[CANCER RESEARCH57. 1344-1352. April 1. 19971
Homozygous Deletions of Human Chromosome 3p in Lung Tumors'
Sean Todd,2
Wilbur
A. Franklin,
Marileila
Vareila-Garcia,
Timothy
Kennedy,
Carl
E. Hilliker,
Jr., Lisa Hahner,
Marshall Anderson, Jonathan S. Wiest, Harry A. Drabkin, and Robert M. Gemmill@
Division of Oncology and the Colorado Cancer Center (S. T., C. E. H.. L H., H. A. D., R. M. G.J and Department of Pathology (W.F.J. University of Colorado Health Sciences
Center, Denver, Colorado 80262, and Lung Cancer institute ofColorado, Denver, Colorado 80218 (T. K.J
ABSTRACT
Cytogenetic and loss of heterozygosity (LOH) studies have demon
strated
that deletions
of chromosome
3p occur at a high frequency
In nfl
forms of lung cancer. To clarify the role of 3p In lung tumorlgenesis and
to more precisely identify targets for positional cloning efforts, we have
performed
3p deletion
analyses
(microsateffite
and fluorescence
in situ
hybridization) in a series of lung cancer cell lines and uncultured tumor
samples. Importantly, we identified homozygous deletions In four uncul
hired tumors and one cell line. Homozygous deletions were found in three
squamous
described
tumors within a region of 3p21 which had previously been
only In ceO lines, a 1—2-megabase homozygous deletion In a
small cell tumor at 3pl2, and a 3pl4.2 homozygous deletion In a non-small
cell lung carcinoma cell line. The detection of homozygous deletions
affecting these multiple regions in uncultured tumor cells substantiates the
belief (previously based on deletions found only in tumor cell lines) that
these sites contain important tumor suppressor genes. Along with previ
ously reported homozygous deletions In a distal portion of3p2l.3, we now
have evidence for four separate regions of 3p which undergo homozygous
deletions In either uncultured lung tumors or cell lines.
likely site for one or more tumor suppressor genes critical to lung
tumor development.
Analysis of 13 polymorphic loci permitted Hibi et a!. (18) to
propose
three separate
and apparently
independent
targets
for LOH in
lung tumors within bands 3p2S, 3p2l.3, and 3cen-pl4. Although Hibi
et a!. (18) did not find any 3p homozygous deletions, such changes
have proven highly useful for defining specific small targets likely to
harbor tumor suppressor genes in other tumor types such as melanoma
(19). Recurrent homozygous deletions have been reported in three
SCLC cell lines within 3p2l.3 (20—23)
whereas one SCLC line
[U2020 (24)] contains a deletion within 3pi2—i3. We have recently
shown that 3p2l.3 harbors two independent sites of homozygous
deletion and observed deletions in the distal region (3p2l.33) in
uncultured tumors of both SCLC and non-SCLC histology (25). In
addition, we and others have observed recurrent homozygous dde
tions in 3pl4.2 in carcinoma cell lines of diverse origin (26—28).
These sites coincide approximately with regions identified by Hibi et
al. (18),
although
this is likely
due to the large size
of the areas
identified in this early study. Similar sites have been shown to
INTRODUCTION
Lung cancer is the leading cause of cancer deaths in the United
States(1), witha typicallyverypoorprognosis.A diversityof genetic
alterations has been documented in lung tumors which affect both
proto-oncogenes
RBI
such as MYC (2) and tumor suppressor genes like
and P53
(3—5). Abnormalities
of specific
chromosomes
and
chromosomal regions are common, including deletions of 3p (6—i1),
Sq (12), 9p (13), l3q and i'7p (14, iS) as well as amplifications on 3q,
5p, 8q, and l7q (16). These
sites are presumed
to harbor
one or more
genes whose specific alteration by deletion, amplification, and/or
mutation constitute important steps on the pathway toward lung tumor
development. Although some of these target genes are known (e.g.,
RB] on l3q, P53 on Yip, and CDKN2 on 9p), the balance remain
unknown.
Deletions affecting the short arm of human chromosome 3 were
first identified in SCLCs4 by cytogenetic analysis (6, 11), confirmed
by LOH studies using polymorphic RFLP markers (10), and subse
quently
extended
to non-SCLCs
(7—9, 14, 17). Tumors
with the small
cell phenotype (pulmonary neuroendocrine cells with expression of
neuron-specific enolase, cortical granules, and neural cell adhesion
molecule) show very high rates of overall 3p LOH (approaching
100%) whereas somewhat lower rates (75%) have been observed in
non-SCLCs. Collectively, these studies strongly implicate 3p as a
Received3/25/96;accepted2/3/97.
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
18 U.S.C. Section 1734 solely to indicate this fact.
I This
investigation
was
supported
by
a grant
from
the
Specialized
Programs
of
Research Excellence of the National Cancer Institute (CA-95008).
2 Supported
by
Postdoctoral
Fellowship
Grant
PF-4l70
from
the
American
Cancer
Society.
3 To
whom
requests
for
reprints
should
be
addressed,
at Division
of
Medical
Oncology,
Bl7l, University of Colorado Health Sciences Center. 4200 East 9th Avenue, Denver, CO
80262. Phone: (303) 315-3556; Fax: (303) 315-8825; E-mail: [email protected].
4 The
abbreviations
used
are:
SCLC,
small
cell
lung
carcinoma;
LOH,
loss
of
het
erozygosity; FISH, fluorescence in situ hybridization; UBL, unbalanced loss; UBG,
unbalanced gain; YAC, yeast artificial chromosome; RER, replication error.
undergo LOH in carcinomas of the kidney (29—31),cervix (32, 33),
head and neck (34), and nasopharynx (35). Data supporting tumor
suppressor function have come from chromosome and DNA transfer
studies suggesting that both the 3p2l.3 region and a more proximal
site including 3pl2—l4 can suppress tumorigenicity of cancer cell
lines in athymic nude mice (36, 37). Although a number of candidate
genes have been identified within some deletions [FHJT (27); GNAI2
(38), semaphorin P1(25), SEMA5 (39), ITGA4L (40)], none have been
confirmed to act as tumor suppressors.
To clarify the role of 3p in lung tumorigenesis and to more
precisely identify targets for positional cloning efforts, we have
utilized a series of 32 highly polymorphic microsatellite loci
(41—44) to examine lung tumors for genetic deletions. Most of
these loci have known genetic and physical map positions (42, 45,
46) and known orders, resulting in relatively unambiguous inter
pretations. Small cell (SCLC) tumors included 33 cell lines and 26
matched tumor/normal SCLC biopsy samples. Two non-SCLC
tumors were also obtained from one of these SCLC patients.
Non-SCLC tumors included 12 cell lines and 23 matched tumor/
normal
biopsy
samples
with squamous
histology.
The results
con
firmed the high rates of 3p deletion in both SCLC and non-SCLC.
One homozygous deletion was detected in a non-SCLC cell line
within 3pl4.2, although this region was never found homozy
gously deleted in uncultured tumors. Importantly, we identified
one SCLC tumor with a homozygous deletion in 3pl2. Three
squamous
carcinomas
were
identified
with homozygous
deletions
in 3p2l.31, coincident with the semaphorin IV and GNAJ2 genes.
These results demonstrate that homozygous deletions occur in at
least a subset of uncultured tumors. Our results also show that
deletions within 3p2l.3l are not restricted to SCLC cell lines and
can occur in non-SCLC tumors with squamous histology. These
data indicate that 3p involvement in lung cancer is complex with
multiple deletion targets and the potential for independent, cumu
lative, or synergistic effects on lung tumor development.
1344
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 1997 American Association for Cancer Research.
CHROMOSOME3p DELETIONSIN LUNG TUMORS
MATERIALS
analyzedCaseSitePAt―Diagnosis'@UH92-5292SCLCvA93-3575SCLCUH92-04206L.
Table I SCLCtumorspecimens
AND METHODS
Polymorphic Loci and PCR Analysis. Primer sequences for the 32 loci
used were obtained from the Genome Database (The Johns Hopkins Univer
sity, Baltimore, MD) or were derived from our previous efforts (41, 44).
lobePSCLCUH92-05l05CerebellumMSCLCUH92-40l9R.
lower
Oligonucleotide primers were obtained from Research Genetics (Huntsville,
AL) and from the Oligonucleotide Synthesis Core Facility of the Colorado
lobePSCLCUH92-9236SCLCUH93-2038R.
upper
Cancer Center. One primer was radioactively
end-labeled
with [y-33P]ATP
using T4 polynucleotide kinase. PCR amplifications occurred in a 20 p1
reaction with 40 ng of template DNA for cell lines and in 10 pA volumes with
less than 1 ng of template for microdissected samples. Template DNA for
micmdissected samples could not be quantitated direcfly; however, aliquots
amplified
in each
reaction
were
estimated
to contain
between
nodeMSCLCVA91-0802Mediastin.
nodeMSCLCvA9l-o8s3BCervical
nodeMSCLCVA9I-2163-4BMarrowMSCLCVA91-21
l2B3Axillary
nodeMSCLCVA91-30361R.
lobePSCLCVA93-0590R.
