Content uploaded by Johanna Maria Margareta van Oers
Author content
All content in this area was uploaded by Johanna Maria Margareta van Oers
Content may be subject to copyright.
Research article
The Journal of Clinical Investigation http://www.jci.org Volume 116 Number 8 August 2006 2201
Mosaicism of activating FGFR3 mutations
in human skin causes epidermal nevi
Christian Hafner,
1
Johanna M.M. van Oers,
2
Thomas Vogt,
1
Michael Landthaler,
1
Robert Stoehr,
3
Hagen Blaszyk,
4
Ferdinand Hofstaedter,
5
Ellen C. Zwarthoff,
2
and Arndt Hartmann
5
1
Department of Dermatology, University of Regensburg, Regensburg, Germany.
2
Department of Pathology, Josephine Nefkens Institute, Erasmus MC,
Rotterdam, The Netherlands.
3
Department of Urology, University of Regensburg, Regensburg, Germany.
4
Department of Pathology,
University of Vermont College of Medicine, Burlington, Vermont, USA.
5
Institute of Pathology, University of Regensburg, Regensburg, Germany.
Epidermal nevi are common congenital skin lesions with an incidence of 1 in 1,000 people; however, their
genetic basis remains elusive. Germline mutations of the FGF receptor 3 (FGFR3) cause autosomal dominant
skeletal disorders such as achondroplasia and thanatophoric dysplasia, which can be associated with acan-
thosis nigricans of the skin. Acanthosis nigricans and common epidermal nevi of the nonorganoid, nonepi-
dermolytic type share some clinical and histological features. We used a SNaPshot multiplex assay to screen
39 epidermal nevi of this type of 33 patients for 11 activating FGFR3 point mutations. In addition, exon 19 of
FGFR3 was directly sequenced. We identified activating FGFR3 mutations, almost exclusively at codon 248
(R248C), in 11 of 33 (33%) patients with nonorganoid, nonepidermolytic epidermal nevi. In 4 of these cases,
samples from adjacent histologically normal skin could be analyzed, and FGFR3 mutations were found to be
absent. Our results suggest that a large proportion of epidermal nevi are caused by a mosaicism of activating
FGFR3 mutations in the human epidermis, secondary to a postzygotic mutation in early embryonic develop-
ment. The R248C mutation appears to be a hot spot for FGFR3 mutations in epidermal nevi.
Introduction
Epidermal nevi show a prevalence of about 1 in 1,000 people and
can be divided into either nonorganoid (keratinocytic) types or
organoid types characterized by hyperplasia of adnexal structures
such as sebaceous glands, sweat glands, and hair follicles. Epider
-
mal nevi of the common, nonorganoid and nonepidermolytic type
are benign skin lesions and may vary in their extent from a single
(usually linear) lesion to widespread and systematized involvement
(Figure 1). They may be present at birth or develop early during
childhood as localized epidermal thickening with hyperpigmenta
-
tion, frequently following the lines of Blaschko. This suggests that
epidermal nevi may be due to mosaicism resulting from postzygot
-
ic mutations in keratinocytes. Mutations of keratins 1 and 10 were
shown to be responsible for a rare subgroup of epidermal nevi,
the linear epidermolytic hyperkeratosis (1, 2). Another variant of
epidermal nevi, the congenital hemidysplasia with ichthyosiform
nevus and limb defects (CHILD)
nevus, is caused by NADPH ste-
roid dehydrogenase-like protein (NSDHL) mutations (Xq28) and
represents a functional X chromosomal mosaicism (3). However,
the genetic basis of the much more common nonorganoid, non
-
epidermolytic keratinocytic epidermal nevi remains elusive.
The FGF receptor (FGFR) family comprises 4 major trans
-
membrane receptor tyrosine kinases (FGFR1–4) and is involved
in embryogenesis, angiogenesis, and tissue homeostasis (4). The
FGFR3 gene contains 19 exons encoding an extracellular region
for ligand binding composed of 3 Ig-like domains, a hydropho
-
bic transmembrane domain and 2 cytoplasmic tyrosine kinase
domains (Figure 2). Alternative splicing of the second half of the
IgIII-like domain (exon 8 versus exon 9) results in the isoforms
IIIb and IIIc. The isoforms show different ligand specificity and
tissue expression.
FGFR3 IIIb is mainly expressed in epithelial cells
while
FGFR3 IIIc is predominantly found in mesenchymal cells (5).
This paper assigns all codon numbers to the open reading frame of
the
FGFR3 IIIb isoform. More than 20 FGFs are known as ligands
now (4). The interaction of the ligand with the receptor requires
the presence of sulfated glycosaminoglycans such as heparin and
leads to the dimerization of the receptor with consecutive phos
-
phorylation of intracellular tyrosine residues in the kinase domain
and activation of intracellular signaling pathways. The autophos
-
phorylation sites of FGFR3 represent potential binding sites for
signaling proteins, for example, with phosphotyrosine binding
(PTB) and Src homology 2 (SH2) domains. Activation of the cyto
-
plasmic region causes phosphorylation of Shp2, PLC
g, ERK1/2,
and STAT3 and also PI3K activation (4). FGFRs can also be acti
-
vated by interaction with EphA4, another receptor tyrosine kinase,
which demonstrates the complexity of FGFR signaling (6).
Activating germline mutations of the
FGFR3 gene result in dwarf-
ism, severe skeletal dysplasia, and craniosynostosis syndromes such
as achondroplasia (ACH), hypochondroplasia (HCH), thanato
-
phoric dysplasia (TD), Crouzon syndrome (CS), Muenke syndrome
(MS), and SADDAN (severe achondroplasia with developmental
delay and acanthosis nigricans) syndrome (7–10). These activating
mutations result in negative chondrocytic growth regulation of
the epiphyseal plates of long bones, causing dwarfism (11). Identi
-
cal mutations are found in different cancer entities and probably
provide proliferative signals (4). FGFR3 signaling in mutated cells
is poorly understood, but previous studies provide some insights.
