Transcriptional Repressor Tbx3 Is Required for the Hormone-Sensing Cell Lineage in Mammary Epithelium
et al. (2014) Transcriptional Repressor Tbx3 Is Required for the Hormone-Sensing Cell Lineage
in Mammary Epithelium. PLoS ONE 9(10): e110191. doi:10.1371/journal.pone.0110191
Transcriptional Repressor Tbx3 Is Required for the Hormone-Sensing Cell Lineage in Mammary Epithelium
Kamini Kunasegaran 0
Victor Ho 0
Ted H-. T. Chang 0
Duvini De Silva 0
Martijn L. Bakker 0
Vincent M. Christoffels 0
Alexandra M. Pietersen 0
Bin He, Baylor College of Medicine, United States of America
0 1 Department of Cellular and Molecular Research, National Cancer Centre Singapore , Singapore, Singapore , 2 Program in Cancer & Stem Cell Biology, Duke-NUS Graduate Medical School Singapore , Singapore, Singapore , 3 Center for Heart Failure Research, Academic Medical Centre , Amsterdam , The Netherlands , 4 Department of Physiology, National University of Singapore , Singapore, Singapore
The transcriptional repressor Tbx3 is involved in lineage specification in several tissues during embryonic development. Germ-line mutations in the Tbx3 gene give rise to Ulnar-Mammary Syndrome (comprising reduced breast development) and Tbx3 is required for mammary epithelial cell identity in the embryo. Notably Tbx3 has been implicated in breast cancer, which develops in adult mammary epithelium, but the role of Tbx3 in distinct cell types of the adult mammary gland has not yet been characterized. Using a fluorescent reporter knock-in mouse, we show that in adult virgin mice Tbx3 is highly expressed in luminal cells that express hormone receptors, and not in luminal cells of the alveolar lineage (cells primed for milk production). Flow cytometry identified Tbx3 expression already in progenitor cells of the hormone-sensing lineage and co-immunofluorescence confirmed a strict correlation between estrogen receptor (ER) and Tbx3 expression in situ. Using in vivo reconstitution assays we demonstrate that Tbx3 is functionally relevant for this lineage because knockdown of Tbx3 in primary mammary epithelial cells prevented the formation of ER+ cells, but not luminal ER- or basal cells. Interestingly, genes that are repressed by Tbx3 in other cell types, such as E-cadherin, are not repressed in hormone-sensing cells, highlighting that transcriptional targets of Tbx3 are cell type specific. In summary, we provide the first analysis of Tbx3 expression in the adult mammary gland at a single cell level and show that Tbx3 is important for the generation of hormone-sensing cells.
Data Availability: The authors confirm that all data underlying the findings are fully available without restriction. All microarray data are available from GEO as
Funding: This work was funded by the Agency for Science, Technology and Research Singapore (www.a-star.edu.sg); grant number SSCC10/022. The funders
had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing Interests: The authors have declared that no competing interests exist.
Tbx3 is a transcriptional repressor with an important role in
embryonic development of the mammary gland and a high
expression in certain breast cancers, but its role in the different cell
types of adult mammary epithelium has yet to be explored .
Mammary gland development starts in the embryo, but the largest
part occurs postnatally. During murine embryogenesis, an
ectodermal mammary placode is induced which develops into a
rudimentary epithelial tree . During puberty, under the
influence of steroid hormones, the epithelial ducts start to elongate
and bifurcate to fill the mammary fat pad . In the adult,
morphogenesis of the mammary gland continues as it is subject to
further branching and the development of lobular structures with
alveoli (milk-producing units) during pregnancy, culminating in
lactation, followed by regression and remodelling to a virgin-like
state after weaning. At a smaller scale, there is even some
alveologenesis and regression under the influence of hormonal
fluctuations during the estrus cycle .
Milk ducts in the adult virgin are bi-layered with a luminal layer
that consists of hormone-sensing cells and cells primed for milk
production (alveolar progenitor cells) and an outer basal layer that
contains mostly contractile myoepithelial cells, but also rare
mammary epithelial stem cells . Both these multipotent stem
cells as well as lineage-restricted populations contribute to
epithelial renewal and alveologenesis . In transplantation
assays, a progenitor that gives rise to all cells types of an alveolus
can be detected , but recent data by several groups [7,9,10]
highlights that in intact mammary glands alveoli are commonly
formed by collaborative outgrowth of cells from at least 3 distinct
lineages. This includes cells from the basal lineage, the luminal
estrogen receptor-negative (ER-) alveolar lineage and the luminal
ER+ hormone-sensing lineage . The latter was unexpected,
since hormone-sensing cells have been considered mature, or
terminally differentiated cells. However several reports have
shown that hormone-sensing cells actively proliferate in vivo in
early pregnancy [11,12]. In addition, ER+ progenitor cells have
recently been identified by cell surface markers and in vitro colony
forming potential [13,14], indicating that it is indeed a separate
lineage. The regulation of the hormone-sensing lineage is
particularly interesting because the majority of breast cancers
express the estrogen receptor [15,16].
Here, we analyzed the role of Tbx3 in the lineage hierarchy of
the adult mammary gland. Tbx3 is one of the earliest markers of
mammary epithelial cells in embryonic development, and in the
absence of Tbx3 embryonic mammary placodes fail to form .
In Tbx3-heterozygote mice, reduced expression of Tbx3 is
sufficient to allow normal mammary gland development ,
although a later study showed that in thoracic mammary glands
epithelial trees occasionally did not form and fewer branches were
observed in the adult glands . In humans, hypomorphic
germline mutations in the Tbx3 gene are the cause of
UlnarMammary Syndrome  in which reduced activity of Tbx3
results in reduced breast development, in addition to other
developmental defects . Thus, even though there appear to be
differential quantitative requirements for Tbx3, Tbx3 plays an
important role in early mammary gland development across
species. Tbx3 is also involved in the embryonic development of
numerous other tissues, including limbs, heart and liver [21,22].
Tbx3 is likely to play a role in adult tissues as well, because
Tbx3 has been implicated as an oncogene [23,24] and was found
overexpressed in cell lines from several cancer types, including
melanoma and hepatoma [24,25]. In breast cancer, high nuclear
and cytoplasmic expression of Tbx3 was found in a subset of cells
in primary tumors  and high transcriptional Tbx3 levels in
human breast tumors correlated strongly with ER expression .
Here, we investigated the expression of Tbx3 in the different
mammary epithelial cell types in the adult mammary gland.