95°C
for5 rainfollowed by touchdowncycles of94°Cfor 1 ruin,annealingfor
temperature
(>90%)S9l-lOl8SAlBrainMSCLCS92-OO8l5AlMediastin.
10—50 cell
equivalents based on comparisons with diluted control templates. PCR was
performed utilizing a combined hot start and touchdown procedure as follows:
1 mm, and 72°C for 1 mm. An initial annealing
lobePSCLCPSL93-8498PSCLC
upper
10°Cabove the
upper
upper lobePSCLC
lower lobePSCLC
(>90%)VA93-3OlOAL
and
carcinomaVA93-3175EL.
optimum for each primer set was ramped at —0.5°C
per cycle until the
appropriate annealing temperature was reached, at which time 10—15addi
tional cycles were performed without ramping. Aliquots of reaction products
were separated on 7% polyacrylamide/8 M urea/32% formamide denaturing
Sq
lower lobePSCLC
and
carcinomaVA93-3
Sq
lobePSCLCVA94-l355CL.
l93AL.
upper
lobePSCLCUH93-4000BrainMSCLCVA9I-3625-2PSCLCVA93-2l03L.
upper
gels and detected by autoradiography overnight. In some cases, biotinylated
primers
were substituted
for [y-33P]ATP-labeled
primers,
and PCR products
were detected using the New England BioLabs Phototope Detection kit.
Alleles were scored by visual inspection of band patterns and ambiguous
results were repeated. In some cases, multiplex analysis was used to provide
internal controls for reaction conditions. Mixtures of two primer sets were first
tested for optimum amplification parameters and for lack of interference.
Optimized
conditions
were then used for analysis
of selected
upper lobePSCLC
(>90%)VA93-2276MR.
lobePSCLCUH93-7779PSCLCVA94-l425ER.
median
median lobePSCLC
a p
primary
tumor;
M,
metastasis;
L,
left;
R,
right;
Sq.
squamous.
b Diagnosis included SCLC or mixed SCLC and squamous carcinoma.
samples.
Cell Line DNA. Thirty-threeSCLC and 12 non-SCLC cell lines were
obtained
from
the National
Cancer
Institute
through
the Colorado
a technique
Cancer
Center Tissue Culture Core. DNA was isolated by detergent lysis and protein
sac K digestion followed
by phenol extraction. The SCLC lines were H60,
H69, H82, Hl28, Hl46, Hl82, Hl87, Hl96, H209, H2l 1, H250, H345, H378,
HSlO (extrapulmonary small cell tumor), H524, H526, H592, H7l 1, H748,
H84l, H865, Hl045, Hl092, HllO5, H1284, Hl339, Hl450, Hl607, Hl622,
Hl876, H1994, H2107, and H2l71. The non-SCLC lines (which included
adenocarcinomas, large cells, mesotheliomas, carcinoids, and some unclassi
fled subtypes)were H28, Hl57, H226, H292, H460, H513, H630, H66l, H892,
H1264, Hl334, and H2l22.
Patient Specimens and Histological Evaluation. Tumor biopsy speci
mens were obtained either at bronchoscopy or thoracotomy from SCLC
bronchoscopy.
DNA at a 1:50 (probe:Cotl)
weight ratio for blocking
with 70% formamide
at 70°C for 2.5 mm to denature chromosomal
BDS image analysis system. The data were analyzed by comparison of
proportions
of nuclei
signal
rations from cultured normal cells.
for tumor cell type. SCLC
samples analyzed are listed in Table 1 and Table 2 lists the squamous
Microdissection. Invasive tumor cells were separated from normal cells by
microdissection. Sections were cut 10 @imthick, stained with H&E, rinsed
and covered
with glycerol.
Cells
of interest
were
scraped from slides using a hand-drawn glass micropipette and transferred to
a microcentrifuge tube containing water. Washed cell aggregates were then
treated with detergent, digested with proteinase K (37°C
overnight), and heated
to 98°Cfor 8 mm to inactivate any remaining proteinase K. Aliquots (1—5
p1)
patterns
using
the standard
normal
approximation for the underlying binomials. This was necessary for case
91-10185
with 90% ethanol,
DNA and
tected using FITC-conjugated avidin and biotinylated goat anti-avidin antibod
ies whereas digoxigenin-labeled probes were detected with sheep anti
were evaluated by two pathologists
briefly
sequences,
then hybridized overnight at 37°C.Extensive posthybridization washes re
moved nonspecific background probe. Probes labeled with biotin were de
histological evaluation, all tissue was fixed in 10% buffered formalin, embed
dad in paraffin,sectionedat 6 pm, and stainedwith H&E. Histologicalslides
independently
of repetitive
denatured, and then preannealed for 30 mm prior to use. Slides were treated
individual
For
carcinomas.
@
et al. (47). Purified nuclei were distributed
observed in a Zeiss Axioskop fluorescence microscope and captured using a
GrandJunction,CO) patients. Some bronchoscopicbiopsies were collected
diagnostic
from Hedley
digoxigenin antibodies conjugated with rhodamine. Fluorescence signals were
(University
of Colorado Health Sciences
Center Specialized
Programs of
Research Excellence,
SPORE) and squamous carcinoma (St. Mary's Hospital,
from patients with a chest mass undergoing
modified
onto slides by cytospin centrifugation. Touch preparations were obtained by
touching frozen tissue blocks against coated slides and fixing in 100% meth
anol at —20°C
for 20 mm.5
Probes were dissolved in hybridization solution along with human Cotl
in which cytospin
tumor nuclei were compared
Nuclei in touch preparations
to cytospin
prepa
were scored only when at least one bright
a-satellite signal was present. This ensured that only nuclei with good hybrid
ization efficiency were included in the analysis. Two cosmid signals were
counted as a single signal if they were in a closely paired (sister chromatid)
configuration. Nuclei with two red and two green signals were considered to
be normal. Those with one of each signal (one red/one green) suggested
balanced loss of the whole chromosome. Those with more than two of each (in
representing DNA from between 10 and 50 cells were used for PCR reactions.
equal proportions) suggested balanced gain of the whole chromosome (three
red/three green, four red/four green, etc.). All other nuclei showed partial
chromosome losses (UBLs) or gains (unbalanced gains) indicated by unequal
Matchednormalcell DNA was obtainedfrom eithermicrodissectedmorpho
numbers of red and green signals (two red/one green, three red/two green, etc.).
logically normal epithelial cells in the same thin section or lymphocytes.
FISH Analysis. Chromosome 3-specific FISH probes were prepared from
cosmids (c98AlO and c59H8; 3p2l.3), YAC (9l4Al 1; 3pl2) and the a-sat
ellite clone D3ZJ as a positive control locus. These DNAs were labeled with
Within this last category were nuclei which showed zero green signals indic
ative of homozygous deletions (one or more red signals/zero green). This
subcategory (one to four red/zero green) was separated out from the UBL
biotin or digoxigenin using nick translation.
Sample slides were preparedusing eitherpurifiednuclei or touch prepara
tions. For nuclei, frozen tumor sections were first microdissected
category to demonstrate the occurrence of homozygous deletions in the tumor
nuclei.
analyses
Since normal nuclei show a small proportion
of the results
were performed
to determine
of abnormal
signals,
significance.
to purify
tumor cells away from normal to as great an extent as possible. Nuclei were
isolated from several normal control cell lines and selected SCLC tumors using
5 M.
Varella-Garcia,
W.
A.
Franklin,
H.
A.
Drabkin,
and
R.
M.
Gemmill,
detection of 3p2l deletions in fresh lung tumors, manuscript in preparation.
1345
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 1997 American Association for Cancer Research.
