Oncogene (1998) 17, 3499 ± 3505
ã 1998 Stockton Press All rights reserved 0950 ± 9232/98 $12.00
http://www.stockton-press.co.uk/onc
Novel regions of allelic deletion on chromosome 18p in tumors of the lung,
brain and breast
Yen Tran1, Khalid Benbatoul1, Karen Gorse2, Sandra Rempel3 Andrew Futreal4, Mark Green5
and Irene Newsham1,2
1
Theodore Gildred Cancer Center, Department of Medicine, University of California-San Diego, La Jolla, California 92093;
Department of Anatomy, Medical College of Virginia at Virginia Commonwealth University, Richmond, Virginia 23298;
3
Department of Neurosurgery, Henry Ford Hospital, Detroit, Michigan 48202; 4Department of Surgery, Duke University,
North Carolina 27710; 5Hollings Cancer Center, University of South Carolina Medical Center, Charleston, South Carolina 29425,
USA
2
Lung cancer is now the number one cause of cancer
death for both men and women. An age-adjusted analysis
over the past 25 years shows that in women speci®cally,
lung cancer incidence is on the rise. It is estimated that
10 ± 20 genetic events including the alteration of
oncogenes and tumor suppressor genes will have occurred
by the time a lung tumor becomes clinically evident. In
an e€ort to identify regions containing novel cancer
genes, chromosome 18p11, a band not previously
implicated in disease, was examined for loss of
heterozygosity (LOH). In this study, 50 matched normal
and NSCLC tumor samples were examined using six
18p11 and one 18q12.3 PCR-based polymorphic markers. In addition, LOH was examined in 29 glioblastoma
pairs and 14 paired breast carcinomas. This analysis has
revealed potentially two regions of LOH in 18p11 in up
to 38% of the tumor samples examined. The regions of
LOH identi®ed included a 2 cm area between markers
D18S59 and D18S476, and a more proximal, 25 cm
region of intermediate frequency between D18S452 and
D18S453. These results provide evidence for the presence
of one or more potential tumor suppressor genes on the
short arm of chromosome 18 which may be involved in
NSCLC, brain tumors and possibly breast carcinomas as
well.
Keywords: loss of heterozygosity; chromosome 18p;
Non-small cell lung cancer; breast cancer; glioblastoma
Introduction
An estimated 178 000 new lung cancer cases are
expected in the United States in 1997 with 80% of
these histologically designated as Non-small cell lung
carcinomas (NSCLC). In the same year, more than
160 400 individuals are expected to die of lung cancer,
making this disease the number one cause of cancer
death for both men and women (Parker et al., 1997).
The 86% mortality rate for NSCLC has remained
fairly constant since 1980 despite advances in cytotoxic
drug development, radiotherapy and patient manage-
Correspondence: I Newsham
Received 7 May 1998; revised 26 June 1998; accepted 3 July 1998
ment. This may be due, in part, to the lack of
information about genetic loci involved in its
pathogenesis. Current theories suggest that as many
as 10 ± 20 events including alterations of oncogenes and
tumor suppressor genes will have occurred by the time
lung cancer is clinically evident (Minna et al., 1991).
Tumor suppressor genes involved in NSCLC have been
located on chromosome bands 3p, (Hung et al., 1995),
5q (Fong et al., 1995a), 9p (Kishimoto et al., 1995;
Center et al., 1993), 11p (Tran and Newsham, 1996;
Bepler and Garcia-Blanco, 1994), 11q (Rasio et al.,
1995), and 17p and 17q (Fong et al., 1995b).
