Identification of dkk4 as a target of Eda-A1/Edar pathway reveals an unexpected role
of ectodysplasin as inhibitor of Wnt signalling in ectodermal placodes
Ingrid Fliniaux, Marja L. Mikkola, Sylvie Lefebvre, Irma Thesleff⁎
Institute of Biotechnology, Developmental Biology Program, University of Helsinki, 00014, Helsinki, Finland
a b s t r a c t a r t i c l ei n f o
Received for publication 10 August 2007
Revised 26 March 2008
Accepted 14 April 2008
Available online 26 April 2008
The development of epithelial appendages, including hairs, glands and teeth starts from ectodermal
placodes, and is regulated by interplay of stimulatory and inhibitory signals. Ectodysplasin-A1 (Eda-A1) and
Wnts are high in hierarchy of placode activators. To identify direct targets of ectodysplasin pathway, we
performed microarray profiling of genes differentially regulated by short exposure to recombinant Eda-A1 in
embryonic eda−/−skin explants. Surprisingly, there were only two genes with obvious involvement in Wnt
pathway: dkk4 (most highly induced gene in the screen), and lrp4. Both genes colocalized with Eda-A1
receptor Edar in placodes of ectodermal organs. They were upregulated upon Edar activation while several
other Wnt associated genes previously suggested as Edar targets were unaffected. However, low dkk4 and
lrp4 expression was retained in the absence of NF-κB signalling in eda−/−hair placodes. We provide evidence
that this expression was dependent on Wnt activity present prior to Eda-A1/Edar signalling. Dkk4 was
recently suggested as a key Wnt antagonist regulating lateral inhibition essential for correct patterning of
hair follicles. Several pieces of evidence suggest Lrp4 as a Wnt inhibitor, as well. The finding that Eda-A1
induces placode inhibitors was unexpected, and underlines the importance of delicate fine-tuning of
signalling during placode formation.
© 2008 Elsevier Inc. All rights reserved.
Ectodermal organ development has been extensively studied
using several models such as hairs, glands, teeth, and feathers (for
reviews see Pispa and Thesleff, 2003; Veltmaat et al., 2003; Mikkola
and Millar, 2006; Lin et al., 2006). Despite their diversity in shape
and function, these organs share common morphological and
molecular features during the early steps of morphogenesis. They
all develop as a result of interactions between ectoderm and
underlying mesenchymal cells (Hardy, 1992). The first sign of
ectodermal organ formation is a local thickening of the epithelium,
the placode, which is accompanied by condensation of the under-
lying mesenchyme. The communication between and within the two
tissues is mediated by several families of signalling molecules
including Wnts, fibroblast growth factors (FGFs), transforming
growth factors-β (TGF-βs), bone morphogenetic proteins (BMPs),
and sonic hedgehog (Shh) which are produced in the placode or the
associated mesenchymal condensate during early morphogenesis,
and are conserved between species (for review see Pispa and
Thesleff, 2003; Mikkola and Millar, 2006). Numerous studies have
shown that a balance between stimulating and inhibiting signals
governs the patterning of ectodermal appendages.
Ectodysplasin is a signalling molecule in the tumor necrosis factor
(TNF) family which triggers a pathway required for the establishment
of the placode in a number of ectodermal organs, and it has been
shown to operate rather early in the hierarchy of signalling molecules
regulating placode formation (Mikkola and Thesleff, 2003; Mustonen
et al., 2004). This pathway is composed of the ligand ectodysplasin
(Eda), Edar, the receptor of the Eda-A1 isoform of ectodysplasin, and
the cytoplasmic Edar-associated death domain adapter protein
(Edaradd). Mutations in any of these three genes cause a syndrome
called hypohidrotic (anhidrotic) ectodermal dysplasia characterized
by defective development of multiple ectodermal organs (Kere et al.,
1996; Monreal et al., 1999). In mice, the mutations in the
corresponding genes are responsible for the mouse mutants tabby,
downless, and crinkled respectively (Srivastava et al., 1997; Headon
and Overbeek, 1999; Headon et al., 2001). Mice carrying these
mutations lack specifically the long guard hairs and show defects in
several exocrine glands and the number and shapes of teeth. Guard
hair follicles which develop as the first wave of mouse hairs at
embryonic day (E) 14, are missing in the mutant mice, whereas
placodes of the next wave giving rise to awl hairs form correctly at
E16 (Vielkind and Hardy, 1996; Laurikkala et al., 2002). Several
studies have shown that the binding of the Eda-A1 isoform to Edar
leads to activation of transcription factor NF-κB (Yan et al., 2000;
Koppinen et al., 2001; Kumar et al., 2001). Mice and humans carrying
mutations perturbing NF-κB activation show similar phenotypes in
ectodermal organs as those with deficiency in Eda-A1/Edar pathway
Developmental Biology 320 (2008) 60–71
⁎ Corresponding author. Fax: +358 9 19159366.
E-mail address: firstname.lastname@example.org (I. Thesleff).
0012-1606/$ – see front matter © 2008 Elsevier Inc. All rights reserved.
Contents lists available at ScienceDirect
journal homepage: www.elsevier.com/developmentalbiology
components (Schmidt-Ullrich et al., 2001; Mikkola and Thesleff,
2003; Orange et al., 2005), suggesting that NF-κB activation is
required for proper Eda-A1/Edar signal transduction. Targeted
epithelial overexpression of Eda-A1 in transgenic mice under the
control of Keratin 14 promoter (K14) results in supernumerary tooth
and mammary placodes, and enlarged or fused hair placodes
(Mustonen et al., 2003, 2004). Moreover, recombinant Eda-A1 protein
fused to the C-terminus of an IgG1 Fc domain (Fc-EdaA1) perma-
nently restores a majority of the defects of eda−/−litters when
injected to eda−/−pregnant mice (Gaide and Schneider, 2003), and
rescues the first wave of hair placode formation when administered
to E13 eda−/−skin in culture (Mustonen et al., 2004).
Wnt signalling plays a crucial role in ectodermal organ placode
formation, most probably via the canonical β-catenin/LEF1/TCF path-
way (Pispa and Thesleff, 2003; Mikkola and Millar, 2006). Canonical
Wnt signalling has been extensively characterized during the last
decades (Clevers, 2006). Wnt ligands activate Frizzled transmembrane
receptors, which leads to the stabilization of cytoplasmic β-catenin
through inhibition of the GSK3β kinase. β-catenin then translocates
to the nucleus where it binds to LEF/TCF transcription factors to
promote transcription of target genes. TOPGAL mice reporting LEF/
TCF and β-catenin activity via β-galactosidase expression show
canonical Wnt signal activity in the basal epithelial cell layer of hair
pregerms and subsequently in cells of dermal condensates suggest-
ing its involvement in the first steps of hair morphogenesis
(DasGupta and Fuchs, 1999; Mikkola and Millar, 2006). Transgenic
overexpression of the Wnt signal inhibitor Dkk1 in ectoderm inhibits
hair and mammary placode formation, and blocks tooth morpho-
genesis before the bud stage (Andl et al., 2002; Chu et al., 2004)
indicating an absolute requirement of Wnt signalling for the
initiation of ectodermal appendage development. Lef1−/−mice lack
whiskers, teeth, and mammary glands, but exhibit a reduced number
of pelage hairs (Van Genderen et al., 1994; Kratochwil et al., 1996),
suggesting a redundancy of LEF/TCF factors during ectodermal organ
Although both Eda and Wnts are high in the hierarchy of signals
required for ectodermal organ development, the relationship
between Wnt and Eda-A1/Edar signalling has remained unclear.