Materials and Methods
Mice used in this study were maintained under protocols which
were approved by the legal authority of the Singhealth
Institutional Animal Care and Use Committee, Singapore. All
procedures were in accordance with its guidelines. Non-terminal
procedures were performed under anesthesia, and all efforts were
made to minimize suffering of the animals. Anesthesia
(Hypnorm:Midazolam:Water = 1:1:2) was administered at 7 ml/kg
and analgesia (Meloxicam) was administered at 0.2 mg/kg.
Further analgesia (Meloxicam) was added to the drinking water
for 23 days after surgery (,0.2 mg/kg). Animals were euthanized
by carbon dioxide inhalation followed by cervical dislocation. All
efforts were taken to prevent animal suffering.
Tbx3tm1(Venus)Vmc (synonym: Tbx3Venus) mice were generated
by inserting the Venus coding and transcription termination/pA
sequences into the start codon of Tbx3. This places Venus under
control of the endogenous Tbx3 locus and prevents expression of
Tbx3 itself. A detailed description of the generation of these
knock-in mice is in preparation and will be described elsewhere
(MLB and VMC). Heterozygous Tbx3+/Venus mice were
maintained on an FVB background. Mice were euthanized by carbon
dioxide inhalation and immediately dissected for thoracic (MG3),
abdominal (MG4 & 5) mammary glands. Animal care and
protocols were in accordance with national and institutional
Cell labeling, flow cytometric analysis &
fluorescence-activated cell sorting (FACS)
Abdominal glands were pooled and processed for single
mammary epithelial cell (MEC) isolation for FACS analysis as
previously described . Fluorochrome-conjugated antibodies
were titrated on primary mammary epithelial cells to ensure
maximal positive signal:background fluorescence ratio.
Antimouse &/or anti-rat compensation beads (BD 552843 and
552845, respectively) were used for single stain antibody controls.
Compensation controls also included two cellular samples:
unstained cells and cells stained with DAPI (Sigma D8417,
USA). Cells were incubated with antibodies on ice for 45 minutes
with agitation every 15 minutes. Samples were then washed with
twice the sample volume and resuspended in L15 (Gibco-Life
Technologies, USA) with 6% FCS (Hyclone, USA) and 200 ng/
mL of DAPI, except non-DAPI compensation controls. All
multiple-labelled samples were gated on FSC-A vs. SSC-A and
doublet discrimination (FSC-H vs. FSC-W & SSC-H vs. SSC-W)
and DAPI negativity. Samples contained anti-CD45 to exclude
lymphocytes from analysis. Cells were analyzed and sorted on a
BD FACS-Aria II containing 355 nm UV, 488 nm blue, 561 nm
yellow-green and 633 nm red lasers. Specific antibodies used and
gating strategy are detailed in File S1.
Synthesis of cDNA & qPCR analysis
For analysis of transcript levels by qPCR in FACS sorted
populations, cells were sorted directly into lysis buffer (10 IU
RNase inhibitor (Invitrogen, USA), 2 mM DTT, 0.15%
Tween20 (Biorad, USA) in 12 mL of nuclease-free water) in PCR tubes
using a Direct Reverse Transcription method as described .
Five-hundred cells were sorted into each tube and Reverse
transcription (RT) was performed using Superscript VILO
(Invitrogen, USA) as per manufacturers protocol. Primers were
designed that span introns to exclude the detection of genomic
DNA and selected for optimum melt curve and amplification
profiles (for primer sequences, see File S2). qPCR was performed
using Sso Fast Evagreen supermix reagent (Biorad #172500) as
per manufacturers protocol.
For microarray analysis, MECs from 3 Tbx3+/Venus adult virgin
mice were pooled and 200,000 luminal VenusHigh and luminal
VenusLow cells were sorted into L15 medium with 6% FCS. After
centrifugation, cell pellets were lysed in Trizol and total RNA was
isolated according to manufacturers protocol. Biotinylated cRNA
was prepared from 250 ng of total RNA with the GeneChip 39
IVT Express Kit according to the manufacturers protocol
(Affymetrix 2008). Following fragmentation, 12.5 ug of cRNA
was hybridized on a GeneChip Mouse Genome 430 2.0 Array for
16 hours at 45 C. The GeneChip was washed and stained in the
GeneChip Fluidics Station 450 (Affymetrix) and scanned using the
GeneChip Scanner 3000 7 G (Affymetrix). A log 2 base
transformation was applied before the data was normalized.
Normalization and centering of all genes for each sample was
performed by BRB Array software. The data is deposited as GEO
Thoracic mammary glands were fixed for 24 hours in 4%
paraformaldehyde and embedded in paraffin wax. Paraffin
sections of 5 mm were prepared and subjected to 1 mM
disodium-EDTA antigen retrieval as described previously .
Primary antibodies used for immunofluorescence are the
following: Cytokeratin 8 (Developmental Studies Hybridoma Bank
TROMA-I, rat, 1:100), Estrogen receptor (Novocastra
NCL-ER6F11, mouse, 1:100), E-Cadherin (BD Biosciences 610181, mouse,
1:250), Progesterone receptor (Abnova MAB9785, rabbit, 1:400),
Smooth muscle actin (Sigma A2547, mouse, 1:1000), Tbx3
(Invitrogen 424800, rabbit, 1:100), turboGFP (Pierce Antibodies
PA522688, rabbit, 1:400), turboGFP (OriGene TA150041, mouse
1:250). Secondary antibodies used at 1:400 dilution are from
Invitrogen: Alexa488-coupled goat anti-mouse (A11029),
Alexa488-coupled goat anti-rabbit (A11034), Alexa568-coupled
goat anti-mouse (A11031) and Alexa568-coupled goat anti-rabbit
(A11036). Additionally, CF633nm-coupled donkey anti-rat
(Biotium 20137-1) was used at 1:400 dilution. Images were acquired
on a Zeiss LSM-710 confocal microscope with a pinhole aperture
of 1 Airy unit.
Colony forming assays
For analysis of colony forming potential, a pool of 46 Tbx3+/
Venus mice between the age of 10 to 16 weeks were used. One
thousand sorted cells were seeded into 22.1 mm (12-well) plate
with 0.75216106 NIH-3T3 feeder cells that had been treated
with mytomycin C (10 ug/ml) (Sigma, M4287). Colonies were
cultured for 56 days in MEC medium, consisting of DMEM:F12
medium (Invitrogen, 11320033) supplemented with Penicillin
Streptomycin (Gibco, 15140), 6% FCS (Hyclone, SV30160.03),
5 ng/ml cholera toxin (Sigma, 8052), 5 ug/ml Insulin (Sigma,
16634) and 10 ng/ml EGF (Sigma, E4127). At the end of the assay
colonies were fixed in 4% paraformaldehyde and stained with
1 mg/ml crystal violet (Sigma, C3886) in 50% dH20: 50%
methanol. Images were acquired on a Olympus IX71 inverted
Lentiviral vector construct & production
Several short hairpins against murine Tbx3 were cloned into a
MSCV-blast vector using a miR30 backbone and tested for knock
down efficiency in HC11 cells. The two best shRNAs (see File S2)
were digested with SalI and MluI and ligated into the GIPZ
cloning vector (Trans-Lentiviral Packaging kit TLP4616, Thermo
Scientific Open Biosystems) digested with XhoI and MluI (New
England Biolabs). Lentiviral particles were produced by
cotransfecting 32 ug of GIPZ plasmid (non-silencing or with a
shRNA against Tbx3), 10 ug of pSuper-Drosha (to reduce
processing of the RNA during packaging) and 30 ug packaging
plasmids (TLP mix) into HEK293T (HCL4517, Thermo Scientific
Open Biosystems) using the calcium phosphate method according
to the manufacturers protocol.
