17b-Estradiol upregulates and activates WOX1/WWOXv1 and
WOX2/WWOXv2 in vitro: potential role in cancerous progression of
breast and prostate to a premetastatic state in vivo
Nan-Shan Chang*,1, Lori Schultz1, Li-Jin Hsu1, Jennifer Lewis1, Meng Su2and Chun-I Sze2
1Guthrie Research Institute, Laboratory of Molecular Immunology, 1 Guthrie Square, Sayre, PA 18840, USA;2Department of
Pathology, University of Colorado Health Sciences Center, Denver, CO 80262, USA
Human WWOX gene encodes a proapoptotic WW domain-
containing oxidoreductase WOX1 (also named WWOX,
FOR2 or WWOXv1). Apoptotic and stress stimuli activate
WOX1 via Tyr33 phosphorylation and nuclear transloca-
tion. WOX1 possesses a tetrad NSYK motif in the C-
(SDR) domain, which may bind estrogen and androgen.
Here, we determined that 17b-estradiol (E2) activated
WOX1, p53 and ERK in COS7 fibroblasts, primary lung
epithelial cells, and androgen receptor (AR)-negative
prostate DU145 cells, but not in estrogen receptor (ER)-
positive breast MCF7 cells. Androgen also activated WOX1
in the AR-negative DU145 cells. These observations suggest
that sex hormone-mediated Tyr33 phosphorylation and
nuclear translocation of WOX1 is independent of ER and
AR. Stress stimuli increase physical binding of p53 with
WOX1 in vivo. We determined here that E2increased the
formation of p53/WOX1 complex and their nuclear
translocation in COS7 cells; however, nuclear translocation
of this complex could not occur in MCF7 cells. By
immunohistochemistry, we determined that progression of
prostate from normal to hyperplasia, cancerous and
metastatic stages positively correlate with upregulation
and activation of WOX1 and WOX2 (FOR1/WWOXv2).
In contrast, breast cancer development to a premetastatic
state is associated with upregulation and Tyr33 phosphor-
ylation of cytosolic WOX1 and WOX2, followed by
significant downregulation or absent expression during
metastasis. These Tyr33-phosphorylated proteins are mostly
located in the mitochondria without translocating to the
nuclei, which is comparable to those findings in cultured
breast cancer cells. Together, sex steroid hormone-induced
activation of WOX1 and WOX2 is independent of ER and
AR, and this activation positively correlates with cancerous
progression of prostate and breast to a premetastatic state.
Oncogene (2005) 24, 714–723. doi:10.1038/sj.onc.1208124
Published online 6 December 2004
Keywords: WWOX; FOR2; WOX1; WOX2; estrogen;
WW domain-containing oxidoreductase WOX1, also
known as WWOX or FOR2 (WWOXv1, GenBank’s
nomenclature), was independently isolated by three
laboratories (Bednarek et al., 2000; Ried et al., 2000;
Chang et al., 2001). WWOX gene encodes this protein.
WOX1 has been shown to be a candidate tumor
suppressor and a proapoptotic protein (Richards,
2001; Chang, 2002a; Chang et al., 2003b; reviews). The
full-length WOX1 possesses two N-terminal WW
domains (containing conserved tryptophan residues), a
nuclear localization sequence (NLS), and a C-terminal
short-chain alcohol dehydrogenase/reductase (SDR)
domain (Chang et al., 2003b). There are at least eight
alternatively transcribed mRNAs of WWOX gene
(Chang et al., 2003b). Whether these transcripts can be
translated successfully into proteins are unknown.
Murine WOX1 was isolated by functional cloning
from hyaluronidase PH-20-treated L929 fibroblasts and
shown to enhance the cytotoxic function of tumor
necrosis factor (TNF) (Chang et al., 2001). Hyaluroni-
dases Hyal-1 and Hyal-2 also induce the expression and
enhance the apoptotic function of WOX1 (Chang,
2002b). WOX1 physically interacts with p53 and both
proteins induce cell death synergistically (Chang et al.,
2001). We have determined that antisense WOX1
mRNA and a dominant negative WOX1 inhibit p53
apoptosis, indicating an important functional relation-
ship for p53 and WOX1 (Chang et al., 2001, 2003a).
This relationship has been further confirmed using
siRNA-targeting WOX1 (Chang et al., unpublished).
Most recently, we determined that UV and anisomycin
induce WOX1 activation, via phosphorylation at
tyrosine 33 (Tyr33) and nuclear translocation (Chang
et al., 2003a). Abrogation of Tyr33 phosphorylation by
mutating this amino-acid residue inhibits WOX1 nucle-
ar translocation (Chang et al., 2003a). Similarly,
alteration of charged amino acid residues in the nuclear
localization sequence (NLS) results in inhibition of
WOX1 nuclear translocation and its-mediated apoptosis
(Chang et al., 2001, 2003a).
WOX1 belongs to the SDR-containing protein family,
which has more than 2000 protein members (Kallberg
et al., 2002; review). Majority of these SDR enzymes
Received 15 March 2004; revised 23 August 2004; accepted 23 August
2004; published online 6 December 2004
*Correspondence: N-S Chang; E-mail: email@example.com
Oncogene (2005) 24, 714–723
& 2005 Nature Publishing Group All rights reserved 0950-9232/05 $30.00
possess a catalytic tetrad of NSYK (Asn–Ser–Tyr–Lys)
(Filling et al., 2002). SDR proteins mediate oxidation/
reduction of lipid hormones and metabolic mediators
(Kallberg et al., 2002). Based on computer simulation,
this NSYK motif in human WOX1 is determined to be
N232, S281, Y293, and K297 (Chang et al., 2003a).
WOX1 and its isoform WOX2 (FOR1/WWOXv2) are
likely to bind sex hormones such as androgen and
estrogen via its NSYK motif.
