Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
00001.tif
(USC Thesis Other)
00001.tif
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
INFORMATION TO USERS
This manuscript has been reproduced from the microfilm master. UMI
films the text directly from the original or copy submitted. Thus, some
thesis and dissertation copies are in typewriter face, while others may be
from any type of computer printer.
The quality of this reproduction is dependent upon the quality of the
copy submitted. Broken or indistinct print, colored or poor quality
illustrations and photographs, print bleedthrough, substandard margins,
and improper alignment can adversely affect reproduction.
In the unlikely event that the author did not send UMI a complete
manuscript and there are missing pages, these will be noted. Also, if
unauthorized copyright material had to be removed, a note will indicate
the deletion.
Oversize materials (e.g., maps, drawings, charts) are reproduced by
sectioning the original, beginning at the upper left-hand comer and
continuing from left to right in equal sections with small overlaps. Each
original is also photographed in one exposure and is included in reduced
form at the back of the book.
Photographs included in the original manuscript have been reproduced
xerographically in this copy. Higher quality 6” x 9” black and white
photographic prints are available for any photographs or illustrations
appearing in this copy for an additional charge. Contact UMI directly to
order.
UMI
A Beil & Howell Information Company
300 North Zeeb Road, Ann Arbor MI 48106-1346 USA
313/761-4700 800/321-0600
INVESTIGATION OF NOVEL TUMOR SUPPRESSOR GENES AND
MECHANISMS OF TUMORIGENESIS IN HUMAN BLADDER CANCER
by
Mirella Gonzalez-Zulueta
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(Biochemistry and Molecular Biology)
December 1995
UMI Number: 9621614
UMI Microform 9621614
Copyright 1996, by UMI Company. All rights reserved.
This microform edition is protected against unauthorized
copying under Title 17, United States Code.
UMI
300 North Zeeb Road
Ann Arbor, MI 48103
UNIVERSITY OF SOUTHERN CALIFORNIA
THE GRADUATE SCHOOL
UNIVERSITY PARK
LOS ANGELES. CALIFORNIA 90007
This dissertation, written by
Mirella Gonzalez-Zulueta
under the direction of k.fZ Dissertation
Committee, and approved by all its members,
has been presented to and accepted by The
Graduate School in partial fulfillment of re
quirements for the degree of
DOCTOR OF PHILOSOPHY
Dean of Graduate Studies
Date S eptem ber 26. 19.95
DISSERTATION COMMITTEE
Chairperson
Mirella Gonzalez-Zulueta
Investigation of novel tumor suppressor genes and mechanisms of
tumorigenesis in human bladder cancer
Deletion mapping of 70 bladder transitional cell carcinomas (TCC) was
performed through microsatellite polymorphism analysis. The data expanded
previous studies and defined two putative tumor suppressor loci on the long arm of
chromosome 9: one mapped to 9q34.1 -qter, and the other mapped to 9pl3-q21.3. Six
low-stage tumors were shown to contain microsatellite instability, which suggested that
alterations in mismatch repair activity may occur early in TCC development.
The role of the pl6/C D K N 2 tumor suppressor gene in bladder TCC
tumorigenesis was investigated. Homozygous deletions and intragenic mutations in
pl6/CDKN2 were found in 54% of bladder tumor-derived cell lines but only in 20%
of uncultured TCCs. The significant difference between cell lines and primary tumors
in pl6/C D K N 2 alterations raised questions on the role of this gene in bladder
tumorigenesis. Subsequently, abnormal methylation of this gene was examined as an
alternative mechanism to inactivate pI6/C D K N 2 function in bladder TCC.
Methylation of the 5' CpG island of pl6/CDKN2 was found in 67% of TCCs and
correlated significantly with gene silencing, which suggested that abnormal
methylation of the 5' CpG island may be a more common mechanism to inactivate
pl6/CDKN2 in TCC. These findings have clear clinical implications since novel
therapeutic strategies can be designed in order to induce the expression of this tumor
suppressor in tumors cells that have inactivated the gene by DNA methylation.
Squamous cell carcinoma (SCC), the most common type of bladder cancer in
some parts of the world, was examined for genetic defects previously shown to be
involved in TCC tumorigenesis. Allelic loss of chromosome 9p and p!6/CDKN2
homozygous deletions and intragenic mutations were detected at a significantly higher
frequency in SCCs than in TCCs. The p53 gene mutation spectrum was also found to
differ in these two tum or types. These data led to the hypothesis that SCC
tumorigenesis follows a genetic alteration pathway distinct from TCC, which may
explain the different behaviors of these two types of bladder cancer.
To my beloved friend
ACKNOWLEDGEMENTS
I am very grateful to those who made this work possible through their
guidance, help and support Specially I would like to thank:
Dr. Peter A. Jones, my thesis advisor, who played a key role in my scientific
education through his teaching of the principles of good research and scientific
work, his constant guidance and support.
Dr. Jane W. Fountain, who gready contributed to this dissertation through
her many helpful ideas, patient manuscript review, and most importandy, her support
and guidance.
Dr. Richard Cote, who shared his knowledge, experience and time in many
helpful discussions.
Atsuko Shibata, Yvonne Tsai, Chuck Spruck, Christina Bender, Allen Yang,
Felicidad Gonzales, who contributed to the experimental performance of this work.
TABLE OF CONTENTS
%
CHAPTER 1: INTRODUCTION TO HUMAN BLADDER CANCER 1
INTRODUCTION I
CANCER AS A GENETIC DISEASE 4
EPIDEMIOLOGY OF BLADDER CANCER 5
FORMS OF BLADDER CANCER 8
Carcinoma in situ 8
Superficial transitional cell carcinoma 8
Muscle-infiltrative transitional cell carcinoma 8
Squamous cell carcinoma 13
STAGING OF BLADDER CANCER 13
TOOLS IN THE GENETIC ANALYSIS OF BLADDER CANCER 16
CHROMOSOMAL ALTERATIONS IN BLADDER TCC 19
TUMOR SUPPRESSOR GENES 21
Allelic losses in bladder cancer 21
Role of chromosome 9 in bladder cancer 25
Alterations in the p53 tumor suppressor gene 31
Alterations in the retinoblastoma gene 33
ONCOGENES 33
MICROSATELLITE INSTABILITY IN BLADDER CANCER 36
GENETIC PATHWAYS TO TCC OF THE BLADDER 37
MOLECULAR EPIDEMIOLOGY 40
CLONAL ORIGIN OF BLADDER CANCER 42
OVERVIEW OF THESIS RESEARCH 44
iv
CHAPTER 2: DELETION MAPPING OF CHROMOSOME 9 IN BLADDER
TRANSITIONAL CELL CARCINOMA USING
MICROSATELUTE POLYMORPHISMS: DEFINITION OF
TWO PUTATIVE TUMOR SUPPRESSOR LOCI
INTRODUCTION “ 47
MATERIALS AND METHODS 50
Tumor samples and DNA isolation 50
Determination of LOH on chromosome 9 by microsatellite analysis 51
RESULTS
Definition of a putative tumor suppressor locus on 9q34.1 -qter 54
Definition of a putative tumor suppressor locus on 9p 13-q21.3 59
Chromosome 9q deletion may be a primary event in superficial
papillary TCC development 59
DISCUSSION 61
CHAPTER 3: MICROSATELUTE INSTABILITY IN BLADDER CANCER
INTRODUCTION 65
MATERIALS AND METHODS 68
Tissue specimens and DNA isolation 68
Analysis of allelic loss on chromosomes 9 and 17p 71
RESULTS 74
DISCUSSION 81
CHAPTER 4: ALTERATIONS OF THE p!6/CDKN2 GENE IN BLADDER
TRANSITIONAL CELL CARCINOMA: HOMOZYGOUS
DELETIONS AND POINT MUTATIONS ARE INFREQUENT
MECHANISMS OF INACTIVATION
INTRODUCTION 86
MATERIALS AND METHODS 90
v
Tumor specimens, cell lines, and DNA isolation 90
Determination of authenticity of the LD cell line series 91
Analysis of point mutations in the pl6/CDKN2 gene 91
Analysis of homozygous deletions of the pl6/CDKN2 gene 94
Detection of homozygous deletions of pl6/CDKN2
m
in cell lines
Detection of homozygous deletions of pl6/CDKN2
in uncultured tumors
Analysis of mutations in the p53 tumor suppressor gene 96
Statistical analysis 98
RESULTS
pl6/CDKN2 homozygous deletions and intragenic point mutations
are frequent in bladder tumor-derived cell lines 100
pl6/CDKN2 homozygous deletions and intragenic point mutations
are infrequent in uncultured TCCs 1 O S
Comparative analysis with p53 gene status 108
DISCUSSION 110
CHAPTER 5: METHYLATION AND EXPRESSION OF THE p!6 AND p!5
CELL CYCLE REGUALTORS: DE NOVO METHYLATION OF
THE 5’ CpG ISLAND IS A FREQUENT MECHANISM OF
INACTIVATION OF THE p!6/CDKN2 TUMOR SUPPRESSOR
GENE IN TRANSITIONAL CELL CARCINOMAS OF THE
BLADDER
114
119
119
120
120
121
121
INTRODUCTION
MATERIALS AND METHODS
Tissue samples
Cell lines
5-Aza-2'-deoxycytidine treatment
Evaluation of cell population doubling time
Reverse transcription (RT)-PoIymerase chain reaction (PCR)
PCR-based methyaltion assay
RESULTS
123
128
Analysis of CpG distribution in exons 1 and 2 of the pl6/CDKN2
and p l5 ,N K 4B genes 128
Methylation and expression of the pI6/CDKN2 and pl5IN K 4B
genes in primary bladder transitional cell carcinomas and cell lines 132
Methylation and expression of the pI6/CDKN2 and plS™*48
genes in colon carcinomas and adjacent normal mucosa 136
Methylation and expression of the pl6/CDKN2 and pl5IN K 4B
genes in non-neoplastic tissue 148
Methylation of exon 1, but not of exon 2, of pl6/CDKN2 is
associated with transcriptional silencing 1 S1
Concordant methylation in both exon 2 CpG islands of
p!6/CDKN2 and p!5IN K 4B 151
Induction of pl6/CDKN2 expression by the demethylating agent
5-Aza-CdR in vitro 154
Activation of p!6/CDKN2 expression by 5-Aza-CdR is associated
with inhibition of tumor cell growth in vitro 156
Induction of pl6/CDKN2 expression by 5-Aza-CdR in vivo 160
DISCUSSION 163
CHAPTER 6: GENETIC ANALYSIS OF SQUAMOUS CELL CARCINOMA OF
THE BLADDER: HIGH FREQUENCY OF CHROMOSOME 9p
ALLELIC LOSS AND p!6/C D K N2 ALTERATIONS
DISTINGUISH IT FROM TRANSITIONAL CELL CARCINOMA
INTRODUCTION 171
174
174
176
178
180
MATERIALS AND METHODS
Tumor samples and DNA isolation
PCR amplification and deletion analysis of pI6/CDKN2
SSCP analysis and sequencing of pI6/CDKN2
Determination of allelic loss
vii
SSCP analysis and sequencing of p53 gene 181
Immunohistochemical analysis of the p53 protein 181
Statistical analysis 182
RESULTS 182
Homozygous deletions of the pl6/CDKN2 locus 182
Mutational analysis of the pl6/CDKN2 gene 188
Allelic loss of chromosome 9 191
Sequence analysis of the p53 gene and allelic loss of chromosome 17p 192
DISCUSSION 196
CHAPTER 7: SUMMARY AND CONCLUSIONS 206
REFERENCES 213
APPENDIX: LIST OF ABBREVIATIONS 238
viii
LIST OF FIGURES
Figure Page
Fig. 1.1. Proportion of SCC in bladder cancer in different parts
of the world 2
Fig. 1.2. Carcinoma in situ 9
Fig. 1.3. Superficial papillary TCC 10
Fig. 1.4. Muscle-ifiltrative TCC 11
Fig. 1.5. SCC of the bladder 14
Fig. 1.6. Staging system of bladder cancer 15
Fig. 1.7. Allelotype of bladder TCC 22
Fig. 1.8. Frequency of LOH on chromosome .9 and
p53 mutations in bladder TCCs 26
Fig. 1.9. Model for the genetic pathways in the development
and progression of bladder TCC 39
Fig. 2.1. Partial allelic losses on chromosome 9 56
Fig. 2.2. Microsatellite analysis of bladder tumor with
subchromosomal deletions in chromosome 9 60
Fig. 3.1. Microsatellite instability in bladder TCC 76
Fig. 3.2. Instability at a trinucleotide repeat in the androgen
receptor gene 78
Fig. 4.1. Mechanism of action of pl6 89
Fig. 4.2. Verification of the origin of the LD cell line series 92
Fig. 4.3. Optimization of PCR-based assay for detection of
pI6/CDKN2 homozygous deletions 97
Fig. 4.4. pl6/CDKN2 mutations and homozygous deletions
in bladder tumor-derived cell lines 99
Fig. 4.5. p!6/CDKN2 homozygous deletions in uncultured TCCs 101
ix
Fig. 4.6. Confirmation of pl6/CDKN2 homozygous deletions
by Southern blot analysis 102
Fig. 5.1. Genomic organization of the p!6/CDKN2 and p i5INK4B
on chromosome 9p21 115
Fig. 5.2. Maps of the p!6/CDKN2 and plSlNK4B genes 122
Fig. 5.3. Comparison of the exon 2 sequences in pl6/CDKN2
and p!5MK4B
125
Fig. 5.4. Optimization of the PCR-based methylation assay 127
Fig. 5.5. Analysis of the CpG distribution in pl6/CDKN2 129
Fig. 5.6. Methylation and expression of pl6/CDKN2 and p!5INK4B
in normal bladder mucosa and bladder tumor 133
Fig. 5.7. Methylation status of exon 2 of p!6/CDKN2 in
bladder TCCs 134
Fig. 5.8. Methylation status of exon 1 of p!6/CDKN2 in colon
carcinomas and adjacent normal mucosa 137
Fig. 5.9. RT-PCR expression analysis of p!6/CDKN2 and
p!5INK4B in colon cancer patients 139
Fig. 5.10. Expression and methylation of p!6/CDKN2 and
p!5INK4B in colon cancer patients 142
Fig. 5.11. Mutation analysis of pl6/CDKN2 in colon carcinomas 145
Fig. 5.12. Analysis of CDK4 expression in colon carcinomas 146
Fig. 5.13. Expression and methylation of p!6/CDKN2 and
pl5lNK4B in non-neoplastic tissues
149
Fig. 5.14. Effect of 5-Aza-CdR treatment on the expression of
p!6/CDKN2 in cell lines 155
Fig. 5.15. Effect of 5-Aza-CdR treatment on the methylation status
of p!6/CDKN2 in cell lines 157
Fig. 5.16. Cell growth and doubling time of EJ and J82 cells
before and after 5-Aza-CdR treatment 159
X
Fig. 5. 17. Induction of pl6/CDKN2 expression in EJ cells
correlates with decreased cell growth rate 161
Fig. 5.18. Induction of pl6/CDKN2 expression in vivo by
5-Aza-CdR treatment 165
Fig. 6.1. Squamous cell carcinoma associated with
Schistosoma infection 175
Fig. 6.2. Optimization of AR amplifications in SCCs 184
Fig. 6.3. Detection of pl6/CDKN2 homozygous deletions
by comparative multiplex PCR 185
Fig. 6.4. Extensive amplification of pl6/CDKN2 to investigate
the presence of normal contaminating cells in the tumor
specimen 187
Fig. 6.5. Microsatellite analysis of SCCs 189
Fig. 6.6. Deletion map of chromosome 9 in SCCs 190
Fig. 6.7. Point mutations in the p53 gene in SCCs 195
Fig. 6.8. Spectra of p53 mutations in bladder cancer 197
Fig. 6.9. IHC analysis of p53 in SCC 198
Fig. 6.10. Comparison of genetic alterations observed in
SCC and TCC of the bladder 200
Fig. 6.11. Proposed model for the molecular progression
of bladder cancer 204
xi
LIST OF TABLES
Table Page
«
Table 1.1. Regions of subchromosomal deletions in chromosome 9
detected in bladder TCCs 28
Table 1.2. Tumor suppressor genes altered in bladder TCCs 30
Table 1.3. Oncogenes altered in bladder TCCs 33
Table 2.1. Microsatellite markers on chromosome 9 analyzed
in this study S3
Table 2.2. Summary of chromosome 9 LOH analysis in 70 TCCs
of the bladder 33
Table 3.1. Stage and grade of 154 TCCs examined 70
Table 3.2. Microsatellite alterations in TCCs of the bladder 72
Table 3.3. Bladder TCCs showing microsatellite alterations 82
Table 4.1. p!6/CDKN2 homozygous deletions and intragenic
mutations in primary TCCs and derived cell lines 103
Table 4.2. Status of the p!6/CDKN2 and p53 genes in
bladder cancer-derived cell lines 104
Table 4.3. Stages of uncultured TCCs containing
pl6/CDKN2 homozygous deletions 106
Table 4.4 Chromosome 9 allelic loss and status of
the pI6/CDKN2 and p53 genes in uncultured TCCs 107
Table 5.1. Analysis of the CpG distribution in pl6/CDKN2
and p lS INK4B 131
i
Table 5.2. De novo methylation of pI6/CDKN2 in primary
bladder TCCs and bladder tumor-derived cell lines 135
Table 5.3. De novo methylation of pl6/CDKN2 in colon cancer 147
Table 5.4. Methylation of exon 1 and exon 2 and expression of
p!6/CDKN2 152
xii
Table 5.5. Comparison of methylation status of exon 2 of
pl6/CDKN2 and pl5INK4* 153
Table 5.6. pl6/CDKN2 expression in EJ tumors of nude rats treated
with 5-Aza-CdR 164
Table 6.1 Status of the pl6/CDKN2 and p53 genes, and allelic loss
of 9p, 9q, and 17p in SCCs 193
xiii
CHAPTER 1
INTRODUCTION TO HUMAN BLADDER CANCER
INTRODUCTION
Bladder cancer is the sixth most frequent neoplasia worldwide (Silverberg and
Boring, 1989; Silverberg and Boring, 1990), affecting predominantly males with a
male:female ratio of approximately 3:1. Epidemiologic studies have revealed the
association of specific risk factors and urothelial tumors, cigarette smoking being the
most important in the Western Hemisphere (Augustine et al., 1988; Talaska et al.,
1991; Spruck et al., 1993). Other risks include infection by the trematode
Schistosoma haematobium (Mostofi et al., 1988) and occupational exposure to certain
chemicals and their metabolites excreted through the urinary tract (Matanoski and
Elliot, 1981), as well as exposure to therapeutic agents, such as phenacetin (Thompson
and Fair, 1989).
Approximately 90% of malignant tumors arising in the urinary bladder in the
United States are transitional cell carcinomas (TCCs), and only about 5% are
classified as squamous cell carcinomas (SCCs) (Mostofi et al., 1973) (Fig. 1.1). In
contrast, in Egypt and regions of the Middle East and Africa, SCCs account for 80%
of all bladder tumors (Mostofi et al., 1988) and have been correlated with infection by
the trematode Schistosoma haematobium.
1
Western Hemisphere Endemic Schistosomiasis Areas
Invasive
TCC 22%
SCC 7%
Superficial Papillary
TCC 63%
TCC 20%
SCC80%
Fig. 1.1. Proportion of SCC in bladder cancer in different parts of the world. While in the Western Hemisphere
SCC represents only 7% of all bladder cancers, in areas of the Middle East and Africa, where infection by the
trematode Schistosoma haematobium is endemic, 80% of bladder cancers are SCC (Mostofi et al., 1988).
Transitional cell carcinoma of the bladder is not a single disease. Based on
morphological evaluation and natural history, TCCs may present in three forms with
distinct behaviors and prognosis: the low-grade, well-differentiated, superficial and
usually papillary tumors, the high-grade, poorly-differentiated, invasive tumors, and
carcinoma in situ (Cis). The first group, including stage Ta and T1 tumors, accounts
for 80% of TCC, while the remaining 20% are stage T2, T3, and T4 lesions at the time
of presentation (Prout, 1977). Up to 80% of patients with a superficial tumor will
have one or more recurrences after initial treatment, and in 10-30% the tumor will
progress to invasive disease (Greene et al., 1973). The ability to predict the clinical
outcome of these patients is hampered by the fact that two morphologically identical
superficial TCCs may behave in different fashions. Thus, new methods and strategies
are needed to identify those patients with superficial tumors who are likely to develop
invasive carcinoma. Cis, although confined to the mucosa, has a worse prognosis than
superficial papillary carcinoma. Without treatment, nearly 80% of patients with Cis
will develop muscle invasion or progress directly to metastatic disease (Utz et al.,
1970).
The variable clinical course of TCC of the bladder suggests that there are
multiple molecular and genetic pathways that lead to the development of this disease.
Although the clinical management of bladder cancer has improved, the mortality rate
has not changed notably in the last S O years. Identification of the genetic changes that
precede the phenotypic alterations in a progressively neoplastic urothelium is essential
to develop early detection strategies and more novel and effective treatments of bladder
tumors according to the nature of their molecular alterations.
3
CANCER AS A GENETIC DISEASE
Cell growth and differentiation are two fundamental aspects of multicellular
existence, and intertwined, with these processes is the phenomenon of unlimited
growth, which is the basis of the neoplastic state. Cancer is believed to result from the
unlimited growth of a given cell which is often due to a block in the ability of cells to
undergo differentiation and/or apoptosis. The genesis of neoplastic diseases is
characterized by its multistep nature, because the cells of a tumor mass that derive
from a single mutated stem cell can acquire subsequent genetic alterations that provide
selective growth advantages. The accumulation of genetic damage results in increased
cell proliferation and tumorigenesis. Multiple mutations appear to be required to
endow a malignant phenotype, and the accumulaton of these events gives neoplastic
cells the ability for tumor progression. (Bishop, 1987; Fearon and Vogelstein, 1990;
Vogelstein and Kinzler, 1993).
Target genes implicated in cellular transformation and tumor progression
belong to two categories: proto-oncogenes and tumor suppressor genes (or growth
suppressor genes). Proto-oncogenes become oncogenes when they are abnormally
activated by either point mutation, amplification, translocation, or even by insertion of
non-eukaryotic sequences such as viral genomes (Solomom et al., 1991). A genetic
alteration in a proto-oncogene resulting in a gain of function usually leads to
uncontrolled cell proliferation. Whereas proto-oncogenes require activating events,
growth suppressor genes need to be inactivated to cause a tumorigenic phenotype
(Commings, 1973). According to Knudson's "two hit" hypothesis (Knudson, 1985),
inactivation of tumor suppressor genes requires functional loss of both copies of the
4
gene which occurs mainly, but not only, through the loss of one allele coupled to
mutation of the remaining allele. Alterations in proto-oncogenes and tumor suppressor
genes seem equally prevalent among human cancers (Bishop, 1987).
EPIDEMIOLOGY OF BLADDER CANCER
The outstanding epidemiologic feature of bladder cancer is the very high rate
in white men. There is a marked excess of bladder cancer in men compared to women
for all racial groups but especially for whites, for which the sex ratio is almost 4:1
(Rossetal., 1988).
Epidemiologists have estimated that more than 50% of bladder cancer cases in
U.S. men and more than 35% of cases in U.S. women are due to either cigarette
smoking or industrial exposure to carcinogens (Cole, 1973). Some studies estimate
that perhaps 55% of all bladder cancer deaths are due to cigarette smoking alone
(Ross et al., 1988). If true, one might expect the incidence of bladder cancer to be
increasing due to increased used of cigarettes and greater industrial pollution since the
turn of the century. The first epidemiologic study of an association between cigarette
smoking and bladder cancer was conducted in 1956 by Lilienfeld (Lilienfeld, 1956).
Most case-control studies have reported a relative risk of about 2 for cigarette smokers
compared to non-smokers, and have observed higher relative risks with increasing
amounts smoked (Ross et al., 1988). The mechanism by which cigarettes induce
bladder cancer is unclear even though cigarette smoke is known to contain numerous
carcinogens. Mice receiving intraoral tobacco tars have a high likelihood of
developing bladder papillomas and carcinomas, and some cyclic N-nitrosamines
derived from tobacco alkaloids are bladder specific carcinogens in animals (Hicks,
1982). Aromatic amines, including the human bladder carcinogen P-naphthylamine,
are also present in tobacco smoke (Holsti and Ermala, 1955).
The class of chemical carcinogens most strongly related to bladder cancer is
the arylamines. Workers in the rubber tire industry have been proposed to have a
high risk of bladder cancer (reviewed in Ross et al., 1988). Beta-naphthylamine has
been used as an antioxidant in the rubber industry. Benzidine, a chemically related
compound widely used in the rubber industry, has now been established as a human
bladder carcinogen as well. In addition to the textile dye and rubber industries,
arylamines have been used in hair dyes and paint pigments. Accordingly, hair
dressers and painters have been shown by some studies to be at high risk of bladder
cancer (Cole, 1973).
In certain geographic areas such as Egypt, North Africa and the Middle East,
bladder cancer is associated with infection by the trematode Schistosoma
haematobium. In those regions, where schistosomiasis is endemic, the risk of
developing bladder cancer is increased three fold (Mostofi, 1991). In fact, bladder
cancer is the most common cancer in men in those areas. The bladder tumors
associated with schistosomiasis are squamous cell carcinomas (SCC), which represent
80% of bladder cancer in bilharzial areas (Fig. 1.1). This type of bladder tumor is
infrequent in the Western Hemisphere and usually associated with chronic infection,
lithiasis and chronic catherization which induce chronic irritation of the urotheliun
(Badawi et al., 1988). The Shistosoma haematobium penetrates the skin of the
human host in its cercariae form, and the adult form migrates to the bladder wall where
eggs are deposited, causing constant irritation to the urothelium (Badawi et al., 1991).
Some investigators have proposed that such increased cell proliferation may potentiate
the mutagenic effect of carcinogens such as N-nitroso compounds, which have been
demonstrated to accumulate in schistosomiasis affected bladders (Badawi et al., 1991).
Studies aimed towards a better understanding of the genetic alterations that underlay
squamous cell carcinoma are described in Chapter 6 of this thesis.
A ten fold increase in bladder cancer incidence is observed in the southwest
region o f Taiwan, where the peripheral vascular disorder known as the "black foot
disease" is endemic. High levels of arsenic in the well water have been implicated in
the genesis of this disease, which is associated with increased incidence of liver and
bladder cancers (Chen et al., 1985; Wu et al., 1989). Arsenic may induce irritation of
the urothelium and potentiate the effect of other environmental carcinogens (Shibata et
al., 1994).
A variety of other potential etiologic factors have been linked to bladder cancer,
but for most of them supportive epidemiologic evidence is either minimal or absent.
Some of those factors include heavy exposure to phenacetin-containing analgesics,
caffeine exposure, and use of artificial sweeteners.
7
FORMS OF BLADDER CANCER
Carcinoma in situ (Cis)
This type of bladder cancer was first described by Melicow (1952) as specific
segments of flat epithelium in tumor-bearing bladders that contained neoplastic cells
but that were separate from the primary tumors (Fig. 1.2). He suggested that these
areas were a distinct cancer entity that represented the earliest stage of bladder cancer.
Cis is a particularly aggressive malignancy since 75 to 80% of patients develop
invasive disease, and 50 to 75% of patients with Cis develop metastases (Daly, 1976).
Superficial transitional cell carcinoma
The majority of patients with bladder cancer in the Western Hemisphere
present with superficial papillary disease (Cutler, 1982). The term "superficial" is
applied to those tumors that are confined to the mucosa (Fig. 1.3). In contrast to
patients with Cis, those with initially papillary tumors confined to the mucosa suffer
muscle invasion in only about 10 to 30% of the cases (Cutler, 1982). However,
superficial papillary tumors recur frequently; about two-thirds of patients who initially
have a single tumor experience recurrence, and 90% of patients who initially have
multiple tumors develop one or more recurrences after initial treatment (Cutler, 1982).
Musde-infiltrative transitional ceil carcinoma
Transitional cell carcinomas that have infiltrated the bladder wall musculature
(Fig. 1.4) have an ominous prognosis, and despite radical therapy 50% of the patients
8
Fig. 1.2. Carcinoma in situ. The appearance of epithelial cells with
irregularly shaped nuclei, and loss of polarity from the basal layers to the
upper layers indicates intraepithelial neoplasia.
9
Fig. 1.3. Superficial papillary transitional cell carcinoma. Cells in
these lesions are confined to the mucosa and generally
homogeneous.
10
Fig. 1.4. Muscle-infiltrative transitional cell carcinoma.
11
develop metastases within 1.5 to 2 years. Even if treated aggressively, ultimate
mortality from cancer has been shown to be greater than 80% within 5 years of
therapy (Lieskovsky et al., 1988).
Major questions in muscle-invasive bladder cancer have focused on the
pathway by which a particular tumor may have progressed. Only 10 to 30% of
initially superficial tumors have been found to progress to rnuscle invasion.
Moreover, the majority of patients with muscle-invasive disease have been found to
have this more advanced stage of tumor at their initial clinical presentation
(Kaye, 1982). Two hypotheses have been develop to explain such observations. The
first postulates that invasive tumors develop sequentially, but rapidly, from superficial
disease, penetrate the bladder wall prior to development of clinically recognizable
symptoms, and metastasize early in the phase of infiltration (Friedell, 1982). The
second hypothesis suggests that the majority of these tumors originate from cells that
follow a separate developmental pathway in which rapid infiltration of the bladder
musculature is the predominant event whereas profusion into the bladder lumen occurs
only later in the life of the tumor (Droller, 1982). Some muscle-invasive tumors are
described as papillary with infiltration in a "broas front" or cohesive mass of cells
(Lieskovsky et al., 1988). Others tumors are described as solid or nodular and appear
to penetrate the bladder wall in a tentacular pattern, in which clusters of tumor cells
extent in finger-like projections (Lieskovsky et al., 1988). These differences between
muscle-invasive tumors may reflect variant pathways of tumor development, the one
characterized predominantly by hyperplasia and cellular proliferation (with possibly
less aggressive behavior) and the other characterized predominantly by dysplasia with
a greater likelihood of infiltration of the bladder wall and distant dissemination. These
12
differences in morphologies and behavior may reflect underlying distinct genetic
alterations in different bladder tumor types.
Squamous cell carcinoma (SCC)
This type of bladder tumor is morphologically very different from TCC,
composed of cells of non-cuboidal shape, and characterized by the presence of large
keratin deposits (Fig. 1.5). SCC is frequent and associated with infection by the
trematode Schistosoma haematobium in regions of Africa and the Middle East,
whereas it is infrequent and associated with long-standing chronic irritation of the
urothelium in the Western Hemisphere. SCC also differs from TCC clinically, since
most of the SCCs have invaded the bladder musculature at the time of diagnosis and
usually have poor prognosis. Most of the studies on the genetic alterations that occur
in bladder cancer have focused exclusively on TCC, and little is known about the
genetic defects involved in SCC tumorigenesis.
STAGING OF BLADDER CANCER
Differences in the behavioral patterns of bladder tumors have led to the
establishment of staging systems to provide an index to prognosis and assist in the
determination of appropriate therapy. The most recent of these systems, the TNM
system, classifies tumors according to the depth of penetration (T) through the bladder
wall at initial presentation, the involvement of lymph nodes (N), and the presence of
distal metastasis. This system (Fig. 1.6) includes distinctions between carcinoma in
13
Fig. 1.5. Squamous cell carcinoma. This type of bladder tumor is
composed of non-cuboidal shaped cells and deposits of keratin.
1 4
Urothelium
Lamina Propria l —~ —
Muaculam Propria
Porivealcal Tissue
T2 Tja bb
Fig. 1.6. Staging system of bladder cancer. A clinically relevant
staging system was derived from observations of the clinical course of
different tumors in association with the depth of penetration through
the bladder wall at initial presentation.
15
situ (Cis or Tis), mucosa confined papillary tumors (Ta), tumors that have invaded the
lamina propria (Tl), tumors that have penetrated the superficial and deep muscle of the
bladder wall (T2-3), and tumors that have invaded the contiguous viscera (T4)
(Lieskovsky etal., 1988). •
TOOLS IN THE GENETIC ANALYSIS OF BLADDER CANCER
The genetic analysis of bladder cancer has advanced through the application of
innovative techniques in molecular biology. Gross chromosomal abnormalities (such
as duplications, amplifications, translocations, monosomies, trisomies, and deletions)
can be detected by cytogenetic studies, which include karyotyping, flow cytometry,
and fluorescence in situ hybridization (FISH). Karyotyping is an approach
extensively used that has provided interesting results; however, this technique has
several drawbacks. First of all, it is generally possible only after tissue culturing,
which has been reported to produce a selective growth of tumor cells with high mitotic
index and loss of chromosomal material (Hopman et al., 1988). Furthermore, these
studies are often impeded by the small number of metaphases, poor banding quality,
and a condensed appearance of chromosomes in general (Hopman et al., 1991). Flow
cytometry analysis is used to establish the ploidy status of tumor cells, which is
frequently altered in cancer. FISH is a very useful method for chromosome analysis,
as it allows detection of chromosomal alterations not detectable by karyotyping, and
dividing cells are not required for this technique (Hopman et al., 1990).
16
Mutations in oncogenes and tumor suppressor genes can be rapidly screened
for by analysis of single strand conformation polymorphisms (SSCP) (Orita et al.,
1989). This method is based upon the differences in the conformations that a wild
type and a mutant single stranded DNA molecule adopt. One nucleotide change is
enough to cause a distinct conformation. After a first step screening by SSCP, the
nature and exact position of the mutations is usually determined by direct DNA
sequencing (Sambrook et al., 1989).
A different approach to screen for genetic mutations includes
immunohistochemistry analysis (IHC) of gene products (Esrig et al., 1994). In
contrast to the techniques described so far, which allow direct analysis of DNA, IHC
detects proteins on histologic sections. Through the use of specific antibodies, IHC
permits the identification of normal and altered proteins.
Chromosomal deletions and allelic loss of tumor suppressor genes can be
identified by two main strategies. Classical studies have used restriction fragment
length polymorphisms (RFLP) analysis, which takes advantage of the presence of
sites for restriction enzymes in the genome whose locations may vary within a given
area in different individuals. Two alleles at a specific locus with different restriction
fragment lengths render that locus heterozygous. The loss of one of the alleles is
detected by the absence of a DNA signal corresponding to the lost allele at a
heterozygous locus, and is described as loss of heterozygosity (LOH). A different
approach, which has become the most useful strategy to detect LOH, is based on
amplification of DNA by the polymerase chain reaction (PCR). It allows for the
distinction between two alleles on the basis of their polymorphic microsatellites.
17
Microsatellites are DNA sequences containing simple repeats that may be either
mononucleotide, dinucleotide, trinucleotide, or tetranucleotide repeats. The number of
repeats at a given microsatellite is varible among individuals, i.e., the microsatellite is
polymorphic. This technique has the advantages of requiring very small amount of
DNA, and allowing a rapid screening of large number of samples.
Early attempts to map small areas of deletion on chromosome 9 in bladder
cancer were hampered by the inadequate coverage of the chromosome by highly
informative polymorphic markers. In particular, studies of the long arm of
chromosome 9 were largely restricted to the distal region (9q34) where several
variable number tandem repeat (VNTR) markers had been mapped (Cairns et al.,
1993; Miyao et al., 1993). Deletion mapping requires the use of highly informative
markers covering all regions of the chromosome. Microsatellite polymorphisms have
proven very useful in the analysis of LOH at chromosomal regions in bladder and
other cancers. The application of PCR to the genetic analysis of bladder cancer, and
cancer in general, has had a deep impact in our understanding of the disease as it
allows for the analysis of minute amounts of DNA from various sources such as fresh
frozen tumor specimens, paraffin-embedded histologic sections from archival tissue
(Shibata et al., 1992), or urine samples containing normal and tumor cells from the
urothelium (Sidransky et al., 1991). Deletion mapping of chromosome 9 by analysis
of dinucleotide repeat polymorphisms is described in Chapter 2 of this thesis.
1 8
CHROMOSOMAL ALTERATIONS IN BLADDER TRANSITIONAL CELL
CARCINOMAS
The concept that changes in chromosomes play a role in the development of
cancer was first suggested in 1890 by von Hansemann (1890), and further refined by
Boveri (1929) who noted that changes in gene function might cause neoplasia.
Boveri's view of the genetics of cancer has been proven to be essentially correct.
Cytogenetic (Gibas et al., 1984; Atkin et al., 1985; Sandberg et al., 1986;
Gibas et al., 1986; Berger et al., 1986; Smeets et al., 1987; Babu et al., 1989) and
molecular analysis from Dr. Jones' group and others (Fearon et al., 1985; Tsai et al.,
1990; Knowles et al., 1994) have shown nonrandom chromosomal alterations in
bladder cancer. The importance of the detection of these chromosomal changes is that
they may reflect activation of oncogenes and inactivation of tumor suppressor genes
that could lead to causation, progression and metastasis of bladder cancer.
Atkin and colleagues (1985) reported abnormalities on chromosomes 1 and 1 1
in 10 bladder TCCs examined. Chromosome 1 changes have been linked to many
cancer types (Atkin et al., 1985), and anomalies in the short arm of chromosome 1 1
(1 lp) have been associated with rhabdomyosarcoma (Scrable et al., 1989), ovarian
carcinoma (Ehlen et al., 1990) and Wilms' tumor (Kaneko et al., 1981). Deletion of
1 lp has been considered to be related to invasive bladder TCCs (Gibas et al., 1984).
In a survey based on a cohort of 15 bladder tumors, Babu et al. found 3p duplication,
trisomy of chromosome 7, and 1 lp deletions as common events in urothelial cancer
(Babu et al., 1987).
19
Isochromosome 5p has been reported in bladder cancer (Gibas et al., 1984;
Gibas et al., 1986). Isochromosomes result from a transverse rather than a
longitudinal division of the centromere during mitosis. Studies with three superficial
bladder TCCs and five -poorly differentiated invasive tumors showed that
isochromosome 5p is a common chromosomal aberration in bladder TCCs (Gibas et
al., 1986). Gibas et al. reported that in a study of 18 tumors from 16 patients
isochromosome 5p was found in 40% of TCCs (Gibas et al., 1986). However,
Smeets et al. (1987) reported that no abnormalities were observed in chromosome 5 in
13 primary TCCs of the bladder.
Monosomy of chromosome 9, i.e., complete loss of one of the two
chromosomes 9, has been reported as the most frequent genetic alteration in bladder
cancer. A total of 5 out of 10 TCCs examined were found to have monosomy 9 in
studies by Gibas (1984), and Vanni and Scarpa (1986). This was the only karyotypic
change in 2 of the cases. Smeets et al. (1987) also identified monosomy of
chromosome 9 in 4 of 8 bladder tumors. Most of the above reported studies were
performed combining superficial and muscle invasive tumors. Gibas et al. (1986)
observed no changes in chromosome 9 in advanced poorly differentiated tumors,
concluding from their results that chromosome 9 changes might be more prevalent in
low-stage, low-grade bladder tumors. In a more recent study, Tyrkus et al. (1992)
karyotyped 17 carcinomas in situ, which are characterized by their aggressive
behavior, and found no chromosome 9 alterations.
20
TUMOR SUPPRESSOR GENES
In 1973 Commings suggested the existance of structural genes active during
embryogenesis but suppressed during cell differentiation by suppressor regulatory
genes (Commings, 1973). Functional loss of both copies of these suppressor genes
may lift the suppression on cellular proliferation, resulting in development of a tumor.
Loss of function of a growth suppressor gene usually requires the alteration of both
copies of the gene, which commonly occurs through point mutations in one of the
alleles, coupled with loss of the other allele. The investigation of nonrandom allelic
losses associated with bladder cancer is important, as these events presumably reflect
the sites of tumor suppressor genes.
Allelic losses in bladder TCC
Loss of heterozygosity (LOH) in chromosomal arms in bladder cancer have
been detected in numerous studies by cytogenetic, RFLP, and microsatellite
polymorphism analyses (Fearon et al., 1985; Tsai et al., 1990; Olumi et al., 1990;
Dalbagni et al., 1993; Habuchi et al., 1993a; Knowles et al., 1993; Knowles et al.,
1994). The results from those studies are compiled in the allelotype shown in Figure
1.7, which is based mainly upon data published by Knowles and colleagues (1994).
LOH has been detected in nearly every single chromosome in at least one bladder
tumor. However, as Figure 1.7 clearly shows, LOH is most commonly detected in
chromosomes 9q, 9p, 17p, Up, 8p, 4p, 3p, 13q, 18q, and 17q. Loss of the Y
chromosome has also been reported as a common event in bladder TCCs, being
2 1
70 -
□ p-arm
■ q-arm
6 0 -
5 0 -
ft
3 4 0 -
u
J
U
J 3 0 -
<
2 0 -
1 0 -
1 2 3 4 S 6 7 8 9 10 1 1 12 13 14 15 16 17 18 19 20 21 22
CHROMOSOME#
Fig. 1.7. Allelotype of bladder TCC showing the frequency of allele loss on each
chromosome arm. This allelotype originally from Knowles et al. (1994) and expanded by
Yang and Jones (1994) has been updated to include other studies (Dalbagni et al., 1993;
KeenandKnowles, 1994).
22
associated with advanced stage, Cis, and poor prognosis (Tyrkus et al., 1992;
Sandberg et al., 1994). It is necessary to determine which genes that may play a role
in bladder tumorigenesis are located on these frequendy deleted regions.
Allelic deletions in bladder TCCs are most frequendy detected in chromosome
9, with an overall frequency of 55%, independent of tumor stage and grade. The role
of chromosome 9 in bladder cancer is discussed in the next section. Deletion
mapping aimed towards the identification of putative tumor suppressor genes on
chromosome 9 important in bladder tumorigenesis is described in Chapter 2 of this
thesis.
LOH in the short arm of chromosome 17 most likely represents the
inactivation of the p53 tumor suppressor gene, located on locus D17S153. In a study
from Dr. Jones' group, Olumi et al., (1990) detected LOH in 17p in 63% of the
invasive bladder TCCs evaluated by RFLP analysis, but none of the low-grade
superficial tumors showed loss of this chromosome. Similar results were obtained by
other groups. Presti et al. (1991) examined 34 TCCs and also found a correlation
between 17p deletions and advanced stage and grade. These findings support the
involvement of 17p in the progression of bladder cancer (Sidransky et al., 1991;
Dalbagni et al., 1993; Habuchi et al., 1993a; Presti et al., 1991; Spruck et al., 1994a).
LOH in the long arm of chromosome 13, including the retinoblastoma gene
locus on 13ql4, was reported by Ishikawa et al. (1991) in 3 of 18 (16%) informative
bladder tumors analyzed, and Cairns etal. (1991) found 28 of 94 (30%) informative
23
cases with LOH at the retinoblastoma gene locus, 26 of these 28 lesions being muscle
invasive tumors.
