Stat3 Mediates Expression of Autotaxin in Breast Cancer
Janeen Azare1, Ashley Doane2, Kenneth Leslie1, Qing Chang1, Marjan Berishaj1, Jennifer Nnoli1, Kevin
Mark1, Hikmat Al-Ahmadie2, William Gerald2, Maryam Hassimi3, Agnes Viale3, Mary Stracke5, David
Lyden4,6*, Jacqueline Bromberg1*
1Department of Medicine, Memorial Sloan Kettering Cancer Center, New York, New York, United States of America, 2Department of Pathology, Memorial Sloan Kettering
Cancer Center, New York, New York, United States of America, 3Genomics Core Laboratory, Memorial Sloan Kettering Cancer Center, New York, New York, United States
of America, 4Department of Pediatrics, Memorial Sloan Kettering Cancer Center, New York, New York, United States of America, 5Laboratory of Pathology, Division of
Clinical Sciences, NCI, National Institutes of Health, Bethesda, Maryland, United States of America, 6Department of Pediatrics, Weill Cornell Medical College, New York,
New York, United States of America
We determined that signal transducer and activator of transcription 3 (Stat3) is tyrosine phosphorylated in 37% of primary
breast tumors and 63% of paired metastatic axillary lymph nodes. Examination of the distribution of tyrosine
phosphorylated (pStat3) in primary tumors revealed heterogenous expression within the tumor with the highest levels
found in cells on the edge of tumors with relatively lower levels in the central portion of tumors. In order to determine Stat3
target genes that may be involved in migration and metastasis, we identified those genes that were differentially expressed
in primary breast cancer samples as a function of pStat3 levels. In addition to known Stat3 transcriptional targets (Twist,
Snail, Tenascin-C and IL-8), we identified ENPP2 as a novel Stat3 regulated gene, which encodes autotaxin (ATX), a secreted
lysophospholipase which mediates mammary tumorigenesis and cancer cell migration. A positive correlation between
nuclear pStat3 and ATX was determined by immunohistochemical analysis of primary breast cancer samples and matched
axillary lymph nodes and in several breast cancer derived cell lines. Inhibition of pStat3 or reducing Stat3 expression led to a
decrease in ATX levels and cell migration. An association between Stat3 and the ATX promoter, which contains a number of
putative Stat3 binding sites, was determined by chromatin immunoprecipitation. These observations suggest that activated
Stat3 may regulate the migration of breast cancer cells through the regulation of ATX.
Citation: Azare J, Doane A, Leslie K, Chang Q, Berishaj M, et al. (2011) Stat3 Mediates Expression of Autotaxin in Breast Cancer. PLoS ONE 6(11): e27851.
Editor: Guenter Schneider, Technische Universita ¨t Mu ¨nchen, Germany
Received June 23, 2011; Accepted October 26, 2011; Published November 28, 2011
Copyright: ? 2011 Azare et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This project was supported by grants from the National Institutes of Health (R01 CA87637), Marjorie and Charles Holloway Foundation, Breast Cancer
Alliance Award, Sussman Family Fund and Lerner Awards to JB and a Cancer Research and Prevention Foundation and Intercultural Cancer Council Fellowship and
a National Cancer Institute, CMBB (Comprehensive Minority Biomedical Branch) Training Award to JA. 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.
* E-mail: firstname.lastname@example.org (JB); email@example.com (DL)
Breast cancer is the most common malignancy diagnosed
among women worldwide . Despite significant improvements in
the diagnosis and treatment of this disease, tumor dormancy
followed by distant recurrences accounts for 90% of all cancer
deaths. Micrometastasis in the blood and bone marrow are the
principal targets for adjuvant therapy [2,3,4,5]. However, these
metastatic cells can evade therapeutic interventions and eventually
lead to recurrence. Clearly understanding the molecular mecha-
nisms underlying the development of metastatic disease is required
in order to treat this fatal disorder effectively.
Stat3 is a transcription factor which is known for its role as an
integrator of cytokine and growth factor signaling . Stat3
activation is dependent upon tyrosine phosphorylation, leading to
dimerization between two Stat3 molecules. Activated Stat3
translocates to the nucleus where it binds to consensus promoter
sequences of target genes and regulates their transcription. In
contrast to normal cells where Stat3 activationis a transient process,
Stat3 is persistently activated in a number of epithelial tumors
including breast cancer and there is increasing evidence demon-
strating that activated Stat3 plays a critical role in the pathogenesis
of breast cancer including metastatic progression and response to
therapy [7,8,9,10,11,12,13,14,15,16,17] [18,19]. Breast tumors
expressing high levels of activated Stat3 are inversely correlated
with a complete pathological response to neo-adjuvant chemother-
apy . Inhibition of Stat3 activation in breast cancer cells inhibits
growth and neo-angiogenesis, and potentiates a response to the
chemotherapeutic agent doxorubicin [16,21,22]. Autocrine IL-6
production, a principal mediator of Stat3 activation in breast
tumors, was found to be elevated in human mammary cancer/stem
cells. Blockade of this signaling pathway reversed the aggressive
features characteristic of basal-like breast cancers [23,24]. In
addition, side-population breast cancer stem-like cells express and
require persistently activated Stat3 for viability and maintenance
. The mechanism(s) by which activated Stat3 mediates its effects
is primarily through its ability to regulate gene transcription.
Although a number of Stat3 target genes including vascular
endothelial growth factor (VEGF), survivin, matrix metalloprotei-
nase-9 (MMP-9) and twist have been identified in primary breast
cancers and cancer-derived cell lines, we were interested in
identifying additional target genes which may participate in
metastatic progression of breast cancer [11,20,26,27,28,29].
PLoS ONE | www.plosone.org1November 2011 | Volume 6 | Issue 11 | e27851
Autotaxin (ATX) or nucleotide pyrophosphatase-phosphodies-
terase 2 (ENPP2), a secreted glycoprotein with lysophospholipase
D activity, promotes cell migration, metastasis, and angiogenesis
through the generation of lysophosphatidic acid (LPA), a lipid
mitogen and motility factor that acts on several G protein-coupled
receptors [30,31,32,33,34] . Elevated levels of ATX have been
demonstrated to play a role in migration and invasion of
glioblastoma, lymphoma, hepatocellular carcinoma, melanoma
and breast cancers, establishing this enzyme as a likely mediator of
metastatic disease [36,37,38,39,40,41,42,43,44,45,46]. Significant-
ly, enforced expression of ATX in metastatic models of breast
cancer enhances osteolytic bone metastases while reduced
expression of ATX inhibits bone metastases through regulation
of osteoclasts . Furthermore, it has recently been shown that
over-expression of either ATX or LPA to the mammary gland
mediates de novo tumorigenesis suggesting the oncogenic nature
of this pathway [48,49].
Here we examined primary breast cancer samples with matched
axillary lymph nodes and observed a greater percentage of lymph
nodes expressing moderate to high levels of pStat3 in contrast to
the primary tumor. Upon further analysis of the primary tumors
we determined that pStat3 levels were frequently highest along the
leading edge of the primary tumor. These data suggested a role for
pStat3 in mediating metastatic spread from the primary tumor to
the axillary lymph nodes. By using gene expression profiling of
primary breast tumors either expressing or lacking pStat3 protein,
we identified a number of potential Stat3 target genes which may
be involved in metastasis including ENPP2 (ATX). The ENPP2
promoter contains multiple Stat3 binding sites and an association
between Stat3 and these binding sites was observed. We
demonstrated a positive correlation between moderate to high
levels of pStat3 and ATX in primary breast tumors and lymph
nodes as well as in several breast cancer derived cell lines.
Activated Stat3 mediated ATX up-regulation and enhanced
migration of breast cancer cell lines. Conversely, inhibition of
Stat3 activity blocked migration with a concomitant decrease in
ATX levels. In summary, we have identified ATX as a putative
novel Stat3 target gene in breast cancer.