17b-estradiol (E2) induces Tyr33 phosphorylation and
nuclear translocation of WOX1 and WOX2 in COS7
We determined whether androgen and estrogen activate
WOX1. When COS7 fibroblasts were exposed to various
amounts of 17b-estradiol (E2, 10–160nM) for 1h, the level
of WOX1 was reduced in the cytosol, but increased in the
nuclei (Figure 1a), indicating that E2 induces WOX1
nuclear translocation. Similarly, the extent of Tyr33
phosphorylation in WOX1 was reduced in the cytosol,
but increased in the nucleus (Figure 1a). Immunostaining
further verified the presence of WOX1 in the nuclei of E2-
stimulated cells (Figure 1a). In parallel, there were
increased levels of phosphorylated p53 (at Ser15) and
extracellular regulated kinase (ERK; at Tyr204) in the
nuclei (Figure 1a). E2 is known to activate ERK
(Keshamouni et al., 2002; Setalo et al., 2002).
In a time course experiment, E2 rapidly induced
nuclear translocation of WOX1 and its isoform WOX2
(FOR1), during exposure of COS7 cells to E2for 20min
(Figure 1a). COS7 is an SV40-transformed cell line,
expressing wild-type p53. Similarly, in nontransformed
primary mink lung epithelial Mv1Lu cells, E2stimulated
phosphorylation and nuclear translocation of both
endogenous WOX1 and p53 in a time-dependent
manner (Figure 1b). Note that compared to the levels
of WOX1, p53 is overexpressed in COS7 cells, but not in
Next, we tested the specificity of E2 in activating
WOX1. Both 6-ketoestradiol and E2had similar effects
on inducing WOX1 activation (data not shown).
Compared to 6-ketoestradiol and E2, estriol had less
effect in activating WOX1 (B50% reduction in nuclear
WOX1 phosphorylation). The results indicate that ‘diol’
in estrogen is critical in activating WOX1.
The quality of our newly developed antibodies against
GST-tagged WOX1, along with our previously de-
scribed antibodies against WOX2 (Sze et al., 2004) and
Tyr33-phosphorylated WOX1 (p-WOX1) (Chang et al.,
2003a), is shown in the Supplementary Information
(Supplementary Figure 1).
Dominant-negative WOX1 abolishes E2-mediated WOX1
activation in COS7 cells
We have previously developed a dominant-negative
WOX1 (dn-WOX1) and shown to block p53 apoptosis
and anisomycin-induced WOX1 phosphorylation at
Tyr33 (Chang et al., 2003a). Transient expression of
GFP-tagged dn-WOX1 blocked E2-induced phosphor-
ylation and nuclear translocation of WOX1 and p53 in
COS7 cells (Figure 1c; data not shown for phosphoryla-
tion). Tyr33-phosphorylation is essential for WOX1
nuclear translocation (Chang et al., 2003a). Expression
of the ectopic GFP-dn-WOX1 or GFP alone in COS7
nuclear translocation of WOX1, p53 and ERK in COS7 fibroblasts
and Mv1Lu epithelial cells. (a) Stimulation of COS7 fibroblasts
with E2(10–160nM) for 1h resulted in activation of WOX1 via
Tyr33 phosphorylation (p-WOX1) and nuclear translocation. By
cell immunostaining, presence of Tyr33-phosphorylated WOX1 in
the nuclei was observed in E2-treated cells. Similarly, E2induced
phosphorylation of nuclear p53 (at Ser15) and ERK (at Tyr204). E2
(40nM) rapidly stimulated nuclear translocation of WOX1 and
WOX2 in COS7 cells in a time-dependent manner. (b) In a time-
course experiment, stimulation of primary mink lung epithelial
Mv1Lu cells with E2 (40nM) for various indicated times also
induced phosphorylation and nuclear translocation of both p53
and WOX1. (c) dn-WOX1 abolished E2-induced nuclear transloca-
tion of WOX1 and p53. COS7 cells were transfected with a dn-
WOX1 construct, GFP-dn-WOX1 (in pEGFP-C1) (Chang et al.,
2003a), or a control GFP construct (pEGFP-C1) by electropora-
tion. After culturing for 48h, the cells were stimulated with E2
(40nM) for 30min. E2induced nuclear translocation of endogenous
WOX1 in GFP-transfected cells, but not in the dn-WOX1-
transfected cells. Expression of the ectopic GFP-dn-WOX1 or
GFP in COS7 cells was observed by fluorescence microscopy
(B60% positive). Representative data were from two experiments
17b-estradiol (E2) stimulates phosphorylation and
17b-Estradiol activates WOX1/WWOXv1
N-S Chang et al
cells was observed by fluorescence microscopy (B60%
Upregulation and activation of WOX1 and family proteins
positively correlate with progression of prostate form
normal to hyperplasia, cancer and metastasis
Both sex steroid hormones and hyaluronidases play
important roles in the development of benign prostatic
hyperplasia and prostatic adenocarcinoma (Lokeshwar
and Rubinowicz, 1999; Lokeshwar et al., 2001; Patel
et al., 2002; Marker et al., 2003). Hyaluronidases PH-20,
Hyal-1 and Hyal-2 induce the expression of WOX1
(Chang et al., 2001; Chang, 2002b), suggesting a likely
role of WOX1 in prostatic cancer and prostatic
By immunohistochemistry, we examined protein
expression in breast and prostate tissue sections using
specific antibodies against (1) the conserved N-terminus
of WOX1 family proteins (Chang et al., 2001), (2) the
unique C-terminus of WOX1 (Sze et al., 2004), (3) the
unique C-terminus of WOX2 (Sze et al., 2004), and (4)
Tyr33 phosphorylation in the first WW domain (p-
WOX1) (Chang et al., 2003a).
