Annals of Oncology original articles

25. Lipton A, Steger GG, Figueroa J et al. Randomized active-controlled phase II
study of denosumab efcacy and safety in patients with breast cancer-related
bone metastases. J Clin Oncol 2007; 25: 44314437.
26. Stopeck AT, Lipton A, Body JJ et al. Denosumab compared with zoledronic acid
for the treatment of bone metastases in patients with advanced breast cancer: a
randomized, double-blind study. J Clin Oncol 2010; 28: 51325139.
27. Cleeland C. Research in cancer pain. What we know and what we need to know.
Cancer 1991; 67: 823827.
28. Rosen LS, Gordon D, Tchekmedyian S et al. Zoledronic acid versus placebo in
the treatment of skeletal metastases in patients with lung cancer and other solid
tumors: a phase III, double-blind, randomized trialthe Zoledronic Acid Lung
Cancer and Other Solid Tumors Study Group. J Clin Oncol 2003;
21: 31503157.
29. Mathias SD, Crosby RD, Qian Y et al. Estimating minimally important differences
for the worst pain rating of the Brief Pain Inventory-Short Form. J Support Oncol
2011; 9: 7278.
30. Weinfurt KP, Anstrom KJ, Castel LD et al. Effect of zoledronic acid on pain
associated with bone metastasis in patients with prostate cancer. Ann Oncol
2006; 17: 986989.
31. Roque IFM, Martinez-Zapata MJ, Scott-Brown M et al. Radioisotopes
for metastatic bone pain. Cochrane Database Syst Rev 2011;
CD003347.

Annals of Oncology 23: 30513057, 2012

doi:10.1093/annonc/mds120
Published online 5 July 2012

Clinicopathological analysis of GATA3-positive breast

cancers with special reference to response to
neoadjuvant chemotherapy
N. Tominaga, Y. Naoi, K. Shimazu, T. Nakayama, N. Maruyama, A. Shimomura, S. J. Kim,
Y. Tamaki & S. Noguchi*
Department of Breast and Endocrine Surgery, Osaka University Graduate School of Medicine, Osaka, Japan

Received 16 December 2011; revised 14 March 2012; accepted 14 March 2012

Background: The aim of this study was to investigate the clinicopathological characteristics of GATA binding protein 3
(GATA3)-positive breast cancers as well as the association of GATA3 expression with response to chemotherapy.
Patients and methods: Tumor specimens obtained before neoadjuvant chemotherapy [ paclitaxel followed by
5-uorouracil/epirubicin/cyclophosphamide)] from breast cancer patients (n = 130) were subjected to
immunohistochemical and mutational analysis of GATA3 and DNA microarray gene expression analysis for intrinsic
subtyping.
Results: Seventy-four tumors (57%) were immunohistochemically positive for GATA3. GATA3-positive tumors were
signicantly more likely to be lobular cancer, estrogen receptor (ER)-positive, progesterone receptor (PgR)-positive, Ki67-
negative, and luminal A tumors. Somatic mutations were found in only three tumors. Pathological complete response (pCR)
was observed in 8 (11%) GATA3-positive tumors and in 22 (39%) GATA3-negative tumors. Multivariate analysis showed that
tumor size, human epidermal growth factor receptor 2 (HER2), and GATA3 were independent predictors of pCR.
Conclusions: GATA3-positive breast cancers showed luminal differentiation characterized by high ER expression and
were mostly classied as luminal-type tumors following intrinsic subtyping. Interestingly, GATA3 was an independent
predictor of response to chemotherapy, suggesting that GATA3 might be clinically useful as a predictor of a poor
response to chemotherapy.
Key words: breast cancer, GATA3, intrinsic subtype, microarray, neoadjuvant chemotherapy, pathological complete
response

introduction motif [1]. GATA3 is a 444-amino-acid transcription factor

GATA binding protein 3 (GATA3) belongs to the GATA
family, which consists of six GATA transcription factors
(GATA16) that bind to the consensus (A/T)GATA(A/G)
*Correspondence to: Prof. S. Noguchi, Department of Breast and Endocrine Surgery,
Osaka University Graduate School of Medicine, 2-2-E10 Yamadaoka, Suita-shi, Osaka
565-0871, Japan. Tel: +81-6-6879-3772; Fax: +81-6-6879-3779;
E-mail: noguchi@onsurg.med.osaka-u.ac.jp
located in 10p15 and possesses an activation domain (exons 2
and 3) and a DNA binding domain (exons 4 and 5) [2].
GATA3 was initially identied as a DNA-binding protein
involved in the activation of transcription at the T-cell receptor
alpha locus and has been shown to be implicated in the
development of T-cells. Later studies have revealed that
GATA3 is also involved in the development of various organs
including the mammary gland [1].