Materials and Methods
Samples and Gene Expression Analysis
Tissue samples were obtained from therapeutic or diagnostic
procedures performed as part of routine clinical management at
Memorial Sloan-Kettering Cancer Center (MSKCC). All research
procedures using human tissue were approved by the MSKCC
institutional review board (06-177). Patients signed an informed
consent demonstrating their interest and willingness to have their
tissues analyzed for research purposes. Tissues were snap frozen in
liquid nitrogen and stored at 280uC. Each sample was examined
sections and enriched for areas of interest by manual trimming
of tissue blocks. Isolation of RNA, cDNA synthesis and gene
expression analysis using the HG-U133A microarray was
previously described . The expression data on these tumor
samples can be accessed using http://caarraydb.nci.nih.gov/
caarray/. We used samples which were definitively pStat3+ versus
pStat3 negative throughout all cell types within the tumor for our
gene expression analysis. Affymetrix.CEL files were analyzed using
tools in Partek Genomic Suite 6.4 software (Partek Inc). The
Robust Multi-array Analysis (RMA) algorithm was used for global
normalization and probeset summarization. To identify differen-
tially expressed genes between two groups (pStat3+ versus
pStat32), a students T-test was performed followed by a 2-fold
change filter. The statistically significant genes (214 gene list of
differentially expressed genes) were used to perform two-way
hierarchical clustering of tumor samples. Pearson dissimilarity
metric and average linkage method were used to calculate the
TMAs and immunohistochemistry
Multi-tissue blocks of formalin-fixed, paraffin- embedded
primary breast cancer and matched axillary lymph nodes
(containing 3 or 4 representative 0.6-mm cores) were prepared
using a tissue arrayer, and immunohistochemistry was performed
as described previously [23,50]. Antigen retrieval, with the use of
citric acid (pH 6.0) at 97uC for 30 minutes, was followed by
treatment with 3% hydrogen peroxide. pSTAT3 (Tyr705) (Cell
Signaling; #9131) and ATX (M. Stracke) were used at 1:200 and
1:1000 respectively [23,43]. The ATX rabbit polyclonal antibody
was affinity purified and generated against a C-terminal peptide
which allows for recognition of all 3 isoforms of ATX as well as
both cytosolic and secreted forms of ATX [51,52,53,54].
Denatured recombinant ATX was used as a blocking agent for
IHC and these controls revealed no staining demonstrating
specificity of this antibody for IHC (data not shown). Secondary
reagents were from DakoCytomation EnVision+ Dual Link
System-HRP (DAB+) kit or the DakoCytomation LSAB+
system-HRP kit. Counterstaining was performed with the use of
hematoxylin. Scoring of the TMAs was performed by 2
independent observers, with a high correlation observed between
scorers (P,0.001) for both pSTAT3 and ATX. For a tumor to be
considered positive for either pSTAT3 or ATX, all 3–4 replicates
in the tissue array had to have a similar staining intensity,
otherwise, it was excluded. pStat3 levels were subjectively graded
as a function of relative nuclear staining intensity: No/low staining
of epithelial cells (0–1+) and moderate/high staining (2–3+). We
also classified pStat3 levels within the stromal and immune
compartment of the tumor and for our gene expression analysis we
excluded those samples which were positive for pStat3 in the
tumor microenvironment yet negative in the epithelial compart-
ment. We classified tumors as expressing no cytoplasmic ATX (0),
low staining (+1) and moderate/high staining (+2–3). Statistical
analyses were done using StatView (SAS Institute). The correlation
between the scores of both scorers and the relation between those
of pSTAT3 and ATX were measured by using the x2 test.
Cell culture and Reagents
Human breast cancer MDA-MB-435, MDA-MB-231, BT474,
MCF-7, 1937, 1806, 1143, 38, 1833 and 4175 cells (ATCC and
kindly provided by J. Massague [55,56,57] were cultured in DME
HG: F12 P+S+NEAA containing 10% fetal bovine serum
(Invitrogen). 1833Stat3Sh and 1833CSh cells were generated
using a Stat3Sh lentiviral construct as well as a scrambled control
construct . Cells were treated with 5 ng/ml oncostatinM
(Chemicon) or 1 mM P6 pan-Jak inhibitor (Calbiochem). ATX
rabbit polyclonal antibody was used at 1:10,000 for western blot
analysis and 1:1000 for IHC (M. Stracke-see description above for
details of antibody) [51,52,53,54]. pStat3 rabbit polyclonal
antibody (Cell Signaling, #9131) was used at 1:1000 for western
blot and 1:200 for IHC. Stat3 rabbit polyclonal antibody (Cell
Signaling, #9132) was used at 1:1000 for western blotting. Anti-
Stat3 antibody for ChIP was used at 1:200 (Santa Cruz, #C-20X).
Tubulin monoclonal antibody (Sigma) was used at 1:10,000.
Vimentin antibody for western blot was used at 1:1000 (Sigma).
pStat5 antibody for western blot was used at 1:1000 (Cell
ATX a Stat3 Target in Breast Cancer
PLoS ONE | www.plosone.org2November 2011 | Volume 6 | Issue 11 | e27851
Cell lysate preparation and Western blotting
Cells were treated with or without 5 ng/ml oncostatinM (OSM;
Chemicon International, Inc.) for 1 h or with or without 1 mM
Pyridone 6 (Calbiochem) for 24 h, and then whole-cell lysis
extracts were obtained: 150 mM NaCl, 10% glycerol,1%
IGEPAL, 0.5% sodium deoxycholate, 2 mM EDTA, 1 mM
NaF, 1 mM Na3VO4, 1 mM Na2MoO4, 0.1% SDS, 20 mM
Tris-HCl, pH 8 with proteinase inhibitors. Western blots analysis
was performed as previously described .
Reverse-transcriptase-PCR (RT-PCR) amplification
Total RNA was isolated from cells treated with or without OSM
for 4 h with or without 1 mM P6 for 24 hr using the RNeasy?
Mini Kit (Qiagen Inc.) according to the manufacturer’s instruc-
tions followed by reverse transcription. The cDNA samples were
amplified by PCR using 32P-labeled deoxynucleosidetriphosphate.
Primers for ATX were: sense, 59-CGT GAA GGC AAA GAG
AAC ACG-39, and antisense, 59-AAA AGT GGC ATC AAA
TAC AGG-39, of the ATX cDNA, producing a 784 bp product.
Primers for the b-actin internal control housekeeping gene were:
sense, 59-CGT GCG TGA CAT TAA GGA GA-39, and
antisense, 59-TGA TCC ACA TCT GCT GGA AG-39,
producing a 450 bp product.
Chromatin immunoprecipitation assays were done on MDA-
MB-435, 1833CSh and 1833S3Sh cells either treated with DMSO
control, OSM for 30 minutes or P6 for 4 hrs.using a chromatin
immunoprecipitation assay kit (Upstate Biotechnology). Stat3-
DNA complexes were precipitated by using anti-Stat3 antibody
(#C-20X, Santa Cruz). Polyclonal IgG antibody was used as a
negative control. Precipitated DNA was amplified by PCR using
primers flanking the gamma-activated sites (GAS) sites at 2821 -
59-CGAAACAAGCTGACAG and 2301 - 59-GGCCCATAA-
CAGTGCATGTTC A 58uC annealing temperature using 30
cycles was used for amplification.
Cell migration assay (Boyden chambers)
MDA-MB-435 and 1833 derived cell lines were preincubated
for 16 h in 1% serum-containing medium and conditioned media
(CM) was collected from these. Migration assay was performed
using a 24-well cell culture plate (Becton Dickinson) with 8.0-mm
pore membrane inserts (Becton Dickinson). To the lower
chambers, conditioned media from MDA-MB-435 and or 1833
derived cell lines either treated for 16 hr with OSM (5 ng/ml) or
P6 (1 mM) was added to the lower chamber. MDA-MB-435 and
1833 derived cell lines (16105) were added to the upper wells, and
the chambers were incubated for 24 h at 37uC. Migrated cells
were visualized with 0.5% crystal violet in 20% methanol and
counted. Each condition was assayed in triplicate, experiments
were performed independently at least three times, and the results
were expressed as the number of cells per field. A one-way analysis
of variance was used to determine significance.