During the progression of prostate from normal to
benign hyperplasia (hypertrophy),
(cancer) and metastasis, significant upregulation of
WOX1, WOX2 and family proteins (if present),
along with their Tyr33 phosphorylation and nuclear
Figure 2 in the Supplementary Information). Shown in
Figure 2 is the increased expression and Tyr33
phosphorylation of WOX1 family proteins (p-WOX1)
during prostate cancer progression to a metastatic
present mostly in perinuclear area. Nuclear transloca-
tion of p-WOX1 was shown at cancerous and metastatic
Similarly, breast cancer progression from normal to a
cancerous stage is also associated with upregulation of
WOX1, WOX2 and family proteins (see Supplementary
Figure 3a and b in Supplementary Information). These
observations are in agreement with other reports
(Driouch et al., 2002; Watanabe et al., 2003). However,
nuclear translocation of WOX1, WOX2 and Tyr33-
phosphorylated proteins did not occur effectively
(Figure 2, Table 1 and Supplementary Figure 3a
and b). Notably, significant downregulation or absent
1, and Supplementary
Increased expression, Tyr33 phosphorylation and nuclear translocation of WOX1 and family proteins (WWOXs) were observed during
progression of prostate from normal to hypertrophy, cancerous and metastatic stages. Tyr33 phosphorylation was also increased
during breast cancer development. However, nuclear translocation of p-WOX1 did not occur effectively. Significant downregulation of
p-WOX1 was observed in metastatic breast cells, whereas this event did not occur in metastatic prostate cancer cells; ?1000
magnification. In negative controls; nonimmune rabbit sera were used
Tyr33-phosphorylated WOX1 (p-WOX1) does not undergo nuclear translocation effectively in breast cancer tissues.
17b-Estradiol activates WOX1/WWOXv1
N-S Chang et al
expression of the WOX1 family proteins and their
Tyr33 phosphorylation was observed in metastatic
breast cancer cells (Figure 2 and Table 1; also see
Supplementary Figure 3a and b). A recent report also
showed that WWOX and FHIT genes are inactivated
coordinately in invasive breast carcinoma (Guler et al.,
By immunofluorescence, WOX1 family proteins are
shown to localize mainly in the mitochondria in the
normal and breast cancer cells (see Supplementary
Figure 3c). Also, WOX1 family proteins in the
mitochondria were Tyr33 phosphorylated (Figure 3),
suggesting that phosphorylation of WOX1 at Tyr33 is
essential for its translocation to the mitochondria.
Mitochondria were stained by using antibodies against
mitochondrial integral proteins such as Hsp60 (Supple-
mentary Figure 3c) and COX4 (cytochrome c oxidase
subunit 4) (Figure 3).
Summarized in Table 1 is the comparison of
the expression of WOX1, WOX2, and total family
proteins (WWOXs), along with their Tyr33 phosphor-
ylation and nuclear translocation, during the develop-
ment of breast and prostate cancers. Furthermore,
to confirm these observations, we determined the
extent of WOX1 expression in adenocarcinoma of
prostate, breast and lung by staining tissue micro-
array slides (from NCI). We determined that WOX1
expression was significantly increased (by 3–6-fold) in
these nonmetastatic adenocarcinoma tissues, compared
to normal subjects (30 cancer samples versus 10
E2stimulates WOX1 activation in androgen receptor-
negative DU145 cells
While the development of prostate and breast cancers
requires sex steroid hormones, we determined whether
E2 induces WOX1 activation in breast and prostate
cancer cells. Prostate DU145 cells are androgen-
resistant and negative for androgen receptors (AR)
(Wollin et al., 1989), and express mutant p53 (Isaacs
et al., 1991). These cells express low levels of mRNA for
estrogen receptor b (ER-b), whereas the protein levels of
ER-b are unknown (Lau et al., 2000; Linja et al., 2003).
Mutant p53 plays a role in generating hormone-
resistance and hormone-independent growth in prostate
LNCaP cancer cells (Burchardt et al., 2001; Nesslinger
Table 1 Localization and expression of WOX1, p-WOX1 and WOX2 in prostate and breast tissuesa
ControlHypertrophy CancerMetastatic cancer
CytosolicNuclear CytosolicNuclear CytosolicNuclear CytosolicNuclear
expression of WWOX/WOX1 family proteins (WWOXs) from each tissue section was determined by comparing to corresponding normal controls.
Prostate cases: control, n¼5; hypertrophy, n¼5; cancer, n¼5; metastatic cancer, n¼5; breast cases: control, n¼5; cancer, n¼5, metastatic
cancer, n¼5.bWWOXv1 and WWOXv2, GenBank’s nomenclature
aScale: ?/+, minimal expression; +, mild expression; ++, moderate expression; +++, abundant expression. Relative
chondria. By immunofluorescence, p-WOX1 (FITC; green) is
shown mainly in the mitochondria, which are located toward the
lumen of mammary gland (?1000). Mitochondria were stained
with antibodies against mitochondrial integral protein COX4
(cytochrome c oxidase subunit 4). Nuclei were stained with DAPI
Tyr33-phosphorylated WOX1 (p-WOX1) in the mito-
17b-Estradiol activates WOX1/WWOXv1
N-S Chang et al
et al., 2003). E2effectively reduced the cytosolic levels of
WOX1 and its Tyr33 phosphorylation in DU145 cells
(Figure 4a). Also, it increased the nuclear levels of
WOX1 and Tyr33 phosphorylation (Figure 4a), indicat-
ing that E2activates WOX1 in these cells. E2increased
nuclear translocation and phosphorylation of mutant
p53 in these cells (Figure 4a). The level of p53 is
relatively higher than that of WOX1 in these cells
(Figure 4a). E2also induced ERK phosphorylation in
both cytosolic and nuclear levels (Figure 4a). These
effects are similar to those observed in COS7 cells
Despite the absence of AR, androsterone, an andro-
gen, activated WOX1 in the nucleus of DU145 cells
(Figure 4b). Androsterone also induced phosphoryla-
tion of ERK, but not p53, at the nuclear level
Together, the above observations clearly demonstrate
that sex hormone-mediated WOX1 activation is inde-
pendent of hormone receptors.
E2could not activate WOX1 and WOX2 in ER-positive
Breast MCF7 cells are ER positive and express wild-
type p53 (Coezy et al., 1982; Gartel et al., 2003).
