Eresearch.jcu.edu.au

EXECUTIVE SUMMARY(1 page maximum)
Please provide an outline of the proposed program of research and the case for award. Significance
Breast cancer affects 10000 Australian women each year, and despite the gains of past decades, breast cancer
remains an intractable disease. The incidence continues to rise, at a rate that increases every year. There are
groups of women for whom available treatments offer little hope, because of the aggressive biological
characteristics of their breast cancer. Finally, development of treatment resistance is common. Bold new
approaches are required to make a major impact on subgroups currently less well served by current treatments,
including premenopausal ER-PR- cancers, and the cancers that now recur after treatment. The lessons learned in
targeting the estrogen pathway can be applied again, as a crucial aspect of estrogen receptor (ER) biology has
been largely overlooked in breast cancer. ER forms part of a highly conserved nuclear receptor (NR) superfamily
of transcription factors. NRs serve as master regulators of virtually every aspect of life, from embryogenesis,
through to reproduction, metabolism and cell death. NRs are highly related, sharing the same structural features,
and utilise the same pool of co-regulator partners to regulate transcription. Many NRs and receptor co-regulators
are expressed in breast cancers, and are linked to clinical features including outcome. Importantly, many NRs are
more frequently expressed in breast cancers than are ER and PR, and their expression is often largely
independent of ER and PR. Moreover, except for the ovarian hormones, the ligands and metabolic signals that
drive these receptors are not linked to menopause and are likely to be circulating throughout life in all women.
Targeting NR pathways provides a promising new avenue in breast cancer management.
Hypotheses
• That the NR family of transcriptional regulators and partner proteins, with fundamental roles in normal physiology, also mediate critical aspects of breast cancer biology • That there are critical points of convergence in the pathways regulated by the NR system in breast cancers, and that these can only be identified using systems biology • That identifying and targeting critical points of convergence in NR pathways in breast cancers will open new directions in breast cancer management, as follows: o NR targets in normal breast are likely to be critical to normal function and amenable to cancer o ER-independent NR targets expressed in breast cancers are likely to be amenable to development of new small molecule agents to selectively target groups of breast cancers with currently poor prognosis and limited treatment options, such as ER-PR- premenopausal breast cancer. o NR pathways that cross-talk to ER signalling will provide new options to interrupt estrogen signalling, particularly upon treatment failure on current endocrine therapies. o Response rates to current endocrine and cytotoxic treatments, particularly in advanced disease with currently few effective options, are likely to be improved by rational combination of existing treatments with currently available NR agonists. Project objective
This application will identify the NR networks active in breast cancers, to identify new predictive targets of
response to treatment in women with ER-PR- premenopausal breast cancer, or women with metastatic disease.
The knowledge will be translated within the granting period, through the application of treatments with existing
NR ligands in women who have failed on endocrine or cytotoxic treatments. Beyond the scope of this funding
period, new rational small molecule agents will be developed and tested in Phase 1/2/3 trials in currently
underserved breast cancer subgroups.
Resources: A major strength of this application is that it brings together established research groups from around
the country, with currently non-overlapping, internationally recognised, programs of research in NR biology.
Tissues and assays: Breast cancer tissues and cohorts, models of human breast and functional assays for NRs are
all in place, Facilities: Low density arrays, gene expression and methylation profiling, animal models, high level
bioinformatics and proteomics are all in place, Project management: Project management, clinical advisory and
international advisory committees are all in place.
DETAILED SCIENTIFIC DESCRIPTION OF THE PROGRAM INCLUDING:
(15 page maximum – applicants are encouraged to be as succinct as possible – utilising the maximum word
count is not a requirement)
7.1 Priorities and objectives of the program 7.2 Details of potential avenues of investigation, including specific project/s, rationale and strategic significance for the prevention and/or management of breast cancer, proposed methods, and key collaborators (networks or partnerships) 7.3 Milestones and timelines, including a Gantt chart Significance
Breast cancer affects 10000 Australian women each year, and breast cancer diagnosis exerts a profound physical,
emotional and financial impact on women, their families and the community. In recognition to this, research over
preceding decades delivered significant gains in better treatments and early diagnosis, and these resulted in a
long-awaited turning point in the 1990s, the beginning of a decline in the death rate from breast cancer.
Despite these advances, breast cancer remains an intractable disease. The causes are still largely unknown, so prevention remains an elusive goal for the majority of women. The incidence continues to rise, at a rate that increases every year and is linked to as yet unknown factors within our affluent lifestyle. There are groups of women for whom available treatments offer little hope, because of the aggressive biological characteristics of their breast cancer. Finally, while a large proportion of women respond to current treatments, development of treatment resistance is also common. Breast cancers are fundamentally reliant on ovarian hormones for full development, and women without ovaries have a dramatically reduced cancer risk. This relationship forms the basis of the use of tamoxifen and the aromatase inhibitors, to interrupt the estrogen pathway, and these modalities have been strikingly successful. The importance of the circulating hormonal environment in modulating breast cancer risk is further underscored by the reduction in breast cancer incidence that followed the sharp decline in use of hormone replacement therapy in the USA in 2003 {Ravdin, 2007 #86}. A similar reduction in incidence over the same time period has been observed in NSW (Cancer Institute NSW 2007). However, while a significant proportion of breast cancers are suitable for antiestrogen-based treatments, younger premenopausal women, and women with aggressive cancers, do not derive benefit, and rational treatments, and treatment predictors, to target these subgroups of breast cancers are required. Because so much progress has been made already in breast cancer, new and bold approaches are required to make a major impact on subgroups currently less well served by current treatments, and the cancers that now recur after treatment. The lessons learned in targeting the estrogen pathway can be applied again, as a crucial aspect of ER biology has been largely overlooked in breast cancer. ER forms part of a nuclear receptor superfamily of transcription factors that are highly conserved in all species from worms to human {Germain, 2003 #92}. Nuclear receptors serve as master regulators of virtually every aspect of life, from embryogenesis, through to reproduction, metabolism and cell death. Nuclear receptors are highly related, sharing the same structural features, and utilise the same pool of co-regulator partners to regulate transcription. Many nuclear receptors and receptor co-regulators are expressed in breast cancers, and are linked to clinical
features including outcome. Importantly, many NRs are more frequently expressed in breast cancers than are ER
and PR, and their expression is often largely independent of ER and PR. Moreover, except for the ovarian
hormones, the ligands and metabolic signals that drive these receptors are not linked to menopause and are likely
to be circulating throughout life in all women.
Strikingly, it appears that most NRs have inhibitory actions in experimental systems and therefore if
appropriately targeted may prove to be fruitful treatments in breast cancer, but this possibility has received little
attention to date. The known structural and functional similarity between nuclear receptor family members, and
their use of the same or related co-regulator proteins, provides an opportunity to identify pathways/targets/critical
decision points of nuclear receptor activity in breast cancer that are common to those regulated by ER, and
distinct from those regulated by ER, yet this has not been pursued. In this project we will identify such targets,
and exploit this knowledge both through clinical application of existing NR modulators, and through
development of new agents to target NR pathways. The promise of this approach is the rational development of
agents that will have utility when resistance to ER-focussed treatments develop, or in combination with existing
treatments. Identification of NR targets distinct from ER holds the real promise of new treatments, for a group of
breast cancers currently with few treatment options.