The Author 2012. Published by Oxford University Press on behalf of the European Society for Medical Oncology.
All rights reserved. For permissions, please email: journals.permissions@oup.com.
original articles Annals of Oncology

GATA3 is critical for the normal development of the Each patient underwent a vacuum-assisted core tumor biopsy
mammary gland. Introduction of GATA3 into a puried (Mammotome 8G; Ethicon Endosurgery; Johnson & Johnson, Cincinnati,
mammary progenitor-enriched population has been shown to OH) under ultrasonographic guidance before neoadjuvant chemotherapy.
induce luminal cell differentiation [3]. In human breast tissue, Tumor biopsy samples were xed in 10% buffered formaldehyde for
GATA3 is expressed in the luminal epithelial cells, and its histological examination and also snap-frozen in liquid nitrogen and kept
expression is also seen in breast cancer [4, 5]. The expression at 80C until use for DNA and RNA extraction. Every patient underwent
breast-conserving surgery or mastectomy after neoadjuvant chemotherapy.
of GATA3 is closely associated with estrogen receptor (ER) and
Patient characteristics entered in this study are shown in Table 1. Informed
forkhead box A1 (FOXA1) expression in breast cancer [4, 6].
consent for the study was obtained from each patient before tumor biopsy.
FOXA1 plays an important role in the transcriptional function
of ER [7]. It has also been suggested that GATA3 may be a
mediator for the transcriptional up-regulation of mucin 1 Table 1. Associations between GATA3 expression and clinicopathological
(MUC1) in breast cancer [8]. These results clearly indicate that parameters
GATA3 plays a pivotal role in the development and
differentiation of luminal epithelial cells, which is consistent Characteristic Number GATA3 P value
with the preferential expression of GATA3 in luminal-type of patients, Negative, n Positive, n
breast cancer [9]. Somatic mutations of GATA3 are reported to n (%)
occur infrequently (4%5%) in breast tumors. Interestingly, Total, n (%) 130 56 (43) 74 (57)
these mutations are mostly seen in ER-positive tumors, Median age, year (range) 52 (2773)
suggesting involvement of GATA3 mutation in the Histological type 0.002a
pathogenesis of luminal-type breast cancer [10, 11]. IDC 119 (92) 56 63
Because of this crucial function of GATA3, GATA3-positive ILC 11 (9) 0 11
tumors are thought to be derived from luminal progenitors and Stage 0.894a
to maintain the luminal epithelial cell phenotype. In fact, IIA 43 (33) 18 25
GATA3-positive tumors are often ER-positive tumors, and a IIB 69 (53) 30 39
large body of studies has demonstrated that GATA3 expression is IIIA 13 (10) 5 8
signicantly associated with a favorable prognosis, suggesting that IIIB 5 (4) 3 2
GATA3 may serve as a new prognostic factor for breast cancer Tumor size 0.624a
T1 7 (5) 2 5
[12]. On the other hand, the relationship of GATA3 expression
T2 98 (75) 44 54
with response to chemotherapy has rarely been studied. Since
T3 20 (15) 7 13
GATA3 expression is associated with ER expression, which has
T4 5 (4) 3 2
been shown to be associated with a poor response to
Lymph node status 0.844b
chemotherapy [13], it is speculated that GATA3 expression is
N0 43 (33) 18 25
also associated with a poor response. Moreover, since GATA3 N1 87 (67) 38 49
induces a variety of luminal epithelial cell-related genes including Histological grade <0.001a
ER, GATA3 may be even more strongly associated with resistance 1 18 (14) 1 17
to chemotherapy. In the study presented here, we therefore tried 2 86 (66) 38 48
rst to elucidate the clinicopathological characteristics of 3 26 (20) 17 9
GATA3-positive tumors and, next, to investigate the association ER status (10%) <0.001b
of GATA3 expression with response to neoadjuvant Positive 77 (59) 13 64
chemotherapy for breast cancer. In addition, we studied whether Negative 53 (41) 43 10
or not GATA3 and FOXA1 could cooperate to induce ER in PgR status (10%) <0.001b
human breast cancer tissues since FOXA1, like GATA-3, has Positive 50 (39) 9 41
been reported to induce ER [6, 7], and also studied the Negative 80 (62) 47 33
correlation of MUC1 with response to neoadjuvant Her2 status (>2.0/FISH) 0.306b
chemotherapy since MUC1 is induced by GATA3 [8] and is Positive 38 (29) 19 19
shown to be implicated in resistance to chemotherapy [14]. Negative 92 (71) 37 55
Ki67 status (20%) 0.017b
Positive 77 (59) 40 37
patients and methods Negative 52 (40) 16 36
Unknown 1 (1) 0 1
patients MUC1 status (10%) 0.024a
Primary breast cancer patients (n = 130) who had been treated with Positive 97 (75) 36 61
neoadjuvant chemotherapy [ paclitaxel (80 mg/m2) weekly for 12 cycles Negative 24 (19) 15 9
followed by combined 5-uorouracil (500 mg/m2), epirubicin (75 mg/m2), Unknown 9 (7) 5 4
and cyclophosphamide (500 mg/m2) every 3 weeks for four cycles (P-
a
FEC)] between 2004 and 2009 at Osaka University Hospital were Fishers exact test.
b
retrospectively recruited for this study. During this period, neoadjuvant Chi-square test.
chemotherapy was indicated for the breast cancer patients with T34 and ER, estrogen receptor; GATA3, GATA binding protein 3; Her2, human
any N or any T and N13, and for those who had large tumors (3 cm) epidermal growth factor receptor 2; IDC, invasive ductal carcinoma; ILC,
relative to the breast and wished to undergo breast-conserving surgery. invasive lobular carcinoma; MUC1, mucin 1; PgR, progesterone receptor.

