Sensitization of BCL-2-expressing breast tumors to chemotherapy by the BH3 mimetic ABT-737.

Abstract

Overexpression of the prosurvival protein BCL-2 is common in breast cancer. Here we have explored its role as a potential therapeutic target in this disease. BCL-2, its anti-apoptotic relatives MCL-1 and BCL-XL, and the proapoptotic BH3-only ligand BIM were found to be coexpressed at relatively high levels in a substantial proportion of heterogeneous breast tumors, including clinically aggressive basal-like cancers. To determine whether the BH3 mimetic ABT-737 that neutralizes BCL-2, BCL-XL, and BCL-W had potential efficacy in targeting BCL-2-expressing basal-like triple-negative tumors, we generated a panel of primary breast tumor xenografts in immunocompromised mice and treated recipients with either ABT-737, docetaxel, or a combination. Tumor response and overall survival were significantly improved by combination therapy, but only for tumor xenografts that expressed elevated levels of BCL-2. Treatment with ABT-737 alone was ineffective, suggesting that ABT-737 sensitizes the tumor cells to docetaxel. Combination therapy was accompanied by a marked increase in apoptosis and dissociation of BIM from BCL-2. Notably, BH3 mimetics also appeared effective in BCL-2-expressing xenograft lines that harbored p53 mutations. Our findings provide in vivo evidence that BH3 mimetics can be used to sensitize primary breast tumors to chemotherapy and further suggest that elevated BCL-2 expression constitutes a predictive response marker in breast cancer.

Full text

Page 1

Sensitization of BCL-2–expressing breast tumors to
chemotherapy by the BH3 mimetic ABT-737
Samantha R. Oakesa,b,1, François Vaillanta,b,1, Elgene Lima,2, Lily Leea,b, Kelsey Breslina, Frank Feleppac,
Siddhartha Debd, Matthew E. Ritchiea,b, Elena Takanod, Teresa Warda, Stephen B. Foxd, Daniele Generalie,
Gordon K. Smytha,f, Andreas Strassera,b, David C. S. Huanga,b, Jane E. Visvadera,b,1,3, and Geoffrey J. Lindemana,g,h,1,3
aThe Walter and Eliza Hall Institute of Medical Research, Parkville, Victoria 3052, Australia;bDepartment of Medical Biology, University of Melbourne,
Parkville, Victoria 3010, Australia;cDepartment of Anatomical Pathology, Royal Melbourne Hospital, Parkville, Victoria 3050, Australia;dPeter MacCallum
Cancer Centre, East Melbourne, Victoria 3002, Australia;eU.O. Multidisciplinare di Patologia Mammaria, Laboratorio di Oncologia Molecolare Senologica, AO.
Istituti Ospitalieri di Cremona, Cremona 26100, Italy;fDepartment of Mathematics and Statistics, University of Melbourne, Parkville, Victoria 3010, Australia;
gDepartment of Medical Oncology, Royal Melbourne Hospital, Parkville, Victoria 3050, Australia; andhDepartment of Medicine, University of Melbourne,
Parkville, Victoria 3010, Australia
Edited by Kornelia Polyak, Dana-Farber Cancer Institute, Boston, MA, and accepted by the Editorial Board June 16, 2011 (received for review March 29, 2011)
Overexpression of the prosurvival protein BCL-2 is common in
breast cancer. Here we have explored its role as a potential
therapeutic target in this disease. BCL-2, its anti-apoptotic relatives
MCL-1 and BCL-XL, and the proapoptotic BH3-only ligand BIM were
found to be coexpressed at relatively high levels in a substantial
proportion of heterogeneous breast tumors, including clinically
aggressive basal-like cancers. To determine whether the BH3
mimetic ABT-737 that neutralizes BCL-2, BCL-XL, and BCL-W had
potential efficacy in targeting BCL-2–expressing basal-like triple-
negative tumors, we generated a panel of primary breast tumor
xenografts in immunocompromised mice and treated recipients
with either ABT-737, docetaxel, or a combination. Tumor response
and overall survival were significantly improved by combination
therapy, but only for tumor xenografts that expressed elevated
levels of BCL-2. Treatment with ABT-737 alone was ineffective,
suggesting that ABT-737 sensitizes the tumor cells to docetaxel.
Combination therapy was accompanied by a marked increase in
apoptosis and dissociation of BIM from BCL-2. Notably, BH3 mim-
etics also appeared effective in BCL-2–expressing xenograft lines
that harbored p53 mutations. Our findings provide in vivo evidence
that BH3 mimetics can be used to sensitize primary breast tumors
to chemotherapy and further suggest that elevated BCL-2 expres-
sion constitutes a predictive response marker in breast cancer.
ABT-263|navitoclax|programmed cell death|mammary|small molecule
inhibitor
B
ally exhibit a triple-negative phenotype, as they lack clinically
significant expression of estrogen receptor (ER), progesterone re-
ceptor (PR), and human epidermal growth factor receptor 2