FISH
CHROMOSOME 3p DELETIONS IN LUNG TUMORS
In tumor 91-3036, proximal short arm markers were retained but distal
markers were lost, suggesting a terminal deletion. Tumor 94-1425 had
only a single marker that showed retention flanked closely by markers
that had lost one allele. Since this case demonstrated a high degree of
microsatellite instability (see below), it was possible that D3S1296
Table 2 Squamous carcinoma specimens analyzed―
Case
Stage
22
23
24
26
27
lilA
II
underwent
30
37
39
43
44
lIlA
II
LIlA
LILA
tumors
were
60
61
62
ll
primary
by the generation
of a new
allele
matching
deletion
of the region in the tumor with the PCR
products arising from contaminating normal DNA as was found at
other loci in some tumors (see below). Alternatively, this case may
retaina smallfragmentof 3p,perhapsfusedwithanotherchromosome
lilA
Lv
63
64
65
66
‘I
All
followed
small homozygous
lilA
45
50
51
52
58
59
LOH
the original through failure of mismatch repair. If true, then 94-1425
would be categorized with the 21 cases that demonstrated loss of one
entire homologue. It is also possible that the retention may reflect a
and resulting
In tumor 91-10185, two adjacent loci (D3S1254 and D3S1776)
retained heterozygosity but were flanked by loci which underwent
loss. This pattern of retention flanked by loss was similar to that
reported by others in tumors that had undergone homozygous dele
tions (49). To test this possibility, both FISH and multiplex PCR
analyses were performed. Nuclei isolated from a frozen tumor block
were hybridized with a mixed probe consisting of a control cosmid
II
and
all
had
squamous
in diploidy for a very limited region surrounding
D3S1296.
histology.
RESULTS
Homozygous 3p Deletions in SCLCs. DNA samples prepared
from cell lines and tumors were assayed using 27 loci distributed
across chromosome 3p as shown in Fig. 1. The concentration of
markers within the proximal portion of the short arm was specifically
chosen to assess this region for genetic change since other investiga
tors had focused on 3p2l (8, 10, 48).
Our initial studies examined 33 SCLC cell lines for heterozygosity,
homozygosity, or homozygous deletions. A typical autoradiogram is
presented in Fig. 14 showing alleles for the locus D3S1480 (3pl4.2)
in 24 lines. Fig. 3A summarizes the complete results for 27 markers in
all of the SCLC lines. Although matching normal controls did not
exist for most of these lines, the preponderance of homozygous loci
strongly suggested that most of one 3p homologue had been deleted in
the majority of lines. Other lines retained heterozygosity for larger
numbers of markers and this precluded interpretation regarding pos
sible allele loss. None of the markers used detected homozygous
deletions, although line H740 (not tested in this study) has been
previously shown to be deleted within 3p2l.3 (21).
The absence of matched normal DNA samples precluded unambig
uous identification of deletion targets in the cell lines. To overcome
these limitations, a series of 26 SCLC tumors with matching normal
DNA samples were studied. Microdissection of both tumor and nor
mal tissues from paraffin-embedded diagnostic biopsies was per
formed to reduce the level of contaminating normal cells in each
tumor sample. Matched pairs of normal and tumor DNA were ana
lyzed with 26 of the 27 loci mentioned above (D3S1251 was omitted),
and an example of this analysis for 8 cases is shown in Fig. 2B
(D3S1284).
All of the informative
cases
(five
of eight)
cM
0
40
Deletion
3p2l.3l
76
82
83
3pl4.2Kidney
Tumor
Breakpoint &
Deletions
95
100
102
U2020 Deletion
underwent
LOH, althoughthe normalcontrolsamplefor case 91-2163failed to
amplify
66
Deletion
104
105
112
113
in this experiment.
Complete data on tumor-normal pairs is presented schematically in
Fig. 3B.In 85% of the cases (22/26), every informative marker under
went LOH (summarized in the last column of Fig. 3B), a result which
Fig. I. Microsatellite loci utilized. Thirty-two highly polymorphic microsatellite mark
era (center) are shown in map order along chromosome 3p with genetic positions (cM)
is in close agreement
indicated on the right and physical location on the left. Three markers mapped within 3pl2
with those obtained from cell lines. Four tumors
(91-10185,
93-2103,91-3036,and 94-1425)retained
heterozygositynearthecentromere,7 derivedfromwithinthe U2020 deletionregion(3pl2—pl3;Ref. 61)
while an additional 10 covered the 3pl3 and proximal 3pl4 regions (45). Several markers
at one or a few loci. Tumor 93-2103 lost the proximal portion of 3p
were targeted to presumed critical sites including the familial kidney cancer (3;8) break
but retained heterozygosity for one distal marker. This result could be point region within 3pl42 (44, 54) and two regions within 3p2l.3 which undergo
independent homozygous deletions (20—23).The five markers indicated with an asterisk
explained by an interstitial deletion, although the extent of distal were utilized on the squamous tumors only. Four homozygous deletion regions (solid
retention could not be well defined because of uninformative markers.
bars) and the familial kidney tumor breakpoint at 3pl4.2 are indicated on the left.
1346
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 1997 American Association for Cancer Research.
CHROMOSOME3p DELETIONSIN LUNG TUMORS
tumor
thatstillwas heterogeneous and contained a subpopulation of cells
Inc@J(n@
ofan retained one copy of the target region. Multiplex PCR analysis
A.
additional sample from this case confirmed the retention of het
erozygosity but failed to show differential amplification between
tumor and normal samples. Since the three analyses described were
performed (of necessity) on different samples of the tumor, the dis
crepancies
heterogeneity.One
are likely due to tumor
lungtumors
SCLC patient (case VA94-l355) had two additional
with squamous histology. Genetic analysis revealed that al
though all three tumors lost all informative 3p markers, the patterns
ofretained
Loc@
D3S1480
(3p14.2)
D3S1286was
allelesweredifferent(Fig. 2C).The large alleleof
@
@
@
@
@
@
in@IAOA
C.
B.
O@
@@@J%Jeach
;@ ;;
;;
5;
;;
j
Z,-
zi-
z@-
z,-
z@-
@
mous
b3
z@
Aø
geneticI:1retained
@
I
B
pevents
failedFig.
Locus
Locus
D,@I@Sll4
D3S1286
D3S1776
(3p13)
(3p25)
tumor
2 retained
the large allele
while
squamous
tumor
the smaller allele. Although the number of informative mark
ers was low in this case, their distribution suggested that the majority
of the short arm was affected by LOH. These results suggest that
multiple primary tumors can develop through independent
affecting the same chromosomes, supporting observations
made by other investigators (50).
Homozygous 3p Deletions in Non-SCLCs. Analysis of 12 non
cell lines (summarized in Fig. 6A) showed multiple sites of
S
;;
ZI-
zi-
V r@@@v-
I9
llll
ll@
“ ii;ii;
retained in the SCLC tumor while the small allele was retained
squamous tumor. However, for D3S1776, the SCLC and squa
homozygosity,
heterozygosity,
and a few loci which failed to amplify.
(3p12)SCLC Although LOH could not be ascertained in the lines, loci which
amplify
2. Microsatellite analysis of tumors. A, SCLC cell lines. Primers for locus
one of which was end labeled with [@y-33PJATP,were used to amplify productstO
24 SCLC cell lines and a normal human control (FS). Samples of products were
were
investigated
further
by multiplex
PCR
for possible
befrom
D3S1480,
homozygous deletion. One cell line (H892) was confirmed to
homozygously deleted for D3S1300 by this analysis. This marker was
.
.
.
.
separated on a denaturing polyacrylamide gel, dried, and autoradiographed. Most lines
not amplifiable ln a multlplex reaction whereas the closely llnked
were homozygous for D3S1480 while three (H2ll, H250, and H510) retained heterozy
absentsected
marker D3S1480 was (data not shown); the remainder of
gosity.
B, microdissected SCLC tumors (uncultured). Eight matched pairs of microdis
normal(N) andtumor(7)DNA sampleswere analyzedwithD3S1284 (3pl3). Five
was shown to be the result of initial failed PCRs. The
of the eight cases were informativewith this markerandall of these underwentdeletional
.
.
.
.
.
ofthree
deletion
in
H892 occurs within 3pl4.2, colncldent with a region
loss. One case (91—2163)failed to amplify from the normal sample. C, case VA94-l355;products
tumorsin onepatient.LOHresultsareshownfortwoloci(D3S1286andD3S1776)
deletion identified in multiple tumor cell lines (28) and
mappingatoppositeendsof3p ina patientwithoneSCLCandtwosquamouscarcinomas the location of the recently described FHIT gene (27).
(5CC). The A allele of D3S1286 was retained in the SCLC whereas the B allele was
andretained
We analyzed 23 microdissected squamous cell carcinomas
retainedin bothsquamouscellcarcinomas.For locusD3S1776(3pl2), theA allelewashomozygous
in the SCLC and 5CC 2, whereas the B allele was retained in 5CC 1. All other
normal DNAs with 24 loci to ascertain genetic change
informative loci in this case (data not shown) underwent LOH, suggesting a complete 3p
.
.
.
.
in this histological
deletion in all three tumors.corresponding
subtype
of lung cancer.