All known missense mutations causing TD I create an unpaired
cysteine residue (12–14), such as the R248C and S249C mutations.
These mutations are localized in the extracellular domain at the
linker region between the Ig-like domains II and III while other
Nonstandard abbreviations used: ACH, achondroplasia; CS, Crouzon syndrome;
FGFR3, fibroblast growth factor receptor 3; HCH, hypochondroplasia; SADDAN,
severe achondroplasia with developmental delay and acanthosis nigricans;
TD, thanatophoric dysplasia.
Conflict of interest: The authors have declared that no conflict of interest exists.
Citation for this article:
J. Clin. Invest. 116:2201–2207 (2006). doi:10.1172/JCI28163.
research article
2202 The Journal of Clinical Investigation http://www.jci.org Volume 116 Number 8 August 2006
mutations causing unpaired cysteine residues such as G372C are
localized at the junction of the extracellular and the transmem
-
brane domain of
FGFR3. The newly created unpaired cysteine
amino acid allows the formation of disulfide bonds between the
extracellular domains of 2 receptors, resulting in homodimeriza
-
tion, increased tyrosine phosphorylation, and ligand-independent
constitutive receptor activation (15, 16). Similar mechanisms were
also shown for other receptor tyrosine kinases, the erythropoietin
receptor and the epidermal growth factor receptor (17, 18). Con
-
stitutive activation in cells bearing the G372C or S373C mutation
results in high basal phosphorylation with significantly increased
constitutive levels of MAPK phosphorylation and c-fos transcrip
-
tion, probably caused by mutant homodimer FGFR3 complexes
(19). Other mutations, such as G382R, which is found in about
97% of ACH patients, affect the transmembrane domain and lead
to the formation of hydrogen bonds between 2 FGFRs, resulting
in constitutive receptor activation (20). The A393E mutation in
CS also affects the transmembrane domain and results in FGFR3
dimer stabilization, measured by the change in the free energy of
the dimerization, thus largely increasing the fraction of dimers
(21). A third class of mutations affect the tyrosine kinase domain,
such as the K652E mutation in TD II or the N542K mutation in
HCH (11). These mutations likely cause conformational changes
in the activation loop that activate the receptor tyrosine kinase
activity and downstream ERK1/2 (22). Other studies suggest that
FGFR3 mutations may delay the downregulation and ligand-medi-
ated internalization of the receptor (23). The phosphorylated
immature form of the mutant receptor accumulates in the endo
-
plasmic reticulum and fails to be degraded (24). The different
degree of receptor activation seems to correlate with the severity of
the phenotype. Stronger activation of the receptor by ligand-inde
-
pendent dimerization via disulfide bonds in TD patients deter
-
mines the more severe phenotype compared with other skeletal
dysplasia syndromes such as ACH and HCH (15, 25).
Somatic activating
FGFR3 mutations have been identified in 40%
of human seborrheic keratoses (26) and in several human cancers
(4), including multiple myeloma (27), urothelial carcinoma (28),
cervix carcinoma (29), and colorectal carcinoma (30). Some of the
skeletal dysplasia syndromes (TD, CS, SADDAN) caused by
FGFR3
mutations are also characterized by marked thickening of the epi
-
dermis. This skin lesion, termed
acanthosis nigricans, and epidermal
nevi share similar histological features, including acanthosis and
papillomatosis (31, 32). Herein we investigate the role of
FGFR3
mutations in common nonorganoid, nonepidermolytic keratino
-
cytic epidermal nevi.
Results
We analyzed 39 common nonepidermolytic, nonorganoid kera
-
tinocytic epidermal nevi of 33 patients using a SNaPshot mul
-
tiplex assay that covered 11
FGFR3 point mutations described
in skeletal dysplasia syndromes and cancer entities (Figure 2).
The following subtypes of keratinocytic epidermal nevi were
Figure 1
Patient 29 displayed a systematized epidermal nevus of the common
soft type with involvement of the face (bilateral), the right scapular region,
the right arm, the right hip, and the right thigh. Abnormalities of the skel-
etal or nervous system were not present. A biopsy was taken from the
epidermal nevus of the right forearm. This epidermal nevus revealed an
R248C FGFR3 gene mutation. DNA isolated from the blood of this patient
revealed WT status at codon 248, excluding a germline mutation.
Figure 2
FGFR3 gene. The position of the muta-
tions covered by the SNaPshot multiplex
assay is indicated. Codons are numbered
according to the FGFR3 IIIb isoform;
potential mutations of the stop codon 809
in exon 19 associated with TD I were ana-
lyzed by direct sequencing. C, C-terminus;
Ig I, Ig II, Ig III, Ig-like domains I–III; N,
N-terminus; TM, transmembrane domain;
TKI, TKII, tyrosine kinase domains I–II.
research article
The Journal of Clinical Investigation http://www.jci.org Volume 116 Number 8 August 2006 2203
included in the study: 28 common soft type, 5 common hard
type, 4 seborrheic-like type, and 2 acrokeratosis verruciformis-
like type. We identified activating
FGFR3 mutations in 11 of 33
(33%) patients and in 16 of 39 (41%) epidermal nevi (Table 1). In
15 of 16 epidermal nevi (94%), the mutation occurred in exon 7
at codon 248 (CGC to TGC), resulting in an Arg to Cys amino
acid substitution. In one epidermal nevus (patient 15), a double
mutation was found in exon 10 at codon 372 (GGC to TGC),
resulting in a Gly to Cys amino acid substitution, and at codon
382 (GGG to AGG), resulting in a Gly to Arg amino acid sub
-
stitution. The epidermal nevus with the double mutation was
a common soft-type nevus and showed no histological abnor
-
malities compared with the other epidermal nevi. All activating
FGFR3 mutations were found in common-
type epidermal nevi (14 common soft type,
and 2 common hard type).