| Tominaga et al. Volume 23 | No. 12 | December 2012

Annals of Oncology
immunohistological examination
Parafn sections (2 m) were subjected to immunohistochemistry for
GATA3, FOXA1, and MUC1. The sections for GATA3 were deparafnized
and incubated with HistoVT one ( pH7.0, citrate buffer; Nacalai Tesque
Inc., Kyoto, Japan) for 40 min at 98C, those for FOXA1 were incubated
with Target Retrieval Solution ( pH6.0, citrate buffer; DAKO Japan Inc.,
Tokyo, Japan) for 20 min at 98C, and those for MUC1 were incubated
with Target Retrieval Solution ( pH6.0, citrate buffer; DAKO Japan Inc.)
for 40 min at 98C. These sections were then incubated with anti-GATA3
antibodies (HG3-31 mouse monoclonal antibody, 1/100; Santa Cruz
Biotechnology Inc., Santa Cruz, CA), anti-FOXA1 antibodies (Q6 mouse
monoclonal antibody, 1/1000; Santa Cruz Biotechnology), or anti-MUC1
antibodies (DF3 mouse monoclonal antibody, 1/100; Abcam Inc., Tokyo,
Japan) for 1 h at room temperature, followed by incubation with the
secondary antibody (Biotin Anti Mouse IgG, 0.33 mg/ml; Jackson
Immunoresearch, West Grove, PA) for 1 h at room temperature, and
subsequent treatment with the avidin-biotinylated peroxidase complex
method for visualization. Positive and negative controls were used for each
run [1517].
Percentages of immunohistochemically positive tumor cells were
determined by counting 1000 tumor cells with WinROOF (Mitani Corp.,
Fukui, Japan). Immunohistochemical ndings for GATA3 and FOXA1
were considered positive when the percentage of positive tumor cells
(nuclear staining) was >10% and >40%, respectively. MUC1 was considered
positive when the percentage of positive tumor cells (membranous
staining) was >10%. These cut-off values for GATA3, FOXA1, and MUC1
were decided so that P value for their correlation with pathological
complete response ( pCR) became minimal. Immunohistochemical
assessments of ER, progesterone receptor (PgR), and Ki67 and FISH of
human epidermal growth factor receptor 2 (HER2) in tumor biopsy
samples were carried out as previously described [18].
evaluation of response to chemotherapy
Pathological response to P-FEC was evaluated by examining the surgical
specimens obtained at surgery. The specimens were cut into 5-mm sections,
which were hematoxylin and eosin stained for the presence or absence of
tumor cells. A complete loss of invasive tumor cells in the primary tumor
site without any lymph node metastasis was classied as pCR.
sequence analysis of GATA3
DNA was extracted with the DNeasy Blood & Tissue Kit (Qiagen, Tokyo,
Japan) from tumor biopsy samples and subjected to mutational analysis for
detection of exons 4, 5, and 6 of GATA3 by direct sequencing [19]. The
primer sequences and PCR conditions are shown in Supplemental Table S1
(available at Annals of Oncology online).
intrinsic subtyping
RNA was extracted with Trizol Reagent (Life Technologies Inc.,
Gaithersburg, MD) from the tumor biopsy samples and subjected to gene
expression analysis using U133 Plus 2.0 array (Affymetrix, Santa Clara,
CA) and intrinsic subtyping (luminal A, luminal B, HER2 enriched, basal
like, normal breast like) according to a previously described method [20].
statistics
Dr. SPSS II (11.0.1J; SPSS Japan Inc., Tokyo, Japan) was used for all
statistical analyses. Associations between the various clinicopathological
parameters and GATA3 expression were assessed by means of the chi-
squared test or Fishers exact test. Associations of the various
clinicopathological parameters with pCR were evaluated by using uni- and
Volume 23 | No. 12 | December 2012
original articles
multivariate analysis (logistic regression analysis). Hierarchal cluster
analysis was carried out according to the method specied for Genomics
Suite 6.5 (version 6.11.0321; Partek Inc., St Louis, MI). Statistical analysis
was two-sided and P < 0.05 was considered to be signicant except for
selection of the genes differentially expressed in GATA3-positive and
GATA3-negative tumors by means of the Welch t-test, where
P < 9.15 107 was considered to be signicant after Bonferroni correction.
results
associations of GATA3 expression with various
clinicopathological parameters
Of the 130 breast tumors, 74 (57%) were
immunohistochemically positive for GATA3. Representative
results of immunohistochemical examinations for GATA3 and
FOXA1 are shown in Figure 1. GATA3-positive tumors were
signicantly more likely to be lobular cancers (P = 0.002), to
show a low histological grade (P < 0.001), and to be ER-positive
(P < 0.001), PgR-positive (P < 0.001), Ki67-negative (P = 0.017),
and MUC1-positive (P = 0.024) tumors. Intrinsic subtyping was
done by means of DNA microarray for the 123 fresh frozen
tumors, which showed 59 luminal A tumors, 28 luminal B
tumors, 14 HER2-enriched tumors, 19 basal-like tumors, and 3
normal breast-like tumors (Figure 2). Since the number of
normal breast-like tumors was so small and since it is suggested
that normal breast-like might result from the contamination by
the normal breast tissue [21], normal breast-like tumors were
not considered in the following analysis. Frequency of GATA3-
