FEBS 29651 FEBS Letters 579 (2005) 3539–3546

Cross-talk between the TGFb and Wnt signaling pathways in

murine embryonic maxillary mesenchymal cells
Dennis R. Warner*, Robert M. Greene, M. Michele Pisano
University of Louisville Birth Defects Center, Department of Molecular, Cellular, and Craniofacial Biology,
501 South Preston Street, Suite 301, Louisville, KY 40292, United States

Received 8 March 2005; revised 11 May 2005; accepted 17 May 2005

Available online 2 June 2005

Edited by Lukas Huber

1. Introduction
Abstract The transforming growth factor beta (TGFb) and
Wnt signaling pathways play central roles regulating embryo-
genesis and maintaining adult tissue homeostasis. TGFb medi- Embryonic development is under the influence of several
ates its cellular effects through types I and II cell surface major signaling pathways, including those initiated by mem-
receptors coupled to the nucleocytoplasmic Smad proteins. bers of the TGFb and Wnt families of secreted growth and dif-
Wnt signals via binding to a cell surface receptor, Frizzled, which ferentiation factors. The TGFb superfamily is composed of
in turn activates intracellular Dishevelled, ultimately leading to three subfamilies: TGFbs, activins/inhibins, and bone morpho-
stabilization and nuclear translocation of b-catenin. Previous genetic proteins (BMPÕs) (for a recent review, see [1]). TGFb s
studies have demonstrated several points of cross-talk between elicit a plethora of biological responses ranging from morpho-
the TGFb and Wnt signaling pathways. In yeast two-hybrid genesis during early embryonic development, to cellular prolif-
and GST-pull down assays, Dishevelled-1 and Smad 3 have been
eration, differentiation, apoptosis, and stimulation of
shown to physically interact through the C-terminal one-half of
Dishevelled-1 and the MH2 domain of Smad 3. The current extracellular matrix synthesis [2–6]. TGFb effects these pro-
study demonstrates that co-treatment of murine embryonic max- cesses by altering patterns of gene expression.
illary mesenchyme (MEMM) cells with Wnt-3a and TGFb leads The TGFb signaling cascade is initiated by binding of TGFb
to enhanced reporter activity from TOPflash, a Wnt-responsive to a type II receptor-serine/threonine kinase (TbRII) that hete-
reporter plasmid. Transcriptional cooperation between TGFb rodimerizes with a type I receptor (TbRI, also a serine/threo-
and Wnt did not require the presence of a Smad binding element, nine kinase), resulting in transphosphorylation of TbRI by
nor did it occur when a TGFb-responsive reporter plasmid TbRII [7]. Activated (phosphorylated) TbRI in turn phosphor-
(p3TP-lux) was transfected. Overexpression of Smad 3 further ylates the Smad proteins, transducers of the TGFb signaling
enhanced the cooperation between Wnt and TGFb while overex- cascade [8]. The Smads are categorized into three distinct
pression of dominant-negative Smads 2 and 3 inhibited this ef-
groups based on function and sequence homology. R-Smads,
fect. Co-stimulation with TGFb led to greater nuclear
translocation of b-catenin, providing explanation for the effect substrates for TbRIs, are further divided into those regulated
of TGFb on Wnt-3a reporter activity. Wnt-3a exerted antiprolif- by TGFb/activin (Smads 2 and 3) and those regulated by
erative activity in MEMM cells, similar to that exerted by BMPÕs (Smads 1, 5, and 8). Smad 4, a ‘‘common’’ Smad uti-
TGFb. In addition, Wnt-3a and TGFb in combination led to syn- lized by multiple TGFb superfamily members, heterodimerizes
ergistic decreases in MEMM cell proliferation. These data dem- with phosphorylated R-Smads resulting in a complex that is
onstrate a functional interaction between the TGFb and Wnt imported into the nucleus [9]. Inhibitory Smads (Smads 6
signaling pathways and suggest that Wnt activation of the and 7), attenuate sustained TGFb signaling by blocking recep-
canonical pathway is an important mediator of MEMM cell tor-activation of R-Smads [10,11]. Once in the nucleus, the co-
growth. Smad/R-Smad complex binds to specific sequences within the

2005 Federation of European Biochemical Societies. Published
promoters of TGFb-responsive genes and either stimulates
by Elsevier B.V. All rights reserved.
or represses transcription. Transcriptional activation by Smads
Keywords: TGFb; Wnt; Orofacial; Cross-talk; Dishevelled; occurs with the aid of nuclear coactivators such as CBP and
Smad; Transcription p300 [12], while transcriptional repression is effected by inter-
action with nuclear corepressors such as SnoN and c-Ski
[13]. Thus, depending, in part, on the particular cellular com-
plement of nuclear coactivators and corepressors, the tran-
scriptional outcome of TGFb signaling will vary. R-Smads
bind to DNA, albeit poorly, via their conserved MH1 do-
*
mains. High affinity binding of Smads to DNA requires addi-
Corresponding author. Fax: +1 502 852 8309. tional cofactors. For example, the proteins OAZ and FAST2
E-mail address: dennis.warner@louisville.edu (D.R. Warner).
are required for Smad-dependent developmental processes by
Abbreviations: TGFb, transforming growth factor beta; TbRI, type I forming transcriptional complexes with Smads in the tran-
TGFb receptor; TbRII, type II TGFb receptor; GSK-3b, glycogen scription of Xvent-2 and goosecoid, respectively [14,15].
synthase kinase 3b; Dvl, Dishevelled; APC, adenomatous polyposis Wnt was initially described in mice as the proto-oncogene
coli; TCF, T-cell factor; LEF, lymphoid enhancer factor; MEMM,
murine embryonic maxillary mesenchyme; GST, glutathione-S-trans-
int-1 [16] and later determined to be identical to Drosophila
ferase; FBS, fetal bovine serum; SDS, sodium dodecyl sulfate; PAGE, wingless [17] (Wnt is a contraction of wingless and int-1).
polyacrylamide gel electrophoresis; PBS, phosphate buffered saline The vertebrate Wnt gene family (currently 19 genes) encodes

