Nature template - PC Word 97 file · Web viewNature template - PC Word 97
Transcript of Nature template - PC Word 97 file · Web viewNature template - PC Word 97
Discovery of MYCN-amplified retinoblastoma with functional
retinoblastoma protein in very young children
Diane E Rushlow, Berber M Mol,* Jennifer Y Kennett,* Stephanie Yee,* Sanja Pajovic,
Brigitte L Thériault, Nadia L Prigoda-Lee, Clarellen Spencer, Helen Dimaras, Timothy W
Corson, Renee Pang, Christine Massey, Roseline Godbout, Zhe Jiang, Eldad
Zacksenhaus, Katherine Paton, Annette C Moll, Claude Houdayer, Anthony Raizis, William
Halliday, Wan L Lam, Paul C Boutros, Dietmar Lohmann, Josephine C Dorsman, Brenda L
Gallie
*These authors share second authorship
Retinoblastoma Solutions and the Toronto Western Hospital Research Institute, Campbell
Family Cancer Research Institute, Princess Margaret Cancer Centre, University Health
Network; Informatics and Biocomputing Platform, Ontario Institute for Cancer Research;
Departments of Hematology/Oncology, Ophthalmology and Visual Science and of
Pathology, Hospital for Sick Children; and Departments of Molecular Genetics,
Ophthalmology, Medical Biophysics, Pathobiology and Lab Medicine, University of Toronto,
Toronto, ON, Canada (D E Rushlow, BSc, S Yee, MSc, S Pajovic, PhD, B L Thériault, PhD, N L
Prigoda-Lee, MSc, C Spencer, BSc, Z Jiang, BSc, E Zacksenhaus, PhD, R Pang, MA, C Massey,
MSc, H Dimaras, PhD, P C Boutros, PhD, W Halliday, MD, Prof B L Gallie, MD); Departments of
Clinical Genetics, Ophthalmology and Pediatric Oncology/Hematology, VU University
Medical Center Amsterdam, Amsterdam, The Netherlands (B M Mol, MSc, A C Moll, MD, Prof J
C Dorsman, PhD); British Columbia Cancer Research Centre and Departments of
Ophthalmology and Pathology & Laboratory Medicine, University of British Columbia,
Vancouver, BC, Canada (J Y Kennett, MSc, K Paton, MD, Prof W L Lam, PhD); Eugene and
Marilyn Glick Eye Institute, Departments of Ophthalmology, Biochemistry and Molecular
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
MYCN amplified retinoblastoma with normal pRb
Biology, Indiana University School of Medicine Indianapolis, Indiana, USA (Timothy W
Corson, PhD); University of Alberta, Cross Cancer Institute, Edmonton, Alberta (Prof Roseline
Godbout, PhD); Service de Génétique Oncologique, Institut Curie and Université Paris
Descartes Paris, France (C Houdayer, PhD); Department of Molecular Pathology, Canterbury
Health Laboratories Christchurch, New Zealand (A Raizis, PhD); and Institut für
Humangenetik, Universitätsklinikum, Essen, Germany (Prof D Lohmann, MD)
Correspondence to Dr Brenda L. Gallie, Campbell Family Cancer Research Institute, Princess Margaret
Cancer Centre, University Health Network, Rm 8-415, 610 University Ave, Toronto, ON, M5G 1M9, Canada,
2
1
2
3
4
5
6
8
9
10
MYCN amplified retinoblastoma with normal pRb
Word count 300/300
Summary
Background Retinoblastoma is the childhood retinal cancer that defined tumour suppressor genes.
By analysing age of diagnosis, Knudson proposed that two “hits” initiate retinoblastoma, later
attributed to mutation of both alleles of the retinoblastoma suppressor gene, RB1, in tumours.
Persons with hereditary retinoblastoma carry a heterozygous constitutional RB1 mutation; one
additional hit initiates retinoblastoma and/or other cancers. Non-hereditary retinoblastoma arises
when both RB1 alleles are damaged in developing retina (RB1-/-).
Methods We analysed clinical data, genomic copy-number, histology, RB1 gene expression and
protein function, and retinal gene expression, in unilateral non-familial retinoblastomas with no
evidence of RB1 mutations (RB1+/+), compared to RB1-/- tumours.
Findings We found that 2·7% (29/1068) of unilateral non-familial retinoblastomas had no evidence
of RB1 mutations (RB1+/+). Surprisingly, half of the RB1+/+ tumours had high-level MYCN oncogene
amplification (28 to 121 copies), while no RB1-/- primary tumours showed MYCN amplification
(p<0·0001). RB1+/+MYCNA cell lines expressed functional RB1 protein. RB1+/+MYCNA tumours had
fewer overall genomic copy-number changes and distinct, aggressive histology. MYCN
amplification was the sole copy-number change detected in one RB1+/+MYCNA retinoblastoma.
Median age at diagnosis of RB1+/+MYCNA tumours was 4·5 months, compared to 24 months for
non-familial unilateral RB1-/- retinoblastoma. Children diagnosed with unilateral non-familial large
retinoblastoma at six months of age or less have >18% chance of RB1+/+MYCNA tumour.
Interpretation Amplification of the MYCN oncogene may initiate RB1+/+MYCNA retinoblastoma
despite normal RB1 genes. Although they are young, it is unlikely that these children carry a
hereditary risk for other retinoblastoma or cancers. Since these tumours may rapidly become
3
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
MYCN amplified retinoblastoma with normal pRb
extraocular, removal of the eye of young children with potential large RB1+/+MYCNA retinoblastoma
is good care.
Funding NCI-NIH; CIHR; Canadian Retinoblastoma Society; Hyland Foundation; Ontario
Ministry of Health and Long Term Care; Toronto Netralya and Doctors Lions Clubs; and
Foundations Avanti-STR and KiKa.