.
This analysis
took advan
tage of several new polymorphic loci which were mapped within or
cell(c98
adjacent to regions known to be homozygously deleted in SCLC
mostmaps
(Fig. 1, asterisk). Amplifications using multiplex PCR (for
from 3p2l, red signal) and YAC 9l4All (green signal) whichlines
provided internal controls and maximized information re
within 3pl2. This YAC is nonchimeric and spans a 930 kbmarkers)
analysisD3S1776
from the limited tumor materials. The results of this
region within the 1—2-megabase interval bounded by D3S1254 andcovery
carcinomasof
summarized in Fig. 6B. Three subsets of squamous
(45). Three tumor sample nuclei displaying the major typesare
informativeently
could be defined: (a) those with complete loss of all
hybridization signals obtained are shown in Fig. 4. Some appar
allhemizygously
normalnuclei revealedtwo copiesof each sequence(Fig. 4A),markers (occurring in 52%); (b) those with complete retention of
retention(Fig. deleted nuclei yielded two red and one green signalmarkers
(13%); and (c) those with loss of most markers but
cellenfly
for a few (34%). These results were different from all non-SCLC
4B) or one red and one green (data not shown), whereas appar
high-stage(Fig.
analyzed in this study except H892. Both low- and
homozygously deleted nuclei yielded two red and no greenlines
subsets.Data
were found in each of these three
4C).tumors
withlymphoblastoid
possible mechanisms could give rise to the eight cases
from 300 tumor and 290 control nuclei (isolated from aSeveral
retention. These regions might reflect actual retention, per
cell line) are presented graphically in Fig. 5A. Asporadic
whenof
striking
difference
between
control
and tumor was the high percentagehaps
resulting from integration of small chromosome fragments
contaminatingone
majority of the homologue was lost. Alternatively,
tumor nuclei with zero signals for the 3pl2 YAC 914A1 1 (Fig. 5A,the
Aa-satellite
DNA could be masking the true state of the genetic region.
to four red/zero green, arrow). The presence of the controlnormal
thedetectionprobe (red signals) indicated that both the hybridization andlocus
undergoing LOH could be interpreted as heterozygous if
reactions were successful. The differences observed werecontaminating
normal DNA contributed an efficiently amplified al
misinterpreted,supporting
A homozygously deleted locus could easily be
highly
significant statistically (z = 6.36, two-tailed P << 0.001),lele.
forgous
since contaminating normal DNA would provide the only template
the hypothesis that this tumor had undergone a homozy
FISHwithin
(49). To test these possibilities, we performed
deletion event within 3pl2. Since the loci involved are containedamplification
retainthat
for the 3p2l.3l region in three tumors that appeared to
the U2020 deletion, this case provided important confirmationanalysis
Twoto
for marker D3S2968 (tumors 23, 26, and 27).
proximal 3p can be a deletional target in tumors and is not uniqueheterozygosity
cosmids derived from 3p2l.3l (c98 and c59) were la
the U2020 cell line. However, the tumor sample also showed aoverlapping
withcompared
with biotin and the control locus (D3ZI) was labeled
higher
proportion ofnuclei with fewer green signals versus red signalsbeled
and the mixture was hybridized to tumor touch prepara
with the normal (Fig. 5A, asterisk), suggesting that thedigoxigenin,
1347
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 1997 American Association for Cancer Research.
CHROMOSOME 3p DELETIONS IN LUNG TUMORS
A. SCLC Cell Lines
B. SCLC Tumors
IndividualLinesI____I
3pter
22
8 14
Lines
IIII4r#‘.4,J'@p@rr@t#Al
D3S1307
D3S1286
0331307
D3S1286
I D3SI100
@
@
p
@p
0351235
10351300
1D3S1450
D3S1233
D3S1217
D3S1261
D3S1296
0331217
3 13
0331566
p
0331598
0331284
I,4#@1 l,4I@
D3S1274
AFMOO1za2
0331274
D331604
0331604
D3S1577
AFM191vd8
D3S1776
0331276
D3S1 101
0331251
CEN
0331261
0331296
D3S1566
0331210
D331562
AFM32OYb5
AFMOOlza2
D3S1254
0331100
0331235
0331233
D3S1210
0331562
AFM@
D3S1598
D3S1284
M32020
I
10331300
t(3;8) j 0331481
I 0331480
t(3;8) ID3sl4al
3p13
Cases
3pter
à U2O2O D3S1577
AFMI91vd8
0331254
ri―@f
0331776
D3S1276
0331101
0331251
L.1_iiu—
HHHH128821461092524
HH
CEN
Cell145021071994
Line#211250510
@
91
VA93 VA91 VA94
10185 2103 3036 1425
Case #
—=LOSS
OF1ALLELE
(LOH)
—= HOMOZGOUS
EJ= HETEROZYGOUS @=
NOTDONE
@
= UNINFORMATiVE
H = HOMOZYGOUS
Fig. 3. Chromosome
= NOT
Complete
LOH
=HETEROZYGOUS
DONE
DELETION (by FISH)
3p deletions in SCLC. A, allele analysis in cell lines. Results from microsatellite analysis are represented for each locus by a box with a filled pattern; I,
homozygous; 0, heterozygous; @,
not done, as indicated. Each vertical stack of boxes symbolizes the overall genetic status for the chromosome arm in individual cases, except for
the last two columnswhich summarizedatafor 8 and 14 lines, respectively.The markersutilized are shown on the left and correspondto those in Fig. 1. B, LOH in microdissected
SCLCtumors.Twenty-sixtumorcasesare presentedusingthe samesymbolismas above,withchangesand additionsas indicated.Twenty-twocasesunderwentLOHfor every
informative marker and are summarized in the last vertical column. H indicates confirmed homozygous deletions.
tions (see “Materialsand Methods―). Between 100 and 300 nuclei
were scored from each tumor, and the resulting numbers were tested
for statistical significance as described above.
The results of these experiments are presented graphically in Fig. 5,
B (case 26) and C (cases 23 and 27). A blind comparison of case 26
with a control sample and three other squamous carcinomas (cases 22,
24, and 37) demonstrated that nearly 25% of the nuclei from this
tumor failed to hybridize with the 3p2l.3 probe (Fig. SB, arrow),
confirming the presence of a homozygous deletion. The FISH data
also revealed a significant proportion of nuclei (13%) that had one red
and one green signal, suggesting the loss of one chromosome 3
homologue in a portion of cells. This latter result indicated that the
tumor was heterogeneous with at least two genetically altered sub
populations. In the second experiment, case 23 had nearly 25% of
nuclei with zero green signals, indicative of a homozygous deletion.
Case 27 had mostlynormalnucleiwith only statisticallyinsignificant
numbers of nuclei showing abnormal patterns. The FISH data thus
confirmed
the presence
of homozygous
deletions
in two of the three
cases that retained heterozygosity for D3S2968. Although the results
for control tumors were generally consistent with the PCR analysis,
FISH data revealed previously hidden details. Case 22 showed a
significant fraction of nuclei with partial 3p deletion (UBL) consistent
with LOH. Case 24 showed no 3p changes by either FISH or PCR
analysis. Case 37 had nearly 30% of nuclei with partial 3p loss (UBL)
consistent with the previously detected 3p LOH. In addition, slightly
more than 10% of the nuclei had no 3p signal, suggesting that a
subclone
of this tumor was also homozygously
deleted
in 3p2l.3.
We
Microsatellite Instability. Our microsatellite analysis permitted
the detection of cases with aberrant bands, suggesting that possible
DNA mismatch repair defects could be present (51). Although none of
the cell lines showed any evidence of altered bands or multiple alleles,
this analysis could not be definitive due to the lack of normal controls
for comparisons. However, three cases of microdissected SCLC tu
mors (11%) showed evidence of multiple or changed alleles. In case
VA94-l425, 7 of 25 (28%) loci tested showed altered alleles in the
tumor DNA. Other cases with apparent instability had 19% and 6% of
loci showing variable bands. In all three cases, the instability pattern
was reproducible; in several other cases, altered bands apparent mi
tially were not reproducible. No examples of instability were detected
in the squamous
carcinomas.
DISCUSSION
The results reported here confirm the high rate of genetic
losses on
chromosome 3p in lung cancer and, importantly, demonstrate the
existence of homozygous deletions in uncultured tumors. The use of
32 highly polymorphic markers has allowed us to detect alterations
which have not been observed in previous studies. Extensive genetic
loss covering
the whole
short arm was present in many SCLC cell
lines and 85% of uncultured tumors. LOH analysis in case VA913036 suggested that the proximal portion of 3p, covering 3p12 and
including much of the U2020 deletion, was not an obligatory region
for genetic loss in SCLC. However, due to the absence of polymor
phic markers at the distal boundary of the U2020 deletion, it was not
note that several additional sites of retention, which might reflect
possible
to determine
whether the deletion
found in VA9I—3036
homozygous deletion, are present in the squamous tumors (Fig. 6B).
overlaps with that of U2020. Although the totality of SCLC data
This possibility needs to be explored further.
supports 3pl2 (as well as the rest of 3p) as a consistent region of loss,
1348
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 1997 American Association for Cancer Research.