Abdominal mammary glands were harvested from 5 wildtype
FVB (experiment 1) and 4 KI (Tbx3+/Venus) donor mice
(experiment 2). The mammary glands were digested to single
cells and plated at 56105 cells/well in a 6-well plate. Cells were
adhered overnight in 3% oxygen and the next day cells were
subjected to spin transduction. Viral supernatant was diluted in
MEC medium at a 2:1 ratio, added to the cells at 1.5 ml/well and
spun at 2000 rpm for 30 minutes at 32uC. After centrifugation,
the cells were returned to the incubator (5% CO2 and 3% O2).
The next morning, cells were washed thrice with PBS and
trypsinized, spun and re-suspended in 1020 ml MEC medium.
Cells from one well were injected into cleared MG4 fat pads of
21day old matched recipient mice and allowed to engraft for 10
weeks . Glands were then harvested, fixed in methacarn (60%
methanol, 30% chloroform, 10% acetic acid) or 4%
paraformaldehyde for 24 hours and embedded in paraffin.
Tbx3 is differentially expressed in mammary epithelial
To investigate Tbx3 expression in the different mammary
epithelial cell (MEC) types, we made use of a reporter mouse strain
in which one of the Tbx3 alleles had been replaced with the gene
for Venus, a variant of yellow fluorescent protein (Tbx3+/Venus,
also referred to as knock-in (KI) mice). We isolated primary MECs
from Tbx3 wildtype and KI females and analyzed them by flow
cytometry. After excluding dead cells, lymphocytes and stromal
cells (see gating strategy in File S1), the epithelial cells segregated
in three distinct peaks based on Venus signal intensity (Figure 1A).
Venus expression accurately reflected Tbx3 transcription in the
three different peaks, with the highest Tbx3 mRNA levels in the
cells with the highest Venus fluorescence (Figure 1B).
To identify the cell type with high Tbx3 expression, we
separated MECs into cells of the luminal layer (CD24hiCD49flo,
blue) and basal layer (CD24luCD49fhi, red, Figure 1C). When the
luminal and basal cellular subsets are plotted separately, it is
apparent that the peak with intermediate Venus signal reflected
basal cells, and that luminal cells were divided into a subset with
high and one with low Tbx3 promoter activity (Figure 1D). This
distribution is quantified in independent animals in Figure 1E.
Plotting the populations with distinct Venus intensities on a
CD24/CD49f contour plot further illustrates the division of the
luminal population based on Tbx3 expression (Figure 1F).
High Tbx3 expression in hormone-sensing cells
Epithelial cells in the luminal layer fall into two main functional
categories, hormone-sensing (HS) cells and alveolar progenitor
cells and these can be separated by flow cytometry using the
additional cell surface markers Sca1 and CD49b (Figure 2A). The
proportion of HS and alveolar progenitor cells in Tbx3+/Venus
mammary epithelium is similar to that in wildtype litter mates
(Figure 2B), indicating that Tbx3 heterozygosity does not affect
the composition of the luminal layer. This could be the result of
relatively high Tbx3 mRNA levels in the heterozygote KI cells
(75% of wildtype cells, Figure 2C), which might suggest that Tbx3
is involved in a negative transcriptional feedback loop. This
experiment also demonstrated that Tbx3 expression is highest in
the HS population (Figure 2C), raising the question whether the
cells with highest Tbx3 expression are all hormone-sensing cells.
Plotting luminal cells based on their Tbx3 expression showed that
indeed almost all VenusHigh cells were part of the HS cell
population whereas almost all VenusLow cells belonged to the
alveolar progenitor cell population (Figure 2D). Similarly,
separating the luminal population based on cell type also showed that
the majority of the HS cell population was VenusHigh and the
alveolar progenitor population was VenusLow (Figure 2E). The
correlation between high Tbx3 expression and a hormone-sensing
cell identity was confirmed by transcriptional analysis by
microarray using cells pooled from three animals and separated
by Venus fluorescence (File S3) and by qPCR on luminal
populations sorted from individual KI animals (Figure 2F).
Luminal cells with low levels of Venus and Tbx3 expressed
variable but high levels of Elf5, a transcription factor that specifies
alveolar cell fate , and beta-Casein, one of the components of
milk (Figure 2E). Luminal cells that expressed high levels of Tbx3
had high levels of Sca1 transcription, in line with the flow
cytometry profiles, and expressed high levels of the estrogen and
progesterone receptor, thereby confirming the hormone-sensing
identity of Tbx3-expressing luminal cells at the molecular level.
Tbx3 expression marks the hormone-sensing cell lineage
In the non-pregnant adult mammary gland, proliferation is
detected mostly in ER- luminal cells, and this is mirrored by their
colony forming potential when plated on feeder layers [13,32].
Similarly, when we sorted mammary epithelial cells from adult
virgin mice, we found that luminal cells form more colonies than
basal cells (Figure 3A, 3B). The luminal colony-forming potential
was derived almost entirely from Tbx3-VenusLow cells (Figure 3C,
3D) as expected based on their alveolar ER- cell identity. Luminal
Figure 1. Fluorescent reporter reveals distinct Tbx3 expression in mammary epithelial cell subsets. (A) Epithelial cells isolated from
mammary glands of wildtype (Tbx3+/+) or knock-in (Tbx3+/Venus) mice show three peaks with different levels of Venus expression (Low, Medium and
High). (B) Mammary epithelial cells (MECs) from 3 independent Tbx3+/Venus animals were sorted according to Venus signal intensity. qPCR on 500
directly lysed cells shows that both Venus and Tbx3 mRNA correlates tightly with Venus fluorescence intensity. (C) MECs were labeled with
fluorescent antibodies against CD24 and a6-integrin (CD49f) to distinguish the luminal (blue) and basal (red) cell populations. (D) When plotted
separately, luminal cells from Tbx3+/Venus mammary glands show two main populations; VenusLow (cells with low Tbx3 expression) and VenusHigh
(cells with high Tbx3 expression). Basal cells express intermediate level of Venus (and Tbx3). (E) Quantification of the percentage of Venus-Low,
-Medium and -High cells in the luminal and in the basal population of Tbx3+/Venus epithelium. Data are presented as mean 6 SD of three individual
adult virgin Tbx3+/Venus animals. (F) Populations gated based on Tbx3 expression (Venus-Low, -Medium and -High) plotted on a CD24/CD49f contour
cells with high Tbx3 expression lacked colony-forming potential
(Figure 3D), consistent with other reports that show that the HS
cell population is non-clonogenic [13,33]. Nevertheless, Shehata
and colleagues recently identified a small subset of
hormonesensing progenitor cells that expressed high levels of
alpha2integrin (Sca1hiCD49bhi) and formed colonies on feeder cells .
