Frequent novel DNA copy number increase in squamous cell head and neck tumors.

Abstract

We have undertaken a study of DNA copy number changes in head and neck squamous cell carcinomas to identify novel DNA copy number changes and to determine the significance of previous findings of cytogenetic alterations in cultured cells. Comparative genomic hybridization was performed on genomic DNA extracted from ten tumors. A novel copy number gain on chromosome 3q26-27 and a loss of chromosome 3p were found at high frequency (> or = 50% of tumors). Many other novel chromosomal copy number changes were identified but occurred at a lower frequency. In addition, our data confirm that DNA copy number changes that frequently occur in cultured cells, such as loss of chromosome 3p, also occur in tumors. Frequently altered loci may encode oncogenes or tumor suppressor genes involved in head and neck squamous cell carcinoma tumorigenesis.

Full text

Page 1

1995;55:3055-3059. Published online July 1, 1995.Cancer Res


Pius M. Brzoska, Nikki A. Levin, Karen K. Fu, et al.
Head and Neck Tumors
Frequent Novel DNA Copy Number Increase in Squamous Cell



Updated Version
http://cancerres.aacrjournals.org/content/55/14/3055
Access the most recent version of this article at:


Citing Articles
http://cancerres.aacrjournals.org/content/55/14/3055#related-urls
This article has been cited by 11 HighWire-hosted articles. Access the articles at:


E-mail alerts
related to this article or journal.Sign up to receive free email-alerts
Subscriptions
Reprints and
.pubs@aacr.orgDepartment at
To order reprints of this article or to subscribe to the journal, contact the AACR Publications
Permissions
.permissions@aacr.orgDepartment at
To request permission to re-use all or part of this article, contact the AACR Publications
American Association for Cancer Research Copyright © 1995
on July 14, 2011cancerres.aacrjournals.orgDownloaded from