Molecular evidence supporting a role for genes on
chromosomal arm 18p in NSCLC is scarce. Several
cytogenetic analyses on primary NSCLC tumor samples
have revealed that chromosome 18 is one of several
chromosomes preferentially lost as a result of tumorigenesis (Testa et al., 1992, 1994). Chromosome 18 was
lost in approximately 35% of the 63 primary tumor
samples analysed cytogenetically. Other signi®cant losses
were measured for chromosomes 9, 13, 21, and 22 (Testa
et al., 1994). Another study of pleural e€usions containing metastatic or invasive NSCLC tumor cells detected
chromosome 18 loss at a frequency of 45% (Lukeis et al.,
1993) in all histological subtypes of lung cancer. In
addition, chromosome 18 structural rearrangements
have been found to be rare events, suggesting that loss
of a gene(s) on this chromosome is an important
mutational event in the tumorigenic process (Merlens
et al., 1997). Previous loss of heterozygosity studies on
NSCLC have identi®ed allelic deletions on chromosome
18q (Shiseki et al., 1996; Nagatake et al., 1996a,b) but
not on chromosome 18p, although often only one 18p
marker has been analysed. However, in one analysis of
29 in®ltrating ductal carcinomas of the breast, LOH on
band 18p was measured at 62% (with 14.5% being
con®ned to the distal region of 18p), indicating a putative
tumor suppressor gene may reside in this region of the
chromosome (Huang et al., 1995).
To further investigate the presence of potential
novel tumor-related genes in chromosomal band
18p11, we examined 50 matched normal and
NSCLC tumor samples as well as 29 glioblastoma
pairs and 14 paired breast carcinoma samples for
loss of heterozygosity (LOH) using six 18p11 and
one 18q12.3 PCR-based polymorphic markers. This
study revealed two potential regions of LOH on
chromsomal band 18p11 in these various tumor
types.
Allelic deletion of chromosome 18p in adult tumors
Y Tran et al
3500
Results
Loss of heterozygosity on chromosome 18p11 in NSCLC
Matched pairs of normal/tumor DNA from 50
NSCLCs, 29 astrocytomas and 14 breast carcinomas
were assessed for allelic deletions at six polymorphic
marker loci spanning chromosome band 18p11 and one
marker in 18q12.3. Our set of NSCLC tumors included
18 squamous cell, 23 adenocarcinoma, two large cell
and seven undi€ subtypes. The overall frequency for
LOH on 18p11 in NSCLC was 38% (19/50). As whole
sample DNA extracts were used for these analyses,
normal in®ltrating cells would be expected to potentially mask a percentage of tumor cells exhibiting
LOH. Most likely therefore our reported frequencies
are under estimates.
Frequencies for individual marker loci in all
NSCLCs ranged from 11 ± 32% with peaks at
D18S476 (32%) in 18p11.3 and D18S452 (21%) in
18p11.2 (Figure 1). As the LOH frequency dropped
between these two markers (10% at D18S52) this
suggests the presence of two potential tumor suppressor genes on chromosome 18p. In our set of tumors,
LOH appeared more frequently in adenocarcinomas
[48% (11/23)] than in squamous cell carcinomas [28%
(5/18)] although the number of individual tumor
subtypes analysed is small. Isolation of the tumor
suppressor genes corresponding to these deleted regions
will determine whether a real subtype speci®city exists
for these LOH regions. One tumor, 2503T, exhibited
microsatellite instability at all markers examined in
18p11 (data not shown) and was not included in the
remaining data analysis.
Examples of allelic patterns for critical NSCLC
samples are shown in Figure 2 with the corresponding densitometrically determined reductions in allelic
ratios. In 7/19 (32%) tumors, allelic deletion was
restricted to the most distal regions of 18p as
indicated by maintenance of heterozygosity at
markers distal to or at D18S453 in 18p11.2. These
results are not evidence however of the absence of
LOH in 18q, as only the 18q12.3 marker D18S57
was analysed in this study. Several other reports
looking at NSCLC and breast carcinomas have
shown that LOH can be present on both the long
and short arms of a chromosome either independently or concommittantly (Huang et al., 1995;
Hampton et al., 1994). Our data indicate that only
11% (2/19) of NSCLC tumors are suggestive of
whole chromosome 18 loss based on the markers
analysed.
Two potential regions of LOH in chromosome 18p11
This analysis raises the possibility that two regions of
LOH exist in NSCLCs and brain tumors. Figure 3
schematically depicts the regions of allelic deletion for
the 19 NSCLC tumors in which LOH was measured
in 18p11. Region 1 (the telomeric region) is de®ned by
markers D18S59 and D18S476, an area encompassing
2 cm. Speci®c examples of Region 1 LOH are shown
in adenocarcinomas 1N/T and 1910B/C, squamous
cell carcinoma 71/70 and undi€erentiated NSCLCs
142/141 and 28N/T. The second region, of intermediate LOH frequency, is located proximal to
D18S452 and distal to D18S453, a region involving
25 cm of sequence. LOH encompassing both Regions
1 and 2 was detected in 68% (13/19) of tumors
whereas 26% (5/19) showed LOH only in the most
distal region. Only one NSCLC tumor (5%) underwent allelic deletion of the second LOH region alone.