Analyzing this ER+ progenitor population separately
demonstrated that the majority of these cells belonged to the Tbx3-VenusHigh
population (94%63%, Figure 3E) and mRNA levels of both Tbx3
and hormone receptors in Sca1hiCD49bhi cells were comparable
to that of the general hormone-sensing cell population (Figure 3F).
In line with previous studies [13,14], the ER+ progenitor
population is distinct in that it also expressed low levels of Elf5
and c-Kit. Luminal cells with high Tbx3 expression (VenusHigh)
that fell into the Sca1hiCD49bhi gate had robust colony forming
potential (Figure 3G, quantified in Figure 3H), confirming the
progenitor potential of this population. These experiments showed
that Tbx3 expression is already high in the earliest recognizable
cell type of the hormone-sensing lineage.
Taken together, analysis of the Tbx3+/Venus reporter by flow
cytometry combined with functional assays demonstrates that in
the luminal layer of mammary epithelium Tbx3 expression
distinguishes the hormone-sensing cell lineage from the secretory
Tbx3 protein expression in intact mammary glands
To further evaluate whether the strict correlation between
transcriptional expression of both Tbx3 and ER that we observed
by FACS holds true at the protein level in intact mammary glands,
we used co-immunofluorescence on paraffin sections from adult
virgin mice. The Tbx3 antibody gave high background staining in
the stroma, but clearly confirmed the existence of cells in the
Figure 2. Tbx3 marks hormone sensing cells. (A) Luminal cells from wildtype mammary glands are separated into hormone-sensing (HS,
Sca1hiCD49blo, purple) and alveolar (Sca1luCD49bhi, orange) subsets based on Sca1 and alpha2-integrin (CD49b) expression. (B) There is no significant
(n.s.) difference in the proportion of hormone-sensing (HS, purple) and alveolar cells (orange) between Tbx3+/+ (wildtype, WT) and Tbx3+/Venus
(Knockin, KI), paired t-test p = 0.53 for HS and p = 0.60 for Alv. (C) Tbx3 mRNA levels in sorted populations as indicated. (D) Tbx3+/Venus luminal cells were first
gated for Low or High Venus expression (see Figure 1D), and then plotted based on Sca1 and CD49b expression. (E) Proportion of hormone-sensing
(HS, purple) and alveolar cells (orange) that are VenusLow (grey) or VenusHigh (green), measured by FACS in 3 independent Tbx3+/Venus animals. (F)
Fold change in mRNA expression in luminal VenusLow (left panel) or VenusHigh (right panel) cells, relative to total luminal population. Data are
presented as mean 6 SD of three adult virgin Tbx3+/Venus animals.
luminal layer of the epithelial ducts with either a strong or
undetectable signal for Tbx3 in the nucleus (Figure 4A). The ER+
cells were 99.7%60.25% Tbx3 positive, with only 4 out of 1391
ER+ cells lacking Tbx3 expression. We did not observe any
Tbx3positive cells that did not express ER in adult virgin mice.
At the transcriptional level, basal cells displayed intermediate
Tbx3 expression (Figure 1D). On paraffin sections, it was difficult
to detect Tbx3 protein in cells of the basal layer (identified by the
basal marker smooth muscle actin, see Figure 4B). This could be
due to post-transcriptional regulation by for instance microRNAs,
but it could also be a technical limitation of the sensitivity of
detection. In the luminal layer, cells strongly express E-cadherin,
including the hormone-sensing cells with high Tbx3 expression
(Figure 4C). Tbx3 was found to repress E-cadherin in melanoma
cells , but it does not appear to do so in mammary epithelial
Previous studies have shown that Tbx3 is involved in
development of the rudimentary mammary gland during
embryogenesis , a process that is steroid hormone independent .
During puberty, steroid hormones induce the elongation of
mammary epithelial ducts. The invasive tips of elongating milk
ducts are called Terminal End Buds (TEBs) and we examined the
pattern of Tbx3 expression in these structures using mammary
glands from pubertal mice. Similar to the adult mammary gland,
Figure 4. Tbx3 expression in epithelial cells in intact mammary glands. (AC) Confocal immunofluorescence on paraffin sections of
mammary glands from wildtype adult virgin mice. (A) Ductal structure probed with antibodies for Tbx3 (green), the estrogen receptor (ER, red) and
the luminal cell marker cytokeratin-8 (CK8, blue). Nuclei are stained with DAPI (grey). (B) Duct probed for Tbx3 (green), the basal marker smooth
muscle actin (SMA, red) and CK8 (blue). (C) Duct probed for Tbx3 (green), E-Cadherin (E-cad, red) and CK8 (blue). Images are representative of staining
performed on paraffin sections of 3 independent animals. Scale bar is 20 mm or 10 mm for the inset.
the majority of ER+ cells in TEBs expressed Tbx3 (92.2%62.2%,
Figure 5A). However in puberty the occurrence of ER+ cells
without Tbx3 (2.6%63.4%) was more prevalent compared to the
adult virgin, and in puberty there were some Tbx3-positive cells
without detectable ER expression (4.3%61.7%). The staining for
Tbx3 appeared more intense in TEBs compared to quiescent adult
mammary tissue (Figure 5B). In puberty, a considerable
proportion of ER+ cells is proliferating , in contrast to the quiescent
adult stage. In early pregnancy, ER+ cells also proliferate  but
Tbx3 levels did not seem elevated during this stage of active
morphogenesis (Figure 5C). At day 3 of pregnancy, there
appeared to be a near perfect correlation between ER and
Tbx3, similar to the adult virgin. However, due to a lower staining
intensity for both ER and Tbx3 at this stage putative cells that
express only one or the other might be more difficult to detect.
Together, these experiments show that Tbx3 expression
consistently distinguishes hormone-sensing cells from ER- luminal
cells, not only by flow cytometry but also in unperturbed
mammary tissue at different developmental stages.