@
@w
CHROMOSOME 3p DELETIONS IN LUNG TUMORS
A
A.
60
Case 91-1 01 85 Homozygous Deletion
@_
5,
Li
0 Control;
N = 290
U Case91-10185;
N = 300
z
w
0
C
5,
U
II
5,
0.
@
@
0
I.r@
.
2R/2G
-
..
@—.
I
1Rh GUBL1-4R/OGOR/i
UBGFISH
-4G
BG
SignalsCase26HomozygousDeletion
B.
B
90
a Control;N —204
@
@Case22;N —255
Case 24; N—ill
a,
U
. Case
26; N —299
z
—;-——;@@:1
@*@@__L@
0
@Case37;
N —170
C
a,
U
a,
0.
2R/2G
1Rh G
UBL
l-4R/OG
BG
UBG
FISHSignals
C.
Case 23 Homozygous Deletion
80
U Control; N —
73
C
aCase23;
N—151
r2Case
27; N—
232
a,
z
0
4'
C
a)
U
—
I
5,
0.
@
fA
@
—F71
UBL
__v@
l-4R/OG ‘ BG
—
UBG
FISHSignals
Fig. 5. FISH analysis of deletions; percentages of normal and deleted nuclei in
seven tumors. Results of FISH analysis for 3p probes are presented as percentages of
nucleiwiththe indicatedpatternsof hybridization.The numberof nucleianalyzedin
each case is indicated. In A, the testing probe was the 3pi2 YAC 9l4All (green) and
the control probe was cosmid c98 (red) which maps within 3p2l. In B and C, the
testingprobewasa mixtureof c98Al0 andc59H8(green),bothof whichmapwithin
the 3p2l .31 homozygous deletion region, whereas the control probe was the a-cen
Fig. 4. FISH analysis on isolated nuclei from SCLC case 91-10185. Cosmid 98 (3p2i)
andYAC 9i4Al I (3pl2)were labeledwith digoxigeninandbiotin,respectively,andused
to analyze for loss or retention of their corresponding regions in whole nuclei. Nuclei were
tromeric sequence D3ZJ (red). The data were divided according to categories of signal
patterns, as indicated; normal nuclei (two red/two green), loss of entire homologue
isolated from a frozen tumor sample and prepared for analysis as described in “Materials
and Methods.―Probe c98 was detected with rhodamine and appears red while the 9l4Al 1
YAC was detected with FITC and appears green. The three panels present three typical
nuclei derived from the tumor sample showing two red and two green signals (A, normal),
(one red/one green), UBL of chromosome 3 (two red/one green, three red/two green,
three red/one green, etc.), nuclei with a homozygous deletion (one to four red/zero
two red and one green (B, hemizygousdeletion),and two red with zero green (C,
green), balanced gain (BG) of chromosome 3 (three red/three green. four red/four
green, etc.), and UBO of chromosome 3 (one red/two green, two red/three green, etc.).
Arrows, cases with significant levels of homozygous deletion; asterisks, cases with a
homozygous deletion).
high frequency of hemizygous deletions.
1349
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 1997 American Association for Cancer Research.
CHROMOSOME 3p DELETIONS IN LUNG TUMORS
B.SquamousTumors
A. Non-SCLC Lines
Case#22
3pter
D3S1 307
D3S1286
@3p21
I D3S1100
I
21
p .
@rox,
D3S1296
D3S1566
D3S1 210
D3S1562
0331233
3p13
D3S1217
D3S1296
D3S1566
D3S1210
AFM320@
D3S1604
p1 —13 AFM191vd8
2 20 D3S1577
D3S1598
D3S1284
12—13I AFMOO1Za
D3S1254
D3S1776
D3S1276
u@o@ol
D3S1 101
D3S1251
D351276
No Complete
LOH
LOH
CEN
H H
H H
H H
H
H H H
H
H
28 157 226 292 460 513 630 661 892 1264 1334 2122
@
I D3S1300
‘, , I D3S1480
D3S1598
D3S1 284
AFMOO1
za2
D3S1274
@
D3S2968
D3S1573
D3S1235
p3.8 •
D3S1481
AFM32Oyb5
@
12
D3S1 100
D3S1217
D3S1261
3p13
3
Cases Cases
D3S1611
p21.3 I D3S1298
(dlst) I D3S1260
1D3S1480
0331233
@
@
@
65
D3S1307
D3S1286
D3S1235
14.2 1D3S1300
t(38) 1D3S1481
,
23 26 27 37 44 45
—=HOMOZYGOUS
=
@
@
= NOT DONE
H
—=LOSS
OF1ALLELE
(LOH)
HOMOZYGOUSDELETION(by PCR)
- HETEROZYGOUS
F@!@
= UNINFORMAATIVE =NOTDETERMINED
H = HOMOZYGOUSDELETION(by FISH)
Fig. 6. Chromosome 3p deletions in non-SCLC. A, cell line summary. Allele analysis in 12 non-SCLC cell lines is shown with symbolism identical to that in Fig. 3, except for noted
changes. A homozygous deletion (If) for D3S1300 was identified in one line, H892. B, allele analysis in microdissected non-SCLC tumors. LOH was performed as described in the
text and is summarized as in Fig. 3. Twelve cases with complete LOH are summarized in the right-hand column while three cases with no LOH are adjacent. The results for the
remaining eight cases are shown individually. Confirmed homozygous deletions are indicated (H); asterisks, particular results which were difficult to determine as undergoing either
loss or retention.
it is also possible
that loss for any 3p site will not be observed
in 100%
of cases and that VA9I -3036 may represent such a case for this
proximal region.
Case 91-10185 provided important confirming evidence that prox
imal 3p deletions occur in uncultured tumors. The loci involved map
within the proximal portion of the 8-megabase U2020 deletion region
(24, 52) and significantly
narrow
the deletion
target.
Initial
PCR,
FISH, and final multiplex PCR data were generated from separate
samples of a brain metastasis from this case. Although the standard
PCR and FISH data supported the presence of a homozygous deletion,
the multiplex PCR data did not. These discrepancies suggested that
the tumor was heterogeneous for the deletion, a conclusion supported
by the FISH data which showed a significant number of nuclei with a
hemizygous deletion for the 3pl2 target probe YAC 9l4Al 1 (Fig.
5A). Heterogeneity implies that the deletion likely occurred subse
quent to the metastatic event and therefore probably was a late event.
However, deletions of the CDKN2 gene on chromosome 9p are known
to be heterogeneous in tumors (53) and yet this loss has a critical role
in abrogating cell cycle control. The low frequency of homozygous
deletions for CDKN2 has been explained by the observation that the
remaining CDKN2 allele is often silenced by hypermethylation (54).
It is also possible the deletion is a consequence of overall genomic
instability rather than a selected event. Case VA91—3036, which
retained heterozygosity for 3pl2, suggested that deletion of 3pl2 may
only occur in a portion of tumors.
Non-SCLC cell lines did not show the extensive loss characteristic
of the majority of lung tumors studied here. A site within 3pl4.2
homology (approximately 50% predicted amino acid identity) to the
yeast enzyme diadenosine tetraphosphate asymmetrical hydrolase
(Aphlp). Abnormal FHIT message was observed using reverse tran
scription-PCR in selected tumor cell lines and uncultured tumors,
leading these investigators to suggest FHff as a candidate tumor
suppressor gene. Sozzi et a!. (57) reported reverse transcription-PCR
analysis in SCLCs, also suggesting FHIT involvement. Three markers
used in our analysis are localized within the FHIT gene (D3S1300,
D3S1481, and D3S1480), and all three show moderate to strong
suppression of heterozygosity and D3S1300 was homozygously de
leted in H892. This region at least is thus implicated in non-SCLC.
However, involvement of the FlIT gene has not been determined. We
note that Thiagalingam et a!. (58) failed to detect FHIT involvement
in colorectal carcinomas. Our own studies have suggested that FilET
expression may simply be affected by high levels of genomic insta
bility in the region (28) due to coincidence with the FRA3B fragile site
and not necessarily a “target―
of loss for cancer development. Con
firmation
of FHJT as a tumor
suppressor
gene awaits
in vivo studies
demonstrating an effect of this gene on growth, differentiation, and/or
cellular
proliferation.