Recent studies have suggested that Eda-A1/Edar signalling mod-
ulates other signal pathways involved in hair development. We and
others have identified two BMP inhibitors, ccn2/ctgf and follistatin, as
well as shh as putative targets of the Eda-A1/Edar pathway (Mou et
al., 2006; Schmidt-Ullrich et al., 2006; Pummila et al., 2007). These
genes and many others were also found in a microarray analysis
comparing transcriptomes from eda−/−and wild-type skin at the
stage of hair placode formation (Cui et al., 2006). Since primary hair
placodes fail to form in eda−/−skin, it is expected that many markers
of the epidermal placode and dermal condensate emerge from this
type of screen, and therefore it is challenging to differentiate
between the direct Eda target genes from those that are secondarily
In an attempt to identify direct target genes of the ectodysplasin
pathway, we performed a microarray profiling of genes differentially
expressed in eda−/−skin after a short exposure to Fc-EdaA1 in vitro.
Surprisingly, among these were only two genes with an apparent
connection to the Wnt pathway: dkk4 and lrp4. In this report, we
show that dkk4 and lrp4 are expressed in ectodermal organ
placodes, and that they are likely to be novel target genes of Eda-
A1/Edar pathway. Using a quantitative approach, we demonstrate
that dkk4 is strongly, and lrp4 moderately upregulated upon
administration of Eda-A1 protein. We also provide evidence
indicating that the expressions of dkk4 and lrp4 depend on Wnt
signalling present prior to Eda-A1/Edar activity in developing hair
placodes. Moreover, our data suggest that Eda is dispensable for hair
follicle induction but is essential for establishing a correct pattern of
primary hair follicles.
Materials and methods
Wild-type females from the NMRI strain were kept by breeding with NMRI males.
The generation and maintenance of the mouse strains used in this study have been
described earlier: eda-deficient mice (also referred as eda−/−; tabby mice, Jackson
Laboratories stock #JR0314) (Pispa et al., 1999); K14-eda mice (Mustonen et al., 2003);
NF-κBREP(Bhakar et al., 2002). K14-eda and NF-κBREPembryos were identified by PCR.
The appearance of a vaginal plug was taken as embryonic day (E) 0, and embryos were
carefully staged according to limb morphogenesis. K14-eda mice are of FVB background
and NF-κBREPof C57Bl/6 background. NF-κBREPmice were bred into eda−/−background
to monitor NF-κB activity in the absence of Eda.
Back skin from E13 NMRI wild-type embryos was dissected and grown on
nuclepore filters at 37 °C for 24 h in a Trowell-type culture containing Dulbecco's
minimum essential medium (DMEM) supplemented with 10% Fetal Calf Serum (FCS),
glutamine and penicillin–streptomycin. Recombinant mouse Dkk4 (R&D Systems) was
administered to the culture medium as indicated in the text.
Embryos were fixed for 30 min in 4% paraformaldehyde, then washed three times
for 10 min in Dulbecco's PBS pH 7.5 containing 2 mM MgCl2and 0.02% Nonidet P-40.
Samples were stained overnight at room temperature with X-gal staining solution
(1 mg/ml X-gal, 5 mM K4Fe(CN)6, 5 mM K3Fe(CN)6, 2 mM MgCl2and 10% Nonidet P-40
in PBS, pH 7.5).
Hanging drop cultures and quantitative RT-PCR
To analyze the induction of gene expression after different treatments indicated
in the text, E14 eda−/−back skins were grown submerged in hanging drops as
described earlier (Pummila et al., 2007). Each skin was dissected in Dulbecco's PBS
pH7.5 and cut in two halves along the midline. For each treatment, one half was used
as control and the other one was exposed to different molecules indicated in the text.
Each skin half was cultured individually in one drop of 40 μl pre-warmed medium
supplemented with purified recombinant 1 to 2 μg/ml of Fc-Eda-A1 (Gaide and
Schneider, 2003), 6-bromoindirubin-3′-oxime (BIO; Calbiochem), both molecules, or
equivalent proportion of molecule dissolvent for controls. Three different batches of
Fc-Eda-A1 protein were used in the current study. The only difference noticed
between the batches was that the onset of induction of the genes analyzed was
somewhat delayed with batch 2, which was used in most of the experiments.
To manipulate Wnt signalling, skin halves were pre-treated for 4 h with the
indicated concentrations of CKI-7, a potent caseine kinase 1 inhibitor (United States
Biological) or the equivalent concentration of the dissolvent. Skin halves were then
transferred into fresh CKI-7 containing medium such that one half was exposed to
2 μg/ml of recombinant Fc-Eda-A1 for 4 h while the other one was used as a control. A
minimum of triplicate samples was analyzed each time.
After the number of hours indicated in the text, tissues from hanging drops
were placed into 350 μl lysis buffer of the RNeasy mini kit (Qiagen) containing 1%
β-mercaptoethanol (Sigma). Total RNA was isolated as specified by the manufac-
turer's instruction and quantified using a nanodrop spectrophotometer. 500 ng of
total RNA was reverse transcribed using 500 ng of random hexamers (Promega) and
100 units of Superscript II (Invitrogen) following the manufacturer's instructions.
For the time-course experiment, quantitative PCR was performed using 2X SYBR-
green PCR master mix (Applied Biosystems), using the default PCR conditions for
the ABI 7000. Data were normalized against ranbp1 and analyzed with Applied
Biosystems' Prism software. For the experiment comparing Fc-Eda-A1 and BIO
effects on gene induction, Lightcycler DNA Master SYBR Green I (Roche) was used
with a Lightcycler 480, and the software provided by the manufacturer was used
for analysis; data were normalized against K14. Primer sequences are available upon
request. Gene expression was quantified by comparing the sample data against a
dilution series of PCR products of the gene of interest.
Microarray experiment and analysis
Pools of three E14 eda−/−half-skins pairs were submerged in hanging drops and
cultured for 1.5 h or 4 h intheabsence or presenceof 2 μg/ml Fc-Eda-A1 followed bytotal
RNA isolation as described above (Fig. 1). Biological triplicates for each condition were
performed. RNA quality and concentration was monitored using a 2100 Bioanalyzer
(Agilent Technologies). RNAs were processed and hybridized on Affymetrix® Mouse
Genome 430 2.0 arrays, and data were analyzed in the Turku Centre for Biotechnology,
(Robust Multi-Array Average) using R/Bioconductor, allowing background correction,
quantile normalization, probe specific correction and summary value computation. 2-
base log-transformed intensity ratio (treated/control) was calculated for each sample
pair. Average of replicate log-ratios and t-test of replicates were calculated. Genes over
1.5-fold differences and with a p-value below 0.05 were chosen for further analysis.
I. Fliniaux et al. / Developmental Biology 320 (2008) 60–71
Promoter analysis and luciferase reporter assays
Mouse and human upstream promoter sequences (−5000 to +1; A of the first
ATG codon being +1) of dkk4 and lrp4 were aligned and analyzed for the presence
of putative NF-κB binding sites with the CONSITE program (http://asp.ii.uib.no:8090/
cgi-bin/CONSITE/consite) using 60% conservation cut-off and window size 25. For
lrp4, sites with an 80% transcription factor score threshold were retained. For
dkk4, transcription factor score thresholds from 75 to 80% were used in the analysis.
For the cloning of mouse dkk4 promoter region containing the putative NF-κB
response elements, a 0.8 kb region upstream of the translation initiation codon was
amplified with the following primers: forward 5′-TGTTTATGCGCCCCACTATT-3′ and
reverse primer including a Hind3 site 5′-CTAGCAAGCTTCCTCAGTCACTCTGGTCTCTCG-
3′. The resulting PCR product was digested with SmaI/Hind3 and cloned into pGL3 basic
luciferase reporter vector (Promega, Madison). For transfection, 293Tcells were seeded
at 1×105cells per well on gelatin coated six well plates. The following day, 950 ng of
each luciferase reporter plasmid was co-transfected with 50 ng of Renilla luciferase
plasmid (pRL-TK, Promega) and 500 ng of expression plasmids for wild-type edar,
mutant edar (edarSleek) lacking the death domain, empty vector (pEF1/Myc-His A)
(Invitrogen) (Koppinen et al., 2001), or 50 ng of wild-type β-catenin in pIRES2-AcGFP1-
Nuc vector and 50 ng of human Lef1 in pcDNA/AmpI vector (kindly provided by Hannu
Sariola) using Fugene6 (Roche) as recommended. The amount of DNA in each
transfection was kept constant by adding pEF1/Myc-His A. The luciferase assays were
performed 24 h after transfection using the Dual-Luciferase Reporter Assay System as
specified by the manufacturer (Promega) and firefly luciferase activities were measured
using Bio Orbit 1253 luminometer and normalized to the Renilla luciferase values. Data
shown are mean±SEM of two independent experiments with duplicate/triplicate
In situ hybridization
Cultured skins were treated with cold methanol before fixation. Whole embryos,
isolated mandibles and methanol treated skins were fixed in 4% paraformaldehyde
overnight, and processed for whole-mount in situ hybridization as described earlier
(Mustonen et al., 2004) by using the InSituPro robot (Intavis AG, Germany).
