Wnt Signaling as a Potential Therapeutic
Target for Frontotemporal Dementia
?Zeljka Korade1,* and Ka ´roly Mirnics1,*
1Department of Psychiatry and Vanderbilt Kennedy Center for Research on Human Development, Vanderbilt University, Nashville,
TN 37232, USA
*Correspondence: firstname.lastname@example.org (?Z.K.), email@example.com (K.M.)
Progranulin mutations result in frontotemporal dementia, but the underlying pathophysiology has remained
largely unexplained. New data by Geschwind and colleagues in this issue of Neuron uncovered that the Wnt/
FZD2 signaling pathway is an early and critical contributor to disease pathology.
neurodegenerative disorder, character-
ized by progressive behavioral, cognitive,
emotional, social, language, and person-
ality deterioration during adult life (van
Swieten and Heutink, 2008). The disease
can arise as a result of mutations in many
genes, including microtubule-associated
protein tau (MAPT), progranulin (GRN),
charged multivescicular body protein 2B
(CHMP2B), and valosin-containing pro-
tein (VCP) (Neumann et al., 2009). Muta-
tions in MAPT and in GRN, both located
on chromosome 17q21, account for
50%–60% of cases of familial FTD. While
the causality of the GRN mutations vis-a `-
vis FTD has been well replicated, limited
the molecular events by which reduced
The study by Geschwind and colleagues
in this issue of Neuron (Rosen et al.,
2011) exploits an impressive cascade of
logical and comprehensive experiments,
and represents the first significant break-
through in this regard.
Progranulin (also known as acrogranin
and epithelin precursor) is a 593 amino
acid secreted glycoprotein that is com-
posed of 7.5 tandem repeats of a 12-
sequence, and the gene is expressed
across a wide variety of tissues, including
lin was first identified as a gene that was
overexpressed in epithelial tumors and
involved in wound healing and inflamma-
tion and did not attract the attention of
neuroscientists for more than a decade:
GRN mutations were first linked to FTD in
2006 by linkage analyses and positional
cloning (Baker et al., 2006). GRN muta-
tions lead to haploinsuficency (Ahmed
et al., 2007), whereby GRN levels are
reduced by approximately 50%, leading
to ubiquitin positive TDP-43 inclusions in
both neurons and glia, but in the absence
of tau pathology (Neumann et al., 2009).
To address the changes associated
with GRN deficiency, the team led by
Geschwind started by developing an
in vitro model using primary human neural
stem cells (hNPC) in which shRNA was
used to diminish GRN levels. Thus, GRN
knockdown ledto robustgeneexpression
changes in the hNPCs, including enrich-
ment in genes related to cell cycling
and ubiquitination. In addition, in GRN-
inhibited neural progenitor cultures, they
observed increased pyknotic nuclei and
activated CASP3 staining, suggestive of
Furthermore, immunostaining for neu-
ronal and glial markers showed that GRN
downregulation in vitro led to reduced
neuronal survival, mimicking the hallmark
neuronal death observed in FTD patients.
To further elucidate the mechanisms
response to GRN downregulation, the
authors tried to uncover the responsible
transcript network. Using Illumina DNA
profile of GRN-inactivated hNPCs, and
found that numerous members of the
Wnt signaling pathway showed dysregu-
lation of transcription, which they vali-
dated with qPCR. The pattern of dysregu-
lation indicated increased activity in the
Wnt pathway (Takahashi-Yanaga and
Sasaguri, 2007) in GRN-inactivated cells.
analysis (WGCNA) (Zhang and Horvath,
2005) was employed. WGCNA allows the
identification of modules of coexpressed
genes, and here it is revealed that alter-
support that mitochondrial and protein
degradation pathways dysfunctions are
a critical part of FTD pathophysiology
(David et al., 2005; Zhang et al., 2009).
In an effort to seek further confirmation
of their findings on diseased brain tissue,
the authors performed WGCNA and
published postmortem microarray data-
set from patients with sporadic FTD,
GRN+ FTD, and matched controls. The
overall results confirmed that the GRN-in-
dant with the postmortem data from FTD
subjects. Furthermore, gene expression
data from cerebellum, cortex, and hippo-
campus of 6-week-old GRN knockout
mice revealed that frizzled homolog 2—
Fzd2 (a receptor that mediates Wnt
signaling) upregulation was one of the
most consistently upregulated genes.
Importantly, this upregulation occurred
well before the appearance of neuropath-
ological alterations or overt neurodegen-
eration in the brains of mutant mice.
