Download Molecular biology of testicular germ cell tumours Rolf Inge Skotheim

Molecular biology of testicular germ cell tumours
New insights into a genetic developmental model
Rolf Inge Skotheim
Department of Genetics
Institute for Cancer Research
The Norwegian Radium Hospital
Department Group of Laboratory Medicine
Faculty of Medicine
University of Oslo
The Research Council of Norway
A thesis for the doctor philosophiae degree, Oslo 2002
© Rolf Inge Skotheim
ISBN 82-8072-065-0
Cover:
Inger Sandved Anfinsen
Series of dissertations submitted to the Faculty of Medicine,
University of Oslo
No. 92
All rights reserved. No part of this publication may be reproduced
or transmitted, in any form or by any means, without permission.
Printed in Norway:
GCS Media AS, Oslo
Publisher:
Unipub AS, Oslo 2003
Unipub AS is a subsidiary company of Akademika AS owned by
The University Foundation for Student Life (SiO)
TABLE OF CONTENTS
ACKNOWLEDGEMENTS ................................................................................................................ 5
PREFACE...................................................................................................................................... 7
LIST OF PAPERS ........................................................................................................................... 9
THE GENETIC MAKEUP OF CANCER .......................................................................................... 11
INTRODUCTION TO TESTICULAR GERM CELL TUMOUR ............................................................. 17
Epidemiology ....................................................................................................................... 17
Risk factors and hereditary predisposition ........................................................................... 18
Histopathology ..................................................................................................................... 19
Treatment and outcome ........................................................................................................ 22
Genome and epigenome ....................................................................................................... 24
AIMS ......................................................................................................................................... 29
RESULTS IN BRIEF ..................................................................................................................... 31
DISCUSSION............................................................................................................................... 35
Hereditary and sporadic TGCTs have similar genetic complements ................................... 35
The TGCT transcriptome ..................................................................................................... 39
Translational genomics by using tissue microarrays............................................................ 43
TGCT candidate genes and their cellular context ................................................................ 46
CONCLUSIONS ........................................................................................................................... 53
FUTURE PROSPECTIVES ............................................................................................................. 55
REFERENCES ............................................................................................................................. 57
ORIGINAL ARTICLES .....................................................................................................................
APPENDICES ..................................................................................................................................
Appendix I. Abbreviations ......................................................................................................
Appendix II. Genes putatively involved in development of TGCT .........................................
ACKNOWLEDGEMENTS
The present work has been carried out at the Department of Genetics, Institute for Cancer
Research at the Norwegian Radium Hospital, and has been financially supported by the
Research Council of Norway.
I will thank my scientific supervisor Ragnhild Lothe for the greatest supervision, her
enthusiasm, constructive criticism, and the well-arranged research projects which made this
thesis possible. I also acknowledge our head of department, Anne-Lise Børresen-Dale, for
providing advanced research facilities, her catching science-mania, and for being my official
link to the Faculty of Medicine. I thank Maja Kraggerud for good collaboration and rewarding
discussions on germ cell tumours. Sharing office with Mr. Diep implies having a good time at
work, and I thank all friends and colleagues within Ragnhild’s research group and the whole
Department of Genetics who make the days easy going.
Much of the laboratory work concerning the cDNA and tissue microarray studies was done in
Anne and Olli-Pekka Kallioniemi’s laboratory at the National Institutes of Health. I
appreciate all help from them and the others in their research group for making my visits
worthy, both on the scientific and social level.
When working with the tissue microarrays, I also had the pleasure to collaborate with several
pathologists, and in particular Head of Department of Pathology, Jahn Nesland and Vera
Abeler have generously spent lots of time by the microscope. During the practical work, I also
had the pleasure to get to know the nice people in the Pathology laboratory.
Most of the work has been concerning investigation of clinical samples, and I am grateful for
the collaboration with the clinicians, Sophie Fosså and Nina Aass, who both meet the patients
and systematically collect the clinical information into well-organised databases.
I would also express gratitude to the rest of the collaborators and co-authors.
Domestically, credit goes to Anne Lise and Sander for coping with a distracted scientist.
Oslo, Dec. 23, 2002
5
PREFACE
The work of this thesis was carried out throughout the first three years of this millennium.
During this time, new high-throughput molecular biological methodologies have come to use
and the draft sequence of the human genome was completed.
Testicular germ cell tumour (TGCT1) is the most common cancer type among young adult
males, and the incidence has been increasing over the past fifty years. The introduction of
cisplatin-based chemotherapy has led to good prognoses for patients diagnosed with TGCT,
but the treatment is not optimal in terms of quality of life. TGCT also sheds biological
enigmas, and research on the molecular mechanisms of TGCT development has relevance for
both normal germ-cell biology and regulation of embryonal differentiation switches, in
addition to the clinical potential of differential diagnosis, prognosis, and treatment of these
patients.
Through five reports, the current thesis investigates the molecular biology of TGCT. A
methodological evaluation of two ways to detect allele specific changes in tumour DNA was
carried out to enable comparisons of data obtained by the two methods. Allelic imbalance
studies are commonly used in the hunt for tumour suppressor genes, and we took advantage of
this approach and found frequent changes within genomic regions with linkage to TGCT
susceptibility. The similarity of the genetic changes between hereditarily predisposed and
sporadic TGCTs made us believe that both groups develop through disruption of the same
molecular pathways. This hypothesis was strengthened by the demonstration that these two
groups of patients have strikingly similar and non-random patterns of genome-wide DNA
copy number changes in their tumours. This work highlighted the significance of increased
copy number of the distal part of chromosome arm 17q, occurring in every second TGCT. We
then focused into that region by a gene expression analysis using cDNA microarrays and
analysed the transcriptional level of all available genes and expressed sequence tags. Several
genes were identified as aberrantly expressed in TGCT. The final work of this thesis took
account for the increasing demand of validation of new potential disease markers, and we
1
See Appendix I for complete list of abbreviations.
7
constructed a tissue microarray which allows for rapid characterisation of new gene and
protein markers within hundreds of testicular tissue samples. The strength of this tool was
demonstrated by the frequently deregulated protein levels of four new candidate genes
recently identified by us, as well as one previously reported candidate gene. We also
identified several associations between the analysed markers and various differentiation steps
of TGCT.
8
LIST OF PAPERS
I Skotheim RI, Diep CB, Kraggerud SM, Jakobsen KS, and Lothe RA (2001). Evaluation
of loss of heterozygosity/allelic imbalance scoring in tumor DNA. Cancer Genetics and
Cytogenetics 127(1): 64-70.
II Skotheim RI, Kraggerud SM, Fosså SD, Stenwig AE, Gedde-Dahl T Jr, Danielsen HE,
Jakobsen KS, and Lothe RA (2001). Familial/bilateral and sporadic testicular germ cell
tumors show frequent genetic changes at loci with suggestive linkage evidence. Neoplasia
3(3): 196-203.
III Kraggerud SM, Skotheim RI, Szymanska J, Eknæs M, Fosså SD, Stenwig AE, Peltomäki
P, and Lothe RA (2002). Genome profiles of familial/bilateral and sporadic testicular
germ cell tumors. Genes Chromosomes and Cancer 34(2): 168-174.
IV Skotheim RI, Monni O, Mousses S, Fosså SD, Kallioniemi OP, Lothe RA, and
Kallioniemi A (2002). New insights into testicular germ cell tumorigenesis from gene
expression profiling. Cancer Research 62(8): 2359-2364.
V Skotheim RI, Abeler VM, Nesland JM, Fosså SD, Holm R, Wagner U, Aass N,
Kallioniemi OP, and Lothe RA. Candidate genes for testicular cancer evaluated by in situ
protein expression analyses on tissue microarrays. Submitted manuscript.
9
THE GENETIC MAKEUP OF CANCER
In normal tissue, a homeostasis ensures that the total cell masses remain more or less constant
through tightly regulated processes of cell growth, proliferation and death (apoptosis).
Disruption of this homeostasis may result in neoplastic growth, and a neoplasm might
eventually form a tumour. Malignant tumours differ from benign tumours in their capacity to
invade and metastasise.
Contemporary with the rediscovery of Mendel’s Laws in the early twentieth century,
Theodore Boveri published his chromosomal theory of heredity, and hence provided a
mechanistic basis for the transmission of hereditary traits explained by Mendel. Based on
observations of abnormal growth of sea-urchin eggs that carry the “wrong” chromosomal
complement, Boveri proposed that tumour growth is based on a similar, but particular,
incorrect combination of chromosomes. This is now known as the somatic mutation theory of
cancer, and paved the way for the field of cancer genetics (Figure 1; ref. 1).
The current biological view of cancer is that, in general, cancer originates from a single cell
and its progeny (i.e. clonal expansion). The cells within this clone accumulate within them a
set of genetic and/or epigenetic changes leading to qualitative and quantitative alterations of
gene products, and hence a variation of phenotypes subjected to selection (Figure 2A; ref. 2).
Hence, tumour development proceed through an analogous process to Darwinian evolution, in
which a succession of genetic changes, each conferring one or another type of growth
advantage, leads to a progressive conversion of normal human cells into cancer cells. It
should however be noted that there are cytogenetic lines of evidence for the existence of
polyclonal cancers (3). The changes a cell requires to turn malignant have been categorised
into six “hallmarks of cancer” (Figure 2B): evasion of apoptosis, self-sufficiency in growth
signals, insensitivity to growth-inhibitory signals, sustained angiogenesis, limitless replicative
potential, and tissue invasion and metastasis (4).
11
Figure 1. Timeline of selected historical hallmarks within genetics and cancer genetics. Original
references: a(5,6), b(7-9), c(10-12), d(13), e(14), f(15), g(16), h(17), i(18), j(19), k(1), l(20,21), m(22,23),
n
(24,25), o(26,27), p(28), q(29,30), r(31,32), s(33), t(34), u(35), v(36), w(37), x(38), y(39), z(40), aa(41),
ab
(42), ac(43,44), ad(45-47), ae(48), af(49).
12
The Genetic Makeup of Cancer
A)
B)
Cancer
Normal
Premalignant
Malignant
Figure 2. Current biological view of cancer. (A) Most cancers are believed to undergo a clonal
expansion, and the genetic or epigenetic events indicated may be of any kind giving the cell a
selective advantage. (B) Ultimately, these changes may fulfil Hanahan and Weinberg’s six hallmarks
of cancer (4).
The multistep phenotypic changes during the tumourigenesis are consequences of dominant
gain-of-function mutations of proto-oncogenes and recessive loss-of-function mutations of
tumour suppressor genes. The multistep process of tumourigenesis fits well with the
observation that most types of cancers have an increased incidence with age, as it takes time
to acquire a sufficient amount of tumourigenic mutations. Calculations of how many such
mutations cancer genomes acquire within cancer relevant genes vastly exceed the number of
mutations possibly reached in cells during normal life spans by normal mutation frequencies
(50). This, in addition to the fact that certain cancer types occur very early in life, led to the
postulation that many cancers develop in cells with a “mutator phenotype” (51). Later this has
been confirmed on the molecular mechanistic level for several genes that normally function in
the maintenance of genetic and genomic stability (50). An example is the defect mismatch
repair system in the hereditary non-polyposis colorectal cancer, where the predisposed
individuals carry germ-line mutations in one out of totally five genes encoding mismatch
repair components (52). This is an inherited mutator phenotype which makes the cells unable
to repair mismatched base pairs, and the accumulation of mutations makes the individual
predisposed to cancer. However, there are also cancer predisposition genes with no known
connection to a mutator phenotype (e.g. RB1, APC, CDKN2A, and CDK4, predisposing to
retinoblastoma, colorectal cancer, and melanoma; refs. 52-54). These genes are also
frequently somatically mutated during development of other cancer types.
13
Mutations in proto-oncogenes and tumour suppressor genes contribute to the malignant
phenotype as the products of these genes have functions related to one or several of the six
hallmarks of cancer. A proto-oncogene may be a promoter of cell growth and division (either
as a cell cycle component or as any upstream stimulating factor), an inhibitor of apoptosis, or
a promoter of angiogenesis. Proto-oncogenes may stimulate tumourigenesis upon both
overexpression and activating mutations (Figure 3). Conversely, tumour suppressor genes
inhibit the tumourigenesis, but loss of expression and mutations causing non-functional
products may promote the malignant phenotype. Tumour suppressor genes usually need to be
inactivated in both alleles in order to be tumourigenic, which often involve point mutation of
one allele (either somatic or in the germ line, as in individuals predisposed to cancer) and
inactivation of the other by loss of chromosomal material, ranging in extent from a sub-band
to the whole chromosome. The second hit has been investigated by numerous loss of
heterozygosity (LOH) studies. In addition, for some tumour suppressor genes, inactivation of
only one of the alleles may also promote tumour growth through haploinsufficiency (55,56).
Figure 3. Examples of molecular changes that activate proto-oncogenes and inactivate tumour
suppressor genes. Both alleles of a tumour suppressor gene usually need to be inactivated to
promote tumourigenesis, but only one of the alleles is illustrated. The exemplified changes are all cisacting events which are heritable through cell generations. Trans-acting factors come in addition,
which may interfere with gene product levels, stability, and activity.
14
The Genetic Makeup of Cancer
The point mutational alterations of proto-oncogenes and tumour suppressor genes only
account for a fraction of the genetic changes in tumours. The tumour genomes are often
highly unstable and undergo whole ploidy changes as well as aneuploidisation and aberrations
within single chromosomes (Figure 3). Aberrations affecting the DNA copy number may
promote tumourigenesis through the subsequent change of transcriptional levels of cancer
relevant genes. Chromosomal amplification leading to overexpression of a proto-oncogene
may have the same effect on the cellular phenotype as an activating mutation of the same
gene.
It has been demonstrated that such chromosomal instability may be caused by numeric and
structural aberrations of centrosomes, which are commonly seen in tumour cells (57).
Centrosome aberrations and chromosomal instability are however expected to enhance one
another. Excess centrosomes could appear through overduplication within a single cell cycle,
through aborted cell division, cell fusion, or de novo genesis (57).
In the same manner as genomic changes may cause genome-wide transcriptional alterations,
aberrant patterns of the epigenome2 also will lead to massive changes in gene expression.
Most cancer epigenomes are hypomethylated but reveal hypermethylation within specific
CpG islands3 compared to epigenomes of normal non-malignant cells (58). The majority of
the CpG islands are located within regulatory elements of genes, and de novo methylation of
residing cytosine residues is associated with transcriptional silencing (59). Such inappropriate
epigenetic silencing of tumour suppressors and DNA repair genes may be just as common as
genetic deletions and point mutations (60).
Functional disruption of DNA (cytosine-5-)methyl transferases may cause aberrant
methylation patterns, and can be seen in analogy with mutator phenotypes, as the number of
aberrantly expressed genes increases. Indeed, a CpG island methylator phenotype has been
described in colorectal carcinogenesis (61).
In recent years, there has been an emergence of high-throughput tools facilitating analyses of
DNA, RNA, and protein levels, often encompassing the whole genome, transcriptome, and
2
Epigenome - the genome-wide DNA methylation and histone modifications.
3
CpG island - regions of the genome with significantly higher than average content of the CpG dinucleotide.
15
proteome. Such global-scale analyses are popularly denoted genomics, transcriptomics, and
proteomics. Similarly, we also have epigenomics. Identification of new cancer genes based on
quantitative alterations may be done routinely on a high-throughput scale by various
microarray technologies. Analogous techniques to identify cancer genes with qualitative
alterations are not yet available, and for those one still need to test one gene at the time.
16
INTRODUCTION TO TESTICULAR GERM CELL TUMOUR
Epidemiology
Testicular cancer is the most common cancer among young adult males in the Western
Industrialised world, and the incidence has increased by three- to four-fold over the past fifty
years (62,63). The current age-adjusted incidence rate of testicular cancer in Norway is 10 per
100.000 males per year (Figure 4).
Figure 4. Incidence rate for testicular cancer compared to that of other cancers in Norway. (A)
Trends in age-standardised incidence rates. (B) Age-specific incidences, 1995-99. The raw data were
obtained from the Norwegian Cancer Registry; http://www.kreftregisteret.no/
Ninety-five percent of the testicular tumours are of germ cell origin, and are hence called
testicular germ cell tumours (TGCT). Females may also develop germ cell tumours (GCTs) in
their gonads (ovarian GCT) and GCTs may also arise extra-gonadally in both sexes.
Three kinds of TGCT exist that are manifest at different times in life. The infantile TGCT are
usually of yolk-sac or choriocarcinoma differentiation (see below for histology). Post-pubertal
TGCT is usually manifest in young adults, and are of both seminoma and nonseminoma
differentiation. Spermatocytic seminoma is the TGCT of elderly men.
Different from the infantile TGCT and spermatocytic seminoma, the post-pubertal TGCT
(from here on and onwards only denoted TGCT) is virtually always seen in connection with a
carcinoma in situ (CIS, or intratubular malignant germ cells) from which they are believed to
originate. Although TGCT becomes clinically manifest only after puberty, studies of
17
epidemiology (63) and of cellular markers (64,65) in human TGCT, as well as graft
experiments in mice (66), indicate that CIS initiates during foetal life from primordial germ
cells (PGCs) or gonocytes (67,68) (Figure 6A).
Risk factors and hereditary predisposition
Patients with history of undescended testis (cryptorchidism) have a well-documented
increased risk of developing TGCT (69). Similarly, other abnormalities of male reproductive
health such as reduced semen quality, infertility, and hypospadias are also associated to
TGCT, and interestingly all these have had similarly increased incidences during the past few
decades (70). Together, these male reproductive problems may all be symptoms of an
underlying testicular dysgenesis syndrome (TDS; Figure 5), of which testicular cancer is the
most severe symptom (71,72). Experimental and epidemiological studies suggest that TDS is
a result of disrupted embryonal programming and gonadal development during foetal life
(72). Even though TDS is thought to commonly affect genetically susceptible individuals, the
increased incidence may be explained due to unfavourable environmental effects that have
been increasingly common over the past few decades. It has been hypothesised that exposure
of the male foetus to high levels of oestrogens, such as diethylstilbestrol, results in the
reproductive defects mentioned in connection to TDS (70).
Figure 5. Components and clinical manifestations of the testicular dysgenesis syndrome.
Modified after (72).
TGCT tends to cluster within families, where brothers and sons to testicular cancer patients
have a 6- to 10-fold increased risk of developing a testicular tumour compared with the
general population (73-78). This observed familial clustering and the high frequency of
bilateral disease (79) suggest that TGCT predisposition is heritable. Shared environment may
of course be an explanation, but a recent Swedish study estimated the contribution of
inherited factors to the causation of testicular cancer to be higher than that of any other cancer
18
Testicular germ cell tumour
type except cancers of the thyroid and the endocrine system (80). Both a segregation analysis
(81) and an analysis based on the frequency of bilateral disease (79) have favoured a recessive
model of inheritance. Both analyses estimated the penetrance among homozygotes for the
malignant genotype to be about 45%. Hence, these two studies propose testicular cancer to be
caused by a relatively common recessive allele (4% allele frequency in the general
population), possibly in combination with polygenic-environmental effects operating
predominantly in recent generations. The calculations also estimated that inherited testicular
cancer susceptibility accounts for one-third of all cases (79). An international testicular cancer
linkage consortium (82) has been collecting families with two or more cases of testicular
cancer and performed a linkage analysis in order to identify testicular cancer susceptibility
loci. This approach has so far resulted in four autosomal regions with suggestive linkage
evidence, namely 3q26-ter, 5q14-22, 12q24.3, and 18q12-ter4 (82), and in one X-linked
region with significant linkage evidence at Xq27.3, named TGCT1 (83).
Murine TGCT susceptibility also fits a recessive model of inheritance, and is similarly to
human TGCT under multigenic control (84,85). However, the mouse models only develop
nonseminoma TGCT, and quite early in life. Hence, this TGCT model system most likely
more relevant to the infantile type of TGCT. The murine TGCT predisposing loci Pgct1 (85),
Ter (86), and Tgct1 (84) at mouse chromosomes 13, 18, and 19, are syntenic to the human
chromosome bands 9q22/5q31-32, 5q31, and 10q21-24, respectively5. The relevant genes are
unknown both in mouse and man.
Histopathology
TGCTs are histologically divided into two main subtypes, seminomas and nonseminomas
(87). Whereas the seminomas resemble the CIS cells, but do not constrain within the
seminiferous tubules and are quite proliferative, the nonseminomas develop through a
pluripotent embryonal carcinoma stage, which may differentiate into cells and tissue types of
all three primary germ layers at various stages of differentiation (somatically differentiated
teratomas and extra-embryonally differentiated choriocarcinomas and yolk sac tumours). This
4
The genome locations are based on the boundary markers given by the original paper, and are updated
according to the June 2002 genome assembly of the UCSC Genome Browser; http://genome.ucsc.edu/
5
The map positions of the markers showing strongest linkage, updated July 2002 according to Mouse Genome
Informatics at http://www.informatics.jax.org/ and the human-mouse homology map at the National Center for
Biotechnology Information (NCBI); http://www.ncbi.nlm.nih.gov/Homology/
19
resemblance of embryogenesis makes the genetics of TGCT relevant also to developmental
biologists.
The origin of the nonseminomas is somewhat disputed. Either, embryonal carcinomas
develop directly from CIS, or they develop through a seminoma stage (Figure 6A). The
former is supported by the fact that differences in centromere numbers and
immunohistochemical markers have been reported between CIS adjacent to seminomas and
CIS adjacent to nonseminomas (88,89). Despite that, there are only few differences
discovered, and to the best of my knowledge, none that have reached statistical significance.
The linear progression model of TGCT tumourigenesis is supported by the presence of
seminomatous components within many nonseminomas. Additionally, the common
observation of nonseminomatous metastases at autopsy in patients who died subsequent to an
orchiectomy that demonstrated only seminoma is consistent with the concept that seminomas
may evolve into other histological subtypes (90,91). A model illustrated by a tetrahedron
(Figure 6B) also describes seminomas as totipotent, and with the capacity to directly
differentiate into all types of nonseminomas (92). The tetrahedron model is supported by the
observations
of
seminomas
with
syncytiotrophoblastic
cells
(otherwise
seen
in
choriocarcinomas), and the presence of hCG within some seminoma cells (92,93).
Examination of TGCT often includes investigation of specific markers. One such TGCT
marker is the isochromosome 12p, i(12p), first identified by Atkin and Baker in 1982 (94).
Because i(12p) is present in more than 80% of the TGCTs it constitutes a useful diagnostic
marker to verify GCT origin (95,96). Serum markers in routine clinical use include alpha
foetoprotein (AFP), human chorionic gonadotropin (CGB), and lactate dehydrogenase (LDH),
of which the levels of all three markers correlate inversely with likelihood of survival (97).
Serum TGCT markers are obtained immediately after orchiectomy, and should, if they are
elevated, be tested for serially after orchiectomy to check whether they decrease according to
the normal decay rates (87).
Yolk sac tumour is the principal source of AFP, but AFP may also be present in some
embryonal carcinomas and teratomas (87). Syncytiotrophoblastic cells are almost exclusively
the source of CGB, and may constitute choriocarcinomas or be present as single cells within
other histological subtypes (87).
20
Testicular germ cell tumour
Figure 6. Histogenetic developmental models of TGCT. (A) By the model most acknowledged
today, carcinoma in situ (CIS) is thought to develop from a foetal primordial germ cell (PGC) or
gonocyte (GC), but the CIS does not develop into invasive TGCT until after puberty (67). Then, a CIS
may develop into both seminoma (Sem) and embryonal carcinoma (EC) cells. The EC cells are
pluripotent, and may differentiate further into various extra-embryonic tissues, like in choriocarcinomas
(CC) and yolk sac tumour (YST), and into somatic tissues (teratomas, Ter). Noteworthy, an alternative
model where CIS develops from a meiotic pachytene spermatocyte has been proposed (98). Whether
embryonal carcinoma (EC) develops directly from CIS, via a seminoma (Sem) stage, or both is also a
debated issue. In the normal situation, gonocytes develop into cells of the spermatogenic lineage,
spermatogonia (SG), spermatocytes (SC), spermatids (ST), and spermatozoa. (B) The more
controversial tetrahedron model (92,93).
Germ cell alkaline phosphatase (ALPPL2) and placental alkaline phosphatase (ALPP) are
both expressed in PGCs, but not in normal adult spermatogenic germ cells. Both ALPPL2 and
ALPP are expressed in CIS, and stay expressed in seminomas and most nonseminoma TGCTs
(Figure 7 and Figure 12, page 45), and are hence used as clinical markers to demonstrate germ
cell origin of tumour cells, as well as to distinguish CIS cells from normal spermatogenic
germ cells. ALPPL2 and ALPP are often jointly denoted PLAP as most antibodies rose
against them recognise epitopes on both of the closely homologous proteins.
21
Figure 7. Examples of immunohistochemical staining of ALPP/ALPPL2 (PLAP) in testicular
tissue cores on a tissue microarray (see page 43 for more information on the tissue microarray
technology). (A) Normal testicular parenchyma with no staining for PLAP. (B) Seminiferous tubules
with CIS cells immunopositive for PLAP. (C) PLAP positive seminoma.
Treatment and outcome
Until the introduction of cisplatin-based chemotherapy in the late 1970s, the survival rate was
low for patients diagnosed with TGCT (Figure 8). At present, virtually all patients with
localised TGCT survive their disease, and nine out of ten patients with metastatic TGCT
survive beyond five years after diagnosis (97). TGCTs may metastasise by both lymphatic
vessels and by the blood stream. The lung, liver, brain, and bone are in decreasing order the
most common sites for distant metastases (90,91). Several different systems for the staging of
testicular tumours are in use (93). The work of this thesis has followed the Royal Marsden
staging (Table 1; ref. 99) which is the system used at the Norwegian Radium Hospital.
Figure 8. Five-year relative survival by period
and stage. Localised TGCT versus TGCT with
regional or distant metastases. The raw data were
obtained from the Norwegian Cancer Registry;
http://www.kreftregisteret.no/
22
Testicular germ cell tumour
Table 1. Royal Marsden staging of TGCT.
Stage
Local disease
Metastatic
disease
I
Localisation
Testis only
IM
No findings of metastases, but positivity for
serum markers after orchiectomy indicate
metastatic disease
II
Involvement of infradiaphragmatic lymph
nodes
III
Supraclavicular or mediastinal involvement,
but with no extralymphatic metastases
IV
Extralymphatic metastases
At the Norwegian Radium Hospital, all patients diagnosed with TGCT have their testis
surgically removed (orchiectomy). Nonseminomas and seminomas each constitute about 50%
of all TGCTs. Tumours that contain both seminoma and nonseminoma components are
regarded as nonseminomas in respect of treatment. About 80% of the seminomas are of
clinical stage I, but 20% of these have micro-metastases, and clinically stage I seminomas are
therefore either treated by prophylactic radiotherapy, chemotherapy, or entered into a “Waitand-see” protocol (Table 2). Of the patients with clinically stage I nonseminomas without
vascular invasion, only 10% have micro-metastases at the time of diagnosis. Therefore the
wait-and-see is preferred to avoid over-treatment of 90% of the patients, but the follow-up is
close to detect the recurrences early. Patients with stage I nonseminomas with vascular
invasion, run a 50% risk of micro-metastases. These patients get adjuvant chemotherapy and
have a subsequent recurrence frequency of 1%. Patients with metastases at diagnosis are
treated with cisplatin based chemotherapy followed by resection of residual disease.
Table 2. Standard treatment, in addition to surgery, and respective five year survival-rates for patients
with local (stage I) and metastatic (stage II-IV) TGCT.
Seminomas
treatment
stage I
prophylactic radiotherapy,
chemotherapy, or wait-andsee
b
stage II-IV
metastases <3cm:
carboplatin + radiotherapy
metastases >3cm: BEPa
Nonseminomas
survival
>99%
treatment
no vascular invasion: wait-and-see
with vascular invasion: BEP
c
93%
survival
a
a
BEP or other cisplatin based
chemotherapy. Resection of
residual disease.
>99%
99%
c
87%
a
BEP, chemotherapy with a combination of bleomycin, etoposid, and cisplatin.
b
Both seminoma and nonseminoma patients with metastases to the cerebrum or bone are usually
given radiotherapy in addition to surgery and chemotherapy.
c
According to the International Germ Cell Cancer Collaborative Group (97).
23
More than 90% of metastases from TGCT have identical histological type to that of their
primary tumours (87,90). According to a meta-analysis of 5862 patients with metastatic GCT
(97), these patients may be subdivided into three prognostic groups with five year survival of
90%, 80%, and 50%. Primary site, number of metastases, metastatic sites, and serum marker
levels (AFP, CGB, and LDH) were the most important independent factors, and were all
included in the definition of the prognostic group classification (97).
Table 2 summarises the stage adapted treatment of TGCT together with the patients’ outcome.
Still, these figures apply only to the more developed countries where less than 10% of patients
diagnosed with testicular cancer die from their disease. According to the database of the
International Agency for Research on Cancer (IARC), the survival rate after testicular cancer
is only about 50% in Africa and South-Eastern Asia6.
Most TGCTs are sensitive to cisplatin-based chemotherapy, but a sub-group of resistant
TGCTs exists. The general treatment responsiveness in TGCT has in some reports been
explained by the expression of high-levels of wild-type TP53 (100). However, a study
analysing matched series of sensitive and resistant TGCTs for TP53 mutations and protein
levels did not find support for this hypothesis (ref. 101 and more refs. therein). A positive
correlation has in one study been reported between microsatellite instability and
chemoresistant TGCTs (102).
Genome and epigenome
Normal diploid germ cell precursors have to undergo a genome amplification when they
develop into CIS, because the CIS cells are generally highly aneuploid with hypertriploid
genomes (103). Seminomas have chromosome numbers similar to their adjacent CIS (103),
whereas nonseminomas usually have lower ploidies, in the hypotriploid range (104,105). Not
all TGCTs pass through a polyploidisation step, as there have been reported a few tumours
with near diploid genomes (106). Whether or not the CIS initiates by a single
endoreduplication or cell fusional event to duplicate its genome, extensive nondisjunction has
6
GLOBOCAN 2000: Cancer incidence, mortality, and prevalence worldwide. International Agency for Research
on Cancer, World Health Organisation; http://www-dep.iarc.fr/globocan/globocan.html
24
Testicular germ cell tumour
to happen afterwards, as the individual chromosome numbers rarely correspond to the ploidy
number.
In addition to being aneuploid, the isochromosome i(12p) is present in more than 80% of the
TGCTs, regardless of histology (94,95). Then again, nonseminomas tend to have a higher
copy number of this aberration than seminomas (106). Most of the TGCTs that lack the i(12p)
have amplified 12p genetic material by other mechanisms (107,108). Hence, chromosome
arm 12p is amplified in virtually all TGCTs, indicating that this is an early event in the TGCT
development. Because the i(12p) isochromosome has genetically identical arms (109), one
could speculate whether one of the long arms of that chromosome is absent from the TGCT
genome. Nevertheless, the notion of retained heterozygosity of at least some polymorphic loci
on chromosome arm 12q in i(12p) positive tumours tells us that aneuploidisation of the
genome has to precede the i(12p) formation (110). Further, cytogenetic analyses of aneuploid
CIS have revealed i(12p) only in a few cases (111,112), and by molecular cytogenetic studies,
increased 12p copy numbers in CIS are infrequently seen (113-116). Hence, gain of 12p is
mainly seen in association with invasive TGCT, and an increase in the 12p copy number may
facilitate survival of the tumour cells outside the seminiferous tubules (68). Both the
chromosomal region 12p11.2-12.1 (115,117-119) and one including 12p13 (107,120,121)
have been reported as smallest amplified regions, but generally, the whole chromosome arm
or even the entire chromosome is present in extra copies in the TGCT genome (121,122).
CIS has generally many of the same chromosomal imbalances as the corresponding invasive
TGCT, but seminomas tend to match their corresponding CIS closer than nonseminomas do
(88,116,123). In a review of cytogenetic analyses of 229 TGCTs, the most common structural
changes were affecting regions on the chromosome arms (in decreasing order of frequency)
12p, 17q, 1p, 1q, 9q, 22q, 6q, and 7p (95). A cytogenetic profile of chromosome losses and
gains in 209 TGCTs has also been published (122), and chromosomal imbalances common to
at least 15% of the TGCTs were gains of 1p36-q44, 3p26-29, 7, 8, 12, 17p11-q25, 20p12-q13,
21, 22p12-13, 22p10-q13, and X, and losses of 1p32-36, 2, 4, 5, 6, 9, 10, 11, 12q10-24, 13,
14, 15, 16, 17, 18, 19, 20p13, 20q12-13, and 22. The G-banding method requires culturing of
tumour cells, and the chromosomes in single nuclei are analysed. By comparative genomic
hybridisation (CGH), the average copy numbers of the DNA sequences from a tumour sample
is analysed (124). Generally, net copy numbers deduced from cytogenetic karyotypes match
the CGH profiles of TGCTs. A summary of the DNA copy number changes in the TGCT
25
genome is presented later (Figure 13, page 47) together with the map positions of genes with
probable relation to TGCT development.
Studies of LOH or allelic imbalance (AI) supplement the cytogenetic and CGH data in TGCT.
These studies usually aim to identify and narrow into regions that potentially harbour tumour
suppressor genes. Polymorphic loci within the chromosome arms 3p, 3q, 5q, 11p, 12q, and
18q seem to undergo particularly frequent allelic changes in TGCT (125-138). Smallest
regions of overlapping deletions have been recognised and suggest TGCT suppressor loci at
3p14 (138), various 5q-regions (134,136), and 12q22 (135). The only identified target gene in
any of these regions is FHIT at 3p14, of which a wide assortment of aberrant transcripts and
reduced protein expression were recently reported in TGCT (138).
Inactivation of the probably most frequently mutated tumour suppressor gene in human
cancer, TP53, referred to as “the guardian of the genome”, is associated to chromosomal
instability in many cancer types. As the TGCT genome is characterised by quite complex
karyotypes, it is somewhat surprising that most TGCTs express abundant levels of wild-type
TP53 (101,131,139-146). However, a few reports of TP53 mutations in TGCT exist
(100,147,148), and one reported four chemotherapy resistant teratoma TGCTs with TP53
mutations (100). Overall, there is a five percent TP53 mutation frequency in TGCT
(sequence-verified
and
non-silent),
taking
271
analysed
tumours
into
account
(100,101,131,140-148).
Microsatellite instability is not commonly seen in TGCTs (149), although site specific
instability has been reported (150,151). In a recent study, rare microsatellite instable TGCTs
were
found
associated
to
chemotherapy
resistance
(102).
In
spite
of
this,
immunohistochemistry of certain mismatch repair factors was neither sensitive nor specific
enough to predict the microsatellite instability status in TGCT (102), which is the case for
colorectal cancer (152), indicating that the phenotype of new microsatellite alleles seen in a
few TGCTs may not be caused by defects in the tested mismatch repair components.
The epigenetics of TGCT may be better understood when looking at some embryological
events. The foetal PGCs are set aside from the rest of the embryo during the epiblast stage, a
26
Testicular germ cell tumour
stage where all cells still are totipotent7 and have low levels of DNA methylation, which
mostly restricts to the parentally imprinted genes (153,154). The PGCs escape from the
epiblast layer just before a major de novo methylation event that is lineage-specific, and
differentially program epiblast cells into their definitive germ layers after gastrulation
(154,155). The PGCs enter extra-embryonic locations where little de novo methylation takes
place, and when the PGCs later in embryogenesis return to the interior of the embryo, they
have maintained their genetic totipotency in the form of a hypomethylated genome. Actually,
an erasure of imprinting takes place in the PGCs, and expression of imprinted genes are
biallelic in the germ line from the time that migratory PGCs enter the embryonic genital ridge,
and new imprinting may not be established until late in gametogenesis (154,156). TGCTs also
consistently express both parental alleles of imprinted genes (157-159), indicating their origin
from cells in which the parental imprinting has been erased. The biallelic expression of
normally imprinted genes in TGCT contrasts the monoallelic expression seen in developing
embryos, which TGCTs resemble histogenetically.
Whereas seminomas and nonseminomas generally have the same genetic alterations, the
epigenomes of the two main subtypes of TGCTs are remarkably different. By a genome-wide
methylation assay by restriction landmark genome scanning (RLGS), CpG island methylation
was virtually absent from seminomas, whereas the methylation level in nonseminomas was
similar to that of other solid tumours (160,161). Seminomas are also hypomethylated
throughout their genomes, within and outside the CpG islands, compared to nonseminomas
(161). Promoter hypermethylation is known to inactivate tumour suppressor genes in cancer.
The cell cycle inhibitor CDKN2A has been shown non-functional through this mechanism in
several cancer types. In TGCT, some report methylation of CDKN2A (162), others do not
(163). The DNA repair gene MGMT has recently been shown to frequently exhibit promoter
hypermethylation in TGCT (163), and although few cases were analysed, this
hypermethylation was associated with lack of MGMT protein expression, supporting that this
epigenetic event is functionally relevant. The combined evidences from the RLGS studies and
the study showing gene-specific inactivation of MGMT by promoter methylation support that
alterations of the TGCT epigenome are most likely an important and general mechanism
involved in deregulation of transcriptional programs in TGCT.
7
Totipotency - ability to differentiate into all other cell types.
27
Summarised, the general TGCT genome is hypo- to hypertriploid, has a complex karyotype
with excess of 12p genetic material, expresses wild-type TP53, has erased parental
imprinting, and an abnormal CpG island methylation pattern.
28
AIMS
TGCT may be looked upon as a disease of the genome, which is invariably altered at multiple
sites. The ultimate goal of our project is to identify such molecular defects and to turn these
discoveries into meaningful biology and clinical utility.
The aims of this thesis were three-fold. First, we wanted to identify genetic changes
associated with development of TGCT. This was achieved by examination of genotypes and
genome-wide copy number changes in a series of primary TGCTs, including both hereditary
and sporadic tumours. These studies involved a methodological study to set guidelines for
scoring of AI in tumour genomes. In search for potential TGCT susceptibility loci, we
compared the genotypes and genome-wide copy number changes in hereditary and sporadic
TGCTs.
Second, and based on the preceding work, we set out to identify target genes within a
frequently altered genomic region. This we would achieve by a detailed gene expression
profiling by use of custom-made cDNA microarrays.
Third, we aimed to validate the importance of newly identified candidate genes/proteins. We
therefore constructed a tissue microarray which is a high-throughput tool to discover
associations between molecular data and subsets of TGCTs with specific biological,
pathological, or clinical characteristics.
29
RESULTS IN BRIEF
Paper I. “Evaluation of loss of heterozygosity/allelic imbalance scoring in tumor DNA.” The
objective of this study was to evaluate how LOH and AI in tumour DNA are scored and to set
guidelines for how the scoring should be done. We found that there are good correlation
between results from the visually scored radioactive labelling protocol and the semiquantitative fluorescent primer protocol. To provide a threshold level for when to score a
tumour genotype as AI by the semi-quantitative protocol, we used the standard deviations of
repeated analysis of 485 constitutional heterozygous genotypes at 20 different dinucleotide
repeat loci. This led to a higher detection frequency than by visual scoring of autoradiographs.
Our data therefore suggest that one should use a different and lower threshold value when
results from both protocols are compared.
Paper II. “Familial/bilateral and sporadic testicular germ cell tumors show frequent genetic
changes at loci with suggestive linkage evidence.” The five genomic regions with suggestive
evidence of linkage to TGCT (82,83) were investigated for genetic changes in tumours. DNA
from matched series of possibly hereditarily predisposed (familial clustering and/or
bilaterality) and sporadic TGCTs were analysed for AI, using the guidelines set by Paper I, in
the autosomal regions, and for locus specific copy number changes in the hemizygous Xlinked region. The autosomal regions had all high frequencies of AI (ranging 36% to 79%),
and gain at the Xq loci was seen in more than 50% of the tumours. Changes at 3q and 12q
were significantly more frequent within nonseminomas than within seminomas. The degree of
Xq amplification varied among the loci in each of 5 tumours, and based on the breakpoints in
these, an overlapping region of highest gains was delineated at Xq28. None of the 5 genomic
regions revealed any particular differences between the hereditary and sporadic tumour
groups. For a subset of the tumours, we had information on the genome-wide DNA copy
numbers (Paper III), and we could therefore tell whether a detected AI most likely was caused
by gain or loss of genetic material. The paper concluded that gain of genetic material at distal
Xq and losses at 5q and 18q contribute to establishment of both seminomas and
nonseminomas, whereas imbalances at 3q as well as gains at distal part of 12q are associated
to nonseminomatous differentiation.
31
Paper III. “Genome profiles of familial/bilateral and sporadic testicular germ cell tumors.”