Tyrosine phosphorylated Stat3 in breast cancer and
matched metastatic axillary lymph nodes
We have described the prevalence of moderate to high levels of
tyrosine phosphorylated Stat3 (pStat3) in primary breast cancers to
be 46% with no association with estrogen, progesterone or Her2/
neu receptor expression . These findings are consistent with
other reports [9,20,59,60,61]. Furthermore, high levels of pStat3
within primary tumors are associated with metastasis to regional
lymph nodes . However, the relationship between pStat3 in
primary tumors and matched axillary lymph node metastases has
not been described. We examined tissue micro-arrays (TMAs) of
38 primary breast tumors with matched axillary lymph node
metastases for the relative expression of pStat3 by immunohisto-
chemistry. pStat3 levels were subjectively graded as a function of
relative nuclear staining intensity: No/low staining of epithelial
cells (0–1+) and moderate/high staining (2–3+) defined negative
and positive pStat3 expression, respectively. 14/38 (37%) of
primary tumor samples were IHC positive (2–3+); while 24/38
(63%) of matched lymph node metastases were IHC postive (2–3+)
(Fig. 1A and B). We also observed that the relative pStat3 IHC
staining was greater in the lymph node metastasis compared to the
corresponding primary tumor in 15/38 specimens, equal in 18/38
and less in 5/38 (Fig. 1B). Given that tumors are heterogeneous,
we hypothesized that a subset of breast cancer cells expressing
activated Stat3 were capable of metastasizing to the axillary lymph
nodes. To this end, we examined pStat3 levels and distribution of
whole tumor sections from the original paraffin tissue blocks from
which the 1 mm cores were obtained to generate the TMAs. We
determined that pStat3 levels were highest along the edge of the
tumor section in .90% of tumors while pStat3 staining within the
lymph node metastases was more uniform in distribution (Fig. 2).
This finding was not however due to an artifact of antigen retrieval
as areas of normal breast surrounding the tumor section were
negative for pStat3 (data not shown).
Identification of Stat3 regulated genes involved in
In order to identify potential Stat3 target genes in breast cancer
which may regulate invasion, migration and metastasis we sought
to identify genes that were differentially expressed in tumor
samples as a function of pStat3 as assessed by IHC. Gene
expression profiling had been performed on 99 primary breast
cancer samples (42 ER2 and 57 ER+) and we previously reported
that 46% of these were immunohistochemically positive for pStat3
[23,50]. Those samples from the TMA which expressed moderate
to high levels of pStat3(2–3+) in tumor epithelial cells versus those
that were negative for pStat3 were further evaluated by examining
pStat3 levels and tissue distribution from sections of the original
paraffin blocks containing the tumor specimens. Many of the
specimens that were negative for pStat3 in the tumor expressed
moderate to high pStat3 levels in surrounding lymphocytes,
endothelial cells or stromal cells (data not shown). However, we
identified 16 specimens which expressed no pStat3 staining in
tumor cells, lymphocytes, endothelial cells or stromal cells. 13/16
of these pStat3 negative specimens were ER negative. Given the
significant effect that ER has on gene expression and the potential
for molecular cross talk between Stat3 and ER signaling, we chose
to restrict our initial analysis to the ER2 subset of tumors in order
to identify genes associated with Stat3 activation.
Of the 42 ER2 tumor samples, 8 were strongly positive for pStat3
(3+ in tumor cells only) and 11 were moderately to strongly positive
for pStat3 (2–3+ in the tumor) but also had many infiltrating pStat3+
stromal, endothelial cells and/or lymphocytes. In contrast, 13 were
negative for pStat3 (in tumor and surrounding stromal, endothelial
in the infiltrating lymphocytes,endothelial orstromal cellsand 5 were
weakly pStat3+ (0–1+) within the tumor cells. Given that our samples
had not undergone laser capture dissection we felt it necessary to
distinguish those samples ‘‘contaminated’’ with pStat3+ stromal and
immune cells in our gene expression analysis. We were interested in
identifying those genes which were differentially expressed predom-
ATX a Stat3 Target in Breast Cancer
PLoS ONE | www.plosone.org3November 2011 | Volume 6 | Issue 11 | e27851
inantly as a function of the expression of pStat3 within the tumor/
epithelial compartment. A microarray statistical analysis of the 8
pStat3+ (2–3+ in tumor cells only) versus the 13 pStat32 (0 in tumor,
endothelial, stromal and lymphocytes) tumor specimens led to the
identification of 214 genes which were differentially expressed
according to pStat3 status (at least two-fold between the means of
pStat3(+) and pStat3(2) cases and a Student’s t-test P,0.05). Of the
cases (Table S1). Differentially expressed genes included several
previously identified Stat3 targets such as Twist, Snail and IL-8
[29,62,63,64,65,66,67,68], and this suggested our analysis at least
partially captured a molecular profile of Stat3 transcriptional
activation in breast cancer. Interestingly, the top ranked differentially
expressed gene according to significance testing (p,0.0001) and fold
change (8.6 fold) was ENPP2 which encodes ATX, a secreted
glycoprotein that promotes cell migration, metastasis, and angiogen-
esis. Two-way hierarchical clustering was performed using expression
of the 214 differentially expressed genes in 21 ER(2) breast cancers
status. This analysis demonstrated a clear separation of these two
groups as expected. The cluster of genes containing ENPP2 is
magnified to illustrate its expression pattern across the samples
(Fig. 3A). We also included the remaining 21 ER2 tumor specimens
in our clustering analysis which either expressed low levels of pStat3
(0–1+) within tumor cells with variable levels of pStat3 within the
‘‘microenvironment’’ (blue rectangles without borders) versus
moderate to high levels of pStat3 (2–3+) within tumor cells also with
variable levels of pStat3 in the microenvironment (orange rectangles
without borders) which also revealed good separation of tumor
specimens as a function of pStat3 status as expected(Fig. 3B). In
addition, hierarchical clustering on all of the tumor specimens for
which pStat3 status had been determined (42 ER2 and 57 ER+)
revealed two sample groups associated with pStat3 status (Fig. 3C).
Within the pStat3 positive group (group 2) ENPP2/ATX levels
appeared to be elevated in both the ER+ (grey rectangles) and the
ER2 (clear rectangles) specimens relative to the pStat3 negative
group. To further explore this expression pattern we repeated the
statistical analysis using all of the tumor specimens to identify
differentially expressed genes as a function of pStat3 (blue versus
(TableS2).Acomparisonofthisgenelist withthe214pStat3gene list
revealed 50 genes overlapping including ENPP2 (Table S3) as
expected. Given the known significance of ENPP2/ATX as a key
in regulating the metastatic potential of pStat3(+) breast cancers
ATX expression in breast cancer correlates with pStat3
By immunohistochemical analysis we next examined the levels
of ATX in the 42 ER2 primary breast tumors. Tumors were
classified as expressing no ATX (0; 24%), low staining (1+; 27%)
and moderate/high staining (2–3+; 49%) (Fig. 4A). A positive
correlation between moderate/highATX and pStat3 levels (2–3+)
was determined by x2analysis (Fig. 4B; p=0.0025). The 38
primary tumors with matched axillary lymph nodes were also
examined for ATX levels. 12/38 (32%) primary tumors expressed
moderate to high (2–3+) levels of ATX; while 22/38 (58%) lymph
nodes were (+2–3) positive for ATX (data not shown). Further-
more, a positive correlation between ATX and pStat3 expression
within lymph nodes was observed p=0.0008 (Fig. 4C). We also
examined the distribution of ATX levels in the whole tumor
section as was performed for pStat3 in Fig. 2. Here we also
observed relatively higher levels of ATX on the edge of the tumor
specimens (Fig. 4D).
