BLo€ıc de Pontual , Delphine Trochet , Franck Bourdeaut , Sophie Thomas ,
Heather Etcheversa, Agnes Ch
Laurence Brugieresc, Arnold M
Isabelle Janoueix-Leroseyb, Jea
s Ge´ne´
s, Hoˆpit
des Ca
stave R
Keywords:
atients with NB due to a germline heterozygous PHOX2B gene muta-
tion are familial and/or syndromic. PHOX2B, at chromosome 4p12, does not lie in a com-
in children. Several lines of evidence support the involvement
of genetic factors in NB, namely, rare familial cases with ver-
tical transmission and the association of NB with other genet-
ically determined congenital malformations of neural crest
origin, such as Hirschsprung disease (HSCR; MIM 142623)
tions predispose to NB.2–5 PHOX2B is a highly conserved
homeotic transcription factor with two alanine tracts of nine
and 20 alanines, C terminal to the homeodomain. The vast
majority of mutations leading to CCHS result in an expansion
of the longer alanine tract. Interestingly, patients harbouring
E U R O P E A N J O U R N A L O F C A N C E R 4 3 ( 2 0 0 7 ) 2 3 6 6 –2 3 7 2
ava i lab le at www.sc iencedi rec t .com
: w* Corresponding author: Tel.: +33 1 44495648; fax: +33 1 44495150.1. Introduction
Neuroblastoma (NB; MIM 256700) is a tumour of the sympa-
thetic nervous system that accounts for 10% of all cancers
and/or Congenital Central Hypoventilation Syndrome (CCHS;
MIM 209880). We recently identified the paired-like homeobox
2B (PHOX2B; MIM 603851) gene as the major disease-causing
gene in CCHS1 and the first gene for which germline muta-Neuroblastoma
PHOX2B
Methylation
Neural crest
monly rearranged locus in NB. To evaluate the role of PHOX2B in sporadic, isolated NB,
we analysed 13 NB cell lines and 45 tumours for expression, mutations of coding and pro-
moter sequences, loss of heterozygosity (LOH), or aberrant hypermethylation of PHOX2B (13
cell lines and 18 tumours). We didn’t identify any mutation but LOH in about 10% of the
cases and aberrant CpG dinucleotide methylation of the 500 bp PHOX2B promoter region
in 4/31 tumours and cell lines (12.9%). Altogether, both germinal and somatic anomalies
at the PHOX2B locus are found in NB.
 2007 Elsevier Ltd. All rights reserved.combinations. Most paUnite´ de Recherches sur les Handicap
Faculte´ de Me´decine, Hoˆpitaux de Pari
bLaboratoire de Pathologie Mole´culaire
cDepartement de Pe´diatrie, Institut Gu
A R T I C L E I N F O
Article history:
Received 2 May 2007
Received in revised form 20 June
2007
Accepted 12 July 2007
Available online 31 August 20070959-8049/$ - see front matter  2007 Elsevi
doi:10.1016/j.ejca.2007.07.016
E-mail address: amiel@necker.fr (J. Amielompretc, Ve´ronique Minardc, Dominique Valteauc,
unnicha, Olivier Delattreb, Stanislas Lyonneta,
nne Amiela,*
tiques de l’Enfant INSERM U-781, et De´partement de Ge´ne´tique, Universite´ Rene´-Descartes,
al Necker-Enfants Malades, 149, rue de Se`vres, 75743 Paris Cedex 15, France
ncers, INSERM U-830, Institut Curie, Paris, France
oussy, Villejuif Cedex, France
A B S T R A C T
Neuroblastoma (NB), an embryonic tumour originating from neural crest cells, is one of the
most common solid tumours in childhood. Although NB is characterised by numerous
recurrent, large-scale chromosome rearrangements, the genes targeted by these imbal-
ances have remained elusive. We recently identified the paired-like homeobox 2B (PHOX2B,
MIM 603851) gene as disease-causing in dysautonomic disorders including Congenital Cen-
tral Hypoventilation Syndrome (CCHS), Hirschsprung disease (HSCR) and NB in variousa a b aMethylation-associated PHOX2
in human neuroblastoma
journal homepageer Ltd. All rights reserved
).gene silencing is a rare event
ww.ejconl ine.com.