Page 2

[CANCER RESEARCH 55, 3055-3059, July 15. 1995]
Frequent Novel DNA Copy Number Increase in Squamous Cell
Head and Neck Tumors1
Pius M. Brzoska, Nikki A. Levin, Karen K. Fu, Michael J. Kaplan, Mark I. Singer, Joe W. Gray, and
Michael F. Christman2
Department
J. W. G.¡,and Department
of RadiationOncology [P. M. B., N. A. L, K. K. F., J. W. G.. M. F. C.I and Division of Molecular
of Otolaryngalogy ¡M.J. K., M. 1. S.J, University of California, San Francisco, California 94143
Cytometry in the Departmentof Laboratory- Medicine [N. A. L,
ABSTRACT
We have undertaken
neck squamous
changes and to determine
netic alterations in cultured cells. Comparative
performedon genomic
number gain on chromosome
found at high frequency
a study of DNA copy number changes in head and
cell carcinomasto identify
the significance of previous findings of cytoge-
novel DNA copy number
genomic hybridization
from ten tumors.
and a loss of chromosome
50' » of tumors). Many other novel chromo
was
DNA extracted
3q26-27
i
A novel copy
3p were
somal copy number
frequency.
that frequently occur in cultured cells, such as loss of chromosome
occur in tumors. Frequently altered loci may encode oncogenes or tumor
suppressor genes involvedin head and neck squamous
tumorigenesis.
changeswere identifiedbut occurredat a lower
In addition, our data confirm that DNA copy number changes
3p, also
cell carcinoma
INTRODUCTION
Forty-two thousand new cases of head and neck cancer occurred in
the United States in 1992 and there were 11,600 deaths (1). Tobacco
and alcohol are the most important etiological
literature cited therein). A variety of genetic alterations and complex
aberrant karyotypeshave been described
from HNSCC1 tumors, including copy number decreases of chromo
some arms 3p (2-4),llq (5, 6), and 18q (7-9) and loss of chromo
somes 9,12, 13, 8, and Y (10). Karyotyping and loss of heterozygosity
analysis revealed rearrangements and translocations on chromosomes
Ip, Iq, 3, 5q, 8, llq,13p, 14p, and 15p (4, 9-11). Point mutations in
the p53 tumor suppressorgene (12-16)
cultured cells. However, culture conditions are known to influence the
karyotype of cells derived from head and neck tumors (10). Therefore,
to corroborate the occurence of the chromosomal aberrations reported
in cell cultures and, thus, confirm their significance in head and neck
tumorigenesis, and to identify novel alterations, molecular cytogenetic
analysis of fresh tumor cells was performed.
We report here the analysis of 10 tumors with the use of CGH (17,
18), a novel cytogeneticmethod that allows the simultaneous
mination of DNA copy number changes in a tumor specimen in a
single experiment. CGH is very powerful because anonymous DNA
copy number changes can be identified rapidly (19-22)
tumors can be analyzed directly without culturing or karyotyping. We
have found that a novel DNA copy number increase at 3q26-27
most frequent cytogenetic alteration in HNSCC. Common DNA copy
number decrease at 3p, which was observed previously
cells, is also found in tumors by CGH. We compare our data to a
agents (Ref. l and
in cultured cells derived
have also been found in
deter
and because
is the
in cultured
Received 12/16/94; accepted 5/16/95.
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
18 U.S.C. Section 1734 solely to indicate this fact.
' This study was supported in part by both the United States Department of Energy
(under contract DE-AC-03-76SF00098)and Imagenetics.
2To whom requests for reprints should be addressed,
Oncology, Universityof California, 1855 Folsom
Francisco. CA 94103.
^ The abbreviations used are: HNSCC, head and neck squamous cell carcinoma; CGH,
comparativegenomichybridization;DAP1, 4,6-diamino-2-phenylindole;
mere; MB, megabase.
in accordance with
at Department
Street, MCB200,
of Radiation
Box 0806, San
Cen, centro
similar CGH analysis of head and neck tumors that was published
recently (23).
MATERIALS AND METHODS
Patients.
head and neck were analyzed by CGH. Specimens were obtained from surgical
resections of the primary tumors or of recurrent tumors as indicated in Table
1.The tumor site and level of differentiation
tumor, a pathologist removed a small specimen
attempting to avoid the inclusion of contaminating
Ten tumors from patients with squamous cell carcinomas of the
are also listed in Table 1. For each
from the tumor mass while
normal cells. The "TNM"
classification
DNA Preparation.
tissue (0.3-0.5
in a solution of 10 ITIM Tris-HCl (pH 8)-100 mM NaCI-1 mM EDTA-0.5 mg/ml
proteinaseK (Boehringer Mannheim, Indianapolis,
following day the dissolved tissue was extracted with phenol-chloroform
four to five times and precipitated with ethanol. The DNA was resuspended
50 JA!of 10 mM Tris-HCl-1 mM EDTA (pH 8), and the concentration
determinedby absorbance at 260 nm.
ComparativeGenomic Hybridization.
essentiallyas described(17). Normal
preparations were denatured in 70% formamide and 2 x SSC [1 x SSC is 0.15
M NaCl and 0.015 sodium citrate (pH 7)] at temperatures
78°Cfor 2.5-10 min depending on the age of the slides. Slides that had been
is based on a system described previously
The DNA was prepared as described (1°). Tumor
g) was cut into small pieces with a razor blade and incubated
(24).
IN) at 45°Cfor 12 h. The
(1:1)
in
was
Hybridizations were performed
human male lymphocytemetaphase
ranging from 72 to
stored at room temperature
frozen generally required longer denaturation
(78°C)for optimal hybridization.
for more than 2 weeks or that had been stored
(10 min) at higher temperatures
The slides were then dehydratedin a se
quence of 70, 85, and 100% ethanol washes. The slides were air dried and
incubated at 37°Cuntil the DNA probes were applied (<5 min). One hundred
and twenty ng of biotinylatedtumor DNA, 120 ng of digoxygenin-labeled
normal reference DNA, and 5 f¿g of Cot-1-blocking
precipitated with ethanol. The pellet was resuspended
and mixed with 7 /J of hybridization
50% formamide,10% dextran sulfate, and 2 x SSC (pH 7). This probe mixture
was denatured for 5 min at 77°C and incubated at 37°Cfor 10 min before being
DNA (GIBCO-BRL)
in 3 /xl of distilled H2O
buffer to give a final concentration
were
of
applied to the metaphase
slides with rubber cement,
chamber for 3 days at 37°C.
Slides were washed at 45°Cin three changes of 50% formamide-2
spreads. Coverslips
and the slides were incubated
were applied and sealed to the
in a humidified
X SSC
(pH 7) followed by two washes with 2 X SSC and one wash with 0.1 X SSC
(10 min in each wash). All subsequent
temperature. The slides were washed for 5 min in 2 X SSC and blocked for 5
min in 2 X SSC-1% BSA (Pentax Fraction V; Sigma). The slides were then
immunohistochemically stained with 5 /ig/ml
chrome; Vector Laboratories, Burlingame,
rhodamine (a red fluorochrome;Boehringer Mannheim) in 2 X SSC-1% BSA
for 30 min in the dark. The slides were then washed in the dark for 10 min
successively in each of the following:
Triton X-100 (Fluka Chemika),4 X SSC and PN (0.1 M NaH2PO4-0.1
Na2HPO4-and 0.1% NP-40). The slides were drained on a paper towel, and 7.5
fil of a solution of 0.1 /IM DAPI in glycerol containing
Coverslips were applied and sealed with nail polish (Sally Hansen's
manipulations were done at room
FITC-avidin
CA) and 2 fig/ml anti-digoxigenin
(a green fluoro-
4 X SSC, 4 X SSC containing0.1%
M
antifade was applied.
Hard as
Nails).
Image Acquisition
Zeiss Axioplan fluorescence microscope. Metaphase images were acquired and
stored as three color images (DAPI, FITC, and rhodamine)
charge-coupleddevice camera (Photometries,
and Processing. The slides were examined with a
with a cooled
Tucson, AZ) interfaced with a
3055
American Association for Cancer Research Copyright © 1995
on July 14, 2011 cancerres.aacrjournals.orgDownloaded from

