Current Drug Targets, 2006, 7, 1411-1419 1411
1389-4501/06 $50.00+.00 © 2006 Bentham Science Publishers Ltd.
GSK3 at the Edge: Regulation of Developmental Specification and Cell Polarization
Leung Kim1 and Alan R. Kimmel2,*
1Department of Biological Sciences, Florida International University, Miami, FL 33199, USA and 2Laboratory of Cellular and
Developmental Biology, NIDDK, National Institutes of Health, Bethesda, MD 20892-8028, USA
Abstract: GSK3 is a multifunctional protein kinase that is pivotal for the regulation of metabolism, the cytoskeleton, and gene expres-
sion. Multicellular eukaryotes utilize GSK3 as a molecular switch to specify distinct cell fates, but also to organize these cells spatially
within the developing organism. We discuss the central role of GSK3 in control of the Wnt, Hedgehog, cAMP (in Dictyostelium), and
other signaling pathways, but also focus on significant new evidence that GSK3 is required to establish cell polarity.
Key Words: Wnt, Frizzled, Hedgehog, Presenilin, PAR, Cell Motility, Chemotaxis, Dictyostelium.
As inferred from its original name glycogen synthase kinase 3,
GSK3 function was initially associated with metabolic pathways.
But, it is now clear that GSK3 activity is required for multicellular
development and for a broad range of cellular functions, including
stem-cell renewal and differentiation, apoptosis, and even circadian
rhythm. Aberrations in GSK3 regulation can promote tumorigene-
sis, neurodegeneration, diabetes and insulin resistance, among other
abnormalities (see [1-5]).
Mammalian cells have three closely related forms of GSK3,
GSK3α, GSK3β, and the minor splice variant GSK3β2. Their cata-
lytic domains are nearly identical and they share many modes of
regulation. Two sites for phosphorylation of GSK3 directly affect
kinase activity. Phosphorylation at S9 of GSK3β (S21 on GSK3α)
by Akt, and other protein kinases, is specifically inhibitory. In gen-
eral, GSK3 requires a “priming” phosphorylation at position +4 in
the downstream position of the optimal consensus sequence
(SXXXpS) for substrate recognition. S9 inhibition appears to func-
tion as an intramolecular pseudo-substrate, blocking access to the
catalytic site  of these “primed” substrates . In the absence of
Akt activation during insulin resistance, GSK3 is active and consti-
tutively represses glycogen synthase. In addition, active GSK3 will
phosphorylate and inhibit IRS-1 (insulin receptor substrate-1), an
essential component of the insulin signaling pathway . Loss of
GSK3 repression exacerbates insulin resistance. In contrast to the
inhibitory action of Akt at S9, tyrosine phosphorylation at Y216 in
the activation loop of GSK3β (Y279 of GSK3α) enhances kinase
activity. While the activation is modest as compared to that usually
observed with other kinases , an explanation became apparent
with the elucidation of the GSK3 structure. In the unphosphorylated
protein, the activation loop is already in an open configuration.
Tyrosine phosphorylation may serve to reposition Y216 through the
interaction with a proximate arginine to increase substrate access
and kinase activity . Data from Dictyostelium (c.i.) indicate that
additional phosphorylation at Y222 is essential for the full activa-
tion of GSK3 .
Alteration in subcellular localization of GSK3 is the other ma-
jor regulatory mode. GSK3 does not have specific subcellular local-
ization motifs, but various extracellular signals can impact the in-
teraction of GSK3 with its target substrates. Other signals influence
the levels of GSK3 in the nucleus, where it can phosphorylate a
variety of transcription factors. Many of these are inhibited, as
*Address correspondence to this author at the Laboratory of Cellular and
Developmental Biology, NIDDK, National Institutes of Health, Bethesda,
MD 20892-8028, USA; Tel: (301) 496-3016; Fax: (301) 496-5239;
phosphorylation by GSK3 can promote nuclear export. The most
widely studied is NFAT (Nuclear Factor of Activated T-cells), a
central regulator of thymic development and immune response,
cardiac morphogenesis, and neural survival [9, 10].
Despite the apparent essential role of GSK3 throughout devel-
opment it came as an initial surprise that GSK3β-deficient mice did
not exhibit an embryonic lethality until mid gestation . Most
certainly, progression through early embryogenesis was “rescued”
by the endogenous expression of GSK3α in these mice. Nonethe-
less, lethality is the result of a liver failure that is consistent with a
hypersensitivity to TNF due to loss of NF-κB activation. The ob-
servations defined a previously unappreciated requirement of GSK3
function and a link to apoptosis and other NF-κB regulated path-
ways [11-13]. The GSK3β-deficient mice provided an additional
view to GSK3 function in vivo. gsk3β+/- mice, which are haploinsuf-
ficient for GSK3β, mimic the behavioral effects observed upon
treatment of control animals with clinical doses of lithium, a GSK3
inhibitor . Lithium has been one of the effective treatments for
human bipolar disorder and these results provide strong support that
a primary target of action may be GSK3.
The variety of cellular and developmental processes that are
controlled by GSK3 is indeed too vast to detail. Rather, this review
will focus on selected GSK3-dependent pathways that have been
well described at the molecular level. Particular attention will be
directed toward mechanisms that organize cell fate specification
and regulate cell polarity and directed cell movement. Since GSK3
is of such critical import, it is subject to continuous discussion and
review; accordingly, emphases here will be limited to the most
GSK3-LINKED PATHWAYS THAT SPECIFY DEVELOP-
Canonical Wnt Signaling and Complexity Beyond
The Wnts are a large family of secreted, lipid-modified, glyco-
proteins  that regulate major developmental and proliferative
events in the metazoa [4, 5, 16]. Loss of Wnt signaling in mammals
disrupts embryonic axis formation, gastrulation, and CNS (central
nervous system), thymic, and limb development. Stem cell prolif-
eration is also regulated by Wnt signaling, and mutations that lead
to the constitutive activation of the pathway promote tumorigenesis
[1-5, 16, 17]. Wnt signaling is multi-facetted and broadly classified
as either canonical or non-canonical and can directly impact tran-
scriptional as well as cytoskeletal processes.