Digoxigenin-labeled probes were detected with BM Purple AP Substrate Precipitating
Solution (Boehringer Mannheim Gmbh, Germany). The following probes were used:
edar and shh (Laurikkala et al., 2002); a 994 base pairs probe specific to the dkk4
sequence (nt 114–1107 of BC018400); a 824 base pairs probe specific to the lrp4
sequence (nt 60–883 of NM_172668). Some stained samples were embedded in 0.5%
gelatin, 30% albumin, 20% sucrose, 2% glutaraldehyde in PBS to be sectioned in 30 μm
thick slices using a vibratome.
Search for ectodysplasin-induced genes by microarray analysis
We have shown earlier that the formation of the first wave of hair
placodes is restored in eda−/−back skin cultures when recombinant
Fc-Eda-A1 protein is administered to the culture medium (Mustonen
et al., 2004). In addition, we reported that Eda-A1/Edar target genes
ctgf/ccn2 and shh are induced already after 1 h of incubation with Fc-
Eda-A1 protein, their expression levels peaking by 4 h (Pummila et
al., 2007). We made use of these findings in designing the protocol
for identification of immediate target genes of Edar (Fig. 1). Back
skins of E14 eda -deficient mouse embryos were cut in two halves
along the midline and these were cultured in hanging drops of
culture medium with or without recombinant Fc-Eda-A1. The
explants were harvested after 1.5 and 4 h and processed for
hybridization on Affymetrix® microarray chips containing approxi-
mately 14,000 well characterized mouse gene probes. Genes
presenting an at least 1.5-fold intensity difference of expression in
treated compared to non-treated skins were considered as putative
Eda target genes. Altogether, 22 genes were found upregulated after
1.5 h, and 142 genes after 4 h, and 16 of them were present in both
arrays. Among the genes upregulated upon Eda treatment, we found
ctgf/ccn2 in both arrays, and follistatin and shh in the 4-hour array.
These genes were previously identified as targets of Eda-A1 (here-
after Eda) (Mou et al., 2006; Pummila et al., 2007), indicating that
our experimental procedure had been successful. In this study, we
focus on genes related to Wnt signalling.
Dkk4 and lrp4 are induced by Eda whereas several other Wnt associated
placodal genes are not
Several Wnt pathway genes have been proposed as possible
Eda targets (Cui et al., 2006).
experiment sorted out only two genes with prior association
with Wnt signalling: dkk4 and lrp4. Dkk4 was the most highly
differentially expressed gene (~30-fold increased) after 4 h of Eda
treatment while lrp4 was only moderately upregulated (~2-fold
In order to validate the accuracy of the microarray data, we
monitored the induction kinetics of dkk4 and lrp4 transcripts after
treatment of eda-deficient skin explants with recombinant Eda
protein by quantitative RT-PCR technique. E14 skin explants were
cut into two halves along the dorsal midline and cultured with or
without Fc-Eda-A1 (Pummila et al., 2007). RNAs were isolated after 1,
3, 5, 7 and 9 h of culture, and the number of dkk4 or lrp4 transcripts
was measured. Dkk4 expression increased rapidly and prominently,
reaching about 2-fold within the first hour, and rising almost
exponentially at least until 9 h when it was 130-fold. Lrp4 expression
increased more moderately showing a rather constant 2–3-fold
induction at all time points (Fig. 2A). The induction of dkk4 at 4 h
was lower than in the microarray data set, most likely due to slight
batch-to-batch variation observed with recombinant Fc-Eda-A1 (see
also Fig. 6L).
Using the same experimental procedure, we monitored the fold
of induction of some other Wnt pathway components which are co-
expressed with edar in nascent hair placodes and have been
proposed to be Eda targets (Andl et al., 2002; Cui et al., 2006), but
did not show any significant changes in our microarray screen (Fig.
2B). These included the Wnt signalling components lef1, β-catenin
and wnt10b, which appear in placodes when hair primordium
morphogenesis is rescued by Fc-EdaA1 treatment in cultured tabby
(eda−/−) skin (Mustonen et al., 2004, and our unpublished data). In
addition we examined the induction of sostdc1 (also known as ec-
todin and wise), a secreted modulator of Wnt and BMP signalling
(Itasaki et al., 2003; Laurikkala et al., 2003; O'Shaughnessy et al.,
Interestingly, our microarray
Fig. 1. Schematic representation of the experimental procedure used to identify direct
targets of Eda-A1/Edar signalling pathway. E14 eda−/−back skins were dissected and
cut in two parts along the midline. Pools of 3 half-skins were cultured in a hanging
drop of control medium, whereas the 3 other halves were cultured in medium
supplemented with 2 μg/ml of Fc-Eda-A1 recombinant protein. After 1.5-hour or 4-
hour incubation, total RNA was extracted from the samples, and processed for
hybridization on Affymetrix® Mouse genome 430 2.0 Arrays. Three replicates were
used for each condition.
I. Fliniaux et al. / Developmental Biology 320 (2008) 60–71
2004) which was previously shown to be overrepresented in wild-
type compared to eda−/−skin (Cui et al., 2006). In contrast to dkk4
and lrp4, none of these genes was upregulated during the course of
the 9-hour experiment. Lef1, and in particular Sostdc1 levels even
tended to decrease upon prolonged exposure to Fc-EdaA1. The fact
that these placode markers were not induced in the time window
analyzed supports the conclusion that genes upregulated upon brief
exposure to recombinant Eda are likely to be direct transcriptional
targets of Edar.
Dkk4 and lrp4 are expressed in placodes during ectodermal organ
morphogenesis and upregulated in placodes of mice overexpressing eda
We next analyzed the expression of dkk4 and lrp4 by whole-mount
in situ hybridization in E14.5 mouse skin, when the placodes of the
first wave are well developed and express many placodal marker
genes including edar (Laurikkala et al., 2002; Headon and Overbeek,
1999). Lpr4, and dkk4 transcripts were restricted to hair placodes,
further supporting their regulation by Edar-dependent signalling
Fig. 2. Validation of dkk4 and lrp4 as putative target genes of Eda-A1/Edar signalling. (A, B) Eda−/−skin explants were cultured in hanging drops in the presence or absence of 2 μg/ml
Fc-Eda-A1, and a time-course analysis induction of dkk4 and lrp4 (A) as well as of other Wnt pathway-associated genes (B) was performed by quantitative RT-PCR. Dkk4 transcription
constantly increased from 1 h to 9 h upon Fc-Eda-A1 stimulus. Lrp4 was upregulated after 3 h of treatment and remained at fairlyconstant levels thereafter whereas no changes were
observed in ectodin, lef1, β-catenin and wnt10b transcripts. (C–H) Dkk4 and lrp4 expression was monitored in E14 skin, and E12 lower jaw and mammary buds in wild-type embryos
by in situ hybridization. (C, E, G) Dkk4 and (D, F, H) lrp4 were expressed in E14 wild-type primary hair follicles placodes, as well as in tooth and mammary buds (numbered) at E12. i,
incisor placode; m, molar placode.; t, tongue. Scale bar: 1 mm in panels C, D; 500 μm in panels E, F; 200 μm in panels G, H.