Page 3

GENETIC ALTERATIONS IN HEAD AND NECK CANCER
Table 1 Tumors analyzed from 10 patients with squamous ceil carcinomaof the head and neck
Case12345678910Clinical
stageT2NOMO/IIT2N1M1/IVT2N1MO/IIIT3NOMO/IIIT3NOMO/IIIT3NUMO/IIIT3NOM1/IVT3N1MO/1IIT2N2MO/IVT3N2M1/IVSexMaleMaleMaleMaleFemaleMaleMaleMaleFemaleMaleAge(yr)84725948676847567175Disease
at time ofsurgeryRecurrentRecurrentRecurrentRecurrentRecurrentPrimaryPrimaryRecurrentPrimaryRecurrentTumor siteLarynxSinusPalateVocal
status
cordLarynxSinusMouth
floorVocal
cordSinusLarynxTumor
differentiationPoor-moderatePoorPoorPoor-moderate,
invasivePoorModerateModerate,
invasiveModerate,
invasiveModerateModerate,
invasivePrevious
radiation
treatmentYesYesYesYesYesNoNoYesN
Sun 4/330
system. Chromosomes
(17). The green:red fluorescence
of the fluorescence
XWOOLZprogram (17). At least three metaphase
determine the ratio of green:red fluorescence for each chromosome.
ten tumor samples were also analyzed
chromes reversed (digoxigenin dUTP for the tumor DNA and biotin dUTP for
the normal DNA). If the green:red ratio was >1.2 or <0.8 the alteration was
scored as a gain or a loss because, in our experience, control CGH experiments
using normal-normalDNAs do not show this much variation in the green:red
fluorescenceratio. The extent and location of each copy number change (based
on the mean fluorescence ratio profiles) were plotted on an ideogram showing
a schematic representationof the G-banding
Heterochromaticchromosome regions (Ceni,
21p, 22p, and Yql2) were not interpretable
ization with Cot-1 DNA in these regions.
work stationequipped
were identified based on the DAPI-banding
ratio profiles were determined by integration
intensity along the axis of each chromosome
with the quantitativeimage processing
patterns
by use of the
spreads were analyzed to
Five of the
with the normal and tumor fluoro-
pattern
Cen9, Cenlo,
because of suppression of hybrid
of the chromosome.
13p, 14p, 15p,
RESULTS
CGH was performed on DNA from 10 HNSCC tumor specimens.
Table 1 shows the clinical stage for each of the patients. Seven of the
patients had received radiation therapy before and three had not at the
time of surgery. The tumors varied somewhat in size and node status,
and three showed métastases.Fig. 1/4 shows a representative
metaphase for case 5. Fig. Ißshows the corresponding
fluorescence ratio profiles. A green:red
indicates a copy number increase of a particular chromosomal region,
whereasa green:redfluorescence
numberdecrease.The tumor shown
increases on chromosomes 3q (Fig. ÕA, red arrows) and 9p. Because
the HNSCC tumor is from a female patient and the normal DNA is
from a male, chromosomeX from the patient DNA shows a 2-fold
increase in copy number resulting in a green appearance of chromo
some X in the metaphase, and chromosome Y, which is absent from the
female tumor, appears to be underrepresented,
chromosome in the metaphase (green:red ratio is much smaller than 1).
A schematic representationof all DNA copy number changes from
the 10 HNSCC cancers analyzed is shown in Fig. 2. Copy number
changes on chromosome 3 occurred in 70% of the cases. Fifty % of
the tumors showed a copy number decrease on chromosome
"p" arm is the top arm of each chromosome
CGH
green:red
ratio of >1.2 fluorescence
ratio of <0.8 indicates
in Fig. 1 has copy number
a copy
resulting in a red Y
3p (the
in Fig. 2) and 50%
showed a copy number increase of chromosome
arm), whereas 30% did not show an alteration on chromosome
Many tumors showed copy number changes on chromosomes
than 3 but only at frequencies of 40% or less. Copy number decrease
of all of chromosomes 19, 22, and Y and copy number decreases on
chromosome arms lip and 16p occurred in 30% of the tumors. The
telomeric regions of chromosomes
in copy number in 40% of the tumors. Copy number increases of
3q (Fig. 2, bottom
3.
other
(Ip, 5q, and 19q) were decreased
chromosome
8q23-24,
A more detailed representation
chromosome 3 for each individual case is shown in Fig. 3. Closer
inspection reveals that the region of copy number increase on 3q in the
individual tumors is not identical but that each spans the region at