(HER2)/ErbB2, and express basal markers such as cytokeratin 5/6
and epidermal growth factor receptor (EGFR) (1, 2). The absence
of ER, PR, and HER2 expression precludes endocrine or anti-
HER2 therapy. Although a number of cytotoxic drugs (including
the taxanes) are efficacious, prognosis remains significantly worse
for basal-like triple-negative cancers than for other tumor subtypes
(3, 4), highlighting the need for new therapeutic strategies.
Impairment of apoptosis is a hallmark of cancer and can result
in resistance to chemotherapy (5). Tumor resistance to apoptosis
is frequently acquired through deregulated expression of BCL-2
family members or mutations in the p53 tumor suppressor
pathway that ablate the ability of this transcription factor to in-
duce BH3-only proteins (such as PUMA and NOXA), which are
critical for the initiation of apoptosis. There are three main
classes of BCL-2 family regulators: the prosurvival BCL-2–like
proteins; the proapoptotic BH3-only ligands, including BIM that
interacts with all prosurvival proteins; and the proapoptotic
multi-BH domain effector proteins, which activate caspases and
lead to cell demolition (5).
BCL-2 has emerged as an important clinical prognostic marker
in breast cancer (6, 7). BCL-2 gene expression is a component of
asal-like tumors account for ∼20% of breast cancers and usu-
a 21-gene expression assay (Oncotype DX) that is increasingly
being used to predict recurrence of hormone receptor-positive,
node-negative breast cancers (8). Although BCL-2 is commonly
associated with ER-positive breast tumors (9), it can also be
expressed in ER-negative tumors. The frequency of BCL-2 ex-
pression in basal-like tumors has not been well-defined. Although
BCL-2 is a favorable prognostic marker, it is noteworthy that
a significant number of patients with BCL-2–positive disease re-
lapseanddie(7).ApotentialroleforBCL-2asatherapeutictarget
in breast cancer has not been explored.
The pivotal role of the BCL-2 family as arbiters of the intrinsic
apoptotic pathway has stimulated considerable interest in de-
veloping anti-cancer agents that specifically act to induce apo-
ptotic cell death. ABT-737, a small molecule that mimics the
action of the proapoptotic BH3-only proteins, has been shown to
bind and neutralize the prosurvival proteins BCL-2, BCL-XL
and BCL-W but not MCL-1 or A1 (10). ABT-737 has demon-
strated killing potency in combination settings in cell line-based
models (11–17). Furthermore, the potency of ABT-737 with di-
verse chemotherapeutic agents has been shown in the case of
primary leukemia cells and in small-cell lung cancer (SCLC)
primary xenografts (18–22).
Here we report elevated expression of BCL-2 among the
various subtypes of breast cancer, including basal-like tumors,
and investigate the clinical relevance of targeting BCL-2 through
the generation of a panel of primary breast tumor xenografts.
Primary tumor xenografts can recapitulate the phenotype, bi-
ological properties, and drug sensitivity of the original primary
tumor, thereby serving as powerful models for preclinical studies
(23). These primary tumor xenografts offer many advantages
over cell line-based xenografts, which do not consistently predict
the clinical potential of drug treatments. Our studies reveal that
ABT-737 potentiates the effects of docetaxel chemotherapy in
basal-like breast cancers with elevated BCL-2 levels, suggesting
that BH3 mimetics in combination therapy have considerable
potential for the treatment of this aggressive cancer subtype and
other BCL-2–expressing tumors.
Author contributions: S.R.O., F.V., E.L., J.E.V., and G.J.L. designed research; S.R.O., F.V.,
E.L., L.L., K.B., F.F., S.D., E.T., T.W., and S.B.F. performed research; D.G. contributed new
reagents/analytic tools; S.R.O., F.V., E.L., L.L., S.D., M.E.R., S.B.F., G.K.S., A.S., D.C.S.H.,
J.E.V., and G.J.L. analyzed data; and S.R.O., J.E.V., and G.J.L. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission. K.P. is a guest editor invited by the Editorial
Board.
Data deposition: The sequence data presented in Fig. S2 has been deposited in the Gene
Expression Omnibus (GEO) database, www.ncbi.nlm.nih.gov/geo (accession no. GSE28570).
1S.R.O., F.V., J.E.V., and G.J.L contributed equally to this work.
2Present address: Dana-Farber Cancer Institute, Boston, MA 02115.
3To whom correspondence may be addressed. E-mail: visvader@wehi.edu.au or lindeman@
wehi.edu.au.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.
1073/pnas.1104778108/-/DCSupplemental.
www.pnas.org/cgi/doi/10.1073/pnas.1104778108 PNAS Early Edition
| 1 of 6
MEDICAL SCIENCES
SPECIAL FEATURE
BREAST CANCER SPECIAL
FEATURE