positive tumors was signicantly higher for luminal A than
luminal B tumors (P = 0.001), HER2-enriched tumors
(P < 0.001), and basal-like tumors (P < 0.001) (Table 2).
associations of GATA3 and FOXA1 with ER
expression
Tumor tissues of 123 patients could be used for simultaneous
analysis of GATA3, FOXA1, and ER. Representative results of
FOXA1 are shown in Figure 1. Of the 71 GATA3-positive
tumors, 52 were positive for FOXA1, and of the 52 GATA3-
negative tumors, 36 were negative for FOXA1. There was a
signicant (P < 0.001) association between GATA3 and FOXA1
expression. Figure 3 shows the relationship between GATA3/
FOXA1 expression and ER expression. Frequency of ER-
positive tumors showed a gradual increase from 14% in
GATA3()/FOXA1() tumors, to 44% in GATA3()/FOXA1
(+) tumors, 79% in GATA3(+)/FOXA1() tumors, and 89% in
GATA3(+)/FOXA1(+) tumors.
mutation analysis of GATA3
DNA from 99 tumors could be used for mutational analysis.
Three somatic mutations, including one missense mutation
(1754 G>A), one insert mutation (1873 CCC insert) and one
deletion mutation (1888-1902 deletion), were identied
(mutation rate, 3%). All three tumors which have somatic
mutation were GATA3-positive, ER-positive tumors, and were
classied as luminal A (n = 2) or luminal B (n = 1) tumors
following intrinsic subtyping. Besides, pCR was not observed
in any of them.
doi:10.1093/annonc/mds120 |
original articles
Figure 1 Immunohistological examination of GATA3 and FOXA1. Representative results of immunohistochemical staining of GATA3 and FOXA1 in
breast tumors and normal breast tissues (200, bar = 20 m) are shown. Nuclear staining of GATA3 and FOXA1 is seen in breast cancer cells ( panels A and
C, respectively) and normal luminal epithelial cells ( panels B and D, respectively). FOXA1, forkhead box A1; GATA3, GATA binding protein 3.
GATA3 for prediction of pathological response
pCR was observed in 8 (11%) of 74 GATA3-positive tumors
and in 22 (39%) of 56 GATA3-negative tumors (P < 0.001)
(Supplemental Figure S1, available at Annals of Oncology
online). Univariate analysis of clinicopathological parameters
showed that tumor size (P = 0.01), ER (P < 0.001), PgR
(P = 0.004), HER2 (P = 0.004), Ki67 (P = 0.027), and GATA3
(P < 0.001), but not FOXA1 and MUC1, were signicantly
associated with pCR rate (Table 3). Multivariate analysis
identied tumor size (P = 0.037), HER2 (P = 0.027), and
GATA3 (P = 0.036) as independent predictors of pCR rate
(Table 3).
differentially expressed genes in GATA3-positive
and GATA3-negative tumors
The 37 genes, corresponding to the 57 probes and differentially
expressed in GATA3-positive and GATA-negative tumors, were
selected by using the P value (9.15 107) determined after
Bonferroni correction. The list of these differentially expressed
genes with annotations is shown in Supplemental Table S3
(available at Annals of Oncology online). Of the 57 probes, 21
(37%) were related to metabolism, 9 (16%) to gene
transcription, 6 (11%) to signal transduction, 5 (9%) to mucin
secretion, 2 (4%) to proliferation, 2 (4%) to DNA repair, and
12 (21%) were unknown.
discussion
For current study, we rst studied the associations between
GATA3 expression and the various conventional
| Tominaga et al.
Annals of Oncology
clinicopathological parameters in order to elucidate the
characteristics of GATA3-positive tumors.
Immunohistochemical examination showed that such tumors
were more likely to be ER-positive and PgR-positive tumors
and were more likely to be luminal A tumors as determined by
gene expression proling. These ndings are consistent with
the function of GATA3 in which it plays a pivotal role in
differentiation into luminal epithelial cells. Similar results for
the preferential expression of GATA3 in luminal A tumors
have also been reported by Chou et al. [22]. Interestingly, we
found that all inltrating lobular cancers were GATA3-positive
as was also found by Albergaria et al. [23]. It has been reported
that GATA3 expression is relatively high in the terminal end
buds from which lobular carcinomas are thought to stem, so
that inltrating lobular cancers are considered to be breast
tumors, which have highly differentiated into luminal cells. In
fact, all inltrating lobular cancers in our study were classied
as luminal A tumors.
Since GATA3 and FOXA1 have been reported to induce ER
synergistically in in vitro studies [22, 24, 25], we investigated
whether such a cooperation can also be observed in human
breast cancers. We found that both GATA3- and FOXA1-
positive tumors showed the highest ER positivity and that both
GATA3- and FOXA1-negative tumors showed the lowest
positivity, while either GATA3- or FOXA1-positive tumors
showed intermediate positivity. These results seem to be
consistent with the in vitro ndings that GATA3 and FOXA1
cooperate to induce ER in breast cancer.
It has been reported that 4%5% of breast cancers harbor
somatic mutations of GATA3, which are predominantly
located in exons 46 [10, 11]. Consistent with this nding, we
Volume 23 | No. 12 | December 2012
Annals of Oncology original articles