Tbx3 is required for the hormone-sensing cell lineage
To determine if Tbx3 expression is functionally relevant for the
hormone-sensing cell lineage, we designed two short hairpins that
target distinct regions of the murine Tbx3 transcript. The
reduction in Tbx3 mRNA by the shRNAs was confirmed in a
mouse mammary epithelial cell line (HC11) after selection of
lentivirally-transduced cells by puromycin (File S4A). Next, freshly
isolated MECs were incubated with lentiviral vectors overnight
and transplanted (without selection) into mammary fat pads
devoid of endogenous epithelium (see File S4B for the
experimental design). Eight to ten weeks after reconstitution, we
examined the identity of the cells that were transduced by
lentiviral vectors (as indicated by their turboGFP (tGFP)
expression). Outgrowths from cells that had been exposed to the control
virus illustrated that the majority of transduced cells were
lineagerestricted progenitors; we rarely found ducts whereby all cell types
were tGFP positive. Instead, we found ducts that contained tGFP+
cells that predominantly belonged to one lineage (either luminal
ER- or ER+ or basal, see File S5). For these experiments, we used
two different pools of donor cells; one consisting of wildtype MECs
Figure 5. Correlation between Tbx3 and ER expression at different developmental stages of postnatal mammary gland
development. Confocal immunofluorescence staining of Tbx3 (green), ER (red) and luminal marker cytokeratin-8 (CK8, blue) on mammary glands
from (A) 5-week old pubertal mice (terminal end bud structure), (B) 10-week old virgin mice (ductal structure) and (C) 3 day pregnant mice (ductal
structure). Nuclei are stained with DAPI (grey). White arrow heads indicate cells in terminal end bud (A) with ER expression but no Tbx3 expression.
Images are representative of staining performed on paraffin sections of 3 independent animals. Scale bar is 20 mm.
and the other consisting of Tbx3+/Venus MECs to potentially
facilitate further knockdown. Importantly, in both cases the pool of
cells transduced with the control vector gave rise to cells of all
different lineages, including ER+ hormone-sensing cells
(Figure 6A&B). Cells with Tbx3 knockdown robustly gave rise to luminal
ER- cells, but there was a strong bias against the generation of
hormone-sensing cells (Figure 6C&D). This was true for both the
wildtype and KI cells and for both short hairpins targeting Tbx3
(Figure 6A). In both control and Tbx3 knockdown outgrowths we
found some tGFP+ cells that contributed to cells of the basal layer
but this proportion was too low for accurate quantification. The
lack of adverse effects of Tbx3 knockdown in the Tbx3Low alveolar
lineage and the strong effect in the Tbx3-positive hormone-sensing
lineage demonstrate that Tbx3 expression is important for the
generation of hormone-sensing cells.
Tbx3 is required for mammary epithelial cell identity early in
embryogenesis, but its role in the different mammary epithelial cell
types of the adult mammary gland had not yet been determined.
Using a novel reporter mouse strain, we show that Tbx3
transcriptional expression is tightly regulated at different levels in
the three main epithelial lineages; high in hormone-sensing cells,
very low in alveolar progenitor cells and intermediate in basal
Based on the striking bimodal distribution of Tbx3 in the
luminal population, we focused our analysis in this study on the
role of Tbx3 in luminal mammary epithelial cells. During
embryonic heart and liver development, Tbx3 plays an important
role in specific lineage choices. For instance, in embryonic heart
development, Tbx3 is involved in a lineage choice between
pacemaker cells and atrial cardiac cells, in which Tbx3 represses
atrial genes in pacemaker cells [35,36]. Ectopic expression of Tbx3
can impose a pacemaker phenotype on atrial cells , showing
that in this context Tbx3 can direct cell fate.
Adult mammary epithelium is actively renewed by stem cells
and lineage-restricted progenitors . The bimodal expression of
Tbx3 in the luminal lineage, together with the expression of Tbx3
in ER+ progenitors, suggested that Tbx3 may play a role in the
lineage choice between the hormone-sensing and the alveolar
lineage. Indeed, we found that knockdown of Tbx3 in primary
mammary epithelial cells strongly reduced the formation of ER+
cells in mammary reconstitution assays, demonstrating that Tbx3
is required for a hormone-sensing cell fate. We have tried to
express a Tbx3 transgene in alveolar progenitor cells to determine
if Tbx3 could impose a hormone-sensing cell identity, but we
found only few cells that expressed Tbx3 ectopically in vivo (data
not shown). This could be due to an anti-proliferative effect of
Tbx3 , however ectopic Tbx3 expression in vitro did not
prevent proliferation of ER- cells. Therefore, we cannot rule out
that the scarcity of Tbx3 overexpressing cells in vivo was due to a
The differential expression of Tbx3 in the luminal lineage raises
the question of which signaling pathways influence Tbx3
transcription. Based on other studies using cell lines combined
with our observations presented in this study, we can speculate
what signals are most likely involved in regulating Tbx3 expression
in the intact adult mammary gland. For instance, in ER+ breast
cancer cell lines, Tbx3 expression was dependent on both estrogen
and FGF signaling . In primary tissue of murine mammary
glands, we found a strong correlation between ER and Tbx3
expression and it is therefore plausible that in adult mammary
epithelium Tbx3 is downstream of estrogen signaling. It was
shown that FGF signaling is required for Tbx3 expression in
mammary epithelial cells in the embryo . At this time the three
distinct adult lineages do not yet exist [6,7] and mammary
development is hormone independent . FGF signaling is also
active in pubertal mammary gland development [39,40] and we
observed strong Tbx3 staining in TEBs, raising the possibility that
FGF signaling also contributes to Tbx3 expression in postnatal
mammary epithelium. However, ER+ and ER- luminal cells
derived from adult mammary glands both express FGF receptors
and a downstream target of FGF signaling, Dusp6, is expressed
even higher in ER- luminal cells (File S6), and it is therefore
unlikely that FGF signaling is responsible for the hormone-sensing
cell specific expression of Tbx3.
Another pathway that may contribute to Tbx3 expression is
TGFb signaling. In MCF12A cells, a non-transformed mammary
epithelial cell line, TGFb directly induced Tbx3 transcription .
Notably, TGFb actively inhibits proliferation of MCF12A cells,
and also prevents proliferation of ER+ cells in the mammary gland
. In MCF12A cells, Tbx3 was required for the
antiproliferative effects of TGFb . This is surprising because
studies with other cell types suggested an oncogenic role of Tbx3.