Chromosome 1 lp is-deleted in approximately 40% of bladder TCCs (Fearon
et al., 1985; Tsai et al., 1990). The tumor suppressor gene associated with Wilms’
tumor (WT1 gene) is located in the short arm of chromosome 11 (Call et al., 1990),
although the WT1 gene has not yet been shown to be inactivated in bladder cancer. In
an attempt to examine any possible correlation of lip LOH with bladder tumor
progression, Dalbagni et al. (1993) analyzed 60 paired bladder tumors and normal
tissue with polymorphic DNA markers and found no association between 1 lp LOH
and the various parameters of poor outcome, including histologic grade, stage, vascular
invasion and lymph node metastasis. Similar results were reported by Presti et al.
(1991) and Habuchi et al. (1993a), who detected 1 lp LOH in 7 of 24 (29%) and 20
of 44 (45%) informative cases, respectively. It appears that such an alteration is not
associated with progression of bladder cancer.
Allelic loss of the short arm of chromosome 3 was not detected in any of the
informative superficial papillary lesions in a study by Presti and colleagues (1991); in
contrast, 18 of 33 (54%) invasive tumors had such alteration. There was a statistically
significant association between the pathological parameters of poor outcome and 3p
LOH. A putative tumor suppressor gene on 3p has been associated with other
genitourinary malignancies, such as renal cell carcinoma (Presti et al., 1991), as well as
with lung cancer (Mori et al., 1989).
24
Role of chromosome 9 in bladder cancer
Cytogenetic, RFLP, and microsatellite analyses have revealed that loss of all or
part of chromosome 9 is the most common chromosomal alteration in bladder cancer
(Fig. 1.7.). Figure 1.8 shows the frequency of 9q LOH reported in bladder TCCs of
different histologic stages. Altogether, 319 out of 580 (55%) bladder TCCs reported
in the literature have shown LOH of chromosome 9. This genetic event is frequent in
all stages of bladder TCC, with the exception of Cis, and chromosome 9 monosomy is
the only recognizable chromosomal change within low-grade, low-stage bladder
tumors in several cases (Freidell et al., 1986; Olumi et al., 1990).
Dr. Jones’ group (Tsai et al., 1990) first reported LOH of chromosome 9q in
67% of 25 advanced stage tumors examined by RFLP analysis, and Olumi et al.
(1990) subsequently detected this event, also by RFLP analysis, in all stages and
grades of bladder TCCs. The LOH of chromosome 9q in low-stage, low-grade
tumors is particularly important, as alterations of other chromosomes occur mainly in
high-stage tumors. Deletions of 9q are also specific for bladder cancer, unlike the
other chromosomal changes which are found in many tumor types. Inactivation of a
tumor suppressor gene on the long arm of chromosome 9 is, therefore, a candidate
initiating event in bladder tumorigenesis. Work described in Chapter 2 of this thesis
was performed in an attempt to identify a minimal region of common deletion on
chromosome 9q which may harbor a tumor suppressor gene.
Interestingly, recent evidence has pointed to the existence of another putative
tumor suppressor gene important in bladder tumorigenesis on the short arm of
25
□ Chromosome 9 LOH
■ p5 3 mutations
e m
23 G
S °
2 f
j s 2
« 8
.fa ! V )
8 a
fa
h o
TCCs
Fig. 1.8. Frequency of LOH in chromosome 9 and p53 gene mutations in bladder
TCCs according to tumor stage, and based upon data reported by Olumi et al.,
(1990), Dalbagni et al., (1993), Habuchi et al., (1993), Spruck et al., (1994a), and
Linnebach et al., (1993).
26
chromosome 9. In a study of 46 bladder TCCs by microsatellite polymorphism
analysis, deletions of the q arm found in some tumors and of the p arm in others
clearly demonstrated two separate areas of loss on this chromosome, and provided
evidence for two putative tumor suppressor genes on chromosome 9, one located on
the q arm and the other on the p arm (Ruppert et al., 1993).
In deletion mapping studies of chromosome 9 in bladder TCCs, identification
of the critical region containing a tumor suppressor gene has been hampered by the
lack of selective or small deletions (Habuchi et al., 1993a; Knowles et al., 1994; Cairns
et al., 1993; Linnenbach et al., 1993; Miyao et al., 1993). Table 1.1 shows the regions
on chromosome 9 implicated in the literature as potential locations for one or more
tumor suppressor genes. Whereas loss of genetic material seems to occur along the
whole long arm of chromosome 9, deletions in the short arm appear to be more
selective. Work from several laboratories has provided evidence that a putative tumor
suppressor gene involved in a number of different malignancies is located within
chromosomal band 9p21-22. This region was shown to be homozygously deleted in
leukemias (Diaz et al., 1990), melanomas (Fountain et al., 1992), gliomas (Olopade et
al., 1992), and lung tumors (Olopade et al., 1993). Subsequently, Cairns et al. (1994a)
detected homozygous deletions of the 9p21-22 region in 10 of 112 primary bladder
tumors, and Stadler and colleagues (1994) reported homozygous loss of the same area
in 4 of 7 bladder carcinoma derived cell lines and in one uncultured bladder tumor.
Two research teams (Kamb et al., 1994a; Nobori et al., 1994) demonstrated
that a gene encoding a previously identified negative cell cycle regulator, a protein
27
Table 1.1. Regions of subchromosomal deletions in chromosome 9 detected in bladder TCCs
Region of Deletion No. of partial deletions/No. of
tumors analized (%)
Reference
9pl2-9q22 9/252 (3.5%) Caims et al., 1993b
9pl2 - 9q34.1 3/18(16%) Linnebach et al., 1993
•
9pl2 - 9q34.1 4/123 (3%) Miyao et al., 1993
9ql3-9q34.1 2/95 (2%) Keen et al., 1994
9q34.1 -9qter 2/49 (4%) Habuchi et al., 1993
9p21 -9p23 4/46 (9%) Ruppert et al., 1993
9p21 - 9p22 10/112(10%) Caims et al., 1994a
9p21 - 9p22 4/7 (57%) *
1/10(10%)*
Stadler et al., 1994*
9p21 3/95 (3%) Keenetal., 1994
* This study examined 7 bladder carcinoma derived cell lines and 10 uncultured tumors.
All the other studies included only uncultured tumors.
named pl6 (Serrano et al., 1993), is located within chromosomal band 9p21 and may
be an important tumor suppressor gene. Supporting the role of the p l6 gene (also
known as CDKN2, pl6IN K 4, or MTS1) as a multiple tumor suppressor is the fact that
46% of 290 cell lines derived from 12 different tumor types (including bladder, breast,
lung, renal carcinomas, melanoma, astrocytoma, glioma, osteosarcoma among others)
had lost both copies of the gene (Kamb et al., 1994a; Nobori et al., 1994). However,
the role of the p l6 gene as a major tumor suppressor, at least in bladder TCC, has
been challenged by two independent groups (Caims et al., 1994b; Spruck et al.,
1994b) who showed that p l6 gene homozygous deletions and intragenic mutations
are frequent in bladder carcinoma derived cell lines but infrequent in primary
uncultured bladder TCCs. In the study performed by our group, which is described in
Chapter 4 of this thesis, both primary bladder TCCs and cell lines were examined
(Table 1.2); while 54% of bladder cancer derived cell lines showed p !6 gene
mutations and homozygous deletions, only 19% of uncultured TCCs had these
genetic changes (Spruck et al., 1994b). Sidransky's group reported only two point
mutations in the p!6 gene among 75 uncultured primary tumors examined, including
bladder, lung, kidney, head and neck, and brain tumors (Caims et al., 1994b).
In contrast to the results obtained in TCCs, analysis of bladder SCCs, which is
described in Chapter 6 of this thesis, revealed that p l6 homozygous deletions and
intragenic mutations occur frequently in this tumor type, leading to the notion that
SCC tumorigenesis follows a distinct genetic pathway. Because of the low rate of
intragenic mutations observed in uncultured TCCs, the homozygous deletions of the
p l6 locus might be only an indicator of inactivation of another tumor suppressor gene
also deleted within band 9p21. However, other mechanisms to inactivate p !6 are
29
Table 1J. Tumor suppressor genes altered in bladder TCCs
Gene Alteration No. of tumors with
change/No. of tumors
examined (%)a
References
p53
Gene mutation 107/259 (41%)
Sidransky et al., 1991
Fujimoto et al., 1992
Habuchi et al., 1993
Spruck et al., 1994
Protein overexpression 39/133 (29%) Sarkis et al., 1993
Wright etal., 1991
RB Allele loss 50/174 (29%) Ishikawa et al., 1991
Caims et al., 1991
Xu et al., 1993
Altered protein expression 82/223 (37%)
Ishikawa et al., 1991
Caims et al., 1991
Cordon-Cardo et al.,
1992
Logothetis et al.,
1992
Xu etal., 1993
p!6 pl6 gene homozygous
deletion and mutations
Csll.lioss lumsis
7/13(54%) 6/31(19%) Spruck et al., 1994b
pl6 gene homozygous
deletion
Cell lines
5/15 (33%) Kamb et al., 1994a
pl6 gene mutations Tumors
1/15 (6%) Caims et al., 1994b
flThese numbers represent compiled data from the indicated references, and are independent
of tumor stage and grade.
30
possible including mutations in intronic and regulatory sequences, and epigenetic
phenomena such as abnormal DNA methylation. The analysis of the p l6 gene
expression and methylation status in normal and transformed human tissues is
described in Chapter 5 of this thesis. This study revealed that aberrant methylation of
the S' CpG island of p !6 leading to transcriptional block may be a more common
mechanism to inactivate the gene since de novo methylation of this island was detected
in 67% of uncultured TCCs.
Alterations in the p53 tumor suppressor gene
Loss of genetic material in the short arm of chromosome 17 has been reported
as a consistent event in high-grade, high-stage bladder TCCs (Tsai et al., 1990; Olumi
et al., 1990; Habuchi et al., 1993a; Knowles et al., 1994; Presti et al., 1991). The p53
tumor suppresssor gene is located on 17pl3.1 and encodes a S3 kilodalton nuclear
phosphoprotein that is involved in transcriptional control and may act as a regulatory
check point in the cell cycle, arresting cells in G1 (Levine et al., 1991; Vogelstein and
Kinzler, 1992). Genetic alterations of the p53 gene are not specific of bladder cancer,
but frequent events in many types of human cancer (Levine et al., 1991; Vogelstein
and Kinzler, 1992).
Mutations in the p53 tumor suppressor gene have been identified in over half
of all high-grade, high-stage bladder TCCs (Fig. 1.8). Sidransky et al. (1991)
reported p53 alterations in 1 1 of 18 (61%) invasive bladder carcinomas, and mutations
in this gene were detected in urine samples of three patients. In another study,
31
Fujimoto etal. (1992) found p53 mutations in 6 of 12 (50%) invasive TCCs, but only
one of 13 superficial tumors was found to have a p53 mutation. In an analysis of
different grade and stage TCCs as well as 23 Cis lesions by Dr. Jones' group (Spruck
et al.„ 1994a), only 1 of 36 Ta tumors was found to contain a p53 mutation, whereas
32% of T1 tumors and 51% of T2 and higher stage tumors contained mutations in
this gene. Among the Cis lesions, 65% contained p53 mutations (Fig. 1.8). This is
very interesting, since the high frequency of p53 mutations in Cis, similar to that in
invasive TCCs, could explain, at a molecular level, the high propensity of this type of
flat lesion to progress.
Several groups have investigated the hypothesis that altered patterns of p53
gene expression are correlated with tumor progression in low stage bladder tumors
and in Cis (Sarkis et al., 1993; Lipponen, 1993; Esrig et al., 1994). Sarkis et al.
(1993) studied a group of patients with early invasive disease (T1 bladder cancer) for
p53 overexpression using immunohistochemistry analysis (because of the prolonged
half life of the mutated p53 proteins, they accumulate in the nucleus and can be
detected with specific antibodies). These patients were followed up for 119 months.
The results from this study suggested that T1 bladder tumors with p53 nuclear
overexpression had a higher rate of disease progression. Similar results were
obtained by Lipponen et al. (1993) in an IHC study of 212 bladder tumors, with a
mean follow-up time of 10 years. In a more recent study conducted by Esrig et al.
(1994) and also based on IHC analysis, outcome of organ confined TCC was
significantly related to p53 nuclear overexpression. These studies point to the
important predictive value of the p53 gene status in bladder cancer.
32
Alterations in the retinoblastoma gene
The retinoblastoma (RB) susceptibility gene is an example of a recessive gene
that can promote tumor growth when both alleles are inactivated. It encodes a 110
kilodalton nuclear phosphoprotein involved in cell cycle regulation. The implication
of alterations in this gene in bladder cancer has been mainly studied by IHC analysis
and LOH analysis (Table 1.2). Because of the large size of this gene (27 exons
spanning 190 kilobases), SSCP analysis and direct DNA sequencing are not
frequently carried out. Cordon-Cardo et al. (1992) reported an altered pattern of
expression of the RB gene (from undetectable to heterogeneous levels of the RB
protein) in 13 of 38 (34%) patients with muscle invasive TCCs, and only one of 10
superficial carcinomas was RB protein negative. Work by other groups (Ishikawa et
al., 1991; Cairns et al., 1991; Logothetis et al., 1992; Xu et al., 1993) has also indicated
that alteration of RB is commonly associated with muscle invasive tumors. Logothetis
et al. (1992) suggested that loss of RB function might be a potentially important
prognostic marker of tumor progression in bladder cancer.
ONCOGENES
Mutations in the Harvey-ras (H-ras) oncogene, located on chromosome 1 lp,
were first detected in the human bladder carcinoma cell line T24 (Capon et al., 1983).
The H-ras gene encodes a cell membrane bound protein of 21 kilodalton, p21,
involved in signal transduction pathways. Mutations in H-ras, found in 15% of solid
tumors, are the most common mutations found in oncogenes (Anderson et al., 1992).
33
These mutations activate p21, leading to transduction of growth signals in the absence
of normal stimuli. The initial expectation that activation of this oncogene might play
an important role in bladder carcinogenesis has decreased, as analyses of uncultured
bladder tumors by Fujita apd Knowles have shown that only about 10% contain a
mutated H-ras gene (Fujita et al., 1985; Knowles et al., 1993) (Table 1.3). There is
however, a more recent report showing a significant higher frequency of H-ras
mutations in bladder TCCs (Ooi et al., 1994). This discrepancy with previous reports
may be explained by the higher sensitivity of the method used to detect mutations.
Increased expression of the H-ras product has been detected in Cis and high grade
tumors but not in hyperplasia or low grade tumors (Viola et al., 1985). This is
consistent with the hypothesis that 1 lp changes are secondary genetic alterations in
the development of bladder cancer. Activated p21 may serve as a marker to
distinguish lesions with high malignant potential (Meyers et al., 1989). Still, the role
of the H-ras oncogene in bladder cancer needs to be more clearly determined.
Trisomy of chromosome 7 has been reported in a few cases of bladder TCCs
(Berger et al., 1983; Babu et al., 1987). The c-erb-1 oncogene is located on this
chromosome. The c-erb-1 gene codes for the epidermal growth factor receptor
(EGFR). The ligand for EGFR, epidermal growth factor, is normally excreted in the
urine (Hirata et al., 1979). Increased levels of EGFR have been detected in invasive
bladder TCCs (Wright et al., 1991) (Table 1.3), and EGFR overexpression has been
correlated with poor prognosis. Interestingly, overexpression of EGFR has been
observed also in the normal urothelium of bladders containing tumors (Messing et al.,
1987; Messing et al., 1990), which has led to the hypothesis that bladder cancer
34
Table 1 J. Oncogenes altered in bladder TCCs
Gene Alteration No. of tumors with
alterations / No. of
tumors examined
(%)a
References
H-ras Gene Mutations 46/273(17%)
Fujita et al., 1985
Knowles et al., 1993
Ooi et al., 1994
Increased Expression of
p21
In all High Grade
Tumors
Viola etal., 1985
EGFR
(c-erb-1)
Increased Expression of
EGFR
124/158 (78%)
Wright et al., 1991
Messing et al., 1987
Messing et al., 1990
Berger et al., 1987
c-myc Increased Expression 22/34 (65%) Kotake et al., 1990
c-src Increased Expression
and Increased Kinase
Activity
26/55 (47%)
18/42 (43%)
Fanning et al., 1992
‘These numbers represent data compiled from the indicated references.
35
patients may have a urothelium that is prone to malignancy by overexpression of
EGFR. However, this hypothesis needs further experimental proof.
The levels of the proteins encoded by the oncogenes c-myc and c-src have also
been shown to be increased in bladder TCCs (Table 1.3). Immunohistochemistry
analyses have detected increased levels of the protein product of several oncogenes in
TCCs of the bladder, however, activating mutations, which directly implicate a gene in
tumorigenesis, have only been found in the H-ras oncogene. The significance of
elevated levels of the product of proto-oncogenes in bladder cancer remains unclear
and may be secondary to a state of increased cell proliferation induced by other
primary genetic events.
MICROSATELLITE INSTABILITY IN BLADDER CANCER
Microsatellites are sequences of polymorphic mononucleotide, dinucleotide,
trinucleotide, or tetranucleotide repeats distributed along the human genome (Weber,
1990; Kwiatkowski et al., 1992). They have proven very useful in the analysis of
LOH of chromosomal regions in human cancers. Abnormalities at microsatellites
were first reported in sporadic and familial cases of colon cancer (Ionov et al., 1993;
Peltomaki et al., 1993; Thibodeau et al., 1993). These abnormalities, or instabilities,
consist in the alteration of the number of repeats of a specific microsatellite in tumor
DNA when compared to normal tissue DNA from the same patient, and they indicate
that replication errors have occurred in these sequences (Peltomaki et al., 1993;
Thibodeau et al., 1993). The persistence of these errors are reflective of the inability
36
of tumor cells to repair mutations. In fact, microsatellite instability has been linked, at
least in colon cancer, to alterations in the hMSH2 gene (Leach et al., 1993). This
gene, located on chromosome 2, is the human homolog of the yeast Mut S gene which
codes for an enzyme involved in DNA repair. Microsatellite instability has also been
detected in bladder TCC, as described in Chapter 3 (Gonzalez-Zulueta et al., 1993). In
a study conducted in part in our laboratory, 6 of 200 TCCs were found to contain
microsatellite instabilities at one or more loci on chromosome 9, and in the androgen
receptor gene on the X chromosome. All 6 cases were low-stage, low-grade tumors,
which suggests that these alterations might be an early event in bladder tumorigenesis.
Similar results were later reported by Linnebach et al. (Linnebach et al., 1994), who
extended the microsatellite analysis to other chromosomes and detected frequent
instabilities on chromosome 2. Orlow et al. (1994) observed this event in 41% of
TCCs, which suggests that microsatellite instability is not a random alteration in
bladder cancer. However, the significance of this event in bladder cancer requires
further investigation.
GENETIC PATHWAYS TO TCC OF THE BLADDER
A common feature of all neoplasias in adults is the multistep nature of the
tumorigenic process. Multiple mutations appear to be required to lead to a malignant
phenotype. It appears that tumor progression is the result of the accumulation of
genetic alterations in neoplastic cells (Fearon and Vogelstein, 1990). Thus, it is
reasonable to think that also in bladder cancer molecular genetic alterations follow a
relative sequence of events leading to cancer progression.
37
Recent genetic research has provided evidence that the histologic and clinically
different forms of bladder TCCs contain distinct molecular defects. Figure 1.9 shows
a genetic model of bladder TCC progression derived from cytogenetic and molecular
genetic analyses (Dalbagni et al., 1993; Jones and Droller, 1993; Spruck et al., 1994a),
and proposed by Dr. Jones' group (Spruck et al., 1994a). First, the model illustrates
that specific genotypic patterns are associated with early and late stages of bladder
TCC. Second, the model proposes the existence of two different genetic pathways for
the development of superficial bladder TCC. Recent microsatellites analyses have
revealed microsatellite abnormalities in bladder TCCs (Gonzalez-Zulueta et al., 1993;
Linnebach et al., 1994; Orlow et al., 1994). These abnormalities appear to be very
early alterations (Fig. 1.9), and may reflect the high degree of genetic instability
present in cancer cells. Papillary low-grade Ta tumors frequently show LOH of
chromosome 9 as the only genetic abnormality detected, which suggests that
inactivation of one or more putative tumor suppressor genes on this chromosome is an
initiating event of altered cell proliferation in the urothelium. In contrast, Cis has been
shown to contain LOH of chromosome 9 infrequently but to have frequent p53 gene
mutations. These observations indicate that there may be two separate pathways in
bladder TCC tumorigenesis that differ in the initiating event: either chromosome 9
deletion or p53 gene alteration. After the transitional cells are transformed, they can
undergo further genetic changes that determine the transition from superficial non-
invasive to invasive disease. Such changes may involve deletions in chromosomes 3p,
4p, 8p, 1 lp, 13q, and 18q.
38
IN C R EA SIN G INVASION
A N D PROGRESSION
Normal
Urothelium
Dysplasia
Papillary
Low Grade
Non-Invasive
Papillary
High Grade
Non-Invasive
Carcinoma
in situ
Lamina Propria
Invasive
Muscle Invasive
Metastasis
Fig. 1.9. Model for the genetic pathways in the development and
progression of bladder TCC based on the model proposed by Dr.
Jones' group (Spruck et al., 1994a). Chromosome 9 allelic losses
(9-) or p53 gene mutations (p53-) define different pathways for
the two types of superficial bladder cancer. Other chromosomal
changes have been associated with TCC progression (Tsai et al.,
1990; Dalbagni et al., 1993; Presti et al., 1991; Cairns et al., 1991;
Xu et al., 1993). Arrow thickness is indicative of the percentage
of tumors that progress to higher stages and grades.
39
Few studies have compared the molecular defects in metastases with those of
the corresponding primary tumors. With this objective, Miyao et al. (1993) in Dr.
Jones' group carried out a study with 14 patients from whom primary tumors and one
or more lymph node metastases were available. The genetic events examined were
LOH of chromosomes 9 and 17p and p53 gene mutations. In all cases there was a
complete concordance between the defects present in the primary and metastatic sites.
These results indicated that the three defects studied existed within the primary tumor
before metastasis had occurred. However, no other genetic markers that could be
associated with metastasis were examined.
MOLECULAR EPIDEMIOLOGY
One of the aims of genetic research in cancer is to determine if specific
patterns of genetic alterations exist in specific tumor types. A common pattern of
molecular changes might not only help in understanding the tumorigenic mechanisms
but also indicate the most likely causative agents.
A distinct pattern of point mutations in the p53 gene has been established in
several cancer types, including colon (Fearon and Jones, 1992), lung, and different
types of bladder cancer (Spmck et al., 1993; Shibata et al., 1994; Gonzalez-Zulueta et
al., 1995a). Thus, colon cancers, in which endogenous carcinogens are thought to
play a significant role (Fearon and Jones, 1992), present frequent transition mutations
(purine to purine or pyrimidine to pyrimidine changes) in the p53 gene, particularly at
the CpG dinucleotide in codon 175. In lung cancer however, where exogenous
40
carcinogens, such as cigarette smoking, appear to be more significant, transversions
(purine to pyrimidine or pyrimidine to purine changes) are very prevalent (Sundaresan
et al., 1992). Interestingly, both types of mutation appear to be equally frequent in
bladder cancer. Spruck et al. (1993) in Dr. Jones' group examined the patterns of p53
gene mutations in smoker and non-smoker patients with bladder TCC. No significant
difference in the type of mutation was found between the two groups. G to C
transversions, unusual in other tumor types, were found to be frequent in both groups,
and codon 280 was shown to be a hot-spot for G to C transversions. Although no
difference in the type of mutation was detected, smokers did have double mutations in
the p53 gene, which were not present in the tumors from non-smokers.
In TCCs obtained from patients chronically exposed to phenacetin (Petersen et
al,, 1995) and arsenic (Shibata et al., 1994) a G to A transition at the CpG dinucleotide
in codon 175 of the p53 gene was detected in a few cases (3 of 8 mutations),
consistent with an endogenous carcinogenic mechanism. Interestingly, the codon 280
mutational hot-spot in TCCs was not observed in these tumors.
The study of SCC of the bladder described in Chapter 6 of this thesis has
revealed a somewhat different pattern of mutations in the p53 gene (Gonzalez-Zulueta
et al., 1995a). No mutations at codon 280 were detected in SCCs, similarly to those
TCCs associated with phenacetin or arsenic exposure. In contrast, G to A transitions
at codon 175 were relatively frequent (3 of 13 mutations detected). SCCs of the
bladder account for approximately 5% of all bladder cancers diagnosed in the
Western Hemisphere (Mostofi et al., 1988), and are often associated with long
standing chronic irritation of the urothelium. In contrast, in Egypt and regions of the
41
Middle East and Africa SCCs account for over 75% of all bladder tumors (Mostofi et
al., 1988), and are thought to be caused by infection by the trematode Schistosoma
haematobium. This microorganism, as well as phenacetin and arsenic, has been
proposed to provide a chronic stimulus for urothelial growth. Shistosomal eggs
which reside in the bladder musculature are a constant source of irritation to the
urothelial lining, leading to a steady turnover of the urothelium. It has been proposed
that the stimulation of urothelium growth could potentiate the action of enviromental
mutagens, such as N-nitroso compounds, which are known to exist at high levels in
endemic regions of schistosomiasis (Hicks et al., 1982; Badawi et al., 1992).
However, the relationship between the mutations at the CpG dinucleotide in codon 175
of the p53 gene and inflamation inducing agents awaits further investigation.
CLONAL ORIGIN OF BLADDER CANCER
Bladder cancer sometimes occurs as a multifocal disease, and excised tumors
often recur. This raises questions about the clonal origin of these tumors. Sidransky
et al. (1992) first showed the monoclonal origin of synchronous multiple bladder
tumors in four cases using a molecular genetic approach based on the analysis of the
X chromosome inactivation pattern in female patients. Similar results were
documented by Habuchi et al. (1993b) who analyzed tumors for p53 gene mutations
in patients whose tumors were resected 1 to 3 years apart. In each single patient, all
tumors contained identical p53 mutations, suggesting that these tumors arose from the
same transformed cell, i. e., they were monoclonal. Miyao et al. (1993) evaluated 20
cases of bladder cancer with multiple tumor sites either synchronously or
42
asynchronously for LOH of chromosomes 9 and 17p and p53 gene mutations. Most
of the synchronous and recurrent tumors contained identical genetic lesions, indicating
that the tumors were most probably monoclonal. These observations appear to oppose
the "Held defect" hypothesis, according to which the individual urothelial cells are
primed to undergo transformation, leading to the multicentricity and recurrence of
bladder cancer.
It has been shown recently (Tsai et al., 1995) that the normal urothelium
consists of relatively large monoclonal patches of cells that have the same X
chromosome inactivation pattern and are most probably derived from the same
precursor cell. If individual cells within the same patch became tumorigenic, the
tumors derived from them would all have the same inactivated X chromosome.
Therefore, clonality studies based on this analysis should be interpreted very carefully,
and ideally several genetic markers should be examined.
43
OVERVIEW OF THESIS RESEARCH
Although the progress in clinical management of bladder cancer has been
substantial, the mortality rate has not changed notably in the last 50 years.
Identification of the genetic changes that precede the phenotypic alterations in a
progressively neoplastic urothelium is essential to develop early detection strategies
and more novel and effective treatments of bladder tumors according to the nature of
their molecular alterations. The primary aim of the research described in this thesis
was to advance in the understanding of the molecular genetic defects involved in
bladder cancer tumorigenesis.
Evidence generated by Dr. Jones' group and others (Tsai et al., 1990; Miyao et
al., 1993; Cairns et al., 1994a; Knowles et al., 1994) indicated the importance of
human chromosome 9 alterations in the genesis and progression of bladder TCC, and
suggested the existence of a tumor suppressor gene(s) on this chromosome. Efforts
to map a minimal region of common deletion on chromosome 9 that might harbor a
tumor suppressor are described in Chapter 2. Deletion mapping of bladder TCCs was
performed through microsatellite polymorphisms analysis, a novel PCR-based
strategy to detect chromosomal deletions. The data expands previous studies and
defines two putative tumor suppressor loci on the long arm of chromosome 9: one
mapped to 9q34.1 -qter, and the other mapped to 9p 13-q21.3.
In the process of allelic loss analysis, microsatellite instability was observed in
TCC of the bladder. The study described in Chapter 3 was the first one to report this
44
type of genetic abnormality in bladder cancer. These alterations, which were later
confirmed in bladder TCC by other groups, reflect inactivation of genes involved in
mismatch repair.
A new tumor suppressor, the p!6/CDKN2 gene, was identified and mapped to
chromosomal region 9p21 (Kamb et al., 1994a; Nobori et al., 1994), an area which is
frequently altered in bladder and other types of cancer. Alterations of the
p!6/CDKN2 gene were initially proposed to be involved in the genesis of many tumor
types. However, pl6/CDKN2's role in human tumorigenesis was controversial since
the initial evidence was based solely on tumor-derived cell lines. Chapters 4 and S
describe the analysis of the role of pl6/CDKN2 in bladder TCC tumorigenesis. First,
homozygous deletions and intragenic mutations were found to be frequent in bladder
tumor-derived cell lines but infrequent in uncultured TCCs. The significant difference
between cell lines and primary tumors in pl6/CDKN2 alterations questioned the role
of this gene in bladder tumorigenesis. Subsequently, other mechanisms to inactivate
pl6/CDKN2 function in bladder TCC were investigated; specifically, the methylation
status of this gene and of pl5IN K 4B , a closely related cell cycle regulator, was
analyzed. Methylation of the 5' CpG island of pl6/CDKN2 was found to correlate
significantly with gene silencing not only in bladder TCCs but also in normal tissues
such as colon mucosa. Most importantly, our data suggested that abnormal
methylation of the 5' CpG island may be a more common mechanism to inactivate
pl6/CDKN2. These findings have clear clinical implications since novel therapeutic
strategies can be designed in order to induce the expression of this tumor suppressor
in tumors cells that have inactivated the gene by DNA methylation. Thus, preliminary
studies aimed to investigate the feasibility of such approach are described in
45
Chapter 5. The effect of 5-aza-2'-deoxycytidine, a potent inhibitor of DNA
methylation, on p!6/CDKN2 expression was examined in vivo and in vitro in cells that
contained an inactive and methylated pl6/CDKN2 gene. The data indicate that a
dormant pl6/CDKN2 gene can be reactivated by 5-aza-2'-deoxycytidine treatment, and
that pl6/CDKN2 induction is associated with retardation of cell proliferation in vitro.
Identification of the genetic alterations involved in the development of SCC,
the most common type of bladder cancer in some parts of the world is described in the
last Chapter. SCCs were examined for genetic defects previously shown to be
involved in TCC tumorigenesis, including allelic loss of chromosomes 9 and 17p, and
alterations in the p!6/CDKN2 and p53 tumor suppressor genes. The data indicate that
SCCs differ from TCCs not only in their epidemiologic and clinical characteristics but
also, and very importantly, in their underlying genetic alterations. Thus, allelic loss of
chromosome 9p and pl6/CDKN2 homozygous deletions and intragenic mutations
were detected at a significantly higher frequency in SCCs than in TCCs. The p53
gene mutation spectrum was also found to differ in these two tumor types. These data
led to the hypothesis that SCC tumorigenesis follows yet another distinct genetic
pathway, and helped to expand the genetic model of bladder cancer initiation and
progression previously proposed by Dr. Jones' group.
46
CHAPTER ?
DELETION MAPPING OF CHROMOSOME 9 IN BLADDER
TRANSITIONAL CELL CARCINOMA USING MICROSATELUTE
POLYMORPHISMS: DEFINITION OF TWO PUTATIVE TUMOR
SUPPRESSOR LOCI
INTRODUCTION
Our understanding of the genetic alterations underlying transitional cell
carcinoma (TCC) of the bladder has expanded over the last decade. Deletions
involving chromosome 9 represent the most frequent genetic alteration identified in
bladder TCCs to date with approximately 60% of superficial and invasive TCC
showing loss of heterozygosity (LOH) at markers on this chromosome, leading to the
belief that chromosome 9 contains one or more suppressor genes involved in bladder
tumorigenesis. With the exception of carcinoma in situ (Cis), these deletions are
present at similar frequencies in bladder tumors of all stages and grades including
grade I and superficial tumors confined to the epithelial cell layer (Tsai et al., 1990;
Cairns et al., 1993; Habuchi et al., 1993a; Miyao et al., 1993; Knowles et al., 1994).
Therefore, chromosome 9 allelic loss not only is a common event in all bladder TCCs
but it may represent an early or initiating event in bladder tumorigenesis (Tsai et al.,
1990; Olumi et al., 1990; Presti et al., 1991; Habuchi et al., 1993; Caims et al., 1993;
Miyao et al., 1993; Linnenbach et al., 1993; Orlow et al., 1994; Knowles et al., 1994).
In contrast, mutations in the p53 tumor suppressor gene and LOH in chromosome
17p, where p53 resides, occur in more advanced and aggressive bladder tumors and
47
are thought to be secondary events important in cancer progression (Tsai et al., 1990;
Olumi et al., 1990; Presti et al., 1991; Habuchi et al., 1993a; Miyao et al., 1993;
Sidransky et al., 1991; Fujimoto et al., 1992). An exception is Cis, a superficial and
flat neoplastic lesion within-the urothelium which has been shown to contain frequent
p53 gene mutations and infrequent chromosome 9 allelic losses. Other genetic
alterations detected in invasive TCCs include allelic losses in chromosomes 3,3,6,8,
11, 13, 18, and are thought to be important in progression (Tsai et al., 1990; Habuchi
et al., 1993a; Dalbagni et al., 1993; Fearon et al., 1985; Knowles et al., 1993; Knowles
et al., 1994; Brewster et al., 1994).
Mapping the precise location of putative tumor suppressor genes on
chromosome 9 involved in bladder tumorigenesis has proven difficult due to the high
frequency of monosomy 9, infrequent partial subchromosomal losses, no familial
form of bladder cancer, and inadequate coverage of chromosome 9 by highly
informative polymorphic markers in early studies. The improvement in the density of
mapped polymorphic simple sequence repeat markers, or microsatellites, on
chromosome 9 (Kwiatkowski et al., 1992) presents the possibility of constructing
deletion maps covering all regions of the chromosome. Microsatellites are tandem
iterations of simple dinucleotide, trinucleotide, or tetranucleotide repeats, and their
usefulness can be attributed to abundancy (Beckman and Weber, 1992; ),
hypervariability (Litt and Luty, 1989), fairly even genomic distribution (Stallings et al.,
1990), and ease of detection by the polymerase chain reaction. Therefore,
microsatellites are now key markers in the analysis of LOH in cancer.
48
Several recent reports have presented evidence for at least two tumor
suppressor genes on chromosome 9, one putative suppressor located on the short arm
(p arm) and another located on the long arm (q arm) of chromosome 9 (Ruppert et al.,
1993; Keen and Knowles,-1994). The minimal common area of subchromosomal
deletion on 9p has been mapped to a 4 cM region between markers D9S171 and
IFNA on band 9p21 (Cairns et al., 1994a; Stadler et al., 1994; Devlin et al., 1994).
This region includes the recently identified tumor suppressor gene pl6/CDKN2,
which encodes a cyclin-dependent kinase inhibitor important in the control of the Gl-
S transition in the cell cycle (Kamb et al., 1994a). In contrast, the precise location of
the putative suppressor gene(s) on 9q remains elusive. Most studies have localized
the proximal limit of a minimal common area of deletion in 9q to either marker D9S18
on 9p21-ql3 or to ASSP3 on 9ql3. The distal limit reported varies from marker
D9S22 on 9q22 to loss of the entire q arm (Habuchi et al., 1993; Cairns et al., 1993;
Miyao et al., 1993; Linnenbach et al., 1993; Ruppert et al., 1993; Keen et al., 1994;
Caims et al., 1994a). It still needs to be determined whether the target regions on 9p
and 9q are deleted concurrently or independently, and if so which event occurs first in
bladder tumorigenesis.
Seventy primary TCCs of the bladder were examined in this study for LOH at
microsatellite markers spanning chromosome 9 in order to progress in the mapping of
the putative tumor suppressor gene(s) on chromosome 9q critical in bladder cancer.
Our results indicate that two putative tumor suppressor loci including 9q exist. One is
mapped to a common minimal area of deletion between D9S171 (9p21) and D9S283
(9q21-22), and a second smaller region of common deletion is telomeric to GSN
49
(9q33). Our results also suggest that deletion on 9q may occur earlier than deletion
on 9p in the formation of noninvasive papillary tumors.
MATERIALS AND METHODS
Tumor samples and DNA isolation
Seventy primary transitional cell carcinomas (TCC) of the bladder were
analyzed. Tumors were obtained from the USC/Kenneth Norris Cancer Center, Los
Angeles, California, and from the Herlev Hospital, University of Copenhagen,
Denmark. TCCs were classified according to Bergkvist classification (1965). Thirty
eight Ta (papillary noninvasive), 11 T1 (lamina propria-invasive) and 21 T2-T4
(muscle-invasive) tumors were examined. Sixteen tumor specimens were obtained
from radical cystectomy. For these specimens, DNA from the corresponding patient's
peripheral blood lymphocytes was used as the normal control. Fifty four samples
were formalin-fixed, paraffin-embedded samples obtained from transurethral resection
of bladder tumor (TURBT) specimens.
High molecular weight DNA was prepared from fresh tumor specimens and
matching blood samples by proteinase K digestion and phenol/chloroform extraction
as described (Marmur, 1961; Bell et al., 1981). DNA isolations from fresh tumor
specimens and from white blood cells were performed by Ms. Mary Peer and Dr.
Yvonne Tsai. DNA from archival paraffin-embedded specimens was isolated by
dissecting tumor and normal tissues from hematoxylin and eosin stained sections as
50
described (Spruck et al., 1993). Briefly, approximately 1000 cells from each normal
and tumor tissue were dissected using sterile scalpels from 8-10 p.m-thick paraffin-
embedded tissue sections. The dissected tissue was suspended in 20-50 pi of buffer
containing 10 mM Tris (pH 7.4), 5mM EDTA, and boiled for 3 min. Proteinase K
(Boehringer Mannheim, Gmb, W. Germany) was added to a final concentration of 1
pg/pl, and the samples were incubated at 56°C overnight. Proteinase K . was
inactivated by boiling for 10 min. Two microliters were used in subsequent PCR
amplifications.
In some cases in which the above-described microdissection technique was
specially difficult due to the limited amount of tumor cells, DNA from tumor tissue
was isolated from archival specimens by Selective UV-Radiation Fractionation
(SURF) technique (Shibata et al., 1992). Approximately 250-500 tumor cells in a
hematoxylin and eosin stained tissue section without cover-slip were carefully dotted
by hand under an optic microscope using a felt-tip pen (Sharpie, fine point, Sanford
Corporation, Belwood, IL). The slides were placed tissue side down on a UV
transluminator (302nm, Model TM-36, 8000 UW/cm2, UVP, San Gabriel, CA) and
exposed to UV light for 30 min to completely fragment the DNA in the ink-
unprotected tissue. Ink-protected tissue dots were scraped with a sterile scalpel into a
microfuge tube and processed as described above.
Determination of LOH on chromosome 9 by microsatellite analysis
DNA from 70 TCCs and corresponding normal controls were examined using
nine microsatellite markers in chromosome 9 (D9S171 and D9S156 on 9p and
51
D9S15, D9S153, D9S283, D9S12, D9S176, GSN and D9S63 on 9q). Loci, primer
sequences and annealing temperatures are specified in Table 2.1. Primer sequences
were retrieved from the Genome Data Base (Welch Medical Library, The Johns
Hopkins University). Each-PCR amplification was performed in a 25 |ll volume and
included 200 of dATP, dGTP, and dTTP, 2.5 pM of dCTP, 0.2 pCi [a-
32pjdCTP (3000 Ci/mmol, ICN Biochemicals, Costa Mesa, California), 10 mM Tris-
HC1 (pH 8.3), 50 mM KC1,1.5 mM MgCl2,1 pM of each primer, and 0.625 units (u)
of Taq polymerase (Boehringer Mannheim, Gmb, W. Germany). Conditions for
PCR amplifications consisted of 94°C for 1 min, 28 cycles of 94°C for 1 min, 56°C for
40 s, and 72°C for 40 s, followed by incubation at 72°C for 1 min. Annealing
temperatures varied for each primer set as indicated in Table 2.1. PCR products from
matched normal and tumor DNAs were separated by electrophoresis in 8%
polyacrylamide/7 M urea gels. Electrophoresis was performed at 75-80 W constant
power, maintaining gel temperature at 55 °C. Gels were dried in a gel dryer (Model
583, Bio-Rad, Richmond, California) and exposed to autoradiogram film (Hyperfilm-
MP, Amersham, Arlington Heights, IL), using one intensifying screen, at -70°C,
overnight. The autoradiographs were assessed visually. Two more intense bands in
the normal DNA represent the presence of two alleles of different size, i.e.,
heterozygosity at the specific locus examined. Multiple bands of decreased intensity
are commonly seen in the analysis of microsatellites, and probably represent the effect
of DNA polymerase slippage in the repeated sequence. Allelic loss was scored for
informative or heterozygous cases when intensity of the signal for a tumor allele was
significandy reduced (by >95%) relative to the matched normal allele.
52
Table 2.1. Microsatellite markers on chromosome 9 analyzed in this study
Chromosomal
Location
Locus Name0 Primer Sequences* Annealing
Temperature
C
9p22 D9S156 ATCACTTTTAACTGAGGCGG
AGATGGTGGTGATAGAGGG
5TC
9p21 D9S171 AGCTAAGTGAACCTCATCTCTGTCT
ACCCTAGCACTGATGGTATAGTCT
56’C
9pl3-p21 D9S169 AGAGACAGATCCAGATCCCA
TAACAACTCACTGATTATTTAAGGC
58’C
9ql3-q21.1 D9S15 TAAAGATTGGGAGTCAAGTA
TTCACTTGATGGTGGTAATC
54°C
9ql3-q22.3 D9S153 TTATGGCAGCCCAAATGGACTA
GCAGAATGTTGCCCAAAACTCA
55°C
9q21.3-q22.1 D9S283 TGAGGCAGGAAAATCACTTG
CCAGTTATACATGTATGGGT
56°C
9q22.3 D9S12 CCTCCTCACACCTCATGTG
AGCTGGGGTGGGGGGAGT
58°C
9q22.3-q31 D9S176 AGCTGGCTGTTGGAGAAA
TGACCAATGGCAGGGTAT
56°C
9q33 GSN CAGCCAGCITTGGAGACAAC
TCGCAAGCATATGACTGTAA
56°C
9q34.1 D9S63 CCGGAAGTTACTCTAGTCTA
TTATAATGCCGGTCAACCTT
54°C
°A11 microsatellites analyzed are dinucleotide repeat polymorphisms of the form (GT)n.