Notably, E2could not activate WOX1, WOX2 and p53
in MCF7, as determined in both Western blotting and
immunostaining (Figure 5a). The concentrations of E2
used in these experiments were increased, compared to
the above experiments. These observations are in
agreement with those findings that WOX1 and WOX2
did not translocate effectively to the nuclei in breast
cancer cells from patients (Figure 2 and Table 1).
Compared to that of WOX1, the p53 level was low in
MCF7 cells (Figure 5a).
Owing to alternative splicing, ER-negative breast
MDA-MB-231 cells do not express the full-length
mRNA coding for wild-type WOX1 (Driouch et al.,
2002). MDA-MB-435S cells express the full-length
WOX1 mRNA (Driouch et al., 2002). These cells are
metastatic and have a dysfunctional p53 (Toillon et al.,
2002; Gartel et al., 2003), and express very low levels of
wild-type WOX1 protein (Figure 5b). These findings are
in agreement with the observations from metastatic
breast specimens, which expressed little or no WOX1
(Figure 2). E2could not effectively induce activation of
WOX1 and ERK in these cells (Figure 5b). E2induced
phosphorylation of nuclear p53 in MDA-MB-231 but
not in MDA-MB-435S cells (Figure 5b). p53 levels in
prostate DU145 cells by E2 and androsterone. (a) Androgen-
resistant, AR-negative prostate DU145 cells were treated with E2
for 1h. E2reduced the cytosolic levels of WOX1, and increased
WOX1 nuclear translocation and phosphorylation. E2stimulated
phosphorylation of ERK and p53. (b) Androsterone increased
phosphorylation of nuclear WOX1 and ERK, but not p53, in
DU145 cells. Representative data were from two experiments
Stimulation of WOX1 activation in AR-negative
MCF7, MDA-MB-231 and MDA-MB-435S cells. (a) Breast MCF7
cells are ER positive and express wild-type p53. E2 could not
activate WOX1, WOX2 and p53 in MCF7, as determined in both
Western blotting and immunostaining (two representative pictures
in each time point). The concentrations of E2 used in this
experiments were increased. (b) ER-negative breast MDA-MB-
231 and MDA-MB-435S cells expressed low levels of WOX1, but
high levels of mutant p53. No apparent E2-mediated activation of
WOX1 and ERK was observed during 1h treatment in both cells,
using indicated concentrations or higher amounts as indicated
above (data not shown). E2induced phosphorylation of nuclear
p53 in MDA-MB-231 but not in MDA-MB-435S cells
Failure of E2in activating WOX1 and WOX2 in breast
17b-Estradiol activates WOX1/WWOXv1
N-S Chang et al
these cells were significantly elevated relative to those of
WOX1 (Figure 5b).
E2could not stimulate WOX1 activation in ER-negative
SK-N-SH neuroblastoma cells are ER-negative, but are
responsive to the neuroprotective effects of E2(Gangolli
et al., 1997; Green et al., 1997; Gridley et al., 1998).
Mechanism of this regard is unknown. E2increased the
levels of cytosolic WOX1 and p53, without inducing
nuclear translocation of these proteins (see Supplemen-
tary Figure 4 in Supplementary Information). Also, E2
could not induce Tyr33 phosphorylation of cytosolic
WOX1, but reduced nuclear WOX1 phosphorylation in
a dose-dependent manner. In agreement with other
report (Watters et al., 1997), E2induced ERK phos-
phorylation at both cytosolic and nuclear levels (Sup-
phosphorylation was observed (Supplementary Figure
4). p53 in these cells are conformationally altered and is
resistant to stress-induced activation (Gaitonde et al.,
2000; Wolff et al., 2001).
Again, as shown above, our findings indicate that
ER and AR are not involved in the hormone-induced
WOX1 activation. Without Tyr33 phosphorylation,
WOX1could not undergo
(Chang et al., 2003a). However, Tyr33-phosphorylated
WOX1 could not translocate to the nuclei in breast
cancer cells, suggesting the presence of cytosolic
inhibitors that prevent E2-induced nuclear translocation
E2induces complex formation and cotranslocation of p53/
WOX1 to the nuclei in COS7 fibroblasts, but not in
We have shown that WOX1 physically interacts with
p53 and JNK1 and the likely presence of a p53/WOX1/
JNK1 complex in vivo (Chang et al., 2003a). Next, we
investigated whether E2 stimulates cotranslocation of
the p53/WOX1/JNK1 complex to the nucleus. COS7
fibroblasts, cultured with 10% fetal bovine serum, were
stimulated with E2for 30min, followed by precipitating
with protein A agarose beads conjugated with IgG
antibodies against p-WOX1, p53 and JNK1, respec-
tively. Precipitating with anti-JNK1 showed the pre-
sence of JNK1 with p53 and WOX1 in the precipitates
from serum-stimulated COS7 cells, suggesting the
presenceof cytosolic p53/WOX1/JNK1
(Figure 6a). E2induced cotranslocation of WOX1 with
p53, but not with JNK1, to the nucleus (Figure 6a). The
nuclear WOX1 was Tyr33-phosphorylated (Figure 6a).
Precipitating with p53 or p-WOX1 also showed presence
of the p53/WOX1 complex in the nuclei of E2-treated
COS7 cells (Figure 6a).