In 4 patients (patients 3, 17, 21, and 32) with
an
FGFR3 mutation (R248C) in the epidermal
nevus, clinically and histologically normal epi
-
dermis adjacent to the nevus showed a WT codon
248, suggesting an epidermal mosaicism of the
FGFR3 mutation and a strong genotype-pheno-
type correlation (Figure 3). In 1 patient (patient
29) with a systematized epidermal nevus display
-
ing the R248C mutation (biopsy was taken from
the right forearm), additional genomic DNA was
isolated from blood. The DNA revealed the WT
codon 248, thus excluding a germline mutation.
Multiple intraindividual epidermal nevi biop
-
sies could be analyzed in 2 patients. One patient
(patient 11a, b) did not show any
FGFR3 muta-
tions. The other patient (patient 33a–e) under
-
went ablative laser treatment of the common
soft-type epidermal nevus at the right side of the
neck (Figure 4). Before treatment, 6 of the scat
-
tered brownish papules were curetted for
FGFR3
mutation analysis. All 6 samples, which were
spatially distant from each other, revealed the
R248C mutation in the SNaPshot analysis.
All
FGFR3 mutations detected so far in skin
lesions (seborrheic keratoses and epidermal nevi)
are associated with TD, CS, and SADDAN syn
-
drome in the germline. The SNaPshot multiplex
assay covered all mutations responsible for these
syndromes except for the stop codon mutation
X809L/G/R/C/W in exon 19 causing TD I (Fig
-
ure 2). For 20 epidermal nevi (nevi of patients 1,
2, 5, 6, 7, 8, 9, 10, 11 [a, b], 14, 16, 17, 20, 23, 27,
28, 30, 31, and 32), DNA was available for fur
-
ther analysis. We additionally sequenced exon 19
of these samples. Seventeen of the 20 epidermal
nevi had not shown any
FGFR3 mutations in the
SNaPshot analysis. However, no further
FGFR3
mutations were detected in exon 19.
Since epidermal nevi represent a heteroge
-
neous skin disorder, we also studied an organ
-
oid type of epidermal nevus, the sebaceous
nevus. This nevus is found almost exclusively
on the scalp or face and is usually present at
birth. Sebaceous nevi are histopathologically
characterized by the presence of large numbers of mature seba
-
ceous glands and papillomatous hyperplasia of the epidermis. We
analyzed 13 sebaceous nevi for
FGFR3 point mutations using the
SNaPshot multiplex assay (Table 2). In contrast to the common
nonorganoid keratinocytic nevi, this organoid type of epidermal
nevus revealed no
FGFR3 mutations. Thus, activating FGFR3 muta-
tions, especially the frequently detected R248C mutation, appear
to be specific for common nonorganoid keratinocytic nevi.
Discussion
Our results indicate that a significant number of epidermal nevi
of the common, nonorganoid, and nonepidermolytic keratinocytic
type are caused by postzygotic mutations in the
FGFR3 gene, which
Table 1
FGFR3 mutations in epidermal nevi
No. Sex Age (yr) Site Subtype Nevus Control Syst.
1 F 1 Head Seborrheic WT – No
2 M 59 Head Acro-verruc WT – –
3 F 9 Left axilla Common hard R248C WT –
4 M 13 Neck Seborrheic WT – No
5 F 36 Right axilla Common soft R248C – No
6 M 17 Head Common soft WT – –
7 M 7 Head Common soft WT – –
8 M 21 Head Common soft WT – –
9 M 15 Head Common soft WT – –
10 F 15 Right shoulder Common hard WT – –
11a M 22 Head Common soft WT – No
11b M 22 Head Common soft WT – No
12 M 20 Left groin Common soft R248C – No
13 F 13 Back Common soft WT – –
14 F 23 Back Common soft WT – –
15 F 5 Left axilla Common soft G372C/G382R – –
16 F 2 Neck Common soft WT – No
17 M 6 Neck Common soft R248C WT Yes
18 M 16 Head Common soft WT – No
19 F 27 Neck Common soft WT – –
20 F 25 Right shoulder Common soft WT – –
21 M 29 Head Common hard R248C WT No
22 F 20 Right arm Common soft R248C – –
23 F 16 Trunk Common hard WT – No
24 M 31 Head Common soft WT – No
25 F 18 Back Seborrheic WT – –
26 M 9 Neck Common soft R248C – Yes
27 M 14 Head Seborrheic WT – No
28 F 9 Head Acro-verruc WT – No
29 F 11 Right forearm Common soft R248C – Yes
30 F 15 Head Common hard WT – No
31 F 16 Neck Common soft WT – –
32 M 23 Back Common soft R248C WT No
33a F 13 Neck Common soft R248C – Yes
33b F 13 Neck Common soft R248C – Yes
33c F 13 Neck Common soft R248C – Yes
33d F 13 Neck Common soft R248C – Yes
33e F 13 Neck Common soft R248C – Yes
33f F 13 Neck Common soft R248C – Yes
Age, age at time of biopsy; site, site of biopsy; control, DNA from clinically and histologically
normal epidermis adjacent to the epidermal nevus; syst., systematized epidermal nevus;
M, male; F, female; acro-verruc, acrokeratosis verruciformis-like; R248C, exon 7, codon 248,
CGC to TGC with Arg to Cys; G372C, exon 10, codon 372, GGC to TGC with Gly to Cys;
G382R, exon 10, codon 382, GGG to AGG with Gly to Arg. Codon numbers refer to the open
reading frame of the FGFR3 IIIb isoform.
research article
2204 The Journal of Clinical Investigation http://www.jci.org Volume 116 Number 8 August 2006
likely occur early in embryonic development in a single keratinocyt-
ic stem cell. The descendants of the mutated stem cell will migrate
along the lines of Blaschko, resulting in the linear or whirled cuta
-
neous patterns usually observed in epidermal nevi (33).