band 3q26-27.Three tumors (nos. 6, 7, and 8) showed a simultaneous
copy number decrease of 3p and a copy number increase of 3q26-27,
two showed only the 3p copy number decrease (nos. 5 and 9), and two
tumors (nos. 2 and 10) showed the 3q26-27
alone.
bands 2q22-24.1, 2q32,3ql3.3,4qll-27,7q31.3,
and 13q21-31 occurred in 30% of the tumors analyzed.
of the copy number changes on
copy number increase
DISCUSSION
The cytogenetic alterations that arise in HNSCCs have been char
acterized extensively (1, 5, 9). They include frequent deletions at 3p,
5q, 8p, 9p, and 18q and frequent gains at 3q, 5p, 7p, 8q, and 1Iq (7).
To corroboratethe occurrenceof the chromosomal
ported in cell cultures and, thus, confirm their significance in head and
neck tumorigenesis, and to identify novel alterations,
togenetic analysis of fresh tumor cells was performed. CGH is very
powerful because it allows one to determine relative copy number of
nearly the entire genome without growing or karyotyping tumor cells.
CGH also allows previously unrecognized
(19, 21, 22, 25). Comparative genomic hybridization
on genomic DNA extracted from ten tumors. A novel copy number
gain on chromosome 3q26-27and a loss of chromosome
found at high frequency (a50%of tumors).
The most frequent DNA copy number changes detected were on
chromosome 3. Several studies have described copy number decreases
of chromosome 3p in cultured cells and primary tumors (2, 4, 26) and
we confirm here that these copy number decreases occur frequently in
primary tumors analyzed by CGH (50% of the cases). Copy number
increases of the entire 3q chromosome
cultured cells at a frequency of 50% (26). Our data confirm the high
incidence of copy number increase on chromosome 3q demonstrating
that classical and molecular cytogenetic methods can yield congruent
findings. Moreover, our data indicate that the minimal common region
of copy number increase on 3q occurs in a small region at 3q26-27.
3q26 is frequently involved in translocations
myelogenous leukemia (27, 28). LAZ3, a Zn finger protein, maps at
3q27. t(3;14) and t(3;4) translocations
disruption of LA7.3 in non-Hodgkin's
aberrations re
molecular cy
alterations to be identified
was performed
3p were
arm have been described in
with chromosome21 in
have been shown to lead to a
lymphoma (29, 30). BCL-6 is
another Zn finger protein that maps to 3q27 (31, 32). Patients with
diffuse, large cell lymphomas often show rearrangements
5' non-coding region of the gene, suggesting that BCL-6 is a proto
oncogene.Because we find that 3q26-27
these two genes are candidates for involvement
within the
is frequently
in tumor progression
amplified,
3056
American Association for Cancer Research Copyright © 1995
on July 14, 2011cancerres.aacrjournals.org Downloaded from

Page 4

GENETIC ALTERATIONS IN HEAD AND NECK CANCER
Fig. 1. A, a representative
from lumor no. 5. Tumor DNA was labeled with biotin
dUTP and normal DNA with digoxigenin
DNA was detected with FITC-avidin (green),
DNA was detected with rhodamine anti-digoxigenin
body (red). Green areas are due to DNA copy number
increases in the tumor, and red areas are due to copy
number decreases. Chromosomes
Red arrow, a common amplification at 3q26-27.
red fluorescence ratio profiles of the HNSCC tumor DNA
shown in A. The green:red ratio is shown as the solid line
in the diagram. Ratio profiles are presented with p arms to
the left and q arms to the right of the vertical mark, which
indicatesthe position of the centromeres.
green:red fluorescence ratio of one; dolled lines, ratios 0.5
(lower) and 1.5 (upper). Numbers to the left of each trace
indicate chromosomenumber: n. number of independent
chromosomesanalyzed. A green:red ratio >1.2 indicates a
copy number increase, and a ratio <0.8 indicates a copy
number decrease in the tumor DNA. The thin solid lines
near the green:red ratio line are SDs.
CGH metaphase with DNA
dUTP. Tumor
and normal
anti
are labeled at the p arm.
B. green:
Dashedline,
W*
6
7
•d
8
n=5
9
»-4
15
ft
16
ioti
-fc
I2L1
Mean ands d.
3057
American Association for Cancer Research Copyright © 1995
on July 14, 2011 cancerres.aacrjournals.orgDownloaded from