Stat3 regulates ATX expression and cell migration in
breast cancer cells
In order to characterize the relationship between activated Stat3
and ATX, we examined a number of breast cancer derived cell
lines for expression of ATX message and protein relative to pStat3
(Fig. 5A, 5B and 5C). MCF-7 and BT474 breast cancer derived
Figure 1. Tyrosine phosphorylated Stat3 in breast cancer and matched metastatic axillary lymph nodes. A. Tissue microarrays (TMAs)
of primary breast tumors (38) were analyzed for nuclear tyrosine-phosphorylated Stat3 (pStat3) by immunohistochemical (IHC) analysis. Examples of
no detectable pStat3 (0), low levels (1+), moderate levels (2+) and high levels (3+) are shown. B. Comparison of pStat3 levels of tumor cells by IHC in
matched primary tumors with axillary lymph nodes. 37% of primary tumors expressed 2–3+ pStat3, while 63% of matched lymph nodes expressed 2–
3+ pStat3. Relative pStat3 IHC staining was greater in the lymph node compared to the corresponding primary tumor in 15/38 specimens, equal in
18/38 and less in 5/38. A representative example of a tumor (Tu) specimen with 1+ pStat3 versus 3+ pStat3 in the corresponding lymph node (Ln) is
ATX a Stat3 Target in Breast Cancer
PLoS ONE | www.plosone.org4 November 2011 | Volume 6 | Issue 11 | e27851
cell lines expressed low levels of pStat3 and no to low expression of
ATX mRNA; while 1937, 1806, 1143, 38, MDA-MB-231, 1833,
4175 and MDA-MB-435 cells expressed high levels of pStat3
protein and ATX mRNA (Fig. S1A and Fig. 5A). MDA-MB-435,
231, 4175 and 1833 cells express activated Stat3 as a consequence
of persistant activation of gp130 signaling . Activated Stat5 has
been reported to regulate some of the same genes as activated
Stat3 [18,19]. We determined that 1833 and to a lesser degree
4175 cells expressed pStat5, while MCF-7, BT474, MDA-MB-231
and MDA-MB-435 cells did not (Fig. S1B and S1C). The role of
pStat5 in the regulation of ATX is not clear. Jak inhibition (P6)
reduced both pStat3 and ATX but had no effect on pStat5 levels
in 1833 cells, suggesting a marginal role for this transcription
factor in ATX regulation.
Stimulation of MDA-MB-435 cells with oncostatin M (OSM), a
growth factor which activates the gp130/Jak/Stat3 pathway, led
to an increase in the levels of activated Stat3 with a concomitant
increase in ATX protein and mRNA levels (Fig. 5B, 5C). In
corcordance with this finding, inhibition of gp130 signaling using a
pan-Jak inhibitor (P6) led to a marked decrease in pStat3 and
ATX levels in all of the described cell lines (Fig. 5B, 5C and data
not shown), suggesting at Stat3 dependent mechanism. Vimentin
(a marker of EMT) levels also increased in response to OSM and
were reduced in response to P6 (Fig. 5B). In order to demonstrate
that Stat3 rather than pan-Jak inhibition was responsible for
mediating ATX expression, we knocked-down Stat3 levels using a
Stat3 ShRNA (S3Sh) in 1833 cells and determined that ATX
protein and RNA (data not shown) levels were decreased
compared to cells infected with a scrambled control ShRNA
(CSh) (Fig. 5B) [58,69].
To investigate the role of Stat3 in the transcriptional regulation
of the human ATX gene, we examined 1000-bp of the ATX
promoter and identified 7 putative Stat3 binding sites (TTN5AA;
N$2GC residues) (Fig. 5D). Cross-linking chromatin immuno-
precipitation assays were carried out to determine whether the
potential Stat3 binding sites within the ATX promoter bound
activated Stat3. Primer sets were used to target the regions of the
ATX promoter corresponding to the Stat3 binding sites found
between 2821 and 2301. Cell extracts from MDA-MB-435 cells
were subjected to chromatin immunoprecipitation (ChIP) analysis
with antibodies to Stat3 or IgG (negative control). Amplification of
the ATX promoter by PCR was observed in the samples expressing
activated Stat3 (Control and OSM stimulated) but not in the P6
treated samples (Fig. 5D). Similarly, ChIP analysis of Stat3 bound
to the ATX promoter was performed on 1833 cells expressing
activated Stat3 (1833CSh) compared to those expressing less Stat3
(1833S3Sh) revealing PCR amplification of the ATX promoter in
the Stat3 expressing line (1833CSh) (Fig. 5D). These data
demonstrate that activated Stat3 can associate with the human
ATX is a known mediator of cell migration, invasion and
angiogenesis principally through the production of LPA . Here
we examined the effect of increasing or decreasing pStat3/ATX
levels on MDA-MB-435 and 1833 cell migration. Migration was
measured using a Boyden chamber assay in the presence of
conditioned media (CM) from MDA-MB-435 or 1833 cells
untreated, stimulated with OSM or treated with P6. We observed
enhanced migration of cells in the presence of CM from OSM
treated cells compared to control CM from MDA-MB-435 cells
and inhibition of migration with CM from P6 treated MDA-MB-
435 and/or 1833 cells (Fig. 5E). We also demonstrated reduced
cell migration in the presence of CM from 1833 cells lacking Stat3
(1833S3Sh) compared to CM from control 1833 cells (Fig. 5E).
Taken together, these data suggest that activated Stat3 upregulates
ATX expression which leads to an increase in cell migration.
Conversely, inhibition of Stat3 leads to a reduction in the
expression of ATX and a corresponding decrease in cell motility.
Stat3 is persistently tyrosine phosphorylated in a large number
of tumors of epithelial origin including breast cancer. However, its
precise role in breast tumorigenesis has been an area of significant
investigation. Examination of the distribution and incidence of
pStat3 levels by immunohistochemical analysis and correlating
these data with prognostic features has led to seemingly
contradictory results. Specifically, elevated levels of phosphorylat-
ed Stat3 in tumor samples from 45 patients with high-risk breast
cancer was correlated with an incomplete response to neo-
adjuvant chemotherapy (i.e. poor prognostic feature) .
However, analysis of tissue microarrays of 346 node-negative
patients revealed that postive phosphorylated Stat3 levels was a
prognostic marker of improved overall survival . A number of
reasons could account for this apparent discrepancy including: 1.
The stage of the disease StageIII  versus StageI  2.
Differences in pStat3 levels in fresh samples  versus archived
Figure 2. Tyrosine phosphorylated Stat3 levels are greater
along the leading edge of tumors. IHC for pStat3 was performed
on the primary tumor specimens (38) from which the cores for the TMA
were derived from. A. A representative example demonstrating
relatively high levels of pStat3 (2+) along the leading edge of the
tumor. B. IHC for pStat3 in the corresponding tissue microarray 0.6 mm
core for both the tumor and matched axillary lymph node.
ATX a Stat3 Target in Breast Cancer
PLoS ONE | www.plosone.org5 November 2011 | Volume 6 | Issue 11 | e27851
samples .1–25 years old . 3. The presence of pStat5, which
has been shown to be associated with more differentiated breast
tumors and with a more favorable prognosis. Furthermore, pStat5
can exert a dominant role over Stat3 . 4. The analysis of whole
tumor blocks  versus 1mm cores .
Our analysis of pStat3 levels of primary tumor sections initially
from tissue microarrays (TMAs) and subsequently from whole
tumor blocks revealed that 20% of those which were initially
scored as low for pStat3 (analysis of the TMA) were in fact positive
along the edge or rim of the tumor adjacent to stromal cells when
the corresponding tumor block was analyzed (Fig. 2). Further-
more, the highest pStat3 levels were almost always found on the
edge of tumor samples. Similar observations have been made in
non-small cell lung cancer samples where pStat3 levels were
concentrated in tumor cells adjacent to non-tumor stromal tissues
. This study is the first to examine the relative levels of pStat3
in primary breast tumors with matched axillary lymph nodes.
Here we show that a greater fraction of lymph nodes express
moderate/high levels of pStat3 compared to the corresponding
primary tumors (Fig. 1B). One explanation for this observation is
that cytokines such as IL-6 which mediate Stat3 activation in
breast cancers are expressed at higher levels in lymphatic tissue
resulting in increased pStat3. However, we did not observe high
IL-6 levels within lymphocytes in the affected lymph nodes, rather
high IL-6 levels were observed within the tumor epithelial cells
which correlated with pStat3 levels (manuscript submitted, JFB)
. An alternative explanation for the relative increase in pStat3
staining in lymph nodes versus primary tumor is that the pStat3
positive cells within primary breast tumors have an increased
migratory and invasive capacity compared to the pStat3 negative
expressing cells. Indeed it has been demonstrated that tumors and
tumor-derived cell lines are heterogeneous in nature leading to the
development of sub-clones with differential capacity to metastasize
to various organs [55,56,57,72].