Of the 13 tumors with LOH involving both Regions 1
and 2, the deletion event was continuous in 31% (4/
13) and discontinuous in the remaining 69% (9/13)
suggesting independent genetic events were responsible
for two thirds of the multiple allelic deletions found in
18p. The prevalence for deletion of both regions in
NSCLC tumors may indicate that alteration of both
tumor suppressor genes is necessary for tumor
formation. It might also be that alteration of the
gene in Region 1 preceeds and/or potentially a€ects
the alteration of sequences positioned in Region 2.
Given the proximity of these two regions however, it
may also be envisioned that genetic alteration in one
region a€ects the stability of nearby sequences via
secondary structures generated during mitotic recombination.
Loss of heterozygosity on 18p11 in gliomas and breast
carcinomas
Figure 1 Frequency of allelic deletion in chromosome band
18p11. Loss of heterozygosity (LOH) was measured at six
polymorphic loci along 18p11. Bars, frequency of LOH as
determined in informative matched normal/tumor pairs of Nonsmall cell lung cancer
To determine if the chromosome 18 LOH regions
were speci®c to NSCLCs or represented alterations
a€ecting sequences which might be important in
other adult neoplasms, we analysed 29 normal/tumor
paired gliomas and 14 normal/tumor paired breast
carcinomas. Glioma specimens were categorized
according to the WHO grading system (Kleihues et
al., 1993) as grade 1 pilocytic astrocytoma (PA),
grade II astrocytoma (A), grade III anaplastic
astrocytoma (AA), and grade IV glioblastoma
multiforme (GBM). Specimens analysed in this
study included 1 PA, 1 A, 8 AAs and 19 GBMs.
Several examples of LOH are shown in Figure 2,
with complete results depicted schematically in
Allelic deletion of chromosome 18p in adult tumors
Y Tran et al
Figure 4. In gliomas, the overall frequency of LOH
was 31%(9/29). LOH in Region 1 alone, Region 2
alone, and both regions together occurred at
approximately the same frequency (33%, 22% and
44% respectively). LOH did not appear to exhibit
grade-speci®city as allelic deletions were detected in
1/2 As, 3/8 AAs and 5/19 GBMs.
Of the 14 invasive ductal breast carcinomas
analysed, LOH was identi®ed in 21% (3/14). No
breast tumors showed LOH encompassing both
Regions 1 and 2 or involving Region 2 alone. The
smallest region of overlap for Region 1 (for tumors 1,
7, and 8) was distal to D18S481 in 18p11.3. In
addition to allelic deletion, 14% of these breast
tumors displayed microsatellite instability (MI) for
multiple chromosome 18p loci. A literature search
revealed only one previous analysis which reported a
MI frequency of 4% for this chromosome in breast
tumors (Huang et al., 1995). No examples of MI were
found in the gliomas studied, consistent with other
published frequencies (1 ± 8%) (Zhu et al., 1996). Only
1/50 (2%) (tumor 2503) of the NSCLC tumors
Figure 2 Examples of tumors critical in de®ning the minimal regions of LOH on chromosome band 18p11 in Non-small cell lung
carcinomas, gliomas and breast carcinomas. Autoradiographs represent normal (N) and matched tumor (T) DNA samples from
patients whose case numbers are shown to the left. Chromosome 18 locus names appear at the top, ordered from most telomeric
(D18S59) to most proximal (D18S453). Densitometrically calculated allelic reductions (D) in the intensity of alleles between tumor
and normal DNA samples is shown below each autoradiograph. A value greater than 0.5 was indicative of LOH; .04-.05 indicated
allelic imbalance; 50.4 was evidence of maintenance of heterozygosity
3501
Allelic deletion of chromosome 18p in adult tumors
Y Tran et al
3502
showed MI and this a€ected all loci analysed.
