Gene expression of estrogen receptor, progesterone receptor and microtubule-associated protein Tau in high-risk early breast cancer: a quest for molecular predictors of treatment benefit in the context of a Hellenic Cooperative Oncology Group trial.

Abstract

Estrogen receptor (ER) and progesterone receptor (PgR) protein expression carry weak prognostic and moderate predictive utility for the outcome of early breast cancer patients on adjuvant chemohormonotherapy. We sought to study the predictive significance and correlations of transcriptional profiling of the ER, PgR and microtubule-associated protein Tau (MAP-Tau) genes in early breast cancer.
Messenger RNA (mRNA) was extracted from 279 formalin-fixed paraffin-embedded breast carcinomas (T1-3N0-1M0) of patients enrolled in the Hellenic Cooperative Oncology Group (HeCOG) trial HE 10/97, evaluating epirubicin-alkylator based adjuvant chemotherapy with or without paclitaxel (E-T-CMF versus E-CMF). Kinetic reverse transcription polymerase chain reaction (kRT-PCR) was applied for assessment of the expression of estrogen receptor, progesterone receptor and MAP-Tau genes in 274 evaluable patients. Cohort-based cut-offs were defined at the 25th percentile mRNA value for ER and PgR and the median for MAP-Tau.
Two hundred and ten patients (77%) were ER and/or PgR-positive by immunohistochemistry (IHC). Positive ER and MAP-Tau mRNA status was significantly associated with administration of hormonal therapy and low grade, while MAP-Tau mRNA status correlated with premenopausal patient status. MAP-Tau strongly correlated with ER and PgR mRNA status (Spearmann r = 0.52 and 0.64, P < 0.001). The observed chance corrected agreement between determination of hormonal receptor status by kRT-PCR and IHC was moderate (Kappa = 0.41) for ER and fair (Kappa = 0.33) for PgR. At a median follow-up of 8 years, univariate analysis adjusted for treatment showed positive ER mRNA status to be of borderline significance for reduced risk of relapse (HR = 0.65, 95% CI 0.41-1.01, P = 0.055) and death (HR = 0.62, 95% CI 0.36-1.05, P = 0.077), while positive MAP-Tau mRNA status was significantly associated with reduced risk of relapse (HR = 0.50, 95% CI 0.32-0.78, P = 0.002) and death (HR = 0.49, 95% CI 0.29-0.83, P = 0.008). In multivariate analysis, only axillary nodal metastases (HR = 2.33, 95% CI 1.05-5.16, P = 0.04) and MAP-Tau mRNA status (HR = 0.46, 95% CI 0.25-0.85, P = 0.01) independently predicted patient outcome. However, MAP-Tau mRNA levels did not predict enhanced benefit from inclusion of paclitaxel in the adjuvant chemotherapy regimen (test for interaction P = 0.99). No correlation was evident between increasing ER and PgR mRNA transcription and increasing benefit from endocrine therapy in 203 ER and/or PgR IHC-positive patients receiving adjuvant hormone therapy (Wald P = 0.54 for ER, 0.51 for PR).
ER gene transcription carries weak predictive significance for benefit from endocrine therapy or for outcome, with no apparent dose-response association. The predictive significance is possibly exerted via MAP-Tau gene expression, an ER-inducible tubulin modulator with strong predictive significance for patient outcome. However, MAP-Tau mRNA did not predict benefit from the addition of a taxane to adjuvant chemotherapy. Further study of the biologic function and utility of MAP-Tau for individualising adjuvant therapy is warranted.

Full text

Page 1

CLINICAL TRIAL
Gene expression of estrogen receptor, progesterone receptor
and microtubule-associated protein Tau in high-risk early breast
cancer: a quest for molecular predictors of treatment benefit
in the context of a Hellenic Cooperative Oncology Group trial
George Pentheroudakis Æ Æ Konstantine T. Kalogeras Æ Æ Ralph M. Wirtz Æ Æ
Irene Grimani Æ Æ George Zografos Æ Æ Helen Gogas Æ Æ Udo Stropp Æ Æ Dimitrios Pectasides Æ Æ
Dimosthenis Skarlos Æ Æ Guido Hennig Æ Æ Epaminondas Samantas Æ Æ Dimitrios Bafaloukos Æ Æ
Pavlos Papakostas Æ Æ Haralabos P. Kalofonos Æ Æ Nicholas Pavlidis Æ Æ George Fountzilas
Received: 13 April 2008/Accepted: 18 July 2008/Published online: 31 July 2008
? Springer Science+Business Media, LLC. 2008
Abstract
progesterone receptor (PgR) protein expression carry
weak prognostic and moderate predictive utility for the
outcome of early breast cancer patients on adjuvant
chemohormonotherapy. We sought to study the predictive
significance and correlations of transcriptional profiling of
the ER, PgRand microtubule-associatedprotein Tau (MAP-
Tau) genes in early breast cancer. Materials and methods
Background
Estrogen receptor(ER)and
Messenger RNA (mRNA)was extracted from279formalin-
fixed paraffin-embedded breast carcinomas (T1-3N0-1M0)
of patients enrolled in the Hellenic Cooperative Oncology
Group (HeCOG) trial HE 10/97, evaluating epirubicin-al-
kylator based adjuvant chemotherapy with or without
paclitaxel (E-T-CMF versus E-CMF). Kinetic reverse tran-
scriptionpolymerasechainreaction(kRT-PCR)wasapplied
for assessment of the expression of estrogen receptor, pro-
gesterone receptor and MAP-Tau genes in 274 evaluable
patients. Cohort-based cut-offs were defined at the 25th
percentile mRNA value for ER and PgR and the median for
MAP-Tau.ResultsTwohundredandtenpatients(77%)were
ER and/or PgR-positive by immunohistochemistry (IHC).
An invited commentary to this article can be found at
doi:10.1007/s10549-008-0176-1.
George Pentheroudakis and Konstantine T. Kalogeras have
contributed equally to this work.
G. Pentheroudakis (&) ? N. Pavlidis
Department of Medical Oncology, Ioannina University Hospital,
Niarxou Avenue, 45500 Ioannina, Greece
e-mail: gpenther@otenet.gr
K. T. Kalogeras ? G. Fountzilas
Department of Medical Oncology, Papageorgiou Hospital,
Aristotle University of Thessaloniki School of Medicine,
Thessaloniki, Greece
K. T. Kalogeras ? I. Grimani
Hellenic Cooperative Oncology Group, Data Office, Athens,
Greece
R. M. Wirtz ? U. Stropp ? G. Hennig
Siemens Healthcare Diagnostics, Cologne, Germany
G. Zografos
Department of Surgery, University of Athens School
of Medicine, Athens, Greece
H. Gogas
Laiko General Hospital, University of Athens School
of Medicine, Athens, Greece
D. Pectasides
Second Department of Internal Medicine, ‘‘Attikon’’ University
Hospital, Athens, Greece
D. Skarlos
Second Department of Medical Oncology, Metropolitan
Hospital, Piraeus, Greece
E. Samantas
Third Department of Medical Oncology, Agii Anargiri Cancer
Hospital, Athens, Greece
D. Bafaloukos
First Department of Medical Oncology, Metropolitan Hospital,
Piraeus, Greece
P. Papakostas
Oncology Department, Hippokration Hospital, Athens, Greece
H. P. Kalofonos
Department of Medicine, University Hospital, Patras Medical
School, Rion, Greece
123
Breast Cancer Res Treat (2009) 116:131–143
DOI 10.1007/s10549-008-0144-9