Stat3 has increasingly been shown to play a role in migration
and invasion of both normal and cancer cells. Stat3 through
LIV1/Snail/Twist controls epithelial-mesenchymal transition in
zebrafish gastrula organizer [66,73]. Activated Stat3 is a direct
transcriptional regulator of Twist in breast cancer which is a
critical regulator of metastatic progression [16,29,74,75]. Condi-
tional deletion of Stat3 in the skin has revealed a requirement for it
in wound healing and tumorigenesis [76,77]. Stat3 knockdown in
murine models of ErbB2-mediated breast carcinoma altered
epithelial adhesion and polarity . Stat3 in conjunction with
NF-kB is a transcriptional regulator of the chemokine RANTES
which was recently shown to be a potent mediator of metastatic
spread of breast cancer cells [78,79]. We have shown that Stat3
regulates expression of MMP-9 in breast cancer and integrin ß6 in
prostate cancer which are critical regulators of invasion and
migration [26,80]. In order to identify additional Stat3 targets
Figure 3. Hierarchical Cluster Analysis by pStat3. A. Two-way hierarchical clustering was performed with the 21 ER2 pStat3 (+/2) samples
using the gene expression values limited to the 214 differentially expressed genes. The region of the dendrogram including ENPP2 is enlarged. B. The
same analysis as described in A was performed with the 41 ER2 samples. Several regions of the dendrogram including ENPP2 were enlarged C. Two-
way hierarchical clustering was performed with all (98) of the tumor specimens categorized by pStat3 status (+/2) using the 214 differentially
expressed genes. The dendogram represents the relationship among samples, and the length of branches represents 1 minus the pearson correlation
coefficient between two samples. Group 1 are primarily pStat3 negative and Group 2 are pStat3 positive. Samples are arranged in columns and genes
in rows. Normalized expression levels are pseudocolored red to indicate transcript levels above the median for that gene across samples, and green
for expression below the median. Color saturation is proportional to the magnitude of expression. Several clusters of differentially expressed genes
including ENPP2 are enlarged. ER and pStat3 status are indicated for each sample (grey versus white, orange versus blue).
ATX a Stat3 Target in Breast Cancer
PLoS ONE | www.plosone.org6November 2011 | Volume 6 | Issue 11 | e27851
responsible for metastatic spread, we used gene expression
profiling of primary breast tumors . We compared the gene
expression profiles of samples expressing moderate/high levels of
pStat3 within carcinoma cells to those with no activated Stat3 in
either carcinoma cells, lymphocytes, stromal cells and endothelial
cells. In addition to the ENPP2/ATX gene, a number of previously
Figure 4. ATX in primary breast tumors and matched axillary lymph nodes positively correlate with pStat3. A. TMAs of 41 ER2 breast
tumors were analyzed for cytoplasmic ATX levels by IHC analysis. Representative tumors of no detectable ATX (0), low levels (+1), moderate levels (+2)
and high levels (+3) are shown. 49% expressed moderate to high levels of ATX (2–3+) whereas 51% expressed low levels of ATX. B. A positive
correlation between (2–3+) pStat3 and (2–3+) ATX levels was observed in the primary tumors (p=0.0025). Examples of both low (0) and high (3+) ATX
with corresponding serial sections for pStat3 are shown. C. A positive correlation between (2–3+) pStat3 and (2–3+) ATX levels was observed in the
38 lymph nodes described in Fig. 1 (p=0.0008). Examples of both low (0) and high (3+) ATX with corresponding serial sections for pStat3 are shown.
D. A representative example of serial sections of a tumor expressing moderate pStat3 and ATX levels on the tumor edge.
ATX a Stat3 Target in Breast Cancer
PLoS ONE | www.plosone.org7November 2011 | Volume 6 | Issue 11 | e27851
identified Stat3 targets were identified using this approach
including including Twist, Snail, Tenascin-C, IL-8 and E-Cadherin
[29,62,63,64,65,66,67,68,69,80]. We have observed an inverse
correlation between pStat3 and E-Cadherin in primary breast
tumors (manuscript under review, JFB). Phoshorylated Stat3 and
Vimentin (a marker of EMT) increased in 435 cells treated with
OSM, while P6 treatment of 1833 cells led to a reduction of both
pStat3 and Vimentin (Fig. 5B). It is interesting that genes over-
expressed in triple negative breast cancers and in highly metastatic
cancer derived cell lines express a number of these same
ATX or ENPP-2 is a secreted lysophospholipase which mediates
production of LPA, a stimulator of breast cancer migration and
whose expression is associated with mammary tumorigenesis, breast
cancer metastasis, angiogenesis and survival [31,38,43,49,83]. MDA-
MB-435 cells express moderate levels of ATX and pStat3, yet its
Figure 5. Stat3 regulates ATX expression and cell migration in breast cancer cells. A. Extracts (20 mg) isolated from MCF7, BT474 (BT),
MDA-MB-231 (231), 1833, 4175 and MDA-MB-435 (435) were analyzed for pStat3, Stat3, ATX and Tubulin by western blot analysis. Total RNA was
isolated from the same cell lines and analyzed for ATX by RT/PCR and normalized to Actin. B. Whole cell extracts (50 mg) isolated from MDA-MB-435
cells treated for 4 hours with dimethyl sulfoxide (C), OSM (5 ng/ml) and P6 (1 mM); 1833 cells treated with DMSO or P6 and 1833 Control Sh (CSh) or
expressing a Stat3 Sh (S3Sh) were analyzed for pStat3, Stat3, Vimentin, ATX and Tubulin by western blot analysis. C. Total RNA isolated from MDA-MB-
435 cells treated for 4 hours with dimethyl sulfoxide C, OSM and P6; and 1833 cells treated with DMSO (C) or P6 was analyzed for ATX by RT/PCR and
normalized to Actin. D. Potential (TTN5AA) Stat3 binding sites are indicated by inverted triangles on the ATX promoter. Arrows indicate the direction
of transcription and nucleotide positions are shown below the diagram. MDA-MB-435 cells treated for 30 minutes with dimethyl sulfoxide (C), OSM
(5 ng/ml) and P6 (1 mM) were subjected to chromatin immunoprecipitation assay using antibodies to to Stat3 (S3) or IgG as a negative control.
Similarly, 1833 Control Sh (1833CSh) or 1833 cells expressing a Stat3Sh (1833S3Sh) were were subjected to chromatin immunoprecipitation assay
using antibodies to Stat3 (S3) or IgG as a negative control. Co-precipitated DNA was amplified by PCR using primers flanking the Stat3 binding sites at
2821 and 2301. The input was 5% of the total. E. Migration of MDA-MB-435 cells in a Boyden chamber assay was determined in the presence of
conditioned media from dimethyl sulfoxide (-), OSM (5 ng/ml) or P6 (1 mM) treated MDA-MB-435 cells. Migration of 1833 cells was determined in a
Boyden chamber assay in the presence of conditioned media from 1833 cells from dimethyl sulfoxide (-) and P6 treated 1833 cells. Migration of 1833
Control Sh (CSh) and 1833 cells expressing a Stat3Sh (S3Sh) was determined in a Boyden chamber assay with conditioned media in the bottom
chamber from the same corresponding cell lines. Results are expressed as the number of cells per field (as determined by crystal violet staining) and
are the mean SD of triplicate values from three independent experiments.
ATX a Stat3 Target in Breast Cancer
PLoS ONE | www.plosone.org8 November 2011 | Volume 6 | Issue 11 | e27851
expression is markedly increased following the addition of a gp130
ligand leading to increased pStat3 levels, which resulted in an
increased chemotactic potential (Fig. 5). There has been some
controversyabout the origin of MDA-MB-435 cellswith some studies
suggesting it may be of melanoma origin . However, a recent
study suggests that the manner in which MDA-MB-435 cells are
propogated can lead to differential expression of epithelial and breast
specific markers as well as melanocyte-specific markers suggesting
lineage infidelity[85,86].Inlight of thiscontroversyweexpanded our
cell line analysis to others (Fig. 5). It was recently reported that
NFAT1 could transcriptionally regulate the ATX gene in integrin
a6ß4 transfected MDA-MB-435 cells . There is an NFAT1
binding site at 2327 (27 bp away from a Stat3 binding site).