Word count 2998/3000
Background
Retinoblastoma set the paradigm for tumour suppressor genes, with Knudson’s classic hypothesis
predicting that two rate-limiting hits initiate this childhood eye cancer.1 The two hits were later
attributed to the retinoblastoma gene (RB1).2 The 40% of children who have bilateral
retinoblastoma carry a heterozygous, constitutional RB1 mutation and are predisposed to
retinoblastoma; one additional hit, damages the second RB1 allele and initiates retinoblastoma,
and/or other cancers later in life. In 70% of tumours, the second hit involves loss of the normal
allele with duplication of the mutant allele (loss of heterozygosity, LOH). The 60% of children with
unilateral non-familial retinoblastoma usually have both RB1 alleles damaged only in developing
retina, but 15% of them carry a heritable, constitutional RB1 mutation. Accepted dogma is that
disruption of both RB1 alleles (RB1-/-) initiates all retinoblastoma.2-4
The heterozygous mutant RB1 allele is identifiable in blood of 95% of bilaterally affected
persons. RB1 mutations or promoter methylation, are detected on both alleles (RB1-/-) in 95% of
unilateral retinoblastomas.5, 6 The possibility that some unilateral retinoblastomas with no detectable
RB1 mutations occur by an independent mechanism has not been previously explored.
We report the first identification of unilateral retinoblastomas with normal RB1 alleles and high-
level MYCN gene amplification (MYCNA). These unilateral RB1+/+MYCNA retinoblastomas are
characterized by distinct histology, few of the genomic copy-number changes characteristic of
4
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
MYCN amplified retinoblastoma with normal pRb
retinoblastoma, and very early age of diagnosis. This newly recognized sub-type of retinoblastoma
has immediate diagnostic, genetic counselling, and therapeutic implications.
5
1
2
3
MYCN amplified retinoblastoma with normal pRb
Methods (details in webappendix)
Clinical samples
Tissues and clinical data were provided for identification of RB1 mutations in clinical care of
children and their families. Research Ethics Board approvals to use de-identified data and tissues
after clinical analyses, are on file at each participating site. Although not required, the Toronto and
Essen patients provided additional informed consent.
Mutation analyses
Standard of care analyses that identify 95% of RB1 (Gen bank accession #L11910) mutant alleles5-7
were applied at each site, including DNA sequencing, quantitative multiplex PCR (QM-PCR) or
Multiplex Ligation-Dependent Probe Amplification (MLPA), and RB1 promoter methylation
testing. Intragenic and closely-linked RB1 microsatellite markers were used to determine zygosity
of the RB1+/+ tumours.
Genome copy-number analyses
Genomic copy-numbers of five genes were determined by QM-PCR (Toronto) or MLPA and single
nucleotide polymorphism (SNP) analyses (Amsterdam). Sub-megabase resolution array
comparative genomic hybridization (aCGH) or SNP array were used to assess overall genomic
copy-numbers and normal cell contamination.
Protein expression studies
Paraffin-embedded sections of retinoblastomas and adjacent normal retinas were stained for full-
length pRb protein (antibodies targeting N- and C-terminal pRb) and N-Myc.8 Western blots were
performed on RB1+/+MYCNA, RB1-/- and control cell lines. Function of pRB was assessed by
6
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
MYCN amplified retinoblastoma with normal pRb
immunoblotting with phospho-specific pRb antibody and co-immunoprecipitation of pRb and its
major effector molecule, E2F1.
RNA gene expression studies
Expression of RB1, MYCN, and genes reflecting the proliferative status and retinal derivation of
tumours were assessed using End-Point Reverse Transcriptase PCR (RT-PCR) and/or Quantitative
Real-Time PCR.
Age of Diagnosis analysis
Ages at diagnosis vs. proportion not yet diagnosed were analysed to assess minimal number of
events for tumour initiation. Likelihood of children having RB1+/+MYCNA tumours at different ages
of diagnosis was estimated.
Role of the funding sources
The sponsors of the study had no roles in study design, data collection, analysis, or interpretation, or
writing of the manuscript. No author was paid to write this article. All authors had full access to all
data in the study; the corresponding author (BLG) had final responsibility for the decision to submit
for publication. NCI-NIH grant 5R01CA118830-05 supported the early discovery at the Canadian
site. Canadian Institutes for Health Research grants (MOP-86731, MOP-77903, MOP-110949)
supported the aCGH studies, and MOP-77710 supported the pRb phosphorylation studies. The
Canadian Retinoblastoma Society, Hyland Foundation and Toronto Netralya and Doctors Lions
Clubs provided critical funding for additional experiments. The Ontario Ministry of Health and
Long Term Care provided infrastructure. Solutions by Sequence supported the overall project, data
analysis and manuscript preparation. The Dutch study was made possible by Avanti-STR and KiKa,
while VUmc provided infrastructure.
7
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
MYCN amplified retinoblastoma with normal pRb
Results
The Toronto lab identifies both tumour mutations (or promoter methylation) in >95% of tumours
from unilateral probands with no known family history. In 3-4% of tumours only one RB1 mutation
is identified, and in 1·6%, no RB1 mutation is found. Low-level mosaicism may account for the 5%
"no mutation found" blood samples from bilaterally affected patients.5 However, undetected RB1
mutations in fully tested tumours (RB1+/+) may include translocations, deep intronic mutations, or
mutations in unknown RB1 regulatory regions.