Page 2

Results
BCL-2 Is Frequently Expressed at Elevated Levels with BIM, MCL-1, and
BCL-XL in Breast Cancer. The expression of BCL-2, BIM, MCL-1,
and BCL-XL was evaluated in a panel of 197 primary breast
tumors that included 60 luminal, 65 basal-like, 24 HER2-posi-
tive, and 29 marker-null tumors, as defined by immunostaining
for ER, PR, HER2, CK5/6, and EGFR (24). Luminal tumors
were characterized by expression of the steroid hormone
receptors ER and/or PR, basal-like tumors by their triple nega-
tive status for ER, PR, and HER2 and positivity for CK5/6 and
EGFR, whereas HER2-positive tumors displayed intense mem-
brane (3+) staining and/or HER2 amplification as determined
by in situ hybridization. Marker-null tumors lacked expression of
all markers. BCL-2 was expressed in a large proportion (83.3%)
of luminal tumors, consistent with previous reports (7, 25, 26),
and in 50.0 and 41.4% of HER2-positive and marker-null breast
tumors, respectively. Within the basal-like subtype, 18.5% were
found to express BCL-2.
BIM, MCL-1, and BCL-XL were widely expressed among the
different breast tumor subtypes: 96.3, 94.4, and 100% in luminal
tumors; 57.6, 89.8, and 93.4% in basal-like tumors; 95.7, 100, and
100% in HER2-positive tumors; and 65.5, 65.5, and 85.2% in
marker-null tumors, respectively. Concurrent scoring of all four
markers was feasible for 159 tumors (Table 1 and Fig. S1).
Overall, 93.5, 94.7, and 96.2% of BCL-2–positive tumors
expressed BIM, MCL-1, and BCL-XL, respectively: BIM and
MCL-1 were almost invariably coexpressed with BCL-2 in lu-
minal and HER2-positive tumors, whereas more than 75% of
basal-like or marker-null BCL-2–positive tumors coexpressed
these proteins. There were higher numbers of BCL-XL–positive
tumors across all subgroups.
Establishment of Human Breast Cancer Xenografts That Recapitulate
the Primary Tumor. A bank of human breast tumor xenografts was
established by serial passage of primary breast tumor fragments
in the cleared mammary fat pads of immunocompromised NOD-
SCID-IL2Rγc−/−mice. Approximately 25% (28 of 112) of pri-
mary tumors successfully engrafted, a frequency similar to that
previously reported (23). The majority of these were derived
from tumors designated as “triple negative.” A schematic dia-
gram of the strategy used to generate primary tumor xenografts
is shown in Fig. 1A. In most cases, the xenografted tumors were
passaged for two or three rounds.
Five breast-tumor xenograft lines (838T, 24T, 315T, 13T, and
806T) were selected for further analysis. Immunohistochemical
analysis of the original patient tumors (Table S1 and Fig. S1)
revealed that 315T contained foci of intense ER and PR staining
(10% of cells), although 838T, 24T, 315T, and 13T had been
reported as triple negative. The primary tumor 806T was HER2-
positive, as determined by immunohistochemistry and the pres-
ence of HER amplification (11-fold) detected by chromogenic in
situ hybridization (CISH), but lacked ER and PR expression.
Weak granular staining for HER2 was also evident in a small
percentage of cells within the 13T and 24T tumors, although
HER2 was not amplified. Notably, all five primary tumors
expressed CK5/6, and all except 315T were strongly positive for
EGFR. Furthermore, gene expression profiling of primary tumor
xenograft mRNA confirmed the basal-like nature of 838T, 24T,
806T, and 13T and revealed that 315T had a luminal B-like
molecular signature (Fig. S2).
Tumor morphology and marker expression were largely main-
tained in the xenograft models although some differences were
apparent (Fig. 1B and Fig. S1). The cellular profile of each
Table 1.Coexpression of BCL-2, BIM, MCL-1, and BCL-XL in primary breast cancer subtypes
Breast cancer
subtype (no.)BCL-2 expression No. of samples (%) BCL-2 histoscore
% BIM positive
(histoscore)
% MCL-1 positive
(histoscore)
% BCL-XL positive
(histoscore)
Luminal (54)+
−
+
−
+
−
+
−
+
−
45 (83.3)
9 (16.7)
11 (50.0)
11 (50.0)
10 (18.5)
44 (81.5)
12 (41.4)
17 (58.6)
78 (49.1)
81 (50.9)
6.89 100 (6.73)
88.9 (4.78)
100 (5.09)
90.0 (5.30)
77.8 (4.78)
58.1 (2.09)
75.0 (2.92)
58.8 (3.65)
93.5 (5.68)
65.8 (3.14)
97.7 (5.98)
88.9 (6.11)
100 (5.00)
100 (6.55)
88.9 (5.00)
92.9 (4.38)
83.3 (3.17)
52.9 (2.82)
94.7 (5.28)
84.8 (4.54)
100 (8.74)
100 (6.89)
100 (8.60)
100 (8.20)
100 (6.00)
89.5 (6.29)
83.3 (4.17)
86.7 (4.80)
96.2 (7.49)
92.1 (6.17)
HER2 (22) 5.09
Basal-like (54)3.70
Marker-null (29)5.00
All tumors (159)5.96
Of 197 tumor samples in the tissue microarrays, scoring of all four markers was feasible for 159 tumors. For BCL-2 expression: (+) positive, (−) negative. The
numbers in parentheses for BIM, MCL-1, and BCL-XL are the mean histoscores. Scoring of samples is described in SI Materials and Methods.
A
B
Fig. 1.
Schematic model of the derivation of human breast tumor xenograft mod-
els. Human breast tumor tissue was implanted into the mammary fat pads
of immunocompromised NOD-SCID-IL2Rγc−/−mice. Upon successful engraft-
ment (explant 1), tumor tissue fragments were prepared and sequentially
passaged in mice (explants 2 and 3). (B) Immunohistochemistry of ER, PR,
HER2, cytokeratin 5/6 (CK5/6), and EGFR expression in breast tumor xeno-
grafts denoted 838T, 24T, 315T, 13T, and 806T. Expression of PR, HER2, CK5/
6, and EGFR replicated that observed in the original patient tumors (Fig. S1),
with the exception of 806T, in which the HER2-amplified fraction failed
to passage. Low-level ER expression was detected in xenografts, possibly
reflecting the use of estradiol pellets. (Scale bars, 100 μm.)
The establishment of a panel of human breast tumor xenografts. (A)
2 of 6
| www.pnas.org/cgi/doi/10.1073/pnas.1104778108 Oakes et al.