Figure 2 Intrinsic subtyping of breast tumors. Classication of 123 breast tumors into intrinsic subtypes (luminal A, luminal B, HER2 enriched, basal-like,
and normal breast like) was done by DNA microarray analysis of gene expression with a previously described method [20]. Hierarchal cluster analysis was
done in each subtype and the results are shown as a heatmap. The genes used for classication is also shown in the Supplementary Table 2 (available at
Annals of Oncology online). Bars at the bottom show the immunohistochemical results for ER, PR, HER2, Ki67, GATA3, and FOXA1 as well as the
pathological response. ER, estrogen receptor; FOXA1, forkhead box A1; GATA3, GATA binding protein 3; HER2, human epidermal growth factor receptor
2; OR, odds ratio; pCR, pathological complete response; PR, progesterone receptor.

Table 2. Associations between GATA3 expression and intrinsic subtype

Subtype GATA3 P valuea

Negative, n Positive, n
Luminal A 9 50
Luminal B 14 14 0.001
Her2 enriched 10 4 <0.001
Basal like 18 1 <0.001
Normal like 1 2 0.416
a
Fishers exact test.
GATA3, GATA binding protein 3; Her2, human epidermal growth factor
receptor 2.

Figure 3 Association of GATA3/FOXA1 with ER expression. Percentages

identied three breast tumors with somatic mutations. All of ER-positive and ER-negative patients in GATA3()/FOXA1(+), GATA3
three tumors were ER positive and were classied as luminal A ()/FOXA1(+), GATA3(+)/FOXA1(), and GATA3(+)/FOXA1(+) breast
tumors (n = 2) and luminal B tumors (n = 1) following tumors are shown. ER, estrogen receptor; FOXA1, forkhead box A1;
intrinsic subtyping. Perou et al. reported that 5 of 111 breast GATA3, GATA binding protein 3.

Volume 23 | No. 12 | December 2012 doi:10.1093/annonc/mds120 |

original articles Annals of Oncology

Table 3. Univariate and multivariate analysis of various predictive factors for response to chemotherapy

Factor Univariate Multivariate

OR 95% CI P valuea OR 95% CI P valueb
Lower Upper Lower Upper
Tumor size (T1,2 versus T3,4) 9.6 1.2 76.9 0.010 9.5 1.1 76.9 0.037
Lymph node metastasis (N1 vs. N0) 1.6 0.6 4.2 0.316
Histological grade (3 versus 1,2) 1.9 0.7 4.9 0.209
ER (negative versus positive) 5.9 2.3 14.9 <0.001 1.8 0.4 9.5 0.450
PgR (negative versus positive) 4.2 1.5 12.0 0.004 1.1 0.2 6.0 0.922
Her2 (positive versus negative) 3.5 1.5 8.4 0.004 3.1 1.1 8.7 0.027
Ki67 (positive versus negative) 2.8 1.1 7.2 0.027 1.4 0.5 4.6 0.535
GATA3 (negative versus positive) 5.3 2.1 13.3 <0.001 3.6 1.1 11.6 0.036
FOXA1 (negative versus positive) 2.1 0.9 4.9 0.085
MUC1 (negative versus positive) 1.9 0.7 5.3 0.217
a
Chi-square test.
b
Logistic regression analysis.
CI, condence interval; ER, estrogen receptor; FOXA1, forkhead box A1; GATA3, GATA binding protein 3; Her2, human epidermal growth factor receptor
2; MUC1, mucin 1; OR, odds ratio; PgR, progesterone receptor.

tumors (5%) were found to harbor somatic mutations and all
these tumors were ER positive [10]. Since GATA3 causes
invasive breast cancer cells to undergo reversal of epithelial
mesenchymal transition, leading to the suppression of cancer
metastasis [26], and it regulates tumor differentiation and
suppresses tumor dissemination in breast cancer [5], GATA3 is
thought to have a tumor suppressive effect. Putting these
results together leads to the speculation that loss of GATA3
function is implicated in the pathogenesis of luminal-type
breast cancer even though its incidence is very low (< 5%).
It is well known that ER positivity is associated with a poor
response to chemotherapy, and, indeed, we found in this study