I. Fliniaux et al. / Developmental Biology 320 (2008) 60–71
(Figs. 2C and D). We then studied their expression in placodes of other
ectodermal organs where edar transcripts are also present, i.e.
mammary placodes and tooth placodes (Pispa et al., 2003; Pummila
et al., 2007). At E12, dkk4 and lrp4 were both intensely expressed in
mammary placodes (Figs. 2E and F) as well as in molar and incisor
placodes (Figs. 2G and H). Thus, dkk4 and lrp4 seem to be general
placode markers of ectodermal appendages and their colocalization
with Edar is in line with regulation by Ectodysplasin signalling.
Next, we examined the expression patterns of dkk4 and lrp4 in
K14-eda-A1 mice characterized by overexpression of the ligand Eda-
A1 throughout the developing ectoderm (Mustonen et al., 2003). In
this transgenic line, hair placodes of the first wave are enlarged and
sometimes fused at E14, as revealed by broader expression of known
placodal markers (Mustonen et al., 2004). Expression of both dkk4 and
lrp4 was more intense and in a larger area in hair placodes of K14-eda-
A1 embryos (Figs. 3A–D, C', D') as compared to wild-type litter mates
(Figs. 3E, E' and F, F'). This suggests that stimulation of Edar signalling
results in increased production of dkk4 and lrp4 transcripts also in
Dkk4 and lrp4 promoters possess NF-κB binding sites and dkk4 promoter
is responsive to Edar
Since several studies have shown that Eda-A1/Edar signalling
activates the NF-κB pathway (Yan et al., 2000; Koppinen et al., 2001;
Kumar et al., 2001) we next examined the promoter regions of dkk4
and lrp4 for the presence of potential NF-κB responsive elements by
computational analysis using mouse and human sequences to find
evolutionary conserved sites (see Materials and methods for details)
(Fig. 4A). Within the lrp4 promoter sequence, 6 distinct putative NF-κB
binding sites were found with a high score. Among those sites, one
located 2271 base pairs upstream of the ATG was 100% identical
between mouse and human and matched perfectly with the NF-κB
consensus binding sequence. For dkk4, one putative site, located 446
base pairs upstream of the ATG was identified by the software, and by
applying a slightly lower transcription factor score threshold 2 more
conserved NF-κB recognition sequences were revealed. Hence these
results suggest that dkk4 and lrp4 might be regulated by NF-κB further
supporting the hypothesis that dkk4 and lrp4 are direct targets of Eda-
To test whether the upstream promoter region of dkk4 that
included the conserved NF-κB sites was responsive to Edar, we
cloned a 0.8 kb region of dkk4 into a luciferase reporter vector. It is
well established that TNF receptors can activate downstream
signalling cascades in a ligand-independent fashion upon over-
expression in transfected cells. We have previously shown that
transfection of 293T cells with wild-type Edar induces the expres-
sion of a NF-κB luciferase reporter while mutant forms of Edar
associated with the ectodermal dysplasia phenotype lead to severely
compromised responses (Koppinen et al., 2001). After 24 h,
Fig. 3. Dkk4 and lrp4 expression is increased in K14-eda mice. Whole-mount in situ hybridization with probes specific to dkk4 (A, C, E) and lrp4 (B, D, F) at E14.5. K14-Eda embryos (A,
B, C, D) revealed a more intense and broader expression of both genes compared to the expression pattern in wild-type litter mates (E, F). (C'–F') Vibratome sections of embryos
stained with in situ hybridization. Scale bars: 1 mm in panels A, B; 200 μm in panels C–F; 50 μm in panels C'–F'.
I. Fliniaux et al. / Developmental Biology 320 (2008) 60–71
transfection of edar led to a moderate but noticeable 2-fold
induction of dkk4 promoter reporter construct as compared to the
control plasmid or edarSleekmutant (Fig. 4B). The modest activation
of the 0.8 kb proximal dkk4 promoter by Edar suggests that other
regulatory regions either further upstream or downstream and not
revealed by computational analysis are essential for the efficient
Eda-A1-dependent induction of dkk4 in vivo. Alternatively, other co-
factors not present in 293T cells may be required for efficient
induction by Edar.
Dkk4 and lrp4 are focally expressed in eda−/−skin in the absence of
Since our results suggested that the Eda-A1/Edar pathway is
upstream of dkk4 and lrp4, we next examined the expression of these
two genes in eda−/−mice. As mice lacking Eda-A1/Edar signalling do
not show hair placodes at E14, nor present patterned expression of
any placodal marker so far analyzed, including edar (Headon and
Overbeek, 1999; Laurikkala et al., 2002), we expected absence of
localized dkk4 and lrp4 expression in eda−/−skin. Surprisingly, both
genes we expressed in pointed irregular patterns in the E14.5 eda−/−
skin, albeit at much lower intensity than inwild-type skin (Figs. 5A, D;
compare to Figs. 2C and D). The pattern of both genes in eda−/−
embryos (Figs. 5B, E) was markedly different from that seen in E14
wild-type skin where the expression was more intense and decorated
the regularly arranged and rather uniformly sized placodes of the first
wave hair follicles (Figs. 5C, F). To our knowledge, dkk4 and lrp4 are
the first placode markers showing a clear periodic expression pattern
in eda mutant skin at E14. These dkk4- and lrp4-positive foci probably
represent the same structures that were recently identified in
histological sections of Edar-deficient E14.5 skin and termed as ‘pre-
placodes’ (Schmidt-Ullrich et al., 2006).
We next studied whether the observed expression of dkk4 and
lrp4 in the eda−/−embryos was dependent on the NF-κB pathway.
We crossed eda−/−mice with mice reporting NF-κB activity via β-
galactosidase expression (Bhakar et al., 2002), allowing us to
monitor in the same embryos the NF-κB activity with X-gal
staining, and the expression patterns of dkk4 and lrp4 by whole-
mount in situ hybridization. At E14, while NF-κBREP/eda−/−embryos
exhibited the same lrp4 and dkk4 expression pattern as eda−/−
embryos (data not shown), they did not show patterned edar
expression (Fig. 5G) nor NF-κB activity in skin (Fig. 5H), in contrast
mice in a wild-type C57Bl/6 background, where
positive Xgal staining was evident in hair placodes (Fig. 5I). This
indicates that NF-κB activity in primary hair placodes depends on
Eda-A1/Edar signalling, as earlier suggested using an independent
NF-κB reporter (Schmidt-Ullrich et al., 2006).
Dkk4 expression was dramatically reduced also in dental placodes
of the lower jaw in eda−/−background. The staining patterns varied,
however, between embryos. In some specimens, the expression was
absent from incisors but faint expression was seen in the molar
placodes (Fig. 5J, compare to Fig. 2G), whereas in other samples weak
staining was present only in incisor placodes (Fig. 5K). Lrp4 expression
did not show any significant difference either in molar or incisor
placodes as compared to the wild-type mice (Fig. 5L compare to 2H).
In addition, dkk4 and lrp4 transcripts were also detected in eda−/−
mammary placodes at E12 (data not shown). These observations
indicate that dkk4 and lrp4 are still induced, although at a lower
extent, after disruption of NF-κB signalling in ectodermal placodes,
suggesting that other signals in addition to ectodysplasin regulate
their expression in vivo.
Dkk4 and lrp4 show similar expression pattern in wild-type and eda−/−
embryos at the time of hair follicle induction
Because dkk4 and lrp4 both showed localized expression in eda−/−
pre-placodes at E14.5, we next examined their expression at the time
when the first hair follicles are induced. At E13.5, both transcripts
were readily observed in the first rows of hair placodes that appear
parallel to mammary line (Figs. 6A–D). The same foci were also
positive for NF-κB reporter activity (Fig. 6E). In addition, some other
discrete and more disorganized stained foci were observed around
the regularly patterned rows. At the same stage, both genes were
also detected in vibrissae placodes and buds (Fig. 6). Interestingly, a
similar pattern of both dkk4 and lrp4 was observed in E13.5 eda−/−
embryos (Figs. 6F–I), whereas again no NF-κB activity was dis-
cernible (Fig. 6J). Taken together, our data suggest that this early
placodal expression of dkk4 and lrp4 is independent of Edar/NF-κB
signalling but subsequently, the establishment of the normal pattern
of primary hair placodes and upregulation of dkk4 and lrp4 fails in
eda mutant embryos.