Canonical Wnt signaling functions primarily by regulating
GSK3 to promote the rapid, cytosolic accumulation of the multi-
functional protein β-catenin (Fig. 1). In quiescent cells, β-catenin is
principally found at the cell cortex in association with cytoskeletal
1412 Current Drug Targets, 2006, Vol. 7, No. 11 Kim and Kimmel
components (e.g. E-cadherin), where it functions to orchestrate
actin dynamics and regulate cell adhesion, cell-cell interactions, and
intercellular communication. Unstimulated cells (Fig. 1A) possess
relatively low levels of cytosolic β-catenin which, in general, is
sequestered to multi-protein scaffolds anchored by Axin and Axin2
(conductin). Within these Axin complexes, β-catenin is phosphory-
lated by CK1α (Casein Kinase I), priming it for additional phos-
phorylations by GSK3β [18, 19]. Polyphosphorylated β-catenin is
then subject to proteasomal destruction [20, 21] that is mediated by
APC (Adenomatous Polyposis Coli) and βTrCP . These Axin
scaffolds are, thus, often denoted as destruction complexes, and
although Axin is also a kinase substrate when complexed with
GSK3, Axin does not appear to be de-stabilized upon phosphoryla-
tion by GSK3.
In addition to its role as a cytoskeletal component, β-catenin
can also function as a transcriptional co-factor; complexes of β-
catenin and Lef/Tcf transcription factors serve as nuclear activators
[22-24] of gene sets that promote specific developmental pathways
or tumorigenesis . Canonical Wnt signaling serves to inhibit β-
catenin phosphorylation and thus promotes the accumulation of an
uncomplexed, cytosolic population of β-catenin that is available for
interaction with and activation of Lef/Tcf family members (Fig.
In the canonical pathway, extracellular Wnt signals are detected
at the cell surface by both the seven-transmembrane Frizzled (Fz)
[25, 26] and the single-pass, LDL receptor-related proteins (LRP)
5/6 [27-29] that act as co-receptors (Fig. 1B). Data suggest that Wnt
treatment disassembles the Axin/GSK3β complexes (see ), thus
decreasing GSK3β phosphorylation of both β-catenin and Axin.
While unphosphorylated β-catenin is consequently stabilized, un-
phosphorylated Axin is subject to proteasomal degradation, further
reducing Axin/GSK3β complex levels [30-33].
While the precise molecular events that transmit Wnt signals
are not well characterized, many additional components are in-
volved (see ). Principal to this is Dishevelled (Dvl/Dsh), which
functions downstream of Fz/LRP but upstream of GSK3 to mediate
β-catenin stabilization and may play an essential, but as yet unde-
fined, role in the disruption of the Axin destruction complex [34-
The structural similarity of Fz to bona fide G protein coupled
receptors had led to early speculation that heterotrimeric Gαβγ
protein signaling may participate in Wnt regulation of GSK3. Al-
though there is good evidence that Fz receptors may interact with G
proteins for Ca+2 signaling, a non-canonical, non-GSK3 Wnt path-
way , until recently, definitive evidence [38, 39] for G protein-
dependent, β-catenin signaling had been elusive. Epistasis experi-
ments in Drosophila now provide strong support that Gαo acts
downstream of Fz, but upstream of Dsh during Wnt regulation of
GSK3/β-catenin signaling ; further, activated Gαo-GTP appears
to function independently of receptor activation in this pathway
. In addition, we have shown that depletion of Gαo and Gαq in
mammalian cells with siRNA will inhibit Wnt-induced stabilization
of β-catenin [40a] and that direct activation of G proteins with
GTPγS can promote β-catenin accumulation in the absence of Wnt
stimulation [40a]. These data are consistent with an intermediary
role for G proteins in signal transduction from Fz/LRP to
Axin/GSK3. Yet, it is not clear why G protein genes had not been
identified previously in various genetic screens for Wnt modifiers.
There may be regulatory paths that potentiate β-catenin stabiliza-
tion independently of Gα-signaling. Potentially, LRP5/6 can trans-
duce a Wnt signal directly to Axin/GSK3 [41-44] by a mechanism
that requires interaction with Fz, but that bypasses Gα activation.
Other evidence exists for a mechanistic parallel. Fz receptors are
characterized by their extended, extracellular N-terminal, cysteine-
rich domain (CRD). Although the CRD is the Wnt-binding region
of Fz, in vivo data in Drosophila indicate that the CRD is not en-
tirely required for Wnt signaling . The direct binding of Wnt to
Fz may not be absolutely essential for signal transduction, but per-
haps can be compensated by interaction with other components of
the receptor complex. To this extent, disruption of Fz/Gα-signaling
may have a less severe effect on Wnt response during development
than would be the loss of an individual Wnt, Fz, or LRP gene mem-
ber. Finally, it should be emphasized that additional non-canonical
Wnt receptor systems (e.g. Ryk, Ror) may function independently
of Fz/Gα-signaling and, thus, also bypass G protein dependency
[46-51]. Still, definitive downstream linkages of these signaling
pathways have not been established.
The Hedgehog (Hh) proteins constitute another family of se-
creted proteins that regulate a broad variety of growth and devel-
opmental processes through the targeted regulation of GSK3 func-
tion. Hh signaling was first described in Drosophila but is now
understood to be essential throughout vertebrate development. Dis-
ruption of component genes in humans leads to neural defects, tu-
morigenesis, and other severe and complex syndromes (see ). In
general Hh and Wnt responses occur in non-overlapping cell popu-
lations, yet their signaling mechanisms have remarkable parallels.
In the absence of Hh protein, the 155 kDa, zinc-finger transcrip-
tion factor Ci [cubitus interruptus in Drosophila; Gli (from glioblas-
toma) in vertebrates] is subject to proteolytic cleavage mediated by
βTrCP/Slimb (supernumerary limbs) in concert with the protea-
some pathway (see ). The resulting product is an N-terminal Ci-
75 fragment that retains the zinc-finger domain, but that functions
largely as a transcriptional repressor. Hh signaling inhibits Ci proc-
essing, thus directing the accumulation of full-length Ci-155 and
the transcriptional activation of Hh-response genes. While it had
been clear that phosphorylation of Ci by the cAMP-dependent pro-
tein kinase (PKA) was essential for proteolytic cleavage in the ab-
sence of Hh, mechanistic links were only poorly understood. Now,
several recent studies have proven an essential role for GSK3 in Hh
signaling and suggest a new model for its antagonistic regulation
In unstimulated cells, Ci-155 is associated with a cytoplasmic
scaffolding complex consisting of Cos2, Fused (Fu), PKA, GSK3,
and CKI (Fig. 2A). Within the complex, PKA phosphorylates Ci-
155 at several sites [54, 55, 61]; each of these in turn serves as a
priming phosphorylation for recognition by both GSK3 and CKI
(Fig. 2B). Hyper-phosphorylation of Ci-155 by all 3 kinases is re-
quired for the proteolytic processing that generates Ci-75. Mutation
of the PKA, the GSK3, or the CKI sites within Ci-155 prevents
proteolysis and mimics activation by Hh [54, 55]. In addition, muta-
tion of Cos2 also suppresses Ci-75 production, presumably through
disruption of the Cos2-Ci-kinase destruction complex.