In MCF7 cells, E2induced the complex formation of
cytosolic p53 and WOX1, whereas WOX1 in this
complex could not undergo nuclear translocation
(Figure 6b). These results are in agreement with the
above observations that WOX1 is mainly present in the
overnight in the presence of 10% fetal bovine serum, were exposed to E2 (40nM) for 30min, followed by processing
immunoprecipitation using cytosolic and nuclear fractions (B500mg protein) and protein A agarose beads conjugated with IgG
antibodies against p53, JNK1, and p-WOX1, respectively. Precipitating with anti-JNK1 or anti-p53 resulted in the presence of JNK1
with p53 and WOX1 in the precipitates from serum-stimulated COS7 cells, suggesting presence of a cytosolic p53/WOX1/JNK1
complex. E2induced cotranslocation of WOX1 with p53, but not with JNK1, to the nucleus. Precipitating with p-WOX1 also showed
presence of the p53/WOX1 complex in the nuclei. (b) In MCF7 cells, E2induced the complex formation of cytosolic p53 and WOX1,
whereas WOX1 in this complex could not undergo nuclear translocation. IP, immunoprecipitation; IgH, IgG heavy chain
E2 induces cotranslocation of p53 and WOX1 to the nuclei of COS7 but not MCF7 cells. (a) COS7 fibroblasts, cultured
17b-Estradiol activates WOX1/WWOXv1
N-S Chang et al
mitochondria of breast cancer cells from patients
By immunohistochemistry, we determined that progres-
sion of prostate from normal to hyperplasia, cancerous
and metastatic stages positively correlate with upregula-
tion and activation of WOX1 (WWOXv1), WOX2
(WWOXv2) and their family proteins (if present). In
comparison, breast cancer development to a premeta-
static stage is also associated with upregulation of these
family proteins, whereas their nuclear translocation
could not occur effectively in breast cells. Majority of
the upregulated proteins are Tyr33-phosphorylated and
located mainly in the mitochondria, suggesting that
Tyr33 phosphorylation is essential for their transloca-
tion to the mitochondria. Interestingly, significant
downregulation or absent expression of these family
proteins is shown in metastatic breast cancer cells, but
not in metastatic prostate cancer cells.
Mitochondrial localization of WOX1 has been shown
in vivo. By electron microscopy, we have recently
demonstrated that constant light induces translocation
of WOX1 to the mitochondria of photoreceptors in rat
eyes (Chen et al., 2004a,b). These mitochondrial WOX1
proteins are also Tyr33-phosphorylated.
While the development of breast and prostate cancers
is hormone-dependent, we discovered that E2activates
WOX1, p53 and ERK in COS7 fibroblasts, primary
lung epithelial cells, and AR-negative prostate DU145
cells, but not in ER-positive breast MCF7 cells.
Androgen also activates WOX1 in DU145 cells. E2
was shown to colocalize and cotranslocate with WOX1
to the nuclei in COS7 cells, as determined by immuno-
fluorescence (data not shown). These observations
suggest that estrogen and androgen are likely to bind
to the NSYK motif in the C-terminus of WOX1, and
that AR and ER are not involved in the E2-induced
Failure of sex hormones in inducing WOX1 nuclear
translocation is shown in MCF7 and other breast cells.
These findings are in agreement with the clinical findings
that Tyr33-phosphorylated WOX1 is located mainly in
the mitochondria and could not translocate to the nuclei
effectively. The mechanisms by which estrogen and
androgen fail to induce WOX1 nuclear translocation in
breast cells are unknown. E2induced the binding of p53
with WOX1 and their nuclear translocation in COS7
cells. Nonetheless, E2-increased p53/WOX1 complex
could not undergo nuclear translocation in MCF7
cells, suggesting presence of a blocker protein(s) that
prevents nuclear translocation. We have shown that
JNK1 binds and blocks cytosolic WOX1 activation
(Chang et al., 2003a). Whether JNK1 contributes a role
to the failure of E2-stimulated WOX1 nuclear transloca-
tion in breast cancer cells remains to be determined.
Other proteins such as p73, a p53 homologue (Aqeilan
et al., 2004), SIMPLE (PIG7) (Ludes-Meyers et al.,
2004) and Tau (Sze et al., 2004) are known to bind
WOX1. p73 and WOX1 appear to induce apoptosis
synergistically (Aqeilan et al., 2004). WOX1 binds Tau
via its C-terminal SDR domain and is likely to play a
critical role in regulating Tau hyperphosphorylation and
formation of neurofibrillary tangles in vivo (Sze et al.,
2004). Whether SIMPLE (PIG7) regulates WOX1
function has not been determined (Ludes-Meyers et al.,
WOX1 is overexpressed in benign prostatic hyperpla-
sia and prostatic cancer, as determined by immunohis-
tochemistry and tissue microarray analysis. Similar
results were also observed in adenocarcinomas of breast
and lung. Most interestingly, metastatic breast cells in
targeted organ sites have significantly reduced levels of
WOX1 and p-WOX1. Whether this is related with
alternative splicing of WWOX mRNA or alteration of
WWOX gene is unknown. ER-negative breast MDA-
MB-231 do not express the full-length mRNA coding
for wild type WOX1, whereas MDA-MB-435S cells
express the full-length WOX1 mRNA (Driouch et al.,
2002). These cells are metastatic and express very low
levels of wild type WOX1 protein. The observations
suggest that a blockade of translation of mRNA to
protein occurs in MDA-MB-435S cells.
Silencing of WOX1 expression may be essential for
cancer cell metastasis. That is, WOX1 protein may
participate in inhibiting cancer cell metastasis. This
notion is supported by the observations that distinct
deleted regions on chromosome segment 16q23–24,
encoding WWOX/FOR2/WOX1, are associated with
metastases in prostate and breast cancers (Caligo et al.,
1998; Elo et al., 1999; Li et al., 1999).
Tyr33 phosphorylation of WOX1 is also significantly
increased in the tissues of prostate and breast adeno-
carcinoma. However, majority of the phosphorylated
WOX1 does not translocate to the nuclei. Indeed, a
large portion of the phosphorylated WOX1 is present in
the mitochondria, suggesting its role in homeostasis in
these organelles. WOX1 possesses a conserved NSYK
motif that may interact with estrogen and androgen
(Chang et al., 2003b). The increased Tyr33-phosphor-
ylation suggests activation of WOX1 by sex hormones
Interestingly, prolonged exposure of COS7 cells to E2
for more than 24h resulted in apoptosis of these
cells (data not shown). Whether the induced apoptosis
is related with nuclear translocation of the proapoptotic
p53 and WOX1 is unknown. Under similar conditions,
E2 could not induce apoptosis of Mv1Lu and other
cells tested in this study. Estrogen is generally con-
sidered to exert neuroprotection (Gangolli et al., 1997;
Green et al., 1997; Gridley et al., 1998; Garcia-Segura
et al., 2001; Wise, 2002) and protect breast tissues from
apoptosis (Gompel et al., 2000; review). However, at
high doses estrogen may induce apoptosis (Song and
Santen, 2003; review). We do not exclude the possibility
that at a physiologic concentration, WOX1 is anti-
apoptotic and essential for cell survival (Chang et al.,
Functionally, ectopic p53 and WOX1 induce apopto-
sis in a synergistic manner (Chang et al., 2001).