Wnt signalling induces dkk4 and lrp4 expression and acts in parallel
with the Eda-A1/Edar pathway
Dkk4 transcription has been reported to be increased in the fore-
brain of mice carrying a dominant-active β-catenin (Ncre/b-catΔex3+/−
embryos), suggesting that it could be downstream of canonical Wnt
activity (Diep et al., 2004). Because Wnt signalling has an established
early function inplacode formation, we tested whether Wnt signalling
might be responsible for the dkk4 and lrp4 expression occurring in
eda−/−skin. We compared the fold of induction of dkk4 and lrp4 by
quantitative real-time PCR in E14 eda−/−back skin treated for 4 h with
Fig. 4. Computational search for NF-κB binding sites in dkk4 and lrp4 promoter regions
and luciferase reporter assays of dkk4 promoter activation. (A) Several putative
conserved NF-κB consensus binding sequences were identified within the 5000 bp
region upstream of ATG in both dkk4 and lrp4 genes. Numbers refer to mouse promoter
sequences. (B) 0.8 kb fragment of mouse dkk4 and upstream promoters encompassing
the NF-κB consensus elements were cloned into a promoter-less pGL3 luciferase
reporter. Wild-type Edar-induced Dkk4 promoter sequence drove the expression of the
reporter moderately when co-expressed with wild-type Edar into 293T cells, and more
strongly when co-transfected with β-catenin and lef1 plasmids.
I. Fliniaux et al. / Developmental Biology 320 (2008) 60–71
either Fc-Eda-A1 or 6-bromoindirubin-3′-oxime (BIO), a pharmacolo-
gical inhibitor specific to glycogen synthase kinase-3β (GSK-3β)
(Meijer et al., 2003) that stimulates canonical Wnt pathway or both
factors together (Fig. 6K). BIO induced dkk4 expression in a dose-
dependent manner from 2.4 to 5.3-fold with BIO concentrations from
5 μM to 20 μM. The addition of both factors together led to an overall
higher induction, corresponding approximately to the sum of each
treatment alone. Similar findings were observed with lrp4 but at a
To address the question whether the induction of dkk4 by Eda-A1
was dependent on intact Wnt/β-catenin signalling, we used the casein
kinase 1 inhibitor CKI-7 (Chijiwa et al., 1989; Price, 2006) to inhibit
canonical Wnt activity. First, E14 eda−/−back skin halves were treated
with 0 to400 μM of CKI-7 for 4 h to block casein kinase 1 activity. After
the pretreatment, the explants were transferred into fresh CKI-7
containing medium so that one half was exposed to Fc-Eda-A1
whereas the other one was used as a control (Fig. 6L). Low
concentration (25 μM) of CKI-7 did not have a gross effect on dkk4
induction by Eda-A1 although prominent variation between samples
was observed. However, higher concentrations of CKI-7 lead to
decreased induction of dkk4 yet a more than 10-fold induction was
evident even at a very high (400 μM) concentration of CKI-7. These
data propose that Wnt signalling has a prominent contribution to the
ability of Eda-A1 to induce dkk4 yet Eda-A1 appears to regulate dkk4
also independent of Wnt activity.
These results indicate that dkk4 and lrp4 are both targets of Eda-
A1/Edar as well as Wnt signalling in developing skin appendages. In
line with these data, dkk4 promoter was recently shown to contain
several LEF/TCF consensus binding sites and was responsive to
ectopically expressed β-catenin and Lef1 in a luciferase reporter
assay (Fig. 4B; Sick et al., 2006; Bazzi et al., 2007).
Dkk4 blocks hair placode formation
Although Dkk1 has an established role as a Wnt inhibitor and its
overexpression in skin prevents completely hair placode formation
of Dkk family, in skin appendage development is poorly understood.
Therefore we examined the effect of Dkk4 on hair placode formation.
We cultured E13.0 (i.e. prior to the formation of the first wave of hair
follicles) wild-type skin for 24 h in medium supplemented with
increasing doses of recombinant Dkk4 (Figs. 7A–D). Placode formation
was monitored by in situ hybridization analysis of shh expression.
Dkk4 inhibited the appearance of placodes in a dose-dependent
manner (Fig. 7E) and the highest Dkk4 concentration (1 μg/mL)
almost completely blocked placode formation. This result indicates
that Dkk4 acts as an inhibitor of Wnt signalling in developing skin.
Previous genome wide studies aiming at discovering Eda target
genes have either compared the expression of genes in adult skin of
mice with altered levels of Eda expression (Cui et al., 2002) or more
recently have analyzed differentially expressed genes at various
developmental stages in control vs. eda-deficient skin (Cui et al.,
2006). However, interpretation of data derived from such experi-
ments is not straightforward as all genes secondarily induced by
primary Eda targets are likely to be differentially expressed as well.
Here we report a novel approach for the identification of direct Eda
target genes. We performed a microarray screen on embryonic eda
null skin after a short-term exposure to recombinant Eda at the time
when primary hair placodes form in wild-type skin.
Dkk4 and lrp4 are likely to be direct Eda target genes
Intriguingly, in our microarray screen we detected only two
differentially expressed genes, dkk4 and lrp4, with an obvious
connection to Wnt signalling while the expression levels of a
number of other Wnt pathway-related genes such as kremen2, dkk1,
wnt5a, wnt6, wnt10b, wnt11, and sostdc1 previously identified as
candidate target genes of Eda (Cui et al., 2006) were not changed. As
Eda and Wnt pathways are both activators of placode formation, we
further analyzed in detail those Wnt-related genes co-expressed
with Edar in developing skin appendages and with an established or
proposed positive role in placode formation. In line with the
microarray data, we found no alterations in β-catenin, wnt10b, or
lef1 mRNA levels upon Eda treatment indicating that these genes,
although not focally expressed in eda mutant skin at E14 (Laurikkala
et al., 2002; Andl et al., 2002), are not directly regulated by Edar.
Interestingly, we noticed a downregulation of sostdc1 after prolonged
Eda treatment, a finding that is in fact consistent with the largely
non-overlapping expression patterns of sostdc1 and edar in vivo
(Laurikkala et al., 2003).
Based on the microarray, RT-qPCR, reporter gene, and expression
analyses in mouse embryos with altered levels of Eda signalling we
conclude that dkk4 and possibly also lrp4 are likely to be direct
transcriptional targets of Edar. Our in situ hybridization analysis
identified both dkk4 and lrp4 as novel markers of all skin appendage
placodes. While we were finalizing our manuscript, similar expression
patterns were also reported byothers (Weatherbee et al., 2006; Sick et
al., 2006; Bazzi et al., 2007). Dkk4 is a poorly characterized member of
the Dickkopf family of Wnt modulators and in the absence of a dkk4-
deficient mouse model its biological function has remained elusive.
However, TOPFLASH reporter assays have shown that Dkk4 can co-
operate with Kremen1/2 to suppress Wnt responses in a manner
analogous toDkk1 (Mao and Niehrs, 2003). In line with these data, our
skin explant studies suggest that Dkk4 acts as a Wnt inhibitor during
The mode of action of Lrp4 is not well established either. The
extracellular domain of Lrp4 resembles that of Lrp5 and Lrp6, the co-
receptors of Wnts involved in canonical Wnt cascade (Clevers 2006).