Page 3

xenograft was similar to that of the corresponding primary pa-
tient sample, as determined by flow cytometric analysis using
antibodies directed against CD49f (α6-integrin) and EpCAM
(CD326/epithelial specific antigen) (27) (Figs. S1 and S3). Weak
nuclear ER staining, however, was detected in four of the xen-
ografted tumors, with highest levels evident in the 315T xeno-
graft line. These findings may reflect the use of estradiol pellets
in the host mice, perhaps leading to the selection of ER-positive
tumor cells or slight activation of ER itself. Substantially lower
levels of HER2 were observed in the 806T primary tumor xe-
nograft compared with the original tumor. Moreover, CISH for
HER2 gene amplification did not reveal any amplification, in-
dicating that selection had occurred against HER2-amplified
cells during xeno-transplantation.
Variable BCL-2 and p53 Expression in Breast Tumor Xenografts. We
next investigated the expression profile of BCL-2 family mem-
bers and p53 among the different breast cancer xenografts by
Western blot analysis (Fig. 2A). BCL-2, BIMELBIML, and BIMS
were found to be differentially expressed, with barely detectable
BCL-2 in the 806T tumor xenografts. NOXA and PUMA, the
two BH3-only proteins that are transcriptionally up-regulated by
p53, were also variably expressed. MCL-1, BCL-XL, and the
apoptotic effectors BAX and BAK were present in all of the
tumor xenografts. Consistent with the Western blot data and
their quantification, immunostaining of the parental primary
tumors revealed strong BCL-2 expression in the 838T and 24T
tumors, focal staining in 13T and 315T tumors, and undetectable
expression in the 806T tumor xenografts (Fig. 2B).
High levels of the tumor suppressor p53 were detected in the
838T tumor xenograft by Western blot analysis and moderate
levels in the 24T and 315T xenograft lines (Fig. 2A). Sequencing
of tumor genomic DNA revealed pathogenic mutations in the
p53 gene in tumors 838T (c.581T > G, 194Leu > Arg, exon 6)
and 24T (c.404T > G, 135Cys > Phe, exon 5), with loss of het-
erozygosity for both mutations. The p53 pathway is therefore
disabled in the 838T and 24T basal-like tumor xenograft models.
Combination Therapy Induces Stable Regression of Breast Cancer.
BecausethreeofthefivexenograftsdisplayedhighlevelsofBCL-2,
we examined whether the BH3 mimetic ABT-737, either alone or
in combination with docetaxel, was efficacious in killing BCL-2–
positivebreasttumors.Adosetitrationwasfirstperformedonmice
bearing 838T and 13T tumors using 50 or 75 mg/kg ABT-737 with
0, 5, 10, or 20 mg/kg docetaxel to establish dose-limiting toxicity.
This was determined by >10% weight loss following therapy and
the extent of ABT-737–induced thrombocytopenia (28). On the
basisofthesestudies,ABT-737at50mg/kganddocetaxelat10mg/
kg were selected as suitable doses for tumor response studies. This
dosage induced tolerable levels of thrombocytopenia (Table S2
and Fig.S4).Notably,docetaxel alone was efficacious inproducing
a tumor response in the 838T tumor xenografts (median survival
was 76 versus 10 d for control vehicle-treated animals; P < 0.0001)
but had little effect in other models (Fig. 3).
ABT-737 administered in combination with docetaxel resulted
in significantly improved animal survival in four of the five tumor
xenograft models compared with treatment with docetaxel alone.
For the models in which little effect was observed with a single
cycle of chemotherapy, no further cycles were administered. The
most profound response was observed in mice bearing the basal-
like tumor xenografts 838T and 24T, in which tumor growth was
dramatically inhibited by thecombination treatment(Fig. 3A) and
animal survival was prolonged (Fig. 3B). A partial response was
evident in mice bearing the basal-like xenograft 13T and the lu-
minal 315Ttumor xenograft. Combination therapy had little effect
on the 806T tumor xenograft, which expresses virtually undetect-
ablelevelsofBCL-2(Figs.2and3).BIMwasexpressedinalltumor
xenografts with high levels present in the most responsive models.
ABT-737 as a single agent was ineffective in all five xenograft
models. Thus, the combination of ABT-737 with docetaxel proved
to be efficacious in eliciting a tumor response and prolonging an-
imal survival in the case of BCL-2–expressing breast cancers.
Given that BCL-2 has been reported to be an estrogen-
responsivegene,tumorresponsewasfurtherevaluatedinthe838T
and 24T xenografts passaged in the presence or absence of es-
trogen pellets. As noted above, the presence of an estrogen pellet
appeared to induce a low level of ER expression in 838T and 24T
xenografted tumors. However, BCL-2 levels did not change
according to Western blot analysis (Fig. S5), and no significant
difference in the rate of tumor growth was observed over several
weeks, irrespective of whether an estrogen pellet was present.
To investigate the durability of the therapeutic response, the
838T tumor xenograft model was evaluated for the acquisition of
resistance over six cycles. Estrogen pellets were not implanted
into mice for these extended experiments as mice succumbed to
renal damage, bladder stone formation, and uterine enlargement
with chronic estrogen exposure, as reported previously (29).
Treatment of mice bearing 838T xenografts with docetaxel alone
showed early recurrence, whereas tumors subjected to combi-
nation therapy remained undetectable until ∼75 d, at which time
8 of 10 tumors resumed growth. Five tumors were re-treated for
a further three cycles, and regression was again evident in four
cases (Fig. S6A). Tumor regression was also evident in mice
bearing 24T tumor xenografts over two cycles of combination
therapy until the end point at 70 d (Fig. 3).
Increased Apoptosis Accompanies the Response to Combination
Therapy. To investigate the extent and kinetics of apoptosis dur-
ingtherapy,micebearing838TxenograftsweretreatedwithABT-
737, docetaxel, or both drugs, and tumors were analyzed at 24 h,
48 h, and 8 d posttreatment (Fig. S6B). Immunohistochemistry
revealed increased numbers of cleaved (i.e., activated) caspase-3
apoptoticfociintumorstreatedwithdocetaxelaloneorcombined
therapy compared with administration of ABT-737 alone or ve-
hicle at 24 and 48 h postinjection [P < 0.001 (n = 3) and P < 0.05
(n = 2), respectively]. Following combination therapy for 8 d, few
apoptotic foci were visible, consistent with tumor regression.
Similarresultswereobserved upontreatment ofmicebearing24T
tumor xenografts with combination therapy (Fig. 4). Docetaxel
treatment of 838T and 24T tumor xenografts exerted inter-
mediate effects compared with combination therapy (P = 0.22
and P < 0.01, respectively). No difference in the numbers of ap-
optotic cells was apparent for the 315T or 806T tumor xenograft
A
B
28
Fig. 2.
grafts. (A) Western blot analysis of p53, BCL-2, MCL-1, BCL-XL, BIM extra long
(EL), long (L) and short (S), PUMA, NOXA, BAK, and BAX in five breast tumor
xenografts:838T,24T,315T,13T,and806T(Left).Probingforβ-ACTINwasused
as a loading control. Densitometry of three independent tumors for each xe-
nograftlinerevealedvariationinthelevelsofBCL-2andBIMexpression(Right).
(B) Immunohistochemistry of BCL-2 expression in primary patient breast
tumors:838T,24T,and806T.(Scalebar,100μm.)(Insets)Adjacentnormaltissue.
Variable expression of BCL-2 family members in breast tumor xeno-
Oakes et al. PNAS Early Edition
| 3 of 6
MEDICAL SCIENCES
SPECIAL FEATURE
BREAST CANCER SPECIAL
FEATURE