17b-Estradiol activates WOX1/WWOXv1
N-S Chang et al
Suppression of WOX1 by antisense mRNA or by
dn-WOX1 abolishes p53 apoptosis (Chang et al.,
2001, 2003a). Most strikingly, all the above tested cells,
except SK-N-SH, demonstrate an inversed relationship
between the expression of WOX1 and p53. That is,
the levels of wild-type p53 are low in MCF7 and
Mv1Lu epithelial cells, yet their expression of WOX1
is high. In contrast, expression of mutant p53 is
increased in a variety of cancer cells, whereas their
WOX1 levels are low. Knockdown of WOX1 expression
by siRNA in L929 fibroblasts abolishes the expression
of p53 and its transcriptional target p21 (Chang et al.,
unpublished). These cells possess wild-type p53. Simi-
larly, suppression of WOX1 expression by siRNA
abrogates p53 and p21 expression in SK-N-SH cells,
suggesting that WOX1 is involved in the regulation of
Aqeilan et al. (2004) demonstrated that ectopic
human WWOX binds ectopic p73. Whether endogenous
p73 binds WWOX is not known. In their experiments,
this ectopic WWOX did not bind ectopic p53. We could
not estimate why both the overexpressed WWOX and
p53 failed to interact. We show that stress stimuli and
sex hormones increase the binding of WOX1 with p53.
Thus, a likely scenario for the failure of p53/WWOX
binding in the report of Aqeilan et al. is due to
experiments performed under nonstressed conditions
Whether WOX1 possesses a dehydrogenase activity is
not known and remains to be determined. Enzymes of
the SDR-containing protein family have been shown to
catalytically interact with estrogens and androgens
(Filling et al., 2002; Kallberg et al., 2002). For instance,
17-hydroxysteroid dehydrogenases (17-HSD) are im-
portant in regulating the biological potency of steroid
hormones by catalysing oxidation or reduction at
position 17 (Filling et al., 2001; He et al., 2001).
However, these enzymes could not translocate to the
nuclei in response to sex hormones. E2-induced WOX1
nuclear translocation is related with the N-terminal NLS
and two conserved phosphorylation sites at Tyr33 and
Tyr61 in the WW domains (Chang et al., 2001, 2003a).
In agreement with our previous observations (Chang
et al., 2003a), E2-induced Tyr33 phosphorylation is
essential for WOX1 nuclear translocation. Whether E2
induces Tyr61 phosphorylation is being determined in
Breast and prostate cancer cells develop sex hormone-
independent growth through unknown mechanisms.
This development could be due to loss, truncation, or
mutation of cell surface or nuclear ER and AR. WOX1
is a proapoptotic protein (Chang et al., 2001, 2003a).
Failure of estrogen or androgen-mediated WOX1
nuclear translocation may allow cancer cell growth.
That is, failure of WOX1 activation is likely to
contribute to hormone-independent growth of breast
and prostate cancer cells. Indeed, WOX1 can either
suppress or enhance the functions of several transcrip-
tion factors that are involved in growth regulation and
apoptosis at the nuclear level (Pugazhenthi et al.,
Materials and methods
Cell lines and chemicals
Cell lines maintained in our laboratory and used in these
studies were monkey kidney COS7 fibroblasts, murine L929
fibrosarcoma cells, human neuroblastoma SK-N-SH cells,
human prostate DU145 cells (AR negative) (Wollin et al.,
1989), and human breast MCF7 (ER positive) (Coezy et al.,
1982), MDA-MB-231 (ER negative) (Xu et al., 1996), and
MDA-MB-435S (ER negative) cells (Bronzert et al., 1982).
MCF7, MDA-MB-231 and MDA-MB-435S were gifts of
Drs John Noti and Mary Cismowski of this Institute.
Water-soluble E2 (1,3,5-estratriene-3,17b-diol; cyclodex-
17b-diol-6-one), estriol (1,3,5-estratriene-3,16a,17b-triol),
and androsterone (5a-androstan-3a-ol-17-one) were purchased
cDNA expression constructs
cDNA constructs for a murine wild-type and a dn-Wox1 were
made in pEGFP-C1 vector (Clontech) (Chang et al., 2001,
2003a). The expressed proteins were tagged with an N-terminal
enhanced green fluorescent protein (EGFP). This dn-WOX1
blocks p53-mediated apoptosis and anisomycin-mediated
WOX1 phosphorylation, but could not inhibit JNK1 activa-
tion (Chang et al., 2003a).
Antibodies and antibody production
We have generated antibodies against (1) an N-terminal
segment (RLAFTVDDNPTKPTTRQRY) of murine WOX1
at amino acid #89–107 (Chang et al., 2001), (2) Tyr33-
phosphorylation at the first WW domain of murine WOX1
TEEKT) at amino acid #28–42 (Chang et al., 2003a), and (3)
the first WW domain (amino acid #28–42) without Tyr33
phosphorylation (Chen et al., 2004b). Additionally, murine
Wox1 cDNA was subcloned in a pGEX2T vector (Amersham
Pharmacia Biotech), and recombinant WOX1 protein tagged
with glutathione-S-transferase (GST) was produced in bacteria
and purified according to the manufacturer’s instruction.
Antibodies against recombinant GST-WOX1 protein were
generated in rabbits as described (Chang et al., 2001, 2003a),
using an EZ Antibody Production and Purification Kit from
Pierce. The generated antisera were shown to interact with
both GST and WOX1.