Two transmembrane proteins, Patched and Smoothened (Smo),
function to transduce the Hh signal (Fig. 2A). Patched is a 12-
transmembrane domain protein and the presumed Hh receptor .
Historically, Patched has been denoted Ptc, but recently several
workers have adopted Ptch to distinguish it from PTC, the bitter-
taste receptor of humans. In the absence of an Hh ligand, Ptc/Ptch is
suggested to inhibit the function of the 7-transmembrane domain
protein Smo (Fig. 2A). Inhibition is still poorly understood mecha-
nistically, but appears to require phosphorylation-dependent events
that promote destabilization and, perhaps, localization to intracellu-
lar vesicles. Although the phosphorylation sites on Smo are re-
markably similar in sequence to those of Ci, phosphorylation of
Smo by GSK3 is not required for function [53, 57]. In contrast,
Smo proteins that lack either the 3 consensus PKA or 3 consensus
CKI sites are not subject to Hh-independent degradation, and are
also unable to support signaling in the presence of Hh (v.i.).
Hh binding to Ptc/Ptch stimulates Smo hyper-phophorylation
and relieves Smo inhibition (Fig. 2C), potentially through stabiliza-
GSK3 and Development Current Drug Targets, 2006, Vol. 7, No. 11 1413
tion and recruitment to the cell surface [56, 58-60]. Activated Smo
is suggested to bind Cos2 in a manner that disassembles the Cos2-
Ci-kinase destruction complex. As a consequence, Ci-155 phos-
phorylation and proteolysis are diminished, production of Ci-75 is
attenuated, and Ci-155 accumulates to promote expression of Hh-
response genes. The similarity in phosphorylated sequences shared
by Smo and Ci suggests a simple mechanism whereby activated,
hyper-phosphorylated Smo may displace and, hence, release Ci-155
from association with the Cos2-kinase complex. Phosphorylation of
Smo by both PKA and CKI (but not by GSK3) is required for Hh-
induced accumulation of Ci-155.
Significant parallels are easy drawn between the Hh and ca-
nonical Wnt pathways. Both utilize GSK3 to regulate the stability
of a core transcription factor, Ci for Hh and β-catenin for Wnt. Un-
der unstimulated conditions, GSK3 is functionally active, directing
the phosphorylation and proteolysis of Ci/β-catenin. Ligand en-
gagement blocks phosphorylation in both pathways, but in neither
is GSK3 activity inhibited, per se; rather, signal transduction serves
to disassemble the Cos2 and Axin destruction complexes releasing
GSK3 from association with its substrate targets and its dependent
priming kinases. One additional link between the Hh and Wnt paths
should be noted. The signal-transducing proteins for each, Smo and
Fz, respectively, share high sequence similarity and belong to the
same subfamily of 7-transmembrane receptors.
Alzheimer’s disease is marked by the elevated accumulation of
the 40 and 42 amino acid amyloid β (Aβ) peptides. Mutation of the
gene for Presenilin (PS) is a major cause of familial, early-onset
Alzheimer’s disease; the PS/Alzeimer’s connection is now well
established. Aβ peptides are generated by sequential cleavages of
the amyloid precursor protein (APP), a single-span transmembrane
protein. The first proteolytic cleavage releases the extracellular
domain and facilitates subsequent intramembranous proteolysis by
γ-secretase; γ-secretase activity requires PS, as well as nicastrin,
Aph1, and Pen2, but the catalytic activity appears to reside within
PS. Dominant mutations in PS can alter the sites for intramembrane
cleveage of APP and increase the relative proportion of Aβ42, the
presumed pathogenic product.
PS is a multi-transmembrane protein that, in addition to exhibit-
ing proteolytic activity, possesses binding domains for both GSK3
and β-catenin. Recent studies have focused on the relationship be-
tween GSK3 and γ-secretase activities and a potential role for PS as
a multi-functional scaffold of GSK3.
Fig. (1). GSK3 regulation of β β-catenin degradation.
A. In the absence of a Wnt ligand, β-catenin is associated with an Axin scaffold in complex with GSK3. Phosphorylation by GSK3 promotes the proteasomal-
mediated degradation (X) of β-catenin. Transcriptional paths that require β-catenin for activation are, thus, inhibited (X). See text for details and additional
B. Wnt binding to the Fz/LRP co-receptor complex leads to the disassembly of the Axin destruction complex, by a process mediated by Dsh. The precise re-
localizations of these various components have not been established. β-catenin is no longer associated with or phosphorylated by GSK3. β-catenin degradation
is attenuated, and accumulated β-catenin is now able to associate with Tcf transcription factors to activate gene expression.
C. A potential β-catenin destruction complex mediated by PS scaffolding and insensitive to Wnt signaling. β-catenin is subject to phosphorylation and degra-
1414 Current Drug Targets, 2006, Vol. 7, No. 11 Kim and Kimmel
Briefly, pharmacological inhibition of GSK3 specifically inhibits
the processing of APP and the accumulation of Aβ peptides. This is
not caused by a global inhibition of γ-secretase, since GSK3 is not
required for the proteolysis of other PS substrates (e.g. Notch, v.i.
[62, 63]), and is not mimicked by overexpression of β-catenin, a co-
factor that may accumulate under conditions of GSK3 inhibition
(v.s.). Further, directed siRNA approaches indicate a requirement
for GSK3α, but not GSK3β, in APP proteolysis, underscoring a
unique and restricted association of GSK3 to PS function. In addi-
tion, although GSK3β is not required for Notch processing, phos-
phorylation of the intracellular domain of Notch by GSK3β does
decrease its stability and suppresses its function (v.i. ).
Linkage between PS and GSK3/β-catenin signaling has been
less clear. PS was identified in genetic screens for novel modifiers
of the Wnt (Wg) pathway in Drosophila [64, 65] and data suggest
that PS acts as a negative regulator of the pathway. PS has also
been identified in a multi-protein complex with PKA, GSK3, and β-
catenin in mammalian cells . Here, again it appears to antago-
nize Wnt signaling (Fig. 1C). In this context, PS functions analo-
gously to Axin, scaffolding β-catenin to GSK3 and a priming
A role for GSK3 in APP processing has been confirmed . kinase (PKA). This complex would promote the de-stabilization of
β-catenin, but would be unresponsive to disassembly by Wnt. How-
ever, depletion of PS (or PKA) by RNAi did not alter β-catenin
(armadillo) levels in Drosophila tissue culture cells .
PS cleaves a diverse group of single-span transmembrane pro-
teins. Other than APP, Notch is among the most well characterized.