The genome-wide DNA copy-number statuses were assessed for 33 TGCTs, including 15
possibly hereditarily predisposed and 18 sporadic tumours, by CGH. Gains of the whole, or
parts of, chromosome 12 were found in all but three tumours. Furthermore, increased copy
numbers of the whole, or parts of, chromosomes 7, 8, 17, and X, and decreased copy numbers
of the whole, or parts of, chromosomes 4, 11, 13, and 18 were observed in at least half of the
tumours. Sixteen smallest regions of overlapping changes were defined on 12 different
chromosomes. The copy number karyotypes of hereditary and sporadic TGCTs were
strikingly similar, suggesting that both groups of tumours develop through the same genetic
pathways. Gains from 15q and 22q were significantly associated with seminomas, whereas
gain of the proximal 17q (17q11.2-21) and high-level amplification from chromosome arm
12p, as well as losses from 10q were associated with nonseminomas.
Paper IV. “New insights into testicular germ cell tumorigenesis from cDNA microarray
analyses.” Paper III demonstrated that chromosome arm 17q is frequently overrepresented in
TGCT genomes. Based on the presumption that genomic regions with common copy number
gains harbour one or several proto-oncogenes, of which increased DNA copy number lead to
increased expression and hence also activity, we searched for relevant overexpressed target
genes on chromosome arm 17q. By using a custom made cDNA microarray containing 636
genes, expressed sequence tags (ESTs), and predicted genes from chromosome 17 to evaluate
the expression levels in 14 TGCTs, one CIS, and three normal testicular tissues, we were able
to list a few genes with consistently high expression in the tumours. Among these, GRB78 and
JUP were the two most highly overexpressed genes. Due to the limited knowledge of altered
gene expression in development of TGCT, we also examined the expression levels of 512
additional genes located throughout the genome. Several genes novel to testicular
tumourigenesis were consistently up- or downregulated, including POV1, MYCL1, MYBL2,
MXI1, and DNMT2. Additionally, the previously reported overexpression of the protooncogenes CCND2 and MYCN were confirmed (164-169). The gene expression profiles were
generally different between seminomas and nonseminomas, and specifically, the average
expression level of GRB7 was significantly higher in nonseminomas than in seminomas,
whereas the expression levels of JUP, MYCL1, and POV1 were highest in seminomas.
8
See Appendix II for complete names of genes putatively related to TGCT.
32
Results in Brief
Paper V. “Candidate genes for testicular cancer evaluated by in situ protein expression
analyses on tissue microarrays.” Even though we had evidence for transcriptional
deregulation of several genes, the limited sample set of the expression profiling (Paper IV)
gave little information on associations to various subgroups of TGCTs. Advances in genomics
and proteomics will bring about long lists of candidate genes to TGCT, that will require
validation and characterisation in large sample sets. For this purpose, we constructed a tissue
microarray with 506 testicular tissue cores from TGCT samples of various histological types,
CIS and normal testicular tissues, punched out from orchiectomy specimens of 279 patients
with TGCT of all clinical stages. We took advantage of this tool to investigate further the in
situ protein expressions of three candidate genes from our expression profiling (JUP, GRB7,
and CCND2; Paper IV), and of the repair enzyme MGMT and tumour suppressor FHIT, two
genes recently identified as candidate TGCT target genes by our research group (138,163).
Whereas JUP, GRB7, and CCND2 immunopositivities were infrequent in normal testis, these
proteins were expressed frequently within subsets of CIS and TGCT. Conversely, expression
of MGMT and FHIT proteins were always present in normal testis, but frequently lost from
CIS and TGCT. An association between CCND2 expression and cryptorchidism was the only
association between the immunostaining and clinical data, but a large number of statistically
strong associations between protein expressions and various histological subtypes
demonstrated the strength of this tool in translational research.
33
DISCUSSION
The discussion is divided into four parts, of which the first three discuss the major findings of
this thesis, and relate these to various quality aspects of the applied methods and compare
these to other available technologies. The last part integrates novel and previously reported
molecular changes of the TGCT genome into established knowledge on three cell-signalling
pathways that may be of importance to development of TGCT.
Hereditary and sporadic TGCTs have similar genetic complements
There are several studies supporting that a subgroup of TGCTs are hereditarily predisposed
(73-81,170), but the genetics underlying this presumption is not known. Some studies have
analysed potential loci selected by an educated guess approach (74,140,171-175), whereas
others have more systematically scanned the genome (82,83,176) in their search for loci with
genetic linkage to TGCT.
To investigate the genetics of TGCT occurring in hereditarily predisposed individuals and
compare it to sporadic TGCT, we selected two tumour series that were comparable with
regard to histological subgroups, percentage of intact tumour tissue, and patients’ age at
diagnosis. By looking further into the most likely loci resulting from linkage studies (82,83),
we were able to detect frequent genetic changes within these genomic regions (Paper II). We
also investigated this tumour series for genome-wide DNA copy number changes by use of
CGH (Paper III). Our AI and CGH studies led to the joint conclusion that similar, if not equal,
genetic pathways are affected during the tumourigenesis of both hereditary and sporadic
TGCT.
The International Testicular Cancer Linkage Consortium (82) found by linkage analyses four
regions with likely locations of TGCT susceptibility genes. Whereas none of the 220 tested
genetic markers gave significant support for a TGCT predisposing locus, four genomic
regions, covering from 8 to 46 Mbp and located at 3q26-ter, 5q14-22, 12q24.3, and 18q12-
35
ter9, showed suggestive evidences of linkage. Later, significant linkage of a 3 Mbp region at
Xq27.3 was reported which were named TGCT1 (83). The matched tumour groups of
hereditarily predisposed and sporadic TGCTs revealed comparably high frequencies of
genetic changes at loci within all five regions. This is in contrast to the expected based on the
recessive model of inheritance where one would expect that one allele was disrupted
constitutionally in the predisposed individuals, and hence, LOH would be revealed through
the loss of the other allele. For the sporadic tumours, one would either see a higher frequency
of genetic changes, as both alleles need to loose their function, or one could see a lower
frequency if sporadic tumours develop through disruption of alternative pathways, not
involving inactivation of the predisposition gene. Hence, the similar frequencies of change,
speak in disfavour of these regions being predisposition loci, but the extent of changes argue
that the regions still may harbour genes important to TGCT, irrespective of the individuals’
predisposition.
Because all four autosomal regions had quite high frequencies of AI (range: 36% to 79%),
one could believe that any region in the genome would be similarly frequently altered.
However, another study also analysed TGCTs for AI within the same four genome regions
(partly overlapping), plus 4 others (137). TGCTs with losses within one or more loci were on
average reported to 45% (range: 44% to 46%) within the four possibly linked chromosome
arms, whereas this average was only 24% (range: 4% to 51%) for the other chromosome arms
(137). This study gave no information on the familial clustering or bilaterality of the TGCTs
investigated.
For the reason that such studies of AI are highly dependent on technical detection limits, we
carried out a detailed investigation of the variation among normal samples analysed multiple
times (Paper I). Therefore, we are confident that we only detected AI due to genetic changes
and not due to technical error. In Paper II we generally detected higher frequencies of AI than
other studies investigating AI within the same regions (129,131-133,137). However, these
studies have analysed AI by visual comparison of autoradiographic gel-bands. But based on
the comparison (Paper I) of this manual detection and the semi-quantitative fluorescent
protocol (used in Paper II), we may apply a second threshold for AI-scoring (QLOH<0.75)
9
The region sizes and locations are based on the boundary markers given by the original paper, and are updated
according to the June 2002 genome assembly of the UCSC Genome Browser; http://genome.ucsc.edu/
36
Discussion
allowing us to compare the results. After doing that, the frequencies of AI in Paper II are in
line with the frequencies of genetic change reported by others (129,131-133,137).
A major advantage of replacing the manual scoring system of the radioactivity protocol with
the semi-automated fluorescence protocol is the speeding up of AI/LOH detection. However,
lately, we have seen the emergence of technologies for further scaling-up the genotyping of
polymorphisms. Instead of using microsatellite markers (short tandem repeats of one to six
bp), single nucleotide polymorphisms (SNPs), may also be used for the same purpose. A main
advantage of using SNP is their abundance, and currently there is on average one SNP
available for every 1.2kb10, about one thousandth of the average distance between the
available microsatellite markers (27; used in Papers I and II). SNPs are often used in
population genetics, linkage studies, and within pharmacogenetics, and several highthroughput technologies on array formats have become available (177,178). These methods
are typically used qualitatively for genotyping, but a few studies have also quantified the
allele intensities of paired genotypes from constitutional and tumour DNA (179,180). By SNP
microarrays one can therefore detect AI genome-wide with high resolution, a technology
highly beneficial for current and future AI-studies.
We also investigated the genome-wide copy number changes of hereditary and sporadic
TGCT by using CGH (Paper III), and complementary to the AI-study (Paper II), this study
also found that the copy number karyotypes are virtually identical between the two groups of
TGCT (Figure 9). This gives further evidence that the two groups develop through disruption
of the same genetic pathways.
Besides the comparisons of hereditary and sporadic TGCT, this study increased significantly
the number of cases in the literature on DNA copy number changes in TGCT (reviewed in
181). The frequent changes detected in this study were generally in accordance with previous
studies, but an exception was the high frequency of gain of chromosome arm 17q. This was
seen in half of our cases, but has not been emphasised in previous CGH studies of TGCT.
Nevertheless, chromosome arm 17q has been listed second, after 12p, for having frequent
structural changes in the TGCT genome (95). By using conservative directions for delineating
smallest regions of overlapping changes, the study revealed as many as sixteen regions on
10
According to the NCBI dbSNP Build 108 (Nov. 6, 2002); http://www.ncbi.nlm.nih.gov/SNP/
37
Figure 9. Genome profiles of familial/bilateral and sporadic TGCTs. The graphs are drawn from
short to long arm direction along each chromosome. The visualisation was facilitated by software tools
made by Chieu Diep.
twelve different chromosomes, all potential locations of genes relevant to TGCT. Several of
the aberrations were associated with histological subtypes. Gains from 15q and 22q were
typically found in seminomas, whereas gains from proximal 17q, high-level amplifications
from 12p, and losses from 10q were most common in nonseminomas. The study was the first
to analyse a series of familial and bilateral TGCTs by CGH. The overall CGH-literature on
TGCT is summarised in Figure 13 (page 47).
All CGH-studies until date on TGCT have applied the classic CGH, hybridising fluorescently
labelled DNA onto normal metaphase chromosomes (124). By this method, copy number
changes affecting chromosome regions of at least 5-10Mbp are detected (182). During the
past few years, a technology has been developed for doing CGH on DNA templates spotted in
a microarray format (183,184). Here, differentially labelled DNA are cohybridised onto glass
slides with DNA vectors spotted in a tiny array pattern. The resolution by using array-CGH is
dependent on the DNA probes spotted onto the array and their genomic localisations. A
genome-wide resolution down to 100-300Mbp may be achieved by using overlapping
bacterial artificial chromosomes (BACs) or P1-derived artificial chromosomes (PACs), or a
gene to gene resolution would be achieved by spotting every gene transcript represented by
cDNA clones. In terms of resolution, CGH on microarrays are therefore superior to CGH on
metaphase chromosomes. However, microarrays for CGH have not yet been made that cover
the whole genome, and they usually consist of tiling paths along specific chromosome
38
Discussion
regions. Hence, for whole genome scans, classical CGH may still be preferable due to
genome coverage, cost, and technical availability.
Although DNA amplification is known to result in overexpression of specific genes, only
recently the impact of DNA copy numbers on gene transcript levels has been reported in
large-scale analyses (185,186). In breast cancer, forty to sixty percent of highly amplified
genes are overexpressed (185,186). One study calculated that eleven percent of the highly
overexpressed genes are amplified (185). This study further highlighted that there may be
several distinct amplicons within certain chromosome arms, which they claimed to be at least
five in the case of chromosome arm 17q in breast cancer (185).
The TGCT transcriptome
From our CGH study (Paper III), it is evident that there are gains and losses of genetic
material in virtually all TGCTs. The vast majority of these DNA copy number changes are
non-random, meaning that the same genomic region has a similar alteration in a substantial
fraction of the TGCTs. The non-randomness of such aberrations most likely reflects a
consequent selective advantage for the cells harbouring them. There have been several studies
investigating the relevance of genomic amplifications to gene transcription (185-190), but few
have been searching for potential target genes of low-level copy number changes, which is a
much more common genetic change in TGCTs and cancer in general. Two independent
studies have investigated the genome-wide associations between DNA copy numbers and
gene expression, and concluded that even low-level copy number changes lead to altered
expression of many genes (185,186). Except for the amplification of 12p, most chromosome
copy number changes in TGCT are rather low-level. One of the most frequent such copy
number changes in TGCT is the gained region on chromosome 17 (Paper III; Figure 10A),
and hence, a detailed transcriptional profiling of this chromosome region was of interest
(Paper IV).
The cDNA microarray study revealed several overexpressed genes throughout the whole
chromosome 17, but GRB7 and JUP, both located within the commonly gained region on
17q, were the two genes with highest average and median expressions among the tumour
samples (Figure 10B). We further analysed GRB7 and JUP by immunohistochemistry on
tissue microarrays, and confirmed their overexpression also on the protein levels (Paper V;
Figure 10C).
39
Figure 10. GRB7 and JUP are overexpressed genes within the commonly gained chromosome
arm 17q. (A) Even though parts of chromosome 17 are overrepresented in every second TGCT
(Paper III; n=33), (B) the chromosome harbours several both up- and downregulated genes, compared
to normal testis, of which GRB7 and JUP were the two most highly upregulated ones (Paper IV). (C)
GRB7 and JUP were validated as upregulated also at their protein levels and the frequencies of
positively stained tissue cores in the tissue microarray are shown (Paper V). Abbreviations: C,
carcinoma in situ; Cc, choriocarcinoma; E, embryonal carcinoma; N, normal testis; S, seminoma; T,
teratoma; Y, yolk sac tumour.
Even though GRB7 and JUP are overexpressed in the majority of TGCTs, there are most
likely additional genes on chromosome arm 17q that are highly and frequently overexpressed.
The applied cDNA microarray contained 636 genes and ESTs located at chromosome 17.
This includes all 201 known genes at the time of construction, as well as 435 ESTs from the
chromosome arm 17q. In the current version of the human genome, as available from the
NCBI web site11, there are 1475 UniGenes on chromosome 17 of which 1041 are located on
the long arm.
Small-scale gene expression studies of TGCTs have indicated overexpression of the protooncogenes MYCN (164) and CCND2 (165-169). Interestingly, several E-boxes (the common
DNA binding site of the MYC family proteins) are found in the promoter region of CCND2.
Additionally, MYC overexpression has been shown to induce chromosomal and
extrachromosomal instability of the CCND2 gene at 12p13 (191). Thus, its tempting to
speculate whether there is a causative link between the overexpression of MYCN and 12p-
11
NCBI Build 30 of the human genome; http://www.ncbi.nlm.nih.gov/genome/guide/human/
40
Discussion
amplification, both phenomena seen in virtually all TGCTs. Interestingly, studies of
neuroblastomas also give evidences for both statistical and structural associations between
MYCN amplification and gain of 17q (192,193). It is known that both these parameters are
commonly present also in TGCT, but there has not been investigated whether there is any
association between them. The part of the cDNA microarray analysis (Paper IV) that
investigated the 512 genes located at other chromosomes than 17, confirmed the
overexpression of the CCND2 and MYCN proto-oncogenes, and revealed several other genes
novel to testicular tumourigenesis as consistently up- or downregulated, including
upregulation of POV1, MYCL1, and MYBL2, and downregulation of DNMT2, MXI1, and
TIMP2.
In addition to the known genes mentioned above, several ESTs were also transcriptionally
deregulated. The putative cancer-related functions of all the known genes mentioned above
and the confirmation of MYCN and CCND2 as overexpressed in TGCT suggest that the
applied cDNA microarrays are sensitive and specific enough to discover oncogenic gene
expression changes in TGCT. Thus, the consistently overexpressed ESTs may also reflect
genes playing important roles in TGCT oncogenesis.
Three of the overexpressed genes from our gene expression analysis (Paper IV) were
validated by real time reverse transcription PCR (RT-PCR). For endogenous control, we used
GAPDH. It has been claimed that this gene should not be used for endogenous control in
TGCT as it is located on chromosome arm 12p. However, in our cDNA microarray analyses
(Paper IV), GAPDH was not overexpressed in the TGCTs. The same was observed by a study
testing several routinely used endogenous control genes where GAPDH had the lowest
variability within testicular cancer and normal adjacent tissues (194). Hence, transcription of
the GAPDH housekeeping gene must be regulated by subtle mechanisms in testicular tissues,
which are not affected by copy numbers.
Even though seminomas and nonseminomas are morphologically quite distinct, they have
many of the same regional genomic disruptions, although frequencies may vary (Paper III).
However, by hierarchical clustering analysis of the gene expression data we demonstrated that
there are individual transcriptional patterns inherent in the two histological subtypes (Paper
IV). This can not be explained in terms of DNA copy number alterations, but differential
DNA methylation patterns could be one additional and most likely explanation as seminomas
and nonseminomas have quite distinct epigenomes (161).
41
Recently, there was published one additional gene expression study of TGCTs which has
taken advantage of microarray technology (190). This study focused onto chromosome arm
12p and analysed DNA copy numbers and gene expression from five and four GCTs. A
cDNA microarray with 8254 spotted ESTs was applied, of which 118 were assigned to 12p
(estimated to represent 28% of all 12p genes12). Nineteen of these were detected as amplified
in at least four of the five tested tumours, and the study then remarkably proceeded with only
the 13 of those 19 amplified ESTs which map to the chromosome region 12p11-12. This is
the region identified by classical CGH as having the highest gain in a few tumours (117-119),
and hence, they disregarded the advantage they could have made out of the much higher
resolution of array-CGH compared to the classical CGH. The expression part of the study
only investigated the 13 amplified genes on 12p11-12. This identified two novel genes, GCT1
and GCT2, as both amplified and overexpressed. The expression levels of the remaining 8241
ESTs were not analysed.
Principally, two different methods are frequently used for large-scale or genome-wide mRNA
expression studies: serial analysis of gene expression (SAGE) and DNA microarrays. By
SAGE (195), it is possible to identify and quantify transcripts on the basis of sequencing.
Short, usually 15-bp sequence tags are isolated from a defined restriction enzyme site near the
3’-end of the cDNAs. These tags contain sufficient sequence information to identify the
transcript from which each tag was derived. The 15-bp tags are concatenated, PCR amplified,
cloned, and sequenced. The abundance of a transcript is estimated by counting the occurrence
of each SAGE tag13. By DNA microarrays, complex nucleic-acid samples are investigated by
hybridisation onto two main types of arrays: in situ synthesised oligonucleotide microarray
(24) and spotted DNA microarrays (25). Oligonucleotide microarrays may contain hundreds
of thousands of ordered, single-stranded synthetic oligonucleotides that are typically 25 bases
in length. Often, each gene is probed by several oligonucleotides. DNA and cDNA samples
are labelled and fragmented before being hybridised to the array. Quantitative estimates of the
transcript abundances can be obtained directly by averaging the signal from all the probes
belonging to one gene. Spotted DNA microarrays usually contain ordered, double-stranded
DNA created by PCR. They correspond to either genomic (BAC- or PAC-microarrays; ref.
12
According to NCBI Build 30 of the human genome; http://www.ncbi.nlm.nih.gov/genome/guide/human/
13
Public SAGE data are available through the NCBI web-site; http://www.ncbi.nlm.nih.gov/SAGE/
42
Discussion
183) or cDNA (cDNA microarray; refs.25,196; used in Paper IV) sequences that have been
spotted onto glass slides. Usually each gene/sequence is represented by one probe. Samples of
mRNA from two sources, often denoted test and reference, are labelled with different
fluorescent dyes, pooled and cohybridised onto the microarray, and quantitative estimates are
based on the dye-ratios.
The SAGE method benefits from not relying on a priori gene predictions and on detecting
absolute expression levels, but is hampered by the rather laborious protocol. DNA
microarrays, on the other hand, rely on known sequences or genetic elements and detect
relative expression levels, but are substantially gaining from their high-throughput. The
oligonucleotide type in common use is mainly the commercial GeneChip® arrays from
Affymetrix (Santa Clara, CA, USA). When a defined project is desired, interesting in
expression of genes within certain chromosome regions (e.g. Paper IV), signalling pathways,
or other selections of specific transcripts, the flexibility of spotted DNA microarrays comes to
its right, where a customary set of genes/clones may be spotted onto the array.
Translational genomics by using tissue microarrays
Advances in genomics and proteomics are dramatically increasing the need to evaluate large
numbers of molecular targets for their diagnostic, predictive or prognostic value in clinical
oncology. While DNA microarrays make it possible to analyse the mRNA expression of
thousands of genes simultaneously, the validation of genes emerging from genome screening
analyses in large series of clinical tumours has become a bottleneck in cancer research.
Analysis of tumour markers has traditionally been accomplished by testing one marker at a
time, starting from a relatively small sample size, as was the case for all studies identifying
the candidate TGCT markers analysed in Paper V. Conventional assays for molecular
pathology are often tedious, and require a lot of tissue, thereby limiting both the number of
tissues and the number of targets that can be evaluated. To evaluate such putative tumour
markers thoroughly, large-scale studies of hundreds of tissue specimens with clinical followup information have to be carried out to demonstrate the significance of the markers.
43
Figure 11. The tissue microarray technology. Tissue cores of 0.6mm in diameter are transferred
from hundreds of donor archival tissue blocks and arrayed into a single recipient tissue microarray
block. Sections from this give rise to tissue microarrays which can be utilised in assays like DNA
fluorescence in situ hybridisation, RNA in situ hybridisation and protein immunohistochemistry. An
example of immunohistochemistry with antibodies against ALPPL2/ALPP is shown.
The tissue microarray technology utilises microscope slides containing hundreds of precisely
arrayed tissue specimens (Figure 11) and has the potential to significantly accelerate studies
seeking for associations between molecular changes and demographic, biological,
pathological, and clinical information (197,198). Tissue microarrays make it possible to study
expression of molecular targets within a large sample set on a single microscope slide, either
at the DNA, RNA, or protein level, by various in situ assays.
In comparison to conventional techniques, tissue microarrays provide a number of
advantages, including preservation of precious tissues, and a substantial time and cost saving
during analysis of molecular targets. The ability to study archival tissue specimens by tissue
microarrays is valuable as such specimens are usually not applicable to molecular genomic or
proteomic surveys. Thousands of archival specimens make up the main bio-banks available
and these are suitable for large retrospective studies, in particular since they easily can be
linked with clinical and follow-up data.
A combination of DNA microarrays with subsequent analyses of candidate genes on a tissue
microarray of large series of clinically well-characterised tumour samples will allow for
identification and validation of new tumour markers for differential diagnosis, prediction, and
prognosis (197,198). The TGCT tissue microarray was constructed to contain all histological
subgroups and clinical stages. A set of relational databases were made for storage of the tissue
microarray data, including information on clinics, histology, research data from other studies,
44
Discussion
archival block numbers, array positions, and the immunohistochemical staining data obtained
with the tissue microarray.
Recently, there was published a set of software tools14 for analysis and archiving of
immunohistochemistry staining data obtained with tissue microarrays (199). This report
demonstrated the applicability of hierarchical clustering analysis to immunohistochemical
data (example from our TGCT data set is shown in Figure 12).
Figure 12. Hierarchical clustering of testicular tissue samples and immunohistochemistry data
from the TGCT tissue microarray. Data from seven antibodies (red bars in the dendrogram; Paper V
and unpublished) were hierarchically clustered. Data regarding histology, year of orchiectomy, and
clinical stage (blue bars) were also included to illustrate correlations with the staining data, but the
genes were clustered independently of these three variables. Text to the right identifies each tissue
core, and the ones coloured red illustrate tissue cores from the same tumours that have clustered
together. Abbreviations: Cc, choriocarcinoma; CIS, carcinoma in situ; EC, embryonal carcinoma; IHC,
immunohistochemistry; N, normal testicular tissue; Orch. year, orchiectomy year; Sem, seminoma;
Ter, teratoma; TGCT, testicular germ cell tumour; YST, yolk sac tumour. For full gene/protein names,
see Appendix II.
14
Freely available from the Stanford tissue microarray website; http://genome-www.stanford.edu/TMA/
45
Although several genes and proteins are known to be up- or downregulated in major or minor
subsets of TGCTs, no new protein tumour markers have yet displayed the level of usefulness
as CGB (E-HCG) and AFP. These two are well-established TGCT markers that can be
detected in the patients’ sera. Additionally, ALPP/ALPPL2 (PLAP) and KIT are well
established diagnostic markers for CIS (68).
Immunohistochemistry using an antibody against ALPP/ALPPL2 was in Paper V used to
discriminate between normal spermatogenic germ cells and CIS. This marker is also present
in many TGCTs, but only small-scale data sets of immunohistochemistry of PLAP in TGCT
have been published. From our tissue microarray, the frequencies of tissue cores positive for
PLAP staining were 89% for the seminomas (n=171), 65% for the embryonal carcinomas
(n=85), 63% for the choriocarcinomas (n=8), 25% for the yolk sac tumours (n=61), and 9%
for the teratomas (n=58; unpublished data). These data demonstrate that expression of ALPP
and/or ALPPL2 is generally turned off as TGCT cells differentiate.
TGCT candidate genes and their cellular context
A list of candidate genes to TGCT development is listed in Appendix II and their loci are
shown in Figure 13. The molecular roles of some of these may be linked together in cellular
pathways, of which the “RAS-, RB-, and WNT-signalling pathways” are discussed in the
following.
The RAS-pathway. Even though KRAS2 is located within a proposed smallest amplified
region on 12p (115,119), it does not have a highly increased expression (Paper IV and ref.
166). High levels of the KRAS2 protein would anyway not be sufficient to effectuate
downstream signalling, as the RAS proteins (HRAS, KRAS2, and NRAS) need to be in an
activated state to pass on cellular signals. Nevertheless, mutations leading to constitutively
activated RAS proteins of both KRAS2 and its homologue NRAS have been detected in
TGCT, but different studies have reported discrepant mutational frequencies, ranging from
none to sixty-five percent (200-204). In a study reviewing the literature on RAS mutations in
TGCT, and analysing an additional series of tumours, the total incidence frequency were
calculated to 11% (204). Although this is not a high mutation frequency, RAS mutations are
significantly present in TGCTs, indicating that the extended RAS-pathway may be of
46
Discussion
Figure 13. The TGCT genome. Red or green chromosome arms have increased or decreased copy
numbers in more than 30% of the TGCTs analysed by comparative genomic hybridisation (CGH).
Note that we have used the opposite colour code to classical CGH to adapt the data to the colour
code used in array-CGH and expression microarrays. The CGH data is based on all studies of a least
two primary TGCTs, reporting the genome-wide copy numbers, and that displays the actual
frequencies (n=122; refs. 116-118,205-207, and Paper III). Commonly overexpressed genes are
written in red, genes with loss of expression in green, and genes that may be mutated or present rare
alleles in blue. Regions with suggestive linkage to human TGCT are indicated by brown bars to the left
of the chromosomes, whereas regions syntenic to susceptibility loci in mice are indicated by yellow
bars. For full gene names, see Appendix II.
importance (Figure 14A). RAS modulates signals from transmembrane tyrosine kinase
growth factor receptors via adaptor proteins, and passes signals further to several oncogenic
RAS-effector pathways (208). The overexpressed GRB7 (Papers IV & V) encodes such an
adaptor protein that contains a RAS-associating like domain (209) and interacts with the
cytoplasmic domain of several activated growth factor receptors, including ERBB2, INSR,
KIT, PDGFRA, and PDGFRB (210-213). Among these, at least KIT and PDGFRA are
expressed at high levels in TGCT (214-218). There are both a close homology and a close
47
Figure 14. RAS-, RB-, and WNT-signalling pathways, simplified. (A) When a growth factor (GF)
binds to a transmembrane growth factor receptor (GFR), the receptor homo-dimerises, and the two
monomers phosphorylate (P) each other on tyrosine residues. The phospho-tyrosine is recognised by
an adaptor protein that again transfers the signal to RAS. Activated RAS then effectuates the cellular
signal to various downstream pathways, many of which act oncogenic. (B) When the levels of D type
cyclins (D1-3) increase in the G1 phase of the cell cycle, they complex with the ubiquitously expressed
cyclin dependent kinases 4 and 6 (CDK4/6), which subsequently may phosphorylate RB1. Native RB1
sequesters E2F transcription factors, but when RB1 is phosphorylated by the cyclin D-CDK4/6
complex, E2F is released and induces cell cycle entry by promoting transcription of various cell cycle
relevant genes. The RB-pathway may be negatively regulated by CDK-inhibitors (CDKI) like the
CDKN2A, -B, -C, and -D. (C) WNT proteins may bind to transmembrane Frizzled receptors, and
through several cytoplasmic relay components, the signal is transduced to E-andJ-catenin (CTNNB1
and JUP), which then are stabilised in the cytoplasm. CTNNB1 and JUP may upon cytoplasmic
accumulation enter the nucleus and complex with TCF/LEF to activate transcription of TCF/LEF
responsive genes, or they may interact with E-cadherin (CDH1) and influence on cell adhesion.
localisation in the genome between PDGFRA and KIT (0.36Mbp in-between15), but there is
no correlation between their expression in TGCT (215). Noteworthy, PDGFRA transcribes
from an alternative promoter in CIS and TGCT which results in a shorter transcript with yet
no assigned function (214,215,219). This alternative promoter use led to the erroneous
scoring of under-expressed PDGFRA in CIS and TGCT by our cDNA microarray analysis,
where PDGFRA was represented by a clone from outside the small transcript (Figure 1 in
Paper IV). Summarised, we forward the hypothesis that most TGCTs have activated
oncogenic RAS effector pathways, either caused by mutated RAS or by high expression of
one or several of its up-stream regulators.
The RB-pathway. Upregulation of D-type cyclins and hence G1-S phase transition in the cell
cycle is a downstream event of the both the RAS effector pathway that signals through the
MAP kinase cascade and the one signalling through PI3K and AKT. This brings us to the G1-
15
June 2002 assembly of the UCSC Genome Browser; http://genome.ucsc.edu/
48
Discussion
S regulating “RB-pathway” (Figure 14B) which is extensively distorted in TGCTs.
Expression of the RB1 tumour suppressor is turned off in virtually all CIS, seminomas, and
embryonal carcinomas, but reexpressed in teratomas, indicating that the gene is not
completely lost, but most likely downregulated at the transcriptional level (220). The RB1
protein sequesters the E2F transcription factors, but when RB1 is phosphorylated by the
cyclin D-CDK4/6 complex, the E2Fs are released, and the cell is committed to cycling.
Knock-out models in rats suggest that inactive RB1 can be compensated by the two RB1 like
proteins RBL1 and RBL2 (p107 and p130; ref. 221). Thus, also other genes/proteins in the
RB-pathway may have to be altered to facilitate G1-S transition.
In TGCTs, CCND2 is frequently both amplified at the DNA level, and overexpressed at the
RNA and protein levels (165-169, and Papers IV and V). The first evidence of CCND2 to be
involved in development of TGCTs came from CCND2 knockout mice (165). Later, mRNA
overexpression of CCND2 were found in a panel of TGCT cell lines (166). CCND2 has
previously been found upregulated in 69% (n=45) of primary TGCTs on the mRNA level
(168). By the tissue microarray analyses we found a comparable frequency (56%; n=278) on
the protein level, and also confirm the lack of association between CCND2 expression and
histological subtype (Paper V). CCND2 mRNA expression has in TGCT been shown to
correlate with its protein binding partner CDK4 (168). Additionally, we have shown that
TGCTs expressing CCND2 more often express GRB7, MGMT, and NKX3A (Paper V and
unpublished data). Another immunohistochemical study analysing 31 TGCTs for expression
of CCND2 (167) found generally higher frequencies of positivity than what we found on the
tissue microarray (Paper V), and they also demonstrated no direct correlation between
CCND2 and Ki67 positivity, meaning that CCND2 expression does not merely reflect the
proliferation status (167). The tissue microarray analysis of CCND2 expression (Paper V) was
however the first to include a large enough sample set to conclude both on presence and
absence of clinical associations.
CCND2 is located within the chromosome band 12p13 which in some TGCTs is the part of
12p with highest amplification (107,120,121, and Paper III). Amplification of 12p is usually
seen only in invasive TGCT (114-116), but has also been detected in a few CIS (111-113).
CCND2 expression is also more often seen in TGCT than in CIS (56% vs. 16%; Paper V).
But because virtually all TGCTs have increased DNA copy numbers of CCND2, it is
uncertain whether amplification of the gene is neither necessary nor sufficient for CCND2
49
expression. However, there are several alternative explanations to induced CCND2 expression
in TGCT, as CCND2 is a down-stream target of both the RAS- and WNT-signalling pathways
(discussed above and below in this section), and also may be induced by MYCN (191), which
is commonly overexpressed in TGCT (ref. 164 and Paper IV).
The cyclin D-CDK4/6 complex may be negatively regulated by a set of CDK-inhibitors.
Several of these, namely CDKN2A, CDKN2C, and CDKN2D, are frequently inactivated in
TGCT (64,162,222,223), and thus add to the effect of overexpressed CCND2 and absent RB1
for deregulation of the RB-pathway in TGCT. The Cyclin D-CDK4/6 complex may also be
negatively regulated by TP53. This tumour suppressor gene is rarely mutated but is still
expressed at high levels in TGCT (100,101,131,139-148).
The WNT-signalling pathway. The WNT-family is highly conserved, and encodes secreted
signaling molecules that regulate cell-to-cell interactions during embryogenesis. Deregulated
WNT-signalling has been implicated in cancer. In cells not exposed to WNT-signal, GSK3B,
AXIN1, and APC will complex and phosphorylate E-catenin (CTNNB1) or J-catenin (JUP),
and ubiquitin mediated proteasome degradation takes place. With WNT-signal, the
GSK3B/AXIN1/APC complex is inhibited, CTNNB1/JUP remain unphosphorylated,
accumulate in the cytoplasm, and may function oncogenic by several means (Figure 14C;
refs. 224,225). One way is by activation of the TCF/LEF transcription factor family,
stimulating the expression of downstream target genes like MYC, CCND1,16 and possibly of
CCND2 (226).
Through our cDNA and tissue microarray studies (Papers IV and V) we have demonstrated
that JUP is expressed at high levels in CIS and TGCT compared to normal spermatogenic
germ cells. However, because we have not investigated PGCs or gonocytes, we can not
conclude that JUP is ectopically overexpressed in CIS and TGCT, or whether the expression
has stayed on high since foetal life. Anyway, a high expression of JUP has been shown to
transform cells in vitro, in contrast to CTNNB1 for which mutation is required (224). Thus,
highly expressed JUP may activate the WNT-signalling pathway and promote testicular
tumourigenesis.
16
Roel Nusse’s WNT-signalling home page, and many references therein; http://www.stanford.edu/~rnusse/
wntwindow.html
50
Discussion
Most colorectal cancers have either activated CTNNB1 due to mutations (227) or inactivated
APC caused by truncating mutations, LOH, or promoter hypermethylation (228). Mutations in
APC have in vitro been shown to cause accumulation of JUP and CTNNB1 (224) and is
suggested to cause chromosomal instability (229). Because TGCTs are generally
chromosomally instable and overexpress JUP, and because APC is located within a region
which both has frequent AI (134,136, and Paper II) and is suggestively linked to hereditary
TGCT (82), it seems interesting to analyse the APC and the WNT-signalling pathway in
TGCT. A direct investigation of known TCF/LEF responsive genes may not be conclusive on
the role of the WNT-signalling pathway in TGCT since it is debated whether JUP and
CTNN1B activate the same set of TCF/LEF responsive genes (225).
JUP and CTNNB1 may also bind to E-cadherin (CDH1), a cell adhesion protein acting as a
suppressor of invasiveness (230). Interestingly, CDH1 has induced expression in
nonseminomas (231) and is lower expressed in embryonal carcinomas of stage II compared to
stage I (232).
WNT5B, WNT8A, and WNT14 expression have been detected in differentiated TGCT cell
lines (233-235). WNT5B is located within chromosome band 12p13, and one can hypothesise
that 12p-amplification may lead to WNT5B overexpression and subsequent downstream
activation of the WNT-signalling pathway.
PTEN is a tumour suppressor essential for embryonic development, and has also been
implicated in the WNT-signalling pathway (236). Pten+/- mice spontaneously develop GCTs
(236), and interestingly, there has been published an abstract of a PTEN germ line mutation in
an adolescent with synchronous testicular and extragonadal GCT (237).
51
CONCLUSIONS
Several molecular changes were identified or characterised by the work of this thesis,
investigating the TGCT genome by a comprehensive functional genomics approach. First, we
analysed for genome-wide copy number changes, delineating chromosome arm 17q as
frequently overrepresented in TGCTs. There next, we focused into that genome region by
gene expression profiling using cDNA microarrays, and identified JUP and GRB7 among
several other genes as overexpressed in TGCT. Finally, we constructed a tissue microarray of
hundreds of testicular samples by which the protein levels of selected candidate genes were
evaluated. Additionally, we have shown that hereditary and sporadic TGCT have strikingly
similar genetic complements. The timing of various common genetic changes is in Figure 15
summarised in a model of the testicular tumourigenesis.
Figure 15. Genetic model of testicular germ cell tumourigenesis. Some of the molecular changes
have not been investigated for all developmental stages (e.g. the foetal stages), and for instance the
changes noted at the transition into foetal carcinoma in situ (CIS) may thus also be present already in
the primordial germ cell (PGC), or may first happen in post-pubertal CIS. Some of the genes noted as
induced in TGCTs were identified as overexpressed compared to normal adult testicular tissue, and
we can therefore not rule out their expression in PGCs. Alterations investigated in the current thesis
are coloured red. Other abbreviations: Cc, choriocarcinoma; EC, embryonal carcinoma; Sem,
seminoma; Ter, teratoma; YST, yolk sac tumour. For full gene/protein names, see Appendix II.
53
REFERENCES
1 Boveri TH (1914). Zur Frage der
Entstehung maligner Tumoren. Verlag von
Gustav Fisher, Jena.
2 Nowell PC (1976). The clonal evolution of
tumor cell populations. Science 194(4260):
23-28.
3 Heim S, Teixeira MR, Dietrich CU, et al.
(1997). Cytogenetic polyclonality in tumors
of the breast. Cancer Genet. Cytogenet.
95(1): 16-19.
4 Hanahan D and Weinberg RA (2000). The
hallmarks of cancer. Cell 100(1): 57-70.
5 Futreal PA, Liu Q, Shattuck-Eidens D, et
al. (1994). BRCA1 mutations in primary
breast and ovarian carcinomas. Science
266(5182): 120-122.
6 Miki Y, Swensen J, Shattuck-Eidens D, et
al. (1994). A strong candidate for the
breast and ovarian cancer susceptibility
gene BRCA1. Science 266(5182): 66-71.
7 Ionov Y, Peinado MA, Malkhosyan S, et al.
(1993). Ubiquitous somatic mutations in
simple repeated sequences reveal a new
mechanism for colonic carcinogenesis.
Nature 363(6429): 558-561.
8 Parsons R, Li GM, Longley MJ, et al.
(1993). Hypermutability and mismatch
repair deficiency in RER+ tumor cells. Cell
75(6): 1227-1236.
9 Thibodeau SN, Bren G, and Schaid D
(1993). Microsatellite instability in cancer of
the proximal colon. Science 260(5109):
816-819.
10 Groden J, Thliveris A, Samowitz W, et al.
(1991). Identification and characterization
of the familial adenomatous polyposis coli
gene. Cell 66(3): 589-600.
11 Kinzler KW, Nilbert MC, Su LK, et al.
(1991). Identification of FAP locus genes
from chromosome 5q21. Science
253(5020): 661-665.
12 Nishisho I, Nakamura Y, Miyoshi Y, et al.
(1991). Mutations of chromosome 5q21
genes in FAP and colorectal cancer
patients. Science 253(5020): 665-669.
13 Baker SJ, Fearon ER, Nigro JM, et al.
(1989). Chromosome 17 deletions and p53
gene mutations in colorectal carcinomas.
Science 244(4901): 217-221.
14 Friend SH, Bernards R, Rogelj S, et al.
(1986). A human DNA segment with
properties of the gene that predisposes to
retinoblastoma and osteosarcoma. Nature
323(6089): 643-646.
15 Cavenee WK, Dryja TP, Phillips RA, et al.
(1983). Expression of recessive alleles by
chromosomal mechanisms in
retinoblastoma. Nature 305(5937): 779784.