0014-5793/$30.00
2005 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.
doi:10.1016/j.febslet.2005.05.024
3540 D.R. Warner et al. / FEBS Letters 579 (2005) 3539–3546

secreted, lipid modified glycoproteins that bind to serpentine and female mice were mated overnight and the presence of a vaginal
receptors encoded by the Frizzled genes [18]. Wnts function plug the following morning was taken as evidence of mating and des-
ignated gestation day 0. Pregnant mice were euthanized on day 13 of
as morphogens regulating early developmental processes such gestation, a critical stage of murine orofacial development, embryos re-
as axis specification, neural patterning, and organ development moved, embryonic maxillofacial tissue microdissected and placed in
[19–22]. In addition, roles for Wnt in proliferation and apop- sterile, cold phosphate-buffered saline followed by gentle trypsiniza-
tosis have been demonstrated experimentally in a number of tion, and plated at a density of 6 · 103 cells/cm2. Experiments utilizing
cell types [23]. primary cultures were initiated within 2 days of plating at which time
the cells were approximately 50% confluent. These primary cultures are
The central protein in the canonical Wnt signaling pathway subsequently designated as murine embryonic maxillary mesenchyme
is b-catenin, identified by its interaction with cadherins and its (MEMM) cells.
role in cell adhesion [24,25]. In the absence of Wnt, intracellu-
lar levels of free b-catenin are kept low through phosphoryla- 2.2. Transfections and luciferase reporter assays
tion by glycogen synthase kinase 3b (GSK3b, [26,27]). For reporter assays, MEMM cells were seeded into 6-well tissue cul-
Phosphorylation of b-catenin by GSK-3b is facilitated by ture plates at a density of 7 · 104 cells/well and maintained in Opti-
MEM (GIBCO Invitrogen Corp., Gaithersburg, MD) supplemented
two negative regulators: axin (conductin) and the tumor sup- with 5% heat-inactivated fetal bovine serum (Sigma, St. Louis, MO)
pressor protein, adenomatous polyposis coli (APC) [28,29]. at 37 C in 5% CO2/95% air. The following day, when plasmid trans-
b-catenin, axin, APC, and GSK3b constitute a ‘‘destruction fections were performed, the cultures were 50–60% confluent. Cell cul-
complex’’ [30]. Phospho-b-catenin binds to the beta transducin tures were transfected with plasmid DNAs (as below) using Effectene
repeat-containing protein (bTrCP) component of this E3 ubiq- transfection reagent (Qiagen, Inc., Valencia, CA) according to the
manufacturerÕs recommendations with the following ratios of reagents:
uitin ligase complex, which ubiquitinates b-catenin, marking it Enhancer:DNA (8:1) and Effectene:DNA (15:1). For Wnt signaling re-
for elimination by the proteasome [31]. Wnt signaling antago- porter assays, the following amounts of plasmids were used, per well of
nizes b-catenin phosphorylation and subsequent proteolysis by a 6-well plate: 1 lg TOPflash or FOPflash (Upstate USA, Inc., Chi-
activating Dishevelled (Dvl) which disrupts formation of this cago, IL), 0.05 lg pRL-CMV (Promega, Madison, WI), 0.25 lg
pCMV-Smad 3, or the dominant-negative mutants, pCMV-Smad 2
destruction complex [32]. The mechanism of Wnt-Frizzled 3S/A and pCMV-Smad 3 3S/A. The dominant-negative mutants of
activation of Dvl involves its hyperphosphorylation and mem- Smads 2 and 3 were created by replacing the C-terminal serine residues
brane translocation [33,34]. Inhibition of b-catenin phosphor- with alanine which prevents receptor-mediated phosphorylation [43].
ylation leads to intracellular accumulation of the protein and The control plasmid pRL-CMV was co-transfected to normalize for
concomitant translocation to the nucleus where it forms a transfection efficiencies within individual experiments. Sixteen hours
post-transfection, cells were washed with fresh culture medium and
complex with transcription factors of the T-cell factor/lym- stimulated with TGFb1 (2 ng/ml, R&D Systems Inc., Minneapolis,
phoid enhancer factor (TCF/LEF) family that then binds to MN) for 24 h. Cells were harvested, lysates prepared, and luciferase
specific DNA sequences to activate gene transcription. Wnt activity determined with the Dual Luciferase Reporter Assay System
also regulates planar cell polarity and convergent extension (Promega) according to the method detailed by the manufacturer. Re-
porter luciferase activity (Firefly) was expressed relative to control
movements through a non-canonical pathway involving acti- luciferase activity (Renilla) for each sample. Each condition was as-
vation of Jun N-terminal kinase (JNK [35,36]). sayed in triplicate and the experiment was performed twice with com-
Wnts and TGFb superfamily members interact to regulate parable results.
the transcription of a number of genes. For example, the TGFb
superfamily member, Activin, and Wnt cooperate to stimulate 2.3. Cell proliferation assay
expression of Siamois in Xenopus [37]. Transcription of Xtwn, MEMM cells were seeded into 24-well plates at a density of 3500
also in Xenopus, requires a transcription complex that includes cells/cm2 and allowed to proliferate for 1–2 days in complete medium
(Opti-MEM containing 5% FBS), at which time they were provided re-
the TGFb-regulated Smad 2 and Smad 3, Smad 4, and TCF/ duced serum media (Opti-MEM containing 0.25% FBS) and main-
LEF [38]. Wnt and decapentaplegic signals are both required tained for 48 h to achieve cell synchrony. Cell cultures were treated
for expression of the Drosophila genes Vestigial and Ultrabitho- with control- or Wnt-3a-conditioned medium (50%, diluted in Opti-
rax and Wnt and TGFb cooperate in formation of SpemannÕs MEM containing 5% FBS), 2 ng/ml TGFb1, or 25 mM LiCl and
2 lCi/ml [3H]thymidine (20 Ci/mmol, Perkin–Elmer Life Sciences, Bos-
organizer in Xenopus [39]. Antagonism between these two sig- ton, MA). Treatment with TGFb1 and LiCl was performed in the pres-
naling pathways can occur through activation of TAK1 by ence of 50% control-conditioned media in order to normalize media
TGFb, which inhibits canonical Wnt signaling in transfected conditions so that a direct comparison of treatment efficacy could be
HEK293 cells [40]. Finally, Axin, has been shown to enhance made. At various time points, cultures were washed three times with
TbRI-dependent Smad phosphorylation [41]. 0.5 ml ice-cold PBS, twice for 10 min each with 0.5 ml ice-cold 10%
(w/v) trichloroacetic acid (TCA), rinsed again three times with 0.5 ml
In the experiments presented herein, functional cross-talk ice-cold PBS and then solubilized for 30 min with 0.5 ml 1 N NaOH.
between the TGFb and Wnt signaling pathways in embryonic Fifty ll glacial acetic acid was added to neutralize each sample and
maxillary mesenchymal cells is demonstrated. Using transcrip- each was mixed with 10 ml Scintiverse scintillation cocktail (Fisher
tion reporter assays combined with previous observations that Chemicals, Fairlawn, NJ) and incorporation of [3H]thymidine into
TCA-precipitable material was determined by liquid scintillation
Smads and Dvl physically interact [42], these data demonstrate spectroscopy.
several levels of cross-talk between TGFb and Wnt in cells de-
rived from developing orofacial tissue.
2.4. Preparation of Wnt-3a conditioned medium
Mouse L-cells stably transfected with the cDNA for Wnt-3a were
obtained from ATCC (Manassas, VA) and control, non-transfected
L-cells were from Dr. Roel Nusse (Stanford University, Stanford,
2. Materials and methods CA). Wnt-3a- and control-conditioned media were prepared according
to the method of Shibamoto and were stored for up to one year with-
2.1. Animals and primary cell cultures out appreciable loss of activity [44]. Using this method, the concentra-
ICR mice were obtained from Harlan (Indianapolis, IN), and tion of soluble Wnt-3a in 100% conditioned media is approximately
housed at a temperature of 22 C with an alternating 12 h light/dark 400 ng/ml [44], however, there were batch-to-batch variations in
cycle and given access to food and water ad libitum. Mature male activity.
D.R. Warner et al. / FEBS Letters 579 (2005) 3539–3546 3541

2.5. Construction of dominant-negative Smad 2 and Smad 3

To create dominant-negative mutants of both Smad 2 and Smad 3,
PCR was utilized using wild-type Smad 2 and Smad 3 cDNA as a tem-
plate in order to change the three carboxyl terminal serine residues to
alanine (Smad 2, -SSMS to -AAMA; Smad 3, -SSVS to -AAVA). This
removed the phosphorylation site necessary for activation [43]. Both
mutants were cloned into pCMV-myc (BD Biosciences) resulting in
fusion of a c-myc epitope to the amino terminus.