Page 4

models in response to combination therapy relative to vehicle
(P = 0.21 and P = 0.97, respectively). Thus, the extent of apoptosis
correlated with response efficacy as measured by tumor growth rate
and animal survival (Figs. 3 and 4).
Western blot analysis of BCL-2, MCL-1, BCL-XL, BIM,
PUMA, NOXA, BAK, and BAX expression in the tumor xeno-
graft models did not reveal any significant differences following
the various treatments (Fig. 4C). However, coimmunoprecipita-
tion coupled with Western blot analysis revealed a rapid re-
duction in the amount of BIM bound to BCL-2 following ex-
posure to docetaxel, ABT-737, or combination therapy in the
most responsive tumor xenograft line 838T (Fig. 4D). A decrease
in BCL-2–associated BIM was also observed in 24T tumor xen-
ografts following treatment with ABT-737 or combination ther-
apy, indicating that perturbation of BIM protein complexes in
these breast tumors may contribute to activation of the intrinsic
apoptotic cascade. Moreover, treatment of the breast cancer cell
line MDA-MB-231, which responded synergistically to combi-
nation therapy, clearly implicated changes in the BCL-2/BIM
complex as a key molecular mechanism underlying the response,
whereas no change was observed in BCL-XL/BIM complexes
(Fig. 4E). Knockdown of either BCL-2 or BIM in MDA-MB-231
cells by siRNAs confirmed the importance of these proteins in
mediating responsiveness to combination therapy (Fig. S7).
Discussion
Here we have established primary breast tumor xenograft mod-
els, including basal-like cancers, that recapitulate the original
patient tumors. Expression analysis revealed that BCL-2 protein
expression was highly variable among this panel of breast tumor
xenografts, thus providing a platform to test the efficacy of the
BH3 mimetic ABT-737 in BCL-2–expressing breast cancers. We
demonstrate here that tumors expressing high BCL-2 respond
effectively to ABT-737 in combination with docetaxel chemo-
therapy, whereas the tumor xenograft model expressing very low
BCL-2 levels showed no response. BIM also appeared to be
more abundant in the responsive tumor xenografts. Conversely,
BCL-XL, another reported target of ABT-737, was expressed in
both responsive and nonresponsive tumors. Notably, significant
inhibition of tumor growth was evident in three basal-like tumor
xenograft lines. Of these, a profound response was observed in
two tumor xenograft lines (838T and 24T) and partial responses
in two further tumor xenograft lines derived from basal-like and
luminal cancers. Although ABT-737 has demonstrated single-
agent efficacy in some cancers (predominantly cell lines), in-
cluding follicular lymphoma and chronic lymphocytic leukemia
(CLL) (10, 30–32), ABT-737 was not effective as a single agent
for any of the primary breast tumor xenograft models tested.
The mechanism by which ABT-737 synergizes with docetaxel
to cause tumor regression is yet to be established but is likely to
involve destabilization of interactions between BIM, BCL-2 and
MCL-1. Combined treatment with docetaxel and ABT-737,
which competitively binds the hydrophobic groove of BCL-2,
BCL-XL, and BCL-W (but not MCL-1 or A1), led to release of
BIM from BCL-2. These findings are consistent with the recent
observation that paclitaxel-induced killing of breast cancer cells
in vitro is associated with competitive displacement of BIM from
anti-apoptotic proteins by BH3-only proteins or ABT-737 (33).
In lymphoma cell lines, ABT-737 responsiveness and BCL-2
dependence correlated with high levels of BIM sequestered by
BCL-2 (34), and in CLL cells from patients, BCL-2 complexed to
BIM proved to be the pivotal target for ABT-737 (31). Fur-
A
B
Fig. 3.
creased tumor growth and increased survival of mice
bearing BCL-2–expressing breast tumor xenografts. (A) Tu-
mor growth curves and (B) Kaplan–Meier survival curves of
mice bearing 838T, 24T, 315T, 13T, and 806T human breast
tumor xenografts treated with vehicle alone (black line),
ABT-737 (50 mg/kg) plus vehicle for docetaxel (green line),
docetaxel (10 mg/kg) plus vehicle for ABT-737 (blue line), or
combined ABT-737 (50 mg/kg) and docetaxel (10 mg/kg,
red line). Estrogen pellets were included for all experiments
except those shown for 838T; no difference in tumor
growth or animal survival was observed for 838T in the
presence or absence of estrogen pellets. Log-rank (Mantel–
Cox) P values are displayed for combination therapy versus
vehicle. Black bars in A indicate a treatment cycle of ABT-
737 (days 1–10) and docetaxel (day 1).
ABT-737 combined with docetaxel results in de-
4 of 6
| www.pnas.org/cgi/doi/10.1073/pnas.1104778108Oakes et al.