Analysis of one SCLC and two squamous tumors from the same
patient showed that all three tumors retained different combina
tions of alleles and thus arose independently. Moreover, the dif
ferent allele combinations retained imply that at least one of the
tumors must have undergone two genetic changes affecting 3p;
either mitotic recombination or two independent deletions affect
ing the different homologues. The data support the field cancer
underwent homozygous deletion in one non-SCLC cell line (H892; ization hypothesis by suggesting that in smokers, the epithelial
Fig. 6A). This region includes the t(3;8) translocation breakpoint
lining of the lung contains many abnormal cells which are mdc
associated with hereditary kidney cancer (55), the 3pl4.2 common
pendently capable of developing into tumors. The data also imply
fragile site (FRA3B; Ref. 56), and has been shown to undergo ho
that loss of one 3p homologue is a key event and must have
mozygous deletion in other tumor cell lines (26, 28). Recently, Ohta
important biological consequences on its own.
et a!. (27) identified a gene (FlIT) within this region which has strong
Our studies on squamous carcinomas showed that 52% lost one
1350
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 1997 American Association for Cancer Research.
CHROMOSOME 3p DELETIONS IN LUNG TUMORS
entire 3p homologue,
13% had no detectable loss, and 35% had
partial loss. Since some loci were difficult to score in a few tumors
(Fig. 6B, asterisks), the true number of cases with complete loss
may be higher. We tested six cases for homozygous deletion by
FISH analysis. Each case retained some heterozygous markers,
including three which retained marker D3S2698 within the 3p2l.3
region. Homozygous deletion was confirmed in a subpopulation of
cells from cases 23, 26, and 37. Case 27 was the only one to retain
heterozygosity for D3S2968 and yet not have a homozygous de
letion. However, this case also showed no loss for distal markers.
These are the first homozygous deletions reported which affect the
3p2l.3 region in uncultured non-SCLC tumor cells. This loss
substantiates the hypothesis that a gene(s) in this region is involved
secondarily
to an initial
3p LOH,
since such partially
deleted
nuclei were observed in each case using FISH. Presumably, the first
event results in unregulated proliferation which becomes more pro
nounced upon loss of the remaining allele. Other tumors which
retained heterozygosity for one or a few markers need to be analyzed
using FISH to determine whether they also represent homozygous
deletions.
defective in DNA mismatch repair (51). Our analysis of45 lung tumor
cell lines and 23 squamous carcinomas provided no support for the
presence of the RER phenotype. However, we did observe 3 of 26
SCLC
tumors
which
showed
reproducible
evidence
for RER.
LOH in lung tumors and showed that homozygous deletions can
three sites within
3pl2,
p14.2,
and p21.31.
Two ofthese
sites
were found homozygously deleted in uncultured tumors whereas
the 3pl4.2 site was observed only in cell lines. As mentioned
previously, there are several additional sites of retention in the
squamous tumors that may reflect homozygous deletions. Included
in this group is a marker within 3pl4.2 (D3S1481) from case 23.
These sites need to be tested using FISH to determine whether they
too represent homozygous deletions in uncultured lung tumors.
Additionally, larger series of microdissected lung tumors with
specific histological features need to be investigated with an cx
panded marker set, including putative target loci on other chromo
somes such as 9p, to determine the frequencies and patterns of
genetic changes associated with each. It is also important to
perform
a similar
detailed
analysis
of deletions
Synthe
and Drs. John Minna, Adi Gazdar, and Bruce Johnson for providing the SCLC
and non-SCLC cell lines.
REFERENCES
1. Ginsberg, R. J., Kris, M. G., and Armstrong, J. 0. Non-small cell lung cancer. In: V.
T. DeVita, Jr., S. Heliman, and S. A. Rosenberg (eds.). Cancer: Principles & Practice
of Oncology, 4th ed., Vol. 1, pp. 673—723,
Philadelphia: J. B. Lippincott, 1993.
2. Ibson, J. M., Waters, J. J., Twentyman, P. R., Bleehen, N. M., and Rabbitts, P. H.
Oncogene amplification and chromosomal abnormalities in small cell lung cancer. J.
3. Mon. N., Yokota, J., Akiyama, T., Sameshima,
in preneoplastic
Y., Okamoto, A., Mizoguchi, H.,
Toyoshima,K.,Sugimura,T., andTerada,M. Variablemutationsof the RBgenein
small-cell lung carcinoma. Oncogene, 5: 1713—1717,
1990.
4. Hensel, C. H.. Hsieh, C. L, Gazdar, A. F., Johnson. B. E., Sakaguchi, A. Y.. Naylor,
S. L., Lee, W. H., and Lee, E. Y. Altered structure and expression of the human
retinoblastoma susceptibility gene in small cell lung cancer. Cancer Res., 50: 3067—
3072, 1990.
5. Miller. C. W., Simon, K., Aslo, A., Kok, K., Yokota, J., Buys, C. H., Terada, M., and
Koeffler, H. P. p53 mutations in human lung tumors. Cancer Rca., 52: 1695—1698,
1992.
6. Whang-Peng,J., Bum, P. A., Jr..KSO-Shan,C. S., Lee, E. C.. Carney,D. N., Gazdar,
A., and Minna, J. D. A nonrandomchromosomalabnormality.del 3p(l4—23),in
human small cell lung cancer(SCLC). Cancer Genet. Cytogenet., 6: 119—134,1982.
7. Brauch, H., Tory, K., Kotler, F., Gazdar, A. F., Penengill,
0. S., Johnson, B.,
Graziano, S., Winton, T., Buys, C. H., Sorenson. G. D., Poiesz, B. J., Minna. 3. D.,
and Zbar.B. Molecularmappingof deletion sites in the shortarmof chromosome3
in human lung cancer. Genes Chromosomes & Cancer, I: 240—246,1990.
8. Kok, K., Osinga, J., Carritt, B., Davis, M. B., van der Hota, A. H., van der Veen,
A. Y., Landsvater,R. M., de Leij, L. F., Berendsen,H. H.. Postmus,P. E., Poppema.
S., andBuys,C. H. C. M. Deletionof a DNAsequenceat the chromosomalregion
3p2l in all major types of lung cancer. Nature (Lond.), 330: 578—581.1987.
9. Brauch, H., Johnson, B., Hovis, J., Yano, T., Gazdar, A., Pettengill, 0. S., Graziano,
S., Sorenson,0. D., Poiesz, B. J., Minna,J., Linehan,M.. and Zbar, B. Molecular
analysis of the short arm of chromosome 3 in small- cell and non-small-cell carci
noma ofthe lung. N. EngI.J. Med.. 317: 1109—1113,
1987.
10. Naylor, S. L., Johnson, B. E., Minna, J. D., and Sakaguchi, A. Y. Loss of heterozy
3p markers in small-cell lung cancer. Nature (Land.), 329:
11. Whang-Peng, J., Kao-Shan, C. S., Lee, E. C., Bunn, P. A., Carney, D. N., Gazdar,
A. F., andMinna,J. D. Specific chromosomedefect associatedwithhumansmall-cell
lung cancer; deletion 3p(l4—23).Science (Washington DC), 215: 181—182,
1982.
12. Miura, I., Graziano, S. L., Cheng, J. Q., Doyle, L. A., and Tests, J. R. Chromosome
alterations in human small cell lung cancer: frequent involvement of 5q. Cancer Res.,
52:
Several other specimens showed altered bands upon initial analyses
but these failed to be reproducible. However, very dilute templates
could yield variable results, as shown recently in hereditary nonpol
yposis colorectal cancer patients (60). The apparent 10% rate of RER
in our SCLC specimens is thus a minimum rate and could be greater.
This investigation confirmed the presence of widespread 3p
affect
Tissue Culture, Oligonucleotide
sis, and Biostatistics Cores. We greatly appreciate the National Cancer Institute
gosity of chromosome
451—454,
1987.
Recent reports have suggested that some lung tumors may be
(11.5%)
The Colorado Cancer Center provided support through several shared
facilities including the Cytogenetics,
Cell.Biochem.,33: 267—288,
1987.
in tumor formation (21 , 36) and proves that homozygous deletions
are not an artifact of culturing tumor cells. Finding these homozy
gous deletions in non-SCLC cells broadens the spectrum of lung
tumor types for which this loss is a critical event, since previous
reports have all been of SCLC cell lines (21, 22, 59). That markers
in this region undergo frequent LOH in a variety of other human
carcinomas supports the general importance of the presumed tumor
suppressor gene in this region.