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Shih C, Shilo BZ, Goldfarb MP, et al.
(1979). Passage of phenotypes of
chemically transformed cells via
transfection of DNA and chromatin. Proc.
Natl. Acad. Sci. U. S. A 76(11): 5714-5718.
Stehelin D, Varmus HE, Bishop JM, et al.
(1976). DNA related to the transforming
gene(s) of avian sarcoma viruses is
present in normal avian DNA. Nature
260(5547): 170-173.
Knudson AG, Jr. (1971). Mutation and
cancer: statistical study of retinoblastoma.
Proc. Natl. Acad. Sci. USA 68(4): 820-823.
Nowell P and Hungerford D (1960).
Chromosomes of normal and leukemic
human leucocytes. J. Natl. Cancer Inst. 25:
85.
Lander ES, Linton LM, Birren B, et al.
(2001). Initial sequencing and analysis of
the human genome. Nature 409(6822):
860-921.
Venter JC, Adams MD, Myers EW, et al.
(2001). The sequence of the human
genome. Science 291(5507): 1304-1351.
Wilmut I, Schnieke AE, McWhir J, et al.
(1997). Viable offspring derived from fetal
and adult mammalian cells. Nature
385(6619): 810-813.
Schnieke AE, Kind AJ, Ritchie WA, et al.
(1997). Human factor IX transgenic sheep
produced by transfer of nuclei from
transfected fetal fibroblasts. Science
278(5346): 2130-2133.
Pease AC, Solas D, Sullivan EJ, et al.
(1994). Light-generated oligonucleotide
arrays for rapid DNA sequence analysis.
Proc. Natl. Acad. Sci. U. S. A 91(11): 50225026.
Schena M, Shalon D, Davis RW, et al.
(1995). Quantitative monitoring of gene
expression patterns with a complementary
DNA microarray. Science 270(5235): 467470.
NIH/CEPH Collaborative Mapping Group
(1992). A comprehensive genetic linkage
map of the human genome. Science
258(5079): 67-86.
Dib C, Faure S, Fizames C, et al. (1996). A
comprehensive genetic map of the human
genome based on 5,264 microsatellites.
Nature 380(6570): 152-154.
Adams MD, Kelley JM, Gocayne JD, et al.
(1991). Complementary DNA sequencing:
expressed sequence tags and human
genome project. Science 252(5013): 16511656.
Olson M, Hood L, Cantor C, et al. (1989). A
common language for physical mapping of
the human genome. Science 245(4925):
1434-1435.
57
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
58
Beckmann JS and Soller M (1990). Toward
a unified approach to genetic mapping of
eukaryotes based on sequence tagged
microsatellite sites. Biotechnology (N. Y. )
8(10): 930-932.
Saiki RK, Scharf S, Faloona F, et al.
(1985). Enzymatic amplification of betaglobin genomic sequences and restriction
site analysis for diagnosis of sickle cell
anemia. Science 230(4732): 1350-1354.
Mullis KB and Faloona FA (1987). Specific
synthesis of DNA in vitro via a polymerasecatalyzed chain reaction. Methods
Enzymol. 155: 335-350.
Mercola KE, Stang HD, Browne J, et al.
(1980). Insertion of a new gene of viral
origin into bone marrow cells of mice.
Science 208(4447): 1033-1035.
Botstein D, White RL, Skolnick M, et al.
(1980). Construction of a genetic linkage
map in man using restriction fragment
length polymorphisms. Am. J. Hum. Genet.
32(3): 314-331.
Cohen SN, Chang AC, Boyer HW, et al.
(1973). Construction of biologically
functional bacterial plasmids in vitro. Proc.
Natl. Acad. Sci. U. S. A 70(11): 3240-3244.
Kleppe K, Ohtsuka E, Kleppe R, et al.
(1971). Studies on polynucleotides. XCVI.
Repair replications of short synthetic DNA's
as catalyzed by DNA polymerases. J. Mol.
Biol. 56(2): 341-361.
Cold Spring Harbor Symposia on
Quantitative Biology (1966). The genetic
code. Includes papers by Ames BN,
Beckwith JR, Brenner S, Khorana HG,
Morgan R, Nirenberg M, Watson JD,
Woese CR, and Yanofsky C.
Watson JD and Crick FHC (1953).
Molecular structure of nucleic acids. Nature
171: 737-738.
Hershey AD and Chase MC (1952).
Independent functions of viral protein and
nucleic acid in growth of bacteriophage. J.
Gen. Physiol. 36: 39-56.
McClintock B (1944). Maize genetics.
Carnègie Inst. Wash. Year Book 43: 127135.
Beadle GW and Tatum EL (1941). Genetic
control of biochemical reactions in
neurospora. Proc. Natl. Acad. Sci. USA 27:
499-506.
Morgan TH, Sturtevant AH, Muller HJ, et
al. (1915). The mechanisms of Mendelian
heredity. Henry Holt, New York.
Sutton WS (1903). The chromosomes in
heredity. Biol. Bull. 4: 231-251.
Boveri TH (1904). Ergebnisse über die
Konstitution der chromatischen Substanz
des Zellkerns. Gustav Fisher, Jena.
Correns CG (1900). Mendels Regel über
das Erhalten der Nachkommenschaft der
Rassenbastarde. Ber. deutsch. Bot. Ges.
18: 158-168.
De Vries H (1900). Sur la loi de disjonction
des hybrides. Compt. Rend. Ac. Sc. Paris
130: 845-847.
47
Tschermak E (1900). Über künstliche
Kreuzung bei Pisum sativum. Ber. deutsch.
Bot. Ges. 18: 232-239.
48 Flemming W (1882). Zell-substanz, Kern
und Zelltheilung. Vogel, Leipzig.
49 Mendel G (1865). Versuche über Pflanzen
- Hybriden. Verlage des Vereines, Brünn.
50 Loeb LA (2001). A mutator phenotype in
cancer. Cancer Res. 61(8): 3230-3239.
51 Loeb LA (1991). Mutator phenotype may
be required for multistage carcinogenesis.
Cancer Res. 51(12): 3075-3079.
52 Müller HH, Heinimann K, and Dobbie Z
(2000). Genetics of hereditary colon cancer
- a basis for prevention? Eur. J. Cancer
36(10): 1215-1223.
53 Liu L, Lassam NJ, Slingerland JM, et al.
(1995). Germline p16INK4A mutation and
protein dysfunction in a family with
inherited melanoma. Oncogene 11(2): 405412.
54 Zuo L, Weger J, Yang Q, et al. (1996).
Germline mutations in the p16INK4a
binding domain of CDK4 in familial
melanoma. Nat. Genet. 12(1): 97-99.
55 Gruber SB, Ellis NA, Rennert G, et al.
(2002). BLM heterozygosity and the risk of
colorectal cancer. Science 297(5589):
2013.
56 Goss KH, Risinger MA, Kordich JJ, et al.
(2002). Enhanced tumor formation in mice
heterozygous for Blm mutation. Science
297(5589): 2051-2053.
57 Nigg EA (2002). Centrosome aberrations:
cause or consequence of cancer
progression? Nat. Rev. Cancer 2(11): 815825.
58 Ehrlich M (2002). DNA methylation in
cancer: too much, but also too little.
Oncogene 21(35): 5400-5413.
59 Esteller M, Corn PG, Baylin SB, et al.
(2001). A gene hypermethylation profile of
human cancer. Cancer Res. 61(8): 32253229.
60 Esteller M (2002). CpG island
hypermethylation and tumor suppressor
genes: a booming present, a brighter
future. Oncogene 21(35): 5427-5440.
61 Toyota M, Ahuja N, Ohe-Toyota M, et al.
(1999). CpG island methylator phenotype
in colorectal cancer. Proc. Natl. Acad. Sci.
U. S. A 96(15): 8681-8686.
62 Devesa SS, Blot WJ, Stone BJ, et al.
(1995). Recent cancer trends in the United
States. J. Natl. Cancer Inst. 87(3): 175182.
63 Bergström R, Adami HO, Mohner M, et al.
(1996). Increase in testicular cancer
incidence in six European countries: a birth
cohort phenomenon. J. Natl. Cancer Inst.
88(11): 727-733.
64 Bartkova J, Thullberg M, Rajpert-De Meyts
E, et al. (2000). Lack of p19INK4d in
human testicular germ-cell tumours
contrasts with high expression during
normal spermatogenesis. Oncogene
19(36): 4146-4150.
References
65
66
67
68
69
70
71
72
73
74
75
76
77
78
Hofmann MC and Millán JL (1993).
Developmental expression of alkaline
phosphatase genes; reexpression in germ
cell tumours and in vitro immortalized germ
cells. Eur. Urol. 23(1): 38-45.
Jiang LI and Nadeau JH (2001). 129/Sv
mice - a model system for studying germ
cell biology and testicular cancer. Mamm.
Genome 12(2): 89-94.
Skakkebæk NE, Berthelsen JG,
Giwercman A, et al. (1987). Carcinoma-insitu of the testis: possible origin from
gonocytes and precursor of all types of
germ cell tumours except spermatocytoma.
Int. J. Androl. 10(1): 19-28.
Rørth M, Rajpert-De Meyts E, Andersson
L, et al. (2000). Carcinoma in situ in the
testis. Scand. J. Urol. Nephrol. Suppl. 205:
166-186.
Forman D, Pike MC, Davey G, et al.
(1994). Aetiology of testicular cancer:
association with congenital abnormalities,
age at puberty, infertility, and exercise.
United Kingdom Testicular Cancer Study
Group. BMJ 308(6941): 1393-1399.
Toppari J, Larsen JC, Christiansen P, et al.
(1996). Male reproductive health and
environmental xenoestrogens. Environ.
Health Perspect. 104(Suppl.4): 741-803.
Boisen KA, Main KM, Rajpert-De Meyts E,
et al. (2001). Are male reproductive
disorders a common entity? The testicular
dysgenesis syndrome. Ann. N. Y. Acad.
Sci. 948: 90-99.
Skakkebæk NE, Rajpert-De Meyts E, and
Main KM (2001). Testicular dysgenesis
syndrome: an increasingly common
developmental disorder with environmental
aspects. Hum. Reprod. 16(5): 972-978.
Tollerud DJ, Blattner WA, Fraser MC, et al.
(1985). Familial testicular cancer and
urogenital developmental anomalies.
Cancer 55(8): 1849-1854.
Forman D, Oliver RT, Brett AR, et al.
(1992). Familial testicular cancer: a report
of the UK family register, estimation of risk
and an HLA class 1 sib-pair analysis. Br. J.
Cancer 65(2): 255-262.
Goldgar DE, Easton DF, Cannon-Albright
LA, et al. (1994). Systematic populationbased assessment of cancer risk in firstdegree relatives of cancer probands. J.
Natl. Cancer Inst. 86(21): 1600-1608.
Heimdal K, Olsson H, Tretli S, et al. (1996).
Familial testicular cancer in Norway and
southern Sweden. Br. J. Cancer 73(7):
964-969.
Dong C and Hemminki K (2001).
Modification of cancer risks in offspring by
sibling and parental cancers from
2,112,616 nuclear families. Int. J. Cancer
92(1): 144-150.
Dong C, Lonnstedt I, and Hemminki K
(2001). Familial testicular cancer and
second primary cancers in testicular
cancer patients by histological type. Eur. J.
Cancer 37(15): 1878-1885.
79
80
81
82
83
84
85
86
87
88
89
90
91
92
Nicholson PW and Harland SJ (1995).
Inheritance and testicular cancer. Br. J.
Cancer 71(2): 421-426.
Czene K, Lichtenstein P, and Hemminki K
(2002). Environmental and heritable
causes of cancer among 9.6 million
individuals in the Swedish Family-Cancer
Database. Int. J. Cancer 99(2): 260-266.
Heimdal K, Olsson H, Tretli S, et al. (1997).
A segregation analysis of testicular cancer
based on Norwegian and Swedish families.
Br. J. Cancer 75(7): 1084-1087.
International Testicular Cancer Linkage
Consortium (1998). Candidate regions for
testicular cancer susceptibility genes.
APMIS 106(1): 64-70.
Rapley EA, Crockford GP, Teare D, et al.
(2000). Localization to Xq27 of a
susceptibility gene for testicular germ-cell
tumours. Nat. Genet. 24(2): 197-200.
Matin A, Collin GB, Asada Y, et al. (1999).
Susceptibility to testicular germ-cell
tumours in a 129.MOLF-Chr 19
chromosome substitution strain. Nat.
Genet. 23(2): 237-240.
Muller AJ, Teresky AK, and Levine AJ
(2000). A male germ cell tumorsusceptibility-determining locus, pgct1,
identified on murine chromosome 13. Proc.
Natl. Acad. Sci. USA 97(15): 8421-8426.
Asada Y, Varnum DS, Frankel WN, et al.
(1994). A mutation in the Ter gene causing
increased susceptibility to testicular
teratomas maps to mouse chromosome
18. Nat. Genet. 6(4): 363-368.
Mostofi FK and Sesterhenn IA (1998).
World Health Organization International
Histological Classification of Tumours:
Histological typing of testis tumours. 2nd
edition. Springer-Verlag, Berlin.
Looijenga LH, Gillis AJ, van Putten WL, et
al. (1993). In situ numeric analysis of
centromeric regions of chromosomes 1,
12, and 15 of seminomas,
nonseminomatous germ cell tumors, and
carcinoma in situ of human testis. Lab.
Invest. 68(2): 211-219.
Rajpert-De Meyts E, Kvist M, and
Skakkebæk NE (1996). Heterogeneity of
expression of immunohistochemical
tumour markers in testicular carcinoma in
situ: pathogenetic relevance. Virchows
Arch. 428(3): 133-139.
Johnson DE, Appelt G, Samuels ML, et al.
(1976). Metastases from testicular
carcinoma. Study of 78 autopsied cases.
Urology 8(3): 234-239.
Bredael JJ, Vugrin D, and Whitmore WF,
Jr. (1982). Autopsy findings in 154 patients
with germ cell tumors of the testis. Cancer
50(3): 548-551.
Srigley JR, Mackay B, Toth P, et al. (1988).
The ultrastructure and histogenesis of male
germ neoplasia with emphasis on
seminoma with early carcinomatous
features. Ultrastruct. Pathol. 12(1): 67-86.
59
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
60
Ulbright TM, Amin MB, and Young RH
(1999). Tumors of the testis, adnexa,
spermatic cord, and scrotum. Third series,
Fascicle 25 edition. Armed forces institute
of pathology, Washington, D.C.
Atkin NB and Baker MC (1982). Specific
chromosome change, i(12p), in testicular
tumours? Lancet 2(8311): 1349.
Sandberg AA, Meloni AM, and Suijkerbuijk
RF (1996). Reviews of chromosome
studies in urological tumors. III.
Cytogenetics and genes in testicular
tumors. J. Urol. 155(5): 1531-1556.
Mukherjee AB, Murty VV, Rodriguez E, et
al. (1991). Detection and analysis of origin
of i(12p), a diagnostic marker of human
male germ cell tumors, by fluorescence in
situ hybridization. Genes Chromosomes.
Cancer 3(4): 300-307.
International Germ Cell Cancer
Collaborative Group (1997). International
Germ Cell Consensus Classification: a
prognostic factor-based staging system for
metastatic germ cell cancers. J. Clin.
Oncol. 15(2): 594-603.
Chaganti RS and Houldsworth J (1998).
The cytogenetic theory of the pathogenesis
of human adult male germ cell tumors.
APMIS 106(1): 80-83.
Hendry WF, Barrett A, McElwain TJ, et al.
(1980). The role of surgery in the combined
management of metastases from
malignant teratomas of testis. Br. J. Urol.
52(1): 38-44.
Houldsworth J, Xiao H, Murty VV, et al.
(1998). Human male germ cell tumor
resistance to cisplatin is linked to TP53
gene mutation. Oncogene 16(18): 23452349.
Kersemaekers A-MF, Mayer F, Molier M, et
al. (2002). Role of P53 and MDM2 in
treatment response of human germ cell
tumors. J. Clin. Oncol. 20(6): 1551-1561.
Mayer F, Gillis AJ, Dinjens W, et al. (2002).
Microsatellite instability of germ cell tumors
is associated with resistance to systemic
treatment. Cancer Res. 62(10): 2758-2760.
De Graaff WE, Oosterhuis JW, de Jong B,
et al. (1992). Ploidy of testicular carcinoma
in situ. Lab. Invest. 66(2): 166-168.
Oosterhuis JW, Castedo SM, de Jong B, et
al. (1989). Ploidy of primary germ cell
tumors of the testis. Pathogenetic and
clinical relevance. Lab. Invest. 60(1): 1421.
Fosså SD, Nesland JM, Pettersen EO, et
al. (1991). DNA ploidy in primary testicular
cancer. Br. J. Cancer 64(5): 948-952.
Rodriguez E, Mathew S, Reuter V, et al.
(1992). Cytogenetic analysis of 124
prospectively ascertained male germ cell
tumors. Cancer Res. 52(8): 2285-2291.
Rodriguez E, Houldsworth J, Reuter VE, et
al. (1993). Molecular cytogenetic analysis
of i(12p)-negative human male germ cell
tumors. Genes Chromosomes. Cancer
8(4): 230-236.
108
109
110
111
112
113
114
115
116
117
118
119
Suijkerbuijk RF, Sinke RJ, Meloni AM, et
al. (1993). Overrepresentation of
chromosome 12p sequences and
karyotypic evolution in i(12p)-negative
testicular germ-cell tumors revealed by
fluorescence in situ hybridization. Cancer
Genet. Cytogenet. 70(2): 85-93.
Sinke RJ, Suijkerbuijk RF, de Jong B, et al.
(1993). Uniparental origin of i(12p) in
human germ cell tumors. Genes
Chromosomes. Cancer 6(3): 161-165.
Geurts van Kessel A, van Drunen E, de
Jong B, et al. (1989). Chromosome 12q
heterozygosity is retained in i(12p)-positive
testicular germ cell tumor cells. Cancer
Genet. Cytogenet. 40(1): 129-134.
Vos A, Oosterhuis JW, de Jong B, et al.
(1990). Cytogenetics of carcinoma in situ
of the testis. Cancer Genet. Cytogenet.
46(1): 75-81.
Van Echten J, van Gurp RJ, Stoepker M, et
al. (1995). Cytogenetic evidence that
carcinoma in situ is the precursor lesion for
invasive testicular germ cell tumors.
Cancer Genet. Cytogenet. 85(2): 133-137.
Ottesen AM, Rajpert-De Meyts E, Leffers
H, et al. High resolution CGH detects extra
chromosome 12p material in most cases of
carcinoma in situ (CIS) adjacent to overt
germ cell tumours, but not in CIS prior to
the tumour development. Conference: 5th
Copenhagen workshop on carcinoma in
situ and cancer of the testis, poster
abstract #30, Copenhagen, Denmark,
2002.
Rosenberg C, van Gurp RJ, Geelen E, et
al. (2000). Overrepresentation of the short
arm of chromosome 12 is related to
invasive growth of human testicular
seminomas and nonseminomas.
Oncogene 19(51): 5858-5862.
Roelofs H, Mostert MC, Pompe K, et al.
(2000). Restricted 12p amplification and
RAS mutation in human germ cell tumors
of the adult testis. Am. J. Pathol. 157(4):
1155-1166.
Summersgill B, Osin P, Lu YJ, et al.
(2001). Chromosomal imbalances
associated with carcinoma in situ and
associated testicular germ cell tumours of
adolescents and adults. Br. J. Cancer
85(2): 213-220.
Korn WM, Oide WD, Suijkerbuijk RF, et al.
(1996). Detection of chromosomal DNA
gains and losses in testicular germ cell
tumors by comparative genomic
hybridization. Genes Chromosomes.
Cancer 17(2): 78-87.
Mostert MM, van de Pol M, Olde WD, et al.
(1996). Comparative genomic hybridization
of germ cell tumors of the adult testis:
confirmation of karyotypic findings and
identification of a 12p-amplicon. Cancer
Genet. Cytogenet. 89(2): 146-152.
Mostert MC, Verkerk AJ, van de Pol M, et
al. (1998). Identification of the critical
region of 12p over-representation in
References
120
121
122
123
124
125
126
127
128
129
130
131
132
testicular germ cell tumors of adolescents
and adults. Oncogene 16(20): 2617-2627.
Blough RI, Vance GH, Henegariu O, et al.
(1998). Characterization of multiple 12p
rearrangements in testicular germ cell
tumor cell line 833K and its subclone 64CP
by chromosome microdissection. Cancer
Genet. Cytogenet. 106(1): 24-29.
Henegariu O, Vance GH, Heiber D, et al.
(1998). Triple-color FISH analysis of 12p
amplification in testicular germ-cell tumors
using 12p band-specific painting probes. J.
Mol. Med. 76(9): 648-655.
Mertens F, Johansson B, Höglund M, et al.
(1997). Chromosomal imbalance maps of
malignant solid tumors: a cytogenetic
survey of 3185 neoplasms. Cancer Res.
57(13): 2765-2780.
Looijenga LH, Rosenberg C, van Gurp RJ,
et al. (2000). Comparative genomic
hybridization of microdissected samples
from different stages in the development of
a seminoma and a non-seminoma. J.
Pathol. 191(2): 187-192.
Kallioniemi A, Kallioniemi OP, Sudar D, et
al. (1992). Comparative genomic
hybridization for molecular cytogenetic
analysis of solid tumors. Science
258(5083): 818-821.
Lothe RA, Fosså SD, Stenwig AE, et al.
(1989). Loss of 3p or 11p alleles is
associated with testicular cancer tumors.
Genomics 5(1): 134-138.
Murty VV, Houldsworth J, Baldwin S, et al.
(1992). Allelic deletions in the long arm of
chromosome 12 identify sites of candidate
tumor suppressor genes in male germ cell
tumors. Proc. Natl. Acad. Sci. USA 89(22):
11006-11010.
Lothe RA, Hastie N, Heimdal K, et al.
(1993). Frequent loss of 11p13 and 11p15
loci in male germ cell tumours. Genes
Chromosomes. Cancer 7(2): 96-101.
Looijenga LH, Abraham M, Gillis AJ, et al.
(1994). Testicular germ cell tumors of
adults show deletions of chromosomal
bands 11p13 and 11p15.5, but no
abnormalities within the zinc-finger regions
and exons 2 and 6 of the Wilms' tumor 1
gene. Genes Chromosomes. Cancer 9(3):
153-160.
Murty VV, Bosl GJ, Houldsworth J, et al.
(1994). Allelic loss and somatic
differentiation in human male germ cell
tumors. Oncogene 9(8): 2245-2251.
Murty VV, Li RG, Houldsworth J, et al.
(1994). Frequent allelic deletions and loss
of expression characterize the DCC gene
in male germ cell tumors. Oncogene 9(11):
3227-3231.
Lothe RA, Peltomäki P, Tommerup N, et al.
(1995). Molecular genetic changes in
human male germ cell tumors. Lab. Invest.
73(5): 606-614.
Al-Jehani RM, Povey S, Delhanty JD, et al.
(1995). Loss of heterozygosity on
chromosome arms 5q, 11p, 11q, 13q, and
133
134
135
136
137
138
139
140
141
142
143
144
145
146
16p in human testicular germ cell tumors.
Genes Chromosomes. Cancer 13(4): 249256.
Peng HQ, Bailey D, Bronson D, et al.
(1995). Loss of heterozygosity of tumor
suppressor genes in testis cancer. Cancer
Res. 55(13): 2871-2875.
Murty VV, Reuter VE, Bosl GJ, et al.
(1996). Deletion mapping identifies loss of
heterozygosity at 5p15.1-15.2, 5q11 and
5q34-35 in human male germ cell tumors.
Oncogene 12(12): 2719-2723.
Murty VV, Montgomery K, Dutta S, et al.
(1999). A 3-Mb high-resolution BAC/PAC
contig of 12q22 encompassing the 830-kb
consensus minimal deletion in male germ
cell tumors. Genome Res. 9(7): 662-671.
Peng HQ, Liu L, Goss PE, et al. (1999).
Chromosomal deletions occur in restricted
regions of 5q in testicular germ cell cancer.
Oncogene 18(21): 3277-3283.
Faulkner SW, Leigh DA, Oosterhuis JW, et
al. (2000). Allelic losses in carcinoma in
situ and testicular germ cell tumours of
adolescents and adults: evidence
suggestive of the linear progression model.
Br. J. Cancer 83(6): 729-736.
Kraggerud SM, Åman P, Holm R, et al.
(2002). Alterations of the fragile histidine
triad gene, FHIT, and its encoded products
contribute to testicular germ cell
tumorigenesis. Cancer Res. 62(2): 512517.
Bartkova J, Bartek J, Lukas J, et al. (1991).
p53 protein alterations in human testicular
cancer including pre-invasive intratubular
germ-cell neoplasia. Int. J. Cancer 49(2):
196-202.
Heimdal K, Lothe RA, Lystad S, et al.
(1993). No germline TP53 mutations
detected in familial and bilateral testicular
cancer. Genes Chromosomes. Cancer
6(2): 92-97.
Peng HQ, Hogg D, Malkin D, et al. (1993).
Mutations of the p53 gene do not occur in
testis cancer. Cancer Res. 53(15): 35743578.
Fleischhacker M, Strohmeyer T, Imai Y, et
al. (1994). Mutations of the p53 gene are
not detectable in human testicular tumors.
Mod. Pathol. 7(4): 435-439.
Riou G, Barrois M, Prost S, et al. (1995).
The p53 and mdm-2 genes in human
testicular germ-cell tumors. Mol. Carcinog.
12(3): 124-131.
Schenkman NS, Sesterhenn IA,
Washington L, et al. (1995). Increased p53
protein does not correlate to p53 gene
mutations in microdissected human
testicular germ cell tumors. J. Urol. 154(2
Pt 1): 617-621.
Guillou L, Estreicher A, Chaubert P, et al.
(1996). Germ cell tumors of the testis
overexpress wild-type p53. Am. J. Pathol.
149(4): 1221-1228.
Heidenreich A, Schenkman NS,
Sesterhenn IA, et al. (1998).
61
147
148
149
150
151
152
153
154
155
156
157
158
159
160
62
Immunohistochemical and mutational
analysis of the p53 tumour suppressor
gene and the bcl-2 oncogene in primary
testicular germ cell tumours. APMIS
106(1): 90-99.
Ye DW, Zheng J, Qian SX, et al. (1993).
p53 gene mutations in Chinese human
testicular seminoma. J. Urol. 150(3): 884886.
Kuczyk MA, Serth J, Bokemeyer C, et al.
(1996). Alterations of the p53 tumor
suppressor gene in carcinoma in situ of the
testis. Cancer 78(9): 1958-1966.
Peltomäki P, Lothe RA, Aaltonen LA, et al.
(1993). Microsatellite instability is
associated with tumors that characterize
the hereditary non-polyposis colorectal
carcinoma syndrome. Cancer Res. 53(24):
5853-5855.
Murty VV, Li RG, Mathew S, et al. (1994).
Replication error-type genetic instability at
1q42-43 in human male germ cell tumors.
Cancer Res. 54(15): 3983-3985.
Faulkner SW and Friedlander ML (2000).
Microsatellite instability in germ cell tumors
of the testis and ovary. Gynecol. Oncol.
79(1): 38-43.
Marcus VA, Madlensky L, Gryfe R, et al.
(1999). Immunohistochemistry for hMLH1
and hMSH2: a practical test for DNA
mismatch repair-deficient tumors. Am. J.
Surg. Pathol. 23(10): 1248-1255.
McCarrey JR(1993). In: Cell and molecular
biology of the testis (Desjardins C and
Ewing LL, Eds.) Oxford University Press,
New York. 58-89.
Li E (2002). Chromatin modification and
epigenetic reprogramming in mammalian
development. Nat. Rev. Genet. 3(9): 662673.
Monk M (1987). Genomic imprinting.
Memories of mother and father. Nature
328(6127): 203-204.
Szabo PE and Mann JR (1995). Biallelic
expression of imprinted genes in the
mouse germ line: implications for erasure,
establishment, and mechanisms of
genomic imprinting. Genes. Dev. 9(15):
1857-1868.
Van Gurp RJ, Oosterhuis JW, Kalscheuer
V, et al. (1994). Biallelic expression of the
H19 and IGF2 genes in human testicular
germ cell tumors. J. Natl. Cancer Inst.
86(14): 1070-1075.
Rachmilewitz J, Elkin M, Looijenga LH, et
al. (1996). Characterization of the
imprinted IPW gene: allelic expression in
normal and tumorigenic human tissues.
Oncogene 13(8): 1687-1692.
Nonomura N, Miki T, Nishimura K, et al.
(1997). Altered imprinting of the H19 and
insulin-like growth factor II genes in
testicular tumors. J. Urol. 157(5): 19771979.
Costello JF, Fruhwald MC, Smiraglia DJ, et
al. (2000). Aberrant CpG-island
methylation has non-random and tumour-
161
162
163
164
165
166
167
168
169
170
171
172
173
type-specific patterns. Nat. Genet. 24(2):
132-138.
Smiraglia DJ, Szymanska J, Kraggerud
SM, et al. (2002). Distinct epigenetic
phenotypes in seminomatous and
nonseminomatous testicular germ cell
tumors. Oncogene 21(24): 3909-3916.
Chaubert P, Guillou L, Kurt AM, et al.
(1997). Frequent p16INK4 (MTS1) gene
inactivation in testicular germ cell tumors.
Am. J. Pathol. 151(3): 859-865.
Smith-Sørensen B, Lind GE, Skotheim RI,
et al. (2002). Frequent promoter
6
hypermethylation of the O -MethylguanineDNA methyltransferase (MGMT) gene in
testicular cancer. Oncogene 21(57): 88788884.
Shuin T, Misaki H, Kubota Y, et al. (1994).
Differential expression of protooncogenes
in human germ cell tumors of the testis.
Cancer 73(6): 1721-1727.
Sicinski P, Donaher JL, Geng Y, et al.
(1996). Cyclin D2 is an FSH-responsive
gene involved in gonadal cell proliferation
and oncogenesis. Nature 384(6608): 470474.
Houldsworth J, Reuter V, Bosl GJ, et al.
(1997). Aberrant expression of cyclin D2 is
an early event in human male germ cell
tumorigenesis. Cell Growth Differ. 8(3):
293-299.
Bartkova J, Rajpert-De Meyts E,
Skakkebæk NE, et al. (1999). D-type
cyclins in adult human testis and testicular
cancer: relation to cell type, proliferation,
differentiation, and malignancy. J. Pathol.
187(5): 573-581.
Schmidt BA, Rose A, Steinhoff C, et al.
(2001). Up-regulation of cyclin-dependent
kinase 4/cyclin D2 expression but downregulation of cyclin-dependent kinase
2/cyclin E in testicular germ cell tumors.
Cancer Res. 61(10): 4214-4221.
Cipriano SC, Chen L, Burns KH, et al.
(2001). Inhibin and p27 interact to regulate
gonadal tumorigenesis. Mol. Endocrinol.
15(6): 985-996.
Han S and Peschel RE (2000). Father-son
testicular tumors: evidence for genetic
anticipation? A case report and review of
the literature. Cancer 88(10): 2319-2325.
Gedde-Dahl TJ, Hannisdal E, Klepp OH, et
al. (1985). Testicular neoplasms occurring
in four brothers. A search for a genetic
predisposition. Cancer 55(9): 2005-2009.
Heimdal K, Andersen TI, Skrede M, et al.
(1995). Association studies of estrogen
receptor polymorphisms in a Norwegian
testicular cancer population. Cancer
Epidemiol. Biomarkers Prev. 4(2): 123-126.
Harries LW, Stubbins MJ, Forman D, et al.
(1997). Identification of genetic
polymorphisms at the glutathione Stransferase Pi locus and association with
susceptibility to bladder, testicular and
prostate cancer. Carcinogenesis 18(4):
641-644.
References
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
Teh BT, Linblad K, Nord B, et al. (1999).
Familial testicular cancer: lack of evidence
for trinucleotide repeat expansions and
association with PKD1 in one family. J.
Med. Genet 36(4): 348-349.
Bianchi NO, Richard SM, Peltomaki P, et
al. (2002). Mosaic AZF deletions and
susceptibility to testicular tumors. Mutat.
Res. 503(1-2): 51-62.
Leahy MG, Tonks S, Moses JH, et al.
(1995). Candidate regions for a testicular
cancer susceptibility gene. Hum. Mol.
Genet. 4(9): 1551-1555.
Kristensen VN, Kelefiotis D, Kristensen T,
et al. (2001). High-throughput methods for
detection of genetic variation.
Biotechniques 30(2): 318-326.
Syvänen AC (2001). Accessing genetic
variation: genotyping single nucleotide
polymorphisms. Nat. Rev. Genet. 2(12):
930-942.
Lindblad-Toh K, Tanenbaum DM, Daly MJ,
et al. (2000). Loss-of-heterozygosity
analysis of small-cell lung carcinomas
using single-nucleotide polymorphism
arrays. Nat. Biotechnol. 18(9): 1001-1005.
Mei R, Galipeau PC, Prass C, et al. (2000).
Genome-wide detection of allelic
imbalance using human SNPs and highdensity DNA arrays. Genome Res. 10(8):
1126-1137.
Skotheim RI and Lothe RA (2003). The
testicular germ cell tumour genome.
APMIS 111(1): 136-151.
Forozan F, Karhu R, Kononen J, et al.
(1997). Genome screening by comparative
genomic hybridization. Trends. Genet.
13(10): 405-409.
Solinas-Toldo S, Lampel S, Stilgenbauer
S, et al. (1997). Matrix-based comparative
genomic hybridization: biochips to screen
for genomic imbalances. Genes
Chromosomes. Cancer 20(4): 399-407.
Pinkel D, Segraves R, Sudar D, et al.
(1998). High resolution analysis of DNA
copy number variation using comparative
genomic hybridization to microarrays. Nat.
Genet. 20(2): 207-211.
Hyman E, Kauraniemi P, Hautaniemi S, et
al. (2002). Impact of DNA Amplification on
Gene Expression Patterns in Breast
Cancer. Cancer Res. 62(21): 6240-6245.
Pollack JR, Sørlie T, Perou CM, et al.
(2002). Microarray analysis reveals a major
direct role of DNA copy number alteration
in the transcriptional program of human
breast tumors. Proc. Natl. Acad. Sci. U. S.
A 99(20): 12963-12968.
Forozan F, Mahlamaki EH, Monni O, et al.
(2000). Comparative genomic hybridization
analysis of 38 breast cancer cell lines: a
basis for interpreting complementary DNA
microarray data. Cancer Res. 60(16):
4519-4525.
Kauraniemi P, Bärlund M, Monni O, et al.
(2001). New amplified and highly
expressed genes discovered in the ERBB2
189
190
191
192
193
194
195
196
197
198
199
200
201
amplicon in breast cancer by cDNA
microarrays. Cancer Res. 61(22): 82358240.
Monni O, Bärlund M, Mousses S, et al.
(2001). Comprehensive copy number and
gene expression profiling of the 17q23
amplicon in human breast cancer. Proc.
Natl. Acad. Sci. USA 98(10): 5711-5716.
Bourdon V, Naef F, Rao PH, et al. (2002).
Genomic and Expression Analysis of the
12p11-p12 Amplicon Using EST Arrays
Identifies Two Novel Amplified and
Overexpressed Genes. Cancer Res.
62(21): 6218-6223.
Mai S, Hanley-Hyde J, Rainey GJ, et al.
(1999). Chromosomal and
extrachromosomal instability of the cyclin
D2 gene is induced by Myc
overexpression. Neoplasia 1(3): 241-252.
O'Neill S, Ekstrom L, Lastowska M, et al.
(2001). MYCN amplification and 17q in
neuroblastoma: evidence for structural
association. Genes Chromosomes. Cancer
30(1): 87-90.
Lastowska M, Nacheva E, McGuckin A, et
al. (1997). Comparative genomic
hybridization study of primary
neuroblastoma tumors. United Kingdom
Children's Cancer Study Group. Genes
Chromosomes. Cancer 18(3): 162-169.
Gerard CJ, Andrejka LM, and Macina RA
(2000). Mitochondrial ATP synthase 6 as
an endogenous control in the quantitative
RT-PCR analysis of clinical cancer
samples. Mol. Diagn. 5(1): 39-46.
Velculescu VE, Zhang L, Vogelstein B, et
al. (1995). Serial analysis of gene
expression. Science 270(5235): 484-487.
Mousses S, Bittner ML, Chen Y, et al.
(2000). Gene expression analysis by cDNA
microarrays. In: Functional Genomics
(Liversey FJ and Hunt SP, Eds.) Oxford
University Press, Oxford. 113-137.
Kononen J, Bubendorf L, Kallioniemi A, et
al. (1998). Tissue microarrays for highthroughput molecular profiling of tumor
specimens. Nat. Med. 4(7): 844-847.
Torhorst J, Bucher C, Kononen J, et al.
(2001). Tissue microarrays for rapid linking
of molecular changes to clinical endpoints.
Am. J. Pathol. 159(6): 2249-2256.
Liu CL, Prapong W, Natkunam Y, et al.
(2002). Software Tools for HighThroughput Analysis and Archiving of
Immunohistochemistry Staining Data
Obtained with Tissue Microarrays. Am. J.
Pathol. 161(5): 1557-1565.
Mulder MP, Keijzer W, Verkerk A, et al.
(1989). Activated ras genes in human
seminoma: evidence for tumor
heterogeneity. Oncogene 4(11): 13451351.
Ganguly S, Murty VV, Samaniego F, et al.
(1990). Detection of preferential NRAS
mutations in human male germ cell tumors
by the polymerase chain reaction. Genes
Chromosomes. Cancer 1(3): 228-232.
63
202
203
204
205
206
207
208
209
210
211
212
213
214
64
Moul JW, Theune SM, and Chang EH
(1992). Detection of RAS mutations in
archival testicular germ cell tumors by
polymerase chain reaction and
oligonucleotide hybridization. Genes
Chromosomes. Cancer 5(2): 109-118.
Ridanpää M, Lothe RA, Önfelt A, et al.
(1993). K-ras oncogene codon 12 point
mutations in testicular cancer. Environ.
Health Perspect. 101(Suppl. 3): 185-187.
Olie RA, Looijenga LH, Boerrigter L, et al.
(1995). N- and KRAS mutations in primary
testicular germ cell tumors: incidence and
possible biological implications. Genes
Chromosomes. Cancer 12(2): 110-116.
Ottesen AM, Kirchhoff M, Rajpert-De
Meyts E, et al. (1997). Detection of
chromosomal aberrations in seminomatous
germ cell tumours using comparative
genomic hybridization. Genes
Chromosomes. Cancer 20(4): 412-418.
Summersgill B, Goker H, Weber-Hall S, et
al. (1998). Molecular cytogenetic analysis
of adult testicular germ cell tumours and
identification of regions of consensus copy
number change. Br. J. Cancer 77(2): 305313.
Rosenberg C, Schut TB, Mostert M, et al.
(1999). Chromosomal gains and losses in
testicular germ cell tumors of adolescents
and adults investigated by a modified
comparative genomic hybridization
approach. Lab. Invest. 79(12): 1447-1451.
Takuwa N and Takuwa Y (2001).
Regulation of cell cycle molecules by the
Ras effector system. Mol. Cell Endocrinol.
177(1-2): 25-33.
Wojcik J, Girault JA, Labesse G, et al.
(1999). Sequence analysis identifies a rasassociating (RA)-like domain in the Ntermini of band 4.1/JEF domains and in the
Grb7/10/14 adapter family. Biochem.
Biophys. Res. Commun. 259(1): 113-120.
Yokote K, Margolis B, Heldin CH, et al.
(1996). Grb7 is a downstream signaling
component of platelet-derived growth
factor alpha- and beta-receptors. J. Biol.
Chem. 271(48): 30942-30949.
Thömmes K, Lennartsson J, Carlberg M, et
al. (1999). Identification of Tyr-703 and
Tyr-936 as the primary association sites for
Grb2 and Grb7 in the c-Kit/stem cell factor
receptor. Biochem. J. 341(Pt 1): 211-216.
Stein D, Wu J, Fuqua SA, et al. (1994).
The SH2 domain protein GRB-7 is coamplified, overexpressed and in a tight
complex with HER2 in breast cancer.
EMBO J. 13(6): 1331-1340.
Kasus-Jacobi A, Bereziat V, Perdereau D,
et al. (2000). Evidence for an interaction
between the insulin receptor and Grb7. A
role for two of its binding domains, PIR and
SH2. Oncogene 19(16): 2052-2059.
Mosselman S, Looijenga LHJ, Gillis AJM,
et al. (1996). Aberrant platelet-derived
growth factor alpha-receptor transcript as a
diagnostic marker for early human germ
215
216
217
218
219
220
221
222
223
224
225
226
227
cell tumors of the adult testis. Proc. Natl.