Lrp4 was shown to inhibit β-catenin-dependent signalling in vitro
(Johnson et al., 2005), and mice bearing null or hypomorphic alleles of
lrp4 present limb defects thought to reflect excess of Wnt activity
(Johnson et al., 2005; Simon-Chazottes et al., 2006; Weatherbee et al.,
2006) suggesting that Lrp4 acts as a Wnt antagonist in developing
ectoderm. However, the kidney abnormalities of lrp4 mutant mice are
reminiscent of Wnt loss-of-function phenotypes. Moreover, Lrp4 is
known to have Wnt-independent functions (May et al., 2007). The
clarification of the role of Lrp4 in skin appendage development must
await the detailed characterization of the lrp4 null phenotype
(Weatherbee et al., 2006).
Wnt activity precedes Eda signalling in nascent hair placodes
The expression of both dkk4 and lrp4 was detectable at E13.5 in
the first rows of hair placodes that emerge both dorsal and ventral
to the mammary line in wild-type embryos. This correlated well
Fig. 5. Dkk4 and lrp4 are expressed in eda-deficient skin at E14.5 in the absence of NF-kB activity. Whole-mount in situ hybridization with a probe specific to dkk4 (A–C, J, K), lrp4
(D–F, L) and edar (G), and whole-mount X-gal staining (H, I). Dkk4 (A, B) and lrp4 (D, E) are expressed in hair ‘pre-placodes’ in E14 eda−/−skin, but less intensely and in a perturbed
pattern compared to primary hair placodes in wild-type E14.5 skin (C, G). Neither edar transcripts (G), nor β-galactosidase activity (H) is detected in E14.5 the skin of NF-/eda−/−
embryos, whereas β-galactosidase activity is present in NF-kBREPprimary hair placodes (I). In E12 eda−/−lower jaws, strongly reduced amounts of dkk4 were observed in dental
placodes (J and K; arrows point to the location of incisor and molar placodes), while no gross changes in lrp4 expression were detected (L). i, incisor placode; m, molar placode; t,
tongue. Scale bars: 1 mm in panels A, D; 200 μm in panels B, C, E, F, G–L.
I. Fliniaux et al. / Developmental Biology 320 (2008) 60–71
I. Fliniaux et al. / Developmental Biology 320 (2008) 60–71
Fig. 6. Dkk4 and lrp4 show a similar expression pattern at E13.5 in lateral skin of wild-type and eda-deficient embryos, and are upregulated by both Wnt/β-catenin and Eda-A1/Edar
signalling. (A–D, F–I) In situ detection of dkk4 and lrp4 inwild-type and eda−/−E13.5 embryos. Expressionwas detected in the first hair follicles (black arrows). In addition, mammary
buds (indicated by numbers) and whiskers (white arrows) are positive. (E, J) Whole-mount X-gal staining of E13.5 NF-κBREPand NF-κBREP/eda−/−embryos. (K) Eda−/−skin explants
were cultured as depicted in Figs.1 and 2A in the absence or presence of 1 μg/ml Fc-Eda-A1, of 5 μM to 20 μM of the GSK3 specific inhibitor BIO, or of 1 μg/ml Fc-Eda-A1 and 5 μM BIO
together. BIO was able to induce dkk4 and lrp4 expression in a dose-dependent manner, and the combination of 1 μg/mlFc-Eda-A1 and 5 μM BIO led to a further increase of dkk4
transcripts. (L) Fc-Eda-A1 treatmentof eda−/−skin led tothe induction of dkk4 expression after inhibition of Wnt/β-cateninwith 100 μM and 400 μM CKI-7. Scale bars: 1 mm inpanels
A, C, F, H; 500 μm in panels B, D, E, G, I, J.
I. Fliniaux et al. / Developmental Biology 320 (2008) 60–71
with the appearance of NF-κB reporter activity. Surprisingly, we
found normal expression of both genes at E13.5 in the absence of
Eda and NF-κB reporter activity. These data unequivocally demon-
strate that Eda is dispensable for hair placode induction which is
thought to take place in response to dermal cues (Mikkola and
Millar, 2006). A day later, both dkk4 and lrp4 were still observable in
a localized manner in eda mutant embryos albeit at strongly
reduced levels. The expression of dkk4 and lrp4 in eda−/−embryos
is likely to be due to Wnt activity based on the fact that their
expression could be induced in embryonic skin by a brief treatment
of BIO, the GSK-3β inhibitor. In line with these data, the proximal
promoter of dkk4 was responsive to β-catenin/Lef1 in reporter
assays and contains multiple Tcf/Lef binding sites (Fig. 4B; Sick et al.,
2006; Bazzi et al., 2007).
Our findings place the Wnt pathway upstream of Eda/NF-κB, but it
remains to be determined whether any of the Eda pathway
Fig. 8. A model for the interplay between Eda and Wnt pathways during primary hair placode formation. Epithelial Wnt activity is involved in the induction of hair follicles, possibly
as an integral part of the response to the primary inductive signal(s) whose molecular identity is still unknown. In the next step, Eda and Wnt pathways may act in parallel in
generation of the correct pattern and stabilization of nascent hair placodes. They both are likely to induce the expression of activators and inhibitors of placode cell fate and share
some common target genes.
Fig. 7. Treatment of wild-type E13 skin with recombinant Dkk4 blocks primary hair placode formation in a dose-dependent manner. Wild-type E13 skin explants were cultured for
24 h in the absence (A) or presence of 0.25 (B), 0.5 (C) or 1 μg/ml (D) recombinant Dkk4 protein, and placode formation was monitored by shh in situ hybridization. A prominent
suppression of placode formation was obtained with 1 μg/ml Dkk4. (E) The percentage of placode forming area compared to the whole explant area was calculated for each culture
condition. Scale bar: 200 μm.
I. Fliniaux et al. / Developmental Biology 320 (2008) 60–71
components are directly regulated by Wnt signalling. We propose a
model (Fig. 8) where epithelial Wnt activity is essential during the
induction stage of hair development, while both Wnt and Eda
pathways may act in parallel in stabilization of placodes during the
propagation of the first wave of hair follicles. It seems that Eda and
Wnt pathways have several common target genes that are character-
ized by multiple conserved NF-κB and Tcf/Lef binding sites inter-
mingled in their promoters while they apparently have also unique
target genes essential for placode formation.
Complex molecular interactions regulate patterning of primary hair
In wild-type embryos the first wave of hair follicles sweeps over
the dorsum and head rapidly and a normal follicle pattern is
established between E13.5 and E14.5. In eda null embryos dkk4 and
lrp4 were expressed in a punctuate manner at E14.5, but with an
obviously irregular spacing indicating a patterning defect. It seems
plausible that Edar signalling that gets restricted to emerging
placodes strengthens the placode identity of the pre-placode cells
and thereby fixes the position of hair follicles. Our in vivo findings
are in line with a model based on in vitro manipulation of skin
explants with recombinant Eda suggesting that initially a labile,
Eda-independent prepattern of primary hair placodes is gener-
ated which is however, subject to subsequent modulation by
Edar activity (Mou et al., 2006). Unexpectedly, our results also
indicate that the reinforcement of placode cell fate by Eda is
not due to a direct positive effect on Wnt signalling although
we cannot rule out the possibility that Eda indirectly promotes
Our findings suggest that in addition to its role in expansion of
placodes (Mustonen et al., 2004), Eda pathway is essential for the
generation of a periodic pattern in the naïve embryonic ectoderm.