Page 2

Positive ER and MAP-Tau mRNA status was significantly
associated with administration of hormonal therapy and low
grade, while MAP-Tau mRNA status correlated with pre-
menopausal patient status. MAP-Tau strongly correlated
with ER and PgR mRNA status (Spearmann r = 0.52 and
0.64, P\0.001). The observed chance corrected agreement
between determination of hormonal receptor status by kRT-
PCRandIHCwas moderate (Kappa = 0.41)forERandfair
(Kappa = 0.33) for PgR. At a median follow-up of 8 years,
univariate analysis adjusted for treatment showed positive
ERmRNAstatustobeofborderlinesignificanceforreduced
risk of relapse (HR = 0.65, 95% CI 0.41–1.01, P = 0.055)
and death (HR = 0.62, 95% CI 0.36–1.05, P = 0.077),
while positive MAP-Tau mRNA status was significantly
associated with reduced risk of relapse (HR = 0.50, 95% CI
0.32–0.78,P = 0.002)anddeath(HR = 0.49,95%CI0.29–
0.83, P = 0.008). In multivariate analysis, only axillary
nodalmetastases(HR = 2.33,95%CI1.05–5.16,P = 0.04)
and MAP-Tau mRNA status (HR = 0.46, 95% CI 0.25–
0.85, P = 0.01) independently predicted patient outcome.
However, MAP-Tau mRNA levels did not predict enhanced
benefit from inclusion of paclitaxel in the adjuvant chemo-
therapy regimen (test for interaction P = 0.99). No
correlation was evident between increasing ER and PgR
mRNA transcription and increasing benefit from endocrine
therapy in 203 ER and/or PgR IHC-positive patients
receivingadjuvanthormonetherapy(WaldP = 0.54forER,
0.51 for PR). Conclusions ER gene transcription carries
weak predictive significance for benefit from endocrine
therapy or for outcome, with no apparent dose-response
association. The predictive significance is possibly exerted
via MAP-Tau gene expression, an ER-inducible tubulin
modulator with strong predictive significance for patient
outcome.However,MAP-TaumRNAdidnotpredictbenefit
from the addition of a taxane to adjuvant chemotherapy.
Further study of the biologic function and utility of MAP-
Tau for individualising adjuvant therapy is warranted.
Keywords
Microtubule ? MAP-Tau ? Predictive value ? Breast cancer
Estrogen receptor ? Progesterone receptor ?
Introduction
Breast adenocarcinoma is the most common malignant
tumor in females with 60–70% of affected patients pre-
senting with localised disease. These women become
disease-free following surgical resection of the tumor and
may be candidates for adjuvant systemic therapy aiming at
eradication of micrometastatic deposits [1]. Estrogen
receptor (ER) protein expression, studied by means of
immunohistochemical (IHC) staining, is the gold standard
for the selection of patients who will be managed with
hormonal therapy, though its predictive value for benefit
from such treatment is modest (30–60%) [2, 3]. Several
clinicopathological parameters, including tumor size,
grade, ER status, axillary nodal status, patient age, and
Human Epidermal growth factor Receptor-type 2 (HER2)
overexpression are taken into consideration in order to
identify which patients are most likely to benefit from
administration of adjuvant cytotoxic chemotherapy and
which drugs should be used [4, 5]. Still, the power of these
complex predictive models is at best moderate: patients
may be undertreated or overtreated. Ultimately, we are
unable to reliably predict the true binary outcome of
recurrence versus no recurrence at the level of the indi-
vidual, as these clinicopathological guidelines do not
capture the rich molecular heterogeneity of the malignancy
and are by no means able to pinpoint the exact clinical
behaviour of breast tumors [6].
Even when conventional IHC studies were used, evi-
dence emerged that incremental ER protein expression was
associated with higher probability of benefit from endocrine
therapy [7]. However, (IHC) studies show considerable
intra- and inter-laboratory variability even when reporting
binary results, due to variation in fixation time, antigen
retrieval and staining methods, observer subjectivity or
image acquisition threshold differences [8–10]. Accord-
ingly, efforts at quantification of ER and progesterone
receptor (PgR) protein expression face problems of tech-
nical feasibility and reproducibility. The advent of robust,
sensitive and reproducible molecular techniques that reli-
ably quantify expression of one or several genes offers
promise for more accurate quantification of ER and PgR.
Reverse-transcriptase polymerase chain reaction (RT-PCR)
techniques analysing messenger RNA (mRNA) extracted
from formalin-fixedparaffin-embedded
recently been developed, able to provide normalised gene
expression data over a broad dynamic range. Preliminary
reports suggest satisfactory correlation of quantitative
mRNA results with categories of IHC protein expression
[11–13]. Still, the prognostic and predictive value of the
continuous range of tumoral ER and PgR gene transcription
data for patient survival and benefit from therapy has not
been examined to date.
Tau is a microtubule-associated protein (MAP-Tau) that
stabilises tubulin polymerisation. Early preclinical and
clinical reports suggest that the MAP-Tau gene is induced
by ER and that expression of Tau protein modulates sen-
sitivity of malignant cells to taxanes [14]. Low levels of
Tau protein expression render mitotic and cytoskeletal
microtubules more sensitive to disruption by paclitaxel,
affecting mitosis and cell signalling. We retrospectively
examined gene expression of ER, PgR and MAP-Tau in
high-risk localised breast carcinomas of women enrolled in
tumors have
132Breast Cancer Res Treat (2009) 116:131–143
123