N C Eeither a missense mutation in the homeodomain or a frame-
shift mutation are the ones at risk of developing NB, whether
or not they have CCHS.6 However, both germline and somatic
PHOX2B coding sequence mutations are rare events in spo-
radic, isolated NB.7,4,8
Epigenetic abnormalities, especially alterations in DNA
methylation, are involved in the development of various adult
tumours. More recent studies have indicated that epigenetic
aberrationsmayalso contribute topaediatric cancerpathogen-
esis. In neuroblastomas, several potential tumour-suppressor
genes have been found to be frequently hypermethylated and
consequently down-regulated; in particular, genes of the
tumour necrosis factor-related apoptosis-inducing ligand
(TRAIL) pathway, CASP89 and the DCR receptors (DCR1, DCR2,
DCR3 and DCR4.10,11 DNA methylation has been shown to re-
duce the binding affinity of sequence-specific transcription
factors while methylation-dependent gene silencing may also
involvealterations in chromatin structure,mediatedbymethyl
binding proteins. Chromosome distribution of the methyl-tar-
geted genes are clustered and the full pattern of methylation
may be generated early in tumourigenesis.12,13 Finally, in tu-
mours carrying a germline mutation, a second-step methyla-
tion of the DNA promoter, if present, occurs exclusively on
the wild-type allele.
The possible involvement of non-coding mutations or pro-
moter methylation of the PHOX2B locus in sporadic neuro-
blastoma has not been evaluated. In this study, we
examined the level of PHOX2B expression and its methylation
in NB. PHOX2B was silenced in a subset of 3/13 NB cell lines
and loss of expression was associated with aberrant 5 0CpG
dinucleotide methylation of the PHOX2B promoter. PHOX2B
promoter methylation was also detected in 2/18 tumuors ana-
lysed. Treatment with the demethyling agent 5-Aza-20-deoxy-
cytidine (5-Aza-dC) restored PHOX2B transcription in PHOX2B-
negative cell lines, showing that gene silencing was due to
aberrant hypermethylation. We conclude that aberrant CpG
dinucleotide methylation of PHOX2B is an alternative mecha-
nism at least as frequent as coding sequence mutations for
inactivation of PHOX2B in sporadic NB.
2. Materials and methods
2.1. Patients
13 NB cell lines and 46 sporadic neuroblastic tumours were
investigated. Patients were staged according to the Interna-
tional Neuroblastoma Staging System (INSS) and included se-
ven with stage I tumour, ten stage II, seven stage III, 15 stage
IV and seven stage IVS. Constitutional and tumour DNA were
extracted using standard protocols. We included a patient
from a two generation family with predisposition to NB due
to a germline PHOX2B gene mutation (NBAF5, Table 1).2
2.2. Sequence analysis
We screened the coding sequence of the PHOX2B gene by di-
rect DNA sequencing, as described elsewhere.1 We studied a
533 bp sequence of the promotor region that is extremely con-
E U R O P E A N J O U R N A L O F C Aserved among species (97% between human and chicken at
nucleotide level). The primer sequences for PHOX2B promotorregion are 5 0-GAAGGGGGAAAACACACAC-3 0 (forward) and 5 0-
CGTAGGCAGAGGAATTGAGG-3 0 (reverse). PCR Direct DNA
sequencing was performed using the fluorometric method
(Big Dye Terminator Cycle Sequencing kit [Applied
Biosystems]).
2.3. LOH analysis
Matched constitutional and tumour DNA samples were PCR
amplified using microsatellitte markers of the Ge´ne´thon data-
base D4S2974 and D4S1536 flanking the PHOX2B locus.
D4S1536 is 4.1Mb centromeric to PHOX2B and D4S2974 is 95
Kb telomeric to PHOX2B. Fluorescent PCR products were elec-
trophoresed and analysed on an automatic sequencer
(ABI377, Applied Biosystems, Foster City, USA).
2.4. Cell culture
Cell lines were cultured at 37 C, 5% CO2, in Dulbecco’s mod-
ified essential medium (DMEM; Invitrogen/Gibco, NY) con-
taining 10% foetal calf serum, 292 lg/ml L-glutamine, 1%
100X MEM (non-essential amino acids medium, Invitrogen/
Gibco), and 0.5% penicillin solution. Primary human neural
crest cells were cultured for 3 weeks in FGF2- and EGF-con-
taining embryonic stem cell medium with modifications
available upon request.