Reported MI frequencies for NSCLC have been
quite variable, ranging from 6.5 ± 34% for primary
tumors (Fong et al., 1995b; Adachi et al., 1995;
Shridhar et al., 1994) and up to 55% in metastases
(Adachi et al., 1995), depending on the markers
analysed. None of these previous analyses on
NSCLCs and brain tumors utilized markers from
chromosome 18p.
Discussion
The results of this study provide evidence for the
presence of one or more tumor suppressor genes on the
short arm of chromosome 18p which may be involved
in NSCLC, brain tumors and possibly breast
carcinomas as well. The analysis also serves to de®ne
the location of these putative tumor suppressor genes
to a 2 cm distal region between D18S59 and D18S476
Figure 3 Schematic representation of the regions of allelic deletion measured in nineteen Non-small cell lung carcinomas (NSCLC),
separated by subtype. The 18p loci analysed are shown telomeric (left) to centromeric (right) along the top. Tumor cases are
identi®ed by the numbers shown to the left. Shaded areas indicate the two distinct regions of LOH de®ned in this analysis. Closed
circles, loss of heterozygosity; Striped circles, allelic imbalance; Open circles, maintenance of heterozygosity; Dotted circles, not
informative; ND, not done
Allelic deletion of chromosome 18p in adult tumors
Y Tran et al
in 18p11.3 and a 25 cm region between D18S452 and
D18S453 in 18p11.2. This is the ®rst description of loss
of heterozygosity a€ecting the short arm of chromosome 18p in either NSCLCs or primary brain tumors.
One previous study on in®ltrating ductal carcinoma of
the breast, showed an overall frequency of 62% LOH
for chromosome 18 with allelic deletion con®ned to
distal 18p in 14.5% of the cases (Huang et al., 1995)
although only a few markers in distal 18p were utilized.
No apparent association between chromosome 18 LOH
and tumor status, tumor size, lymph node involvement
or metastases was detected.
Tumors with LOH maintained heterozygosity for at
least one marker on chromosome 18 in 16/19 (84%) of
NSCLC tumors and in 7/9 (78%) of brain tumors.
Such retention indicates that loss of the whole
chromosome 18 is a rare event in NSCLC. Intrachro-
mosomal deletion has also been reported as the
majority of LOH events on chromosome 11 in lung
and breast cancer (Winqvist et al., 1995; Shridhar et
al., 1994). This is in contrast to the mutational events
leading to LOH on chromosome 3 in NSCLC
(Rabbitts et al., 1990) and chromosome 10 in
malignant meningiomas (Rempel et al., 1993) where
whole chromosome loss rather than regional deletion
appears to be the underlying mechanism of choice.
By analysing a series of six markers in 18p11, two
potential regions of LOH were discovered. Region 1
(the telomeric region) is de®ned by markers D18S59
and D18S476, an area representing 2 cm. Region 2 is
proximal to D18S452 and distal to D18S453 in
chromosome band 18p11.2, was of intermediate
frequency, and encompassed 25 cm of sequence. In
NSCLCs, two-thirds as many tumors displayed LOH
Figure 4 Schematic representation of the regions of allelic deletion discovered in nine astrocytomas and three breast carcinomas.
The 18p loci analysed are as in Figure 3. Tumor cases are identi®ed by the numbers shown to the left. Shaded areas indicate distinct
regions of LOH. Circles are as described in Figure 3. MI, microsatellite instability; ND, not done; A, Astrocytoma; AA, Anaplastic
Astrocytoma; GBM, Glioblastoma multiforme
3503
Allelic deletion of chromosome 18p in adult tumors
Y Tran et al
3504
in both Regions 1 and 2 (68%) as for Region 1 alone
(26%) whereas only 5% (1/29) underwent allelic
deletion of the second LOH region alone (Table 1).