The prevailing paradigm to explain the appearance of periodicity
from an initially homogenous state (such as skin prior to hair follicle
development) is the reaction–diffusion model (Nagorcka, 1984; Jung
et al., 1998; Maini et al., 2006; Stark et al., 2007) originally
formulated by Turing (Turing, 1952). In its simplest form, the
model presumes the presence of two soluble factors: an activator
that promotes its own production/activity as well as the production
of its own inhibitor. When the inhibitor diffuses more rapidly and/or
decays faster than the activator, foci (placodes) with high activator
and inhibitor concentrations will appear. High local levels of the
activator will amplify the signal (in placodal cells) while the inhibitor
will prevent the reaction in the surroundings (i.e. impedes
neighbouring cells from adopting the follicle fate). The molecular
origin of periodic patterning of hair follicles remained obscure for
long. Recent studies have suggested that the interplay between Wnts
and their inhibitors Dkk1/4 plays a key role (Andl et al., 2002; Sick et
al., 2006). However, the identification of Eda as a potent inducer of
dkk4 expression adds more complexity to the simple mathematical
equations where hair follicle spacing was explained by Wnt-Dkk
specific reciprocal interactions only (Sick et al., 2006). Moreover, it is
apparent that the restriction of responsiveness to Eda by BMPs
through downregulation of edar expression in interplacodal cells is
another central aspect involved in hair follicle patterning (Mou et al.,
2006). Thus, the reaction–diffusion seems to involve multiple
signalling pathways with cross-talks, as already implied by Jiang et
Previous studies have established the role of Eda in several later
aspects of skin appendage development such as tooth morphogenesis,
nipple formation, and hair shaft production (Mikkola and Thesleff,
2003). More recent findings by us and others combined with those
reported here indicate that Eda plays a prominent role in placode
spacing and expansion by modulating BMP and Wnt pathways. The
identification of Dkk4, a placode inhibitor, as the most highly induced
gene in our microarray screen was a startling finding. The next task is
to recognize those Eda target genes that mediate its activator
functions. Although BMP inhibitors CTGF/CCN2 and follistatin (Mou
et al., 2006; Pummila et al., 2007) probably contribute to these
functions, other yet uncharacterized target genes are likely to be
We thank Riikka Santalahti, Raija Savolainen and Merja Mäkinen
for their excellent technical assistance, Géraldine Dejean for the
cloning of the lrp4 probe, Pascal Schneider (Department of Biochem-
istry, University of Lausanne, Switzerland) for providing recombinant
Fc-Eda-A1 protein, and Philip Barker (McGill University, Montreal,
Canada) for the NF-κBREPmice. This work was supported by the
Academyof Finland and Sigrid Juselius Foundation. I. F. is a recipientof
a Marie Curie Intra-European fellowship, and S. L. a Marie Curie Early
Stage Training fellow.
Andl, T., Reddy, S.T., Gaddapara, T., Millar, S.E., 2002. WNT signals are required for the
initiation of hair follicle development. Dev. Cell 2, 643–653.
Bazzi, H., Fantauzzo, K.A., Richardson, G.D., Jahoda, C.A., Christiano, A.M., 2007. The Wnt
inhibitor, Dickkopf 4, is induced by canonical Wnt signaling during ectodermal
appendage morphogenesis. Dev. Biol. 305, 498–507.
Bhakar, A.L., Tannis, L.L., Zeindler, C., Russo, M.P., Jobin, C., Park, D.S., MacPherson, S.,
Barker, P.A., 2002. Constitutive nuclear factor-kappa B activity is required for central
neuron survival. J. Neurosci. 22, 8466–8475.
Chijiwa, T., Hagiwara, M., Hidaka, H., 1989. A newly synthesized selective casein
kinase I inhibitor, N-(2-aminoethyl)-5-chloroisoquinoline-8-sulfonamide, and
affinity purification of casein kinase I from bovine testis. J. Biol. Chem. 264,
Chu, E.Y., Hens, J., Andl, T., Kairo, A., Yamaguchi, T.P., Brisken, C., Glick, A., Wysolmerski,
J.J., Millar, S.E., 2004. Canonical WNT signaling promotes mammary placode
development and is essential for initiation of mammary gland morphogenesis.
Development 131, 4819–4829.
Clevers, H., 2006. Wnt/beta-catenin signaling in development and disease. Cell 127,
Cui, C.Y., Durmowicz, M., Tanaka, T.S., Hartung, A.J., Tezuka, T., Hashimoto, K., Ko, M.S.,
Srivastava, A.K., Schlessinger, D., 2002. EDA targets revealed by skin gene
expression profiles of wild-type, Tabby and Tabby EDA-A1 transgenic mice. Hum.
Mol. Genet. 11, 1763–1773.
Cui, C.Y., Hashimoto, T., Grivennikov, S.I., Piao, Y., Nedospasov, S.A., Schlessinger, D.,
2006. Ectodysplasin regulates the lymphotoxin-beta pathway for hair differentia-
tion. Proc. Natl. Acad. Sci. U. S. A. 103, 9142–9147.
DasGupta, R., Fuchs, E., 1999. Multiple roles for activated LEF/TCF transcription
complexes during hair follicle development and differentiation. Development
Gaide, O., Schneider, P., 2003. Permanent correction of an inherited ectodermal
dysplasia with recombinant EDA. Nat. Med. 9, 614–618.
Hardy, M.H., 1992. The secret life of the hair follicle. Trends Genet. 8, 55–61.
Headon, D.J., Overbeek, P.A., 1999. Involvement of a novel Tnf receptor homologue in
hair follicle induction. Nat. Genet. 22, 370–374.
Headon, D.J., Emmal, S.A., Ferguson, B.M., Tucker, A.S., Justice, M.J., Sharpe, P.T., Zonana,
J., Overbeek, P.A., 2001. Gene defect in ectodermal dysplasia implicates a death
domain adapter in development. Nature 414, 913–916.
Itasaki, N., Jones, C.M., Mercurio, S., Rowe, A., Domingos, P.M., Smith, J.C., Krumlauf, R.,
2003. Wise, a context-dependent activator and inhibitor of Wnt signalling.
Development 130, 4295–4305.
Jiang, T.X., Widelitz, R.B., Shen, W.M.,Will, P., Wu, D.Y., Lin, C.M., Jung, H.S., Chuong, C.M.,
2004. Integument pattern formation involves genetic and epigenetic controls:
feather arrays simulated by digital hormone models. Int. J. Dev. Biol. 48, 117–135.
Johnson, E.B., Hammer, R.E., Herz, J., 2005. Abnormal development of the apical
ectodermal ridge and polysyndactyly in Megf7-deficient mice. Hum. Mol. Genet.14,
Jung, H.S., Francis-West, P.H., Widelitz, R.B., Jiang, T.X., Ting-Berreth, S., Tickle, C.,
Wolpert, L., Chuong, C.M., 1998. Local inhibitory action of BMPs and their
relationships with activators in feather formation: implications for periodic
patterning. Dev. Biol. 196, 11–23.
Kere, J., Srivastava, A.K., Montonen, O., Zonana, J., Thomas, N., Ferguson, B., Munoz, F.,
Morgan, D., Clarke, A., Baybayan, P., Chen, E.Y., Ezer, S., Saarialho-Kere, U., de la
Chapelle, A., Schlessinger, D., 1996. X-linked anhidrotic (hypohidrotic) ectodermal
dysplasia is caused by mutation in a novel transmembrane protein. Nat. Genet. 13,
Koppinen, P., Pispa, J., Laurikkala, J., Thesleff, I., Mikkola, M.L., 2001. Signaling and
subcellular localization of the TNF receptor Edar. Exp. Cell Res. 269, 180–192.
Kratochwil, K., Dull, M., Farinas, I., Galceran, J., Grosschedl, R., 1996. Lef1 expression is
activated by BMP-4 and regulates inductive tissue interactions in tooth and hair
development. Genes Dev. 10, 1382–1394.
I. Fliniaux et al. / Developmental Biology 320 (2008) 60–71
Kumar, A., Eby, M.T., Sinha, S., Jasmin, A., Chaudhary, P.M., 2001. The ectodermal Download full-text
dysplasia receptor activates the nuclear factor-kappaB, JNK, and cell death
pathways and binds to ectodysplasin A. J. Biol. Chem. 276, 2668–2677.
Laurikkala, J., Pispa, J., Jung, H.S., Nieminen, P., Mikkola, M., Wang, X., Saarialho-Kere, U.,
Galceran, J., Grosschedl, R., Thesleff, I., 2002. Regulation of hair follicle development
by the TNF signal ectodysplasin and its receptor Edar. Development 129,
Laurikkala, J., Kassai, Y., Pakkasjärvi, L., Thesleff, I., Itoh, N., 2003. Identification of a
secreted BMP antagonist, ectodin, integrating BMP, FGF, and SHH signals from the
tooth enamel knot. Dev. Biol. 264, 91–105.