Four of the six tumors studied by FISH were found to be highly
heterogeneous, usually with two genetically abnormal subpopulations
in addition to the diploid nuclei (two red/two green) which were
presumed to be normal. Clearly, the homozygous deletions had to
occur
ACKNOWLEDGMENTS
1322—1328, 1992.
13. Bepler, G.. and Garcia-Blanco, M. A. l'hree tumor-suppressor regions on chromo
some lIp identified by high-resolution deletion mapping in human non-small-cell
lung cancer. Proc. Nail. Acad. Sci. USA, 91: 5513—5517.
1994.
14. Mori, N., Yokota, J., Oshimura, M., Cavenee, W. K., Mizoguchi, H., Noguchi, M.,
Shimosato.Y.,Sugimura,T..andTerada,M.ConcordantdeletionsofchromosomeIp
and loss of heterozygosity
for chromosomes
13 and 17 in small cell lung carcinoma.
Cancer Res., 49: 5130—5135,1989.
15. Yokota, J., Wads, M., Shimosato, Y.. Terada,M., and Sugimura,T. Loss of bet
erozygosity on chromosomes 3, 13, and 17 in small-cell carcinoma and on chromo
some 3 in adenocarcinoma
1987.
of the lung. Proc. Natl. Aced. Sci. USA, 84: 9252—9256,
16. Ried, T., Petersen,I., Holtgreve-Grez,H., Speicher,M. R., Schrock.E., du Manoir,
S., andCremer,T. Mappingof multipleDNA gains and losses in primarysmall cell
lungcarcinomasby comparativegenomichybridization.CancerRes., 54: 1801—1806,
1994.
17. Yokota, J., Mori, N., Akiyama, T., Shimosato, Y., Sugimura, T., and Terada, M.
Multiple genetic alterations in small-cell lung carcinoma. mt. Symp. Princess Taka
matsu Cancer Res. Fund, 20: 43—48,1989.
18. Hibi. K.. Takahashi. T., Yamakawa, K., Ueda, R., Sekido, Y., Ariyoshi. Y.. Suyama,
M., Takagi, H., and Nakamura, Y. Three distinct regions involved in 3p deletion in
human lung cancer. Oncogene, 7: 445—449,1992.
19. Kamb, A., Gruis, N. A., Weaver-Feldhaus, J.. Liu, Q., Harshman, K.. Tavtigian, S. V.,
Stockert, E., Day, R. S., Johnson, B. E., and Skolnick, M. H. A cell cycle regulator
potentially involved in genesis of many tumor types. Science (Washington DC), 264:
436—440, 1994.
20. Yamakawa, K., Takahashi, T., Horio, Y., Murata, Y., Takahashi, E., Hibi. K.,
Yokoyama, S., Ueda, R., and Nakamura, Y. Frequent homozygous deletions in lung
cancer cell lines detected by a DNA marker located at 3p2l.3-p22. Oncogene. 8:
327—330,1993.
21. Daly, M. C., Xiang, R. H., Buchbagen, D., Hensel, C. H., Garcia. D. K., Killary.
A. M., Minna,J. D., and Naylor, S. L. A homozygousdeletion on chromosome3 in
a smallcell lungcancercell line correlateswith a regionof tumorsuppressoractivity.
Oncogene, 8: 1721—1729.
1993.
lesions, since such information may, when correlated with clinical
outcomes, provide very useful prognostic information to improve
patient care.
22. Kok, K.. van den Berg, A., Veldhuis, P. M.. van der Veen, A. Y., Franke, M.,
1351
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 1997 American Association for Cancer Research.
CHROMOSOME3p DELETIONSIN LUNG TUMORS
Schoenmakers,E. F., Hulsbeek,M. M., van der Hout,A. H., de Leij, L., van de Ven,
W., andBuys, C. H. C. M. A homozygousdeletionin a smallcell lungcancercell line
42. Gyapay, G., Morissene, J., Vignal, A., Dib, C., Fizames. C., Millasseau, P., Marc, S.,
Bemardi, G., Lathrop, M.. and Weissenbach, J. The 1993-94 Genethon human
genetic linkage map. Nat. Genet., 7: 246—339,1994.
involving a 3p2l region with a marked instability in yeast artificial chromosomes.
Cancer Res., 54: 4183—4187,1994.
43. Smith, D. I., Glover, T. W., Gemmill, R., Drabkin. H., O'Connell,
J., Yamakawa,K., andOgawa,A. Characterizationof an 800 kb regionat 3p22—p2l.3
that was homozygously deleted in a lung cancer cell line. Hum. Mol. Genet., 3:
1341—1344.
1994.
24. Rabbitts, P., Bergh, J.. Douglas, J., Collins, F., and Waters, J. A submicroscopic
homozygous deletion at the D3S3 locus in a cell line isolated from a small cell lung
carcinoma. Genes Chromosomes & Cancer, 2: 231—238,1990.
25. Roche, J., Boldog, F., Robinson, M., Robinson, L., Varella-Garcia, M., Swanton, M.,
3 mapping 1994. Ann Arbor, Michigan, May 8—9.1994. Cytogenet. Cell Genet., 68:
125-146,1995.
44. Roche, J., Whisenant, E., Boldog, F., Loeb, D., Vance, J. M., and Drabkin, H.
Dinucleotiderepeatsflankingthe renalcarcinomabreakpointat 3p14.2. Hum. Mol.
Genet., 3: 215, 1994.
45. Gemmill, R. M., Chumakov, I., Scott, P., Waggoner, B., Rigault, P., Cypser, J., Chen,
Q., Weissenbach,J., Gardiner,K., Wang, H., Pekarski,Y., Le Gall, I., Le Paslier,D.,
Waggoner, B.. Fishel, R., Franklin, W., Gemmill, R., and Drabkin, H. Distinct 3p2l.3
deletions in lung cancer and identification of a new human semaphorin. Oncogene,
12: 1289—1297.1996.
26. Kastury, K., Baffa, R., Dnick, T., Ohta, M., Cotticelli, M. G., Inoue, H., Negrini, M.,
Rugge, M., Huang, D., Croce, C. M., Palazzo, J., and Huebner, K. Potential gastro
intestinal tumor suppressor locus at the 3pl'4.2 FRA3B site identified by homozygous
deletions in tumor cell lines. Cancer Res., 56: 978—983,1996.
Guillon, S., Li, E., Robinson, L., Hahner, L., Todd, S., Cohen, D., and Drabkin, H. A.
An second-generationYAC-contig map for humanchromosome3. Nature(Lond.),
377S:
humangenome. Nature(Lond.), 366: 698—701,1993.
47. Hedley, D. W., Friedlander, M. L., Taylor, I. W., Rugg, C. A., and Musgrove, E. A.
Method for analysis of cellular DNA content of paraffin-embedded
z..Mori,
M.,McCue,
P.,Druck,
T.,Croce,
C.M.,andHuebner,
K.TheFHJT
gene, 48.
breakpoint,is abnormalin digestivetractcancers.Cell,84: 587—597,
1996.
28. Boldog, F., Gemmill, R. M., West, J., Robinson, M., Robinson, L., Li, E., Roche, J.,
Todd. S., Lundstrom, R.. Jacobson, J., Klinger, H., and Drabkin, H. A. 3pl4
homozygous deletions and sequence analysis of FRA3B. Hum. Mol. Genet., 6:
193-203, 1997.
29. Ogawa, 0., Kakehi, Y., Ogawa, K.. Koshiba, M., Sugiyama, T., and Yoshida, 0.
Allelic loss at chromosome 3p characterizes clear cell phenotype of renal cell
6q in primary human small cell lung cancer. Cancer Res., 54: 2322—2326,1994.
50. Sozzi, G., Miozzo, M., Pastorino, U., Pilotti, S., Donghi, R., Giarola, M., DeGregorio,
L., Manenti,G., Radice,P., Minoletti,F., Della Ports,G., andPierotti,M. A. Genetic
evidence for an independent origin of multiple preneoplastic and neoplastic lung
lesions. Cancer Res., 55: 135-140, 1996.
51. Shridhar, V., Siegfried, J., Hunt, J., del Mar Alonso, M., and Smith, D. 1. Genetic
instability of microsatellite
of chromosome 3 in renal-cell cancer correlates with the cytological tumour type. mt.
J. Cancer,53: 353—357.
1993.
31. mar, B., Branch, H., Talmadge, C., and Linehan, M. Loss of alleles of loci on the short
arm of chromosome 3 in renal cell carcinoma. Nature (Lond.), 327: 721—724,1987.
32. Chung, G. T., Huang, D. P., Lo, K. W., Chan, M. K., and Wong, F. W. Genetic lesion
in the carcinogenesis of cervical cancer. Anticancer Res., 12: 1485—1490,
1992.