Several findings support the notion that the detected
FGFR3
mutations are causative for the development of epidermal nevi:
(a) The detected mutations R248C and, in one case, G372C are
known to occur in TD I patients and act as dominant germline
mutations in this lethal skeletal dysplasia syndrome. Affected
individuals are often stillborn or die as neonates. It has previously
been postulated that such lethal mutations may only survive by
mosaicism (34). The R248C mutation results in strong constitu
-
tive and ligand-independent receptor activation (15), excluding
the possibility that the detected mutations have no functional
consequence in the investigated skin samples. (b) Mutations
responsible for TD I (including R248C and G372C) have been
shown to cause another benign skin lesion, the seborrheic kerato
-
sis (26). Epidermal nevi and seborrheic keratoses share many clini
-
cal and histopathological features. Another activating mutation
(S249C) associated with TD I causes thickening of the skin and
verrucous skin tumors in a mouse model. The murine skin lesions
resemble seborrheic keratoses (and likewise epidermal nevi) clini
-
cally and histologically (26). (c) A TD I patient with an R248C
mutation and long-term survival developed acanthosis nigricans,
a skin disorder that is clinically and histologically very similar to
epidermal nevi (10). Both acanthosis nigricans and epidermal nevi
show acanthosis, papillomatosis, hyperkeratosis, and basal
hyperpigmentation (31, 32). A potential overlap between
the 2 skin disorders is acknowledged since both acanthosis
nigricans type of epidermal nevus and nevoid acanthosis
nigricans have been described (32, 35, 36). In our series of
epidermal nevi, an acanthosis nigricans-like type was not
included. A female patient with a mosaicism of R248C (25%
of the blood lymphocytes were affected by the mutation)
developed disseminated thickening and hyperpigmenta
-
tion of the skin consistent with acanthosis nigricans (37).
(d) Some patients with epidermal nevus syndrome showed
typical keratinocytic epidermal nevi, skeletal abnormalities,
and the occurrence of urothelial carcinoma at an early age
(38–40). This correlation of epidermal nevus and urothelial
carcinoma is thought to be nonstochastic.
FGFR3 muta-
tions are frequent events in papillary urothelial carcinoma
(29, 41). These findings and the skeletal abnormalities rem
-
iniscent of skeletal dysplasia syndromes strongly suggest
that such patients feature a mosaicism of activating
FGFR3
mutations, which in turn cause epidermal nevi of the skin,
skeletal changes, and bladder cancer. (e) In one patient of
our series (patient 33), multiple intraindividual samples of
the epidermal nevus localized on the neck could be analyzed.
All 6 samples revealed the same R248C mutation, suggest
-
ing the presence of a scattered
FGFR3 mutation mosaicism in the
skin of this patient following the lines of Blaschko. (f) The stron
-
gest support that the detected mutations are indeed causative for
epidermal nevi is that 4 patients with an R248C mutation in the
epidermal nevus showed the WT allele in clinically and histologi
-
cally normal epidermis adjacent to the epidermal nevus. This also
suggests mosaicism of an
FGFR3 mutation in the epidermis and a
strong genotype-phenotype correlation.
In our series, 15 of 16 epidermal nevi with an
FGFR3 mutation
displayed the R248C mutation, resulting in a cytosine-thymine
substitution (CGC to TGC). This C to T transition is a typical
example of deamination of methylated cytosines leading to CpG
dinucleotide depletion. The reason for this mutational hot spot
is unknown. One could speculate that the development of epi
-
Figure 3
Patient 32 had a common soft-type epidermal nevus on his back. Two biop-
sies were taken from the epidermal nevus and the adjacent normal skin.
The epidermal nevus histologically showed the typical acanthosis and papil-
lomatosis (H&E staining; original magnification, ×40) and an R248C muta-
tion in the SNaPshot analysis. In contrast, the clinically and histologically
normal epidermis revealed a WT status for codon 248. This result suggests
a strong genotype-phenotype correlation and the presence of a mosaicism
of the FGFR3 mutation in the epidermis of this patient.
Figure 4
Patient 33 revealed a common soft-type epidermal nevus at the right
side of the neck consisting of scattered brownish papules following
the lines of Blaschko. Before ablative laser treatment, 6 papules dis-
tant from each other were curetted for FGFR3 mutation analysis. The
green peaks represent WT codon 375, the black peaks WT codon 248.
The R248C mutation is characterized by a red peak. All 6 samples
displayed the R248C mutation. In the bottom right corner, control DNA
with WT codon 248 is shown.
research article
The Journal of Clinical Investigation http://www.jci.org Volume 116 Number 8 August 2006 2205
dermal nevi with acanthosis and papillomatosis of the epidermis
requires strong activation of the receptor and not all activating
FGFR3 mutations are capable of this strong activation. Indeed,
Naski et al. found that the R248C mutation activates the receptor
more strongly than the G382R mutation responsible for ACH (15).
Stronger activation of the receptor is thought to result in a more
severe phenotype in TD compared with other skeletal dysplasia
syndromes (14). According to this theory, mosaicism of
FGFR3
mutations other than R248C may also be present in the skin, but
they may be insufficient to induce epidermal nevi. This hypoth
-
esis is supported by Logie et al. (26), who found that seborrheic
keratoses sharing some histological characteristics with epidermal
nevi are caused by acquired somatic
FGFR3 mutations. All somatic
mutations detected so far in seborrheic keratoses (R248C, S249C,
G372C, S373C, Y375C, K652E, and K652M) are associated with
TD and SADDAN syndrome in germline. Remarkably, none of
the mutations associated with ACH or HCH were found in this
series of 62 seborrheic keratoses. Patients with TD and SADDAN
syndrome show acanthosis nigricans whereas ACH and HCH are
usually not associated with this skin lesion except for in 1 reported
case (42). ACH is the most common cause of dwarfism in humans,
with an incidence of 1:15,000 to 1:40,000 of live births, and the
mutated nucleotide 1138 in ACH is thought be the most muta
-
ble nucleotide in the human genome described so far (9, 43). The
R248C mutation is the most common mutation in TD (44), but
it remains elusive why other mutations causing TD and SADDAN
syndrome (S249C, S373C, Y375C, K652E, K652M) were not
detected in our series of epidermal nevi.