^Primer sequences for each microsatellite marker were obtained from the Genome
Data Base (Welch Medical Library, The Johns Hopkins University). All sequences
are given in S' to 3' orientation.
C PCR conditions used for each primer pair varied only in the annealing temperature,
which is indicated in this table.
53
RESULTS
Table 2.2 shows the results obtained from the analysis of chromosome 9
allelic loss in 70 TCCs. Overall, 30% (21/70) of the tumors retained all informative
markers on both chromosomal arms, whereas 37% (26/70) showed LOH at all
informative markers examined indicating monosomy 9. LOH at all informative
markers in 9q with retention at all markers in 9p was found in 11% (9/70) of TCC,
and 3% (2/70) showed LOH in the 9p markers with retention at all informative
markers on 9q. Partial deletions on 9q, based on informative microsatellites, was
demonstrated in 18% (13/70) of tumors.
Definition of a putative tumor suppressor locus on 9q34.1-qter
Table 2.2 demonstrates the accumulation of chromosome 9 losses in more
invasive tumors. Partial deletions in 9q were observed most frequently (24%) in Ta
tumors (papillary, non-invasive), while 18% of T1 tumors (lamina propria-invasive)
and only 10% of T2 tumors (muscle-invasive) showed partial losses on chromosome
9q. On the contrary, monosomy 9 was more common in the invasive tumors. Figure
2.1 shows the results obtained for those tumors in which partial losses were detected.
The nine tumors represented in Fig. 2.1a contained allelic loss in markers in distal 9q,
and retained one or more proximal 9q markers. Four of these tumors (# 3,4,11, 15)
showed LOH in D9S63 with retention in GSN, therefore the minimal common region
of LOH delimited by these four tumors mapped to 9q34.1-qter, telomeric to GSN in
9q33.
54
Table 2.2. Summary of chromosome 9 LOH analysis" in 70 TCCs* of the bladder
Ta Tl >T2 Total
9 retention* 15/38 (39.5)* 2/11 (18) 4/21 (19) 21/70 (30)
9 monosomy** 10/38 (26.0) 5/11 (46) 11/21 (52) 26/70 (37)
9q loss* 4/38 (10.5) 1/11 (9) 3/21 (14) 8/70 (11)
9p lossf 0/38 (0) 1/11 (9) 1/21 (5) 2/70 (3)
Partial 9q& 9/38 (24) 2/11 (18) 2/21 (10) 13/70 (18)
"Table 2 is based on data from informative microsatellite markers.
*TCCs were classified according to Bergkvist et al. (1965) in superficial non-invasive (Ta),
lamina propria invasive (Tl), and muscle-invasive (>T2) tumors.
c9 retention represents no LOH in any marker on 9p or 9q.
d9 monosomy represents LOH in all informative markers on 9p and 9q.
*9q loss represents LOH in all informative 9q markers with retention in 9p markers.
)9p loss represents LOH in all informative 9p markers with retention in all 9q markers.
^Partial 9q represents LOH in some 9q markers with retention in other 9q markers.
^Percentages in parenthesis.
Chromosome 9 LOH analysis was performed by Dr. A. Simoneau, Dr. C. Spruck and Dr. M.
Gonzalez-Zulueta.
Fig. 2 . 1 . Microsatellite analysis revealed that 13 TCCs contained areas of partial
allelic loss in chromosome 9 associated with the q arm. Those areas represent two
regions where putative tumor suppressor genes for bladder cancer may reside.
Chromosome 9 and the markers analyzed are indicated. The tumor stage is
designated below the corresponding tumor #. D retained; I LOH; 13 not
informadve; § not done.
(a) Nine tumors showed LOH in markers in distal 9q while retained markers more
proximally. Four of these tumors (# 3,4, 11,15) showed LOH at D9S63 on 9q34.1
with retention at GSN on 9q33, presenting evidence that a putative tumor suppressor
gene resides telomeric to GSN. (b) Seven tumors showed LOH at proximal 9q
markers while retained markers more distally. The common minimal interstitial
deletion is flanked by D9S171 on 9p21 and D9S283 on 9q21-22, presenting
evidence for a second region where a putative tumor suppressor gene may reside.
Microsatellite analysis of TCCs was performed by Dr. A. Simoneau, Dr. C. Spruck,
and Dr. M. Gonzalez-Zulueta.
56
Tumor# 3 4 15 1 1 5 10 13 12 14
Stage H i 111 12 Ik Ik Ik Ik II T k
9p
24
23
22
21
13
12
11
11
D9S156
D9S171
S S S S S i
9 q 221
r
D9S15
D9S153
D9S12
D9S176
GSN
D9S63
/✓ /
/✓ /
□
■ l IL
□
□ n m
T u m o r# 7 8 9 6 3 10 11
S tag e P l Ik P2 Pa Tb Hi T Ji
D9S156
D9S171
D9S15
D9S153
D9S12
D9S176
GSN
D9S63
P7TT1
V / /
’ ///
' / / /
V //
’ / / /
7 / /
H I
II I I
I I
□ □ □ □
1
Definition of a putative tumor suppressor locus on 9pl3-q21.3
Figure 2.1 b shows the results for 7 tumors which had LOH in proximal 9q
markers with retention in distal 9q markers. Tumors # 3, 8, 10 and 11 were further
analyzed with additional microsatellites in an attempt to map more finely this proximal
region of deletion. Only the marker at D9S283 provided helpful information in
defining the minimal common loss by showing retention in tumor #11, thus
narrowing the region of proximal LOH in 9q. Therefore, the smallest common area of
deletion was defined by tumor # 1 1 (Fig. 2.2) which retained all informative markers
in 9p and markers in 9q distal to D9S283, but contained allelic loss in more proximal
9q markers. This region between D9S171 and D9S283 expands approximately 51
cM. Interestingly, three tumors (# 3, 10, 11) contained both proximal and distal areas
of partial LOH and retained heterozygosity in markers between those two putative
tumor suppressor loci (Fig. 2.1 and 2.2).
Chromosome 9q deletion may be a primary event in superficial papillary
TCC development
As illustrated in Fig. 2.1, four of the nine Ta tumors with partial 9q deletions
had concomitant LOH in 9p markers (tumors # 3, 4, 5, 6) and four did not (tumors #
8,11,13,14). None of the Ta tumors examined showed LOH in markers on 9p with
retention in markers on 9q; however, two invasive tumors showed LOH in only 9p
markers. In contrast, four Ta tumors showed allelic loss in markers on 9q with
retention in all informative markers on 9p (Fig. 2.1 and 2.2).
59
s a n s . = = == s =333 333 s a a 3 3 3
IB I B ■
n o n Df Sl SS DK17 C GS N D7 S4 3
NT NT NT NT NT
*
M
Fig. 2 . 2 . Representative examples of results from the allelic loss analysis of DNA
from tumor #11. DNA from normal and tumor tissue was microdissected from 8
pm-thick paraffin-embedded tissue sections. The data obtained from this
papillary, non-invasive tumor are of interest because: a) they helped to limit the
smallest area of LOH on proximal 9q to 9pl3-q21.3; b) they showed that two
regions of LOH separated by an area of retention were detected in some cases; c)
they indicated that 9q deletion may be the primary event in superficial papillary
TCC, as markers on 9q showed LOH whereas markers on 9p showed no allelic
loss. Chromosome 9 with the location of the informative markers for this case are
represented. N, DNA from normal tissue; T, DNA from the corresponding tumor
tissue. Two more intense bands in the normal DNA represent the presence of two
alleles of different size, i.e., heterozygosity at the specific locus examined.
Multiple stattered bands of decreased intensity are commonly seen in the analysis
of microsatellites, and probably represent the effect of DNA polymerase slippage
in the repeated sequence. LOH is scored for informative or heterozygous cases
when intensity of die signal for a tumor allele is significantly reduced (by £95%)
relative to the matched normal allele. The microsatellite polymorphisms shown in
this figure illustrate that areas of LOH (D9S153 and D9S63) were separated by an
interstitial region of retention of heterozygosity (D9S176 and GSN).
60
DISCUSSION
The aim of this study was to further map a common minimal region of LOH
on the q arm of chromosome 9 which could lead to the localization of a tumor
suppressor gene important in bladder tumorigenesis. Microsatellite markers on the p
arm, D9S171 or D9S156, were analyzed to evaluate the presence of monosomy 9
which was scored as such when LOH was detected at all informative markers on the q
and p arms. Overall, 37% (26/70) of TCCs showed monosomy 9 and 32% (23/70) of
tumors had subchromosomal deletions. These data are consistent with a recently
published study in which microsatellite analysis showed that 32% and 22% of TCCs
contained monosomy 9 and subchromosomal deletions, respectively (Keen et al.,
1994). Perhaps the higher overall frequency of subchromosomal partial LOH
detected in our study was due to the fact that our analysis included a large proportion
(54%) of non-invasive papillary tumors.
Eighteen percent of TCCs (13/70) showed partial LOH at markers on
chromosome 9q, and two distinct regions of LOH including 9q were apparent. Nine
tumors contained LOH at marker D9S63 on 9q34, the most telomeric marker
analyzed in this study, and four of these tumors retained the adjacent marker GSN on
9q33 (Figure 2.1a). Therefore, the GSN marker flanks a region on the tip of 9q
where a putative tumor suppressor gene may reside. This is of interest because the
loss of the ABO antigen, which lies approximately 6 cM telomeric to D9S63 on 9q34
(Povey et al., 1994), has been suggested as a marker for bladder cancer progression
(Olumi et al., 1990). Consistent with our observations, Orlow et al. (1994) also
reported a high frequency (45%) of allelic loss at markers associated with band
9q34.1 in human bladder tumors. The tuberous sclerosis gene (TSC1 gene) has been
mapped to band 9q34, distal to D9S149 and proximal to D9S114 (Povey et al., 1994),
and the hereditary hemorrhagic telangiectasia gene (HHT gene) resides on band 9q33-
34 between D9S61 and D9S63 (McDonald et al., 1994). Both genes are involved in
syndromes of angiodysgenesis. The tuberous sclerosis syndrome presents as
angiofibromas and angiomyolipomas in 40-80% of patients, who have a higher than
average incidence of renal cell carcinoma (Glassberg et al., 1992). McDonald et al.
(1994) speculated that the HHT gene may be a tumor suppressor gene encoding an
antiangiogenesis factor. Angiogenesis is vital for solid tumor development, hence a
gene involved in angiogenesis regulation may be a candidate tumor suppressor gene.
It is also of interest that a putative cyclin dependent kinase gene, PITALRE, has
recently been mapped to 9q34.1 (Bullrich et al., 1995). However, the role of this new
gene in cell proliferation control has yet to be determined.
Another region for a putative bladder cancer tumor suppressor gene is limited
by D9S171 on 9p21 and D9S283 on 9q21-22. Seven tumors contained LOH at
markers proximal to 9p21 with retention of heterozygosity at one or more markers
distal to 9q21-22. Bemues and colleagues reported on one case of bladder cancer
with a deletion of this region, del (9ql 1 q21.1), by karyotype analysis (Bemues et al.,
1993). Several candidate tumor suppressor genes reside on proximal 9q. For
example, the human growth arrest specific gene 1 (gasl) has been mapped to 9q21.3-
q22 proximal to D9S12 (Del Sal et al., 1994; Wicking et al., 1995). Overexpression
of gasl has been shown to block cell proliferation in the T24 bladder carcinoma cell
line (Del Sal et al., 1994). Similarly, the annexin 1 gene which likely plays a role in
cell regulation maps to 9ql 1 -q21.2 (Bemues et al., 1993). Excluded from this region
62
is the nevoid basal cell carcinoma syndrome gene (NBCC gene) or Gorlin’s
syndrome gene, which has long been considered a possible candidate tumor
suppressor important for bladder cancer (Gailani et al., 1992; Shanley et al., 1995).
The NBCC gene maps to 9q22.3-q31 between D9S180 and D9S196, approximately 7
cM telomeric to the region defined in our analysis. Other genes excluded are the
EES1 gene (multiple self healing squamous epithelioma), the XPAC gene (xeroderma
pigmentosum complementation group A) and the FACC gene (Fanconi anemia
complementation group C) (Povey et al., 1994). The region proposed by our study
does not include the previously defined tumor suppressor locus at 9p21 where the
p!6/CDKN2 tumor suppressor gene resides. Although pI6/CDKN2 has been shown
to be frequently altered in bladder cell lines (Spruck et al., 1994b) and in squamous
cell carcinoma of the bladder (Gonzalez-Zulueta et al., 1995; Chapter 6), p!6/CDKN2
mutations and homozygous deletions are infrequent in primary TCCs (Spruck et al.,
1994b; Cairns et al., 1994b).
There are most likely various genes on chromosome 9 involved in bladder
tumorigenesis. Knowles and colleagues reported on two possible target sites for
suppressors on the p arm based on data from one tumor (Keen et al., 1994), and our
results indicate the existence of at least two suppressor loci involving the q arm. It is
interesting that the clinical scenarios associated with allelic loss on chromosome 9q
involve epithelial tumors: NBCC, EES, XPAC, bladder TCC, squamous cell
carcinomas of the head and neck, non-small cell carcinoma of the lung, and ovarian
carcinomas contain 9q LOH in 35 to 54% of cases analyzed (Ah-See et al., 1994;
Merlo et al., 1994; Osborne et al., 1994; Cliby et al., 1993).
63
It is not yet clear whether the target regions on 9p and 9q are deleted
concurrently or independently, and if so which event occurs first in bladder
tumorigenesis. Our results suggest that allelic losses on 9q may be preferentially
involved in non-invasive papillary tumor initiation since no Ta tumors showed LOH in
only 9p, whereas four Ta tumors showed 9q allelic losses with 9p retention. The Ta
tumors with partial deletion on 9q were evenly distributed between those with and
without concomitant 9p loss. In contrast, two invasive tumors showed LOH in only
9p markers. Therefore, the inactivation of a gene located on chromosome 9q may be
an early event in bladder TCC tumorigenesis. However, the number of cases in our
study is small and further analysis is necessary to determine the timing of inactivation
of these putative tumor suppressor genes.
64
CHAPTER 3
MICROSATELUTE INSTABILITY IN BLADDER CANCER
INTRODUCTION
In cancer cells mutation of a tumor suppressor gene is often accompanied by
loss of the non-mutated allele and large regions of flanking DNA (Fearon et al.,
1985). Loss of an allele from tumor DNA can be detected with polymorphic markers
for the affected locus. In tumor samples in which the marker is informative, i.e.
heterozygous, deletion of the locus on one chromosome results in loss of
heterozygosity (LOH) for the polymorphism. Analysis of LOH in cancer may,
therefore, indicate the chromosomes in which tumor suppressor genes are located.
More detailed analysis of chromosomal arms with several polymorphic markers may
identify a common region of deletion and allow localization of the tumor suppressor
gene within a small area of the chromosome.
Early attempts to map small areas of deletion on chromosome 9 in bladder
cancer were hampered by the inadequate coverage of the chromosome by highly
informative polymorphic markers. In particular, studies of the long arm of
chromosome 9 were largely restricted to the distal region (9q34) where several
variable number tandem repeat (VNTR) markers had been mapped (Cairns et al.,
1993; Miyao et al., 1993). Deletion mapping requires the use of highly informative
markers covering all regions of the chromosome. Microsatellite polymorphisms have
65
proven very useful in the analysis of LOH at chromosomal regions in bladder and
other cancers. Microsatellites are sequences of polymorphic mononucleotide,
dinucleotide, trinucleotide, or tetranucleotide repeats distributed along the genome.
Microsatellite polymorphisms are ideal for deletion mapping since they are highly
informative (Litt and Luty, 1989), abundant (Beckman and Weber, 1992), fairly evenly
distributed throughout the genome (Stallings et al., 1990), and easy to analyze by the
polymerase chain reaction (PCR).
Alterations at microsatellites have been reported in some inherited diseases and
in some types of cancer. These alterations, or instability, consist in the expansion or
reduction in the number of repeats of a specific microsatellite (Nelson and Warren,
1993). Expansion of a trinucleotide repeat is responsible for the fragile X syndrome
(Kremer et al., 1991), spino-bulbar muscular atrophy (Biancalana et al., 1992),
myotonic dystrophy (Mahadevan et al., 1992), Huntington's Disease (The
Huntington's Disease Collaborative Research Group, 1993), and spinocerebellar ataxia
type 1 (Orr et al., 1993). Genetic instability at microsatellites has also been shown in
some inherited syndromes in which there is an associated predisposition to cancer,
such as xeroderma pigmentosum, ataxia telangiectasia and Bloom's syndrome
(German et al., 1989). Dinucleotide repeat alterations were first described in
hereditary non-polyposis colon cancer (HNPCC) and were linked to predisposition to
colorectal cancer (Ionov et al., 1993; Aaltonen et al., 1993; Thibodeau et al, 1993).
Changes in microsatellite repeats in HNPCC were shown to be variable, ranging from
2 base pairs changes to larger alterations (Ionov et al., 1993; Thibodeau et al, 1993).
These alterations were attributed to non-repaired replication errors, or RER (Aaltonen
et al., 1993), the persistance of which is reflective of the inability of tumor cells to
66
repair mutations. In fact, microsatellite instability has been linked, at least in colon
cancer, to alterations in any one of four genes (hMSH2, hMLHl, hPMSl and
hPMS2) all of which encode homologs of the microbial mismatch repair proteins
MutS and MutL (Leach et al., 1993). The hMSH2 gene codes for a MutS homolog,
whereas hMLHl, hPMSl, and hPMS2 encode homologs of MutL (Papadopoulos et
al., 1994). A mismatch-binding heterodimer has been recently isolated from Hela
cells by virtue of its ability to restore mismatch repair to nuclear extracts of hMSH2-
deficient colorectal tumor cells (Drummond et al., 1995; Palombo et al., 1995;
Papadopoulos et al., 1995). This mismatch-binding heterodimer consists of two
distinct proteins, the 100-kilodalton hMSH2 and the 160-kilodalton GTBP (G/T
binding protein). Alterations in either protein lead to deficient mismatch-binding
activity and, therefore, confer hypermutability and cancer predisposition. Tumor cells
deficient in mismatch repair activity exhibit a marked instability of microsatellite
sequences.
Work from various groups (Vanni et al., 1986; Tsai et al., 1990; Olumi et al.,
1990; Knowles et al., 1993; Habuchi et al., 1993; Dalbagni et al., 1993) indicates that
deletions of chromosome 9 is the most frequent genetic alteration in bladder TCC, and
evidence from Dr. P. A. Jones' group suggests that LOH of chromosome 9 is an early
event in the generation of non-invasive papillary TCC but not in carcinoma in situ of
the bladder (Spruck et al., 1994a). In contrast, mutations in the p53 tumor suppressor
gene and allelic losses of chromosome I7p, where p53 resides, are commonly seen in
carcinoma in situ and in invasive tumors (Spruck et al., 1994a).
67
We have used (GT)n repeat polymorphisms located on human chromosomes
9 and 17p to determine the sequence of molecular defects occurring in a series of 200
TCC of the bladder. In parallel, a trinucleotide repeat polymorphism within the
androgen receptor gene which resides on the X chromosome was used to determine
the clonal origin of some of these tumors. In the process of microsatellite analysis,
changes in these markers were apparent in some TCCs including four tumors with
dinucleotide repeat alterations in chromosome 9, and three tumors with changes in the
trinucleotide repeat in the androgen receptor gene. It is interesting that all the tumors
with microsatellite alterations were low stage, tumors, suggesting that the genomic
instability giving rise to these changes might occur as an early event in bladder
tumorigenesis.
MATERIALS AND METHODS
Tissue specimens and DNA isolations
Two hundred transitional cell carcinomas of the bladder were analyzed in this
study. TCC specimens were obtained from hospitals in Los Angeles County,
California (n = 90), from the Herlev Hospital in Copenhagen, Denmark (n = 64), and
from the Johns Hopkins tumor bank, Baltimore (n = 46). 112 were fresh-frozen and
88 were paraffin-embedded tissues. Tumors were graded according to the criteria of
Bergkvist et al. (1965) and Lieskovsky et al. (1988), and staged according to the
tumor-nodes-metastasis staging system (Bears et al., 1988). 154 specimens were
analyzed in Dr. P. A. Jones' laboratory by Mirella Gonzalez-Zulueta at the
68
USC/Norris Cancer Center, whereas 46 tumors were examined in Dr. D. Sidransky's
laboratory at the Johns Hopkins University. The distribution of TCCs examined in
Dr. Jones' laboratory is presented in Table 3.1.
High molecular weight DNA was prepared from fresh tumor specimens and
matching blood samples by proteinase K digestion and phenol/chloroform extraction
as described (Marmur, 1961; Bell et al., 1981). DNA isolations from fresh tumor
specimens and from white blood cells were performed by Ms. Mary Peer and Dr.
Yvonne Tsai. DNA from archival paraffin-embedded specimens was isolated by
dissecting tumor and normal tissues from hematoxylin and eosin stained sections as
described (Spruck et al., 1993). Briefly, approximately 1000 cells from each normal
and tumor tissue were dissected using sterile scalpels from 8-10 pm-thick paraffin-
embedded tissue sections. The dissected tissue was suspended in 20-50 pi of buffer
containing 10 mM Tris (pH 7.4), 5mM EDTA, and boiled for 3 min. Proteinase K
(Boehringer Mannhein, Germany) was added to a final concentration of 1 pg/pl, and
the samples were incubated at 56°C overnight. Proteinase K was inactivated by
boiling for 10 min. Two microliters were used in subsequent PCR amplifications.
In some cases in which the above-described microdissection technique
was specially difficult due to the limited amount of tumor cells, DNA from tumor
tissue was isolated from archival specimens by Selective UV-Radiation Fractionation
(SURF) technique (Shibata et al., 1992). Approximately 250-500 tumor cells in a
hematoxilin and eosin stained tissue section without cover-slip were carefully dotted
by hand under an optic microscope using a felt-tip pen (Sharpie, fine point, Sanford
Corporation, Belwood, IL). The slides were placed tissue side down on a UV
69
Table 3.1. Stage and grade of 154 TCCs examined
Stage and grade No. of tumors analyzed
Carcinoma in situ (Cis) 23
Superficial (Ta), grade II 43
Superficial (Ta), grade HI 6
Lamina propria-invasive (Tl), grade III 16
Muscle-invasive (T2-T4), grade HI 66
70
transluminator (302nm, Model TM-36, 8000 UW/cm2, UVP, San Gabriel, CA) and
exposed to UV light for 30 min to completely fragment the DNA in the ink-
unprotected tissue. Ink-protected tissue dots were scraped with a sterile scalpel into a
microfuge tube and processed as described above.
Analysis of allelic loss on chromosomes 9 and 17p
Tumor DNA was examined for genetic alterations at seven separate
microsatellites, five localized on chromosome 9 (D9S59, D9S63, D9S64, D9S146,
D9S156), one on chromosome 17p (D17S513), and one on the X chromosome
(androgen receptor gene locus). Analysis of the microsatellite in the androgen
receptor gene was performed by Dr. Yvonne Tsai. Loci D9S59, D9S63, and D9S64
were analyzed for 156 tumors (Table 3.2); loci D9S146 and D9S156 were analyzed
for 49 tumors; locus D17S513 was analyzed for 90 tumors; the androgen receptor
gene locus was analyzed for 25 tumors. Dinucleotide repeat polymorphisms of the
form (GT)n on chromosomes 9 and 17p were analyzed by PCR amplification as
described (Kwiatkowski et al., 1992; Oliphant et al., 1992). PCR amplifications were
performed in 25 pi volumes, containing 10 mM Tris-HCl (pH 8.3), 50 mM KC1, 1.5
mM MgCl2,1 pM primers (each), 200 pM of each deoxynucleoside triphosphate, 0.2
pCi [a-3 2 P]dCTP (3000 Ci/mmol, ICN Biochemicals, Costa Mesa, California), and
0.625 u of Taq polymerase (Boehringer Mannheim, GmbH, W. Germany). PCR
conditions were: 94°C for 1 min, 28 cycles at 94°C for 1 min, 56°C for 40 s, and 72°C
for 40 s, followed by incubation at 72°C for 2 min. PCR products from matched
normal and tumor DNAs were analyzed by electrophoresis in denaturing 8%
polyacrylamide-7M urea gels. Gels were dried in a gel dryer (Model 583, Bio-Rad,
71
Table 3.2. Microsatellite alterations in TCCs of the bladder
Locus
analyzed
# of tumors
examined0
# o f tumors showing
m icrosatellite
alterations
D9S59 156 1
D9S63 156 2
D9S64 156 1
D9S146 49 3
D9S156 49 3
D17S513 90 0
Androgen
receptor gene
25 1
a Several of the total 200 different tumors were examined for more
than one locus.
72
Richmond, California) and exposed to autoradiogram film (Hyperfilm-MP,
Amersham, Arlington Heights, IL), using one intensifying screen, at -70°C, overnight.
For informative (i.e., heterozygous) cases, allelic loss was scored when intensity of the
signal for a tumor allele was significantly reduced (by > 95%) relative to the matched
normal allele.
The sequences of primers used are: locus D9S59, 5'-TTA CAC TAT ACC
AAG ACT CC-3' and 5-AAG GGA ATT CAT CCC CTG CT-3'; locus D9S63, 5 -
TTA TAA TGC CGG TCA ACC TT-3’ and 5-CCG GAA GTT ACT CTA GTC TA-
3’; locus D9S64, 5 -GAA GGG CTC TTT ATT AAC TGA T-3' and 5'- AAC CTG
GGC GAC ACA GCA A-3'; locus D9S146, 5-TGC AAT CAA ATT CCC AGC-3’
and 5 -GAG GTG ACA TCT GGA ATT-3'; locus D9S156, 5 -ATC ACT TTT AAC
TGA GGA GG-3' and 5-AGA TGG TGG TGA ATA GAG GG-3'; locus D17S513,
5-TTC ACT TGT GGG CTG CTG TC-3' and 5-TAA GAA AGG CTC CCA CAA
GCA-3'. The trinucleotide repeat polymorphism, (CAG)n, in the androgen receptor
gene (Cutler Allen et al., 1992) was analyzed by PCR, performed in a final volume of
25 pi, containing 50 ng of genomic DNA, 1 pM of each oligonucleotide primer (5-
GTG CGC GAA GTG ATC CAG AA-3' and 5 -TCT GGG ACG CAA CCT CTC
TC-3'), 200 pM of the non-radioactive deoxynucleotides, 2 pCi of [a-3 2 P] dCTP, 10
mM Tris-HCl (pH 8.3), 50 mM KC1, 1.5 mM MgC12, 0.01% gelatin and 1 unit of
Taq DNA polymerase (Boehringer Mannhein Biochemicals, Indianapolis, IN).
Twenty-four cycles of 94°C for 1 min, 60°C for 1 min, and 72°C for 1.5 min were
performed with the initial denaturation step and final elongation step lengthened to 2
min and 3 min, respectively. PCR products were resolved on 5% polyacrylamide-7M
urea gels for 2.5 h at 60 W.
73
RESULTS
Microsatellite repeat polymorphisms were used to investigate allelic deletions
on chromosome 9 in 200 TCCs and on chromosome 17p in 90 TCCs. The
microsatellite banding patterns observed for 194 of these cases either showed no
changes between normal and tumor DNA, or loss of an allele in the tumor DNA.
Figure 3.1a shows examples of the same banding pattern being present in the DNAs
isolated from a paraffin-embedded superficial papillary grade II tumor and adjacent
normal tissue for two (GT)n repeat polymorphisms at loci D9S59 and D9S64 (Fig.
3. la, Lanes 1, 2, 5, and 6); an example of LOH detected in the same tumor DNA at
locus D9S156 is shown in Fig. 3.1a, Lane 10. However, differences between normal
and tumor DNA banding patterns were observed in the same specimen at loci D9S63
and D9S146 (Fig. 3.1a, Lanes 3, 4, and 7, 8). These differences consisted of shifts
in the electrophoretic mobilities of (GT)n dinucleotide repeat fragments, reflecting a
minor 2 bp expansion of the repeat at locus D9S63 and a >2 bp expansion at locus
D9S146. These alterations were not due to polymerase errors during PCR
amplification since results were reproducible in replicate assays, and in mixing
reactions in which tumor DNA was added to normal DNA from other patients.
Figure 3.1& shows the results obtained for patient B, whose lamina propria
invasive grade HI tumor DNA, obtained from fresh-frozen tissue, showed changes in
banding patterns at all five loci analyzed on chromosome 9 (Fig.3.1 b, Lanes 2, 4, 6, 8,
10). The alterations detected in this tumor, similarly to the changes observed in the
tumor from patient A, were of two types: a single 2 bp shift was observed at loci
74
Fig. 3.1. Dinucleotide repeat polymorphisms in normal (N) and tumor (T) tissue
from patients with transitional cell carcinoma of the bladder. The microsatellite
markers are located at loci D9S59, D9S63, D9S64, D9S146, and D9S156 in
chromosome 9. Each normal allele is represented by a major band surrounded by
several other lighter bands, (a) (GT)n repeat polymorphisms in patient A. The
superficial low grade papillary tumor DNA in lanes 4 and 8 showed microsatellite
abnormalities at loci D9S63 and D9S146, respectively. Loci D9SS9 and D9S64 in
lanes 1,2 and 5,6 respectively showed the same banding pattern for normal and tumor
DNA. Locus D9S156 revealed LOH in tumor DNA in lane 10. (b) (GT)n repeat
polymorphisms in patient B. The lamina propia invasive grade in tumor DNA
showed alterations in all five loci examined in chromosome 9. Lanes 2 and 8 show a 2
bp shift at loci D9SS9 and D9S146, respectively. Lanes 4, 6 and 10 contain tumor
DNA presenting larger alterations of the allele sizes at loci D9S63, D9S64 and
D9S156.
75
A
NT N T N T N T N T
1 2 3 4 5 6 7 8 9 10
D9S59 D 9S63 D9S64 D9S146 D 9S156
B
NT NT N T N T N T
i
J 2_ 3______4 5_____6 7 8 9 10
D9S59 D9S63 D9S64 D9S146 D 9S156
76
D9S59 and D9S146 (Fig. 3. lb, Lanes 2, and 8), and larger alterations (>2 bp) were
detected at loci D9S63, D9S64, and D9S156 (Fig. 3.1 b. Lanes 4, 6, and 10). To rule
out the possibility of specimen contamination or sample switching, we obtained and
analyzed paraffin-embedded tumor and normal material from patient B. When results
from the paraffin-embedded material were compared with those obtained from the
fresh-frozen tumor the same alteration was detected in DNA from both sources of
tumor tissue.
Figure 3.2 shows the results of the analysis of a trinucleotide repeat
polymorphism in the androgen receptor gene obtained for patients C and D. The
tumors from these patients revealed changes in the (CAG)n repeat at the androgen
receptor gene locus. Fig. 3.2a shows that the two tumor specimens from patient C,
superficial (Ta) grade ID and lamina propria invasive (T1) grade HI, contained a new
shortened 275 bp allele that was not present in the normal tissue DNA. These
results for patient C were confirmed in tumor DNA obtained from paraffin-embedded
material. The new fragment represented a deletion within the trinucleotide repeat
element. Sequencing of each of the three alleles revealed that the first two, present in
the normal DNA, contained 24 and 17 CAG repeats respectively, while the third, new
allele contained only 9 CAG repeats. In contrast, an expansion within the trinucleotide
repeat element is shown in Fig. 3.2b for the tumor obtained from patient D.
Table 3.2 presents the different microsatellites examined in this study,
indicating the number of TCC analyzed for each microsatellite and the proportion of
tumors that contained instability at each microsatellite. Loci D9S146 and D9S156,
located on chromosome 9p, showed the highest frequency of instability (6%),
77
Fig. 3.2. Trinucleotide repeat polymorphisms in the androgen receptor gene in the X
chromosome in TCC patients. This analysis was performed by Dr. Yvonne Tsai, (a)
(CAG)n repeat polymorphism in patient C. Allele sizes are indicated in base pairs.
Two alleles of 320 and 299 bp are present in the normal DNA (N). A new, truncated
allele of 275 bp appears in the superficial grade m tumor DNA (Ta) and in the lamina
propria invasive grade m tumor DNA (Tl). (b) (CAG)n repeat polymorphism in
patient D. The tumor DNA (T) shows a major expansion (>2 bp) within the
trinucleotide repeat. This specimen also showed alterations in (GT)n repeats on
chromosome 9.
78
N Ta T1
9
t
•
1 •
, - ; V ' *
«
• •
. 320 bp
299 bp
- 275 bp
79
D
N T
8 0
followed by the trinucleotide repeat in the androgen receptor gene (4%). The markers
examined on chromosome 9q (D9S59, D9S63, D9S64) were shown to be altered in
only 0.6 to 1.3% of the tumors. In contrast to the results obtained for chromosome 9
and the androgen receptor gene in the X chromosome, the analysis of locus D17S513
in chromosome 17p did not reveal any microsatellite alteration in 90 TCCs analyzed.
Table 3.3 summarizes the stages and grades as well as the genetic alterations
detected in the tumors in which microsatellite changes were found. All six RER+
tumors were low stage (Ta-Tl), grades n-in, with three tumors (from patients C and
E) showing alterations in only one of the seven loci analyzed, while three other tumors
(from patients A, B, and D) revealed alterations in more than one locus.
DISCUSSION
Our data showed thar genomic instability as measured by changes in
microsatellite repeats occurs in TCC of the bladder. Since alterations were detected in
low-stage TCCs, including two low-grade tumors, genomic instability might be an
early event in bladder tumorigenesis. The low number of bladder tumors in which we
observed microsatellite changes could be due to the fact that only seven markers were
analyzed, and five of them were located on the same chromosome. Thus, our results
may reflect only a small part of a genome-wide instability in bladder cancer. Similar
results were reported by Linnebach et al (Linnebach et al., 1994), who extended the
microsatellite analysis to other chromosomes and detected frequent instabilities on
81
Table 33. Bladder TCCs showing microsatellite alterations
Patient Tumor Tumor # Loci Altered/# Names of Loci Altered Microsatellite Alteration
Stage Grade Loci Examined
i
A Ta n 2/7 D9S63, D9S146 Minor and major
expansions of (GT)n repeat
B Tl m 5/7
D9S59, D9S63, D9S64,
D9S146, D9S156
Minor and major
expansions and contractions
of (GT)n repeat
c
Ta ID 1/7
Androgen receptor gene
in X chromosome
Major truncation
of (CAG)n repeat
Tl m
1/7
Androgen receptor gene
in X chromosome Major truncation
of (CAG)n repeat
D Ta n 5/7
D9S63, D9S64, D9S146,
D9S156, Androgen
receptor gene
Minor and major
contractions of (GT)n repeat;
major expansion of (CAG)n
repeat
E Tl H I 1/7
C
D9S156 Major expansion of
(GT)n repeat
chromosome 2. Orlow et al. (1994) observed this event in 41% of TCCs, which
suggests that microsatellite instability is not a random alteration in bladder cancer.
The tumor obtained from patient B was particularly interesting in that it
contained microsatellite changes in all five loci analyzed on chromosome 9. This
specimen was obtained from a patient who had a history of a kidney TCC and a
ureteral TCC resected seven and three years, respectively, before diagnosis of the
tumor analyzed in our study, suggesting an association between somatic instability in
chromosome 9 and susceptibility to multiple primary TCCs.
The data obtained for patient C were informative in relation to the timing of the
instability in tumorigenesis since the superficial (Ta) tumor showed LOH for
chromosome 9 whereas the lamina propria-invasive (Tl) tumor, which was excised 10
months later, had retention for this chromosome (Miyao et al., 1993). Because both
tumors contained the new truncated trinucleotide repeat fragment in the androgen
receptor gene, it is likely that they were derived from the same transformed cell and
that the allelic loss in chromosome 9 occurred after the instability, resulting in a faster
growing tumor which was detected earlier. The development of a shortened CAG
repeat allele in the androgen receptor gene may be a manifestation of un-repaired
replication errors (RER) and genomic instability associated with transformation. RER
during tumor development may also result in expansions of trinucleotide repeats, such
as those which occurred in the tumor from patient D. In contrast to tumors from
patient C, the tumor from patient D contained an expansion of the (CAG)n repeat in
the androgen receptor gene, and also showed expansions and deletions at (GT)n
repeats on chromosome 9.
83
Genetic alterations similar to the ones detected in our study have been reported
in virtually all tumors that occur in patients with hereditary nonpolyposis colorectal
cancer (HNCC) (Ionov et al., 1993; Aaltonen et al., 1993; Peltomaki et al., 1993).
Microsatellite changes have not only been linked to familial predisposition to colon
cancer, but have also been detected in sporadic colon carcinomas (Aaltonen et al.,
1993; Peltomaki et al., 1993). The changes in repeat lengths that we detected in
bladder carcinomas were of two types: major alterations (>2 bp) and minor alterations
(2 bp change) in the repeat fragment size, similar to the changes observed in colon
cancer (Peltomaki et al., 1993). The tumors in which we detected microsatellite
alterations could be grouped according to the number of loci affected: 3 tumors
showed DNA alterations at multiple loci, and 3 tumors showed alterations at only 1
locus.
Our findings that microsatellite instability is present as an apparently early
event in the development of bladder cancer, for which a hereditary predisposition has
never been described, suggested that this kind of genetic alteration might be common
to sporadic human cancers and that RER+ tumors might not be unique to HNPCC
families. In fact, since our initial report on microsatellite instability in bladder cancer
(Gonzalez-Zulueta et al., 1993), other sporadic tumor types, such as gastric and
ovarian cancers, have been shown to manifest similar genomic instability (for reviews,
Modrich, 1994; Loeb, 1994). The. degree of microsatellite instability detected in
bladder cancer and other sporadic tumors is low compared to that observed in
sporadic colon cancer and HNPCC cases. It has been shown (Modrich, 1994) that
mutations in mismatch repair genes, or MMR genes, including hMSH2, hMLH2,
hPMSl, and hPMS2 genes, can cause HNPCC. A new member of the MMR family
has been recently isolated, the G/T binding protein or GTBP (Drummond et al., 199S;
Palombo et al., 1995; Papadopoulos et al., 1995). The GTBP gene maps to human
chromosome 2pl6 within 1 megabase of the hMSH2 gene, and codes for a 160-
kilodalton polypeptide that-binds to G/T mismatches as a heterodimer with hMSH2.
Papadopoulos et al. (1995) have shown that while germline mutations of hMSH2
account for about 50% of total HNPCC cases, germline GTBP mutations in HNPCC
are rare. GTBP mutations may therefore not cause sufficient genetic instability to
result in predisposition to tumor formation. The observation that the genomic
instability observed in cells with GTBP deficiency is less severe than in cells with
other MMR gene defects raises the possibility that the low degree of microsatellite
instability detected in certain sporadic cancers, including bladder cancer, may be
associated to alterations in GTBP rather than in other MMR genes. Still, it remains to
be determined whether GTIiP defects play a role in instability in bladder and other
cancer types. If so, it would raise the possibility that the molecular defects responsible
for genetic instability could be predicted from the spectra of mutations in the tumors.
Such predictive power would have important implications for cancer diagnosis and
treatment.
85
CHAPTER4
ALTERATIONS OF THE pl6/CDKN2 GENE IN BLADDER TRANSITIONAL
CELL CARCINOMA: HOMOZYGOUS DELETIONS AND POINT
MUTATIONS ARE INFREQUENT MECHANISMS OF INACTIVATION
INTRODUCTION
LOH at markers on chromosome 9 occurs in approximately 60% of primary
superficial and invasive TCCs of the bladder, and appears to be the earliest and most
common identifiable genetic event in this cancer type (Tsai et al., 1990; Miyao et al.,
1993; Dalbagni et al., 1993; Cairns et al., 1993; Knowles et al., 1994; Yang and Jones,
1995; Gonzalez-Zulueta and Jones, 1995). An exception is carcinoma in situ where
chromosome 9 deletions are rare and p53 tumor suppressor gene mutations are
common (Spmck et al., 1994a). The majority of tumor specimens examined to date
exhibit LOH for every chromosome 9 marker tested, confirming earlier cytogenetic
observations of monosomy 9 in this disease (Perucca et al., 1990). Recent reports on
partial independent deletions on either the p or the q arm of chromosome 9 have
suggested the presence of at least two tumor suppressor genes important in TCC
tumorigenesis located on both arms of chromosome 9 (Ruppert et al., 1994; Keen and
Knowles, 1994). In a small minority of tumors, 9p LOH occurs with retention of
heterozygosity on the long arm (Ruppert et al., 1993). The minimal region of LOH
on the short arm has been mapped to band 9p21, and homozygous deletions of this
region have been demonstrated in approximately 10% of TCCs (Cairns et al., 1994a,
and bladder-derived cell lines (Stadler et al., 1994). Deletions of the 9p21
86
chromosomal region are not unique to bladder TCC and have in fact been described in
a large number of different tumors including melanoma (Fountain et al., 1992), lung
cancer (Olopade et al., 1993), mesothelioma (Cheng et al., 1993), head and neck
carcinomas (van der Riel et al., 1994), and gliomas (Dreylin et al., 1993).
Chromosomal region 9p21,0.3 cM centromeric to the interferon-alpha (IFNA) gene
cluster, is a target of frequent chromosome inversions and translocations (Diaz et al.,
1988; Lukeis et al., 1990). This tumor suppressor gene locus is unique in that a large
number of homozygous deletions have been described (Fountain et al., 1992; Merlo et
al., 1994; Coleman et al., 1994; Nobori et al., 1994).
Recently, a candidate tumor suppressor gene, p!6/CDKN2 (also pl6IN K 4 or
MTS1), was localized to chromosome 9p21 (Kamb et al., 1994; Nobori et al., 1994).
The Multiple Tumor Suppressor 1 (MTS1) gene was isolated from genomic clones
from a wide variety of cancer-derived cell lines which mapped to the smallest common
region of homozygous deletion within band 9p21. This gene encodes a 16-kilodalton
protein that contains four conserved ankyrin motifs and functions as a specific
inhibitor of the cyclin D-CDK4 and -CDK6 protein kinase complexes (Serrano et al.,
1993). Expression of the D-type cyclins and their associated kinases are regulated by
extracellular signals, and their transcription is induced by stimulation of cells with
mitogens (Sherr, 1993). The cyclin D-associated kinases integrate extracellular
signals with cellular DNA synthesis by phosphorylating proteins that control rate-
limiting functions in mid to late G1 (Sherr, 1993). One of the critical events
controlled by cyclin D-cyclin dependent kinases complexes appears to be the initial
inactivating phosphorylation of the retinoblastoma protein, pRB, which triggers the
release of pRB-associated transcription factors that are necessary for S-phase entry
87
(Fig. 4.1) (Matsushime et al., 1992). By specifically inhibiting cyclin D-CDK4
activity, p!6/CDKN2 would be predicted to limit the growth-promoting activity of
mitogenic signals.
m
Bi-allelic, or homozygous, deletion of p!6/CDKN2 has been identified in SO to
85% of cell lines derived from human tumors of skin, bladder, lung, breast, kidney,
head and neck, gliomas, leukemias (Kamb et al., 1994a; Nobori et al., 1993; Okamoto
et al., 1994). Furthermore, melanoma cell lines that did not contain pl6/CDKN2
homozygous deletions did show intragenic mutations in this gene (Kamb et al.,
1994a). Also, a nonsense mutation in pl6/CDKN2 was found in a lymphoblastoid
cell line derived from a patient with dysplastic nevus syndrome, which is believed to
represent a precursor of malignant melanoma (Nobori et al., 1994).