Acad. Sci. USA 93(7): 2884-2888.
Palumbo C, van Roozendaal K, Gillis AJ,
et al. (2002). Expression of the PDGF
alpha-receptor 1.5 kb transcript, OCT-4,
and c-KIT in human normal and malignant
tissues. Implications for the early diagnosis
of testicular germ cell tumours and for our
understanding of regulatory mechanisms.
J. Pathol. 196(4): 467-477.
Strohmeyer T, Peter S, Hartmann M, et al.
(1991). Expression of the hst-1 and c-kit
protooncogenes in human testicular germ
cell tumors. Cancer Res. 51(7): 1811-1816.
Rajpert-De Meyts E and Skakkebæk NE
(1994). Expression of the c-kit protein
product in carcinoma-in-situ and invasive
testicular germ cell tumours. Int. J. Androl.
17(2): 85-92.
Izquierdo MA, van der Valk P, van Ark-Otte
J, et al. (1995). Differential expression of
the c-kit proto-oncogene in germ cell
tumours. J. Pathol. 177(3): 253-258.
Mosselman S, Claesson-Welsh L,
Kamphuis JS, et al. (1994).
Developmentally regulated expression of
two novel platelet-derived growth factor
alpha-receptor transcripts in human
teratocarcinoma cells. Cancer Res. 54(1):
220-225.
Strohmeyer T, Reissmann P, CordonCardo C, et al. (1991). Correlation between
retinoblastoma gene expression and
differentiation in human testicular tumors.
Proc. Natl. Acad. Sci. USA 88(15): 66626666.
Mulligan G and Jacks T (1998). The
retinoblastoma gene family: cousins with
overlapping interests. Trends. Genet.
14(6): 223-229.
Heidenreich A, Gaddipati JP, Moul JW, et
al. (1998). Molecular analysis of
P16(Ink4)/CDKN2 and P15(INK4B)/MTS2
genes in primary human testicular germ
cell tumors. J. Urol. 159(5): 1725-1730.
Bartkova J, Thullberg M, Rajpert-De Meyts
E, et al. (2000). Cell cycle regulators in
testicular cancer: loss of p18INK4C marks
progression from carcinoma in situ to
invasive germ cell tumours. Int. J. Cancer
85(3): 370-375.
Kolligs FT, Kolligs B, Hajra KM, et al.
(2000). Gamma-catenin is regulated by the
APC tumor suppressor and its oncogenic
activity is distinct from that of beta-catenin.
Genes. Dev. 14(11): 1319-1331.
Barker N and Clevers H (2000). Catenins,
Wnt signaling and cancer. Bioessays
22(11): 961-965.
Megason SG and McMahon AP (2002). A
mitogen gradient of dorsal midline Wnts
organizes growth in the CNS. Development
129: 2087-2098.
Morin PJ, Sparks AB, Korinek V, et al.
(1997). Activation of beta-catenin-Tcf
signaling in colon cancer by mutations in
References
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
beta-catenin or APC. Science 275(5307):
1787-1790.
Esteller M, Sparks A, Toyota M, et al.
(2000). Analysis of adenomatous polyposis
coli promoter hypermethylation in human
cancer. Cancer Res. 60(16): 4366-4371.
Fodde R, Kuipers J, Rosenberg C, et al.
(2001). Mutations in the APC tumour
suppressor gene cause chromosomal
instability. Nat. Cell Biol. 3(4): 433-438.
Van Aken E, De Wever O, Correia da
Rocha AS, et al. (2001). Defective Ecadherin/catenin complexes in human
cancer. Virchows Arch. 439(6): 725-751.
Saito T, Katagiri A, Watanabe R, et al.
(2000). Expression of E-cadherin and
catenins on testis tumor. Urol. Int. 65(3):
140-143.
Heidenreich A, Sesterhenn IA, Mostofi FK,
et al. (1998). Prognostic risk factors that
identify patients with clinical stage I
nonseminomatous germ cell tumors at low
risk and high risk for metastasis. Cancer
83(5): 1002-1011.
Saitoh T and Katoh M (2001). Molecular
cloning and characterization of human
WNT5B on chromosome 12p13.3 region.
Int. J. Oncol. 19(2): 347-351.
Wakeman JA, Walsh J, and Andrews PW
(1998). Human Wnt-13 is developmentally
regulated during the differentiation of
NTERA-2 pluripotent human embryonal
carcinoma cells. Oncogene 17(2): 179-186.
Saitoh T and Katoh M (2001). Molecular
cloning and characterization of human
WNT8A. Int. J. Oncol. 19(1): 123-127.
Di Cristofano A, Pesce B, Cordon-Cardo C,
et al. (1998). Pten is essential for
embryonic development and tumour
suppression. Nat. Genet. 19(4): 348-355.
Psiachou-Leonard E. PTEN mutation in an
adolescent with synchronous - testicular
and retroperitoneal - germ cell tumor and
the Bannayan-Riley-Ruvalcaba syndrome.
Conference: Oncogenomics, poster
abstract C-69, Dublin, Ireland, 2002.
Englert CR, Baibakov GV, and EmmertBuck MR (2000). Layered expression
scanning: rapid molecular profiling of tumor
samples. Cancer Res. 60(6): 1526-1530.
Ziauddin J and Sabatini DM (2001).
Microarrays of cells expressing defined
cDNAs. Nature 411(6833): 107-110.
Mousses S, Kallioniemi A, Kauraniemi P,
et al. (2002). Clinical and functional target
validation using tissue and cell
microarrays. Curr. Opin. Chem. Biol. 6(1):
97-101.
Jacobsen GK and Jacobsen M (1983).
Alpha-fetoprotein (AFP) and human
chorionic gonadotropin (HCG) in testicular
germ cell tumours. A prospective
immunohistochemical study. Acta Pathol.
Microbiol. Immunol. Scand. [APMIS] 91(3):
165-176.
Millán JL and Manes T (1988). Seminomaderived Nagao isozyme is encoded by a
243
244
245
246
247
248
249
250
251
252
253
254
255
germ-cell alkaline phosphatase gene. Proc.
Natl. Acad. Sci. USA 85(9): 3024-3028.
Roelofs H, Manes T, Janszen T, et al.
(1999). Heterogeneity in alkaline
phosphatase isozyme expression in human
testicular germ cell tumours: An enzyme/immunohistochemical and molecular
analysis. J. Pathol. 189(2): 236-244.
Millán JL, Eriksson A, and Stigbrand T
(1982). A possible new locus of alkaline
phosphatase expressed in human testis.
Hum. Genet. 62(4): 293-295.
Satie AP, Rajpert-De Meyts E, Spagnoli
GC, et al. (2002). The cancer-testis gene,
NY-ESO-1, is expressed in normal fetal
and adult testes and in spermatocytic
seminomas and testicular carcinoma in
situ. Lab. Invest. 82(6): 775-780.
Zafarana G, Gillis AJ, van Gurp RJ, et al.
(2002). Coamplification of DAD-R, SOX5,
and EKI1 in human testicular seminomas,
with specific overexpression of DAD-R,
correlates with reduced levels of apoptosis
and earlier clinical manifestation. Cancer
Res. 62(6): 1822-1831.
Strohmeyer D, Langenhof S, Ackermann
R, et al. (1997). Analysis of the DCC tumor
suppressor gene in testicular germ cell
tumors: mutations and loss of expression.
J. Urol. 157(5): 1973-1976.
Zeeman A-M, Stoop H, Boter M, et al.
(2002). VASA is a specific marker for both
normal and malignant human germ cells.
Lab. Invest. 82(2): 159-166.
Eyzaguirre E and Gatalica Z (2002). Loss
of fhit expression in testicular germ cell
tumors and intratubular germ cell
neoplasia. Mod. Pathol. 15(10): 10681072.
Katagiri A, Tomita Y, Nishiyama T, et al.
(1993). Immunohistochemical detection of
P-glycoprotein and GSTP1-1 in testis
cancer. Br. J. Cancer 68(1): 125-129.
Ryberg D, Heimdal K, Fosså SD, et al.
(1993). Rare Ha-ras1 alleles and
predisposition to testicular cancer. Int. J.
Cancer 53(6): 938-940.
Siegrist S, Feral C, Chami M, et al. (2001).
hH-Rev107, a class II tumor suppressor
gene, is expressed by post-meiotic
testicular germ cells and CIS cells but not
by human testicular germ cell tumors.
Oncogene 20(37): 5155-5163.
Strohmeyer T, Reese D, Press M, et al.
(1995). Expression of the c-kit protooncogene and its ligand stem cell factor
(SCF) in normal and malignant human
testicular tissue. J. Urol. 153(2): 511-515.
Bokemeyer C, Kuczyk MA, Dunn T, et al.
(1996). Expression of stem-cell factor and
its receptor c-kit protein in normal testicular
tissue and malignant germ-cell tumours. J.
Cancer Res. Clin. Oncol. 122(5): 301-306.
Luo LY, Rajpert-De Meyts E, Jung K, et al.
(2001). Expression of the normal epithelial
cell-specific 1 (NES1; KLK10) candidate
tumour suppressor gene in normal and
65
256
257
258
259
260
261
262
263
66
malignant testicular tissue. Br. J. Cancer
85(2): 220-224.
Chang A, Yousef GM, Jung K, et al.
(2001). Identification and molecular
characterization of five novel kallikrein
gene 13 (KLK13; KLK-L4) splice variants:
differential expression in the human testis
and testicular cancer. Anticancer Res.
21(5): 3147-3152.
Von Eyben FE, de Graaff WE, Marrink J, et
al. (1992). Serum lactate dehydrogenase
isoenzyme 1 activity in patients with
testicular germ cell tumors correlates with
the total number of copies of the short arm
of chromosome 12 in the tumor. Mol. Gen.
Genet. 235(1): 140-146.
Von Eyben FE (2001). A systematic review
of lactate dehydrogenase isoenzyme 1 and
germ cell tumors. Clin. Biochem. 34(6):
441-454.
Bouras M, Tabone E, Bertholon J, et al.
(2000). A novel SMAD4 gene mutation in
seminoma germ cell tumors. Cancer Res.
60(4): 922-928.
Aubry F, Satie AP, Rioux-Leclercq N, et al.
(2001). MAGE-A4, a germ cell specific
marker, is expressed differentially in
testicular tumors. Cancer 92(11): 27782785.
Eid H, Institoris E, Geczi L, et al. (1999).
mdm-2 expression in human testicular
germ-cell tumors and its clinical value.
Anticancer Res. 19(4C): 3485-3490.
Datta MW, Macri E, Signoretti S, et al.
(2001). Transition from in situ to invasive
testicular germ cell neoplasia is associated
with the loss of p21 and gain of mdm-2
expression. Mod. Pathol. 14(5): 437-442.
Kraft HJ, Mosselman S, Smits HA, et al.
(1996). Oct-4 regulates alternative platelet-
derived growth factor alpha receptor gene
promoter in human embryonal carcinoma
cells. J. Biol. Chem. 271(22): 1287312878.
264 Qiao D, Zeeman AM, Deng W, et al.
(2002). Molecular characterization of hiwi,
a human member of the piwi gene family
whose overexpression is correlated to
seminomas. Oncogene 21(25): 3988-3999.
265 Sugihara A, Saiki S, Tsuji M, et al. (1997).
Expression of Fas and Fas ligand in the
testes and testicular germ cell tumors: an
immunohistochemical study. Anticancer
Res. 17(5B): 3861-3865.
266 Braendstrup O, Bols B, and Jensen L
(1999). Fas and Fas-ligand expression in
seminomatous testes. APMIS 107(4): 431436.
267 Hara S, Miyake H, Nakamura I, et al.
(2001). Increased Fas ligand expression in
the tumor tissue and serum of patients with
testicular germ cell tumors with
seminomatous elements. Urology 58(3):
471-476.
268 Kersemaekers AM, van Weeren PC,
Oosterhuis JW, et al. (2002). Involvement
of the Fas/FasL pathway in the
pathogenesis of germ cell tumours of the
adult testis. J. Pathol. 196(4): 423-429.
269 Schmelz HU, Abend M, Kraft K, et al.
(2002). Fas/Fas ligand system and
apoptosis induction in testicular carcinoma.
Cancer 95(1): 73-81.
270 Takayama H, Takakuwa T, Tsujimoto Y, et
al. (2002). Frequent Fas gene mutations in
testicular germ cell tumors. Am. J. Pathol.
161(2): 635-641.
Paper I
Skotheim RI, Diep CB, Kraggerud SM, Jakobsen KS, and Lothe RA
Evaluation of loss of heterozygosity/allelic imbalance scoring in tumor DNA
Cancer Genetics and Cytogenetics, 2001, 127(1): 64-70
blank
Cancer Genetics and Cytogenetics 127 (2001) 64–70
Evaluation of loss of heterozygosity/allelic imbalance
scoring in tumor DNA
Rolf I. Skotheima,b, Chieu B. Diepa, Sigrid M. Kraggeruda, Kjetill S. Jakobsenb,
Ragnhild A. Lothea,*
a
Department of Genetics, Institute for Cancer Research, The Norwegian Radium Hospital, Montebello, N-0310 Oslo, Norway
b
Division of General Genetics, Biological Institute, University of Oslo, P.O. Box 1031 Blindern, N-0315 Oslo, Norway
Received 23 August 2000; received in revised form 7 November 2000; accepted 20 November 2000
Abstract
Loss of heterozygosity and allelic imbalance in tumors are usually detected by either radioactive
labeling of PCR products with subsequent scoring of autoradiographs or by a semi-quantitative
fluorescence-based protocol. Polymorphic microsatellite loci are the most common marker type
used in these studies. Even though no consensus exists as to how to evaluate such data, results are
often compared directly between studies applying the two different protocols. In the present
study, we analyzed twice by each protocol three loci in 60 blood/tumor pairs, finding good correlation between the results obtained by the two methods. However, a higher sensitivity and the possibility to correct for stutter peaks were among several advantages inherent in the fluorescence labeling approach. In addition, we determined the cut-off level for allelic imbalance scoring by the
fluorescent primer protocol, by repeated analysis of 485 constitutional heterozygous genotypes at
20 different dinucleotide repeat loci. Based on the standard deviation, we found that allelic imbalance should be scored whenever the peak height of one allele in tumor DNA is reduced to less
than 0.84 of its value in constitutional DNA, relative to the other allele. Applying this cut-off
value, more imbalances are detected than by the visual scoring of autoradiographs. Our data
therefore suggest that a lower threshold value (0.75) must be used when results from both fluorescent and radioactive assays are compared. © 2001 Elsevier Science Inc. All rights reserved.
1. Introduction
Loss of heterozygosity (LOH) means loss of one allele at a
constitutional heterozygous locus. Non-random LOH in a
certain tumor type indicates the map position of a tumor suppressor gene (TSG) whose loss promotes neoplastic progression [1,2]. The loss can be the second and fatal event of the
total functional knockout of the TSG, in which the other allele has been mutated or imprinted in an earlier event, somatically or constitutionally. Cavenee et al. [3] showed, by
Southern blot analysis, LOH at the tumor suppressor locus
RB1 in retinoblastomas from patients carrying a germ line
mutation of the RB1 gene, and thus experimentally confirmed
Knudson two-hit hypothesis for inactivation of a TSG [4].
Detection of a skewed intensity ratio between two alleles
at a locus is described as allelic imbalance (AI) for the tumor in question. AI may reflect the complete loss of one al* Corresponding author. Tel.: 47 22 93 44 15; fax: 47 22 93 44 40
E-mail address: [email protected] (R.A. Lothe).
lele that is masked by the presence of normal cells, by tumor
heterogeneity, or by non-clonal loss. Increased DNA copy
number will also reveal an AI pattern.
Screening of tumors for LOH and AI is widely used as a
tool for trapping TSGs [1], and was initially done by Southern blots with restriction fragment length polymorphism
(RFLP) or variable number of tandem repeats (VNTR)
probes [5,6]. Later, amplification of specific RFLPs by PCR
with consecutive restriction digestion was applied [7] before PCR of highly polymorphic microsatellite loci became
common in use. The microsatellite loci usually display a
high fraction of heterozygotes, and they are abundant genome-wide. In the most recent version of Généthon’s human genetic map [8], 5264 (CA)n dinucleotide repeats were
positioned with an average interval size of 1.6 cM.
In most AI/LOH studies so far, the PCR products have
been labeled by incorporation of radioactive nucleotides,
followed by electrophoresis in a polyacrylamide slab gel
and detection by autoradiography [7]. In such a protocol,
usually two to four primer pairs are multiplexed in the PCR,
0165-4608/01/$ – see front matter © 2001 Elsevier Science Inc. All rights reserved.
PII: S0165-4608(00)00433-7
R.I. Skotheim et al. / Cancer Genetics and Cytogenetics 127 (2001) 64–70
limited by the requirement of about 20 bp size-interval between each locus. The AI/LOH is usually determined by visual examination of the X-ray films, but semi-quantitative
determination can be performed by phosphor imaging.
In an alternative protocol, one primer in each primer pair
is fluorescently labeled, and the PCR products are detected
by automatic slab gel sequencers [9,10] or capillary electrophoresis [11]. Primers targeting loci with overlapping size
ranges are labeled with different fluorescent dyes, and PCRproducts from several loci can be pooled and assayed together. Analyzes of up to 24 loci simultaneously have been
described [12]. By this method, semi-quantitative data are
obtained, and thus the degree of AI can be calculated. However, the chosen cut-off values for designation of AI differ
among previous studies [13–17]. Secondly, the presence of
normal cells in the tumor samples is ignored in some studies, and finally, the contribution of shadow bands/stutter
peaks from one allele to the measurement of a neighboring
allele should be considered in the quantitation of imbalance.
These aspects may account for different results obtained
among comparable studies.
We have compared the radioactivity- and fluorescencebased methods by analyzing the same microsatellite markers twice by each method in two series of human blood/tumor pairs, and found good correlation between the results of
the two protocols. However, the fluorescence-based protocol has some advantages, and it is more sensitive to detect
minor imbalances. Based on our results, we propose guidelines on how to score and interpret the obtained results.
2. Materials and methods
2.1. Samples
DNA from 30 fresh frozen biopsies of colorectal carcinomas (CRC) and 30 testicular germ cell tumors (TGCT), as
well as DNA from corresponding peripheral blood samples
from each patient were analyzed.
The 30 CRC samples were selected from a series of tumors,
each known to contain a high tumor fraction (microscopic
evaluation of hematoxylin and eosin stained sections), with
an estimated mean of 84% (range: 62–97%)[18]. The series
of CRCs had previously been scored for AI/LOH at 18q21
loci, others than those included in the present study (unpublished data). Since changes at 18q21 loci usually reflect deletion of large areas or even the whole chromosome arm in
CRCs, those data could be used to select samples expected
to show LOH (n10), AI (n10), and retained heterozygosity (n10) at other loci residing within the same chromosome band.
TGCTs are histologically heterogeneous and may contain several tumor components. The samples included in
this study had a mean tumor fraction of 73% (range: 25–
95%; unpublished data), as visually evaluated by a pathologist from hematoxylin and eosin stained sections of the frozen biopsies used for DNA isolation. The 30 TGCTs were a
65
selected subset from a study of 51 tumors investigated by
capillary electrophoresis with fluorescent primer labeling at
22 loci (unpublished data), including the three loci used in
the present study. We chose tumors with imbalances ranging the whole scale, from complete LOH to retained heterozygosity, for the three selected loci.
2.2. Selection of microsatellite markers
Both CRC and TGCT are known to exhibit frequent AI/
LOH for loci at, or near the chromosome band 18q21
[19,20], and thus the (CA)n microsatellite loci D18S460 and
D18S554 were included. A third locus, D3S2748, a (CA)n
dinucleotide repeat localized at 3q28 [21], was also analyzed. The three markers do not have overlapping size
ranges; D3S2748 gives fragments in the size range 72 to
106 base pairs (bp), D18S460 range from 177 to 191 bp,
and D18S554 from 210 to 228 bp. Primer sequences for
PCR amplification, fragment sizes, and map positions were
found from the Human Genome Database [22] and the
Généthon linkage map [8].
The three markers were PCR amplified in multiplex from
blood (constitutional) and tumor DNA of the 60 cancer patients. Independent PCR amplifications were performed in
duplicate for each of the two different protocols.
2.3. Protocol for radioactive labeling of PCR amplified
microsatellites
A total reaction volume of 10 l consisted of
1xGeneAmp®PCR buffer containing 1.5 mM MgCl2 (Applied Biosystems, Foster City, CA, USA), 2 to 5 pmol of
each primer (Research Genetics, Inc., Huntsville, AL, USA),
200 M each of dATP, dGTP and dTTP, 2.5 M unlabeled
dCTP (Amersham Pharmacia Biotech Inc., London, UK),
0.7 Ci[-32P]dCTP (Amersham P. B.), 0.4 units AmpliTaq
DNA Polymerase (Applied Biosystems), and 25 or 50 ng
DNA template from blood or tumor tissue, respectively.
The PCR was carried out in a 96-well format using an
MJ PTC-200 thermocycler (MJ Research, Inc., Watertown,
MA, USA). Two minutes of denaturation at 94C was followed by 27 cycles of 30 s denaturation at 94C, 75 s annealing at 55C and 15 s elongation at 72C, before 6 min final extension at 72C.
The PCR products were mixed with gel loading buffer
and denatured for a few minutes at 95C before they were
left on ice. The products were loaded on to a 6% polyacrylamide gel in a 35 cm 43 cm BRL S2 Sequencing Gel Electrophoresis Apparatus (Life Technologies, Inc., Gaithersburg, MD, USA). One lane was reserved for an MspI
digested pBR322 size standard (New England Biolabs, Beverly, MA, USA) labeled with [-32P]dATP (Amersham
Pharmacia Biotech). The gel was run at 2.1 kV, 150 mA
for about 100 min in 0.5xTBE, and then transferred to a
Whatman 3MMChromatography paper (Whatman International Ltd., Maidstone, UK), covered with plastic film and
66
R.I. Skotheim et al. / Cancer Genetics and Cytogenetics 127 (2001) 64–70
dried for 25 min at 80C in a Speed Gel SG210D vacuum
drier (Savant Instruments, Inc., Farmingdale, NY, USA).
Several exposures of Fuji Medical x-ray Films (Fuji Photo
Film Co., Ltd., Tokyo, Japan) were taken in a cassette without intensifying screen, and developed in an Agfa Curix 60
developer (Agfa-Gevaert N.V., Mortsel, Belgium).
Visual evaluation of the autoradiographs was performed
independently by two of the authors. The tumor genotypes
were compared against their corresponding normal genotypes and designated as LOH, AI, retained heterozygosity
or homozygosity.
2.4. Fluorescent primer protocol
The same PCR protocol as described above was applied
with two modifications: the CA strand primers were labeled
in 5 end with the fluorochromes HEX, TET or 6-FAM
(DNA Technology AS, Aarhus, Denmark), and all four
dNTPs had the same concentration, 200 M.
One microliter PCR product was mixed with 0.5 l GeneScan-350 [TAMRA] Size Standard (Applied Biosystems)
in 12 l deionized formamide, CH3NO (Kodak Eastman
Chemical Company, New Haven, CT, USA), and denatured
for 3 min at 95C, before quick cooling on ice.
Up to 96 samples at the time were then ready for capillary electrophoresis using an ABI PRISM™310 Genetic
Analyzer (Applied Biosystems). The samples were electrokinetically injected at 15 kV, for 1 to 10 s, into a 50 m 47 cm capillary (Applied Biosystems). The PCR products
were separated by 15 kV at 60C through Applied Biosystems’ polymer POP4, which was automatically refilled prior
to each new injection. The platina electrodes were immersed
in 310 Genetic Analyzer Buffer with EDTA (Applied Biosystems). Run time was set to 23 min to include two size
standard peaks longer than the PCR products of interest.
The results were analyzed by the ABI PRISM 310 GeneScan 3.1 software, and then exported to Genotyper 2.1
(both Applied Biosystems) for semi-automatic allele calling. This software uses pre-programmed marker-categories
and filters to ignore the non-template adenine- and stutter
peaks. Regardless, the electropherograms must be visually
inspected to ensure correct allele calling. Measured peak
heights have shown to be more reproducible than the peak
areas by the software’s allele calling system [23], and are
also recommended for quantitation of PCR products by the
instrument’s manufacturer (Applied Biosystems). Thus,
peak heights (relative fluorescence units) were exported to
Microsoft Excel where a semi-quantitative expression of the
degree of allelic imbalance, QLOH [6], and further statistical
analyses were performed. QLOH is calculated from the measured peak heights by dividing the allele ratio in tumor
DNA (t1/t2) by the allele ratio in constitutional (normal)
DNA (n1/n2); QLOH (t1/t2)/(n1/n2). When this ratio gives a
value greater than one, QLOH is set to the inversion. Thus,
QLOH range from 0 to 1, indicating total loss to retained heterozygosity, respectively.
2.5. Reproducibility test of the fluorescent primer protocol
In order to estimate the reproducibility and accuracy of
results obtained by the fluorescence-based protocol, we analyzed 22 (CA)n repeat markers in 51 blood DNA samples.
Fifty-one samples were heterozygous at 6 or more loci, representing a total of 485 genotypes for analysis. These were
amplified twice, and QLOH values with expectation values of
1.00 were calculated for determination of standard deviation
among samples with retained heterozygosity.
2.6. Stutter peak corrections by the fluorescent primer
protocol
In addition to the main allele, amplification of microsatellites generates products referred to as shadow-bands (seen
on a gel) or stutter-peaks (electropherogram) due to polymerase slippage during elongation [24,25]. Usually, the additional fragments are one to four repeat units shorter than
the allele, and when the size of the two alleles differs by one
repeat unit, the stutter from the longer allele will contribute
significantly to the shorter allele’s main peak/band. As outlined in the Results and Discussion sections, this phenomenon can be corrected for by the fluorescent primer protocol.
3. Results
3.1. Visual scoring of autoradiograms
The visual scoring of 180 genotypes (3 loci in 60 pairs of
blood/tumor DNA) on X-ray films resulted in 130 informative cases (LOH, AI or retained heterozygotes), 38 constitutionally homozygotes, and 12 cases not scoreable. The 12
were unsuccessfully analyzed a third time using the same
standard PCR conditions.
Inter-observer variation was found for 4 of the 168 scoreable genotypes. One locus in one tumor was scored as homozygous by one observer, and heterozygous by the other.
In the three other instances, the conflicting scorings were AI
versus retained heterozygosity. The corresponding QLOH
values for these were 0.63, 0.80 and 0.89.
3.2. Semi-quantitative scoring by the fluorescence-based
protocol
All blood/tumor pairs were successfully genotyped at all
loci, and semi-quantitative expressions (QLOH) for the degree of imbalances were calculated as described in Fig. 1.
The QLOH values were consistent between the two repetitive
analyzes, with a mean difference of only 0.03.
3.3. Comparison of results obtained by the two protocols
Fig. 2 shows a clear correlation between the results obtained by the two methods. The gel-scored categories
(LOH, AI, and retained heterozygosity) are shown as
curves, and the corresponding QLOH values obtained by the
fluorescent primer protocol can be read along the x-axis.
With the exception of one outlier, the samples visually
R.I. Skotheim et al. / Cancer Genetics and Cytogenetics 127 (2001) 64–70
67
Fig. 3. Distribution of QLOH values among samples with simulated retained
heterozygosity. The expected value of QLOH for normal heterozygotes is
1.00. DNA from 51 blood samples, known to be heterozygous at 20 (CA)n
dinucleotide loci, were amplified independently twice. The second amplifications from the blood samples were simulated as tumors with retained heterozygosity. Calculation of 485 QLOH values resulted in a one-sided normal
distribution with standard deviation 0.083.
Fig. 1. Allelic imbalance (AI) in tumor DNA. The QLOH values correspond
to the degree of allelic imbalance, and range from 0.00 to 1.00, reflecting
complete LOH to retained heterozygosity, respectively. The QLOH value is
calculated as a ratio between the allele ratios in tumor and constitutional
DNA. The peak heights are measured in relative fluorescence units. In this
example [t1/t2]/[n1/n2] [2862/443]/[2458/2067] 5.43, but since this
ratio is greater than 1, the QLOH value is set to be the inverse. Thus, QLOH 0.18, showing a reduction of one allele’s intensity, from constitutional to
tumor DNA, by 82% relative to the other allele.
scored as LOH corresponded to QLOH values ranging from
0.00 to 0.34 (n25), AI from 0.22 to 0.79 (n52), and the
retained heterozygotes from 0.73 to 1.00 (n53).
3.4. Normal variation between individual runs using the
fluorescent primer protocol
The replicate analysis of the 485 heterozygous genotypes
enabled us to calculate 485 QLOH values of samples with
known retained heterozygosity. They showed a one-sided
normal distribution with standard deviation 0.083 (Fig. 3).
The standard deviation calculated for each locus ranged
from 0.057 (D18S460) to 0.135 (D18S57).
3.5. Correction for stutter peaks by the fluorescent primer
protocol
In the direct comparison of the two protocols, as outlined
here, we did not take stutter peaks/bands into account. After
correction for the contribution of stutter peaks from neighbor
alleles by the fluorescent primer protocol (Fig. 4), some of the
samples scored as retained heterozygotes from the autoradiographs got adjusted QLOH values as low as 0.65. Thus, tumors
with up to 35% reduction of one allele peak relative to the
other allele, comparing blood and tumor DNA, can be scored
as false negative by visual evaluation of autoradiographs.
3.6. Comparison of syntenic loci
Thirty-one samples were informative at both 18q loci,
and the same genotypes, LOH, AI or retained heterozygosity, were found in all but three samples by the radiolabeling
protocol. One tumor showed retained heterozygosity at
D18S460 and AI at D18S554, and in each of the two other
tumors, one locus revealed AI, and the other LOH. By the
fluorescence-based protocol, all samples had similar QLOH
values at the two syntenic loci, differing on average by 0.05.
This shows good reproducibility of measured QLOH, independent of the chosen marker, and when correcting for stutter, the average deviation was decreased to 0.03.
Fig. 2. Visually scored tumor genotypes compared to their distribution in
the QLOH range. The curves show the visually scored tumor genotypes as
LOH, AI or retained heterozygosity. Their corresponding QLOH values are
indicated along the x-axis. With the exception of one outlier, the samples
scored as LOH by visual inspection of autoradiographs had QLOH values
ranging from 0.00 to 0.34, AI from 0.22 to 0.80, and retained heterozygotes
from 0.73 to 1.00 by the fluorescent primer protocol.
4. Discussion
4.1. Cut-off values for determination of AI and LOH
The results obtained by the two methods are usually in
agreement with each other. Doubt of whether to visually
68
R.I. Skotheim et al. / Cancer Genetics and Cytogenetics 127 (2001) 64–70
Fig. 4. Comparison of genotyping by visual scoring of autoradiographs and semi-quantitative detection of PCR products labeled by fluorescence. The TC44
tumor was scored as LOH at D18S554 by both protocols. The TC148 tumor was scored as retained heterozygosity by the visual scoring, and as an AI by the
fluorescence-based protocol. The interference of the stutter peak from the longer allele to the main peak of the shorter allele is obvious here. By the fluorescence-based protocol, such contribution from stutter can be corrected for. The height of a stutter peak compared its main peak is calculated, and then the main
peak’s height, as it would have been without contribution from the neighboring allele’s stutter, can be estimated. For the example above, the stutter peaks have
about one fourth of the true alleles’ heights. For TC148 the 624-allele contributes with about 156 (624/4) relative fluorescent units to the 802-allele’s peak
height. Thus, we ‘normalize’ this peak height to 646 (802–156). Without stutter correction D18S554 in TC148 gets a QLOH [620/623]/[802/624] 0.77.
After correcting for stutter the value decreases to [464/623]/[646/624] 0.72.
score a tumor genotype as AI or as retained heterozygosity
is restricted to those with QLOH values in the 0.7 to 0.8 range
by the fluorescent primer protocol. Samples with QLOH
0.4 are usually scored as LOH by visual examination of
autoradiograms, although some are interpreted as AI.
From the reproducibility test of QLOH, we found that
samples with retained heterozygosity are one-sided normal
distributed with standard deviation of 0.08 from the expected value 1.00. Thus, a sample with retained heterozygosity has a 95% probability to get a QLOH value higher than
0.84. All samples with suspected AI are confirmed, and
hence, the probability that a heterozygous tumor sample
gets a QLOH 0.84 both times is 0.0025, given that the technical errors of duplicate analyzes are independent of each
other. Thus, a cut-off value at 0.84 provides safe interpretation of AI. However, we noticed some locus variation, with
maximum standard deviation of 0.135 for D18S57. But
even for this locus, a confirmed QLOH
0.84 has a 95%
probability to be due to imbalance.
Using 0.84 as a cut-off value for AI will result in detection
of more samples with AI than visual inspection of autoradiographs does. Thus, when results obtained by both methods are
compared, one can not use this cut-off value. Fig. 2 suggests
that a QLOH value of 0.75 separates visually scored AI from retained heterozygosity, and 0.30 separates AI from LOH. Usage of 0.75 as a cut-off for AI ensures the correctness of direct
comparison of fluorescence-based studies with results from
visual scoring of autoradiographs, despite the exclusion of
some samples with most likely imbalances (0.75–0.84).
Because AI and LOH often span huge chromosomal regions, or results from loss of whole chromosomes or chromosome arms, we expect correlation between the results
from the two syntenic 18q loci. Three samples showed different genotypes by the visual evaluation, whereas no such
differences were found between the loci by the fluorescence-based protocol. Does this mean that artificial breakpoints in the AI pattern are more likely to occur by the radioactive labeling protocol? At least, this emphasizes some
important aspects for designation of minimal common deleted regions (i.e., the smallest region of overlap, SRO).
Rules for determination of SROs have been outlined by
Thorstensen et al. [26]. In brief, more than one marker
should show the imbalance, the results must be confirmed at
both sides of the chromosomal breakpoints, and the SRO
should rely on more than a single tumor.
4.2. Contribution of DNA from normal cells
An obstacle to automated scoring of AI/LOH is the cellular and genetic heterogeneity in most tumor samples. One
approach to determine which QLOH values that should be
considered as AI is to plot the number of QLOH that falls into
different histogram groups. The distribution has in some
studies shown to be bimodal [6,9,14], and the samples residing in the lower QLOH are designated as AI and the others as
retained heterozygosity. The cut-off values found by this
approach will depend on the tumor type and percentage of
tumor cells in each sample, and thus a general value for
R.I. Skotheim et al. / Cancer Genetics and Cytogenetics 127 (2001) 64–70
scoring of AI can not be set. However, if the tumor content
is comparable among the samples, a reasonable cut-off
value can be found for the specific study. Such a test would
not be feasible in the present study because of our criteria of
sample selection, aiming to include tumors with QLOH values evenly distributed across the 0 to 1 range.
For interpretation of the results, the fraction of tumor
cells in the sample should be evaluated prior to isolation of
the DNA. It is easier to interpret the data if the tumor samples are rich in tumor cells. Micro-dissection, flow sorting
of cells and whole genome amplification may be useful
tools to enrich the tumor content, and thus increase the sensitivity of LOH detection [23].
4.3. Interpretation of QLOH
As mentioned, in informative cases, AI should be scored
if the peak height of an allele in tumor is reduced to less
than 0.84 of its value in the normal DNA, relative to the
other allele.
When the number of PCR cycles are kept within the logarithmic phase of amplification, the measured QLOH is approximately proportional to the allele ratio in the template
[23,27]. This indicates that different QLOH limits should be
used for scoring of LOH according to the amount of normal
cells in the tumor biopsy, e.g., if QLOH.50 for a sample
where the estimated fraction of normal cells is 50%, this indicates a total LOH. However, this does not imply that QLOH
is proportional to actual number of cells with retained heterozygosity, because tumor cells with LOH are not necessarily disomic for the chromosome harboring the locus in
question, and thus each tumor cell can contribute with more
than one copy of the retained allele.
In addition to presence of normal cells, the QLOH can reflect tumor heterogeneity. If a tumor shows loss of two different chromosomal regions, and markers in one of the regions give consistent lower QLOH, it is reasonable to assume
that the loss of this region was an earlier event during tumor
progression than deletion of the other. Values indicating AI
can in addition to partial loss also reflect genomic gain/amplification. In order to distinguish between loss and gain,
co-amplification of a locus believed to be inert in the tumor
type in question can be included in the PCR. Peak heights of
the investigated loci are then related to those of the control
to see whether an imbalance is due to loss or gain. Further
interpretation of AI results can also be made by a combination with other methods like comparative genomic hybridization (CGH), fluorescence in situ hybridization (FISH)
and karyotyping.
A barely detectable AI is not necessarily unimportant. It
may for instance reflect a minor clone with metastatic abilities. However, genomic instability is a typical phenotype in
cancer, and minor imbalances are more likely to be an effect
of, rather than causing tumor growth. Further, this implies
the choice of a somewhat strict QLOH value for scoring of
AI, provided that the presence of normal cells is low.
69
4.4. Stutter peak feature of amplified microsatellites
The stutter peaks/bands are detected by both protocols,
and may cause incorrect scoring if the size difference between the alleles is small. In such instances, one may score
heterozygotes as false homozygotes, and vice versa, and imbalances may remain undetected. This is a problem mainly
related to the visual scoring protocol. When the two alleles
of a microsatellite differs by one (and in less extent 2) repeat units, the minus one repeat peak of the longer allele
gives additional signal to the shorter one. An allele loss here
will result in an artificial high QLOH value because of the
contribution from the other allele’s stutter peak, and by the
visual gel-scoring, an imbalance can be masked, and LOH
can be reduced to AI. However, this can be corrected for by
the fluorescent primer protocol (Fig. 4).
Stutter peaks are in general a minor problem for microsatellites with longer repeat units, like tetra- and pentanucleotide repeats, and thus such markers give on average
slightly more reproducible QLOH values. This is however
marker/sequence dependent, and some dinucleotide repeats
may perform equally well as tetra- and pentanucleotide repeats do [28]. The locus pattern shown on an electropherogram (main allele stutter) may also be of help to identify
the true alleles from possible unspecific PCR products. But
most important, dinucleotide repeats are usually chosen for
AI/LOH studies because of their abundance and genomewide distribution. Apart from the (A)n mononucleotide repeat, they are the most common microsatellites in the human
genome [29], and more than 5000 have been genetically
mapped with an average interval size of 1.6 cM.
4.5. Advantages with the fluorescent primer protocol
The obvious advantage with the fluorescence-based
method versus visual scoring is the more objective genotyping with instrumental quantitation of PCR products. Further, the quantitation gives the opportunity to correct for
stutter contribution. Genotypes on autoradiographs can also
be quantified by optical densitometry [6,30], but this detection is not as linear with respect to time and amount of radioactivity [31] as fluorescence detected with laser is [12].
Our results show that the fluorescent protocol is more reproducible and able to detect minor imbalances than visual
evaluation of X-ray films. It also seemed to be more sensitive, as all samples were scoreable at all loci by this protocol, whereas 12 out of 180 could not be determined by the
radioactive labeling protocol.
The radiolabeling protocol is neither as adaptable to high
throughput analysis as the fluorescence-based protocol. By
the latter protocol, PCR products can be pooled before separation in addition to the multiplexing of markers in the PCR
mix. The processing of gels and the scoring of autoradiographs are also rate-limiting steps in conventional scoring
of LOH and AI. These labor-intensive procedures are eliminated by capillary electrophoreses, where slab gel preparation is replaced by automated capillary filling, and sample
70
R.I. Skotheim et al. / Cancer Genetics and Cytogenetics 127 (2001) 64–70
loading is replaced with electrokinetic injection. Samples
requiring a new run can be automatically re-injected and analyzed by the instrument without the need to remake gels
and new loading of samples. Sizing is also more reproducible by the fluorescent protocol. The size standard has reserved its own fluorescent dye, and is added to, and electrophoresed together with the samples (internal size standard).
Thus, we avoid the mobility shift problem associated with
external size standards. Standard deviation of sizing in capillary electrophoreses has been reported to be less than 0.2
bp [11]. Finally, by applying a fluorescence-based protocol,
the potential hazard of ionizing radiation is avoided.
Acknowledgments
R.I.S. and S.M.K. are research fellows of the Research
Council of Norway and the Norwegian Cancer Society
(NCS), respectively. This study was also supported by
grants from the Norwegian Health and Rehabilitation legacy (C.B.D.) and the NCS (R.A.L.).
References
[1] Lasko D, Cavenee W, Nordenskjold M. Loss of constitutional heterozygosity in human cancer. Annu Rev Genet 1991;25:281–314.