PS processing of Notch releases the intracellular C-terminal domain
(ICD) that functions as an essential transcriptional activator for
embryogenesis, and ps1-/-ps2-/- mice carrying mutations that fully
ablate PS function exhibit an embryonic Notch-like lethality. Thus,
it has been difficult to analyze an in vivo role for PS in GSK3/β-
catenin regulation in mammals. However, ps1+/-ps2-/- mice that are
partially deficient for PS survive and their derived fibroblasts ex-
hibit significantly elevated levels of β-catenin as compared with
wild-type controls . These data are consistent with a scaffold
function for PS that promotes GSK3-mediated degradation of β-
catenin (Fig. 1C).
Although most developmental control appears to be directed
towards the negative regulation of GSK3, two organisms, Cean-
Fig. (2). GSK3 regulation of Ci-155 processing.
A. In the absence of an Hh ligand, Ci-155 is in a Cos2 scaffold complex with GSK3, PKA, CKI, and Fu. Phosphorylation of Ci-155 by GSK3, PKA, and CKI
promotes Ci-155 proteolysis and the production of C-75. (See text for details and additional components.)
B Consensus sites for Ci-155 phosphorylation by PKA, GSK3, and CKI. PKA recognizes the sequence RRXS, creating a priming phosphorylation recognized
by GSK3 (at position +4 within the consenus site, SXXXpS) and by CKI (at position -3, pSXXS).
C. Upon Hh binding to Ptc, Smo becomes activated and promotes disassembly of the Cos2 destruction complex. Ci-155 is no longer associated with or phos-
phorylated by GSK3. Ci155 degradation is attenuated and now activates expression of the Hh-responsive genes.
GSK3 and Development Current Drug Targets, 2006, Vol. 7, No. 11 1415
horabditis elegans and Dictyostelium discoideum, utilize an activat-
ing signal to specify cell fate determination (Fig. 3).
In the nematode C. elegans, endoderm and mesoderm fates
derive from an EMS precursor cell, which arises after the second
cell division of the embryo. The EMS cell is oriented along an ante-
rior/posterior axis and its subsequent division generates an anterior
MS and a posterior E cell that, respectively, define mesoderm and
endoderm lineages. These cell fates are regulated, at least in part,
by Wnt signal induction from the neighboring P2 cell [69-72]. Two
mutations that inhibit E cell differentiation define the pathway.
MOM-2 (more mesoderm) is a Wnt homolog produced in P2 cells,
and MOM-5 is a Fz receptor expressed in EMS cells (Fig. 3). In
addition, a Dsh protein is required for the intracellular transduction
of the Wnt signal. However, Wnt signaling through Fz and Dsh
does not repress GSK3 function as in the canonical pathway, but
requires active GSK3. Thus, disruption of GSK3 in C. elegans does
not promote endoderm induction as would occur if GSK3 were
inhibited by Wnt, but produces a MOM (more mesoderm) pheno-
type (Fig. 3) that parallels the loss-of-function mutations in mom-2
(Wnt) and mom-5 (Fz). β-catenin and Tcf homologs function down-
stream of GSK3, but again their genetic interactions differ from that
of the canonical Wnt pathway [69-72]. Despite the requirement for
active GSK3 during E/MS specification, C. elegans utilizes canoni-
cal Wnt pathway members, including PRY-1, a highly diverged but
functionally active Axin ortholog, to regulate neuroblast and vulval
Dictyostelium provides further evidence for an activating path-
way for GSK3 during cell fate determination (Fig. 3). There are 2
major cell classes at terminal differentiation of Dictyostelium.
These are the spore and stalk cells which derive from the non-
terminally differentiated prespore and prestalk progenitors (see
). GSK3 acts as developmental switch to establish these cell
fates (Fig. 3).
In Dictyostelium, GSK3 is not regulated by a Wnt signal (mak-
ing it extremely non-canonical). Instead, the prestalk/prespore axis
is specified by signaling via secreted cAMP that binds and activates
a family of 7-transmembrane domain, cell-surface receptors
(CARs), which are distantly related to Fz. CAR3, the related tyro-
sine kinases ZAK1 and ZAK2, and GSK3 are required for prespore
cell fate differentiation, but inhibit prestalk differentiation ([8, 17,
74-77] Kim and Kimmel, unpublished). ZAK1/2 are activated by
cAMP/CAR3 and, in turn, ZAK1/2 are required to phosphorylate
and activate GSK3. The sites phosphorylated by ZAK1 have been
mapped to the activation loops of both Dictyostelium GSK3 and
mammalian GSK3β and these phospho-tyrosines are required for
GSK3 activation in vivo and in vitro [8, 75]. It should be noted that
a Src-related tyrosine kinase functions in concert with GSK3 to
activate MS differentiation in C. elegans . Although the target
of Src has not yet been determined, GSK3 is a potential candidate
that may be subject to an activating tyrosine phosphorylation in this
pathway (Fig. 3).
In contrast to the role of CAR3, Dictyostelium that are deficient
for CAR4 exhibit the phenotypic loss of prestalk cell gene expres-
sion and expansion of prespore gene expression (Fig. 3). car4-nulls
also have persistent tyrosine phosphorylation and hyper-activation
of GSK3 [8, 17, 74, 75]. Developmental defects of car4-nulls can
be partially rescued by biochemical and genetic reduction of GSK3
activity in vivo, indicating that GSK3 is a downstream negative
target of CAR4. The absence of CAR4 has no affect on cAMP/
CAR3 regulated activation of ZAK1, indicating that CAR4 does not
Fig. (3). The role of GSK3 for cell fate determination in C. elegans and Dictyostelium.
In contrast to the inactivating signal transmitted to GSK3 in the canonical Wnt pathway (see Fig. 1), C. elegans requires an active GSK3 to regulate EMS
fates. See text for details.
In Dictyostelium, GSK3 acts as developmental switch to establish the fates of the prespore and prestalk cells, the two major developmental progenitors. Acti-
vated GSK3 is required for the prespore pathway, whereas de-activated GSK3 promotes prestalk differentiation. The data are consistent with phosphoryla-
tion/de-phosphorylation of GSK3 as the mechanism for activation/de-activation of GSK3. The different levels of tyrosine phosphorylation of GSK3 resulting
from the antagonistic actions of the tyrosine kinases ZAK1/2 or the CAR4-PTPase constitute the core of the regulatory machinery on GSK3 activity in the
context of cell fate decision regulated by the 7-TM CARs. See text for details.
1416 Current Drug Targets, 2006, Vol. 7, No. 11 Kim and Kimmel
inhibit GSK3 by repressing ZAK1, the activating tyrosine kinase of
GSK3. Rather, CAR4 activates a specific PTPase. This has been
confirmed in biochemical assays using a specific phosphoGSK3
substrate . These data are consistent with phosphorylation/de-
phosphorylation as a mechanism for activation/de-activation of
GSK3. The different levels of tyrosine phosphorylation of GSK3
resulting from the antagonistic actions of the tyrosine kinases
ZAK1/2 or the CAR4-PTPase constitute the core of the regulatory
machinery on GSK3 activity in the context of cell fate decision by
the 7-TM CARs in Dictyostelium (Fig. 3).