A second related cyclin D-CDK4 specific inhibitor, pl5IN K 4B (or MTS2), is
located adjacent to pl6/CDKN2 on chromosome 9p21 and is codeleted in a high
proportion of established human cancer cell lines (Kamb et al., 1994). Transcription
of the pl5IN K 4B gene is rapidly induced in response to stimulation of cells with
transforming growth factor (3 ; thus, its inhibition of cyclin D-dependent kinase activity
appears to play a critical role in the growth inhibitory signals initiated by this cytokine
(Hannon et al., 1994).
Like p i6, p21 is also a presumptive negative regulator of the cell cycle which
has been proposed to act by blocking the G1 to S transition through inhibition of
cyclin-dependent kinases (Xiong et al., 1993). The expression of p21 is upregulated
by the p53 tumor suppressor gene, a gene that is mutated in approximately 50% of
88
TGF-B
GROWTH BLOCKED
G1/S TRANSITION
Fig. 4.1. Mechanism of action of p i6. The pl6/CDKN2 gene product,
p i6, is a cell cycle negative regulator that functions at the G1 to S
transition. pl6 binds to cdk4 and inhibits complex formation between
cdk4 and cyclin D. Decreased cyclin D-cdk4 activity results in deficient
phosphorylation of pRB and deficient release of transcription factors
(TF) that stimulate entry into the S phase of the cell cycle.
89
human cancers (Harris, 1993). Because the p53 gene is so frequently mutated, it may
play a leading role in tumorigenesis by regulating a variety of CDK inhibitors
including p21 and pl6.
In our study we sought to determine whether homozygous deletion or
mutational inactivation of pl6/CDKN2 contributes to TCC formation or progression
since the chromosomal location of pl6/CDKN2 to a frequently altered region in
bladder TCC, and p!6/CDKN2 role as a negative regulator of the cell cycle make it an
attractive candidate tumor suppressor gene important in bladder tumorigenesis. In
addition, we considered that examination of tumors and cell lines for mutations in
other tumor suppressor genes may provide important information about the pathways
of tumorigenesis through which pl6/CDKN2 acts. Therefore, we analyzed uncultured
bladder TCCs and bladder cancer-derived cell lines for homozygous deletions and
intragenic point mutations in pl6/CDKN2, and, additionally, we conducted a
comparative analysis with p53 gene mutations.
MATERIALS AND METHODS
Tumor specim ens, cell lines, and DNA isolation
Forty fresh-frozen TCC specimens obtained from the USC/Norris Cancer
Center, and thirteen bladder cancer-derived cell lines were used in this study. Tumor
specimens included: 1 1 superficial papillary TCCs (Ta), 9 lamina propria-invasive
TCCs (Tl), and 19 muscle-invasive TCCs (T2-4). The LD series of cell lines was
90
established in Dr. Jones laboratory by Dr. Louis Dubeau and Ms. Mary Peer. The
remaining cell lines were obtained from the American Type Tissue Culture Collection
(Rockville, MD). DNA was isolated from each sample by proteinase K digestion and
phenol/chloroform extraction as described (Bell et al., 1981). All DNA isolations
were performed by Ms. Mary Peer and Dr. Yvonne Tsai.
Determination of authenticity of the LD cell line series
The origins of the LD series of cell lines established from fresh tumors in our
laboratory was confirmed by analysis and comparison of highly polymorphic
microsatellite markers in DNA from each cell line, the tumor from which the line was
derived, and the white blood cells of the patient from whom the tumor was obtained.
Microsatellites used were D9S171, D9S1S6, and GSN located on chromosome 9, and
D17S153 on chromosome 17. PCR amplifications and analysis were performed as
described in Chapter 2. An example of this analysis is shown in Figure 4.2.
Analysis of point mutations in the pl6/CDKN2 gene
Exon 2 of the pl6/CDKN2 gene was screened for point mutations by Single
Strand Conformation Polymorphism analysis (SSCP analysis) and direct DNA
sequencing. Only exon 2 was examined because it constitutes 66% of the
pl6/CDKN2 coding sequence and most of the mutations described in this gene have
been observed within exon 2 (Kamb et al., 1994). SSCP analysis was performed by
Dr. Charles Spruck. For SSCP analysis 20 ng of DNA from each sample were
subjected to two different primary rounds of PCR in order to examine the S'-, middle,
or 3'-segments of exon 2. For analysis of the S'-segment, the primers used in the first
91
I ________________II________________II_________________I
D9S156 D9S171 GSN
Fig. 4.2. Verification of the origin of the LD cell line
series. This series of cell lines was derived from cultured
tumors in Dr. Jones' laboratory by Ms. Mary Peer and Dr.
Louis Dubeau. In order to exclude any error, cross
contamination between lines, or sample switching,
polymorphic microsatellite markers on chromosome 9
(D9S156, D9S171, GSN) were amplified and compared in
DNA isolated from cell line, the tumor (T) from which it
was derived, and the normal blood lymphocytes (N) from
the patient from whom the tumor was obtained.
92
round of PCR were: 5'-GTG GGG TGC TTG GCG GTG AG-3' (in intron 1) and 5’ -
GGT ACC GTG CGA CAT CGC-3' (in exon 2); the primers used in the second
round of PCR were: 5 -CAT TCT GTT CTC TCT GGC AGG-3' and 5-CAC CAG
CGT GTC CAG GAA-3'. ‘For the middle and 3'-segments, the first round of PCR
was common and used the following primers: 5-GAC CCGTGC ACG ACG CT-3'
within exon 2 and 5-TGA GCT TTG GAA GCT CTC AG-3' in intron 2. The second
PCR round for the middle fragment used the primers 5-GAC CCG TGC ACG ACG
CT-3' and 5-GGT ACC GTG CGA CAT CGC-3'. For the 3'-segment, the primers
used in the second PCR round were 5-TGG ACC TGG CTG AGG AG-3' and 5'-
CAA ATT CTC AGA TCA TCA GTC CTC-3'. Conditions for the first
amplifications were 94°C for 2 min, 17 cycles of 94°C for 1 min, 55°C for 45 sec, and
72°C for 1 min, followed by incubation at 72°C for 2 min. Conditions for the second
amplifications were similar to those used for the first amplifications except that the
annealing temperature was 58°C, and 28 amplification cycles were performed. 2 fxCi
[a-32P]dCTP (3000 Ci/mmol, ICN Biochemicals, Costa Mesa, California) were
included in the second amplifications. Radiolabeled secondary PCR products were
diluted 1:5 with Sequenase Kit stop solution (US Biochemicals, Cleveland, OH),
boiled for 5 min, and immediately loaded onto nondenaturing gels (6%
polyacrylamide, 10% glycerol, and lx Tris-Borate-EDTA buffer). Electrophoresis
was performed at 30W constant power at room temperature with a cooling fan for
approximately 7 h. Gels were dried for 15 min, and exposed to X-ray film
(Amersham Hyperfilm MP, Arlington Heights, IL) with an intensifying screen at
-70°C overnight. Each reaction was performed at least twice.
93
Exon 2 of p!6/CDKN2 was sequenced in all samples. DNA sequencing was
performed by Dr. Atsuko Shibata. DNA was first amplified without labeled
deoxynucleoside triphosphate using the same primers as indicated for the primary
amplifications in the SSCP analysis and the amplification products were subjected to
electrophoresis in 2% agarose gels and further purified using Mermaid (Bio 101, La
Jolla, CA). Sequencing of the purified DNA fragments was performed bidirectionally
with the same primers used in the second PCR amplifications in the SSCP analysis
and the Sequenase version 2.0 kit (US Biochemicals, Cleveland, OH). Sequencing
reaction products were resolved in 8% polyacrylamide gels containing 7M urea, and
were visualized by autoradiography.
When base changes were detected by SSCP and direct sequencing, exon 2 of
pl6/CDKN2 from the matching normal blood lymphocytes or muscle tissue was also
sequenced to determine whether the observed base changes were true mutations or
polymorphisms.
Analysis of homozygous deletions of the pl6/CDKN2 gene
Detection of homozygous deletions of d!6/CDKN2 in cell lines. pl6/CDKN2
homozygous deletions were initially assessed by the absence of a detectable PCR
product as determined by SSCP analysis. All potential deletions were confirmed by
trying to amplify pl6/CDKN2 under the conditions described above with the
exception that the number of cycles was increased to 20 in the primary PCR, and to 40
in the secondary PCR. Additionally, homozygous deletions were further confirmed in
some of cell lines by Southern blot analysis. The integrity of the DNA from the lines
94
with p!6/CDKN2 homozygous deletions was verified by PCR amplification of
polymorphic microsatellite markers located on chromosome 9.
Detection of homozygous deletions of d!6/CDKN2 in uncultured tumors.
Homozygous deletions in primary, or uncultured, TCCs were examined by PCR
amplification using primers complementary to intronic sequences flanking exon 2 of
pl6/CDKN2. Primer sequences were: 5'-GTG GGG TGC TTG GCG GTG AG-3'
(sense) and 5-TGA GCT TTG GAA GCT CTC AG-3' (antisense). 100 ng of DNA
were amplified in 25 pl-volume reactions containing the same reagents as described
above. PCR conditions consisted of 1 step of 94°C for 3 min, 22 cycles of 94°C for
lmin 15 sec, 58°C for 30 sec, and 72°C for 40 sec, followed by incubation at 72°C for
2 min. An internal control was included for each sample amplification, which
consisted in the amplification of the same amount of DNA (100 ng) from each sample
with primers specific for the GSN locus on chromosome 9q (Chapter 2; Kwiatkowski
et al., 1992). Control amplifications were performed for 22 cycles under conditions
described by Kwiatkowski et al. (1992). PCR products were separated on 1.8%
agarose gels, transferred onto Zetaprobe nitrocellulose membranes (BIO-RAD,
Hercules, CA) in 0.4 N NaOH, and probed using either the full length pl6/CDKN2
cDNA or a 30 mer poly-GT oligonucleotide. The plasmid blueScript (pSK)
containing the pl6/CDKN2 cDNA was kindly provided by Dr. David Beach.
The optimal PCR conditions with regard to DNA template input and number
of cycles were initially established based on achieving amplifications that were within
the linear portion of the PCR. Linearity of the PCR amplifications was determined in
independent experiments by increasing amounts of input DNA and increasing number
95
of cycles as shown in Figure 4.3. A linear relationship between the intensities of the
PCR products, after hybridization to the appropriate a -32P-labeled probe, and the
amount of input DNA or number of cycles was established. Homozygous deletions
of pl6/CDKN2 were scored when the relative intensity of the pl6/CDKN2 product
was <10% of the marker control intensity.
The purity of each tumor was estimated by histopathological examination of
specimen sections or by LOH analysis of microsatellite markers on chromosome 9.
Most of the p!6/CDKN2 homozygous deletions were confirmed by Southern
blot analysis performed by Ms. Felicidad Gonzales. Ten micrograms of genomic
DNA were digested with 4 units/pg of the restriction endonucleases EcoRl or Hindm,
electrophoresed on 0.7% agarose gels, transferred in 0.4 N NaOH onto nitrocellulose
membranes, and hybridized to radio-labeled pl6/CDKN2 cDNA. Membranes were
washed once in 2x SSC, 0.1% SDS at room temperature for 13 min, three times in
0.2x SSC, 0.1% SDS at 45°C for 15 min, and once in O.lx SSC, 0.1% SDS at 50°C
for IS min.
Analysis of mutations in the p53 tumor suppressor gene
The p53 gene mutational status had been previously determined in Dr. Jones'
laboratory for most of the tumors analyzed in this study. DNA from the cell lines and
some selected tumors was analyzed by SSCP with primers specific for exons 5
through 8 of the p53 gene as described (Spruck et al., 1994a). Briefly, exons 5
through 8 were individually amplified in two serial PCR rounds using primers
96
Fig. 4 3 . Autoradiographs showing the optimization of the PCR-based
assay used to detect homozygous deletions of pl6/CDKN2 in uncultured
TCCs. Top, PCR-cycle curve for pl6/CDKN2 amplification indicating a
linear range of amplification through 27 cycles. 100 ng of DNA input
were used in each reaction with increasing number of cycles. The
number of cycles is indicated at the top. ND, no DNA control. 22 cycles
were used in the screening of TCCS for pl6/CDKN2 homozygous
deletions. Bottom, PCR-cycle curve for GSN locus amplifications used
as a control for the pl6/CDKN2 amplifications. A linear range of
amplification is evident though 27 cycles with 100 ng of input DNA in
each reaction. 22 cycles were used in GSN amplifications from TCCs.
97
complementary to intronic sequences. Radiolabeled secondary PCR products were
diluted 1:8 in loading buffer and separated on non-denaturing 6% polyacrylamide gels
containing 10% glycerol. Samples showing aberrant migration patterns in SSCP gels
were sequenced using procedures identical to that described above for the
p!6/CDKN2 gene and to that described in Spruck et al., 1994a.
Statistical analysis
Statistical analysis was performed using Fisher's exact test, and all p-values
obtained were two-sided.
RESULTS
Point mutations in exon 2 of pl6/CDKN2 were detected by SSCP analysis
and direct sequencing of DNA from all uncultured TCCs and cell lines. A complete
concordance between the SSCP and sequencing results was observed in all cases. A
representative example of a SSCP result is shown in Fig. 4.4. This figure shows that
while no PCR product was obtained from lines LD583, LD605, RT4, and LD71
suggesting p!6/CDKN2 homozygous deletions, an abnormal migration pattern was
detected in the DNA from line LD600, indicating a sequence change. This change
was demonstrated to be a fiAC (Asp)— >AAC (Asn) mutation at nucleotide 214
present in line (LD600) as well as in the tumor from which the cell line was derived
(Fig. 4.4). The absence of this base change in the DNA from the patient's blood
lymphocytes indicated that it was a true somatic mutation. Homozygous deletions of
98
: • : :
N LD600 T
A C 6 T ACGT ACGT
« s
i
•<-GAC (Asp) to AAC (Asn)
nucleotide position 214
Fig. 4.4. pl6/CDKN2 mutations and homozygous deletions in bladder tumor-
derived cell lines were detected by SSCP analysis (top) and direct sequencing
(bottom). The SSCP autoradiograph showed an abnormal migration pattern in
line LD600, indicating a sequence change. The absence of PCR product from
lines LD583, LD605, RT4, and LD71 indicated homozygous deletions of
pl6/CDKN2 which were later confirmed by Southern blot analysis (see
Materials and Methods). N, DNA from peripheral blood lymphocytes that
contained a wild-type pl6/CDKN2 exon 2 sequence. Direct sequencing
(bottom) of DNA from line LD600, tumor from which it was derived (T), and
normal blood lymphocytes from the patient (N) showed that a true somatic
mutation was present in the cell line and the tumor at position 214 in exon 2 of
pl6/CDKN2 (nucleotide numbering according to Serrano et al., 1993). SSCP
analysis and direct DNA sequencing were performed by Dr. Charles Spruck and
Dr. Atsuko Shibata, respectively.
99
pl6/CDKN2 in the uncultured TCCs were detected by a PCR-based assay performed
under optimal conditions for amplifications in the linear range of the reactions as
described in the Material and Methods section. p!6/CDKN2 homozygous deletions
in the cell lines were detected initially by SSCP (Fig. 4.4) and confirmed by PCR-
based assay as described above. Figure 4.5 shows a representative example of this
assay in which the intensity of the pl6/CDKN2 amplified product was compared to
that of an internal standard represented by a microsatellite marker at the GSN locus
located on chromosome 9. Figure 4.5 demonstrates that pl6/C D K N 2 was
homozygously deleted in both the LD605 cell line and the tumor from which it was
derived. pl6/CDKN2 homozygous deletions were confirmed by Southern blot
analysis as shown in Figure 4.6.
pl6/C D K N 2 homozygous deletions and intragenic point mutations are
frequent in bladder tumor-derived cell lines
Homozygous deletions and intragenic point mutations in p!6/CDKN2 were
detected in 54% (7/13) bladder tumor-derived cell lines (Tables 4.1 and 4.2). Five of
the 13 cell lines examined contained both copies of pl6/CDKN2 deleted, and 2 of the
13 lines contained point mutations in exon 2 of pl6/CDKN2. Cell line LD600 was
found to have a QAC (Asp)~> AAC (Asn) transition mutation at nucleotide position
214 (Fig. 4.4) (nucleotide numbering according to Serrano et al., 1993). Cell line
SCABER contained a 2-base deletion at nucleotides 227-228 which results in a
frameshift. A previous study by Kamb et al. (1994a) reported 33% of homozygous
deletions in bladder cancer-derived cell lines.
100
pl6-
GSN (9q)-
|
GSN (9q)
P *
Fig. 4.5. pl6/CDKN2 homozygous deletions in uncultured TCCs were detected
by a PCR-based assay. Top, exon 2 of pl6/CDKN2 was amplified by PCR for 22
cycles using primers complementary to intronic sequences flanking exon 2. A
PCR product is detected in the blood lymphocytes DNA (N) after probing with
radio-labeled pl6/CDKN2 cDNA; no product was observed in the tumor (T)
obtained from this patient nor in the line LD605 derived from the tumor, indicating
pl6/CDKN2 homozygous deletions. Middle, PCR amplification of the GSN locus
on chromosome 9q was used as control. These PCR products were separated on
agarose gels, blotted and probed with a polyGT-oligonucleotide. Bottom, Purity of
the tumor was estimated by LOH analysis at the GSN locus using a microsatellite
marker. These radio-labeled PCR products were separated on denaturing
polyacrylamide gels. Allelic loss is present in the tumor (T) and its derived cell
line (LD605); this allelic loss indicates a high degree of purity of the tumor
specimen with little amount of contaminating normal cell DNA.
101
pMHZIO (9q)
Fig. 4.6. Homozygous deletions of pl6/CDKN2 were confirmed by Southern
blot analysis performed by Ms. Felicidad Gonzalez. Top, radio-labeled
pl6/CDKN2 cDNA was used as a probe to confirm pl6/CDKN2 homozygous
deletions in tumor Tl and derived cell line LD605 (also shown in Fig. 4.5), and
in tumors T2 and T3. Tumor T4 retained pl6/CDKN2. N, DNA from the
corresponding normal blood lymphocytes. Bottom, a probe specific for the
polymorphic locus MHZ 10, located on chromosome 9q, was used as control for
DNA loading.
Table 4.1. pl6/CDKN2 homozygous deletions and intragenic mutations in
primary TCCs and derived cell lines.
Total number
examined
Number with
pl6/CDKN2
homozygous
deletions
Number with
pl6/CDKN2
point mutations
Percentage with
pl6/CDKN2
alterations
Cell lines 13 5 2 54%
Uncultured
tumors
40 7 1 20%
103
Table 4.2. Status0 of the pl6/CDKN2 and p53 genes in bladder cancer-derived
cell lines
Cell line pl6/CDKN2 status p53 status
LD71 HD WT
LD137 WT Mutated
LD583 HD* WT
LD600 Mutated0 Mutated
LD605 H D <* WT
RT4 HD WT
J82 WT Mutated
T24 WT° WT
HT1376 WT Mutated
HT9 WT Mutated
TCCSup WT Mutated
UMUC3 HD Mutated
SCABER Mutated WT
°WT, wild-type; HD, homozygous deletion
*The corresponding tumor also contained CDKN2 homozygous deletion
cThe corresponding tumor also contained identical CDKN2 mutation: QAC->AAC
at nucleotide 214
^The corresponding tumor also contained CDKN2 homozygous deletion
*T24 line contains a Ha-ras mutation
104
pl6/CDKN2 homozygous deletions and intragenic point mutations are
infrequent in uncultured TCCs
Homozygous deletions of p!6/CDKN2 were detected in 17.5% (7/40) bladder
TCCs, and one somatic mutation was found in one tumor (Table 4.1). Therefore,
pl6/CDKN2 alterations (20%) were 2.7 times less frequent in primary or uncultured
tumors than in bladder-derived cell lines. This difference was found to be statistically
significant (P= 0.024, two sided).
Table 4.3 demonstrates that pl6/CDKN2 alterations were found in TCCs of all
stages. All the tumors (7 of 40) containing pl6/CDKN2 homozygous deletions also
showed allelic deletions at additional loci on chromosome 9 (Fig. 4.5, Table 4.4),
which indicated monosomy 9 in these tumors. One point mutation was detected in
exon 2 of pl6/CDKN2 in one tumor. Two tumors contained a £CG (Ala)— >ACG
«
(Thr) change at nucleotide position 436, which was determined to be a natural
polymorphism since the same base change was detected in the peripheral blood
lymphocytes DNA from the corresponding patients. It is of interest that the same 436
base change and a QCA (Ala)~>ICA (Ser) change at position 373 were detected in
the blood lymphocytes DNA from two other patients whose bladder tumors were
shown to contain pl6/CDKN2 homozygous deletions. The fact that the changed
pl6/CDKN2 alleles in these patients were not maintain in their tumors suggests that
those changes provided no growth advantage to the tumor, supporting their
polymorphic nature. The 436 polymorphism occurs outside the ankyrin motifs of the
pl6 protein, and has been detected in approximately 4% of the normal population
(Kamb et al., 1994b).
105
Table 4.3. Stages of uncultured TCCs containing pl6/CDKN2 homozygous
deletions
Stage # of tumors examined # of tumors containing
pl6/CDKN2 homozygous
deletions (%)
Superficial (Ta) 1 1 2 (18%)
Lamina propria-invasive
(Tl)
9 1 (11%)
Muscle-invasive (T2-4) 19 4 (21%)
106
Table 4.4. Chromosome 9 allelic loss and status of the pl6/CDKN2 and p53
genes in uncultured TCCs
Tumor
number
Chromosome 9a pl6/CDKN2 status* p53 status
1 ret wt wt
2 LOH wt wt
3 ret wt wt
4 LOH HD wt
5 LOH wt wt
6 LOH wt wt
7 ret wt wt
8 LOH HD wt
9 LOH wt wt
10 ret wt wt
11 LOH wt MUTATED
12 LOH wt wt
13 LOH wt MUTATED
14 LOH wt wt
15 LOH HD wt
16 ret wt MUTATED
17 ret wt wt
18 LOH wt MUTATED
19 LOH wt wt
20 LOH wt MUTATED
21 LOH wt wt
22 LOH wt wt
23 LOH wt MUTATED
24 LOH HD MUTATED
25 LOH HD wt
26 LOH wt wt
27 ret wt wt
28 LOH wt wt
29 ret wt wt
30 LOH HD MUTATED
31 LOH wt wt
32 LOH wt MUTATED
33 LOH wt wt
34 LOH wt wt
35 LOH wt wt
36 LOH wt wt
37 ret HD wt
38 ret wt wt
39 ret wt wt
40 LOH MUTATED MUTATED
a ret, retention of heterozygosity; LOH, loss of heterozygosity
* wt, wild-type; HD, homozygous deletion
107
To determine whether the pl6/CDKN2 alterations present in 54% of the cell
lines were a tissue culture-induced event or whether those alterations were already
present in the initial tumors from which the lines were derived, we examined three
cases in which DNA was available from the cell line, the originating tumor and the
corresponding patient’ s peripheral blood lymphocytes (Table 4.2). The pl6/CDKN2
point mutation detected in line LD600 was also found in the corresponding uncultured
tumor but not in the patient's lymphocyte DNA (Fig. 4.4), and the p!6/CDKN2
homozygous deletions in lines LD583 and LD605 where also present in their
corresponding tumors (Fig. 4.5 and 4.6). These data suggest that the p!6/CDKN2
alterations were somatic changes that occurred in the primary tumors and were
maintained in the derived cultures.
Comparative analysis with p53 gene status
The status of the pl6/CDKN2 and p53 genes were examined and compared in
each of the bladder cancer-derived cell lines (Table 4.2) and uncultured tumors (Table
4.4) to investigate the pathways through which p!6/CDKN2 might act. Seven of 13
lines contained pl6/CDKN2 alterations, and also 7 of 13 lines showed p53 gene
mutations. Only two cell lines (LD600 and UMUC3) were found to have
concomitant alterations in both genes, and only the line T24, which has a mutant ti
ros gene (Capon et al., 1983), showed no alteration in either gene. The remaining ten
cell lines were found to contain an alteration in one of the two genes (Table 4.2).
Table 4.4. shows that 3 of the 8 uncultured tumors in which pl6/CDKN2 was altered
contained a concomitant p53 gene mutation.
108
DISCUSSION
A newly described class of cyclin-dependent kinase (CDK) inhibitory proteins
may provide a connection-between tumor suppression and cell cycle regulation.
Known CDK inhibitors include p21W A F 1 (El-Deiry et al., 1993), p27K IP 1 (Polyak et
al., 1994), and the more recently isolated pl6IN K 4, or pl6/CDKN2 (Serrano et al.,
1993), and pl5IN K 4 B (Hannon et al., 1994). Alterations of these inhibitors would be
expected to result in abnormal cell proliferation due to loss of important growth
control function.
Four distinct lines of evidence promote the candidacy of p!6/CDKN2 as a
tumor suppressor gene: (a) it maps to 9p21, a chromosome region frequently deleted
in many tumor types including bladder transitional cell carcinomas; (b) homozygous
deletions and intragenic mutations of pl6/CDKN2 are frequent in a wide variety of
tumor-derived cell lines (Kamb et al., 1994a; Nobori et al., 1994); (c) mutations in
p!6/CDKN2 appear to underlie some cases of familial malignant melanoma (Kamb et
al., 1994b; Hussussian et al., 1994); (d) the p!6/CDKN2 product specifically inhibits
cyclin D-CDK4 activity in vitro (Hussusian et al., 1994), suggesting that the pl6
protein may be a negative regulator of proliferation. Structural or functional loss of
pl6 could therefore lead to abnormal cell growth.
Homozygous deletions of pl6/CDKN2 were initially reported in 33% of
bladder cancer-derived cell lines (Kamb et al., 1994a); however, that observation does
not answer the question of whether pl6/CDKN2 is the target of 9p21 disruption in
primary bladder TCC. Therefore, we sought to determine whether p!6/CDKN2
109
represents the tumor suppressor gene important in bladder tumorigenesis thought to
reside on chromosome 9p21.
In our series of primary TCCs, 20% (8 of 40) were shown to contain
pl6/CDKN2 alterations. In 7 tumors both alleles of pl6/CDKN2 were deleted, and 1
tumor contained a missense mutation in exon 2 of pl6/CDKN2 (Fig. 4.4). This
intragenic mutation was the only direct evidence of the potential involvement of
pl6/CDKN2 in TCC tumorigenesis. According to Knudson's hypothesis (Knudson,
1971), if pl6/CDKN2 was the target of the deletion, tumors with a 9p LOH should
have intragenic mutations in pl6/CDKN2. However, only 1 of 29 tumors with
chromosome 9 LOH showed pl6/CDKN2 mutations. Similar results were reported
by Cairns et al. (1994b), who detected pl6/CDKN2 mutations in 1 of 15 TCCs with
chromosome 9 LOH. It is intriguing that the frequency of pl6/CDKN2 alterations in
bladder TCC (20%) is significantly lower than the 70-80% of 9p21 allelic loss
observed in these tumors. Therefore, our observations, and those by others, suggest
that pl6/CDKN2 is not the primary target of 9p deletions in primary bladder TCC and
a second locus on 9p is involved in bladder cancer formation. Alternatively, other
mechanisms of pl6/CDKN2 inactivation that do not involve deletion or intragenic
point mutations might explain the relative low frequency of pl6/CDKN2 alterations
detected in TCCs; for example, mutations in promoter or intronic sequences that affect
mRNA splicing or stability could lead to pl6/CDKN2 inactivation. The possibility
also exists that relatively small changes in pl6 levels may be important in the behavior
of dividing cells; if true, loss of a single p!6/CDKN2 allele could affect the
proliferation properties of cells. Lastly, epigenetic phenomena such as methylation
may play a role in controlling the expression of this gene. In fact, the presence of a
110
CpG island in the pl6/CDKN2 promoter and exon 1 sequences make this gene an
excellent candidate to be transcriptionally controlled by DNA methylation (see
Chapter 5).
The frequency of pl6/CDKN2 homozygous deletions and point mutations in
bladder cancer-derived cell lines was 2.7 times higher than that in primary bladder
TCCs. This difference may be an effect of the genetic changes needed for the
establishment of cell lines from tumors. Because only approximately 10% of in vitro
explanted bladder tumors result in the establishment of cell lines, specific genetic
alterations which confer a growth advantage in vitro would be expected to be over
represented in cell cultures. Perhaps tumor cells with pl6/CDKN2 alterations may
have a selective growth advantage in tissue culture, as has been reported for
neuroblastoma cells with N-myc amplification and other tumor cells with regard to
p53 (Greenblaff et al., 1994). Three cases in our study supported this notion, as each
of the tumors from which 3 cell lines were derived (LD583, LD600, and LD605)
(Table 4.2) was shown to contain an identical p!6/CDKN2 alteration as that found in
the corresponding cell line. The significantly (P=0.024) higher frequency of
pl6/CDKN2 alterations in bladder cell lines than in primary tumors raises questions
about the initial cell line-based predictions (Kamb et al., 1994; Nobori et al., 1994) that
p!6/CDKN2 may be altered in human tumors more frequently than any previously
identified gene, including the p53 tumor suppressor gene which appears to be mutated
in over 30% of human tumors. Our observations indicate that inactivation of
pJ6/CDKN2 by homozygous deletion or intragenic point mutation does not play a
common role in bladder transitional cell carcinoma tumorigenesis. However, loss of
pl6/CDKN2 function by these mechanisms may be involved in development of other
111
tumor types; for example, somatic mutations and homozygous deletions of
pl6/CDKN2 have been shown to occur frequently in not only carcinomas of the
pancreas (Caldas et al., 1994) and esophagus (Mori et al., 1994), but also in squamous
cell carcinoma of the bladder (Gonzalez-Zulueta et al., 1995; Chapter 6) which differs
significantly from bladder transitional cell carcinoma in incidence, risk factors, and
clinical and pathological characteristics.
When comparing the pl6/CDKN2 and p53 tumor suppressor gene status in
the 13 bladder cancer-derived cell lines (Table 4. 2), 10 of them showed alterations in
only one of the two genes, and only two lines contained concomitant alterations in
both genes. Both pl6 andp53 proteins are involved in cell cycle control: pl6 directly
interacts and inhibits CDK4 (Serrano et al., 1993), while p53 indirectly inhibits CDKs
through upregulation of p21, a universal CDK inhibitor (El-Deiry et al., 1994). The
exclusive occurrence of either p!6/CDKN2 or p53 alterations in 10 of 13 cell lines
suggests that inactivation of either gene may be sufficient to promote growth in vitro.
In a study of bladder and melanoma cell lines, Gruis et al. (1995) detected mutations
in both pl6/CDKN2 and p53 in 2 of 16 bladder lines and in 3 of 29 melanoma lines,
concluding that p 16 and p53 must operate in different pathways, each of which may
be important to frank malignancy (Gruis et al., 1995)
Germline mutations within p!6/CDKN2 have been detected in a proportion of
familial melanoma patients (Hussussian et al., 1994; Kamb et al., 1994b), and recent
reports contribute to enlarge the list of tumor types in which pl6/CDKN2 inactivation
appears to be involved. p!6/CDKN2 was found to be mutated in 64% (16 of 25) of
primary biliary tract cancers (Yoshida et al., 1995), and loss of p!6 expression appears
112
to correlate with invasive stage of tumor progression in sporadic melanoma (Reed et
al., 1995) and non-small cell lung cancer (Okamoto et al., 1995). However, in bladder
TCC the possibility remains that the pl6/CDKN2 maps to a hypermutable locus prone
to homozygous deletions aiid that another tumor suppressor gene(s) involved in this
tumor type resides in the same region. The only other gene identified to date in the
9p21 area of homozygous deletion is pl51 N K 4 B , although no inactivating mutations
were detected in this gene in melanoma cell lines (Kamb et al., 1994), and its role in
tumorigenesis has yet to be demonstrated.
Confirmation of pl6/CDKN2's role in bladder TCC tumorigenesis will require
not only functional assays in which introduction of a wild type copy of pl6/CDKN2
in bladder tumor cells restores cell proliferation control, but also investigation of the
role that DNA methylation may play in controlling p!6/CDKN2 expression.
113
CHAPTER 5
METHYLATION AND EXPRESSION OF TH E pl6A N D plS CELL CYCLE
REGULATORS: DE NOVO METHYLATION OF THE 5' CpG ISLAND IS A
FREQUENT MECHANISM OF INACTIVATION OF THE pl6/CDKN2
TUMOR SUPPRESSOR GENE IN TRANSITIONAL CELL CARCINOMAS
OF THE BLADDER
INTRODUCTION
The pl6 protein inhibits cyclin-dependent kinases 4 and 6 (CDK4, CDK6),
which are key regulators of the progression of eukaryotic cells through the Gi phase
of the cell cycle (Serrano et al., 1993). The pl6/C D K N 2 gene, resides on
chromosome band 9p21 (Fig. 5.1), a region frequently altered in diverse tumor types
(Fountain et al., 1992; van der Riet et al., 1994). The high frequency of p!6/CDKN2
alterations initially reported in tumor-derived cell lines (Kamb et al., 1994; Nobori et
al., 1994) and in some tumor types, such as familial melanoma, pancreatic
adenocarcinoma, and esophageal carcinoma (Hussussian et al., 1994; Caldas et al.,
1994; Mori et al., 1994), along with the role of pl6/CDKN2 in cell cycle control made
this gene an excellent candidate for a tumor suppressor. Emerging evidence suggests
that pl6/CDKN2 alterations are involved in the progression of certain tumor types
(Okamoto et al.,1995; Reed et al., 1995). Another related cell cycle regulator, the
p l5 1N K 4B gene, is also located on 9p21 (Fig. 5.1) and it has been shown to be
upregulated in human keratinocytes following treatment with transforming growth
factor-P (TGF-P), suggesting that it may be an effector of TGF-P-mediated cell cycle
arrest (Hannon et al., 1994). Although most of the pl6/CDKN2 homozygous
114
9q 9p
21
m
2 i
telomere centromere
10Kb 20Kb
E 1 E2
E1P
E1a E2 E3
-p l5INK4B 1 ! -----------P16/CDKN2-
Fig. 5 . 1 . Genomic organization and location of pl6/CDKN2 and pl5IN K 4 B on chromosome 9p21. Diagram
adapted from Mao et al., 1995. Coding exons are designated as black rectangles (pl6/CDKN2) and grey
rectancles (pl5,N K 4B ). E la, exon 1 of pl6/CDKN2, where the transcript translated into pl6 protein is
initiated; E2, exon 2; E3, exon 3; Eip, coding sequence which initiates a novel transcript (P). It remains to be
determined whether transcript P is translated. Transcription could initiate from different promoters, or the
mRNA could be derived from a single promoter and then alternatively spliced to generate the pl6 transcript
(or a) and the p transcript (Stone et al., 1995; Mao et al., 1995).
deletions also include the pl5IN K 4 B locus, no intragenic mutations in pl5IN K 4 B have
been reported to date and its role in human tumorigenesis has yet to be demonstrated.
Hemizygous deletions of 9p21 are a frequent genetic alteration in transitional
cell carcinomas (TCC) of the bladder (Caims et al., 1994a; Stadler et al., 1994) which
suggests that a tumor suppressor gene for bladder cancer may reside in this area.
However, pl6/CDKN2 homozygous deletions and intragenic mutations occur at low
frequencies in uncultured TCCs (Chapter 4; Spruck et al., 1994b; Caims et al.,
1994b), although they are common in uncultured squamous cell carcinomas of the
bladder (Chapter 6; Gonzalez-Zulueta et al., 1995). Also, no homozygous deletions of
the pl6/CDKN2 locus have been described in colon cancer-derived lines in contrast to
a wide variety of tumor-derived cell lines (Kamb et al., 1994). This low rate and
absence of p!6/CDKN2 homozygous deletions and intragenic point mutations in
bladder TCCs and colon cancer cell lines may reflect that pl6/CDKN2 inactivation
occurs in these tumor types by alternative mechanisms. Thus, transcriptional
repression by DNA methylation of promoter and 5' regulatory sequences may be a
pathway to inactivate the pl6/CDKN2 and pl5IN K 4 B genes.
Global changes in DNA methylation patterns are known to occur during
tumorigenesis, and gene silencing has been associated with methylation of CpG
islands located in, or near, promoters and 5' regulatory regions (Hermann et al., 1994;
Issa et al., 1994). CpG islands are G+C rich regions which show a higher frequency
of CpG dinucleotides than is normally seen in the vertebrate genome, and which are
not methylated in the germline (Bird, 1986). With the exception of islands in genes
on the inactive X chromosome (Singer-Sam and Riggs, 1993), Alu and LI sequences
116
(Liu and Schmid, 1993), and some imprinted genes (Ferguson-Smith et al., 1993),
CpG islands are usually unmethylated in normal somatic cells. In contrast,
widespread methylation of CpG islands occurs in autosomal genes during oncogenic
transformation (Jones and Buckley, 1990; Laird and Jaenisch, 1994). The evidence
linking transcriptional suppression with methylation of CpG islands located in, or
near, promoters is strong. It has been shown that methylation in a number of known
CpG islands is associated with transcriptional repression (Hermann et al., 1994; Issa
et al., 1994; Ottaviano et al., 1994), as well as altered chromatin structure (Antequera et
al., 1989; Rideout et al., 1994). Furthermore, promoters silenced by DNA methylation
can be reactivated in many cases by treatment with the drug 5-Aza-2'-deoxycytidine
(5-Aza-CdR) which is a well-established inhibitor of DNA methylation (Jones and
Taylor, 1980). While it is clear that promoter methylation can be associated with
transcriptional repression, many genes show extensive methylation of highly CpG rich
sequences, such as Alu and LI elements, in non-promoter regions and still are actively
transcribed (Magewu, 1994).
Recent advances in understanding the de novo methylation process during
embryonic development include the description of "methylation centers" which initiate
the spread of methylation (Mummanemi et al., 1993; Mummanemi et al., 199S), and
the finding that Spl sites block methylation of CpG islands (Brandeis et al., 1994;
Macleod et al., 1994). However, it has been difficult to determine the factors that
govern the establishment of de novo methylation patterns because it is not always
possible to separate the extent of DNA methylation from the levels of gene
expression. In this study we have taken advantage of the fact that the pl6/CDKN2
and p 1 5IN K 4 B genes contain a high degree of homology in their exon 2 sequences yet
117
are subject to different control (Hannon et at., 1994) to determine whether cells that
had undergone de novo methylation of one gene sequences also showed methylation
of the other.
For our study we first analyzed the p!6/CDKN2 and pl5,N K 4B sequences to
determine whether they fulfill the established criteria for CpG islands (Gardiner-
Garden and Frommer, 1987). A schematic diagram of the genomic organization of
pl6/CDKN2 and pl5,N K 4 B is presented in Figure 5.1. We observed that the exon 1
coding sequences of the pl6/CDKN2 and pl5IN K 4B genes reside within 5' CpG
islands, and their exon 2 regions are 83% homologous and also constitute CpG
islands. We hypothesized that abnormal DNA methylation might be an alternative
mechanism of inactivation of the pl6/CDKN2 and pl5,N K 4B genes in bladder and
colon tumors. To test our hypothesis, we examined the methylation status and
expression levels of the pl6/CDKN2 tumor suppressor gene and the pl5,N K 4B cell
cycle regulator in normal tissues, cell lines, bladder TCCs and colon carcinomas.
Samples from normal tissues included sperm (n=l), colon epithelium from individuals
without colon cancer (n=4), bladder urothelium (n=l), kidney (n=l), and peripheral
blood lymphocytes (n=l). We also obtained 18 bladder transitional cell carcinoma
specimens, as well as normal colonic mucosa (n=10) and corresponding tumor
specimens (n=10) from patients with colon cancer, and one ulcerative colitis specimen.
The data suggest that expression of the p!6/CDKN2 tumor suppressor gene, but not
expression of the pl5,N K 4 B cell cycle regulator, is controlled by methylation of its 5’
CpG island, and that de novo methylation of this island is a mechanism for
pl6/CDKN2 inactivation in bladder transitional cell carcinomas. In contrast, the
pl6/CDKN2 5' CpG island is methylated and the gene silenced in normal colonic
118
mucosa rather than in colon carcinomas. Therefore, we present evidence for
methylation of the CpG island in this autosomal gene in normal colonic mucosa. Our
observations also have implications for understanding the mechanisms for the
establishment of de novo methylation patterns.
MATERIALS AND METHODS
Tissue samples
Sperm DNA (n=l) was isolated from donated sample. Peripheral blood
lymphocytes (n=l) and normal kidney (n=l) were obtained from a patient with
bladder cancer. Normal colon mucosa (n=4) was obtained from individuals who had
undergone coleostomy for a non-cancerous process. One ulcerative colitis specimen
was obtained from a patient affected with ulcerative colitis. Colon carcinoma
specimens (n=10) and matched normal colonic mucosa (n=10) were obtained from 10
patients with colon cancer. Colon tumors were graded according to Dukes'
classification (Crissman and Barwick, 1993), and ranged from stage B1 to C2.
Bladder transitional cell carcinomas (n=18) were obtained from 18 patients, and their
stages ranged from Ta to T4 according to Bergkvist classification (Bergkvist et al.,
1965). DNA was isolated by proteinase K digestion and phenol/chloroform
extraction as described previously (Bell et al., 1981). Most DNA isolations were
performed by Ms. Mary Peer and Dr. Yvonne Tsai.
119
Cell lines
EJ, HT1376 and J82 (bladder transitional cell carcinoma-derived cell lines)
and SW480 (colon carcinoma-derived cell line) were obtained from the American
Type Culture Collection, Rockville, MD. EJ and HT1376 lines were cultured in
DMEM medium supplemented with 10% FCS and 5% penicilin/streptomicin. J82
and SW480 lines were cultured in MEM medium supplemented with 10% FCS, 5%
penicilin/streptomicin, non-essential amino acids, and sodium pyruvate. All cell lines
were cultured at 37°C and in the presence of 5% C02- DNA and total RNA were
isolated from non-confluent exponentially growing cells.
5-Aza-2'-deoxycytidine treatment
Cell lines. Cells were plated at lOfylOO mm dish, and treated 24 h later with 5-
Aza-2-deoxycytidine (5-Aza-CdR) (Sigma Chemical Co., St. Louis, MO, USA). The
concentrations used were 10'7 M, 10-6 M, and 10 s M. The medium was changed 24
h after addition of the drug and every 3 days thereafter. RNA and DNA were isolated
9 to 12 days after treatment.