[2] Fearon ER, Cho KR, Nigro JM, Kern SE, Simons JW, Ruppert JM,
Hamilton SR, Preisinger AC, Thomas G, Kinzler KW, Vogelstein B.
Identification of a chromosome 18q gene that is altered in colorectal
cancers. Science 1990;247:49–56.
[3] Cavenee WK, Dryja TP, Phillips RA, Benedict WF, Godbout R, Gallie BL, Murphree AL, Strong LC, White RL. Expression of recessive
alleles by chromosomal mechanisms in retinoblastoma. Nature 1983;
305:779–84.
[4] Knudson AGJ. Mutation and cancer: statistical study of retinoblastoma. Proc Natl Acad Sci USA 1971;68:820–3.
[5] Lothe RA, Nakamura Y, Woodward S, Gedde-Dahl TJ, White R.
VNTR (variable number of tandem repeats) markers show loss of
chromosome 17p sequences in human colorectal carcinomas. Cytogenet Cell Genet 1988;48:167–9.
[6] Solomon E, Voss R, Hall V, Bodmer WF, Jass JR, Jeffreys AJ, Lucibello FC, Patel I, Rider SH. Chromosome 5 allele loss in human colorectal carcinomas. Nature 1987;328:616–19.
[7] Merlo GR, Cropp CS, Callahan R, Takahashi T. Detection of loss of
heterozygosity in tumor DNA samples by PCR. Biotechniques 1991;
11:166–70.
[8] Dib C, Faure S, Fizames C, Samson D, Drouot N, Vignal A, Millasseau P, Marc S, Hazan J, Seboun E, Lathrop M, Gyapay G, Morissette J, Weissenbach J. A comprehensive genetic map of the human
genome based on 5,264 microsatellites. Nature 1986;380:152–4.
[9] Hampton GM, Larson AA, Baergen RN, Sommers RL, Kern S, Cavenee WK. Simultaneous assessment of loss of heterozygosity at multiple microsatellite loci using semi-automated fluorescence-based detection: subregional mapping of chromosome 4 in cervical
carcinoma. Proc Natl Acad Sci USA 1996;93:6704–9.
[10] Cawkwell L, Bell SM, Lewis FA, Dixon MF, Taylor GR, Quirke P.
Rapid detection of allele loss in colorectal tumours using microsatellites and fluorescent DNA technology. Br J Cancer 1993;67:1262–7.
[11] Mansfield ES, Vainer M, Enad S, Barker DL, Harris D, Rappaport E,
Fortina P. Sensitivity, reproducibility, and accuracy in short tandem
repeat genotyping using capillary array electrophoresis. Genome Res
1996;6:893–903.
[12] Schwengel DA, Jedlicka AE, Nanthakumar EJ, Weber JL, Levitt RC.
Comparison of fluorescence-based semi-automated genotyping of
multiple microsatellite loci with autoradiographic techniques. Genomics 1994;22:46–54.
[13] Larson AA, Kern S, Curtiss S, Gordon R, Cavenee WK, Hampton
GM. High resolution analysis of chromosome 3p alterations in cervical carcinoma. Cancer Res 1997;57:4082–90.
[14] MacGrogan D, Levy A, Bostwick D, Wagner M, Wells D, Bookstein
R. Loss of chromosome arm 8p loci in prostate cancer: mapping by
quantitative allelic imbalance. Genes Chromosomes Cancer 1994;10:
151–9.
[15] Wang Y, Hung SC, Linn JF, Steiner G, Glazer AN, Sidransky D,
Mathies RA. Microsatellite-based cancer detection using capillary array electrophoresis and energy-transfer fluorescent primers. Electrophoresis 1997;18:1742–9.
[16] Tomlinson I, Ilyas M, Johnson V, Davies A, Clark G, Talbot I, Bodmer W. A comparison of the genetic pathways involved in the pathogenesis of three types of colorectal cancer. J Pathol 1998;184:148–52.
[17] Canzian F, Salovaara R, Hemminki A, Kristo P, Chadwick RB, Aaltonen LA, de la Chapelle A. Semiautomated assessment of loss of
heterozygosity and replication error in tumors. Cancer Res 1996;56:
3331–7.
[18] Meling GI, Lothe RA, Borresen AL, Hauge S, Graue C, Clausen OP,
Rognum TO. Genetic alterations within the retinoblastoma locus in
colorectal carcinomas. Relation to DNA ploidy pattern studied by
flow cytometric analysis. Br J Cancer 1991;64:475–80.
[19] Murty VV, Li RG, Houldsworth J, Bronson DL, Reuter VE, Bosl GJ,
Chaganti RS. Frequent allelic deletions and loss of expression characterize the DCC gene in male germ cell tumors. Oncogene 1994;9:
3227–31.
[20] Vogelstein B, Fearon ER, Hamilton SR, Kern SE, Preisinger AC,
Leppert M, Nakamura Y, White R, Smits AM, Bos JL. Genetic alterations during colorectal-tumor development. N Engl J Med 1988;319:
525–32.
[21] Stoilova D, Child A, Desai SP, Sarfarazi M. Refinement of the locus
for autosomal dominant juvenile optic atrophy to a 2 cM region on
3q28. Ophthalmic Genet 1997;18:1–6.
[22] GDB (TM) Human Genome Database [database online]. Johns Hopkins University, Baltimore, Maryland, USA. 1990-, updated daily.
Available from Internet: URL http://www.gdb.org/.
[23] Paulson TG, Galipeau PC, Reid BJ. Loss of heterozygosity analysis
using whole genome amplification, cell sorting, and fluorescencebased PCR. Genome Res 1999;9:482–91.
[24] Hauge XY, Litt M. A study of the origin of ‘shadow bands’ seen
when typing dinucleotide repeat polymorphisms by the PCR. Hum
Mol Genet 1993;2:411–15.
[25] Smeets HJ, Brunner HG, Ropers HH, Wieringa B. Use of variable
simple sequence motifs as genetic markers: application to study of
myotonic dystrophy. Hum Genet 1989;83:245–51.
[26] Thorstensen L, Qvist H, Heim S, Liefers GJ, Nesland JM, Giercksky
KE, Lothe RA. Evaluation of 1p losses in primary carcinomas, local
recurrences and peripheral metastases from colorectal cancer patients.
Neoplasia 2000;2:514–22.
[27] Stolovitzky G, Cecchi G. Efficiency of DNA replication in the polymerase chain reaction. Proc Natl Acad Sci USA 1996;93:12947–52.
[28] Shaw SH, Carrasquillo MM, Kashuk C, Puffenberger EG, Chakravarti
A. Allele frequency distributions in pooled DNA samples: applications to mapping complex disease genes. Genome Res 1998;8:111–23.
[29] Beckmann JS, Weber JL. Survey of human and rat microsatellites.
Genomics 1992;12:627–31.
[30] Zauber NP, Sabbath-Solitare M, Marotta SP, Mcmahon L, Bishop
DT. Comparison of allelic ratios from paired blood and paraffinembedded normal tissue for use in a polymerase chain reaction to assess loss of heterozygosity. Mol Diagn 1999;4:29–35.
[31] Laskey RA, Mills AD. Quantitative film detection of 3H and 14C in
polyacrylamide gels by fluorography. Eur J Biochem 1975;56:335–41.
blank
Paper II
Skotheim RI, Kraggerud SM, Fosså SD, Stenwig AE, Gedde-Dahl T Jr, Danielsen
HE, Jakobsen KS, and Lothe RA
Familial/bilateral and sporadic testicular germ cell tumors show frequent
genetic changes at loci with suggestive linkage evidence
Neoplasia, 2001, 3(3): 196-203
blank
RESEARCH ARTICLE
Neoplasia . Vol. 3, No. 3, 2001, pp. 196 – 203
196
www.nature.com/neo
Familial/Bilateral and Sporadic Testicular Germ Cell Tumors Show
Frequent Genetic Changes at Loci with Suggestive Linkage
Evidence1
Rolf I. Skotheim* y, Sigrid M. Kraggerud*, Sophie D. Fosså z, Anna E. Stenwig x, Tobias Gedde -Dahl jr. {,
Håvard E. Danielsen x, Kjetill S. Jakobsen y and Ragnhild A. Lothe*
Department of *Genetics and xPathology, Institute for Cancer Research, and Department for zOncology and
Radiotherapy, The Norwegian Radium Hospital, Montebello, Oslo N -0310, Norway; yDivision of General Genetics,
Biological Institute, University of Oslo, P.O. Box 1031, Blindern, Oslo N -0315, Norway; {Institute of Forensic
Medicine, National Hospital, University of Oslo, Oslo N -0027, Norway
Abstract
Testicular germ cell tumor ( TGCT ) is the most
common tumor type among adolescent and young
adult males. Familial clustering and bilateral disease
are suggestive of a genetic predisposition among a
subgroup of these patients, but susceptibility genes
for testicular cancer have not yet been identified.
However, suggestive linkage between disease and
genetic markers has been reported at loci on
chromosome arms 3q, 5q, 12q, 18q, and Xq. We
have analyzed primary familial / bilateral ( n = 20 ) and
sporadic ( n = 27 ) TGCTs, including 28 seminomas
and 19 nonseminomas, for allelic imbalance ( AI )
within the autosomal regions. DNA from all tumors
were analyzed by fluorescent polymerase chain
reaction of 22 polymorphic loci at 3q27 - ter,
5q13 – 35.1, 12q21 - ter, and 18q12 – ter. All tumor
genotypes were evaluated against their corresponding constitutional genotypes. The percentages of
TGCTs with genetic changes at 3q, 5q, 12q, and
18q, were 79%, 36%, 53% and 43%, respectively.
The frequencies at 3q and 12q in nonseminomas
were significantly higher than in seminomas
( P = .003 and P = .004 ). In order to evaluate changes
at hemizygous Xq loci, five loci were analyzed by
co - amplification with an autosomal reference
marker known to reveal retained heterozygosity in
the tumor DNA. Gain of Xq sequences was seen
in more than 50% of the tumors. The degree of
amplification varied among the loci in each of five
tumors, and based on these breakpoints, a common region of overlapping gains was found at
Xq28. No significant differences were found between the frequencies of genetic changes in
familial / bilateral versus sporadic tumors, an observation speaking in disfavor of the existence of a
single susceptibility gene for TGCT in any of the
analyzed regions. Our data suggest that gain of
genetic material at distal Xq and losses at 5q and
18q contribute to establishment of seminomas,
whereas imbalances at 3q as well as gain at distal
part of 12q are associated with further progression
into nonseminomas. Neoplasia ( 2001 ) 3, 196 – 203.
Keywords: allelic imbalance, familial cancer, loss of heterozygosity, susceptibility
gene, testicular germ cell tumor.
Introduction
Testicular germ cell tumor ( TGCT ) is the most common
malignancy among young white males, and the incidence is
increasing rapidly [ 1 – 3 ] . TGCTs are subdivided into two
main histological entities: the undifferentiated seminomas,
and the nonseminomas, composed of embryonic neoplastic
germ cells, which mimic the histogenesis of an early embryo.
Seminomas are believed to arise from a carcinoma in situ
stage, and may develop into nonseminomas [ 4,5 ] . TGCTs
are characterized by overrepresentation of chromosome arm
12p, often through the presence of isochromosome 12p
[ 6,7 ] , and nonrandom losses and gains of certain chromosomes [ 5,8 – 11 ] . TGCTs are nearly always hyperdiploid,
and are frequently in the triploid range [ 5,12 ] .
The cause of TGCT remains unknown. Increased
incidence over time and correlation with socioeconomic
class point toward influence of environmental factors. The
observed familial clustering of TGCT, particularly among
brothers, may be due to their exposure to similar environments, in utero, or as children [ 13 – 16 ] . However, the
four - fold increased risk for father – son transmission
indicates a genetic predisposition [ 14 ] . Men with GCT in
one testis are at increased risk of developing a contralateral malignancy [ 17 ] . The presence of bilateral neoplastic
changes supports a genetic susceptibility for TGCT, but is
Abbreviations: AI, allelic imbalance; CGH, comparative genomic hybridization; GCT, germ
cell tumor; ITCLC, international testicular cancer linkage consortium; LOH, loss of
heterozygosity; TGCT, testicular germ cell tumor
Address all correspondence to: Ragnhild A. Lothe, Department of Genetics, Institute for
Cancer Research, The Norwegian Radium Hospital, Montebello, Oslo N - 0310, Norway.
E-mail: [email protected]
1
This study was supported by grants from the NCS ( R. A. L. ).
Received 28 December 2000; Accepted 24 February 2001.
Copyright # 2001 Nature Publishing Group All rights reserved 1522-8002/01/$17.00
Genetic Changes in Testicular Germ Cell Tumors
Skotheim et al.
also consistent with exposure to environmental carcinogens. Statistical analyses by Nicholson and Harland [ 18 ]
suggest that patients with bilateral disease carry the same
genetic predisposition as familial cases, and that approximately one third of all men with TGCT is genetically
predisposed to the disease.
The International Testicular Cancer Linkage Consortium
( ITCLC ) analyzed 220 polymorphic microsatellite loci
throughout the autosomal genome in selected families with
two or more cases of testicular cancer. None of the markers
showed conclusive evidence of a close map position to a
TGCT predisposing gene, but loci on chromosome arms 3q,
5q, 12q, and 18q showed suggestive linkage to the disease
[ 19 ] . Recently, Rapley et al. [ 20 ] found significant linkage
between markers at Xq27 and TGCT within a subset of
TGCT families ( hLOD = 4.7 ).
In the present study, series of familial / bilateral and
sporadic TGCTs, comparable according to histology and
percentage of tumor cells, were analyzed for somatic
alterations at polymorphic microsatellite loci, within and near
the five candidate regions.
Materials and Methods
Samples from the TGCT Patients
Primary tumor biopsies and corresponding peripheral
blood samples were obtained from 47 Norwegian TGCT
patients. The patients were grouped into cases of familial
and / or bilateral TGCT ( n = 20 ) and cases of sporadic
cancer ( n = 27 ). Among the 20 familial / bilateral TGCTs,
13 were bilateral, 11 had affected family members, and
thus, 4 had both bilateral tumors and familial occurrence of
the disease. Four of the familial / bilateral TGCTs were from
patients with history of cryptorchidism. Median age at
diagnosis was 29 years for the familial / bilateral group and
30 for the sporadic.
Three 5 m sections were taken from different parts of
each frozen tumor sample prior to DNA isolation. The
sections were stained with hematoxylin and eosin and
visually evaluated by light microscopy. The various tumor
197
components were described according to the WHO’s
recommendations [ 21 ] , and percentage of intact neoplastic
cells was estimated for each section. An average of the three
sections per tumor sample was calculated. Among all
tumors, an average of 75% tumor cells was found ( range
30 – 100% ). The familial / bilateral and sporadic tumor groups
were comparable according to histology and estimated
percentage of tumor cells. A total of 28 seminomas included
13 familial / bilateral and 15 sporadic tumors, and among the
19 nonseminomas, 7 were familial / bilateral and 12 sporadic
TGCTs.
DNA was isolated from blood and tumor tissues by
applying the phenol / chloroform extraction principle [ 22 ] .
Microsatellite Analyses
Throughout the five candidate regions suspected to carry
a TGCT susceptibility gene [ 19,20,23 ] , we investigated
markers at 27 microsatellite loci ( Figure 1 ). Primer sequences and allele diversities were obtained from the Human
Genome Database [ 24 ] and the Généthon human linkage
map [ 25 ] .
3q27 - ter Five members of a cancer - prone Canadian
kindred who all developed TGCT [ 26 ] shared a common
haplotype for three markers in the 3q telomeric region. We
analyzed the same three markers, D3S1601, D3S2748, and
D3S1265, which are all located in the 3q27 - ter candidate
region [ 19 ] .
5q13 – 35.1 The candidate region at 5q suggested by
ITCLC lies between the markers D5S428 ( maps together
with the more informative marker D5S617 ) and D5S421.
Leahy et al. [ 23 ] suggested a target region between
D5S428 and D5S409. The marker D5S346 is closely
located to adenomatous polyposis coli ( APC ) [ 27 ] , a
candidate tumor - suppressor gene on 5q21 [ 28,29 ] . Additional three markers were included to flank and refine this
candidate region.
12q21 - ter The ITCLC results showed increasing linkage
evidence along the long arm of chromosome 12, as the
Figure 1. Map positions of the analyzed microsatellite markers. All markers have ( CA )n dinucleotide repeats, except D5S1456 that has a ( GATA )n tetranucleotide
repeat. Numbers to the left of each ideogram indicate the chromosome bands. Numbers to the right of each autosomal genetic map indicate the sex - averaged map
distance between the markers in centi Morgan ( cM ). For the X chromosome, this value represents the female recombination ( fcM ) value, based on the Généthon
human linkage map [ 25 ] .
Neoplasia . Vol. 3, No. 3, 2001
Genetic Changes in Testicular Germ Cell Tumors
markers became more distal [ 19 ] . We therefore analyzed
three markers in the q telomeric region ( 12q24.3 ) as well as
one more proximal marker.
18q12 – ter The suggestive linkage evidence at 18q spanned
several chromosome bands. D18S554 at 18q23 was found
to be the marker with the overall highest linkage score
( nonparametric linkage = 1.6 ) in the ITCLC study [ 19 ] . This
and two flanking markers, D18S58 and D18S461, were
included in the present study. Five additional markers
mapping to 18q12 – 21 were also analyzed due to the
clustering of putative tumor - suppressor genes ( e.g. DCC,
SMAD2, and DPC4 ) in this region.
Xq27 - ter Five markers were chosen to cover and flank the
Xq27 region defined by Rapley et al. [ 20 ] .
Polymerase chain reaction ( PCR ) conditions The 10 l
reaction volume consisted of 1 GeneAmp PCR buffer with
1.5 mM MgCl2 ( Applied Biosystems, Foster City, CA, USA ),
2 to 5 pmol of each primer ( DNA Technology, Aarhus,
Denmark ), 200 M each of the four dNTPs ( Amersham
Pharmacia Biotech., London, UK ), 0.4 units AmpliTaq DNA
Polymerase ( Applied Biosystems, Foster City, CA, USA ),
and 50 ng DNA template. The forward primers were 50 labeled with HEX, TET, or 6 - FAM fluorochromes. Three
primer pairs were multiplexed in each PCR.
The PCR was carried out in a 96 - well format using an MJ
PTC - 200 thermocycler ( MJ Research, Watertown, MA,
USA ). Two minutes of denaturation at 948C was followed
by 27 cycles of 30 seconds denaturation at 948C, 75 seconds
annealing at 558C, and 15 seconds elongation at 728C,
before 6 minutes final extension at 728C.
Detection of PCR products PCR products from two
multiplex reactions ( 20.8 l ) were pooled to allow
capillary electrophoresis of six loci simultaneously. This
was further mixed with 0.5 l GeneScan - 350 [ TAMRA ]
Size Standard ( Applied Biosystems, Foster City, CA,
USA ) in 12 l deionized formamide, CH3NO ( Kodak
Eastman Chemical, New Haven, CT, USA ), followed by
capillary electrophoresis on an ABI PRISM 310 Genetic
Analyzer ( Applied Biosystems, Foster City, CA, USA ).
The samples were electrokinetically injected for 1 to 20
seconds into a 4750 m capillary, and electrophoresed
at 15 kV for 23 minutes. The resulting electropherograms
represented relative intensities of four different fluorescent
dyes with respect to electrophoresis time ( i.e., sizes of
DNA fragments ). The softwares GeneScan 3.1 and
GenoTyper 2.1 ( Applied Biosystems, Foster City, CA,
USA ) were used to analyze the electropherograms, before
the allele peak heights were further exported to Microsoft
Excel.
Determination of allelic imbalance ( AI ) and loss of heterozygosity ( LOH ) A semiquantitative expression of AI, Q LOH,
was calculated as the ratio of the allele intensity ratios in
tumor and blood ( constitutional ) DNA, as in [ tumor allele 1 /
Skotheim et al.
197
tumor allele 2 ] / [ blood allele 1 / blood allele 2 ] . When this
value was greater than one, Q LOH was set to be the inverse.
For designation of AI at a locus, we required two independent
amplifications of the specific marker where both showed
Q LOH values less than or equal to 0.84 [ 30 ] . The mean
Q LOH value was used further. The 0.84 cut - off value was
determined due to the standard deviation of Q LOH among
samples with retained heterozygosity ( SD = 0.083 ). This
gives a probability of 99.75% that a scored AI is real, and not
due to technical error, given independence between the
errors of repeated PCRs [ 30 ] .
The TGCTs comprise a heterogeneous group of
neoplasms, both with respect to different tumor components, and the varying presence of normal cells in the
tumor biopsies. These factors must be considered when
scoring LOH. In the present study, LOH was scored when
Q LOH was less than or equal to the estimated fraction of
normal cells in the tumor biopsy. The latter is of course
somewhat subjective, but still, this way of LOH scoring is
safer than the usual practice, designating all tumors as
LOH if their Q LOH values are below a certain fixed
threshold value. However, no matter how low the Q LOH
value is, it is still possible that it reflects gain of one allele,
and not loss of the other. Therefore, we obtained
additional information on the nature of our AIs by
comparing our results to those of a separate study,
analyzing 33 of the same tumors by comparative genomic
hybridization ( CGH ) [ 31 ] .
Determination of the results for the X chromosome markers
An AI approach is not possible for investigation of X
chromosome markers in male tumors because of their
constitutional hemizygosity. Together with the X markers,
we therefore co - amplified an autosomal reference marker
with Q LOH value known to be close to 1.00. We then
compared the peak heights of the X markers with the peak
heights of the reference, in both blood and tumor DNA, to
see whether the X markers were over - or underrepresented in tumor DNA, compared to the reference. The
results were always confirmed by a second independent
PCR.
Results
Analysis of AI and LOH at Autosomal Loci
Forty - seven TGCTs were analyzed for AI and LOH at 22
autosomal polymorphic loci covering four autosomal candidate regions for TGCT susceptibility. The distributions of the
tumors’ average Q LOH values are shown in Figure 2. The
frequencies of tumors showing alterations ( i.e., confirmed
Q LOH0.84 ) at one or more loci at 3q, 5q, 12q, or 18q were
79%, 36%, 53%, and 43%, respectively ( Table 1 ). The
frequency of changes in the 3q region was significantly
higher than for each of the 5q, 12q, and 18q regions
( P < .001, P = .009, and P < .001, respectively ).
LOH was found in 32%, 21%, 9%, and 28% of the tumors
at the 3q, 5q, 12q, and 18q loci, respectively ( Table 1 ). The
Neoplas ia
Vol. 3, No. 3, 2001
Genetic Changes in Testicular Germ Cell Tumors
Skotheim et al.
199
Figure 2. The distribution of Q LOH ( x - axis ) for each of the analyzed autosomal regions. For all tumors, an average Q LOH value was found along each of the four
investigated autosomal regions, and the distributions are shown in the histograms above, e.g., if a tumor showed 0.39, 0.43, and 0.41 for the three loci at
chromosome arm 3q, the average value of 0.41 contributed to the bar representing Q LOH values from 0.4 to 0.5 in the 3q histogram. The y - axis shows the number of
tumors in each histogram group. The figure illustrates the infrequent LOH at chromosome arm 12q, where only one average Q LOH value is less than 0.5. For the
chromosome arms 3q, 5q, and 18q, there were 10, 9, and 11 tumors with average Q LOH values less than 0.5, respectively. Few tumors have average Q LOH value
near 1.0 at 3q loci, in contrast to the other regions.
LOH frequency at the 12q loci was significantly lower than for
each of the 3q, 5q, and 18q regions ( P = .005, .05, and .02,
respectively ).
Breakpoints in the AI / LOH pattern within the investigated
regions were seen in six tumors. At the 5q region, two tumors
showed retained heterozygosity at D5S644, but AI at the
more distal markers. At 12q, one tumor revealed retained
heterozygosity at D12S81, but increasingly stronger AI
toward the telomere ( Figure 3A ). Another tumor showed AI
at D12S324, but retained heterozygosity for the flanking
markers. At 18q, two tumors showed either AI or retained
heterozygosity at D18S57, and LOH or AI at the more distal
markers, respectively.
Familial / Bilateral versus Sporadic
The overall frequencies of tumors showing AI or LOH in all
investigated regions were 51% among the familial / bilateral
and 55% among the sporadic tumors. No significant differences were seen comparing the familial / bilateral and the
sporadic tumors for genetic changes within the individual
regions ( Table 1 ). For 3q and 12q, AI / LOH was found in
85% and 59% of the sporadic tumors, whereas 70% and
45%, respectively, among the familial / bilateral ( P = .21 and
P = .33 ).
Seminomas versus Nonseminomas
The overall number of changes was significantly higher
among the nonseminomas than for the seminomas
( P < .001 ). The frequencies of genetic changes at 3q and
12q in nonseminomas ( 100% and 79%, respectively ) were
significantly higher than in seminomas ( 64% and 36%;
P = .003 and P = .004, respectively ).
Analysis of X Chromosome Loci
Thirty - eight of the 47 pairs of blood / tumor DNA were
analyzed at five loci on the X chromosome. In general, the
peak heights showed increased values from blood to tumor,
relative to their co - amplified autosomal reference markers.
Though heterozygous ( Q LOH > 0.84 ), the reference markers
may still have altered copy numbers in tumor, and thus, the X
markers’ status as gained or lost is not definite by this
approach. More interesting are the observed breakpoints
between the peak heights of neighboring X chromosome
markers, relative to their common reference marker. Five
tumors with such breakpoints were seen, and altogether,
Table 1. Frequencies of Tumors Showing AI, LOH, and the Total Frequency of Change ( AI + LOH ).
3q
5q
12q
18q
All tumors
( n = 47 ) ( % )
Familial / bilateral
( n = 20 ) ( % )
Sporadic
( n = 27 ) ( % )
Seminomas
( n = 28 ) ( % )
Nonseminomas
( n = 19 ) ( % )
AI
47
30
60
36
LOH
32
40
26
29
63
37
Total
79
70
85
64
100
AI
15
20
11
18
11
LOH
21
25
19
18
26
Total
36
45
30
36
37
AI
45
40
48
29
68
11
LOH
9
5
11
7
Total
53
45
59
36
79
AI
15
15
15
14
16
LOH
28
25
30
21
37
Total
43
40
44
36
53
Neoplasia . Vol. 3, No. 3, 2001
Genetic Changes in Testicular Germ Cell Tumors
Skotheim et al.
199
Figure 3. Chromosome 12 alterations in a mixed TGCT. ( A ) The electropherograms of three markers amplified in blood ( constitutional ) and tumor DNA show the
allele intensities in relative fluorescence units ( y - axis and peak heights in boxes below the alleles ). The tumor showed gradually stronger AI ( decreasing Q LOH )
toward the distal 12q loci. A second PCR of the same markers and templates confirmed the results, and showed Q LOH values of 0.20, 0.97, and 0.77. The fourth
investigated 12q marker, D12S357 ( not shown ), was constitutionally homozygous, and thus not informative. ( B ) CGH of the tumor showed gain of the whole
chromosome with additional amplification of two regions. The central curve shows the average fluorescence ratio of 14 chromosomes between tumor and reference
DNA, whereas the two flanking curves represent the 95% confidence interval. The gain of the short arm might reflect the isochromosome 12p, a frequent and
characteristic aberration in germ cell tumors, but interestingly, the distal part of the long arm is also amplified. A Q LOH value of 0.19 ( as for D12S367 ) will almost
exclusively, in any AI / LOH study, be interpreted as LOH. However, upon comparison with CGH data, we see that the AI in this tumor is most likely caused by
amplification of genetic material ( complete CGH — copy number karyotypes — for all tumors will be published elsewhere; Ref. [ 31 ] ).
they showed increased gain toward the more distal markers
( Figure 4 ).
Figure 4. Closing in on TGCT1. This panel shows the pattern of the five
tumors with breakpoints in their X marker peak heights, compared to their
reference peak heights. The filled circles indicate markers with increased gain
compared to their neighboring markers ( open circles ). DXS1193 was the only
marker showing increased gain in all these tumors.
Discussion
AI in TGCTs
AI studies of TGCT are complicated by tumor heterogeneity, where both tumor tissues of different histologic types
and also normal cells often are intermingled. Furthermore,
the tumor can be genetically heterogeneous within a
morphologically homogeneous component. Thus, AI can be
the result of LOH masked by both normal cells and by other
subclones of the tumor with retained heterozygosity. Qualitative and semiquantitative histological examinations of
tumor cross sections ensure better interpretation of AI
studies. In this study, the percentage of intact tumor tissue
was estimated in all biopsies used for DNA isolation, and this
was taken into account when scoring LOH among the AI
cases. However, detection of AI can also reflect gain of one of
the alleles. For better interpretation of the AIs in our study, we
therefore compared our data with the results of a separate
study [ 31 ] , where 33 of the same tumors were analyzed by
CGH. That study showed net loss at chromosome arms 5q
and 18q, in 48% and 52% of the tumors, respectively, and
none of the tumors revealed gain. At distal 12q and 3q, 60%
and 12% of the tumors showed gain, and none showed loss.
When comparing these results with the present study, one
Neoplas ia
Vol. 3, No. 3, 2001
Genetic Changes in Testicular Germ Cell Tumors
Skotheim et al.
201
should bear in mind the different resolutions inherent in the
two methods. Furthermore, skewed intensities between the
homologues, but unchanged overall copy number ( as is the
case with uniparental disomy ), are only detected by the AI
approach, whereas simultaneous gain of both homologues is
only revealed by CGH.
We reported at the 2000 AACR Annual Meeting [ 32 ] the
high frequencies of AI at 3q, 5q, 12q, and 18q. This was
recently confirmed by Faulkner et al. [ 33 ] who reported
frequencies of LOH comparable to our frequencies of AI.
However, this study did not take into account that the
imbalances also might reflect gain of genetic material.
reflects gain, rather than loss, of genetic material. The
significantly higher proportion of AI among nonseminomas
than among seminomas indicates that this gain is involved in
the progression of seminomas into nonseminomas.
The results of the ITCLC study showed increasing
linkage evidence for the 12q markers as they approached
the telomere [ 19 ] . This correlates with the gain at distal
12q seen by CGH in some of the tumors ( Figure 3B ).
Microsatellites are underrepresented in subtelomeric
regions [ 38 ] , and analyses of more distal markers could
reveal even stronger evidence of linkage in the ITCLC
study.
Frequent Changes at 3q are Due to Both Loss and Trisomy
The overall frequency of AI was significantly higher for
the 3q loci than for any other investigated loci. However,
only 4 of 33 TGCTs investigated by CGH showed changes
( all gain ) at distal 3q. This indicates that AI detected at 3q
usually reflects trisomy, which will be hidden by a CGH
approach by a near triploid background. This is in keeping
with previous cytogenetic findings [ 5 ] . Furthermore, studies
have indicated trisomy 3 to be more frequent in nonseminomas than in seminomas [ 5,7 ] , which is in agreement with our AI data for the 3q loci. However, the
investigated 3q loci also showed the highest frequencies
of LOH, indicating that a substantial share of the changes at
3q is not caused by trisomy. In addition, the group of tumors
with lowest Q LOH values was not correlated with any
aberration seen by CGH, indicating the involvement of a
relatively small region. Because the seminomas and nonseminomas both show similar and high frequencies of LOH
for the 3q markers, the loss of genetic sequences on
chromosome 3 is most likely an early event in TGCT
development, and the 3q27 - ter candidate region may
harbor a TGCT suppressor gene.
AI / LOH at Syntenic Loci
Due to the low number of breakpoints in the AI / LOH
pattern, we have not defined any smallest region of
overlapping imbalances within the autosomal regions. This
might indicate that it is not the loss or gain of one
particular gene, but the unison copy number change of
several genes along a chromosomal region that is
important for TGCT development. The clustering of known
tumor - suppressor genes at 5q21 and 18q21 supports this
theory.
AI at 5q and 18q is Due to Loss of Genetic Material
Our AI data, together with the corresponding CGH results,
give evidence that the frequent imbalances at 5q and 18q
result from loss of genetic material. The observed LOH
frequencies at these genomic regions are in keeping with
previous studies [ 34 – 36 ] , and are similar in both seminomas and nonseminomas. Thus, loss of genetic material
from these regions appears to be an early event in the TGCT
development.
AI at 12q Loci is Due to Gain of Genetic Material
Isochromosome 12p, i( 12p ), is present in more than 80%
of human TGCTs [ 7 ] . However, Rodriguez et al. [ 7 ]
hypothesized that the pathogenetic trigger in TGCT is not
the gain of 12p, but the simultaneous loss of a putative
tumor - suppressor gene at 12q. LOH has previously been
reported in 50% of TGCTs at one or more loci along 12q [ 37 ] .
In the present study, we show a similar frequency of AI ( 55% )
at 12q loci. However, the frequent gain of 12q sequences
seen by CGH, and not a single event of loss, together with
significantly lower frequencies of LOH than at all the other
investigated regions, suggests that AI scored at 12q loci
Neoplasia . Vol. 3, No. 3, 2001
Closing in on TGCT1?
Our results on the X chromosome are in agreement with
molecular cytogenetic studies on TGCT, showing a general
overrepresentation of the X chromosome in the tumor DNA
[ 39 – 41 ] , and thus indicating the existence of one or more
genes on the X chromosome which, upon up - regulation,
contributes to TGCT development. Recently, Rapley et al.
[ 20 ] found evidence for a TGCT susceptibility locus,
TGCT1, at Xq27, between the markers DXS8028 and
FMR1Di ( 2.5 female cM proximal to DXS1215 ). However,
this region was limited by only one recombination event on
each side of the region. Five of our investigated tumors
showed breakpoints in the amplification level among the
investigated X chromosome markers ( Figure 4 ). Although
speculative, one may hypothesize from these somatic
changes that TGCT1 may have a more distal map position,
or a second target gene is present on Xq, distal of DXS1215
and TGCT1.
Similar Frequencies of Genetic Changes between Familial /
Bilateral and Sporadic TGCTs Speak in Disfavor of One
Single Susceptibility Gene
A segregation analysis on TGCT families and an
analysis based on the frequency of bilateral disease gave
evidence for an autosomal recessive inheritance mode
[ 18,42 ] . Individuals with familial / bilateral TGCT may thus
have inherited two inactive alleles of a tumor - suppressor
gene with limited penetrance. Those with sporadic TGCT
are then thought to be heterozygous for the gene, and
somatic mutation, imprinting, and loss are possible second
steps in the total inactivation of the tumor - suppressor
gene.
The fact that none of the candidate regions showed
significantly different frequencies of genetic changes
Genetic Changes in Testicular Germ Cell Tumors
between the familial / bilateral and the sporadic tumor
groups speaks in disfavor of the existence of one single
TGCT susceptibility gene. However, the high frequencies
of genetic changes within the investigated regions
suggest their importance in the development of primary
TGCTs. One may hypothesize that different genes
located within the different candidate regions are
responsible for the predisposition in different individuals,
or that several genes together give an elevated risk of
TGCT.
Based on the model seminomas arise from carcinomas
in situ, and may develop into nonseminomas [ 5 ] , our data
suggest that gain of genetic material at distal Xq, and
losses at 5q and 18q, contribute to establishment of
seminomas, whereas imbalances at 3q and gain at distal
part of 12q are associated with further progression into
nonseminomas.
Acknowledgements
R. I. S. and S. M. K. are research fellows of the Research
Council of Norway and the Norwegian Cancer Society
( NCS ), respectively.
References
[1] Devesa SS, Blot WJ, Stone BJ, Miller BA, Tarone RE, and Fraumeni
JFJ ( 1995 ). Recent cancer trends in the United States. J Natl Cancer
Inst 87, 175 – 182.
[2] Ekbom A, and Akre O ( 1998 ). Increasing incidence of testicular cancer
— birth cohort effects. APMIS 106, 225 – 229.
[3] Hoff WE, Tretli S, and Fosså SD ( 1995 ). Trends in incidence of
testicular cancer in Norway 1955 – 1992. Eur J Cancer 31A, 2044 –
2048.
[4] Raghavan D, Sullivan AL, Peckham MJ, and Neville AM ( 1982 ).
Elevated serum alpha - fetoprotein and seminoma: clinical evidence for
a histologic continuum? Cancer 50, 982 – 989.
[5] De Jong B, Oosterhuis JW, Castedo SM, Vos A, and te Meerman GJ
( 1990 ). Pathogenesis of adult testicular germ cell tumors. A cytogenetic model. Cancer Genet Cytogenet 48, 143 – 167.
[6] Atkin NB, and Baker MC ( 1982 ). Specific chromosome change, i( 12p ),
in testicular tumours? Lancet 2, 1349.
[7] Rodriguez E, Mathew S, Reuter V, Ilson DH, Bosl GJ, and Chaganti RS
( 1992 ). Cytogenetic analysis of 124 prospectively ascertained male
germ cell tumors. Cancer Res 52, 2285 – 2291.
[8] Castedo SM, de Jong B, Oosterhuis JW, Seruca R, Idenburg VJ, Dam
A, te MG, Koops HS, and Sleijfer DT ( 1989 ). Chromosomal changes in
human primary testicular nonseminomatous germ cell tumors. Cancer
Res 49, 5696 – 5701.
[9] Castedo SM, de Jong B, Oosterhuis JW, Seruca R, te MG, Dam A, and
Schraffordt KH ( 1989 ). Cytogenetic analysis of ten human seminomas.
Cancer Res 49, 439 – 443.
[10] Lothe RA, Peltomäki P, Tommerup N, Fosså SD, Stenwig AE,
Børresen AL, and Nesland JM ( 1995 ). Molecular genetic changes in
human male germ cell tumors. Lab Invest 73, 606 – 614.
[11] Murty VV, Bosl GJ, Houldsworth J, Meyers M, Mukherjee AB, Reuter V,
and Chaganti RS ( 1994 ). Allelic loss and somatic differentiation in
human male germ cell tumors. Oncogene 9, 2245 – 2251.
[12] Chaganti RS, Rodriguez E, and Bosl GJ ( 1993 ). Cytogenetics of male
germ cell tumors. Urol Clin North Am 20, 55 – 66.
[13] Heimdal K, Olsson H, Tretli S, Flodgren P, Børresen AL, and Fosså SD
( 1996 ). Familial testicular cancer in Norway and southern Sweden. Br J
Cancer 73, 964 – 969.
[14] Forman D, Oliver RT, Brett AR, Marsh SG, Moses JH, Bodmer JG,
Chilvers CE, and Pike MC ( 1992 ). Familial testicular cancer: a report of
the UK family register, estimation of risk and an HLA class 1 sib - pair
analysis. Br J Cancer 65, 255 – 262.
Skotheim et al.
203
[15] Goldgar DE, Easton DF, Cannon - Albright LA, and Skolnick MH
( 1994 ). Systematic population - based assessment of cancer risk in
first - degree relatives of cancer probands. J Natl Cancer Inst 86, 1600 –
1608.
[16] Tollerud DJ, Blattner WA, Fraser MC, Brown LM, Pottern L, Shapiro E,
Kirkemo A, Shawker TH, Javadpour N, and O’Connell K ( 1985 ).
Familial testicular cancer and urogenital developmental anomalies.
Cancer 55, 1849 – 1854.
[17] Reinberg Y, Manivel JC, Zhang G, and Reddy PK ( 1991 ). Synchronous
bilateral testicular germ cell tumors of different histologic type.
Pathogenetic and practical implications of bilaterality in testicular germ
cell tumors. Cancer 68, 1082 – 1085.
[18] Nicholson PW, and Harland SJ ( 1995 ). Inheritance and testicular
cancer. Br J Cancer 71, 421 – 426.
[19] International Testicular Cancer Linkage Consortium ( 1998 ). Candidate
regions for testicular cancer susceptibility genes. APMIS 106, 64 – 70.
[20] Rapley EA, Crockford GP, Teare D, Biggs P, Seal S, Barfoot R,
Edwards S, Hamoudi R, Heimdal K, Fosså SD, Tucker K, Donald J,
Collins F, Friedlander M, Hogg D, Goss P, Heidenreich A, Ormiston W,
Daly PA, Forman D, Oliver TD, Leahy M, Huddart R, Cooper CS,
Bodmer JG, and Easton DF ( 2000 ). Localization to Xq27 of a
susceptibility gene for testicular germ cell tumours. Nat Genet 24,
197 – 200.
[21] Sobin LH ( 1977 ). Standardization and the histopathology of tumours.
Histopathology 1, 87 – 92.
[22] Kunkel LM, Smith KD, Boyer SH, Borgaonkar DS, Wachtel SS, Miller
OJ, Breg WR, Jones HWJ, and Rary JM ( 1977 ). Analysis of human Y chromosome – specific reiterated DNA in chromosome variants. Proc
Natl Acad Sci USA 74, 1245 – 1249.