Of the 13 tumors with LOH involving both Regions 1
and 2, the frequency of separate or continuous LOH in
Regions 1 and 2 was di€erent for di€erent subtypes of
lung tumors. For adenocarcinomas, 70% (6/9) were
indicative of continuous loss while 30% (3/9) showed
two independent regions of LOH. In contrast, only
25% (1/4) of tumors with a squamous histology had
continuous allelic deletion whereas 75% (3/4) exhibited
two separate LOH regions. It will be necessary to
examine this particular characteristic in an expanded
group of tumors but it may re¯ect a di€erence in
timing for the mutation of loci or a di€erence in the
mutational mechanisms a€ecting chromosome 18p in
the cellular precursors to these speci®c lung tumor
subtypes. The prevalence for deletion of both regions
in NSCLC tumors may also indicate that alteration of
both tumor suppressor genes is necessary for tumor
formation. Since in 18/19 NSCLCs LOH never occurs
in Region 2 without concomitant LOH in Region 1,
one might propose that alteration of a gene in Region
1 preceeds and/or potentially a€ects the alteration of
sequences positioned in Region 2 in NSCLCs.
In brain tumors, the overall LOH frequency was
31% (9/29). LOH encompassing both Regions 1 and 2
was twice as frequent (44%) as for Region 1 alone
(22%), much the same as for NSCLC. However, 3/9
brain tumors (33%) had measurable LOH in Region 2
only, an uncommon event in NSCLCs and breast
tumors. Additionally, continuous LOH of both regions
was much more frequent (75%) than separate LOH
events (25%) in brain tumors whereas these fractions
were 31% and 69% in NSCLCs. And ®nally, LOH in
Region 1 (3/14) alone was found only in breast
carcinomas although the number of tumors analysed
here was small. The tumor- and subtype-speci®c
regional LOH di€erences between gliomas, breast
carcinomas and NSCLCs suggests these two putative
tumor suppressor genes might play roles of varying
importance in the tumorigenic process in di€erent
tissues. These di€erences may also re¯ect mutational
mechanisms speci®c to the individual tumor types, be
suggestive of a preferential order of alteration for
genes, or result from variating environmental exposures
for the representative organ tissues.
The possibility of multiple LOH regions on 18p for
both lung and brain tumors presents a scenario where
LOH at these di€erent loci could potentially occur on
the same chromosome 18 concurrently or on separate
chromosomal homologs. In breast carcinomas, some
Table 1
Regional LOH speci®city on Chromosome 18p
individual tumors have been found which exhibit
chromosome 11p and 11q LOH at varying frequencies
indicating that tumors can in fact be heterogeneous for
these events and suggesting that multiple genetic events
a€ecting the same chromosome can act independently
of one another, contributing to the stepwise development of the fully malignant phenotype (Winqvist et al.,
1995). It would be of interest to analyse NSCLC
tumors with distinct regions of LOH with regard to
their inherited chromosome 18 haplotypes, to determine if the individual LOH regions are in fact
independent events occurring on the same or separate
chromosomes 18.
The biological importance of these putative loci and
their clinical relevance in the pathogenesis of adult
neoplasms is at present unclear. Their signi®cance will
most appropriately be addressed once the corresponding tumor suppressor genes from these two distinct
regions are isolated and analysed for functionality and
mutational status. The narrow 2 cm LOH area
reported for the most distal Region 1 should add
focus to the search for such genes and help to decipher
the mechanisms responsible for a€ecting chromosomal
band 18p in tumors of the lung, brain and breast.
Materials and methods
Patient materials
Primary Non-small cell lung carcinoma and patientmatched normal lung tissue samples were obtained from
50 unselected patients upon surgical resection at the UCSD
Medical Center as well as through the Cooperative Human
Tissue Network Western Division at Case Western Reserve
University. All samples, after surgical removal, were
immediately snap-frozen in liquid nitrogen and stored at
7808C until further use. Each fresh lung tumor and
accompanying normal tissue was histopathologically
characterized at the time of surgery with informed consent
and the corresponding pathology reports have been
obtained.
Glioma specimens were divided immediately upon surgical
removal. Partial tumor specimens were ®xed in 10% bu€ered
formalin and embedded in paran for pathological diagnosis. The remaining tumor tissue was immediately snapfrozen in liquid nitrogen and later processed for DNA as
previously reported. Blood specimens were collected at the
time of surgery and stored at 48C. Peripheral blood
lymphocyte DNA was isolated within 3 days using a DNA
extraction kit according to manufacturer's instructions
(Stratagene). Institutional IRB-approved informed consent
was obtained from all patients or the patient's guardian.
Pathology reports were reviewed and specimens selected at
random on the basis of availability of both normal and
tumor DNA.