In clinical work in Toronto, we found seven unilateral retinoblastoma tumours with no RB1
mutations, no promoter methylation, and no RB1 LOH (RB1+/+). We used QM-PCR to measure, in
RB1+/+ tumours, copy-numbers of genes at 6p, 1q, 16q and 2p that are characteristically gained or
lost in RB1-/- retinoblastoma tumours.9 To our surprise, 5/7 RB1+/+ tumours showed dramatic MYCN
oncogene amplification (MYCNA). To validate this observation, we collaborated with RB1 clinical
laboratories in Germany, France, and New Zealand, to study similar RB1+/+ tumours (table 1). After
initial studies, we learned that the Amsterdam group had independently discovered three
RB1+/+MYCNA tumours. We report our combined data after statistical analysis showed that
frequencies were not different (p = 0·08) among the five labs. Retinoblastoma T101, predicted
clinically and pathologically to be an RB1+/+MYCNA tumour (figure 4C), had 40 copies of MYCN,
and is included in some analyses.
Since our standard sensitivity to find the RB1 mutation in blood of bilaterally affected persons is
>95%,5-7 the probability of finding no RB1 mutations in a tumour with no RB1 LOH is equivalent to
the probability of missing two independent RB1 mutations in one sample (0·05 x 0·05) or 0·25%.
By combining our data on 1068 unilateral non-familial tumours, we identified 29 RB1+/+tumours
(2·7%), 10-fold more than expected (p = 6 x 10-45) (table 1). This suggested that some RB1+/+
tumours might originate by a mechanism other than Knudson’s two RB1 mutations.
8
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
MYCN amplified retinoblastoma with normal pRb
To characterize copy numbers of genes commonly gained or lost in retinoblastomas,9 we used
QM-PCR (Toronto) or MLPA/SNP (Amsterdam) analyses. MYCN copy-number was elevated in
27/30 (90%) RB1+/+ vs. 60/93 (65%) RB1-/- retinoblastomas (p = 3·4 x 10-4) (table S6). No RB1-/-
primary tumours showed over >10 MYCN copies. In comparison, 16/30 (53%) RB1+/+ tumours
showed high-level MYCN amplification (28 to 121 copies of MYCN), (RB1+/+MYCNA
retinoblastoma) (p <0·0001) (tables 1, S6, figure 1). The remaining 14 RB1+/+ tumours showed
between 2 and 10 MYCN copies. For ten children with RB1+/+MYCNA tumours, normal DNA from
available blood showed two MYCN copies.
Compared to RB1-/- retinoblastoma, the 16 RB1+/+MYCNA tumours showed lower frequency of
copy-number changes in four other genes characteristic of retinoblastoma: gain of oncogenes KIF14
(62% vs. 19%; p = 0·002), DEK and E2F3 (57% vs. 6%; p = 0·0002) and loss of tumour suppressor
gene CDH11 (56% vs. 13%; p = 0·002) (tables S6, S7). SNP analysis indicated that the level of
normal cell contamination in the RB1+/+MYCNA tumours was low and similar to RB1-/- tumours
(table S2).
To probe for other genomic gains/losses, we studied DNA from 48 unilateral retinoblastomas by
aCGH10 and 3 by SNP analysis (Amsterdam) (14 RB1+/+MYCNA, 12 RB1+/+, 25 RB1-/-(+)) (tables S6,
S8, figure 2A). None of the RB1+/+MYCNA retinoblastomas showed any evidence of copy-number
changes or translocations11 at the RB1 locus. Except for MYCN copy-number, aCGH (figure 2A)
confirmed a reduced frequency in RB1+/+MYCNA retinoblastomas of the specific genomic copy-
number changes characteristic of RB1-/- retinoblastomas (figure 1, table S7).9 The RB1+/+MYCNA
retinoblastomas also showed overall significantly fewer altered bp and aCGH clones than the RB1-/-
retinoblastomas (p = 0·033) (figure S3, table S8).
MYCN amplicons of 14 (11 by aCGH, 3 by SNP analysis) RB1+/+MYCNA retinoblastomas, as
well as one RB1+/- tumour (T33) with 73 copies of MYCN, and one RB1-/- primary tumour (RB381)
9
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
MYCN amplified retinoblastoma with normal pRb
with 9·2 copies of MYCN, were narrow, spanning 1·1 to 6·3 Mbp encompassing MYCN (figures S4,
5, table S8). Importantly, the sole copy-number change detected in one RB1+/+MYCNA
retinoblastoma (E4) was 48 copies of 2p24·2-24·3 encompassing MYCN. The minimal common
amplicon defined by RB1+/+(-) MYCNA retinoblastomas T33 and P2 spanning 513 kbp, contained only
MYCN. RB1+/+MYCNA retinoblastomas T5 and P2 also defined a minimal common amplicon
including only MYCN. Of the remaining 35 tumours tested by aCGH, 24 unilateral tumours showed
no gain or loss at MYCN, and 11 had moderate gain spanning a broad region of at least 28 Mbp of
chromosome 2p, too large to meet the definition of amplification.12
Three (23%) RB1+/+MYCNA tumours showed 17q21·3-qter or 17q24·3-qter gain; two
RB1+/+MYCNA tumours showed 11q loss. Both regions are commonly altered in neuroblastoma,13, 14
but rare in RB1-/- retinoblastoma (present and published data15, 16). Other changes in RB1+/+MYCNA
retinoblastomas not often seen in RB1-/- retinoblastoma were gains at 14q and 18q, and losses at 11p
(figures 2A, S5).
Retinoblastoma T33 (RB1+/-) showed high-level MYCN amplification and loss of one copy of
most of 13q, including RB1; we suspect that amplification of MYCN initiated proliferation, followed
by 13q deletion. Since T33 showed numerous characteristics of RB1+/+MYCNA tumours, we included
T33 with MYCNA retinoblastomas in many analyses (figures 2, 4A, S4, S6).