2.5. PHOX2B expression in NB cell lines and tumours
Total RNA was isolated from cell lines and tumours by use of
RNAzolB (Invitrogen). PHOX2B expression was first obtained
from microarray on Affymetrix HG-U133 Plus 2.0 arrays. Data
from 55 neuroblastic tumours were normalised using the GC-
RMA method. Detailed methods will be detailed elsewhere (I.
Janoueix-Lerosey and O. Delattre, manuscript in preparation).
In addition, RT-PCR detection of PHOX2B mRNA was per-
formed in order to validate these data in 13 NB cell lines
and eight tumours (Fig. 1a). First strand cDNA synthesis was
performed on 2 lg of total RNA in a volume of 20 ll by use
of [RNA kit, Applied Biosystems] and oligo(dT) primers. The
specific primers used for mRNA amplification were designed
within exons 2 and 3, as follows: 5 0-GAGGCGCGAGTCCA
GGTGTGGTTC-3 0 (forward) and 5 0-CGACAATAGCCTTGGGCC-
TACCCG-3 0 (reverse). Expression analysis was performed in
a 25-ll PCR reaction containing 1 ll of cDNA, 1 ll dNTPs (2.5
mmol/l each), 0.5 ll of each specific primers (150 ng/ll), and
0.2 ll Taq polymerase (5 U/ll; Invitrogen). PCR conditions were
standard with an annealing temperature of 69 C. PCR prod-
ucts were loaded on a 2% agarose gel and directly visualised
under UV illumination.
2.6. Analysis of PHOX2B gene hypermethylation by
bisulfite DNA sequencing
Genomic DNAwas isolated from cell lines and primary tissues
by standard procedures. Bisulfite treatment and DNA
sequencing were performed as described.14 The primer se-
quences for PHOX2B are 5 0-AAATGTAATTTATAAGATGTTT
R 4 3 ( 2 0 0 7 ) 2 3 6 6 –2 3 7 2 2367TTTTTTTG-3 0 (forward) and 5 0-CACACTACTTAAAAATAATAA
AAATTAAAT-30 (reverse). PCR conditions were standard with

um
R. M
lific
L
N CTable 1 – Presentation of a series of 13 NB cell lines and 18 t
Affymetrix HG U133 Plus 2.0 arrays (Affy Exp) and/or RT-PC
well as prognostic factors (stage, loss of 1p36, N-MYC amp
Affy Exp RT-PCR Mutation
Cell lines CLB-BAR 3679 + –
CLB-GA 6956 + –
CLB-GE 5725 + –
CLB-MA 7820 + –
CLB-PE 387 + –
GIMEN 6 – –
IMR32 8839 + –
KCNR 5181 + –
SJNB12 11010 + –
SKNAS 2169 + –
SKNSH 23 – 721–740del20nt
SKNBE 6 – –
TR14 2534 + –
Tumours NBAF15 6775 ND –
NBAF21 136 – –
NBAF26 8259 + –
NBAF38 23 – –
NBAF39 6293 ND –
2368 E U R O P E A N J O U R N A L O F C Aan annealing temperature of 58 C. PCR products were loaded
on a 2% agarose gel and directly visualised under UV illumi-
nation. PCR products were cloned in a TA cloning vector (Pro-
mega, Madison, WI) and ten individual clones were
sequenced for each sample.
2.7. Treatment of NB cells with 5-Aza-dC
Cells were seeded, allowed to attach over a 24 h period, and
treated for 72 h with the demethyling reagent 5-Aza-dC (Sig-
ma) at a final concentration of 1 lM. After the treatment per-
iod the medium was removed and RNAs were extracted.