2.6. Nuclear translocation of b-catenin

To determine cytosolic and nuclear levels of b-catenin in MEMM
cells, primary cultures were established in 10 cm tissue culture plates
(2 plates per treatment) and grown to 70% confluence. Cultures were
serum-deprived for 16 h in DME/F12 (Sigma), and then stimulated
with control- or Wnt-3a-conditioned media in the presence or absence Fig. 1. Activation of the canonical Wnt signaling pathway in MEMM
of 2 ng/ml TGFb1 for 3 h. Cells were washed twice with ice-cold PBS cells. Primary cultures of MEMM cells were established and seeded
and scraped in 1 ml PBS. Cells were collected by centrifugation at into 6-well plates at a density of 7.5 · 104 cells/well. Cells were allowed
2000 · g for 2 min and cytosolic and nuclear protein fraction prepared to reach 50–60% confluence and were transfected with 1 lg of the
using the NE-PER nuclear and cytoplasmic extraction kit from Pierce TCF-reporter plasmid, TOPflash, or a mutated version, FOPflash,
(Rockford, IL) according to the manufacturerÕs instructions. Five- which does not respond to Wnt signaling, and 50 ng pRL-CMV, which
10 lg were analyzed by SDS–PAGE and immunoblotting with anti- served as a control to ascertain transfection efficiency. Twenty-four
active b-catenin antibody (clone 8E4; Upstate Cell Signaling Solutions, hours post transfection, cells were treated with control- or Wnt-3a-
Lake Placid, NY), which recognizes non-phosphorylated b-catenin. conditioned media (50%, diluted in 5% FBS/Opti-MEM) for an
Quantitative assessment of translocation was determined by densitom- additional 24 h. Cell lysates were prepared and assayed for Firefly
etry using NIH Image 1.63f. luciferase (expressed from TOPflash or FOPflash) or Renilla luciferase
activity (from pRL-CMV). Data are expressed as the ratio of Firefly
2.7. Statistical analyses luciferase activity to that of Renilla luciferase. Each data point
represents the means ± standard deviation of triplicate determinations.
Data from reporter and proliferation assays was analyzed using
Wnt-3a conditioned medium stimulated transcription specifically from
one-way ANOVA followed by BonferroniÕs multiple comparison test.
TOPflash (P < 0.05, one-way ANOVA and BonferroniÕs test), with no
P-values less than or equal to 0.05 were considered significant.
effect on FOPflash.

3. Results

3.1. Stimulation of a Wnt reporter construct in MEMM cells

In order to determine the ability of MEMM cells to respond
to signals through the canonical Wnt signaling pathway (i.e.
through b-catenin and TCF/LEF), primary cultures of
MEMM cells were transfected with the reporter construct,
TOPflash. This construct contains six copies of an optimized
TCF-consensus binding site upstream of Firefly luciferase.
FOPflash, which contains mutations in each of the six TCF
consensus sites rendering it insensitive to Wnt signaling, was
transfected into MEMM cells as a negative control [45].
Twenty-four hours following transfection, cells were treated
for an additional 24 h with control- or Wnt-3a-conditioned
medium (50% concentration) and luciferase activity measured.
Wnt-3a was chosen because it has been shown to activate the
canonical Wnt pathway through stabilization of b-catenin
[44]. Additionally, Wnt-3a was found to be present in the
developing secondary palate (manuscript in preparation). As Fig. 2. TGFb and Wnt cooperate to stimulation transcription from a
reporter plasmid containing TCF binding sites. Primary cultures of
shown in Fig. 1, Wnt-3a effectively stimulated transcription MEMM cells were transfected with TOPflash, FOPflash, or the p3TP-
of luciferase from the TOPflash reporter construct, but not lux reporter plasmids and pRL-CMV. Cultures were incubated for
from FOPflash, demonstrating that MEMM cells contain a 24 h and then stimulated with 50% conditioned medium (control or
functional canonical Wnt signaling pathway. Wnt-3a) in the presence or absence of 2 ng/ml TGFb1 for 24 h.
Induced-expression of luciferase was measured as detailed in Section 2
and is reported as the means ± standard deviation of triplicate
3.2. Cross-talk between the Wnt-3a and TGFb signaling determinations. Wnt-3a treatment resulted in an increase in transcrip-
pathways in MEMM cells tion from TOPflash, but not FOPflash or p3TP-lux. TGFb treatment
resulted in transcription only from p3TP-lux. Co-treatment of MEMM
In order to define a functional interaction between the Wnt- cells with both Wnt-3a and TGFb led to a synergistic response from
3a and TGFb signaling pathways in MEMM cells, cultures TOPflash (P < 0.05, as compared to Wnt-3a alone, one-way ANOVA
were transfected with TOPflash, FOPflash, or p3TP-lux, a and BonferroniÕs test) but not p3TP-lux.
luciferase reporter plasmid highly responsive to Smad-medi-
ated TGFb signaling [46] and stimulated with 50% Wnt-3a
conditioned medium and/or 2 ng/ml TGFb. The results shown sults presented in Fig. 1. However, co-treatment with both
in Fig. 2 demonstrate that Wnt-3a stimulation led to activation 50% Wnt-3a conditioned medium and 2 ng/ml TGFb led to
of TOPflash but not FOPflash, which is consistent with the re- an enhanced activation of TOPflash (P < 0.05 compared to
3542 D.R. Warner et al. / FEBS Letters 579 (2005) 3539–3546