[23] Leahy MG, Tonks S, Moses JH, Brett AR, Huddart R, Forman D,
Oliver RT, Bishop DT, and Bodmer JG ( 1995 ). Candidate regions for
a testicular cancer susceptibility gene. Hum Mol Genet 4, 1551 –
1555.
[24] GDB ( TM ) Human Genome Database ( database online ). Johns
Hopkins University, Baltimore, Maryland, USA, 1990. Updated daily.
Available from Internet: URL http: / / www.gdb.org / .
[25] Dib C, Faure S, Fizames C, Samson D, Drouot N, Vignal A,
Millasseau P, Marc S, Hazan J, Seboun E, Lathrop M, Gyapay G,
Morissette J, and Weissenbach J ( 1996 ). A comprehensive genetic
map of the human genome based on 5264 microsatellites. Nature
380, 152 – 154.
[26] Goss PE, and Bulbul MA ( 1990 ). Familial testicular cancer in five
members of a cancer - prone kindred. Cancer 66, 2044 – 2046.
[27] Spirio L, Joslyn G, Nelson L, Leppert M, and White R ( 1991 ). A CA
repeat 30 – 70 kB downstream from the adenomatous polyposis coli
( APC ) gene. Nucleic Acids Res 19, 6348.
[28] Leppert M, Dobbs M, Scambler P, O’Connell P, Nakamura Y, Stauffer
D, Woodward S, Burt R, Hughes J, and Gardner E ( 1987 ). The gene
for familial polyposis coli maps to the long arm of chromosome 5.
Science 238, 1411 – 1413.
[29] Bodmer WF, Bailey CJ, Bodmer J, Bussey HJ, Ellis A, Gorman P,
Lucibello FC, Murday VA, Rider SH, and Scambler P ( 1987 ).
Localization of the gene for familial adenomatous polyposis on
chromosome 5. Nature 328, 614 – 616.
[30] Skotheim RI, Diep CB, Kraggerud SM, Jakobsen KS, and Lothe RA
( 2001 ). Evaluation of loss of heterozygosity / allelic imbalance in tumor
DNA. Cancer Genet Cytogenet 127(1), 64 – 70.
[31] Kraggerud SM, Skotheim RI, Szymanska J, Eknæs M, Fosså SD,
Stenwig AE, Peltomäki P, and Lothe RA ( 2001 ). Familial / bilateral and
sporadic testicular germ cell tumors analyzed by comparative genomic
hybridization. Submitted.
[32] Skotheim RI, Kraggerud SM, Stenwig AE, Fosså SD, and Lothe RA
( 2000 ). Testicular germ cell tumors show frequent genetic changes
at 3q, 5q, 12q and 18q by allelic imbalance studies and comparative
genomic hybridization. Proc Am Assoc Cancer Res 41, 420
( Abstract ).
[33] Faulkner SW, Leigh DA, Oosterhuis JW, Roelofs H, Looijenga LH, and
Friedlander ML ( 2000 ). Allelic losses in carcinoma in situ and testicular
germ cell tumours of adolescents and adults: evidence suggestive of
the linear progression model. Br J Cancer 83, 729 – 736.
[34] Peng HQ, Liu L, Goss PE, Bailey D, and Hogg D ( 1999 ). Chromosomal
deletions occur in restricted regions of 5q in testicular germ cell cancer.
Oncogene 18, 3277 – 3283.
[35] Murty VV, Li RG, Houldsworth J, Bronson DL, Reuter VE, Bosl GJ, and
Chaganti RS ( 1994 ). Frequent allelic deletions and loss of expression
characterize the DCC gene in male germ cell tumors. Oncogene 9,
3227 – 3231.
Neoplas ia
Vol. 3, No. 3, 2001
Genetic Changes in Testicular Germ Cell Tumors
Skotheim et al.
[36] Peng HQ, Bailey D, Bronson D, Goss PE, and Hogg D ( 1995 ). Loss of
heterozygosity of tumor - suppressor genes in testis cancer. Cancer
Res 55, 2871 – 2875.
[37] Murty VV, Houldsworth J, Baldwin S, Reuter V, Hunziker W, Besmer P,
Bosl G, and Chaganti RS ( 1992 ). Allelic deletions in the long arm of
chromosome 12 identify sites of candidate tumor - suppressor genes in
male germ cell tumors. Proc Natl Acad Sci USA 89, 11006 – 11010.
[38] Weissenbach J ( 1993 ). Microsatellite polymorphisms and the genetic
linkage map of the human genome. Curr Opin Genet Dev 3, 414 –
417.
[39] Summersgill B, Goker H, Weber - Hall S, Huddart R, Horwich A, and
Shipley J ( 1998 ). Molecular cytogenetic analysis of adult testicular
germ cell tumours and identification of regions of consensus copy
number change. Br J Cancer 77, 305 – 313.
Neoplasia . Vol. 3, No. 3, 2001
203
[40] Rosenberg C, Schut TB, Mostert M, Tanke H, Raap A, Oosterhuis JW,
and Looijenga L ( 1999 ). Chromosomal gains and losses in testicular
germ cell tumors of adolescents and adults investigated by a modified
comparative genomic hybridization approach. Lab Invest 79, 1447 –
1451.
[41] Mostert MM, van de Pol M, Olde WD, Suijkerbuijk RF, Geurts vK, van
Echten J, Oosterhuis JW, and Looijenga LH ( 1996 ). Comparative
genomic hybridization of germ cell tumors of the adult testis:
confirmation of karyotypic findings and identification of a 12p amplicon. Cancer Genet Cytogenet 89, 146 – 152.
[42] Heimdal K, Olsson H, Tretli S, Fosså SD, Børresen AL, and
Bishop DT ( 1997 ). A segregation analysis of testicular cancer
based on Norwegian and Swedish families. Br J Cancer 75, 1084 –
1087.
Paper III
Kraggerud SM, Skotheim RI, Szymanska J, Eknæs M, Fosså SD, Stenwig AE,
Peltomäki P, and Lothe RA
Genome profiles of familial/bilateral and sporadic testicular germ cell tumors
Genes Chromosomes and Cancer, 2002, 34(2): 168-174
blank
GENES, CHROMOSOMES & CANCER 34:168 –174 (2002)
DOI 10.1002/gcc.10058
Genome Profiles of Familial/Bilateral and Sporadic
Testicular Germ Cell Tumors
Sigrid Marie Kraggerud,1 Rolf I. Skotheim,1 Jadwiga Szymanska,2 Mette Eknæs,1 Sophie D. Fosså,3
Anna E. Stenwig,4 Päivi Peltomäki,2 and Ragnhild A. Lothe1*
1
Department of Genetics, Institute for Cancer Research, The Norwegian Radium Hospital, Oslo, Norway
Division of Human Cancer Genetics, Comprehensive Cancer Center, Ohio State University, Columbus, Ohio
3
Department of Oncology and Radiotherapy, The Norwegian Radium Hospital, Oslo, Norway
4
Department of Pathology, The Norwegian Radium Hospital, Oslo, Norway
2
In order to investigate the genetics of testicular germ cell tumors (TGCTs), we examined 33 TGCTs, including 15
familial/bilateral and 18 sporadic tumors, using comparative genomic hybridization. The frequencies of the histological subtypes
were comparable between the two groups. Gains of the whole or parts of chromosome 12 were found in 30 tumors (91%).
Furthermore, increased copy number of the whole or parts of chromosomes 7, 8, 17, and X, and decreased copy number of
the whole or parts of chromosomes 4, 11, 13, and 18 were observed in 50% of the tumors. Sixteen smallest regions of
overlapping changes were delineated on 12 different chromosomes. The chromosomal copy numbers of familial/bilateral and
sporadic TGCTs were comparable, suggesting similar genetic pathways to disease in both groups. However, significant
differences were observed between the two main histological subgroups. Gains from 15q and 22q were associated with
seminomas (P ⫽ 0.005 and P ⫽ 0.02, respectively), whereas gain of the proximal 17q (17q11.2–21) and high-level amplification
from chromosome arm 12p, and losses from 10q were associated with nonseminomas (P ⬍ 0.001, P ⫽ 0.04, and P ⫽ 0.03,
respectively).
© 2002 Wiley-Liss, Inc.
INTRODUCTION
Testicular germ cell tumor (TGCT) is the most
common malignancy among young white males,
and the incidence is increasing (Devesa et al., 1995;
Bergstrom et al., 1996). Although most TGCTs are
sporadic, some are bilateral and/or associated with a
positive family history of TGCT, suggesting hereditary predisposition. So far, susceptibility genes
for TGCT are not known, although suggestive
linkage between disease and genetic markers has
been reported at loci on 3q, 5q, 12q, 18q, and Xq
(International Testicular Cancer Linkage Consortium, 1998; Rapley et al., 2000). TGCTs are subdivided into two main histological entities, seminomasB andB nonseminomas.B SeminomasB are
believed to arise from a carcinoma in situ stage, and
may develop into nonseminomas (de Jong et al.,
1990). Seminomas and nonseminomas are typically
hypertriploid and hypotriploid, respectively. Irrespective of histological subtype, TGCTs are characterized by the presence of isochromosome 12p.
Overrepresentation of chromosomes 7, 8, 12, and X
and underrepresentation of chromosomes 11, 13,
18, and Y (Sandberg et al., 1996) are also frequently
seen. Whole genome analyses by comparative
genomic hybridization (CGH) have been published for some TGCTs (Korn et al., 1996; Mostert
et al., 1996; Ottesen et al., 1997; Summersgill et al.,
1998; Rosenberg et al., 1999). Although several
©
2002 Wiley-Liss, Inc.
recurrent copy number changes have been reported for TGCTs in general, to our knowledge, no
familial tumors and only 3 cases of bilateral tumor
(Ottesen et al., 1997) have previously been analyzed by CGH. In the present study, the copy
number karyotypes obtained by CGH in a series of
familial/bilateral TGCTs were compared with
those of sporadic TGCTs to evaluate and compare
their genetic constituents.
MATERIALS AND METHODS
Patients and Samples
The tumor material consisted of 33 freshly frozen TGCTs, obtained from 33 adult Norwegian
TGCT patients. The tumors were grouped into
familial and/or bilateral (n ⫽ 15) and sporadic
TGCTs (n ⫽ 18). Among 15 patients with familial/
bilateral TGCTs, 8 had bilateral cancer, 9 had
affected family members (first-degree relatives: 8
cases; second-degree: 1 case), and thus 2 patients
had both bilateral tumors and affected family members. Cryptorchidism was reported for patients
Supported by: The Norwegian Cancer Society and the Research
Council of Norway.
*Correspondence to: Ragnhild A. Lothe, Department of Genetics,
Institute for Cancer Research, The Norwegian Radium Hospital,
Montebello, N-0310 Oslo, Norway. E-mail: [email protected]
Received 2 October 2001; Accepted 27 November 2001
FAMILIAL/BILATERAL VS. SPORADIC TGCTs
within both groups (2 in the familial/bilateral and 1
in the sporadic group). Median age at diagnosis was
31 years for the patients in the familial/bilateral
group and 28.5 for the sporadic (average ages: 32
and 31 years, respectively).
Three 5 ␮m sections were taken from different
parts of each tumor sample before DNA isolation. The sections were stained with hematoxylin and eosin (HE) and visually evaluated by
light microscopy. The various tumor components
were described according to recommendations of
the WHO (Mostofi and Sobin, 1976), and the
percentage of intact neoplastic tissue was estimated for each of the three HE-stained sections.
The familial/bilateral and sporadic tumor groups
were comparable according to histology and percentage of intact tumor tissue. Among the 33
TGCT samples, 20 were classified as seminomas:
15 were pure seminomas (7 familial/bilateral and
8 sporadic tumors), and 5 were seminoma components from combined tumors (2 familial/bilateral and 3 sporadic tumors). Thirteen of the
tumors were nonseminomas (6 familial/bilateral
and 7 sporadic tumors). The percentage of intact
tumor tissue for all tumors was, on average, 73%
(range: 30 –100%). Only 6 of the 33 TGCTs were
estimated to have less than 50% tumor tissue.
Despite an apparently low tumor percentage, all
6 showed aberrations by CGH. The 5 seminoma
samples from combined cases, in which only the
seminoma component was present in the frozen
biopsy used for DNA isolation, were included in
the seminoma group for statistical analyses.
Twenty-two of the tumors were stage I (12 pure
seminomas, 5 seminoma samples from combined
cases, and 5 nonseminomas), whereas the remaining 11 tumors were stages II, III, and IV
(n ⫽ 6, 2, and 3, respectively, of which 3 were
pure seminomas and 8 nonseminomas).
DNA Isolation and Measurements
DNA was isolated from freshly frozen TGCT
tissue (n ⫽ 33), peripheral blood lymphocytes both
from a healthy male donor (reference DNA) and a
healthy female donor (negative control), and a tumor with known aberrations [case 347 in Lothe et
al., 1995] (positive control) by extraction with phenol/chloroform followed by ethanol precipitation.
The DNA concentration of each sample was measured in a 1 ⫻ 10⫺4 mg/ml Hoechst solution
(Hoechst 33258) with a TKO 100 fluorometer
(Hoefer Scientific Instruments, San Francisco,
CA).
169
Comparative Genomic Hybridization
The CGH method, initially described by Kallioniemi et al. (1992), was used with the modifications described by Kraggerud et al. (2000). Briefly,
test and reference DNA were labeled in a nicktranslation reaction with a mixture of two fluorochrome-conjugated nucleotides (FITC-12-dCTP
and FITC-12-dUTP for tumor DNA, and Texas
Red-6-dCTP and Texas Red-6-dUTP for normal
DNA). Equal amounts (1 ␮g) of labeled tumor and
reference DNA, and 20 ␮g Cot-1 DNA, were hybridized onto normal, denatured metaphase
spreads and incubated for 2–3 days at 37°C. Finally, the slides were counterstained in an antifade
solution with DAPI (4⬘,6-diamino-2-phenylindole)
and Vectashields H-1200, and analyzed in a fluorescence microscope (Zeiss Axioplan, Oberkochen,
Germany). Single-color images (FITC, Texas Red,
and DAPI) of metaphase chromosomes were sequentially acquired with a Cohu 4900 CCD (12-bit gray
scale) camera, using Cytovision software and hardware (Applied Imaging, Newcastle, UK). Chromosomes were identified based on their inverted DAPI
banding, and fluorescence ratio profiles (green to red
fluorescence) were calculated for each chromosome
before data from at least 14 representative copies of
each chromosome (range: 14 –22) were combined and
average ratio profiles with 95% confidence intervals
generated for each tumor. The Y chromosome and
centromeric and pericentromeric heterochromatic regions were not evaluated.
Upper and lower threshold values of 1.17 and
0.83, respectively, were used to determine the
gains and losses of DNA sequences. These cutoff
values correspond to gain or loss of one chromosome homolog in 50% of the cells analyzed, given
a triploid tumor genome. Chromosome regions
showing a fluorescence ratio of 2.0 or more were
classified as amplified.
Certain chromosome regions are known to be
problematic by CGH. Artificial copy number
changes have been observed at 1p33-ter, 16p, 17p,
19, and 22 (Kallioniemi et al., 1994; el-Rifai et al.,
1997). In this study, a modification of the original
CGH protocol was applied, using a mixture of fluorochrome-labeled nucleotides (fluorochromedUTP and -dCTP) during nick translation, which
ensures a reduction of false positive signals in these
“problem areas” (el-Rifai et al., 1997). In addition,
the analysis of each tumor profile was performed
relative to the profile of the negative control in the
same experiment, further supporting that the detected changes within these regions are present in
the tumor.
170
KRAGGERUD ET AL.
Figure 1. Copy number changes in TGCTs. Schematic diagram of the chromosomal gains (bars to the
right) and losses (bars to the left) identified in the 33 TGCTs analyzed by CGH. SROs are indicated with
gray boxes.
Smallest Region of Overlap (SRO)
Overlapping chromosome regions, altered in
more than 40% of the tumors (i.e., 14 tumors),
were evaluated for SROs. At least 3 tumors were
required to identify each of the two borders of an
SRO of chromosomal gain or loss. If an apparent
border was defined by less than 3 samples, the
SRO was expanded to the next “break point.” This
was the case for the proximal border of the 4q SRO,
and the 8q and Xq SROs (Fig. 1). Tumors presenting CGH chromosome profiles showing a harlequin
pattern, that is, two or more areas with gains or
losses along a given chromosome arm, were not
accepted as informative with regard to defining
SRO borders.
Statistical Analysis
Pearson chi-square tests (two-sided) were performed to evaluate the differences observed in our
results. However, when one of the expected values
was less than 5, we applied Fisher’s exact test
(two-sided). Values of P 0.05 were interpreted as
statistically significant. Statistical analyses on the
differences between the histological subgroups
(i.e., seminomas vs. nonseminomas) were performed for two alternatives of the seminoma specimens: (1) seminoma samples (i.e., both pure seminomas and the 5 seminoma samples from
combined tumors), and (2) pure seminomas.
RESULTS
The copy number changes detected by CGH in
TGCTs are summarized in Figure 1, and some
examples are given in Figure 2. Aberrations were
observed in 32 of the 33 TGCTs (97%). The
TGCTs showed, on average, 14 copy number
changes per case (range: 0 –28), representing 9
gains (range: 0 –17) and 5 losses (range: 0 –12). The
most frequent aberration, gain of chromosome arm
12p, was seen in 29 tumors (88%). Other frequent
FAMILIAL/BILATERAL VS. SPORADIC TGCTs
171
C
O
L
O
R
Figure 2. Examples of chromosomes and chromosome regions altered in TGCTs. Color images and CGH ratio profiles are presented.
The straight vertical black line in the presentation of the CGH profile
represents the fluorescence ratio equal to 1, whereas the vertical lines
to the left (red) and right (green) each represent ratio deviations of
0.25. The central line in the CGH profile shows the average fluorescence ratio along the chromosome (at least 14 homologs were evalu-
ated), and the flanking curves (brown) represent the 95% confidence
interval. To the left, examples of gains from chromosome 12 are shown.
Far left, the chromosome 12 profile from the tumor showing the
highest 12p (12p12) amplification observed in the present series. Loss
of 13q21–31 in a seminoma is shown in the midposition. Gain of
17q23-ter in a seminoma and gain of 17pter– q21 in a nonseminoma are
shown to the right.
Figure 3. Copy number changes in familial/bilateral TGCTs (n ⫽ 15) (A) and in sporadic TGCTs (n ⫽
18) (B). Chromosomal gains and losses are shown as bars to the right and bars to the left, respectively, of
the ideograms. The copy number changes in nonseminomas and seminomas are presented in blue and red,
respectively, and bold lines reflect amplifications.
changes, which appeared in 30% or more of the
TGCTs, were gains of the following whole or parts
of chromosome arms: 1p (46%), 1q (49%), 2p
(33%), 7p (70%), 7q (76%), 8p (61%), 8q (61%), 12q
(64%), 14q (30%), 17q (67%), 19p (42%), 19q
(30%), 21q (42%), 22q (39%), Xp (52%), and Xq
(52%); and losses of the whole or parts of 4p (39%),
4q (64%), 5p (36%), 5q (49%), 9p (33%), 11p (33%),
11q (64%), 13q (79%), and 18q (52%). Sixteen
SROs were identified, corresponding to gains of
1q21–23, 7p14-ter, 7q11.2–22, 8p22-ter, 8q12–23,
12p, 12q24.1-ter, 17q11.2–21, 17q24-ter, 19p13.1ter, and Xq21-ter; and losses of 4q21–27, 5q11.2–
23, 11q14 –22, 13q21–31, and 18q12-ter (Fig. 1).
Amplified regions of the whole or parts of 12p were
observed in 13 (39%) of the tumors. Seven of these
tumors showed amplification of whole 12p, and 6
tumors had amplifications restricted to 12p12-ter
and 12p13 (n ⫽ 2 and n ⫽ 4, respectively). Amplification of 14q24-ter and X were observed in one
tumor each (Fig. 3).
Familial/Bilateral vs. Sporadic Tumors
No significant differences in CGH patterns were
observed between the familial/bilateral and the
sporadic tumor groups (Fig. 3). The number of
losses and gains per tumor were comparable between the groups. The familial/bilateral cases
showed, on average, 15 changes/case (range: 0 –28),
representing 10 gains and 5 losses, whereas the
sporadic tumors revealed 13 changes/case (range:
4 –20), 8 gains and 5 losses. Although not statistically significant, certain changes were restricted to
only one of the two tumor groups. Gains from 6q
172
KRAGGERUD ET AL.
and 11q, and losses from 14q, were observed in two
familial/bilateral tumors each, whereas losses from
chromosome 20 were observed in only two sporadic
tumors. Furthermore, gain of 16p was found in 4
(27%) familial/bilateral tumors, but in only 1 (6%)
of the sporadic tumors (P ⫽ 0.15).
Nonseminomas vs. Seminomas
Statistically significant differences were observed between the two main histological subgroups. Among nonseminomas, 38% (5/13) showed
losses from 10q (overlapping region: 10q11.2–21),
whereas only 7% (1/20) of the seminoma samples, a
pure seminoma (1/15), showed loss of this region
(P ⫽ 0.03 and P ⫽ 0.07, respectively). Gains from
15q were not observed in any of the nonseminomas, but were seen in 45% (9/20) of the seminoma
samples and in 47% (7/15) of the pure seminoma
cases (P ⫽ 0.005 and P ⫽ 0.007, respectively). Gain
of the whole or parts of the proximal 17q SRO,
17q11.2–21, was observed in 85% (11/13) of the
nonseminomas, but in only 15% (3/20) of the seminoma samples and 13% (2/15) of the pure seminomas (both P ⬍ 0.001). For the SRO at distal 17q,
17q24-ter, the frequencies of gains, of the whole or
parts of this region, were not statistically different
between the groups. However, gain of this region
appeared more frequently among nonseminomas
(77%, n ⫽ 10) than among seminoma samples
(50%, n ⫽ 10) and pure seminomas (53%, n ⫽ 8).
Gain from 22q was seen in 2 (16%) nonseminomas,
11 (55%) seminoma samples, and 8 (53%) pure
seminomas (P ⫽ 0.02 and P ⫽ 0.06). Amplification
of the whole or parts of 12p was observed in 8
(62%) nonseminomas, compared to 5 (25%) seminoma samples and 4 (27%) pure seminomas (P ⫽
0.04 and P ⫽ 0.06).
DISCUSSION
Changes were observed in all but one of the
tumors, a seminoma from the familial/bilateral
group (Fig. 1). This tumor was described by the
pathologist to contain 95% tumor tissue, but with
massive lymphocyte infiltration. Thus, normal
DNA extracted from the lymphocytes might have
masked changes present in the seminoma cells. By
reviewing the previous CGH studies (Korn et al.,
1996; Mostert et al., 1996; Ottesen et al., 1997;
Summersgill et al., 1998; Rosenberg et al., 1999),
gains of the whole or parts of chromosomes 1, 2, 7,
8, 12, 14, 15, 17, 20, 21, 22, and X; and losses of the
whole or parts of chromosomes 4, 5, 11, 13, and 18
are each found in 30% of the primary tumors
analyzed, in two or more of the studies. The
present study confirmed these findings, although
gains from 15q were observed in only 27% of our
TGCTs. In addition, losses from 9p and gains from
chromosome 19 were observed in more than 30% of
the present tumor series. The changes observed at
these chromosomes are in agreement with previous
cytogenetic and single CGH studies (Castedo et
al., 1989a,b; van Echten et al., 1995; Korn et al.,
1996; Rosenberg et al., 1999). However, chromosome 19 is one of the problematic areas in CGH
analyses, and should be evaluated with caution.
Familial/bilateral and sporadic TGCTs were
found to have comparable patterns of copy number
changes (Fig. 3). However, gains from 6q and 11q
and losses from 14q were restricted to familial/
bilateral tumors, although seen in only two tumors
each. Gains from 6q were previously found in some
primary TGCTs (n ⫽ 11), whereas gains from 11q
and losses from 14q have been observed in only
one tumor each (Korn et al., 1996; Ottesen et al.,
1997; Summersgill et al., 1998; Rosenberg et al.,
1999). Only 3 bilateral cases were previously studied (Ottesen et al., 1997), and thus the importance
of these copy number changes in familial/bilateral
disease remains unknown. Gains from 6q have
been proposed to be associated with chemotherapy
resistance (Rao et al., 1998; Summersgill et al.,
1998; Hiorns et al., 1999). However, for one of the
present two cases with 6q gain, clinical patient
information was available, although no chemotherapy resistance was observed. It has been suggested
that gains of 9q22.1–22.2 and losses of 16q13–21
are associated with stage II tumors (Ottesen et al.,
1997), but in the present study the only tumor
showing these alterations was stage I, and no specific chromosomal alteration was associated with
stage II tumors.
Comparison of the genome profiles of the nonseminomas and seminomas revealed several important differences. First, gains from 15q and 22q were
significantly associated with seminomas, which is
in agreement with previous FISH (Looijenga et al.,
1993), CGH (Korn et al., 1996; Ottesen et al., 1997;
Summersgill et al., 1998), and cytogenetic studies
(Castedo et al., 1989a,b; van Echten et al., 1995).
Second, gain of the proximal SRO at 17q (17q11.2–
21) was preferentially seen in nonseminomas (P ⬍
0.001). A statistically significant association between a higher copy number of chromosome 17
and nonseminomas has been reported in a previous
cytogenetic study (van Echten et al., 1995). Gain of
17q, proximal 17q material in particular, is likely to
involve genes encoding oncogenic proteins associated with the progression from seminomas to nonseminomas. Increased copy number of 17q is also
found in other solid tumors, and amplified regions
173
FAMILIAL/BILATERAL VS. SPORADIC TGCTs
overlapping with our 17q SROs have been designated (Lothe et al., 1996; Barlund et al., 1997;
Kokkola et al., 1997). The 17q genes need further
investigation, and a chromosome 17 cDNA microarray study of our TGCTs is in progress. Third,
chromosome 10 was reported to be preferentially
lost in nonseminomas (Rosenberg et al., 1999), and
in our investigation nonseminomas were associated
with loss of an overlapping region at 10q (10q11.2–
21; P ⫽ 0.03). However, P values close to 0.05
should be treated with caution because many statistical tests were performed in this study.
Gain of 12p was the most frequent aberration,
occurring in 88% of the TGCTs. The SROs at
chromosome 12, 12p, and 12q24.1-ter were in
agreement with those previously found in ovarian
germ cell tumors (Kraggerud et al., 2000), indicating common genetic patterns in germ cell tumors of
females and males. We accepted the 12q24.1-ter
region in spite of our criterion excluding tumors
with a harlequin pattern (see Materials and Methods) from SRO determinations. This criterion was
included to ensure that SROs were not defined
from tumors with CGH profiles fluctuating around
the detection limit. However, the gain of proximal
12q material was clearly an extension of the gain of
12p, and the gain of 12q24.1-ter was a separate
region of gain not only caused by small profile
fluctuations. Gain of the distal 12q material can also
be seen in a few tumors (⬃5) from illustrations in
previous CGH reports (Ottesen et al., 1997; Rosenberg et al., 1999), but has not previously been
identified as an SRO in TGCTs. In contrast to the
fact that gain of 12p material was observed in
nearly all TGCTs, amplification of 12p sequences
was clearly associated with nonseminomas. Cytogentic studies have reported more copies of i(12p)
in nonseminomas than in seminomas (Rodriguez et
al., 1992), and high-level amplifications of 12p11.2–
12.1 have been reported by CGH (Suijkerbuijk et
al., 1994; Korn et al., 1996; Mostert et al., 1996,
1998; Rao et al., 1998). We found the highest levels
of amplification usually toward the 12p telomere,
and some tumors showed amplification of only
12p12-ter or 12p13. It has been suggested that the
amplification of 12p allows the TGCT cells to
survive outside their microenvironment and that
12p genes inhibit apoptosis (Roelofs et al., 2000),
although the actual genes involved remain unknown.
The minimal region of loss at 13q found in our
TGCTs (13q21–31) was distal to the RB1 locus.
Previous findings of loss of heterozygosity and reduced expression of RB1 in TGCTs may be because of the fact that it is often lost together with
the relevant gene(s) in the SRO at 13q21–31. In
other studies of TGCT (Korn et al., 1996; Mostert
et al., 1996; Ottesen et al., 1997; Summersgill et al.,
1998; Rosenberg et al., 1999), losses from 13q have
been reported with 13q31-ter and 13q22–32 as the
overlapping deleted regions (Mostert et al., 1996;
Rosenberg et al., 1999). Although our study did not
confirm the previously suggested association of 13q
gains with nonseminomas (Summersgill et al.,
1998), the relevance of the loss of 13q material in
tumorigenesis is supported by studies of other solid
tumors. It is reported as the most frequently lost
region in tumors from familial prostate cancer patients (Rokman et al., 2001), and aggressive prostate cancers are associated with loss of 13q21 (Dong
et al., 2000). This region may also harbor a breast
cancer susceptibility locus (Kainu et al., 2000).
In conclusion, familial/bilateral and sporadic
TGCTs were demonstrated to have the same nonrandom genome-wide pattern of copy number
changes, suggesting their development through the
same genetic pathways. The development of seminomas is strongly associated with gains from 15q
and 22q, whereas gain of 17q11.2–21 and high-level
amplifications of 12p material are important for the
progression into nonseminomas.
ACKNOWLEDGMENTS
This research was supported, in part, by grants
from the Norwegian Cancer Society (to S.D.F. and
R.A.L.). S.M.K. and R.I.S. are research fellows of
the Norwegian Cancer Society and the Research
Council of Norway, respectively.
REFERENCES
Barlund M, Tirkkonen M, Forozan F, Tanner MM, Kallioniemi O,
Kallioniemi A. 1997. Increased copy number at 17q22– q24 by
CGH in breast cancer is due to high-level amplification of two
separate regions. Genes Chromosomes Cancer 20:372–376.
Bergstrom R, Adami HO, Mohner M, Zatonski W, Storm H, Ekbom
A, Tretli S, Teppo L, Akre O, Hakulinen T. 1996. Increase in
testicular cancer incidence in six European countries: a birth
cohort phenomenon. J Natl Cancer Inst 88:727–733.
Castedo SM, de Jong B, Oosterhuis JW, Seruca R, Idenburg VJ,
Dam A, te Meerman G, Koops HS, Sleijfer DT. 1989a. Chromosomal changes in human primary testicular nonseminomatous
germ cell tumors. Cancer Res 49:5696 –5701.
Castedo SM, de Jong B, Oosterhuis JW, Seruca R, te Meerman G,
Dam A, Schraffordt Koops H. 1989b. Cytogenetic analysis of ten
human seminomas. Cancer Res 49:439 – 443.
de Jong B, Oosterhuis JW, Castedo SM, Vos A, te Meerman GJ.
1990. Pathogenesis of adult testicular germ cell tumors. A cytogenetic model. Cancer Genet Cytogenet 48:143–167.
Devesa SS, Blot WJ, Stone BJ, Miller BA, Tarone RE, Fraumeni
JFJ. 1995. Recent cancer trends in the United States. J Natl
Cancer Inst 87:175–182.
Dong JT, Chen C, Stultz BG, Isaacs JT, Frierson HFJ. 2000.
Deletion at 13q21 is associated with aggressive prostate cancers.
Cancer Res 60:3880 –3883.
el-Rifai W, Larramendy ML, Bjorkqvist AM, Hemmer S, Knuutila
S. 1997. Optimization of comparative genomic hybridization using
fluorochrome conjugated to dCTP and dUTP nucleotides. Lab
Invest 77:699 –700.
Hiorns LR, Seckl MJ, Paradinas F, Sharp SY, Skelton LA, Brun-
174
KRAGGERUD ET AL.
strom G, Newlands ES, Kelland LR, Leyland-Jones B. 1999. A
molecular cytogenetic approach to studying platinum resistance.
J Inorg Biochem 77:95–104.
International Testicular Cancer Linkage Consortium. 1998. Candidate regions for testicular cancer susceptibility genes. APMIS
106:64 –70.
Kainu T, Juo SH, Desper R, Schaffer AA, Gillanders E, Rozenblum
E, Freas-Lutz D, Weaver D, Stephan D, Bailey-Wilson J, Kallioniemi OP, Tirkkonen M, Syrjakoski K, Kuukasjarvi T, Koivisto
P, Karhu R, Holli K, Arason A, Johannesdottir G, Bergthorsson
JT, Johannsdottir H, Egilsson V, Barkardottir RB, Johannsson O,
Haraldsson K, Sandberg T. 2000. Somatic deletions in hereditary
breast cancers implicate 13q21 as a putative novel breast cancer
susceptibility locus. Proc Natl Acad Sci USA 97:9603–9608.
Kallioniemi A, Kallioniemi OP, Sudar D, Rutovitz D, Gray JW,
Waldman F, Pinkel D. 1992. Comparative genomic hybridization
for molecular cytogenetic analysis of solid tumors. Science 258:
818 – 821.
Kallioniemi OP, Kallioniemi A, Piper J, Isola J, Waldman FM, Gray
JW, Pinkel D. 1994. Optimizing comparative genomic hybridization for analysis of DNA sequence copy number changes in solid
tumors. Genes Chromosomes Cancer 10:231–243.
Kokkola A, Monni O, Puolakkainen P, Larramendy ML, Victorzon
M, Nordling S, Haapiainen R, Kivilaakso E, Knuutila S. 1997.
17q12–21 amplicon, a novel recurrent genetic change in intestinal
type of gastric carcinoma: a comparative genomic hybridization
study. Genes Chromosomes Cancer 20:38 – 43.
Korn WM, Olde Weghuis DEM, Suijkerbuijk RF, Schmidt U, Otto
T, du Manoir S, Geurts van Kessel AD, Harstrick A, Seeber S,
Becher R. 1996. Detection of chromosomal DNA gains and losses
in testicular germ cell tumors by comparative genomic hybridization. Genes Chromosomes Cancer 17:78 – 87.
Kraggerud SM, Szymanska J, Abeler VM, Kaern J, Eknaes M, Heim
S, Teixeira MR, Trope CG, Peltomaki P, Lothe RA. 2000. DNA
copy number changes in malignant ovarian germ cell tumors.
Cancer Res 60:3025–3030.
Looijenga LH, Gillis AJ, Van Putten WL, Oosterhuis JW. 1993. In
situ numeric analysis of centromeric regions of chromosomes 1,
12, and 15 of seminomas, nonseminomatous germ cell tumors, and
carcinoma in situ of human testis. Lab Invest 68:211–219.
Lothe RA, Slettan A, Saeter G, Brøgger A, Børresen AL, Nesland
JM. 1995. Alterations at chromosome 17 loci in peripheral nerve
sheath tumors. J Neuropathol Exp Neurol 54:65–73.
Lothe RA, Karhu R, Mandahl N, Mertens F, Saeter G, Heim S,
Børresen-Dale AL, Kallioniemi OP. 1996. Gain of 17q24-qter
detected by comparative genomic hybridization in malignant tumors from patients with von Recklinghausen’s neurofibromatosis.
Cancer Res 56:4778 – 4781.
Mostert MC, Verkerk AJ, van de Pol M, Heighway J, Marynen P,
Rosenberg C, van Kessel AG, van Echten J, de Jong B, Oosterhuis
JW, Looijenga LH. 1998. Identification of the critical region of
12p over-representation in testicular germ cell tumors of adolescents and adults. Oncogene 16:2617–2627.
Mostert MMC, van de Pol M, Olde Weghuis D, Suijkerbuijk RF,
Geurts van Kessel A, van Echten J, Oosterhuis JW, Looijenga
LHJ. 1996. Comparative genomic hybridization of germ cell tumors of the adult testis: confirmation of karyotypic findings and
identification of a 12p- amplicon. Cancer Genet Cytogenet 89:
146 –152.
Mostofi FK, Sobin LH. 1976. International histological classification
of tumors (No. 16): histological typing of testis tumors. Geneva:
World Health Organization.
Ottesen AM, Kirchhoff M, De-Meyts ER, Maahr J, Gerdes T, Rose
H, Lundsteen C, Petersen PM, Philip J, Skakkebaek NE. 1997.
Detection of chromosomal aberrations in seminomatous germ cell
tumours using comparative genomic hybridization. Genes Chromosomes Cancer 20:412– 418.
Rao PH, Houldsworth J, Palanisamy N, Murty VVVS, Reuter VE,
Motzer RJ, Bosl GJ, Chaganti RSK. 1998. Chromosomal amplification is associated with cisplatin resistance of human male germ
cell tumors. Cancer Res 58:4260 – 4263.
Rapley EA, Crockford GP, Teare D, Biggs P, Seal S, Barfoot R,
Edwards S, Hamoudi R, Heimdal K, Fosså SD, Tucker K, Donald
J, Collins F, Friedlander M, Hogg D, Goss P, Heidenreich A,
Ormiston W, Daly PA, Forman D, Oliver TD, Leahy M, Huddart
R, Cooper CS, Bodmer JG, Easton DF. 2000. Localization to
Xq27 of a susceptibility gene for testicular germ-cell tumours. Nat
Genet 24:197–200.
Rodriguez E, Mathew S, Reuter V, Ilson DH, Bosl GJ, Chaganti
RSK. 1992. Cytogenetic analysis of 124 prospectively ascertained
male germ cell tumors. Cancer Res 52:2285–2291.
Roelofs H, Mostert MC, Pompe K, Zafarana G, van Oorschot M, van
Gurp RJ, Gillis AJ, Stoop H, Beverloo B, Oosterhuis JW, Bokemeyer C, Looijenga LH. 2000. Restricted 12p amplification and
RAS mutation in human germ cell tumors of the adult testis. Am J
Pathol 157:1155–1166.
Rokman A, Koivisto PA, Matikainen MP, Kuukasjarvi T, Poutiainen
M, Helin HJ, Karhu R, Kallioniemi OP, Schleutker J. 2001.
Genetic changes in familial prostate cancer by comparative
genomic hybridization. Prostate 46:233–239.
Rosenberg C, Schut TB, Mostert M, Tanke H, Raap A, Oosterhuis
JW, Looijenga L. 1999. Chromosomal gains and losses in testicular germ cell tumors of adolescents and adults investigated by a
modified comparative genomic hybridization approach. Lab Invest 79:1447–1451.
Sandberg AA, Meloni AM, Suijkerbuijk RF. 1996. Reviews of chromosome studies in urological tumors. III. Cytogenetics and genes
in testicular tumors. J Urol 155:1531–1556.
Suijkerbuijk RF, Sinke RJ, Weghuis DE, Roque L, Forus A, Stellink F, Siepman A, van de Kaa C, Soares J, Geurts van Kessel A.
1994. Amplification of chromosome subregion 12p11.2–p12.1 in a
metastasis of an i(12p)-negative seminoma: relationship to tumor
progression? Cancer Genet Cytogenet 78:145–152.
Summersgill B, Goker H, Weber-Hall S, Huddart R, Horwich A,
Shipley J. 1998. Molecular cytogenetic analysis of adult testicular
germ cell tumours and identification of regions of consensus copy
number change. Br J Cancer 77:305–313.
van Echten J, Oosterhuis JW, Looijenga LH, van de Pol M,
Wiersema J, te Meerman GJ, Schaffordt Koops H, Sleijfer DT, de
Jong B. 1995. No recurrent structural abnormalities apart from
i(12p) in primary germ cell tumors of the adult testis. Genes
Chromosomes Cancer 14:133–144. AQ1: Editor: Please provide
date.
blank
Paper IV
Skotheim RI, Monni O, Mousses S, Fosså SD, Kallioniemi OP, Lothe RA, and
Kallioniemi A
New insights into testicular germ cell tumorigenesis from gene expression
profiling
Cancer Research, 2002, 62(8): 2359-2364
blank
[CANCER RESEARCH 62, 2359 –2364, April 15, 2002]
New Insights into Testicular Germ Cell Tumorigenesis from Gene
Expression Profiling1
Rolf I. Skotheim, Outi Monni, Spyro Mousses, Sophie D. Fosså, Olli-P. Kallioniemi, Ragnhild A. Lothe,2 and
Anne Kallioniemi
Department of Genetics, Institute for Cancer Research [R. I. S., R. A. L.] and Department for Oncology and Radiotherapy [S. D. F.], The Norwegian Radium Hospital, N-0310
Oslo, Norway; Biomedicum Biochip Center, Biomedicum Helsinki, FIN-00290 Helsinki, Finland [O. M.]; Cancer Genetics Branch, National Human Genome Research Institute,
NIH, Bethesda, Maryland 20892 [O. M., S. M., O-P. K., A. K.]; and Laboratory of Cancer Genetics, Institute of Medical Technology, University of Tampere and Tampere
University Hospital, FIN-33520 Tampere, Finland [A. K.]