Nude rats. 3 million EJ cells were injected through laparostomy directly into
the bladder wall of 4-week old nu/nu rats with a 27-gauge needle as described
(Simoneau et al., manuscript in preparation). Bladder tumors were allowed to grow
for 8 weeks, and 5-Aza-CdR treatments initiated. Each animal received one daily
intraperitoneal injection of 600 or 900 |X g of 5-Aza-CdR for five days. Animals were
sacrified 24 h after the last treatment, and their bladders harvested for analysis of
p!6/CDKN2 expression in the EJ cells-induced tumors. The presence of bladder
120
tumors was macroscopically assesed, tumor size was determined and total bladder
RNA was isolated.
Evaluation of cell doubling time
Five pairs of 60 mm tissue culture dishes were seeded with 104 EJ cells each,
and the number of cells present in each pair of dishes was counted every day, at the
same time, for five days. The doubling time of the 5-Aza-CdR treated cells was
determined 9 to 12 days after treatment when cells had recovered from the toxic efect
of the drug and were growing exponentially .
Reverse Transcription (RT)-Polymerase Chain Reaction (PCR)
Total RNA was isolated as described (Chomczynski and Sacchi et al., 1987).
Two and a half pg total RNA was reverse transcribed using random hexamers, dNTPs
(Boehringer Mannheim, Germany), and Superscript II reverse transcriptase (GIBCO
BRL) in a 25 pi reaction volume as previously described (Eversole-Cire et al., 1993).
cDNA was amplified using primers specific for the pl6/CDKN2 gene, the pl5,N K 4B
gene, or the GAPDH gene, which was used as as a control. The primers used for
pl6/CDKN2 amplifications were located in exon 1 and exon 2 of the gene, flanking
intron 1 (Fig. 5.2). Their sequences were: 5'-AGC CTT CGG CTG ACT GGC
TGG-3' (sense), and 5 -CTG CCC ATC ATC ATG ACC TGG A-3' (antisense).
Conditions for pl6/CDKN2 amplifications were: 94°C for 3 min, 22 cycles of 94°C
for 1 min, 56# C for 30 sec, 72°C for 40 sec, followed by incubation at 72°C for 1 min.
Primers for p l5 IN K 4 B were: 5 -TGA TGA TGG GCA GCG CCC GC-3' (sense) and
5'-CGG CTG GGG AAC CTG GCG TCA-3' (antisense) (Fig. 5.2). Conditions for
121
CCC
pl6/CDKN2
p l5 INK4B
Fig. 5.2. Schematic maps of the pl6/CDKN2 and pl5IN K 4 B genes. Coding
sequences are represented as boxes in which the size of the exon in base pairs is
indicated. The methylation status was examined at the restriction endonuclease
sites indicated in the maps: FnuDU. (F), SacII (s), Hpzll (H), CfoI (C), Smal (S),
Nael (N), or Mspl (M). The primers used in the PCR-based methylation assay
flank the restriction sites and are represented as arrows above each of the genes.
The primers used in the RT-PCR expression analysis are represented as arrows
below each of the genes.
122
pl5IN K 4B amplifications were: 95°C for 3 min, 22 cycles of 95°C for 1 min 40 sec,
60’C for 45 sec, and 72°C for 40 sec, followed by incubation at 72°C for 2 min.
Primers for the GAPDH gene were: 5'-CAG CCG AGC CAC ATC G-3' (sense), and
5 -TGA GGC TGT TGT CAT ACT TCT C-3' (antisense). Conditions for GAPDH
amplifications were: 94°C for 1 min, 22 cycles of 94°C for 1 min, 58*C for 30 sec,
72°C for 45 sec, followed by incubation at 72°C for 1 min. Each PCR was performed
with an amount of cDNA equivalent to 100 ng of RNA, and PCRs for pl6/CDKN2
and p l5 IN K 4B contained 10% DMSO (final concentration). Under the above
conditions, all amplifications were in the linear range of the assay. PCR products
were resolved on 2% agarose gels. DNA was transferred to a nylon membrane
(Zetaprobe; Bio-Rad, Richmond, CA) via alkali transfer. pl6/CDKN2 blots were
hybridized to radiolabeled pl6/CDKN2 cDNA. pl5IN K 4 B blots were hybridized to a
radiolabeled or digoxigenin-labeled internal oligonucleotide. GAPDH blots were
hybridyzed to radiolabeled GAPDH cDNA. Quantitation of PCR products was
performed by scanning autoradiographs with an LKB UltraScan XL Laser
densitometer. The ratio between CDKN2/GAPDH and pl5IN K 4B /GAPDH signals
was obtained for each sample. All reactions were done at least twice, and all were
controlled with the omission of reverse transcriptase.
PCR-based methylation assay
A PCR assay relying on the inability of some restriction enzymes to cut
methylated sequences (Singer-Sam et al., 1990) was used to analyze the methylation
status of the first and second exons of the pl6/CDKN2 and pl5IN K 4 B genes. The
sites examined were: one FnuDU, one 5acII, one Hpall, and two Cfol sites in exon 1
123
of pl6/CDKN2, and two Noel sites in exon 1 of p l S ^ 48 (Fig. 5.2). Due to the high
sequence similarities (Fig. 5.3), identical restriction sites could be examined in exon 2
of p!6/CDKN2 and pl5IN K 4 B which included one Smal, four Hpall, and six Cfol
sites. DNA digests were* performed according to the manufacturer directions
(Boehringer Mannheim, Indianapolis, Ind.). DNA (1 pg) was digested for 2 h, with
10 U of enzyme per pg of DNA. 50 ng of the digested DNA was amplified with
primers flanking the restriction sites (Fig. 5.2). The primer set used for methylation
analysis of pi6/CDKN2 exon 1 was: 5 -AGC CTT CGG CTG ACT GGC TGG-3'
(sense) and 5'-CTG GAT CGG CCT CCG ACC GTA-3' (antisense), under the
following conditions: 94°C for 3 min, 21 cycles of 94°C for 1 min, 55°C for 30 sec,
72°C for 40 sec, followed by 72°C for 1 min. Primers used for pl6/CDKN2 exon
2were: 5 -CTG CTT GGC GGT GAG GGG G-3' (sense), and 5 -CCT CAC CTG
AGG GAC CTT C-3' (antisense); conditions were the same as for exon 1 except for
an annealing temperature of 57°C. The primer set used for methylation analysis of
p l5iN K 4B exon 1 was: 5'-AGC TGA GCC CAG GTC TCC TA-3’ (sense), and 5'-
CGC CTCCCG AAA CGG TTG AC-3’ (antisense). Conditions were: 95°C for 3
min, 21 cycles of 94°C for 1 min 40 sec, 58°C for 45 sec, 72°C for 45 sec, followed by
72°C for 2 min. Primers for pl5IN K 4 B exon 2 were: 5'-TCT TTA AAT GGC TCC
ACC T-3* (sense) and 5'-CTC CCC GTT GGC AGC CTT C-3’ (antisense);
conditions were the same as for pl51 N K 4 B exon 1 amplifications. Under the indicated
PCR conditions, amplifications with each primer set were in the linear range of the
assay as determined by cycle curves and DNA concentration curves performed to
establish the optimal conditions of the assay (Fig. 5.4). PCR products were resolved
on 2% agarose gels, blotted by alkali transfer, and hybridized as described for the RT-
PCR products.
Fig. S3. Nucleotide sequence of exon 2 and flanking intronic sequences in
the pl6/CDKN2, (or MTS1, top) and pl5lN K 4B, (or MTS2, bottom) genes.
The first 260 nucleotides of exon 2 of pl6/CDKN2 are 99% homologous to
exon 2 of pl5lN K 4 B . The start and the end of both exons are indicated by
arrows.
125
tfTSl (co p ), MTS2 (bottom)
10 20 30 40 SO 60 70
0 I 1 1 I I
D | I - | MGCCQGCMCTC
10 20 30 40 SO 60 70
80 90 100 110 120 130 140 ISO
1 I l» I I I I I I
----------------------------------- ^ijQ ^TQ Q Q C A G CG C C C G R G TG G C Q tjA G C TG C TG C TtX rrC C A C G G C G C G G A G C C C
i i u i n
80 90 100 110 120 130 140 ISO
160 170 180 190 200 210 220 230
I I I I I I I I
160 170 180 190 200 210 220 230
240 250 260 270 280 290 300 310
I______i______ i ______ I I I I_________ I
I I I I I I I I
240 2S0 260 270 280 290 300 310
320 330 340 350 360 370 380 390
1 ^ I rL rrrrr yJtym gllO tM V
• • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • « • • • • • • • • • •
G C G ftDjTTG C A O G G TftC CTCC G C A C A G C C A C G G G G G ftC TG A CG O C A figriC O O C R G C C G C C C A C A A C G A CT ra vr r n v r
I I I I J I I I >
320 330 340 350 ^ 360 370 380 390
400 410 420 430 440 450
1 I I I I
CCCICAGGrTGAGGACTGATGATCTGAGAAnTCjTACYCTGAGAGCTrCCAAAGCTCA
• •• • • • • • • • • • • | «
TACCCAMTI'CCCACCCCCACCCACCTRATTCGATGAAGGCTGCCAACCSGtXJAGCGG
I I I I I I
400 410 420 430 440 450
126
18 2 0 22 24 2 6 28
WMU MU MU M U M U M U M
# of cycles
pl6/CDKN2
Fig. 5.4. Representative example of the optimization of the PCR-based
methylation To prevent overcycling, the optimal cycle number was determined
for each of the primer sets by performing a cycle curve using undigested template
(U) and template digested with Mspl restriction enzyme (M) which is
methylation-insensitive. The optimal cycle number was determined to be that
which produced a detectable band of minimal intensity after hybridization of PCR
products, and that did not produce any signal from the Mspl-digested DNA.
Increasing number of cycles in the pl6/CDKN2 exon 2 PCR amplifications
indicated a linear range of amplification through 28 cycles. 100 ng of DNA input
were used in each reaction with increasing number of cycles. The number of
cycles is indicated at the top. WM, molecular weight marker; U, undigested
DNA; M, Afrpl-digested DNA.
127
PCR-based methylation analysis using restriction enzymes may be subject to
variability if the DNA digestions are not complete, and if amplifications are performed
for an excesive number of cycles. To rule out the possibility of incomplete restriction,
all samples were digested twice with each of the enzymes in independent experiments.
PCR amplifications from each of the duplicate digests were repeated at least twice to
ensure reproducibility of the results. To prevent overcycling, the optimal cycle
number was determined for each of the primer sets by performing a cycle curve using
undigested template and template digested with Mspl restriction enzyme which is
methylation-insensitive (Fig. 5.4). The optimal cycle number was determined to be
that which produced a detectable band of minimal intensity after hybridization of PCR
products, and that did not produce any signal from the Mspl-digested DNA. For
every site examined, an undigested DNA control and a Mspl-digested DNA control
were included.
RESULTS
Analysis of CpG distribution in exons 1 and 2 of the pl6/CDKN2 and
p !5iN K 4B g e n e s
We first analyzed the CpG distribution in the first and second exons of
pl6/CDKN2 and pl5IN K 4B . A representative example of this analysis is shown in
Figure 5.5, which presents the CpG distribution in exons 1 and 2 of pl6/CDKN2.
Table 5.1 shows that the exon 2 regions of both genes and exon 1 and the 5'UTR of
p l5 iN K 4 B ^ CpG islands based upon the criteria of Gardiner-Garden and Frommer
128
Fig. 5.5. Analysis of the CpG distribution in pl6/CDKN2. The graphs represent
the ratio between the observed and the expected number of CpG dinucleotides
(colums), and the percentage (%) of G+C (line) in overlapping windows of 20
nucleotides. The observed/expected CpG ratio was calculated according to Gardiner-
Garden and Frommer (1987) as follows:
number of CpG
number of C x number of G ) where N is the total number of nucleotides in
the sequence being analyzed. A moving average value for % G+C and for Obs/Exp
CpG was calculated for each sequence, using 20 bp window (N=20), moving across
the sequence at 15 bp intervals. According to Gardiner-Garden and Frommer, a
region of >200 bp constitutes a CpG island if it contains a moving % of G+C over
50, and a moving obs/exp CpG ratio higher than 0.6. Top, CpG distribution in exon
1 of pl6/CDKN2. Exon 1 of pl6/CDKN2 is too small to satisfy the criteria for a
CpG island, however, given its G+C content over 50% and a observed/expected CpG
dinucleotide ratio greater than 0.6, it very likely resides within an island which extents
into the promoter sequences of the gene. Bottom, CpG distribution in exon 2 and
flanking intronic sequences.
129
!
Obs/Exp CpG
pl6/CDKN2 ex o n 1 (bp)
Obs/Exp CpG
30-50
45-65
60-80
75-95
90-110
105-125
8
t
$
o
ft
0
1
■ 3
o
*
G+C
Table S.l. Analysis of CpG distribution in exons 1 and 2 of pl6/CDKN2 and
p l5IN K 4B
Sequence Examined0 %G+C Obs/Exp CpG
pl6 1 -141 nt 71.5 0.9
EXON 1
(including
5'UTR)
pl5 1 - 477 nt 70 0.93
pl6 142 - 400 nt 72 1.1
EXON 2
pl5 478 - 736 nt 72 1.1
a The nucleotide positions derive from the numbering scheme used in Serrano et
al. (1993) and Hannon etal.(1994).
1 3 1
(1987), because they are larger than 200 bp, have a G+C content over 50%, and an
observed/expected CpG dinucleotide ratio greater than 0.6. Exon 1 of pl6/CDKN2 is
too small to satisfy the criteria for a CpG island, however, given its G+C content over
50% and a observed/expected CpG dinucleotide ratio greater than 0.6 (Table 5.1),
exon 1 of pl6/CDKN2 very likely resides within an island which extents into the
promoter sequences of the gene.
Methylation and expression of the pl6/CDKN2 and p l5IN K 4B genes in primary
bladder transitional cell carcinomas and ceil lines
The methylation status of several CpG sites within the first and second exons
of pl6/CD KN2 and pl5IN K 4B was examined in normal bladder mucosa, 18
uncultured transitional cell carcinomas (TCCs) of the bladder, and 3 bladder
carcinoma-derived cell lines. The sites were analyzed by a PCR-based assay (Singer-
Sam et al., 1990) using methylation-sensitive endonucleases for DNA cleavage
followed by PCR amplification of the digested DNAs (Fig. 5.2 and 5.6). The normal
urothelium showed no methylation in either exon of either gene, as no PCR product
was generated following digestion with the corresponding enzyme (Fig. 5.6). In
contrast, 2 of 3 (67%) cell lines and 12 of 18 (67%) TCCs showed de novo
methylation of exon 1 of p!6/CDKN2, since PCR products were generated after
cutting with the methylation-sensitive enzymes FnuDU and SaclI (Fig. 5.6). Figure
5.6 also shows that while no methylation was observed in exon 2 of pl6/CDKN2 in
the normal bladder mucosa, extensive methylation of this CpG island occurred in the
bladder tumor. Exon 2 of pl6/CDKN2 was found to be de novo methylated in 15 of
18 (83%) uncultured TCCs and in 3 of 3 cell lines (Table 5.2 and Fig. 5.7).
132
I
n o r m a l ’“ -ffillo lllftu u
BLADDER n
MUCOSA
BLADDER
TUMOR
PCR-based Methylation Analysis Expression
Exonl Exon 2
U F s H C M US HC M
s
U N M
E
■ K "
□
Fig. 5.6. PCR methylation analysis and RT-PCR expression analysis of the pl6 and p i5 genes in normal and tumor
tissues. Results for normal urothelium and one of the bladder tumors examined are shown. DNA was digested with either
FnuDll (F), SacII (s), //pall (H), Qol (C), Smal (S), Nad (N), or AfspI (M). One hundred nanograms of digested or
undigested (U) DNA was amplified with the primer sets indicated by the arrows in the maps. Solid circles represent
methylated sites, partially solid circles indicate partial methylation, and empty circles represent unmethylated sites.
U >
H S
I
H H
1 1
i
C c c
I i
cc
Fig. 5.7. Methylation status of exon 2 of pl6/CDKN2 in uncultured bladder
TCCs. 66% of TCCs of all stages showed de novo methylation of exon 2 of
pl6/CDKN2 which, in contrast to methylation of exon 1, was not associated with
transcriptional silencing of the gene. A diagram of exon 2, the sites examined and
the primers used in the analysis are indicated at the top. Hypermethylation of exon
2 is shown in tumors of different stage.
134
Table 5.2. De novo methylation of pl6/CDKN2 in primary bladder TCCs and
bladder tumor-derived cell lines
Methylation of exon \a Methylation of exon 2
Bladder TCCs 12/18 (67%) 15/18 (83%)
Bladder cell lines 2/3 (67%) 3/3 (100%)
^Number of specimens showing de novo methylation of at least the FnuDU and
SacE sites/number of specimens analyzed.
135
RNA for expression analysis was available from normal urothelium, 3 of the
uncultured tumors, and 3 cell lines. A perfect correlation between methylation of exon
1 of pl6/CDKN2 and transcriptional inactivity of the gene was found, as 1 of the 3
tumors (Fig. 5.6) and 2 of 3 cell lines showed methylation of exon 1 with no
expression of p!6/CDKN2, whereas the normal urothelium (Fig. 5.6), two of three
tumors, and one of three cell lines showed no methylation of exon 1 and high levels of
p!6/CDKN2 mRNA. In contrast, no correlation was found between the methylation
status of exon 2 and expression levels of pl6/CDKN2 or between expression and
methylation of either exon of the pl5IN K 4 B gene in uncultured TCCs and cell lines.
Methylation and expression of the pl6/CDKN2 and p l5 IN K 4B genes in colon
carcinomas and adjacent normal mucosa
Figure 5.8 shows representative examples of the 10 colon cancer cases
examined, from which normal colonic epithelium and the corresponding colon tumor
were available. The FnuDll, Sacll and Hpa.IL sites in exon 1 of pl6/CDKN2 were
fully methylated in the normal colonic mucosa of patient A, and all five sites examined
were methylated in the normal mucosa of patient C. In contrast, no PCR product was
obtained from the corresponding colon tumor DNA of the same individuals following
enzyme digestion with any of the four methylation-sensitive enzymes used, or the
control endonuclease Mspl (Fig. 5.8). Thus, extensive methylation in exon 1 of
pl6/CDKN2 was present in these patients' normal colonic epithelium and was
associated with lack of expression of pl6/CDKN2, while the corresponding tumors
showed no methylation of this region and expressed the gene (Fig. 5.9). In contrast,
136
Fig. 5.8. Methylation status of exon 1 of the pl6/CDKN2 gene (or p l6 IN K 4) in
normal colon mucosa and colon carcinoma of patients A and C. PCR amplifications
were performed with primers flanking the FnuDU, SacU, HpaU, and Cfol sites after
DNA digestion with the corresponding methylation-sensitive endonuclease.
Presence of a PCR product indicates methylation at the site examined. Non-digested
DNA was used as control for 100% methylation. DNA digested with Ms pi
endonuclease, which is insensitive to methylation of the internal C in the sequence
CCGG, was used as control for absence of methylation.
137
I
NORMAL TUMOR
PATIENT A
TUMOR NORMAL
u >
o o
Fig. 5.9. RT-PCR analysis showed remarkable differences between pl6/CDKN2
(pfflN K 4 ) and pl5m K 4B expression in 4 of 10 colon cancer patients (patients A, C, D,
and E). Expression of the GAPDH gene was used as control. An amount of cDNA
equivalent to 100 ng of RNA was used in each reaction. 22 cycles were carried out
for pl6/CDKN2, pl5IN K 4B , and GAPDH amplifications. N, normal colon mucosa.
T, colon tumor.
139
p !6 'N K 4
p l 5 INK4B
GAPDH
A C D E
N T N T N T N T
m J g mL
Wl “ “
-
i
m - - 4• •
— —
—
----------
the CpG island in exon 2 of pl6/CDKN2 underwent extensive de novo methylation in
these tumors. While exon 1 of pl5IN K 4 B was methylated at both Nael sites in normal
and tumor tissue, exon 2 of pl5,N K 4 B also underwent extensive de novo methylation
in the tumor.
Methylation of exon 1 of pl6/CDKN2 was observed in the normal mucosa of
6 of the 10 colon cancer patients examined, as well as in 1 of the 10 colon tumors
(Fig. 5.10). Figure 5.10 shows that in 7 of 10 cases pl6/CDKN2 expression was
undetectable in the normal colonic epithelium of patients with colon cancer, and only 3
patients (F, I, J) showed substantial expression of the gene in their normal mucosa.
All samples were found to contain a wild-type pl6/CDKN2 gene by SSCP analysis
(Fig. 5.11), and the tumors showed no upregulation of CDK4 expression (Fig. 5.12).
Furthermore, the levels of pl6/CDKN2 expression were markedly increased in the
tumors of 50% (5 of 10) of the patients (A to E) when compared to the normal
mucosa (Figs. 5.9 and 5.10). Variable levels of pl6/CDKN2 expression were
observed in the remaining cases and, importantly, in no case was exon 1 methylated
and the pl6/CDKN2 gene expressed. De novo methylation of the CpG island in exon
2 of pl6/C D K N 2 was observed in 7 of 10 (70%) colon tumors (Table 5.3).
Interestingly, an inverse pattern of expression of p!6/CDKN2 and pl5IN K 4 B was also
observed in cases A, C, D, E, (Figs. 5.9 and 5.10) that is, pl6/CDKN2 expression was
undetectable in the normal mucosa but significantly upregulated in the corresponding
tumor, whereas pl5,N K 4 B expression was higher in the normal tissue when compared
to the tumor. In contrast to pl6/CDKN2, no clear correlation was observed between
methylation of exon 1 of pl5/IN K 4B and expression (Fig. 5.10). In addition,
pl5«NK4B mRNA was detected in most (9 of 10) of the normal colonic
141
Fig. 5.10. Summary of mRNA expression and methylation status of (a) the
pl6/CDKN2 and (b) pl5IN K 4B genes in colon carcinomas and adjacent normal
mucosa of patients A to J. mRNA levels were detected by RT-PCR; GAPDH
expression was used as control. Bars and numbers represent pl6/GAPDH or
pl5/GAPDH ratio as determined with an LKB UltraScan XL Laser densitometer.
Methylation status of exon 1 of both genes was determined by a PCR-based assay
with primers flanking the FnuDU (F), SacII (S), HpaU (H), and Cfol (C) sites in
p l6 1N K 4 exon 1, and primers flanking the NaeI (N) sites in pl51 N K 4 B exon 1 .
Diagrams of exon 1 and restriction sites are not drawn to scale. O , unmethylated site.
• , fully methylated site. Partially filled circles indicate partially methylated sites, as
determined by scanning with an LKB UltraScan XL Laser densitometer.
142
C O L O N C A N C E R C A S E S
PATIENT
■ Normal mucosa
□ Colon tumor
METHYLATION OF p!6 EXON 1
F S H cc
• • •
O
ooo O
• • • O
ooo o
• • • •
ooo o
•
ooo o
888
•
o
8 8 8
o
o
OOO o
ooo o
• • • o
• • • o
888
o
o
OOO 0
o oo o
A
B
C
D
E
F
G
H
I
J
p!6 EXPRESSION
10
_ l _
12
-I
C O L O N C A N C E R C A S E S
I Normal mucosa
PATIENT
*
METHYLATION OF pl5 EXON
N N
3
O
A
3
3
B
•
•
C
3
3
D
3
3
E
•
3
F
3
•
G
•
•
H
0 I
O
A
• J
•
□ Colon tumor
p!5 EXPRESSION
Fig. 5.11. pl6/CDKN2 mutational analysis in colon carcinomas was performed by
SSCP analysis. No mutations were present in exon 2 of pl6/CDKN2, as indicated
by the absence of band migration shifts in either the 5' (top) or the 3' (bottom)
segments of the exon. SSCP analysis was performed as described in Chapter 4.
145
Fig. 5.12. Cyclin-dependent kinase 4 (CDK4) expression was examined by RT-
PCR in colon carcinomas and adjacent normal mucosa of patients A through J.
Amplifications were performed for 23 cycles with primers complementary to exon
2 (forward primer) and exon 3 (reverse primer) of CDK4 as described by He et al.,
1994. The 356-bp CDK4-amplification products were separated by electrophoresis
on a 1% agarose gel. N, normal colon mucosa; T, colon tumor.
146
Table 5.3. De novo methylation of pl6/CDKN2 in colon cancer
Methylation of exonla Methylation of exon 2
Colon carcinomas
1/10 (10%)*
7/10 (70%)
dum ber of specimens showing de novo methylation of at least the FnuDU and 5acII
sites/number of specimens analyzed.
*The corresponding normal colonic mucosa of this tumor showed identical
methylation pattern.
147
mucosas and, in general, decreases in the level of expression were observed in the
tumors (Figs. 5.9 and 5.10).
Methylation and expression of the pl6/CDKN2 and p l5 IN K 4B genes in non
neoplastic tissue
The methylation status and expression levels of the pl6/CDKN2 and
pl5IN K 4B genes were also examined in several normal tissues including sperm,
kidney, peripheral blood lymphocytes, and colonic epithelium from individuals
without colorectal cancer (Fig. 5.13). Partial methylation of the FnuDU. and SacII
sites in pl6/CDKN2 exon 1, and complete methylation of the 2 NaeI sites within
p!5iNK4B exon i was seen jn sperm DNA. DNA from kidney was not methylated at
the sites examined in either exon of either gene. Lymphocyte DNA showed no
methylation in exon lof pl6/CDKN2 and partial methylation in exon 1 of pl5IN K 4 B .
The normal colonic epithelium from the individuals without colon cancer showed
methylation of exon 1 of pI6/CDKN2 which correlated with lack of pl6/CDKN2
expression in this tissue (Fig. 5.13). pl5,N K 4 B exon 1 was partially methylated in this
normal colon mucosa. The results showing methylation of exon 1 of pl6/CDKN2 in
the normal colon mucosa of individuals without colon cancer were similar to those
obtained in one case of ulcerative colitis (Fig. 5.13) as well as in the normal colonic
epithelium of 6 of the 10 patients with colon cancer examined where exon 1 of
pl6/CDKN2 was extensively methylated (example shown in Fig. 5.8, and summary of
results in Fig. 5.10).
148
Fig. 5.13. mRNA expression and methylation status of the pl6/CDKN2 and
p l5tN iC 4 B genes in non-neoplastic tissues. mRNA levels were detected by RT-PCR;
GAPDH expression was used as control. Bars and numbers represent pl6/GAPDH
or pl5/GAPDH ratio as determined with an LKB UltraScan XL Laser densitometer.
Methylation status of exon 1 (■ ■ ) of both genes was determined by a PCR-based
assay with primers flanking the FnuDU (F), SacII (S), HpaU (H), and Cfol (C) sites
in pl6IN K 4 exon 1, and primers flanking the NaeI (N) sites in pl5IN K 4B exon 1 .
Diagrams of exon 1 and restriction sites are not drawn to scale. O, unmethylated site.
# , fully methylated site. Partially filled circles indicate partially methylated sites, as
determined by scanning with an LKB UltraScan XL Laser densitometer, n.d., not
determined.
149
NON-NEOP1A STIC TISSUE
METHYLATION OF p!6 EXON 1
F S H
—
• • o
ooo
ooo
ooo
ooo
cc
o
o
o
o
o
Colon
Ulc. Colitis
Kidney
Lymphocyte
Sperm
p!6 EXPRESSION
10
_L_
12
_l
METHYLATION OF pl5 EXON 1
N N
o
o
o
o
•
Colon
Ulc. Colitis
Kidney -
Lymphocyte - n d
Sperm - n d
p!5 EXPRESSION
Methylation of exon 1, but not of exon 2, of pl6/CDKN2 is associated with
transcriptional silencing
Table 5.4 is a compilation of the analyses of mRNA expression and
methylation status of exons 1 and 2 of the pl6/CDKN2 gene conducted in a total of
30 specimens from which both RNA and DNA was available, including 4 normal
tissues, 10 colon cancer cases (normal and tumor from each case), 3 bladder tumors,
and 3 bladder cell lines. Overall, exon 1 of pl6/CDKN2 was methylated in 11 of 30
(37%) specimens, and remarkably, no case revealed expression of the gene with
methylation of this exon. On the other hand, exon 1 was not methylated in 17 of 30
(57%) specimens which expressed pl6/CDKN2, while the remaining 2 specimens did
not show methylation of exon 1 nor expression of the gene. The correlation between
methylation of both the FnuDU and 5acII sites in exon 1 and transcriptional silencing
of pi6/CDKN2 was statistically significant (P=0.00001) (Table 5.4). In contrast,
there was not a significant association between methylation of exon 2 and
transcriptional repression of pl6/CDKN2 (P=0.340), since 12 of 30 (40%) specimens
showed extensive methylation of exon 2 with pl6/CDKN2 expression (Table 5.4).
Concordant methylation in both exon 2 CpG islands of pl6/CDKN2 and
p 15lN K 4B
Upon further analysis of our panel of normal tissues, tumors and cell lines we
observed similar patterns of methylation of the CpG islands in exon 2 of these related
genes, i.e., when exon 2 of pl6/CDKN2 was methylated, exon 2 of pl5IN K 4 B was also
methylated. Table 5.5 shows that in 32 of 43 (74%) specimens analyzed
151
Table 5.4. Methylation of exon 1 and exon 2 and expression of the pl6/CDKN2 gene
EXON1 EXON 2
UNMETHYLATED METHYLATED UNMETHYLATED METHYLATED
pl6 EXPRESSED 17 0
•
7 12
p!6 NOT EXPRESSED 2 11B 3 86
aP= 0.00001, two-sided
*P=0.340, two-sided
Table 5.5. Comparison of methylation status0 of exon 2 of pl6/CDKN2 and p l5 lN K 4B
pl6/CDKN2
METHYLATION
p l 5INK4B
METHYLATION
FULL PARTIAL NONE*
FULL 32/43* (74%) 0/43 0/43
PARTIAL 0/43 0/43 3/43 (7%)
NONE 0/43 3/43 (7%) 5/43(12%)
°The methylation status of exon 2 of the two genes was compared based on the results for the unique Sma I
site within the region examined. Results for this site were used in the comparison because there is only one
Sntal site in exon 2 of both genes and, therefore, the methylation status of a specific CpG site can be clearly
determined by the PCR-based method used.
*The methylation status of exon 2 of both genes was examined in a total of 43 samples, including 14
normal tissues, 26 tumors and 3 cell lines.
(including normal and tumor tissues, and cell lines), both islands were methylated in a
similar fashion, whereas only 6 of 43 (14%) cases showed discordant methylation
status of these islands in the two genes. The remaining 5 of 43 (12%) cases showed
unmethylated sites in both islands. The parallel patterns of methylation of these two
homologous sequences in two genes which show some differential regulation of
expression, suggest that the DNA sequence may be important in directing the
methylation process.
Induction of pl6/CDKN2 expression by the demethylating agent 5-Aza-CdR
In order to test whether demethylation of pl6/CDKN2 would result in its
upregulation and to further investigate the hypothesis that methylation of the S' CpG
island of p!6/CDKN2 represses transcription of the gene, two bladder carcinoma-
derived cell lines (EJ and HT1376) that did not express pl6/CDKN2 and showed
extensive methylation of exon 1, and one colon carcinoma-derived cell line (SW-480)
which did not express p!6/CDKN2, were treated with the demethylating agent 5-Aza-
CdR (Jones and Taylor, 1980). These cell lines were shown to contain a wild-type
pl6/CDKN2 sequence (Chapter 4; Spruck et al., 1994b). pl6/CDKN2 expression
was induced in all three cell lines after treatment with lp.M of 5-Aza-CdR for 24 h
(Fig. 3.14). cDNA from the J82 cell line was analyzed simultaneously as a positive
control for gene expression because this line was determined to express high levels of
p!6/CDKN2, with an unmethylated exon 1. As a control to exclude the possibility
that pl6/CDKN2 induction in the 5-Aza-CdR-treated cells was due to cell damage, the
EJ cell line was also treated with the DNA synthesis inhibitor 1-p-D-
arabinofuranosylcytosine (AraC), which does not inhibit DNA methylation (Jones and
154
cc
•a
o
C O
N
<
in
+
UJ
2
9
o
T”
C C
■ o
o
(0
N
<
m
+
to to
2
©
cc
■o
o
C O
N
<
in
C M _ r - co in
« j u o © © [2
. C O C O
T ^ r* <
I- H 5 ^
CO CO
p16
* »
p15
GAPDH
Fig. 5.14. Effect of 5-aza-CdR treatment on cell lines. Two bladder
carcinoma-derived cell lines (EJ and HT1376), and one colon carcinoma-
derived cell line (SW-480) were treated for 24 h with the indicated doses of
the DNA methylation inhibitor 5-aza-CdR. RT-PCR analysis of pl6/CDKN2
and pl51 N K 4 B expression was performed in the indicated cell lines before and
after 5-aza-CdR treatment. The J82 line was included as a positive control for
pl6/CDKN2 and pl5IN K 4 B expression. The EJ line was also treated with
AraC, which does not inhibit DNA methylation. pl6/CDKN2 expression was
induced in all three cell lines by 5-aza-CdR. GAPDH was used as a control
for cDNA integrity and input.
1 5 5
Taylor, 1980). Treatment with this drug did not induce pl6/CDKN2 expression,
suggesting that pl6/CDKN2 induction after 5-Aza-CdR treatment was associated with
demethylation of the pl6/CDKN2 gene. In contrast to the results obtained for
pl6/CDKN2, pis™*™ expression remained at similar levels in untreated and 5-Aza-
CdR-treated cells (Fig. 5.14).
Demethylation of CpG sites in exon 1 of pl6/CDKN2 occurred in EJ and
HT1376 cells treated with 5-Aza-CdR, including the SczcII and the H paVL site in both
lines, and also at least one of the two Cfol sites in the EJ line (Fig. 5.15). However,
the FnuDU site remained methylated in both treated cell lines. Partial demethylation
of exon 2 was also observed in the 5-Aza-CdR-treated EJ line, and more extensive
demethylation of this exon was observed in the 5-Aza-CdR-treated HT1376 line (Fig.
5.15).
Activation of pl6/C D K N 2 expression by 5-Aza-CdR is associated with
inhibition of tumor cell growth in vitro
To assess the effect of pl6/CDKN2 induction by 5-Aza-CdR on cell growth,
the population doubling time of the untreated and 5-Aza-CdR-treated EJ cells was
determined. The J82 cell line was included as control because it expresses
p!6/CDKN2. The results of the cell counts are presented in Fig. 5.16, and the
doubling times are indicated. EJ cells in which pl6/CDKN2 expression was induced
after 5-Aza-CdR treatment showed a marked increase in their population doubling
time (from 20 h to 43 h). However, the J82 cell line, that expressed a wild-type
p!6/CDKN2 gene before and after 5-Aza-CdR treatment, did not show any difference
156
Fig. 5.15. Effect of 5-Aza-CdR treatment on cell lines. PCR-based methylation
analysis of the pl6/CDKN2 (or p l6) and pl51 N K 4 B (or pl5) genes in the EJ and
HT1376 cell lines before and after 5-Aza-CdR treatment. Schematic maps of the
pl6/CDKN2 and pl5,N K 4 B genes with the sites analyzed are presented. DNA was
digested with either FnuDU (F), 5acII (s), HpaU (H), Cfol (C), Smal (S), Nael (N),
or MspI (M). One hundred nanograms of digested or undigested (U) DNA was
amplified with the primer sets indicated by the arrows in the maps. Solid circles
represent methylated sites, partially solid circles indicate partial methylation, and
empty circles represent unmethylated sites. The absence of a circle on a site indicates
that, although it was examined, its precise methylation status could not be determined
due to the presence of more than one restriction site for the same enzyme in the
region flanked by the primer sets used in the assay.
157
PI
El CELL LINE
ni6 ir tif i m m n
nis ,, imwtn
SAY.ACdR
+ SAYACdR
p !6
Pis
SAYACdR
HT1376 CELL LINE
nit 1ft.. f f t f f i f f f
PIS , , H ftlftff?
pis n tinlftiii
+ SAYACdR
PCR-based Methylation Analysis
E xon I E xon 2
U K s H C M It S H C M
H N M
□
EJ growth before and after 5-Aza-CdR treatment
• Aza: 20 h
+ Aza: 43 h
too-
X
£
E
3
e
10-
s
2 4 0 I 3 5 6
Tima (days)
J82 growth before and after 5-Aza-CdR treatment
100
- Aza: 26 h
+ Aza: 26 h
O
X
10-
£
E
3
e
0 2 1 3 4 5 6
Tima (days)
Fig. 5.16. Cell growth and doubling times of EJ (top) and
J82 (bottom) cell lines before and after treatment with 10-6 M
of 5-Aza-CdR.
159
in growth rate after the treatment. These results suggested that the increase in the
doubling time of treated EJ cells was not a consequence of 5-Aza-CdR toxic effect,
but rather was associated with the activation of the pl6/CDKN2 gene in these cells.
The duration of pl6/CDKN2 expression in EJ cells after induction with 5-
Aza-CdR was examined. Treated cells were maintained in culture and passaged
before reaching confluency. pl6/CDKN2 expression was tested and cell population
doubling time estimated in each consecutive passage. The results obtained for eight
different passages are presented in Fig. 5.17. The data showed that p!6/CDKN2
mRNA levels in EJ cells remained high for 3 passages after induction by 5-Aza-CdR,
decreased in the fourth passage, and were undetectable after six passages. The
doubling time of EJ cells increased two-fold after 5-Aza-CdR treatment, and correlated
with pl6/CDKN2 induction since we observed a progressive increase in the cell
growth rate as p!6/CDKN2 expression levels decreased (Fig. 5.17). These results
show that there is an association between 16/CDKN2 induction by 5-Aza-CdR
treatment and inhibition of cell growth proliferation.
Induction of pl6/CDKN2 expression by 5-Aza-CdR in vivo
To determine if 5-Aza-CdR could induce pl6/CDKN2 expression in vivo, EJ
cells (that do not express pI6/CDKN2) were implanted in the bladder of 4 week-old
nu/nu rats through laparostomy and direct injection into the bladder wall with a 27-
gauge needle. Eight weeks after implantation of the cells, 5-Aza-CdR was
administered daily for five days via intraperitoneal injections. 24 h after the last
injection rats were sacrificed and their bladders obtained. Human p!6/CDKN2
160
Fig. 5.17. Induction of pl6/CDKN2 expression in EJ cells by 5-Aza-CdR treatment
correlates with a decreased cell growth rate. EJ cells were maintained in culture after
treatment with 10' 6 M 5-Aza-CdR. pl6/CDKN2 mRNA levels and cell doubling
times were examined in consecutive passages (p). Reactivation of pl6/CDKN2
expression correlated with increased doubling times. After 6 passages pl6/CDKN2
expression was not detectable, and cell growth rate increased.
161
p16 expression and doubling times
in non-treated and 5-aza-CdR-treated EJ cells
p16 mRNA -►
GAPDH mRNA -►
Doubling tim e ------- ► 20 43 43.5 40 37 22.5 22
(hours)
pi 6 exp.
0.75
c
o
(0
(0
s 0.5
a
x
0)
< 0
a
0.25 -
a. a
C M
Q.
cn
a a
ce
a
c o
a
in
EJ + 5-aza-CdR
162
expression was examined by RT-PCR analysis with primers that flank intron 1 of the
human p l 6 gene (the 5-primer is complementary to exon 1 and the 3'-primer is
complementary to exon 2 sequences) and do not amplify rat sequences. The results
obtained are shown in Table 5.6 and Figure 5.18. Seven of 9 bladders presented
tumors, and no significant difference in the tumor sizes was detected in the animals
that received 5-aza-CdR treatment when compared with the non-treated ones. 5-aza-
CdR treatment resulted in induction of pl6/CDKN2 expression in four of the six
treated rats, including 3 of 3 that received daily doses of 600 |xg of 5-Aza-CdR and 1
of 3 that received daily doses of 900 |ig of 5-Aza-CdR. These results indicated that
pl6/CDKN2 expression can be induced by in vivo treatment with 5-Aza-CdR.
DISCUSSION
Our results show a highly significant correlation (P=0.00001, two sided)
between methylation of exon 1 of pl6/CDKN2 and transcriptional silencing of the
gene, since none of the normal and tumor specimens that showed methylation of this
exon (11/30) expressed p!6/CDKN2, and 17 of 30 specimens expressed pl6/CDKN2
without evident methylation of exon 1 (Table 5.4). These data suggest that
methylation of the 5' CpG island of p!6/CDKN2 might be a mechanism to control
expression of this tumor suppressor gene. It remains to be seen whether the
methylation is causally related to the silencing of expression. However, additional
evidence that methylation of the 5' CpG island is involved in pl6/CDKN2 silencing
comes from the results obtained from the treatment of three cell lines that did not
express p!6/CDKN2 with the inhibitor of DNA methylation 5-Aza-CdR (Fig. 5.14).
163
Table 5.6. pl6/CDKN2 expression in EJ tumors of nude rats treated
with 5-aza-CdR
Sample 5-aza-CdR
m
Tumor size
pl6/CDKN2
mRNA
A1 no 3 mm tumor no
B3 no no evident
tumor
no
Cl no haemorragic
7mm tumor
no
A2 600 pg i.p./day, 5
days
haemorragic
6 mm tumor
yes
A3 600 pg i.p./day, 5
days
haemorragic
5mm tumor
yes
A4 600 pg i.p./day, 5
days
5 mm tumor yes
B1 900 jig i.pVday, 5
days
4 mm tumor no
B2 900 pg i.p./day, 5
days
haemorragic
7mm tumor
yes
B4 900 pg i.p./day, 5
days
no evident
tumor
no
164
Induction of p 1 6 exp ression in vivo
by 5-aza-CdR treatment
I 1
« o
O o
z +
EJ
a )
■ D
£ I
0 0
O C C E
o c
■ o
o
0
N
0
i n
o
600 pg 900 |ig
A1 B3 C1 A2 A3 A4 B1 B2 B4
p 1 6 mRNA
hGAPDH mRNA
«#
Fig. 5.18. Induction of pl6/CDKN2 expression in vivo by 5-Aza-CdR
treatment. pl6/CDKN2 mRNA levels were examined in EJ cells-derived tumors
in nude rats after 5-Aza-CdR treatment. Treatment consisted of daily
intraperitoneal injections of 600 pg or 900 pg for 5 days. The human GAPDH
gene expression was used as control for RNA loading and integrity.
1 6 5
The failure of AraC, a DNA synthesis inhibitor, to activate pl6/CDKN2 expression
and the observation that the 5-Aza-CdR induced pl6/CDKN2 expression was
accompanied by a decrease of exon 1 methylation is consistent with S' CpG island
methylation playing a role in transcriptional silencing of pl6/CDKN2.