Wnt-3a alone) (Fig. 2). The increase appeared to be specific to

the TCF reporter plasmid, as no effect of Wnt-3a on transcrip-
tion from p3TP-lux was detected either in the presence or ab-
sence of TGFb. TGFb alone had no affect on luciferase
transcription from TOPflash or FOPflash, nor did Wnt-3a
alone have any affect on transcription from p3TP-lux. To fur-
ther examine cross-talk between the TGFb and Wnt-3a signal-
ing pathways in MEMM cells, and to determine if this
interaction is Smad-dependent, Smad 3 was overexpressed in
MEMM cells also transfected with TOPflash. Cells were then
stimulated with Wnt-3a conditioned medium with or without
TGFb. Overexpression of Smad 3 led to a generalized increase
in basal TOPflash reporter activity (Fig. 3). However, the effect
is not specific to TOPflash, as basal activities from FOPflash
are also elevated. Smad 3 overexpression led to an increase Fig. 4. Inhibition of TGFb signaling through overexpression of
in transcriptional activity from TOPflash in response to Wnt- dominant-negative mutants of Smads 2 and 3 blocks cooperation
between TGFb and Wnt-3a in MEMM cells. Primary cultures of
3a, from 1.5-fold to 1.9-fold, compared to control values MEMM cells were transfected with TOPflash or FOPflash, pRL-
(Fig. 3). Furthermore, the transcriptional activation observed CMV, and either pCMV-myc (empty vector) or pCMV-myc vectors
with Wnt-3a and TGFb was enhanced in the presence of excess that contained dominant-negative mutants of Smad 2 and Smad 3
Smad 3 (7-fold stimulation over basal, compared to 2-fold (Dom.-neg.). Transfected cell cultures were treated with control- or
Wnt-3a-conditioned medium in the presence or absence of 2 ng/ml
when Smad was not overexpressed, P < 0.01 vs. Wnt-3a
TGFb1, and luciferase activity measured and normalized to that from
alone), suggesting that Smads can mediate cross-talk between pRL-CMV. The inset is a Western blot with anti-myc antibodies
TGFb and Wnt. Interestingly, overexpression of Smad 3 also demonstrating expression of both myc-Smad 2 3S/A and myc-Smad 3
led to increases in transcription of TOPflash by TGFb, to a le- 3S/A. Inhibition of TGFb signaling with dominant-negative mutants
vel actually greater than that stimulated by Wnt-3a alone. of Smads 2 and 3 not only led to an increased ability of Wnt-3a to
induce transcription from TOPflash, but also blocked the ability of
Although Smad 3 led to increased basal activity from FOP- TGFb to enhance Wnt-3a-mediated transcription. Transfection with
flash, no specific effect of Wnt-3a conditioned media or TGFb dominant-negative mutants of Smad 2 and Smad 3 led to a general
was observed on luciferase levels expressed from this plasmid. increase in activity from FOPflash, but there was no specific activation
To further examine the Smad dependence on the synergistic from Wnt-3a or TGFb stimulation.
activation of TOPflash by Wnt-3a and TGFb, dominant-neg-
ative mutants of both Smad 2 and Smad 3 were overexpressed
in TOPflash- or FOPflash-transfected MEMM cells and then nant-negative mutants of Smad 2 and 3 contain mutations of
the cultures were stimulated with Wnt-3a conditioned media the carboxyl-terminal serine residues that prevent TGFb recep-
in the presence or absence of TGFb (Fig. 4). These domi- tor-mediated phosphorylation and effectively block signaling
through the TGFb pathway [43]. Inhibition of TGFb signaling
by dominant-negative mutants of Smad 2 and Smad 3 blocked
the combined activation of TOPflash by Wnt-3a and TGFb
(Fig. 4). Similar to the results with overexpression of wild-type
Smad 3 (Fig. 3), overexpression of dominant-negative Smads 2
and 3 resulted in an enhanced ability of Wnt-3a to increase
transcription on TOPflash (from 1.6-fold to 3.5-fold). The
dominant-negative Smad 2 and Smad 3 mutants blocked the
low level of TGFb-mediated stimulation of TOPflash reporter
assay observed in this experiment. Although mediating a small
increase in basal activity, Smad 2 3S/A and Smad 3 3S/A had
no effect on FOPflash reporter activity. These data suggest that
Smads may act as a cofactor in the Wnt signaling pathway and
that Smad phosphorylation may not be necessary for enhance-
ment of Wnt signaling, however it is required for TGFb-med-
iated effects on TOPflash transcription.
Fig. 3. Overexpression of Smad 3 in MEMM cells leads to enhanced
Wnt-3a-TGFb-mediated transcription. Primary cultures of MEMM 3.3. Cross-talk between TGFb and Wnt downstream of
cells were transfected with the TOPflash or FOPflash reporter Frizzled and Dvl
construct, pRL-CMV, and the expression vector pCMV-myc without To determine if the effect of TGFb on Wnt-3a-transcrip-
a cDNA insert (empty vector) or with full-length Smad 3 followed by
tional activity requires signaling through the Frizzled receptors
treatment with 50% conditioned medium (control or Wnt-3a) with or
without 2 ng/ml TGFb1 for 24 h. The inset is a Western blot with anti- and Dvl, MEMM cells were transfected with TOPflash or
myc antibodies demonstrating expression of myc-Smad 3. Overex- FOPflash and stimulated with 25 mM LiCl in the presence or
pression of Smad 3 led to a marked increase in both basal and absence of 2 ng/ml TGFb. Lithium ions directly inhibit
stimulated transcription of TOPflash and the cooperation between GSK-3b and stimulate the canonical pathway by increasing
TGF and Wnt-3a in stimulating transcription was enhanced. Although
overexpression of Smad 3 led to increased activity from FOPflash, b-catenin levels independent of receptor activation [47]. LiCl
there was no effect from treatment with Wnt-3a conditioned media was effective in stimulating luciferase activity from TOPflash
either in the presence or absence of TGFb. (P < 0.001 vs. NaCl control) (Fig. 5), with a response higher
D.R. Warner et al. / FEBS Letters 579 (2005) 3539–3546 3543

Table 1
Nuclear translocation of b-catenin in MEMM cells
Treatment Densitometric Average fold-change
units
Exp. 1 Exp. 2
Control 163 419 –
TGFb 185 359 No-change
Wnt-3a 219 506 1.2 ± 0
Wnt-3a + TGFb 289 636 1.6 ± 0.1
Primary cultures of MEMM cells were serum-starved for 16 h and
stimulated with control- or Wnt-3a conditioned media ±2 ng/ml
TGFb1 for 3 h. Cells were harvested and total nuclear protein was
Fig. 5. Synergistic transcriptional activation of TOPflash by TGFb isolated as described in Section 2. Nuclear protein (5 lg) was analyzed
and LiCl. Primary cultures of MEMM cells were transfected with by SDS–PAGE and Western blotting with anti-b-catenin (active, non-
TOPflash or FOPflash, and pRL-CMV. Twenty-four hours post phosphorylated). The level of b-catenin present in the nucleus was
transfection, cells were treated with either control- or Wnt-3a- determined by densitometry with NIH image 1.63 and data are pre-
conditioned medium, 25 mM NaCl or 25 mM LiCl in the presence sented from two independent experiments. The data are also expressed
or absence of 2 ng/ml TGFb1 for 24 h. Each data point represents the as the average fold-change ± range, relative to control values.
means ± standard deviation of the luciferase activity from triplicate
transfections. LiCl efficiently stimulated transcription of TOPflash
(P < 0.001, compared to NaCl), and when combined with TGFb, lead
to a marked increase in luciferase transcription (P < 0.001, compared the effect of various treatments on cell proliferation was ascer-
to LiCl alone). For comparison, Wnt-3a ± TGFb was also tested and tained by measuring [3H]thymidine incorporation into DNA.
was consistent with earlier observations (not shown). No effect was
As has been previously demonstrated, TGFb exerts an anti-
observed with FOPflash.
proliferative effect on MEMM cells [52] (Fig. 6). Treatment
with Wnt-3a conditioned medium inhibited MEMM cell pro-
liferation to a similar extent as that seen in response to 2 ng/
than that obtained with Wnt-3a conditioned media (not ml TGFb. Co-treatment of MEMM cells with both Wnt-3a
shown). It is unclear why stimulation of MEMM cells with conditioned medium and TGFb resulted in a further reduction
Wnt-3a results in a lower response than with LiCl, however, in the rate of proliferation of MEMM cells (P < 0.05 com-
the receptor expression (Frizzled) and concentration of Wnt- pared to Wnt-3a or TGFb alone; Mann–Whitney test). To
3a in the conditioned media may combine to result in lower determine if the effect observed with Wnt-3a is mediated
signals. Nevertheless, simultaneous treatment with LiCl and through the canonical pathway, the effect of 25 mM LiCl on
TGFb, resulted in a greater level of luciferase activity than that cell proliferation was examined (Fig. 6). Although 25 mM LiCl
observed with LiCl alone (P < 0.001). These data demonstrate inhibited cell proliferation, it was not as effective as Wnt-3a
that the cooperation between TGFb and Wnt-3a is likely med- conditioned medium in inhibiting MEMM cell proliferation.
iated downstream of Frizzled and Dvl. Whether this indicates that Wnt-3a utilizes alternate pathways,