Additional antibodies used in Western blottings were
against the following proteins: full-length p53 (FL-393; Santa
Cruz Biotechnology), Tyr204-phosphorylated ERK (Santa
Cruz Biotechnology), Ser15-phosphorylated p53 (Oncogene
Sciences), JNK1 (Santa Cruz Biotechnology), and a-tubulin
(Accurate Chemicals). Also, goat anti-human WWOX IgG
antibodies (N-19) were gifts of Santa Cruz Biotechnology.
Western blotting and coimmunoprecipitation
Coimmunoprecipitation was performed as previously de-
scribed (Chang et al., 2001, 2003a). Briefly, COS7 cells were
treated with E2for various indicated times to activate WOX1.
These cells were washed with ice-cold Tris-buffered saline
(TBS; pH7.4) containing 50mM EDTA (TBS/EDTA), and the
cytosolic and nuclear fractions of these cells were prepared
using NE-PER Nuclear and Cytoplasmic Extraction reagents
(Pierce). Endogenous WOX1, JNK1 and p53 were captured by
specific IgG antibodies against p-WOX1 (Chang et al., 2003a),
p53 (FL-393; Santa Cruz Biotechnology), or JNK1 (Santa
17b-Estradiol activates WOX1/WWOXv1
N-S Chang et al
Cruz Biotechnology) in the presence of protein A-agarose
beads (Pierce). Proteins bound by the agarose beads were then
washed with TBS/EDTA containing 0.5% NP-40, and
analysed by SDS–PAGE and Western blotting using indicated
Human prostate and breast tissues, immunohistochemistry and
Medically and legally consent postmortem prostate tissues
(five each) from patients with benign prostatic hypertrophy
(BPH), prostatic adenocarcinoma, prostatic metastasis, or age-
matched normal prostate tissues were obtained from the
autopsy service at the Department of Pathology, University of
Colorado Health Sciences Center. Similarly, breast tissues (five
each) from patients with normal breast, ductal adenocarcino-
ma, or metastatic cancer were examined. All tissues were
processed and treated following standard histology protocols
that include antigen retrieval, blocking of endogenous
peroxidase and nonspecific antigen binding (Sze et al., 2004).
The tissue sections were incubated with aliquots of antibodies
against the C-terminus of WOX1 or WOX2, the first WW-
domain of WOX1, or the Tyr33-phosphorylated WOX1
(1:200 final dilution in TBS, pH 7.4) at 41C overnight.
Aliquots of biotinylated secondary antibodies were then added.
Color development was performed using an LSABþHRP kit
(DAKO). In negative controls, tissue sections were stained
with aliquots of prebled rabbit sera. Where indicated, synthetic
peptides (100mM) were premixed with the above-mentioned
antisera (5min at room temperature) prior to immunostaining.
Also, immunofluorescence was performed using breast and
prostate tissue sections. Primary antibodies used in the
immunofluorescence were against: (1) Hsp60 (Stressgen), (2)
COX4 (Clontech), (3) WWOX (N-19) (Santa Cruz Biotech-
nology), and (4) Tyr33 phosphorylation in the first WW
domain (p-WOX1) (Chang et al., 2003a,b). Secondary
antibodies were (1) anti-rabbit IgG FITC, (2) anti-goat IgG
rodamine, and (3) anti-mouse IgG Texas Red.
Cancer tissue microarray
Multitumor tissue microarray slides were from the Tissue
Array Research Program, National Cancer Institute. These
slides were immunostained with antibodies against the N-
terminal region of WOX1 (Chang et al., 2001), and a
secondary FITC-conjugated anti-rabbit IgG. Each tissue
microarray slide contained a total of 500 sections of normal
tissues, CNS tumors, melanoma, lymphoma, and cancers of
ovarian, breast, colon, lung and prostate. Expression of
WOX1 was quantified by the extent of emitted green
Research was supported by the Department of Defense
(DAMD17-03-1-0736) to NSC. We thank Ms Terri Zimmer
for antibody production in rabbits, and Dr Robert I Richards
of the University of Adelaide, Australia, in contributing to the
production of antibodies against WOX1/FOR2 and WOX2/
FOR1. LJH is currently a visiting scientist from the National
Cheng Kung University Medical College, Taiwan, and
supported by the Ministry of Education, Taiwan, Republic
of China (Grant 91-B-FA09-1-4). JL was a Guthrie Scholar
from the Wilkes University, Wilkes-Barre, PA.
Aqeilan RI, Pekarsky Y, Herrero JJ, Palamarchuk A, Letofsky
J, Druck T, Trapasso F, Han SY, Melino G, Huebner K
and Croce CM. (2004). Proc. Natl. Acad. Sci. USA, 101,
Bednarek AK, Laflin KJ, Daniel RL, Liao Q, Hawkins KA
and Aldaz CM. (2000). Cancer Res., 60, 2140–2145.
Bronzert DA, Hochberg RB and Lippman ME. (1982).
Endocrinology, 110, 2177–2182.
Burchardt M, Burchardt T, Shabsigh A, Ghafar M, Chen
MW, Anastasiadis A, de la Taille A, Kiss A and Buttyan R.
(2001). Prostate, 48, 225–230.
Caligo MA, Polidoro L, Ghimenti C, Campani D, Cecchetti D
and Bevilacqua G. (1998). Int. J. Oncol., 13, 177–182.
Chang N-S. (2002a). Int. J. Mol. Med., 9, 19–24 (review).
Chang N-S. (2002b). BMC Cell Biol., 3, 8.
Chang N-S, Doherty J and Ensign A. (2003a). J. Biol. Chem.,
Chang N-S, Doherty J, Ensign A, Lewis J, Heath J, Schultz L,
Chen ST and Oppermann U. (2003b). Biochem. Pharmacol.,
66, 1347–1354 (review).
Chang N-S, Pratt N, Heath J, Schultz L, Sleve D, Carey GB
and Zevotek N. (2001). J. Biol. Chem., 276, 3361–3370.
Chen ST, Chuang JI, Cheng CL, Hsu LJ and Chang N-S.