RB1+/+MYCNA retinoblastomas expressed functional RB1 protein (pRB). The RB1+/+MYCNA
retinoblastomas and the expected retinal cell types expressed pRb (ganglion cells, specific inner
nuclear cells, and cone photoreceptors)17 and showed nuclear staining for C-terminal (figure 3A, S6)
and N-terminal (data not shown) epitopes, while RB1-/- and RB1+/- tumours were negative or stained
weakly, depending on their RB1 mutations. Western blots and immunoprecipitation in 3
RB1+/+MYCNA cell lines (A3, RB522, T101) derived from RB1+/+MYCNA primary retinoblastomas,
demonstrated functional nuclear pRb that was full-length (figures S6, S7), normally hypo- and
hyperphosphorylated (figure 3B, S7), and bound to E2F1 (figure 3C), the major target of pRb in
10
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
MYCN amplified retinoblastoma with normal pRb
control of the cell cycle.18RB1 transcripts were detected by end-point and real-time RT-PCR in the 3
RB1+/+MYCNA primary retinoblastomas for which mRNA was available (figures 3D, S8, table S9).
In contrast, most RB1-/- retinoblastomas expressed low RB1 transcript levels.
RB1+/+MYCNA primary retinoblastomas (figure 3A) and three derived cell lines (data not shown)
stained strongly for N-Myc protein. MYCN and KI67 transcripts (indicative of proliferation) were at
low levels in adult retina, at higher levels in fetal retina and RB1-/- retinoblastomas without MYCN
amplification, and at very high levels in primary RB1+/+MYCNA retinoblastoma and RB1-/- cell lines
with MYCN amplification (figure 3A, D, S8A, table S8). RB1+/+MYCNA tumours showed reduced
expression of the oncogene KIF14,9 in contrast to normal fetal retina, and RB1-/- primary
retinoblastomas and cell lines which over-express KIF14 (figure 3F, table S9).
RB1+/+MYCNA tumours expressed embryonic retinal cell markers consistent with a retinal origin.
The mRNAs of cone cell marker X-arrestin19 and CRX, a marker of retinal and pineal lineage
tumours, (strongly expressed in RB1-/- retinoblastoma but not in neuroblastoma,20) were expressed in
fetal retina, human adult retina, and RB1+/+MYCNA and RB1-/- primary retinoblastomas (figure S8B).
Children with RB1+/+MYCNA retinoblastomas were much younger at diagnosis than children with
unilateral RB1-/- retinoblastomas. The median age at diagnosis of 17 children (with 16 RB1+/+MYCNA
and one RB1+/-MYCNA [T33] tumours) was 4·5 months, significantly younger than children with
unilateral sporadic RB1-/- (24·0 months, p<10-4) or RB1+/+ (21·5 months, p<10-4) retinoblastomas
(figure 4A, tables S5A & S6). We estimate that 18·4% of children diagnosed with non-familial,
unilateral retinoblastoma at age six months or younger will have RB1+/+MYCNA retinoblastoma
(table S5B).
Analysis of age at diagnosis vs. proportion not yet diagnosed led Knudson to propose that two
hits initiate retinoblastoma.1 Our data from 79 unilaterally affected RB1-/- patients was consistent
with Knudson’s model and fit a two-hit curve, representative of two independent mutational events
in a tumour suppressor gene (figure 4B). Similar analysis for RB1+/+MYCNA tumours was
11
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
MYCN amplified retinoblastoma with normal pRb
inconclusive: while the data points for twelve children diagnosed before 10 months approximated
the calculated one-hit curve, ages for the older children deviated (figure 4B).
RB1+/+MYCNA tumours showed distinctive histology, displaying undifferentiated cells with large,
prominent, multiple nucleoli, and necrosis, apoptosis, and little calcification, similar to other
MYCNA embryonic tumours, such as neuroblastoma21 (figures 4C, S6). The Flexner-Wintersteiner
rosettes22 and nuclear molding characteristic of prototypic RB1-/- retinoblastoma (figure 4D) were
absent.
Clinically, the RB1+/+MYCNA retinoblastomas were large and invasive, considering the very
young age of these children (figures 4C, E, S7, S9). Three RB1+/+MYCNA retinoblastomas (RB522,
T101, and A3) from enucleated eyes grew rapidly into cell lines, unlike the RB1-/- retinoblastomas
that grow poorly in tissue culture. The cell lines still showed no RB1 mutations and maintained the
MYCN amplification of the primary, confirming minimal normal cell contamination of the primary
tumour. One RB1+/+MYCNA retinoblastoma had already invaded the optic nerve past the cribriform
plate at age 11 months, a feature of aggressive disease23, 24 (figure 4E). However, all the children
with RB1+/+MYCNA tumours were cured by removal of their affected eye with no adverse outcomes;
none developed retinoblastomas in their other eye.
12
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
MYCN amplified retinoblastoma with normal pRb
Discussion
Knudson’s analysis of retinoblastoma was fundamental to the concept that normal genes suppress
cancer,1 leading to the identification of the RB1 tumour suppressor gene,2 widely assumed to initiate
all retinoblastoma. We now describe a previously unrecognized type of retinoblastoma with no
evidence of RB1 mutations and functional pRb, likely initiated by amplification of the MYCN
oncogene.
Our international study of 1,068 unilateral retinoblastomas revealed a distinct RB1+/+MYCNA
subtype comprising 1·4% of unilateral, non-familial retinoblastomas. At least two previously
described tumours may also be RB1+/+MYCNA retinoblastomas,25 although RB1 genetic status was
not defined. Despite the low incidence of RB1+/+MYCNA retinoblastoma, RB1+/+MYCNA tumours
were independently discovered and characterized in the Toronto and Amsterdam labs, with different
patient cohorts and technologies.