GSK3 AND CELL POLARITY
Analysis of P2/EMS interactions in C. elegans provided new
perspectives into GSK3 regulation of cell specification. However,
Wnt signaling to the EMS cells has complexity beyond the control
of transcriptional events [69, 78]. Prior to cell division, the EMS
precursor is organized along an anterior/posterior (A/P) axis. Ini-
tially however, the mitotic spindle is not aligned with the A/P axis.
The spindle must rotate during mitosis to adopt the appropriate A/P
orientation. Spindle rotation and re-orientation relies on the same
upstream members of the MOM-2/MOM-5 (i.e. Wnt/Fz) signaling
pathway that specify E/MS fates (see Figs. 3 and 4). Both pathways
require, but diverge at GSK3. Thus, spindle rotation occurs inde-
pendently of activated gene expression.
Other developmental aspects in C. elegans have demonstrated a
requirement for GSK3 in organizing cell polarity. Early embryo-
genesis in C. elegans involves a series of asymmetric cell divisions.
“par” embryos that are defective in partitioning are unable to un-
dergo normal asymmetric divisions, with resulting defects in differ-
entiation. Genetic screens for such abnormalities identified six PAR
genes essential to polarize cells for partitioning. Five of these genes
are conserved in mammals, where they also regulate cell polariza-
Par6 and Par3 are PDZ domain-containing proteins that form a
heterotrimer with aPKCζ (an atypical protein kinase C). Mutations
in any of these genes disrupt cell polarity in C. elegans, Drosophila,
and vertebrates. The Par6 complex [79, 80] is intracellularly polar-
ized to the anterior pole of the C. elegans zygote, but also to the
cell-cell junctions of epithelial cells where it regulates api-
cal/basolateral specificity, to nascent axons on neurites, and in cer-
tain cells migrating toward a directional signal.
Astrocyte migration involves cellular protrusions that are driven
by microtubule elements. In “scratch” assays that mimic wound
healing, primary astrocytes re-orient and migrate to repair the abra-
sion. Scratch-induced migration causes a re-orientation of the cen-
trosome towards the leading edge and the formation of microtubule
protrusions. cdc42 and Rac1, members of the Rho family of small
GTPases, also become activated and re-localized to the leading
edge of migrating astrocytes. Active, GTP-bound cdc42 will induce
cytoskeletal rearrangements through interaction with WASp (Wis-
cott-Aldrich syndrome protein) and PAK and their reorganization
of actin filaments. But, cdc42 has not been generally associated
with establishing microtubule polarity. Recent data now connect
cdc42 to microtubule assembly through the regulation and function
of Par6/PKCζ and GSK3 (Fig. 4).
Par6 binding to cdc42 at the leading edge promotes the activa-
tion of the associated aPKCζ leading to the phosphorylation of
GSK3β at S9 and its consequent inactivation [80, 81]. Inhibition of
cdc42 or of aPKCζ blocks GSK3β phosphorylation, microtubule
orientation, and astrocyte migration. Expression of the non-
phosphorylatable, S9A variant of GSK3β is similarly insensitive to
scratch induction. Remarkably, expression of a constitutively-active
(GTP-locked) cdc42 or a kinase-inactive GSK3β also disrupts cell
polarization [80, 81]. The global inhibition of GSK3β does not
prevent cell protrusions, but it induces them along the entire cell
periphery, emphasizing the significance of localized activation and
(GSK3 inhibition) for cell polarization.
edge. In addition to its role as a mediator of β-catenin degradation
during canonical Wnt signaling (see Fig. 1), APC can associate
with the plus-ends of microtubules to regulate their stability and
dynamics. Expression of an APC fragment that lacks a microtubule
interacting domain blocks centrosome re-orientation in migrating
astrocytes [80, 81].
Localized inhibition of GSK3β by Par6/aPKCζ is also required
to establish and maintain neuronal polarity (Fig. 4 [82-87]). Axons
and dendrites constitute distinct structural elements that define the
polarized neuron. The Par6/aPKCζ complex localizes to a single tip
of the immature neurite and ultimately specifies the axon . An
APC fragment, which will interact with microtubules but not with
phosphorylated GSK3, acts as a dominant-negative for axon forma-
tion. Likewise, expression of the non-phosphorylatable (“constitu-
tively-active”) S9A GSK3β will also inhibit axon formation,
whereas global inhibition of GSK3 promotes multiple axons and
transforms pre-existing dendrites into new axons . Semaphorin
3A (Sema 3A), a molecule that suppresses axonal growth, promotes
GSK3 de-phosphorylation and its consequent activation at the lead-
ing edge of the growing neuron .
Studies on another Par protein have provided an additional
linkage for GSK3 regulation of cell polarity. Several mammalian
epithelial cell culture models will spontaneously polarize when
confluent. Apical brush borders form and sort from basolateral
markers. Par6/Par3 and cdc42-GTP modulate this process, but an-
other Par protein, Par4, is also essential [88, 89]. The Par-4 ortholog
in humans is LKB1, a tumor suppressor linked to Peutz-Jeghers
cancer syndrome. LKB1 is activated by two proteins, STRAD and
MO25, and the induced expression of STRAD will activate LKB1
and rapidly remodel and polarize individual epithelial cells prior to
confluence. Conversely, depletion of LKB1 in these epithelial cells
will delay the spontaneous formation of apical/basolateral polarity
LKB1 is itself multifunctional. Like Par6, it can associate with
aPKCζ to promote the phosphorylation and inhibition of GSK3β.
Although it has been controversial whether enzymatic inhibition of
GSK3β may synergize with the Wnt effects on GSK3 (see Fig. 1) to
increase β-catenin stabilization (see ), recent data clearly indi-
cate that LKB1-mediated phosphorylation of GSKβ will activate β-
catenin dependent transcriptional pathways [90, 91]. Part of the
LKB1 effect could be mediated by the Par1 kinase, a substrate of
LKB1. Indeed, Par1 is suggested to facilitate Wnt signaling .