3.4. TGFb co-stimulation with Wnt-3a leads to increased

b-catenin nuclear translocation
One explanation for the positive effect of TGFb on Wnt-3a-
mediated TOPflash reporter activity could be due to enhanced
nuclear translocation of b-catenin. When MEMM cells were
stimulated with Wnt-3a conditioned media, increased nuclear
levels of b-catenin were detected with an antibody specific
for non-phosphorylated (i.e. activated) b-catenin (Table 1).
Little effect by TGFb alone on b-catenin levels was detected.
In the presence of 2 ng/ml TGFb, stimulation with Wnt-3a
conditioned media led to a small, but reproducible (1.5–1.8-
fold) increase in the level of activated b-catenin found in the
nucleus. As b-catenin is a transcriptional co-factor, these rela-
tively small changes in nuclear levels may lead to much larger Fig. 6. TGFb and Wnt-3a cooperate to inhibit MEMM cell prolifer-
ation. Primary cultures of MEMM cells were seeded into 24-well plates
responses seen at the transcriptional level. Since Smads have
at a density of 7000 cells/well. One day after seeding, medium was
been shown to bind b-catenin [48], it is possible that a Smad- replaced with reduced-serum medium (0.25% FBS) and incubated for
b:catenin complex formed by stimulation with both TGFb an additional 48 h to synchronized the cells. Control- or Wnt-3a-
and Wnt-3a leads to the increase in nuclear b-catenin observed conditioned medium (50%, diluted in 5% FBS/Opti-MEM) containing
in this experiment. the TGFb1 vehicle (0.1% BSA, 4 mM HCl), 2 ng/ml TGFb1, 25 mM
NaCl, or 25 mM LiCl was then added to the cells. In addition, all
medium contained 2 lCi/ml [3H]thymidine (20 Ci/mmol). At the
3.5. Cross-talk between TGFb and Wnt-3a in regulation of indicated time following addition of TGFb or Wnt-3a, the cells were
MEMM cell proliferation harvested and the amount of TCA-insoluble [3H]thymidine was
TGFb and Wnt are both able to alter cell proliferation [49– determined and expressed as cpm ± standard deviation of four
samples. TGFb and Wnt-3a inhibited MEMM cell proliferation
51]. TGFb negatively regulates the proliferation of MEMM (P < 0.001, compared to control at 24 and 48 h). The combination of
cells [52]. To examine the effect of Wnt-3a on MEMM cell pro- Wnt-3a and TGFb led to a small synergistic decrease in MEMM cell
liferation, cells were synchronized by serum deprivation and proliferation (P < 0.05).
3544 D.R. Warner et al. / FEBS Letters 579 (2005) 3539–3546