that the pCR rate of ER-positive tumors (12%) was
signicantly lower than that of ER-negative tumors (40%). The
precise reason for this association still remains to be
investigated but one hypothesis is that ER serves as a marker of
luminal epithelial differentiation and that well-differentiated
tumors are generally thought to be less likely to respond to
chemotherapy than less differentiated tumors. We next
examined the relationship between GATA3 expression and
pCR rate because GATA3 is a master gene for luminal
differentiation and can thus be expected to reect the luminal
differentiation status more accurately than ER expression and
may serve as a better marker of resistance to chemotherapy
than ER. We were able to demonstrate that GATA3-positive
tumors were signicantly less likely to achieve pCR, and
multivariate analysis results showed that GATA3 positivity was
signicantly associated with a lower pCR rate as well as that
this association was independent of other clinicopathological
parameters including ER. To the best of our knowledge, ours is
the rst report to demonstrate that GATA3 is a better predictor
than ER of response to chemotherapy.
Since GATA3 induces a variety of genes during luminal
differentiation, we speculated that some of these might play a
signicant role in the resistance to chemotherapy. We therefore
investigated the genes differentially expressed in GATA3-
positive and GATA3-negative tumors and were able to identify
37 genes implicated in apoptosis, DNA repair, cell cycle
regulation, and metastasis (Supplemental Table S3, available at
Annals of Oncology online). The inclusion of trefoil factor 1
(TFF1) in these genes was noteworthy since it was signicantly
highly expressed in GATA3-positive tumors and is known to
inhibit the cytochrome C- or Fas-induced activation of caspase
8 [27]. This suggests that TFF1 may play a signicant role in
the resistance to chemotherapy through the inhibition of
apoptosis. It has also been suggested that phosphoserine
aminotransferase 1 (PSAT1) is implicated in chemoresistance
by decreasing the apoptotic response in colon cancer cell [28].
Moreover, anterior gradient 2 homolog (AGR2) is reportedly
implicated in resistance to gemcitabine in pancreatic cancer
cell lines [29] and cyclin D1 (CCND1) is also reported to play
a signicant role in resistance to docetaxel in in vitro [30]. In
addition, high expression of CCND1 in GATA3-positive
tumors is compatible with the recent nding that CCND1 is a
direct transcriptional target of GATA3 in neuroblastoma tumor
cells [31]. However, the precise mechanism of the resistance of
GATA3-positive tumors to chemotherapy still needs to be
investigated.
A limitation of this study is that the GATA3 cut-off value
was decided so that the P value for the correlation between
GATA3 and pathological response would become minimal,
implying that this could be a hypothesis-raising study and thus
the predictive value of GATA3 is overrated. Our present
observation needs to be validated in future. However, the fact
that the median value of GATA3 was also 10% seems to
indicate that the association of GATA3 expression with
resistance to chemotherapy may be plausible since
dichotomization using a median value is a less biased method.
In conclusion, our results showed that GATA3-positive
tumors possessed a phenotype of luminal A tumors and
GATA3 mutations were found in only 3% of breast tumors.
Furthermore, GATA3 was found to be an independent
predictor of response to chemotherapy for human breast
cancer, suggesting that GATA3 may be clinically useful as a