The overall results prove, beyond any
doubt, that the GRN+ FTD pathology is at
least in part mediated through dysregula-
tion of the Wnt signaling pathway and
that these changes are in place before
the onset of neurodegenerative changes
(Figure 1). Furthermore, their results imply
that the mitochondrial and protein degra-
dation pathways are a first consequence
deficitand thatthe inflammatory,synaptic,
and other associated changes represent
Neuron 71, September 22, 2011 ª2011 Elsevier Inc.
downstream evolution of the
primary neuronal progenitors,
postmortem data, transgenic
mouse models, and superb
data mining strategies are an
extremely powerful combina-
tion of research tools. Yet,
regardless of the wealth of
the presented data, a num-
ber of questions remain un-
First, how is GRN exactly
regulating the Wnt signaling
pathway? Noncanonical Wnt
signaling pathways driven by
AP1, cJun, and NFAT did not
show significant changes in
the current study, and the
exact relationship between
GRN-Wnt signaling is an in-
triguing topic of further inves-
tigations. Assessing the role
of genes like Tcf7l2, a key
mediator of canonical Wnt
signaling, might be fruitful, as
dnTcf7l2 (a truncated Tcf7l2
enin and therefore acts as
a potent dominant-negative
Wnt antagonist. Such experi-
ments might help to map out
the pathway between GRN
and Wnt and their regulators,
present in the brain from early embryonic
most prominent and progressive in the
sixth and seventh decade of life, and
what are the compensatory mechanisms
that ‘‘burn out’’ by late adulthood?
Clearly, GRN+ FTD has two phases: a
latent, and asymptomatic phase, when
resses over time, but cellular adaptational
mechanisms can compensate for the
detrimental effects of GRN haploinsuffi-
ciency. Over time, the compensatory
mechanisms fail, cellular damage accu-
mulates, and FTD pathology and symp-
toms evolve. The compensatory mecha-
nisms that keep the disease ‘‘in check’’
for half a century are poorly understood,
and it is not known if this compensation
is mediated through a Wnt-dependent
signaling pathway. However, it is very
likely that this part of the protective-adap-
tational response will involve additional,
non-Wnt-dependent processes (Kumar-
Singh, 2011), potentially including growth
endogenous neuroprotection (Saragovi
et al., 2009).
with a rationale that the new drugs should
either prevent formation or increase clear-
ance of these protein aggregates (Troja-
nowski et al., 2008). So, could modulation
of the Wnt signaling pathway achieve this
goal? Regardless of the enticing findings
of the current study, there is no clear-cut
answer to this question, and one can only
be cautiously optimistic. Wnt/b-catenin
signaling is widespread in the
whole body (from brain to
bone and muscle), and it is
modulation of the Wnt path-
way might result in numerous
fects (Takahashi-Yanaga and
Sasaguri, 2007). In addition,
the in vitro cell line and in vivo
mouse models might not fully
recapitulate the critical fea-
tures of the human disease.
Finally, the most beneficial
effect of Wnt pathway modu-
during the latent phase of the
disease: any beneficial effect
of Wnt modulation could be
diminished by the time that
the diagnosis is made and/or
the inflammatory and degen-
erative changes arise.
Given that disease patho-
neuronal and glial changes,
what is the relationship be-
tween these two deficits?
phages and microglia were
cells in vitro, and that GRN-
deficient hippocampal slices
deprivation of oxygen and
glucose (Yin et al., 2010).
Thus, while the present re-
sults by Rosen et al. (2011) argue for a
strong neuronal pathology in response to
reduced GRN levels, early contribution
of glial dysfunction to the FTD pathology
cannot be excluded. Both glia and neu-
rons express GRN from early develop-
ment (Ahmed et al., 2007), and microglia
lacking GRN may become activated, trig-
gering neuronal-glial interactions that can
further accelerate neuronal degeneration
and cell death.
Finally, what is the convergence of the
early-activated molecular pathways be-
tween the various forms of FTD, and
more widely, across all dementias? The
findings from Rosen et al. (2011), as well
as previous reports suggest that the
molecular pathophysiology, regardless
of the genetic cause, might share signifi-
cant molecular commonalities between
ated Disease Ph
Wnt signaling dysregula?on
(Canonical↑ - Inhibitory↓)
Altered cellAltered cell
Cell death and
FTD phenotype: Cogni?ve, emo?onal, social,
language and personality deteriora?on
onic life. However, the clinical symptoms of disease arise more than half
acenturylater. Initially, the disease is in its latent, compensated phase. During
this time the pathophysiological events slowly progress, but compensatory
mechanisms presumably prevent the emergence of the disease phenotype.
In this latent disease phase, progranulin deficiency triggers a complex dysre-
gulation of the Wnt signaling pathway, where gene products belonging to the
stimulatory, canonical Wnt pathway are upregulated, while negative regula-
of mitochondrial energy metabolism, inefficient protein degradation and
altered cell cycling. At this phase, Wnt dysregulation might be, at least
partially, a compensatory event, which is likely to become detrimental over
a prolonged period of time. The neurodegenerative phase is characterized
by lysosomal alterations, appearance of complex inflammatory processes,
disrupted synaptic transmission, myelination defects, and appearance of
TDP-43 inclusions, which jointly lead to neuronal death. During this neurode-
generative phase Wnt signaling changes are likely to be detrimental to brain
function, rather than compensatory. The molecular pathology and cell loss
ultimately result in functional disturbances and clinical diagnosis of FTD.