However, a related study found that LKB1 expression appeared to
antagonize the pathway rather than activate it . Perhaps addi-
tional complexities explain the apparent contradictions . There
are multiple Par1 kinases and multiple functions for the individual
Par1 proteins. Activation by LKB1 may redirect Par1 localization
or complex formation. It should also be noted that AMPK (AMP
protein kinase) is a member of the Par1 family and a downstream
target of LKB1. AMPK is also positive regulator of TOR (target of
rapamycin), and although new TOR complexes have been identified
that regulate the actin cytoskeleton and cell polarity , the direct
input of LKB1 and AMPK into these TOR processes has not been
Downstream of GSK3, APC is also recruited to the leading
GSK3 participates in multiple pathways that regulate cell fate
specification and cell polarization. While these events have been
considered distinct, the most recent results suggest an intimate in-
terrelationship between them. It may now be prudent to re-examine
several non-canonical Wnt pathways that have been thought to
function independently of GSK3 . Planar cell polarity (PCP)
establishes wing bristle and ommatidia polarity in Drosophila and
convergent movement during vertebrate gastrulation. PCP is regu-
lated through Fz/Dsh, but does not require β-catenin stabilization.
Mis-expression of GSK3 can also cause polarity defects , but
GSK3 and Development Current Drug Targets, 2006, Vol. 7, No. 11 1417
whether this is related to an alteration in PCP, to mis-regulation of
Par, or an interaction between the pathways remains to be explored.
Astrocyte migration differs widely from another type of di-
rected cell movement, chemotaxis (see ). Yet, a linkage to
GSK3 regulation is potentially related. When Dictyostelium or neu-
trophils migrate within a chemoattractant gradient, they preferen-
tially localize PI3K and activated Akt to their leading edges (see
[74, 96, 97]). The antagonistic actions at the leading edge and rear
of a chemotaxing cell are essential to maintain polarity and direc-
tionality and to suppress spurious lateral movement (see [74, 96,
97]). Microtubule assembly is critical to regulate these balanced
and localized activities [98, 99]. Disruption of the microtubule net-
work in neutrophils suppresses leading edge formation and cell
polarization in response to a chemoattractant signal, yet broadly
activates Rho and Rho kinase, regulators that normally specify the
rear of chemotaxing cells [98, 99].
While there may be a similar downregulation of GSK3β activ-
ity via Akt at the leading edge of neutrophils as in migrating astro-
cytes (see Fig. 4), a dependency on GSK3 inhibition and microtu-
bule regulation in chemotaxing cells is not proven. Note that the N-
terminus of GSK3 in Dictyostelium is unrelated to that of GSK3β,
and although it is Ser-rich, phosphorylation (or inhibition) by Akt
has not been examined. However, one observation in Dictyostelium
may further the connection of GSK3 function to cell polarity during
chemotaxis. As would be consistent, disruption of GSK3 in Dictyos-
telium leads to defects at the backs of chemotaxing cells (Kim and
Kimmel, unpublished). It is interesting to consider that localized
activation of GSK3 by tyrosine phosphorylation may also influence
cell polarization by restricting microtubule formation to promote
essential posterior functions during chemotaxis.
NOTE ADDED IN PROOFS
A phosphorylated motif within the intracellular domain of
LRP5/6 can function as a docking site for Axin . Data now
indicate that GSK3β directs this phosphorylation , which in-
duces the subsequent phosphorylation of LRP5/6 by casein kinase 1
and recruits Axin binding [101, 102]. The rapid disruption of Axin/
GSK3β interactions, upon Wnt stimulation [40a], may promote
LRP5/6 phosphorylation and potentiate the destabilization of the
Axin destruction complex .
We are indebted to continuous and insightful discussions with
Drs. Xunxian Liu, Yingzi Yang, Jeffrey Rubin, and Pamela
 Jope, R.S. and Johnson, G.V. (2004) Trends Biochem. Sci., 29, 95-
 Meijer, L.; Flajolet, M. and Greengard, P. (2004) Trends Pharma-
col. Sci., 25, 471-480.
 Frame, S. and Cohen, P. (2001) Biochem. J., 359, 1-16.
 Logan, C.Y. and Nusse, R. (2004) Annu. Rev. Cell Dev. Biol., 20,
 Moon, R.T.; Bowerman, B.; Boutros, M. and Perrimon, N. (2002)
Science, 296, 1644-1646.
 Doble, B.W. and Woodgett, J.R. (2003) J. Cell Sci., 116, 1175-
 Liberman, Z. and Eldar-Finkelman, H. (2005) J. Biol. Chem., 280,
 Kim, L.; Harwood, A. and Kimmel, A.R. (2002) Dev. Cell, 3, 523-
 Crabtree, G.R. and Olson, E.N. (2002) Cell, 109(Suppl), S67-79.
 Benedito, A.B.; Lehtinen, M.; Massol, R.; Lopes, U.G.; Kirch-
hausen, T.; Rao, A. and Bonni, A. (2005) J. Biol. Chem., 280,
 Hoeflich, K.P.; Luo, J.; Rubie, E.A.; Tsao, M.S.; Jin, O. and
Woodgett, J.R. (2000) Nature, 406, 86-90.
 Bachelder, R.E.; Yoon, S.O.; Franci, C.; de Herreros, A.G. and
Mercurio, A.M. (2005) J. Cell Biol., 168, 29-33.
 Takada, Y.; Fang, X.; Jamaluddin, M.S.; Boyd, D.D. and Aggar-
wal, B.B. (2004) J. Biol. Chem., 279, 39541-39554.
 O'Brien, W.T.; Harper, A.D.; Jove, F.; Woodgett, J.R.; Maretto, S.;
Piccolo, S. and Klein, P.S. (2004) J. Neurosci., 24, 6791-6798.
 Willert, K.; Brown, J.D.; Danenberg, E.; Duncan, A.W.; Weiss-
man, I.L.; Reya, T.; Yates, J.R., 3rd and Nusse, R. (2003) Nature,
 Veeman, M.T.; Axelrod, J.D. and Moon, R.T. (2003) Dev. Cell, 5,
 Kim, L. and Kimmel, A.R. (2000) Curr. Opin. Genet. Dev., 10,
 Amit, S.; Hatzubai, A.; Birman, Y.; Andersen, J.S.; Ben-Shushan,
E.; Mann, M.; Ben-Neriah, Y. and Alkalay, I. (2002) Genes Dev.,
Fig. (4). GSK3 and cell polarity.
The GSK3 regulatory modes for several cell polarization systems are compared and contrasted. See text for details and additional components.
1418 Current Drug Targets, 2006, Vol. 7, No. 11 Kim and Kimmel
 Liu, C.; Li, Y.; Semenov, M.; Han, C.; Baeg, G.H.; Tan, Y.; Zhang,
Z.; Lin, X. and He, X. (2002) Cell, 108, 837-847.
Hart, M.J.; de los Santos, R.; Albert, I.N.; Rubinfeld, B. and Po-
lakis, P. (1998) Curr. Biol., 8, 573-581.