3. Results
3.1. Mutation and LOH at the PHOX2B locus in
neuroblastoma
As previously reported, a heterozygous frameshift mutation
was identified in the SK-N-SH cell line (721-740del20nt,3). No
mutation, either somatic or germinal, of the PHOX2B coding
sequence and promoter conserved region were identified in
the other 12 NB cell lines and the 45 individuals with sporadic
NB. A heterozygous missense mutation (R100L) was identified
in patient NBAF5 originating from a family predisposed to NB
NBAF40 7462 ND –
NBAF42 7557 ND –
NBAF43 3415 + –
NBAF45 7089 + –
NBAF46 1394 + –
NBAF5 1053 ND 299G > T(R100L)
NBAF50 7811 ND –
NBAF54 5166 ND –
NBAF56 12059 ND –
NBAF59 10706 + –
NBAF61 10684 ND –
NBAF64 5349 ND –
NB10 ND – –ours for which PHOX2B expression data were analysed on
utations, LOH and methylation analyses are presented as
ation; A, Amplified; NA, Non Amplified)
PHOX2B
OH Methylation Stage INSS Loss of 1p36 N-MYC
– – 4 + A
– – 4 + NA
– – 4 + A
– – 4 + A
– – 2 – A
– + 4 + NA
– – 2 + A
– – 4 + A
– – 3A + NA
– – 4 + NA
– – 4 – NA
– + 4 + A
– – 3 + A
– – 2B – NA
– + 3 – NA
– – 4S – NA
+ – 2A – NA
– – 2B – NA
E R 4 3 ( 2 0 0 7 ) 2 3 6 6 –2 3 7 2over three generations. No second molecular event could be
identified in tumoural DNA (Table 1). We detected three
known synonymous base substitutions, 552C > T(S184S),
750G > A (A250A) and 870C > A (P290P). In one patient, the
P290P variant was heterozygous in constitutional DNA and
homozygous in tumoural DNA. LOH was confirmed in this
patient and identified in four other tumours (5/46 tumours,
10.8 %) with fluorescent microsatellite markers.
3.2. PHOX2B expression in NB cell lines and tumours
PHOX2B expression levels were obtained from Affymetrix HG
U133 Plus 2.0 arrays. High level of PHOX2B expression was
found in 10/13 cell lines and 15/17 tumourswhereas no expre-
sion could be detected in three cell lines and two tumours
(Table 1). PHOX2B expression was examined by RT-PCR in 13
NB cell lines and eight tumours and two neural crest-derived
tissues: adrenal gland medulla (MSR) and human trunk-level
neural crest cells. These primary, non-transformed cells were
derived from a human embryo at 28 days of development
with a normal karyotype. Loss of expression was confirmed
by the absence of PHOX2B mRNA transcripts in 3/13 cell lines
(SK-N-SH, SK-N-BE and GIMEN) whereas PHOX2B cDNAs were
visualised in all other cell lines and non-cancerous adrenal
(MSR) and neural crest cells (NC) (Fig. 1a). PHOX2B was
+ – 4S – NA
– – 4 – NA
– – 2B
– – 1 – NA
– – 2A – NA
– – 1 – NA
– – 2B – NA
+ – 4S – NA
– – 4S + NA
+ – 2A – NA
+ – 4S – NA
– – 4 + NA
– + 2B – NA

CL
-
B-
G
A
CL
-
B-
G
A
CL
-
B-
G
A
N C ESK
-
N-
SH
SK
-
N-
B
E
M
SR
N
C CL
-
B
-G
E
TR
14
CL
-
B
-M
A
CL
-
B
-B
AR
SK
-
N-
SH
SK
-
N-
B
E
M
SR
N
C CL
-
B
-G
E
TR
14
CL
-
B
-M
A
CL
-
B
-B
AR
SK
-
N-
SH
SK
-
N-
B
E
M
SR
N
C CL
-
B
-G
E
TR
14
CL
-
B
-M
A
CL
-
B
-B
ARa
E U R O P E A N J O U R N A L O F C Asilenced in three tumours: one ganglioneuroma (NBAF21), one
ganglioneuroblastoma (NBAF38) and one NB (NB10) (Fig. 1a).