Page 3

a prospective, randomised Hellenic Cooperative Oncology
Group (HeCOG) trial of adjuvant chemotherapy in order to
study their correlation and concordance with IHC expres-
sion of the relevant proteins as well as their utility for
predicting clinical benefit from adjuvant hormonal therapy
and taxane-containing chemotherapy.
Materials and methods
Clinical trial
TheHeCOGprospectivetrialHE10/97randomisedatotalof
595 high-risk (T1-3N1M0 or T3N0M0) breast cancer
patients in the period 1997–2000, to receive either four
cycles of epirubicin followed by four cycles of intensified
cyclophosphamide, methotrexate and 5-fluorouracil combi-
nation chemotherapy (E-CMF) or three cycles of epirubicin
followed by three cycles of paclitaxel and three cycles of
intensified CMF (E-T-CMF) [15]. Chemotherapy cycles
were administered every 2 weeks and patients received
granulocyte-colony stimulating factor (GCSF) support. The
clinical protocol and translational research studies were
approvedbytheHeCOGProtocolReviewCommittee,bythe
InstitutionalReviewBoardofKyanousStavrosHospitaland
the AHEPA University Hospital. All patients provided
written informed consent prior to enrolment.
Molecular and immunohistochemical studies
This was a retrospective translational research study among
patients who had been enrolled in a prospective clinical
trial. Accordingly, collection of formalin-fixed paraffin-
embedded (FFPE) tumor tissue samples was possible in
279 patients only, due to logistical/organisational barriers.
An experimental method based on proprietary magnetic
beads from Siemens Healthcare Diagnostics (Cologne,
Germany) resulted in mRNA isolation from 274 FFPE
blocks. For all tumor samples included in the analysis the
number of malignant cells represented at least 75% of all
nucleated cells per section as verified by hematoxylin-eosin
staining. Kinetic reverse transcription polymerase chain
reaction (kRT-PCR) was applied for assessment of the
expression of estrogen receptor (ESR1), progesterone
receptor and MAP-Tau genes using gene specific Taq-
ManTMbased Primer/Probe sets. Forty cycles of nucleic
acid amplification were applied and the cycle threshold
(CT) values of the target genes were identified. CT values
were normalized by subtracting the CT value of the
housekeeping gene RPL37A from the CT value of the
target gene (DCT). RNA results were then reported as 40-
DCT values, which would correlate proportionally to the
mRNA expression level of the target gene.
In short, each FFPE slide (5 lm thick) was deparaffi-
nized in xylol and ethanol, the pellet was washed with
ethanol and dried at 55?C for 10 min. The pellet was then
lysed and proteinised overnight at 55?C with shaking. After
adding a binding buffer and the magnetic particles (Sie-
mens Healthcare Diagnostics, Cologne, Germany) nucleic
acids were bound to the particles within 15 min at room
temperature. On a magnetic stand, the supernatant was
aspirated and the beads were washed several times with a
washing buffer. After adding elution buffer and incubating
for 10 min at 70?C the supernatant was aspirated on a
magnetic stand without touching the beads. After normal
DNAse I treatment for 30 min at 37?C and inactivation of
DNAse I the solution was used for RT-PCR. The Primer/
Probe sets used for amplification of the target genes were
the following:
ER Probe ATGCCCTTTTGCCGATGCA
Forward Primer GCCAAATTGTGTTTGATGGAT
TAA
Reverse Primer GACAAAACCGAGTCACATCAG
TAATAG
PgR Probe TTGATAGAAACGCTGTGAGCTCGA
Forward Primer AGCTCATCAAGGCAATTGGTTT
Reverse Primer ACAAGATCATGCAAGTTATCAA
GAAGTT
MAP-Tau Probe AGACTATTTGCACACTGCCGCCT
Forward Primer GTGGCTCAAAGGATAATATC
AAACAC
Reverse Primer ACCTTGCTCAGGTCAACTGGTT
The quality and quantity of RNA was checked by
measuring absorbance at 260 and 280 nm. Pure RNA has
an A260/A280 ratio of 1.9–2.0. Human reference total
RNA pooled from ten human cell lines (Stratagene, La
Jolla, California, USA) was used as a positive control.
RNA-free DNA extracted from tumor tissues was used as a
negative control.
ER and PgR protein expression was evaluated immu-
nohistochemically in the pathology department of the
different centers involved in the study, by means of
streptavidin–biotin complex immunoperoxidase assay on 5
lm sections from FFPE tumor tissue. Tumor samples were
categorised as ER or PgR-positive if at least 1% of
malignant cells stained for ER or PgR protein, respectively.
Statistical methodology
In the absence of a reliable gold standard and following
distributional studies, we used the 25th percentile of
observed hormonal receptor mRNA expression levels and
the median of observed MAP-Tau mRNA levels as
thresholds for categorization of tumors to positive or
negative cases. The chosen cut-offs were found to be
Breast Cancer Res Treat (2009) 116:131–143 133
123

Page 4

close to the natural cut-offs in distributional studies
(Fig. 1a–c). Moreover, cut-offs on the quartiles offer
easily interpretable, reproducible and objective results.
Fisher’s exact test was applied to compare standard
clinical and pathological factors between ER, PgR and
MAP-Tau mRNA levels. Spearman’s rank correlation
was used as a measure of association between the three
biomarkers. The Mann–Whitney test was used to com-
pare MAP-Tau levels with ER and PgR expression
levels.
Overall percentages of agreement, uncorrected for
agreement by chance, were calculated and given along with
the 95% confidence intervals. The degree of agreement
corrected by chance was estimated using the Kappa coef-
ficient. Kappa estimations were interpreted following the
guidelines described by Landis and Koch [16]: j B 0, poor
agreement; j[0 but B0.20, slight agreement; j[0.20 but
B0.40, fair agreement; j [ 0.40 but B0.60, moderate
agreement; j[0.60 but B0.80, substantial agreement; and
j[0.80–1.00, almost perfect agreement.
Disease-free survival (DFS) was defined as the interval
from study entry to first locoregional recurrence, first
distant metastasis, contralateral breast cancer, secondary
neoplasm, death from the disease, or death from any
cause non-related to breast cancer, whichever occurred
first [17]. Overall survival (OS) was measured from study
entry until death from any cause. Surviving patients were
censored at the date of last contact. Estimation of disease-
free and overall survival was performed using the Kap-
lan–Meier product-limit method and differences between
survival curves were assessed with the log-rank test. The
Cox proportional hazards model was used to assess the
strength of the association of OS and DFS with various
clinical and histological variables in the presence of
treatmentgroup, ER, PgR,
expression levels. A backward selection procedure with
removal criterion P [ 0.10, identified a subclass of sig-
nificant variables among the following: menopausal status
(pre versus post), tumor grade (I–II versus III-Undiffer-
entiated), tumor size (B2 cm vs. 2–5 cm vs.[5 cm), and
number of positive lymph nodes (0–3 vs. C4). Multivar-
iate P values were based on Wald statistics. Results of
this study were presented according to reporting recom-
mendations for tumor marker prognostic studies [18].
andMAP-Tau mRNA
Fig. 1 Distribution of ER (a),
PgR (b) and MAP-Tau (c)
mRNA levels in the studied
population. Solid vertical lines
on the graphs correspond to the
chosen cut-offs (25th percentile
for ER and PgR and the median
for MAP-Tau) and dashed lines
to the natural cut-offs as
determined by the
corresponding distributions
134Breast Cancer Res Treat (2009) 116:131–143
123