Introduction
Breast development and function is influenced by steroid hormones
The human breast is a remarkably dynamic organ that displays dramatic changes in structure and function
associated with every phase of development, from initiation of breast development in embryonic life, to
regression of the organ after the menopause {Howard, 2000 #78}. The adult breast contains epithelial cells
embedded in a complex stroma that adopts different characteristics dependent on its location, and epithelial and
stromal constituents are embedded in fat.
The key milestones in breast development take place prenatally, by 28 weeks gestation, when the two populations
of epithelial cells in the breast first become discernible, and in puberty, when the most striking proliferative
expansion observed during a woman’s lifetime takes place {Howard, 2000 #78}. More minor proliferative
episodes are observed during the menstrual cycle, and in pregnancy. The systemic endocrine environment is
active during the major developmental milestones in the human breast {Howard, 2000 #78}, implicating
hormonal signalling in this process, and animal studies provide support for the absolute requirement for estrogen
(E) {Hewitt, 2005 #77} and progesterone (P), working in concert with pituitary and other circulating and local
factors, in mammary gland development and function {Conneely, 2003 #76}. Other circulating steroids,
particularly androgens and glucocorticoids, also contribute to normal breast development, but their roles have
received relatively little attention to date.
The initiating events in breast cancer are unknown
The emergence of breast cancer in later life is consistent with the accepted multistep model of carcinogenesis.
The nature of the initial genetic changes remains largely undefined, except in the case of the less than 10% of
breast cancers that are familial. Germline mutations in one of the BRCA susceptibility genes are associated with
markedly increased breast cancer risk {Antoniou, 2003 #75}. In the majority of breast cancers however,
environmental factors such as ionising radiation or exposure to carcinogens are likely initial triggers. The initial
genetic hits probably occur in puberty, when the most rapid expansion of the breast takes place, and invasive
cancer is most commonly diagnosed in women over 50. Consequently, cells with genetic changes, which are
likely to remain morphologically undetectable for a number of years, will be exposed for an extended period to
systemic influences from the endocrine, dietary and metabolic environment.
Breast cancers are comprised of multiple subgroups
The continuous morphogenesis of the breast underscores its capacity for undergoing remodelling under the
influence of a variety of factors, and this plasticity is also thought to contribute to its susceptibility to
carcinogenic transformation. Breast cancer is not a single disease, and subgroups with different biological
features, responses to treatment and outcome, have been described. Breast cancer subgroups are postulated to
reflect alternative pathways of epithelial differentiation during carcinogenesis {Abd El-Rehim, 2004 #74} and
there is now overwhelming evidence that the biological character of breast cancers is established early in disease
inception {Lacroix, 2004 #79}. Breast cancers of higher clinical grade, or larger size, or involving spread to
lymph nodes, have long been associated with poorer survival. More recently, gene expression profiling has
further delineated breast cancers into subgroups characterised by, for example, ER expression, or expression of
basal markers, or erbB2 expression {Sorlie, 2001 #80}. These subgroups have distinct clinical outcomes, and
their identification illustrates the power of high throughput approaches to uncover new features of breast cancer
biology.
Role of ER and PR in defining clinically important subgroups of breast cancer
ER is frequently expressed in breast cancer, with over 70% of newly diagnosed cancers being ER+
(Osborne1998). ER expression is the hallmark of a subgroup of breast cancers with distinctly better prognostic
features than ER- cancers. Around half of ER+ cancers also express the progesterone receptor (PR), and this
subgroup of cancers has the best prognosis of all, and overall the best response to endocrine treatments.
Notwithstanding these encouraging features of ER+ breast cancer, which have been exploited in the successful
use of estrogen-targeted treatments including tamoxifen and aromatase inhibitors, ER+ breast cancers frequently
recur after adjuvant treatment, and only a small proportion of these advanced cancers will respond to available
treatment options.
Breast cancers lacking ER have poorer prognosis and fewer treatment options
Although ER+ breast cancers are common, the cancers lacking ER exert a disproportionate impact on the patient
population. ER- breast cancers are commonly diagnosed in younger women, and, with the exception of the small
proportion with amplification of the tyrosine kinase gene erbB2, for whom Herceptin is a helpful modality,
treatment options are primarily cytotoxic chemotherapy, which is blunt and largely ineffective as a cure.
Recurrences are common in this subgroup of women.
Prognostic and predictive markers in breast cancer
The ER+ breast cancers that develop treatment resistance, ER- breast cancers, particularly in younger women,
and cancers that recur in all age groups, are currently underserved with existing treatment options, and this is due
in part to the lack of factors that can accurately predict treatment response in these subgroups. Currently, a
number of biological features of breast tumours are evaluated, including lymph node status, tumour size, clinical
grade, expression of ER and the receptor tyrosine kinase ErbB2, to provide prognostic information on tumour
biology and the likelihood of recurrence. In addition to these routinely used features, a very large number of
factors have been identified and validated that correlate with tumour biology and recurrence. These include a
large number of molecules identified by virtue of their roles in some aspect of cell physiology. However, while
prognostic factors have been helpful in identifying breast cancers with worrying risks of recurrence, their clinical
impact in identification of subgroups that benefit from treatment has been questionable.
In order to identify patients likely to benefit from specific treatments, factors that can predict treatment outcome,
independent of prognosis, are required, and the best candidates are factors that are targets of treatment, or can
modulate treatment targets. In breast cancer, the only robust predictive factors are ER, which predicts response to
ER-targeted treatments, and ErbB2, which predicts response to Herceptin. Use of these predictive markers in
selecting patient populations for treatment has made a major impact to date, highlighting the likelihood that
identification of additional treatment targets as predictive markers will make a major contribution to breast cancer
management.
ER and PR are members of the nuclear receptor superfamily
Although the link between ovarian function and breast cancer has focused attention on steroid receptors, it is less
frequently acknowledged that steroid receptors form part of a nuclear receptor superfamily of fundamental
importance in inter- and intra-cellular signalling. Nuclear receptors (NRs) are a family of transcription factors
with evolutionary origins in fungi, and highly conserved in all species from worms to human {Germain, 2003
#92}. They are distinguished from other transcriptional factors in having the capacity to bind to ligands including
steroids (estrogen, testosterone etc), thyroid hormones, dietary lipids, oxysterols, fatty and bile acids and vitamin-
related ligands, thereby translating physiological, metabolic and nutritional signals into gene regulation in a
cell/organ specific manner. Nuclear receptors regulate complex genetic programmes often consisting of thousands
of genes, with temporally and spatially regulated expression, that enable nuclear receptors to serve as master
regulators of virtually every aspect of life, from embryogenesis, through to reproduction, metabolism and cell
death {Germain, 2003 #92}. Nuclear receptors also serve as platforms for integration of signals from diverse
pathways, by virtue of the extensive cross-talk between nuclear receptor and other signalling cascades, so-called
because NRs act as targets of other signalling pathways, and in turn, regulate the activities of such pathways
{Germain, 2003 #92}.
NRs share a common structure of 5-6 domains with specific functions, and the high degree of conservation within
the ligand and DNA-binding domains forms the basis for the grouping and nomenclature of the 6 NR subgroups
identified to date {Nuclear Receptors Nomenclature, 1999 #87}.
NRs regulate transcription via sequence-specific DNA binding, and all NRs recognize derivatives of the same
core hexameric DNA motif, the NR response element. Response element recognition requires dimerisation of
NRs, and association with co-regulators is essential for signal transmission to the transcriptional machinery.