3.3. Methylation and silencing of PHOX2B in NB cell lines
and tumours
Promoter-associated CpG islands of PHOX2B were analysed
by methyl sequencing in a panel of 13 cell lines, 18 tumours
1 2 3 4 5 6 7 8
Forward primer
SKNSH
CLBGE
CLBBAR
CLBMA
TR14
SKNBE
CLBGA
GIMEN
KCNR
CLBPE
SKNAS
SJNB12
IMR32
NB10
NBAF21
NBAF38
NBAF15
NBAF26
NBAF38
NBAF39
NBAF40
NBAF45
NBAF5
NBAF54
NBAF59
NBAF64
NBAF42
NBAF43
NBAF46
NBAF50
NBAF56
Ce
lll
in
es
Tu
m
o
rs
a
m
pl
es
Ce
lll
in
es
Tu
m
o
rs
a
m
pl
es
b
N
B
AF
59
N
B
AF
43
N
B
AF
45
N
B
AF
38
N
B
AF
46
N
B
AF
21
N
B
AF
59
N
B
AF
43
N
B
AF
45
N
B
AF
38
N
B
AF
46
N
B
AF
21
N
B
AF
59
N
B
AF
43
N
B
AF
45
N
B
AF
38
N
B
AF
46
N
B
AF
21
Fig. 1 – (a) PHOX2B mRNA in NB cell lines. PHOX2B mRNA is det
(MSR) and non-transformed human neural crest cells (NC). No e
SKNBE and GIMEN. (b) Bisulfite sequencing analysis of the meth
microgram of tumour DNA was denatured by sodium hydroxide
unmethylated cytosines to uracil. Bisulfite treated DNA was am
promotor is 100% methylated in SKNBE and GIMEN but not in o
two primary tumours NBAF21 and NB10. All ten clones showed
s non-methylated cytosine.G
IM
EN
H
2OSK
-
N-
AS
SJ
-
N
-
B
12
IM
R
3
2
K
C-
N
-
R
CL
-
B
-P
E
PHOX2B
G
IM
EN
H
2OSK
-
N-
AS
SJ
-
N
-
B
12
IM
R
3
2
K
C-
N
-
R
CL
-
B
-P
E
G
IM
EN
H
2OSK
-
N-
AS
SJ
-
N
-
B
12
IM
R
3
2
K
C-
N
-
R
CL
-
B
-P
E
R 4 3 ( 2 0 0 7 ) 2 3 6 6 –2 3 7 2 2369and two neural crest-derived tissues: adrenal gland medulla
(MSR) and human trunk-level neural crest cells. All 12 CpG
dinucleotides were 100% methylated (10/10 clones) in 2/13
cell lines (SK-N-BE and GIMEN) and in 2/18 tumours (NBAF21
and NB10) (Fig. 1b). As sequencing of the PCR products
showed that cytosines outside the CpG sites were converted
to thymine, an incomplete bisulfite conversion could be ru-
led out.
9 10 11 12
Reverse primer
Actin
PHOX2B
N
B
AF
26
N
B
10
N
B
AF
26
N
B
10
N
B
AF
26
N
B
10
Actin
ected in 10/13 cell lines as in human adrenal gland medulla
xpression could be observed in three NB cell lines; SKNSH,
ylation status of the 12 CpG from the start site. One
and modified by sodium bisulfite treatment, which converts
plified using specific primers (added in methods). PHOX2B
ther cell lines. All the 12 CpG dinucleotide are methylated in
identical methylation in each sample. •methylated cytosine

N C3.4. Restoration of PHOX2B expression in negative cell
lines by 5-Aza-dC
5-Aza-dC, a methyltransferase inhibitor, was used to investi-
gate whether PHOX2B expression could be restored cell lines
for which PHOX2B was not expressed. PHOX2B mRNA expres-
sion levels, analysed by RT-PCR, were restored in SK-N-BE and
GIMEN cell lines, while no re-expression was observed for SK-
N-SH.