in addition to the canonical pathway to inhibit cell prolifera- tion that Wnt had no effect on the expression of gelatinase
tion, or that LiCl is not as efficacious as Wnt-3a is unclear. B, a TGFb-responsive gene, in MEMM cells, either in the pres-
LiCl may also have other cellular effects, such as increasing ence or absence of TGFb (data not shown).
inositol lipids, which may influence its efficacy in this assay The observed cooperation between TGFb and Wnt in
[53]. These data do suggest, however, that Wnt-3a can inhibit MEMM cells on transcription of TOPflash is not consistent
proliferation of MEMM cells, in part, through control of with an earlier report examining the functional interaction be-
b-catenin levels. The cooperation between TGFb and Wnt-3a tween these two pathways. Labbé et al. demonstrated that
in regulating MEMM cell proliferation may not extend to TGFb and Wnt-3a cooperate in the regulation of the Xenopus
other cellular phenomena as no cooperation was observed in homeobox gene Xtwn, but demonstrated that binding sites for
TGFb-induced up-regulation of gelatinase B (data not shown). both Smads and TCF are required, not only in the natural pro-
This latter observation is consistent with reporter assays dem- moter, but also in artificial plasmid constructs (e.g. TOPflash,
onstrating no effect of Wnt-3a conditioned medium on tran- [56]). However, in this case the Wnt signaling pathway was
scription from the TGFb-responsive reporter, p3TP-Lux stimulated by overexpression of LEF1 in a cell line containing
(Fig. 2). activating mutations of b-catenin (HepG2) and may not reflect
endogenous activation by the natural ligand, Wnt.
A recent report demonstrated that treatment of mesenchy-
4. Discussion mal progenitor cells with TGFb resulted in transcriptional
activation of TOPflash and nuclear translocation of b-catenin
TGFb and Wnt activate two well-characterized signaling in the absence of exogenous Wnt signaling [57]. These data
pathways that function during embryonic development and suggest that responses to TGFb, Wnt, and combined
in adult tissue homeostasis. The TGFbs and Wnts bind to cell TGFb-Wnt treatment may be tissue-specific since no consis-
surface receptors that elicit nuclear translocation of a tran- tent effect of TGFb on TOPflash was found in MEMM or
scription factor that associates with additional nuclear proteins Mv1Lu cells.
to promote specific changes in gene expression. Cell-specific Since nuclear translocation of b-catenin leads to activation
interpretation of extracellular signals depends, in part, on inte- of transcription, any changes in either its degradation or nucle-
gration between distinct signaling pathways. Signaling path- ar import would lead to alterations in transcription. Co-treat-
way crosstalk is now recognized as a critical mechanism by ment of MEMM cells with TGFb and Wnt-3a conditioned
which inputs from multiple pathways can cooperate in gene media led to increased levels of b-catenin found in the nucleus,
transcription, or the means by which one signaling pathway and may provide one mechanism for the greater level of Wnt-
can silence a competing or antagonistic pathway. Multiple lev- 3a-mediated TOPflash reporter activity conferred by TGFb.
els of cooperation between the TGFb and Wnt signaling path- Although the increases seen when TGFb was included with
ways in regulating gene expression have been documented. For Wnt-3a were small, relatively minor changes in levels of tran-
example, Wnt and TGFb cooperate to induce the expression of scription factors could potentially lead to larger transcriptional
Vestigial and Ultrabithorax [54,55]. Labbé et al. [56] demon- changes. Whether Smads and b-catenin form a complex that is
strated that Smads 2, 3, and 4 bind to LEF1, to synergistically more efficiently imported into the nucleus, or whether Smads
activate the Xtwn promoter in Xenopus. Previously, it was and b-catenin independently move into the nucleus and form
demonstrated that Smad 3 binds to Dvl-1 in a region encom- a more effective transcriptional complex awaits further investi-
passing most of the PDZ domain and the entire DEP domain gation. As the ultimate cellular effect and mechanisms of sig-
[42]. naling by both TGFb and Wnt-3a are complex, other
The data presented in this report is the first demonstration mechanisms mediating cross-talk between these two pathways
that MEMM cells respond to Wnt by increasing transcription may be operational and thus the simple increase in nuclear
through the canonical pathway. Activation of the cell polarity b-catenin levels may be insufficient to explain this phenome-
pathway (through activation of JNK) was also tested using non. The increase in b-catenin, however, may explain in part
JNK activation-specific antibodies, however, no activation of the increase in TOPflash reporter activity seen when MEMM
JNK was observed in response to Wnt-3a or Wnt-5a (data cells are treated with both TGFb and Wnt-3a.
not shown). Due to the difficulties in obtaining functional The data presented in the current report supports a model
Wnts, others were not tested for their ability to stimulate the whereby Smad is a common cofactor for both the TGFb and
canonical and cell polarity pathways in MEMM cells. Wnt signaling pathways. It is possible that the observed inter-
TGFb alone activated TOPflash in some experiments (e.g. action between TGFb and Wnt is mediated by CREB binding
Figs. 3 and 4), which is consistent with previous reports dem- protein (CBP), since CBP, through its interaction with Smads
onstrating that TGFb is capable of activating this reporter [12], TCF [58], and b-catenin [59] is a common transcriptional
construct in other cell types [57]. This effect was dependent co-activator for both TGFb and Wnt. Although the role(s) of
upon Smads and activation of the TGFb pathway. However, Wnts in orofacial development remain largely unknown, the
overexpression of Smad alone (either wild-type or dominant- above data suggest that Wnts, in particular those stimulating
negative mutants), even in the absence of TGFb was sufficient the canonical pathway, are important mediators of cell prolif-
to render MEMM cells more responsive to Wnt. These data eration since Wnt-3a exerted a strong antiproliferative re-
suggest that Smads have the ability to serve as a cofactor in sponse on MEMM cells. This may be particularly significant
the Wnt signaling pathway but may not require activation of for development of the secondary palate, a primary research
the TGFb pathway. The inability of Wnt-3a to elicit a tran- focus in this laboratory, since alterations in mesenchyme cell
scriptional response from p3TP-lux suggests that Wnt-3a can- proliferation may lead to clefts of the secondary palate
not stimulate transcription from a promoter consisting of only [60,61]. In addition, synergistic cross-talk between Wnt-3a
a Smad binding element. This is consistent with the observa- and TGFb on promoters containing TCF binding sites may al-
D.R. Warner et al. / FEBS Letters 579 (2005) 3539–3546 3545