Page 5

thermore, only cells dependent on BH3-activator proteins bound
to BCL-2 but not MCL-1 were found to be sensitive to ABT-
737–induced killing. Thus, in primary breast cancers that express
elevated BCL-2 levels, chemo-sensitization by ABT-737 may
require BIM occupation of BCL-2 to prime cancer cells for
death. Interestingly, acquired resistance in SCLC xenografts was
linked to decreased expression of BCL-2, BAX, and BIM, and a
perturbation of BCL-2–BIM complexes (18). Although MCL-1
expression in the breast cancer xenograft models did not corre-
late with responsiveness, a role for MCL-1 cannot be excluded.
The p53 status of basal-like breast cancers may also impact on
sensitivity to a BH3 mimetic and chemotherapy, as the most
responsive of the human breast tumor xenograft lines harbored
p53 mutations. Indeed, p53 mutations in breast tumors have
been linked with responsiveness to paclitaxel (35), although an in
silico analysis comparing changes in gene expression with clinical
drug response did not find an association between p53 status and
chemoresponsiveness (36). Notably, it has been speculated that
ABT-737 is more effective in tumors harboring p53 mutations,
because ABT-737 acts downstream of p53 where an intact apo-
ptotic cascade is present (37). A key point to emerge from this
study is that combination therapy using BH3 mimetics can be
very effective in tumors harboring p53 mutations.
Our data reveal that BCL-2 expression in breast cancer may serve
as a predictive biomarker for responsiveness to ABT-737 combined
with docetaxel chemotherapy. Interestingly, BCL-2 was found to be
highly expressed in ∼20% of basal-like primary breast cancers and
was often associated with abundant BIM. Expression of these pro-
teins may identify a subset of breast cancer patients who are likely
to benefit most from treatment with the orally bioavailable BH3
mimeticABT-263(Navitoclax).BCL-XLandMCL-1levelsmayalso
contribute to modulating the response to ABT-737, as exemplified
by their roles in other tumor types. Similar to ABT-737, ABT-263
can induce complete tumor regression in certain xenograft models,
either as a single agent or combined with clinically relevant drugs
(38), and is currently undergoing several phase I/II clinical trials.
In summary, we have developed preclinical models of primary
breast cancer, including basal-like and luminal tumor xenografts,
and provide in vivo evidence that ABT-737 can be used to sen-
sitize primary breast tumors to taxane therapy. Combination
therapy resulted in more durable responses and improved overall
survival. Our results provide a rationale for the development of
clinical protocols evaluating ABT-263 as an adjunct to conven-
tional chemotherapy in BCL-2–expressing basal-like and luminal
breast cancers.
Materials and Methods
Detailed methods are described in SI Materials and Methods.
Human Xenograft Establishment. Mice were anesthetized with ketamine/
xylazine (200 mg/kg)/(20 mg/kg) i.p., and analgesia was administered with
carprofen (5 mg/kg) subcutaneously. Human breast tumor fragments (0.5–1
mm × 0.5–1 mm × 5–8 mm) were inserted into inguinal mammary fat pads
of 3- to 4-wk-old NOD-SCID-IL2Rγc−/−female mice. Silastic estrogen pellets
(0.5 mg) were prepared as previously described (39) and implanted sub-
cutaneously at the time of surgery. Following engraftment, tumors were
A
B
C
D
E
25
Fig. 4.
creased apoptosis and dissociation of BIM from
BCL-2. (A) Immunohistochemical staining for
cleaved caspase-3 of 838T, 24T, 315T, and 806T
xenografts treated with vehicle alone, ABT-737
(50 mg/kg) plus vehicle for docetaxel, docetaxel
(10 mg/kg) plus vehicle for ABT-737, or com-
bined ABT-737 (50 mg/kg) and docetaxel (10 mg/
kg) collected at 24 h. (Scale bar, 50 μm.) (B)
Analysis of the number of cleaved caspase-3–
positive apoptotic foci [*P < 0.01, **P < 0.001,
and NS (not significant) of each column mean
compared with vehicle]. (C) Western blot analy-
sis of BCL-2, MCL-1, BCL-XL, BIM extra long (EL),
long (L) and short (S), PUMA, NOXA, BAK, and
BAX in 838T, 24T, 315T, and 806T tumor xeno-
grafts collected at 24 h after combined treat-
ment with ABT-737 and docetaxel. Probing for
β-ACTIN was used as a loading control. (D) BCL-2
immunoprecipitates from 838T and 24T tumor
lysates collected 24 h posttreatment with ABT-
737 and/or docetaxel were analyzed by Western
blotting using antibodies against BIM and BCL-2.
(E) BCL-2, BIM, and BCL-XL immunoprecipitates
from MDA-MB-231 cells collected 24 h post-
treatment with ABT-737 and/or docetaxel and
analyzed by Western blotting using antibodies
against BIM, BCL-2, or BCL-XL.
Combination therapy results in in-
Oakes et al.PNAS Early Edition
| 5 of 6
MEDICAL SCIENCES
SPECIAL FEATURE
BREAST CANCER SPECIAL
FEATURE