NRs in health and disease
NRs have fundamental roles in promoting and maintaining human health, and disruption of NR pathways is
implicated in diseases including diabetes, obesity, inflammation and cancer. These roles are underscored by the
therapeutic efficacy and range of medicinal compounds associated with dysfunctional hormone signalling {Barry,
2005 #69; Suzuki, 2004 #68}, and pharmaceuticals directed at nuclear receptor targets and the steroidogenic
pathways have utility in reproduction, inflammation, cancer, osteoporosis, diabetes, and cardiovascular disease.
In addition to ligands for the steroid receptors, PPARγ agonists are in extensive use for the treatment of diabetes,
retinoids are used in the treatment of acne and also in the treatment of promyelocytic leukaemia. Calcitriol is used
to treat osteomalacia and osteoporosis. Fibrates, which act through PPARα, are used in the treatment of lipid
disorders.
Steroid receptors with a role in breast cancer
AR: Although androgens are perceived as male sex steroid hormones, women produce androgens, from the ovaries and adrenal glands, at a level equivalent to two-thirds of that synthesized in men {Labrie, 2003 #21}. The extra-gonadal sources of androgen production highlight the fact that androgens circulate in women throughout life, even upon cessation of ovarian function at menopause. Androgen excess suppresses breast development (reviewed in {Labrie, 2003 #21}), whereas mice lacking a functional androgen receptor (AR) display defective mammary gland development {Yeh, 2003 #22}. The Adelaide investigators, and others, have demonstrated AR immunoreactivity in the epithelial cells of the breast, with little or no expression of AR in the myoepithelium or stroma {Janssen, 1994 #25; Moinfar, 2003 #23; Ruizeveld de Winter, 1991 #24; Zhuang, 2003 #26}. AR is at least as frequently expressed in breast cancers as ER, and its expression is correlated with ER and PR expression. There is emerging evidence that the androgen signaling pathway plays a critical protective role in breast cancer growth {Birrell, 1998 #18; Bry, 2000 #82; Langer, 1990 #20; Liao, 2002 #19}. AR positive breast cancers are associated with better prognosis and response to hormone therapy {Birrell, 1995 #27; Bryan, 1984 #37; Langer, 1990 #20}. In a preliminary study of breast cancer patients, the Adelaide investigators have shown an association between AR positivity and survival (p<0.01) with an odds ratio of 3.39, suggesting that loss of AR is a determinant of the aggressiveness of breast tumors. Androgens have a predominantly inhibitory effect on the growth of breast cancer cells, both in vitro and in vivo {Birrell, 1995 #27; Dauvois, 1991 #32; de Launoit, 1991 #30; Greeve, 2004 #31; Hackenberg, 1991 #28; Ortmann, 2002 #33}, potentially due to induction of apoptosis {Kandouz, 1999 #34; Lapointe, 1999 #35}, suggesting that androgen signaling pathways may be useful therapeutic targets and effective alternatives to current treatments. ERβ: The identification of a second ER gene in 1996, ERβ, stimulated a flurry of activity including exploration
of its role in breast biology and cancer {Harris, 2007 #81; Speirs, 2004 #11}. Expression of ERβ is generally
lower than that of ERα in breast cancer, but unlike ERα, ERβ is expressed in a variety of tissues beyond just the
epithelium. In normal tissue, its expression is widespread and although the literature is at times contradictory,
loss of ERβ expression has generally been associated with malignancy and a worse prognosis. The evidence that
ERβ is anti-proliferative is consistent with the concept that it may be a tumour suppressor gene; epigenetic
silencing has been observed in breast cancer. Variants of ERβ have been described which may have a role in
tamoxifen resistance but this remains to be defined. Although ERβ responds to estradiol with an efficacy similar
to that of ERα, it is significantly more sensitive to the so-called phytoestrogens. The role of ERβ may differ
between ERα positive and ERα negative breast cancers {Skliris, 2006 #10}, and its role in both the genesis and
treatment of breast cancer remains to be clearly defined.
Mineralocorticoid and glucocorticoid receptors: Mineralocorticoids and glucocorticoids, secreted by the adrenal,
play fundamentally important roles in normal physiology, and, significantly, are detected throughout life in the
systemic circulation. The mineralocorticoid (MR) and glucocorticoid (GR) receptors respond to the adrenal
cortical steroid hormone cortisol, the MR also responds to aldosterone primarily in epithelial tissues. Pre-receptor
selectivity of the MR for aldosterone in sodium transporting epithelial tissues is conferred by the enzyme 11β-
hydroxysteroid dehydrogenase type II (11βHSD2) which converts cortisol to its inactive metabolite, cortisone.
An extensive literature describes the fundamental role that cortisol plays in the growth, development and
maturation of the mammary epithelium. Both GR and MR are present in human breast tumours {Martin, 1981
#3} and 66% of breast tumour samples contain 11βHSD2 {Koyama, 2001 #5}. MR and 11βHSD2 are co-
localised in breast cancer {Sasano, 1997 #4}. Cortisol has been reported to be anti-proliferative in breast cancer
cell lines, with this response being modulated by the levels of 11βHSD2 expression {Koyama, 2001 #5; Lipka,
2004 #2}.
Other NRs expressed in breast cancer
NRs expressed in breast cancer, in addition to the steroid receptors outlined above, include the vitamin D receptor
(VDR) {Eisman, 1986 #88}, the thyroid hormone receptor (TR) {Silva, 2002 #14; Conde, 2006 #13}, the retinoid
(RAR) and retinoid-X (RXR) receptors, ER-related orphan receptors (ERRs), the farnesoid X bile acid
homeostasis receptor (FXR) {Swales, 2006 #85}, the steroid and xenobiotic receptor (SXR) and the pregnane X
receptor (PXR), involved in xenobiotic detoxification and multidrug resistance {Miki, 2006 #90}. Many of the
orphan NRs, so-called because they lack known ligands (including ERRs, COUP-TFII, LRH-1 etc) are
abundantly expressed in breast tissue and aberrantly expressed in breast cancer cells.
NR coregulators
The discovery of NR coregulators has dramatically changed our understanding of NR action. The first NR
coregulator, SRC-1 (Steroid Receptor Coactivator 1), was cloned 10 years ago {Onate, 1995 #45} and found to
be the first of a large family of transcription factors that do not bind DNA, but rather, bind directly or indirectly to
NRs to mediate their transcriptional efficiency. NR coregulators have been divided functionally into coactivators,
which enhance transcription, and corepressors, which repress transcription. Once the NR is bound to the DNA,
coregulators are recruited to perform all of the subsequent reactions needed to induce or repress expression of
genes {Lonard, 2006 #46}. Many of the coregulators contain enzymic capacity important for regulating specific
aspects of gene expression, including chromatin modification and remodeling, initiation of transcription,
elongation of RNA chains, RNA splicing, and finally, termination of the transcriptional response. Recent
evidence suggests that coregulators can control cellular reactions outside the nucleus such as mRNA translation,
mitochondrial function, and motility {Wu, 1999 #47; Yu, 2007 #48}.