4. Discussion
This study of a series of 13 cell lines and 46 neuroblastic tu-
mours aimed to determine the role of PHOX2B in sporadic
NB. PHOX2B, not known so far for being either a tumour-sup-
pressor gene or a proto-oncogene, is the first predisposing
gene identified in NB. Most mutations occur in familial and
syndromic NB cases. A heterozygous germline mutation was
found in about 20% of published pedigrees and reached 50%
when HSCR is associated with NB either in the index case
or relatives.2,5,7,4 However, germline PHOX2B mutations are
rare in sporadic isolated NB (1/215 cases in the series reported
by Van Limpt and 2/86 in the series reported by McConville).3,8
Somatic PHOX2B mutations have also been described in NB.3
In the series of 45 sporadic NB cases we report, no PHOX2B
gene mutations were identified either in the coding sequence
or the promoter region. LOH at the PHOX2B locus could be
detected in 5/46 cases. This ratio of about 10% has to be bal-
anced with the known high rate of chromosomal rearrange-
ment in NB. It is worth noting that LOH of the short arm of
chromosome 4 is found in roughly 20% of NB, but PHOX2B is
centromeric to the smallest region of overlap (SRO).15,16
We subsequently asked the question of epigenetic events
at the PHOX2B locus and identified clonal aberrant CpG island
methylation of the promoter in 2/13 NB cell lines and 2/18 tu-
mours (ganglioneuroma in one case and NB stage II in one
case). The demethylating drug 5-Aza-dC re-induced PHOX2B
expression in NB cell lines, suggesting that PHOX2B methyla-
tion correlates with gene silencing. Moreover, it is well estab-
lished that DNA demethylating agents as 5-Aza-dC induce
adrenergic differentiation in NB cell lines17 and have an anti-
proliferative effect in mouse NB models.18 Hypermethylation
and downregulation of potential tumour-suppressor genes
such as genes involved in cell-cycle control or apoptosis is of-
ten associated with a poor outcome in NB.9,19,20
We first considered PHOX2B as a potential tumour-sup-
pressor gene since, i) at least some mutations identified in
syndromic or familial NB cases (i.e. missense mutations of
the homeodomain) are likely loss-of-function mutations, ii)
a ‘second hit’ model has been proposed in NB21 and, iii) some
constitutional LOH at chromosome 4p encompass the
PHOX2B locus.4 However, in patients with germinal PHOX2B
mutation, neither a second mutation nor LOH or aberrant pro-
moter hypermethylation in the tumours have been identified
as reported earlier.2
PHOX2B has been shown to promote differentiation by
controlling G1-S transition during cell cycle of sympathetic
neuroblast precursors22 and is an essential regulator of nor-
2370 E U R O P E A N J O U R N A L O F C Amal autonomic nervous system development.23 Mice with a
homozygous inactivation of Phox2b fail in proper differentia-tion of the sympathetic nervous system. A gain-of-function
or a dominant negative effect of PHOX2B frameshift muta-
tions is not, however, ruled out. When tested, mutant tran-
scripts were present and stable.2 While mutant proteins
localised to the nucleus, we observed some ability to bind
DNA for two of the three PHOX2B frameshift mutations tested
in vitro, although transactivation of the dopamine beta-
hydroxylase promoter was always severely impaired.24 More-
over, Bachetti and coworkers reported an increased transacti-
vation of the PHOX2A promoter for proteins resulting from
frameshift mutations when compared to the wild type pro-
tein.25 PHOX2B is expressed not only in neural crest cells but
also in mature sympathetic tissue of adrenal gland (Fig. 1a
and1). Methylation-associated repression of PHOX2B could re-
sult in a differentiation block of sympathetic neuroblasts.
Interestingly, NB10 patient had HSCR. We found aberrant
homozygous PHOX2B promoter methylation on the tumour
sample DNA while neither a mutation nor a deletion could
be identified on either germinal or tumour DNA. Interestingly,
PHOX2B was not methylated on lymphocyte DNA. We could
speculate that PHOX2B methylation occurred during embry-
onic development and is responsible for both HSCR and NB
development. As malignant transformation of cells could
happen at different stages of tissue maturation, aberrant
methylation may contribute to the diversity that character-
ises NB and other genes of the RET-PHOX2B pathway could
be implicated. Interestingly, aberrations in the p53/MDM2/
p14 (ARF) pathway have recently been described in the two
NB cell lines where PHOX2B was found methylated26; a p53
mutation in SKNBE(2)C and aberrant p14 (ARF) methylation
in GIMEN.
Promoter methylation is not only implicated in silencing
tumour-suppressor genes but also in regulation of develop-
mental pathways during embryogenesis.27 Abnormal methyl-
ation in NB could result from abnormalities in this process in
a self-renewing multipotent stem cell becoming the malig-
nant progenitor of this neural crest cancer. Within a single tu-
mour, cell phenotypes are characteristic of embryonic
structures, particulary neuroblasts, Schwann cells and mela-
nocytes. Cellular heterogeneity and maturation stage corre-
late with clinical stage and prognosis of the disease.