Page 5

Fig. 2. Idiogram of DMA copy number changes in
10 head and neck tumors. Bars on the left side of the
chromosomes indicate copy number decreases;
on the right side show copy number increases. Num
bers on top of the bars refer to the tumors described in
Table 1. The idiogram shows all copy number changes
found.
bars
GENETICALTERATIONSIN HEAD AND NECK CANCER
Loss Gain
98763 3689
3676
873863
9698767
12
73 6e678763 763
13 14 15 1617 1819 20 2122
in squamous cell head and neck cancer. Studies of the expression level
of the respective proteins or the mRNAs might shed light on their
possible involvement in HNSCC tumorigenesis.
Our data are quite consistent with a recently published CGH anal
ysis of 13 head and neck tumors (23). In that study the most frequent
copy number gain was found to be 3q26-qter, and the most frequent
loss was 3p. These are also the most common alterations found in our
study, although our data suggest a somewhat narrower region for the
common copy number gain (3q26-27).
was found in 8 of 13 tumors by Speicher et al. (23) whereas we found
5p increase in only 2 of 10 tumors. The difference may simply reflect
the small number of tumors analyzed. Both studies suggest that copy
number increase in the 3q26-27 region may be an important event in
head and neck tumorigenesis. Amplification
observed recently in several other tumor types, including brain, blad-
A copy number gain on 5p
of 3q26 has also been
Loss
98764
Gain
267810
Common region of
copy number gain
der, ovarian, and cervical cancer.4 Whether a common oncogene is
involved in tumorigenesis
through identification
In HNSCC tumors, copy number increase of BCL-I and INT-2 on
Ilql3 and of the epidermal growth factor receptor on 7p have been
reported (5, 6, 33). Only one of our patients [and two in the Speicher
et al. (23) study] showed a copy number increase in this region. Most
likely copy number increases of those loci are below the detection
level of CGH. The ampliconsize of the INT-2 regions has been
reported to range from <1 to 4.5 MB (34). The minimal size of copy
number changes that can be detected by CGH is about 2 MB (for
50-fold copy number increase) to 10 MB (for 5-fold copy number
increase) for cultured cells. In CGH analysis of tumors, these values
may be higher if the tumor is contaminated
in our analysis we analyzed regions of the tumors judged to contain
few contaminating normal cells based on pathology. CGH on cultured
HNSCC cells with known amplicon size could be used to further
delineate the sensitivity of the technique.
We also found copy number increases
chromosomes but at lower frequencies.
changes occurred at a frequency of <40%. Other studies have found
higher frequency copy numberdecreases
chromosomesin cultured cells derived from head and neck tumors (4,
10, 11, 35). The reason for the discrepancy could lie in the difference
between primary tumors and cultured cells. However, because about
50% of HNSCC tumors have p53 mutations, one might expect that the
resulting genomic instability would generate a large number of alter
ations in at least this fraction of tumors. Thus, an alternative expla
nation is that the resulting alterations are relevant to tumorigenesis but
are below the limits of detection of CGH. Additional experiments are
required to resolve this issue. Our data show that the other chromo
somal copy number changes are qualitatively the same as described in
the literature. The high frequency of loss of chromosome 3p and gain
of chromosome 3q26-27indicates the likely involvement
loci in HNSCC tumor formation and/or tumor progression.
in all the tumor types could be determined
of the relevant cDNA(s).
by normal cells, although
and decreases
All of these copy number
on other
and increaseson other
of these
Fig. 3. Idiogram of the copy number changes on chromosome
of the chromosomesindicate copy number decreases; bars on the right side, copy number
increases. Numbers on lop of the bars refer to the tumors described in Table 1.
3. Bars on the left side
4 J. Gray, unpublishedobservation.
305S
American Association for Cancer Research Copyright © 1995
on July 14, 2011 cancerres.aacrjournals.orgDownloaded from