Coregulators expressed in breast cancer
Coregulators, by definition, are present at rate-limiting levels in the nucleus so modification of expression and/or
activity can lead to profound alterations in NR signaling. Not surprisingly, therefore, they can influence human
disease, and there are a number of examples of disrupted co-regulator expression/function in androgen-
insensitivity syndrome and in prostate cancer {Adachi, 2000 #57; Zhou, 2003 #51}. In breast cancer, there is
increased expression of AIB-1/SRC-3 {Schiff, 2003 #50}, and evidence that differential expression and/or
activity of coactivators/corepressors in a given cell can modulate the agonist versus antagonist activity of
tamoxifen (Tam) on ER activity {Smith, 1997 #49}. AIB-1/SRC-3 appears to be a target of several kinases,
including p38, JNK, erbB2 and ERK1/2, and AIB-1/SRC-3 phosphorylation is critical for its in vivo oncogenic
effect in oncogenesis models. Thus AIB-1/SRC-3 responds to a variety of cellular signals by site-specific
phosphorylation, thereby integrating several key growth and homeostatic pathways.
There is substantial evidence implicating various other coregulators in breast growth and tumorigenesis. SRA
(Steroid receptor RNA regulator, an RNA coregulator) is upregulated in 90% of breast tumor samples {Lanz,
2003 #54}. The Perth node investigators have recently identified a novel corepressor, SLIRP, that is widely
expressed in human cancer, including breast cancer {Hatchell, 2006 #55}, and whose intracellular distribution,
predominantly mitochondrial, suggests a much broader role in the functional biology of breast cancer cell
metabolism. Data from experimental models shows that antisense to SRA reduces ER -dependent gene
expression in MCF-7 cells {Cavarretta, 2002 #56}.
Association of NRs expressed in breast cancer with prognostic factors and outcome
VDR expression in breast cancers is very common, and its expression is independent of ER, PR and GR and
nodal status {Eisman, 1986 #88}. VDR expression is associated with longer disease-free survival ( ). Vitamin D
is anti-proliferative in breast cancer cell lines in vitro {Swami, 2003 #8} with some evidence indicating an effect
of VDR on the ER {Swami, 2000 #7}, suggesting that vitamin D signaling pathways may be useful therapeutic
targets.
Thyroid hormone plays a fundamental role in development and in metabolism. There are two thyroid receptor
genes, which have frequently altered and/or absent expression in breast cancers {Silva, 2002 #14; Conde, 2006
#13}. LOH and/or biallelic inactivation by hypermethylation of the region of chromosome 3p that includes TRβ
{Li, 2002 #15} supports the view that TR is likely to be a tumour suppressor. TRα is linked to c-erbB2 on
chromosome 17q and several studies have found co-amplification in breast tumours of high metastatic potential
{Tavassoli, 1989 #17}. Co-amplification has also been associated with mutations {Futreal, 1994 #16}. Also
linked to the TR gene loci are the retinoic acid receptor (RAR)α and β genes {Yang, 2002 #12}.
SXR/PXR are expressed in breast cancers but not normal breast, and their expression is associated with tumour
grade, and lymph node status {Miki, 2006 #90}. If the tumours are ER+, then SXR/PXR expression is also
associated with increased levels of the proliferation marker Ki67. PXR expression is inversely associated with ER
expression {Dotzlaw, 1999 #89}.
ERR expression is significantly associated with an increased risk of recurrence and adverse clinical outcome. A
similar propensity was also found in the group of breast cancer patients who received tamoxifen therapy after
surgery, suggesting ERR expression is a potent prognostic factor in human breast carcinoma {Suzuki, 2004
#68}. ERR expression in human breast tumors correlates with ErbB2 expression {Ariazi, 2007 #91}.
The co-regulator SRA is similarly expressed in ER+ and ER- breast cancers, and an SRA splice variant is
aberrantly expressed in some breast cancer and associated with a worse prognosis {Murphy, 2000 #52}. Elevated
levels of the ER coactivator AIB1 are associated with decreased disease-free survival (Zhao2003).
Association of NRs and co-regulators with treatment resistance
The ER co-activator AIB1 is associated with tamoxifen resistance in patients: in Tam-treated patients, high AIB-
1/SRC-3 conferred a worse disease free survival, and when AIB-1/SRC-3 was coexpressed with HER2, women
had a very poor outcome (Osborne2003). Another co-regulator associated with impaired tamoxifen response is
BCAS3, which is a target of the ER coregulator MTA1. MTA1 overexpression is associated with decreased
disease-free survival, and BCAS3 elevation is associated with impaired tamoxifen response. SRA coactivates
Tam-antagonized ER, consistent with a role in Tam-resistance {Coleman, 2004 #53}
The orphan nuclear receptor chicken ovalbumin upstream promoter transcription factors (COUP-TFI and II) have
been implicated in tamoxifen resistance in experimental systems. A recent study has demonstrated COUP-TFII
expression is attenuated in tamoxifen resistant cell lines derived from tamoxifen sensitive MCF-7 human breast
cancer cells, suggesting that decreased COUP-TFII expression correlates with the acquisition of tamoxifen
resistance in human breast cancer cell lines and that COUP-TFII operates in the growth inhibitory cascade induce
by tamoxifen {Riggs, 2006 #72}.NRs contribute to breast cancer risk in populations
NRs contribute to breast cancer risk in populations
As predicted from the association of AR and better prognosis in breast cancers, a reduction in androgen signaling
is associated with increased breast cancer risk. Levels of the androgen regulated kallikreins in nipple aspirate
fluid are lower in women with breast cancer than in healthy, age-matched women {Sauter, 2004 #36}. A number
of epidemiological studies have suggested an inverse link between sunlight/vitamin D and the incidence of breast
cancer {Cui, 2006 #6}, and polymorphisms in the VDR are associated with breast cancer risk (Chen2005 – to
get). Polymorphisms in the peroxisome proliferator receptor PPAR are associated with increase risk of alcohol-
related breast cancer (Vogel2007 – to get). Increased exposure to the dietary phytoestrogens
(lignans/enterolignans) that modulate ERR activity is associated with decreased risk of ER+PR+ postmenopausal
breast cancer {Touillaud, 2007 #71}.
Local production of NR ligands - steroidogenic enzymes.
NRs respond to circulating cues in exerting endocrine effects. They also respond to cues produced within target
cells, or cells neighbouring target cells and these autocrine or paracrine effects are particularly relevant in breast,
where local production of steroids takes place. In postmenopausal women, where ovarian steroid biosynthesis has
ceased, the estrogen-dependent development of breast cancer is driven by the local production of estrogens within
the breast adipose by the aromatase gene, CYP19, which converts androgens to estrogen. Its expression is 3-4-
fold enhanced in breast cancers, by the expression and interaction of various transcription factors, including the
orphan NR liver receptor homologue-1 (LRH-1), peroxisome proliferator-activated receptor-gamma (PPAR )
coactivator-1alpha (PGC-1α) (a powerful coactivator of LRH-1), and cyclic-AMP response element binding
protein (CREB).
NRs serve as platforms integrating diverse pathways – cross-talk between NR pathways
NR pathways that influence ER signalling Progesterone interacts with ER signalling pathways in all reproductive tissues. In the uterus, estrogen promotes proliferation and augments PR levels, and progesterone in turn attenuates the estrogen signal via downregulation of ER, and inhibition of ER-mediated proliferation. In the human breast, the interaction between estrogen and progesterone is less clear – in the mouse mammary gland, both E and P are required for cell type specific and
development stage specific proliferation, and PR levels are not sensitive to estrous cycle fluctuations in serum
estrogen concentration. In experimental systems of breast cancer PR decreases ER activity by mechanisms
analogous to those described in the uterus.