Primary breast cancer tissue along with matching
peripheral blood lymphocytes were obtained with informed
consent for tissue banking from patients receiving care at
Duke University Medical Center. All tumors were characterized as in®ltrating ductal carcinomas by pathological
examination. Samples were stored frozen until DNA
extraction.
LOH Region
Lung (50)a
Glioma (29)
Breast (14)
Region 1
Region 2
Region 1+2
26%
5%
68%
22%
33%
44%
21%
0%
0%
Region 1+2
Continuous
Region 1+2
Discontinuous
31%
75%
±
PCR analysis
69%
25%
±
DNA from each normal/tumor pair was subjected to PCR
analysis using six chromosome 18p markers including
D18S59,
D18S476,
D81S481,
D18S52,
D18S452,
D18S453, and one 18q12.3 marker D18S57. The likely
a
Number in parentheses represents total number of normal/tumor
pairs analysed
Allelic deletion of chromosome 18p in adult tumors
Y Tran et al
order of these markers and estimated interval distances are
based on both GeÂneÂthon and CEPH linkage data (Rojas et
al., 1995; Buetow et al, 1994; Gyapay et al., 1994). Detailed
ampli®cation information for all polymorphic markers
used in this study were obtained from the Genome
DataBase (Johns Hopkins University Medical School,
Baltimore, MD, USA). Conditions for denaturation,
annealing, and extension for each primer pair were
essentially as reported except that 0.3 ± 1 unit of Perfect
Match (Stratagene) was added to each reaction to optimize
the equal ampli®cation of alleles. Primer end-labeling,
ampli®cation reactions, and resolution of PCR products by
denaturing electrophoresis were carried out as previously
described (Besnard-GueÂrin et al., 1996).
a densitometer, the relative ratio of normal and tumor
alleles was ®rst normalized and then compared. When the
allelic ratio in the tumor DNA was reduced by more than
50% (D0.5) from that found in the normal DNA, this
sample was denoted as having LOH at that locus. Allelic
ratios reduced by between 40 ± 50% (D0.4 ± D0.5) were
scored as allelic imbalances and interpreted as LOH with
residual allele signals from contaminating normal tissue.
Tumors with ratios less than 40% (5D0.4) were assigned
as maintaining heterozygosity. By choosing a stringent
allelic loss ratio (0.5, as compared to 0.3 in most other
analyses) for determination of LOH, false negative LOH
resulting from allele overlap with polymerase slippage
bands should be minimized.
LOH analysis
LOH for tumor DNA samples was assessed ®rst by visual
inspection followed by densitometric con®rmation as
described previously (Hampton et al., 1994). When
discordant results were obtained between visual and
densitometric interpretations, PCR analyses were repeated
with multiple exposures and a consensus result reported.
Several chromosome 18p markers consistently resulted in
unequally ampli®able alleles in the normal samples. Using
Acknowledgements
This work is dedicated to the memory of Candace L
Andersson. The authors wish to acknowledge Dr Oliver
BoÈgler for his assistance in the densitometric analyses and
for his critical reading of the manuscript. This work was
supported in part by funds from the UCSD Cancer Center
Foundation, NIH grants 5-P30-CA23100-13, 1 P20CA66215-01 and R29-CA77730-02 (IN).
References
Adachi J-I, Shiseki M, Okazaki T, Ishimaru G, Noguchi M,
Hiroashi S and Yakota J. (1995). Gene Chrom & Cancer,
14, 301 ± 306.
Bepler G and Garcia-Blanco M. (1994). Proc. Natl. Acad.
Sci. USA, 91, 5513 ± 5517.
Besnard-GueÂrin C, Newsham I, Winqvist R and Cavenee W.
(1996). Human Genet., 97, 163 ± 170.
Buetow K, Wever J, Ludwigsen S, Scherpbier-Heddema T,
Duyk G, Sheeld V, Zhenyuan W and Murray J. (1994).
Nat. Genet., 6, 391 ± 393.
Center R, Lukeis R, Dietzsch E, Gillespie M and Garson O.
(1993). Genes Chrom & Cancer, 7, 47 ± 53.
Fong K, Zimmerman PV and Smith P. (1995a). Cancer Res.,
55, 220 ± 223.