| Tominaga et al. Volume 23 | No. 12 | December 2012

Annals of Oncology
predictor of a poor response to chemotherapy. However, our
ndings need to be validated by a future study including a
larger number of patients.
funding
This work was supported, in part, by a grant for The 3rd-Term
Comprehensive 10-Year Strategy for Cancer Control (026)
from the Ministry of Health, Labour and Welfare, Japan.
disclosures
SN received research funding from Bristol-Myers Squibb and
Pzer and honoraria from Pzer. SJK and TN received
honoraria from Bristol-Myers Squibb and Pzer. The
remaining authors have declared no conicts of interest.
references
1. Fang SH, Chen Y, Weigel RJ. GATA-3 as a marker of hormone response in breast
cancer. J Surg Res 2009; 157: 290295.
2. Yang Z, Gu L, Romeo PH et al. Human GATA-3 trans-activation, DNA-binding, and
nuclear localization activities are organized into distinct structural domains. Mol
Cell Biol 1994; 14: 22012212.
3. Asselin-Labat ML, Sutherland KD, Barker H et al. Gata-3 is an essential regulator
of mammary-gland morphogenesis and luminal-cell differentiation. Nat Cell Biol
2007; 9: 201209.
4. Kouros-Mehr H, Slorach EM, Sternlicht MD, Werb Z. GATA-3 maintains the
differentiation of the luminal cell fate in the mammary gland. Cell 2006; 127:
10411055.
5. Kouros-Mehr H, Bechis SK, Slorach EM et al. GATA-3 links tumor differentiation
and dissemination in a luminal breast cancer model. Cancer Cell 2008; 13:
141152.
6. Kouros-Mehr H, Kim JW, Bechis SK, Werb Z. GATA-3 and the regulation of the
mammary luminal cell fate. Curr Opin Cell Biol 2008; 20: 164170.
7. Bernardo GM, Lozada KL, Miedler JD et al. FOXA1 is an essential determinant of
ERalpha expression and mammary ductal morphogenesis. Development 2010;
137: 20452054.
8. Abba MC, Nunez MI, Colussi AG et al. GATA3 protein as a MUC1 transcriptional
regulator in breast cancer cells. Breast Cancer Res 2006; 8: R64.
9. Demir H, Turna H, Can G, Ilvan S. Clinicopathologic and prognostic evaluation of
invasive breast carcinoma molecular subtypes and GATA3 expression. J BUON
2010; 15: 774782.
10. Usary J, Llaca V, Karaca G et al. Mutation of GATA3 in human breast tumors.
Oncogene 2004; 23: 76697678.
11. Arnold JM, Choong DY, Thompson ER et al. Frequent somatic mutations of
GATA3 in non-BRCA1/BRCA2 familial breast tumors, but not in BRCA1-, BRCA2-
or sporadic breast tumors. Breast Cancer Res Treat 2010; 119: 491496.
12. Yoon NK, Maresh EL, Shen D et al. Higher levels of GATA3 predict better survival
in women with breast cancer. Hum Pathol 2010; 41: 17941801.
13. Tokuda E, Seino Y, Arakawa A et al. Estrogen receptor-alpha directly regulates
sensitivity to paclitaxel in neoadjuvant chemotherapy for breast cancer. Breast
Volume 23 | No. 12 | December 2012
original articles
Cancer Res Treat 2011; [epub ahead of print], doi:10.1007/s10549-10011-
11758-x.
14. Ren J, Agata N, Chen D et al. Human MUC1 carcinoma-associated protein
confers resistance to genotoxic anticancer agents. Cancer Cell 2004; 5:
163175.
15. Ciocca V, Daskalakis C, Ciocca RM et al. The signicance of GATA3
expression in breast cancer: a 10-year follow-up study. Hum Pathol 2009; 40:
489495.
16. Ademuyiwa FO, Thorat MA, Jain RK et al. Expression of Forkhead-box protein
A1, a marker of luminal A type breast cancer, parallels low Oncotype DX 21-
gene recurrence scores. Mod Pathol 2010; 23: 270275.
17. Matsukita S, Nomoto M, Kitajima S et al. Expression of mucins (MUC1, MUC2,
MUC5AC and MUC6) in mucinous carcinoma of the breast: comparison with
invasive ductal carcinoma. Histopathology 2003; 42: 2636.
18. Naoi Y, Kishi K, Tanei T et al. Prediction of pathologic complete response to
sequential paclitaxel and 5-uorouracil/epirubicin/cyclophosphamide therapy
using a 70-gene classier for breast cancers. Cancer 2011; 117:
36823690.
19. Muroya K, Hasegawa T, Ito Y et al. GATA3 abnormalities and the phenotypic
spectrum of HDR syndrome. J Med Genet 2001; 38: 374380.
20. Oshima K, Naoi Y, Kishi K et al. Gene expression signature of TP53 but not
its mutation status predicts response to sequential paclitaxel and 5-FU/
epirubicin/cyclophosphamide in human breast cancer. Cancer Lett 2011; 307:
149157.
21. Parker JS, Mullins M, Cheang MC et al. Supervised risk predictor of
breast cancer based on intrinsic subtypes. J Clin Oncol 2009; 27:
11601167.
22. Chou J, Provot S, Werb Z. GATA3 in development and cancer differentiation: cells
GATA have it!. J Cell Physiol 2010; 222: 4249.
23. Albergaria A, Paredes J, Sousa B et al. Expression of FOXA1 and GATA-3 in
breast cancer: the prognostic signicance in hormone receptor-negative tumours.
Breast Cancer Res 2009; 11: R40.
24. Nakshatri H, Badve S. FOXA1 in breast cancer. Expert Rev Mol Med 2009;
11: e8.
25. Badve S, Nakshatri H. Oestrogen-receptor-positive breast cancer: towards
bridging histopathological and molecular classications. J Clin Pathol 2009; 62:
612.
26. Yan W, Cao QJ, Arenas RB et al. GATA3 inhibits breast cancer metastasis
through the reversal of epithelial-mesenchymal transition. J Biol Chem 2010;
285: 1404214051.
27. Buron N, Guery L, Creuzot-Garcher C et al. Trefoil factor TFF1-induced protection
of conjunctival cells from apoptosis at premitochondrial and postmitochondrial
levels. Invest Ophthalmol Vis Sci 2008; 49: 37903798.
28. Vie N, Copois V, Bascoul-Mollevi C et al. Overexpression of phosphoserine
aminotransferase PSAT1 stimulates cell growth and increases chemoresistance
of colon cancer cells. Mol Cancer 2008; 7: 14.
29. Brychtova V, Vojtesek B, Hrstka R. Anterior gradient 2: a novel player in tumor
cell biology. Cancer Lett 2011; 304: 17.
30. Kastl L, Brown I, Schoeld AC. miRNA-34a is associated with docetaxel
resistance in human breast cancer cells. Breast Cancer Res Treat 2011; 131:
445454.
31. Molenaar JJ, Ebus ME, Koster J et al. Cyclin D1 is a direct transcriptional
target of GATA3 in neuroblastoma tumor cells. Oncogene 2010; 29:
27392745.
doi:10.1093/annonc/mds120 |