Androgens oppose the growth-stimulatory effects of estrogen-mediated activation of ERα in primates in vivo
{Zhou, 2000 #41}. Androgens decrease expression of ERα and decrease ERα activity in breast cancer cells
{Poulin, 1989 #39; Panet-Raymond, 2000 #40}. Thus, the extent that androgens oppose estrogen signalling in
breast tissue may play a critical role in controlling cellular proliferation and maintaining tissue homeostasis in the
breast.
ERRs modulate estrogen dependent transcription but do not bind native estrogens. ERR , ERR and ERR
modulate estrogenic action by (dimeric and monomeric) binding to estrogen response elements, and novel
extended ERE half sites, respectively. A variety of synthetic compounds and many phytoestrogens that positively
modulate ERR activity have been described. Interestingly, organochlorine pesticides, diethylstilbestrol (the
synthetic estrogen) and the selective estrogen receptor modulators (4-hydroxytamoxifen and 4-
hydroxytoremifene) antagonize ERR activity. Moreover, several studies have reported CAR and SXR/PXR (the
xenobiotic NRs) modulate ER dependent gene regulation in cell and gene/target specific manner {Min, 2006
#70}. This orphan NR subgroup provides an opportunity to therapeutically modulate ER signalling at a different
step in the regulatory cascade {Ariazi, 2006 #84}.
Cross-talk between other NR pathways
Both the MR and the GR cross-talk with the PR in breast cancer cells {Leo, 2004 #1}, emphasising the
importance of more clearly defining their respective signalling pathways in the mammary epithelium.
RXR ligands have been repeatedly shown in experimental systems to augment the effects of PPAR agonists, and
combined activation of both pathways leads to a dramatic increase in apoptosis (Elstner2001). These findings
emphasise the feasibility of using combinations of NR ligands to increase inhibition of breast cancer cell growth.
ERR phosphorylation is modulated by epidermal growth factor (EGF) signalling. This suggested crosstalk
between the phosphoprotein, ERR , and EGF signalling can selectively regulate ERR target genes in breast
cancer cells {Barry, 2005 #69}. EGFR is also linked to expression RAR , which is lost in breast cancers and is a
putative tumour suppressor: downregulation of EGFR leads to reactivation of RAR in experimental systems.
Cross-talk between NR pathways, co-regulators and miRNAs
miRNAs – powerful modulators of gene expression with a rapidly emerging role in human cancer Approximately 98% of the transcriptional output in humans consists of non-protein-coding RNAs (ncRNAs). miRNAs are small single-stranded ncRNAs of ~22 nucleotides (nt) that comprise a family of >450 different transcripts, which contribute ~1-4% of all expressed human genes {Bartel, 2004 #58}. miRNAs negatively regulate their targets by augmenting mRNA decay or by decreasing translation {Bartel, 2004 #58; Esquela-Kerscher, 2006 #59}. Expression studies, including data from the Perth investigators, indicate that miRNAs can downregulate large numbers of target mRNAs (>100), consistent with a broad range of functional effects {Lim, 2005 #60; Webster, 2007 #83}. Moreover, these studies imply that miRNAs could control expression of ~ 1/3 of all human mRNAs. miRNAs and cancer – specifically breast cancer Several lines of evidence suggest that miRNAs play important roles in human cancer and more than half of miRNAs are located at sites in the human genome that are frequently amplified, deleted or rearranged in cancer. Some miRNAs are decreased in cancer {Calin, 2002 #61}; breast and prostate cancer miRNA expression signatures predict targets comprising genes enriched for tumor suppressors and oncogenes {Volinia, 2006 #62}; synthetic miRNA inhibitors inactivate miRNAs in tumors and can slow growth {Krutzfeldt, 2005 #63}; and conversely, overexpression of miRNAs that function as tumor suppressors, can abrogate growth {Scott, 2007 #64}. miR-21 acts as an oncogene in multiple cancers: it is highly overexpressed in breast cancer {Si, 2007 #66}. Interesting links are emerging between the NR signaling pathway, coregulators and miRNAs. For example, miR-17-5p regulates breast cancer cell growth by inhibiting expression of AIB-1/SRC-3, via translational repression {Hossain, 2006 #65}. MiR-17-5p was low in breast cancer cell lines, and downregulation of AIB-1 by miR-17-5p decreased ER-mediated, as well as ER-independent, gene expression and proliferation of breast cancer cells. This suggests that miR-17-5p has a role as a tumor suppressor in breast cancer cells, and illustrates remarkable convergence of the NR signaling pathways and miRNAs. • Current treatment regimens provide little advantage to some subgroups of breast cancer patients, particularly ER-PR- premenopausal women, and women who develop resistance to treatment and face the challenge of living with advanced disease. New approaches are required to reduce the impact of breast cancer in these subgroups. • NR, their coregulators, and critical regulators including miRNAs are important contributors to breast tumor growth, and to development of resistance to endocrine therapy. While these highly related transcription regulators interact at a number of levels and are likely to form functional networks in target cells, their combined expression and interaction in breast cancer is almost completely unstudied. • Targeting of NR pathways is now common in diabetes, inflammation, osteoporosis and cardiovascular disease, highlighting the fact that NR ligands are tolerated and effective in human physiology. However, NR ligands have largely been overlooked in clinical breast cancer. Project objective
This application will identify the NR networks active in breast cancers, to identify new predictive targets of
response to treatment in women with ER-PR- premenopausal breast cancer, or women with metastatic disease.
Because the components of NR networks are so closely interlinked, their simultaneous evaluation will provide
the power required to identify critical players in breast cancer biology. The knowledge will be translated within
the granting period, through the application of treatments with existing NR ligands in women who have failed on
endocrine or cytotoxic treatments. Beyond the scope of this funding period, new rational small molecule agents
will be developed and tested in Phase 1/2/3 trials in currently underserved breast cancer subgroups.
Hypotheses
• That the nuclear receptor family of transcriptional regulators and associated partner proteins, with fundamental roles in normal physiology, also mediate critical aspects of breast cancer biology • That there are critical points of convergence in the pathways regulated by the nuclear receptor system in breast cancers, and that these can only be identified using systems biology • That identifying and targeting critical points of convergence in nuclear receptor pathways in breast cancers will open new directions in breast cancer management, as follows: o Nuclear receptor targets in normal breast are likely to be critical to normal function and amenable o ER-independent nuclear receptor targets expressed in breast cancers are likely to be amenable to development of new small molecule agents to selectively target groups of breast cancers with currently poor prognosis and limited treatment options, such as ER-PR- premenopausal breast cancer. o Nuclear receptor pathways that cross-talk to ER signalling will provide new options to interrupt estrogen signalling, particularly when treatment failure on current endocrine therapies occurs. o Response rates to current endocrine and cytotoxic treatments, particularly in advanced disease with currently few effective options, are likely to be improved by rational combination of existing treatments with currently available NR agonists. Research Plan
A major strength of this application is that it brings together established research groups from around the country,
with currently non-overlapping, internationally recognised, programs of research in NR biology and
bioinformatics. Significant expertise in all aspects of the project is in place, the approaches, facilities, and models
are all established, and the CIs are motivated now to combine their expertise to bring new insights into the
management of breast cancer.