Our analysis of uncultured bladder tumors and bladder cell lines indicates that
the p!6/CDKN2 tumor suppressor gene is frequently subject to de novo methylation
of the S' CpG island in transitional cell carcinomas of the bladder. The frequency
(67%) of bladder TCCs showing de novo methylation of the pl6/CDKN2 5’ CpG
island could explain the low rate at which pl6/CDKN2 homozygous deletions and
intragenic mutations are found in this tumor type (Chapter 4; Spruck et al., 1994b).
Thus, our data indicates that, in contrast to squamous cell carcinomas of the bladder in
which pl6/CDKN2 is frequently deleted (Chapter 6 ; Gonzalez-Zulueta et al., 1995), de
novo methylation of the 5' CpG island of p!6/CDKN2 may be the most common
mechanism of inactivation of this tumor suppressor gene in transitional cell
carcinomas of the bladder.
Our study is the first to show that methylation of the 5' CpG island of
pl6/CDKN2 occurred not only in bladder tumors and bladder cell lines, but also in
normal colon mucosa. Methylation of this region was observed in the normal colonic
epithelium from five patients who had segments of the colon removed for reasons
other than cancer development, and also in the normal colon mucosa of several (6 of
10) patients with colon cancer (Figs. 5.9 and 5.10). This is striking, as methylation of
CpG islands in autosomal genes is generally restricted to genes located on the inactive
X chromosome (Singer-Sam and Riggs, 1993), imprinted genes (Ferguson-Smith et
166
al., 1993), or genes that undergo de novo methylation during cell line inmortalization
and tumorigenesis (Jones and Buckley, 1990; Rideout et al., 1994; Laird and Jaenisch,
1994). To date, methylation of CpG islands of autosomal genes in normal tissues has
only been reported for the estrogen receptor gene in normal colonic mucosa, where
methylation of this gene was observed to arise as a direct function of age (Issa et al.,
1994). It will be important to determine if the methylation and transcriptional
silencing of pJ6/CDKN2 observed in some colon specimens are involved in normal
cell function or are precursors to tumorigenesis.
It was surprising to find that, in contrast to bladder TCCs, de novo methylation
of the 5' CpG island of p!6/CDKN2 was not observed in the colon carcinomas
examined. Furthermore, in 6 of the 10 colon carcinomas p!6/CDKN2 expression was
strongly upregulated with no methylation of exon 1, and in 3 of these 6 cases we
observed a clear inverse expression pattern of the pl6/CDKN2 and pl5IN K 4 B genes,
with pl6/CDKN2 upregulation and pl5,N K 4B downregulation in the tumors when
compared to the corresponding normal mucosa (Fig. 5.10). pl6/CDKN2 appears to
accumulate in cells in which the function of pRb has been compromised (Serrano et
al., 1993; Yeager et al., 1993). It would be of interest to examine the status of pRb in
our panel of colon carcinomas, although pRb alterations have not been reported in
colon cancer. Hannon et al. have shown a strong upregulation of pl5,NK4B
expression by TGF-P in human keratinocytes (Hannon et al., 1994), and have
proposed p l5IN K 4B role as an effector of TGF-P-mediated cell cycle arrest. In
normal colonic epithelium the upper 2/3 of the crypts are composed of quiescent cells
that are rapidly being replaced by new cells migrating from the lower 1/3 of the crypt
(Crissman and Barwick, 1993). The appreciable levels of pl5IN K 4B expression
167
detected in the normal colon mucosas in our study may indicate that this cell cycle
regulator has a role in inducing quiescence in the majority of cells of the normal
colonic epithelium. Markowitz et al. have recently reported a high frequency of
mutations in the TGF-p type II receptor in colon carcinomas (Markowitz et al., 1995),
therefore, it would be important to examine the status of this receptor in our panel of
colon carcinomas to determine if the observed plS84* 48 downregulation in a subset
of these tumors correlates with TGF-P type II receptor mutations.
Our results also show that extensive de novo methylation of the CpG islands
in exon 2 of both pl6/CDKN2 and plS^ 48 occurred in tumors and cell lines which,
in contrast to methylation of pl6/C D K N 2 exon 1, was independent of the
transcriptional activities of the genes. It has been shown (Magewu, 1994) that highly
CpG rich sequences, such as Alu and LI elements, in non-promoter regions of many
genes are extensively methylated and do not block transcription. A similar situation
appears to occur in the CpG islands in exon 2 of p!6/CDKN2 and plS84*48.
It has been difficult to determine the factors that govern the establishment of
de novo methylation patterns. The expression-independent methylation of the two
highly homologous CpG islands in the exon 2 regions of the p l 6 and pl5IN K 4B
genes provides a good opportunity to study the mechanisms responsible for the
establishment of de novo methylation. Our observation that these two homologous
regions follow nearly identical patterns of methylation at the sites examined suggests
the importance of the specific DNA sequence in driving the de novo methylation
process. The existence of cis-acting signals that direct de novo methylation has been
previously described in other situations (Selker et al., 1987; Szyf et al., 1986;
168
Mummanemi et al., 1993; Mummanemi et al., 1993). A portable signal causing DNA
methylation was described in Neurospora crassa (Selker et al., 1987), and a
“methylation center” has been shown to direct the methylation of surrounding sites
in the promoter region of the mouse adenine phosphoribosyltransferase gene
(Mummanemi et al., 1993). Similarly, integrated virus-induced de novo methylation
of flanking host sequences has been described (Jahner and Jaenisch, 1985; Toth et al.,
1990), and it has been shown that spreading of methylation into flanking sequences
begins from specific nonrandom regions of the viral genome (Kuhlmann and
Doerfler, 1983). Specific sites within exon 2 of p!6/CDKN2 and pl5IN K 4B may be
recognized by the de novo methylation activity, leading to subsequent spreading of
methylation in this region, and ultimately resulting in similar methylation patterns in
both genes.
Evidence has accumulated that dysregulation of genes controlling the cell cycle
may contribute to the malignant progression of cells. Premature entry of a cell into
the next phase of the cell cycle may result in incomplete repair of DNA damage and
subsequent genomic instability. The p i6 protein can complex with cyclin D1-CDK4
and inhibit its interaction with pRb, thus retarding passage through the cell cycle.
Inactivation of pl6/CDKN2 may increase the activity of CDK4, resulting in
hyperphosphorylation of the Rb protein. The cell cycle will be driven from G1 to S
and cell proliferation will be enhanced. Our results and those of others (Merlo et al.,
1995; Hermann et al., 1995) suggest that hypermethylation associated with
transcriptional silencing may be the primary mode of p!6/CDKN2 inactivation in
cancers such as bladder, lung, head and neck, and breast cancer. Our data also
indicate that p!6/CDKN2 expression can be restored in vitro and in vivo in cells that
169
contain a methylated p!6/CDKN2 gene by treatment with 5-Aza-CdR, and that
reactivation of pl6/CDKN2 is associated with cell growth inhibition. These
observations, which are being confirmed with experiments to directly link cell growth
inhibition to pl6/CDKN2 reactivation, have clear clinical implications, and DNA
methylation inhibitors, such as 5-Aza-CdR, may have application in novel cancer
therapy strategies.
170
CHAPTER 6
GENETIC ANALYSIS OF SQUAMOUS CELL CARCINOMA OF THE
BLADDER: HIGH FREQUENCY OF CHROMOSOME 9p ALLELIC LOSS
AND CDKN2 ALTERATIONS DISTINGUISH IT FROM TRANSITIONAL
CELL CARCINOMA
INTRODUCTION
Approximately 90% of malignant tumors of the urinary bladder in the
Western Hemisphere are transitional cell carcinomas (TCCs), while only about 7% are
classified as squamous cell carcinomas (SCCs), which are often associated with long
standing chronic irritation of the urothelium and have poor prognosis (Mostofi et al.,
1988). In contrast, in Egypt and some regions of the Middle East and Africa, SCC is
not only the most common bladder cancer but also the most common cancer in men,
accounting for more than one-fourth of all cancer cases and 80% of bladder cancers
(Eagan, 1989). This frequent occurrence of SCC has been postulated to be caused, in
part, by endemic infection by the trematode Schistosoma haematobium. Some
investigators have suggested that this organism may potentiate the mutagenic action of
some environmental carcinogens through its chronic irritation of the urothelial lining
(Badawi et al., 1992).
SCC differs from TCC not only in its histology, geographic distribution, and
risk factors but also in its clinical behavior and prognosis. At the time of diagnosis
SCC is more advanced than TCC, with deep muscle invasion almost always present
171
and common extravesical extension. SCC is typically more aggressive than TCC, with
a lower overall 5-year survival rate (Eagan et al., 1989). Because they differ at the
histopathological and clinical levels, their underlying genetic defects may also be
different. To date, most genetic studies on bladder cancer have focused on TCC, and
little is known about the genetic alterations that occur in SCC tumorigenesis.
A new putative tumor suppressor gene, p!6/CDKN2 (also MTS1, or pl61 N K 4 ),
has been mapped to band 21 of the short arm of chromosome 9 (9p21) (Kamb et al.,
1994; Nobori et al., 1994). Molecular and cytogenetic abnormalities of 9p21 have
been reported in several malignancies and tumor-derived cell lines, including
melanoma (Fountain et al., 1992; Petty et al., 1993), glioma (Olopade et al., 1992a),
leukemia (Olopade et al., 1992b), lung cancer (Olopade et al., 1993), head and neck
squamous cell carcinoma (van der Riet et al., 1994) and TCC of the bladder (Cairns et
al., 1994a; Stadler et al., 1994). The p!6/CDKN2 gene encodes the previously
identified p l6-kilodalton protein, a cell cycle regulatory protein that binds to CDK4
and inhibits the catalytic activity of the CDK4-cyclin D enzymes (Serrano et al.,
1993). The CDK4-cyclin D complexes control passage through the Gi phase into
the S phase of the cell cycle by phosphorylation of cellular factors, one of which is the
retinoblastoma protein (Matsushime et al., 1992). Initial studies demonstrating
frequent homozygous deletions and point mutations of pl6/CDKN2 in a variety of
tumor-derived cell lines raised the possibility that p!6/CDKN2 might be an important
tumor suppressor gene present on 9p21 (Kamb et al., 1994; Nobori et al., 1994).
Additionally, somatic mutations of pl6/CDKN2 were found in 14 of 27 squamous cell
carcinomas of the esophagus (Mori et al., 1994), and Caldas et al. (1994) reported
homozygous deletions and point mutations of p!6/CDKN2 in 10 and 11, respectively,
172
of 27 pancreatic adenocarcinomas. In contrast, more recent studies performed on
primary uncultured tumors such as TCC of the bladder (Spruck et al., 1994b; Cairns
et al., 1994b), lung, brain, kidney (Caims et al., 1994b), head and neck (Zhang et al.,
1994), breast (Xu et al., 1994), sporadic melanoma (Ohta et al., 1994), astrocytoma
(Ueki et al., 1994), and malignant mesothelioma (Cheng et al., 1994) have reported
pl6/CDKN2 alterations in a significantly smaller percentage of uncultured tumors.
These data suggest that p!6/CDKN2 alterations may be important in only certain
tumor types, and that loss of pl6/CDKN2 function may give a selective advantage
during in vitro establishment of cell lines from tumor tissue (Spruck et al., 1994),
leading to an overrepresentation of cultured tumor cells carrying these defects. Recent
results indicate that loss of CDNK2 may be found during tumor progression
(Okamoto et al., 1995; Reed et al., 1995).
Molecular genetic and cytogenetic analyses have shown that loss of
chromosome 9 is the most frequent alteration in urothelial TCC (Tsai et al., 1990;
Knowles et al., 1994), and there is evidence that two putative tumor suppressor genes
involved in TCC development may reside on this chromosome (Miyao et al., 1993;
Keen and Knowles, 1994). Mutations in the p53 tumor suppressor gene are also
commonly found in TCC, and have been associated with an aggressive clinical course
(Sidransky et al., 1991; Fujimoto et al., 1992; Sarkis et al., 1993; Spruck et al., 1994a;
Esrig et al., 1994). The p53 tumor suppressor gene maps to the short arm of
chromosome 17 and encodes a nuclear phosphoprotein of 53 kD which is also
involved in cell cycle control (Levine et al., 1991).
173
The molecular genetic alterations involved in the tumorigenesis of SCC of the
bladder are not currently well defined. We obtained SCC specimens from Egypt, a
high-risk area where this tumor type is associated with schistosomiasis affected
bladders, and from Sweden,* a low-risk area where SCC is often found in paraplegics,
who are frequently victims of chronic urinary tract infections and lithiasis. The aims
of this study were to understand the genetic defects underlying SCC development, and
to determine if SCC and TCC, which are clinically and histopathologically distinct, are
also genetically different. Alterations examined included allelic loss in chromosome
arms 9p, 9q and 17p, and point mutations in the pl6/CD KN2 and p53 tumor
suppressor genes on these chromosomes.
MATERIALS AND METHODS
Tumor samples and DNA isolation
Thirty-one SCCs of the bladder were obtained for the present study. Nineteen
tumors were obtained from patients treated at the Urology and Nephrology Center in
Mansoura, Egypt, and 12 tumors were obtained from patients treated at the University
Hospital in Uppsala, Sweden. Specimens were high-stage, poorly-differentiated
tumors. SCC specimens were analyzed histopathologically for evidence of
schistosomal infection. Schistosoma eggs were found embedded in the bladder
musculature of 7 specimens from Egypt (Fig. 6.1), while not detected in any of the
specimens from Sweden.
174
Fig. 6.1. Schistosoma haematobium eggs (arrows) embedded in the urothelium of a patient with invasive
squamous cell carcinoma (arrowhead). SCC specimens were analyzed histopathologically for evidence of
schistosomal infection. Schistosome eggs were found embedded in the bladder musculature of 7 specimens
from Egypt, while not detected in any of the specimens from Sweden.
DNA was isolated from formalin-fixed, paraffin-embedded tissue specimens
as previously described (Chapter 2; Spruck et al., 1994a). Briefly, tumor and normal
tissues were microdissected from 8 -1 0 pm thick hematoxylin and eosin stained
sections using sterile scalpels. Tissue was microdissected from three sections of each
specimen. Approximately 1000 cells from normal and tumor tissue were taken for
analysis. The tissue was resuspended in lOmM Tris-HCl and SmM EDTA, and
treated at 56°C, overnight, with proteinase K (Boehringer Mannheim, GmbH, W.
Germany) at a final concentration of 1 mg/ml.
Sufficient amounts of DNA were obtained from 12 SCCs in order to perform
a complete genetic study, i.e., analysis of alterations in pl6/CDKN2, p53, chromosome
9p, chromosome 9q, and chromosome 17p. However, due to extremely limited tumor
material in the remaining 19 SCCs DNA was only analyzed for alterations in p53,
chromosome 9q, and chromosome 17p.
PCR amplification and deletion analysis of p i 6/CDKN2
Nested PCR was used to amplify exon 2 of the pl6/CDKN2 gene in DNA
isolated from archival tissue. For the first round of PCR the primers used were: 5'-
GAC CCG TGC ACG ACG CT- 3' (within exon 2), and 5 -TGA GCT TTG GAA
GCT CTC AG -3' (in intron 2). For the second round of PCR the primers used were:
5 -GAC CCG TGC ACG ACG CT-3' and 5 -GGT ACC GTG CGA CAT CGC-3'
(both within exon 2). A microsatellite sequence within the androgen receptor gene
located on the X chromosome was selected as control template (Cutler et al., 1992).
The primer sets used for amplifying the microsatallite DNA were: S'-GTG CGC GAA
176
GTG ATC CAG AA-3' and 5 -TCT GGG ACG CAA CCT CTC TC-3' for the first
round of PCR, and 5’ -AGA GGC CGC GAG CGC AGC ACC TC-3’ and 5-GCT
GTG AAG G'lT GCT GTT CCT CAT-3' for the second round of PCR. Reactions
were performed in a 50 pi volume and included primer sets for both the control and
the pl6/CDKN2 templates, at a final concentration of lpM. Reaction mixtures also
contained 5% DMSO, 10 mM Tris-HCl (pH 8.3), 50 mM KC1, 1.5 mM MgCl2,400
pM of each deoxynucleoside triphosphate, and 5 units (u) of Taq polymerase
(Boehringer Mannheim, GmbH, W. Germany). Conditions for the first PCR were:
94°C for 2 min, 17 cycles of 94°C for 1 min, 55°C for 40 sec, and 72*C for 1 min,
followed by incubation at 72°C for 2 min. One microliter of each reaction product was
subjected to a second round of PCR using the same conditions except that 28 cycles
were performed, 60°C was used as the annealing temperature, and 2 pCi [a-32P]dCTP
(3000 Ci/mmol, ICN Biochemicals, Costa Mesa, California) were added to each
reaction mixture. Under these conditions, reactions were determined to be in the
exponential phase of DNA amplification (M. Gonzalez-Zulueta, unpublished data).
PCR products were separated by electrophoresis in nondenaturing 8 % polyacrylamide
gels. Gels were dried in a gel dryer (Model 583, Bio-Rad, Richmond, California) and
exposed to autoradiogram film (Hyperfilm-MP, Amersham, Arlington Heights, IL),
using one intensifying screen, at -70°C, overnight. Homozygous deletion was scored
as present if the p!6/CDKN2 signal was less than 5% of the signal from the control
locus, as determined by densitometry analysis. Each reaction was performed at least
twice to minimize the possibility of artifacts.
177
SSCP analysis and sequencing of p!6/CDKN2
All specimens were screened for mutations in exons 1 and 2 of pl6/CDKN2.
These two exons constitute 99% of the pl6/CDKN2 coding sequence, and the
majority of mutations described to date have been detected in these exons. Exon 1
was sequenced with primers that flank the coding sequence and the acceptor-splice
sites. Each specimen was first amplified with the following primers: S'-GAA GAA
AGA GGA GGG GCT GG-3' (sense), and 5’ -GCG CTA CCT GAT TCC AAT TC-3'
(antisense). One microliter from this primary PCR was then amplified with the
internal primers: 5-CTC GGC GGC TGC GGA GA-3’ (sense), and 5'-TCC CCT
GCT CCC GCT GC-3' (antisense). The last two oligonucleotides were used for
direct sequencing of the PCR products in both directions.
Exon 2 was first analyzed by single-strand conformation polymorphism
(SSCP). Mutant controls were included in each SSCP gel to verify the sensitivity of
the method. These controls consisted of three bladder tumors with a G to A base
change at nt 214 of the p!6/CDKN2 gene, a G to T base change at nt 373, and a G to
A base change at nt 436, respectively. Each sample was subjected to two different
primary rounds of PCR in order to examine the S'-, middle, or 3-segments of exon 2.
For analysis of the S'-segment, the primers used in the first round of PCR were: 5'-
GTG GGG TGC TTG GCG GTG AG-3' (in intron 1) and 5'-GGT ACC GTG CGA
CAT CGC-3' (in exon 2); the primers used in the second round of PCR were: 5'-CAT
TCT GTT CTC TCT GGC AGG-3' and 5 -CAC CAG CGT GTC CAG GAA-3'. For
the middle and 3-segments, the first round of PCR was common and used the
following primers: 5-GAC CCG TGC ACG ACG CT-3' within exon 2 and 5-TGA
178
GCT TTG GAA GCT CTC AG-3' in intron 2. The second PCR round for the middle
fragment used the primers 5'-GAC CCG TGC ACG ACG CT-3' and 5-GGT ACC
GTG CGA CAT CGC-3’. For the 3'-segment, the primers used in the second PCR
round were 5-TGG ACC TGG CTG AGG AG-3‘ and 5-CAA ATT CTC AGA TCA
TCA GTC CTC-3'. Conditions for the first amplifications were 94°C for 2 min, 17
cycles of 94°C for lmin, 55°C for 45 sec, and 72*C for 1 min, followed by incubation
at 72°C for 2 min. Conditions for the second amplifications were similar to those used
for the first amplifications except that the annealing temperature was 58°C, and 28
amplification cycles were performed. 2 pCi [a-32P]dCTP (3000 Ci/mmol, ICN
Biochemicals, Costa Mesa, California) were included in the second amplifications.
Radiolabeled secondary PCR products were diluted 1:5 with Sequenase stop solution
(US Biochemicals, Cleveland, OH), boiled for 5 min, and immediately loaded onto
nondenaturing gels (6 % polyacrylamide, 10% glycerol, and lx Tris-Borate-EDTA
buffer) (39). Electrophoresis was performed at 30W constant power at room
temperature with a cooling fan for approximately 7 h. Gels were dried for 15 min, and
exposed to X-ray film (Amersham Hyperfilm MP, Arlington Heights, IL) with an
intensifying screen at -70°C overnight. Each reaction was performed at least twice.
DNA from samples that showed migration shifts by SSCP was reamplified
without labeled deoxynucleoside triphosphate, and the amplification products were
subjected to electrophoresis in 2% agarose gels and further purified using Mermaid
Kit (Bio 101, La Jolla, CA). Sequencing of the purified DNA fragments was
performed bidirectionally with the same primers used in the second PCR
amplifications and the Sequenase version 2.0 kit (US Biochemicals, Cleveland, OH).
Sequencing reaction products were resolved in 8% polyacrylamide gels containing
179
7M urea, and were visualized by autoradiography. SSCP analysis and sequencing
reactions were performed by Dr. Atsuko Shibata.
Determination of allelic loss
Loss of heterozygosity in chromosomes 9 and 17p was determined by PCR
amplification of polymorphic dinucleotide repeat sequences. Microsatellite markers
examined were D9S156, D9S162, IFNA, D9S171, D9S169, and D9S161 on
chromosome 9p; D9S153, D9S59 and D9S63 on chromosome 9q, and D17S513 on
chromosome 17p. Primer sequences for the analysis of microsatellite markers were
obtained from The Genome Data Base (Welch Medical Library, The Johns Hopkins
University). PCR amplifications were performed in 25 pi volumes, containing 10 mM
Tris-HCl (pH 8.3), 50 mM KC1, 1.5 mM MgCl2 , 1 pM primers (each), 200 pM of
each deoxynucleoside triphosphate, 0.2 pCi [a-32P]dCTP (3000 Ci/mmol, ICN
Biochemicals, Costa Mesa, California), and 0.625 u of Taq polymerase (Boehringer
Mannheim, GmbH, W. Germany). PCR conditions were: 94°C for 1 min, 34 cycles at
94°C for 1 min, 56°C for 40 s, and 72°C for 40 s, followed by incubation at 72°C for
2 min. PCR products from matched normal and tumor DNAs were analyzed by
electrophoresis in denaturing 8 % polyacrylamide-7M urea gels, followed by
autoradiography. For informative (i.e., heterozygous) cases, allelic loss was scored
when intensity of the signal for a tumor allele was significantly reduced (by > 95%)
relative to the matched normal allele.
180
SSCP analysis and sequencing of p53 gene
Exons 5-8 of the p53 gene were screened for mutations by SSCP analysis.
Each exon was individually amplified in 2 serial PCRs using nested primer pairs.
m
Conditions used for PCR and SSCP have been described previously (Spruck et al.,
1994a).
Exons which showed migration shifts by SSCP were reamplified without
labeled nucleoside triphosphate and the amplification products were subjected to
electrophoresis in 2% agarose gels and purified further using Mermaid Kit (Bio 101,
La Jolla, CA). Sequencing was performed using the Sequenase Kit version 2.0 (US
Biochemicals, Cleveland, OH) and p55-specific primers. Sequencing reaction
products were processed and analyzed as described for the pl6/CDKN2 gene. SSCP
analysis and sequencing reactions were performed by Dr. Atsuko Shibata and Dr.
Petra Ohneseit.
Immunohistochemical analysis of the p53 protein
Deparaffinized, rehydrated tissue sections were treated with 0.3% hydrogen
peroxide in PBS to block the endogenous peroxidase activity . Each section was then
subjected to microwave antigen retrieval in deionized water for 3.5 min at 700W in a
domestic microwave oven (Miele 696) operating at 245GHz. After 10 min of
preincubation in 1% BSA in lx PBS, the sections were incubated for 1 h with the
primary mouse monoclonal p53 antibody DO-7 (DAKO, Glostrup, Denmark) at
37°C. After rinsing the sections three times in lx PBS, a biotinylated rabbit anti-
181
mouse monoclonal antibody (DAKO, Glostrup, Denmark) was applied, followed by
further rinsing in lx PBS. Finally, the sections were incubated with the avidin-biotin
complex (ABC) (DAKO, Glostrup, Denmark) and developed for 15 min using 0.05%
diaminobenzidine (Sigma,-St. Louis, MO) diluted in lx PBS as a chromogen. All
antibodies were diluted 1:200 in 1% BSA in lx PBS and all incubations were
performed at room temperature for 30 min except for those with the primary antibody.
Counterstaining was done with hematoxylin. Control sections were processed in
parallel, and these included replacement of the specific antibody solution with PBS
alone and staining of known positive and negative sections. The evaluation of the
results was based on both the extent and the intensity of the staining.
Immunohistochemical analysis was performed by Christer Busch.
Statistical analyses
All statistical analyses were performed using Fisher's exact test, and all p-
values obtained were two-sided. Statistical analysis was performed by Dr. Atsuko
Shibata.
RESULTS
Homozygous deletions of the p!6/CDKN2 locus
We attempted to amplify exons 1 and 2 and flanking intronic sequences of
pl6/CDKN2 by PCR to investigate the status of the pl6/CDKN2 gene in SCCs.
Sufficient amounts of DNA were obtained from 12 SCCs in order to perform a
182
complete genetic study, i.e., analysis of alterations in pl6/CDKN2, p53, chromosome
9p, chromosome 9q, and chromosome 17p . However, due to extremely limited tumor
material in the remaining 19 SCCs DNA was only analyzed for alterations in p53,
chromosome 9q, and chromosome 17p. Our initial failure to amplify pl6/CDKN2 in 6
(4 from Egypt, 2 from Sweden) of 12 SCCs using two different primer sets suggested
the possibility of homozygous deletions. Confirmation of homozygous pl6/CDKN2
deletions by Southern blot analysis was not possible because insufficient amounts of
DNA are obtained from microdissected tissue excised from hematoxylin-eosin stained
histologic sections. Therefore, a series of experiments was performed to exclude
other possible causes for these results, such as template degradation or the presence of
inhibitors in the DNA preparations. A polymorphic microsatellite sequence located
within the androgen receptor gene (AR) was chosen as an internal control template
because the primer set used routinely in our laboratory for analysis of this sequence
produces a PCR fragment which is longer (299-320 bp) than that produced by the
p!6/CDKN2 exon 2 and exon 1 primer sets (257 bp and 141 bp, respectively).
Optimal conditions for amplification of both templates were first determined (Fig.
6.2). Subsequently, comparative multiplex PCR revealed that while the AR signal in 6
tumors was of equal intensity to that in matched normal controls and in tumors that
retained pl6/CDKN2, the signal for pl6/CDKN2 was absent in DNA from all 6
tumors (Fig. 6.3). Only more extensive amplification resulted in a pl6/CDKN2 signal
from tumor DNA (Fig. 6.4), which probably reflects the contamination of tumor
specimens with small number of normal cells. The large number of cycles (18 cycles
in the first round of PCR followed by 40 to 45 cycles in the second round of
nested PCR) needed to produce a signal from the p!6/CDKN2 gene in these 6
cases suggested a high degree of tumor cell purity, which was
183
s e e # 25 28 31 32 35 37 29
Serial
dilutions
Fig. 6.2. Example of the optimization of PCR conditions for amplification of the
androgen receptor gene in SCC specimens. Serial dilutions of DNA input were
performed before two rounds of PCR amplifications. 17 cycles were performed in
the first PCR, and 28 cycles in the second round of amplification.
Fig. 6.3. Comparative multiplex PCR showing homozygous deletions of
pl6/CDKN2 (or CDKN2) in 6 SCCs. The androgen receptor gene (A.R.)
was used as control. Primer sets for the A.R. gene and the CDKN2 gene
were included in the reactions. Numbers represent the SCC specimens. N,
normal DNA; T, tumor DNA. Matched normal and tumor DNAs were
obtained from the same hematoxylin-eosin stained section. SCC# 29 retained
CDKN2. SCC# 25, 28, 31, 32, 35, and 37 contained homozygous deletion of
CDKN2. Normal DNAs from tumors 25 and 28 were not available.
185
31 32 35 37
29T 25T 28T N T
( DKN2
00
O n
S C C # 25 28 31 32 35 37
# of cycles 35 40 45 50 55 3540 45 50 55 3540 45 50 55 3540 45 50 55 35 40 45 50 55 3540 45 50 55
Fig. 6.4. Autoradiograph demonstrating that only extensive amplification (18
cycles in the first round of PCR followed by 40 to 45 cycles in the second round
of nested PCR) resulted in a pl6/CDKN2 signal from DNA of SCCs in which
pl6/CDKN2 homozygous deletions were detected. This amplification signal most
likely derives from normal cells contaminating the tumor specimen. The number
of cycles in the first round of PCR was 18 for all reactions; increasing number of
cycles was used in the second round of PCR, which is indicated in the figure.
consistent with 100% allelic signal loss for markers in chromosome 9p (Fig. 6.5).
The validity of this comparative multiplex PCR assay has been previously established
by Cairns et al. (1994a) in a study in which homozygous deletions at 9p21 were
detected by this method in 10 primary TCCs of the bladder.
Multiplex PCR experiments, positive amplification of at least 7 markers on
chromosome 9 in all 6 cases, as well as the fact that pl6/CDKN2 sequences were
amplifiable from the DNA of normal cells microdissected from the same histologic
sections as the tumor cells, indicated that the pl6/CD KN2 gene was indeed
homozygously deleted in 6 of 12 (50%) SCCs examined (Fig. 6.6). Interestingly, this
frequency is 2.8 times higher than that we previously reported for pl6/CDKN2
alterations in primary TCCs (P<0.01, two-sided) (Spruck et al., 1994b).
Mutational analysis of the pl6/CDKN2 gene
Direct sequencing of exon 1 in the SCCs with no homozygous deletions of
pl6/CDKN2 (6 of 12) revealed no mutations in any of the specimens. SSCP analysis
of exon 2 of the p!6/CDKN2 gene in 12 SCCs revealed aberrant fragment migration
in 2 cases (SCC# 33, and 36) compared to the normal pattern. DNA sequencing of
the amplified product from SCC# 33 revealed that one of the two alleles of
p!6/CDKN2 contained a QAG (Glu) to AAG (Lys) missense base change at
nucleotide position 199 (nucleotide position numbering according to Serrano et al.,
1993); the other allele contained two separate base changes: a £GA (Arg) to IG A
(Stop) change at position 166 and a GA£ (Asp) to GAA (Glu) change at position
246. The pl6/CDKN2 DNA from the normal tissue of this patient showed the wild
188
31 32 35 37
N T N T NT NT
D 9S156
Fig. 6.5. Microsatellite analysis of SCC# 31, 32, 35, and 37, which contain
homozygous deletions of pl6/CDKN2. N, normal DNA; T, tumor DNA.
LOH is evident at D9S156. Allelic loss in the tumors is 100%, indicating a
high degree of tumor purity in these specimens.
189
SCC#
25 28 31 32 35 37 18 19 29 33 34 36
D9S156
DK162
CDKN2
D95171
D9S153 f~~~1 f~~~1
[ = □ c r i □ pppp
m m
^P
ni ni ni
m ni m
a O
Fig. 6.6. Deletion map of chromosome 9 in SCCs. SCC specimen # is indicated at the top. Markers analyzed and their relative
genetic locations are illustrated on the left. Numbers between markers are consensus intermarker distances in cM (sex-average)
(Kwiatkowski et al., 1993). All markers except cl.b are microsatellite markers; cl.b is a STS (Sequence Tagged Site) derived
marker (Weaver-Feldhaus et al., 1994). # homozygous deletion of pl6/CDKN2, O retention of pl6/CDKN2, t-Jretenlion of cl .b,
® homozygous deletion of cl.b, I — I retention of heterozygosity, i&3 LOH, homozygous deletion of microsatellite marker,
type sequence for both alleles, indicating that the base changes were not
polymorphisms. The somatic nonsense mutation 166C --> T has been previously
detected as a germline mutation in a familial melanoma kindred (Hussussian et al.,
1994). DNA sequencing of the amplified product from SCC# 36 revealed a fiCA
(Ala) to XCA (Ser) missense base change at nucleotide position 373. No base
changes were detected in the corresponding normal tissue DNA.
A llelic loss o f chromosome 9
Loss of heterozygosity (LOH) was analyzed by PCR amplification of 6
microsatellite markers located at 9p21-22 to determine whether loss of 9p loci
surrounding the pl6/C D K N2 gene occurred in the set of SCCs for which
pl6/CDKN2 was examined (Fig. 6.6). Remarkably, 9 of 1 1 (82%) informative SCCs
examined showed LOH of at least one of the markers in this region. Figure 6.6 shows
that the homozygous deletions involving the pl6/CDKN2 gene extended to the loci
IFNA, D9S171, and D9S169, located approximately 0.5, 4, and 8 Mb from
p!6/CDKN2, respectively (Kwiatkowski et al., 1993). In 5 of the 6 cases, the
homozygous deletions extended in a centromeric direction and included D9S171
(SCC# 31, 32, 35, 37), or both D9S171 and D9S169 (SCC# 28). In contrast, in the
remaining SCC which harbored a p!6/CDKN2 homozygous deletion (SCC# 25), the
area of homozygous deletion extended towards the telomere to include the IFNA
marker, but did not include any of the markers centromeric to the pl6/CDKN2 locus
(Fig. 6.6). Thus, the region of homozygous deletion including the pl6/CDKN2 gene
ranged from <1 Mb (in SCC# 25) to 8 Mb (in SCC# 28). The results from SCC# 25
indicate that the targeted locus is telomeric to the cl.b marker, while results from
191
SCCs# 28, 31, 32, 35 and 37 define the distal border of the critical region as
centromeric to the IFNA marker. The region defined by these two markers spans
about 1 Mb and contains the pl6/CDKN2 gene and the more recently isolated
p!5iNK4B gene, which is a member of the p!6/CDKN2 family (Hannon et al., 1994).
The status of the long arm of chromosome 9 was also determined by PCR-
based analysis of dinucleotide repeat polymorphisms. LOH in chromosome 9q was
detected in 5 of 27 (18%) informative SCCs (Table 6.1). Interestingly, 8 of 10 (80%)
informative SCCs showed LOH in 9p with retention of all markers in 9q (Table 6.1,
and Fig. 6.6), a relatively rare event in TCC tumorigenesis (Cairns et al., 1994a; Miyao
et al., 1993; Ruppert et al., 1993; Keen and Knowles, 1994). Furthermore, all tumors
containing pl6/CDKN2 alterations showed retention of chromosome 9q markers.
These data strongly support the importance of chromosome 9p alterations in the
development of SCC.
Sequence analysis of the pS3 gene and allelic loss of chromosome 17p
SSCP analysis of exons 5 through 8, direct sequencing, and
immunohistochemical analysis were used to determine the status of the p53 gene in
SCCs. Table 6.1 shows that 12 of 20 (60%) SCCs contained p53 gene mutations,
including 8 of 15 (53%) tumors from Egypt and 4 of 5 (80%) tumors from Sweden.
One tumor from Egypt (SCC# 37) was found to contain two separate p53 mutations:
a T£A (Ser) to TfiA (Stop) mutation at codon 183, and a CTQ (Leu) to CT£ (Leu)
mutation at codon 130 that did not alter the amino acid sequence of the protein (Fig.
6.7). No obvious differences were observed in terms of the position or type of p53
192
Table 6.1. Status of the p!6/CDKN2 and p53 genes, and allelic loss of 9p, 9q, and
17p in SCC of the bladder
specimens0 pl6/CDKN2b 9pc
9q
p53
17p
25 S HD no LOH no LOH nd nd
28 E HD ni no LOH nd ni
31 E HD LOH no LOH wt ni
32 E HD LOH no LOH wt no LOH
35 S HD LOH no LOH nd ni
37 E HD LOH no LOH MT LOH
18 S wt no LOH no LOH nd nd
19 E wt LOH no LOH MT LOH
29 E wt LOH no LOH nd no LOH
33 E MT LOH no LOH nd no LOH
34 E wt LOH no LOH wt LOH
36 E MT LOH ni nd LOH
2 E nd* nd* LOH MT ni
3 E nd* nd* no LOH MT ni
5 E nd* nd* LOH MT LOH
6 E nd* nd* no LOH wt ni
9 E nd* nd* no LOH MT no LOH
H E nd* nd* no LOH MT no LOH
13 E nd* nd* no LOH wt ni
16 E nd* nd* no LOH wt no LOH
22 E nd* nd* no LOH wt no LOH
26 E nd* nd* ni MT ni
4S nd* .nd* no LOH MT ni
7 S nd* nd* LOH MT ni
8S nd* nd* no LOH nd ni
14 S nd* nd* ni MT ni
193
Table 6.1. (continued).
15 S nd* nd* LOH MT no LOH
20S nd* * nd* no LOH nd ni
23S nd* nd* LOH nd ni
24S nd* nd* no LOH nd ni
30 S nd* nd* ni wt ni
Alterations LOH LOH Mutations LOH
8/12(67%) 9/11(82%) 5/27(18%) 12/20(60%)
5/13(38%)
aE, specimen from Egypt; S, specimen from Sweden
f c HD, homozygous deletion; wt, wild type; MT, mutant; nd, not determined
cLOH, loss of heterozygosity; ni, non-informative
*DNA from these specimens was not available for analysis of pl6/CDKN2
alterations and 9p allelic loss
194
Fig. 6.7. Autoradiograph showing that identical CGC --> CAC point
mutations were identified at codon 175 of the p53 gene in 2 SCC (tumors #
3 and #5). WT= wild-type control. DNA sequencing was performed by
Dr. Atsuko Shibata.
195
mutations comparing the Swedish and Egyptian tumors. Figure 6.8 shows that 11/13
mutations (85%) were transitions, 10 (91%) of which were G:C to A:T base changes.
Five transition mutations occurred at CpG dinucleotides, indicative of the deamination
of 5-methylcytosine (Jones-et al., 1992). Three of the mutations were identical CQC
(Arg) to CAC (His) base changes at codon 175, which have not been reported
previously in the analysis of TCCs except in some bladder tumors from patients
exposed to inflammation-inducing agents, such as phenacetin and arsenic (Petersen et
al., 1993; Shibata et al., 1994). On the other hand, no mutation was found at codon
280 in SCCs, a previously identified mutational hotspot in TCCs (Fig. 6.8).
Eight of the 12 tumors with detectable missense mutations were found to stain
positive for the p53 protein (Fig. 6.9), and 4 tumors stained positive for the p53
protein but had no detectable mutations by SSCP analysis within exons 5-8 of the
gene, possibly indicating mutations outside of the exons examined.
LOH in chromosome 17p was observed in 5 of 13 (38%) informative cases.
Three of the 5 cases with chromosome 17p LOH were found to contain p53 gene
mutations. In addition, three cases harbored a p53 mutation without concomitant loss
of chromosome 17p markers.
DISCUSSION
Most studies on the molecular genetics of bladder cancer have focused on
TCC, and little is known about the genetic changes that occur in SCC, whose
196
I
140 1(0 110 200 220 240 2 (0 200 200 320
. Codon
1 ;
A *
\ W ? V V ▼ Bladder (SCC)
. v v Y I v
K>
< c
D >
w V 7 I TV
t
Bladder (SCC.
A A AA A A A A A
Tahm noee,
and
phanacetln)
w
Fig. 6.8. Spectra of p53 mutations in bladder cancer. Top: squamous cell carcinomas (this study).
Middle: squamous (this study) and transitional cell carcinomas associated with inflammation-inducing
agents (Petersen et al., 1993; Shibata et al., 1994; Habuchi et al., 1993). Bottom: summary of mutations in
transitional cell carcinomas (Fujimoto et al., 1992; Spruck et al., 1993; Habuchi et al., 1993). ▼: transition
at CpG dinucleotide; V : transition at non-CpG site; A : transversion.
S ' P f c , ' '
» ' \ i V * ' * . ■ * _ ' + . « n # 1
4 ^ . * * — V-A^*V
* « ■ • • . * ’ , % • • ' f
r • % 9' ; .
_ - -
• ' . J T
>a
*
* 4
• • • •
» 4 *
f ••
Fig. 6.9. Positive immunohistochemical staining of the p53 protein in a
squamous cell carcinoma specimen. The mouse monoclonal p53 antibody DO-7
(DAKO, Glostrup, Denmark) was used. Immunohistochemical analysis was
performed by Christer Busch.
1 9 8
geographical distribution as well as the associated risk factors are significantly
different to those of TCC. In the present study we sought not only to understand the
genetic alterations involved in SCC development, but also to determine if the tumor
tissue in SCC and TCC of the bladder show differences at the molecular level.
Comparison of the genetic changes in SCC observed in this study with those
reported for TCC (Fig. 6.10) revealed that distinct alterations exist in both tumor
types. Notably, the status of chromosome 9 is quite different between SCC and TCC.
Allelic loss in 9p was observed twice as frequently in SCC (9/11) as in TCC (35/90)
(unpublished TCC data from our group) (P=0.001, two-sided), whereas allelic loss in
9q was three times less frequent in SCC. Also the frequency of hemi- and
homozygous deletions in 9p with retention of all markers in 9q was extremely high
(91%, 10/11) in SCC (Fig. 6.10). In sharp contrast, the majority of TCCs described
in the literature show monosomy 9, and very few cases (10%, 11/110) (Miyao et al.,
1993; Cairns et al., 1994a; Ruppert et al., 1993; Keen and Knowles, 1994; Linnenbach
et al., 1993) retain all the chromosome markers in the q arm and lose only some
markers in the p arm (P<0.001, two-sided).
SCC and TCC were also found to differ in the frequency of pl6/CDKN2
alterations (P=0.009, two-sided), with SCC containing homozygous deletions and
point mutations within this gene three times more frequently (8/12) than what we
previously observed in TCC (8/40) (Chapter 4; Spruck et al., 1994b) (Fig. 6.10). It is
remarkable that the frequency of these defects in SCC is similar to that we observed
for bladder carcinoma derived cell lines (Spruck et al., 1994b). Reciprocal
pl6/CDKN2 expression and retinoblastoma protein (pRb) inactivation has been
199
100
90
80
70
60
% 50
40
30
20
10
0
Fig. 6.10. Comparison of genetic alterations observed in SCC and TCC of the
bladder. Data for SCC were derived solely from this study. Data for TCC are a
compilation from previous reports (Caims et al., 1994a; Stadler et al., 1994; Miyao
et al., 1993; Knowles et al., 1994; Ruppert et al., 1993; Keen and Knowles, 1994;
Sidransky et al., 1991; Fujimoto et al., 1992; Sarkis et al., 1993; Spruck et al.,
1994a).