Page 5

Results
Patient and tumor characteristics
Patients included in the present study were comparable in
terms of basic characteristics (age, menopausal status, ER
status, PgR status, proportion of patients receiving hor-
monal treatment, number of positive nodes and tumor size)
with patients from the HE10/97 trial that were not included.
Two-hundred and seventy-four patients (median age 51)
harboured breast carcinomas, most commonly hormone-
receptor positive by IHC (77% of patients), T2-3 (70%) of
moderate to poor differentiation (96%), from which neo-
plastic FFPE material underwent molecular analysis for
gene expression of ER, PgR and MAP-Tau. After modified
radical mastectomy or breast conserving surgery and nodal
dissection, malignant metastases in axillary lymph nodes
were present in all but four cases, with presence of four or
more involved nodes in 77% of patients. Slightly more
patients were managed with adjuvant E-CMF chemother-
apy than E-T-CMF. With the exception of grade, basic
clinicopathological characteristics were well balanced
when the 274 analysed patients were stratified by adjuvant
chemotherapy arm (Table 1). Following the completion of
adjuvant chemotherapy, 92% of the 274 women received
some form of hormonal therapy, mostly tamoxifen. Among
women with IHC ER-negative tumors who took tamoxifen,
a proportion had PgR-unknown or PgR-positive tumors,
while some received hormonal therapy despite having both
ER and PgR-negative tumors, a practice not followed
anymore in HeCOG participating centres.
Correlation of molecular and clinicopathological
parameters
Although mRNA isolation was successful in 274 out of 279
tumor blocks, the study of several other genes in com-
panion projects resulted in availability of high-quality
mRNA in only 271 (MAP-Tau) and 263 (ER, PgR) blocks
respectively. The correlations of ER, PgR and MAP-Tau
mRNA status (positive or negative) with standard clinical
and pathologic factors are shown in Table 2. The associ-
ations of ER mRNA and MAP-Tau mRNA status with
administration of adjuvant hormonal therapy and histo-
logical grade were statistically significant. Histological
grade was also significantly associated to PgR mRNA
status. Moreover, MAP-Tau mRNA status was signifi-
cantly associatedwithpatient
Specifically, patients with negative ER, PgR or MAP-Tau
mRNA levels had significantly higher grade (ER negative:
78% vs. positive: 41%, P\0.001; PgR negative: 67% vs.
positive: 45%, P = 0.003; MAP-Tau negative: 64% vs.
positive: 36%, P\0.001). Significantly more patients with
menopausal status.
positive ER and MAP-Tau mRNA levels received adjuvant
hormonal therapy (ER negative: 85% vs. positive: 95%,
P = 0.012; MAP-Tau negative: 85% vs. positive: 98%,
P \ 0.001). Premenopausal patients were significantly
more numerous in patients with positive MAP-Tau mRNA
levels (MAP-Tau negative: 43% vs. positive: 58.5%,
P = 0.015). On the contrary, mRNA status of any of ER,
PgR and MAP-Tau did not exhibit significant correlations
with patient age, adjuvant chemotherapy arm (E-CMF
versus E-T-CMF), administration of adjuvant radiotherapy,
tumor size, and number of involved axillary lymph nodes.
Tumoral MAP-Tau mRNA status was significantly
associated with both ER (Spearmann rank correlation
r = 0.52, P \ 0.001) and PgR mRNA status (Spearmann
rank correlation r = 0.64, P \ 0.001). In fact, positive
MAP-Tau mRNA expression was seen in 67% of cases
with positive ER mRNA expression versus only 3% of
cases with negative ER mRNA expression (Fisher exact
test, P \ 0.001). Similarly, positive MAP-Tau mRNA
expression was observed in 61% of cases with positive PgR
mRNA expression versus 20% of cases with negative PgR
mRNA status (Fisher exact test, P\0.001). The distribu-
tion of tumoral MAP-Tau mRNA levels in accordance to
ER mRNA status is shown in Fig. 2.
Concordance between immunohistochemical
and molecular hormonal receptor studies
Two hundred and ten patients (77%) were ER and/or PgR-
positive by IHC. The observed overall agreement between
the determination of hormonal receptor status by kRT-PCR
(mRNA) and IHC was 77% (95% confidence interval (CI):
71.5%–82%) for ER and 71% (95% CI: 65%–79%) for
PgR. Among ER-mRNA positive tumors, 82.5% stained
positively for the ER protein whereas 17.5% did not.
Among ER-mRNA negative tumors, 61% were IHC neg-
ative but 39% stained positively for the ER protein. These
rates of concordance and disagreement between the RT-
PCR and IHC methodologies were quite similar for the
PgR parameter as well: among PgR-mRNA positive
tumors, 26% exhibited negative protein staining, while
among PgR-mRNA negative tumors 37.5% stained posi-
tively for the PgR protein.
Despite the good observed agreement between the two
methods, the Kappa statistic (chance-corrected measure of
agreement) between kRT-PCR and IHC was 0.41 (95% CI:
0.29–0.54) for ER, a marker of moderate agreement, and
0.33 (95% CI: 0.20–0.45) for PgR, a marker of fair agree-
ment between methodologies. Using the 25th percentile of
the ER and PgR mRNA values as the threshold and con-
sidering the more widely used IHC protein expression as the
benchmark, the relatively new methodology of kRT-PCR
had satisfactory sensitivity (85.5–86.5%) and positive
Breast Cancer Res Treat (2009) 116:131–143135
123

Page 6

Table 1 Patient and tumor
characteristics
aDistribution of grade differs
significantly (P\0.001). This
result is identical to the
corresponding result in the
clinical paper
N (%)
P
N = 274 E-T-CMF
N = 121
E-CMF
N = 153
Age (range)51 (22.5–76) 52 (28–76) 51 (22–76)0.873
Menopausal status0.274
Premenopausal140 (51) 57 (47)83 (54)
Postmenopausal 134 (49) 64 (53)70 (46)
Type of operation 0.758
MRM 221 (81) 99 (82)122 (80)
Breast conserving surgery53 (19)22 (18)31 (20)
Interval from operation (weeks)
\2
2–4
[4
Adjuvant HT
0.116
43 (16)17 (14)26 (17)
128 (47) 65 (54)63 (41)
103 (38) 39 (32)64 (42)
0.387
No 23 (8)8 (7) 15 (10)
Yes251 (92) 113 (93) 138 (90)
Tamoxifen218 (87)100 (88.5) 118 (86) 0.575
LH-RH107 (43) 49 (43) 58 (42)0.898
Aromatase inhibitors11 (4)5 (4)8 (6)0.777
Other3 (1) 1 (1)2 (1) 0.999
Unknown 3 (1)2 (2)1 (1) 0.590
Adjuvant RT 0.765
No 57 (21)24 (20)33 (22)
Yes 214 (78)96 (79)118 (77)
Unknown3 (1)1 (1) 2 (1)
Estrogen receptor status (IHC)0.589
Negative 78 (28.5)32 (26) 46 (30)
Positive191 (70) 87 (72)104 (68)
Unknown5 (2) 2 (2)3 (2)
Progesterone receptor status (IHC)0.999
Negative95 (36) 42 (35) 53 (35)
Positive 171 (62) 75 (62)96 (63)
Unknown 8 (3)4 (3)4 (3)
Tumor size (cm)0.83
B2 83 (30)35 (29)48 (31)
2–5
[5
137 (50) 63 (52) 74 (48)
54 (20)23 (19) 31 (20)
Gradea
0.001
I 11 (4) 8 (7)3 (2)
II 125 (46)41 (34)84 (55)
III135 (49)69 (57) 66 (43)
Undifferentiated2 (1)2 (2) 0 (0)
Missing1 (0.4) 1 (1)0 (0)
Nodes removed
Median (range) 20 (4–59)20 (5–59)20 (4–53) 0.950
Positive nodes
Median (range)6 (0–54) 7 (0–54)6 (0–49)0.463
0–363 (23)24 (20)39 (25.5) 0.312
4 or more211 (77) 97 (80)114 (74.5)
136 Breast Cancer Res Treat (2009) 116:131–143
123