We also identified 14 RB1+/+ retinoblastomas without MYCN amplification. aCGH showed
genomic gains and losses distinct from either RB1-/- or RB1+/+MYCNA retinoblastomas, but we
lacked RB1+/+ tumours for gene expression or protein studies. In particular, these RB1+/+
retinoblastomas showed frequent gain of 19p and q, 17p and q, 2p, and the telomeric end of 9q. The
RB1+/+ group appears heterogeneous and merits further study.
Our paper highlights how molecular diagnostics can identify novel malignancies that elude
histopathological recognition. Although RB1+/+MYCNA retinoblastomas were not previously
recognized as distinct, they resemble large nucleolar neuroblastomas with MYCN-amplification and
poor outcome.21 Like MYCN-amplified neuroblastomas,14 RB1+/+MYCNA tumours showed less
complex genomic copy-number alterations than tumours without MYCN amplification, suggesting
that MYCNA may be the critical driver of malignancy. However, the early age of diagnosis of
RB1+/+MYCNA tumours may allow less time to accumulate genomic alterations. Although whole
13
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
MYCN amplified retinoblastoma with normal pRb
genome sequencing recently suggested that point mutations (other than in the RB1 gene) are few in
RB1-/- retinoblastoma,26 loss of RB1 induces mitotic changes and lagging chromosomes,27 leading to
genomic instability. RB1 loss in retinoblastoma and pre-malignant retinoma8, 16 is also associated
with specific DNA copy-number changes, that are less frequent in RB1+/+MYCNA retinoblastomas.
Are RB1+/+MYCNA tumours truly retinoblastomas? RB1+/+MYCNA retinoblastomas arise in, and
express markers of embryonic retina, meeting the definition of retinoblastoma as ‘a blast cell
tumour arising from the retina’.19, 20, 28 We demonstrate intact RB1 genes in primary RB1+/+MYCNA
retinoblastomas, with strong nuclear pRb antibody staining. Three RB1+/+MYCNA cell lines
expressed full length RB1 mRNA, and inactive hyperphosphorylated and active
hypophosphorylated pRb that binds to E2F1, indicating normal functional pRb18. The possibility of
normal cell contamination masking an RB1 mutations in the RB1+/+MYCNA tumours is ruled out by
the cell lines that maintain the RB1+/+ genotype of the primary tumour, the very high level MYCN
amplification, histology with the only normal cells being vascular, and strong histological and
biochemical evidence that pRb is functional.
The very early presentation of RB1+/+MYCNA retinoblastomas, lack of RB1 mutations, functional
pRb and high MYCN-amplification in a relatively copy-number stable genome, suggests that these
retinoblastomas arise by somatic MYCN oncogene amplification in a retinal progenitor cell. This is
supported by one RB1+/+MYCNA tumour in which MYCN amplification was the only identified
genomic copy-number change. The RB1+/+MYCNA retinoblastomas are already large in very young
children, so they likely initiate earlier in retinal development than RB1-/- retinoblastomas. In
children with an RB1 mutation of equivalent age, the tumours are usually much smaller (figure 4F).
How MYCN-amplification is initiated, and whether MYCN-amplification alone suffices to initiate
these retinoblastomas remains to be formally demonstrated.
In retinal development the N-myc protein can support cell division in the absence of activating
E2fs; presumably unregulated MYCN expression associated with high-level gene amplification in
14
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
MYCN amplified retinoblastoma with normal pRb
RB1+/+MYCNA tumours promotes cell division indirectly through inactivation of pRb.29 The relative
genomic stability of RB1+/+MYCNA retinoblastomas suggests that anti-N-Myc therapeutic agents30
may avoid emergence of drug resistance acquired through progressive genomic rearrangements in
RB1-/- retinoblastoma. Our preliminary MYCN knock down experiments resulted in rapid death of
RB1+/+MYCNA cell lines. We predict that children less than one year of age with extra-ocular
retinoblastoma around the world23, may have RB1+/+MYCNA retinoblastomas that might benefit from
anti-N-Myc therapy. This idea is consistent with our (BG) anecdotal observation during review of
retinoblastoma pathology slides in Kenya, of RB1+/+MYCNA histology in an orbital recurrence,
confirmed to have 40 copies of MYCN.
Young age at diagnosis of unilateral retinoblastoma is frequently interpreted as an indication of
heritable retinoblastoma, and is currently cited as a reason to try to cure the cancer without
enucleation. However, attempts to salvage an eye with a large RB1+/+MYCNA retinoblastoma could
be dangerous; in our study, prompt removal of the unilateral affected eyes was curative with no
adverse outcomes. The young patients with RB1+/+MYCNA tumours in our study had large,
aggressive tumours, including invasion into the optic nerve. At similar young ages, the usual
hereditary RB1-/- tumours are very much smaller, detected only by active surveillance (figure 4F).
While our data predict that 18·4% of children diagnosed with non-familial, unilateral
retinoblastoma at age six months or younger will have RB1+/+MYCNA retinoblastoma (table S4B), if
tumour size is also considered, the risk will be higher, and RB1+/+MYCNA tumours may be clinically
predictable, facilitating prompt removal of these eyes with good outcomes for the children.
Standard care for unilateral non-familial retinoblastoma is identification of the RB1 mutant
alleles in tumour, and examination of blood to determine whether either mutant allele is germline.
Our study suggests that when no RB1 mutation is detected in a retinoblastoma tumour, determining
MYCN copy-number may assist on-going care. The diagnosis of RB1+/+MYCNA retinoblastoma
15
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
MYCN amplified retinoblastoma with normal pRb
strongly suggests non-hereditary disease with normal population risks for retinoblastomas in the
other eye and other cancers throughout life.