□ TCC
1
9pLOH 9qLOH 9pdeletion CDKN2 p53
9q retention
200
documented (Yeager et al., 1995; Shapiro et al., 1995; Lukas et al., 1995). Therefore,
it would be interesting to examine the status of pRb in our SCC specimens.
Unfortunately, all histologic sections obtained from Egypt and Sweden for this study
were hematoxilin-eosin stained and used for tissue microdisection for DNA isolation.
Interestingly, a high frequency of p!6/CDKN2 alterations has also been reported in
uncultured squamous cell carcinomas of the esophagus (Mori et al., 1994), and in
pancreatic adenocarcinomas (Caldas et al., 1994). Clinically, these three tumor types
have a poor prognosis. All the homozygous deletions detected on 9p21-22 contained
but were not limited to pl6/CDKN2, extending towards the centromere up to 4 to 8
Mb in 5 tumors, and towards the telomere up to 0.5 Mb in 1 tumor, possibly
implicating other genes in the deleted region. In parallel to our observations, Cheng et
al. (1994) have reported two separate regions of homozygous loss within 9p21-22 in
malignant mesothelioma, one proximal to and the other distal to p!6/CDKN2. The
minimal area of homozygous deletion defined by our analysis was localized to a
region of aproximately 1 Mb, where the pl6/CDKN2 and pl5,N K 4B genes reside.
This deleted region in SCC of the bladder may be smaller than that previously mapped
in TCC of the bladder (Caims et al., 1994a; Stadler et al., 1994) and other tumor types
(Olopade et al., 1992; van der Riet et al., 1994), which spanned about 10 Mb.
Homozygous deletions, however, do not clearly define the exact gene(s) targeted in
tumorigenesis. In our study, intragenic mutations within the pl6/CDKN2 gene were
also detected in two unrelated SCCs, suggesting that pl6/CDKN2 may well be the
9p21 gene critical in SCC development. Specifically, the pl6/CDKN2 mutations
detected in two SCCs lie within the p 16 protein ankyrin repeats, which are motifs
involved in protein-protein interactions and presumed to be crucial for the pl6 protein
function (Yang et al., 1995). Tumor suppressor genes are typically characterized by
201
functional inactivation of both alleles, and all the pl6/CDKN2 alterations found in our
panel of SCCs represented homozygous changes. Therefore, these findings are
consistent with inactivation of both alleles of the pl6/CDKN2 putative tumor
suppressor gene, and suggest that pl6/CDKN2 may play a key role in SCC
tumorigenesis.
Another interesting difference between SCC and TCC was the type and
distribution of p53 mutations in the two tumor types (Fig. 6.8). Overall, 77% (10/13)
of p53 mutations detected in SCCs were G:C to A:T transitions, whereas only 47% of
the previously reported mutations in TCCs were of this type (Fujimoto et al., 1992;
Sarkis et al., 1993; Spruck et al., 1993; Spruck et al., 1994a). Additionally, half of the
p53 mutations observed in SCCs were transition mutations at CpG dinucleotides,
indicative of deamination of 5-methylcytosine (Jones et al., 1992), and 60% of those
mutations were identical CQC (Arg) to CAC (His) base changes at codon 175 (Fig.
6.8, top). This mutation has been previously observed in a few TCCs obtained from
patients chronically exposed to phenacetin (Petersen et al., 1993) and arsenic (Shibata
et al., 1994) (Fig. 6.8, middle), but has not been previously reported in the analysis of
TCCs which are not linked to exposure to known risk factors (Fig. 6.8, bottom).
Additionally, the codon 280 mutational hotspot reported in non exposure-linked TCCs
(Spruck et al., 1994) was not observed in the SCCs analyzed in this study or in those
tumors associated with phenacetin or arsenic exposure. These agents, like the
potential risk factors of SCCs, have been proposed to produce chronic urothelial
damage and growth stimulation.
202
In the schistosomiasis affected bladder, Schistosoma eggs reside in the
bladder musculature providing a constant source of irritation and damage to the
urothelial lining, which proposedly leads to a steady turnover of the urothelium
(Badawi et al., 1992). Other sources of long-standing chronic irritation of the
urothelium are lithiasis, chronic urinary tract infections, and chronic catheterization
which are also associated with increased risk of SCC. The stimulation of urothelium
growth could potentiate the action of environmental mutagens, such as N-nitroso
compounds, which are known to be at measurably higher levels in the endemic regions
of schistosomiasis (Hicks, 1982; Badawi et al., 1992). Interestingly, N-nitroso
compounds are known to form a variety of adducts with DNA, the most common of
which is the alkylation of guanine at the O& position (O'Connor et al., 1979).
Following replication, these adducts have been shown to induce specifically G to A
mutations in DNA (Saffhill et al., 1985), as those observed in 77% (10/13) of the
mutations detected in our panel of SCCs.
Differences in molecular defects have been observed between superficial
papillary TCC and carcinoma in situ of the bladder (Spruck et al., 1994a), leading to
the concept that two molecular pathways exist in bladder TCC tumorigenesis. Our
results suggest that SCC tumorigenesis may follow yet another genetic pathway
distinct from those leading to TCC, which likely involves the preferential loss of a
tumor suppressor gene, possibly pl6/CDKN2, on chromosome 9p (Fig. 6.11). Thus,
the different histologic and clinical characteristics of these two types of urothelial
cancers may reflect different underlying genetic mechanisms of tumorigenesis.
Divergent molecular pathways may also exist in other tumor types, such as
rhabdomyosarcomas in which the embryonal and the alveolar forms typically contain
203
Normal
Urothelium
lillary Low GradewPapillary High Grade
Non-In vasive -Sk Non-Invasive
TCC J L TCC j
Squamous Cell
Metaplasia
Lamina Propria Invasive
TCC
Squamous Cell
Carcinoma
Muscle Invasive
TCC
- ! f = a s'‘ ' ,,s S
Metastasis J
Fig. 6.11. Proposed model for bladder cancer progression based on chromosome
9 deletion data (9-) and p53 gene mutation data (p53-) from the present study and
a previously reported study (Spruck et al., 1994a). This model is an expansion of
a previously proposed model for TCC progression (Spruck et al., 1994a) and
includes a new proposed genetic pathway for SCC development. TCCs are
stratified in light-shaded boxes according to their invasion into the bladder wall.
Squamous cell metaplasia and SCC are stratified in dark-shaded boxes.
Chromosome 9q deletions (9q-) arid p53 gene mutations (p53-) have been shown
to occur early in TCC tumorigenesis leading to superficial papillary tumors (Ta)
or to carcinoma in situ (Cis), respectively (Spruck et al., 1994a); further
alterations would lead to progression of Ta and Cis into more invasive forms of
TCC. SCC tumorigenesis may follow a distinct genetic pathway which likely
involves the preferential loss of a tumor suppressor gene on chromosome 9p (9p-
), such as the pl6/CDKN2 gene, and mutations in the p53 gene (p53-).
204
different chromosomal alterations (Scrable et al., 1989). Similarly, squamous cell
carcinoma of the lung has been shown to be genetically distinct from adenocarcinoma
(Sato et al., 1994).
Our study contributes to a better understanding of the genetic characteristics
of SCC, which may eventually lead to improvement in diagnosis and therapy of SCC.
Potentially, SCC diagnosis could be attained by detection of p!6/CDKN2 and
chromosome 9p alterations in exfoliated cells present in the urine of individuals at a
high risk of developing SCC, such as those infected by Schistosoma or those
suffering from chronic urinary tract infections and lithiasis.
205
CHAPTER 7
SUMMARY AND CONCLUSIONS
It is essential to understand the etiology and pathogenesis of a disease to treat
it rationally. Cancer is a genetic disease in which the progressive accumulation of
molecular defects results in the neoplastic phenotype. Therefore, an increasing
understanding of the molecular genetic mechanisms of the neoplastic process should
lead to earlier diagnosis, more precise prognosis and novel and effective treatments.
Although human bladder cancer is one of the most prevalent cancers worldwide and
the most common cancer in men in some parts of the world, the molecular genetic
alterations involved in bladder tumorigenesis have not begun to be understood until
recently. Further insight into the genetic nature of bladder cancer has been gained by
the use of modem molecular biology and immunohistochemistry techniques. The
research work described in this thesis was designed to investigate the genetic changes
underlying bladder cancer.
Molecular abnormalities have been previously identified in bladder transitional
cell carcinomas by Dr. Jones' group and others (Tsai et al., 1990; Olumi et al., 1990;
Sidransky et al., 1991; Cairns et al., 1993; Habuchi et al., 1993a; Miyao et al., 1993;
Knowles et al., 1994; Spruck et al., 1994a). Loss of genetic material on chromosome
9, reflecting inactivation of a tumor suppressor gene, is thought to be an initiating
event in superficial papillary tumors which have a more proliferative than invasive
tendency (Spruck et al., 1994a). Intense research to locate the tumor suppressor
206
genes on this chromosome is underway. Part of those efforts are presented in
Chapter 2 of this thesis, which describes the deletion mapping of chromosome 9 in a
large number of primary transitional cell carcinomas. This study used novel
molecular biology techniques that allowed for the analysis of allelic losses of
chromosome 9 in both fresh-frozen and archival tissue specimens. Microsatellite
analysis revealed that 37% of the primary bladder TCCs examined were monosomic
for chromosome 9, confirming previous findings by the Jones' group and others.
Eighteen per cent of the tumors were found to contain partial subchromosomal losses
on chromosome 9, and further analysis led to the definition of two minimal areas of
common deletion on this chromosome in our panel of bladder TCCs. One area of
common deletion was mapped to the distal portion of the long arm of chromosome 9,
telomeric to the GSN locus on 9q33. The second region of deletion was mapped to
9pl3-q21.3 between markers D9S171 and D9S283. These results not only expanded
previously reported data but also indicated that two tumor suppressor loci may be
associated with the long arm of chromosome 9 in bladder TCC. Several genes reside
within these two loci on chromosome 9. Further characterization of those loci and
investigation of already identified genes in those regions will help in the search of a
tumor suppressor(s) involved in bladder TCC tumorigenesis.
Microsatellite polymorphism analysis revealed that genomic instability occurs
in bladder TCC, as reflected by alterations in microsatellites in tumor specimens. The
study described in Chapter 3 was the first one to report on microsatellite instability in
bladder cancer. This type of genetic alteration is associated with deficient DNA
mismatch repair activity, which leads to hypermutability and cancer predispodition
(Leach et al., 1993; Drummond et al., 1993). All the specimens that showed
207
microsatellite instability were low-grade, low-stage tumors, which suggests that this
type of genetic alteration may occur early in bladder TCC tumorigenesis.
Alterations in chromosomal band 9p21 have been shown to occur frequently
in bladder and other cancer types (Fountain et al., 1992; Olopade et al., 1993; Stadler
et al., 1994; van der Riet et al., 1994; Dreylin et al.,1995). The search for a tumor
suppressor on this region led to the identification of the p!6/CDKN2 gene (Kamb et
al., 1994; Nobori et al., 1994), which was initially proposed to be involved in the
genesis of many tumor types and has been more recently shown to be important in
familial melanoma tumorigenesis (Hussussian et al., 1994), and in the progression of
sporadic melanoma (Reed et al., 1995) and lung cancer (Okamoto et al., 1995). Since
LOH at the 9p21 region is a frequent genetic alteration in bladder TCCs (Cairns et al.,
1994a), the role of the pl6/CDKN2 gene as a tumor suppressor in this type of bladder
cancer was investigated. In the study presented in Chapter 4, uncultured primary
TCCs and bladder cancer-derived cell lines were examined for alterations in the
pl6/CDKN2 suppressor gene, including homozygous deletions and intragenic
mutations. Only 7 of 40 (17%) uncultured tumors contained p!6/C D K N 2
homozygous deletions, and one tumor (2%) showed an intragenic mutation. These
alterations were found in tumors of various stages and grades, which suggested that
these type of pl6/CDKN2 changes were not involved in a specific step of bladder
tumorigenesis. Moreover, the frequency of these type of defects was found to be
almost three times higher in cell lines (54%) than in uncultured tumors (20%). These
results indicated that homozygous deletions and intragenic mutations of pl6/CDKN2
are not a frequent mechanism of inactivation of this gene in uncultured TCCs.
However, these alterations are significantly more frequent in cell lines and may
208
provide a selective advantage for growth in vitro. Various groups have also detected
pl6/CDKN2 homozygous deletions and intragenic mutations at a higher frequency in
cell lines than in uncultured tumors (Cheng et al., 1994; Xu et al., 1994; Ohta et al.,
1994; Zhan et al., 1994), and the issue of pl6/CDKN2 involvement in in vitro growth
remains an open matter of debate. Cairns et al. (1994b) also reported infrequent
pl6/CDKN2 homozygous deletions and point mutations in bladder TCCs and other
tumor-types containing allelic loss of markers on 9p21. These observations
questioned pl6/CDKN2 gene as the target for LOH in this region. Other genes with a
tumor suppressor function may be located in this area. Alternatively, pl6/CDKN2
may be more frequently inactivated by other mechanisms, such as mutations in gene
regulatory sequences, or abnormal DNA methylation of promoter and regulatory
sequences.
The role of DNA methylation in controlling p!6/CDKN2 expression was
investigated, and is presented in Chapter 5. A highly significant correlation between
methylation of the 5' CpG island of pl6/CDKN2 and transcriptional inactivity was
detected in bladder uncultured TCCs, cell lines and normal colon mucosa, whereas no
association was found between methylation of exon 2 of pl6/CDKN2 and gene
silencing. Twelve of 18 (67%) of primary bladder TCCs were shown to contain a
methylated pl6/CDKN2 5'CpG island, which suggested that abnormal methylation of
this region may be a more common mechanism to inactivate p!6/CDKN2. These
results may explain the discrepancy between the reported frequency of 9p21 allelic
loss and pl6/CDKN2 homozygous deletions and intragenic mutations in bladder
TCCs. Inactivation of p!6/CDKN2 by methylation of its 5' CpG island has also been
proposed as a frequent mechanism in other tumor types in which 9p21 LOH occurs
209
frequently but p!6/CDKN2 homozygous deletions or intragenic mutations are rare
(Merlo et al., 1995; Herman et al., 1995). Although the role of abnormal methylation
in silencing p!6/CDKN2 awaits direct and definitive proof, our findings and those of
others have important and inmediate therapeutic implications. Preliminary evidence
from the Jones group (Chapter 5; Gonzalez-Zulueta et al., manuscript in preparation)
indicates the possibility of using 5-Aza-2'-deoxycytidine, a DNA methylation
inhibitor, to induce the expression of the pl6/CDKN2 tumor suppressor gene in
neoplastic cells that have inactivated the gene by methylation. Thus, reactivation of
p!6/CDKN2 has been shown in vitro and in vivo by treatment with 5-Aza-
2'deoxycytidine, and preliminary results indicate that reexpression of the gene is
associated with inhibition of cell proliferation. Experiments to prove the direct
involvement of pl6/CDKN2 induction in growth inhibition are in progress. These
findings may lead to the design of novel therapeutic strategies involving the use of 5-
Aza-2'-deoxycytidine in human cancers in which methylation of pl6/CDKN2 plays a
role in tumorigenesis and progression.
Most of the studies on the molecular genetic alterations underlying bladder
cancer have focused on TCCs, and little is known about SCC tumorigenesis.
Although SCC represents only 1% of bladder cancers in the Western Hemisphere, it
is the most common bladder cancer and the most common cancer in men in other
parts of the world, such as areas in which schistosomiasis is endemic. The efforts to
understand the genetic defects that occur in SCC tumorigenesis are described in
Chapter 6. Specimens from SCC-high-risk (Egypt) and low-risk areas (Sweden)
were explored for genetic alterations previously involved in bladder TCCs. The
results indicated that the clinically and morphologically different SCC and TCC also
210
differ at the molecular genetic level. Homozygous deletions and sequence mutations
in the pl6/CDKN2 gene were found in 67% (8/12) of SCCs, a frequency three times
higher than that reported for uncultured TCCs. Hemi- and homozygous deletions in
9p, where pl6/CDKN2 resides, were found in 91% (11/12) of uncultured SCCs, while
only about 39% (35/90) of TCCs showed these losses (P=0.001). Interestingly,
deletions in 9p with retention of 9q were found in 91% (10/11) of SCCs compared to
only 10% (11/110) of TCCs (P<0.001) reported in the literature. These results
suggested that a putative tumor suppressor gene on 9p, possibly pl6/CDKN2, may
contribute to SCC tumorigenesis. The frequency of p53 mutations in SCCs was
similar to that reported for invasive TCCs, but the type and position of mutations
differed between the two tumor types. The data indicated that SCC and TCC differ in
their genetic alterations, suggesting that pathological differences between the two
tumors of the bladder epithelium may be explained, at least in part, by distinct
underlying genetic defects. These findings have important clinical implications, as
development of diagnostic and therapeutic strategies for SCC of the bladder based on
its distinct genetic alterations is warranted. Thus, SCC diagnosis could be attained by
detection of pl6/CDKN2 and chromosome 9p alterations in exfoliated cells present in
the urine of individuals at high risk of developing SCC, such as those infected by
Schistosoma or those suffering from chronic urinary tract infection and lithiasis.
Certainly there are events, in the genetic pathways leading to bladder
tumorigenesis that remain unknown. The primary goal of the research described in
this thesis was to advance the task of establishing a correlation between specific
genotypic alterations and specific phenotypic changes. This correlation needs to be
translated into diagnostic and therapeutic advances in order to increase the quality of
211
life and life expectancy of bladder cancer patients. Overall, the work described in this
thesis has contributed to a better understanding of the genetic changes involved in
bladder cancer, and specifically, has provided evidence that novel diagnostic and
therapeutic strategies can be achieved. Detection of pl6/CDKN2 alterations in the
urine of patients with SCC could be applied for early diagnosis of this bladder cancer
in populations at high risk of developing this tumor type. Furthermore, the efficacy of
therapeutic strategies aimed towards reactivation of a dormant, methylated
pl6/CDKN2 gene with chemotherapeutic agents, as well as gene therapy approaches
will have to be tested in human clinical trials.
212
REFERENCES
Aaltonen, L., A., Peltomaki, P., Leach, F., S., Sistonen, P., Pylkkanen, L., Mecklin, J-P.,
Jarvinen, H., Powell, S., M., Jen, J., Hamilton, S., R., Petersen, G., M., Kinzler, K., W.,
Vogelstein, B., de la Chapelle, A. Clues to the pathogenesis of familial colorectal
cancer. Science, 260: 812-816,1993.
Ah-See, K., W., Cooke, T., G., Pickford, I., R., Soutar, D., Balmain, A. An allelotype
of squamous carcinoma of the head and neck using microsatellite markers. Cancer
Res., 54: 1617-1621,1994.
Ahlering, T., E., Dubeau, L., Jones, P., A. A new in vivo model to study invasion and
metastasis of human bladder carcinoma. Cancer Res, 47: 6660-6665,1987.
Anderson, J., A., Irish, J., C., Ngan, B., Y. Prevalence of ras oncogene mutations in
head and neck carcinomas. J. Otolaryngol., 21:321-326,1992.
Antequera, F., D., Macleod, A., P., Bird, A. Specific protection of methylated CpGs in
mammalian nuclei. Cell, 58:509-517,1989.
Antequera, F., J., Boyes, Bird, A. High levels of de novo methylation and altered
chromatin structure at CpG islands in cell lines. Cell, 62:503-514,1990.
Atkin, N., B., Baker, M., C. Cytogenetic study of ten carcinomas of the bladder:
Involvement of chromosomes 1 and 11. Cancer Genet. Cytogenet., 15:253-268,1985.
Augustine, A., Herbert, J., R., Kabat, G., C., Wynder, E., L. Bladder cancer in relation
to cigarette smoking. Cancer Res., 48:4405-4408,1988.
Babu, V., R., Lutz, M., D., Miles, B., J., Farah, R., N., Weiss, L., Van, D., D. Tumor
behavior in transitional cell carcinoma of the bladder in relation to chromosomal
markers and histopathology. Cancer Res., 47:6800-6805,1987.
Babu, V., R., Miles, B., J., Cemey, J., C., Weiss, L., van Dyke, D., L. Chromosome
21q22 deletion: A specific chromosome change in a new bladder cancer subgroup.
Cancer Genet. Cytogenet., 38:127-129,1989.
213
Badawi, A., F., Mostafa, M., H., Aboul-Azm, T., Haboubi, N., Y., O'Connor, P., J.,
Cooper, D., P. Promutagenic methylation damage in bladder DNA from patients with
bladder cancer associated with schistosomiasis and from normal individuals.
Carcinogenesis, 13: 877-881,1992
Barlow, D., P., Stoger, R.„ Herrmann, B., G., Saito, K., Schweifer, N. The mouse
insulin-like growth factor type-2 receptor is imprinted and closely linked to the Tme
locus. Nature, 349: 84-87,1991
Baylin, S., B., Hoppener, J., W., DeBustros, A., Steenbergh, P., H., Lips, C., J., Nelkin,
B. D. DNA methylation patterns of the calcitonin gene in human lung cancers and
lymphomas. Cancer Res., 46: 2917-2922,1986.
Baylin, S., B., Fearon, E., R., Vogelstein, B., DeBustros, A., Sharkis, S., J., Burke, P.,
J., Staal, S., P., Nelkin, B., D. Hypeimethylation of the 5' region of the calcitonin gene
is a property of human lymphoid and acute myeloid malignancies. Blood, 70: 412-
417,1987.
Bears, O., H., Henson, D., E., Hutter, R., V., P., Myers, M., H. (eds.). American Joint
Committee on Cancer. Manual for staging of cancer, 3rd ed. pp. 193-198,
Philadelphia: J.P. Lippincott, 1988.
Beckman, J., Weber, J., L. Survey of human and rat microsatellites. Genomics, 12:
627-631,1992.
Bell, G., I., Karam, J., H., Rutter, W., J. Polymorphic DNA region adjacent to the 5'
end of the human insulin gene. Proc. Natl. Acad. Sci. USA, 78: 5759-5763,1981.
Berger, M., S., Greenfield, C., Gullick, W., J., Haley, J., Downward, J., Neal, D., E.,
Harris, A., L., Waterfield, M., D. Evaluation of epidermal growth factor receptors in
bladder tumors. Br J Cancer, 56: 533-537,1987.
Berger, C., S., Sandberg, A., A., Todd, I., A., D., Pennington, R., D., Haddad, F., S.,
Hecht, B., K., Hecht, F. Chromosomes in kidney, ureter, and bladder cancer. Cancer
Genet. Cytogenet., 23: 1-24,1986.
Bergkvist, A., Ljungquist, A., Moberger, G. Classification of bladder tumors based on
cellular pattern. ActaChir. Scan., 130: 371-378,1965.
Bemu6s, M., Casadevall, C., Miro, R., Caballfn, M. R., Gelabert, A., and Egozcue, J. A
case of transitional cell carcinoma of the bladder with a del(9)(ql lq21.2). Cancer
Genet. Cytogenet., 69:76-77,1993.
214
Biancalana, V., Serville, F., Pommier, J., Julien, J., Hanauer, A., and Mandel, J.L.
Moderate instability of the trinucleotide repeat in spino bulbar muscular atrophy.
Hum. Molec. Gen., 1: 255-258, 1992.
Bishop, J., M. The molecular genetics of cancer. Science, 235:305-311,1987.
Boveri, T. The Origin of Malignant Tumors. London, England, Bailliere, 1929.
Brandeis, M., Frank, D., Keshet, I., Siegfried, Z., Mendelsohn, M., Nemes, A., Temper,
V., Razin, A., Cedar, H. Spl elements protect a CpG island from de novo methylation.
Nature, 371: 435-438,1994.
Brewster, S., F., Gingell, J., C., Browne, S., Brown, K., W. Loss of heterozygosity on
chromosome 18q is associated with muscle-invasive transitional cell carcinoma of the
bladder. Br. J. Cancer, 70: 697-700,1994.
Bullrich, F., MacLachlan, T., K., Sang, N., Druck, T., Veronese, M. L., Allen, S., L.,
Chiorazzi, N., Koff, A., Heubner, K., Croce, C., M., Giordano, A. Chromosomal
mapping of members of the cdc2 family of protein kinases, cdk3, cdk6, PISSLRE, and
PITALRE, and a cdk inhibitor, p27Kipl, to regions involved in human cancer. Cancer
Res., 55: 1199-1205,1995.
Cairns, P., Proctor, A., J., Knowles, M., A. Loss of heterozygosity at the RB locus is
frequent and correlates with muscle invasion in bladder carcinoma. Oncogene, 6:2305-
2309, 1991.
Caims, P., Shaw, M. E., Knowles, M. A. Initiation of bladder cancer may involve
deletion of a tumour-suppressor gene on chromosome 9. Oncogene, 8: 1083-1085,
1993.
Caims, P., Shaw, M., E., Knowles, M., A. Preliminary mapping of the deleted region
of chromosome 9 in bladder cancer. Cancer Res., 53:1230-1232,1993.
Caims, P., Kaori, T., Eby, Y., Sidransky, D. Homozygous deletions of 9p21 in
primary human bladder tumors detected by comparative multiplex polymerase chain
reaction. Cancer Res., 54: 1422-1424,1994.
Caims, P., Mao, L., Merlo, A., Lee, D., J., Schwab, D., Eby, Y., Tokino, K., van der
Riet, P., Blaugrund, J., E., Sidransky, D. Rates of pl6 (MTS1) mutations in primary
tumors with 9p loss. Science , 265:415-416, 1994.
215
Caldas, C„ Hahn, S.,A., da Costa, L.,T., Redston, M.,S., Schutte, M., Seymour, A., B.,
Weinstein, C., L., Hruban, R., H., Yeo, C., J., Kern, S., E. Frequent somatic mutations
and homozygous deletions of the pl6 (MTS1) gene in pancreatic adenocarcinoma.
Nature Genet., 8: 27-32,1994.
Call, K., M., Glaser, T., Ito, £., Y., Buckler, A., J., Pelletier, J., Haber, D., A., Rose, E.,
A., Krai, A., Yeger, H., Lewis, W., H., Jones, C., Houseman, D., E. Isolation and
characterization of a zinc finger polypeptide gene at the human chromosome 11
Wilms’ tumor locus. Cell, 60: 509-520,1990.
Capon, D., J., Chen, E., Y., Levinson, A., D., Seeburg, P., H., Goeddel, D., V. Complete
nucleotide sequences of the T24 human bladder carcinoma oncogene and its normal
homologue. Nature, 302: 33-37, 1983.
Chen, C.J., Lin, T.M., and Wu, H.Y. Malignant neoplasms among the residents of a
black foot disease-endemic area in Taiwan: high arsenic artesian well water and
cancers. Cancer Res., 45: 5895-5899,1985.
Cheng, J., Q., Jhanvar, S., C., Klein, W., M, Bell, D., W., Lee, W., C., Altomare, D., A.,
Nobori, T., Olopade, O., I., Buckler, A., J., Testa, J., R. p l6 alterations and deletion
mapping of 9p21-p22 in malignant mesothelioma. Cancer Res., 54: 5547-5551,1994.
Cheng, J., Q., Jhanwar, S., C., Lu, Y., Y., Testa, J.R. Homozygous deletions within
9p21-p22 identify a small critical region of chromosomal loss in human malignant
mesotheliomas. Cancer Res., 53: 4761-4763,1993.
Chomczynski, P., Sacchi, N. Single-step method of RNA isolation by acid
guanidinium thiocyanate-phenol-chloroform extraction. Analytical Biochem., 162:
156-159,1987.
Cliby, W., Ritland, S., Hartmann, L., Dodson, M., Hailing, K. C., Keeney, G., Podratz,
K. C., and Jenkins, R., B. Human epithelial ovarian cancer allelotype. Cancer Res.,
53: 2393-2398,1993.
Cole, P. A population based study of bladder cancer. In: Host Environment
Interactons in the Etiology of Cancer in Man. Doll, R., Vodopija, I., (eds.),
International Agency for Research on Cancer, Lyon, pp. 83-87,1973.
Coleman, A., Fountain, J.W., Nobori, T., Olopade, O.I., Robertson, G., Houseman,
D.E., and Lugo, T.G. Distinct deletions of chromosome 9p associated with melanoma
versus glioma, lung cancer, and leukemia. Cancer Res., 54: 344-348,1994.
216
Coinings, D., E. A. general theory of carcinogenesis. Proc. Natl. Acad. Sci. USA,
70:3324-3328,1973.
Cordon-Cardo, C., Wartinger, D., Petrilak, D, Dalbagni, G., Fair, W., R., Fuks, Z.,
Reuter, V., E. Altered expression of the retinoblastoma gene product is a prognostic
indicator in bladder cancer. J. Nat. Cancer Inst., 84:1251-1256, 1992.
Cress, W., D., Johnson, D., G., Nevins, J., R. A genetic analysis of the E2F1 gene
distinguishes regulation by Rb, p i07, and adenovirus E4. Mol. Cell Biol., 13: 6314-
6325,1993.
Crissman, J., D., Barwick, K., W. In Pathologic staging of colon and rectal cancer.
Wanebo, H. (ed.). Colorectal cancer. pp57-74. Mosby: St. Louis, 1993.
Crowther, P., J., Doherty, J., P., Linsenmeyer, M., E., Williamson, M., R., Woodcock,
D., M. Revised genomic consensus for the hypermethylated CpG island region of the
human LI transposon and integration sites of full length LI elements from
recombinant clones made using methylation-tolerant host strains. Nucleic Acids Res.,
19: 2395-2401,1991.
Cutler Allen, R., Zoghbi, H.Y., Moseley, A.B., Rosenblatt, H.M., and Belmont, W.
Methylation of Hpa fi and Hha I sites near the polymorphic CAG repeat in the human
androgen receptor gene correlates with X chromosome inactivation. Am. J. Hum.
Genet., 51: 1229-1239, 1992.
Cutler, S., G., Heney, N., M., Friedell, G., H. Longitudinal study of patients with
bladder cancer: factors associated with disease recurrence and progression. In: AUA
Monographs, Vol. I: Bladder Cancer. Cooper, E.H., and Williams, R.E. eds., Williams
and Wilkins, Baltimore, pp. 35-46,1982.
Dalbagni, G., Presti, J., Reuter, V., Fair, W. R., Cordon-Cardo, C. Genetic alterations
in bladder cancer. Lancet, 342:469-471,1993.
Daly, J.J. Carcinoma in situ of the urothelium. Urol. Clin. North Am., 3: 87,1976.
Del Sal, G., Collavin, L., Ruaro, M. E., Edomi, P., Saccone, S., Della Valle, G.,
Schneider, C. Structure, function, and chromosome mapping of the growth-
suppressing human homologue of the murine gasl gene. Proc. Natl. Acad. Sci. USA,
91: 1848-1852,1994.
217
Devlin, J., Keen, A. J., Knowles, M. A. Homozygous deletion mapping at 9p21 in
bladder carcinoma defines a critical region within 2cM of IFNA. Oncogene, 9: 2757-
2760, 1994.
Diaz, M., O., Rubin, C., M., Harden, A., Ziemin, S., Larsen, R., A., Le-Beau, M., M.,
Rowley, J., D. Deletions of interferon genes in acute lymphoblastic leukemia. N.
Eng. J. Med., 322: 77-82,1990.
Diaz, M., O., Ziemin, S., Le-Beau, M., M., Pitha, P., Smith, S., D., Chilcote, R., R.,
Rowley, J., D. Homozygous deletion of the alpha- and beta 1-interferon genes in
human leukemia and derived cell lines. Proc. Natl. Acad. Sci. USA, 85: 5259-5263,
1988.
Dreylin, M., H., Bohlander, S., K., Adeyanju, M., O., Olopade, O., I. Detection of
CDKN2 deletions in tumor cell lines and primary glioma by interphase fluorescence
in situ hybridization. Cancer Res., 55: 984-988,1995.
Droller, M., J. Bladder Cancer. Monogr. Urol., 3: 131,1982.
Drummond, J., T., Li, G., M., Longley, M., J., Modrich, P. Isolation of an hMSH2-
pl60 heterodimer that restores DNA mismatch repair to tumor cells. Science, 268:
1909-1912,1995.
Eagan, J., W., Jr. Urothelial neoplasms: Urinary bladder. In: Gary, F.H. (ed.).
Uropathology. pp.793-811. New York: Churchill Livinston, 1989
Ehlen, T., Dubeau, L. Loss of heterozygosity on chromosomal segments 3p, 6q, and
1 lp in human ovarian carcinomas. Oncogene, 5:219-223,1990.
El-Deiry, W., S., Tokino, T., Levy, D., B., Parsons, R., Trent, J., M., Kinzler, K., W.,
and Vogelstein, B. WAF1, a potential mediator of p53 tumor suppression. Cell, 75:
817-825, 1993.
Esrig, D., Elmajian., D., Groshen, S., Freeman, J., A., Stein, J., P., Chen, S., C.,
Nichols, P., W., Skinner, D., G., Jones, P., A., Code, R. Accumulation of nuclear p53
and tumorprogression in bladder cancer. N. Engl. J. Med., 331: 1259-1264,1994.
Esrig, D., Spruck, C., H., Nichols, P., Chaiwun, B., Steven, K., Groshen, S., Chen, S.,
C., Skinner, D., G., Jones, P., A., Cote, R., J. p53 nuclear accumulation correlates with
mutations in the p53 gene, tumor grade, and stage in bladder cancer. Am. J. Path., 143:
1389-1397, 1993.
218
Eversole-Cire, P., Ferguson-Smith, A., Sasaki, H., Brown, K., D., Cattanach, B., M.,
Gonzales, F., Surani, M., A., Jones, P., A. Activation of an imprinted Igf 2 gene in
mouse somatic cell cultures. Mol. Cell. Biol., 13:4928-4938., 1993.
Fanning, P., Bulovas, K., Saini. K., S., Libertino, J., A., Joyce, A., D., Summerhayes, I.,
C. Elevated expression of pp60c' src in low grade human bladder carcinomas. Cancer
Res., 52:1457-1462, 1992.
Fearon, E. R., Feinberg, A. P., Hamilton, S. H., Vogelstein, B. Loss of genes on the
short arm of chromosome 1 1 in bladder cancer. Nature, 318: 377-380,1985.
Fearon, E., R., Jones, P., A. Progressing toward a molecular description of colorectal
cancer development. FASEB J., 6:2783-2790, 1992.
Fearon, E., R., Vogelstein, B. A genetic model for colorectal tumorigenesis. Cell, 61:
759-767, 1990.
Fearon, E., R., Feinberg, A., P., Hamilton, S., H., Vogelstein, B. Loss of genes on the
short arm of chromosome 1 1 in bladder cancer. Nature, 318: 377-380,1985.
Ferguson-Smith, A., Sasaki, C., H., Cattanach, B., M., Surani, M., A. Parental-origin-
specific epigenetic modification of the mouse H19 gene. Nature, 362: 751-755,1993
Fountain, J., W., Karayiogou, M., Emstoff, M., S., Kirkwood, J., M., Vlock, D., R.,
Titus-Emstoff, L., Bouchard, B., Vijayasaradhi, S., Houghton, A., N., Lahti, J., Kidd,
V., J., Housman, D., E., Dracopoli, N., C. Homozygous deletions within human
chromosome band 9p21 in melanoma. Proc. Natl. Acad. Sci. USA, 89: 10557-10561,
1992.
Freidell, G., H., Soloway, M., Hilgar, A. et al. Summary on workshop on carcinoma in
situ of the bladder. J. Urol., 136:1047,1986.
Friedell, G., J. The role of urinary tract cytology in the detection and clinical
mangement of bladder cancer. In: Bladder Cancer, AUA Monographs, Bonney, W.,
W., Prout, G., R. (eds.), Williams and Wilkins, Baltimore, p. 49,1982.
Fujimoto, K., Yamada, Y., Okajima, E., Kakizoe, T., Sasaki, H., Sugimura, T., Terada,
M. Frequent association of p53 gene mutation in invasive bladder cancer. Cancer
Res., 52: 1393-1398,1992.
219
Fujimoto, K., Yamada, Y., Okajima, E., Kakizoe, T., Sasaki, H., Sugimura, T., Terada,
M. Frequent association of p53 gene mutation in invasive bladder cancer. Cancer Res.,
52: 1393-1398,1992.
Fujita, J., Srivastava, S., K., Kraus, M., H., Rhim, J., S., Tronick, S., R., Aaronson, S.,
A. Frequency of molecular ^Iterations affecting ras protooncogenes in human urinary
tract tumors. Proc. Nad. Acad. Sci. USA, 82:3849-3853,1985.
Gailani, M., R., Bale, S., J., Leffell, D., J., DiGiovanna, J., J., Peck, G., L., Poliak, S.,
Drum, M., A., Pastakia, B., McBride, O., W., Kase, R., Greene, M., Mulvihill, J., J.,
Bale, A., E. Developmental defects in Gorlin syndrome related to a putative tumor
suppressor gene on chromosome 9. Cell, 69: 111-117,1992.
Gardiner-Garden, M., Frommer, M. CpG islands in vertebrate genomes. J. Mol. Biol.,
196: 261-282, 1987.
German, J., Bloom, D., Passarge, E. Bloom’s Syndrome. XII. Report from the
Registry for 1987. Clin. Genet., 35: 57-69,1989.
Gibas, Z., Prout, G., R., Connelly, J., G., Ponds, J., E., Sandberg, A., A. Non random
chromosomal changes in transitional cell carcinoma of the bladder. Cancer Res.,
44:1257, 1984.
Gibas, Z., Prout, G., R., Pontes, J., E., Sandberg, A., A. A possible specific
chromosome change in transitional cell carcinoma of the bladder. Cancer Genet.
Cytogenet., 19:229-238, 1986.
Glassberg, K., I. Renal dysplasia and cystic disease of the kidney. In: Walsh, P., C.,
Redk, A., B., Stamey, T., A., and Vaughan, E., D. (eds). Campbell’s Oncology Sixth
Edition, pp. 1443-1495. USA: Saunders, 1992.
Gonzalez-Zulueta, M., Ruppert, J., M., Tokino, K., Tsai, Y., C., Spruck, C., H., Miyao,
N., Nichols, P., W., Hermann, G., G., Horn, T., Steven, K., Summerhayes, I., C.,
Sidransky, D., and Jones, P., A. Microsatellite instability in bladder cancer. Cancer
Res., 53: 5620-5623,1993.
Gonzalez-Zulueta, M., Shibata, A., Ohneseit, P., F., Spruck, C., H., Bush, C., Shamaa,
M., Elbaz, M., Nichols, P. W., Malmstrom, P-U., Jones, P. A. High frequency of
chromosome 9p allelic loss and CDKN2 tumor suppressor gene alterations in
squamous cell carcinoma of the bladder. J. Natl. Cancer hist., 87: 1383-1393.
220
Gonzalez-Zulueta, M., Bender, C., M., Yang, A.,S., Nguyen, T., T., Beart, R., W., Van
Tomout, J., M., and Jones, P., A. Methylation of the S' CpG island of the
pl6/CDKN2 tumor suppressor gene in normal and transformed human tissues
correlates with gene silencing. Cancer Res., 55: in press, 1995.
Gonzalez-Zulueta, M., and Jones, P., A. Molecular genetic alterations in transitional
cell Textbook of Genitourinary Oncology. Vogelzang, N.J., Scardino, P.T., Shipley,
W.V., and Coffey, D.S. (eds.) Williams & Wilkins Publications, 1995. (In press).
Gorlin, R. J. Nevoid basal-cell carcinoma syndrome. Medicine, 66: 98-113,1987.
Greenblatt, M., S., Bennet, W., P., Hollstein, M., Harris, C., C. Mutations in the p53
tumor suppresor gene: Clues to cancer etiology and molecular pathogenesis. Cancer
Res., 54: 4855-4878,1994.
Greene, L., F., Hanash, K., A., Farrow, K., M. Benign papilloma or papillary
carcinoma of the bladder? J. Urol., 110:205-207, 1973.
Gruis, N., A., Weaver-Feldhaus, J., Liu, Q„ Y., Frye, C., Eeles, R., Orlow, I., Lacombe,
L., Poncecastaneda, V., Lianes, P., Kamb, A. Genetic evidence in melanoma and
bladder cancer that pl6 and p53 function in separate pathways of tumor suppression.
Am. J. Pathol., 148, 1199-1209, 1995.
Habuchi, T., Ogawa, O., Kakehi, Y., Ogura, K., Koshiba, M., Hamazaki, S., Takahashi,
R., Sugiyama, T., Yoshida, O. Accumulated allelic losses in the development of
invasive urothelial cancer. Int. J. Cancer, 53: 579-584,1993.
Habuchi, T., Takahashi, R., Yamada, H., Ogawa, O., Kakehi, Y., Ogura, K., Hamazaki,
S., Toguchida, J., Ishizaki, K., Fujita, J., Sugiyama, T., Yoshida, O. Influence of
cigarette smoking and schistosomiasis on p53 gene mutation in urothelial cancer.
Cancer Res., 53: 3795-3799, 1993.
Hannon, G., J., Beach, D. pl5IN K 4B is a potential effector of TGF-P-induced cell
cycle arrest. Nature, 371: 257-260, 1994.
Hermann, J., G., Latif, F., Weng, W., Lerman, M., I., Zbar, B., Liu, S., Samid, D.,
Duan, D.-S., Gnarra, J., R., Linehan, W., M., Baylin.S. B. Silencing of the VHL tumor
suppressor gene by DNA methylation in renal carcinoma. Proc. Natl. Acad. Sci. USA,
91: 9700-9704, 1994.
Hermann, J., G., Merlo, A., Mao, L., Lapidus, R.,G., Issa, J-P., J., Davidson, N., E.,
Sidransky, D., and Baylin, S.B. Inactivation of the CDKN2/pl6/MTS 1 gene is
221
frequently associated with aberrant DNA methylation in all common human cancers.
Cancer Res., 55: in press, 1995.
Hicks, R., M. Nitrosamines as possible etiological agents in bilharzial bladder cancer.
In: Magee, P., N. (ed.) Nitrosamines and human cancer (Banbury report number 12).
pp.445-471. New York: Cold Spring Harbor Lab. Press, 1982.
Hirata, Y., Orth, D., N. Epidermal growth factor (urogastrone) in human fluids: size
heterogeneity. J. Clin. Endocrimol. Metab., 48: 673-678,1979.
Holsti, L. R., Ermala, P. Papillary carcinoma of the bladder in mice obtained after
peroral administration of tobacco tar. Cancer, 8: 679-682,1955.
Hopman, A., H., N., Moesken, O., Smeets, A., W., G., B., Pairwels, R., P., E., Vooijs,
P., Ramaekers, F., C., S. Numerical chromosome 1,7,9, and 11 aberrations in bladder
cancer detected by in situ hybridization. Cancer Res., 51: 644-651,1991.