Kitagawa, M.; Hatakeyama, S.; Shirane, M.; Matsumoto, M.;
Ishida, N.; Hattori, K.; Nakamichi, I.; Kikuchi, A. and Nakayama,
K. (1999) EMBO J., 18, 2401-2410.
Cong, F.; Schweizer, L.; Chamorro, M. and Varmus, H. (2003)
Mol. Cell Biol., 23, 8462-8470.
Takemaru, K.; Yamaguchi, S.; Lee, Y.S.; Zhang, Y.; Carthew,
R.W. and Moon, R.T. (2003) Nature, 422, 905-909.
Tolwinski, N.S. and Wieschaus, E. (2004) PLoS Biol., 2, E95.
Bhanot, P.; Brink, M.; Samos, C.H.; Hsieh, J.C.; Wang, Y.; Macke,
J.P.; Andrew, D.; Nathans, J. and Nusse, R. (1996) Nature, 382,
Yang-Snyder, J.; Miller, J.R.; Brown, J.D.; Lai, C.J. and Moon,
R.T. (1996) Curr. Biol., 6, 1302-1306.
Wehrli, M.; Dougan, S.T.; Caldwell, K.; O'Keefe, L.; Schwartz, S.;
Vaizel-Ohayon, D.; Schejter, E.; Tomlinson, A. and DiNardo, S.
(2000) Nature, 407, 527-530.
Tamai, K.; Semenov, M.; Kato, Y.; Spokony, R.; Liu, C.; Ka-
tsuyama, Y.; Hess, F.; Saint-Jeannet, J.P. and He, X. (2000) Na-
ture, 407, 530-535.
Pinson, K.I.; Brennan, J.; Monkley, S.; Avery, B.J. and Skarnes,
W.C. (2000) Nature, 407, 535-538.
Willert, K.; Shibamoto, S. and Nusse, R. (1999) Genes Dev., 13,
Yamamoto, H.; Kishida, S.; Kishida, M.; Ikeda, S.; Takada, S. and
Kikuchi, A. (1999) J. Biol. Chem., 274, 10681-10684.
Salic, A.; Lee, E.; Mayer, L. and Kirschner, M.W. (2000) Mol.
Cell, 5, 523-532.
Lee, E.; Salic, A.; Kruger, R.; Heinrich, R. and Kirschner, M.W.
(2003) PLoS Biol., 1, E10.
Noordermeer, J.; Klingensmith, J.; Perrimon, N. and Nusse, R.
(1994) Nature, 367, 80-83.
Itoh, K.; Antipova, A.; Ratcliffe, M.J. and Sokol, S. (2000) Mol.
Cell Biol., 20, 2228-2238.
Smalley, M.J.; Sara, E.; Paterson, H.; Naylor, S.; Cook, D.; Jayati-
lake, H.; Fryer, L.G.; Hutchinson, L.; Fry, M.J. and Dale, T.C.
(1999) EMBO J., 18, 2823-2835.
Wang, H.Y. and Malbon, C.C. (2003) Science, 300, 1529-1530.
Liu, T.; DeCostanzo, A.J.; Liu, X.; Wang, H.; Hallagan, S.; Moon,
R.T. and Malbon, C.C. (2001) Science, 292, 1718-1722.
Liu, X.; Liu, T.; Slusarski, D.C.; Yang-Snyder, J.; Malbon, C.C.;
Moon, R.T. and Wang, H. (1999) Proc. Natl. Acad. Sci. USA, 96,
Katanaev, V.L.; Ponzielli, R.; Semeriva, M. and Tomlinson, A.
(2005) Cell, 120, 111-122.
Liu, X.; Rubin, J.S. and Kimmel, A.R. (2005) Curr. Biol., 15,
Mao, J.; Wang, J.; Liu, B.; Pan, W.; Farr, G.H., 3rd; Flynn, C.;
Yuan, H.; Takada, S.; Kimelman, D.; Li, L. and Wu, D. (2001)
Mol. Cell, 7, 801-809.
Schweizer, L. and Varmus, H. (2003) BMC Cell Biol., 4, 4.
He, X.; Semenov, M.; Tamai, K. and Zeng, X. (2004) Develop-
ment, 131, 1663-1677.
Brennan, K.; Gonzalez-Sancho, J.M.; Castelo-Soccio, L.A.; Howe,
L.R. and Brown, A.M. (2004) Oncogene, 23(28), 4873-84.
Chen, C.M.; Strapps, W.; Tomlinson, A. and Struhl, G. (2004)
Proc. Natl. Acad. Sci. USA, 101, 15961-15966.
He, X. (2004) Cell, 118, 668-670.
Deshpande, R.; Inoue, T.; Priess, J.R. and Hill, R.J. (2005) Dev.
Biol., 278, 118-129.
Inoue, T.; Oz, H.S.; Wiland, D.; Gharib, S.; Deshpande, R.; Hill,
R.J.; Katz, W.S. and Sternberg, P.W. (2004) Cell, 118, 795-806.
Oishi, I.; Suzuki, H.; Onishi, N.; Takada, R.; Kani, S.; Ohkawara,
B.; Koshida, I.; Suzuki, K.; Yamada, G.; Schwabe, G.C.; Mundlos,
S.; Shibuya, H.; Takada, S. and Minami, Y. (2003) Genes Cells, 8,
Lu, W.; Yamamoto, V.; Ortega, B. and Baltimore, D. (2004) Cell,
 Yoshikawa, S.; McKinnon, R.D.; Kokel, M. and Thomas, J.B.
(2003) Nature, 422, 583-588.
Cohen, M.M., Jr. (2003) Am. J. Med. Genet. A., 123, 5-28.
Apionishev, S.; Katanayeva, N.M.; Marks, S.A.; Kalderon, D. and
Tomlinson, A. (2005) Nat. Cell Biol., 7, 86-92.
Jia, J.; Amanai, K.; Wang, G.; Tang, J.; Wang, B. and Jiang, J.
(2002) Nature, 416, 548-552.
Price, M.A. and Kalderon, D. (2002) Cell, 108, 823-835.
Zhang, W.; Zhao, Y.; Tong, C.; Wang, G.; Wang, B.; Jia, J. and
Jiang, J. (2005) Dev. Cell, 8, 267-278.
Jia, J.; Tong, C.; Wang, B.; Luo, L. and Jiang, J. (2004) Nature,
Jia, J.; Tong, C. and Jiang, J. (2003) Genes Dev., 17, 2709-2720.
Lum, L.; Zhang, C.; Oh, S.; Mann, R.K.; von Kessler, D.P.; Tai-
pale, J.; Weis-Garcia, F.; Gong, R.; Wang, B. and Beachy, P.A.
(2003) Mol. Cell, 12, 1261-1274.
Ruel, L.; Rodriguez, R.; Gallet, A.; Lavenant-Staccini, L. and
Therond, P.P. (2003) Nat. Cell Biol., 5, 907-913.