However, no known synergistic associations between these two
transcription factors have been described. In addition, v-Jun
expression in chick embryo fibroblasts led to an increase in ATX
mRNA levels, though a direct transcriptional mechanism was not
determined . Stat3 has been shown to interact with the AP-1
proteins c-jun and c-fos, however no ideal AP-1 binding sites were
identified within the ATX promoter region . Although our data
demonstrates a direct interaction between Stat3 and the ATX
promoter, other Stat3 targets or upregulated transcripts (VEGF,
NRP1 and CD36) may be participating in the transcriptional
regulation of ATX [54,89]. In summary we have determined that
activated Stat3 expression is heterogeneous in distribution in primary
breast tumors with the highest levels found on the tumor edge.
Furthermore, a relative increase in pStat3 staining in matched
axillary lymph nodes versus primary tumor suggesting that pStat3
positive cells have an increased capacity to metastasize to lymph
nodes. Finally, using gene expression profiling of primary tumors, we
identified ENPP2/ATX as a novel Stat3 target gene involved in cell
migration and invasion.
negative breast cancer cells. A. Extracts (20 mg) isolated from
1937, 1806, 1143, 38 and MD-MB-231 cell lineswere analyzed for
pStat3, ATX and Actin by western blot analysis. B. Extracts (20 mg)
isolated from MCF7, BT474 (BT), MDA-MB-231 (231), 1833, 4175
and MDA-MB-435 (435) were analyzed for pStat3, Stat3, pStat5,
(mg) isolated from MDA-MB-435 cells treated for 4 hours with
dimethyl sulfoxide (C), OSM (5 ng/ml) and P6 (1 mM); 1833 cells
Stat3 regulation of ATX expression in triple
pStat5, Stat5, ATX and Tubulin by western blot analysis.
breast cancer as a function of pStat3. Amicroarraystatistical
analysis of the 8 pStat3+ versus the 13 pStat32 tumor specimens
identified 214 differentially expressed genes according to pStat3
status (at least two-fold between the means of pStat3(+) and
pStat3(2) cases and a Student’s t-test P,0.05). Of the 214
differentially expressed genes, 150 genes were over-expressed and
64 genes were under-expressed in pStat3(+) cases relative to
Differential gene expression in primary ER2 2
cancer asa functionofpStat3. A microarray statistical analysis
of 54 pStat32 versus 45 pStat3+ tumor specimens resulted in
identifying 136 genes which were differentially expressed accord-
ing to pStat3 status (at least 1.5 fold between the means of
pStat3(+) and pStat3(2) cases and a Student’s t-test P,0.05). Of
the 136 differentially expressed genes, 115 were over-expressed
and 21 were under-expressed in pStat3(+) cases relative to
Differential gene expression in primary breast
as a function of pStat3. A comparison of differentially
expressed genes in ER2 breast cancer (214 genes) versus all
(ER2 and ER+) breast cancers (136 genes) as a function of pStat3
revealed 50 genes overlapping including ENPP2.
Differential gene expression in breast cancer
We are grateful for the expert assistance from Marina Asher for
Conceived and designed the experiments: JA AD JN DL JB. Performed the
experiments: JA AD KL JN QC MB KM HA MH AV. Analyzed the data:
JA AD KL JN QC MB KM HA WG MH AV MS DL JB. Contributed
reagents/materials/analysis tools: WG MH AV MS. Wrote the paper: JA
AD DL JB.
1. Parkin DM, Bray F, Ferlay J, Pisani P (2005) Global cancer statistics, 2002. CA
Cancer J Clin 55: 74–108.
2. Pantel K, Brakenhoff RH (2004) Dissecting the metastatic cascade. Nat Rev
Cancer 4: 448–456.
3. Cristofanilli M, Budd GT, Ellis MJ, Stopeck A, Matera J, et al. (2004)
Circulating tumor cells, disease progression, and survival in metastatic breast
cancer. N Engl J Med 351: 781–791.
4. Braun S, Vogl FD, Naume B, Janni W, Osborne MP, et al. (2005) A pooled
analysis of bone marrow micrometastasis in breast cancer. N Engl J Med 353:
5. Muller V, Stahmann N, Riethdorf S, Rau T, Zabel T, et al. (2005) Circulating
tumor cells in breast cancer: correlation to bone marrow micrometastases,
heterogeneous response to systemic therapy and low proliferative activity. Clin
Cancer Res 11: 3678–3685.
6. Bromberg J (2002) Stat proteins and oncogenesis. J Clin Invest 109: 1139–1142.
7. Ranger JJ, Levy DE, Shahalizadeh S, Hallett M, Muller WJ (2009) Identification
of a Stat3-dependent transcription regulatory network involved in metastatic
progression. Cancer Res 69: 6823–6830.
8. Garcia R, Yu CL, Hudnall A, Catlett R, Nelson KL, et al. (1997) Constitutive
activation of Stat3 in fibroblasts transformed by diverse oncoproteins and in
breast carcinoma cells. Cell Growth Differ 8: 1267–1276.
9. Garcia R, Bowman TL, Niu G, Yu H, Minton S, et al. (2001) Constitutive
activation of Stat3 by the Src and JAK tyrosine kinases participates in growth
regulation of human breast carcinoma cells. Oncogene 20: 2499–2513.
10. Turkson J, Kim JS, Zhang S, Yuan J, Huang M, et al. (2004) Novel
peptidomimetic inhibitors of signal transducer and activator of transcription 3
dimerization and biological activity. Mol Cancer Ther 3: 261–269.
11. Gritsko T, Williams A, Turkson J, Kaneko S, Bowman T, et al. (2006) Persistent
activation of stat3 signaling induces survivin gene expression and confers
resistance to apoptosis in human breast cancer cells. Clin Cancer Res 12: 11–19.
12. Kotha A, Sekharam M, Cilenti L, Siddiquee K, Khaled A, et al. (2006) Resveratrol
inhibits Src and Stat3 signaling and induces the apoptosis of malignant cells
containing activated Stat3 protein. Mol Cancer Ther 5: 621–629.
13. Leslie K, Lang C, Devgan G, Azare J, Berishaj M, et al. (2006) Cyclin D1 is
transcriptionally regulated by and required for transformation by activated
signal transducer and activator of transcription 3. Cancer Res 66: 2544–2552.
14. Selander KS, Li L, Watson L, Merrell M, Dahmen H, et al. (2004) Inhibition of
gp130 signaling in breast cancer blocks constitutive activation of Stat3 and
inhibits in vivo malignancy. Cancer Res 64: 6924–6933.
15. Alvarez JV, Febbo PG, Ramaswamy S, Loda M, Richardson A, et al. (2005)
Identification of a genetic signature of activated signal transducer and activator
of transcription 3 in human tumors. Cancer Res 65: 5054–5062.
ATX a Stat3 Target in Breast Cancer
PLoS ONE | www.plosone.org9November 2011 | Volume 6 | Issue 11 | e27851
16. Ling X, Arlinghaus RB (2005) Knockdown of STAT3 expression by RNA
interference inhibits the induction of breast tumors in immunocompetent mice.
Cancer Res 65: 2532–2536.
17. Guo W, Pylayeva Y, Pepe A, Yoshioka T, Muller WJ, et al. (2006) Beta 4
integrin amplifies ErbB2 signaling to promote mammary tumorigenesis. Cell
18. Hedvat M, Huszar D, Herrmann A, Gozgit JM, Schroeder A, et al. (2009) The
JAK2 inhibitor AZD1480 potently blocks Stat3 signaling and oncogenesis in
solid tumors. Cancer Cell 16: 487–497.
19. Walker SR, Nelson EA, Zou L, Chaudhury M, Signoretti S, et al. (2009)
Reciprocal effects of STAT5 and STAT3 in breast cancer. Mol Cancer Res 7:
20. Diaz N, Minton S, Cox C, Bowman T, Gritsko T, et al. (2006) Activation of
stat3 in primary tumors from high-risk breast cancer patients is associated with
elevated levels of activated SRC and survivin expression. Clin Cancer Res 12:
21. Turkson J, Zhang S, Palmer J, Kay H, Stanko J, et al. (2004) Inhibition of
constitutive signal transducer and activator of transcription 3 activation by novel
platinum complexes with potent antitumor activity. Mol Cancer Ther 3:
22. Gariboldi MB, Ravizza R, Molteni R, Osella D, Gabano E, et al. (2007)
Inhibition of Stat3 increases doxorubicin sensitivity in a human metastatic breast
cancer cell line. Cancer Lett 258: 181–188.