Our findings challenge the long-standing dogma that all retinoblastomas are initiated by RB1
gene mutations (summary figure 4G). RB1+/+MYCNA retinoblastoma provides an intriguing contrast
to classical retinoblastoma, of immediate importance to patients.
Panel: Research in Context
Systematic Review
We systematically searched the published, peer-reviewed literature on PubMed
(http://www.ncbi.nlm.nih.gov/pubmed/) using the search terms “retinoblastoma”, “initiation” and
“genetics”; “retinoblastoma tumour genetics”; “retinoblastoma development”; and “retinoblastoma
initiation”. We reviewed publications with a main focus on genetic initiation and development of
human retinoblastoma. We found no data that challenged Knudson’s 1971 conclusion that two rate-
limiting events1, later shown to be loss of both RB1 gene alleles,2-5, 7 are essential but not necessarily
sufficient for development of retinoblastoma. We found no data to suggest another form of
retinoblastoma.
Interpretation
Our collaborative studies identify a previously unrecognized disease: retinoblastoma with normal
RB1 genes, apparently driven by MYCN oncogene amplification. This newly recognised form of
retinoblastoma has immediate clinical implications for patients. RB1+/+MYCNA retinoblastomas are
diagnosed by molecular study of the tumour after removal of the eye of very young children with
unilateral non-familial disease. The children with RB1+/+MYCNA retinoblastoma and their relatives
are predicted to be at normal population risk for other cancers. Attempts to salvage the eye on the
assumption of heritable disease in young children may incur high treatment morbidity and failure to
cure these aggressive oncogene-driven retinoblastomas.
16
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
MYCN amplified retinoblastoma with normal pRb
17
1
MYCN amplified retinoblastoma with normal pRb
Tables
Test Site Total RB1+/+ Proportion RB1+/+Number
RB1+/+MYCNA
Proportion
RB1+/+MYCNA
Canada 441 7 1·6% 5 1·1%
Germany 400 12 3·0% 4 1·0%
France 150 5 3·3% 2 1·3%
New Zealand 30 2 6·7% 1 3·3%
The Netherlands 47 3 6·4% 3 6·4%
Total 1068 29 2·7% 15# 1·4%
*Fisher's exact tests indicates that percentage of unilateral tumours with RB1+/+ is not related to site (p = 0·08).
Table 1. Frequency of RB1+/+ unilateral retinoblastoma at five international diagnostic labs
18
1
2
3
4
MYCN amplified retinoblastoma with normal pRb
Figure Legends
Figure 1: Copy numbers of genes commonly altered in RB1-/-
retinoblastoma
(A) Box-plot of genomic copy-numbers determined by QM-PCR for indicated
genes in unilateral retinoblastoma, grouped by RB1 mutation status. T33 is an
outlier of the RB1+/- group, showing an RB1+/+MYCNA-like profile. Gray line, 2-
copies. (B) Heat-map for copy number for the profile genes (red, increased, and
blue, decreased copy-number; gray, 2 copies; white, not tested; n, number in
each group). Recently discovered RB1+/+MYCNA tumours not in (A) are indicated
by black triangles.
Figure 2: Genomic copy number alterations
(A) aCGH on 12 RB1+/+MYCNA (including T33) (red), 12 RB1+/+ (blue), 13 RB1+/-
(brown), and 11 RB1-/- (green) tumours; gains, right, and losses, left of
chromosome; minimal commonly gained/lost regions in RB1-/- tumours boxed;
*normally occurring copy number variations. Tumour T33 shows loss of most of
13q. (B) The minimal amplicon of 513 kbp is defined by two MYCNA tumours
(pink band); MYCN copy number by QM-PCR, red italics; aCGH individual
probes, green bars.
19
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
MYCN amplified retinoblastoma with normal pRb
Figure 3: Expression of RB1 and MYCN
(A) Staining of adjacent retina and RB1+/+MYCNA or RB1 mutant retinoblastoma
for N-Myc protein and pRB (C-terminus antibody); T, tumour; INL, inner nuclear
layer retina. (B) Western blots with pRb antibody that recognizes both hypo-
and hyperphosphorylated pRb, phospho-Rb (Ser795) antibody, and E2F1
antibody. (C) Cell lysates were immunoprecipitated with antibodies to mouse
IgG (negative control), pRb or E2F1, and western blots performed with
antibodies to pRb and E2F. (D) Real-time RT-PCR for RB1, MYCN and KIF14 in
human fetal (FR) and adult (HR) retina, RB1+/+MYCNA, and RB1-/- primary
tumours and cell lines; triplicate measurements normalized against GAPDH,
relative to FR; MYCN DNA copy-numbers in italics; #, KIF14 not done.
20
1
2
3
4
5
6
7
8
9
10
11
12
MYCN amplified retinoblastoma with normal pRb
Figure 4: Clinical features of children with RB1+/+MYCNA tumours
(A) Children with RB1+/+MYCNA retinoblastoma are diagnosed significantly
younger than children with RB1-/- tumours (p<0·0001, Wilcoxon rank sum test).
(B) The Knudson plot of proportion not yet diagnosed vs age at diagnosis, using
birth as a surrogate for initiation, showing a two-hit curve (blue) and a one-hit
curve (red) for children with RB1-/- tumours or RB1+/+MYCNA tumours;
scatterplot does not distinguish identically aged children. (C) Fundus image of
a large RB1+/+MYCNA unilateral tumour, extending from optic nerve (white
arrow) to anterior border of retina (double arrows) in a 4 month-old child with
characteristic calcification on ultrasound, and round nuclei with prominent
large multiple nucleoli on pathology, in comparison to (D) RB1-/- tumour
showing classic Flexner-Wintersteiner rosettes and nuclear molding;
hematoxylin-eosin staining. (E) RB1+/+MYCNA retinoblastoma in an 11 month-old
child (A2) with extra-ocular extension into the optic nerve (arrows) (2·5x,
hematoxylin-eosin staining). (F) In comparison, in 3·5 month-old child with
heritable RB1-/- retinoblastoma, a small tumour is present in the inner nuclear
layer of the retina on optical coherent tomography (OCT). (G) Schema of data
establishing RB1+/+MYCNA retinoblastoma as a novel disease; months (m), data
figures (f) and tables (t) indicated in grey on left.