Fong K, Kida Y, Zimmerman P, Ikenaga M and Smith P.
(1995b). Cancer Res., 55, 4268 ± 4272.
Gyapay G, Morissette J, Vignal A, Dib C, Fizames C,
Millasseau P, Marc S, Bernardi G, Lathrop M and
Weissenbach J. (1994). Nat. Genet., 7, 246 ± 339.
Hampton G, Mannermaa A, Winqvist R, Alavaikko M,
Blanco G, Taskinen P, Kiviniemi H, Newsham I, Cavenee
W and Evans G. (1994). Cancer Res., 54, 4586 ± 4589.
Huang TH-M, Yeh P L-H, Martin MB, Straub RE, Gilliam
TC, Caldwell CW and Skibba JL. (1995). Diag. Mol.
Path., 4, 6672.
Hung J, Kishimoto Y, Sugio K, Virmani A, McIntire D,
Minna J and Gazdar A. (1995). JAMA, 273, 558 ± 563.
Kishimoto Y, Sugio K, Hung J, Virmani A, McIntire D,
Minna J and Gazdar A. (1995). J. Natl. Cancer Inst., 87,
1224 ± 1229.
Kleihues P, Burger PC and Scheitnauer BW. (1993). J. Brain
Pathol., 3, 255 ± 268.
Lukeis R, Ball D, Irving L, Garson OM and Hasthorpe S.
(1993). Genes Chrom & Cancer, 8, 262 ± 269.
Merlens F, Johansson B, Hoglund M and Mitelman F.
(1997). Cancer Res., 57, 2765 ± 2780.
Minna J, Maneckjee R, D'Amico D, et al. (1991). Cold
Spring Harbor, NY: Cold Spring Harbor Laboratory
Press.
Nagatake M, Takagi Y, Osada H, Uchida K, Mitsudomi T,
Saji S, Shimokata K, Takahashi T and Takahashi T.
(1996a). Cancer Res., 56, 2718 ± 2720.
Nagatake M, Osada H, Kondo M, Uchida K, Nishio M,
Shimokata K, Takahashi T and Takahashi T. (1996b).
Cancer Res., 56, 1886 ± 1891.
Parker S, Tong T, Bolden S and Wingo P. (1997). Ca Cancer
J Clin., 47, 5 ± 27.
Rabbitts P, Douglas J, Daly M, Sundaresan V, Fox B,
Haselton P, Wells R, Albertson D, Waters J and Bergh J.
(1990). Genes Chrom & Cancer, 1, 95 ± 105.
Rojas K, Silverman GA, Hudson Jr JR and Overhauser J.
(1995). Genomics, 25, 329 ± 330.
Rasio D, Negrini M, Manenti G, Dragani T and Croce C.
(1995). Cancer Res., 55, 3988 ± 3991.
Rempel SA, Schwechheimer K, Davis RL, Cavenee WK and
Rosenblum ML. (1993). Cancer Res., 53, 2386 ± 2392.
Shiseki M, Kohno T, Adachi J-I, Okazaki T, Otsuka T,
Mizoguchi H, Noguchi M, Hirohashi S and Yokota J.
(1996). Genes Chrom & Cancer, 17, 71 ± 77.
Shridhar V, Siegfried J, Hunt J, del Mar Alonso M and Smith
DI. (1994). Cancer Res., 54, 2084 ± 2087.
Testa JR, Siegfried JM, Liu Z, Hunt JD, Feder MM, Litwin
S, Zhou J-Y, Taguchi T and Keller SM. (1994). Genes
Chrom & Cancer, 11, 178 ± 194.
Testa JR and Siegfried JM. (1992). Cancer Res., 52, 2702s ±
2706s.
Tran YK and Newsham IF. (1996). Cancer Res., 56, 2916 ±
2921.
Winqvist R, Hampton GM, Mannermaa A, Blanco G,
Alavaikko M, Kiviniemi H, Taskinen PJ, Evans GA,
Wright FA, Newsham I and Cavenee WK. (1995). Cancer
Res., 55, 2660 ± 2664.
Zhu J, Guo S-Z, Beggs AH, Maruyama T, Santarius T,
Dashner K, Olsen N, Wu JK and Black P. (1996).
Oncogene, 12, 1417 ± 1423.
3505