ABSTRACT
3
We have shown recently that about half of the human TGCTs reveal
DNA copy number increases affecting two distinct regions on chromosome
arm 17q. To identify potential target genes with elevated expressions
attributable to the extra copies, we constructed a cDNA microarray
containing 636 genes and expressed sequence tags from chromosome 17.
The expression patterns of 14 TGCTs, 1 carcinoma in situ, and 3 normal
testis samples were examined, all with known chromosome 17 copy numbers. The growth factor receptor-bound protein 7 (GRB7) and junction
plakoglobin (JUP) were the two most highly overexpressed genes in the
TGCTs. GRB7 is tightly linked to ERBB2 and is coamplified and coexpressed with this gene in several cancer types. Interestingly, the expression
levels of ERBB2 were not elevated in the TGCTs, suggesting that GRB7
might be the target for the increased DNA copy number in TGCTs.
Because of the limited knowledge of altered gene expression in the development of TGCTs, we also examined the expression levels of 512 additional genes located throughout the genome. Several genes novel to testicular tumorigenesis were consistently up- or down-regulated, including
POV1, MYCL1, MYBL2, MXI1, and DNMT2. Additionally, overexpression
of the proto-oncogenes CCND2 and MYCN were confirmed from the
literature. The overexpressions were for some of the target genes closely
associated to either seminoma or nonseminoma TGCTs, and hierarchical
cluster analysis of the gene expression data effectively distinguished
among the known histological subtypes. In summary, this focused functional genomic characterization of TGCTs has lead to the identification of
new gene targets associated with a common genomic rearrangement as
well as other genes with potential importance to testicular tumorigenesis.
teristic of virtually all TGCTs (4, 5). In addition, specific gains and
losses from several other chromosomal regions have been described.
Although molecular studies have shown some genes to be altered at
the DNA and/or expression levels in a limited number of TGCTs, the
target genes reflecting the nonrandom chromosomal aberrations remain unknown (for a review of TGCT genetics, see Refs. 6 and 7).
We have demonstrated recently by a genome-wide copy number
analysis using CGH that sequences on chromosome arm 17q are
frequently overrepresented in TGCTs, and two common regions of
copy number increase were identified (8). Gain of the proximal
region, 17q11– q21, is preferentially observed in nonseminomas,
whereas gain of the distal region, 17q24 – qter, is common to all
TGCTs (8). Nonrandom gain at 17q has also been reported in several
other cancer types (9 –13).
To identify differentially expressed genes on chromosome 17 in
TGCTs, we analyzed a series of TGCTs and normal testicular samples
using a custom-made cDNA microarray with a comprehensive chromosome 17 coverage (14). All analyzed tumors had been studied
previously by CGH, and thus, the expression profiles could be related
to the DNA copy numbers. The expression levels of 512 additional
genes mapping elsewhere in the genome, including many cancerrelated genes, were also analyzed in the same set of TGCT samples.
MATERIALS AND METHODS
Tumor and Cell Line Samples. Eighteen testicular tissue samples were
analyzed, including 8 pure seminomas, 6 nonseminomas, 1 carcinoma in situ
(from the vicinity of a nonseminoma), and 3 normal samples. The tumors were
INTRODUCTION
selected from a series of primary TGCTs analyzed previously by CGH (8).
3
TGCT is the most common malignancy among adolescent and Four of the 8 seminomas had gains at distal 17q, including the 17q24 – qter
young adult Caucasian males, and the incidence has been steadily region (Fig. 1A). The 6 nonseminomas (4 embryonal carcinomas and 2 imincreasing over the past 50 years (1, 2). TGCTs are classified into two mature teratomas) had large gains at chromosome 17, all including the 17q11–
main histological subtypes, seminomas and nonseminomas, and there q21 region. The use of these samples in cDNA microarray experiments was
are two models describing their development from carcinomas in situ approved by the Regional Committee for Medical Research Ethics in Norway
and the NIH Office of Human Subjects Research.
(3). Either both subtypes develop independently from carcinomas in
A pool of two breast cancer cell lines, HBL100 and MDA-436 (American
situ, or they develop as a continuum where seminomas may progress Type Culture Collection, Manassas, VA), was used as a common reference in
further into nonseminomas.
the cDNA microarray experiments. These cell lines were selected based on the
TGCTs are generally in the triploid range, and isochromosome 12p facts that they show no increase in copy number at 17q and express most genes
or gain of DNA sequences from chromosome arm 12p is a charac- on the cDNA microarray to some extent (10, 14).
cDNA Microarray Experiments. The construction of the cDNA microarray with comprehensive chromosome 17 coverage has been described previReceived 12/27/01; accepted 2/14/02.
ously by Monni et al. (14). The microarray consisted of printed PCR products
The costs of publication of this article were defrayed in part by the payment of page
charges. This article must therefore be hereby marked advertisement in accordance with
from 636 sequence-verified IMAGE cDNA clones (Research Genetics, Hunts18 U.S.C. Section 1734 solely to indicate this fact.
ville, AL), including 201 known genes from the entire chromosome 17 and 435
1
Supported by the Research Council of Norway (to R. I. S.) and the Norwegian Cancer
ESTs from the 17q arm. An additional 512 sequence-verified IMAGE cDNA
Society (to R. A. L.).
2
clones were placed on the array, representing transcribed sequences located
To whom requests for reprints should be addressed, at Department of Genetics,
Institute for Cancer Research, The Norwegian Radium Hospital, N-0310 Oslo, Norway.
elsewhere in the genome. Eighty-eight of these were housekeeping genes and
Phone: 47-22934415; Fax: 47-22934440; E-mail: [email protected].
were used for calibration among the different experiments (15), 162 were a
3
The abbreviations used are: TGCT, testicular germ cell tumor; CCND2, cyclin D2; CGH,
selection of known or putative cancer-related genes, and 262 were a collection
comparative genomic hybridization; DNMT2, DNA (cytosine-5)-methyltransferase 2;
ERBB2, v-erb-b2 avian erythroblastic leukemia viral oncogene homolog 2; EST, expressed
of genes and ESTs from chromosome 10.
sequence tag; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; GRB7, growth factor
Preparation and printing of the cDNA clones on glass slides, probe prepareceptor-bound protein 7; JUP, junction plakoglobin; LLGL2, lethal giant larvae (Drosophila)
rations, hybridizations, and image generation and analyses were performed as
homolog 2; MYBL2, B-Myb; MYCL1, L-Myc; MYCN, N-Myc; MXI1, MAX-interacting
described (16). Briefly, mRNA was isolated from the test samples by using the
protein 1; PDE6G, phosphodiesterase 6G, cGMP-specific, rod, gamma; POV1, prostate cancer
Trizol reagent (Life Technologies, Inc., Rockville, MD) and oligo(dT)25 dynaoverexpressed gene 1; RT-PCR, reverse transcription-PCR.
2359
GENE EXPRESSION IN TESTICULAR GERM CELL TUMORS
Fig. 1. Genetic changes in testicular germ cell
tumors. A, genomic copy number gains on chromosome 17 as seen by comparative genomic hybridization. Each colored bar represents gain from the
corresponding chromosome segment in the tumor
indicated below. B, sample tree (dendrogram) from
the hierarchical cluster analysis of 18 testicular
samples. Letters below the dendrogram represent
the tissue sources: N, normal testis; C, carcinoma in
situ; I, immature teratoma; E, embryonal carcinoma, and S, seminoma. The vertical distances on
the dendrogram reflect the relatedness of neighboring samples. A gene expression map with pseudocolors coding for the normalized ratios of up- and
down-regulated genes in TGCTs located on chromosome 17 (C) and elsewhere in the genome (D) is
shown. Transcribed sequences are presented in
rows, and each experiment (cDNA sample) is
shown in columns. Thus, each cell in the matrix
represents the expression level of a single transcript
in a single sample. Numbers in parentheses behind
the ESTs provide the IMAGE clone ids. E, color
coding for the normalized expression ratios (expression relative to the average expressions in the
three normal samples).
beads (Dynal Biotech, Oslo, Norway) according to the manufacturers’ specifications. From the reference cell lines, mRNA was isolated directly by using
FastTrack 2.0 mRNA isolation kit (Invitrogen, Carlsbad, CA). Labeled cDNA
was synthesized from 1–3 or 5 g mRNA (test or reference, respectively) in
an oligo(dT)-primed polymerization with SuperScript II reverse transcriptase
(Life Technologies) in the presence of either Cy3 (test) or Cy5 (reference)
labeled dUTP (Amersham Pharmacia, Piscataway, NJ). The Cy3-labeled test
cDNA from the various testicular samples and Cy5-labeled reference cDNA
were mixed and simultaneously hybridized to the cDNA microarray.
The fluorescence intensities at the targets were detected by a laser confocal
scanner (Agilent Technologies, Palo Alto, CA). For each array element, a ratio
between the relative fluorescence intensities of the test and reference was
calculated. This ratio was divided by the average expressions of the 88
housekeeping genes, giving a calibrated ratio. The calibrated ratio was then
normalized by dividing it by the average calibrated ratio of the three normal
testicular samples. Thus, this normalized ratio reflects relative up- or downregulated gene expression from normal to neoplastic testicular tissues.
Statistics. A two-tailed t test for equality of means was used to calculate the
statistical significance of differences in expression levels between different
groups of samples. The hierarchical cluster analysis was done on successful
gene elements (i.e., clones where all experiments had spot sizes ⬎100 area
units and fluorescence intensities stronger than 200 fluorescence units) with
more than 4-fold differential expression within the sample set. The resulting
data were hierarchically clustered by both gene and sample sides (501 clones
and 18 experiments). The average-linkage clustering method was used with
Pearson’s correlation similarity measure. Before calculation of the correlation
between two genes or samples, the original ratio was log transformed, followed
by subtracting the mean from the ratio. The sample tree (dendrogram) is drawn
with “real” instead of fixed distances (in-house cluster analysis software at the
National Human Genome Research Institute, NIH).
Validation by Real-Time RT-PCR. We used real time RT-PCR (TaqMan
system; Applied Biosystems, Foster City, CA) to validate the mRNA expression levels of three genes (GRB7, JUP, and POV1) in 10 testicular samples (3
normal testicular tissues and 7 TGCTs). In this quantitative RT-PCR, a
dual-labeled fluorescent probe is degraded concomitant with PCR amplification. Input target mRNA levels are calculated from the time (measured in PCR
cycles) at which the reporter fluorescent emission increases beyond a threshold
level, as measured by an ABI PRISM 7700 Sequence Detector (Applied
Biosystems).
Primers and probes targeting the mRNA sequences (Table 1) were designed
using the Primer Express software (Applied Biosystems). cDNA synthesized
from 50 ng of mRNA was used as PCR template in a total volume of 25 l
containing 200 nM of each oligonucleotide primer and probe (MedProbe, Oslo,
Norway), 0.2 mM of each deoxynucleotide triphosphate, 1 ⫻ TaqMan buffer,
6 mM MgCl2, 1.25 units of AmpliTaq Gold, 0.25 units of AmpErase UNG (all
Applied Biosystems), and 0.8% glycerol. The PCR program was initiated by 2
2360
GENE EXPRESSION IN TESTICULAR GERM CELL TUMORS
Table 1 Primers and probes used for real time RT-PCR
The probes were 3⬘ labeled with TAMRA and 5⬘ labeled with 6-FAM (GRB7, JUP, and POV1) or JOE (GAPDH).
Gene
Forward primer
Reverse primer
Probe
GRB7
JUP
POV1
GAPDH
TGG CCT CTC GGT CTG TAC AAA
CCA AAA ACA TAA AGC GAT AAT AAT AAA ACA C
AAC CCC TAA CCC AGG ACA CAG
GAA GGT GAA GGT CGG AGT C
GGC AGG GAA TTA TGG GAG
CCC CAT TTC CCG CAC AT
AGA GAC ACA GCC CTC CTT TCA G
GAA GAT GGT GAT GGG ATT TC
CGT GAA ACC GCC TGG GCT GC
CTG CTT GGA CCT CCC CCA GCC
TGG CAC CTC AGG CCC CTT TCC T
CAA GCT TCC CGT TCT CAG CC
min at 50°C and 10 min at 95°C before 40 thermal cycles, each of 15 s at 95°C
and 1 min at 60°C.
Primers and probe targeting GAPDH were multiplexed together with primers and probes targeting each gene of interest. For both the test genes and
GAPDH, standard curves were made from which relative expression values
were calculated. The expression levels of the genes of interest were then
calibrated by dividing by the expression of GAPDH. Again, division by the
average values of the three normal testicular samples normalized all calibrated
expression values, and thus, these values were comparable with the normalized
ratios from the microarray experiments.
RESULTS
The expression levels of 636 chromosome 17-specific transcripts
as well as 512 transcripts located elsewhere in the genome were
determined in 18 testicular tissue samples by cDNA microarrays.
Hierarchical cluster analysis with a set of 501 differentially expressed genes separated the TGCT samples according to their
known histological subgroups (Fig. 1B). The single carcinoma in
situ sample, representing a precursor stage, was most closely
related to the normal testis specimens. The seminomas formed a
single cluster, whereas within the nonseminomas, immature teratomas and embryonal carcinomas clustered into separate groups. A
comprehensive gene expression map for the 51 genes that were
differentially expressed at a 0.01 significance level and had on
average ⬎3-fold up- or down-regulation across all tumors, or
within a histological subgroup, is shown in Fig. 1, C and D.
To identify up-regulated genes from the two regions with frequent
copy number increase on chromosome 17, the normalized ratios of the
genes were plotted as a function of their physical map positions (Fig.
2). This visualization indicated that not all transcripts located in a
region with increased copy number show increased expression. At the
proximal region (17q11– q21), GRB7 and JUP were consistently the
most overexpressed transcripts in the TGCTs (Fig. 1C). At the distal
region (17q24 – qter), LLGL2, PDE6G, and EST (IMAGE clone
124915) were the most up-regulated transcripts (Fig. 1C). Among the
overexpressed genes on 17q, GRB7 was significantly more expressed
in nonseminomas and JUP in seminomas (both P ⬍ 0.01).
For the clones mapping elsewhere in the genome, the most upregulated transcribed sequences in TGCTs, i.e., the highest average
normalized ratios across all tumor samples, were in decreasing order
MYBL2, CCND2, MYCN, POV1, EST (272938), and MYCL1. The
average expression levels of POV1 and MYCL1 were significantly
higher in seminomas than in nonseminomas (P ⬍ 0.01). The expression data also revealed several genes, such as MXI1 and DNMT2, that
were down-regulated in the TGCTs (on average ⬎3-fold downregulated and P ⬍ 0.01 for differential expression between normal
testis and TGCTs; Fig. 1, C and D).
The expression levels of GRB7, JUP, and POV1 were validated in
10 samples by real time RT-PCR, and overexpression in tumors
(compared with normal testicular samples) were seen by both methods
(Fig. 3).
DISCUSSION
Increased DNA copy number is a common mechanism for overexpression of genes promoting neoplastic and malignant cell behavior.
In TGCTs, frequent DNA copy number increase of several chromosomal regions has been observed (17, 18). However, not much is
known with regard to the specific genes that are targeted for overexpression. Recently, we identified two novel regions on chromosome
arm 17q with common copy number increase in TGCT (8). In the
present study, we used gene expression analysis by cDNA microarrays as a high throughput method to identify potential target genes in
these two regions. The comprehensive coverage of the microarray
enabled us to determine which genes were overexpressed in TGCT, as
compared with normal testicular tissue, and therefore most likely to be
involved in driving the genomic alteration. Furthermore, the microarray was constructed to include several additional genes with known or
Fig. 2. Expression levels of transcripts on chromosome 17. Normalized expression values of
genes localized on chromosome 17 were plotted as
a function of their physical map positions (megabasepairs from p-telomere, obtained from the University of California, Santa Cruz database, http://
genome.ucsc.edu/). Individual data points were
connected with a line. The chromosome ideogram
is shown only for approximate visual comparison.
Examples of an embryonal carcinoma with gain
from the proximal gained region in TGCT (17q1221; A), an immature teratoma with extra copies of
the whole chromosome 17 (B), a seminoma with
gain at the distal region (17q24 – qter; C) and a
seminoma with no copy number changes on chromosome 17 (D) are shown.
2361
GENE EXPRESSION IN TESTICULAR GERM CELL TUMORS
Fig. 3. Validation of cDNA microarray results by real time RT-PCR. The expression
levels of GRB7, JUP, and POV1 were analyzed by real-time RT-PCR in 10 of the same
mRNA samples used for cDNA microarray analyses. Relative mRNA levels in TGCTs
and normal testicular tissues are indicated by f and 䡬 respectively. The results from both
methods are shown as normalized ratios (i.e., expression levels relative to the average of
the three normal testicular samples).
putative cancer-related functions, which makes the present study the
most extensive expression analysis of potentially cancer-promoting
genes in TGCT.
The cDNA microarray-based expression survey in TGCTs and
normal testis samples revealed several novel results:
(a) the hierarchical cluster analysis of the cDNA microarray data
grouped the samples according to their correct histological subtypes.
This is rather surprising, taken into account the limited number and
highly selected nature of the transcripts included in this analysis, and
might indicate that genes located on chromosome 17 are fundamental
for the biological characteristics of these tumors.
(b) Comparison of the microarray expression data and the DNA
copy number increases along chromosome 17 as determined by CGH
showed that most genes were not transcriptionally up-regulated because of extra DNA copies. These results are in line with our previous
data on breast cancer (14) and indicate that increased gene copy
number does not always lead to increased gene expression.
In the present study, we have identified overexpressed genes located in the two common regions of copy number increase on chromosome 17 in TGCTs. At the proximal region (17q11– q21), the
cDNA microarray survey showed consistent overexpression of the
GRB7 and JUP genes. GRB7 is closely linked to the ERBB2 oncogene
(20 kb apart4), and has been shown frequently coamplified and coexpressed with ERBB2 in breast, esophageal, and gastric cancers (19 –
22). Interestingly, the expression of ERBB2 was not elevated in any of
the TGCT samples. To our knowledge, this represents the first example where increased copy number at the ERBB2 locus does not lead to
transcriptional activation of ERBB2. Furthermore, this indicates that
other genes at this locus, such as GRB7, are critical for the development of TGCTs and possibly to other tumor types with ERBB2
amplification.
GRB7 encodes an adaptor protein that through its Src homology 2
domain interacts with the cytoplasmic domain of the growth factor
receptor ERBB2 (19). Thus, increased expression of one of these
proteins may be sufficient to promote tumor development. GRB7 also
binds to several other tyrosine kinase growth factor receptors, including KIT, platelet-derived growth factor receptor, RET, and INSR
(23–26), as well as to cytoplasmic tyrosine kinases (27, 28). The KIT
proto-oncogene has previously been suggested to play a role in TGCT
development, both attributable to increased expression (29), and by its
4
involvement in survival, proliferation, and migration of primordial
germ cells (30). Additional knowledge about GRB7 that strengthens
its potential importance in TGCT development is the RAS-associating-like domain (31) and its role in cell migration (32, 33). Interestingly, the expression of GRB7 in esophageal carcinomas is related to
metastatic progression (34).
JUP is also located within the proximal 17q region gained in
TGCTs. It was up-regulated in tumors, with and without genomic gain
by CGH. JUP belongs to the catenin family and encodes a submembranous junctional plaque protein in both desmosomes and intermediate junctions. It may have oncogenic potential through its function
in the Wnt signaling pathway (35, 36), although the importance of this
pathway in TGCT remains to be elucidated.
At the distal 17q region frequently gained in TGCTs, 17q24 – qter,
the cDNA microarray analyses implicated the LLGL2 and PDE6G
genes, as well as an uncharacterized EST (124915), as consistently
up-regulated in the TGCTs. The Drosophila orthologue to LLGL2,
1(2)gl, functions as a tumor suppressor (37). However, the function
may be different in germ cells, because 1(2)gl is required for survival
of germ-line cells in Drosophila (38), and our results show that
LLGL2 is up-regulated in human TGCT. Furthermore, LLGL2 are
abundantly represented in some cDNA libraries derived from human
lung and prostate tumors.5 PDE6G encodes an effector protein involved in phototransduction in the eye (39), and to our knowledge, no
cancer-related function has been linked to this gene.
From the genes located elsewhere in the genome, three human
homologues of avian retroviral oncogenes, MYCN at 2p24.1, MYCL1
at 1p34.3, and MYBL2 at 20q13.1, were among the most overexpressed genes in the TGCTs. The chromosomal locations of MYCN
and MYCL1 are both within regions that are gained in approximately
one-third of all TGCTs, whereas the locus of MYBL2 is rarely involved in copy number changes (8). All three gene products are
localized to the nucleus and function as transcriptional transactivators.
The MYCN overexpression has been detected previously in TGCT
(40). Remarkably, studies of neuroblastomas give evidence for both
statistical and structural associations between MYCN overexpression
and gain of 17q21–ter (12, 41). Furthermore, several E-boxes (the
common DNA binding site of the MYC family proteins) are found in
the promoter region of CCND2, and MYC overexpression has been
shown to induce chromosomal and extrachromosomal instability of
the CCND2 gene at 12p13 (42).
Gain of chromosome arm 12p, often through the presence of
isochromosome 12p (4), is the most common genetic aberration in
TGCTs. Two smallest regions of overlapping amplifications on 12p
have been suggested, one at 12p13 (8) and one more proximal region
(43), harboring the candidate genes CCND2 and KRAS2, respectively.
We showed that both genes were transcriptionally up-regulated in
TGCTs, but CCND2 significantly more than KRAS2. Activating mutations of KRAS2 have only rarely been detected in TGCTs (44, 45),
further reducing the importance of this proto-oncogene in TGCTs.
The observed overexpression of CCND2 is in keeping with a study by
Houldsworth et al. (46), where CCND2 had the highest increased
expression among a set of six candidate genes on 12p, and with a
study by Bartkova et al. (47), finding CCND2 protein expression
related to early stages of TGCTs.
The POV1 gene at 11q12 was highly expressed in all seminomas
and in the carcinoma in situ, but neither in the normal samples nor in
the nonseminomas. Thus, independently of developmental model, this
gene may be an early-onset gene in the development of seminoma
5
Internet address: http://genome.ucsc.edu/ (Aug. 6, 2001 freeze).
2362
Internet address: http://www.ncbi.nlm.nih.gov/UniGene/.
GENE EXPRESSION IN TESTICULAR GERM CELL TUMORS
TGCTs. In analogy, precursor lesions of prostate cancer have shown
increased expression of POV1 (48).
In addition to DNA copy number, other means of transcriptional
control are obviously important to TGCT development. These may
include regulation of transcription factors and disruption of the DNA
methylation pattern. Hence, the gene products of MXI1 and DNMT2
are potential candidates because of their consistently down-regulated
mRNA levels demonstrated by our microarray survey. The MXI1
protein is a transcriptional repressor through its binding to MAX, a
MYC heterodimerization partner (49). Thus, by its competition for
MAX, MXI1 antagonizes the MYC transcription factors, of which we
have shown increased mRNA levels in TGCTs for both MYCN and
MYCL1. Interestingly, the MXI1 mouse homologue is mapped within
the region with highest score in a genome-wide linkage analysis
targeting TGCT susceptibility loci in mice (50, 51). Furthermore, the
MXI1 gene is commonly mutated or deleted in prostate carcinomas
(52). The DNMT2 gene has strong sequence homology to the DNA(cytosine-5)-methyltransferases, although its catalytic activity has yet
to be demonstrated (53). Additionally, DNMT2 is transcriptionally
down-regulated in colorectal, stomach, and hepatocellular cancers
(54, 55).
In summary, the present study has identified altered expression of
several genes that are novel to testicular tumorigenesis. The increased
copy numbers observed at 17q11– q21 in TGCTs are associated with
overexpression of GRB7, and in contrast to other tumor types, not with
overexpression of the ERBB2 oncogene. In addition, JUP, MYCN,
MYCL1, MYBL2, CCND2, and POV1 are consistently overexpressed in
TGCTs compared with their expressions in nonneoplastic testicular tissue. Furthermore, our data show clear gene expression differences between seminoma and nonseminoma TGCTs. The average expression
level of GRB7 was significantly higher in nonseminomas than in seminomas, whereas the expressions of JUP, MYCL1, and POV1 were highest
in seminomas. The putative cancer-related functions of all these genes, in
addition to the previously reported overexpression of MYCN and CCND2
in TGCT (40, 46), suggest that the applied cDNA microarrays are
sensitive and specific enough to discover oncogenic gene expression
changes in TGCT. Thus, the consistently overexpressed ESTs may also
reflect genes playing important roles in TGCT oncogenesis. In conclusion, this focused functional genomic characterization has lead to the
identification of new gene targets associated with a common genomic
rearrangement in TGCT and of genes with potential clinical impact
worthy further investigation.
REFERENCES
1. Bergström, R., Adami, H. O., Mohner, M., Zatonski, W., Storm, H., Ekbom, A.,
Tretli, S., Teppo, L., Akre, O., and Hakulinen, T. Increase in testicular cancer
incidence in six European countries: a birth cohort phenomenon. J. Natl. Cancer Inst.,
88: 727–733, 1996.
2. Devesa, S. S., Blot, W. J., Stone, B. J., Miller, B. A., Tarone, R. E., and Fraumeni,
J. F. J. Recent cancer trends in the United States. J. Natl. Cancer Inst., 87: 175–182,
1995.
3. Dieckmann, K. P., and Skakkebæk, N. E. Carcinoma in situ of the testis: review of
biological and clinical features. Int. J. Cancer, 83: 815– 822, 1999.
4. Atkin, N. B., and Baker, M. C. Specific chromosome change, i(12p), in testicular
tumours? Lancet, 2: 1349, 1982.
5. Rodriguez, E., Houldsworth, J., Reuter, V. E., Meltzer, P., Zhang, J., Trent, J. M.,
Bosl, G. J., and Chaganti, R. S. Molecular cytogenetic analysis of i(12p)-negative
human male germ cell tumors. Genes Chromosomes Cancer, 8: 230 –236, 1993.
6. Chaganti, R. S., and Houldsworth, J. Genetics and biology of adult human male germ
cell tumors. Cancer Res., 60: 1475–1482, 2000.
7. Looijenga, L. H., and Oosterhuis, J. W. Pathogenesis of testicular germ cell tumours.
Rev. Reprod., 4: 90 –100, 1999.
8. Kraggerud, S. M., Skotheim, R. I., Szymanska, J., Eknæs, M., Fosså, S. D., Stenwig,
A. E., Peltomäki, P., and Lothe, R. A. Genome profiles of familial/bilateral and
sporadic testicular germ cell tumors. Genes Chromosomes Cancer, in press, 2002.
9. Lothe, R. A., Karhu, R., Mandahl, N., Mertens, F., Sæter, G., Heim, S., BørresenDale, A. L., and Kallioniemi, O. P. Gain of 17q24 – qter detected by comparative
genomic hybridization in malignant tumors from patients with von Recklinghausen’s
neurofibromatosis. Cancer Res., 56: 4778 – 4781, 1996.
10. Forozan, F., Mahlamaki, E. H., Monni, O., Chen, Y., Veldman, R., Jiang, Y., Gooden,
G. C., Ethier, S. P., Kallioniemi, A., and Kallioniemi, O. P. Comparative genomic
hybridization analysis of 38 breast cancer cell lines: a basis for interpreting complementary DNA microarray data. Cancer Res., 60: 4519 – 4525, 2000.
11. Kokkola, A., Monni, O., Puolakkainen, P., Larramendy, M. L., Victorzon, M.,
Nordling, S., Haapiainen, R., Kivilaakso, E., and Knuutila, S. 17q12–21 amplicon, a
novel recurrent genetic change in intestinal type of gastric carcinoma: a comparative
genomic hybridization study. Genes Chromosomes Cancer, 20: 38 – 43, 1997.
12. Lastowska, M., Nacheva, E., McGuckin, A., Curtis, A., Grace, C., Pearson, A., and
Bown, N. Comparative genomic hybridization study of primary neuroblastoma tumors. United Kingdom Children’s Cancer Study Group. Genes Chromosomes Cancer, 18: 162–169, 1997.
13. Pack, S. D., Karkera, J. D., Zhuang, Z., Pak, E. D., Balan, K. V., Hwu, P., Park, W. S.,
Pham, T., Ault, D. O., Glaser, M., Liotta, L., S. D., and Wadleigh, R. G. Molecular
cytogenetic fingerprinting of esophageal squamous cell carcinoma by comparative
genomic hybridization reveals a consistent pattern of chromosomal alterations. Genes
Chromosomes Cancer, 25: 160 –168, 1999.
14. Monni, O., Bärlund, M., Mousses, S., Kononen, J., Sauter, G., Heiskanen, M.,
Paavola, P., Avela, K., Chen, Y., Bittner, M. L., and Kallioniemi, A. Comprehensive
copy number and gene expression profiling of the 17q23 amplicon in human breast
cancer. Proc. Natl. Acad. Sci. USA, 98: 5711–5716, 2001.
15. Chen, Y., Dougherty, E. R., and Bittner, M. L. Ratio-based decisions and the
quantitative analysis of cDNA microarray images. J. Biomed. Optics, 2: 364 –374,
1997.
16. Mousses, S., Bittner, M. L., Chen, Y., Dougherty, E. R., Baxevanis, A., Meltzer, P. S.,
and Trent, J. M. Gene expression analysis by cDNA microarrays. In: F. J. Liversey
and S. P. Hunt (eds.), Functional Genomics, pp. 113–137. Oxford: Oxford University
Press, 2000.
17. Chaganti, R. S., Rodriguez, E., and Bosl, G. J. Cytogenetics of male germ-cell tumors.
Urol. Clin. North Am., 20: 55– 66, 1993.
18. Rosenberg, C., Schut, T. B., Mostert, M., Tanke, H., Raap, A., Oosterhuis, J. W., and
Looijenga, L. Chromosomal gains and losses in testicular germ cell tumors of
adolescents and adults investigated by a modified comparative genomic hybridization
approach. Lab. Investig., 79: 1447–1451, 1999.
19. Stein, D., Wu, J., Fuqua, S. A., Roonprapunt, C., Yajnik, V., D’Eustachio, P.,
Moskow, J. J., Buchberg, A. M., Osborne, C. K., and Margolis, B. The SH2 domain
protein GRB-7 is co-amplified, overexpressed and in a tight complex with HER2 in
breast cancer. EMBO J., 13: 1331–1340, 1994.
20. Kauraniemi, P., Bärlund, M., Monni, O., and Kallioniemi, A. New amplified and
highly expressed genes discovered in the ERBB2 amplicon in breast cancer by cDNA
microarrays. Cancer Res., 61: 8235– 8240, 2001.
21. Tanaka, S., Mori, M., Akiyoshi, T., Tanaka, Y., Mafune, K., Wands, J. R., and
Sugimachi, K. Coexpression of Grb7 with epidermal growth factor receptor or
Her2/erbB2 in human advanced esophageal carcinoma. Cancer Res., 57: 28 –31,
1997.
22. Akiyama, N., Sasaki, H., Ishizuka, T., Kishi, T., Sakamoto, H., Onda, M., Hirai, H.,
Yazaki, Y., Sugimura, T., and Terada, M. Isolation of a candidate gene, CAB1, for
cholesterol transport to mitochondria from the c-ERBB-2 amplicon by a modified
cDNA selection method. Cancer Res., 57: 3548 –3553, 1997.
23. Thömmes, K., Lennartsson, J., Carlberg, M., and Rönnstrand, L. Identification of
Tyr-703 and Tyr-936 as the primary association sites for Grb2 and Grb7 in the
c-Kit/stem cell factor receptor. Biochem. J., 341: 211–216, 1999.
24. Yokote, K., Margolis, B., Heldin, C. H., and Claesson-Welsh, L. Grb7 is a downstream signaling component of platelet-derived growth factor - and -receptors.
J. Biol. Chem., 271: 30942–30949, 1996.
25. Pandey, A., Liu, X., Dixon, J. E., Di Fiore, P. P., and Dixit, V. M. Direct association
between the Ret receptor tyrosine kinase and the Src homology 2-containing adapter
protein Grb7. J. Biol. Chem., 271: 10607–10610, 1996.
26. Kasus-Jacobi, A., Bereziat, V., Perdereau, D., Girard, J., and Burnol, A. F. Evidence
for an interaction between the insulin receptor and Grb7. A role for two of its binding
domains, PIR and SH2. Oncogene, 19: 2052–2059, 2000.
27. Han, D. C., and Guan, J. L. Association of focal adhesion kinase with Grb7 and its
role in cell migration. J. Biol. Chem., 274: 24425–24430, 1999.
28. Keegan, K., and Cooper, J. A. Use of the two hybrid system to detect the association
of the protein-tyrosine-phosphatase, SHPTP2, with another SH2-containing protein,
Grb7. Oncogene, 12: 1537–1544, 1996.
29. Strohmeyer, T., Reese, D., Press, M., Ackermann, R., Hartmann, M., and Slamon, D.
Expression of the c-kit proto-oncogene and its ligand stem cell factor (SCF) in normal
and malignant human testicular tissue. J. Urol., 153: 511–515, 1995.
30. Mauduit, C., Hamamah, S., and Benahmed, M. Stem cell factor/c-kit system in
spermatogenesis. Hum. Reprod. Update, 5: 535–545, 1999.
31. Wojcik, J., Girault, J. A., Labesse, G., Chomilier, J., Mornon, J. P., and Callebaut, I.
Sequence analysis identifies a ras-associating (RA)-like domain in the N-termini of
band 4.1/JEF domains and in the Grb7/10/14 adapter family. Biochem. Biophys. Res.
Commun., 259: 113–120, 1999.
32. Han, D. C., Shen, T. L., and Guan, J. L. Role of Grb7 targeting to focal contacts and
its phosphorylation by focal adhesion kinase in regulation of cell migration. J. Biol.
Chem., 275: 28911–28917, 2000.
33. Manser, J., Roonprapunt, C., and Margolis, B. C. elegans cell migration gene mig-10
shares similarities with a family of SH2 domain proteins and acts cell nonautonomously in excretory canal development. Dev. Biol., 184: 150 –164, 1997.
34. Tanaka, S., Sugimachi, K., Kawaguchi, H., Saeki, H., Ohno, S., and Wands, J. R.
Grb7 signal transduction protein mediates metastatic progression of esophageal
carcinoma. J. Cell Physiol., 183: 411– 415, 2000.
2363
GENE EXPRESSION IN TESTICULAR GERM CELL TUMORS
35. Kolligs, F. T., Kolligs, B., Hajra, K. M., Hu, G., Tani, M., Cho, K. R., and Fearon,
E. R. -catenin is regulated by the APC tumor suppressor and its oncogenic activity
is distinct from that of -catenin. Genes Dev., 14: 1319 –1331, 2000.
36. Barker, N., and Clevers, H. Catenins, Wnt signaling and cancer. Bioessays, 22:
961–965, 2000.
37. Török, I., Hartenstein, K., Kalmes, A., Schmitt, R., Strand, D., and Mechler, B. M.
The 1(2)gl homologue of Drosophila pseudoobscura suppresses tumorigenicity in
transgenic Drosophila melanogaster. Oncogene, 8: 1537–1549, 1993.
38. De Lorenzo, C., Strand, D., and Mechler, B. M. Requirement of Drosophila I(2)gl
function for survival of the germline cells and organization of the follicle cells in a
columnar epithelium during oogenesis. Int. J. Dev. Biol., 43: 207–217, 1999.
39. Pittler, S. J., Baehr, W., Wasmuth, J. J., McConnell, D. G., Champagne, M. S., van
Tuinen, P., Ledbetter, D., and Davis, R. L. Molecular characterization of human and
bovine rod photoreceptor cGMP phosphodiesterase -subunit and chromosomal
localization of the human gene. Genomics, 6: 272–283, 1990.
40. Shuin, T., Misaki, H., Kubota, Y., Yao, M., and Hosaka, M. Differential expression
of protooncogenes in human germ cell tumors of the testis. Cancer, (Phila.), 73:
1721–1727, 1994.
41. O’Neill, S., Ekstrom, L., Lastowska, M., Roberts, P., Brodeur, G. M., Kees, U. R.,
Schwab, M., and Bown, N. MYCN amplification and 17q in neuroblastoma: evidence
for structural association. Genes Chromosomes Cancer, 30: 87–90, 2001.
42. Mai, S., Hanley-Hyde, J., Rainey, G. J., Kuschak, T. I., Paul, J. T., Littlewood, T. D.,
Mischak, H., Stevens, L. M., Henderson, D. W., and Mushinski, J. F. Chromosomal
and extrachromosomal instability of the cyclin D2 gene is induced by Myc overexpression. Neoplasia, 1: 241–252, 1999.
43. Roelofs, H., Mostert, M. C., Pompe, K., Zafarana, G., van Oorschot, M., van Gurp,
R. J., Gillis, A. J., Stoop, H., Beverloo, B., Oosterhuis, J. W., Bokemeyer, C., and
Looijenga, L. H. Restricted 12p amplification and RAS mutation in human germ cell
tumors of the adult testis. Am. J. Pathol., 157: 1155–1166, 2000.
44. Mulder, M. P., Keijzer, W., Verkerk, A., Boot, A. J., Prins, M. E., Splinter, T. A., and
Bos, J. L. Activated ras genes in human seminoma: evidence for tumor heterogeneity.
Oncogene, 4: 1345–1351, 1989.
45. Ridanpää, M., Lothe, R. A., Önfelt, A., Fosså, S. D., Børresen, A. L., and HusgafvelPursiainen, K. K-ras oncogene codon 12 point mutations in testicular cancer. Environ.
Health Perspect., 101 (Suppl. 3): 185–187, 1993.
46. Houldsworth, J., Reuter, V., Bosl, G. J., and Chaganti, R. S. Aberrant expression of
cyclin D2 is an early event in human male germ cell tumorigenesis. Cell Growth
Differ., 8: 293–299, 1997.
47. Bartkova, J., Rajpert-De Meyts, E., Skakkebæk, N. E., and Bartek, J. D-type cyclins
in adult human testis and testicular cancer: relation to cell type, proliferation,
differentiation, and malignancy. J. Pathol., 187: 573–581, 1999.
48. Cole, K. A., Chuaqui, R. F., Katz, K., Pack, S., Zhuang, Z., Cole, C. E., Lyne, J. C.,
Linehan, W. M., Liotta, L. A., and Emmert-Buck, M. R. cDNA sequencing and
analysis of POV1 (PB39): a novel gene up-regulated in prostate cancer. Genomics,
51: 282–287, 1998.
49. Zervos, A. S., Gyuris, J., and Brent, R. Mxi1, a protein that specifically interacts with
Max to bind Myc-Max recognition sites. Cell, 72: 223–232, 1993.
50. Collin, G. B., Asada, Y., Varnum, D. S., and Nadeau, J. H. DNA pooling as a quick
method for finding candidate linkages in multigenic trait analysis: an example
involving susceptibility to germ cell tumors. Mamm. Genome, 7: 68 –70, 1996.
51. Matin, A., Collin, G. B., Asada, Y., Varnum, D., and Nadeau, J. H. Susceptibility to
testicular germ-cell tumours in a 129.MOLF-Chr 19 chromosome substitution strain.
Nat. Genet., 23: 237–240, 1999.
52. Prochownik, E. V., Eagle, G. L., Deubler, D., Zhu, X. L., Stephenson, R. A., Rohr,
L. R., Yin, X., and Brothman, A. R. Commonly occurring loss and mutation of the
MXI1 gene in prostate cancer. Genes Chromosomes Cancer, 22: 295–304, 1998.
53. Robertson, K. D. DNA methylation, methyltransferases, and cancer. Oncogene, 20:
3139 –3155, 2001.
54. Kanai, Y., Ushijima, S., Kondo, Y., Nakanishi, Y., and Hirohashi, S. DNA methyltransferase expression and DNA methylation of CpG islands and peri-centromeric
satellite regions in human colorectal and stomach cancers. Int. J. Cancer, 91: 205–
212, 2001.
55. Saito, Y., Kanai, Y., Sakamoto, M., Saito, H., Ishii, H., and Hirohashi, S. Expression
of mRNA for DNA methyltransferases and methyl-CpG-binding proteins and DNA
methylation status on CpG islands and pericentromeric satellite regions during human
hepatocarcinogenesis. Hepatology, 33: 561–568, 2001.