21
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
MYCN amplified retinoblastoma with normal pRb
Contributors
Diane E Rushlow recognized the initial connection between MYCNA amplification and RB1
mutation status, performed literature search and QM-PCR analysis and supervised RB1 mutation
analysis, coordinated collaborations with the other sites and was the major contributor to manuscript
preparation. Berber M Mol recognized the RB1 and MYCN mutation status of the Amsterdam
samples by MLPA and SNP array analysis, performed immunohistochemistry, Western blots and
co-immunoprecipitations. Jennifer Y Kennett performed aCGH and analysed aCGH data. Stephanie
Yee performed analysis of aCGH data, the MYCNA alignment, immunohistochemistry and reverse
transcriptase PCR. Sanja Pajovic performed literature search, reverse transcriptase PCR and
immunohistochemistry. Brigitte L Thériault performed literature search and RNA expression
studies. Nadia L Prigoda-Lee performed literature search, statistical analysis and contributed to
figure and manuscript preparation. Clarellen Spencer performed immunohistochemistry. Helen
Dimaras and Timothy W Corson performed literature searches, assisted in data analysis and
conceptualization of discussion, and contributed to figure and manuscript preparation. Renee Pang
performed statistical and bioinformatic analyses on the aCGH data. Christine Massey performed
statistical analyses. Roseline Godbout discovered and characterised the first tumour with MYCN
amplification and normal pRb (RB522) (now RB1+/+MYCNA) long before anyone believed her; she
provided the cell line and data and Western blot showing functional pRb. Zhe Jiang and Eldad
Zacksenhaus performed Western blots to demonstrate functional phosphorylated pRb. Katherine
Paton and Annette C Moll provided clinical images and material, and conceptual discussion. Claude
Houdayer and Anthony Raizis provided RB1 mutation analysis, and clinical features. William
Halliday recognized and characterized the unique histological features of the RB1+/+MYCNA
retinoblastomas and prepared digital images for publication. Wan L Lam supervised Jennifer
22
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
MYCN amplified retinoblastoma with normal pRb
Kennet and the aCGH experiments. Paul C Boutros performed detailed and novel analysis of the
aCGH data, and statistical analyses throughout the project, and supervised Renee Pang. Dietmar
Lohmann performed literature search, provided RB1 mutation analysis, and contributed to figure
construction and development of concepts. Josephine C Dorsman coordinated the Amsterdam
study, and, with Berber Mol who she supervised, recognized the RB1 and MYCN mutation status of
the Amsterdam samples. Brenda L Gallie supervised overall, performed literature search, provided
critical guidance on all components of the project, and contributed extensively to figure and
manuscript preparation. All authors contributed to manuscript preparation.
Conflicts of interest
BLG was part-owner of Solutions by Sequence. All other authors declare that they have no conflicts
of interest.
Acknowledgments
This study was conducted with the support of the Ontario Institute for Cancer Research to PCB
through funding provided by the Government of Ontario. BMM was funded by a grant from
CCA/V-ICI/ Avanti-STR (to JCD, J. Cloos and ACM), the Dutch research was also funded in part
by KIKA (JCD, H. te Riele, J. Cloos, ACM). SY was funded by the Vision Science Research
Program of the University Health Network and the University of Toronto. RP was funded in part by
a Great West Life Studentship from Queen’s University School of Medicine. We thank Leslie
MacKeen for the montage of RetCam images in figure 3B. We thank Dr. Valerie White of U.
British Columbia for providing clinical and pathological details and images. We thank Cynthia
Vandenhoven for the clinical images in figure 4F. We thank members of the VU University
Medical Center/The Netherlands Cancer Institute, Institut Curie, Institut für Humangenetik, Toronto
retinoblastoma teams and other, wise colleagues for useful discussions. We thank the children and
families who donated tissues for these studies for the benefit of future families.
23
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
2425
MYCN amplified retinoblastoma with normal pRb
REFERENCES
1. Knudson AG. Mutation and cancer: statistical study of retinoblastoma. Proceedings of the National Academy of Science, USA 1971; 68(4): 820-3.
2. Friend SH, Bernards R, Rogelj S, Weinberg RA, Rapaport JM, Albert DM, et al. A human DNA segment with properties of the gene that predisposes to retinoblastoma and osteosarcoma. Nature 1986; 323(6089): 643-6.
3. Cavenee WK, Hansen MF, Nordenskjold M, Kock E, Maumenee I, Squire JA, et al. Genetic origin of mutations predisposing to retinoblastoma. Science 1985; 228(4698): 501-3.
4. Lohmann DR, Gallie BL. Retinoblastoma: Revisiting the model prototype of inherited cancer. Am J Med Genet 2004; 129C(1): 23-8.
5. Rushlow D, Piovesan B, Zhang K, Prigoda-Lee NL, Marchong MN, Clark RD, et al. Detection of mosaic RB1 mutations in families with retinoblastoma. Hum Mutat 2009; 30(5): 842-51.
6. Lohmann D, Gallie B, Dommering C, Gauthier-Villars M. Clinical utility gene card for: Retinoblastoma. Eur J Hum Genet 2011; 19(3).
