clin Oncol 2010;28:1684–1691.
34 Hanley K, Wang J, Bourne P, et al. Lack of expression of 41 Blows FM, Driver KE, Schmidt MK, et al. Subtyping of
androgen receptor may play a critical role in transfor- breast cancer by immunohistochemistry to investigate
mation from in situ to invasive basal subtype of high- a relationship between subtype and short and long
grade ductal carcinoma of the breast. Hum Pathol term survival: a collaborative analysis of data for
2008;39:386–392. 10,159 cases from 12 studies. PLoS Med 2010;7:
35 Meijnen P, Peterse JL, Antonini N, et al. Immunohis- e1000279.
tochemical categorisation of ductal carcinoma in situ 42 Cheang MCU, Chia SK, Voduc D, et al. Ki67
of the breast. Br J Cancer 2008;98:137–142. index, HER2 status, and prognosis of patients with
36 Gucalp A, Traina TA. Triple-negative breast cancer: luminal B breast cancer. J Natl Cancer Inst 2009;101:
role of the androgen receptor. Cancer J 2010;16:62–65. 736–750.

Modern Pathology (2011) 24, 924–931