However, down-regulation of PHOX2B was not always associ-
ated with hypermethylation. In SK-N-SH, the mutation is het-
erozygous and no PHOX2B expression could be detected. This
result is discordant with the one reported by van Limpt and
coworkers who observed the expression of both wild-type
and mutant cDNAs in this cell line. However, they observed
a PHOX2B gene silencing in the SHEP cell line, a stable sub-
clone of SK-N-SH and also carrying the frameshift mutation.
One could speculate on aberrant hypermethylation outside
of the promoter region studied. However, it is worth noting
that no expression of H-ASH1 could be detected either in
the SK-N-SH cell line or in NBAF38 tumour (data not shown).
This could also contribute to PHOX2B down-regulation. Final-
ly, one can speculate on the effects of other mechanisms such
as micro RNAs Fig. 2.
Genetic heterogeneity in predisposition to neuroblastic tu-
mours is likely. The high reocurence of loss of heterozygosity
E R 4 3 ( 2 0 0 7 ) 2 3 6 6 –2 3 7 2for chromosomes 1p36 and 11q23 supported the existence of
putative tumour-suppressor genes within these regions.

G i
(+)
r b
N C EHowever, these loci do not segregate with NB in most familial
cases. On the other hand, linkage analysis has focused atten-
tion on 16p12-13 and 4p16.28,16 Several studies have clearly
demonstrated that PHOX2B is a major predisposing gene in
syndromic NB cases (i.e. associated with other autonomic
dysfunction).6,3,29 Aberrant methylation of the PHOX2B pro-
moter seems to be an alternative mechanism as frequent as
mutation in sporadic NB cases and argues that loss-of-func-
tion by haploinsufficiency is the NB-predisposing
mechanism.
Fig. 2 – (a) RT-PCR: Demethylation of the PHOX2B promoter Cp
SKNBE and GIMEN. SKNBE, GIMEN and SKNSH were treated
three days. (b) Methyl sequencing results of PHOX2B promotoE U R O P E A N J O U R N A L O F C AConflict of interest statement
None declared.
Acknowledgement
We thank Fe´rechte´ Encha Razavi for providing adrenal
tissues.
R E F E R E N C E S
1. Amiel J, Laudier B, Attie-Bitach T, et al. Polyalanine expansion
and frameshift mutations of the paired-like homeobox gene
PHOX2B in congenital central hypoventilation syndrome. Nat
Genet 2003;33(4):459–61.
2. Trochet D, Bourdeaut F, Janoueix-Lerosey I, et al. Germline
mutations of the paired-like homeobox 2B (PHOX2B) gene in
neuroblastoma. Am J Hum Genet 2004;74(4):761–4.
3. van Limpt V, Schramm A, van Lakeman A, et al. The Phox2B
homeobox gene is mutated in sporadic neuroblastomas.
Oncogene 2004;23(57):9280–8.4. Perri P, Bachetti T, Longo L, et al. PHOX2B mutations and
genetic predisposition to neuroblastoma. Oncogene
2005;24(18):3050–3.
5. Mosse YP, Laudenslager M, Khazi D, et al. Germline PHOX2B
mutation in hereditary neuroblastoma. Am J Hum Genet
2004;75(4):727–30.
6. Trochet D, O’Brien LM, Gozal D, et al. PHOX2B genotype
allows for prediction of tumor risk in congenital central
hypoventilation syndrome. Am J Hum Genet 2005;76(3):421–6.
7. van Limpt V, Chan A, Schramm A, et al. Phox2B mutations
and the Delta-Notch pathway in neuroblastoma. Cancer Lett
slands restores expression in PHOX2B-negative NB cell lines,
or not treated (–) with 1 lmol/l 5 0-aza-20-deoxycytidine for
efore (1) and after treatment (2) by 5-azacytidine.
R 4 3 ( 2 0 0 7 ) 2 3 6 6 –2 3 7 2 23712005;228(1–2):59–63.
8. McConville C, Reid S, Baskcomb L, et al. PHOX2B analysis in
non-syndromic neuroblastoma cases shows novel mutations
and genotype-phenotype associations. Am J Med Genet A
2006;140(12):1297–301.