Lin, C.M., Jiang, T.X., Widelitz, R.B., Chuong, C.M., 2006. Molecular signaling in feather
morphogenesis. Curr. Opin. Cell Biol. 18, 730–741.
Maini, P.K., Baker, R.E., Chuong, C.M., 2006. The Turing model comes of molecular age.
Science 314, 1397–1398.
Mao, B., Niehrs, C., 2003. Kremen2 modulates Dickkopf2 activity during Wnt/LRP6
signaling. Gene 302, 179–183.
May, P., Woldt, E., Matz, R.L., Boucher, P., 2007. The LDL receptor-related protein (LRP)
family: an old family of proteins with new physiological functions. Ann. Med. 39,
Meijer, L., Skaltsounis, A.L., Magiatis, P., Polychronopoulos, P., Knockaert, M., Leost, M.,
Ryan, X.P., Vonica, C.A., Brivanlou, A., Dajani, R., Dajani, R., Crovace, C., Tarricone, C.,
Musacchio, A., Roe, S.M., Pearl, L., Greengard, P., 2003. GSK-3-selective inhibitors
derived from Tyrian purple indirubins. Chem. Biol. 10, 1255–1266.
Mikkola, M.L., Thesleff, I., 2003. Ectodysplasin signaling in development. Cytokine
Growth Factor Rev. 14, 211–224.
Mikkola, M.L., Millar, S.E., 2006. The mammary bud as a skin appendage: unique
and shared aspects of development. J. Mammary Gland Biol. Neoplasia 11,
Monreal, A.W., Ferguson, B.M., Headon, D.J., Street, S.L., Overbeek, P.A., Zonana, J.,
1999. Mutations in the human homologue of mouse dl cause autosomal
recessive and dominant hypohidrotic ectodermal dysplasia. Nat. Genet. 22,
Mou, C., Jackson, B., Schneider, P., Overbeek, P.A., Headon, D.J., 2006. Generation of the
primary hair follicle pattern. Proc. Natl. Acad. Sci. U. S. A. 103, 9075–9080.
Mustonen, T., Pispa, J., Mikkola, M.L., Pummila, M., Kangas, A.T., Pakkasjärvi, L., Jaatinen,
R., Thesleff, I., 2003. Stimulation of ectodermal organ development by Ectodyspla-
sin-A1. Dev. Biol. 259, 123–136.
Mustonen, T., Ilmonen, M., Pummila, M., Kangas, A.T., Laurikkala, J., Jaatinen, R., Pispa, J.,
Gaide, O., Schneider, P., Thesleff, I., Mikkola, M.L., 2004. Ectodysplasin A1 promotes
placodal cell fate during early morphogenesis of ectodermal appendages. Devel-
opment 131, 4907–4919.
Nagorcka, B.N., 1983–1984. Evidence for a reaction–diffusion system as a mechanism
controlling mammalian hair growth. Biosystems 16, 323–332.
Orange, J.S., Levy, O., Geha, R.S., 2005. Human disease resulting from gene mutations
that interfere with appropriate nuclear factor-kappaB activation. Immunol. Rev.
O'Shaughnessy, R.F., Yeo, W., Gautier, J., Jahoda, C.A., Christiano, A.M., 2004. The WNT
signalling modulator, Wise, is expressed in an interaction-dependent manner
during hair-follicle cycling. J. Invest. Dermatol. 123, 613–621.
Pispa, J., Thesleff, I., 2003. Mechanisms of ectodermal organogenesis. Dev. Biol. 262,
Pispa, J., Jung, H.S., Jernvall, J., Kettunen, P., Mustonen, T., Tabata, M.J., Kere, J., Thesleff, I.,
1999. Cusp patterning defect in Tabby mouse teeth and its partial rescue by FGF.
Dev. Biol. 216, 521–534.
Pispa, J., Mikkola, M.L., Mustonen, T., Thesleff, I., 2003. Ectodysplasin, Edar and
TNFRSF19 are expressed in complementaryand overlapping patterns during mouse
embryogenesis. Gene Expr. Patterns 3, 675–679.
Price, M.A., 2006. CKI, there's more than one: casein kinase I family members in Wnt
and Hedgehog signaling. Genes Dev. 20, 399–410.
Pummila, M., Fliniaux, I., Jaatinen, R., James, M.J., Laurikkala, J., Schneider, P., Thesleff, I.,
Mikkola, M.L., 2007. Ectodysplasin has a dual role in ectodermal organogenesis:
inhibition of Bmp activity and induction of Shh expression. Development 134,
Schmidt-Ullrich, R., Aebischer, T., Hülsken, J., Birchmeier, W., Klemm, U., Scheidereit, C.,
2001. Requirement of NF-kappaB/Rel for the development of hair follicles and other
epidermal appendices. Development 128, 3843–3853.
Schmidt-Ullrich, R., Tobin, D.J., Lenhard, D., Schneider, P., Paus, R., Scheidereit, C., 2006.
NF-kappaB transmits Eda A1/EdaR signalling to activate Shh and cyclin D1
expression, and controls post-initiation hair placode down growth. Development
Sick, S., Reinker, S., Timmer, J., Schlake, T., 2006. WNT and DKK determine hair follicle
spacing through a reaction–diffusion mechanism. Science 314, 1447–1450.
Simon-Chazottes, D., Tutois, S., Kuehn, M., Evans, M., Bourgade, F., Cook, S., Davisson,
M.T., Guénet, J.L., 2006. Mutations in the gene encoding the low-density
lipoprotein receptor LRP4 cause abnormal limb development in the mouse.
Genomics 87, 673–677.
Stark, J., Andl, T., Millar, S.E., 2007. Hairy math: insights into hair-follicle spacing and
orientation. Cell 128, 17–20.
Srivastava, A.K., Pispa, J., Hartung, A.J., Du, Y., Ezer, S., Jenks, T., Shimada, T., Pekkanen, M.,
Mikkola, M.L., Ko, M.S., Thesleff, I., Kere, J., Schlessinger, D., 1997. The Tabby
phenotype is caused by mutation in a mouse homologue of the EDA gene that
reveals novel mouse and human exons and encodes a protein (ectodysplasin-A)
with collagenous domains. Proc. Natl. Acad. Sci. U. S. A. 94, 13069–13074.
Turing, A.M.,1952. The chemical basis of morphogenesis. Philos. Trans. R. Soc. Lond. 237,
Van Genderen, C., Okamura, R.M., Fariñas, I., Quo, R.G., Parslow, T.G., Bruhn, L.,
Grosschedl, R., 1994. Development of several organs that require inductive
epithelial–mesenchymal interactions is impaired in LEF-1-deficient mice. Genes
Dev. 8, 2691–2703.
Veltmaat, J.M., Mailleux, A.A., Thiery, J.P., Bellusci, S., 2003. Mouse embryonic
mammogenesis as a model for the molecular regulation of pattern formation.
Differentiation 71, 1–17.
Vielkind, U., Hardy, M.H., 1996. Changing patterns of cell adhesion molecules during
mouse pelage hair follicle development. 2. Follicle morphogenesis in the hair
mutants, Tabby and downy. Acta Anat. (Basel) 157, 183–194.
Weatherbee, S.D., Anderson, K.V., Niswander, L.A., 2006. LDL-receptor-related protein 4
is crucial for formation of the neuromuscular junction. Development 133,
Yan, M., Wang, L.C., Hymowitz, S.G., Schilbach, S., Lee, J., Goddard, A., de Vos, A.M.,
Gao, W.Q., Dixit, V.M., 2000. Two-amino acid molecular switch in an epithelial
morphogen that regulates binding to two distinct receptors. Science 290,
I. Fliniaux et al. / Developmental Biology 320 (2008) 60–71