ter gene transcription profiles (i.e. some genes may require in- regulate TGFb-dependent transcription through the forkhead
puts from both Wnt and TGFb) or TGFb may modulate the DNA-binding protein FAST2. Mol. Cell 2, 109–120.
[16] Nusse, R. and Varmus, H.E. (1982) Many tumors induced by the
strength of the Wnt-b-catenin/TCF signal. Although the bio- mouse mammary tumor virus contain a provirus integrated in the
logical effects of Wnt and TGFb, and cross-talk between the same region of the host genome. Cell 31, 99–109.
two, has been defined in a cell culture system, the ultimate goal [17] Rijsewijk, F., Schuermann, M., Wagenaar, E., Parren, P., Weigel,
is to understand their roles in embryogenesis. Thus, data from D. and Nusse, R. (1987) The Drosophila homolog of the mouse
the experiments presented in the present report using cultured mammary oncogene int-1 is identical to the segment polarity gene
wingless. Cell 50, 649–657.
embryonic mesenchyme cells can be used to design experimen- [18] Bhanot, P., Brink, M., Samos, C.H., Hsieh, J.C., Wang, Y.,
tal approaches to elucidate signaling mechanisms in craniofa- Macke, J.P., Andrew, D., Nathans, J. and Nusse, R. (1996) A new
cial development in vivo. member of the Frizzled family from Drosophila functions as a
Wingless receptor. Nature 382, 225–230.
Acknowledgements: The authors thank Dr. Roel Nusse (Stanford Uni- [19] Zecca, M., Basler, K. and Struhl, G. (1996) Direct and long-range
versity) for mouse L-cells and Dr. Joan Massagué (Memorial Sloan- action of a wingless morphogen gradient. Cell 87, 833–844.
Kettering Cancer Center) for p3TP-lux. This work was supported, in [20] Parr, B.A. and McMahon, A.P. (1995) Dorsalizing signal Wnt-7a
part, by NIH grants DE12363 (to M.M.P.), DE12858, DE05550, required for normal polarity of D-V and A-P axes of mouse limb.
and P20 RR017702 from the COBRE Program of the National Center Nature 374, 350–353.
for Research Resources (all to R.M.G.), and the Commonwealth of [21] Ikeya, M., Lee, S.M., Johnson, J.E., McMahon, A.P. and
Kentucky Research Challenge Trust Fund. Takada, S. (1997) Wnt signalling required for expansion of
neural crest and CNS progenitors. Nature 389, 966–970.
[22] Pandur, P., Lasche, M., Eisenberg, L.M. and Kuhl, M. (2002)
Wnt-11 activation of a non-canonical Wnt signalling pathway is
References required for cardiogenesis. Nature, 636–641.
[23] Chen, S., Guttridge, D.C., You, Z., Zhang, Z., Fribley, A., Mayo,
[1] Shi, Y. and Massagué, J. (2003) Mechanisms of TGF-beta M.W., Kitajewski, J. and Wang, C.Y. (2001) Wnt-1 signaling
signaling from cell membrane to the nucleus. Cell 113, 685–700. inhibits apoptosis by activating beta-catenin/T cell factor-medi-
[2] Zorn, A.M., Butler, K. and Gurdon, J.B. (1999) Anterior ated transcription. J. Cell Biol. 152, 87–96.
endomesoderm specification in Xenopus by Wnt/beta-catenin [24] Ozawa, M., Baribault, H. and Kemler, R. (1989) The cytoplasmic
and TGF-beta signalling pathways. Dev. Biol. 209, 282–297. domain of the cell adhesion molecule uvomorulin associates with
[3] Edwards, D.R., Murphy, G., Reynolds, J.J., Whitham, S.E., three independent proteins structurally related in different species.
Docherty, A.J.P., Angel, P. and Heath, J.K. (1987) Transforming EMBO J. 8, 1711–1717.
growth factor beta modulates the expression of collagenase and [25] Nagafuchi, A. and Takeichi, M. (1989) Transmembrane control
metalloproteinase inhibitor. EMBO J. 61, 1899–1904. of cadherin-mediated cell adhesion: a 94 kDa protein functionally
[4] Roberts, A.B., Anzano, M.A., Wakefield, L.M., Roche, N.S., associated with a specific region of the cytoplasmic domain of E-
Stern, D.F. and Sporn, M.B. (1985) Type beta transforming cadherin. Cell Regul. 1, 37–44.
growth factor: a bifunctional regulator of cellular growth. Proc. [26] Peifer, M., Pai, L.M. and Casey, M. (1994) Phosphorylation of
Natl. Acad. Sci. USA 82, 119–123. the Drosophila adherens junction protein Armadillo: roles for
[5] Rotello, R.J., Lieberman, R.C., Purchio, A.F. and Gerschenson, wingless signal and zeste-white 3 kinase. Dev. Biol. 166, 543–556.
L.E. (1991) Coordinated regulation of apoptosis and cell prolif- [27] van Noort, M., Meeldijk, J., van der Zee, R., Destree, O. and
eration by transforming growth factor beta 1 in cultured uterine Clevers, H. (2002) Wnt signaling controls the phosphorylation
epithelial cells. Proc. Natl. Acad. Sci. USA 88, 3412–3415. status of beta-catenin. J. Biol. Chem. 277, 17901–17905.
[6] Ignotz, R.A. and Massague, J. (1985) Type beta transforming [28] Ikeda, S., Kishida, S., Yamamoto, H., Murai, H., Koyama, S.
growth factor controls the adipogenic differentiation of 3T3 and Kikuchi, A. (1998) Axin, a negative regulator of the Wnt
fibroblasts. Proc. Natl. Acad. Sci. USA 82, 8530–8534. signaling pathway, forms a complex with GSK-3beta and beta-
[7] Ventura, F., Doody, J., Liu, F., Wrana, J.L. and Massagué, J. catenin and promotes GSK-3beta-dependent phosphorylation of
(1994) Reconstitution and transphosphorylation of TGF-beta beta-catenin. EMBO J. 17, 1371–1384.
receptor complexes. EMBO J. 13, 5581–5589. [29] Rubinfeld, B., Albert, I., Porfiri, E., Fiol, C., Munemitsu, S. and
[8] Kretzschmar, M., Liu, F., Hata, A., Doody, J. and Massagué, J. Polakis, P. (1996) Binding of GSK3beta to the APC-beta-catenin
(1997) The TGF-beta family mediator Smad1 is phosphorylated complex and regulation of complex assembly. Science 272, 1023–
directly and activated functionally by the BMP receptor kinase. 1026.
Genes Dev. 11, 984–995. [30] Kishida, S., Yamamoto, H., Ikeda, S., Kishida, M., Sakamoto, I.,
[9] Lagna, G., Hata, A., Hemmati-Brivanlou, A. and Massagué, J. Koyama, S. and Kikuchi, A. (1998) Axin, a negative regulator of
(1996) Partnership between DPC4 and SMAD proteins in TGF- the wnt signaling pathway, directly interacts with adenomatous
beta signalling pathways. Nature 383, 832–836. polyposis coli and regulates the stabilization of beta-catenin. J.
[10] Imamura, T., Takase, M., Nishihara, A., Oeda, E., Hanai, J., Biol. Chem. 273, 10823–10826.
Kawabata, M. and Miyazono, K. (1997) Smad6 inhibits signalling [31] Aberle, H., Bauer, A., Stappert, J., Kispert, A. and Kemler, R.
by the TGF-beta superfamily. Nature 389, 549–551. (1997) beta-catenin is a target for the ubiquitin-proteasome
[11] Souchelnytskyi, S., Nakayama, T., Nakao, A., Morén, A., pathway. EMBO J., 3797–3804.
Heldin, C.H., Christian, J.L. and ten Dijke, P. (1998) Physical [32] Li, L., Yuan, H., Weaver, C.D., Mao, J., Farr, G.H., Sussman
and functional interaction of murine and Xenopus Smad7 with 3rd, D.J., Jonkers, J., Kimelman, D. and Wu, D. (1999) Axin and
bone morphogenetic protein receptors and transforming growth Frat1 interact with Dvl and GSK, bridging Dvl to GSK in Wnt-
factor-receptors. J. Biol. Chem. 273, 25364–25370. mediated regulation of LEF-1. EMBO J. 18, 4233–4240.
[12] Pouponnot, C., Jayaraman, L. and Massagué, J. (1998) Physical [33] Axelrod, J.D., Miller, J.R., Shulman, J.M., Moon, R.T. and
and functional interaction of SMADs and p300/CBP. J. Biol. Perrimon, N. (1998) Differential recruitment of Dishevelled
Chem. 273, 22865–22868. provides signaling specificity in the planar cell polarity and
[13] Stroschein, S.L., Wang, W., Zhou, S., Zhou, Q. and Luo, K. Wingless signaling pathways. Genes Dev. 12, 2610–2622.
(1999) Negative feedback regulation of TGF-beta signaling by the [34] Lee, J.S., Ishimoto, A. and Yanagawa, S. (1999) Characterization
SnoN oncoprotein. Science 286, 771–774. of mouse dishevelled (Dvl) proteins in Wnt/Wingless signaling
[14] Hata, A., Seoane, J., Lagna, G., Montalvo, E., Hemmati- pathway. J. Biol. Chem. 274, 21464–21470.
Brivanlou, A. and Massagué, J. (2000) OAZ uses distinct DNA- [35] Yamanaka, H., Moriguchi, T., Masuyama, N., Kusakabe, M.,
and protein-binding zinc fingers in separate BMP-Smad and Olf Hanafusa, H., Takada, R., Takada, S. and Nishida, E. (2002)
signaling pathways. Cell 100, 229–240. JNK functions in the non-canonical Wnt pathway to regulate
[15] Labbe, E., Silvestri, C., Hoodless, P.A., Wrana, J.L. and convergent extension movements in vertebrates. EMBO J. 3, 69–
Attisano, L. (1998) Smad 2 and Smad 3 positively and negatively 75.
3546 D.R. Warner et al. / FEBS Letters 579 (2005) 3539–3546