9. Teitz T, Wei T, Valentine MB, et al. Caspase 8 is deleted or
silenced preferentially in childhood neuroblastomas with
amplification of MYCN. Nat Med 2000;6(5):529–35.
10. Abe M, Ohira M, Kaneda A, et al. CpG island methylator
phenotype is a strong determinant of poor prognosis in
neuroblastomas. Cancer Res 2005;65(3):828–34.
11. Banelli B, Di Vinci A, Gelvi I, et al. DNA methylation in
neuroblastic tumors. Cancer Lett 2005;228(1–2):37–41.
12. Esteller M, Fraga MF, Guo M, et al. DNA methylation patterns
in hereditary human cancers mimic sporadic tumorigenesis.
Hum Mol Genet 2001;10(26):3001–7.
13. Ushijima T, Okochi-Takada E. Aberrant methylations in
cancer cells: where do they come from? Cancer Sci
2005;96(4):206–11.
14. Herman JG, Graff JR, Myohanen S, et al. Methylation-specific
PCR: a novel PCR assay for methylation status of CpG islands.
Proc Natl Acad Sci U S A 1996;93(18):9821–6.
15. Caron H, van Sluis P, de Kraker J, et al. Allelic loss of
chromosome 1p as a predictor of unfavorable outcome in
patients with neuroblastoma. N Engl J Med 1996;334(4):225–30.
16. Perri P, Longo L, Cusano R, et al. Weak linkage at 4p16 to
predisposition for human neuroblastoma. Oncogene
2002;21(54):8356–60.
17. Okuse K, Mizuno N, Matsuoka I, et al. Induction of cholinergic
and adrenergic differentiation in N-18 cells by differentiation

agents and DNA demethylating agents. Brain Res 1993;626
(1-2):225–33.
18. Bartolucci S, Estenoz M, Longo A, et al. 5-Aza-2’-
deoxycytidine as inducer of differentiation and growth
inhibition in mouse neuroblastoma cells. Cell Differ Dev
1989;27(1):47–55.
19. van Noesel MM, van Bezouw S, Voute PA, et al. Clustering of
hypermethylated genes in neuroblastoma. Genes Chromosomes
Cancer 2003;38(3):226–33.
20. Yang Q, Zage P, Kagan D, et al. Association of epigenetic
inactivation of RASSF1A with poor outcome in human
neuroblastoma. Clin Cancer Res 2004;10(24):8493–500.
21. Knudson Jr AG, Meadows AT. Developmental genetics of
neuroblastoma. J Natl Cancer Inst 1976;57(3):675–82.
22. Dubreuil V, Hirsch MR, Pattyn A, et al. The Phox2b
transcription factor coordinately regulates neuronal cell cycle
exit and identity. Development 2000;127(23):5191–201.
23. Pattyn A, Morin X, Cremer H, et al. The homeobox gene
Phox2b is essential for the development of autonomic neural
crest derivatives. Nature 1999;399(6734):366–70.
24. Trochet D, Hong SJ, Lim JK, et al. Molecular consequences of
PHOX2B missense, frameshift and alanine expansion
mutations leading to autonomic dysfunction. Hum Mol Genet
2005;14(23):3697–708.
25. Bachetti T, Borghini S, Ravazzolo R, et al. An in vitro approach
to test the possible role of candidate factors in the
transcriptional regulation of the RET proto-oncogene. Gene
Expr 2005;12(3):137–49.
26. Carr J, Bell E, Pearson AD, et al. Increased frequency of
aberrations in the p53/MDM2/p14(ARF) pathway in
neuroblastoma cell lines established at relapse. Cancer Res
2006;66(4):2138–45.
27. Theise ND, Krause DS. Toward a new paradigm of cell
plasticity. Leukemia 2002;16(4):542–8.
28. Maris JM, Weiss MJ, Mosse Y, et al. Evidence for a hereditary
neuroblastoma predisposition locus at chromosome
16p12–13. Cancer Res 2002;62(22):6651–8.
29. Maris JM. The biologic basis for neuroblastoma heterogeneity
and risk stratification. Curr Opin Pediatr 2005;17(1):
7–13.
2372 E U R O P E A N J O U R N A L O F C A N C E R 4 3 ( 2 0 0 7 ) 2 3 6 6 –2 3 7 2