Resources
Breast cancer tissues and cohorts
Fresh breast cancers will be obtained upon application from the Breast Cancer Tissue Bank, and from existing
sources in the Adelaide, Perth and Sydney nodes. Clinical annotation on these specimens includes detailed
pathology information, epidemiological features, treatments received and follow-up data. Archival cohorts
include TMAs of primary cancers with known clinical outcome from the Perth node, and metastatic breast
cancers, treated with either endocrine or cytotoxic agents, and known clinical outcome, from the Westmead node.
Models of human breast
A major limitation in the study of normal human breast to date has been a lack of suitable experimental models
that can adequately recapitulate the normal structure and hormone responsiveness of the tissue.
Explant model. The Adelaide investigators have developed a new method of culturing normal human breast
tissue that retains the structure of the tissue and its hormone responsiveness. Normal tissue obtained from surgery
is finely diced and cultured as whole tissue pieces on gelatine sponges, and this maintains excellent histological
structure and steroid receptor expression in fixed paraffin-embedded breast tissues. Tissues are transported on ice
from surgery to the laboratory in RPMI medium, where they are dissected into small pieces of approximately
3mm x 3mm. The tissue are cultured on small pieces of pre-hydrated gelatine sponge, incubated at 37°C for 48-
96 hours then cryopreserved or fixed in paraformaldehyde and paraffin-embedded.
3D model. The Sydney investigators have developed a model in which primary normal breast tissue undergoes a
morphogenic program akin to that observed in vivo, with formation of polarised structures after culture in 3D.
Progenitor cells are retained in this model, and morphogenesis arises from progenitor expansion. ER and PR are
co-expressed in the 3D model. Normal breast epithelial cells isolated by collagenase digestion are mixed with
MCDB medium, then added to GFR Matrigel, pipetted into chamber slides, overlaid with MCDB medium and
incubated for 10-12 days at 37C with a medium change every 2-3 days then cryopreserved or fixed in
paraformaldehyde and paraffin-embedded.
Functional assays for NRs
Each of the geographic nodes includes laboratories, led by one of the CIs, with a strong basic focus in specific,
but distinct, aspects of NR biology. All the molecular and cell biological assays and cell lines required for the
discovery of targets in Aim 3 and the functional studies in Aim 4 are in place.
Facilities
Low density arrays: IMB Facility, ABI/microfluidic cards on the ABI7900 high throughput real time instrument. Gene expression and methylation profiling: WAIMR Facility, using the Illumina system. Animal models: The European Conditional Mouse Mutagenesis Program (EUCOMM) is part of the International Knockout Consortium, which will systematically target and mutate all 20,000 mouse genes in embryonic stem cells. By 2010 the majority of these genes will have been mutated in ES cells, and a significant number of mice will be available. CI Rosenthal, who leads the Monterotondo EUCOMM centre, will be developing selected Cre
mice as part of the Consortium, will be housing frozen ES cells from the collection at the Australian Regenerative
Institute in Victoria.
Antibodies: The National Monoclonal Antibody Technology Facility, to be based in Victoria, has particular
expertise in development of antibodies to phosphoproteins.
Informatics: An integrated approach to information management is critical to this project, as indeed to any project
involving high-throughput data generation at multiple sites, large complex data sets, computational inference and
collaborative analysis. There will be four components of project bioinformatics: (i) Management of data
generated in project, (ii) Integration of project data with other information sources; extended ontologies, (iii) Data
analysis including algorithms & software for pathways & network, (iv) Support of collaboration in data
generation, access and analysis.
We will deliver the required experience, breadth of skills and data infrastructure through a dedicated
bioinformatics node. Day-to-day operations will be embedded in a specialised facility, Queensland Facility for
Advanced Bioinformatics (QFAB), which will implement and manage technologies to support real-time
collaboration in data analysis and knowledge discovery across all project teams.
Technology development
Bioinformatics moving rapidly; many new analysis tools, datasets etc. will appear during the lifetime of this
project, particularly in literature mining, high-throughput omic technologies including DNA resequencing, data
integration, automated inferencing and network analysis. Do we need to identify a technology awareness /
development / adoption strategy?
Project management and Advisory Committees
Project management
The CIs will form the management team, and will appoint a project manager to implement project objectives.
Collaboration between the nodes and the central database will be determined, but could be via a Twiki™ or
similar [Mark]. Communication will also be facilitated by a twice-annual face-to-face management meeting, and
an annual face-to-face meeting of all project participants.
Clinical Advisory Committee
This will be chaired by Rick Kefford, and will include clinicians (named AIs) from the Sydney, Melbourne,
Brisbane, Adelaide and Perth nodes. Its roles will be to advise in the selection of clinical cohorts for Aim1, to
develop and conduct the Phase 2 trials in Aim 6 and finally to participate in ongoing project evaluation including
annual project meetings.
International Advisory Group
This will be chaired by John Funder, and will include leaders in NR biology (Bert O’Malley, Baylor College), in
high throughput analysis of breast cancers (Edison Liu, Singapore Genomics Institute), in clinical breast cancer
(Mitch Dowsett, David Huntsman), in epidemiology of hormones and breast cancer (Malcolm Pike) and in xx
[George] (Frank Gannon). Its role will be to advise on strategic directions, synergism and funding possibilities,
provision of independent oversight, and participation in an annual face-to-face meeting.
Aim 1: To identify NRs, coregulators, and associated critical regulators in breast cancers.
Overall approach:
1) We will isolate RNA from matched human breast cancer cohorts, and normal human breast tissues, and
quantitate levels of expression of NRs, coregulators, miRNAs, histone deacetylases and steroidogenic enzymes.
We will validate identified transcripts at the protein level. We will also validate the targets identified in the initial
breast cancer cohorts in large cohorts of breast cancers, by immunostaining of TMAs.
2) In parallel, we will use bioinformatics to search existing breast cancer gene expression datasets for expression
of NRs, coregulators, miRNAs, histone deacetylases and steroidogenic enzymes. These existing datasets are
unlikely to contain all the NR-related genes of interest, as existing high throughput array technologies do not
have the sensitivity to reliably detect low abundance genes such as NR family members and their coregulators.
However, if expression data are found, these will provide useful information in clinical cohorts other than those
selected for this study.
3) The data from our discovery program and the existing datasets will be analysed bioinformatically, to identify
combinations of gene, pathway and regulatory networks likely to be active in breast cancers.
Detailed methods:
Clinical cohorts: selection and matching: Fresh frozen samples (n=20 per group). Group 1, ER+PR+
postmenopausal breast cancers; Group 2, ER-PR- postmenopausal breast cancers. Fresh breast cancer tissue with
clinical annotation will be obtained by application to the Breast Cancer Tissue Bank. Group 3, normal breast
tissue from postmenopausal women. Normal breast tissue will be obtained from current sources in NSW and SA.
Breast cancer groups will be matched for histological grade, and patient age. Clinical cohorts will also be
matched by gene expression profiling using the Illumina system. For each matched cohort, expression profiles
will be analysed using unsupervised heuristics (hierarchical clustering, k-means, SOM, SOTA) to identify the
maximally coherent subset of size N=10 within each. (Using the Illumina Bead Studio data analysis package
and/or other packages as required.)
Measurement of targets in clinical cohorts: Low density arrays: RNA will be isolated from tissues (n=10 per
group) using Qiagen RNEasy kits, cDNA will be synthesised and loaded onto TaqMan low density arrays
(LDAs) Nuclear Receptor Panel containing primer probe sets for 48 nuclear receptors. Subsequent custom arrays
to measure known isoforms of identified NRs will be developed. Custom TaqMan LDAs with 384 primer sets for
NR isoforms, co-regulators, histone deacetylase inhibitors and steroidogenic enzymes will be designed using the
TaqMan library of 47,000 validated targets, and selection will be guided in part by selection of likely partners for
the NRs identified on the Nuclear Receptor Panel. Real time PCR data will be quantified using relevant software.