33. Yokota. J., Tsukada, Y., Nakajima, T., Gotoh, M., Shimosato, Y., Mon. N., Tsuno
kawa, Y., Sugimura, T., and Terada, M. Loss of heterozygosity on the short arm of
chromosome 3 in carcinoma of the uterine cervix. Cancer Res., 49: 3598—3601,1989.
34. Latif, F., Fivash, M., Glenn, G., Tory. K., Orcutt, M. L., Hampsch, K., Delisio, J.,
Lerman, M., Cowan, 3., Beckeu, M., and Weichselbaum, R. Chromosome 3p dele
tions in head and neck carcinomas: statistical ascertainment of alleic loss. Cancer
Res., 52: 1451—1456,1992.
35. Huang, D. P., Lo, K. W., Choi, P. H., Ng. A. Y., Tsao, S. Y., Yiu, G. K., and Lee,
36. Killary. A. M., Wolf, M. E., Giambernardi, T. A., and Naylor, S. L Definition of a
tumorsuppressorlocus withinhumanchromosome3p2l—p22.
Proc.Natl. Acad. Sci.
USA, 89: 10877—10881,
1992.
37. Sanchez. Y.. el-Naggar, A.. Pathak, S.. and Killary, A. M. A tumor suppressor locus
within 3pl4—pl2mediates rapid cell death of renal cell carcinoma in vivo.Proc. Nati.
Aced. Sci. USA, 91: 3383—3387,
1994.
38. Magovcevic, I., Ang, S-L., Seidman, JG., Tolman, Ci., Neer, El., and Morton, CC.
Regional localization of the human G protein alpha i2 (GNAI2) gene: assignment to
3p2l anda relatedsequence(GNAI2L)to l2pl2—pl3.Genomics,12: 125—129,
1992.
in the 3p2l.3 small cell lung cancer deletion region and demonstrate distinct expres
in many non-small
cell lung carcinomas.
52. Todd, S., Roche, J., Hahner, L., Bolin, R., Drabkin, H. A., and Gemmill, R. M. YAC
contigs coveringan 8 megabaseregionof 3p deletedin the smallcell lung cancercell
line U2020. Genomics, 25: 19—28,1995.
53. Cairns, P., Mao, L., Merlo, A., Lee, D. J., Schwab, D., Eby, Y., Tokino, K.. van der
Riet, P.. Blaugrund, J. E., and Sidransky. D. Rates of p16 (MTSJ) mutations in
primary tumors with 9p loss. Science (Washington DC), 265: 415—416, 1994.
54. Merlo, A., Herman, J. 0.. Mao, L, Lee, D. J., Gabrielson, E., Burger, P. C., Baylin,
S. B., and Sidransky,D. 5' CpG island methylationis associatedwith transcriptional
silencing of the tumour suppressor pl6/CDKN2IMTSI in human cancers. Nat. Med.,
1: 686—692, 1995.
55. Boldog. F. L., Gemmill, R. M., Wilke, C. M., Glover, T. W., Nilsson, A. S.,
Chandrasekharappa, S. C.. Brown, R. S., Li, F. P., and Drabkin, H. A. Positional
cloning of the hereditary renal carcinoma 3;8 chromosome translocation breakpoint.
Proc. Nail. Aced. Sci. USA, 90: 8509-8513, 1993.
56. Glover, T. W., Coyle-Morris, J. F., Li, F. P.. Brown, R. S., Berger, C. S., Gemmill,
R. M.. andHecht,F. Translocationt(3;8)(pl4.2;q24.l) in renalcell carcinomaaffects
expression of the common fragile site at 3p14(FRA3B) in lymphocytes. Cancer
Genet. Cytogenet., 31: 69—73,1988.
57. Sozzi, 0., Veronese, M. L, Negrini, M., Baffa, R., Cotticelli, M. G.. Inoue, H., Pilotti,
S., Dc Gregorio,L., Pastorino,U., Pierotti,M. A., Ohta,M.,, Huebner,K., andCroce,
C. M. The FHJTgene 3pl&2 is abnonnalin lung cancer.Cell, 85: 17—26,
1996.
58. Thiagalingam, S., Lisitsyn, N. A., Hamaguchi, M., Wigler, M. H., Willson, J. K. V.,
Markowitz,S. D., Leach,F. S., Kinzler,K. W., and Vogelstein, B. Evaluationof the
39. Sekido, Y., Bader, S.. Latif, F., Chen, J. Y., Dub, F. M., Wei, M. H., Albanesi, J. P.,
Lee,C. C., Lerman,M.I., andMinna,J. D. HumansemaphorinsA(V)andIVreside
sequences
Cancer Res., 54: 2084—2087,1994.
J. C. Loss of heterozygosityon the short arm of chromosome3 in nasopharyngeal
carcinoma. Cancer Genet. Cytogenet., 54: 91—99,1991.
pathological
materialusingflowcytometry.J. Histochem.Cytochem.,31: 1333—1335,
1995.
Mooibroek, H., Osinga, J., Postmus, P. E., Cariitt, B., and Buys, C. H. Loss of
heterozygosity for a chromosome 3 sequence presumably at 3p2l in small cell lung
cancer. Cancer Genet. Cytogenet.. 27: 361—365,1987.
49. Merlo, A., Gabrielson, E., Mabry, M., Voilmer, R., Baylin, S. B., and Sidransky, D.
Homozygous deletion on chromosome 9p and loss of heterozygosity on 9q, 6p, and
carcinoma.CancerRes.,51: 949—953,
1991.
30. vanderHout,A. H.. vandenBerg,E., vanderVlies,P., Dijkhuizen.T., Storkel,S.,
Oosterhuis, J. W., de Jong, B., and Buys. C. H. Loss ofheterozygosity at the short arm
299—319, 1995.
46. Cohen, D.. Chumakov, I., and Weissenbach, J. A first-generation physical map of the
27. Ohta, M., Inoue, H., Cotticelli, M. G., Kastury, K., Baffa, R., Palazzo, J., Siprashvili,
spanning the chromosome 3pl4.2 fragile site and renal carcinoma-associated t(3;8)
P., and Naylor,
S. L. Reportandabstractsof the fifth internationalworkshopon humanchromosome
23. Murata, Y., Tamari, M., Takahashi, T., Horio. Y., Hibi, K., Yokoyama, S., Inazawa,
FlIT
gene
in colorectal
cancers.
Cancer
Rca.,
56:
2936—2939.
1996.
59. Wei, M. H., Latif, F., Bader, S., Kashuba, V., Chen, J. Y., Duh, F. M., Sekido, Y.,
Lee, C. C., Geil, L., Kuzmin, I., Zabarovsky, E., Klein, G., Thar, B., Minna, J. D., and
Lerman, M. I. Construction of a 600-kilobase cosmid clone contig and generation of
a transcriptional map surrounding the lung cancer tumor suppressor gene (TSG) locus
on humanchromosome3p2l.3: progresstowardthe isolation of a lung cancerTSG.
sion pattems. Proc. Nail. Acad. Sci. USA, 93: 4120—4125, 1996.
40. Hibi, K., Yamakawa, K., Ueda, R., Horio, Y., Murata, Y., Tamari, M., Uchida, K.,
Takahashi, T., and Nakamura, Y. Aberrant upregulation of a novel integrin alpha
subunit gene at 3p2l3 in small cell lung cancer. Oncogene, 9: 61 1—619,1994.
Cancer Rca., 56: 1487—1492,1996.
41 . Todd, S., Naylor, S. L., Drabkin, H. A., and Gemmill, R. M. Dinucleotide repeat
polymorphism (D3S1776) on human chromosome 3p. Hum. Mol. Genet., 3: 841,
1994.
60. Parsons, R., Li, G. M., Longley, M., Modrich, P., Liu, B., Berk, T., Hamilton, S. R.,
Kinzler, K. W., and Vogelstein, B. Mismatch repair deficiency in phenotypically
normal human cells. Science (Washington DC), 268: 738—740,1995.
1352
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 1997 American Association for Cancer Research.
Homozygous Deletions of Human Chromosome 3p in Lung
Tumors
Sean Todd, Wilbur A. Franklin, Marileila Varella-Garcia, et al.
Cancer Res 1997;57:1344-1352.
Updated version
E-mail alerts
Reprints and
Subscriptions
Permissions
Access the most recent version of this article at:
http://cancerres.aacrjournals.org/content/57/7/1344
Sign up to receive free email-alerts related to this article or journal.
To order reprints of this article or to subscribe to the journal, contact the AACR Publications
Department at [email protected].
To request permission to re-use all or part of this article, contact the AACR Publications
Department at [email protected].
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 1997 American Association for Cancer Research.