23. Berishaj M, Gao SP, Ahmed S, Leslie K, Al-Ahmadie H, et al. (2007) Stat3 is
tyrosine-phosphorylated through the interleukin-6/glycoprotein 130/Janus
kinase pathway in breast cancer. Breast Cancer Res 9: R32.
24. Sansone P, Storci G, Tavolari S, Guarnieri T, Giovannini C, et al. (2007) IL-6
triggers malignant features in mammospheres from human ductal breast
carcinoma and normal mammary gland. J Clin Invest 117: 3988–4002.
25. Zhou J, Wulfkuhle J, Zhang H, Gu P, Yang Y, et al. (2007) Activation of the
PTEN/mTOR/STAT3 pathway in breast cancer stem-like cells is required for
viability and maintenance. Proc Natl Acad Sci U S A 104: 16158–16163.
26. Dechow TN, Pedranzini L, Leitch A, Leslie K, Gerald WL, et al. (2004)
Requirement of matrix metalloproteinase-9 for the transformation of human
mammary epithelial cells by Stat3-C. Proc Natl Acad Sci U S A 101:
27. Niu G, Wright KL, Huang M, Song L, Haura E, et al. (2002) Constitutive Stat3
activity up-regulates VEGF expression and tumor angiogenesis. Oncogene 21:
28. Hsieh FC, Cheng G, Lin J (2005) Evaluation of potential Stat3-regulated genes
in human breast cancer. Biochem Biophys Res Commun 335: 292–299.
29. Lo HW, Hsu SC, Xia W, Cao X, Shih JY, et al. (2007) Epidermal growth factor
receptor cooperates with signal transducer and activator of transcription 3 to
induce epithelial-mesenchymal transition in cancer cells via up-regulation of
TWIST gene expression. Cancer Res 67: 9066–9076.
30. Clair T, Lee HY, Liotta LA, Stracke ML (1997) Autotaxin is an exoenzyme
possessing 59-nucleotide phosphodiesterase/ATP pyrophosphatase and ATPase
activities. J Biol Chem 272: 996–1001.
31. Nam SW, Clair T, Kim YS, McMarlin A, Schiffmann E, et al. (2001) Autotaxin
(NPP-2), a metastasis-enhancing motogen, is an angiogenic factor. Cancer Res
32. Stracke ML, Clair T, Liotta LA (1997) Autotaxin, tumor motility-stimulating
exophosphodiesterase. Adv Enzyme Regul 37: 135–144.
33. Chen M, O’Connor KL (2005) Integrin alpha6beta4 promotes expression of
autotaxin/ENPP2 autocrine motility factor in breast carcinoma cells. Oncogene
34. Stracke M, Liotta LA, Schiffmann E (1993) The role of autotaxin and other
motility stimulating factors in the regulation of tumor cell motility. Symp Soc
Exp Biol 47: 197–214.
35. van Meeteren LA, Moolenaar WH (2007) Regulation and biological activities of
the autotaxin-LPA axis. Prog Lipid Res 46: 145–160.
36. Cooper AB, Wu J, Lu D, Maluccio MA (2007) Is Autotaxin (ENPP2) the Link
between Hepatitis C and Hepatocellular Cancer? J Gastrointest Surg 11:
37. Kishi Y, Okudaira S, Tanaka M, Hama K, Shida D, et al. (2006) Autotaxin is
overexpressed in glioblastoma multiforme and contributes to cell motility of
glioblastoma by converting lysophosphatidylcholine to lysophosphatidic acid.
J Biol Chem 281: 17492–17500.
38. Noh JH, Ryu SY, Eun JW, Song J, Ahn YM, et al. (2006) Identification of large-
scale molecular changes of Autotaxin(ENPP2) knock-down by small interfering
RNA in breast cancer cells. Mol Cell Biochem 288: 91–106.
39. Hoelzinger DB, Mariani L, Weis J, Woyke T, Berens TJ, et al. (2005) Gene
expression profile of glioblastoma multiforme invasive phenotype points to new
therapeutic targets. Neoplasia 7: 7–16.
40. Yano Y, Hayashi Y, Sano K, Shinmaru H, Kuroda Y, et al. (2003) Expression
and localization of ecto-nucleotide pyrophosphatase/phosphodiesterase I-3 (E-
NPP3/CD203c/PD-I beta/B10/gp130RB13-6) in human colon carcinoma.
Int J Mol Med 12: 763–766.
41. Kehlen A, Englert N, Seifert A, Klonisch T, Dralle H, et al. (2004) Expression,
regulation and function of autotaxin in thyroid carcinomas. Int J Cancer 109:
42. Song J, Jie C, Polk P, Shridhar R, Clair T, et al. (2006) The candidate tumor
suppressor CST6 alters the gene expression profile of human breast carcinoma
cells: down-regulation of the potent mitogenic, motogenic, and angiogenic factor
autotaxin. Biochem Biophys Res Commun 340: 175–182.
43. Yang SY, Lee J, Park CG, Kim S, Hong S, et al. (2002) Expression of autotaxin
(NPP-2) is closely linked to invasiveness of breast cancer cells. Clin Exp
Metastasis 19: 603–608.
44. Lee HY, Bae GU, Jung ID, Lee JS, Kim YK, et al. (2002) Autotaxin promotes
motility via G protein-coupled phosphoinositide 3-kinase gamma in human
melanoma cells. FEBS Lett 515: 137–140.
45. Zhang G, Zhao Z, Xu S, Ni L, Wang X (1999) Expression of autotaxin mRNA
in human hepatocellular carcinoma. Chin Med J (Engl) 112: 330–332.
46. Yang Y, Mou L, Liu N, Tsao MS (1999) Autotaxin expression in non-small-cell
lung cancer. Am J Respir Cell Mol Biol 21: 216–222.
47. David M, Wannecq E, Descotes F, Jansen S, Deux B, et al. (2010) Cancer cell
expression of autotaxin controls bone metastasis formation in mouse through
lysophosphatidic acid-dependent activation of osteoclasts. PLoS One 5: e9741.
48. Jonkers J, Moolenaar WH (2009) Mammary tumorigenesis through LPA
receptor signaling. Cancer Cell 15: 457–459.
49. Liu S, Umezu-Goto M, Murph M, Lu Y, Liu W, et al. (2009) Expression of
autotaxin and lysophosphatidic acid receptors increases mammary tumorigen-
esis, invasion, and metastases. Cancer Cell 15: 539–550.
50. Doane AS, Danso M, Lal P, Donaton M, Zhang L, et al. (2006) An estrogen
receptor-negative breast cancer subset characterized by a hormonally regulated
transcriptional program and response to androgen. Oncogene 25: 3994–4008.
51. Giganti A, Rodriguez M, Fould B, Moulharat N, Coge F, et al. (2008) Murine
and human autotaxin alpha, beta, and gamma isoforms: gene organization,
tissue distribution, and biochemical characterization. J Biol Chem 283:
52. Murata J, Lee HY, Clair T, Krutzsch HC, Arestad AA, et al. (1994) cDNA
cloning of the human tumor motility-stimulating protein, autotaxin, reveals a
homology with phosphodiesterases. J Biol Chem 269: 30479–30484.
53. Stracke ML, Krutzsch HC, Unsworth EJ, Arestad A, Cioce V, et al. (1992)
Identification, purification, and partial sequence analysis of autotaxin, a novel
motility-stimulating protein. J Biol Chem 267: 2524–2529.
54. Ptaszynska MM, Pendrak ML, Bandle RW, Stracke ML, Roberts DD (2008)
Positive feedback between vascular endothelial growth factor-A and autotaxin in
ovarian cancer cells. Mol Cancer Res 6: 352–363.