Page 7

predictive
(45–53%) and negative predictive value (61–62.5%). Met-
rics of concordance between the two methods for ER and
PgR determination are summarised in Table 3.
value(74–82.5%) butmodest specificity
Clinicopathological/molecular variables and patient
outcome
Survival status of the 274 patients was updated in October
2007. At the time of analysis (February 2008), the median
follow up time was 8 years. During this time, 90 patients
(33%) had developed a relapse and 62 patients (23%) had
died. Malignant relapses occurred in 43% and deaths in
31% of the ER-mRNA negative cases versus 30% relapsed
and 21% dead patients with ER-mRNA positive tumors.
Similarly, malignant relapses were seen in 41% and deaths
in 29% of patients harbouring MAP-Tau negative tumors
Table 2 Relationship between mRNA ER, PgR and MAP-Tau levels and standard clinical and pathological factors
ER mRNA level (N = 263) PgR mRNA level (N = 263)MAP-Tau mRNA level (N = 271)
NegativePositive
P-valueNegativePositive
P-valueNegativePositive
P-value
Age
Median 5052 0.31 51510.90 52500.27
Range27.5–74 22.5–7622.5–7427–7622.5–74.527–76
N (%)
N (%)
P-value
N (%)
N (%)
P-value
N (%)
N (%)
P-value
Menopausal status
Premenopausal34 (52)101 (51) 0.8931 (47)104 (53) 0.4859 (43)79 (58.5)
0.015
Postmenopausal31 (48)97 (49) 35 (53)93 (47)77 (57) 56 (41.5)
Treatment group
E-T-CMF 32 (49)89 (45)0.57 26 (39) 95 (48)0.25 59 (43)59 (44)1.0
E-CMF 33 (51)109 (55) 40 (61)102 (52) 77 (57) 76 (56)
Adjuvant HT
No 10 (15)10 (5)
0.012
8 (12)12 (6) 0.11 20 (15)3 (2)
\0.001
Yes 55 (85)188 (95) 58 (88)185 (94) 116 (85)132 (98)
Adjuvant RT
No 17 (27)37 (19) 0.2117 (27) 37 (19)0.21 33 (25)24 (18)0.18
Yes46 (73) 160 (81)47 (73)159 (81) 100 (75)111 (82)
Italic values denote relationships statistically significant at P\0.05
Fig. 2 MAP-Tau mRNA levels in ER mRNA positive and negative
cases (Mann–Whitney test, P\0.001)
Table 3 Agreement between kRT-PCR and IHC status of ER, PgR
and ER/PgR
ER
(%)
PgR
(%)
ER/PgR
(%)
Observed agreementa
77 7182
Agreement expected by chance
Kappa coefficientb
6157.5 69
4133 41.5
kRT-PCR
Sensitivity86.585.5 91
Specificity53 45 47
PPV82.5 7486.5
NPV6162.5 59
aProportion of agreement observed between IHC and kRT-PCR
methods
bKappa coefficients are expressed as percentages
NPV, negative predictive value; PPV, positive predictive value
The third column ER/PgR represents hormone receptor status (any of
the two receptors ER, PgR positive versus both ER and PgR negative)
Breast Cancer Res Treat (2009) 116:131–143 137
123

Page 8

versus 23% relapses and 16% deaths among those with
MAP-Tau positive carcinomas. In contrast, 33% of PgR-
mRNA negative cases suffered a relapse and 21% died
versus 33.5% relapsed and 24% dead patients with PgR-
mRNA positive tumors.
Univariate Cox regression analysis, adjusted for treat-
ment, showed a non-significant trend for decreased risk of
death in patients with ER-mRNA positive tumors (hazard
ratio (HR) = 0.62, 95% CI: 0.36–1.05, Wald-P = 0.077),
as well as a trend towards a decreased risk of malignant
relapse (HR = 0.65, 95% CI: 0.41–1.01, Wald-P = 0.06),
in comparison to those affected by ER-mRNA negative
breast carcinomas. In contrast, the status of tumoral PgR
mRNA did not carry any predictive significance for either
risk of death or malignant relapse. Patients with MAP-Tau
mRNA positive tumors had a significantly decreased risk of
death compared with those with MAP-Tau mRNA negative
tumors (HR = 0.49, 95% CI: 0.29–0.83, Wald-P = 0.008)
and a significantly decreased risk of disease relapse
(HR = 0.50, 95% CI: 0.32–0.78, Wald-P = 0.002). Kap-
lan–Meier survival curves representing the probability of
survival and disease-free survival as a function of MAP-
Tau mRNA status are presented in Fig. 3 (log-rank test for
DFS P = 0.002, for OS P = 0.006). However, neither
MAP-Tau mRNA nor ER mRNA expression status were
predictive for benefit from adjuvant treatment with paclit-
axel (E-T-CMF arm) for either disease-free survival or
overall survival (DFS and OS tests for interaction Wald
P = 0.65 and P = 0.73 for MAP-Tau, P = 0.32 and
P = 0.99 for ER, respectively). Restricting analysis to ER
positive patients, MAP-Tau mRNA expression status
remained non-predictive for benefit from adjuvant treat-
ment with paclitaxel (DFS and OS tests for interaction
Wald P = 0.50 and P = 0.88 for MAP-Tau among ER
positive patients).
The impact of paclitaxel on OS and DFS in MAP-Tau
negative and MAP-Tau positive patients are shown in
Fig. 4a and b, respectively. Same analysis among ER posi-
tive patients is presented in Figs. 5a and 5b, respectively.
Multivariate Cox regression analysis (Table 4), includ-
ing menopausal status, tumor grade, size, number of
positive lymph nodes, treatment group, ER, PgR, and
MAP-Tau mRNA expression levels, revealed that only
MAP-Tau and number of involved axillary nodes remained
significant independent predictors of OS, while MAP-Tau,
number of positive nodes and tumor grade independently
predicted DFS. More specifically, positive MAP-Tau was
associated with a decreased risk of death (HR = 0.46, 95%
CI: 0.25–0.85, Wald-P = 0.01). As expected, four or more
positive nodes were associated with an increased risk of
death (HR = 2.33, 95% CI: 1.05–5.16, Wald-P = 0.04).
Regarding DFS, positive MAP-Tau mRNA status was
associated with a decreased risk of relapse (HR = 0.53,
95% CI: 0.32–0.89, Wald-P = 0.02), while four or more
positive nodes (HR = 2.30, 95% CI: 1.21–4.38, Wald-
P = 0.01) and higher tumor grade (HR = 1.56, 95% CI:
0.98–2.50, Wald-P = 0.06) significantly increased the risk
of malignant recurrence. In both multivariate models
described above, MAP-Tau mRNA status was not predic-
tive for the benefit of adjuvant treatment with paclitaxel
either for OS or for DFS (test for interaction: Wald-
P = 0.99 and Wald-P = 0.92, respectively).
As one of the advantages of quantitative measurement of
hormonal receptor mRNA expression is a continuous
dynamic range of output, we sought to determine whether
increasing levels of hormone receptor mRNA correlated
Fig. 3 Overall survival (OS) (a) (log rank P = 0.006) and DFS (b)
(log rank P = 0.002) by MAP-Tau mRNA expression
138Breast Cancer Res Treat (2009) 116:131–143
123