Epidermal nevus syndromes are characterized by epidermal
nevi, abnormalities of the skeletal and nervous system, and
rarely, some associated cancer entities (45, 46). The occurrence
of papillary bladder cancer and skeletal abnormalities in epider
-
mal nevus syndromes (38–40) suggests that
FGFR3 is a promis-
ing candidate gene for epidermal nevus syndromes that may be
caused by a more widespread mosaicism of
FGFR3 mutations. If
the
FGFR3 mosaicism involves the germ cells in those patients,
the offspring should show a TD phenotype. However, we did not
find any reports of patients with systematized epidermal nevi or
epidermal nevus syndrome and an offspring with TD in the litera
-
ture. Another interesting fact is that acanthosis nigricans is pre
-
dominantly observed in intertriginous areas in skeletal dysplasia
syndromes. The reason for this preferential localization remains
unclear since the entire skin carries the mutation. There seems
to be a preponderance of intertriginous localization (neck, axilla,
groin) also in the epidermal nevi with
FGFR3 mutations in our
series (Table 1). Additional cofactors in intertriginous areas may
favor the development of acanthosis and papillomatosis on the
basis of activating
FGFR3 mutations.
One epidermal nevus in our series displayed a double mutation,
the G372C mutation known from TD I and the typical G382R
mutation known from ACH. It remains elusive whether the G372C
mutation alone, which causes a stronger activation of the FGFR3
receptor according to the associated skeletal dysplasia syndromes
(13–15), was able to cause the epidermal nevus or whether the
combination of both mutations was necessary for the induction
of the nevus. The possibility that the nevus was mainly caused by
the G382R mutation seems unlikely since ACH patients, in con
-
trast to TD patients, usually do not develop acanthosis nigricans
(10, 42). Double mutations in the
FGFR3 gene have been reported
in urothelial carcinoma (28, 47, 48). To our knowledge, germline
compound heterozygosity for mutations associated with TD and
ACH has not been described. The epidermal nevus carrying the
double mutation displayed no histopathological abnormalities.
Each of the major 3 cell types present in our microdissected
samples (keratinocytes, dendritic cells, and melanocytes) may
theoretically be the carrier of the detected mutations, but kera
-
tinocytes represent the likely host cell type. Logie et al. studied a
mouse model in which the S249C mutation (causing TD I) was
targeted to the basal layer of the epidermis using the keratin 5 pro
-
moter (26). Keratin 5 is a marker for basal keratinocytes and not
expressed in melanocytes or dendritic cells. The transgenic mice
developed thickening of the skin and verrucous skin tumors with
histological features similar to epidermal nevi. The basal hyper
-
pigmentation mediated by melanocytes would be a secondary phe
-
nomenon if the mutations affect keratinocytes. However, further
studies are needed to prove this hypothesis.
It is unclear why the same mutation (R248C) can cause both
seborrheic keratoses and epidermal nevi.
FGFR3 mutations exhibit
pleiotropic effects ranging from inherited skeletal dysplasia syn
-
dromes and benign skin tumors to cancer. The cell type–specific
involvement of different signaling pathways such as Ras/MAPK
and STAT as well as the FGFR3-dependent recruitment of cell-spe
-
cific second receptors such as EphA4 (6) may determine the effect
of activating
FGFR3 mutations in each cell type (4).
Mutations of
FGFR3 in the urothelium are significantly associ-
ated with benign urothelial papillomas (49) and with low-grade
and low-stage pTa G1/2 tumors, which rarely progress (41, 48,
50). These findings support the concept that mutations of the
FGFR3 IIIb isoform can induce proliferation and tumor forma-
tion but are associated with a low malignant potential. This
would be consistent with our findings of
FGFR3 mutations in
epidermal nevi that represent a benign skin disorder and show
signs of hyperproliferation (namely acanthosis and papilloma
-
tosis) but bear no malignant potential.
Several small molecule tyrosine kinase inhibitors of FGFR3, such
as PKC412, PD173074, and SU5402, are already available. They
have been used in vitro and in animal models to inhibit the growth
of multiple myeloma cell lines with activating
FGFR3 mutations
(51–54). PKC412 is currently being evaluated in phase II trials for
acute myeloid leukemia patients (55). This molecule effectively
inhibits the tyrosine kinase activity of FGFR3, as shown by the
Table 2
FGFR3 mutation analysis in sebaceous nevi
No. Sex Age (yr) Site Sebaceous nevus
1 M 17 Head WT
2 M 15 Head WT
3 F 7 Head WT
4 M 12 Head WT
5 M 33 Head WT
6 M 17 Head WT
7 M 7 Head WT
8 M 14 Head WT
9 M 13 Head WT
10 F 16 Head WT
11 M 12 Head WT
12 M 32 Head WT
13 M 13 Head WT
research article
2206 The Journal of Clinical Investigation http://www.jci.org Volume 116 Number 8 August 2006
inhibition of proliferation in hematopoietic cells transformed by
a mutant
FGFR3 gene (55). A broader use of this drug in other
disorders associated with an increased activity of FGFR3 has been
suggested. This may also include benign skin lesions such as seb
-
orrheic keratoses (26) and epidermal nevi. The current standard
therapy for epidermal nevi and seborrheic keratoses is surgery or
ablative laser treatment, which is often associated with irreversible
scar formation. Topical application of tyrosine kinase inhibitors of
FGFR3 may obviate the need for surgical intervention.
In summary, a large proportion of human epidermal nevi are
caused by mosaicism of postzygotic activating
FGFR3 mutations
in the human epidermis. The R248C mutation appears to be a hot
spot for
FGFR3 mutations in epidermal nevi. Additional studies
are needed to investigate other regions of
FGFR3 and different
receptor tyrosine kinases as possible mutational targets in epi
-
dermal nevi and elucidate functional aspects of enhanced FGFR3
signaling in the skin.