For instance, Tbx3 was found to directly bind and repress the
tumor suppressor Arf in mouse embryonic fibroblasts , thus
preventing activation of p53. However, Tbx3 might have other
target genes in the mammary gland, since both Arf and p21
expression are robustly detectable on the microarray of sorted
primary Tbx3-VenusHigh cells (File S6). Moreover, loss of Arf or
p53 did not rescue developmental defects of mammary placodes in
Tbx3-mutants . Apart from Arf and p21 not many direct
targets for Tbx3 have been described. Tbx3 targets in the mouse
heart include genes for gap junctions and ion channels  and in
embryonic stem cells Tbx3 binds promoters of genes that are also
regulated by pluripotency factors , underscoring that Tbx3
likely regulates different target genes depending on the cellular
context . Indeed, genes that are repressed by Tbx3 such as
Ecadherin in melanoma cells  and p21 in promoter assays ,
are highly expressed in Tbx3-positive mammary epithelial cells
(Figure 4C and File S3). Determining the cistrome of Tbx3
specifically in hormone-sensing cells will likely help to unravel the
role of Tbx3 in normal and malignant mammary gland
In primary human breast tissue obtained from reduction
mammoplasties, Tbx3 expression is also highest in a population
of luminal cells that expresses high levels of steroid receptors ,
indicating that Tbx3 is a marker for hormone-sensing cells in
human mammary epithelium as well. Interestingly, this population
is also predominantly non-clonogenic , which raises the
possibility that Tbx3 has indeed an anti-proliferative role in
hormone-sensing cells in normal breast tissue. This would be in
line with data from sequencing breast cancer genomes; Tbx3
mutations are found specifically in ER+ breast cancers and these
mutations are predicted to result in loss of function [46,47]. This
indicates that Tbx3 might actively prevent tumorigenesis of
hormone-sensing cells resulting in selection pressure to lose that
function. However, it is possible that in tumors that have retained
wildtype Tbx3 expression, Tbx3 is involved in promoting
migration  or paracrine stimulation of proliferation .
Considering these potential conflicting roles of Tbx3 in breast
tumorigenesis, it will be important to further characterize the
specific role of Tbx3 in breast tissue before considering inhibition
of Tbx3 for cancer therapy .
In summary, we provide the first characterization of Tbx3
expression at the single cell level in the adult mammary gland.
By different methods we demonstrate that Tbx3 is highly
expressed in the hormone-sensing cell lineage and is functionally
required for the generation of hormone-sensing cells during
mammary morphogenesis. Our data highlight that Tbx3 is likely
to regulate a distinct set of target genes depending on the cellular
context. Given the current uncertainty about an anti- or
protumorigenic role of Tbx3 in breast cancer, it will be interesting
to determine the relevant targets for Tbx3 specifically in
File S1 FACS sorting of primary MECs. Antibodies used in
FACS sorting for separating the different mammary epithelial
populations. (B) Gating strategy for FACS analysis and sorting.
File S2 Primer sequences. (A) Polymerase chain reaction
(PCR) primers used for gene expression quantification by
quantitative PCR (qPCR). (B) Target sequences for the short
hairpins in Tbx3, and the short hairpin against Drosha used
during viral vector production.
File S3 Top 100 genes highest in luminal Venus High vs
Low cells. Tab 1. Top 100 genes highest expressed in luminal
VenusHigh cells compared to luminal VenusLow cells based on
Affymetrix microarray using cells pooled from three animals that
were separated by Venus fluorescence. Several described markers
for hormone-sensing cells are highlighted in bold. Tab 2. Top 100
genes highest expressed in luminal VenusLow cells compared to
luminal VenusHigh cells. Several described markers for alveolar
cells are highlighted in bold.
File S4 Transplantation of lentivirally-transduced
MECs. (A) mRNA levels of Tbx3 from puromycin-selected
HC11 that were transduced with either empty vector or short
hairpins targeting Tbx3. (B) Experimental set up for lentiviral
transduction of MECs and subsequent transplantation into cleared
mammary fat pads of 21-day old recipient mice. (C) For each
condition a small aliquot of cells was plated on coverslips while the
File S5 Examples of transduced lineage-restricted
progenitors. Paraffin sections of mammary outgrowths of MECs
transduced with lentiviral vectors. Transduced cells are identified
with an antibody staining against tGFP (green), luminal cells are
identified by cytokeratin 8 (blue) and HS cells are identified by the
estrogen or progesterone receptor (ER or PR, red). (A) Example of
an outgrowth containing transduced cells that belong to the
luminal alveolar (ER-negative) lineage (tGFP+CK8+ER-, white
arrow head). (B) Example of an outgrowth containing transduced
cells that belong to the luminal hormone-sensing lineage (tGFP+
CK8+PR+, white arrow). (C) Example of an outgrowth containing
transduced cells that belong to the basal lineage (tGFP+CK8-ER-,
white arrow head). (D) Transplanted fat pads were fixed with
either paraformaldehyde (PFA) or methacarn. Representative
File S6 Selected genes from microarray. Expression of
FGF receptors & ligands, cell cycle inhibitors and E-cadherin in
luminal VenusHigh and VenusLow cells (Affymetrix log2 values).
The authors thank Jen Nee Goh, Kakaly Ghosh, Bryan Tan and Ken
Chow for technical assistance, Mathijs Voorhoeve for helpful discussions
and critical reading of the manuscript and the Duke-NUS Genomics
Facility for performing the microarray.
Conceived and designed the experiments: KK TC AP. Performed the
experiments: KK VH TC DD. Analyzed the data: KK VH TC DD AP.
Contributed reagents/materials/analysis tools: MB VC. Wrote the paper:
1. Douglas NC , Papaioannou VE ( 2013 ) The T-box transcription factors TBX2 and TBX3 in mammary gland development and breast cancer . J Mammary Gland Biol Neoplasia 18 : 143 - 147 . doi:10.1007/s10911- 013 - 9282 -8.
2. Propper AY , Howard BA , Veltmaat JM ( 2013 ) Prenatal morphogenesis of mammary glands in mouse and rabbit . J Mammary Gland Biol Neoplasia 18 : 93 - 104 . doi:10.1007/s10911- 013 - 9298 -0.
3. Brisken C , O'Malley B ( 2010 ) Hormone action in the mammary gland . Cold Spring Harbor Perspectives in Biology 2: a003178 . doi:10.1101/cshperspect.a003178.
4. Meier-Abt F , Bentires-Alj M ( 2014 ) How pregnancy at early age protects against breast cancer . Trends Mol Med 20 : 143 - 153 . doi:10.1016/j.molmed. 2013 . 11.002.
5. Rios AC , Fu NY , Lindeman GJ , Visvader JE ( 2014 ) In situ identification of bipotent stem cells in the mammary gland . Nature 506 : 322 - 327 . doi:10.1038/ nature12948.