Page 9

with increasing benefit from adjuvant hormonal therapy in
a ‘‘dose-dependent’’ manner. We tested this hypothesis in
203 patients harbouring breast carcinomas immunohisto-
chemically positive for ER and/or PgR, who received
adjuvant hormonal manipulations. Data were split into four
equal groups, using the quartiles as cut off points.
Using Cox regression analysis, we examined the asso-
ciation of very low (ER, PgR mRNA expression B first
quartile), low (first quartile\ER, PgR mRNA expression
B second quartile), moderate (second quartile\ER, PgR
mRNA expression B third quartile), and high (ER, PgR
mRNA expression [ third quartile) mRNA levels of ER
and PgR with disease-free survival. There was no evidence
of association between increasing ER or PgR mRNA
expression and benefit from hormonal therapy (Wald P
values adjusted for treatment group: P = 0.54 for ER
mRNA and P = 0.505 for PgR mRNA). In comparison to
tumors with very low ER mRNA expression, the hazard
ratios for malignant recurrence in tumors with low, mod-
erate and high expression were 1.08 (95% CI: 0.51–2.31,
Wald-P = 0.835), 1.61(95%
P = 0.19) and 1.12 (95% CI: 0.52–2.39, Wald-P = 0.77),
CI:0.79–3.29, Wald-
respectively. Similarly, in comparison to tumors with very
low PgR mRNA expression, the hazard ratios for relapse in
patients with low, moderate and high mRNA-expressing
tumors were 1.06 (95% CI: 0.54–2.10, Wald-P = 0.86),
0.61 (95% CI: 0.28–1.32, Wald-P = 0.21) and 0.97 (95%
CI: 0.48–1.97, Wald-P = 0.93), respectively.
Discussion
In our patient cohort, the concordance between immuno-
histochemical and kRT-PCR determination of ER and PgR
status was good (77% for ER, 71% for PgR) but clearly
inferior to rates of agreement reported by Badve et al. (88–
91%), Bianchini et al. (94%) and Gong et al. (88–90%)
[11–13]. After excluding chance agreement, as estimated
by the Kappa coefficient, only fair to moderate agreement
was found between IHC and RT-PCR determination of ER
and PgR. However, in most studies the investigators
actually intended to identify those kRT-PCR mRNA cut-
off values that would provide the highest concordance with
the IHC status of the target protein.
Fig. 4 OS (left) and DFS
(right) for E-T-CMF (red line)
and E-CMF (blue line) in MAP-
Tau negative and MAP-Tau
positive patients
Breast Cancer Res Treat (2009) 116:131–143 139
123

Page 10

On the contrary, we regarded IHC ER and PgR protein
expression determination as the most widely used analyti-
cal assay, but not the gold standard, in view of its moderate
predictive value for benefit from endocrine therapies, semi-
quantitative nature of results and inter- or intra-laboratory
reporting variability [2, 3, 8–10]. Using distributional
studies of mRNA values and taking into account hormonal
receptor protein staining as a gross template, we sought to
study the predictive value and correlation of a more con-
servative threshold, the 25th percentile of mRNA values, in
order not to miss an effect of low, but potentially biolog-
ically significant level of messenger RNA.
With the use of this threshold, among patients with
hormone receptor mRNA-positive tumors, 17–25% lacked
immunohistochemical protein staining. This observation
could be the result of mRNA failing to translate to protein
(alternate splicing, aberrant polyadenylation, microRNA-
mediated or other form of translational silencing) or rapid
catabolism of the hormone receptor protein. Protein deg-
radation in the tumor block by inadequate fixation,
problematic antigen retrieval, protein staining underesti-
mationin oldarchival material,orfinally, RNA
amplification of normal tissue along with the tumor could
also be the culprits for absence of protein staining [9, 19–
21]. Conversely, when tumoral hormone receptor mRNA
was categorised as negative, immunohistochemical protein
staining was evident in at least 1% of tumor cells in 37–
39% of cases. This discrepancy could reflect the difficulty
capturing short-lived, sensitive to degradation mRNA from
FFPE tumor blocks, as well as the different fixation and
turnover times of ribonucleic acids in relation to proteins
[22, 23]. Accumulation of the protein due to aberrant
catabolism may also be a factor.
Despite the moderate at best agreement between IHC
and kRT-PCR, the ER mRNA status, according to the
objective threshold values we set, did show a trend for
predictive significance. Estrogen receptor protein expres-
sion was shown to be a weak prognostic factor for survival
and a moderate predictive factor for benefit from hormonal
manipulations in patients with early breast cancer [1]. Its
prognostic information may be carried by a multi-gene
signature, for which the ER gene itself is either a key
regulator or simply the best-known surrogate marker.
Pusztai et al. recently used DNA microarrays in order to
Fig. 5 OS (left) and DFS
(right) for E-T-CMF (red line)
and E-CMF (blue line) in MAP-
Tau negative and MAP-Tau
positive patients restricted to ER
positive patients
140Breast Cancer Res Treat (2009) 116:131–143
123

Page 11

profile mRNA transcripts from snap-frozen breast carci-
nomas of 38 patients [24]. They reported that ER positive
tumors, as assessed by IHC were strongly associated with
transcriptional activity of 105 genes involved in regulation
of transcription, cellular proliferation, differentiation,
intracellular trafficking of proteins, and signal transduction,
in contrast to ER negative tumors. Most of these genes are
not known to be regulated by ER, suggesting that differ-
ences between IHC ER positive and negative tumors are
not attributable to the presence or absence of ER function,
but rather reflect different molecular phenotypes. These
findings were supported by identification of other gene
signatures associated with the presence of ER (Paik et al.,
22 genes and Symmans et al., 200 genes) [25, 26]. How-
ever, the validity of the above findings was challenged by
the low number of tumor samples, the high number of
assessed genes and technical variances implicating statis-
tical problems (i.e. over-fitting).
In our study of patients with high-risk, predominantly
node-positive early breast carcinomas, ER mRNA positive
status had a favourable prognostic significance for
decreased risk of relapse and death, which was of border-
line statistical significance. Ninety two percent of the
patients received some form of adjuvant hormonal therapy,
including as many as 69% of ER mRNA negative patients.
Moreover, all did receive chemotherapy. Accordingly, it is
unclear whether the transcriptional activity of ER carries
prognostic significance for outcome irrespectively of the
therapy administered, in which case it would reflect
inherent biologic characteristics of the tumor, or whether it
possesses predictive utility for benefit from antineoplastic
or hormonal therapy.
We sought to exploit the availability of a dynamic range
of continuous mRNA data to look for an association of
increasing hormone receptor gene transcription with
enhanced benefit from anti-estrogen therapy but found
none. It is unlikely that sample size precluded identification
of a modest ‘‘dose-response’’ effect, in view of the absence
of any trend for association. Presence of hormone receptor
mRNA levels above a certain biological threshold may
well be irrelevant, if low levels ensure presence of an
amount of protein adequate for biologic effects [27].
Alternatively, increasing concentrations of mRNA may not
linearly translate to protein synthesis, since it is modulated
by several other regulatory mechanisms [28]. Ultimately,
ER and PgR may not be the sole or determining genes that
define response or resistance to hormonal therapy.
We found that ER, PgR and MAP-Tau mRNA positive
status correlated to low histological grade. In view of recent
evidence suggesting that high-grade tumors are driven by a
distinct genetic programme, rather than simply lacking
some markers of differentiation, tumors with active ER,
PgR or MAP-Tau gene function may represent a group with
indolent characteristics in terms of proliferation, anaplasia,
invasion, and cell cycle control [29]. The association of
tumoral MAP-Tau gene transcription with premenopausal
patient status, a marker of more aggressive tumor biology,
may be a random one in a relatively small sample or may
suggest that a subgroup of premenopausal women harbour
breast carcinomas of more benign biology. Alternatively,
this could be a result of high plasma estrogen levels in
premenopausal patients, with increased expression of MAP-
Tau in ER positive tumors of comparatively benign nature,
compared to the more aggressive ER negative, potentially
BRCA1 or BRCA2-positive, premenopausal breast cancer
Table 4 Estimated hazard ratios (HRs) and 95% confidence intervals
(CIs) for OS and DFS—multivariate analysis
HR 95% CIWald P-value
Overall survival
Treatment group
E-T-CMF1
E-CMF1.09 0.65–1.820.74
Positive nodes
0–31
C4 2.33 1.05–5.16 0.04
MAP-Tau mRNA levels
Negative1
Positive 0.460.25–0.85 0.01
ER mRNA levels
Negative1
Positive0.67 0.33–1.360.26
PgR mRNA levels
Negative1
Positive1.780.86–3.670.12
Disease-free survival
Treatment group
E-T-CMF1
E-CMF1.150.74–1.780.53
Positive nodes
0-31
C42.301.21–4.380.01
Tumor grade
I–II1
III-Undifferentiated1.56 0.98–2.500.06
MAP-Tau mRNA levels
Negative 1
Positive0.530.32–0.890.02
ER mRNA levels
Negative1
Positive0.870.46–1.630.66
PgR mRNA levels
Negative1
Positive1.330.72–2.460.36
Breast Cancer Res Treat (2009) 116:131–143 141
123