Hopman, A., H., N., Raap, A., K., Landegent, J., E., Wiegant, J., Boerman, R., H., Van
Der Ploeg M. Non-radioactive in situ hybridization. In: Van Leeuwen, Buijs, Pool and
Pach (eds.) Molecular neuroanatomy. Amsterdam, Elsevier Science Publishers BV,
43,1988.
Hopman, A., H., N., Ramaekers, F., C., S., Vooijs, P. Interphase cytogenetics of solid
tumors. In: Polak, J., M., McGee, J., O., D. (eds.) In situ hibridization principles and
practices. Oxford University Press, Oxford, 1990.
Hussussian, C., J., Struewing, J., P., Goldstein, A., M., Higgins, P., A., Ally, D., S.,
Sheahan, M., D., Clark, W., H., Jr., Tucker, M., A.,. Dracopoli, N., C. Germline pl6
mutations in familial melanoma. Nature Genet., 8: 15-21,1994.
Ionov, Y., Peinado, M., A., Malkhosyan, S., Shibata, D., Perucho, M. Ubiquitous
somatic mutations in simple repeated sequences reveal a new mechanism for colonic
carcinogenesis. Nature, 363: 558-561,1993.
Ishikawa, J., Xu, H., J., Hu, S., X., Yandell, D., W., Maeda, S., Kamidono, S., Benedict,
W., F., Takahashi, R. Inactivation of the Retinoblastoma gene in human bladder and
renal cell carcinomas. Cancer Res., 51: 5736-5743,1991.
Issa, J.-P., J., Ottaviano, Y., L., Celano, P., Hamilton, S., R., Davidson, N., E., Baylin,
S., B. Methylation of the estrogen receptor CpG island links ageing and neoplasia in
human colon. Nature Genetics, 7: 536-540,1994.
222
Jahner, D., Jaenisch, R. Retrovirus-induced de novo methylation of flanking host
sequences correlates with gene inactivity. Nature, 315: 594-597,1985.
Jones, P,. A., Droller, M. J. Pathways of development and progression in bladder
cancer: New correlations between clinical observations and molecular mechanisms.
Semin. Urol., XL 177-192,1993.
Jones, P., A., Buckley, J., D. The role of DNA methylation in cancer. Adv. Cancer
Res., 54: 1-23,1990.
Jones, P., A., Rideout, W., M., Shen, J.-C., Spruck, C,. H., Tsai, Y., C. Methylation,
mutation and cancer. BioEssays, 14: 33-36,1992.
Jones, P., A., Taylor, S. M. Cellular differentiation, cytidine analogs and DNA
methylation. Cell, 20: 85-93,1980.
Kamb, A., Gruis, N., A., Weaver-Feldhaus, J., Liu, Q., Harshman, K., Tavtigian, S., V.,
Stocket, E., Day, R., S. IE, Johnson, B., E., Skolnick, M., H. A cell cycle regulator
potentially involved in genesis of many tumor types. Science, 264:436-439,1994.
Kamb, A., Shattuck-Eidens, D., Eeles, R., Liu, Q., Gruis, N. A., Ding, W., Huusey, C.,
Tran, T., Miki, Y., Weaver-Feldhaus, J., McClure, M., Aitken, J. F., Anderson, D. E.,
Bergmann, W., Frants, R., Goldar, D. E., Green, A., MacLennan, R., Martin, N. G.,
Meyer, L. J., Youl, P., Zone, J. J., Skolnick, M. H., and Cannon-Albright, L. A.
Analysis of the p i6 gene (CDKN2) as a candidate for the chromosome 9p melanoma
susceptibility locus. Nature Genet., 8: 22-26,1994.
Kaneko, Y., Egues, M., C., Rowley, J., D. Interstitial deletion of short arm of
chromosome 11 limited to Wilms' tumor cells in patients without aniridia. Cancer
Res., 41: 4577-4578, 1981.
Kaye, K., W., Lange, P., H. Mode of presentation of invasive bladder cancer:
Reassessment of the problem. J. Urol., 128: 31,1982.
Keen, A. J., Knowles, M., A. Definition of two regions of deletion on chromosome 9
in carcinoma of the bladder. Oncogene, 9:2083-2088,1994.
Knowles, M., A., Elder, P., A., Williamson, M., Caims, J., P., Shaw, M., E., Law, M.,
G. Allelotype of human bladder cancer. Cancer Res., 54: 531-538,1994.
223
Knowles, M., A., Shaw, M., E., Proctor, A., J. Deletion mapping of chromosome 8 in
cancers of the urinary bladder using restriction fragment length polymorphisms and
microsatellite polymorphisms. Oncogene, 8: 1357-1364,1993.
Knudson, A., G. All in the family. Nature Genet., 5:103-104,1993.
Knudson, A.G. Mutation and cancer: statistical study of retinoblastoma. Proc. Natl.
Acad. Sci. USA, 68: 820-823,1971.
Kotake, T., Saiki, S., Kinouchi, T., Shiku, H., Nakayama, E. Detection of the c-myc
gene product in urinary bladder cancer. Jpn. J. Cancer Res., 81:1198-1201,1990.
Kremer, E., J., Pritchard, M., Lynch, M., Yu, S., Holman, K., Baker, E., Warren, S.T.,
Schlessinger, D., Sutherland, G.R., and Richards, R.I. Mapping of DNA instability at
the fragile X to a trinucleotide repeat sequence p(CCG)n. Science, 252: 1711-1714,
1991.
Kuhlmann, I., Doerfler, W. Loss of viral genomes from hamster tumor cells and
nonrandom alterations in patterns of methylation of integrated adenovirus type 12
DNA. J. Virol., 47:631-636,1983.
Kwiatkowski, D., J., Armour, J., Bale, A., E., Fountain, J., W., Goudie, D., Haines, J.,
L., Knowles, M., A., Pilz, A., Slaugenhaupt, S., Povey, S. Report on the second
international workshop on human chromosome 9. Cytogenet. Cell. Genet., 64: 94-
106, 1993.
Kwiatkowski, D., J., Henske, E., P., Weimer, K., Ozelius, L., Gusella, J., F., Haines, J.
Construction of a GT polymorphism map of human 9q. Genomics, 12: 229-240,
1992.
Laird, P., W., Jaenisch, R. DNA methylation and cancer. Human Mol. Genet., 3:
1487-1495,1994.
Leach, F., S., Nicolaides, N., C., Papadopoulos, N., Liu, B., Vogelstein, B. Mutations
of a mutS homolog in hereditary nonpolyposis colorectal cancer. Cell, 75:1215-1225,
1993.
Levine, A., J., Momand, J., Finlay, C., A. The p53 tumor suppressor gene. Nature, 351:
453-456,1991.
Li, E., Beard, C., Jaenisch, R. Role for DNA methylation in genomic imprinting.
Nature, 366: 362-365,1993.
224
Li, E., Bestor, T., H., Jaenisch, R. Targeted mutation of the DNA methyltransferase
gene results in embryonic lethality. Cell, 69:915-926,1992.
Li, Y., Nichols, M., A., Shay, J., W., Xiong, Y. Transcriptional repression of the D-type
cyclin-dependent kinase inhibitor pl6 by the retinoblastoma susceptibility gene
product pRb. Cancer Res., 54: 6078-6082,1994.
Lieskovsky, G., Ahlering, T., Skinner, D., G. Diagnosis and staging of bladder
cancer. In: Diagnosis and Management of Genitourinary Cancer, Lieskovsky, G. and
Skinner, D.G., (eds.), W.B. Saunders Co., Philadelphia, PA, pp. 262-280,1988.
Lilienfeld, A., J., M., L., Moore, G., E. The association of smoking with cancer of the
urinary bladder in humans. Arch. Intern. Med., 98: 129-135,1956.
Linnebach, A., J., Pressler, L., B., Seng, B., A., Kimmel, B., S., Tomaszewski, J., E.,
Malkowicz, S., B. Characterization of chromosome 9 deletions in transitional cell
carcinoma by microsatellite assay. Hum. Molec. Genetics, 2: 1407-1411, 1993.
Linnebach, A., J., Robbins, S., L., Seng, B., A., Tomaszewski, J., E., Pressier, L., B.,
Malkowicz, S., B. Urothelial carcinogenesis. Nature, 367: 419-420,1994.
Lipponen, P., K. Over-expression of p53 nuclear oncoprotein in transitional-cell
bladder cancer and its prognostic value. Int. J. Cancer, 53:365-370,1993.
Litt, M., Luty, J.A. A hypervariable microsatellite revealed by in vitro amplification of
a dinucleotide repeat within the cardiac muscle actin gene. Am. J. Hum. Genet., 44:
397-401,1989.
Liu, W.-M., Schmid, C., W. Proposed roles for DNA methylation in Alu
transcriptional expression and mutational inactivation. Nucleic Acids Res., 21: 1351-
1359, 1993.
Loeb, L., A. Microsatellite instability: Marker of a mutator phenotype in cancer.
Cancer Res., 54: 5059-5063,1994.
Logothetis, C., J., Xu, H., J., Ro, J., Y., Hu, S., X., Sahin, A., Orodonez, N., Benedict,
W., F. Altered expression of Retinoblastoma protein and known prognostic variable in
locally advanced bladder cancer. J. Nat. Cancer Inst., 84: 1256-1261,1992.
Lukas, J., Parry, D., Aagaard, L., Mann, D., Bartkova, J., Strauss, M., Peters, G.,
Bartek, J. Retinoblastoma-protein-dependent cell-cycle inhibition by the tumor
suppressor p i6. Nature, 375: 503-506,1995.
225
Lukeis, R., Irving, L., Garson, M., Hasthorpe, S. Cytogenetics of non-small cell lung
cancer: analysis of consistent ono-random abnormalities. Genes Chrom. Cancer, 2:
116-124,1990.
Macleod, D., Charlton, J., Bird, A., P. Spl sites in the mouse aprt gene promoter are
required to prevent methylafion of the CpG island. Genes and Development, 8: 2282-
2292,1994.
Magewu, A., N. Ph.D. thesis. University of Southern California, Los Angeles, 1994.
Mahadevan, M., Tsilfidis, C., Sabourin, L., Shutler, G., Amemiya, C., Jausen, G.,
Neville, C., Narang, M., and Barcelo, J. Myotonic Dystrophy mutation: an unstable
CTG repeat in the 3’ untranslated region of the gene. Science, 255:1253-1255,1992.
Marmur, J. A. Procedure for the isolation of deoxyribonucleic acid from
microorganisms. J. Mol. Biol., 3: 208-218,1961.
Matanoski, G., M., Elliot, E., A. Bladder cancer epidemiology. Epidemiology Rev., 3:
203,1981.
Matsushime, H., Ewen, M., E., Strom, D., K., Kato, J., Y., Hanks, S., K., Roussel, M.,
F., Sherr, C., J. Identification and properties of an atypical catalytic subunit (p34PSK-
J3/cdk4) for mammalian D type G1 cyclins. Cell, 71: 323-334,1992.
McDonald, M., T., Papenberg, K., A., Ghosh, S., Glatfelter, A., A., Biesecker, B., B.,
Helmbold, E., A., Markel, D., S., Zolotor, A., McKinnon, W., C., Vanderstoep, J., L.,
Jackson, C., E., Iannuzzi, M., Collins, F., S., Boehnke, M., Porteous, M., E.,
Guttmacher, A., E., Marchuk, D., A. A disease locus for hereditary haemorrhagic
telangiectasia maps to chromosome 9q33-34. Nature Genet., 6: 107-204,1994.
Melicow, M., M. Histological study of vesical urothelium intervening between gross
neoplasm in total cystectomy. J. Urol., 68: 261,1952.
Merlo, A., Gabrielson, E., Askin, F., and Sidransky, D. Frequent loss of chromosome
9 in human primary non-small cell lung cancer. Cancer Res., 54: 640-642, 1994.
Merlo, A., Gabrielson, E., Mabry, M., Baylin, S., B., Sidransky, D. Homozygous
deletion on chromosome 9p and loss of heterozygosity on 9q, 6p, and 6q in primary
human small cell lung cancer. Cancer Res., 54: 2322-2326,1994.
Messing, E., M. Clinical implications of the expression of epidermal growth factor
receptors in human transitional cell carcinoma. Cancer Res., 50:2530-2537,1990.
226
Messing, E., M., Hanson, P., Ulrich, P., Erturk, E. Epidermal growth factor
interactions with normal and malignant urothelium: in vivo in situ studies. J. Urol.,
138: 1329-1335, 1987.
Meyers, F., J., Gumerlock, P., H., Kokoris, S., P, de Vere-White, R., W., McCormick,
F. Human bladder and colop carcinomas contain activated ras p21: Specific detection
of twelfth codon mutants. Cancer, 63: 2177-2181,1989.
Miyao, N., Tsai, Y., C., Lemer, S., P., Olumi, A., F., Spruck, C., H., Gonzalez-Zulueta,
M., Nichols, P., W., Skinner, D., G., Jones, P., A. Role of chromosome 9 in human
bladder cancer. Cancer Res., 53:4066-4070,1993.
Modrich, P. Mismatch repair, genetic stability, and cancer. Science, 266: 1959-1956,
1994.
Mori, N., Yokota, J., Oshimura, M., Cavenee, W., K., Mizoguchi, H., Noguchi, M.,
Shimosato, Y., Sugimura, T., Terada, M. Concordant deletions of chromosome 3p and
loss of heterozygosity for chromosomes 13 and 17 in small cell lung carcinoma.
Cancer Res., 49:5130-5135,1989.
Mori, T., Miura, K., Aoki, T., Nishihira, T., Mori, S., Nakamura, Y. Frequent somatic
mutation of the MTS1/CDK4I (Multiple Tumor Suppressor/Cyclin-dependent Kinase
4 Inhibitor) gene in esophageal squamous cell carcinoma. Cancer Res., 54: 3396-
3397, 1994.
Mostofi, F., K., Davis, C., J., Sesterhenn, I., A. Pathology of tumors of the urinary
tract. In: Skinner D., G., Lieskovsky, G. (eds.): Diagnosis and Management of
Genitourinary Cancer. Philadelphia, W.B. Saunders, pp 83-117,1988.
Mostofi, F., K., Sobin, L., H., Torloni, H. Histologic typing of urinary bladder tumors.
International histological classification of tumors, Vol. 10. Geneva: The World Health
Organization, 1973.
Mummaneni, P., Bishop, L., P., Turker, M., S. A cis-acting element accounts for a
conserved methylation pattern upstream of the mouse adenine
phosphoribosyltransferase gene. J. Biol. Chem., 268: 552-558,1993.
Mummaneni, P., Walker, K.,A., Bishop, P., L., Turker, M., S. Epigenetic gene
inactivation induced by a cis-acting methylation center. J. Biol. Chem. 270: 788-792,
1995.
227
Nasmyth, K. Control of the yeast cell cycle by the Cdc28 protein kinase. Curr. Opin.
Cell Biol., 5: 166-179,1993.
Nelson, D., L., Warren, S., T. Trinucleotide repeat instability: when and where? Nature
Genetics, 4: 107-108,1993.
Nobori, T., Miura, K., Wu, D., J., Lois, A., Takabayashi, K., Carson, D., A. Deletions
of the cycle-dependent kinase-4 inhibitor gene in multiple human cancers. Nature,
368: 753-756,1994.
Nyce, J., Liu, L., Jones, P., A. Variable effects of DNA-synthesis inhibitors upon
DNA methylation in mammalian cells. Nucleic Acids Res., 14: 4357-4367., 1986
O’ Connor, P., J., Saffhill, R., Margison, G., P. N-nitroso compounds: biochemical
mechanisms of action. In: Emmelog, P., Kriek, E. (eds.) Environmental
carcinogenesis, occurrence, risk evaluation and mechanisms, pp.73-96.
Amsterdam:Elsevier, 1979.
Ohta, M., Nagai, H., Shimizu, M., Rasio, D., Berd, D., Mastrangelo, M., Singh, A., D.,
Shields, J., A., Shields, C., L., Croce, C., M. Rarity of somatic and germ line mutations
of the cyclin-dependent kinase 4 inhibitor gene, CDK4I, in melanoma. Cancer Res.,
54: 5269-5272, 1994.
Okamoto, A., Demetrick, D.J., Spillare, E.A., Hagiwara, K., Hussain, S.P., Bennet,
W.P., Forrester, K., Gerwin, B., Serrano, M., Beach, D.H., and Harris, C.C. Mutations
and altered expression of pl6IN K 4 in human cancer. Proc. Natl. Acad. Sci. USA, 91:
11045-11049,1994.
Okamoto, A., Hussain, S., P., Hagiwara, K., Spillare-E-A. Rusin-M-R. Demetrick, D.,
J., Serrano, M., Hannon, G., J., Shiseki, M., Zariwala, M., Harris, C., C. Mutations of
the pl6/INK4/MTSl/CDKN2, pl5/MTS2, and pl8 genes in primary and metastatic
lung cancer. Cancer Res 55: 1448-1451,1995.
Oliphant, A., R., Wright, E., C., Swensen, J., Grvis, N., A., Goldar, D., Skolnick, M., H.
Dinucleotide repeat polymorphism at the D17S513 locus. Nucleic Acids Res., 19:
4794, 1992.
Olopade, O., I., Bohlander, S., K., Pomykala, H., Maltepe, E., Van Melle, E., Le Beau
M., M., Diaz, M., O. Mapping of the shortest region of overlap of deletions of the
short arm of chromosome 9 associated with human neoplasia. Genomics, 14: 431-
443,1992.
228
Olopade, O., I., Jenkins, R., B., Ransom, D., T., Malik, K., Pomykala, H., Nobori, T.,
Cowan, J., M., Rowley, J., D., Diaz, M., O. Molecular analysis of deletions of the
short arm of chromosome 9 in human gliomas. Cancer Res., 52: 2523-2529, 1992.
Olopade, O., I., Buchhagen, D., L., Malik, K., Sherman, J., Nobori, T., Bader, S., Nau,
M., M., Gazdar, A., F., Minna, J., D., Diaz, M., O. Homozygous loss of the interferon
gene defines the critical region on 9p that is deleted in lung cancer. Cancer Res., 53:
2410-2415,1993.
Olumi, A., F., Skinner, E., C., Tsai, Y., C., Jones, P., A. Molecular analysis of human
bladder cancer. Semin, in Urol., 8: 270-277,1990.
Olumi, A., F., Tsai, Y., C., Nichols, P., W., Skinner, D., G., Cain, D., R., Bender, L., I.,
Jones, P., A. Allelic loss of chromosome 17p distinguishes high grade from low
grade transitional cell carcinomas of the bladder. Cancer Res., 50: 7081-7083,1990.
Ooi, A., Herz, F., Li, S., Cordon-Cardo, C., Fradet, Y., Mayall, B., H., NIH Maker
Network for Bladder Cancer. Ha-ras codon 12 mutation in papillary tumors of the
urinary bladder. A retrospective study. Int. J. Oncol., 4: 85-90,1994.
Orita, M., Suzuki, Y., Sekiya, T., Hayashi, K. Rapid and sensitive detection of point
mutations and DNA polymorphisms using the polymerase chain reaction. Genomics,
5: 874-879,1989.
Orlow, I., Lianes, P., Lacombe, L., Dalbagni, G., Reuter, V. E., and Cordon-Cardo, C.
Chromosome 9 allelic losses and microsatellite alterations in human bladder tumors.
Cancer Res., 54: 2848-2851,1994.
Orr, H., T., Chung, M., Banfi, S., Kwiatkowski Jr., T.J., Servadio, A., Beaudet, A., L.,
McCall, A., E., Duvick, L., A., Ranum, L., P., W., Zoghbi, H., Y. Expansion of an
unstable trinucleotide CAG repeat in spino cerebellar ataxia type 1. Nature Genetics,
4: 221-226,1993.
Osbome, R., J., Leech, V. Polymerase chain reaction allelotyping of human ovarian
cancer. Br. J. Cancer, 69:429-438,1994.
Ottaviano, Y., L., Issa, J.-P., Pari, F., F., Smith, H., S., Baylin, S., B., Davidson, N., E.
Methylation of the estrogen receptor gene CpG island marks loss of estrogen receptor
expression in human breast cancer cells. Cancer Res., 54: 2552-2555,1994.
229
Otterson, G., A., Kratzke, R., A., Coxon, A., Kim, Y., W., Kaye, F., J. Absence of
p!6lNK4 protein is restricted to the subset of lung cancer lines that retains wildtype
RB. Oncogene, 9,3375-3378,1994.
Palombo, F., Gallinari, P., Iaccarino, I., Lettieri, T., Hughes, M., Truong, O., Hsuan J.,
J., Jiricny J. GTBP, a 160-kilodalton protein essential for mismatch binding activity in
human cells. Science, 268:1912-1914,1995.
Papadopoulos, N., Nicolaides, N., C., Wei, Y.-F., Ruben, S., M., Carter, K., C., Rosen,
C., A., Haseltine, W., A., Fleischmann, R., D., Fraser, C., M., Adams, M., D., Venter, J.,
C., Hamilton, S., R., Petersen, G., M., Watson, P., Lynch, H., T., Peltomaki, P.,Watson,
P., de la Chapelle, A., Kinzler, K., W., Vogelstein, B. Mutation of a mutL homolog in
hereditary colon cancer. Science, 263,1625-1629,1994.
Papadopoulos, N., Nicholaides, N., C., Liu, B., Parsons, R., Willson, J., K., Kinzler,
K., W., Jiricny, J., Vogelstein, B. Mutations of GTBP in genetically unstable cells.
Science, 268: 1915-1917.
Peltomaki, P., Aaltonen, L., A., Sistonen, P, Pylkkanen, L., Mecklin, J., P., Jarvinen, H.,
Green, J., S., Jass, J., R., Weber, J., L., Leach, F., S., Vogelstein, B:, de la Chapelle, A.
Genetic mapping of a locus predisposing to human colorectal cancer. Science, 260:
810-812,1993.
Perucca, D., Szepetowske, P., Simon, M., P., Gaudray, P. Molecular genetics of
human bladder carcinomas. Cancer Genet. Cytogenet., 49: 143-156,1990.
Petersen, I., Ohgali, H., Ludeke, B., I., Kleihues, P. p53 mutations in phenacetin-
associated human urothelial carcinomas. Carcinogenesis, 14: 2119-2122,1993.
Petty, E., M., Gibson, L., H., Fountain, J., W., Bolognia, J., L., Yang-Feng, T., L.,
Housman, D., E., Bale, A., E. Molecular definition of a chromosome 9p21 germ-line
deletion in a woman with multiple melanomas and a plexiform neurofibroma:
implications for 9p tumor-supressor gene(s). Am. J. Hum. Genet., 53:96-104,1993.
Polyak, K., Lee, M.H., Koff, A., Roberts, J.M., Tempst, P., and Massague, J. Cloning
of p27, a cyclin-dependent kinase inhibitor and a potential mediator of extracellular
antimitogenic signals. Cell, 78: 59-66,1994.
Povey, S., Armour, J., Famdon, P., Haines, J. L., Knowles, M., Olopade, F., Pilz, A.,
White, J. A., Members of the Utah Genome Center Genetic Marker and Mapping
Group, and Kwiatkowski, D. J. Report on the Third International Workshop on
Chromosome 9 held at Queens’ College, Cambridge, UK. Ann. Hum. Genet., 58:
177-250,1994.
230
Presti Jr., J. C., Reuter, V. E., Galan, T., Fair, W. R., and Cordon-Cardo, C. Molecular
genetic alterations in superficial and locally advanced human bladder cancer. Cancer
Res., 51: 5405-5409,1991.
Prout, G., R. Jr. Bladder carcinoma and a TNM system of classification. J. Urol., 117:
583-590,1977.
Reed, J., A., Loganzo, F., Shea, C.R., Walker, G., J., Flores, J., F., Glendening, M.,
Bogdani, J., Shiel, M., J., Haluska, F., G., Fountain, J., W., and Albino, A., P. Loss of
expression of the pl6/cyclin-dependent kinase inhibitor 2 tumor suppressor gene in
melanocytic lesions correlates with invasive stage of tumor progression. Cancer Res.,
55: 2713-2718,1995.
Rideout, W., M., Eversole-Cire, P., Spruck, C., H., Hustad, C., M., Coetzee, G., A.,
Gonzales, F., Jones, P., A. Progressive increases in the methylation status and
heterochromatinization of the myoD CpG island during oncogenic transformation.
Mol. Cell. Biol., 14: 6143-6152, 1994.
Riggs, A., D. X-inactivation, differentiation, and DNA methylation. Cytogenet. Cell
Genet., 14: 9-25,1975.
Ross, R., K., Paganini-Hill, A., Herderson, B., E. Epidemioloy of bladder cancer. In:
Diagnosis and Management of Genitourinary Cancer. Skinner, D., G., Lieskovsky,
G., (eds.), W.B. Saunders Company, Philadelphia, PA, pp. 23-31,1988.
Ruppert, J., M., Tokino, K., Sidransky, D. Evidence for two bladder cancer suppressor
loci on human chromosome 9. Cancer Res., 53: 5093-5095,1993.
Saffhill, R., Margison, G., P., O'Connor, P., J. Mechanisms of carcinogenesis induced
by alkylating agents. Biochem. Biophys. Acta, 823: 111-145,1985.
Sambrook, J., Fritsch, E., F., Maniatis, T. Molecular Cloning: A Laboratory Manual,
Cold Spring Harbor Laboratory Press, New York, 1989.
Sandberg, A., A. Chromosome changes in bladder cancer: Clinical and other
correlations. Cancer Genet. Cytogenet., 19:163-175,1986.
Sandberg, A., A., Berger, C., S. Review of chromosome studies in urological tumors
II. Cytogenetics and molecular genetics of bladder cancer. J. Urol., 151: 545-560,
1994.
2 3 1
Sarkis, A., S., Dalbagni, G., Cordon-Cardo, C., Zhang, Z., F., Sheinfeld, J., Fair, W.,
R., Herr, H., W., Reuter, V., E. Nuclear overexpression of p53 protein in transitional
cell bladder carcinoma: a marker for disease progression. J. Nat. Cancer Inst., 85:53-
59,1993.
Sato, S., Nakamura, Y., Tsuchiya, E. Difference of allelotype between squamous cell
carcinoma and adenocarcinoma of the lung. Cancer Res., 54: 5652-5655,1994.
Scrable, H., Witte, D., Shimada, H, Seemayer, T. Sheng, W., W., Soukup, S., Koufos,
A., Houghton, P., Lampkin, B., Cavenee, W. Molecular differential pathology of
rhabdomyosarcoma. Gene. Chromo. Cancer, 1:23-35,1985.
Selker, E., U., Jensen, B., C., Richardson, G., A. A portable signal causing faithful
DNA methylation de novo in Neurospora crassa. Science, 238: 48-53, 1987.
Serrano, M., Hannon, G., J., Beach, D. A. new regulatory motif in cell-cycle control
causing specific inhibition of cyclin D/CDK4. Nature, 366: 704-707, 1993.
Shanley, S., M., Dawkins, H., Wainwright, B., J., Wicking, C., Heenan, P., Eldon, M.,
Searle, J., Chenevix-Trench, G. Fine deletion mapping on the long arm of
chromosome 9 in sporadic and familial basal cell carcinomas. Hum. Mol. Genet., 4:
129-133,1995.
Shapiro, G., I., Edwards, C., D., Kobzik, L., Godleski, J., Richards, W., Sugarbaker,
D., J., Rollins, B., J. Reciprocal Rb inactivation and pl6IN K 4 expression in primary
lung cancers and cell lines. Cancer Res., 55: 505-509,1995.
Sherr, C., J. Mammalian G1 cyclins. Cell, 73: 1059-1065,1993.
Shibata, A., Ohneseit, P., F., Tsai, Y., C., Sprack, C., H., Nichols, P., W., Chiang, H.,
S., Lai, M., K., Jones, P., A. Mutational spectrum of the p53 gene in bladder tumors
from the endemic area of black foot disease in Taiwan. Carcinogenesis, 15: 1085-
1087,1994.
Shibata, D., Hawes, D., Li, Z., H., Hernandez, A., R., Spruck, C., H., Nicholls, P., W.
Specific genetic analysis of microscopic tissue after selective ultraviolet radiation
fractionation and the polymerase chain reaction. Am. J. Path., 141: 539-543,1992.
Sidransky, D., Frost, P., Von Eschenbach, A., Oyasu, R., Preisinger, A., C., Vogelstein,
B. Clonal origin of bladder cancer. N. Engl. J. Med., 326: 737-740, 1992.
232
Sidransky, D., Von Eschenbach, A., Tsai, Y., C., Jones, P., A., Summerhayes, I.,
Marshall, F., Paul, M., Green, P., Hamilton, S., R., Frost, P., Vogelstein, B.
Identification of p53 gene mutations in bladder cancers and urine samples. Science,
252,706-709,1991.
Silverberg, E., Boring, C., G, Squires, T., S. Cancer statistics. CA, 39: 1-26,1989.
Silverberg, E., Boring, C., C., Squires, T., S. Cancer statistics. CA, 40: 1,1990.
Singer-Sam, J., Riggs A., D. X chromosome inactivation and DNA methylation. EXS,
64: 358-384,1993.
Singer-Sam, J., Yang, T., P., Mori, N. Tanguay, R., L., Le Bon, J., M., Flores, J., C.,
Riggs, A., D. DNA methylation in the 5' region of the mouse PGK-1 gene and a
quantitative PCR assay for methylation, p. 285-289. In: G. A. Clawson, D. B. Willis,
A. Weissbach, and P. A. Jones (eds.), Nucleic acid methylation, vol. 128. Alan R.
Liss, Inc, New York, 1990.
Smeets, W., Pauwells, R., Laarakkers, L., Debruyne, F., Geraedts, J. Chromosomal
analysis of bladder cancer III. Nonrandom alterations. Cancer Genet. Cytogenet.,
29:29-41, 1987.
Solomon, E., Borrow, J., Goddard, A. Chromosome aberrations and cancer. Science,
254: 1153-1160, 1991.
Spruck, C., H., Ohneseit, P., F., Gonzalez-Zulueta, M., Esrig, D., Miyao, N., Tsai, Y.,
C., Lemer, S. P., Schmutte, C., Yang, A., S., Cote, R., Dubeau, L., Nichols, P., W.,
Hermann, G., G., Steven, K., Horn, T., Skinner, D., G., and Jones, P., A. Two
molecular pathways to transitional cell carcinoma of the bladder. Cancer Res., 54:
784-788, 1994.
Spruck, C., H., Gonzalez-Zulueta, M., Shibata, A., Simoneau, A., R., Lin, M.-F.,
Gonzales, F., Tsai, Y., C., Jones, P., A. pl6 gene in uncultured tumors. Nature, 370:
183-184,1994.
Spruck, C., H., Rideout, W., R., Olumi, A., F., Ohneseit, P., F., Yang, A., S., Tsai, Y.,
C., Nichols, P., W., Horn, T., Hermann, G., G., Steven, K., Ross, R., K., Yu, M., C.,
Jones, P., A. Distinct pattern of p53 mutations in bladder cancer: relationship to
tobacco usage. Cancer Res., 53: 1162-1166,1993.
233
Stadler, W., M., Sherman, J., Bohlander, S., K., Roulston, D., Dreyling, M., Rukstalis,
D., Olopade, O., I. Homozygous deletions within chromosomal bands 9p21-22 in
bladder cancer. Cancer Res., 54: 2060-2063,1994.
Stallings, R., L., Tomey, D., C., Hildebrand, C., E., Longmire, J., L., Deaven, L., L., Jett,
J., H., Doggett, N., A., Moysis, R., K. Physical mapping of human chromosomes by
repetitive sequence fingerprinting. Proc. Natl. Acad. Sci. USA, 87: 6218-6222,1990.
Sundaresan, V., Ganly, P., Hasleton, P., Rudd, R., Sinha, G., Bleehen, N., M., Rabbitts,
P. p53 and chromosome 3 abnormalities, characteristic of malignant lung tumors, are
detectable in preinvasive lesions of the bronchus. Oncogene, 7: 1989-1997,1992.
Szyf, M., Schimmer, B., P., Seidman, J., G. Nucleotide sequence specific de novo
methylation in a somatic murine cell line. Proc. Natl. Acad. Sci. USA, 86: 6853-6857,
1986.
Talaska, G., Al-Juburi, A., Z., S., S., and Kadlubar, F., F. Smoking related carcinogen-
DNA 8-yl)-4-aminbiphenyl as a major product. Proc. Natl. Acad. Sci. USA, 88:5350-
5354,1991.
Tam, S., W., Shay, J., W., Pagano, M. Differential expression and cell cycle regulation
of the cyclin-dependent kinase 4 inhibitor pl6IK N *. Cancer Res., 54: 5816-5820,
1994.
The Huntington’s Disease Collaborative Research Group. A novel gene containing a
trinucleotide repeat that is expanded and unstable on Huntington's Disease
chromosomes. Cell, 72: 371-383, 1993.
Thibodeau, S., N., Bren, G., Schaid, D. Microsatellite instability in cancer of the
proximal colon. Science, 260: 816-819,1993.
Thompson, I., M., Fair, W., R. The epidemiology of bladder cancer. AUA Update Ser,
8:210-215, 1989.
Toth, M., Muller, U., Doerfler, W. Establishment of de novo methylation patterns:
transcription factor binding and deoxycytidine methylation at CpG sequences in an
integrated adenovirus promoter. J. Mol. Biol., 214: 673-683, 1990.
Tsai, Y., C., Nichols, P., W., Hiti, A., L., Williams, Z., Skinner, D., G., Jones, P., A.
Allelic losses of chromosome 9,11, and 17 in human bladder cancer. Cancer Res., 50:
44-47,1990.
234
Tsai, Y., C., Simoneau, A., R., Spruck, C., H., Nichols, P., W., Steven, K., Buckley, J.,
D., Jones, P., A. Mosaicism in human epithelium: Macroscopic monoclonal patches
cover the urothelium. J. Urol., 153,1697-1700,1995.
Tsai, Y., C., Nichols, P., W., Hiti, A., L., Williams, Z., Skinner, D., G., Jones, P., A.
Allelic losses of chromosome 9,11, and 17 in human bladder cancer. Cancer Res, 50:
44-47, 1990.
Tyrkus, M., Powell, I., Fakr, W. Cytogenetic studies of carcinoma in situ of the
bladder: pronostic implications. J. Urol., 148:44-46,1992.
Ueki, K., Rubio, M., P., Ramesh, V., Correa, K., M., Rutter, J., L., von Deimling, A.,
Buckler, A., J., Gusella, J., F., Louis, D., N. CDKN2 mutations are rare in primary
human astrocytomas with allelic loss of chromosome 9p. Human Mol. Genet., 3:
1841-1845, 1994.
Utz, D., C., Hanash, K., A., Farrow, G., M. The plight of the patient with carcinoma in
situ of the bladder. J. Urol., 103:160-164,1970.
van der Riet, P., Nawroz, H., Hruban, R., H., Corio, R., Tokino, K., Koch, W.,
Sidransky, D. Frequent loss of chromosome 9p21-22 early in head and neck cancer
progression. Cancer Res., 54: 1156-1158,1994.
Vanni, R., Scarpa, R., M. (Correspondence) Gibas et al. Non random chromosomal
changes in transitional cell carcinoma of the bladder. Cancer Res., ,46:4873,1986.
Vertino, P., M., Issa, J.-P., Pereira-Smith, O., M., Baylin, S., B. Stabilization of DNA
methyltransferase levels and CpG island hypermethylation precede SV40-induced
immortalization of human fibroblasts. Cell Growth Differentiation, 5: 1395-1402,
1994.
Viola, M., V., Fromowitz, F., Oravez, S., Deb, S., Schlom, J. ras Oncogene p21
expression in premalignant lesions and high grade bladder carcinoma. J. Exp. Med.,
161: 1213-1218, 1985.
Vogelstein, B., Kinzler, K. p53 function and dysfunction. Cell, 70: 523-526,1992.
Vogelstein, B., Kinzler, K., W. The multistep nature of cancer. TIG, 9: 138-141,1993.
von Hansemann, D. Uber asymetrische Zellteilung in Epithelkrebsen und deren
biologische Bedeutung. Arch. Pathol. Anat. Physiol., 119: 299-326,1890.
235
Weaver-Feldhaus, J., Gruis, N., A., Neuhausen, S., Le Paslier, D., Stocker, E.,
Skolnick, M., Kamb, A. Localization of a putative tumor suppressor gene by using
homozygous deletions in melanomas. Proc. Natl. Acad. Sci. USA, 91: 7563-7567,
1994.
Weber, J., L. Informativeness of human (dC-dA)n (dG-dT)n polymorphisms.
Genomics, 7:524-530,1990.
Wicking, C., Breen, M., Negus, K., Berkman, J., Evdokiou, A., Cowled, P., Chenevix-
Trench, G., and Wainwright, B. The human growth-arrest-specific gene GAS1 maps
outside the candidate region of the gene for nevoid basal cell carcinoma syndrome.
Cytogenet. Cell Genet., 68: 119-121,1995.
Wright, C., Mellon, K., Johnston, P., Lane, D., P., Harris, A., L., Home, C,. H., W.,
Neal, D., E. Expression of mutant p53, c-erbB-2 and the epidermal growth factor
receptr in transitional cell carcinoma of the human urinary bladder. Br. J. Cancer, 63:
967-970,1991.
Wu, M., M., Kuo, T., L. Hwang, Y., H., Chen, C., J. Dose-response relation between
arsenic concentration in well water and mortality from cancers and vascular diseases.
Am. J. Epidemiol., 130: 1123-1132, 1989.
Xu, H., J., Caims, P., Hu, S., X., Knowles, M., A., Benedict, W., F. Loss of RB protein
expression in primary bladder cancer correlates with loss of heterozygosity at the RB
locus and tumor progression. Int. J. Cancer, 53: 781-784, 1993.
Xu, L., Sgroi, D., Sterner, C., J., Beauchamp, R., L., Pinney, D., M., Keel, S., Ueki, K.,
Rutter, J., L., Buckler, A., J., Louis, D., N., Gusella, J., F., Ramesh, V. Mutational
analysis of CDKN2 (MTSl/pl6,N K 4) in human breast carcinomas. Cancer Res., 54:
5262-5264, 1994.
Xu, L., Sgroi, D., Sterner, C., J., Beauchamp, R., L., Pinney, D., M., Keel, S., Ueki, K.,
Rutter, J., L., Buckler, A., J., Louis, D., N., Gusella, J., F., Ramesh, V. Mutational
analysis of CDKN2 (MTSl/pl6,N K 4) in human breast carcinomas. Cancer Res, 54:
5262-5264, 1994.
Yan, Z., Hsu, S., Winawer, S., and Friedman, E. Transforming growth factor pi
(TGFp-1) inhibits retinoblastoma gene expression but not pRB phosphorylation in
TGFP-l-growth stimulated colon carcinoma cells. Oncogene, 7: 801-805,1992.
Yang, A, Jones, P., A. The molecular progression of bladder cancer. In: Basic research
in urological oncology. Luciani, L. (ed.), Karger A.G., Basel, Switzerland. In press.
236
Yang, R., Gombart, A., F., Serrano, M., Koeffler, H., P. Mutational effects on the
pi 6® ^* tumor suppressor protein. Cancer Res., 55: 2503-2506,1995.
Yeager, T., Stadler, W., Belair, C., Puthenveettil, J., Olopade, O., Reznikoff, C.
Increased pl6 levels correlate with pRb alterations in human urothelial cells. Cancer
Res., 55: 493-497,1995. ,
Yoshida, S., Takeshi, T., Ichikawa, Y., Hanai, S., Suzuki, H., Hori, M., Fukao, K.,
Miwa, M., and Uchida, K. Mutations of pl6IN K 4/CDKN2 and pl5IN K 4B /MTS2 in
genes in biliary tract cancers. Cancer Res., 55: 2756-2760,1995.
Zhang, S.-Y., Klein-Szanto, A., J., P., Sauter, E., R., Shafarenko, M., Mitsunaga, S.,
Nobori, T., Carson, D., A., Ridge, J., A., Goodrow, T., L. Higher frequency of
alterations in the pl6/CDKN2 gene in squamous cell carcinoma cell lines than in
primary tumors of the head and neck. Cancer Res., 54: 5050-5053,1994.
237
APPENDIX
LIST OF ABBREVIATIONS
A Adenine
5-A.za-CdR 5-Aza-2'-deoxycytidine
AR Androgen Receptor
AraC 1-p-D-arabinofuranosylcytosine
bp Base Pair
BCC Basal Cell Carcinoma
C Cytosine
cDNA Complementary DNA
CDK Cyclin-dependent Kinase
CDKN2 Cyclin-dependent Kinase Inhibitor 2
CiS Carcinoma in situ
DMEM Dulbecco's Modified Eagle's Medium
DMSO Dimethyl Sulfoxide
DNA Deoxyribonucleic Acid
EDTA Ethylenediamine Tetraacetic Acid
EGF Epidermal Growth Factor
FISH Fluorescence in situ Hybridization
G Guanosine
G1 Gap Phase-1
GTBP G/T Binding Protein
h Hour(s)
hMLHl Human MutL Homolog 1
hMSH2 Human MutS Homolog 2
HNPCC Hereditary Non-Polyposis Colon Carcinoma
m e - Immunohistochemistry
INF Interferon
KD Kilodalton
LOH Loss of Heterozygocity
M Molar
MEM Minimum Essential Medium Eagle
min Minute(s)
mM Millimolar
mRNA Messenger RNA
MTS1 Multiple Tumorsuppressor 1
MTS2 Multiple Tumorsuppressor 2
M l
Microliter
p l 5 INK4B
15-kilodalton Cyclin-dependent Kinase-4-Inhibitor B
P16IN K 4 16-kilodalton Cyclin-dependent Kinase-4-Inhibitor
P53 53-kilodalton Tumor Suppressor Gene
PCR Polymerase Chain Reaction
pRB Retinoblastoma Protein
RFLP Restriction Fragment Length Polymorphism
RT Reverse Transcription
s e e Squamous Cell Carcinoma
sec Second(s)
SSCP Single Strand Conformation Polymorphism
239
SURF Selective Ultraviolet Radiation Fractionation
T Thymine
TCC Transitional Cell Carcinoma
TF - Transcription Factor
TGF-P Transforming Growth Factor P
U Unit
UV Ultraviolet
VNTR Variable Number Tandem Repeat
W Watts
WBC White Blood Cell
WT Wilms' Tumor
240
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
00001.tif
PDF
00001.tif
PDF
00001.tif
PDF
00001.tif
PDF
00001.tif
PDF
00001.tif
PDF
00001.tif
PDF
00001.tif
PDF
00001.tif
PDF
00001.tif
PDF
00001.tif
PDF
00001.tif
PDF
00001.tif
PDF
00001.tif
PDF
00001.tif
PDF
00001.tif
PDF
00001.tif
PDF
00001.tif
PDF
00001.tif
PDF
00001.tif
Asset Metadata
Core Title
00001.tif
Tag
OAI-PMH Harvest
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-oUC11256642
Unique identifier
UC11256642
Legacy Identifier
9621614