Methods
Microdissection. Thirty-nine histologically confirmed nonepidermolytic
epidermal nevi of the common nonorganoid type of 33 patients with
varying degrees of skin involvement were retrieved from the histol-
ogy files of the Department of Dermatology, University of Regensburg.
Informed consent for the scientific use of the material and photographs
had been obtained from all patients according to the guidelines of the
ethics committee of the University of Regensburg and the Declaration of
Helsinki. Classical symmetrical acanthosis nigricans in addition to the
epidermal nevi was not seen in any of the study patients, and no signs
of skeletal dysplasia or associated cancer were observed. We classified
the different subtypes of epidermal nevi according to a previous study
(32). The common keratinocytic type was subdivided into soft and hard
types according to the degree of hyperkeratosis. The characteristics of
the patients and their nevi are shown in Table 1. We also investigated an
organoid epidermal nevus type, the sebaceous nevus, because epidermal
nevi represent a heterogeneous group of lesions. Thirteen sebaceous nevi
were retrieved from the histology files of the Department of Dermatology
of the University of Regensburg (Table 2).
Sections of 10-mm thickness were microdissected manually from par-
affin-embedded epidermal nevi tissues with a needle under an inverted
microscope. We dissected the acanthotic epidermis of the nevi containing
mainly keratinocytes but also small numbers of melanocytes and dendritic
cells. In sebaceous nevi, both the hypertrophic sebaceous glands and the
acanthotic epidermis were dissected. Clinically and histologically normal
epidermis adjacent to the nevus was microdissected in 4 patients with com-
mon keratinocytic nevi to serve as a source of control DNA.
DNA isolation. DNA was isolated following standard protocols. In brief,
about 25–50 mg formalin-fixed paraffin-embedded tissue was microdis-
sected for each sample. The microdissected tissue was digested with pro-
teinase K overnight in lysis buffer, and DNA isolation was performed with
the High Pure PCR Template Preparation Kit (Roche Diagnostics) accord-
ing to the manufacturer’s protocol. The DNA of each tissue was eluted in a
volume of 200 ml elution buffer. The amount of isolated DNA ranged from
5 mg to 30 mg for each sample.
SNaPshot assay. A previously described SNaPshot multiplex assay, based
on the SNaPshot Multiplex System assay (Applied Biosystems), was used to
screen for activating FGFR3 point mutations (56). We used 2 ml template
DNA for the multiplex PCR. The SNaPshot multiplex PCR assay can detect
mutations with an input DNA amount of only 1 ng genomic DNA (56).
Three regions of interest in exons 7, 10, and 15 comprising 11 FGFR3 muta-
tions were amplified in 1 multiplex PCR, followed by extension of muta-
tion-specific primers with a labeled dideoxynucleotide. Two new antisense
primers were added to the original assay to screen for mutations S373C
(5′-T19GAGGATGCCTGCATACACAC-3′) and G382R (5′-T56GAACAG-
GAAGAAGCCCACCC-3′). Concentrations for those primers used in the
multiplex assay were 1.0 and 0.6 pmol/ml, respectively. Thus, screening
could be performed for 11 known mutations found in bladder tumors and
other noncutaneous epithelial malignancies (R248C, S249C, G372C, S373C,
Y375C, G382R, A393E, K652E, K652M, K652Q, and K652T; codons are
numbered according to the open reading frame of the FGFR3 IIIb isoform,
which is predominantly present in epithelial cells). This mutation spectrum
also covers the most frequently found FGFR3 mutations in skeletal dysplasia
syndromes. Extended primers were separated by capillary electrophoresis in
an automatic sequencer, and the presence or absence of a mutation was indi-
cated by the incorporated WT or mutant labeled dideoxynucleotide. When a
mutation is present, a second peak from the mutated nucleotide will appear
next to the WT peak in the electropherogram (see Figures 3 and 4). However,
the assay is not quantitative due to the different emission efficiencies of the
labels. Mutations were confirmed by a second independent reaction.
FGFR3 sequence analysis. In addition to the SNaPshot analysis, exon 19 of
the FGFR3 gene was directly sequenced. This exon contains the potential
mutation at stop codon 809 associated with TD I. We were able to analyze
20 epidermal nevi. The other samples failed due to limited DNA amounts.
We used the forward primer 5′-CCTGTCGGCGCCTTTGGAGCAG-3′ and
the reverse primer 5′-CAGACCAAAGCTCTGTAGCT-3′ to generate a 235
bp PCR product of exon 19 containing the stop codon 809. Sequence anal-
ysis was performed following standard protocols.
Acknowledgments
We thank our patients for their participation and Alois Eckl for
the helpful support. The excellent technical work of Monika Ker
-
scher, Kirstin van der Keur, Anne Pietryga-Krieger, Lydia Kuenzel,
and Nadine Wandtke is gratefully acknowledged.
Received for publication February 8, 2006, and accepted in revised
form May 16, 2006.
Address correspondence to: Christian Hafner, Department of Der
-
matology, University of Regensburg, Franz-Josef-Strauss-Allee 11,
93042 Regensburg, Germany. Phone: 49-941-944-9610; Fax: 49-
941-944-9611; E-mail: christian.hafner@klinik.uni-regensburg.de.
1. Rothnagel, J.A., et al. 1992. Mutations in the rod
domains of keratins 1 and 10 in epidermolytic
hyperkeratosis. Science. 257:1128–1130.
2. Paller, A.S., et al. 1994. Genetic and clinical mosa-
icism in a type of epidermal nevus.
N. Engl. J. Med.
331:1408–1415.
3. Konig, A., Happle, R., Bornholdt, D., Engel, H., and
Grzeschik, K.H. 2000. Mutations in the NSDHL
gene, encoding a 3beta-hydroxysteroid dehydro-
genase, cause CHILD syndrome.
Am. J. Med. Genet.
90:339–346.