Analysis of low-density array data: Primary analysis of expression data (fold statistics, gene identification) will
be carried out using GeneSpring GX (Agilent) and the results uploaded via secure protocol to QFAB for detailed
annotation as described under Aim 2. Highly capable tiered backup is in place at QFAB including offsite
mirrored storage and tape archive. QFAB will be an early adopter of developing Australian standards of
authentication, authorisation and data grid technologies.
Validation of gene expression data: Protein: antibodies against the genes identified by low density arrays will
be obtained or prepared and immobilised on antibody arrays, which will be hybridised with protein extracts from
the same clinical material from which RNA was isolated. Validation in clinical cohorts: this will be done in two
ways. Firstly, the gene expression profiles on the larger series, used to match the clinical cohorts, will be re-
examined for expression of targets identified by LDAs. Secondly, antibodies against identified targets will be
used to immunostain TMAs containing hundreds of breast cancers, including ER+PR+ and ER-PR- cancers.
Targets identified in the small, carefully selected discovery cohorts above, and that are generalisable by being
found also in these larger cohorts, will be investigated further.
Bioinformatic analysis of existing breast cancer datasets: Expression microarray data will be obtained via
online archives maintained at Pittsburgh [1], MIT [2], Stanford [3], Duke [4], North Carolina [5], Institut Curie
[6] and elsewhere. Where necessary, UniProt IDs will be assigned. Known and potential NR family members, co-
regulators, histone deacetylases and components of steroidogenic pathways will be identified directly (from
available metadata) or via pathway databases (KEGG, Reactome) and ontology terms. Potential miRNA
regulators will be identified from literature [e.g. 7,8] and a proprietary target database (ARC Centre of Excellence
in Bioinformatics, unpublished). Further pathways and networks will be inferred using Ingenuity Pathway
Analysis [9] and results ported to Cytoscape for visualisation outside the ambit of the IPA license. It is highly
likely that existing GO, FMA/AO and SOFG/SAEL ontologies will have to be extended to cover endocrine-
specific resources; for this we will use the OWL language, BioPAX Level 2 format, and the Protégé tool with
appropriate plug-ins.
Outcome: Identification of NRs, partner and related proteins and steroidogenic enzymes that are expressed in
ER+PR+, ER-PR- breast cancer, and normal breast. This will provide a distinction between NRs expressed in
cancer and normal tissue. It will reveal NRs that are independent of ER expression, candidates for new
approaches to treatment. It will also reveal NRs that are linked with ER expression and may be candidates for
cross-talk with ER in breast cancers and therefore may provide new routes to target the ER pathway.
Aim 2: To refine the identified NR family members to select those most likely to be critical in breast cancer
Overall approach:
In Aim 1 we will have comprehensively identified the NR family members, associated partner proteins and
related enzymes that are present in breast cancer and normal breast, and where possible will have assigned them
to intracellular pathways and sub-networks. In Aim 2 we will refine this list, based on a number of criteria:
Patterns of expression: genes/pathways/networks that are more highly expressed in breast cancer compared with
normal breast will be initially selected. Those that are differentially expressed between ER+PR+ and ER-PR-
cancers will also be selected.
Known involvement in breast cancer: genes that have been implicated in breast cancer biology in experimental
studies, using cell lines or animal models, may be implicated in human disease.
Hypervariability within the population: NR and associated partners that are highly polymorphic are likely to be
good candidates for contributing to inter-individual differences in breast cancer treatment response and outcome.
Detailed methods:
Determination of gene expression pattern: The low-density array data from Aim 1, together with results from analysis of existing datasets, will yield a list of differentially expressed NRs and related molecules. Each molecule in this list will be extensively annotated for intron/exon structure, splice variants, promoters, hormone-responsive elements, miRNA binding sites, post-translational modification motifs, trafficking/localisation motifs, mouse and rat orthologs, pathway membership, protein-protein interactions, gene product ontology, SNPs, haplotype data, expression in different cohorts and tissues, and known disease involvement. Annotation will rely on federated data within SRS [10] and will be automated to ensure that new information is integrated. This central data resource will be available to all project groups via a secure, platform-independent portal. Known involvement in breast cancer: Comparison of gene, transcript and protein IDs will identify cross-matching between genes identified in Aim 1 and those implicated in experimental models of breast cancer. Hypervariability: polymorphisms in the gene set identified in Aim 1 will be found by HapMap searching. Identification of NR pathways with highest probability of involvement in clinical breast cancer: The pathways and networks identified above (Ingenuity) will be vertex-annotated to show known and potential correlation with clinical disease, e.g. involvement in experimental models of cancer, differential expression, and presence of hormone-related promoters, miRNA target sites, haplotypes and SNPs. It may be possible to infer extensions to these pathways and networks using predictive software tools. This annotation will capture specialist knowledge across the project. On this basis pathways will be sorted into those potentially involved in NR signalling in clinical disease, and those less likely to be involved. Each component of the former will be provisionally classified as either a driver or a bystander based on whether or not it constitutes a hub (highly connected vertex) in the pathway.
Outcome: The dataset derived in Aim 1 will be refined, to eliminate candidates that are unlikely to represent
breast cancer targets amenable to intervention, for instance those expressed predominantly in normal breast. The
refined target list will also begin to identify drivers, rather than bystanders, of the effects of NRs in the breast, for
instance those that are in critical locations within pathways containing a number of candidates. The ultimate
outcome of this Aim will be the description of the NR pathways that may be active in breast cancer.
Aim 3: To identify targets of NR pathways in the breast
Overall approach: NR pathways selected in Aim 2 will be activated/inhibited using agonists/antagonists in
culture systems, and target genes/pathways will be identified by high throughput gene expression profiling, by
phosphoarrays, methylation arrays, analysis of miRNAs. For NR family members that do not require ligands for
activation, over-expression and knock out strategies, either using siRNA or dominant negative mutants, combined
with the analysis approaches above, will be used to identify likely targets. Various model systems will be used,
including normal breast in 3D and explant culture, breast cancer cell lines, primary cultures of breast cancers, and
mouse models. Analysis of the data obtained will be focussed on identifying NR targets that are critical nodes in
NR signalling networks in the breast.
Detailed methods:
Models: Human breast and cell lines: normal breast models, and ER+PR+ and ER-PR- cell lines in routine use in
CIs’ labs, will be used.
Mammary specific Cre recombinase mice, and floxed NR alleles: We will obtain mammary specific deletion of
NR/partner genes either through classical genetic means, or by topical delivery of Cre recombinase via intraduct
injection [Nadia please clarify – currently feasible?]. Ortholog mapping between human and mouse will build on
gene pairs identified in ENSEMBL and the UCSC Genome Browser (both of which are mirrored within QFAB,
the latter uniquely within Australia) and will include a unique high-resolution (~20 amino acids) map uniquely
available through QFAB.
Treatments: Ligand treatment: Model systems will be exposed to NR agonists, guided by the NR pathways
shown to be implicated in Aims above, but could include: steroid hormones, vitamin D, thyroid hormone,
retinoids, bile acids, adipokines, xenobiotics. Treatments will include 1, 6, 24h exposure times. Model systems
will also be exposed to relevant NR antagonists and inhibitors. For NRs without known ligands, overexpression
and/or dominant negative mutants will be transfected into cell lines.