Neuron 71, September 22, 2011 ª2011 Elsevier Inc.
the various forms of early onset demen-
tias. To underscore this point, it was sug-
gested almost a decade ago that drugs
that both inhibit the cell cycle and rescue
Wnt activity could provide novel Alz-
et al., 2003). Thus, the accumulating
evidence suggests that the effect of
various FTD-causing mutations and other
dementias converge on a few, common
intracellular pathways including but not
limitedto Wntsignaling. Usingconverging
approaches across hNPC, transgenic
animal models and human postmortem
brains, we should attempt to decipher the
earliest commonalities between the tran-
scriptome/signaling disturbances across
various forms of early-onset dementias.
Consistent data mining with WGCNA
(Zhang and Horvath, 2005) could be
crucial for a success of such an effort,
as over the last several years WGCNA
has arisen as a very powerful, function-
based network analysis tool.
A great study always opens up new
research avenues and highlights the
most important, missing knowledge. The
current study is no exception to this rule,
and the findings of Rosen et al. (2011)
indicate a clear path to the most intriguing
future experiments—and hopefully pro-
vide us with a good foundation for devel-
opment of long-awaited, efficacious ther-
apies for early-onset dementias.
Ahmed, Z., Mackenzie, I.R., Hutton, M.L., and
Dickson, D.W. (2007). J. Neuroinflammation 4, 7.
Gass, J., Rademakers,R., Lindholm, C., Snowden,
J., Adamson, J., Sadovnick, A.D., Rollinson, S.,
et al. (2006). Nature 442, 916–919.
Bhandari, V., Palfree, R.G., and Bateman, A.
(1992). Proc. Natl. Acad. Sci. USA 89, 1715–1719.
Caricasole, A., Copani, A., Caruso, A., Caraci, F.,
Iacovelli, L., Sortino, M.A., Terstappen, G.C., and
Nicoletti, F. (2003). Trends Pharmacol. Sci. 24,
Schuessel, K., Keil, U., Rizzu, P., Ravid, R., Drose,
S., Brandt, U., Muller, W.E., et al. (2005). J. Biol.
Chem. 280, 23802–23814.
D.C., Hauptmann,S., Scherping, I.,
Kumar-Singh, S. (2011). J. Mol. Neurosci. 10.1007/
Neumann, M., Tolnay, M., and Mackenzie, I.R.
(2009). Expert Rev. Mol. Med. 11, e23.
Rosen, E.Y., Wexler, E.M., Versano, R., Coppola,
G., Gao, F., Winden, K.D., Oldham, M.C., Martens,
L.H., Zhou, P., Farese, R.V., Jr., and Geschwind,
D.H. (2011). Neuron 71, this issue, 1030–1042.
Saragovi, H.U., Hamel, E., and Di Polo, A. (2009).
Curr. Alzheimer Res. 6, 419–423.
Takahashi-Yanaga, F., and Sasaguri, T. (2007). J.
Pharmacol. Sci. 104, 293–302.
Trojanowski, J.Q., Duff, K., Fillit, H., Koroshetz, W.,
Kuret, J., Murphy, D., and Refolo, L. (2008). Alz-
heimers Dement. 4, 89–93.
van Swieten, J.C., and Heutink, P. (2008). Lancet
Neurol. 7, 965–974.
Yin, F., Banerjee, R., Thomas, B., Zhou, P., Qian,
L., Jia, T., Ma, X., Ma, Y., Iadecola, C., Beal,
M.F., et al. (2010). J. Exp. Med. 207, 117–128.
Zhang, B., and Horvath, S. (2005). Stat. Appl.
Genet. Mol. Biol. 4, Article17.
Zhang, Y.J., Xu, Y.F., Cook, C., Gendron, T.F.,
Roettges, P., Link, C.D., Lin, W.L., Tong, J., Casta-
nedes-Casey, M., Ash, P., et al. (2009). Proc. Natl.
Acad. Sci. USA 106, 7607–7612.
Patterning Spinal Motor Activity
in the Absence of Synaptic Excitation
Steven A. Crone1and Kamal Sharma1,*
1Department of Neurobiology, University of Chicago, Chicago, IL 60637, USA
et al., in this issue of Neuron highlights the importance of spinal cord inhibitory interneurons in generating
motor activity by showing that they can generate alternating flexor-extensor motor neuron firing in the
absence of glutamatergic synaptic input.
Watch any animal run and it is easy to
appreciate that animal movement is
rhythmic and exquisitely coordinated.
Spinal neural networks comprising excit-
atory and inhibitory interneurons are
thought to generate the locomotor rhythm
and control the pattern of movement.
Theseneuralnetworks areableto orches-
trate the movement across multiple joints
in each leg as the animal moves. In terms
of neural computations, this is not an
easy task. Movement at each joint is
made possible by two sets of muscles
that antagonize each other, and their
contraction moves the joint in opposite
directions. These muscles are activated
in a stereotypic, rhythmic fashion when
an animal is walking or running.
rhythmic motor output and coordinate the
activity of antagonistic muscles? Simple
models of neural networks are useful
tools to conceptualize the essential orga-
nizational principles of complex neural
networks. More than a quarter century
ago, Miller and Scott proposed such
a simple model that could initiate and
sustain coordinated flexor-extensor motor
output (Figure 1A) (Miller and Scott, 1977).
In this model, motor activity is initiated by
Neuron 71, September 22, 2011 ª2011 Elsevier Inc.