4. L’Hote, C.G., and Knowles, M.A. 2005. Cell respons
-
es to FGFR3 signalling: growth, differentiation and
apoptosis. Exp. Cell Res. 304:417–431.
5. Scotet, E., and Houssaint, E. 1995. The choice
between alternative IIIb and IIIc exons of the FGFR-3
gene is not strictly tissue-specific. Biochim. Biophys.
Acta. 1264:238–242.
6. Yokote, H., et al. 2005. Trans-activation of EphA4
and FGF receptors mediated by direct interactions
between their cytoplasmic domains.
Proc. Natl.
Acad. Sci. U. S. A. 102:18866–18871.
7. Meyers, G.A., Orlow, S.J., Munro, I.R., Przylepa,
K.A., and Jabs, E.W. 1995. Fibroblast growth factor
receptor 3 (FGFR3) transmembrane mutation in
Crouzon syndrome with acanthosis nigricans.
Nat.
Genet. 11:462–464.
8. Muenke, M., et al. 1997. A unique point mutation
in the fibroblast growth factor receptor 3 gene
(FGFR3) defines a new craniosynostosis syndrome.
Am. J. Hum. Genet. 60:555–564.
9. Vajo, Z., Francomano, C.A., and Wilkin, D.J. 2000.
The molecular and genetic basis of fibroblast
research article
The Journal of Clinical Investigation http://www.jci.org Volume 116 Number 8 August 2006 2207
growth factor receptor 3 disorders: the achondro-
plasia family of skeletal dysplasias, Muenke cranio-
synostosis, and Crouzon syndrome with acanthosis
nigricans.
Endocr. Rev. 21:23–39.
10. Baker, K.M., Olson, D.S., Harding, C.O., and
Pauli, R.M. 1997. Long-term survival in typical
thanatophoric dysplasia type 1.
Am. J. Med. Genet.
70:427–436.
11. Webster, M.K., and Donoghue, D.J. 1997. FGFR
activation in skeletal disorders: too much of a good
thing.
Trends Genet. 13:178–182.
12. Tavormina, P.L., et al. 1995. Another mutation that
results in the substitution of an unpaired cysteine
residue in the extracellular domain of FGFR3 in
thanatophoric dysplasia type I.
Hum. Mol. Genet.
4:2175–2177.
13. Tavormina, P.L., et al. 1995. Thanatophoric dys-
plasia (types I and II) caused by distinct mutations
in fibroblast growth factor receptor 3.
Nat. Genet.
9:321–328.
14. Rousseau, F., et al. 1996. Missense FGFR3 muta
-
tions create cysteine residues in thanatophoric
dwarfism type I (TD1). Hum. Mol. Genet. 5:509–512.
15. Naski, M.C., Wang, Q., Xu, J., and Ornitz, D.M. 1996.
Graded activation of fibroblast growth factor recep
-
tor 3 by mutations causing achondroplasia and
thanatophoric dysplasia. Nat. Genet. 13:233–237.
16. d’Avis, P.Y., et al. 1998. Constitutive activation of
fibroblast growth factor receptor 3 by mutations
responsible for the lethal skeletal dysplasia thanato
-
phoric dysplasia type I. Cell Growth Differ. 9:71–78.
17. Watowich, S.S., et al. 1992. Homodimerization and
constitutive activation of the erythropoietin receptor.
Proc. Natl. Acad. Sci. U. S. A. 89:2140–2144.
18. Sorokin, A., Lemmon, M.A., Ullrich, A., and Sch-
lessinger, J. 1994. Stabilization of an active dimeric
form of the epidermal growth factor receptor by
introduction of an inter-receptor disulfide bond.
J. Biol. Chem. 269:9752–9759.
19. Adar, R., Monsonego-Ornan, E., David, P., and
Yayon, A. 2002. Differential activation of cysteine-
substitution mutants of fibroblast growth factor
receptor 3 is determined by cysteine localization.
J. Bone Miner. Res. 17:860–868.
20. Webster, M.K., and Donoghue, D.J. 1996. Constitu-
tive activation of fibroblast growth factor receptor
3 by the transmembrane domain point mutation
found in achondroplasia. EMBO
J. 15:520–527.
21. Li, E., You, M., and Hristova, K. 2006. FGFR3 dimer
stabilization due to a single amino acid pathogenic
mutation. J. Mol. Biol. 356:600–612.
22. Lievens, P.M., Roncador, A., and Liboi, E. 2006.
K644E/M FGFR3 mutants activate Erk1/2 from
the endoplasmic reticulum through FRS2alpha
and PLCgamma-independent pathways.
J. Mol. Biol.
357:783–792.
23. Monsonego-Ornan, E., Adar, R., Feferman, T.,
Segev, O., and Yayon, A. 2000. The transmem
-
brane mutation G380R in fibroblast growth fac-
tor receptor 3 uncouples ligand-mediated receptor
activation from down-regulation.
Mol. Cell. Biol.
20:516–522.
24. Lievens, P.M., Mutinelli, C., Baynes, D., and Liboi,
E. 2004. The kinase activity of fibroblast growth
factor receptor 3 with activation loop mutations
affects receptor trafficking and signaling.
J. Biol.
Chem. 279:43254–43260.
25. Bellus, G.A., et al. 2000. Distinct missense muta-
tions of the FGFR3 lys650 codon modulate receptor
kinase activation and the severity of the skeletal dys
-
plasia phenotype. Am. J. Hum. Genet. 67:1411–1421.
26. Logie, A., et al. 2005. Activating mutations of the
tyrosine kinase receptor FGFR3 are associated with
benign skin tumors in mice and humans.
Hum. Mol.
Genet. 14:1153–1160.
27. Chesi, M., et al. 1997. Frequent translocation
t(4;14)(p16.3;q32.3) in multiple myeloma is asso
-
ciated with increased expression and activating
mutations of fibroblast growth factor receptor 3.
Nat. Genet. 16:260–264.
28. van Rhijn, B.W., et al. 2002. Novel fibroblast growth
factor receptor 3 (FGFR3) mutations in bladder
cancer previously identified in non-lethal skeletal
disorders.