55. Kang Y, Siegel PM, Shu W, Drobnjak M, Kakonen SM, et al. (2003) A
multigenic program mediating breast cancer metastasis to bone. Cancer Cell 3:
56. Minn AJ, Gupta GP, Siegel PM, Bos PD, Shu W, et al. (2005) Genes that
mediate breast cancer metastasis to lung. Nature 436: 518–524.
57. Minn AJ, Kang Y, Serganova I, Gupta GP, Giri DD, et al. (2005) Distinct
organ-specific metastatic potential of individual breast cancer cells and primary
tumors. J Clin Invest 115: 44–55.
58. Ho HH, Ivashkiv LB (2006) Role of STAT3 in type I interferon responses.
Negative regulation of STAT1-dependent inflammatory gene activation. J Biol
Chem 281: 14111–14118.
59. Watson CJ, Miller WR (1995) Elevated levels of members of the STAT family of
transcription factors in breast carcinoma nuclear extracts. Br J Cancer 71:
60. Cotarla I, Ren S, Zhang Y, Gehan E, Singh B, et al. (2004) Stat5a is tyrosine
phosphorylated and nuclear localized in a high proportion of human breast
cancers. Int J Cancer 108: 665–671.
61. Dolled-Filhart M, Camp RL, Kowalski DP, Smith BL, Rimm DL (2003) Tissue
Microarray Analysis of Signal Transducers and Activators of Transcription 3
(Stat3) and Phospho-Stat3 (Tyr705) in Node-negative Breast Cancer Shows
Nuclear Localization Is Associated with a Better Prognosis. Clin Cancer Res 9:
62. de la Iglesia N, Konopka G, Lim KL, Nutt CL, Bromberg JF, et al. (2008)
Deregulation of a STAT3-interleukin 8 signaling pathway promotes human
glioblastoma cell proliferation and invasiveness. J Neurosci 28: 5870–5878.
63. Oka M, Sakaguchi M, Okada T, Nagai H, Ozaki M, et al. (2010) Signal
transducer and activator of transcription 3 upregulates interleukin-8 expression
at the level of transcription in human melanoma cells. Exp Dermatol 19: e50–55.
64. Katoh M (2009) Transcriptional mechanisms of WNT5A based on NF-kappaB,
Hedgehog, TGFbeta, and Notch signaling cascades. Int J Mol Med 23:
65. Shen H, Qin H, Guo J (2009) Concordant correlation of LIV-1 and E-cadherin
expression in human breast cancer cell MCF-7. Mol Biol Rep 36: 653–659.
66. Yamashita S, Miyagi C, Fukada T, Kagara N, Che YS, et al. (2004) Zinc
transporter LIVI controls epithelial-mesenchymal transition in zebrafish gastrula
organizer. Nature 429: 298–302.
67. Cheng GZ, Zhang WZ, Sun M, Wang Q, Coppola D, et al. (2008) Twist is
transcriptionally induced by activation of STAT3 and mediates STAT3
oncogenic function. J Biol Chem 283: 14665–14673.
68. Sullivan NJ, Sasser AK, Axel AE, Vesuna F, Raman V, et al. (2009) Interleukin-
6 induces an epithelial-mesenchymal transition phenotype in human breast
cancer cells. Oncogene 28: 2940–2947.
69. Leslie K, Gao SP, Berishaj M, Podsypanina K, Ho H, et al. (2010) Differential
interleukin-6/Stat3 signaling as a function of cellular context mediates Ras-
induced transformation. Breast Cancer Res 12: R80.
70. Pamuklar Z, Federico L, Liu S, Umezu-Goto M, Dong A, et al. (2009)
Autotaxin/lysopholipase D and lysophosphatidic acid regulate murine hemo-
stasis and thrombosis. J Biol Chem 284: 7385–7394.
ATX a Stat3 Target in Breast Cancer
PLoS ONE | www.plosone.org 10November 2011 | Volume 6 | Issue 11 | e27851
71. Dauer DJ, Ferraro B, Song L, Yu B, Mora L, et al. (2005) Stat3 regulates genes Download full-text
common to both wound healing and cancer. Oncogene 24: 3397–3408.
72. Gupta GP, Minn AJ, Kang Y, Siegel PM, Serganova I, et al. (2005) Identifying
site-specific metastasis genes and functions. Cold Spring Harb Symp Quant Biol
73. Yamashita S, Miyagi C, Carmany-Rampey A, Shimizu T, Fujii R, et al. (2002)
Stat3 Controls Cell Movements during Zebrafish Gastrulation. Dev Cell 2:
74. Ma L, Teruya-Feldstein J, Weinberg RA (2007) Tumour invasion and metastasis
initiated by microRNA-10b in breast cancer. Nature 449: 682–688.
75. Yang J, Mani SA, Weinberg RA (2006) Exploring a new twist on tumor
metastasis. Cancer Res 66: 4549–4552.
76. Chan KS, Sano S, Kiguchi K, Anders J, Komazawa N, et al. (2004) Disruption
of Stat3 reveals a critical role in both the initiation and the promotion stages of
epithelial carcinogenesis. J Clin Invest 114: 720–728.
77. Sano S, Itami S, Takeda K, Tarutani M, Yamaguchi Y, et al. (1999)
Keratinocyte-specific ablation of stat3 exhibits impaired skin remodeling, but
does not affect skin morphogenesis [In Process Citation]. Embo J 18:
78. Karnoub AE, Dash AB, Vo AP, Sullivan A, Brooks MW, et al. (2007)
Mesenchymal stem cells within tumour stroma promote breast cancer
metastasis. Nature 449: 557–563.
79. Yang J, Liao X, Agarwal MK, Barnes L, Auron PE, et al. (2007) Unpho-
sphorylated STAT3 accumulates in response to IL-6 and activates transcription
by binding to NFkappaB. Genes Dev 21: 1396–1408.
80. Azare J, Leslie K, Al-Ahmadie H, Gerald W, Weinreb PH, et al. (2007)
Constitutively activated Stat3 induces tumorigenesis and enhances cell motility
of prostate epithelial cells through integrin beta 6. Mol Cell Biol 27: 4444–4453.
81. Kim MY, Oskarsson T, Acharyya S, Nguyen DX, Zhang XH, et al. (2009)
Tumor self-seeding by circulating cancer cells. Cell 139: 1315–1326.
82. Montserrat N, Gallardo A, Escuin D, Catasus L, Prat J, et al. (2011) Repression
of E-cadherin by SNAIL, ZEB1, and TWIST in invasive ductal carcinomas of
the breast: a cooperative effort? Hum Pathol 42: 103–110.
83. Mills GB, Moolenaar WH (2003) The emerging role of lysophosphatidic acid in
cancer. Nat Rev Cancer 3: 582–591.
84. Rae JM, Creighton CJ, Meck JM, Haddad BR, Johnson MD (2007) MDA-MB-
435 cells are derived from M14 melanoma cells–a loss for breast cancer, but a
boon for melanoma research. Breast Cancer Res Treat 104: 13–19.
85. Sellappan S, Grijalva R, Zhou X, Yang W, Eli MB, et al. (2004) Lineage
infidelity of MDA-MB-435 cells: expression of melanocyte proteins in a breast
cancer cell line. Cancer Res 64: 3479–3485.
86. Chambers AF (2009) MDA-MB-435 and M14 cell lines: identical but not M14
melanoma? Cancer Res 69: 5292–5293.
87. Black EJ, Clair T, Delrow J, Neiman P, Gillespie DA (2004) Microarray analysis
identifies Autotaxin, a tumour cell motility and angiogenic factor with
lysophospholipase D activity, as a specific target of cell transformation by v-
Jun. Oncogene 23: 2357–2366.
88. Ginsberg M, Czeko E, Muller P, Ren Z, Chen X, et al. (2007) Amino acid
residues required for physical and cooperative transcriptional interaction of
STAT3 and AP-1 proteins c-Jun and c-Fos. Mol Cell Biol 27: 6300–6308.
89. Geretti E, Shimizu A, Klagsbrun M (2008) Neuropilin structure governs VEGF
and semaphorin binding and regulates angiogenesis. Angiogenesis 11: 31–39.
ATX a Stat3 Target in Breast Cancer
PLoS ONE | www.plosone.org11November 2011 | Volume 6 | Issue 11 | e27851