Page 6

minced and digested in 150 U/mL collagenase (Sigma) and 50 U/mL hyal-
uronidase (Sigma) for 1–1.5 h at 37 °C. The resulting organoid suspension
was sequentially digested with 0.25% trypsin 1 mM EGTA and 5 mg/mL
dispase (Roche Diagnostics) for 1 min at 37 °C. A single-cell suspension was
obtained by filtration (40 μm), and, where required, red blood cells were
removed by lysis. Cell sorting is described in SI Materials and Methods.
Immunohistochemistry. Human breast cancer xenografts were collected and
fixed in 4% paraformaldehyde before embedding in paraffin. Sections were
subjected to antigen retrieval and then incubated with antibodies against ER
(Novocastra), PR (Novocastra), HER2 (Dako), cytokeratin 5/6 (Dako), EGFR
(EGFR.25, Novocastra), BCL-2 (BCL-2–100; Alexis/Enzo), BIM (3C5; Alexis/Enzo),
MCL-1 (Alexis/Enzo), or BCL-XL (54H6; Cell Signaling) for 30 min at room
temperature, followed by biotinylated anti-IgG secondary antibodies (Vector
Labs). Signal detection was performed using ABC Elite (Vector Labs) for 20 min
and 3,3′-diaminobenzidine (Dako) for 5 min at room temperature.
Tumor Monitoring and Chemotherapy Administration. Cohorts of 24–32 female
mice were seeded with single-cell suspensions of human breast tumors.
Treatment was initiated when the tumor volume reached 100–150 mm3. For
survival studies, mice were randomized into four groups: vehicle for both
docetaxel and ABT-737, docetaxel (10 mg/kg) plus vehicle for ABT-737, ABT-
737 (50 mg/kg) plus vehicle for docetaxel, and ABT-737 (50 mg/kg) plus
docetaxel (10 mg/kg). ABT-737 (or vehicle) was injected i.p. daily for 10 d,
and a single dose of docetaxel was injected i.p. 4 h after the initial injection
of ABT-737. A cycle of treatment refers to 21 d after the initiation of
treatment. For short-term experiments, mice were collected at 24 h after
injection of docetaxel. Eye bleeds were performed at 2, 8, or 11 d after
initiation of treatment, and whole-blood count analysis was performed us-
ing an ADVIA 120 hematological analyzer (Bayer). Tumors were collected at
the ethical end-point volume of 500 mm3.
Western Blot Analysis. Tumors were homogenized in lysis buffer (20 mM
Tris·HCL, 135 mM NaCl, 1.5 mM MgCl2, 1 mM EGTA, 1% Triton X-100, and
10% glycerol). Western blot analysis was performed using 30 μg protein lysate
per lane; membranes were probed with antibodies (listed in SI Materials
and Methods), and detection was performed using HRP-conjugated anti-IgG
secondary antibodies and ECL (GE Healthcare Life Sciences). Densitometry was
performed using the mean gray value of inverted scanned images of Western
blots in triplicate.
ACKNOWLEDGMENTS. We are grateful to K. Lowes, B. Pal, J. Corbin,
M. Chapman, and the Royal Melbourne Hospital Tissue Bank staff for expert
technical assistance; K. Stoev and A. Morcom for animal care; and S. Mihaj-
lovic and E. Tsui for histology support. We thank P. Bouillet for discussions
and Abbott Laboratories (S. Rosenberg, S. Elmore, and colleagues) and Gen-
entech (W. Fairbrother) for providing ABT-737 and discussions. Primary tu-
mor samples were provided by the Victorian Cancer Biobank (VCB). This
work was supported by the Victorian Government through the Victorian
Cancer Agency/Victorian Breast Cancer Research Consortium (J.E.V., G.J.L.,
and S.B.F.) and an Operational Infrastructure Support grant, the National
Health and Medical Research Council Australia (NHMRC; to J.E.V., G.J.L., A.S.,
and D.C.S.H.) and the Australian Cancer Research Foundation. S.R.O. was
supported by the NHMRC and the National Breast Cancer Foundation
(NBCF); E.L. by the NHMRC and NBCF; L.L. by the NHMRC and the Cancer
Therapeutics CRC; S.D. by the VCB; and A.S., D.C.S.H., J.E.V., and G.J.L. by
NHMRC Research Fellowships.
1. Reis-Filho JS, Tutt AN (2008) Triple negative tumours: A critical review. Histopathol-
ogy 52:108–118.
2. Cheang MC, et al. (2008) Basal-like breast cancer defined by five biomarkers has
superior prognostic value than triple-negative phenotype. Clin Cancer Res 14:
1368–1376.
3. Carey LA, et al. (2007) The triple negative paradox: Primary tumor chemosensitivity of
breast cancer subtypes. Clin Cancer Res 13:2329–2334.
4. Dent R, et al. (2007) Triple-negative breast cancer: Clinical features and patterns of
recurrence. Clin Cancer Res 13:4429–4434.
5. Adams JM, Cory S (2007) The Bcl-2 apoptotic switch in cancer development and
therapy. Oncogene 26:1324–1337.
6. Callagy GM, et al. (2006) Bcl-2 is a prognostic marker in breast cancer independently
of the Nottingham Prognostic Index. Clin Cancer Res 12:2468–2475.
7. Dawson SJ, et al. (2010) BCL2 in breast cancer: A favourable prognostic marker across
molecular subtypes and independent of adjuvant therapy received. Br J Cancer 103:
668–675.
8. Paik S, et al. (2004) A multigene assay to predict recurrence of tamoxifen-treated,
node-negative breast cancer. N Engl J Med 351:2817–2826.
9. Bhargava V, Kell DL, van de Rijn M, Warnke RA (1994) Bcl-2 immunoreactivity in
breast carcinoma correlates with hormone receptor positivity. Am J Pathol 145:
535–540.
10. Oltersdorf T, et al. (2005) An inhibitor of Bcl-2 family proteins induces regression of