Eur. J. Hum. Genet. 10:819–824.
29. Cappellen, D., et al. 1999. Frequent activating
mutations of FGFR3 in human bladder and cervix
carcinomas.
Nat. Genet. 23:18–20.
30. Jang, J.H., Shin, K.H., and Park, J.G. 2001. Muta-
tions in fibroblast growth factor receptor 2 and
fibroblast growth factor receptor 3 genes associ-
ated with human gastric and colorectal cancers.
Cancer Res. 61:3541–3543.
31. Schwartz, R.A. 1994. Acanthosis nigricans. J. Am.
Acad. Dermatol. 31:1–19; quiz 20–22.
32. Su, W.P. 1982. Histopathologic varieties of epider
-
mal nevus. A study of 160 cases. Am. J. Dermatopathol.
4:161–170.
33. Happle, R. 1993. Mosaicism in human skin. Under-
standing the patterns and mechanisms. Arch. Der-
matol. 129:1460–1470.
34. Happle, R. 1987. Lethal genes surviving by mosa-
icism: a possible explanation for sporadic birth
defects involving the skin. J. Am. Acad. Dermatol.
16:899–906.
35. Curth, H.O. 1976. Unilateral epidermal naevus
resembling acanthosis nigricans.
Br. J. Dermatol.
95:433–436.
36. Kim, M.Y., Lee, J.S., Kim, H.O., and Park, Y.M. 2004.
A case of nevoid acanthosis nigricans.
Acta Derm.
Venereol. 84:234–235.
37. Hyland, V.J., et al. 2003. Somatic and germline
mosaicism for a R248C missense mutation in
FGFR3, resulting in a skeletal dysplasia distinct
from thanatophoric dysplasia.
Am. J. Med. Genet. A.
120:157–168.
38. Rongioletti, F., and Rebora, A. 1991. Epidermal nevus
with transitional cell carcinomas of the urinary
tract. J. Am. Acad. Dermatol. 25:856–858.
39. Rosenthal, D., and Fretzin, D.F. 1986. Epidermal
nevus syndrome: report of association with transi
-
tional cell carcinoma of the bladder. Pediatr. Dermatol.
3:455–458.
40. Garcia de Jalon, A., et al. 2004. Epidermal naevus
syndrome (Solomon’s syndrome) associated with
bladder cancer in a 20-year-old female.
Scand. J.
Urol. Nephrol. 38:85–87.
41. Kimura, T., et al. 2001. The incidence of thanato-
phoric dysplasia mutations in FGFR3 gene is high
-
er in low-grade or superficial bladder carcinomas.
Cancer. 92:2555–2561.
42. Van Esch, H., and Fryns, J.E. 2004. Acanthosis nigri
-
cans in a boy with achondroplasia due to the clas-
sical Gly380Arg mutation in FGFR3. Genet. Couns.
15:375–377.
43. Wilkie, A.O. 1997. Craniosynostosis: genes and
mechanisms. Hum. Mol. Genet. 6:1647–1656.
44. Passos-Bueno, M.R., et al. 1999. Clinical spectrum
of fibroblast growth factor receptor mutations.
Hum. Mutat. 14:115–125.
45. Vujevich, J.J., and Mancini, A.J. 2004. The epider-
mal nevus syndromes: multisystem disorders.
J. Am. Acad. Dermatol. 50:957–961.
46. Vidaurri-de la Cruz, H., Tamayo-Sanchez, L.,
Duran-McKinster, C., de la Luz Orozco-Covarru-
bias, M., and Ruiz-Maldonado, R. 2004. Epidermal
nevus syndromes: clinical findings in 35 patients.
Pediatr. Dermatol. 21:432–439.
47. Jebar, A.H., et al. 2005. FGFR3 and Ras gene muta
-
tions are mutually exclusive genetic events in uro-
thelial cell carcinoma. Oncogene. 24:5218–5225.
48. van Rhijn, B.W., et al. 2003. Molecular grading of
urothelial cell carcinoma with fibroblast growth
factor receptor 3 and MIB-1 is superior to patho
-
logic grade for the prediction of clinical outcome.
J. Clin. Oncol. 21:1912–1921.
49. van Rhijn, B.W., Montironi, R., Zwarthoff, E.C.,
Jobsis, A.C., and van der Kwast, T.H. 2002. Fre
-
quent FGFR3 mutations in urothelial papilloma.
J. Pathol. 198:245–251.
50. Billerey, C., et al. 2001. Frequent FGFR3 mutations
in papillary non-invasive bladder (pTa) tumors.
Am. J. Pathol. 158:1955–1959.
51. Mohammadi, M., et al. 1997. Structures of the
tyrosine kinase domain of fibroblast growth fac-
tor receptor in complex with inhibitors. Science.
276:955–960.
52. Trudel, S., et al. 2004. Inhibition of fibroblast
growth factor receptor 3 induces differentia
-
tion and apoptosis in t(4;14) myeloma. Blood.
103:3521–3528.
53. Paterson, J.L., et al. 2004. Preclinical studies of
fibroblast growth factor receptor 3 as a therapeu
-
tic target in multiple myeloma. Br. J. Haematol.
124:595–603.
54. Grand, E.K., Chase, A.J., Heath, C., Rahemtulla, A.,
and Cross, N.C. 2004. Targeting FGFR3 in mul
-
tiple myeloma: inhibition of t(4;14)-positive cells
by SU5402 and PD173074. Leukemia. 18:962–966.
55. Chen, J., et al. 2005. FGFR3 as a therapeutic target
of the small molecule inhibitor PKC412 in hema-
topoietic malignancies. Oncogene.
24:8259–8267.
56. van Oers, J.M., et al. 2005. A simple and fast
method for the simultaneous detection of nine
fibroblast growth factor receptor 3 mutations in
bladder cancer and voided urine.
Clin. Cancer Res.
11:7743–7748.