Analyses: The effect of switching on NR pathways will be determined by identifying gene targets, their
methylation status, their posttranslational modification by phosphorylation, which is a critical aspect of NR
signalling, and finally miRNAs that may be targeting NR pathways. High throughput gene expression arrays:
using llumina Human-6 V2 with bioinformatic analysis as described above. Phosphoarrays: Phosphorylation
status of NR targets will be determined using antibodies to phospho forms of identified proteins. miRNAs: NR
effects on miRNAs will be determined by real time RT-PCR of total RNA using TaqMan miRNA Human
Assays. Methylation arrays: using the Illumina GoldenGate assay technology.
Data integration: For a given NR pathway, integrating the data from the RNA, protein, phosphorylation and
microRNA readouts, will be done initially, then combined bioinformatic analysis will be used to identify critical
nodes in NR networks. Identification of critical nodes in NR networks in the breast: The inferred molecular
interaction networks will be extended based on data obtained from these experiments and platforms. Vertices that
are both highly connected (i.e. drivers as per Aim 2) and appear in more than one sub-network will be identified
as critical nodes in NR signalling and will be subjected to more-intensive database analysis and literature review.
Where sub-networks are proximal but do not overlap, secure inferencing lies beyond the range of current
technology; however, novel semantic inferencing technologies (e.g. using RDF and SPARQL) are under active
development (including within the ARC Centre of Excellence in Bioinformatics) and could become available
later in the project.
Outcome: The pathways identified in previous Aims will be switched on or off, to identify the downstream
cellular targets of NRs in the breast. Systems biology will be used to develop a detailed picture of the networks
regulated by NRs in the breast, and therefore to uncover points of convergence in NR pathways likely to be
critically involved in breast cancer.
Aim 4: To validate the role of critical NR pathway components in experimental models of breast cancer
Overall approach: The outcome of previous aims will have delivered candidate, highly selected, small sub-
groups of NRs, their effectors and downstream targets. The role of these critical NR pathway components in
breast cancer biology will be demonstrated using mouse models and cell lines models. NR pathway components
will be selectively removed in the mammary glands of mice, and effect on susceptibility to carcinogenic
transformation determined. NR pathway components will be removed in primary models and cell lines described
above, using siRNA or dominant negative constructs and effects on cancer-associated phenotypes including cell
proliferation, invasion and motility determined. There will be particular emphasis on identifying NR pathways
that cross-talk to ER signalling pathways, as a means of selecting NRs that may be active in breast cancers that
have developed resistance to tamoxifen or aromatase inhibitors.
Detailed methods:
Models: Mammary carcinogenesis models: the impact of deleting NRs on mammary cancer susceptibility will be
determined using mice with an established increased susceptibility and looking at whether that is modified by NR
targeting. Models could include MMTV-driven oncogene mice or p53 null mice. Cell line models: NR targets
will be removed using siRNA or dominant negative approaches, dependent on the target, in breast cancer cell
lines and in normal breast models, using various approaches including retroviral transduction, to introduce
targeting vectors into cells.
Analyses: Gene expression profiles: will be compared between wild-type and null mice, and wild-type and null
cell lines as outlined above. Functional assays: cell proliferation, invasion and motility will be measured using
standard cell biological assays in normal breast and breast cancer cell lines.
Outcome: NR pathways with a demonstrated role in altering the biology of breast cancer will be identified. The
mouse studies will reveal NR pathways that are required for mammary carcinogenesis. The cell line models will
identify which NR pathways play a role in regulating breast cancer biology.
Aim 5: To identify NR pathway components with predictive power in breast cancer
Overall approach: Critical NR pathway components will be measured in clinical cohorts currently underserved
with treatment options, to identify NR components, or combinations of NR components, that are particularly
applicable to these subgroups and therefore may be suitable targets for novel treatment approaches. Clinical
cohorts initially examined will be ER-PR- premenopausal breast cancers, the subgroup in younger women with
least favourable outcomes, and metastatic breast cancers. We will also examine the status of selected components
in more biologically diverse breast cohorts, including premalignant lesions such as ADH and PDWA, DCIS (low
and high grade), grade 1, grade 2 and grade 3 invasive breast cancers with known node status, to determine
whether particular combinations of NR components also link to other breast cancer subgroups. The association
with existing pathophysiological features of breast cancers, including expression of known prognostic and
predictive factors, will be determined.
Detailed methods:
Cohorts: Initially, patients who have metastasised on endocrine or cytotoxic treatment. Other subgroups to
follow.
Analysis: Prior aims will identify the exact analytes for this aim, for instance whether transcript or protein. If
transcript, RNA will be isolated from FFPE archival specimens and expression of candidate genes performed
using the Illumina DASL system and custom probe panels. If protein, immunostaining will be done of TMAs
prepared from cohort FFPE material.
Data integration: Bioinformatics will be used for combined analysis of clinical follow-up data and presence of
NR pathway components, to derive clinically important subgroups of NR pathway components for use as
treatment targets in specific breast cancer subgroups. [Details to follow. Main point will be that using
bioinformatic techniques, the project will ask whether any components of the networks heretofore identified are
critical predictive factors for the clinical prognosis of breast cancer, and if so, why. This should direct us to
pathways that are especially promising targets for treatment.]
Outcome: This aim will determine NR pathway components, singly or in combination, that are expressed in
subgroups of breast cancers with currently few treatment options. Additionally, NR targets with a role in breast
cancer biology demonstrated in model systems will be identified in a wide range of clinically and biologically
diverse breast lesions. Taken together, this aim will identify potential targets for clinical testing in underserved
breast cancer subgroups, and will identify targets in other subgroups to which novel agents should be developed.
Aim 6: To develop new breast cancer treatments by targeting NR pathways
Approach: The current widespread clinical use of NR modulators in other diseases offers the possibility of
immediate translation of the findings of this study into new treatments for breast cancer, in particular for ER-PR-
premenopausal breast cancers, and advanced disease, where there are few effective modalities. NR pathway
components identified in Aim 5 as being expressed in these subgroups, and with existing clinically approved
modulators, will be targeted in Phase 2 trials, alone and in combination with existing treatments such as taxanes.
Beyond the scope of the current granting period, new small molecule modulators of promising NR targets will be
developed and tested for efficacy via Phase 1/2/3 trials.
Detailed approach:
We will initiate Phase 2 combination trials, for example of taxane/NR modulators, in metastatic patients who
have failed on taxanes. Patients will be recruited from Westmead, Adelaide and Perth nodes.
Outcomes: Within this funding period, it should be feasible to begin combination treatment in suitable advanced
disease cohorts with NR agonists and existing therapies, and effects on outcome determined. Beyond the scope of
this application, but feasible within 10 years, forming partnerships with appropriate drug companies for the
development of new NR modulators will be initiated.
Milestones and timelines, including a Gantt chart 2010 2011
Clinical cohorts: selection and matching Measurement of targets in clinical cohorts Bioinformatic analysis of existing breast cancer datasets Identification of NR pathways with highest probability of involvement in clinical breast cancer

Source: http://eresearch.jcu.edu.au/staff/deployments/plone-1/qfab/proposals-and-designs/Cancer%20Research%20Project%20Description.pdf