7. Houdayer C, Gauthier-Villars M, Lauge A, Pages-Berhouet S, Dehainault C, Caux-Moncoutier V, et al. Comprehensive screening for constitutional RB1 mutations by DHPLC and QMPSF. Hum Mutat 2004; 23(2): 193-202.
8. Dimaras H, Khetan V, Halliday W, Orlic M, Prigoda NL, Piovesan B, et al. Loss of RB1 induces non-proliferative retinoma: increasing genomic instability correlates with progression to retinoblastoma. Human molecular genetics 2008; 17(10): 1363-72.
9. Corson TW, Gallie BL. One hit, two hits, three hits, more? Genomic changes in the development of retinoblastoma. Genes Chromosomes Cancer 2007; 46(7): 617-34.
10. Ishkanian AS, Malloff CA, Watson SK, DeLeeuw RJ, Chi B, Coe BP, et al. A tiling resolution DNA microarray with complete coverage of the human genome. Nat Genet 2004; 36(3): 299-303.
11. Watson SK, deLeeuw RJ, Horsman DE, Squire JA, Lam WL. Cytogenetically balanced translocations are associated with focal copy number alterations. Hum Genet 2007; 120(6): 795-805.
12. Myllykangas S, Bohling T, Knuutila S. Specificity, selection and significance of gene amplifications in cancer. Semin Cancer Biol 2007; 17(1): 42-55.
13. O'Neill S, Ekstrom L, Lastowska M, Roberts P, Brodeur GM, Kees UR, et al. MYCN amplification and 17q in neuroblastoma: evidence for structural association. Genes Chromosomes Cancer 2001; 30(1): 87-90.
14. Mosse YP, Diskin SJ, Wasserman N, Rinaldi K, Attiyeh EF, Cole K, et al. Neuroblastomas have distinct genomic DNA profiles that predict clinical phenotype and regional gene expression. Genes Chromosomes Cancer 2007; 46(10): 936-49.
15. Chen D, Gallie BL, Squire JA. Minimal regions of chromosomal imbalance in retinoblastoma detected by comparative genomic hybridization. Cancer Genet Cytogenet 2001; 129(1): 57-63.
16. Sampieri K, Amenduni M, Papa FT, Katzaki E, Mencarelli MA, Marozza A, et al. Array comparative genomic hybridization in retinoma and retinoblastoma tissues. Cancer Sci 2009; 100(3): 465-71.
17. Spencer C, Pajovic S, Devlin H, Dinh QD, Corson TW, Gallie BL. Distinct patterns of expression of the RB gene family in mouse and human retina. Gene Expr Patterns 2005; 5(5): 687-94.
18. Templeton DJ, Park SH, Lanier L, Weinberg RA. Nonfunctional mutants of the retinoblastoma protein are characterized by defects in phosphorylation, viral oncoprotein association, and
24
1
23456789
10111213141516171819202122232425262728293031323334353637383940414243444546
MYCN amplified retinoblastoma with normal pRb
nuclear tethering. Proceedings of the National Academy of Sciences of the United States of America 1991; 88(8): 3033-7.
19. Murakami A, Yajima T, Sakuma H, McLaren MJ, Inana G. X-arrestin: a new retinal arrestin mapping to the X chromosome. FEBS Lett 1993; 334(2): 203-9.
20. Terry J, Calicchio ML, Rodriguez-Galindo C, Perez-Atayde AR. Immunohistochemical Expression of CRX in Extracranial Malignant Small Round Cell Tumours. The American journal of surgical pathology 2012; 36(8): 1165-9.
21. Tornoczky T, Semjen D, Shimada H, Ambros IM. Pathology of peripheral neuroblastic tumours: significance of prominent nucleoli in undifferentiated/poorly differentiated neuroblastoma. Pathol Oncol Res 2007; 13(4): 269-75.
22. Flexner S. A peculiar glioma (neuroepithelioma?) of the retina. Johns Hopkins Hosp Bull 1891; 2: 115.
23. Dimaras H, Kimani K, Dimba EA, Gronsdahl P, White A, Chan HS, et al. Retinoblastoma. Lancet 2012; 379(9824): 1436-46.
24. Chantada GL, Casco F, Fandino AC, Galli S, Manzitti J, Scopinaro M, et al. Outcome of patients with retinoblastoma and postlaminar optic nerve invasion. Ophthalmology 2007; 114(11): 2083-9.
25. Lillington DM, Goff LK, Kingston JE, Onadim Z, Price E, Domizio P, et al. High level amplification of N-MYC is not associated with adverse histology or outcome in primary retinoblastoma tumours. Br J Cancer 2002; 87(7): 779-82.
26. Zhang J, Benavente CA, McEvoy J, Flores-Otero J, Ding L, Chen X, et al. A novel retinoblastoma therapy from genomic and epigenetic analyses. Nature 2012; 481(7381): 329-34.
27. Manning AL, Longworth MS, Dyson NJ. Loss of pRB causes centromere dysfunction and chromosomal instability. Genes Dev 2010; 24(13): 1364-76.
28. Kobayashi M, Takezawa S, Hara K, Yu RT, Umesono Y, Agata K, et al. Identification of a photoreceptor cell-specific nuclear receptor. Proceedings of the National Academy of Sciences of the United States of America 1999; 96(9): 4814-9.
29. Chen D, Pacal M, Wenzel P, Knoepfler PS, Leone G, Bremner R. Division and apoptosis of E2f-deficient retinal progenitors. Nature 2009; 462(7275): 925-9.
30. Hook KE, Garza SJ, Lira ME, Ching KA, Lee NV, Cao J, et al. An integrated genomic approach to identify predictive biomarkers of response to the aurora kinase inhibitor PF-03814735. Molecular cancer therapeutics 2012; 11(3): 710-9.
25
123456789
101112131415161718192021222324252627282930313233
34
35