Page 12

tumors. In line with this, MAP-Tau gene transcription was
strongly associated with ER and PgR gene transcription, an
observation that confirms prior evidence that MAP-Tau is
an ER-inducible gene. Tau contains an imperfect estrogen
response element upstream of its promoter and is an
estrogen-inducible protein in cultured neurons and in the
MCF7, ZR75 and T47D breast cancer cell lines [30, 31].
Interestingly, MAP-Tau is the target of multiple signalling
pathways and thereby may represent an important check-
point for microtubule and cellular functions beyond its
interplay with hormonal activities.
Microtubules are intrinsically dynamic polymers com-
posed of ab-tubulin heterodimers and constitute essential
components of the mitotic spindle and cytoskeleton.
Microtubule dynamics control mitosis and cellular motility.
Moreover, microtubules serve as scaffolds for signalling
molecules,genetranscriptionandcellcycleactivity[14,32].
Several proteins bind to microtubules and either stabilize or
destabilize their polymerisation. Microtubule-associated
protein Tau, one of the most extensively studied, is found in
neurons, breast epithelium and breast cancer. Preclinical
experiments have shown that the phosphorylated form of
MAP-Tau binds to tubulin, resulting in promotion and sta-
bilisation of microtubules. MAP-Tau can compete with
paclitaxel for binding to the inner surface pocket of micro-
tubules, a finding that suggests that microtubules assembled
in the presence of Tau may be less susceptible to taxane
binding and pharmacological hyperpolymerisation [33].
Indeed, Rouzier et al. reported that MAP-Tau gene expres-
sion was inversely associated with pathologic complete
response (pCR) of breast carcinomas to neoadjuvant paclit-
axel-based chemotherapy. Among 82 patients treated with
T-FAC induction regimen, pCR occurred in 44% of MAP-
Tau mRNA-negative tumors, but only in 17% of MAP-Tau
mRNA-positive ones (P = 0.04) [34].
We report here for the first time in the literature that
positive MAP-Tau mRNA expression strongly predicts
decreased risk of relapse and death in patients with high-
risk early breast cancer, suggesting that it carries either
prognostic information for patient outcome or predictive
information for benefit from antineoplastic and/or hor-
monal therapy. In fact, the observed weak predictive utility
of ER mRNA expression was lost in multivariate analysis,
suggesting that it may be exerted via induction of MAP-
Tau gene transcription. The critical regulatory properties of
MAP-Tau for microtubule dynamics make it a potential
key protein in controlling mitosis, migration and cell cycle
activity, possibly resulting in its strong predictive signifi-
cance for benefit from antineoplastic and/or hormonal
therapy.
In contrast to the findings reported by Rouzier et al.,
MAP-Tau mRNA status did not predict clinical benefit
from taxanes in our cohort, as it had the same predictive
utility in both E-CMF treated and E-T-CMF treated
patients. Selection of a different cut-off was unlikely to
change this finding and our exploratory analyses using
different threshold values confirmed this. Andre et al.
found high MAP-Tau gene transcriptional activity to be a
bifunctional predictor of endocrine sensitivity and chemo-
therapy resistance [35]. As the majority of the E-T-CMF
patients in our cohort received endocrine therapy after
chemotherapy, it is possible that any predictive significance
of low MAP-Tau mRNA expression for taxane benefit was
cancelled out by resistance to hormonal therapy, resulting
in no detectable differences in outcome. Moreover, in the
entire HE10/97 trial, the addition of taxanes reduced the
hazard of death only in ER-negative patients [15]. Given
the correlation of ER and MAP-Tau gene activity, it is
probable that ER negative patients who benefit from tax-
anes are mostly MAP-Tau negative. However, this effect
may have been lost in our cohort, as the number of patients
evaluated for MAP-Tau mRNA expression in the present
study was reduced to less than half, due to limitations in
tissue availability. Finally, it is always possible that MAP-
Tau has no true predictive value for response to taxanes, or
that other molecular characteristics abrogate it: tubulin
mutations, variable expression
expression of multidrug resistance proteins, or BCL2 may
contribute to taxane resistance in tumors with low MAP-
Tau expression. Indeed, Rody et al. performed gene
expression profiling in breast carcinomas from 50 patients
treated with taxane-based induction chemotherapy and
failed to find any predictive significance of MAP-Tau gene
expression for pCR [36].
To conclude, we showed the feasibility of obtaining
quantitative mRNA expression data of ER, PgR and MAP-
Tau from widely available FFPE breast tumor tissue,
though the choice of biologically relevant threshold values
needs insights from the fields of molecular and systems
biology. The use of a dynamic range of continuous hor-
mone receptor mRNA data has yet no established value in
quantifying the benefit from endocrine manipulations, as
mRNA levels are modulated by several other parameters
before exerting biological effects. MAP-Tau is an ER-
associated gene that carries a strong predictive significance
for patient outcome, though its effects on sensitivity to
chemotherapy and hormonal therapy are still unclear.
Study of larger and more homogeneous populations of
breast cancer patients is warranted in order to unravel these
issues and exploit their potential predictive utility for
devising tailored therapeutic strategies.
oftubulinisoforms,
Acknowledgements
coordinators for data collection, Ms. E. Fragou and Ms. D. Katsala for
study monitoring, Ms. Th. Spinari for tissue sample collection and
Ms. M. Moschoni for data management. Supported by HeCOG
research grant HE R 10/97.
The authors wish to thank all HeCOG study
142 Breast Cancer Res Treat (2009) 116:131–143
123