(2004a). Neuroscience in press.
Chen ST, Chuang JI, Wang JP, Tsai MS, Li H and Chang
N-S. (2004b). Neuroscience, 124, 831–839.
Coezy E, Borgna JL and Rochefort H. (1982). Cancer Res., 42,
Driouch K, Prydz H, Monese R, Johansen H, Lidereau R and
Frengen E. (2002). Oncogene, 21, 1832–1840.
Elo JP, Harkonen P, Kyllonen AP, Lukkarinen O and Vihko
P. (1999). Br. J. Cancer, 79, 156–160.
Filling C, Berndt KD, Benach J, Knapp S, Prozorovski T,
Nordling E, Ladenstein R, Jornvall H and Oppermann U.
(2002). J. Biol. Chem., 277, 25677–25684.
Filling C, Wu X, Shafqat N, Hult M, Martensson E, Shafqat J
and Oppermann UC. (2001). Mol. Cell. Endocrinol., 171,
Gaitonde SV, Riley JR, Qiao D and Martinez JD. (2000).
Oncogene, 19, 4042–4049.
Gangolli EA, Conneely OM and O’Malley BW. (1997).
J. Steroid Biochem. Mol. Biol., 61, 1–9.
Garcia-Segura LM, Azcoitia I and DonCarlos LL. (2001).
Prog. Neurobiol., 63, 29–60 (review).
Gartel AL, Feliciano C and Tyner AL. (2003). Oncol Res., 13,
Gompel A, Somai S, Chaouat M, Kazem A, Kloosterboer HJ,
Beusman I., Forgez P, Mimoun M and Rostene W. (2000).
Steroids, 65, 593–598 (review).
Green PS, Bishop J and Simpkins JW. (1997). J. Neurosci., 17,
Gridley KE, Green PS and Simpkins JW. (1998). Mol.
Pharmacol., 54, 874–880.
Guler G, Uner A, Guler N, Han SY, Iliopoulos D,
Hauck WW, McCue P and Huebner K. (2004). Cancer,
He XY, Merz G, Yang YZ, Mehta P, Schulz H and Yang SY.
(2001). Eur. J. Biochem., 268, 4899–4907.
Isaacs WB, Carter BS and Ewing CM. (1991). Cancer Res., 51,
Kallberg Y, Oppermann U, Jornvall H and Persson B. (2002).
Eur. J. Biochem., 269, 4409–4417.
Keshamouni VG, Mattingly RR and Reddy KB. (2002).
J. Biol. Chem., 277, 22558–22565.
17b-Estradiol activates WOX1/WWOXv1
N-S Chang et al
Lau KM, LaSpina M, Long J and Ho SM. (2000). Cancer Download full-text
Res., 60, 3175–3182.
Li C, Berx G, Larsson C, Auer G, Aspenblad U, Pan Y,
Sundelin B, Ekman P, Nordenskjold M, van Roy F and
Bergerheim US. (1999). Genes Chromosomes Cancer, 24,
Linja MJ, Savinainen KJ, Tammela TL, Isola JJ and Visakorpi
T. (2003). Prostate, 55, 180–186.
Lokeshwar V and Rubinowicz D. (1999). Prostate Cancer PD,
Lokeshwar VB, Rubinowicz D, Schroeder GL, Forgacs E,
Minna JD, Block NL, Nadji M and Lokeshwar BL. (2001).
J. Biol. Chem., 276, 11922–11932.
Ludes-Meyers JH, Kil H, Bednarek AK, Drake J, Bedford MT
and Aldaz CM. (2004). Oncogene, 23, 5049–5055.
Marker PC, Donjacour AA, Dahiya R and Cunha GR. (2003).
Dev. Biol., 253, 165–174 (review).
Nesslinger NJ, Shi XB and deVere White RW. (2003). Cancer
Res., 63, 2228–2233.
Patel S, Turner PR, Stubberfield C, Barry E, Rohlff CR,
Stamps A, McKenzie E, Young K, Tyson K, Terrett J, Box
G, Eccles S and Page MJ. (2002). Int. J. Cancer, 97, 416–424.
Richards RI. (2001). Trends Genet., 17, 339–345 (review).
Ried K, Finnis M, Hobson L, Mangelsdorf M, Dayan S,
Nancarrow JK, Woollatt E, Kremmidiotis G, Gardner A,
Venter D, Baker E and Richards RI. (2000). Hum. Mol.
Genet., 9, 1651–1663.
Setalo Jr G, Singh M, Guan X and Toran-Allerand CD.
(2002). J. Neurobiol., 50, 1–12.
Song RX and Santen RJ. (2003). Apoptosis, 8, 55–60 (review).
Sze CI, Su M, Pugazhenthi S, Jambal P, Hsu LJ, Heath J,
Schultz L and Chang N-S. (2004). J Biol Chem., 279,
Toillon RA, Chopin V, Jouy N, Fauquette W, Boilly B and Le
Bourhis X. (2002). Breast Cancer Res Treat., 71, 269–280.
Watanabe A, Hippo Y, Taniguchi H, Iwanari H, Yashiro M,
Hirakawa K, Kodama T and Aburatani H. (2003). Cancer
Res., 63, 8629–8633.
Watters JJ, Campbell JS, Cunningham MJ, Krebs EG and
Dorsa DM. (1997). Endocrinology, 138, 4030–4033.
Wise PM. (2002). Trends Endocrinol Metab., 13, 229–230
Wolff A, Technau A, Ihling C, Technau-Ihling K, Erber R,
BoschFX and Brandner
Wollin M, FitzGerald TJ, Santucci MA, Menon M, Longcope
C, Reale F, Carlson J, Sakakeeny MA and Greenberger JS.
(1989). Radiother. Oncol., 15, 285–293.
Xu Y, Song J, Berelowitz M and Bruno JF. (1996).
Endocrinology, 137, 5634–5640.
Supplementary information accompanies the paper on Oncogene website (http://www.nature.com/onc)
17b-Estradiol activates WOX1/WWOXv1
N-S Chang et al