[36] Boutros, M., Paricio, N., Strutt, D. and Mlodzik, M. (1998) [50] Kuhnert, F., Davis, C.R., Wang, H.T., Chu, P., Lee, M., Yuan,
Dishevelled activates JNK and discriminates between JNK J., Nusse, R. and Kuo, C.J. (2004) Essential requirement for Wnt
pathways in planar polarity and wingless signaling. Cell 94, signaling in proliferation of adult small intestine and colon
109–118. revealed by adenoviral expression of Dickkopf-1. Proc. Natl.
[37] Crease, D.J., Dyson, S. and Gurdon, J.B. (1998) Cooperation Acad. Sci. USA 101, 266–271.
between the activin and Wnt pathways in the spatial control of [51] Willert, K., Brown, J.D., Danenberg, E., Duncan, A.W., Weiss-
organizer gene expression. Proc. Natl. Acad. Sci. USA 95, 4398– man, I.L., Reya, T., Yates, J.R.r. and Nusse, R. (2003) Wnt
4403. proteins are lipid-modified and can act as stem cell growth factors.
[38] Labbé, E., Letamendia, A. and Attisano, L. (2000) Association Nature 423, 448–452.
of Smads with lymphoid enhancer binding factor 1/T cell- [52] Linask, K.K., DÕAngelo, M., Gehris, A.L. and Greene, R.M.
specific factor mediates cooperative signaling by the trans- (1991) Transforming growth factor-beta receptor profiles of
forming growth factor- and Wnt pathways. Proc. Natl. Acad. human and murine embryonic palate mesenchymal cells. Exp.
Sci. USA 97, 8358. Cell. Res. 192, 1–9.
[39] Nishita, M., Hashimoto, M.K., Ogata, S. and Laurent, M.N. [53] Sun, G.Y., Navidi, M., Yoa, F.G., Lin, T.N., Orth, O.E., Stubbs,
(2000) Interaction between Wnt and TGF-signalling pathways E.B.J. and MacQuarrie, R.A. (1992) Lithium effects on inositol
during formation of SpemannÕs organizer. Nature 403, 781–785. phospholipids and inositol phosphates: evaluation of an in vivo
[40] Ishitani, T., Ninomiya-Tsuji, J., Nagai, S., Nishita, M., Menegh- model for assessing polyphosphoinositide turnover in brain. J.
ini, M., Barker, N., Waterman, M., Bowerman, B., Clevers, H., Neurochem. 58, 290–297.
Shibuya, H. and Matsumoto, K. (1999) The TAK1-NLK- [54] Klein, T. and Arias, A.M. (1999) The vestigial gene product
MAPK-related pathway antagonizes signalling between b-catenin provides a molecular context for the interpretation of signals
and transcription factor TCF. Nature 399, 798–802. during the development of the wing in Drosophila. Development
[41] Furuhashi, M., Yagi, K., Yamamoto, H., Furukawa, Y., 126, 913–925.
Shimada, S., Nakamura, Y., Kikuchi, A., Miyazono, K. and [55] Riese, J., Yu, X., Munnerlyn, A., Eresh, S., Hsu, S.C., Gros-
Kato, M. (2001) Axin facilitates Smad3 activation in the schedl, R. and Bienz, M. (1997) LEF-1, a nuclear factor
transforming growth factor signaling pathway. Cell. Biol. 21, coordinating signaling inputs from wingless and decapentaplegic.
5132–5141. Cell 88, 777–787.
[42] Warner, D.R., Pisano, M.M., Roberts, E.A. and Greene, R.M. [56] Labbé, E., Letamendia, A. and Attisano, L. (2000) Association of
(2003) Identification of three novel Smad binding proteins Smads with lymphoid enhancer binding factor 1/T cell-specific
involved in cell polarity. FEBS Lett. 539, 167–173. factor mediates cooperative signaling by the transforming growth
[43] Macias-Silva, M., Abdollah, S., Hoodless, P.A., Pirone, R., factor- and Wnt pathways. Proc. Natl. Acad. Sci. USA 97, 8358–
Attisano, L. and Wrana, J.L. (1996) MADR2 is a substrate of the 8363.
TGFbeta receptor and its phosphorylation is required for nuclear [57] Tuli, R., Tuli, S., Nandi, S., Huang, X., Manner, P.A., Hozack,
accumulation and signaling. Cell 87, 1215–1224. W.J., Danielson, K.G., Hall, D.J. and Tuan, R.S. (2003)
[44] Shibamoto, S., Higano, K., Takada, R., Ito, F., Takeichi, M. and Transforming growth factor-b-mediated chondrogenesis of hu-
Takada, S. (1998) Cytoskeletal reorganization by soluble Wnt-3a man mesenchymal progenitor cells involves N-cadherin and
protein signalling. Genes to Cells 3, 659–670. mitogen-activated protein kinase and Wnt signaling cross-talk.
[45] Korinek, V., Barker, N., Morin, P.J., van Wichen, D., de Weger, J. Biol. Chem. 278, 41227–41236.
R., Kinzler, K.W., Vogelstein, B. and Clevers, H. (1997) Consti- [58] Waltzer, L. and Bienz, M. (1998) Drosophila CBP represses the
tutive Transcriptional Activation by a -Catenin-Tcf Complex in transcription factor TCF to antagonize Wingless signalling.
APC/ Colon Carcinoma. Science 275, 1784–1787. Nature 395, 521–525.
[46] Wrana, J.L., Attisano, L., Carcamo, J., Zentella, A., Doody, J., [59] Hecht, A., Vleminckx, K., Stemmler, M.P., van Roy, F. and
Laiho, M., Wang, X.F. and Massagué, J. (1992) TGF beta signals Kemler, R. (2000) The p300/CBP acetyltransferases function as
through a heteromeric protein kinase receptor complex. Cell 71, transcriptional coactivators of beta-catenin in vertebrates. EMBO
1003–1014. J. 19, 1839–1850.
[47] Stambolic, V., Ruel, L. and Woodgett, J.R. (1996) Lithium [60] Ito, Y., Yeo, J.Y., Chytil, A., Han, J., Bringas, P.J., Nakajima, A.,
inhibits glycogen synthase kinase-3 activity and mimics Wingless Shuler, C.F., Moses, H.L. and Chai, Y. (2003) Conditional
signalling in intact cells. Curr. Biol. 6, 1664–1668. inactivation of Tgfbr2 in cranial neural crest causes cleft palate
[48] Hu, M.C., Piscione, T.D. and Rosenblum, N.D. (2003) Elevated and calvaria defects. Development 130, 5269–5280.
SMAD1/b-catenin molecular complexes and renal medullary [61] Zhang, Z., Song, Y., Zhao, X., Zhang, X., Fermin, C. and Chen,
cystic dysplasia in ALK3 transgenic mice. Development 130, Y. (2002) Rescue of cleft palate in Msx1-deficient mice by
2753–2766. transgenic Bmp4 reveals a network of BMP and Shh signaling in
[49] G.J., H., D. Beach, p15INK4B is a potential effector of TGF- the regulation of mammalian palatogenesis. Development 129,
beta-induced cell cycle arrest. Nature 371 (1994) 257–261. 4135–4146.