Lefers, M.A. and Holmgren, R. (2002) Curr. Biol., 12, R422-423.
Phiel, C.J.; Wilson, C.A.; Lee, V.M. and Klein, P.S. (2003) Nature,
Foltz, D.R.; Santiago, M.C.; Berechid, B.E. and Nye, J.S. (2002)
Curr. Biol., 12, 1006-1011.
Noll, E.; Medina, M.; Hartley, D.; Zhou, J.; Perrimon, N. and
Kosik, K.S. (2000) Dev. Biol., 227, 450-464.
Cox, R.T.; McEwen, D.G.; Myster, D.L.; Duronio, R.J.; Loureiro,
J. and Peifer, M. (2000) Genetics, 155, 1725-1740.
Kang, D.E.; Soriano, S.; Xia, X.; Eberhart, C.G.; De Strooper, B.;
Zheng, H. and Koo, E.H. (2002) Cell, 110, 751-762.
Matsubayashi, H.; Sese, S.; Lee, J.S.; Shirakawa, T.; Iwatsubo, T.;
Tomita, T. and Yanagawa, S. (2004) Mol. Cell Biol., 24, 2012-
Tournoy, J.; Bossuyt, X.; Snellinx, A.; Regent, M.; Garmyn, M.;
Serneels, L.; Saftig, P.; Craessaerts, K.; De Strooper, B. and Hart-
mann, D. (2004) Hum. Mol. Genet., 13, 1321-1331.
Schlesinger, A.; Shelton, C.A.; Maloof, J.N.; Meneghini, M. and
Bowerman, B. (1999) Genes Dev., 13, 2028-2038.
Thorpe, C.J.; Schlesinger, A.; Carter, J.C. and Bowerman, B.
(1997) Cell, 90, 695-705.
Bei, Y.; Hogan, J.; Berkowitz, L.A.; Soto, M.; Rocheleau, C.E.;
Pang, K.M.; Collins, J. and Mello, C.C. (2002) Dev. Cell, 3, 113-
Maduro, M.F.; Meneghini, M.D.; Bowerman, B.; Broitman-
Maduro, G. and Rothman, J.H. (2001) Mol. Cell, 7, 475-485.
Korswagen, H.C.; Coudreuse, D.Y.; Betist, M.C.; van de Water, S.;
Zivkovic, D. and Clevers, H.C. (2002) Genes Dev., 16, 1291-1302.
Kimmel, A.R. and Firtel, R.A. (2004) Curr. Opin. Genet. Dev., 14,
Kim, L.; Liu, J. and Kimmel, A.R. (1999) Cell, 99, 399-408.
Harwood, A.J.; Plyte, S.E.; Woodgett, J.; Strutt, H. and Kay, R.R.
(1995) Cell, 80, 139-148.
Plyte, S.E.; O'Donovan, E.; Woodgett, J.R. and Harwood, A.J.
(1999) Development, 126, 325-333.
Walston, T.; Tuskey, C.; Edgar, L.; Hawkins, N.; Ellis, G.; Bow-
erman, B.; Wood, W. and Hardin, J. (2004) Dev. Cell, 7, 831-841.
Brajenovic, M.; Joberty, G.; Kuster, B.; Bouwmeester, T. and
Drewes, G. (2004) J. Biol. Chem., 279, 12804-12811.
Etienne-Manneville, S. and Hall, A. (2003) Curr. Opin. Cell Biol.,
Etienne-Manneville, S. and Hall, A. (2003) Nature, 421, 753-756.
Ciani, L.; Krylova, O.; Smalley, M.J.; Dale, T.C. and Salinas, P.C.
(2004) J. Cell Biol., 164, 243-253.
Shi, S.H.; Cheng, T.; Jan, L.Y. and Jan, Y.N. (2004) Curr. Biol.,
Yoshimura, T.; Kawano, Y.; Arimura, N.; Kawabata, S.; Kikuchi,
A. and Kaibuchi, K. (2005) Cell, 120, 137-149.
Jiang, H.; Guo, W.; Liang, X. and Rao, Y. (2005) Cell, 120, 123-
Eickholt, B.J.; Walsh, F.S. and Doherty, P. (2002) J. Cell Biol.,
Zhou, F.Q.; Zhou, J.; Dedhar, S.; Wu, Y.H. and Snider, W.D.
(2004) Neuron, 42, 897-912.
GSK3 and Development Current Drug Targets, 2006, Vol. 7, No. 11 1419 Download full-text
 Baas, A.F.; Kuipers, J.; van der Wel, N.N.; Batlle, E.; Koerten,
H.K.; Peters, P.J. and Clevers, H.C. (2004) Cell, 116, 457-466.
Baas, A.F.; Smit, L. and Clevers, H. (2004) Trends Cell Biol., 14,
Ossipova, O.; Bardeesy, N.; DePinho, R.A. and Green, J.B. (2003)
Nat. Cell Biol., 5, 889-894.
Clements, W.K. and Kimelman, D. (2003) Nat. Cell Biol., 5, 861-
Ossipova, O.; Dhawan, S.; Sokol, S. and Green, J.B. (2005) Dev.
Cell, 8, 829-841.
Spicer, J.; Rayter, S.; Young, N.; Elliott, R.; Ashworth, A. and
Smith, D. (2003) Oncogene, 22, 4752-4756.
Martin, D.E. and Hall, M.N. (2005) Curr. Opin. Cell Biol., 17, 158-
Tomlinson, A.; Strapps, W.R. and Heemskerk, J. (1997) Develop-
ment, 124, 4515-4521.
 Kimmel, A.R.; Parent, C.A. and Gough, N.R. (2004) Sci. STKE,
Kimmel, A.R. and Parent, C.A. (2003) Science, 300, 1525-1527.
Niggli, V. (2003) J. Cell Sci., 116, 813-822.
Xu, J.; Wang, F.; Van Keymeulen, A.; Rentel, M. and Bourne, H.R.
(2005) Proc. Natl. Acad. Sci. USA, 102, 6884-6889.
Tamai, K.; Zeng, X.; Liu, C.; Zhang, X.; Harada, Y.; Chang, Z. and
He, X. (2004) Mol. Cell, 13, 149-156.
Davidson, G.; Wu, W.; Shen, J.; Bilic, J.; Fenger, U.; Stannek, P.;
Glinka, A. and Niehrs, C. (2005) Nature, 438, 867-872.
Zeng, X.; Tamai, K.; Doble, B.; Li, S.; Huang, H.; Habas, R.;
Okamura, H.; Woodgett, J. and He, X. (2005) Nature, 438, 873-
Quaiser, T.; Anton, R. and Kuhl, M. (2006) Bioessays, 28, 339-
Received: August 11, 2005
Accepted: April 20, 2006