6. Van Keymeulen A , Rocha AS , Ousset M , Beck B , Bouvencourt G , et al. ( 2011 ) Distinct stem cells contribute to mammary gland development and maintenance . Nature 479 : 189 - 193 . doi:10.1038/nature10573.
7. van Amerongen R , Bowman AN , Nusse R ( 2012 ) Developmental stage and time dictate the fate of Wnt/b-catenin-responsive stem cells in the mammary gland . Cell Stem Cell 11 : 387 - 400 . doi:10.1016/j.stem. 2012 .05.023.
8. Boulanger CA , Wagner K-U , Smith GH ( 2005 ) Parity-induced mouse mammary epithelial cells are pluripotent, self-renewing and sensitive to TGFbeta1 expression . Oncogene 24 : 552 - 560 . doi:10.1038/sj.onc.1208185.
9. S ale S , Lafkas D , Artavanis-Tsakonas S ( 2013 ) Notch2 genetic fate mapping reveals two previously unrecognized mammary epithelial lineages . Nat Cell Biol 15 : 1 - 11 . doi:10.1038/ncb2725.
10. Chang TH-T , Kunasegaran K , Tarulli GA , De Silva D , Voorhoeve PM , et al. ( 2014 ) New insights into lineage restriction of mammary gland epithelium using parity-identified mammary epithelial cells . Breast Cancer Res 16 : R1. doi:10.1186/bcr3593.
11. Beleut M , Rajaram RD , Caikovski M , Ayyanan A , Germano D , et al. ( 2010 ) Two distinct mechanisms underlie progesterone-induced proliferation in the mammary gland . Proceedings of the National Academy of Sciences 107 : 2989 - 2994 . doi:10.1073/pnas.0915148107.
12. Mastroianni M , Kim S , Kim YC , Esch A , Wagner C , et al. ( 2010 ) Wnt signaling can substitute for estrogen to induce division of ERa-positive cells in a mouse mammary tumor model . Cancer Lett 289 : 23 - 31 . doi:10.1016/j.canlet. 2009 .07.012.
13. Shehata M , Teschendorff A , Sharp G , Novcic N , Russell A , et al. ( 2012 ) Phenotypic and functional characterization of the luminal cell hierarchy of the mammary gland . Breast Cancer Res 14 : R134. doi:10.1186/bcr3334.
14. Regan JL , Kendrick H , Magnay F-A , Vafaizadeh V , Groner B , et al. ( 2012 ) cKit is required for growth and survival of the cells of origin of Brca1-mutationassociated breast cancer . Oncogene 31 : 869 - 883 . doi:10.1038/onc.2011.289.
15. Hammond MEH , Hayes DF , Dowsett M , Allred DC , Hagerty KL , et al. ( 2010 ) American Society of Clinical Oncology/College Of American Pathologists guideline recommendations for immunohistochemical testing of estrogen and progesterone receptors in breast cancer . Journal of Clinical Oncology 28 : 2784 - 2795 . doi:10.1200/JCO. 2009 .25.6529.
16. Santagata S , Thakkar A , Ergonul A , Wang B , Woo T , et al. ( 2014 ) Taxonomy of breast cancer based on normal cell phenotype predicts outcome . J Clin Invest 124 : 859 - 870 . doi:10.1172/JCI70941.
17. Davenport TG , Jerome-Majewska LA , Papaioannou VE ( 2003 ) Mammary gland, limb and yolk sac defects in mice lacking Tbx3, the gene mutated in human ulnar mammary syndrome . Development 130 : 2263 - 2273 .
18. Jerome-Majewska LA , Jenkins GP , Ernstoff E , Zindy F , Sherr CJ , et al. ( 2005 ) Tbx3, the ulnar-mammary syndrome gene, and Tbx2 interact in mammary gland development through a p19Arf/p53-independent pathway . Dev Dyn 234 : 922 - 933 . doi:10.1002/dvdy.20575.
19. Bamshad M , Lin RC , Law DJ , Watkins WC , Krakowiak PA , et al. ( 1997 ) Mutations in human TBX3 alter limb, apocrine and genital development in ulnar-mammary syndrome . Nat Genet 16 : 311 - 315 . doi:10.1038/ng0797- 311 .
20. Bamshad M , Le T , Watkins WS , Dixon ME , Kramer BE , et al. ( 1999 ) The spectrum of mutations in TBX3: Genotype/Phenotype relationship in ulnarmammary syndrome . Am J Hum Genet 64 : 1550 - 1562 . doi:10.1086/302417.
21. Washkowitz AJ , Gavrilov S , Begum S , Papaioannou VE ( 2012 ) Diverse functional networks of Tbx3 in development and disease . Wiley Interdiscip Rev Syst Biol Med 4 : 273 - 283 . doi:10.1002/wsbm.1162.
22. Lu dtke TH-W, Christoffels VM , Petry M , Kispert A ( 2009 ) Tbx3 promotes liver bud expansion during mouse development by suppression of cholangiocyte differentiation . Hepatology 49 : 969 - 978 . doi:10.1002/hep.22700.
23. Brummelkamp TRT , Kortlever RMR , Lingbeek MM , Trettel FF , MacDonald MEM , et al. ( 2002 ) TBX-3, the gene mutated in Ulnar-Mammary Syndrome, is a negative regulator of p19ARF and inhibits senescence . J Biol Chem 277 : 6567 - 6572 . doi:10.1074/jbc.M110492200.
24. Rodriguez M , Aladowicz E , Lanfrancone L , Goding CR ( 2008 ) Tbx3 represses e-cadherin expression and enhances melanoma invasiveness . Cancer Res 68 : 7872 - 7881 . doi:10.1158/ 0008 - 5472 . CAN-08-0301.
25. Renard C -A, Labalette C , Armengol C , Cougot D , Wei Y , et al. ( 2007 ) Tbx3 is a downstream target of the Wnt/beta-catenin pathway and a critical mediator of beta-catenin survival functions in liver cancer . Cancer Res 67 : 901 - 910 . doi:10.1158/ 0008 - 5472 . CAN-06-2344.
26. Yarosh W , Barrientos T , Esmailpour T , Lin L , Carpenter PM , et al. ( 2008 ) TBX3 is overexpressed in breast cancer and represses p14 ARF by interacting with histone deacetylases . Cancer Res 68 : 693 - 699 . doi:10.1158/ 0008 - 5472 . CAN-07-5012.
27. Fillmore CM , Gupta PB , Rudnick JA , Caballero S , Keller PJ , et al. ( 2010 ) Estrogen expands breast cancer stem-like cells through paracrine FGF/Tbx3 signaling . Proc Natl Acad Sci USA 107 : 21737 - 21742 . doi:10.1073/ pnas.1007863107.