solid tumours. Nature 435:677–681.
11. Chipuk JE, et al. (2008) Mechanism of apoptosis induction by inhibition of the anti-
apoptotic BCL-2 proteins. Proc Natl Acad Sci USA 105:20327–20332.
12. Cragg MS, et al. (2008) Treatment of B-RAF mutant human tumor cells with a MEK
inhibitor requires Bim and is enhanced by a BH3 mimetic. J Clin Invest 118:3651–3659.
13. Kim KW, Moretti L, Mitchell LR, Jung DK, Lu B (2009) Combined Bcl-2/mammalian
target of rapamycin inhibition leads to enhanced radiosensitization via induction of
apoptosis and autophagy in non-small cell lung tumor xenograft model. Clin Cancer
Res 15:6096–6105.
14. Li M, et al. (2010) Combined inhibition of Notch signaling and Bcl-2/Bcl-xL results in
synergistic antimyeloma effect. Mol Cancer Ther 9:3200–3209.
15. Okumura K, Huang S, Sinicrope FA (2008) Induction of Noxa sensitizes human
colorectal cancer cells expressing Mcl-1 to the small-molecule Bcl-2/Bcl-xL inhibitor,
ABT-737. Clin Cancer Res 14:8132–8142.
16. Paoluzzi L, et al. (2008) The BH3-only mimetic ABT-737 synergizes the antineoplastic
activity of proteasome inhibitors in lymphoid malignancies. Blood 112:2906–2916.
17. Witham J, et al. (2007) The Bcl-2/Bcl-XL family inhibitor ABT-737 sensitizes ovarian
cancer cells to carboplatin. Clin Cancer Res 13:7191–7198.
18. Hann CL, et al. (2008) Therapeutic efficacy of ABT-737, a selective inhibitor of BCL-2,
in small cell lung cancer. Cancer Res 68:2321–2328.
19. Kang MH, et al. (2007) Activity of vincristine, L-ASP, and dexamethasone against
acute lymphoblastic leukemia is enhanced by the BH3-mimetic ABT-737 in vitro and
in vivo. Blood 110:2057–2066.
20. Kohl TM, et al. (2007) BH3 mimetic ABT-737 neutralizes resistance to FLT3 inhibitor
treatment mediated by FLT3-independent expression of BCL2 in primary AML blasts.
Leukemia 21:1763–1772.
21. Kuroda J, et al. (2007) Apoptosis-based dual molecular targeting by INNO-406,
a second-generation Bcr-Abl inhibitor, and ABT-737, an inhibitor of antiapoptotic Bcl-
2 proteins, against Bcr-Abl-positive leukemia. Cell Death Differ 14:1667–1677.
22. Mason KD, et al. (2009) The BH3 mimetic compound, ABT-737, synergizes with a
range of cytotoxic chemotherapy agents in chronic lymphocytic leukemia. Leukemia
23:2034–2041.
23. Marangoni E, et al. (2007) A new model of patient tumor-derived breast cancer
xenografts for preclinical assays. Clin Cancer Res 13:3989–3998.
24. Nielsen TO, et al. (2004) Immunohistochemical and clinical characterization of the
basal-like subtype of invasive breast carcinoma. Clin Cancer Res 10:5367–5374.
25. Bertolo C, et al. (2008) Differences and molecular immunohistochemical parameters
in the subtypes of infiltrating ductal breast cancer. Am J Clin Pathol 130:414–424.
26. Callagy GM, Webber MJ, Pharoah PD, Caldas C (2008) Meta-analysis confirms BCL2 is
an independent prognostic marker in breast cancer. BMC Cancer 8:153.
27. Lim E, et al. (2009) Aberrant luminal progenitors as the candidate target population
for basal tumor development in BRCA1 mutation carriers. Nat Med 15:907–913.
28. Mason KD, et al. (2007) Programmed anuclear cell death delimits platelet life span.
Cell 128:1173–1186.
29. Kang JS, et al. (2009) Low dose estrogen supplementation reduces mortality of mice
in estrogen-dependent human tumor xenograft model. Biol Pharm Bull 32:150–152.
30. Certo M, et al. (2006) Mitochondria primed by death signals determine cellular ad-
diction to antiapoptotic BCL-2 family members. Cancer Cell 9:351–365.
31. Del Gaizo Moore V, et al. (2007) Chronic lymphocytic leukemia requires BCL2 to se-
quester prodeath BIM, explaining sensitivity to BCL2 antagonist ABT-737. J Clin Invest
117:112–121.
32. Konopleva M, et al. (2006) Mechanisms of apoptosis sensitivity and resistance to the
BH3 mimetic ABT-737 in acute myeloid leukemia. Cancer Cell 10:375–388.
33. Kutuk O, Letai A (2008) Alteration of the mitochondrial apoptotic pathway is key
to acquired paclitaxel resistance and can be reversed by ABT-737. Cancer Res 68:
7985–7994.
34. Deng J, et al. (2007) BH3 profiling identifies three distinct classes of apoptotic blocks
to predict response to ABT-737 and conventional chemotherapeutic agents. Cancer
Cell 12:171–185.
35. Kandioler-Eckersberger D, et al. (2000) TP53 mutation and p53 overexpression for
prediction of response to neoadjuvant treatment in breast cancer patients. Clin
Cancer Res 6:50–56.
36. Sekine I, Shimizu C, Nishio K, Saijo N, Tamura T (2009) A literature review of molecular
markers predictive of clinical response to cytotoxic chemotherapy in patients with
breast cancer. Int J Clin Oncol 14:112–119.
37. Cory S, Huang DC, Adams JM (2003) The Bcl-2 family: Roles in cell survival and on-
cogenesis. Oncogene 22:8590–8607.
38. Tse C, et al. (2008) ABT-263: A potent and orally bioavailable Bcl-2 family inhibitor.
Cancer Res 68:3421–3428.
39. Laidlaw IJ, et al. (1995) The proliferation of normal human breast tissue implanted
into athymic nude mice is stimulated by estrogen but not progesterone. Endocri-
nology 136:164–171.
6 of 6
| www.pnas.org/cgi/doi/10.1073/pnas.1104778108Oakes et al.