2364
Paper V
Skotheim RI, Abeler VM, Nesland JM, Fosså SD, Holm R, Wagner U, Aass N,
Kallioniemi OP, and Lothe RA
Candidate genes for testicular cancer evaluated by in situ protein expression
analyses on tissue microarrays
Submitted manuscript
blank
Submitted manuscript, 2002, pp. 1-12
Candidate genes for testicular cancer evaluated by in situ protein
expression analyses on tissue microarrays
Rolf I. Skotheim,1 Vera M. Abeler,2 Jahn M. Nesland,2 Sophie D. Fosså,3 Ruth Holm,2
Urs Wagner,4,5 Nina Aass,3 Olli-P. Kallioniemi,4,6 and Ragnhild A. Lothe1,*
Departments of 1Genetics and 2Pathology, Institute for Cancer Research, and 3Department of Medical
Oncology and Radiotherapy, The Norwegian Radium Hospital, N-0310 Oslo, Norway; 4Cancer
Genetics Branch, National Human Genome Research Institute, National Institutes of Health,
Bethesda, Maryland; 5Institute for Pathology, University of Basel, Basel, Switzerland; and 6Medical
Biotechnology Group, VTT Technical Research Centre of Finland and University of Turku, Turku,
Finland.
By use of high-throughput molecular technologies, the number of genes and proteins
potentially relevant to testicular germ cell tumor (TGCT) and other diseases will
increase rapidly. In a recent transcriptional profiling, we demonstrated overexpression
of GRB7 and JUP in TGCTs, and confirmed the reported overexpression of CCND2. We
also have recent evidences for frequent genetic alterations of FHIT and epigenetic
alterations of MGMT. To evaluate whether expression of these genes are related to any
clinicopathological variables we constructed a tissue microarray with 506 testicular
tissue cores from 278 patients diagnosed with TGCT, covering various histological
subgroups and clinical stages. By immunohistochemistry we found that JUP, GRB7, and
CCND2 proteins were rarely present in normal testicular tissue, but frequently
expressed at high levels in TGCT. Additionally, all carcinomas in situ were JUP
immunopositive. MGMT and FHIT were expressed by normal testicular tissues, but at
significantly lower frequencies in TGCT. Except for CCND2, the expressions of all
markers were significantly associated with various TGCT subtypes. Furthermore, the
immunoreactivity of FHIT was strongly associated with those of GRB7 and MGMT. In
summary, we have developed a high-throughput tool for evaluation of TGCT markers,
and utilized this to validate five candidate genes whose protein expressions were indeed
deregulated in TGCT.
seminomas resemble CIS cells, but do not
constrain within the tubules and are quite
INTRODUCTION
proliferative, the nonseminomas develop
Testicular germ cell tumor (TGCT)
through a pluripotent embryonal carcinoma
of adolescent and young adult males are
stage, which may differentiate into cells and
classified into two main histological
tissue types of all three primary germ layers
1
subtypes, seminomas and nonseminomas,
at various stages of differentiation
and both types develop from premalignant
(somatically differentiated teratomas and
carcinoma in situ (CIS; intratubular
extra-embryonally differentiated choriomalignant germ cells).2 Whereas the
carcinomas and yolk sac tumors). Thus,
*
Address correspondence to Ragnhild A. Lothe, Department of Genetics, Institute for Cancer Research, The
Norwegian Radium Hospital, N-0310 Oslo, Norway. Phone: +47 22934415; Fax: +47 22934440; E-mail:
[email protected]
1
Skotheim et al.
tumor development in the testis mimics the
embryogenesis and makes TGCT an
interesting model also for developmental
biology.
The TGCT genomes are usually
hypo- or hypertriploid with extensive
chromosome losses and gains.3 Virtually all
TGCTs have extra copies of chromosome
arm 12p, often seen as an isochromosome.46
There are also other chromosomal copy
number alterations occurring at high frequencies, implying the existence of genes
within them with relevance to TGCT
development. Furthermore, epigenetically
deregulated gene expression through
aberrant CpG island methylation seems to
be common in TGCTs.7
We have in three recent reports on
the genetics and epigenetics of TGCT
gained evidence for specific target genes in
TGCT.8-10 A cDNA microarray study,
mainly focusing on the frequently overrepresented chromosome arm 17q,11
revealed growth factor receptor-bound
protein 7 (GRB7) and junction plakoglobin
(JUP; Ȗ-catenin) as transcriptionally
overexpressed in TGCT.8 This study also
confirmed the overexpression of cyclin D2
(CCND2).12-15 By a candidate gene
approach, we found epigenetic alterations
of the DNA repair gene O6-methylguanineDNA methyltransferase (MGMT) and
frequent allelic imbalances in the
chromosome band 10q26 that harbor this
gene.9 Although in a limited series,
methylation of MGMT was found to be
associated with lack of protein expression.16
Further, we reported the fragile histidine
triad (FHIT) gene, located within the
commonly deleted region on 3p14, to have
aberrant splice variants and downregulated
expression in TGCT.10
2
All these studies reported novel
target genes in TGCT, but because of the
limited sample sizes (range 14 to 70
TGCTs), only few conclusions could be
drawn in relation to clinicopathological
variables.
Tissue microarrays facilitate the
validation of candidate genes/proteins, as a
larger series of samples are evaluated, and
thus give statistically strong data to
associations between genotypes or phenotypes and clinicopathological variables.17-19
We therefore constructed a tissue microarray from archival blocks of a large series
of primary TGCTs of various clinical
stages, including all histological subtypes,
as well as CIS and normal testicular tissues.
We have evaluated the in situ protein
expressions of the candidates JUP, GRB7,
CCND2, MGMT, and FHIT in this set of
testicular tissue samples and correlated the
results with clinical and pathological
variables.
MATERIALS AND METHODS
Tumor material and the tissue
microarray technology
A tissue microarray block was
constructed with 506 testicular tissue cores
punched from formalin fixed and paraffinembedded orchiectomy specimens from
278 TGCT patients. The distribution of the
histological subtypes is shown in Table 1.
One to five tissue cores of different
histological subtypes from each TGCT/
patient were transferred into the array
block, reflecting the number of histological
components. Forty-five of the tissue cores
were replicates of the same histological
subtype from the same tumor, and were
included for validation of heterogeneity and
consistency of staining.
TGCT tissue microarray
Table 1. Histological subtypes of the 506 testicular tissue cores in
the tissue microarray.
Histology
Orchiectomy
specimens (n=)
Tissue cores (n=)
Normal testis
21
24
Carcinoma in situ
21
21
Seminoma
167
184
Embryonal carcinoma
99
102
Choriocarcinoma
16
16
Yolk sac tumor
62
69
Teratoma
75
90
278*
506
Total
*
The total number of TGCTs/patients is lower than the sum of each
histological subtype as there often are tissue cores of several
histological subtypes provided from each orchiectomy specimen.
According to the Royal Marsden
staging system (stage I: non-metastatic
TGCT; stage II-IV: metastatic TGCT ),20
there were 174 patients classified as stage I,
53 stage II, 13 stage III, and 38 stage IV.
Patients without clinically demonstrated
metastases underwent retroperitoneal lymph
node dissection or followed a surveillance
program. Patients with metastases received
cisplatin based chemotherapy followed, in
the majority of cases, by resection of
residual masses. The patients underwent
orchiectomy between 1981 and 1999, and
all patients were followed up until death or
May 2002. All TGCT samples were
available from the archive of Department of
Pathology of The Norwegian Radium
Hospital. The study was approved by the
Regional Committee for Medical Research
Ethics (S-00201, 150800).
Sections of up to 10 tissue blocks
from each orchiectomy specimen were
stained with hematoxylin and eosin, and
light microscopically examined by an
expert pathologist on germ cell tumors (V.
A.). The best areas for tissue punching were
marked. The tissue microarray was
assembled using a robotic tissue microarrayer. Briefly, cylindrical tissue cores
with 0.6mm diameter were transferred from
the donor archival tissue blocks and arrayed
into an empty recipient paraffin block,
building up the tissue microarray.17
The Instrumedics (Instrumedics,
Hackensack, NJ, USA) tape-transfer
method was used to transfer 4Pm sections
of the tissue microarray to glass slides.
Hematoxylin and eosin stained tissue
microarray sections were evaluated to
check for consistency with the originally
assigned histology. Histological classification was performed according to the
WHO recommendations.1 Distinction of
CIS from normal tissue was assisted by
immunohistochemical staining of a parallel
section using anti-PLAP antibodies,
targeting germ cell and placental alkaline
phosphatases, extensively present in CIS
but not in normal spermatogenic germ
cells.2,21
Immunohistochemistry
stained
Tissue microarray sections were
with the biotin–streptavidin3
Skotheim et al.
peroxidase
method
(Supersensitive
Immunodetection System, LP000-UL,
Biogenex, San Raman, CA) and OptiMax
Plus Automated Cell Staining System
(BioGenex). One tissue microarray section
for each antibody was deparaffinized and
rehydrated, and high temperature antigen
retrieval was performed by microwave oven
at 900W. The slides were then incubated
with 1% hydrogen peroxide (H2O2) for 10
min to block the endogenous peroxidase
activity before incubation with the
polyclonal antibodies GRB7 (N-20, sc-607,
1:100, 2 Pg IgG/ml, Santa Cruz Biotechnology, Inc. [SCB], Santa Cruz, CA),
CCND2 (C-17, sc-181, 1: 400, 0.5 Pg
IgG/ml, SCB), MGMT (C-20, sc-8825,
1:200, 1 Pg IgG/ml, SCB), FHIT (ZP54,
1:100, 5 Pg IgG/ml, Zymed Laboratories,
Inc., South San Francisco, CA), and
monoclonal antibodies JUP (clone 15,
1:300, 0.8 Pg IgG2a/ml, Nota Bene
Scientific ApS, Hellebæk, Denmark) and
PLAP (clone 8A9, 1:20, IgG1k, Novocastra
Laboratories Ltd., Newcastle, UK) for 30
min at room temperature. Afterwards, the
sections were incubated for 20 min with
multilink
biotinylated
anti-immunoglobulins (1:30; BioGenex) and 20 min
with streptavidin peroxidase (1:30; BioGenex). Finally, the sections were stained
for 5 min with 0.05% of the peroxidase
substrate 3’3-diaminobenzidine tetrahydrochloride (DAB) freshly prepared in 0.05 M
Tris–HCl buffer at pH=7.6 containing
0.01% H2O2, before being counterstaining
with
hematoxylin,
dehydrated,
and
mounted. Negative controls consisted of
replacement
of
primary
polyclonal
antibodies with normal rabbit IgG at the
same concentration as the polyclonal
antibodies and replacement of primary
monoclonal
antibodies
with
mouse
4
myeloma protein of the same subclass and
concentration as the monoclonal antibodies.
All controls gave satisfactory results.
The JUP immunostaining was
membranous and/or cytoplasmic (Figure 1).
Cases with any membranous and/or
moderate to strong cytoplasmic staining
were scored as positive. The GRB7
immunostaining was membranous and/or
cytoplasmic. Cases with moderate or strong
staining in tumor cells were scored as
positive. The CCND2 immunostaining was
nuclear, though some cytoplasmic staining
was seen in the negative normal testicular
tissues. Cases with staining of more than
5% of the nuclei were considered positive.
The MGMT immunostaining was nuclear,
and cases with staining of more than 5% of
the nuclei were considered positive. The
FHIT immunostaining was cytoplasmic and
was evaluated by a composite score of
intensity (1, weak/absent; 2, moderate; 3,
strong) multiplied by the fraction of
positive cells (1, <10%; 2, 10-50%; 3,
>50%), where cases with composite score
at three or below were regarded FHIT
negative.10,22
A tumor was considered positive
when one or more of the tumor tissue cores
from that specific tumor were positive.
Equally, when more than one tissue core of
a specific histological component of a
tumor was present on the array, that specific
component was considered positive if at
least one of the samples were scored
positive.
RESULTS
In total, the immunohistochemical
analyses of JUP, GRB7, CCND2, MGMT,
and FHIT resulted in 433, 414, 424, 430,
and 418 scored tissue cores, from 256, 254,
TGCT tissue microarray
Figure 1. TGCT tissue microarray, immunohistochemical staining. One negative and one positive
tissue core (0.6mm diameter) are shown for each of the five analyzed proteins. The colored squares
designate the histological subtype of each sample specified by the color code of the histograms. The
histograms, again, indicate the frequencies of positive cases for each histological subtype.
259, 260, and 255 TGCTs, respectively.
The frequencies of positive staining for
each antibody and histological subtype are
shown in Figure 1.
The frequencies of positive tissues
according to histological subtypes are
illustrated in Figure 2A. For JUP, only 13%
of the normal tissues stained positive,
compared to 100% and 94% for CIS and
5
Skotheim et al.
seminomas (p=1x10-5 and p=1x10-10,
respectively). The frequency of JUP
positives in nonseminomas was 77%, which
is significantly lower than among both CIS
and seminomas (Figure 2B; p=6x10-4 and
p=6x10-6).
For GRB7, 19% of the normals were
scored positive, compared to 55% of the
CIS and 56% of the TGCTs (p=0.04 and
p=4x10-3, respectively). The frequency of
GRB7 immunoreactivity in seminomas
(42%) was significantly different from that
in nonseminomas (66%; p=1x10-4).
For CCND2, all normal tissues were
negative, whereas 16% of the CIS samples
and 56% of the TGCTs were positive
(p=5x10-6; normal to invasive tumor). The
frequency of CCND2 immunoreactivity
was
similar
in
seminomas
and
nonseminomas (p=1.00).
For MGMT, all normal tissues had
positive spermatogenic cells, whereas the
percentages of positives among CIS,
seminomas and embryonal carcinomas were
47%, 16%, and 6% (p-values 8x10-5, 6x1013
, and 4x10-16 when compared to the
normal testicular tissues). However, among
the yolk sac tumors and teratomas, 49% and
44% were positive, meaning a significant
re-expression of MGMT upon differentiation from embryonal carcinoma (p=4x10-9
and p=2x10-7).
For FHIT, all normal tissues and
76% of CIS were positive. Both seminomas
(41%) and embryonal carcinomas (38%)
were positive in significantly fewer cases
than both normal tissues (p=3x10-6 and
p=2x10-6) and CIS (p=9x10-3 and p=6x103
).
When we compared the pure seminomas with those from combined tumors,
no significant differences were seen in the
expression patterns of the five analyzed
6
proteins (Figure 2B). The various
antibodies also revealed comparable
staining in CIS samples from seminomas
and nonseminomas (Figure 2C). Within the
teratomas, the epithelial components were
significantly more frequent positive for JUP
and GRB7 than the mesenchymal
components (Figure 2D).
None of the markers were
significantly associated with clinical stage
or mortality (Figure 2E and 2F). Among
TGCTs from patients with history of
undescended testis (cryptorchidism; Figure
2G), 40% were positive for CCND2, as
compared to 60% among TGCTs from
patients with no history of cryptorchidism
(p=0.03).
The immunostaining results evaluated for several combinations (Figure 2H),
and the strongest associations were seen
between FHIT positives and tumors
positive for GRB7 and MGMT (p=6x10-8
and p=3x10-4), followed by associations
between CCND2 positives and tumors
positive for MGMT and GRB7 (p=0.002
and p=0.006).
DISCUSSION
As the human genome gets
unraveled and high-throughput molecular
technologies are utilized, the number of
genes with putative relation to various
diseases, including TGCT,
increases
dramatically.3 Hence, there is a need for
validation of the new putative disease
markers, but so far, the studies on TGCT
have analyzed too few samples to really
pinpoint
significant
associations
to
clinicopathological variables.
By the present study we have taken
advantage of the tissue microarray
technology,17-19 and by transferring more
TGCT tissue microarray
Figure 2. Expression profiles of five TGCT candidate genes according to various clinical and
pathological subgroups. The 506 testicular tissue cores, derived from orchiectomy specimens of 278
TGCT patients, were subgrouped according to several criteria. Each line represents one subgroup of
TGCT, and the colored squares illustrate the frequencies of immunopositive cases for the different
markers. Parenthesized numbers specify the number of analyzable tumors/patients from each
subgroup
than 500 cylindrical testicular tissue cores
into a single recipient block, we developed
a tool enabling us to analyze TGCT
candidate target genes in a large series of
samples of all histological subtypes and
stages, linked to a database with relevant
clinical,
pathological
and
genetic
information. We have used this tool to
7
Skotheim et al.
examine the protein expression of five
TGCT candidate genes. Four of these (JUP,
GRB7, MGMT, and FHIT) were recently
targeted by us,8,10,16 and the fifth is the
CCND2 candidate gene on chromosome
arm 12p.8,12-15
JUP and GRB7 showed high
expression in TGCT but not in normal
testicular tissues in a cDNA microarray
study focusing into chromosome arm 17q,8
which is over-represented in every second
TGCT.11 By transcriptional profiling using
DNA microarrays, many genes are usually
analyzed in a relatively small sample set,
often identifying a molecular signature of
the tumor in question, but with weak
statistics regarding the individual genes.
However, by analyzing the two candidate
genes JUP and GRB7 further on the TGCT
tissue microarray, we are confident about
their overexpression in TGCTs, also on
their protein levels, and evidence was
provided for their differential expression
across the various histological subtypes of
TGCT.
JUP belongs to the catenin family
and may have oncogenic potential through
its function in the WNT signaling
pathway.23,24 The tissue microarray data
demonstrated that JUP protein is rarely
expressed in normal spermatogenic germ
cells, even though it is expressed in
virtually all CIS and seminomas, and in
most nonseminomas. However, it remains
to be elucidated whether induction of JUP
expression is an initial event in
development of CIS, or if JUP is already
expressed in the fetal gonocytes, the germ
cell precursors of which CIS is believed to
originate from.2,25 The fact that the WNT
pathway is involved in embryogenesis,
which again is mimicked by the testicular
tumorigenesis, makes components of this
8
pathway
interesting
candidates
for
examination. The JUP binding partner Ecadherin is expressed in embryonal
carcinomas26,27 and the JUP-homolog Ecatenin is expressed in both normal and
malignant testicular tissues,27 but the
impact of WNT signaling in TGCT remains
poorly understood.
GRB7 encodes an adaptor protein
that through its SH2-domain interacts with
the cytoplasmic domain of several tyrosine
kinase growth factor receptors, including
ERBB2, KIT, PDGFR, RET, and INSR,28-32
as well as with cytoplasmic tyrosine
kinases.33,34 GRB7 also has a RASassociating-like domain,35 and plays a role
in cell migration.36,37 The TMA results for
GRB7
confirm
that
positive
immunostaining is more frequent in CIS,
seminomas, and nonseminomas than in
normal testicular tissues, but with the
highest frequency in nonseminomas. But
still, seminoma components within
combined TGCTs were not more often
positive than the pure seminomas. Among
teratomas, the epithelial part was generally
positive, but not the mesenchymal
component. In breast, esophageal, and
gastric cancers GRB7 is often coamplified
and co-overexpressed with ERBB2.28,38-40
Although we do not have copy number data
for these two genes, the combined CGH
and cDNA microarray analyses of TGCT
showed that this chromosome region is
overrepresented,11 but with overexpression
only of GRB7,8 suggesting that GRB7
interacts with another main target than
ERBB2 in these cells.
CCND2 is located at chromosome
arm 12p, and several studies have noted its
high expression in TGCT,8,12-15 which most
likely reflects the DNA sequence copy
number gains, seen in virtually all TGCTs,
TGCT tissue microarray
and often as an i(12p).4,6,41 However,
CCND2 can also be induced down-stream
of several molecular pathways such as RAS
and WNT-signalling.42,43 We noted that
56% of the TGCTs in our series were
immunopositive for CCND2, which is
somewhat lower than the frequency found
by Bartkova and coworkers (n=31, 81%).14
In CIS, the frequency of CCND2 positives
was intermediate between the always
negative normals and the TGCT samples.
One might speculate whether we
underestimate the frequency of CCND2
positives in CIS, as there are fewer CIS
nuclei in each tissue core than for instance
seminoma nuclei in a seminoma tissue core.
However, as parallel sections were stained
with antibodies against germ cell and
placental alkaline phosphatases, extensively
present in CIS but absent in normal
spermatogenic germ cells, we saw that there
were usually about 50 CIS cells in each CIS
tissue core. Among the invasive tumors, we
confirmed that the CCND2 expression is
not associated with histological subtype
(p=1.0), which is previously reported for
the mRNA level.15 CCND2 mRNA
expression has been shown to correlate with
the mRNA expression of its protein binding
partner CDK4,15 and in the present study
we demonstrated that its protein expression
correlated to those of GRB7 and MGMT.
Interestingly, we found in our series that
TGCTs of patients with history of
cryptorchidism had a lower frequency of
CCND2 immunoreactivity. However, the
biological impact is an enigma.
MGMT is a DNA repair gene which
we recently demonstrated to be frequently
inactivated in TGCT by promoter
hypermethylation.9,16 In the present study
we have shown that also the protein product
is silenced, in particular in seminomas and
embryonal carcinomas. However, the
MGMT protein seems to be re-expressed
upon further differentiation of embryonal
carcinoma into choriocarcinoma, yolk sac
tumor, and teratomas. Hence, a reversible
silencing mechanism seems plausible,
which fits well with the observed
hypermethylation of the MGMT promoter.9
FHIT is also a newly identified
target gene in TGCT.10 We confirmed that
the FHIT protein is downregulated in half
of the TGCTs compared to normal
testicular tissue. The downregulation seems
to take place when CIS is transformed into
invasive
TGCT
(p=0.009).
The
immunoreactivity of FHIT was strongly
associated with those of GRB7 and MGMT.
However, we have here failed to confirm
the associations proposed by the initial
study10 between reduced FHIT expression
and metastasis (present study, p=0.8) and
that mesenchymal components of teratomas
have more frequently reduced expression
compared to the epithelial components
(present study, p=0.5). Ten whole-mount
sections of TGCTs that also were present
on the tissue microarray were analyzed for
FHIT staining, yielding the same score in
eight out of the ten cases. Hence, it is likely
that the conflicting conclusions of this and
the previous FHIT study were not due to
the different technologies, but rather due to
the limited sample size and borderline
significance levels of the initial study.
For all tested markers, the
frequencies of immunoreactive cases were
similar for pure seminomas and seminoma
components of combined TGCTs. Thus,
this gives evidence for seminomas of both
groups to be evaluated together in the same
category in molecular studies of TGCT.
Additionally, this speaks in favor of
seminomas developing through the same
9
Skotheim et al.
molecular-pathological pathway irrespectively of whether it is pure or in combination
with nonseminoma components.
In summary, we have constructed a
TGCT tissue microarray on which we have
evaluated the protein expression of five
candidate target genes. We found that JUP
was upregulated and MGMT was
downregulated upon initiation of CIS, and
that upregulation of CCND2, downregulation of FHIT, and further downregulation of MGMT were related to
development of invasive tumors. GRB7 is
upregulated and JUP downregulated in the
transition into embryonal carcinoma, no
matter whether embryonal carcinomas
develop directly from CIS or through a
seminoma stage. Additionally, MGMT is
re-expressed during further differentiation
of embryonal carcinomas. Hence, we have
demonstrated that our tissue microarray
enables high-throughput evaluation of
TGCT markers, and we have utilized this
tool to validate five TGCT candidate genes
whose protein expressions were indeed
deregulated.
3.
4.
5.
6.
7.
8.
9.
10.
ACKNOWLEDGEMENTS
This work was supported by grants
from the Norwegian Cancer Society
(R.A.L.). R.I.S. is a research fellow for the
Research Council of Norway. We
acknowledge Liv Inger Håseth for mining
the tissue archive and Ellen Hellesylt for
her assistance in the immunohistochemistry
part.
11.
12.
13.
REFERENCES
1. Mostofi FK, Sesterhenn IA: World Health
Organization
International
Histological
Classification of Tumours: Histological
typing of testis tumours. Berlin SpringerVerlag, 1998
2. Rørth M, Rajpert-De Meyts E, Andersson L,
Dieckmann KP, Fosså SD, Grigor KM,
10
14.
Hendry WF, Herr HW, Looijenga LH,
Oosterhuis JW, Skakkebæk NE: Carcinoma
in situ in the testis. Scand J Urol Nephrol
Suppl 2000,166-186
Skotheim RI, Lothe RA: The testicular germ
cell tumour genome. APMIS 2003, 111:136151
Atkin NB, Baker MC: Specific chromosome
change, i(12p), in testicular tumours?
Lancet 1982, 2:1349
Rodriguez E, Mathew S, Reuter V, Ilson DH,
Bosl GJ, Chaganti RS: Cytogenetic analysis
of 124 prospectively ascertained male germ
cell tumors. Cancer Res 1992, 52:22852291
Rodriguez E, Houldsworth J, Reuter VE,
Meltzer P, Zhang J, Trent JM, Bosl GJ,
Chaganti
RS:
Molecular
cytogenetic
analysis of i(12p)-negative human male
germ cell tumors. Genes Chromosomes
Cancer 1993, 8:230-236
Smiraglia DJ, Szymanska J, Kraggerud SM,
Lothe RA, Peltomäki P, Plass C: Distinct
epigenetic phenotypes in seminomatous
and nonseminomatous testicular germ cell
tumors. Oncogene 2002, 21:3909-3916
Skotheim RI, Monni O, Mousses S, Fosså
SD, Kallioniemi O-P, Lothe RA, Kallioniemi
A: New insights into testicular germ cell
tumorigenesis from gene expression
profiling. Cancer Res 2002, 62:2359-2364
Smith-Sørensen B, Lind GE, Skotheim RI,
Fosså SD, Fodstad Ø, Stenwig AE,
Jakobsen KS, Lothe RA: Frequent promoter
6
hypermethylation of the O -MethylguanineDNA methyltransferase (MGMT) gene in
testicular cancer. Oncogene 2002, 21:8878
Kraggerud SM, Åman P, Holm R, Stenwig
AE, Fosså SD, Nesland JM, Lothe RA:
Alterations of the Fragile Histidine Triad
Gene, FHIT, and Its Encoded Products
Contribute to Testicular Germ Cell
Tumorigenesis. Cancer Res 2002, 62:512517
Kraggerud SM, Skotheim RI, Szymanska J,
Eknæs M, Fosså SD, Stenwig AE,
Peltomäki P, Lothe RA: Genome profiles of
familial/bilateral and sporadic testicular
germ cell tumors. Genes Chromosomes
Cancer 2002, 34:168-174
Sicinski P, Donaher JL, Geng Y, Parker SB,
Gardner H, Park MY, Robker RL, Richards
JS, McGinnis LK, Biggers JD, Eppig JJ,
Bronson RT, Elledge SJ, Weinberg RA:
Cyclin D2 is an FSH-responsive gene
involved in gonadal cell proliferation and
oncogenesis. Nature 1996, 384:470-474
Houldsworth J, Reuter V, Bosl GJ, Chaganti
RS: Aberrant expression of cyclin D2 is an
early event in human male germ cell
tumorigenesis. Cell Growth Differ 1997,
8:293-299
Bartkova
J,
Rajpert-De
Meyts
E,
Skakkebæk NE, Bartek J: D-type cyclins in
adult human testis and testicular cancer:
relation
to
cell
type,
proliferation,
differentiation, and malignancy. J Pathol
1999, 187:573-581
TGCT tissue microarray
15. Schmidt BA, Rose A, Steinhoff C,
Strohmeyer T, Hartmann M, Ackermann R:
Up-regulation of cyclin-dependent kinase
4/cyclin D2 expression but down-regulation
of cyclin-dependent kinase 2/cyclin E in
testicular germ cell tumors. Cancer Res
2001, 61:4214-4221
16. Smith-Sørensen B, Lind GE, Skotheim RI,
Fosså SD, Fodstad Ø, Stenwig AE, Lothe
6
RA: Alterations of the O -methylguanineDNA methyltransferase (MGMT) contribute
to development of testicular cancer.
Oncogene 2002 (in press)
17. Kononen J, Bubendorf L, Kallioniemi A,
Bärlund M, Schraml P, Leighton S, Torhorst
J, Mihatsch MJ, Sauter G, Kallioniemi OP:
Tissue microarrays for high-throughput
molecular profiling of tumor specimens. Nat
Med 1998, 4:844-847
18. Moch H, Schraml P, Bubendorf L, Mirlacher
M, Kononen J, Gasser T, Mihatsch MJ,
Kallioniemi OP, Sauter G: High-throughput
tissue microarray analysis to evaluate genes
uncovered by cDNA microarray screening in
renal cell carcinoma. Am J Pathol 1999,
154:981-986
19. Torhorst J, Bucher C, Kononen J, Haas P,
Zuber M, Köchli OR, Mross F, Dieterich H,
Moch H, Mihatsch M, Kallioniemi OP,
Sauter G: Tissue microarrays for rapid
linking of molecular changes to clinical
endpoints. Am J Pathol 2001, 159:22492256
20. Hendry WF, Barrett A, McElwain TJ,
Wallace DM, Peckham MJ: The role of
surgery in the combined management of
metastases from malignant teratomas of
testis. Br J Urol 1980, 52:38-44
21. Hofmann MC, Millán JL: Developmental
expression of alkaline phosphatase genes;
reexpression in germ cell tumours and in
vitro immortalized germ cells. Eur Urol 1993,
23:38-45
22. Greenspan DL, Connolly DC, Wu R, Lei RY,
Vogelstein JT, Kim YT, Mok JE, Munoz N,
Bosch FX, Shah K, Cho KR: Loss of FHIT
expression in cervical carcinoma cell lines
and primary tumors. Cancer Res 1997,
57:4692-4698
23. Kolligs FT, Kolligs B, Hajra KM, Hu G, Tani
M, Cho KR, Fearon ER: Gamma-catenin is
regulated by the APC tumor suppressor and
its oncogenic activity is distinct from that of
beta-catenin. Genes Dev 2000, 14:13191331
24. Barker N, Clevers H: Catenins, Wnt
signaling and cancer. Bioessays 2000,
22:961-965
25. Skakkebæk NE, Berthelsen JG, Giwercman
A, Muller J: Carcinoma-in-situ of the testis:
possible origin from gonocytes and
precursor of all types of germ cell tumours
except spermatocytoma. Int J Androl 1987,
10:19-28
26. Heidenreich A, Sesterhenn IA, Mostofi FK,
Moul JW: Prognostic risk factors that identify
patients
with
clinical
stage
I
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
nonseminomatous germ cell tumors at low
risk and high risk for metastasis. Cancer
1998, 83:1002-1011
Saito T, Katagiri A, Watanabe R, Tanikawa
T, Kawasaki T, Tomita Y, Takahashi K:
Expression of E-cadherin and catenins on
testis tumor. Urol Int 2000, 65:140-143
Stein D, Wu J, Fuqua SA, Roonprapunt C,
Yajnik V, D'Eustachio P, Moskow JJ,
Buchberg AM, Osborne CK, Margolis B: The
SH2 domain protein GRB-7 is co-amplified,
overexpressed and in a tight complex with
HER2 in breast cancer. EMBO J 1994,
13:1331-1340
Pandey A, Liu X, Dixon JE, Di Fiore PP,
Dixit VM: Direct association between the
Ret receptor tyrosine kinase and the Src
homology 2-containing adapter protein
Grb7. J Biol Chem 1996, 271:10607-10610
Yokote K, Margolis B, Heldin CH, ClaessonWelsh L: Grb7 is a downstream signaling
component of platelet-derived growth factor
alpha- and beta-receptors. J Biol Chem
1996, 271:30942-30949
Thömmes K, Lennartsson J, Carlberg M,
Rönnstrand L: Identification of Tyr-703 and
Tyr-936 as the primary association sites for
Grb2 and Grb7 in the c-Kit/stem cell factor
receptor. Biochem J 1999, 341:211-216
Kasus-Jacobi A, Bereziat V, Perdereau D,
Girard J, Burnol AF: Evidence for an
interaction between the insulin receptor and
Grb7. A role for two of its binding domains,
PIR and SH2. Oncogene 2000, 19:20522059
Keegan K, Cooper JA: Use of the two hybrid
system to detect the association of the
protein-tyrosine-phosphatase,
SHPTP2,
with another SH2-containing protein, Grb7.
Oncogene 1996, 12:1537-1544
Han DC, Guan JL: Association of focal
adhesion kinase with Grb7 and its role in
cell migration. J Biol Chem 1999,
274:24425-24430
Wojcik J, Girault JA, Labesse G, Chomilier
J, Mornon JP, Callebaut I: Sequence
analysis identifies a ras-associating (RA)like domain in the N-termini of band 4.1/JEF
domains and in the Grb7/10/14 adapter
family. Biochem Biophys Res Commun
1999, 259:113-120
Manser J, Roonprapunt C, Margolis B: C.
elegans cell migration gene mig-10 shares
similarities with a family of SH2 domain
proteins and acts cell nonautonomously in
excretory canal development. Dev Biol
1997, 184:150-164
Han DC, Shen TL, Guan JL: Role of Grb7
targeting to focal contacts and its
phosphorylation by focal adhesion kinase in
regulation of cell migration. J Biol Chem
2000, 275:28911-28917
Akiyama N, Sasaki H, Ishizuka T, Kishi T,
Sakamoto H, Onda M, Hirai H, Yazaki Y,
Sugimura T, Terada M: Isolation of a
candidate gene, CAB1, for cholesterol
transport to mitochondria from the c-ERBB-
11
Skotheim et al.
2 amplicon by a modified cDNA selection
method. Cancer Res 1997, 57:3548-3553
39. Tanaka S, Mori M, Akiyoshi T, Tanaka Y,
Mafune K, Wands JR, Sugimachi K:
Coexpression of Grb7 with epidermal
growth factor receptor or Her2/erbB2 in
human advanced esophageal carcinoma.
Cancer Res 1997, 57:28-31
40. Kauraniemi P, Bärlund M, Monni O,
Kallioniemi A: New amplified and highly
expressed genes discovered in the ERBB2
amplicon in breast cancer by cDNA
microarrays. Cancer Res 2001, 61:82358240
41. Suijkerbuijk RF, Sinke RJ, Meloni AM,
Parrington JM, van Echten J, de Jong B,
Oosterhuis JW, Sandberg AA, Geurts van
12
Kessel
A:
Overrepresentation
of
chromosome 12p sequences and karyotypic
evolution in i(12p)-negative testicular germcell tumors revealed by fluorescence in situ
hybridization. Cancer Genet Cytogenet
1993, 70:85-93
42. Dey A, She H, Kim L, Boruch A, Guris DL,
Carlberg K, Sebti SM, Woodley DT,
Imamoto A, Li W: Colony-stimulating factor1 receptor utilizes multiple signaling
pathways to induce cyclin D2 expression.
Mol Biol Cell 2000, 11:3835-3848
43. Megason SG, McMahon AP: A mitogen
gradient of dorsal midline Wnts organizes
growth in the CNS. Development 2002,
129:2087-2098
Appendices
Appendix I. Abbreviations
Appendix II. Genes putatively involved in development of TGCT
blank
Appendix I. Abbreviationsa
AI
allelic imbalance
BAC
bacterial artificial chromosome
BEP
bleomycin, etoposid, and cisplatin
bp
base pair
cDNA
complementary DNA
CGH
comparative genomic hybridisation
CIS
carcinoma in situ
EST
expressed sequence tag
GCT
germ cell tumour
kb
kilo base pair (103)
LOH
loss of heterozygosity
Mbp
mega base pair (106)
mRNA
messenger RNA
NCBI
National Center for Biotechnology Information
PAC
P1-derived artificial chromosome
PCR
polymerase chain reaction
ref
reference
RLGS
restriction landmark genome scanning
RT-PCR
reverse transcription-PCR
SAGE
serial analysis of gene expression
siRNA
small interfering RNA
SNP
single nucleotide polymorphism
TDS
Testicular dysgenesis syndrome
TGCT
testicular germ cell tumour
a
Abbreviated gene names of TGCT related genes are listed in Appendix II.
Appendix II. Genes putatively involved in development of TGCT
For full names of other gene symbols, please refer to the GeneCards database at the Weizmann
Institute of Science, Rehovot, Israel; http://bioinformatics.weizmann.ac.il/cards/
Gene
symbolb
Alias
AFP
Locusc
References
D-fetoprotein
4q13.3
97,241
ALPP
PLAP
alkaline phosphatase, placental
2q37.1
65,242,243
ALPPL2
GCAP
alkaline phosphatase, placental-like 2
2q37.1
65,242-244
CCND2
cyclin D2
12p13.32
165-169 and
Papers IV & V
CCNE1
cyclin E1
19q12
168,223
CDH1
E-cadherin
cadherin 1, type 1, E-cadherin (epithelial)
16q22.1
231,232
CDKN2A
p16INK4A/
p14ARF
cyclin-dependent kinase inhibitor 2A
9p21.3
162,163,222
CDKN2C
p18INK4C
cyclin-dependent kinase inhibitor 2C
1p32.3
223
CDKN2D
p19
INK4D
cyclin-dependent kinase inhibitor 2D
19p13.2
64
CGB
E-hCG
chorionic gonadotropin, E polypeptide
19q13.33
97,241
CTAG1
NY-ESO-1
cancer/testis antigen 1
Xq28
245
DAD-Rd
DAD1-related gene
12p12.1
246
DCC
deleted in colorectal carcinoma
18q21.2
130,247
DEAD/H box polypeptide 4
5q11.2
248
DNMT2
DNA (cytosine-5)-methyltransferase 2
10p13
Paper IV
FHIT
fragile histidine triad gene
3p14.2
138,249 and Paper V
DDX4
VASA
GCT1
d
LOC51026
germ cell tumour 1
12p12.1
190
GCT2
d
FLJ10637
germ cell tumour 2
12p11.23
190
GRB7
growth factor receptor-bound protein 7
17q12
Papers IV & V
GSTP1
glutathione S-transferase S
11q13.3
173,250
HRAS
v-Ha-ras Harvey rat sarcoma viral
oncogene homolog
11p15.5
202,251
HRAS-like suppressor 3
11q13.1
252
junction plakoglobin
17q21.2
Papers IV & V
4q12
216-218,253,254
12q21.32
254
Ha-RAS
HRASLS3 HREV107
JUP
b
Gene name
J-catenin
KIT
c-KIT
KITLG
MGF, SCF
v-kit Hardy-Zuckerman 4 feline sarcoma
viral oncogene homolog
KIT ligand
Approved by the HUGO Gene Nomenclature Committee; http://www.gene.ucl.ac.uk/nomenclature/
According to the UCSC June 2002 assembly of the human genome; http://genome.ucsc.edu/
d
No symbol approved by the HUGO Gene Nomenclature Committee for these genes (Dec. 23, 2002).
c
Gene
symbolb
Alias
Gene name
Locusc
References
KLK10
NES1
kallikrein 10
19q13.33
255
KLK13
KLK-L4
kallikrein 13
19q13.33
256
KRAS2
K-RAS
v-Ki-ras2 Kirsten rat sarcoma 2 viral
oncogene homolog
12p12.1
115,119,200,202-204
LDHB
LDH
lactate dehydrogenase B
12p12.1
97,257,258
LLGL2
lethal giant larvae (Drosophila) homolog 2 17q25.1
Paper IV
MADH4
MAD, mothers against decapentaplegic
homolog 4
18q21.1
259
MAGEA4
melanoma antigen, family A, 4
Xq28
260
MDM2
Mdm2, transformed 3T3 cell double
minute 2, p53 binding protein (mouse)
12q15
101,143,145,261,262
MGMT
O6-methylguanine-DNA methyltransferase 10q26.3
163 and Paper V
MXI1
MAX interacting protein 1
10q25.2
Paper IV
20q13.12
Paper IV
1p34.2
Paper IV
2p24.3
164 and Paper IV
1p13.2
200-204
platelet-derived growth factor receptor, D
4q12
214,215,219,263
12q24.33
264
SMAD4
MYBL2
B-MYB
MYCL1
L-MYC
MYCN
N-MYC
NRAS
N-RAS
PDGFRA
v-myb myeloblastosis viral oncogene
homolog (avian)-like 2
v-myc myelocytomatosis viral oncogene
homolog 1, lung carcinoma derived
v-myc myelocytomatosis viral related
oncogene, neuroblastoma derived
v-ras neuroblastoma RAS viral oncogene
homolog
PIWIL1
HIWI
piwi-like 1 (Drosophila)
POU5F1
OCT-4
POU domain, class 5, transcription factor1 6p21.33
215,263
POV1
prostate cancer overexpressed gene 1
11q12.1
Paper IV
PTEN
phosphatase and tensin homolog
10q23.31
236,237
RB1
retinoblastoma 1
13q14.2
220
10q23.31
265-270
1q24.3
265-269
tissue inhibitor of metalloproteinase 2
17q25.3
Paper IV
tumour protein 53
17p13.1
100,101,131,139,141148,262
TNFRSF6 FAS
TNFSF6
FAS ligand
TIMP2
TP53
p53
tumour necrosis factor receptor
superfamily, member 6
tumour necrosis factor (ligand)
superfamily, member 6
blank