Page 13

References
1. EBCTC Group (2005) Effects of chemotherapy and hormonal
therapy for early breast cancer on recurrence and 15-year survival:
an overview of the randomised trials. Lancet 365:1687–1717
2. Bonneterre J, Thurlimann B, Robertson JFR et al (2000) Anas-
trozole versus tamoxifen as first-line therapy for advanced breast
cancer in 668 post-menopausal women: results of the Tamoxifen
or Arimidex Randomised Group Efficacy and Tolerability study.
J Clin Oncol 18:3748–3757
3. Mouridsen H, Gershanovich M, Sun Y et al (2001) Superior
efficacy of letrozole versus tamoxifen as first-line therapy for
postmenopausal women with advanced breast cancer: results of a
phase III study of the International Letrozole Breast Cancer
Group. J Clin Oncol 19:2596–2606
4. Henderson IC, Patek AJ (1998) The relationship between prog-
nostic and predictive factors in the management of breast cancer.
Breast Cancer Res Treat 52:261–288
5. Trudeau M, Charbonneau F, Gelmon K et al (2005) Selection of
adjuvant chemotherapy for treatment of node-positive breast
cancer. Lancet Oncol 6:886–898
6. Pusztai L, Mazouni C, Anderson K, Wu Y, Symmans F (2006)
Molecular classification of breast cancer: limitations and poten-
tial. Oncologist 11:868–877
7. Harvey JM, Clark GM, Osborne CK et al (1999) Estrogen
receptor status by immunohistochemistry is superior to the
ligand-binding assay for predicting response to adjuvant endo-
crine therapy in breast cancer. J Clin Oncol 17:1474–1481
8. Layfield LJ, Goldstein N, Perkinson KR et al (2003) Interlabo-
ratory variation in results from immunohistochemical assessment
of estrogen receptor status. Breast J 9:257–259
9. Rhodes A, Jasani B, Barnes DM et al (2000) Reliability of immu-
nohistochemical demonstration of estrogen receptors in routine
practice:interlaboratoryvarianceinthesensitivityofdetectionand
evaluation of scoring systems. J Clin Pathol 53:125–130
10. Rudiger T, Hofler H, Kreipe HH et al (2002) Quality assurance in
immunohistochemistry:resultsofaninterlaboratorytrialinvolving
172 pathologists. Am J Surg Pathol 26:873–882
11. Bianchini G, Zambetti M, Pusztai L et al (2007) Use of estrogen
receptor expression by quantitative RT-PCR to identify an ER-
negative subgroup by IHC who might benefit from hormonal
therapy. 2007 Breast Cancer Symposium; Abstract 106
12. Badve SS, Baehner FL, Gray R et al (2007) ER and PR assess-
ment in ECOG 2197: comparison of locally determined IHC with
centrally determined IHC and quantitative RT-PCR. 2007 Breast
Cancer Symposium; Abstract 87
13. Gong Y, Yan K, Anderson K, Sotiriou C, Andre F et al (2007)
Determination of estrogen-receptor status and ERBB2 status of
breast carcinoma: a gene-expression profiling study. Lancet
Oncol 8:203–211
14. Bhat KMR, Setaluri V (2007) Microtubule-associated proteins as
targets in cancer chemotherapy. Clin Cancer Res 13:2849–2854
15. FountzilasG, SkarlosD,DafniU,GogasH,BriasoulisE et al(2005)
Postoperative dose-dense sequential chemotherapy with epirubicin,
followedbyCMFwithorwithoutpaclitaxel,inpatientswithhigh-risk
operablebreastcancer:arandomisedphaseIIIstudyconductedbythe
Hellenic Cooperative Oncology Group. Ann Oncol 16:1762–1771
16. Landis JR, Koch GG (1977) The measurement of observer
agreement for categorical data. Biometrics 33:159–174
17. Hudis TL et al (2007) Proposal for standardized definitions for
efficacy end points in adjuvant breast cancer trials: the STEEP
system. J Clin Oncol 25:2127–2132
18. McShane LM, Altman DG, Sauerbrei W et al (2005) Reporting
recommendations for tumor marker prognostic studies. J Clin
Oncol 23:9067–9072
19. Goldstein NS, Ferkowicz M, Odish E et al (2003) Minimum
formalin fixation time for consistent estrogen receptor immuno-
cytochemical staining of invasive breast carcinoma. Am J Clin
Pathol 120:86–92
20. Rhodes A (2003) Quality assurance in immunohistochemistry.
Am J Surg Pathol 27:1284–1285
21. Morris KV (2008) RNA-mediated transcriptional gene silencing
in human cells. Curr Top Microbiol Immunol 320:211–224
22. Ravo M, Mutarelli M, Ferraro L, Grober OM, Paris O et al (2008)
Quantitative expression profiling of highly degraded RNA from
formalin-fixed paraffin-embedded breast tumour biopsies by oli-
gonucleotide microarrays. Lab Invest 88:430–440
23. Van Maldegem F, De Wit M, Morsink F, Musler A, Weegenaar J
et al (2008) Effects of processing delay, formalin fixation and
immunohistochemistry on RNA recovery from formalin-fixed
paraffin-embedded tissue sections. Diagn Mol Pathol Jan 28
(Epub ahead of print)
24. Pusztai L, Ayers M, Stec J, Clark E, Hess K et al (2003) Gene
expression profiles obtained from fine-needle aspirations of breast
cancer reliably identify routine prognostic markers and reveal
large-scale molecular differences between estrogen-negative and
estrogen-positive tumours. Clin Cancer Res 9:2406–2415
25. Paik S, Shak S, Tang G et al (2004) A multigene assay to predict
recurrence of tamoxifen-treated, node-negative breast cancer. N
Engl J Med 351:2817–2826
26. SymmansWF,SotiriouC,AndersonSKetal(2005)Measurements
ofestrogenreceptorandreportergenesfrommicroarraysdetermine
receptor status and time to recurrence following adjuvant tamoxi-
fen therapy. Breast Cancer Res Treat 94(suppl 1):308a
27. Webb SE, Roberts SK, Needham SR, Tynan CJ, Rolfe DJ et al
(2008) Single-molecule imaging and fluorescence lifetime
imaging microscopy show different structures for high- and low-
affinity EGFR in A431 cells. Biophys J 94:803–819
28. Yasuda J, Hayashizaki Y (2008) The RNA continent. Adv Cancer
Res 99:77–112
29. Ramaswamy S, Tamayo P, Rifkins R et al (2001) Multiclass
cancer diagnosis using tumour gene expression signatures. Proc
Natl Acad Sci USA 98:15149–15154
30. West M, Blanchette C, Dressman H, Huang E, Ishida S et al
(2001) Predicting the clinical status of human breast cancer by
using gene expression profiles. Proc Natl Acad Sci USA
98:11462–11467
31. Matsuno A, Takekoshi S, Sanno N, Utsunomiya H, Ohsugi Y
et al (1997) Modulation of protein kinases and microtubule-
associated proteins and changes in ultrastructure in female rat
pituitary cells: effects of estrogen and bromocriptine. J Histo-
chem Cytochem 45:805–813
32. Jordan MA, Wilson L (2004) Microtubules as a target for anti-
cancer drugs. Nat Rev Cancer 4:253–265
33. Kar S, Fan J, Smith MJ, Goedert M, Amos LA (2003) Repeat
motifs of Tau bind to the insides of microtubules in the absence
of taxol. EMBO J 22:70–77
34. Rouzier R, Rajan R, Wagner P, Hess KR, Gold DL et al (2005)
Microtubule-associated protein tau: a marker of paclitaxel sen-
sitivity in breast cancer. PNAS 102:8315–8320
35. Andre F,Hatzis C,Anderson K, SotiriouS, Mazouni Cetal (2007)
Microtubule-associated protein tau is a bifunctional predictor of
endocrine sensitivity and chemotherapy resistance in estrogen-
receptor positive breast cancer. Clin Cancer Res 13:2061–2067
36. Rody A, Karn T, Gatje R, Ahr A, Solbach C et al (2007) Gene
expression profiling of breast cancer patients treated with doce-
taxel, doxorubicin, and
GEPARTRIO trial: HER2, but not topoisomerase II alpha and
microtubule-associated protein tau, is highly predictive of tumour
response. Breast 16:86–93
cyclophosphamide within the
Breast Cancer Res Treat (2009) 116:131–143143
123