28. Tarulli GA , De Silva D , Ho V , Kunasegaran K , Ghosh K , et al. ( 2013 ) Hormone-sensing cells require Wip1 for paracrine stimulation in normal and premalignant mammary epithelium . Breast Cancer Research 15 : R10. doi:10.1186/bcr3381.
29. Ho V , Yeo SY , Kunasegaran K , De Silva D , Tarulli GA , et al. ( 2013 ) Expression analysis of rare cellular subsets: direct RT-PCR on limited cell numbers obtained by FACS or soft agar assays . BioTechniques 54 : 208 - 212 . doi:10.2144/ 000114019.
30. Smalley MJ ( 2010 ) Isolation, culture and analysis of mouse mammary epithelial cells . Methods Mol Biol 633 : 139 - 170 . doi:10.1007/978- 1 - 59745 - 019 -5_ 11 .
31. Oakes SR , Naylor MJ , Asselin-Labat M-L , Blazek KD , Gardiner-Garden M , et al. ( 2008 ) The Ets transcription factor Elf5 specifies mammary alveolar cell fate . Genes Dev 22 : 581 - 586 . doi:10.1101/gad.1614608.
32. Sleeman KE , Kendrick H , Robertson D , Isacke CM , Ashworth A , et al. ( 2007 ) Dissociation of estrogen receptor expression and in vivo stem cell activity in the mammary gland . J Cell Biol 176 : 19 - 26 . doi:10.1083/jcb.200604065.
33. Michalak EM , Nacerddine K , Pietersen A , Beuger V , Pawlitzky I , et al. ( 2013 ) Polycomb group gene Ezh2 regulates mammary gland morphogenesis and maintains the luminal progenitor pool . STEM CELLS 31 : 1910 - 1920 . doi:10.1002/stem.1437.
34. Li M , Fu X , Ma G , Sun X , Dong X , et al. ( 2012 ) Atbf1 Regulates Pubertal Mammary Gland Development Likely by Inhibiting the Pro-Proliferative Function of Estrogen-ER Signaling. PLoS ONE 7: e51283 . doi:10.1371/ journal.pone. 0051283 .t001.
35. Hoogaars WMH , Tessari A , Moorman AFM , de Boer PAJ , Hagoort J , et al. ( 2004 ) The transcriptional repressor Tbx3 delineates the developing central conduction system of the heart . Cardiovasc Res 62 : 489 - 499 . doi:10.1016/ j.cardiores. 2004 .01.030.
36. Hoogaars WMH , Engel A , Brons JF , Verkerk AO , de Lange FJ , et al. ( 2007 ) Tbx3 controls the sinoatrial node gene program and imposes pacemaker function on the atria . Genes Dev 21 : 1098 - 1112 . doi:10.1101/gad.416007.
37. Li J , Weinberg MS , Zerbini L , Prince S ( 2013 ) The oncogenic TBX3 is a downstream target and mediator of the TGF-b1 signaling pathway . Mol Biol Cell 24 : 3569 - 3576 . doi:10.1091/mbc. E13-05-0273.
38. Eblaghie MC , Song S-J , Kim J-Y , Akita K , Tickle C , et al. ( 2004 ) Interactions between FGF and Wnt signals and Tbx3 gene expression in mammary gland initiation in mouse embryos . J Anat 205 : 1 - 13 . doi:10.1111/j.0021- 8782 . 2004 .00309.x.
39. Parsa S , Ramasamy SK , De Langhe S , Gupte VV , Haigh JJ , et al. ( 2008 ) Terminal end bud maintenance in mammary gland is dependent upon FGFR2b signaling . Dev Biol 317 : 121 - 131 . doi:10.1016/j.ydbio. 2008 .02.014.
40. Lu P , Ewald AJ , Martin GR , Werb Z ( 2008 ) Genetic mosaic analysis reveals FGF receptor 2 function in terminal end buds during mammary gland branching morphogenesis . Dev Biol 321 : 77 - 87 . doi:10.1016/j.ydbio. 2008 .06.005.
41. Ewan KBR , Oketch-Rabah HA , Ravani SA , Shyamala G , Moses HL , et al. ( 2005 ) Proliferation of estrogen receptor-alpha-positive mammary epithelial cells is restrained by transforming growth factor-beta1 in adult mice . Am J Pathol 167 : 409 - 417 .
42. Lingbeek ME , Jacobs JJL , van Lohuizen M ( 2002 ) The T-box repressors TBX2 and TBX3 specifically regulate the tumor suppressor gene p14ARF via a variant T-site in the initiator . J Biol Chem 277 : 26120 - 26127 . doi:10.1074/ jbc.M200403200.
43. van den Boogaard M , Wong LYE , Tessadori F , Bakker ML , Dreizehnter LK , et al. ( 2012 ) Genetic variation in T-box binding element functionally affects SCN5A/SCN10A enhancer . J Clin Invest 122 : 2519 - 2530 . doi:10.1172/ JCI62613.
44. Han J , Yuan P , Yang H , Zhang J , Soh BS , et al. ( 2010 ) Tbx3 improves the germline competency of induced pluripotent stem cells . Nature 463 : 1096 - 1100 . doi:10.1038/nature08735.
45. Hoogaars WMH , Barnett P , Rodriguez M , Clout DE , Moorman AFM , et al. ( 2008 ) TBX3 and its splice variant TBX3 + exon 2a are functionally similar . Pigment Cell Melanoma Res 21 : 379 - 387 . doi:10.1111/j.1755- 148X.2008.00461.x.
46. Stephens PJ , Tarpey PS , Davies H , Van Loo P , Greenman C , et al. ( 2012 ) The landscape of cancer genes and mutational processes in breast cancer . Nature 486 : 400 - 404 . doi:10.1038/nature11017.
47. Ellis MJ , Ding L , Shen D , Luo J , Suman VJ , et al. ( 2012 ) Whole-genome analysis informs breast cancer response to aromatase inhibition . Nature : 1 - 8 . doi:10.1038/nature11143.
48. Peres J , Davis E , Mowla S , Bennett DC , Li JA , et al. ( 2010 ) The Highly Homologous T-Box Transcription Factors, TBX2 and TBX3, Have Distinct Roles in the Oncogenic Process . Genes & Cancer 1 : 272 - 282 . doi:10.1177/ 1947601910365160.
49. Lu J , Li X-P , Dong Q , Kung H-F , He M-L ( 2010 ) TBX2 and TBX3: the special value for anticancer drug targets . Biochim Biophys Acta 1806 : 268 - 274 . doi:10.1016/j.bbcan. 2010 .07.001.